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
|
|---|---|---|---|---|---|
The Drosophila Nonspecific Lethal ( NSL ) complex is a major transcriptional regulator of housekeeping genes . It contains at least seven subunits that are conserved in the human KANSL complex: Nsl1/Wah ( KANSL1 ) , Dgt1/Nsl2 ( KANSL2 ) , Rcd1/Nsl3 ( KANSL3 ) , Rcd5 ( MCRS1 ) , MBD-R2 ( PHF20 ) , Wds ( WDR5 ) and Mof ( MOF/KAT8 ) . Previous studies have shown that Dgt1 , Rcd1 and Rcd5 are implicated in centrosome maintenance . Here , we analyzed the mitotic phenotypes caused by RNAi-mediated depletion of Rcd1 , Rcd5 , MBD-R2 or Wds in greater detail . Depletion of any of these proteins in Drosophila S2 cells led to defects in chromosome segregation . Consistent with these findings , Rcd1 , Rcd5 and MBD-R2 RNAi cells showed reduced levels of both Cid/CENP-A and the kinetochore component Ndc80 . In addition , RNAi against any of the four genes negatively affected centriole duplication . In Wds-depleted cells , the mitotic phenotypes were similar but milder than those observed in Rcd1- , Rcd5- or MBD-R2-deficient cells . RT-qPCR experiments and interrogation of published datasets revealed that transcription of many genes encoding centromere/kinetochore proteins ( e . g . , cid , Mis12 and Nnf1b ) , or involved in centriole duplication ( e . g . , Sas-6 , Sas-4 and asl ) is substantially reduced in Rcd1 , Rcd5 and MBD-R2 RNAi cells , and to a lesser extent in wds RNAi cells . During mitosis , both Rcd1-GFP and Rcd5-GFP accumulate at the centrosomes and the telophase midbody , MBD-R2-GFP is enriched only at the chromosomes , while Wds-GFP accumulates at the centrosomes , the kinetochores , the midbody , and on a specific chromosome region . Collectively , our results suggest that the mitotic phenotypes caused by Rcd1 , Rcd5 , MBD-R2 or Wds depletion are primarily due to reduced transcription of genes involved in kinetochore assembly and centriole duplication . The differences in the subcellular localizations of the NSL components may reflect direct mitotic functions that are difficult to detect at the phenotypic level , because they are masked by the transcription-dependent deficiency of kinetochore and centriolar proteins .
The spindle is a microtubule ( MT ) -based highly dynamic molecular machine that mediates chromosome segregation during cell division . Spindle formation requires many proteins that specifically bind and/or regulate MT assembly and dynamics , and also some proteins that are associated with the chromatin during interphase [1] . These latter proteins dissociate from the chromosomes upon mitotic entry and return to the nucleus during telophase; they have therefore functional roles in both interphase chromatin and in mitotic spindle assembly . Examples of these proteins are the components of the human KAT8-associated nonspecific lethal ( KANSL ) complex , which includes at least seven proteins: KANSL1 , KANSL2 , KANSL3 , MCRS1 , PHF20 , WDR5 and the MOF/KAT8 histone acetyltransferase . The KANSL complex localizes in the nucleus of interphase cells where it regulates transcription of a specific set of genes and contributes to stem cell identity [2 , 3] . Mutations in KANSL1 dominantly induce the Koolen-de Vries syndrome characterized by mental retardation and peculiar facial features , and mutations in KANSL2 have been associated with intellectual disabilities [4–6] . Studies on human cells and Xenopus egg extracts have shown that during mitosis , KANSL1 , KANSL3 and MCRS1 re-localize from the chromatin to the MT minus ends of the mitotic spindle , playing essential roles in spindle assembly and chromosome segregation [7 , 8] . Also WDR5 associates with the spindle during mitosis , and its RNAi-mediated depletion leads to spindle defects and chromosome misalignment [9] . Thus , the subunits of the KANSL complex , which is restricted to the nucleus during interphase , after the nuclear envelope breakdown redistribute to spindle structures where they are thought to play mitotic functions . The Drosophila Nonspecific Lethal ( NSL ) complex is the fly counterpart of the KANSL complex . It includes seven conserved subunits: Nsl1/Wah ( KANSL1 ) , Dgt1/Nsl2 ( KANSL2 ) , Rcd1/Nsl3 ( KANSL3 ) , Rcd5 ( MCRS1 ) , MBD-R2 ( PHF20 ) , Wds ( WDR5 ) and Mof ( MOF/KAT8 ) . MBD-R2 , Rcd5 , Nsl1 , and Mof colocalize in the interbands of polytene chromosomes of third instar larvae , and MBD-R2 physically interacts with the histone modifying complexes Trx/MLL and the Nup98 nucleoporin [10 , 11] . Wds is not only a member of the NSL complex , but is also part of several chromatin complexes including the ATAC histone acetyltransferase and the Trx/MLL histone methyltransferase [9 , 12 , 13] . The NSL complex associates with the promoters of more than 4 , 000 housekeeping genes , indicating that it acts as a major transcriptional regulator [10 , 12 , 14] . As in the case of their human counterparts , depletion of the NSL complex members results in mitotic defects . Genome-wide RNAi screens performed in S2 cells showed that Dgt1 , Rcd1 and Rcd5 control mitotic centrosome behavior . Dgt1 ( diminished γ-tubulin 1 ) is required for γ-tubulin recruitment , Rcd1 ( Reduction in Cnn dots 1 ) for centriole duplication , and Rcd5 ( Reduction in Cnn dots 5 ) for both centriole duplication and pericentriolar material ( PCM ) recruitment at the centrosomes [15 , 16] . However , the mechanisms through which Dgt1 , Rcd1 and Rcd5 regulate centriole duplication and centrosome maturation are currently unknown . Another RNAi-based screen in S2 cells showed that Rcd1 and MBD-R2 are required for mitotic chromosome segregation , but the mechanisms leading to this phenotype are also unknown [17] . Here , we analyzed the mitotic phenotypes caused by Rcd1 , Rcd5 , MBD-R2 or Wds depletion in greater detail . Cells depleted of these proteins showed a common defect in centrosome duplication , confirming the centrosome phenotype described earlier for Rcd1 and Rcd5 RNAi cells [16] . However , we found that Rcd1 , Rcd5 and MBD-R2 RNAi cells , and to a lesser extent wds RNAi cells , also exhibit defect in chromosome alignment and segregation . Accordingly , Rcd1- , Rcd5- and MBD-R2-depleted cells displayed reduced levels of both Cid and Ndc80 . Cid is a centromere-specific histone variant homologous to CENP-A that is required for kinetochore assembly [18] , and Ndc80 is the protein that directly mediates MT attachment to the kinetochore [19] . We generated stable S2 cell lines for the inducible expression of Rcd1-GFP , Rcd5-GFP , MBD-R2-GFP or Wds-GFP . We show that these tagged proteins largely rescue the mitotic phenotype caused by RNAi-mediated depletion of their endogenous counterparts , and localize to the nucleus of interphase cells , as expected for transcription factors . However , during mitosis each protein relocalizes to a specific set of mitotic structures , including the centrosomes , the kinetochores and different regions of the midbody . Our results suggest that the common mitotic phenotypes generated by the depletion of the four NSL components are primarily due to reduced transcription of genes encoding centromere/kinetochore components and genes required for centriole duplication . The different mitotic localizations of Rcd1 , Rcd5 , MBD-R2 and Wds further suggest that these proteins might have acquired some direct mitotic roles .
We first analyzed the centrosomal phenotype of S2 cells exposed to Rcd1 , Rcd5 , MBD-R2 or wds double stranded RNAs ( dsRNAs ) for 5 days . To check for RNAi efficiency we performed Western blotting of cell extracts using antibodies directed to Rcd1 , MBD-R2 or Wds . These assays showed that the antibodies specifically recognize bands of the expected molecular weights ( MWs ) that are strongly reduced after RNAi ( Fig 1A–1C ) . The efficiency of RNAi against Rcd5 was demonstrated by RT-qPCR , which revealed a drastic reduction of the Rcd5 transcript level ( Fig 1D ) . RNAi cells were fixed and stained for DNA , tubulin and the centrosomal marker Spd-2 [20 , 21] . We examined only cells that appeared to have the basic karyotype of the S2 cells ( ~ 12 chromosomes [22] ) and did not consider cells that are clearly “polyploid” that are common in S2 cell cultures . However , several S2 cells with the basic karyotype have either a single centrosome or more than two centrosomes ( Fig 1E ) . To ascertain whether depletion of Rcd1 , Rcd5 , MBD-R2 or Wds affects centrosome structure and duplication , we scored the Spd-2 signals in prometaphase and metaphase cells from 3 independent experiments . In this analysis , we counted cells showing no centrosomes , a single centrosome , two centrosomes , or more than 2 centrosomes . We did not subdivide the cells with more than 2 centrosomes in subclasses ( i . e . cells with different centrosome numbers ) , because multiple centrosomes tend to overlap or cluster at the spindle poles and are therefore difficult to count . In addition , we did not take into account cells showing very small Spd-2 signals , because it was unclear whether they represented examples of PCM fragmentation or were produced by immunostaining artifacts . In Rcd1 , Rcd5 , MBD-R2 and wds RNAi cells , the frequencies of cells showing 0–1 centrosomes were significantly increased compared to controls , although in Wds-depleted cells this increase was more modest that in the other RNAi cells ( Fig 1F ) . The frequency of cells with more than two centrosomes was either significantly decreased ( Rcd1 , Rcd5 , MBD-R2 RNAi cells ) or unchanged ( wds RNAi ) compared to control ( Fig 1F ) . We also measured the fluorescence intensity of Spd-2 signals in prometaphase and metaphase cells showing two centrosomes; we found that in Rcd1 , MBD-R2 and wds RNAi cells the average fluorescence of these signals is not significantly different from controls , while in Rcd5-depleted cells it was slightly but significantly higher ( p < 0 . 01; Wilcoxon Signed-Rank Test ) than in control ( Fig 1G and S1 Data ) . Consistent with these findings , we found that the asters of these RNAi cells are indistinguishable from those of control cells . We note that the results on Rcd1 RNAi cells are fully consistent with those of Dobbelaere et al . [16] , who showed that depletion of Rcd1 reduces the centrosome number without affecting PCM recruitment . Dobbelaere et al . also showed that Rcd5 depletion negatively affects centrosome duplication but also reduces Cnn recruitment at centrosomes . We confirmed that Rcd5 deficiency reduces centrosome number but we found that it leads to a small increase in the fluorescence intensity of centrosome-associated Spd-2 . We do not know the reasons for this apparent discrepancy but it is possible that the centrosomes of Rcd5-depleted cells are defective in Cnn recruitment but have normal or slightly increased ability to associate with Spd-2 . In Drosophila S2 cells , which often contain multiple centrosomes and can divide even in the absence of centrosomes , centrosome loss could arise either through defective centrosome separation during prophase or defective centriole duplication during interphase [16 , 23] . Failure in centrosome separation in prophase would lead to formation a bipolar spindle with two centrosomes ( each containing a pair of centrioles ) at one pole and no centrosome at the other . Because S2 cells can assemble fully functional monastral spindles [24] , a defect in prophase centrosome separation is expected to ultimately lead to an increase in both cells with zero and with more than two centrosomes . In the case of defective centriole duplication during interphase , the expectation is instead an increase in cells with 0–1 centrosomes and a decrease in cells with more than 2 centrosomes . Thus , our results ( Fig 1F ) are consistent with the second alternative , and suggest that depletion of the NSL components primarily impairs centriole duplication . To confirm and extend these results , we counted the number of centrioles in interphase cells depleted of Rcd1 , Rcd5 , MBD-R2 or Wds . As centriole marker we decided to use Asterless ( Asl ) , a protein that is thought to link the centriole wall with the PCM [25–27] . However , preliminary tests with antibodies directed to Asl or to other centriolar components revealed that in our hands they do not stain the centrioles in all interphase cells . We thus constructed an S2 cell line that expresses Asl-GFP under the control of the copper-inducible Metallothionein A ( MtnA ) promoter , and used it in the RNAi experiments against the NSL genes . As Asl overexpression causes centriole overduplication during the S phase [26] , we limited the impact of this event by treating both RNAi and control cells with CuSO4 for only 12 hours before fixation ( the cell cycle of S2 cells lasts approximately 24 hours ) . We stained control and RNAi cells with an anti-GFP antibody and scored them for the number of Asl signals present in interphase cells ( Fig 1H ) . We did not take into account cells with zero signals because they could be cells in which the Asl-GFP expression was not sufficiently strong to mark the centrioles . RNAi-mediated depletion of Rcd1 , Rcd5 , MBD-R2 or Wds resulted in significant increases in interphase cells showing a single Asl signal compared to control . We also found significant decreases in cells with more than 2 signals in Rcd1 , MBD-R2 and wds RNAi cells; the frequency of Rcd5 RNAi cells with multiple centrioles was also reduced compared to control but not significantly ( p = 0 . 07 ) ( Fig 1H and 1I ) . Collectively , these results reinforce the conclusion that the NSL components are required for centriole duplication . An analysis of mitotic division in several independent experiments revealed that Rcd1- , Rcd5- and MBD-R2-depleted cells exhibit very similar phenotypes . They display an approximately ten-fold reduction in the frequency of anaphases compared to controls and very high frequencies of PMLES ( prometaphase-like cells with elongated spindles ) figures ( Fig 2A–2C ) . PMLES ( formerly called pseudo-ana-telophases , abbreviated with PATs [17] ) are peculiar mitotic figures that contain late anaphase/early telophase-like spindles associated with chromosomes that are improperly comprised of both sister chromatids and are usually scattered along the spindles; in some cases PMLES exhibit central spindle-like structures and irregular cytokinetic rings ( Fig 2B and S1 Fig; see also refs . [17 , 28] ) . Notably , several PMLES exhibit arched spindles , which are presumably the consequence of an excessive spindle elongation within the constraints imposed by the plasma membrane . Despite their ana-telophase-like spindle structure , PMLES contain high levels of Cyclin B , which is normally degraded at the beginning of anaphase ( Fig 2A and 2B ) . Thus , PMLES appear to be in a pre-anaphase stage as far as sister chromatid separation and Cyclin B degradation are concerned , but they are nevertheless permissive of typical telophase events , such as central spindle assembly and initiation of cytokinesis . To further characterize the PMLES , we measured the length of their spindle axis and compared it with the length of the other mitotic figures . In Rcd1 , Rcd5 and MBD-R2 RNAi cells the prometaphase and metaphase spindles were morphologically normal but slightly longer ( ~10% ) than their control counterparts ( Fig 2D , S2 Data ) . PMLES spindles of RNAi cells were instead substantially longer than prometaphase/metaphases spindles ( ~ 60% ) and anaphases spindles ( ~ 25% ) ( Fig 2D and 2E ) . The average length of the PMLES spindles was either slightly shorter or similar to the length of the spindles of control telophases ( Fig 2E ) , indicating that most PMLES can attain the maximum spindle elongation that is normally achieved by S2 cells . The slight increase in the prometaphase and metaphase spindle length observed in RNAi cells is likely to reflect the presence of some cells in the initial stages of evolution towards a PMLES configuration . Although Rcd1 , Rcd5 and MBD-R2 RNAi cells exhibit very low anaphase frequencies ( Fig 2C ) , they show relatively high frequencies of mitotic figures with telophase-like spindles , characterized by the presence of a constricted central spindle and decondensed chromosomes at the cells poles; about a third of these cells displayed lagging chromosomes between the cell poles ( Fig 2B and S1D Fig ) . Because the chromosomes in these cells are decondensed , it is not possible to discern whether they contain one or two sister chromatids . Nonetheless , we favor the idea that many of the “telophases” observed in Rcd1- , Rcd5- and MBD-R2-depleted cells are in fact PMLES that managed to progress further through the mitotic process and undergo chromosome decondensation as normally occurs in telophase . Regardless of the nature of these telophase-like cells , it is clear that Rcd1 , Rcd5 and MBD-R2 are all required for sister chromatid separation and chromosome segregation . Cells depleted of Wds displayed a mitotic phenotype qualitatively similar but quantitatively milder than that observed in Rcd1- , Rcd5- or MBD-R2-deficient cells ( Figs 1F and 2C-2E ) . Thus , although the degree of RNAi-mediated depletion of the Wds protein is comparable to that observed for the other NSL proteins ( Fig 1C ) , in wds RNAi cells both the centrosome and chromosome segregation phenotypes are substantially milder than those observed in Rcd1 , Rcd5 or MBD-R2 RNAi cells . Collectively our findings indicate that Rcd1- , Rcd5- and MBD-R2-depleted cells are severely defective in chromosome segregation . Depletion of Wds also perturbed chromosome segregation but caused a relatively mild defect . This chromosome segregation phenotype cannot be ascribed to centrosome defects , as abundant evidence indicates that Drosophila mitotic spindle assembly and functioning does not require the centrosomes [15 , 17 , 29–31] . We have previously observed frequent PMLES in S2 cells depleted of the centromere-specific histone H3 Cid/Cenp-A or the kinetochore components Ndc80 , Nuf2 and Kmn1 [17 , 32] . PMLES have been also observed in S2 cells depleted of Mast/Orbit that has a role in MT-kinetochore attachment [33] , or depleted of the Klp67A kinesin-like protein , which represses MT plus end growth and is required for proper MT binding to kinetochores [32] . More recently , PMLES were observed in S2 cells depleted of the Sf3A2 and Prp31 splicing factors that have direct roles in MT-kinetochore interactions [28] . These results suggest that the PMLES found in Rcd1 , Rcd5 and MBD-R2 RNAi cells are a consequence of a defective kinetochore-MT attachment . To test this possibility , we first analyzed the effects of Rcd1 , Rcd5 , MBD-R2 or Wds depletion on the intracellular concentration of the core centromere component Cid/CenpA that is required for kinetochore assembly [18] , and Ndc80 that is directly responsible for MT-kinetochore attachment [19 , 34–36] . Western blotting analysis showed that in Rcd1- , Rcd5- and MBD-R2-depleted cells Cid is considerably reduced compared to controls ( 32% , 36% and 40% of the control level , respectively ) ; the level of Cid was also reduced in Wds-depleted cells but to a lesser extent ( 78% ) ( Fig 3A , S3 Data ) . In addition , Western blotting on extracts from Rcd1 , Rcd5 , and MBD-R2 RNAi cells showed substantial reductions of Ndc80 compared to mock-treated cells ( 32% , 35% and 52% of the control level , respectively ) , while in wds RNAi cells the Ndc80 level was comparable to that of controls ( Fig 3A , S3 Data ) . To extend the analysis at the subcellular level we focused on Rcd1 , Rcd5 and MBD-R2 RNAi cells , which exhibit strong reductions in the Cid content compared to both control and Wds-depleted cells . The Cid protein is detectable both at the centromeres of mitotic chromosomes and in interphase nuclei; nuclei of control cells exhibit 3–6 foci corresponding to clustered centromeres ( Fig 3B ) . In Rcd1 , Rcd5 and MBD-R2 RNAi cells , the frequencies of interphase nuclei devoid of Cid signals were significantly higher than in controls ( Fig 3B and 3C ) . We also examined prometaphases and metaphases with a basic karyotype for the presence of Cid signals . In general , RNAi cells displayed Cid signals of lower intensity compared to controls; they also showed a great variability in the number of detectable signals , ranging from zero to more than 20 signals ( in a cell with 12 chromosomes the maximum number of Cid signals is 24 ) . In Rcd1- , Rcd5- and MBD-R2-depleted cells , the frequencies of cells with 0–9 signals were drastically increased compared to controls , in which this type of cells are virtually absent ( Fig 3C ) . The similarity of the mitotic phenotypes observed in Rcd1- , Rcd5- and MBD-R2-depleted cells might be a consequence of an interdependence of these proteins . Indeed , it has been previously shown that depletion of Rcd5 in salivary glands leads to a severe reduction in Rcd1 , while Rcd1 depletion results only in a slight reduction in Rcd5 [10] . Similar results were obtained in SL-2 cells , where RNAi-mediated depletion of Rcd5 caused a reduction in Rcd1 , whereas Rcd1 depletion did not substantially affects the Rcd5 level [10] . In both salivary glands and SL-2 cells , MBD-R2 depletion did not affect the stability of the other members of the complex [10] , while the Wds levels were similar to those of controls in Rcd1 , Rcd5 or MBD-R2 RNAi cells [10] . We found that RNAi against Rcd5 leads to a small reduction in Rcd1 also in the S2 cell line used here . Rcd5 RNAi cells displayed a small reduction in MBD-R2 , while MBD-R2 depletion did not substantially affect the Rcd1 level ( Fig 4А ) . Consistent with these results , RT-qPCR showed that RNAi against each of the four NSL genes studied here does not substantially affect transcription of the others , which , with the exception of wds upon Rcd5 RNAi , are expressed at slightly higher rates compared to control ( Fig 4B ) . These results suggest that the phenotypes observed in Rcd1- , Rcd5- , MBD-R2- and Wds-depleted cells are largely due to the deficiency of each individual factor and do not reflect interdependencies among these proteins . In the attempt of detecting a mitotic localization of the NSL subunits , we first tested the commercial anti-MBD-R2 and anti-Wds antibodies and our anti-Rcd1 mouse antibody . All these antibodies specifically recognize bands of the expected MWs that are strongly reduced in RNAi cells ( Figs 1A–1C and 4A ) . However , immunostaining using the same antibodies showed that the anti-MBD-R2 and anti-Wds antibodies stain weakly only the interphase nuclei but do not decorate any mitotic structure , while the anti-Rcd1 did not work at all in indirect immunofluorescence . We then generated stable cell lines expressing Cherry-tubulin and any of Rcd1-GFP , Rcd5-GFP , MBD-R2-GFP or Wds-GFP tagged proteins , all under the control of the copper-inducible MtnA promoter . We exposed these cells for 12–14 hours to different concentrations of CuSO4 ( 0 . 1 , 0 . 25 , 0 . 4 and 0 . 5 mM ) and examined them under a confocal fluorescence microscope . All GFP-tagged proteins showed strong nuclear signals but also displayed clear but different mitotic localizations ( Figs 5 and 6 ) . In prometaphase and metaphase cells , Rcd1-GFP and Rcd5-GFP proteins were no longer associated with chromatin , but accumulated at the centers of the asters/centrosomes and were occasionally weakly enriched at the spindle area . In anaphase and telophase cells , the accumulation at the centrosomes was reduced compared to the previous mitotic phases . In telophase cells , Rcd1-GFP and Rcd5-GFP concentrated in the daughter nuclei as expected for transcription factors , but were also enriched at the midbody , the structure that connects the two daughter cells during late telophase and cytokinesis ( Fig 5 ) . The midbody contains bundled antiparalled MTs with their plus ends overlapping at the center of the structure . The MT overlapping area is often dark ( dark zone ) after staining with anti-tubulin antibodies because it is enriched in proteins that block antibody binding to tubulin [37] . We see a dark zone also in living cells expressing Cherry-tubulin , most likely because the same proteins that prevent antibody binding quench the Cherry-tubulin fluorescence . Interestingly , while Rcd1 was excluded from the dark zone and enriched at both sides of this region , Rcd5 was specifically accumulated in the dark zone at center of the midbody ( Fig 5 ) . After CuSO4 induction , MBD-R2-GFP and Wds-GFP were strongly enriched in interphase nuclei but showed very different localization patterns in mitotic cells . MBD-R2-GFP localized exclusively at the chromosomes and did not show accumulations at either the centrosomes or the midbody ( Fig 6A ) . In contrast , Wds-GFP was enriched at several mitotic structures ( Fig 6B and 6C ) . In all mitotic phases , Wds-GFP accumulated in a discrete region of a specific chromosome . This GFP signal was also detected in both telophase and interphase nuclei of living cells , and was sufficiently strong to be detected in fixed cells with well spread chromosomes ( Fig 6B and 6C ) . This allowed us to localize the Wds-GFP accumulation on a specific region of an acrocentric chromosome characterized by a highly DAPI-fluorescent pericentric heterochromatin ( Fig 6C ) . The DAPI staining pattern of this chromosome and the localization of the GFP signal along the chromosome suggest that Wds might associate with the nucleolus organizer of the X chromosome in both mitotic cells and interphase nuclei [22] . We also observed an enrichment of Wds-GFP at the centrosomes; this enrichment was clearly visible in most prophase , prometaphase and metaphase cells , but was hardly detectable in anaphases and telophases . In addition , in most late prometaphase and metaphase cells , Wds-GFP was enriched at structures that are likely to correspond to the centromeres/kinetochores ( Fig 6B ) . This localization is transient , and was never observed in the other mitotic phases . In all telophase cells , Wds-GFP was accumulated at the midbody dark zone . The cells shown in Figs 5 and 6 were treated for 12–14 hours with 0 . 4 mM CuSO4 , as this is the optimal concentration for a clear visualization of both the GFP-tagged protein and Cherry-tubulin . However , we were able to see mitotic accumulations of GFP-tagged proteins also after 12–14 hours induction with 0 . 1 mM CuSO4 . Importantly , at all CuSO4 concentrations , we consistently observed the same localization patterns as those shown in Figs 5 and 6 . Thus , the mitotic localization of each protein appears to be a characteristic feature of the protein independent of its intracellular concentration . To obtain further insight into this issue , we also performed a Western blotting analysis to determine the levels of the endogenous and the GFP-tagged forms of Rcd1 , MBD-R2 and Wds after CuSO4 induction ( S2 Fig ) . We could not carry out this analysis for cells expressing Rcd5-GFP due to the unavailability of an anti-Rcd5 antibody . These experiments showed that MBD-R2-GFP and its endogenous counterpart were expressed at similar levels after induction with 0 . 1 or 0 . 25 mM CuSO4 , while induction with 0 . 5 mM CuSO4 led to a limited MBD-R2-GFP overexpression ( S2A Fig , S4 Data ) . In MtnA-Rcd1-GFP-bearing cells , the endogenous and the GFP-tagged protein were expressed at similar levels after induction with 0 . 1 mM CuSO4 , but the GFP protein was 2 . 5- and 3 . 9-fold more abundant than the normal protein after induction with 0 . 25 and 0 . 5 mM CuSO4 , respectively ( S2B Fig , S4 Data ) . Wds-GFP was expressed at relatively high levels at all CuSO4 concentrations , with the tagged protein showing expression levels 3-4-fold higher than that of the corresponding endogenous protein ( S2C Fig , S4 Data ) . These results strongly suggest that Rcd1 and MBD-R2 localizations are independent of the concentration of the individual proteins . The lack of biochemical data on Rcd5-GFP and the fact the Wds-GFP is 3 to 4 times more abundant than the endogenous protein do not permit us to exclude that the mitotic localizations of these proteins could be partially affected by their intracellular quantity . However , an analysis of live cells expressing either Rcd5-GFP or Wds-GFP strongly suggests that this is not the case . In cells induced with 0 . 1 mM CuSO4 , there is great variability in the levels of the GFP-proteins expressed by the individual dividing cells . Nonetheless , regardless their degree of GFP fluorescence , all mitotic cells showed the same Rcd5-GFP- or Wds-GFP-specific accumulations ( see Figs 5 and 6 ) , suggesting that the mitotic localization of these proteins is largely independent of their intracellular concentration . We also fixed the cells treated for 12 hours with either 0 . 1 or 0 . 5 mM CuSO4 and stained them with anti-GFP and anti-tubulin antibodies . Cells expressing Rcd1-GFP , Rcd5-GFP , MBD-R2-GFP or Wds-GFP fixed with standard formaldehyde- and paraformaldehyde-based procedures ( see Materials and Methods ) showed an evident GFP staining of interphase nuclei . However , the localization of these proteins on the mitotic apparatus was fixation-dependent but independent of the concentration of CuSO4 . In Rcd1-GFP and Rcd5-GFP expressing cells , we did not detect any clear accumulation of the tagged proteins at either the centrosomes or the central spindles ( S3 Fig ) . This prevented double immunostaining experiments aimed determining whether Rcd1 and Rcd5 precisely colocalize with the centrosomes . Previous work has shown that both Rcd1 and Rcd5 are required for centriole duplication in S2 cells but failed to detect accumulations of these proteins at the centrosomes [16] . However , in these studies cells expressing GFP-tagged Rcd1 or Rcd5 were fixed with 4% paraformaldehyde , a treatment that likely disrupted centrosomal localization of the GFP-tagged proteins [16] . In contrast , fixed cell expressing MBD-R2 GFP showed a strong and specific enrichment of the tagged protein at the chromosomes ( S3 Fig ) . Lastly , after fixation , Wds-GFP was consistently enriched at a specific chromosomal region in metaphase and anaphase cells and at the dark zone of the midbody during telophase . The fixation-resistant localization of Wds on a discrete chromosomal region is a likely example of mitotic chromosome bookmarking . Such bookmarking occurs when transcription factors remain associated with chromosomes during mitosis so as to facilitate reactivation of a subset of genes in the subsequent cell cycle [38] . The observation that the GFP-tagged components of the NSL complex exhibit different localization patterns during mitosis raises the question of whether these proteins have the same functions as their non-tagged counterparts . To address this question we performed RNAi using dsRNAs that target only the 5ʹ and 3ʹ untranslated regions ( UTRs ) of the Rcd1 , Rcd5 , MBD-R2 and wds endogenous genes ( see S1 Table ) but not the coding sequences ( CS ) of the GFP-tagged transgenes . We specifically asked whether the expression of the GFP-tagged NSL proteins rescues the mitotic effects caused by treatments with the corresponding UTR dsRNAs . In cells expressing the GFP-tagged proteins , dsRNAs targeting of CS resulted in very strong phenotypic effects comparable to those observed in normal cells treated with the same CS dsRNAs ( compare Fig 2C with S2 Table ) . dsRNAs targeting the UTR sequences were less effective than CS dsRNAs in inducing mitotic defects ( Fig 2C , S2 Table ) , consistent with the fact that the CS used for RNAi are considerably longer than the corresponding UTRs ( S1 Table ) . However , when RNAi with UTR dsRNAs was performed in cells expressing the corresponding GFP-tagged proteins ( induced by 0 . 1 mM CuSO4; see Materials and Methods ) the mitotic effects were substantially milder than those observed in cells that express only the endogenous proteins ( S2 Table ) . In summary , the data reported in the S2 Table show that the GFP-tagged forms of Rcd1 , Rcd5 , MBD-R2 and Wds rescue the phenotypic defects caused by depletion of the endogenous proteins . This suggests that the GFP-tagged NSL components are largely functional and supports the view that their different localizations during mitosis reflect the normal localizations of their untagged counterparts . The finding that Rcd1 , Rcd5 , MBD-R2 and Wds exhibit different localization patterns during mitosis , and yet cause very similar phenotypes when depleted , suggests the hypothesis that these proteins might act together in regulating the expression of mitotic genes . Previous work has shown that at least four components of the NSL complex ( Nsl1 , Rcd1 , Rcd5 and MBD-R2 ) bind the active promoters of more than 4 , 000 constitutively expressed genes , suggesting that NSL acts as complex to specifically upregulate this type of genes [10 , 12 , 14] . However , it appears that NSL depletion results in diminished expression of only a subset of the genes to which it is bound [12] . To address the possibility that depletion of the NSL components affects mitotic gene transcription , we exploited the published ChIP and gene expression datasets generated in S2 cells [12 , 14] to ask whether the NSL complex binds the promoters and regulates the expression of genes encoding ( i ) centromere and kinetochore proteins ( cid , Cenp-C , Mis12 , Nnf1a , Nnf1b , Kmn1/Nsl1 , Ndc80 , Nuf2 , Spc25/Mitch and Spc105R/KNL1 ) , ( ii ) factors that mediate centriole duplication ( ana2 , asl , SAK , Sas-4 , Sas-6 ) and , and ( iii ) components of the spindle assembly сheckpoint ( SAC ) machinery ( Mad1 , mad2 , Bub1 , Bub3 , BubR1 , Zw10 , rod , Zwilch , cmet , nudE ) . We examined the SAC genes not only as a term of comparison with centromere/kinetochore and centriole duplication genes but also to gather information on whether the SAC is compromised in the absence of functional NSL complex . We found that the promoters of all these genes are bound by at least two components of the NSL complex ( Rcd1 and MBD-R2; Table 1 and S3 Table ) . In addition , interrogation of published datasets [12] revealed that RNAi-mediated Nsl1 depletion results in reduced transcription of most of these genes ( Table 1 ) . Among the centromere/kinetochore genes , the strongest reductions in transcription were observed for Nnf1b , Mis12 and cid , which were transcribed at 6 . 1% , 6 . 9% and 17 . 5% of the control level , respectively . The genes specifying SAC functions were generally under-transcribed , with rod and Zwilch showing particularly reduced transcription levels ( below 30% the control level ) , suggesting the SAC could be partially compromised in cells depleted of the NSL components . Finally , all genes required for centriole duplication showed reduced transcription , with Sas-6 , Sas-4 and asl transcripts reduced to 9 . 5% , 26 . 3% , and 30 . 3% of those of controls , respectively ( Table 1 ) . These data prompted us to determine the transcription levels of cid , Ndc80 , Nnf1b , Mis12 and Spc25/Mitch in cells depleted of the NSL components . RT-qPCR showed that the transcripts of all these genes are substantially reduced in Rcd1 , Rcd5 and MBD-R2 RNAi cells , but only weakly reduced in wds RNAi cells . In cells depleted of Rcd1 , Rcd5 or MBD-R2 , the Nnf1b and Mis12 transcripts were below ( 1 ) Cen , centromere; Kin , kinetochore; SAC , spindle assembly checkpoint; Ced , centriole duplication . ( 2 ) ChIP data on Rcd1 and MBD-R2 are from [14]; ChIP data on Nsl1 are from [12] . We did not include Kmn2 in the table because it was not included in the analyses of [14] and [12] . Transcription factor binding to the gene promoter region is indicated by +; for a quantitative analysis of promoter binding see S3 Table; NA , data not available . ( 3 ) Gene expression data are from [12] . ( 4 ) The values reported are the means of a number ( indicated between brackets ) of independent experiments . 30% of the control level , while the cid transcripts were approximately 50% of the control level; the Spc25/Mitch and Ndc80 transcripts were roughly 70% of those of controls ( Table 1 ) . Importantly , the reductions of the transcript levels detected by RT-qPCR , although quantitatively lower , are absolutely proportional to those previously observed by microarray experiments ( Table 1 and [12] ) . The transcription level of Ndc80 detected by RT-qPCR does not match the strong reduction of the protein seen by Western blotting ( Fig 3A ) . However , it is possible that this discrepancy is due to the downregulation of Nuf2 transcription ( Table 1 ) , which might affect the quantity of the Ndc80 protein , as Nuf2 and Ndc80 are mutually dependent for their stability [34 , 39] . We also determined whether RNAi against Rcd1 , Rcd5 , MBD-R2 and wds affects the transcription of asl , Sas4 and Sas-6 ( Table 1 ) . Sas-6 , SAK and Ana2 form a conserved core module required for Drosophila centriole duplication [40 , 41] . Consistent with the published ChIP data ( Table 1 ) , we found that in Rcd1 , Rcd5 and MBD-R2 RNAi cells the Sas-6 transcripts are substantially reduced compared with controls , ranging from 20% to 37% of the control level ( Table 1 ) . Reductions in the asl and Sas-4 transcripts were less pronounced , ranging from 40–56% and 63–66% of the control levels , respectively . In wds RNAi cells the abundance of Sas4 and Sas-6 was slightly reduced while the level of asl transcripts was not affected ( Table 1 ) . Thus , our direct measure of the transcript abundance determined by RT-qPCR is fully consistent with both the published datasets and with our hypothesis that both the centriole duplication and chromosome segregation phenotypes are caused by reduced transcription of specific mitotic gene sets .
The finding that Rcd1 , Rcd5 , MBD-R2 and Wds exhibit different localization patterns during mitosis , and yet cause very similar phenotypes when depleted raised the hypothesis that these proteins regulate mitosis by controlling the transcription of mitotic genes . This hypothesis is strongly supported by previous [12 , 14] and current ( see Table 1 ) findings indicating that Rcd1 , Rcd5 or MBD-R2 depletion results in a substantial downregulation of several genes involved in centromere/kinetochore assembly and centriole duplication . This hypothesis is further corroborated by the observation that Wds depletion , which causes a limited reduction in mitotic gene transcription , also results in a milder mitotic phenotype compared to those caused by Rcd1 , Rcd5 or MBD-R2 deficiency . It has been previously shown that the NSL complex specifically binds the promoters of most housekeeping genes and activates a large subset of these genes . It has been further shown that subunits of the NSL complex co-localize at promoters of the target genes , and that the complex acts as a single functional unit [10 , 14] . Consistent with these results , we found that RNAi-mediated silencing of Rcd1 , Rcd5 or MBD-R2 results in identical defects in sister chromatid separation . These defects lead to PMLES , which have been previously observed in cells defective in MT-kinetochore interactions [17 , 28 , 32 , 33] . Analysis of published datasets revealed that transcription of both cid and several kinetochore protein-coding genes is downregulated in Nsl1-depleted cells ( Table 1; see [12] ) . Moreover , we showed that in Rcd1 , Rcd5 and MBD-R2 RNAi cells the transcription levels of cid , Mis12 and Nnf1b , and to lesser extent , those of Ndc80 and Spc25/Mitch , are reduced with respect to controls . The same RNAi cells displayed fewer Cid signals and reduced levels of the Cid and Ndc80 proteins compared to control . Thus , these results collectively suggest that Rcd1 , Rcd5 and MBD-R2 work together in interphase to regulate proper transcription of multiple genes encoding centromere and kinetochore components . We propose that reduced transcription of these genes disrupts proper kinetochore assembly , impairing kinetochore-MT interaction . There are at least two considerations that support this interpretation . First , there is a clear hierarchy in recruitment of the kinetochore proteins . Cid is required for recruitment of all kinetochore proteins; localization of the Mis12 complex ( Mis12 , Nnf1a , Nnf1b and Knm1 ) and Spc105R/KNL1 are interdependent , while recruitment of the Ndc80 complex ( Ndc80 , Nuf2 , Spc25R/Mitch ) requires both Spc105R/KNL1 and the Mis12 complex . Second , even components of the same complex such as Ndc80 and Nuf2 are mutually dependent for their stability/localization [34–36 , 42] . These complex dependency relationships suggest that even relatively modest reductions in kinetochore proteins can generate synthetic effects leading to kinetochore dysfunction . We have shown that Rcd1 and Rcd5 accumulate at the centrosomes while MBD-R2 localization is restricted to the chromosomes . Nonetheless , depletion of each of the three NSL components leads to a clear defect in centrosome duplication . Thus , it is unlikely that this centrosomal phenotype is caused by a direct effect of the NSL complex proteins on centrosome behavior . Here again , the most likely explanation is that defective centrosome duplication is due to reduced transcription of multiple genes required for proper centriole structure and duplication , such as ana2 , asl , SAK , Sas-4 and Sas-6 [23 , 40 , 41 , 43 , 44] . The products of these genes also show dependencies in their recruitment at centrioles . For example , the SAK kinase recruits the centriole cartwheel components Sas-6 and Ana2 that are required for recruitment of Sas-4 , which in turn recruits Asl [41 , 44] , suggesting that multiple , even modest , depletions of these proteins can impair the centriole duplication machinery . Thus , we propose that the centrosome duplication phenotype elicited by RNAi against Rcd1 , Rcd5 or MBD-R2 , is due to reduced transcription of target genes required for centriole duplication . Mitotic defects caused by depletion of transcription factors have been previously observed both in flies and mammals . For example , mutations in genes encoding subunits of the Drosophila TFIIH transcription complex have been previously shown to disrupt mitosis . The TFIIH holo-complex is comprised of two subcomplexes , a 7-proteins core complex ( XPB , XPD , p62 , p52 , p44 , p34 and p8 ) and the CAK ( Cdk7 , CycH and MAT1 ) complex [45 , 46] . Mutations in marionette ( mrn ) that encodes Drosophila p52 cause defects in mitotic chromosome condensation and integrity in larval brains [47] . Recently it has been also shown that embryos produced by Drosophila females depleted of p8 , p52 , XPB or Cdk7 often exhibit mitotic defects; these defects include poorly condensed chromosomes , disorganized spindles , isolated centrosomes and chromosomes not associated with the spindle . These mitotic aberrations have been attributed to transcriptional downregulation of a set of critical genes . Specifically , embryos from p8 mutants showed reduced transcription of 104 genes that encode factors involved in DNA replication or mitosis [48] . Another interesting example pointing to an involvement of transcription factors in mitotic regulation is provided by studies on the human MMXD complex , which includes the MIP18 , MMS19 and XPD proteins; XPD is also a component of TFIIH complex and is responsible for Xeroderma pigmentosum ( XP ) . MMS19 or MIP18 knockdown moderately reduces the XPD level but does not affect the levels of the other TFIIH subunits , suggesting that the MMXD complex can function independently of the TFIIH complex . MIP18 , MMS19 and XPD localize to the mitotic spindle of human cells and their siRNA-mediated depletion results in monopolar and multipolar spindles; similar aberrant spindles were also observed in cells from XP patients [49] . However , the molecular mechanisms leading to this mitotic phenotype are not fully understood , and both a direct mitotic role of the MMXD complex and possible subtle defects in global transcription were considered [49] . Strong mitotic defects have also been observed in human cells upon siRNA-mediated inactivation of ERG , a transcription factor of the Erg family . However , in this case the mitotic defects have been attributed to failure to degrade the Aurora A and B mRNAs , and to a consequent excess of these mitotic kinases [50] . While our data suggest that the chromosome segregation and the centrosome duplication phenotypes result from reduced transcription of mitotic genes during interphase , the localization of the NSL proteins to mitotic structures raises the possibility that these proteins might also play direct roles during cell division . A direct mitotic role of the KANSL complex components has been suggested in several studies on human cells . In HeLa cells , MCRS1 ( Rcd5 ) , KANSL1 ( Nsl1/Wah ) and KANSL3 ( Rcd1/Nsl3 ) co-localize with the centrosomes at the center of the asters . KANSL1 and KANSL3 bind the MT minus ends , while MCRS1 does not directly bind MTs but is recruited at the minus ends by KANSL3 [7 , 8] . RNAi-mediated silencing of the MCRS1 , KANSL1 or KANSL3 resulted in very similar phenotypes consisting of misaligned chromosomes , a prolonged arrest in a prometaphase-like state , often followed by mitotic catastrophe [7 , 8] . It has been suggested that MCRS1 , KANSL1 and KANSL3 stabilize MTs , favor chromosome-driven MT formation , and promote correct kinetochore fiber dynamics [8] . WDR5 , the human orthologue of Wds , has been also implicated in mitosis . WDR5 associates with the mitotic spindles [9] and is particularly enriched at the midbody dark zone [51] . Most interestingly , WDR5 depletion results in many cells that are highly reminiscent of the PMLES observed here . In most WDR5-depleted cells , the chromosomes did not properly align in the metaphase plate but are dispersed throughout unusually long anaphase-like spindles without undergoing sister chromatid separation , suggesting a defect in kinetochore-MT connection [9] . It has been also reported that RNAi-mediated knockdown of WDR5 impairs completion of cytokinesis leading to multinucleated cells [51] . We have shown that MBD-R2 remains associated with the chromosomes throughout mitosis , just like some components of the TFIIH complex that localize to the chromosomes of dividing nuclei in Drosophila embryos [48] . However , we found that Rcd1 ( KANSL3 ) and Rcd5 ( MCRS1 ) are enriched at the centrosomes and at the midbodies , with Rcd5 accumulating in the midbody dark zone and Rcd1 excluded from this region but enriched at its sides . Finally , we showed that during mitosis Wds remains associated with a specific chromosome region that is likely to correspond to the nucleolus organizer; in addition , it accumulates at the centrosomes , the kinetochores and the midbody dark zone . These results raise the question of whether the accumulations of Rcd1 , Rcd5 and Wds at the centrosomes/asters and the midbody reflect direct roles of these proteins during mitosis . It is unlikely that Rcd1 , Rcd5 or Wds localization at the centrosomes is related to centrosome duplication , because defects in this process have been also observed in cells depleted of MBD-R2 , which fails to accumulate at the centrosomes . It is also unlikely that these three proteins are required for aster formation , as in Rcd1- , Rcd5- or Wds-depleted cells these structures are morphologically normal . Postulating centrosomal functions for Rcd1 , Rcd5 or Wds is extremely difficult because the centrosomes contain hundreds of proteins that do not appear to serve canonical centrosome functions , and it remains to be determined whether these proteins fulfill centrosome-related regulatory functions [52 , 53] . Postulating direct roles of Rcd1 , Rcd5 and Wds at the midbody is even more difficult . Their enrichment at the midbody suggests an involvement in cytokinesis . However , we did not notice any significant increase in the frequency of binucleated cells after Rcd1 , Rcd5 or wds RNAi , and none of these genes was detected in genome-wide RNAi-based screens aimed at identifying Drosophila genes required for cytokinesis [54 , 55] . Moreover , proteomic analyses have shown that the midbody contains many proteins with no obvious roles in the execution of cytokinesis , and it is currently unknown whether they regulate some aspects of the process [56] . Studies on the human homologues of Rcd1 , Rcd5 and Wds have shown that depletion of these proteins leads to chromosome segregation defects reminiscent of those observed in Drosophila cells [7 , 8 , 9] . These studies did not address the roles of Rcd1 , Rcd5 and Wds on centrosome duplication and did not check whether their depletion leads to a reduced transcription of critical mitotic genes . However , it has been reported that KANSL3 ( Rcd1 ) and MCRS1 ( Rcd5 ) bind to MTs and physically interact with the TPX2 and MCAK spindle proteins , suggesting a direct participation of both KANSL components in the mitotic process [7 , 8] . It has been also shown that WDR5 ( Wds ) interacts with the mitotic kinesin Kif2A and with the MLL complex that binds to MTs and regulates spindle assembly and chromosome segregation [9] . Although these findings do not exclude that downregulation of KANSL3 , MCRS1 and WDR5 lowers the expression of a number of mitotic genes , they suggest that these KANSL subunits play direct mitotic roles . Likewise , we cannot exclude that Rcd1 , Rcd5 , MBD-R2 and Wds have some minor direct roles in chromosome segregation in Drosophila S2 cells . We have shown that Wds localizes to the kinetochores and it is possible that also small , cytologically undetectable , amounts of Rcd1 , Rcd5 , and MBD-R2 localize and function at kinetochores . The human WDR5 protein localizes to dark zone of the midbody and is required for abscission during cytokinesis [51] . It is conceivable that Rcd1 , Rcd5 and Wds play some functions in cytokinesis , and that these functions cannot be detected because they are masked by the chromosome segregation phenotype leading to PMLES . Alternatively , Rcd1 , Rcd5 and Wds might play roles in cytokinesis that do not lead to a complete failure of the process; for example they could regulate the timing of the events underlying cytokinesis either in its early or late stages , such as central spindle assembly and MT severing during abscission . In conclusion , our data suggest that depletion of components of the NSL complex negatively affects centriole duplication and kinetochore assembly through downregulation of genes required for these processes . However , we have also shown that Rcd1 , Rcd5 and Wds accumulate at the centrosomes and the midbody suggesting possible moonlighting functions for these proteins during mitosis . It is generally accepted that protein moonlighting occurred through the gradual transition from the original function to a novel function , an evolutionary process that involves the coexistence of two functions in the same protein . Because these functions should not be conflicting it has been also posited that evolution of moonlighting functions is favored when they are exerted in different cellular compartments [57 , 58] . The components of the conserved NSL/KANSL complex are therefore in a very favorable situation to evolve secondary mitotic functions , as transcription is strongly reduced during cell division [59 , 60] . In both Drosophila and human cells , the NSL/KANSL components appear to localize to mitotic structures . However , the localizations and functions of the NSL subunits are not identical to those of their KANSL counterparts . We would like to speculate that the components of the KANSL and NSL complexes are both evolving towards the acquisition of direct mitotic functions . However , while in human cells the functions of these proteins are integral to the mitotic process , in Drosophila they are not yet essential for mitosis . Although , further analyses are required to compare the mitotic functions of the KANSL and NSL complexes , the current results suggest their components are evolving secondary mitotic functions that are partially different .
All DNA templates for synthesis of dsRNAs specific to the CS or the UTRs of the MBD-R2 , Rcd1 , Rcd5 and wds genes were amplified by PCR from a pool of cDNAs obtained from ovaries of 3-day-old wild-type females and from 0–2 hour wild-type embryos ( for the primer sequences used , see S1 Table ) The PCR products were purified using spin columns ( BioSilica; http://biosilica . ru/ ) . Synthesis of dsRNAs was done as described earlier [61] , with the following minor modifications . Heating of the synthesized dsRNAs to 65°C and the subsequent slow cooling to room temperature were done before treatment with DNaseI; also , the phenol/chloroform extraction was omitted . S2 cells free from mycoplasma contamination were cultured in 39 . 4 g/L Shields and Sang M3 Insect medium ( Sigma ) supplemented with 0 . 5 g/L KHCO3 and 20% heat-inactivated fetal bovine serum ( FBS ) ( Thermo Fisher Scientific ) at 25°C . S2 cells expressing GFP-tagged proteins were cultured in 39 . 4 g/L Shields and Sang M3 Insect medium supplemented with 2 . 5 g/L bacto peptone ( Difco ) , 1 g/L yeast extract ( Difco ) and 5% heat-inactivated FBS at 25°C . RNAi treatments were carried out as follows . 1×106 cells were plated in 1 ml of serum-free medium in a well of a six-well culture dish ( TPP ) and 30 μg of CS dsRNA or 40 μg of UTR dsRNA ( S1 Table ) was added to each well . After a 1 hour incubation , 2 ml of the medium supplemented with 20% or 5% heat-inactivated FBS was added to each well and cells were grown for 3 days . After that , the second dose of the same dsRNA ( 30 μg of CS dsRNA or 40 μg of UTR dsRNA ) was added to each sample and cells were grown for 2 additional days . In the case of the rescue experiments shown in S2 Table , together with the second dose of dsRNA , we added CuSO4 to the final concentration of 0 . 1 mM in the medium . Control S2 cell samples were prepared in the same way , but without addition of dsRNA . Gene-specific primers were designed by using Primer-BLAST ( https://www . ncbi . nlm . nih . gov/tools/primer-blast/ ) or Primer3 ( http://bioinfo . ut . ee/primer3-0 . 4 . 0/primer3/ ) software; primer sequences are provided in S4 Table . For each primer pair , the efficiency was determined by construction of a standard curve using dilutions of the cDNA prepared from S2 cells according to [62] ( S4 Table ) . Total RNA was isolated from control and dsRNA-treated S2 cells using RNAzol RT reagent ( MRC ) according to the manufacturer’s instructions . Genomic DNA was eliminated using the RapidOut DNA Removal Kit ( Thermo Fisher Scientific ) . Reverse transcription was performed with the RevertAid reverse transcriptase ( Thermo Fisher Scientific ) using 2 μg of total RNA in the presence of 2 U/μl of RNaseOut Recombinant RNase Inhibitor ( Thermo Fisher Scientific ) . qPCR was carried out using BioMaster HS-qPCR SYBR Blue ( 2× ) reagent kit ( Biolabmix; http://biolabmix . ru/en/ ) and CFX96 Real-Time PCR Detection System ( Bio-Rad ) . We used the following thermal cycling conditions: 5 minutes at 95°C , followed by 39 cycles of 15 seconds at 95°C , 30 seconds at 60°C , and 30 seconds at 72°C . Data were collected during each extension phase . Negative control templates ( water and cDNA synthesized without reverse transcriptase ) were included in each run . Measurements of gene expression were done at least in two biological replicates , each with three ( or , in the case of the negative controls , in two ) technical replicates . The relative mRNA quantification was determined using the ΔΔCq method . mRNA expression levels were normalized to those of the housekeeping gene RpL32 . A 488-aa portion of Rcd1 ( corresponding to amino acids 133–620 of GenPept accession no . NP_610927 . 3 ) was expressed as GST-fusion in E . coli and subsequently purified as described in [63] . The purified GST-Rcd1 fusion protein was used to immunize mice . Immunization was performed at the Center for Genetic Resources of Laboratory Animals , Institute of Cytology and Genetics SB RAS . Polyclonal antibodies were affinity purified from serum as reported previously [63] . S2 cells were collected by centrifugation and pellets were lysed in either RIPA buffer ( Sigma ) containing 1× Halt Protease and Phosphatase Inhibitor Cocktail ( Thermo Fisher Scientific ) or in lysis buffer ( 50 mM Hepes-KOH pH 7 . 6 , 1 mM MgCl2 , 1 mM EGTA , 1% Triton X-100 , 45 mM NaF , 45 mM β-glycerophosphate , 0 . 2 mM Na3VO4 ) in the presence of a cocktail of protease inhibitors ( Roche ) . Cell extracts were pelleted at 15 , 000g for 15 minutes at 4°C and the supernatants were analyzed by Western blotting . Lysates were run on an 8% or a 10% SDS-PAGE and transferred to an Amersham Protran Supported 0 . 45 μm Nitrocellulose Blotting Membrane ( GE Healthcare ) by wet or semi-dry transfer . Membranes were blocked for 30 minutes in 2% dry milk in PBT ( PBS with 0 . 1% TritonX-100 ) . Membranes were incubated overnight using following primary antibodies: mouse anti-Rcd1 ( 1:500 , this study ) , rabbit anti-MBD-R2 ( 1:1000 , Novus Biologicals 49940002 ) , rabbit anti-Wds ( 1:1000 , Novus Biologicals 40630002 ) , rabbit anti-Ndc80 ( 1:1000; a gift of M . Goldberg , Cornell University ) , rabbit anti-Cid ( 1:500; Active Motif 39713 ) , mouse anti-α-tubulin ( 1:5000 , Sigma T6199 ) , rabbit anti-beta-actin , ( 1:1000 , Invitrogen PA5-16914 ) , mouse anti-Lamin Dm0 ( 1:3500 , Developmental Studies Hybridoma Bank ADL67 . 10 ) and mouse anti-β-actin-HRP-conjugated ( 1:5000 , Santa Cruz Biotechnology sc-47778 HRP ) . The non-HRP-conjugated primary antibodies were detected with HRP-conjugated anti-mouse or anti-rabbit IgGs , using either the ECL detection kit ( GE Healthcare ) or the Novex ECL Chemiluminescent Substrate Reagent Kit ( Thermo Fisher Scientific ) following the manufacturer’s protocols . Full-length CS of MBD-R2 ( nucleotides ( NTs ) 123–3629 , GenBank accession no . NM_169461 . 3 , but with 1433T>C , 1436G>C and 1893G>A NT substitutions ) , Rcd1 ( NTs 295–3492 , GenBank accession no . NM_137083 . 4 ) , Rcd5 ( NTs 101–1834 , GenBank accession no . NM_139595 . 3 , but with 748C>A NT substitution ) , wds ( NTs 300–1382 , GenBank accession no . NM_080245 . 5 , but with 698T>C and 827C>T NT substitutions ) and asl ( NTs 98–3079 , GenBank accession no . NM_141300 . 2 , but with 376C>G , 2046A>G , 2872C>T and 3041A>G NT substitutions ) were cloned in a piggyBac transposon-based plasmid vector upstream of and in frame with the enhanced GFP ( for simplicity , referred to as GFP ) CS . The plasmids also contained a blasticidin-resistance cassette and the sequence encoding mCherry-αTub84B ( Cherry-tubulin ) fluorescent fusion protein . The expression of all fluorescent fusion proteins is under the control of the copper-inducible MtnA promoter . S2 cells co-transfected with a plasmid encoding the fluorescent fusion proteins and a plasmid encoding piggyBac transposase were cultured in 39 . 4 g/L Shields and Sang M3 Insect medium supplemented with 2 . 5 g/L bacto peptone , 1 g/L yeast extract , 5% heat-inactivated FBS and 20 μg/ml blasticidin ( Sigma ) for two weeks at 25°C . The antibiotic was then removed from the culture medium . All cells were free from mycoplasma contamination . To induce expression of fluorescent fusion proteins , cells were grown in the presence of different CuSO4 concentrations ( 0 . 1 , 0 . 25 , 0 . 4 or 0 . 5 mM ) for 12–14 hours before in vivo analysis or fixation . All procedures were performed at room temperature . 2×106 S2 cells were centrifuged at 800g for 5 minutes , washed in 2 ml of PBS ( Sigma ) , and fixed for 10 minutes in 2 ml of 3 . 7% formaldehyde in PBS . Fixed cells were spun down by centrifugation , resuspended in 500 μl of PBS and placed onto a clean side using Cytospin 4 cytocentrifuge ( Thermo Fisher Scientific ) at 900 rpm for 4 minutes . The slides were immersed in liquid nitrogen , washed in PBS , incubated in PBT ( PBS with 0 . 1% TritonX-100 ) for 30 minutes and then in PBS containing 3% BSA for 30 minutes . The slides were then immunostained using the following primary antibodies , all diluted in PBT: mouse anti-α-tubulin ( 1:500 , Sigma T6199 ) , rabbit anti-Spd-2 ( 1:4 , 000 , [21] ) , rabbit anti-Cid ( 1:300 , Abcam ab10887 ) , rabbit anti-CycB ( 1:100 , [64] ) , and mouse anti-Rcd1 ( 1:50 , this study ) . Primary antibodies were detected by incubation for 1 hour with goat FITC-conjugated anti-mouse IgG ( 1: 30 , Sigma F8264 ) or goat Alexa Fluor 568-conjugated anti-rabbit IgG ( 1: 350 , Invitrogen A11077 ) . In the attempt to stain GFP-tagged proteins with anti-GFP antibodies , cells expressing these proteins were collected as described above , and fixed ( i ) for 10 minutes in 2 ml of 3 . 7% formaldehyde in PBS or ( ii ) for 15 minutes in 4% paraformaldehyde in PBS , or ( iii ) for 10 minutes with 8% formaldehyde ( methanol-containing ) in PBS ( Sigma ) . The slides obtained as described above were then stained with mouse anti-α-tubulin ( 1:500 , Sigma T6199 ) and either with rabbit anti-GFP ( 1:200 , Invitrogen A11122 ) or chicken anti-GFP ( 1:200 , Invitrogen PA1-9533 ) , which were detected by Alexa Fluor 568-conjugated goat anti-mouse IgG ( 1:300 , Invitrogen A11031 ) , Alexa Fluor 488-conjugated goat anti-rabbit IgG ( 1:300 , Invitrogen A11034 ) or Alexa Fluor 488-conjugated goat anti-chicken IgG ( 1:300 , Invitrogen A11039 ) , respectively . These procedures stained the interphase nuclei but did not result in clear immunostaining of the spindle-associated GFP-tagged proteins . All slides were mounted in Vectashield antifade mounting medium with DAPI ( Vector Laboratories ) to stain DNA and reduce fluorescence fading . Images were obtained on ZeissAxioImager . M2 using an oil immersion EC Plan-Neofluar 100x/1 . 30 lens ( Carl Zeiss ) and captured by 506 mono ( D ) High Performance camera ( Carl Zeiss ) . The spindle length in S2 cells was measured with the ZEN 2012 ( Carl Zeiss ) software , using the “Spline curve” tool and measure function . We considered only cells that did not appear to be polyploid with respect to the basic karyotype of S2 cells . To measure the spindle length in cells at different mitotic stages we drew a freehand line between the two poles along the spindle axis . The data obtained for each spindle type ( prometaphase/metaphase; anaphase , telophase and PMLES ) were compared using the Wilcoxon Signed-Rank Test and plotted using the BoxPlotR program ( http://shiny . chemgrid . org/boxplotr/ ) . Cells carrying a transgenic construct encoding Cherry-tubulin and either MBD-R2-GFP , Rcd1-GFP , Rcd5-GFP or wds-GFP were grown for 12–14 hours in the presence of different concentrations of CuSO4 ( 0 . 1 , 0 . 25 , 0 . 4 or 0 . 5 mM ) . 500 μl aliquots of cell suspensions ( 5×105 cells/ml ) were then transferred to cell chambers ( Invitrogen A-7816 ) containing coverslips treated with 0 . 25 mg/ml concanavalin A ( Sigma-Aldrich C0412 ) placed on the bottom of the chambers . Observations were performed between 20 and 120 minutes after cell plating in the chamber at a Zeiss LSM 710 confocal microscope , using an oil immersion 100×/1 . 40 plan-apo lens and the ZEN 2012 software . To estimate the promoter binding by Rcd1/Nsl3 and MBD-R2 , we first calculated the genome-wide distributions of these proteins . BAM files with Illumina sequencing reads obtained in ChIP-seq profiling of Rcd1 and MBD-R2 ( and the corresponding Input ) in S2 cells aligned to the Drosophila melanogaster genome release 5 ( dm3 ) [14] were downloaded from ArrayExpress ( https://www . ebi . ac . uk/arrayexpress/ ) ( accession number: E-MTAB-1085 ) . The data were transformed to the BED format using convert2bed in BEDOPS toolkit ( version 2 . 4 . 35 ) ( http://bedops . readthedocs . io/en/latest/index . html ) [65] and genomic positions of aligned reads were converted to Drosophila melanogaster genome release 6 ( dm6 ) using UCSC LiftOver tool ( http://genome . ucsc . edu/cgi-bin/hgLiftOver ) . Next , we divided the genome ( only sequences of chromosomes X , 2L , 2R , 3L , 3R and 4 were taken for the analysis ) into bins of equal size ( 100 bp ) and counted the number of reads in each bin . Then , we converted the counts in reads per million ( RPM ) values and calculated log2-transformed MBD-R2/Input and Rcd1/Input ChIP-seq ratios . Only bins with finite values were used for the further analysis ( 1 , 124 , 416 bins for log2 ( MBD-R2/Input ) and 1 , 125 , 776 bins for log2 ( Rcd1/Input ) ) . To estimate the promoter binding by Nsl1 we first calculated the genome-wide distribution of the protein . Scaled log2 ( ChIP/Input ) microarray-based data for two replicates of Nsl1 in S2 cells [12] were downloaded from GEO ( https://www . ncbi . nlm . nih . gov/geo/ ) ( accession number: GSE30991 ) . Genomic positions of ChIP microarray probes were converted from Drosophila melanogaster genome release 4 ( dm2 ) to 6 ( dm6 ) using FlyBase Drosophila Sequence Coordinates Converter ( http://flybase . org/convert/coordinates ) and the log2 ( ChIP/Input ) ratios from the replicates were averaged . We next retrieved gene annotation data from Ensembl BioMart release 91 ( http://www . ensembl . org/index . html ) . Promoters were arbitrary defined as gene regions spanning from -1000 to +101 bp relative to the transcription start sites ( TSS ) . To measure the levels of Rcd1 , MBD-R2 and Nsl1 binding to promoters , we identified all 100-bins ( in the case of Rcd1 and MBD-R2 ) or microarray probes ( in the case of Nsl1 ) that overlap with each promoter by 1 bp or more . Then , for each promoter , we averaged the log2 ChIP values of such bins or microarray probes . If there were more than one TSS for a gene , their log2 ChIP values were averaged as well . The exact values obtained are reported in S1 Table; in Table 1 we indicate with a “+” symbol all promoters that are enriched in Rcd1 , MBD-R2 or Nsl1 ChIP samples compared to the rest of the genome ( in nearly all cases , the promoter sequences analyzed are within the 5% of the most Rcd1- , MBD-R2- or Nsl1-enriched genomic sequences ) . To measure the effects of Nsl1 deficiency on gene expression , normalized log2-transformed microarray-based data for gene expression in GST-RNAi ( control; in three replicates ) or nsl1 RNAi ( in two replicates ) S2 cells [12] were downloaded from GEO ( accession number: GSE30991 ) and the replicates were averaged . Next , we identified microarray probes that belong to each gene , averaged their values , and calculated the percentage of gene expression in Nsl1-depleted cells compared to control .
|
The Drosophila Nonspecific Lethal ( NSL ) complex is a conserved protein assembly that controls transcription of more than 4 , 000 housekeeping genes . We analyzed the mitotic functions of four genes , Rcd1 , Rcd5 , MBD-R2 and wds , encoding NSL subunits . Inactivation of these genes by RNA interference ( RNAi ) resulted in defects in both chromosome segregation and centrosome duplication . Our analyses indicate that RNAi against Rcd1 , Rcd5 or MBD-R2 reduces transcription of genes involved in centromere/kinetochore assembly and centriole replication . During interphase , Rcd1 , Rcd5 , MBD-R2 and Wds are confined to the nucleus , as expected for transcription factors . However , during mitosis each of these proteins relocates to specific mitotic structures . Our results suggest that the four NSL components work together as a complex to stimulate transcription of genes encoding important mitotic determinants . However , the different localization of the proteins during mitosis suggests that they might have acquired secondary “moonlighting” functions that directly contribute to the mitotic process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"centrosomes",
"rna",
"interference",
"metaphase",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"animals",
"dna",
"transcription",
"animal",
"models",
"mitosis",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"epigenetics",
"centrioles",
"cellular",
"structures",
"and",
"organelles",
"drosophila",
"telophase",
"research",
"and",
"analysis",
"methods",
"genetic",
"interference",
"animal",
"studies",
"chromosome",
"biology",
"gene",
"expression",
"insects",
"arthropoda",
"biochemistry",
"rna",
"eukaryota",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2019
|
RNAi-mediated depletion of the NSL complex subunits leads to abnormal chromosome segregation and defective centrosome duplication in Drosophila mitosis
|
We have investigated the underlying mechanism by which direct cell–cell contact enhances the efficiency of cell-to-cell transmission of retroviruses . Applying 4D imaging to a model retrovirus , the murine leukemia virus , we directly monitor and quantify sequential assembly , release , and transmission events for individual viral particles as they happen in living cells . We demonstrate that de novo assembly is highly polarized towards zones of cell–cell contact . Viruses assembled approximately 10-fold more frequently at zones of cell contact with no change in assembly kinetics . Gag proteins were drawn to adhesive zones formed by viral Env glycoprotein and its cognate receptor to promote virus assembly at cell–cell contact . This process was dependent on the cytoplasmic tail of viral Env . Env lacking the cytoplasmic tail while still allowing for contact formation , failed to direct virus assembly towards contact sites . Our data describe a novel role for the viral Env glycoprotein in establishing cell–cell adhesion and polarization of assembly prior to becoming a fusion protein to allow virus entry into cells .
The ability of retroviruses to utilize and manipulate cell–cell contact for the purpose of efficient transmission contributes to the spread of infection and the progression to diseases such as leukemia and AIDS . In vitro , cell-to-cell transmission of the human immunodeficiency virus ( HIV ) is 100–10 , 000-fold more efficient under conditions of direct cell–cell contact as compared to cell-free virus [1]–[4] . The spread of the human T cell leukemia virus 1 ( HTLV-1 ) depends on contacts between lymphocytes , and little cell-free infectivity is released into the culture supernatant [5] . The enhancement of infectivity by cell–cell contact has been suggested to reflect the proximal coupling of virus assembly and entry machineries [6]–[9] . Indeed , morphological analyses have revealed HIV and HTLV-1 antigens clustering at cell–cell contact zones between antigen-presenting cells and T cells , as well as between infected and uninfected T cells [6] , [10]–[13] . These cell–cell contacts are specifically enriched in microtubules , actin , and adhesion factors , and are designated as “virological” or “infectious” synapses due to their resemblance to the immunological synapse [10]–[12] . In addition to broad synaptic contacts , thin filopodial connections called cytonemes or nanotubes are utilized by retroviruses for the purpose of cell–cell spread [14]–[17] . Importantly , live-cell imaging has confirmed the direct transfer of retroviruses from one cell to another via both , thin filopodial connections and broad virological synapses [14] , [18] . Here , we have applied live four-dimensional ( 4D ) imaging in order to dissect the sequential stages of retroviral assembly , release , and transmission in real time for the model retrovirus murine leukemia virus ( MLV ) . Our data reveal that after the establishment of contacts between infected and uninfected cells , the majority of virus particle assembly is initiated at sites of cell–cell contact . This bias in the site of virus production did not reflect any changes in particle assembly kinetics . Instead , contact-polarized assembly was dependent on signaling from the cytoplasmic tail of viral Env . In sum , we provide evidence that the initiation of retroviral assembly is directed towards infectious cell–cell interfaces , and identify the cytoplasmic tail of Env as a critical viral determinant for efficient intercellular spread .
We used spinning disc confocal microscopy to visualize individual budding and cell-to-cell transmission events of retroviruses in three-dimensional space over time ( 4D ) . Compared to conventional confocal microscopes that contain a single pinhole , the Yokogawa SCU10 scan head used in our system contains about 20 , 000 microlenses that rotate at 1 , 800 rpm and allow the capture of confocal images at high speed and with little photobleaching . This allows the fast acquisition of Z-stacks of images over a long period of time , thereby recording the spatial information ( 3D ) over time ( 4D ) . The 4D imaging allowed us to monitor the dynamics of virus assembly and release , as well as to follow the cell-to-cell transmission of viral particles . For these studies , we monitored the assembly and spread of the model retrovirus MLV because it allowed us to perform precise single-particle tracking ( Figure 1 ) [14] . We have also made attempts to apply single-particle tracking to the transmission of HIV-1 , but these have been impeded by the greater tendency of HIV-1 particles to aggregate at sites of cell–cell contact into big button or ring-shaped clumps of Gag punctae ( unpublished data ) [17] , [18] . We first tested the ability of 4D imaging to detect de novo MLV assembly in the absence of target cells . HEK293 cells were transfected with plasmids encoding the viral components MLV GagPol , Gag-YFP , Env , and genome [20] . A GagPol to Gag-YFP ratio of 10∶1 allowed the production of fully infectious fluorescently labeled MLV viruses [20] , [21] . Six hours following transfection , we identified cells that displayed a few YFP-positive punctae and monitored them by spinning disc confocal microscopy ( Figure 2A ) . We detected the appearance of single fluorescent punctae that gradually intensified and then abruptly disappeared or underwent diffusive movement along the plasma membrane ( Figure 2A , Video S1 ) . Fluorescent punctae were tracked from the time they appeared , and their fluorescence intensity and XYZ coordinates were measured over time . The analysis showed that the fluorescence intensity of these punctae increased from background level over time . Once maximum intensity was reached , the punctae either abruptly dropped to background level or plateau undergoing diffusive movement along the plasma membrane ( Video S1 ) . For example , fluorescent punctae B and D presented in Figure 2A disappeared shortly after reaching maximum intensity , while punctae A and C remained associated with the cell surface for approximately 30 min prior to their sudden disappearance ( Figure 2B ) . The observed increase in intensity followed by either disappearance or diffusive movement along the cell surface is consistent with the interpretation that these events represent de novo assembly followed by release of viral particles . We defined the assembly time as the amount of time that each particle took to achieve maximum intensity from background levels . For example , for the particles presented in Figure 2A , the time of assembly varied from 8 to 30 min ( Figure 2B ) . Interestingly , analysis of the spatial information indicated that some particles formed at the glass–plate interface , completed assembly , and then migrated to the dorsal face of the cell prior to their disappearance ( shown for particle A in Figure 2C , Video S1 ) . Combined , 25 of the 35 particles ( ∼70% ) monitored in these experiments disappeared during the time of imaging . Of these , 60% vanished shortly after completion of assembly ( Figure 2D ) . The other 40% remained associated with the plasma membrane for extended lengths of time , often undergoing rapid movements prior to release ( Figure 2E ) . Similar results were obtained for COS-1 cells ( Figure S1 ) . In sum , the detection of fluorescent punctae that grow in intensity before either abruptly disappearing or undergoing diffusive movement along the plasma membrane is consistent with the interpretation that 4D imaging can detect individual retroviral budding events . In order to study MLV cell-to-cell transmission in cell culture , we explored whether the virus spreads by a contact-dependent mechanism or whether cell-free virus dominates viral transmission . To distinguish between either modes , we cocultured infected and uninfected cells in a viscous 1% methyl cellulose solution previously demonstrated to slow the diffusion of large particles such as viruses [22] . We applied a quantitative assay that is based on an intron-regulated MLV luciferase reporter ( inLuc ) , in which the expression of luciferase is prevented in producer cells and restricted to newly infected target cells ( D . Mazurov , G . Heidecker , P . A . Lloyd , D . Derse , unpublished data ) . Interestingly , whereas 1% methyl-cellulose completely blocked infection with cell-free virus , the spread of MLV infectivity in cocultures of virus-producer cells and uninfected target cells was unaffected ( Figure 3 ) . The resistance of viral spread to 1% methyl cellulose was independent of the cell type used and was observed for transmission between the producer HEK293 and COS-1 cells and target cells such as rat XC , NIH3T3 , and HEK293 cells expressing the MLV receptor mCAT1 . Thus , despite the ability of MLV to be released into the medium ( Figure 2 ) , these results suggested that the predominant mode of MLV spread was through direct cell–cell transmission at physical interfaces ( Figure 3 ) . We next applied these experimental conditions to monitor virus assembly in the context of cell-to-cell transmission . In addition to viral components , we cotransfected virus producer cells to express dynamin2-CFP , a marker protein that accumulates at cell–cell contacts [14] . Contact zones form specifically between Env and receptor expressing cells [14] . They are characterized by dynamin-containing endocytic areas where target cell membranes are anchored in the infected cell ( Figure S2 ) [14] . Although dynamin2 , expressed in the producer cell , has been implicated in the infectivity of HIV [23] , transiently expressed wild-type or dominant-negative dynamin2-CFP did not affect the efficiency of virus cell-to-cell transmission , but facilitated the easy identification of cell–cell contacts ( Figure S3 ) . XC cells expressing a CFP-tagged version of the MLV receptor mCAT1 ( mCAT1-CFP ) were used as target cells [24] . XC cells were chosen because they exhibit a spread and dynamic peripheral actin cytoskeleton that assists visualization of distinct structural features at the cell–cell interface [14] . Five hours posttransfection , we initiated coculture of producer cells generating YFP-labeled MLV and target cells . Following 1 h of coculture , the accumulation of dynamin2-CFP and receptor-CFP molecules allowed us to clearly identify cell–cell contacts between producer and target cells ( green ) . Strikingly , we observed a large number of MLV particles ( red ) emerging from the region of cell–cell contact ( green ) ( Figure 4A , Video S2 ) . Spatial analysis demonstrated that particles were formed at the edge of the producer cell contacting the target cell and then moved up towards the cell body of the target cell ( shown for particle E in Figure 4B , Video S2 ) . Correlative fluorescence and scanning microscopy confirmed that all observed fluorescent punctae correlated to single 100–150-nm viral particles ( Figure 5 ) [25] . Single-particle tracking was used to identify de novo virus assembly events in virus-producing cells . These particles were then tracked over time , and their YFP ( red ) and CFP ( green ) fluorescence intensity as well as XYZ coordinates were measured ( Figure 4C for particle E , and Figure 4D ) . The motility ( blue ) of each fluorescent spot was determined using the distance traveled between consecutive XYZ coordinates . Such an analysis revealed that assembly of viral particles ( red ) was specifically initiated in adhesive zones characterized by an accumulation of receptor and dynamin ( green ) ( Figure 4C for particle E , and Figure 4D ) . Following completion of assembly , most particles were released from producer cells to migrate towards the target cell body ( Figure 4C and 4D ) . Being able to reliably detect de novo assembly events , we next asked whether the assembly events occurred “in” or “out” of cell–cell contact zones . Towards this end , we first identified all de novo MLV assembly events in virus-producing cells ( blue crosses in Figures 6 and 7 and Video S3 ) . Then , we defined contact zones in the virus-producing cell as the region enriched in dynamin2-CFP and receptor mCAT1-CFP . Because the contact zones are dynamic over time , the surface area of the contact zone ( red line ) as well as the noncontact zone ( white line ) were measured for the 37 time points when de novo virus assembly events were detected ( Figure 6A , Table S1 ) . This analysis revealed 44 assembly events in the contact zone and eight outside of the contact zone ( Video S3 ) . To calculate the overall assembly frequency per surface unit ( in square micrometers ) , the number of assembly events observed in either zone was divided by their average surface area in all the 37 frames with assembly events ( Figure 6B , left panel , Table S1 ) . The ratio of the normalized assembly frequency occurring in or out of the contact zone served as an indicator for the fold enhancement of assembly at contact zone . For the cell–cell contact shown in Figure 6 , this analysis revealed a striking 54 . 5-fold enhancement of MLV assembly in zones of cell–cell contact ( Figure 6B , left panel , Table S1 ) . A simplified approach whereby all frames of the time-lapse video were overlaid into a single image to define a larger contact zone ( Figure 6B , right panel ) still revealed a 14 . 7-fold enhancement ( Figure 6B , Table S1 ) . Although the non–time-resolved analysis clearly underestimated the stimulation of assembly at sites of cell–cell contact , it proved to be a rapid and reliable method that allowed the quantification of a large set of time-lapse videos . We applied this method to analyze additional contacts between HEK293 cell and XC cell expressing mCAT1-CFP . The observed stimulation of assembly at sites of cell–cell contact varied between 6- and 18-fold and averaged 11-fold for nine representative contacts ( Figure 7A–7I , Table S2 , Videos S3 , S4 , and S5 ) . A similar enhancement of assembly was observed for cocultures of MLV producing COS-1 cells and XC target cells ( Figure 7L–7N , Table S2 , Video S8 ) as well as for coculture of MLV producing HEK293 cells and HEK293 cells expressing mCAT1-CFP ( Figure 7K , Table S2 , Video S7 ) . This enhancement of assembly at cell–cell contact was independent of expression of mCAT1-CFP in target cells or dynamin2-CFP in producer cells ( Figure 7J and 7K , Table S2 , Videos S6 and S7 ) . Thus , 4D imaging of MLV assembly in the absence and presence of cell–cell contact revealed a striking enhancement of MLV assembly at sites of cell–cell contact . To understand the nature of enhancement of virus assembly at sites of cell–cell contact , we first tested the possibility that assembly is accelerated by contact . Comparative analysis of MLV assembly events in the presence or absence of cell–cell contact revealed that the average MLV assembly time was similar , 14 . 6 and 15 . 6 min , respectively ( Figure 8A ) . The p-value of 0 . 3012 indicated that both values did not significantly differ . The distribution of assembly time for MLV assembly events in the presence or absence of cell–cell contact was also similar , with the most frequently observed assembly time ranged between 9 and 12 min ( Figure 8B ) . Thus , the process of virus assembly is not accelerated at sites of cell–cell contact and proceeds within 14–15 min irrespective of the location . We next asked the question whether a local increase in Gag concentration leads to the enhancement of assembly at sites of cell–cell contact . To test this possibility , HEK293 cells were transfected with an assembly-deficient MLV provirus lacking the capsid domain ( MLVΔCA-GFP ) . This Gag mutant exists intracellularly as a monomer and does not form virus particles . Interestingly , MLVΔCA-GFP was recruited to sites of cell–cell contact ( Figure 9 ) . These results indicate that the local concentration of monomeric Gag is increased at contact zones . These data , taken together with our earlier observation that virus assembly is specifically initiated in contact zones ( Figures 4 , 6 , and 7 ) , suggest that nucleation of assembly , the rate-limiting step of many polymerization reactions , is enhanced at sites of cell–cell contact . Polarization in other biological systems is governed by adhesion proteins that redirect protein sorting towards sites of cell–cell contact to establish polarity [26] . Intriguingly , during virus cell-to-cell transmission , the establishment of cell–cell contact is driven by a high-affinity interaction between viral Env glycoprotein and receptor mCAT1 [14] . Consequently , Env accumulates at sites of cell–cell contact ( Figure 10 ) . Given that the cytoplasmic tail domains of transmembrane adhesion proteins can contribute to establishing cellular polarity , we deleted the cytoplasmic tail of Env in order to determine whether it plays a role in the polarization of assembly . MLV Env glycoproteins are single-pass transmembrane proteins , and their cytoplasmic tails have been shown to regulate Env fusogenicity [27]–[29] . Because C-tail deletion can lead to a high degree of cell–cell fusion in infected cultures , we deleted the histidine residue at position 8 of Env ( Env ΔH8 ) , known to suppress Env fusogenicity without compromising receptor binding [30] , [31] . The formation of cell–cell contacts and polarization of virus assembly to contact sites appeared unaltered for EnvΔH8 ( Figure 11A , Table S3 , Video S9 ) . In contrast , polarized assembly was completely abolished for Env ΔH8 lacking the cytoplasmic tail ( Env ΔH8ΔCT ) despite efficient formation of cell–cell contacts that were indistinguishable from wild-type Env ( Figure 11B , Table S3 , Video S10 ) . Although we cannot exclude the possibility that contact zone dynamics are altered , these data suggest a model whereby direct or indirect signaling via the cytoplasmic tail of Env directs Gag trafficking to sites of cell–cell contact . Long-term imaging experiments of several hours allowed us to monitor polarized assembly in the context of the formation and dissociation of cell–cell contacts . In this case , when imaging was initiated , we could readily observe completely assembled viral particles randomly located at the plasma membrane of the producer cells . However , in response to the establishment of cell–cell contact , we observed that assembly was coordinated with cell-to-cell transmission . We observed that cell-to-cell transmission proceeds in four phases ( Videos S11 and S12 ) . A representative cell–cell contact established between a virus-producing Cos-1 cell and the receptor expressing target cell is presented in Figure 12A ( Videos S11 and S12 ) . Both Dynamin2-CFP and mCAT1-CFP ( green ) accumulated together at sites of contact during Phase I ( Figure 12A ) . In Phase II , de novo MLV assembly ( Gag-YFP , red ) was induced at contact zones , and numerous bright particles were generated . The assembly frequency at these sites of cell–cell contact was elevated as compared to the occasionally observed assembly of few viral particles outside of contact zones . During Phase III , viruses were released and moved along filopodial bridges towards the target cell ( Figure 12A ) . Finally , in Phase IV , virus transmission was stopped by the apparent down-regulation of receptor/dynamin complexes at contacts , resulting in cell separation ( Figure 12A ) . Over the period of 8 . 5 h , we observed four consecutive “waves” of contact , polarized assembly , virus transmission , and cell separation ( Figure 12B , Videos S11 and S12 ) . Quantitative analysis of the CFP-labeled receptor/dynamin2 ( green ) and Gag-YFP fluorescence ( red ) for each wave indicated that the establishment of contact preceded virus assembly ( Figure 12B , Video S12 ) . The average composite of these four transmission events allowed us to generalize our observations ( Figure 12C ) . It took approximately 30 min to establish cell–cell contact before the first virus assembled . Assembly of individual viruses proceeded in approximately 10 min . In the subsequent transmission phase , which lasted approximately 30 min , additional viruses assembled at the contact site and moved towards target cells . Finally , transmission was terminated due to contact down-regulation ( Figure 12C ) . In this system , the establishment and maintenance of cell–cell contact lasted approximately 1 h , whereas assembly was relatively swift , proceeding in approximately 10 min . Taken together , long-term imaging demonstrated that virus assembly at the plasma membrane of infected cells can be polarized in response to the establishment of cell–cell contact , reinforcing the notion of a contact-induced switch from random to polarized assembly .
It has long been known that retroviral spreading is more efficient when cells can physically interact with each other [1]–[4] , [8] . Applying 4D imaging and single-particle tracking , we have demonstrated that the murine leukemia virus can redirect virus assembly to sites of cell–cell contact for transmission to neighboring cells . As such , our results support a model of polarized assembly as the primary cause for the accumulation of viral particles at zones of cell–cell contact ( Model II in Figure 1 ) . Our data contribute to the emerging picture that several steps of the viral life cycle are efficiently coordinated at sites of cell–cell contact . Future work will reveal to what extent our model applies to other viruses and experimental conditions . Our work is based on the ability of spinning disc confocal microscopy to detect de novo assembly and monitor the subsequent spatial movement of completely assembled particles . Applying a cautious definition of contact zones , our visual approach revealed an approximately 10-fold enhancement of virus assembly at sites of cell–cell contact . In the absence of cell–cell contact , particle release from producer cells into the culture supernatant was observed , consistent with the production of cell-free virus . Yet , in the context of coculture , MLV assembly was strongly directed towards sites of cell–cell contact , followed by efficient transmission to target cells . These data indicated that although assembly occurs randomly at plasma membrane , assembly becomes polarized following the establishment of cell–cell contact . In an effort to understand the mechanism of the enhancement of assembly at sites of cell–cell contact , we observed no acceleration of assembly . On average , the assembly time observed for MLV in HEK293 cells was approximately 15 min , slower in comparison to the approximately 8 min observed for HIV in HeLa cells [32] . MLV assembly was even slower in COS-1 cells , averaging 20 . 2 min for 79 events , suggesting that assembly time varies depending on the cell type ( Figure S1 , unpublished data ) . Future experiments carried out in the same cell type in parallel are required to address the observed differences between HIV and MLV . Although virus assembly per se was not accelerated at the sites of cell–cell contact , Gag proteins that drive virus particle assembly were recruited to cell–cell contacts . An elevation of Gag levels at contact sites may increase the frequency of nucleation , thereby enhancing virus assembly . The polarization of assembly required the cytoplasmic tail of the viral Env glycoprotein . Evidence for a communication between the cytoplasmic tail of retroviral Env and Gag proteins has been reported [33]–[36] . Env expressed in polarized cells such as MDCK cells and neurons can relocalize Gag [37]–[39] . In this work , we demonstrate that the establishment of cell–cell adhesion following Env/receptor interactions can break symmetry and establish polarity in otherwise nonpolarized fibroblasts . Future work is needed to understand whether the communication between Env and Gag is direct or indirect . Our results reinforce similarities between virological and biological synapses in that the establishment of cell–cell adhesion is followed by polarization and the directed delivery of ligands towards sites of cell–cell contact [26] . Our data suggest that the MLV Env glycoprotein functions analogously to a cellular adhesion protein that establishes cell–cell contact and polarizes cells . Intriguingly , once MLV Env is packaged into virions , during or soon after virus budding , the cytoplasmic tail is cleaved off by the viral protease [40]–[42] . As such , the viral protease transforms an adhesion protein into a highly fusogenic fusion protein to mediate virus-to-cell fusion . This mechanism represents yet another clever adaptation and utilization of cellular principles by viruses to favor efficient viral spreading .
Plasmid encoding MLV GagPol , MLV LTR-LacZ , MLV Gag-YFP , MLV Env-YFP , mCAT1-CFP , and dynamin2-CFP were described previously [14] , [20] . Plasmids encoding mutant Friend MLV EnvΔH8 was a gift from J . Cunningham ( Harvard Medical School , Boston , MA ) . The cytoplasmic tail of Env ΔH8 was truncated at the viral protease cleavage site by PCR-based mutagenesis to generate MLV Env ΔH8ΔCT . This truncation has also been designated R peptide minus mutant [28] , [42] . CA and Pol coding regions were deleted , and the GFP coding region was fused to C-terminal of NC Full-length Friend MLV genome to generate a mutant provirus that expresses GFP-fused deltaCA Gag as well as Env . HEK293 , COS-1 , and NIH 3T3 cells were maintained in DMEM high glucose ( Invitrogen ) containing 10% FBS plus Pen/Strep/Glutamine . Rat XC sarcoma cells were grown in MEM ( Invitrogen ) with 10% FBS plus Pen/Strep/Glutamine . XC cells stably expressing mCAT1-CFP and HEK293 cells stably expressing mCAT1 were selected using G418 ( Invitrogen ) and twice FACS-sorted for mCAT1 surface expression . Monoclonal mouse anti-human dynamin antibody ( BD Biosciences ) was used to stain endogenous dynamin in virus-producing cells . For live confocal imaging , virus-producing cells and target cells were cocultured in MatTek glass-bottom plates that were pretreated with 0 . 2 mg/ml fibronectin ( Invitrogen ) for 10 min at room temperature . To generate HEK293 cells producing fluorescently labeled MLV , cells were transfected in 24-well plates using 800 ng of total DNA ( 244 ng of MLV Env or MLV Env mutant , 254 ng of MLV GagPol , 26 ng of MLV Gag-YFP , 244 ng of MLV LTR-LacZ , and 32 ng of dynamin2-CFP ) and 2 µl of Lipofectamine 2000 ( Invitrogen ) per well . To generate COS-1 cells producing fluorescently labeled MLV , cells were transfected in six-well plates using 1 , 200 ng of total DNA ( 366 ng of MLV Env , 381 ng of MLV GagPol , 39 ng of MLV Gag-YFP , 366 ng of MLV LTR-LacZ , and 48 ng of dynamin 2-CFP ) and 3 . 6 µl of FuGene 6 reagent ( Roche ) per well . At 4 h post HEK293 transfection and 22 h post Cos-1 transfection , virus-producing cells were replated in fibronectin-coated MatTek plates . One hour later , XC cells expressing mCAT1-CFP were added to start coculture; 1 h post initiation of coculture , live imaging was performed using the 60× objective of a Volocity spinning disc confocal microscope equipped with an environmental chamber ( LIVE CELL; Pathology Devices ) and a Nikon Perfect Focus . We took advantage of the Perfect Focus to simultaneously image multiple cell–cell contacts over time . All time-lapse videos were edited using Volocity , Openlab software ( Improvision/PerkinElmer ) and ImageJ . Videos were saved for presentation in QuickTime format using Sorensen 3 compression for videos . Single-particle tracking of 986 particles , analyzed in this work , was performed using the Quantitation software package from Volocity ( Improvision/PerkinElmer ) . Fluorescent punctae were identified and their YFP and CFP fluorescence intensity as well as XYZ coordinates determined over time . Additional analysis and data presentations were performed following the export of datasets into Microsoft Excel . Correlative fluorescence and scanning electron microscopy was essentially as previously described [24] . Briefly , cells were cocultured on MatTek dishes carrying an etched grid for cell re-identification ( MatTek ) . Immediately after live imaging , cells were fixed in 4% PFA , washed three times with PBS , and then returned to the wide-field fluorescence microscope . Samples were subsequently processed for scanning electron microscopy . Cells were fixed for 30 min with 2 . 5% glutaraldehyde/2% paraformaldehyde in 100 mM cacodylate buffer ( pH 7 . 4 ) , rinsed three times with 100 mM cacodylate buffer , and dehydrated through a graded ethanol series . After washing three times with hexamethyldisilazane ( EMS ) , cells were dried for 5 min at 60°C and coated with platinum . The grid was used to re-identify regions of interest and the area analyzed using a FEI ESEM scanning electron microscope ( Philips ) . To measure MLV cell-to-cell transmission , we applied a quantitative assay that is based on an intron-regulated MLV luciferase reporter ( inLuc ) , in which the expression of luciferase is prevented in producer cells and restricted to newly infected target cells ( D . Mazurov , G . Heidecker , P . A . Lloyd , D . Derse , unpublished data ) . To generate virus-producing cells , HEK293 producer cells were transfected with plasmids encoding the MLV inLuc reporter , MLV Gag Pol , and MLV Env . HEK293 cells were transfected in 24-well plates using 800 ng of total DNA ( 266 ng of MLV Env , 267 ng of MLV GagPol , and 267 ng of MLV LTR-inLuc ) and 2 µl of Lipofectamine 2000 ( Invitrogen ) per well . COS-1 cells were transfected in six-well plates using 1 , 200 ng of total DNA ( MLV Env 400 ng , MLV GagPol 400 ng , and MLV LTR-inLuc 400 ng ) and 3 . 6 µl of FuGene 6 reagent ( Roche ) . At 6 h post HEK293 transfection and 24 h post COS-1 transfection , producer cells were cocultured with target cells at a 2∶1 ratio for 24 h in the absence or presence of 1% methyl cellulose . At the end of coculture , cells were lysed and the luciferase activity measured using a Berthold Technologies Centro LB960 Luminometer .
|
Retroviruses such as the human immunodeficiency virus are known to spread much more efficiently under conditions of direct cell–cell contact as compared to cell-free conditions . How cell–cell contact stimulates virus spreading is poorly understood . In this study , we apply four-dimensional imaging ( 3D space over time ) of a model retrovirus to directly monitor and quantify the events of assembly , release , and transmission of individual viral particles in real time in living cells . Our work reveals that after contacts are established between virus-producing cells and uninfected target cells , the majority of virus particle assembly is initiated at sites of cell–cell contact . The ability of the virus to direct assembly of its particles towards sites of cell–cell contact is dependent on the presence of the cytoplasmic tail of the viral envelope glycoprotein . When this cytoplasmic tail was deleted , virus assembly at cell–cell contacts was no longer enhanced . This study contributes to an emerging model in which several steps of the viral life cycle are efficiently coordinated at sites of cell–cell contact , thereby promoting the spreading of viral infection to neighboring cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/virion",
"structure,",
"assembly,",
"and",
"egress",
"cell",
"biology",
"virology",
"microbiology",
"neuroscience"
] |
2009
|
Assembly of the Murine Leukemia Virus Is Directed towards Sites of Cell–Cell Contact
|
The indigenous people of the Tibetan Plateau have been the subject of much recent interest because of their unique genetic adaptations to high altitude . Recent studies have demonstrated that the Tibetan EPAS1 haplotype is involved in high altitude-adaptation and originated in an archaic Denisovan-related population . We sequenced the whole-genomes of 27 Tibetans and conducted analyses to infer a detailed history of demography and natural selection of this population . We detected evidence of population structure between the ancestral Han and Tibetan subpopulations as early as 44 to 58 thousand years ago , but with high rates of gene flow until approximately 9 thousand years ago . The CMS test ranked EPAS1 and EGLN1 as the top two positive selection candidates , and in addition identified PTGIS , VDR , and KCTD12 as new candidate genes . The advantageous Tibetan EPAS1 haplotype shared many variants with the Denisovan genome , with an ancient gene tree divergence between the Tibetan and Denisovan haplotypes of about 1 million years ago . With the exception of EPAS1 , we observed no evidence of positive selection on Denisovan-like haplotypes .
Adaptation to high altitude is among the most notable examples of natural selection in our species . Tibetans exhibit an exceptional capacity to compensate for decreased oxygen availability ( hypoxia ) , UV exposure , cold , and limited food sources as exemplified by continuous habitation at greater than 4000 meters above sea level on the Qinghai-Tibetan Plateau over several millennia[1–4] . Many candidate genes have been highlighted as likely contributors to Tibetans’ trans-generational success at high altitude[5–9] . These findings are largely based upon the overlap of a priori genes , specifically those involved in hypoxia tolerance , and genomic footprints of adaptation identified in patterns within single nucleotide polymorphism ( SNP ) microarrays[10–12] or exome sequencing data[13] . In a few cases , putatively adaptive tagging variants in hypoxia inducible factor ( HIF ) pathway genes ( EPAS1[10 , 13] , EGLN1 and PPARA[12] ) are associated with relatively lower hemoglobin concentration exhibited by many Tibetans at altitude . With the exception of the EGLN1 locus[14 , 15] , however , precise functional variants within these genomic regions are unknown . Recent advances in whole-genome sequencing ( WGS ) technology provide progress toward identifying the functional variants that underlie high-altitude adaptation . Rather than focusing on a relatively small portion of the genome using candidate gene approaches , WGS selection analyses often have sufficient power to comprehensively interrogate the genome in an unbiased manner . Moreover , WGS offers the opportunity to perform exhaustive searches for adaptive variation , yielding complete surveys of non-protein coding regulatory , conserved , and structural variants possibly tagged or missed in microarray analyses . WGS data are also important from an evolutionary standpoint , providing unbiased insights into long-standing questions regarding adaptive processes , including the role of adaptive introgression , as recently revealed through the targeted sequence comparisons between Tibetans and the Denisovan genome at the EPAS1 gene locus[16] . To identify regions contributing to high-altitude adaptation and regions with archaic introgressions with high resolution , we performed a comprehensive genomic analysis of WGS data in 27 Tibetans[17] . We estimated the demographic history of Tibetans using MSMC and ∂a∂i , and performed the Composite of Multiple Signals test for fine-mapping loci and variants with recent selective sweep . We also estimated genome-wide levels of archaic admixture , and generated a fine-scale map of Denisovan-like introgression in Tibetans .
We conducted principal components analysis ( PCA ) on all SNVs with observed non-reference allele frequency greater than 5% among the combined samples of Tibetans and five 1000 Genomes Project[18] populations ( Peruvian , Punjabi , Chinese , Yoruba and European ) ( Supplementary Methods ) . Tibetans and Han Chinese were closely clustered in the first 4 principal components ( S1A and S1B Fig ) similar to our previous observation[12] , but were clearly differentiated along 5th principal component ( S1C Fig ) . Similarly , ADMIXTURE analysis[19] ( Supplementary Methods ) on the same six populations shows that when the number of ancestral populations ( K ) was set to less than 6 , Tibetans and Han Chinese were genetically similar . However , when K was set to 6 or greater , these two populations exhibited distinct genetic profiles ( S2 Fig ) . The estimated proportion of Tibetan genetic ancestry was greater than 99% for 19 out of the 27 Tibetan individuals . For the remaining Tibetan individuals , the average estimated proportion of Tibetan genetic ancestry was 83 . 1% . For the demographic history analysis , we excluded Tibetan individuals with less than 99% Tibetan genetic ancestry . FST between Tibetans and Chinese based on our SNV data was 0 . 0148 using the Hudson estimator[20] and 0 . 0149 using the Weir and Cockerham estimator [21] , similar to our previous analyses based on SNP microarray data[22] , which is slightly higher than the FST between Chinese and Japanese populations but lower than the FST between pairs of HapMap populations from different continents [23] . We next annotated the functional impact of SNVs that are frequent in Tibetans but uncommon in other populations . We selected all germline SNVs with an alternative allele frequency above 10% in Tibetans but below 1% among Yoruba , Han Chinese , and Europeans from the 1000 Genomes Project . In total , we identified 10 , 702 such SNVs , among which there were 65 nonsynonymous variants that occurred within 58 genes ( S1 Table ) . We predicted the impact of these protein-altering SNVs on genome functions using PolyPhen-2 and the Conservation-controlled Amino Acid Substitution Matrix ( CASM ) in VAAST 2 . 0[24] . Of the 65 nonsynonymous variants , Polyphen-2 predicted 12 to be functional and CASM predicted a distinct set of 3 variants to be functional . One of the variants , Chr2:233244223 A->C ( p . T104P; all genomic coordinates in this manuscript are relative to GRCh37 reference sequence ) , is in the gene ALPP with an observed frequency of 0 . 14% in dbSNP and 14 . 81% in Tibetans . ALPP encodes an alkaline phosphatase , which is mainly expressed in placental and endometrial tissues . Rare nonsynonymous variants in ALPP have previously been associated with decreased risk of spontaneous abortion and in vitro fertilization failure[25] . We inferred the demographic history of Tibetans using the multiple sequential Markovian Coalescent ( MSMC ) [26] ( Fig 1 ) and ∂a∂i[27] ( Fig 2; Table 1 ) ( Supplementary Methods ) . MSMC estimated that the relative cross coalescence rate ( a measurement of the amount of gene flow between two populations ) between Han and Tibetans fell below 80% around 7 kya ( bootstrap 95% CI: 3kya to 10kya ) . We then derived a more detailed demographic model using ∂a∂i , which predicted that the initial divergence between Han and Tibetan was much earlier , at 54kya ( bootstrap 95% C . I 44 kya to 58 kya ) . However , for the first 45ky , the two populations maintained substantial gene flow ( 6 . 8x10-4 and 9 . 0x10-4 per generation per chromosome ) . After 9 . 4 kya ( bootstrap 95% C . I 8 . 6 kya to 11 . 2 kya ) , the gene flow rate dramatically dropped ( 1 . 3x10-11 and 4x10-7 per generation per chromosome ) , which is consistent with the estimate from MSMC . FST predicted by the best-fitting demographic model in ∂a∂i produced estimates consistent with observed FST ( 0 . 0147 using Hudson’s FST estimator and 0 . 0148 using the Weir and Cockerham’s estimator ) . We used the Composite of Multiple Signals ( CMS ) test[28] to map regions under positive selection in the Tibetan genome ( Fig 3 ) . At each genomic variant , CMS evaluates the likelihood of five test statistics ( iHS , XP-EHH , ΔiHH , FST and ΔDAF ) under two models: a null model that assumes neutral evolution and an alternative model that incorporates a variety of scenarios involving recent strong positive selection ( see S1 Supplementary Methods ) . CMS then combines information from the five test statistics in a Bayesian framework to create an aggregated CMS score . CMS and its components ( iHS , FST and XP-EHH ) have been successfully applied to identify adaptive sweeps originating both from the same population[12 , 28 , 29] and from archaic introgression[30–32] . We first calculated the CMS score at each SNV using the 27 Tibetan genomes , with 62 Han Chinese genomes from the 1000 Genomes ( 1KG ) Project[18 , 33] included as a comparison population ( Supplementary Methods ) . In order to evaluate the statistical significance of our top candidates , we performed coalescence simulations with cosi[34] under a neutral model , using the demographic model that we estimated with ∂a∂i[27] ( Supplementary methods ) . Comparing the CMS scores from our observed data and the null distribution from coalescence simulations , we identified 377 candidate SNVs with FDR < 0 . 3; 100 of these SNVs are within 350kb of EGLN1 and 199 are within 350kb of EPAS1 . We listed all the SNVs with FDR<0 . 3 in S2 Table . Because CMS scores between nearby SNVs are often correlated due to linkage disequilibrium , we also performed a region-based CMS test by dividing the genome into 200kb consecutive regions[12] , with the test statistic equal to the highest CMS score within each region . Table 2 lists the top 10 regions with the highest CMS scores . The top regions are in the proximity of EPAS1 and EGLN1 genes , which were previously reported to be responsible for high-altitude adaptation among Tibetans[12 , 13] . One of the regions overlaps the gene VDR , which encodes the vitamin D3 receptor . Mutations in this gene can cause an abnormality in vitamin D metabolism , which may in turn lead to rickets[35] and preterm birth[36] . We examined the evidence of positive selection on the three SNVs with the best-CMS scores in this 200kb region . Based on the Complete Genomics ( CG ) 1000 Genome panel data[18 , 33] , the alternative alleles of top 2 SNVs ( Chr12:48363253 G->C and Chr12:48359527 ) are infrequent in Han Chinese ( MAFs: 1 . 6% and 3 . 2% ) , but the MAFs were comparable between Tibetans ( 31 . 5% and 31 . 5% ) and Europeans ( 32 . 1% and 31 . 4% ) . The third ranking SNV ( Chr12: 48337328 C->T ) has a minor allele frequency of 2 . 4% in Han Chinese , 0 . 0% in Europeans , 17 . 0% in Yoruba and 25 . 9% in Tibetans . The iHS value at this location is 4 . 22 standard deviations higher than the genome-wide mean , which is about 13 , 000 times more likely under the alternative model than the null . Fig 4a displays the haplotype map in Tibetans surrounding this region . The focal SNV is 394 bp upstream of the first exon of the transcript AK309587 , overlapping the binding site of three transcription factors: TCF7L2 , MAFK and MAFF[37] . Both a DNaseI hypersensitivity assay and H3K27me3 histone marker signature in the K562 cell line indicates this is a potential noncoding regulatory region . Another top 10 ranking region overlapped the KCTD12 gene ( potassium channel tetramerization domain containing 12 ) , a component of the G-protein-coupled receptors for γ-aminobutyric acid ( GABA ) receptor[38] . This region contains 3 SNVs with FDR < 0 . 05 from our single-variant CMS test . The strongest signal ( Chr13: 77399462 ) among these three SNVs has an observed minor allele frequency of 18 . 5% in Tibetans , 0 . 06% in dbSNP ( build 142 ) and 0% among Yoruba , Europeans , and Han Chinese in the CG 1000 Genomes panel . The normalized integrated haplotype score ( iHS ) value at this location was 4 . 90 standard deviations higher than the genome-wide mean , which is ~48 , 000 times more likely to occur under the alternative model than under the null model . This indicates that the haplotype homozygosity around the derived allele is much higher than around the ancestral allele , a scenario consistent with strong positive selection . The three SNVs are located in genomic regions with potential regulatory function . Chr13: 77399462 T->C and Chr13:77398842 G->A overlap with histone mark H3K27Ac annotation and are within enhancer regions , while Chr13:77405193 G->A is at a DNase I hypersensitive site and has a strong signature of H3K27me3 in the K562 ( myelogenous leukemia ) cell line[37 , 39] . We next examined Gene Ontology ( GO ) [40] annotations associated with genes in the top 0 . 2% of 200kb windows . We evaluated the statistical significance for over-representation of any GO term by repeated sampling without replacement from all 200kb regions in our dataset ( Table 3 ) . The following GO terms related to hypoxia responses were over-represented: GO:0071456 ( cellular response to hypoxia ) , GO:0036294 ( cellular response to decreased oxygen levels ) , and GO:0071453 ( cellular response to oxygen levels ) . In addition to two genomic regions surrounding EPAS1 and EGLN1 , we identified one additional gene related to hypoxia response in GO:0071456: prostaglandin I2 synthase ( PTGIS ) . PTGIS converts prostaglandin H2 to prostacyclin , an effective vasodilator and inhibitor of platelet activation[41] . The SNV with the highest CMS score in this region ( Chr20: 48175598 C->T ) has a normalized XP-EHH value of 3 . 67 , which was 51 times more likely under the alternative model than the null . The derived allele frequency was 85 . 2% in Tibetan , 42 . 7% in Han Chinese , 32 . 1% in Europeans , and 8 . 9% in Yoruba . The SNV is in the first intron of PTGIS , overlapping DNase I hypersensitivity regions , potential ZNF217 transcription factor binding sites and H3K36me3 histone mark signatures in the NT2-D1 ( pluripotent embryonic carcinoma ) cell line[37] . Because of the lower variant-calling qualities at genomic insertions and deletions ( indels ) , we could not obtain reliable results from haplotype-based tests ( XP-EHH , iHS and ΔiHH ) on indels . Instead , we used the Population Branch Statistic ( PBS ) [13] to search for signatures of positive selection . To calculate the significance levels of the PBS scores , we estimated the null distribution on the same simulation dataset as in the CMS test . We found that indels with the best PBS signals tended to fall into regions having the highest CMS scores . Specifically , 6 out of the top 10 indels are also within the top 10 200kb regions with the highest CMS scores , all in the EPAS1 and EGLN1 regions ( S3 Table; p = 1 . 1x10-15 ) . We fine-mapped SNVs under positive selection in the EPAS1 region by combining the CMS scores with the population allele frequencies in major continental population groups ( Europeans , Han , Native Americans and Yoruba ) ( Fig 4b ) . We discovered 199 SNVs with FDR < 0 . 3 from the CMS test ( CMSq<0 . 3 ) in the EPAS1 region ( S4 Table ) , which spans 321 kb from the first intron of EPAS1 to 257 kb downstream of the last exon of EPAS1 . Of these 199 SNVs , 64 were in the intronic region and the remaining 135 were in the 3’ downstream region . These SNVs are enriched for alleles present in the Denisovan genome and rare in non-Tibetan populations ( MAF<0 . 05 among Yoruba , Europeans , Native Americans and Asians ) . Specifically , 16 . 6% of EPAS1 CMSq<0 . 3 SNVs meet this criterion , as compared to 0 . 3% genome-wide ( p<2x10-16 ) . A previous study identified a 32 . 7kb region ( Chr2: 46567916–46600661 ) in EPAS1 that was highly differentiated between Han and Tibetans and also has a high proportion of Denisovan-like variants[16] . In our CMS test result , 2 out of the top 3 SNVs with the highest CMS score in the EPAS1 region ( 46597756 and 46598025 on Chromosome 2 ) were in this 32 . 7kb region . After excluding rare variants ( SNVs with MAF <0 . 05 in both Han and Tibetans ) , the average FST between Han and Tibetans in this 32 . 7kb region was 0 . 28 . However , the haplotype extends far beyond the previously reported 32 . 7 kb region and contains genomic regions at least 150 kb downstream of EPAS1 . Fig 4b illustrates the Tibetan haplotypes in EPAS1 and its 3’ region . The linkage disequilibrium between SNVs ( both Denisovan-like and not ) was high among Tibetans ( S5 Table ) , suggesting a slow decay of haplotype homozygosity . The small indel with the highest PBS score was a 4bp deletion located in the 2nd intron of EPAS1 ( Chr2:46577800–46577803 ) . Three out of the four nucleotides affected by this deletion are highly conserved in mammals . This indel overlaps an activating H3K27Ac mark in seven cell lines and the binding sites of three transcription factors: the polymerase subunit POLR2A , the HIF co-activator EP300[42] , and GATA2 , which is key in the control of erythroid differentiation[43] . The observed frequency of the derived allele ( deletion ) was 62 . 9% in Tibetans , 0 . 8% in Han Chinese and 0% in Europeans , corresponding to a PBS score of 0 . 600 , which was ranked 3rd among all the genomic variant ( indels and SNVs ) in the EPAS1 gene . In addition to the 32 . 7 kb block , we also identified a second 39 . 0kb candidate block downstream of EPAS1 coding region on the haplotype ( Chr2: 46675505–46714553 ) ; this block was highly differentiated between Han and Tibetans and enriched for Denisovan variants ( S3 Fig ) . Within this block , 88 . 2% ( 60 out of 68 ) of common Tibetan SNVs ( derived allele frequency >50% ) are present in the Denisovan genome , as compared to 81 . 4% ( 48 out of 59 ) in the previously reported 32 . 7kb region . The proportion of SNVs shared with the Neanderthal genome is 26 . 5% ( 18 out of 68 ) for the new 39 . 0kb block as compared to 25 . 4% ( 15 out of 59 ) for the previous 32 . 7kb block . In this block , the average FST between Han and Tibetans for SNVs common in at least one of the two populations is 0 . 304 ( as compared to 0 . 279 for the previously reported region ) . Intriguingly , this region also fully contains a 3 . 4kb Tibetan-enriched deletion region that showed footprints of positive selection in a previous study[44] ( Fig 4b ) . Among our Tibetan samples , this deletion has an allele frequency of 59 . 2% and is in high LD ( r2 = 0 . 86 ) with the SNV with the highest CMS in the EPAS1 region ( Chr2: 46597756; S6 Table ) . We estimated the origin age of the selective sweep on the adaptive EPAS1 haplotype to be 12 kya ( 95% CI: 7 kya to 28 kya ) using a maximum likelihood approach based on coalescence simulations of the demographic model described in Table 1 ( see S1 Supplementary Methods for details ) . Our estimate is consistent with a previous estimate of 12 kya to 15 kya[44] . Using this point estimate of selection start time and assuming that the adaptive haplotype indeed originated from Denisovans or a closely related population , we further inferred that the introgression of this haplotype most likely occurred between 32 and 12 kya . We also estimated the gene-tree divergence time between the adaptive haplotype in Tibetans and the available Denisovan DNA sequence to be 997 kya ( 95% CI: [868 , 1 , 139] ) . From the gene-tree divergence and calibrated models of Denisovan demographic history[45] , we estimated a population divergence time of 868 kya ( 95%CI: 952 kya to 238 kya ) between the Densivan population and the archaic population that contributed the adaptive EPAS1 haplotype . Previously , we identified two non-synonymous variants in the EGLN1 gene ( c . 12C>G , p . D4E; c . 380G>C , p . C127S ) which together act as a co-adapted gene complex contributing to high-altitude adaptation in Tibetans[14] . In the current dataset , the c . 12C>G variant was absent from any other populations in the 1KG dataset and had high FST ( 0 . 70 ) and ΔDAF ( 0 . 70 ) values , indicative of strong positive selection . The c . 380G>C variant has a derived allele frequency of 77 . 8% ( 95% CI: [66 . 7% , 88 . 9%] ) in Tibetans and 44 . 9% ( 95% CI: [40 . 0% , 50 . 0%] ) in Han Chinese , consistent with previous findings[14] . The CMS score is -60 . 43 for c . 12C>G and -72 . 08 for c . 380G>C , both ranked among the top 0 . 01% variants genome-wide . We also identified a novel rare EGLN1 variant ( c . C358T , p . P120L ) within one Tibetan individual , who also carries the c . 380G>C but not c . 12C>G variant . In our dataset this variant has a frequency of 1 . 85% ( 95% CI: [0 . 05% , 9 . 89%] ) in Tibetans , 0 . 81% ( 95% CI: [0 . 02% , 4 . 41%] ) in Han Chinese and is absent from other populations in the 1KG dataset . To evaluate whether any genomic regions with Denisovan introgression other than EPAS1 may contribute to Tibetan high-altitude adaptation , we first calculated D-statistics[46] to evaluate the overall level of Denisovan admixture in the Tibetan genome . When comparing Tibetans to Yoruba , we observed a D of 0 . 011 ( 95% CI: 0 . 0068 to 0 . 0163 ) , indicating a significant excess of Denisovan admixture among Tibetans . However , we observed no significant difference in D between Tibetans and Han Chinese ( D = -0 . 0007 , 95% CI -0 . 0027 to 0 . 0021 ) , demonstrating that the genome-wide level of Denisovan admixture does not differ substantially between Tibetans and Han Chinese , and is consistent with a previous report[47] . Using the Q statistic of Rogers and Bohlender[48] , we estimated that the proportion of Denisovan admixture in Tibetans is 0 . 4% ( 95% CI: 0 . 2% to 0 . 6% , see Methods and S4 Fig ) . Briefly , Q is an f3 ratio estimator of admixture[48] . Given a sample from three populations X , Y and N , so that X and Y are more recently diverged from each other than either is from N , the calculation of Q uses a ratio of expectations to calculate the total admixture from N into Y , in a similar way as other f3-ratio estimators[46] . Q has the advantage of requiring only a single archaic individual , and uses external information about ghost admixture to correct bias due to admixture from an unsampled archaic population ( Neanderthal ) . To investigate the variation in Denisovan admixture across the genome among Tibetans , we first used an LD-based statistic S*[32] to identify introgressed regions originated from Denisovans . S* searches for haplotypes that share a substantial amount of similarity with the archaic genome and are present in the population of interest but not in the comparison population ( see S1 Supplementary Methods ) . In our study , we used Han Chinese as the comparison population and Denisovan alleles as the archaic genome to search for Denisovan haplotypes absent among the Han Chinese genomes and present in at least one Tibetan genome ( Supplementary Methods ) . In total , S* identified 660 50kb candidate regions using this procedure ( S5 Fig; S7 Table ) . As an alternative approach , we also developed a new statistic , D* , which normalizes the D-statistics to enable the identification of admixed regions with calibrated Type I error ( Methods ) . We used D* to compare Tibetans to Han Chinese for all 200kb genomic regions ( S5 Fig ) . With this procedure , we identified 11 regions with significant proportions of Denisovan admixture after multiple-testing correction ( FDR < 0 . 30 ) . The region containing EPAS1 ranked second , with D* = 6 . 2 ( p = 6 . 1 x 10−6 ) . Six of 11 regions identified by D* overlapped with the 660 regions identified by S* , including the region containing EPAS1 ( Table 4 ) . Next , to test for evidence of adaptive Denisovan introgression , we compared the location of the Denisovan DNA regions identified by either S* or D* to the regions with the highest CMS scores . The top candidate identified by this analysis is EPAS1 . Since we were interested in identifying novel adaptive Denisovan introgression , we removed the EPAS1 region and evaluated the statistical significance of the observed number of overlaps . To calculate the p-value , we permuted the CMS scores of all 200kb genomic regions to generate the distribution of overlaps under the null hypothesis that no additional Denisovan-like DNA was positively selected in Tibetans . When we compared the Denisovan-like genomic regions to the top 0 . 2% of CMS regions , we found no overlapping region with D* and only one with S* ( Chr2:42200001–42400000 ) . In the top 1% of CMS regions , we identified no overlapping regions with D* and three overlapping regions ( Chr2: 42200001–42400000; Chr1: 79200001–79400000; Chr1:1200001–1400000 ) with S*; however , the p-values from all tests were non-significant . Therefore , we find no evidence that additional Denisovan-like DNA has contributed to Tibetan high-altitude adaptation . We summarized the CMS scores of Denisovan variants that are present in Tibetans but uncommon among major population groups in Fig 3b .
Our demographic history estimates from ∂a∂i suggest that population structure existed between the Tibetan and Han ancestral subpopulations as early as 44 to 58 kya , with high rates of admixture maintained between the two subpopulations until around 9 kya . This agrees well with the findings of Lu et al . [49] that the early colonization of the Tibetan plateau occurred between 62kya and 38kya , and the post-Last Glacial Maximum ( LGM ) arrival at the Tibetan plateau could be between 15kya and 9kya . Moreover , archaeological evidence [1 , 50] suggests that the Tibetan plateau was occupied during the Late Pleistocene , roughly coinciding with the time of Han-Tibetan separation in our demographic model . Interestingly , MSMC did not detect a decrease in relative cross coalescence rate between Han and Tibetans until 3 to 9 kya . We hypothesized that this discrepancy was due to an inability of MSMC to differentiate between panmixia and high rates of gene flow between Tibetan and Han Chinese populations between 9kya and 54kya . To test this hypothesis , we conducted coalescent simulations to evaluate the ability of MSMC to accurately estimate the Tibetan-Han divergence time under a model where the two populations diverged at 54 kya but experienced a large amount of gene-flow ( 6 . 8x10-4 and 9 . 0x10-4 per generation per chromosome ) until 9 kya . As shown in S6 Fig , MSMC was indeed unable to detect the ancestral population split at 54 kya . The observed pattern of estimated relative cross-coalescence rate was consistent between the observed data and simulations based on the best-fitting ∂a∂i model , with no decrease in the relative cross-coalescence rate prior to 10 kya . Conversely , when we conducted coalescence simulation using the simplified demographic model predicted by MSMC ( i . e . , Han and Tibetan separated 7 , 000 years ago ) and ran ∂a∂i , we found that ∂a∂i accurately recovered the simulated divergence date ( S8 Table ) , providing support for the ∂a∂i model with an earlier divergence date but a high migration rate between the two populations until approximately 9 kya . Our admixture analysis in the EPAS1 region confirmed findings from previous studies that a common Tibetan haplotype in this region contains excessive Denisovan-like DNA[16] , but also showed that on a genome-wide level , the amount of Denisovan admixture in Tibetans ( 0 . 4% ) is similar to that of Han Chinese . This suggests that at least one major introgression occurred from the archaic population to the common ancestor of Han and Tibetans . The high-altitude adaptive haplotype in the EPAS1 region may have been acquired prior to Han-Tibetan divergence , but was either lost or present at low frequencies in Han due to the lack of selection . This explanation agreed with the hypothesis proposed by Huerta-Sanchez et al[16] , who posit that the EPAS1 haplotype may have been introduced prior to the separation of Han and Tibetans , considering the presence of the Tibetan EPAS1 haplotype in a single Han Chinese individual . Alternatively , it is possible that the low frequency of the Tibetan EPAS1 haplotype in China is the result of relatively recent migration from Tibet[16] and that the haplotype was originally introduced into Tibet after the Han-Tibetan divergence . This alternative explanation is supported by our 95% C . I . for the introduction of the EPAS1 haplotype into the ancestral Tibetan population ( 32 to 12 kya ) which is later than the previous estimates for the date of Denisovan admixture into modern humans ( 44 . 0 to 54 . 0 kya ) [51] . Other than the EPAS1 haplotype , we observed no evidence of other Denisovan-like DNA segments contributing to high-altitude adaptation in Tibetans , given that we did not detect significant overlaps between introgressed regions and the top 1% of CMS regions . However , we cannot rule out the possibility that with larger sample sizes , subtle adaptive introgression signals may be detected . In addition to the previously reported 32 . 7kb haplotype block in EPAS1[16] , we identified a much larger haplotype that contains both the EPAS1 genic region and at least 150kb 3’ of EPAS1 . However , the extensive LD within the EPAS1 and its 3’ region prevented us from pinpointing the location of the adaptive mutation . Interestingly , the selected haplotype ( Fig 4b; S3 Fig ) contains many Denisovan-like and non-Denisovan-like variants , and many top candidate SNVs in our analysis , as well as the previously reported 3 . 4kb Tibetan enriched deletion , were absent from the high-coverage Denisovan genome . This observation is reflected in the estimated population divergence time of 952 to 238 kya between the population represented by the Denisovan reference genome and the archaic population that admixed with Tibetans ( Supplementary Methods ) . This estimate is broadly consistent with previous evidence that the archaic Denisovan-like population that admixed with modern human populations separated from the population represented by the Denisovan reference genome between 276 kya and 403 kya[52] . Thus , the advantageous EPAS1 haplotype shows far greater similarity to Denisovans than would be expected in the absence of archaic admixture , but substantial mismatch should be expected given the apparent diversity of archaic populations outside of Africa[53] . Our CMS test combines five statistical tests of positive selection ( iHS , XP-EHH , FST , ΔiHH and ΔDAF ) , offering a more robust performance across a wide variety of scenarios compared to any single test[28]; this allowed us to identify novel candidate genes contributing to adaptations on the Tibetan plateau . One of the candidate gene for adaptive selection is VDR , a gene encoding nuclear hormone receptor for 1 , 25 dihydroxyvitamin D3 and functions in Vitamin D metabolism . Vitamin D is a secosteroid nutrient that plays an important role in calcium homeostasis and bone mineralization , and mainly comes from two sources: 1 ) through skin exposure to sunlight and 2 ) food , such as fish , milk and egg yolk . Deficiency in vitamin D leads to impairment of bone mineralization and skeletal deformities , known as rickets in children and osteomalacia and osteoporosis in adults [54] . Low level of vitamin D has also been associated with cancers , diabetes , autoimmune diseases , hypertension , and infectious disease[35 , 55] . In a study involving vitamin D status in a cohort of 63 Tibetans [54] , the proportion of nomadic Tibetans with vitamin D deficiency ( 25 ( OH ) D <75 nmol/L ) was 100% with 80% of people having severe deficiency ( 25 ( OH ) D < 30nmol/L ) ; the proportion of non-nomadic Tibetans with vitamin D deficiency ranges from 40% to 83% . Consistently , in another study , 61% of Tibetan children suffer from rickets and 51% have stunted growth[56] . Such a high prevalence of vitamin D deficiency may be explained by the traditional Tibetan diet consisting of barley , yak meat and butter tea which are poor sources of vitamin D , and clothing habits in cold temperatures which allow for minimal skin exposure to the sunlight [54] . Therefore , we hypothesize that VDR gene is positively selected to compensate for the lack of vitamin D , the mechanism of which remains to be determined . Previously , it has been shown that genomic regions bound by VDR were under adaptive selection in the human genome[57] . A second promising candidate identified in our CMS test is PTGIS , a gene associated with three hypoxia-response GO terms ( 0071456 , 0036294 , 0071453 ) and regulated by hypoxia-inducible factor 1 ( HIF-1 ) . PTGIS encodes prostaglandin I2 synthase , which participates in the synthesis of prostanoid . Previous studies have shown that the expression of PTGIS is induced by hypoxic conditions in human lung fibroblast cells and cancer cell lines[58] , and can activate vascular endothelial growth factor . Therefore , it is possible that a Tibetan-unique genomic variant may induce vasodilation and angiogenesis in response to hypoxia by altering the expression of PTGIS . The CMS test also identified KCTD12 as a new high-altitude adaptation candidate . Previously , it has been shown in B16 murine melanoma cells that KCTD12 is up-regulated in hypoxic conditions[59] , and thus it may play a role in the hypoxic responses . This gene is predominantly expressed in fetal organs , suggesting a potentially important role in early development , but dramatically lower levels in adult tissue including brain and lung[60] . KCTD12 is differentially expressed in human pulmonary endothelial cells upon 48 hours of hypoxia exposure[59] and noted as one of 40 CpG sites with the greatest difference in methylation levels between highland and lowland Oromo Ethiopians[61] . Our functional annotation of Tibetan genomic variants identified a frequent protein-coding SNV in ALPP , a gene associated with pregnancy losses[25] . Since Tibetans tend to have higher birth weight in high altitude environments compared to lowland native populations living at similar altitudes[62] , it is possible that the nonsynonymous variant in ALPP may affect the birth weight , development , and pregnancy outcomes in Tibetans . For every 1000 meters of altitude , birth weight decreases an average of 100 grams due to restricted fetal growth[63–66] , with a three-fold increase in the number of infants born small for gestational age ( SGA ) at high altitude[67] . Weights of infants born at high altitudes to mothers of highland origin , however , are greater than those of lowland origin . This is shown specifically in infants born to Bolivian versus European women in the Andes , whose birth weights are about 300g higher in the former group , while no difference is reported by ethnicity at low altitude[67] . Whether these putatively adaptive markers play roles in this process remains to be determined . In summary , we performed a comprehensive genomic analysis on whole-genome sequence data from 27 Tibetan individuals . Our analyses detected evidence of population structure between the ancestral Han and Tibetan subpopulations beginning between 44 and 58 kya , although admixture rates between the two subpopulations remained high until around 9 kya . The Denisovan EPAS1 haplotype introgressed into the Tibetan population between 12 and 32 kya , and positive adaptive pressure on this haplotype began between 7 and 28 kya . We summarized the dates of important demographic events in Fig 5 . The Our CMS test identified novel candidate genes for high-altitude adaptation including KCTD12 , VDR and PTGIS , and also generated a list of candidate variants within the EPAS1 gene region . We estimated that 0 . 4% of the Tibetan genome are introgressed DNA from Denisovans , although EPAS1 is probably the only introgressed locus that was influenced by strong positive selection in Tibetans . Our study provided a rich genomic resource of the Tibetan population and generated hypotheses for future positive selection tests .
Our study subjects included 27 Tibetan individuals recruited from Tibetans living in San Diego , California ( n = 3 ) , Salt Lake City , Utah ( n = 5 ) , and the United Kingdom ( n = 19 ) . The sample type and place of origin information are listed in S9 Table . Whole-genome sequencing and variant-calling was performed by CG ( v2 . 0 for the first 17 samples and v2 . 5 for the remaining 10 samples ) . Variants with genotyping quality less than 30 ( i . e . , more than 1 calling error among 1000 variants ) were converted to missing genotype calls , and sites with more than 5% missing genotype rate were removed . To identify additional variants enriched for higher genotyping error rates , we compared 62 genomes that had been sequenced in both CG public genome data[33 , 68] and 1KG Project[18] Phase I data . We identified 652 , 899 SNVs with more than 5% discordant genotype calls between these two datasets , and removed them from all further analyses . We then performed haplotype phasing and imputation using Shapeit2[69] . After previous steps , 10 , 405 , 415 SNVs remained for subsequent analyses . The CMS test combines the signals from five different tests ( ΔiHH , iHS , XP-EHH , FST and ΔDAF ) to create a single test statistic[28] . On each variant , CMS calculates the posterior probability of the variant being selected in a naïve Bayes framework by taking the product of the posterior probabilities from each of the five tests . A higher CMS score is consistent with a stronger signal of positive selection . Our implementation of the CMS test closely follows the recommendations given in the previous report[28] . Briefly , we first performed coalescence simulations using cosi[34] based on the demographic model estimated by ∂a∂i . To simulate genetic data under various scenarios of selective sweeps , we used the following selection coefficients ( s ) : 0 . 02 , 0 . 03 and 0 . 04; the following start times of selection: 0 , 100 , 200 , 300 and 400 generations ago ( the last one represents the oldest selection start time that cosi may simulate within our model ) ; and the following non-reference allele frequencies at the conclusion of the sweep: 0 . 2 , 0 . 4 , 0 . 6 and 0 . 8 . In addition , we also simulated the scenario where no selective sweep had occurred , which corresponded to the null model in the CMS test . In each situation , we performed 1000 simulations . We calculated the CMS scores for all SNVs that passed our quality-control procedures . To calculate their p-values , we also ran the CMS test over the null model that we simulated above . The p-value for a certain CMS score was calculated as the proportion of SNVs in the null model that has the same or higher CMS scores . The D statistic tests whether the amount of archaic admixture in one population exceeds that of another population by examining variant allele frequency spectrum in the case population , control population , an archaic population and an outgroup population[70] . In our study , these were selected as Tibetans , Han Chinese/Yoruba , Denisovan and chimpanzee . Let p^1i , p^2i , p^3i , and p^4i , be the allele frequency of the i-th variant in the case , control , archaic and outgroup populations , respectively . D is defined as: D ( P1 , P2 , P3 , O ) =∑i=1n[ ( 1−p^1i ) p^2ip^3i ( 1−p^4i ) − ( 1−p^2i ) p^1ip^3i ( 1−p^4i ) ]∑i=1n[ ( 1−p^1i ) p^2ip^3i ( 1−p^4i ) + ( 1−p^2i ) p^1ip^3i ( 1−p^4i ) ] ( 1 ) We estimated the confidence interval of D using bootstrap sampling with 200 replicates . Specifically , we divided the whole genome into 1MB blocks , and then randomly sampled with replacement the same number of 1MB blocks as in the actual data , each time calculating a sample D statistic . The 2 . 5% and 97 . 5% quantiles of the 200 D statistics values were used as the two end points of the 95% confidence interval . We developed the D* statistic as a complement to S* to identify local genomic regions with excess archaic admixture in a population . We first intended to use the D statistic to identify 200kb genomic regions with archaic admixture in the case population , but noticed that the variance of the D-statistic was high if the number of archaic variants in a target region was small . This sensitivity of the D-statistic to regions of low archaic gene flow has been noted previously , and alternatives to the D-statistic have been suggested[71] . To illustrate this sensitivity , consider the extreme scenario where only one archaic variant exists in a 200kb genomic region , and this variant was present in only one case genome and absent from all control genomes . In this situation , D will be the largest possible value Eq ( 1 ) despite the fact that the evidence supporting an archaic admixture in this region is poor . This property of D makes it unsuitable for prioritizing candidate introgression regions , since the majority of regions with high D values will be those with few archaic variants in modern human genomes . Previous work has suggested using direct estimators of the quantity of gene flow as an alternative to a method like D-statistics due to the sensitivity of the statistic to population history and other exogenous parameters[71] . However , D-statistics and the related f-statistics vary only in magnitude as a result of these biases , and will only deviate from zero , in expectation , as a result of gene flow[46] . Normalization of D-statistics preserves their desirable characteristics , while eliminating a source of bias . To normalize D-statistics , we calculated the following statistic ( U ) for each 200kb genomic region: U ( P1 , P2 , P3 , O ) =∑i=1n[ ( 1−p^1i ) p^2ip^3i ( 1−p^4i ) + ( 1−p^2i ) p^1ip^3i ( 1−p^4i ) ] ( 2 ) As can be seen , U is the denominator of the D in Eq ( 1 ) and is negatively correlated with the standard deviation associated with D ( S7 Fig ) . To create a normalized D-statistic ( D* ) that can be used to compare admixture signals across regions , we sort all 200-kb genomics regions according to their U values and then divide the regions into 20 equal-sized groups . Then D* of a 200kb region is equal to its D value divided by the within-group standard deviation . We estimated the significance of a given D* by comparing it to the null distribution of D* generated by coalescence simulation . Because only one Denisovan genome is available , we use the statistic Q of Rogers and Bohlender ( Equation 11 of [48] ) , interpreting the primary source of introgression as Denisovan rather than Neanderthal . The expectation of Q depends not only on the fraction , mD , of Denisovan admixture but also on mN , the fraction of ghost admixture from another archaic such as Neanderthal . Equating observed and expected values defines mD as an implicit function of mN . To evaluate this function , we assumed the parameter values in table 3 of [48] . The result is shown as a solid black line in S4 Fig . We repeated this process , swapping the roles of Neanderthal and Denisovan , to estimate mN as a function of mD . This result is shown as a solid red line in S4 Fig . The intersection provides a simultaneous estimate of mD and mN . The database of Genotypes and Phenotypes ( dbGAP ) accession number for the whole-genome sequencing data reported in this paper is phs001338 . This study was approved by the Institutional Review Board at University of Utah and at the University of California , San Diego , by the Berkshire Clinical Research Ethics Committee , UK , and by the Western Institutional Review Board ( WIRB ) . Informed consent was obtained from all participants .
|
The Tibetan population has been residing on high plateau for thousands of years and developed unique adaptation to the local environment . To investigate the demographic history of Tibetans and search for possible adaptive genetic variants , we performed whole-genome sequencing of 27 Tibetan individuals . We found evidence of genetic separation between Han and Tibetans around since 44 and 58 thousand years ago; however , these two populations maintained a high rate of gene flow until 9 thousand years ago . In addition to replicating two previously discovered candidate genes ( EGLN1 and EPAS1 ) for high altitude adaptation , we also found three new candidate genes , including PTGIS , VDR and KCTD12 . We confirmed the high similarity of EPAS1 gene region between Tibetans and Denisovans , but did not detect any evidence of high altitude adaptation from Denisovan gene alleles otherwise .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"han",
"chinese",
"demography",
"population",
"genetics",
"genetic",
"mapping",
"ethnicities",
"population",
"biology",
"tibetan",
"people",
"introgression",
"comparative",
"genomics",
"people",
"and",
"places",
"haplotypes",
"heredity",
"evolutionary",
"processes",
"genetics",
"population",
"groupings",
"biology",
"and",
"life",
"sciences",
"gene",
"flow",
"genomics",
"evolutionary",
"biology",
"genomics",
"statistics",
"computational",
"biology"
] |
2017
|
Evolutionary history of Tibetans inferred from whole-genome sequencing
|
In the primary visual cortex of primates and carnivores , functional architecture can be characterized by maps of various stimulus features such as orientation preference ( OP ) , ocular dominance ( OD ) , and spatial frequency . It is a long-standing question in theoretical neuroscience whether the observed maps should be interpreted as optima of a specific energy functional that summarizes the design principles of cortical functional architecture . A rigorous evaluation of this optimization hypothesis is particularly demanded by recent evidence that the functional architecture of orientation columns precisely follows species invariant quantitative laws . Because it would be desirable to infer the form of such an optimization principle from the biological data , the optimization approach to explain cortical functional architecture raises the following questions: i ) What are the genuine ground states of candidate energy functionals and how can they be calculated with precision and rigor ? ii ) How do differences in candidate optimization principles impact on the predicted map structure and conversely what can be learned about a hypothetical underlying optimization principle from observations on map structure ? iii ) Is there a way to analyze the coordinated organization of cortical maps predicted by optimization principles in general ? To answer these questions we developed a general dynamical systems approach to the combined optimization of visual cortical maps of OP and another scalar feature such as OD or spatial frequency preference . From basic symmetry assumptions we obtain a comprehensive phenomenological classification of possible inter-map coupling energies and examine representative examples . We show that each individual coupling energy leads to a different class of OP solutions with different correlations among the maps such that inferences about the optimization principle from map layout appear viable . We systematically assess whether quantitative laws resembling experimental observations can result from the coordinated optimization of orientation columns with other feature maps .
Neurons in the primary visual cortex are selective to a multidimensional set of visual stimulus features , including visual field position , contour orientation , ocular dominance , direction of motion , and spatial frequency [1] , [2] . In many mammals , these response properties form spatially complex , two-dimensional patterns called visual cortical maps [3]–[25] . The functional advantage of a two dimensional mapping of stimulus selectivities is currently unknown [26]–[28] . What determines the precise spatial organization of these maps ? It is a plausible hypothesis that natural selection should shape visual cortical maps to build efficient representations of visual information improving the ‘fitness’ of the organism . Cortical maps are therefore often viewed as optima of some cost function . For instance , it has been proposed that cortical maps optimize the cortical wiring length [29] , [30] or represent an optimal compromise between stimulus coverage and map continuity [31]–[44] . If map structure was largely genetically determined , map structure might be optimized through genetic variation and Darwinian selection on an evolutionary timescale . Optimization may , however , also occur during the ontogenetic maturation of the individual organism for instance by the activity-dependent refinement of neuronal circuits . If such an activity-dependent refinement of cortical architecture realizes an optimization strategy its outcome should be interpreted as the convergence towards a ground state of a specific energy functional . This hypothesized optimized functional , however , remains currently unknown . As several different functional maps coexist in the visual cortex candidate energy functionals are expected to reflect the multiple response properties of neurons in the visual cortex . In fact , consistent with the idea of joint optimization of different feature maps cortical maps are not independent of each other [8] , [10] , [19] , [23] , [42] , [45]–[48] . Various studies proposed a coordinated optimization of different feature maps [31] , [33] , [34] , [37] , [38] , [40]–[42] , [44] , [49]–[51] . Coordinated optimization appears consistent with the observed distinct spatial relationships between different maps such as the tendency of iso-orientation lines to intersect OD borders perpendicularly or the preferential positioning of orientation pinwheels at locations of maximal eye dominance [8] , [10] , [19] , [23] , [42] , [45] , [47] . Specifically these geometric correlations have thus been proposed to indicate the optimization of a cost function given by a compromise between stimulus coverage and continuity [33] , [35] , [38] , [40] , [42] , [44] , a conclusion that was questioned by Carreira-Perpinan and Goodhill [52] . Visual cortical maps are often spatially complex patterns that contain defect structures such as point singularities ( pinwheels ) [6] , [12] , [53] , [54] , [55] or line discontinuities ( fractures ) [13] , [56] and that never exactly repeat [3]–[10] , [12]–[25] , [57] . It is conceivable that this spatial complexity arises from geometric frustration due to a coordinated optimization of multiple feature maps in which not all inter-map interactions can be simultaneously satisfied [51] , [58]–[61] . In many optimization models , however , the resulting map layout is spatially not complex or lacks some of the basic features such as topological defects [29] , [51] , [58] , [62] , [63] . In other studies coordinated optimization was reported to preserve defects that would otherwise decay [51] , [58] . An attempt to rigorously study the hypothesis that the structure of cortical maps is explained by an optimization process thus raises a number of questions: i ) What are the genuine ground states of candidate energy functionals and how can they be calculated with precision and rigor ? ii ) How do differences in candidate optimization principles impact on the predicted map structure and conversely what can be learned about an hypothetical underlying optimization principle from observations on map structure ? iii ) Is there a way to analyze the coordinated organization of cortical maps predicted by optimization principles in general ? If theoretical neuroscience was able to answer these questions with greater confidence , the interpretation and explanation of visual cortical architecture could build on a more solid foundation than currently available . To start laying such a foundation , we examined how symmetry principles in general constrain the form of optimization models and developed a formalism for analyzing map optimization independent of the specific energy functional assumed . Minima of a given energy functional can be found by gradient descent which is naturally represented by a dynamical system describing a formal time evolution of the maps . Response properties in visual cortical maps are arranged in repetitive modules of a typical spatial length called hypercolumn . Optimization models that reproduce this typical length scale are therefore effectively pattern forming systems with a so-called ‘cellular’ or finite wavelength instability , see [64]–[66] . In the theory of pattern formation , it is well understood that symmetries play a crucial role [64]–[66] . Some symmetries are widely considered biologically plausible for cortical maps , for instance the invariance under spatial translations and rotations or a global shift of orientation preference [51] , [63] , [67]–[71] . In this paper we argue that such symmetries and an approach that utilizes the analogy between map optimization and pattern forming systems can open up a novel and systematic approach to the coordinated optimization of visual cortical representations . A recent study found strong evidence for a common design in the functional architecture of orientation columns [3] . Three species , galagos , ferrets , and tree shrews , widely separated in evolution of modern mammals , share an apparently universal set of quantitative properties . The average pinwheel density as well as the spatial organization of pinwheels within orientation hypercolumns , expressed in the statistics of nearest neighbors as well as the local variability of the pinwheel densities in cortical subregions ranging from 1 to 30 hypercolumns , are found to be virtually identical in the analyzed species . However , these quantities are different from random maps . Intriguingly , the average pinwheel density was found to be statistical indistinguishable from the mathematical constant up to a precision of 2% . Such apparently universal laws can be reproduced in relatively simple self-organization models if long-range neuronal interactions are dominant 3 , 70–72 . As pointed out by Kaschube and coworkers , these findings pose strong constraints on models of cortical functional architecture [3] . Many models exhibiting pinwheel annihilation [51] , [58] or pinwheel crystallization [62] , [63] , [73] were found to violate the experimentally observed layout rules . In [3] it was shown that the common design is correctly predicted in models that describe long-range interactions within the OP map but no coupling to other maps . Alternatively , however , it is conceivable that they result from geometric frustration due to inter-map interactions and joint optimization . In the current study we therefore in particular examined whether the coordinated optimization of the OP map and another feature map can reproduce the quantitative laws defining the common design . The presentation of our results is organized as follows . First we introduce a formalism to model the coordinated optimization of complex and real valued scalar fields . Complex valued fields can represent for instance orientation preference ( OP ) or direction preference maps [14] , [24] . Real valued fields may represent for instance ocular dominance ( OD ) [1] , spatial frequency maps [20] , [45] or ON-OFF segregation [74] . We construct several optimization models such that an independent optimization of each map in isolation results in a regular OP stripe pattern and , depending on the relative representations of the two eyes , OD patterns with a regular hexagonal or stripe layout . A model-free , symmetry-based analysis of potential optimization principles that couple the real and complex valued fields provides a comprehensive classification and parametrization of conceivable coordinated optimization models and identifies representative forms of coupling energies . For analytical treatment of the optimization problem we adapt a perturbation method from pattern formation theory called weakly nonlinear analysis [64]–[66] , [75]–[78] . This method is applicable to models in which the spatial pattern of columns branches off continuously from an unselective homogeneous state . It reduces the dimensionality of the system and leads to amplitude equations as an approximate description of the system near the symmetry breaking transition at which the homogeneous state becomes unstable . We identify a limit in which inter-map interactions that are formally always bidirectional become effectively unidirectional . In this limit , one can neglect the backreaction of the complex map on the layout of the co-evolving scalar feature map . We show how to treat low and higher order versions of inter-map coupling energies which enter at different order in the perturbative expansion . Second we apply the derived formalism by calculating optima of two representative low order examples of coordinated optimization models and examine how they impact on the resulting map layout . Two higher order optimization models are analyzed in Text S1 . For concreteness and motivated by recent topical interest [3] , [79] , [80] , we illustrate the coordinated optimization of visual cortical maps for the widely studied example of a complex OP map and a real feature map such as the OD map . OP maps are characterized by pinwheels , regions in which columns preferring all possible orientations are organized around a common center in a radial fashion [53] , [55] , [81] , [82] . In particular , we address the problem of pinwheel stability in OP maps [51] , [71] and calculate the pinwheel densities predicted by different models . As shown previously , many theoretical models of visual cortical development and optimization fail to predict OP maps possessing stable pinwheels [29] , [51] , [58] , [62] . We show that in case of the low order energies , a strong inter-map coupling will typically lead to OP map suppression , causing the orientation selectivity of all neurons to vanish . For all considered optimization models , we identify stationary solutions of the resulting dynamics and mathematically demonstrate their stability . We further calculate phase diagrams as a function of the inter-map coupling strength and the amount of overrepresentation of certain stimuli of the co-evolving scalar feature map . We show that the optimization of any of the analyzed coupling energies can lead to spatially relatively complex patterns . Moreover , in case of OP maps , these patterns are typically pinwheel-rich . The phase diagrams , however , differ for each considered coupling energy , in particular leading to coupling energy specific ground states . We therefore thoroughly analyze the spatial layout of energetic ground states and in particular their geometric inter-map relationships . We find that none of the examined models reproduces the experimentally observed pinwheel density and spatially aperiodic arrangements . Our analysis identifies a seemingly general condition for interaction induced pinwheel-rich OP optima namely a substantial bias in the response properties of the co-evolving scalar feature map .
We model the response properties of neuronal populations in the visual cortex by two-dimensional scalar order parameter fields which are either complex valued or real valued [53] , [83] . A complex valued field can for instance describe OP or direction preference of a neuron located at position . A real valued field can describe for instance OD or the spatial frequency preference . Although we introduce a model for the coordinated optimization of general real and complex valued order parameter fields we consider as the field of OP and as the field of OD throughout this article . In this case , the pattern of preferred stimulus orientation is obtained by ( 1 ) The modulus is a measure of the selectivity at cortical location . OP maps are characterized by so-called pinwheels , regions in which columns preferring all possible orientations are organized around a common center in a radial fashion . The centers of pinwheels are point discontinuities of the field where the mean orientation preference of nearby columns changes by 90 degrees . Pinwheels can be characterized by a topological charge which indicates in particular whether the orientation preference increases clockwise or counterclockwise around the pinwheel center , ( 2 ) where is a closed curve around a single pinwheel center at . Since is a cyclic variable in the interval and up to isolated points is a continuous function of , can only have values ( 3 ) where is an integer number [84] . If its absolute value , each orientation is represented only once in the vicinity of a pinwheel center . In experiments , only pinwheels with a topological charge of are observed , which are simple zeros of the field . OD maps can be described by a real valued two-dimensional field , where indicates ipsilateral eye dominance and contralateral eye dominance of the neuron located at position . The magnitude indicates the strength of the eye dominance and thus the zeros of the field correspond to the borders of OD . In this article , we view visual cortical maps as optima of some energy functional . The time evolution of these maps can be described by the gradient descent of this energy functional . The field dynamics thus takes the form ( 4 ) where and are nonlinear operators given by , . The system then relaxes towards the minima of the energy . The convergence of this dynamics towards an attractor is assumed to represent the process of maturation and optimization of the cortical circuitry . Various biologically detailed models have been cast to this form [35] , [51] , [85] . All visual cortical maps are arranged in repetitive patterns of a typical wavelength . We splitted the energy functional into a part that ensures the emergence of such a typical wavelength for each map and into a part which describes the coupling among different maps . A well studied model reproducing the emergence of a typical wavelength by a pattern forming instability is of the Swift-Hohenberg type [65] , [86] . Many other pattern forming systems occurring in different physical , chemical , and biological contexts ( see for instance [75]–[78] ) have been cast into a dynamics of this type . Its dynamics in case of the OP map is of the form ( 5 ) with the linear Swift-Hohenberg operator ( 6 ) , and a nonlinear operator . The energy functional of this dynamics is given by ( 7 ) In Fourier representation , is diagonal with the spectrum ( 8 ) The spectrum exhibits a maximum at . For , all modes are damped since and only the homogeneous state is stable . This is no longer the case for when modes on the critical circle acquire a positive growth rate and now start to grow , resulting in patterns with a typical wavelength . Thus , this model exhibits a supercritical bifurcation where the homogeneous state looses its stability and spatial modulations start to grow . The coupled dynamics we considered is of the form ( 9 ) where , and is a constant . To account for the species differences in the wavelengths of the pattern we chose two typical wavelengths and . The dynamics of and is coupled by interaction terms which can be derived from a coupling energy . In the uncoupled case this dynamics leads to pinwheel free OP stripe patterns . How many inter-map coupling energies exist ? Using a phenomenological approach the inclusion and exclusion of various terms has to be strictly justified . We did this by symmetry considerations . The constant breaks the inversion symmetry of inputs from the ipsilateral ( ) or contralateral ( ) eye . Such an inversion symmetry breaking could also arise from quadratic terms such as . In the methods section we detail how a constant shift in the field can eliminate the constant term and generate such a quadratic term . Including either a shift or a quadratic term thus already represents the most general case . The inter-map coupling energy was assumed to be invariant under this inversion . Otherwise orientation selective neurons would , for an equal representation of the two eyes , develop different layouts to inputs from the left or the right eye . The primary visual cortex shows no anatomical indication that there are any prominent regions or directions parallel to the cortical layers [67] . Besides invariance under translations and rotations of both maps we required that the dynamics should be invariant under orientation shifts . Note , that the assumption of shift symmetry is an idealization that uncouples the OP map from the map of visual space . Bressloff and coworkers have presented arguments that Euclidean symmetry that couples spatial locations to orientation shift represents a more plausible symmetry for visual cortical dynamics [68] , [87] , see also [88] . The existence of orientation shift symmetry , however , is not an all or none question . Recent evidence in fact indicates that shift symmetry is only weakly broken in the spatial organization of orientation maps [89] , [90] . A general coupling energy term can be expressed by integral operators which can be written as a Volterra series ( 10 ) with an -th . order integral kernel . Inversion symmetry and orientation shift symmetry require to be even and that the number of fields equals the number of fields . The lowest order term , mediating an interaction between the fields and is given by i . e . ( 11 ) Next , we rewrite Eq . ( 11 ) as an integral over an energy density . We use the invariance under translations to introduce new coordinates ( 12 ) This leads to ( 13 ) The kernel may contain local and non-local contributions . Map interactions were assumed to be local . For local interactions the integral kernel is independent of the locations . We expanded both fields in a Taylor series around ( 14 ) For a local energy density we could truncate this expansion at the first order in the derivatives . The energy density can thus be written ( 15 ) Due to rotation symmetry this energy density should be invariant under a simultaneous rotation of both fields . From all possible combinations of Eq . ( 15 ) only those are invariant in which the gradients of the fields appear as scalar products . The energy density can thus be written as ( 16 ) where we suppress the argument . All combinations can also enter via their complex conjugate . The general expression for is therefore ( 17 ) From all possible combinations we selected those which are invariant under orientation shifts and eye inversions . This leads to ( 18 ) The energy densities with prefactor to do not mediate a coupling between OD and OP fields and can be absorbed into the single field energy functionals . The densities with prefactors and ( also with and ) are complex and can occur only together with ( ) to be real . These energy densities , however , are not bounded from below as their real and imaginary parts can have arbitrary positive and negative values . The lowest order terms which are real and positive definite are thus given by ( 19 ) The next higher order energy terms are given by ( 20 ) Here the fields and enter with an unequal power . In the corresponding field equations these interaction terms enter either in the linear part or in the cubic nonlinearity . We will show in this article that interaction terms that enter in the linear part of the dynamics can lead to a suppression of the pattern and possibly to an instability of the pattern solution . Therefore we considered also higher order interaction terms . These higher order terms contain combinations of terms in Eq . ( 19 ) and are given by ( 21 ) As we will show below examples of coupling energies ( 22 ) form a representative set that can be expected to reproduce experimentally observed map relationships . For this choice of energy the corresponding interaction terms in the dynamics Eq . ( 9 ) are given by ( 23 ) with and denoting the complex conjugate . In general , all coupling energies in , and can occur in the dynamics and we restrict to those energies which are expected to reproduce the observed geometric relationships between OP and OD maps . It is important to note that with this restriction we did not miss any essential parts of the model . When using weakly nonlinear analysis the general form of the near threshold dynamics is insensitive to the used type of coupling energy and we therefore expect similar results also for the remaining coupling energies . Numerical simulations of the dynamics Eq . ( 9 ) , see [63] , [91] , with the coupling energy Eq . ( 22 ) and are shown in Fig . 1 . The initial conditions and final states are shown for different bias terms and inter-map coupling strengths . We observed that for a substantial contralateral bias and above a critical inter-map coupling pinwheels are preserved from random initial conditions or are generated if the initial condition is pinwheel free . Without a contralateral bias the final states were pinwheel free stripe solutions irrespective of the strength of the inter-map coupling . We studied Eq . ( 9 ) with the low order inter-map coupling energies in Eq . ( 22 ) using weakly nonlinear analysis . We therefore rewrite Eq . ( 9 ) as ( 24 ) where we shifted both linear operators as , . The constant term in Eq . ( 9 ) is replaced by a quadratic interaction term with , see Methods . The uncoupled nonlinearities are given by , while and are the nonlinearities of the low order inter-map coupling energy Eq . ( 23 ) . We study Eq . ( 24 ) close to the pattern forming bifurcation where and are small . We therefore expand both control parameters in powers of the small expansion parameter ( 25 ) Close to the bifurcation the fields are small and thus nonlinearities are weak . We therefore expand both fields as ( 26 ) We further introduced a common slow timescale and insert the expansions in Eq . ( 24 ) and get ( 27 ) and ( 28 ) We consider amplitude equations up to third order as this is the order where the nonlinearity of the low order inter-map coupling energy enters first . For Eq . ( 27 ) and Eq . ( 28 ) to be fulfilled each individual order in has to be zero . At linear order in we get the two homogeneous equations ( 29 ) Thus and are elements of the kernel of and . Both kernels contain linear combinations of modes with a wavevector on the corresponding critical circle i . e . ( 30 ) with the complex amplitudes , and , . In view of the hexagonal or stripe layout of the OD pattern shown in Fig . 1 , is an appropriate choice . In the following sections we assume i . e . the Fourier components of the emerging pattern are located on a common circle . To account for species differences we also analyzed models with detuned OP and OD wavelengths in part ( II ) of this study . At second order in we get ( 31 ) As and are elements of the kernel . At third order , when applying the solvability condition ( see Methods ) , we get ( 32 ) We insert the leading order fields Eq . ( 30 ) and obtain the amplitude equations ( 33 ) For simplicity we have written only the simplest inter-map coupling terms . Depending on the configuration of active modes additional contributions may enter the amplitude equations . In addition , for the product-type coupling energy , there are coupling terms which contain the constant , see Methods and Eq . ( 40 ) . The coupling coefficients are given by ( 34 ) From Eq . ( 33 ) we see that inter-map coupling has two effects on the modes of the OP pattern . First , inter-map coupling shifts the bifurcation point from to . This can cause a potential destabilization of pattern solutions for large inter-map coupling strength . Second , inter-map coupling introduces additional resonant interactions that for instance couple the modes and their opposite modes . In case of the inter-map coupling terms in dynamics of the modes are small . In this limit the dynamics of the modes decouples from the modes and we can use the uncoupled OD dynamics , see Methods . When we scale back to the fast time variable and set , we obtain ( 35 ) The amplitude equations are truncated at third order . If pattern formation takes place somewhat further above threshold fifth order , seventh order , or even higher order corrections are expected to become significant and can induce quantitative modifications of the low order solutions . If third order approximate solutions exhibit degeneracies or marginal stability , higher orders of perturbation theory will qualitatively change the solutions . However , none of the solutions found in the studied models was only marginally stable . This suggests that the obtained solutions are in general structurally stable . A derivation of amplitude equation with higher order inter-map coupling energies is presented in Text S1 . Using symmetry considerations we derived inter-map coupling energies up to eighth order in the fields , see Eq . ( 19 ) , Eq . ( 20 ) , and Eq . ( 21 ) . Which of these various optimization principles could reproduce realistic inter-map relationships such as a uniform coverage of all stimulus features ? We identified two types of optimization principles that can be expected to reproduce realistic inter-map relationships and good stimulus coverage . First , product-type coupling energies of the form . These energies favor configurations in which regions of high gradients avoid each other and thus leading to high coverage . Second , gradient-type coupling energies of the form . In experimentally obtained maps , iso-orientation lines show the tendency to intersect the OD borders perpendicularly . Perpendicular intersection angles lead to high coverage as large changes of the field in one direction lead to small changes of the field in that direction . To see that the gradient-type coupling energy favors perpendicular intersection angles we decompose the complex field into the selectivity and the preferred orientation . We obtain ( 36 ) If the orientation selectivity is locally homogeneous , i . e . , then the energy is minimized if the direction of the iso-orientation lines ( ) is perpendicular to the OD borders . In our symmetry-based analysis we further identified terms that are expected to lead to the opposite behavior for instance mixture terms such as . Pinwheels are prominent features in OP maps . We therefore also analyze how different optimization principles impact on the pinwheel positions with respect to the co-evolving feature maps . The product-type coupling energies are expected to favor pinwheels at OD extrema . Pinwheels are zeros of and are thus expected to reduce this energy term . They will reduce energy the most when is maximal which should repel pinwheels from OD borders , where is zero . Also the gradient-type coupling energy is expected to couple the OD pattern with the position of pinwheels . To see this we decompose the field into its real and imaginary part ( 37 ) At pinwheel centers the zero contours of and cross . Since there and are almost constant and not parallel the energy can be minimized only if is small at the pinwheel centers , i . e . the extrema or saddle-points of . From the previous considerations we assume all coupling coefficients of the energies to be positive . A negative coupling coefficient can be saturated by higher order inter-map coupling terms . In the following , we discuss the impact of the low order inter-map coupling energies on the resulting optima of the system using the derived amplitude equations . The corresponding analysis for higher order inter-map coupling energies is provided in Text S1 . As indicated by numerical simulations and weakly nonlinear analysis of the uncoupled OD dynamics , see Methods , we discussed the influence of the OD stripe , hexagon , and constant solutions on the OP map using the coupled amplitude equations derived in the previous section . A potential backreaction onto the dynamics of the OD map can be neglected if the modes of the OP map are much smaller than the modes of the OD map . This can be achieved if . We first give a brief description of the uncoupled OP solutions . Next , we study the impact of the low order coupling energies in Eq . ( 22 ) on these solutions . We demonstrate that these energies can lead to a complete suppression of orientation selectivity . In the uncoupled case there are for two stable stationary solutions to the amplitude equations Eq . ( 35 ) , namely OP stripes ( 38 ) and OP rhombic solutions ( 39 ) with , , an arbitrary phase , and . In the uncoupled case the angle between the Fourier modes is arbitrary . The stripe solutions are pinwheel free while the pinwheel density for the rhombic solutions varies as and thus . For the rhombic solutions pinwheels are located on a regular lattice . We therefore refer to these and other pinwheel rich solutions which are spatially periodic as pinwheel crystals ( PWC ) . In particular , we refer to pinwheel crystals with as rhombic spatial layout as rPWC solutions and pinwheel crystals with a hexagonal layout as hPWC solutions . Without inter-map coupling , the potential of the two solutions reads , thus the stripe solutions are always energetically preferred compared to rhombic solutions . In the following we study three scenarios in which inter-map coupling can lead to pinwheel stabilization . First , a deformation of the OP stripe solution can lead to the creation of pinwheels in this solution . Second , inter-map coupling can energetically prefer the ( deformed ) OP rhombic solutions compared to the stripe solutions . Finally , inter-map coupling can lead to the stabilization of new PWC solutions . For the low order interaction terms the amplitude equations are given by , with the potential ( 40 ) with the uncoupled contributions ( 41 ) The coupling coefficients read , , , , where is the angle between the wavevector and . We first studied the impact of the low order product-type coupling energy . Here , the constant enters explicitly in the amplitude equations , see Eq . ( 40 ) and Eq . ( 68 ) . When using a gradient-type inter-map coupling energy the interaction terms are independent of the OD shift . In this case , the coupling strength can be rescaled as and is therefore independent of the bias . The bias in this case only determines the stability of OD stripes , hexagons or the constant solution .
In this study we presented a symmetry-based analysis of models formalizing that visual cortical architecture is shaped by the coordinated optimization of different functional maps . In particular , we focused on the question of whether and how different optimization principles specifically impact on the spatial layout of functional columns in the primary visual cortex . We identified different representative candidate optimization principles . We developed a dynamical systems approach for analyzing the simultaneous optimization of interacting maps and examined how their layout is influenced by coordinated optimization . In particular , we found that inter-map coupling can stabilize pinwheel-rich layouts even if pinwheels are intrinsically unstable in the weak coupling limit . We calculated and analyzed the stability properties of solutions forming spatially regular layouts with pinwheels arranged in a crystalline array . We analyzed the structure of these pinwheel crystals in terms of their stability properties , spatial layout , and geometric inter-map relationships . For all models , we calculated phase diagrams showing the stability of the pinwheel crystals depending on the OD bias and the inter-map coupling strength . Although differing in detail and exhibiting distinct pinwheel crystal phases for strong coupling , the phase diagrams exhibited many commonalities in their structure . These include the general fact that the hexagonal PWC phase is preceded by a phase of rhombic PWCs and that the range of OD biases over which pinwheel crystallization occurs is confined to the stability region of OD patch solutions . Our analytical calculations of attractor and ground states close a fundamental gap in the theory of visual cortical architecture and its development . They rigorously establish that models of interacting OP and OD maps in principle offer a solution to the problem of pinwheel stability [51] , [71] . This problem and other aspects of the influence of OD segregation on OP maps have previously been studied in a series of models such as elastic net models [34] , [37] , [41] , [42] , [51] , [59] , self-organizing map models [38] , [40] , [44] , [49] , [50] , spin-like Hamiltonian models [58] , [93] , spectral filter models [94] , correlation based models [95] , [96] , and evolving field models [97] . Several of these simulation studies found a higher number of pinwheels per hypercolumn if the OP map is influenced by strong OD segregation compared to the OP layout in isolation or the influence of weak OD segregation [51] , [93] , [97] . In such models , large gradients of OP and OD avoid each other [40] , [49] . As a result , pinwheel centers tend to be located at centers of OD columns as seen in experiments [19] , [45] , [46] , [48] , [95] . By this mechanism , pinwheels are spatially trapped and pinwheel annihilation can be reduced [51] . Moreover , many models appear capable of reproducing realistic geometric inter-map relationships such as perpendicular intersection angles between OD borders and iso-orientation lines [59] , [94] , [95] . Tanaka et al . reported from numerical simulations that the relative positioning of orientation pinwheels and OD columns was dependent on model parameters [95] . Informative as they were , almost all of these previous studies entirely relied on simulation methodologies that do not easily permit to assess the progress and convergence of solutions . Whether the reported patterns were attractors or just snapshots of transient states and whether the solutions would further develop towards pinwheel-free solutions or other states thus remained unclear . Moreover , in almost all previous models , a continuous variation of the inter-map coupling strength was not possible which makes it hard to disentangle the contribution of inter-map interactions from intrinsic mechanisms . The only prior simulation study of a coordinated optimization model that tracked the number of pinwheels during the optimization process did not provide evidence that pinwheel annihilation could be stopped but only reported a modest reduction in annihilation efficiency [51] . From this perspective , the prior evidence for coordination induced pinwheel stabilization appears relatively limited . Our analytical results leave no room to doubt that map interactions can stabilize an intrinsically unstable pinwheel dynamics . They also reveal that interaction of orientation preference with a stripe pattern of OD is per se not capable of stabilizing pinwheels . Independent of its predictions , our study clarifies the general mathematical structure of interaction dominated optimization models . To the best of our knowledge our study for the first time describes an analytical approach for examining the solutions of coordinated optimization models for OP and OD maps . Our symmetry-based phenomenological analysis of conceivable coupling terms provides a general classification and parametrization of biologically plausible coupling terms . To achieve this we mapped the optimization problem to a dynamical systems problem which allows for a perturbation expansion of fixed points , local minima , and optima . Using weakly nonlinear analysis , we derived amplitude equations as an approximate description near the symmetry breaking transition . We identified a limit in which inter-map coupling becomes effectively unidirectional enabling the use of the uncoupled OD patterns . We studied fixed points and calculated their stability properties for different types of inter-map coupling energies . This analysis revealed a fundamental difference between high and low order coupling energies . For the low order versions of these energies , a strong inter-map coupling typically leads to OP map suppression , causing the orientation selectivity of all neurons to vanish . In contrast , the higher order variants of the coupling energies do generally not cause map suppression but only influence pattern selection , see Text S1 . We did not consider an interaction with the retinotopic map . Experimental results on geometric relationships between the retinotopic map and the OP map are ambiguous . In case of ferret visual cortex high gradient regions of both maps avoid each other [42] . In case of cat , however , high gradient regions overlap [18] . Such positive correlations cannot be easily treated with dimension reduction models , see [98] . It is noteworthy that our phenomenological analysis identified coupling terms that could induce an attraction of high gradient regions . Such terms contain the gradient of only one field and can thus be considered as a mixture of the gradient and the product-type energy . Our results indicate that a patchy layout of a second visual map interacting with the OP map is important for the effectiveness of pinwheel stabilization by inter-map coupling . Such a patchy layout can be easily induced by an asymmetry in the representation of the corresponding stimulus feature such as eye dominance or spatial frequency preference . In spatial frequency maps , for instance , low spatial frequency patches tend to form islands in a sea of high spatial frequency preference [45] . Also in cat visual cortex the observed OD layout is patchy [99]–[103] . In our model , the patchy layout results from the overall dominance of one eye . In this case , OD domains form a system of hexagonal patches rather than stripes enabling the capture and stabilization of pinwheels by inter-map coupling . The results from all previous models did not support the view that OD stripes are capable of stabilizing pinwheels [51] , [93] , [97] . Our analysis shows that OD stripes are indeed not able to stabilize pinwheels , a result that appears to be independent of the specific type of map interaction . In line with this , several other theoretical studies , using numerical simulations [51] , [93] , [97] , indicated that more banded OD patterns lead to less pinwheel rich OP maps . For instance , in simulations using an elastic net model , the average pinwheel density of OP maps interacting with a patchy OD layout was reported substantially higher ( about 2 . 5 pinwheels per hypercolumn ) than for OP maps interacting with a more stripe-like OD layout ( about 2 pinwheels per hypercolumn ) [51] . Several lines of biological evidence appear to support the picture of interaction induced pinwheel stabilization . Supporting the notion that pinwheels might be stabilized by the interaction with patchy OD columns , visual cortex is indeed dominated by one eye in early postnatal development and has a pronounced patchy layout of OD domains [104]–[106] . Further support for the potential relevance of this picture comes from experiments in which the OD map was artificially removed resulting apparently in a significantly smoother OP map [44] . In this context it is noteworthy that macaque visual cortex appears to exhibit all three fundamental solutions of our model for OD maps: stripes , hexagons , and a monocular solution , which are stable depending on the OD bias . In the visual cortex of macaque monkeys , all three types of patterns are found near the transition to the monocular segment , see [106] and Fig . ( 8 ) . Here , OD domains form bands in the binocular region and a system of ipsilateral eye patches at the transition zone to the monocular region where the contralateral eye gradually becomes more dominant . If pinwheel stability depends on a geometric coupling to the system of OD columns one predicts systematic differences in pinwheel density between these three zones of macaque primary visual cortex . Because OD columns in the binocular region of macaque visual cortex are predominantly arranged in systems of OD stripes our analysis also indicates that pinwheels in these regions are either stabilized by other patchy columnar systems or intrinsically stable . One important general observation from our results is that map organization was often not inferable by simple qualitative considerations on the energy functional . The organization of interaction induced hexagonal pinwheel crystals reveals that the relation between coupling energy and resulting map structure is quite complex and often counter intuitive . We analyzed the stationary patterns with respect to intersection angles and pinwheel positions . In all models , intersection angles of iso-orientation lines and OD borders have a tendency towards perpendicular angles whether the energy term mathematically depends on this angle , as for the gradient-type energies , or not , as for the product-type energies . Intersection angle statistics thus are not a very sensitive indicator of the type of interaction optimized . Mathematically , these phenomena result from the complex interplay between the single map energies and the interaction energies . In case of the low order gradient-type inter-map coupling energy all pinwheels are located at OD extrema , as expected from the used coupling energy . For other analyzed coupling energies , however , the remaining pinwheels are located either at OD saddle-points ( low order product-type energy ) or near OD borders ( higher order gradient-type energy ) , in contrast to the expection that OD extrema should be energetically preferred . Remarkably , such correlations , which are expected from the gradient-type coupling energies , occur also in the case of the product-type energies . Remarkably , in case of product type energies pinwheels are located at OD saddle-points . which is not expected per se and presumably result from the periodic layout of OP and OD maps . Correlations between pinwheels and OD saddle-points have not yet been studied quantitatively in experiments and may thus provide valuable information on the principles shaping cortical functional architecture . Our results demonstrate that , although distinct types of coupling energies can leave distinguishing signatures in the structure of maps shaped by interaction ( as the OP map in our example ) , drawing precise conclusions about the coordinated optimization principle from observed map structures is not possible for the analyzed models . In the past numerous studies have attempted to identify signatures of coordinated optimization in the layout of visual cortical maps and to infer the validity of specific optimization models from aspects of their coordinated geometry [34] , [37] , [38] , [40]–[42] , [44] , [49]–[51] , [58] , [59] , [93] . It was , however , never clarified theoretically in which respect and to which degree map layout and geometrical factors of inter-map relations are informative with respect to an underlying optimization principle . Because our analysis provides complete information of the detailed relation between map geometry and optimization principle for the different models our results enable to critically assess whether different choices of energy functionals specifically impact on the predicted map structure and conversely what can be learned about the underlying optimization principle from observations of map structures . We examined the impact of different interaction energies on the structure of local minima and ground states of models for the coordinated optimization of a complex and a real scalar feature map such as OP and OD maps . The models were constructed such that in the absence of interactions , the maps reorganized into simple stripe or blob pattern . In particular , the complex scalar map without interactions would form a periodic stripe pattern without any phase singularity . In all models , increasing the strength of interactions could eventually stabilize qualitatively different , more complex , and biologically more realistic patterns containing pinwheels that can become the energetic ground states for strong enough inter-map interactions . The way in which this happens provides fundamental insights into the relationships between map structure and energy functionals in optimization models for visual cortical functional architecture . Our results demonstrate that the structure of maps shaped by inter-map interactions is in principle informative about the type of coupling energy . The organization of the complex scalar map that optimizes the joined energy functional was in general different for all different types of coupling terms examined . We identified a class of hPWC solutions which become stable for large inter-map coupling . This class depends on a single parameter which is specific to the used inter-map coupling energy . Furthermore , as shown in Text S1 , pinwheel positions in rPWCs , tracked while increasing inter-map coupling strength , were different for different coupling terms examined and thus could in principle serve as a trace of the underlying optimization principle . This demonstrates that , although pinwheel stabilization is not restricted to a particular choice of the interaction term , each analyzed phase diagram is specific to the used coupling energy . In particular , in the strong coupling regime substantial information can be obtained from a detailed inspection of solutions . In the case of the product-type coupling energies , the resulting phase diagrams are relatively complex as stationary solutions and stability borders depend on the magnitude of the OD bias . Here , even quantitative values of model parameters can in principle be constrained by analysis of the map layout . In contrast , for the gradient-type coupling energies , the bias dependence can be absorbed into the coupling strength and only selects the stationary OD pattern . This leads to relatively simple phase diagrams . For these models map layout is thus uninformative of quantitative model parameters . We identified several biologically very implausible OP patterns . In the case of the product-type energies , we found orientation scotoma solutions which are selective to only two preferred orientations . In the case of the low order gradient-type energy , we found OP patterns containing pinwheels with a topological charge of 1 which have not yet been observed in experiments . If the relevant terms in the coupling energy could be determined by other means , the parameter regions in which these patterns occur could be used to constrain model parameters by theoretical bounds . The information provided by map structure overall appears qualitative rather than quantitative . In both low order inter-map coupling energies ( and the gradient-type higher order coupling energy , see Text S1 ) , hPWC patterns resulting from strong interactions were fixed , not exhibiting any substantial dependence on the precise choice of interaction coefficient . In principle , the spatial organization of stimulus preferences in a map is an infinite dimensional object that could sensitively depend in distinct ways to a large number of model parameters . It is thus not a trivial property that this structure often gives essentially no information about the value of coupling constants in our models . The situation , however , is reversed when considering the structure of rPWCs . These solutions exist and are stable although energetically not favored in the absence of inter-map interactions . Some of their pinwheel positions continuously depend on the strength of inter-map interactions . These solutions and their parameter dependence nevertheless are also largely uninformative about the nature of the interaction energy . This results from the fact that rPWCs are fundamentally uncoupled system solutions that are only modified by the inter-map interaction . As pointed out before , preferentially orthogonal intersection angles between iso-orientation lines and OD borders appear to be a general feature of coordinated optimization models in the strong coupling regime . Although the detailed form of the intersection angle histogram is solution and thus model specific , our analysis does not corroborate attempts to use this feature to support specific optimization principles , see also [52] , [107] , [108] . The stabilization of pinwheel crystals for strong inter-map coupling appears to be universal and provides per se no specific information about the underlying optimization principle . In fact , the general structure of the amplitude equations is universal and only the coupling coefficients change when changing the coupling energy . It is thus expected that also for other coupling energies , respecting the proposed set of symmetries , PWC solutions can become stable for large enough inter-map coupling . Our analysis conclusively demonstrates that OD segregation can stabilize pinwheels and induce pinwheel-rich optima in models for the coordinated optimization of OP and OD maps when pinwheels are intrinsically unstable in the uncoupled dynamics of the OP map . This allows to systematically assess the possibility that inter-map coupling might be the mechanism of pinwheel stabilization in the visual cortex . The analytical approach developed here is independent of details of specific optimization principles and thus allowed to systematically analyze how different optimization principles impact on map layout . Moreover , our analysis clarifies to which extend the observation of the layout in physiological maps can provide information about optimization principles shaping visual cortical organization . The common design observed in experimental OP maps [3] is , however , not reproduced by the optima of the analyzed optimization principles . Whether this is a consequence of the applied weakly nonlinear analysis or of the low number of optimized feature maps or should be considered a generic feature of coordinated optimization models will be examined in part ( II ) of this study [91] . In part ( II ) we complement our analytical studies by numerical simulations of the full field dynamics . Such simulations allow to study the rearrangement of maps during the optimization process , to study the timescales on which optimization is expected to take place , and to lift many of the mathematical assumptions employed by the above analysis . In particular , we concentrate on the higher order inter-map coupling energies for which the derived amplitude equations involved several simplifying conditions , see Text S1 .
We studied the intersection angles between iso-orientation lines and OD borders . The intersection angle of an OD border with an iso-orientation contour is given by ( 63 ) where denotes the position of the OD zero-contour lines . A continuous expression for the OP gradient is given by . We calculated the frequency of intersection angles in the range . In this way those parts of the maps are emphasized from which the most significant information about the intersection angles can be obtained [19] . These are the regions where the OP gradient is high and thus every intersection angle receives a statistical weight according to . For an alternative method see [10] . We studied how the emerging OD map depends on the overall eye dominance . To this end we mapped the uncoupled OD dynamics to a Swift-Hohenberg equation containing a quadratic interaction term instead of a constant bias . This allowed for the use of weakly nonlinear analysis to derive amplitude equations as an approximate description of the shifted OD dynamics near the bifurcation point . We identified the stationary solutions and studied their stability properties . Finally , we derived expressions for the fraction of contralateral eye dominance for the stable solutions .
|
Neurons in the visual cortex form spatial representations or maps of several stimulus features . How are different spatial representations of visual information coordinated in the brain ? In this paper , we study the hypothesis that the coordinated organization of several visual cortical maps can be explained by joint optimization . Previous attempts to explain the spatial layout of functional maps in the visual cortex proposed specific optimization principles ad hoc . Here , we systematically analyze how optimization principles in a general class of models impact on the spatial layout of visual cortical maps . For each considered optimization principle we identify the corresponding optima and analyze their spatial layout . This directly demonstrates that by studying map layout and geometric inter-map correlations one can substantially constrain the underlying optimization principle . In particular , we study whether such optimization principles can lead to spatially complex patterns and to geometric correlations among cortical maps as observed in imaging experiments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"physics",
"general",
"physics",
"computational",
"neuroscience",
"interdisciplinary",
"physics",
"biology",
"computational",
"biology",
"sensory",
"systems"
] |
2012
|
Coordinated Optimization of Visual Cortical Maps (I) Symmetry-based Analysis
|
The factors that determine the characteristic seasonality of influenza remain enigmatic . Current models predict that occurrences of influenza outside the normal surveillance season within a temperate region largely reflect the importation of viruses from the alternate hemisphere or from equatorial regions in Asia . To help reveal the drivers of seasonality we investigated the origins and evolution of influenza viruses sampled during inter-seasonal periods in Australia . To this end we conducted an expansive phylogenetic analysis of 9912 , 3804 , and 3941 hemagglutinnin ( HA ) sequences from influenza A/H1N1pdm , A/H3N2 , and B , respectively , collected globally during the period 2009-2014 . Of the 1475 viruses sampled from Australia , 396 ( 26 . 8% of Australian , or 2 . 2% of global set ) were sampled outside the monitored temperate influenza surveillance season ( 1 May – 31 October ) . Notably , rather than simply reflecting short-lived importations of virus from global localities with higher influenza prevalence , we documented a variety of more complex inter-seasonal transmission patterns including “stragglers” from the preceding season and “heralds” of the forthcoming season , and which included viruses sampled from clearly temperate regions within Australia . We also provide evidence for the persistence of influenza B virus between epidemic seasons , in which transmission of a viral lineage begins in one season and continues throughout the inter-seasonal period into the following season . Strikingly , a disproportionately high number of inter-seasonal influenza transmission events occurred in tropical and subtropical regions of Australia , providing further evidence that climate plays an important role in shaping patterns of influenza seasonality .
Human influenza virus is characterized by a marked seasonality in temperate regions , where the virus exhibits a distinct annual peak in epidemic activity during the winter months [1] . However , in subtropical and tropical regions transmission patterns are often less clearly defined [2] . Indeed , the epidemiological dynamics of influenza in tropical regions are more commonly characterized by continuous low rates of disease throughout the year , with semi-annual epidemics in some regions [3–5] . Two main categories of factor appear to drive these complex patterns of influenza seasonality [1 , 6 , 7]: those associated with changes in host behavior , perhaps driven by differing environmental conditions , and changes in the physical properties of the virus ( or host ) that reflect large-scale environmental changes . Examples of potential behavioral drivers of influenza seasonality include patterns of school attendance or crowding indoors during inclement weather [8–10] , while potential environmental factors may include changes in humidity and temperature that affect virion stability and hence virus survival [11–13] . Understanding the patterns of viral activity in different climatic regions and at different times is central to revealing the determinants of influenza seasonality . However , those studies undertaken to date have generally focused on patterns of virus transmission within the defined influenza season in temperate regions [14] . Indeed , there is a marked absence of studies of viral transmission and evolution outside of the usual time-scale of influenza seasons ( with an epidemiological study of the inter-seasonal 2010/2011 period in Australia an exception [15] ) , even though these may provide an important perspective on influenza seasonality . Inter-seasonal influenza is generally thought to involve the importation of an influenza virus from a locality either in the alternate hemisphere where the influenza season is current [6] , or from the tropics where low levels of virus may circulate year-round [2 , 3] , particularly the densely populated regions of East and South-East Asia [16] . However , once in a local inter-seasonal period , it is expected that migrant viruses will not be able to achieve onward transmission in the population due to unfavorable behavioral or environmental conditions . As a consequence , inter-seasonal occurrences of influenza are generally assumed to represent sporadic importations that do not play a major role in global viral transmission dynamics [14 , 17] . Theoretically , viral lineages may also persist over the inter-seasonal period in specific populations . To date , however , there is limited evidence for local multi-seasonal viral persistence . For example , phylogenetic analysis has provided evidence for the persistence of lineages of pandemic influenza A/H1N1pdm in West Africa ( although not within a single country ) [18] and Vietnam [19] , as well as highly pathogenic A/H5N1 avian influenza viruses , again in South-East Asia and Africa [20] . However , such instances are rare , with little evidence for the persistence of A/H3N2 influenza virus [16–17 , 21–22] , even in cases where there is evidence for transmission into the summer months [23] , although it is possible that this in part reflects poor or biased sampling . Herein we focus on the patterns and dynamics of inter-seasonal influenza in a single country–Australia–and ask how viruses sampled during this period are related to those sampled on a global scale . Previous studies of influenza in Australia have considered influenza circulation at a regional level with a focus on seasonal influenza , with relatively little consideration of inter-seasonal activity [15 , 24] . However , the availability of both epidemiological and sequence data makes Australia an ideal study site for a wider study of influenza seasonality , allowing us to reveal the origin and spread of influenza viruses during the inter-seasonal period . In light of the possible role played by climate in driving influenza seasonality [6] , we also sought to identify potential links between climatic factors and the occurrence of inter-seasonal influenza , a task made possible by the broad range of climatic zones in Australia ( Fig 1 ) . Accordingly , we performed an expansive phylogenetic analysis of a global data set of hemagglutinin ( HA ) glycoprotein coding sequences from over 17 , 500 influenza A/H1N1pdm , A/H3N2 and B virus sequences sampled over a five-year period from 2009–2014 , of which 396 ( 2 . 2% ) were sampled during the inter-seasonal period in Australia . In particular , we sought to determine whether these inter-seasonal viruses were seeded from global sources ( i . e . were imports ) or were part of a lineage able to transmit for extended periods within Australia , including temperate regions of the country . We focused on 2009–2014 as laboratory testing of samples increased following the 2009 human A/H1N1 pandemic ( H1N1pdm ) , yielding a relatively rich data set .
Full-length HA sequences were collected both during the temperate seasonal and inter-seasonal period in Australia during 1 Jan 2009–31 Jan 2014 . The seasonal model for influenza is perhaps less strictly descriptive of the shape and peak of influenza occurrence in Australia compared to wholly temperate countries in the northern hemisphere , although still of importance in predicting when the peak of influenza incidence will occur ( usually late July/August ) and when to implement vaccination campaigns . The influenza season in temperate Australia ( where the bulk of the population resides and which contains the main transportation hubs ) was therefore defined as occurring between 1 May and 31 October , particularly as surveillance activities focus on this as the peak time for influenza viral activity [24] . Consequently , inter-seasonal events were classified here as those that occurred between 1 November and 30 April . The more complex pattern in the less populous tropical northern Australia is then discussed in relation to these temperate seasonal boundaries . Respiratory samples or influenza virus isolates were collected from National Influenza Centres and laboratories across all eight Australian states and territories . Clinical samples and isolates were passaged in MDCK cells ( ATCC CCL-34 ) and resultant viruses had their HA genes sequenced as previously described [27] . Briefly , virus samples were cultured in Madin-Darby canine kidney ( MDCK ATCC CCL-34 ) cells as previously described [27] and RNA was extracted from the virus isolates using QIAamp Viral RNA Mini Kit ( Qiagen ) according to the manufacturer’s instructions . RT-PCR using MyTaq One-Step RT-PCR kit ( Bioline ) with subtype specific HA primers ( primer sequences available upon request ) . RT-PCR products were purified by ExoSAP-IT ( GE Healthcare ) and sequenced with Big Dye Terminator Reaction Mix ( Applied Biosystems ) and run on ABI 3500 XL following the manufacturer’s instructions . All sequences generated here have been deposited at the Global Initiative on Sharing all Influenza Data ( GISAID; http://platform . gisaid . org/ ) database and assigned accession numbers as listed in S1 Table . These data comprised 148 pandemic A/H1N1 ( A/H1N1pdm ) sequences , 66 A/H3N2 sequences , and 39 influenza B sequences . The sequences generated here were combined with 456 A/H1N1pdm , 274 A/H3N2 and 276 influenza B HA sequences from Australia sampled during the same time period ( 2009–2014 , including both seasonal and inter-seasonal data ) and downloaded from GISAID . To place the Australian sequences in the context of global influenza virus genetic diversity , data sets of full-length HA sequences from each influenza virus subtype ( A/H1N1pdm , A/H3N2 , and B ) sampled worldwide during 2009–2014 were compiled using GISAID and GenBank ( http://www . ncbi . nlm . nih . gov/genbank/ ) . Only sequences for which the collection date was known were included . Duplicate sequences ( i . e . collected from the same location on the same date and which appear to differ only in passage history ) were excluded to improve computational tractability . Consequently , final HA data sets of 9 , 912 A/H1N1pdm sequences , 3 , 804 A/H3N2 sequences and 3 , 941 influenza B virus sequences , with total numbers of Australian seasonal/inter-seasonal sequences of 546/149 for A/H1N1pdm , 244/108 for A/H3N2 , and 292/139 for influenza B , were utilized for phylogenetic analysis . Sequences of each influenza subtype were aligned separately using the MAAFT program ( v7 . 017; [28] ) available through Geneious ( v7 . 1 . 3; http://www . geneious . com/ ) , followed by manual adjustment . To obtain an initial view of the phylogenetic relationships of each of the three data sets , phylogenetic trees were inferred using the maximum likelihood ( ML ) method available in RAxML [29] . This returned the best tree from 20 replicates inferred using the general time-reversible ( GTR ) substitution model with a gamma ( Γ ) distribution of among-site rate variation . All parameter values were estimated from the empirical data ( available upon request ) . To assess the reliability of key nodes ( i . e . those pertinent to inter-seasonal influenza virus transmission in Australia ) , a second ML phylogenetic analysis was undertaken on clades comprising largely Australian samples , this time incorporating bootstrap resampling . Accordingly , clades were selected for further analysis if they contained isolates spanning a date range outside of the influenza surveillance season . We then inferred 1000 replicate ML trees in PhyML [30] using the GTR+Γ substitution model described above . The resulting phylogenies were visualized using FigTree ( v1 . 4 . 1; available at: http://tree . bio . ed . ac . uk/software/figtree/ ) . Only bootstrap values >70% were considered as statistically robust [31] . To determine the epidemiological context of influenza viruses sampled inter-seasonally in Australia , we classified them into five categories ( denoted a-e ) based on their phylogenetic position within the trees described above ( schematic epidemiological patterns are shown in Fig 2; see below ) . We assumed that Australian sequences that cluster together in the context of the global background denoted transmission events that most likely occurred within Australia , although every event type ( category ) necessarily began with an importation into Australia . Finally , geographic location data ( to the nearest town/city ) was available for the majority of the inter-seasonal sequences , and was used as a proxy for location of viral infection . The five categories of epidemiological events were: Import: If a single Australian isolate fell in a clade containing only globally sampled sequences , with no close phylogenetic relationship to other Australian sequences , then it was considered to result from an importation causing only limited onward transmission in Australia . Although most imports were only sampled once , and hence may have experienced only limited transmission in Australia , some comprised lineages of multiple sequences such that they were clearly transmitted for extended periods inter-seasonally . Although , as noted above , all transmission events studied here began with an importation event , this category as defined here only considers transmissions within the inter-seasonal period . Herald: If an Australian isolate was closely related to global sequences , but clustered with and preceded a number of other Australian isolates from the forthcoming influenza season , it was considered to be part of a “herald” event . Hence , we assumed that herald lineages persisted from within the inter-seasonal period into the next influenza season following an international importation . Straggler: If an inter-seasonal Australian isolate exhibited a close phylogenetic relationship to Australian isolates from the preceding influenza season , but did not persist into the following season , then it was considered to be part of a “straggler” event . Persistent: Isolates were classified as being part of a persistence event if they were part of a cluster of Australian isolates that occurred inter-seasonally , with only a limited number of related global sequences , and which were related to those from both the preceding and subsequent seasons such that they are indicative of continual local inter-seasonal transmission . Two-tailed: Finally , viral lineages that transmitted before , during and after the normal influenza season , such that they contain herald , seasonal and straggler components but which were not classified as persistent as they were not maintained throughout the inter-seasonal period , were termed “two-tailed” lineages . To explore the effect of climate on influenza seasonality in Australia , we used the Australian Bureau of Meteorology’s climate classifications . These recognize three commonly used methods for classifying the climate of mainland Australia: ( i ) temperature/humidity maps , compiled from data collected nationally over the period 1961–1990 and comprising six key zones; ( ii ) modified Köppen maps , which again show six major groups , but with 27 sub-classifications; and ( iii ) seasonal rainfall levels , which again identify six major climate groups [25] . The six key groups identified by the temperature/humidity maps were used as the basis of climate classification for cities and towns in this study ( Fig 1 ) . The locations of interest here were classified consistently in all three methods , with the only differences reflecting levels of rainfall and summer temperatures . For instance , Townsville and Darwin are both classified as “tropical” in the Köppen system and as having “hot , humid summers” in the temperature/humidity zones , whereas Sydney and Canberra are both classified as “temperate” in the Köppen system , but their summers are differentially classified as “warm” and “mild/warm” in the temperature/humidity zones [25] .
Our phylogenetic analysis of 17 , 657 HA gene sequences sampled globally during 2009–2014 ( S1–S3 Data ) revealed that multiple lineages of each influenza virus established extended transmission chains in both temperate and tropical Australia during inter-seasonal periods , including persistence in the case of influenza B viruses ( Table 1 ) . To examine the patterns of inter-seasonal transmission and evolution in more detail , we performed focused phylogenetic analyses of each possible transmission type ( Table 1 and Fig 3 ) . All five types of inter-seasonal events–import , herald , straggler , persistent , two-tailed–were observed . Of these , importation was the most common and represented 58 . 4% of all inter-seasonal sequences in A/H1N1pdm , 50 . 9% in A/H3N2 , and 37 . 4% in influenza B virus ( Table 1 ) . Frequencies for each event type seemingly differed between influenza A subtypes A/H1N1pdm and A/H3N2 , and between influenza A and B ( Table 1 ) . However , because of relatively limited and non-systematic sampling available here it was unclear whether this represented a fundamental difference in epidemiological dynamics between subtypes [17] or ascertainment bias . Although influenza transmission events have previously been documented outside of seasonal boundaries [23 , 32] , the inter-seasonal events documented here were lengthier than expected , included those of potential evolutionary importance such as local persistence , and sometimes involved viruses sampled from both tropical and temperate regions within Australia . According to our phylogenetic analysis , 70 Australian viruses appeared in herald events , denoting viruses that appeared prior to the local influenza season and continued to transmit into the full season , while 115 could be classified as stragglers that continued to transmit beyond the end of the season . Notably , although we observed imports , stragglers , and heralds within Australia for both A/H3N2 and A/H1N1pdm , no evidence for inter-seasonal persistence was observed in either virus . Rather , the best evidence for persistence came from a clade of influenza B virus ( Victoria lineage ) that included 92 Australian isolates , of which 17 were collected during the summer of 2010/2011 ( Fig 3D ) . This clade appeared to persist locally within the 2010 season ( earliest date isolated: 11 August 2010 , sequence B/Victoria/503/2010 ) and through the 2011 season ( latest date isolated: 13 October 2011 , B/Sydney/34/2011 ) , with most major branches showing bootstrap support values over 70% . Importantly , eight isolates from this event , spanning an eight-month period linking the 2010 and 2011 seasons , clustered strongly together to the exclusion of non-Australian sequences ( 94% bootstrap support ) . The earliest virus within this group was isolated in Townsville ( in tropical northern Australia ) at the end of the 2010 season ( 14 October 2010 ) . Three inter-seasonal isolates followed , two from Darwin ( also located in tropical northern Australia ) in January 2011 and another from Townsville in April of that year . In the 2011 season , the earliest sequence was again from Townsville ( 02 May 2011 ) , followed by isolations in South Australia ( a temperate region; 01 June 2011 ) , Brisbane ( subtropical; 08 June 2011 ) and Townsville ( 20 June 2011 ) . Earlier sequences , which appeared to be related to this cluster albeit with weaker bootstrap support , were all from 2010 , and came from Sydney ( temperate ) , and Brisbane . Despite incomplete global sampling , clearly the most parsimonious explanation with the phylogenetic data in hand is that the influenza B viruses in question have transmitted locally for the full duration of the Australian summer , and spread within both tropical and temperate regions . The 2010/2011 , 2012/2013 and 2013/2014 inter-seasonal periods in Australia were characterized by increased numbers of laboratory-confirmed notifications of influenza to the National Notifiable Diseases Surveillance System ( NNDSS ) ( Fig 4 ) . It was previously concluded that the high number of notifications between the 2010 and 2011 seasons likely reflected a genuine increase in disease , magnified by increased laboratory testing of samples following the 2009 H1N1 pandemic , and that this may have been a normal fluctuation in levels of inter-seasonal influenza [15] . Notably , although there was no obvious association between the number of influenza notifications and the types of inter-seasonal transmission events that occurred , our phylogenetic analysis did reveal the presence of a persistent lineage of influenza B virus and two-tailed seasonal lineages of influenza A/H1N1 during the 2010/2011 inter-seasonal period , as well as frequent other inter-seasonal transmission events . The availability of geographical data enabled us to determine whether there was an association between the inter-seasonal transmission and climatic zone within Australia ( Table 1 ) . Although the majority of the Australian population resides in temperate regions , our analysis is striking in that the proportions for inter-seasonal influenza were strongly skewed towards sub-tropical and tropical zones , with a disproportionate number of inter-seasonal events falling within Australia’s tropical and subtropical regions ( 98 and 101 sequences , respectively ) , as opposed to the temperate regions ( 197 sequences ) ( Table 1 ) . Accordingly , the ratio of inter-seasonal isolates sampled in ‘temperate: subtropical: tropical’ zones is ‘2:1:1’ , while the size of the population in each of these zones is in the ratio 43:7:1 [26] . Although there are likely biases in those samples selected for sequencing , these would not necessarily result in an increase in apparent inter-seasonal transmission . Indeed , low levels of sampling are likely to under-estimate the true occurrence of inter-seasonal influenza transmission , such that these results err on the conservative side . Accordingly , this result , coupled with the appearance of persistent virus transmission events centering on tropical Darwin and Townsville , suggests that there may be a link between viral persistence and climate . In particular , while the seasonal portions of these extended transmission chains tend to occur in temperate regions , the inter-seasonal transmission period appears to be more frequent in hot and humid localities , also suggesting that there is a high degree of connectivity between these regions .
Our study benefits from the fact that we were able to examine the inter-seasonal dynamics of influenza virus in climatically diverse areas . Indeed , our study was noteworthy in that we observed both the inter-seasonal transmission of influenza in temperate regions of Australia ( including the most southerly state of Tasmania; Fig 3E ) , and lengthy inter-seasonal transmission chains centered on sparsely populated tropical and sub-tropical areas , although often closely related to seasonally transmitting viruses from temperate regions . Although this in part clearly reflects aspects of population mobility , including fly-in and fly-out workers , if population mobility were the major factor explaining persistence we might expect to see more frequent persistence in temperate zones where the populations are large , dense , and mobile ( including the main national and international transportation hubs ) . Hence , that we observed greater levels of extended inter-seasonal transmission in the tropical and sub-tropical zones where populations are smaller and less dense is more consistent with a climate-driven effect . However , it is evident that further surveillance of tropical , sub-tropical and other non-temperate climatic zones is needed year-round to ascertain whether persistence can be definitively linked to certain climatic zones . The climate-driven model of influenza transmission generally considers absolute temperature and relative humidity ( RH ) to be the driving factors of influenza transmission , although it is has also been suggested that absolute humidity ( AH ) is a better predictor of influenza virus survival and transmission , or moderates transmission mechanisms [11 , 37] . We necessarily focused on RH as these are the data provided by the Australian Bureau of Meteorology [11 , 25] , so that the role of AH is difficult to assess here . In tropical Darwin where we observe viral persistence , temperature and RH do not vary widely , with average daily temperatures ranging from only 30 to 33°C throughout the year [25] . The most important difference between the seasonal and inter-seasonal period in Darwin is in the amount of rainfall . From May–October an average monthly rainfall of 14 . 1 mm is recorded , whereas the equivalent value for November–April is 241 . 4 mm [25] . Other measures of climatic variability , such as temperature and relative humidity , seem not to differ greatly between the seasonal and inter-seasonal periods . For example , the average maximum/minimum temperature for May–October is 32 . 2°C/22 . 07°C , while for November–April it is 33 . 1°C/25 . 1°C , with the average 9am/3pm readings for RH at 63 . 17%/48% from May–October and 74 . 83%/65 . 67% for November–April [25] . Experimental studies using guinea pig models have found that the aerosol transmission of influenza virus was blocked or inefficient at 30°C and intermediate-high humidity ( 50–80% RH ) [11 , 12] , such that it was favored in cool and dry conditions . It is therefore unclear how influenza virus transmission occurs in the tropics . It is possible that much transmission occurs indoors , for example mediated by air-conditioning , or that contact transmission is more efficient in warm humid conditions , for instance if droplets of mucus that contain virus desiccate at a lower rate in high humidity . Studies on the relationship between absolute humidity and influenza virus survival and transmission have provided evidence for increased survival and transmission at both low and high absolute humidity , suggesting a potential bimodal relationship [37 , 38] . Large-scale epidemiological studies have also shown that the association between peaks in influenza activity and climatic variables , such as temperature , humidity and solar radiation , varies with latitude , being strongest at latitudes higher than 25°S , and no significant association between 12 . 5–25°S [39] . Within our study population , Darwin sits at 12 . 4°S , Townsville at 19°S , Brisbane at 27°S , while all temperate cities are at higher latitudes ( at around 33°S ) . Accordingly , there should be no significant association between influenza peaks and climatic variables in the Australian tropics . Clearly , further work is needed to determine whether the populations in tropical and sub-tropical areas have high rates of inter-seasonal influenza due to increased host susceptibility , or environmental factors specific to their location , or some combination of these factors . Previous studies of influenza seasonality [40 , 41] have described Australia as a “temperate” country , presumably as the most populous cities are located within temperate zones . However , this does not take into account Australia’s large climatic diversity , nor the epidemiology of influenza within populations outside temperate zones which can be more complex than a single seasonal peak . Whatever the classification method utilized , it is important that studies of influenza epidemiology in Australia reflect its climatic complexity , particularly as models of influenza seasonality are being used to inform vaccination strategies within Australia [42 , 43] . Influenza circulation is clearly complex , involving an interplay between climate , viral movement , population mobility , and aspects of population immunity and susceptibility . We have revealed an unexpectedly important role for inter-seasonal influenza transmission in both tropical and temperate regions . While it is clear that the vast majority of influenza cases occur during the temperate influenza season , the thresholds of these seasons may change with variations in seasonal climatic factors such as temperature and humidity year-to-year . A greater focus on the occurrence and determinants of inter-seasonal influenza may provide data central to determining the key drivers of influenza seasonality .
|
Human influenza virus commonly causes disease in the winter months of temperate countries , but exhibits more complex patterns in tropical localities . Most studies of this complex seasonality have only considered viruses sampled within the “normal” influenza season . To help reveal the drivers of influenza seasonality we utilized viruses sampled outside of the normal influenza season , focusing on Australia which is characterized by a wide range of climates . Using a phylogenetic approach we revealed more complex patterns of influenza transmission than previously anticipated , particularly that the virus is able to transmit for extended periods and even persist locally within Australia throughout the virus “off-season” . In addition , we found that inter-seasonal influenza was more frequent in tropical and sub-tropical than temperate regions , adding weight to theories that climate likely plays an important role in influenza seasonality .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Inter-Seasonal Influenza is Characterized by Extended Virus Transmission and Persistence
|
The Escherichia coli transcription system is the best characterized from a biochemical and genetic point of view and has served as a model system . Nevertheless , a molecular understanding of the details of E . coli transcription and its regulation , and therefore its full exploitation as a model system , has been hampered by the absence of high-resolution structural information on E . coli RNA polymerase ( RNAP ) . We use a combination of approaches , including high-resolution X-ray crystallography , ab initio structural prediction , homology modeling , and single-particle cryo-electron microscopy , to generate complete atomic models of E . coli core RNAP and an E . coli RNAP ternary elongation complex . The detailed and comprehensive structural descriptions can be used to help interpret previous biochemical and genetic data in a new light and provide a structural framework for designing experiments to understand the function of the E . coli lineage-specific insertions and their role in the E . coli transcription program .
RNA in all cellular organisms is synthesized by a complex molecular machine , the DNA-dependent RNA polymerase ( RNAP ) . In bacteria , the catalytically competent core RNAP ( subunit composition α2ββ'ω ) has a molecular mass of ∼400 kDa . Evolutionary relationships for each of the bacterial core subunits have been identified between all organisms from bacteria to man [1]–[3] . These relationships are particularly strong between the two largest subunits , β' and β , which contain colinearly arranged segments of conserved sequence ( Figure 1 ) [3] . These conserved segments are separated by relatively nonconserved spacer regions in which large , lineage-specific gaps or insertions can occur [3] , [4] . The functional significance of these lineage-specific differences is poorly understood due to a lack of correlated biochemical and structural information . The bulk of our biochemical and genetic knowledge on bacterial RNAP comes from studies of Escherichia coli ( Eco ) RNAP but all of our high-resolution structural information comes form Thermus RNAPs [5]–[8] as Eco RNAP has not been amenable to X-ray crystallography analysis . The Eco and Thermus β and β' subunits harbor large sequence insertions ( >40 amino acids ) that are not present in the other species and are not shared across bacterial species ( Figure 1 ) [3] . For example , the Eco β' subunit contains β'-insert-6 ( or β'i6 , using the lineage-specific insert nomenclature of Lane et al . [3] ) , a 188-residue insertion in the middle of the highly conserved “trigger loop . ” On the other hand , the Thermus β' subunit lacks β'i6 but contains β'i2 ( 283 residues ) . High-resolution structures of both of these lineage-specific inserts reveal that they comprise repeats of a previously characterized fold , the sandwich-barrel hybrid motif ( SBHM ) [9] , [10] . Similarly , the Eco β subunit harbors three large insertions missing in Thermus , βi4 ( 119 residues ) , βi9 ( 99 residues ) , and βi11 ( 54 residues ) , whereas the Thermus β subunit harbors βi12 ( 43 residues ) . In some respects , the high-resolution Thermus RNAP structures have served as good models to interpret the functional literature obtained from biochemical , biophysical , and genetic studies of Eco RNAP [11] , [12] . Nevertheless , a complete molecular model of Eco core RNAP has not been available due to the absence of high-resolution structural information on the Eco β subunit lineage-specific inserts . The most detailed structural studies of Eco RNAP have come from cryo-electron microscopy ( cryo-EM ) analysis of helical crystals at about 15 Å-resolution [13] . This cryo-EM reconstruction of Eco core RNAP could be interpreted in detail by fitting the Taq core RNAP X-ray structure , revealing a large distortion of the structure ( opening of the active site channel by more than 20 Å ) due to intermolecular contacts in the helical crystals . Strong electron density for Eco βi9 was present in the cryo-EM reconstruction , but weak density for Eco βi4 and Eco β'i6 indicated these domains were flexible in the context of the helical crystals [13] . Most previous EM reconstructions of various forms of Eco RNAP have not revealed information concerning the lineage-specific inserts ( for instance , see [14] ) . A recent 20 Å-resolution , negative-stain EM reconstruction of an activator-dependent transcription initiation complex containing Eco RNAP [15] allowed the positioning of the Eco β'i6 crystal structure [10] , but the lack of structural information on the other Eco lineage-specific inserts prevented the detailed interpretation of additional densities present in the reconstruction [15] . In this study , we used a combination of structural approaches to generate a complete molecular model of Eco core RNAP . We determined two new high-resolution X-ray crystal structures of Eco RNAP β subunit fragments that include Eco βi4 and βi9 and used an ab initio method to predict the structure of the small Eco βi11 [16] . The three available X-ray crystal structures of Eco RNAP fragments ( the two structures determined herein and the structure of Eco β'i6 [10] ) and the predicted structure of Eco βi11 were incorporated into a homology model of Eco core RNAP . Finally , we used cryo-EM imaging combined with single-particle image analysis to obtain a low-resolution structure of the solution conformation of Eco core RNAP in which densities corresponding to lineage-specific insertions could be clearly identified . Flexible-fitting of the Eco RNAP homology model into cryo-EM densities generated a complete molecular model of Eco core RNAP and an Eco RNAP ternary elongation complex ( TEC ) .
The lineage-specific insert βi4 ( previously named β dispensable region 1 , or βDR1 , or SI1 in the literature [13] , [17] , [18] ) , located between bacterial shared regions βb6 and βb7 ( using the bacterial RNAP common region nomenclature of Lane et al . [3] ) in the β2 domain ( Figure 1 ) [5] , [19] , was predicted to comprise from one to six tandem repeats of a structural motif termed the β-β' module 2 ( BBM2 ) [4] . The βi4 of Acidobacteria , Mollicutes , and Proteobacteria ( including Eco ) was predicted to comprise two tandem BBM2 repeats [3] . Eco βi4 comprises β residues 225–343 ( Figure 2A ) . We prepared a construct comprising the Eco β2 domain including βi4 inserted within it ( Eco β residues 152–443 , hereafter called Eco β2-βi4 ) . After reductive methylation [20] , the protein formed crystals that diffracted X-rays to 1 . 6 Å-resolution ( Table 1 ) . The structure was solved by single-anomalous dispersion using a dataset collected from crystals of selenomethionyl-substituted protein [21] and refined to an R/Rfree of 0 . 209/0 . 229 at 1 . 6 Å-resolution ( Table 1 , Figures 2 , S1 ) . As expected , the Eco β2 ( Eco β residues 151–224 and 344–445 ) and the Thermus β2 ( Taq or Tth β residues 138–325 ) domains have similar overall structures ( Figure S2 ) . A superimposition of the two domains over 100 residues ( excluding flexible loops connecting secondary structural elements ) yields a root-mean-square deviation in α-carbon positions of 1 . 68 Å . Significant differences in the structures include: ( i ) the loop connecting the first two β-strands of the β2 domain , where Eco has a 5-residue insertion ( Eco β residues 164–168 , disordered in our structure ) , and ( ii ) the loop connecting the last two α-helices of the β2 domain , which includes a 7-residue insertion present in Taq β ( Taq β residues 293–299; Figures 2A , S2 ) . The βi4 domain is inserted at the surface of the β2 domain distal to the connection with the RNAP ( Figure 2B ) . A 3-residue segment of Taq β ( Taq β 212–214 ) is replaced by the 119-residue Eco βi4 ( Figure 2A ) . The Eco βi4 folds into a compact , cylinder-shaped domain about 22 Å in diameter and about 50 Å in length ( Figures 2B , 2C ) . The compact domain is connected to the β2 domain by two short connector loops ( Eco β 225–226 and 337–345 ) . The βi4 domain packs against β2 , resulting in the burial of a modest 618 Å2 of surface area . As predicted [4] , the Eco βi4 includes two tandem BBM2 motifs ( Figure 2A , 2C ) . The lineage-specific insert βi9 ( previously named β dispensable region 2 , or βDR2 , or SI2 in the literature [13] , [18] , [22] , [23] ) is located between bacterial shared regions βb13 and βb14 [3] at the base of the flap domain ( Figure 1 ) [5] , [19] . The βi9 is found in Acidobacteria , Aquificae , Bacteriodetes , Chlamydiae , Chlorobi , Planctomycetes , Proteobacteria ( including Eco ) , and Nitrospirae [3] . Eco βi9 comprises β residues 938–1042 ( Figure 3A ) . A construct comprising the Eco flap domain ( Eco β 831–1057 ) , including βi9 , was crystallized as a complex with bacteriophage T4 gp33 ( K . -A . F . T . , P . Deighan , S . Nechaev , A . Hochschild , E . P . Geiduschek , S . A . D . , in preparation ) . The structure was solved by a combination of molecular replacement ( using the Taq flap domain as a search model ) and single-anomalous dispersion using data collected from selenomethionyl-substituted protein ( Table S1 , Figure S3 ) [21] . The complete structure was refined to an R/Rfree of 0 . 264/0 . 291 at 3 . 0 Å-resolution . T4 gp33 interacts primarily with the flap-tip and does not make any interactions with βi9 . These and further details of the complex with T4 gp33 will be described elsewhere ( K . -A . F . T . , P . Deighan , S . Nechaev , A . Hochschild , E . P . Geiduschek , S . A . D . , in preparation ) . The βi9 domain is inserted at the base of the flap domain , near the C-terminal connection of the flap with the rest of the RNAP and distal to the flap-tip ( Figure 3B ) . A 6-residue segment of Taq β ( Taq β 809–814 ) is replaced by the 105-residue Eco βi9 ( Figure 3A ) . The Eco βi9 comprises two long , parallel α-helices of 38 and 32 residues ( Eco β 943–980 and 1006–1037 , respectively ) with a short , hook-like connecting segment ( residues 981–1005 ) at the end distal to the flap ( Figure 3B ) , forming an apparently rigid structure reminiscent of a hook-and-ladder that extends nearly 65 Å out from the flap domain . The βi9 is connected to the flap domain by two connector loops ( Eco β 938–942 and 1038–142 ) but makes minimal interactions with the flap itself . The structure does not appear to conform to the β-β' module 1 motif ( BBM1 , similar to the BBM2 motif , Figure 2C ) predicted for βi9 [4] . The 105-residue Eco βi9 is at the lower end of the size range for βi9 sequences , which ranges from 105 residues in some Proteobacteria to 143 residues in some Bacteriodetes . An alignment of 307 non-redundant βi9 sequences ( see Dataset S1 ) reveals that the two long , ladder α-helices do not harbor insertions; all of the insertions occur in the hook-like connector at the distal end of βi9 ( Figure 3A ) . Therefore , we conclude that βi9 has a conserved core structure with the two ladder α-helices of conserved length . We generated a single-particle cryo-EM ( spEM ) reconstruction of Eco RNAP by analyzing ∼42 , 000 images of Eco RNAP particles preserved in vitreous ice ( Figures 4A , S4–S6 ) . Initial image orientation parameters were determined using a 35 Å-resolution RNAP model based on the Taq core RNAP X-ray structure [5] . Final refinement of image orientation parameters by projection matching yielded a structure of Eco RNAP with a 0 . 5 Fourier-shell cutoff resolution of ∼11 . 2 Å ( Figure S4 ) . Nevertheless , information beyond about 14 Å resolution was very weak , and so the figures and analysis described herein were performed on a low-pass Fourier-filtered map [24] , [25] . Although the cryo-EM grids were prepared with samples of Eco RNAP holoenzyme ( core RNAP plus the promoter-specificity σ70 subunit ) , the σ70 subunit apparently dissociated during grid preparation as density corresponding to σ70 was completely absent . Dissociation during cryo-EM sample preparation has been noted for other macromolecular complexes [26] and is also consistent with reports of dissociation constants for the σ70/core RNAP complex as high as 200–300 nM ( the RNAP concentration used here was about 200 nM ) . The spEM reconstruction showed Eco core RNAP in a conformation similar to that observed in Thermus X-ray structures but with clear density corresponding to βi4 , βi11 , and β'i6 ( Figures 4A , S5 , S6 ) . In order to interpret the spEM map of Eco core RNAP , we generated a homology model of Eco core RNAP using the core component of the T . thermophilus ( Tth ) RNAP holoenzyme structure ( PDB ID 1IW7 ) [7] as a template . The locations of the Eco lineage-specific insertions βi4 , βi9 , βi11 , and β'i6 ( absent in Thermus ) were left as gaps in the Eco sequences . Thermus-specific inserts βi12 and β'i2 ( Figure 1 ) were also removed from the structural template . The crystal structures of Eco β2-βi4 ( Figure 2B ) and βflap-βi9 ( Figure 3B ) were spliced into the resulting homology model by superimposition of the overlapping β2 and βflap domains , respectively . At this stage , the Eco RNAP model was readily fit manually into the spEM map . The spEM map contained clear density corresponding to βi4 , but density for βi9 was absent . Density for the ω subunit as well as the C-terminal helix of β' were also absent . In addition , extra density not accounted for by the homology model was present for βi11 and β'i6 . An ab initio predicted structure of the short βi11 ( see below ) was placed into the corresponding density to fill in the gap in the Eco β sequence between 1121 and 1181 . The crystal structure of Eco β'i6 ( PDB ID 2AUK ) [10] was readily fit manually into excess density in the vicinity of its insertion point in β' . Two criteria were used to determine the orientation of β'i6 with respect to the rest of the RNAP . First , although β'i6 comprises a tandem repeat of two SBHM domains , the C-terminal SBHM domain ( SBHMb ) [10] harbors larger insertions between the core SBHM β-strands , making β'i6 asymmetric in shape . The asymmetry is clearly seen in the spEM density as well ( see Figure 4A , top view ) . Moreover , only one orientation of β'i6 allows connection to the gap in the Eco β' sequence ( between residues 940 and 1132 ) without severe distortion . The positioned β'i6 was readily connected to the open ( unfolded ) trigger-loop ( TL ) conformation of the model . Flexible-fitting of the final Eco RNAP model ( excluding ω , the C-terminal 41 residues of β' , and βi9 ) into the spEM map was performed using YUP . SCX [27] , resulting in a superb fit of the conserved RNAP as well as of the lineage-specific inserts ( excluding βi9; Figures 4A , S5 , S6 ) . In order to position βi9 in the context of the entire RNAP structure , we used our previously determined helical cryo-EM map of Eco core RNAP ( hEM ) and fit of the Taq core RNAP X-ray crystal structure [13] since the hEM map contains strong density for βi9 . The βflap portion ( excluding the flexible flap-tip ) of the Eco βflap-βi9 crystal structure ( Figure 3B ) was superimposed on the Taq βflap domain in the context of the Taq RNAP fit into the hEM density . The resulting position of βi9 did not correspond to the hEM density ( light orange , βi9 in Figure 4B ) but was fit into the density by a rotation of about 35° ( orange , βi9' in Figure 4B ) . This positioning of βi9 is consistent with the location of positive difference density observed in the context of the helical crystals due to a 234-residue insertion between Eco β residues 998 and 999 ( red dot , Figure 4B ) . The Eco core RNAP model was completed by adding back the C-terminal segment of β' as well as ω ( in accordance with the Thermus RNAP structures ) . The Eco core RNAP model was then used as the basis for generating a homology model of an Eco TEC , using the Tth TEC crystal structure ( open TL conformation , PDB ID 2O5I ) [8] . For both models , the lineage-specific inserts ( βi4 , βi9 , βi11 , β'i6 for Eco; β'i2 and β'i12 for Tth ) were removed . The nucleic acids present in the Tth crystal structure were fixed during the modeling . The Eco lineage-specific inserts were added back to the resulting TEC model ( according to their positions in the Eco core RNAP model ) , and missing portions of the nucleic acids ( the upstream double-stranded DNA , and the nontemplate strand of the DNA within the transcription bubble ) were modeled according to Korzheva et al . [28] .
RNAPs harboring deletions or insertions within βi4 support cell growth and retain basic in vitro transcription function , leading to its designation as “dispensable region I” of the β subunit [17] . Nevertheless , careful studies of a nearly precise βi4 deletion ( deletion of Eco β 226–350 ) revealed defects [18] . The purified Δβi4-RNAP showed only very mild defects , or no defects at all , in a number of in vitro tests [17] , [18] . In vivo , however , the Δβi4-RNAP was unable to support cell growth at 42°C and could only support slow growth at 30°C . In our model of the Eco TEC , βi4 extends out from the β2 domain roughly in the direction of the downstream double-stranded DNA ( Figure 5 ) . However , βi4 is unlikely to interact directly with the downstream DNA to form part of an extended DNA binding channel since βi4 tilts away from the DNA , creating a roughly 15 Å gap between itself and the DNA . Moreover , the solvent-exposed surface of βi4 , including the entire surface facing the DNA , is highly acidic ( Figure 5 , front view ) , except for a “neutral patch” that arises from three conserved residues , Eco β R268 , R272 , and R275 ( Figure 5 , top view ) . These positions are conserved as basic residues ( either R or K ) in 98% , 91% , and 91% of the sequences , respectively , in an alignment of 316 non-redundant βi4 sequences ( containing only “Eco-like” βi4 sequences comprising two BBM2 domains; see Dataset S2 ) and may comprise an interaction determinant for an as yet unidentified regulatory factor . The bacteriophage T4 Alc protein interacts with the host Eco RNAP [30] and causes premature transcription termination on Eco DNA while allowing Eco RNAP-mediated transcription of phage DNA containing 5-hydroxymethylcytosine [31] . Eco paf mutants ( prevent Alc function ) have been mapped to the rpoB gene encoding the RNAP β subunit [17] , [32] . Eco β mutants R368H , R368C , and a double mutant ( P345S/P372L ) display the paf phenotype , possibly by directly preventing Alc interaction with RNAP [17] . These mutations lie within a region of the β subunit that could be deleted without disrupting basic transcription function [17] but are not , in fact , contained within βi4 ( Figure 2A ) . Two of the mutated positions ( 368 and 372 ) lie within βb7 , a region shared among all bacterial RNAPs ( Figure 2A ) [3] . In our structural model of the Eco RNAP TEC , βR368 and βP372 lie within a structural feature that sits at the entrance of the main RNAP active site channel , inside the “V” formed by the upstream and downstream DNA of the TEC ( Figure 5 , channel and front views ) . These residues are not near any nucleic acids in the TEC ( the closest approach is for the backbone carbonyl of βP372 , which is 15 Å away from the nontemplate DNA phosphate backbone at the -10 position ) but could comprise part of an Alc binding determinant on the RNAP [17] . The 19 kDa Alc protein bound in this vicinity ( Figure 5 , channel and front views ) would be well positioned to distinguish the presence of cytosine or 5-hydroxymethylcytosine in either the downstream double-stranded DNA ( where the 5-hydroxymethyl moiety would be exposed in the major groove ) or the single-stranded non-template DNA in the transcription bubble . RNAPs harboring deletions or insertions within βi9 support cell growth and retain in vitro transcription function , leading to its designation as “dispensable region II” of the β subunit [17] , [22] , [23] , [33] . Nevertheless , careful studies of a precise βi9 deletion ( deletion of Eco β 938–1040 ) revealed defects [18] . The purified Δβi9-RNAP showed only very mild defects , or no defects at all , in a number of in vitro tests [18] . The βi9 contains the epitope for the PYN-6 monoclonal antibody and , consistent with in vitro tests showing little effect of deleting βi9 on normal RNAP function , RNAP can be immobilized using the PYN-6 antibody but remains active for in vitro transcription [22] . In vivo , however , the Δβi9-RNAP was unable to support cell growth in minimal media [18] . Our crystal structure of the Eco βflap-βi9 suggests that βi9 is attached to the flap via flexible linkers and does not make a significant , stable interaction with the flap ( Figure 3B ) , suggesting that βi9 is highly flexible in its orientation with respect to the flap . Indeed , the position of βi9 in the βflap-βi9 crystal structure appears to be determined by packing interactions with neighboring , symmetry-related molecules . In keeping with this , there is no density for βi9 in the spEM reconstruction ( Figures 4A , S5 , S6 ) . However , in our previous hEM reconstruction of Eco RNAP , strong density consistent with βi9 was observed , and this density was shown to correspond to βi9 through a helical reconstruction of a mutant RNAP harboring a large insertion between positions 998 and 999 [23] . In the helical crystals , the packing of a neighboring , symmetry-related RNAP molecule restricts the range of positions available to βi9 , allowing its visualization ( Figure 4B ) . Fitting βi9 into the corresponding density in the hEM reconstruction required a large change in the position of βi9 with respect to the flap , but the final model fits very well into the density and is also consistent with the EM localization results [23] , which were not used as a constraint in the fitting ( Figure 4B ) . This model for the position of βi9 in the context of the entire RNAP is presented as an example of a particular orientation that is possible for βi9 ( since it was observed in the helical crystals ) , but the evidence indicates that βi9 does not adopt a particular conformation with respect to the RNAP but can access a wide range of positions ( Figure 6 ) . The modeled position of βi9 is not near any nucleic acids in the TEC or in the open promoter complex [34] . Moreover , the solvent-exposed surface of βi9 is primarily acidic ( Figure S7 ) . Interestingly , an alignment of 307 non-redundant βi9 sequences ( see Dataset S1 ) reveals that conserved , solvent-exposed residues are all displayed on the back face of the “ladder , ” opposite the “hook” ( Figure S7 ) . Conserved features of this face comprise charged residues D959 ( conserved as D or E in 97% of the sequences ) , E962 ( D/E , 95% ) , R974 ( K/R , 89% ) , K1032 ( K/R , 95% ) , and K1035 ( K/R , 94% ) , and one conserved hydrophobic residue , I966 . These features suggest that this face of the ladder may serve as an interaction determinant for as yet unidentified regulatory factors . D959 and K1032 participate in an apparently conserved salt bridge . Predictably , a number of conserved hydrophobic residues participate in the hydrophobic core of the domain , either between the ladder and the hook ( L979 , L989 ) or in the packing interface between the two ladder helices ( L1029 , I1036 ) . The lineage-specific insert βi11 is located between bacterial shared regions βb14 and βb15 ( Figures 1 , 7A ) [3] . The βi11 is found in Acidobacteriaceae , Aquificae , and Proteobacteria ( including Eco ) [3] . In each bacterial species where it is found , βi11 has a length ranging from 54–69 residues . Comparing Taq with Eco , a 5-residue segment of Taq β ( Taq β 895–899 ) is replaced by the 59-residue Eco βi11 , comprising Eco β residues 1122–1180 ( Figure 7A ) . Although a construct corresponding to Eco RNAP βi11 overexpressed and was well behaved , we were unable to obtain crystals suitable for X-ray analysis . The Robetta server ( http://robetta . bakerlab . org/ ) provided an ab initio predicted structure of this short , 59-residue fragment ( Figure S8 ) that is consistent with a number of observations from our structural and sequence analyses: The βi11 was only recently recognized as a distinct , lineage-specific insertion [3] , [4] . To our knowledge , no information on the effects of deletions or mutations in this region is available . Inspection of the spEM map and the aligned X-ray structure of Taq core RNAP in the region of the β subunit between shared regions βb14 and βb16 revealed a clear discrepancy that corresponds to Taq βi12 ( Figure 7B ) . In our Eco RNAP model , the Taq βi12 was removed and the resulting gap was connected by the loop corresponding to Eco β residues 1200–1207 . The predicted structure of Eco βi11 ( Figure S8 ) was then spliced between Eco β residues 1121 and 1181 and oriented to fit into the EM density , resulting in a good fit . The resulting location of Eco βi11 clashed with the position of the β-subunit N-terminus , which was redirected to relieve the clash ( Figure 7B ) . While the large Eco lineage-specific insertions βi4 and βi9 appear to play only peripheral roles in RNAP function , and the complete deletion of either one results in relatively minor growth defects [18] , β'i6 plays a more important role in Eco RNAP function . Complete deletion , or even partial deletion , of β'i6 is not viable [18] , [35] . Complete deletion causes a severe defect in RNAP assembly , both in vivo and in vitro [18] , [35] , but the in vivo–assembled Δβ'i6-RNAP can be obtained from cells simultaneously overexpressing the other RNAP subunits [18] , and partial deletions of β'i6 can be assembled in vitro [35] . Biochemical studies of enzymes with complete or partial β'i6 deletions reveal a number of severe defects . The Δβ'i6-RNAP forms dramatically destabilized open promoter complexes [18] . RNAPs harboring partial deletions in β'i6 are defective in transcript cleavage and have a dramatically reduced transcript elongation rate at subsaturating NTP concentrations [35] . Antibody binding to epitopes within β'i6 inhibit transcription as well as intrinsic transcript cleavage [35] , [36] . The β'i6 plays a central role in the pausing/termination behavior of elongating Eco RNAP [18] , [35] . Full or partial deletions in β'i6 result in RNAPs with dramatically altered pausing behavior [18] , [35] . A genetic screen for termination-altering mutants in Eco RNAP uncovered 10 positions scattered throughout β'i6 [37] . These profound effects of β'i6 on Eco RNAP function are likely due to its insertion in the middle of a critical and highly conserved structural feature of the RNAP , the so-called “trigger-loop” ( TL ) , which connects two highly conserved α-helices ( TL-helices 1 and 2 , TLH1 and TLH2; Figures 1 , 8 ) . The TLHs , in turn , interact with another central structural element , the bridge-helix ( BH; Figure 8B ) . The TL tends to be unstructured ( open ) in RNAP and in the substrate-free TEC but is found in a structured conformation ( closed ) where it makes many direct contacts with the incoming NTP substrate in the TEC [38] , [39] . The TL has been proposed to cycle between open and closed conformations at each nucleotide addition step to promote rNTP substrate recognition , enzyme fidelity , and possibly catalysis [38]–[42] . Microcin J25 ( MccJ25 ) is a bactericidal 21-residue peptide that inhibits transcription by binding bacterial RNAP within the secondary channel [43]–[46] . Based on saturation mutagenesis of Eco rpoC ( the gene encoding the RNAP β' subunit ) , MccJ25 does not contact β'i6; most amino acid substitutions that yield strong resistance against MccJ25 lie in the BH and the TL [43] , [44] , [46] . Nevertheless , a deletion of β'i6 perturbs the effects of MccJ25 [46] , likely through the effects of the β'i6 deletion on the TL conformation . Our positioning of β'i6 in the spEM density ( Figures 4 , S5 , S6 ) and its connections with the open TL conformation ( Figure 8B ) are similar to the results of Hudson et al . [15] . The β'i6 sits outside the RNAP active site channel and makes extensive interactions with the β'-jaw ( Figure 8B ) . The N-terminal SBHM domain of β'i6 ( SBHMa ) faces the secondary channel , consistent with the results of crosslinks mapped from backtracked TECs ( in which the 3′-end of the RNA transcript is extruded out the secondary channel ) between analogs incorporated into the RNA 3′-end and the N-terminal region of β'i6 [28] . SBHMb faces the downstream double-stranded DNA-binding channel ( Figures 5 , 8 ) but does not contact the DNA; the closest approach between the DNA and β'i6 is 16 Å ( between β'D1073 and the nontemplate strand backbone phosphate at +14 ) . Moreover , β'i6 is highly acidic over its entire solvent-exposed surface , including the region facing the downstream double-stranded DNA ( Figure 5 , front view ) . Although β'i6 connects readily to the open conformation of the TL via extended linkers ( Figure 8B ) , modeling suggests it would not be able to connect with the closed TL conformation in the modeled position , a conclusion also reached by Hudson et al . [15] . Since the folding of the TL is required for interactions between highly conserved TL-residues and the incoming nucleotide substrate [19] , [38] , [39] , it is likely that the position of β'i6 must change to accommodate the folded TL conformation at each nucleotide addition step of the transcription cycle . During bacteriophage T7 infection , the Eco RNAP β' subunit is phosphorylated by the phage-encoded kinase Gp0 . 7 [47] , and the site of phosphorylation has been identified as a single amino acid in β'i6 , T1068 ( Figures 5 , 8 ) [48] . Phosphorylation at this site appears to affect pausing , as well as ρ-dependent termination behavior , of Eco RNAP [48] . This site is in the β'i6 loop that makes the closest approach to the downstream DNA , but as discussed above , this region is nevertheless not in close contact with the DNA . The surface is already overall acidic ( Figure 5 , front view ) , so it seems unlikely that phosphorylation at this site affects RNAP function by affecting interactions with the downstream DNA . An understanding of the basic principles of transcription and its regulation has been garnered largely through detailed study of the transcription system of one organism , Eco , which has served as a model for understanding transcription at the molecular and cellular level for more than four decades . The detailed and comprehensive structural description of Eco core RNAP and an Eco RNAP TEC presented here sheds new light on the interpretation of previous biochemical and genetic data . Moreover , the molecular models provide a structural framework for designing future experiments to investigate the function of the Eco RNAP lineage-specific insertions and their role in the Eco transcription program , allowing a fuller exploitation of Eco as a model transcription system .
Eco β2-βi4 was amplified by the polymerase chain reaction from the Eco rpoB expression plasmid pRL706 [49] and cloned between the NdeI and BamHI sites of a pET28a-based expression plasmid , creating pSKB2 ( 10-His ) Ecoβ2-βi4 , encoding Eco β2-βi4 with an N-terminal PreScission protease ( GE Healthcare ) cleavable His10-tag . The pSKB2 ( 10-His ) Ecoβ2-βi4 was transformed into Eco BL21 ( DE3 ) cells . After growing transformed cells in LB medium with kanamycin ( 50 µg/ml ) at 37 °C to an A600 nm = 0 . 6 , isopropyl β-D-1-thiogalactopyranoside was added to a final concentration of 1 mM and cells were grown for an additional 3 h at 37 °C . Cells were harvested by centrifugation , resuspended in lysis buffer ( 20 mM Tris-HCl , 0 . 5 M NaCl , 0 . 5 mM β-mercaptoethanol , 5% v/v glycerol , 0 . 5 mM phenylmethanesulphonylfluoride ) , lysed in a continuous-flow French press ( Avestin ) , and clarified by centrifugation . The protein was purified by HiTrap Ni2+-chelating affinity chromatography ( GE Healthcare ) and the His10-tag was removed using PreScission protease ( GE Healthcare ) . The sample was further purified by a second , subtractive HiTrap Ni2+-chelating affinity chromatography step to remove uncleaved His10-tagged protein and the His10-tag released from the cleaved product , and gel filtration chromatography ( Superdex 75 , GE Healthcare ) . The purified protein was concentrated to 17 mg/ml by centrifugal filtration ( VivaScience ) and exchanged into storage buffer ( 10 mM Tris-HCl , pH 8 . 0 , 0 . 15 M NaCl , 1 mM DTT ) , and stored at –80 °C . Selenomethionyl-substituted protein was prepared by suppression of methionine biosynthesis [50] and purified by using similar procedures . Reductive methylation of lysine residues was performed as described [20] . Crystals were grown at 22°C in sitting drops using vapor diffusion by mixing equal volumes of protein solution ( 0 . 5 µl at 6 mg/ml in storage buffer ) and crystallization solution ( 0 . 2 M potassium-sodium tartrate , 20% PEG3350 ) . Crystals ( irregular plates ) appeared after a few days and grew to a maximum size of about 200×100×50 µm in 1 wk . Crystals were prepared for cryo-crystallography by a quick soak in cryo-solution ( 0 . 2 M potassium-sodium tartrate , 35% PEG3350 ) , then flash frozen and stored in liquid nitrogen . Diffraction data were collected at beamline X3A at the National Synchrotron Light Source ( NSLS , Brookhaven , NY ) and processed using HKL2000 [51] . Six of seven possible Se sites were located within the asymmetric unit using the anomalous signal from the Se1 dataset ( Table 1 ) using SHELX [52] . Heavy atom refinement , phasing , and density modification calculations were performed with SHARP [53] using the single-wavelength anomalous dispersion data to 1 . 9 Å-resolution from the Se1 dataset , as well as the 1 . 6 Å-resolution Se2 dataset ( Table 1 ) , yielding an excellent map that allowed automatic building of almost the entire structure using ARP/wARP [54] . Iterative cycles of refinement and model building were carried out using Coot [55] and RefMac5 [56] . The final model was refined to an R/Rfree of 0 . 209/229 at 1 . 6 Å-resolution ( Rfree was calculated using 5% random data omitted from the refinement ) . 97 . 5% of residues fall in the most favored regions of the Ramachandran plot , while no residues are in disallowed regions . The Eco βflap-βi9 ( Eco β residues 831–1057 ) was co-expressed with bacteriophage T4 gp33 [57] as a single operon from a modified pET29a vector [58] and the complex was purified using standard procedures ( K . -A . F . T . , P . Deighan , S . Nechaev , A . Hochschild , E . P . Geiduschek , S . A . D . , in preparation ) . Selenomethionyl-substituted complex was produced by suppression of methionine biosynthesis [50] . Crystals of the complex were grown at 22°C in sitting drops using vapor diffusion by mixing equal volumes of protein solution ( 1 µl at 7 . 5–12 mg/ml in 10 mM Tris-HCl , pH 8 . 0 , 150 mM NaCl , 1% v/v glycerol , 1 mM β-mercaptoethanol , 1 mM DTT ) and crystallization solution ( 0 . 2 M tri-potassium citrate , 20% w/v PEG3350 ) . Crystals were prepared for cryo-crystallography by slow exchange into cryo-solution ( 0 . 2 M tri-potassium citrate , 20% w/v PEG3350 , 20% v/v ethylene glycol ) , then flash frozen and stored in liquid nitrogen . Diffraction data were collected at beamline X3A at the NSLS ( Brookhaven , NY ) and processed using HKL2000 ( Table S1 ) [51] . A molecular replacement solution was obtained using the Native amplitudes ( Table S1 ) with a search model consisting of a homology model of the Eco βflap based on the Taq βflap generated using MODELLER ( the search model excluded the flexible flap-tip ) [59] . The molecular replacement phases were used to locate four Se sites from the anomalous signal of the Se dataset ( Table S1 ) . Heavy atom refinement , phasing , and density modification calculations were performed with SHARP [53] using the single-wavelength anomalous dispersion data from the Se dataset ( Table S1 ) yielding an interpretable map ( Figure S3 ) . Iterative cycles of refinement and model building were carried out using Coot [55] and RefMac5 [56] . The final model was refined to an R/Rfree of 0 . 265/0 . 291 at 3 . 0 Å-resolution ( Rfree was calculated using 5% random data omitted from the refinement ) . 95 . 25% of residues fall in the most favored regions of the Ramachandran plot , while no residues are in disallowed regions . Purification of Eco core RNAP from an overexpression system was performed as described [60] . This results in highly pure Eco RNAP core enzyme , which is deficient in the ω subunit . Eco RNAP holoenzyme was prepared by incubating core RNAP ( 3 mg/ml in 10 mM Tris-HCl , pH 8 , 0 . 2 M NaCl , 0 . 1 mM EDTA , 5 mM DTT ) with a 5-fold molar excess of σ70 for 30 min at room temperature . For cryo-EM , a 5 µl sample ( 0 . 1 mg/ml in the same buffer ) was applied to a Quantifoil grid coated with holey carbon film previously made hydrophilic by glow-discharge . The grid was blotted with filter paper and then immediately plunged into liquid ethane slush . The sample was imaged at 50 , 000× magnification with a Tecnai F20 transmission electron microscope operating at 200 kV . Micrographs displaying minimal astigmatism were digitized at a 14 µm interval ( corresponding to 2 . 8 Å on the image ) using a Zeiss SCAI flat-bed densitometer ( ZI/Carl Zeiss ) . Individual particles were selected by eye and windowed in 90×90 pixel images . Defocus values were estimated from digitized micrographs using ctfit ( EMAN ) [61] . We generated a spEM reconstruction of Eco RNAP by analyzing ∼42 , 000 cryo-images of Eco RNAP particles ( Figures 4A , S4–S6 ) . Particle image orientation parameters were approximately determined using reference projections of a volume generated by low-pass filtration of the Taq core RNAP X-ray structure [5] to 35 Å-resolution . We used a previously devised protocol in which image orientation parameters are iteratively refined by cycling through sets comprising relatively small numbers of reference projections [62] . After a large number of iterations ( 130 ) using the SPIDER software package [63] , we obtained a structure in which well-defined densities not present in the original model volume were apparent . Further refinement of image orientation parameters by projection matching [64] using the SPARX software package [25] yielded a structure of Eco core RNAP with a 0 . 5 Fourier-shell cutoff resolution of about 11 . 2 Å ( Figure S4 ) . For further analysis , the map was Fourier filtered using an ahyperbolic tangent low-pass filter [24] as implemented in the SPARX software package [25] with a stop-band frequency of 0 . 28 and a fall-off of 0 . 45 . Alignments for the Eco lineage-specific insertions ( see Datasets S1–S3 ) were created using the bacterial lineage-specific insertions alignments from Lane et al . [3] as a starting point . The final alignments were created by iterative cycles in which sequences that did not match the Eco domains were removed , followed by re-alignment with MUSCLE [65] or PCMA [66] . Electron Microscopy Data Bank: The single-particle cryoEM reconstruction volume has been deposited under ID code EMD-5169 . Protein Data Bank: Atomic coordinates and structure factors for Eco RNAP β2-βi4 have been deposited under accession code 3LTI . The EM-fitted coordinate model of Eco core RNAP has been deposited under accession code 3LU0 . The coordinates of the Eco RNAP TEC model are available in the Supporting Information ( Dataset S4 ) .
|
Transcription , or the synthesis of RNA from DNA , is one of the most important processes in the cell . The central enzyme of transcription is the DNA-dependent RNA polymerase ( RNAP ) , a large , macromolecular assembly consisting of at least five subunits . Historically , much of our fundamental information on the process of transcription has come from genetic and biochemical studies of RNAP from the model bacterium Escherichia coli . More recently , major breakthroughs in our understanding of the mechanism of action of RNAP have come from high resolution crystal structures of various bacterial , archaebacterial , and eukaryotic enzymes . However , all of our high-resolution bacterial RNAP structures are of enzymes from the thermophiles Thermus aquaticus or T . thermophilus , organisms with poorly characterized transcription systems . It has thus far proven impossible to obtain a high-resolution structure of E . coli RNAP , which has made it difficult to relate the large collection of genetic and biochemical data on RNAP function directly to the available structural information . Here , we used a combination of approaches—high-resolution X-ray crystallography of E . coli RNAP fragments , ab initio structure prediction , homology modeling , and single-particle cryo-electron microscopy—to generate complete atomic models of E . coli RNAP . Our detailed and comprehensive structural models provide the heretofore missing structural framework for understanding the function of the highly characterized E . coli RNAP .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biophysics/transcription",
"and",
"translation",
"biophysics/macromolecular",
"assemblies",
"and",
"machines"
] |
2010
|
Complete Structural Model of Escherichia coli RNA Polymerase from a Hybrid Approach
|
Vulval development in Caenorhabditis elegans serves as an excellent model to examine the crosstalk between different conserved signaling pathways that are deregulated in human cancer . The concerted action of the RAS/MAPK , NOTCH , and WNT pathways determines an invariant pattern of cell fates in three vulval precursor cells . We have discovered a novel form of crosstalk between components of the Insulin and the RAS/MAPK pathways . The insulin receptor DAF-2 stimulates , while DAF-18 PTEN inhibits , RAS/MAPK signaling in the vulval precursor cells . Surprisingly , the inhibitory activity of DAF-18 PTEN on the RAS/MAPK pathway is partially independent of its PIP3 lipid phosphatase activity and does not involve further downstream components of the insulin pathway , such as AKT and DAF-16 FOXO . Genetic and biochemical analyses indicate that DAF-18 negatively regulates vulval induction by inhibiting MAPK activation . Thus , mutations in the PTEN tumor suppressor gene may result in the simultaneous hyper-activation of two oncogenic signaling pathways .
PTEN ( Phosphatase and TENsin homologue ) is the second-most frequently somatically mutated tumor suppressor gene in human cancer . PTEN is often inactivated in glioblastoma , melanoma , prostate and endometrial neoplasia [1] . Germline mutations in PTEN are also known to cause a variety of rare syndromes , collectively known as the PTEN hamartoma tumor syndromes ( PHTS ) [2] . Cowden syndrome is the best-described syndrome within PHTS , with approximately 80% of patients carrying germline PTEN mutations [3] . The main reported function of PTEN is as a lipid phosphatase , which dephosphorylates PhosphatidylInositol ( 3 , 4 , 5 ) -trisphosphate ( PIP3 ) at position 3 , thereby inhibiting the insulin pathway [4] . However , PTEN can also act as a dual-specificity tyrosine and serine/threonine protein phosphatase [5] , [6] . The catalytic phosphatase domain of PTEN ( amino acids 123–131 ) contains several conserved amino acids , mutations of which affect the efficiency and specificity of the phosphatase activity [7] . One such mutation is G129E , which causes PTEN to lose its lipid phosphatase activity while retaining most of its protein phosphatase activity [4] , [5] . Using the G129E mutation , numerous reports have provided evidence for a crucial role of PTEN protein phosphatase activity in regulating cell migration , invasion and spreading independently of the canonical insulin signaling pathway . For example , PTEN G129E regulates cell migration , spreading , and the formation of focal adhesions [8] . Moreover , PTEN G129E binds and de-phosphorylates the Focal Adhesion Kinase FAK in vitro [8] . In glioblastoma cells injected into nude mice , PTEN G129E expression inhibits cell invasion , accompanied by decreased FAK phosphorylation without changing the activity of the AKT kinase [6] , [9] . Additionally , PTEN binds and dephosphorylates the adaptor protein Shc to modulate cell motility [10] . The daf-18 gene encodes the single PTEN ortholog in C . elegans [11] . Under favorable growth conditions , C . elegans larvae pass through four distinct larval stages termed L1 to L4 . However , under conditions of starvation or overcrowding , the L1 larvae enter a long-lived , stress resistant alternative developmental stage called the dauer larva stage . DAF-18 PTEN controls entry into the larval dauer stage , life span , neurite outgrowth and cell-cycle progression , mainly by inhibiting the insulin signaling pathway [12]–[14] . Human PTEN can functionally replace C . elegans DAF-18 to rescue the daf-18 ( lf ) DAuer Formation defective ( DAF-d ) phenotype [15] . In the presence of abundant food , binding of various insulin ligands to the DAF-2 insulin receptor causes the activation of AGE-1 , the only type I phosphatyidlinositol-3-kinase ( PI3K ) encoded by the C . elegans genome [11] . AGE-1 phosphorylates PI ( 4 , 5 ) P2 to PI ( 3 , 4 , 5 ) P3 , which acts as a second messenger . PIP3 then activates the AKT-1 and AKT-2 kinases that phosphorylate and thereby inhibit the FOXO transcription factor DAF-16 [16] . In the absence of the insulin signal , non-phosphorylated DAF-16 enters the nucleus and activates genes promoting entry into the dauer stage [17] . The main reported function of DAF-18 PTEN is to antagonize the insulin pathway by dephosphorylating PIP3 [14] . Loss of daf-18 thus leads to hyper-activation of the insulin pathway and a DAF-d phenotype , while loss of daf-2 or age-1 function leads to a DAuer Formation constitutive ( DAF-c ) phenotype . Recent evidence indicates that similar to mammalian PTEN , C . elegans DAF-18 can also act as a protein phosphatase to regulate insulin-independent events . For example , DAF-18 PTEN directly binds and dephosphorylates the ephrin receptor tyrosine kinase VAB-1 to regulate oocyte maturation in the hermaphrodite germline [18] . Moreover , multiple genes causing synthetic lethality in combination with daf-18 ( lf ) have been identified , pointing to additional functions of DAF-18 besides its role in insulin signaling [19] . The development of the C . elegans hermaphrodite vulva , the egg-laying organ , is one of the best-characterized models for organogenesis [20] . The interplay between the conserved RAS/MAPK , NOTCH , and WNT signaling pathways specifies two distinct vulval cell fates . Beginning in the L2 stage , the gonadal Anchor Cell ( AC ) releases the EGF ligand LIN-3 , which activates the EGF receptor homolog LET-23 in the six Vulval Precursor Cells ( VPCs ) . The VPC located nearest the AC ( P6 . p ) receives most of the inductive LIN-3 EGF signal and hence exhibits the strongest activity of the RAS/MAPK pathway , allowing P6 . p to adopt the primary ( 1° ) vulval cell fate . Consequently , P6 . p produces a lateral signal that activates the LIN-12 NOTCH signal in the neighboring VPCs P5 . p and P7 . p . Notch signaling in these two VPCs induces them to adopt the secondary ( 2° ) cell fate and at the same time blocks transduction of the inductive LIN-3 signal by inhibiting MAPK activation . The 1° VPC P6 . p and the 2° VPCs P5 . p and P7 . p each go through three rounds of cell division to generate 22 cells that form the vulva . The remaining three distal VPCs ( P3 . p , P4 . p and P8 . p ) adopt the non-vulval 3° fate , which divide once and then fuse with the surrounding hypodermis ( hyp7 ) . Mutations in genes encoding components of the RAS/MAPK , NOTCH , and WNT signaling pathways change the patterning of the VPC fates , which can readily be quantified . For example , mutations that hyperactivate the RAS/MAPK pathway cause the induction of more than three VPCs , resulting in a Multivulva ( Muv ) phenotype . On the other hand , mutations that cause a decrease in RAS/MAPK pathway activity , lead to the induction of fewer than three VPCs , a phenotype called Vulvaless ( Vul ) . In this work , we have discovered and characterized a new mode of crosstalk between components of the insulin and the RAS/MAPK pathways . Using genetic and biochemical epistasis analyses , we found that the insulin receptor DAF-2 stimulates while DAF-18 PTEN inhibits RAS/MAPK signaling in the VPCs . Surprisingly , part of the inhibitory activity of DAF-18 on the RAS/MAPK pathway is independent of its PIP3 lipid phosphatase activity and does not involve further downstream components of the insulin pathway . Our results indicate that DAF-18 negatively regulates vulval induction most likely by inhibiting MAP kinase MPK-1 signaling .
In our previous work , we reported first evidence for a crosstalk between the DAF-2 insulin receptor and the RAS/MAPK signaling pathway during vulval development [21] . To further investigate this interaction , we performed a systematic epistasis analysis by combining various mutations in the insulin and RAS/MAPK signaling pathways and quantifying the levels of vulval induction ( Table 1 ) . As reported previously , a reduction-of-function ( rf ) mutation in the insulin receptor daf-2 partially suppressed the Muv phenotype of let-60 ras gain-of-function ( gf ) animals ( Table 1 , rows 1 , 2 ) [21] . Conversely , a loss-of-function ( lf ) mutation in the PTEN homolog daf-18 strongly enhanced the let-60 ( gf ) Muv phenotype ( Table 1 , row 3 ) . Moreover , daf-18 ( lf ) suppressed the vulvaless ( Vul ) phenotype caused by mutations in the EGF receptor let-23 or in its cofactor lin-2 , which activates the RAS/MAPK signaling pathway in the VPCs ( Table 1 , rows 4–7 ) . Since DAF-18 PTEN counteracts the type I phosphatidyl-inositol-3 kinase ( PI3K ) AGE-1 that transduces the insulin signal downstream of DAF-2 , we tested if an age-1 ( lf ) mutation could revert the enhanced vulval induction caused by daf-18 ( lf ) . Surprisingly , age-1 ( lf ) only partially suppressed the increase in vulval induction caused by daf-18 ( lf ) , both in the let-60 ( gf ) and the lin-2 ( lf ) backgrounds ( Table 1 , rows 8 , 9 ) . Also , the daf-2 ( rf ) mutation only partially reverted the enhancement of the let-60 ( gf ) Muv phenotype by daf-18 ( lf ) ( Table 1 , row 10 ) , suggesting that DAF-18 inhibits vulval induction to some extent independently of DAF-2 and AGE-1 . However , mutations in further downstream components of the DAF-2 insulin pathway had no detectable effect on vulval induction . For example , a gf mutation in akt-1 , which encodes one of the two AKT homologues transducing the insulin signal downstream of AGE-1 [16] , did not suppress the let-23 ( rf ) Vul phenotype ( Table 1 , row 11 ) , and a lf mutation in akt-1 did not suppress the daf-18 ( lf ) let-60 ( gf ) Muv phenotype ( Table 1 , row 12 ) . To examine a possible redundancy between the two akt genes , we performed akt-2 RNAi in daf-18 ( lf ) let-60 ( gf ) ; akt-1 ( lf ) triple mutants , but observed no reduction in vulval induction compared to the RNAi controls ( Table 2 , rows 1 , 2 ) . Also , a lf mutation in daf-16 , which encodes a FOXO transcription factor that is negatively regulated by the insulin pathway , did not enhance the let-60 ( gf ) Muv phenotype ( Table 1 , row 13 ) . We further tested the role of AGE-1 during vulval development . Since age-1 ( lf ) leads to a constitutive dauer phenotype ( DAF-c ) that is maternally rescued , homozygous age-1 ( lf ) worms could only be analyzed in the F1 progeny of heterozygous age-1 ( lf ) /+ parents or when rescued by the daf-16 ( lf ) mutation . Our analysis indicated that age-1 also exhibits a partial maternal rescue in vulval induction since the homozygous age-1 ( lf ) ; let-60 ( gf ) progeny obtained from heterozygous age-1 ( lf ) /+ parents displayed similar levels of vulval induction as let-60 ( gf ) single mutants ( Table 1 , row 14 ) . In contrast , homozygous age-1 ( lf ) ; let-60 ( gf ) double mutants maintained in the daf-16 ( lf ) background exhibited a partially suppressed Muv phenotype , though age-1 ( lf ) suppressed the let-60 ( gf ) Muv phenotype to a lesser extent than daf-2 ( rf ) ( Table 1 , row 15 , p-value≤0 . 05 compared to row 2 ) . Taken together , the genetic analysis indicates that the DAF-2 insulin receptor promotes and DAF-18 PTEN inhibits vulval induction . DAF-2 and DAF-18 both regulate vulval induction through AGE-1-dependent as well as AGE-1-independent pathways that do not utilize the canonical insulin pathway downstream of AGE-1 . AGE-1 is the only C . elegans member of the type I family of PI3Ks , which convert PI ( 4 , 5 ) P2 into PI ( 3 , 4 , 5 ) P3 . To further investigate the AGE-1-independent effect of DAF-18 on vulval induction , we tested the roles of alternative PI3Ks that can phosphorylate PIs at position 3 . vps-34 encodes a type III PI3K , which catalyzes the production of PI ( 3 ) P1 , and piki-1 encodes a type II PI3K involved in the engulfment of apoptotic cell corpses [22] . In order to examine the role of these alternative PI3Ks during vulval induction , we performed RNAi against vps-34 and piki-1 in age-1 ( lf ) ; daf-18 ( lf ) ; lin-2 ( lf ) animals and tested for a further reduction of vulval induction . RNAi to vps-34 and piki-1 has been previously shown to be effective in different tissues [23] , [24] . Neither vps-34 nor piki-1 RNAi caused any significant reduction in the number of induced VPCs when compared to control ( gfp ) RNAi animals ( Table 2 , rows 3–5 ) . Furthermore , vps-34 and piki-1 RNAi in an age-1 ( lf ) ; daf-18 ( lf ) let-60 ( gf ) background did not cause a decrease in vulval induction ( Table 2 , rows 6–8 ) . It thus seems unlikely that an alternative PI3K acts redundantly with AGE-1 during vulval induction , further supporting our observation that DAF-18 regulates vulval induction not only by regulating PIP3 levels but also via a lipid phosphatase-independent activity . To characterize the nature of the cell fate transformation caused by daf-18 ( lf ) , we quantified the levels of the EGL-17::CFP reporter , whose expression is induced by RAS/MAPK signaling in the 1° vulval cell lineage [25] and of the LIP-1::GFP reporter whose expression is induced by LIN-12 NOTCH signaling in the 2° vulval cell lineage [26] . daf-18 ( lf ) let-60 ( gf ) double mutants showed an increased frequency of adjacent VPC descendants expressing high levels of EGL-17::CFP when compared to let-60 ( gf ) single mutants ( Figure 1A–1D ) . Specifically , in daf-18 ( lf ) ; let-60 ( gf ) double mutants 27% of adjacent VPC descendants showed strong EGL-17::CFP expression ( i . e . at least 50% of the signal intensity seen in the P6 . p lineage ) versus 16% in let-60 ( gf ) single mutants ( Figure 1G ) . Furthermore , while the 2° P5 . p and P7 . p descendants in the wild-type displayed weak ( i . e . less than 50% of the P6 . px signal intensity ) EGL-17::CFP expression in 22% of the cases , 52% of daf-18 ( lf ) single mutants showed EGL-17::CFP expression in the 2° cells ( Figure 1G ) . Besides the slight increase in EGL-17::CFP expression , the morphology of the vulval invagination at the L4 larval stage was changed in daf-18 ( lf ) let-60 ( gf ) double mutants . The vulval invagination formed by the P5 . p to P7 . p descendants of most let-60 ( gf ) single mutants resembles the single invagination formed in the wild-type ( Figure 1E ) . In daf-18 ( lf ) let-60 ( gf ) double mutants , on the other hand , the P5 . p to P7 . p descendants were often completely detached from the cuticle , resulting in an abnormal shape of the vulval invagination ( Figure 1F ) ( 37% detached P5 . p and/or P7 . p descendants in daf-18 ( lf ) let-60 ( gf ) versus 3% detached in let-60 ( gf ) , n = 54 and n = 35 , respectively ) . A detachment of the P5 . p and P7 . p descendants from the cuticle is indicative of a 2° to 1° cell fate transformation as it has been observed in mutants exhibiting elevated MAPK activity in the 2° lineage [27] . In contrast to the 1° fate marker EGL-17::CFP , expression of the 2° fate marker LIP-1::GFP was not changed in daf-18 ( lf ) mutants . In particular , LIP-1::GFP levels in the P5 . p and P7 . p descendants were unchanged in daf-18 ( lf ) let-60 ( gf ) double mutants compared to let-60 ( gf ) single mutants ( Figure 1H–1J ) . Thus , daf-18 ( lf ) enhances specification of the 1° cell fate and causes a partial 2° to 1° fate transformation in P5 . p and P7 . p without affecting the strength of the lateral LIN-12 NOTCH signal . Since EGFR/RAS/MAPK signaling induces the 1° vulval cell fate and daf-18 ( lf ) mutants exhibited an increased expression of the 1° cell fate marker EGL-17::CFP , we sought to determine at what level DAF-18 inhibits the activity of the EGFR/RAS/MAPK signaling pathway . For this purpose , we performed further epistasis analysis combining daf-18 ( lf ) with mutations in different components of the RAS/MAPK pathway . Even though daf-18 ( lf ) single mutants showed no obvious changes in the vulval fate pattern ( Table 3 , rows 1 , 2 ) , daf-18 ( lf ) increased the levels of vulval induction in most of mutants in the RAS/MAPK pathway tested , confirming that DAF-18 negatively regulates the RAS/MAPK signaling during vulval induction . For example , when combined with mutations in different positive regulators of the RAS/MAPK pathway such as let-23 ( rf ) , lin-2 ( lf ) or lin-45 ( rf ) , daf-18 ( lf ) significantly suppressed the Vul phenotype of these mutants ( Table 3 , rows 5–8 and 11–12 ) . In particular , daf-18 ( lf ) suppressed a lf mutation in the RAS-GEF sos-1 when assayed in the let-60 ( gf ) background to rescue the lethality caused by sos-1 ( lf ) , placing daf-18 function downstream of sos-1 ( table 3 , rows 9–10 ) . However , since vulval induction in sos-1 ( lf ) ; let-60 ( gf ) animals is partly sensitive to the inductive anchor cell signal [28] , we cannot exclude the possibility that DAF-18 might inhibit RAS/MAPK signaling through a SOS-1 independent branch of the pathway . As an exception , daf-18 ( lf ) did not suppress the Vul phenotype of lin-3 ( rf ) mutants ( Table 3 , rows 3–4 ) , suggesting that daf-18 ( lf ) alone is not sufficient to activate the RAS/MAPK pathway in the absence of the inductive AC signal . Furthermore , daf-18 ( lf ) did not affect the completely penetrant Vul phenotype caused by mpk-1 ( lf ) ( Table 3 , rows 13–14 ) . Taken together , our epistasis analysis indicates that DAF-18 inhibits RAS/MAPK signaling downstream of or in parallel with the RAS-GEF SOS-1 and upstream or at the level of the MAP kinase MPK-1 . Activation of the RAS/MAPK pathway results in an increased phosphorylation and activity of the downstream effectors RAF , MAPK kinase ( MEK ) and MAPK . We thus examined if daf-18 ( lf ) mutants exhibit elevated levels of MEK and MAPK phosphorylation . Western blots of extracts from L4 larvae were probed with antibodies against mono-phosphorylated MEK ( pMEK-2 ) and di-phosphorylated MAPK ( dpMPK-1 ) . Although the C . elegans genome encodes two MEK genes , MEK-1 and MEK-2 , the phosphorylation site in human MEK to which the phospho-MEK antibody was raised ( S217/S221 ) is only conserved in C . elegans MEK-2 . Thus , we were able to specifically detect phosphorylated MEK-2 in whole animal extracts . Wild-type and daf-18 ( lf ) L4 larvae contained only low levels of pMEK-2 and dpMPK-1 that could not be reliably quantified . As expected , let-60 ( gf ) single mutants contained significantly higher levels of pMEK-2 and dpMPK-1 than wild-type larvae ( Figure 2A and 2C ) . However , we observed no further increase in pMEK-2 levels in daf-18 ( lf ) let-60 ( gf ) double compared to let-60 ( gf ) single mutants ( Figure 2B ) . In contrast , dpMPK-1 levels were around two-fold increased daf-18 ( lf ) let-60 ( gf ) compared to let-60 ( gf ) mutants ( Figure 2D ) . Together with the genetic epistasis data presented above , the increase in MPK-1 phosphorylation in the absence of a significant change in MEK-2 phosphorylation indicates that DAF-18 most likely inhibits vulval induction at the level of MPK-1 . Finally , the fact that we observed increased MPK-1 phosphorylation in total worm lysates suggests a global regulation of the RAS/MAPK pathway by DAF-18 , probably including the germline . To further investigate the role of DAF-18 during vulval induction , we constructed a translational reporter by inserting a gfp cassette at the 3′ end of the ORF in a genomic daf-18 fragment ( Figure 3 ) . This DAF-18::GFP reporter rescued both the dauer defective ( DAF-d ) phenotype ( data not shown ) as well as the vulval phenotypes of daf-18 ( lf ) with similar efficiency as a 6 . 5 kb genomic fragment spanning the entire daf-18 locus ( Figure 4 ) . DAF-18::GFP expression was observed in many tissues during all larval stages , including the developing vulva , the uterus , the ventral nerve cord and the distal tip cells ( data not shown ) . In particular , equal levels of DAF-18::GFP expression were detected in the six VPCs of L2 larvae , and expression persisted in the descendants of the induced VPCs until the Pn . pxxx stage ( Figure 3 ) . Interestingly , the sub-cellular localization of DAF-18::GFP changed over the course of vulval development . In the VPCs of L2 larvae prior to and during induction ( Pn . p stage ) , DAF-18::GFP was predominantly localized in the cytoplasm and the nucleus ( Figure 3A , 3B and 3B′ ) . However , at the subsequent stages ( Pn . px to Pn . pxx stages ) , DAF-18::GFP became increasingly localized to the plasma membrane of the vulval cells ( Figure 2C , 2D and 2D′ ) . Plasma membrane staining peaked at the “Christmas tree” ( Pn . pxxx ) stage , when almost all the protein appeared to be localized to the membranes and nuclear staining was reduced to very low levels ( Figure 3E and 3F ) . Since the DAF-18::GFP reporter was also expressed in tissues surrounding the vulval cells , we examined whether tissue-specific expression of DAF-18 in the VPCs reduces vulval induction . For this purpose , we expressed daf-18 cDNA fused to gfp under the control of the Pn . p cell-specific lin-31 promoter , which is active in the VPCs before and during vulval induction [29] ( Plin-31::daf-18 cDNA::gfp::unc-54 3′ UTR ) . Indeed , introduction of the lin-31::daf-18::gfp transgene into daf-18 ( lf ) ; let-23 ( rf ) animals repressed vulval induction with similar efficiency as the daf-18::gfp reporter or a genomic daf-18 rescue construct ( Figure 4 ) . Besides the vulval cells , the DAF-18::GFP reporter was also expressed at the L3 to L4 larval stages in several cells of the uterus , which is part of the somatic gonad ( Figure 3 ) . We thus tested if the daf-18-mediated repression of vulval induction requires the gonad by ablating the Z1 to Z4 somatic gonad and germline precursor cells at the L1 stage . In gonad-ablated daf-18 ( lf ) let-60 ( gf ) double mutants , vulval induction was higher than in gonad-ablated let-60 ( gf ) single mutants , indicating that DAF-18 represses vulval induction independently of a signal from the gonad ( Table 3 , rows 15 , 16 ) . Thus , DAF-18 probably acts predominantly in the VPCs to inhibit MAPK signaling during vulval induction . Mammalian PTEN acts as a lipid phosphatase as well as a dual-specificity protein phosphatase [5] , [6] . Moreover , a recent report has shown that C . elegans DAF-18 can act as a protein phosphatase inhibiting signaling by the VAB-1 ephrin receptor during oocyte maturation [18] . The G129E mutation in the catalytic center of human PTEN eliminates the lipid phosphatase activity , while retaining the protein phosphatase activity [7] . The corresponding glycine 174 residue in C . elegans DAF-18 was therefore mutated to glutamic acid in the daf-18 genomic rescue construct . To quantify the rescuing activity of the daf-18 wild-type ( daf-18 wt ) and the G174E mutated lipid phosphatase mutant ( daf-18 G174E ) , these constructs were expressed in the daf-18 ( lf ) let-60 ( gf ) and let-23 ( rf ) ; daf-18 ( lf ) backgrounds , and vulval induction was quantified . As expected , expression of daf-18 wt rescued both the DAF-d ( data not shown ) and vulval phenotypes of daf-18 ( lf ) ( Figure 5 ) . In contrast , daf-18 G174E did not rescue the DAF-d phenotype ( [15] and own observation ) , yet exhibited a weaker , though significant rescuing activity of the vulval induction phenotype ( Figure 5 ) . These results indicate that the DAF-18 protein and lipid phosphatase activities each play independent roles in negatively regulating the RAS/MAPK pathway and that both activities are required for the full inhibition of vulval induction by DAF-18 .
The Insulin pathway is a key regulator of development , reproduction , and life span in metazoans . In this study , we have discovered a new form of cross-talk between the Insulin and RAS/MAPK pathways during vulval development . Signaling by the Insulin receptor DAF-2 positively regulates MAPK activation . Surprisingly , the effect of DAF-2 on vulval development does not involve activation of the canonical Insulin pathway . DAF-2 signaling regulates vulval induction in at least two distinct manners , through AGE-1 dependent and independent pathways ( Figure 6 ) . One possible explanation for the AGE-1-independent branch of DAF-2 signaling is supported by mammalian data , which suggest that the Insulin receptor can directly stimulate RAS activation by recruiting GRB2 and the RAS-GEF SOS [30] , [31] . Also in C . elegans , LET-60 RAS was found to act downstream of the DAF-2 Insulin receptor to modulate the effects of Insulin signaling during entry into the Dauer stage [32] . Furthermore , we found that the PTEN ortholog DAF-18 strongly inhibits RAS/MAPK signaling . Vulval induction in daf-18 ( lf ) let-60 ( gf ) double mutants reaches levels similar to those seen in the strongest Muv mutants such as lin-15AB ( lf ) [33] . The increase in RAS/MAPK signaling in daf-18 ( lf ) mutants could be partially reverted by loss of the PI3K activity , suggesting that elevated levels of PIP3 do stimulate RAS/MAPK signaling but cannot explain all the functions DAF-18 exerts during vulval induction . Accordingly , the inhibitory activity requires both the lipid and protein phosphatase activities of DAF-18 . PIP3 acts as a second messenger that activates multiple downstream targets . One major PIP3 target in the Insulin pathway is the AKT kinases , which phosphorylate and thereby inhibits the FOXO transcription factor DAF-16 . However neither akt-1 , akt-2 nor daf-16 mutations had any detectable effect on vulval induction . Thus , PIP3 must act via other targets to stimulate RAS/MAPK signaling . Increased levels of PIP3 in the plasma membrane could , for example , enhance the recruitment of an alternative GEF that activates RAS signaling in parallel to the RAS-GEF SOS-1 [28] ( Figure 6 ) . However , we observed that prior to and at early stages of vulval induction , DAF-18::GFP was localized predominantly in the cytoplasm and nucleus of the VPCs , while membrane localization of DAF-18 only became apparent at later stages . Previous observations of mammalian PTEN localization suggested that PTEN performs different functions depending on its sub-cellular localization [34] . It has been proposed that the lipid phosphatase activity is important for the cytoplasmic and membrane functions of mammalian PTEN , while the protein phosphatase activity is rather required for its nuclear functions [34] , [35] . The nuclear localization of PTEN in mammalian cells is often associated with cell-cycle arrest in G1 and accompanied by decreased levels of ERK phosphorylation . Prior to vulval induction , the VPCs are maintained in a long G1 arrest lasting the entire L2 stage [36] . It is therefore possible that initially DAF-18 acts predominantly in the nucleus as a protein phosphatase that negatively regulates vulval induction . Indeed , Western blot analysis revealed elevated levels of dpMPK-1 in daf-18 ( lf ) mutants , supporting our model that DAF-18 –directly or indirectly- blocks MAPK activation ( Figure 6 ) . In humans , PTEN is one of the most frequently mutated tumor suppressor genes . However , not all disease phenotypes associated with loss of PTEN can be explained by hyper-activation of the Insulin pathway alone . Thus , PTEN must have other functions that are independent of its inhibitory activity in the Insulin pathway . Accordingly , Suzuki and Han [19] observed many synthetic phenotypes in C . elegans daf-18 ( lf ) mutants , including embryonic lethality and sterility , which are independent of DAF-16 FOXO and do not involve DAF-2 InsR signaling . Our work highlights the importance of C . elegans DAF-18 PTEN in regulating a range of biological processes and may serve as a basis to better understand the multiple roles human PTEN plays during cancer initiation and progression . Thus , single mutations in the PTEN tumor suppressor may result in the simultaneous hyper-activation of several oncogenic signaling pathways .
Standard methods were used for maintaining and manipulating Caenorhabditis elegans [37] . Animals were cultured at 20°C and the wild-type strain is the Bristol N2 strain . Information regarding the mutants used in this study can be found on WormBase ( http://www . wormbase . org ) . Mutations used according to their linkage group: LG I: daf-16 ( mu86 ) , LG II: age-1 ( mg44 ) , let-23 ( sy1 ) , LG III: daf-2 ( e1370 ) , mpk-1 ( ga117 ) , LG IV: let-60 ( n1046 ) , daf-18 ( ok480 ) , lin-3 ( e1417 ) , lin-45 ( sy96 ) , LG V: akt-1 ( mg144gf ) , akt-1 ( ok525lf ) , LG X: lin-2 ( n397 ) , sos-1 ( s1031 ) , unc-46 ( e177 ) to cis link sos-1 ( s1031 ) , LG X: lin-2 ( n397 ) , gap-1 ( ga133 ) . Transgenes used: syIs59[Pegl-17::cfp] , zhIs4[Plip-1::gfp] , zhEx382[daf-18 genomic] zhEx343[daf-18::gfp] , zhEx358[Plin-31::daf-18::gfp] , zhEx344[daf-18 G174E] . pIN05 ( daf-18 genomic wt ) was made by cloning the whole genomic fragment of daf-18 starting 1 . 3 kb upstream of the ATG and ending 0 . 5 kb downstream of the STOP and cloning to pGEM-T . pIN03 ( daf-18 genomic G174E ) was made by fusion PCR [38] of two overlapping fragments of the whole genomic daf-18 starting 1 . 3 kb upstream of the ATG and ending 0 . 5 kb downstream of the STOP using primers which contain the mutation G174E ( GGC to GAA ) in the overlapping region and cloning to pGEM-T . pIN17 ( Plin-31::daf-18 cDNA::gfp::unc-54 3′UTR ) was made by amplifying daf-18 cDNA::gfp from a previously cloned plasmid with primers containing NotI sites on both ends , digestion with NotI and cloning into the NotI site of the pB253 plasmid containing the lin-31 enhancer and promoter . The daf-18 genomic translational GFP reporter was made using fusion PCR of three parts by inserting a gfp cassette in frame between the last exon and the 3′ UTR into a genomic fragment encompassing 1 . 3 kb of 5′ regulatory sequences and the complete daf-18 coding sequences . Sequences of the primers used for the different constructs can be found in Table S1 . Worms expressing extra-chromosomal transgenic arrays were generated by microinjection of DNA into young adult worms [39] . pIN03 ( zhEx344 ) , pIN05 ( zhEx382 ) and pIN17 ( zhEx358 ) were injected at a concentration of 50 ng/µl . The fusion PCR daf-18 genomic::gfp ( zhEx343 ) translational reporter was injected at a concentration of 30 ng/ul . Co-markers used were either pCFJ90 ( Pmyo-2::mCherry ) at a concentration of 2 ng/ul or pTJ1157 ( Plin-48::gfp ) at a concentration of 50 ng/ul . Final concentration of injected DNA was filled up to 150 ng/ul using the empty plasmid pBluescript-KS . DAF-18::GFP , Plip-1::GFP and Pegl-17::CFP expression were observed under fluorescent light illumination with either a Leica DMRA microscope equipped with a cooled CCD camera ( Hamamatsu ORCA-ER ) or Olympus BX61 with Q Imaging Fast 1394 Retiga 2000R camera ( Q Imaging Inc . , Canada ) controlled by the Openlab 5 . x software ( Improvision/PerkinElmer ) . Animals were mounted on 4% agarose pads in M9 solution with 20 mM tetramisole hydrochloride . Quantification of fluorescence levels was performed under the same picture acquisition settings for all conditions examined . Vulval induction was scored by examining worms at the L4 stage under Nomarski optics as described [40] . The number of VPCs that had adopted a 1° or 2° Vulval fate was counted for each animal and the induction index was calculated by dividing the number of 1° or 2° induced cells by the number of animals scored . Statistical analysis was performed using a t-test for independent samples . RNA interference analysis ( RNAi ) was performed by feeding animals dsRNA-producing bacteria as described previously [41] . Around 10 to 20 P0 animals at the L1 larval stage were transferred to plates containing RNAi bacteria grown on 3 mM IPTG . Vulval induction was scored in the F1 generation ( or the P0 for akt-2 RNAi ) at the L4 larval stage to count the number of induced VPCs . gfp , akt-2 , piki-1 and vps-34 RNAi clones were all taken from the Ahringer RNAi library . Forty-five animals at the L4 stage were placed into 15 µl of 1× SDS sample buffer , lysed at 95°C for 5 min , centrifuged at 14 , 000 rpm for 2 min and the supernatant was loaded on 10% acrylamide gels , which were analyzed by Western blotting . Anti-phospho-MEK1/2 ( S217/S221 ) and Anti-MEK1/2 ( D1A5 ) antibodies were purchased from Cell Signaling Technology ( Beverly , MA ) . Anti-di-phosphorylated ERK-1&2 ( M8159 ) antibody was purchased from Sigma-Aldrich ( St . Louis , MO ) , and anti-ERK 2 ( K-23 ) antibody was purchased from Santa Cruz Biotechnology ( Santa Cruz , CA ) . Quantification of the bands was performed using the gel quantification plugin in ImageJ software [42] . The ratios between phosphorylated and total MEK-2 and MPK-1 levels , respectively , were calculated and normalized for each independent experiment to the ratios measured in the let-60 ( gf ) single mutants .
|
The human tumor suppressor PTEN is mutated in many different types of cancer . Using the roundworm C . elegans as a model to study how cells communicate during animal development , we discovered a new mechanism by which PTEN inhibits the activity of the oncogenic RAS/MAPK signaling pathway . Focusing on the development of the vulva , the egg-laying organ of the hermaphrodite , as a model to investigate the regulation of RAS/MAPK signaling , we could distinguish between two distinct inhibitory activities of PTEN on the RAS/MAPK signaling pathway . On the one hand , PTEN acts as a lipid phosphatase that inhibits the production of PIP3 , which in turn stimulates RAS/MAPK signaling . On the other hand , PTEN acts as a protein phosphatase that negatively regulates RAS/MAPK signaling by inhibiting signal transduction at the level of the MAPK , which is a key component in the pathway . Understanding the detailed molecular mechanism by which the PTEN tumor suppressor homolog regulates signal transduction in C . elegans can help predict the consequences of mutations in human PTEN for cancer development in humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"gene",
"networks",
"animal",
"genetics",
"basic",
"cancer",
"research",
"cell",
"differentiation",
"gene",
"function",
"animal",
"models",
"oncology",
"developmental",
"biology",
"caenorhabditis",
"elegans",
"model",
"organisms",
"molecular",
"development",
"biology",
"signaling",
"genetics",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics",
"cell",
"fate",
"determination"
] |
2012
|
PTEN Negatively Regulates MAPK Signaling during Caenorhabditis elegans Vulval Development
|
A significant percentage of young men are infertile and , for the majority , the underlying cause remains unknown . Male infertility is , however , frequently associated with defective sperm motility , wherein the sperm tail is a modified flagella/cilia . Conversely , a greater understanding of essential mechanisms involved in tail formation may offer contraceptive opportunities , or more broadly , therapeutic strategies for global cilia defects . Here we have identified Rab-like 2 ( RABL2 ) as an essential requirement for sperm tail assembly and function . RABL2 is a member of a poorly characterized clade of the RAS GTPase superfamily . RABL2 is highly enriched within developing male germ cells , where it localizes to the mid-piece of the sperm tail . Lesser amounts of Rabl2 mRNA were observed in other tissues containing motile cilia . Using a co-immunoprecipitation approach and RABL2 affinity columns followed by immunochemistry , we demonstrated that within developing haploid germ cells RABL2 interacts with intra-flagella transport ( IFT ) proteins and delivers a specific set of effector ( cargo ) proteins , including key members of the glycolytic pathway , to the sperm tail . RABL2 binding to effector proteins is regulated by GTP . Perturbed RABL2 function , as exemplified by the Mot mouse line that contains a mutation in a critical protein–protein interaction domain , results in male sterility characterized by reduced sperm output , and sperm with aberrant motility and short tails . Our data demonstrate a novel function for the RABL protein family , an essential role for RABL2 in male fertility and a previously uncharacterised mechanism for protein delivery to the flagellum .
Infertility affects at least 1 in 20 men of reproductive age [1] and for the majority , the underlying causal mechanism remains unknown . This , and the absence of effective male-based contraceptives , stems from a fundamental lack of knowledge of the genes and pathways required to form functional sperm . Spermatozoa are produced within the seminiferous epithelium of the testis through a series of processes including stem cell renewal , meiosis and a radical differentiation process , termed spermiogenesis , wherein haploid germ cells are transformed into highly polarized cells with the potential for motility and fertilization . The mammalian sperm tail , like motile cilia and flagella from all species , contains an axoneme at its core composed of a 9+2 microtubule arrangement . The axoneme develops from a centriole/basal body at the base of the sperm head and functions to metabolize ATP and generate microtubule sliding and motility [2] . Unlike the majority of other cilia however , the sperm tail possesses peripherally arranged accessory structures including the fibrous sheath and outer dense fibers which impart directionality to tail beating , protection against shearing forces , and in the case of the fibrous sheath is a scaffold for enzymes involved in glycolysis and the generation of at least a proportion of the ATP required as fuel for axoneme movement [3] . The mechanisms by which the sperm tail is assembled remain almost completely unknown . Defects in sperm axoneme function result in asthenospermia ( abnormal sperm motility ) [4] . Global defects in motile axoneme function result in primary ciliary dyskinesia ( PCD ) , a syndrome characterized by variable presentations of recurrent respiratory tract infections , male infertility , dextrocardia ( Kartegener's syndrome ) and hydrocephalus [5] . Using a random mutagenesis approach , we have identified RABL2 as being essential for sperm tail function and male fertility . RABL proteins are a poorly characterized sub-family of the Ras GTPase superfamily originally discovered in Trypanosomes and Chlamydomonas as an essential component of the intra-flagellar transport ( IFT ) particles required for primary cilia function [6] , [7] . Here we have demonstrated that RABL2 is essential for sperm flagella , a motile cilia assembly . Biochemically , RABL2 function is regulated by GTP , it binds to components of the IFT complex B machinery and is involved in the delivery a set of cargo protein either to , or within , the developing flagellum .
In an effort to identify genes critically involved in male fertility , we used N-ethyl-nitrosourea ( ENU ) to randomly mutate the mouse genome . Mice were mutated on a C57BL6 background then outbreed onto the CBA strain for two generations to facilitate mutation mapping , after which mice were maintained on a mixed background through inter-crossing . Mouse lines carrying mutations causing male sterility were identified using breeding trials [8] , [9] . The Mot line presented with male sterility with a frequency of one in four individuals , and was thus strongly suggestive of a recessive mutation . Mapping narrowed the causal mutation to a region on chromosome 15 ( bp 64 , 938 , 858 and 93 , 141 , 531 ) containing 49 genes . Of these genes , 32 were expressed within the testis as indicated in EST expression databases , and were thus potentially causal in the Mot phenotype . The protein-coding regions and intron-exon boundaries of all 32 genes were amplified and sequenced and a single homozygous A to G mutation was identified in exon 5 of the Rabl2 gene in all affected males ( Rabl2Mot/Mot ) ( Figure 1A ) . No other mutations were found within the linkage region . Unaffected ( fertile ) males possessed either homozygous wild type alleles ( Rabl2WT/WT ) or were heterozygous for the wild type and Mot allele ( Rabl2WT/Mot ) . The Mot mutation resulted in the substitution of an aspartic acid ( D , negatively charged ) for a glycine ( G , non-polar ) at amino acid 73 of the predicted RABL2 isoforms 1 ( ENSMUST00000058058 ) and 2 ( ENSMUST00000023294 ) , while the predicted isoform 3 ( ENSMUST00000094056 ) would be unchanged as a consequence of exons 3–5 being removed by splicing ( Figure 1A–1B ) . Thus , Rabl2Mot/Mot males were sterile because of the single amino acid substitution in isoforms 1 and 2 of RABL2 . Of note , the frequency and the 100% association between the D73G mutation and male sterility was unchanged following 7 generations of backcrossing onto a pure C57BL6 background adding further weight to the identity of the Rabl2 mutation as causal of the phenotype ( data not shown ) . As indicated results contained herein were generated using mixed background mice . Quantitative PCR analysis on testes of different ages during post-natal development and the establishment of spermatogenesis revealed high levels of isoform 2 and lower levels of isoform 1 mRNAs expression ( Figure 2A ) . Both isoforms 1 and 2 were most highly expressed from post-natal day 18 when haploid germ cells first appear in the germinal epithelium ( Figure 2B–2C ) . Immunofluorescent labelling , using a monoclonal antibody generated against RABL2 isoform 2 , confirmed RABL2 localized predominantly to the haploid compartment of the testis ( Figure 2D ) . Within epididymal sperm , RABL2 was localized to the mid-piece of the tail ( Figure 2D ) . Pre-absorption of the antibody with the immunizing peptide prior to immunofluorescent labelling resulted in the elimination of staining ( Figure 2D inset ) , thus supporting the specificity of immunolablelling . Immunfluorescent labelling of testis sections from Rabl2Mot/Mot animals resulted in a similar localization of RABL2 as that observed in wild type samples ( Figure 2D ) . All Rabl2Mot/Mot males examined were sterile when paired with wild type females ( n>20 for periods of up to 6 months ) . With the exception of uniform male sterility , Rabl2Mot/Mot males were outwardly healthy , had normal body weights ( Figure S1 ) and displayed normal mating behaviour up until at least 10 weeks of age when tissues were harvested for analysis . In order to define the cellular cause of male sterility , mice were phenotyped using the strategy outlined in Borg et al . [10] . Testes from Rabl2Mot/Mot males contained all germ cell types ( Figure 3A–3B ) , however , testes weights were reduced by 15% compared to wild type littermates ( 91 . 8 mg versus 78 . 3 mg , p = 0 . 0094 , n = 6 per group ) ( Figure 3C ) . An analysis of daily sperm production , as indicated by the number of Triton X-100-resistant nuclei , revealed output was reduced by 49% in Rabl2Mot/Mot males ( 1 . 82×106 versus 0 . 93×106 , p = 0 . 046 , n = 6 ) ( Figure 3D ) . Collectively , these data indicate that germ cells were being lost during the latter part of spermatogenesis wherein they contribute relatively little to the overall testis weight . Significant numbers of spermatozoa did reach the epididymis in Rabl2Mot/Mot males , and were thus available for ejaculation and fertilization ( Figure S1B–S1C ) . Sperm structure appeared superficially normal ( Figure 3E versus Figure 3F ) . Similarly , electron microscopy revealed no obvious structural abnormalities in either the axoneme and accessory structures ( outer dense fibers , fibrous sheath and mitochondrial sheath ) ( Figure 3I–3J , n = 2 different genotype animal , 10 sperm per animal ) . Computer assisted sperm analysis of sperm from Rabl2Mot/Mot males , however , revealed low total motility ( 70 . 3% versus 51 . 8% , p = 0 . 0114 , n = 6 per genotype ) and minimal progressive motility ( 44 . 1% versus 7 . 4% , p = 0 . 000002 , n = 6 per genotype ) ( Figure 3G ) . These data strongly suggest that Rabl2Mot/Mot males were sterile as a consequence of an inability of sperm to ascend the female reproductive tract following mating . Sequence analysis identified Rabl2 as an uncharacterized member of the poorly characterized Rab-like clade of the Ras GTPase superfamily [11] . As indicated in Figure S2 , RABL2 contains significant sequence similarity , and a likely evolutionary origin , to members of the RAB family , however , it forms a distinct branch within the superfamily – hence the name ‘RAB-like’ . RABL proteins are defined by being closely related to , but excluded from the Rab clade due to the absence of one or more subfamily specific factors [12]–[14] . In particular , the predicted RABL2 protein from multiple organisms lacks the canonical C-terminal prenylation signal , is not identified as a Rab protein by Rabifier [14] and is excluded from the Rab clade on phylogenetic analysis using a selected training set of Rab sequences that encompass the diversity of Rab proteins in eukaryotes [15] . Although the precise biochemistry remains to be defined , recent data has suggested a role for two RABL proteins in cilia/flagella development . In particular , inactivation of Rabl5 in Trypanosoma brucei resulted in stunted and immotile flagella [6] , and IFT27 ( RABL4 ) is a core component of the IFT ( intra-flagellar transport ) complex B machinery with a role in the anterograde delivery of proteins from the growing flagella tip in Chlamydomonas and RABL4 is also found in mammals [16]–[18] . In common with other members of the complex B , Ift27 RNA interference results in stunted cilia [7] . A RABL2 gene is visible within the human genome on chromosome 22q13 . 33 ( called RABL2B ) [19] ( 88 . 7% protein identity ) and orthologues are present in many species , including the flagellated green algae Chlamydomonas ( 49% identity ) and Trypanosoma species ( 51% identity ) ( Figure 1B ) . In Homo sapiens we note that RABL2 has a paralogues expansion ( RABL2A ) . With only four amino acid changes between the paralogues the functional significance is unclear . The notable exception to the existence of RABL2 orthologues in species containing cilia is in Drosophila , where an orthologue was not identified . The absence of Rabl2 orthologues in organisms lacking cilia/flagella , as defined through a reverse BLAST search , is suggestive of an evolutionarily conserved role for RABL2 in cilia/flagella function , and is highly similar to the phylogenetic distribution of the bona fide IFT factor IFT22/RABL5 [6] . RT-PCR for RABL2A and RABL2B using whole human testis mRNA indicated that both paralogues are expressed in the testis ( data not shown , MKOB ) . EST expression data suggests that RABL2A is expressed in a wide range of tissues including the brain , uterus , testis , lung , eye and prostate as well as a range of cancerous tissues ( Unigene entry Hs . 446425 ) . EST expression data suggests that RABL2B is predominantly expressed in the brain , testis and uterus with lesser amounts in a range of other tissues ( Unigene entry Hs . 584862 ) . Of note , aspartic acid 73 is conserved in all likely orthologues in all species examined and is thus suggestive of it having a critical role in RABL2 function ( Figure 1B ) . Western blotting of extract from equal number of sperm Rabl2WT/WT versus Rabl2Mot/Mot males revealed comparable levels of protein , indicating the phenotype was unlikely to be due to mRNA or protein instability , but rather the specific amino acid change ( Figure 3H ) . These data also demonstrate that the D73 mutation does not impede RABL2 entry into the flagella/cilia compartment . This conclusion is also supported by the immunofluorescent localization of RABL2 in wild type versus mutant testis tissue sections ( consideration should be given to the decreased numbers of elongated spermatids in mutant animals ) ( Figure 2D ) . In common with the RAB GTPases , RABL2 possess all five consensus motifs required for GTP binding ( Figure 1C ) , but it lacks the C-terminal prenylation signal required for membrane interactions classically associated with RAB function [20] . The possession of motifs involved in GTP/GDP binding in RAB proteins ( Figure 1C ) raised the possibility that RABL2 may also cycle between a GTP-bound ‘active’ form and a GDP-bound ‘inactive’ form . In order to predict the effect of the Mot mutation on RABL2 function , and thus the underlying biochemical cause of the observed male sterility , the RABL2 protein sequence was aligned with the structure of multiple related Ras GTPases . Analyses revealed that the Mot mutation occurred in a RABL-specific amino acid in the center of a β-sheet critically involved in mediating a range of protein-protein interactions for the entire superfamily ( for example RAB1B in complex with the guanidine dissociation factor and guanidine exchange factor , GEF , DRRA: pdb identifier 3JZA ( Figure S3 ) . In the context of RAB proteins , GEF proteins function to facilitate the exchange of GDP for GTP in the nucleotide binding site and thus , convert RABs from an inactive state to an active state wherein they are capable of binding to a specific set of effector proteins . If the same biochemistry is maintained in the RABL sub-group , we hypothesize that the Mot mutation would impede RABL2 binding to partner proteins , including with its GEF , and lead to the decreased conversion of inactive GDP-bound RABL2 into active GTP-bound RABL2 . Ultimately , this would lead to the reduced delivery of effector proteins to target locations , in this instance the sperm tail . The sequence homology between RABL2 and RABL4 in Chlamydomonas and RABL5 in Trypanosomes and the sperm tail phenotype in homozygous Mot mutant males , raised the possibility that RABL2 may also interact with components of the IFT machinery and have a role in defining sperm tail ( motile cilia ) length . In order to investigate this hypothesis , we perform immunoprecipitation using testis homogenate with specific IFT complex B component antibodies and then probed for RABL2 using Western blotting . Data revealed an interaction between RABL2 and all of IFT27 , IFT81 and IFT172 ( Figure 4A ) . Of interest , binding to IFT27 and IFT81 was apparently equal in the presence of GTP or GDP . IFT172 however , preferentially bound to GDP-RABL2 . The significance of preferential binding to inactive RABL2 is currently unknown . Immunofluorescent microscopy revealed the co-localization of RABL2 and each of IFT27 , IFT81 and IFT172 within the mid-piece of elongated spermatids within the testis ( Figure 4B–4E ) . Free ( not co-localized with RABL2 ) IFT proteins were frequently observed within the principal piece of developing sperm , thus , raising the possibility that the annulus which sits at the junction between the mid- and principal-pieces may act as a barrier to RABL2 movement . A role for RABL2 in defining sperm tail length was further supported by a 17% reduction in the length of sperm tails from Rabl2Mot/Mot males compared to those from wild type males ( 93 µm versus 111 µm , p<0 . 0001 , n = 4 per genotype ) ( Figure 3K ) . Collectively , these data support a role for RABL2 in the assembly of the sperm tail . At present however , it is not possible to determine whether RABL2 is specifically involved in protein transport along the entire length of the cilia compartment , or if the sperm tail phenotype is secondary to a more generalized defect in RABL2-mediated ( in conjunction with IFT components ) protein transport in the cytosol ( including the mid-piece of the sperm ) . As indicated above , the presence of the 5 motifs characteristic of GTP binding in RABs raises the possibility that RABL2 may cycle between a GTP-bound active state and a GDP-bound inactive state . If true , active GTP-bound RABL2 would be expected to bind to a set of effector proteins and deliver them to the developing sperm tail compartment , as the major site of RABL2 localization and the mediator of the Rabl2Mot/Mot phenotype . In order to test this hypothesis and identify putative effector proteins , recombinant RABL2 ( isoform 2 ) was produced in E . coli and conjugated to agarose in either a GTP-bound or GDP-bound state and incubated with adult mouse testes extracts . Following extensive washing , protein bands preferentially bound to ‘active’ GTP-bound RABL2 were identified using mass spectrometry ( Figure S4 ) . 89 proteins were identified ( Table S1 ) . Of these , five were chosen for further analysis based on known roles in fertility or cilia function , the availability of analytical reagents and confirmation of preferential binding to the GTP-bound form of RABL2 ( Figure 5A ) . Putative effector proteins analyzed included: ATP6V1E1 which is a protein exchanger localized to several ciliated tissues including the olfactoary epithelium [21]; the microtubule plus end trafficking protein EB1 which is involved in centrosome function and primary cilia development in the retina [22]; HK1 which is a component of the fibrous sheath of the sperm tail with a role in glycolysis [23]; the chaperone HSP4AL which has a role in prophase I of meiosis and haploid germ cell development [24]; and LDHC which is also a component of the glycolytic pathway and localized to the fibrous sheath [25] . Of note , genetic or chemical inhibition of LDHC or HK1 results in reduced ATP generation and the inhibition of axoneme microtubule sliding and thus , sperm immotility [23] , [25] . Specific interactions between RABL2 and all five effector proteins were confirmed by co-immunoprecipitation using antibodies directed against the effector proteins ( Figure 5B ) . These data indicate that in common with RAB proteins , GTP-bound RABL2 binds to a specific set of cargo proteins . The effect of the Mot mutation on sperm development and the ultimate fate of the effector proteins was then tested through an examination of the relative effector proteins content , using immunofluorescence and Western blotting , on sperm collected from wild type and Rabl2Mot/Mot mice ( n = 3 mice per genotype ) . Of note , sperm from wild type and Rabl2Mot/Mot mice were immunolabelled and photographed in parallel and under identical conditions . Within sperm from wild type mice ATP6V1E1 , EB1 , HK1 and LDHC were localized primarily to the principal piece of the sperm tail ( Figure 6 ) . Less intense staining for HK1 was observed within the mid-piece of the tail . HSPA4L staining was observed as speckled staining along both the mid- and principal pieces of the sperm tail . In addition , all of ATP6V1E1 , EB1 and HSPA4L were observed in the peri-acrosomal region of the sperm head ( Figure 6 ) . Consistent with the hypothesis that the Mot mutation would lead to decreased delivery of effector proteins , sperm tails from Rabl2Mot/Mot mice contained relatively lesser amounts of all of ATP6V1E1 , EB1 , HK1s , HSP4AL and LDHC than sperm from wild type mice ( Figure 6 ) . The preferential localization of RABL2 to the mid-piece and of the effector proteins to the principal piece of the sperm tail under normal conditions and the residual localization of HK1 to the mid-piece , but not the principal piece , in sperm from Rabl2Mot/Mot males is suggestive of a role for RABL2 in the delivery of effector proteins up to the annulus , but not beyond . By contrast to the apparent decreased content of ATP61E1 , EB1 and HSPA4L in the sperm tail , the association with the peri-acrosomal region of the sperm head was not obviously changed . This data is consistent with IFT transport being a specific requirement for protein transport into the cilia/flagella compartment . Decreased effector protein content in sperm from Rabl2Mot/Mot males compared to sperm from wild types was confirmed by Western blotting ( Figure 6 ) . As a consequence of the apparent specific requirement of RABL2 for tail development , the relative effector protein content per sperm as measured by Western blotting needs to be qualified as it measured total sperm content i . e . head and tail . Regardless , all effector proteins , with the exception of EB1 , were reduced . Collectively , these data demonstrate that RABL2 function is regulated by GTP binding and that the D73 mutation compromised effector protein delivery into the growing sperm tail . The presence of residual levels of effector within the Rabl2Mot/Mot sperm suggest that the Mot mutation is either hypomorphic or some functional redundancy exists with other proteins . In order to define the distribution of RABL2 in the mouse , and thus tissues wherein additional pathology may be anticipated , a tissue survey for Rabl2 expression was undertaken using quantitative PCR methods . Rabl2 isoform 1 and 2 mRNA are both enriched within the male germ line ( Figure 2E–2F ) . Both forms were however , widely expressed and notably within other tissues containing motile cilia including the lung , trachea , brain , ovary and kidney . Isoform 3 was not detected in any tissues examined ( data not shown ) . Through the use of random mutagenesis we have identified RABL2 as an evolutionarily conserved protein with an essential role in male fertility . As far as we can discern , the Mot line is the first model of Rabl2 dysfunction in any species . Here we have demonstrated that RABL2 binds , in a GTP-regulated manner , to a specific set of effector proteins including key proteins involved in cilia development and function and delivers them into the growing sperm tail . Herein we have also defined the first component of a pathway by which components of the fibrous sheath are transported into the tail . Further , analogous to a recent report on the function of RABL5 in Chlamydomonas reinhardtii [26] , RABL2 binds components of the IFT complex B in mammals , and RABL2 dysfunction results in shortened sperm tails . It is the absence of the effector proteins that likely mediates the sterility observed within the Mot mouse line . The reproductive phenotype observed in the Mot mouse lines is remarkably similar to that seen in a sub-group of infertile men . Such men would be classified as having oligoasthenospermia i . e . decreased sperm output and severe motility defects but normal sperm morphology , and would usually be investigated for primary ciliary dyskinesia ( PCD ) [1] , [27] . PCD is usually a recessive syndrome occurring in 1 in 20 , 000–60 , 000 live births ( Mendelian Inheritance in Man no . 232 , 650 ) and is most frequently characterized by recurrent lung disease ( from childhood ) , sinusitis and male infertility in adults [5] , [28] . More variably PCD is associated with hydrocephalus , laterality defects and polycystic kidney disease . Known causes of human infertility include mutations in the axoneme component genes DNAI2 , DNAH5 , DNAH11 , DNAAF2 and LRRC50 , RSPH4A and RSPH9 which contribute to the development of the central tubules of the axoneme , TXNDC which is a thioredoxin and the coiled coiled proteins encoded by CCDC39 and CCDC40 which are involved in the early stages of axoneme assembly [29] , [30] . Collectively however , the underlying aetiology remains unknown in ∼60% of cases . Several additional candidate genes have been identified using mouse studies e . g . Pacrg [31] , Spef2 [32] , Cby [33] , and Pcdp1 [34] . It is clear however , from both human and mouse studies that the composition of cilia ( including motile cilia ) varies subtly between tissues , thus leading to a spectrum of clinical presentations in ciliopathies ( http://v3 . ciliaproteome . org , [35] ) . Regardless , Rabl2 expression data suggests that RABL2 may be a candidate primary ciliary dyskinesia gene .
Animal procedures were approved by the Australian National University and Monash University Animal Experimentation Ethics Committees and performed in accordance with Australian NHMRC Guidelines on Ethics in Animal Experimentation . Point mutant mice were generated on a C57BL/6 background and outbred to CBA for two generations before being inter-crossed as described previously [8] . In order to identify lines containing sterility causing mutations , eight G3 brother-sister pairs per line were co-housed and the presence of pups monitored . If no pups were observed following six weeks , mice were re-paired with wild type partners to determine the origin of the infertility . The presence of copulatory plugs was monitored as an indication of mating behaviour . Lines wherein male sterility was observed in a ratio of approximately one in four with apparently normal mating behaviour were analyzed further . The sterility causing mutation was mapped using a SNP-based method . Genomic DNA from five affected males was hybridized onto Affymetrix 5K mouse SNP Chips at the Australian Genome Research Facility and sequences compared to wild type C57BL6 and CBA sequences . The linkage interval was narrowed using additional mice and SNPs ( www . well . ox . ac . uk/mouse/INBREDS/ ) . SNP typing was performed using the Amplifluor SNP Genotyping System ( Chemicon ) and plates read in a BMG Fluostar Optima fluorescent microplate reader . The sequence of testis expressed candidate genes was determined by sequencing all protein coding exons and ∼50 bp of flanking introns through the Australian Genome Research Facility . Following the identification of the phenotype causing mutation , mice were genotyped using the Amplifluor system using a wild type-specific primer 5′-GAAGGTCGGAGTCAACGGATTACAGAGTTGTGTTCTTGTTGCAGA-3′ , a mutant allele primer 5′-GAAGGTGACCAAGTTCATGCTAGAGTTGTGTTCTTGTTGCAGG-3′ , an antisense primer , 5′-AGCCTTGTGGTAGTAGGAAGCA-3′ and Platinum Taq DNA Polymerase ( Invitrogen ) : 1 cycle , 95°C , 4 min; 35 cycles , 95°C , 10 sec; 60°C , 20 sec: 72°C , 40 sec and a final extension at 72°C , 3 min . Mot infertility was classified using the regime outlined in [10] . Daily sperm outputs were determined using the Triton X-100 nuclear solubilization method as described previously [36] . Sperm motility was assessed using computer assisted sperm analysis ( n = 6/genotype ) [37] and ultra-structure using electron microscopy ( n = 2/genotype , average 10 sperm/mouse ) [38] . Cauda epididymal sperm tail length was measured ( 10 weeks old , n = 4/genotype ) following staining with hematoxylin and eosin . 40 tails/mouse were measured using Imaging technology MetaMorph software ( Molecular Devices ) . To identify RABL2 orthologues , the mouse RABL2 isoform 2 sequence was used as a query in organism specific BLAST searches against protein sequence databases at NCBI . The highest scoring hit was reverse BLASTed against the mouse predicted proteome . Only sequences where mouse RABL2 came up as the highest identify were considered as containing RABL2 orthologues . The species databases searched were Homo sapiens , Macaca mulatta , Trichomonas viginalis , Tetrahymena thermophila , Anolis carolinesis , Monodelphis domestica , Ornithorhynchus anatinus , Danio rerio , Gallus gallus , Xenopus ( Silurana ) tropicalis , Leishmania major , Chlamydomonas reinhardtii , Amphimedon queenslandica , Saccoglossus kowalevskii , Cryptosporidium parvum , Theileria parva , Dictyostelium discoideum , Caenorhabitis elegans , Cryptococcus neoformans , Guillardia theta , Saccharomyces cerevisiae , Physcomitrella patens , Arabidopsis thaliana , Nasonia vitripennis , Plasmodium falciparum , Drosophilia melanogaster , Nematostella vectensis , Monosiga brevicollis , Thalassiosira pseudonana , Phytophthora ramorum , Naegleria gruberi , Toxoplasma gondii , Cyanidioschyzon merolae , Paramecium tetraurelia , Trypanosome cruzi , Trypanosome brucei and Eimeria tenella . A selection of RABL2 orthologues were aligned using ClustalW2 ( http://www . ebi . ac . uk/Tools/msa/clustalw2/ ) . Mus musculus RABL2 transcript 2 sequence ( ENSMUST00000023294 , CCDS49704 ) was aligned with eight other taxa sequences obtained from NCBI ( http://www . ncbi . nlm . nih . gov/ ) . Sequences included: Homo sapiens ( NP_009013 . 1 and NP_001003789 . 1 ) ; Gallus gallus ( XP_424473 . 1 ) ; Xenopus ( Silurana ) tropicalis ( NP_001072451 . 1 ) ; Danio rerio ( NP_001038428 . 1 ) ; Anolis carolinensis ( XP_003229013 . 1 ) ; Leishmania major ( XP_001685503 . 1 ) ; Chlamydomonas reinhardtii ( XP_001697212 . 1 ) ; and Trypanosome brucei ( XP_829561 . 1 ) . Selected RABL2 orthologues were aligned using muscle against a master RAB dataset that contained representatives of all RAB subfamilies previously identified as present in the last eukaryotic common ancestor , together with representative Ran sequences as an outgroup [15] , [39] . The alignment was edited in Mesquite to remove regions of high divergence , and a phylogenetic tree built using Mr Bayes v3 . 2 and RAxML v7 . 0 . 3 [40] , [41] . Highly divergent sequences were deleted and a second round of analysis performed . These data confirm that RABL2 falls within the RAB-like GTPase grouping , and are monophyletic , and hence a distinct paralogous family . The relative expression of Rabl2 during post-natal testis development ( day 0–70 ) and other adult tissues was defined using quantitative PCR using Brilliant Fast SYBR Green QPCR Master Mix ( Stratagene ) in the Agilent Mx 3000P QPCR System: 95°C , 2 min; 95°C , 5 sec; 60°C , 20 sec; 95°C , 2 min; then 72°C , 20 sec , for 50 cycles . Primers were: Rabl2201 5′-TCGATTAGCTGTGGCTTACAAA-3′ and 3′-CCTGTAAAACCTCGTCCATGA-5′; Rabl2202 5′-CTGCACCTGGGTGACAGTAA-3′ and 3′-CATCTTGGGAAGGGAAACAA-5′ . 18S expression was used as a reference in post-natal testis expression analysis . Differential expression was analyzed using the 2ΔΔCT method [42] . N = 3 separate mice per age . Full-length Rabl2 ( isoform 2 ) was amplified from wild type testes cDNA using the primers: 5′-ATAACCCGGGGTTCTGCAGGGGACAGAAACAGGCA-3′ and 5′-ATCTCCATGGTTAGGAAGGAGATGGGCTCTTG-3′ , then cloned into the XmaI and Ncol sites of pET32a ( Novagen ) . Recombinant RABL2 was produced and the histidine tag removed as described previously [43] . N-terminal tagged GST-RABL2 was generated using the Gateway cloning system ( Life Technologies ) according to manufacturer's instruction using pDEST15 vector and primers Rabl2-attB1: 5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTAGAAAACCTGTATTTTCAGGGCGCAGGGGACAGAAACAGG-3′ and Rabl2-attB2: 5′-AGCCCATCTCCTTCCTAATACCCAGCTTTCTTGTACAAAGTGGTCCCC -3′ with pDOR222 and pDEST15 . Recombinant protein was produced as described above . RABL2 mouse monoclonal antibodies were generated against recombinant RABL2 isoform 2 at the Monash University Antibody Technology Facility . Clones were amplified and purified as described [44] . Immunoglobulin was purified from media using Protein G FF sepharose beads ( GE Healthcare ) , dialyzed against PBS and the concentration determined using the DC assay ( Biorad ) . RABL2 and IFT protein immuno-localization was done on fixed frozen testis sections and sperm smear post-fixed in 4% paraformaldehyde . Primary antibodies included RABL2 ( RG2–G8 ) 6 µg/ml , IFT27 4 µg/ml ( Santa Cruz ) , IFT81 10 µg/ml ( Santa Cruz ) , IFT172 6 µg/ml ( Santa Cruz ) . Effector proteins were localized on 1∶1 methanol:acetone fixed sperm [45] . Primary antibodies included: HK1 0 . 4 µg/ml ( Sigma ) , ATP6V1E1 4 µg/ml ( Abcam ) , EB1 4 µg/ml ( Santa Cruz ) , HSPA4L 4 µg/ml ( Abcam ) , LDHC 0 . 5 µg/ml ( Sigma ) . Secondary antibodies included: donkey anti-mouse Alexa Fluor 555 , donkey anti-rabbit Alexa Fluor 488 , Alexa Fluor 555 and donkey anti-goat Alexa 555 ( Life Technologies ) . DNA was visualized using 1 µg/ml 4′ , 6-diamidino-2-phenylindole ( DAPI ) . Images were taken with Nikon C1 Eclipse C1 plus 90i upright automated microscope or Leica SP5 5 channels confocal invert microscope in the Monash University Microimaging Facility . Different excitation lasers ( 408 nm , 488 nm or 561 nm ) were used depending on Alexa fluor dye conjugated to secondary antibody . The specificity of immunolabelling was determined by staining parallel sections in the absence of the primary antibody . The specificity of RABL2 staining was determined by pre-absorbing the antibody with a 500-fold molar excess of the immunizing peptide prior to immunochemistry in parallel with a non-preabsorbed positive control . Putative RABL2 testis effector proteins were identified using RABL2 affinity columns loaded with either GTP or GDP to create active and inactive RABL2 as described previously [46] . Eluted proteins were visualized on 12% SDS-PAGE gels using Coomassie brilliant blue . Proteins that preferentially bound to GTP-RABL2 were excised and sequenced by nano-LC ESI MS/MS at the Australian Proteomics Analysis Facility . Only proteins with two or more unique peptides matches with a Mascot scores of at least 40 were considered as putative effector proteins . Of the 89 non-redundant proteins identified , five were chosen for further analysis based on known roles in sperm or cilia function and the availability of analytical reagents . Preferential binding to GTP-RABL2 was confirmed by Western blotting on additional column eluates and by co-immunoprecipitation using antibodies directed against putative effector proteins . Briefly , 2 mg of adult mouse testis lysate was incubated with 4 µg of the target antibody: HK1 ( Santa Cruz ) , ATP6V1E1 ( Abcam ) , EB1 ( Santa Cruz ) , HSPA4L ( Abcam ) , LDHC ( Abnova ) supplemented with 100 µM GTPγS ( Jena Bioscience ) . Interacting proteins were captured by protein G magnetic beads ( Millipore ) and eluted with 0 . 1 M glycine , pH 2 . 7 . Protein binding to RABL2 was assessed by Western blotting . Negative control reactions included incubation with immunoglobulin of the appropriate species and at a matching concentration . The relative content of RABL2 and effector proteins in sperm was determined using Western blotting of protein from 2×106 sperm per lane from wild type versus mutant mice ( n = 3 , each lane containing sperm from one mouse ) . Samples were separated by 12% SDS-PAGE and transferred onto PVDF membrane . Membrane was blocked using 5% skim milk in PBS followed by incubation with the primary antibody overnight at 4°C . Samples were probed with the following antibodies separately . ATP6V1E1 4 µg/ml , EB1 0 . 4 µg/ml , HSPA4L 4 µg/ml , HK1 0 . 04 µg/ml , LDHC 2 µg/ml . Sample loading was normalized using pro-acrosomal binding protein ( ACRBP ) content and values were averaged over the three mice and compared between genotypes . A sperm head protein was used for normalization as the use of a tail protein e . g . actin , would have skewed the data as a consequence of short sperm tails being a component of the phenotype . Positive band intensity was measured and analyzed by ImageJ ( http://imagej . nih . gov/ ) . P values of <0 . 05 were considered as significant . The ACRBP antibody was raised against the peptides CEMNELYDDSWRSQSTG and CLLRNQNRKMSRMR in goats as described previously [8] , [47] and used at a dilution of 1 in 100 , 000 . The potential for RABL2 to interact with components of the IFT pathway was explored using co-immunoprecipitations with 1 mg of adult testis lysate incubated with 4 µg of IFT172 , IFT81 , IFT27 ( Santa Cruz ) antibodies as discussed above . Immunoprecipitated complexes were separated on 12% SDS-PAGE gels then probed for RABL2 ( 5 µg/ml ) in a Western blot as described above . Student's t-test was used to compare the means of two populations using Graphpad Prism 5 . 0 . P values<0 . 05 was used to define statistical significance .
|
A greater understanding of the mechanism of male fertility is essential in order to address the medical needs of the 1 in 20 men of reproductive age who are infertile . Conversely , there remains a critical need for additional contraceptive options , including those that target male gametes . Towards the aim of filling these knowledge gaps , we have used random mutagenesis to produce the Mot mouse line and to identify RABL2 as an essential regulator of male fertility . Mice carrying a mutant Rabl2 gene are sterile as a consequence of severely compromised sperm motility . Using biochemical approaches we have revealed that RABL2 binds to components of the intraflagellar transport machinery and have identified a number of RABL2 binding ( effector ) proteins . The presence of the Mot mutation in RABL2 leads to a significantly compromised ability to deliver binding proteins into the sperm tail . RABL2 is predominantly produced in male germ cells; however , lower levels are notably produced in organs that contain motile cilia ( hair like structures involved in fluid/cell movement ) , thus raising the possibility that RABL2 may be involved in a broader set of human diseases collectively known as primary cilia dyskinesia .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"medicine",
"urology",
"developmental",
"biology",
"model",
"organisms",
"genetics",
"and",
"genomics",
"biology",
"genomics",
"respiratory",
"medicine",
"pulmonology"
] |
2012
|
RAB-Like 2 Has an Essential Role in Male Fertility, Sperm Intra-Flagellar Transport, and Tail Assembly
|
Centrosome amplification ( CA ) is a common feature of human tumours and a promising target for cancer therapy . However , CA’s pan-cancer prevalence , molecular role in tumourigenesis and therapeutic value in the clinical setting are still largely unexplored . Here , we used a transcriptomic signature ( CA20 ) to characterise the landscape of CA-associated gene expression in 9 , 721 tumours from The Cancer Genome Atlas ( TCGA ) . CA20 is upregulated in cancer and associated with distinct clinical and molecular features of breast cancer , consistently with our experimental CA quantification in patient samples . Moreover , we show that CA20 upregulation is positively associated with genomic instability , alteration of specific chromosomal arms and C>T mutations , and we propose novel molecular players associated with CA in cancer . Finally , high CA20 is associated with poor prognosis and , by integrating drug sensitivity with drug perturbation profiles in cell lines , we identify candidate compounds for selectively targeting cancer cells exhibiting transcriptomic evidence for CA .
The centrosome , an organelle composed of two centrioles surrounded by a pericentriolar protein matrix , is the major microtubule-organising centre of animal cells , hence being pivotal for several fundamental cellular processes , including signalling , cell polarity , division and migration [1–4] . Each centrosome duplicates once per cell cycle to ensure bipolar spindle assembly and successful chromosome segregation [5 , 6] . Centrosomes are thus implicated in the maintenance of genome stability . Centrosome amplification ( CA ) –the presence of more than two centrosomes—is a common feature in cancer [7] . Supernumerary centrosomes generate multipolar mitosis and consequent genome instability [6 , 8–10] , they can accelerate and promote tumourigenesis in vivo [11–13] and promote cellular invasion and metastatic behaviour [14–17] . However , CA’s pan-cancer prevalence , molecular role in tumourigenesis and therapeutic value remain poorly understood , largely due to the technical challenges associated with profiling such small cellular structures in human cancer tissues . For instance , quantifying centrosome numbers and abnormalities is often hampered by the limited thickness of formalin-fixed and paraffin-embedded tissue sections , preventing the imaging of entire cells [18] . In addition , three-dimensional imaging and analysis are mandatory , but cumbersome and time consuming [19] . To at least partially circumvent those challenges , we propose the estimation of CA based on the expression levels of CA-associated genes . Recently , proof-of-principle gene-expression-based CA signatures have been developed [20–23] , the most comprehensive one being CA20 , based on the expression of TUBG1 , which encodes the most abundant centrosomal protein , and 19 other genes whose overexpression has been experimentally shown to induce CA [23] . This signature was proposed to reflect CA levels in tumour samples and shown to have a prognostic value in two independent breast cancer cohorts [23] . In the present study , we used CA20 to estimate relative CA levels across 9 , 721 tumour and 725 matched-normal samples of 32 cancer types from The Cancer Genome Atlas ( TCGA ) , thereby revealing the first pan-cancer landscape of CA-associated gene expression . We show the association of CA20 with distinct breast cancer clinical and molecular features . We also break down the independent associations of CA20 with different sorts of genomic instability across cancer types . Finally , we show that high CA20 is associated with poor clinical outcome in different cancer types , having identified candidate compounds for selectively targeting cancer cells exhibiting transcriptomic evidence for this hallmark of cancer .
To estimate relative CA levels in human samples , we used CA20 , a score based on the expression of 20 genes experimentally associated with CA [23] , as a surrogate . We quantified CA20 across the transcriptomes ( profiled by RNA-seq ) of 9 , 721 tumour and 725 matched-normal samples spanning 32 cancer types from TCGA ( Fig 1a , S1 Table ) . CA20 correlates with the predicted proliferation rates of TCGA tumour samples [24] ( Spearman’s correlation coefficient , r = 0 . 4 , p-value < 2 . 2e-16; S1a Fig ) , as expected , given that some of the CA20 genes encode for proteins involved in cell proliferation . Cervical ( CESC ) , testicular ( TGCT ) and oesophageal ( ESCA ) cancers show high CA20 , contrasting with lower scores in kidney ( KIRP , KICH and KIRC ) and prostate ( PRAD ) cancers ( Fig 1b ) . Some cancer types , such as low-grade glioma ( LGG ) and breast invasive carcinoma ( BRCA ) , exhibit high variability of CA20 , concordantly with previous observations that the proportion of cells with CA in breast tumours ranges from 1 to 100% [7 , 25] depending on the tumour subtype [26] . We also observed significant differences in CA20 between specific cancer types with the same tissue of origin . Although all kidney cancers have low CA20 scores , kidney renal papillary cell carcinoma ( KIRP ) shows a lower score than the other types ( p-value < 0 . 0001 , Wilcoxon rank-sum test; S1b Fig ) . Similarly , glioblastoma multiforme ( GBM ) , skin cutaneous melanoma ( SKCM ) and lung squamous cell carcinoma ( LUSC ) show higher CA20 than low-grade glioma ( LGG ) , uveal melanoma ( UVM ) and lung adenocarcinoma ( LUAD ) , respectively ( p-value < 0 . 0001 for all comparisons , Wilcoxon rank-sum test; S1b Fig ) . We note that squamous cell carcinomas have higher CA20 within cervical ( CESC ) and oesophageal ( ESCA ) cancers ( p-value < 0 . 001 and < 0 . 01 , respectively , Wilcoxon rank-sum test; S1c Fig ) , suggesting that the observed differences are indeed associated to the different cell types of origin and not only to differences between tissue of origin . Since CA has been considered a hallmark of tumour cells [7] , we tested the difference of CA20 between tumour and matched-normal samples . Indeed , tumour samples have higher CA20 levels in all 15 cancer types with both sample types available ( at least 10 samples of each type; False Discovery Rate ( FDR ) < 0 . 0001 , Wilcoxon rank-sum test; Fig 1c ) . In addition , using linear regression analyses with proliferation rate as an additional covariate , we found that CA20 is higher in tumour samples , either when considering all cohorts together ( linear regression p-value < 0 . 0001 , using cohort as an additional covariate; S2 Table ) or per individual cohort ( FDR < 0 . 0001 for all cohorts; S2 Table ) , independently of proliferation rate , discarding the suggestion of CA20 being its mere surrogate . These results emphasise CA as a hallmark of cancer . Breast cancer is one of the best studied cancer types , with large cohorts of clinically annotated tumour samples available [27 , 28] , and where the CA20 score was developed [23] . In addition , CA has been frequently correlated with aggressive features in breast cancer [6 , 25 , 26 , 29] . Given that we observed high variability of CA20 in TCGA breast tumour samples , we sought to investigate in more detail the relationship between CA20 and different breast cancer molecular features in that cohort . CA20 is higher in tumours than in normal breast samples ( p-value < 0 . 0001 , Wilcoxon rank-sum test; Fig 2a ) and we found higher levels of CA20 in invasive tumours from ductal histologic subtype ( the most common , accounting for 90% of tumours ) [30] when compared with lobular ones ( p-value < 0 . 0001 , Wilcoxon rank-sum test; Fig 2b ) . The difference between ductal and lobular subtypes is consistent in non-triple negative breast tumours ( p-value < 0 . 0001 , Wilcoxon rank-sum test; S2d Fig ) , as well as in samples from tumour stages II and III ( p-value < 0 . 0001 and < 0 . 01 , respectively , Wilcoxon rank-sum test; S2e Fig ) . We also tested the differences in CA20 between the different PAM50 molecular subtypes , derived based on a 50-gene classifier [31] . Basal-like breast tumours have the highest CA20 scores ( p-value < 0 . 0001 , p-value < 0 . 0001 , and p-value < 0 . 001 for contrasts with luminal A , luminal B , and HER2-enriched , respectively , Wilcoxon rank-sum test; Fig 2c ) . This is in line with our recent work experimentally showing that basal-like breast cancers have indeed more CA than luminal ones [26] . We also observed a strong difference between luminal subtypes , with higher CA20 in luminal B samples ( p-value < 0 . 0001 , Wilcoxon rank-sum test; Fig 2c ) . Moreover , we tested the association between CA20 and tumour stage , having found a significant CA20 increase from stage I to stage II ( p-value < 0 . 0001 , Wilcoxon rank-sum test; Fig 2d ) , but no significant changes between subsequent stages ( Fig 2d ) . All associations between CA20 and breast cancer histology , PAM50 molecular subtypes and tumour stage remain significant within both low and high proliferating tumours ( samples divided by the median of estimated proliferation rates; S2a–S2c Fig ) . All the aforementioned associations were validated in an independent cohort ( Fig 2e–2h , S2f and S2g Fig and S3 Table ) , comprising 144 normal and 1 , 992 tumour breast samples from the Molecular Taxonomy of Breast Cancer International Consortium ( METABRIC ) [28] . We still tested the association between CA20 and the METABRIC integrative clusters , 10 molecular subgroups defined based on joint clustering of copy number and gene expression data [28] . CA20 varies across integrative clusters ( p-value < 0 . 0001 , Fligner-Killeen test ) and is particularly enriched in cluster 10 ( FDR < 0 . 0001 , Wilcoxon rank-sum test , for comparisons with each of all the other clusters; S2h Fig ) , characterized by high proportion of basal-like tumours , high genomic instability , high rate of TP53 mutations , chromosome arm 5q deletions and very poor prognosis in the short term [28] . We complementarily analysed the frequency of CA in human breast carcinomas from the different PAM50 molecular subtypes , comprising 29 luminal A , 3 luminal B , 3 HER2 and 13 basal-like tumours ( Fig 2i and S4 Table ) . Concordantly with TCGA and METABRIC results , we observed a higher percentage of cells with supernumerary centrioles in luminal B ( average of 27% ) than in luminal A carcinomas ( 7%; p-value < 0 . 05 , Wilcoxon rank-sum test; Fig 2j and S3 Fig ) . Moreover , basal-like ( 25% ) display higher levels of CA than luminal A tumours ( p-value < 0 . 0001 , Wilcoxon rank-sum test ) . Despite the reduced number of luminal B samples , our patient data support CA20 as a good surrogate of CA levels and the suggestion that CA is more frequent in luminal B than in luminal A human breast carcinomas . CA and consequent multipolar mitoses have been associated with aneuploidy , genomic instability and tumourigenesis for more than a century [32 , 33] . Using the available quantitative characterization of aneuploidy in TCGA [34] , we found that CA20 is higher in samples with genome doubling ( p-value < 0 . 0001 , Wilcoxon rank-sum test; Fig 3a ) and positively correlated with their aneuploidy score ( measured as the total number of altered—gained or lost—chromosome arms; Spearman’s correlation coefficient , r = 0 . 44 , p-value < 2 . 2e-16; Fig 3b ) . Although CA20 is positively correlated with both chromosomal deletions and amplifications ( Spearman’s correlation coefficient , r = 0 . 41 and 0 . 36 , p-value < 2 . 2e-16 , respectively; S4a and S4b Fig ) , it is more strongly associated with chromosomal deletions ( p-value < 2 . 2e-16 , t-test for z-transformed coefficients; see also S4c Fig ) . Given the known association between loss of p53 and CA [6 , 7 , 35] and the recent observation that p53 null cells tend to have an enrichment of chromosome losses over gains [36] , we tested the hypothesis that the observed association between CA20 and chromosomal deletions could be linked to TP53 mutations . However , the increase in the proportion of deletions per sample from low to high CA20 samples is consistent within both TP53 wild-type and mutated samples ( p-value < 0 . 0001 and < 0 . 05 , respectively , Wilcoxon rank-sum test; S4d and S4e Fig ) , showing it is independent of TP53 mutations ( two-way ANOVA p-value for interaction = 0 . 6; S4d Fig ) . Investigating the hypothesis that CA20-associated aneuploidy levels could vary between chromosomes , we identified 20 chromosome arms whose deletion ( 10 arms ) or amplification ( 10 arms ) was enriched in tumour samples with higher CA20 ( linear regression , FDR < 0 . 05; Fig 3c and S2 and S5 Tables ) . The strongest associations were with loss of 5q , 16p and 7p . Interestingly , 5q deletion was previously associated with CA20-high basal-like breast tumours [27 , 37–40] and METABRIC integrative cluster 10 [28] ( Fig 2c and 2g and S2h Fig ) . The association between CA20 and 5q deletion remains when removing the breast cancer cohort ( linear regression p-value < 2 . 2e-16; S5 Fig and S2 Table ) . This observation raises the question if matched-normal samples of the analysed tumour samples have a CA20 signal predictive of those 5q , 16p and 7p deletions . We tested this hypothesis by comparing the CA20 levels between normal samples ( with intact tested chromosomal arms ) whose matched tumours lost 5q , 16p or 7p , with those with tumours with amplifications or no alterations in those chromosomal arms . We found that normal samples whose matched tumours lost 5q or 16p exhibit higher CA20 scores ( p-value < 0 . 01 and < 0 . 05 , respectively , Wilcoxon rank-sum test; S6 Fig ) , therefore suggesting that a CA20 increase may precede those chromosomal abnormalities . In addition to tumour aneuploidy , CA20 is positively correlated with mutation burden , number of somatic Copy Number Alterations ( CNA ) and number of clones per tumour ( Spearman’s correlation coefficient , r = 0 . 48 , 0 . 47 and 0 . 43 , respectively , p-value < 2 . 2e-16 for all; Fig 3d–3f ) . All these associations are independent of cell proliferation ( linear regression p-values < 1e-8 for all; S2 Table and S7 Fig ) . We found that the correlation with mutation burden holds for different types of mutations ( silent , missense , splice site and nonsense ) , as well as for mutations shown to be pathogenic ( data from ClinVar https://www . ncbi . nlm . nih . gov/clinvar/ ) in all diseases and particularly in cancer ( S8 Fig ) . Since these genomic instability features are likely correlated between each other , we applied multiple linear regression analyses across 1050 tumour samples ( from 12 different cancer types; minimum of 30 and average of 88 samples per cohort ) with information for those 4 covariates ( S6 Table ) . We identified independent positive associations between CA20 and all genomic instability features , with stronger association for CNAs ( linear regression p-values = 1 . 3e-5 , 7 . 2e-4 , 5 . 3e-10 and 6 . 4e-3 for aneuploidy , mutation burden , CNA and number for clones , respectively; Fig 3g and S2 Table ) . These associations remain significant when proliferation rate is used as an additional covariate in the regression ( p-values = 2 . 3e-5 , 7e-4 , 2 . 4e-9 and 0 . 03 for aneuploidy , mutation burden , CNA and number for clones , respectively; S2 Table ) . We performed similar analyses per TCGA cohort and identified a group of cancer types where CA20 is mostly associated with CNA and aneuploidy ( prostate adenocarcinoma , glioblastoma multiforme , bladder urothelial carcinoma , and brain low-grade glioma; Fig 3g and S9 Fig; S2 Table ) . Although CA has been globally associated with genomic instability , these results highlight CNA as the main associated feature and show that these associations differ between cancer types . Point mutations are one of the most common types of mutational events that impact the stability of a cancer genome . We examined the pan-cancer association between CA20 and somatic mutations in 14 , 589 genes and found 752 whose mutations are associated with CA20 ( FDR < 0 . 05; Fig 4a and S2 and S7 Tables ) . Most significant associations of mutated genes with the CA20 score are positive , consistently with its correlation with higher mutation burden ( Fig 3d ) , and enriched in cancer driver genes ( Gene Set Enrichment Analysis ( GSEA ) [41 , 42] p-value < 0 . 001 , using a list of 299 cancer driver genes derived from TCGA’s PanCancer analysis [43]; S10a Fig ) . TP53 shows the strongest association ( linear regression p-value < 0 . 0001; Fig 4a ) , with positive correlations for the majority of cancer types surveyed ( 10 out of 17 cancer types with at least 20 mutated samples; FDR < 0 . 05; Fig 4b ) , therefore putatively extending the reported association between loss of p53 and CA [6 , 7 , 35] to 10 different cancer types . The second strongest positive association is with tumour suppressor pRb ( RB1 ) , whose acute loss has been found to induce CA [44] . Unexpectedly , the strongest negative association is with E-cadherin ( encoded by CDH1 ) , meaning CDH1-mutated samples have lower CA20 levels . Given its tumour suppressor role in cancer and the fact that its mutations mostly induce loss of function [45] , this result suggests loss of E-cadherin is associated with lower CA in human tumours , which is contrary to what have been reported in epithelial cancer cells [46] . GSEA on genes whose mutations are associated with CA20 found that they are enriched in cancer-associated pathways and Wnt/β-catenin signalling ( S10c–S10f Fig ) . As only a small fraction of somatic mutations represent driver events , we repeated the pan-cancer analysis of association between CA20 and somatic mutations using likely driver mutations from the Cancer Genome Interpreter ( https://www . cancergenomeinterpreter . org/mutations ) [45] . Within the tested 33 genes with at least 10 mutated samples , we found three ( TP53 , PIK3CA and EGFR ) whose driver mutations are associated with CA20 ( FDR < 0 . 05; S10b Fig and S2 and S8 Tables ) , TP53 being again the strongest association . Overall , we show that CA20 is associated with both passenger and driver mutational spectra in cancer , with particular enrichment in cancer driver genes and Wnt/β-catenin signalling . CA has still been proposed as a driver of genomic instability [11] . We thus wondered if the DNA mutation spectrum associated with CA was similar to specific signatures of somatic mutations caused by different mutational processes in cancer [47] . We therefore retrieved the contribution of the 30 published mutational signatures for each TCGA tumour sample from mSignatureDB [48] and uncovered three of them positively associated with CA20: signature 3 , associated with BRCA1/2 mutations; signature 13 , attributed to APOBEC activity; and signature 4 , characteristic of smoking’s mutational pattern ( FDR < 0 . 05; S11 Fig ) . As these signatures are likely confounded with genomic instability , we performed multiple linear regression on CA20 including , as independent variables , the mutational signature and the four aforementioned genomic instability features: aneuploidy , mutation burden , CNA and number of clones per tumour ( S2 Table ) . Signature 1 , linked with ageing and characterised by C>T substitutions ( S12a Fig ) , and its “reverse” ( T>C substitution bias ) Signature 12 , found mainly in liver cancer ( S12b Fig ) , are respectively positively and negatively associated ( FDR < 0 . 05 ) with CA20 ( Fig 4c ) , independently of other types of genomic instability and even when proliferation rate is added as a variable ( FDR = 0 . 051 for both signatures ) . To evaluate the putative causality of CA20-associated mutations ( Fig 4a ) , we interrogated the Connectivity Map ( CMap ) database of signatures [49] about the impact of each of the 3 , 799 gene knock-downs on the CA20 gene set in human cancer cell lines . The resultant connectivity scores ( S9 Table ) , ranging from 100 ( CA20 up-regulation ) to -100 ( CA20 down-regulation ) , were compared with the pan-cancer association between somatic mutations in the cognate genes and CA20 ( Fig 4d ) . We thereby identified 6 genes with a putative causal effect on CA20 scores ( |connectivity score| > 80; Fig 4d ) : P2RY12 , RB1 , ITSN1 and MYCBP2 are putative inhibitors of CA ( their knock-down up-regulate CA20 genes ) , whereas ABCC5 and COPA are putative promoters of CA ( their knock-down down-regulate CA20 genes ) . Although acute loss of pRb ( encoded by RB1 ) has been found to induce CA [44] , confirming pRb as a CA inhibitor , to our knowledge none of the remaining genes identified herein has been previously associated with CA . They are therefore interesting candidates for future functional studies . Genes from a manually curated list of centriole duplication factors ( 93 genes , including only 10 from the CA20 signature; S10 Table ) are enriched in negative CMap knock-down scores ( GSEA p-value < 0 . 001; Fig 4e ) , suggesting they are indeed needed for cells to express CA-associated genes . Using the MSigDB’s Hallmark Gene Sets library [50] , we identified unfolded protein response and mitotic spindle as significantly enriched in genes whose knock-down showed negative scores , i . e . CA20 down-regulation ( GSEA FDR < 0 . 05; Fig 4f ) . This association suggests that mitotic spindle components activate CA-associated genes and/or that cells highly expressing CA-associated genes may be less likely to survive when their mitotic spindle is perturbed . CA has been associated with poor patient prognosis in a variety of cancer types [7] . We therefore tested CA20’s association with overall patient’s survival across 31 TCGA cancer types with more than 40 samples each , finding high CA20 significantly associated with worse prognosis in 8 different cancer types ( FDR < 0 . 05 , log-rank test; Fig 5a and S11 Table ) . This result supports the potential of CA20 for prognostic-based patient stratification . Hypoxia is a potent microenvironmental factor promoting genetic instability and malignant progression [51–53] . Given that hypoxia has been shown to enhance centrosome aberrations in breast cancer [54 , 55] , we investigated whether CA20 is associated with the relative hypoxia levels in TCGA tumour samples , given by a previously established surrogate metagene expression signature [56] . We found a positive correlation between CA20 and the hypoxia score ( Spearman’s correlation coefficient , r = 0 . 61 , p-value < 2 . 2e-16; Fig 5b ) that is independent of genomic instability ( linear regression p-value = 7 . 8e-9; S2 Table ) . We further confirmed that this association is independent of estimated proliferation rates ( linear regression p-value = 5 . 6e-7 when proliferation rate is added as a covariate to the regression; S13a Fig and S2 Table ) . We also performed this linear regression analysis for each of the 12 TCGA cohorts with information for all covariates and identified three cancer types ( glioblastoma multiforme , lung adenocarcinoma and bladder urothelial carcinoma ) where hypoxia is positively associated ( FDR < 0 . 05 ) with CA20 ( Fig 5c; S2 Table ) . Although a tumour is also composed by stromal and immune cells [57] , the association between CA and tumour cellular composition has not been addressed yet . CA20 is associated with lower stromal ( Spearman’s correlation coefficient , r = -0 . 52 , p-value < 2 . 2e-16; Fig 5d ) and immune ( Spearman’s correlation coefficient , r = -0 . 34 , p-value < 2 . 2e-16; S13c Fig ) cell infiltration in TCGA . However , pan-cancer linear regression analyses revealed that only the negative association with stromal infiltration is independent of genomic instability ( linear regression p-value = 2 . 7e-6 and 0 . 24 , for stromal and immune , respectively; S2 Table ) . The same was observed when including proliferation rate as an additional covariate ( linear regression p-value = 1 . 2e-4 and 0 . 21 , respectively; S13b and S13d Fig and S2 Table ) . We have also performed similar analyses for each of the 5 TCGA cohorts with information for all covariates and found that CA20 is significantly associated ( FDR < 0 . 05 ) with lower stromal infiltration in head and neck and lung cancers ( Fig 5e ) , with lower immune infiltration in glioblastoma , and with higher immune infiltration in head and neck cancer ( S13e Fig ) , all independently of genomic instability ( S2 Table ) . CA is a hallmark of cancer cells and hence an appealing target in cancer therapy . In order to identify compounds that could target cancer cells with such abnormality , we have employed CA20 to estimate relative CA levels in 823 human cancer cell lines from the Cancer Therapeutics Response Portal ( CTRP ) [58] ( S12 Table ) , for which both transcriptomic and drug-sensitivity profiles are publicly available . Correlation analyses between CA20 and drug-sensitivity ( in Area Under the dose-response Curve , AUC ) for 354 compounds revealed 81 negatively correlated with CA20 ( FDR < 0 . 05 , Spearman’s correlation; Fig 6a and S13 Table ) , i . e . higher CA20 was associated with lower drug AUC and , therefore , higher drug activity . The enrichment of negative correlations ( S14 Fig ) may reflect the bias for cancer-targeting compounds in CTRP . These results suggest several candidate compounds to selectively kill cancer cells with CA , such as 3-CI-AHPC , CD-437 , STF-31 , methotrexate , BI-2536 and clofarabine ( Fig 6b ) . The first three are probes , methotrexate and clofarabine are U . S . Food and Drug Administration ( FDA ) -approved drugs for several cancer types ( https://www . cancer . gov/about-cancer/treatment/drugs/methotrexate ) and paediatric acute lymphoblastic leukemia ( https://www . cancer . gov/about-cancer/treatment/drugs/fda-clofarabine ) , respectively . Interestingly , BI-2536 has been in clinical trials for several solid and liquid tumours ( https://clinicaltrials . gov/ct2/results ? cond=&term=bi+2536&cntry=&state=&city=&dist ) and is an inhibitor of polo-like kinase 1 ( PLK1 ) , whose inhibition has already been associated with CA suppression [59 , 60] . Complementarily , we mined the CMap database to identify compounds that could impact the CA20 score and therefore putatively reduce/increase CA levels . We calculated the impact of 2 , 837 compounds on the CA20 transcriptomic levels in human cancer cell lines ( S14 Table ) and identified some whose activity drove CA20 up-regulation ( putative CA promoters; S15 Fig ) , such as VEGFR2-kinase-inhibitor-IV , dienestrol ( oestrogen receptor agonist ) and sulforaphane ( anticancer agent in clinical trials for Bladder , Breast , Lung and Prostate cancers; https://clinicaltrials . gov/ct2/results ? cond=sulforaphane&Search=Apply&recrs=d&age_v=&gndr=&type=&rslt= ) . We also identified compounds that down-regulated CA20 , such as two CDK inhibitors ( purvalanol-a and aminopurvalanol-a ) , JAK3-inhibitor-VI , etoposide ( topoisomerase and cell cycle inhibitor ) and CD-437 ( agonist of RARG , retinoic acid receptor gamma; Fig 6c ) . For the 164 drugs tested in both datasets , we observed a positive correlation between their CA20/sensitivity correlations in CTRP and their CMap scores ( Spearman’s correlation coefficient , r = 0 . 26 , p-value = 8 . 3e-4; Fig 6d ) , indicating that drugs selectively targeting cells with higher CA20 are reducing the expression of these genes , possibly by killing the abnormal cells in the tumour cell population . These complementary approaches uncovered RARG’s agonist CD-437 as the strongest candidate for targeting CA . Moreover , drugs targeting coagulation factor II ( F2R ) , farnesyltransferase ( FNTA and FNTB ) , ubiquitin isopeptidases ( USP13 and USP5 ) , DNA topoisomerase II alpha ( TOP2A ) and cyclin-dependent kinases ( CDKs ) are also promising candidates ( Fig 6d ) . Given cell proliferation’s association with CA20 ( S1a Fig ) , we have tested the association between its estimated rates across TCGA primary tumour samples and the expression of the 164 compounds’ predicted target genes ( merging this information from the CTRP ( S13 Table ) and CMap datasets ( S14 Table ) ) , using linear regression analyses with cohort as additional covariate ( S2 Table ) . The resultant coefficients ( S15 Table ) are not correlated with CMap’s average scores of the respective compounds ( Spearman’s correlation coefficient , r = 0 . 016 , p-value = 0 . 84; S16a Fig ) , but are correlated with their CTRP’s Spearman correlation coefficients ( Spearman’s correlation coefficient , r = -0 . 26 , p-value = 9e-04; S16b Fig ) , i . e . compounds selective for cells with high CA20 are predicted to target genes positively associated with proliferation in TCGA tumour samples . Nevertheless , predicted target genes of several compound candidates from our analyses do not show strong association with proliferation ( S16c and S16d Fig ) . These results need to be considered when prioritizing candidate compounds for further experiments aiming to target cancer cells through CA .
CA is known to promote tumourigenesis but its molecular role therein remains elusive and , although it is also suggested to be a promising target for cancer therapy , CA’s prevalence in different types of cancer and therapeutic value in the clinic are still pretty much unprobed . Using the CA20 signature and TCGA RNA-seq data , we characterise the landscape of CA-associated gene expression in a broad range of cancer types , thereby demonstrating the potential of using gene expression-based signatures in multi-omic and clinical data integrative approaches to investigate the biological and medical relevance of their respective cellular and molecular processes . Despite the lack of a full direct experimental validation of CA20 as a surrogate of CA levels , our observations are very consistent with known CA’s features , namely CA20’s upregulation in cancer [7] and in basal-like breast tumours [26] , and its association with the knock-down of centriole duplication factors , genomic instability [11] , loss of p53 [6 , 7 , 35] and pRB [44] , hypoxia [54 , 55] and worse patient’s prognosis [7] . In addition , we found that luminal B breast tumours have higher prevalence of CA than luminal A ones , concordantly with the observed differences in the CA20 score between the two molecular subtypes in two independent cohorts . Finally , we have analysed two transcriptomic datasets of multiciliogenesis , where cells escape centriole number regulation to generate hundreds of centrioles during differentiation [61] , and found that CA20 increases during the centriole overduplication stage , resuming basal levels afterwards ( S17 Fig ) , suggesting CA20 as a marker of active amplification . These observations vouch for the present proof-of-concept study to pave the way for more in-depth and bona fide findings when CA’s transcriptomic signature is experimentally refined . Moreover , here we already propose novel hypotheses that will trigger studies aiming at a more comprehensive understanding of the role of CA in cancer . We observed higher CA-associated gene expression in cancer samples of squamous cell origin than in adenocarcinomas , suggesting that their different cell types of origin can have different CA’s prevalence and/or ways to cope with this abnormality . Previous work has indeed shown that CA triggers spontaneous squamous cell carcinomas , lymphomas and sarcomas , but not adenocarcinomas , in mice [11] . We also show that breast invasive carcinoma samples have high variability on CA20 , concordantly with previous observations [7 , 25] , that is related to their distinct clinical and molecular features . We had recently shown that basal-like breast carcinomas have higher CA than luminal tumours [26] , but here we report for the first time an upregulation of CA-associated genes in tumours from both invasive ductal histologic subtype and luminal B molecular subtype . We validated the CA20-based predictions by quantitatively analysing centrosome numbers in human breast carcinoma samples , where we found that indeed CA is more prevalent in luminal B than luminal A tumours , providing a novel insight into the differences between these two hormone-receptor positive molecular subtypes . Given the limited number of luminal B samples in our cohort , more extensive analyses are necessary to confirm this association . Our data show that centrosome amplification is associated with breast cancer clinical features and endorses the potential of using a gene-expression-based signature for patient stratification . CA-associated gene expression upregulation is positively correlated with different types of genomic instability , like aneuploidy , mutation burden , CNA and tumour heterogeneity . In particular , CA20 is more strongly associated with chromosomal deletions than amplifications , independently of TP53 mutations . We speculate that this association may be due to the impact of CA in cellular genomic stability having non-random genomic “hot spots” . In fact , through a more detailed analysis , we found an association with alterations in specific chromosomal arms , that may be due to the localisation of genes encoding for regulators of CA20 genes therein and/or to those arms’ higher susceptibility to the genomic instability triggered by centrosome abnormalities . The latter is supported by recent work showing that human chromosome mis-segregation is not random and can be biased by inherent properties of individual chromosomes [62] , and also by our observation that normal samples whose matched tumours lost 5q or 16p have higher CA20 predictive of those deletions ( S6 Fig ) . Moreover , we characterised the DNA mutation spectrum associated with CA20 and found it to be enriched in C>T mutations , a signature characteristic of ageing , with which centrosome aberrations have also been associated [63–67] . Genes whose mutations are associated with CA20 are enriched in cancer driver genes , and particularly in Wnt/β-catenin signalling . Wnt/β-catenin signalling components interact with the centrosome [68] and a previous study has demonstrated that mutant β-catenin induces centrosome aberrations in normal epithelial cells and is required for CA in cancer cells [69] . Our results extend this previous association to human cancer samples , suggesting mutations in β-catenin might contribute to the observed CA in cancer . Finally , we show the usefulness of a novel approach whereby we integrated information on genes whose somatic mutations are associated with CA20 in TCGA tumour samples with the impact of their knock-downs on the CA20 expression in human cancer cell lines , aiming at unveiling candidate molecular players in CA in cancer . Concordantly with previous work on CA [7] , we observed that high CA20 is associated with poor patient’s survival in several cancer types . Furthermore , we found a positive correlation between CA20 and hypoxic levels in glioblastoma multiforme that is particularly interesting , due to its highly hypoxic microenvironment and HIF-1α levels [70] , also shown to enhance migration and invasion of its tumour cells [71 , 72] . Given the observed association between CA and invasion of tumour cells [15 , 17] , an exciting hypothesis is hypoxia-induced invasion being mediated through CA . When looking at the tumour cellular composition , we found that tumours with high CA20 have lower stromal and immune cell infiltration , although the latter is not independent of tumour genomic instability and proliferation rate . Detailed studies aiming to decouple these effects could provide relevant molecular insights when considering immunotherapy , alone or in combination with genotoxic and/or anti-proliferative therapeutic approaches . Moreover , by pioneering the integration of drug sensitivity with drug perturbation profiles in human cancer cell lines , we identify candidate compounds for selectively targeting cancer cells exhibiting transcriptomic evidence for CA . These compounds could be particularly useful in the treatment of cancer types we identified as having high CA and to whose current therapy patients respond poorly . For instance , their potential in specifically targeting basal-like and luminal B breast tumours could be assessed by taking advantage of resources like patient-derived tumour xenografts [73] . The observed ability of cells carrying extra centrosomes to manipulate the surrounding tumour cells and promote their invasiveness [15 , 17] suggests that targeting the former may be clinically more impactful . Given CA’s cancer-specificity , the compounds identified herein could underlie the development of novel targeted cancer therapeutic options .
The study with human samples was conducted under the national regulative law for the handling of biological specimens from tumour banks , with samples being exclusively used for research purposes in retrospective studies , and was approved by the ethics committee of the Hospital Xeral-Cies , Vigo , Spain . Informed consent was obtained from all human participants . Publicly available RNAseqV2 ( quantified through RNA-seq by Expectation Maximization ) [74] and clinical data for 9 , 721 tumour and 725 matched-normal samples from The Cancer Genome Atlas ( TCGA; https://cancergenome . nih . gov/ ) were downloaded from Firebrowse ( http://firebrowse . org/ ) . Gene expression ( read counts ) data were quantile-normalized using voom [75] . For each sample , the CA20 score was calculated as the sum of the across-sample ( including both tumours and matched-normal samples ) normalized ( log2 median-centred ) expression levels of the CA20 published signature genes [23]: AURKA , CCNA2 , CCND1 , CCNE2 , CDK1 , CEP63 , CEP152 , E2F1 , E2F2 , LMO4 , MDM2 , MYCN , NDRG1 , NEK2 , PIN1 , PLK1 , PLK4 , SASS6 , STIL and TUBG1 ( Fig 1a ) . Predicted proliferation rates of each TCGA tumour sample were retrieved from [24] ( n = 9 , 568 ) . Whole genome doubling ( corresponding to 0 , 1 and ≥ 2 genome doubling events in the clonal evolution of the cancer ) , aneuploidy ( both aneuploidy score—number of altered chromosome arms—and alterations per chromosome arm ) and mutation burden characterizations were retrieved from [34] ( n = 9 , 166 ) . Since the chromosomal arm status was not available for TCGA normal samples , we have selected only those with no CNA in the chromosomal arms tested , to make sure they are intact . CNA ( n = 8 , 879; copy number levels were derived with the GISTIC algorithm [76] and considered as CNA if having a score lower than -1 ( loss ) or higher than 1 ( gain ) ) and mutation ( n = 7 , 120; including classification as silent , missense , splice site or nonsense ones ) processed data were downloaded from Firebrowse ( http://firebrowse . org/ ) . Mutations were classified as likely pathogenic and pathogenic based on ClinVar database’s ( https://www . ncbi . nlm . nih . gov/clinvar/ ) variant summary annotation ( ftp://ftp . ncbi . nlm . nih . gov/pub/clinvar/tab_delimited/variant_summary . txt . gz; accessed in November 12th 2018 ) , and 5 , 601 likely driver mutations were obtained from the Cancer Genome Interpreter ( https://www . cancergenomeinterpreter . org/mutations; accessed in November 12th 2018 ) [45] . The list of 299 cancer driver genes was retrieved from [43] . Intra-tumour heterogeneity data , measured by the number of clones per sample , were retrieved from [77] ( n = 1 , 080 ) . The mutational signature profiles were retrieved from mSignatureDB [48] ( n = 9 , 004 ) . The predicted fraction of stromal ( stromal score ) and immune ( immune score ) cells in TCGA tumour samples ( n = 2 , 463 ) was retrieved from [78] . We used the scores calculated based on RNASeqV2 expression levels . Importantly , no CA20 gene was used by the authors to infer those cell proportions [78] . TCGA tumour samples were analysed for hypoxic status based on expression of 95 genes included in the hypoxia 99-metagene signature [56] . The four missing genes are three ( LOC149464 , LOC56901 and TIMM23 ) for which expression levels were not available and NDRG1 , excluded for being part of the CA20 gene signature . The hypoxia score was calculated like the CA20 score . Additional clinical information for TCGA breast tumour samples was retrieved from [27] . Normalized gene expression data for 1992 primary breast tumours and 144 normal breast tissue samples from the Molecular Taxonomy of Breast Cancer International Consortium ( METABRIC ) [28] were retrieved from European Genome-Phenome Archive ( EGAC00001000484 ) . Gene expression was profiled with Illumina HT-12 v3 microarrays , with probe-level intensity values being mean-summarised per gene . The CA20 score was calculated as for the TCGA dataset . Clinical information for the same samples was downloaded from cBioPortal ( http://www . cbioportal . org/ ) [79] . Quantification of CA in breast cancer samples was performed as described in [26] . Briefly , formalin-fixed and paraffin-embedded human breast carcinoma samples were consecutively retrieved from the files of the Department of Pathology , Hospital Xeral-Cies , Vigo , Spain . This series comprises 29 luminal A , 3 luminal B , 3 HER2 and 13 basal-like tumours . Some of these samples had already been used in one of our recent studies [26] . The status of the oestrogen receptor ( ER ) , progesterone receptor ( PR ) , epidermal growth factor receptor 2 ( HER2 ) , antigen Ki67 , and the basal markers epidermal growth factor receptor , cytokeratin 5 , cytokeratin 14 , P-cadherin and Vimentin was previously characterized for all tumour cases . According to their immunoprofile , breast tumour samples were classified as luminal A ( ER+ , PR+ , HER2− and Ki67− ) , luminal B ( ER+ , PR+ , HER2 overexpressing or Ki67+ ) , HER2 ( ER- , PR- , HER2 overexpressing ) or basal-like carcinomas ( ER− , PR− , HER2− , basal marker+ ) . Representative tumour areas were carefully selected and at least two tissue cores ( 0 . 6 mm in diameter ) were deposited into a tissue microarray . This study was conducted under the national regulative law for the handling of biological specimens from tumour banks , with samples being exclusively used for research purposes in retrospective studies . Informed consent was obtained from all human participants . For immunofluorescence staining , 3 μm-thick tissue sections were deparaffinised in Clear-Rite-3 ( Thermo Scientific , USA , CA ) and rehydrated using a series of solutions with decreasing concentrations of ethanol . High temperature ( 98 °C , 60 min ) antigenic retrieval with Tris-EDTA pH = 9 . 0 ( LeicaBio systems , UK ) was performed , followed by incubation with UltraVision protein block ( Thermo Scientific ) for 30 min at room temperature . The slides were , afterwards , incubated with mouse anti-GT335 ( 1/800 dilution , Adipogen Ref . AG- 20B-0020-C100 ) and rabbit anti-pericentrin ( 1/250 dilution , Abcam AB4448 ) in UltraAb diluent ( Thermo Scientific ) overnight at 4 °C . The sections were then washed three times , 5 min per wash , with 1× PBS + 0 . 02% Tween20 before a 1 h room temperature incubation with the secondary antibodies , anti-IgG rabbit coupled to Alexa 488 and anti-IgG mouse coupled to Alexa-594 ( Invitrogen ) , diluted at 1/500 in PBS . Finally , sections were washed extensively with 1× PBS + 0 . 02% Tween20 and then counterstained and mounted with Vectashield containing DAPI ( VectorLabs , CA , USA ) . Imaging was performed on a Zeiss Imager Z1 inverted microscope , equipped with an AxioCam MRm camera ( Zeiss ) and ApoTome ( Zeiss ) , using the ×100 1 . 4 NA Oil immersion objective . Images were taken as Z-stacks in a range of 10–14 μm , with a distance between planes of 0 . 3 μm , and were deconvolved with AxioVision 4 . 8 . 1 software ( Zeiss ) . Only the structures positive for GT335 ( centriolar marker ) and pericentrin ( PCM marker ) were analysed and scored . Between 5 and 107 cells were analysed for each patient and cells with more than 4 centrioles were considered as having CA ( S4 Table ) . Normalized gene-level expression and drug sensitivity ( n = 481 compounds ) data for 823 human cancer cell lines from the Cancer Therapeutics Response Portal ( CTRP ) v2 were retrieved from [58] . The CA20 score was calculated as for the aforementioned datasets . Compounds with more than 20% of missing data ( n = 127 ) were removed from the analyses . Area Under the dose-response Curve ( AUC ) was used as the metric of cell line’s drug sensitivity , measured over a 16-point concentration range . Note that lower AUC means higher drug activity . The Connectivity Map ( CMap ) database of signatures [49] was interrogated using CA20 genes as an individual query in the CLUE L1000 tool ( https://clue . io/l1000-query#individual , login required; CA20 genes were used as putative UP-regulated genes ) . For each of the 9 human cancer cell lines profiled within the Touchstone dataset ( PC3 , VCAP , A375 , A549 , HA1E , HCC515 , HT29 , MCF7 and HEPG2 ) , a connectivity score was computed per perturbation ( gene knock-down , gene overexpression , small molecule administration ) [49] , reflecting its effect on the expression of CA20 genes ( except for SASS6 , not profiled in this dataset ) . We calculated an average connectivity score per perturbation by averaging the 9 cell lines’ connectivity scores in order to have a more robust connectivity score that can be used across different cell types and tissues . Two types of perturbations were analysed: 3 , 799 gene knock-downs and 2 , 837 compounds . The Broad compound ID was used to match the 164 compounds tested by CMap and CTRP , so that the results of the analyses of the two datasets could be combined . Normalized gene expression data for adult mouse airway epithelial cells during multiciliogenesis ( triplicates for three different time points: days 0 , 2 and 4 ) was retrieved from [80] ( GEO dataset accession GSE73331 ) . The CA20 score was calculated as for the TCGA dataset . The transcriptomic alterations between non-ciliating mouse tracheal epithelial cells and those undergoing differentiation , through transition to an air-liquid interface culture ( ALI ) , and harvested at four ( ALI+4 ) or twelve ( ALI+12 ) days , were retrieved from [81] . Those probe-level transcriptomic alterations were mean-summarised per gene . Spearman’s correlations were performed using the cor . test R function ( method = ‘‘spearman” ) [82] . The difference between two Spearman’s correlations was tested using the paired . r function from R package psych [83] . Wilcoxon rank-sum tests were performed using the wilcox . test R function [82] . Multiple linear regression modelling was implemented using the lm function from R package limma [84] . Covariate collinearity was tested using the corvif function from [85] , in which all covariates had a variance inflation factor below 2 . All equations and respective statistics are shown in S2 Table . We have normalised the genomic instability covariates using z-scores ( number of standard deviations from the mean ) to account for differences in the prevalence of aneuploidy , mutation burden , CNA and number of clones per cohort . Fligner-Killeen test was implemented using the fligner . test R function [82] . Proportions tests were performed using the prop . test R function [82] . Two-way ANOVA was done using the aov R function [82] . Unsupervised hierarchical clustering of the multiple linear regression results per cancer type was performed using the heatmap . 2 function from R package gplots [86] . Genes ranked according to the knock-down connectivity score were analysed for pathway enrichment using Gene Set Enrichment Analysis [41 , 42] with default parameters . We used a list of 299 cancer driver genes from [43] , a manually curated list of centriole duplication factors ( 93 genes , including 10 from the CA20 signature; S10 Table ) , gene sets retrieved from the KEGG pathway database ( https://www . kegg . jp/ ) and the MSigDB’s Hallmark Gene Sets library [50] . Those with a False Discovery Rate ( FDR ) lower than 5% were considered significant . Dividing patients into two subgroups by CA20 median value , the significance of differences in prognostic was estimated using Kaplan−Meier plots and log-rank tests , per cancer type , through R package survival [87] . To calculate the expected Spearman’s correlation coefficients and p-values used in the quantile-quantile ( Q-Q ) plot ( S14 Fig ) , we permutated 1000 times the drug-sensitivity ( in AUC ) of all compounds across cell lines and , for each permutated dataset , we calculated the respective CA20-AUC Spearman’s correlations . The expected values were obtained by median-summarizing the ranked 1000 permutations’ results .
|
Centrosome amplification , i . e . an increased number of centrosomes—structures that exist inside cells , is a hallmark of cancer cells and therefore an Achilles' heel for the development of innovative therapies that specifically target tumour cells , sparing healthy ones . To exploit centrosome amplification’s clinical potential , it is crucial to understand its role in cancer development and to identify compounds for its selective targeting . These are challenging tasks due to the technical difficulty of profiling centrosome amplification in cells . In this study , we circumvent those challenges by computationally analysing the expression of 20 genes known to promote centrosome amplification across nearly 10 , 000 tumours of over 30 cancer types , thereby estimating their relative centrosome amplification levels . We found that those genes are indeed highly active in tumours and associated with prognosis in different cancer types . We also show that those genes’ expression is associated with instability in the structure of cancer cells’ chromosomes and identify candidate drugs for selectively targeting those cells . Our work therefore demonstrates the potential of computational analyses of large volumes of cancer molecular and clinical data to elucidate cellular and molecular mechanisms of tumour development and propose novel therapeutic options in oncology .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"cancer",
"genomics",
"medicine",
"and",
"health",
"sciences",
"breast",
"tumors",
"squamous",
"cell",
"lung",
"carcinoma",
"genetic",
"networks",
"statistics",
"carcinomas",
"cancers",
"and",
"neoplasms",
"basic",
"cancer",
"research",
"oncology",
"regression",
"analysis",
"mathematics",
"network",
"analysis",
"adenocarcinomas",
"head",
"and",
"neck",
"tumors",
"research",
"and",
"analysis",
"methods",
"head",
"and",
"neck",
"squamous",
"cell",
"carcinoma",
"computer",
"and",
"information",
"sciences",
"lung",
"and",
"intrathoracic",
"tumors",
"mathematical",
"and",
"statistical",
"techniques",
"head",
"and",
"neck",
"cancers",
"linear",
"regression",
"analysis",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"squamous",
"cell",
"carcinomas",
"physical",
"sciences",
"genomics",
"statistical",
"methods",
"genomic",
"medicine"
] |
2019
|
Pan-cancer association of a centrosome amplification gene expression signature with genomic alterations and clinical outcome
|
How stable synchrony in neuronal networks is sustained in the presence of conduction delays is an open question . The Dynamic Clamp was used to measure phase resetting curves ( PRCs ) for entorhinal cortical cells , and then to construct networks of two such neurons . PRCs were in general Type I ( all advances or all delays ) or weakly type II with a small region at early phases with the opposite type of resetting . We used previously developed theoretical methods based on PRCs under the assumption of pulsatile coupling to predict the delays that synchronize these hybrid circuits . For excitatory coupling , synchrony was predicted and observed only with no delay and for delays greater than half a network period that cause each neuron to receive an input late in its firing cycle and almost immediately fire an action potential . Synchronization for these long delays was surprisingly tight and robust to the noise and heterogeneity inherent in a biological system . In contrast to excitatory coupling , inhibitory coupling led to antiphase for no delay , very short delays and delays close to a network period , but to near-synchrony for a wide range of relatively short delays . PRC-based methods show that conduction delays can stabilize synchrony in several ways , including neutralizing a discontinuity introduced by strong inhibition , favoring synchrony in the case of noisy bistability , and avoiding an initial destabilizing region of a weakly type II PRC . PRCs can identify optimal conduction delays favoring synchronization at a given frequency , and also predict robustness to noise and heterogeneity .
Several lines of evidence indicate that synchronous activity in the hippocampal formation is important for learning and memory . Coherent activity arises when animals are in states of active locomotion and information acquisition [1] , [2] . Disabling coherent theta activity leads to memory impairment [3] , [4] . Synchronous oscillations at gamma frequency have been implicated in binding of sensory experiences [5] and attention [6] . Computational models incorporating nested theta-gamma oscillations are well-suited to associative and sequence-learning tasks [7] , [8] , underscoring the potential importance of synchronous activity . Although many studies have analyzed systems of coupled oscillators , few [9] have incorporated the physical constraints of axonal conduction delays; therefore , there is a gap in our understanding of how distal neural modules can synchronize [10] that will be addressed in the proposed work . We use circuits constructed from stellate cells and pyramidal cells from the entorhinal cortex ( EC ) in rats in order to search for general principles of synchronization in the presence of conduction delays that may include multiple intervening synapses . Stellate cells in particular have been implicated as potential theta pacemakers [11] . Previously , Netoff et al . [12] used the Dynamic Clamp [13] , [14] to measure the spike time response curves ( STRC ) for isolated layer 2 stellate cells in entorhinal cortex . The spike time response curve plots the change in cycle period due to a synaptic input as a function of the point in the cycle at which the input is received; in this study , we normalize the change in cycle period by the intrinsic period and call this the phase resetting curve ( PRC ) . Using a strictly phenomenological criterion , Type I STRCs and PRCs contain either advances or delays whereas Type II contain both [15] . The STRCs ( and PRCs ) observed in response to excitation consisted of advances at most , but not all phases . The resetting was nearly zero at phases of zero and one with a peak near the center . There was a small region of small delays at very early phases; the presence of this region makes them weakly type II rather than Type I [15] . For inhibition , the PRCs consisted of only delays , hence they were Type I , but instead of having a peak in the center , the delays were monotonically increasing with phase . Netoff et al . [12] used the dynamic clamp to construct hybrid circuits of two biological neurons coupled by artificial synapses , or of one biological and one model cell . Their study , like ours , did not make any inferences regarding the response of the neurons to very weak pulses , but instead used the measured STRC directly to predict network activity , under the assumption that the pulsatile nature of the coupling made it likely that the effect of an isolated synaptic input was not changed by the mutual coupling within the network . The method successfully predicted that with no delays incorporated in the circuit , mutually excitatory circuits of stellate cells synchronized , whereas mutually inhibitory cells fired in antiphase . Recently , Woodman and Canavier [16] derived existence and stability criteria for 1∶1 phase locking in a network of two oscillators reciprocally pulse-coupled with conduction delays . Pulse coupled means that the interaction between the coupled oscillators takes the form of brief pulses that can be approximated by delta functions with infinitesimal duration . The locking point for each oscillator is defined as the phase within its own cycle at which it receives an input from its partner during a one to one periodic locking in the network . For synchronous modes in circuits of identical oscillators , the phase at which an input is received by each oscillator is the delay divided by the intrinsic period of the oscillator , so increasing the delay above zero shifts the locking point along the PRC from zero phase to larger values of the phase . The key characteristic of the PRC that determines whether a 1∶1 locking such as synchrony is stable is the slope of the PRC at the locking point . Thus , depending upon the shape of the PRC , some delays will produce stable synchrony whereas others will not . Here we extend the work of Netoff et al . [12] on two coupled oscillators to include conduction delays , and to investigate the robustness of the synchronous solution to the heterogeneity and noise inherent in biological networks .
It is not practical to think that we can map out in detail the exact connectivity and characterize in detail every oscillatory element in brain circuits responsible for the synchrony that may underlie cognition . Instead , we seek to understand the general principles that underlie collective synchronous activity . Thus our general approach , illustrated schematically in Fig . 1 , was to catalogue the representative characteristics of our circuit elements ( EC neurons ) , then to use the range of characteristics of the individual components to predict and explain the range of collective activities observed when the individual components are connected in a network with conduction delays . To this end , we constructed very simple circuits in which we had complete control over the connectivity between the biological circuit components . We used the Dynamic Clamp both to characterize the synchronization tendencies of individual neurons ( Fig . 1A1 ) and to build simple networks ( Fig . 1B1 ) . The Dynamic Clamp is an electrophysiological method that allows one or more living cells to interface with a computer in real time . In the instances that spontaneous synaptic activity was observed in the biological neurons , the inputs were blocked pharmacologically ( see Methods ) , and then virtual synapses were created as follows . The dynamic clamp sampled the membrane potential Vmem in the soma of the biological neurons every 100 µs , then calculated and injected a synaptic current into each of the form ISYN = g ( t ) ( Vmem−Esyn ) as described in the Methods . The same type of virtual synapse was used to characterize the phase response curve for each oscillator as the virtual synapses used in the hybrid networks . We measured PRCs using dynamic clamp experiments ( Fig . 1A2 ) by applying either an excitatory or inhibitory synaptic current at various stimulus intervals ( ts ) after a spike to determine the recovery time ( tr ) until the next spike . The duration of the perturbed cycle ( Pj = ts+tr ) is in general different from the duration of the average free running unperturbed cycle Pi . The normalized difference in cycle period is the phase resetting , which was calculated by the equation fj ( φ ) = ( Pj−Pi ) /Pi , where Pj is the length of the cycle that contains the perturbation and plotted as a function of the phase at which the input was applied ( Fig . 1A3 ) . The phase ( φ ) is estimated by normalizing the stimulus interval ts by the average intrinsic period Pi . The phase resetting was quite noisy , and curve fitting was used to determine the general shapes of the phase resetting that we can expect to encounter in these cells . We then used the Dynamic Clamp to build simple two neuron networks ( Fig . 1B1 ) . The time-dependent synaptic conductance waveform g ( t ) was triggered in this case by a spike in the partner with an adjustable delay . Example voltage traces from a hybrid circuit experiment ( Fig . 1B2 ) show a one-to-one locking in which there are two measurable time lags: the interval ( time lag 1 ) between a spike in neuron 1 and the next spike in neuron 2 , and the interval ( time lag 2 ) between a spike in neuron 2 and the next spike in neuron 1 . The average values of these time lags were measured for all pairs at different values of conduction delay between the neurons . Delays and time lags were normalized by the uncoupled period of the two neurons , which was set as nearly as possible to a single constant value using DC current ( see Methods ) . The main goal of this study was to use the PRCs that were typically observed experimentally in order to account for the degree of synchronization actually observed in hybrid circuits , without knowing the exact PRC for each neuron in every circuit . We measured a total of 24 PRCs ( 17 using a virtual excitatory synapse and 7 using an inhibitory one ) . Fig . 2 shows that two general classes of PRCs were observed for both inhibitory and excitatory coupling . In the convention used in this paper , a positive value of phase resetting means that the cycle period was lengthened , causing a delay before the next spike is emitted . A negative value corresponds to an advance in the time that a spike is emitted . Type I PRCs consist of either all advances or all delays , whereas Type II PRCs have a mix of the two [15] . Here we use these categories in a purely phenomenological sense , and make no implications regarding the excitability type [17] or bifurcation structure [18] . We found that for our data , the order of the best polynomial fit operationally allowed us to categorize Type I and Type II PRCs; those with a single extremum ( implying a second order polynomial ) in the best fit were consistent with Type I , whereas those with a higher order polynomial best fit were more consistent with Type II . Specifically , the 7 excitatory Type I PRCs exhibited only advances and a single inhibitory Type I PRC exhibited only delays . Ten excitatory Type II PRCs exhibited small delays at very early phases and advances at all other phases , and 6 inhibitory Type II PRCs exhibited small advances at very early phases and delays at all other phases . Thus the Type II PRCs were only weakly Type II . The best fit curve is an estimate of the mean PRC , and the mean plus or minus a single standard deviation is shown ( thin curves ) to give an idea of the phase dependence of the variability observed in the phase resetting . Consistent with [12] , for inhibitory PRCs the variance was not strongly phase dependent , but for excitatory PRCs the variability decreased at late phases . Excitatory Type I PRCs had a negative slope at early phases , but a positive slope late , whereas the opposite was true for inhibition . Inhibitory PRCs did not appear to return to zero at a phase approaching one , also consistent with those reported by Netoff et al . [12] . Excitatory Type II PRCs had an initial and final positive slope but negative in the middle , whereas inhibitory Type II PRCs had an initial region of negative slope followed by a large region of positive slope . A second region of negative slope was sometimes observed at very late phases . The slopes have important implications for stability of phase locking; in our convention , negative slopes are destabilizing and positive slopes are stabilizing . Prior to turning the coupling on between two neurons , steady current was injected to cause the neurons to fire repetitively at similar frequencies ( 7∼10 Hz ) . Synchronization within a circuit was evaluated by constructing histograms of the time lags observed while the neurons were coupled via the dynamic clamp . Composite data in Fig . 3 from two excitatory hybrid circuits illustrate representative firing patterns observed in this type of circuit . Fig . 3A shows the peaks in the time lag histograms associated with the firing patterns illustrated in Fig . 3B . In a synchronous mode , one time lag is zero and the other is equal to the normalized network period . Synchrony was not in general observed in excitatory hybrid circuits with small delays . Instead , modes with one time lag that was roughly equal to the delay were observed at short delays . If the neurons are sufficiently homogeneous , then either neuron can lead , resulting in bistability between two firing patterns . We call this mode ( Fig . 3B1 ) a leader/follower mode [19]–[21] because the firing of the leader evokes a spike in the follower ( but not vice versa ) after a delay equal to the time lag . In the first cycle , the red neuron leads , but leader switching is frequently observed , and the blue neuron leads in the last two cycles . Since the free-running periods of the two neurons were adjusted to be as nearly equal as possible , noise induced leader switches presumably due to bistability are not surprising . As the delays were increased to intermediate values , the leader/follower pattern transitioned to a near anti-phase mode ( Fig . 3B2 ) . At delays greater than half the period of the slower neuron , a sharp transition to synchrony was observed in which one value of the time lag was quite close to zero . Since there are two types of PRCs , a two-neuron circuit may be composed of two type I cells , two type II cells , or one of each . In order to determine if the observed activity could be explained using the PRCs , we used previously published theoretical methods [16] to predict the time lags corresponding to stable one-to-one lockings for each combination of phase resetting curves , using the representative examples from Fig . 2A and initially assuming that both neurons in the circuit had the same intrinsic period . We need to make the following assumptions in order to use PRCs to analyze network behavior of coupled neurons . 1 ) Each neuron is a pacemaker , i . e . a limit cycle oscillator , and remains so in the neural circuit . 2 ) The effect of single perturbation decays before the next input is received . This implies that the perturbed neuron returns immediately back to the limit cycle , otherwise the phase would be undefined when the input is received . 3 ) The perturbations that the neuron receives in a closed loop configuration are similar to ones that are used to generate the open loop PRCs . Given these assumptions , we can calculate the periodic one-to-one locked firing patterns that are consistent with the phase resetting tendencies of both neurons at any given value of conduction delay as illustrated in Fig . 4 . The stimulus and recovery intervals can easily be calculated under these assumptions for any value of the phase φ , thus these intervals can be considered a function of the phase at which an input is received . The stimulus interval is Piφ , and from the definition of the phase resetting we can obtain that the recovery interval is Pi−Piφ+Pi f ( φ ) . Although we can calculate stimulus and recovery intervals for any arbitrary phase , only certain pairs of phases ( φ1 , φ2 ) , where the subscript indicates the neuron receiving the input at a given phase , can satisfy the periodicity constraints for a one to one locking . In the presence of delays , it takes k cycles , where k is an integer , for the firing of one neuron to affect the next firing time in the same neuron . Fig . 4A1 and A2 illustrate the periodicity constraints for k = 1 and Fig . 4B1 and B2 illustrates them for k = 2 . Briefly , twice the delay value plus the response interval ( 2δ+tri ) in one neuron by definition must be equal to the stimulus interval in the other neuron plus k−1 times the network period ( tsi+ ( k−1 ) ( tsi+tri ) ) ; this is true for both neurons resulting in two separate criteria that must both be satisfied in a one to one locking . For each neuron ( neuron 1 in black and neuron 2 in red ) we can plot these quantities at each phase as in Fig . 4C in order to find the intersections . The axes are selected so that at the intersections the abscissa and ordinate values for the red and black curves are equal so both periodicity conditions are satisfied . Furthermore , we can use the slopes of the phase resetting curve at the locking points to determine whether the firing patterns are stable and therefore observable in the presence of noise . If the slope of the black curve in Fig . 4C is steeper than that of the red curve at the point of intersection , the point is stable; otherwise it is unstable ( see Fig . S1 ) . The slopes of the PRC curve at the locking points determine the slopes of the red and black curves at the intersection points corresponding to one to one lockings; generally a positive slope of the PRC in our convention is stabilizing and a negative one is destabilizing ( for an exact treatment see [16] ) . An intuitive explanation can be given for a slight perturbation from synchrony in a two neuron circuit; the neuron that fires too early receives an input at phase greater than the locking point , so for a positive slope at the locking point it is delayed more ( or advanced less ) than it would be at the locking point and therefore fires less early in the next cycle causing convergence to the locking point . The final step in the prediction method is to use the algebraic relationships depicted in Fig . 4A and 4B to determine the values of the time lags ( see Fig . 4D ) given the stimulus and recovery intervals and the delay values . Fig . 4C and 4D specifically illustrate the PRC prediction method for two neurons with unit period and type I PRCs illustrated in Fig . 2 A1 . At a delay that is 0 . 04 times the period ( Fig . 4C1 ) , there are three possible periodic one to one lockings , all with a k value of 1 , as in the firing pattern shown in Fig . 4A; for that delay value , no other k values produce an intersection . The filled circle in the center is unstable and is ignored . The two open circles indicate modes in which the observable time lags are unequal . However , since the two neurons are identical , there are two firing patterns corresponding to these two time lags because either neuron can lead . This can account for the bistable leader follower mode observed experimentally in Fig . 3A . At a longer delay that is 0 . 40 times the period ( Fig . 4C2 ) , the antiphase mode with two identical time lags , again at a k value of 1 , becomes stable . At even longer delays of 0 . 80 times the period ( Fig . 4C3 ) , the only intersection appears in the plot for a k value of 2 , as in the firing pattern shown in Fig . 4B . This intersection produces a stable synchronous mode with one time lag equal to zero and the other equal to the network period PN , which is equal to the sum of the time lags ( tl1 and tl2 ) as well as the sum of the stimulus and recovery intervals in either neuron . The prediction results at each delay value are summarized in Fig . 4D . The X symbols show the values of the time lags calculated from the intersection points , and the gray circles show the network period , which is the sum of the times lags . For the antiphase mode at a delay of 0 . 40 , for example , the two time lags overlay each other at exactly half the network period . Since this study was not limited to weak coupling , the network period can differ quite noticeably from the intrinsic period ( assumed to be equal to one in this example ) because of the resetting experienced by each neuron in the network . Consequently , the two points at each value of delay are not constrained to have a sum equal to one . A total of twelve hybrid circuits were constructed from eight pairs of biological neurons coupled by excitation , of which four pairs were coupled at two different conductance values . Clear peaks in the histograms indicating one-to-one phase locking with preferred time lags were evident in all but one experiment . We excluded the data from that experiment , which happened to be from one of the pairs in which experiments were conducted at two conductance values . The circuit that did not lock had the weakest conductance value used in any experiment , and was apparently too weak to induce phase locking at any value of delay recorded . Data from the eleven phase-locked circuits is summarized in Fig . 5A , using a different symbol for each circuit . The summary data shows the same dependence of the observed firing pattern on the conduction delay that was clearly illustrated in Fig . 3 . We then compared the results of the PRC prediction method for all possible combinations of PRC type . The predicted values of the normalized time lags ( X symbols ) for each of the three cases are plotted in Fig . 5B , C and D . For the circuits that contain at least one Type I neuron ( Fig . 5B and D ) , the predicted activity follows the same trends as the experimental data in Figs . 3A and 5A . In particular , Fig . 5B and D show prominent leader-follower behavior for normalized delays less than 0 . 2–0 . 3 . For circuits of two identical Type II neurons , synchrony rather than leader follower mode was predicted for short delays , but if the periods were allowed to vary by a few percent ( 4% in the circles in Fig . 5C ) , then early synchrony was disrupted and the general trends observed in the experimental data , including an approximate leader/follower mode at short delays , were restored . Therefore the experimental results for excitatory hybrid circuits are quite consistent from what is expected using PRC theory . More importantly , the use of the PRC methods allows us to gain insight into why the observed firing patterns are favored . Specifically , synchrony is only observed for zero delay and delays that are longer than half the intrinsic periods . First we will discuss why synchrony is not observed at short delays , and then we explain why it is observed for longer ones . With no delay , the locking point for synchrony is at a phase of zero because each neuron affects the other immediately upon spike initiation at a phase of zero . As the delay is increased , the locking point is moved to the right along the PRC [16] to a phase equal to the delay divided by the intrinsic period ( see Fig . 6A ) . The slope of the Type I PRC shown in Fig . 2A1 is negative for over half the cycle and therefore destabilizes synchrony for delays less than half an intrinsic period . Thus we would not expect synchrony for short delays in any circuit that contains a Type I PRC . On the other hand , the Type II PRC shown in Fig . 2A2 has an early region of stable positive slope for phases less than about 0 . 15 , so we might expect synchrony in the cases in which the hybrid circuit happens to contain two neurons with Type II PRCs . This synchrony , however , results from the symmetry of two identical , identically coupled oscillators [22] , and was easily disrupted in Fig . 5C by the introduction of heterogeneity in the period . It is striking , however , that for all panels in Fig . 5 , long delays ( greater than 0 . 6 to 0 . 8 of the intrinsic period depending upon the specific example ) produced robust synchronization that was not disrupted by heterogeneity either in the hybrid circuits ( Fig . 5A ) or in the coupled PRCs ( Fig . 5C ) . Not coincidentally , this robust synchrony occurs when the locking point nears the causal limit region of the PRCs towards the end of a cycle when an excitation almost immediately evokes a spike . Thus the magnitude of the phase advance is equal to the fraction of the cycle remaining at the time the input is given ( 1−φ ) . Under our sign convention , this produces a linear region ( φ−1 ) in the PRC with a positive slope of one that is strongly stabilizing . In order to explain why early but not late synchrony was disrupted by heterogeneity , we quantified the degree to which small deviations from synchrony caused by heterogeneity can be quantified in terms of the PRC . Assuming that the PRC for two neurons is identical , but that they have a small difference in intrinsic period , we can derive this expression for the nonzero time lag ε when the synchronous solution is disrupted by the unequal intrinsic periods P1 and P2 and a difference Δδ in the delays δ , where δ represents the average of the two delays:where f is the phase resetting and the prime indicates the slope . This expression is valid for small ε+Δδ/2 so that the PRC can be linearized ( see Fig . 6B and the derivation in Text S1 ) . For identical neurons , the numerator in the fractional term is zero , which allows exact synchrony for equal delays ( Δδ = 0 ) . In Fig . 5C , equal delays but non-identical intrinsic periods cause a nonzero phase lag in the near synchronous solution ( circles versus crosses ) . Increasing the slope of the phase resetting curve by increasing the conductance is not an effective strategy to minimize the time lags because very large positive slopes are also destabilizing . For the synchronous solution , the absolute value of one minus the sum of the slopes at the two locking points ( one for each neuron ) needs to be less than one for stability [16] , [23] . At late phases for strong coupling , the PRCs in the vicinity of the locking point are essentially the same and linear at the causal limit , so the expression given above for the nonzero time lag is quite valid . At the causal limit , f ( δ/Pi ) = δ/Pi−1 and the numerator goes to zero even with different intrinsic periods or PRCs as long as the locking point for both is on the causal limit . Because the slope of each PRC is one in this region , the time lag ε reduces to ( Δδ ) /2 , which is zero for identical delays . Exactly on the causal limit , synchrony becomes neutrally stable in theory; however , the causal limit cannot be physically achieved because some time must elapse between an action potential in the leader and the one it evokes in the follower . We previously mentioned that an abrupt transition to synchrony ( Fig . 3A ) was observed between intermediate ( Fig . 3B2 ) and long delays ( Fig . 3B3 ) . This abrupt transition as the delay was increased was sometimes accompanied by an abrupt increase in the tightness of the phase locking . Fig . 7A1 shows the histogram of the times lags for the antiphase mode illustrated in Fig . 3B2 in a hybrid circuit with a normalized delay of about 0 . 57 . Rather than plotting both times lags on the same axis , as in Figs . 3A , 4D and 5 , here we have plotted time lag 1 ( see Fig . 1B2 ) as positive and time lag 2 as negative so that we get two distinct peaks for antiphase . Circular statistics ( see Methods ) showed that the circuit was locked at a network phase of 0 . 5 with an R2 = 0 . 7 . Fig . 7B1 shows a histogram of the time lags for the synchrony illustrated in Fig . 3B3 at a normalized delay of about 0 . 71 . The histogram for synchrony has a peak at zero and peaks at ± the network period depending on which neuron is considered to fire first , but in this case one peak is smaller than the other indicating that the faster neuron fired first more often , breaking symmetry . The peaks for synchrony had a narrower width indicating tighter locking than in the antiphase example , as confirmed by circular statistics indicating a network phase of 0 for synchrony with an R2 = . 87 . The transition from antiphase to synchrony is evident in each panel of Fig . 5 where a clear antiphase mode with equal or roughly equal time lags at intermediate delays is replaced by synchrony as the delays are lengthened by at most 10% or 15% of the period from the clear antiphase mode . We can explain the narrowing of the histogram peaks by assuming that in this example , both cells in the hybrid circuit had identical Type II PRCs as in Fig . 7C . In this hypothetical circuit , for an antiphase mode the locking point for the individual is not at 0 . 5 phase but rather at the phase that satisfies φ = 0 . 5+f ( φ ) /2+δ/Pi , because of contributions from nonzero phase resetting and from the conduction delays . For a normalized delay of 0 . 55 , the locking point corresponding to antiphase has shifted far enough to the right to “wrap around” a phase of one and land on the initial stable branch of the PRC with positive slope ( filled circle labeled A in Fig . 7C ) . This branch is quite noisy . As the normalized delay is increased , antiphase loses stability as the locking point moves onto the middle unstable branch of the PRC . The locking point for the individual neurons in a synchronous mode is not at zero phase , but rather at the normalized delay value . As the normalized delay value increases to 0 . 9 , the locking point for synchrony ( open circle marked B in Fig . 7C ) falls in the causal limit region of the phase resetting curve ( dashed line ) where an excitatory input reliably triggers a spike with short latency , reducing the noisy variability . We suspected that the sudden decrease in the width of the histogram of network phases observed in the transition from antiphase to synchrony ( Fig 7A1 compared to 7B1 ) could be accounted for by a switch in the locking point on the PRC from a region of high variability to a region of lower variability . In order to test this possibility , we constructed a noisy iterated map ( see Methods ) based on the PRC [12] , [24] by initializing each neuron in a simulated hybrid circuit at an arbitrary phase , polling the neurons to see which one would fire next , updating the phase of the partner to the firing time , keeping track of input emission and delayed arrivals , and resetting the phases appropriately when an input arrived . The phase resetting was a random Gaussian variable with the mean and the variance determined at each phase by the experimental data . Previous such maps [12] , [24] did not include the greater complexity encountered in the presence of delays . The noisy map produced the broad histogram peaks shown in Fig . 7A2 for the antiphase mode with a normalized delay of 0 . 55 . Circular statistics gave an R2 = . 86 at a network phase of 0 . 5 . On the other hand , the noisy map produced the narrow histogram peaks shown in Fig . 7B2 for the synchronous mode with a normalized delay of 0 . 9 . Circular statistics gave an R2 = . 98 at a network phase of 0 , confirming that the phase dependent variance of the PRC and specifically the decrease in the variance at very late phases , can provide a possible explanation for the tighter phase locking that is sometimes observed in the transition to synchrony as the delay is increased . Fig . 8A shows the summary data for 6 hybrid circuits coupled by inhibition . The tendency was to exhibit antiphase at the shortest delays and near synchrony at slightly longer delays of up to 0 . 7 times the intrinsic period , the largest values explored in these circuits . At the longest delays examined , out of the three cell pairs tested at these delays , two pairs ( X symbols and open squares ) exhibited a transition from near synchrony to near anti-phase as the delay was increased , suggesting that if longer delays were applied the other cell pairs would have undergone the transition as well . Many of the time lags are quite close to zero , across a broad range of phases including relatively early phases , in contrast to the hybrid circuits coupled by excitation . Once again we used each possible combination of PRC type within a circuit to predict how the observed time lags should vary as the delay is increased , and the predicted pattern of the dependence of network activity on the delay shown in Fig . 8B–D conforms to the overall pattern obtained experimentally in Fig . 8A . For a circuit with two identical neurons with Type I PRCs ( X symbols in Fig . 8B ) synchronous modes are predicted for a large range of phases corresponding to normalized delays from 0 to 0 . 75 , because the slope of the Type I PRC ( Fig . 2B1 ) is positive in that range . Introducing heterogeneity in the form of a 4% difference in cycle period only slightly perturbs the synchronous solution ( gray filled circles in Fig . 8B ) , except at very short normalized delays ( <0 . 1 ) . For some normalized delays ( less than about 0 . 1 and between 0 . 5 and 0 . 7 ) , antiphase is bistable with synchrony , so either mode could theoretically be observed depending upon the initial conditions . The same level of heterogeneity disrupts early but not late bistability . For a circuit with two identical neurons with Type II PRCs ( X symbols in Fig . 8C ) , synchronous modes are predicted at phases between about 0 . 1 and 0 . 85 because the slope of the Type II PRC is positive in that region . There is also a region of bistability with antiphase from about 0 . 1 to 0 . 3 and from about 0 . 6 to 0 . 85 . Heterogeneity in the form of a 4% difference in intrinsic periods did not severely disrupt synchrony , although synchrony in the Type I circuits in Fig . 8B was more robust than the Type II in Fig . 8C . This level of heterogeneity reduces but does not eliminate the bistable regions . The predicted time lags in Fig . 8B and Fig . 8C are in qualitative agreement with the summarized experimental results in Fig . 8A . The final possibility , a circuit with one cell with a Type I PRC and other with a Type II PRC , does not synchronize at any delay when the periods are matched ( predicted time lags indicated by X symbols in Fig . 8D ) , so we conclude that this type of circuit was not represented in the hybrid circuits that we constructed . Nonetheless , there was a region in which the time lag was fairly flat over a range of delays ( about 0 . 1 to 0 . 6 ) . If heterogeneity in the periods is introduced by slowing down the Type II neuron , this has the effect of matching the network periods because the Type II neuron is in general delayed less ( or advanced more ) than the Type I neuron at the same phase . This in turn pushes the circuit toward synchrony ( filled gray circles in Fig . 8D ) and is an alternate , but less likely explanation , of why near synchronization was universally observed , because every effort was made to match the periods . We conclude that the tendency to exhibit antiphase rather than synchrony at short delays is attributable to the initial region of negative slope in Type II inhibitory PRCs ( Fig . 2B2 ) and to the scarcity of Type I PRCs combined with the vulnerability of very early synchrony in circuits with Type I PRCs to heterogeneity . There is , however , another contributing factor that relies on the discontinuity consistently observed in PRCs measured for strong inhibition [12] , [23] ( see Fig . 2B1 and B2 ) . A strong inhibition applied immediately before a spike would have occurred in the absence of the inhibition ( at a phase just less than one ) consistently delays the next spike much more than one applied immediately after a spike ( at a phase just greater than zero ) . Consequently , in the absence of conduction delays , synchrony is unstable; if one neuron happens to fire just before the other neuron was going to spike , the second neuron is substantially delayed and synchrony is disrupted ( as illustrated in Fig . S2 ) . Therefore simply using the slope of the PRC to determine whether synchrony will be stable [16] , [23] , [25] is not sufficient if the PRC is discontinuous . The addition of a short conduction delay removes this discontinuity and stabilizes synchrony . In some cases the apparent discontinuity is due to resetting that is manifested in the second [26] rather than the first cycle after the perturbation , but that is not the case here . Overall , inhibitory coupling in these neurons favors synchronous activity at shorter normalized delays ( 0 . 1 to 0 . 7 ) than excitatory coupling ( >0 . 5 ) . Given that all but one of the inhibitory PRCs were Type II , the most likely circuit configuration for the hybrid circuits is to be comprised of two Type II neurons , so the solution structure illustrated in Fig . 8C should predominate . In the next section , we present evidence that the solution structure illustrated in Fig . 8C does indeed predominate . Fig . 9A shows data from a single representative hybrid circuit . A sharp transition from near antiphase to near synchrony is observed at a normalized delay of about 0 . 19 . This transition could occur if the circuit is comprised of two neurons with Type II PRCs as in Fig . 2B2 , as the delays are increased so the locking point for synchrony acquires a positive slope . Fig . 9B1 shows the voltage traces for each of the two neurons in the hybrid circuit firing in antiphase for a short delay corresponding to antiphase at a normalized delay of 0 . 09 . At a normalized delay of 0 . 19 delay ( Fig . 9B2 ) switching between near synchrony and antiphase is observed consistent with the prediction of bistability in Fig . 8C . At a normalized delay of 0 . 31 near synchronous activity is observed ( Fig . 9B3 ) . The right hand side of Fig . 10D shows the time lags ( black circles ) predicted for stable modes from the PRC exactly as in the homogeneous case for two cells with Type II PRCs as shown in Fig . 8C . However , here we keep track of the two time lags in the circuit separately as in the histograms in Fig . 7A and B . The black circles correspond to predicted stable modes . In this figure , we also show the predicted unstable modes ( red diamonds ) , because they form the boundaries between bistable modes . The solution branches ( black circles and red diamonds ) at ±0 . 6 on the y-axis correspond to the antiphase mode whereas the peaks at zero and near ±1 correspond to synchrony . For normalized delays between about 0 . 1 and 0 . 25 , synchrony is bistable with antiphase , but any time lags that fall between the red diamonds and the time lags for antiphase will converge to antiphase . At the beginning of the bistable regime this includes almost all time lags . As the delay is increased , the domain that converges to antiphase shrinks , and the one that converges to synchrony grows . As before , we used a noisy map based on the measured PRCs , in this case the Type II PRC shown in Fig . 2B2 , and the time lags produced by the noisy map are shown as gray circles . The left part of Fig . 10D shows the histograms produced by the noisy map for two identical neurons with Type II PRCs . At a normalized delay value of 0 . 09 ( line marked “A” in panel D ) only the two peaks associated with antiphase are observed . At a normalized delay value of 0 . 19 ( line marked “B” in panel D ) , there are five peaks: two that sample the antiphase mode and three that sample the synchronous mode . At a normalized delay value of 0 . 31 ( line marked “C” in panel D ) , only three peaks corresponding to synchrony remain . We compare the experimentally observed histograms ( Fig . 10A1 ) associated with the near antiphase mode in Fig . 9B1 , the bistable mode ( Fig . 10B1 ) observed in Fig . 9B2 and the near synchronous mode ( Fig . 10C1 ) observed in Fig . 9B3 with the corresponding histograms from Fig . 10D , with the slight difference that 4% heterogeneity in period was introduced ( see Figs . 10A2 , B2 and C2 ) . The heterogeneity was introduced in order to match the asymmetry in the experimentally observed histograms . The excellent correspondence between experiments and the noisy map based on the PRC is convincing evidence that the hybrid circuit exhibits bistability and that these circuits are well characterized using only the information in the PRCs under the assumption of pulsatile coupling .
Much previous work on coupled oscillators with delays ( see Discussion in [16] ) has relied upon specific models with a specific form of coupling [19] , [20] , [29]–[32] . An alternative PRC-based approach based on the assumption of weak coupling [33] , [34] implies that the weak coupling only slightly perturbs the intrinsic period of each oscillator , which is clearly not the case for near-causal-limit synchrony . A novel approach [9] did not presume a one to one locking between oscillators to explain gamma synchrony at a distance , but instead proposed a very specific alternate mechanism that is dependent on spike doublets that emerge as a consequence of delays and on both excitatory and inhibitory effects at both sites . Our approach reveals that the observed dynamics are very much dependent upon PRC shape and will of course vary depending upon the model and the coupling type . The neurons in the present study can be very successfully characterized as periodic oscillators if sufficient background excitation is provided . The mechanisms proposed herein for synchronization at a distance are predicated on pulsatile coupling and predictable from the PRC . Their applicability is subject to experimental verification in specific instances , but may be broadly applicable as described below . Although the superficial entorhinal cortex ( EC ) is a relatively well studied region of the mammalian brain , the dominant mode of communication among the EC stellate cells we study is controversial . Based on sharp-electrode recordings in brain slices , Dhillon and Jones [35] argued that EC stellate cells are not directly connected , which implies that they communicate with each other via inhibitory interneurons that introduce a polysynaptic delay . The putative effect of this delay was previously predicted to be synchronizing based upon a similar PRC-based approach [36] . More recently , Kumar and colleagues [37] used uncaging techniques to attempt to map connectivity within the superficial EC , and argued that EC stellate cells are connected directly , via excitatory synapses , with high probability . Our unpublished data collected using visual guidance and thus allowing recordings from EC stellate cells that lie very near each other , found no direct connectivity and are thus compatible with the results of Dhillon and Jones [35] . A previous study [12] showed that with no delay , mutual excitation produced synchrony whereas mutual inhibition gave rise to an antiphase mode . The present results show that these results are substantially altered by the presence of delays and support a model in which somewhat distant EC stellate cells , with polysynaptic communication delays of 5 ms or more , should synchronize best in the theta frequency band by driving inhibitory intermediaries and thus effectively inhibiting each other . Our results suggest that monosynaptic excitatory connections between stellate cells cannot support synchrony robustly , although they could support a nearly synchronous state with very small conduction delays . In the context of the larger cortico-hippocampal circuits in which these cells participate , the longest biologically plausible delays also arise via polysynaptic pathways . For example , a direct hippocampal-prefrontal pathway has a conduction velocity of 0 . 6 m/s for a conduction delay of 16 ms [38] , [39] . However , hippocampal activation by stimulation of the CA1 area elicited bursts in prefrontal cortex with a latency of 80 to 100 ms [40] implicating polysynaptic pathways in the delay . Resonant loops created by interconnected brain regions with accumulated transmission and activation delays on the order of 150 ms have been hypothesized to be important for the formation and retrieval of memories across cortico-hippocampal circuits [41] , [42] and could contribute to phase locking at theta frequencies . Zero phase lag synchronization in the presence of presumably symmetric inter-hemispheric delays was observed between pairs of multiunit responses from area 17 in the left and right hemispheres of cats with an intact corpus callosum . The locking was disrupted when the corpus callosum was severed [43] ) , indicating that mutual coupling was responsible for the phase locking . The locking was at gamma frequency ( 40–50 Hz ) , and the interhemispheric delays were on the order of 4–6 ms , or about a sixth to a third of a gamma cycle . Since the projection neurons from this region are excitatory , and the type of phase resetting curves expressed by the EC cells in this study would not support locking at those delay values for excitation , we predict that the relevant PRCs for interhemispheric communication between V17 areas have a significantly different shape that the ones observed in this study . In another example , also with presumably symmetric time delays , synchronization at gamma frequency was observed with a time lag of less than a millisecond between two sites separated by up to 4 mm in hippocampal slices as a result of tetani simultaneously applied at the two sites [44] . If the conduction velocity is as slow as 300 µm/ms [45] , the total delay between two hippocampal neurons 4 mm apart ( including a 1 ms synaptic delay ) could be 14 ms , more than half of a gamma cycle . Thus the PRCs similar to the ones observed in this study could produce such synchronization . Finally , synchronization was observed in a computational model [46] between gamma modules with similar frequencies in the presence of conduction delays up to 8 ms . For all of the cited examples , it is possible that synchronization may result from the mutual pulse coupling of oscillators; however , the oscillator may be a group of neurons rather than a single neuron . In order to apply the theoretical frame work used in this study to such cases , the relevant PRC becomes a property of the oscillatory unit rather than of an individual neuron . The conduction velocity in axons can be modulated [47] , leaving open the possibility of a self-regulatory mechanism that adjusts delays to compensate for heterogeneity and to induce synchronization under the appropriate conditions . These results can be generalized to larger networks in several ways . First , instead of reciprocal coupling between only two oscillators , these methods may generally apply to two coupled populations if the dynamics of the population can be approximated by those of a representative neuron [25] , [48] , [49] . For the second type of generalization , two ( or more neurons ) reciprocally coupled via a central hub neuron [50] , [51] , like the networks of neurons presented in this study with direct reciprocal connections , possess the symmetry required for synchronization at a distance , but the robustness of this architecture to heterogeneity and noise has not yet been characterized . Finally , we can generalize to large fully connected networks with delays . Stability of the in phase synchronous state for two neurons translates to stability of the fully synchronized large network state ( provided that the aggregate input received by each neuron is not too strong [25] ) . In the networks we studied , in the absence of delays , mutual excitation led to synchrony whereas mutual inhibition led to antiphase locking [12] . One of the most interesting aspects of our study is that the presence of delays that were a small fraction of the period inverted these results such that mutual inhibition favored synchrony whereas mutual excitation was desynchronizing .
|
Individual oscillators , such as pendulum-based clocks and fireflies , can spontaneously organize into a coherent , synchronized entity with a common frequency . Neurons can oscillate under some circumstances , and can synchronize their firing both within and across brain regions . Synchronized assemblies of neurons are thought to underlie cognitive functions such as recognition , recall , perception and attention . Pathological synchrony can lead to epilepsy , tremor and other dynamical diseases , and synchronization is altered in most mental disorders . Biological neurons synchronize despite conduction delays , heterogeneous circuit composition , and noise . In biological experiments , we built simple networks in which two living neurons could interact via a computer in real time . The computer precisely controlled the nature of the connectivity and the length of the communication delays . We characterized the synchronization tendencies of individual , isolated oscillators by measuring how much a single input delivered by the computer transiently shortened or lengthened the cycle period of the oscillation . We then used this information to correctly predict the strong dependence of the coordination pattern of the firing of the component neurons on the length of the communication delays . Upon this foundation , we can begin to build a theory of the basic principles of synchronization in more complex brain circuits .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"biology",
"computational",
"biology"
] |
2012
|
Short Conduction Delays Cause Inhibition Rather than Excitation to Favor Synchrony in Hybrid Neuronal Networks of the Entorhinal Cortex
|
Dog rabies annually causes 24 , 000–70 , 000 deaths globally . We built a spreadsheet tool , RabiesEcon , to aid public health officials to estimate the cost-effectiveness of dog rabies vaccination programs in East Africa . RabiesEcon uses a mathematical model of dog-dog and dog-human rabies transmission to estimate dog rabies cases averted , the cost per human rabies death averted and cost per year of life gained ( YLG ) due to dog vaccination programs ( US 2015 dollars ) . We used an East African human population of 1 million ( approximately 2/3 living in urban setting , 1/3 rural ) . We considered , using data from the literature , three vaccination options; no vaccination , annual vaccination of 50% of dogs and 20% of dogs vaccinated semi-annually . We assessed 2 transmission scenarios: low ( 1 . 2 dogs infected per infectious dog ) and high ( 1 . 7 dogs infected ) . We also examined the impact of annually vaccinating 70% of all dogs ( World Health Organization recommendation for dog rabies elimination ) . Without dog vaccination , over 10 years there would a total of be approximately 44 , 000–65 , 000 rabid dogs and 2 , 100–2 , 900 human deaths . Annually vaccinating 50% of dogs results in 10-year reductions of 97% and 75% in rabid dogs ( low and high transmissions scenarios , respectively ) , approximately 2 , 000–1 , 600 human deaths averted , and an undiscounted cost-effectiveness of $451-$385 per life saved . Semi-annual vaccination of 20% of dogs results in in 10-year reductions of 94% and 78% in rabid dogs , and approximately 2 , 000–1 , 900 human deaths averted , and cost $404-$305 per life saved . In the low transmission scenario , vaccinating either 50% or 70% of dogs eliminated dog rabies . Results were most sensitive to dog birth rate and the initial rate of dog-to-dog transmission ( Ro ) . Dog rabies vaccination programs can control , and potentially eliminate , dog rabies . The frequency and coverage of vaccination programs , along with the level of dog rabies transmission , can affect the cost-effectiveness of such programs . RabiesEcon can aid both the planning and assessment of dog rabies vaccination programs .
Rabies causes an estimated 25 , 000–70 , 000 human deaths annually , with about 90% of those deaths due to dog rabies [1–2] . Human rabies can be prevented through prompt post exposure prophylaxis ( PEP ) [3–5]; however , human rabies vaccine and immune globulin , needed for PEP , are frequently unavailable or unaffordable in developing countries with the highest burden of human rabies exposure [3 , 6] . Controlling dog rabies through large-scale dog vaccination programs effectively reduces human rabies mortality [7–10] . Previous studies have modeled dog rabies transmission and probabilities of human death after contact with a rabid animal [7 , 8 , 11–16] , as well as estimating the cost effectiveness of specific dog rabies control programs [7 , 17–21] . There are few tools available , however , that public health decision makers can readily use to estimate the impact and the cost-effectiveness of dog rabies control programs in their jurisdictions . We extend the existing literature by presenting an easy-to-use spreadsheet-based tool , called RabiesEcon , which public health officials can use to calculate the costs-and-benefits of dog rabies vaccination programs , including the number of averted rabid dogs and human rabies cases . We use RabiesEcon to estimate the impact and cost-effectiveness of dog rabies vaccination programs in a representative East African population of 1 million . Input values can be readily changed to represent almost any country or region , and thus RabiesEcon can provide public health officials with essential data for decision making related to controlling dog rabies .
RabiesEcon is a spreadsheet-based tool ( S1 Appendix ) that incorporates a mathematical ( deterministic ) model of dog-dog and dog-human rabies transmission to estimate dog and human rabies cases averted , and the cost per human rabies death averted and per year of life gained ( YLG ) due to dog rabies vaccination programs . We used RabiesEcon to estimate the cost-effectiveness of dog rabies vaccination programs in an illustrative East Africa human population of 1 million in a mixture of urban and rural settings . Because there are insufficient data from a single country in Africa for every input in RabiesEcon , we used data from a number of African countries , primarily Chad , Malawi and Tanzania ( Table 1 ) . We estimated , based on published measurements of dog ownership in East Africa [2 , 20] , that the modeled population has approximately 82 , 000 dogs ( 36 , 500 in urban setting , 45 , 700 in rural setting ) ( Table 1 ) . We chose East Africa as an example because recently published studies demonstrated the feasibility of conducting dog rabies vaccination programs in this region [20–23] . We built RabiesEcon to include a separate sub-model for each sub-region , urban and rural . Each sub-model calculates the number of dog rabies , human deaths and impact of dog vaccinations and PEP for that sub-region , using data relevant to the sub-regions ( Table 1 ) . The results from each sub-region are then summed and presented as a total for the entire area being studied . We compared three different dog rabies vaccination options: no vaccination , annual vaccination of 50% of all dogs , and semi-annual vaccination of 20% of dogs . We included , for each vaccination option , two dog rabies transmission scenarios: low ( 1 . 2 dogs infected per infectious dog ) and high ( 1 . 7 dogs infected per infectious dog ) ( see later for further details ) . We used several published sources of demographic , epidemiological , and economic data ( Table 1 ) . We used a government perspective ( government-as-payer ) . We assessed the impact of the interventions over a 10-year period , and we discounted all future costs and benefits ( including lives saved ) at a rates of 3% and 16% [24] . The later discount rate was derived from the weighted average yield to maturity for 10-year Bank of Tanzania Treasury bonds in October 2017 . ( https://www . bot . go . tz/financialmarkets/aspSmartUpload/TBondsResults . asp: accessed May 10 , 2018 ) . A user of RabiesEcon can alter almost all the input values . Our illustrative East African example includes urban and rural settings , using a population of approximately 1 million , with 2/3 of that population in an urban setting and 1/3 in a rural setting ( Table 1 ) . We set the total area occupied by this population at approximately 2 , 000 sq . km . , with approximately 200 sq . km . being urban ( Table 1 ) . These urban and rural settings allow for differences in human and dog population densities , and resultant differences in risk of rabies transmission ( Table 2 ) . We used , based on published studies , a rate of human to dog population of 18 . 1:1 for the urban areas and 7 . 4:1 for the rural areas [2 , 20 , 30] ( Table 1 ) . We used a previously published model [7] as a basis for our mathematical model of rabies transmission incorporated into RabiesEcon ( for equations , see S2 Appendix ) . We provide in Tables 1 and 2 , and S2 Appendix ( Table 1 ) , a list of inputs used in the transmission model . The model uses one-week time steps . The introduction of rabies into a previously uninfected dog population initially results in large oscillations in the estimated weekly number of rabid dogs . We therefore , to make it easier to facilitate comparisons between no vaccination and dog vaccination programs , programmed into RabiesEcon a process to calculate a “steady state” of a near-constant number of annual cases of canine rabies in a “no vaccination” scenario . We did this by programming RabiesEcon to run an initial 10 , 000 weeks ( S2 Appendix and Table 1 shows the specific parameters used ) . Because the risk of dog rabies transmission depends on a number of variables , such as the density of dogs and bites per rabid dog when attacking susceptible dog , we included in our analyses of each vaccination program two scenarios , low and high , of rates dog-to-dog rabies transmission [8 , 12 , 30] . We calculated the number of dogs infected per infectious dog as follows: Number of dogs infected per infectious dog ( Ro ) = Number of bites from infectious dog to susceptible dog x risk of infection per bite from infectious dog . Based on data from Tanzania , we used a range of 2 . 4–3 . 8 bites per infectious dog [8] . We then , to provide a range of Ro values from 1 . 1 to 1 . 7 ( Table 2 ) , assumed a value of 0 . 45 as the risk of infection per bite from infectious dog ( S2 Appendix . Table 1 ) . The range of values of Ro used closely follows the range reported by Hampson et al [8] , when they reviewed the literature of canine rabies transmission dynamics . The number of dogs infected by an infectious dog ( Ro value ) is likely impacted by factors such as dog density and percentage of dogs that are unconfined ( free roaming ) . The relationship between those and Ro is not well measured . Thus , any value chosen or calculated becomes a proxy for the impact of those other factors . We note that deterministic models , of the type used to build RabiesEcon , allow for the number of infectious dogs to be reduced to less than 1 ( e . g . , 0 . 5 infectious dog ) , but still able to transmit . This can result in “pop up” outbreaks of dog rabies in later years . We retained this factor for two reasons; It can be interpreted as mimicking , to a degree , the risk of importation of a rabid animal from outside , or the incomplete recording of all rabid dogs within , the dog rabies control area . And , users of RabiesEcon can easily ignore those “pop-up” outbreaks that occur in years well beyond the chosen analytic horizon ( e . g . , if the user runs a scenario in which dog rabies is eliminated by year 6 , “pop up” of cases in , say , year 16 can be assumed to be due to the mechanics of the model ) . As stated earlier , we compared a no vaccination option to two dog vaccination options ( annual vaccination of 50% of all dogs , and semi-annual vaccination of 20% of dogs ) ( Table 3 ) . The 50% annual coverage rate reflects , approximately , the average rate found by Jibat et al when they reviewed dog rabies vaccination coverage in Africa as reported in 16 published papers [31] . The 20% rate for semi-annual vaccination represents a potentially cheaper alternative ( i . e . , 10% less dogs are vaccinated ) . However , because the high turnover of dog populations ( due to a combination of short life expectancy and high dog birth rate–Table 1 ) , an annual vaccination program may result in up to 1/3 of vaccinated dogs dying in the interim between vaccinations programs . A smaller , but more frequent , semi-annual vaccination program may result in almost the same percentage of vaccinated dogs as with the annual program . These dog vaccination Options are illustrative , and can be readily changed by a user . We examine , in the sensitivity analysis , the impact of increasing the vaccination rate to the World Health Organization recommended level of 70% [2 , 3 , 12] . We assumed that dog rabies vaccine , when correctly administered , was 95% effective , similar to the effectiveness in humans [4] . Following Zinsstag et al . [7] , we included waning immunity in dogs vaccinated against rabies ( Table 3 ) . Because dog birth rate greatly influences dog-to-dog rabies transmission [7 , 29] , we included in the dog vaccination options concurrent dog population control programs , in which annually 7 . 5% of the intact male dogs were neutered ( Table 3 ) . We assumed that , for a user-defined percentage of male dogs neutered , there will be an equal percentage reduction in the number of dog litters , and thus a reduced dog population . We based this percentage on half the percentage of castrated male dogs observed in a survey of 150 dog-owning households in Machakos , Kenya [29] . We halved the percentage observed in Machakos because that was a relatively small survey , and our experience is that dog neutering programs in Africa are frequently under-resourced and thus do not impact large portions of the dog populations . We altered this assumption in our sensitivity analysis ( see later ) . We assumed , based on recent data from Haiti ( which faces rabies control resource constraints similar to many countries in Africa ) , that dogs with rabies symptoms would be immediately euthanized , and a small percentage ( 0 . 7% ) of the brains from those animals would be laboratory tested for rabies ( Table 3 ) . We further assumed that 5% of all dog-human bites would be investigated for potential rabies transmission [19] . Finally , we assumed that 21% of dog bite victims would start post-exposure prophylaxis ( PEP ) ( see later , Table 4 ) . We assumed a 95% efficacy when PEP is given as per recommended protocols , [4] . We altered in our sensitivity analyses the percentage of dog bite victims who receive PEP ( see later ) . We used , when modeling the dog vaccination strategies , the following three assumptions . Dog rabies is endemic ( i . e . , near steady state ) in the region being analyzed . Second , mass vaccination campaigns last 10 weeks , each year ( or 10 weeks twice per year if bi-annual ) . Third , the dog population can only increase to a maximum of 5% per year , which is near the lower limit measured by Kitala et al . in Machakos District , Kenya [28] . Kitala et al stated that the dog population in Machakos was growing at a rate faster than normally encountered in Africa . We calculated the cumulative 10-year totals of the number of rabid dogs , human rabies deaths and YLG with and without the rabies vaccination programs . We also estimated the 10-year total cost of each program . To calculate the cost-effectiveness over 10 years of each vaccination option per human death averted , we used the following formula: Costperhumandeathaverted=Costsofdogvaccinationprogram−costsincurredwithnovaccinationprogramNumberofhumandeathswithoutvacciantionprogram−humandeathswithvaccinationprogram For estimates of cost per case averted over more than 1 year ( e . g . , 10 years ) , each component of the formula was first summed , then the overall result calculated ( e . g . , for a 10 year cost of human death averted , the 10 year cost for dog vaccination program was summed separately , then added into the formula ) . When discounting was applied , each component was individually discounted to year 1 . We used a similar formula to calculate the cost per YLG , assuming that the average age of dog-rabies related death is 10 years of age [28] , and that life expectancy at age 10 is approximately 53 years [27] ( Table 1 ) ( Additional details in S2 Appendix , Note #2 ) . In addition to presenting all our results based on two different scenarios of low and high dog-to-dog rabies transmission ( Table 2 ) , we conducted the following sensitivity analyses . First , we examined the impact on estimates of rabid dogs in the high transmission scenario by changing the percentage of dogs neutered during the vaccination programs from 7 . 5% ( Table 3 ) to either 0% or 20% , assuming use of vaccination Option 1 ( 50% dogs vaccinated annually ) . Second , we calculated the number of rabid dogs if 0% , 20% , 50% , and 70% of the dog population were vaccinated annually , over a 30-year period . The 70% level is the World Health Organization ( WHO ) recommended minimum level of rabies vaccination needed to ensure dog rabies elimination [2 , 3 , 12] . We also considered the value of increasing PEP coverage from the base case of 21% ( Table 3 ) to 99% . Assuming that the effectiveness of PEP is 95% ( Table 4 ) , and that all those exposed comply with the full PEP regime , such a strategy would be designed to prevent almost all loss of human life to dog rabies , without the cost of large-scale dog rabies vaccination programs . Because such a strategy would have to continue without cessation due to the unceasing threat of rabid dogs , we calculated the results for both 10 years ( as for the other analyses in this paper ) , and for 30 years . Finally , we noted that the rate of onward dog-to-dog transmission is a crucial factor in estimating the spread of dog rabies and the consequent benefits of vaccinating dogs against rabies . We therefore conducted a multivariable analysis in which we made simulations changes in the following 4 variables that most directly impact the number of rabid dogs in our scenarios ( Table 1 ) . Annual percentage dogs vaccinated ( 30% , 40% , 50%—baseline 50% ) ; Dog birth rate ( 550 and 350/1 , 000 dogs–baseline 676/1 , 000 ) ; Dog life expectancy ( 3 . 0 and 2 . 5 years–baseline 3 . 0 years ) ; and , initial rate of dog-to-dog transmission , Ro ( 1 . 2 , 1 . 5 , 1 . 8 –baseline 1 . 2 ) . To simplify , when running this sensitivity analysis , we only used the values for the “urban” setting ( Table 1 ) ( i . e . , “turned off” rural settings ) . The range of annual percentage of dogs vaccinated was based on observations that these are the levels of coverage need to begin to observe “notable” reductions , but not guaranteed elimination , of human rabies deaths [1] . The estimate birth rate of 550/1 , 000 dogs was based on the lower 99% confidence interval from N’Djamena , Chad [7] . The lower estimate of 350/1 , 000 dogs came from birth rates for young dogs ( ≤ 12 months of age ) in rural Machakos District , Kenya [29] . The lower estimate of life expectancy is based on data from N’Djamena , Chad [7] . The Ro values examined are similar to those in Table 2 , which we derived from the review by Hampson et al . [8] .
Without a vaccination program , in the illustrative example there would be approximately 4 , 500 ( low rabies transmission ) to 6 , 500 ( high transmission ) rabid dogs per year , totaling approximately 44 , 000–65 , 000 rabid dogs over ten years ( Fig 1 and Table 6 ) . In the low rabies transmissions scenario , dog rabies vaccination options resulted in almost complete control of dog rabies within 5 years , with 10 years total reductions of approximately 42 , 600–41 , 200 rabid dogs , for dog vaccination Options 1 and 2 respectively ( Fig 1 and Table 6 ) . Such control remained for more than 10 years ( assuming the vaccination programs continued ) ( Fig 1 ) . In the high transmission scenario , the 10 year total reductions of rabid dogs were approximately 47 , 800–50 , 300 , for vaccination programs Options 1 and Options 2 , respectively ( Table 6 ) . Dog rabies cases begin to increase , for both options , at year 6 , and thereafter the number of cases fluctuates , albeit always lower than “no vaccination” option ( Fig 1 ) . Note that , in the high transmission scenario , vaccination Option 2 results in fewer rabid dogs , despite a lower total of dogs vaccinated ( Fig 1 , Table 6 ) . This is because , with the relatively high birth rate and short life spans of dogs in East Africa ( Table 1 ) , more frequent vaccination programs ( i . e . , twice per year ) protect a relatively larger portion of living dogs ( i . e . , dogs are vaccinated closer to the time of their birth , and thus here is a smaller pool of dogs susceptible to rabies ) . Human rabies deaths , without a vaccination program , total approximately 2 , 100–2 , 900 over 10 years ( Table 6 ) . The impact of vaccination programs on human deaths follows the same pattern as that for numbers of rabid dogs ( Fig 2 ) . The number of human deaths averted , under low rabies transmissions scenario , range from 2 , 100–2 , 000 for vaccination programs Options 1 and 2 , respectively . The deaths averted under the high transmission scenario range from approximately 1 , 600 ( Option 1 ) and 1 , 900 ( Option 2 ) deaths averted ( Table 6 ) . The reason why more deaths are averted in the high transmission scenario with vaccination option 2 , compared to vaccination Option 1 , is the same as the previously given explanation for the relatively lower number of dog rabies cases occurring under the same vaccination program ( Table 6 ) . The 10-year total program cost for dog vaccination Option 1 ( annual vaccination of 50% of the dog population ) was $1 . 4 million to $1 . 2 million , and Option 2 ( 20% of the dog population vaccinated ) cost $1 . 2 million to $1 . 2 million ( Table 6 ) . The no vaccination option would cost the government , over 10 years , approximately $0 . 4 million to $0 . 6 million . The undiscounted 10 year cost-effectiveness for Option 1 vaccination program ranged from $451-$385 per death averted ( low and high rabies transmission , respectively ) and $8-$6 per YLG ( Table 6 ) . The undiscounted cost-effectiveness for vaccination Option 2 were similar ( Table 6 ) . Reducing in the high transmission scenario the percent of dogs neutered , from 7 . 5% to 0% , during each vaccination program ( 50% dogs vaccinated , high rabies transmission scenario ) causes the rise in dog rabies cases to start 1 year earlier ( Fig 3 ) . Neutering 20% of the dogs delays by 3 years , compared to the 7 . 5% dogs neutered , any increase in dog rabies cases ( Fig 3 ) . Comparing the impact of percentage of dogs vaccinated over 30 years , in a low dog rabies transmission scenario , both 50% and 70% vaccination rates essentially eliminate dog rabies within 3 years , and maintain that rabies-free state for 30 years ( Fig 4 ) . This assumes no re-introduction of rabies from outside the area in which dog vaccination programs are initiated . In contrast , with high dog-to-dog disease transmission , 50% dogs vaccinated will result in outbreaks of dog rabies at year 6 , with cases occurring every year thereafter ( Fig 4 ) . An annual vaccination rate of 70% may result in an outbreak of rabies at approximately year 20 . The importance of the level of dog rabies transmission ( low versus high ) is consistent with previous findings [8 , 12 , 30] . Further , due to the linear relationships in dog-to-human transmission built into RabiesEcon ( S2 Appendix , Note #1 ) , as the number of rabid dogs decreases , the number of human deaths will also proportionately decrease . The impact of increasing PEP coverage from the base case of 21% to 99% is shown in Table 7 . In the no vaccination scenario , increasing effective coverage to 99% greatly reduces the number of human deaths , over a 30 year span , from approximately 7 , 900 to approximately 100 ( Table 7 ) . Such coverage , however , increases costs to approximately $6 . 9 million . Because 50% dog vaccination , in the low transmission scenario , will effectively eliminate dog rabies with 10 years ( Fig 1 ) , it is feasible to assume that the dog vaccination program will , if not entirely cease , be greatly reduced . The no vaccination at 99% PEP coverage , while greatly reducing number of deaths , has to continue indefinitely because to risk of human rabies does not reduce . Thus , it may be more relevant to compare the $6 . 9 million costs of 30 year no vaccination , 99% PEP costs to the $1 . 8 million costs of 10 year dog vaccination , 99% PEP program ( Table 7 ) . When we simultaneously changed the 4 variables that most impact the number of rabid dogs , we found that the most important variables were the Ro , the dog birth rate , and dog -life expectancy ( Fig 5 ) . Whenever dog birthrate was cut from the baseline value of 676/ 1 , 000 dogs to 350/ 1 , 000 dogs , any level of vaccination included in the analyses eliminated dog rabies ( Fig 5 ) . However , combining higher levels of dog birth rate and life expectancy ( 550 births/ 1 , 000 dogs and 3 . 0 years ) with higher levels of Ro ( 1 . 5 and 1 . 8 ) dog rabies may not be eliminated within 10 years ( Fig 5 ) . This suggests that dog rabies vaccination programs can benefit from any concurrent program that can effectively reduce dog birth rates . We note , however , that there are few examples from developing countries of such dog-population control programs being started and successfully maintained .
We estimate that vaccinating 20% ( semi-annually ) or 50% of an East African dog population will result in a cost-effectiveness of approximately $300–$450 per human death averted , and less than $10 per YLG . Our results were sensitive to the degree of dog-dog transmission ( Fig 1 and Fig 5 ) . For example , assuming that one infectious dog infects 1 . 2 other dogs allows our Option 2 ( 20% dogs vaccinated , semi-annually; low transmission scenario ) to essentially eliminate dog rabies in a 10 year period . But , if it is assumed that one infectious dog infects 1 . 7 other dogs ( +40% increase in risk of transmission; high transmission scenario ) , even vaccinating 50% of dogs annually is insufficient to eliminate dog rabies ( though there would still be fewer rabid dogs than the no vaccination option ) . In the high transmissions scenario , it requires 70% of dogs vaccinated to eliminate dog rabies for at least 20 years . Our results are similar to those of Bögel and Meslin , who found that dog vaccination , combined with administration of post-exposure prophylaxis to persons with a dog bite injury is more cost-effective than post-exposure prophylaxis alone [42] . Our estimates of the epidemiological impact of vaccinating 50% of the dog population are very similar to those of Coleman and Dye [12] . They used a mathematical model to estimate that dog rabies could be eliminated by vaccinating 39 to 57% of a dog population , with upper 95% confidence intervals of 55 and 71% , respectively [12] . They also estimated that achieving the WHO target of 70% of dogs vaccinated against rabies would give a 96 . 5% probability of preventing an outbreak . Zinsstag et al estimated that , in Chad , mass dog vaccination programs would result in a cost-effectiveness of $596 per human death averted in year 10 of a program ( applying a 5% discount rate ) [7] . Mindekem et al . , reporting on dog rabies vaccination program in Chad run in 2012 and 2013 , calculated a cost-effectiveness of $121 per Disability Life Year saved ( when death is almost the only outcome from a case of human rabies , Disability Life Year saved and YLG are almost equivalent ) [21] . It is noted that some have estimated higher Ro values than those we used ( Table 2 ) . Kitala et al estimated a higher value of 2 . 44 in Machokas District Kenya [16] . But , their 95% Confidence Interval of 1 . 52–3 . 36 spans the values that we used ( Tables 1 and 6 , Figs 1–5 ) . In a separate paper , they stated that their higher incidences of dog rabies are “… probably both a function of better case reporting… and a very high relative incidence of disease” [43] . Our model and estimates have some limitations . There is the previously mentioned mechanics of the mathematical model that allows for the number of infectious dogs to be reduced to less than 1 ( e . g . , 0 . 5 infectious dog ) , but still able to transmit . However , users of RabiesEcon can easily ignore those “pop-up” outbreaks that occur in years well beyond the chosen analytic horizon ( e . g . , Figs 1 and 5 ) . Another two important limitations are that the results can be , as demonstrated in the sensitivity analyses , greatly influenced by the values used to define the risk of dog-to-dog transmission ( e . g . , Fig 5 ) . In many instances , public health official using RabiesEcon may not have ready access to reliable estimates from their locale for all the inputs required . The other important limitation is that , as a deterministic model , RabiesEcon does not contain any built-in uncertainty . Thus , to correct for such imitations , users of RabiesEcon are greatly encouraged to conduct extensive sensitivity analyses , with a primary aim to determine which variables most likely influence the outcomes of interest , and at what point changes in modeled outcomes may change public health decisions . Other limitations derive from the fact that RabiesEcon calculates economic evaluations from the perspective of the government . Potential benefits accruing to others are not included . For example , Okell et al found that villagers in the Oromia region of Ethiopia considered rabies to be the zoonotic disease of greatest risk to both human and their livestock [44] . Jibat et al found that , in Ethiopia , rabies can cause a loss of 1–2 ( range: 1–5 ) head of cattle in affected herds , and the value of such losses ranges from $147 up to $1 , 140 , depending up the agricultural system ( mixed crop-livestock or pastoral ) [45] . In many parts of Africa , cattle are often sold at the end of their productive life . Their productive life includes being used for draft , which affects household income , labor and ultimately food security [46] . Thus , the value of cattle lost to rabies used by Jibat et al may be conservatively low . Public health decision makers , when using the results from RabiesEcon , will likely want to also consider including the value of such other benefits , even if they do not directly impact government budgets . Programs designed to notably reduce , even eliminate , human dog rabies deaths have to rely on the expansion of dog rabies vaccination coverage . Human PEP does save lives , but it can be relatively expensive and it is difficult to ensure that all persons potentially exposed to dog rabies have timely access to PEP [6 , 34 , 41 , 47 , 48] . It may well be difficult to implement-and-maintain PEP programs over several years that achieve 99% coverage ( Table 7 ) . Expansion of dog rabies vaccination programs require local , political , and economic support [9 , 49 , 50] . Anyiam et al have proposed a novel method to fund the required expansion of dog rabies vaccination programs [51] . They suggest that the government sell “development impact bonds” to private investors for the initial expansion . Assuming that the expanded vaccination program produces the anticipated results , then more traditional funding sources , such as the World Bank , African Development Bank , donor organizations , and the government can repay the bonds and continue funding the additional years of vaccination program . In this manner , banks , donors and the government only fund the program once a positive impact ( i . e . , success ) has been demonstrated . It will require negotiations as to the premium needed by investors to accept the initial risk . To attract investors to such a funding scheme will require estimates of disease burden without intervention , costs of intervention , and impact of intervention . RabiesEcon can be used to provide such estimates . Equally important to ensuring the success of the any dog rabies vaccination program is community involvement . The price of dog rabies vaccination to dog owners can notably reduce the willingness and/ or ability of dog owners to pay for dog vaccinations [31 , 52] . Dog owners also have to understand the need to maintain the vaccination status of their dogs–reduction in cases of rabid dogs and human rabies deaths may lead to complacency , and thus increased risk of either an outbreak or a re-introduction of rabies ( as modeled in Figs 1B , 4B and 5 ) . As dog rabies vaccination programs expand , and more dogs are vaccinated , there are other factors , beyond the current scope of RabiesEcon , which will need to be considered . These factors include the need for increased surveillance as cases of dog rabies decline . Such increased surveillance is needed to rapidly respond to any outbreak , or re-introduction , of dog rabies . It is possible that community involvement in such enhanced surveillance will be needed to ensure that such surveillance is successful [19 , 40 , 43] . Further , as cases of dog rabies decrease , there will likely be a financial benefit to health care payers ( e . g . , government agencies ) from improving the quality of screening human dog bite victims to receive PEP [34] . The goal of such screening would be to reduce the number of “false positives” ( i . e . , those who aren’t infected with rabies , but still receive PEP ) , whilst ensuring that there is no increase in the number of “false negatives” ( i . e . , those who are infected with rabies , but do not receive PEP ) . In conclusion , as demonstrated by the example and results presented here , RabiesEcon can help translate the complex set of factors affecting dog rabies transmission and human deaths due to dog rabies into readily understood estimates of impact-of-vaccination and cost-effectiveness . RabiesEcon is sufficiently flexible that a user can enter the relevant data ( Tables 1–5 ) from almost any country or locale , and thus estimate in costs-and-benefits of a dog rabies control program almost anywhere in the world . Such data may aid the expansion of dog rabies vaccination programs , and thus potentially aid the eventual elimination of dog rabies .
|
Dog rabies causes , globally , approximately 55 , 000 human deaths per year . Mass vaccination programs can control dog rabies . We built a spreadsheet-based tool , RabiesEcon , to aid public health officials in planning large-scale dog rabies vaccination programs . We used RabiesEcon to estimate the cost-effectiveness of dog rabies control programs in East Africa for a human population of one million ( approximately 2/3 urban , 1/3 rural ) . We evaluated three different vaccination options: no vaccination , annual vaccination of 50% of dogs , and semi-annual vaccination of 20% of dogs . Over a 10-year period , no dog vaccination results in approximately 44 , 000–65 , 000 rabid dogs and 2 , 000 human deaths . Annually vaccinating 50% of dogs for 10 years resulted in approximately 42 , 000–48 , 000 fewer rabid dogs and approximately 2 , 000–1 , 600 fewer human deaths . These reductions cost approximately $450-$385 per life saved . Semi-annual vaccination of 20% of dogs for 10 years resulted in approximately 41 , 000–50 , 000 fewer rabid dogs and approximately 2 , 000–1 , 900 fewer human deaths . These reductions cost approximately $400–$300 per life . In certain scenarios , 70% of dogs vaccinated eliminated dog rabies . Dog rabies vaccination programs can control , and potentially eliminate , dog rabies .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"cost-effectiveness",
"analysis",
"economic",
"analysis",
"post-exposure",
"prophylaxis",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"social",
"sciences",
"dogs",
"animals",
"mammals",
"vaccines",
"preventive",
"medicine",
"rabies",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"africa",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"zoonoses",
"economics",
"people",
"and",
"places",
"population",
"metrics",
"birth",
"rates",
"eukaryota",
"prophylaxis",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"amniotes",
"organisms"
] |
2018
|
Cost-effectiveness of dog rabies vaccination programs in East Africa
|
Network models are routinely downscaled compared to nature in terms of numbers of nodes or edges because of a lack of computational resources , often without explicit mention of the limitations this entails . While reliable methods have long existed to adjust parameters such that the first-order statistics of network dynamics are conserved , here we show that limitations already arise if also second-order statistics are to be maintained . The temporal structure of pairwise averaged correlations in the activity of recurrent networks is determined by the effective population-level connectivity . We first show that in general the converse is also true and explicitly mention degenerate cases when this one-to-one relationship does not hold . The one-to-one correspondence between effective connectivity and the temporal structure of pairwise averaged correlations implies that network scalings should preserve the effective connectivity if pairwise averaged correlations are to be held constant . Changes in effective connectivity can even push a network from a linearly stable to an unstable , oscillatory regime and vice versa . On this basis , we derive conditions for the preservation of both mean population-averaged activities and pairwise averaged correlations under a change in numbers of neurons or synapses in the asynchronous regime typical of cortical networks . We find that mean activities and correlation structure can be maintained by an appropriate scaling of the synaptic weights , but only over a range of numbers of synapses that is limited by the variance of external inputs to the network . Our results therefore show that the reducibility of asynchronous networks is fundamentally limited .
While many aspects of brain dynamics and function remain unexplored , the numbers of neurons and synapses in a given volume are well known , and as such constitute basic parameters that should be taken seriously . Despite rapid advances in neural network simulation technology and increased availability of computing resources [1] , memory and time constraints still lead to neuronal networks being routinely downscaled both on traditional architectures [2] and in systems dedicated to neural network simulation [3] . As synapses outnumber neurons by a factor of 103 − 105 , these constitute the main constraint on network size . Computational capacity ranges from a few tens of millions of synapses on laptop or desktop computers , or on dedicated hardware when fully exploited [4 , 5] , to 1012 − 1013 synapses on supercomputers [6] . This upper limit is still about two orders of magnitude below the full human brain , underlining the need for downscaling in computational modeling . In fact , any brain model that approximates a fraction of the recurrent connections as external inputs is in some sense downscaled: the missing interactions need to be absorbed into the network and input parameters in order to obtain the appropriate statistics . Unfortunately , the implications of such scaling are usually not investigated . The opposite type of scaling , taking the infinite size limit , is sometimes used in order to simplify equations describing the network ( Fig 1A ) . Although this can lead to valuable insights , real networks in the human brain often contain on the order of 105 − 107 neurons ( Fig 1B ) , too few to simplify certain equations in the limit of infinite size . This is illustrated in Fig 1C using as an example the intrinsic contribution to correlations due to fluctuations generated within the network , and the extrinsic contribution due to common external inputs to different neurons in random networks . Although the intrinsic contribution falls off more rapidly than the extrinsic one , it is the main contribution up to large network sizes ( around 108 for the given parameters ) . Therefore , taking the infinite size limit and neglecting the intrinsic contribution leads to the wrong conclusions: The small correlations in finite random networks cannot be explained by the network activity tracking the external drive [7] , but rather require the consideration of negative feedback [8] that suppresses intrinsically generated and externally imprinted fluctuations alike [9] . Taking the infinite size limit for analytical tractability and downscaling to make networks accessible by direct simulation are two separate problems . We concentrate in the remainder of this study on such downscaling , which is often performed not only in neuroscience [10 , 11 , 12 , 13] but also in other disciplines [14 , 15 , 16 , 17] . Neurons and synapses may either be subsampled or aggregated [18]; here we focus on the former . One intuitive way of scaling is to ensure that the statistics of particular quantities of interest in the downscaled network match those of a subsample of the same size from the full network ( Fig 1D ) . Alternatively , it may sometimes be useful to preserve the statistics of population sums of certain quantities , for instance population fluctuations . We here focus on the preservation of mean population-averaged activities and pairwise averaged correlations in the activity . We consider both the size and temporal structure of correlations , but not distributions of mean activities and correlations across the network . Means and correlations present themselves as natural quantities to consider , because they are the first- and second-order and as such the most basic measures of the dynamics . If it is already difficult to preserve these measures , it is even less likely that preserving higher-order statistics will be possible , in view of their higher dimensionality . However , other choices are possible , for instance maintaining total input instead of output spike rates [19] . Besides being the most basic dynamical characteristics , means and correlations of neural activity are biologically relevant . Mean firing rates are important in many theories of network function [20 , 21] , and their relevance is supported by experimental results [22 , 23] . For instance , neurons exhibit orientation tuning of spike rate in the visual system [24] and directional tuning in the motor system [25] , and sustained rates are implicated in the working memory function of the prefrontal cortex [22] . Firing rates have also been shown to be central to pattern learning and retrieval in highly connected recurrent neural networks [21] . Furthermore , mean firing rates distinguish between states of arousal and attention [26 , 27] , and between healthy and disease conditions [28] . The relevance of correlations is similarly supported by a large number of findings . They are widely present; multi-unit recordings have revealed correlated neuronal activity in various animals and behavioral conditions [29 , 30 , 31] . Pairwise correlations were even shown to capture the bulk of the structure in the spiking activity of retinal and cultured cortical neurons [32] . They are also related to information processing and behavior . Synchronous spiking ( corresponding to a narrow peak in the cross-correlogram ) has for example been shown to occur in relation to behaviorally relevant events [33 , 34 , 35] . The relevance of correlations for information processing is further established by the fact that they can increase or decrease the signal-to-noise ratio of population signals [36 , 37] . Moreover , correlations are important in networks with spike-timing-dependent plasticity , since they affect the average change in synaptic strengths [38] . Correspondingly , for larger correlations , stronger depression is needed for an equilibrium state with asynchronous firing and a unimodal weight distribution to exist in balanced random networks [39] . The level of correlations in neuronal activity has furthermore been shown to affect the spatial range of local field potentials ( LFPs ) effectively sampled by extracellular electrodes [40] . More generally , mesoscopic and macroscopic measures like the LFP and fMRI depend on interneuronal correlations [41] . Considering the wide range of dynamical and information processing properties affected by mean activities and correlations , it is important that they are accurately modeled . We allow the number of neurons N and the number of incoming synapses per neuron K ( the in-degree ) to be varied independently , generalizing the common type of scaling where the connection probability is held constant so that N and K change proportionally . It is well known that reducing the number of neurons in asynchronous networks increases correlation sizes in inverse proportion to the network size [19 , 42 , 43 , 44 , 45] . However , the influence of the number of synapses on the correlations , including their temporal structure , is less studied . When reducing the number of synapses , one may attempt to recover aspects of the network dynamics by adjusting parameters such as the synaptic weights J , the external drive , or neurotransmitter release probabilities [11 , 19] . In the present work , spike transmission is treated as perfectly reliable . We only adjust the synaptic weights and a combination of the neuronal threshold and the mean and variance of the external drive to make up for changes in N and K . A few suggestions have been made for adjusting synaptic weights to numbers of synapses . In the balanced random network model , the asynchronous irregular ( AI ) firing often observed in cortex is explained by a domination of inhibition which causes a mean membrane potential below spike threshold , and sufficiently large fluctuations that trigger spikes [46] . In order to achieve such an AI state for a large range of network sizes , one choice is to ensure that input fluctuations remain similar in size , and adjust the threshold or a DC drive to maintain the mean distance to threshold . As fluctuations are proportional to J2 K for independent inputs , this suggests the scaling J ∝ 1 K ( 1 ) proposed in [46] . Since the mean input to a neuron is proportional to J K , Eq ( 1 ) leads , all else being equal , to an increase of the population feedback with K , changing the correlation structure of the network , as illustrated in Fig 2 for a simple network of inhibitory leaky integrate-and-fire neurons ( note that in this example we fix the connection probability ) . This suggests the alternative [42 , 44 , 45] J ∝ 1 K , ( 2 ) where now the variance of the external drive needs to be adjusted to maintain the total input variance onto neurons in the network . For a given network size N and mean activity level , the size and temporal structure of pairwise averaged correlations are determined by the so-called effective connectivity , which quantifies the linear dependence of the activity of each target population on the activity of each source population . The effective connectivity is proportional to synaptic strength and the number of synapses a target neuron establishes with the source population , and additionally depends on the activity of the target neurons . Effective connectivity has previously been defined as “the experiment and time-dependent , simplest possible circuit diagram that would replicate the observed timing relationships between the recorded neurons” [47] . In our analysis we consider the stationary state , but at different times the network may be in a different state exhibiting a different effective connectivity . The definition of [47] highlights the fact that identical neural timing relationships can in principle occur in different physical circuits and vice versa . However , with a given model of interactions or coupling , the activity may allow a unique effective connectivity to be derived [48] . We define effective connectivity in a forward manner with knowledge of the physical connectivity as well as the form of interactions . We show in this study that with this model of interactions , and with independent external inputs , the activity indeed determines a unique effective connectivity , so that the forward and reverse definitions coincide . This complements the groundbreaking general insight of [47] . We consider networks of binary model neurons and networks of leaky integrate-and-fire ( LIF ) neurons with current-based synapses to investigate how and to what extent changes in network parameters can be used to preserve mean population-averaged activities and pairwise averaged correlations under reductions in the numbers of neurons and synapses . The parameters allowed to vary are the synaptic weights , neuronal thresholds , and the mean and variance of the external drive . We apply and extend the theory of correlations in randomly connected binary and LIF networks in the asynchronous regime developed in [7 , 8 , 9 , 42 , 45 , 49 , 50 , 51 , 52 , 53] , which explains the smallness and structure of correlations experimentally observed during spontaneous activity in cortex [54 , 55] , and we compare analytical predictions of correlations with results from simulations . The results are organized as follows . In “Correlations uniquely determine effective connectivity: a simple example” we provide an intuitive example that illustrates why the effective connectivity uniquely determines correlation structure . In “Correlations uniquely determine effective connectivity: the general case” we show that this one-to-one relationship generalizes to networks of several populations apart from degenerate cases . In “Correlation-preserving scaling” we conclude that , in general , only scalings that preserve the effective connectivity , such as J ∝ 1/K , are able to preserve correlations . In “Limit to in-degree scaling” we identify the limits of the resulting scaling procedure , demonstrating the restricted scalability of asynchronous networks . “Robustness of correlation-preserving scaling” shows that the scaling J ∝ 1/K can preserve correlations , within the identified restrictive bounds , for different networks either adhering to or deviating from the assumptions of the analytical theory . “Zero-lag correlations in binary network” investigates how to maintain the instantaneous correlations in a binary network , while “Symmetric two-population spiking network” considers the degenerate case of a connectivity with special symmetries , in which correlations may be maintained under network scaling without preserving the effective connectivity . Preliminary results have been published in abstract form [56] .
In this section we give an intuitive one-dimensional example to show that effective connectivity determines the shapes of the average pairwise cross-covariances and vice versa . For the following , we first introduce a few basic quantities . Consider a binary or spiking network consisting of several excitatory and inhibitory populations with potentially source- and target-type-dependent connectivity . For the spiking networks , we assume leaky integrate-and-fire ( LIF ) dynamics with exponential synaptic currents . The dynamics of the binary and LIF networks are respectively introduced in “Binary network dynamics” and “Spiking network dynamics” . We assume irregular network activity , approximated as Poissonian for the spiking network , with population means να . For the binary network , ν = ⟨n⟩ is the expectation value of the binary variable . For the spiking network , we absorb the membrane time constant into ν , defining ν = τm r where r is the firing rate of the population . The external drive can consist of both a DC component μα , ext and fluctuations with variance σ α , ext 2 , provided either by Poisson spikes or by a Gaussian current . The working points of each population , characterized by mean μα and variance σ α 2 of the combined input from within and outside the network , are given by μ α = ∑ β J α β K α β ν β + μ α , ext ( 3 ) σ α 2 = ∑ β J α β 2 K α β ϕ β + σ α , ext . 2 ( 4 ) with ϕ ≡ { ( 1 - ⟨ n ⟩ ) ⟨ n ⟩ for binary ν for LIF , ( 5 ) where Jαβ is the synaptic strength from population β to population α , and Kαβ is the number of synapses per target neuron ( the in-degree ) for the corresponding projection ( we use ≡ in the sense of “is defined as” ) . We call σ α , ext 2 “external variance” in the following , and the remainder “internal variance” . The mean population activities are determined by μα and σα according to Eqs ( 39 ) and ( 67 ) . Expressions for correlations in binary and LIF networks are given respectively in “First and second moments of activity in the binary network” and “First and second moments of activity in the spiking network” . As a one-dimensional example , consider a binary network with a single population and vanishing transmission delays . The effective connectivity W is just a scalar , and the population-averaged autocovariance a and cross-covariance c are functions of the time lag Δ . We define the population-averaged effective connectivity as W = w ( J , μ , σ ) K , ( 6 ) where w ( J , μ , σ ) is an effective synaptic weight that depends on the mean μ [Eq ( 3 ) ] and the variance σ2 [Eq ( 4 ) ] of the input . For LIF networks , w = ∂rtarget/∂rsource is defined via Eq ( 68 ) and can be obtained as the derivative of Eq ( 67 ) . Note that we treat the effective influence of individual inputs as independent . A more accurate definition of the population-level effective connectivity , beyond the scope of this paper , could be obtained by also considering combinations of inputs in the sense of a Volterra series [57] . When the dependence of w on J is linearized , the effective connectivity can be written as W = S ( μ , σ ) J K , ( 7 ) where the susceptibility S ( μ , σ ) measures to linear order the effect of a unit input to a neuron on its outgoing activity . In our one-dimensional example , W quantifies the self-influence of an activity fluctuation back onto the population . Expressed in these measures , the differential equation for the covariance function [Eq ( 52 ) ] takes the form τ 1 - W d d Δ c ( Δ ) = - c ( Δ ) + W 1 - W a ( Δ ) N , ( 8 ) with initial condition [from Eq ( 41 ) ] ( 1 - W ) c ( 0 ) = W a N , ( 9 ) which is solved by c ( Δ ) = a N ( 1 - W ) e W - 1 τ Δ - a N e - Δ τ . ( 10 ) Eq ( 10 ) shows that the effective connectivity W together with the time constant τ of the neuron ( which we assume fixed under scaling ) determines the temporal structure of the correlations . Furthermore , since a sum of exponentials cannot equal a sum of exponentials with a different set of exponents , the temporal structure of the correlations uniquely determines W . Hence we see that there is a one-to-one correspondence between W and the correlation structure if the time constant τ is fixed , which implies that preserving correlation structure under a reduction in the in-degrees K requires adjusting the effective synaptic weights w ( J , μ , σ ) such that the effective connectivity W is maintained . If , in addition , the mean activity ⟨n⟩ is kept constant this also fixes the variance a = ⟨n⟩ ( 1 − ⟨n⟩ ) . Eq ( 10 ) shows that , under these circumstances with W and a fixed , correlation sizes are determined by N . More generally , networks consist of several neural populations each with different dynamic properties and with population-dependent transmission delays dαβ . Since this setting does not introduce additional symmetries , intuitively the one-to-one relationship between the effective connectivity and the correlations should still hold . We here show that , under certain conditions , this is indeed the case . Instead of considering the covariance matrix in the time domain , for population-dependent dynamic properties we find it convenient to stay in the frequency domain . The influence of a fluctuating input on the output of the neuron can to lowest order be described by the transfer function H ( ω ) . This quantity measures the amplitude and phase of the modulation of the neuronal activity given that the neuron receives a small sinusoidal perturbation of frequency ω in its input . The transfer function depends on the mean μ [Eq ( 3 ) ] and the variance σ2 [Eq ( 4 ) ] of the input to the neuron . We here first consider LIF networks; in the Supporting Information we show how the results carry over to the binary model . In “First and second moments of activity in the spiking network” , we give the covariance matrix including the autocovariances in the frequency domain , C ‾ ( ω ) = C ( ω ) + A ( ω ) , as C ¯ ( ω ) = ( 𝟙 - M ( ω ) ) - 1 A ( 𝟙 - M T ( - ω ) ) - 1 , ( 11 ) where M has elements Hαβ ( ω ) Wαβ . If C ‾ ( ω ) is invertible , we can expand the inverse of Eq ( 11 ) to obtain C ¯ α β - 1 ( ω ) = ∑ γ ( 𝟙 α γ - M γ α ( - ω ) ) A γ - 1 ( 𝟙 γ β - M γ β ( ω ) ) = δ α β ( 1 - W α α e i ω d α α 1 - i ω τ α ) A α - 1 ( 1 - W α α e - i ω d α α 1 + i ω τ α ) + ( δ α β - 1 ) [ ( 1 - W α α e i ω d α α 1 - i ω τ α ) A α - 1 W α β e - i ω d α β 1 + i ω τ α + W β α e i ω d β α 1 - i ω τ β A β - 1 ( 1 - W β β e - i ω d β β 1 + i ω τ β ) ] + ∑ γ ≠ α , β W γ α e i ω d γ α 1 - i ω τ γ A γ - 1 W γ β e - i ω d γ β 1 + i ω τ γ , ( 12 ) where we assumed the transfer function to have the form H ( ω ) = e − i ω d α β 1 + i ω τ α , which is often a good approximation for the LIF model [45] . In the second step we distinguish terms that only contribute on the diagonal ( α = β ) , those that only contribute off the diagonal ( α ≠ β ) , and those that contribute in either case . For α = β , only the first and last terms contribute , and we get C ¯ α α - 1 = A α - 1 - W α α A α ( e - i ω d α α 1 + i ω τ α + e i ω d α α 1 - i ω τ α ) + ∑ γ W γ α 2 A γ - 1 1 + ω 2 τ γ 2 . ( 13 ) If we want to preserve C ‾ , this fixes Aα and thereby also Wαα , since it multiplies terms with unique ω-dependence . For α ≠ β , we obtain C ¯ α β - 1 = W α β A α e - i ω d α β ( - 1 1 + i ω τ α + W α α e i ω d α α 1 + ω 2 τ α 2 ) + W β α A β e i ω d β α ( - 1 1 - i ω τ β + W β β e - i ω d β β 1 + ω 2 τ β 2 ) + ∑ γ ≠ α , β W γ α W γ β A γ e i ω ( d γ α - d γ β ) 1 + ω 2 τ γ 2 . ( 14 ) With Aα fixed , this additionally fixes Wαβ , in view of the unique ω-dependence it multiplies . Since C ( ω ) = C ‾ ( ω ) − A , a constraint on A necessary for preserving C ‾ ( ω ) may not translate into the same constraint when we only require the cross-covariances C ( ω ) to be preserved . However , C ( ω ) and C ‾ ( ω ) have identical ω-dependence , as they differ only by constants on the diagonal ( approximating autocorrelations as delta functions in the time domain [45] ) . To derive conditions for preserving C ( ω ) , we therefore ignore the constraint on A but still require the ω-dependence to be unchanged . A potential transformation leaving the ω-dependent terms in both Eqs ( 13 ) and ( 14 ) unchanged is Aα → kAα , Wαβ → kWαβ , Wαα → kWαα , but this only works if τα = τγ , dαα − dαβ = dγα − dγβ for some γ , and if the terms for the corresponding γ are also transformed to offset the change in W α α W α β A α − 1; or if some of the entries of W vanish . The ω-dependence of C ‾ and C would otherwise change , showing that , at least in the absence of such symmetries in the delays or time constants , or zeros in the effective connectivity matrix ( i . e . , absent connections at the population level , or inactive populations ) , there is a one-to-one relationship between covariances and effective connectivity . Hence , preserving the covariances requires preserving A and W except in degenerate cases . Note that the autocovariances and hence the firing rates can be changed while affecting only the size but not the shape of the correlations , but that the correlation shapes determine W . Even in case of identical transfer functions across populations , including in particular equal transmission delays and identical τ , the one-to-one correspondence between effective connectivity and correlations can be demonstrated except for a narrower set of degenerate cases . The argument for d = 0 proceeds in the time domain along the same lines as “Correlations uniquely determine effective connectivity: a simple example” , using the fact that for a population-independent transfer function , the correlations can be expressed in terms of the eigenvalues and eigenvectors of the effective connectivity matrix ( cf . “First and second moments of activity in the binary network” and “First and second moments of activity in the spiking network” ) . For general delays , a derivation in the frequency domain can be used . Through these arguments , we show in the Supporting Information that the one-to-one correspondence holds at least if W is diagonalizable and has no eigenvalues that are zero or degenerate . If the working point ( μ , σ ) is maintained , the one-to-one correspondence between the effective connectivity and the correlations implies that requiring unchanged average covariances leaves no freedom for network scaling except for a possible trade-off between in-degrees and synaptic weights . In the linear approximation W ( J , μ , σ ) = S ( μ , σ ) JK , this trade-off is J ∝ 1/K . When this scaling is implemented naively without adjusting the external drive to recover the original working point , the covariances change , as illustrated in Fig 3B for a two-population binary network with parameters given in Table 1 . The results of J ∝ 1/K scaling with appropriate adjustment of the external drive are shown in Fig 3C . The scaling shown in Fig 3B also increases the mean activities ( E: from 0 . 16 to 0 . 23 , I: from 0 . 07 to 0 . 11 ) , whereas that in Fig 3C preserves them . If one relaxes the constraint on the working point while still requiring mean activities to be preserved , the network does have additional symmetries due to the fact that only some combination of μ and σ needs to be fixed , rather than each of these separately . This combination is more easily determined for binary than for LIF networks , for which the mean firing rates depend on μ and σ in a complex manner [cf . Eq ( 67 ) ] . When the derivative of the gain function is narrow ( e . g . , having zero width in the case of the Heaviside function used here ) compared to the input distribution , the mean activities of binary networks depend only on ( μ − θ ) /σ [9] . Changing σ while preserving ( μ − θ ) /σ leads for a Heaviside gain function to a new susceptibility S′ = ( σ/σ′ ) S [cf . Eq ( 43 ) ] . For constant K , if the standard deviation of the external drive is changed proportionally to the internal standard deviation , we have σ ∝ J and thus J′ S′ = JS , implying an insensitivity of the covariances to the synaptic weights J [52] . In particular , this symmetry applies in the absence of an external drive . When K is altered , this choice for adjusting the external drive causes the covariances to change . However , adjusting the external drive such that σ′/σ = ( J′ K′ ) / ( JK ) , the change in S is countered to preserve W and correlations . This is illustrated in Fig 3D for J ∝ 1 / K , which is another natural choice , as it preserves the internal variance if one ignores the typically small contribution of the correlations to the input variance ( [9] Fig 3D illustrates the smallness of this contribution for an example network ) . This is only one of a continuum of possible scalings preserving mean activities and covariances ( within the bounds described in the following section ) when the working point and hence the susceptibility are allowed to change . We now show that both the scaling J ∝ 1/K for LIF networks ( for which we do not consider changes to the working point , as analytic expressions for countering these changes are intractable ) , and correlation-preserving scalings for binary networks ( where we allow changes to the working point that preserve mean activities ) are applicable only up to a limit that depends on the external variance . For the binary network , assume a generic scaling K′ = κK , J′ = ιJ and a Heaviside gain function . We denote variances due to inputs from within the network and due to the external drive respectively by σ int 2 and σ ext 2 . The preservation of the mean activities implies S′ = ( σ/σ′ ) S as above , where σ 2 = σ ext 2 + σ int 2 . To keep SJK fixed we thus require σ int 2 ′ + σ ext 2 ′ = ( ι κ ) 2 ( σ int 2 + σ ext 2 ) σ ext 2 ′ = ι 2 κ [ ( κ - 1 ) σ int 2 + κ σ ext 2 ] , ( 15 ) where we have used σ int ′ ≈ ι κ σ int in the second line . For σext = 0 this scaling only works for κ > 1 , i . e . , increasing instead of decreasing the in-degrees . More generally , the limit to downscaling occurs when σ ext ′ = 0 , or κ = σ int 2 σ int 2 + σ ext 2 , ( 16 ) independent of the scaling of the synaptic weights . Thus , larger external and smaller internal variance before scaling allow a greater reduction in the number of synapses . The in-degrees of the example network of Fig 3 could be maximally reduced to 73% . Note that ι could in principle be chosen in a κ-dependent manner such that σ ext 2 is fixed or increased instead of decreased upon downscaling , namely ι ≥ σ ext 2 κ 2 σ ext 2 + κ ( κ − 1 ) σ int 2 . However , Eq ( 16 ) is still the limit beyond which this fails , as ι then diverges at that point . Note that the limit to the in-degree scaling also implies a limit on the reduction in the number of neurons for which the scaling equations derived here allow the correlation structure to be preserved , as a greater reduction of N compared to K increases the number of common inputs neurons receive and thereby the deviation from the assumptions of the diffusion approximation . This is shown by the thin curves in Fig 3C , 3D . Now consider correlation-preserving scaling of LIF networks . Reduced K with constant JK does not affect mean inputs [cf . Eq ( 3 ) ] but increases the internal variance according to Eq ( 4 ) . To maintain the working point ( μ , σ ) , it is therefore necessary to reduce the variance of the external drive . When the drive consists of excitatory Poisson input , one way of keeping the mean external drive constant while changing the variance is to add an inhibitory Poisson drive . With K′ = K/ι and J′ = ιJ , the change in internal variance is ( ι − 1 ) σ int 2 , where σ int 2 is the internal variance due to input currents in the full-scale model . This is canceled by an opposite change in σ ext 2 by choosing excitatory and inhibitory Poisson rates r e , ext = r e , 0 + ( 1 - ι ) σ int 2 τ m J ext 2 ( 1 + g ) , ( 17 ) r i , ext = ( 1 - ι ) σ int 2 τ m J ext 2 g ( 1 + g ) , ( 18 ) where re , 0 is the Poisson rate in the full-scale model , and the excitatory and inhibitory synapses have weights Jext and −g Jext , respectively . Eqs ( 17 ) and ( 18 ) match Eq ( E . 1 ) in [45] except for the 1 + g in the denominator , which was there erroneously given as 1+g2 . Since downscaling K implies ι > 1 , it is seen that the required rate of the inhibitory inputs is negative . Therefore , this method only allows upscaling . An alternative is to use a balanced Poisson drive with weights Jext and − Jext , choosing the rate of both excitatory and inhibitory inputs to generate the desired variance , and adding a DC drive Iext to recover the mean input , r e , ext = r i , ext = r e , 0 2 + ( 1 - ι ) σ int 2 2 τ m J ext 2 , ( 19 ) I ext = τ m r e , 0 J ext . ( 20 ) In this manner , the network can be downscaled up to the point where the variance of the external drive vanishes . Substituting this condition into Eq ( 15 ) , the same expression for the minimal in-degree scaling factor Eq ( 16 ) is obtained as for the binary network . In this section , we show that the scaling J ∝ 1/K , which maintains the population-level feedback quantified by the effective connectivity , can preserve correlations ( within the bounds given in “Limit to in-degree scaling” ) under fairly general conditions . To this end , we consider two types of networks: 1 . a multi-layer cortical microcircuit model with distributed in- and out-degrees and lognormally distributed synaptic strengths ( cf . “Network structure and notation” ) ; 2 . a two-population LIF network with different mean firing rates ( parameters in Table 2 ) . For both types of models , we contrast the scaling J ∝ 1/K with J ∝ 1 / K , in each case maintaining the working point given by Eqs ( 3 ) and ( 4 ) . Fig 4 illustrates that the former closely preserves average pairwise cross-covariances in the cortical microcircuit model , whereas the latter changes both their size and temporal structure . Fig 5 demonstrates the robustness of J ∝ 1/K scaling to the firing rate of the network . In this example , both the full-scale network and the downscaled networks receive a balanced Poisson drive producing the desired variance , while the mean input is provided by a DC drive . By changing the parameters of the external drive , we create two networks each with irregular spiking but with widely different mean rates ( 3 . 3 spikes/s and 29 . 6 spikes/s ) . Downscaling only the number of synapses but not the number of neurons , both the temporal structure and the size of the correlations are closely preserved . Reducing the in-degrees and the number of neurons N by the same factor , the correlations are scaled by 1/N . Hence , the correlations of the full-scale network of size N0 can be estimated simply by multiplying those of the reduced network by N/N0 . In contrast , J ∝ 1 / K changes correlation sizes even when N is held constant , and combined scaling of N and K can therefore not simply be compensated for by the factor N/N0 . In the high-rate network , the spiking statistics of the neurons is non-Poissonian , as seen from the gap in the autocorrelations ( insets in Fig 5B , 5D ) . Nevertheless , J ∝ 1/K preserves the correlations more closely than J ∝ 1 / K , showing that the predicted scaling properties hold beyond the strict domain of validity of the underlying theory . Although it is not generally possible to keep mean activities and correlations invariant upon downscaling , transformations may be found when only one aspect of the correlations is important , such as their zero-lag values . We illustrate this using a simple , randomly connected binary network of N excitatory and γN inhibitory binary neurons , where each neuron receives K = pN excitatory and γK inhibitory inputs . The parameters are given in Table 3 . The linearized effective connectivity matrix for this example is W = S ( μ , σ ) J K ( 1 - γ g 1 - γ g ) . ( 21 ) When the threshold θ is ≤ 0 , the network is spontaneously active without external inputs . In the diffusion approximation and assuming stationarity , the mean zero-lag cross-covariances between pairs of neurons from each population can be estimated from Eq ( 41 ) ( see also [52] ) [ ( 1 0 0 1 ) - W e 2 ( 2 - γ g - γ g 1 1 - 2 γ g ) ] ( c ee c ii ) = W e a N ( 1 - g ) c ei = c ie = 1 2 ( c ee + c ii ) , ( 22 ) where the subscripts e and i respectively denote excitatory and inhibitory populations . Moreover , We is the effective excitatory coupling , W e = S ( μ , σ ) J K , ( 23 ) with S the susceptibility as defined in Eq ( 43 ) . Furthermore , a is the variance of the single-neuron activity , a = ⟨ n ⟩ ( 1 - ⟨ n ⟩ ) , ( 24 ) which is identical for the excitatory and inhibitory populations . The mean input to each neuron is given by [cf . Eq ( 3 ) ] , μ = J K ( 1 - γ g ) ⟨ n ⟩ , ( 25 ) and , under the assumption of near-independence of the neurons , the variance of the inputs is well approximated by the sum of the variances from each sending neuron [cf . Eq ( 4 ) ] , σ 2 = J 2 K ( 1 + γ g 2 ) ⟨ n ⟩ ( 1 - ⟨ n ⟩ ) . ( 26 ) Finally , the mean activity can be obtained from the self-consistency relation Eq ( 39 ) . Eq ( 22 ) shows that , when excitatory and inhibitory synaptic weights are scaled equally , the covariances scale with 1/N as long as the network feedback is strong ( We ≫ 1 ) , ( for this argument , we assume that ⟨n⟩ is held constant , which may be achieved by adjusting a combination of θ and the external drive ) . Hence , conventional downscaling of population sizes tends to increase covariances . We use Eq ( 22 ) to perform a more sophisticated downscaling ( cf . Fig 6 ) . Let the new size of the excitatory population be N′ . Eq ( 22 ) shows that the covariances can only be preserved when a combination of We , γ , and g is adjusted . We take γ constant , and apply the transformation W e → f W e ; g → g ′ . ( 27 ) Solving Eq ( 22 ) for f and g′ yields ( cf . Fig 6B ) f = a c ee N ′ + γ c ii 2 ( c ee - c ii ) W e [ ( a N ′ + c ee ) ( a N + γ c ii ) - γ 4 ( c ee + c ii ) 2 ] ( 28 ) g ′ = c ee ( c ee - c ii ) - 2 a N ′ c ii γ c ii ( c ee - c ii ) + 2 a N ′ c ee . ( 29 ) The change in We can be captured by K → f K as long as the working point ( μ , σ ) is maintained . This intuitively corresponds to a redistribution of the synapses so that a fraction f comes from inside the network , and 1 − f from outside ( cf . Fig 6A ) . However , the external drive does not have the same mean and variance as the internal inputs , since it needs to make up for the change in g . The external input can be modeled as a Gaussian noise with parameters μ ext = K J ( 1 - γ g ) ⟨ n ⟩ - f K J ( 1 - γ g ′ ) ⟨ n ⟩ ( 30 ) σ ext 2 = K J 2 ( 1 + γ g 2 ) a - f K J 2 ( 1 + γ g ′ 2 ) a , ( 31 ) independent for each neuron . An alternative is to perform the downscaling in two steps: First change the relative inhibitory weights according to Eq ( 29 ) but keep the connection probability constant . The mean activity can be preserved by solving Eq ( 39 ) for θ , but the covariances are changed . The second step , which restores the original covariances , then amounts to redistributing the synapses so that a fraction f ˜ comes from inside the network , and 1 − f ˜ from outside , where the external ( non-modeled ) neurons have the same mean activity as those inside the network . This mean activity is negative , as the balanced regime implies stronger inhibition than excitation . Note that f ˜ ≠ f , since We changes already in the first step . The requirement that inhibition dominate excitation places a lower limit on the network size for which the scaling is effective . The reason is that g decreases with network size , so that a bifurcation occurs at g = 1/γ , beyond which the only steady states correspond to a silent network or a fully active one . We have seen that the one-to-one relationship between effective connectivity and correlations does not hold in certain degenerate cases . Here we consider such a degenerate case and perform a scaling that preserves mean activities as well as both the size and the temporal structure of the correlations under reductions in both the number of neurons and the number of synapses . The network consists of one excitatory and one inhibitory population of LIF neurons with a population-independent connection probability and vanishing transmission delays . Due to the appearance of the eigenvalues in the numerator of the expression for the correlations in LIF networks [cf . Eqs ( 70 ) and ( 71 ) ] , such networks are subject to a reduced number of constraints when W has a zero eigenvalue , as this leaves a freedom to change the corresponding eigenvectors . Furthermore , identically vanishing delays greatly simplify the equations for the covariances . The single-neuron and network parameters are as in Table 2 except that , here , N = 10 , 000 , J = 0 . 2 mV , and the external drive is chosen such that the mean and standard deviation of the total input to each neuron are μ = 15 mV , σ = 10 mV . Furthermore , the delay is chosen equal to the simulation time step to approximate d = 0 , which we assume here . The effective connectivity matrix for this network is W = w K ( 1 - γ g 1 - γ g ) , ( 32 ) where w = ∂rtarget/∂rsource is the effective excitatory synaptic weight obtained as the derivative of Eq ( 67 ) . Here , we take into account the dependence of w on J to quadratic order . The inhibitory weight is approximated as gw to allow an analytical expression for the relative inhibitory weight in the scaled network to be derived . The left and right eigenvectors are v 1 = 1 1 − γ g ( 1 − γ g ) , u 1 = 1 1 − γ g ( 1 1 ) corresponding to eigenvalue L = w K ( 1 − γ g ) and v 2 = 1 1 − 1 γ g ( 1 − 1 ) , u 2 = 1 1 − 1 γ g ( 1 1 γ g ) corresponding to eigenvalue 0 . The normalization is chosen such that the bi-orthogonality condition Eq ( 47 ) is fulfilled . A transformed connectivity matrix should have the same eigenvalues as W , and can thus be written as W ′ = w ′ K ′ ( 1 - b c - b c ) ( 33 ) where b = 1 c [ 1 - w K w ′ K ′ ( 1 - γ g ) ] . ( 34 ) Denote the new population sizes by N1 and N2 . Equating the covariances before and after the transformation yields using Eq ( 71 ) and Ajk = vjT A vk [cf . Eq ( 49 ) ] , a 1 N + γ g 2 a 2 N ( 1 - γ g ) 2 ( 2 - 2 L ) ( 1 1 1 1 ) + a 1 N + g a 2 N ( 2 - γ g - 1 γ g ) ( 2 - L ) ( 1 1 γ g 1 1 γ g ) = a 1 N 1 + b 2 a 2 N 2 ( 1 - b c ) 2 ( 2 - 2 L ) ( 1 c c c 2 ) + a 1 N 1 + b c a 2 N 2 ( 2 - b c - 1 b c ) ( 2 - L ) ( 1 1 b c c b ) . ( 35 ) In Eq ( 35 ) we have assumed that the working points , and thus a1 and a2 , are preserved , which may be achieved with an appropriate external drive as long as the corresponding variance remains positive . The four equations are simultaneously solved by N 1 = N w ′ K ′ a 1 ( 2 - L ) w K g a 2 ( w K - w ′ K ′ ) + a 1 [ 2 w ′ K ′ - w K ( w ′ K ′ - γ g w K ) ] N 2 = N a 2 w K L ( L - w ′ K ′ - 2 ) + 2 w ′ K ′ g a 2 ( 2 - L ) + ( w ′ K ′ - w K ) ( a 1 + g a 2 ) c = 1 , ( 36 ) where w′ K′ may be chosen freely . Thus , the new connectivity matrix reads W ′ = w ′ K ′ ( 1 w K w ′ K ′ ( 1 - γ g ) - 1 1 w K w ′ K ′ ( 1 - γ g ) - 1 ) , ( 37 ) which may also be cast into the form W ′ = w ′ K ′ ( 1 - γ ′ g ′ 1 - γ ′ g ′ ) , ( 38 ) where γ′ = N2/N1 and g ′ = w ′ K ′ − L w ′ K ′ γ ′ . When the populations receive statistically identical external inputs , we have a1 = a2 = r , since the internal inputs are also equal . Fig 7 illustrates the network scaling for the choice w′ = w . Results are shown as a function of the relative size N1/N of the excitatory population . External drive is provided at each network size to keep the mean and standard deviation of the total inputs to each neuron at the level indicated . The mean is supplied as a constant current input , while the variability is afforded by Poisson inputs according to Eqs ( 17 ) and ( 18 ) ( Fig 7D ) . It is seen that the transformations ( Fig 7B ) are able to reduce both the total numbers of neurons and the total number of synapses ( Fig 7C ) while approximately preserving covariance sizes and shapes ( Fig 7E , 7F ) . Small fluctuations in the theoretical predictions in Fig 7E are due to the discreteness of numbers of neurons and synapses , and deviations of the effective inhibitory weight from the linear approximation g w . The fact that the theoretical prediction in Fig 7F misses the small dips around t = 0 may be due to the approximation of the autocorrelations by delta functions , eliminating the relative refractoriness due to the reset . The numbers of neurons and synapses increase again below some N1/N , and diverge as g′ becomes zero . This limits the scalability despite the additional freedom provided by the symmetry .
By applying and extending the theory of correlations in asynchronous networks of binary and networks of leaky integrate-and-fire ( LIF ) neurons , our present work shows that the scalability of numbers of neurons and synapses is fundamentally limited if mean activities and pairwise averaged activity correlations are to be preserved . We analytically derive a limit on the reducibility of the number of incoming synapses per neuron , K ( the in-degree ) , which depends on the variance of the external drive , and which indirectly restricts the scalability of the number of neurons . Within these restrictive bounds , we propose a scaling of the synaptic strengths J and the external drive with K that can preserve mean activities and the size and temporal structure of pairwise averaged correlations . Mean activities can be approximately preserved by maintaining the mean and variance of the total input currents to the neurons , also referred to as the working point . The temporal structure of pairwise averaged correlations depends on the effective connectivity , a measure of the effective influence of source populations on target populations determined both by the physical connectivity and the working point of the target neurons . When the dependence of the effective connectivity on the synaptic strengths J is linearized , it can be written as SJK , where S is the susceptibility of the target neurons ( quantifying the change in output activity for a unit change in input ) . Scalings and analytical predictions of pairwise averaged correlations are tested using direct simulations of randomly connected networks of excitatory and inhibitory neurons . Our most important findings are: The population-level effective connectivity matrix and pairwise averaged correlations are linked by a one-to-one mapping except in degenerate cases . Therefore , with few exceptions , any network scaling that preserves the correlations needs to preserve the effective connectivity . The most straightforward way of simultaneously preserving mean activities and pairwise averaged correlations is to change the synaptic strengths in inverse proportion to the in-degrees ( J ∝ 1/K ) , and to adjust the variance of the external drive to make up for the change in variance of inputs from within the network . Other scalings , such as J ∝ 1 / K , can in principle also preserve both mean activities and pairwise averaged correlations , but then change the working point ( hence the neuronal susceptibility determining the strength of stimulus responses , and the degree to which the activity is mean- or fluctuation-driven ) , and are analytically intractable for LIF networks due to the complicated dependence of the firing rates and the impulse response on the mean and variance of the inputs . When downscaling the in-degrees K and scaling synaptic strengths as J ∝ 1/K , the variance of inputs from within the network increases , so that the variance of external inputs needs to be decreased to restore the working point . This is only possible up to the point where the variance of the external drive vanishes . The minimal in-degree scaling factor equals the ratio between the variance of inputs coming from within the network , and the total input variance due to both internal inputs and the external drive . The same limit to in-degree scaling holds more generally for scalings that simultaneously preserve mean activities and correlations . Thus , in the absence of a variable external drive , no downscaling is possible without changing mean activities , correlations , or both . Within the identified restrictive bounds , the scaling J ∝ 1/K , where the external variance is adjusted to maintain the working point , can preserve mean activities and pairwise averaged correlations also in asynchronous networks deviating from the assumptions of the analytical theory presented here . We show this robustness for an example network with distributed in- and out-degrees and distributed synaptic weights , and for a network with non-Poissonian spiking . For a sufficiently large change in in-degrees , a scaling that affects correlations can push the network from the linearly stable to an oscillatory regime or vice versa . Transformations derived using the diffusion approximation are able to closely preserve the relevant quantities ( mean activities , correlation shapes and sizes ) in simulated networks of binary and spiking neurons within the given bounds . Reducing the number of neurons only increases correlation magnitudes without affecting their structure in this approximation . However , strong deviations from the assumptions of the diffusion approximation can cause also correlation structure to change in simulated networks under scalings originally constructed to maintain correlation structure . This occurs for instance when a drastic reduction in network size is coupled with a less than proportional reduction in in-degrees , leading to large numbers of common inputs and increased synchrony . Thus , the scalability of the number of neurons with available analytical results is indirectly limited by the minimal in-degree scaling factor . In conclusion , we have identified limits to the reducibility of neural networks , even when only considering first- and second-order statistical properties . Networks are inevitably irreducible in some sense , in that downscaled networks are clearly not identical to their full-scale counterparts . However , mean activity , a first-order macroscopic quantity , can usually be preserved . The present work makes it clear that non-reducibility already sets in at the second-order macroscopic level of correlations . This does not imply a general minimal size for network models to be valid , merely that each network in question needs to be studied near its natural size to verify results from any scaled versions . Our analytical theory is based on the diffusion approximation , in which inputs are treated as Gaussian noise , valid in the asynchronous irregular regime when activities are sufficiently high and synaptic weights are small . Moreover , external inputs are taken to be independent across populations , and delays and time constants are assumed to be unchanged under scaling . A further assumption of the theory is that the dynamics is stationary and linearly stable . The one-to-one correspondence between effective connectivity and correlations applies with a few exceptions . For non-identical populations with different impulse responses , an analysis in the frequency domain demonstrates the equivalence under the assumption that the correlation matrix is invertible . An argument that assumes a diagonalizable effective connectivity matrix extends the equivalence to identical populations apart from cases where the effective connectivity matrix has eigenvalues that are zero or degenerate . The equivalence of correlations and effective connectivity ties in with efforts to infer structure from activity , not only in neuroscience [59 , 60 , 61 , 62 , 63 , 64 , 65 , 66] but also in other disciplines [67 , 68 , 69] , as it implies that one should in principle be able to find the only—and therefore the real—effective connectivity that accounts for the correlations . Within the same framework as that used here , [65] shows that knowledge of the cross-spectrum at two distinct frequencies allows a unique reconstruction of the effective connectivity matrix by splitting the covariance matrix into symmetric and antisymmetric parts . The derivation considers a class of transfer functions ( the Fourier transform of the neuronal impulse response ) rather than any specific form , but the transfer function is taken to be unique , whereas the present work allows for differences between populations . Furthermore , we here present a more straightforward derivation of the equivalence , not focused on the practical aim of network reconstruction , and clarify the conditions under which reconstruction is possible . In practice , using our results to infer structure from correlations may not be straightforward , due to both deviations from the assumptions of the theory and problems with measuring the relevant quantities . For instance , neural activity is often nonstationary [70] , transfer functions are normally not measured directly , and correlations are imperfectly known due to measurement noise . Furthermore , inference of anatomical from functional connectivity ( correlations ) is often done based on functional magnetic resonance imaging ( fMRI ) measurements , which are sensitive only to very low frequencies and therefore only allow the symmetric part of the effective connectivity to be reliably determined [66] . The presence of unobserved populations providing correlated input to two or more observed populations can also hinder inference of network structure . Thus , high-resolution measurements ( e . g . , two-photon microscopy combined with optogenetics to record activity in a cell-type-specific manner [71 , 72] ) of networks with controlled input ( e . g . , in brain slices ) hold the most promise for network reconstruction from correlations . The effects on correlation-based synaptic plasticity of scaling-related changes in correlations may be partly compensated for by adjusting the learning parameters . For instance , an increase in average correlation size with factor 1/N without a change in temporal shape may be to some extent countered by reducing the learning rate by the same factor . Changes in the temporal structure of the correlations are more difficult to compensate for . When learning is linear or slow , so that the learning function can be approximated as constant ( independent of the weights ) , the mean drift in the synaptic weights is determined by the integral of the product of the correlations and the learning function [73 , 74] . Therefore , this mean drift may be kept constant under a change in correlation shapes by adjusting the learning function such that this product is preserved for all time lags . However , given that the expression for the correlations is a complicated function of the network parameters , the required adjustment of the learning function will also be complex . Moreover , the effects of this adjustment on precise patterns of weights are difficult to predict , since the distribution of correlations between neuron pairs may change under the proposed scalings , and this solution does not apply when learning is fast and weight-dependent . The groundbreaking work of [46] identified a dynamic balance between excitation and inhibition as a mechanism for the asynchronous irregular activity in cortex , and showed that J ∝ 1 / K can robustly lead to a balanced state in the limit N → ∞ for constant K/N . However , it is not necessary to scale synaptic weights as 1 / K in order to obtain a balanced network state , even in the limit of infinite network size ( and infinite K ) . For instance , J ∝ 1/K can retain balance in the infinite size limit in the sense that the sum of the excitatory and inhibitory inputs is small compared to each of these inputs separately . To retain irregular activity with this scaling one merely needs to ensure a variable external drive , as the internal variance vanishes for N → ∞ . Moreover , in binary networks with neurons that have a Heaviside gain function ( a hard threshold ) identical across neurons , one does not even need a variable drive in order to stay in a balanced state [46 , p . 1360] . This can be seen from a simple example of a network of N excitatory and γN inhibitory neurons with random connectivity with probability p , where J = J0/N > 0 is the synaptic amplitude of an excitatory synapse , and −gJ the amplitude of an inhibitory synapse . The network may receive a DC drive , which we absorb into the threshold θ . The summed input to each cell is then μ = pNJ ( 1 − γg ) n , where n ∈ [0 , 1] is the mean activity in the network . For a balanced state to arise , the negative feedback must be sufficiently strong , so that the mean activity n settles on a level where the summed input is close to the threshold μ ≃ θ . This will always be achieved if pJ0 ( 1 − γg ) < θ < 0: in a completely activated network ( n = 1 ) the summed input is below threshold , and in a silent network ( n = 0 ) the summed input is above threshold , hence the activity will settle close to the value n ≃ θ/[pJ0 ( 1 − γg ) ] . As the variance of the synaptic input decreases with network size , the latter estimate of the mean activity will become exact in the limit N → ∞ . The underlying reason for both 1/K and 1 / K scaling to lead to a qualitatively identical balanced state is the absence of a characteristic scale on which to measure the synaptic input: the threshold is hard . Only by introducing a characteristic scale , for example distributed values for the thresholds , the 1/K scaling with a DC drive will in the large N limit lead to a freezing of the balanced state due to the vanishing variance of the summed input , while with either 1 / K scaling , or 1/K scaling with a fluctuating external drive , the balanced state is conserved . In [46] , J ∝ 1 / K refers not only to a comparison between differently-sized networks , but also to the assumption that approximately K excitatory synapses need to be active to reach spike threshold . However , this is also not a necessary condition for balance , which can arise for a wide range of synaptic strengths relative to threshold , as long as inhibition is sufficiently strong compared to excitation . As discussed in “Correlation-preserving scaling” , with appropriately chosen external drive , J even drops out of the mean-field theory for binary networks with a Heaviside gain function altogether [52] . The difficulty in the interpretation of the [46] results illustrates a more general point: The primary goal of scaling studies is to identify the mechanisms governing network dynamics . Nevertheless , these studies usually also specify requirements on the robustness of the mechanism , leading to scaling laws for network parameters that may be more restrictive than a description of the mechanism per se . An example is the robustness to strong synapses , defined such that activation of ∼ K excitatory synapses suffices to reach threshold in the absence of an external drive [46 , p . 1324] . This scenario was considered in order to create a condition under which dynamic balance is clearly necessary for achieving asynchronous irregular activity in balanced random networks , since combined inputs would otherwise far exceed the threshold . However , dynamic balance can arise also with weak synapses , e . g . , with strength ∼ 1/K of the distance to threshold . Without questioning the value of scaling studies , which can distill essential mechanisms and are sometimes possible where finite-size analytical descriptions are intractable , this shows that scaling laws need to be interpreted with care . The issue of the interrelation between network size , synaptic strengths , numbers of synapses per neuron , and activity is embedded in the wider context of anatomical and physiological scaling laws observed experimentally . In homeostatic synaptic plasticity , synaptic strengths are adjusted in a manner that keeps the activity of the postsynaptic neurons within a certain operating range [75 , 76 , 77] . Since postsynaptic activity depends not only on the strength of inputs but also on their number , this may induce a correlation between synaptic strengths and in-degree . In line with this hypothesis , excitatory postsynaptic currents ( EPSCs ) at single synapses were found to be inversely related to the density of active synapses onto cultured hippocampal neurons [78] , and the size of both miniature EPSCs and evoked EPSCs between neurons decreased with network size and with the number of synapses per neuron in patterned cultures [79] , although contrasting results have also been reported [80 , 81] . In the development of a mammal , the neuronal network grows by orders of magnitude and is continuously modified . For instance , the amplitude of miniature EPSCs is reduced in a period of heightened synaptogenesis in rat primary visual cortex [82] . During such developmental processes , some functions are conserved and new functions emerge . This balance between stability and flexibility is an intriguing theoretical problem . Here , network scaling is deeply related to biological principles . Our results open up a new perspective for analyzing and interpreting such biological scaling laws . Certainly , most network models will not fit neatly into the categories considered here , and detailed models often provide valuable insights regardless of whether they are scaled in a systematic manner . Nevertheless , it is usually possible to at least mention whether and how a particular model is scaled . When the results are not amenable to mathematical analysis , we suggest investigating through simulations of networks of different sizes how essential characteristics depend on numbers of neurons and synapses ( the relevant characteristics depend on the model at hand , and do not necessarily include mean activities or correlations ) . Thus , while both the investigation of the infinity limit and the exploration of downscaled networks remain powerful methods of computational neuroscience , we argue for a more careful approach to network scaling than has hitherto been customary , making the type of scaling and its consequences explicit . Fortunately , in neuroscience full-scale simulations are now becoming routinely possible due to the technological advances of recent years .
We verify analytical results for networks of binary neurons and networks of spiking neurons using direct simulations performed with NEST [83] revisions 10711 and 11264 for the spiking networks and revision 11540 for the binary networks . For simulating the multi-layer microcircuit model , PyNN version 0 . 7 . 6 ( revision 1312 ) [84] was used with NEST 2 . 6 . 0 as back end , single-threaded on 12 MPI processes on a high-performance cluster . All simulations have a time step of 0 . 1 ms . Spike times in the microcircuit model are constrained to the grid . The other spiking network simulations use precise spike timing [85] . In part , Sage was used for symbolic linear algebra [86] . Pre- and post-processing and numerical analysis were performed with Python . For both the binary and the spiking networks , we derive analytical results where both the number of populations Npop and the population-level connectivity are arbitrary . Specific examples are given of networks with a single , inhibitory population , or with two populations ( one excitatory , one inhibitory ) with either population-specific or population-independent connectivities . In addition , we discuss a multi-layer spiking cortical microcircuit model consisting of 77 , 169 neurons with approximately 3 × 108 synapses , with eight populations ( 2/3E , 2/3I , 4E , 4I , 5E , 5I , 6E , 6I ) and population-specific connection probabilities [58] , slightly adjusted to enhance the asynchrony of the activity . The adjustments consist of replacing normally by lognormally distributed weights with the same mean and with coefficient of variation 3; and using 4 . 5 instead of 4 as the relative strength of synapses from 4I to 4E compared to excitatory synaptic strengths . Besides distributed synaptic strengths , the model has binomially distributed in- and out-degrees , and normally distributed delays ( clipped at the simulation time step ) , thereby deviating from the assumptions of our analytic theory . It thus serves to evaluate the robustness of our analytical results to such deviations from the underlying assumptions . In all cases , pairs of populations are randomly connected . In the binary and one- and two-population LIF network simulations , in-degrees are fixed and multiple directed connections between pairs of neurons ( multapses ) are disallowed . In the multi-layer microcircuit model , in-degrees are distributed and multapses are allowed . In case of population-specific connectivities , we denote the ( unique or mean ) in-degree for connections from population β to population α by Kαβ , and synaptic strengths by Jαβ . Population sizes are denoted by Nα . For the example networks with population-independent connection probability , we denote the size of the excitatory population by N , the in-degree from excitatory neurons by K = pN , and the size of the inhibitory relative to the excitatory population by γ , so that the inhibitory in-degree is γK . Synaptic strengths are also taken to only depend on the source population , and are written as J for excitatory and −gJ for inhibitory synapses . We denote the activity of neuron j by nj ( t ) . The state nj ( t ) of a binary neuron is either 0 or 1 , where 1 indicates activity , 0 inactivity [7 , 42 , 87] . The state of the network of N such neurons is described by a binary vector n = ( n1 , … , nN ) ∈ {0 , 1}N . We denote the mean activity by ⟨nj ( t ) ⟩t , where the average ⟨⟩t is over time and realizations of the stochastic activity . The neuron model shows stochastic transitions ( at random points in time ) between the two states 0 and 1 . In each infinitesimal interval [t , t + δt ) , each neuron in the network has the probability 1 τ δ t to be chosen for update [88] , where τ is the time constant of the neuronal dynamics . We use an equivalent implementation in which the time points of update are drawn independently for all neurons . For a particular neuron , the sequence of update points has exponentially distributed intervals with mean duration τ , i . e . , update times form a Poisson process with rate τ−1 . The stochastic update constitutes a source of noise in the system . Given that the j-th neuron is selected for update , the probability to end in the up state ( nj = 1 ) is determined by the gain function Fj ( n ( t ) ) = Θ ( ∑k Jjk nk ( t ) − θ ) which in general depends on the activity n of all other neurons . Here θ denotes the threshold of the neuron and Θ ( x ) the Heaviside function . The probability of ending in the down state ( nj = 0 ) is 1 − Fj ( n ) . This model has been considered previously [42 , 87 , 89] , and here we follow the notation introduced in [87] that we also employed in our earlier works . We skip details of the derivation here that are already contained in [9] . The combined distribution of large numbers of independent inputs can be approximated as a Gaussian 𝓝 ( μ , σ2 ) by the central limit theorem . The arguments μ and σ are the mean and standard deviation of the synaptic input noise , together referred to as the working point [cf . Eqs ( 3 ) and ( 4 ) ] . The stationary mean activity of a given population of neurons then obeys [7 , 9 , 46 , 52] ⟨ n ⟩ = ⟨ F ( n ) ⟩ ≃ ∫ - ∞ ∞ Θ ( x - θ ) N ( μ , σ 2 , x ) d x = ∫ θ ∞ N ( μ , σ 2 , x ) d x = 1 2 erfc ( θ - μ ( ⟨ n ⟩ ) 2 σ ( ⟨ n ⟩ ) ) . ( 39 ) This equation needs to be solved self-consistently because ⟨n⟩ influences μ , σ through interactions within the population itself and with other populations . When network activity is stationary , the covariance of the activities of a pair ( j , k ) of neurons is defined as cjk ( Δ ) = ⟨δnj ( t + Δ ) δnk ( t ) ⟩t , where δnj ( t ) = nj ( t ) − ⟨nj ( t ) ⟩t is the deviation of neuron j’s activity from expectation , and Δ is a time lag . Instead of the raw correlation ⟨nj ( t + Δ ) nk ( t ) ⟩t , here and for the spiking networks we measure the covariance , i . e . , the second centralized moment , which is also identical to the second cumulant . To derive analytical expressions for the covariances in binary networks in the asynchronous regime , we follow the theory developed in [7 , 9 , 42 , 52 , 53] . We first consider the case of vanishing transmission delays d = 0 and then discuss networks with delays . Let c α β = 1 N α N β ∑ j ∈ α , k ∈ β , j ≠ k c j k ( 40 ) be the covariance averaged over disjoint pairs of neurons in two ( possibly identical ) populations α , β , and a α = 1 N α ∑ j ∈ α a j the population-averaged single-neuron variance aj ( Δ ) = ⟨δnj ( t + Δ ) δnj ( t ) ⟩t . Note that for α = β there are only Nα ( Nα − 1 ) disjoint pairs of neurons , so cαα differs from the average pairwise cross-correlation by a factor ( Nα − 1 ) /Nα , but we choose this definition because it slightly simplifies the population-level equations . For sufficiently weak synapses and sufficiently high firing rates , and when higher-order correlations can be neglected , a linearized equation relating these quantities can be derived for the case d = 0 ( [42] Eqs ( 9 . 14 ) – ( 9 . 16 ) ; [7] supplementary material Eq ( 36 ) , [9] Eq ( 10 ) ) , 2 c α β = ∑ γ ( W α γ c γ β + W β γ c γ α ) + W α β a β N β + W β α a α N α . ( 41 ) Here , we have assumed identical time constants across populations , and W α β = S ( μ α , σ α ) J α β K α β ( 42 ) is the linearized effective connectivity . The susceptibility S is defined as the slope of the gain function averaged over the noisy input to each neuron [9 , 52 , 53] , reducing for a Heaviside gain function to S ( μ , σ ) = 1 2 π σ e - ( μ - θ ) 2 2 σ 2 . ( 43 ) With the definitions c ¯ α β ≡ 1 N α N β ∑ j ∈ α , k ∈ β c j k = c α β + δ α β a α N α ( 44 ) P α β ≡ δ α β - W α β ( 45 ) Eq ( 41 ) is recognized as a continuous Lyapunov equation P c ¯ + ( P c ¯ ) T = 2 diag ( a α N α ) ≡ 2 A , ( 46 ) which can be solved using known methods . Let vj , uk be the left and right eigenvectors of W , with eigenvalues λj and λk , respectively . Choose the normalization such that the left and right eigenvectors are biorthogonal , v j T u k = δ j k . ( 47 ) Then multiplying Eq ( 46 ) from the left with vjT and from the right with vk yields ( 1 - λ j ) v j T c ¯ v k + v j T c ¯ v k ( 1 - λ k ) = 2 v j T A v k . ( 48 ) Define m j k ≡ v j T m v k , ( 49 ) for m = c , c ‾ , A . Then solving Eq ( 48 ) for c ‾ gives c ¯ = ∑ j , k 2 A j k 2 - λ j - λ k u j u k T , ( 50 ) as can be verified using Eq ( 47 ) . This provides an approximation of the population-averaged zero-lag correlations , including contributions from both auto- and cross-correlations . To determine the temporal structure of the population-averaged cross-correlations , we start from the single-neuron level , for which the correlations approximately obey ( [53] Eq ( 29 ) ) τ d d Δ c j k ( Δ ) + c j k ( Δ ) = ∑ i w j i c i k ( Δ ) , Δ ≥ 0 , ( 51 ) where wij is the neuron-level effective connectivity ( wij = Si Jij if a connection exists and wij = 0 otherwise ) . This equation also holds on the diagonal , j = k . To obtain the population-level equation , we use Eqs ( 40 ) and ( 44 ) and count the numbers of connections , which yields a factor Kαβ for each projection . Eq ( 51 ) then becomes τ d d Δ c ¯ ( Δ ) = - P c ¯ ( Δ ) , Δ ≥ 0 . ( 52 ) This step from the single-neuron to the population level constitutes an approximation when the out-degrees are distributed , but is exact for fixed out-degree [8 , 53] . The correlations for Δ < 0 are determined by c ‾ α β ( − Δ ) = c ‾ β α ( Δ ) . With the definition Eq ( 49 ) , Eq ( 52 ) yields τ d d Δ c ¯ j k ( Δ ) = ( λ j - 1 ) c ¯ j k ( Δ ) Δ ≥ 0 . ( 53 ) Using the initial condition for c ‾ from Eq ( 50 ) and multiplying Eq ( 53 ) by uj ukT , summing over j and k , we obtain the solution c ¯ ( Δ ≥ 0 ) = ∑ j , k 2 A j k 2 - λ j - λ k u j u k T e λ j - 1 τ Δ . ( 54 ) The shape of the autocovariances is well approximated by that for isolated neurons , A e − Δ τ , with corrections due to interactions being O ( 1/N ) [42] . Substituting this form in Eq ( 54 ) leads to c ( Δ ≥ 0 ) = ∑ j , k 2 A j k 2 - λ j - λ k u j u k T e λ j - 1 τ Δ - A e - Δ τ , ( 55 ) equivalent to [42] Eq ( 6 . 20 ) . Note that this equation still needs to be solved self-consistently , because the variance of the inputs to the neurons , which goes into S ( μ , σ ) , depends on the correlations . However , correlations tend to contribute only a small fraction of the input variance in the asynchronous regime ( cf . [9] Fig 3D ) . The accuracy of the result Eq ( 55 ) is illustrated in Fig 3A for a network with parameters given in Table 1 by comparison with a direct simulation . Note that the delays were not zero but equal to the simulation time step of 0 . 1 ms , sufficiently small for the correlations to be well approximated by Eq ( 55 ) . Now consider arbitrary transmission delay d > 0 , and let both d and the input statistics be population-independent . This case is most easily approached from the Fourier domain , where the population-averaged covariances including autocovariances can be approximated as [53] C ¯ ( ω ) = ( H ( ω ) - 1 - W ) - 1 2 τ A ( H ( - ω ) - 1 - W T ) - 1 . ( 56 ) Here , H ( ω ) is the transfer function H ( ω ) = e - i ω d 1 + i ω τ , ( 57 ) which is equal for all populations under the assumptions made . The transfer function is the Fourier transform of the impulse response , which is a jump followed by an exponential relaxation , h ( t ) = Θ ( t - d ) 1 τ e - t - d τ , ( 58 ) where Θ is the Heaviside step function . For the case of population-independent H ( ω ) , Fourier back transformation to the time domain is feasible , and was performed in [53] for symmetric connectivity matrices . Here , we consider generic connectivity ( insofar as consistent with equal H ( ω ) ) , and again use projection onto the eigenspaces of W to obtain a form similar to Eq ( 55 ) , i . e . , insert the identity matrix ∑ j u j v j T = 𝟙 ( 59 ) both on the left and on the right of Eq ( 56 ) , and Fourier transform to obtain 2 π c ¯ ( Δ ) = ∫ - ∞ + ∞ C ¯ ( ω ) e i ω Δ d ω = ∫ - ∞ + ∞ e i ω Δ ∑ j , k u j 1 H ( ω ) - 1 - λ j 2 τ v j T A v k 1 H ( - ω ) - 1 - λ k u k T d ω = 2 τ ∑ j , k u j u k T A j k ∫ - ∞ + ∞ f j k ( ω ) e i ω Δ d ω with f j k ( ω ) ≡ 1 H ( ω ) - 1 - λ j 1 H ( - ω ) - 1 - λ k . ( 60 ) In the third line of Eq ( 60 ) , we used Ajk = vjT A vk and collected the frequency-dependent terms for clarity . The exponential eiωΔ does not have any poles , so the only poles stem from fjk , which we denote by zl ( λj ) and the corresponding residues by Resj , k[zl ( λj ) ] . We only need to consider Δ ≥ 0 , since the solution for negative lags follows from c ‾ ( Δ ) = c ‾ T ( − Δ ) . The equation can then be solved by contour integration over the upper half of the complex plane , as the integrand vanishes at ω → +i∞ . Stability requires that the poles of the first term of Eq ( 60 ) lie only in the upper half plane ( note that the linear approximation we have employed only applies in the stable regime ) . The poles of the second term correspondingly lie in the lower half plane and hence need not be considered . For d > 0 , the locations of the poles are given by [53] Eq ( 12 ) , z l ( λ j ) = i τ - i d W l ( λ j d τ e d / τ ) , ( 61 ) where Wl is the lth of the infinitely many branches of the Lambert-W function defined by x = W ( x ) eW ( x ) [90] . For d = 0 , the poles are z ( λ j ) = − i τ ( λ j − 1 ) . Using the residue theorem thus brings Eq ( 60 ) into the form c ¯ ( Δ ≥ 0 ) = 2 τ i I ( γ ) ∑ j , k , l u j u k T A j k Res j , k [ z l ( λ j ) ] e i z l ( λ j ) Δ = ∑ j , k , l a j k l u j u k T e i z l ( λ j ) Δ , with a j k l ≡ 2 τ i I ( γ ) A j k Res j , k [ z l ( λ j ) ] , ( 62 ) where I ( γ ) = 1 is the winding number of the contour γ around the poles . To see that Eq ( 62 ) reduces to Eq ( 55 ) when d = 0 , substitute the poles in the upper half plane z ( λ j ) = − i τ ( λ j − 1 ) with residues [iτ ( 2 − λj − λk ) ]−1 and note that c ( Δ ) = c ‾ ( Δ ) − A ( Δ ) . When the input statistics and hence transfer functions are population-specific , Eq ( 56 ) becomes C ¯ ( ω ) = ( 𝟙 - M ( ω ) ) - 1 D ( ω ) ( 𝟙 - M T ( - ω ) ) - 1 , ( 63 ) D ( ω ) ≡ diag ( { 2 τ α a α N α ( 1 + ω 2 τ α 2 ) } α = 1 … N pop ) , ( 64 ) where Mαβ ( ω ) = Hαβ ( ω ) Wαβ . The spiking networks consist of single-compartment leaky integrate-and-fire neurons with exponential current-based synapses . The subthreshold dynamics of neuron i is given by τ m d V i d t = - V i + I i ( t ) , τ s d I i d t = - I i + τ m ∑ j J i j s j ( t - d ) , ( 65 ) where we have set the resting potential to zero without loss of generality , and absorbed the membrane resistance into the synaptic current Ii , in line with previous works [45 , 91] . Bringing back the corresponding parameters , the dynamics reads τ m d V ˜ i d t = - ( V ˜ i - E L ) + R m I ˜ i ( t ) , τ s d I ˜ i d t = - I ˜ i + τ s ∑ j J ˜ i j s j ( t - d ) . ( 66 ) Thus , our scaled synaptic amplitudes Jij in terms of the amplitudes J ˜ i j of the synaptic current due to a single spike are J i j = R m τ s / τ m J ˜ i j . Here , τm and τs are membrane and synaptic time constants , EL is the leak or resting potential , Rm is the membrane resistance , d is the transmission delay , I ˜ i = I i / R m is the total synaptic current , and s j = ∑ k δ ( t − t k j ) are the incoming spike trains . When Vi reaches a threshold θ , a spike is assumed , and the membrane potential is clamped to a level Vr for a refractory period τref . Threshold and reset potential in physical units are shifted by the leak potential ( θ = θ ˜ − E L , V r = V r ˜ − E L ) , showing that the assumption EL = 0 in Eq ( 65 ) does not limit generality . The intrinsic dynamics of the neurons in the different populations are taken to be identical , so that population differences are only expressed in the couplings . An approximation of the stationary mean firing rate of LIF networks with exponential current-based synapses was derived in [91] , r = ( τ m π ∫ V r - μ σ + α 2 τ s τ m θ - μ σ + α 2 τ s τ m Ψ ( s ) d s ) - 1 , Ψ ( s ) = e s 2 ( 1 + erf ( s ) ) , α = 2 | ζ ( 1 2 ) | , ( 67 ) where the summed synaptic input is characterized by a Gaussian noise with first moment μ and second moment σ2 based on the diffusion approximation , and ζ is the Riemann zeta function . For the covariances , we follow and extend the theory developed in [45 , 53] , starting with the average influence of a single synapse . Assuming that the network is in the asynchronous state , and that synaptic amplitudes are small , the synaptic influences can be averaged around the mean activity rj of each neuron j . These influences are characterized by linear response kernels hjk ( t , t′ ) defined as the derivative of the density of spikes of spike train sj ( t ) of neuron j with respect to an incoming spike train sk ( t′ ) , averaged over realizations of the remaining incoming spike trains s\sk that act as noise . In the stationary state , the kernel only depends on the time difference t − t′ , giving ⟨ s j ( t ) | s k ⟩ s \ s k = r j + ∫ - ∞ t h j k ( t - t ′ ) ( s k ( t ′ ) - r k ) d t ′ , h j k ( t - t ′ ) = ⟨ δ s j ( t ) δ s k ( t ′ ) ⟩ s \ s k ≡ w j k h ( t - t ′ ) , ( 68 ) where δsj ≡ sj − rj is the j-th centralized ( zero-mean ) spike train . Here , wjk is the integral of hjk ( t − t′ ) , and h ( t − t′ ) is a normalized function capturing its time dependence , which may be source- and target-specific . The dimensionless effective weights wjk are determined nonlinearly by the synaptic strengths Jjk , the single-neuron parameters , and the working point ( μj , σj ) ( cf . [45] Eq ( A . 3 ) but note that β as given there has a spurious factor J ) . We approximate the impulse response by the form Eq ( 58 ) , where τ is now an effective time constant depending on the working point ( μj , σj ) and the parameters of the target neurons . This form of the impulse response , corresponding to a low-pass filter , appears to be a good approximation in the noisy regime when the neuron fires irregularly . In the mean-driven regime ( μ ≫ σ ) the transfer function of the LIF neuron is known to exhibit resonant behavior with a peak close to its firing rate . In this regime a single exponential response kernel is expected to be a poor approximation ( see , e . g . , [92] Fig 1 ) . In general , the source population dependence of Eq ( 58 ) comes in through the delay d , and the target population dependence through both τ and d . As for binary networks with delays , the average pairwise covariance functions cij ( Δ ) ≡ ⟨δsi ( t + Δ ) δsj ( t ) ⟩t are most conveniently derived starting from the frequency domain . In case of identical transfer functions for all populations , the matrix of average cross-covariances is given by [53] Eq ( 16 ) minus the autocovariance contribution , C ( ω ) = ( H ( ω ) - 1 - W ) - 1 W A W T ( H ( - ω ) - 1 - W T ) - 1 + ( H ( ω ) - 1 - W ) - 1 WA + A W T ( H ( - ω ) - 1 - W T ) - 1 . ( 69 ) Here , W contains the effective weights of single synapses from population β to population α times the corresponding in-degrees , wαβ Kαβ; and A contains the population-averaged autocovariances , which we approximate as δ α β r α N α , with rα the mean firing rate , as also done in [45] . In [53] , Eq ( 69 ) was written using a more general diagonal matrix instead of A , to help clarify close similarities between binary and LIF networks and Ornstein-Uhlenbeck processes or linear rate models; however , for LIF networks , this diagonal matrix corresponds precisely to the autocovariance matrix . We chose the form Eq ( 69 ) because it separates terms that vanish at either ω → i∞ or ω → −i∞ depending on Δ . This facilitates Fourier back transformation , as contour integration with an appropriate contour can be used for each term . To perform the Fourier back transformation , we apply the same method as used for the binary network . Let vj , uj be the left and right eigenvectors of the connectivity matrix W , and λj the corresponding eigenvalues . Insert ∑j uj vjT = 𝟙 into Eq ( 69 ) on the left and right , and Fourier transform , 2 π c ( Δ ) =∫ - ∞ + ∞ C ( ω ) e i ω Δ d ω = ∫ - ∞ + ∞ e i ω Δ { ∑ j , k u j λ j H ( ω ) - 1 - λ j v j T A v k λ k H ( - ω ) - 1 - λ k u k T + ∑ j , k u j λ j H ( ω ) - 1 - λ j v j T A v k u k T + ∑ j , k u j v j T A v k λ k H ( - ω ) - 1 - λ k u k T } d ω . ( 70 ) As for the binary case , we only need to consider Δ ≥ 0 , as the solution for Δ < 0 is given by c ( Δ ) = cT ( −Δ ) . The contour can then be closed over the upper half plane , where the term containing only H ( −ω ) has no poles due to the stability condition . When Δ < d , the contour for the term containing only H ( ω ) can also be closed in the lower half plane where it has no poles , so that the corresponding integral vanishes . Analogously , the integral of the term with only H ( −ω ) vanishes when 0 > Δ > −d . Therefore , the second and third terms represent ‘echoes’ of spikes arriving after one transmission delay [53] . For Δ = 0 and d > 0 , only the first term contributes , and the contour can be closed in either half plane . As before , the poles are given by Eq ( 61 ) for d > 0 , and by z ( λ j ) = ∓ i τ ( λ j − 1 ) for d = 0 . The residue theorem yields a solution of the form Eq ( 62 ) , the only difference being the precise form of the residues , and the fact that we here consider c as opposed to c ‾ . In the absence of delays , an explicit solution can again be derived . For Δ > 0 , the poles inside the contour are z ( λ j ) = − i τ ( λ j − 1 ) corresponding to the terms with H ( ω ) −1 . The residue corresponding to λ j H ( ω ) − 1 − λ j is λ j i τ , and the term λ k H ( − ω ) − 1 − λ k is finite and evaluates at the pole to λ k 2 − λ j − λ k . Using Ajk = vjT A vk we get c ( Δ > 0 ) = ∑ j , k A j k τ λ j ( 2 - λ j ) 2 - λ j - λ k u j u k T e λ j - 1 τ Δ , ( 71 ) which is reminiscent of but not identical to Eq ( 55 ) for the binary network . Note that Eq ( 71 ) for the LIF network corresponds to spike train covariances with the dimensionality of 1/t2 due to [Ajk] = [1/t] and the factor 1/τ , whereas the covariances for the binary network are dimensionless . The population-specific generalization of Eq ( 69 ) reads C ( ω ) = ( 𝟙 - M ( ω ) ) - 1 M ( ω ) A M T ( - ω ) ( 𝟙 - M T ( - ω ) ) - 1 + ( 𝟙 - M ( ω ) ) - 1 M ( ω ) A + A M T ( - ω ) ( 𝟙 - M T ( - ω ) ) - 1 , ( 72 ) where M ( ω ) has elements Hαβ ( ω ) Kαβ wαβ , as before . The covariance matrix including autocovariances can be more simply written as C ¯ ( ω ) = ( 𝟙 - M ( ω ) ) - 1 A ( 𝟙 - M T ( - ω ) ) - 1 . ( 73 ) The only difference compared to the expression Eq ( 63 ) for the binary network is the form of the diagonal matrix , here analogous to white output noise in a linear rate model , whereas the binary network resembles a linear rate model with white noise on the input side , which is passed through the transfer function before affecting the correlations [53] . An alternative description of the spiking dynamics can be obtained by considering a system of linear coupled rate equations that produces the same moments to second order as the spiking dynamics [53] . The convolution equation y j ( t ) = ∑ k ∫ h j k ( t - t ′ ) y k ( t ′ ) d t ′ + x j ( t ) with ⟨ x j ( t ) x k ( s ) ⟩ = δ j k r j δ ( t - s ) , ( 74 ) with pairwise uncorrelated white noises xj and the response kernel hjk given by Eq ( 68 ) can be shown to yield a cross-covariance matrix of the form Eq ( 69 ) by considering the Fourier transform of Eq ( 74 ) , written in matrix notation as Y ( ω ) = H ( ω ) W Y ( ω ) + X ( ω ) . ( 75 ) We can expand the latter equation into eigenmodes by multiplying from the left with the left-sided eigenvector vk of W and by writing the general solution as a linear combination of right-sided eigenmodes Y ( ω ) = ∑j ηj ( ω ) uj to obtain ( with the bi-orthogonality relation vkT uj = δkj ) η k ( ω ) = H ( ω ) λ k η k ( ω ) + v k T X ( ω ) η k ( ω ) = 1 1 - λ k H ( ω ) v k T X ( ω ) . ( 76 ) The latter equation shows that the same poles z ( λk ) that appear in the covariance function Eq ( 70 ) also determine the evolution of the effective rate equation . Moreover , transforming Eq ( 76 ) back to the time domain , we see with η k ( t ) = i ∑ poles z ( λ k ) Res ( 1 1 - λ k H ( z ) , z ( λ k ) ) v k X ( z ) e i z ( λ k ) t that the eigenmodes have a time evolution determined by eiz ( λk ) t . Hence the imaginary part of the pole z ( λk ) controls whether the mode is exponentially growing ( Im ( z ) < 0 ) or decaying ( Im ( z ) > 0 ) , while the real part determines the oscillation frequency .
|
Neural networks have two basic components: their structural elements ( neurons and synapses ) , and the dynamics of these constituents . The so-called effective connectivity combines both components to yield a measure of the actual influence of physical connections . Previous work showed effective connectivity to determine correlations , which quantify the co-activation of different neurons . Conversely , methods for estimating network structure from correlations have been developed . We here extend the range of networks for which the mapping between effective connectivity and correlations can be shown to be one-to-one , and clarify the conditions under which this equivalence holds . These findings apply to a class of networks that is often used , with some variations , to model the activity of cerebral cortex . Since the numbers of neurons and synapses in real mammalian brains are vast , such models tend to be reduced in size for simulation purposes . However , our findings imply that if we wish to retain the original dynamics including correlations , effective connectivity needs to be unchanged , from which we derive scaling laws for synaptic strengths and external inputs , and fundamental limits on the reducibility of network size . The work points to the importance of considering networks with realistic numbers of neurons and synapses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Scalability of Asynchronous Networks Is Limited by One-to-One Mapping between Effective Connectivity and Correlations
|
Although the family of genes encoding for olfactory receptors was identified more than 15 years ago , the difficulty of functionally expressing these receptors in an heterologous system has , with only some exceptions , rendered the receptive range of given olfactory receptors largely unknown . Furthermore , even when successfully expressed , the task of probing such a receptor with thousands of odors/ligands remains daunting . Here we provide proof of concept for a solution to this problem . Using computational methods , we tune an electronic nose to the receptive range of an olfactory receptor . We then use this electronic nose to predict the receptors' response to other odorants . Our method can be used to identify the receptive range of olfactory receptors , and can also be applied to other questions involving receptor–ligand interactions in non-olfactory settings .
The mammalian sense of smell is able to detect molecules at levels as low as a few parts per trillion , as well as recognize and discriminate thousands of volatile molecules of diverse structure . Smelling begins when airborne odorants traverse the aqueous mucus layer covering the nasal epithelium , and bind to receptor proteins ( ORs ) within the membrane of cilia stemming from olfactory neurons . These receptors exhibit the characteristic structural features of the super family of G-protein–coupled receptors . In mammals , the repertoire of ORs is extremely large , consisting of over a thousand different subtypes [1–4] . It is currently held that many odorants are recognized by more than one receptor type , and most receptors recognize multiple odorants [5 , 6] . Olfactory neurons are found in abundance ( 10–100 million ) within the sensory surface , and it is thought that all sensory cells expressing the same receptor type project their axons onto two ( or more ) topographically fixed glomeruli in the olfactory bulb [7–10] . Therefore , the number of glomeruli is estimated to be between 1 , 000 and 2 , 000 , namely a reflection of the number of different receptor types . Thus , the receptive field of a glomerulus—which is defined as the stimulus range to which it responds—is equivalent to the molecular receptive range of the olfactory receptor expressed by its innervating neurons . As implied by the above-described organization of the peripheral olfactory system , the initial key to elucidating olfactory coding lies in elucidating the receptive range of given olfactory receptors . Indeed , this goal has been the subject of intensive research in laboratories and computation centres worldwide . Computational methods typically try to estimate the binding affinity based on the structures of the ligand and receptor [11–14] . While unequivocally promising , the complexity of the computations and the paucity of GPCR receptors with known structure put limits on the power of this approach . In turn , finding the receptive range for a given OR experimentally is difficult because ORs expressed in heterologous cells are typically retained in the endoplasmic reticulum [15] . Furthermore , even when successfully expressed [16–18] , the stimulus set size used is rather small compared to the large set of possible stimuli , and therefore capable of covering only a small portion of the receptive ranges . Here we set out to ask whether we could use artificial olfaction to predict ligand–receptor binding affinity , thereby suggesting a fast and cheap alternative to time-consuming computations or tedious experiments . This possibility has far-reaching consequences for drug design and might find applications in odor communication as well [19] . The most straightforward tools of artificial olfaction are electronic noses ( eNoses ) [20 , 21] . These are analytic devices that play a constantly growing role as general-purpose odor analyzers . The main component of an eNose is an array of non-specific chemical sensors . An analyte stimulates many of the sensors in the array , and elicits a characteristic response pattern . The sensors inside the eNose are made using diverse technologies , but in all cases a certain physical property is measured and a set of signals is generated . Electronic noses have been used intensively for many applications , mostly for classification tasks , with considerable success . For example , eNoses have been used in medical applications [22–25] , for environmental control [26] , for quality assessment of food products [27–31] , in the car manufacturing industry [32] , in predicting biological activity of alcohols [33] , and even in predicting human percept [34] . The results of these applications imply that a good eNose captures enough information on an odorant to allow its discrimination from other odorants [35] . A considerable part of this information is likely to overlap with the factors that determine the interaction strength with the biological receptor , such as relevant structural motifs , surface physical shape , and charge distribution [21] . Consequently , we hypothesized that eNose fingerprints capture information that is relevant to the strength of the biological interaction . In the first systematic study of OR receptive range , Araneda et al . [16] measured the response of the rat I7 OR to 90 pure chemicals , and divided the results into four groups according to the level of response to the odorant stimulus: high , medium , low , and no response . In a later and more comprehensive effort , Hallem et al . [36] measured the response of 24 Drosophila ORs to a set of 110 odorants . Here we set out to ask whether we could tune an eNose to these results such that the eNose could then predict the receptor response .
Encouraged by the prediction accuracy for the activity of the rat I7 receptor from only a small subset of chemicals , we set out to further test this approach on more OR data . A recent study measured the response of 24 Drosophila ORs to a set of 110 odorants [36] . We therefore took upon ourselves to measure these chemicals using the same eNose . We first measured 70 chemicals from [36] using the MOSESII eNose . Each sample was measured 3–4 times . Since in some cases the chemicals measured did not elicit a strong response , we modified the eNose measurement technique by increasing the temperature and the flow rate ( see Methods ) . This insured that almost all chemicals elicited a clear response . We than asked if we could predict the activity of each of the 24 receptors from our eNose signals . To find a statistically good learning rule that can predict the correct class of an unknown sample , the training set must contain a substantial number of examples from each class . Out of the 24 ORs measured , 12 responded to less than 15% of the odorants ( for example , OR 2a responded weekly to only three odorants ) . We therefore ignored these receptors and did not try to develop a learning rule for them . To find the learning rule for the remaining 12 Ors , we applied the perceptron learning scheme again . We grouped the odorants into two classes . The first contained all the odorants that elicited a weak response or no response ( less than 50 spikes , as defined in Hallem et al . ) , and the second class contained all odorants that had a medium or high response ( above or equal 50 ) . Thanks to the larger dataset , we could now use simpler feature extraction methods compared to those we used for the I7 receptor . Each sample was represented by a vector of size 120 . This vector contained four datapoints from each of the 16 signals ( see Methods ) , and 56 values that represent the 28 ratios between the maximum values of the eight signals from the metal-oxide ( MOX ) sensors , and 28 ratios of the eight signals from the quartz microbalance ( QMB ) sensors . Presenting a vector of size 120 to a neural network increases its computation time and might reduce the success rate . We thus applied PCA and took the eigenvectors that covered 99 . 9% of the variance . This reduced the size of the input vector by a factor of 2/3 . To test our learning rule , we applied a “leave 10 groups out” procedure . We performed each test 500 times . On each test , we randomly removed ten odorants and all their repeated measurements; this set was defined as the test set . The remaining odorants were used as a training set . We then trained a perceptron . The success rate for each receptor was defined as the average success rate of all the 500 runs . The results are depicted in Figure 3 . As can be seen , in almost all the receptors we tested the success rates were above 70% , with the best success rate being 86% for receptor ORs 7a , 59b , and 67c , and the lowest being 63% for receptor OR 35a . We tested the significance of the result using a χ2 test ( by comparing the expected 2-classes distribution of the null hypothesis to the observed 2-classes distribution of the learned rule . The null hypothesis is to always predict the class most abundant ) . We also tested the significance by using a 2-samples t-test to compare the mean success rate of the null hypothesis versus the learned rule . As can be seen in Figure 3 , in ten out of 12 receptors we tested , the predictive power was significant ( χ2 test , p < 0 . 01 ) . To further test the strength of this method , we applied a leave group out test of various group sizes . As seen Figure 4 , the prediction rates slightly diminish when larger group sizes were used . To further verify that the results cannot be attributed to chance , we tested the ability to predict a randomly created OR response . To do this , we randomly shuffled the response vector of each OR and then tested if we can learn the pseudo created OR response using the same algorithm . We repeated this randomization process 30 times for each receptor and calculated the average prediction success rate for the 12 randomly created receptors , which turned out to be 56% , a value not different from chance ( T ( 22 ) = −1 . 04 , p = 0 . 31 ) and significantly below our results with real OR responses ( T ( 22 ) = −6 . 05 , p = 0 . 0001 ) . The most convincing way to test such empirically developed rules is to use data that was not used in the rule building set . To this end , we set out to test an additional 21 odorants that were tested in [36] , but were not part of the 70 odorants used above . We measured each of these 21 odorants 3–4 times using our eNose , with the same parameters and outlier removal scheme as used above . This process generated 54 samples , representing 2–3 measurements of each of the 21 odorants . We then checked the prediction success rates of our previously learned rules on these new odorants . The results are depicted in Figure 5 . The average prediction rate was 77 . 07% , where the maximum was 88 . 8% . In 3/4 of the ORs , we obtained a success rate of above 70% , while in three ORs we obtained a success rate of ∼64% ( ORs 67a , 43b , and 35a ) . The prediction values of a few sample ORs of the list of 21 odorants is detailed in Table 1 . In other words , we were able to predict the response to odorants that were not used in our original model building set .
Predicting the response of an OR from the molecular features of an odorant ligand has met limited success . As the number of molecular features involved in the receptor-ligand interaction can be very high and the number of subsets and relations between these features grows very fast , the task of identifying the important features and relations is daunting . With this in mind , here we quantified the response to an odorant in synthetic sensors , in order to obtain an indirect measure of the forces and parameters involved . Our results suggest that an eNose can be used to predict , with reasonable accuracy , the response of an OR to an unknown odorant . This provides a link between artificial and biological olfaction . The fact that we could not duplicate the successful learning process for the pseudo-receptors lends further credence to this link . The observation that eNoses capture information that is relevant to biological interaction is promising in that they can provide a fast path for analysing the relationship between odorant and receptor , and can be extended to other kinds of biological interaction , outside the realm of olfaction . The prediction rates for 21 new samples using the 12 ORs tested ranged from 64% to 89% . One may ask what underlies this variability in classification power . Because we could not explain these differences from the receptor response characteristics or the learning scheme ( we tried several other learning methods , but in general the results were similar ) , we hypothesize that one source of the differences might be the noise in the measurements—of both the olfactory response and the eNose . Another likely source is that the ORs for which we could find only weak classifiers are tuned to molecular features that are not captured by the current eNose technology . This implies that an eNose with a greater number of sensor modules will provide more information and will improve the ability to extract more accurate rules . Finally , in this respect , it is noteworthy that even when the prediction rates are not high ( but significantly differ from the null hypothesis rule ) they can be further improved using boosting methods ( from M . Kearns , Thoughts on hypothesis boosting , unpublished manuscript , December 1988 ) , that turn a collection of weak classifiers into a single , highly accurate , classifier . Such boosting could also be achieved by using 2–3 independent eNoses . One may argue that our solution provides a sort of “black box”: we can tell which odorants may elicit a receptor response , but we can't say why . In other words , we have not directly generated insight regarding the molecular features dominating the response of a particular receptor . We would argue , however , that our results remain valuable even in the absence of mechanistic understanding: experimentally , such predictions may significantly contribute to the study of the olfactory system by providing a path to stimulus selection . If one wants to drive the olfactory system in order to probe its function systematically , one can use this method to judiciously select stimuli . Outside of the experimental context , this method may pave a unique path to ligand identification within a clinical framework . If one has in hand a particular target ligand , one can now use this method to identify additional potential ligands . Although one may indeed remain in the dark as to why these ligands are effective , one would nevertheless have potential ligands in hand . Finally , and critically in this respect , a key advantage in our approach is that the use of an eNose allows the application of the method to mixtures of odorants , which is not feasible for methods that quantify the response by recognizing the individual molecular features . All that said , this method may also set a path toward more mechanistic understanding as well . Such understanding may be achieved by in-depth examination of the particular sensors that dominate the response in one case or another , or by systematic examination of ligands that are grouped by this method , asking what molecular features are common to these ligands . Although such ligand groupings could be equally obtained by targeting odorants at expressed biological receptors , using an eNose for this task is orders-of-magnitude easier , faster , and cheaper .
The list of the 39 pure chemicals we used , and their interaction strength with the I7 rat OR , appear in Table 2 . Out of the 90 chemicals used in the I7 experiments , only 14 had a high or a medium response . We managed to measure 11 of them ( some of them could not be measured by the eNose ) . From the remaining low responding or non-responding 76 chemicals , we were able to obtain 28 . These numbers constitute a good representation of both response groups . We were unable to obtain the additional odorants because they were not identified by CAS , and were thus of ambiguous identity . We used 91 out of the 110 chemicals used in the Drosophila ORs experiments ( various technical issues such as low boiling temperature prevented us from obtaining or measuring the remaining 19 odorants ) . The list of the 91 pure chemicals used appears in Table 3 . The MOSES II eNose we used contains eight metal-oxide ( MOX ) sensors and eight quartz microbalance ( QMB ) sensors . MOX and QMB are two very different sensor technologies that together capture many facets of the ligand's nature . In the I7 experiments , the samples were put in 20-ml vials in an HP7694 headspace sampler , which heated them to 40 °C and injected the headspace content into the MOSES II with a flow rate of 25 ml/litter . There , each analyte was first introduced into the QMB chamber , whence it flowed through to the 300 °C heated MOX chamber . The injection lasted 30 s , and was followed by a 15 min purging stage using synthetic air . Each chemical was measured in batches , with a single batch containing ten successive measurements . In total , we performed 390 measurements . Each odorant was measured at the same level of humidity and temperature . Each single measurement consisted of 16 time-dependent signals , corresponding to the 16 sensors . In the Drosophila 24 ORs experiment , we heated the oven to 50 °C and increased the flow to 40 ml/litter . These parameter changes increased the number of chemicals that elicit a strong response . To avoid the problem of conditioning , we put a blank vial before every measurement and we cleaned the system using steamed air after each run . We measured each sample 3–4 times . In the first experiment , we used the Lorentzian method . Although this method outperforms simple methods commonly used in eNose applications , the Lorenzian method is , however , a lengthy process in which all abnormal signals need to be fixed , as described in [39 , 40] . Due to the large number of odorants used in the Drosophila experiment we decide to use a different feature extraction method that is similar to the Lorentzian but is easier to apply to all kind of signals . This method extracts four parameters from each signal . These parameters are: the signal max value and the time it reaches it , the time the signal reaches the half max value on the decay part and on the rise part . These four parameters are similar to the four parameters used in the Lorentzian model . In many cases the signal max value can change considerably between measurements of the same odorants; however , the relative height of the eight sensors is kept . Thus , to capture this behaviour we added to each odorant representation the 28 possible ratios of the eight MOX signals and 28 ratios of the eight QMB signals . We thus ended up with 120 features for each odorant . To ask whether this feature extractions method is a good representation of the odorants , we clustered the 273 eNose measurements we had into 70 classes and tested how many odorants fall into other classes . Out of the 273 measurements , 85% clustered either to their class or to the class closest to their class , suggesting that this feature extraction method is a good representation of the odorants . We used the Matlab [43] implementation of perceptron . We presented the training set to the perceptron and calculated the training error . Since perceptron is guaranty to converge only for linearly separable problems , we stopped the training if there was no improvement in three consecutive epochs , or when we reached 100 epochs .
|
A key goal in biology is to identify specific ligands for specific receptors . One example is where the ligand is a drug . In turn , in the olfactory system the ligand is the odorant that binds to olfactory receptors . There are many olfactory receptor types , and which odorants will activate which receptors remains largely unknown . One way to answer this is to systematically vary the molecular features of ligands and to measure the olfactory receptor response . However , the vast number of molecular features and their combinations renders such an effort potentially unsolvable . Here , rather than looking at the trees ( each molecular feature ) , we looked at the forest ( the smell they generate ) . We used a device called an electronic nose that generates a patterned response to odorants . We then obtained the response to a set of odorants that are known to activate a particular olfactory receptor , and we used this pattern to predict the response of that receptor to other odorants . We found that , on average in three out of four we could predict the response of olfactory receptors . This result provides a new method for probing the olfactory system , and also suggests a novel method for identifying potential drugs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"none",
"computational",
"biology"
] |
2008
|
Predicting the Receptive Range of Olfactory Receptors
|
Where human African trypanosomiasis ( HAT ) patients are seen , failure to microscopically diagnose infections by Trypanosoma brucei gambiense in blood smears and/or cerebrospinal fluid ( CSF ) in the critical early stages of the disease is the single most important factor in treatment failure , a result of delayed treatment onset or its absence . We hypothesized that the enhanced sensitivity of detergent-enhanced loop-mediated isothermal amplification ( LAMP ) will allow for point of care ( POC ) detection of African trypanosomes in the CSF of HAT patients where the probability for detecting a single parasite or parasite DNA molecule in 1 μL of CSF sample is negligible by current methods . We used LAMP targeting the multicopy pan-T . brucei repetitive insertion mobile element ( RIME LAMP ) and the Trypanosoma brucei gambiense 5 . 8S rRNA-internal transcribed spacer 2 gene ( TBG1 LAMP ) . We tested 1 μL out of 20 μL sham or Triton X-100 treated CSFs from 73 stage-1 and 77 stage-2 HAT patients from the Central African Republic and 100 CSF negative controls . Under sham conditions , parasite DNA was detected by RIME and TBG1 LAMP in 1 . 4% of the stage-1 and stage-2 gambiense HAT CSF samples tested . After sample incubation with detergent , the number of LAMP parasite positive stage-2 CSF’s increased to 26% , a value which included the 2 of the 4 CSF samples where trypanosomes were identified microscopically . Unexpected was the 41% increase in parasite positive stage-1 CSF’s detected by LAMP . Cohen’s kappa coefficients for RIME versus TBG1 LAMP of 0 . 92 ( 95%CI: 0 . 82–1 . 00 ) for stage-1 and 0 . 90 ( 95%CI: 0 . 80–1 . 00 ) for stage-2 reflected a high level of agreement between the data sets indicating that the results were not due to amplicon contamination , data confirmed in χ2 tests ( p<0 . 001 ) and Fisher’s exact probability test ( p = 4 . 7e-13 ) . This study detected genomic trypanosome DNA in the CSF independent of the HAT stage and may be consistent with early CNS entry and other scenarios that identify critical knowledge gaps for future studies . Detergent-enhanced LAMP could be applicable for non-invasive African trypanosome detection in human skin and saliva or as an epidemiologic tool for the determination of human ( or animal ) African trypanosome prevalence in areas where chronically low parasitemias are present .
In East Africa , the tsetse fly-transmitted protozoan parasite Trypanosoma brucei rhodesiense causes acute human African trypanosomiasis ( HAT/ sleeping sickness ) [1] . Transmitted from animals to man , T . b . rhodesiense infection is a zoonosis characterized by relatively high parasite loads . Over 97% of all HAT cases occur in West and Central Africa where the disease is caused by T . b . gambiense , which causes chronic disease with intermittent parasitemias characterized by low parasite numbers [2 , 3] . In 2016 , the number of patients with HAT reported by the World Health Organization ( WHO ) was fewer than 2 , 000; however , with many unreported cases , the estimate of actual number of infected people in the remaining endemic countries in Africa is probably higher [4] . History has shown that HAT reappears at epidemic cyclic intervals as parasites from chronically infected individuals without clinical signs of disease or harbored in animal reservoirs re-emerge back into the population [5–7] . The disease is marked by an early systemic hemolymphatic stage-1 phase , where the clinical symptoms and signs are easily confused with those of other infectious diseases ( i . e . malaria , viral syndromes ) . Left untreated , the parasites invade the central nervous system ( CNS; stage-2 ) , a process that usually takes weeks to months with T . b . rhodesiense or months to years for T . b . gambiense infections . Both parasites cause white matter encephalitis ( leukoencephalitis ) that belies neuropathologic manifestations that lead to death if untreated [reviewed in [8 , 9]] . Night time insomnia and day time drowsiness , which give the disease its name , are the most characteristic neurologic signs of gambiense HAT . However , the somnolence and other late stage mental signs are less common in Rhodesian disease , although there may be mental slowness or dullness and drowsiness or coma in terminal disease [10] . A key issue in the diagnosis and treatment of HAT is to distinguish reliably CNS involvement with HAT from the early stage disease . Accurate staging of HAT is critical because failure to treat a patient with CNS involvement using stage-2 drugs will lead inevitably to death from the disease , yet inappropriate CNS treatment in an early-stage patient carries a high risk of unnecessary drug toxicity and potentially death . The diagnosis and staging of HAT in the rural clinical setting where most patients are found , is time consuming , difficult and still relies largely on the microscopic detection of parasites in clinical samples ( blood smear , lymph , CSF ) . While this approach is inherently insensitive , it is still considered the unofficial gold-standard for specific diagnosis [11–13] . Where the disease is hyperendemic and since trypanosomes can be difficult to detect in CNS HAT , especially in the late stage [14] , failure by these methods to correctly classify stage-2 from stage-1 disease is probably the single most important contributor to disease progression and treatment failure [11] . While T . b . rhodesiense detection in blood is frequently successful , for T . b . gambiense infections , where only a few parasites are present in the peripheral circulation or in CSF , a thorough search is required , but is time consuming and subjective . Concentration techniques such as double centrifugation or mini-anion exchange columns ( mAECT ) are usually necessary [11–13] . Because of the inherent difficulties associated in detection of parasites in gambiense HAT patient CSF samples , examination of white cell count/protein concentrations suggestive of chronic meningoencephalitis is required . Because CNS involvement is often-silent , staging relies on lumbar puncture to assess chronic meningoencephalitis , especially in field screening wherein few cases have neurological signs [14 , 15] . CSF leukocyte counts are scored according to stage-2 cut-offs recommended by WHO [12] . Detection of trypanosomes in CSF does not define ‘chronic’ CNS infection , since the immune system may also destroy the parasite [14 , 16] . Hence , determination of persistent disease to eliminate parasite reservoirs in the population remains an unmet challenge . Assay sensitivity for T . b . gambiense detection even by molecular tests , i . e . polymerase chain reaction ( PCR ) and loop-mediated isothermal amplification ( LAMP ) [17–23] , is often limited by the stoichiometric presence of the parasite in the assayed sample . Cox et al . [24] reported the difficulty to establish true trypanosome ( T . congolense , T . vivax , T . b . brucei ) prevalence in blood spotted on paper cards using specific PCR detection tools in indigenous African zebu cattle due to chronically low parasitemias [24] . Because parasite DNA was unevenly distributed across the card , a single punch from an FTA card was insufficient to confirm infectivity: i , e . the stochastic sampling effect results in underestimation of prevalence [24] . The same stoichiometric apply to blood/CSF-based molecular assays sufficiently sensitive to detect DNA below the content of a single parasite; i . e . the detection limit of the assay is still restricted by the number of parasites present in the volume of sample assayed [22] . Remarkably , the answer was simple: i . e . LAMP assays that recognize multi-copy gene targets for trypanosome DNA are dramatically enhanced by sample pretreatment with detergents to lyse or solubilize their DNA prior to assay [22] . By pre-lysing cells with detergent before application of parasite-specific LAMP primers and amplification that recognize multi-copy gene targets we have been able to markedly improve the detection of parasite genomic DNA by LAMP . Using human CSF spiked with trypanosomes as direct source of DNA template , we found that detergent-enhanced LAMP assay targeting multi-copy trypanosome genes reached analytical sensitivities about 100 to 1000-fold or lower [22] . Similar increases in LAMP assay analytical sensitivity were also found using DNA extracted from filter paper cards containing blood pretreated with detergent before card spotting , or using DNA extracted from blood samples spotted on detergent-pretreated air-dried cards for improved assay reproducibility [22] . Hayashida et al [25] later showed that RNA could also be amplified directly from detergent-lysed blood samples . Here we assess and show that detergent-enhanced LAMP is a simple point of care ( POC ) molecular assay platform for T . b . gambiense parasite detection in clinical samples where chronically low parasitemias are expected . As stage determination relies on lumbar puncture to examine CSF for trypanosomes confirming neurological invasion [15] , our findings are intriguing in that the data provide evidence for detection of genomic T . b . gambiense DNA in the CSF independent of HAT stage . Overall , it is predicted that this simple technological advance will greatly improve POC pathogen detection including those trypanosomes in the so-called aparasitemic individuals who also may or may not be seropositive for the parasites and environmental monitoring .
A de-identified cohort of 150 clinical samples was obtained from HAT patients during studies in Central African Republic ( Batangafo focus , 2001 ) under the direction of “Programme National de Lutte contre la Trypanosomose Humaine Africaine” ( PNLTHA ) ( S1 Table ) . Written informed consent was received from these subjects prior to enrollment and/or from their parents or guardians for participants below 18 years of age . All patients in the collection were screened for clinical signs and specifically for neurological and psychiatric disturbances . Samples from patients testing positive for microscopic presence of malaria ( blood smear ) , filariasis ( blood examination by capillary tube centrifugation ) , schistosomiasis ( when blood was detected in urine ) , as well as by retrospective testing of stored samples for HIV and syphilis , were excluded . All clinical samples that remained after the above clinical diagnostic procedures were aliquoted and stored in liquid nitrogen before being transported on dry ice to the Institute of Tropical Neurology at Limoges University . The anonymized samples were archived and stored at -70°C and at no time was there a break in the cold-chain until the aliquoted samples were used for the LAMP assay for this study . Because it is not ethical to perform a lumbar puncture on a healthy person , we used negative control human CSF obtained as discarded de-identified clinical samples from The Johns Hopkins Hospital Microbiology laboratory that were obtained from patients with neurological manifestations with approval of the Johns Hopkins Medicine Institutional Review Board ( IRB ) . Gambiense HAT was confirmed by a positive card agglutination test for trypanosomiasis ( CATT ) and with trypanosomes microscopically identified in blood and posterior cervical lymph nodes if latter were enlarged , and on CSF smears based on WHO guidelines adapted for populations where the prevalence of the HAT is high- ( > 1% ) areas [12] . Based on the prevalence of HAT in the area our cohort was obtained , our decision pathways were based on a published algorithm ( Fig 1 in ref [26] ) modified with a focus on having a real "stage-1" group . We considered all patients with presence of trypanosomes in blood or lymph and with less than 5 white blood cells ( WBC ) /μL ) , no trypanosomes in CSF and no neurological signs as stage-1 . All others at stage-2 , so patients with 10 cells without trypanosomes in CSF were considered stage-2 . A positive CATT at ≥1:16 dilution with documented trypanosomes ( either by microscopy or by mAECT ) was our reference standard for confirmed gambiense HAT . Positive CATT ≥1:16 without evidence of trypanosomes was considered serologic HAT . Positive CATT ≥1:4 but <1:16 without evidence of trypanosomes was considered possible HAT . Negative CATT ( <1:4 ) was considered not HAT ( control ) . Patients with HAT and CSF with evidence of IgM ( nephelometric detection ) [27] , or trypanosomes ( microscopic detection ) , or increased CSF white cell blood count ( > 5 WBC/μL ) , or relapse after stage-1 HAT treatment , were classified as stage-2 HAT . Patients presenting with HAT without IgM or trypanosomes in the CSF and with CSF cell counts ≤5 WBC/μL who had not relapsed after stage-1 treatment were also considered stage-1 HAT . The presence or absence of trypanosomes ( by microscopy or mAECT ) dictated whether stage-1 and stage-2 HAT cases were confirmed or serologic , respectively . Overall , the final cohort consisted of de-identified archived CSFs from 150 HAT patients—includes 95 adults; 21 patients between the ages of 12 to 17; and 34 patients <12 years of age—clinically defined as stage-1 ( 73 samples ) or stage-2 ( 77 samples ) . Negative control samples were available from 100 patients . Two LAMP primer sets targeting the pan-T . brucei—500 copy—repetitive insertion mobile element ( RIME ) of subgenus Trypanozoon ( GenBank Accession No . K01801 ) ( RIME LAMP ) and the—200 copy—T . b . gambiense 5 . 8S rRNA-internal transcribed spacer 2 ( 5 . 8S-ITS2 ) gene ( GenBank Accession No . AF306777 ) ( TBG1 LAMP ) were used ( S2 Table ) . The analytical specificity for trypanosome DNA and sensitivity ( equivalent to 0 . 01 parasite or less ) for RIME and TBG1 LAMP are well documented [20 , 22 , 23 , 28 , 29] . All LAMP primers ( Forward and Backward Primers F3 and B3; Forward and Backward Inner Primers FIP and BIP; and Forward and Backward Loop Primers LF and LB ) were synthesized and HPLC-purified by Integrated DNA Technologies ( IDT ) . A 10% ( w/v ) Triton X-100 stock solution was made by adding 1 g Triton X-100 to a final volume of 10 mL DNase/RNase free water ( Qiagen ) . The clinical CSF samples were adjusted to contain 1/20 volume of 10% Triton X-100 ( final concentration 0 . 5% Triton ) , or 1/20 volume deionized water ( untreated sham CSF ) [22] . The CSFs were assayed immediately or after a 60 min incubation at ambient temperature to allow for complete detergent lysis prior to LAMP as we previously described [22] . All LAMP reactions using commercially available kits ( Eiken Chemical Co , Japan ) were previously optimized for reagent concentration , reaction time and temperature in real-time in a Loopamp real-time turbidimeter LA320C ( Teramecs , Tokyo , Japan ) as previously described [17 , 22 , 29 , 30] . Briefly , the reaction contained 12 . 5 μL of 2x LAMP buffer ( 40 mM Tris-HCl [pH 8 . 8] , 20 mM KCl , 16 mM MgSO4 , 20 mM [NH4]2SO4 , 0 . 2% Tween 20 , 1 . 6 M Betaine , 2 . 8 mM of each deoxyribonucleotide triphosphate ) , 1 . 0 μL primer mix ( 5 pmol each of F3 and B3 , 40 pmol each of FIP and BIP ) or 1 . 3 μL when LF and LB ( 20 pmol each ) were included , 1 μL ( 8 units ) Bst DNA polymerase ( New England Biolabs , Tokyo , Japan ) and 1 μL of human CSF . Final volumes were adjusted to 25 μL with water . LAMP reactions monitored for 60 min by measuring turbidity in real-time as previously described [22] were conducted in duplicate and at optimal reaction temperatures , 62°C for RIME LAMP and 63°C for TBG1 LAMP , prior to termination at 80°C for 5 min . We considered precipitation occurring after a reaction time of 60 minutes to be nonspecific artifacts . For end-point analysis ( S1 Fig ) the amplified products were analyzed using the E-Gel high throughput DNA electrophoresis system with ethidium bromide or SYBR green incorporated into the gels ( Invitrogen ) , or after addition of hydroxy naphthol blue ( HNB ) to monitor the sample color change from violet to sky blue , a readout unaffected by detergent , has been interpreted by independent observers as the easiest to see [23] . As with any DNA amplification method , standard precautions for avoiding template contamination [31] also apply for LAMP-based assays . Thus , the reactions were assayed in 4 blocks with each block containing 8 samples with 2 no template controls for every 6 samples assayed . A false positive response from any no template or non-HAT CSF control negated the entire 32-reaction run and the samples were re-assayed . Statistical significance between data sets obtained with RIME versus TBG1 LAMP was determined using the Cohen’s kappa coefficient ( Vassar Stats; http://vassarstats . net/kappa . html ) as a measure of agreement between Triton X-100 pretreated patient CSF samples that were positive or negative for trypanosomes . For reference , the following Kappa coefficients and levels of agreement are as follows: 0 . 80 to 1 . 00 = Very good agreement , 0 . 60 to 0 . 80 = Good agreement , 0 . 40–0 . 60 = Moderate agreement , 0 . 20–0 . 40 = Fair agreement , while values <0 . 20 = Poor agreement . Additionally , we applied the χ2 test and Fisher’s exact probability test to the 2x2 contingency table generated comparing the two approaches ( Vassar Stats; http://vassarstats . net/tab2x2 . html ) .
Using pan-T . brucei RIME LAMP as our ‘gold standard’ LAMP assay , conventional WHO staging concepts predicted that a significant number of the stage-2 CSF samples would be LAMP positive , while the stage-1 CSFs would yield predominantly negative results . To establish baseline clinical control values , RIME and TBG1 LAMP were performed on at least 100 negative control CSF samples obtained from the Johns Hopkins Hospital ( Baltimore , MD USA ) . All control CSF samples were found to be RIME- and TBG1 LAMP-negative whether or not they were pre-incubated for 1 h with detergent prior to assay to allow for optimal sample lysis [22] . We then tested the archived CSFs from 150 HAT patients—includes adults and children the between the ages of 3 to 17 years ) —clinically defined as stage-1 ( 73 samples ) or stage-2 ( 77 samples of which 4 were microscopically parasite positive ) ( S1 Table ) . Overall , under sham conditions using only 1 μL sample and pan-T . brucei RIME LAMP , we found that 1/73 ( 1 . 4% ) of the stage-1 samples tested positive possibly due to the presence of circulating free DNA ( cfDNA ) , as the patient was previously treated for HAT ( S1B Table; see below ) and 2/77 ( 2 . 6% ) of the stage-2 HAT CSF samples tested were parasite DNA-positive by pan-T . brucei RIME LAMP . We have shown that release of parasite DNA by 0 . 5% Triton required between 30 and 60 min incubation [22] . While the percentage of trypanosome DNA-positive CSFs by RIME LAMP conducted immediately after Triton X-100 addition increased 2-fold , LAMP conducted 60 min after detergent addition revealed that 21 out of 77 ( 27 . 3% ) stage-2 CSFs tested were parasite DNA-positive ( Table 1A ) . In patients with HAT , especially with T . b . rhodesiense , it is more difficult to detect trypanosomes in the CSF compared with the blood where the parasitaemia is generally high [9] though much less so in T . b . gambiense . Remarkably , even under our stringent assay conditions whereby only 1 μL of CSF was directly assayed , RIME LAMP with detergent did identify 2 ( 1 adult ( ID 8 ) and 1 pediatric <12 years of age ( ID 107 ) ; S1 Table ) of the 4 patient CSFs in which trypanosomes were identified microscopically . Somewhat unexpected was the dramatic 1 . 4% ( 2/73 ) to 42 . 5% ( 31/73 ) increase in the number of parasite DNA-positive CSFs for stage-1 samples ( Table 1A ) . It has been suggested that molecular tests ( PCR ) may not be suitable for post-treatment follow-up of HAT cure because of persistence of trypanosome cfDNA that may lower test specificity [32–34] . Released into the bloodstream as a result of cell death , necrosis , or by release by viable cells , cfDNA has been found in many disease conditions [35] . In our cohort , 22 samples ( 9 stage-1 and 13 stage-2 ) were from patients that had previously been treated ( with some trypanocide ) for HAT prior to sample collection ( S1B Table ) . While 8 out of 9 stage-1 and 5 out of 13 stage-2 Triton-pretreated samples were RIME LAMP positive , one stage-1 sample , ID2 , and two stage-2 samples , ID1 and ID12 , were strongly LAMP positive under sham condition . In fact , this was the only sham sample in the entire 150 sample cohort positive for trypanosome DNA ( Compare S1A Table to S1B Table ) . A re-analysis of the 128 samples from HAT patients who never received any trypanocide prior to collection ( S1A Table ) , also showed that while none of the stage-1 samples were LAMP-positive , 3 . 1% of the stage-2 HAT CSF samples tested were parasite DNA-positive by RIME LAMP , a value predicted provided that the samples on average had 1 parasite / 20 μL sample assayed: i . e . 1/20 x 64 = 3 . 2% . RIME LAMP conducted on these Triton pretreated CSF samples showed a 25% ( 16/64 ) increase in the number stage-2 CSFs as well as a 35 . 9% ( 23/65 ) increase in the number of parasite DNA positive stage-1 samples ( Table 1B ) . These findings were confirmed by repeating the experiments in the absence or presence of Triton using TBG1 LAMP targeting the T . b . gambiense specific 5 . 8S-ITS2 gene ( S1 Table and Table 1 and Fig 1 ) . When properly executed , detergent-enhanced LAMP is able to detect very low parasite numbers , but the assay’s sensitivity is a potential drawback because of risk for amplicon contamination . Thus , we used the Cohen’s kappa test to compare the LAMP data from two unrelated trypanosome gene targets . Based on data derived from the entire cohort , Cohen’s kappa test identified a high level of agreement between RIME and TBG1 LAMP in samples that had been pre-incubated with Triton X-100 prior to assay ( the most sensitive condition ) when testing CSFs , thus indicating that the results are not due to amplicon contamination ( Table 2 ) . Kappa coefficients for RIME versus TBG1 LAMP of 0 . 92 ( 95%CI: 0 . 82–1 . 00 ) for stage-1 and 0 . 90 ( 95%CI: 0 . 80–1 . 00 ) for stage-2 reflect a high level of agreement between the data set , data confirmed in χ2 tests ( p<0 . 001 ) and Fisher’s exact probability test ( p = 4 . 7e-13 ) . Inclusion of the 100 negative control CSFs strengthens the level of agreement: 0 . 94 ( 95%CI: 0 . 87–1 . 00 ) for stage-1 and 0 . 92 ( 95%CI: 0 . 83–1 . 00 ) for stage-2 . Re-analysis of Kappa after removing the 22 samples from patients previously treated with trypanocide ( data from S1B Table only ) did not significantly change the kappa coefficients for RIME versus TBG1 LAMP for stage-1 and for stage-2 HAT; 0 . 90 ( 95%CI: 0 . 78–1 . 00 ) for stage-1 , 0 . 88 ( 95%CI: 0 . 75–1 . 00 ) for stage-2 , and 0 . 89 ( 95%CI: 0 . 81–0 . 98 ) for all stages .
Overall , the evidence support the idea that current WHO criteria need to go beyond the prevailing notion that the transition from acute to CNS stage disease entails crossing brain barriers to enter the CNS to initiate inflammation that define clinical staging for CNS [41] . Whether detection of trypanosome genomic DNA by detergent-enhanced LAMP in the CSF independent of HAT stage provides supportive clinical evidence for early CNS entry , reflects parasite contamination of CSF with infected blood during lumbar puncture , and/or other scenarios , identify critical knowledge gaps for future studies . Controlled laboratory and field validation of detergent LAMP [22] or the adaptation of more recent ultrasensitive LAMP assays based on DNA-protein chimeras called LAMPoles for detection of non-nucleic based targets [52] that include HAT staging biomarker molecules [53 , 54] that take into account saliva , skin and subcutaneous fat-dwelling trypanosomes may help facilitate development of better algorithms for HAT diagnosis and monitoring . Further , the detection of parasite DNA regardless of sample source by detergent LAMP may help provide for very early T . b . gambiense HAT diagnosis regardless of staging . The use of these LAMP-based platforms as epidemiologic tools for the determination human African trypanosome prevalence in areas where chronically low parasitemias are present or in areas where there are high numbers of asymptomatic ( carriers ) patients should also be considered .
|
Human African trypanosomiasis is a fatal disease ( if untreated ) spread by bloodsucking tsetse flies . These protozoan parasites first enter the lymph and blood to invade many organ systems ( early stage sleeping sickness ) . Weeks to months later , the parasites invade the brain causing a wide variety of neurological symptoms ( late stage sleeping sickness ) . In rural clinical settings , diagnosis still relies on the detection of these microbes in blood and cerebrospinal fluid ( CSF ) by microscopy . LAMP , or loop-mediated isothermal amplification of DNA , is a technique that can specifically detect very small amounts of DNA from an organism . We previously showed that by simply adding detergent during sample preparation , the analytical sensitivity of LAMP targeting many gene copies is greatly improved , presumably because DNA is released from the pathogen cells and dispersed through the sample . We demonstrated proof of principle using pathogenic trypanosomes in different human body fluids ( CSF or blood ) and showed that this simple modification should be applicable for diagnosis of other microbial infections where cells are sensitive to detergent lysis . After completion of the above published study , we tested a collection of clinical CSF samples from African patients diagnosed with early or late stage sleeping sickness based on current World Health Organization ( WHO ) guidelines . For proof-of-concept we tested only a single microliter of detergent-treated CSF to test for late stage disease . We predicted that a significant number of the late stage samples would be LAMP positive , while the early stage CSFs would yield predominantly negative results . Instead , our study detected trypanosome DNA in patient CSF independent of African sleeping sickness stage , results that may be consistent with early brain entry and other scenarios that identify critical knowledge gaps for future studies .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"nervous",
"system",
"african",
"trypanosomiasis",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"saliva",
"protozoans",
"materials",
"science",
"neglected",
"tropical",
"diseases",
"surfactants",
"infectious",
"diseases",
"detergents",
"zoonoses",
"protozoan",
"infections",
"trypanosomiasis",
"trypanosoma",
"eukaryota",
"blood",
"anatomy",
"central",
"nervous",
"system",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"materials",
"cerebrospinal",
"fluid",
"organisms"
] |
2019
|
Using detergent-enhanced LAMP for African trypanosome detection in human cerebrospinal fluid and implications for disease staging
|
Mechanically gated ion channels convert sound into an electrical signal for the sense of hearing . In Drosophila melanogaster , several transient receptor potential ( TRP ) channels have been implicated to be involved in this process . TRPN ( NompC ) and TRPV ( Inactive ) channels are localized in the distal and proximal ciliary zones of auditory receptor neurons , respectively . This segregated ciliary localization suggests distinct roles in auditory transduction . However , the regulation of this localization is not fully understood . Here we show that the Drosophila Tubby homolog , King tubby ( hereafter called dTULP ) regulates ciliary localization of TRPs . dTULP-deficient flies show uncoordinated movement and complete loss of sound-evoked action potentials . Inactive and NompC are mislocalized in the cilia of auditory receptor neurons in the dTulp mutants , indicating that dTULP is required for proper cilia membrane protein localization . This is the first demonstration that dTULP regulates TRP channel localization in cilia , and suggests that dTULP is a protein that regulates ciliary neurosensory functions .
The auditory system allows animals to communicate and obtain information about their environment . The hearing organs transform sound into an electrical signal through a process called mechanotransduction , the conversion of a mechanical force impinging on a cell into an intracellular signal [1] . Although the recent discovery of several molecules involved in mechanotransduction allows interpretation of the biophysical properties of the mechanotransduction process for hearing [2] , many additional molecular players in auditory development and function are waiting to be unveiled . Drosophila melanogaster has been suggested as a model organism to study the fundamental process of hearing [3] , [4] . Hearing in the fly is necessary for the detection of courtship songs [5]–[7] . Male-generated courtship song causes females to reduce locomotion and enhances female receptivity , whereas it causes males to chase each other [8] . The ability to hear courtship songs is ascribed to Johnston's organ ( JO ) in the second antennal segment . Near-field sounds rotate the sound receiver; the third antennal segment and the arista and this rotation of the antennal receiver transmits mechanical forces to the JO in the second antennal segment , which is connected to the third antennal segment by a thin stalk [9] . Each JO sensilla consists of two or three chordotonal neurons and several supporting cells . The outer dendritic segments of the JO neurons are compartmentalized cilia which are directly connected to the antennal sound receiver via extracellular caps . The distortion of the junction between the second and third segment stretches the cilia and stimulates the JO neurons . Several transient receptor potential ( TRP ) channels have been shown to be required for Drosophila hearing transduction and amplification [4] , [10]–[14] . Mutation in nompC , the Drosophila TRPN channel , resulted in substantial reduction of sound-evoked potentials [4] . Reports showing that NompC and TRP-4 ( the C . elegans ortholog of NompC ) are bona fide mechanotransduction channels support the idea that NompC is the Drosophila hearing transducer [15] , [16] . Two Drosophila TRPV channel , inactive ( iav ) and nanchung ( nan ) , mutants showed complete loss of sound-evoked action potentials [11] . However , they have not been considered to be the hearing transduction complex per se; rather they are thought to be required to amplify the electric signal generated by the hearing transduction complex , since Iav and Nan reside in the proximal cilia which are distant from the distal cilia where NompC is localized and mechanical force is directly transmitted [17] , [18] . A recent study which employed a new method to measure subthreshold signals from the JO neurons suggested the opposite possibility that the TRPV ( Iav and Nan ) complex is the hearing transduction complex modulated by TRPN ( NompC ) [14] . Although the exact roles of each TRP in Drosophila hearing are still controversial , it is clear that TRPN and TRPV have essential and distinct roles in Drosophila hearing . Several attempts have been made to identify molecular players regulating the function of the ciliated mechanoreceptor neurons . Gene expression profiling identified chordotonal organ-enriched genes from campaniform mechanoreceptors , developing embryo chordotonal neurons , and the second antennal segment [19]–[21] . Alternatively , chordotonal neuron-specific genes were identified by searching for regulatory factor X ( RFX ) -binding sites , because ciliogenesis of the chordotonal neurons mainly depends on the RFX transcription factor [22] . However , so far only a limited number of genes involved in TRP channel localization in the JO neuron cilia have been identified and characterized , including axonemal components and intraflagellar transports ( IFTs ) [17] , [23] . IFTs are indispensable for the formation and maintenance of cilia as well as for the transport of proteins along the microtubules in and out of the cilia [24]–[26] . Therefore , mutation of many of the characterized genes results in not only delocalization of the TRPs but also profound structural abnormality in cilia , rendering it difficult to delineate the gene functions specific to TRP localization . Tubby is the founding member of Tubby-like proteins ( TULPs ) [27] . Loss-of-function of the Tubby gene exhibits adult-onset obesity , retinal degeneration , and hearing loss in mice . The Drosophila genome encodes one Tubby homolog called King tubby ( hereafter designated dTULP ) , which shares approximately 43% amino acid identity with mouse Tubby ( Figure S1A ) [28] . At the embryonic stage , dTULP is expressed in various types of neuronal cells including the chordotonal neurons . Although previous expression analyses and bioinformatic approaches detected dTulp in the chordotonal organs , its presence did not attract much interest because of its distribution in various neuronal cell types [22] . In this study , we aimed to investigate the novel molecular function of dTULP in Drosophila hearing . dTULP is localized to the well-defined ciliary structure of Drosophila auditory organs . Loss of dTULP has no effect on the ciliary structure of the JO neurons , but NompC and Iav localization in cilia was severely altered . These data demonstrate a new role of dTULP as a regulator of TRP localization in the hearing organs .
To test whether dTULP plays a role in Drosophila hearing , we generated two dTulp mutant alleles by ends-out homologous recombination [29] . The first allele was dTulp1 , which harbours a deleted C-terminal containing the conserved “tubby domain” ( residues 220 to 460; Figure 1A ) . The second allele , dTulpG , was generated by replacing an N-terminal portion of the dTULP coding region ( residues 18 to 261; Figure S1B ) with GAL4 coding sequences at the site corresponding to the initiation codon of the short splicing variant of dTulp . Genomic PCR analyses showed that the dTulp genomic locus was deleted in dTulp1 and dTulpG flies ( Figure 1B and Figure S1B ) . We raised antibodies to dTULP , which recognized a 51 kDa protein as predicted in wild-type fly extracts on a Western blot , and confirmed that dTULP was not detected in dTulp1 and dTulpG fly extracts ( Figure 1C ) . Both alleles are homozygous viable and fertile . Since both dTulp1and dTulpG mutant alleles showed postural problems and uncoordinated movement , we performed a climbing assay . Flies were banged down to the bottom of a vertical tube and the percentage of the flies climbing above half of the height of the vertical tube within 10 seconds was recorded as the climbing index . dTulp1 , dTulpG , and transheterozygote flies exhibited a decreased climbing index compared to control flies ( Figure 1D ) . Introduction of a P[acman] clone containing the dTULP coding region ( CH321-59C17 ) in the dTulp1 mutant background rescued this phenotype [30] . These data suggested that dTulp mutants may have functional defects in the JO neurons [13] . To check for hearing defects in dTulp mutant flies , we recorded extracellular sound-evoked potentials in wild-type and dTULP-deficient flies . Sound-evoked potentials were completely abolished in dTulp1 , dTulpG , and dTulp1 in trans with a deletion that completely removed dTulp , Df ( 2R ) BSC462 . Genomic rescue using the P[acman] clone produced sound-evoked potentials similar to those in the wild-type , suggesting that the hearing defect was specifically due to dTulp ablation ( Figure 1E and 1F ) . To test whether dTULP is expressed in the JO neurons , we first attempted to take advantage of the GAL4/UAS system using the dTulpG allele . However , the GAL4 reporter inserted in dTulpG was not expressed . This may be caused by inserting GAL4 at the site corresponding to the initiation codon of the short splicing variant of dTulp rather than the long splicing variant . Therefore , we performed immunohistochemistry with dTULP antibodies . We found that dTULP was expressed in the cilia as well as the cell body of the chordotonal neurons ( Figure 2A , left ) . We did not detect dTULP immunoreactivity in the JO neurons in dTulp1flies , indicating that the immunosignal is specific for dTULP ( Figure 2A , right ) . To further characterize the ciliary localization of dTULP , we compared the localization of dTULP with that of Iav and NompC . The subcellular localization of Iav and NompC are in the proximal and distal cilia , respectively , in a mutually exclusive manner ( Figure 2B ) [11] , [17] , [18] , [31] . dTULP staining extended from the proximal to distal cilia with a much weaker signal observed in the distal portion ( Figure 2C and 2D ) . The mouse Tubby protein has been reported to shuttle from the plasma membrane to the nucleus upon Gq-coupled G protein-coupled receptor ( GPCR ) activation [32] . dTULP was also detected in the cell body as well as the nucleus in the JO neurons ( Figure 2A and Figure S7B ) . We also found that dTULP was expressed in other types of sensory neurons with cilia ( Figure S2 ) . To examine whether the dTulp mutants have developmental defects in the JO neuron structure , we observed the expression of a membrane-targeted GFP ( UAS-mCD8:GFP ) driven by the pan-neuronal promoter ( elav-GAL4 ) in the JO neurons . We found no gross structural abnormalities in dTulp1flies ( Figure 3A ) . Electron microscopy of the JO revealed that most dTulp mutants had normal ciliary ultrastructure ( Figure 3B ) . Approximately 9 . 3% of chordotonal scolopidia appeared abnormal in terms of cilia number or cap-cilia connections ( Figure S3 ) . In addition , we did not observe any discernible changes in the expression of the dendritic cap protein NompA , which transmits mechanical stimuli to the distal segment of chordotonal neurons in dTulp mutants ( Figure 3C ) [33] . These observations suggest that structural changes in the JO cannot account for severe hearing impairment in dTulp mutants . Mutations of trps , including iav and nompC , cause hearing defects in Drosophila [4] , [11] . To investigate the possibility that dTULP controls the expression of TRPs and other genes which are indispensable for Drosophila hearing , we performed quantitative PCR analysis of such genes and no significant differences in expression levels were present between wild-type and dTulp1 antennae ( Figure S4 ) . This suggested that dTULP plays other roles in Drosophila chordotonal neurons rather than as a transcription factor that controls transcription of known hearing related genes , although we cannot exclude the possibility that dTULP regulates the expression of hearing related genes we did not survey . Next we examined the ciliary localization of Iav and NompC in the dTulp mutants . Surprisingly , Iav was not localized to the proximal cilia in dTulp1 flies ( Figure 4A ) . Furthermore , NompC , characteristically localized to the distal cilia ( Figure 2B ) , was redistributed toward the proximal cilia ( Figure 4B ) . Spacemaker ( Spam ) is an extracellular protein which protects cells from massive osmotic stress [34] . Localization of Spam was also altered in dTulp mutants from its two typical locations: the luminal space adjacent to the cilia dilation and the scolopidium base ( Figure 4C ) [35] . Introduction of the dTulp+ transgene rescued the localization of Iav , NompC , and Spam ( Figure 4 ) . IFTs are involved in the localization of Iav , NompC , and Spam [17] , [23] . Because IFT mutants show similar phenotypes to the dTulp mutant , we investigated the localization of IFT proteins in dTULP-deficient flies . Ciliary localization of the two IFTs , NompB ( the ortholog of human IFT-B , IFT88 ) and RempA ( the ortholog of human IFT-A , IFT140 ) , was unaffected in dTulp1 mutants ( Figure S5 ) . To further address the functional relationship between dTULP and IFTs , we examined distribution of dTULP in three IFT ( nompB , rempA , and oseg1 ) mutants and a retrograde motor dynein heavy chain ( beethoven ) mutant . Although the rempA , oseg1 , and beethoven mutants show different degrees of defective cilia structure , dTULP is localized to the deteriorated cilia of each mutant , suggesting that rempA , Oseg1 , and beethoven are not required for the transport of dTULP into the cilia ( Figure S6A–S6C ) . Since the nompB mutant does not develop cilia structure , dTULP was present in the inner segment at a high level ( Figure S6D ) [36] . However , it is possible that other IFTs may play a role for dTULP ciliary localization even though the IFTs we examined are not involved in ciliary localization of dTULP . Mammalian Tubby have two distinct domains: nuclear localization signal ( NLS ) and phosphoinositide ( PIP ) -binding domain . An NLS , which allows Tubby to translocate into the nucleus , resides in the N-terminal region of Tubby [32] . Recently , a short stretch of amino acids including the NLS in TULP3 , a mammalian member of the Tubby-like protein family , has been reported as an IFT-A binding domain [37] . A PIP-binding domain in the C-terminal tubby domain allows Tubby to be localized under the inner leaflet of the plasma membrane through binding to specific phosphoinositides . These domains are also conserved in dTULP ( Figure 5A ) . In order to investigate the mechanism by which dTULP regulates the ciliary localization of Iav and NompC , we introduced mutations into the putative NLS/IFT-binding ( dTULPmutA ) , PIP-binding domain ( dTULPmutB ) , or both domains ( dTULPmutAB ) of dTulp cDNA and generated UAS-wild-type dTulp ( UAS-dTulpwt ) , UAS-dTulpmutA , UAS-dTulpmutB , and UAS-dTulpmutAB transgenic flies , respectively . To eliminate positional effects , all transgenes were integrated into the same loci using site-specific recombination with an attP landing site on the third chromosome [38] . To test the effect of each mutation on the subcellular localization of dTULP , we examined the subcellular localization of dTULPwt , dTULPmutA , and dTULPmutB in Drosophila salivary glands . dTULPwt was detected mainly in the plasma membrane and nucleus ( Figure S7A ) . Mutations in the NLS/IFT-binding domain or PIP-binding domain of dTULP resulted in significant exclusion from the nucleus or accumulation in the nucleus , respectively , which suggested that the NLS/IFT-binding and PIP-binding properties of mouse Tubby are conserved in dTULP in Drosophila salivary glands ( Figure S7A ) . However , the localization of dTULPwt , dTULPmutA , and dTULPmutB in the JO neurons in terms of the cell body and nuclear distribution was virtually the same ( Figure S7B ) . These data suggested that dTULP is not shuttled between the plasma membrane and the nucleus in the JO neurons and these domains may have other functions in the JO neurons rather than controlling the translocation of dTULP from the plasma membrane to the nucleus . To evaluate the functional consequences of each mutation , we expressed dTULPwt , dTULPmutA , dTULPmutB , or dTULPmutAB in the JO neurons of dTulp1 flies . The expression of dTULPwt in the dTulp mutant background restored the distribution and the expression level of Iav and NompC similar to those of wild type ( Figure 5B and 5C ) . The expression of dTULPmutA or dTULPmutB rescued the Iav trafficking defect of the dTulp mutant , but the expression levels of Iav in the proximal cilia in dTULPmutA- or dTULPmutB-expressing flies were reduced compared to those of dTULPwt-expressing flies ( Figure 5B and 5E ) . NompC localization to the distal cilia in dTULPmutA- or dTULPmutB-expressing flies was similar to that in dTULPwt-expressing flies ( Figure 5C ) . dTULPmutAB could not rescue the Iav or NompC localization defects of the dTulp mutant . This difference was not due to the expression levels of the mutant dTulp transgene since the expression levels of mutant forms of dTULP were similar to those of wild-type dTULP ( Figure S8 ) . Next , we examined whether the different degrees of rescue of Iav and NompC localization was due to differential ciliary trafficking of variant forms of dTULP . The ciliary expression level of dTULPmutB was similar to that of dTULPwt , whereas the ciliary expression levels of dTULPmutA and dTULPmutAB were reduced compared with those of dTULPwt ( Figure 5D and 5F ) . These data suggested that the putative NLS/IFT-binding domain of dTULP has a regulatory function to control the trafficking of dTULP into the cilia . Consistent with immunohistochemical analyses , dTULPwt fully rescued the hearing defect of the dTulp mutant . dTULPmutA and dTULPmutB restored a partial function and dTULPmutAB had no such activity ( Figure 5G and 5H ) .
In the current study , we demonstrate that dTULP is a cilia trafficking regulator in the Drosophila hearing system . Mutation of dTulp results in hearing loss due to the mislocalization of two TRP channels , Iav and NompC , which are ciliary membrane proteins . In addition , Spam , whose localization is dependent on the IFT machinery , is also mislocalized in dTulp mutants . How does dTULP regulate the ciliary distribution of TRPs in the JO neurons ? Several studies have shown that mutations in IFT machinery or cilia components result in mislocalization of Iav , NompC , and Spam , along with abnormal axonemal structure [17] , [23] . It is notable that , in contrast to IFT or cilia component mutants , ciliogenesis and maintenance appear normal in dTULP-missing flies . Furthermore , the altered distribution of Iav , NompC , and Spam in dTulp mutants was not due to the mislocalization of IFT proteins , since the localization of two IFTs ( NompB and RempA ) was normal in dTulp mutants ( Figure S6 ) . These data suggest that dTULP acts downstream of the IFTs to regulate TRP localization . Even though the mutation of dTulp affected the trafficking of both Iav and NompC , the compartmentalized ciliary localization of Iav and NompC is differentially regulated by dTULP . An individual mutation in either the putative IFT- or PIP-binding domain reduced Iav expression levels in cilia , whereas NompC localization was not altered until both domains were mutated . Even after the double mutations in both domains of dTULP , NompC is still situated inside the cilia , but in abnormal locations . These findings demonstrate that ciliary entry of NompC is not dependent on dTULP while the distal ciliary localization of NompC is dependent on dTULP . One possibility is that dTULP allows NompC to disengage from the IFT complex at the distal cilia so that NompC is enriched in the distal cilia through the mechanism that required both IFT- and PIP-binding domains . It is also possible that the distal ciliary localization of NompC is regulated by an unidentified factor ( s ) whose ciliary localization is dTULP-dependent as is Iav . Both the putative IFT- and PIP-binding domains play important roles in the proper Iav distribution in cilia , but they appear to have different roles . Even though the IFT- or PIP-binding mutant forms of dTULP could only partially rescue the ciliary levels of Iav to the similar extent , the mutation of the IFT-binding domain reduced the ciliary levels of dTULP while disruption of the PIP-binding domain had no effect on the ciliary levels of dTULP . These findings suggest that two domains play distinct roles in the regulation of the ciliary localization of Iav . The IFT-binding domain is the motif required for the ciliary entry for dTULP , and the PIP-binding domain is not related to dTULP ciliary entry itself , rather it affects recruitment of Iav-containing preciliary vesicles to dTULP . By these two linked steps , Iav localization to cilia would be facilitated by dTULP . In mammals , IFT-A directs the ciliary localization of TULP3 through physical interaction between TULP3 and the IFT-A core complex ( WDR19 , IFT122 , and IFT140 ) , and in turn , promotes trafficking of GPCR to the cilia . Indeed , the depletion of individual IFT-A core complex components affects the ciliary localization of TULP3 , which results in the inhibition of GPCR trafficking to the cilia [37] . It appears that dTULP and TULP3 have the similar molecular mechanisms to regulate ciliary membrane proteins . However , unlike TULP3 , dTULP ciliary access is not dependent on IFT-A . dTULP ciliary trafficking was not affected by the mutation of Oseg1 ( an ortholog of human IFT-A , IFT122 ) or rempA ( an ortholog of human IFT-A , IFT140 ) . Furthermore , the presence of dTULP in cilia did not determine the normal localization of Iav . For example , in the rempA mutant , even when dTULP was localized to the cilia ( Figure S6B ) , Iav was not found in cilia [23] . Taken together , dTULP facilitates the relay of preciliary vesicles to the IFT complex at the base of cilia rather than moving together with ciliary membrane proteins into the cilia as an adaptor between IFT and cargo . dTULP may have other additional roles in cilia , which needs to be explored in the future . Based on our finding that dTULP but not Iav could be found in cilia of IFT mutants , it is also possible that recruitment of Iav-containing preciliary vesicles requires dTULP and additional unknown factors , whose function is altered in IFT mutants . Thus , Iav-containing preciliary vesicles may not be able to form stable interactions with dTULP and IFTs . After the cloning of the Tubby gene two decades ago , one promising hypothesis has been that Tubby is a transcription factor , since Tubby translocates to the nucleus upon GPCR activation and the N-terminal region of Tubby has transactivation potentials [32] , [39] . However , candidate target genes for Tubby have not been identified . Tubby is thought to have additional functions including vesicular trafficking , insulin signaling , endocytosis , or phagocytosis [40]–[43] . It is still not clear how these molecular functions lead to the in vivo phenotypes observed in the tubby mouse . Meanwhile , several studies have hinted at possible connections between the phenotypes of tubby mutant mice and ciliary dysfunction . Tubby mice phenotypes comprise syndromic manifestations that are commonly observed in ciliopathies such as Bardet-Biedle syndrome [44] and Usher syndrome [45] , [46] . Recently , GPCR trafficking into neuronal cilia was reported to be misregulated in tubby mice [47] . Mutation of Tulp1 , a member of the TULPs , in human and mice , exhibits retinal degeneration due to the mislocalization of rhodopsin [48] . TULP3 represses Hedgehog signalling , which is a crucial signalling cascade in cilia , via the regulation of the ciliary localization of GPCRs [49] . Our current study provides additional supports for the idea that TULPs play an important role in ciliary signalling and that the tubby mouse syndrome might be due to the ciliary defects . In contrast to mammalian cells , only specialized cell types have the ciliary structure in Drosophila , and the expression of dTULP is not restricted to organs with the ciliary structure , which suggested that dTULP may have other roles not related to the ciliary function [28] . For example , dTULP mediates rhodopsin endocytosis in Drosophila photoreceptor cells which do not have cilium in contrast to its mammalian counterpart [50] . In summary , we demonstrate an intriguing role of dTULP in governing the ciliary localization of TRP proteins . This is the first in vivo evidence showing that dTULP may have important roles in the maintenance of ciliary functions by regulating the localization of ciliary proteins , thereby maintaining sensory functions .
All fly stocks were maintained in regular laboratory conditions ( conventional cornmeal agar molasses medium , 12-h light/12-h dark cycle at 25°C , 60% humidity ) . Iav-GFP and NompA-GFP were reported previously [13] , [33] . RempA-YFP and NompB-GFP were from M . Kernan . Y . Jan and M . Noll provided UAS-NompC:GFP and Poxn-GAL4 , respectively . Df ( 2R ) BSC462 , elav-GAL4 , UAS-mCD8:GFP , AB1-GAL4 , F-GAL4 , and Orco-GAL4 were from the Bloomington Stock Center ( Bloomington , IN ) . We employed ends-out homologous recombination to generate dTulp mutant alleles . To make the dTulp1 allele , 3 kb genomic DNA at the 5′ and 3′ ends of the tubby domain ( 220 to 460 residues ) coding sequence was PCR amplified from w1118 and subcloned into the pw35 vector . The primer sequences for the 5′ homologous arm of the pw35 vector are 5′-AAAGCGGCCGCCACCGGTGACATCCTCATGTTC-3′ and 5′-AAAGCGGCCGCGTTGCATCACGAACTGGTCGATATTG-3′ . The primer sequences for the 3′ homologous arm of the pw35 vector are 5′-TGAGCTGGCTGGGATCCTCGGGTTGG-3′ and 5′-GTGGATCCTTCCTGGTTGGCATCACGTTGAC-3′ . To generate the dTulpG allele , we used the pw35GAL4loxP vector in which GAL4 and white are flanked by loxP sequences so the cassette can be removed by introducing Cre recombinase . We subcloned the 3 kb of genomic DNA from each of the 5′ and 3′ ends of the dTULP coding region ( 18 to 261 residues ) into the pw35GAL4loxP vector . The primer sequences for the 5′ homologous arm of the pw35GAL4loxP vector are5′-ACAGATCTCACCGTCGCCTGGCTCAGTGCCC-3′ and 5′-GTGGTACCCAGCTGGCGCTGCAAAGCAGTTAAATC-3′ . The primer sequences for the 3′ homologous arm of the pw35GAL4loxP vector are5′-AAAGCGGCCGCGTGGGTTATTGATAGTGATCCTCTA-3′ and 5′-AACCGCGGCGTACAGAATACTCCCTGTTCATGTCT-3′ . We generated transgenic flies by germ line transformation ( BestGene Inc . , Chino Hills , CA ) and screened for the targeted alleles as described previously [51] . Targeted alleles were subjected to outcross for five generations into a w1118 genetic background . We amplified dTulp cDNAs from cDNA clones ( RE38560 ) with PCR and subcloned the fragments into the pUASTattB vector . These constructs were subjected to further modification . We generated the dTulpmutA and dTulpmutB mutant constructs using site-directed mutagenesis to change the sequence encoding R23QKR to L23AAA , and K292LR to A292LA , respectively . The dTulpmutAB construct was generated by introducing the mutation corresponding to dTulpmutB into the dTulpmutA construct . To generate genomic rescue transgenic flies , we obtained the BAC clones CH321-59C17 from the BACPAC Resource Center ( Oakland , CA ) and used these as genomic rescue constructs . Transgenic flies were generated using PhiC31 integrase-mediated transgenesis on the third chromosome to minimize position effect ( Bloomington stock number 24749 ) . Sound-evoked potentials were recorded as described by Eberl et al [4] . Briefly , the fly's antennal sound receivers were stimulated by computer-generated pulse songs . Neuronal responses were detected using a recording electrode inserted in the junction between the first and second antennal segment and a reference electrode was inserted in the dorsal head cuticle . The signals were subtracted with a DAM50 differential amplifier ( World Precision Instruments , Sarasota , FL ) and digitized using Superscope 3 . 0 software ( GW Instruments , Somerville , MA ) . Each trace represents the average responses to 10 stimuli . For whole-mount staining , antennae were dissected at the pupa stage and the labellum and legs were prepared at the adult stage . Salivary glands were dissected from third instar larvae . For antenna sections , fly heads were embedded in OCT medium and 14 µm frozen cryostat sections were collected . Dissected tissues and sections were fixed for 15 min with 4% paraformaldehyde in 1× PBS containing 0 . 2% TritonX-100 ( PBS-T ) and washed three times with PBS-T . The fixed samples were blocked for 30 min with 5% heat-inactivated goat serum in PBS-T and incubated overnight at 4°C in primary antibodies diluted in the same blocking solution . The tissues were washed three times for 10 min with PBS-T and incubated for 1 h at room temperature in secondary antibodies diluted 1∶500 in blocking solution . Following three washes with PBS-T , the samples were mounted with Vectashield ( Vector Laboratories , Burlingame , CA ) and examined using a Zeiss LSM710 confocal microscope ( Jena , Germany ) . To quantify Iav-GFP and dTULP expression levels in cilia , all samples were prepared at the same time and all confocal images were obtained under the same conditions . The pixel intensity of each protein was measured using Zen Software ( Jena , Germany ) . Iav-GFP intensity was measured without immunostaining . Rabbit dTULP antibodies were raised by injecting animals with a purified His-tagged dTULP fusion protein ( residue 95–339 ) , followed by affinity purification . The primary antibodies were used in immunohistochemistry at the following dilutions: rabbit anti-dTULP , 1∶400; 22C10 , 1∶200 ( Hybridoma Bank , University of Iowa ) ; 21A6 , 1∶200 ( Hybridoma Bank ) ; rabbit anti-Orco , 1∶1 , 000 ( gift from L . Vosshall ) ; rabbit anti-NompC , 1∶20; rabbit anti-GFP , 1∶1 , 000 ( Molecular Probes , Eugene , OR ) ; mouse anti-GFP , 1∶500 ( Molecular Probes ) . The secondary antibodies used were Alexa 488- , Alexa 568- , and Alexa 633-conjugated anti-mouse or anti-rabbit IgG ( Molecular Probes; 1∶500 ) . DNA and actin were visualized by DAPI and Alexa Fluor 633 Phalloidin ( Molecular Probes ) staining , respectively . Fly head or antennae lysates from each genotype were subjected to electrophoresis on SDS-polyacrylamide gels and transferred onto polyvinylidene fluoride membranes . The membranes were blocked for 1 h with 5% nonfat milk plus 0 . 1% Tween-20 . Membrane-bound proteins were analyzed by immunoblotting with primary antibodies against dTULP ( 1∶1 , 000 ) and tubulin ( Hybridoma Bank , 1∶2 , 000 ) . Fly heads were dissected and fixed in 2% paraformaldehyde , 2 . 5% glutaraldehyde , 0 . 1 M cacodylate , and 2 mM CaCl2 , pH 7 . 4 . The tissue was embedded in LR white resin . Thin sections were cut , mounted on formvar-coated single slot nickel grids , counterstained with uranyl acetate and lead citrate , and examined on a Hitachi H-7500 electron microscope ( Hitachi , Tokyo , Japan ) . Total RNA was extracted from adult antennae using Trizol reagent ( Invitrogen , Carlsbad , CA ) . cDNA was generated from 0 . 5 µg of RNA from each genotype using the SuperScript III First Strand Synthesis System ( Invitrogen ) . Quantitative PCR was performed using an ABI7500 real-time PCR machine ( Applied Biosystems , Foster City , CA ) and the ABI SYBR green system . Transcript levels were normalized to rp49 as an internal control and the ΔCT ( CT = threshold cycle ) method was used to calculate the relative amount of mRNAs . The primers used for qRT-PCR are listed in Table S1 . Fifteen 3- to 6-day-old flies were placed in an empty fly food vial . The climbing index is the fraction of flies that climb halfway up the vials in 10 s after being tapped down to the bottom of the tube . We performed each experiment twice and used the average of the two trials to calculate the climbing index . Data shown are the mean ± SEM . To compare two sets of data , unpaired Student's t-tests were used . ANOVA with the Tukey post-hoc test was used to compare multiple sets of data . Asterisks indicate statistical significance .
|
Tubby is a member of the Tubby-like protein ( TULP ) family . Tubby mutations in mice ( tubby mice ) cause late-onset obesity and neurosensory deficits such as retinal degeneration and hearing loss . However , the exact molecular mechanism of Tubby has not been determined . Here we show that Drosophila Tubby homolog , King tubby ( dTULP ) , regulates ciliary localization of transient receptor potential protein ( TRP ) . dTULP-deficient flies showed uncoordinated movement and complete loss of sound-evoked action potentials . dTULP was localized in the cilia of chordotonal neurons of Johnston's organ . Two TRP channels essential for auditory transduction , Inactive and NompC , were mislocalized in the cilia of chordotonal neurons in the dTulp mutants , indicating that dTULP is required for proper cilia membrane protein localization . This is the first demonstration that dTULP regulates TRP channel localization in cilia , and thus provides novel insights into the pathogenic mechanism of tubby mice .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
dTULP, the Drosophila melanogaster Homolog of Tubby, Regulates Transient Receptor Potential Channel Localization in Cilia
|
A major component of ex vivo amyloid plaques of patients with dialysis-related amyloidosis ( DRA ) is a cleaved variant of β2-microglobulin ( ΔN6 ) lacking the first six N-terminal residues . Here we perform a computational study on ΔN6 , which provides clues to understand the amyloidogenicity of the full-length β2-microglobulin . Contrary to the wild-type form , ΔN6 is able to efficiently nucleate fibrillogenesis in vitro at physiological pH . This behavior is enhanced by a mild acidification of the medium such as that occurring in the synovial fluid of DRA patients . Results reported in this work , based on molecular simulations , indicate that deletion of the N-terminal hexapeptide triggers the formation of an intermediate state for folding and aggregation with an unstructured strand A and a native-like core . Strand A plays a pivotal role in aggregation by acting as a sticky hook in dimer assembly . This study further predicts that the detachment of strand A from the core is maximized at pH 6 . 2 resulting into higher aggregation efficiency . The structural mapping of the dimerization interface suggests that Tyr10 , His13 , Phe30 and His84 are hot-spot residues in ΔN6 amyloidogenesis .
β2-microglobulin ( β2m ) is a 99-residue protein with a typical immunoglobulin fold comprising seven anti-parallel β-strands stabilized by a disulfide bridge ( Fig . 1 ) [1] . Upon dissociation from the MHC-I heavy chain , human β2m ( Hβ2m ) is catabolised in the kidneys . In individuals undergoing long-term hemodialysis the clearance process is strongly impaired and the levels of Hβ2m in the serum can increase up to 60-fold [2] . The progressive accumulation of Hβ2m in the osteoarticular system , presumably driven by its affinity for type-I collagen [3] , eventually leads to amyloid assembly and the onset of dialysis-related amyloidosis ( DRA ) , a pathological condition characterized by tissue erosion and destruction [4] . The wild-type Hβ2m ( WT Hβ2m ) does not form amyloid fibrils in vitro in the absence of ex vivo amyloid seeds [5] , or additional factors such as Cu2+ [6] , [7] or TFE [8] . This limitation makes the determination of the aggregation mechanism of Hβ2m in physiological conditions ( 37°C , pH 7 . 5 ) a particularly challenging conundrum . A major contribution towards its solution was the identification [9] , [10] , and atomic-level structural characterization [11] , of an intermediate state ( representing from 3 . 7±1 . 4% [10] up to ∼14±8% [9] of the equilibrium population ) in the folding pathway of WT Hβ2m . The intermediate was termed IT because its main structural trait is a non-native trans isomerization of Pro32 . Enhanced fibrillogenesis in physiological conditions ( including the ability to elongate and/or nucleate amyloid fibril assembly ) has been observed in connection with an increase in the equilibrium concentration of IT [10] , [12] , [13] , indicating that IT is highly amyloidogenic and a key player in Hβ2m fibrilogenesis . While relevance of IT for β2m fibrillogenesis is widely acknowledged [9] , [10] , [14]–[16] , alternative intermediate states , which are less native-like than IT , become relevant under different experimental conditions [8] , [17]–[19] . Furthermore , a variety of environmental aspects have been found to directly affect the process of fibril formation by β2m including solubility , supersaturation and ultrasonication/agitation effects [20] . Recently , the single point mutant Asp76Asn ( D76N ) , a naturally occurring variant of Hβ2m , was associated with the late onset of a fatal hereditary systemic amyloidosis characterized by extensive visceral amyloid deposits . However , and contrary to what occurs in DRA , in this newly discovered disease the plasma concentration of Hβ2m is not augmented [21] , [22] . In vitro studies have shown that the Asp76Asn mutant is highly amyloidogenic , displaying an abundant ( ∼25% ) equilibrium population of IT under physiological conditions [22] . Another recently reported single point variant of β2m for which fibrillation occurs without seeding under physiological conditions is the Arg3Ala ( R3A ) mutant [23] . The latter , however , has not yet been associated with any conformational disorder . In this work we focus on ΔN6 , a truncated form of Hβ2m , lacking the first six N-terminal residues . This variant is potentially relevant because it represents ∼30% of ex vivo amyloid deposits extracted from DRA patients [24] . Radford and co-workers proposed that ΔN6 is a structural mimic of IT because it populates a conformational state that reproduces the conformational features of IT and represents 90% of ΔN6's in vitro equilibrium population [11] . While there is a broad agreement regarding the ability of ΔN6 to prime the fibrillar conversion of WT Hβ2m in vitro under physiological conditions , the mechanism by which it occurs is not consensual . In particular , Eichner and Radford proposed that monomeric ΔN6 conformationally converts WT Hβ2m into an amyloidogenic state in a mechanism akin to prion conversion [11] , while Bellotti and coworkers challenged the prion-like hypothesis by reporting that the WT Hβ2m does not fibrillate with monomeric ΔN6 but rather with preassembled fibrils of ΔN6 [22] . However , and independently of these controversies , it is widely accepted that ΔN6 alone is able to efficiently nucleate fibrillogenesis in physiological conditions ( tlag∼35 days , 80 mM ) [11] , [12] , [25] . Furthermore , it displays an enhanced amyloidogenicity at pH 6 . 2 ( tlag∼15 days , 80 mM ) [11] , i . e . , in conditions compatible with the mildly-acidic character of the synovial fluid of DRA patients [26] . It has been suggested that the aggregation potential of ΔN6 stems from its unique ability to populate one or more aggregation-prone intermediate states [11] . Therefore , a complete picture of the aggregation mechanism of ΔN6 requires disclosing the process according to which it aggregates de novo starting from the self-association of aggregation-prone monomeric states . Addressing this challenge via molecular simulation is the major goal of the present work . By studying the early stage of aggregation of ΔN6 one expects to get insights into the amyloidogenicity of the full-length protein . The large size of the system and the long timescales involved in the process of protein aggregation strongly restrict the use of classical molecular dynamics ( e . g . based on the AMBER or GROMOS force fields ) to explore it . For this reason researchers have been developing coarse-grained approaches to study protein aggregation [27] . One example is the symmetrised Gō potential used to study domain-swapping ( DS ) [28] . In DS two monomers exchange identical structural elements or “domains” to form a strongly bound dimer . Since the DS hypothesis is based on the association of two monomers into dimers , the mechanism of fibril formation from more than two identical proteins is still unclear . On the other hand since DS is a manifestation of concomitant folding and binding it requires the use of a simulation framework where the two processes compete directly with each other via a force field that accounts for competing intra- and intermolecular interactions [28] . Hβ2m fibrillogenesis has been reported to be initiated by dimerization of monomers [6] , [12] , [16] , [29]–[31] including DS [32] . Here , however , we will not study the dimerization of Hβ2m that may result from DS . Instead , our goal in this study is to explore the early stage of the aggregation mechanism of the truncated variant ΔN6 that may occur as a side-effect of protein folding . More precisely , if aggregation-prone intermediate states ( including highly native-like species ) are populated along the folding pathway of ΔN6 , they may start interacting with each other ( e . g . via solvent-exposed hydrophobic residues ) thus triggering the amyloid cascade . These aggregation-prone intermediates are a by-product of the folding process and likewise their formation is exclusively driven by intra-molecular interactions . Inter-molecular interactions will only start operating once the monomers representative of the intermediate state get within interaction range , which may eventually lead to their self-association into dimers . Our study seeks to explore this type of ( de novo ) aggregation route for the variant ΔN6 of Hβ2m by highlighting its topological aspects , i . e . , the predictions reported here are strictly structure-based . In doing so , we use a three-stage computational protocol based on an array of tools as detailed in the Methods section . In the first stage , following our previous studies [33] , [34] , replica-exchange discrete molecular dynamics ( RE-DMD ) simulations of a full atomistic Gō model [35] are used to study the folding transition and to identify intermediate states in the folding pathway of ΔN6 . The adopted level of structural resolution encompasses the effect on the folding mechanism of detailed atomic native contacts and fully takes into account side-chain packing , a fundamental ingredient of the folding process . Combined with RE-DMD simulations , the full atomistic Gō model enables equilibrium sampling of the conformational space , a task far beyond the possibilities of routinely used classical molecular dynamics protocols , especially for a system of the size of Hβ2m . While this simulation procedure captures the fundamental features of the folding process [36] , it fails to include others . In particular , it neglects the effects of the pH , an important environmental parameter . Indeed , it is well known that changes in pH can induce conformational changes of varying degree , ranging from structural fluctuations to modifications in secondary structure content [37]–[39] . Furthermore , in the case of ΔN6 the pH turns out to be a particularly relevant parameter because – as stated before – the aggregation potential of this variant is remarkably sensitive to pH changes [11] . To identify the molecular roots of this dependence , it is thus crucial to establish how the pH affects the structure of the relevant conformational states , because structural changes at the monomer level ( e . g . the reorganization of aromatic side chains , which are bulky and therefore natural players in the establishment of intermolecular interactions ) will directly affect monomer association and ultimately dictate aggregation performance . In the second stage of our computational protocol , we investigate how the pH modulates an intermediate state's structure . Leaving aside its possible effect on large-scale conformational dynamics , we can afford to accurately capture the effect of pH by employing constant pH molecular dynamics ( CpHMD ) with explicit titration [37] , [38] , [40]–[43] . In doing so one also obtains conformations representative of the intermediate state with the most accurate representation of side-chain and backbone geometries , which is a requirement of the Monte Carlo ensemble docking ( MC-ED ) [33] protocol whose predictions depend critically on the structural accuracy of the analysed structures . The MC-ED is a low-resolution protocol that highlights the role of shape complementarity , a major driver of protein aggregation [44] , [45] . It takes pairs of monomer conformations obtained with CpHMD to generate two ensembles of putative dimer structures ( one obtained from monomers at equilibrated pH 6 . 2 and another at pH 7 . 2 ) where the number of residue pairs within interaction distance is maximized and the number of excluded volume interactions is minimized . The number of contacts thus evaluated provides a measure of the quality of the geometric matching between the two monomers . Therefore the MC-ED method allows predicting the residues that are most likely to trigger dimerization in an ensemble of dimers whose interface was optimized for shape complementarity . The MC-ED allows analyzing the association of an exceedingly large number of monomeric conformations while discriminating between the dimer structures that are prone to further oligomerize from those that are not . This is an important point because protein conformations are not static entities . Indeed , they have a dynamic nature leading to structural variability even for the native state . Thus , two pairs of self-associating conformations ( representing the same conformational state ) will not form exactly the same docking interface upon dimerization . By exhaustively docking thousands of equilibrated conformations collected from CpHMD simulations at pH 6 . 2 and 7 . 2 this work provides a probabilistic structurally resolved picture of the dimerization interface of the identified intermediate state , the native state of ΔN6 and also the native state of WT Hβ2m at physiological and near physiological pH . In doing so , it recapitulates and rationalizes previous experimental observations , and draws new insights into the aggregation mechanism of ΔN6 , including the prediction of aggregation hot spots .
The free energy ( FE ) surfaces at the folding temperature ( Tf ) , evaluated with the WHAM method [46] , reveal a well-defined intermediate basin for ΔN6 that is not present in the FE surfaces of Hβ2m ( Fig . 2A ) . The native and intermediate states populate ∼38% and 11% of the equilibrium ensemble at Tf , respectively . To isolate and structurally characterize the intermediate state populated by ΔN6 , which we term ΔN6-I , we performed structural clustering over an ensemble of conformations collected from DMD simulations at fixed temperature ( ∼Tf ) . The intermediate species preserves the trans-isomerization of Pro32 ( as a consequence of the native-centric character of the Gō potential ) and exhibits an unstructured/disordered strand A detached from a fairly conserved core region comprising residues 21 to 94 ( i . e . strands B–G and connecting loops ) ( Fig . 2B ) . The detachment of strand A from the protein core and its structureless nature are likely the result of a smaller number of native interactions involving this secondary structural element , which decreases by 27% with regard to that observed in the full length species ( Fig . S5 ) . The evaluation of solvent accessible surface area ( SASA ) per residue reveals that 62% of the hydrophobic core residues become highly solvent-exposed in ΔN6-I with SASA exhibiting a 3- up to 7-fold increase in Leu7 , Val9 , Leu23 and Trp95 , all located at the termini . Phe30 on the BC-loop , and Ile35 in strand C are also significantly more exposed to the solvent in the intermediate state ( Fig . 2C ) . These observations are particularly relevant because the exposure of aggregation-prone hydrophobic patches has been pointed out as a hallmark of protein aggregation ( reviewed in [47] ) and suggest that the identified intermediate state has a high aggregation potential . We conjecture that strand A , by being exposed to the solvent , will be a particularly important structural motif for the early aggregation stage of ΔN6 . The Gō potential adopted in this work does not predict an equilibrium population of a similar full-length species , with a conserved core region and detached strand A , across the folding transition of the WT variant . However , there is experimental evidence that the amyloid-transition of the full-length Hβ2m is concomitant with a detachment of the N-terminal strand A [48] triggered by an acidic pH [19] , [49]–[51] or Cu2+ binding [7] , [48] , [52] . Therefore it is likely that the full-length Hβ2m may undergo a similar conformational transition . The ΔN6-I intermediate state identified with the Gō model highlights the importance of native topology in determining the folding space . In order to investigate how the pH modulates the structure of the intermediate state , and , in particular , the degree of solvent-exposure of the unstructured strand A , we set up a series of CpHMD simulations that used the intermediate conformation as a topological template . This means that a structurally refined version of ΔN6-I was prepared by taking into account the structural information provided by the native-centric model . The refined structure was then used as the starting conformation in CpHMD simulations at pH 7 . 2 and pH 6 . 2 . The analysis of conformational ensembles taken from the equilibrated parts of CpHMD trajectories reveals that pH 6 . 2 has a striking effect on the region comprising strand A and the AB-loop . In ΔN6-I this region deviates significantly from its original position in the native structure as indicated by the large ( mean ) RMSD ( 16 Å ) obtained after optimally superimposing each analyzed intermediate conformation over the native core region . The recorded RMSD at pH 6 . 2 represents an increase of 20% from that observed at pH 7 . 2 indicating a distinctively higher degree of solvent exposure at lower pH . On the other hand , the core region is better preserved at pH 6 . 2 with the ( mean ) RMSD decreasing up to 40% relative to pH 7 . 2 ( Table S1 ) . The increased solvent-exposure of the N-terminal region ( comprising residues 6 to 20 ) at pH 6 . 2 can be tentatively explained on the basis of a more favorable electrostatic contribution to the free energy of solvation at this pH . At a physiological/near-physiological pH Arg12 and Lys19 are permanently protonated , Glu16 is mostly deprotonated ( only ∼0 . 3% protonation at pH 6 . 2 ) , but the protonation state of His13 changes given the similarity between the medium pH and the average pKa of its imidazole ring ( 6 . 0 ) . Our data hints at the possibility of a direct connection between His13's protonation state and the SASA of the N-terminal region ( comprising both strand A and the AB-loop ) at pH 6 . 2 ( Fig . S1 ) . Therefore , one can argue that the higher degree of protonation of His13 at pH 6 . 2 leads to an increased solvent-exposure of that region which is concomitant with a favorably-enhanced electrostatic contribution to its free energy of solvation ( it has been shown that protonation of the imidazole side chain produces a substantial increase of that histidine's absolute solvation free energy [53] ) . Seminal studies carried out by Miranker and co-workers have emphasized the importance of dimerization in the aggregation pathway of a mutational variant of Hβ2m ( P32A ) [6] . The latter is a structural mimic of an intermediate state ( M* ) , which shares with IT ( and , therefore , with ΔN6 ) important structural features . Of note , a trans peptidyl-prolyl His31-Ala32 bond and the re-packing of several aromatic side chains within the hydrophobic core including Phe30 and Leu62 . Furthermore , Eichner and Radford have shown that the M* intermediate elutes at a retention volume identical to that of IT [12] . More recently , studies carried out by Radford and co-workers on the mutational variants P5G and ΔN6 ( for which the intermediate IT represents respectively ∼60% and 90% of the equilibrium ensemble in physiological conditions ) showed that the first assembled oligomeric state in the amyloid pathway of both mutants is a dimer of IT monomers [12] . Motivated by these findings , we carried out an exhaustive study of the dimerization interface in ΔN6 via MC-ED simulations . We mapped the dimerization interfaces of the intermediate state ( ΔN6-I ) and of the native state ( ΔN6-N ) at pH 7 . 2 and 6 . 2 by docking pairs of conformations obtained from CpHMD simulations under those pH conditions . We also investigated the dimerization interface of the native state of WT Hβ2m ( WT-N ) as a control experiment . In Figure 3 we report density histograms ( DH ) of the number of intermolecular contacts at pH 7 . 2 ( Fig . 3A ) and pH 6 . 2 ( Fig . 3B ) for the dimers of the analyzed species . This property provides a quantitative measure of the quality of the geometric matching between the two monomers because each dimer conformation was optimized for maximum number of interactions and minimum number of excluded volume interactions . Each dimer interface was thus optimized for shape complementarity , a property that is considered a major driver of protein-protein association [44] , [45] . In this sense the density histograms provide insight regarding the dimerization potential of each species . In the DHs the vertical lines indicate the mean , and the mode corresponds to the highest point of the distribution ( representing the most probable number of intermolecular interactions in the population of dimers ) . In order to facilitate the comparison of these data , Figures 3C–E separately report the DHs for the two considered pH values . The analysis of the DHs reveals important findings . First , the high similarity between the curves obtained for the WT-N species suggests that it should conserve its dimerization propensity upon changing the pH from 7 . 2 to 6 . 2 ( Fig . 3C ) . Since WT-N is the most populated state [9] , [10] , [15] of the in vitro equilibrium population in physiological conditions , our observation is consistent with the conservation of the aggregation behavior of Hβ2m across this pH range [25] . Our analysis further suggests that at the molecular level this behavior may be rooted in the conformational robustness exhibited by the monomeric form of WT-N across the analyzed pHs ( Table S2 ) , and , in particular , points out the importance of the protective role played by the N-terminus in maintaining the hydrophobic balance that stabilizes the native state [25] . Second , ΔN6-I forms dimers with number of intermolecular contacts given by the mode ( mean ) with a probability that is up to ∼52% ( 60% ) higher than in the WT-N at pH 6 . 2 ( for ΔN6-N this probability goes up to 51% at pH 7 . 2 ) ( Fig . 3D and Fig . 3E ) . This is consistent with the higher amyloidogenicity of ΔN6 at physiological/near physiological pH . Third , when the pH decreases from 7 . 2 to 6 . 2 , the mean and the mode decrease marginally for both ΔN6 conformational states . However , in the case of ΔN6-I , this mild decrease goes in tandem with a significant increase ( up to 10% ) in the probability of formation of the corresponding dimers ( Fig . 3E ) . On the other hand , for ΔN6-N , the dimer conformations representative of the mean and mode are less probable at pH 6 . 2 than at pH 7 . 2 ( Fig . 3D ) . Since it is likely that further oligomerization will be limited by nucleation of dimers , both measures ( i . e . mean and mode ) predict that ΔN6-I plays a major role in amyloid formation at pH 6 . 2 . In order to pinpoint the regions of the protein that are most likely to start dimerization , we have constructed probability contact maps for the dimer interfaces ( Fig . S2 ) . The probability of each intermolecular contact was evaluated by counting the number of times the contact is present in the ensemble of dimers that was used to determine the DH . We have also analyzed several representative dimer conformations ( i . e . conformations with number of intermolecular contacts corresponding to the mode and tail of the DHs ) , to gauge their importance for further oligomerization ( Fig . 4 and Fig . S3 ) . We have chosen to analyze the structure of ‘mode’ dimers for consistency reasons , i . e . , because they exhibit the most likely number of intermolecular contacts ( and a minimal number of excluded volume interactions ) in the dimer interface , a property that quantifies the degree of geometric matching and shape complementarity of the interfaces in the ensemble of MC-ED generated dimers . On the other hand , the analysis of ‘tail’ dimers is particularly pertinent for the WT-N species because a unique feature of its DH is a rather extended tail indicating the formation of dimers with the strongest geometric matching ( Fig . 3C ) . Since shape complementarity is a major driver of protein aggregation and is maximized for ‘tail’ dimers it is important to establish if/how the existence ‘tail’ dimers may affect the aggregation performance of Hβ2m . We find that at both pH values dimerization of WT-N is majorly driven by the DE-loop ( especially residues 56–60 ) ( Fig . S2A ) . The analysis of several dimer conformations representative of the mode of the DH reveals that the most likely dimerization interface involves the DE-loop of one monomer that associates with the second monomer in several possible spots ( Fig . S3A ) . On the other hand , the DE-loop directed interfaces of the strongly packed dimers are more specific , being based on loop-loop interactions involving the BC and DE aromatic-rich regions ( Fig . S3B ) . Since the latter become unavailable for subsequent interaction , further oligomerization ( via addition of another monomer ) appears to be restricted to the potentially adhesive residues located on the EF-loop ( e . g . Phe70 and Tyr78 ) and in the C-terminus ( Trp95 ) . Our analysis therefore predicts that the recruitment of the aromatic-rich regions in the WT-N best geometrically matched dimers' interfaces renders these dimeric entities particularly soluble thus lowering their aggregation potential ( soluble dimerization was recently found to be a possible dead-end for aggregation in Ref . [54] ) . In order to gauge the importance of residue 76 ( located in the EF-loop ) for the dimerization of Hβ2m we have selected a representative mode dimer of WT Hβ2m and used the program SCAP ( included in the Jackal package [55] ) to replace ( in each monomer ) the original amino acid Asp by an Asn thus mimicking the mutation that occurs in the systemic amyloidosis characterized by extensive visceral amyloid deposits . We then computed the electrostatic potential at the interface of both dimers ( i . e . with and without the mutation ) ( Fig . S4 ) . Our results are consistent with the enhanced amyloidogenicity observed in vitro for Asp76Asn with regard to Hβ2m [21] , [22] because they indicate that the mutation contributes to stabilize the dimer by decreasing the amount of repulsive electrostatic interactions between the EF-loop of one monomer with the DE-loop of the second monomer . In the case of ΔN6-N the pH has a modulating effect on the dimerization interface . First , the DE-loop is no longer the major player in dimerization , as both the AB- and BC-loops gain significant importance ( Fig . S2B; Fig . 4C and Fig . 4E ) . The most important structural element for dimerization in ΔN6-I at pH 6 . 2 is the unstructured and detached strand A together with the adjoining AB-loop ( Fig . S2C ) . Furthermore , inspection of several dimer conformations with number of intermolecular contacts equal to the mode ( Fig . 4B and Fig . 4D; Fig . S3C and Fig . S3D ) reveals that strand A facilitates fibril growth by imposing a rather straightforward oligomerization pattern . Indeed , strand A acts as a sticky ‘hook’ that recruits another monomer by interacting with its DE- , EF- or FG-loops ( Fig . 4B and Fig . 4D; Fig . S3C and Fig . S3D ) thereby leaving the second monomer's strand-A available for further growth . These ‘sticky hook’ interactions driven by strand A clearly drive a preferred oligomerization direction that could coincide with that of the amyloid fibril axis ( Fig . 4I ) . Whenever monomer association involves strand A-strand A interactions , the resulting dimers can still grow via the BC- and DE-loops ( Fig . 4F; Fig . S3D ) . In order to identify putative hot spots for aggregation , we computed the probability of intermolecular interaction per residue in the subset of the 50 most frequent intermolecular interactions . The latter were identified by taking the ensemble of dimers used in the evaluation of the corresponding DH . Pairs involving an aromatic amino acid and His or Lys dominate in ΔN6-I dimers at pH 6 . 2 . In WT-N ( Fig . 5A ) the distinctive predominance of interactions involving the DE-loop illustrates this region's importance for dimerization . The relevance of the DE-loop in different experimental conditions has been acknowledged by several authors [11] , [56] , [57] , including a recent study by Eisenberg and co-workers which reported a hinge motif in dimers of Hβ2m based on DE-loop swapping at pH 8 ( in the presence of DTT ) [32] and another study by Rennella et al . which reported the formation of Hβ2m dimers with a head-to-head arrangement of monomers driven by DE-loop interactions [16] . The aromatic residues Phe56 , Trp60 ( located on the DE loop's tip ) , Phe62 , Tyr63 and the aliphatic Leu65 are expected dimerization spots because they assist the docking of Hβ2m onto the MHC-I heavy chain [2] . Phe62 , Tyr63 and Leu65 were further shown to play an important role in fibril nucleation at acidic pH 2 . 5 [58] . The importance of Phe56 and Trp60 in β2m oligomer assembly based on D-D strand association ( pH∼7 ) was reported in several studies [7] , [57] . Of note , Trp60 was found to be the residue involved in the largest number of intermolecular contacts in Molecular Dynamics simulations that studied intermolecular interactions establishing between monomers of β2m [59] . The results reported here recapitulate that , with the exception of Leu65 , DE-loop aromatic residues are important drivers of monomer association in Hβ2m . We further find that lowering the pH reduces the importance of residues Phe56 , Lys58 and Phe62 , while Trp60 becomes a particularly promiscuous interaction hub at pH 6 . 2 ( Fig . 5A ) . However , the analysis of dimer's conformations whose formation is triggered by this region of the protein indicates that further oligomerization is not straightforward ( Fig . S3A and Fig . S3B ) . In other words , our results suggest that while the DE-loop is certainly important for dimerization , the amyloid route that is triggered by this structural element is not the most efficient one . Since the native state is the dominant conformational state in physiological ( and near physiological ) pH , this observation rationalizes the little amyloidogenic character of WT Hβ2m in those conditions . Also , in line with this idea , we find that in ΔN6 , which is considerably more amyloidogenic than Hβ2m , the importance of Trp60 ( and nearby residues ) is substantially reduced , especially in ΔN6-I at pH 6 . 2 . This observation is particularly relevant because at this pH the cleaved mutant is more amyloidogenic . In ΔN6-N ( Fig . 5B ) , the region comprising the AB-loop ( residues 10 to 20 ) exhibits an increased probability to form intermolecular contacts at both pH values . Due to increased solvent exposure of the BC-loop ( residues 28 to 34 ) ( Fig . 2C ) , there is also a significant enhancement of the participation of Phe30 and His31 ( especially at pH 6 . 2 ) . Direct involvement of the BC-loop in ΔN6 oligomer assembly ( in physiological and near physiological pH ) was reported in Ref . [11] . Furthermore , His31 was found to be a major contributor to the intermolecular contacts established in association interfaces within hexamers of H13F [7] and tetramers of DCIM50 , the E50C Hβ2m mutant disulfide-linked homodimer [57] , in physiological conditions ( in the presence of Cu2+ or 20% TFE and fibril seeds , respectively ) . It was also observed to be a component of the non-covalent interface between two ΔN6 nanobody-trapped domain-swapped dimers ( pH 5 . 0 ) in the respective crystal asymmetric unit along with the aromatic residues Phe56 and Trp60 [31] . The region comprising the end of strand F and the FG-loop ( residues 84 to 90 ) – which is not involved in dimerization of Hβ2m – becomes especially relevant for ΔN6-N ( pH 7 . 2 ) and ΔN6-I ( at both pHs ) ( Fig . 5C ) . Interestingly , the stretch of amino acids 83N-89Q was implicated in the nanobody-driven domain-swapping aggregation of ΔN6 [31] and was shown to fibrillate into amyloid in a highly acidic pH 2 . 0 [60] . At neutral pH , both the BC- and DE-loops of ΔN6-N and ΔN6-I deviate significantly from their native positions ( Table S1 and Table S3 ) . The cleaved N-terminus , which is more detached from the core in ΔN6-I , facilitates such conformational migration . Consequently , His84 located in the FG-loop ( adjoined to the BC-loop ) , is more solvent exposed in ΔN6-N and ΔN6-I than in WT-N thus becoming an important interaction hub in the mutant's dimers ( Fig . 5B and Fig . 5C ) . In the ΔN6-N dimer interfaces , His84 preferentially interacts with Phe56 and Trp60 at pH 7 . 2 , while interaction with Tyr10 becomes relevant in ΔN6-I , especially at pH 6 . 2 ( see next section for further details ) . The low amyloidogenicity of the Hβ2m mutational variant ΔN6/H84A in physiological or near-physiological pH [11] may thus reflect the absence of relevant interactions involving His84 . Taken together , these findings thus point out to a direct participation of His84 in Hβ2m association , which adds up to a proposed indirect effect according to which His84 helps maintaining the trans-isomerization of Pro32 thus enhancing the population of IT [61] . In ΔN6-I ( Fig . 5C ) residues Tyr10 and His13 , located in strand A and start of the AB-loop , gain importance especially at pH 6 . 2 . Previous studies reported the participation of these residues in association interfaces within hexamers of H13F [7] and tetramers of DCIM50 [57] . Furthermore , a single Tyr residue can act as the sole driving force triggering self-aggregation of a short polyalanine peptide ( through cation - π and π-stacking interactions ) [62] . His84 and Phe30 maintain their relevance for dimerization at both pHs . There is , however , a noticeable increase in the importance of Trp60 at pH 7 . 2 ( in comparison with ΔN6-N ) . This happens because there is a larger migration of the DE-loop from the core region facilitating the participation of Trp60 in dimerization ( Table S1 ) . At pH 6 . 2 the EF-loop ( residues 71–77 ) , especially residues Glu74 and Lys75 , also gains importance in dimer association ( the EF-loop is not involved in dimerization in the WT-N ) . Overall , the most important feature of ΔN6-I dimerization is the striking increase in importance of strand A relative to WT-N . Here we identify the interaction partners of the predicted dimerization hot spots ( Fig . 6 ) and pinpoint specific interactions ( e . g . aromatic π-stacking , cation-π , and hydrophobic ) that may contribute to efficiently stabilize the dimerization interface in vitro and in vivo ( see Table S4 and description therein of the 50 most frequent intermolecular contacts in ΔN6-I dimers at pH 6 . 2 ) . Indeed , while the force field used in the MC-ED simulations does not explicitly take into account specific types of interactions , it is reasonable to determine if the predicted ( structured-based ) dimers meet the geometric requirements for the occurrence of such interactions ( e . g . Cation-π interactions require that at least one of the atoms of the aromatic ring is located no further than ∼4 . 5 Å from one of the atoms carrying the net or partial positive charge – in His the positive charge can be located in the atoms Nδ1 , Nε2 , or Cε1 of the imidazole ring . In the present Gō model the maximum contact distance is ∼4 . 7 Å . Therefore , every contact between one His or Arg or Lys and one aromatic residue is within the Cation-π interaction distance ) .
In recent years the identification and structural characterization of intermediate states for folding and aggregation [33] , [69] has greatly contributed to a better understanding of the relation between the folding and aggregation landscapes [70] . The identification of these states in association with proteins of medical interest is of paramount importance . Indeed , not only it contributes to solve their aggregation mechanism but it also strengthens the need of including protein homeostasis as a therapeutic target for conformational diseases [71] . The work reported here illustrates how the combination of computational methods with different levels of resolution provides a unique opportunity to analyze the aggregation pathway and formulate testable predictions thus contributing to clarify the relation between folding and aggregation . This study focused on the truncated mutant ΔN6 of protein Hβ2m and its dimerization mechanism . While the results reported here help gaining insights into the fibrillogenesis mechanism of the parent species , they do not entail an exclusive role of the truncated species in the actual fibrillogenesis pathway of the full-length protein . Our study predicts the existence of an intermediate state for folding and aggregation in ΔN6 . The intermediate preserves the trans-isomerization of Pro32 that characterizes IT and a new structural trait: an unstructured strand A that detaches significantly from a fairly conserved core region comprising residues 21 to 94 . The new intermediate state identified here represents a conformational excursion of the native state extending the loss of native structure already detected in the amyloidogenic intermediate IT [10] , [11] . The association of an unstructured/detached strand A with the onset of fibrillogenesis in β2-microglobulin was originally proposed by Verdone and co-workers [48] , and subsequent studies have linked this structural trait with acidic pH [19] , [49]–[51] or Cu2+ binding [7] , [48] , [52] . Therefore , it is likely that a similar conformational pathway may occur also with the full-length protein despite remaining undetected in the simulations carried out in this study . That this should the case is in fact demonstrated by the fibrillogenesis of the mutants D76N and R3A at neutral pH and without seeding . A lack of structure in one or both termini is a common feature shared by intermediates states that link the folding and aggregation landscapes [33] , [69] , [72] . Results reported here indicate that ΔN6 dimerizes with higher probability than WT Hβ2m , in line with its higher in vitro amyloidogenic potential and further predict that at pH 6 . 2 the intermediate state ΔN6-I identified in this work becomes the key player in ΔN6 dimerization . We find that the region comprising strand A and the AB-loop is critical for dimerization ( especially at pH 6 . 2 ) and , presumably , to further oligomerization as well . Eichner and Radford reported a set of resonances in strand A and in the AB-loop of ΔN6 that shifted significantly at physiological pH depending on protein concentration , which is consistent with their involvement in aggregation . Interestingly , most of the chemical shifts of strand A are not defined because the residues resonate in a crowded region of the spectrum [11] . This may be taken as an indication of conformational liability for this part of the protein , making the NMR characterization of the proposed intermediate state a particularly challenging task . The ΔN6 ( -N and I ) dimers depicted in this work provide direct access to the atomic-level details associated to the participation of the N-terminal and BC-loop regions in Hβ2m oligomer assembly . In particular , our results reinforce the importance of the direct involvement of both regions in oligomerization which has been previously observed for several Hβ2m mutational variants usually in association with an enhanced amyloidogenicity [7] , [11] , [57] . The study of the dimerization interface we carried out for the WT form also recapitulates previous experimental findings . Namely , they reinforce the relevance of the DE-loop aromatic amino acids as important drivers of monomer association in Hβ2m . We found that monomer association driven by this region of the protein results into dimers of ( WT ) Hβ2m with a head-to-head arrangement of monomers that is similar to what is observed by other authors [16] , [59] . The current work establishes that this ( WT ) Hβ2m mode of monomer association is such that further oligomerization is not straightforward . This is in agreement with the reported limited oligomerization of the native WT in physiological conditions [12] , [56] . The comparative study of the dimerization interfaces we carried out for ΔN6 and WT Hβ2m allowed the prediction of putative dimerization hot spots in the truncated form . Residues Phe30 , Phe56 , Trp60 and Trp95 are universally important interaction hubs across the three species and pHs . They are able to establish a myriad of associations via their aromatic side-chains ranging from hydrophobic to the more stabilizing cation-π and π-stacking interactions . Trp60 is always highly promiscuous , Phe30 becomes distinctively important for ΔN6-N and ΔN6-I at pH 6 . 2 , with Trp95 assuming a more important role in WT-N . Finally , our results highlight the involvement of His84 in important interactions within ΔN6-N and ΔN6-I dimers therefore contributing to rationalize the low amyloidogenicity observed in vitro ( at physiological and near-physiological pH ) for the Hβ2m mutational variant ΔN6/H84A [11] .
We consider a full atom representation where each non-hydrogen atom is taken as a hard sphere of unit mass . The atom's size is defined by scaling the relevant van der Waals ( vdW ) radius by a factor α<1 [73] . Protein energetics is given by excluded volume interactions ( which forbid hard-core clashes ) , bonded interactions , and non-bonded ( or contact ) interactions , all of which are all modeled by discontinuous , piecewise constant interaction potentials . Contact interactions are represented by a square-well potential whose depth is given by Gō energetic [35] . Thus , if atoms i and j are located in residues which are separated by more than two units of backbone distance the interaction parameter between them , εij , is given by ( 1 ) In the expression above σ = α ( r0i+r0j ) is the hard-core distance , r0i is the vdW radius of atom i , λ is a scaling factor that controls the range of attractive interactions , and Δij = −1ε ( where ε is the energy unit ) if i and j are in contact in the native conformation and is 0 otherwise . We followed Ref . [74] in treating the energetics of the disulfide bond in the same manner as we treat the other contact interactions . We set α = 0 . 80 and λ = 1 . 6 in order to have a well-behaved folding transition [73] , [74] . This choice of parameters sets a cut-off distance of 4 . 7 Å ( for methyl carbon ) , and leads to 957 native contacts in Hβ2m ( PDB ID: 2XKS ) and 899 native contacts in the ΔN6 mutant ( PDB ID: 2XKU ) . The native contacts are distributed within the elements of secondary structure as reported in Fig . S5 . The total energy of a conformation is computed as the sum over all atom pairs , ( 2 ) Further information about the adopted model can be found in Refs . [33] , [34] , [75] . Temperature is measured in units of ε/kB . The folding transition is explored with a discrete ( or discontinuous ) Molecular Dynamics ( DMD ) engine [76] and correct equilibrium sampling is achieved by using a standard replica-exchange ( RE ) Monte Carlo method [77] with a temperature grid that was calibrated to ensure a high acceptance probability ( >90% ) for the RE moves and replica ‘round-trips’ ( i . e . moving from the top to the bottom of the temperature grid and back ) with a mean cycle time of ∼50 RE moves . The equilibrated part of each simulation consisted of ∼5×1010 events per replica , and was used to collect uncorrelated data for the thermodynamic calculations . The folding ( or melting ) transition Tf is usually estimated as the temperature at which the heat capacity Cv attains its maximum value . Here , the Cv is computed from the mean squared fluctuations in energy at each temperature considered in the RE simulations , in accordance with the definition . To compute the free energy as a function of different reaction coordinates ( E , Rg , RMSD ) we have used the weighted histogram analysis method ( WHAM ) [46] . In order to isolate and structurally characterize the intermediate state populated by the ΔN6 truncated variant we started by running extensively long ( up to 2 . 4×1011 events ) DMD simulations at fixed temperature T ( with T located within the transition region ) . A total number of three trajectories were considered . For each trajectory , a conformational ensemble ( with up to 30k elements ) was constructed by picking up equilibrated conformations ( i . e . conformations sampled beyond the first folding transition ) . Subsequently , each conformational ensemble was analyzed with the k-means clustering algorithm of Brooks and co-workers as implemented in the MMTSB toolset [78] . The clustering radius cutoff was set to 9 Å ( whenever the trajectories sampled both the native and intermediate basins ) or 5–6 Å ( if only the native basin was sampled ) . We performed CpHMD simulations at pH 7 . 2 and 6 . 2 . The simulations of WT-N and ΔN6-N started from their NMR structures ( PDB ID: 2XKS and 2XKU , respectively ) and those of ΔN6-I started from five conformations that were built from the intermediate state obtained in DMD simulations . As reported previously ( page 4 and Fig . 2 ) the intermediate state predicted by the full atomistic Gō model has two important structural features: it preserves the native core structure of ΔN6 ( the RMSD of the region comprising strands B-G plus connecting loops to the same region in the native structure is 3 Å ) and it exhibits a detached and unstructured strand A . To construct the starting conformations for the CpHMD we have firstly detached strand A from the core of the native conformation of ΔN6-N using PyMol ( http://www . pymol . org ) and subsequently relaxed these conformations via classical MD . All backbone dihedral angles modified were confirmed to be in Ramachandran allowed regions [79] . Relaxed conformations with an RMSD of 3 Å of the core region ( measured to the core of the native structure ) and five representative positions of strand A ( that are consistent with those found in the ensemble of conformations representative of the intermediate basin ) ( see Fig . 2A ) were then used as starting conformations for the CpHMD . By adopting this procedure one obtains conformations representative of the intermediate state with the most accurate/realistic representation of side-chain and backbone geometries , which is a requirement for the Monte Carlo ensemble docking protocol ( see below ) because the quality of the method's prediction depends critically on the structural accuracy of the analysed structures . We performed 30 simulations of 100 ns ( 3 systems , 2 pH values and 5 replicates ) . All simulations were performed using the stochastic titration constant-pH MD method implemented for the GROMACS package , developed by Baptista et al . [37] , [38] , [40]–[43] . The stochastic titration method consists essentially of a MM/MD simulation in which the protonation states of the protein are periodically replaced with new states sampled by Monte Carlo ( MC ) using Poisson-Boltzmann ( PB ) derived free energy terms . All His and acidic ( Asp , Glu and C-ter ) residues were titrated at all simulated pH values . Each constant-pH MD cycle was 2 ps long and the solvent relaxation step was 0 . 2 ps long . The MM/MD steps were performed using GROMACS 4 . 0 . 7 [80]–[82] and the GROMOS96 54A7 force field [83] . The leap-frog algorithm was used with a 2 fs time step . The structures were surrounded by 13641 SPC [84] water molecules in a rhombic dodecahedral box with periodic boundary conditions . The non-bonded interactions were treated using a twin-range cutoff of 8/14 Å and updating the neighbor lists every 10 fs . Electrostatic long range interactions were treated with a generalized reaction field [85] with a relative dielectric constant of 54 [86] and an ionic strength of 0 . 1 M [41] . The Berendsen coupling [87] was used to treat temperature ( 310 K ) and pressure ( 1 bar ) with coupling constants of 0 . 1 and 0 . 5 , respectively . Solvent and solute were separately coupled to the temperature bath . Isothermal compressibility of 4 . 5×10−5 bar−1 was used . All bonds were constrained using the LINCS algorithm . The PB/MC calculations were done as previously described [88] . The MEAD 2 . 2 . 0 [89] software package was used for PB calculations . The atomic charges and radii [88] were taken from the GROMOS96 54A7 force field . A dielectric constant of 2 for the protein and 80 for the solvent were used . Grid spacing of 0 . 25 , 1 . 0 and 2 . 0 Å were used in the finite difference focusing procedure [90] . The molecular surface was determined using a rolling probe of 1 . 4 Å and the Stern layer was 2 Å . The temperature was 310 K and the ionic strength was 0 . 1 M . The MC calculations were performed using the PETIT ( version 1 . 5 ) [91] software with 105 steps for each calculation . Each step consisted of a cycle of random choices of protonation state ( including tautomeric forms ) for all individual sites and for pairs of sites with a coupling above 2 . 0 pKa units [91] , [92] , followed by the acceptance/rejection step according to Metropolis criterion [93] . Several tools from the GROMACS software package [80]–[82] were used for analysis and others were developed in-house . The DSSP criterion [94] was used to assign the secondary structure . The MC-ED method highlights the role of shape complementarity , which is a major driver of protein aggregation [44] , [45] . The ultimate goal of the MC-ED [33] is to predict which parts of the protein are most likely to form geometrically matched protein-protein interfaces upon monomer self-association , and which residues may be critical for the onset of dimerization ( i . e . dimerization hot-spots ) . This It is based on the assumption that any pair of monomers ( representative of a specific conformational state , e . g . , the intermediate state of ΔN6-I equilibrated at pH 6 . 2 or 7 . 2 ) may a priori dimerize should they come into interaction distance and on the importance of interface shape complementarity in protein-protein association . This assumption translates into building an ensemble of random pairs , over which the propensity to form geometrically matched interfaces will be analyzed statistically , as the starting point of the method . The random pairing introduces no bias and it is physically reasonable , since there is evidence that monomers approach each other via a long-range hydration force of enthalpic origin acting on the hydrophilic residues [95] , before short-range , local hydrophobic interactions initiate dimerization and a well-packed interface may eventually be formed . The MC protocol employed to dock the two monomers represents protein conformations as rigid bodies and uses a series of random translations and rotations along the so-called docking axis ( which is the axis that a priori guarantees a higher number of intermolecular contacts ) combined with two cost functions that exclusively take into account packing interactions . For each pair of randomly selected conformations the MC returns an optimized docking interface with a maximum number of structural interactions ( i . e . intermolecular contacts ) and a minimum number of excluded volume interactions ( i . e . atomic clashes ) . The detailed chemical structure of each amino acid is taken into account in the full atomistic protein representation and also to establish the intermolecular contact map ( two amino acids are considered to be in contact in the dimer if any two atoms , whose size is given by the corresponding vdW radii , are within the interacting distance defined by the intramolecular Gō potential ) . To construct the density histograms the MC is applied consecutively to random pairs until the mean and the standard deviation of the number of intermolecular contacts converge . This typically amounts to dock up to 5000 pairs of conformations per studied species . The DHs are computed by counting ( and normalizing ) the number of dimers assigned to each bin of intermolecular contacts . The DHs provide a probabilistic description of the ensemble of random pairs from the point of view of the number of contacts of the geometrically matched interface each pair can form . They are used for comparison of dimerization propensity between species , and the ensemble of dimers that generates each DH is further used to identify the most likely structural parts of the protein which are key players in dimer formation , including the aggregation hot-spots . In this regard , since the heterogeneity in the amino acid interactions resulting from their different chemical nature is not taken into account into the corresponding cost function , the dimerization hot-spots will mostly depend on the quantity ( and not on the quality ) of intermolecular interactions established by each residue .
|
Dialysis-related amyloidosis ( DRA ) is a conformational disease that affects individuals undergoing long-term haemodialysis . In DRA the progressive accumulation of protein human β2-microglobulin ( Hβ2m ) in the osteoarticular system , followed by its assembly into amyloid fibrils , eventually leads to tissue erosion and destruction . Disclosing the aggregation mechanism of Hβ2m under physiologically relevant conditions represents a major challenge due to the inability of the protein to efficiently nucleate fibrillogenesis in vitro at physiological pH . On the other hand , ΔN6 , a truncated variant of Hβ2m , which is also a major component of ex vivo amyloid deposits extracted from DRA patients , is able to efficiently form amyloid fibrils de novo in physiological conditions . This amyloidogenic behavior is dramatically enhanced in a slightly more acidic pH ( 6 . 2 ) compatible with the mild acidification that occurs in the synovial fluid of DRA patients . In this work , an innovative three-stage methodological approach , relying on an array of molecular simulations , spanning different levels of resolution is used to investigate the initial stage of the de novo aggregation mechanism of ΔN6 in a physiologically relevant pH range . We identify an intermediate state for folding and aggregation , whose potential to dimerize is enhanced at pH 6 . 2 . Our results provide rationalizations for previous experimental observations and new insights into the molecular basis of DRA .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"physics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2014
|
A Simulated Intermediate State for Folding and Aggregation Provides Insights into ΔN6 β2-Microglobulin Amyloidogenic Behavior
|
Mosquito-borne viruses—such as Zika , chikungunya , dengue fever , and yellow fever , among others—are of global importance . Although vaccine development for prevention of mosquito-borne arbovirus infections has been a focus , mitigation strategies continue to rely on vector control . However , vector control has failed to prevent recent epidemics and arrest expanding geographic distribution of key arboviruses , such as dengue . As a consequence , there has been increasing necessity to further optimize current strategies within integrated approaches and advance development of alternative , innovative strategies for the control of mosquito-borne arboviruses . This review , intended as a general overview , is one of a series being generated by the Worldwide Insecticide resistance Network ( WIN ) . The alternative strategies discussed reflect those that are currently under evaluation for public health value by the World Health Organization ( WHO ) and represent strategies of focus by globally recognized public health stakeholders as potential insecticide resistance ( IR ) -mitigating strategies . Conditions where these alternative strategies could offer greatest public health value in consideration of mitigating IR will be dependent on the anticipated mechanism of action . Arguably , the most pressing need for endorsement of the strategies described here will be the epidemiological evidence of a public health impact . As the burden of mosquito-borne arboviruses , predominately those transmitted by Aedes aegypti and A . albopictus , continues to grow at a global scale , new vector-control tools and integrated strategies will be required to meet public health demands . Decisions regarding implementation of alternative strategies will depend on key ecoepidemiological parameters that each is intended to optimally impact toward driving down arbovirus transmission .
International public health workers are challenged by a burden of mosquito-borne arboviral diseases despite best efforts in control programs . An estimated 4 billion people live in areas at risk for dengue virus transmission alone [1] . Well-documented successes indicate that rigorously applied vector control using existing interventions can reduce arbovirus transmission and disease [2 , 3]; however , the degree to which such strategies may have prevented epidemics and the spread of arbovirus diseases is not well understood due to lack of evidence [4] . Despite existing interventions , epidemics and spread of arbovirus diseases continue . The reasons for this are complex but include inadequate program implementation; ineffective coverage; lack of human , financial and infrastructural capacity; insecticide resistance ( IR ) ; political will; and inability to scale . Integrated approaches and advancements in development of alternative strategies have been of renewed focus . This review provides an overview of strategies under development for the control of arbovirus mosquito vectors , focusing primarily on Aedes aegypti and A . albopictus .
A primary strategy for arbovirus outbreak control , such as dengue , is the use of synthetic chemicals with quick-action killing of adult vectors using space spraying [2 , 5 , 6] . The majority of recommended insecticides are of the pyrethroid chemical class , creating challenges to preventing selection pressure on susceptible mosquito populations as well as the control of pyrethroid-resistant vectors [5] . Regarding arbovirus vector population management , specifically of A . aegypti , larval control has long been proposed and implemented as a primary strategy [7] , including applications of chemical and microbial larvicides , insect growth regulators ( IGRs ) , and bacterial toxins [8] . Biological agents used against immatures include predatory copepods , fish , and Toxorhynchites larvae . Arguably , the greatest obstacle to A . aegypti larval control success is dependency on the ability to detect , access , and eliminate or treat domiciliary—often cryptic—breeding sites , a challenging and costly task that often leads to low coverage . In addition , their reduced efficiency in some occasions limits their widespread adoption [9] .
New tools are being developed on the premise that significant health benefit can be demonstrated in at least two endemic settings , aiming at niche roles rather than becoming the default intervention across a wide range of settings . Priority is given to tools that will improve current interventions in areas where they are challenged , either due to vector behaviors that prevent mosquito interaction with the intervention , IR , and/or residual disease transmission [5] . Several new strategies and product classes are under review by the World Health Organization Vector Control Advisory Group ( WHO VCAG ) [10] ( Fig 1 ) . We have used the current VCAG portfolio as a basis for our “inclusion criteria” for interventions described in this review . Specifically , we review strategies for which evaluations ( 1 ) have previously been conducted against arbovirus mosquito vectors demonstrating evidence of entomological success ( e . g . , autodissemination using entomopathogenic fungi or pyriproxyfen , pyrethroid treated traps , attractive targeted sugar baits [ATSB] , Wolbachia , genetic manipulation ) and/or ( 2 ) are actively underway against arbovirus vectors ( e . g . , spatial repellents , treated materials , sterile insect technique [SIT] ) . We did not include strategies currently being evaluated only against anopheline vectors ( e . g . , insecticide-treated eave tubes ) . The exception is gene drive due to historical theory and/or evaluations of entomological impact , expectation of broad utility of this strategy across disease vectors , and potential to overcome challenges posed by traditional genetic manipulation . In addition , the VCAG met in Geneva in November 2016 to review new potential vector control based on “genetic manipulation of mosquitoes through gene drive technology to reduce vector populations and transmission”; therefore , gene drive is part of the portfolio of vector control tools ( VCTs ) under consideration by WHO [11] . WHO has formally recognized some of these strategies for arbovirus control and encouraged testing in affected countries following appropriate monitoring and evaluation procedures [12] . Among the VCAG criteria for facilitating WHO recommendation , arguably the most critical is epidemiological evidence to endorse full-scale implementation . Funding to support rigorous pilot trials for generating preliminary evidence supportive of large-scale clinical trials as well as for randomized cluster trials themselves must be forthcoming . Where funding is not made available , support for delivering the requisite evidence base by other means , such as analyses of historical data captured during public health exercises , must be advocated . Other VCAG criteria include entomological correlates of protection , acceptability , and manufacturability ( Table 1 ) .
Entomopathogenic Ascomycetes fungi , especially Metarhizium anisopliae and Beauveria bassiana , have been suggested for control of both larval and adult stages of dengue vectors [13] . Fungal longevity ( duration of efficacy once applied ) may be one obstacle , with another being delivery of killing dose to target insect; therefore , formulation optimization is critical . Additionally , entomopathogens have fared badly in agriculture because they simply cannot compete with the costs and efficacy of chemical insecticides . This may change as insecticide regulation becomes more difficult and IR begins to dominate . An approach to circumvent the difficulty of locating and treating immature habitats is autodissemination , wherein dispersal and transfer of actives is carried out by contaminated adult mosquitoes [14] . Contamination can occur through treated materials [15] or dissemination stations such as modified ovitraps [16 , 17] . Once contaminated , mosquitoes disperse the agent in subsequent contacts with untreated surfaces . Autodissemination can exploit polygamic behavior whereby treated males can contaminate multiple females during mating [18] . It is important to note that there must be amplification in coverage between the lure and aquatic habitat to achieve benefit beyond killing offspring of only contaminated adults . An efficacy trial recently conducted in the United States showed a decline of A . albopictus populations following the field release of males contaminated with pyriproxyfen compared to untreated field sites [18] , and the combination of pyriproxyfen autodissemination with the SIT ( see below ) , a concept termed “Boosted SIT , ” is being explored [19] . Pilot interventions with pyriproxyfen have given promising results [16 , 17 , 20] . Design of new types of dissemination stations and other actives [21] as well as new formulations ( e . g . , IGRs in combination with bacterial toxins of spinosad or fungi B . bassiana ) may further improve effectiveness of autodissemination . Despite growing evidence of entomological efficacy , data requirements to demonstrate public health value are lagging ( Table 1 ) . Spatial repellents are products designed to release volatile chemicals into an air space , induce insect behavior modification to reduce human-vector contact , and thereby reduce pathogen transmission [22] . A spatial repellent product category is currently in Stage 3 of the VCAG assessment scheme , establishing proof-of-principle of efficacy through clinical trials ( Fig 1 ) . The application of spatial repellents at the household level through a consumer product market offers a “bottom-up” user-centric strategy for enhanced uptake ( coverage ) , potentially overcoming challenges of scale [23] . Evidence that community-led campaigns can impact transmission is needed . Spatial repellents may also be implemented through a donor-subsidized market similar to that used for insecticide-treated bed nets against malaria . Studies have demonstrated that chemicals currently recommended for vector control can elicit varied responses dependent on concentration [24] . For example , pyrethroids used in space spraying are applied at predominantly toxic levels to kill mosquitoes that land on treated surfaces or through contact with forced dispersal of formulated droplets . Other pyrethroids—labeled spatial repellents—such as transfluthrin and metofluthrin , are highly volatile at ambient temperature and disperse passively to repel , inhibit host-seeking , or kill mosquitoes depending on the chemical concentration gradient in the air space [25 , 26] . A range of spatial repellent products has demonstrated reduction in human−vector contact , and coils have been shown to contribute to reduction in human malaria infection [27 , 28] . Historically , the mode of action ( MoA ) of spatial repellent products has been focused on “movement away from the chemical source without the mosquito making physical contact with the treated surface” ( deterrency ) , an expanded concept that reflects the complexity in defining spatial repellents , includes chemical actions that interfere with host detection and/or disrupt blood-feeding , and was established by WHO in 2013” [29] . Knowledge gaps exist about how spatial repellents work , including exact molecular and physiological mechanisms [30] , the hereditary basis by which spatial repellent traits are maintained in populations , and the relationship of response intensity with IR [31] . All are vital characterizations required in discovery and optimization of spatial repellent compounds and formulated products . Despite this , demonstration of efficacy in pyrethroid-resistant A . aegypti [32] , as well as enhanced A . aegypti attraction response of gravid females to experimental ovipostion sites following exposure [33] , may offer insights into complementary or synergistic roles for spatial repellents in integrated vector management strategies . Mosquito traps have served for decades as effective surveillance tools but have only recently been considered under VCAG as a control strategy ( Fig 1 ) . For a trap to be an efficient tool for vector elimination , it must be highly sensitive and specific for a target species . The most effective traps rely on a combination of attractant cues such as light , heat , moisture , carbon dioxide , and synthetic chemicals for host attraction [34] . The Centers for Disease Control and Prevention ( CDC ) miniature light trap was introduced in 1962 [35] , but major advances have been made with newer traps that have incorporated innovations such as carbon dioxide plumes and counter flow geometry [36] . Another trap recently developed is the Mosquito Magnet Pro ( MMPro ) , which is commercially available and uses propane , making it affordable and easily deployable for the general public . Expanding on the trap concept for arbovirus mosquito vector population reduction was the introduction of the lethal ovitrap . These traps are designed to attract and kill egg-bearing females . The attractive baited lethal ovitrap ( ALOT ) has shown promise in both laboratory and field settings for significantly reducing Aedes populations . A prospective nonrandomized field trial of the ALOT trap was conducted in two cohorts of Iquitos , Peru . One year into the trial , dengue incidence as measured by febrile surveillance was 75% lower in the intervention area compared to the control cohort [37] . Although there have been a number of unsuccessful attempts to document significant reductions in vector densities using lethal ovitraps , it appears that sufficient coverage with an appropriate number of traps per unit area is key to this strategy’s success [38] . Other research has demonstrated potential of the baited gravid ( BG ) -Sentinel trap as a vector-control tool [39] . Studies with area-wide use of autocidal gravid ovitrap ( AGO ) in Puerto Rico have shown 80% reduction in densities of female A . aegypti for up to 1 year [40] . In Brazil , significant reductions in densities of gravid A . aegypti by the biogents passive gravid Aedes trap ( BG-GAT ) [39] was achieved . Incorporating a distribution model of peridomestic lethal ovitraps to remove gravid females from the vector population is anticipated to complement current A . aegypti control campaigns focused on source reduction of larval habitats . As naturally preferred oviposition sites are removed from the environment , artificial traps may become more effective at reducing vector populations; however , routine monitoring and/or retrieving of traps , if no longer used , must be incorporated to avoid traps becoming potential breeding sites in scaled control programs . In anticipation of success based on this concept , WHO has currently developed guidelines for efficacy testing of traps against arbovirus vectors to include indicators [http://apps . who . int/iris/bitstream/handle/10665/275801/WHO-CDS-NTD-VEM-2018 . 06-eng . pdf ? ua=1] . The ATSB product class is currently at Stage 2 of the VCAG evaluation pathway ( Fig 1 ) . The success of an ATSB strategy relies on attracting mosquitoes to and having them feed on toxic sugar meals sprayed on plants or used in bait stations . The use of sugar feeding to reduce mosquito populations was first reported by Lea in 1965 [41] and then other studies [42–49] in A . albopictus , other culicines , and sand flies . ATSBs for both indoor and outdoor control of mosquito vectors may impact populations by direct mortality induced by feeding on an insecticide bait , and/or , dissemination through a bait of mosquito pathogens or nonchemical toxins [50] . Because both female and male mosquitoes require sugar throughout the adult lifespan , the potential effects of this intervention on a vector population may be dramatic but will depend on feeding behavior ( e . g . , readiness to sugar feed indoors ) . Bait solutions are composed of sugar , an attractant , and an oral toxin . Toxins tested include boric acid [51] , spinosad , neonicotinoids , and fipronil [52] , as well as several other classes of insecticide [49] . Some attractants are focused on locally acquired sugars , juices , and fruit as mosquitoes may be selective toward carbohydrates originating from their geographic range , although more “general” attractants have been developed more recently [48 , 53] . A cumulative effect of ATSBs on an anopheline mosquito population was demonstrated for an area with alternative sugar sources resulting in delayed population level lethality as compared to sugar-poor areas [54] , suggesting its utility in arid environments . A single application of an ATSB affected Anopheles sergentii density , parity , survival , and hence vectorial capacity [54] . Potential drawbacks of this strategy might be the effects on nontarget sugar-feeding organisms , as well as the high coverage required . Environmental , safety , and cost estimates are being explored as part of the WHO VCAG requirements for demonstration of public health value ( Table 1 ) . Insecticide-treated materials ( ITMs ) can provide bite protection by killing or repelling vectors . ITMs , in the form of clothing , can be worn outside during the day , at work , or in school , offering protection where current mosquito control strategies , such as bed nets , may not . Military and other commercial companies have used insecticide-treated clothing to protect their workers from biting arthropods , and ITMs have reduced the incidence of vector-borne diseases such as malaria and leishmaniasis [55] . Currently , an ITM strategy for vector control is in Stage 1 of VCAG evaluations ( Fig 1 ) , with permethrin ( a pyrethroid ) being the only active ingredient used in ITMs due to requirements of meeting human safety profiles [55] . Permethrin ITMs have demonstrated personal protection against A . aegypti mosquitoes in various laboratory experiments [56–58] , reduction of A . aegypti human biting rates by 50% ( partial limb coverage ) to 100% ( when fully clothed ) in semifield trials [57 , 58] , and an 80% reduction in A . aegypti densities after just one month of children wearing permethrin treated school uniforms [59] . Models have estimated permethrin-treated uniforms could reduce dengue infections by up to 55% in the most optimistic scenarios [60] . Despite the potential for ITMs to control arbovirus vectors , current application techniques and formulations have limited efficacy under general use as permethrin washes out of material after several washes and is degraded by UV and heat exposure [56 , 61] . Novel formulations are needed to achieve long-lasting , effective release of permethrin under anticipated use . A technology being developed to address this challenge is microencapsulation , which binds deeper within the fabric and increases insecticide stability , allowing for a more consistent and extended release of the active ingredient [62] . Novel active ingredients that have far-ranging efficacy compared to permethrin and/or represent a different chemical class ( including natural ingredients ) will be needed to overcome biting on exposed skin not covered by ITMs , as well as to address pyrethroid resistance , respectively [63] . As with other alternative strategies under evaluation , it will be imperative that epidemiological evidence be generated in robust trial designs before wider implementation of ITMs can be recommended . Efficacy of ITMs will be dependent on user compliance; therefore wearable technologies must be acceptable to target populations ( see “Acceptability and compliance of alternative methods” section ) . The sterile insect technique ( SIT ) is based on the release of sterilized male insects , traditionally by means of irradiation , to suppress vector mosquito populations . SIT induces random lethal dominant mutations in the germ cells , which acts on the eggs in the female to prevent fertilization [64] . The concept is that sterile males will mate with wild females without producing any offspring . A major challenge to scale implementation has been building infrastructure in endemic settings to support mass rearing of the target vector . New technologies for mass rearing of mosquitoes , especially Aedes , are currently available [65–67]; however , improvement in the sterilization process is still needed to avoid somatic damage , resulting in reduction of longevity , problems with sexual vigor , and overall male activity [68] . Although encouraging results of SIT have been obtained with A . albopictus [69] , operational cost still constitutes a significant barrier to large-scale rearing facilities in endemic countries . The release of insects with dominant lethality ( RIDL ) strategy reduces vector populations ( self-limiting approach ) through individuals carrying a transgenic construct , which acts on the late larval stage and the pupae to prevent survival to imago . In contrast to both SIT and Wolbachia-based population suppression ( see “Gene drives” section ) , for RIDL technology , eggs must become fertilized for subsequent impact . The engineered effector gene is homozygous , repressible dominant lethal , and activates its own promoter in a positive feedback loop but can be regulated using an external activator . The construct also includes a reporter gene resulting in RIDL insects expressing a visible fluorescent marker for easy screening of transgenic and hybrid individuals before and after an intervention [70] . RIDL transgenic constructs have been successfully integrated into A . aegypti laboratory strains by Oxitec , which has performed open release trials with the OX513A strain since 2009 in Brazil , Cayman Islands , and Panama , with approvals pending for trials in the US and India . In Brazil , Panama , and Cayman , wild A . aegypti populations were reduced by more than 90% after release of the OX513A strain [71 , 72] . The sustainability of this reduction will depend on methods to avoid a new population increase from the remaining insects , hatching of dried eggs , and migration from uncontrolled areas . A continued monitoring system will provide a better overview of the impact of this method in long-term suppression of A . aegypti . Although the RIDL strategy has advanced to VCAG Stage 3 ( Fig 1 ) , a concern was that the reduction of A . aegypti could favor its replacement by A . albopictus . It should be noted that suppression of a population by any technique might encourage invasion and replacement by competitors . In Panama , however , six months after the OX513A A . aegypti releases ceased , there was no evidence of either expansion or augmented density of A . albopictus where both species occurred in sympatry [72] . Wolbachia is a natural intracellular bacterial symbiont found in at least 60% of insects known to alter reproduction of its host . Present in the female germline of an infected insect , Wolbachia is maternally transmitted to offspring . It can induce cytoplasmic incompatibility ( CI ) , where mating between Wolbachia-infected males and uninfected females yields eggs that fail to develop . Regular releases of male mosquitoes infected with a Wolbachia strain not present in the wild-type mosquito could theoretically reduce the viability of eggs in the field and lead to population suppression . Alternately , Wolbachia strains may cause a decrease in the vectorial capacity of the vector ( pathogen interference ) —directly by interfering with competence or indirectly by shortening lifespan [73] . The wMel Wolbachia strain currently being assessed by VCAG , in Stage 3 of evaluations , is intended for population replacement that interferes with ability to transmit pathogens ( Fig 1 ) . Regarding population suppression , though it was reported almost 50 years ago that Culex pipiens fatigans was eradicated via CI in a village in Myanmar ( then Burma ) , it was only in 1971 that the involvement of Wolbachia was reported [74] . Successes in species-specific suppression using inundative male releases have been demonstrated in semifield and full-field trials involving Aedes polynesiensis in French Polynesia [75] and A . albopictus in Lexington , Kentucky , US [76] . Releases involving Wolbachia−A . aegypti have also started in California , Thailand , Singapore , and Australia . The safety of the approach has been thoroughly evaluated and reported [77] . Although results of field studies have been promising , full-scale sustainable deployment of a Wolbachia strategy requires more developmental efforts . Besides the need for an arsenal of Wolbachia vector strains that offers excellent CI , there is a need for active community engagement to gain public acceptance and a need to scale up the production of large numbers of Wolbachia mosquitoes through automation and optimization . The population replacement strategy is based on the capacity of Wolbachia to invade and persist in wild mosquito populations , decreasing their vector competence [78] . Advantages of a replacement strategy , at its full potential , is the lack of requirement for consistent and continuing releases or reliance on community engagement to achieve near universal coverage . In the case of A . aegypti , the Wolbachia evaluation of a population replacement strategy has achieved important goals , such as a fruit fly Wolbachia strains can invade and sustain themselves in mosquito populations , reduce adult lifespan , affect mosquito reproduction , and interfere with pathogen replication [79] . So far three Wolbachia transinfection strains have been used in A . aegypti: wAlbB [80] introduced from A . albopictus , and wMel and wMelPop from Drosophila melanogaster [81] . The wMel Wolbachia strain has the ability to reduce A . aegypti vectorial capacity to dengue and chikungunya viruses [81 , 82] , an encouraging result recently extended to Zika virus , as indicated by experimental infection and transmission assays in wMel-infected mosquitoes [83] . The World Mosquito Program aims to promote research in arbovirus control by releasing Wolbachia-infected A . aegypti in dengue-affected communities ( www . eliminatedengue . com ) . The first release occurred in Australia in 2011 . Two natural populations were successfully invaded after 10 weekly releases , nearly reaching fixation five weeks after releases stopped [84] . This high infection rate has been maintained through 2017 without further input . The substitution of a natural A . aegypti population by a wMel-infected A . aegypti one prompted a scale up to other countries , including Vietnam , Indonesia , Brazil , and Colombia . Gene drives are transgenic constructs that possess the property to invade populations of the target species , even when conferring a fitness cost . The concept applied to mosquito control was proposed by Austin Burt as early as 2003 [85] and has since been a topic of research to spread a desired trait in mosquito species . Current gene drive designs are based on the Clustered Regularly Interspaced Short Palindromic Repeats—CRISPR-associated protein 9 ( CRISPR-Cas9 ) system . The transgenic element must be inserted precisely in the sequence that it is designed to cleave . To achieve this , a “cassette” is integrated by “gene knock-in” [86] . The drive cassette thus becomes heritable and able to “drive” ( Fig 2 ) . Alternatively , docking sites can be knocked in the gene of interest and subsequently serve as acceptors for the gene drive cassette [87] . Two strategies are under evaluation in VCAG Step 1 based on gene drives—population replacement and population suppression ( Fig 1 ) . In the latter , a drive element can be designed to insert into and inactivate a sex-specific fertility gene , suppressing the population as the resulting “sterility allele” increases its frequency . This approach was illustrated by a recent proof-of-principle laboratory study in the African malaria vector Anopheles gambiae [87] and could rapidly be adapted to A . aegypti and A . albopictus . Gene drives may prove to function less efficiently in the field than in the laboratory [88] due to the selection of pre-existing refractory variants in the guide RNA ( gRNA ) targets . Should gene drives function almost as efficiently in the field as in theory , the suppression approach carries the potential to eradicate the target species altogether , although insect population substructuring would probably allow the existence of residual pockets of intact populations [89 , 90] . Gene-drive designs for population suppression are not yet optimal for release in the field; a female sterility-spreading gene drive will only function optimally if heterozygous females are fully fertile , which is not the case of the published designs . Optimized promoters that restrict Cas9 activity to germline tissues , or Cas9 variants with shorter half-lives , are avenues to explore to alleviate these limitations . A careful evaluation of the ecological impact of species eradication must be considered . Gene-drive animals represent a new class of genetically modified organism ( GMO ) , the safety of which will need to be evaluated according to new criteria compared to traditional GMOs . This and other considerations , such as horizontal transfer , are the focus of a recent US National Academy of Sciences report on gene drives [91] . Gene drives can also be applied in population replacement strategies—using a drive construct to confer mosquito resistance to a given pathogen resulting in a population that becomes pathogen resistant as the genetic invasion progresses . Proof-of-principle has been demonstrated in the laboratory using the Asian malaria vector Anopheles stephensi [92] . In Aedes , antiviral constructs could be designed similarly , targeting one or several viruses . Candidate antiviral factors include RNA interference ( RNAi ) constructs , overexpressed components of the antiviral response , or RNA-targeting molecular scissors such as the recently identified CRISPR-Cas-like system C2c2 [93] . In all cases , it will be important to ascertain increasing resistance to a given virus does not render mosquitoes more susceptible to another . There are other product categories that have received recent attention but are in early development and not under VCAG assessment to include acoustic larvicides [94] and RNAi [95] .
IR in arbovirus vectors , such as A . aegypti , is considered a major obstacle to successful control [5] . The incidence of IR has increased rapidly in recent years [96] , and concerns regarding environmental impact caused from insecticide residues continue [97] . This highlights the need for alternative methods to better manage arbovirus vector populations while mitigating selection pressure on existing IR genes [98] . Other WIN reviews will describe IR distribution , mechanisms , and management; we present here complementary considerations on the integration of alternative strategies for resistance . The use of nonchemicals , or chemicals with completely different MoAs ( i . e . , alternate target site ) , will potentially have a greater impact on insecticide resistance management ( IRM ) , such as using IGRs for larval and pyrethroids for adult A . aegypti control , respectively [99] . Combination or rotations of unrelated compounds can ( in theory ) mitigate the occurrence of resistance and/or delay the selection process if already present at low levels . This may not hold true when considering metabolic resistance mechanisms by which one enzyme degrading one insecticide with a particular MoA may also degrade another insecticide with a different MoA . Reciprocally , two insecticides with the same MoA may not be degraded by the same enzymes ( no cross metabolic resistance ) . The authors acknowledge that any tool is vulnerable to resistance development; therefore , robust monitoring and evaluation should accompany implementation . The potential of example alternative strategies to impact IRM is indicated in Table 2 , whereas details of resistance risks are presented in Fig 3 . There are a number of ways in which alternative strategies can control insecticide-resistant vectors , including different MoAs or different application methods ( i . e . , oral versus contact ) . Tests are available to measure and predict this "antiresistance" potential at several stages of product development; however , evidence is missing , for example , on the impact of target site resistance to pyrethroids on efficacy of transfluthrin ( a pyrethroid spatial repellent ) to reduce human infections in settings with resistant vector populations , or the efficacy of pyriproxyfen against insects overexpressing cytochrome P450s capable of metabolizing pyrethroids . In addition to cross resistance and efficacy assessments against pyrethroid-resistant mosquitoes , the introduction of novel active ingredients into VCTs requires careful monitoring of possible resistance development . Strategies used for IRM and the introduction of new VCTs is adopted from agricultural practices [100] , for which another WIN review is forthcoming . IRM should be applied within integrated vector management ( IVM ) approaches [101] , which is defined as the rational decision-making process for the optimal use of resources for vector control [102] and includes the use of multiple complementary tools . An important point to stress is that a novel vector-control product may be more useful via its role in IRM in a certain setting with high levels of resistance risk , even if it is of a similar efficacy to existing interventions ( “non inferior” ) .
The importance of characterizing the biological profile ( genetic structure , insecticide susceptibility ) of the local vector population should be considered in the context of IVM and IRM using alternative strategies . For instance , the first release of Wolbachia in Rio de Janeiro , Brazil , did not succeed , because the spread of the bacteria was not self-sustained . Although wMel Wolbachia was already hosted by around 60% of the local A . aegypti after 20 weeks of insertion , this frequency rapidly dropped when releases stopped . This may likely have occurred because the released wMel-infected A . aegypti strain was susceptible to pyrethroids , the class of insecticide largely used by residents in dwellings . It was known that the natural A . aegypti population of that locality was highly resistant to pyrethroids . When a new colony of A . aegypti , as resistant as the field population , was infected with wMel and released in the field for 24 weeks , 80% of local A . aegypti presented the bacteria , and this index is continually increasing [103 , 104] . Additional characterizations of importance include male feeding behavior for optimizing ATSB and genotyping key phenotypes to identify targets for gene drives .
Social science , effective training of staff , and capacity building have major roles in the deployment of novel VCTs . New interventions must be acceptable to the communities in which it will be used . One possible barrier is that homeowners might perceive the intervention as ineffective or harmful to themselves or their environment , and adoption of the strategy might be compromised . For example , populations may be reluctant to use juvenoids ( e . g . , pyriproxyfen ) because the insecticide has a late killing action on pupae , leaving live larvae in the container . Regardless of how efficacious a control strategy may prove to be during proof-of-concept , it is critical to assess the potential barriers and/or acceptance of local populations in which implementation trials may occur early in the development phase , even within a limited participant population [105 , 106] . More robust surveys can be used once an intervention prototype is available [107] . Such surveys can be used to assess factors associated with adoption and maintenance behaviors and identify barriers to its correct and consistent use to ensure ( or improve ) product sustainability .
Regulatory considerations are required early in the development process to ensure data requirements are met for safety and use ( Table 1 ) . If a tool under investigation does not fall under an existing product category ( i . e . , is a new class ) , a new regulatory framework and/or data requirements may be warranted . It is important to note , regulation can be as substantial an obstacle to an effective intervention as any gap in research and development ( R&D ) . Where regulations are prohibitive , options for arbovirus control will become even more constrained; this may be most evident when a lack of expertise and/or infrastructure exists around the new tool ( e . g . , biotechnology and gene drives ) . Historically , the WHO Pesticide Evaluation Scheme ( WHOPES ) was the body responsible for review of product evaluation and recommendations [108] . However , a reform in the regulatory process for evaluation of new products has been made with the recent agreement by WHO to adopt the Innovation to Impact ( I2I ) initiative in 2017 [109] . The I2I is expected to accelerate the evaluation process , increase transparency , and improve quality assurances , similar to the framework already adopted for drug and vaccine prequalifications [110] . Development of new guidelines or modifications of existing ones for efficacy testing of alternative VCTs are anticipated to be needed as novel MoA are exploited . Likewise , if the formulated product under investigation is not registered in the study area for which it will be evaluated , experimental use permits or similar must be obtained before investigations begin . Proprietary protection for evaluation of product formats still under development should follow industry specifications .
Arboviruses transmitted by Aedes mosquitoes represent major international public-health concerns that will surely require a range of integrated interventions to be effectively controlled . As the scope of arboviruses continues to grow , development and evaluation of alternative vector-control products and strategies are critical to pursue . Following endorsement by global , national , and local authorities , effective strategies will have to be locally adapted to take into account the biology of the vector and virus transmission intensity , as well as human and financial resources . This review focuses on alternative strategies mainly for control of A . aegypti and A . albopictus because these two species are arguably the primary arbovirus vectors in the world . Alternative strategies will provide additional options for arbovirus control and potentially add value to existing strategies; however , until operational effectiveness and frameworks for use are in hand , further optimization of current strategies is warranted , to include innovative delivery methods of existing products ( e . g . , targeted indoor residual spraying [111] ) .
|
International public health workers are challenged by the burden of arthropod-borne viral diseases , to include mosquito-borne arboviruses transmitted by Aedes aegypti and A . albopictus due in part to lack of sustainable vector control and insecticide resistance ( IR ) , as well as the inability to scale up and sustain existing interventions for prevention of urban epidemics . As a consequence , there has been increasing interest to advance the development of alternative methods . This review provides a general overview of alternative vector-control strategies under development for the control of arbovirus mosquito vectors and highlights how each could offer innovative public health value . Considerations to regulations , acceptance , and sustainability are also provided .
|
[
"Abstract",
"Introduction",
"Rationale",
"for",
"developing",
"alternative",
"strategies",
"Outlook",
"on",
"alternative",
"strategy",
"development",
"Alternative",
"strategies",
"Considerations",
"for",
"introducing",
"alternative",
"strategies",
"Knowledge",
"base",
"on",
"mosquito",
"biology",
"and",
"behavior",
"Acceptability",
"and",
"compliance",
"of",
"alternative",
"methods",
"Regulatory",
"considerations",
"for",
"alternative",
"control",
"strategies",
"Conclusion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"animals",
"wolbachia",
"viruses",
"review",
"infectious",
"disease",
"control",
"insect",
"vectors",
"bacteria",
"infectious",
"diseases",
"agrochemicals",
"aedes",
"aegypti",
"arboviral",
"infections",
"disease",
"vectors",
"insects",
"agriculture",
"arthropoda",
"insecticides",
"mosquitoes",
"eukaryota",
"arboviruses",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"viral",
"diseases",
"organisms"
] |
2019
|
Alternative strategies for mosquito-borne arbovirus control
|
Gli3 is a major regulator of Hedgehog signaling during limb development . In the anterior mesenchyme , GLI3 is proteolytically processed into GLI3R , a truncated repressor form that inhibits Hedgehog signaling . Although numerous studies have identified mechanisms that regulate Gli3 function in vitro , it is not completely understood how Gli3 function is regulated in vivo . In this study , we show a novel mechanism of regulation of GLI3R activities in limb buds by Gata6 , a member of the GATA transcription factor family . We show that conditional inactivation of Gata6 prior to limb outgrowth by the Tcre deleter causes preaxial polydactyly , the formation of an anterior extra digit , in hindlimbs . A recent study suggested that Gata6 represses Shh transcription in hindlimb buds . However , we found that ectopic Hedgehog signaling precedes ectopic Shh expression . In conjunction , we observed Gata6 and Gli3 genetically interact , and compound heterozygous mutants develop preaxial polydactyly without ectopic Shh expression , indicating an additional prior mechanism to prevent polydactyly . These results support the idea that Gata6 possesses dual roles during limb development: enhancement of Gli3 repressor function to repress Hedgehog signaling in the anterior limb bud , and negative regulation of Shh expression . Our in vitro and in vivo studies identified that GATA6 physically interacts with GLI3R to facilitate nuclear localization of GLI3R and repressor activities of GLI3R . Both the genetic and biochemical data elucidates a novel mechanism by Gata6 to regulate GLI3R activities in the anterior limb progenitor cells to prevent polydactyly and attain proper development of the mammalian autopod .
Understanding the developmental mechanisms that regulate progenitor cells to generate organs with specific morphology and function is a central topic in developmental biology . The vertebrate limb has been serving as an excellent system for such studies . In particular , mesenchymal progenitor cells in limb buds are specified , patterned and expanded to generate each skeletal element with a distinct morphology at each defined position to create the stereotypical limb skeletal system . The mammalian autopod possesses five digits , termed as d1-d5 , in an anterior to posterior order . The number and identity of digits have been used as a readout of specification , patterning , and proliferative expansion of progenitor cells [1] . Sonic Hedgehog ( Shh ) is expressed in the zone of polarizing activity ( ZPA ) , located at the posterior mesenchyme of the limb bud , and acts as a major regulatory molecule for limb development [1 , 2] . Anterior-posterior specification of digit progenitors is regulated by the concentration and duration of progenitor exposure to SHH [3–6] . SHH also regulates the proliferative expansion of mesenchymal progenitor cells to generate a sufficient number of cells to develop into cartilage condensations [7 , 8] . Accordingly , ectopic expression of Shh in the anterior portion is associated with preaxial polydactyly , which is characterized by the formation of ectopic digits in the anterior of the limb [9] . By contrast , the most anterior digit ( d1 ) develops in a SHH-independent manner [10 , 11] . Recent studies have shown that anterior genetic programs , such as Irx3-Irx5 and Sall4 , are required for development of d1 , at least in part , by excluding SHH signaling from the anterior mesenchyme [12 , 13] . The glioma-associated oncogene family ( GLI ) proteins are zinc finger DNA binding proteins , which play diverse roles in animal development and diseases [14] . Among the three Gli genes , Gli3 encodes a bi-functional molecule , acting as both an activator ( GLI3A ) and a repressor ( GLI3R ) , whose balance depends on Hedgehog signaling [14] . In the presence of Hedgehog ligands , its signal transduction at primary cilia causes inhibition of proteolytic processing of GLI3 [15] . This results in the accumulation of a full-length activator form of GLI3 ( GLI3A ) in the posterior mesenchyme . In contrast , in the absence of Hedgehog signaling , GLI3 is subjected to proteolysis , generating a truncated repressor form ( GLI3R ) , which accumulates in the anterior mesenchyme . Because GLI1 lacks a repressor domain and GLI2 predominantly functions as an activator [16 , 17] , GLI3R is the major GLI repressor in the limb [18] . Consistent with the important function of Gli3 in limb development , its mutations cause developmental defects in mice and humans [19–21] . In particular , Gli3-/- mice develop polydactyly [21] . Genetic studies in mice demonstrated that a predominant function of Gli3 is to repress Hedgehog signaling target genes [22 , 23] . Furthermore , it has been shown that the balance of GLI3A and GLI3R regulates digit number and identity [24–26] . Numerous studies have shown that multiple mechanisms regulate GLI3 functions in vitro , such as posttranslational modifications , degradation , cytoplasmic retention , and primary cilium-mediated processing ( reviewed in [14 , 27 , 28] ) . In vivo studies in mice demonstrated that Gli3 genetically interacts with Hox genes , Zic3 and Alx4 during limb development [29–31] . Despite these studies , the in vivo control of Gli3 function during proper limb development is still to be elucidated . The Gata family of zinc finger transcription factors is an important regulator of tissue and organ development . The Gata family is subdivided into the Gata1/2/3 subfamily and the Gata4/5/6 subfamily , which show expression in hematopoietic cell lineages and meso-endoderm lineages , respectively [32 , 33] . In particular , Gata6 is essential for endoderm formation and is also involved in the development of various mesoderm- and endoderm-derived organs , such as the cardiovascular system and pancreas [34–37] . Moreover , a recent study suggested that Gata6 functions as a negative regulator of Shh expression in limb buds by binding to its limb bud-specific cis-regulatory element , ZRS [38] . In this study , we found that broad deletion of Gata6 in the limb mesenchymal progenitors caused hindlimb-specific preaxial polydactyly , which is associated with ectopic SHH signaling in the anterior hindlimb bud . We discovered that Gata6 and Gli3 genetically interact to regulate normal patterning of the hindlimb . Furthermore , we show that direct association of GATA6 with GLI3R promoted nuclear localization and transcriptional repressor activity of GLI3R . Our work identified that genetic and biochemical interactions between Gata6 and Gli3 act as essential mechanisms to regulate GLI3R activity for proper autopod patterning .
Prior studies have identified expression of Gata6 in developing limb buds [38–40] . Gata6 null embryos die during gastrulation [34 , 35]; therefore , we inactivated Gata6 in the meso-endoderm by using the conditional allele of Gata6 ( Gata6fl ) [41] and the Tcre line , which recombines in the early meso-endoderm [42] . We found that Tcre; Gata6fl/fl mutants ( hereafter referred to as Gata6 cKO ) die around E12 . 5–14 . 5 with broad hemorrhage ( Fig 1A and 1E ) . This result is consistent with a former study , demonstrating a role of proper dosage of Gata4 and Gata6 for vessel integrity [43] . We found that Gata6 cKO embryos exhibited polydactyly in the hindlimb , while forelimbs seem to be unaffected ( Fig 1A–1C , 1E and 1G , S1 Table ) . Alcian blue staining demonstrated that the mutant hindlimbs possess patterned digits , d1-d5 , and an extra digit on the anterior edge , which morphologically resembles d1 . Based on the position and morphology , tarsal and metatarsal elements were also patterned . Two ectopic tarsal elements , likely the navicular and medial cuneiform , were present proximally to the ectopic 1st metatarsal ( Fig 1D and 1H ) . These observations indicate that the autopod is patterned along the anterior-posterior axis , and the absence of Gata6 induces the formation of an extra anterior digit with the associated tarsal and metatarsal elements . Preaxial polydactyly is known to be associated with ectopic Sonic Hedgehog ( Shh ) expression in the anterior margin . At E10 . 5 , we detected posteriorly-localized Shh expression without ectopic anterior expression ( n = 4 , 39–40 somite stage , Fig 1I and 1Q ) . Consistent with this normal expression , Hoxd13 ( n = 3 ) and Hand2 ( n = 6 ) , upstream regulators of limb bud Shh expression [44] , were normally expressed in the posterior mesenchyme ( Fig 1J , 1K , 1R and 1S ) . However , Gli1 ( n = 3 ) and Patch1 ( n = 3 ) , targets of Hedgehog signaling , were detected in the anterior margin of Gata6 cKO hindlimb buds ( Fig 1L , 1M , 1T and 1U ) . Expression of anterior marker genes , such as Alx4 ( n = 3 ) , Gli3 ( n = 4 ) and Irx3 ( n = 3 ) , were not significantly affected in Gata6 cKO hindlimb buds ( Fig 1N–1P and 1V–1X ) . We also examined gene expression at a later stage . At E11 . 5 , we detected ectopic Shh expression in the anterior border of Gata6 cKO hindlimbs ( n = 4 , S1 Fig ) . Consistent with evident ectopic Shh expression , expression of Hoxd13 ( n = 3 ) , Gli1 ( n = 6 ) , Ptch1 ( n = 6 ) and Gremlin1 ( n = 3 ) was also detected in the anterior margin . This data indicates that ectopic Hedgehog signaling became evident at E10 . 5 in Gata6 cKO hindlimb buds , although ectopic Shh expression was undetectable . At a later stage ( E11 . 5 ) , ectopic Shh expression became evident and all SHH targets , examined in this study , were detected in the anterior margin . Shh expression is negatively regulated in the anterior margin by various genes . Thus , we examined expression of negative regulators of Shh expression . In addition to Alx4 and Gli3 ( Fig 1 ) [23 , 45] , expression of Etv4 ( n = 3 ) , Etv5 ( n = 5 ) , Tulp3 ( n = 3 ) , Twist1 ( n = 3 ) , whose loss can cause ectopic Shh expression in the anterior margin [46–52] , did not show evident alteration ( S2 Fig ) . Therefore , it is unlikely that these genes account for the preaxial polydactyly phenotype in Gata6 cKO hindlimbs . If ectopic Shh expression accounts for the preaxial polydactyly in Gata6 cKO hindlimbs , we would expect that reducing Shh dosage might rescue the phenotype . Therefore , we genetically reduced Shh dosage from the Gata6 cKO background using the Shh null allele [2] . Gata6 cKO; Shh+/- mutants did not survive beyond E12 . 5 , thus , we examined expression of SHH target genes ( Gli1 and Ptch1 ) and expression of Sox9 , an early marker of chondrogenic condensation [53] . Removing one allele of Shh from the Gata6 cKO background resulted in posteriorly restricted expression of Gli1 and Ptch1 , and the ectopic anterior expression became undetectable ( n = 4 , Fig 2A–2C and 2E–2G ) . However , ectopic chondrogenic condensation in the anterior portion was still detected by Sox9 expression at E12 . 5 ( n = 3 , Fig 2I–2K ) . Removing both alleles of Shh from the Gata6 cKO background resulted in the loss of Gli1 and Ptch1 expression and single digit condensation , the same phenotype as Shh-/- limbs ( n = 3 , Fig 2D , 2H and 2L ) [10 , 11] . These results indicate that Shh functions downstream of Gata6 during preaxial polydactyly development . However , ectopic chondrogenic condensation in the anterior portion of Gata6 cKO; Shh+/- hindlimbs suggests that additional mechanisms could be involved in the preaxial polydactyly in Gata6 cKO hindlimbs . GLI3 is a major regulator of Hedgehog signaling , and thus , Gli3 might be involved in preaxial polydactyly in Gata6 cKO hindlimbs . To test this hypothesis , we genetically removed Gli3 from the Gata6 cKO background . Gli3+/- hindlimbs developed a small spike in the anterior region [21 , 54] , while most of the Tcre; Gata6+/fl hindlimbs were indistinguishable from the wild-type hindlimbs at E14 . 5–15 . 5 ( Fig 3F–3H , Table 1 ) . Tcre; Gata6+/fl; Gli3+/- compound heterozygous hindlimbs developed an extra digit in the anterior region ( Fig 3I ) . Unexpectedly , we also found that this interaction operates in forelimbs . Gli3+/- forelimbs developed d1 , which was associated with small ectopic cartilage condensation at the distal tip . Contrary to this , Tcre; Gata6+/fl; Gli3+/- compound heterozygous forelimbs developed an evident extra digit with incomplete penetrance ( Fig 3A–3D ) or an extra digit that partially fused with endogenous d1 with incomplete penetrance ( S2 Table ) . These results demonstrate a genetic interaction between Gli3 and Gata6 in fore- and hind-limbs . Because the Gata6 cKO limb phenotype was evident in hindlimbs , we focused the following analysis on hindlimbs . Ectopic Shh expression can cause preaxial polydactyly , therefore , we examined Shh expression at E11 . 5 . We detected a small domain of anterior ectopic Shh expression in Gli3-/- hindlimbs ( n = 3/6 , Fig 3O ) , as previously reported [23] . By contrast , Tcre; Gata6+/fl; Gli3+/- compound heterozygous hindlimbs did not exhibit anterior ectopic Shh expression ( n = 6 ) , similar to wild-type , Tcre; Gata6+/fl ( n = 6 ) and Gli3+/- ( n = 6 ) hindlimb buds ( Fig 3K–3N ) . Therefore , preaxial polydactyly in Tcre; Gata6+/fl; Gli3+/- compound heterozygous limbs were unlikely to be caused by ectopic Shh expression . Given that GLI3R prevents ectopic digit formation in the anterior portion [55] , these results suggest that an interaction between Gata6 and Gli3 contributes to GLI3R activities . Gata6 cKO; Gli3+/- embryos do not survive beyond E12 . 5 , therefore , we further examined the interaction between Gata6 and Gli3 by visualizing digit condensation by Sox9 in situ hybridization . Both Gli3+/- and Tcre; Gata6+/fl hindlimbs exhibited similar expression patterns to wild-type hindlimbs at E12 . 5 ( Fig 3P–3R ) . Correlating with preaxial polydactyly at E15 . 5 , Gata6 cKO and Tcre; Gata6+/fl; Gli3+/- compound heterozygous hindlimbs exhibited ectopic anterior digit condensation ( Fig 3S and 3T ) . Gata6 cKO; Gli3+/- hindlimbs were slightly underdeveloped and exhibited seven digit condensations ( n = 2/6 , Fig 3U ) , distally-fused condensation ( n = 2/6 , Fig 3V ) or one extra anterior condensation , similar to Gata6 cKO hindlimbs ( n = 2/6 ) . Formation of multiple extra digits and distal fusion of cartilage condensation are characteristics of Gli3-/- limbs [21] . Therefore , we speculate that the Gata6 cKO; Gli3+/- genotype may be in conditions similar to the Gli3-/- genotype in hindlimbs . These results further support the idea that loss of Gata6 leads to reduction of GLI3R activities . In order to further characterize the Gata6-Gli3 interaction , we examined gene expression at E11 . 5 . Expression of Gli1 and Patch1 was posteriorly restricted in wild-type , Tcre; Gata6+/fl and Gli3+/- hindlimbs ( Fig 4A–4C and 4H–4J ) . Hindlimbs with the Tcre; Gata6+/fl; Gli3+/- , Gata6 cKO , Gata6 cKO; Gli3+/- or Gli3-/- genotypes exhibited anterior ectopic expression of these genes ( Fig 4D–4G and 4K–4N ) . The ectopic expression domain was larger in Gata6 cKO and Gata6 cKO; Gli3+/- hindlimb buds than that in Tcre; Gata6+/fl; Gli3+/- and Gli3-/- hindlimbs , likely due to ectopic Shh expression in the Gata6 cKO background . Pax9 , whose expression requires high levels of GLI3R activities [56] , was detected in the anterior of wild-type and Tcre; Gata6+/fl hindlimbs , and was reduced in Gli3+/- hindlimb buds ( Fig 4O–4Q ) . In Tcre; Gata6+/fl; Gli3+/- , Gata6 cKO , Gata6 cKO; Gli3+/- hindlimbs , Pax9 expression was undetectable , similar to Gli3-/- hindlimbs ( Fig 4R–4U ) . These alterations of gene expression at E11 . 5 are consistent with the idea that GLI3R activities were reduced in hindlimbs with the Tcre; Gata6+/fl; Gli3+/- , Gata6 cKO and Gata6 cKO; Gli3+/- genotypes . Ectopic Shh expression in the Gata6 cKO background could affect gene expression patterns in hindlimb buds . Therefore , we set up in vitro experiments to further investigate how Gata6 regulates Gli3 function . We first set up luciferase reporter assays using 12xGLI-binding site luciferase [31] . Transfecting a C-terminally truncated form of human GLI3 that could function as GLI3R caused significant reduction of the reporter activities , while transfecting human GATA6 did not affect the reporter activities . Co-transfecting GLI3R and GATA6 caused further reduction of the reporter activities ( Fig 5A ) . These results are consistent with the in vivo data and support the idea that GATA6 functionally interacts with and contributes to GLI3R activities . Next , we tested whether GATA6 and GLI3R physically interact by co-immunoprecipitation assays . HEK293T cells were transfected with Flag-tagged GATA6 , Myc-tagged GLI3R or GFP . Flag-GATA6 and Myc-GLI3R were co-immunoprecipitated , demonstrating that GATA6 and GLI3R can interact ( Fig 5B and 5C ) . We also confirmed that the interaction occurs in vivo . GLI3R was detected in immunoprecipitated complex from E10 . 25–10 . 5 wild-type hindlimb buds using ant-GATA6 ( Fig 5D ) . To further characterize their interaction , we mapped the GLI3R interaction domain in GATA6 . For this purpose , we generated serial deletion mutants ( Fig 5E ) , and performed co-immunoprecipitation assays with Flag-GLI3R . The ΔN1 and ΔN2 mutants showed a strong interaction with Flag-GLI3R . The ΔN3 and ΔC1 mutants exhibited weak interaction , and we did not detect any interactions of Flag-GLI3R with ΔN4 and ΔC2 ( Fig 5F ) . We also generated intra-molecular deletion mutants . These mutants lack the GLI3R-binding domain ( GBD ) 1 , which includes the second putative transactivation domain ( ΔGBD1 ) , or both GBD1 and GBD2 ( ΔGBD1/2 ) . We did not detect any interaction of ΔGBD1/2 with GLI3R , although ΔGBD1 exhibited a weak interaction with GLI3R ( Fig 5G ) . These results suggest that the zinc finger domain 1 ( ZFD1 ) is critical to interact with GLI3R . The weak interaction of ΔN3 , ΔC1 and ΔGBD1 , which possess the ZFD1 , also suggests that both the N- and C-terminal regions around the ZFD1 contribute to the interaction with GLI3R , in collaboration with the ZFD1 . Our analyses indicated the presence of genetic and physical interactions between Gata6 and Gli3 . Given that both GATA6 and GLI3R act as transcription factors , we next examined subcellular localization of these proteins after co-transfecting HEK293T cells with Flag-GLI3R and either full length or mutant forms of Myc-GATA6 . We observed three patterns of localization ( Fig 6A and 6B , S3 Fig ) . First , co-transfection of either full length GATA6 , ΔN1-GATA6 or ΔN2-GATA6 , which can interact with GLI3R and possess the nuclear localization signal ( NLS ) , resulted in predominant nuclear localization of both GLI3R and GATA6 . Second , we co-transfected ΔN3-GATA6 or ΔN4-GATA6 , which possess the NLS , but have either very weak or undetectable interactions with GLI3R . In these transfection assays , GLI3R localization became either predominantly cytoplasmic or localized similarly in both the cytoplasm and nucleus , although GATA6 was predominantly detected in the nucleus . Third , we co-transfected ΔC1-GATA6 or ΔC2-GATA6 , which lack the NLS and have very weak or undetectable interactions with GLI3R . We detected GATA6 predominantly in the cytoplasm , consistent with the lack of NLS . GLI3R was also predominantly located in the cytoplasm or located similarly in the nucleus and cytoplasm . These results indicate a correlation between GLI3R nuclear localization and nuclear GATA6 that possesses a GLI3R-interaction ability . This correlation suggests that physical association between GATA6 and GLI3R contributes to nuclear localization and the repressor activities of GLI3R . We next tested this idea in vivo by examining GLI3R nuclear localization . The earliest molecular alteration in Gata6 cKO hindlimb buds in our study is ectopic Gli1 and Ptch1 expression at E10 . 5 ( Fig 1 ) . Therefore , we re-examined Gata6/GATA6 expression , although their mRNA expression patterns were examined in previous studies [38–40] . Gata6 mRNA was detected in the anterior-proximal part of hindlimb buds at E10 . 25 ( 34 somite stage ) ( S4A Fig ) , but the strong signals in endoderm -derived tissues seem to mask the limb bud signals . Therefore , we also performed immunofluorescence of GATA6 in combination with limb bud mesenchyme markers , such as Fibroblast growth factor10 ( FGF10 ) [57] or Dual specificity phosphatase6 ( DUSP6 ) [58–60] . Co-staining with these markers on transverse sections indicates that GATA6 is present in the ventral side of the proximal region in anterior hindlimb buds at E10 . 25 ( S4B and S4C Fig ) . The GATA6 signal was undetectable in limb buds in the middle-posterior region . In the anterior proximal region of limb buds at E10 . 25 , we detected GLI3R predominantly in the nucleus or similarly in the nucleus and cytoplasm ( Fig 6C , 6E , 6G and 6I ) . By contrast , Gata6 cKO hindlimb buds showed a reduced percentage of cells with predominant nuclear GLI3R signals . Accordingly , we detected an increased percentage of cells with nuclear/cytoplasmic GLI3R ( Fig 6D , 6H and 6I ) . Western blot analysis of nuclear extracts from the anterior part of hindlimb buds showed reduced GLI3R levels in Gata6 cKO , compared to wild-type embryos ( Fig 6J ) . Although the presence of nuclear GLI3R in Gata6 cKO hindlimb buds indicates Gata6-independent GLI3R nuclear localization mechanisms in the anterior mesenchyme , reduced GLI3R levels provide evidence that Gata6 contributes to GLI3R nuclear localization . These results are consistent with the in vitro data , and further support the idea that Gata6 regulation of GLI3R nuclear localization contributes to GLI3R activities during normal limb development .
In this study , we found hindlimb-specific preaxial polydactyly in Gata6 mutants . The skeletal phenotype of Gata6 mutants was restricted to hindlimbs , and the forelimbs developed normally . Several possibilities would account for such limb type-specific phenotypes . For instance , a recent study showed that Gata4 is differentially expressed in forelimb buds ( high ) and hindlimb buds ( low ) [38] . Gata4 and Gata6 are functionally redundant during heart development and for vascular integrity [36 , 43]; therefore , Gata4 might compensate for loss of Gata6 in forelimb buds [38] . Another possibility is that differences in the sensitivity to Hedgehog signaling contribute to different phenotypes in fore- and hind-limbs . It is suggested that levels of Hedgehog signaling are higher in hindlimb mesenchyme than forelimb mesenchyme [12] , and that hindlimbs are more sensitive to changes in the levels of Hedgehog signaling . Higher Hedgehog signaling , in combination with reduced GLI3R , might have contributed to hindlimb-specific polydactyly in Gata6 cKO . This idea is consistent with ectopic digit formation in Tcre; Gata6+/fl; Gli3+/- forelimbs , in which GLI3R activities would be lower than and SHH signaling levels would be higher than Gli3+/- forelimbs . These two scenarios are not mutually exclusive , and they might cooperate together to ensure proper Hedgehog signaling and pentadactyly in mammalian limbs . Our study proposes two mechanisms by which Gata6 regulates proper autopod patterning . One mechanism is by enhancing GLI3R activities to repress Hedgehog signaling in the anterior mesenchyme , and the other is by negative regulation of Shh expression in the anterior mesenchyme . Genetic studies have shown that preaxial polydactyly is associated with ectopic expression of Shh in the anterior mesenchyme [9] . Expression of Shh is positively and negatively regulated in the posterior and anterior mesenchyme , respectively . Twist1 , Alx4 , Gli3 , Tulp3 and Etv4-Etv5 act as negative regulators , for their loss of function caused ectopic Shh expression [23 , 47 , 49 , 51 , 61] . Genetic and biochemical studies have shown that Hand2 and Hoxd13 positively regulate Shh expression through the limb bud-specific cis-regulatory element , ZRS [44 , 62] . Anterior Shh expression could be induced by loss of negative regulators or ectopic expression of positive regulators [63] . Given that these regulators did not exhibit significant alteration in Gata6 cKO hindlimb buds , the preaxial polydactyly phenotype in Gata6 cKO limbs is unlikely to be induced through these genes . A recent study suggested that Gata6 represses Shh in the limb through binding to ZRS [38] . Our data is consistent with this report , and demonstrated that Shh and its targets are ectopically expressed in Gata6 cKO hindlimb buds at E11 . 5 . Restoration of normal expression pattern of Gli1 and Ptch1 in Gata6 cKO; Shh+/- hindlimbs also supports the idea that Gata6 is upstream of Shh . The second role is repressing ectopic Hedgehog signaling by enhancing repressor function of Gli3 . Ectopic Shh expression in the Gata6 cKO background affects data interpretations; however , compound heterozygous mutant analyses could enable separate analysis of the two mechanisms and support the second mechanism . Previous studies have shown Gli3 to genetically interact with other genes during limb development . Studies on Hox genes suggested that the Gli3-/- polydactyly phenotype is mediated by Hoxd9 and Hoxd10 [29 , 64] . In addition , it has been shown that polydactyly of Gli3-/- limbs becomes milder on the Alx4-/- or Zic3-/- background [30 , 31] , which suggested that the Gli3-/- polydactyly phenotype requires Alx4 or Zic3 . In contrast to these reports , loss of one allele of Gata6 enhanced the polydactyly phenotype of Gli3+/- hindlimbs . Therefore , unlike previous genetic studies , our study identified Gata6 as a negative factor for polydactyly development . Given that GLI3R prevents extra-digit formation in the anterior mesenchyme [55] , our results suggest that Gata6 cooperates with GLI3R activities . It is believed that d1 develops in a Shh-independent manner , while development of d2-d5 requires Shh [5 , 6 , 10 , 11] . Genetic manipulation of Gli3 in mice provided evidence that high levels of GLI3R in the anterior of limb buds is necessary for proper d1 development and ensuring pentadactyly [24 , 55 , 65] . Expression pattern of Pax9 , which requires high levels of GLI3R [56] , indicates that Gata6 contributes to GLI3R activities in the anterior of hindlimb buds . In particular , Pax9 was undetectable in Tcre; Gata6+/fl; Gli3+/- hindlimb buds , similar to Gata6 cKO and Gli3-/- hindlimb buds . These altered expression pattern of Pax9 correlates with ectopic digit condensation and preaxial polydactyly , and further supports the idea that Gata6 cooperate with Gli3 for proper GLI3R activities in the anterior of hindlimb buds . How does Gata6 cooperate with Gli3 ? Our data support the idea that GATA6 physically interacts with GLI3R , facilitates the nuclear localization of GLI3R , and enhances the repressor activities of GLI3R . Reduced nuclear GLI3R localization in Gata6 cKO hindlimb supports the idea that this interaction-mediated nuclear GLI3R localization would also occur in vivo . A recent study showed that Gata4 , 5 , and 6 can repress Gli-dependent reporter activation in vitro [66] . This study suggested that GATA inhibits SHH-dependent GLI activator function by protein interaction in the chick presomitic mesoderm . Based on this report and our study , GATA might modulate both GLI3R ( this study ) and SHH-dependent GLI activator [66] in a context-dependent manner . Since expression of Gata genes is reported in other Gli3-positive developing tissues , such as the branchial arch , somite and central nervous system [16 , 67 , 68] , Gata regulation of GLI3R might be a shared mechanism during the development of other organs .
Animal breeding was performed according to the approval by the Institutional Animal Care and Use Committee of the University of Minnesota . Compressed CO2 gas from a cylinder followed by cervical dislocation was the methods of euthanasia for mice . All efforts were made to minimize suffering . The mouse lines for Gata6fl [41] , Gli3- [69] and Tcre [42] were maintained on a mixed genetic background . Skeletal preparation was done as previously published [70] . Whole mount in situ hybridization was done as previously published [13] . The full-length human GATA6 construct and the human GLI3 construct were published [31 , 71] . The GLI3R construct was generated by deleting the 3’ part of full-length cDNA , and cloned into 3xFlag CMV7 . GATA6 deletion constructs were generated by PCR-based cloning and cloned in pcDNA3 . 1 or pCS2 . For in vitro analysis , cells were fixed with 4% PFA for two hours at room temperature , washed with PBS and stained with anti-Flag ( Sigma , M2 , F3165 , dilution 1:500 ) and anti-Myc tag ( Abcam , ab9106 , dilution 1:500 ) antibodies . For in vivo analysis , embryos were fixed for two hours in 4% PFA at 4C , washed with cold PBS , and cryosectioned with the OCT compound at 14 μm thickness . Sections were stained according to a standard procedure [13] without heat-induced epitope retrieval . Anti-GATA6 ( R&D Systems , AF1700 , dilution 1:400 ) and anti-GLI3R ( Clone 6F5 , dilution 1:200 ) [15 , 72] were used . Alexa fluorophore-labelled secondary antibodies were obtained from Invitrogen ( 1:1000 dilution ) . Fluorescent confocal images were obtained by using Zeiss LSM 710 laser scanning microscope system ( Carl Zeiss Microscopy ) , and analyzed using ZEN2009 software ( Carl Zeiss Microscopy ) . For subcellular localization analysis in vitro , images were acquired form six arbitrary areas from two plates . Nuclear/cytoplasmic localization of GLI3R and GATA6 was blindly evaluated in cells that were doubly transfected with GLI3R and GATA6 ( or its mutants ) except for samples that are transfected with GLI3R alone . For in vivo samples , nuclear/cytoplasmic localization of GLI3R was evaluated similarly in the anterior-proximal domain where GATA6 signals in wild-type hindlimb buds were detected . In Gata6 cKO embryos , the anterior-proximal domain , similar to wild-type embryos , was selected for GLI3R subcellular localization . The quantification was performed similar to in vitro samples . In order to clarify GATA6 localization in hindlimb bud mesenchyme , GATA6 was simultaneously detected with limb bud mesenchyme markers , such as FGF10 or DUSP6 . Wild-type embryos were fixed , washed and cryosectioned as described above . Sections were simultaneously stained by anti-GATA6 ( R&D AF1700 or Cell Signaling #5851 , dilution 1:1 , 600 ) and anti-FGF10 ( Santa Cruz , sc-7917 , dilution 1:100 ) or anti-DUSP6 ( Sigma , Clone 3G2 , dilution 1:200 ) . Sections were reacted with Alexa fluorophore-labelled secondary antibodies , and fluorescent signals were detected by Zeiss LSM 710 according to a standard procedure [13] . NIH3T3 cells in 48-well plates were transfected with the 12xGLI-binding site-TK minimum promoter-luciferase [31] with pRL-TK , GATA6 and/or GLI3R expression constructs by using Fugene6 ( Promega ) . Forty hours after transfection , cells were subjected to analysis using the Dual-Luciferase Reporter Assay System ( Promega ) . Experiments were performed in triplicate , and statistical significance was analyzed by One-way ANOVA followed by the Tukey’s comparison . HEK293T cells were transfected with expression constructs by using the standard calcium phosphate method . Cell lysates , prepared after two days , were passed through 25 gauge syringes to ensure protein extraction from the nucleus , and co-immunoprecipitation assays were performed by using Dynabeads protein G ( Invitrogen ) and anti-Flag ( Sigma , M2 , F3165 , 2μg ) or anti-Myc tag ( Abcam , ab9106 , 1 μg ) antibodies . Proteins were resolved by SDS-PAGE , transferred to PVDF membranes ( Millipore , MA , USA ) , reacted with anti-Myc tag or anti-Flag antibodies , followed by HRP goat anti-mouse or rabbit IgG , and a chemiluminescence detection . For co-immunoprecipitation assays with in vivo samples , hindlimb buds were collected from wild-type embryos at E10 . 25–10 . 5 . After pooling , the samples were lysed and subjected to co-immunoprecipitation procedures [73] using anti-GATA6 ( Cell Signaling , #5851 ) and Dynabeads protein G . The protein complex was eluted , and detected by Western using anti-GLI3 ( R&D Systems , AF3690 , dilution 1:100 ) and the PicoLUCENT PLUS HRP detection kit ( G-Bioscience ) according to the manufacturer’s instructions . For nuclear GLI3R detection by Western , anterior one third of hindlimb buds at E10 . 25–10 . 5 were collected , and the nuclear fraction was prepared after dissociating cells by using the NE-PER kit ( Thermo Fischer ) according to the manufacturer’s instructions . The nuclear extracts were analyzed by Western using anti-GLI3 ( R&D Systems , AF3690 ) and anti-Histone H3 ( Abcam , ab-1791 ) .
|
Gli3 is a major regulator of Hedgehog signaling in the limb , where Gli3 counteracts Sonic hedgehog ( Shh ) for patterning and proliferative expansion of limb progenitor cells . In the anterior limb mesenchyme , GLI3 is proteolytically processed into GLI3R , a truncated repressor form that inhibits Hedgehog signaling . In this study , we show a novel mechanism of regulation of GLI3R activities in limb buds by Gata6 , a member of GATA transcription factor family . Conditional inactivation of Gata6 in mice caused formation of an extra digit in the anterior hindlimbs , a common congenital limb malformation . This phenotype was associated with ectopic Hedgehog signaling activation , and later ectopic Shh expression , in the anterior of hindlimb buds . We show that Gata6; Gli3 compound heterozygous mutants developed anterior extradigit without ectopic Shh expression , indicating there to be an additional and prior mechanism before ectopic Shh activation that induces extradigit formation . We identified that GATA6 physically interacts with GLI3R and that the interaction facilitates nuclear localization of GLI3R and repressor activities of GLI3R . Therefore , our study identified a novel mechanism by Gata6 to regulate GLI3R activities in the anterior limb mesenchyme to prevent extra digit formation and proper development of the mammalian autopod .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"gene",
"regulation",
"developmental",
"biology",
"regulator",
"genes",
"organism",
"development",
"gene",
"types",
"embryos",
"cellular",
"structures",
"and",
"organelles",
"limb",
"development",
"embryology",
"limb",
"buds",
"gene",
"expression",
"organogenesis",
"subcellular",
"localization",
"hedgehog",
"signaling",
"signal",
"transduction",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cell",
"signaling"
] |
2016
|
Gata6-Dependent GLI3 Repressor Function is Essential in Anterior Limb Progenitor Cells for Proper Limb Development
|
The thalamus , a crucial regulator of cortical functions , is composed of many nuclei arranged in a spatially complex pattern . Thalamic neurogenesis occurs over a short period during mammalian embryonic development . These features have hampered the effort to understand how regionalization , cell divisions , and fate specification are coordinated and produce a wide array of nuclei that exhibit distinct patterns of gene expression and functions . Here , we performed in vivo clonal analysis to track the divisions of individual progenitor cells and spatial allocation of their progeny in the developing mouse thalamus . Quantitative analysis of clone compositions revealed evidence for sequential generation of distinct sets of thalamic nuclei based on the location of the founder progenitor cells . Furthermore , we identified intermediate progenitor cells that produced neurons populating more than one thalamic nuclei , indicating a prolonged specification of nuclear fate . Our study reveals an organizational principle that governs the spatial and temporal progression of cell divisions and fate specification and provides a framework for studying cellular heterogeneity and connectivity in the mammalian thalamus .
A large portion of the vertebrate brain is organized so that groups of neurons that share similar properties are arranged in aggregates called nuclei . Developmental mechanisms underlying the formation of such nuclear brain structures are much less understood compared with the mechanisms that regulate the formation of laminar structures , including the cerebral cortex and retina . The thalamus contains dozens of nuclei that play crucial roles in controlling cortical functions [1 , 2] . Most thalamic nuclei are populated by excitatory neurons that project axons to the cerebral cortex . Some of these nuclei receive substantial subcortical inputs and convey sensory and motor information to specific areas in the neocortex , whereas other nuclei receive inputs mostly from the cortex and serve as a hub for cortico-cortical communications . Thalamic nuclei can be distinguished by morphology and gene expression as well as by areal and laminar specificity of their axonal projections to the cortex . Genetic fate-mapping at the population level revealed that neurons that populate cortex-projecting nuclei are derived from a progenitor domain marked by the proneural basic helix-loop-helix ( bHLH ) transcription factors NEUROG1 and NEUROG2 ( Fig 1A , pTH-C domain; [3] ) . In addition to these “excitatory nuclei , ” the thalamus also contains “inhibitory nuclei” that are exclusively composed of GABAergic neurons , which include the intergeniculate leaflet ( IGL ) and ventral lateral geniculate ( vLG ) nucleus . Notably , recent studies identified a small progenitor domain ( rostral thalamic progenitor [pTH-R] domain ) that is located rostral to the caudal thalamic progenitor ( pTH-C ) domain during neurogenesis and expresses the proneural bHLH factors ASCL1 and TAL1 . Population-level fate-mapping studies using these markers found that pTH-R contributes to IGL and vLG nuclei ( Fig 1A; [3 , 4] ) . However , due to the complex spatial arrangement of all thalamic nuclei and the narrow time window during which neurons for these nuclei are produced [5] , little is known about how individual progenitor cells undergo divisions and fate specification and eventually produce neurons and glia that populate different thalamic nuclei . As demonstrated in invertebrate brains [7] , determination of precise cell lineage at the single-cell level is key to understanding the molecular and cellular mechanisms that underlie the formation of complex brain structures like the mammalian thalamus . In this study , we performed clonal analyses of neural progenitor cells of the thalamus using mosaic analysis with double markers ( MADM ) -based genetic lineage tracing [8 , 9] . Quantitative analysis of clone size and distribution obtained with 2 inducible Cre drivers allowed us to determine the division patterns of individual neuroepithelial and radial glial cells and their differential contributions to distinct thalamic nuclei . Furthermore , we used a third Cre driver to label more differentiated , basally dividing progenitor cells and captured the fate of their progeny generated in the last few cell divisions . These results reveal how patterning , cell division , and cell fate specification are coupled in a spatially and temporally specific manner to generate distinct thalamic nuclei during development .
To investigate the temporal progression of thalamic nucleogenesis in mice , we first performed ethynyldeoxyuridine ( EdU ) birth dating by labeling progenitor cells during the S phase of the cell cycle at embryonic day 9 . 5 ( E9 . 5 ) , E10 . 5 , E11 . 5 , E12 . 5 , E13 . 5 , and E14 . 5 . We analyzed the distribution of EdU-positive cells at postnatal day 0 ( P0 ) ( Fig 1B , S1 Fig ) . No labeling was detected in the thalamus with EdU injections at E9 . 5 . With EdU injections at E10 . 5 , robust labeling was detected in ventral medial geniculate ( MGv ) , dorsal lateral geniculate ( dLG ) , lateral posterior ( LP ) , and lateral dorsal ( LD ) nuclei . With the exception of these earliest-generated nuclei and the latest-generated , paraventricular ( PV ) nucleus , most thalamic nuclei became postmitotic between E11 . 5 and E13 . 5 . The lateral-to-medial neurogenic gradient was evident throughout the thalamus , consistent with previous studies [5 , 10] . Some nuclei ( e . g . , dLG , LP , LD ) were generated over at least 48 h ( E10 . 5 to E12 . 5 ) , whereas others ( e . g . , centromedian [CM] , reuniens/rhomboid [Re/Rh] , mediodorsal [MD] ) were generated within 1 d around E13 . 5 . Thus , thalamic neurogenesis occurs over a short span of embryonic development and likely involves a complex pattern of division and fate specification at the level of individual progenitor cells . In order to determine the temporal and spatial dynamics of neural progenitor cells in the thalamus , we performed genetic lineage tracing at the clonal level by using the MADM system ( S2A Fig; [8 , 9] ) . When the frequency of Cre-mediated recombination is kept low by titrating the tamoxifen dosage , the sparse Green/Red ( G/R ) clones generated by “G2-X”–type segregation ( see S2A Fig for definition ) would allow us to determine the division pattern and fate of the 2 daughter cells that are derived from the original progenitor cell undergoing recombination . The yellow ( Y ) clones produced by “G2-Z”–segregation of chromosomes can be defined as hemiclones because their sister clones are not labeled . In this study , 3 inducible Cre driver lines , Gli1CreERT2 , Olig3CreERT2 , and Neurog1CreERT2 were combined with MADM-11 reporters . Both Gli1 and Olig3 are expressed in neuroepithelial and radial glial cells in the thalamus [6] . Gli1 is a direct target gene of Sonic Hedgehog signaling pathway and is expressed highly in the rostral part of the thalamus as well as in other regions in the forebrain , including the prethalamus and hypothalamus . OLIG3 is expressed in the entire thalamic ventricular zone and in the zona limitans intrathalamica ( ZLI ) , a secondary organizer located immediately rostral to the thalamus ( Fig 1A ) , whereas its expression is minimal in other forebrain regions ( [3]; see Fig 1A ) . In the thalamus , NEUROG1 is largely excluded from intracellular domain of NOTCH ( Notch-ICD ) –expressing neuroepithelial and radial glial cells and is expressed in basally dividing progenitor cells in the pTH-C domain [11] . To validate the recombination patterns of each CreER driver , we performed a population analysis of reporter mice . The Rosa-H2B-GFP reporter line crossed with the Gli1CreERT2 driver , and the Ai6 ZSGreen reporter crossed with Olig3CreERT2 or Neurog1CreERT2 driver , showed labeling of progenitor cells within the thalamic primordium 1 d after tamoxifen administration ( E10 . 5 for Gli1 , E11 . 5 for Olig3 , and E12 . 5 for Neurog1 ) . At E18 . 5 , labeling was detected throughout the thalamus ( Fig 2A ) . We then crossed each CreER driver with MADM-11GT/GT mice , and the CreER+/−; MADM-11GT/GT mice were crossed with MADM-11TG/TG mice . One day after the tamoxifen administration , MADM brains for all 3 CreER drivers had G/R and Y clones ( see S2A Fig for definition ) in thalamic primordium ( Fig 2B ) , and the apical and basal processes of the labeled cells in Gli1CreERT2 MADM brains were positive for Nestin , a neuroepithelial and radial glial cell marker ( S2B Fig ) . Many clones were produced in the thalamus of Gli1CreERT2 MADM brains with tamoxifen administration at E9 . 5 . In contrast , Olig3CreERT2 MADM clones were obtained only after tamoxifen administration at E10 . 5 or later . This is likely due to the difference in the onset of the expression of the Cre drivers . In addition , expression of Gli1 has a rostral-high , caudal-low gradient within the thalamus [6] , and Gli1 MADM clones tended to be concentrated in the rostral part of the thalamus . Thus , using both Gli1 and Olig3 CreER drivers would allow us to fully trace the lineage of spatially and temporally divergent neuroepithelial and radial glial cell populations in the thalamus . In addition , the Neurog1 CreER driver would allow labeling of progenitor cells at a later stage of neurogenesis . Both Gli1 and Neurog1 CreER drivers labeled progenitor cells widely in the forebrain , whereas the Olig3 driver was specific to the thalamus ( Fig 2B ) . By adjusting the dosage of tamoxifen for each CreER driver , we achieved a low probability of labeling a pair of progeny generated by a single recombined progenitor cell in the thalamus ( Fig 2C ) . The percentage of hemispheres that had at least one G/R or Y clone 1 d after the Cre activation was less than 25% ( 22% for Gli1CreER , 19% for Olig3CreER , and 22% for Neurog1CreER ) , and almost no hemispheres had 2 clones ( 2% for Gli1CreER , 3% for Olig3CreER , and 5% for Neurog1CreER ) , which validated our methodology for clonal analysis of thalamic progenitor cells . For the Olig3CreER driver , we analyzed 186 hemispheres ( 27 litters ) of CreER+ E18 . 5 mice and 112 hemispheres ( 23 litters ) of CreER+ P21 mice . Out of these brains , we detected 72 and 22 G/R clones , respectively . For the Gli1CreER driver , we analyzed 194 hemispheres ( 37 litters ) of CreER+ E18 . 5 mice and 112 hemispheres ( 29 litters ) of CreER+ P21 mice . Out of these brains , we found 27 and 24 G/R clones , respectively . For the Neurog1CreER driver , we analyzed 83 hemispheres ( 17 litters ) of CreER+ E18 . 5 mice and found 55 G/R clones . In order to avoid analyzing potentially mixed clones that arise from 2 or more recombined cells , we excluded Olig3CreERT2 E18 . 5 brains from 2 litters in which every hemisphere had fluorescently labeled cells in the thalamus . A major advantage of using the MADM system for lineage tracing is that it allows for analysis of the division pattern of the recombined single progenitor cell based on the numbers of green cells and red cells in each G/R clone . In early embryonic neocortex , neuroepithelial cells undergo symmetric proliferative divisions to increase the pool size of progenitor cells . As neurogenesis starts , progenitor cells divide asymmetrically and produce a self-renewed radial glial cell and a more differentiated cell , either a postmitotic neuron or an intermediate progenitor cell ( IPC ) . IPCs in embryonic neocortex express the T-box transcription factor TBR2 and divide symmetrically in basal locations away from the surface of the lateral ventricle [12–15] . Although TBR2 is not expressed in embryonic thalamus , basally dividing cells that express the proneural bHLH transcription factors NEUROG1 and NEUROG2 do exist [11] . Because neocortical IPCs also express NEUROG2 [16] , we hypothesized that Neurogenin-expressing , basally dividing cells in the thalamus are IPCs . In E18 . 5 Neurog1CreERT2 MADM brains , 78% of G/R clones had 4 or fewer cells ( Fig 3A ) , and all labeled cells had a neuronal morphology . Thus , a majority of Neurog1-expressing thalamic progenitor cells are intermediate neuronal progenitor cells with a limited proliferative capacity . In contrast to the small size of Neurog1 MADM clones , most clones in Gli1 and Olig3 MADM brains had more than 10 cells ( Fig 3B ) . Large G/R clones with more than 4 cells in both green and red progeny are likely to originate from a recombined progenitor cell that produced 2 radial progenitors , not Neurog1-expressing progenitor cells . We defined these clones as symmetric proliferative ( Fig 3C; “Sym Pro” ) . G/R clones in which either green or red hemiclone contained 4 or fewer cells were defined as asymmetric self-renewing/neurogenic , which likely produced one radial progenitor and an IPC , or one radial progenitor and a postmitotic neuron ( Fig 3C , “Asym Neuro” ) . G/R clones with 4 or fewer cells in both hemiclones were defined as symmetric neurogenic ( Fig 3C , “Sym Neuro” ) . Both Gli1 and Olig3 clones exhibited a combination of the three division patterns described in Fig 3C ( Fig 3D ) . When recombination was induced in Gli1CreER mice at E9 . 5 , 57% of the clones were of the symmetric proliferative type , indicating that a majority of recombined progenitor cells at E9 . 5 were neuroepithelial cells undergoing symmetric expansions . In Olig3CreER mice , this fraction was 42% at E10 . 5 and 22% at E11 . 5 . In addition , at E11 . 5 , 61% of the clones were of the symmetric neurogenic type . The average clone size was smaller for a later Cre activation ( 22 . 5 for E10 . 5 and 9 . 5 for E11 . 5 ) . The size of the minority hemiclones in asymmetric clones ( Fig 3E ) and the composition of symmetric neurogenic clones ( Fig 3F ) both showed a large variation . These results together indicate a transition of cell division mode in the thalamus from symmetric proliferative to asymmetric to symmetric neurogenic , similar to the neocortex [17] . To investigate the spatial organization of clonal units during thalamic nucleogenesis , we performed 3D reconstructions of serial brain sections containing all labeled cells ( Fig 4A and S3 Fig ) [18] and assigned each labeled cell to a specific thalamic nucleus using a published atlas [19] , data from our previous study [20] , and a newly generated custom atlas ( S4 Fig ) . Asymmetric G/R clones generated 11 . 9 ± 0 . 87 cells ( Fig 4B ) that populated 3 . 3 ± 0 . 26 nuclei ( Fig 4C ) . This demonstrates that each thalamic radial glial cell produces neurons that populate multiple nuclei after asymmetric divisions . The average number of nuclei populated by progeny of asymmetrically dividing progenitor cells was significantly smaller than the number for symmetric proliferative clones ( Fig 4C ) , suggesting that the transition from symmetric proliferative to asymmetric division restricts the nuclear fate potential . Within asymmetric clones , “minority” hemiclones ( i . e . , G or R hemiclones containing 4 or fewer cells ) exhibited significantly more fate restriction than “majority” hemiclones based on the average number of populated nuclei ( Fig 4C ) . However , some minority hemiclones still populated more than one nuclei , implying that the fate of thalamic nuclei may not be fully restricted even in terminally dividing cells . The 3 principal sensory nuclei ( ventral posterior [VP] , dLG , MGv ) are among the earliest-born in the thalamus ( Fig 1B ) . Approximately 61% of majority hemiclones were composed of both principal sensory nuclei and other later-born nuclei , whereas only 12% of them had cells only in principal sensory nuclei . Forty-four percent of minority hemiclones showed an exclusive contribution to principal sensory nuclei ( Fig 4D ) . This suggests that thalamic radial glial cells sequentially generate neurons that populate different nuclei as they undergo asymmetric divisions . This pattern is reminiscent of the sequential generation of deep- and upper-layer neurons from individual radial glial cells in the embryonic neocortex [17] . We next tested whether individual radial glial clones contribute to specific subsets of thalamic nuclei . To analyze our results computationally , we generated a data matrix in which rows and columns represented individual G/R clones and thalamic nuclei , respectively . The numeric value in each cell of the matrix was computed ( see Materials and methods ) so that it indicated the frequency of progeny in a given clone that was located in each thalamic nucleus . This configuration is analogous to the matrix used for single-cell RNA sequencing analysis , in which rows , columns , and values represent individual cells , genes , and expression levels , respectively . We then applied principal component analysis ( PCA ) or unsupervised hierarchical clustering analysis for classification of thalamic MADM clones . PCA of asymmetric G/R clones identified 5 clusters that show distinct patterns of nuclear contribution ( Fig 5A ) . Three of the 5 clusters had major contributions to principal sensory nuclei ( yellow , green , and red clusters as shown in Fig 5A ) . Of these , the yellow cluster contained clones that contributed to principal sensory nuclei and other nuclei , including ventromedial ( VM ) and Re/Rh . Clones in the green cluster included cells in principal sensory nuclei and the medial ventral field ( S4C Fig; see Materials and methods for definition ) . VM and Re/Rh and the medial ventral field are all located medial to principal sensory nuclei , and neurons in these nuclei are born later than those in principal sensory nuclei ( Fig 1B ) . In 7 out of 8 clones in the green cluster , the minority hemiclones contributed exclusively to principal sensory nuclei . This result suggests that asymmetrically dividing thalamic radial progenitor cells first generate early-born , principal sensory nuclei and then late-born , medially located nuclei . It also suggests a strong lineage relationship among the 3 principal sensory nuclei , which share many properties , including laminar and areal patterns of axon projections to primary sensory areas of the neocortex [1 , 21] , gene expression [22 , 23] , and patterns of propagating spontaneous activity during embryogenesis [24] . The other two ( blue and grey ) clusters included clones that have few cells in principal sensory nuclei ( Fig 5A ) . All the clones in the grey cluster contributed to the ventrolateral ( VL ) nucleus . The remaining nuclei represented in this cluster were diverse and included CM/intermediodorsal ( CM/IMD ) , MD , and Re/Rh . Many of these nuclei were located more dorsally and caudally within the thalamus compared with principal sensory nuclei , demonstrating that nuclei that are close to each other in physical location are likely to share the same lineage . This result provides single-cell-based evidence consistent with our previous population-based study , in which gradients of gene expression within the pTH-C domain corresponded to preferential generation of progeny along the rostro-ventral to caudo-dorsal axis of the thalamus [3] . The blue cluster contained clones that were relatively more diverse . Here , 2 clones contained cells in IGL and either MGv or VP . The IGL nucleus receives axonal projections from non-image-forming retinal ganglion cells and is known to be derived from the pTH-R domain at the population level [3 , 4 , 25 , 26] . The surprising coexistence of pTH-C-derived cells ( in principal sensory nuclei ) and pTH-R-derived cells ( in IGL ) in the same clones suggests that , during early stages , some progenitor cells might still be uncommitted to either the pTH-C or the pTH-R fate . The existence of clone clusters suggests that some thalamic nuclei are more likely to share the cell lineage than other nuclei . In order to assess the clustering of clones and the “lineage distance” between nuclei , we performed unsupervised hierarchical clustering analysis and generated a heatmap . Because a majority of the clones that contributed to late-born nuclei were of the symmetric proliferative type , we included both asymmetric and symmetric G/R clones in this analysis in order to cluster the entire set of thalamic nuclei . We found that certain groups of nuclei were more likely to originate from common progenitor cells at E18 . 5 ( Fig 5B ) . Examples of additional clusters of nuclei identified in this analysis include ( 1 ) VM , VL , CM/IMD , and Re/Rh; ( 2 ) LD and centrolateral/paracentral ( CL/PC ) ; ( 3 ) LP , MD , and PV; ( 4 ) MGv , dLG , Po , medial ventral field , and parafascicular ( PF ) ; and ( 5 ) vLG and IGL . These clusters of nuclei are generally arranged radially and are located at similar rostro-caudal and dorso-ventral positions within the thalamus . The presence of lineage-based clusters of nuclei shown here suggests that the distribution of progeny generated from thalamic progenitors appears to follow distinct nuclei-specific patterns . To determine whether the clustering of clones based on cell lineage reflects the distinct positions of starting progenitor cells within the embryonic thalamus , we took advantage of the finding that some clones , especially the ones labeled at E9 . 5 in the Gli1CreER line , retained a cell at the ventricular surface with a long radial process spanning the entire thalamic wall . In order to use the locations of these cells as a proxy for the position of the founder progenitor cell at early embryonic stages , we defined 4 domains of these radial glial cells within the E18 . 5 thalamus ( Fig 6A ) . “Dorsal-rostral ( DR ) ” corresponded to the most rostral domain in relatively dorsal sections , whereas the “dorsal-middle ( DM ) ” and “dorsal-caudal ( DC ) ” domains were located progressively more caudally . The “ventral ( V ) ” corresponded to the domain in more ventral sections near the basal plate . Two-dimensional shapes of clones derived from each of the 4 domains are shown in Fig 6B . A hierarchical clustering analysis revealed that clones containing radial glial cells in the DR and DM domains heavily contributed to principal sensory nuclei ( Fig 6C ) ; many of these clones contained cells in both principal sensory nuclei and more medially located nonsensory nuclei , including VM , Re/Rh , and the medial ventral field . Some DR clones contained cells in both “excitatory nuclei” composed of cortex-projecting neurons and “inhibitory nuclei” containing GABAergic neurons . Thus , DR and DM clones appear to correspond to those belonging to the yellow , green , and red clusters in Fig 5A . DC and V clones had a large contribution to dorsally located , nonsensory nuclei and a small contribution to principal sensory nuclei except MGv . These results suggest that thalamic nuclei that are related in ontogeny originate from similar locations within the embryonic thalamus and support the model wherein individual thalamic radial glial cells form a radial column of progeny that spans a defined subset of thalamic nuclei . Out of the 46 G/R P21 clones obtained from Gli1CreER and Olig3CreER drivers , 17 clones were composed of both neurons and glia . Clones containing glia were distributed in multiple PCA clusters ( glia-containing clones in asymmetric G/R clones are indicated by an asterisk in Fig 5A ) , suggesting that the ontogeny of thalamic glial cells is heterogeneous and that gliogenesis is not restricted to specific thalamic progenitor cell populations . Many postnatal MADM brains contained glial cells . The average number of glial cells was 24 . 7 ± 11 . 4 cells in the P21 glia-containing clones . Although the number of glial cells did not correlate with the number of neurons in the P21 clones , our data suggested a significant linear correlation between the number of glia and total cells ( S5A Fig ) . Among all of the postnatal clones , 3 were composed mostly of glial cells , suggesting the existence of glial lineage-specific radial glial precursors . The distribution analysis of glial number in the P21 clones showed that the 3 clones with a large number of glia were outliers ( S5B Fig ) . After removing these clones , we did not find any correlation between the glial number and the neuronal number or between glial number and total cell number in glia-containing clones ( S5C Fig ) . The Gli1 clones contained more glia than Olig3 clones ( S5D Fig ) , which might be due to the earlier labeling of radial glial precursor in the Gli1 clones ( at E9 . 5 ) than in Olig3 clones ( E10 . 5 or E11 . 5 ) . Bipotent radial glial progenitors producing neurons and astrocytes ( N+A ) generated more neurons than unipotent progenitors producing neurons only ( S4E Fig ) . We did not collect sufficient clones derived from bipotent progenitors producing neurons and oligodendrocytes ( N+O ) and tripotent progenitors producing all 3 types of neural cells ( N+A+O ) to compare with unipotent progenitors . The majority of symmetric proliferative and asymmetric neurogenic P21 clones contained both neurons and glia ( S5F Fig ) . Together , our data suggest that a majority of radial glial cells in different progenitor domains are multipotent and generate glial cells by P21 . Analysis of asymmetric clones of Gli1CreER and Olig3CreER MADM brains revealed minority hemiclones that included cells in more than one thalamic nuclei ( Fig 4C ) . This suggested that progenitor cells undergoing their last few rounds of divisions are still not yet specified to the fate of a single nucleus . As described above ( Fig 3A ) , a large majority of Neurog1CreER MADM clones contained 4 or fewer cells with a neuronal morphology , indicating that Neurog1-expressing thalamic progenitor cells are IPCs . Taking further advantage of the MADM system , we analyzed the division patterns and fates of these cells . At the population level , timed activation of Cre recombinase at E11 . 5 resulted in expression of the ZSGreen reporter in the entire thalamus except the medial-most region including PV , CM , and Re/Rh nuclei ( Figs 2A and 7A ) . These medial nuclei began to be labeled by ZSGreen with the Cre activation at E12 . 5 ( Fig 7A ) . This is 1 d before they become robustly labeled through EdU birth dating ( Fig 1B ) . Similarly , VL starts to be labeled by ZSGreen at E11 . 5 and by EdU at E12 . 5 , indicating that , at least for the above thalamic nuclei , progenitor cells express Neurog1 before they undergo the final division . We next performed clonal lineage tracing with MADM using the Neurog1 CreER driver . Activation of the Cre recombinase at E11 . 5 , E12 . 5 , E13 . 5 , or E14 . 5 followed by analysis at E18 . 5 revealed that a vast majority of MADM clones contained 4 or fewer cells ( 43 out of 55 G/R clones; Fig 3A ) . Many of these clones ( 13 out of 36 G/R clones with 2–4 cells ) contained neurons that populated more than one thalamic nuclei ( Fig 7B and 7C ) . Clones covering multiple nuclei were abundant with E11 . 5 or E12 . 5 activation ( 9 out of 30 clones ) , but there were some with the activation at E13 . 5 or E14 . 5 ( 4 out of 13 clones; Fig 7D ) . Four of the multi-nuclei clones spanned 2 or all 3 principal sensory nuclei , while others included late-born nuclei like PV and Re/Rh . By taking advantage of the dual-color labeling of the progeny in the MADM system , we also captured 58 terminal divisions for which we could determine whether the last division generated cells in the same nucleus or 2 different nuclei . Among these 58 final divisions , 13 generated cells in 2 different nuclei , of which 4 were in dLG and VP and 1 was in MGv and VP ( Fig 7E ) , consistent with the ontogenic proximity of these nuclei revealed in the analysis of Gli1CreER and Olig3CreER MADM samples ( Fig 5 ) . For individual nuclei , we found that 4 out of 8 terminal divisions that produced a neuron in dLG also produced a neuron in another nucleus ( all in VP ) . Out of 13 terminal divisions that produced a neuron in VP , 7 produced a neuron in another nucleus ( 4 in dLG , 1 in MGv , 1 in LP , 1 in CL/PC ) . Thus , dLG and VP nuclei are particularly close to each other in cell lineage , and their fates may not diverge at least until the final progenitor cell division . In contrast , only 2 out of 9 terminal divisions that produced a neuron in MGv also produced a neuron in another nucleus ( 1 each in VP and LP ) , suggesting that the MGv lineage is largely specified by the time of the final division of Neurog1-expressing IPCs . Taken together , analysis of Neurog1CreER clones demonstrated that the thalamus contains IPCs and that many of these cells are not yet specified for the fate of a specific nucleus . Even the last division of such progenitor cells can generate cells in 2 distinct nuclei , indicating prolonged mechanisms of nuclear fate specification in developing thalamus .
Thalamic neurogenesis occurs over a short period of time during embryonic development . In mice , neurons in the earliest-born thalamic nuclei , including dLG and MGv , are first generated at around E10 . 5 , while later-born nuclei are born by E13 . 5 with an overall “outside-in” pattern [5] ( Fig 1B ) . Due to the short time span of thalamic neurogenesis and the spatially complex arrangement of dozens of nuclei , it was previously unknown whether each neural progenitor cell produces a limited number of progeny of a single nuclear identity or sequentially generates a large number of neurons that populate multiple nuclei with different birthdates . With the MADM approach , we defined clones that originated from an asymmetrically dividing radial glial cell and found that each asymmetric clone contains an average of 11 . 9 cells that populated 3 . 3 thalamic nuclei . This result provides a general principle of thalamic nucleogenesis in which a single radial glial cell produces many neurons that populate multiple nuclei . By distinguishing the differentially colored hemiclones derived either from a self-renewed radial glia or a differentiated daughter cell , we found that many radial glial cells labeled at E9 . 5 , E10 . 5 , or E11 . 5 first produced neurons that populate early-born , principal sensory nuclei , including dLG , VP , and MGv . Then , in many clones , the same lineage produced neurons that occupy more medially located , later-born nuclei such as Re/Rh and VM ( Fig 1B ) . A recent study by Gao and colleagues [17] analyzed neurogenesis in the neocortex using the MADM approach and found that each asymmetrically dividing radial glia sequentially produces an average of 8 . 4 cells that populated multiple layers in an inside-out fashion . This indicates that , despite the difference between nuclear versus laminar organizations of these two structures , the thalamus and neocortex might share a basic strategy in cell division and differentiation . In order to elucidate the heterogeneity among individual cell lineages , we performed PCA of asymmetric MADM clones obtained with Gli1CreER and Olig3CreER drivers . This revealed several clusters of clones , each of which contributed to a shared set of thalamic nuclei ( Fig 5A ) . Out of the 5 clusters identified , 3 showed significant contributions to principal sensory nuclei—dLG , VP , and MGv . One of these clusters ( yellow ) was composed of principal sensory nuclei and later-born nuclei that lie medial to principal sensory nuclei , indicating that some radial glia are capable of sequentially generating cells that contribute to different sets of nuclei located along the radial axis . The green cluster included few nuclei other than the principal sensory nuclei . Thus , some radial glial lineages may be depleted once the early-born sensory nuclei are generated . The third such cluster ( red ) had a strong contribution to MGv , the principal auditory nucleus . These clones contributed to a different set of late-born nuclei , including PF and Po , from the clones of green and yellow clusters . The remaining 2 clusters of clones had more contributions to caudo-dorsally located nuclei than principal sensory nuclei . Thus , although the principal sensory nuclei generally shared a progenitor cell lineage , the actual pattern of nuclear contribution differed between clones . Clustering of nuclei based on contributions by both symmetric and asymmetric G/R clones revealed additional nuclear groups that are located more caudo-dorsally than principal sensory nuclei ( Fig 5B ) . Recently , Shi and colleagues ( 2017 ) reported MADM-based lineage tracing in the postnatal mouse thalamus using the NestinCreER driver [27] . They used an unsupervised hierarchical clustering method to cluster both MADM clones and thalamic nuclei based on the distribution of the cells of each clone in different nuclei . The patterns of clone clusters are generally similar to ours ( Fig 5A and 5B ) . They found that the 3 principal sensory nuclei and the motor nuclei ventroanterior ( VA ) /VL are ontogenetically segregated from 3 other clusters ( MD/Po/LP/CM etc . ; VM/Re; and anteromedial ( AM ) /anteroventral ( AV ) /PV/LD , etc . ) . Sherman and colleagues [2 , 28] have proposed that the 3 principal sensory and VA/VL nuclei in the thalamus receive strong subcortical input and relay the information to primary sensory and motor areas ( “first-order” nuclei ) . Therefore , Shi and colleagues concluded that the first-order and higher-order nuclei , which receive most driver input from the cortex , are ontogenetically segregated . Our study found clusters of nuclei ( Fig 5B ) that are similar to those detected by Shi and colleagues . However , there were some differences in the results as well as interpretation . First , our analysis did not find a strong clustering of VL with principal sensory nuclei . Instead , VL more closely segregated with CM , Re , and VM ( Fig 5A and 5B ) , indicating that the ontogenetic relationship of thalamic nuclei does not necessarily follow the distinction between the first-order versus higher-order nuclei . In addition , another first-order nucleus , the anterior nuclear complex—particularly the anterodorsal ( AD ) nucleus , which receives subcortical driver input from the mammillary body [2 , 28 , 29]—was ontogenetically distant from the principal sensory nuclei both in our study and in Shi and colleagues [27] . Thus , we propose that the ontogenetic relationship between thalamic nuclei are defined primarily by their rostro-caudal and dorso-ventral positions instead of functions . Patterns of gene expression [20 , 22 , 23 , 30] , laminar specificity of efferent projections to the cortex [21] , and sources and types of afferent projections [31] are likely dependent on both cell lineage and postmitotic , potentially extrinsic mechanisms . In addition to simply clustering clones and nuclei , we attempted to correlate the locations of original progenitor cells with their nuclear contributions by analyzing clones that retained radial glial cells at E18 . 5 ( Fig 6 ) . This analysis revealed that the location of the original radial glia predicts the set of nuclei it populates . For example , rostrally located clones ( shown as “DR” and “DM” clones in Fig 6A ) strongly contributed to principal sensory nuclei , whereas more caudally located , DC clones contributed to caudo-dorsal nuclei , including Po , PF , MD , and CM/IMD . Some of the most rostrally located ( DR ) clones included both principal sensory nuclei and IGL/vLG , the 2 nuclei that are mostly populated by inhibitory neurons . This was unexpected because previous population-based lineage studies found that IGL and vLG nuclei originate from a distinct progenitor domain , pTH-R—not pTH-C , which contributes to principal sensory nuclei [3 , 4 , 6 , 26] . Although we did not determine whether the neurons in pTH-C-derived principal sensory nuclei that shared the lineage with those in IGL/vLG are glutamatergic or GABAergic , there are almost no GABAergic neurons in VP and MGv in mice , and those in dLG are derived from the midbrain and appear only postnatally [25] . In addition , we observed this type of clone not only at E18 . 5 but also at P21 , making it unlikely that all the neurons in the pTH-C domains in clones of this cluster at E18 . 5 were still en route to their final destinations in pTH-R-derived nuclei . Thus , it is likely that some clones do contribute to both glutamatergic pTH-C nuclei and GABAergic pTH-R nuclei . One possible mechanism for this is that some early progenitors are still uncommitted for either the pTH-C or pTH-R fate . This is consistent with the fact that the pTH-R domain emerges only after neurogenesis starts in the thalamus [6] . The location of the originating radial glia was not only correlated with the set of nuclei it populates but also the temporal patterns of cell division . Most of the clones that contained caudo-dorsally located , late-born nuclei such as MD , CM , or PV nuclei were found to be of the symmetric proliferative type ( 26 out of 32 clones containing cells in these nuclei were of symmetric type ) , demonstrating that , upon early Cre activation , progenitor cells that produce neurons of these nuclei are still expanding the pool size by symmetric proliferative divisions ( S6A Fig ) . Analysis of clones containing a residual radial glia ( Fig 6 ) indicated that cells of MD , CM , and PV nuclei are derived from caudo-dorsally located progenitor cells . In contrast , progenitor cells in the rostral part contributed to principal sensory nuclei , many of which underwent an asymmetric division immediately after the recombination at E9 . 5 to E11 . 5 ( 28 out of 77 clones containing cells in these nuclei were of asymmetric type ) ( S6A Fig ) . Our results indicate that the temporal pattern of transition from symmetric to asymmetric divisions differs among progenitor domains within the thalamus , where the transition occurs earlier in rostro-ventrally located progenitor cells than in the caudo-dorsal cells and produces a cohort of earlier-born nuclei . Our previous study showed that , similar to the neocortex , the thalamus has a large population of progenitor cells dividing basally away from the surface of the third ventricle [11] . However , the division patterns and the fate of these cells had remained unknown . By using the Neurog1CreER driver for clonal lineage analysis of these basal progenitor cells in the thalamus , we found that most of the labeled clones contained 4 or fewer cells of neuronal morphology . This demonstrates that Neurog1-expressing basal progenitor cells in the thalamus are neuron-generating IPCs . The existence of IPCs fits the model of sequential neurogenesis in the thalamus , in which each radial glial cell divides a limited number of times during a short neurogenic period and generates IPCs , and that these IPCs divide once or twice to expand the neuronal output . In the MADM study by Gao and colleagues [17] , the average size of asymmetric clones in the embryonic neocortex was 8 . 4 when they were defined as those with 3 or fewer G or R cells . Using the same criteria , the number for the embryonic thalamus in the current study was 11 . 4 , which indicates that the asymmetrically dividing radial glia in the thalamus produce more progeny within a shorter period of neurogenesis than the neocortical counterpart . Many of the Neurog1CreER MADM clones populated 2 to 3 different thalamic nuclei , demonstrating that each round of radial glia division provides a pool of neurons that share similar birthdates and sometimes populate multiple nuclei . This is an efficient strategy to produce a large variety of neuronal types with a limited number of radial glial cells in a short period of time ( summary schematics in S6B Fig ) . We further took advantage of the dual-color labeling of the progeny of IPCs and determined whether the terminal division of these cells produced neurons in the same or different thalamic nuclei . Although 78% of the final divisions resulted in neurons in the same nucleus , the remaining 22% produced neurons in 2 different nuclei . The most frequent pair of these nuclei was VP and dLG , the 2 principal sensory nuclei that share patterns of gene expression [22 , 23 , 32] and functional interactions [24] . Thus , the last division of thalamic IPCs may not be strictly symmetric , producing 2 neurons that have different nuclear fates . Alternatively , the division could be symmetric , but postmitotic mechanisms determine the fate of the 2 daughter neurons . Further studies will be needed to distinguish these possibilities .
All animal procedures used in this study were performed in accordance with the protocol approved by the Institutional Animal Care and Use Committee of Johns Hopkins University School of Medicine and the University of Minnesota ( protocol number: 1403-31417A ) . Gli1CreERT2 [33] , Olig3CreERT2 [34] , and Neurog1CreERT2 [35 , 36] mutants were found to provide an optimal labeling of thalamic progenitor cells at a clonal level when combined with the MADM-11 system [8] . We bred Gli1CreERT2/+; MADM-11GT/GT mice , Olig3CreERT2/+; MADM-11GT/GT mice or Neurog1CreERT2/+; MADM-11GT/GT mice with MADM-11TG/TG mice . A single dose of tamoxifen ( 132 mg/kg body weight for Gli1CreERT2 , 24 mg/kg or 34 mg/kg for Olig3CreERT2 and Neurog1CreERT2 ) was administered to pregnant females ( intraperitoneal injection for Gli1CreERT2 and oral gavage for Olig3CreERT2 and Neurog1CreERT2 ) at various time points ( embryonic day 9 . 5 ( E9 . 5 ) for Gli1CreERT2 , E10 . 5 or E11 . 5 for Olig3CreERT2 and E11 . 5 to E14 . 5 for Neurog1CreERT2 ) . Embryos or postnatal pups were fixed and frozen . Each Cre driver line was also bred with H2B-GFP ( for Gli1CreERT2/+ ) or Ai6 ZSGreen reporter ( for Olig3CreERT2 and Neurog1CreERT2/+ ) mice for a population-based lineage tracing with tamoxifen administration ( 60 mg/kg body weight for Olig3CreERT2 and Neurog1CreERT2/+ , 132 mg/kg body weight for for Gli1CreERT2/+ ) . For the birth dating study , EdU ( Invitrogen or Carbosynth ) was injected intraperitoneally ( 50 mg/kg body weight ) into pregnant CD1 female mice at E9 . 5 , E10 . 5 , E11 . 5 , E12 . 5 , E13 . 5 , or E14 . 5 , and pups were perfusion fixed on the day of birth ( P0 ) . Serial coronal brain sections ( 40 μm in thickness ) through the entire forebrain were cut by a cryostat or a sliding microtome and were immunostained with the following antibodies: anti-GFP ( 1:200 or 500; goat; Rockland ) , anti-GFP ( 1:500; chicken; Aves Labs ) , anti-RFP ( 1:200 or 1 , 000; rabbit; Rockland; to detect tdTomato ) , anti-Nestin ( 1:500; chicken; Aves Labs ) , anti-goat/chicken Cy2 , anti-rabbit Cy3 , and anti-goat Cy5 ( 1:200 or 500; donkey; Jackson ImmunoResearch ) . EdU was visualized on cryosections in the detection solution ( 5 uM Sulfo-Cy3 azide [Lumiprobe] , 0 . 1 M Tris pH 7 . 5 , 4 mM copper sulfate , 100 mM sodium ascorbate ) for 30 min after permeabilization in 0 . 5% Triton-X100 for 30 min . Consecutive sections covering individual clones were imaged using Zeiss LSM 710 confocal microscope ( Carl Zeiss ) . For 3D reconstruction , optical stacks from the entire diencephalon were serially aligned along the rostro-caudal axis using Reconstruct 1 . 1 . 0 ( J . C . Fiala , NIH ) , followed by import into Imaris ( Bitplane ) for further analysis . We adopted the axial nomenclature used in the prosomeric model of forebrain organization ( [3 , 37]; Fig 1A ) . In this system , the rostral border of the thalamus is the ZLI , and in the frontal view of cross sections , the bottom part ( closer to the ZLI ) is ventral-rostral and the top part ( closer to the pretectum ) is dorsal-caudal . Axial nomenclature of E18 . 5 sections ( e . g . , Fig 6A ) also follows this rule in order to retain consistency . For identification of thalamic nuclei within tissue sections , we referred to “Atlas of the Developing Mouse Brain at E17 . 5 , P0 and P6” ( [19] and [20] ) . Furthermore , we performed in situ hybridization [3] on E18 . 5 and P21 sections of wild-type brains using the same orientation and thickness as used for the MADM brains and generated a custom atlas ( S4A and S4B Fig ) . The mRNA probes for Gbx2 , Calb2 , RORα , Foxp2 , Igsf21 , Slc6a4 , and Zyx were used to label thalamic nuclei . At E18 . 5 , Gbx2 was expressed in PV , MD , CL , CM , Re/Rh , Po , LP , VM , MGv , and the medial ventral field ( shown by a white asterisk ) , which is not labeled in the available atlas [19] or book [1] . Because this region is marked by ZSGreen in Olig3CreERT2: Ai6 reporter mice at E18 . 5 ( S4C Fig ) , we consider it to be thalamus derived and included it in the analysis; Calb2 was expressed in PV , LD , CL , CM , Re/Rh , Po , LP , and the medial ventral field; RORα was expressed in AD , MD , dLG , VP , and MGv; Foxp2 was expressed in PV , LD , CM , Re/Rh , MD , Po , LP , and PF; Isgf21 was expressed in AD , AV , AM , PV , LD , MD , Po , LP , VP , dLG , MGv , and PF; Slc6a4 was expressed in VP , dLG , and MGv; and Zyx was expressed in AV , PV , LD , MD , VL , VM , CM , Re/Rh , Po , and PF . At P21 , expression patterns remained essentially unchanged from E18 . 5 for all markers except RORα , which was expressed in an additional set of nuclei including PT , AD , AV , AM , LD , VL , VM , Po , LP , and MGd . Each clone was first assigned a vector of values representing its cellular contribution to each nucleus ( nucleus X , number of labeled cells in nucleus X ) . Analogous to analysis of large datasets of gene expression in single-cell RNA sequencing in which each cell has a corresponding vector of values based on relative expression levels for each gene ( gene X , expression level of gene X ) , we performed logarithmic transformation , followed by PCA and unsupervised hierarchical clustering analysis . For PCA , the clonal distribution data were rescaled such that the average number of cells in each nucleus was centered to 0 with principal components as normalized eigenvectors of the covariance matrix of the cellular distributions . The clones were then ordered according to their contribution to the variance in the dataset . For unsupervised hierarchical clustering analysis , we calculated Euclidean distance among clones for plotting dendrograms and heatmaps . Because there were no negative values in the dataset , the Canberra distance was further analyzed for hierarchical clustering with Ward’s linkage , which yielded similar results . Computational analysis of MADM clones was performed using Excel and R software . GraphPad Prism version 5 . 0 was used for other statistical tests . In Fig 4C , statistical significance was assessed by two-tailed unpaired Student t tests ( *P < 0 . 05; **P < 0 . 01; ***P < 0 . 001 ) . For the non-normal data , statistical significance was evaluated by Mann Whitney test ( *P < 0 . 05; **P < 0 . 01 ) .
|
The thalamus—a brain structure commonly associated with relaying sensory information between cortex and other regions—is organized into many cell clusters called nuclei . Each thalamic nucleus is populated by neurons with distinct patterns of gene expression and connections to other brain regions and plays a distinct role in cortical functions . In this study , we performed an analysis of developing cells in the thalamus , using the mosaic analysis with double markers ( MADM ) method in mice , a technique that allows the labeling of descendants of dividing cells . Using 3 different transgenic mouse lines allowed us to determine the cell lineage of thalamic progenitor cells at different locations and stages of differentiation . By genetically labeling single progenitor cells , we measured how cell division and maturation occurs during the brief time span when neurons are generated . Our data also show how neurons eventually contribute to multiple nuclei across the thalamus . The organizational principles that we found in the thalamus might apply to the development of other brain structures that are composed of multiple nuclei .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"neurogenesis",
"cell",
"division",
"analysis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"brain",
"cloning",
"neuroscience",
"stem",
"cells",
"bioassays",
"and",
"physiological",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"developmental",
"neuroscience",
"cell",
"analysis",
"animal",
"cells",
"thalamus",
"molecular",
"biology",
"cellular",
"neuroscience",
"thalamic",
"nuclei",
"anatomy",
"cell",
"biology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2018
|
In vivo clonal analysis reveals spatiotemporal regulation of thalamic nucleogenesis
|
Acute effects of sex steroid hormones likely contribute to the observation that post-pubescent males have shorter QT intervals than females . However , the specific role for hormones in modulating cardiac electrophysiological parameters and arrhythmia vulnerability is unclear . Here we use a computational modeling approach to incorporate experimentally measured effects of physiological concentrations of testosterone , estrogen and progesterone on cardiac ion channel targets . We then study the hormone effects on ventricular cell and tissue dynamics comprised of Faber-Rudy computational models . The “female” model predicts changes in action potential duration ( APD ) at different stages of the menstrual cycle that are consistent with clinically observed QT interval fluctuations . The “male” model predicts shortening of APD and QT interval at physiological testosterone concentrations . The model suggests increased susceptibility to drug-induced arrhythmia when estradiol levels are high , while testosterone and progesterone are apparently protective . Simulations predict the effects of sex steroid hormones on clinically observed QT intervals and reveal mechanisms of estrogen-mediated susceptibility to prolongation of QT interval . The simulations also indicate that acute effects of estrogen are not alone sufficient to cause arrhythmia triggers and explain the increased risk of females to Torsades de Pointes . Our results suggest that acute effects of sex steroid hormones on cardiac ion channels are sufficient to account for some aspects of gender specific susceptibility to long-QT linked arrhythmias .
In the past decade , studies have suggested that female gender is an independent risk factor for long-QT ( LQT ) dependent cardiac arrhythmias [1]–[3] . Since the differences in QT intervals in males and females appear from the time of puberty [4] , [5] , sex steroid hormone effects on cardiac repolarization have been implicated . Clinical studies have found no difference in QT interval in male and female children , but shorter QT intervals in men versus women under age 50 [4] . The international Long QT syndrome ( LQTS ) registry 1998 reported that females had higher risk of a first cardiac event between 15 and 40 years [6] . Moreover , clinical findings observed that more than 68% of drug-induced torsade de pointes ( TdP ) occur in women [7]–[9] . It is known that one way that sex steroid hormones cause functional physiological changes is via transcriptional regulation . Sex hormones may bind to sex hormone receptors and then translocate into the nucleus . In the nucleus , a ligand-bound sex hormone receptor acts a transcription factor by binding to the promoter region of genes containing a hormone responsive element ( HRE ) , leading to regulation of gene expression . For example , in the heart , lipocalin-type prostaglandli D synthase ( L-PDGS ) has been found to be transcriptionally upregulated by estradiol and estrogen receptor ( ER ) [10] . This genomic action requires several hours before the effects can be observed . In addition to the genomic pathway , sex steroid hormones may induce a rapid activation of mitogen-activated protein kinase ( MAPK ) leading to transcription factor activation [11] , [12] as well as activation of membrane bound endothelial nitric oxide synthase ( eNOS ) [13] , [14] . Interestingly , recent studies have demonstrated that sex steroid hormones may also act acutely and rapidly modulate cardiac ion channel activity directly via a PI3K/Akt/eNOS pathway [15]–[17] . Testosterone induced phosphorylation of the Ser/Thr kinase Akt and eNOS leads to NO synthase 3 ( NOS3 ) activation and production of nitric oxide ( NO ) [15] . NO leads to s-nitrosylation of cysteine residues on the channel underlying the slow delayed rectifier K+ current ( IKs ) [17] . L-type Ca2+ current ( ICa , L ) is conversely suppressed by NO via a cGMP dependent pathway . Regulation of IKs and ICa , L by testosterone is dose-dependent [15] and leads to shortening of action potential duration ( APD ) [15] and QT intervals [18]–[20] . In adult men , the serum testosterone level is reported to be 10 to 35 nM [21] , however circulating levels of testosterone begin to decline in men as young as 40 [22] . QT intervals are shorter in adult men than in adult women until around the age of 50 [4] , suggesting a likely role for circulating testosterone . In females , progesterone fluctuates through the menstrual cycle . The reported serum progesterone level is 2 . 5 nM in the follicular phase and 40 . 6 nM in the luteal phase [23] . It was recently shown by Nakamura et al . that progesterone increases IKs current through the NO production pathways and prevents cAMP-enhancement of ICa , L [16] . The apparent result of acute effects of progesterone and testosterone is to shorten ventricular repolarization and diminish incidence of arrhythmias [15] , [16] , [20] , [24] . Recently , experiments have suggested protective effects of testosterone against arrhythmia . In vivo experiments show that orchiectomized male rabbits treated with dihydrotestosterone ( DHT ) had shorter QT interval and APD90 compared to non-DHT treated rabbits [18] , [20] . Also , experiments in testosterone treated female animals have shown that DHT reduces drug-induced arrhythmia by dofetilide [24] . The acute effects of estradiol result in suppression of human ether-a-go-go-related gene ( hERG ) underlying the rapid delayed rectifier current ( IKr ) by directly binding to the channel , altering channel kinetics and reducing current [25] . Kurokawa and co-workers showed that 17β-estradiol ( E2 ) increases the channel rate of closure ( deactivation ) and lessens repolarizing current . They also showed that in the presence of E2 , hERG is more sensitive to block by drugs . The group proposed that aromatic centroid of E2 may be responsible for increasing the sensitivity of hERG block by E4031 via interaction with the aromatic side chain of Phe656 and aromatic rings of the hERG blocker . Because 1 ) the concentration of E2 is not constant through the menstrual cycle , but rather fluctuates from the peak follicular phase serum E2 level of 1 nM to 0 . 7 nM in the luteal phase , and 2 ) E2 has dramatic effects on sensitivity to hERG block within this range , it stands to reason that susceptibility to drug-induced arrhythmia by hERG block may vary through the menstrual cycle . Although studies have shown that female hormones estradiol and progesterone have opposite effects on cardiac repolarization: E2 prolongs QT intervals , and progesterone reduces QT interval [16] , [25] , [26] , the question of whether normal hormonal fluctuations are sufficient to account for variability in QT during the menstrual cycle in not known . Neither are the effects of physiological concentrations of hormones on arrhythmia susceptibility well understood . Some studies do report that dynamic fluctuations in QT intervals during the menstrual cycle are related to changes in susceptibility to TdP risk [27] , [28] . Other studies in postmenopausal women also suggest the importance of female hormones as estrogen hormone replacement therapy prolongs QT intervals and increases arrhythmia risk [26] , [29] , [30] . Other data have not found marked fluctuation in QT interval during specific phases of the menstrual cycle [28] , [31] , [32] . Burke et al . , ( 1997 ) found that the corrected QT ( QTc ) interval does not significantly change through menstrual cycle in pre-menopausal women; however , QTc is reduced in the luteal phase after autonomic blockade [31] . A study of drug-induced QT prolongation during the menstrual cycle observed that QTc did not vary during the menstrual cycle , but QTc shortening was more pronounced in the luteal phase with ibutilide application [28] . Nonetheless , both the clinical and experimental data suggest that women have both longer QT intervals than men and are more likely to develop long-QT dependent arrhythmias and TdP arrhythmias [9] , [28] . Women are especially susceptible to increased arrhythmia risk in response to QT-prolongation drugs [9] , [28] , [33] , [34] . It is a major challenge to specifically determine the relationship between sex steroid hormones and arrhythmia susceptibility in males and females since the cardiac system is extraordinarily complex . In order to attribute risk to a particular parameter , in this case physiologically relevant concentrations of sex steroid hormones , the specific effect must be studied in isolation without other perturbations to the system . This is the strength of the computational approach that we employ . In the present study , we focus on acute effects of sex steroid hormones on cardiac ion channel targets . We use guinea pig models that incorporate the effects of hormones measured experimentally from guinea pig , and then can test these changes specifically within the complex cellular and tissue milieu . Importantly , we use the model to make predictions about the effects of physiological concentrations of sex steroid hormones on gender specific cardiac physiology parameters and arrhythmia susceptibility . Some recent experimental studies investigating functional effects of sex hormones on cardiac function have utilized hormone concentrations in the micromolar range that is orders of magnitude higher than the nanomolar physiological circulating concentration of E2 [35] . This is a critical consideration because micromolar concentrations of E2 are apparently cardioprotective via effects on L-type Ca2+ current ( ICa , L ) . Although high hormone concentrations may be relevant during phases such as pregnancy , a recent study showed that E2 at 1 nM did not have significant effects on IKs or ICa , L [25] . Our model simulations reproduce observed fluctuations of QT through the menstrual cycle in females in both cell and tissue-level . Simulations also predict that effects of testosterone and progesterone on ion channels hasten repolarization and protect from drug-induced arrhythmias .
To investigate the acute effects of sex steroid hormone on cardiac electrophysiology and arrhythmia susceptibility , we developed a computational model that mimicked the conditions employed experimentally so that we could directly validate our model by comparison to experimental measurements . Experiments were conducted in isolated ventricular myocytes from Langendorff-perfused adult female guinea pigs , so that they were free of endogenous neuronal and hormonal effects . The isolated cells were then incubated with human physiological concentrations of hormones for 10 min . and the effects of hormones on cardiac ion channels were measured . A range of cardiac ion channels were screened for functional changes induced by sex steroid hormones , but acute effects of progesterone were found only to modify IKs [16] while testosterone primarily increased IKs and inhibited ICa , L [15]; acute E2 treatment only significantly suppressed IKr current [25] . We utilized the experimentally observed effects of physiological concentrations of sex-steroid hormones in adult women and men and incorporated these functional changes into our computational models ( described in detail in Supplemental Text S1 ) . Experiments [25] show that E2 primarily affects the conductance of IKr , and has a minor , but measurable and significant effect on slowing channel activation kinetics . To simulate the experimentally observed IKr current reduction by E2 ( Figure 1A – right ) , we scaled the IKr conductance and incorporated the minor effects of E2 on the voltage dependence ( not shown ) of IKr in the model ( Figure 1A – left ) . E2 at 1 nM reduced IKr tail current in a dose-dependent manner , but did not affect the time course of tail current decay ( Figure 1A ) . Unlike the direct effects of E2 on IKr , progesterone modulates the IKs through non-genomic activation of eNOS . We used experimental data [16] ( Figure 1B – left traces ) to scale the conductance of ionic currents in the model to incorporate effects of progesterone on IKs . Progesterone-induced IKs enhancement is concentration-dependent as shown in Figure 1B . Experimentally recorded and simulated dose-response curves for progesterone effects on IKs tail current amplitude is shown in Supplemental Figure S1 . IKs current was simulated with different concentrations corresponding to progesterone concentrations at various points in the menstrual cycle ( 0 nM – control case , 2 . 5 nM – follicular phase , 40 . 6 nM – luteal phase and 100 nM - maximal experimental concentration ) during a voltage pulse from −40 mV to +50 mV . Note that the effect of progesterone on IKs is nearly saturated at a concentration of 40 . 6 nM , corresponding to the peak value during the luteal phase of the menstrual cycle ( indicated by the near overlay of the 100 nM curve ) . Like progesterone , testosterone modifies cardiac ion channels comprising IKs and ICa , L via eNOS production of NO . We used the same method as above to incorporate experimental ratios of control conductance for testosterone . Dose-dependent effects of testosterone on IKs enhancement and ICa , L suppression are shown in Figure 1C for experiments ( top ) and simulated currents ( lower panels ) . Simulated IKs and ICa , L are compared to experimentally recorded guinea pig IKs and ICa , L using the same protocol . Cells were depolarized to test potential +50 mV for 3 . 5 seconds and then repolarized to −40 mV to record IKs . ICa , L was experimentally recorded during a voltage step from −40 mV to 0 mV . Testosterone strongly enhances IKs current ( Figure 1C – left traces ) at 10 nM while high concentrations of testosterone ( 300 nM ) markedly suppress ICa , L ( Figure 1C – right traces ) . Like humans , many studies have demonstrated that female guinea pigs have slower repolarization than male guinea pigs [1] , [36] . To examine the contribution of sex-steroid hormones on the ventricular action potential duration ( APD ) , we included the effects of E2 , progesterone and testosterone on membrane currents and simulated action potentials ( APs ) in three cell types . Figure 2 shows APs for the 50th beat at 1000 ms pacing rate in M cells . Simulated APs of epicardial and endocardial cells are described in Supplemental Figure S2 . E2-induced IKr suppression contributes to APD prolongation in a dose-dependent manner ( Figure 2A ) . A low concentration of E2 ( 0 . 1 nM ) , corresponding to the early follicular phase of the menstrual cycle , has slight effects on APD compared with control case ( from 233 to 235 ms — 0 . 86% prolongation ) . However , a concentration of E2 corresponding to the late follicular phase of the menstrual cycle ( prior to ovulation ) ( 1 . 0 nM ) prolonged APD ( 250 ms ) by 7 . 3% ( Figure 2A ) . This value is in good agreement with the observed APD prolongation in guinea pig myocytes in patch-clamp experiments with E2 incubation ( 11±1% ) [25] . Figure 2B shows that progesterone reduced APD in a concentration-dependent manner ( 222 ms — 4 . 7% reduction at 2 . 5 nM corresponding to the follicular phase; 212 ms — 9 . 0% reduction at 40 . 6 nM , corresponding to the luteal phase ) , which agrees with patch-clamp experimental data ( 6 . 3% reduction at 40 . 6 nM ) [16] . To investigate the combined effects of E2 and progesterone as they fluctuate during the normal menstrual cycle on the cardiac action potential , we used clinically measured concentrations of hormones at three discrete phases of the menstrual cycle ( early follicular , late follicular and luteal ) . During the early follicular stage , E2 = 0 . 1 nM , progesterone = 2 . 5 nM , during the late follicular stage , E2 = 1 . 0 nM , progesterone = 2 . 5 nM and during the luteal stage , E2 = 0 . 7 nM , progesterone = 40 . 6 nM [23] . As see in Figure 2C , the simulations predict longer APD in the late follicular phase ( 233 ms ) than in the early follicular ( 223 ms — 4 . 3% reduction ) . Simulations predict shortest APD in the luteal phase ( 218 ms — 6 . 4% reduction ) , consistent with experimental observations ( ≈11% shortening ) [36] . We also simulated changes in APD at two physiological concentrations of testosterone ( 10 nM and 35 nM ) shown in Figure 2D , which reflect the normal low and high ranges found in post-pubescent pre-senescent males [21] . The simulations predict marked APD shortening by 10 . 7% ( 208 ms ) and 15 . 9% ( 196 ms ) at 10 and 35 nM testosterone , respectively . We next computed the effects of sex-steroid hormones in a one-dimensional strand of coupled M cells ( results from other cell types are shown in Supplemental Figure S2 ) to determine the effects of hormones in an electrotonically coupled system ( Figure 3 ) . We also computed spatial gradients of depolarization and repolarization to generate a pseudo ECG electrogram ( Figure 3B ) . APs were initiated via a stimulus applied to the first cell and then propagated from top to bottom along the 1 cm fiber . Figure 3A show that the first cell fired first and then repolarized first . The effects of E2 on IKr leads to dose-dependent APD prolongation in the simulated tissue ( Figure 3A ) , and results in a longer QT interval in the presence of 1 nM ( 7 . 7% prolongation ) from 260 ms ( Figure 3A-i — 0 nM sex-steroid hormone ) to 280 ms as seen in Figure 3B ( top panel ) . Also , the simulations clearly show progesterone shortened APD in a dose-dependent manner ( 3A-iv 2 . 5 nM , and 3A-v 40 . 6 nM ) . The corresponding computed electrograms from the fibers in Figure 3B ( lower panel ) demonstrates the progesterone-induced QT interval reduction from 260 ms ( control case ) to 250 ms ( 3 . 8% — iv ) and 240 ms ( 7 . 7% — v ) . A recent clinical study has observed that the QT intervals fluctuate during the menstrual cycle , suggesting that progesterone may reverse effects of the estrogen-induced QT prolongation [27] . Figure 4A represents the results of simulations in a 1D cable at combined hormone concentrations observed during various phases of the menstrual cycle . Simulations show a QT interval reduction of 10 and 20 ms in the luteal phase compared to the early and late follicular phases , respectively ( Figure 4B — top panel ) , which agree with the clinically observed QT shortening ( ≈10 ms shortening in the luteal phase compared to the follicular phase ) [27] . The models demonstrate that despite the presence of E2 ( 0 . 7 nM ) during the luteal phase , high progesterone ( 40 . 6 nM ) results in luteal phase shortening of APD and a QT interval ( on the pseudo-ECG ) reduction of 4% ( from early follicular phase ) and 7 . 7% ( from late follicular phase ) . The experimental study from Liu et al . suggested the QT intervals were significantly shorter ( 11 . 3% ) in male than in female rabbits [37] . In Figure 4A-iv and Figure 4A-v , our simulations show the effects of testosterone on APD in simulated one-dimensional tissue . The model predicts that testosterone-induced faster repolarization and caused QT interval reduction to 230 ms ( 11 . 5% shortening — case iv ) and 220 ms ( 15 . 3% shortening — case v ) compared with the late follicular phase ( 260 ms ) in Figure 4B . We also ran these simulations in the presence of 10 nM and 35 nM testosterone and 0 . 1 nM E2 , which is estimated as the average circulating concentration of E2 in men [28] ( shown in Supplemental Figure S3 ) . In the presence of E2 , QT intervals increase by 10 ms , corresponding to 7 . 7% ( 10 nM ) and 11 . 5% ( 35 nM ) shortening compared to the late follicular phase in females . Experimental evidence suggests that in the presence of physiological concentrations of E2 , the potency of IKr block by drugs is increased [25] . This finding may explain , in part , the increased susceptibility of females to drug-induced arrhythmias [8] , [9] . Hence , we next tested the effect of E2 on IKr suppression induced by the IKr channel blocker E-4031 and investigated the effects of female hormones on drug-induced arrhythmia susceptibility . Experimental results [25] shown in Figure 5A ( top ) illustrate that E2 ( 1 nM ) considerably increased the suppression of hERG by E-4031 ( light gray line ) . However DHT did not greatly change the drug-induced inhibition of hERG current ( dark gray line ) . We then obtained measured ratios of IKr conductance in the presence of E-4031 and E2 or DHT from the experimental data and used these values to simulate dose-dependence curves for IKr suppression by E-4031 ( control — black line ) and after addition of 1 nM E2 ( light gray line ) and DHT 3 nM ( dark gray line ) ( Figure 5A — lower panel ) . In Figure 5B ( top panel ) , we show a simulation of a one-dimensional strand of coupled M cells ( 100 cells ) in the late follicular phase during E-4031 treatment , where the model predicts the most dramatic APD and QT interval prolongation . At 10 nM E-4031 , the simulated tissue-level APD is shorter with testosterone application ( 250 ms — 3 nM ) compared with APD in the presence of female hormones ( 280 ms — E2 = 1 . 0 nM , progesterone = 2 . 5 nM ) as seen in Figure 5B . The pseudo ECG ( 5B — lower traces ) shows that QT interval is substantially longer in the late follicular phase ( case i ) than with testosterone treatment ( case ii ) . The exact mechanism of TdP induction is unclear , but it is thought that pause-induced early afterdepolarizations ( EADs ) can underlie TdP initiation [38] , [39] . Hence , we performed a series of simulations to investigate pause-dependent LQT syndrome and its association with arrhythmia susceptibility in the presence of male and female hormones . Single M cells were paced for 10 beats of BCL at 1000 ms ( s1 ) followed by a premature beat ( s2 ) with varying s1–s2 intervals and then a long pause of varying duration as indicated . Our simulations show no EADs ( APD>450 ms ) occurred during the late follicular phase with no drug application ( Figure 6A — left panel ) or with the application of E-4031 in the presence of testosterone 3 nM ( middle ) during a short-long pacing protocol . However in the absence of sex-steroid hormones , EADs were generated by addition of 10 nM E-4031 when the pause interval was very long ( >2500 ms ) ( right panel ) . In Figure 6B , we investigated the short-long pacing induced EAD by E-4031 in the late follicular phase , where the concentration of E2 is highest , after pacing at three basic cycle length ( 500 , 750 and 1000 ms ) . This pacing sequence triggered EADs over a wide range of pauses in all three conditions . APDs of the s3 ( post pause ) beat are notably lengthened with increasing basic cycle lengths from 500 ms to 1000 ms ( Figure 6B — left panel to right panel ) . Severe EADs were induced at 1000 ms pacing length with a pause greater than 1500 ms ( 6B — right panel ) . The point in Figure 6B ( right ) indicates an EAD that was triggered following a pause of 1500 ms and s1–s2 interval of 810 ms during baseline pacing length of 1000 ms . We have carried out the simulations in a coupled one-dimensional M-cell tissue ( 6B — lower panels ) using the same protocol , and observed propagation of the EAD in the tissue . These simulations suggest that 3 nM testosterone is sufficient to prevent EAD development in the presence of E-4031 10 nM . However , in females , during the late follicular phase of the menstrual cycle , the increased concentration of estrogen appears to exacerbate drug-induced TdP arrhythmias . Finally , to test the potential for E2-exacerbated EADs to trigger reentrant arrhythmia activity in 2D heterogeneous tissue , we carried out a series of simulations with varying combinations of hormones and/or drug application . The simulated tissue was stimulated along one edge , and a point stimulus was applied to induce an ectopic beat during a short-long-short sequence described in Supplemental Text S1 . Figure 7 shows the results of simulations in four cases at indicated time points . In the absence of hormones or drugs , an initiated wave propagates in all directions , and no reentry occurs ( first row ) . The same behavior is observed following drug application alone ( E-4031 ) and with testosterone application alone ( DHT 10 nM ) . However , when E2 ( 1 nM ) is present ( bottom row ) , the M-cell region is preferentially prolonged ( due to the effect of E2 on the background of less repolarizing current that defines this region ) , which prevents the wavefront from crossing the refractory M-cell region . Instead , the wave propagates leftward at first - until the M-cell region repolarizes , and allows the wave to first cross the M-cell region and then slowly turn to the right . The slowly traveling wavefront ( Na+ channels are only partially recovered following the prolonged action potential initiated by previous stimulus ) begins a cycle of reentry – turning around and continuing to propagate on the wake of the preceding wave ( Figure 7 – bottom panel ) . In Figure 8A ( top ) , the simulations suggest no reentrant activity during the late follicular phase of the menstrual cycle ( progesterone 2 . 5 nM and 1 nM E2 ) . However , when 10 nM E-4031 is applied during the late follicular phase , a spiral wave is readily induced ( Figure 8A – middle ) . We also tested the effects of male hormone ( testosterone ) in the presence of E-4031 . Figure 8A ( bottom ) shows that testosterone 3nM with 10 nM E-4031 did not trigger reentry activity . Interestingly , the induction of a spiral wave in the presence of E-4031 during the late follicular phase of the menstrual cycle is a robust phenomenon . Reentry was introduced in this condition when the ectopic stimulus was applied in the subendocardial or subepicardial region – although not in the M-cell region ( not shown ) . The position of the stimulus is also not critical . Figure 8B shows the effect of a point stimulus applied in the middle of endocardial tissue , leading to the initiation of a pair of counter-rotating spiral waves ( Supplemental Figure S4 – protocol 2 ) .
Here , we demonstrate the acute effects of sex steroid hormones in model cells and tissue , from physiological blood concentration to channel interaction , to their effects on APD and tissue dynamics . We used a computational approach to examine the role for acute application of sex steroid hormones on susceptibility to cardiac arrhythmias . The benefit of this approach is that it allows us to investigate the consequences of hormones on cardiac ion channels in isolation , so that observed changes can be specifically attributed to them . We simulated the acute effects of sex steroid hormones on cardiac cell and tissue dynamics and on fluctuations of QT interval . It has been shown that progesterone enhances IKs , which counterbalances the IKr reduction by E2 . Because estrogen and progesterone dominate in the different phases of menstrual cycle , simulations show that during the late follicular phase ( prior to ovulation ) of the menstrual cycle , QT interval is longer than in the luteal phase when progesterone is increased , which is consistent with the clinical observation by Nakagawa et al [27] . Notably , the fluctuations in QT interval during the menstrual cycle predicted by our model are within a relatively narrow range of 20 ms , which approximates the clinically assessed standard deviation in pooled QT intervals for a patient population assessed at each phase of the menstrual cycle [31] . One explanation is that such an analysis is unlikely to be sensitive enough to observe significant individual differences in QT intervals as they are fluctuating throughout the menstrual cycle , since biological variability between patients may be larger than fluctuations in individual patients . Here we demonstrated that increasing testosterone reduced the APD and QT interval in a dose-dependent manner ( Figure 2 and 4 ) by enhancing IKs and inhibiting ICa , L current . Moreover , differences in APD become more pronounced between E2 treatment and testosterone treatment when cycle length ≥800 ms - shown in Supplemental Figure S5 . Taken all together , these results suggest that sex hormones influence cardiac repolarization in a dose- and cycle length-dependent manner . This is consistent with experimental studies of gender-related differences on cycle length-dependent QT [2] , [37] . Since clinical findings suggest female gender is an independent risk factor for TdP arrhythmias and previous experimental studies have shown that E-4031 induced greater prolongation in E2-treated than in DHT-treated animals [2] , [37] , [40] , we investigated the potential for E2 to exacerbate and testosterone to ameliorate arrhythmias in the presence of IKr block . We used simulations to probe these effects and ask if the presence of these two hormones at physiological concentrations play a key role on gender differences in drug-induced LQTS . We incorporated the experimentally measured combined effects of E2 and E-4031 on IKr and then simulated them on cardiac tissue dynamics . Since drug-induced TdP is often observed following a short-long-short pacing sequence in clinical studies [38] , [41] , [42] , we explored tissue dynamics using such a protocol . Although we did not observe the development of arrhythmogenic EADs during the late follicular phase of the menstrual cycle ( when E2 concentration is at its peak ) , addition of the IKr blocker E-4031 resulted in EAD formation in the late follicular phase for a wide range of pacing protocols ( Figure 6 ) . The model simulations also suggested that E-4031 treatment in late follicular phase could lead to initiation of spiral wave reentrant arrhythmias ( Figure 8 ) . These simulations imply that at certain phases of the menstrual cycle , elevated levels of E2 may put females at risk for drug-induced arrhythmias – in particular by agents that bind to the promiscuous drug target hERG . Furthermore , we demonstrated that progesterone has a protective effect against E2-induced LQT syndrome ( Figure 7 and 8 ) . Spiral waves were not initiated in the presence of low concentrations of progesterone ( 2 . 5 nM – late follicular phase ) . This suggests that progesterone may play an important role in protecting against arrhythmia in females . Unlike the apparent pro-arrhythmic effects of E2 in the presence of Ikr block , testosterone was shown in our simulations to prevent E-4031 induced EADs during a short-long-short pacing protocol ( Figure 6 and 8 ) . In the present study , our computational investigation demonstrates the acute effects of progesterone , estradiol and testosterone on cardiac ion channels that are critical for the rate of cardiac repolarization and resultant QT intervals . The models successfully simulates the effects of progesterone and testosterone on cardiac IKs and ICa , L . These two hormones hasten repolarization , albeit to different extents , and reduce QT interval and susceptibility to LQT-linked arrhythmias . On the contrary , E2 increased QT interval and propensity for TdP arrhythmias by reducing repolarizing current via IKr . There are several limitations to this study that must be noted . First , we deliberately focused on the effects of acute application of physiological concentrations of sex steroid hormones here , but this means that we have neglected the effects of chronic hormone application . Several studies have suggested that chronic exposure to sex hormones may alter the response of the tissue to acute application of sex hormones [35] , alter expression of ICaL ( in a species dependent manner ) [43]–[47] and may cause structural remodeling of the myocardium [45] . We ran simulations incorporating the measured differences by Verkerk [43] in human ICaL between males and females ( since the goal was not to examine guinea pig sex differences ) . The results of these simulations are in Supplemental Figure S6 . As expected , the additional Ca2+ current in the female prominently exacerbated the APD prolongation associated with E2 . These effects may even be expected to increase in human females where primary repolarizing K+ currents , especially IKs , are apparently less prominent than in guinea pig [48] . However , progesterone effects on IKs in human may offset some of the observed increase in ICaL in human females compared to human males ( described above ) . This issue should be addressed in future studies when the human data are more complete . The guinea pig model is also lacking transient outward K+ currents ( Ito ) – a subset of channels that have not yet shown to be affected by acute application of sex steroid hormones . In summary , our study suggests that a computational approach to investigating effects of physiologically circulating hormones can be useful to test and predict their contribution to gender differences in cardiac arrhythmia susceptibility . Moreover , the findings from our model simulations suggest the potential utility of progesterone as a therapeutic agent for inherited and acquired forms of Long-QT Syndrome and that progestin-only contraceptives be given special consideration for their potential amelioration of LQT risk among pre-menopausal women . Finally , the link between estrogen containing hormone replacement therapy among post-menopausal women and increased incidence of adverse cardiac events needs to be investigated in the context of acute hormone effects on ion channels .
The guinea pig cardiac cell model was chosen in this study because we used the guinea pig ventricular myocytes experimental results reported in Ref . 11 , 12 , and 21 . We modified the Faber-Rudy cardiac cell model [49] . The IKr channel was replaced with Markov model of wild type based on Clancy and Rudy so that E2 effects on activation gating could be readily incorporated . Although the level of detail in terms of description of gating by the IKr Markov model or the H-H IKr model is the same , the key difference is that the Markov model takes into account the property of coupling between discrete states , while H-H presumes independence between gating processes . This difference is only relevant in the setting of perturbation to one gating process – precisely what is observed when E2 is present , where activation gating is exhibits a small positive shift in voltage dependence [50] . All the simulation code was in C/C++ and run on Mac Pro 3 . 0 GHz 8-Core computers . The time step was set to 0 . 0005 ms during AP upstroke , otherwise the time step was 0 . 01 ms . Numerical results were visualized using MATLAB R2007a by The Math Works , Inc . Details of computational models and simulation protocols can be found in Supplemental Text S1 .
|
It is well known that female gender is an independent risk factor for some types of cardiac arrhythmias . However , it has been difficult to determine how much of a role physiological concentrations of circulating sex steroid hormones play in gender linked arrhythmia susceptibility because the cardiac system is so extraordinarily complex . Here we employ a computational strategy , based on experimental measurements , to tease out the individual contributions of estrogen , progesterone and testosterone on cardiac electrical behavior and then make predictions about their effects in combination and in the presence of drugs . The computational models convincingly reproduce observed fluctuations of QT intervals ( as recorded on the ECG ( electrocardiogram ) , the QT interval reflects the time period between ventricular excitation and relaxation ) through the menstrual cycle in females and effects of testosterone on ECG parameters . Our simulations also predict that testosterone and progesterone are protective against drug-induced arrhythmias , while estrogen likely exacerbates the breakdown of normal cardiac electrical activity in the presence of QT-prolonging drugs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cardiovascular",
"disorders/arrhythmias,",
"electrophysiology,",
"and",
"pacing",
"physiology/cardiovascular",
"physiology",
"and",
"circulation",
"computational",
"biology/systems",
"biology",
"cardiovascular",
"disorders/cardiovascular",
"diseases",
"in",
"women"
] |
2010
|
Acute Effects of Sex Steroid Hormones on Susceptibility to Cardiac Arrhythmias: A Simulation Study
|
Epstein-Barr virus ( EBV ) is closely associated with nasopharyngeal carcinoma ( NPC ) , a human malignancy notorious for its highly metastatic nature . Among EBV-encoded genes , latent membrane protein 1 ( LMP1 ) is expressed in most NPC tissues and exerts oncogenicity by engaging multiple signaling pathways in a ligand-independent manner . LMP1 expression also results in actin cytoskeleton reorganization , which modulates cell morphology and cell motility— cellular process regulated by RhoGTPases , such as Cdc42 . Despite the prominent association of Cdc42 activation with tumorigenesis , the molecular basis of Cdc42 activation by LMP1 in NPC cells remains to be elucidated . Here using GST-CBD ( active Cdc42-binding domain ) as bait in GST pull-down assays to precipitate active Cdc42 from cell lysates , we demonstrated that LMP1 acts through its transmembrane domains to preferentially induce Cdc42 activation in various types of epithelial cells , including NPC cells . Using RNA interference combined with re-introduction experiments , we identified FGD4 ( FYVE , RhoGEF and PH domain containing 4 ) as the GEF ( guanine nucleotide exchange factor ) responsible for the activation of Cdc42 by LMP1 . Serial deletion experiments and co-immunoprecipitation assays further revealed that ectopically expressed FGD4 modulated LMP1-mediated Cdc42 activation by interacting with LMP1 . Moreover , LMP1 , through its transmembrane domains , directly bound FGD4 and enhanced FGD4 activity toward Cdc42 , leading to actin cytoskeleton rearrangement and increased motility of NPC cells . Depletion of FGD4 or Cdc42 significantly reduced ( ∼50% ) the LMP1-stimulated cell motility , an effect that was partially reversed by expression of a constitutively active mutant of Cdc42 . Finally , quantitative RT-PCR and immunohistochemistry analyses showed that FGD4 and LMP1 were expressed in NPC tissues , supporting the potential physiologically relevance of this mechanism in NPC . Collectively , our results not only uncover a novel mechanism underlying LMP1-mediated Cdc42 activation , namely LMP1 interaction with FGD4 , but also functionally link FGD4 to NPC tumorigenesis .
Epstein–Barr virus ( EBV ) is a human γ-herpesvirus that is closely associated with many human malignancies , including nasopharyngeal carcinoma ( NPC ) , Burkitt's lymphoma , T-cell lymphoma , and gastric carcinoma [1] . NPC , which is prevalent in Taiwan and southeastern China , is a human squamous cell cancer notorious for its highly metastatic nature [2] . In NPC , EBV infection is predominantly latent and viral gene expression is restricted . Among the expressed viral genes , latent membrane protein 1 ( LMP1 ) is detected in most NPC tissues [3] . LMP1 has oncogenic properties to transform rodent fibroblast cell lines [4] , [5] and promote cell growth in soft agar [6] . LMP1 is a 62-kDa integral membrane protein composed of a short N-terminal domain , six transmembrane domains , and a 200-amino-acid ( aa ) cytoplasmic tail at the C-terminus [7] . By mimicking TNFR ( tumor necrosis factor receptor ) family members , LMP1 through its cytoplasmic tail engages TRAFs ( TNFR-associated factors ) and TRADD ( TNFR-associated death domain protein ) to transduce multiple signaling pathways , including nuclear factor-kappa B ( NF-κB ) -mediated transcription [8] and the c-Jun amino-terminal kinase ( JNK ) pathway [9] , [10] . Unlike TNFR-based signaling , however , LMP1 appears to signal in a ligand-independent fashion relying on its N-terminus and transmembrane domains to self-associate in the lipid rafts [11]–[13] . As a result , LMP1 is a constitutively active receptor [14]–[16] . In addition to growth transformation , LMP1 has also been linked to regulation of the actin cytoskeleton . In lymphocytes , LMP1 expression leads to the formation of membrane protrusions and membrane ruffling , which involve actin reorganization [5] . In Swiss 3T3 fibroblasts , LMP1 is capable of inducing the assembly of actin-rich surface protrusions called filopodia [17] . Moreover , the LMP1-induced formation of filopodia in fibroblasts can be abolished by a dominant-negative mutant of Cdc42 , a member of the Rho ( Ras-homology ) GTPase family [17] , implying that LMP1 is capable of activating Cdc42 . Rho GTPases , mainly comprising members of the Cdc42 , Rac and Rho subfamilies , actively regulate various actin-dependent functions such as cell migration , adhesion , cytokinesis , axon guidance , and phagocytosis in all eukaryotic cells [18] . Cdc42 in particular is well known to regulate actin filament ( F-actin ) organization and vesicle trafficking [19]–[21] . Like all GTPases , Cdc42 acts as a binary switch cycling between an inactive ( GDP-bound ) and an active ( GTP-bound ) conformational state . The activation of Cdc42 is mediated by guanine nucleotide exchange factors ( GEFs ) , which convert the GDP-bound form of Cdc42 to the GTP-bound form [20] , [22] , [23] . Activated Cdc42 , in turn , binds to its downstream effectors , such as the Wiskott-Aldrich syndrome protein ( WASP ) [24] , [25] , through which it ultimately generates a variety of cellular effects [26] , [27] . Not surprisingly , given the prominent role of Cdc42 in so many aspects of cell biology , aberrant activation of Cdc42 ( or dysfunction of Cdc42 GEFs ) results in pathogenesis , including tumorigenesis and tumor progression , cardiovascular diseases , diabetes , and neuronal degenerative diseases [28] , [29] . FGD4 ( FYVE , RhoGEF and PH domain-containing 4 ) , also known as Frabin ( FGD1-related F-actin binding protein ) , like FGD2 and FGD3 , is a Cdc42-specific GEF that shows significant sequence homology to FGD1 , which was originally discovered by positional cloning as the gene responsible for a human X-linked skeletal disease called faciogenital dysplasia [30]–[33] . It has been revealed that mutations in the gene encoding FGD4 cause an inherited neurological disease commonly referred to as Charcot-Marie-Tooth ( CMT ) disease , a type of hereditary motor and sensory neuropathy [34] . All FGD proteins possess a similar domain organization , whereas each FGD has a unique N-terminal region [35] . FGD4 consists of an N-terminal FAB ( F-actin-binding ) domain , a DH ( Dbl homology ) domain containing the principal GEF catalytic unit , and multiple phosphoinositide-binding domains , including two PH ( pleckstrin homology ) domains and an FYVE ( Fab1 , YOTB , Vac1 , and EEA1 ) domain at the C-terminus [30] , [35] . Accordingly , FGD4 likely couples the actin cytoskeleton to the cellular membrane by localization to the membrane and simultaneously binding F-actin . In fibroblasts , it has been shown that rat Fgd4 binds along the sides of F-actin through the FAB domain and directly induces activation of Cdc42 in the vicinity of actin structures , resulting in actin reorganization [30] . However , the mechanisms by which external or internal stimuli transduce the signals to activate FGD4 largely remain unclear . In the present study , we sought to investigate the activation of Cdc42 by LMP1 in NPC cells , which are physiologically relevant to EBV . Importantly , we uncovered a novel mechanism underlying LMP1-mediated Cdc42 activation , showing that LMP1 physically interacts with FGD4 , leading to functional consequences associated with NPC tumorigenesis and tumor progression .
To assess the effect of LMP1 on Cdc42 activation in cells , we carried out GST-pull-down assays using GST-CBD ( containing the active Cdc42-binding domain of WASP ) as bait to precipitate active Cdc42 in lysates of 293 Tet-On cells , in which the expression of LMP1 was induced by doxycycline ( Dox ) . As shown in Figure 1A , LMP1 expression led to a 4 . 8-fold increase in the level of active Cdc42 compared with the control without affecting total Cdc42 expression levels . Moreover , LMP1-mediated Cdc42 activation was specific since activation of Rac1 and RhoA , two related members of Rho GTPase family , was not significant . To corroborate these phenomena in EBV-associated cells , we conducted similar GST pull-down assays using nasopharyngeal epithelial cells ( NP69 ) and four NPC cell lines , each of which expressed LMP1 or empty vector ( control ) . Consistent with the results obtained in 293 Tet-On cells , LMP1 expression led to a 3 . 3-fold increase in active Cdc42 in NP69 cells ( Figure 1B ) , and induced 4 . 6- , 6 . 7- , 10 . 2- , and 10 . 6-fold increases in active Cdc42 in four tested NPC cell lines ( TW02 , TW01 , TW04 , and TW06 ) , respectively ( Figure 1C ) . In contrast , there was no evidence for Rac1 or RhoA activation by LMP1 in NP69 cells or NPC cells ( Figure 1B and 1C ) . Collectively , these results demonstrate that LMP1 preferentially induces Cdc42 activation in various types of epithelial cells . To dissect the functional domains of LMP1 responsible for its activation of Cdc42 , we generated a series of LMP1 deletion constructs ( Figure 2A ) and examined their effects on Cdc42 activation in NPC cells . Deletion mutants lacking transmembrane domains 3 and 4 ( ΔTM3/4 ) or 3–6 ( ΔTM3–6 ) exhibited substantially impaired ability to activate Cdc42 , producing 3 . 1- and 3 . 5-fold increases , respectively , compared to the 9 . 7-fold increase induced by full-length LMP1 ( Figure 2B ) . In contrast , deletion of the entire C-terminal cytoplasmic tail ( ΔCT ) of LMP1 did not interfere with Cdc42 activation ( a 9 . 3-fold vs . a 9 . 7-fold increase ) , indicating that C-terminus-dependent LMP1-transduced signaling pathways are not involved in this event . To confirm the importance of the transmembrane domains of LMP1 in Cdc42 activation , we replaced this region with the transmembrane domain of a TNFR member , CD40 ( denoted CD40CT in Figure 2A ) . As shown in Figure 2C , the resulting chimera failed to activate Cdc42 ( 1 . 3-fold increase vs . 7 . 0-fold increase for chimeric and wild-type LMP1 , respectively ) , indicating that the transmembrane domains of LMP1 are required for Cdc42 activation . To verify whether the LMP1-induced activation of Cdc42 was associated with remodeling of the actin filaments , we next conducted the immunofluorescence staining using NPC cells which expressed Flag-LMP1 or its various transmembrane domains-truncated forms . As shown in Figure 2D , expression of Flag-LMP1 and its C-terminus-deleted form ( ΔCT ) led to formation of microspike-like actin structures ( filopodia ) at the plasma membrane and actin bundles at the perinuclear regions ( Golgi apparatus ) . In contrast , expression of the transmembrane domains-truncated form ΔTM3/4 or ΔTM3–6 , or the chimera CD40CT failed to induce the actin remodeling as described above . The truncated form ΔTM1/2 remained a moderate ability for the actin remodeling compared to the ΔTM3–6 or the CD40CT chimera , correlating with the extent of Cdc42 activation . Since the transmembrane domains confer properties on LMP1 that distinguish it from TNFR family members [13] , [14] , we next investigated whether the TNFR members , TNFR and IL-1R ( interleukin-1 receptor ) , were capable of activating Cdc42 in NPC cells . Compared with vehicle controls , treatment of NPC cells with the TNFR ligand TNF-α ( 50 or 100 ng/ml ) or the IL-1R ligand IL-1α ( 10 ng/ml ) for 30 min had little effect on Cdc42 activation ( Figure S1A ) and on actin remodeling ( Figure S1B ) . While NPC cells seemed to be insensitive to IL-1α induction of NF-κB signaling , positive controls showed that TNF-α induced degradation of inhibitor kappa B ( IκBα ) protein ( Figure S1A ) and subsequent nuclear translocation of a NF-κB subunit p65 ( Figure S1B ) , indicating activation of NF-κB signaling [36] and confirming the functional integrity of TNF-α in this experimental setting . These data conclusively demonstrated that the action of LMP1 was distinct from that of TNFR with respect to Cdc42 activation in NPC cells . To better elucidate the nature of LMP1-mediated Cdc42 activation , we sought to dissect the spatial pattern of Cdc42 activation upon LMP1 expression . For this aim , we generated an EGFP-CBD construct containing the active Cdc42-binding domain of WASP to locate active Cdc42 inside cells . Following transfection of 293 Tet-On cells with pEGFP-CBD , LMP1 expression was induced by doxycycline ( Dox ) and cells were analyzed by confocal microscopy . As shown in Figure S2A ( upper panel ) , LMP1 expression led to EGFP-CBD co-distribution at LMP1-resident sites ( i . e . , Golgi apparatus and plasma membrane ) . In contrast , a more homogenous distribution of EGFP-CBD was evident in cells in which LMP1 expression was not induced ( bottom panel ) , implying Cdc42 activation at LMP1-resident sites . We next verified whether LMP1 expression resulted in distribution of Cdc42 to LMP1-resident sites using NPC cells expressing EGFP-LMP1 or the truncated form ΔTM3–6 . As shown in Figure S2B , a portion of Cdc42 was co-distributed with EGFP-LMP1 ( the inset ) rather than with the ΔTM3–6 form in NPC cells . These data suggest the possibility that certain factors that act upstream of Cdc42-activation cascades may co-localize with LMP1 and participate in this event . On the basis of the above observation ( Figures 2 and S2 ) and the preferential activation of Cdc42 by LMP1 ( Figure 1 ) , we hypothesized that LMP1 modulates a Cdc42 GEF and thereby activates Cdc42 . To verify this , we first surveyed the literature for candidate Rho GEFs that show specificity toward Cdc42 and localize to the Golgi apparatus and the plasma membrane , where LMP1 was expressed as well ( Figure 2E ) . A subset of GEFs that meet such criteria , including FGD1 [37] , FGD3 [32] , FGD4 [38] , intersectin-1 ( ITSN1 ) [39] , and DOCK9 ( zizimin ) [40] , was selected for further evaluation . The expression of these GEFs at the transcriptional level was validated in NPC cells using quantitative reverse transcription-polymerase chain reaction ( RT-PCR; data not shown ) . To evaluate the effects of these GEFs on LMP1-mediated Cdc42 activation , we performed RNA interference using small interfering RNA ( siRNA ) to deplete each GEF from LMP1-expressing NPC cells , followed by precipitation of active Cdc42 , as described above . The knockdown efficiency of each siRNA toward its targeted GEF was analyzed by quantitative RT-PCR ( Figure S3A ) and Western blotting . As shown in Figure 3A , knockdown of FGD4 ( siFGD4 ) reduced the LMP1-induced Cdc42 activation from the 2 . 3-fold increase observed in control siRNA ( siCtrl ) -treated cells to a 0 . 8-fold increase . In contrast , knockdown of each of the other GEFs had little effect on LMP1-induced Cdc42 activation . A quantitative analysis of data from five independent experiments further reinforced the inhibitory effect of FGD4 knockdown on LMP1-induced Cdc42 activation ( Figure 3B ) . To reproduce the phenomenon in epithelial cells other than NPC cells , we carried out similar knockdown experiments using 293 Tet-On cells with or without Dox induction . As shown in Figure 3C , knockdown of FGD4 consistently reversed LMP1-mediated Cdc42 activation , reducing the fold-increase from 2 . 3 to 0 . 5 , indicating that this effect was not restricted to NPC cells . To reinforce the functional role of FGD4 , we re-introduced FGD4 expression by ectopically expressing FGD4 ( Myc-FGD4 ) in FGD4-depleted NPC cells that co-expressed LMP1 , and then assessed the activation of Cdc42 by LMP1 . As shown in Figure 3D , depletion of FGD4 consistently eliminated LMP1-mediated Cdc42 activation , reducing Cdc42 activation from a 3 . 0-fold increase to a 1 . 0-fold increase . In contrast , re-introduction of FGD4 expression attenuated this reduction , limiting it to a 2 . 2-fold increase . Collectively , these results confirm that FGD4 indeed plays a role in the activation of Cdc42 by LMP1 . As noted , FGD4 depletion had no effect on the basal level of Cdc42 activation in cells lacking LMP1 ( Figure 3 ) . Accordingly , the data demonstrate that FGD4 is primarily involved in mediating LMP1-induced Cdc42 activation . What little has been learned to date about the function of FGD4 has mainly been gleaned from experiments involving manipulation of rat Fgd4 [30] , [35] , [38] , [41] . Despite the high sequence homology between rat Fgd4 and human FGD4 ( Figure S3B ) , it is uncertain whether the functions of both proteins are identical . To characterize the function of human FGD4 in NPC cells , we generated a series of truncated human FGD4 constructs ( Figure 4A ) to examine their effects on Cdc42 activation ( Figure 4B–4D ) and actin remodeling ( Figure S4A ) . As shown in Figure 4B , expression of Myc-FGD4 ( FL ) activated Cdc42 compared with the vector control , increasing the level of active Cdc42 by 3 . 3- ( right panel ) and 7 . 8-fold ( left panel ) in two different experiments . The immunofluorescence staining revealed that expression of Myc-FGD4 induced filopodia formation at the plasma membrane ( Figure S4A ) . Deletion of the FAB domain ( ΔFAB ) of FGD4 did not affect the activation of Cdc42 ( Figure 4C ) and the induction of filopodia ( Figure S4A ) . In contrast , deletion of the DH domain ( ΔDH vs . FL; PH1–2 vs . ΔFAB ) substantially impaired the activity of FGD4 toward Cdc42 ( Figure 4B and 4C ) as well as induction of actin remodeling ( Figure S4A ) and recruitment of Cdc42 ( Figure S4B ) , indicating that the DH domain is essential for a full FGD4 activity . Moreover , deletion of the PH1-to-PH2 domains ( denoted FAB–DH ) caused a functional impairment of FGD4 ( Figures 4B , 4C , and S4A ) but left its ability to recruit Cdc42 unchanged ( Figure S4B ) , suggesting some aspect of the function of PH1-to-PH2 domains , such as membrane targeting , is also needed for a full FGD4 activity . To investigate how FGD4 affects LMP1-mediated Cdc42 activation , we assessed the activation of Cdc42 by LMP1 in NPC cells expressing LMP1 together with various forms of FGD4 . As shown in Figure 4B , ex`pression of LMP1 alone led to activation of Cdc42 compared with vector-transfected cells , increasing active Cdc42 levels by 5 . 1- ( right panel ) and 8 . 7-fold ( left panel ) in two different experiments . Co-expression of LMP1 with the full-length FGD4 and with the ΔFAB form further augmented LMP1-mediated Cdc42 activation by 1 . 2- and 1 . 3-fold , respectively ( Figure 4B ) , in agreement with the intact activities of these FGD4 proteins ( Figure 4C ) . In contrast , co-expression of LMP1 with the truncated forms of FGD4 that exhibited impaired FGD4 activities ( ΔDH , PH1–2 , and FAB–DH ) instead reduced LMP1-mediated Cdc42 activation on average by 62% , 50% , and 39% , respectively ( Figure 4D ) , revealing that FGD4 indeed acts downstream of LMP1 to modulate Cdc42 activation . On the basis of the known ability of PH domains to bind to phosphoinositides as well as proteins [42] , [43] and our observation that LMP1-mediated Cdc42 activation occurred at LMP1-resident sites ( Figure S2 ) , we speculated that LMP1 likely interacted with FGD4 . To verify the interaction between LMP1 and FGD4 , we performed co-immunoprecipitation assays using NPC cells co-expressing LMP1 and various forms of FGD4 . As shown in Figure 4E , LMP1 was co-precipitated with full-length , ΔFAB , and ΔDH forms of FGD4 , indicating that LMP1 interacted with FGD4 in a manner that did not require FAB or DH domains . Consistent with this , the FAB–DH form showed an impaired ability to interact with LMP1 , in contrast to the PH1–2 form , which was sufficient for interaction with LMP1 . Collectively , these results revealed that the PH1-to-PH2 domains of FGD4 are mainly responsible for the interaction of FGD4 with LMP1 . To corroborate the impact of the LMP1-FGD4 interaction on LMP1 activation of Cdc42 , we next examined if the transmembrane domains of LMP1 , which were necessary for inducing Cdc42 activation ( Figure 2B and 2C ) , were responsible for the interaction with FGD4 . To accomplish this , we performed co-immunoprecipitation assays using anti-Myc affinity resins to precipitate Myc-FGD4-associated protein complexes in lysates of NPC cells co-expressing Myc-FGD4 plus Flag-LMP1 or its chimera , CD40CT . As shown in Figure 4F , wild-type LMP1 but not CD40CT was co-precipitated with Myc-FGD4 , indicating that the transmembrane domains of LMP1 are required for its interaction with FGD4 . To explore the role of this protein-protein interaction in FGD4 activity , we assessed Cdc42 activation by Myc-FGD4 in the presence of LMP1 or CD40CT . The results showed that co-expression of Myc-FGD4 with LMP1 increased the level of active Cdc42 by 1 . 8-fold compared with expression of FGD4 alone ( Figure 4F ) . In contrast , co-expression of FGD4 with CD40CT did not promote Cdc42 activation ( 1 . 1-fold ) . Taken together , these data revealed that LMP1 interacts with and coordinates the activity of FGD4 . To investigate whether Myc-FGD4 was co-precipitated with LMP1 in a reciprocal way , we conducted co-immunoprecipitation assays using anti-Flag affinity resins to precipitate Flag-LMP1-associated protein complexes in lysates of NPC cells co-expressing Myc-FGD4 with various forms of LMP1 . As shown in Figure 5A , Myc-FGD4 was co-precipitated with Flag-LMP1 but not with the CD40CT chimera , consistent with the result shown in Figure 4F . In contrast , Myc-FGD4 was not co-precipitated with the truncated form ΔTM3/4 or ΔTM3–6 , indicating that the transmembrane domains 3 and 4 were the minimal region required for LMP1 interaction with FGD4 . Moreover , deletion of the short N-terminal domain ( ΔNT ) did not affect LMP1 co-precipitation of Myc-FGD4 ( Figure 5A ) . We next corroborate the interaction between endogenous FGD4 and LMP1 by co-immunoprecipitation assays using NPC cells expressing various forms of LMP1 . The resulting cell lysates were subsequently incubated with an anti-FGD4 antibody coupled with protein G beads to precipitate endogenous FGD4 . As shown in Figure 5B , Flag-LMP1 was co-precipitated with FGD4 , indicating a physical interaction between FGD4 and LMP1 . Notably , neither the CD40CT chimera nor the LMP1 truncated form lacking the transmembrane domains 3 and 4 ( ΔTM3/4 and ΔTM3–6 ) could be detected in the FGD4-associated protein complexes , confirming the observation that LMP1 mainly relies on its transmembrane domains 3 and 4 to interact with FGD4 . To further investigate whether LMP1 directly interacted with FGD4 , we conducted in vitro affinity chromatography assays using GST-FGD4 as bait to precipitate in vitro-translated , [35S]methionine-labeled LMP1 . As shown in Figure 5C , in vitro-translated , [35S]methionine-labeled LMP1 was precipitated by GST-FGD4 , indicating a direct interaction between LMP1 and FGD4 . To buttress the direct interaction between LMP1 and FGD4 within live cells , we carried out an advanced bioluminescence resonance energy transfer ( BRET2 ) assay using NPC cells . This technology uses Renilla luciferase ( Rluc ) as the donor molecule and a GFP2 as the acceptor molecule in an assay analogous to fluorescence resonance energy transfer ( FRET ) but without the need for the use of an excitation light source . As shown in Figure 5D , the BRET signal was evident in cells co-expressing Rluc-LMP1 and GFP2-FGD4 , but not in the controls ( Rluc/GFP2 , Rluc-LMP1/GFP2 , and Rluc/GFP2-FGD4; P = 0 . 005 , 0 . 016 , and 0 . 004 , respectively , vs . Rluc-LMP1/GFP2-FGD4 ) . Moreover , expression of Flag-LMP1 in combination with Rluc-LMP1/GFP2-FGD4 competitively decreased the BRET ratio ( P = 0 . 004; paired t-test ) , providing evidence for specificity . To investigate whether LMP1 affected the localization of FGD4 through protein-protein interaction , we performed subcellular fractionation using postnuclear extracts of NPC cells expressing LMP1 or the ΔTM3–6 form ( Figure 5E ) , and performed immunofluorescence staining of FGD4 using NPC cells expressing various forms of LMP1 ( Figure 6A ) . As shown in Figure 6A , LMP1 appeared to recruit a fraction of FGD4 to the perinuclear regions ( Golgi apparatus ) where LMP1 was localized , compared to a primary cytoplasmic localization of FGD4 in control vector-transfected cells ( data not shown ) . In contrast , replacement ( CD40CT ) or deletion of the transmembrane domains ( in particular ΔTM3/4 and ΔTM3–6 ) of LMP1 substantially impaired its co-localization with FGD4 . Moreover , expression of full-length LMP1 led to FGD4 redistribution from fractions 2–5 to fractions 2–9 compared with the vector-expressing cells ( Figure 5E ) ; however , expression of the ΔTM3–6 form appeared not to result in this redistribution , suggesting that LMP1 interaction with FGD4 affects the intracellular distribution of FGD4 . Concomitantly , LMP1 expression led to a notable redistribution of β-actin ( a component of F-actin ) from fractions 1–18 to fractions 1–28 ( Figure 5E ) , implying a rearrangement of actin filaments . In contrast , expression of the ΔTM3–6 had no effect on this event , agreeable with the immunofluorescence staining data in Figure 2D . This redistribution was not due to a general alteration of protein localization , because the distribution of caveolin-1 ( CAV1 ) was comparable in all examined cells . While efforts to detect traces of Cdc42 in all fractions were not successful , these data functionally linked the LMP1-FGD4 interaction with actin redistribution , suggesting that LMP1 induces actin rearrangement by enhancing FGD4 activity toward Cdc42 . To investigate whether the above events contributed to cell motility , we conducted transwell migration assays using NPC cells expressing various forms of LMP1 . As shown in Figure 6B , LMP1 expression clearly induced cell motility compared with the vector control ( P<0 . 001; paired t-test ) ; however , deletion of the transmembrane domains apparently impaired this ability of LMP1 ( P<0 . 01; paired t-test ) . Moreover , the CD40CT chimera failed to induce cell motility compared with the vector control ( P<0 . 005; paired t-test ) , correlating with its eliminated activation of Cdc42 ( a 0 . 6-fold increase; Figure 6C ) . To verify the requirement of FGD4 and Cdc42 in LMP1-induced actin rearrangement , we used RNA interference to deplete FGD4 or Cdc42 from NPC cells , followed by expression of GFP-LMP1 and staining for F-actin . As shown in Figure 7A , expression of GFP-LMP1 but not GFP alone in control siRNA ( siCtrl ) -treated NPC cells resulted in filopodia formation at the plasma membrane . In contrast , knockdown of either FGD4 or Cdc42 led to the disappearance of such actin substructures resulting from GFP-LMP1 expression , demonstrating the equivalent roles of FGD4 and Cdc42 in LMP1-induced actin rearrangement . To investigate whether the above events result in increased cell motility , we next performed transwell migration assays using NPC cells co-expressing Flag-LMP1 and siRNA specific for FGD4 or Cdc42 . The data revealed that , compared to vector-transfected cells , expression of LMP1 in control siRNA ( siCtrl ) -treated cells increased cell motility ( Figure 7B , P = 0 . 0001 ) . This LMP1-induced cell motility was reduced by approximately 50% by knockdown of either FGD4 ( P = 0 . 008 ) or Cdc42 ( P = 0 . 005; Figure 7B ) . This reduced cell motility was associated with a decreased level of LMP1 activation of Cdc42; knockdown of FGD4 decreased the fold-increase in active Cdc42 from 4 . 7 to 0 . 9 , and knockdown of Cdc42 decreased active Cdc42 to undetectable levels ( Figure 7C ) . Despite its specificity toward Cdc42 , FGD4 has also been shown to indirectly promote Rac1 activation [44] . To clarify whether Rac1 activation was involved in LMP1-induced cell motility , we assessed the level of active Rac1 using NPC cells prepared from similar experiments as described above . The results revealed that neither knockdown of FGD4 nor reconstitution of FGD4 can affect the activation of Rac1 in NPC cells ( Figure S5A and S5B ) . In addition , knockdown of Cdc42 had no effect on the activation of Rac1 or RhoA ( Figure S5C ) . Furthermore , neither knockdown of Rac1 nor knockdown of RhoA can affect the LMP1-induced cell motility ( Figure S5D and S5E ) . Taken together , these data conclusively demonstrated that Rac1 and RhoA are not involved in LMP1-induced cell migration of NPC cells . To buttress the requirement of Cdc42 activity for the LMP1-induced cell motility , we conducted the transwell assays similar to those used in Figure 7B , except we co-expressed Flag-LMP1 with Flag-Cdc42DN ( a dominant negative mutant ) in control siRNA-treated cells to block Cdc42 activity , or co-expressed Flag-LMP1 with Flag-Cdc42CA in FGD4 siRNA-treated cells to bypass the effect of FGD4 depletion . As shown in Figure 7D , co-expression of LMP1 with Cdc42DN reduced LMP1-induced cell motility by approximately 50% ( P = 0 . 02; paired t-test ) , a reduction reminiscent of the effect of FGD4 depletion ( Figure 7B ) . Conversely , co-expression of LMP1 with Cdc42CA partially reversed the reduction in cell motility caused by FGD4 depletion ( Figure 7D , P = 0 . 004 ) , revealing that active Cdc42 was responsible for mediating the cell migration triggered by the LMP1-FGD4 axis . To explore the physiological relevance of LMP1 and FGD4 in NPC , we assessed LMP1 and FGD4 expression in NPC specimens . We first performed quantitative RT-PCR analyses using mRNAs isolated from specimens from 13 NPC patients and 14 controls . As shown in Figure 8A , LMP1 mRNA expression was exclusively detected in NPC specimens . Moreover , FGD4 mRNA was expressed at a higher level in NPC specimens than in controls ( 0 . 422±0 . 173 vs . 0 . 131±0 . 135; P = 0 . 0005; two-tailed Mann Whitney test ) ; however , no correlation was observed between the levels of FGD4 and LMP1 mRNA ( P = 1; Spearman test ) . To detect FGD4 and LMP1 proteins in NPC specimens , we next conducted immunohistochemical staining for FGD4 and LMP1 on consecutive NPC tissue sections from 48 NPC cases . Among them , both FGD4 and LMP1 were detectable in 29 cases and no correlation was shown between the levels of FGD4 and LMP1 ( P = 0 . 85; Spearman test ) . Despite this , it was notable that LMP1 and FGD4 exhibited similar staining patterns , prominently at the cell membrane , as shown in three representative cases ( Figure 8B ) . These data provide a potential physiological relevance for the LMP1-FGD4 interaction in NPC tissues . Collectively , our data reveal that LMP1 induces Cdc42 activation by directly binding to FGD4 , promoting actin rearrangement and , ultimately , cell migration .
Although Puls et al . have previously linked Cdc42 activation to LMP1-induced actin rearrangement in fibroblasts [17] , the molecular basis of Cdc42 activation by LMP1 and its role in EBV-associated malignancy remained to be elucidated . Here , we demonstrated that LMP1 induces Cdc42 activation in NPC cells via a direct interaction with the Cdc42-specific GEF , FGD4 , leading to actin remodeling and increased cell motility . We further verified the expression of LMP1 and FGD4 in NPC specimens , not only providing support for the physiological relevance of this mechanism but also linking FGD4 to tumorigenesis for the first time . Pathogenic microbes , including viruses , commonly hijack the host cell processes , such as cytoskeleton reorganization , to benefit their own survival in aspects of attachment , entry into cells , movement within and between cells , as well as vacuole formation and remodeling [45] , [46] . In addition to LMP1 , several viral proteins have been documented to affect the function of Rho protein , which is highly involved in the regulation of cytoskeleton organization . For instance , the E6 oncoprotein of high-risk human papilloma virus type 16 interacts with a binding partner of a GEF ( ARHGEF16 ) to coordinate Cdc42 activation [47] , and the Nef protein of human immunodeficiency virus type 1 recruits the GEF Vav1 into plasma membrane microdomains , where it associates with and activates Cdc42/PAK2 ( p21-activated kinase 2 ) [48] . To our knowledge , our evidence that LMP1 elicits Cdc42 activation via direct binding to a Cdc42-specific GEF is the first such demonstration for a viral oncoprotein . It has been proposed that FGD4 is targeted to a preexisting specific actin structure through its FAB domain [35] . LMP1 interaction with FGD4 likely promotes recruitment of FGD4 to the sites where LMP1 is present ( Figure 5E and 6A ) and thereby elicits Cdc42 activation in the vicinity of the actin structure associated with FGD4 ( Figures 2D and S2 ) . This process ultimately results in spatial reorganization of the actin cytoskeleton and regulates cell morphogenesis and cell motility [35] , [49] , in line with our observation that LMP1 induces the formation of filopodia at the cell surface ( Figures 2D and 7A ) and promotes cell migration via FGD4/Cdc42 ( Figure 7B–7E ) . In this study , we demonstrated that LMP1 directly interacts with FGD4 ( Figure 5C and 5D ) and enhances FGD4 activity toward Cdc42 ( Figure 4F ) . The LMP1-FGD4 interaction requires the transmembrane domains 3 and 4 of LMP1 and the PH1-to-PH2 domains ( phosphoinositide-binding domains ) of FGD4 ( Figures 4E , 4F , 5A , and 5B ) , indicating that the membrane/lipid association of both proteins may allow or enhance their interaction . We have previously identified that upon synthesis in the endoplasmic reticulum ( ER ) , LMP1 , through its transmembrane domains 3–6 , interacts with PRA1 ( the prenylated Rab acceptor 1 ) for transport from the ER to the Golgi apparatus [50] , an intracellular compartment where LMP1 primarily induces signaling pathways [50] , [51] . Deletion of the LMP1 transmembrane domains ( in particular 3–6 ) or knockdown of PRA1 leads to LMP1 retention in the pre-Golgi compartment ( Figures 2D , 5E , 6A , and S2B; [50] ) concomitant with a reduction of Cdc42 activation ( Figures 2B , 2C , and 6C; data not shown ) , implying that the proper localization of LMP1 is needed for its full activation of Cdc42 . These data elicit an argument that the impaired FGD4 interaction of the truncated LMP1 ( ΔTM3/4 and ΔTM3–6 ) actually arises from the impaired trafficking process instead of deletion of specific transmembrane domains . Here we propose that LMP1 interacts with FGD4 via the transmembrane domains 3 and 4 , and subsequently recruits and/or enhances FGD4 association with the membrane during the Golgi-directed trafficking of LMP1 ( Figure 8C ) . This model is supported by several lines of data as the following . First , expression of LMP1 but not its truncated form ( in particular ΔTM3–6 ) leads to FGD4 redistribution toward the LMP1-containing fractions as demonstrated by subcellular fractionation ( Figure 5E ) and immunofluorescence staining ( Figure 6A ) . These data suggest that LMP1 interaction with FGD4 indeed occurs prior to LMP1 localization to specific compartments . Second , the short N-terminus and the transmembrane domains 1 and 2 of LMP1 have been shown to be required for targeting of LMP1 to the lipid raft [52] , [53] . Deletion of these domains ( denote ΔNT and ΔTM1/2 ) also impedes the Golgi-directed trafficking of LMP1 ( data not shown; Figures 2D and 6A ) but does not abolish LMP1 interaction with FGD4 ( Figure 5A and 5B ) , revealing that this interaction is not restricted at certain compartments but rather relies on the transmembrane domains 3 and 4 of LMP1 . Agreeably , the CD40CT chimera fails to interact with FGD4 ( Figures 4F , 5A , and 5B ) despite its intact membrane localization ( Figures 2D and 6A ) . Like some GEFs whose activation is stimulated by protein-protein interaction [54]–[56] , it is assumed that FGD4 activation is potentiated by interaction with LMP1 or self-oligomerization ( Figure S4C ) . Further localization of LMP1 and FGD4 to lipid-rich compartments ( the Golgi apparatus and the plasma membrane ) likely strengthens the LMP1-FGD4 interaction and the membrane association of FGD4 , leading to an enhanced FGD4 activity toward Cdc42 ( the model in Figure 8C ) . Conceivably , the lipid in the membrane may enhance stimulation of FGD4 activities . In line with this , deletion of the PH1-to-PH2 domains of FGD4 substantially impairs its activation of Cdc42 ( Figure 4C ) , indicating that membrane/lipid association is involved in the regulation of FGD4 activity . As the truncated LMP1 ( ΔTM3/4 and ΔTM3–6 ) still retains ∼30% activation of Cdc42 , we speculate that the targeting of LMP1 to the lipid raft ( via the N-terminus and the transmembrane domains 1 and 2 ) may indirectly assist FGD4 to associate with the lipid/membrane . Further investigation will be needed to clarify the mechanisms in greater details . Importantly , LMP1-induced Cdc42 activation can be blunted by knockdown of FGD4 ( Figure 3 ) or overexpression of functionally impaired forms of FGD4 ( Figure 4D ) , indicating that LMP1 acts upstream of FGD4 rather than in parallel with it to induce Cdc42 activation . We propose that the functionally impaired forms of FGD4 ( ΔDH and PH1-2 ) inhibit LMP1-induced Cdc42 activation by competitively impeding the interaction between LMP1 and full-length FGD4 ( Table S1 ) . Intriguingly , the FAB-DH form of FGD4 , which interacts poorly with LMP1 , could still inhibit LMP1-induced Cdc42 activation , probably by competing with full-length FGD4 for access to Cdc42 . Apart from this , phosphatidylinositol 3-kinase ( PI3K ) has been linked to the translocation of FGD4 during infection of the enteric parasite Cryptosporidium parvum [57] . In this case , C . parvum infection induces recruitment of FGD4 to the host cell-parasite interface; this process , which results in Cdc42 activation , is dependent on PI3K and is required for C . parvum-induced actin remodeling and cellular invasion [57] . Although it has been demonstrated that LMP1 is able to act through its C-terminus to activate PI3K [58] , our study found that this region is dispensable for LMP1 activation of Cdc42 ( Figure 2B ) . Moreover , treatment of LMP1-expressing cells with the PI3K inhibitors , wortmannin and LY294002 , did not block the LMP1-induced Cdc42 activation ( data not shown ) , indicating that the LMP1-associated functional regulation of FGD4 involves the LMP1-FGD4 interaction instead of PI3K activity . We noted that LMP1 , together with FGD4 , is expressed heterogeneously at the plasma membrane as well as the Golgi apparatus ( Figure 6A ) and induces Cdc42 activation at these sites ( Figure S2A ) . It has been suggested that restricted localization of active Cdc42 is important for its distinct functions [59] , suggesting the possibility that LMP1 induction of Cdc42 activation at the plasma membrane and the Golgi apparatus serves distinct purposes . Given that Cdc42 also controls the intracellular protein trafficking , including the Golgi-to-ER retrograde transport [60] , [61] and protein exit from the trans-Golgi network [19] , [62] , it is conceivable that the LMP1-FGD4-Cdc42 cascade may have a regulatory role in intracellular protein transport apart from cell migration . We speculate that LMP1-induced Cdc42 activation may attenuate Golgi-to-ER retrograde protein transport [60] , promoting LMP1 retention at the Golgi apparatus and sustaining LMP1-mediated signaling . Further investigation will be needed to dissect the interplays between LMP1 and FGD4 within distinct compartments and elucidate how their functions are coordinated to affect cellular processes . Although FGD4 has been implicated in neural development [29] , [63] , [64] , we here delineate a potential role for FGD4 in NPC progression that is associated with LMP1 . We verified the expression of LMP1 and FGD4 in NPC tissues at both mRNA and protein levels using the quantitative RT-PCR and immunohistochemistry , respectively ( Figure 8A and 8B ) . Intriguingly , FGD4 mRNA expression appears to be elevated in cancerous tissues compared to the normal controls ( Figure 8A ) , although the underlying mechanism remains to be identified . In any case , it is conceivable that higher levels of FGD4 in NPC tissues lead to an increase in FGD4 function . In addition , it was recently shown that FGD1 , which is functionally related to FGD4 , is up-regulated in human prostate and breast cancer , and regulates cancer cell invasion by modulating Cdc42 activation in a cell model [65] . Accordingly , our findings highlight the importance of elucidating the mechanism by which FGD4 and its related proteins are dysregulated in tumor development .
This research followed the tenets of the Declaration of Helsinki and all subjects signed an informed consent approved by Institutional Review Board of Chang Gung Memorial Hospital before their participation in this study and for the use of tissue samples collected before treatment . NPC-TW01 , -TW02 , -TW04 and -TW06 cell lines , which had been established using NPC biopsy specimens collected from four NPC patients , respectively [66] , [67] , were cultured in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) at 37°C in a humidified 5% CO2 environment . The inducible LMP1-expressing 293 cell line ( 293 Tet-On ) , generated previously [50] , was grown on 1% collagen-coated dishes and maintained in DMEM supplemented with 10% FBS , 100 µg/ml G418 , and 50 µg/ml hygromycin . The human embryonic kidney cell line HEK293 , obtained from American Type Culture Collection ( ATCC; CRL 1573 ) , was grown on 1% collagen-coated dishes and cultured in DMEM containing 10% equine serum . The nasopharyngeal epithelium cell line ( NP69 ) and a stable LMP1-expressing NP69 cell line ( NP69-LMP1 ) were generous gifts from Dr . Sai-Wah Tsao ( University of Hong Kong , China ) . The growth medium used for NP69 and NP69-LMP1 cells has been previously described in detail [68] . Unless specified , all the reagents were purchased from Invitrogen ( Carlsbad , CA , USA ) . Freshly frozen biopsied tissues from 13 NPC patients and 14 control individuals with nasosinusitis , and slide-mounted consecutive NPC tissue sections from 48 NPC patients were collected at Chang Gung Memorial Hospital ( Lin-Kou , Taiwan ) . Clinical data for the NPC patients are presented in Table S2 . The anti-LMP1 monoclonal antibody ( S12 ) was affinity purified from a hybridoma . Mouse anti-Flag ( M2 ) and anti-HA ( 12CA5 ) antibodies were purchased from Sigma-Aldrich ( St . Louis , MO , USA ) ; a mouse anti-Myc tag ( 9E10 ) antibody was purchased from Cell Signaling Technology ( Danvers , MA , USA ) ; mouse anti-human Cdc42 and anti-Rac1 antibodies were purchased from BD Transduction Laboratories ( BD Biosciences , San Jose , CA , USA ) ; a mouse anti-human RhoA antibody was purchased from Santa Cruz Biotechnologies , Inc . ( Santa Cruz , CA , USA ) . Rabbit antibodies against human DOCK9 , intersectin-1 , FGD3 , and CAV1 were purchased from Santa Cruz Biotechnologies; rabbit anti-human FGD1 antibody was purchased from GeneTex ( Irvine , CA , USA ) ; rabbit anti-human FGD4 antibody was purchased from both Novus Biologicals ( Littleton , CO , USA ) and GeneTex ( Irvine , CA , USA ) . Fluorescein isothiocyanate ( FITC ) - , tetramethylrhodamine isothiocyanate ( TRITC ) - , and horseradish peroxidase ( HRP ) -conjugated secondary antibodies were purchased from BD Transduction Laboratories . N-terminally Flag-tagged LMP1 as well as its truncated derivatives were generated by ligation of PCR-amplified DNA fragments to HindIII/BamHI-treated pCMV2-Flag ( Kodak ) as described previously [50] . GFP-tagged LMP1 was generated by ligation of DNA fragments to HindIII/BamHI-treated pEGFP-C3 ( BD Biosciences ) . Full-length and truncated FGD4 constructs were generated by PCR using human FGD4 cDNA as a template; the resulting DNA fragments were subsequently inserted into pCMV2-Flag at the EcoR1/XbaI sites or into pCMV-Myc ( BD Biosciences ) at the KpnI/XbaI sites . Glutathione S-transferase ( GST ) -tagged full-length FGD4 construct was created by ligating the respective PCR-amplified DNA fragments into EcoRI/XhoI-treated pGEX 4T . 1 ( BD Biosciences ) . Plasmids encoding the GST fusion proteins , GST-CBD ( containing the active Cdc42-binding domain of Wiskott-Aldrich syndromewith protein , WASP; aa 201–321 ) , GST-PBD ( containing the active Rac1-binding domain of PAK1; aa 70–132 ) , and GST-RBD ( containing the active RhoA-binding domain of Rhotekin; aa 7–113 ) were gifts from Dr . Jacques Bertoglio ( INSERM U461 , Faculté de Pharmacie-Paris Sud , Chatenay-Malabry , France ) . pEGFP-CBD was generated by subcloning the DNA fragments encoding the CBD of WASP into pEGFP-C3 . Flag-tagged , constitutively active ( CA ) and dominant-negative ( DN ) versions of Cdc42 , Cdc42L61 and Cdc42N17 , respectively , were generated by site-directed mutagenesis using primers bearing the desired sequence changes . The DNA fragments were subsequently inserted into pCMV2-Flag at EcoRI/BamHI sites . For BRET2 assays , N-terminally Rluc-tagged LMP1 and GFP2-tagged FGD4 were generated by ligation of PCR-amplified DNA fragments into KpnI/BamHI-treated pRluc ( h ) -C2 and KpnI/XbaI-treated pGFP2-C1 vectors ( PerkinElmer Life and Analytical Sciences , MA , USA ) , respectively . Primer sequences used for cloning are provided in Table S3 . Unless specified , plasmid transfections were carried out using Lipofectamine ( Invitrogen ) , according to the manufacturer's instructions . Cells were incubated for 24 h prior to further treatments . For RNA interference , cells were co-transfected with 25 nM siRNAs as a set of four duplexes ( SMARTpool ) directed against a specific gene , together with plasmids indicated elsewhere , using Lipofectamine 2000 ( Invitrogen ) according to the manufacturer's instructions . Cells transfected with non-targeting duplexes ( siCtrl ) were used as a negative control . Cells were incubated for 48 h before further treatments , and the efficiency of gene silencing was estimated by quantitative RT-PCR and Western blotting . All siRNAs were purchased from Dharmacon ( CO , USA ) except Rac1 siRNAs , which were purchased from Invitrogen . Total RNA was purified using TRIzol ( Invitrogen ) and 1 µg of each sample was reverse transcribed using ImProm-II and Oligo ( dT ) 15 primers ( Promega ) . The resulting cDNAs were analyzed using the FastStart DNA Master SYBR Green I reagent ( Roche , Germany ) on a LightCycler ( Roche ) , according to the manufacturer's instructions . The reactions were incubated at 95°C for 10 min , followed by 45 cycles of 95°C for 10 s , 60°C for 5 s and 72°C for 10 s . Primer sequences used in RT-PCR experiments are presented in Table S4 . The levels of individual target mRNAs were normalized to the level of glyceraldehyde-3-phosphate dehydrogenase ( GADPH ) mRNA in each sample . pGEX construct-transformed Escherichia . coli ( strain BL21 ) were grown to mid-exponential phase , induced for 4 h with 0 . 5 mM isopropylthiogalactoside and lysed by sonicating in PBST buffer ( 1× phosphate-buffered saline [PBS] with 2 mM EDTA , 0 . 1% β-mercaptoethanol , 0 . 2 mM phenylmethylsulfonyl fluoride [PMSF] , and 5 mM benzamidine ) . The GST fusion proteins were purified from bacterial lysates by incubation with glutathione-coupled Sepharose beads ( Amersham Biosciences ) . The cellular level of the GTP-bound form of Cdc42 was determined using the GST-CBD pull-down assays . Briefly , at given incubation time after transfection , cells were cultured in serum-free media for an additional 6 h and lysed in Nonidet P-40 ( NP-40 ) lysis buffer ( 1% NP-40 , 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 1 mM Na3VO4 , 5 mM EDTA pH 8 . 0 , 10% glycerol , 10 µg/ml leupeptin , 10 µg/ml aprotinin , 1 mM PMSF ) . After centrifugation , the resulting lysates ( 1 mg protein ) were incubated with 20 µg of immobilized GST-CBD proteins at 4°C for 1 h . CBD-bound proteins were centrifuged , washed three times in NP-40 lysis buffer , boiled in sodium dodecyl sulfate ( SDS ) sample buffer and analyzed by Western blotting , as described below , using a Cdc42-specific antibody . GST-PBD and GST-RBD were used for precipitation of active forms of Rac1 and RhoA , respectively , following a similar procedure . Full-length LMP1 cDNA was constructed and cloned into pcDNA3 . 1 for in vitro transcription/translation . [35S]methionine-labeled proteins were generated using the TNT Coupled Reticulocyte Lysate System ( Promega , Madison , WI , USA ) , following the manufacturer's recommendation . Briefly , a reaction mixture ( total volume , 100 µl ) containing 50 µl of TNT Rabbit Reticulocyte Lysate , 4 µl of TNT Reaction Buffer , 2 µl of TNT T7 RNA Polymerase , 2 µl of Amino Acid Mixture Minus Methionine ( 1 mM ) , 2 µl of RNasin Ribonuclease Inhibitor ( Promega ) , 2 µg of pcDNA3 . 1-LMP1 , 4 µl of [35S]methionine ( 10 mCi/ml; Izotop , Hungary ) , 4 µl of Canine Pancreatic Microsomal Membrances ( Promega ) , and 30 µl of nuclease-free water was incubated at 30°C for 90 min . Equal amounts of the in vitro-translated , [35S]methionine-labeled LMP1 protein were incubated with 20 µg of GST or GST-FGD4 immobilized on glutathione-Sepharose beads , rotated overnight at 4°C , and then washed five times with NP-40 lysis buffer . The bound LMP1 proteins were resolved by SDS-PAGE on a 10% gel followed by autoradiography . Cells expressing Flag-tagged and/or Myc-tagged proteins were extracted in 1% NP-40 lysis buffer as described above and fractionated by centrifugation ( 10 , 000 g , 15 min at 4°C ) to obtain cell lysates . For co-immunoprecipitation , cell lysates ( 500 µg protein ) from NPC cells expressing Flag- or Myc-tagged proteins were incubated with 40 µl of a 50% ( w/v ) slurry of anti-Flag M2 agarose ( Sigma-Aldrich ) or with 25 µl of a 50% ( w/v ) slurry of anti-Myc affinity matrix ( Sigma-Aldrich ) , rotated overnight at 4°C , and then washed five times with NP-40 lysis buffers . For co-immunoprecipitation of endogenous FGD4 and LMP1 , cell lysates ( 1 mg protein ) from NPC cells expressing LMP1 were incubated with 2 µg of anti-FGD4 IgG or control IgG together with 20 µl of a 50% ( w/v ) slurry of protein G ( Amersham Biosciences ) , and further processed using procedures similar to those above . The resulting protein products were eluted with SDS sample buffer and analyzed by Western blotting using appropriate primary and secondary antibodies . Cells were lysed in 1% NP-40 lysis buffer as described above . Protein concentrations were determined using the Protein Assay Reagent ( Bio-Rad , CA , USA ) , and equal amounts of proteins ( 30–50 µg/lane ) were resolved by SDS-polyacrylamide gel electrophoresis ( SDS-PAGE ) on 7 . 5%–12% polyacrylamide gels . The proteins were then electro-transferred onto nitrocellulose ( NC ) membranes ( Amersham Biosciences ) . After blocking with 5% non-fat powdered milk in TBS , membranes were incubated with the respective primary antibodies overnight at 4°C . Membranes were then incubated with the appropriate HRP-conjugated secondary antibody for 1 h at room temperature . Protein bands were detected using enhanced chemiluminescence reagents ( Pierce ECL , Thermo Scientific ) and Fuji SuperRx film . NPC-TW01 cells grown on a 10 cm-dish were transfected with plasmids encoding donor Renilla luciferase ( Rluc ) -tagged LMP1 ( 1 µg ) and acceptor GFP2-tagged FGD4 ( 3 µg ) . Cells were detached at 24 h after transfection and washed with Dulbecco's Phosphate Buffered Saline ( D-PBS; Invitrogen ) and then resuspended in D-PBS to a final density of approximately 2×106 cells/ml . Approximately 1×105 cells/well were distributed in a 96-well white polystyrene microplate ( Conig , NY , USA ) . The DeepBlueC coelenterazine substrate ( PerkinElmer Life and Analytical Sciences ) was added to a final concentration of 5 µM , and bioluminescence emission was monitored immediately using a Fluoroskan Ascent FL microplate fluorometer ( Thermo Electron Corporation , MA , USA ) , which allows the sequential integration of signals detected in 410-nm and 515-nm windows . The BRET2 ratio is calculated as the following ratio: ( emission of transfected cells at 515 nm – emission of non-transfected cells at 515 nm ) / ( emission of transfected cells at 410 nm – emission of non-transfected cells at 410 nm ) . The expression level of each fusion protein was analyzed by Western blotting with appropriate antibodies . NPC-TW04 cells grown on a 10-cm dish were transfected with 1 µg of pFlag-LMP1 , pFlag-LMP1ΔTM3–6 , or pFlag-CMV2 vector and incubated for 24 h . Cells were then homogenized and centrifuged , and the resulting supernatant was layered onto a continuous sucrose gradient ( 10%–45% sucrose ) and centrifuged for 1 h at 55 , 000 rpm . using an SW55 rotor ( Beckman , Fullerton , CA , USA ) . The fractions were collected manually from the top of the gradient and 30 µl of every other fraction was subjected to Western blot analysis . Details of this assay have been described previously [50] , [69] . Cells grown on poly-L-lysine-coated coverslides were fixed with 4% formaldehyde , and permeabilized and blocked with 0 . 1% saponin containing 1% BSA for 20 min at room temperature . For co-staining of Flag-LMP1 and FGD4 , cells were incubated with anti-Flag antibody ( M2; 1∶200 dilution ) and anti-FGD4 antibody ( GeneTex; 1∶50 dilution ) for 2 h at room temperature , followed by incubation with the respective fluorophore-conjugated secondary antibody for 45 min . For actin filament staining , cells were incubated in TRITC-conjugated phalloidin ( 50 µg/ml; Sigma-Aldrich ) for 1 h after fixation . Nuclei were stained with 4′ , 6-diamidino-2-phenylindole ( DAPI; Sigma-Aldrich ) . All coverslides were mounted with the Vectashield reagent ( Vector Laboratories Inc . , CA , USA ) and visualized by confocal microscopy using a ZEISS LSM510 META laser-scanning confocal microscope ( Carl Zeiss , Germany ) with a 63×1 . 32 NA oil-immersion objective . For detection of LMP1 and FGD4 , consecutive slide-mounted NPC sections were first treated with proteinase K at room temperature for 15 min . Endogenous peroxidase activity was inhibited by incubating with 3% H2O2 ( DAKO ) . Nonspecific binding was blocked with Antibody Diluent and Background Reducing Component ( DAKO ) . Sections were then incubated with anti-FGD4 ( GeneTex; 1∶50 dilution ) and anti-LMP1 ( S12 , 1∶15 dilution ) antibodies at room temperature for 1 h . After a washing step , a HRP-conjugated secondary antibody was added and sections were incubated at room temperature for 20 min . Tissue sections were then treated with DAB reagent ( DAKO ) ; 3 , 3′-diaminobenzidine tetrahydrochloride was used as a chromogen . All images were acquired on an Olympus BX51 microscope ( Olympus , Japan ) . Expression of LMP1 and FGD4 was evaluated according to the simplified H score system [70] , which is based on the percentage of cell staining: 3 ( ≥90% ) , 2 ( 50%–89% ) , 1 ( 10%–49% ) , or 0 ( 0%–9% ) , and the intensity of cell staining: 3 ( high ) , 2 ( moderate ) , 1 ( low ) , or 0 ( no cell staining ) . The two scores were multiplied by each other and then divided by three to get the final score . The motility of NPC-TW02 cells was evaluated by transwell migration assays using a chemotaxis chamber ( Corning , NY , USA ) . Using calcium phosphate precipitation , NPC-TW02 cells were transfected with 1 µg of plasmid for Flag-LMP1 , its transmembrane domain-truncated forms , CD40CT chimera , or empty vector . For knockdown experiments , NPC-TW02 cells were co-transfected with siRNAs ( 37 . 5 nM ) directed against FGD4 ( siFGD4 ) , Cdc42 ( siCdc42 ) or non-targeting duplexes ( siCtrl ) , plus 1 µg of pFlag-LMP1 or pCMV-Flag vector . In a subset of assays , cells were co-transfected with pFlag-LMP1 ( 0 . 75 µg ) plus pFlag-Cdc42DN or pFlag-Cdc42CA ( 1 . 5 µg each ) . The oligonucleotides were mixed thoroughly in 250 µl of solution A ( 136 . 7 mM NaCl , 19 . 2 mM HEPES , pH 6 . 95 ) , 2 . 5 µl of solution B ( 57 . 6 mM Na2HPO4 ) , and 12 . 5 µl of solution C ( 2 . 5 M CaCl2 ) . After 30-min incubation at room temperature , the mixtures were added to cells and incubated for 6 h at 37°C . Twenty-four hours after transfection , cells were trypsinized and washed twice with serum-free DMEM , and then resuspended in 100 µl of serum-free DMEM and seeded into the insert chamber ( 3×105 cells ) . After a 20-h incubation , cells that had migrated to the opposite side of the insert ( immersed in DMEM supplemented with 10% FBS ) in the lower well were fixed and stained with crystal violet ( 1% crystal violet and 5% formaldehyde in 70% ethanol ) for 30 min , followed by washing twice with double-distilled H2O ( ddH2O ) to remove the background staining . The number of migrating cells was counted in images acquired at 200× magnification for each experiment and analyzed with NIH Image J software . Quantitative data were presented as means ± SDs for five independent experiments . Significance between groups was calculated using two-tailed paired t-tests . For clinical specimens , significance between groups was calculated using two-tailed Mann-Whitney tests . Correlation between groups was analyzed using two-tailed Spearman tests . A P-value<0 . 05 was considered statistically significant . The Entrez Gene ID numbers for genes or proteins described in this study are as follows: 5176215 ( LMP1 ) , 121512 ( FGD4 ) , 998 ( Cdc42 ) , 5879 ( Rac1 ) , 387 ( RhoA ) , 60 ( β-actin ) , 2245 ( FGD1 ) , 89846 ( FGD3 ) , 6453 ( Intersectin-1 ) , 23348 ( DOCK9 ) , 857 ( CAV1 ) , 7124 ( TNF ) , 3552 ( IL-1α ) , 958 ( CD40 ) .
|
Epstein-Barr virus ( EBV ) is closely associated with human malignancies , including nasopharyngeal carcinoma ( NPC ) . Among EBV-expressed genes , latent membrane protein 1 ( LMP1 ) has been detected in most NPC tissues and has the ability to transform cell growth and drive cell migration , both of which are highly associated with tumorigenesis and tumor progression . Previous reports have demonstrated that cell migration primarily involves cytoskeleton rearrangement , and the RhoGTPase Cdc42 is known to actively mediate such rearrangement processes . Using LMP1-expressing NPC cells , we discovered that LMP1 induces Cdc42 activation by directly binding to FGD4 , a positive regulator of Cdc42 , thereby promoting motility of NPC cells . The observed correlation between FGD4 and LMP1 expression in NPC tissues provides support of physiological relevance . Notably , FGD4 has recently been shown to be responsible for a type of inherited neural disease . Our findings not only provide a novel insight into EBV pathogenesis , but also suggest a role for FGD4 in tumorigenesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"proteomics"
] |
2012
|
Epstein-Barr Virus-Encoded LMP1 Interacts with FGD4 to Activate Cdc42 and Thereby Promote Migration of Nasopharyngeal Carcinoma Cells
|
This work introduces a number of algebraic topology approaches , including multi-component persistent homology , multi-level persistent homology , and electrostatic persistence for the representation , characterization , and description of small molecules and biomolecular complexes . In contrast to the conventional persistent homology , multi-component persistent homology retains critical chemical and biological information during the topological simplification of biomolecular geometric complexity . Multi-level persistent homology enables a tailored topological description of inter- and/or intra-molecular interactions of interest . Electrostatic persistence incorporates partial charge information into topological invariants . These topological methods are paired with Wasserstein distance to characterize similarities between molecules and are further integrated with a variety of machine learning algorithms , including k-nearest neighbors , ensemble of trees , and deep convolutional neural networks , to manifest their descriptive and predictive powers for protein-ligand binding analysis and virtual screening of small molecules . Extensive numerical experiments involving 4 , 414 protein-ligand complexes from the PDBBind database and 128 , 374 ligand-target and decoy-target pairs in the DUD database are performed to test respectively the scoring power and the discriminatory power of the proposed topological learning strategies . It is demonstrated that the present topological learning outperforms other existing methods in protein-ligand binding affinity prediction and ligand-decoy discrimination .
Arguably , machine learning has become one of the most important developments in data science and artificial intelligence . With its ability to extract features of various levels hierarchically , deep convolutional neural networks ( CNNs ) have made breakthroughs in image processing , video , audio , and computer vision [1 , 2] , whereas recurrent neural networks have found success in analyzing sequential data , such as text and speech [3–6] . Deep learning algorithms are able to automatically extract high-level features and discover intricate patterns in large data sets . In general , one of the major advantages of machine learning algorithms is their ability to deal with large and diverse data sets and uncover complicated relationships . Recently , machine learning has become an indispensable tool in biomolecular data analysis and structural bioinformatics . Almost every computational problem in molecular biophysics and biology , such as the predictions of solvation free energy , solubility , partition coefficient , protein-ligand binding affinities , mutation induced protein stability change , molecular multipolar electrostatics , virtual screening , etc . , has machine learning based approaches that are either parallel or complementary to their physics based counterparts . The success of deep learning has fueled the rapid growth in several areas of biological science [3 , 5 , 6] , including bioactivity of small-molecule drugs [7–10] and genetics [11 , 12] , where large data sets are available . A key component of a learning machine based on biomolecular structures is featurization , that is translating the 3D structures of biomolecules to features . While the degrees of freedom of the original biomolecular structures are large and vary among different molecules , it is almost inevitable that information loss happens with dimension reduction during featurization . Besides the choice of learning models , the performance of a predictor heavily depends on how the features are extracted . Although deep learning has been known to be powerful for the automatic extraction of features from original inputs such as images , deep learning based models directly taking biomolecules as inputs are not as competitive as the state-of-art machine learning models with carefully designed features , due to the intrinsic complexity of biomolecules [13] . Biomolecules can be characterized by geometric features , electrostatic features , high-level ( residue and global level ) features , and amino-acid sequence features based on physical , chemical , and biological understandings [14] . Geometric features , such as coordinates , distances , angles , surface areas [15–17] and curvatures [18–21] , are important descriptors of biomolecules [22–24] . However , geometric features often involve too much structural detail and are frequently computationally intractable for large biomolecular data sets . Electrostatic features include atomic partial charges , Coulomb potentials , atomic electrostatic solvation energies , and polarizable multipolar electrostatics [25] . These descriptors become essential for highly charged biomolecular systems , such as nucleic acid polymers and some protein-ligand complexes . High-level features refer to pKa values of ionizable groups and neighborhood amino acid compositions , such as the involvement of hydrophobic , polar , positively charged , negatively charged , and special case residues . Sequence features consist of secondary structures , position-specific scoring matrix ( PSSM ) , and co-evolution information . Sequence features and annotations provide a rich resource for bioinformatics analysis of biomolecular systems . Topology offers a new unconventional representation of biomolecules . Topology can describe biomolecules in a variety of ways [26] . Some of the most powerful topological features are obtained from multi-component persistent homology or element specific persistent homology ( ESPH ) [14 , 27] . Recently , we carried out a comprehensive comparison of the performance of geometric features , electrostatic features , high-level features , sequence features and topological features , for the prediction of mutation induced protein folding free energy changes of four mutation data sets [14] . Surprisingly , topological features outperform all the other features [14] . Unlike geometry , topology is well known for its power of simplification to geometric complexity [28–35] . The global description generated by classical topology is based on the concept of neighborhood and connectedness . If a space can be continuously deformed to another , they are considered to have the same topological features . In this sense , topology can not distinguish between a folded protein and its unfolded form if only covalent bonds are considered . Such property prevents the use of classical topology for the characterization of biomolecular structures . Instead of using topology to describe a single configuration of connectivity , persistent homology scans over a sequence of configurations induced by a filtration parameter and renders a sequence of topological invariants , which partially captures part of geometric features . Persistent homology has been applied to biomolecular systems in our earlier works [26] . In mathematics , persistent homology is a relatively new branch of algebraic topology [29 , 36] . When dealing with proteins and small molecules , it is conventional to consider atoms as point clouds . For a given point cloud data set , one type of persistent homology turns each point into a sphere with their radii systematically increasing . The corresponding topological invariants and their persistence over the varying radius values can be computed . Therefore , this method embeds multiscale geometric information in topological invariants to achieve an interplay between geometry and topology . Consequently , persistent homology captures topological structures continuously over a range of spatial scales . It is called persistent homology because at each given radius , topological invariants , i . e . , Betti numbers , are practically calculated by means of homology groups . In the past decade , much theoretical formulation [37–46] and many computational algorithms [47–52] have been developed . One-dimensional ( 1D ) topological invariants generated from persistent homology is often visualized by persistence barcodes [53 , 54] and persistence diagrams [55] . In recent years , multidimensional persistence has attracted much attention [43 , 56] in hope that it can better characterize the data shape when there are multiple measurements of interest . Persistent homology has been applied to various fields , including image/signal analysis [57–62] , chaotic dynamics verification [63 , 64] , sensor networks [65] , complex networks [66 , 67] , data analysis [68–72] , shape recognition [73–75] , and computational biology [76–79] . Compared with traditional computational topology [80–82] and/or computational homology , persistent homology inherently adds an additional dimension , i . e . , the filtration parameter . The filtration parameter can be used to embed important geometric or quantitative information into topological invariants . As such , the importance of retaining geometric information in topological analysis has been recognized [83] , and persistent homology has been advocated as a new approach for handling big and high dimensional data sets [54 , 68 , 84–86] . Recently , we have introduced persistent homology for mathematical modeling and/or prediction of nano-particles , protein unfolding , and other aspects of biomolecules [26 , 87] . We proposed the molecular topological fingerprint ( TF ) to reveal topology-function relationships in protein folding and protein flexibility [26] . We established some of the first quantitative topological analyses in our persistent homology based predictions of the curvature energy of fullerene isomers [87 , 88] . We have also shown correlation between persistence barcodes and energies computed with physical models during molecular dynamics experiments [26] . Moreover , we have introduced the first differential geometry based persistent homology that utilizes partial differential equations ( PDEs ) in filtration [88] . Most recently , we have developed a topological representation to address additional measurements of interest , by stacking the persistent homology outputs from a sequence of frames in molecular dynamics or a sequence of different resolutions [89 , 90] . We have also introduced one of the first uses of topological fingerprints for resolving ill-posed inverse problems in cryo-EM structure determination [91] . In 2015 , we constructed one of the first integrations of topology and machine-learning and applied it to protein classification involving tens of thousands of proteins and hundreds of tasks [92] . We also developed persistent-homology based software for the automatic detection of protein cavities and binding pockets [93] . Despite much success , it was found that persistent homology has a limited characterization power for proteins and protein complexes , when applied directly to biomolecules [92] . Essentially , biomolecules are not only complex in their geometric constitution , but also intricate in biological constitution . In fact , the biological constitution is essential to biomolecular structure and function . Persistent homology that is designed to reduce the geometric complexity of a biomolecule neglects biological information . To overcome this difficulty , we have introduced multi-component persistent homology or element specific persistent homology ( ESPH ) to recognize the chemical constitution during the topological simplification of biomolecular geometric complexity [14 , 27 , 94] . In ESPH , the atoms of a specific set of element types in a biomolecule are selected so that specific chemical information , such as hydrophobicity or hydrophilicity , is emphasized in each selection . Our ESPH is not only able to outperform other geometric and electrostatic representations in large and diverse data sets , but is also able to shed light on the molecular mechanism of protein-ligand binding , such as the relative importance of hydrogen bond , hydrophilicity and hydrophobicity at various spatial ranges [27] . The objective of the present work is to further explore the representability and reduction power of multi-component persistent homology for biomolecules and small molecules . To this end , we take a combinatorial approach to scan a variety of element combinations and examine the characterization power of these components . Additionally , we also propose a multi-level persistence to study the topological properties of non-covalent bond interactions . This approach enables us to devise persistent homology to describe the interactions of interest between atoms that are connected by weak non-covalent bonds and delivers richer representation especially for small molecules . Moreover , realizing that electrostatics are of paramount importance in biomolecules and to enhance the power of our topological representation , we introduce electrostatic persistence , which embeds charge information in topological invariants , as a new class of features in multi-component persistent homology . The aforementioned approaches can be realized via the modification of the distance matrix with a more abstract setting , for example , Vietoris-Rips complex . The complexity reduction is guaranteed in the 1D topological representation of 3D biomolecular structures . Obviously , the multi-component persistent homology representation of biomolecule leads to a higher machine learning dimensionality compared to the original single component persistent homology for a biomolecule . Therefore , it is subject to overfitting or overlearning problem in machine learning theory . Fortunately , gradient boosting trees ( GBT ) method is relatively insensitive to redundant high dimensional topological features [14] . Finally , since the components can be arranged as a new dimension ordered by their feature importance , multi-component persistent homology barcodes are naturally a two-dimensional ( 2D ) representation of biomolecules . Such a 2D representation can be easily used as image-like input data in a deep CNN architecture , with different topological dimensions , i . e . , 0 , 1 , and 2 , being treated as channels . Such a topological deep learning approach addresses the nonlinear interactions among important element combinations while keeping the information from less important ones . Barcode space metrics , such as bottleneck distance and more generally , Wasserstein distance [95 , 96] , offer a direct description of similarity between molecules and can be readily used with nearest neighbor regression or kernel based methods . The performance of Wasserstein distance for protein-ligand binding affinity predictions is examined in this work . After assessing the new method’s ability to represent small molecules and protein-compound complexes , the derived model is used for virtual screening . Virtual screening computationally screens a collection of small molecules to identify those who can potentially bind to the protein target . There are mainly two types of virtual screening which are ligand-based and structure-based . Ligand-based approaches depend on a measurement of similarity among small molecules using either 2D or 3D structural information of small molecules . Structure-based approaches attempt to dock the small molecule candidate to the protein target and determine if the candidate is a potential ligand based on the top docking poses . The performance of structure-based virtual screening methods heavily depends on the quality of the docking method and the accuracy of the post-docking scoring method . Our effort focuses on the development of a topology based method for the latter part . It has been shown that using machine learning or deep learning based methods to rescore the docking poses can significantly boost the performance [97 , 98] . For the models such as ensemble of trees and classical neural networks , carefully constructed features are needed . For example , a neural network based method NNScore uses a collection of derived features such as the count of hydrogen bonds and electrostatics of close contacts to describe the protein-compound complex [97] . Another class of deep learning based methods feed lower level features to deep neural networks and relies on the neural networks to automatically extract higher-level features . For example , DeepVS first computes features on each atom involved in the docking interface and feed this information to a deep neural network starting with convolution layers to hierarchically extract higher-level features [98] . The rest of this manuscript is organized as follows . Section Methods is devoted to introducing methods and algorithms . We present multi-component persistent homology , multi-level interactive persistent homology , vectorized persistent homology representation and electrostatic persistence . These formulations are crucial for the representability of persistent homology for biomolecules . Machine learning algorithms associated with the present topological data analysis are briefly discussed . Results are presented in Section Results . We first consider the characterization of small molecules . More precisely , the cross-validation of protein-ligand binding affinities prediction via solely ligand topological fingerprints is studied . We illustrate the excellent representability of our multi-component persistent homology by a comparison with a method using physics based descriptors . Additionally , we investigate the representational power of the proposed topological method on a few benchmark protein-ligand binding affinity data sets , namely , PDBBind v2007 , PDBBind v2013 , PDBBind v2015 and PDBBind v2016 [99] . These data sets contain thousands of protein-ligand complexes and have been extensively studied in the literature . Results indicate that multi-component persistent homology offers one of most powerful representations of protein-ligand binding systems . The aforementioned study of the characterization of small molecules and protein-ligand complexes leads to an optimal selection of features and models to be used for virtual screening . Finally , we consider the directory of useful decoys ( DUD ) database to examine the representability of our multi-component persistent homology for virtual screening to distinguish actives from non-actives . The DUD data set used in this work has a total of 128 , 374 ligand-target and decoy-target pairs containing 3961 active ligand-target pairs , and involves 40 protein targets from six families . A large number of state-of-the-art virtual screening methods have been applied to this data set . We demonstrate that the present multi-component persistent homology outperforms other methods with reported results on this benchmark . This paper ends with a conclusion .
In this section , we address the representation of small molecules by element specific persistent homology , especially the proposed multi-level persistent homology designed for small molecules . In this section , we develop topological representations of protein-ligand complexes . In this section , we examine the performance of the proposed method for the main application in this paper , which is structure-based virtual screening which involves protein-compound complexes obtained by attempting to dock the candidates to the target proteins . The dataset is much larger than the two applications on protein-ligand binding affinity prediction which makes parameter tuning very time consuming . Therefore , the best performing procedures in ligand-based binding affinity prediction and protein-ligand-complex-based binding affinity prediction are applied in this virtual screening application .
We conduct several experiments on ligand based protein-ligand binding affinity prediction in this section which leads to the final models . To examine the strength and weakness of different sets of features and models , we first show a statistics fact of the S1322 data set of 7 protein clusters in Fig 2 . The details of the S1322 data set is given in Section Results/Ligand based protein-ligand binding affinity prediction . All the gradient boosting trees models take the setup described in Section Methods/Machine learning algorithms/Gradient boosting trees . Having demonstrated the representational power of the present topological learning method for characterizing small molecules , we further examine the method on the task of characterizing protein-ligand complex . Biologically , we consider the same task , i . e . , the prediction of protein-ligand binding affinity , with a different approach that is based on the structural information of the protein-ligand complexes . Only gradient boosting trees and deep convolutional neural network algorithms are used in this section . All the gradient boosting trees models take the setup described in Section Methods/Machine learning algorithms/Gradient boosting trees . In the present topological learning study , we use four versions of PDBBind core sets as our test sets . For each test set , the corresponding refined set , excluding the core set , is used as the training set . In our final model TopVS reported in Table 6 , we use topological descriptors of both protein-compound interactions and only the compounds ( i . e . , ligands and decoys ) and take a consensus model on top of several ensemble of trees models and a deep learning model . We have also tested the behavior of our topological learning model TopVS-ML using either one of the aforementioned descriptions . The tests are done with TopVS-ML because that TopVS-DL is much more time consuming . When only topological descriptor of small molecules are used , which falls into the category of ligand-based virtual screening , an AUC of 0 . 81 is achieved . For the topological learning model using only the descriptions of protein-ligand interactions , an AUC of 0 . 77 is achieved . An AUC of 0 . 83 is obtained with a model combining both sets of descriptors which is better than each individual performance , suggesting that the two groups of descriptors are complementary to each other and are both important for achieving satisfactory results . The marginal improvement made by protein-compound complexes maybe due to the various docking quality . Similar situation was encountered by a deep learning method [98] . For the targets with high quality results by Autodock Vina ( AUC of ADV > 0 . 8 ) , the ligand-based features achieve an AUC of 0 . 81 and the complex-based features achieve an AUC of 0 . 86 . On the other hand , for the targets with low quality results by Autodock Vina ( AUC of ADV < 0 . 5 ) , the ligand-based features achieve an AUC of 0 . 82 and the complex-based features achieve an AUC of 0 . 74 . The results of these cases are listed in S1 Text , Tables H and I . This observation suggests that the performance of features describing the interactions and the geometry of protein-compounds complexes highly depends on the quality of docking results . Our model with small molecular descriptors delivers an AUC of 0 . 81 , which is comparably well to the other top performing methods . The performance of this model is also competitive in the regime of protein-ligand binding affinity prediction based on experimentally solved complex structures as is shown in Section Discussion/Ligand based protein-ligand binding affinity prediction . These results suggest that topology based small molecule characterization proposed in this work is potentially useful in other applications involving small molecules , such as predictions of toxicity , solubility and partition coefficient of small molecules . Persistent homology is a relatively new branch of algebraic topology and is one of the main tools in topological data analysis . The topological simplification of biomolecular systems was a major motivation of the earlier persistent homology development [29 , 36] . Persistent homology has been applied to computational biology [76 , 77 , 77–79] , including our efforts [26 , 87–91 , 93] . However , the predictive power of primitive persistent homology was limited in early topological learning applications [92] . To address this challenge , we have recently introduced element specific persistent homology to retain chemical and biological information during the topological abstraction of biomolecules [14 , 27 , 94] . The resulting topological learning approach offers competitive predictions of protein-ligand binding affinity and mutation induced protein stability changes . However , persistent homology based approaches for small molecules have not been developed and its representability and predictive powers for the interaction of small molecules with macromolecules have not been extensively studied . The present work further introduces multi-component persistent homology , multi-level persistent homology and electrostatic persistence for chemical and biological characterization , analysis and modeling . Multi-component persistent homology takes a combinatorial approach to create possible element specific topological representations . Multi-level persistent homology allows tailored topological descriptions of any desirable interaction in biomolecules which is especially useful for small molecules . Electrostatic persistence incorporates partial charges that are essential to biomolecules into topological invariants . These approaches are implemented via the appropriate construction of the distance matrix for filtration . The representation power and reduction power of multi-component persistent homology , multi-level persistent homology and electrostatic persistence are validated by two databases , namely PDBBind [99] and DUD [107 , 108] . PDBBind involves more than 4 , 000 high quality protein-ligand complexes and DUD contains 128 , 374 compound-target pairs . Two classes of problems are used to test the proposed topological methods , including the prediction of protein-ligand binding affinities and the discrimination of active ligands from decoys ( virtual screening ) . In both problems , we examine the representability of proposed topological learning methods on small molecules , which are somewhat more difficult to describe by persistent homology due to their chemical diversity , variability and sensitivity . Additionally , these methods are tested on their ability to handle the full protein-ligand complexes . Advanced machine learning methods , including Wasserstein metric based k-nearest neighbors ( KNNs ) , gradient boosting trees ( GBT ) , random forest ( RF ) , extra trees ( ET ) and deep convolutional neural networks ( CNN ) are utilized in the present work to facilitate the proposed topological methods , rendering advanced topological learning algorithms for quantitative and qualitative biomolecular predictions . The thorough examination of the method on the prediction of binding affinity for experimentally solved protein-ligand complexes leads to a structure-based virtual screening method , TopVS , which outperforms other methods . The feature sets introduced in this work for small molecules and protein-ligand complexes can be extended to other applications such as 3D-structure based prediction of toxicity , solubility , and partition coefficient for small molecules and complex structure based prediction of protein-nucleic acid binding and protein-protein binding affinities .
The concept of persistent homology is built on the mathematical concept of homology , which associates a sequence of algebraic objects , such as abelian groups , to topological spaces . For discrete data such as atomic coordinates in biomolecules , algebraic groups can be defined via simplicial complexes , which are constructed from simplices , generalizations of the geometric notion of nodes , edges , triangles and tetrahedrons to arbitrarily high dimensions . Homology characterizes the topological connectivity of geometric objects in terms of topological invariants , i . e . , Betti numbers , which are used to distinguish topological spaces by counting k-dimensional holes . Betti-0 , Betti-1 and Betti-2 , respectively , represent independent components , rings and cavities in a physical sense . In persistent homology , the generators in the homology groups are tracked along with a filtration parameter , such as the radius of a ball or the level set of a hypersurface function , that continuously varies over a range of values . Therefore , persistent homology is induced by the filtration . For a given biomolecule , the change and the persistence of topological invariants over the filtration offer a unique characterization . These concepts are very briefly discussed below . For more detailed theory and algorithms , the interested readers are referred to a book on computational topology [117] . The development of persistent homology was motivated by its potential in the dimensionality reduction , abstraction and simplification of biomolcular complexity [36] . In the early applications of persistent homology to biomolecules , emphasis was given on major or global features ( long-persistent features ) to derive descriptive tools . For example , persistent homology was used to identify the tunnel in a Gramicidin A channel [36] and to study membrane fusion [118] . For the predictive modeling of biomolecules , features of a wide range of scales might all be important to the target quantity [26] . At the global scale , the biomolecular conformation should be captured . At the intermediate scale , the smaller intra-domain cavities need to be identified . At the most local scale , the important substructures should be addressed , such as the pyrrolidine in the side chain of proline . These features of different scales can be reflected by barcodes with different centers and persistences . Therefore , applications in biomolecules can make a more exhaustive use of persistent homology [26 , 87] , compared to some other applications where only global features matter while most local features are mapped to noise . Earlier use of persistent homology was focused on qualitative analysis . Only recently had persistent homology been devised as a quantitative tool [26 , 87] . While the aforementioned applications are descriptive and regression based analysis , we have also applied persistent homology to predictive modeling of biomolecules [92] . However , biomolecules are both structurally and biologically complex . Their geometric and biological complexities include covalent bonds , non-covalent interactions , effects of chirality , cis and trans distinctions , multi-leveled protein structures , and protein-ligand and protein-nucleic acid complexes . Covering a large range of spatial scales is not enough for a powerful model . The biological details should also be explored . We address the underlying biology and physics by modifying the distance function and selecting various sets of atoms according to element types , to describe different interactions . Some biological considerations are discussed in this section . One important issue is how to protect chemical and biological information during the topological simplification . As mentioned earlier , one should not treat different types of atoms as homogeneous points in a point cloud data . To this end , element specific persistent homology or multi-component persistent homology has been proposed to retain biological information in topological analysis [14 , 27 , 94] . The element selection is similar to a predefined vertex color configuration for graphs . When all atoms are passed to persistent homology algorithms , the information extracted mainly reflects the overall geometric arrangement of a biomoelcule at different spatial scales . By passing only atoms of certain element types or of certain roles to the persistent homology analysis , different types of interactions or geometric arrangements can be revealed . In protein-ligand binding modeling , the selection of all carbon atoms characterizes the hydrophobic interaction network whilst the selection of all nitrogen and/or oxygen atoms characterizes hydrophilic network and the network of potential hydrogen bonds . In the protein structural analysis , computation on all atoms can identify geometric voids inside the protein which may suggest structural instability and computation on only Cα atoms reveals the overall structure of amino acid backbones . In addition , combination of various selections of atoms based on element types provides very detailed description of the biomolecular system and the hidden relationships from the structure to function can then be learned by machine learning algorithms . This may lead to the discovery of important interactions not realized as a prior . This can be realized by passing the set of atoms of the selected element types to the persistent homology computation . This concept is used with the various definitions of distance matrix discussed as follows . Biomolecular systems are not only complex in geometry , but also in chemistry and biology . To effectively describe complex biomolecular systems , it is necessary to modify the filtration process . There are three commonly used filtrations , namely , radius filtration , distance matrix filtration , and density filtration , for biomolecules [26 , 90] . A distance matrix defined with smoothed cutoff functions was proposed in our earlier work to deal with interactions within a spatial scale of interest in biomolecules [26] . In the present work , we introduce more distance matrices to enhance the representational power of persistent homology and to cover some important interactions that were not covered in our earlier works . The distance matrices can be used with a more abstract construction of simplicial complexes , such as Vietoris-Rips complex . Barcode representation of topological invariants offers a visualization of persistent homology analysis . In machine learning analysis , we convert the barcode representation of topological invariants into structured feature arrays for machine learning . To this end , we introduce two methods , i . e . , counts in bins , barcode statistics , and persistence diagram slice and statistics , to generate feature vectors from sets of barcodes . These methods are discussed below . Python code is given in S1 Code for the generation of features used in the final models in the Results section . Three machine learning algorithms , including k-nearest neighbors ( KNN ) regression , gradient boosting trees and deep convolutional neural networks , are integrated with our topological representations to construct topological learning algorithms .
|
Conventional persistent homology neglects chemical and biological information during the topological abstraction and thus has limited representational power for complex chemical and biological systems . In terms of methodological development , we introduce advanced persistent homology approaches for the characterization of small molecular structures which can capture subtle structural difference . We also introduce electrostatic persistent homology to embed physics in topological invariants . These approaches encipher physics , chemistry and biology , such as hydrogen bonds , electrostatics , van der Waals interactions , hydrophobicity and hydrophilicity , into topological fingerprints which , although cannot literally recast into physical interpretations , are ideally suitable for machine learning , particularly deep learning , rendering topological learning algorithms . In terms of applications , we construct a structure-based virtual screening model which outperforms other existing methods . This competitive model on the DUD database is derived by assessing the performance of a comprehensive collection of topological approaches proposed in this work and introduced in our earlier work , on the PDBBind database . The topological features constructed in this work can readily be applied to other biomolecular problems where the characterization of proteins or small molecules is needed .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biotechnology",
"machine",
"learning",
"algorithms",
"chemical",
"compounds",
"neural",
"networks",
"small",
"molecules",
"applied",
"mathematics",
"electricity",
"neuroscience",
"organic",
"compounds",
"simulation",
"and",
"modeling",
"algorithms",
"electrostatics",
"mathematics",
"algebra",
"artificial",
"intelligence",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"chemistry",
"physics",
"chemical",
"elements",
"topological",
"invariants",
"organic",
"chemistry",
"algebraic",
"topology",
"biology",
"and",
"life",
"sciences",
"topology",
"physical",
"sciences",
"machine",
"learning"
] |
2018
|
Representability of algebraic topology for biomolecules in machine learning based scoring and virtual screening
|
The evidence for the existence of genetic susceptibility variants for the common form of hypertension ( “essential hypertension” ) remains weak and inconsistent . We sought genetic variants underlying blood pressure ( BP ) by conducting a genome-wide association study ( GWAS ) among African Americans , a population group in the United States that is disproportionately affected by hypertension and associated complications , including stroke and kidney diseases . Using a dense panel of over 800 , 000 SNPs in a discovery sample of 1 , 017 African Americans from the Washington , D . C . , metropolitan region , we identified multiple SNPs reaching genome-wide significance for systolic BP in or near the genes: PMS1 , SLC24A4 , YWHA7 , IPO7 , and CACANA1H . Two of these genes , SLC24A4 ( a sodium/potassium/calcium exchanger ) and CACNA1H ( a voltage-dependent calcium channel ) , are potential candidate genes for BP regulation and the latter is a drug target for a class of calcium channel blockers . No variant reached genome wide significance for association with diastolic BP ( top scoring SNP rs1867226 , p = 5 . 8×10−7 ) or with hypertension as a binary trait ( top scoring SNP rs9791170 , p = 5 . 1×10−7 ) . We replicated some of the significant SNPs in a sample of West Africans . Pathway analysis revealed that genes harboring top-scoring variants cluster in pathways and networks of biologic relevance to hypertension and BP regulation . This is the first GWAS for hypertension and BP in an African American population . The findings suggests that , in addition to or in lieu of relying solely on replicated variants of moderate-to-large effect reaching genome-wide significance , pathway and network approaches may be useful in identifying and prioritizing candidate genes/loci for further experiments .
Genome wide association studies ( GWAS ) on large scale population samples have been remarkably successful in uncovering novel susceptibility loci for a wide range of complex human diseases including type 2 diabetes , coronary artery disease , dyslipidemia , breast cancer , obesity-related traits , prostate cancer and Crohn's disease [1] . These notable success stories represent significant advances in the global effort to understanding the genetic basis of common human diseases . However , this has not been the case for hypertension , a common human disease affecting over one billion people worldwide [2] and a major contributor to cerebrovascular accidents , myocardial infarction , congestive cardiac failure and chronic renal disease [3] , [4] . The earliest published GWAS that specifically sought associations for hypertension and/or BP traits ( the Wellcome Trust Case Control Consortium ( WTCCC ) [5] and the Diabetes Genetics Initiative ( DGI ) [6] studies ) did not find any genetic variant significantly associated with hypertension at the genome wide level . While these two studies have some limitations , these negative findings have strengthened the notion that multiple rare independent variants may account for a large fraction of BP variation [7] , a situation in which GWAS ( designed to work best in “common disease , common variant” scenarios ) would be less useful . A further note is that these studies were conducted in European populations and it is unknown if similar studies in populations with non-European ancestry would yield different insights . In the present study , we conducted a GWAS of BP among African Americans enrolled in the Washington DC metropolitan region of the United States . In comparison with other population groups in the United States , African Americans suffer a disproportionate burden of hypertension and its complications . A priori , we considered that: ( 1 ) Gene variants associated with BP variation among normotensive individuals may not be exactly the same set identified as those associated with persistently elevated blood pressure ( i . e . “hypertension” ) ; ( 2 ) Since the clinical definition of hypertension utilizes elevation of either the systolic blood pressure ( SBP ) or diastolic blood pressure ( DBP ) , those with hypertension are a heterogenous group comprising those with isolated SBP elevation , those with isolated DBP elevation and those with both . This heterogeneity is likely to be reflected in genetic associations for each of these traits ( SBP , DBP , hypertension ) ; ( 3 ) Individual response to hypertension treatment varies greatly thereby making it a real possibility that statistical adjustment of SBP and DBP for treatment ( e . g . adding a fixed quantity to measured BP ) among treated hypertensive individuals [8] , may mask real associations in GWAS . ( 4 ) The evidence so far from GWAS of hypertension and BP suggest that there may be few or no variants with large effects , implying that p values may be modest compared to those reported for other traits . For these reasons , we chose to: 1 ) conduct a case-control association study for hypertension; 2 ) conduct an association study for SBP and DBP among normotensive individuals; 3 ) use pathway-based analyses of the GWAS data to determine if the variants most strongly associated with BP phenotypes cluster in pathways and networks that are of biological relevance to BP regulation . Using this strategy , we hoped to maximize the chances of discovering loci influencing hypertension susceptibility and/or normal BP control .
Ethical approval for the study was obtained from the Howard University Institutional Review Board ( IRB ) . All subjects provided written informed consent for the collection of samples and subsequent analysis . This study was conducted according to the principles expressed in the Declaration of Helsinki . The subjects studied were all participants in the Howard University Family Study ( HUFS ) , a population based family study of African Americans in the Washington metropolitan area . The major objectives of the HUFS were to: 1 ) enroll and examine a randomly ascertained cohort of African-American families , along with a set of unrelated individuals , from the Washington DC metropolitan area to study the genetic and environmental basis of common complex diseases including hypertension , obesity and associated phenotypes; 2 ) to characterize study participants for anthropometry ( including weight , height , waist and hip circumferences , body composition measures ) and BP; and 3 ) evaluate the association between genetic variants and selected traits ( hypertension , BP and obesity ) . Participants were sought through door-to-door canvassing , advertisements in local print media and at health fairs and other community gatherings . In order to maximize the utility of this cohort for the study of multiple common traits , families were not ascertained based on any phenotype . During a clinical examination , demographic information was collected by interview . Weight , height , waist circumference and hip circumference were measured using standard methods as follows: Weight was measured in light clothes on an electronic scale to the nearest 0 . 1 kg , and height was measured with a stadiometer to the nearest 0 . 1 cm . Body mass index ( BMI ) was computed as weight in kg divided by the square of the height in meters . Waist circumference was measured to the nearest 0 . 1 cm at the narrowest part of the torso as seen from the anterior aspect . BP was measured in the sitting position using an oscillometric device ( Omron ) . Three BP readings were taken with a ten minute interval between readings . The reported SBP and DBP readings were the average of the second and third readings . Pulse pressure ( PP ) was calculated as the difference between the SBP and DBP . Hypertension status was defined as SBP> = 140 mmHg and/or DBP> = 90 mmHg and/or treatment with antihypertensive medication . In the overall cohort , the frequency of hypertension was 35% and among those that were hypertensive , 64% were on antihypertensive medication at the time of the study . Genome-wide genotyping was performed using the Affymetrix® Genome-Wide Human SNP Array 6 . 0 [9] . DNA samples were prepared and hybridized following the manufacturer's instructions . After processing , chips were scanned and genotype calls were made using the Birdseed 2 algorithm [9] , [10] . All samples used in the analysis achieved a chip wide call rate of ≥95% . Individual SNPs were excluded if they had a call rate of less than 95% ( n = 41 , 885 ) across all individuals , a minor allele frequency < = 0 . 01 ( n = 19 , 154 ) or had a Hardy-Weinberg equilibrium ( HWE ) test p of <1×10−3 ( n = 6 , 317 ) . The current analysis focused on the 808 , 465 autosomal SNPs that passed these filters . The average call rate for this set of SNPs in these individuals was 99 . 5% . The concordance of blind duplicates was 99 . 74% . Focused , lower-throughput genotyping for replication was carried out using Sequenom Homogenous MassEXTEND or iPLEX Gold SBE assays at the National Human Genome Research Institute ( NHGRI ) . Evidence for population stratification or structure was sought by conducting non parametric clustering of genotypes using the AWClust algorithm [11] . All the subjects formed one cluster with a few outliers . Individuals identified as outliers were removed before association analysis , which in this case resulted in the removal of 7 individuals from a sample of 1024 individuals , for a final sample size of 1017 individuals . Further checks were conducted during the association analysis on the 1017 participants as follows: first , the genomic control ( GC ) method was used to compute the genomic inflation factor for each analysis and was determined to be 1 . 007 for hypertension , 1 . 001 for SBP and 0 . 998 for DBP , showing minimal evidence of inflation of the test statistic due to stratification . As expected , the GC-adjusted test statistics were virtually identical to the unadjusted values . Second , a Q-Q plot was used to visualize the distribution of the test statistic for each trait analysis and these again showed no evidence of population stratification . Finally , principal components ( PC ) were computed using the eigenstrat method [12] . Based on examination of the scree plot ( shown in Figure S1 ) , the first two PCs were retained and used as covariates during the association analysis in order to adjust for any potential residual population stratification . Hypertension was analyzed as a binary trait ( cases versus controls ) using a logistic regression model under an additive model with adjustment for age , sex , BMI , and the first 2 PCs of the genotypes . Given that treatment for hypertension alters BP values , we conducted the association analysis for SBP and DBP in two ways . First , a normotensives-only analysis was carried out using linear regression models with age , sex , BMI , and the first 2 PCs of the genotypes as covariates . This approach was designed to uncover any BP associated loci without the “noise” effect of treatment . Second , an analysis of the whole dataset was carried out using the same covariates and also adjusting for the effect of treatment . All association analyses were performed using the PLINK software package , v1 . 04 [13] . Association for the replication sample of 980 unrelated non-diabetic West Africans enrolled as part of the Africa America Diabetes ( AADM ) Study [14] , [15] was done the same way . P-values for the discovery ( African American ) sample and the replication ( West African ) samples were combined using the Meta-Analysis Tool for genome-wide association scans , METAL ( http://www . sph . umich . edu/csg/abecasis/Metal/ ) . The METAL algorithm calculates a z-statistic for each marker summarizing the magnitude and direction of the effect relative to the reference allele in each sample and then calculates an overall z-statistic and p value from the weighted average of the statistics . Weights are proportional to the square-root of the sample size of each study . SNPs that showed an association p-value less than 1e-04 for each trait were mapped to genes within 5 kB using Ensembl ( http://www . ensembl . org ) . The resulting gene list for the hypertension phenotype and for SBP and DBP , each with corresponding Entrez IDs , were entered into MetaCore ( http://www . genego . com ) and tested for enrichment in Maps , Diseases , Gene Ontology ( GO ) processes and GeneGO processes . MetaCore uses a hypergeometric model to determine the significance of enrichment .
The subjects comprised 1017 individuals ( 419 men , 598 women ) , including 509 cases of hypertension and 508 normotensive controls . Hypertensive subjects were older ( mean age 54 years versus 41 years ) and heavier ( mean BMI 31 . 7 kg/m2 versus 29 . 3 kg/m2 ) than the normotensive subjects . As expected , mean BP was higher and showed more variance among hypertensive compared to normotensive subjects ( Table 1 ) . The distribution of association p-values ( Manhattan plot ) for the three traits is shown in Figure 1 and the QQ plots in Figure 2 . The ten top scoring SNPs for association with hypertension are shown in Table 2 . The SNP with the lowest p-value ( 5 . 10×10−7 ) for this trait was rs9791170 located on chromosome 5 . This intergenic SNP is about 6 kbp upstream of the P4HA2 ( GeneID 8974 ) gene . However , it did not show genome-wide significance ( Bonferroni-corrected p = 0 . 412 ) for association with hypertension; neither did any of the other SNPs ( see Table S1 for a list of the top-scoring associations for hypertension as a binary trait ) . In contrast to the hypertension results , the T allele of the rs5743185 SNP , an intronic SNP in the PMS1 ( GeneID 5378 ) gene , was strongly associated with SBP ( nominal p = 2 . 09×10−11 , Bonferroni-corrected p = 1 . 69×10−5 ) among normotensive individuals . Other SNPs that showed significant association with SBP among normotensive individuals , each with a Bonferroni-corrected p value of ≤0 . 05 , include: rs3751664 ( a non-synonymous coding SNP in CACNA1H ( GeneID 8912 ) ) , rs11160059 ( an intronic SNP inSLC24A4 ( GeneID 123041 ) ) , rs17365948 ( an intronic SNP in YWHAZ ( GeneID 7534 ) ) , rs12279202 ( an intronic SNP in IPO7 ( GeneID 10527 ) ) and rs1687730 ( an intergenic SNP , 12 kb from AL365365 . 23 , a pseudogene ) , – Table 3 . Repeating these analyses for the whole sample , with adjustment for treatment effects , did not change the top-scoring characteristics of these six SNPs ( as shown in Table S2 ) . The mean effect size on SBP associated with the at-risk alleles of these six SNPs ( estimated from the linear model adjusted for age , sex , BMI and PCs among normotensive individuals only ) was ∼5–6 mmHg . If independent , each SNP significant after Bonferroni-correction correction would be associated with ∼5% of the variance in SBP . The full list of the top-scoring associations for SBP is shown in Table S3 . Haplotype analysis did not show any haplotype association that reached the significance of the single locus analyses ( data not shown ) . Two-locus interaction analyses between the SNPs that were significant or marginally so did not show any significant interactions , with the lowest p-value 0 . 115 ( between rs17315498 and rs11160059 ) . For DBP , the A allele of rs1867226 ( an intronic SNP in PRC1 ( GeneID 9055 ) ) showed the lowest p-value ( 5 . 8×10−7 ) . However , neither this nor any other association reached genome wide significance ( Table 3; see Table S4 for list of top-scoring SNPs for DBP ) . Pathway analysis revealed a number of significant pathways and processes that are associated with SBP and DBP ( Table 4 ) . Examination of each of these pathways and processes showed annotations with obvious cardiovascular implications ( for example , Development_PIP3 signaling in cardiac myocytes , Transport_Potassium transport and Development_Blood vessel morphogenesis ) and several pathways and processes that are enriched for genes involved in hypertension and/or blood pressure regulation . As a case in point , the top scoring pathway – Development_Role of HDAC and calcium/clamodulin-dependent kinase ( CaMK ) in control of skeletal myogenesis- ( Figure 3 ) contains the calcium-gated channels CACNA1E and CACNA1H , IGF-1 , and AKT , each of which is known to play a role in mechanisms of BP regulation , hypertension and/or complications of hypertension ( including left ventricular hypertrophy ) [16]–[21] . The top-scoring pathways for hypertension alone are shown in Table S5 . A total number of 17 SNPs were carried forward for replication in the sample of 980 unrelated non-diabetic West Africans ( 366 HTN cases , 614 normotensive subjects; mean age 49 ( SD 12 ) years , mean BMI 25 . 1 ( SD 6 ) kg/m2 ) enrolled as part of the Africa America Diabetes Mellitus ( AADM ) study [14] , [15] . These 17 SNPs comprised the top-scoring seven SNPs for SBP , the top scoring three SNPs for DBP , two SNPs that had low p-values ( p<1×10−4 ) for both SBP and DBP , as well five of the top-scoring SNPs for HTN as a dichotomous trait . Five ( rs5743185 , rs3751664 , rs12279202 , rs11659639 and rs6543012 ) were monomorphic in the West African sample . The results for the other twelve SNPs analyzed under an additive model and with adjustment for age , sex , BMI , ethnic group and treatment for hypertension ( adjustment for treatment for SBP and DBP only ) are shown in Table 5 . Three SNPs ( rs1867226 , rs1550576 and rs8039294 ) were significant at a p-value of <0 . 05 among the West Africans . The combined analysis showed that five SNPs , including rs11160059 ( SLC24A4 ) , were significantly associated with the trait and with the same direction of effect in both samples . Two recent GWA studies [22] , [23] identified the STK39 and CDH13 genes as being significantly associated with BP . We therefore looked for evidence for association of SNPs in these genes with SBP and DBP in the present study . Each of these genes showed multiple SNPs associated with SBP and DBP at a p<0 . 05 ( Table 6 ) . Of note , STK39 had many more significantly associated SNPs ( 9/136 for SBP , 33/136 for DBP ) than would be expected by chance at a nominal p value of 0 . 05 ( 7/136 ) . All of the STK39 SBP-associated SNPs and 24 of the 33 DBP-associated SNPs were in the LD bins 1 and 2 ( chr2:168 , 699 , 002-168 , 788 , 544 ) reported in the Amish . We also looked for in silico replication of this study's top SBP-associated SNPs in the Diabetes Genetics Initiative ( DGI ) [6] GWAS , which to our knowledge , was the first published GWAS for BP . Out of the five genes harboring the top scoring SNPs for SBP in this study , three had variants with low p-values associated with SBP under the same additive model in the DGI ( Table S6 ) . These were SLC24A4 ( rs7142084 , p = 0 . 0017 ) , IPO7 ( rs7480643 , p = 0 . 009 ) and PMS1 ( rs3791767 , p = 0 . 014 ) .
While this paper was under review , two GWAS for hypertension , SBP and DBP in subjects of European descent were published [35] , [36] . One of these studies [36] also reported finding significant hits in the CACNB2 gene for hypertension and DBP , a gene with a high-scoring variant for hypertension in the present study ( Table S1 ) and in the PMS1 gene for hypertension and SBP , which scored highly for SBP in this study ( Table 3 ) .
|
Despite intense research , the genetic risk factors for essential hypertension and blood pressure ( BP ) regulation have not been identified with consistency . We conducted a genome wide association scan using over 800 , 000 genetic markers in an African American sample of 1 , 017 adults in the Washington , D . C . , area of the United States . We found evidence to suggest that genetic variants in several genes , including PMS1 , SLC24A4 , YWHA7 , IPO7 , and CACANA1H , are significantly associated with systolic BP levels . From our previous knowledge of human physiology , two of these genes have potential roles to play in BP regulation . The evidence for genetic variants influencing diastolic BP levels and hypertension status was weaker and inconclusive . To our knowledge , this is the first study that has used a genome-wide association approach to study hypertension and BP in an African American population , a minority group that experiences hypertension more frequently and more severely than other population groups in the United States . The findings will be useful to other researchers seeking to advance our understanding of the genetic factors that influence BP with the hope that these insights will eventually translate to new and better treatment options for hypertension in African Americans and other global populations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"cardiovascular",
"disorders/hypertension",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/complex",
"traits"
] |
2009
|
A Genome-Wide Association Study of Hypertension and Blood Pressure in African Americans
|
Many protein-coding genes identified by genome sequencing remain without functional annotation or biological context . Here we define a novel protein-coding gene , Nmf9 , based on a forward genetic screen for neurological function . ENU-induced and genome-edited null mutations in mice produce deficits in vestibular function , fear learning and circadian behavior , which correlated with Nmf9 expression in inner ear , amygdala , and suprachiasmatic nuclei . Homologous genes from unicellular organisms and invertebrate animals predict interactions with small GTPases , but the corresponding domains are absent in mammalian Nmf9 . Intriguingly , homozygotes for null mutations in the Drosophila homolog , CG45058 , show profound locomotor defects and premature death , while heterozygotes show striking effects on sleep and activity phenotypes . These results link a novel gene orthology group to discrete neurological functions , and show conserved requirement across wide phylogenetic distance and domain level structural changes .
The biological functions of many protein-coding genes remain unknown , often despite conservation across considerable evolutionary distance . Such “orphan” molecules may include genes whose functions affect fitness but not overt phenotypes in experimental settings , genes that affect overt but under-studied phenotypes , and/or genes that have been difficult to study due to unusual molecular properties , such as low level expression or poor physical annotation in genomes of well-studied organisms . Forward genetics offers one entry point into functional studies by highlighting those genes whose alterations produce measurable effects [1 , 2] . A key challenge in de-orphanizing novel genes is to integrate their functions with observable sites of action , organism-level phenotypes and known cellular activities . Nmf9 and its homologs were essentially unknown prior to our studies . The nmf9 mutation was recovered in a N-ethyl-N-nitrosourea ( ENU ) screen at the Neuroscience Mutagenesis Facility ( NMF ) of the Jackson laboratory [3 , 4] based on tremor and vestibular phenotypes . Several exons of the gene we identify as Nmf9 had been systematically annotated as Ankfn1 ( for ANK and FN3 domain containing ) , while other exons annotated as separate genes were identified only by clone names , based on non-overlapping partial cDNA clones from high-throughput screens [5 , 6] . We selected nmf9 for study as part of a long-standing interest in mice with unusual tremors and ataxias [7–11] . The unusual conservation pattern we identified among Nmf9 homologs , along with recent innovations in genome editing [12–18] , motivated additional studies to test conservation of function in mice and flies , including the generation of equivalent mutations at the same amino acid position in both species . While our work was in progress , two groups reported identification of the Drosophila homolog in very different contexts . A transposon-based screen for low-sleeping mutants by Mark Wu and colleagues found insertions in a poorly conserved , variably included 5’ region of the gene , but concluded based on antibody studies that the resulting wide-awake ( wake ) alleles were null [19] . The same group also found that the fly gene interacts with and promotes cell surface localization of the fly GABA-A receptor encoded by Resistance-to-dieldrin ( Rdl ) , through physical and genetic interactions . Independently , an RNAi screen for genes required for proper segregation of Numb during asymmetric cell division of sensory organ precursors carried out by Juergen Knoblich and colleagues identified the same gene , which they named Banderuola ( Bnd ) . Intragenic deletion of most of the Bnd coding sequence resulted in pupal and early adult lethality [20] . This group also showed physical , genetic and functional interactions with Discs-large 1 ( Dlg1 ) , a prototypical membrane-associated guanylate kinase ( MAGUK ) . We will refer to the Drosophila gene by its current annotation symbol , CG45058 , and to 5’ P element alleles as wake . Here we use positional cloning , expression-guided behavioral studies , deep evolutionary constraint , and genome editing to define novel activities of Nmf9 and its Drosophila homolog . Positional cloning of mouse nmf9 identified a single point mutation at a splice donor sequence , resulting in exon skipping of a frame-shifting exon in Nmf9 . The expression pattern of Nmf9 in mouse brain suggested neural circuits that might be compromised in mutant animals . Behavioral tests confirmed deficits in vestibular function , fear learning , and circadian behavior . Female mice were more severely affected than males for several phenotypes . Sliding window analysis of relative constraint across the protein coding sequence showed that the skipped exon encodes the most evolutionarily constrained peptide sequence among Nmf9 homologs and genome editing of a single glycine to alanine at this site was sufficient to generate a non-complementing allele . Genome editing of the Drosophila homolog at three different sites produced null alleles that resulted in premature lethality and severe locomotor retardation in homozygotes . Remarkably , heterozygous flies showed mild activity and sleep-related phenotypes . Our studies thus show the first genetic , behavioral and molecular information for mammalian Nmf9 , highlight functional importance of the most conserved sequence among its homologs , and resolve competing views regarding CG45058 null phenotypes . These genetic , behavioral , and comparative studies provide a foundation for understanding activities of Nmf9 homologues in broad context .
The nmf9 mutation was recognized in a chemical mutagenesis screen based on tremor and vestibular phenotypes . In our evaluation , vestibular signs included circling , nodding , and head tilt ( S1 Video ) and abnormal landing and forced swim tests . Not every mutant was abnormal in every test , but all mutant animals were abnormal in at least one test and by visual inspection relative to co-isogenic non-mutant littermates . Visible head nodding , hyperactivity , and tremor were enhanced by light vertical acceleration or extended handling . In the landing test , animals were suspended by their tails and scored for trunk curling and attempts to rotate , rather than reaching for ground , by an investigator blinded to genotype . Mutant animals showed a significant increase in frequency of both trunk curling and rotation compared to littermate controls ( Fig 1A ) . In the forced swim test , most control littermates swam with their snout above water for ≥1 min . , while nmf9 mutants typically were unable to remain righted above the surface and had to be rescued before 30 sec . to prevent drowning ( Fig 1B ) . Differences between genotypes were more pronounced in females than males for vestibular phenotypes . The frequency and severity of circling and head nodding in mutant animals increased progressively from 21 days to 6 months ( Fig 1C and 1D ) , though hyperactivity and tremor did not ( Fig 1E and 1F ) ; these phenotypes were essentially absent from control littermates . Histologically , however , mutant animals had grossly normal inner ear structures and did not show hearing impairment ( S1 Fig ) , suggesting a mature functional , rather than a gross morphological , basis for vestibular defects . We mapped the nmf9 mutation by recombination in ~1000 F2 progeny from crosses between B6-nmf9 and four mapping strains ( Fig 2A ) . Intercross progeny included 332 from AKR/J , 245 from BALB/cJ , 371 from C3H/HeJ , and 119 from DBA/2J that were typed by PCR ( [21] and S1 Table ) . The data confirmed a fully penetrant , recessive phenotype with no evidence of segregating modifier genes . Exclusion mapping placed nmf9 within a 1 . 3 Mb interval ( Chr11: 88 , 240 , 426–89 , 538 , 743 in GRCm38/mm10 assembly ) that included several well-annotated genes . We previously estimated the frequency of induced nucleotide mutations in the NMF screen as ~1/Mb [4] , which predicts a very low probability of confounding functional mutations in an interval of this size [22] . Sanger sequencing of all canonical and EST exons ( Fig 2B ) in mutant and littermate control identified a single mutation: a G-to-A transition in the splice donor U1 binding site of a frame-shifting exon of predicted gene Ankfn1 ( Fig 2C ) . RT-PCR , homologous cDNAs from other species , and detailed in situ hybridization patterns indicated that the major transcript of this gene comprised both Ankfn1 and 4932411E22Rik annotations as well as additional 5’ exons ( Figs 2D and S2 ) . We refer to this transcript and locus as Nmf9 to avoid confusion with the narrower definition of Ankfn1 . RT-PCR across EST and predicted exons showed low overall abundance ( requiring polyA+ RNA for detection from whole brain ) , skipping of the frame-shifting exon adjacent to the mutated U1 site , and variable utilization of alternative 5’ ends ( Fig 2D ) . RNA gel blot hybridization showed a single major size form and confirmed that while the transcript had very low abundance in control brain , the mutant transcript had still lower levels in nmf9 homozygote brains , consistent with predicted nonsense-mediated decay for the frame-shifted splice product ( Fig 2E ) . While it remains possible that other transcripts exist at low levels , our data support a single major open reading frame , with some variation at the 5’ end . To define potential sites of Nmf9 action , we examined its pattern of expression . In situ hybridization to both male and female embryos and RT-PCR from dissected tissue showed Nmf9 expression largely restricted to the nervous system . In vestibular inner ear , Nmf9 expression fully overlapped Atoh1 ( Fig 3A ) , a well-studied marker for hair-cell progenitors [23 , 24] . In situ hybridization showed that Nmf9 was expressed as early as E14 . 5 in inner ear , nasal epithelium , ventricular zone , and the spinal cord ( S2 Fig ) . In adult brain , both our data and data from the Allen Brain Atlas [25] showed expression enriched in a few centers ( Fig 3B ) , including the accessory olfactory bulb ( OB ) , piriform cortex ( PC ) , lateral septum ( LS ) , amygdala ( AMY ) , suprachiasmatic nucleus ( SCN ) , and modest enrichment in ventral medial hypothalamus ( VMH ) . Probes corresponding to Ankfn1 and 4932411E22Rik annotated exons showed the same detailed pattern in our data ( S2 Fig ) and in comparable data from the Allen Brain Atlas , further supporting a single transcription unit . Identification of Nmf9 pattern in mouse brain allows behavioral tests of neurological function for each of the major sites of expression , providing a structured approach to defining additional phenotypes in nmf9 mice . To probe the functional significance of Nmf9 expression in neuroanatomical structures outside of the vestibular system , we tested nmf9 and littermate ( +/+ ) control mice on selected behavioral tasks . Each task included enough male and female animals to assess potential sex differences , as female mice had shown earlier and stronger vestibular defects than males . A summary of statistical analyses is presented in S2 Table . Homologs of Nmf9 across wide taxonomic boundaries showed strong patterns of conservation that highlight select domains , as well as lineage-restricted modulation of domain architecture ( Figs 6A and S4 ) . Homologs were identified by reciprocal BLAST/BLAT searches in sequenced genomes of nearly all metazoan lineages , including placozoa and porifera , and in at least some choanoflagellates and filasterea , sister groups to animals that diverged from the lineage leading to animals after the split between animals and fungi . Choanoflagellate ( M . brevicollis , S . rosetta ) and filasterea ( C . owczarzaki ) homologs included an N-terminal CRIB domain ( associated with binding to Cdc42/Rac subfamily small GTPases ) , a C-terminal Ras-association ( RA ) domain , or neither . A single instance in fungi–comprising only a choanoflagellate-like copy of the conserved , non-motif domains–was found in Mortierella verticillata , but not in basal or sister lineages , and might therefore represent a horizontal gene transfer event rather than an earlier origin of the gene genealogy . Sequenced invertebrate animal genomes had a single Nmf9 homolog , except for an apparent loss in the urochordate lineage ( 0/4 species ) . Invertebrate homologs included an RA domain , except for the single placozoa sequence and a few genomes with incomplete assembly , but none included a CRIB domain . Jawed vertebrates basal to mammals contained two homologs: an ancestral copy with the RA domain and a derived copy without it . Mammalian genomes had only the derived copy . Most animal homologs occur in poorly annotated regions of their respective genome assemblies , limiting the number of complete sequences available among homologs we examined . Analysis of evolutionary constraint among 14 animal species spaced by known evolutionary distance [26] confirmed conservation of the annotated ankyrin and fibronectin type 3 motifs , but also predicts three additional regions of unknown biological function under equal or stronger constraint ( Fig 6B ) . Novel domain 2 was the most conserved region in the entire protein , including a GLYLGYLK sequence that is nearly invariant among animals and whose first glycine was the most highly conserved residue ( Fig 6C ) . This region contained no motif annotation in current databases nor predicted post-translational modification sites . In a detailed analysis of 113 homologous sequences , Domain 2 was the most highly conserved sequence both among ancestral homologs and among derived homologs , with the GLYLGYLK signature providing the strongest sliding window support in both ( S5 Fig ) . In situ hybridization studies in diverse animal species suggested the potential for a conserved neuronal function . In contrast to mouse , the homolog in D . melanogaster appeared more broadly expressed during development , including both CNS and tissues outside of the nervous system ( S6 Fig ) . Expression in the sea anemone N . vectensis also showed neural expression: in the planula , highest expression was in the apical tuft , a larval chemosensory organ , and in the pharyngeal nerve ring; in the polyp stage highest expression was in the endoderm , the predominant region of the net-like adult nervous system [27] . Ancestral and derived homologs may also differ in expression pattern . In the model fish D . rerio , the ancestral homolog appeared broadly expressed , as in D . melanogaster , while the derived homolog appeared more strictly localized , including distinct expression in the developing inner ear . To confirm the gene identification of Nmf9 and to test the functional importance of conserved domain 2 , we used CRISPR/Cas9-mediated genome editing in mouse one-cell embryos [17 , 28] to target mutations to the conserved GLYLGYLK region . We recovered 26 G0 animals , of which 24 survived to adulthood . Sequencing of a 550-bp PCR product encompassing the target site showed that 21 G0 animals were edited on both alleles , 4 were edited on only one allele , and one was not determined . Among 21 edited on both alleles , 17 appeared to be homozygous ( or possibly heterozygous to an allele that precluded amplification , such as a large deletion ) . G0 mice with both alleles edited to frameshift or other clearly deleterious lesions phenocopied the overt vestibular phenotypes of nmf9 ( Table 1 ) . Four predicted pathogenic mutations were used for complementation tests with the original nmf9 allele ( Fig 7A and Table 2 ) . In aggregate , G1 and later progeny carrying one edited allele heterozygous to the original nmf9 mutation failed to complement in both the landing test and forced swim test ( Fig 7B and 7C ) , confirming correct gene identification of nmf9 . The first glycine of the GLYLGYLK sequence was conserved through metazoa and holozoan sister groups . Strikingly , an alanine substitution at this residue failed to complement nmf9 . Although homozygotes for this allele did not show a strong phenotype , G-to-A/nmf9 heterozygotes had the tremor , hyperactivity , and vestibular dysfunction characteristic of nmf9 homozygotes ( Table 2 and S2 Video ) . These results support the functional importance of the highly conserved domain at a neurological level , even under the relaxed constraints of inbred laboratory mice . To test conservation of function at an organismal level , we similarly edited the Drosophila homolog of Nmf9 , CG45058 , at conserved sites in three different exons: the first ANK repeat , the FN3 domain , and conserved domain 2 ( Fig 8A ) . Apparent null and other predicted deleterious mutations induced at each site showed a consistent , severe adult locomotor phenotype and reduced viability . Newly eclosed homozygotes were predominantly stationary and most could be placed unanesthetized on a cardboard surface without the animal attempting to fly or walk ( Fig 8B , 8C and 8D ) . Most held their wings at unusual postures , either up or slightly down from the horizontal plane . When these animals did move , they typically fell over onto their backs and kicked their legs without apparent coordination . Mutant adults had very poor survival , even in uncrowded , horizontal vials ( Fig 8E ) . In addition to three independent sets of alleles with consistent phenotype , deficiency mapping with two distinct alleles induced at conserved domain 2 confirmed that the severe locomotor defects map to the CG45058 locus as a loss of function ( S7 Fig ) . These data conflict with an interpretation that wake alleles of this locus , which do not have strong locomotor phenotypes , are null [19] . In contrast , these data are fully consistent with the bnd intragenic deletion allele as null [20] . Since wake alleles were previously reported to have reduced sleep and nmf9 mice had abnormal circadian behaviors , we made related measurements for heterozygotes of our engineered null mutations . Consistent with previous reports , wake alleles showed decreased sleep in both sexes ( Fig 9 ) . While male null heterozygotes also showed reduced daily sleep , surprisingly , females showed slightly increased daily sleep . Null heterozygotes also had increased latencies to first sleep bout after lights on ( daytime sleep latency ) , and heterozygote females had increased latency to first sleep bout after lights off ( nighttime sleep latency ) as well . Increased latencies were consistent with , but less severe than , wake alleles ( Fig 9 ) . Reduced sleep and increased sleep latencies could not be accounted for by an increase in rate of movement since heterozygotes for engineered null alleles showed moderately reduced waking activity levels , as did P-element wake alleles ( Fig 8F ) . Statistical summaries for these tests are given in S5 Table . Together , these results show a strong requirement for CG45058 function for viability , activity level , and sleep-related measures; show sexual dimorphism for impact on daily sleep; and resolve a conflict between previous reports of CG45058 null phenotypes .
Through a combination of genetic and molecular approaches , we showed that Nmf9 encodes a novel yet highly conserved protein that is functionally important to a distinct set of neurological phenotypes in both mice and flies . Starting with the mutation , we identified Nmf9 transcripts using positional cloning and experimental validation of predicted exons . RT-PCR and in situ hybridization experiments defined timing and location of gene expression . Sites of expression predicted , in addition to readily observed vestibular abnormalities , substantial phenotypes in fear learning and circadian behavior that were not obvious from the initial description or gross observation . Sequence information from >100 animal homologs identified distinct regions of strong evolutionary constraint , including ANK and FN3 motifs and three novel segments that lack motif annotation . Induced mutations at most conserved peptide sequence among the novel conserved domains produced strong phenotypes in both mice and flies , including a single glycine to alanine substitution in mice that failed to complement the original nmf9 allele . Among animal and non-animal Holozoan homologues some strongly predicted functional motifs appeared to be modular across phyla , including two small GTPase-binding motifs , CRIB and RA . Among animals , the Nmf9 homology group appears to have been lost only in urochordates , based on four genomes available in that taxon . The N-terminal CRIB domain was found only in choanoflagellate and filasterean sequences and was either gained in these lineages or lost in the lineage leading to animals . The RA domain appears to have been lost in a derived paralog during or soon after a gene duplication event at the basal vertebrate lineage and the ancestral paralog was lost in the basal mammalian lineage . Nmf9 showed a complex pattern of expression in the vestibular system , olfactory system , and regions of the brain implicated in satiety and metabolism , innate anxiety , fear learning , and circadian rhythm . We tested the functional integrity of each of these systems with a battery of standardized behavioral assays . Deficits in vestibular , circadian and fear learning measures were clear despite significant sex differences . Abnormalities in measures of anxiety and appetitive behavior were nominally significant , but confounded by hyperactivity in the mutant mice–particularly after sustained handling , which may suggest either an anxiety-related aspect or perhaps feedback from disturbed sensory input from the vestibular system . Although mutants and non-mutant littermates were easily distinguished by overall behavior , mutant phenotypes for any single formalized test were not fully penetrant and tended toward higher variance than control littermate values , which suggests that Nmf9 may be important to the robustness of these pathways , but not essential to their basic function . On this interpretation , loss of Nmf9 activity in AMY , LS , and the olfactory system may not be disruptive enough to these circuits for behavioral phenotypes to be detected on a standard inbred background , but might have consequences that would be subject to selection under the more competitive fitness constraints of wild populations . Although we observed striking Nmf9 expression in cortical ventricular zones , the adult cortex appeared normal on gross inspection , with no obvious loss of cells in any of the Nmf9-enriched sites . Females were more severely affected than males by the nmf9 mutation in several tests ( Figs 1 and 4 ) . While both sexes express Nmf9 RNA , we have not explicitly tested for quantitative differences in specific pattern elements . Sex-dependence is known for several mouse behaviors both at baseline and in response to perturbations [29] . Interestingly , mutations of Ntrk2 ( TrkB ) also induce vestibular defects that are substantially more pronounced in female mice [30] , although any shared mechanism remains to be explored . Our analysis of CG45058 in Drosophila resolves a conflict in the current literature and extends both sets of previous findings . While our work was nearing completion , two other groups reported putative null mutations in CG45058 , but with very different primary outcomes . In a P-element screen for sleep phenotypes , Mark Wu and colleagues reported viable P-element and imprecise excision alleles , which they termed wide-awake ( wake ) based on increased latency to sleep and decreased total sleep by mutant animals [19] . In contrast , through an RNAi screen for regulators of asymmetric cell division , Juergen Knoblich and colleagues reported pupal-lethal gene deletion alleles of CG45058 , which they termed banderuola ( bnd ) based on the cytological appearance of dividing sensory organ precursor cells [20] . A major difference between these studies is the nature of the alleles examined . CG45058 has several annotated transcripts , which primarily differ in transcriptional start site and utilization of 5’ exons that are not well-conserved outside of Diptera . The wake alleles occur in the variant 5’ region of the gene and appear likely to affect only those transcripts , but not several others that include all of the conserved protein coding sequences . Alternatively , the bnd gene deletion allele might have removed regulatory sequences or barriers that influence expression of neighboring genes resulting in a synthetic phenotype . Our results resolve this conflict by creating discrete loss-of-function alleles induced at three distinct , functionally important sites in the predicted protein , encoded by separate exons that are each included in all known CG45058 transcripts . These alleles strikingly reduced adult viability and locomotor function of surviving adults , demonstrating that this is the null phenotype , in support of Mauri et al . Heterozygotes for these null alleles had reduced waking activity , increased latencies to first sleep bout , and effects on total sleep , confirming and extending the behavioral results of Liu et al . for wake alleles . Our data show sexually dimorphic effects on total sleep in null heterozygotes , in contrast to decreased sleep found in both sexes of wake mutants . This allelic difference suggests either distinct cellular functions of CG45058 isoforms or , perhaps more likely , isoform-specific expression in cell types that impinge on sleep regulation . While the Nmf9 orthology group shows extraordinary conservation–including highly constrained sequence domains that had no prior functional or motif annotation–the reorganization of putative GTPase-binding domains across major phylogenetic boundaries , loss of the homology group in urochordates , and duplication and rescission events in the vertebrate lineage all suggest a degree of adaptive plasticity that may be reflected in both shared and lineage-specific requirements . Indeed , mutations in the Nmf9 homologues of both mouse and fly resulted in phenotypes related to daily cycles of activity , as well as locomotor abnormalities , but with some clear differences . While mouse nmf9 null alleles showed hyperactivity and tremor that increased over the course of six months , fly mutations at distinct conserved domains ( ANK , FN3 , and conserved domain 2 ) produced severe locomotor retardation and early death , within one week of eclosion . These studies lay a foundation for understanding both common functions of Nmf9 homologs and lineage-restricted activities that might relate to derived loss of the RA domain or other lineage-specific features . How differences in domain architecture and expression patterns reshape the functional networks in which Nmf9/CG45058 proteins act will surely be of interest in the evolutionary development of the nervous system .
Coisogenic C57BL/6J–nmf9 mice were obtained from the Neuroscience Mutagenesis Facility ( NMF ) and AKR/J , BALB/cJ , C3H/HeJ and DBA/2J mapping partners from production colonies at the Jackson Laboratory . Conventional exclusion mapping was performed as described [11 , 31] . PCR primers for new genetic markers are given in S1 Table . New markers were also typed on a smaller C57BL/6J–nur12 x BALB/cJ cross obtained from NMF . C57BL/6J–nmf9 mutant line was maintained on C57BL/6J and genotyped by sequencing . Northern blots were performed by standard methods [32] , essentially as described [33] . RNA from whole brains was extracted with Trizol reagent ( Life Technologies ) . Poly ( A ) + RNA was isolated on Oligo ( dT ) cellulose type 7 ( Amersham Biosciences ) . Concentrations and integrity were verified by spectrophotometry and gel electrophoresis . 8 . 5 μg poly ( A ) + RNA per sample was electrophoresed through a denaturing formaldehyde/agarose gel and transferred to Hybond-N+ membrane ( Thermo Fisher ) . Size standard was removed prior to transfer and imaged by ethidium bromide fluorescence . Probes were synthesized from two cloned fragments of Nmf9 by random priming in the presence of 32P-dCTP and hybridized overnight . Blots were exposed to a phosphor screen for 5 days before quantitative imaging with a Storm 860 instrument ( Molecular Dynamics ) . Quantitative RT-PCR was performed using SybrGreen fluorescence quantification on a BioRad CFX96 instrument . Expression was relative to Gapdh and RP49 in Drosophila , and to GAPDH and SDHA in zebrafish . Mouse tissues were processed by a standard method [34] as previously modified [8] . Briefly , embryos were fixed in formalin , adult mouse brain in 4% PFA . Samples were cryoprotected in 30% sucrose and sectioned at 20 μm . Slides were treated 15’ with boiling sodium citrate , followed by acetic anhydride in triethanolamine prior to hybridization with dioxigenin-labeled RNA probes at 65°C overnight . After hybridization and washing , sections were blocked with 5% normal donkey serum ( Jackson ImmunoResearch ) in PBSTX ( 0 . 2% Triton-X100 in PBS ) for 1 hour and incubated with anti-digoxigenin-AP Fab fragments ( Roche ) at 1:2000 overnight at 4°C . Whole mount in situ in zebrafish was performed as described [35] . Samples were fixed in 4% PFA . Proteinase K was used for antigen retrieval at 10 μg/mL for 1 hour at 37°C . Hybridization was at 65°C for 48 hours . Samples were blocked with 2% normal donkey serum , 2mg/mL BSA ( NEB ) in PBTx and incubated with 1:5000 anti-dig-AP Fab fragments overnight at 4°C . Hybridization to whole mount fly embryos was performed as described [36] . Briefly , embryos were fixed in methanol and treated with Proteinase K at 37C for 7 minutes . Embryos were hybridized at 55°C overnight . Samples were blocked with 1:10 Western Blocking Reagent ( Roche ) in PBTwx ( 0 . 1% Tween 20 and 0 . 1% Triton-X in PBS ) and incubated with 1:1000 anti-dig-AP Fab fragments for 2 hours at 4°C . All probes were prepared by in vitro transcription from linearized plasmid templates and diluted in hybridization buffer prior to use . Sea anemone in situ hybridization was performed as described [37] . Probe templates for all species were generated by RT-PCR and cloned into appropriate vectors for in vitro transcription . All behavioral tests were performed on mice between 2 to 6 months of age , with wild-type littermates as control to the mutant animals , by investigators blinded to the animals’ genotypes . Protein sequences homologous to mouse Nmf9 were retrieved through NCBI tblastn , UCSC Genome Browser , Ensembl , and JGI web sites and made use of GNOMON and ENSEMBL transcript predictions . Identified fragments were used to search the full length sequences in surrounding genomic sequences where available ( see S4 Table ) . For homologs inferred purely through genomic sequences , a combination of BLAT [41] , BLAST [42] and GENSCAN [43] programs were used to predict open reading frames . All database searches were performed before February 2013 . For the full-length protein analysis , the best-annotated sequence of each evolutionary branch was used . For the high resolution analysis sequences with more than 50% gaps in that domain were excluded . Predicted cDNA sequences were translated using ExPASy Translate [44] and aligned using MUSCLE [45 , 46] and the result of the amino acid alignment used to manually correct the nucleic acid alignment . Conservation rates were calculated with Datamonkey [47] using codon data type and universal genetic code with neighbor joining tree . Motif search was performed with Motif Scan [48] , Scansite 2 . 0 Scan Motif [49] , Scan Prosite ( ExPAsy ) , InterProScan 4 [50] , Pfam [51] , and SMART 6 [52] . Protein modification scan was performed using The Sulfinator ( ExPASy ) for sulfination and YinOYang [53] for glycosylation and phosphorylation . Genome-edited mice were generated essentially as described [17 , 28] . Briefly , in-vitro synthesized Cas9 mRNA , sgRNA , and ssDNA homology-directed repair oligos were co-injected as a cocktail into C57BL/6 one-cell embryos at the Moores UCSD Cancer Center Transgenic Mouse Shared Resource . Oligonucleotide sequences are listed in S6 Table . sgRNA templates and ssDNA repair oligonucleotides were synthesized as Ultramers by IDT . All procedures were approved by the UCSD IACUC . RNA and DNA reagents for fly injection were prepared as above . Conserved Domain 2 mutants were generated by co-injection of in-vitro synthesized Cas9 , sgRNA , and ssDNA repair oligo into w1118 embryos . ANK and FN3 mutants were generated by co-injection of sgRNA and repair oligonucleotides into Cas9-expressing embryos ( PBac<y[+mDint2] = vas-Cas9>VK00027 , Bloomington Stock Number 51324 ) as described by [54] . All fly embryo injections were performed by Rainbow Transgenic Flies , Inc . ( Camarillo , CA ) . Oligonucleotide sequences are listed in S6 Table . Mutations were verified by Sanger sequencing of PCR products encompassing targeted genomic loci . Transposon mutant and deficiency lines for CG45058 were obtained from the Bloomington Stock Center according to the following stock numbers: wakeEY02219 [15858] , wakeKG02188 [14082] , wakeKG08407 [15129] , wakeMI02905 [37162] , Df ( 3R ) Exel6273 [7740] , Df ( 3R ) ED6085 [150049] , Df ( 3R ) ED6091 [9092] , Df ( 3R ) BSC527 [25055] , Df ( 3R ) Exel6192 [7671] , Df ( 3R ) BSC618 [25693] , and Df ( 3R ) ED6090 [150614] . Flies were raised at room temperature ( ~21°C ) on standard cornmeal/molasses . Within 1–2 days of eclosion , flies were assayed for wing orientation and the ability to maintain a standing position . For sleep assays , 1–5 day old flies were loaded into 65x5 mm glass tubes containing 5% sucrose and 2% agarose and entrained to a 12hr:12hr light:dark ( LD ) cycle for 2 days before recording sleep/wake patterns using the Drosophila Activity Monitoring System ( Trikinetics , Waltham , MA ) . Sleep was defined as 5 minutes of inactivity and , along with waking activity , was measured at 25°C using custom software as previously described [55] . Mice were euthanized by CO2 inhalation or by perfusion or organ removal under deep anesthesia with tribromoethanol ( avertin ) . Fish embryos were euthanized by fixation or snap-frozen in E3 media . All vertebrate animal procedures were approved by the University of California San Diego Institutional Animal Care and Use Committee ( IACUC ) . The University of California San Diego is AAALAC accredited , AAALAC institutional number 000503 .
|
Genome sequencing projects have identified large numbers of genes that encode proteins of unknown function . Many of these genes show strong evolutionary conservation , predicting important and well-conserved functions . A fraction of these show strong conservation of core domains but dynamic changes in other domains , predicting both conserved and lineage-dependent functions . Here we identify neurological functions associated with one such gene identified by a forward genetic screen in mice . We use recently developed genome editing tools both to confirm the mouse studies and to test comparative functions in a model insect , the fruit fly Drosophila melanogaster . Each of these species has a single homolog of this gene family , but differ by inclusion of a ras-association ( RA ) domain present in most invertebrate species but missing in mammals . Null mutations in both mice and flies produce neurological phenotypes , but while the mouse phenotypes are comparatively mild ( vestibular deficits , mild tremor , hyperactivity , mild circadian phenotypes and abnormal fear learning–but normal viability and breeding ) , null flies rarely survive to adulthood and surviving flies have severe locomotor deficits . Interestingly , heterozygous flies have significant sleep-related phenotypes . Together , our results provide a detailed first look at comparative function for a gene lineage with an unusual evolutionary history .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Nmf9 Encodes a Highly Conserved Protein Important to Neurological Function in Mice and Flies
|
Pathogen genome sequencing can reveal details of transmission histories and is a powerful tool in the fight against infectious disease . In particular , within-host pathogen genomic variants identified through heterozygous nucleotide base calls are a potential source of information to identify linked cases and infer direction and time of transmission . However , using such data effectively to model disease transmission presents a number of challenges , including differentiating genuine variants from those observed due to sequencing error , as well as the specification of a realistic model for within-host pathogen population dynamics . Here we propose a new Bayesian approach to transmission inference , BadTrIP ( BAyesian epiDemiological TRansmission Inference from Polymorphisms ) , that explicitly models evolution of pathogen populations in an outbreak , transmission ( including transmission bottlenecks ) , and sequencing error . BadTrIP enables the inference of host-to-host transmission from pathogen sequencing data and epidemiological data . By assuming that genomic variants are unlinked , our method does not require the computationally intensive and unreliable reconstruction of individual haplotypes . Using simulations we show that BadTrIP is robust in most scenarios and can accurately infer transmission events by efficiently combining information from genetic and epidemiological sources; thanks to its realistic model of pathogen evolution and the inclusion of epidemiological data , BadTrIP is also more accurate than existing approaches . BadTrIP is distributed as an open source package ( https://bitbucket . org/nicofmay/badtrip ) for the phylogenetic software BEAST2 . We apply our method to reconstruct transmission history at the early stages of the 2014 Ebola outbreak , showcasing the power of within-host genomic variants to reconstruct transmission events .
Understanding transmission is important for devising effective policies and measures that limit the spread of infectious diseases . In recent years , affordable whole genome sequencing has provided unprecedented detail on the relatedness of pathogen samples [1–4] . Consequently , accurately inferring transmission between hosts is becoming more feasible . However , this requires robust statistical approaches that make use of the full extent of genetic and epidemiological data available . Here , we present a new approach that makes use of within-host genetic variation and epidemiological data to infer transmission . A number of approaches have been developed that reconstruct transmission from genetic data . The number of substitutions between samples from different hosts can be used to rule out transmission [5–7] , or the phylogenetic tree of the pathogen samples can be used as a proxy for the transmission history [8 , 9] . However , while the phylogenetic signal can be very informative of transmission , it can also be misleading [10 , 11] , due to within-host variation that can generate discrepancies between the phylogenetic and epidemiological relatedness of hosts , and can bias estimates of infection times [12 , 13] . In recent years a number of methods have been proposed explicitly modelling both the transmission process and within-host pathogen genetic evolution to infer transmission events [11 , 13–28] . Some of these methods use epidemiological data and genetic sequences from pathogen samples , and ignore within-host evolution and other causes of phylogenetic discordance with transmission history [14–19 , 21–23] . Methods that explicitly model pathogen evolution within hosts and within an outbreak [13 , 20 , 24 , 25 , 27] generally assume , among other things , that samples provide individual and reliable pathogen haplotypes . This is often true for bacteria that are sampled and cultured before being sequenced , but it is mostly false for viruses and bacteria that are sequenced directly from samples without culturing . In fact , in these cases the sequencing process delivers reads coming from the different pathogen haplotypes that constitute the within-host pathogen population , and it is often very hard ( if not impossible ) to reconstruct complete haplotypes from these reads . In such cases , within-sample genetic variation is often neglected , and a single haplotype ( which we call the consensus sequence of the sample ) is built . While this procedure might lead to errors ( and maybe biases ) , it also certainly discards a very informative part of the available genetic data , because within-sample genetic variants can be very informative of epidemiological distance , direction of transmission , time from infection and transmission bottleneck intensity ( see [29–32] and Fig 1 ) . Furthermore , it is generally assumed that the pathogen does not recombine , so that a single phylogeny describes the evolutionary history of the whole genome , but this assumption does not fit highly recombinant pathogens such as HIV [33] . For these reasons , a few approaches have recently been proposed that use within-host genetic variants to reconstruct transmission [30 , 32] . Here , we propose a new Bayesian approach called BadTrIP ( BAyesian epiDemiological TRansmission Inference from Polymorphisms ) that not only uses within-sample genetic variants ( from possibly multiple samples per host ) to reconstruct transmission ( including directionality and time of infection ) , but also combines this information with epidemiological data and an explicit model of within-host pathogen population evolution and transmission . We use the phylogenetic models with polymorphisms PoMo [36–38] to model population evolution along branches of the transmission tree; thanks to this , our transmission tree and phylogenetic tree are the same entity , and within-host evolution and recombination ( resulting from a single primary infection , not multiple infections ) do not create discrepancies that make statistical inference hard and computationally demanding [24 , 25 , 27] . We also explicitly model transmission bottlenecks , with one parameter defining the intensity of the bottleneck , and therefore the number of pathogen particles that establish a new population at transmission . Another feature of our approach is that we assume that different genomic positions are unlinked , an assumption also made by other methods using within-host variants [30 , 32]; most coalescent-based methods assume instead no recombination at all . Because of our assumption of no linkage , we expect our approach to work well when recombination is strong enough to break linkage between genetic variants in the same host , or when the evolutionary rate is slow so that very few new mutations originate with each new transmission . BadTrIP is implemented as an open-source package for the Bayesian phylogenetic software BEAST2 [39] , and as such , it can be freely installed , used , and modified . We compare the performance of BadTrIP , of the shared variants-based clustering ( SVC ) method of [30] , and of the coalescent-based method SCOTTI [13] on simulated data and on a real dataset from the early stages of the 2014 Ebola outbreak [40] . These applications show that BadTrIP has high accuracy to reconstruct transmission thanks to its explicit model of population evolution , the use of within-host genetic variants , and the inclusion of epidemiological data , and can provide important information to understand and limit the spread of infectious disease . In the rest of the manuscript , we refer to a “host” as any entity that can contain and transmit a pathogen . Typically a host is a human within a community or nosocomial outbreak , or patients , but the concept of host can also be generalised for example to farms within a livestock outbreak . We will refer to the collection of all pathogens of the type under consideration within an individual host at a certain time as a “pathogen population” ( for example all Ebola virions within an infected host , excluding non-Ebola pathogens and Ebola virions from other hosts ) . We will call a “pathogen unit” a single pathogen individual within a population , for example an individual bacterial cell or an individual virion . We call a pathogen population “polymorphic” at a particular genome position if pathogen units with different nucleotides at that position are present in the population; in this case , we also call the considered genome position a “genetic variant” .
Methods to reconstruct transmission that account for within-host evolution usually have to deal with the complex task of modelling and inferring the discrepancies between the transmission tree and the pathogen phylogenetic trees [13 , 20 , 24 , 25 , 27] . We avoid this complication by adopting and adapting a substitution model , PoMo [36–38] , that describes population evolution along the branches of a species ( or population ) tree . In this model , a virtual population , similar to a Moran model [41] without selection and with fixed population size , evolves by accumulating random changes in nucleotide frequencies ( genetic drift , eventually resulting in the fixation of polymorphic sites ) , and new mutations resulting in new polymorphic sites . Different genome positions are modelled as completely unlinked . The adoption of such a population genetic model within a transmission tree structure means that the phylogenetic tree and the transmission tree are now the same entity , and that each point of the tree represents the state of the pathogen population at a certain time within a host ( Fig 2 ) . Each bifurcation in the tree represents a transmission event , where the pathogen population splits in two groups: one remaining in the current host , and a small sub-population colonising a new host . We use a population bottleneck at time of transmission for the colonising branch to better model the transmission process . Our method uses two sources of information: epidemiological and genetic data . Epidemiological data is in the form of dates: the times when genetic samples are collected ( it is possible to give any number of samples ≥ 0 for any host , even no sample at all ) and a time interval for each host describing when it can contribute to the outbreak . Each host can only be infected , be sampled , and can infect other hosts within its time interval [13] . Genetic data from each sample is in the form of nucleotide counts: for each position of the genome , for a certain sample , the model expects the number of times each of the four nucleotides is observed in the reads ( for example: 59 As , 0 Cs , 12 Gs , 1 Ts ) . We assume that reads are sampled with replacement from the pathogen population according to the ( hidden ) true nucleotide frequencies , and we model the sequencing error . This in particular means that sites without any sequencing coverage , or with very low coverage , are also allowed , and that differently from similar approaches ( i . e . [30 , 32] ) we don’t require the specification of a minimum genetic variant frequency threshold . While in our model we make the strong assumption that sites are completely unlinked , we test the performance of our approach with simulations in which we explicitly model within-host recombination events and we assume that a limited number of individuals in the pathogen population is sequenced . We even simulate scenarios in the total absence of recombination ( complete linkage ) to measure the robustness of our method . We simulate a broad range of scenarios: different transmission bottleneck severities ( weak vs . strong ) , different amounts of genetic information , different recombination and mutation rates , different sequencing coverage levels , different sequencing error rates , and different virtual population sizes . We give further details on the model used and the simulations in the Materials and Methods section . To test the accuracy of our new method BadTrIP in inferring transmission events , and to compare it to previous methods [13 , 30] , we simulated pathogen evolution within outbreaks and sample sequencing , and we used different methods to reconstruct the transmission history from sequencing and epidemiological data . To simulate pathogen evolution , first we simulated an outbreak using SEEDY [42] ( we used a fixed population of 15 hosts , one initial case , and a basic reproduction number of 1 . 43 , see Materials and methods ) ; then , we translated the transmission history into a population history , and simulated within-population pathogen coalescent , recombination and mutation with fastsimcoal2 [43] . Throughout the simulations each host in the outbreak is sampled exactly once . We measure the accuracy of a method as the frequency with which the correct transmission source of each host is inferred to be the most likely a posteriori . We also give a measure of how well calibrated [44] methods are by counting how often the correct source is in the 95% posterior credible set , defined as the minimum set of sources with cumulative probability ≥ 95% such that all sources in the set have higher posterior probability than all sources outside of it . BadTrIP shows elevated accuracy in detecting the correct source of transmission ( between 50% and 90% ) and calibration ( between 80% and 100% ) , in particular compared to the SVC approach ( accuracy between 20% and 45% and calibration between 45% and 95% ) , see Fig 3 . This shows that the use of epidemiological data and an explicit model of evolution can help to reconstruct transmission . Using alternative statistics for accuracy and calibration leads to similar patterns ( Fig F in S1 Text ) . BadTrIP also shows more accuracy than the coalescent approach SCOTTI ( accuracy between 25% and 70% ) . The latter method appears very conservative in this application ( calibration between 95% and 100% ) . SCOTTI uses the same epidemiological information as BadTrIP , but a different format of genetic data and a different model of genetic evolution . In fact , like most coalescent-based approaches , SCOTTI requires a full haplotype to be given for each sample; in these simulations we used the consensus sequence of a sample as its haplotype for SCOTTI , discarding within-sample genetic variation . The fact that SCOTTI has strictly less genetic information available than BadTrIP can explain why generally it has less accuracy and is more conservative , however it is not the only factor at play , another being recombination . For example in the scenario with 1x coverage BadTrIP seems to have higher accuracy than SCOTTI , despite the two methods having the same information available: this can be explained with the fact the SCOTTI wrongly assumes that there is no recombination . Similarly , the simulations suggest that the accuracy gap between SCOTTI and BadTrIP reduces with no recombination , and increase at high recombination: this fits well with the fact BadTrIP assumes no linkage between genomic positions , while SCOTTI assumes complete linkage ( no recombination ) . While these results are very suggestive and fit with our expectations , we also have to warn that for each individual scenario we have 10–20 simulated outbreaks , so while the general patterns are clear , the specific patterns of each scenario are subject to considerable uncertainty . Comparing the base scenario with the one with almost no mutation , we see that BadTrIP accuracy drops from about 80% to about 50%; this drop hints to the contribution given by genetic data to the inference of transmission . Also , since in the latter scenario almost no genetic information is available , it also suggests what is the contribution of epidemiological information alone . Calibration of BadTrIP seems to increase as mutation rate decreases , one probable contributing factor being that as mutation rate decreases the effect of genetic linkage on the pathogen evolutionary dynamics decreases ( neither method models genetic linkage ) , or possibly as a result of the increased uncertainty on the evolutionary process . The complete absence of recombination seems to negatively affect calibration in BadTrIP , but the difference is not dramatic ( from about 90% to about 80% ) suggesting that even in the worst case scenario of complete absence of recombination BadTrIP can still provide meaningful inference and posterior distributions . Accuracy of all methods seems to decrease with decreasing mutation rate , as is expected because of the reduced genetic information . However , very high mutation rates ( to the point that about half the genome , of length 5kb , is polymorphic within the outbreak ) do not seem to improve inference , probably because of saturation . Accuracy of BadTrIP seems higher ( around 10% difference ) in the presence of a strong bottleneck ( small inoculum ) than a weak bottleneck ( large inoculum ) , while calibration seems almost unaffected; this probably happens because , with strong bottlenecks , polymorphisms are unlikely shared between hosts , and so polymorphisms leading to substitutions ( see Fig 1B ) become more informative for identifying infectors . An increase in coverage ( from 40x to 100x ) does not seem to bring improvement in accuracy or calibration to BadTrIP; on the other hand , when a single uniform colony is sequenced ( which is equivalent to reducing coverage to 1x , and therefore removing information on within-host genetic variation ) , accuracy seems moderately reduced ( ≈ 10% ) but not calibration . Introducing sequencing error ( 0 . 2% of mis-called bases , slightly more than what typical for high-throughput DNA sequencing [45] ) accompanied by reduced coverage ( 20x ) and genome length ( 1kb ) still seems to result in elevated accuracy ( 72 . 5% ) and calibration ( 97 . 5% ) . Increasing the PoMo virtual population size ( from 15 to 25 , while the actual simulated population size remains 1000 ) showed negligible effects on the inference . BadTrIP also infers the time of infection . Calibration seems to increase with recombination , and to decrease with mutation ( Fig 4 ) , probably again an effect of our assumption of no linkage . Also , very high mutation rates seem to reduce the error in time inference , as do high coverage and virtual population size . The running time of BadTrIP is affected by the number of genetic variants present in the alignment and by the number of hosts present in the outbreak ( Fig A in S1 Text ) . The number of variants affect the number of likelihoods that need to be calculated at each MCMC step , while the number of hosts affects the size of the transmission/population tree ( so both the computational and statistical complexity of BadTrIP ) . However , the time required to complete an analysis is not always a linear function of these two quantities: at low mutation rates BadTrIP requires similar times for different outbreak sizes . The reason is probably that with less data there is more uncertainty ( in particular in the posterior distribution of the mutation rate ) , and so it takes longer to explore the the parameter space effectively . Overall , it takes a few hours to completely investigate an outbreak of moderate size ( one or two dozen hosts ) with BadTrIP . To demonstrate the applicability of BadTrIP and the advantage of using a model that combines epidemiological and within-sample genetic variation data , we use BadTrIP to infer transmission within the early cases of the 2014 Ebola outbreak in Sierra Leone . We use data published by Gire and colleagues [40] and previously analysed with the SVC method by Worby and colleagues [30] . One of the factors that make this dataset important to this study is the presence of within-host variants shared by multiple hosts , with one genetic variant that was even shared by eleven hosts [40] . While classical approaches based on consensus sequences would struggle to accommodate such data , in particular due to their assumption of strong transmission bottleneck that would not allow the transmission of variants , BadTrIP can accommodate such features , and such shared genomic variants are expected to increase the resolution of our transmission history inference . We investigate a collection of 62 samples with associated time and location of sampling . As observed by previous researchers , the number of substitutions ( and more generally the number of SNPs ) within this partial outbreak is very limited , and as such we expect to see a lot of uncertainty in the inference [30]; furthermore , all the samples were collected over a time interval of two months , and we assume transmission from a host to be possible from three weeks prior to three weeks following the sample collection , so the epidemiological data are also not very informative . Indeed , we see that most of the cases are inferred by BadTrIP to have a flat distribution of possible infectors , with highest per-infectee values generally under 30% posterior probability ( Fig 5 ) . However , we also see that BadTrIP identifies some pairs of infector-infectee with very high posterior probabilities ( Fig B in S1 Text ) . These pairs not only generally fit with the geographical epidemiological data , with most transmission with posterior probability > 50% happening within chiefdoms ( with two exceptions discussed later ) , but also with the SVC inference [30] . Of these , transmission from EM119 to G3770 was inferred by Worby and colleagues [30] using consensus sequence genetic distance , while transmission from EM096 to G3679 , from G3826 to G3827 , from G3820 to G3838 , from EM110 to G3809 , and from G3729 to G3795 was inferred with the help of shared within-host genetic variants . All highly likely transmission pairs in [30] are also inferred by BadTrIP , but there are some highly likely transmission events inferred by BadTrIP that were not detected by SVC . For example , transmission from G3834 to G3817 is inferred by BadTrIP and is supported by a 3% frequency variant within G3834 that becomes fixed in G3817; however , such a variant fixation , attributable to the transmission dynamics described in Fig 1B , is not informative in the SVC method [30] and was further ignored due to the imposition of a 5% variant frequency threshold that we could avoid thanks to our explicit model of sequence evolution and sequencing error . Other cases similar to the latter are the inferred transmissions from EM110 to G3856 , from EM110 to G3822 , and from EM111 to G3724 . Cross-chiefdom transmissions inferred by BadTrIP with elevated posterior distributions are from EM110 in the chiefdom of Jawie , district of Kailahun , to G3856 in the chiefdom of Nongowa , district of Kenema; and from G3834 in the chiefdom of Kpeje to G3817 in the chiefdom of Jawie , both in the district of Kailahun . Neither of them had a high probability in [30] , but they are both supported by low-frequency variants becoming fixed in the recipient . Our inference of the sequencing error rate ϵ is extremely low ( 2 ⋅ 10−7 < ϵ < 7 ⋅ 10−7 ) consistent with the thorough filtering steps adopted by Gire and colleagues [40] prior to within-host variant calling .
Methods to infer transmission histories within outbreaks are important to determine the causes of transmission , and to limit and prevent future outbreaks . Genomic pathogen data from an outbreak reveals in detail the genetic relatedness of pathogens from different cases . Most methods to infer transmission from pathogen genetic data require full haplotypes , but it is often not possible to reconstruct haplotypes due to pathogen recombination and short or inaccurate reads . This leads in many cases to discarding information regarding within-sample genetic diversity , and only use a sample consensus . In recent years two methods have been proposed to infer transmission from genetic distances between samples and shared within-sample variants [30 , 32] . Here we presented BadTrIP , a Bayesian approach to transmission inference that makes use of within-sample variants and allows inference of transmission direction and time . Compared to other similar methods [30 , 32] , our approach has the advantage of implementing an explicit model of pathogen population evolution , transmission and sequencing , of allowing the inclusion of epidemiological data ( sampling times and host exposure times ) , of not requiring minimum thresholds for within-host variant frequencies , of accounting for sequencing errors , and of being implemented as part of an open source phylogenetic package ( BEAST2 [39] ) . These aspects can result in more applicability , but also , as we have seen in our simulations , in greater accuracy . Compared to existing methods based on the coalescent ( e . g . [13 , 24 , 25 , 28] ) BadTrIP does not require the reconstruction of haplotypes and consensus sequences , but instead uses data of within-sample genetic variability , therefore having access to important information that can reveal otherwise cryptic transmission events . Using simulations , we show that our approach achieves higher accuracy and calibration than SVC [30] , has more accuracy than the coalescent-based method SCOTTI [13] used on consensus sequences of pathogen population genetic data , and can reliably identify likely transmission histories . The comparison between BadTrIP and SCOTTI is particularly interesting , because it shows us that reducing the genetic data of a within-host pathogen population to a single consensus sequence leads not only to the loss of within-host genetic diversity information , but can also lead to errors by ignoring recombination and weak transmission bottlenecks . Also , using a dataset of the early 2014 Ebola outbreak in Sierra Leone , and making use of information of within-sample variation and an explicit population evolution model , BadTrIP could infer previously unidentified likely transmission events , including transmissions between different geographic locations . BadTrIP infers transmission from both epidemiological time data and pathogen genetic data . In most circumstances , both types of data are extremely useful , and we see in our simulations that removing genetic data information leads to a loss of ≈ 30% accuracy , and similarly the epidemiological data is expected to provide ≈ 40% accuracy ( the baseline accuracy without any data is expected to be around 10% in our simulations ) . However , the contribution of the two types of data will be extremely dependent on the particular context at hand . As we showed in our simulations , BadTrIP can account for uninformative genetic data , with which it still provides meaningful inference . Our approach can however also account for uninformative epidemiological data: in the absence of exact dates , the user can specify arbitrarily large exposure intervals , allowing hosts to be infected any time by any host; as with the lack of genetic data , in this case we would also expect a significant drop in the accuracy of our method . Despite these results , BadTrIP also has limitations , for example its model of genetic linkage . By assuming that all sites are unlinked , our model could be poorly calibrated in cases where there is no within-host recombination but high within-host mutation , causing strong correlations between inherited variants that are not expected in our model . However , we show in our simulations that our method is robust in a large variety of scenarios , including in the absence of recombination and with reads coming from few pathogen units . Another limitation is that our approach is generally not fast enough to deal with very large datasets , and , at the current stage , application is recommended to outbreaks with fewer than 100 cases . Also , BadTrIP is only applicable to the case where all hosts in the outbreak have been observed . In fact , our current implementation does not allow to infer the number of non-observed hosts ( hosts for which there is no sample or epidemiological data ) . However , BadTrIP does allow to model non-sampled hosts with epidemiological data , or a fixed number of non-observed hosts ( such hosts could be given uninformative epidemiological data , such as exposure intervals without ends ) . The assumption that all cases are observed or sampled is very common among transmission inference methods [11 , 14–20 , 23–26 , 28] , but it limits their applicability . Extending our method to infer the presence of possible non-sampled and non-observed intermediate hosts would be relatively straightforward and would increase the method’s applicability , but it would also lead to a significant increase in the statistical complexity and computational demand ( but see [13 , 27] ) . Another scenario that is not accounted for in our model is multiple infections of the same host ( one host being infected by multiple sources , or by the same source multiple times ) . This scenario can be relatively frequent in many viruses , for example HIV [46] , but it is very hard to model in our context as it would require the use of a population network ( see e . g . [47] ) instead of a population phylogeny , which would make likelihood calculation more computationally demanding . Another similarly looking and equally concerning problem is sample contamination . We recommend sequencing data to be tested for possible contaminations and multiple infections using methods such as PHYLOSCANNER [48] prior to being investigated with BadTRiP . In our Ebola dataset we found no obvious pattern of mixed infection or contamination ( like an excess of similar frequency SNPs in one sample ) . However , none of these approaches would detect multiple infections from closely related cases . BadTrIP uses a very simple model of sequencing error , only accounting for the two most common nucleotides at a given position and sample . This sequencing error model would probably have sub-optimal performance when sequencing error rate is high ( e . g . with Nanopore sequencing technologies ) and coverage is high or mutation rate is elevated . In these circumstances , a more realistic and computationally demanding model of sequencing error might be preferable . Similarly , our model of evolution only allows 2 alleles for one genome position in one host at one time . If mutation rates are so high that more than 2 alleles are frequently present simultaneously in the same host , time and position , then our model could have sub-optimal performance . However , our approach can still account for the more common scenario where a site has more than 2 alleles but not all in the same host: for example if at a certain position host 1 has a fixed A , host 2 has a polymorphism with A and C , and host 3 has a polymorphism with C and G . BadTrIP does not account for selective pressure , which could sometimes cause errors , for example by creating homoplasies due to the same mutation appearing multiple times in different hosts , or by the same polymorphism being maintained by balancing selection . However , our approach weighs information from both fixed substitutions and polymorphic variants , so the same mutation appearing in different genetic backgrounds will not be as nearly as misleading as for the SVC method ( which gives much more weight to shared variants than to genetic distances ) . We assume that within-host population sizes are constant after an initial expansion . Size fluctuations in all hosts are unlikely to cause problems , as the PoMo drift rate would in this case represent the average drift rate in hosts . On the other hand , if fluctuations only happen in certain hosts , so that different hosts have different average drift rates , it might have averse effects on the estimate of infection times . As our model is implemented in BEAST2 , it is possible to specify a broad range of models of genomic variation in substitution rates which could at least partly account for the effects of selection . An additional feature that could be added to BadTrIP is indel evolution . For example , by assuming an infinite sites mutation model , indel data could be reduced to 0–1 states , and a PoMo matrix with two alleles instead of 4 nucleotides could be used . This approach could be useful to complement SNP data , but would only work at relatively low indel rates . Finally , it is possible that errors in the bioinformatic processing of reads , for example mapping errors , cause the identification of the same spurious genetic variants in multiple hosts . We therefore encourage the investigations of genetic variants shared by many hosts to assess their biological plausibility . In the future we will work to solve some of the limitations of BadTrIP , in particular to reduce its computational demand and to model non-sampled non-observed hosts . In conclusion , we have presented a new method that addresses the urgent need for software to efficiently and accurately analyse genomic and epidemiological data , in particular taking advantage of within-sample genetic variants to identify transmission pairs and reconstruct direction and time of infection . BadTrIP can be used in a broad range of outbreaks , and will be important for devising effective strategies to fight the spread of infectious disease .
We model each host as a deme d ∈ D that can be colonised by a pathogen population , with total number of hosts-demes being nD . Each deme d is associated with an exposure interval limited by an introduction time id ∈ ( −∞ , +∞] and a removal time rd ∈ [−∞ , +∞ ) , with rd < id ( we consider time backward as typical in coalescent theory ) ; the host only contributes to the outbreak within this interval , which is determined by the epidemiological data . In the least informative scenario where no information on host d exposure is provided , it is assumed that d is exposed for the whole outbreak ( id = +∞ and rd = −∞ ) . We will denote as X the collection of exposure times . Each host-deme starts off as non-colonised and is colonised ( infected ) at some time td between id and the time that the first sample is collected from d ( if no sample is collected from d , then we require only td > rd ) . Also , unless d is the first host to be infected in the outbreak , d is infected by another host in the outbreak Id ≠ d , such that r I d < t d < t I d , that is , d is infected after Id is infected , but before Id reaches its removal time . If d is indeed the first case of the outbreak , then Id is assigned the ∅ ( we assume ∅ ∉ D ) . We assume for simplicity that transmission between any pair of hosts and at any time is equally likely , as long as it is consistent with the epidemiological data . A transmission event of host d at time td is inconsistent with the epidemiological data if td is outside the exposure interval of d or its infector Id , or if d is sampled , infects another host , before td . Given the epidemiological data , some infector-infectee pairs are a priori more likely than others , depending on the length of time that a transmission between them is allowed . Each host is also provided with a ( possibly empty ) set of samples , Sd . Each sample s consists of a sampling time ts and genetic data Gs . Each sample s in Sd has to be collected after d is infected ( ts < td ) and before d is removed ( ts > rd ) . Assuming that the genome is L bases long , then the genetic data Gs of every sample s has to be in the form of a list of L quadruples , with for example the quadruple for genome position i being Gsi = ( ai , ci , gi , ti ) , the four positive natural values being the numbers of A’s , C’s , G’s and T’s observed at position i in the sample . If there is no read mapping to position i in sample s , then its quadruple is simply Gsi = ( 0 , 0 , 0 , 0 ) . We denote the set of all sequencing data as G . All hosts share a common parameter B ( with real positive values ) describing the intensity of the transmission bottlenecks associated with transmission events . Generally , the value of B can be inferred jointly with other model parameters , however its interpretation in terms of the size of the transmission inoculum is not straightforward . T denotes the transmission-population tree consisting of all sampling times , all infection times and all infectors of each host , and μ denotes the pathogen evolution model ( described below ) . An example of tree T and of model parameters is given in Fig C in S1 Text . We aim to sample from the following joint posterior distribution with a Monte Carlo Markov Chain approach: P ( T , μ , B | G , X ) ∝ P ( G | T , μ , B ) P ( T | X ) P ( μ ) P ( B ) . ( 1 ) P ( μ ) and P ( B ) are the prior probabilities for respectively the substitution model and the bottleneck size , which can be chosen arbitrarily by the user . We ignore the prior for the transmission tree P ( T|X ) as in [13] . P ( G|T , μ , B ) is the likelihood of the sequences given the genealogy and substitution model , and is calculated as described below , using an adaptation of [36–38] to transmission trees . So once we calculate the likelihood P ( G|T , μ , B ) , we can use Eq 1 with an MCMC to infer a posterior distribution of infection times , infectors , bottleneck size and substitution model parameters . Here , we make use of a phylogenetic model for population evolution , PoMo [36–38] , to model mutation and drift in the within-host pathogen populations; also , we extend the model to include transmission bottlenecks and sequencing errors . Sequence evolution is usually modelled along phylogenetic trees , which can differ from the transmission tree [13] . However , PoMo describes evolution along species ( or population ) trees , and the population tree of a pathogen within an outbreak corresponds to the transmission tree T described in the previous section . If we consider the pathogen community within a host d as a population , we see that this population exists from time of infection td , when it originates from a split with the population of its infector Id . So , transmission events corresponds to timed splits in the population tree , similar to the bifurcations of a species tree . However , one difference is that the split is asymmetrical , as we assume that the pathogen population size is not affected at td in Id , but at the start of the branch leading to d it undergoes a bottleneck of intensity B . All events in the tree are timed in real time ( e . g . , days ) with some values fixed ( for samples ) and some values inferred in the MCMC ( infection times ) . We use a procedure very similar to the Felsenstein pruning algorithm [49] to calculate the likelihood of the genetic data over the tree . First of all , the substitution process along the branches of the transmission-population tree is not a simple DNA substitution process , but is similar to a 4-allelic Moran model [41] with mutation . We assume we have a continuous-time Markov process along each branch of the tree , where the state space is not made by the four nucleotides , as is typical , but by all 1- and 2-allelic states possible for a population of N units . Typical values of N that we use here are 15 or 25 , that is , we describe evolution of a large within-host pathogen population ( possibly with billions of units ) with a small virtual within-host population of N units . Such an approximation generally leads to reasonably good results as long as we rescale the mutation rates between the real and the virtual population [36–38] . N here is not estimated , but is fixed by the user . Lower values of N are expected to reduce the computational demand of the method , but can result in lower accuracy . The states of our Markov process always include the four fixed states , where only one nucleotide is present in the population . In addition , they also include six groups of polymorphic states , where two nucleotides are present in the virtual population at the same site at the same time . Each group corresponds to one of the six unordered pairs of nucleotides ( {A , C} , {A , G} , {A , T} , {C , G} , {C , T} , {G , T} ) and contains N − 1 states: if the two nucleotides present in the population are n1 and n2 , then such N − 1 states are the ones in which the population contains i times nucleotide n1 and N − i times nucleotide n2 , for 0 < i < N . So in total our state space is of size 4 + 6 ( N − 1 ) . Our substitution rate matrix is sparse , in that we only allow one unit in the virtual population to change at the time . So , from a fixed state with nucleotide n1 , a instantaneous move is only possible to one of the three states with N − 1 times nucleotide n1 and one time any other nucleotide n2 different from n1 . Such moves correspond to mutation events , and we represent their rates as μ n 1 , n 2 . Instead , if we are already in a polymorphic state with i times nucleotide n1 and N − i times nucleotide n2 , we only allow nucleotide counts to instantaneously change by one , so an instantaneous move is only possible to the state with i + 1 times nucleotide n1 and N − i − 1 times nucleotide n2 , or to state i − 1 times nucleotide n1 and N + 1 − i times nucleotide n2 ( one of these two latter states might be a fixed state ) . The instantaneous rate at which such changes happen is i ( N - i ) N 2 R which corresponds to the rate of genetic drift; here R scales the rate of drift in the virtual population in units of real time; the rate of drift in the virtual population also depends on N , and it represents the rate of drift in a real pathogen population , which in turn depends on the pathogen effective population size , the pathogen generation time , and the time unit . All other non-diagonal substitution rates are set to 0 . All these states and rates constitute the substitution process E . The rate matrix is further described in Fig D in S1 Text . Our model only allows 2 alleles to be present in one host at one time at one position . This can be unrealistic where mutation rates are extremely high , or selection favours several variants at the same site . The likelihood of T is calculated starting from the hosts in the outbreaks who don’t infect others ( the leaves of the transmission tree ) . For such leaves , the likelihood is first calculated from the latest sample ( if no sample is present , then the likelihood of such leaf at time of their transmission is 1 for every state ) . Given any state of our substitution process with nucleotides n1 and n2 with respectively abundances i and N − i in the virtual population ( here for generality i can also be 0 ) , given a sample and site at which the nucleotides with the highest coverage are x1 with coverage c1 , and x2 with coverage c2 ( we ignore the nucleotides with lower counts for numerical stability , and in case of a tie random nucleotides are selected from the tying ones ) , then the likelihood of this state at this sample and site is approximated as: P ( c1 , x1 , c2 , x2|i , n1 , N−i , n2 , ϵ ) == ( Ix1=n1 ( i ( 1−ϵ ) N+ ( N−i ) ϵ3N ) +Ix1=n2 ( ( N−i ) ( 1−ϵ ) N+iϵ3N ) +Ix1≠n1 , x1≠n2*ϵ3 ) c1·· ( Ix2=n1 ( i ( 1−ϵ ) N+ ( N−i ) ϵ3N ) +Ix2=n2 ( ( N−i ) ( 1−ϵ ) N+iϵ3N ) +Ix2≠n1 , x2≠n2*ϵ3 ) c2·· ( c1+c2c1 ) ( 2 ) Where ϵ is a parameter describing the sequencing error rate . Here , due to sequencing errors and to random sampling of reads from the pathogen population , the observed alleles c1 and c2 are allowed be different from the alleles n1 and n2 in the virtual population . We assume that each read has the same probability to represent any of the individuals in the virtual population , and that there is a probability ϵ that the considered position of the read is a sequencing error ( in which case any of the three wrong nucleotides is equally likely to be on the read ) . ( i ( 1 - ϵ ) N + ( N - i ) ϵ 3 N ) is the probability to see a n1 nucleotide: the first part is the probability that the read comes from an individual in the virtual population with nucleotide n1 at the given position and that no sequencing error happened; the right end part represents the probability that the virtual individual had a different nucleotide but there was a sequencing error . ϵ can be estimated with the other model parameters as we do with the real data and with the simulations including sequencing error . For all other simulations we set ϵ = 0 . This sequencing model assumes that there are at most 2 alleles in the reads data for one sample at one position . If more than 2 alleles are observed , then only the counts from the 2 most common alleles are retained . Along branches of T , the likelihood is updated using the matrix exponential of E . At bifurcations ( corresponding either to internal samples or transmission events ) the likelihood is also updated according to the classical pruning algorithm , but at transmission events an extra step is added . A new drift-only substitution matrix ED is defined by setting the mutation rates in E to 0 . Then , we describe a bottleneck as a branch of length B along which the population evolves under drift alone , that is , under ED . The length B does not count toward the branch lengths in real time , so that changing the intensity of the bottleneck does not affect the timing of the events in T . Under this model , a more intense bottleneck , corresponding to a small transmission inoculum , will be represented by a longer bottleneck branch , so a larger B . If we have a transmission event from Id to d at time td , we first calculate the likelihood within population Id up to right before time td ( likelihoods are updated backward in time ) , then within population d up to right before time td , then we update the likelihood within d using the bottleneck branch , and finally we multiply the two likelihoods in d and Id to obtain the likelihood in Id right after td ( again backward in time ) . This backward-in-time likelihood update process is terminated after the transmission event of the index case , and before its bottleneck we assume state equilibrium frequencies . We now describe an example of likelihood calculation in Fig E in S1 Text . We use typical BEAST2 scalar proposals for B , ϵ and E , which , given a constant s and a random uniform real number 0 < u < 1 , propose to scale the given parameter by a factor of s + ( 1/s − s ) u; the reciprocal of this factor is the Hastings ratio of the proposal . We also define below five new operators ( proposals ) for updating our transmission-population tree . We will now give a very informal intuition of why the above proposals make an irreducible MCMC . We will focus on the transmission history , and not on B , ϵ or E , but the extension is trivial . We will discuss intuitively how it is possible to move from any given tree T to a specific tree T ˜ ( the tree we use as a starting point of the MCMC ) . As proposals are reversible , this is sufficient to have irreducibility . T ˜ is defined as the tree where the host with the earliest introduction time is the index case; each non-index host d in T ˜ is infected by the host Id with the earliest introduction time i I d among those with an exposure overlap to d; in T ˜ infection time td of any host d is set to id ( for a more formal proof an infinitesimal interval after id might be considered ) . Starting from T , we first approach T ˜ by moving host h , the index case in T ˜ , up from its starting position in T by using repeatedly operator four . At each step , before applying operator four , we use operator one to move th up to make sure it is the first infectee of its infector , and that it is infected before the first sample of its infector is collected . Repeating these two steps long enough , h is guaranteed to become the root , at which point we can apply operator one to make sure its infection time is the same as in T ˜ . We then proceed to apply a similar strategy iteratively on all other hosts in order based on their introduction time ( from earlier to latest ) . We stop when we obtain T ˜ . To test the accuracy of our new method BadTrIP in inferring transmission events , and to compare it with the SVC method [30] , we simulated pathogen evolution within outbreaks and sample sequencing , and we used different methods to reconstruct the transmission history from sequencing and epidemiological data . To simulate pathogen evolution , first we simulated an outbreak using SEEDY [42] with a host population of 15 hosts and an infection rate of 0 . 1 per day , a recovery rate 0 . 07 per day , and conditionally accepting only outbreaks that achieve a minimum total of 10 infected cases . Given these parameters , SEEDY will start at time 0 with one infected individual in the community of 15 hosts . Each day every infected host has a 0 . 1 chance of infecting any other host , and a 0 . 07 chance of recovering ( recovered hosts are no more infectious or infectable ) . If the outbreak runs out of infected hosts before a total of 10 hosts are infected , the simulation is repeated . We then took the outbreak simulated by SEEDY and translated the transmission history into a population history , assuming a within-host pathogen population size of 1000 and using fastsimcoal2 [43] to simulate pathogen coalescent , recombination and mutation with scenario-dependent parameters . fastSimCoal2 is an approximate coalescent simulator implementing the sequential Markov coalescent model [50 , 51] with cross-over recombination . This model describes viral recombination more appropriately than bacterial recombination , for which a coalescent simulation software modeling gene conversion is preferable [52 , 53] . The use of a coalescent simulator with recombination is also the main difference with the simulations made by [30] , where within-host recombination was not allowed . Within each infection we assume that the population size is constantly 1000 individuals , but at the time of transmission we assume an instantaneous population bottleneck ( founding population size either 1 or 5 individuals depending on the scenario ) . At the time of a transmission ( simulated by SEEDY ) the whole infectee population is , backward in time , merged with the infector population . We observed that some times , in particular at high recombination rates , fastSimCoal can crash: if this happens we simply repeat the fastSimCoal2 simulation with a different seed . Throughout all simulations each host was sampled exactly once . We define a basic group of simulations ( called “base” ) , and nine variants , in each of which one or two aspects of the base group of simulations is modified . In “base” we simulated about 300–500 SNPs ( counting also variants present at very low frequency in just one host ) or 45 substitutions per outbreak ( which might be typical for HIV but high for many other pathogens ) , recombination rate 10 times higher than the mutation rate , complete bottlenecks ( no transmission of within-host genetic variants ) , homogeneous read coverage of 40x , no sequencing error , PoMo virtual population size of 15 , all equal mutation rates , and genome size of 5 kb . The eleven variant settings are: We ran 10 replicates for all scenarios , and 20 for “base” , “weak bottleneck” and “no recombination” ( some scenarios are more computationally demanding due to the effect of recombination on coalescent simulations and of genetic diversity on transmission inference ) . For each repeat in each scenario we ran a completely different simulation with different seeds resulting in different transmission and coalescent histories , even when outrbeak or coalescent parameters do not change across scenarios . We ran the BadTrIP MCMC for 5 ⋅ 105 steps for each replicate , sampled from the posterior every 100 steps and with a 20% burn-in . We specified in BadTrIP the true simulated sampling time and removal time of each host , while we specified as introduction time of each host its infection time minus one quarter of the mean duration of infection ( so that the true infection time is contained within the exposure time of the host ) . For SCOTTI we used the same epidemiological data and options as for BadTrIP , except that we ran the MCMC for 2 ⋅ 106 steps for each replicate . We did not allow unobserved cases in SCOTTI . We measured accuracy as the frequency with which the correct transmission source of each host is inferred by a method to be the most likely a posteriori . We also measured calibration as how often the correct transmission source is the the 95% posterior credible set ( the minimum set of sources with cumulative probability ≥ 95% such that all sources in the set have higher posterior probability than all sources outside of it ) . We also used the SVC method [30] to infer transmission from simulated data . This method consists of selecting , for each host d , the set of possible infectors as those cases with most shared variants with d , or , if d does not share variants with other hosts , the cases with the smallest consensus genetic distance from d . If a single possible infector is found , it is assigned 100% posterior probability , otherwise if multiple possible infectors are found they are assigned the same posterior probability . For example , if 4 cases all have 2 shared genetic variants with d , and all other cases have fewer than 2 , than each of those 4 cases is assigned a posterior probability of 25% of infecting d . This is very different from BadTrIP , which always weighs the information from shared variants , genetic distances , and epidemiological data simultaneously from all cases . So , the 4 cases sharing 2 genetic variants with d can have very different posterior probabilities in BadTrIP of being infectors of d , depending on the other data . For example , if one of these 4 cases has very high genetic distance from d , or epidemiological data incompatible with a transmission to d , BadTrIP would infer very low ( or null ) probability of it being the infector of d . We use sequencing and epidemiological data published by Gire and colleagues [40] and analysed by Worby and colleagues [30] . In particular , we use information from sampling dates , nucleotide frequencies and sequencing coverage . We specify the introduction date ( removal date ) of each host as its sampling date minus ( plus ) 21 days . This means that we allow each host to be infected at most 21 days before it being sampled , and to infect others at most 21 days after being sampled . We ran the BadTrIP MCMC until an effective sample size of 1000 was reached for each parameter and for the posterior probability ( requiring ≈ 3 . 5 million MCMC steps ) . to reduce the computational time required we subsampled the reads from each sample to obtain a per-base coverage of at most 100 . BadTrIP is distributed as an open source package for the Bayesian phylogenetic software BEAST2 [39] . It can be downloaded from https://bitbucket . org/nicofmay/badtrip/ or via the BEAUti interface [54] of BEAST2 .
|
We present a new tool to reconstruct transmission events within outbreaks . Our approach makes use of pathogen genetic information , notably genetic variants at low frequency within host that are usually discarded , and combines it with epidemiological information of host exposure to infection . This leads to accurate reconstruction of transmission even in cases where abundant within-host pathogen genetic variation and weak transmission bottlenecks ( multiple pathogen units colonising a new host at transmission ) would otherwise make inference difficult due to the transmission history differing from the pathogen evolution history inferred from pathogen isolets . Also , the use of within-host pathogen genomic variants increases the resolution of the reconstruction of the transmission tree even in scenarios with limited within-outbreak pathogen genetic diversity: within-host pathogen populations that appear identical at the level of consensus sequences can be discriminated using within-host variants . Our Bayesian approach provides a measure of the confidence in different possible transmission histories , and is published as open source software . We show with simulations and with an analysis of the beginning of the 2014 Ebola outbreak that our approach is applicable in many scenarios , improves our understanding of transmission dynamics , and will contribute to finding and limiting sources and routes of transmission , and therefore preventing the spread of infectious disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"sequencing",
"techniques",
"taxonomy",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"population",
"genetics",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"molecular",
"biology",
"techniques",
"dna",
"population",
"biology",
"genetic",
"epidemiology",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"epidemiology",
"evolutionary",
"systematics",
"molecular",
"biology",
"evolutionary",
"genetics",
"nucleotide",
"sequencing",
"biochemistry",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"dna",
"recombination",
"evolutionary",
"biology"
] |
2018
|
Bayesian reconstruction of transmission within outbreaks using genomic variants
|
Pseudomonas aeruginosa ( P . aeruginosa ) is an opportunistic pathogen chronically infecting the lungs of patients with chronic obstructive pulmonary disease ( COPD ) , pneumonia , cystic fibrosis ( CF ) , and bronchiectasis . Cif ( PA2934 ) , a bacterial toxin secreted in outer membrane vesicles ( OMV ) by P . aeruginosa , reduces CFTR-mediated chloride secretion by human airway epithelial cells , a key driving force for mucociliary clearance . The aim of this study was to investigate the mechanism whereby Cif reduces CFTR-mediated chloride secretion . Cif redirected endocytosed CFTR from recycling endosomes to lysosomes by stabilizing an inhibitory effect of G3BP1 on the deubiquitinating enzyme ( DUB ) , USP10 , thereby reducing USP10-mediated deubiquitination of CFTR and increasing the degradation of CFTR in lysosomes . This is the first example of a bacterial toxin that regulates the activity of a host DUB . These data suggest that the ability of P . aeruginosa to chronically infect the lungs of patients with COPD , pneumonia , CF , and bronchiectasis is due in part to the secretion of OMV containing Cif , which inhibits CFTR-mediated chloride secretion and thereby reduces the mucociliary clearance of pathogens .
Respiratory infections are the greatest cause of disease worldwide [1] . A study by the World Health Organization ( WHO ) determined that the global disease burden of lung infections exceeds that of HIV/AIDS , cancer and heart disease and has since 1990 [1] . P . aeruginosa , an opportunistic human pathogen is commonly associated with respiratory infections , particularly nosocomial , ventilator-associated infections and pseudomonal pneumonia in immunocompromised patients , including cystic fibrosis , chronic obstructive pulmonary disease ( COPD ) , ventilator-associated pneumonia , community-acquired pneumonia , and bronchiectasis patients . We have previously demonstrated that outer membrane vesicles ( OMV ) secreted by P . aeruginosa deliver multiple virulence factors into host human airway epithelial cells via a mechanism involving OMV fusion with airway cell plasma membrane lipid rafts and trafficking via an N-WASP induced actin pathway to deliver OMV cargo directly to the host cytoplasm [2] . This provides a mechanism for P . aeruginosa to alter host cell biology without the need for contact with airway epithelial cells , an important consideration in respiratory diseases where P . aeruginosa resides primarily in the mucus layer above the host airway epithelium [2] . Cif , a virulence factor secreted in OMV by clinical isolates of P . aeruginosa , was first described for its ability to decrease the apical membrane expression of the cystic fibrosis transmembrane conductance regulator ( CFTR ) and to reduce chloride secretion [3] , [4] , [5] . The Cif-induced reduction in the apical membrane abundance of CFTR in airway epithelial cells is due to an inhibition of the recycling of endocytic vesicles containing CFTR back to the plasma membrane and redirection of these vesicles to the lysosome where CFTR is degraded . The mechanism by which Cif reduces the recycling of endocytic vesicles containing CFTR is currently unknown , and thus , characterizing this mechanism was the goal of the present study . Many bacteria-derived effectors regulate host pathways , including intracellular vesicular trafficking and ubiquitination , which targets proteins for degradation in the lysosome and proteasome [6] , [7] . Pathogens frequently target the ubiquitination/deubiquitination systems of host cells to suppress the innate immune response and enhance pathogen colonization [8] , [9] . For example , Salmonella produces a DUB , SslE , which reduces the cytotoxicity of Salmonella in macrophages [6] . Altered ubiquitin signaling involves the delivery by the pathogen of a DUB into host cells , thereby reducing the host response to bacterial pathogens . In a recent study , we demonstrated that USP10 , a host cell DUB , deubiquitinates CFTR in endosomes , thereby reducing the lysosomal degradation of CFTR , and maintaining cell and plasma membrane CFTR [10] . However , the effect of Cif on USP10 has not been examined . Thus , the goal of this study was to test the hypothesis that Cif inhibits USP10 , which increases the amount of ubiquitinated CFTR that is degraded in lysosomes , thereby reducing cell and plasma membrane CFTR level . The data in this report demonstrates that Cif stabilizes an inhibitory effect of Ras-GAP SH3 domain binding protein-1 ( G3BP1 ) on USP10 , thereby reducing USP10 mediated deubiquitination of CFTR and increasing the degradation of CFTR in lysosomes . This is the first example of a bacterial toxin that regulates a host DUB . We propose that the ability of P . aeruginosa to chronically infect the lungs of patients with CF , pneumonia , COPD , and bronchiectasis is due in part to the secretion of OMV containing Cif , which inhibits CFTR mediated chloride secretion and thus , reduces the mucociliary clearance of pathogens .
To elucidate the mechanism for the reduction of apical membrane CFTR , we first examined the time course of the effect of Cif on the amount of CFTR in the plasma membrane . To this end , purified P . aeruginosa outer membrane vesicles ( OMV ) containing the Cif toxin were applied to the apical face of polarized human airway epithelial cells . P . aeruginosa OMV were isolated from an overnight culture and diluted to approximate the OMV produced by 108 bacteria . A bacterial count of 108 to 1010 is relevant because this is the bacterial density often detected in CF patient respiratory secretions [11] . After P . aeruginosa OMV treatment , CFTR was measured in cell lysates by Western blot analysis and in the apical plasma membrane by cell surface biotinylation followed by Western blot analysis . Cif rapidly ( 30–60 minutes ) decreased the apical membrane abundance of CFTR and subsequently reduced CFTR protein levels in the cell lysate ( Figure 1A ) . OMV purified from P . aeruginosa clinical isolates and applied to airway epithelial cells also significantly reduced the apical membrane and cell lysate abundance of CFTR ( Figure 1B ) . By contrast , OMV isolated from P . aeruginosa lacking Cif had no effect on CFTR [2] . CFTR has a long half-life at the apical plasma membrane ( 8–24 hours ) because it is efficiently recycled back to the plasma membrane after it is removed by endocytosis [12] , [13] , [14] , [15] . Thus , the ability of Cif to rapidly ( 30–60 minutes ) reduce CFTR abundance in human airway epithelial cells suggests that Cif enhances the endocytic removal of CFTR from the apical plasma membrane and/or reduces the recycling of endocytic vesicles containing CFTR back to the plasma membrane . In a previous publication , we demonstrated that Cif did not alter the endocytic rate of CFTR , but dramatically reduced the recycling of endocytic vesicles containing CFTR back to the plasma membrane [3] . In addition , our previous publication also demonstrated that Cif did not alter the abundance of other apical membrane proteins , like the transferrin receptor or the GPI-anchored protein , gp114 [3] . To begin to elucidate the mechanism whereby Cif altered CFTR trafficking , endosomes were isolated by density gradient purification from polarized human airway epithelial cells that had been treated with either OMV containing Cif or OMV from the cif mutant ( control , in which the cif gene had been deleted ) . These experiments revealed that in control cells , CFTR co-immunoprecipitated with Rab5a and Rab11a , a finding consistent with previous reports that CFTR is localized primarily in early endosomal ( Rab5a-labelled ) and recycling endosomal ( Rab11a-labelled ) compartments ( Figure 1C , [12] , [13] , [14] , [16] , [17] , [18] ) . Although addition of Cif-containing OMV to polarized human airway epithelial cells did not change the amount of CFTR in early endosomes ( Rab5a compartment ) , Cif dramatically shifted the distribution of CFTR from recycling endosomes ( Rab11a-compartment ) to the Rab7a , late endosomal compartment ( Figure 1C ) . These results support the conclusion that Cif redirects CFTR from endosomes that recycle to the plasma membrane to a degradative pathway . To investigate further the trafficking of CFTR in the presence of Cif , we followed the movement of CFTR through intracellular compartments via differential centrifugation and Optiprep gradient fractionation . In these experiments , apical membrane CFTR was biotinylated and the portion of CFTR that started at the apical membrane was tracked as a function of time after exposure to OMV containing or lacking Cif . In control cells ( treated with OMV lacking Cif ) , most CFTR was present in the membrane fraction , but CFTR was also present in the endosomal and a small portion in the lysosomal fraction ( Figure 1D ) . Addition of Cif-containing OMV reduced the amount of CFTR in the plasma membrane and in endosomes , and increased the amount of CFTR in the lysosomal compartment ( Figure 1D ) . These results are consistent with the results presented in Figure 1C demonstrating that Cif reduces apical plasma membrane CFTR by redirecting CFTR from the recycling endocytic pathway ( i . e . , Rab 11a ) to a lysosomal degradative pathway ( i . e . , Rab 7 and lysosomes ) . To provide additional support for the observation that Cif redirects CFTR to lysosomes for degradation , Cif-containing OMV were incubated in combination with control ( vehicle ) or the lysosomal inhibitors chloroquine or ammonium chloride . Both chloroquine and ammonium chloride reduced the Cif-mediated degradation of CFTR ( Figure 2A ) . By contrast , the proteasomal inhibitor lactacystin had no effect on the Cif-mediated degradation of CFTR ( Figure 2A ) . Thus , Cif increased the degradation of CFTR in the lysosome , a conclusion consistent with the results presented in Figures 1C and D . The next series of experiments were conducted to elucidate the cellular mechanism whereby Cif redirected CFTR to lysosomes for degradation . First , we tested the hypothesis that Cif increased the amount of ubiquitinated CFTR , since it is known that ubiquitinated CFTR is degraded in the lysosome [10] , [19] . Airway cells were treated with Cif-containing OMV for various time points , in the presence of a lysosomal inhibitor that prevents ubiquitinated CFTR from being degraded . CFTR was then immunoprecipitated and western blot analysis was performed using ubiquitin antibodies . Figure 2B demonstrates that Cif increased the amount of ubiquitinated CFTR with a time course that is concomitant with a decrease in the amount of CFTR ( Figure 2C ) . Several observations support the view that Cif increased the multi-ubiquitination of CFTR , rather than mono-ubiquitination or poly-ubiquitination . First , the ubiquitin antibody FK1 , which only recognizes poly-ubiquitinated CFTR , did not identify ubiquitinated CFTR in the presence of Cif treatment ( Figure S1 ) . Second , the ubiquitin antibody , FK2 , which recognizes mono- and multi-ubiquitinated proteins , recognized immunoprecipitated CFTR ( Figure 2B ) . Third , the molecular weight of ubiquitinated CFTR in the presence of Cif increased by ∼40 kDa , an amount greater than 8 kDa , the molecular weight of a single ubiquitin moiety . Thus , taken together these results are consistent with the conclusion that Cif enhances the degradation of CFTR primarily by increasing the amount of multi-ubiquitinated CFTR , and its subsequent degradation in the lysosome . Cif may increase the amount of multi-ubiquitinated CFTR , and thereby its degradation in lysosomes , by activating an E3 ligase and/or by inactivating a DUB . To determine the mechanism by which Cif increases the amount of multi-ubiquitinated CFTR , Cif was applied to polarized human airway epithelial cells in OMV , followed by immunoprecipitation of Cif to identify Cif-interacting proteins . Mass spectrometry of the immunoprecipitated proteins revealed interaction of Cif with several DUBs , including Ubiquitin Specific Protease-10 ( USP10 ) and USP34 ( data not shown ) . We previously reported that the DUB USP10 deubiquitinates CFTR in early endosomes of human airway epithelial cells [10] . To determine if Cif inhibits the activity of USP10 , and thereby increases the amount of ubiquitinated CFTR , we used a DUB activity assay to measure USP10 activity in early endosomes ( EE ) of airway epithelial cells [10] , [20] , [21] , [22] . The DUB activity assay employs a HA-UbVME probe that forms an irreversible , covalent bond only with active DUBs . Identification of DUBs covalently linked to the HA-UbVME probe was achieved by immunoprecipitation of the HA-UbVME-DUB complex using an anti-HA monoclonal antibody followed by SDS-PAGE and western blot analysis for the DUB of interest . Using this assay , we demonstrated that USP10 activity was inhibited 49±4% by Cif ( Figure 3A ) . Cif did not alter the activity of other EE-resident DUBs including USP8 or USP34 , thereby demonstrating the specificity of Cif in EE for USP10 ( Figure 3A ) . Moreover , silver stain analysis of the DUB activity assay revealed that Cif does not alter the activity of any DUBs in addition to USP10 ( data not shown ) . To provide additional support for the hypothesis that Cif inactivation of USP10 is responsible for the increase in the amount of ubiquitinated CFTR and its lysosomal degradation , we examined the amount of multi-ubiquitinated CFTR following siRNA knockdown of USP10 . siRNA-mediated reduction of USP10 protein expression ( by 76±4% , Figure 3C ) increased the amount of ubiquitinated CFTR ( Figure 3B , [10] ) . These data are consistent with the view that Cif reduces USP10 activity and thereby increases the amount of multi-ubiquitinated CFTR and its degradation in the lysosome . The next set of experiments was designed to elucidate how Cif inhibits USP10 activity . Three observations suggest that Cif may inhibit USP10 activity by stabilizing the interaction between USP10 and G3BP1 , which inhibits USP10 DUB activity , and also reduces the interaction between USP10 and CFTR . First , published studies have shown that USP10 and G3BP1 interact in yeast and mammalian systems [23] , [24] , [25] . Second , in U2OS bladder cancer cells , the interaction between G3BP1 and USP10 inactivates the deubiquitinating enzyme activity of USP10 [23] . Third , our preliminary mass spectrometry experiments revealed that Cif and G3BP1 interact ( data not shown ) . Thus , studies were conducted to determine if Cif inhibits USP10 activity by stabilizing the interaction between USP10 and G3BP1 , and by reducing the interaction between USP10 and CFTR . As shown in Figure 4A , USP10 interacts with CFTR in the early endosomes isolated from airway epithelial cells , and Cif reduces this interaction by 50±6% . Moreover , Cif increased the interaction between G3BP1 and USP10 by 210±12% ( Figure 4B ) . Thus , these observations reveal that Cif stabilizes an interaction between USP10 and G3BP1 , and also reduces the interaction between USP10 and CFTR . To provide additional support for the immunoprecipitation studies demonstrating that Cif increases the interaction between G3BP1 and USP10 in the early endosomal compartment , we performed bimolecular fluorescence complementation ( BiFC ) studies . BiFC utilizes two half yellow fluorescent protein ( YFP ) sequences fused to two hypothetical interacting proteins ( USP10 and G3BP1 ) . If two fusion proteins interact ( e . g . , 1-154YFP-USP10 ( YN-USP10 ) and 155-238YFP-G3BP1 ( YC-G3BP1 ) ) , a full YFP protein is formed and fluorescence is detected via confocal microscopy [26] . Moreover , by co-transfecting cells with organelle markers , it is possible to identify the compartment where USP10 and G3BP1 interact . In control experiments transient transfection of any one of the eight half-YFP constructs alone in airway epithelial cells did not result in fluorescence in the yellow channel , as expected ( Figure 4C and not shown ) . By contrast , co-transfection of both constructs ( YN-USP10 and YC-G3BP1 ) yielded YFP fluorescence ( Figure 4C ) . Combinations of USP10 and G3BP1 constructs with different orientation of the half YFP protein in the fusion proteins ( N- or C-terminus ) achieved varying degrees of BiFC fluorescence ( Figure 4C and not shown ) . Experiments in Figures 4d–f were performed with the constructs yielding maximal BiFC fluorescence intensity under control conditions , YN-USP10 and YC-G3BP1 ( Figure 4C ) . To provide additional support for our immunoprecipitation studies that the interaction between USP10 and G3BP1 occurs in early endosomes ( Figure 4B ) , BiFC experiments were performed in cells transduced with a baculovirus system expressing eRFP-labeled Rab5a ( an early endosomal protein ) to label early endosomes . Co-localization of the BiFC signal with the early endosomal marker was quantified by intensity correlation analysis using Nikon Elements Software . Mander's overlap coefficients of 0 . 84±0 . 06 demonstrated a high degree of co-localization of the USP10-G3BP1 pair with early endosomes ( Figure 4D ) . Treatment of co-transfected ( YN-USP10 and YC-G3BP1 ) airway epithelial cells with Cif-containing OMV resulted in an increase in BiFC signal , confirming an increased ( 4 . 35±0 . 30 fold ) interaction between USP10 and G3BP1 in the presence of Cif ( Figure 4E ) compared to cells treated with control ( Cif mutant OMV ) . Co-localization studies revealed that the BiFC signal ( USP10-G3BP1 interaction ) was localized to the early endosomal compartment ( Rab5a ) in airway epithelial cells treated with Cif-containing OMV ( Figure 4F , Mander's overlap coefficient of 0 . 76±0 . 11 ) . These studies support the conclusion that Cif stabilizes the interaction between USP10 and G3BP1 in the early endosomes . Finally , if Cif stabilizes an inhibitory interaction between G3BP1 and USP10 , silencing G3BP1 should reduce the Cif-induced decrease in USP10 activity as well as the Cif-induced increase in ubiquitinated CFTR , and its degradation in lysosomes . siRNA-mediated knockdown of G3BP1 protein expression ( by 50%±3% ) eliminated the ability of Cif to inhibit USP10 activity ( Figure 5A ) . If G3BP1 knockdown prevented the Cif-mediated inhibition of USP10 activity , we would predict that knockdown of G3BP1 would also eliminate the increase in multi-ubiquitinated CFTR and degradation of CFTR induced by Cif . Knockdown of G3BP1 did , in fact , block the Cif-mediated increase in the amount of multi-ubiquitinated CFTR ( Figure 5B ) and enhanced lysosomal degradation of CFTR ( Figure 5C ) . Notably , siRNA-mediated knockdown of G3BP1 significantly increased the abundance of CFTR , most likely because endogenous G3BP1 inhibits USP10 ( Figure 5C ) . Two additional siRNA target sequences for G3BP1 also abrogated the Cif-mediated degradation of CFTR ( Figure S2 ) . Taken together , these data confirm our hypothesis that Cif , by enhancing G3BP1 interaction with USP10 , inhibits the ability of USP10 to interact with , and deubiquitinate CFTR .
To our knowledge this is the first report demonstrating that a bacterial toxin , Cif ( PA2934 ) , regulates a host protein ( USP10 ) involved in ubiquitination and lysosomal degradation of an ion channel ( CFTR ) , and thereby modulates the ability of airway epithelial cells to secrete chloride , an important component of the innate immune response in the lung . The data in this report demonstrates that Cif alters the intracellular trafficking of CFTR , redirecting CFTR from endosomes where CFTR recycles to the plasma membrane to lysosomes where CFTR is degraded . Our data support the conclusion that Cif stabilizes an inhibitory effect of G3BP1 on USP10 , thereby reducing its ability to deubiquitinate CFTR and increasing the degradation of CFTR in lysosomes ( Figure 6 ) . These data suggest that the ability of P . aeruginosa to chronically infect the lungs of patients with COPD , pneumonia , CF , and bronchiectasis is due in part to the secretion of OMV containing Cif , which inhibits CFTR-mediated chloride secretion and thus , diminishes the clearance of respiratory pathogens by the mucociliary escalator . Ubiquitin modification of proteins regulates numerous cell processes , including protein degradation , intracellular protein trafficking , and cell signaling [27] . Ubiquitination and deubiquitination are dynamic and regulated processes; over 1000 E3 ligases , which attach ubiquitin moieties to substrate proteins , are encoded by the human genome , whereas ∼90 DUBs remove ubiquitin from substrate proteins [6] , [7] . Thus , the amount of an ubiquitinated protein , and thereby the amount of the protein that is degraded in lysosomes or the proteasome , is regulated by the balance between the activity of E3 ligases and DUBs . Importantly , the abundance of a given protein is thought to be regulated by a small subset of E3 ligases and DUBs [6] , [7] . At the present time , two E3 ligases ( Nedd4-2 and c-Cbl ) have been shown to regulate CFTR trafficking . In pancreatic cells , Nedd4-2 regulates CFTR abundance [28] , but preliminary studies from our laboratory demonstrate that silencing Nedd4-2 does not alter the amount of CFTR in human airway epithelial cells ( unpublished data ) . Recently , we demonstrated that c-Cbl ubiquitinates CFTR in airway epithelial cells [29]; however , siRNA studies from our laboratory demonstrated that siRNA knockdown of c-Cbl did not inhibit the Cif-mediated increase in CFTR degradation ( unpublished data ) . Thus , it is unlikely that Cif increases the amount of ubiquitinated CFTR by activating Nedd4-2 or c-Cbl . However , it cannot be ruled out that Cif may be activating an unknown E3 ligase to increase the amount of ubiquitinated CFTR . Additional experiments , beyond the scope of this study , are required to determine if other E3 ligases ubiquitinate CFTR in airway epithelial cells and to determine if Cif regulates the activity of these E3 ligases . Approximately 90 DUBs remove ubiquitin from target proteins [6] , [7] . The effect of these ligases and DUBs are known to be highly specific . For example , in this and a previous study we observed that neither USP34 , nor USP8 deubiquitinate CFTR , and only USP10 activity was inhibited by Cif [10] , [30] . Our results suggest a model whereby in steady-state conditions , USP10 activity is regulated by dynamic interactions between its target protein , CFTR and its negative regulator , G3BP1 . Upon P . aeruginosa infection and OMV delivery of Cif into host cells , the interaction of USP10 with its negative regulator , G3BP1 , is stabilized and USP10 is sequestered from interaction with CFTR . Thus , CFTR remains multi-ubiquitinated in early endosomes and is targeted for lysosomal degradation ( Figure 6 ) . It is not currently known which ubiquitin linkages USP10 targets , but in our study it appears to target the multi-ubiquitinated CFTR . A previous study reported that G3BP1 interacts directly with the N-terminus of USP10 , a common protein-protein interaction domain in the USP family of DUBs for regulation of DUB activity [23] . While our mass spectrometry data revealed an interaction between Cif and G3BP1 , additional studies are needed to determine if the interaction is direct or indirect , and to elucidate how Cif stabilizes the inhibitory protein complex between USP10 and G3BP1 . Given the importance of the host ubiquitin degradation system in regulating basic cell biology , it is not surprising that many pathogens have evolved to target the ubiquitin pathway to promote their colonization of the host . Pathogen effects on E3 ligases to modify host cell function are well documented , but only recently has pathogen manipulation of the deubiquitinating machinery of the host been investigated [6] , [7] , [8] , [9] , [31] . Several bacterial species have been shown to encode deubiquitinating enzymes , the majority playing a role in dampening the host inflammatory response [6] , [32] , [33] , [34] , [35] . To date , one other host DUB ( i . e . , in addition to USP10 ) has been targeted by a bacterial species . The host-encoded DUB , Cylindromatosis ( CYLD ) , is regulated indirectly by bacterial pathogens through changes in its gene expression , but a mechanism for the altered gene expression has yet to be reported [36] , [37] , [38] , [39] . Infection with Haemophilus influenzae or Eschericia coli induces CYLD expression , which down-regulates the NF-κB inflammatory pathway . CYLD -/- mice have a hypersensitivity to infection with both Haemophilus influenzae and Eschericia coli [37] , [38] , [39] . On the other hand , the CYLD -/- mice experience acute lung injury and increased lethality in response to Streptococcus pneumoniae infection [36] . These opposing effects of CYLD DUB activity in response to different pathogens suggest a potential complexity in targeting host DUBs for therapeutic purposes to combat infection . To enable therapeutic development targeting bacterial effector proteins , and thereby bacterial infections , a better understanding of the mechanism of action of the bacterial effectors is required . The crystal structure of the Cif toxin has recently been solved and shows homology with the α/β hydrolase family of bacterial enzymes [40] , [41] . Cif catalyzes the hydrolysis of epoxide compounds , with specific activity against epibromohydrin and cis-stilbene oxide . Interestingly , mutations to the active site of Cif that reduce epoxide hydrolase activity also reduce the effect of Cif on CFTR degradation [40] . Current studies are underway to elucidate the mechanism by which the epoxide hydrolase activity of Cif promotes the inactivation of USP10 , via enhancement of G3BP1 interaction , leading to the degradation of CFTR . The data in this manuscript is relevant to clinical infections by P . aeruginosa since the Cif toxin is expressed by clinical isolates of P . aeruginosa and in OMV isolated from CF and pseudomonal pneumonia patients ( [4] , unpublished data ) . Accordingly , taken together with previous studies on Cif , the data in this paper are consistent with the view that Cif-mediated reductions in CFTR abundance ( ∼60% ) and chloride secretion ( ∼60% ) by human airway epithelial cells [4] would be expected to reduce mucociliary clearance in the airway , a critical mechanism of the innate immune response to eliminate P . aeruginosa and other pathogens from the airway of patients with COPD , ventilator-associated pneumonia , CF and bronchiectasis [42] , [43] . In addition to the clinical relevance of the Cif toxin on host mucociliary clearance and innate immune defense this study reports data identifying Cif as the first bacterial toxin that inactivates a host DUB . Understanding the mechanism by which bacterial toxins alter host cell biology provides the basis for therapeutic development to inhibit toxin function and potentially reduce bacterial pathogenesis .
The role of Cif in CFTR degradation was studied in human airway epithelial cells ( CFBE41o- cells , homozygous for the ΔF508 mutation ) stably expressing wt-CFTR ( hereafter called airway epithelial cells ) . The derivation and characterization of these cells have been described in detail by several laboratories [13] , [44] . Airway epithelial cells between passages 18 and 27 were grown and polarized in an air-liquid interface culture at 37°C for 6–9 days , as described [13] . To identify active DUBs in airway epithelial cells we used a chemical probe screening approach designed and described in detail by Dr . Hidde Ploegh [20] , [21] , [45] and recently published by our laboratory [10] . The specificity of the HA-UbVME probe for active DUBs was confirmed with the addition of N-ethylmaleimide ( 10 µM ) , which inhibits cysteine protease DUBs , during the labeling reaction [20] , [21] , [45] . To determine if USP10 is expressed in early endosomes , differential centrifugation and fractionation techniques were used to isolate early endosomes from CFBE cells using a protocol adapted from Butterworth et al . [46] . Briefly , polarized CFBE cells , grown on 24 mm permeable membrane supports , were scraped into phosphate-buffered saline , pelleted , and resuspended in 600 µl of HEPES buffer ( 250 mM sucrose , 10 mM HEPES , 0 . 5 mM EDTA at pH 7 . 4 containing protease inhibitors ( Roche Diagnostic Corp . , Indianapolis , IN ) ) . The cells were homogenized with a Dounce homogenizer and passed through a 22-gauge needle 20 times . Following a low speed spin ( 3 , 000× g ) , the post-nuclear supernatant was diluted 1∶1 with 62% sucrose in HEPES buffer and placed at the bottom of a 4 . 4 ml ultracentrifuge tube ( Sorvall , Ashville , NC ) . 1 . 5 ml of 35% sucrose in HEPES buffer was layered on top followed by 1 . 5 ml of 25% sucrose in HEPES buffer and 0 . 5 ml of HEPES buffer . The gradients were centrifuged in a TH-660 rotor at 167 , 000× g for 75 min at 4°C , and the interfaces were collected to isolate the early endosomal fractions . Western blot analysis for various Rab GTPases was used to confirm purity of the early endosomal fraction [10] . To assess the trafficking of Cif through the endocytic trafficking pathway , we used differential centrifugation and an Optiprep continuous gradient to separate the plasma membrane , endosome , and lysosome fractions from airway epithelial cells , a protocol adapted from a previous study [47] . Plasma membrane ( Na/K ATPase ) , endosome ( early endosomal antigen-1 , EEA-1 ) , and lysosome ( LAMP-1 ) resident proteins were used to identify these compartments in the fractionations to identify the localization of CFTR . To assess the amount of ubiquitinated CFTR in airway epithelial cells , a protocol was adapted from Urbe et al . [48] and recently published by our laboratory [10] . To determine if CFTR interacts with USP10 in the early endosome ( EE ) , USP10 was immunoprecipitated from EE fractions isolated from airway epithelial cell lysates by methods described previously in detail [12] . The biochemical determination of plasma membrane CFTR was performed by domain selective cell surface biotinylation using EZ-Link Sulfo-NHS-LC-Biotin ( Pierce ) , as described previously in detail [49] . USP10 and G3BP1 protein expression was selectively reduced using siRNA purchased from Qiagen ( Valencia , CA ) , by methods described previously [12] . In brief , airway epithelial cells were seeded at 0 . 1×106 on 24 mm Transwell permeable membrane supports and on day 4 , post-seeding , cells were transfected with HiPerfect transfection reagent according to the manufacturer's protocol ( Qiagen , Valencia , CA ) . Sequences for siRNAs are: siUSP10 sense 5′ CACAGCUUCUGUUGACUCUTT 3′; siG3BP1 #1 sense 5′ GGAGGAGUCUGAAGAGATT 3′; siG3BP1 #2 sense 5′ CCCUGGUUCCAACAGAAUGTT 3′; siG3BP1 #3 sense 5′ GAAAGAAAUCCACAGGAAATT 3′; siNegative scrambled sense 5′ UUCUCCGAACGUGUCACGU 3′ . Cells were studied on day 8 post-seeding ( i . e . , 4 days after transfection with siRNA ) . The sequences encoding YFP ( 1–154 ) ( yellow fluorescent protein-1–154-peptide ) and YFP ( 155–238 ) were kindly provided by Dr . Tom K . Kerppola ( University of Michigan Medical School , Ann Arbor , MI , U . S . A . ) [50] , [51] and cloned onto the N- or C-terminal end of the human USP10 to produce USP10-YN , USP10-YC , YN-USP10 and YC-USP10 . The human G3BP1 was also fused to the same YFP sequences to produce G3BP1-YN , G3BP1-YC , YN-G3BP1 and YC-G3BP1 . The constructs were purchased from OriGene ( pCMV6 vectors , Rockville , MD ) and verified by DNA sequencing . CFBE-WT cells seeded at 0 . 1×106 on collagen-coated , glass-bottom MatTek dishes , were transfected with 1 µg of a single USP10 and G3BP1 BiFC construct using the Effectene transfection reagent , according to manufacturer's protocol ( Qiagen , Valencia , CA ) . All combinations of the USP10 and G3BP1 BiFC fusion proteins were transfected in pilot experiments and the single combination demonstrating maximum BiFC fluorescence was used in remaining experiments ( YN-USP10 and YC-G3BP1 ) . Two days post-transfection , cells were infected with a baculovirus expressing a RFP-Rab5a plasmid ( Organelle Lights Endosomes-RFP , Molecular Probes , Invitrogen ) , according to the manufacturer's instructions . Samples were incubated in the presence or absence of Cif-containing OMV for 15 min at 37°C and then fixed with 4% paraformaldehyde in PBS for imaging . Z-stack images ( 0 . 4 µm sections ) of labeled cells were acquired with a Nikon Sweptfield confocal microscope ( Apo TIRF 100x oil immersion 1 . 49 NA objective ) fitted with a QuantEM:512sc camera ( Photometrics , Tuscon , AZ ) and Elements 2 . 2 software ( Nikon , Inc . ) . YFP fluorescence emission was measured at 535/30 nm and the fluorescence of Rab5a-RFP , a red fluorescent protein , was measured at 610/30 nm . Experiments were repeated three times , with ten fields imaged for each experiment . The antibodies used were: mouse anti-ezrin antibody , mouse anti-G3BP1 , mouse anti-GFP antibody ( BD Biosciences , San Jose , CA ) ; mouse anti-HA antibody ( Santa Cruz Biotechnology , Santa Cruz , CA ) ; mouse anti-Ubiquitin antibodies ( clones FK2 and FK1 ) ( BioMol , Plymouth Meeting , PA ) ; rabbit anti-USP10 antibody , rabbit anti-USP34 , rabbit anti-USP8 ( Bethyl Laboratories , Montgomery , TX ) ; horseradish peroxidase-conjugated goat anti-mouse and goat anti-rabbit secondary antibodies ( Bio-Rad , Hercules , CA ) . CFTR antibodies were used as described previously [10] . All antibodies and reagents were used at the concentrations recommended by the manufacturers or as indicated in the figure legends . Statistical analysis of the data was performed using Graphpad Prism version 4 . 0a for Macintosh ( Graphpad , San Diego , CA ) . When appropriate , experimental triplicates were performed and all replicates were expressed as a percentage of control before the mean was determined . Means were compared using a t-test or ANOVA followed by Tukeys test , as appropriate . P<0 . 05 was considered significant . Data are expressed as the mean ± SEM . To show data distribution , 95% confidence intervals are presented in the figure legends . Cif ( PA2934 , NP 251624 . 1 ) ; USP10 ( NP 005144 . 2 ) ; G3BP1 ( NP 005745 . 1 ) ; Rab5a ( NP 004153 . 2 ) ; Rab7a ( NP 004628 . 4 ) ; Rab11a ( NP 004654 . 1 ) ; CFTR ( NP 000483 . 3 )
|
In this manuscript , we present a detailed mechanistic study of how a secreted P . aeruginosa virulence factor disrupts mucociliary clearance in the lung by inactivating a host cell deubiquitinating enzyme ( USP10 ) , thereby facilitating the degradation of the CFTR secretory chloride channel , which reduces mucociliary clearance by human airway epithelial cells . To our knowledge , this is the first report of a bacterial toxin that alters host innate immune defense by hijacking host proteins involved in the ubiquitin proteolytic system . P . aeruginosa is recognized as the most common bacterial pathogen in ventilator-associated pneumonia and is commonly isolated from community-acquired pneumonia patients . While P . aeruginosa can cause devastating acute infections , many studies have documented that P . aeruginosa can also generate chronic infections in immunocompromised individuals , including cystic fibrosis , chronic obstructive pulmonary disease and bronchiectasis patients . In these patients , the ineffective host immune response to the bacterial colonization is thought to play a large role in deteriorating lung function and ultimately the death of the patient . Our findings have significant implications to the study of P . aeruginosa infections in the lung , particularly the ability of P . aeruginosa to disrupt critical host innate immune response mechanisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"infectious",
"diseases/nosocomial",
"and",
"healthcare-associated",
"infections",
"microbiology/immunity",
"to",
"infections",
"respiratory",
"medicine/respiratory",
"infections",
"cell",
"biology/membranes",
"and",
"sorting",
"infectious",
"diseases/respiratory",
"infections",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections"
] |
2011
|
A Pseudomonas aeruginosa Toxin that Hijacks the Host Ubiquitin Proteolytic System
|
Human breast cancer has been characterized by extensive transcriptional heterogeneity , with dominant patterns reflected in the intrinsic subtypes . Mouse models of breast cancer also have heterogeneous transcriptomes and we noted that specific histological subtypes were associated with particular subsets . We hypothesized that unique sets of genes define each tumor histological type across mouse models of breast cancer . Using mouse models that contained both gene expression data and expert pathologist classification of tumor histology on a sample by sample basis , we predicted and validated gene expression signatures for Papillary , EMT , Microacinar and other histological subtypes . These signatures predict known histological events across murine breast cancer models and identify counterparts of mouse mammary tumor types in subtypes of human breast cancer . Importantly , the EMT , Adenomyoepithelial , and Solid signatures were predictive of clinical events in human breast cancer . In addition , a pan-cancer comparison revealed that the histological signatures were active in a variety of human cancers such as lung , oral , and esophageal squamous tumors . Finally , the differentiation status and transcriptional activity implicit within these signatures was identified . These data reveal that within tumor histology groups are unique gene expression profiles of differentiation and pathway activity that stretch well beyond the transgenic initiating events and that have clear applicability to human cancers . As a result , our work provides a predictive resource and insights into possible mechanisms that govern tumor heterogeneity .
One of the hallmarks of breast cancer is tumor heterogeneity at both the histological and genomic level . The histological type of the tumor refers to the morphological and cytological patterns evident within a tumor . There are a large number of special tumor histologies recognized for breast cancer [1 , 2] including lobular , cribriform and several other types . The most frequently observed tumor histology is the invasive ductal carcinoma [3] . Similarly , there is a large degree of genomic heterogeneity in human breast cancer , which has been classified using gene expression analysis . Classification of breast tumors into their molecular subtypes based on unique gene expression profiles has led to tumors being described according to their “intrinsic subtype”: Basal-like , Luminal A , Luminal B , Her-2 enriched , Claudin Low and Normal-like breast group [4–6] . Importantly , these intrinsic subtypes of breast cancer provide a basis by which researchers can classify tumor heterogeneity . Importantly , recent work has identified the gene expression relationships between intrinsic subtypes of human breast cancer and specific histological types of breast cancer [2] . Chief amongst their findings was that within intrinsic subtypes of cancer were multiple histological types of cancer . For example , both medullary and metaplastic breast cancer were categorized as claudin low . Further , individual tumors of the same tumor histological types corresponded to different intrinsic subtypes of breast cancer . For example , some medullary tumors were classified as basal and others were categorized as claudin low . These findings suggest that gene expression methods may do better job of organizing tumors into similar disease entities[2] . Collectively , these studies demonstrate that histological and genomic heterogeneity present in breast cancer establishes a complex array of distinct subtypes of tumors[2 , 4] . With this complexity , modeling breast cancer in vivo requires numerous preclinical models that effectively mimic the multiple factors inherent to human breast cancer progression and parallel the molecular profiles of human breast cancer subtypes . While the use of human cell lines and patient derived xenografts offer the opportunity to study human breast cancer in vivo , they rely on immunocompromised hosts . The use of genetically engineered mouse models of cancer offer the advantage and the opportunity to study tumor progression in an immuno-competent system . As a result , a major focus has been to establish which genetically engineered mouse models have parallels in human breast cancer . [7] . Expanding upon these findings with additional tumor models and samples , numerous reports have documented mouse and human counterparts at the level of gene expression [8–12] . In addition , copy number variation at the chromosome [13]and gene level [14]has been predicted from expression data and examined similarity to human breast cancer . The gene level CNV predictions demonstrated that chromosomal alterations were associated with histological subtypes[14] . With gene expression similarities to human breast cancer , a critical need remains to address how the tumor histology of mouse mammary tumors is related to gene expression programs . As seen in human breast cancer , a large number of histological subtypes have been observed for mouse mammary tumors [15] . This includes glandular , acinar , cribriform , papillary , solid , squamous , fibroadenoma , adenomyoepithelioma , adenosquamous , microacinar , adenocarcinoma , comedoadenocarcinoma , and medullary [8 , 15–17] . Prior characterization of mouse models illustrates a number of mouse models with varied histological subtypes present across the population of tumors . For example , amongst Myc initiated tumors , epithelial to mesenchymal ( EMT ) -like , papillary , microacinar , solid , and squamous tumors were observed [18] . Comparison of mouse and human histological subtypes reveals key differences , for example squamous tumors are not frequently observed in human breast cancer [1 , 3] . As such , it is critical to begin to understand how mouse and human tumor pathologies impact the genomic relationships between mouse models and human breast cancer . To address the need to characterize the genomic patterns defining histological subtypes to allow a mouse / human comparison we have examined a wide spectrum of mouse model tumors . In previous work we observed that unsupervised hierarchical clustering of Myc initiated tumors resulted in subclasses that correlated with their histology [19] . Further , even in the presence of loss of the activator E2F transcription factors , clustering arranged tumors according to histology , rather than genotype[20] . This suggested that there are unique gene expression components inherent to histological subtypes apart from the initiating oncogenic events . Using gene expression data from histologically annotated mouse mammary tumors initiated by different oncogenic events , we have developed gene expression signatures that define tumors with squamous or adenosquamous , EMT-like , microacinar , solid , papillary , or adenomyoepithelial tumor histology . Applying these signatures to our published database [9] of mouse mammary tumors we scored mouse tumors for histology , tested which cell signaling pathways tightly correlate with tumor histology , and investigated signature relationships to human breast cancer . Together , this data demonstrates robust signatures that can be used to predict tumor histology and further our understanding of human breast cancer heterogeneity .
To build a gene expression signature that could identify specific histological types of tumors , we utilized publicly available gene expression data that was annotated for mammary tumor histology for each sample analyzed on array . For each tumor type that we built signatures for , histology is described in Table 1 according to descriptions from expert pathologists [15 , 21] . Using significance analysis of microarrays ( SAM ) , we identified genes uniquely and consistently differentially expressed in a specific tumor histology in a training dataset . For example , we utilized histological classifications of tumors from our previous study of the MMTV-PyMT mouse model where squamous , microacinar , and papillary tumors arise [16] ( Fig 1A ) . Using SAM , we filtered out genotype differences to identify genes consistently differentially regulated and intrinsic to the squamous identity ( Fig 1B ) . Focusing only on the genes detected in all four comparisons , we identified 184 genes upregulated in squamous tumors . We did not detect any genes that were consistently downregulated in this comparison . We tested the performance of these genes on the training data using unsupervised hierarchical clustering . As expected , this separated adenosquamous tumors from papillary and microacinar tumors regardless of E2F status ( Fig 1C ) . To validate these genes , we then tested performance on a separate dataset of histologically annotated tumors ( MMTV-Myc tumors ) . Unsupervised hierarchical clustering separated Myc-induced squamous tumors from non-squamous tumors ( Fig 1D ) and importantly gene set enrichment analysis ( GSEA ) showed that Myc induced squamous tumors were significantly enriched for upregulation of the squamous signature genes derived from the MMTV-PyMT tumor dataset ( Fig 1E , Normalized Enrichment Score or NES = 1 . 48 , nominal p-value = 0 . 0 , FDR q-value = 0 . 029 , fwer p-value = 0 . 016 ) . This illustrated the squamous signature genes as robust and valid with the ability to properly classify squamous tumors in another gene expression dataset and in tumors initiated by a different oncogene . Using a very similar approach , we generated gene expression signatures for EMT-like tumors ( S1 Fig ) , microacinar tumors ( S2 Fig ) , papillary tumors ( S3 Fig ) , solid tumors ( S4 Fig ) , and tumors with adenomyoepithelial ( S5 Fig ) content . In each case , potential signature genes were identified using SAM ( q-value ≤ 5% ) doing multiple comparisons between the target tumor histology and other tumor types in the dataset . Unsupervised hierarchical clustering and GSEA was used on a separate histologically annotated dataset to validate the signature . As additional validation of our signatures , we examined individual genes for prior association with histological types in the literature . As shown in Table 2 , several of the squamous signature genes have been shown to be markers for squamous tumors and keratinocytes . Similarly , many of the traditional markers ( such as Zeb1 , vimentin , E-Cadherin ) of EMT were captured in our signatures . In addition , genes from the papillary and adenomyoepithelial signatures also had been observed as markers of these histologies . Together , the ability to detect known histological subtypes across datasets and mouse models as well as the historical use of several individual genes depicts these signatures as robust classifiers of mouse mammary tumor histology . Importantly , each of the histological signatures is provided as a supplemental file ( S1 File ) in GSEA “ . gmt” format as a predictive resource . To test our gene expression signatures in mouse mammary tumors , we utilized two published mouse mammary tumor model databases [9 , 22] . To identify the most likely histology of each tumor in the dataset , we utilized single sample GSEA ( ssGSEA ) and ordered tumors according to their highest scoring signature . With this approach , we observed tumors with robust expression of signature genes for each histology ( Fig 2 and S6 Fig ) . In addition , there were also tumors that did not show strong expression patterns for a particular histology signature , likely indicating a different histology without a predictor . With application of these signatures , we see evidence for profound histological heterogeneity both across and within mouse models . For example , Myc , PyMT , Wnt , Large T , and p53 lines had tumors with a squamous prediction . Indeed , no histological prediction was represented by a single mouse model and most mouse models ( as categorized by the driver event ) showed histological heterogeneity . For example , Wnt , Met , and Myc induced tumor models presented tumors with high scores for each of the other histological subtypes , consistent with reports of histological heterogeneity in these models [18 , 23 , 24] . Alternatively , other models had a preponderance of a particular histological outcome . This is best represented by the Wap-Int3 and Notch induced tumors which were predominantly enriched for the papillary signature . Another model , H-Ras initiated tumors favored microacinar and solid nodular outcomes . Interestingly , models featuring inducible expression of an oncogene , showed elevation of the EMT signature in the recurrent tumors ( S7 Fig ) ; consistent with prior reports [25 , 26] . Finally , predictions organized into figures for each individual mouse model are provided as additional material ( S2–S28 Files ) . As an additional test of the validity and capability of our signatures , we itemized tumors that had been individually annotated for a particular histology by a pathologist ( see blue bars above heatmap , Fig 2 ) . Overall , the pathologist based classification of individual tumors and the classification predicted by the expression signatures demonstrated a high degree of agreement . In addition to this , we cross-referenced the literature to determine whether any of the predicted histologies for a given mouse mammary tumor had been observed in reports for that model . As shown in Table 3 , many of the predicted histological match reports for tumors from individual mouse models . Finally , MMTV-Myc tumors with mixed histology ( multiple histological components within a single tumor ) were noted to have strong scores for individual histology signatures . Thus , we examined matched H&E sections and find that in 89% of samples , the predicted histology was present in at least half the section and 100% concordance where the predicted histology was present in some part of the sample ( S8 Fig ) . Thus , these signatures demonstrate the ability to resolve intra-tumor heterogeneity by identifying the dominant histological component of the tumor being transcriptomically profiled . Importantly , all scores for tumors in each dataset are provided for download ( S29 File ) . With our large dataset and robust performance of the histology signatures , we aimed to test for relationships between tumor histology and other features of mammary gland differentiation . To enable these comparisons , we used the histological classifications made by ssGSEA for each tumor and used standard GSEA to test for enrichment of other signatures in a comparison of predicted tumor histological subtypes ( S30 File ) . We noted prominent associations between histological classes of tumors and signatures for mammary cell types . As shown in Fig 3A , squamous , EMT , and tumors with high adenomyoepithelial content showed high expression signatures for mammary stem cells and mammary basal cells . Amongst these , EMT tumors displayed features most concordant with mammary stem cells . Squamous tumors showed the highest expression of the mammary basal signatures and had gene expression features ( S9A and S9B Fig ) that suggests these tumors are further along the differentiation spectrum than EMT tumors but not as differentiated as other histology types ( S9C and S9D Fig ) . Papillary tumors were more luminal progenitor-like , showing moderate expression of both mammary stem cell and luminal progenitor cell signatures . Finally , the microacinar and solid tumors showed gene expression patterns consistent with those found in mature luminal cells ( Fig 3A ) . We also evaluated the relationships of signatures of breast cancer subtype ( Fig 3B ) [5] . Squamous tumors had highest expression of signatures for basal subtypes of breast cancer . As expected , EMT tumors showed high expression of a signature for claudin low subtypes and showed less luminal or basal-like features . Papillary and tumors with high predicted adenomyoepithelial content showed more moderate expression of all signatures for subtype; while microacinar and solid tumor types had high expression of signatures for luminal breast cancer . As a whole , this suggests a range of differentiation states across histological types . We next tested for relationships between histologies and specific features with tumor progression ( Fig 3C ) . Consistent with prior studies[27] , EMT tumors showed high expression of the hallmark angiogenesis signature . In addition , microacinar and solid tumors exhibit low expression of this signature . In addition to angiogenesis , it was interesting to note differential expression of breast cancer metastasis signatures in these mouse mammary tumor types . The ‘Vantveer Breast Cancer Metastasis Up’ signature was high in microacinar tumors and low in EMT tumors , while EMT tumor showed expression of other metastasis signatures . In addition , squamous tumors showed lower expression of metastasis signatures . Together , this suggests differences in metastatic capacity and mechanism for individual tumor histologies . In addition to phenotypic features , we also tested for key molecular aspects of each tumor histology . In many cases , the histology signatures themselves provide insight into key molecular features , as key signaling molecules were present in several of the signatures . Fig 4A , shows elevation of several pathways consistent with the relationships already detected . For example , Hedgehog and Wnt signaling in squamous tumors[28–31] . In addition , several pathways are shared between histology types . For example , EMT and squamous tumors share high expression of Kras signatures . Microacinar and solid tumors share Erbb2 signature expression , AKT1 signaling via MTOR signature expression , and very low expression of Vhl targets . Examining transcription factors ( Fig 4B ) , a number of key relationships are predicted . Some are known markers , such as TP63 in squamous and Zeb1 , Yap , and Ets transcription factors in EMT tumors are noted . However , unexpected relationships were also present such as Esr1 in microacinar tumors . Despite similarities luminal features , it is interesting to note that the E2F1 signatures distinguishes solid tumors and microacinar tumors . Signature genesets were also tested for overrepresentation in curated pathway databases , offering predictions of additional pathways of interest for each type of tumor histology ( S10 Fig and S11 Fig ) Examining potential miRNAs with GSEA ( Fig 4C ) , suggests tumor types where miRNAs may be actively expressed or lost . For example , mir-202 , mir-17-3p , mir-517 targets are highly expressed in EMT tumors and lowly expressed in the more luminal tumors . Mir-486 was also interesting as its targets showed low expression almost exclusive to microacinar and EMT tumors . Similarly , mir-133A showed evidence for repression in papillary tumors . Taken together , these data suggest a number of key molecular features from pathways , transcription factors , miRNAs for each tumor histology . Given high expression of human breast cancer signatures in certain histologies ( ie- luminal signatures in microacinar ) , we tested whether any of the mouse tumor histology signatures were enriched in subtypes of human breast cancer using the Metabric dataset[32] . As shown in Fig 5A , a portion of the squamous signature was highly expressed in basal tumors . This suggests that mouse mammary squamous tumors are basal-like , but human basal tumors are not known to be squamous . However , human basal tumors and mouse squamous tumors shared similarly high expression of well-studied pathway ligands within the squamous signature ( shared high expression of Wnt10a , Wnt6 , Bmp2 , Bmp7 , and Jag2 ) . Moreover , testing the 45 common highly expressed genes for overrepresentation in pathway signatures indicates possible shared activation of Hedgehog , Wnt , and Bmp pathways in mouse squamous tumors and human basal breast cancer ( S12 Fig ) . Similarly , a subset of the genes that are highly expressed in microacinar tumors were highly expressed in luminal subtypes . Amongst these microacinar genes , many have previously been associated with luminal breast cancer and are also amongst genes that define mature luminal cells ( S13 Fig ) . Finally , both genesets ( up and down ) that define EMT tumors were significantly expressed in claudin low tumors S14A and S14B Fig ) . This result is consistent with numerous reports that mouse EMT tumors are molecularly similar to claudin low tumors[7 , 8 , 12 , 26 , 33] . Together , these data further define appropriate mouse counterparts for study of human breast cancer . With high expression of signature genes in certain subtypes of human breast cancer , it was important to test whether these signatures displayed predictive capacity of clinical events across human breast cancer patients . As shown by Kaplan-Meier analysis , high expression of the EMT and adenomyoepithelial signatures are associated with acceleration of tumor relapse in basal-like breast cancer ( Fig 5B and 5C respectively ) . Adenomyoepithelial signatures were also associated with relapse and earlier onset of distant metastasis in Lum B breast cancer ( Fig 5D , S14C Fig respectively ) , while having high expression of the solid signature was protective in luminal B ( Fig 5E ) . Finally , high expression of the papillary signature genes were associated with accelerated progression to distant metastasis in Her-2 enriched breast cancer ( S14D Fig ) . Together , these results suggest potential mouse tumor types for investigating these human counterparts and prognostic features . Since some of the histology types observed in mouse mammary tumors are often found in other human cancers ( ie- squamous lung tumors , papillary thyroid tumors ) , we sought to test whether the mouse signatures were enriched in other human cancer types . We utilized public gene expression data from the gene expression omnibus and mediated batch effects according to established protocol[9] . Using unsupervised hierarchical clustering arranged many of the tumors with squamous histology across lung , oral , melanoma , and esophageal cancer types into the same cluster with high expression of our murine squamous signature ( Fig 6A , green cluster ) and GSEA testing showed significant enrichment in these tumors ( Fig 6B ) . While the mouse mammary tumor squamous signature extended to other human cancers , the murine papillary signature was not highly expressed in human papillary tumors . The other murine signature with enrichment in human cancers were the EMT signatures that showed concordant expression in a subset of melanoma and metastatic melanoma tumors ( Fig 6A , blue cluster ) . As expected , GSEA showed significant enrichment in these tumors ( Fig 6C , S15 Fig ) . Given that we have itemized many similarities between gene expression profiles of mouse human tumors across cancer types , we tested for unifying features at the level of transcriptomic indicators of pathway activity and differentiation . The concise summary is shown in Fig 7 and more detailed results are available in S16–S18 Figs . Collectively , we observed that murine mammary tumors from the EMT histopathology is similar to human tumors from claudin low breast and melanoma . This includes having gene expression features similar to those found in stem cells and having Kras pathway activity . Mouse and human squamous tumors share enrichment of basal-cell genes and HRas pathway activity , and while similar pathways are active in human basal breast tumors , basal-like breast tumors were enriched for upregulation of luminal progenitor cell genes . We were unable to find human counterparts for the murine papillary tumors in our analyses . For mouse mammary microacinar and solid tumors , luminal features were observed , and like human luminal tumors , enrichment for luminal cell signatures were detected; complete with high expression of ER-target genes . Taken as a whole , these observations suggest that many features of murine tumor histologies are conserved from mouse to human and across several different cancer types .
In this study , we generated and validated signatures for specific histologies that are observed in mouse mammary tumor models . Both training and validation sets utilized prior histological annotations from expert pathologists from a number of studies . With our signature generation and validation approach ( Fig 1 , S1–S5 Figs ) , we show that features of tumor histology span oncogenic mouse models of cancer and human cancers ( Figs 5–7 ) . As shown in Fig 2 and Table 3 , these signatures were predictive of known historical observations for tumor models in our dataset . Thus , we believe these signatures to be a valuable resource tool and have provided our signatures in gene set enrichment analysis format . With a robust capacity to identify tumor histology types , we used this platform to investigate and predict molecular features of each tumor histology all the way from broad features such as differentiation , to specific molecular aspects such as pathway , transcription factor , and miRNA utilization . Based on prior studies , relationships between the mouse EMT signature and claudin low tumors [12 , 33–35] were expected . In addition , recent reports have highlighted cases of melanoma that classify as claudin-low[27 , 36 , 37] . Unlike the relationship between human basal breast cancer and mouse squamous tumors , this relationship is likely due to similar histologies; as breast and melanoma claudin low tumors , like EMT tumors , have been reported to contain spindle-shaped cells[2] . Importantly , evidence of stem-cell like properties and Kras activation was identified in each of these cancer types . Activating mutations in Kras have been observed in mouse EMT[18 , 35 , 38] , however in human breast cancer , the prevalence is somewhat low , as COSMIC[39] reports 80 instances that lack intrinsic subtype information ( with the exception of MDA-MB-231 cells that are claudin low ) . In the case of melanoma , it seems likely measures for Kras activity stem from downstream activating mutations in Braf , which are common to melanomas ( COSMIC reports 44% of melanomas with Braf mutations ) . Together , these data suggest events affiliated with the Kras pathway are important to the EMT / claudin low outcome . We also detected relationships between squamous tumors and human basal breast cancer that seemed to stem from shared activity of multiple pathways . These shared pathways , such as Hras and hedgehog signaling , seem to come from activation events outside of mutations of those genes as both COSMIC and C-Bio-Portal illustrate a low incidence for DNA events on these genes . Although , as reported by TCGA , 32% of basal-like breast cancers harbor amplifications of Kras[40]; suggesting the Hras signature maybe measuring Kras activity in these tumors . Regardless , the shared activation of key pathways supports the use of squamous tumors as a tool for investigating human basal breast cancer at the pathway level . Mouse microacinar tumors showed gene expression traits that define luminal breast cancers . At the pathway level , the relationships between mouse microacinar and human luminal breast cancers is still somewhat perplexing . While both the mouse and human tumors show strong expression of mature luminal cell differentiation signatures and activation of several pathways , mouse microacinar tumors also show activation of Erbb2 signaling , which is traditionally associated with the Her-2 enriched subtype of breast cancer . Furthermore , the microacinar tumors showed high expression of signatures for estrogen receptor signaling . Yet , mouse mammary tumors are notoriously ER-negative by IHC . Indeed , this does draw comparisons to the human setting where Her-2 negative tumors still classify as Her-2 enriched in intrinsic profiling despite the IHC diagnosis as Her-2 negative[41] . This might indicate that similarities to luminal breast cancer are achieved by expression of estrogen receptor target genes by a mechanism other than estrogen receptor itself . Interestingly , several of our mouse histology gene signatures were prognostic in specific intrinsic subtypes of human breast cancer . For example , luminal B tumors with high expression of the solid tumor signature displayed prolonged times to relapse . This finding is particularly of note in light of the recent finding that HER2+ tumors with luminal B gene expression profiles benefitted significantly from trastuzumab[42]; similarly , we note elevation of a Her-2 ( Erbb2 ) signature in murine solid tumors that also have luminal expression profiles ( Fig 4A ) . However , our solid signature was not predictive of prognosis in Her-2 enriched tumors , suggesting the criticality of other pathways differentially regulated between luminal and Her-2 enriched tumors . High levels of EMT and adenomyopithelial signatures were associated with accelerated relapse in basal-like breast cancers ( likely identifying basal-like tumors with claudin-low like properties ) . Indeed , relapse following chemotherapy is common in these tumors [43] and other work has shown an association of EMT phenotypes with chemo-resistance , in part due to lower rates of proliferation and apoptosis[44 , 45] . Taken together , these findings are of particular significance because they may specify high risk patients where alternative therapies may be necessary . In addition , these signatures may suggest appropriate mouse models for testing new therapeutic strategies . The fact that the same histological fates are often achieved despite differences in oncogenic events , genetic background , and promoter ultimately questions the mechanism ( s ) for development of a particular tumor histology . Examination of mammary cell differentiation signatures across tumors revealed unique differentiation states within each tumor histology . Indeed , it is tempting to infer that this indicates the cell of origin leading to tumor initiation and that this cell of origin ultimately drives histological outcome . Indeed , work using the PyMT model suggests that cell of origin plays a role in histological outcome [46] . Yet , more recent work counters that while cell of origin still might be a factor , the initiating oncogenic mutation plays a large role in the histological outcome[47] . In light of these findings[47 , 48] and our study , one might envision that the particular combination of pathways that are activated could commit cells into a specific differentiation state . Alternatively , this could also cause selective outgrowth of specific populations of cells . Ultimately either case would result in tumors forming a particular histology . In support of this , previous work using an inducible Myc mouse model showed that after Myc withdrawal , tumors regressed , and then recurred with tumors mainly being EMT or squamous with activating mutations in Kras [25] . In part , activation of Kras was thus associated with development of these characteristic tumor pathologies . In addition , we and others have observed Kras activation in both of these histological types in other models [35 , 49 , 50] . Indeed , our work presented here might provide predictions as to which differentiation state of cells and which pathways drive the formation of particular histologies . While our method provides robust classification of tumors in our large dataset , there is one important application guideline we wish to highlight . Due to gene centering techniques that are often employed with normalization of gene expression data , predicting tumor histology should be done in settings with adequate tumor heterogeneity or done using methods that adjust for skewed pathological data . In cases where heterogeneity across the dataset is low , we recommend batch adjusting[51 , 52] to combine datasets of interest with large datasets such as our own[9] or others [7 , 10] prior to employing gene signatures . In addition , we wish to refer to the work of Zhao et al which describes in great detail the issues surrounding gene centering and classification of homogenous cohorts while providing alternative approaches for solving such issues[53] . Indeed , the technicalities of gene centering on skewed molecular datasets highlight the necessity of conventional classification methods such as IHC and pathology of H&E stained sections to enable proper data handling . Finally , though traditional tumor classification methods are essential , the gene-signature based classification method here offers several key advantages . First , intra-tumor heterogeneity presents challenges for accurate interpretation of the data that cannot always be addressed by conventional methods . Illustrating this , we examined a large number of tumors presenting mixed tumor histology where the portion of the tumors analyzed on microarray displayed a gene expression profile representing a major histological class . Importantly , the histology predicted by our gene expression signatures were concordant with the major component present in the associated histological section . Therefore , these signatures represent an important tool for resolving mixed cases and ensuring molecular profiles match the expected histology from H&E . Another advantage over conventional methods is the reduced variance in the clinical classification of tumors and classifying cases where histology might be misleading . This is demonstrated by Her-2 enriched Her-2-IHC negative tumors in human breast cancer [41] and ER-target gene enrichment in ER-IHC-negative microacinar tumors from mouse mammary tumor models . Finally , we demonstrate the ability of gene signatures to tie tumor cell phenotypes and functions to supporting pathways that represent therapeutic targets beyond the capacity of IHC . It is our hope that this work’s correlation of gene expression signatures to specific cell biology in the form of tumor histopathologies may provide useful inroads to understanding tumor subtype , tumor progression , and for identifying specific therapeutic strategies aimed at the biological processes upon which the tumor cells depend .
Previously published gene expression data were derived from mouse and human tumors and done in accordance to the ethics statements as reported in their respective publications . Details for assembling the mouse mammary tumor model databases can found [9 , 22] . For the squamous signature , the training data was derived from squamous and non-squamous MMTV-PyMT tumors; this data is deposited on GEO Datasets GSE104397 [54] . All animal work has been conducted according to national and institutional guidelines . These tumors were prepared by isolation of RNA samples from flash frozen tumors using the Qiagen RNeasy kit after roto-stator homogenization . RNA was submitted to the Michigan State University Genomics Core facility for gene expression analysis using Mouse 430A 2 . 0 Affymetrix arrays . The validation set for the squamous signature was from MMTV-Myc tumors found under GSE30805 and GSE15904[8] . The training dataset for generation of the EMT signature is published , GSE30805 and GSE15904 [8] . The validation dataset for the EMT signature can be found GSE41601[12] . Generation of the microacinar signature was done by dividing the published dataset[8] into training and validation sets with random sample selection . The training dataset for generation of the papillary signature is published , GSE30805 and GSE15904 [8] . The validation sets for the papillary signature were from Array Express E-MEXP-3663 [55] and gene expression omnibus GSE20614[56]; batch effects between datasets were mediated using combat[52] . The solid tumor signature was generated using the training dataset GSE41601[12] and validated using GSE73073[57] . Finally the signature for adenomyoepithelial content was generated using from Array Express E-MEXP-3663 [55] , filtered using GSE69290 [58] , and validated on GSE37223[59] . Gene expression data for human squamous and non-squamous tumors was accessed on the Gene Expression Omnibus under the following accession numbers: GSE10245 , GSE10300 , GSE14020 , GSE17025 , GSE18520 , GSE2034 , GSE20347 , GSE21422 , GSE21653 , GSE2280 , GSE2603 , GSE27155 , GSE27678 , GSE29044 , GSE30219 , GSE30784 , GSE3292 , GSE33630 , GSE3524 , GSE35896 , GSE37745 , GSE39491 , GSE39612 , GSE43580 , GSE45670 , GSE4922 , GSE50081 , GSE51010 , GSE6532 , and GSE7553 . These datasets were normalized using Affymetrix Expression Console . Bayesian Factor Regression Methods ( BFRM ) [60] was used to combine datasets and remove batch effects . ( http://www . stat . duke . edu/research/software/west/bfrm/download . html ) . Gene expression signatures were derived using significance analysis of microarrays [61] to detect the genes that were differentially regulated for each tumor histology as illustrated in Fig 1 and S1–S5 Figs . Venn diagrams were generated using online tool available at the following URL: http://bioinformatics . psb . ugent . be/webtools/Venn/ . Unsupervised hierarchical clustering was done using Cluster 3 . 0 and Java Tree View . The color scheme for the heatmap and sample legends were made using Matlab . Gene set enrichment analysis [62] and single sample gene set enrichment analysis was done by converting our gene expression data and gene lists to the specified file formats and using these available modules hosted by Gene Pattern[63] . Tumors sorting for Fig 2 was by sorting tumors for the maximum single sample GSEA score for upregulated genes of any histological type . Pathway and transcription factor overrepresentation analysis was done using Innate-DB[64] and using the Broad Institute’s molecular signatures database ‘investigate gene sets’ web tool[65] . Kaplan-Meier analysis was done using the http://geneanalytics . duhs . duke . edu/Surv_sig . html tool . Samples were assigned to groups based on being above or below the median population value .
|
We developed predictive gene signatures that identify specific histological mouse mammary tumor subtypes with high fidelity to expert pathologist classifications . As a result , these signatures are a powerful tool for classification , particularly in cases of intratumor heterogeneity; where confounding results arise from differences in the tumor portions sent for pathology and separately for molecular analysis . Further , we show that despite differences in the tumor initiating oncogene , histological subtypes in mouse mammary tumor are unified in their transcriptomic profiles and activation of cell signaling pathways . We find that these transcriptomic profiles and activation of key signaling pathways are not only conserved in human breast cancer , but also other human cancer types . Further , the EMT , Adenomyoepithelial and Solid signatures were prognostic in specific human breast cancer subtypes . Indeed , this work provides a new and needed perspective on how mouse models relate to specific human breast cancer subtypes by showing that the tumor histology of the mouse mammary tumor is far more important than the initiating oncogenic event in terms of how the mouse mirrors a specific human subtype .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"breast",
"tumors",
"animal",
"models",
"of",
"disease",
"cancers",
"and",
"neoplasms",
"oncology",
"animal",
"models",
"histology",
"model",
"organisms",
"experimental",
"organism",
"systems",
"bioassays",
"and",
"physiological",
"analysis",
"research",
"and",
"analysis",
"methods",
"animal",
"studies",
"gene",
"expression",
"melanomas",
"breast",
"cancer",
"mouse",
"models",
"microarrays",
"anatomy",
"genetics",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Histological subtypes of mouse mammary tumors reveal conserved relationships to human cancers
|
Discovery in developmental biology is often driven by intuition that relies on the integration of multiple types of data such as fluorescent images , phenotypes , and the outcomes of biochemical assays . Mathematical modeling helps elucidate the biological mechanisms at play as the networks become increasingly large and complex . However , the available data is frequently under-utilized due to incompatibility with quantitative model tuning techniques . This is the case for stem cell regulation mechanisms explored in the Drosophila germarium through fluorescent immunohistochemistry . To enable better integration of biological data with modeling in this and similar situations , we have developed a general parameter estimation process to quantitatively optimize models with qualitative data . The process employs a modified version of the Optimal Scaling method from social and behavioral sciences , and multi-objective optimization to evaluate the trade-off between fitting different datasets ( e . g . wild type vs . mutant ) . Using only published imaging data in the germarium , we first evaluated support for a published intracellular regulatory network by considering alternative connections of the same regulatory players . Simply screening networks against wild type data identified hundreds of feasible alternatives . Of these , five parsimonious variants were found and compared by multi-objective analysis including mutant data and dynamic constraints . With these data , the current model is supported over the alternatives , but support for a biochemically observed feedback element is weak ( i . e . these data do not measure the feedback effect well ) . When also comparing new hypothetical models , the available data do not discriminate . To begin addressing the limitations in data , we performed a model-based experiment design and provide recommendations for experiments to refine model parameters and discriminate increasingly complex hypotheses .
Optimization algorithms attempt to find a parameter set ( or point , i . e . a value for each uncertain parameter ) that gives the best value for some objective defining model fitness , typically the error between model predictions and data; they are commonly identified as either local or global methods ( Figure 2B illustrates these as applied in this study ) . Local optimization starts at a specific parameter set and selects a search direction and step based on the gradient , i . e . how much the error changes with small parameter changes . Global methods use the fitness evaluated for a sampling of parameter sets to then select new samples expected to improve ( algorithmic details vary ) . Qualitative data , such as the fluorescent images of the germarium , define predominantly binary fitness criteria; either the model outputs satisfy the observation or not . They provide no gradient information and discontinuous changes in fitness that may be difficult to identify . Optimization procedures are likely to fail to see where a better solution might lie if a sample did not happen to be placed there . As a result , biological model parameters have typically been estimated either using only data that is quantitative , or by the modeler manually adjusting parameters based on intuition , a very time-consuming process . To design a general procedure for optimization to qualitative data , we considered past efforts in several fields that have addressed aspects of the problem . We predominantly build on the Optimal Scaling method , reviewed below , but it is informative to comment on alternative techniques available . In statistics , regression to qualitative data has a long history [7] , but in contrast with the mechanistic biological context , only minimal models are used . These statistical models are typically linear with some assumed structure on the data ( i . e . a function such as logit or probit is applied to the model values ) . Thresholds are defined to subdivide the continuous model output into intervals , and map each interval to a discrete qualitative output ( e . g . high and low , or a phenotype name ) . The reliance on model linearity limits the immediate utility of past statistical approaches for the non-linear models at hand . In complex model analysis , behavior discrimination [8] has recently been described to define thresholds among different model behaviors , but could be applied to model tuning with qualitative data . It relies on mathematical descriptions of each qualitative behavior to create quantitative metrics to evaluate how near a model is to satisfying each behavior . Defined behaviors can range from simple thresholds to complex time-dependent relationships . A conceptual compromise , Optimal Scaling [9] is an older approach that originated in the social sciences . Similar to behavior discrimination , it evaluates a distance from the point of satisfaction , but is more directly oriented toward model tuning . It also resembles statistical regression problems , but while its past use has been with simple models , it is more generally applicable to complex cases ( i . e . non-linear models ) . For a particular model output , Optimal Scaling uses regression to estimate the optimal quantitative values likely to have generated the qualitative data , i . e . the best-case fit to that model output . While each of these approaches estimates a quantitative fitness , Optimal Scaling offers particularly broad applicability and a focus on the feasible values of the real system . The Optimal Scaling process is illustrated in Figure 2A , and details are provided in Methods . Each time a model output is considered , Optimal Scaling defines quantitative values to replace the qualitative observations; we refer to these as surrogate data ( illustrated as blue circles in Figure 2A , right ) . The surrogate data are intended to represent what could have existed in the true system . To evaluate the best-case fitness to the given model output , the surrogate data values are optimized within the constraint that they still satisfy the qualitative observations ( constraints shown as shaded boxes in Figure 2A ) . The quantitative error between these optimal values and the model output then defines the model fitness , and may be used as the objective for existing optimization techniques . As originally presented for regression of simple models in the social sciences , Optimal Scaling is alternated with a least squares optimization of parameter values [9] , [10] . However , for more complex models , the necessary convexity of that optimization scheme can not be guaranteed . Instead , to apply global and multi-objective optimization techniques , we nest the optimal scaling step fully within the parameter estimation problem [1] ( i . e . optimal scaling is performed explicitly for every parameter set evaluated ) . For details on the optimization process , see Methods . The Optimal Scaling procedure addresses model fitness to the qualitative distributions from germarium images ( examples in Figure 1C ) , but the uncertainty among the different observation types remains . For quantitative data , the trade-off between satisfying each type would be informed by measured experimental variance . For these qualitative data , we suggest that the problem can be viewed as having multiple objectives , i . e . fitting each type of data as a separate objective ( as described in general in [1] ) . In this way , the risk of bias in estimating a single best parameter set is mitigated and a more complete perspective on model performance constructed by evaluating the continuous trade-off among fitting the different data types . An approach that originated in economics and is commonly applied in design optimization , the multi-objective Pareto optimality concept focuses on determining a well-spaced set of points describing this trade-off , each of which corresponds to an optimal point for a different weighting among objectives [11] . Therein , a point is considered Pareto optimal if no other points improve one objective without compromising another . Evaluating a set of Pareto optimal points ( termed the Pareto front , demonstrated in Figure 2C ) comes at a significant computational cost . It is useful to minimize the dimension of the multiobjective problem and group the most similar data together . While we consider the germarium data grouped by the type of observation , data can be grouped in a variety of other ways as suits the problem at hand , including the quality of data ( e . g . nominal , ordinal , ratiometric , etc . ) , or the measurement technique used . The Pareto front is described by plotting the Pareto optimal points on the objective space ( e . g . fitness to wild type data vs . to mutant data ) . Reflected in its placement and curvature , the Pareto front shows the trade-off between objectives , such as how much wild type fitness must be sacrificed to better fit mutants . Accordingly , we can then use the Pareto front to compare the performance of different models . In this study we analyze simplified spatio-temporal models of the germarium subject to a compiled group of available qualitative data by estimating quantitative fitness through Optimal Scaling . To robustly capture data-consistent model behavior , we use multi-objective optimization to estimate a group of representative model parameter sets ( Representatives ) . With this approach , we are able to refine predictive estimates of system behavior , discriminate among multiple models , and estimate the merit of future experiments . To develop and demonstrate the approach in the germarium , we compare alternative regulatory networks generated by a naive screen , as well as mechanistic hypotheses informed by current evidence , including a model based on previous work [12] . We then estimate Representative parameter sets ( Pareto points , in this study ) for each model and discriminate among models based on their simultaneous fitness to published qualitative protein and mRNA distribution data from wild type and mutant organisms . Using the Representatives for each model , we assess which data and parameters should be considered in expanding on the current models , and estimate which future experiments will be most informative by model-based experiment design .
Data were compiled from published images of protein expression across wild type and mutant germaria . All of the data used ( Table 1 ) are qualitative , giving relative expression of proteins , as shown in Figure 1 . Phenotype data are common and indicate fusome morphology ( e . g . Figure 1C , image on right , showing in red the spectrosomes as round and fusomes as branched ) . We correlate the fusome development to Brat expression as an indicator of differentiation . We consider the germarium divided into 4 regions: GSC , CB , Cyst and Posterior . The mapping of these regions onto our 1 dimensional models is shown in Figure 1A lower , indicated by color ( See also Figure S1 in Supporting Information , Text S1 , and refer to Methods for modeling ) . We provide this color map as a reference for model outputs throughout the analysis . Example qualitative interpretations of data are provided below each image in Figure 1C , with a reference to the 1-D model . We note that Bam is known to be repressed by RBP9 , which is present in the posterior region of the germarium [23] . However , the regulation of RBP9 remains unknown . We neglect this posterior repression on Bam in the data , as it is outside the scope of the models we test . Examining the qualitative interpretations , it is apparent that each observation provides only loose constraints , emphasizing the importance of considering many such observations simultaneously . To separate the different types of observations used , we divide data among three categories and independently evaluate model satisfaction of ( 1 ) Wild Type observations , ( 2 ) Mutant observations , and ( 3 ) Behavioral observations . The Behavioral category includes both dynamic constraints , specifying how quickly the cells must respond , and negative phenotypes observed in mutants , which reflect robustness to some perturbations ( indicated in Table 1 ) . These categories were chosen both for biological interest and to aggregate data expected to be similar . For example , Mutants commonly exhibit an all or nothing response over the entire germarium , while Wild Type responses are more graded . We developed a new approach to search for Representatives that best satisfy qualitative data , which incorporates three elements: ( 1 ) the novel application of Optimal Scaling to quantitatively estimate model fitness , ( 2 ) global optimization to select a single best solution for each objective , and ( 3 ) multi-objective optimization to find a set of Representatives irrespective of weighting among objectives . Our implementation of these techniques is illustrated in Figure 2 . For details on each of these processes , consult Methods . The quantification of model fitness by Optimal Scaling in this study is represented in Figure 2A . The procedure generates surrogate data ( blue circles ) that are required to lie within intervals that ensure consistency with qualitative data ( shaded boxes ) . Model error is then calculated as a relative sum of squared error between surrogates and the model output ( green line ) . Note that error is only non-zero when surrogates cannot be perfectly aligned with the model output , as in cell positions 3 and 4 in Figure 2 A . The optimization problem in Optimal Scaling is to select the intervals and surrogates that minimize the model error for a given model output . The global parameter estimation process is depicted in Figure 2B . In this study , we address non-linear spatio-temporal systems with a minimum of 10 states and 18 uncertain parameters . When estimating parameters , dense parameter screening is infeasible and gradient-based searches are not expected to reliably arrive at a global solution , but identify local optima instead . To proceed , we employ a hybrid semi-deterministic approach comprising a sparse global screen followed by a multi-start gradient search . It is important to keep in mind that for these models , available optimization techniques do not guarantee globally optimal or unique solutions ( note the unidentified local minimum in Figure 2B right ) . Finally , to generate the set of representative model parameters , we use multi-objective optimization to find points on the Pareto front , as illustrated in Figure 2C . Here , we determine the Pareto points ( the Representatives ) using the Normalized Normal Constraint ( NNC ) method [24] ( Figure 2C , right ) , with modifications to suit the problem at hand and take advantage of global screening ( see Methods for more details ) . This method performs multiple single-objective gradient searches , with each restricted to lie on a different line so that resulting points are well spaced ( dashed lines in Figure 2C , right ) . The multi-objective approach reliably determines a set of Representatives for the germarium models . The Pareto front identified for the Core regulatory network ( as depicted in Figure 1B ) is shown in Figure 3 . The front is quite convex ( toward the Utopian point ) , but with a significant trade-off between Wild Type and Mutant fitness ( 2nd from left ) . Behavioral fitness closely matches Wild Type ( 3rd from left , note the very small scale ) and exhibits a similar trade-off with Mutant fitness ( right ) . To illustrate fitness , Figure 3B presents examples of both well and poorly fit observations , for the Pareto point nearest the Utopian ( arrows in Figure 3A ) , chosen by Euclidean distance to estimate a midpoint in the trade-off ( fitness at nearby Pareto points was similar , data not shown ) . Most of the observations are satisfied , or nearly satisfied , at this point . The two largest misfits are pMad in a dMyc mutant with ectopic dMyc expression , and pMad in a Brat mutant ( arrows in Figure 3B ) . Examining the data and results for the Brat mutant leads to two important comments . First , we note that the interpretation of the Brat mutant phenotype may be overly aggressive ( i . e . too many cells designated with high pMad ) , due to the discretization of the germarium into the 4 regions considered in this study . The Cyst region extends throughout the 2–8 cell cysts ( cell 3–9 in the 1D model ) , but the indications from data of high pMad expression past the CB do not clearly extend throughout 8 cell cysts [12] . Second , while pMad signaling in the Brat mutant extends beyond the CB , that in Bam mutants does not [25] , suggesting either an unknown regulatory interaction or inconsistency among experiments . To evaluate the Core model in our framework , we compare alternative connections of its regulatory elements , pMad , Bam , Nos , and Brat ( different model structures , i . e . rewiring of network edges ) . Through a simple network inference problem focused on Wild Type data only , we performed a broad screen of alternative networks and identified a set of feasible networks to more thoroughly evaluate , shown in Figure 4 . Considering only Wild Type fitness , we tested the ∼65 k alternatives with only inhibitory connections and additionally performed searches for alternatives that include activation , beginning with 250 k samples . Refer to Methods and Supporting Information ( Text S1 ) for details . Due to the sparse qualitative data , many networks ( hundreds ) were identified as capable of fitting Wild Type data . To refine this large group , we relied on the principle of parsimony , preferring simple networks ( i . e . those with fewer connections ) . Most of the acceptable networks were nested ( i . e . contained simpler acceptable networks plus additional connections ) . From these , we identified five parsimonious variants containing no simpler acceptable networks . We additionally included two networks with extra connections , chosen arbitrarily , to provide a comparison for trade-offs in more complex , but uninformed , models . ( ‘Alt6’ , ‘Alt7’ ) . We compared these networks against the Core , using all available data to generate a Pareto front for each . Pareto fronts determined for each of the alternative networks are shown superimposed in Figure 5A ( between 11 and 46 Representatives per network ) . All networks fail to fully satisfy the data . Examining the Wild Type vs . Mutant projection to compare performance among networks , the Core network dominates most alternatives ( Figure 5A , left ) . However , networks Alt1 and Alt4 perform very similarly to the Core , dominating it at some points . To more closely compare these three models , we examine fitness to individual data ( Figure 5B , plots from the Representatives nearest the Utopian , arrows in Figure 5A ) . For reference , we also present results from Alt3 , which performs poorly ( e . g . compare top plots , where Nos is observed uniformly high in Bam mutants ) . In the Nos mutant where Brat data are uniformly high , Alt3 fails while Alt4 performs quantitatively better ( Figure 5B 2nd row ) . However , the qualitative decrease in the anterior region for Alt4 indicates that its structure is less consistent with the mutant phenotype than the Core or Alt1 ( compare 2nd plots for each network , decrease in Alt4 indicated by arrow ) . The Core and Alt1 networks perform quite similarly , with only minor differences among the unfit data ( Figure 5B , compare 3rd and bottom plots ) . All of the networks compared failed to fit these data . The similar performance of these two networks is explained by similarity in structure . The only difference between the two is that Alt1 lacks feedback of Brat upon Mad ( Figure 4A ) . The basic structure of the Core network is thus well supported , but the data provide poor support for the feedback component . These comparisons and the relative lack of support for feedback exemplify how sparse and qualitative data can be limiting , even when evaluated quantitatively . Rather than suggesting that the well observed feedback element is not involved , this study indicates that the readily available data from genetic experiments are not sensitive to feedback on Mad . Instead , biochemical evidence indicates the repression of Mad in the presence of Brat ( with Pumilio as a cofactor ) in a Drosophila S2 cell line [12] . While such data can be applied directly to define a model , it is not an explicit observation of the germarium that can be compared to simulations . Furthermore , to better understand the system and build parsimonious models , we encourage considering feasible alternatives to the observed interactions , and asking what is necessary for the system to function , i . e . if elements are indispensable , redundant or unimportant . The example experiment design provided below suggests other genetic experiments in the germarium that may be more sensitive to the feedback on Mad . We constructed four hypothetical networks that include additional regulatory mechanisms , as discussed in recent literature . Each contains the Core network along with additional components and interactions ( Figure 6A ) . For simplicity with the current model structure , we do not consider mechanisms based on cell-cell contact and adhesion . We include both the Core and Alt1 networks in the analysis , as they perform nearly indistinguishably . Pareto fronts are presented superimposed in Figure 6B . As indicated by the overlap of all fronts , no clear improvements are made by the hypothetical networks , based on the data at hand . The only indication of improved fitness is a lower error achieved by the Ago model for Wild Type and Behavioral data , while relatively well fit to Mutant data ( examine left and right plots , respectively , at the Mutant anchor point indicated by arrows ) . However , no clear improvements are apparent in individual outputs for the Ago model ( data not shown ) . The lack of clear discrimination among models indicates that the currently available data is inadequate to distinguish the expanded mechanisms tested . Beyond model performance , we use the identified Representatives ( Pareto points ) to assess the relative influence of each observation and parameter as we consider future model development . We examine the distribution of model error to identify which observations are not yet consistently satisfied , and the distribution of parameter sensitivity to identify influential parameters . Using the Representatives , we are able to perform a simple model based experiment design , aiming to estimate the most informative experiments from a set of hypothetical perturbations and measurements . Each Representative of each model produces an individual estimate of the system response in a novel experiment . Potential experiments can then be selected to reduce uncertainty in model parameters , in model outputs or to discriminate among competing model structures . To consider different expectations from data as well as different modeling goals , we present a small variety of approaches to the experiment design problem . First , we focus on a realistic case , expecting qualitative protein distributions , as with current data . Second , we consider a more ideal scenario expecting quantitative distributions of protein concentration . In each , we rank experiments by their utility in discriminating among models and contrast with a ranking focused on refining parameter estimates . In all cases , we correlate utility with variance of the predicted observations , either among models or Representatives , as greater differences are more likely to be discernible . This is an approach implemented previously [36] , [37] , also known as a Maximally Informative Next Experiment [38] and satisfying G-optimality [39] . To illustrate the rankings , the top experiments in each design are presented by heatmaps in Figure 7 , color intensity indicating the relative information gain expected , based on the objective ( e . g . variance with parameters , for reducing uncertainty ) . Refer to Methods for details on the experiment design procedure and calculation of objectives . For each design , the landscape of objective values over all of the experiments considered exhibits a sharp peak , indicating the importance of carefully selecting the experiment ( Supporting Information , Figures S15–S17 in Text S1 ) . Selected experiments from the designs for qualitative data are shown in Figure 8 , where upper panels display expected qualitative predictions and lower plots provide predictions from all Representatives for each model , normalized for visibility . Note that these experiment designs represent a limited range of feasible experiments in this system . More exhaustive model based experiment design carries the promise of more finely resolving system function ( e . g . by considering experiments beyond basic genetic perturbations ) , but is beyond the scope of the current work . In this study we have presented a quantitative model analysis based on qualitative data , via multi-objective optimization with Optimal Scaling fitness estimates . Through our analysis of stem cell regulation in the Drosophila germarium , we have demonstrated the estimation of a set of representative parameter sets , discrimination among multiple models , and model-based experiment design . Using the newly developed process to study the germarium , we have shown the extent to which the existing data employed can discriminate among hypothetical regulatory mechanisms . Current qualitative mRNA and protein image data support the serial inhibition of the ( previously presented ) Core network , but not the feedback element , which is well evidenced in biochemical data . These data do not distinguish among the more complex mechanisms proposed . Toward future modeling , we indicated data that have yet to be satisfied , model parameters that influence fitness , and presented an example experiment design to improve model discrimination . Based on the limited discrimination expected in the experiment designs performed , we recommend first aiming to reduce parameter uncertainty , e . g . by measuring Nos in Nos +/− Dpp +/− . We also recommend pursuing quantitative measurements for Dpp or pMad , as feasible . The designs presented also indicate a variety of other potential experiments . Beyond these initial experiments however , we recommend a more thorough experiment design with careful attention to the feasibility and cost of different experiments . The framework we have developed offers benefits in a wide range of applications . In principle , it is appropriate for any mathematical modeling problem where some or all data are limited to qualitative observations . Naturally , there is particular potential for gains in biological applications , where highly complex systems are prevalent . With the Drosophila germarium as a prime example , developmental biology presents many potential applications as it focuses on pattern formation and spatio-temporal behavior , as in the organization of body axes , limbs , and organ structures [44] . In the broader context of biology and medicine , a variety of fields exhibit similar problems and may also benefit from more widespread use of qualitative data in mathematical modeling studies such as this one . General examples include mechanobiology [45] , neurobiology [46] , [47] , and tissue engineering [48] . We would like to emphasize that the techniques developed in this work accommodate uncertainty in data . If all data can be taken in a rigorously quantitative format , the Optimal Scaling procedure is unnecessary . We anticipate that these techniques will be most valuable when including historical data and when employing new measurements that are not yet refined enough to ensure quantitative reporting .
The aggregate dataset of observations on the anterior germarium was assembled from published literature only . Sources were identified by a primary search of combinations of the terms Drosophila , germarium , GSC , bam , brat , nos , and mad . Searches were performed via the search engine Google Scholar and the databases Medline , PubMed , and Science Citation Index . A secondary search identified additional data sources from references within and articles citing the primary findings . Sources were screened for experiments and relevant data . Data are recorded under a variety of conditions , including genetic mutation and overexpression . Some data were excluded to limit the computational cost of simulations , especially from overexpression studies ( e . g . expression via the yeast Gal4-UAS system [49] ) where the increase of expression over wild type is highly uncertain and requires optimization of experimental parameters . Qualitative data were defined by subjective ( visual ) review of figures and by the interpretations presented by the original authors ( e . g . pMad expression is ranked high in a region because its image intensity there appears clearly greater than elsewhere in the same image , with deference to any declared observations made in the published text ) . Data repeated in multiple works were included one time in the aggregate set , as the observation best representing the consensus from the field . Many data were recorded via fluorescent immunochemistry , which can be ratiometric ( i . e . linearly related to the protein concentration ) and is often used quantitatively after normalization . However , it is important to consider that the quality of data relies on the entire experiment , not just the final measurement type . The linearity of the data , which is required to reliably normalize , cannot be assured without express guarantees both that the experimental reaction steps were designed to preserve a linear relationship and that the images available accurately present the original intensity values . Many of the experiments aggregated for this study employed enzyme linked visualization assays not originally intended for quantitative comparison or modeling , so controls were not presented to ensure that the reactions remained linear . In addition , the germarium is composed of a soft tissue with a high degree of geometric variability between images , limiting the ability to combine multiple images by geometric registration and evaluate measurement uncertainty . Accordingly , all data were treated as ordinal , which reflects the subjective evaluations presented in the source literature . To correlate the Phenotype data to Brat expression , we evaluate the mean Brat concentration over the past 6 hours ( expecting unmodeled delays , and a cell cycle less than 24 hours [50] , [51] ) . Accordingly , data observed with a fusome are assigned a higher rank than those with a spectrosome . Models of the anterior germarium were designed to represent the system as presented in Figure 1A ( see Figure S1 in Supporting Information , Text S1 ) . The models consider secretion of Dpp into the extracellular space , diffusion , receptor binding , and protein levels within each cell , according to the internal regulatory network . Alternative models only differ in the intracellular regulatory network , with the exception that the Diff model includes a secreted molecule not modeled otherwise . Optimal Scaling constructs a set of surrogate data , which may take any value within the qualitative constraints observed . The Optimal Scaling problem selects surrogate data that minimize error from the model . Because absolute error values may vary with the scale of the model output , we use squared relative errors , where are model outputs , are surrogate data , is a small constant to enforce finite values , and is the net variation across the model geometry , where indexes cell position . The inclusion of penalizes error in flat model outputs , i . e . common trivial fitness compromises . The final error is the square root of the sum of average error over all cells in each observation domain , , for . Our experiment designs for qualitative data are performed by exhaustively evaluating all of the experiments we consider . To estimate the qualitative observation for each simulation , we apply the surrogate data interval boundary constraints from our Optimal Scaling formulation . These designs then differ only in the objective by which we rank experiments . In all cases , the goal is to maximize the objective value . For each objective , the color intensity plotted in Figure 7 is determined by mapping between RGB colors ( low ) and ( high ) , relative to the other values in the same heatmap . Toward discrimination , we rank by the variance over models ( ) of expected predictions ( 4 ) or the overlap in prediction distributions , using the Jaccard index ( 5 ) . refers to the set of Representatives for model . The number of Representatives identified per model varied from 11 to 46 . The objective is defined over the 1-D space of the model ( 17 cells after removing the CC ) , and we aggregate to a scalar by evaluating the mean of each model region and taking the sum . We define the index for , which indicates the cell positions for these regions of the model ( as in Figure 1A ) , as well as the number of cells in each ( i . e . ) . ( 4 ) ( 5 ) To calculate the Jaccard index , is the number of Representatives in model , while is the number of Representatives of model that predict an output also predicted for model , and vice versa for . To refine parameters , we rank experiments by the sum over models of the variance of predictions among Representatives ( 6 ) . ( 6 ) In the experiment design for quantitative data , we use the local sensitivity results previously discussed , which approximate . In the objective for discrimination , we aggregate by taking the mean sensitivity across Representatives ( sensitivities with inconsistent sign will cancel ) , and rank by variance among models ( 7 ) . To refine parameters , we evaluate the sensitivity variance over Representatives and rank by the sum of this variance over models ( 8 ) . ( 7 ) ( 8 )
|
We developed a process to quantitatively fit mathematical models using qualitative data , and applied it in the study of how stem cells are regulated in the fruit fly ovary . The available published data we collected are fluorescent images of protein and mRNA expression from genetic experiments . Despite lacking quantitative data , the new process makes available quantitative model analysis techniques to reliably compare different models and guide future experiments . We found that the current consensus regulatory model is supported , but that the data are indeed insufficient to address more complex hypotheses . With the quantitatively fit models , we evaluated hypothetical experiments and estimated which future measurements should best refine or test models . The model fitting process we have developed is applicable to many biological studies where qualitative data are common , and can accelerate progress through more efficient experimentation .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"bioengineering",
"systems",
"biology",
"developmental",
"biology",
"signal",
"transduction",
"biological",
"systems",
"engineering",
"developmental",
"signaling",
"stem",
"cells",
"stem",
"cell",
"niche",
"biology",
"computational",
"biology",
"molecular",
"cell",
"biology",
"engineering",
"signaling",
"in",
"selected",
"disciplines"
] |
2014
|
Model-Based Analysis for Qualitative Data: An Application in Drosophila Germline Stem Cell Regulation
|
The burden of typhoid in sub-Saharan African ( SSA ) countries has been difficult to estimate , in part , due to suboptimal laboratory diagnostics . However , surveillance blood cultures at two sites in Nigeria have identified typhoid associated with Salmonella enterica serovar Typhi ( S . Typhi ) as an important cause of bacteremia in children . A total of 128 S . Typhi isolates from these studies in Nigeria were whole-genome sequenced , and the resulting data was used to place these Nigerian isolates into a worldwide context based on their phylogeny and carriage of molecular determinants of antibiotic resistance . Several distinct S . Typhi genotypes were identified in Nigeria that were related to other clusters of S . Typhi isolates from north , west and central regions of Africa . The rapidly expanding S . Typhi clade 4 . 3 . 1 ( H58 ) previously associated with multiple antimicrobial resistances in Asia and in east , central and southern Africa , was not detected in this study . However , antimicrobial resistance was common amongst the Nigerian isolates and was associated with several plasmids , including the IncHI1 plasmid commonly associated with S . Typhi . These data indicate that typhoid in Nigeria was established through multiple independent introductions into the country , with evidence of regional spread . MDR typhoid appears to be evolving independently of the haplotype H58 found in other typhoid endemic countries . This study highlights an urgent need for routine surveillance to monitor the epidemiology of typhoid and evolution of antimicrobial resistance within the bacterial population as a means to facilitate public health interventions to reduce the substantial morbidity and mortality of typhoid .
Typhoid fever is a systemic infection caused by the Gram-negative bacterium Salmonella enterica serovar Typhi ( S . Typhi ) that continues to be a serious global health problem and a major cause of morbidity and mortality in low-middle income countries [1] . It is estimated that the yearly incidence of typhoid fever exceeds 20 million cases , with over 200 , 000 deaths [2 , 3] . Defining the burden of typhoid fever is a challenge in settings where there are few diagnostic microbiology facilities , with diagnosis often based on clinical history of fever , malaise , and abdominal pain . Unfortunately , these symptoms have considerable overlap with several other febrile illnesses and clinical diagnosis is therefore inaccurate [4] . Nigeria is one of the most densely populated countries in Africa with large areas of urban development . Thus , it is perhaps surprising that little reliable data are available on microbial culture of the etiologic agents of bacteremia in children or adults . This poses a challenge for data comparison with other regions , including other sub-Saharan African countries where such data are available [5–7] . In general , febrile illnesses among children in Nigeria are presumed by clinicians to be caused by malaria , which is still very common in many parts of the country . Only if fever persists following an empiric course of anti-malarials , is typhoid then considered as a potential cause of infection [8] . In studies from central and northwest Nigeria [9] , we found that S . Typhi was the commonest cause of bloodstream infections in children , particularly in those living in the proximity of Abuja city located in central Nigeria . Until recently , molecular epidemiological studies on S . Typhi were compromised by a lack of genetic resolution , limiting the ability to define the population structure of the bacteria and identify transmission patterns . This is because S . Typhi is a relatively monomorphic pathogen with limited genome variation [10] . However , sequencing-based approaches have facilitated the stratification of S . Typhi into multiple genotypes [11] ( see Wong et al . 2016 , under review in Nature Communications , NCOMMS-15-25823 , manuscript included ) . Whole genome sequencing in particular can unequivocally identify phylogenetic relationships with important genetic traits such as antimicrobial resistance [12] . Here we report whole genome-based analysis of 128 bloodstream isolates of S . Typhi from children residing in two regions of Nigeria , and compared these with data from other countries in Africa , including the West African subregion .
Nigeria has a population of approximately 177 million people making it the most populous country in sub-Saharan Africa [13] . The two study sites in Nigeria were the Federal Capital Territory ( FCT ) and Kano . The FCT is a federal territory in central Nigeria and covers a land area of 8 , 000 square kilometers . It is the home of the capital city Abuja , a “planned” city , built in the 1980s . It was officially made Nigeria’s capital in 1991 replacing the previous capital in Lagos . In 2006 , the population was estimated at 1 . 7 million [14] . The FCT continue to experience rapid population growth; it has been reported that some areas around Abuja have been growing at an annual rate of 20–30% , and the current population may be as high as 5 . 7 million [14] . The rapid spread of squatter settlements and shantytowns in and around the city limits contribute to this rapid growth . The rainy season begins in April and ends in October . Within this period there is a brief interlude of Harmattan , occasioned by the Northeast Trade Wind , with the main features of dust haze , intensified coldness and dryness . The annual total rainfall for the FCT is in the range of 1 , 100 to 1 , 600 mm . The population is diverse , with increasing representation from the major ethnic groups of Hausa , Yoruba , and Igbos following the development of the FCT and relocation of the federal capital [15] . Of note , there is also perennial malaria transmission , mostly due to Plasmodium falciparum , and the HIV prevalence is 7 . 5% amongst pregnant women attending antenatal clinics [16] . Kano is the capital of Kano state in northwest Nigeria . According to the 2006 census , Kano state has a population of 9 . 38 million , which is comprised predominantly of Hausa and Fulani ethnic groups [17] . It is recognized as one of the fastest growing cities in Nigeria with a population density of about 1 , 000 inhabitants per km2 . It lies within the Sahel savannah region with daily mean temperature of about 30–33°C during the dry months of March to May and 10°C during the autumn months of September to February . Rainy season varies from year to year , but typically commences in May and ends in October , with an average annual rainfall of 600mm . The dry season starts from November to April [18] . The entire state is within the meningococcal disease belt and malarial transmission is seasonal [17] . HIV prevalence among women attending antenatal clinic is 1 . 3% [16] . The enrolment sites at FCT are as previously described [9 , 15] . Briefly , children aged less than 5 years were enrolled from primary , secondary and tertiary healthcare facilities on presentation with an acute febrile illness and symptoms suggestive of sepsis . In Kano , we enrolled children from Aminu Kano Teaching Hospital ( AKTH ) , Hasiya Bayero Pediatric Hospital and Murtala Specialist Hospital . While AKTH serves as a tertiary referral center , the other two facilities provide primary and secondary healthcare services . The combined outpatient attendance for children at these three facilities is about 1 , 000 daily . Both study sites included patients from the newer settlements on the outskirt of Abuja and around Kano where the level of sanitation is poor and access to potable water limited . A structured questionnaire was used to collate the clinical information . Study data were collected and managed using REDCap electronic data capture tools hosted at the University of Nebraska Medical Center [19] . IBM SPSS for statistics was used for data analysis . Dichotomous variables were analyzed using χ2 or χ2 for trend tests [20] . Clinical information was collected using a structured questionnaire after obtaining a signed informed consent from the child’s parent or legal guardian . This study was approved by the ethics committees of the FCT , National Hospital Abuja , Zankli Medical Center , Federal Medical Center Keffi , Aminu Kano Teaching Hospital , and UNMC , Omaha Institutional Review Board . Blood sampling and processing were as previously described [9 , 15] . Briefly , we utilized only aerobic blood culture bottles and held cultures in the Bactec 9050 incubator for a maximum of 5 days . Bacteria were identified by a combination of colony morphology and biochemical assays . For example , the API 20E system ( bioMérieux , France ) was used to identify Enterobacteriacae . Antimicrobial susceptibility profiles of the bacteria were determined by the Kirby-Bauer disk diffusion test using standard interpretative criteria [21] for locally available antimicrobials ( amoxicillin , co-amoxiclav , ceftazidime , ceftriaxone , nalidixic acid , ciprofloxacin , ofloxacin , sulfamethoxazole , trimethoprim-sulfamethoxazole , chloramphenicol , tetracycline , streptomycin , gentamicin , kanamycin , azithromycin , imipenem ) in order to provide immediate management of patients . Bacterial isolates were stored in skimmed milk at -70°C and further characterized at the Clinical Microbiology Laboratory of the University of Nebraska Medical Center ( UNMC ) . Antimicrobial susceptibility testing was performed at the UNMC Microbiology laboratory using the Epsilometer test ( Etest; bioMérieux , France ) according to standard methods . Minimum inhibitory concentration ( MIC ) values were interpreted according to Clinical Laboratory Standards Institute ( CLSI ) standards [21] . Due to the lack of CLSI standards , a streptomycin MIC of ≥16 mg/L was considered resistant in these studies . All Salmonella isolates were identified to the serotype level using the Bioplex 200 ( Bio-Rad ) as previously described using the CDC standard Salmonella molecular serotyping protocol [22–24] . A total of 128 S . Typhi isolates were identified in these studies for whole genome sequencing . S . Typhi DNA was prepared using the Wizard Genomic DNA Kit ( Promega , Madison , WI , USA ) as per manufacturer’s instructions . Index-tagged paired end Illumina sequencing libraries were prepared as previously described [25] . These were combined into pools each containing 96 uniquely tagged libraries and sequenced on the Illumina Hiseq2000 or Miseq platforms ( Illumina , San Diego , CA , USA ) according to manufacturer’s protocols to generate tagged 100 or 150 base pair ( bp ) paired-end reads with an insert size of 300–400 bp . Sequence reads were deposited in the European Nucleotide Archive under accession ERP005877 and a full list of accession numbers for each sample is available in S1 Table . Sequence data from 1 , 831 additional S . Typhi isolates from 63 countries , generated previously in the same manner ( Wong et al . 2015 ) [12] , were also included in the study ( reads are available in the European Read Archive under accession ERP001718 ) . For analysis of single nucleotide polymorphisms ( SNPs ) , the paired-end reads were mapped to the reference genome of S . Typhi CT18 ( accession number AL513382 ) , including the chromosome and plasmids pHCM1 and pHCM2 [26] , using SMALT ( version 0 . 7 . 4 ) ( http://www . sanger . ac . uk/resources/software/smalt/ ) . SNPs were identified as previously described , using samtools mpileup [27] and filtering with a minimum mapping quality of 30 and a quality ratio cut-off of 0 . 75 [25] . The allele at each locus in each isolate was determined by reference to the consensus base in that genome , using samtools mpileup [27] and removing low confidence alleles with consensus base quality ≤20 , read depth ≤5 or a heterozygous base call . SNPs called in phage regions , repetitive sequences ( 354 kbp; ~7 . 4% of bases in the S . Typhi CT18 reference chromosome , as defined previously [10] ) or recombinant regions ( ~180 kbp; <4% of CT18 reference chromosome , identified using an approach described previously [25 , 28] ) were excluded , resulting in a final set of 23 , 300 chromosomal SNPs . The maximum likelihood ( ML ) phylogenetic tree was built from 23 , 300 SNP alignment of 1 , 961 isolates , including one S . Paratyphi A ( accession number ERR326600 ) to provide an outgroup for tree rooting . We used RAxML ( version 7 . 0 . 4 ) [29] with the generalized time-reversible model and a Gamma distribution to model site-specific rate variation ( the GTR+ substitution model; GTRGAMMA in RAxML ) . Support for the ML phylogeny was assessed via 100 bootstrap pseudo-replicate analyzes of the alignment data . The ML trees were displayed and annotated using iTOL [30 , 31] . Plasmids and acquired antimicrobial resistance genes were detected , and their precise alleles determined , using the mapping-based allele typer SRST2 [32] together with the ARG-Annot database of antimicrobial resistance genes [33] and the PlasmidFinder database of plasmid replicons [34] . SRST2 was also used to identify mutations in the gyrA , gyrB , parC and parE genes that have been associated with resistance to quinolones in Salmonella and other Gram-negative bacteria [35–38] .
Blood cultures were performed for the evaluation of 10 , 133 acutely ill children , aged 0–60 months , from September 2008 until April 2015 , in the FCT ( including Abuja ) and Kano located in central and northwest Nigeria , respectively [9] . At FCT 6 , 082 children were enrolled between June 2012 and March 2015 , of whom 457 ( 8% ) had clinically significant bacteremia . Of these 110 ( 24% ) had invasive salmonellosis , consisting of S . Typhi in 84 cases and non-typhoidal salmonellae ( NTS ) in 26 cases . In Kano from January 2014 until April 2015 clinically significant bacteremia was detected in 609 ( 15% ) of 4 , 051 children: salmonellae accounted for 364 ( 60% ) of 609 cases , of which 296 were S . Typhi and 68 were NTS . Across both regions Salmonella species accounted for 24–60% of bacteremia with S . Typhi being the most common serovar isolated with a total of 380 isolates ( 76–79% ) [9] . A selection of one hundred and twenty-two S . Typhi from the FCT and six from Kano , all isolated between 2008–2013 , were randomly selected and sequenced via Illumina HiSeq and MiSeq ( see Methods ) . The genomes of the Nigerian isolates were compared to that of the S . Typhi CT18 reference strain and a previously published global collection of approximately 2 , 000 S . Typhi isolates [12] . A phylogeny was built by extracting single nucleotide polymorphisms ( SNPs ) from the whole genome sequences , excluding likely recombination events and repetitive sequences that could confound phylogenetic analysis as described in Methods . The SNP data were also used to assign each isolate to one of 62 previously defined genotypes; details of the source and genotype of all Nigerian isolates is given in Table 1 and S1 Table . The distribution of the 128 Nigerian S . Typhi within the global phylogenetic tree is shown in S1 Fig . This global phylogeny includes 238 isolates from other countries in Africa , and the Nigerian isolates all cluster with other African isolates . Detailed phylogenetic relationships amongst the 366 African isolates are shown in Fig 1 , and an interactive version of the phylogeny and map are available for exploration online at http://microreact . org/project/styphi_nigeria . The majority of Nigerian S . Typhi ( 84/128 , 66% ) belonged to genotype 3 . 1 . 1 ( these isolates were assigned to H56 under the old typing scheme of Roumagnac et al ( 2006 ) [11] ) . This dominant genotype is relatively common across Africa , predominantly western and central countries ( Fig 1 ) . The Nigerian isolates formed a tight phylogenetically clustered subgroup within the 3 . 1 . 1 subclade ( Fig 1 ) , suggesting recent local expansion , and included isolates from both Abuja and Kano , suggesting intra-country transmission . Interestingly , in the wider African collection genotype 3 . 1 . 1 was represented by isolates from neighboring Cameroon and across West Africa ( Benin , Togo , Ivory Coast , Burkina Faso , Mali , Guinea and Mauritania ) suggesting long-term inter-country exchange within the region ( Fig 1 ) . Most of the remaining isolates belonged to four other genotypes , indicating that these are also established genotypes in circulation at the study sites in Nigeria . These genotypes , highlighted in Fig 1 , are 2 . 2 . 0 ( n = 13 ) , 2 . 3 . 1 ( n = 8 ) , 4 . 1 . 0 ( n = 8 , H52 under the old scheme ) and 0 . 0 . 3 ( n = 7 , H12 ) . Nigerian isolates of genotypes 2 . 2 . 0 and 2 . 3 . 1 were closely related to isolates from neighboring Cameroon and West African countries and not found elsewhere , supporting regional transmission similar to the dominant genotype 3 . 1 . 1 ( see map in Fig 1 ) , while genotype 4 . 1 . 0 was more widespread across Africa . Interestingly genotype 0 . 0 . 3 ( previously identified in India and Malaysia ) , which accounted for >5% of Nigerian isolates , maps very close to the root of the global S . Typhi tree , suggestive of older circulating isolates . A further six other genotypes were also detected amongst the Nigerian isolates , represented by 1–2 isolates each ( Table 1 ) . Of note , genotype 4 . 3 . 1 ( H58 ) , which has become dominant elsewhere in sub-Saharan Africa and accounts for the majority of antimicrobial resistant typhoid globally , was not detected in the Nigerian studies . Fig 2 shows the proportion of S . Typhi isolates that were resistant to one or more antimicrobials , and the proportion that were multidrug-resistant ( MDR; defined as resistance to ampicillin , chloramphenicol and trimethoprim-sulfamethoxazole ) , each year from 2008–2013 . The majority of isolates were MDR throughout this period ( Fig 2 ) . Fig 3 and Table 2 show the distribution of antimicrobial resistance determinants in the Nigerian isolates . Most of the 3 . 1 . 1 ( H56 ) isolates carried genes encoding resistance to ampicillin , chloramphenicol , tetracycline and sulfamethoxazole ( blaTEM-1 , catA1 , tetB , dfrA15 , sul1 ) . These were located on an IncHI1 plasmid , similar to that commonly found in MDR S . Typhi 4 . 3 . 1 ( H58 ) . The same profile was identified in a single isolate of 0 . 0 . 3 , indicative of local plasmid transfer between the co-circulating genotypes . Genotype 2 . 3 . 1 isolates were found to carry IncHI1 plasmids encoding these resistance genes , as well as resistance determinants sul2 and strAB . An IncHI1 plasmid carrying blaTEM and tetB was also identified in one 2 . 2 . 0 isolate . Interestingly , nine genotype 3 . 1 . 1 isolates lacked the IncHI1 plasmid . However , four of these carried plasmids of other incompatibility groups . Three isolates ( 3135STDY5861338; 3135STDY5861351; 3135STDY5861282 ) harbored a novel IncY plasmid ( blaTEM-198 , catA1 , tetB , dfrA14 , sul1 ) and one ( 3135STDY5861242 ) harbored a plasmid-related to the Kpn3 plasmid ( blaTEM-198 , tetAR , dfrA14 , sul1 , sul2 , strAB and also qnr-S , which mediates fluoroquinolone resistance ) . Thus , plasmid-mediated MDR is common in Nigerian S . Typhi from the regions under study . We identified only six S . Typhi isolates with quinolone resistance-associated mutations in gyrA ( one with S83F; five with S83Y ) . The affected isolates were all of the dominant genotype 3 . 1 . 1 , including the three that carried IncY plasmids and three that carried IncHI1 plasmids . No other polymorphisms were detected in the quinolone resistance determining regions of the gyrA or parC genes of Nigerian S . Typhi isolates .
Vanessa K . Wong1 , 2 , Stephen Baker3 , 4 , 5 , Derek Pickard1 , Julian Parkhill1 , Andrew J Page1 , Nicholas A . Feasey6 Robert A . Kingsley1 , 7 , Nicholas R . Thomson1 , 5 , Jacqueline A . Keane1 , François-Xavier Weill8 , Simon Le Hello8 , Jane Hawkey9 , 10 , 11 , David J . Edwards9 , 11 , Zoe A . Dyson9 , 11 , Simon R . Harris1 , Amy K . Cain1 , James Hadfield1 , Peter J . Hart12 , 13 , Nga Tran Vu Thieu3 , Elizabeth J . Klemm1 , Robert F . Breiman14 , 15 , 16 , Conall H . Watson17 , Samuel Kariuki1 , 14 , Melita A . Gordon18 , 19 , Robert S . Heyderman20 , 19 , Chinyere Okoro1 , 2 , Jan Jacobs21 , 22 , Octavie Lunguya23 , 24 , W . John Edmunds17 , Chisomo Msefula19 , 25 , Jose A . Chabalgoity26 , Mike Kama27 , Kylie Jenkins28 , Shanta Dutta29 , Florian Marks30 , Josefina Campos31 , Corinne Thompson3 , 4 , Stephen Obaro32 , 33 , 34 , Calman A . MacLennan1 , 12 , 35 , Christiane Dolecek3 , 4 , Karen H . Keddy36 , Anthony M . Smith36 , Christopher M . Parry37 , 38 , Abhilasha Karkey39 , E . Kim Mulholland5 , 40 , James I . Campbell3 , 4 , Sabina Dongol39 , Buddha Basnyat39 , Amit Arjyal39 , Muriel Dufour41 , Don Bandaranayake42 , Take N . Toleafoa43 , Shalini Pravin Singh44 , Mochammad Hatta45 , Robert S . Onsare14 , Lupeoletalalelei Isaia46 , Guy Thwaites3 , 4 , Paul Turner4 , 47 , 48 , Sona Soeng48 , John A . Crump49 , Elizabeth De Pinna50 , Satheesh Nair50 , Eric J Nille51 , Duy Pham Thanh3 , Mary Valcanis52 , Joan Powling52 , Karolina Dimovski52 , Geoff Hogg52 , Thomas R . Connor53 , Jayshree Dave54 , Niamh Murphy54 , Richard Holliman54 , Armine Sefton55 , Michael Millar55 , Jeremy Farrar3 , 4 , Alison E . Mather56 , Ben Amos57 , Grace Olanipekun58 , Huda Munir59 , Roxanne Alter60 , Paul D . Fey60 , Kathryn E Holt9 , 11 and Gordon Dougan1
|
Typhoid fever , a serious bloodstream infection caused by the bacterium Salmonella Typhi , is a major cause of disease and death around the world . There have been limited data on the epidemiology of typhoid in many countries in sub-Saharan African , including Nigeria . Recent evidence , however , showed that typhoid was an important cause of bacteraemia in children residing in two regions of Nigeria . Here , we analyzed the whole genome sequences of 128 S . Typhi isolates from two studies in order to elucidate the population structure and characterize the genetic components of antimicrobial resistance . We found that the multiple S . Typhi genotypes identified were closely related to other S . Typhi from neighboring regions of Africa and that multidrug resistance ( MDR ) was common among these isolates , and in many cases was associated with the IncHI1 plasmid known to cause MDR typhoid . These results provide evidence that typhoid was established in Nigeria as a result of several independent introductions into the country and that there has been extensive exchange of S . Typhi in and around the region of West Africa . This study emphasizes the importance of surveillance to improve our understanding of the epidemiology of typhoid , which is needed to underpin public health measures to reduce the spread of disease and facilitate patient management .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"geographical",
"locations",
"microbiology",
"vertebrates",
"salmonella",
"typhi",
"animals",
"mammals",
"bacterial",
"diseases",
"enterobacteriaceae",
"global",
"health",
"pharmacology",
"united",
"kingdom",
"bacteria",
"bacterial",
"pathogens",
"africa",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"antimicrobial",
"resistance",
"medical",
"microbiology",
"microbial",
"pathogens",
"salmonella",
"typhoid",
"nigeria",
"people",
"and",
"places",
"cats",
"microbial",
"control",
"biology",
"and",
"life",
"sciences",
"europe",
"amniotes",
"organisms"
] |
2016
|
Molecular Surveillance Identifies Multiple Transmissions of Typhoid in West Africa
|
Interorgan lipid transport occurs via lipoproteins , and altered lipoprotein levels correlate with metabolic disease . However , precisely how lipoproteins affect tissue lipid composition has not been comprehensively analyzed . Here , we identify the major lipoproteins of Drosophila melanogaster and use genetics and mass spectrometry to study their assembly , interorgan trafficking , and influence on tissue lipids . The apoB-family lipoprotein Lipophorin ( Lpp ) is the major hemolymph lipid carrier . It is produced as a phospholipid-rich particle by the fat body , and its secretion requires Microsomal Triglyceride Transfer Protein ( MTP ) . Lpp acquires sterols and most diacylglycerol ( DAG ) at the gut via Lipid Transfer Particle ( LTP ) , another fat body-derived apoB-family lipoprotein . The gut , like the fat body , is a lipogenic organ , incorporating both de novo–synthesized and dietary fatty acids into DAG for export . We identify distinct requirements for LTP and Lpp-dependent lipid mobilization in contributing to the neutral and polar lipid composition of the brain and wing imaginal disc . These studies define major routes of interorgan lipid transport in Drosophila and uncover surprising tissue-specific differences in lipoprotein lipid utilization .
Lipoproteins allow the transport of lipids between different organs . In humans , perturbed lipoprotein levels correlate with metabolic disease , but to which extent they contribute to tissue pathology is unclear . Animals synthesize a huge variety of lipids that form cellular membranes , function as signaling molecules , and constitute the major storage and transport form of energy . The lipid composition of different cell types and tissues is important for biological function . To what extent do lipoproteins influence these cellular properties ? Mammals have two types of apolipoproteins that scaffold particles with different functions [1] . Several proteins of the exchangeable apolipoprotein family , including apoA-I , scaffold high-density lipoproteins ( HDL ) , which mediate reverse cholesterol transport . ApoB scaffolds very low-density lipoproteins ( VLDL ) and chylomicrons , which are secreted by the liver and gut , and deliver fat and sterols to peripheral tissues . Mammalian apoB acquires lipid in producing cells by a process requiring MTP [2] , [3] . In humans , MTP deficiency blocks secretion of apoB-containing lipoproteins , resulting in abetalipoproteinemia [4] . This causes fatty liver , intestinal lipid malabsorption , and defects in peripheral tissue function including ataxia , retinal degeneration and anemia [5] . On the other hand , elevated levels of apoB-containing lipoproteins are a hallmark of metabolic syndrome , a pathological condition comprising wide-ranging dysfunctions in different tissues . These include obesity , diabetes , heart disease and increased risk of dementia [6] , [7] . Mammalian tissue culture cells preferentially derive fatty acids and cholesterols from lipoproteins , but can switch to endogenous synthesis if lipoproteins are not provided [8] , [9] . However , it is not clear to what extent autonomous synthesis suffices for different tissues to maintain a normal lipidome in vivo . Furthermore , while the influence of dyslipidemia on the plasma lipidome has been well studied , less attention has been paid to organism-wide changes in tissue lipid composition . Advances in lipid mass spectrometry are only beginning to make such studies possible [10] . To investigate how lipoproteins influence tissue lipid composition requires a system where lipoproteins can be manipulated in a time and tissue-dependent manner . Drosophila genetics could provide a tool to easily control lipoprotein levels in an organism whose metabolism shares many similarities with that of mammals [11] , [12] . In Drosophila , the molecular mechanisms controlling storage and mobilization of neutral lipid in cellular lipid droplets resemble mammalian pathways [13] , [14] . This similarity even extends to the progression of metabolic diseases caused by dysfunctions in lipid metabolism [15] . The major lipoproteins of Drosophila and other insects , the lipophorins ( Lpp ) , are similar to mammalian apoB-containing lipoproteins; their scaffolding apolipoproteins , the apolipophorins ( apoLpp ) , are members of the apoB family , which is conserved throughout the animal kingdom [16] . Moreover , Drosophila lipoprotein receptors resemble those of mammals [17] . The low-density lipoprotein ( LDL ) receptor homologues LpR1 and LpR2 promote Lpp internalization [18] , but also appear to increase cellular neutral lipid storage by non-endocytic mechanisms [19] . Similarly , the role of heparan sulfate proteoglycans as endocytic lipoprotein receptors is conserved between mammals and flies [20] , [21] . Insect lipoproteins have been best studied in Manduca sexta and Locusta migratoria , large insects amenable to biochemical and physiological manipulations [22] , [23] . It has been proposed that insects produce Lpp exclusively in the fat body [24] , which functions analogously to mammalian liver and white adipose tissue [25] . However , Lpp can take up neutral lipid from both fat body and gut when added externally to explanted tissues . This lipidation mechanism requires activity of another lipoprotein , Lipid Transfer Particle ( LTP ) [26] , [27] , [28] , and appears to differ from the MTP-dependent lipidation of mammalian apoB in the secretory pathway of producing cells . In Drosophila , initial studies have shown that Lpp knock-down in the fat body causes accumulation of neutral lipid in the gut , providing in vivo support for models developed from physiological experiments in other insects [29] . However , whether Drosophila might produce particles similar to the LTP of other insects has not been addressed , because the gene ( s ) encoding insect LTP apolipoproteins have been unknown . The Drosophila genome does encode a homologue of MTP , and this protein facilitates secretion of mammalian apoB and locust apoLpp in heterologous tissue culture systems [30] , [31] . This raises the possibility that Drosophila Lpp assembly might resemble that of mammalian apoB-containing lipoproteins , but the requirement for MTP has never been examined in vivo . Here , we characterize lipoprotein function in Drosophila . We identify three circulating lipoproteins in Drosophila larvae , and analyze their source , functions , and mechanisms of secretion and lipid loading . The development of high-resolution shotgun lipidomics for the first time allows the precise and comprehensive quantification of many different lipid species with a sensitivity that makes it possible to study individual Drosophila tissue lipidomes . Therefore , we have combined the power of Drosophila genetics with mass spectrometry to investigate systematically and quantitatively how lipoproteins influence tissue lipid composition .
We started our study of Drosophila lipoprotein metabolism with a genome search for potential apolipoproteins . Many proteins involved in interorgan lipid transport harbor vitellogenin-N domains , including apoB , MTP and vitellogenins [16] , [32] , [33] . BLAST searches with the vitellogenin-N domain of human apoB yield four fly genes: apolpp [34] and mtp [31] , as well as two novel genes , CG15828 and CG31150 ( Figure 1A ) . apolpp and CG15828 seem to have arisen by gene duplication of an ancestral insect apoB homologue ( Figure 1B ) . As will be shown below , the protein encoded by CG15828 assembles a lipoprotein that functions similarly to a lipid transfer particle , LTP , identified in Locusta and Manduca [35] , [36] . We therefore refer to it as apoLTP . CG31150 , which has been recently shown to be mutated in crossveinless d ( cv-d ) [37] , is most closely related to vitellogenins ( Figure 1B ) . To ask whether the fly genome encoded exchangeable apolipoproteins like those scaffolding mammalian HDL , we searched for sequences similar to apolipoproteins A-I and E . No Drosophila protein had significant homology . Neither was there a homologue of apoLipophorin III , a structurally related exchangeable apolipoprotein in Locusta and Manduca [38] . Thus , there is no evidence for apolipoproteins of this family in Drosophila . To ask which of the vitellogenin-N domain proteins might scaffold lipoproteins , we fractionated hemolymph from feeding third instar larvae in isopycnic gradients and probed for apoLpp , apoLTP and Cv-d . These proteins are all present in circulation ( Figure S1A ) , and fractionate at densities consistent with different degrees of lipidation ( Figure 1C ) . ApoLpp is posttranslationally cleaved into apoLII and apoLI , which assemble the lipoprotein Lpp ( 1 . 13–1 . 14 mg/ml ) . Like apoLpp , apoLTP harbors furin cleavage sequences C-terminal to the vitellogenin-N domain , and is cleaved into two polypeptides , which we denote apoLTPII and apoLTPI . ApoLTPII and apoLTPI assemble a higher density ( 1 . 23 mg/ml ) lipoprotein , LTP . Cv-d is poorly lipidated ( 1 . 24 mg/ml ) , consistent with its similarity to vitellogenins , which contain little lipid [39] . To investigate the relative amounts of Lpp , LTP and Cv-d in circulation , we used silver and Coomassie staining to detect them in hemolymph fractionated by density and size ( Figure 1D , Figure S1B ) . These methods detect two prominent bands corresponding to apoLII and apoLI . ApoLTPI and Cv-d are detectable , but much less abundant . Each of these proteins is also present in embryos and adults ( Figure S1C ) . We do not detect any other proteins in low-density fractions , suggesting that no other abundant lipoproteins exist in larval hemolymph . To assess the amount of lipid associated with each lipoprotein , we quantified hemolymph lipids in different density fractions by shotgun mass spectrometry . More than 95% of hemolymph lipids co-fractionate with Lpp ( Figure 1E , Figure S1D ) . The fractions containing LTP and Cv-d account for less then 1% and 0 . 5% of hemolymph lipids , respectively . Thus , Lpp carries the bulk of lipids in circulation . The most abundant Lpp lipids are diacylglycerol ( DAG ) ( 70 mol% ) and phosphatidylethanolamine ( PE ) ( 20 mol% ) ( Figure 1F ) . Sterols comprise 5 mol% of Lpp lipids; the remainder includes triacylglycerol ( TAG ) , phosphatidylcholine ( PC ) , ceramide ( Cer ) and ceramide phosphorylethanolamine ( CerPE ) . More than 95% of Lpp DAG species have a combined acyl chain length of only 26 or 28 carbons , suggesting they contain mostly medium-chain fatty acid residues ( 12 or 14 carbons ) ( Figure 1G ) . This differs strikingly from Lpp phospholipids , which have a combined acyl chain length of 32 to 36 carbons , suggesting they contain long-chain fatty acid residues ( 16 or 18 carbons ) . It was more difficult to assess the lipid composition of the low-abundant LTP and Cv-d . However , we noted that fractions containing LTP are relatively enriched in sphingolipids with hydroxylated fatty acids ( Figure S1E ) . Mammalian apoB-containing lipoproteins are secreted from both liver and gut . While the insect fat body similarly secretes Lpp , the insect gut does not appear to produce Lpp [23] , [24] . To ask which of the different Drosophila lipoproteins were produced by the fat body or gut , we used reverse transcription PCR to determine the presence of apolpp , apoltp , and cv-d transcripts in these organs . apolpp , apoltp , and cv-d transcripts are readily detectable in the fat body ( Figure S2 ) . In contrast , none of them can be detected in the gut . Thus , the Drosophila larval gut does not produce any of these lipoproteins . To ask what fraction of circulating Lpp , LTP and Cv-d was derived from the fat body , we blocked their production in this tissue by RNAi . Fat body-specific knock-down strongly reduces levels of Lpp , LTP and Cv-d in the hemolymph , establishing this organ as the major source of circulating lipoproteins ( Figure 2A–2C ) . mtp transcripts are readily detectable in the fat body , similar to what we observed for transcripts of the different apolipoproteins ( Figure S2 ) . To ask whether production of Lpp , LTP or Cv-d depended on MTP , we knocked down MTP in the fat body by RNAi and examined hemolymph lipoproteins . MTP RNAi causes the buildup of the uncleaved full-length apolipoprotein precursors apoLpp and apoLTP in the fat body , and strongly reduces hemolymph Lpp and LTP levels ( Figure 2D ) . Thus , Drosophila MTP has a conserved function in the production of apoB-family lipoproteins in vivo . However , MTP RNAi does not reduce levels of Cv-d in circulation; thus , not all proteins with vitellogenin-N domains depend on MTP for their release to circulation . These data distinguish the vitellogenin-like lipoprotein Cv-d from canonical vitellogenins in other organisms , whose release is promoted by MTP [40] , [41] . Although Lpp originates in the fat body , we previously showed that its knock-down causes the buildup of neutral lipid in the gut [29] . To ask whether Lpp , or other lipoproteins , are recruited to the gut , we assessed lipoprotein levels in this organ by Western blotting . These experiments show that Lpp and LTP are readily detectable in the gut ( Figure 2E ) . LTP is most abundant in gastric caecae , and in subsets of the anterior and posterior midgut ( Figure 3A ) . Lpp has a broader distribution , but is enriched in the same regions . Interestingly , neutral lipid droplets are most abundant in these same regions of the anterior and posterior midgut , suggesting they may be involved in dietary lipid mobilization . To confirm that LTP and Lpp in the gut are derived from the fat body , we blocked Lpp and LTP secretion from the fat body using tissue-specific MTP RNAi . This strongly reduces their levels in the gut ( Figure 2D ) . Thus , Lpp and LTP produced in the fat body enter circulation and are recruited to the gut . To ask whether either Lpp or LTP were required to export lipid from the gut to circulation , we studied mtp mutant larvae , which arrest in the first larval instar and do not secrete these lipoproteins ( Figure S3A–S3D ) . Loss of MTP increases the size and number of neutral lipid droplets in the anterior and posterior midgut and , more moderately , in the gastric caecae ( Figure 3B ) . We wondered whether lipoprotein production in the fat body would be sufficient for the mobilization of lipids from the gut . To investigate this , we restored MTP activity specifically in the fat body of mtp mutant larvae ( Figure S3E ) . This reduces size and number of lipid droplets in the midgut to wild-type levels , indicating that Lpp and LTP production in the fat body suffices to effect lipid export from the gut ( Figure S3F ) . To investigate individual requirements for Lpp and LTP for lipid export from the gut , we blocked their production by mutation , or by RNAi-mediated knock-down of the respective apolipoprotein . apolpp mutants die as embryos; apolpp is transcribed in the embryonic yolk ( Figure S3G ) [34] , suggesting that embryonic lethality might result from a failure to mobilize maternal lipid stores . Initiating Lpp RNAi during larval stages reduces Lpp levels by more than 90% ( Figure S3H ) and dramatically enlarges neutral lipid droplets in the gut ( Figure 3C ) . Interestingly , expression of two independent Lpp RNAi constructs also reduces LTP levels in the hemolymph and in all organs , including the fat body ( Figure 2E , Figure S3H , and data not shown ) . Thus , LTP production and/or turnover may be influenced by Lpp . Similar to mtp mutants , apoltp mutants arrest in early larval development ( Figure S3I–S3L ) . Surprisingly , although LTP carries such a small proportion of hemolymph lipids , apoltp mutant larvae also accumulate large neutral lipid droplets in the gut ( Figure S3M ) . The lipid accumulation caused by loss of LTP appears identical to that caused by loss of MTP or Lpp . RNAi-mediated knock-down of LTP in the fat body reduces levels of LTP by more than 95% ( Figure S3H ) and causes a gut phenotype indistinguishable from that of apoltp mutants ( Figure 3C ) . LTP RNAi does not reduce the amount of circulating Lpp ( Figure 2E ) ; thus , gut lipid accumulation in LTP RNAi animals cannot be caused by reduced Lpp levels . Taken together , these data suggest that LTP must act with Lpp in the gastric caecae and subsets of the anterior and posterior midgut to effect dietary lipid mobilization . Unlike Lpp and LTP , the vitellogenin-like protein Cv-d is barely detectable in the gut ( Figure 2D ) , and loss of Cv-d does not cause obvious perturbations in gut lipid export ( Figure S4A , S4B ) . Cv-d RNAi does not prevent the development of fertile flies , and Cv-d is not enriched in embryos ( Figure S1C ) , suggesting that unlike canonical vitellogenins it also does not function in embryonic lipid metabolism . This is consistent with the finding that unrelated yolk proteins constitute the major storage proteins in the Drosophila embryo [42] . We were surprised that LTP knock-down produced such a dramatic lipid accumulation phenotype in the gut , because LTP carries a minor proportion of hemolymph lipids . We therefore wondered whether LTP acted catalytically to promote lipid export from the gut to Lpp . To ask whether loss of LTP might change the lipid composition of circulating Lpp , we compared Lpp from wild-type and LTP RNAi hemolymph by density gradient centrifugation . Indeed , LTP RNAi increases Lpp density ( Figure 4A ) , suggesting that Lpp particles contain less lipid when LTP is absent . To directly test the idea that gut lipids are loaded onto Lpp by LTP , we incubated explanted guts with different combinations of lipoproteins . We collected hemolymph either from LTP RNAi animals or from lipid-starved animals . Both treatments reduce Lpp lipid content ( see Figure S9A ) . However , the hemolymph from lipid-starved animals contains LTP . We incubated the different hemolymph preparations with guts dissected from LTP RNAi animals . After 4 h , we recovered the hemolymph and analyzed Lpp density . In the presence of LTP , Lpp shifts to lower densities after incubation with explanted guts ( Figure 4B ) . In contrast , Lpp density does not change when guts are incubated with LTP-free hemolymph . We conclude that LTP facilitates lipid export from the gut to Lpp . As the LTPs of other insects were shown to have similar lipid transfer activity [26] , and also resemble Drosophila LTP in density and in the size of two apolipoprotein subunits [35] , [36] , we suspect that these LTPs are scaffolded by orthologous apoB-family proteins . Manduca LTP can facilitate lipid exchange between Lpp and the fat body . Stimulating adult fat bodies with adipokinetic hormone yields net transfer from the fat body to Lpp [28] . In contrast , in feeding Manduca larvae , net lipid flux is from Lpp to the fat body [27] . To ask whether LTP from Drosophila larvae promoted loading of Lpp with fat body lipids , we incubated hemolymph containing lipid-poor Lpp derived from LTP RNAi animals with either wild-type fat bodies or LTP RNAi fat bodies . In contrast to the gut , we did not observe net lipid transfer from fat body to Lpp , regardless of the presence of LTP ( Figure S5A ) . Thus , LTP does not appear to promote the loading of Lpp with neutral lipids at the fat body in feeding larvae , similar to the situation in Manduca . Previous work demonstrated that Manduca LTP could function as a carrier that shuttles lipids between donor and acceptor lipoproteins [43] . This led us to wonder whether Drosophila LTP acted as an intermediate in the transfer of lipids from the gut to Lpp . If so , then loss of Lpp might trap these lipids in LTP particles . To test this , we asked whether removing Lpp altered LTP density . Indeed , LTP shifts to lower density fractions in hemolymph from Lpp RNAi animals ( Figure 4A , Figure S5B ) , suggesting that lipids are loaded onto LTP before being transferred to Lpp . While MTP transfers lipids to apoB in the secretory pathway of producing cells [2] , [3] , LTP must function differently in the gut , since both Lpp and LTP are recruited there from circulation . To investigate the LTP/Lpp lipid transfer mechanism , we examined their subcellular localization in the lipid droplet-rich regions of the posterior midgut . Lpp accumulates both in the overlying muscle layer that surrounds the gut , and on the basal ( outward facing ) side of absorptive enterocytes ( Figure 4C , 4D ) . In contrast , LTP is not detectable in muscle but accumulates strongly in basal regions of enterocytes . Its subcellular localization extends apically to abut the level where lipid droplets are found . There is little obvious subcellular colocalization of LTP and Lpp in the midgut . The subcellular localization of LTP raised the possibility that it was internalized by enterocytes; we therefore wondered whether endocytosis was required for lipid mobilization from the gut . To address this , we induced expression of dominant negative dynamin in enterocytes and monitored neutral lipid droplets at different times following induction . Within 3 h , neutral lipid droplets accumulate over a broader region of the posterior midgut , and their number within individual enterocytes increases ( Figure 4E ) . By 6 h , most of the posterior midgut is filled with large lipid droplets , similar to the gut of lipoprotein-deficient larvae . Within the same time frame , LTP shifts its subcellular localization to accumulate at the cell boundaries of enterocytes ( Figure 4F ) . In contrast , the subcellular distribution of Lpp appears unaltered . This suggests that endocytosis of LTP may be required for lipid mobilization from the gut . A model consistent with these data is that LTP is internalized by enterocytes , loaded with lipids in an endocytic compartment , and subsequently transfers its lipid cargo to Lpp . We note that 24 h after induction of dominant negative dynamin , lipid droplets in the gut are strongly reduced ( ) . Since dynamin blocks not only endocytosis , but also some plasma membrane delivery routes , we suspect that blocking endocytosis for longer periods of time might compromise delivery of proteins involved in lipid uptake . The vast majority of lipids in the hemolymph are carried by Lpp , and Lpp RNAi reduces the amount of all hemolymph lipid species over 10-fold ( Figure 5H ) . To ask which lipids depended LTP for their transfer to Lpp , we quantified changes in hemolymph lipids of LTP RNAi animals . LTP RNAi specifically reduces the levels of medium-chain DAG ( DAG 26 , DAG 28 ) and sterols ( Figure 5H , Figure S5D ) by about 70% . In contrast , levels of PE , the major polar Lpp lipid , are not changed . The amounts of several minor Lpp lipid classes ( PC , TAG , sphingolipids ) increase slightly . These data suggest that LTP specifically facilitates loading of DAG and sterols onto Lpp . We reasoned that cargo transferred to Lpp by LTP might specifically accumulate in the gut upon either LTP or Lpp knock-down . We therefore asked whether DAG and sterol increased under these conditions . Wild-type guts contain both the medium-chain DAG found in Lpp and smaller amounts of long-chain DAG ( Figure 5B , Figure S6 ) whose combined acyl chain length resembles those of cellular and Lpp phospholipids ( Figure 5A ) . Upon LTP or Lpp RNAi , medium-chain DAG increases 5–8 fold with respect to polar lipids ( Figure 5D , 5E ) . Long-chain DAG increases moderately , and contributes less to the total elevation in gut DAG . These data confirm that the gut uses LTP to export medium-chain DAG to Lpp . Although the gut must also be the source of Lpp sterol ( Drosophila are sterol auxotrophs ) , sterols do not accumulate in this organ upon lipoprotein knock-down ( Figure S7A , S7B ) . It is possible that sterol esterification may increase when export is blocked; however our current methods do not allow us to detect sterol esters . Strikingly , loss of either LTP or Lpp also causes a strong increase in TAG in the gut ( Figure 5F ) . Since the minor amounts of TAG normally present Lpp particles do not decrease upon LTP RNAi ( Figure 5H ) , this cannot reflect a block in TAG export to Lpp . Interestingly , medium-chain TAG is most strongly elevated ( Figure 5G ) . This suggests that some medium-chain fatty acids eventually exported as DAG can be stored as TAG , when export from the gut is blocked . We wondered whether incorporation of medium-chain fatty acids into TAG was an obligate intermediate in the production of medium-chain DAG . To address this , we perturbed the two well-characterized lipolytic systems known to be required for hydrolysis of TAG in the Drosophila fat body – the Adipocyte Triglyceride Lipase homologue , Brummer ( Bmm ) , and the Adipokinetic Hormone Receptor ( AKHR ) regulated lipase system [13] , [44] . We first quantified gut TAG and DAG species in bmm and akhr mutant larvae . Loss of either Bmm or AKHR causes TAG accumulation that is biased towards medium-chain species – similar to the effect of lipoprotein knock-down ( Figure 5I , Figure S8A , S8B ) . Thus , the gut requires both lipolytic systems to mobilize TAG at a normal rate . Despite this , neither perturbation reduces medium-chain DAG in the hemolymph ( Figure 5K , Figure S8C ) . Levels of medium-chain DAG in the gut actually increase slightly ( Figure 5J , Figure S8D ) . We conclude that even under these conditions where TAG storage is favored over lipolysis , the gut can supply normal levels of DAG to Lpp . To ask whether these lipolytic pathways might function redundantly to generate Lpp DAG , we knocked down Bmm and AKHR in enterocytes , alone and in combination , and measured resulting changes in hemolymph DAG . While neither knock-down alone reduces hemolymph DAG ( similar to the effect of single mutants ) simultaneous knock-down lowers hemolymph DAG by 15% ( Figure S8E ) . These data suggest that some Lpp medium-chain DAG is derived from medium-chain TAG by Bmm and AKHR-dependent lipolysis . However , other mechanisms suffice to generate the majority of Lpp medium-chain DAG . We were intrigued by the distinctive fatty acid composition of Lpp DAG . The combined acyl chain length in these DAG species ( 26–28 carbons ) differs not only from that of phospholipids , but also from that of dietary lipids – both contain almost exclusively long-chain fatty acids ( 32–36 carbons ) ( Figure 1G , Figure 5A; M . Carvalho et al . , submitted ) . These observations raise questions about the source of the medium-chain fatty acids in Lpp DAG . One possibility is that they are derived from dietary fatty acids by processing mechanisms such as chain length shortening . Alternatively , they may be generated de novo from non-lipid dietary components such as sugars . To determine the contribution of dietary lipids to Lpp medium-chain DAG , we compared levels of hemolymph DAG in lipid-fed and lipid-starved animals . Lipid starvation increases the density of hemolymph Lpp ( Figure S9A ) . Thus , lipid-starved animals produce similar amounts of Lpp , but these particles contain less lipid . Furthermore , lipid starvation reduces the ratio of medium-chain DAG to polar lipids in Lpp by about 2-fold ( Figure S9B ) . However , lipid starvation affects Lpp density and DAG content more mildly than LTP RNAi , which reduces the ratio of DAG to polar lipid by about 3-fold . This raises the possibility that only part of the medium-chain DAG loaded onto Lpp by LTP is derived from dietary lipids . To explore the contribution of endogenous synthesis , we asked to what extent neutral lipid accumulation in the gut of Lpp RNAi larvae depended on dietary lipids . We knocked down Lpp in lipid-fed and lipid-starved larvae , and quantified neutral lipids in the gut . Although lipid starvation slightly reduces the amount of TAG and DAG in wild-type guts , these lipids accumulate dramatically when Lpp levels are reduced - both in lipid-fed and in lipid-starved larvae ( Figure 6A , 6B ) . On both diets , neutral lipid species containing medium-chain fatty acids increase most strongly in response to Lpp RNAi ( Figure S9C ) . This supports the idea that part of the medium-chain DAG present in Lpp derives from endogenous synthesis in the gut . Drosophila Fatty Acid Synthase ( FAS ) can synthesize medium-chain fatty acids [45] , raising the possibility that it generates fatty acids for the medium-chain DAG present in Lpp . To investigate the contribution of fatty acid synthesis in the gut to Lpp DAG , we knocked down FAS in this organ , and quantified hemolymph DAG . FAS RNAi decreases Lpp DAG by 30% ( Figure 6C ) . Therefore , even when fatty acids are supplied by the diet , a significant proportion of Lpp DAG contains fatty acids derived from endogenous synthesis in the gut . The insect fat body is a major site of lipid synthesis , storage and export [25] . How does the balance of lipid import and export affect the lipid composition of this organ ? We first examined the contribution of lipid delivery from the gut by blocking this transport route through RNAi-mediated LTP knock-down . Fat bodies of LTP RNAi animals contain much less sterol than those of wild-type , consistent with the sterol auxotrophy of Drosophila ( Figure S7A , S7C ) . Interestingly , LTP RNAi fat bodies also contain much less medium-chain DAG ( Figure 6D , 6E ) . Thus , the fat body does not maintain medium-chain DAG levels when export from the gut is blocked . To what extent is delivery of lipids from the gut required to build TAG stores in the fat body ? The wild-type fat body contains large amounts of TAG with predominantly long-chain fatty acids – unlike the gut , which contains similar amounts of medium-chain and long-chain TAG ( Figure 5C , Figure S6 ) . Removing lipids from the diet does not reduce the amount of TAG stored in the fat body ( Figure S9D ) . Thus , endogenous synthesis of fatty acids from non-lipid sources suffices to build TAG stores in the fat body . Impairing lipid delivery from the gut by LTP RNAi reduces fat body TAG storage by 30% ( Figure 6F , 6G ) but does not obviously alter the morphology of neutral lipid droplets ( Figure 6H ) . Interestingly , blocking both Lpp-dependent lipid import and export from the fat body using Lpp RNAi does not lower fat body TAG levels at all . Taken together , these data suggest that homeostatic mechanisms in the fat body , presumably involving endogenous lipid synthesis , can compensate for reduced lipid delivery from the gut . The fat body produces Lpp particles rich in long-chain PE species with a combined acyl chain length of 32 or 34 carbons , and 1 double bond ( PE 32∶1 and 34∶1; Figure S10A ) , which are also major phospholipid species in fat body membranes . We wondered how impaired Lpp production might affect the levels of these PE species . Loss of Lpp , but not LTP , increases the level of PE 32∶1 in the fat body about 1 . 5-fold ( Figure 6I ) . However , this is modest compared with the 5-fold increase in DAG in the gut that occurs upon lipoprotein knock-down . Since the gut expands its stores of medium-chain TAG when export of DAG is blocked , we considered the possibility that the fat body might increase its stores of long-chain TAG when PE export is blocked . However , this is not the case; Lpp RNAi does not significantly increase the amount of long-chain TAG in the fat body ( Figure 6G ) . These data indicate that homeostatic mechanisms in the fat body maintain TAG storage in a narrow range . Our previous work showed that Lpp is required for neutral lipid storage in the wing imaginal disc [29] . We wondered whether LTP-dependent mobilization of gut lipids onto Lpp contributed to the TAG and DAG stores of the wing disc . To address this , we quantified different species of neutral lipids in wild-type , Lpp RNAi and LTP RNAi wing discs . Wild-type wing disc cells contain medium-chain DAG , like that in Lpp , and also long-chain DAG ( Figure 5B , Figure S6 ) . Loss of either Lpp or LTP specifically depletes medium-chain DAG without affecting long-chain DAG ( Figure 7A , 7B ) . This suggests that a large fraction of the medium-chain DAG in wing disc cells is derived from the gut via LTP and Lpp . Total TAG levels in wing disc cells are reduced 60% by Lpp RNAi and 40% by LTP RNAi ( Figure 7C ) . Consistent with this , Lpp RNAi strongly reduces the size and number of neutral lipid droplets in the wing disc , while LTP RNAi has more modest effects ( Figure 7E ) . These data confirm that Lpp-mediated lipid delivery is needed for wing disc cells to store normal levels of TAG . They further indicate that mobilization of DAG from the gut via LTP and Lpp contributes to wing disc TAG stores . However , other Lpp lipids partially support wing disc TAG storage , at least when gut lipid mobilization is blocked . Medium-chain TAG species are reduced equally by Lpp and LTP RNAi , whereas the effects of LTP RNAi on long-chain TAG species is weaker ( Figure 7D ) . This raises the possibility that Lpp long-chain PE , which does not decrease upon LTP RNAi , contributes fatty acids to long-chain TAG stored in wing disc cells . The Drosophila brain is shielded from the circulation by a blood-brain barrier similar to that of mammals . Nevertheless , Lpp crosses this barrier [46] , raising the possibility that cells in the brain may derive lipids from the circulation . To investigate this , we analyzed lipid profiles in the brains of wild-type , LTP RNAi and Lpp RNAi animals . Lpp RNAi decreases medium-chain DAG , but not long-chain DAG in the brain ( Figure 7F , 7G ) , and reduces total brain TAG by 60% ( Figure 7H , 7I ) . In contrast , LTP RNAi does not affect neutral lipids in brain cells significantly . Thus , normal TAG storage in the brain requires Lpp-mediated lipid delivery , but it is not limited by the LTP-dependent loading of Lpp with DAG in the gut . To what extent do different tissues synthesize their own membrane lipids ? Do they also rely on Lpp to deliver some membrane lipids ? Because Drosophila cannot synthesize sterols , it is unsurprising that sterol accumulation in peripheral tissues depends on Lpp and LTP ( Figure S7A ) [18] . To investigate whether Lpp was important for delivery of other membrane lipids , we quantified the polar lipid composition of tissues from LTP and Lpp RNAi animals . Wing disc and gut from Lpp RNAi animals contain about 20% less PE 32∶1 and PE 34∶1 than those of wild-type ( Figure 7J , Figure S10B ) . Interestingly , these PE species are not only major membrane constituents , but are also precisely those species that are most abundant in Lpp ( Figure S10A ) . In contrast , we did not observed a reduction in any species of PC , which is abundant in membranes , but only a minor component of Lpp ( Figure S10C ) . These observations suggest that wing disc and gut cells cannot completely compensate for the loss of Lpp-derived PE species by increasing endogenous PE synthesis . They further raise the possibility that Lpp PE species might be directly incorporated into cell membranes without remodeling . In contrast to wing disc and gut , the brain can maintain normal levels of all phospholipids including PE 32∶1 and PE 34∶1 even when Lpp levels are strongly reduced by RNAi ( Figure 7K , Figure S10C , S10D ) . Thus , while the brain requires Lpp delivery to store normal levels of TAG , its phospholipid composition is autonomously controlled .
How disturbed lipoprotein metabolism affects lipid composition in individual organs is insufficiently understood . Drosophila could provide a useful model in which to study this problem , but its lipoproteins had not been well characterized . Here , we outline the basic features of the lipoprotein metabolism of Drosophila larvae , and its relevance for tissue-specific fat storage and membrane lipid composition ( Figure 8 ) . The major inter-organ lipid transport routes in Drosophila are executed by a single lipoprotein – Lpp , which is scaffolded by the apoB homologue apoLpp . Its major polar lipid constituents are long-chain PE and sterols , and its major neutral lipid is medium-chain DAG . Lpp lipidation takes place in two consecutive steps , which require distinct lipid transfer proteins , MTP and LTP , and take place in different organs – fat body and gut . ApoLpp is translated and lipidated in the fat body by an MTP-dependent mechanism , resulting in the formation of dense Lpp particles rich in PE . These are recruited to the gut , where they are further loaded with DAG and sterols through the activity of LTP . Thus , although Lpp originates in the fat body , it is loaded both with fat body and gut lipids . Lipidation of mammalian apoB , like that of Drosophila apoLpp , proceeds in two distinct steps , formation of primordial phospholipid-rich lipoprotein particles , and subsequently acquisition of bulk neutral lipid [2] . However , this process occurs entirely in the secretory pathway of producing cells . MTP has been proposed to be required both for initial transfer of phospholipids , and for the recruitment of TAG to the ER lumen for incorporation into lipoproteins [47] , [48] , [49] , [50] . Interestingly , Drosophila MTP has been shown to promote the secretion of apoB-containing lipoproteins from COS cells , and to transfer phospholipids , but not TAG , between liposomes [31] , [51] . This suggested that MTP acquired the ability to transfer TAG in the vertebrate lineage . Our experiments show that Drosophila MTP is required for the production of the two Drosophila apoB-family lipoproteins Lpp and LTP in vivo; they further show that MTP is insufficient to load Lpp with normal quantities of DAG , the major neutral lipid of Lpp . These data support the idea that MTP originally evolved to promote the assembly of phospholipid-rich apoB-family lipoproteins . The novel Drosophila apoB-family lipoprotein LTP shares many properties with the Lipid Transfer Particle purified from the hemolymph of several insects , including Manduca and Locusta [35] , [36] . The scaffolding proteins of Drosophila LTP , apoLTPI and apoLTPII , are generated from a single precursor , apoLTP . Orthologous apoB-family proteins of other insects are therefore plausible candidates for the scaffolding proteins of their LTP particles . Insect LTPs were shown to contain a third , small protein subunit , apoLTPIII [36] , [52] . Our biochemical experiments do not address whether Drosophila LTP might contain an apoLTPIII subunit , because LTP is of such low abundance that silver staining barely detects the much larger apoLTPI . Sequence analysis of apoLTP does not suggest the existence of a protease cleavage site that could give rise to a protein of the size of apoLTPIII , and neither apoLTPI nor apoLTPII antibodies detect an additional protein of this size . Thus , if apoLTPIII exists in Drosophila , it is not likely to be derived from the apoLTP precursor . The function of LTP as a lipid transfer protein rather than a carrier of bulk hemolymph lipid uncovers surprising evolutionary plasticity of the apoB lipoprotein family . Insect LTPs have been studied in vitro in a wide range of systems [38] . In different contexts , they have been shown to facilitate the exchange of DAG and phospholipids between Lpp and fat body or gut [26] , [27] , [28] , and even between insect and human lipoproteins of different densities [52] , [53] , [54] . Our studies of feeding Drosophila larvae indicate that only a subset of the lipid transfer activities of LTP may be relevant under specific metabolic conditions in vivo . LTP moves DAG and sterols from the larval gut onto Lpp . However , it does not facilitate significant net transfer of fat body lipids to Lpp . Consistent with this , radiolabeling experiments showed that the rate of DAG transfer from larval Manduca fat body to Lpp exceeds the rate of the reverse process [27] . This may reflect a dominance of nutritional lipid uptake and fat storage in feeding larvae . Although we have been unable to identify a Drosophila HDL-like lipoprotein , we note that LTP and Lpp share some functional features with mammalian HDL , despite being scaffolded by unrelated apolipoproteins . Together , Lpp and LTP mediate efflux of sterols from the gut to circulation . Conceivably , other tissues that recruited both lipoproteins might efflux sterol for reverse transport . While it is clear that dietary lipids do contribute to Lpp DAG , the gut does not directly incorporate dietary fatty acids into DAG destined for export . The long-chain fatty acids that predominate in the diet strikingly differ from the medium-chain fatty acids in Lpp DAG ( M . Carvalho et al . , submitted ) . A possible explanation is that the gut remodels dietary fatty acids , conceivably via limited β-oxidation . Interestingly , the gut is also a lipogenic organ and a significant fraction of the medium-chain fatty acids found in Lpp DAG derives from de novo fatty acid synthesis in this organ . In more primitive animals , such as Caenorhabditis elegans , lipid uptake , storage and lipogenesis all occur in the gut [55] . More complex animals , including Drosophila , have developed separate organ systems for lipid storage and lipogenesis . However , our data show that this separation of functions is not absolute in the fly . Rather , other nutrients such as amino acids or sugars might be partially converted to lipid by the gut , instead of being transported intact into circulation . It would be interesting to ask what circumstances favor this conversion . Intriguingly , de novo lipogenesis has been observed in the mammalian gut , especially under conditions of insulin resistance , and has been proposed to contribute to the postprandial dyslipidemia observed in this state [56] . Drosophila may be a useful model to explore this problem . Gut and fat body differ in how they respond to blockage of lipid export to Lpp . Enterocytes vastly and rapidly expand their normally moderate stores of medium-chain DAG and TAG . This occurs even in the absence of dietary lipids , when exported lipids are derived from endogenous fatty acid synthesis . Thus , the gut has a flexible capacity for lipid storage . In contrast , the larval fat body maintains its neutral lipid stores within tight limits . When lipoprotein transport is blocked , endogenous lipid synthesis from other dietary components may suffice to build the large TAG stores of this organ . Furthermore , even though the fat body normally supplies the entire animal with large amounts of lipoproteins , TAG stores hardly increase when Lpp is not produced . Homeostatic mechanisms must maintain fat body TAG levels . In this way , the fat body differs from the gut , which accumulates fat when lipoprotein export is blocked , similar to mammalian gut and liver [5] . Peripheral tissues cannot maintain normal TAG levels in the absence of Lpp . The wing disc depends on Lpp for a large fraction of its fat stores . Interestingly , our work indicates that lipid delivery from the fat body and gut differently contributes to wing disc neutral lipids . TAG species containing medium-chain fatty acids depends on LTP and Lpp-mediated DAG mobilization from the gut . TAG species containing long-chain fatty acids also depend on Lpp-mediated lipid delivery , but are less affected by a blockage of DAG export from the gut . As Lpp is produced in the fat body , this suggests that long-chain TAG in wing discs may be derived from lipids supplied by the fat body . The most abundant source of long-chain fatty acids in Lpp is PE , which raises the possibility that wing discs use Lpp phospholipids to build cellular fat stores . Consistent with this , cultured murine hepatocytes convert a significant fraction of LDL or HDL-derived PC to TAG [57] , [58] , although the in vivo relevance of this pathway remains to be explored . However , Lpp still contains reduced amounts of medium-chain DAG when LTP-mediated lipid loading is impaired . Thus , long-chain fatty acids in wing disc TAG might also derive from elongation of medium-chain fatty acids . Interestingly , although medium-chain DAG is the most abundant lipid transported through circulation , tissues store only minor amounts of neutral lipid containing medium-chain fatty acids . This would be consistent with the idea that tissues either elongate these fatty acids or subject them to β-oxidation . The brain also requires Lpp-mediated lipid delivery to build its TAG stores . Interestingly , the brain stores normal levels of TAG when gut lipid mobilization is inhibited . While this does not exclude the possibility that the brain may directly acquire lipids from the gut under normal conditions , it indicates that TAG levels in this organ are more resistant to fluctuations in nutritional conditions than those in the wing disc . In addition to providing fatty acids for neutral lipid storage , lipoproteins also influence the phospholipid composition of wing disc and gut: Lpp knock-down specifically reduces those PE species that are most abundant in Lpp . This suggests that Lpp might directly deliver PE to the cellular membranes of wing disc and gut . It further raises the possibility that phospholipid synthesis in other tissues could have organism-wide effects on membrane lipid composition . Since PE-rich Lpp particles are assembled in the fat body , this tissue is a likely source of these lipids . However , the brain does not depend on Lpp to maintain its normal membrane phospholipid composition . Furthermore , our previous work suggested that the brain is more resistant to sterol depletion than other tissues [59] . In general , these data indicate that the lipid composition of the brain is more tightly and autonomously controlled than that of other tissues . In mammals , cellular lipid synthesis and lipid supply from circulation are coordinated through the SREBP pathway [60] , [61] . Since Drosophila SREBP is regulated by PE instead of sterols [62] , it will be interesting to explore whether altered PE levels in Lpp-deprived wing discs might activate SREBP signaling and increase lipid synthesis or lipoprotein uptake . If true , coordination of cellular lipid synthesis with lipid supply through lipoproteins is an evolutionarily conserved function of the SREBP pathway . Lipoproteins transport large amounts of lipids through circulation – including many of the polar and neutral lipid species present in cells . Our data indicate that in Drosophila , individual organs utilize lipoprotein-derived lipids not only for fat storage but also for membrane homeostasis . ApoB-deficient human patients , and patients with dyslipidemia suffer from various abnormalities in peripheral tissues . Our data suggest that it may be worthwhile to examine how these perturbations alter the membrane lipid composition of affected tissues .
Flies were raised on food containing yeast , yeast extract , soy peptone , sucrose , and fructose . For lipid starvation experiments , larvae were fed lipid-depleted food , supplemented with sterols [59] . Unless otherwise stated , feeding late third instar larvae were used for experiments . ApoLII and apoLI antibodies were described previously [20] , [29] . Mouse anti-α-tubulin was provided by Sigma . ApoLTPII , apoLTPI and MTP antibodies were raised in rabbit , Cv-d antibodies in guinea pig . Inducible RNAi lines were crossed with lines harboring heat-shock-inducible flippase and the GAL4 driver . RNAi was initiated by heat shock during early larval development . Controls are stage-matched larvae from the same cross lacking the GAL4 driver . RNAi was driven with Adh-GAL4 , which is mostly active in the fat body . In experiments addressing the contribution of the fat body to hemolymph lipoprotein levels , RNAi was driven with Lpp-GAL4 , which is exclusively active in the fat body during larval stages . RNAi driven with either Adh-GAL4 or Lpp-GAL4 reduced hemolymph Lpp and LTP to the same extent . Larvae were bled in PBS . Hemocytes and cell fragments were removed by centrifugation for 30 min at 1500 g and subsequently for 30 min at 16 000 g . Note that stronger centrifugation pellets a large fraction of hemolymph lipoproteins . Isopycnic centrifugation in KBr gradients was essentially performed as described [63] . Immunostaining and Nile red staining of tissues was performed as described [20] , [29] . For co-staining with antibodies , lipid droplets were visualized with BODIPY 493/503 ( Invitrogen ) . Explanted guts from early 3rd instar LTP RNAi larvae were incubated for 4 h at 25°C with hemolymph prepared from LTP RNAi larvae or lipid-starved larvae , diluted in Grace's insect medium . Subsequently , lipoprotein density was determined by isopycnic centrifugation and immunoblotting . A dominant negative allele of dynamin ( shibire K44A ) was expressed in enterocytes in a time-controlled manner with MyoIA-GAL4 , Tubulin-GAL80TS . Lipids from hemolymph and tissue homogenates were extracted and analyzed by shotgun mass spectrometry in positive ion mode as described in [59] . Sterols were quantified according to [64] . For supporting mass spectrometry data , see Figure S11 . For more detailed protocols , fly strains , and the generation of mtp and apoltp mutants , RNAi transgenes and antibodies see Text S1 .
|
Lipoproteins transport both dietary and endogenously synthesized lipids between different organs . Lipoprotein dysfunction is associated with many medical disorders , including cardiovascular disease , but the mechanisms underlying pathogenesis are unclear . Simple animal models would be valuable , therefore , to understand basic functions of lipoprotein and their influence on tissue lipids . Here , we develop the fruit fly Drosophila melanogaster as a genetically tractable model to study lipoprotein metabolism . We characterize the major Drosophila lipoproteins , the lipids they transport through circulation , and the mechanisms by which they acquire lipid cargo from different organs . By genetically blocking specific inter-organ lipid transport routes , we uncover surprising tissue-specific differences in lipoprotein lipid utilization . Lipoproteins deliver lipids from fat body and gut to wing imaginal discs , which utilize them to build their fat stores . In addition , lipoproteins provide a significant fraction of membrane phospholipids to wing disc and gut cells . In contrast , fat storage in the brain does not require lipoprotein-mediated delivery of lipids from the gut , and the brain phospholipid composition can be maintained independently of lipoproteins . Our studies define basic features of Drosophila lipoprotein metabolism and suggest novel mechanisms for how lipoproteins might affect animal tissue function in general .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"lipoprotein",
"metabolism",
"animal",
"genetics",
"gene",
"function",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"lipid",
"classes",
"lipid",
"metabolism",
"tissue",
"distribution",
"lipids",
"proteins",
"lipoproteins",
"biology",
"biochemistry",
"genetics",
"apolipoprotein",
"genes",
"metabolism",
"genetics",
"and",
"genomics"
] |
2012
|
Lipoproteins in Drosophila melanogaster—Assembly, Function, and Influence on Tissue Lipid Composition
|
Young children who contract Ebola Virus Disease ( EVD ) have a high case fatality rate , but their sources of infection and the role of breastfeeding are unclear . Household members of EVD survivors from the Kerry Town Ebola Treatment Centre in Sierra Leone were interviewed four to 10 months after discharge to establish exposure levels for all members of the household , whether or not they became ill , and including those who died . We analysed a cohort of children under three years to examine associations between maternal illness , survival and breastfeeding , and the child’s outcome . Of 77 children aged zero to two years in the households we surveyed , 43% contracted EVD . 64 children and mothers could be linked: 25/40 ( 63% ) of those whose mother had EVD developed EVD , compared to 2/24 ( 8% ) whose mother did not have EVD , relative risk adjusted for age , sex and other exposures ( aRR ) 7·6 , 95%CI 2·0–29·1 . Among those with mothers with EVD , the risk of EVD in the child was higher if the mother died ( aRR 1·5 , 0·99–2·4 ) , but there was no increased risk associated with breast-feeding ( aRR 0·75 , 0·46–1·2 ) . Excluding those breastfed by infected mothers , half ( 11/22 ) of the children with direct contact with EVD cases with wet symptoms ( diarrhoea , vomiting or haemorrhage ) remained well . This is the largest study of mother-child pairs with EVD to date , and the first attempt at assessing excess risk from breastfeeding . For young children the key exposure associated with contracting EVD was mother’s illness with EVD , with a higher risk if the mother died . Breast feeding did not confer any additional risk in this study but high risk from proximity to a sick mother supports WHO recommendations for separation . This study also found that many children did not become ill despite high exposures .
Young children experience a high case fatality rate from Ebola , but the incidence of Ebola Virus Disease ( EVD ) in children appears to be lower than in adults . [1–4] Young children may have limited exposure outside the home , but within the household maintaining hygiene in young children is difficult , although efforts may be made to keep children away from those who are sick . For very young children who need to be fed and held , contact with sick caregivers may be unavoidable . Breastfeeding is a possible additional source of infection for young children: Ebola has been found in breast milk , but the risk to breastfed babies and the contribution of breastfeeding to transmission is poorly understood . [5 , 6] An investigation of household contacts following the Ebola outbreak in Gulu , Uganda in 2000 included five infants whose mother had EVD: three of four infants who were breastfed developed EVD . [7] The other infant was reported to have been separated from his mother early in the course of her illness and remained well; it is not clear if he was breastfed . Two recent systematic reviews of transmission of Ebola did not did not mention risks associated with breastfeeding . [8 , 9] As part of a study of transmission patterns in Sierra Leone we collected data on exposure patterns and outcomes of all individuals present in the households of EVD survivors . In this analysis we sought to identify likely sources of infection and characterise risk of transmission to young children , including those breastfed by mothers with EVD .
The study was approved by the Sierra Leone Ethics and Scientific Review Committee and the Ethics Committee of the London School of Hygiene & Tropical Medicine . At the interview , individual written informed consent to participate in the study was sought from all adults , and from parents or guardians for children ( < 18 years ) , with assent from children of 12 years or older .
One hundred and fifty one survivors were discharged from Kerry Town Ebola Treatment Centre from November 2014 through March 2015 , of whom 138 were still living in the Western Area of Sierra Leone when sought for interview in July-September 2015 . Twelve were uncontactable and a further two were known to have bad relationships with their households so were not approached . We contacted and interviewed 123 Kerry Town survivors , living in 94 households . Only one contacted survivor refused to be interviewed , and only two of 526 household members refused to participate . A further 37 members were not available to attend the interview . Some households also included survivors who had been treated in other facilities . The households contained 77 children aged less than three years: 43% ( 33/77 ) got EVD , including four who fitted the case definition but were not diagnosed at the time . The risk of EVD was 54% ( 13/24 ) in those under one year; 40% ( 12/30 ) in those aged one year and 35% ( 8/23 ) in those aged two years ( p-value for trend = 0 . 2 ) . The risk was slightly higher in males than in females: 51·4% ( 18/35 ) vs 35·7% ( 15/42 ) , p = 0 . 2 . Three of the children were primary or co-primary cases in their household . Overall , 24 children under three years died of EVD , giving a case fatality rate of 73%: 85% ( 11/13 ) , 75% ( 9/12 ) and 50% ( 4/8 ) at ages under-one , one , and two respectively ( p-value for trend = 0·1 ) . Among the 77 children were 13 whose mothers were not present ( including two mothers who had died in other households ) , or were not clearly identified: six ( 46% ) of these children developed EVD and five died compared to 27 cases ( 42% ) and 19 deaths among the 64 children who could be linked to their mothers . Details of the mother-child pairs for whom the outcome of both mother and child are known are shown in Table 1 for the 40 whose mothers had EVD , in Table 2 for the 24 whose mothers had no symptoms , and in summary for all 64 in Table 3 . The highest level of exposure is shown , in terms of direct or indirect exposure to those with EVD in the home or outside . None of the children had direct contact with dead bodies . Breastfeeding was taken as the highest exposure if the mother had EVD unless the child developed symptoms before or at the same time as the mother . EVD in the children was much more likely among those whose mother had EVD ( 25/40 , 63% ) than among those whose mother did not get EVD ( 2/24 , 8% , risk ratio ( RR ) 7·5 , 95% confidence interval ( CI ) 1 . 9–28 . 9 , p<0 . 0001 , Table 3 ) . The RR remained high after adjusting for age and sex of the child ( RR 9·4 , 95% CI 2·6–34·0 ) , and after additionally adjusting for maximum exposure level ( RR 7·6 , 95%CI 2·0–29·1 ) . Household crowding and sanitation were not associated with EVD in the child , and adjusting for them made little difference to the results . After adjusting for mother’s EVD status and exposure levels , the risk of EVD in the child decreased with age ( Table 3 ) . After adjusting for mother’s EVD , age , and sex , there was no effect of exposure level . Among those whose mother had EVD , excluding the two pairs in which the children were ill first , the risk of EVD in the child was higher if the mother died ( 79% vs 50% , Table 3 ) , giving a relative risk of 1·6 ( 95% CI 0·97–2·6 ) . This association was similar after adjusting for the child’s age and sex and additionally for exposure level . Of the 13 children who did not get EVD whose mother survived , five had contact with the mother when she was a wet case and five only when she was a dry case ( unknown for three ) . As the only child over two years who was breastfed got ill at the same time as the mother and was therefore excluded , the analysis of breastfeeding was restricted to the under two’s . We also excluded the child who became ill first ( Table 1 ) , leaving 26 children . The proportion of children with EVD was very similar in those who were or were not breast fed ( 69% vs 70% , Table 3 ) , RR 0·98 , 0·58–1·7 . There was no evidence of increased risk from breastfeeding after adjusting for age and sex ( RR 0·76 , 0·46–1·2 ) or for whether the mother died ( Table 3 ) . The analyses were re-run excluding the six mother-child pairs for which either the mother or the child was classified as having EVD on the basis of symptoms ( Table 1 ) . The associations with having a mother with EVD ( fully adjusted RR 6 . 5 , 1·6–26·0 ) and with breastfeeding ( fully adjusted RR 0·74 ( 0·47–1·2 ) were similar to the main analysis , but the effect of having a mother who died of Ebola was lost ( fully adjusted RR 1·3 , 0·76–2·1 ) . The analyses were also rerun using Poisson regression . The results were similar to the main analysis . Among the children under three years whose mother did not get EVD , only two children got EVD . Both were aged under one year , from households with many EVD cases ( Table 2 ) , and both were reported to have had close contact with wet cases in the household . Seven other children whose mother did not have EVD and 4 whose mother had EVD but were not breastfed , had direct contact with wet cases and did not get ill . Overall , excluding children breastfed by mothers with EVD , half ( 11/22 ) of the children who had direct contact with wet cases or fluids remained well . These contacts included sharing beds with and embracing close relatives who suffered from vomiting and/or diarrhoea .
Among the very young children in this study the risk of EVD depended largely on whether their mother developed EVD , with an additional risk for those whose mothers died of Ebola . The high risk in those with sick mothers is expected , and the higher risk in those with mothers who died may reflect higher viral loads and/or viral shedding in these mothers . The low risk in children in Ebola-affected households when the mother was not ill is surprising , and cannot all be explained by low exposure in the children . Overall , nearly two thirds of under-three year olds had direct contact with wet cases in the household or their body fluids . While the risk of disease decreased with decreasing exposure , half of the young children with direct exposure to wet cases remained well . Only three children were deliberately sent out of the household to reduce exposure , and for all three there was some exposure before they left . The opportunities for households to protect children from exposure are limited , particularly as more and more cases arise , and young children share beds with sick relatives . While a ‘no touch policy’ may be understood by older children , it is impossible to explain to an infant . Among children whose mothers had EVD , being breastfed did not appear to increase the risk . Numbers were small and risks were already high in this group so there was limited power to detect an association . Current WHO guidelines recommend that asymptomatic breastfed infants of Ebola-infected mothers should be separated from their mothers and replacement fed . [15] Although we found no excess risk from breastfeeding , further studies , ideally with larger , pooled datasets , are needed to assess this further before suggesting any changes to the recommendation . The high risk from proximity to a sick mother supports the need for separation . The children in this study all came from households with at least one survivor . This may mean small households and households with fewer cases are underrepresented , as there would be a lower chance for small households to include a survivor , and households in which all cases of EVD died are missed . This might underestimate the case fatality rate and overestimate attack rates , but should not bias the relative risks by age and exposure . This study shows the remarkable resilience of some young children despite apparent exposure to Ebola . This could be dose-related—we do not know the actual viral exposure through contact or breastfeeding—but in other contexts some people seem to be infected from minimal exposures . Relative resistance to Ebola could be influenced by genetic factors , [16] though the correlation between infections in mothers and children is more likely to reflect exposure patterns than shared genes . It is possible that there is some protection through maternal antibody from breastfeeding ( perhaps more in mothers who survive ) that counteracts any increased risk from transmission via breastmilk . This is much the largest study of mother-child pairs with EVD to date , and the first attempt to assess any excess risk from breastfeeding . By visiting households after transmission had ceased and talking to all members we were able to determine exposure in much more detail than is possible in an acute epidemic situation . And because we included all children in these households , including those who were not sick , we have been able to calculate age and exposure-specific attack rates . In these households the risk to young children was largely dependent on whether their mother had EVD , regardless of whether they were breastfed .
|
Our study is the first to quantify sources of infection and describe risk of transmission of Ebola to young children . We found that the risk of a child under three developing Ebola disease was low unless their mother had EVD , and that the risk was particularly high if their mother died of EVD . But we found no additional risk from breastfeeding . WHO recommends separating asymptomatic breast-fed infants from their mothers if they develop Ebola , and using formula feeding . We support the need for separation because of the high risk related to proximity , but more research is needed to more fully understand this , particularly given the importance of breast-feeding in preventing other childhood illnesses . We also found young children in Ebola-affected households whose mothers were not ill had a surprisingly low risk of developing EVD which was not all explained by low exposure to the virus . Many children stayed well despite having direct contact with EVD patients with diarrhoea , vomiting or bleeding who are considered the most infectious . We hope these findings will provide impetus for more detailed studies into age-related response to the Ebola virus .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"neonatology",
"children",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"maternal",
"health",
"vomiting",
"pathogens",
"tropical",
"diseases",
"microbiology",
"ebola",
"hemorrhagic",
"fever",
"pediatrics",
"viruses",
"diarrhea",
"age",
"groups",
"physiological",
"processes",
"filoviruses",
"women's",
"health",
"signs",
"and",
"symptoms",
"rna",
"viruses",
"gastroenterology",
"and",
"hepatology",
"infants",
"neglected",
"tropical",
"diseases",
"families",
"viral",
"hemorrhagic",
"fevers",
"infectious",
"diseases",
"medical",
"microbiology",
"microbial",
"pathogens",
"people",
"and",
"places",
"breast",
"feeding",
"diagnostic",
"medicine",
"ebola",
"virus",
"anatomy",
"physiology",
"viral",
"pathogens",
"population",
"groupings",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"hemorrhagic",
"fever",
"viruses",
"organisms"
] |
2016
|
Effects of Mother’s Illness and Breastfeeding on Risk of Ebola Virus Disease in a Cohort of Very Young Children
|
Twilight is characterised by changes in both quantity ( “irradiance” ) and quality ( “colour” ) of light . Animals use the variation in irradiance to adjust their internal circadian clocks , aligning their behaviour and physiology with the solar cycle . However , it is currently unknown whether changes in colour also contribute to this entrainment process . Using environmental measurements , we show here that mammalian blue–yellow colour discrimination provides a more reliable method of tracking twilight progression than simply measuring irradiance . We next use electrophysiological recordings to demonstrate that neurons in the mouse suprachiasmatic circadian clock display the cone-dependent spectral opponency required to make use of this information . Thus , our data show that some clock neurons are highly sensitive to changes in spectral composition occurring over twilight and that this input dictates their response to changes in irradiance . Finally , using mice housed under photoperiods with simulated dawn/dusk transitions , we confirm that spectral changes occurring during twilight are required for appropriate circadian alignment under natural conditions . Together , these data reveal a new sensory mechanism for telling time of day that would be available to any mammalian species capable of chromatic vision .
The ability to predict and adapt to recurring events in the environment is fundamental to survival . Organisms across the living world achieve this using endogenous circadian clocks [1–3] . However , if such clocks are to fulfil their ethological function they need to be regularly reset to local time . This is achieved by sensory inputs that report changes in the physical environment providing a useful proxy for time of day . By far the best characterised of these input pathways is that recording the diurnal change in the overall quantity of light reaching the earth’s surface ( irradiance ) . In the case of mammals , a dedicated retino-hypothalamic projection brings this visual information to the brain’s “master” clock in the suprachiasmatic nuclei ( SCN ) [4–7] . The retino-hypothalamic projection is formed by a unique class of retinal ganglion cells ( RGCs ) , which are intrinsically photosensitive thanks to their expression of melanopsin [4] . Although these so-called ipRGCs can therefore entrain the clock even in the absence of the conventional rod and cone photoreceptors , all photoreceptor classes are capable of influencing the clock in intact animals [8–13] . This arrangement has previously been considered only insofar as it allows the clock to respond to changes in irradiance . However , the inclusion of cones in this pathway allows for the possibility that the clock could also receive information about changes in the spectral composition of light ( colour ) [14] . There has previously been speculation that such colour signals could provide a reliable method of telling time of day [3 , 15] but , to date , there has been no direct test of that possibility in mammals . The ability to discriminate colour relies on comparing the relative activation of photopigments with divergent spectral sensitivities . In mammals , this task is achieved via differential processing of cone photoreceptor signals in the retina [16] . At least 90% of mammalian species are believed capable of this form of colour discrimination [17] which , with the exception of Old World primates , allows for dichromatic vision . Thus , most mammals express just two distinct classes of cone opsin , one maximally sensitive to short wavelengths ( ultraviolet–blue ) and a second with peak sensitivity to longer wavelengths ( green–red ) [18] . Here we show that , in mice , this primordial colour discrimination axis ( equivalent to human blue–yellow colour vision ) is an influential regulator of SCN activity , essential for appropriate circadian timing relative to the natural solar cycle .
We first set out to determine whether changes in spectral composition associated with the earth’s rotation could provide reliable information about solar angle that the circadian clock could use to estimate time of day . To this end , we obtained high resolution measurements of natural variations in spectral irradiance across multiple days ( Manchester , August–October 2005 , n = 36 d ) . As expected , these measurements revealed highly predictable changes in both irradiance and spectral composition as a function of solar angle ( Fig 1A and 1B ) . In particular , we observed a progressive enrichment of short-wavelength light across negative solar angles: a result of the increasing amount of ozone absorption and consequent Chappuis band filtering of green–yellow light when the sun was below the horizon [14] . We next calculated the extent to which this change in spectral composition was detectable to the mammalian visual system . Taking the mouse as a representative species , we employed previously validated approaches to quantify the relative excitation of its ultraviolet and medium wavelength sensitive ( UVS/MWS ) cone opsins [12 , 19 , 20] . This analysis revealed a robust change in the ratio of excitation between the two pigments ( Fig 1C ) that would constitute substantial changes in apparent colour along the blue–yellow axis . These changes were restricted to the twilight transition , with the UVS:MWS ratio fairly invariant throughout the day , indicating that measuring the change in colour could provide a useful method of tracking the progression of dawn and dusk . To ascertain how reliably this blue–yellow colour signal alone could be used to estimate phase of twilight , relative to simple measures of irradiance , we next compared the day-to-day variability of colour and irradiance measurements across our dataset ( Fig 1D ) . Surprisingly , we found that colour was in fact more predictive of sun position across twilight ( -7 to 0° below horizon ) than was irradiance ( 78 . 5 ± 0 . 1% versus 75 . 8 ± 0 . 1% of variance explained by solar angle; mean ± SD ) . Accordingly , for any fixed solar angle , the range of observed colour values was considerably more tightly clustered than those for irradiance . These observations most likely reflect the fact that cloud cover can change overall brightness quite dramatically , but exerts only relatively minor effects on spectral composition . Importantly , then , measuring colour could provide a more reliable estimate of the approach of night or day than measuring irradiance . However , while the mammalian circadian clock is certainly known to respond to diurnal variations in irradiance [8–13] , there has been no investigation of whether the SCN also receives colour signals . Accordingly , we next asked whether the central clock showed electrophysiological responses to changes in colour by recording extracellular activity in the mouse SCN . In order to identify colour-sensitive cells , we set out to generate test stimuli which differentially modulated the UVS and MWS mouse cone opsins . While , in principle , producing such stimuli is straightforward , the close spectral sensitivity of mouse opsins makes it difficult to achieve this aim without concomitant changes in the activation of rods and/or melanopsin . To circumvent this problem , we employed a well-validated transgenic model in which the native mouse MWS opsin is replaced by the human long-wavelength sensitive ( LWS ) opsin ( Opn1mwR; [8 , 12 , 19] ) . Cones in these animals develop and function normally , with LWS opsin expression entirely recapitulating that of the native MWS opsin [21] . Importantly , however , the resultant shift in cone spectral sensitivity in Opn1mwR mice facilitates the generation of stimuli that provide selective modulation of individual opsin classes [22] . Using this Opn1mwR model , we first established a background lighting condition ( using a three-primary LED system ) , whose spectral composition recreated a wild-type mouse’s experience of natural daylight ( S1A Fig ) . We next designed a set of manipulations of this background spectrum that allowed us to modulate excitation of one or both cone opsins without any concomitant change in rod or melanopsin activation . Under these conditions , we were then able to unambiguously distinguish colour-sensitive neurons based on the following criteria: ( 1 ) the presence of larger responses to chromatic versus achromatic changes in cone excitation and ( 2 ) responses of opposite sign to selective activation of UVS and LWS opsin in isolation ( i . e . , excitatory/ON versus inhibitory/OFF ) . To achieve the largest possible change in colour , we started by selectively modulating UVS and LWS cone opsin excitation in antiphase ( “colour”; S1B Fig ) . We then compared responses to this stimulus with those elicited by one in which the change in UVS and LWS opsin activation occurred in unison ( “brightness”; S1B Fig ) . Any spectrally opponent cells should be more responsive to the “colour” as opposed to “brightness” condition . We found that 17/43 SCN units ( from 15 mice ) that responded to these stimuli showed a significant preference for the pair in which UVS and LWS activation was modulated in antiphase ( Fig 2A and 2B; paired t test , p<0 . 05 , n = 17 ) . As there were an additional 26 visually responsive units that did not respond to either of these analytical stimuli ( paired t tests , p>0 . 05 ) , these data indicate that at least one quarter of light-responsive SCN neurones show chromatic opponency . Interestingly , we found that cone inputs exerted a much more powerful influence over the firing activity of cells exhibiting a preference for chromatic stimuli relative to than achromatic cells ( Fig 2B; absolute change for responses of chromatic cells = 8 . 1 ± 2 . 3 spikes/s versus 1 . 9 ± 0 . 5 spikes/s for achromatic cells , n = 17 and 26 respectively; t test: p<0 . 01 ) . Moreover , we found that the spiking activity for the majority ( 13/17 ) of colour-sensitive SCN neurons was highest during the stimulus phase biased towards UVS opsin activation ( Fig 2A ) , and that these cells exhibited especially robust and sustained changes in firing ( Fig 2A and 2B ) . Our data above therefore indicate that cone inputs constitute a dominant influence on the firing activity of colour-sensitive SCN neurons and that most of these cells exhibit blue-ON/yellow-OFF colour opponency . We confirmed this by selectively modulating brightness for each of these cone opsins independently ( stimulus shown in S2A Fig ) ; as expected , these cells reliably increased firing in response to selective increases in UVS opsin activation and decreased firing following increases in LWS opsin activation ( Fig 2C ) . Conversely , the remaining colour-sensitive cells exhibited the opposite preference ( yellow-ON/blue-OFF; Fig 2C ) . An aspect of mouse retinal organisation that poses a challenge to colour vision is that most cones in this species co-express UVS and MWS opsin [23] . The exceptions are rare “primordial S-cones” that only express UVS opsin [24] and peripheral cones that may express either pigment alone [25] . One might expect that chromatic opponency would rely on comparisons between these rare single pigment cones . If this were the case , then responses to LWS- and UVS-specific stimuli of defined contrast should be insensitive to changes in the spectral composition of the background light . In fact , we found that this was not the case ( S2A–S2C Fig ) , indicating involvement of the more common opsin co-expressing cones in the chromatic responses of SCN neurons . By contrast with chromatic SCN neurons , none of the cells identified as achromatic exhibited any overt OFF response to selective activation of either UVS or LWS cone opsin . Instead , these achromatic cells exhibited pure ON responses to stimuli targeting one or both opsin classes , such that on average the population showed little bias towards UVS/LWS opsin-driven responses under background spectra resembling natural daylight ( Fig 2C ) . Adjusting the background spectra to equalise basal activation of the two cone opsins skewed responses in favour of UVS opsin , however ( S2D Fig ) , consistent with previous suggestions that ipRGCs are relatively enriched in the UVS opsin-biased dorsal retina [26] . We next asked whether colour opponent cells also received irradiance information from the melanopsin-expressing ipRGCs that dominate retinal input to the SCN [4 , 6 , 7] . To this end , we used changes in spectral composition to selectively modulate melanopsin excitation ( see Methods; 14/15 mice above tested with these stimuli ) . When presented with large steps in melanopsin excitation ( 92% Michelson contrast ) generated in this way , “blue”-ON cells showed slow and sustained increases in firing ( Fig 3A; peak response = 3 . 2 ± 0 . 8 spikes/s above baseline; paired t test , p<0 . 01 , n = 13 ) , as previously described for melanopsin-driven responses [8 , 27 , 28] . The behaviour of the rare “yellow” ON cells to this stimulus was variable ( Fig 3B; n = 4 ) , while colour-insensitive cells showed the expected excitatory response ( Fig 3A; peak response = 1 . 8 ± 0 . 2 spikes/s above baseline; paired t test , p<0 . 01 , n = 23 ) . These data therefore reveal that both chromatic and achromatic cells have access to melanopsin-dependent information about irradiance . Of note , for the smaller changes in opsin excitation applied above ( 70% Michelson; ~0 . 75 log units ) , we found that the inclusion of melanopsin contrast had little impact on the integrated cellular response . Thus for both chromatic and achromatic populations , responses evoked by spectrally neutral increases in irradiance ( “energy” ) were very similar in magnitude to those observed where changes in irradiance were restricted to just cone opsins ( Fig 3A; subtraction: energy − “brightness” ) . This was true even for steady-state components of the SCN response ( last 1 s of step ) —we found no significant difference in responses to two conditions ( paired t test; p>0 . 05 for both blue-ON and achromatic populations ) . Thus SCN responses to relatively modest changes in light intensity and/or spectral composition are , in fact , dominated by those originating with cones . How then do chromatic and irradiance responses interact to encode time of day under more natural conditions ? To address this question , we produced stimuli that recreated , for Opn1mwR mice , the change in irradiance and colour experienced by wild-type ( green cone ) mice across the twilight to daylight transition ( Fig 4A ) . We presented these as discrete light steps from darkness , to simulate the challenge in telling time of day faced by a rodent emerging from a subterranean burrow to sample the light environment . Due to their scarcity , we were unable to determine the behaviour of yellow-ON cells under these conditions . However , blue ON cells reliably exhibited a near linear increase in firing rate as a function of simulated solar angle ( Fig 4B and 4C; n = 9 from 7 mice ) , indicating that their sensitivity is well suited to track changes in colour/irradiance occurring across the twilight to daylight transition . Interestingly , the range of solar angles to which these neurons responded was substantially greater than that for achromatic cells recorded in the same set of mice ( Fig 4D and 4E; see also S3 Fig; n = 8 ) indicating that they may be an especially important source of temporal information for the clock around twilight . To determine the extent to which this ability of blue-ON cells to encode solar angle relied upon their chromatic opponency , we next presented stimuli that recreated the natural change in irradiance over twilight but in which colour was invariant . Two versions of these stimuli were produced , in which colour was fixed either to that at the lowest solar angle for which data was available ( “night” ) or to that recorded in daylight ( “day”; Fig 4A ) . Whereas achromatic cells were unable to distinguish between these two stimulus sets ( Fig 4D and 4E; F-test , p = 0 . 72 ) , the relationship between solar angle and blue-ON cell firing rate was consistently disrupted under these conditions ( Fig 4B and 4C; F-test , p = 0 . 009; see also S3A Fig ) . Thus , firing was reliably higher for “night” and lower for “day” conditions than appropriate for that time of day . These effects are consistent with the blue-ON nature of the chromatic units and confirm that these cells employ a combination of colour and irradiance signals in order to encode time of day . These electrophysiological recordings indicate a significant fraction of neurons in the SCN convey information about changes in spectral composition occurring during natural twilight . We hypothesised , therefore , that by improving the SCN’s ability to estimate solar angle , activation of the colour mechanism would influence the phasing of circadian rhythms under natural conditions . To determine whether this was indeed the case , we scaled up our twilight stimuli to produce an artificial sky that could be presented to freely moving mice over many days in their home cage . We aimed then to compare the phase of circadian rhythms ( assayed using body temperature telemetry ) under exposure to lighting conditions that recreated natural changes in irradiance across dawn/dusk transitions , with or without the associated alterations in colour ( S4 Fig; “irradiance only” twilight replicated “night” spectral composition ) . To maximise our ability to detect changes in phasing under these conditions , we modelled the temporal profile of these photoperiods on the extended twilight of a northern-latitude summer ( S4C Fig ) . To allow us to readily separate irradiance and colour elements , we undertook these experiments in Opn1mwR mice . Importantly , however , we designed the stimuli to recreate the change in colour across twilight that is experienced by normal , wild-type mice . We found that the inclusion of colour significantly altered the phase of circadian entrainment . Peak body temperature occurred consistently later when irradiance and colour elements of twilight were included compared to the irradiance signal alone ( Fig 5A; 31 ± 8 min; paired t test , p = 0 . 003; n = 10 ) . This distinction was absent in mice lacking cone phototransduction ( Cnga3-/- , [29 , 30]; Fig 5B; 6 ± 9 min; paired t test , p = 0 . 51; n = 9 ) , confirming that it originated with cone-dependent colour coding , rather than any differences in the pattern of rod/melanopsin activation between the two photoperiods . As further confirmation that these differences in body temperature cycles reflected an action on the timing of central clock output , we also monitored SCN firing rate rhythms in a subset of mice via ex vivo multielectrode array recordings . We and others have previously shown that the distribution of daily electrical activity patterns among individual SCN neurons encodes photoperiod duration , resulting in broad phase distributions under summer days [31 , 32] . Consistent with this work , peak multiunit firing ( sampled across small groups of neurons ) in the ex vivo SCN of twilight-housed mice was widely distributed across recording epochs corresponding to projected day . Importantly , in line with our body temperature data , this distribution was centred around the middle of the projected day for Opn1mwR mice exposed to “natural” twilight ( Fig 6A; n = 124 SCN electrodes from seven slices ) but shifted substantially earlier when mice were housed under twilight that lacked changes in colour ( Fig 6B; p<0 . 001 , bootstrap percentiles; n = 170 SCN electrodes from six slices ) . A similarly early phase of peak SCN electrical activity was also observed in slices prepared from Cnga3-/- individuals housed under natural twilight ( S5 Fig; p<0 . 001 versus Opn1mwR , bootstrap percentiles ) , confirming that the cone-dependent colour signal is indeed required for appropriate biological alignment with twilight . We also found that , across the three groups , Opn1mwR mice exposed to “irradiance-only” twilight exhibited a significantly broader distribution of SCN phasing ( Brown-Forsythe test , p = 0 . 01 ) , suggesting that the inappropriate cone signals under this photoperiod partially impair SCN synchrony .
Here we demonstrate that the mammalian clock has access to information about not just the amount but also the spectral composition of ambient illumination , in the form of a cone-dependent colour opponent input that reports blue–yellow colour . The idea that chromatic signals associated with twilight might provide important cues for circadian photoentrainment has been proposed previously [3 , 15] . However , the significant technical challenges inherent in distinguishing the influence of changes in colour versus brightness have left the specific role of colour untested , until now . Our work thus represents the first demonstration that colour-opponent signals influence the circadian clock in any mammalian species . It is clear , from the long history of housing animals under artificial lighting , that colour signals are not necessary for circadian entrainment per se . However , our data indicates that most mammals could use colour [18 , 33–35] to provide additional information about sun position , above that available from simply measuring irradiance . Our entrainment experiments likely underestimate the importance of that colour signal under field conditions as they lack the daily variation in cloud cover that makes irradiance-alone a less reliable indicator of time of day . Nevertheless , even under these conditions , we find a significant impact of the twilight spectral change on the phasing of entrained rhythms . This reveals that spectral opponency contributes to the most fundamental function of the entrainment mechanism , ensuring correct timing of physiological and behavioural rhythms . Given the nature of the change in spectral composition , we might expect that it would be available to any species capable of comparing the activity of short with middle/longer wavelengths . Our own subjective experience is that the event is detectable to humans , and a chromatic opponency equivalent to that described here could account for previous reports of subadditivity for polychromatic illumination in human melatonin suppression [36 , 37] . It is also noteworthy that the majority of mammalian species have retained the short and mid-/long-wavelength cone opsins required to detect changes in spectral composition associated with twilight ( for a detailed discussion of the exceptions to this rule see [17 , 18] ) . Similarly , earlier studies have identified the capacity for blue–yellow colour discrimination in the pineal/parietal organs of a number of non-mammalian vertebrates , including reptiles , amphibians , and fish ( for review see [38] ) . By directly influencing melatonin secretion , chromatic signals are thus presumably also a key component of the neural mechanisms responsible for appropriate alignment of non-mammalian physiology relative to dawn and dusk . Alongside our present data , it appears then that the use of colour as an indicator of time of day is an evolutionarily conserved strategy , perhaps even representing the original purpose of colour vision . The specific sensory properties of the circadian photoentrainment mechanism in mammals have long remained a subject of debate [2] . SCN neurons are known to receive input from all major classes of retinal photoreceptor [8–13] . However , since “cone-only” mice do not reliably entrain to conventional light–dark cycles , current models posit that photoentrainment is primary driven by a combination of rod and melanopsin inputs [11 , 12] . By contrast , the proposed role of cones has been to allow the clock to track relatively high frequency changes in light—a signal that does not appear to play much role in circadian entrainment under conditions most commonly employed in the laboratory ( but see [12] ) . Our data thus establish an important new role for cones in photoentrainment , one which would not be apparent under standard laboratory conditions but will act as an essential regulator of biological timing in more natural settings . Insofar as most retinal input to the clock is provided by ipRGCs [4 , 6 , 7] the appearance of colour opponency in this subset of retinal ganglion cells would provide a simple explanation for the chromatic responses of SCN neurons observed here . Colour opponency has not yet been documented in mouse ipRGCs [39] , but has been reported in primates [40] ( although it is unknown whether any of these cells project to the SCN ) . Interestingly , the dominant form of spectral opponency we observe here in the mouse SCN ( blue-ON/yellow-OFF ) is opposite to that reported for primate ipRGCs and , most recently , for chromatic response of the pupil in humans [41] . While this would , by no means , rule out a role for chromatic influences on the human circadian system , it is also currently unclear whether such yellow-ON/blue-OFF responses are a characteristic feature of all primate ipRGCs . Indeed , such behaviour certainly appears inconsistent with the sensory properties of human melatonin regulation , which seems to exhibit a short- rather than long-wavelength bias [42] . Of course , alternative possibilities to that outlined above are that colour information reaches SCN neurons via the small number of non-melanopsin-expressing RGC inputs or is generated by a mechanism distinct from the conventional retinal colour processing circuitry . Previous work indicates that asymmetries in the gradient of cone opsin expression in the mouse retina could impose an indirect form of chromatic bias for stimuli larger than the cells’ receptive field [43] . Alternatively , opponent responses in the SCN may be generated centrally , e . g . , via local processing or indirect visual input from the intergeniculate leaflet . Indeed , based on our identification of a rare yellow-ON cell exhibiting inhibitory responses to melanopsin contrast , we speculate that central processing could contribute to at least some of the responses reported here . Regardless of their biological origin , chromatic signals provide the SCN with additional information about solar angle , above that available from measuring brightness alone , allowing the clock to appropriately time its output under natural photoperiods . Based on the widespread capacity for colour vision among mammals ( and the previous identification of colour opponent ipRGCs in primates ) , we suggest related mechanisms are likely to be broadly applicable across many mammalian species .
All animal use was in accordance with the Animals ( Scientific Procedures ) Act of 1986 ( United Kingdom ) . Electrophysiological experiments were performed under urethane anaesthesia; other procedures were conducted under isfluorane anaesthesia . Unless otherwise stated , animals used in this study ( homozygous Opn1mwR and Cnga3-/- mice ) were housed under a 12-h dark/light cycle at a temperature of 22°C with food and water available ad libitum . Spectral irradiance measurements ( 280–700 nm , 0 . 5 nm bins ) were collected in Manchester , UK ( Lat . : 53 . 47 , Long . : -2 . 23 , Elevation 76 m ) every minute across the solar cycle using a METCON diode array spectroradiometer contained within a temperature stabilised weatherproof housing . The global entrance optics was levelled and mounted at a rooftop monitoring site , providing a horizon relatively clear of obstructions , the entrance optics being connected to the spectrometer by way of a 600 μm diameter 5 m long optical fibre . Instrument calibrations were carried out with reference to spectral irradiance standards , traceable to NIST ( National Institute of Standards and Technology , United States ) . Instrument dark counts were observed to be spectrally flat and were removed by subtracting the mean value for wavelengths <290 nm ( where no ground-level solar signal is present ) . Data analysed were spectral irradiance measurements collected between 31 August and 14 October 2005 ( 41 d ) . Due to gaps in the data collection record , we were able to extract from these 71 complete dawn/dusk transitions ( from 36 d ) . No attempt was made to select data on the basis of weather condition although the period was broadly representative in comparison to relevant climatological averages . For each twilight transition , we first calculated the average spectral irradiance profile as a function of solar angle relative to the Horizon ( 0 . 5° bins , 2–5 measurements/bin ) . We restricted this analysis to solar angles greater than 7° below the horizon , since our detector was specifically optimised to obtain measurements across light intensities encountered through civil twilight to daytime ( making night-time measurements less reliable ) . We next converted these spectral irradiance profiles into effective photon fluxes as experienced by mouse opsin proteins , using established and validated procedures [19 , 20 , 22] based on Govardovski visual pigment templates [44] and published values for mouse lens transmission [45] . Calculations presented in the manuscript were based on the following peak sensitivity ( λmax ) : UVS cone opsin-365 nm , Melanopsin-480 nm , Rhodopsin-498 nm , MWS cone opsin 511 nm . The resulting series” of photon flux versus solar angle values for each opsin were then analysed individually or in combination ( additive or as ratios ) . Specifically , we calculated the percentage of variance for the dataset in question that was explained by sun position , using the following calculation ( with N representing the total number of data points , K the number of dawn/dusk observations and P the number of solar angle bins ) : Var|θ=100K∑h=1P ( X¯h−X¯ ) 2∑i=1N ( Xi−X¯ ) 2 Since there was no apparent difference in photon flux versus solar angle profiles obtained during dawn or dusk transitions , we pooled these data for the above analysis , treating each as an independent observation . For comparisons of colour versus irradiance based estimates of solar angle ( Fig 1D and associated text ) , irradiance was defined as effective photon flux at UVS+MWS cone opsins . Values obtained using other mouse opsins ( singly or in combination ) produced essentially identical results . For the aforementioned comparisons , estimates of mean and standard deviation for Var|θ were obtained based on bootstrap replicates ( every possible combination of 69 out of the total 71 dawn/dusk observations ) . Similar analysis to those described above , but performed using only observations taken at either dawn or dusk also produced essentially identical results . Calculations of ““blue–yellow”“colour index ( Fig 1C and 1D ) were based on the ratio of MWS:SWS cone opsin activation ( [MWS+LWS]/SWS for human visual system ) . Urethane ( 1 . 55 g/kg ) anaesthetised adult ( 60–120 d ) male Opn1mwR mice were prepared for stereotaxic surgery as previously described [8] . Recording probes ( Buszaki 32L; Neuronexus , MI , US ) consisting of four shanks ( spaced 200 μm ) , each with eight closely spaced recordings sites in diamond formation ( intersite distance 20–34 μm ) were coated with fluorescent dye ( CM-DiI; Invitrogen , Paisley , UK ) and then inserted into the brain 1 mm lateral and 0 . 4 mm caudal to bregma at an angle of 9° relative to the dorsal-ventral axis . Electrodes were then lowered to the level of the SCN using a fluid-filled micromanipulator ( MO-10 , Narishige International Ltd . , London , UK ) . After allowing 30 min for neural activity to stabilise following probe insertion , wideband neural signals were acquired using a Recorder64 system ( Plexon , TX , US ) , amplified ( x3000 ) and digitized at 40 kHz . Action potentials were discriminated from these signals offline as “virtual”-tetrode waveforms using custom MATLAB ( The Mathworks Inc . , MA , US ) scripts and sorted manually using commercial principle components based software ( Offline sorter , Plexon , TX , US ) as described previously [46] . Surgical procedures were completed 1–2 h before the end of the home cage light phase , such that electrophysiological recordings spanned the late projected day-early projected night , an epoch when the SCN light response is most sensitive . Cells were initially characterised as light responsive on the basis of responses to bright mono and polychromatic light steps ( 10–30 s dur . ; intensity >1014 photons/cm2/s ) . Once visual responsiveness was confirmed , experimental stimuli were applied as described below . Following the experiment , accurate electrode placement was confirmed histologically as described previously [8] . Projected anatomical locations of light response units reported in this study are presented in S6 Fig . All visual stimuli were delivered in a darkened chamber from a custom built source ( Cairn Research Ltd , Kent , UK ) consisting of independently controlled UV , blue and amber LEDs ( λmax: 365 , 460 , and 600 nm respectively ) . Light was combined by a series of dichroic mirrors and focused onto a 5 mm diameter piece of opal diffusing glass ( Edmund Optics Inc . , York , UK ) positioned <1 mm from the eye ( contralateral to the recording probe for SCN recordings ) . LED intensity was controlled by a PC running LabView 8 . 6 ( National instruments ) . Light measurements were performed using a calibrated spectroradiometer ( Bentham instruments , Reading , UK ) . LED intensity was initially calibrated ( using the principles described above ) to recreate for Opn1mwR individuals the effective rod , cone and melanopsin excitation experienced by a wild-type ( green cone ) mouse visual system under typical natural daylight ( average values from our environmental data above at a solar angle 3° above the horizon; S1A Fig ) . We also carefully calibrated differential modulations in the intensity of each LED to produce stimuli that independently varied in apparent brightness for one or both cone opsin classes ( either in unison or antiphase ) with no apparent change in rod or melanopsin excitation ( S1B Fig ) . In each case , brightness for the stimulated opsin was varied by ±70% , to produce an overall 4 . 7-fold increase in intensity of between “bright” and “dim” phases of the stimulus . Transitions between the two stimulus phases occurred smoothly over 50ms ( half sinusoid profile ) . We also applied stimuli that selectively modulated melanopsin excitation ( ±92% ) , without changing effective cone excitation . These later also , in principle , modulated apparent brightness for rod photoreceptors ( ±84% ) , however we think a rod contribution to the resulting responses unlikely owing to the high background light levels ( 14 . 9 rod effective photons/cm2/s ) and our previous work suggesting that rods have little influence on acute electrophysiological light responses in the SCN [8] . Indeed , similar stimuli evoke very little response in the lateral geniculate nuclei of melanopsin knockout animals [22] . In a subset of experiments ( 7/15 ) we also applied a second set of stimuli designed to recreate various stages of twilight , using our calculations of the effective photon fluxes experienced by mouse opsins at solar angles between -7 and 3° relative to the horizon . These were applied as light steps ( 30 s ) from darkness in random sequence with an interstimulus interval of 2 min . To confirm whether elements of the resulting responses were dependent on spectral composition , these stimuli were interspersed with two additional stimulus sets which were identical except that irradiance for the UVS opsin was fixed at a constant ratio relative to LWS ( mimicking either day or night spectral composition ) . For behavioural experiments , we used similar principles to generate photoperiods that smoothly recreated our measured changes in twilight illumination , with ( “natural” ) or without the associated change in spectral composition ( irradiance-only: spectra fixed to mimic “night” ) . Stimuli were generated by an array of three violet ( 400 nm ) and three amber ( 590 nm ) high-power LEDs ( LED Engin Inc . , San Jose CA , US ) placed behind a polypropylene diffusing screen covering the top of the cage . The combination of multiple LEDs allowed a larger range of brightness ( from dark up to approximately 25 W/m2 for the violet and 10 W/m2 for the amber ) . Intensity of each LED was independently controlled by a voltage controlled driver ( Thorlabs Inc . , Newton NJ , US ) . The light intensity modulation signals were provided by a PC running Labview through a voltage output module ( National Instruments ) , and followed a temporal profile that recreated the sun’s progression during a northern latitude summer ( calculations based on Stockholm , Sweden; Lat: 59 , Long: 18 , Elevation 76 m , 20 June 2013; total twilight duration = 2 . 3 h ) . To determine the impact of twilight spectral changes on mouse entrainment , female Opn1mwR and Cnga3-/- mice ( housed under an 18:6 light–dark [LD] cycle ) were first implanted with iButton temperature loggers ( Maxim , DS1922L-F5# ) . To reduce weight and size , these were dehoused and encapsulated in a 20% Poly ( ethylene-co-vinyl acetate ) and 80% paraffin mixture as described by Lovegrove [47] . For implantation , mice were anaesthetised with isoflurane ( 1%–5% in O2 ) and the temperature logger implanted into the peritoneal cavity . Following surgery , animals were given a 0 . 03 mg/kg subcutaneous dose of buprenorphine and allowed to recover for at least 9 d in 18:6 LD before the start of the experiment . The timing of lights off under this cycle was designated as Zeitgeber time ( ZT ) 12 and the timing of experimental photoperiods were set to align their midnight ( ZT15 ) with this square wave LD cycle . Following recovery , group housed mice ( five per cage ) were transferred to the natural twilight photoperiod . The cage environment contained an opaque plastic hide , allowing the animals to choose their own light sampling regime . After 14 d , mice were then returned to 18:6 LD for a further 14 d and finally transferred to the “irradiance-only” twilight photoperiod . At the end of the experiment , mice were culled by cervical dislocation and temperature loggers recovered . Temperature data ( recorded in 30 min time bins ) was processed by upsampling to 5 min resolution ( cubic spline interpolation ) , Gaussian smoothing ( SD = 45 min ) , and normalisation as a fraction of daily temperature range . Phase of entrainment was estimated as the timing of peak body temperature from that individual’s daily average profile ( calculated from the last 9 d in each photoperiod ) . Opn1mwR and Cnga3-/- mice were housed under twilight stimuli of either “natural” or “night” composition ( as described above ) for at least 14 d prior to experiments . Mice were removed from the home cage 30–60 min after the end of the dawn transition ( ~ZT19 ) and culled by cervical dislocation followed by decapitation . The brain was then rapidly removed , mounted onto a metal stage and cut using a 7000 smz vibrating microtome ( Campden Instrument , UK ) in ice-cold ( ~4°C ) sucrose-based slicing solution composed of ( in mM ) : sucrose ( 189 ) ; D-glucose ( 10 ) ; NaHCO3 ( 26 ) ; KCl ( 3 ) ; MgSO4 ( 5 ) ; CaCl2 ( 0 . 1 ) ; NaH2PO4 ( 1 . 25 ) ; oxygenated with 95% O2/5% CO2 mixture . Coronal brain slices containing the SCN ( 350 μm ) were then immediately transferred into a petri dish containing oxygenated artificial cerebrospinal fluid ( aCSF ) composed of ( in mM ) : NaCl ( 124 ) ; KCl ( 3 ) ; NaHCO3 ( 24 ) ; NaH2PO4 ( 1 . 25 ) ; MgSO4 ( 1 ) ; glucose ( 10 ) ; CaCl2 ( 2 ) ; slices were then left to rest at room temperature ( 22 ± 1°C ) . Approximately 30 min after slice preparation , slices were placed , recording side down , onto 6x10 perforated multielectrode arrays ( pMEAs; Multichannel Systems , MCS , Germany ) . Slices were visualised under the microscope and photos were taken with a GXCAM-1 . 3 camera ( GX Optical , UK ) in order to confirm appropriate slice placement over pMEA electrode sites . Slices were held in place by both the suction via the MEA perforations and a harp slice grid ( ALA Scientific Instruments Inc . , US ) . The pMEA recording chamber was continuously perfused with pre-warmed oxygenated aCSF ( 34 ± 1°C ) to both slice surfaces at a rate of 2 . 5–3 ml/min . Neural signals were acquired as time-stamped action potential waveforms using a USB-ME64 system and a MEA1060UP-BC amplifier ( MCS , Germany ) . Signals were sampled at 12 . 5 kHz , High pass filtered at 200 Hz ( second order Butterworth ) with a threshold of usually at -16 . 5 μV . Recordings were maintained for a total duration of 30 h . At the end of each recording , slices were treated with bath applications of 20μM NMDA to confirm maintained cell responsiveness , followed by 1 μM TTX to confirm acquired signals exclusively reflected Na+-dependent action potentials . All drugs were purchased from Tocris ( UK ) , kept as stock solutions at -20°C ( dissolved in dH2O ) , and were diluted to their respective final concentrations directly in pre-warmed , oxygenated aCSF; all drugs were bath applied for 5 min . Multiunit action potential firing rates detected at electrodes located within the SCN region were then selected for further analysis . Data were subsequently binned ( 60 s ) and smoothed via boxcar averaging ( width: 2 h ) to determine the timing of peak activity . Channels where peak firing did not decay by >50% within ±12 h , or where peak firing was less than 0 . 2 spikes/s were excluded from this analysis , such that on average 22 ± 2 SCN electrodes were analysed for each experiment . Based on peak firing rates observed ( mean ± SEM: 6 . 9 ± 0 . 5 spikes/s ) we estimate these typically represent recordings from less than four neurons . To assess for significant differences in the timing of population activity under our different experimental conditions , we drew 1 , 000 samples of 100 randomly selected neurons from each condition ( Opn1mwR “natural” , Opn1mwR “night” , Cnga3-/- “natural” ) . By calculating the circular mean phase for each sample , we thus obtained estimates of the probability that the observed population means differed by chance .
|
Animals use an internal brain clock to keep track of time and adjust their behaviour in anticipation of the coming day or night . To be useful , however , this clock must be synchronised to external time . Assessing external time is typically thought to rely on measuring large changes in ambient light intensity that occur over dawn/dusk . The colour of light also changes over these twilight transitions , but it is currently unknown whether such changes in colour are important for synchronising biological clocks to the solar cycle . Here we show that the mammalian blue–yellow colour discrimination axis provides a more reliable indication of twilight progression than a system solely measuring changes in light intensity . We go on to use electrical recordings from the brain clock to reveal the presence of many neurons that can track changes in blue–yellow colour occurring during natural twilight . Finally , using mice housed under lighting regimes with simulated dawn/dusk transitions , we show that changes in colour are required for appropriate biological timing with respect to the solar cycle . In sum , our data reveal a new sensory mechanism for estimating time of day that should be available to all mammals capable of chromatic vision , including humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Colour As a Signal for Entraining the Mammalian Circadian Clock
|
Entry of human immunodeficiency virus type 1 ( HIV-1 ) commences with binding of the envelope glycoprotein ( Env ) to the receptor CD4 , and one of two coreceptors , CXCR4 or CCR5 . Env-mediated signaling through coreceptor results in Gαq-mediated Rac activation and actin cytoskeleton rearrangements necessary for fusion . Guanine nucleotide exchange factors ( GEFs ) activate Rac and regulate its downstream protein effectors . In this study we show that Env-induced Rac activation is mediated by the Rac GEF Tiam-1 , which associates with the adaptor protein IRSp53 to link Rac to the Wave2 complex . Rac and the tyrosine kinase Abl then activate the Wave2 complex and promote Arp2/3-dependent actin polymerization . Env-mediated cell-cell fusion , virus-cell fusion and HIV-1 infection are dependent on Tiam-1 , Abl , IRSp53 , Wave2 , and Arp3 as shown by attenuation of fusion and infection in cells expressing siRNA targeted to these signaling components . HIV-1 Env-dependent cell-cell fusion , virus-cell fusion and infection were also inhibited by Abl kinase inhibitors , imatinib , nilotinib , and dasatinib . Treatment of cells with Abl kinase inhibitors did not affect cell viability or surface expression of CD4 and CCR5 . Similar results with inhibitors and siRNAs were obtained when Env-dependent cell-cell fusion , virus-cell fusion or infection was measured , and when cell lines or primary cells were the target . Using membrane curving agents and fluorescence microscopy , we showed that inhibition of Abl kinase activity arrests fusion at the hemifusion ( lipid mixing ) step , suggesting a role for Abl-mediated actin remodeling in pore formation and expansion . These results suggest a potential utility of Abl kinase inhibitors to treat HIV-1 infected patients .
HIV-1 enters cells in a pH-independent manner by fusion at the plasma membrane or from within endosomes [1]–[3] . HIV-1 entry requires multiple conformational changes in the HIV-1 glycoprotein , and rearrangement of the actin cytoskeleton . These events are triggered by binding of the viral envelope ( Env ) surface subunit gp120 to the primary receptor CD4 and one of two chemokine coreceptors , CCR5 or CXCR4 [1] , [4] . This interaction activates signaling events in the cell , similar to those initiated by natural ligands , such as Ca2+ mobilization , activation of RhoGTPases , and phosphorylation of tyrosine kinases , pyk2 , Zap70 and p56lck [4]–[6] . Rho family GTPases , which include the Cdc42 , Rac , and Rho subfamilies , are responsible for regulating signaling from membrane receptors to the actin cytoskeleton . The Rho sub-family stimulates myosin based contractility , and drives the formation of stress fibers and focal adhesions . The Rac sub-family stimulates lamellipodia and membrane ruffles , and the Cdc42 subfamily stimulates the formation of filopodia [7]–[9] . We showed that HIV-1 Env binding to target cells induces activation of Rac , stimulates membrane ruffles and lamellipodia , and fusion is inhibited by dominant negative Rac [4] , [10] . Furthermore , HIV-1 Env-induced Rac activation depends on activation of Gαq , phospholipase C ( PLC ) , Ca2+ mobilization , protein kinase C ( PKC ) , pyk2 and the GTPase Ras [5] . In the current study we identified the fusion-specific effectors of Rac required for actin cytoskeleton rearrangements that mediate membrane fusion and entry . Guanine nucleotide exchange factors ( GEFs ) activate GTPases , facilitating the GDP to GTP switch , and regulate their downstream effects by participating in scaffolding protein complexes , thereby linking GTPase activity to specific effectors [7]–[9] . HIV-1 Env-induced Rac activation is mediated by a specific Rac GEF , either Tiam-1 or Trio [10] , [11] . There are multiple effectors of Rac , including serine/threonine kinases , lipid kinases , actin-binding proteins , and adaptor/scaffold molecules [7] , [12] . PAK is a downstream effector of Rac and Cdc42 that promotes stabilization of actin networks . Another downstream effector of Rac that nucleates actin polymerization is the Arp2/3 complex . The Arp2/3 complex is activated by the Wave2 complex through IRSp53 , an adaptor protein that binds Rac and Wave2 [7] . The Wave2 complex includes Rac-associated protein 1 , Nck-associated protein , Abl-interacting protein 2 , and heat shock protein C300 . Wave2 also associates with Abl , and Abl-mediated phosphorylation of Wave2 promotes its activation [13] , [14] . In addition to determining which Rac effectors are critical for membrane fusion , we studied the steps in the membrane fusion process affected by these signaling molecules . These data demonstrate that the Wave2 signaling complex and Abl are required for Env-mediated membrane fusion , entry , and infection and that Abl kinase inhibitors arrest the fusion process at hemifusion .
To determine whether Abl , Trio , or Tiam-1 were required for HIV-1 Env-mediated cell-cell fusion , expression of these proteins was down regulated by RNA interference ( RNAi ) in U87 . CD4 . CCR5 cells . Cells expressing siRNA were then mixed with BSC40 cells expressing different Env subtypes and Env-dependent cell-cell fusion was measured . Transfection of target cells with siRNA to Tiam-1 and Abl decreased levels of Env-mediated cell-cell fusion by an average of 79±5% and 74±5% respectively for both HIV-1 R5 and dual-tropic Env-subtypes ( Figure 1A , left ) . There was no significant fusion observed with CCR5 expressing target cells and X4 Env expressing cells with or without siRNA , as expected . The decrease in the levels of fusion correlated well with the decreased steady-state level of Tiam-1 , and Abl as detected by immunoblot ( Figure 1C ) . A siRNA directed against Trio had no effect on Env-induced cell-cell fusion despite a 70% reduction in expression of the Trio protein ( Figure 1A and C ) . To determine whether Tiam-1 and Abl are acting exclusively upstream of Rac , a constitutively active Rac mutant , RacV12 was expressed in siRNA transfected cells . Expression of RacV12 in cells expressing siRNA to Tiam-1 reversed the effects of this siRNA on fusion , suggesting that Tiam-1 is functioning upstream of Rac . In contrast , levels of fusion in cells expressing RacV12 and siRNA to Abl were only 53±1% that of cells expressing RacV12 and control siRNA , suggesting a role for Abl upstream and downstream of Rac ( Figure 1A , right ) . Tiam-1 binds to the Rac and Cdc42 effector IRSp53 , enhancing IRSp53 binding to Rac and activation of the Wave2 scaffolding complex [15] . To determine the role of these Rac effectors in Env-mediated membrane fusion , their expression was down regulated by RNAi in U87 . CD4 . CCR5 cells . The siRNA expressing cells were mixed with Env-expressing cells and cell-cell fusion was measured . Expression of siRNA to IRSp53 , Wave2 , and Arp3 decreased fusion by 74±5% 77±4% and 78±4% , respectively . The decrease in fusion with these siRNAs was not overcome by expression of RacV12 , suggesting that these proteins are required downstream of Rac ( Figure 1B ) . The decrease in levels of fusion correlated with the decrease in protein expression in cells expressing these siRNAs , as seen by immunoblot ( Figure 1C ) , and each siRNA was specific for its target protein ( Figure S1A ) . Treatment of cells stably expressing siRNA resistant Arp3 , with Arp3 targeted siRNA had no effect on Env-mediated cell-cell fusion ( Figure 1D , E ) . In contrast , with untransfected cells , and cells stably expressing siRNA resistant Arp3 , treatment with siRNA to Rac decreased fusion by 75±5% and 76±3% respectively ( Figure 1D ) . These results show that the effects of RNAi on fusion were specific to inhibition of their target molecules . To demonstrate the role of Tiam-1 , Abl , Rac , IRSp53 , Wave2 and Arp3 in virus-cell fusion , their expression was down regulated by RNAi in TZM-BL cells , a derivative of HeLa cells that express CD4 , CCR5 , and CXCR4 , and these cells were then used in a Vpr-Blam assay [16] , [17] . In this assay siRNA expressing cells were mixed for 90 min with HIV-1 strains with cores carrying a β-lactamase ( BlaM ) -Vpr chimera , and pseudotyped with Env from ADA ( R5 ) , YU2 ( R5 ) or HXB2 ( X4 ) , and fusion was quantified by measuring the cytosolic activity of viral core-associated BlaM [18] . Expression of siRNA to Tiam-1 , Abl , Rac , IRSp53 , Wave2 , and Arp3 decreased virus-cell fusion by an average of 80±4% , 83±1% , 76±4% , 82±6% , 77±3% and 82±6% , respectively , for HIV-1 R5 and X4 Env subtypes ( Figure 1F ) . These results show that activation of the Wave2 signaling complex is required for Env-dependent cell-cell fusion and virus-cell fusion . Since treatment of cells with Abl targeted siRNA led to a decrease in Env-dependent cell-cell fusion and virus-cell fusion we wanted to determine whether treatment of target cells with commercially available Abl kinase inhibitors , imatinib ( IMB ) , nilotinib ( NIL ) , and dasatinib ( DAS ) , block fusion . IMB is a relatively specific inhibitor of Bcr-Abl , Abl , Arg , and class III receptor tyrosine kinases . NIL is an Abl kinase inhibitor 20–50 fold more potent than IMB at inhibiting Abl . DAS , originally designed as a Src family kinase inhibitor , antagonizes Abl , ephrin and platelet-derived growth factor receptor kinases , and kit . DAS is 300 fold more potent than IMB at inhibiting Abl [19] , [20] . To determine the concentrations of these Abl kinase inhibitors that inhibit Abl kinase activity and Env-mediated cell-cell fusion , without non-specific effects , Abl kinase activity , trypan blue analysis , vaccinia virus infection , and T7 polymerase activity were measured in addition to Env-dependent cell-cell fusion ( Figure S1B , S2 , and data not shown ) . Treatment of U87 . CD4 . CCR5 cells with 10 uM IMB , 500 nM NIL , and 300 nM DAS for 1 h prior to and during 3 h incubation with Env-expressing cells decreased Env-mediated cell-cell fusion by an average of 95±2% , 92±5% , and 92±6% , respectively , and Abl kinase activity by 85–87% ( Figure 2A and S1B ) . The CCR5 inhibitor TAK-779 , which completely blocks Env-mediated cell-cell fusion and infection of CCR5 expressing cells , was included as a control , and it decreased Env-dependent cell-cell fusion by 99±1% and Env-mediated Abl kinase activation by 98% ( Figure 2A and S1B ) . Similar results were observed with U87 . CD4 . CXCR4 cells treated with CXCR4 inhibitor AMD3100 and Abl kinase inhibitors and incubated with cells expressing HIV-1 X4 or dual-tropic Env subtypes ( Figure S3A ) . There was no decrease in T7 polymerase activity , or localization of CD4 and CCR5 on the cell surface ( Figure S4 and data not shown ) . Expression of RacV12 in U87 . CD4 . CCR5 cells treated with IMB , NIL and DAS increased the level of fusion by an average of 3 . 5-fold ( * , P<0 . 05 ) compared to treated cells without RacV12 , suggesting a role of Abl kinase activity upstream of Rac ( Figure 2B ) . To determine the effect of these Abl kinase inhibitors on Env-induced Rac activation , U87 . CD4 . CCR5 cells were treated with inhibitors for 1 h prior to mixing with BSC40 cells expressing no HIV-1 Env , HIV-1 X4 Env , or HIV-1 R5 Env for 30 minutes in the presence of inhibitor . The mismatched X4 Env , that does not induce Rac activation in CCR5 expressing cells , and the CCR5 inhibitor TAK-779 , which completely blocks Env-mediated Rac activation in CCR5 expressing cells , were included as controls [5] . Env-induced Rac activation was abolished in cells treated with TAK-779 , and all three of the Abl kinase inhibitors ( Figure 2C ) . To validate these effects in a relevant HIV-1 target cell , peripheral blood lymphocytes ( PBLs ) , which express CD4 , CCR5 and CXCR4 , were used as the target cell population in an Env-dependent cell-cell fusion assay . Treatment of PBLs with IMB , NIL , and DAS decreased fusion by an average of 92±1% , 92±3% , and 99 . 5±1% , respectively , for HIV-1 R5 , dual-tropic and X4 Env subtypes ( Figure 2D ) . The CCR5 inhibitor TAK-779 , as expected , completely blocked fusion mediated by R5 Env-expressing cells , inhibited fusion mediated by dual-tropic Env by 56±2% , and had no effect on fusion mediated by X4 Env ( Figure 2D ) . A long term infection assay was also performed where PBLs were infected with 150 ng of the X4 HIVHXB2 virus after 1 h preincubation with no inhibitor , DMSO , 10 µM IMB , 250 nM NIL , or 75 nM DAS . After 3 h , virus and inhibitors were washed off , inhibitors were added back and the plate was incubated at 37° for 21 days with addition of the inhibitors every 24 h . After 21 days the samples were assayed for cell viability and p24 antigen content . Treatment with IMB , NIL , and DAS decreased cell viability of HIVHXB2 infected cells by 17±4% , 8±5% , and 8±3% respectively and decreased infection by 52% , 51% and 94% compared to DMSO treated cells ( Figure S5 ) . To validate the specificity of these effects , we performed an Env-dependent cell-cell fusion assay with cells stably expressing two different drug resistant Bcr-Abl mutants ( Y253F and T315I ) , or expressing wild type ( WT ) Bcr-Abl [21] . Expression of the drug resistant Bcr-Abl mutants but not WT Bcr-Abl resulted in recovery of fusion ( Figure 2E ) , demonstrating that the effects of these inhibitors on Env-dependent cell-cell fusion are specific to inhibition of Abl . To confirm these results using virus particles with relevant levels of virus-associated glycoprotein , we used a virus-dependent cell-cell fusion assay based on the ability of virus particles to bridge two cells and allow transfer of cytoplasmic contents , and we also used the Vpr-BlaM assay described above [4] , [10] . For the virus-dependent cell-cell fusion assay we used two populations of U87 . CD4 . CCR5 cells , one expressing the T7 polymerase and the other expressing the β-galactosidase ( β-gal ) gene under the T7 promoter . Both populations were incubated with inhibitors for 1 h prior to 3 h incubation with R5 virus HIVYU2 . In this assay , controls included untreated and inhibitor treated cells that were not incubated with virus , the CCR5 inhibitor TAK-779 , and T-20 which blocks entry by inhibiting the conformational change in HIV-1 gp41 required for fusion [17] . R5 Virus-dependent cell-cell fusion was reduced by an average of 94±3% in cells treated with IMB , DAS , and NIL compared to cells treated with DMSO alone , and treatment with TAK-779 and T-20 completely inhibited fusion ( Figure 2F ) . Treatment of U87 . CD4 . CXCR4 cells incubated with the X4 virus HIVHXB2 with AMD3100 IMB , NIL , and DAS decreased virus-dependent cell-cell fusion by 88±7% , 98 . 6±1% , 87±5% , and 96±17% , respectively ( Figure S3B ) . For the Vpr-BlaM assay , TZM-BL cells were treated with 1 µM AMD3100 , 1 µM TAK-779 , 10 µM IMB , 500 nM NIL and 150 nM DAS for 1 hr prior to and during the 90 min incubation with HIV-1 Vpr-BlaM viruses expressing R5 and X4-tropic Env . AMD3100 treatment decreased X4-Vpr-BlaM activity by 84±1% , but had no effect on R5-Vpr-BlaM activity . TAK-779 treatment decreased R5-Vpr-BlaM activity by an average of 89±2% , but had no effect on X4-Vpr-BlaM activity , as expected . However , treatment of TZM-BL cells with IMB , NIL , and DAS decreased virus-cell fusion by an average of 81±4% , 89±5% , and 90±1% , respectively , for both HIV-1 R5 and X4 Env subtypes ( Figure 2G ) . These results together with the results of the Env-dependent and virus-cell fusion assay demonstrate that Abl kinase is required for HIV-1 entry mediated by CXCR4 and CCR5 . To determine whether the Wave2 signaling complex and Abl are required exclusively for HIV-1 entry , or virus-induced fusion and infection in general , we examined infection with HIV-1 versus A-MLV Env ( A-MLV-ENV-HIV-1 ) or VSV-G pseudotyped HIV-1 ( VSV-G-HIV-1 ) using the TZM-BL assay . HIV-1 Env induces pH independent virus-cell fusion to facilitate entry , whereas viruses pseudotyped with VSV-G or A-MLV Env induce pH-dependent clathrin mediated endocytosis or caveolin-mediated endocytosis , respectively [22]–[25] . TZM-BL cells , a derivative of HeLa cells that express CD4 , CCR5 , CXCR4 , and luciferase ( luc ) under the control of the HIV-1 LTR , were pretreated with the 10 µM IMB , 500 nM NIL and 150 nM DAS for 1 h prior to incubation with virus for 3 h , and a subsequent 24 h incubation with inhibitor only [16] , [17] . The CCR5 inhibitor TAK-779 , the CXCR4 inhibitor AMD3100 , and ammonium chloride ( NH4Cl ) which inhibits endosomal acidification required for VSV-G mediated entry , were included as controls [22] , [23] , [26] . The top two panels of Figure 3A shows that treatment with IMB , NIL , and DAS decreased infection with R5 HIVYU2 virus and X4 HIVHXB2 virus by an average of 91±7% , 88±4% , and 91±5% , respectively , comparable to the reductions observed with Env-dependent cell-cell fusion , virus-dependent cell-cell fusion and virus-cell fusion ( Figure 2 ) . The Abl kinase inhibitors had no effect on infection of TZM-BL cells with A-MLV-ENV-HIV-1 or VSV-G-HIV-1 , but treatment of cells with NH4Cl blocked infection with VSV-G-HIV-1 as expected ( Figure 3A , bottom two panels ) . These data show that Abl-kinase inhibitors were able to block HIV-1 Env-mediated fusion specifically and had no effect on infection via pH-dependent clathrin-mediated or caveolin-mediated endocytosis , and post-entry steps were not affected by these inhibitors . To test the effect of Wave2 complex targeted siRNAs on infection , TZM-BL cells were transfected with 200 nM control siRNA or siRNA directed towards Tiam-1 , Trio , Abl , IRSp53 , Wave2 and Arp3 . These cells were incubated with virus for 3 h , and media alone for 24 h . The decreased levels of HIV-1YU2 and HIV-1HXB2 infection of TZM-BL cells expressing siRNA targeted to Tiam-1 , Abl , IRSp53 , Wave2 , and Arp3 were comparable to levels of Env-mediated cell fusion with U87 . CD4 . CCR5 cells expressing these siRNAs , whereas siRNA to Trio had no effect ( Figure 3B , top two panels ) . Steady state levels of target proteins in cells expressing targeted siRNAs were decreased to similar levels as in U87 cells ( Figure 1C and data not shown ) . Infection of TZM-BL cells with A-MLV-ENV-HIV-1 or VSV-G-HIV-1 was not affected by expression of the targeted siRNAs , suggesting that Tiam-1 , Abl , IRSp53 , Wave2 , and Arp3 are required for HIV-1 Env-mediated entry and are not necessary for post-fusion steps in the virus life cycle ( Figure 3B , bottom 2 panels ) . HIV-1 Env-induced fusion , and release of the viral capsid into the cytosol is a multistep process . First , gp120 binds to CD4 inducing conformational changes in gp120 , and actin cytoskeletal rearrangements in the target membrane that bring the coreceptor CCR5 or CXCR4 into close proximity with CD4 . Next , coreceptor binding to gp120 triggers conformational changes in gp41 to produce a prebundle conformation that inserts into the target cell membrane , allowing lipid mixing or hemifusion , and then pore formation . Additional conformational changes induce formation of the gp41 6-helix-bundle which prevents pore closure and facilitates pore enlargement and full fusion [2] , [27] , [28] . To determine which step ( s ) in the membrane fusion process are blocked by the Abl kinase inhibitors , we examined the effect on infection of membrane curving agents . Oleic acid ( OLA ) , chlorpromazine ( CPZ ) , and trifluoperazine ( TFP ) are lipid analogs that insert into the inner leaflet of the cell membrane . OLA induces negative curvature in the membrane that promotes formation of a hemifusion intermediate ( i . e . lipid mixing ) , but cannot induce pore formation if there is a block at hemifusion . CPZ and TFP are membrane-permeable weak bases that partition into inner leaflets of cell membranes , induce positive curvature , and relieve a block at hemifusion [29]–[32] . To determine the effect of inhibitors and lipid analogs on HIV-1 infection , TZM-BL cells were treated with 1 µM AMD3100 , 1 µM TAK-779 , 10 µM IMB , 500 nM NIL , and 150 nM DAS for 1 h , prior to and during 1 h incubation with no virus , HIVΔENV , R5 HIVYU2 , X4 HIVHXB2 , A-MLV-ENV-HIV-1 , or VSV-G-HIV-1 . After 1 h , cells were treated with CPZ or TFP for 1 min or OLA for 5 min , followed by 2 h incubation with inhibitor and virus , and subsequent 24 h incubation with inhibitor only . Addition of CPZ and TFP to cells treated with Abl kinase inhibitors and infected with HIVYU2 or HIVHXB2 resulted in an 8 fold increase in infection compared to inhibitor treated cells infected in the absence of lipid analogs ( Figure 4A ) , The exogenous cone shaped lipid OLA , which induces negative curvature of the membrane resulting in lipid mixing , had no affect on infection ( Figure 4A ) . TAK-779 mediated inhibition of HIVYU2 infection and AMD3100 mediated inhibition of HIVHXB2 infection was not affected by these lipid analogs . No increase in luc activity was observed with lipid analog treatment of cells infected with HIVΔENV versus no virus , indicating that Env is required to observe an increase in infection ( Figure S6A ) . Treatment of A-MLV-ENV-HIV-1 and VSV-G-HIV-1 infected cells with CPZ and TFP decreased overall infection by 2 fold and had no effect on cells treated with Abl kinase inhibitors , indicating that the increase in HIV-1 infection observed with Abl kinase inhibitor treated cells was specific ( Figure 4A , lower panels ) . CPZ also partially reversed the inhibitory effects of nilotinib as measured by the Vpr-BlaM assay ( Figure S6B ) . Similar increases in virus-dependent cell-cell fusion were observed when U87 . CD4 . CCR5 cells were treated with inhibitors and lipid analogs and HIVYU2 mediated fusion was measured after 3 h ( Figure S6C ) . Cells were also incubated with the lipid analogs in the absence of HIVYU2 to account for the effects of these agents on the cells and on T7 polymerase activity . Addition of OLA did not increase fusion in cells treated with any of the inhibitors ( Figure S6B ) . To confirm the results obtained with the Abl kinase inhibitors we incubated TZM-BL cells transfected with Tiam-1 , Abl , Rac , IRSp53 , Wave2 , and Arp3 targeted siRNA , for 1 h with no virus , HIVΔENV , R5 HIVYU2 , or X4 HIVHXB2 . After 1 h cells were treated with CPZ for 1 min or OLA for 5 min , followed by 2 h incubation with virus , and subsequent 24 h incubation with media alone . As with the Abl kinase inhibitors , treatment of siRNA transfected cells with CPZ increased infection by an average 8 . 4 fold compared to untreated cells , and OLA had no effect ( Figure 4B ) . These results suggest that inhibition of Tiam-1 , Abl , Rac , IRSp53 , Wave2 or Arp3 arrests fusion at hemifusion , preventing pore formation , pore enlargement and content mixing . To confirm that Abl kinase inhibitors cause arrest at hemifusion , we used a modification of a fusion assay described previously [33] . CHO-K1 cells that lack expression of the lipid ganglioside GM1 , were engineered to express GFP and the HIV-1ADA ( R5 ) Env protein . U87 . CD4 . CCR5 cells were used as the target cell , and lipid mixing was detected when GM1 , detected by a TRITC-conjugated form of cholera toxin β-subunit ( CTX ) , was transferred from the target cell to CHO-K1-GFP cells . Complete fusion is detected when cells express GM1 , GFP , and are multinucleated . Quantification was performed for three independent experiments and the percentage of hemifused GFP+ , GM1+ cells with single nuclei and the percentage of multinucleated fully fused cells was enumerated for 68 cells from each condition ( Figure 5 , S7 , and Table S1 ) There were 83 . 1±10 . 9% hemifused cells with IMB-treated cells mixed with HIVADA-expressing CHO-K1 cells ( Figure 5 ) , compared to DMSO treated cells with 22 . 3±4 . 9% hemifused cells and 75 . 5±6 . 2% fully fused cells . With no HIV-1 Env or with the addition of TAK-779 there was little or no hemifusion or full fusion ( Figure 5 ) . To demonstrate the effects of the lipid analog CPZ on HIV-1 Env mediated cell-cell fusion and to observe the effect of CPZ and the Abl kinase inhibitors on A-MLV Env or VSV-G induced cell-cell fusion we treated U87 . CD4 . CCR5 cells with DMSO , TAK-779 , or IMB for 1 hr prior to incubation with CHO-K1 cells expressing no Env , HIVADA , A-MLV Env or VSV-G for 1 hr . After 1 h cells were treated with CPZ for 1 min and OLA for 5 min , then washed and incubated with inhibitor for an additional 2 h prior to fixation and GM1 staining . Incubation of IMB treated cells with HIVADA and CPZ promoted the transition from hemifusion to full fusion as expected ( Figure S8 ) . Fusion of A-MLV Env and VSV-G Env expressing cells with U87 . CD4 . CCR5 cells was unaffected by treatment with IMB or CPZ ( Figure S9 ) and all Env-mediated fusion was unaffected by OLA treatment ( data not shown ) . These results confirm that Abl kinase activity is required at a post-hemifusion step for HIV-1 Env mediated fusion and entry .
Dynamic regulation of the actin cytoskeleton is required for fusion of biological membranes . Multiple reports have demonstrated that actin remodeling is required for HIV-1 mediated fusion and entry [4] , [5] , [10] , [11] , [34]–[36] . Some studies showed that treatment of target cells expressing physiologically relevant levels of receptor and coreceptor with the actin filament capping drug cytochalasin D prevented the formation of the gp120-CD4-coreceptor complex [35] , [37] , [38] . Another more recent study , demonstrated a role for CD4 and coreceptor-mediated filamin-A interactions in receptor clustering that is dependent on RhoA and ROCK mediated phosphorylation of ADF/cofilin [34] . Previous work from our lab with the actin filament stabilizing drug jasplakinolide and the actin monomer sequestering drug latrunculin A ( LA ) suggested a role for actin remodeling at a post binding step in fusion [4] . To further substantiate the role of actin polymerization in HIV-1 entry , we treated cells with 1 µM LA and 5 µM latrunculin B ( LB ) . Both drugs blocked HIV-1 fusion for multiple cell types , as measured by the Env-dependent cell-cell fusion assay , the virus-dependent cell-cell fusion assay , the virus-cell fusion assay , and infection ( Figure S10 ) . Our previous data demonstrated that the GTPase Rac was activated by HIV-1 Env ligation of CCR5 , resulting in membrane ruffles and lamellipodia in the target cell membrane . Inhibition of this activation by dominant negative Rac or by a Rac GEF inhibitor completely abolished Env-dependent cell-cell fusion , virus dependent cell-cell fusion and infection [4] , [5] , [10] , [11] . Our lab went on to show that Env-induced Rac activation is mediated by Gαq and its downstream effectors , including Ras . Other studies showed that Ras promotes Rac activation via direct interaction with Tiam-1 , or by phosphatidylinositol 3-kinase ( PI3K ) -mediated activation of Tiam-1 [39] . Env-dependent Rac activation likely occurs through the first mechanism , since treatment of target cells with PI3K inhibitors had no effect on Env-dependent cell-cell fusion [40] . The nonreceptor tyrosine kinase , Abl , modulates actin upstream and downstream of Rac [41] , [42] . In the current study , we used siRNAs and specific inhibitors to show that the activity of Abl kinase is required both upstream and downstream of Rac for Env-induced membrane fusion . Upstream of Rac , Abl phosphorylation of the Ras GEF complex promotes the activity of the Rac GEF Tiam-1 , which was shown in the current study to be required for HIV-1 fusion . Downstream of Rac , Abl promotes phosphorylation and activation of Wave2 and its interaction with the Arp2/3 complex , events also demonstrated here to be critical for HIV-1 infection , but not VSV-G or A-MLV Env-mediated infection . These results suggest that these signaling mediators are important for HIV-1 Env mediated entry , are not necessary for pH dependent clathrin or caveolin-mediated endocytosis , and are not required at post-entry steps in the virus life cycle . There is some conflict in the literature as to the location and mechanism of virus cell fusion . A recent report used microscopic imaging to track HIV-1 Env-pseudotyped MLV virus particles and observed virus-membrane fusion in endosomes [3] . This study also showed that virus-cell fusion and infection were inhibited in the presence of the dynamin inhibitor dynasore ( DYN ) which is known to block both clathrin and caveolin-mediated endocytosis [3] . The results in our current study suggest that fusion is occurring via a mechanism that is distinct from that of VSV ( clathrin-mediated endocytosis ) or A-MLV ( caveolin-mediated endocytosis ) . In order to address this conundrum , we treated cells with the dynamin inhibitor DYN , and then used these cells for the Env-dependent cell-cell fusion assay , the virus-dependent cell-cell fusion assay , the virus-cell fusion assay and the TZM-BL infection assay . DYN treatment decreased HIV-1 Env-mediated infection and virus-cell fusion by an average of 58±7% and 50±3% respectively ( Figure S10 and Figure S11 ) . However treatment with DYN decreased A-MLV-Env-HIV-1 infection and VSV-G-HIV-1 infection by 75±5% and 89±1% respectively , showing that the affect on HIV-1 Env-mediated infection was not as significant ( Figure S11 ) . DYN treatment also decreased Env-dependent cell-cell fusion and virus-dependent cell-cell fusion by 53±8% and 50±10% , respectively which was unexpected since these assays both measure cell-cell plasma membrane fusion . Dynamins are a group of large GTPases that are involved in multiple processes in addition to endocytosis , such as vesicle transport , cytokinesis , organelle division , cell movement and cell signaling [43]–[45] . Therefore , the inhibition observed with the dynamin inhibitor DYN could be due to nonspecific effects on cellular processes . In support of this conclusion , a recent study used the Rev-dependent indicator cell line Rev-CEM to study the effects of DYN on HIV-1 replication and VSV-G-HIV-1 infection [44] . Using this assay they observed a dosage dependent decrease in VSV-G-HIV-1 infection with DYN treatment but did not see any decrease in HIV-1 infection [44] . These results as well as the results in Figure 3 show a clear distinction between HIV-1 Env-mediated entry and VSV-G- and A-MLV-mediated entry . The current study also showed that the block in fusion caused by inhibition of Tiam-1 , Abl , Rac , IRSp53 , Wave2 and Arp3 occurs after hemifusion and before cytoplasmic mixing . This conclusion was based on the 1 ) confocal microscopy demonstration that addition of IMB to the fusion reaction allowed membrane but not cytoplasmic mixing , and 2 ) observation that lipid analogs that overcome a block at hemifusion overcame inhibition of HIV-1 virus dependent cell fusion , virus-cell fusion and infection caused by Abl kinase inhibitors and siRNA expression . These results support a model whereby HIV-1 Env binding to CCR5 stimulates activation of Gαq resulting in activation of Rac and activated Rac interacts with IRSp53 . IRSp53 promotes Rac activation of the Wave2 complex , which is also activated by Abl , and activated Wave2 induces subsequent activation of Arp2/3-mediated actin rearrangements which facilitate pore formation , pore enlargement , and entry of HIV-1 . Many microbial pathogens depend on Abl family kinases to mediate efficient infection of their targeted host , including Shigella flexneri , enteropathogenic Escherichia coli , Helicobacter pylori , Anaplasma phagocytophilum , coxsackievirus , poxvirus , and murine AIDS virus . Abl kinases are involved in pathogen entry , intracellular movement , and exit from target cells; proliferation of target cells; and phosphorylation of microbial effectors . Many of these processes involve reorganization of the target cell actin cytoskeleton and depend on the same signaling pathways as HIV-1 [4] , [5] , [46] . Discovery of these signaling mediators as fundamental components of microbial pathogenesis provides new targets for therapeutic intervention . The clinical application of IMB , NIL , and DAS , which block deregulated Abl kinases in leukemia patients , demonstrate that inhibition in vivo is possible with manageable side effects [19] , [20] . In addition IMB has been shown to be an effective inhibitor of anti-apoptotic pathways induced by HIV-1 in macrophages [47] . Most current antiviral therapies target viral proteins and mutation of the virus leads to therapy resistance . Therefore , using inhibitors that target host signaling proteins essential for HIV-1 entry may be an efficient new strategy for treatment of infected patients .
U87 . CD4 . CCR5 cells are astroglioma cells expressing CD4 , CCR5-GFP or HA-CCR5 . U87 . CD4 . CXCR4 cells are astroglioma cells expressing CD4 and CXCR4-GFP . CHO-K1 cells ( ATCC ) were grown in F-12K media with 10% serum and other cells maintained as described [48] . pMSCVneo-WT , Y253F , and T315I Bcr-Abl were gifts from Dr . R . Van Etten [21] . The siRNA resistant mutations were generated in Arp3 based on sequences obtained from Santa Cruz Biotechnology , Inc ( SCBT , Santa Cruz , CA , ) , by PCR-mediated mutagenesis of a sub fragment that was sequenced to confirm the presence of mutations before sub cloning into the corresponding cDNA . WT and mutant cDNAs were cloned in pcDNA3 . 1+zeo for expression by transduction . IMB , NIL , and DAS were from LC Laboratories and were used at 10 uM , 500 nM , and 300 nM respectively unless indicated; CPZ ( 0 . 5 mM ) , TFP ( 0 . 3 mM ) , OA ( 100 nM ) , OLA ( 50 uM ) and NH4Cl ( 50 mM ) were from Sigma; TAK-779 ( 1 uM ) , and T-20 ( 10 ug/ml ) were from the AIDS Research and Reference Reagent Program . The control siRNA constructs ( non-targeting 20–25 nt siRNA designed as a negative control ) , the siRNA constructs and antibodies used for Western blots were from SCBT [5] . The siRNA constructs were transfected using GeneEraser siRNA Transfection Reagent or Lipofectamine RNAiMAX Transfection Reagent according to the manufacturer's instructions ( Stratagene , La Jolla , CA , Invitrogen , Carlsbad , CA ) . Wild-type ( WT ) vaccinia ( WR strain ) and recombinant vaccinia viruses expressing β-galactosidase ( vCB21R ) , T7 polymerase ( vPT7-3 ) , constitutively active Rac GTPase ( vRacV12 ) , or HIV-1 Env proteins were described [48] . HIV with R5 YU2 or X4 HXB2 Env in HIVNL4-3 backbone were generated from 293T cells; some were pseudotyped with amphotropic murine leukemia virus ( MLV ) or vesicular stomatitis virus ( VSV ) glycoproteins [5] . TZM-BL assays were performed as described [5] . For the BlaM assay pseudoviruses were produced by co-transfecting 293T cells with HIVNL4-3ΔVpr expressing YU2 , ADA , or HXB2 Env and BlaM-Vpr expressing pMM310 vector . Transfected 293T cell supernatants were harvested 48 h postlipofection , filtered , and assayed for p24 antigen content by enzyme-linked immunosorbent assay . Viruses were resuspended in culture media , aliquoted and stored at −80°C . TZM-bl cells were serum starved for 24–36 h then plated ( 4×104 cells/well ) in 96-well plates in complete media overnight . Cells were treated with indicated concentrations of inhibitors for 1 hr prior to and during 90 min incubation with DEAE-dextran ( 20 µg/ml ) alone or DEAE-dextran ( 20 µg/ml ) and 150 ng p24 HIVYU2Vpr-BlaM , HIVADAVpr-BlaM , or HIVHXB2Vpr-BlaM . After 90 min virus and media were aspirated off cells and 100 ul 1X Lysis and Detection Solution was added to wells ( LyticBlazer-BODIPY FL , Invitrogen ) . The plate was incubated at room temperature in the dark overnight . The BlaM activity was quantified using TECAN fluorescence plate reader ( Tecan , Switzerland ) . The extent of virus-cell fusion was measured with excitation centered at 485 nm and emission centered at 535 nm . The green signal for samples incubated with no inhibitors or inhibitors and no virus was subtracted as background from their respective virus treated samples . TZM-BL cells were serum starved for 12–24 h then plated overnight in complete media in 96 well plate at 2×104 cells per well . Cells were treated for 1 h with indicated concentrations of inhibitors prior to addition of media alone or 150 ng p24 of HIVYU2 HIVHXB2 or VSVG or A-MLV-pseudotyped HIV in the presence of 20 ug/ml DEAE-dextran for 3 h at 37°C . After 3 h cells were washed 3 times with PBS and inhibitors were added in fresh media . Following a 24 h incubation cells were lysed and luciferase ( luc ) units determined . Infected wells and uninfected wells with inhibitor were compared to wells with no inhibitor . For the TZM-Bl assay with lipid analogs serum starved TZM-BL cells were treated with inhibitors for 1 h , then 150 ng of indicated virus was added for 1 h prior to treatment with CPZ or TFP for 1 min or OLA for 5 min . Cells were washed three times with PBS and virus and inhibitors were added back . After 2 h cells were washed two times with PBS and incubated in inhibitor overnight and luc activities were measured . PBMCs that were isolated and stimulated as previously described [5] . They were plated at 5×105 cells per well in 96 well plate and were treated with 10 µM IMB , 250 nM NIL , or 75 nM DAS for 1 h prior to addition of 150 ng p24 of HIVHXB2 in the presence of 20 ug/ml DEAE-dextran for 3 h at 37°C . After 3 h cells were washed three times with PBS and incubated in inhibitor for 24 h . Inhibitors were added back at the same concentration every 24 h for three weeks . 100 ul of supernatant was collected every fourth day and all samples were assayed for p24 antigen content by enzyme-linked immunosorbent assay . Two separate plates were set up under the exact same conditions and one plate was used for p24 measurement and the other was incubated with 20 ul cell viability substrate per 100 ul of sample ( Promega , Madison , WI ) . Envelope-mediated and virus-dependent fusion assays were described . Average fusion compared to untreated control reactions were detected by β-galactosidase activity ± standard deviation [5] . To account for any effect of inhibitors on vaccinia virus infection and/or on T7 polymerase function , vCB21R and vPT7-3 co-infected cells were similarly treated with inhibitors . Concentration curves were performed with all of the inhibitors to determine the concentration that resulted in the maximum decrease in fusion without altering vaccinia virus infection or T7 polymerase activity . Hemifusion assays were performed with 2×106 CHO-K1 cells nucleofected with a GFP expression plasmid , and after 24 h infected with vaccinia virus expressing HIVADA Env or no Env . After 16 h , 4×105 U87 . CD4 . CCR5 . HA cells were added for 3 h , fixed with paraformaldehyde , stained with TRITC-conjugated CTX-555 ( EMD ) , and analyzed on a 510 Meta LSM confocal microscope . Fusion and infectivity results were compared using a two-tailed t-test . All p values , unless indicated , were <0 . 03 .
|
Patients infected with HIV-1 are currently treated with highly active antiretroviral therapy ( HAART ) that efficiently suppresses the virus but does not cure the infection . HIV-1 envelope activates Rac-mediated actin cytoskeleton rearrangements in the target cell that promote membrane fusion and entry . We discovered that these rearrangements require activation of the actin polymerization machinery including the tyrosine kinase Abl . We also showed that Abl kinase inhibitors imatinib , nilotinib , and dasatinib , current drug therapies for chronic myeloid leukemia , block HIV-1 entry and infection . These results suggest that these inhibitors might be appropriate drugs for treatment of HIV-1 . This strategy of using inhibitors that disable host signaling proteins rather than viral proteins , essential for pathogen survival , may have a general efficacy in developing drugs to combat HIV-1 and other pathogens that acquire drug resistance .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/host",
"invasion",
"and",
"cell",
"entry",
"virology/immunodeficiency",
"viruses",
"virology/antivirals,",
"including",
"modes",
"of",
"action",
"and",
"resistance",
"virology"
] |
2010
|
Role of Abl Kinase and the Wave2 Signaling Complex in HIV-1 Entry at a Post-Hemifusion Step
|
Nematodes of the genus Caenorhabditis enter a developmental diapause state after hatching in the absence of food . To better understand the relative contributions of distinct regulatory modalities to gene expression changes associated with this developmental transition , we characterized genome-wide changes in mRNA abundance and translational efficiency associated with L1 diapause exit in four species using ribosome profiling and mRNA-seq . We found a strong tendency for translational regulation and mRNA abundance processes to act synergistically , together effecting a dramatic remodeling of the gene expression program . While gene-specific differences were observed between species , overall translational dynamics were broadly and functionally conserved . A striking , conserved feature of the response was strong translational suppression of ribosomal protein production during L1 diapause , followed by activation upon resumed development . On a global scale , ribosome footprint abundance changes showed greater similarity between species than changes in mRNA abundance , illustrating a substantial and genome-wide contribution of translational regulation to evolutionary maintenance of stable gene expression .
Animals of diverse genera react to unfavorable growth conditions by entering developmentally arrested states known as diapause [1] , [2] . Nematodes of the genus Caenorhabditis can enter and exit diapause at several developmental time points , allowing populations to reproduce through boom and bust cycles of nutrient availability . At least four specific programs of developmental arrest and resumption have been identified , each accompanied by unique morphological and gene-regulatory responses [3]–[6] . In newly-hatched Caenorhabditids , entry and exit from L1 diapause can be controlled in large and synchronous populations by depriving or providing food . C . elegans L1 diapause responses have been well characterized at the level of mRNA biogenesis , with developmental state changes associated with substantial transcriptional changes , the accumulation of RNA polymerase at gene promoters , and alternative splicing [7] , [8] . Translational regulation is also expected to contribute significantly to major developmental transitions . We selected four nematode species for investigation of the translation and mRNA-level gene regulatory program associated with L1 diapause exit: two hermaphroditic species , Caenorhabditis elegans and C . briggsae , and two gonochoristic ( male/female ) species , C . remanei and C . brenneri ( Fig . 1A ) . These four species exhibit highly similar morphologies and developmental timing despite significant genomic sequence divergence [9] . For each species , we applied mRNA-seq and ribosome profiling [10] to populations of arrested L1 diapause larvae and to populations harvested three hours after the first food encounter that signals the animals to exit diapause and commence development ( Fig . 1B; sequencing summary statistics: Table S1 ) . Samples from the two conditions are denoted “diapause” or “developing” throughout . All sequence reads and processed count data are available from the Gene Expression Omnibus ( http://www . ncbi . nlm . nih . gov/geo ) via accession number GSE48140 . Our data provide gene-by-gene measurements at two levels of expression: ( i ) mRNA-seq data measures the relative steady-state abundances of mRNAs in the transcriptome ( abundances are products of mRNA biogenesis and decay ) , ( ii ) ribosome profiling allows the counting of ribosome-protected fragments ( RPFs ) derived from each mRNA , with each fragment corresponding to one active ribosome and thus to an instance of peptide synthesis [10] . For brevity , we use “translatome” [11] to describe the population of mRNA fragments undergoing translation at a given time point , with the relative abundances captured by RPF counts . We emphasize that the abundance of RPFs reflects the input of two biological parameters: steady-state mRNA abundance and ribosome binding ( translation efficiency ) . As diapause entails a substantial decrease in overall translational activity [12] , it is additionally important to note that all measurements we infer from the data are relative . Changes in RPF levels between L1 diapause and developing states thus represent relative changes in the commitment of available translational resources between the two conditions .
As a starting point for our analysis , we sought to examine the general character of the gene expression changes associated with the transition from L1 diapause to development . mRNA-seq measurements indicated that diapause exit triggered substantial remodeling of transcriptome composition in the four species , similar to the response previously described in C . elegans [7] , with thousands of differentially expressed transcripts and expression changes spanning three orders of magnitude ( Fig . 2A , Fig . S1 ) . As expected , transcriptome changes ( changes in steady-state mRNA levels ) were accompanied by dramatic shifts in the composition of the translatomes ( Fig . 2B , Fig . S1 ) . The combination of transcriptome and translatome data from common samples allowed us to compare of the relative magnitude of the two levels of response . Fig . 2C shows a comparison of the frequency distributions of mRNA and RPF abundance changes for the four species . In each case , we observed a significantly broader distribution for RPF changes , consistent with a regulatory response in which changes in translation efficiencies and mRNA levels taken together constitute a larger magnitude overall response than that seen at the level of mRNA abundance alone ( Fig . 2C , p<2e-16 for all comparisons ) . Overall , between the four species we found that ∼15–30% of well-expressed transcripts showed a >2-fold change in relative mRNA abundance , and ∼30–45% of transcripts changed >2-fold in relative RPF abundance . The transcriptome-wide tendency for RPF level changes to exceed mRNA changes could in principle have resulted from ( i ) translational changes that act synergistically with and amplify mRNA abundance changes , or ( ii ) translational changes of large magnitude whose directions are somewhat or predominantly independent of the direction of mRNA abundance changes . While the first scenario may seem most likely from a cellular energetic perspective , genome-wide studies in a variety of systems have revealed varying degrees of coordination ( and lack of coordination ) of transcriptome and translational responses to a range of stimuli [11] , [13]–[15] . To determine which of these scenarios best match our data , we compared changes in RPF level to changes in mRNA level on a gene-by-gene basis . As RPF levels represent the combined input of processes affecting mRNA abundance ( e . g . , transcription , decay ) and translational regulation , we reasoned that , for a given transcript , a change in RPF level in the same direction and of a greater magnitude than the change in mRNA level represents a case in which translation and mRNA abundance processes act in concert ( a “concordant” change; equivalent to “homodirectional” in [15] ) . Conversely , if the change in RPF level is of lesser magnitude or in the opposite direction to the change in mRNA level , this represents a situation in which translational regulation is acting in opposition to mRNA abundance processes ( a “discordant” change; Fig . 3A ) . Comparing transcript-wise changes in mRNA and RPF levels from our data , we found that concordant changes were overwhelmingly favored over discordant changes in the four species by a ratio of 2 . 8–3 . 7∶1 ( Fig . 3B–E ) . Thus during the feeding-induced transition from L1 diapause to active development , these four species apparently utilize a shared regulatory logic: an amplified global gene expression response produced by synergistic changes in mRNA abundance and translational control . We next asked to what degree expression changes were similar between species at the gene level . Studies of gene expression conservation have largely focused on mRNA levels [16]–[20] , though several groups have reported superior conservation of orthologous protein abundances , implying an important role for translational or post-translational mechanisms in maintaining optimal gene expression during evolution [21] , [22] . We began by comparing feeding-induced changes in mRNA abundance for ortholog pairs identified between the four species . mRNA abundance changes for well-expressed transcripts correlated strongly in all pair-wise species comparisons ( Fig . 4A , Fig . S2 ) . Observed correlation coefficients were highly significant and ranged from 0 . 63 to 0 . 74 ( Spearman's rho , Fig . 4B , E ) . We next examined the between-species correlation of changes in RPF levels . For each pair-wise species comparison , we observed significantly stronger correlations for changes in RPF abundance than for changes in mRNA abundance , with correlation coefficients in the range of 0 . 76 to 0 . 85 ( Fig . 4C–E , Fig . S2 ) . To complement pair-wise correlations , we examined expression changes within ortholog groups for which an ortholog could be assigned in each of the four species ( the four genes together constituting a “four-way” ortholog group ) . We found that overall expression divergence within four-way ortholog groups was significantly greater for mRNA abundance changes than for RPF changes ( Fig . 4F , p<2e-16 ) , and that this difference disappeared after randomly shuffling ortholog groupings ( Fig . S3 ) . Together , these results demonstrate that , for the L1 diapause program in these nematode species , comparisons accounting for translational regulation reveal a greater level of overall gene expression conservation than is observed at the level of mRNA abundance alone . A significant resulting inference is that alterations in translational control and processes affecting mRNA abundance can compensate for one another during evolution to achieve stable protein expression . The substantial observed conservation underscored the functional significance of expression changes during L1 diapause exit . We therefore sought to compare general properties of gene expression in the arrested and developing states , and specifically to identify features that distinguished the two states . Principal components analysis ( PCA ) is a technique that takes data featuring many variables , such as gene expression data , and extracts a series of linear combinations of the individual variables ( called a “principal component” ) that explains a substantial fraction of the variance between samples , with each successive component explaining less variance than the previous component . We applied PCA separately to ortholog abundance data from mRNA-seq and RPF datasets , including all arrested and developing samples from the four species in the analysis . For mRNA-seq data , the first two principal components provide poor separation of samples by condition or species ( Fig . 5A ) . For RPF data , we observe a clear separation of diapause samples from developing samples on both of the first two principal components ( Fig . 5B , Fig . S4 ) . The clean separation achieved with RPF data suggests that translatomes of animals from different species in the same nutritional/developmental state are more similar than translatomes from animals of the same species in opposite states . For RPF data , the first principle component explains a majority of the between-sample variance ( 86 . 4% ) , and higher values on this component are associated with developing samples . We identified a set of transcripts that were exceptionally highly weighted on the first component . Overlaying these transcripts on RPF fold-change plots revealed that these transcripts largely corresponded , in each species examined , to a cluster of highly-expressed and strongly up-regulated genes independently identified ( by visual inspection ) as a group of genes of interest ( Fig . 5C , Fig . S5 ) . We investigated the identities of the transcripts making up this group and found that the significant majority corresponded to ribosomal protein genes , along with several core translation factors and a small number of additional genes including ubiquitin , heat shock proteins , and the RACK1 homolog ( Table S2 ) . These transcripts also formed a readily-identifiable cluster in mRNA-seq data , but with substantially weaker up-regulation ( Fig . 5C , Fig . S5 ) . Direct comparison of fold-changes for ribosomal proteins in mRNA and RPF data showed that up-regulation was significantly stronger at the RPF level , with average fold-changes >10 compared to ∼2-fold up-regulation at the mRNA level , indicating that the differential representation of these genes in the translatome was primarily due to translational regulation ( Fig . 5D ) . We also found that non-ribosomal components of the core translation apparatus showed a significant trend towards up-regulation ( Fig . 5C , red ) , with contributions from both increased mRNA abundance and translation efficiency ( Fig . 5E ) . Genes of the translation apparatus include many of the most highly-expressed transcripts . The coordinated up-regulation of these transcripts thus constitutes an extraordinary re-allocation of cellular energetic resources . From the raw counts for mapped RPF reads , we infer that , in C . elegans , approximately 3% of ribosomes are bound to a ribosomal protein transcript during L1 diapause . Three hours after feeding , more than 20% of all ribosomes are engaged in translating ribosomal proteins ( Table S3 ) . For the translation apparatus as a whole , this figure jumps from ∼4 . 5% in L1 diapause to nearly 30% in fed animals ( Table S3 ) . This striking change suggests that a central feature of the gene regulatory response to L1 diapause exit is to prioritize existing translational resources to building up the animal's capacity for protein synthesis . In addition to translation genes , several categories of functionally related genes were enriched among transcripts whose RPF levels were significantly higher in developing animals . These include genes involved in promoting growth , development , ribosome biogenesis , the proteasome , and mitochondrial genes ( Table S4 ) . These enrichments were exceptionally consistent between the four species examined ( Fig . S6A ) . In contrast , functional enrichments among genes with higher RPF levels in the L1 diapause state were generally weaker and less-consistent between species than those seen for transcripts up-regulated after feeding ( Table S4 , Fig . S6B ) . Manual inspection revealed a number of interesting genes that showed conserved higher expression during diapause , including nhr-49 ( required for adult reproductive diapause [3]–[6] ) , genes involved in autophagy and dauer formation , superoxide dismutases , and several heat-shock proteins ( Table S5 ) . Ribosomal protein genes were subject to qualitatively similar regulation in each of the four species examined , i . e . , modest increases in mRNA abundance and strong increases in translation efficiency . This led us to ask whether other groups of functionally related genes tended to share regulatory patterns across species . To this end , we defined the “translational component” of regulation as the ratio of the change in translation efficiency to the change in mRNA abundance ( see Materials and Methods and Text S1 ) . Examining the distribution of log ( 2 ) -transformed translational component scores for differentially-expressed transcripts in the four species revealed unimodal distributions in which a majority of transcripts ( 71–91% ) were subject to “mixed” regulation , with mRNA abundance and translational regulation each accounting for at least 25% of the change in RPF level ( Fig . 6A ) . Despite the overall similarities between the species , species-specific differences were evident . Notably , the broader distribution evidenced in C . briggsae indicated a greater tendency for transcripts to be primarily regulated either translationally or at the level of mRNA abundance , while the narrow distribution of C . brenneri suggested a trend toward coordinated regulation ( Fig . 6A ) . While the results indicate a potential quantitative difference in the balance between transcriptional and translational regulation in different species , the coordination between these two regulatory modalities is evident in all four species . We examined the distribution of translational components for transcripts corresponding to up-regulated functional gene categories and found that these categories showed remarkably similar profiles across the four species ( Fig . 6B ) . For example , transcripts corresponding to proteasome components showed minimal contributions from translational regulation in each species , whereas ribosomal components , as demonstrated previously , exhibited large translational components ( Fig . 6B ) . Likewise , spliceosome components favored strong contributions from mRNA accumulation changes in every species , while transcripts of the non-ribosomal translation machinery showed a broad distribution , indicating varied contributions from mRNA abundance and translation for this category . The between-species similarity of the relative contributions of mRNA abundance and translational control to regulation of functionally related transcripts suggests that the transcriptional and translational control networks underlying these changes may also be conserved in these species . Our results point to a key role for translational control in the transition from L1 diapause to active development in these Caenorhabditis species . Translational regulation affects the expression of thousands of transcripts , and the patterns of regulation are well conserved between species at the genome-wide , functional , and gene level . Highly conserved translational modulation of certain sets of related transcripts , notably the ribosomal protein genes , implies that translational control programs may remain largely intact despite significant genome sequence divergence . A recent study reported the persistence of a pool of translationally repressed ribosomal protein mRNAs in yeast undergoing glucose starvation [13] , suggesting that translational suppression of this class of genes may be a somewhat generalized feature of eukaryotic stress responses . In summary , we describe a system in which evolutionarily diverged species maintain a common program of mRNA abundance and translational efficiency changes that cooperatively drive the dynamic reallocation of gene expression resources to traverse a shared developmental and environmental transition .
Strains were obtained from the Caenorhabditis Genetics Center: C . elegans N2 bristol , C . briggsae AF16 , C . remanei PB4641 , C . brenneri PB2801 . Embryos were hatched in sterile S-complete liquid media , starved for 24 hours , and half were supplied with E . coli HB101 . Samples were frozen in liquid nitrogen after 24 hours of starvation and after three hours of feeding . Three full biological replicates were prepared for each species . mRNA-seq and ribosome profiling were carried out as described in [23] , [24] , with modifications as described in Text S1 . Sequencing was performed on Illumina's HiSeq 2000 machine . Raw sequence reads were trimmed of adaptor sequence and mapped using Bowtie 0 . 12 . 7 [25] to the appropriate species' genomes and coding sequence with genomic flanking sequence , and screened for quality . For between-species comparisons , count normalization was performed with the EdgeR package [26] and orthologs were assigned using inParanoid [27] . Differential expression was determined using the DESeq package [28] . All additional analysis was carried out with custom Perl scripts and using the R computing environment [29] . Figures were created with R [30] , [31] . Expression divergence for four-way orthologs was calculated by first normalizing log fold changes for each species by mean and standard deviation , then calculating each pair-wise species-species difference and taking the mean of the resulting differences . Principal components analysis was carried out using the prcomp function in R . Ontology analysis was performed using the web-based DAVID knowledge tool [32] . A more extensive description of the methods can be found in Text S1 .
|
Working with a set of four related animal species , we have studied a conserved developmental and metabolic transition at the level of protein production and regulation of RNA levels . Strikingly , regulatory effects at the level of RNA accumulation and protein synthesis act together to achieve the observed metabolic shift . In addition to a general conservation of the underlying basis for the regulation of individual genes , alterations of these two processes—mRNA production and protein synthesis—can compensate for one another during evolution to maintain stable amounts of functional gene products . A salient feature of the observed regulation was the storage of idle mRNAs encoding key members of the protein synthesis machinery during metabolic arrest ( diapause ) . Maintenance of this pool facilitates re-activation upon feeding , with the rapid regeneration of protein synthesis capacity an early and critical function during adaptation to a major metabolic shift .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
Conserved Translatome Remodeling in Nematode Species Executing a Shared Developmental Transition
|
Protein C inhibitor ( PCI ) is a heparin-binding serine proteinase inhibitor belonging to the family of serpin proteins . Here we describe that PCI exerts broad antimicrobial activity against bacterial pathogens . This ability is mediated by the interaction of PCI with lipid membranes , which subsequently leads to their permeabilization . As shown by negative staining electron microscopy , treatment of Escherichia coli or Streptococcus pyogenes bacteria with PCI triggers membrane disruption followed by the efflux of bacterial cytosolic contents and bacterial killing . The antimicrobial activity of PCI is located to the heparin-binding site of the protein and a peptide spanning this region was found to mimic the antimicrobial activity of PCI , without causing lysis or membrane destruction of eukaryotic cells . Finally , we show that platelets can assemble PCI on their surface upon activation . As platelets are recruited to the site of a bacterial infection , these results may explain our finding that PCI levels are increased in tissue biopsies from patients suffering from necrotizing fasciitis caused by S . pyogenes . Taken together , our data describe a new function for PCI in innate immunity .
Our early response to an invading pathogen relies to a major part on our innate immune system . In order to sense and fight an infection , the human host has developed an arsenal of pattern recognition proteins that interact with so-called pathogen associated molecular patterns or PAMPs ( for a review see [1] ) . Pattern recognition proteins have two major tasks . Some , like toll-like receptors , evoke an inflammatory response , such as the induction of proinflammatory cytokines [2] , while others are involved in the direct killing of the pathogen . For instance , there are scavenger receptors that can act as phagocytic receptors mediating direct non-opsonic uptake of pathogenic microbes and/or their products [3] . However , there are also pattern recognition proteins , such as complement , and antimicrobial peptides that fall into both categories . For example , the anaphylatoxin peptide C3a is a potent chemoattractant for phagocytes , but also has a direct antimicrobial effect [4]; other examples include chemotactic chemokines and neuropeptides [5] , [6] . The mode of the antimicrobial action of these substances is often based on their ability to penetrate the cell wall of the pathogen , which eventually leads to membrane disruption followed by cytosolic leakage and ultimately to the death of the targeted organism . The number of antimicrobial peptides/proteins ( AMPs ) is constantly increasing and today more than 880 have been described [7] . In order to display their activity , many AMPs must first be released from their precursor molecules . Probably one of the best-studied mechanisms is release of LL-37 from cathelicidin hCAP-18 by the action of proteinase 3 [8] . Notably , in some cases an entire protein can exploit its antimicrobial activity without any prior processing . Thus , proteins such as bactericidal/permeability increasing protein , azurocidin and histidine-rich glycoprotein have been reported to function as antimicrobial agents ( for reviews see [9] , [10] , [11] ) . It is noteworthy that many of these proteins have an affinity for heparin . Protein C inhibitor ( PCI ) is a heparin-binding serine proteinase inhibitor [12] . As indicated by its name , PCI was originally reported as an inhibitor of activated protein C , a blood coagulation factor . Later it was reported that PCI is also found , apart from plasma , in tears , saliva , cerebral spinal fluid , breast milk , seminal plasma , and amniotic fluid ( for a review see [13] ) . Recently , it was described that human PCI is efficiently internalized by neutrophils and targeted to the nucleus [14] . Interestingly , the authors also found that internalized PCI promotes phagocytosis of bacteria . As PCI apparently has an affinity for lipids , we set about to analyze its interaction with bacterial membranes . To this end we performed a number of experiments demonstrating for the first time that human PCI is a potent antibacterial reagent .
Previous work has shown that SEK20 , a peptide derived from PCI ( SEKTLRKWLK MFKKRQLELY ) , and LL-37 ( LLGDFFRKSK EKIGKEFKRI VQRIKDFLRN LVPRTES ) , have a broad antimicrobial activity against pathogens such as Candida albicans , Enterococcus faecalis , Escherichia coli , Proteus mirabilis , and Pseudomonas aeruginosa [15] . In addition to these pathogens we find in the present study that SEK20 is also able to kill Bacillus subtilis , Staphylococcus aureus , and Streptococcus pyogenes . In concordance with the previous report , the antimicrobial activity of SEK20 was as efficient as that of LL-37 ( Table 1 ) . Figure S1 shows the effect of SEK20 , LL-37 , and GDK25 ( a control peptide derived from human high molecular weight kininogen [16] ) on E . coli and S . pyogenes bacteria which were used throughout this study . Both pathogens are frequently isolated from patients suffering from severe acute infectious diseases . The broad antimicrobial activity of SEK20 and its positive net-charge ( pI = 10 . 3 ) [15] , suggest that the peptide does not interact with species-specific surface proteins of these pathogens , but rather targets their cell membranes , which is also the point of attack for many other antimicrobial peptides ( for a review see [17] ) . We therefore tested the effect of SEK20 in a permeabilization assay by employing unilamellar anionic liposomes [18] . To this end , liposomes were treated with SEK20 and LL-37 and the subsequent release of carboxyfluorescein was monitored . Figure 1A shows that SEK20 like LL-37 permeabilizes liposomes , suggesting that SEK20 has membrane lytic activity . A negative side effect of some antimicrobial peptides is that they not only act on bacterial or fungal surfaces , but also lyse and ultimately kill eukaryotic cells . We therefore tested the effect of SEK20 on human erythrocytes and found that SEK20 , in contrast to LL-37 , had no hemolytic activity ( Figure 1B ) . Similar results were obtained when measuring the LDH release from HaCaT keratinocytes , where LL-37 demonstrated significant release at higher concentrations , but SEK20 did not ( Figure 1C ) . Taken together , the results show that SEK20 is a potent antimicrobial peptide with a broad specificity , but less toxicity for eukaryotic cells than LL-37 . In order to become active , many antimicrobial peptides such as LL-37 must be released from a precursor molecule [19] . In some cases , however , this processing is not required and the entire protein is antimicrobial by itself , most likely because the antimicrobial region is surface exposed as is the case for histidine-rich glycoprotein and azurocidin [20] , [21] . The three-dimensional structure of PCI has been resolved and a closer examination revealed that a region spanning the SEK20 sequence forms a hairpin loop sticking out at the amino terminal part of the protein [22] . It was therefore tempting to speculate that PCI is by itself antibacterial . To test this , the effect of PCI on Gram-negative ( E . coli ) and Gram-positive ( S . pyogenes ) bacteria was investigated in viable count assays . Figure 2A shows that PCI kills E . coli bacteria very efficiently , while its antimicrobial activity towards S . pyogenes bacteria is slightly reduced when compared with SEK20 or LL-37 ( Figure 2C ) . Notably , the antimicrobial effect of PCI was dose-dependent in both cases ( Figure 2B and 2D ) . Next , we treated PCI with several proteinases ( activated human protein C , factor Xa , plasma kallikrein , thrombin , elastase , cathepsin G , and proteinase 3 ) in order to exclude the possibility that the activity of PCI was achieved upon proteolytic processing of the protein and subsequent release of a SEK20-containing peptide . To this end , Western blot experiments with antibodies against PCI and SEK20 revealed that the protein is resistant to proteolytic degradation and a fragment spanning the SEK20 peptide has not been released when incubated with these enzymes ( data not shown ) . Additional analysis by negative staining electron microscopy showed that protease treatment did not affect PCI's antimicrobial activity ( Figure S2 ) . These findings are in line with reports showing that also other antimicrobial proteins such as azurocidin are resistant to proteolysis [23] . Finally , we performed binding assays with radiolabeled PCI in order to investigate the interaction between the entire PCI molecule and the bacterial surface . We found that the two bacteria strains tested , E . coli ( 12% binding of added radio-labeled protein ) and S . pyogenes ( 27% binding of added radio-labeled protein ) , were able to assemble PCI on their surface . When 125I-PCI bound to streptococci was eluted from the bacterial surface and run on SDS-PAGE followed by auto-radiographic analysis , we did not find any signs of degradation ( data not shown ) , implying that the interaction of PCI with bacteria does not trigger truncation or processing of the protein . Taken together our findings show that the entire PCI molecule has antimicrobial activity and no cleavage of the protein is needed to generate this effect . In the next series of experiments we wished to test whether PCI is able to permeabilize bacteria by the same mechanism as SEK20 . Thus , liposomes were incubated with PCI and the release of carboxyfluorescein was recorded . The results show that PCI , like SEK20 , has the ability cause a concentration dependent release of carboxyfluorescein ( Figure 2E ) . To investigate the effect of PCI on the cell wall of E . coli and S . pyogenes , bacteria were treated with PCI and analyzed by negative staining microscopy . As seen in Figure 3 , this treatment evoked significant membrane destruction , which was followed by the extravasation of cytosolic content as detected by the release of oligonucleotides ( Figure 3I ) . An antimicrobial effect was also seen when PCI was used at 100 nM which reflects its concentration in human plasma ( Figure S3 ) . Thus , our results show that PCI mediates its antimicrobial activity by perforating the bacterial cell membrane followed by the efflux of intracellular material and subsequent death of the bacteria . In humans , PCI is found in many fluids and secretions including plasma , seminal plasma , urine , sweat , saliva , tears , milk , and cerebrospinal fluid ( for a review see [24] ) . Many antimicrobial peptides/proteins do not display their full activity in a physiological environment and require special conditions such as low salt concentration or low pH . To investigate whether this also applies for PCI , we studied the interaction of PCI with AP1 bacteria in human plasma . In contrast to E . coli , S . pyogenes bacteria are not phagocytozed in human blood and therefore we focused on the Streptococci only throughout the rest of this study . In a first series of experiments , we tested whether AP1 bacteria can absorb PCI from human plasma . To this end , bacteria and normal human plasma were incubated for 1 h at room temperature . After a centrifugation step to separate plasma and bacteria and a washing step , bacteria-bound plasma proteins were recovered by an acid wash and subjected to Western blot analysis with anti PCI antibodies . Figure 4 shows that PCI was recruited onto the surface of AP1 bacteria and a depletion of the protein from human plasma was also recorded . Based on these findings we wanted to explore whether PCI can exert its antimicrobial activity in plasma . In order to do so , we compared the growth of AP1 bacteria in normal and PCI-deficient plasma . As seen in Figure 5 , bacterial proliferation is significantly accelerated when PCI has been removed from human plasma , suggesting that PCI is a relevant antimicrobial agent in human blood . PCI is contained within the alpha-granules of platelets and may be released on activation [25] . It has previously been reported that M1 protein from Streptococcus pyogenes can stimulate platelet activation and that activated platelets are present at the site of streptococcal infection [26] . We therefore set about to determine whether PCI can be localized on platelets in response to the physiological activation ( ADP ) or bacterial activation ( M1 protein ) . Unstimulated platelets had background levels of PCI on their surface . Following ADP activation , 11% of the platelet population had PCI on their surface and this was further increased to 17% on treatment with M1 protein ( Figure 6 ) . A proportion of PCI is likely to be released directly into the plasma , however our attempts to quantify platelet derived PCI in plasma failed due to technical limitations . We can therefore not differentiate between PCI released from platelets and PCI acquired from plasma and subsequently bound to platelets . These results do however demonstrate that PCI can be accumulated on the platelet surface during streptococcal infection and this may give rise to an increased PCI concentration at the local site of infection . To test whether PCI levels increase at an infectious site , we analyzed biopsies derived from four patients with necrotizing fasciitis caused by S . pyogenes and a healthy control by immunohistochemistry . Bacterial colonization was demonstrated by employing an antibody against S . pyogenes , which demonstrates the presence of streptococci in the biopsy from the patient with necrotizing fasciitis , but not in the healthy control ( Figure 7 ) . When biopsies were immunostained for PCI , it was found that the protein was enriched at the infectious site , while it was not detectable in the control biopsy . PCI recruitment to the site of infection was localized in cellular infiltrates especially in areas with a lot of bacteria . To investigate these areas by confocal microscopy , biopsies were immunostained with antibodies against S . pyogenes and PCI . Once again the micrographs show that PCI and Streptococci are distributed throughout the cellular infiltrate ( Figure 8A–C ) , and are most commonly co-localized ( Figure 8D ) .
In the present study we show that PCI has a broad antibacterial activity towards many Gram negative and Gram positive microorganisms . Importantly , proteolytic processing of PCI is not required , since the protein is released into plasma in its antibacterial active form . This is in contrast to many other antimicrobial peptides/proteins that have to be generated from a precursor by the action of proteolytic active enzymes . Indeed , our data show that treatment of PCI with a panel of human proteinases including neutrophil-borne elastase , proteinases 3 or cathepsin G , does not diminish the antimicrobial activity of PCI . This feature has a major advantage , since due to neutrophil recruitment and their subsequent degranulation , neutrophil-derived proteinases are enriched at the infected site . Notably , especially in deep tissue infections such as necrotizing fasciitis with severe tissue degradation , extremely high proteolytic activity is recorded which is not only caused by host proteinases , but also by bacterial virulence factors ( for a review see [27] ) . The increased protease activity at the site of infection may lead to an inactivation of antibacterial peptides that are not proteolytic resistant ( for a review see [28] ) . This has been shown for instance for LL-37 , which is probably cleaved by a streptococcal cysteine proteinase at the infected site in patients with severe tissue infections and completely broken down in wound fluids from patients with chronically infected venous ulcers [29] , [30] . Our preliminary results with wound fluids from patients with chronically infected venous ulcers show that unlike LL-37 , PCI is protected from degradation in this proteolytically potent environment ( Malmström , Schmidtchen , and Herwald , unpublished results ) , suggesting that PCI is not only resistant to degradation by the host , but also by bacterial proteinases . Analysis of biopsies from patients suffering from streptococcal necrotizing fasciitis reveal an accumulation of PCI at the infectious site and confocal microscopy studies show that most of the PCI is found co-localized with streptococci . These data confirm our in vitro and ex vivo results showing that PCI has an affinity for bacterial surfaces and they also allow the assumption that the protein exerts antibacterial activity at an infectious site . As PCI is found at low concentrations in plasma ( 4 to 6 µg/ml ) an active transport of the proteins to or its up-regulation at the site of infection is required . Previous work has shown that platelets can infiltrate the infected site of patients suffering from soft tissue infections caused by S . pyogenes [26] . As the alpha-granules of platelets constitute a storage of PCI [25] , we propose a model where PCI is released from platelets that are recruited to the site of infection . Unfortunately , we were not able to distinguish whether PCI was released from the α-granules or from the surface of platelets that have absorbed PCI from plasma . However , considering that platelets contain very little PCI ( 160 ng PCI/2×109 cells ) [25] , it seems more likely that they absorbed it from plasma . At this point we cannot exclude that PCI synthesis is induced at the infectious site and therefore , in ongoing studies we address whether PCI generation can be triggered in different cell lines , including HepG2 cells , upon stimulation with inflammatory substances , such as IL-1β , IL-6 , or TNFα . The affinity of PCI for negatively charged lipids may help explain the mechanism by which the bacterial cell wall is disrupted . As visualized by negative staining electron microscopy , the interaction of PCI with bacteria leads to destruction of the bacterial cell wall and the release of cytosolic content , eventually leading to bacterial killing . Recently , Baumgärtner et al . reported that PCI is avidly internalized by human polymorph nuclear neutrophils ( PMNs ) , and this involves phosphatidyl-ethanolamines [14] . The authors also reported that this interaction enhances the uptake of E . coli bacteria and , thus they suggest an important role for PCI in innate immunity [14] . Considering that the PCI-evoked release of bacterial cytosolic content will cause additionally inflammatory reactions at the infected site , it seems plausible that the host has developed a counteracting mechanism . Uptake of PCI-opsonized bacteria by PMNs before their destruction may lead to decreased inflammatory reactions , while still guaranteeing an efficient killing of the pathogen . Thus a synergistic effect of PCI resulting in bacterial recognition by PCI and their subsequent uptake by phagocytic cells followed by intracellular killing , is an attractive concept that would lead to a clearance of the infection and a dampening of inflammatory responses . Taken together , our studies show a novel function for PCI as an antimicrobial agent against a broad arsenal of bacterial pathogens . This is mediated by the ability of PCI to interact with lipids leading to the efflux of bacterial cytosolic content . When analyzing tissue biopsies we find an accumulation of PCI at the infectious site . These findings suggest an important and novel role of PCI in innate immunity .
The Human Subjects Review Committee of the University of Toronto and of Lund University approved the studies , and written , informed consent from the patients and volunteers was received . Fresh frozen plasma from healthy individuals were obtained from the blood bank at Lund University Hospital , Lund , Sweden , and kept frozen at −80°C until use . Protein C Inhibitor deficient plasma was prepared as described [31] . M1 protein was purified from the supernatant of S . pyogenes MC25 , as previously described [32] . SEK20 ( SEKTLRKWLKMFKKRQLELYL ) and LL-37 ( LLGDFFRKSK EKIGKEFKRI VQRIKDFLRN LVPRTES ) were synthesized by Innovagen AB , Lund , Sweden . The purity ( >95% ) and molecular weight of these peptides was confirmed by mass spectral analysis ( MALDI . TOF Voyager ) . Recombinant protein C inhibitor was purified as previously described [33] . Escherichia coli 37 . 4 and ATCC25922 , Pseudomonas aeruginosa ATCC27853 , Staphylococcus aureus ATCC29213 , Bacillus subtilis ATCC6633 bacterial isolates , and the fungal isolate Candida albicans ATCC90028 were grown as described elsewhere [34] . The Streptococcus pyogenes AP1 ( 40/58 ) strain of the M1 serotype was provided by the WHO ( World Health Organization ) Streptococcal Reference Laboratory in Prague , Czech Republic and cultured as previously described [35] . Radial diffusion assays were performed as described previously [36] . Briefly , bacteria were grown to mid-logarithmic phase in 10 ml full strength 3% ( w/v ) tryptic soy broth ( TSB ) ( Becton Dickinson , Franklin Lakes , NJ , USA ) . The bacteria were washed once in 10 mM Tris , pH 7 . 4 and then 2×106 CFU were added to 5 ml of the underlay agarose gel consisting of 0 . 03% ( w/v ) TSB , 1% ( w/v ) low-electroendosmosistype ( Low-EEO ) agarose ( Sigma ) and a final concentration of 0 . 02% ( v/v ) Tween-20 . The underlay was poured into a Petri dish . After the agarose had solidified , 4 mm diameter wells were punched and 6 µl of test samples were added to each well . Samples were allowed to diffuse into the gel for 3 h in 37°C and then the underlay gel was covered with 5 ml overlay ( 6% TSB , 1% Low-EEO agarose ) . Antibacterial activity was visualized as a clear zone around each well after overnight incubation at 37°C of the plate . The activities of the peptides are presented as diameter of clear zone-well diameter . EDTA-blood was centrifuged at 800 g for 10 min and plasma and buffy coat removed . Erythrocytes were washed three times and resuspended in 5% PBS , pH 7 . 4 . The cells were incubated with end-over-end rotation for 1 h at 37°C in the presence of SEK20 or LL-37 ( 3–60 µM ) . 2% Triton X-100 ( Sigma-Aldrich ) served as positive control . The samples were then centrifuged at 800 g for 10 min . Hemoglobin release was measured as the absorbance at λ 540 nm and the values are expressed as % of TritonX-100 induced hemolysis . Dry lipid films were prepared by dissolving dioleoylphosphatidylcholine ( Avanti Polar Lipids , Alabaster , AL ) ( 30 mol% ) , dioloeolphosphatidic acid ( 30 mol % ) and cholesterol ( Sigma , St Louis , MO ) ( 40 mol% ) in chloroform , and removing the solvent by evaporation under vacuum overnight . Subsequently , buffer solution containing 10 mM Tris , pH 7 . 4 , was added together with 0 . 1 M carboxyfluorescein ( CF ) ( Sigma , St Louis , MO ) . After hydration , the lipid mixture was subjected to eight freeze-thaw cycles consisting of freezing in liquid nitrogen and heating to 60°C . Unilamellar liposomes with a diameter of about 140 nm ( as found with cryo-TEM and dynamic light scattering; results not shown ) were generated by multiple extrusions through polycarbonate filters ( pore size 100 nm ) mounted in a LipoFast miniextruder ( Avestin , Ottawa , Canada ) . Untrapped carboxyfluorescein was then removed by filtration through two subsequent Sephadex G-50 columns with the relevant Tris buffer as eluent . Both extrusion and filtration was performed at 22°C . In the liposome leakage assay , the well known self-quenching of CF was used . Thus , at 100 mM CF is self-quenched , and the recorded fluorescence intensity from liposomes with entrapped CF is low . On leakage from the liposomes , released CF is dequenched , and hence fluoresces . The CF release was determined by monitoring the emitted fluorescence at 520 nm from a liposome dispersion ( 10 mM lipid in 10 mM Tris pH 7 . 4 ) . An absolute leakage scale is obtained by disrupting the liposomes at the end of the experiment through addition of 0 . 8 mM Triton X100 ( Sigma , St Louis , MO ) , thereby causing 100% release and dequenching of CF . A SPEX-fluorolog 1650 0 . 22-m double spectrometer ( SPEX Industries , Edison , NJ ) was used for the liposome leakage assay . HaCaT keratinocytes were grown to confluence in 96 well plates ( 3000 cells/well ) in DMEM , 10% FCS . The medium was removed and the cells were subsequently washed with 100 µl DMEM . 100 µl of SEK20 or LL-37 ( 0 , 3 , 6 , 30 , 60 µM ) diluted in DMEM were added in triplicates . The LDH based TOX-7 kit ( Sigma-Aldrich ) was used to measure the viability of the cells . Bacteria were cultivated overnight in Todd-Hewitt medium ( TH; Difco ) at 37°C . 250 µl were transferred to 10 ml TH-medium and grown to mid-log phase ( A620≈0 . 4 ) . The bacterial solution was washed 3 times in 10 mM Tris , 5 mM glucose , pH 7 . 4 ( Tris-HCl buffer ) . 5 µl ( 2×106 cfu/ml ) was added to the respective antimicrobial substance at a total volume of 17 µl and incubated for 1 h at 37°C . After the incubation 500 µl Tris buffer was added to the mixture and 100 µl was transferred to Todd-Hewitt broth agar plate to determine the bacterial growth . Plates were incubated overnight at 37°C and the number of colony forming units ( cfu ) was determined . The antimicrobial effect of PCI against E . coli and S . pyogenes was analyzed by negative staining and electron microscopy as previously described [37] . Bacteria were diluted to a 1% solution with TBST ( 20 mM Tris , 150 mM NaCl , 0 . 05% Tween , pH 7 . 4 ) and 10 µl incubated with PCI at a concentration of 4 µM for 90 min at 37°C . 5 µl aliquots were adsorbed onto carbon-coated grids for 1 min , washed with two drops of water , and stained on two drops of 0 . 75% uranyl formate . The grids were rendered hydrophilic by glow discharge at low pressure in air . Specimens were observed in a Jeol JEM 1230 electron microscope operated at 60 kV accelerating voltage . Images were recorded with a Gatan Multiscan 791 CCD camera . Bacteria , grown to mid-exponential growth phase were washed and resuspended in PBST ( PBS+0 . 05% Tween-20 ) . 250 µl of the bacterial solution ( 2×109 bacteria/ml ) was incubated with 1 . 5 ml citrate plasma for 1 h at 37°C . The bacterial cells were collected , washed with PBST including 0 . 5 M NaCl and bound proteins eluted with 0 . 1 M glycine–HCl , pH 2 . 0 . The pH of the eluted material was raised to 7 . 5 by addition of 1 M Tris . Precipitated material was dissolved in SDS sample buffer and subjected to Tricine-SDS gel electrophoresis and Western blot analysis . Samples were boiled for 5 min in an equal volume of sample buffer containing 2% SDS and 5% 2-mercaptoethanol and run on SDS-PAGE . Bio-rad kaleidoscope prestained standards were used as molecular weight markers . Separated proteins were transferred to polyvinylidene difluoride ( PVDF ) membranes ( Amersham Biosciences ) . Membranes were blocked with PBST ( PBS+0 . 05% Tween-20 ) containing 5% dry milk powder ( blocking buffer ) overnight at 4°C and then incubated with primary antibodies ( rabbit anti-PCI K88032 1∶1000 ) in blocking buffer for 1 h at 37°C . After a washing step with PBST with 0 . 35 M NaCl , the membranes were incubated with HRP-conjugated secondary antibodies ( goat anti-rabbit IgG 1∶10000 ) in blocking buffer for 1 h at 37°C . The membranes were washed and bound antibodies detected by chemiluminescence . S . pyogenes was cultivated in human plasma and PCI-deficient plasma at 37°C . Bacterial suspensions were diluted 10000 times in 10 mM Tris , containing 5 mM glucose ( pH 7 . 4 ) and transferred to TH-agar plates at different time points ( 0 , 4 and 8 h ) to monitor bacterial growth . Plates were incubated overnight at 37°C and the number of bacteria determined . Blood samples were collected from healthy donors who had not taken antiplatelet medication in the previous ten days . Five ml of blood was collected into citrated vacuum tubes . Platelet-rich plasma ( PRP ) was prepared by centrifugation at 150×g for 10 minutes . Twenty µl of PRP was incubated at room temperature with 25 µl of HEPES buffer pH 7 . 4 , either in the presence or absence of 5 µM adenosine diphosphate ( ADP ) or M1 protein ( 1 µg/ml ) . After 5 minutes , primary antibodies were added ( rabbit anti-PCI K88032 1∶100 ) and incubated for 5 minutes . Five µl of fluorochrome conjugated secondary antibody ( anti Rabbit FITC ) was then added and after 5 minutes the incubation was stopped by addition of 0 . 5% formaldehyde in ice cold PBS . Samples were analysed using a FACSCalibur flow cytometer in logarithmic mode with a gate setting on the platelet population . 50 , 000 cells were acquired and analysed using Cell Quest software ( Becton Dickinson ) . A snap-frozen tissue biopsy collected from patients with necrotizing fasciitis caused by group A streptococcus ( GAS ) was stained and compared with a snap-frozen punch biopsy taken from a healthy volunteer . The biopsies were cryostat-sectioned to 8 µm , fixed in 2% freshly prepared formaldehyde in PBS and stained . Tissue sections were initially blocked with 10% fetal calf serum in Earl's balanced salt solution ( BSS ) with 0 . 1% saponin for 30 minutes at room temperature , followed by additional blocking with 1% H2O2 in BSS-saponin and an avidin and biotin blocking ( Vector laboratories ) . Primary antibodies were diluted in BSS solution containing saponin and 0 . 02% NaN3 and incubated over night at room temperature . S . pyogenes bacteria were identified with a polyclonal rabbit antiserum specific for the Lancefield group A carbohydrate ( diluted 1∶10 , 000; Difco ) , while PCI was identified with a polyclonal rabbit antiserum ( 1 µg/ml ) K88032 . After incubation , tissue sections were washed and blocked with 1% normal goat serum in BSS-saponin before addition of biotinylated goat anti-rabbit IgG ( diluted 1∶500 , Vector laboratories ) diluted in 1% normal goat serum in BSS-saponin . Avidin-peroxidase solution was added ( Vectastain-Elite; Vector Laboratories ) and the color reaction developed by the addition of 3 , 3-diaminobenzidine ( Vector Laboratories ) followed by counterstaining with hematoxylin . PCI and streptococci were also visualized with immunoflourescence stainings . The tissue was initially blocked with an avidin/biotin blocking kit ( Vector laboratories ) whereafter PCI was identified with a monoclonal mouse antibody ( 5 µg/ml ) API-93 followed by a blocking step with 1% normal goat serum . This was followed by incubation with a biotinylated antibody against mouse IgG ( diluted 1∶600 , Dako ) , and subsequently with a streptavidin-conjugated fluorophor ( Alexa 594 , diluted 1∶500 , Molecular Probes ) . After yet another round of blocking with avidin/biotin streptococci were detected using a biotinylated polyclonal rabbit antiserum specific for the Lancefield group A carbohydrate ( 16 µg/ml , Difco ) followed by incubation with a second streptavidin-conjugated fluorophore ( Alexa 488 , diluted 1∶600 , Dako ) . All antibodies and fluorochromes were diluted in PBS-saponin-BSA-c , while washes were done with PBS-saponin . The immunofluorescence stainings were evaluated by a Leica confocal scanner TCP SP II coupled to a Leica DMR microscope . Statistical analysis was performed using GraphPadPrism 4 . 00 . For viable count data the p value was determined using a one sample T test with a theoretical mean set as 100% survival based on untreated samples . For flow cytometry data the p value was determined using students T test .
|
The innate immune system is an integral part of our battle against an invading pathogen . Antimicrobial peptides and proteins partake in this fight due to their ability to perforate the bacterial cell wall , which eventually will cause the efflux of bacterial cytosolic content and efficient bacterial killing . Protein C inhibitor ( PCI ) is a multifunctional heparin-binding serpin which has been implicated in a number of pathological conditions , including severe infectious diseases . Here we show that PCI is a potent antimicrobial agent that is able to destroy the bacterial cell wall and thereby cause death of the bacteria . Our study also shows that in contrast to many other antimicrobial peptides , processing of PCI is not required since the full length protein exerts its antimicrobial activity , and we present data demonstrating that PCI is enriched at the infected site of patients suffering from severe streptococcal infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology/immune",
"response",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"infectious",
"diseases/bacterial",
"infections",
"immunology/immunity",
"to",
"infections",
"microbiology/medical",
"microbiology",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] |
2009
|
Protein C Inhibitor—A Novel Antimicrobial Agent
|
Drosophila melanogaster head development represents a valuable process to study the developmental control of various organs , such as the antennae , the dorsal ocelli and the compound eyes from a common precursor , the eye-antennal imaginal disc . While the gene regulatory network underlying compound eye development has been extensively studied , the key transcription factors regulating the formation of other head structures from the same imaginal disc are largely unknown . We obtained the developmental transcriptome of the eye-antennal discs covering late patterning processes at the late 2nd larval instar stage to the onset and progression of differentiation at the end of larval development . We revealed the expression profiles of all genes expressed during eye-antennal disc development and we determined temporally co-expressed genes by hierarchical clustering . Since co-expressed genes may be regulated by common transcriptional regulators , we combined our transcriptome dataset with publicly available ChIP-seq data to identify central transcription factors that co-regulate genes during head development . Besides the identification of already known and well-described transcription factors , we show that the transcription factor Hunchback ( Hb ) regulates a significant number of genes that are expressed during late differentiation stages . We confirm that hb is expressed in two polyploid subperineurial glia cells ( carpet cells ) and a thorough functional analysis shows that loss of Hb function results in a loss of carpet cells in the eye-antennal disc . Additionally , we provide for the first time functional data indicating that carpet cells are an integral part of the blood-brain barrier . Eventually , we combined our expression data with a de novo Hb motif search to reveal stage specific putative target genes of which we find a significant number indeed expressed in carpet cells .
The development of complex organs is often accompanied by extensive cell- and tissue rearrangements . Initially simple cells undergo profound morphological changes such as extensive cell fusions of muscle precursor cells to form syncytial muscle fibers [1] or the formation of highly polarized neurons from initially uniform neuroectodermal cells [2 , 3] . Other cell types , such as germ cells , first migrate long distances before coming to rest in the developing gonads [4] . Although these cell-type specific processes need to be tightly controlled and coordinated with those of other cell types of the same and neighboring organs , the molecular mechanisms involved are still poorly understood . The development of the adult Drosophila melanogaster head and the visual system has been proven to be an excellent model to study the coordination of different developmental processes [5–10] . The adult D . melanogaster head is composed of the compound eyes ( the main visual system ) , the three dorsal ocelli , the antennae , the ventral mouthparts and the head capsule that connects these organs and encloses the brain [11] . Most of these structures develop during larval stages from eye-antennal imaginal discs , which originate from about 20 cells that are specified at embryonic stages [12–14] . Throughout larval development , the eye-antennal discs grow extensively by cell proliferation resulting in discs composed of more than 15 , 000 cells at the beginning of pupation [7 , 15] . During the first two larval stages , the initially uniform disc is subdivided into an anterior antennal and a posterior retinal compartment by the action of two opposing morphogen gradients , which subsequently activate genes responsible for antennal development [16 , 17] and the retinal determination genes [15 , 18–24] . Approximately at the same time when the retinal part of the disc and the antennal region separate during the early L2 stage , the maxillary palp is defined in the ventral portion of the antennal part [25–27] . Once the eye-antennal disc is subdivided into the different organ precursors , cells within each compartment start to differentiate at L2/early L3 stages . In the retinal region , a differentiation wave that is established in the posterior most part of the equator region moves anteriorly . This wave is accompanied by a morphologically visible indentation , the so-called morphogenetic furrow ( MF ) [28] . In the region of the MF the expression of the proneural gene atonal ( ato ) becomes restricted to regularly spaced single cells posterior to the furrow [29 , 30] . Those cells are destined to become R8 photoreceptors , which subsequently recruit R1-R7 photoreceptors and associated cell types , such as cone and pigment cells from the surrounding cells [31–33] . The axons of successively forming photoreceptor cells need to be connected to the optic lobes to allow a functional wiring of the visual system with the brain . All axons are collected at the basal side of the eye-antennal disc and guided through the optic stalk throughout the L3 stage . This process is supported by retinal glia cells , which originate mainly by proliferation from 6–20 glia cells located in the optic stalk prior to photoreceptor differentiation [34–36] . These retinal glia cell types include migratory surface glia ( including perineurial and subperineurial glia cells ) and wrapping glia . Triggered by the presence of developing photoreceptor cells , the retinal glia cells enter the eye-antennal disc through the optic stalk and migrate towards the anterior part of the disc , always remaining posterior to the advancing morphogenetic furrow [34–36] . When photoreceptors differentiate , the contact of their growing axons with perineurial glia cells triggers the reprogramming of these glia cells into differentiated wrapping glia , which extend their cell membranes to ensheath bundles of axons that project to the brain lobes through the optic stalk [34 , 37 , 38] . The basally migrating perineurial glia cells and the wrapping glia ensheathed projecting axons are separated by two large polyploid carpet cells , each of them covering half of the retinal field [34] . The two carpet cells form septate junctions and express the G protein-coupled receptor ( GPCR ) encoded by the moody locus , both characteristics of the subperineurial surface glia type [34 , 39] . While subperineurial glia cells located in the brain remain there to form the blood-brain barrier , the carpet cells are thought to originate in the optic stalk [36] , and during L2 and early L3 stages they migrate into the eye-antennal disc . Later during pupal stages , they migrate back through the optic stalk to remain beneath the lamina neuropil in the brain . However , so far it is not known , whether carpet cells or other retinal glia cell types eventually contribute to the formation of the blood-eye barrier , the retinal portion of the blood-brain barrier [40 , 41] . The carpet cells thus share features of subperineurial glia , but their extensive migratory behavior and their function in the eye-antennal disc suggest that these cells may exhibit distinct cellular features . However , so far , no carpet cell specific regulator has been identified that may be involved in specifying carpet cell fate . Although eye-antennal disc growth and patterning , and especially retinal determination and differentiation , are among the most extensively studied processes in D . melanogaster , a systematic understanding of involved genes and their potential genetic and direct interactions is limited to the late L3 stage in the context of retinal differentiation [42–49] . Similarly , recent attempts to incorporate existing functional and genetic data into a gene regulatory network context covers mainly retinal determination and differentiation processes [50] . So far , a comprehensive profiling of gene expression dynamics throughout eye-antennal disc development is missing . The same holds true for the molecular control of retinal glia cell development . While the transcriptome of adult surface glia in the brain has been analyzed [51] , retinal glia cells have not been comprehensively studied yet . Here we present a dynamic genome wide expression analysis of D . melanogaster eye-antennal disc development covering late L2 to late L3 stages . We show that the transition from patterning to differentiation is accompanied by extensive remodeling of the transcriptional landscape . Furthermore , we identified central transcription factors that are likely to regulate a high number of co-expressed genes and thus key developmental processes in the different organ anlagen defined in the eye-antennal disc . One of these central factors is the C2H2 zinc-finger transcription factor Hunchback ( Hb ) [52] that has been extensively studied in D . melanogaster during early axis determination and segmentation [53 , 54] . It is also well-known for its role in the regulation of temporal neuroblast identity during embryogenesis , as it determines first-born identity in the neural lineage [55 , 56] . Here we show for the first time that hb is expressed in carpet cells and loss of function experiments suggest that its activity is necessary for carpet cell formation and/or migration and consequently for blood-brain barrier integrity . Eventually , we reveal putative Hb target genes and confirm that bioinformatically predicted targets are indeed expressed in developing carpet cells .
Although compound eye development and retinal differentiation are among the most intensively studied processes in D . melanogaster , a comprehensive understanding of the underlying gene expression dynamics is still missing to date . To identify the genes expressed during D . melanogaster eye-antennal disc development and their expression dynamics , we performed RNA-seq on this tissue at three larval stages covering the process of retinal differentiation that is marked by the progression of the morphogenetic furrow . The late L2 stage ( 72h after egg laying , AEL ) represents the initiation of differentiation , at mid L3 stage ( 96h AEL ) the morphogenetic furrow is in the middle of the retinal field and the late L3 stage ( 120h AEL ) represents the end of morphogenetic furrow progression ( Fig 1A ) . Multidimensional scaling clustering clearly indicated that the largest difference in gene expression ( dimension 1 ) was between L2 eye-antennal discs ( 72h AEL ) and L3 eye-antennal discs ( 96h and 120h AEL ) ( S1 Fig ) . After filtering out not expressed and very lowly expressed genes , we observed that 9 , 194 genes were expressed at least in one of the three sequenced stages . As anticipated by the multidimensional scaling plot ( S1 Fig ) , the number of genes that changed their expression between 72h AEL and 96h AEL was much larger than between 96h AEL and 120h AEL , ( Fig 1A ) . In only 24 hours , during the transition from L2 to L3 , 60% of the expressed genes changed their expression significantly . In the transition from mid L3 to late L3 , in contrast , only 22% of the genes underwent a change in their expression . The observation that many more genes were significantly differentially expressed during the transition from late L2 stages to mid L3 stages may in part also be caused by the fact that only female discs were analyzed between 96h and 120h AEL , while we compared mixed males and females at 72h AEL with only females at 96h AEL during the first transition . Since about one third of all genes in D . melanogaster show signs of sex-specific expression [57] , the differentially expressed genes in the first transition may also include some male or female biased genes . However , in a GO term analysis , we only found two terms associated with the search term “sex” ( see Materials and Methods ) , namely “sex differentiation” ( GO:0007548; padj = 5 . 35e-5; 47 genes ) and “dosage compensation” ( GO:0007549; padj = 2 . 61e-3; 15 genes ) among those genes that were higher expressed at 96h AEL . None of the 165 terms was enriched in those genes higher at 72h AEL and among differentially expressed genes in the 96h vs . 120h comparison . It has been shown that early eye-antennal disc stages are mainly characterized by patterning processes that subdivide the initially uniform disc into the individual organ anlagen [6 , 15 , 18 , 25 , 58–61] . Within organ-specific domains , further patterning processes define for instance the dorsal ventral axis in retinal field [62–64] or the proximal-distal axis of the antennae [65] . Additionally , the discs grow extensively throughout L1 and L2 stages mainly by cell proliferation [15 , 66] . Accordingly , the genes active at the late L2 stage were mostly involved in metabolic processes and generation of energy ( Fig 1B ) . At the end of L2 stages , the patterning processes are mostly concluded and differentiation starts within each compartment . For instance , in the retinal field the progression of the differentiation wave is accompanied by a reduction in cell proliferation [8 , 10] . Therefore , mostly genes related to cell differentiation , nervous system development , pattern specification and compound eye development were significantly up-regulated at the mid L3 stage ( Fig 1B ) . During the transition from the mid L3 stage to the late L3 stage again genes related to metabolism and energy production were down-regulated ( Fig 1C ) . This may be explained by the fact that at 96h AEL the disc has not yet reached its final size , and cells anterior to the morphogenetic furrow still proliferate [10] . Also , directly behind the morphogenetic furrow one last synchronous cell division takes place to give rise to the last cells of the photoreceptor clusters ( R1 , R6 and R7 ) [10 , 60] . In the light of ongoing differentiation , genes active at the late L3 stage belonged to GO categories related to differentiation processes ( Fig 1C ) . At this stage , terms related to processes taking place late during eye-antennal disc development [10] such as R7 cell differentiation or pigment metabolic process ( Fig 1C ) were also obtained . In summary , the pairwise comparison of expression levels of genes expressed at the three studied stages of eye-antennal disc development recapitulate the key transition from pattering and proliferation processes to differentiation . In order to better characterize the different expression dynamics of the expressed genes , we performed a co-expression clustering analysis based on Poisson Mixture models [67] . Manual comparison of the different outputs showed that the 13 clusters predicted by one of the models ( Djump , [68] ) were non-redundant and sufficiently described all the expression profiles present in the data . A total of 8 , 836 genes could be confidently placed in one of these clusters ( maximum a posteriori probability ( MAP ) > 99% ) . We ordered the predicted 13 clusters according to their expression profile ( Fig 2 ) : four clusters contained clearly early expressed genes , two of them contained genes expressed only at 72h AEL ( cluster 1 and 2 ) and two contained genes predominantly expressed early , but also with low expression at 96h and/or 120h AEL ( clusters 3 and 4 ) ; one cluster showed down-regulation at 96h AEL , but a peak of expression again at 120h AEL ( cluster 5 ) ; the genes in the largest clusters showed almost constant expression throughout the three stages ( clusters 6 and 7 ) ; one cluster showed constant expression at 72h AEL and 96h AEL and down-regulation at 120h AEL ( cluster 8 ) ; one cluster showed a peak of expression at 96h AEL ( cluster 9 ) and four clusters contained genes with predominantly late expression , one with high and constant expression at 96h AEL and 120h AEL ( cluster 10 ) , two with up-regulation in both transitions ( cluster 11 and cluster 12 ) and one with genes expressed only at 120h AEL ( cluster 13 ) . A GO enrichment analysis for the genes in the individual clusters showed that the genome-wide co-expression profiling and subsequent ordering of the clusters recapitulated the consecutive biological processes that take place during eye-antennal disc development with a great resolution ( Fig 2 , S1 Table ) . For instance , we found genes related to energy production mainly in clusters 2 and 3 , while genes more specific for terms related to mitosis and cell cycle were found in clusters 4 , 8 and 9 . For example , cluster 8 grouped genes that were similarly high expressed at 72h and 96h AEL , and their expression decreases at 120h AEL . The enrichment of GO terms related to DNA replication and cell cycle control ( Fig 2; S1 Table ) , corresponds with the fact that active proliferation takes place at these stages [60] . Thus , other genes that were grouped in this cluster , but for which no previous knowledge is available , are likely also related to these biological functions . Genes up-regulated in the later stages were separated in more specific clusters , and most of the enriched GO terms were related to differentiation and neuron and eye development . For instance , cluster 10 contained the more general term “imaginal disc development” , while cluster 12 showed enrichment for “compound eye morphogenesis” , and cluster 13 was the only with enriched terms related to pupation processes and pigmentation . In order to get an overview of the presence of key eye developmental genes in the individual clusters , we mapped the clusters onto the FlyODE network ( Fig 3 ) that comprises manually curated interactions of genes involved in retina development [50] . Only ten of the 146 genes of this network were not present in any of the 13 clusters , either because they did not pass the expression level or the cluster assignment thresholds ( grey nodes in Fig 3 ) . Apart from the cell death regulator reaper ( rpr ) ( cluster 3 ) , all other genes of this network are present in clusters 7 to 13 ( Fig 3 ) , that represent genes with constant expression throughout the three stages ( clusters 7 to 9 ) or with increasing expression at later stages ( clusters 10 to 13 ) ( Fig 2 ) . This observation is in accordance with the nature of the FlyODE network that covers mainly stages of retinal differentiation from the third instar to the adult [50] . The most prominent retinal determination genes [7 , 10] eyeless ( ey ) , sine oculis ( so ) , dachshund ( dac ) and optix/six3 are present in clusters 7 to 9 , while eyes absent ( eya ) , twin of eyeless ( toy ) and eye gone ( eyg ) were found in clusters 10 and 11 . Among the genes expressed during late L3 stages in clusters 12 and 13 we found central genes involved in photoreceptor and neuronal differentiation [10] such as atonal ( ato ) , glass ( gl ) and prospero ( pros ) ( Figs 2 and 3 ) . Members of well-known developmental signaling pathways such as the Wnt , BMP and Hh pathways , EGFR , Notch and cell cycle related genes ( e . g . CycE ) were present in cluster 10 ( Fig 3 ) that grouped genes with similarly high expression at 96h AEL and 120h AEL ( Fig 2 ) . Among genes which steadily increased in expression throughout the three studied stages ( cluster 12 ) , we found for instance Delta ( Dl ) , which is one of the Notch receptor ligands [69] and has been shown to fulfill different roles during eye development [66 , 70 , 71] . Also , anterior open ( aop ) ( also known as yan ) , a member of the JNK signaling pathway [72] , which is described to repress photoreceptor differentiation [73] and also to determine R3 photoreceptor identity [74] was present in this cluster . Although we sequenced the entire eye-antennal discs , we found many GO terms related to eye development with high enrichment scores , while very few GO terms specific for antenna and maxillary palps were observed ( e . g . in cluster 12 “eye development” appears with p = 4 . 38e-24 , “antennal development” with p = 4 . 37e-08 and no GO terms related to maxillary palps were found ) ( S1 Table ) . This finding may be a result of much more extensive research on eye specific developmental processes in comparison to the other organs that develop from the same imaginal disc . However , many GO terms related to leg formation and proximodistal pattern formation were highly enriched in the genes in cluster 9 ( “proximaldistal pattern formation” with p = 4 . 48e-05 ) , cluster 10 ( “leg disc development” with p = 7 . 85e-20 ) and cluster 12 ( “leg disc development” with p = 4 . 32e-12 ) ( S1 Table ) . Since antennae and maxillary palps are serially homologue to thoracic appendages , pathways involved in leg , antenna and maxillary palp development are likely to share key regulators [17 , 65 , 75–80] , suggesting that genes found in these clusters may also play a role in antenna or maxillary palp development . The assignment of all expressed genes to their corresponding cluster is available along with the GEO submission number GSE94915 and on this website http://www . evolution . uni-goettingen . de/posnienlab/clusterSearch/search . html . In summary , we showed that clusters with early expressed genes mainly represent metabolic and energy related processes , while clusters with late expressed genes represent more organ specific differentiation and morphogenetic processes . Therefore , we provide a dataset that contains the complete gene expression dynamics underlying the fundamental change from predominantly patterning and proliferation processes to the onset of differentiation from late L2 stages to late L3 larval instars [7 , 28] . The co-expression of genes observed in the 13 clusters may be a result of co-regulation by the same transcription factors or combinations thereof . In order to test this hypothesis and to reveal potential central upstream regulators , we used the i-cisTarget method [81] to search for enrichment of transcription factor binding sites in the regulatory regions of the genes within each of the 13 clusters ( Fig 2 , S2 Table ) . As basis for this enrichment analysis various experimental ChIP-chip and ChIP-seq datasets were used , namely those published by the modENCODE Consortium [82] , the Berkeley Drosophila Transcription Network Project [83] and by the Furlong Lab [84 , 85] . One of the most noticeable results of the statistical ranking analysis was that genes in 9 of the 13 clusters showed significant enrichment for Nejire binding sites ( Fig 2 ) . Nejire ( also known as CREB-binding protein ( CBP ) ) is a zinc-finger DNA binding protein that functions as a co-activator that acts as bridge for other transcription factors to bind specific enhancer elements [86–88] , which can explain why we find it to regulate such many target genes . Nejire/CBP has been shown to be involved in many processes during eye development and patterning in D . melanogaster [89 , 90] and mutations in this gene cause the Rubinstein-Taybi syndrome in humans [91] that among others is characterized by extensive problems during retinal development [92] . Similarly , Pannier was found enriched to regulate the genes of many clusters ( clusters 2 , 4 , 6 , 7 , 8 , 9 and 11 ) ( Fig 2 ) . This GATA transcription factor is involved in the establishment of the dorsal-ventral axis of the retinal field of the discs during early L1 and L2 stages [93 , 94] , while later during L2 and L3 stages it is known to have a role in defining the head cuticle domain by repressing eye determination genes [64 , 94] . In both cases , Pannier is found in a very upstream position of the respective gene regulatory networks that define these cell fates [64 , 95] . Besides these highly abundant transcription factors , the clusters with genes predominantly expressed at later stages were also enriched for transcription factors already known to play a role in eye-antennal disc development . For instance , a significant number of Sloppy-paired 1 ( Slp1 ) target genes are up-regulated at L3 stage ( cluster 12 , Fig 2 ) and this transcription factor is known to play a critical role in establishing dorsal-ventral patterning of the eye field in the eye-antennal disc [96] . A function of Daughterless ( Da ) ( identified in cluster 13 , Fig 2 ) is also described: it is expressed in the morphogenetic furrow , it interacts with Atonal and is necessary for proper photoreceptor differentiation [97] . Finally , Snail ( enriched in cluster 1 and 13 ) and Twist ( enriched in cluster 12 ) ( Fig 2 ) were previously identified as possible repressors of the retinal determination gene dac [98] and our results indicate that they regulate also other genes during eye-antennal disc development . Cluster 5 contained genes that show a peak in expression at 72h AEL and 120h AEL stages , which precede major stage transitions from L2 to L3 and from L3 to pupa stage , respectively . These transitions are characterized by ecdysone hormone pulses before larval molting and pupation [99] . The only potential transcription factor binding site that was significantly enriched was that of the Ecdysone Receptor ( EcR ) , that has been shown to be expressed in the eye-antennal disc in the region of the progressing morphogenetic furrow [100] . Intriguingly , we did not find an enrichment for known transcription factors such as So or Ey . The most obvious explanation is that there is no ChIP-seq data available ( e . g . for Ey ) or that the available data for these factors ( e . g . for So [48] ) is not included as transcription factor binding information in the current i-cisTarget database [81 , 101] . The identification of well-known transcription factors suggests that the applied clustering approach indeed allows identifying key regulators of various processes taking place throughout eye-antennal disc development . Interestingly , we identified a few generally well-known upstream factors for which a potential role during eye-antennal disc development has not yet been described . For instance , in clusters of very early expressed genes , we found an enrichment of motifs for the transcription factor Caudal ( Cad ) ( cluster 1 and 2 ) and the Hox protein Fushi tarazu ( Ftz ) ( cluster 1 ) ( Fig 2 ) . Caudal is a downstream core promoter activator [102] and very recently it has been found that it cooperates with Nejire to promote the expression of the homeobox gene fushi tarazu ( ftz ) [103] . A Caudal-like transcription factor binding motif has been identified within Sine oculis ( So ) bound DNA fragments as identified by ChIP-seq [48] , suggesting that So and Cad may co-regulate potential target genes in the eye-antennal disc . The MADS-box transcription factor Myocyte enhancer factor 2 ( DMef2 ) was predicted to regulate genes found in clusters 2 , 4 and 12 ( Fig 2 ) . Using two independent Dmef2-Gal4 lines to drive GFP expression , we confirmed expression of Dmef2 in lose cells attached to the developing eye-antennal discs ( S2 Fig ) . Eventually , we found an enrichment of potential target genes of the C2H2 zinc-finger transcription factor Hunchback ( Hb ) in clusters 12 and 13 , which are active mainly during mid and late L3 stages . Since GO terms enriched in these two clusters suggested an involvement in retinal development or neurogenesis ( Fig 2 , S1 Table ) , we examined a potential function of Hb in the eye-antennal disc in more detail . Using in-situ hybridization we found hb expression in two cell nuclei at the base of the optic stalk in the posterior region of late L3 eye-antennal discs ( S3A Fig ) . With a Hb antibody we also detected the Hb protein in these two basally located nuclei ( S3B and S3C Fig ) . DNA staining with DAPI showed that the Hb-positive nuclei are bigger than those of surrounding cells , suggesting that they are polyploid . Additionally , we tested two putative Gal4 driver lines obtained from the Vienna Tile library [104] ( VT038544; S3D Fig and VT038545; S3E Fig ) . Both lines drove reporter gene expression in the two polyploid nuclei as described above . Note that both lines also drove the typical hb expression in the developing embryonic nervous system , but not the early anterior expression [105 , 106] . The regulatory region covered by the two Gal4 driver lines is located at the non-coding 3’ end of the hb locus accessible to DNA-binding proteins at embryonic stages 9 and 10 [83] ( S4 Fig ) , a time when early-born neuroblasts express hb [55] . The lack of the early anterior expression may be explained by the fact that the DNA region covered by the driver lines does not seem to be bound by Bicoid during early embryonic stages [83] ( S4 Fig ) . Based on these findings , we are confident that the regions covered by the two Gal4 driver lines ( VT038544 and VT038545 ) recapitulate native hb expression . The basal location of the hb-positive cells suggests that they may be retinal glia cells . Co-expression of hb with the pan-glial marker Repo ( Fig 4A–4C ) further supported this suggestion . Previous data has shown that two polyploid retinal subperineurial glia cells ( also referred to as carpet cells ) cover the posterior region of the eye-antennal disc [34 , 36] . In order to test , whether hb may be expressed in carpet cells , we first investigated the expression of the subperineurial glia marker Moody [107] and we found a clear co-localization with Hb ( Fig 4D ) . Carpet cells migrate through the optic stalk into the eye-antennal disc during larval development [34 , 36] . Therefore , we followed the expression of the hb driver lines throughout late L2 and L3 larval stages ( Fig 4A–4C ) . Already at the L2 stage , we could easily recognize the hb-positive cell nuclei by their large size ( Fig 4A and 4A”; see also size quantification in Fig 5D ) . We could corroborate that these cells indeed migrated through the optic stalk during late L2 and early L3 stages ( Fig 4A and 4B ) , and then entered the disc and remained basally in the posterior region of the disc , flanking the optic stalk ( Fig 4C and 4C” ) . As previously observed for carpet cells [34] , we never found hb-positive cell nuclei in the midline of the retinal field . Taken together , these data show that hb is expressed in two polyploid retinal subperineurial glia cells ( carpet cells ) that enter the basal surface of the eye-antennal disc through the optic stalk during larval development . The expression of hb in carpet cells suggested an involvement in their development . To test this hypothesis , we examined loss of Hb function phenotypes based on RNA interference ( RNAi ) driven specifically in subperineurial glia cells ( moody-Gal4 driving UAS-hbdsRNA ) . Of four tested UAS-hbdsRNA lines we used the most efficient line ( see Materials and Methods ) for the RNAi knock-down experiments . Additionally , we investigated eye-antennal discs of a temperature sensitive mutant ( hbTS ) [108] . Since Hb is necessary during embryogenesis [53 , 54] , the analyzed flies were kept at 18°C during egg collection and throughout embryonic development , and they were only transferred to the restrictive temperature of 28°C at the L1 stage . We quantified the size differences between carpet cell nuclei and neighboring glia cells by measuring the nucleus area . The mean area of a Repo-positive glia cell was 30 . 42 μm2 , while the carpet cell nuclei had a mean area of 86 . 10 μm2 ( Fig 5A and 5D ) . Based on these clear results , the presence or absence of carpet cell nuclei could be confidently identified by α-Repo staining because of their significantly larger size ( see Fig 5A and 5D ) . The most common phenotype observed in late L3 eye-antennal discs of RNAi and mutant flies was the absence of one or both carpet cell nuclei ( Fig 5A–5C ) . In wild type animals , we could unambiguously identify two carpet cell nuclei in 72% of the eye-antennal discs . In 21% of the analyzed discs , we found only one carpet cell nucleus ( Fig 5E ) . In contrast , in 35% to 40% of the studied Hb loss of function discs only one carpet cell nucleus was observed ( Fig 5B and 5E ) . This single polyploid Repo-positive nucleus was significantly larger than carpet cell nuclei in discs that contained two carpet cell nuclei ( Fig 5D; 129 . 80 μm2 compared to 86 . 10 μm2 if two carpet cells were present ) . Additionally , the single carpet cell nuclei were mostly located in the midline of the retinal field ( Fig 5B ) . No carpet cell nuclei could be observed in 24% and 38% of the eye-antennal discs originating from moody>>hbdsRNA and hbTS flies , respectively ( Fig 5C and 5E ) . Note that we obtained comparable results when we expressed the hb dsRNA in all glia cells ( repo>>hbdsRNA ) or only in subperineurial glia cells ( moody>>hbdsRNA ) . To identify larval stages at which Hb function is crucial for carpet cell development , we transferred hbTS flies to the restrictive temperature of 28°C at 24h AEL ( early L1 stage ) , at 48h AEL ( late L1 ) , at 72h AEL ( late L2 ) or 96h AEL ( mid L3 stage ) and assessed the presence of polyploid Repo-positive carpet cell nuclei in late L3 eye-antennal discs , respectively . In all cases , we found a significant reduction of the number of carpet cell nuclei when compared to control discs ( Fig 5F ) . Although no clear significant differences in the number of carpet cells was detected between the consecutive experiments ( S5 Fig ) , our results show that Hb function is necessary for the presence of polyploid carpet cell nuclei throughout larval development . The observed loss of polyploid carpet cell nuclei could be a result of either the loss of the entire carpet cells , incomplete migration into the eye-antennal disc or loss of the polyploidy . To distinguish between these options , we tested whether also the carpet cell membranes were affected upon loss of hb expression , in addition to the polyploid nuclei . To this aim , we expressed hbdsRNA specifically in subperineurial glia cells with a moody-Gal4 driver line together with a strong membrane marker ( 20xUAS-mCD8::GFP ) to label the extensive carpet cell membranes ( moody>>20xmCD8::GFP; moody>>hbdsRNA , Fig 6 ) . In the eye-antennal discs where two carpet cells could be identified , moody expressing membranes were always present in the optic stalk and in most cases ( 95% ) they spanned the entire posterior region of the eye-antennal disc up to the morphogenetic furrow ( Fig 6A and 6D ) . In discs with only one clear carpet cell nucleus , we observed a decrease both in the cases where mCD8::GFP was present in the optic stalk ( 82% ) and even more in the retinal field ( 73% ) ( Fig 6B and 6D ) . Of the discs with no clear polyploid carpet cell nuclei ( Fig 6C and 6D ) , also fewer ( 89% ) showed moody expressing membranes in the stalk , although also the difference was not significant compared to controls . However , a very significant decrease in the percentage of discs with moody expressing membranes in the retinal field was observed in this case ( 22% ) . Subperineurial glia cells cover the entire surface of the brain from larval stages onwards . They are an integral part of the protective blood-brain barrier by establishing intercellular septate junctions [109] . The blood-brain barrier prevents the substances that circulate in the hemolymph to enter the brain and helps maintaining the proper homeostatic conditions of the nervous system [110] . Since it has been shown that the carpet cells migrate through the optic stalk towards the brain during pupal stages [40] , we tested , whether the loss of hb expression in developing carpet cells had an effect on the integrity of the blood-brain barrier . To this aim , we injected fluorescently labeled dextran into the abdomen of moody>>hbdsRNA adult flies and scored the presence of this dye in the retina of the flies . Animals with a properly formed blood-brain barrier showed a fluorescent signal in their body , but not in the retina ( Fig 7A ) . However , in animals that had an incomplete blood-brain barrier , the dextran penetrated into the retina and fluorescence was observed in the compound eyes ( Fig 7A’ ) . Since it is known that blood-brain barrier permeability can increase after exposure to stress conditions [111 , 112] , we only scored animals that survived 24h after the injection of dextran . In most cases , the two eyes of an individual presented different fluorescent intensities , and even no fluorescence in one eye but strong signal in the other . Therefore , we scored each eye separately . moody>>hbdsRNA flies had a significantly higher rate of fluorescent retinas ( p = 8 . 08e-7 , χ2 test ) , indicating that their eyes were not properly isolated from the hemolymph circulating in the body cavity ( Fig 7B ) . In summary , our loss of function experiments further confirmed a central role of Hb in carpet cell development . Besides impaired retinal glia cell migration and axon guidance , we showed that upon loss of Hb function also the blood-brain barrier integrity is disrupted . Since we have identified Hb because of an increase in expression of its target genes during 96h and 120h AEL stages and hb itself is only expressed in carpet cells , we also investigated , whether some of the targets were expressed in these cells . Using available ChIP-chip data for Hb from the Berkeley Drosophila Transcription Network Project ( BDTNP ) [83] , we generated a high confidence list of 847 putative Hb target genes ( see Materials and Methods for details ) , of which 585 were expressed in eye-antennal discs at least in one of the studied stages . More precisely , we found that 267 of these genes were differentially expressed in the transition from 72h to 96h AEL and only 52 were differentially expressed between 96h and 120h AEL ( Fig 8 , S4 Table ) . In both cases , most of these genes were up-regulated , suggesting that Hb mainly activates target gene expression in the eye-antennal disc . Focusing only on those target genes that resulted in the identification of Hb in our clustering approach ( see above ) , we found that 77 of the 585 expressed putative Hb targets were present in clusters 12 and 13 . We searched the GO terms for biological functions of these 77 genes and found that 17 code for transcription factors and up to 25 code for proteins integral to the cell membrane . A number of GO terms were related to neuronal development and eye development and to note is the presence of genes known to be related to glia cell migration and endoreduplication ( S5 Table ) . Based on their annotated GO terms , predicted or known cellular location and the availability of driver lines and antibodies , we selected 13 of these target genes and tested if they were expressed in carpet cells at 120h AEL . For 8 out of the 13 selected targets we found no clear expression related to carpet cells ( archipelago ( ago ) , Delta ( Dl ) , knirps ( kni ) , rhomboid ( rho ) , roundabout 3 ( robo3 ) , Sox21b , Src oncogene at 64B ( Src64B ) and thickveins ( tkv ) ) . This could be because they were false positives , but they could also be expressed at earlier stages than analyzed here or the used driver constructs did not include the regulatory regions to drive expression in carpet cells . Tkv for instance could still be an interesting candidate because it has been implicated in retinal glia cell development [113] . brinker ( brk ) , Cadherin-N ( CadN ) , cut ( ct ) , Fasciclin 2 ( Fas2 ) and sprouty ( sty ) showed expression in carpet cells ( Fig 9 ) . brinker ( brk ) was very broadly expressed in the eye-antennal disc ( S6C and S6D Fig ) . Although we could only observe expression in one of the two cells in every eye-antennal disc we analyzed , CadN is clearly expressed in carpet cells ( Fig 9A ) . Recent data demonstrated that CadN , a Ca+ dependent cell adhesion molecule , is necessary for the proper collective migration of glia cells [114] , a key feature of carpet cells . As it has previously been published , Cut is expressed in subperineurial glia cells ( Fig 9B ) [115] . The Cut protein is present in carpet cells already at L2 stage and remains until late L3 stage ( Fig 9B ) . It has been shown that Cut is necessary for proper wrapping glia differentiation ( note that the wrapping glia expression is found in another focal plane in Fig 9B ) and to correctly form the large membrane processes that these cells form [115] . Interestingly , carpet cells have a similar morphology , with very large membrane surface and extensive processes that reach to the edge of the retinal field . In contrast , retinal perineurial glia do not have this morphology and do not express Cut . Also , Fas2 ( Fig 9C ) and sty ( Fig 9D ) were clearly expressed in carpet cells as well as in several other cells in the eye-antennal disc . Sty and Fas2 are negative regulators of the EGFR signaling pathway that is involved in retinal glia cell development and photoreceptor differentiation [116–121] . Please note that we used antibodies to detect target genes if they were available . Given the nature of the CadN and Fas2 proteins [122 , 123] , we detected broad membrane associated expression for both proteins ( S6 Fig ) . Therefore , we decided to study the expression of those two candidates using available Gal4 lines ( see above and Fig 9A and 9C ) . For both lines , we found at least one high confidence Hb motif ( S6 Table ) . In summary , we showed that 5 of the 13 computationally predicted Hb target genes that we tested were expressed in carpet cells , suggesting that our bioinformatic pipeline allows the identification of new potential regulators of carpet cell development .
Although compound eye development is one of the most extensively studied processes in D . melanogaster , a comprehensive understanding of genome wide gene expression dynamics is still missing . A previous genome wide expression study based on Microarray data has shown that between 5 , 600 and 6 , 100 genes are active in the retinal part of the eye-antennal disc at the late L3 instar stage [49] . Given that we studied the entire eye-antennal discs at three time points , our observation that 9 , 194 of all annotated D . melanogaster genes are expressed in at least one stage is likely to provide a meaningful estimate of the number of active genes throughout eye-antennal disc development . The pairwise comparison of the studied stages ( Fig 1 ) , as well as the more defined clustering approach ( Fig 2 ) , revealed an extensive remodeling of the transcriptomic landscape during eye-antennal disc development . It has previously been shown that co-expression of genes is likely to be a result of regulation by similar or even the same transcription factors [124–128] . This basic assumption has been successfully used to identify central transcriptional regulators in developmental gene regulatory networks [44 , 85 , 129–133] . Here , we identified several transcriptional regulators , such as Pnr , Nejire , Slp1 , Da , Snail , Twist and EcR ( Fig 2 ) that have previously been shown to regulate various aspects of eye-antennal disc development , but we also found new transcriptional regulators . For instance , the MADS-box transcription factor Myocyte enhancer factor 2 ( DMef2 ) was predicted to regulate a number of genes in various clusters . The detection of Dmef2 expression in lose cells attached to the developing eye-antennal discs ( S2 Fig ) confirmed that our result is not an artefact , but rather specific . Since , DMef2 is crucial for the development of muscle and heart tissue [134] these cells could be precursors of future head muscles . However , some recent findings could also hint towards an important role of this transcription factor in eye development since DMef2 has been implicated in circadian behavior [135] and larval eye and adult ocelli function [136] . These findings certainly encourage additional research on the possible role of DMef2 in photoreceptor cell development . Taken together , the combination of dynamic gene expression clustering and upstream factor enrichment provides an excellent basis for the identification of potential new regulators involved in a given biological process . We identified a number of putative direct transcriptional regulators , although the ChIP-seq experiments that identified the direct interaction of a transcription factor with its target genes were not specifically performed in eye-antennal disc tissue at the stages we studied here . Indeed , most data available in current databases is based on experiments in embryonic or adult stages [82 , 83 , 85 , 137 , 138] . Interestingly , the enrichment of Caudal in clusters 1 and 2 ( S2 Table ) is based on data from a ChIP-seq experiment performed in adult flies [82] , but does not represent an experiment performed in embryos [83] . This could indicate that Caudal has very different downstream targets during embryogenesis compared to its target genes at later stages . Although this observation may also indicate that the parameters and thresholds used in the different ChIP-seq experiments are very different , a large degree of tissue and stage specific target genes is expected . In the light of this specificity , we may miss eye-antennal disc specific target genes in our survey , but we are confident that one can identify a representative set of target genes to justify further tissue and stage specific ChIP-seq experiments if necessary . We failed to identify enriched motifs for classical retinal determination genes such as So , Ey or Glass ( Gl ) in any of the 13 clusters , due to the fact that respective ChIP-seq data is either not available or not included in the i-cisTarget database . Therefore , our analysis is biased towards those transcription factors for which ChIP-seq data is publicly available . We performed a preliminary power analysis to evaluate the potential to pick up an enrichment of So binding sites . We identified those clusters that contained 112 previously identified putative So/Eya target genes [47] and we found 76 of them distributed throughout all clusters ( S3 Table ) . The most representative clusters were cluster 2 with 17 ( of 656 genes in the cluster ) and cluster 10 with 15 ( of 935 genes in the cluster ) So/Eya target genes . Given that the significantly enriched transcription factors with relatively low NES scores of 3 . 7 ( Mef2 in cluster 2 ) and 3 . 5 ( dTcf in cluster 10 ) ( Fig 2 ) are predicted to regulate 57 and 559 target genes , respectively , this analysis very likely would not provide enough power to pick up a statistically significant enrichment . However , we detected an enrichment for Nejire although its putative target genes are present in 9 of 13 clusters . One explanation could be that the Nejire protein is detected ubiquitously in the entire eye disc [90] where it acts as a co-regulator [86–88] and thus influences the expression of many target genes . In contrast , classical retinal determination genes are dynamically expressed in more restricted domains , what is recapitulated by a highly complex regulation of these genes [139 , 140] and many different functions throughout eye-antennal disc development . Previous studies aiming at the identification of target genes resulted in limited lists of putative target genes for So ( 112 high confidence targets [47] ) and Ey ( 20 direct targets in the retinal part of the eye-antennal disc [46] ) suggesting that retinal determination genes regulate a restricted number of genes in a highly context dependent manner . The combination of our genome wide expression analysis with functional experiments revealed a novel role of the C2H2 zinc-finger transcription factor Hunchback ( Hb ) in carpet cell development . Carpet cells have been shown to be a sub-population of the subperineurial glia cells due to shared key cellular features , such as the formation of extensive septate junctions [34] . However , a well-established subperineurial glia cell driver ( NP2276 [141] ) does drive reporter gene expression in brain subperineurial glia , but we could not detect expression in carpet cells . In contrast , we only detected hb expression in carpet cells and not in any subperineurial glia cells in the larval brain ( results confirmed both using immunostaining and two driver lines ( VT038544 and VT038545 ) ) . Additionally , if we compare our list of putative Hb target genes with the 50 genes enriched in adult blood-brain barrier surface glia [51] , we only find Fas2 to be present in both datasets . All these data suggest that carpet cells are indeed a retina specific subperineurial glia cell type that is molecularly very distinct from brain subperineurial glia cells . The carpet cells function as a scaffold for undifferentiated retinal perineurial glia cells , which migrate into the disc to find the nascent axons of differentiating photoreceptors [34] and to guide them through the optic stalk [35] . In accordance with these roles , our Hb target gene analysis revealed many candidate target genes with GO terms related to axon guidance ( S5 Table ) . Upon loss of Hb function , the most obvious phenotype was the lack of polyploid cell nuclei and carpet cell membranes in the eye-antennal discs . In contrast to our results , previous studies have shown that carpet cell ablation or a reduction of their size causes over migration of perineurial glia cells anterior to the morphogenetic furrow [34 , 113] . The corresponding experiments are based on the induction of cell death in moody expressing cells [34] and thus affect not only the carpet cells , but also for instance all other subperineurial glia of the brain . Since hb is specifically expressed carpet cells , the phenotype obtained here may be more specific . It remains to be studied , however , how carpet cells and subperineurial glia cells of the brain may interact to regulate perineurial glia cell migration in the eye-antennal discs . Carpet cells possess a high level of plasticity since the loss of one carpet cell from the retinal field resulted in one larger carpet cell ( i . e . larger polyploid nucleus ) located in the midline of the eye field . In these cases , also no perineurial glia cell over migration could be observed and the membrane of this single cell seemed to sufficiently cover the full retinal field indicating that a single carpet cell could compensate the loss of the other one . The loss of hb expression affected the integrity of the blood-eye barrier , a subset of the blood-brain barrier ( Fig 7 ) . This phenotype was not as striking as previously published for moody mutant flies [39] , where all surface glia cells were affected , suggesting that the carpet cells may indeed only contribute to the retinal part of the blood-brain barrier ( i . e . the blood-eye barrier ) . Additionally , in our analysis of the moody positive cell membranes present in the optic stalk in hbRNAi flies we found still carpet cell membranes in the optic stalk in up to 80% of the knock-down discs ( Fig 6D ) . This is a similar proportion as we find in our blood-brain barrier assay ( Fig 7 ) . The largest portion of the blood-brain barrier is already established by the end of embryogenesis [142 , 143] , while the eye-antennal disc and developing photoreceptors seem to be accessible for the hemolymph during larval and very early pupal stages . Indeed , the final closure of the blood-brain barrier in the region where the optic stalk contacts the brain ( i . e . the blood-eye barrier ) is only established late during pupal development [109 , 144] . The rather late formation of the blood-eye barrier may be related to the dual role of carpet cells , which first migrate into the eye-antennal disc and only during pupal stages migrate back into the optic stalk towards the brain lobes . By mid-pupa stages they are located at the base of the brain lamina [40] , where they remain throughout adult life and form septate junctions that isolate the brain and retina from the hemolymph [109] . The use of the newly analyzed driver lines , which drive expression specifically in the carpet cells represent an excellent starting point to study the migration of these cells throughout late larval and pupae stages in more detail . The lack of ployploid large carpet cells during larval stages and the loss of blood-brain barrier integrity could either indicate a central role of Hb in specifying carpet cell identity entirely or a more specific role in defining aspects of carpet cell identity such as polyploidy and/or its migratory behavior . Based on our data , we propose the following cellular functions: First , the lack of polyploid nuclei could hint towards a role of Hb in regulating the extensive endoreplication process necessary to generate such huge cell nuclei . Indeed , in our target gene analysis we found archipelago ( ago ) as one potential target . Ago has been shown to induce degradation of CyclinE ( CycE ) [145] , a crucial prerequisite for efficient endoreplication cycles [146] . Second , Hb could be involved in establishing the migratory behavior of carpet cells . In the list of putative Hb target genes , we found many genes with GO terms related to cell migration and , some even specifically with the “glia cell migration” term . Additionally , many of the identified Hb target genes are involved in the epidermal growth factor ( EGF ) pathway that has various roles in development and cancer [147–149] including cell division , differentiation , cell survival and migration [150 , 151] . The list of Hb target genes up-regulated at 96h and 120h AEL in eye-antennal disc development includes both positive ( rhomboid , Star and CBP ) and negative regulators ( Fasciclin2 and sprouty ) of this pathway . Fas2 and sprouty are specifically expressed in carpet cells ( Fig 9C and 9D ) , suggesting that Hb may actively influence the migratory behavior of carpet cells by activating genes involved in EGFR signaling regulation . Another putative target gene of Hb is Ets98b ( Fig 8B ) , the ortholog of which has recently been shown to induce ectopic cell migration upon misexpression during early embryonic development in the common house spider Parasteatoda tepidariorum [152] . Finally , Nejire has been shown to be involved in glia cell migration in the peripheral nervous system in Drosophila [153] , and we found it as putative regulator of genes in cluster 13 . It remains to be established , whether Nejire and Hb may collaborate during carpet cell development . Although carpet cells fulfill fundamental functions , it is still unclear where these cells originate from . Based on observations by Choi and Benzer ( 1994 ) using the enhancer trap line M1-126 , these cells originate in the optic stalk where they are present at late L2 stage [36] . It has also been proposed that carpet cells may originate from a pool of neuroblasts in the neuroectoderm during embryogenesis [154] or in the optic lobes [155] . A clonal analysis using the FLP-out system suggests that various retinal glia cell types , including the carpet cells , originate from at least one mother cell at L1 larval stage [35] . Since only one polyploid cell nucleus seems to originate from one clone [35] and we show that in some loss of Hb imaginal discs only one polyploid cell nucleus is present , we propose that the two carpet cells may originate independently probably from two mother cells defined during L1 stages . Our observation that loss of Hb function resulted in loss of polyploid carpet cell nuclei when hbTS mutant flies were transferred to the restrictive temperature during the L1 larval stage , further supports an involvement of Hb during this stage ( Fig 5 and S5 Fig ) . The exact role of Hb , however , is still unclear and will require further in-depth analyses . The newly analyzed driver lines in combination with the extensive list of potential Hb target genes , of which many are likely to be expressed in this specific glia cell type , represents a valuable resource to address the questions concerning the origin of these cells in more detail . The identification of Hb has only been possible because we studied the dynamic expression profiles of all genes expressed during eye-antennal disc development . Since the RNA levels of hb in the entire eye-antennal disc are negligible , we could identify Hb as central factor only through the expression profiles of its putative target genes , which are steadily up-regulated throughout development . This up-regulation is very likely due to the large cell bodies of the carpet cells , which need to produce high amounts of cytosolic or membrane bound proteins . At earlier stages , carpet cells are not yet in the eye-antennal discs , and hb expression could only been have identified by studies focused on the optic stalk . Moreover , we could show that refining the putative Hb target genes by incorporating the expression data results in a list that contains genes with GO terms highly specific for the putative function of Hb in carpet cells . Following the stepwise identification of putative target genes , we could confirm a high number of those experimentally . All these findings demonstrate that the combination of high throughput transcript sequencing with a ChIP-seq data based transcription factor enrichment analysis can reveal previously unknown factors and their target genes , and therefore increase the number of connections within the underlying developmental GRNs . Other studies have searched for regulating transcription factors that were in the same co-expression clusters as its targets genes [44] . However , upstream orchestrators do not necessarily have the same expression levels as their targets . Therefore , the combination of ChIP-seq methods with RNA-seq co-expression analyses has proven to be a powerful tool to identify new developmental regulators that can complement other studies based on reverse genetics .
D . melanogaster ( OregonR ) flies were raised at 25°C and 12h:12h dark:light cycle for at least two generations and their eggs were collected in 1h windows . Freshly hatched L1 larvae were transferred into fresh vials in density-controlled conditions ( 30 freshly hatched L1 larvae per vial ) . At the required time point , eye-antennal discs of either only female larvae ( 96h and 120h AEL ) or male and female larvae ( 72h AEL ) were dissected and stored in RNALater ( Qiagen , Venlo , Netherlands ) . Please note that a mix of males and females was dissected for the 72h AEL samples because it is not possible to morphologically distinguish the sex at this stage . 40–50 discs were dissected for the 120h samples , 80–90 discs for the 96h samples and 120–130 discs for the 72h samples . Three biological replicates were generated for each sample type . Total RNA was isolated using the Trizol ( Invitrogen , Thermo Fisher Scientific , Waltham , Massachusetts , USA ) method according to the manufacturer’s recommendations and the samples were DNAseI ( Sigma , St . Louis , Missouri , USA ) treated in order to remove DNA contamination . RNA quality was determined using the Agilent 2100 Bioanalyzer ( Agilent Technologies , Santa Clara , CA , USA ) microfluidic electrophoresis . Only samples with comparable RNA integrity numbers were selected for sequencing . Library preparation for RNA-seq was performed using the TruSeq RNA Sample Preparation Kit ( Illumina , catalog ID RS-122-2002 ) starting from 500 ng of total RNA . Accurate quantitation of cDNA libraries was performed using the QuantiFluordsDNA System ( Promega , Madison , Wisconsin , USA ) . The size range of final cDNA libraries was determined applying the DNA 1000 chip on the Bioanalyzer 2100 from Agilent ( 280 bp ) . cDNA libraries were amplified and sequenced ( 50 bp single-end reads ) using cBot and HiSeq 2000 ( Illumina ) . Sequence images were transformed to bcl files using the software BaseCaller ( Illumina ) . The bcl files were demultiplexed to fastq files with CASAVA ( version 1 . 8 . 2 ) .
|
The development of different cell types must be tightly coordinated , and the eye-antennal imaginal discs of Drosophila melanogaster represent an excellent model to study the molecular mechanisms underlying this coordination . These imaginal discs contain the anlagen of nearly all adult head structures , such as the antennae , the head cuticle , the ocelli and the compound eyes . While large scale screens have been performed to unravel the gene regulatory network underlying compound eye development , a comprehensive understanding of genome wide expression dynamics throughout head development is still missing to date . We studied the genome wide gene expression dynamics during eye-antennal disc development in D . melanogaster to identify new central regulators of the underlying gene regulatory network . Expression based gene clustering and transcription factor motif enrichment analyses revealed a central regulatory role of the transcription factor Hunchback ( Hb ) . We confirmed that hb is expressed in two polyploid retinal subperineurial glia cells ( carpet cells ) . Our functional analysis shows that Hb is necessary for carpet cell development and we show for the first time that the carpet cells are an integral part of the blood-brain barrier .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"nervous",
"system",
"gene",
"regulation",
"regulatory",
"proteins",
"blood-brain",
"barrier",
"dna-binding",
"proteins",
"departures",
"from",
"diploidy",
"animals",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"transcription",
"factors",
"eyes",
"drosophila",
"research",
"and",
"analysis",
"methods",
"transcriptional",
"control",
"proteins",
"gene",
"expression",
"head",
"polyploidy",
"insects",
"arthropoda",
"biochemistry",
"eukaryota",
"anatomy",
"central",
"nervous",
"system",
"genetics",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"organisms"
] |
2018
|
Dynamic genome wide expression profiling of Drosophila head development reveals a novel role of Hunchback in retinal glia cell development and blood-brain barrier integrity
|
Both NK cells and CTLs kill virus-infected and tumor cells . However , the ways by which these killer cells recognize the infected or the tumorigenic cells are different , in fact almost opposite . CTLs are activated through the interaction of the TCR with MHC class I proteins . In contrast , NK cells are inhibited by MHC class I molecules . The inhibitory NK receptors recognize mainly MHC class I proteins and in this regard practically all of the HLA-C proteins are recognized by inhibitory NK cell receptors , while only certain HLA-A and HLA-B proteins interact with these receptors . Sophisticated viruses developed mechanisms to avoid the attack of both NK cells and CTLs through , for example , down regulation of HLA-A and HLA-B molecules to avoid CTL recognition , leaving HLA-C proteins on the cell surface to inhibit NK cell response . Here we provide the first example of a virus that through specific down regulation of HLA-C , harness the NK cells for its own benefit . We initially demonstrated that none of the tested HSV-2 derived microRNAs affect NK cell activity . Then we show that surprisingly upon HSV-2 infection , HLA-C proteins are specifically down regulated , rendering the infected cells susceptible to NK cell attack . We identified a motif in the tail of HLA-C that is responsible for the HSV-2-meduiated HLA-C down regulation and we show that the HLA-C down regulation is mediated by the viral protein ICP47 . Finally we show that HLA-C proteins are down regulated from the surface of HSV-2 infected dendritic cells ( DCs ) and that this leads to the killing of DC by NK cells . Thus , we propose that HSV-2 had developed this unique and surprising NK cell-mediated killing strategy of infected DC to prevent the activation of the adaptive immunity .
Human Natural killer ( NK ) cells comprise approximately 5–15% of peripheral blood lymphocytes . They kill infected or transformed cells and can also contribute to the activation of the adaptive immunity through the secretion of cytokines and chemokines [1] . Additionally , NK cells regulates adaptive immunity through the killing of autologous immune cells including activated T cells and DCs [2] . They can also kill autologous self cells such as beta cells [3] and stellate cells [4] . The activity of NK cells is controlled by the balance of signals delivered by inhibitory and activating receptors [5] , [6] . Thus , NK cells can be activated by induction in the expression of activating ligands and/or by reduction in the expression of inhibitory ligands [7] . A group of NK inhibitory receptors interacts specifically with MHC class I ( MHC-I ) proteins . These receptors prevent the NK cell-mediated attack of normal cells , whereas cells with compromised MHC-I expression become susceptible to NK cell-mediated killing [8] . The MHC-I molecules in humans comprise the classical HLAs: HLA-A , HLA-B and HLA-C , and the non-classical HLA-E , HLA-F and HLA-G molecules [9] . Practically all of the HLA-C alleles can be divided into two groups , in terms of NK cell recognition , based on the residue located at position 80 [10] . The HLA-C1 group , that includes for example HLA-Cw3 and HLA-Cw7 , is characterized by the presence of asparagine in position 80 and is recognized by the KIR2DL2 receptor . The HLA-C2 group , which includes proteins such as HLA-Cw4 and HLA-Cw6 , is characterized by the presence of lysine in position 80 and is recognized by the KIR2DL1 receptor [8] , [10] , [11] . Since , virtually all of the HLA-C molecules belong to either group 1 or group 2 it is thought that the HLA-C molecules were probably developed to primarily inhibit the NK cell activity . In marked contrast , cytotoxic T lymphocytes ( CTLs ) execute their cytolytic activity upon interaction with MHC-I proteins . Stable peptide/MHC-I complexes are assembled in the ER and transported to the cell surface where they are inspected by the T-cell receptors ( TCR ) of CTLs [12] . If the MHC-I proteins carry foreign peptides , derived from viral or tumor antigen , then the cells carrying these complexes will be killed by CTLs . Hence , MHC-I has a dual opposite role with regard to innate and adaptive immunity as on the one hand it inhibits NK cells activity and on the other hand it activates CTLs . Viruses in general and especially viruses that establish long-term infections in their hosts , have evolved a number of mechanisms to down regulate MHC-I expression [13] . However , a general down regulation of MHC-I proteins is problematic because although the infected cells will be protected from killing by CTLs , they will become more susceptible to NK cells attack . Therefore , some sophisticated viruses such as HIV developed mechanisms to specifically down regulate certain HLA alleles such as HLA-A and HLA-B , which are especially important for HLA-TCR interaction and leave on the cell surface the HLA-C proteins to inhibit the NK cell activity [14] , [15] . Herpes simplex virus type 2 ( HSV-2 ) is a double-stranded DNA virus , belonging to the Herpesviridae family . Its genome is very large ( 155 kb ) , encoding for at least 84 genes [16] . Each year around 500 , 000 people in the United States are infected with HSV-2 , and at least 22% of the population has a latent infection [17] . NK cells play an important role in fighting HSV-2 infections . Lower rate of mice survival and higher viral titers were observed in the vaginal mucosa of IL-15−/− deficient mice that lacks NK cells and in mice in which NK cells were depleted [18] . Moreover , NK cells were shown to be the main source of IFN-γ secreted in the mice genital tract in the first three days following HSV-2 infection [18] , [19] and in agreement with this , IFN-γ deficient mice were more susceptible to HSV-2 infection than wild type mice [18] . Finally , it was shown recently that genital HSV-2 infection induces short-term NK cell memory [20] . Herpes viruses are well known for their sophisticated immune evasion mechanisms . However , not much is known about the immune evasion strategies of HSV-2 . The best example with this regard is probably the HSV-2 ICP47 protein , a viral immediate early protein that blocks peptide binding to the transporter associated with antigen processing ( TAP ) and consequently reduces MHC-I expression [21] . The efficiency however of ICP47 is cell type dependent and in cells that express high levels of TAP ( such as antigen-presenting cells ( APC ) ) ICP47 poorly inhibits MHC-I antigen presentation [21] , [22] . Here we describe a novel and unique immune evasion mechanism of HSV-2 that is specific to humans as we demonstrate that HSV-2 , unlike any other virus known to date , specifically down regulates the expression of HLA-C . We identify the mechanism leading to this unexpected and surprising down regulation and we show that it leads to NK cell-mediated killing of HSV-2 infected DCs . We suggest that the virus had developed this unique strategy to interfere with activation of the adaptive immunity response .
We have previously demonstrated that the miRNAs of several herpes viruses: HCMV , KSHV and EBV target the stress-induced ligands of NKG2D to avoid attack by NK cells [23] , [24] . To test whether the miRNAs of HSV-2 are also involved in the regulation of NK cell killing we expressed 21 of HSV-2 microRNAs ( miRNAs ) in various human tumor cell lines that endogenously express NK cells ligands ( Figure 1 , summarizes the human cell lines that were transduced with each of the 21 miRNAs of HSV-2 and ( Figure 2 ) ) . To our disappointment , none of the HSV-2 miRNAs affected the expression of MHC-I , MICA , MICB , ULBP1 , ULBP2 , ULBP3 , PVR and ICAM ( data not shown , Figure 1 and selected examples in Figure 2 ) . To determine whether HSV-2 has developed other strategies to escape NK cell cytotoxicity we infected primary Human Foreskin Fibroblast ( HFF ) cells either with HSV-2 or with a UV-inactivated HSV-2 virus which is able to infect the cells but is unable to de novo produce viral gene transcripts . As can be seen in figure 3 , HSV-2-infected HFF cells were killed more efficiently than the uninfected ones . Similar results were obtained at various time points post infection ( data not shown ) . This enhanced killing was due to the production of viral gene products as infection with a UV-inactivated virus did not enhance the NK cell killing of HFF cells ( Figure 3 ) . Thus , surprisingly and in contrary of what we have expected , HSV-2 viral products are involved in enhancing NK cell cytotoxicity . The enhanced killing observed following HSV-2 infection might be explained either by a reduction in the expression of inhibitory ligands or by an up regulation of activating ligands . To investigate this we infected the HeLa cell line ( since it expresses numerous NK activating ligands ) with HSV-2 and monitored the expression of various NK cell ligands , at different time points following infection . The infection was performed with low MOI of 0 . 1 and therefore expression of the various ligands could be examined over relatively long period of time ( 144 hours , Figure 4 ) until the cells had died due to the infection . As can be seen in figure 4 , 48 hours following HSV-2 infection we observed some reduction in the expression of several NK ligands , but this reduction was not consistent as it was not observed 144 hours post infection . Therefore we did not consider this reduction meaningful . The only consistent change we observed over time was the reduction in the expression of MHC-I proteins ( Figure 4 ) . Similar results were observed when HFF cells were used ( data not shown ) . To examine whether the HSV-2-mediated MHC-I down regulation is a general mechanism or whether it affects only specific MHC-I alleles we have used the class I negative cell line 721 . 221 ( 221 ) and 721 . 221 cells that were transfected with various MHC-I proteins: two HLA-A alleles -A2 and -A3 , two B alleles -B8 and -B73 , three HLA-C alleles -Cw3 , -Cw4 and -Cw6 and two non-classical MHC-I proteins HLA-E and HLA-G ( Figure 5A ) . The various transfectants were infected with HSV-2 ( MOI of 0 . 5 ) and the levels of the various HLA class I proteins were evaluated . Remarkably , a marked reduction of HLA-C expression was observed following HSV-2 infection , whereas little or no reduction in the expression levels of other MHC proteins was noted ( Figure 5A ) . No reduction of HLA-C expression was detected when a UV-inactivated virus was used ( data not shown ) . The specific reduction of HLA-C expression was not restricted to a certain HSV-2 virus strain , as infection with 2 other clinical isolates of HSV-2 gave similar results ( Figure 5B ) . Furthermore , although HSV-1 and HSV-2 are approximately 50% homologous [25] , the specific down regulation of HLA-C was not observed following HSV-1 infection ( Figure 5B ) . We next tested the kinetics of HLA-C down regulation and observed a substantial reduction in HLA-C expression as early as 4 hours post infection ( Figure 5C ) . Sometimes we also observed a slight reduction in other MHC-I alleles such as HLA-B73 . However , this reduction was not consistent and as can be seen in figure 5D ( which summarizes several FACS experiments performed with infected 721 . 221 transfectants ) , the HSV-2-mediated reduction of MHC-I proteins other than HLA-C was minimal . To test if the reduction in HLA-C expression will affect NK cell inhibition we infected the parental 221 cells , 221 cells expressing B8 and 221 cells expressing Cw6 with HSV-2 and verified that the infection lead , as above , to specific down regulation of HLA-C ( Figure 6A ) . Infected and non-infected cells were next subjected to killing by KIR2DL1-positive NK clones ( example is shown in Figure 6A , bottom ) as this receptor recognizes HLA-Cw6 [10] , [26] . As expected , inhibition of NK cytotoxicity was observed only with uninfected cells expressing HLA-Cw6 ( Figure 6B ) . This inhibition resulted from the interaction between KIR2DL1 and Cw6 as it was abrogated following blockage by anti-KIR2DL1 antibody ( Figure 6B ) . Importantly , upon HSV-2 infection the inhibition was lost and the infected Cw6 expressing cells were killed as efficiently as all other cells ( Figure 6B ) . Similar results were obtained with additional KIR2DL1-positive NK clones ( data not shown ) . To elucidate the mechanism leading to the HSV-2-mediated down regulation of HLA-C we concentrated on the cytoplasmic tail of HLA-C molecules . This is because a similar , yet opposite , tail-dependent mechanism is used by the HIV Nef protein to down regulate HLA-A and HLA-B molecules and not HLA-C [14] . We replaced the tail of HLA-Cw4 with the tail of HLA-A2 and the tail of HLA-Cw6 with the tail of HLA-G . The various proteins were expressed in 221 cells resulting in the generation of the following transfectants: 221 Cw4 , 221 Cw6 , 221 Cw4 tail of A2 and 221 Cw6 tail of HLA-G . All transfectants were next infected with HSV-2 and the expression levels of the various HLA class I proteins were evaluated . As can be seen in figure 7 , HSV-2 indeed specifically targets the tail of HLA-C because when it was replaced , either with the tail of A2 or with the HLA-G tail , the down regulation of HLA-C proteins ( either Cw4 or Cw6 ) was abolished . To identify the HLA-C residue/s that are involved in the HSV-2-mediated HLA-C down regulation we searched for residues which are found in the tail of HLA-C alleles and absent in other HLA class I molecules . Four such residues were found: C320 , N327 , E334 and I337 ( Figure 8A ) . To study which of the four amino acids is important for the HSV-2-mediated HLA-C down regulation , we inserted point mutation ( s ) in each of the HLA-C-specific amino acid residues of HLA-Cw6 converting them into the corresponding amino acids that are present in the cytoplasmic tail of HLA-B73 and other B alleles ( Cw6 C320Y , Cw6 N327D , Cw6 E334V , for some unknown reason expression of the Cw6 I337T mutation was not detected on the cell surface ) . The reciprocal mutations in which the HLA-C residues were inserted into the B73 proteins were also performed ( B73 Y320C , B73 D327N , B73 V334E and B73 T337I ) . All 7 proteins were expressed in 721 . 221 cells and the expression of the corresponding molecules was evaluated before and after HSV-2 infection . Importantly , we discovered that residue 334 ( E in Cw6 ) is the primary residue involved in the HSV-2-mediated HLA-C down regulation . This is because the Cw6 down regulation was almost completely abolished when this residue was mutated ( Figure 8B ) . And vice versa , mutating the B73 residue in position 334 from V to E resulted in efficient down regulation of B73 following HSV-2 infection ( Figure 8B ) . As above , although sometimes , little down regulation of HLA-B was also observed following HSV-2 infection ( Figure 8B ) , this down regulation was much weaker as compared to Cw6 and the most efficient down regulation of B73 was observed only when its V residue at position 334 was replaced with E . Thus , glutamic acid in position 334 of Cw6 is targeted by HSV-2 . It is well establish that the ICP47 proteins of HSV-2 inhibit TAP in epithelial cells and that this consequently leads to HLA down regulation [22] . However , it is also known that APCs are not sensitive to the ICP47-mediated TAP inhibition . Thus , it seemed to us unlikely that the virus will develop a protein the inefficiently inhibits TAP in APCs and we were wondering whether ICP47 has additional functions which are TAP-independent . Furthermore , it is known that often different immune evasion mechanisms are utilized by the same viral protein [27]–[29] . Therefore , we hypothesized that ICP47 might be responsible for the specific HLA-C down regulation that occurs following HSV-2 infection . To investigate this , we cloned the ICP47 cDNA into a lenti virus vector that also contains GFP and transduced 221 cells and various 221 transfectants expressing various MHC-I proteins . The transduction efficiency ( as determined by the GFP expression ) was high ( data not shown ) and the analysis was performed on the GFP-positive cells . As can be seen in figure 9A when ICP47 was expressed in 721 . 221 cells expressing HLA-A2 , HLA-B73 or HLA-B8 a modest down regulation was observed . Importantly , the expression of HLA-C was completely abolished following the ICP47 transduction ( Figure 9A ) . To test whether this complete ICP47-dependent reduction of Cw6 is dependent on the E334 residue we transduced 221 cells expressing either Cw6 E334V or B73 V334E with lenti viruses expressing ICP47 . The transduction efficiency was very high ( data not shown ) and the analysis was performed on the GFP-positive cells . As can be seen in figure 9B , the complete ICP47-mediated reduction that was observed in 221 Cw6 cells was not observed the 221 Cw6 E334V cells and some expression of Cw6 E334V was maintained on the cell surface even in the presence of ICP47 ( compare figures 9A and B ) . The reciprocal picture was seen in the transduced 221 HLA-B73 cells . When ICP47 was expressed in 221 B73 V334E cells a complete down regulation was observed in contrast to 221 HLA-B73 cells in which only partial down regulation was detected ( compare figures 9A and B ) . Thus , the ICP47 protein of HSV-2 down regulates HLA expression using two different mechanisms: ( i ) through TAP inhibition ( a general , inefficient mechanism that function primarily in non-APCs ) and ( ii ) through its interaction with glutamic acid residue present in position 334 of the cytoplasmic tail of HLA-C ( a specific HLA-C-dependent mechanism ) . We next wonder why a virus would develop a mechanism ( specific down regulation of HLA-C ) that enables a better killing of the infected cells . Our assumption was that HSV-2 induces a selective down regulation of HLA-C proteins , primarily in APCs , to interfere with the activation of the adaptive immunity . Indeed it is known that DCs express the HSV-2 entry receptors , HVEM and nectin-2 and are therefore highly susceptible to HSV infection [30] . Hence we decided to test whether HSV-2 infection of DC will result in specific HLA-C downregulation and consequently increased NK cell killing . Testing for selective down regulation of HLA-C on DC is difficult due to the lack of HLA-C specific mAbs . Therefore , to overcome this problem , we used fusion proteins composed from the extra cellular portions of KIR2DL1 and KIR2DL2 ( that recognizes specifically the entire spectrum of HLA-C proteins [8] ) fused to human IgG1 ( KIR2DL1-Ig and KIR2DL2-Ig , respectively ) . The general expression levels of MHC-I proteins were determined by using the pan anti-MHC-I mAb W6/32 . As can be seen in figure 10A , HSV-2 infection of DC resulted in reduced expression of HLA-C , as it abolished the binding of both KIR2DL1-Ig and KIR2DL2-Ig fusion proteins . This specific reduction was observed as early as 8 hours following HSV-2 infection . In contrast , little or no reduction in the general expression levels of MHC-I proteins was observed . This is probably because the levels of HLA-C is only around 10% of that of HLA-A and HLA-B proteins [31] . Similar results were obtained with other DCs derived from additional donors ( data not shown ) . Next , we tested whether the down regulation of HLA-C will affect the NK cell mediate killing . For that we incubated primary bulk NK cell cultures with HSV-2-infected and uninfected DCs and stained the NK cells for the expression of CD107a ( LAMP-1 , a marker for cytoplasmic granules release and NK cell degranulation ) . As can be seen , following DC infection , 34% and 41% of the NK cells degranulated ( as evidence from the CD107a staining ) , 8 and 18 hours post infection respectively , while incubation with uninfected DCs led to only 3% degranulation ( Figure 10B ) .
NK cells and viruses co-evolved for millions of years . This co-existence led to the development of opposite mechanisms in which the NK cells try to kill the infected cells and the viruses on the other hand try to escape such killing [32] . Famous among these viruses are viruses of the Herpesviride that had developed numerous immune evasion mechanisms such as the use of viral miRNAs in order to escape NK cell attack [23] , [24] , [33] . Hence , we were quite surprised to observe that no changes were detected in the expression of the NK ligands tested , after transduction with 21 miRNAs of HSV-2 and that increased NK cell killing was observed following HSV-2 infection . Indeed it was shown that NK cells play an important role in controlling HSV-2 infections . It was previously shown that IL-15−/− mice , which lack NK and NKT cells , were more susceptible to HSV-2 infection than mice lacking both T and B cells [18] . In addition , it was demonstrated that CCR5 deficient mice suffer from increased susceptibility to genital challenge with HSV-2 . The absence of CCR5 had no significant impact on T-cell mobilization or recruitment to sites of infection , rather , NK cell infiltration was diminished , and a reduction in NK cell activity was observed [34] . Finally , it was recently shown that human NK cells contributed significantly to the stimulation of CD4 T lymphocytes through direct presentation of HSV1/2 antigens to CD4 T lymphocytes [35] . To understand the reasons accounting for the increased killing of HSV-2 infected cells , we evaluated the expression of various NK ligands following infection and observed reduction in the expression levels of MHC-I proteins in infected cells . This finding was quite expected as it was shown that HSV-2 encodes for the ICP47 protein that binds TAP and therefore leads to a general MHC-I down regulation [36] . However , ICP47 does not efficiently down regulate MHC-I in cells expressing high levels of TAP such as B cells and DCs [22] . Indeed we observed , when using the EBV transformed B cell line 721 . 221 and BM-derived human DCs that HSV-2 infection leads to an allele-specific down regulation of HLA-C , which consequently rendered the infected 721 . 221 cells and DC susceptible to NK cell killing . Moreover , we showed that HSV-2 through its ICP47 protein down regulates HLA-C by targeting the glutamic acid residue present in position 334 of the cytoplasmic tail of HLA-C . An interesting question is why HLA-C contains this residue and why it is not mutated to avoid this virus tactic . One possible answer is that this residue also regulates the expression of HLA-C under normal conditions and therefore it is indispensible . Indeed , HLA-C is expressed in lower levels as compared with HLA-A and HLA-B [31] and it was shown that the surface expression of HLA-C is regulated through multiple mechanisms , some of which are unknown [37] . Cellular pathways/proteins that are important for immune evasion of herpes viruses are often targeted by several viral proteins and sometimes even by viral proteins combined with viral miRNAs , as we have shown regarding the NKG2D ligands [23] , [24] , [38] . With regard to MHC-I it was shown , for example , that HCMV encodes at least four proteins ( US2 , US3 , US6 and US11 ) that function to reduce MHC-I expression . HCMV also encodes for the UL18 protein that binds the inhibitory receptor ILT-2 and for the UL40 protein that provides a leader peptide , which is presented by HLA-E [15] . Furthermore , occasionally a single viral protein possesses multiple immune evasion properties [27]–[29] . For example , the latent membrane protein 1 ( LMP1 ) of Epstein-Barr virus ( EBV ) mimics CD40 signaling , functions as a viral oncogene , affects cell-cell contact , cytokine and chemokine production and antigen presentation [29] . Another example is the K5 protein that functions as E3 ligase and degrades several immune-related ligands [27] , [39] . Thus we hypothesized that the ICP47 protein which was shown previously to affect MHC-I expression in non-APC is responsible for the specific HLA-C down regulation observed in APC . Importantly we demonstrate that when ICP47 was expressed in cells expressing Cw6 a complete down regulation was observed and when the E334 residue of Cw6 was converted to V the expression of Cw6 E334V was detected on the cell surface even in the presence of ICP47 . The reciprocal picture was seen with regard to HLA-B73 . The expression of Cw6 E334V in the presence of ICP47 was not completely restored probably because ICP47 still function as TAP inhibitor in these B cells ( as it expressed in high levels ) . In addition , mutating the E334 residue into V did not completely restored the expression of Cw6 in cells infected with HSV-2 , suggesting that additional elements in the tail of Cw6 might be involved in the ICP47-mediated down regulation . We therefore propose that ICP47 affects HLA expression by using two different mechanisms 1 ) Via inhibition of the TAP transporter which leads to a general down regulation of various HLA-I proteins especially in cells that express low levels of TAP and 2 ) Via a specific down regulation of HLA-C molecules , a mechanism that is primarily dependent on the E334 residue located in the HLA-C tail that occurs in cells expressing high levels of TAP such as APC . It was shown that immature DCs express unknown ligands for the NK killer receptors NKp46 and NKp30 and that they are killed by NK cells , while mature DCs are protected from killing because they express high levels of MHC-I proteins [40] . We demonstrated here that the down regulation of HLA-C following DC infection with HSV-2 renders them susceptible to NK cell killing . We therefore suggest that in the infected tissue the virus uses ICP47 to avoid CTL attack and in DCs it uses ICP47 to specifically target HLA-C and consequently to render DCs susceptible to NK cell elimination . By killing DCs which are central activators of T helper cells and CTLs , through cross presentation , HSV-2 avoids adaptive immune responses . Interestingly , a similar example was described before . The ICP10 protein of HSV-2 induces apoptosis in immune cells , while its expression protects epithelial cells from apoptosis [41] . HSV-2 is a human pathogen that co-evolved with its human host for millions of years . Thus , the virus had developed mechanisms to avoid specifically the human immune response . ICP47 , for example , blocks TAP in human fibroblasts , however , almost no inhibition of the mouse TAP is observed in a variety of mouse cells , unless ICP47 is applied in high concentrations ( 50 to 100-fold higher than those required to inhibit the human TAP ) [21] , [22] . More importantly , the new mechanism that we have discovered here in which the virus target HLA-C proteins is unique to humans as HLA-C is not expressed in the mouse . This novel human-specific immune evasion mechanism in which the virus harness NK cells for its own benefit demonstrates yet again the sophistication of herpesviruses in general and of HSV-2 in particular .
The NK cells that were used in this study were obtained from the blood of healthy voluntaries . The intuitional Helsinki committee of Hadassah approved the study ( Helsinki number 0030-12-HMO ) . All subjects provided a written informed consent . RNA artificial hairpins that function as orthologs of pre-miRNA hairpins were generated by using the pTER vector . Two complementary specific oligonucleotides were annealed , phosphorylated using T4 polynucleotide kinase as was previously described [42] and inserted into the pTER vector . The artificial hairpin and H1 RNA polymerase III promoter were excised from the vector and cloned into the lentiviral vector SIN18- pRLL-hEFIap-EGFP-WRPE [43] . The lentiviral vector contains GFP , thus allowing the simultaneous expression of both reporter GFP and miRNA . Lentiviral viruses were produced by transient three-plasmid transfection as described [43] . The KIR2DL1-Ig and KIR2DL2-Ig fusion proteins were generated in COS-7 cells and purified by affinity chromatography using a protein G column , as described previously [44] . The various 721 . 221 expressing HLA proteins were stained using W6/32 mAb directed against MHC-I , anti human HLA-E mAb ( clone MEM-E07 , Serotec ) and anti human HLA-G mAb ( clone MEM-G9 , Serotec ) . Anti-MICA , anti-MICB , anti-ULBP1 , anti-ULBP2 , anti-ULBP3 , anti- ICAM1 and anti-PVR antibodies were all purchased from R&D Systems ( Minneapolis ) . The staining of cell lines was visualized using a secondary Allophycocyanin ( APC ) conjugated goat anti-human Abs ( Jackson ImmunoResearch Laboratories , West Grove , PA ) and a secondary APC conjugated goat anti-mouse Abs ( Jackson ImmunoResearch Laboratories , West Grove , PA ) . The 721 . 221 , RKO , DU145 , HeLa , Hep3b , 293T and MCF7 cell lines were used . 721 . 221 transfectants ( A2 , A3 , B8 , B73 , Cw3 , Cw4 , Cw6 , HLA-E , HLA-G , Cw4 tail of A2 and Cw6 tail of HLA-G ) were generated previously as described [14] . Infection with three HSV-2 clinical isolations and HSV-1 17+/pR20 5/5 strain was done at a multiplicity of infection ( MOI ) as described in each experiment . Primary NK cells from healthy donors were isolated from PBLs using the human NK cell isolation kit and the autoMACS instrument ( Miltenyi Biotec , Auburn , CA ) according to the manufacturer's instruction . They were grown as described [10] . KIR2DL1 positive NK clones were identified by flow cytometry using the anti-KIR2DL1 mAb HP3E4 . For mutating the MHC-I proteins , we amplified two overlapping fragments of the gene by PCR . The upstream fragment was amplified using a gene-specific 5′ primer ( including the BamHI restriction site ) and an internal 3′ primer that contains the mutation . The downstream fragment was amplified using an internal 5′ primer containing the mutation and a gene-specific 3′ primer ( including Bsp1407I restriction site ) . Next , both purified fragments were mixed together with the 5′ primer and the 3′ primer to generate the mutated full-gene cDNA . Mutations in the end of the 3′ were performed by using the gene-specific 5′ primer and 3′ primer that includes the mutation . The various cDNAs were cloned into the lentiviral vector SIN18-pRLL-hEFIp-EGFP-WRPE by removing the GFP gene and inserting the HLA gene and used to infect 721 . 221 cells . Lentiviral viruses were produced by transient three-plasmid transfection as described [43] . These viruses were used to transduce 721 . 221 cells in the presence of polybrene ( 5 g/ml ) . The cytotoxic activity of NK cells against target cells was assessed in 5-hour 35S-release assays , as described [45] E:T ratio was 10∶1 . The final concentration of the blocking antibodies ( anti-KIR2DL1 mAb HP3E4 ) was 2 . 5 µg/ml . NK cells ( 5×105 ) were co-incubated with DCs in a ratio of 1∶1 in the presence of 0 . 2 µg of a Bioten-conjugated CD107a Ab ( 1D4B; Southern Biotechnology Associates , Birmingham , AL ) for 2 h . Bioten-conjugated CD153 Ab was used as control , the staining was visualized using a secondary R-Phycoerythrin ( PE ) -conjugated streptavidin ( Jackson ImmunoResearch Laboratories , West Grove , PA ) . NK cells were identified by APC-conjugated CD56 mAb ( HCD56; Biolegend , San Diego , CA ) . Afterward; cells were washed and analyzed by FACS . The ICP47 the gene was amplified by PCR from cDNA of 221 HSV-2 infected cells . The cDNA was cloned into the lentiviral vector DsRed ( − ) wich also includes a reporter GFP . Lentiviral viruses were produced by transient three-plasmid transfection as described [43] these viruses were used to transduce 721 . 221 cells in the presence of polybrene ( 5 g/ml ) .
|
Approximately 20% of all humans are latently and asymptomatically infected with HSV-2 . This suggests that the virus developed mechanisms to avoid immune cell detection; many of which are still unknown . Infected cells are killed mainly by two lymphocyte populations; NK cells and CTLs that belong to the innate and the adaptive immunity , respectively . While the killing machinery of these two cell types is similar , almost identical , the ways by which they discriminate between infected and uninfected cells is different . CTLs are activated , primarily by DCs , to become effector cells . They then recognize virus-derived peptides in the groove of MHC class I molecules and eliminate the virally infected cells . In contrast , NK cells recognize infected cells through several NK cell activating receptors , while the recognition of MHC class I proteins by NK cells leads to inhibition of NK cell killing . Viruses , such as HIV , developed mechanisms to interfere with the function of both NK cells and CTLs via targeting of specific MHC class I proteins . Here we show that HSV-2 developed a MHC class I-dependent mechanism in which the virus , through specific targeting of HLA-C by the viral protein ICP47 , harness the NK cells for its own benefit , probably to avoid the activation of adaptive immune response .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"infectious",
"diseases",
"immunology",
"biology"
] |
2013
|
HSV-2 Specifically Down Regulates HLA-C Expression to Render HSV-2-Infected DCs Susceptible to NK Cell Killing
|
Plasmodium falciparum is a highly lethal malaria parasite of humans . A major portion of its life cycle is dedicated to invading and multiplying inside erythrocytes . The molecular mechanisms of erythrocyte invasion are incompletely understood . P . falciparum depends heavily on sialic acid present on glycophorins to invade erythrocytes . However , a significant proportion of laboratory and field isolates are also able to invade erythrocytes in a sialic acid-independent manner . The identity of the erythrocyte sialic acid-independent receptor has been a mystery for decades . We report here that the complement receptor 1 ( CR1 ) is a sialic acid-independent receptor for the invasion of erythrocytes by P . falciparum . We show that soluble CR1 ( sCR1 ) as well as polyclonal and monoclonal antibodies against CR1 inhibit sialic acid-independent invasion in a variety of laboratory strains and wild isolates , and that merozoites interact directly with CR1 on the erythrocyte surface and with sCR1-coated microspheres . Also , the invasion of neuraminidase-treated erythrocytes correlates with the level of CR1 expression . Finally , both sialic acid-independent and dependent strains invade CR1 transgenic mouse erythrocytes preferentially over wild-type erythrocytes but invasion by the latter is more sensitive to neuraminidase . These results suggest that both sialic acid-dependent and independent strains interact with CR1 in the normal red cell during the invasion process . However , only sialic acid-independent strains can do so without the presence of glycophorin sialic acid . Our results close a longstanding and important gap in the understanding of the mechanism of erythrocyte invasion by P . falciparum that will eventually make possible the development of an effective blood stage vaccine .
The erythrocyte invasion mechanisms of P . falciparum are varied and complex . Erythrocytes are rich in surface glycophorins which contain sialic acid . Earlier studies demonstrated that invasion of erythrocytes could be inhibited by treatment of erythrocytes with neuraminidase , which removes sialic acid , or blocked by purified glycophorin A [1]–[3] . In addition , erythrocytes genetically deficient in glycophorin A [En ( a- ) ] , glycophorin B ( S-s-U- ) , or sialic acid ( Tn ) showed reduced invasion compared to normal cells [1] , [4] . These studies suggested that sialic acid and the peptide backbones of glycophorin A and B play a role in the invasion of erythrocytes by P . falciparum . Glycophorin C was later added to this list [5] . However , some malaria strains exhibit sialic acid-independent invasion [6]–[8] , which is not affected by the absence of glycophorin A or B [6] or by antibodies against glycophorin A [7] , but it is trypsin sensitive [6] , [9] , [10] . The putative sialic acid-independent trypsin-sensitive receptor on erythrocytes has been referred to as “X” [9] . The relevance of the sialic acid-independent pathway has been shown by the demonstration that many field isolates can utilize this invasion pathway [11] , [12] . Therefore , the identification of the host and parasite molecules involved in this pathway is a necessary step towards the ultimate goal of developing a malaria vaccine that effectively blocks erythrocyte invasion . We considered CR1 to be a good candidate for being “X” because , like “X” , CR1 is known to be highly sensitive to trypsin and does not contain sialic acid [13] . In addition , CR1 is used by several pathogens as a receptor for entry into host cells [14] , [15] . Structurally , it is a 200 kD integral membrane protein found on all erythrocytes and most leukocytes and it is composed of 30 complement control protein ( CCP ) modules which can be organized , based on their degree of homology , into long homologous repeats ( LHR ) A–D [16] . It also serves as a co-factor to Factor I for the inactivation of C3b and accelerates the decay of C3 and C5 convertases and thus protects cells from autologous complement attack [17] . CR1 also mediates the binding of mature infected erythrocytes ( schizonts ) to uninfected erythrocytes ( rosetting ) [18] .
In order to test the hypothesis that CR1 is the sialic acid-independent receptor on erythrocytes we used the P . falciparum sialic acid-independent laboratory strain 7G8 [6] . We initially used a chicken polyclonal anti-human CR1 that recognizes CR1 on red cells specifically as measured by flow cytometry ( Figure 1A and 1B ) and is capable of immunoprecipitating CR1 from a red cell lysate ( Figure 1C ) . Incubation of neuraminidase-treated erythrocytes with either sCR1 [19] or anti-human CR1 Fab blocked 7G8 invasion of neuraminidase-treated erythrocytes in a dose-dependent manner ( Figure 2A–D ) but had no discernible effect on the invasion of untreated erythrocytes . In order to determine whether we were working at excess concentrations of antibodies and sCR1 we measured the invasion inhibition by anti-CR1 antibody and sCR1 under increasing starting parasitemia . We observed stable inhibition of invasion within a wide range of starting parasitemias ( Figure S1 ) . Next , to further test the specificity of our findings and narrow down the binding site within CR1 we used a panel of monoclonal antibodies directed against various CR1-defined CCPs: J3D3 , E11 , To5 , and J3B11 . J3B11 and J3D3 were the most effective in blocking invasion ( Figure 2E ) . These monoclonal antibodies are known to bind to adjacent CCPs within the C3b binding sites and the CR1 binding site may overlap these two epitopes [20] . J3B11 also interferes with the binding of PfEMP-1 [21] , the putative malaria ligand that mediates rosetting and cytoadherence of schizonts to endothelial cells [18] , [22] . Although published data suggest that J3D3 and To5 bind within the same group of CCPs [20] , the fact that To5 showed minimal , if any , inhibition suggests that finer epitope mapping of this antibody may demonstrate that its epitope is distinct from J3D3 . E11 recognizes CCPs that also contain the epitopes for J3D3 and J3B11 but its epitope seems to be more conformationally dependent [20] . To further confirm our results and to look for synergy between J3B11 and J3D3 we repeated these studies using these monoclonals singly or in combination ( Figure 3 ) . The results showed that at 1 µg/ml there was some synergistic effect resulting from the combination of antibodies . However , at concentrations of 5 µg/ml or higher J3B11 was more effective than J3D3 and there was no synergy . Therefore , J3B11 appears to be s binding closest to the ligand binding site in CR1 . In an effort to test whether our results represent a general phenomenon , we examined several additional P . falciparum sialic acid-independent laboratory strains ( HB3 , 3D7 , and Dd2NM ) [8] , [9] . In two of the three laboratory strains ( HB3 and 3D7 ) anti-CR1 Fab or sCR1 resulted in a reduction of invasion of untreated erythrocytes that was statistically significant ( Figure 4 ) . Interestingly , the irrelevant anti-CD55 monoclonal antibody 4D3 reduced invasion of untreated erythrocytes by HB3 but had no effect on the other strains . In contrast to untreated erythrocytes , invasion of neuraminidase-treated cells was consistently reduced in all strains by the use of polyclonal or monoclonal antibodies against CR1 as well as by sCR1 . Monoclonal and polyclonal antibodies were least effective in blocking HB3 invasion and most effective in blocking 3D7 invasion . sCR1 was usually more effective than antibodies in reducing the sialic acid-independent invasion of red cells ( >90% for HB3 and 3D7 , and 74% for Dd2NM ) . Chicken IgY Fab control had a paradoxical effect by increasing the invasion of HB3 . Altogether , these results suggest that the use of CR1 as an erythrocyte receptor is common in sialic acid-independent strains , but strains such as HB3 and Dd2NM may utilize additional sialic acid-independent receptors . In addition to the laboratory strains mentioned above , we examined the role of CR1 in the invasion of red cells in three sialic acid-independent wild strains from Kenya ( Table S1 ) . Like the laboratory strains tested , the invasion of neuraminidase-treated red cells was inhibited by sCR1 and chicken anti-CR1 . However , unlike laboratory strains , we observed modest but significant inhibition of invasion of intact red cells especially with sCR1 . To demonstrate direct interaction between merozoites and CR1 , we performed immunofluorescent microscopy using freshly released merozoites . In agreement with previous observations [23] , CR1 appeared as a punctate or speckled pattern on red cells which disappeared upon trypsinization ( Figure S2 ) . Merozoites were observed interacting directly with CR1 on erythrocytes both in neuraminidase-treated and untreated control cells ( Figure 5 , Figure S3 , and Supplementary 3D Videos S1 , S2 , S3 , S4 , S5 , and S6 ) . Although not all merozoites were seen to interact directly with CR1 , examples of this interaction were easy to find . When interaction was observed , CR1 staining always appeared to be more intense around the merozoite , suggesting that capping may be occurring . In addition , while the interaction often appeared to be focal , in some instances merozoites seemed to be partially or completely coated by CR1 ( Figure S3B and S3D ) . We observed no clearly discernible difference between 7G8 and the sialic acid-dependent strain Dd2 by immunofluorescent microscopy ( data not shown ) , suggesting that both strains can interact with CR1 or that this technique lacks sensitivity to detect differences in the way the two strains interact with CR1 . To further document the interaction of merozoites with CR1 , we incubated enriched schizonts and late trophozoites with polystyrene microspheres coated with sCR1 , BSA , glycophorin A , or fetuin . Figure 6 shows that merozoites bind to CR1-coated microspheres with greater frequency than to microspheres coated with BSA , glycophorin A , or fetuin . Further , the binding to sCR1-coated microspheres was inhibited by chicken anti-CR1 whereas the binding to BSA-coated microspheres was not inhibited by polyclonal rabbit anti-BSA . Based on our findings , we predicted a correlation between CR1 expression level and the level of P . falciparum invasion , especially in neuraminidase-treated cells . In lieu of the absence of individuals with complete CR1 deficiency , we took advantage of the fact that CR1 is known to have two quantitative alleles ( H and L ) that are co-dominant and result in the distribution of CR1 levels into low ( LL ) , medium ( HL ) , and high ( HH ) expressors [16] . We used flow cytometry to determine the CR1 median fluorescence intensity ( MFI ) of 75 healthy individuals . Erythrocytes of 27 individuals , spanning the range of low , medium , and high expressors , were used for invasion assays . A positive correlation was found between CR1 MFI and invasion for both intact and neuraminidase-treated red cells . Furthermore , this relationship was much stronger in neuraminidase-treated cells than in untreated cells ( Figure 7 ) , which is consistent with a greater role of CR1 as a receptor in the absence of sialic acid . Finally , we compared the P . falciparum invasion of human CR1 transgenic mouse erythrocytes [23] to wild-type mouse erythrocytes . In the mouse , CR1 is expressed mostly on B cells and not on red cells [24] . Although mouse glycophorins contain sialic acid , P . falciparum invades mouse erythrocytes at a lower rate than human erythrocytes [25] probably due to structural differences between human and mouse glycophorins [26] . P . falciparum 7G8 invaded CR1 transgenic mouse erythrocytes preferentially over wild-type erythrocytes ( Figure 8 ) . Anti-CR1 monoclonal J3B11 and polyclonal anti-CR1 Fab reduced the invasion of transgenic erythrocytes by 7G8 down to wild-type levels ( Figure 8B ) . While the sialic acid-dependent strain Dd2 also showed increased invasion of CR1 transgenic mouse erythrocytes , following neuraminidase treatment the invasion decreased dramatically compared to that of 7G8 ( Figure 8C ) . These experiments suggest that both sialic acid-independent and dependent strains are able to interact with CR1 , but for the latter CR1-mediated invasion may rely on the presence of intact sialic acid on glycophorin . To further explore the interaction of sialic acid-dependent strains with CR1 , we compared the ability of anti-CR1 and sCR1 to inhibit the invasion of neuraminidase-treated red cells in sialic acid-dependent strains Dd2 and FVO and sialic acid-independent strains 7G8 and 3D7 ( Figure 9 ) . While anti-CR1 and sCR1 inhibited 75–90% of invasion in 3D7 and 7G8 , the magnitudes of inhibition in Dd2 and FVO were only 30–50% . Therefore , the invasion of neuraminidase-treated red cells by sialic acid-dependent strains , although minimal , is much less sensitive to inhibition by anti-CR1 and sCR1 .
We have shown in a logical and systematic manner that CR1 is a sialic acid-independent receptor for P . falciparum . This molecule was previously dismissed when the invasion of erythrocytes expressing very low levels of CR1 ( Helgeson phenotype ) by sialic acid-independent strain Dd2SNM ( similar to Dd2NM used here ) was not completely abolished after neuraminidase treatment [27] . These results may be explained by the efficiency of very low levels of CR1 in promoting sialic acid-independent invasion and/or by the presence of an alternative receptor used by Dd2NM ( Figure 2 ) . Using the prototypical sialic acid-independent strain 7G8 , we demonstrated that sCR1 as well as polyclonal and monoclonal antibodies directed against CR1 drastically reduced or eliminated sialic acid-independent invasion while control proteins and antibodies did not . Similar results were obtained with the sialic acid-independent laboratory strains 3D7 , HB3 , and Dd2NM . However , for HB3 and Dd2NM CR1 did not appear to be the only sialic acid-independent receptor since these strains showed some residual invasion of neuraminidase-treated red cells in the presence anti-CR1 antibodies and sCR1 . In addition , we have shown that the sialic acid-independent invasion of three field isolates from Kenya can also be inhibited by these reagents . In two of the four laboratory strains tested here we were able to show significant inhibition of intact erythrocytes with anti-CR1 Fab and/or sCR1 ranging from 20–30% and similar results were obtained using wild isolates from Kenya . Although sCR1 is present in serum , its levels are much lower than those required for inhibition of merozoite invasion [28] . These findings suggest that for many malaria strains CR1 is an alternative receptor to glycophorins on intact red cells . The partial inhibition of invasion of intact red cells by anti-CR1 and sCR1 is not surprising given that there are in the order of 106 molecules of glycophorin A per erythrocyte [29] compared to an average of 600 of CR1 . The correlation between invasion and CR1 expression level further suggests that the contribution of CR1 to the overall invasion is determined by the number of CR1 molecules present on the red cell . In addition , invasion pathways may work in a hierarchical manner and less dominant pathways only become more prominent in the face of silencing of more dominant ones [30] . Since there is no complete absence of CR1 among humans , to show the effect of complete absence of CR1 on invasion we compared the invasion of wild-type mouse red cells , which do not express CR1 , and mouse transgenic red cells expressing human CR1 . Using this system we showed that expression of CR1 rendered red cells more susceptible to P . falciparum invasion . Interestingly , both sialic acid-independent and dependent strains showed increased invasion of CR1 transgenic mouse red cells but the latter seemed to require the intact sialic acid of glycophorin for optimal CR1-mediated invasion . This is also supported by the finding that invasion of neuraminidase-treated red cells by sialic acid-dependent strains is less sensitive to inhibition by anti-CR1 and sCR1 ( Figure 9 ) . One possible explanation for these observations is that the CR1 ligand in sialic acid-dependent strains may require interaction with both CR1 and glycophorin to mediate invasion optimally . The involvement of merozoite ligands in both sialic-dependent and independent invasion has been postulated [31] and our findings are consistent with this concept . The ability of merozoites to interact with CR1 was shown by immunofluorescence microscopy and without question by the demonstration that merozoites can bind preferentially to sCR1-coated polystyrene microspheres ( Figure 6 ) . That this interaction is mediated by a receptor-ligand interaction was demonstrated by the inhibition induced by chicken anti-CR1 . By contrast , the interaction of merozoites with BSA-coated microspheres was not inhibited by rabbit anti-BSA . Surprisingly , binding to glycophorin A-coated microspheres was relatively poor , suggesting that glycophorin may be less effective than CR1 in mediating merozoite attachment . Regardless , these results confirm that merozoites can interact directly with CR1 . P . falciparum has at its disposal an extensive array of ligands each of which is involved in one or more invasion pathways defined mostly by enzymatic treatment of erythrocytes . The ligands so far identified belong to two major families of proteins: the erythrocyte-binding like ( EBL ) family [8] , which contains EBA-175 [32] , and the recently identified reticulocyte-binding like ( RBL ) family of proteins: PfRh1 , PfRh2a , PfRh2b , PfRh3 , and PfRh4 [33] . Two studies have reported that the expression of PfRh4 correlates with sialic acid-independent invasion [34] , [35] . Antibodies against this protein blocked sialic acid-independent invasion in one study [36] but not in other [37] . In addition , parasite knockouts of PfRh2a and PfRh2b show decreased sialic acid-independent invasion suggesting that these molecules may also be involved in this pathway [31] . Additional work will be needed to determine whether CR1 serves as receptor for any of the PfRh ligands . There are several scenarios under which the CR1-mediated invasion pathway could become more prominent . One scenario may occur when the parasite encounters red cell or glycophorin variants that lack the receptors for the dominant pathway . This may explain why the CR1-mediated pathway plays such a major role in the invasion of CR1-transgenic mouse red cells . A second situation may involve the development of a natural immune response by the host against the dominant pathway which may drive switching to a less dominant pathway or selection of parasites that use alternative pathways . Already some evidence from the field suggests that this mechanism may actually be at play in endemic populations [38] . Thirdly , and most importantly , immunologic pressure induced by vaccination against the dominant ligands involved in the glycophorin-dependent pathway but not against the CR1-mediated pathway may lead to selection of strains that are more reliant on the latter [39] . Therefore , it is imperative that ligands from all the major invasion pathways be represented in a future malaria blood stage vaccine . The demonstration that CR1 is a sialic acid-independent receptor of P . falciparum will facilitate the identification of its ligand ( s ) and the development of a vaccine that effectively blocks red cell invasion .
Strains HB3 and 7G8 were obtained from the Malaria Research and Reference Reagent Resource Center ( ATCC , Manasas , VA , USA ) . Strain 3D7 Oxford was provided by the Walter Reed Army Institute of Research ( WRAIR ) through the generosity of David Haynes . Dd2NM was derived as described from a Dd2 stock at the WRAIR . Wild strains were obtained from malaria-infected children in western Kenya under approved protocols . Parasite cultures were maintained and synchronized in O+ blood with 10% heat inactivated serum using temperature cycling . Neuraminidase treatment was as described except that erythrocytes at 50% hematocrit in RPMI 1640 were incubated in 250 mU/ml of Vibrio cholera neuraminidase ( Sigma-Aldrich , St . Louis , MO , USA ) . For human red cells , the effectiveness of digestion was verified by loss of binding of mouse anti-human glycophorin A/B clone E3 ( Sigma-Aldrich ) used at a dilution of 1∶3375 with a secondary FITC-conjugated anti-mouse IgG ( Sigma-Aldrich ) at a dilution of 1∶50 . Polyclonal chicken anti-CR1 ( Accurate Chemical & Scientific Corp . , Westbury , NY , USA ) and purified chicken IgY ( Biomeda , Foster City , CA , USA ) were digested with the use of a commercial kit ( Thermo Fisher Sicentific , Rockford , IL , USA ) to obtain Fab fragments . The final stocks of Fab antibody were adjusted to a concentration of 80 µg/ml . The following IgG1 monoclonal antibodies directed against CR1 were also used: J3D3 ( Biomeda Corp . , Foster City , CA , USA ) , To5 and E11 ( Accurate ) , and J3B11 ( a generous gift of Dr . Jacques Cohen , Hospital Robert Debré , Reims , France ) . An IgG1 irrelevant monoclonal ( R&D Systems , Minneapolis , MN , USA ) and anti-CD55 monoclonal ( clone NaM16-4D3 , IgG1 ) ( Santa Cruz Biotechnology Inc . , Santa Cruz , CA , USA ) were used as additional negative controls for human erythrocytes . For experiments with mouse erythrocytes , we used as negative control an IgG2a rat monoclonal antibody directed against the complement receptor 1 related protein Y ( Crry ) ( Becton-Dickinson , San José , CA , USA ) found on the surface of mouse erythrocytes . Human blood was obtained in citrate phosphate dextrose ( Sigma-Aldrich ) and washed with RPMI to remove the buffy coat . 100 µl of packed red cells was incubated in 1 ml of RPMI 1640 with 500 µg/ml TPCK-treated trypsin ( Sigma-Aldrich ) for one hour at 37°C and washed ×3 with RPMI by centrifuging at 2 , 000 rpm for 5 min . Following the last wash , the pellet was resuspended in 1 ml of double deionised water containing protease inhibitor cocktail ( Sigma-Aldrich ) and allowed to sit in ice for 30 min . The ghosts were recovered by centrifugation at 10 , 000 rpm for 10 min and the pellet was washed twice more with deionised water . Finally , the pellet was solubilized with 200 µl of IP lysis buffer ( Thermo Fisher Scientific ) containing protease inhibitor cocktail ( Sigma-Aldrich ) and allowed to sit in ice for 30 min after which it was stored at −20°C until used . An additional 100 µl of packed red cells was similarly lysed and solubilized without prior enzyme treatment . Aproximately 100 µg of polyclonal chicken anti-CR1 or chicken IgY were covalently linked to each of two 100 µl volumes of AminoLink Plus Coupling Resin ( 50% slurry ) following the manufacturer's instructions ( Thermo Fisher Scientific ) . 500 µl of IP lysis buffer containing 5 µl of lysate from trypsin-treated or untreated red cells in the presence of protease inhibitor cocktail was added to each chicken anti-CR1 and IgY-linked resin pellets followed by incubation at room temperature for two hours with constant mixing and overnight at 4°C . The following day , the resins were washed exhaustively with IP lysis buffer and the trapped proteins were eluted with 50 µl low pH elution buffer followed by neutralization with 5 µl of 2 M Tris . 10 µl of each eluate was loaded onto separate lanes of a 4–12% Nupage-Novex Bis-Tris denaturing gel ( Invitrogen Corp . , Carlsbad , CA , USA ) followed by silver staining ( Thermo Fisher Scientific ) . Schizont-infected erythrocytes were separated from uninfected erythrocytes by magnet-activated cell sorting ( MACS; Miltenyi BioTec , Bergisch Gladbach , Germany ) using “LS” columns . The magnetized columns were equilibrated with 3 ml of RPMI 1640+0 . 2% sodium bicarbonate ( elution buffer ) . Typically , a 24 ml culture was processed in three 8 ml aliquots . After loading each 8 ml of culture onto the magnetized column , it was washed and the infected erythrocytes eluted after demagnetization with 3 ml of elution buffer each time . The total infected erythrocytes eluted were pelleted and resuspended in complete medium ( RPMI 1640 , 10% O+ plasma , 0 . 2% NaHCO3 , 25 mM Hepes ) . O+ human erythrocytes , wild-type C57BL/6 mouse erythrocytes , or CR1 transgenic mouse erythrocytes in C57BL/6 background , were plated in duplicate wells of a 96-well plate at a hematocrit of 2–4% in complete medium and were inoculated with schizont-infected erythrocytes to achieve a parasitemia of 0 . 3 to 5% for human red cells and 4–7% for mouse red cells . Plasma inactivated by heating at 56°C for 45 min was used in most experiments although we have not observed a difference in invasion or invasion inhibition between heat-inactivated and non-inactivated plasma ( data not shown ) . For microscopic readout human red cell invasion , only experiments in which the control untreated erythrocytes gave 2% or higher parasitemia were included in the analysis . Because much lower invasion rates were seen with mouse red cells , all levels of parasitemias were accepted . Blocking antibodies were added to uninfected erythrocytes prior to the addition of schizont-infected erythrocytes . For blocking with soluble CR1 ( sCR1 ) ( AVANT Immunotherapeutics , Inc . , Needham , MA , USA ) , bovine serum albumin , fetuin , and α-2-macroglobulin ( Sigma-Aldrich ) were used as negative control proteins depending on their availability . After a 16–22-hour incubation the ring stage parasites in 1 , 000 erythrocytes were counted in thin smears stained with Giemsa . The microscopist was always blinded to the experimental group designation of each slide . In some experiments , parasitemia was measured by flow cytometry ( see below ) . When comparing invasion across different donors , the parasitemia was normalized to the parasitemia of a single erythrocyte donor used as control in each assay using the formula , where “CorrParS” is the corrected parasitemia of the sample , “ParS” is the uncorrected parasitemia of the sample , “ParCMean” is the mean parasitemia of all the control samples , and “ParC” is the uncorrected parasitemia of the control for that sample . For detection of CR1 , a 4% hematocrit cell suspension was incubated with chicken anti-human CR1 Fab at 8 µg/ml for 30 min at RT and washed three times with RPMI 1640 . The cell sample was incubated with a 1∶50 dilution of FITC-conjugated goat anti-chicken IgG ( Sigma-Aldrich ) for 30 min at RT and washed three times with RPMI and stored in 1% paraformaldehyde at 4°C until acquisition . During acquisition , erythrocytes were gated on the basis of their forward and side scatter characteristics and the median fluorescence intensity ( MFI ) was measured using logarithmic amplification . To correct for day-to-day variation when comparing the CR1 MFI among a series of donors , the MFI was normalized to the mean of a control sample that was included in all the assays using the formula , where “CorrMFIs” and “MFIs” are the corrected and uncorrected sample MFI respectively , “MFIcmean” is the mean of all the MFI values of the standard control , and “MFIc” is the mean of the control obtained in parallel with the sample . For detection of parasitized erythrocytes , cells were incubated in 1 µg/ml Hoechst 33342 ( Invitrogen ) prior to fixation . The background staining of an uninfected red cell sample was always subtracted . Mouse red cells were differentiated from human red cells by use of PE-Cy5-labeled rat anti-mouse glycophorin ( clone TER-119 , Becton-Dickinson , San José , CA , USA ) at a dilution of 1∶100 . At least 10 , 000 erythrocytes were acquired for each sample . Acquisition was done using a LSRII flow cytometer ( Becton-Dickinson ) equipped with a UV laser and analysis was done using Winlist v5 . 0 ( Verity Software , Topsham , ME , USA ) . For determination of % parasitemia ( %P ) of mouse red cells using flow cytometry we used the following formulawhere %PEinf = %Hoechst positive erythrocytes in malaria culture , %NEinf = %Hoechst negative erythrocytes in malaria culture , %PEui = % Hoechst positive erythrocytes in uninfected culture , and %NEui = % Negative erythrocytes in uninfected culture . We always observed good correlation between flow cytometry results and microscopy . Proteins were linked to 6 µm carboxylated polystyrene microspheres using the PolyLink coupling kit ( Polyciences Inc . , Warrington , PA , USA ) . 10 µg of each protein was incubated overnight at room temperature with 50 µl of a 2 . 6% suspension of activated microspheres . The following day , the microspheres were washed thrice with wash buffer by centrifugation at 1 , 000×g for 5 min . The microspheres were then resuspended in PBS containing 100 µg/ml of BSA and incubated at room temperature for 1 hr to block any remaining binding sites . Finally , the microspheres were washed thrice again in PBS and stored at 4°C in 500 µl of PBS until used . Coating of microspheres with proteins was confirmed by flow cytometry using polyclonal antibodies against each protein ( data not shown ) . 7G8 late trophozoites and schizonts were purified by Percoll gradient centrifugation [40] . Approximately 1×106 infected red cells ( >80% parasitemia ) were incubated with 8×105 microspheres in duplicate wells of a 96-well tissue culture plate containing 120 µl of complete medium and 25 µg/ml gentamicin . Chicken polyclonal anti-CR1 ( 10 µl of 1 µg/ml in PBS ) or rabbit anti-BSA ( Sigma-Aldrich ) ( 5 µl of 20 µg/ml in PBS ) was added to separate duplicate wells . The plate was then incubated overnight at 37°C in a sealed gas impermeable bag containing 5% CO2 , 5% O2 , and 90% N2 . The following day , Giemsa-stained smears of each well were prepared to confirm merozoite release and attachment . An equal volume of PBS 2% paraformaldehyde containing 4 µg/ml Hoechst 33342 ( Invitrogen Corp . ) was added to each well and the plate was then stored at 4°C until acquisition took place . Acquisition was carried out on an LSR II ( Becton-Dickinson ) using the violet and 488 laser lines ( Pennsylvania State University College of Medicine Flow Core ( www . hmc . psu . edu/core/flow/overview . htm ) . The microspheres were gated based on their forward and side scatter characteristics using logarithmic amplification . The % microspheres that showed Hoechst staining was determined . Following overnight culture of malaria-infected red cells with sCR1 or BSA-coated microspheres , a 50 µl aliquot of each culture was centrifuged at 14 , 000 rpm for a few seconds and resuspended in 1∶100 chicken anti-CR1 in complete culture medium and incubated at room temperature for two hours . The microspheres were then washed twice with PBS and resuspended in 1∶100 DyLight 488-labeled goat anti-chicken IgY ( KPL , Inc . , Gaithersburg , MD , USA ) in PBS and incubated for 2 hours at room temperature . After one wash in culture medium , the pellet was resuspended in 2% paraformaldehyde with 4 µg/ml of Hoechst 33342 ( Invitrogen ) and incubated overnight at 4°C . The following day , 5 µl of microsphere suspension was spotted onto a slide , dried , and fixed with methanol . The microspheres were observed under a Leica TCS SP2 AOBS confocal microscope using the 405 and 488 laser lines ( www . hmc . psu . edu/core/microscopy/confocal . htm ) . Schizont-infected erythrocytes were incubated with 10 µg/ml of leupeptin ( Sigma-Aldrich ) for 8–12 hrs in complete medium at 37 °C , and then washed three times by centrifugation ( 400×g/4 min ) and resuspension in RPMI 1640 ( Sigma-Aldrich ) containing 25 mM Hepes and 0 . 2% NaHCO3 . Fresh uninfected erythrocytes , normal or neuraminidase-treated , were then added to the parasitized erythrocytes to a final hematocrit of 2% and parasitemia of 2% in complete medium and allowed to incubate for 30–60 minutes until at least 50% of the parasite population were membrane-enclosed merozoites . At this point , they were diluted to 0 . 2% hematocrit and added to 14 mm collagen-coated No . 1 coverslips ( MatTek Corporation , Ashland , MA , USA ) . After incubating for 30 min at 37 °C to allow for attachment of the erythrocytes to the collagen , the cells were fixed and prepared for immunofluorescence using methods described by Tokumasu and Dvorak [41] . Briefly , erythrocytes were cross-linked with 50 mM dimethyl suberimidate dihydrochloride ( DMS ) ( Thermo Fisher Scientific ) in a buffer containing 100 mM sodium borate buffer ( pH 9 . 5 ) and 1 mM MgCl2 for 1 hr at RT , followed by fixation with 1% paraformaldehyde in PBS for 1 hr at RT . DMS and paraformaldehyde were quenched by incubating with 0 . 1 M glycine in PBS , pH 7 . 4 , for 1 hr at RT . After washing in PBS for 5 min , the erythrocytes were blocked with 3% BSA in PBS with 0 . 2% Tween 20 ( blocking buffer ) for 30 min at RT . Chicken polyclonal anti-CR1 Fab ( 8 µg/ml ) , and rabbit polyclonal anti-glycophorin A IgG ( Accurate ) ( diluted 1∶10 for neuraminidase-treated cells and 1∶50 for normal cells ) were added in blocking buffer for 1 hr at RT . Following three washes with PBS with 0 . 2% Tween 20 , the secondary antibodies goat anti-chicken IgG-Alexa 488 , and goat anti-rabbit IgG-Alexa 555 ( Santa Cruz ) were diluted 1∶50 in blocking buffer and incubated for 30 min at RT . Nuclei were stained with Hoechst 33342 ( 2 µg/ml in PBS ) ( Invitrogen ) for 15 min at RT . After a final wash in PBS with 0 . 2% Tween 20 , the coverslips were dried , mounted with Antifade ( Biomeda ) , and sealed to standard microscope slides . The fluorescence z-series were collected on a LEICA DM RXA fluorescence microscope at 100× under oil immersion . Prior to reconstruction , the images were unsharp masked to remove the background and blurry parts of the images and , subsequently , contrast enhanced using TIFFany3 image processing software ( Caffeinesoft , Inc . , www . caffeinesoft . com ) . This method provides highly satisfactory results for small sharp features in the images as verified by comparison to the original data sets . The 3D image reconstructions from 40-plane image stacks were performed with BICViewer ( Bioinstrumentation Center , USUHS , Bethesda , Maryland , USA ) . Blood from human CR1 transgenic and wild-type mice in C57BL/6 background was obtained by cardiac puncture following deep anesthesia with inhaled isoflurane . Cardiac blood was anticoagulated with citrate phosphate dextrose , shipped overnight from the University of Massachusetts in Worcester , Massachusetts , to the Walter Reed Army Institute of Research in Silver Spring , Maryland , and used immediately upon arrival . Collection of human blood samples for this study was conducted according to the principles expressed in the Declaration of Helsinki and under protocols approved by the Human Use Research Committee of the Walter Reed Army Institute of Research and/or the National Ethics Review Committee of the Kenya Medical Research Institute . All patients provided written informed consent for the collection of samples and subsequent analysis . All animals were handled in strict accordance with good animal practice as defined by the National Institutes of Health and the Association for the Assessment and Accreditation of Laboratory Animal Care International ( AAALAC ) under a protocol approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee . Analysis was done using SPSS v11 . 5 ( SPSS Inc . , Chicago , IL , USA ) . The general linear model , an analysis of variance procedure , was used to test for equality of means across several experimental groups taking into account matching across groups by date of assay . If the overall F test was significant , Dunnett's pairwise multiple comparison t-test was used to compare each experimental group mean to the control mean . Spearman rank correlation was used to study the relationship between CR1 level and invasion . The paired samples t-test or the non-parametric Wilcoxon signed rank test , for small samples , was used for the comparison of two paired samples . All tests were 2-sided with α≤0 . 05 .
|
Plasmodium falciparum malaria is a blood parasite that lives for the most part inside red cells . It is responsible for the death of 1-2 million people every year . The mechanisms by which the parasite invades red cells are complex and not completely understood . For many years it has been known that proteins called glycophorins are used by the parasite to gain entry into the red cell . However , the existence of another protein that allows entry independent of glycophorins has been suspected for nearly as long . The identity of the alternative protein has been a mystery difficult to solve . In this article we present strong evidence that the alternative protein is the complement receptor 1 . The complement receptor 1 is a well-studied protein that is known to be important in protecting red cells from attack by the host immune system as well as suspected of having other roles in the development of malaria complications . The recognition of the additional role of complement receptor 1 in red cell invasion will allow the definitive identification of malaria proteins that interact with it and that could be used in a future vaccine cocktail to block red cell invasion .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunology",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] |
2010
|
Complement Receptor 1 Is a Sialic Acid-Independent Erythrocyte Receptor of Plasmodium falciparum
|
Epistatic genetic interactions are key for understanding the genetic contribution to complex traits . Epistasis is always defined with respect to some trait such as growth rate or fitness . Whereas most existing epistasis screens explicitly test for a trait , it is also possible to implicitly test for fitness traits by searching for the over- or under-representation of allele pairs in a given population . Such analysis of imbalanced allele pair frequencies of distant loci has not been exploited yet on a genome-wide scale , mostly due to statistical difficulties such as the multiple testing problem . We propose a new approach called Imbalanced Allele Pair frequencies ( ImAP ) for inferring epistatic interactions that is exclusively based on DNA sequence information . Our approach is based on genome-wide SNP data sampled from a population with known family structure . We make use of genotype information of parent-child trios and inspect 3×3 contingency tables for detecting pairs of alleles from different genomic positions that are over- or under-represented in the population . We also developed a simulation setup which mimics the pedigree structure by simultaneously assuming independence of the markers . When applied to mouse SNP data , our method detected 168 imbalanced allele pairs , which is substantially more than in simulations assuming no interactions . We could validate a significant number of the interactions with external data , and we found that interacting loci are enriched for genes involved in developmental processes .
The simultaneous perturbation of two epistatically interacting genes leads to a phenotype that is not expected based on the phenotypes of the individual genes . Understanding these phenomena is indispensable for explaining multi-factorial traits and diseases [1] . In addition , epistatic interactions provide important insights into the functional organization of molecular pathways [2] , [3] . Much effort has therefore been put into the development of methods to discover epistatic interactions , mostly in linkage and association studies [1] , [4]–[10] . Epistasis is always defined with respect to a specific phenotype and describes a non-additive interaction effect of two genes on that phenotype . Most gene interaction studies explicitly measure a phenotype such as growth rate or viability [11]–[14] . However , one can also study implicit phenotypes by searching for the over- or under-representation of certain allele pairs in a given population . Such allele pairs are examples of Dobzhansky-Müller incompatibilities: they establish a fitness bias in favor of individuals inheriting the over-represented allele combination [15] . In their most extreme form such incompatibilities are embryonic lethal . Genes harboring these alleles are clearly in epistasis , as none of the alleles alone has a fitness effect . Only the presence of specific allele pairs in one individual exposes the phenotype . In this context , an implicit phenotype is a trait that is not explicitly measured in the sample but whose regulators can still be inferred from the genotype data . Whereas several such incompatibilities are known in plants ( see [16] and references therein ) , only very few allele incompatibilities have been reported in mammals [17] , [18] . A small number of recent studies have explored this idea for the genome-level identification of epistatic interactions: if a large number of individuals is genotyped at a large number of genomic positions , it becomes possible to test all allele pairs for over- and under-representation in that population [18]–[20] . For example , [19] provide a map of distant linkage disequilibrium ( LD ) in mouse recombinant inbred lines ( RIL ) giving some indication about the distribution of imbalanced allele pair frequencies in the genome . However , even though some methodological progress has been made [18] , previous studies could hardly identify a significant number of interactions . The main obstacle is the humongous number of statistical hypotheses tested when comparing all markers in a genome against all markers . When correcting for multiple hypothesis testing one is usually left with very few or even no significant allele pairs . Here , we propose to address this problem by exploiting the additional information gained from studying family trios . We show that by analyzing a sufficiently large number of individuals with known family structure it becomes possible to detect substantially more interactions than what is expected if all markers were independent . Our method , called “Imbalanced Allele Pair frequencies ( ImAP ) ” , relies on sequence data only , making it applicable to the many already available SNP studies without the need for additional phenotype measurements . ImAP is based on inspecting contingency tables that track the frequencies of all possible two-locus allele combinations in heterozygous individuals ( assuming a diploid genome ) . The test that we propose is similar to a test in that it compares the observed frequencies in this table to expected frequencies assuming independence . However , our version corrects the expected frequencies for confounding factors such as family structure or allelic drift [21] . In a population of heterozygous mice with known family structure genotyped at markers we identify LD block pairs with imbalanced alleles . Using simulations we can show that this number is significantly larger than expected under the null hypothesis even after correcting for multiple hypothesis testing . The significance of the top scoring interactions between the LD blocks could be independently confirmed using a large collection of RIL . The number of significant allele pair imbalances that we detected is surprisingly large and was not expected based on the published evidence . We have made the top interactions identified with ImAP available as Tables S3 and S4 .
The core step of ImAP consists of a -type test comparing the observed frequency of the joint occurrence of a certain diallelic genotype in one locus together with a certain genotype in a second locus with the frequency expected based on the genotypes of the parents under the null hypothesis ( i . e . assuming no epistasis , Figure 1 ) . The two loci are required to be distant enough from each other in order not to get false positive results due to local linkage . This results in a score quantifying the deviation of allele pair frequencies from their expected values that is already corrected for inherent population structure . Subsequently , the significance of the scores is assessed with a permutation approach using pseudo-controls that are derived from the genotypes that parents could have transmitted to their offspring . We apply this framework in two steps: First , we only analyze genomic blocks with high local LD using representative markers . In a second step we drill down to individual marker pairs . To further verify our results , we established a simulation procedure that mimics the mating structure of the pedigree under the assumption of independence . We applied ImAP to search for potential epistatic interactions using an outbred heterogeneous stock ( HS ) of mice that was established by crossing eight inbred lines [22] . We are using the genotype data of individuals that were genotyped at markers . Importantly , the pedigree of these individuals is almost completely known . The HS consists of families , some of which are large , while others are only nuclear families . These families were derived from mating pairs of mice from the original stock after more than generations of random mating . Genotypes were obtained with the Illumina BeadArray platform achieving call rates of , the genotyping accuracy was greater than [22] . After removing individuals with more than missing data , we were left with individuals . In addition , we excluded markers with more than missing values and/or a minor allele frequency ( MAF ) less than . Since we observed a rather poor quality of the genotypes on the X chromosome with relatively few markers passing the quality criteria , we discarded data from this chromosome altogether . The filtering resulted in markers used for the subsequent analysis . We did not have to discard any SNPs due to lack of Hardy-Weinberg equilibrium as is generally done in genome-wide association studies . Instead , ImAP corrects for the disequilibrium ( see Methods ) . In the first run of our analysis , out of markers had correction factors greater than or smaller than . There are several explanations for the deviation from Hardy-Weinberg equilibrium , for example natural selection , genetic drift or segregation distortion [21] , [23] . Even though it might not be possible to distinguish the source of disequilibrium , our correction can be applied anyway . When applying ImAP to the HS mouse data , we limited our analysis to markers residing on different chromosomes in order to exclude local LD [18] . An alternative approach would have been to determine local LD first and subsequently apply ImAP to regions outside local LD . As described in the Methods , we first applied ImAP to a reduced set of markers , one per LD block . Figure 2 shows the spatial distribution of the interactions at the level of LD blocks in a genome-wide map . As expected , most block pairs do not interact . At a p-value cutoff of we identify interactions between distinct loci ( i . e . LD blocks ) . This p-value corresponds to an FDR of ( Benjamini-Hochberg procedure [24] ) . Although we did not achieve very low FDR values , they were still markedly lower than in five simulated data sets . In two out of these the minimum FDR was above . Most of the loci only interact with one other locus , only loci participate in more than interactions ( Figure S5 ) . Not surprisingly , there are more significant interactions between large chromsomes with many measured markers than between small chromosomes ( Figure 3 ) . However , we also found markable differences in the relative number of interactions per chromosome . Especially chromosomes , and incorporate more loci carrying allelic incompatibilities than other chromosomes . To see whether the number of interactors per chromosome is different from what would be expected by chance , we simulated the interacting marker pairs times and compared the distribution of the number of interactors per chromosome to the observed values . At a nominal significance level , three chromosomes ( , , and ) differ from their expected values . At this significance level , we expect less than one chromosome to differ significantly by chance . Hence , there is significant variation of the number of interacting LD blocks between chromosomes . In order to rule out the possibility of false positive findings due to increased numbers of missing values or small MAF on some markers , we compared the distributions of missing values and MAF between block representatives from significant block pairs to those of non-significant pairs ( Figures S3 and S4 ) . There are no significant differences between the proportion of missing values ( Wilcoxon rank sum test , p-value ) . The MAF tends to be even higher in the significant blocks compared to the other blocks . Thus , our results are not biased by missing genotypes or differences in MAF . The histograms in Figure 4 compare the distribution of the p-values that we obtained by applying ImAP to the orignal block representative data with those resulting from five simulations . While the histograms of the simulated data sets resemble those of uniformly distributed p-values under the null hypothesis , the original data show a clear peak in the low p-value range . The simulated pedigrees contain significantly less interactions with low p-values than the real data ( one-sided Kolmogorov-Smirnov test ) . The p-value distribution of the observed genotypes is also significantly different from a uniform distribution ( one-sided Kolmogorov-Smirnov test , ) . This is not the case for all but one of the simulations ( ) . Taken together this confirms that there are more imbalances in allele pair frequencies than expected by chance . This difference between the real and simulated data can now be quantified to make suggestions about the number of true allelic incompatibilities in the HS mouse population . For example , at ( corresponding to an ) we find between and more significant block pairs in the original data compared to the simulations . As can be seen in the inset of Figure 2 , each chromosome pair exhibits only few such interacting pairs that are often surrounded by less significant markers due to local linkage . To further increase the resolution in these interesting regions , we performed fine mapping of all marker pairs in the significant block pairs . For the second step of the analysis we chose all LD blocks that were involved in at least one significant interaction . There might be one or more interacting markers within each LD block and the above analysis does not reveal which markers within a region are involved in the interactions . We repeated the calculation of the test statistics , null distribution and p-values with all markers in those blocks to find the SNP pairs with the highest signal in each significant block pair . This resulted in marker pairs with a ( Tables S3 and S4 ) , since each block pair could contain more than one significant marker pair . Note that the interpretation of the newly calculated p-values has to be done with care since a large number of the tested marker pairs is already assumed to be interacting ( they were chosen from interacting LD blocks ) and because markers inside LD blocks are highly correlated ( i . e . not independent ) . Therefore , it is difficult to correct for multiple hypothesis testing . However , we can still use the p-values to rank the interactions , i . e . to identify the most likely interacting marker inside each LD block . Only few allele incompatibilites in mouse have been reported so far [17] , [18] . We are not aware of any analysis that quantitatively examines the number of such interactions that can be expected in the whole genome . An overview of the distribution of allele imbalances in RIL is given in [19] . The authors inferred the correlation between locus pairs as a measure for distant LD . The strains used in this study are partly identical to the progenitors of the HS stock . Thus , it is reasonable to assume at least partial overlap of incompatible locus pairs between our study and the RIL data . We therefore investigated the distant LD of markers that were genotyped in the RIL as well as in the HS mice . We downloaded the genotype data for inbred mouse strains ( www . genenetwork . org ) and recalculated the as well as the MAF of the common markers . This allowed us to apply the same quality constraints ( ) to the RIL data as to the HS genotypes . Moreover , only marker pairs on different chromosomes were considered . After the filtering , markers constituting informative pairs were used for the analysis . Figure S6 compares the overall distribution of distant linkage disequilibrium in the RIL data with that of markers with high ImAP scores . There is a significant difference between the background distribution of of common marker pairs on different chromosomes and the of the top ImAP pairs ( one-sided Kolmogorov-Smirnov test , ) . Marker pairs with a significant ImAP score tend to be more in distant LD than other marker pairs . More specifically , out of the marker pairs have an absolute correlation above . Thus , a significant number of interactions obtained from the HS can independently be confirmed in a different set of mouse populations . We investigated if the genes mapping to loci that participate in high ranking interactions are enriched for relevant Gene Ontology ( GO ) categories [25] . ImAP detects interactions between markers , not genes . Thus , in order to perform such analysis we have to assign gene functions to markers . A conservative solution to this problem is to assign to a marker the functions of all genes encoded between the flanking markers and . If there actually exists a functional enrichment among genes causing allele incompatibilities this enrichment will be ‘diluted’ due to this procedure . However , since we do not know the causal genes a priory there is no other rigorous way of performing such GO enrichment . This strategy also prevents a bias in GO enrichment due to local gene clusters with similar annotation . We further restricted the enrichment analysis to interacting pairs whose table contained exactly one cell with a zero entry . This corresponds to locus pairs where one allele pair combination was not observed at all in the sample and can thus be assumed to be lethal . We reasoned that genes involved in such an interaction have functions related to organism development . The mapping of genes and their associated GO terms to these markers resulted in markers having at least one GO term assigned to them . Seventy three of these markers are involved in one of the significant interactions . The enrichment test was conducted using the topGO algorithm [26] . An advantage of topGO is that it corrects for multiple hypothesis testing , particularly taking into account the nested structure of the GO tree . Since the multiple hypothesis testing correction is inherent in the algorithm , the authors suggest to use the unadjusted p-values as a ranking criterion . We call all terms significant with a based on the “weighting” algorithm of topGO . The top ranking GO biological process terms for the original data as well as for an exemplary simulation are shown in the Supporting Material ( Tables S1 and S2 ) . We found more significant and more relevant GO terms in the original data compared to the simulation . As expected , many of the significant GO terms are related to developmental processes such as germ cell layer development and development of brain , lung and epithelium . A lot of interesting terms had p-values just above the threshold of ( e . g . stem cell maintenance ( ) , anterior/posterior axis specification ( ) or determination of left/right symmetry ( ) ) . This analysis shows that markers participating in interactions are enriched for relevant GO categories . One might also expect that pairs of interacting markers share similar functions . However , we did not observe that interacting markers share GO categories more often than expected by chance ( data not shown ) . Epistatic interactions affecting the viability of an organism often bridge parallel pathways [2] , [3] . The assumption underlying this between-pathway model is the existence of functional redundancy among pathways . A decrease in functionality of only one of two genes operating in two redundant pathways still allows for regulation of the downstream process through the second alternative pathway . However , if both genes are dysfunctional , both pathways will be disrupted , which may lead to a severe phenotype ( i . e . an epistatic interaction between the two genes ) . Therefore , two genes in the same pathway should share some of their interaction partners , namely those in a functionally similar pathway [27] . Thus , the interaction profiles of genes in the same pathway should be correlated ( Figure 5A ) . Here , we are interested in markers having a significant number of common interactors . In order to find such groups of markers with similar interaction profiles , we compared the marker interaction profiles from the ImAP analysis using the congruence score [28] . It is calculated as the negative transformed p-value of a hypergeometric test for the number of shared interaction partners . Thus , the score relates the number of interactions shared between two markers to the total number of interactions each single marker participates in [28] . Since here we are analyzing interaction profiles ( i . e . all interactions of a given marker rather than single interactions ) we chose a less stringent cutoff value for interacting block pairs ( ) . Even though using the more stringent cutoff of also yielded more correlated interaction pairs in the real data than in the simulations , choosing a higher cutoff increases the difference between real and simulated data . The fraction of block pairs with congruence scores is higher in the original data than in the five simulations ( Figure S7 ) . This difference between the proportions is significant in four out of five cases for a significant congruence score ( ) . Thus , interaction profiles are more consistent in the real data compared to our simulations . An important and nontrivial step in any genetic mapping study is to identify the causal genes encoded in the significant loci . Additional , independent genomic information has been widely used to prioritize genes in a genetic region of interest [29]–[31] . Here , we are using expression data for prioritizing candidate genes at interesting loci . It is likely that several of the allele incompatibilities are caused through functionally relevant changes of gene expression between the minor and major alleles at the two loci [32] . We used expression data from three tissues ( lung , liver , hippocampus ) measured in a subset of the HS mice ( , and individuals , respectively ) . For each marker we considered all genes encoded in the region defined by the flanking markers . We then filtered for genes showing significant expression differences between individuals carrying the major versus minor alleles . This analysis was performed independently for each marker using one-way ANOVA with the three possible genotypes as levels . Each genotype had to be observed in at least individuals . Among the top scoring ImAP pairs , we found , and pairs where each locus contained at least one differentially expressed gene ( ) in the hippocampus , liver and lung data sets , respectively . locus pairs were associated with the same differentially expressed genes in all three tissues . Among the marker pairs with a congruence score greater than there are , and locus pairs containing at least one differentially expressed gene in the hippocampus , lung and liver data , respectively . Figure 5B shows an example of such a marker pair . The putatively causal genes Fgf10 and Btrc showed differential expression ( ) in the hippocampus . The two genes are critically involved in the development of several tissues such as lung , mammary gland , tooth or telencephalon [33]–[37] . This is consistent with the GO terms we found to be enriched among the top scoring ImAP pairs ( Table S1 ) . Btrc is an inhibitor of Sonic Hedgehog ( Shh ) signaling , which is involved in the development of the lung and the telencephalon [38] . Both , Fgf10 and Shh signaling are involved in development of anatomical structures and are known to influence each other [39] . According to our gene expression analysis , the minor allele of Fgf10 leads to a reduced expression of this gene while individuals carrying the minor allele of Btrc show a higher expression than individuals with the major allele . Since Btrc is an inhibitor of Shh signaling , this implies that both minor alleles reduce Hedgehog signaling . The Btrc and Fgf10 loci share ImAP interactions . One of them involves a locus on chromosome containing , among others , the homeobox transcription factor Nkx2 . 1 , which is indispensable for lung and telencephalon development . Depending on the cell type and developmental stage Nkx2 . 1 either interacts with the Fgf10 and Shh pathway [38] , [40] or it independently acts in parallel [41] . Thus , the reduced activity of Hedgehog signaling in carriers of the minor Btrc or Fgf10 alleles may be rescued by a fully functional Nkx2 . 1 . The ImAP analysis suggests that the combination of the minor allele at the Nkx2 . 1 locus together with minor alleles at either the Btrc or Fgf10 locus leads to an embryonic lethal phenotype , presumably due the loss of the buffering effect of Nkx2 . 1 .
We present a new approach to infer epistatic interactions on a genome-wide scale in family data using sequence information only . The method scans all marker pairs in the genome for deviation from the expected allele pair frequencies resulting in a list of putative pairs featuring an allele incompatibility . Relying on sequence data only is an advantage compared to existing methods for the inference of gene-gene interactions , since the approach can readily be applied to existing SNP data . There is no need for resource- and cost-intensive phenotype measurements . Regression and methods have been proposed in the past for the identification of epistatic interactions [1] , [7] , [9] , [10] , [42] , [43] and the two approaches have been shown to be interconvertible [44] . We chose a -based approach , which makes the fewest assumptions about the underlying genetic model [45] . Which ever way the detection of allele incompatibilities is performed , the key notion is to implement means for accounting for the confounding factors and to remove single-marker effects ( e . g . leading to a deviation from Hardy-Weinberg equilibrium ) . Only after accounting for these confounding factors we got an appreciable number of significant allele incompatibilities . We identified substantially more interacting loci than expected by chance , which is first evidence that we detect true ‘signal’ . Further , we could show that interacting marker pairs are enriched for genes involved in developmental processes and a significant number of interactions could be validated using independent external data . Due to the very large number of pairs tested , finding a large number of interactions with low p-values even in the simulations is expected . However , at low p-values we observed significantly more interactions in the original data than in any of the simulations; e . g . at we found at least interactions more than in any of the simulations . Considering that so far virtually no allele incompatibilities between mouse strains were reported , this is a surprisingly large number . Suitable statistical tools for the detection of allele incompatibilities at a genomic scale did not exist so far . Hence , this study presents first evidence about the extent of allele incompatibilities in model populations such as the HS . Although the number of interactions we identified might not seem immense , it partly explains the difficulties faced when breeding recombinant inbred lines [19] . For example , during the generation of the Collaborative Cross , a multiparental recombinant inbred strain panel , of the initial lines were lost during the first three to five generations of inbreeding [46] . ImAP helps better understand these issues and it can reveal potential biases in the breeding process that might be introduced due to allele incompatibilities . Future work should also include haplotype information . Local haplotypes have been inferred for the HS population in terms of probability of inheritance from any of the eight founder strains [47] . I . e . haplotypes are expressed as vectors of probabilities . Consideration of these haplotypes would considerably increase the complexity of the analysis ( thereby also increasing the number of hypotheses tested ) , but it might further improve the accuracy . An epistatic interaction is always defined with respect to a specific phenotype . In this study the phenotype is implicit , hidden . Indeed , looking for allele pairs that are underrepresented in the HS population reveals the genotype of the non-existing individuals . Therefore , the hidden phenotypes should relate to any biological processes affecting the fertilisation , the development or the viability of an individual and thus prevent its appearance in the population . Interestingly , top scoring marker pairs are enriched for genes involved in these expected phenotypes . It is not immediately obvious how our findings translate to human populations [48] , [49] . Although we are working with outbred mice , they were derived from genetically distinct inbred strains . These founder strains differ at at least genomic positions ( SNPs and structural variations ) [50] . It is likely that many of the incompatibilities that we see in the HS developed in the inbred founder strains used for establishing the HS . Even though allele incompatibilities cannot evolve in mixing populations , also human races have been in isolation for more than generations [51]–[53] . Hence , it is possible that an appreciable number of incompatibilities exist in the human species . [54] have shown that incompatibilities in yeast can manifest already after relatively few ( approximately ) generations . Again , also that finding is not easily transfered to mammals , as the speed of such process will depend on several factors , including recombination- and mutation rates . As the number of family trios that is being fully sequenced increases [55] , [56] , we expect that our framework will be applicable to human populations within the next years to address these questions .
The calculation of the test statistic can be divided into several steps which are depicted in Panel A of Figure 1 . 1 . Let be the set of all individuals for which we have genotype information on the individuals themselves and their parents . This set might differ between markers due to missing values . Hence , for each marker only these trios are taken into account for which there are no missing values in the genotypes of both the parents and the offspring . 2 . For each individual in , calculate the probability to inherit each genotype based on the genotypes of the parents . This calculation is based on Mendelian laws . Let be the genotype indicator of a diploid individual on marker . , can take one of the three values ( AA ) , ( Aa ) , ( aa ) , where A is the major allele and a the minor allele on marker . is the corresponding expected genotype probability . The expected genotype of individual on marker is derived from the genotypes of the parents under the assumption of equal chances of inheriting each of the two possible alleles from each of the parents . The resulting probabilities for all possible parental genotype combinations are shown in Table 1 . 3 . Correct the expected genotypes for possible confounding factors such as segregation distortion . There might be a preference in the inheritance of a certain genotype on one marker in the population which is independent of interaction effects , e . g . if this genotype leads to increased fitness . In order to correct the expected frequencies for allele selection that is independent of other loci we multiply each individual's expected genotype by the ratio of the sample-wide observed and expected frequencies for the corresponding marker ( based on all samples ) : ( 1 ) Normalize the corrected expectation so that the probabilities for each marker sum up to : ( 2 ) This guarantees an adjustment of expected allele frequencies in cases where the observed frequency of a marker in the population deviates from the theoretically expected values . 4 . Next , the observed and expected number of times each combination of genotypes of two markers appears , can be inferred . Let be the observed frequency of the genotype combination on markers and , the corresponding expected frequency . They are obtained by summing over all individuals : ( 3 ) ( 4 ) Using the product of the marginal probabilities of each single marker genotype for calculating the probability of the genotype combination mimics the assumption of no epistatic effects under the null hypothesis . This step results in the tables in the boxes “observed genotype combination” and “expected genotype combination” in Figure 1 . 5 . Finally , a -like test statistic can be obtained by first calculating the squared difference of observed and expected frequencies for each genotype combination of two markers and divided by the corresponding expected frequency . The final score for a marker pair is the sum of these values over all nine possible genotype combinations , ( 5 ) The significance of the imbalances observed for each marker pair is assessed with a permutation approach based on pseudo-controls . This approach has already been adopted in related problems [57] . The general outline of the procedure is shown in Figure 1B . For each parent-child trio we infer the four genotypes that the child could have inherited at each marker assuming independence . These are then randomly combined to pseudo-offspring genomes in which each of the possible marker pair - genotype combinations could in principle appear . Calculations were done using the R package trio [58] . We use these pseudo genotypes to assess the significance of the test statistics of each marker pair by calculating an empirical marker-specific null distribution based on permutations . The permutation p-value is calculated as the fraction of pseudo-control test statistics exceeding the observed score . FDR is obtained with the Benjamini-Hochberg approach [24] . In an earlier version of our analysis pipeline we calculated the p-values based on the distribution . The degrees of freedom were obtained by using the actual number of genotypes present in the population for each marker pair , . The degrees of freedom are then calculated as . However , we found that the distribution of these parametric p-values differed conditional on the minor allele frequencies ( MAF ) of the markers , as shown in Figure S1 . The distribution based p-values tend to be too conservative when the MAF is small . The underlying cause is a shift in the distribution of the test statistics depending on the MAF ( Figure S2 ) . This phenomenon was greatly reduced when we changed to the permutation based p-value calculation as can be seen in Figure S1 . In order to speed up the calculations but still retain an acceptable resolution of loci with potentially interacting genes , we pursued the following strategy . In a first run of ImAP we split the data into blocks of high linkage disequilibrium ( LD ) . This is again done with the package trio , which provides an algorithm to estimate LD block borders in parent-offspring data . Afterwards , one representative marker is chosen randomly among all markers with a minimum number of missing values in each LD block and the test is applied to all possible combinations of these representatives on different chromosomes . The restriction to markers on different chromosomes is applied to rule out false positive results due to local linkage disequilibrium . Subsequently , we identify all block pairs which were assigned an FDR below and repeat the analysis using all markers from those blocks . In this way we restrict testing of individual marker pairs to genomic regions that are suggestive for interactions . Finally , we select the highest scoring marker pairs from each locus pair as the ‘interacting pairs’ . This two-step approach allows for an accurate mapping of epistatic interactions over the whole genome by simultaneously restricting the number of tests and the computing time to a more reasonable level . The pseudo-control data was used to compute p-values . In order to also correct for multiple hypothesis testing and for testing for any other possible biases in our data we simulated the mating process in the mouse population assuming independence of the markers but adhering to the original pedigree structure . The simulation starts with the first generation of mice for which we have genotype information ( F0 generation ) . Using fastPHASE [59] we infer the haplotypes of these individuals . fastPHASE is based on the notion that haplotypes cluster into locally restricted groups which can be described using a Hidden Markov model . As opposed to other methods , fastPHASE assumes that due to recombination events the group membership changes continuously across the chromosome and not only at the block borders . Obtaining the haplotypes of the F0 generation allows us to initialize the mating process . For each mother and father of an F1 individual we start with randomly choosing whether they pass on the maternal or the paternal allele of the first marker on a chromosome to the offspring . Then , using either general or sex-specific recombination rates ( Supplementary Material in [22] ) , we sample whether the second marker is inherited from the same chromosome or whether a recombination took place during meiosis . This procedure is continued until a complete chromosome is assembled that is passed on to the offspring . The whole process is repeated until all generations are simulated . Subsequently , we randomly add genotyping errors ( making sure we do not introduce any Mendelian errors ) as well as the same missing values as in the original data . Since the simulation only accounts for local linkage but not for any other influences on allele frequencies , these data should not contain any true gene-gene interactions . The proportion of false positive findings should be comparable to the original data due to the same error rates and missing values .
|
Elucidating non-additive ( epistatic ) interactions between genes is crucial for understanding the molecular mechanisms of complex diseases . Even though high-throughput , systematic testing of genetic interactions is possible in simple model organisms , such screens have so far not been successful in mammals . Here , we propose a computational screening method that only requires genotype information of family trios for predicting genetic interactions . We tested our framework on a set of more than 2 , 000 heterozygous mice and found 168 imbalanced allele pairs , which is substantially more than expected by chance . We confirmed many of these interactions using data from recombinant inbred lines . The number of significant allele pair imbalances that we detected is surprisingly large and was not expected based on the published evidence . Our framework sets the stage for similar work in human trios .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genome",
"analysis",
"tools",
"genomics",
"genetics",
"population",
"genetics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Systematic Detection of Epistatic Interactions Based on Allele Pair Frequencies
|
Staphylococcus aureus is a leading cause of endovascular infections . This bacterial pathogen uses a diverse array of surface adhesins to clump in blood and adhere to vessel walls , leading to endothelial damage , development of intravascular vegetations and secondary infectious foci , and overall disease progression . In this work , we describe a novel strategy used by S . aureus to control adhesion and clumping through activity of the ArlRS two-component regulatory system , and its downstream effector MgrA . Utilizing a combination of in vitro cellular assays , and single-cell atomic force microscopy , we demonstrated that inactivation of this ArlRS—MgrA cascade inhibits S . aureus adhesion to a vast array of relevant host molecules ( fibrinogen , fibronectin , von Willebrand factor , collagen ) , its clumping with fibrinogen , and its attachment to human endothelial cells and vascular structures . This impact on S . aureus adhesion was apparent in low shear environments , and in physiological levels of shear stress , as well as in vivo in mouse models . These effects were likely mediated by the de-repression of giant surface proteins Ebh , SraP , and SasG , caused by inactivation of the ArlRS—MgrA cascade . In our in vitro assays , these giant proteins collectively shielded the function of other surface adhesins and impaired their binding to cognate ligands . Finally , we demonstrated that the ArlRS—MgrA regulatory cascade is a druggable target through the identification of a small-molecule inhibitor of ArlRS signaling . Our findings suggest a novel approach for the pharmacological treatment and prevention of S . aureus endovascular infections through targeting the ArlRS—MgrA regulatory system .
Staphylococcus aureus , a Gram-positive bacterial pathogen , is a leading cause of infections worldwide . The most severe forms are bloodstream infections , such as sepsis , endocarditis , or thrombophlebitis , characterized by their high mortality and frequent disease sequelae [1 , 2] . The crucial site in endovascular infections is the endothelium—the inner lining of blood vessels , which orchestrates the host response to bloodstream pathogens . Endothelial dysfunction is central to most of the infection-caused pathology [3 , 4] . Therefore , adhesion to endothelium is an essential element of S . aureus virulence during bloodstream infections . Adhesion allows S . aureus to establish stable infection foci at the vessel walls , leading to the aforementioned dysregulation of endothelial functions , vascular leak , formation of intravascular vegetations , and metastatic spread to surrounding tissues [5–7] . S . aureus survival in the bloodstream is aided by its ability to form clumps of tightly packed cells , which protects bacteria in aggregates from immune attacks , increases antibiotic resistance , and allows a coordinated secretion of virulence factors [8] . Both S . aureus clumping and endothelial adhesion depend on interaction with host molecules . Clumping is caused mainly by linking of neighboring bacteria through binding both ends of fibrinogen molecule , an abundant blood dimeric protein . S . aureus binds fibrinogen with its surface adhesins , including the clumping factors ( ClfA , ClfB ) or , to a lesser extent , fibronectin binding proteins ( FnbpA , FnbpB ) [8] . Clumping is also likely a prerequisite for the induction of coagulation and formation of fibrin-coated S . aureus aggregates , which are essential for pathogenicity [9] . Adhesion of S . aureus to endothelium is a multifactorial process , and various host ligands in vessel walls may act as anchoring points . Von Willebrand factor ( vWF ) , present on activated endothelium , is bound by staphylococcal vWF-binding protein ( vWbp ) , or to a lesser extent by protein A [10–12] . In addition to vWbp and protein A , ClfA , FnbpA , FnbpB , and teichoic acid can also mediate the direct adhesion of S . aureus to endothelium , though the ligands and mechanisms of this adhesion are not entirely clear [11 , 13–19] . At the sites where endothelial layer is already damaged , underlying collagen of the basal membrane is exposed , to which S . aureus binds either directly with the collagen adhesin ( Cna ) , or indirectly with vWF bridging between collagen and bacteria [10 , 20] . At the sites of endothelial inflammation or damage , clots and micro-clots composed of fibrin ( ogen ) , fibronectin and platelets appear , providing additional binding sites for S . aureus adhesins [21] . Animal experiments with S . aureus mutants deficient in the ability to clump , coagulate , or adhere , demonstrate that these properties are essential for pathogenicity in sepsis and endocarditis [8 , 9 , 17 , 22–26] . The complicated interaction of individual S . aureus surface proteins with host ligands , and their extensive functional redundancy , however , pose a challenge for identification of the optimal treatment target . Therefore , targeting entire regulatory pathways instead of single adhesins might hold the best therapeutic promises . Despite the importance of clumping and adhesion , little is known about the regulation of these processes during endovascular infection , and contradictory observations have been reported regarding well-known global regulatory systems Agr and SarA [27–31] . Recently , a new regulatory cascade of ArlRS–MgrA has been identified [22 , 32] that consists of a two-component regulatory system ArlRS ( a transmembrane sensory kinase ArlS and the response regulator ArlR ) , which in turn drives the expression of MgrA , a cytoplasmic DNA-binding transcription regulator . MgrA acts as a final effector and controls the gene expression pattern [22] . Inactivation of this cascade through deletion of arlRS or mgrA blocks S . aureus clumping in plasma [22 , 23] . This phenotype was associated with upregulation of multiple surface proteins , including some well-known adhesins ( FnbB and protein A ) , but also a subset of unusual giant surface proteins Ebh , SasG and SraP ( predicted size ≈1 . 1 , ≈0 . 2 and ≈0 . 2 MDa , respectively ) [22] . Therefore , a question about the mechanism of the observed inhibition of clumping remains , leading us to hypothesize that this changed repertoire of surface proteins might also have a pronounced effect on S . aureus endothelial adhesion . In this work , we show that the ArlRS–MgrA signaling cascade is a major regulator of S . aureus clumping and adhesion in physiologically relevant conditions , both in in vitro models and in vivo . This effect is exerted not through a direct regulation of proteins involved in clumping and adhesion , but indirectly through control of a set of giant surface-bound proteins ( Ebh , SasG and SraP ) that shield other surface proteins from binding their ligands . This reveals a novel mechanism regulating S . aureus adhesion , and identifies the ArlRS–MgrA cascade as a potential treatment target in endovascular infections .
We have previously observed that deletion of ArlRS–MgrA regulatory cascade inhibits clumping of S . aureus in the presence of fibrinogen [22 , 23] . Considering the inhibitory impact on cell-to-cell interactions , we hypothesize that ΔarlRS and ΔmgrA mutants might also have defects in adhesion to surface-bound fibrinogen , and possibly also to other endovascular ligands . This was tested using S . aureus USA300 LAC strain , a prototypical community-associated MRSA strain representing the widespread USA300 clonal lineage [33] . When adhesion to a fibrinogen-coated surface was measured in vitro , the ΔarlRS and ΔmgrA mutant strains showed a pronounced adhesion deficit ( Fig 1A ) . This adhesion phenotype could be reversed by complementation through chromosomal integration of the missing elements of the signaling cascade at the phage 11 integration site . In case of the complemented ΔarlRS , the adhesion was even slightly improved in comparison to the wild-type ( WT ) baseline ( Fig 1A ) . The adhesion deficit was not limited to fibrinogen: deletion of ΔarlRS or ΔmgrA also strongly decreased binding to fibronectin and vWF , the additional ligands involved in binding to the vessel wall ( Fig 1B and 1C ) . The LAC strain lacks the cna gene encoding a collagen adhesin [33] , and is accordingly incapable of adhesion to collagen . As collagen is another important ligand in the vessel wall , we tested MW2 , a USA400-lineage community-acquired MRSA strain , which contains and expresses the cna gene [33] . The MW2 ΔarlRS and ΔmgrA deletions inhibited adhesion to collagen ( Fig 1D ) , as well as to fibrinogen , fibronectin and vWF ( S1A–S1C Fig ) , confirming the observations in USA300 strain . When various other S . aureus strains were screened for their adhesion to fibrinogen , the same pattern emerged . Introduction of an ΔmgrA mutation into MRSA clinical isolates #103 and #132 from the USA100 lineage ( a lineage common in hospital-acquired bloodstream infections ) , USA200 lineage strain MRSA252 , as well as in MSSA strains Newman , HG001 , and 502a , all caused a decrease in adhesion ( Fig 1E ) . Overall , our findings demonstrate that a deletion of the ArlRS–MgrA cascade across different S . aureus lineages has a pronounced inhibitory effect on adhesion to a diverse array of potential endovascular ligands . We previously observed that ΔarlRS and ΔmgrA mutants upregulate expression of a subset of very large surface proteins Ebh , SraP , and SasG , normally repressed by the ArlRS–MgrA cascade [22] . As USA300 strains have truncated and non-functional SasG [22] , the effects of SasG are best studied in strains such as USA400 MW2 . The notable upregulation of these giant surface proteins led us to hypothesize they were responsible for inhibiting adhesion in ΔarlRS and ΔmgrA strains . To test this question , we incorporated deletions of the ebh , sraP , and sasG genes into the ΔmgrA mutant background . Results confirmed that these three proteins all are jointly responsible for the observed adhesion defects ( Fig 2A–2D ) . The relative contribution of each protein differed slightly depending on the strain and adhesion ligand . Nevertheless , only a simultaneous deletion of ebh and sraP in LAC ( Fig 2A–2C ) , or additionally sasG in MW2 ( Fig 2D , S1D–S1F Fig ) , restored the normal adhesion to various ligands in the ΔmgrA mutants , phenocopying the WT parent strains . Deletion of these proteins in the WT background had no effect on adhesion , in line with their low expression in the strains with a functional MgrA [22] . This parallels the previously reported role of Ebh , SraP , and SasG in inhibition of clumping [22] . Therefore , upregulation of this set of surface proteins controlled by the ArlRS–MgrA cascade is the main cause for both the lack of clumping and the inhibited adhesion . We performed additional experiments to identify the mechanism of restored adhesion in the ΔmgrA mutant with Ebh , SraP , and SasG removed . Deletion of ClfA ( responsible for fibrinogen binding ) in LAC caused a similar decrease of fibrinogen adhesion both in the WT background and in the ΔmgrA Δebh sraP::Tn triple mutant ( S2A Fig ) . In case of MW2 , deletion of collagen adhesin Cna led to a similar decrease in WT background and in ΔmgrA mutant lacking Ebh , SraP , and SasG ( S2B Fig ) . These observations suggest that conventional surface adhesins are still present on surface of the ArlRS–MgrA cascade mutants , and that when the giant surface inhibitory proteins are removed , conventional adhesins function normally , as in the WT strains . The presence of conventional adhesins ClfA and Cna on cell surface was further confirmed by their detection in western blots of the cell wall fraction of the S . aureus cells , and in their direct detection on cell surfaces with immunofluorescence microscopy , when characteristic surface “doughnut” staining pattern was observed ( S2C–S2F Fig ) . Therefore , Ebh , SraP and SasG probably exert their anti-adhesive and anti-clumping effect through blocking or masking the activity of the neighboring conventional surface adhesins . Surface charge and hydrophobicity affect adhesive properties of microorganisms . However , only a modest increase in surface hydrophobicity was observed in LAC ΔarlRS and ΔmgrA mutants . This increase persisted in the ΔmgrA strain with deleted giant surface inhibitory proteins ( S3A Fig ) , which otherwise phenocopies the WT strain in respect to clumping and adhesion ( Fig 2A–2C ) , demonstrating that the changed hydrophobicity does not correlate with changes in clumping and adhesion . No significant changes in the relative negative charge on cell surface were noted ( S3B Fig ) . These findings support our hypothesis that the giant surface inhibitory proteins exert their effect by specifically interfering with the conventional adhesins , and not through non-specific changes to the overall cell surface hydrophobicity or charge . To further investigate the anti-adhesive phenotypes , the SasG was expressed in the LAC strain using a tetracycline-inducible promoter . As the amount of the expressed SasG increased ( following increased doses of added anhydrotetracycline to the strain carrying tet-inducible pRMC2-SasG ) , the level of clumping decreased ( Fig 3A ) . The same effect was observed for adhesion to fibrinogen , though the degree of inhibition was relatively small , demonstrating that inhibition of adhesion needs higher level of SasG expression than inhibition of clumping . When pALC2073-SasG , a tet-inducible relative of pRMC2 allowing for higher expression levels , was used , a clear dose-dependent inhibition of adhesion by SasG was demonstrated ( Fig 3B ) . Further , the addition of anhydrotetracycline had no effect on the strains carrying empty expression vectors ( S4A and S4B Fig ) , confirming that it was specifically the induced SasG , not the antibiotic , that changed the phenotype . Together , this showed that the amount of giant surface proteins correlates with the anti-adhesive and anti-clumping properties . Addition of a soluble , recombinant SasG to the assay , at an amount comparable to or even exceeding the one induced on the S . aureus surface ( 1–5 μg/ml ) , produced no observable effect ( Fig 3A and 3B ) . This demonstrated that for their inhibitory activity , the giant proteins have to be present directly on the cell surface , possibly interacting with the neighboring conventional adhesins . In addition to the dose- and localization-dependence , the observed inhibitory effect was also dependent on the size of the inhibitory proteins . A series of truncations were introduced in ebh in the chromosome of LAC ( in the ΔmgrA sraP::Tn background ) , preserving the Ebh secretion signal and transmembrane/cytoplasmic domains responsible for protein anchoring , resulting in a series of progressively shorter Ebh variants ( Fig 3C ) . All of the protein variants were correctly displayed on the cell surface as evidenced by the staining of whole cells with anti-Ebh antibody , which produced a characteristic “doughnut” pattern of surface staining , and by Ebh Western analyses ( S5A–S5C Fig ) . When properties of this mutant series were tested , there was a correlation of the phenotype with the size of the Ebh construct . As Ebh truncations reached smaller sizes , Ebh lost its capacity to inhibit clumping and aggregation ( Fig 3D and 3E ) , with the first effects visible after ≈1 . 1 MDa protein was truncated to ≈0 . 8 MDa , and a more pronounced effect after truncation to ≈0 . 5 MDa . The observed inhibition of clumping and adhesion caused by expression of giant surface proteins raised the possibility that these might also cause detachment of individual bacteria from already established clumps or infectious foci . Indeed , when expression of sasG was induced in an already clumped LAC , it led to a marked destruction of existing clumps ( Fig 4A ) . Similar , though less dramatic , effect occurred when expression of sasG was induced in LAC that was already adherent to fibrinogen . Expression of sasG caused a small degree of detachment , and prevented further accumulation of bacteria on fibrinogen-coated surface ( Fig 4B ) . To simulate conditions occurring during dissemination from an infected vegetation in vasculature , we measured the effect of sasG induction on dissemination from an infected plasma clot under shear . In this setting , the strain expressing SasG disseminated from inside the infected clot to the surrounding medium quicker and to a greater degree than the strain lacking sasG ( Fig 4C ) . No differences in unclumping , detachment , nor in dissemination from the clot between strains were observed in absence of induction ( S6A–S6C Fig ) , confirming that expression of sasG ( as opposed to the mere presence of the gene ) was necessary to cause these effects . Altogether , these experiments demonstrate that regulation of giant surface proteins by the ArlRS–MgrA cascade might potentially play a role not only in prevention of clumping and attachment , but also possibly participate in regulation of S . aureus dissemination . Conventional assays measure the “bulk” properties of the entire S . aureus population , but do not offer insight into behavior of individual cells or the molecular forces governing adhesion . To understand the molecular events behind the non-adherent phenotype in ΔarlRS and ΔmgrA mutants , the interaction forces of single cells were studied by atomic force microscopy ( AFM ) [34] . Individual bacterial cells were immobilized on AFM cantilevers and used to probe the fibrinogen-covered surfaces , while measuring adhesion forces ( Fig 5A ) . In case of the LAC WT strain , a high frequency of adhesion events was observed , with strong force peaks of 2 , 368 ± 946 pN ( n = 4 , 406 adhesive curves from a total of 14 cells ) being found in most of the obtained force curves ( Fig 5B and 5C ) . These very strong adhesion forces are in the range of those measured previously with the same assay for staphylococcal adhesins [35–37] . They are consistent with the high-affinity “dock , lock and latch” and “collagen hug” ligand binding mechanisms , which involve dynamic conformational changes of the adhesins resulting in highly stabilized adhesin-ligand complexes [38 , 39] . These binding forces were missing in the cells of ΔarlRS and ΔmgrA mutants , where both the frequency and the observed forces of adhesion were dramatically reduced , especially in the ΔmgrA strain ( Fig 5B and 5C ) . These results provide direct evidence that fully functional fibrinogen-binding adhesins are exposed on the surface of the S . aureus WT cells , and that their “dock , lock and latch” binding activity is strongly inhibited in the ΔarlRS and ΔmgrA mutant strains . To analyze the distribution of adhesins on the cell surface , bacteria were mapped with the AFM tips functionalized with the fibrinogen ( Fig 5D ) . At a short interaction time ( 100 ms ) , multiple receptors interacting with the fibrinogen were detected across the surface of LAC WT , but very few receptors were accessible for fibrinogen binding on the surface of ΔmgrA mutant cells ( Fig 5E ) , consistent with the whole cell data ( Fig 5B and 5C ) . Only when the contact time was extended ( 500 ms ) , binding of fibrinogen by surface receptors on the ΔmgrA was detected ( Fig 5E ) . This supports a model where adhesins are present in the mutant cells , but their interaction with ligands is decreased by shielding caused by the neighboring giant surface proteins . The large size of the putative inhibitory proteins appearing on the surface of the ΔarlRS and ΔmgrA mutants might indicate that they act as a kind of “molecular bumper” , non-specifically preventing interaction of bacteria with surrounding surfaces . To test if such nonspecific anti-adhesive mechanisms are present in these mutants , adhesive forces between individual S . aureus cells immobilized at the AFM tip and different abiotic surfaces ( hydrophobic / hydrophilic / charged—corresponding to CH3 , OH , and COO- substrate notation , respectively ) were recorded ( Fig 5F ) . Both the adhesion frequency and adhesion forces ( Fig 5G ) were predictably smaller than in the case of specific adhesion to fibrinogen . We saw no substantial differences in interaction with abiotic surfaces between the LAC WT cells and the ΔmgrA cells . The same was true when observing the approach curves ( Fig 5H ) , indicating that there were no major differences in long-range electrostatic and steric interactions between the strains . All together , these data suggest that ArlRS–MgrA regulated factors hinder ligand adhesion . According to our other findings ( e . g . Fig 2 ) , these factors are likely the giant surface proteins under control of this regulatory system . The most probable scenario is that the inhibitory proteins appearing on the surface of the ΔarlRS and ΔmgrA mutants are shielding the neighboring adhesins through crowding , perhaps forcing them into suboptimal conformation or preventing the three-dimensional conformational changes needed for ligand binding . Adhesion in vasculature during infection occurs in the presence of the shear force of bloodflow , with shear stress ranging from below 5 dyn/cm2 to around 50 dyn/cm2 , depending on the location [40] . To translate observations from the static microwell assays into a system mimicking real-life conditions , adhesion of S . aureus was studied in the flow chambers at a shear stress of 10 dyn/cm2 , comparable to that observed in the human body . Under these conditions , LAC readily adhered to fibrinogen , with a slightly weaker adhesion to fibronectin and collagen ( the latter for MW2 ) , and a yet weaker adhesion to vWF . Irrespective of the baseline level of the WT adhesion , the ΔarlRS and ΔmgrA mutants displayed decreased adhesion to all of the ligands ( Fig 6A–6D ) . This adhesion deficit was much more pronounced than the one in static assays . Deletion of the giant surface protein genes ebh , sraP ( and sasG , for the MW2 strain ) , which are de-repressed in the ΔarlRS and ΔmgrA mutants , restored the phenotypes to the WT level ( Fig 6A , 6B and 6D ) or markedly increased the adhesion ( Fig 6C ) , demonstrating that effect of the ArlRS-MgrA cascade on adhesion is mediated mainly through these giant surface proteins . Unlike most of the in vitro assays , where single proteins and coated surfaces are used for the adhesion , the fibrin-embedded infectious foci at the vessel wall ( as well as developing micro-thrombi at the sites of endovascular damage ) are composed of multiple protein types with intricate three-dimensional architecture . This meshwork of fibrin and other proteins could trap individual bacteria in a non-specific way , promoting subsequent specific adhesion . To test how ΔarlRS and ΔmgrA mutants would behave in such conditions , their adhesion to real human plasma clots under shear was tested . In line with other results , the adhesion of the ArlRS–MgrA cascade mutants to plasma clots was significantly reduced compared to the WT strains ( Fig 6E and 6F ) . Similar as the experiments with single purified ligands , the decreased adhesion to plasma clot in the ΔarlRS and ΔmgrA mutants could be restored to WT levels by introducing additional deletions of the giant surface proteins Ebh , SraP , and SasG ( Fig 6E and 6F ) . This showed that also in the complex environment of a three-dimensional clot , these giant proteins are the main effectors of the observed lack of adhesion in the ArlRS-MgrA cascade mutants . Many of the ligands used by S . aureus for attachment during endovascular infection would be present either in the vicinity , or directly on the endothelial cells lining the vessel . Therefore , we measured the ability of mutant strains to adhere to human endothelial cell monolayers under physiological shear of 10 dyn/cm2 . Under these conditions , LAC bound mainly to vWF multimers displayed on the endothelial surface , and to a smaller extent directly to the endothelial cells ( S7 Fig ) . Under these conditions , the differences between WT and mutants were clearly visible . Irrespective of the strain background , ΔarlRS and ΔmgrA mutants attached significantly less than the respective LAC and MW2 WT parents ( Fig 7A and 7B ) . This lack of adhesion was reversed by additional deletion of genes for the giant surface proteins ( Fig 7A and 7B ) . Thus , the ArlRS–MgrA cascade appears to be a master regulator of S . aureus adhesion to a wide variety of ligands and conditions that could occur in vasculature , and this effect appears to be mediated by the giant surface proteins Ebh , SraP , and SasG . To investigate the behavior of S . aureus inside vasculature in vivo , we used intravital microscopy tracking fluorescent bacteria in a damaged mouse carotid artery . Following a chemical injury to the endothelial layer , the GFP-expressing LAC circulating in the bloodstream rapidly adhered both directly to the damaged vessel wall , and to the thrombus developing in the lumen over the injured area ( Fig 8A ) . Strains with ΔarlRS and ΔmgrA mutations failed to adhere to the damaged murine vessel in vivo , consistent with our panel of in vitro assays . Mutant strains began to accumulate in the artery only after a prolonged time , potentially because the developing thrombus at that time occluded most the vessel’s lumen , allowing for unspecific trapping of circulating bacteria in regions of stagnant flow . Nevertheless , even at these later time points , the strains with an inactivated ArlRS–MgrA cascade failed to accumulate to the same extent as the WT ( Fig 8A ) . This demonstrates that hypotheses formed in our in vitro model systems translate to a real infectious event , and that ArlRS–MgrA is indeed crucial for regulating adhesion to vessel walls and overall behavior of S . aureus inside vasculature . To confirm these observations in another in vivo setting , we used intravital microscopy to observe attachment of GFP-expressing S . aureus to mesenteric endothelium activated with a Ca2+-ionophore . The LAC WT strain attached to and formed large clots obstructing entire lumen of the large mesenteric vessels ( Fig 8B ) . These large clots were observed to some degree after injection of the ΔarlRS mutant , but were absent after injection of the ΔmgrA mutant . The ΔmgrA didn’t attach at all , or attached to vessel walls without occluding the vessel’s lumen ( Fig 8B ) . Differences between the strains were even more pronounced in smaller vessels . Tiny clumps and individual bacteria of WT strain attached within small vessels , while neither ΔarlRS nor ΔmgrA mutants attached to these microvessels ( Fig 8B ) . Overall , the ArlRS–MgrA cascade played an important role in attachment of S . aureus to mouse vasculature in various in vivo conditions and sites . Currently , there are no known inhibitors of the ArlRS two-component system . We screened a 2320-compound small molecule library for potential new inhibitors with a previously described fluorescent reporter in which sGFP is expressed under the control of the ArlRS-dependent P2 promoter of mgrA [22] . This led to five potential hits ( biochanin A , protoporphyrin IX , haematoxylin , chlorophyllide , and quercetin ) . After secondary testing of these candidates through Ebh production dot-blot assays [22 , 23] , and a final confirmation with qPCR measuring mgrA expression , biochanin A was identified as the best candidate affecting the ArlRS–MgrA cascade at doses of 40 μM and 80 μM ( Fig 9A ) . When assayed for antimicrobial activity , biochanin A had an MIC >320 μM , approaching the solubility limit of the compound and well above the ArlRS–MgrA inhibitory doses . At sub-MIC levels , biochanin A had no impact on S . aureus growth at 40 μM , but caused a delay at 80 μM ( S8 Fig ) . Since biochanin A can reduce mgrA expression at 40 μM ( Fig 9A ) , we decided to further explore its potential as an ArlRS inhibitor . Biochanin A was able to inhibit the adhesion of S . aureus LAC to fibrinogen in a dose-dependent manner ( Fig 9B ) , and it also triggered a significant delay in plasma-mediated clumping ( Fig 9C ) . To further investigate biochanin’s inhibitory effect on ArlRS signaling , and to estimate how much of the observed effect is due to its impact on the ArlRS-MgrA cascade , we performed additional experiments . We attempted to counteract the anti-adhesive effect of biochanin A by expressing mgrA from a plasmid , independently from the ArlRS-MgrA regulatory cascade . Complementation with additional MgrA had no effect on adhesion to fibrinogen in absence of biochanin A , but it significantly relieved part of the inhibition caused by addition of biochanin A ( Fig 9D ) . A similar effect was observed with strains lacking giant surface proteins . Deletion of ebh and sraP had no effect on adhesion to fibrinogen in absence of biochanin A , but it led to a significantly better adhesion in the presence of this inhibitor ( Fig 9E ) . While biochanin A may have some off-target effects on S . aureus , its anti-adhesive impact is attributable to a large extent to the inhibition of the mgrA expression ( approximately 30–40% ) and the subsequent upregulation of the giant surface proteins . Altogether , the above results suggest that ArlRS is a druggable target , and that inhibition of ArlRS could be a feasible way to intervene in S . aureus endovascular infections .
S . aureus bloodstream infections remain a major challenge . In the US alone , each year over 100 , 000 patients are hospitalized with S . aureus sepsis , and about 5 , 000 with S . aureus endocarditis , with incidences rising [41–43] . The mortality of these S . aureus infections remains remarkably high at 15–50% [44] , with survivors suffering from debilitating disease sequelae [1] . As such , it is crucial to develop novel therapies for endovascular infections [43] , and this pressing need can be met only through understanding the mechanisms underlying S . aureus virulence . In this study , we demonstrated that S . aureus clumping in plasma and adhesion to vessel walls are controlled by a regulatory cascade composed of the ArlRS two-component system and the downstream MgrA regulator . Inactivation of this single cascade was sufficient to completely inhibit clumping and adhesion . Notably , deletion of ArlRS–MgrA components had the same phenotypic effect irrespective of the S . aureus strain background used . Previous observations already suggested involvement of ArlRS and MgrA in pathogenicity of sepsis and endocarditis [22 , 23 , 45–47] , and recently ArlRS and MgrA were identified to be parts of a single signaling cascade [22] . However , the mechanism behind involvement of this cascade in virulence remained unclear . Our study answers this question by linking the ArlRS–MgrA cascade to the tangible elements of the disease pathogenesis: clumping and adhesion . ArlRS–MgrA regulates both of these crucial infectious processes . We demonstrate the impact of ArlRS–MgrA both at the macroscale , and at the cellular and molecular level . Historically , bacterial adhesion was studied in static assays . Static conditions make detailed analysis of mechanisms feasible , but they differ from real-life conditions inside the host . Within the human vasculature , S . aureus attaches to vessel walls while withstanding the shearing force of the blood flow . This shear modulates conformation of bacterial surface proteins and alters their adhesive properties . Indeed , S . aureus adhesion to endothelium and endothelial ligands changes depending on the flow and shear stress [10 , 30 , 48–50] . In experiments where we mimicked conditions in the vasculature , the anti-adhesive effect of deletions in ArlRS–MgrA cascade become even more apparent than in static assays . Under physiological shear stress , inactivation of the ArlRS–MgrA cascade led to a complete lack of adhesion to typical endovascular ligands . The ΔarlRS and ΔmgrA mutants also showed defects in adhesion to plasma clots and endothelial layer under shear stress , representing real-life conditions where adhesion occurs in the context of complex mixtures of proteins or neighboring cells , not just single purified ligands . These behaviors of ΔarlRS and ΔmgrA mutants were replicated in two mouse models of endovascular adhesion , confirming the in vitro observations . These collective findings clearly demonstrate the ArlRS–MgrA signaling cascade as the major regulator of adhesion . Noteworthy , the ArlRS–MgrA cascade does not regulate adhesion and clumping through altered expression of the well-known S . aureus adhesins previously implicated in pathogenesis ( e . g . ClfA , ClfB , Coa , vWbp , FnbpA , FnbpB , protein A , or Cna ) [22] . Instead , this cascade regulates expression of a subset of giant surface proteins ( Ebh , SraP and SasG ) that inhibit clumping and binding to a vessel wall . Of these , only SasG was previously reported to interfere with S . aureus binding to fibronectin and fibrinogen [51] , but no such activity was previously known for Ebh and SraP . The most likely scenario explaining the phenotypes of defects in ArlRS–MgrA mutants is that Ebh , SraP , and SasG shield the neighboring adhesins through crowding , perhaps forcing them into suboptimal conformation or preventing the three-dimensional conformational changes needed for ligand binding . This is fully consistent with the notion that ligand binding by staphylococcal adhesins through the “dock , lock , and latch” or “collagen hug” mechanisms involves multiple structural changes [38 , 39] . Inhibitory proteins will prevent optimal fitting of the interacting molecules , thus compromising the formation of stabilized complexes . This would explain both the requirement for the large size and cell-surface localization of the inhibitory proteins , their lack of unspecific effects on interaction of S . aureus with ligand-free surfaces , and their mode of action through the specific inhibition of the neighboring surface adhesins . This model is also in agreement with our single-cell force spectroscopy data . Through such crowding , a simultaneous inhibition of binding to numerous ligands can be achieved , preventing clumping and adhesion irrespective of what adhesins are present on the surface . Remarkably , the ArlRS–MgrA cascade achieves such profound phenotypic effects through regulation of only three surface proteins . Overall , data collected in this study led us to construct a model linking ArlRS–MgrA signaling , clumping , and adhesion ( Fig 10 ) . When the ArlRS–MgrA cascade is active , giant surface proteins are repressed , and only conventional adhesins are present on bacterial surface . These bind fibrinogen and other common vascular ligands , leading to cells clumping and adhering to vessel walls ( Fig 10A ) . When the ArlRS–MgrA cascade is inactive , conventional adhesins are still present , but giant proteins Ebh , Srap and SasG are de-repressed , and their appearance on bacterial surface interferes with ligand binding by the adhesins . This leads to abrogation of clumping and adhesion ( Fig 10B ) . Unable to clump and adhere , S . aureus will be vulnerable to host attacks and will be gradually cleared from circulation . Lack of adhesion to vessel wall will prevent formation of localized infectious foci and endovascular vegetations , and will not allow for endothelial damage and dysfunction to occur , explaining recent observations of decreased killing caused by ΔarlRS mutant strain in cultured endothelial cells [52] . Thus , many conventional elements of disease pathogenesis in bloodstream infections will be derailed simultaneously . Identification of the ArlRS-MgrA cascade as one of the master regulators of adhesion and clumping makes it a desirable drug target , though question remains about the role of this cascade in the normal life cycle of S . aureus . In our assays the phenotype of the ΔmgrA mutants was frequently more pronounced than the one of the ΔarlRS mutants , which results from MgrA being the direct regulator of gene expression , and from low levels of constitutive expression of mgrA even in the absence of ArlRS [22] . Similarly , effects of the AlrRS-MgrA regulatory cascade during naturally occurring infection might also be more gradual than in the case of ΔmgrA mutants , depending on the level of the ArlRS activity in response to the infection environment . Various other systems , such as agr and sarA have been implicated in regulation of staphylococcal adhesins , and ability to both display and remove adhesins from cell surface in response to environmental conditions is important for S . aureus virulence [53] . The ability of ArlRS-MgrA regulation to change the surface properties of S . aureus without a need for individual regulation of each adhesin make it especially useful for adaptation to changing environment during infection and colonization . As the signal detected by the ArlS is still unknown , we are unable to confidently identify conditions in which S . aureus uses this regulation . We speculate that the ArlRS-MgrA cascade may be employed for regulation of detachment , or for swift adaptation of bacteria to a non-vascular environment . Our data suggested , that in addition to prevention of clumping and adhesion , the expression of Ebh , SraP and SasG may be used by S . aureus as means of unclumping and/or detachment from a surface . Considering the very strong force of ligand-adhesin bond in S . aureus , we envision that display of giant surface proteins probably does not cause immediate detachment on its own . Rather , as the clumped/adherent state is being continuously disrupted by cells divisions and/or mechanical forces of the environmental shear , the detached or newly divided cells expressing giant surface proteins will be unable to re-attach . These detached cells will continuously disseminate from the infectious foci , shifting the overall balance towards an unclumped/detached phenotype . ArlRS–MgrA may therefore be thought of as a regulatory switch between two distinct phenotypes: ( i ) an adhesive/clumping phenotype that is optimal for establishing infectious foci inside host’s vasculature ( MgrA levels are high ) ; and ( ii ) a detaching/spreading phenotype necessary for other lifestyles ( MgrA levels are low ) . Notably , this phenomenon of detachment/spread is possibly specific only for the bloodstream , and for sites in the body where “conventional” host matrix molecules ( fibrinogen , fibronectin , collagen ) are abundant . At other sites , Ebh , SraP and SasG may act not as “anti-adhesins” , but rather as “alternative adhesins” . SasG is known to bind nasal epithelial cells and induce biofilm formation [22 , 51 , 54 , 55] , and SraP binds to sialylated receptors on platelets and lung epithelium [56 , 57] . Binding partners are currently unknown for the N-terminal domain of Ebh , and although there are reports of an internal repeat binding fibronectin [58] , it seems likely that Ebh does have another host ligand that remains to be identified . Altogether , this leads us to hypothesize that turning off the ArlRS–MgrA cascade might not only allow S . aureus to move to another body site , but also to adapt to a non-endovascular , presumably skin or mucosal environment during colonization , where the higher levels of Ebh , SraP , and SasG could facilitate adherence to relevant local ligands , and where “conventional” adhesins would not be needed . As ArlRS is also involved in regulation of S . aureus metabolism [59 , 60] , it is possible that ArlRS–MgrA cascade takes part in a more general adaptation to these novel niches . The ongoing effort at identification of the signals turning the cascade on and off will probably allow to elaborate this model in the future . The important S . aureus virulence functions identified for the ArlRS–MgrA cascade suggest it has potential as a drug target . As shown herein , cascade inactivation renders S . aureus unable to clump in plasma or to adhere to vessel wall , leaving the pathogen exposed to host defenses and clearance from circulation . In this work , we identified biochanin A as a modest ArlRS inhibitor that can interfere with S . aureus adhesion and clumping , providing a proof-of-concept that ArlRS is a druggable treatment target . However , a more specific inhibitor would be necessary for a clinically meaningful outcome . These findings highlight the potential for future identification of more potent inhibitors and understanding the mechanism of their action on the ArlRS two-component system . There are reported putative MgrA inhibitors showing promising effects [8 , 61 , 62] , though their specificity for MgrA was not established . Our studies and these previous reports underscore that the ArlRS–MgrA cascade should be considered as the potential target for treatment and prevention of endovascular infections . In conclusion , our work identifies the ArlRS–MgrA cascade as the key regulator of S . aureus adhesion and clumping , both in vitro and in vivo in the bloodstream . This regulation is likely achieved through control of expression of the giant surface proteins Ebh , SraP , and SasG , which our data suggest interfere with conventional surface adhesins binding their host ligands , such as fibrinogen , fibronectin , vWF and collagen . This mechanism of adhesion control—not through directly regulating adhesins , but by inducing inhibitory proteins—adds a new level of sophistication to mechanisms by which S . aureus modulates its virulence . It is possible , that the ArlRS–MgrA cascade is one of the master switches controlling transition of S . aureus from one body site to another , or from colonization to infection . In order to improve our understanding of ArlRS–MgrA signaling , it will be important to determine the complete list of processes controlled by this cascade , identify the environmental cues sensed by ArlS , and uncover the ligands possibly bound by Ebh , SraP , and SasG . Finally , disruption of the ArlRS–MgrA cascade has potential as a therapeutic strategy , which could ameliorate the enormous burden currently imposed by S . aureus bloodstream infections .
The animal experiments were approved by the University of Iowa Institutional Animal Care and Use Committee , protocol number 6041727 , and by the University of Colorado Institutional Animal Care and Use Committee , protocol number 00439 . Both the University of Iowa and the University of Colorado are AAALAC accredited , and their centralized facilities meet and adhere to the standards in the “Guide for the Care and Use of Laboratory Animals . ” Human plasma used in clumping assays was obtained from the already existing anonymized plasma collection from healthy donors at the University of Iowa Inflammation Program , with all necessary institutional approvals . Plasma was diluted 1:1 with heparin/dextran sulfate to prevent clotting ( this mixture referred to as 100% plasma ) . Plasma for other assays was obtained from University of Iowa DeGowin Blood Center , as an anonymized apheresis plasma anticoagulated with the acid citrate dextrose solution . No regulatory approval was needed for using this blood product . Bacterial strains and plasmids used are listed in the Table 1 . Two USA100 clinical isolates from bloodstream infections in the University of Iowa Hospitals and Clinics were kindly donated by Dr . Daniel Diekema . For all experiments , S . aureus cultures were grown in a brain-heart infusion broth ( BHI ) at 37°C , with shaking to an OD600 = 1 . 5 , washed with phosphate buffered saline ( PBS ) , resuspended in the same volume of PBS , and directly used for assays or further adjusted to a desired concentration as needed . Escherichia coli cultures were grown in LB broth or agar . Antibiotics were added to media as needed: chloramphenicol ( Cam ) , 10 μg/ml; erythromycin ( Erm ) , 5 μg/ml; tetracycline ( Tet ) , 1 μg/ml; ampicillin ( Amp ) , 100 μg/ml , spectinomycin ( Spec ) , 50 μg/ml . Collagen and 6- and 96-well plates were purchased from Corning , bacteriological growth media and bovine serum albumin ( BSA ) from Research Products International , and vWF ( Humate-P ) from CSL Behring . Human Umbilical Vein Endothelial Cells ( HUVEC ) from pooled donors and cell culture reagents were from Lonza , and the cells were cultured in EGM-2 endothelial growth medium according to manufacturer’s instructions . Biochanin A was from MicroSource Discovery Systems . All other chemicals and proteins , unless noted otherwise , were from Sigma-Aldrich . Escherichia coli strains , restriction enzymes , DNAse I , T4 DNA ligase , shrimp alkaline phosphatase , Gibson Assembly Master Mix , and Phusion DNA polymerase were from New England Biolabs ( NEB ) . DNA and RNA isolation kits were from Invitrogen , Qiagen and Ambion . In case of S . aureus , lysostaphin treatment step was added to the DNA isolation protocol . Plasmids were electroporated into S . aureus RN4220 as described previously [75] . Bacteriophage transduction between S . aureus strains was performed with phage 80α or 11 as described previously [76] . Oligonucleotides were ordered from Integrated DNA Technologies; their sequences are listed in the S1 Table . DNA sequencing was performed by the Genomics Division , University of Iowa . S . aureus LAC expressing a secreted recombinant His6-tagged SasG ( rSasG ) from a pHC90 plasmid was inoculated 1:100 from an overnight culture into 500 ml of tryptic soy broth and grown for 16 h with 100 ng/ml of anhydrotetracycline as an inducer . The spent medium was harvested by centrifugation and filtration through a 0 . 2 μl filter , and concentrated with a 100 kDA cut-off Ultracell ultrafiltration disc ( Millipore ) to enrich for the rSasG ( ≈200 kDa ) . The protein was further purified on the Ni-charged Bio-Scale Mini Nuvia IMAC Cartridge ( Bio-Rad ) according to manufacturer’s protocol , dialyzed into PBS , and stored at -80°C . To assess the expression of SasG in S . aureus suspensions used for various assays , bacteria from 1 ml bacteria suspension were lysed by resuspending in 100 μl of a Tris-buffered saline with 0 . 5 U/ml of lysostaphin and 10 U/ml of DNAse I , and incubating for 1h at 37°C . Cell debris was collected by centrifugation and lysate supernatants were analyzed by SDS-PAGE , stained with silver or Coomassie stains . The rabbit anti-Ebh serum , and the rabbit polyclonal anti-ClfA antibodies ( kindly donated by Joan Geoghegan , Trinity College Dublin ) were described previously [23 , 80] . Rabbit anti-Cna serum was produced by Pacific Immunology by immunizing rabbits with a recombinant peptide consisting of amino acids 174–296 of the Cna molecule ( Cna174-296 ) . To clone and purify the Cna174-296 peptide , the corresponding sequence was generated by PCR with primers JK79 and JK80 using genomic DNA of S . aureus MW2 as template , digested with NheI and EcoRI , ligated into pTEV5 to generate pJK11 , from which it was expressed as His6-N-terminally tagged protein in E . coli ER2566 , and purified on HIS-Select nickel affinity resin ( Sigma ) . Measurement of S . aureus clumping in the presence of plasma was performed as described previously [22 , 23] . Human plasma was added to 1 . 5 ml of bacterial suspension in PBS to a final concentration of 2 . 5% v/v , and the mixture was vortexed and incubated at a room temperature for 2h . Clumping , resulting in the sedimentation of clumps , was measured by removing 100 μl aliquot from the top of the tube and measuring its OD600 . The % of clumping was calculated as the % decrease from the OD600 at time 0 . Cell culture 96-well plates were coated with human fibrinogen , fibronectin , vWF or type I rat collagen by filling with 20 μg/ml protein solution in PBS ( or 0 . 2% acetic acid for collagen ) and incubating overnight at 4°C . Afterwards , plates were washed with PBS , blocked by incubating with 5% BSA in PBS for 2h at 37°C , and washed again . Wells were filled with 100 μl of bacterial suspensions in PBS at OD600 = 1 . 0 , and incubated for 1h at 37°C . Afterwards , wells were washed and dried , the adherent bacteria were stained with 0 . 1% crystal violet , the bound stain was solubilized with 33% acetic acid , and the adhesion was measured as the OD570 of the resulting solution . Previously described models [10 , 81] were adapted to study S . aureus adhesion under flow conditions . Channels of μ-slides VI0 . 4 ( ibiTreat surface , Ibidi ) were coated by filling with the protein solutions ( as for the static adhesion assay , except that 50μg/ml protein concentrations were used ) and incubating overnight at 4°C . Afterwards , channels were washed with PBS , blocked by filling with 5% BSA in PBS for 2h at 37°C , and washed again . Channels were attached to a peristaltic pump ( model 205S , Watson-Marlow ) and perfused for 10 min with suspensions of GFP-expressing bacteria adjusted to OD600 = 0 . 65 in PBS , at a physiological shear stress of 10 dyn/cm2 ( shear rate of 1000 s-1 ) , calculated as described previously [82] . Afterwards , unbound bacteria were washed away by perfusing with PBS for 10 min , and the remaining attached bacteria were fixed with 2% phosphate-buffered formaldehyde for 15 min . Bacteria adhering to channels were imaged using BZ-X710 fluorescent microscope ( Keyence ) , with images taken of 5 random location in each channel , and subsequently processed using FIJI software [83]: auto thresholding ( MaxEnthropy method ) was applied to identify and quantify the image area occupied by the adherent bacteria . Adhesion to endothelial cells under flow was measured in a same way , with HUVECs ( at passages 4–6 ) cultivated in the channels until forming a monolayer , and activated by incubation with 0 . 1 mM calcium-ionophore A23187 in EBM-2 medium for 10 min at 37°C [10 , 81] . Subsequently , the channels were perfused at 37°C for 15 min with S . aureus suspensions of OD600 = 1 . 0 in DMEM , washed by 10 min perfusion with DMEM , and fixed . When needed , vWF on endothelial surface was visualized by immunostaining as described previously [84] , except that secondary antibodies were labelled with Alexa Fluor 568 . Human plasma clots were generated in 6-well cell culture plates by mixing 500 μl human plasma with 200 μl of PBS containing 6U/ml thrombin and 0 . 1 mM CaCl2 , incubating overnight at 4°C , and washing with PBS , resulting in a layer of clot covering the bottom of the well . Bacterial suspensions in PBS at 5×103 CFU/ml were added to the wells , 2 ml/well . Plates were incubated for 15 min at 37°C in a shaking microplate incubator ( Stuart ) at 500 rpm , inducing an estimated maximal shear of ≈20 dyn/cm2 , calculated as described previously [85] . Afterwards , wells were washed extensively with PBS , filled with a melted tryptic soy agar , and adherent CFUs were counted after an overnight incubation at 37°C . Immunofluorescence microscopy was used to visualize various proteins on S . aureus surface . Bacteria in PBS were allowed to attach to Superfrost Plus microscope slides ( Fisher ) , and afterwards were fixed for 15 min with 2% formaldehyde solution , blocked with 5% BSA in PBS for 1 h , stained with primary antibody at 4°C overnight and secondary Alexa488-conjugated goat anti-rabbit IgG antibody for 1h ( all antibodies diluted in 5% BSA in PBS ) , mounted with Fluoroshield Mounting Medium with DAPI ( Abcam ) , and observed with BZ-X710 fluorescent microscope . All LAC strains used for imaging lacked protein A ( Δspa or spa::φNΣ ) . In case of MW2 strains , 1% human serum was added to blocking solution and antibody solutions to saturate protein A and block unspecific binding . Western blot of proteins sheared from S . aureus surface with anti-Ebh serum was described previously [22] . For Western blot of cell-wall anchored proteins , the cell wall fraction was prepared as described previously [86] and stained with anti-ClfA or anti-Cna antibodies analogously as for the anti-Ebh staining . All LAC strains used lacked protein A ( Δspa , Δspa::tetM , or spa::φNΣ ) . In case of MW2 strains , 1% human serum was added to all solutions to block unspecific binding to protein A . The relative surface hydrophobicity was measured using the microbial adhesion to hydrocarbon ( hexadecane ) assay , as described previously [87] , except that addition of ammonium sulfate ( needed for the gram-negative bacteria in the original protocol ) was omitted . The relative surface charge was measured by electrostatic interaction chromatography on a Dowex 1×8 anion exchange resin , 100–200 mesh , as described previously [88] . Fibrinogen was immobilized on substrates via self-assembled monolayers of thiols . Practically , gold-coated glass coverslips were immersed overnight in an ethanol solution containing 1 mM of 10% 16-mercaptododecahexanoic acid/90% 1-mercapto-1-undecanol , rinsed with ethanol , and dried with nitrogen . They were then immersed for 30 min into a solution containing 10 mg×ml-1 N-hydroxysuccinimide ( NHS ) and 25 mg×ml-1 1-ethyl-3- ( 3-dimethylaminopropyl ) -carbodiimide ( EDC ) , rinsed with ultrapure water ( ELGA LabWater ) , incubated with 0 . 5 mg×ml-1 human fibrinogen for 1 h , rinsed further with PBS buffer , and immediately used without de-wetting . The abiotic substrates were obtained by immersing gold-coated glass coverslips overnight in an ethanol solution containing 1 mM of dodecanethiol ( Sigma ) for CH3 substrates , 1mM of 1-mercapto-1-undecanol ( Sigma ) for OH substrates or 1 mM of 16-mercaptododecahexanoic acid ( Sigma ) for COO- substrates . These surfaces were rinsed in ethanol and dried with nitrogen before to use . Bacterial cell probes were obtained as previously described [89 , 90] . Briefly , colloidal probes were obtained by attaching single silica microsphere ( 6 . 1 μm diameter , Bangs laboratories ) on triangular shaped tipless cantilevers ( NP-O10 , Microlevers , Bruker Corporation ) with UV-curable glue ( NOA 63 , Norland Edmund Optics ) using a Nanoscope VIII multimode AFM ( Bruker Corporation , Santa Barbara , USA ) . These colloidal probes were then incubated for 1 h in a 10 mM TBS ( pH 8 . 5 ) containing 4 mg×ml-1 of dopamine hydrochloride ( 99% ) . They were then rinsed in TBS and directly used for cell probe preparation . The nominal spring constant of the colloidal probe cantilever was ~0 . 06 N×m-1 as determined by the thermal noise method . 50 μl of diluted bacterial suspension in exponential phase was deposited into a petri dish containing fibrinogen-coated or abiotic surfaces at a distinct location , and filled with 3 ml of PBS . Single bacterium was attached on the center of the colloidal probes using a Bioscope Catalyst AFM ( Bruker Corporation , Santa Barbara , USA ) equipped with a Zeiss Z1 Axio Observer and a model C10600 Hamamatsu camera . Cell probes were used to measure cell–substrates interaction forces at room temperature ( 20°C ) , using an applied force of 250 pN , a constant approach-retraction speed of 1 μm×s-1 and a contact time of 100 ms . Data were analyzed using the Nanoscope software from Bruker ( Santa Barbara , USA ) . Adhesion forces were obtained by calculating the maximum adhesion force for each curve . The results from independent measurements were merged . For each condition , experiments were repeated for at least 3 times with independent cultures . Gold cantilevers ( OMCL-TR4 , Olympus Ltd . , Tokyo , Japan ) with a nominal spring constant of ~0 . 02 N×m-1 were functionalized with fibrinogen using thiols chemistry as previously described for substrates . The spring constants of the cantilevers were measured using the thermal noise method . Bacteria from exponential phase culture were immobilized by mechanical trapping into porous polycarbonate membranes ( Millipore , Billerica , USA ) with a pore diameter of 0 . 8 μm . After filtering a cell suspension , the membrane was rinsed with PBS , cut into piece ( 1 x 1 cm2 ) and attached to a steel sample puck using a small piece of double-face adhesive tape . The mounted sample was transferred into the AFM liquid cell while avoiding de-wetting . Bare tips were first used to localize and image individual cells and then replaced by functionalized tips . Adhesion maps were obtained by recording 32-by-32 force-distance curves on areas of 500 x 500 nm2 , using an applied force of 250 pN , a constant approach-retraction speed of 1 μm×s-1 and a contact time of 100 or 500 ms . Data were analyzed using the Nanoscope software from Bruker ( Santa Barbara , USA ) . Adhesion forces were calculated considering the last peak for each curve and adhesive events are displayed as light pixels . For each condition , experiments were repeated for at least 3 times with independent cultures . Previously described intravital microscopy method was used [91 , 92] . Female C57BL/6 mice , 8–11 week old , were anesthetized with katamine/xylazine , and GFP-expressing bacteria ( 2 . 2×108 CFU/mouse in 50 μl PBS ) were injected intravenously through the retro-orbital plexus . The common carotid artery was exposed and 1×2 mm piece of Whatman paper saturated with 10% FeCl3 was applied to it to induce endothelial injury and thrombosis . Afterwards the field was washed with PBS , and thrombus formation and S . aureus attachment were monitored with a Nikon upright microscope equipped with a high-speed electron-multiplying camera . Images were collected at determined time points , and the percent of the artery area occupied by the adherent fluorescent S . aureus was were calculated using FIJI image analysis software [83] . A previously described intravital microscopy method was used [10 , 81] . Female 7-week old C57BL/6 mice were anesthetized with a ketamine/xylazine . Their peritoneal cavity was opened via midline abdominal incision , and their mesenteric circulation was exposed . To activate mesenteric endothelium , 5 μl of a 10 mM Ca2+-ionophore A23187 was topically applied to the surface of the mesentery . Afterwards , GFP-expressing bacteria ( 1 . 2×108 CFU/mouse in 50 μl PBS ) were injected intravenously through the retro-orbital plexus . After 30 minutes , all visible adhesion of S . aureus to mesenteric vasculature was assessed and captured with a Nikon upright microscope with a digital camera . If no adhesion was observed , representative images of vessels were captured . A previously described reporter of ArlRS activity—that is S . aureus USA300 LAC strain carrying reporter plasmid pHC68 in which expression of sGFP is driven by the P2 promoter of mgrA ( P2 promoter is entirely dependent on arlRS activity for its expression , and it accounts for ≈90% of all mgrA transcripts in the cell ) , thus allowing for measurement of GFP fluorescence as the proxy for ArlRS activity in bacteria growing in a presence of tested compounds [22] was used to screen the “Spectrum Collection” library ( MicroSource Discovery Systems ) , consisting of 2320 small molecules , for potential ArlRS inhibitors . The reporter strain was grown in 96-well plates in TSB with erythromycin in the presence of 50 μM of each compound . Plates were grown at 37 °C with shaking in a Stuart incubator , and growth ( OD600 ) and fluorescence were monitored over 24 h . Potential hits were further evaluated by growing S . aureus USA300 LAC Δspa in TSB with 100 μM compounds for 24h , and assaying level of expressed Ebh in culture supernatants by the dot-blot , as described previously [22] . Biochanin A used for experiments were dissolved in DMSO as a vehicle to create stock solutions . To determine antibacterial activity of biochanin , its MIC was measured with broth microdilution method as detailed by Clinical and Laboratory Standards Institute ( CLSI ) M07-A10 guidelines , except that BHI was used as a growth medium . To study the effect of biochanin A on S . aureus growth , overnight cultures of S . aureus were diluted 100× in fresh BHI supplemented with biochanin A ( or DMSO vehicle control at the same volume ) at 40 μM and 80 μM , incubated with shaking at 37°C , and OD600 of the culture measured at regular time intervals . For all the other experiments , S . aureus was grown as for other assays , to an OD600 = 1 . 5 , but biochanin A ( or DMSO vehicle control at the same volume ) was added to the medium to a final concentration of 40 μM or 80 μM . Afterwards , bacteria were used either for the RNA isolation and qPCR , for the clumping assay , or for the static fibrinogen adhesion assay as described above . For experiments investigating mechanism of action of biochanin A , additional S . aureus strains were used: LAC strain lacking giant surface proteins ( Δebh sraP::φNΣ ) , and LAC strains expressing mgrA from a xylose-inducible plasmid pHC187 ( or empty control plasmid pEPSA5 ) grown in presence of 1% xylose . Bacteria were washed with ice-cold PBS , and RNA was purified with commercial kits as described before [22] . To perform a quantitative PCR ( qPCR ) , cDNA was generated with the High Capacity cDNA Reverse transcription Kit ( Applied Biosystems ) . Primers JK45 and JK46 were used for mgrA , and 41995031X and 419950330X for DNA gyrase ( gyrB ) , as described previously [22] , with primer efficiencies of 97% and 88% , respectively . qPCR was performed by amplifying 20 ng of cDNA in 20 μl total reaction volume with iTaq Universal SYBR Green Supermix ( Bio-Rad ) in CFX96 Touch Real-Time PCR System ( Bio-Rad ) under the following conditions: 3 min at 95°C , 40 cycles of 15 s at 95°C and 30 s at 55°C , followed by a dissociation curve . No template and no reverse transcription controls were performed in parallel . Data were analysed and Cq were determined with CFX Manager 3 . 1 ( Bio-Rad ) . Expression was normalized to that of gyrB , and values represent three biological replicates . For all assays data were pooled from at least two independent experiments . Differences between S . aureus strains were analyzed by ANOVA with a Dunnett’s multiple comparisons post-test . In case of strains expressing increasing levels of SasG , increasingly truncated Ebh , or exposed to increasing concentration of ArlRS inhibitor , the differences were analyzed by ANOVA with a post-test for linear trend . All CFU data were normalized by log transformation before statistical analysis . Two-tailed p values were calculated . Prism 7 ( GraphPad Software ) was used for statistical calculations .
|
Adhesion is central to the success of Staphylococcus aureus as a bacterial pathogen . We describe a novel mechanism through which S . aureus alters adhesion to ligands by regulating expression of giant inhibitory surface proteins . These giant proteins shield normal surface adhesins , preventing binding to ligands commonly found in the bloodstream and vessel walls . Using this unique regulatory scheme , S . aureus can bypass the need for individualized regulation of numerous adhesins to control overall adhesive properties . Our study establishes the importance of these giant proteins for S . aureus pathogenesis and demonstrates that a single regulatory cascade can be targeted for treating infections .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"fibrinogen",
"pathogens",
"microbiology",
"staphylococcus",
"aureus",
"collagens",
"plasmid",
"construction",
"dna",
"construction",
"molecular",
"biology",
"techniques",
"glycoproteins",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"adhesins",
"microbial",
"physiology",
"staphylococcus",
"medical",
"microbiology",
"proteins",
"microbial",
"pathogens",
"molecular",
"biology",
"endovascular",
"infections",
"blood",
"plasma",
"biochemistry",
"bacterial",
"physiology",
"blood",
"anatomy",
"virulence",
"factors",
"physiology",
"biology",
"and",
"life",
"sciences",
"vascular",
"medicine",
"glycobiology",
"organisms"
] |
2019
|
Staphylococcus aureus adhesion in endovascular infections is controlled by the ArlRS–MgrA signaling cascade
|
Motor learning has been extensively studied using dynamic ( force-field ) perturbations . These induce movement errors that result in adaptive changes to the motor commands . Several state-space models have been developed to explain how trial-by-trial errors drive the progressive adaptation observed in such studies . These models have been applied to adaptation involving novel dynamics , which typically occurs over tens to hundreds of trials , and which appears to be mediated by a dual-rate adaptation process . In contrast , when manipulating objects with familiar dynamics , subjects adapt rapidly within a few trials . Here , we apply state-space models to familiar dynamics , asking whether adaptation is mediated by a single-rate or dual-rate process . Previously , we reported a task in which subjects rotate an object with known dynamics . By presenting the object at different visual orientations , adaptation was shown to be context-specific , with limited generalization to novel orientations . Here we show that a multiple-context state-space model , with a generalization function tuned to visual object orientation , can reproduce the time-course of adaptation and de-adaptation as well as the observed context-dependent behavior . In contrast to the dual-rate process associated with novel dynamics , we show that a single-rate process mediates adaptation to familiar object dynamics . The model predicts that during exposure to the object across multiple orientations , there will be a degree of independence for adaptation and de-adaptation within each context , and that the states associated with all contexts will slowly de-adapt during exposure in one particular context . We confirm these predictions in two new experiments . Results of the current study thus highlight similarities and differences in the processes engaged during exposure to novel versus familiar dynamics . In both cases , adaptation is mediated by multiple context-specific representations . In the case of familiar object dynamics , however , the representations can be engaged based on visual context , and are updated by a single-rate process .
Object manipulation is an essential feature of everyday human behavior [1] . It represents a challenge for the motor system because grasping an object changes the relationship between the motor commands and the resulting movement of the arm [2] , [3] , [4] . Skillful manipulation thus requires the rapid adaptation of motor commands to the particular dynamics of the object . This adaptation can be facilitated by using stored knowledge such as an internal model of object dynamics [5] , [6] , [7] , [8] . Previous studies have examined the representation of dynamics using tasks in which subjects adapt their reaching movements to novel and unusual force-fields applied to the hand by robotic interfaces [9] , [10] , [11] , [12] , [13] , [14] . In these tasks , the force-field alters the normal dynamics of the arm , inducing movement errors which reduce gradually across many trials . Several models have been developed to explain how errors on each trial result in the gradual acquisition of an internal representation of the perturbing dynamics [15] , [16] , [17] , [18] , [19] , [20] . This approach is based on state-space models in which , typically , the internal state represents an estimate of the perturbation . The state estimate is updated after each trial based on the error experienced on the previous trial . Recently , a model which includes two internal states has been proposed [17] . The two states adapt independently at different rates ( one fast , one slow ) and sum to produce an estimate of the perturbation . Importantly , this dual-rate model can reproduce phenomena observed experimentally that a single-rate model cannot [17] , [19] , [ for an alternative view see 21] . State-space models have also been used to account for generalization , in which the movement direction ( the kinematic context ) varies across trials [15] , [19] . In this case , each movement direction has its own state in the model , representing the estimate of the perturbation associated with a movement in that direction . A generalization function specifies how an error experienced during a movement in one direction affects the states associated with other directions . These models can successfully reproduce the trial-by-trial changes in performance and the pattern of generalization observed during tasks that involve novel dynamics and visuomotor perturbations . The novel dynamics described above typically take subjects tens to hundreds of trials to learn [22] . In contrast , when interacting with everyday objects subjects adapt to the familiar dynamics much more rapidly . For example , when lifting an object of unknown mass , subjects adapt their predictive load and grip forces within just a few trials [5] , [23] , [24] , [25] , [26] , [27] , [28] , [29] . Given appropriate visual and other contextual cues subjects can even generate appropriate motor commands on the very first trial [5] . Rapid adaptation of grip force is also observed during bimanual object manipulation , when subjects pull on an object with one hand while stabilizing it with the other [30] , [31] . In this case , grip force adaptation can be shown to be context-specific , being locally confined to the movement direction in which the object is experienced [31] . As such , adaptation to the dynamics associated with familiar objects also appears to manifest context-dependent generalization . Moreover , the adaptation , while rapid , nevertheless occurs progressively as a result of trial-by-trial errors associated with an internal estimate of object parameters . This suggests that context-dependent state-space models may also be applicable to adaptation associated with manipulating objects with familiar dynamics . In a previous study , we examined adaptation to familiar object dynamics by presenting subjects with a virtual hammer-like tool [32] . The task involved rotating the object while keeping the grasp point stationary ( Figure 1A to C ) . This required subjects to generate a torque to rotate the object as well as a force to stabilize the grasp point . The dynamics were simulated using a novel robotic manipulandum ( the WristBOT [33] ) which can produce forces and torques that depend on the translational position and angular rotation of its vertical handle . The visual orientation of the object could also be varied from trial to trial . Results showed that subjects generate anticipatory forces in the direction appropriate for the visually-presented orientation of the object , even before they had been exposed to the dynamics . This suggests that subjects have pre-existing knowledge of the structural form of the dynamics that can be recalled based on vision . Moreover , when exposed to the dynamics of a specific object , subjects rapidly adapt the magnitude of their anticipatory forces over the first few trials to be appropriate for the particular mass of the object . To probe the representation of the dynamics , we examined the force magnitude at novel visual orientations of the object , where the dynamics had not been experienced . Consistent with previous studies , both of novel dynamic perturbations and familiar object dynamics , we showed that generalization of force magnitude was context-specific , being limited to orientations close to those at which the object had been experienced . However , in contrast to previous studies , the kinematics of the movements in our task were unchanged , and the pattern of generalization observed depended only on the visual context ( the orientation ) of the object . The aim of the current study was to determine whether a state-space model could account for rapid context-specific adaptation to the familiar dynamics of everyday objects and whether such adaptation would be mediated by a single-rate or dual-rate process . Specifically , we developed a state-space model that includes multiple context-dependent states and a generalization function tuned to the visual orientation of the object . The results show that a single-rate context-dependent model reproduces the time-course of adaptation and de-adaptation , along with the context-specific behavior observed in three experiments from our previous study . In these previous experiments , subjects were exposed to the dynamics of the object at a single orientation . The model also makes predictions with regards to exposure at multiple orientations which we confirm in two new experiments .
The study was approved by the local ethics committee and 68 subjects provided informed consent before participating . A preliminary analysis of experiments 1 , 2 and 3 has been previously reported [32] . We include the basic methods for these experiments here for completeness . In the current study , we extend our previous analysis , run additional subjects on a new condition in Experiment 2 , and perform three new experiments ( 4 , 5 and S1 ) . In addition , we apply state-space models to the data from all experiments . Subjects were seated at a virtual reality system and grasped the handle of a planar robotic manipulandum ( the WristBOT ) with their right hand ( Figure 1A ) . The WristBOT can generate translational forces in the horizontal plane and a rotational torque about its vertical handle [33] . This allowed us to provide haptic feedback ( forces and torques ) of the dynamics of a simulated object ( Figure 1B ) . A virtual reality display system provided visual feedback associated with the object and the task ( Figure 1C ) . The object was a small hammer-like tool which consisted of a mass on the end of a rigid rod . The task was to grasp the object by the handle at the base of the rod and rotate it back and forth between visual targets . Subjects were told that the object might wobble during the rotation and that they should try to maintain the handle as still as possible within the central home region . We ensured that the wrist operated near the midpoint of its range of motion so that a comfortable posture was adopted . Subjects maintained this posture throughout the experiment . The visual feedback was provided by a projection system that overlaid a visual image in the plane of the movement as previously described [33] . The visual object ( Figure 1B ) consisted of a circular handle ( radius 0 . 5 cm ) attached to a 4 cm square mass by an 8 cm rod ( width 0 . 2 cm ) . The position and orientation of the object was determined by the position and orientation the WristBOT handle . The home region was a 1 cm radius disc and the start and end targets for rotation were oriented rectangles ( 0 . 6 by 2 . 5 cm ) continuous with the disc ( Figure 1B ) . The orientation of the object could be varied between trials . For a given object orientation , subjects performed trials that alternated between clockwise ( CW ) and counter-clockwise ( CCW ) rotations of amplitude 40° between the two targets . The targets were ±20° relative to the central orientation at which the object was said to be presented . For example , when the object was presented at −90° the CW and CCW targets were at −70° and −110° , respectively . A trial began with the handle of the object stationary within the home region and the rod of the object aligned with either the CCW target or the CW target . The movement was cued by a tone and the appearance of the second target . The trial ended when subjects had rotated the object to reach the second target . Subjects were required to make the movement within 400 ms . They were warned if they took longer and had to repeat the trial if the movement exceeded 500 ms . Rest breaks ( 30–60 s ) were given every 3–5 minutes . The WristBOT simulated the dynamics of an object ( Figure 1B ) that consisted of a point mass ( m ) at the end of a rigid rod ( length r ) . In all experiments the object had a rod length of 8 cm and , except were explicitly stated , the mass was equal to 1% of the subject's body mass ( 1% BM ) . When rotating the object , subjects experienced a torque ( ) that depended on the angular acceleration ( ) of the handle: ( 1 ) Subjects also experienced a force at the handle that consisted of two orthogonal components . The first component was associated with the tangential force ( ) , which depended on the angular acceleration of the handle: ( 2 ) Note that acts in the direction of the tangential acceleration of the mass , and is thus perpendicular to the rod . The second component was associated with the centripetal force ( ) , which depended on the angular velocity ( ) of the handle: ( 3 ) Note that acts perpendicular to the tangential velocity of the mass ( towards the centre of rotation ) , and is thus parallel to the rod . The resultant force vector ( ) experienced by the subject at the handle is given by: ( 4 ) where is the two-dimensional force vector ( in the coordinate system of the WristBOT , as specified in Figure 1B ) , is the angle of the rod ( 0° is aligned with the y-axis ) and is a 2×2 clockwise rotation matrix . To avoid the need to compute angular velocity and acceleration , the dynamics were approximated by a simulation in which the point mass was attached to the end of the rod by a stiff spring ( 3000 N/m ) . Translation and rotation of the object caused the spring to stretch , which then generated forces ( and torques ) on the handle . At the same time , these forces were used to update the state of the simulated mass . A small amount of damping was applied to the mass to prevent oscillations ( 7 N m−1 s ) . An analysis of the kinematics and the forces and torques generated by the WristBOT during the task verified that this approximation accurately captured the dynamics of the object ( see Figure S1 in Text S1 ) . A specific description of the individual experiments can be found in the following sections . In general , each experiment consisted of multiple trials in which the visual orientation and dynamics of the object could be varied from trial to trial . Subjects always experienced the torque associated with rotating the object on every trial , whereas the forces experienced could be varied across trials in three different ways . On “exposure” trials , subjects experienced the full dynamics of the object such that the manipulandum produced the forces associated with rotating the object . On these trials , movement of the handle during the rotation was caused by the sum of the forces produced by the object and the forces produced by the subject . Specifically , if subjects produced forces that exactly opposed those produced by the object , the handle would remain stationary during the rotation ( as per the task requirements ) . On “zero-force” trials , the forces associated with rotating the object were turned off and the manipulandum produced no forces . On these trials the handle was free to move and any forces produced by subjects as they rotated the object resulted in a displacement of the handle . Finally , on “error-clamp” trials , the manipulandum simulated a stiff two-dimensional spring ( 1000 N/m ) centered on the handle position at the start of the trial . On these trials , any forces produced by the subject as they rotated the object were recorded as equal but opposite forces generated by the spring . These error-clamp trials minimize kinematic errors [34] and thus minimize adaptation ( or de-adaptation ) , allowing the anticipatory forces produced by subjects as they rotate the object to be assessed . The position and orientation of the handle and the force and torque generated by the manipulandum were saved at 1000 Hz for offline analysis using Matlab ( R14 , The MathWorks Inc . , Natick , MA , USA ) . Two measures were used to characterize the trial-by-trial performance of the subjects during the task . On zero-force and exposure trials , the peak displacement of the handle was measured , relative to its position at the start of the trial . The peak displacement of the handle ( in cm ) is a measure of error , because the task required subjects to keep the base of the object as still as possible during the rotation . A peak displacement of zero would thus indicate perfect performance . On error-clamp trials , the forces produced by subjects were measured . Subjects produce these forces in order to oppose the perturbing forces generated by the object . The peak force can be regarded as a measure of the subject's estimate of the mass of the object . As such , for a given error-clamp trial , we divide the peak force produced by the subject by the peak force which would have been generated by the object . We refer to this dimensionless ratio as the adaptation , which has a value of 1 if subjects produce forces which exactly compensate for the mass of the object . All statistical tests were performed using Matlab . All t-tests were paired and two-tailed . Various state-space models have been recently proposed to explain adaptation to dynamic [15] , [16] , [17] , [20] and kinematic ( visuomotor ) perturbations [18] , [19] . These models have yet to be applied to object manipulation , especially in light of experimental results that characterize trial-by-trial adaptation to object dynamics [29] , [30] , [32] . In the simplest case , the state-space adaptation model takes the form of a single-rate ( single-state ) model ( SRM ) as follows: ( 5 ) where x ( n ) is the state of the model on trial n , α is the retention constant , β is the learning-rate constant , and e is the error given by: ( 6 ) In the case of dynamic and visuomotor perturbation experiments , f is typically taken to be a dimensionless value which represents the magnitude and sign of the external perturbation [17] , [19] . In our case , f is the mass of the object . For simplicity we consider f , x and e in Equation 5 and 6 to be dimensionless quantities , where x can be thought of as the subject's internal estimate of the mass and e can be thought of as the error in this estimate . Where reported , values of x in the model are referred to as adaptation . To facilitate comparison with the equivalently named experimental quantity described above , adaptation in the model is expressed as the ratio of the estimated mass ( x ) to the actual mass ( f ) . In all experiments , f was set to the experimental mass of the object , which ( as described above ) was 1% BM in all experiments , except where explicitly stated . Recently , a dual-rate ( dual-state ) model ( DRM ) has been successfully applied to the results from both dynamic [17] and visuomotor [19] perturbations: ( 7 ) The rate-specific states sum together to produce the net output state of the model: ( 8 ) In these previous studies , the rate-specific states and the relative values for their retention and learning-rate constants have been associated with fast and slow adaptation processes [17] , [19] . In the current study , we fit both the SRM and DRM to the results of Experiment 1 ( details in the following section ) in order to determine whether adaptation to familiar object dynamics is mediated by a single-rate or dual-rate process . The single-rate and dual-rate models described above are context-independent models . To explain context-specific adaptation , multiple-state context-dependent versions of these models have been proposed [15] , [19] . The multiple-context form of the single-rate model ( MCSRM ) in the current study is given as follows: ( 9 ) where z ( n ) is a vector of the context-specific states on trial n and c is the context-selection vector ( described below ) . The net output state of the model is the sum of the context-specific states , weighted by the context-selection vector: ( 10 ) where x ( n ) is the net output state of the model on trial n . As originally described by Lee and Schweighofer [19] , the context-selection vector c defines which context is active on a given trial . It does this in a binary manner . Specifically , the element in c associated with the current context is 1 and all other elements are 0 . In this previous study , the context was the direction of movement in a visuomotor perturbation task . In the current study , the context is the visual orientation of the object specified in increments of 22 . 5° . As such , the context-dependent state vector z and the context-selection vector c contain 16 elements ( covering 360° in 22 . 5° steps ) . In our previous study , we reported a Gaussian-tuned pattern of generalization across different visual orientations of the object [32] . This graded pattern of generalization cannot be reproduced by the binary context-selection vector from the Lee and Schweighofer model , described above . Rather , we use a Gaussian function tuned to the visual orientation of the object to specify the 16 elements of the context-selection vector . The shape of the function ( Figure 1D ) is defined by two parameters which specify the standard deviation of the Gaussian ( σ ) and its offset ( d ) . The function is normalized to be 1 at the current orientation θ ( n ) and decays to d at θ ( n ) ±180° . The context-selection vector ( c ( n ) in Equations 9 and 10 ) thus becomes c ( θ ( n ) , σ , d ) , which is given by: ( 11 ) The function N ( x , σ ) in Equation 11 is a zero-mean Gaussian , as follows: ( 12 ) The function a ( θ ) in Equation 11 specifies a 16 element vector ( adjusted to the circular range of ±180° ) which centers the tuning function at the current orientation , as follows: ( 13 ) The MCSRM with the Gaussian context-selection function described above has 4 parameters ( MCSRM4 ) . In order to test our assumption that the context-selection function was Gaussian in form , we also considered a model in which the individual elements of c were free parameters . Assuming symmetry and a fixed value of 1 at the current orientation , 8 parameters defined the context-selection vector in this version of the model to give a total of 10 parameters ( MCSRM10 ) . In the context-independent versions of the model ( SRM and DRM , Equations 5 to 8 ) , the object is experienced at a single orientation and the error ( e ) is calculated simply as the difference between the actual mass of the object ( f ) and the subject's estimate of the mass ( x ) . However , in the context-dependent versions of the model ( MCSRM4 and MCSRM10 , Equations 9 and 10 ) , the object can be experienced at multiple orientations . This complicates the calculation of error because the displacement of the object during the task will be influenced by the compliance of the arm , which varies for perturbations in different directions [35] , [36] . To account for this , the calculation of error in the context-dependent versions of the model includes a compliance term which was determined experimentally . Specifically , we define a compliance-dependent error function for the model as follows: ( 14 ) As in the original error function ( Equation 6 ) , the error is due to the difference between the actual mass of the object ( f ) and the subject's estimate of the mass ( x ) . In this case , however , the error is the product of this difference with the compliance factor ( k ) and the gain factor ( g ) . The subscripts and superscripts on k in Equation 14 allow the compliance to vary for different orientations of the object ( θ ) and for positive and negative errors , respectively ( k+ when f>x , k− when f<x ) . This latter feature of the function allows the compliance to be different during adaptation and de-adaptation . Specifically , during adaptation , displacement of the hand is due to the object producing net forces on the subject . In contrast , during de-adaptation , displacement of the hand is due to the subject producing net forces on the object . The compliance can be different in each case . The value of k for a range of orientations was determined experimentally in a separate group of subjects ( see Experiment S1 in Text S1 for full details ) . It then remained fixed for all other experiments and subject groups . Because the peak displacement for these different subject groups could vary over a small but critical range , the gain factor ( g ) in Equation 14 was included as a free parameter in those models which implemented the compliance function . This parameter was close to 1 in all cases ( range 0 . 86 to 1 . 14 ) . As for peak displacement in the experiments , the compliance-dependent error has units cm . When fitting the various models to experimental data , parameters were estimated by a non-linear least-squares procedure performed in Matlab ( lsqnonlin ) . The absolute error output of the model ( from either Equation 6 or 14 ) was fit to the peak displacement trials series for each experiment . The mean peak displacement across subjects was used because the data for individual subjects was too noisy to obtain reliable fits . Confidence intervals for parameter estimates were calculated using a boot-strap procedure [17] . Specifically , the boot-strap was performed using 1 , 000 unique combinations drawn with replacement from the subject pool for each experiment . The model was fit separately to the mean peak displacement trial series for each of the 1 , 000 unique combinations of subjects . The 95% confidence intervals were calculated as the 2 . 5 and 97 . 5 percentile values from the distribution for each parameter obtained across the 1 , 000 individual fits . In the case where more than one model was fit to the experimental data , model selection was performed using the Bayesian Information Criterion ( BIC ) . The BIC for a particular model combines a “reward” for the goodness of fit with a “penalty” for the number of free parameters: ( 15 ) where σ2e is the variance in the residual errors of the fit , k is the number of free parameters and n is the number of data points ( the number of trials ) . Taking the difference in BIC values for two competing models approximates half the log of the Bayes factor [37] . A BIC difference of greater than 4 . 6 ( a Bayes factor of greater than 10 ) is considered to provide strong evidence in favor of the model with the lower BIC value [38] . The first experiment was designed to examine adaptation and de-adaptation in a single context ( Figure 2A ) . Subjects ( n = 8 ) were presented with the object at 0° . They performed an initial familiarization block of 48 pre-exposure trials during which the forces associated with the dynamics of the object were turned off ( zero-force trials ) . The main experiment consisted of 320 trials and began with a second block of 48 pre-exposure ( zero-force ) trials . Subjects then experienced the full dynamics of the object ( both torques and the perturbing forces of the object ) during an exposure phase of 224 trials . In the final 128 trials of the exposure phase , one error-clamp trial was inserted randomly every 8 trials for a total of 16 error-clamp trials ( 8 CW and 8 CCW trials ) . During the final post-exposure phase of 48 trials , the object forces were again turned off ( zero-force trials ) . Subjects were given rest breaks , randomly timed to occur every 3–5 minutes . We fit both the single-rate and the dual-rate context-independent models ( SRM and DRM ) to the trial series of the normalized mean peak displacement from Experiment 1 . Specifically , the mean peak displacement was calculated for each trial across subjects and then normalized across trials so that the mean of the pre-exposure phase was 0 and the maximum error across all trials was 1 . To characterize the time constant of adaptation ( during the exposure phase ) and de-adaptation ( during the post-exposure phase ) , we fit a single exponential function to each individual subject's data: ( 16 ) where n is the trial number . We calculated the mean time constant ( t ) across subjects . The second experiment was designed to examine the context-dependent pattern of generalization across multiple object orientations after exposure to the dynamics of the object at a single orientation ( Figure 3A ) . Two groups of subjects ( n = 12 in each group ) were first exposed to the object at 0° ( group 1 ) or 180° ( group 2 ) for 64 trials . They were then presented with multiple blocks of 30 trials in which they were first partially de-adapted with 8 zero-force trials presented at one of five possible probe orientations ( group 1: 0 , −22 . 5 , −45 , −90 and 180°; group 2: 180 , −157 . 5 , −135 , −90 and 0° ) . They were then re-exposed to the full dynamics of the object for 18 trials at the training orientation . The first 2 trials immediately before and immediately following the zero-force de-adaptation trials were error-clamp trials presented at the training orientation , during which the forces produced by subjects were measured . Probe orientations were presented in a pseudo-random order such that each probe orientation was presented once per cycle , with subjects performing 3 cycles ( 3 cycles×5 blocks per cycle = 15 blocks ) . The peak displacement of the handle during the first 4 de-adaptation trials was used as a measure of the magnitude of the anticipatory forces produced by subjects at each probe orientation . In addition , the peak displacement during the first 4 re-exposure trials ( at the training orientation ) , was also measured . Note that the probe orientations for each group ( 0° and 180° ) represented identical steps relative to the training orientation ( 0 , 22 . 5 , 45 , 90 and 180° ) . Rest breaks were given every 3–5 minutes as in Experiment 1 . We fit the 4 parameter and 10 parameter versions of the single-rate multiple-context model ( MCSRM4 and MCSRM10 , respectively ) to the trial series of the mean peak displacement across subjects . Both functions implemented the compliance-dependent error function ( Equation 14 , described above ) . The 4 parameter version of the MCSRM included the retention and learning-rate constants ( α and β ) as well as the width ( σ ) and offset ( d ) parameters for the Gaussian context-selection function . The 10 parameter version of the model included the 2 rate constants as well as 8 values specifying the individual elements of the context-selection vector , as described above . The compliance factor in the compliance-dependent error function ( k in Equation 14 ) was estimated by exposing subjects to the object at multiple orientations ( see Experiment S1 in Text S1 for full details ) . Briefly , subjects ( n = 12 ) experienced the object at 5 different orientations ( 0 , −45 , −90 , −135 and 180° ) . They were adapted and then de-adapted to the object dynamics multiple times at each orientation . A modified version of the SRM was then fit to the mean peak displacement trial series . In this modified SRM , the original error function ( Equation 6 ) was replaced by the compliance-dependent error function ( Equation 14 ) . The model fit the α and β parameters of the SRM ( Equation 6 ) along with 10 parameters which defined the compliance factor k . The third experiment was designed to examine adaptation to objects of different mass ( Figure 4A ) . Specifically , subjects ( n = 8 ) were exposed to 3 object masses ( 0 . 7% , 1 . 0% and 1 . 3% of the subject's body mass ) . The object was presented in separate blocks of 90 trials for each mass ( in pseudo-randomized order ) . Each block began with an exposure phase of 60 trials during which subjects experienced the object at the training orientation of 0° . In the subsequent 30 trials , one error-clamp trial was inserted randomly every 5 trials . For these error-clamp trials , the object was presented at 0° ( the training orientation ) or at −90° ( the transfer orientation ) . By examining the forces generated by the subjects on error-clamp trials we could assess their adaptation to the particular mass of the object at the training orientation and at the novel probe ( transfer ) orientation . Rest breaks were given every 3–5 minutes as in Experiment 1 . Simulated results were generated for this experiment using the selected MCSRM and its best-fit parameters from Experiment 2 . The fourth experiment was designed to examine adaptation when the object alternated sequentially between two different contexts ( Figure 5A ) . Subjects ( n = 8 ) performed blocks of 24 trials with the object presented at 180° and 0° consecutively across pairs of blocks . A cycle thus consisted of a pair of blocks at 180° and 0° . An initial pre-exposure cycle and a final post-exposure cycle were performed ( 2 blocks×24 trials per block = 48 trials ) . These consisted of zero-force trials as in Experiment 1 . The exposure phase included 9 cycles ( 18 blocks for a total of 432 trials ) . The entire experiment consisted of 11 cycles ( 11 cycles×2 blocks per cycle×24 trials per block = 528 trials ) . Rest breaks were given every 3–5 minutes as in Experiment 1 . Predictions for this experiment were generated using the selected MCSRM and its best-fit parameters from Experiment 2 . The fifth experiment was similar to Experiment 4 and was also designed to examine adaptation when the object switched between different contexts ( Figure 6A ) . In this case , the object was presented at five different orientations . Subjects ( n = 8 ) performed multiple exposure cycles that consisted of blocks of 20 trials in which the object was presented at one of five orientations ( 0° , −45° , −90° , −135° , 180° ) . The first two and the last two trials of each block were error-clamp trials . The remaining 16 trials were under the full dynamics of the object . A cycle consisted of a sequence of 5 blocks with each orientation presented once ( in a pseudo-random order ) . Subjects performed two pre-exposure cycles ( error-clamp trials but only 4 trials: 2 cycles×5 blocks×4 trials = 40 trials ) , five exposure cycles ( 5 cycles×5 blocks×20 trials = 500 trials ) and two post-exposure cycles ( as in the pre-exposure ) . This gave a total of 580 trials . Rest breaks were given every 3–5 minutes as in Experiment 1 . Predictions for this experiment were generated using the selected MCSRM and its best-fit parameters from Experiment 2 . The model and experimental results for Experiment 5 were analyzed as follows . Values for the two measures ( peak displacement and adaptation ) were binned for each block based on the absolute change in orientation relative to the previous block . The analysis included two bins; one for absolute changes in orientation of 45° and another for absolute changes of 90° or larger ( denoted 90+° ) . The 90+° bin was chosen because relative changes in orientation greater than 90° did not occur often enough to allow separate bins . Data from the first exposure cycle was not included because it represented the initial adaptation to the object dynamics . Two previous studies of object manipulation are relevant to the model presented in the current study . These previous studies examined grip force adaptation during a bimanual object manipulation task . Specifically , they characterized the time-course of adaptation and de-adaptation of grip force [30] and the pattern of generalization following single-context and dual-context exposure [31] . The MCSRM was fit concurrently to data from these four experiments ( see Figure S6 and further details in Text S1 ) .
The first experiment examined adaptation and de-adaptation of subjects ( n = 8 ) to the dynamics of the object at a single orientation ( Figure 2A ) . During the initial pre-exposure ( zero-force ) phase of the experiment , displacements of the handle were small ( initial light grey shaded plot in Figure 2B ) . The mean peak displacement across the final 8 trials of the pre-exposure phase was 0 . 23±0 . 08 cm ( subject mean ± standard deviation ) . In the exposure phase ( dark grey shaded plot in Figure 2B ) , upon introduction of the forces associated with the rotational dynamics of the object , displacement increased markedly on the first exposure trial to 1 . 13±0 . 24 cm , falling rapidly over subsequent trials . The mean peak displacement across the final 8 trials of the exposure phase was 0 . 43±0 . 15 cm , significantly larger than the final peak displacement for the pre-exposure phase ( two-tailed paired t-test , p<0 . 02; see inset of Figure 2B ) . During the post-exposure phase , when the object forces were again turned off , displacement increased on the first post-exposure trial to 0 . 80±0 . 31 cm , falling rapidly over subsequent trials ( final light grey shaded plot in Figure 2B ) . The mean peak displacement across the final 8 trials of the post-exposure phase was 0 . 26±0 . 13 cm , which was not significantly different from the final peak displacement for the pre-exposure phase ( two tailed paired t-test , p = 0 . 26; see inset of Figure 2B ) . The mean peak displacement across subjects was normalized for model fitting by subtracting the mean displacement across the final 8 trials of the pre-exposure phase and dividing by the maximum displacement across all trials . Peak displacement data normalized in this way for model fitting is referred to simply as “error” ( black trace in Figure 2C ) . Two alternative models were fit to the single-context experiment . The first was the context-independent single-rate model ( SRM ) described in Equation 5 , which has two free parameters . The second was the context-independent dual-rate model ( DRM ) described in Equations 7 and 8 , which has four free parameters . The SRM and DRM are equivalent to the single-state single-rate model and two-state multi-rate model of Smith et al [17] , respectively . Model parameters were estimated by a least squares minimization function in Matlab . Model selection was performed using BIC ( see Methods and Equation 15 ) . Both models were able to reproduce the time-course of adaptation and de-adaptation in the experimental data ( Figure 2C shows the SRM fit; Figure S3-A in Text S1 shows the fit for both SRM and DRM ) . The best-fit parameters for the SRM were α = 0 . 9513 and β = 0 . 2150 ( R2 = 0 . 7131 ) . The best-fit parameters for the DRM were α1 = 0 . 9808 , β1 = 0 . 0139 , α2 = 0 . 9453 , β2 = 0 . 2053 ( R2 = 0 . 7153 ) . The difference in BIC values for the two models was 8 . 3 , providing strong evidence in favor of the SRM ( see Methods ) . Thus , despite the slightly better fit to the data achieved by the DRM , its additional parameters were not justified . The 95% confidence limits for the SRM parameters were α = 0 . 9090–0 . 9796 and β = 0 . 1553–0 . 2668 ( estimated using a boot-strap procedure , see Methods ) . The 95% confidence limits for the model fit to the experimental data are shown in Figure 2C ( pink shading on the red line ) . The BIC analysis , which selected the SRM over the DRM , suggests that adaptation to familiar object dynamics is mediated by a single-rate process . An analysis of the exponential time constants for adaptation and de-adaptation provided further evidence for a single-rate process . Specifically , we fit exponential functions ( Equation 16 ) to the adaptation and de-adaptation phases of Experiment 1 individually for each subject . The mean time constant for adaptation was not significantly different from the mean time constant for de-adaptation ( adaptation: 3 . 3±1 . 1 trials; de-adaptation: 4 . 0±1 . 1 trials; two tailed paired t-test , p = 0 . 21; see inset of Figure 2C ) . Thus , we found no evidence for the phenomenon referred to as fast de-adaptation which is characteristic of adaptation to novel dynamics [17] , and dual-rate adaptation processes [17] . We found further evidence to support a single-rate process by performing simulations and an extensive search of the DRM parameter space ( see Text S1 for full details ) . First , the DRM parameters obtained from fitting the model to Experiment 1 exhibit neither spontaneous recovery nor savings when the appropriate experiments are simulated ( Figure S3 in Text S1 ) . These phenomena are characteristic of dual-rate adaptation processes [17] and their absence is thus consistent with a single-rate process . Second , an analysis of the DRM parameter space shows a wide range of solutions which provide a good fit to Experiment 1 ( Figure S4 in Text S1 ) . Such parameter redundancy would be expected if a DRM is fit to data generated by a single-rate process . In contrast , the SRM solutions which provide a good fit to Experiment 1 are confined to a narrow region of the parameter space ( inset of Figure S4 in Text S1 ) . Finally , we show that by excluding fast de-adaptation , results from Experiment 1 constrain the best-fit DRM solutions to a single-rate subspace which excludes spontaneous recovery and savings ( Figure S5 in Text S1 ) . This is especially striking given the wide range occupied by these solutions in DRM parameter space ( compare Figure S4 and Figure S5 in Text S1 ) . The second experiment examined context-specific de-adaptation at a range of probe orientations after exposure at a single training orientation ( Figure 3A ) . Subjects initially adapted to the dynamics of the object at 0° ( group 1 , n = 12 ) or 180° ( group 2 , n = 12 ) . They then completed multiple de-adaptation blocks at one of five probe orientations ( 0° , 22 . 5° , 45° , 90° and 180° relative to the training orientation ) . After each de-adaptation block , subjects were re-exposed to the dynamics at the original training orientation . Peak displacement during de-adaptation blocks at the various probe orientations gave a measure of context-dependent transfer . Peak displacement during the subsequent re-exposure blocks gave a measure of the context-dependent de-adaptation associated with each probe orientation . Analysis of the two error-clamp trials immediately before each de-adaptation block ( at the training orientation ) , shows no difference in adaptation as a function of the probe orientation ( ANOVA F ( 4 , 55 ) = 0 . 084 , p = 0 . 696 for group 1 and F ( 4 , 55 ) = 0 . 151 p = 0 . 988 for group 2 ) . Therefore , although each subject experienced the probe orientations in a different sequence , they were in a similar state of adaptation at the start of each de-adaptation block . We therefore felt justified in averaging the peak displacement data for the three de-adaptation blocks at each orientation within and then across subjects for the 2 groups . This allowed us to construct a composite trial series which included a single de-adaptation and re-exposure block for each orientation ( see black traces in Figure 3B and E ) . Consistent with the previous single-context experiment , during initial exposure to the full object dynamics , subjects adapted rapidly to the perturbing forces of the object ( E0° and E180° in Figure 3B and E ) . During the de-adaptation block at each probe orientation ( Dθ° ) , peak displacement on the initial trials was small for probe orientations that were far removed from the training orientation , increased progressively as probe orientations approached that of the training orientation , and reached a maximum value at the training orientation itself . Moreover , for probe orientations close to or at the training orientation , the initially large errors decreased rapidly across the 8 de-adaptation trials . This orientation-dependence for the initial de-adaptation error can be seen in Figure 3C and F ( black symbols are means across the first 4 de-adaptation trials for each orientation ) . A similar pattern was observed during the re-exposure block ( R0° and R180° in Figure 3B and E ) . As can be seen in Figure 3D and G ( black symbols are means across the first 4 re-exposure trials for each orientation ) , the smallest re-exposure errors occurred after de-adaptation trials at probe orientations far removed from the training orientation , whereas progressively larger errors were observed as the de-adaptation orientation approached the training orientation . As with de-adaptation errors , re-exposure errors fell rapidly as subjects quickly re-adapted to the full dynamics of the object ( Figure 3B and E ) . As reported previously [32] , the orientation-dependent behavior described above can be well captured by a half Gaussian centered on the training orientation ( see black lines in Figure 3C , D , F and G ) . This motivated the use of a Gaussian tuning function in the multiple-context version of the model . Specifically , the individual elements of the context-selection vector ( c in Equations 9 and 10 ) were set according to a Gaussian function centered on the current orientation of the object ( see Figure 1D ) . This version of the model had 4 parameters ( MCSRM4 ) : the two rate constants ( α and β ) and two additional parameters which specified the width ( σ ) of the Gaussian tuning function and its offset ( d ) . To test our assumption that the tuning function was Gaussian in form , we also fit a version of the model in which the individual elements of the context-selection vector were free parameters . This version of the model had 10 parameters ( MCSRM10; see Methods for full details ) . Model selection was performed using BIC , as described above . Both models were fit concurrently to the data from both groups ( 0° and 180° ) . The best-fit parameters for the MCSRM4 were α = 0 . 9811 , β = 0 . 0451 , σ = 26 . 3° d = 0 . 09 ( R2 = 0 . 8523; see model fit in Figure 3B and E ) . Fitting the MCSRM10 yielded similar values for the rate parameters ( α = 0 . 9812 , β = 0 . 0624 ) with a small improvement in the fit ( R2 = 0 . 8581 ) . The individual values fit for the context-selection vector are plotted in the inset of Figure 3B ( black points ) and are shown with the Gaussian tuning function from the MCSRM4 fit for comparison ( red line ) . The difference in BIC values between the two models was 23 . 8 , providing strong evidence in favor of the MCSRM4 ( see Methods ) . As such , allowing the individual elements of the context-selection vector to vary independently provided only a small ( and unjustified ) improvement in the ability of the model to fit the experimental data . Thus , the assumption that the tuning function is Gaussian in form appears to be valid . The 95% confidence limits for the MCSRM4 parameters were α = 0 . 9760–0 . 9870 , β = 0 . 0300–0 . 0540 , σ = 24 . 2–30 . 1° and d = 0 . 08–0 . 13 ( estimated using a boot-strap procedure , see Methods ) . The 95% confidence limits for the model fits to the experimental data are also shown in Figure 3B and 3E ( pink shading on the red line ) . Note that values for the β parameter differ substantially between Experiment 1 ( β = 0 . 2150 ) and Experiment 2 ( β = 0 . 0451 ) . This lower value for Experiment 2 is expected because in the MCSRM the multiple states associated with neighboring contexts contribute to adaptation . In contrast , in the SRM a single context-independent state is responsible for adaptation , which results in a larger value for the β parameter . Peak displacement data from the MCSRM4 fit was analyzed in the same way as the experimental data in order to produce plots summarizing the orientation-dependent behavior of the model . Results from this analysis of the model data can be compared with the equivalent analysis of the peak displacement data from the experiment . In both cases , the de-adaptation error ( Figure 3C and F ) and the re-exposure error ( Figure 3D and G ) from the model well captured the orientation-dependent behavior seen experimentally ( red versus black symbols and lines ) . In addition to fitting the MCSRM4 concurrently to the data for both groups of subjects , we also fit the model independently to data from each group . This allowed us to test the generality of the parameters with different groups of subjects who experienced the object across a different range of orientations . Importantly , the parameters were similar when the model was fit separately to data from the two groups . For group 1 ( training at 0° ) the best-fit parameters were α = 0 . 9798 , β = 0 . 0635 , σ = 24 . 4° d = 0 . 15 ( R2 = 0 . 8598 ) and for group 2 ( training at 180° ) the best-fit parameters were α = 0 . 9817 , β = 0 . 0412 , σ = 25 . 9° d = 0 . 06 ( R2 = 0 . 8956 ) . This represents an important validation for the model because the peak displacement data varies substantially between the two groups ( compare especially Figure 3B and E , and Figure 3C and F ) . The third experiment examined adaptation to objects of varying mass ( Figure 4A ) . In separate blocks , subjects ( n = 8 ) were exposed ( training orientation 0° ) to the dynamics of objects with three different masses ( 0 . 7% , 1 . 0% and 1 . 3% of the subject's body mass ) . Adaptation was then examined using error-clamp trials at the training orientation and a novel transfer orientation ( −90° ) . Because the mass of the object varied in this experiment , the level of adaptation was expressed relative to the 1 . 0% BM object . Adaptation increased with the mass of the object at both the training orientation ( Figure 4B , blue squares ) and the transfer orientation ( Figure 4B , red circles ) . Subject-by-subject linear fitting yielded slopes that were significantly different between the training and transfer orientation ( 0 . 59±0 . 17 for training , 0 . 22±0 . 15 for transfer; two-tailed paired t-test p<0 . 005; data previously reported [32] ) . A simulation of Experiment 3 was performed using the best-fit parameters for the MCSRM4 determined from Experiment 2 . The model reproduced the pattern of adaptation seen experimentally ( compare Figure 4B and C ) and made a prediction regarding the final peak displacement associated with each mass . We confirmed this prediction in a new analysis ( Figure 4D ) . Specifically , the model predicted that the final peak displacement would increase with the mass of the object . This was confirmed statistically in an analysis of the experimental data ( ANOVA F ( 2 , 21 ) = 0 . 0839 p<0 . 005 ) , showing that subjects tolerate larger errors for heavier objects . The experiments described thus far have involved exposing subjects to the full dynamics of the object in a single context ( a single orientation ) . Even in experiments 2 and 3 , in which the object was presented at multiple orientations , exposure to the full dynamics was restricted to a single training orientation . In the case of exposure at two or more orientations , the MCSRM makes several predictions . These predictions are tested in two new experiments described below . In the case of dual-context adaptation , in which the object alternates between two orientations in short blocks ( 180° and 0°; Figure 5A ) , the MCSRM4 makes two predictions . First , due to the relatively narrow tuning curve in the model ( 26 . 3° ) , there should be little transfer between orientations separated by more than 60° . Therefore , adaptation during the initial exposure for each context should be largely independent . As such , after first adapting to the dynamics at 180° , subjects should adapt essentially “from scratch” to the dynamics at 0° , with little benefit ( transfer ) from the exposure at 180° . The model similarly predicts that de-adaptation for each context in the post-exposure phase should be largely independent . As such , after first de-adapting to the dynamics at 180° , subjects will have to de-adapt essentially “from scratch” to the dynamics at 0° . Second , the model has a relatively small retention constant ( α = 0 . 9811 ) which means that non-active contexts should de-adapt quickly . Specifically , after the initial adaptation at both orientations , and as the blocks continue to alternate , there will be some amount of de-adaptation in the non-active context . As such , during subsequent blocks at a particular orientation , there will be a small about of re-adaptation within each block . The results from the MCSRM4 simulation , using the best-fit parameters from Experiment 2 , well matched the peak displacement trial series ( mean across subjects; n = 8 ) for the experimental data ( Figure 5B; R2 = 0 . 5996 ) . ) . In some cases , the model over- or under-estimated the experimental peak displacement by a few millimeters ( for example , the first few trials in the initial exposure blocks , see Figure 5C ) . However , the model results are a simulation ( not a fit ) based on parameters obtained from different groups of subjects . As such , small discrepancies , especially for the large displacements associated with initial exposure , are not unexpected . With regards to the first prediction , the experiment confirmed that during the initial adaptation there would be little transfer between the initial exposure in the first context and the initial exposure in the second context ( see Figure 5C ) . The largely independent adaptation at each orientation was confirmed statistically by comparing peak displacement values on the first and last trials of each of the initial exposure blocks ( E180° and E0° in Figure 5C , p<0 . 0001 in both cases ) . In addition , upon transition to the second context , peak displacement increased significantly , consistent with the limited transfer predicted by the model ( Figure 5C , E180° to E0° p<0 . 0001; two-tailed paired t-test ) . The experiment also confirmed that during the final post-exposure blocks there would be little transfer of de-adaptation between the two contexts ( see Figure 5D ) . The largely independent de-adaptation at each orientation was confirmed statistically by comparing peak displacement values on the first and last trials of each of the post-exposure blocks ( Figure 5D , p = 0 . 0141 for D180° , p = 0 . 0051 for D0°; two-tailed paired t-tests ) . In addition , upon transition to the second context , peak displacement increased significantly , consistent with the limited transfer predicted by the model ( Figure 5D , D180° to D0° p = 0 . 0230; two-tailed paired t-test ) . With regards to the second prediction , the experiment confirmed that , after the initial two adaptation blocks , there would be a small amount of re-adaptation within each block , due to de-adaptation in the non-active context . The re-adaptation within each block can be seen in the peak displacement trial series ( Figure 5B ) , although individual blocks are noisy . The re-adaptation is best appreciated and statistically confirmed when multiple blocks are averaged for each orientation ( Figure 5E ) . Comparing the first and last trial in the average block trial series for each orientation shows that a significant reduction in error occurred within the block ( Figure 5E , p = 0 . 012 for E180° , p = 0 . 001 for E0°; two-tailed paired t-tests ) . Due to the varying compliance associated with the object at 180° and 0° ( see Figure S2 in Text S1 ) , the model made a third prediction regarding the peak displacement and adaptation associated with each orientation . Specifically , the relatively higher compliance associated with the object at 180° should result in a higher peak displacement for 180° blocks than for 0° blocks . This was statistically confirmed when peak displacement was averaged across blocks for each orientation ( Figure 5F , p = 0 . 003 ) . In addition , the model predicts that the higher peak displacement experienced at 180° should drive a larger adaptation to the dynamics of the object at this orientation , relative to 0° . This was similarly confirmed when error-clamp trials were analyzed ( Figure 5G , p<0 . 001 ) . The prediction described above regarding de-adaptation in non-active contexts was investigated further in Experiment 5 which examined the case of multiple-context adaptation ( Figure 6A ) . In this experiment , subjects ( n = 8 ) experienced the object at 5 orientations ( 0° , −45° , −90° , −135° , 180° ) , with each orientation presented in blocks of 20 trials . Five cycles of 5 blocks were performed and a block for each orientation was presented once per cycle in pseudo-random order . The MCSRM4 ( using the best-fit parameters from Experiment 2 ) predicts that the degree of de-adaptation that occurs in a particular non-active context will depend on how far it is removed from the active context . For example , during an exposure block in which the object is presented at −45° , the amount of de-adaptation occurring at 0° and −90° will be less than that occurring at −135° and 180° . This is because for orientations close to the active context , some amount of adaptation occurs ( due to spread of adaptation from the active context to its non-active neighbors ) . This spread of adaptation from the active context offsets the effects of de-adaptation ( which occurs equally in all contexts ) . In contrast , at orientations further removed from the active context , the effects of de-adaptation dominate . As described in the Methods , the peak displacement and adaptation data from the model and experiment were binned based on the relative change in orientation between consecutive blocks . This yielded two bins ( 45° and 90+° ) . The model predicted that the initial adaptation at the start of each block should be larger for the 45° bin than for the 90+° bin , whereas the final adaptation at the end of each block should be the same for the two bins . This prediction was confirmed in the experimental results . Specifically , the initial adaptation ( measured using two error-clamp trials at the start of each block ) was larger for the 45° bin than for the 90+° bin ( Figure 6B; p = 0 . 003 two-tailed paired t-test ) , whereas the final adaptation ( measured using two error-clamp trials at the end of each block ) was the same for the two bins ( Figure 6C; p = 0 . 93 ) . Model predictions were also confirmed in the experimental peak displacement trial series for the binned blocks ( Figure 6D ) . Specifically , the higher level of adaptation for the 45° bin relative to the 90+° bin was reflected by a lower peak displacement for the 45° bin relative to the 90+° bin ( p = 0 . 015 , two-tailed paired t-test between the first trial of the 45° and 90+° bins , as shown in Figure 6D ) . By the end of the block this difference in peak displacement between the two bins had reduced to non-significant levels ( p = 0 . 172 , two-tailed paired t-test between the last trial of the 45° and 90+° bins , as shown in Figure 6D ) . In additional , the degree of re-adaptation within the binned blocks also confirmed model predictions . Specifically , a small ( non-significant ) amount of re-adaptation occurred during the block for the 45° bin ( p = 0 . 077 , two-tailed paired t-test between the first and last trials of the 45° block , as shown in Figure 6D ) , whereas a larger ( significant ) amount of re-adaptation occurred during the block for the 90+° bin ( p = 0 . 040 , two-tailed paired t-test between the first and last trials of the 90+° block , as shown in Figure 6D ) . The results from the two multiple-context adaptation experiments described above show that , even when subjects receive constant exposure to the object dynamics , non-active contexts de-adapt in a manner consistent with the MCSRM . The MCSRM4 also successfully captured results from two previous studies of object manipulation ( see Figure S6 in Text S1 ) . These studies examined grip force adaptation during a bimanual manipulation task . The model concurrently fit the data from four experiments across the two studies , two which characterized the time-course of adaptation and de-adaptation ( Figure S6-B and S6-C in Text S1 ) [30] and two which characterized the pattern of context-specific generalization ( Figure S6-D and S6-E in Text S1 ) [31] . The best-fit parameters were α = 0 . 8122 , β = 0 . 2568 , σ = 18 . 2° and d = 0 . 16 ( R2 = 0 . 8419 ) . In addition , to facilitate comparison across the different object manipulation tasks considered in the current study , exponential functions were fit to two studies which have characterized the time-course of adaptation . The first study was the bimanual task described above [30] , which yielded a time constant for grip force adaptation of 0 . 9 trials . The second study examined adaptation during a task in which subjects lifted an object with an asymmetrically offset centre of mass [29] , which yielded a time constant for compensatory torque adaptation of 0 . 9 trials ( see Figure S7 in Text S1 ) .
We have used a context-dependent state-space model to examine how subjects adapt when manipulating objects with familiar dynamics . The model reproduces results from our previous study [32] , including the time-course of adaptation and de-adaptation ( Figure 2 ) as well as the context-specific behavior associated with exposure to the dynamics of the object at a single orientation ( Figure 3 and 4 ) . In addition , adaptation and de-adaptation were found to occur at similar rates , which we show to be diagnostic of a single-rate process . Thus , in contrast to the dual-rate process thought to underlie adaptation to novel dynamics [17] , adaptation to familiar dynamics appears to be mediated by a process which adapts at a single rate . We also confirm predictions of the model with two new experiments in which subjects were exposed to the dynamics of the object at multiple orientations ( Figure 5 and 6 ) . A key aspect of the model is that separate states are associated with different visual contexts of the object . This context-specific state represents the subject's estimate of object mass for a particular visual orientation . The state ( mass estimate ) is updated based on the kinematic error experienced on each trial and a generalization function determines how errors in the active context affect the states associated with the non-active contexts . In addition , each state undergoes spontaneous trial-based decay independent of the current context . Finally , by considering a version of the model in which the individual elements of the context-selection vector were free parameters , we show that the generalization function is Gaussian in form . Having shown that the model can account for our previous results obtained during exposure to the dynamics of the object at a single orientation , we test predictions of the model with two new experiments in which subjects experienced the dynamics at multiple orientations . We first examined the case of dual-context exposure , in which the object alternated between two different orientations ( 0° and 180° ) . The model correctly predicted that , due to the relatively narrow generalization function , adaptation and de-adaptation would be partially independent in each context ( Figure 5C and D , respectively ) . Specifically , after initial adaptation ( or de-adaptation ) in the first context , the model correctly predicted that there would be little benefit ( transfer ) to the second context . In addition , the spontaneous decay in the model predicted that small decreases in performance would occur each time the context alternated ( Figure 5E ) . This partial de-adaptation in the non-active context was further examined in the second new experiment , in which the object switched randomly between five different orientations . In this case , the model correctly predicted that partial de-adaptation would occur in all non-active contexts , but would be greater for those contexts furthest removed from the active one ( Figure 6 ) . Taken together , the modeling and experimental results from the current study provide further support that internal models of familiar object dynamics are mediated my multiple context-specific representations . Whereas a dual-rate process is thought to mediate adaptation to novel dynamics [17] , results from the current study suggest that adaptation to the familiar dynamics of everyday objects is mediated by a single-rate process . Specifically , in the case of novel dynamics , a dual-rate model ( DRM ) has been proposed which has fast and slow adaptation processes acting in parallel [17] . Such a model explains various phenomena observed during adaptation to novel dynamic perturbations , such as fast de-adaptation , spontaneous recovery and savings ( see Figure S3 and the associated discussion in Text S1 ) . In Experiment 1 of the current study ( Figure 2 ) , we found no evidence for fast de-adaptation ( see inset of Figure 2C ) . However , due to methodological constraints associated with our task , we were unable to test for spontaneous recovery and savings . Rather , in simulations and an extensive search of the DRM parameter space ( see Figure S4 and Figure S5 of Text S1 ) , we show that the absence of fast adaptation is a diagnostic feature of single-rate adaptation processes . Specifically , we show that DRM solutions which do not exhibit fast adaptation , exhibit neither spontaneous recovery nor savings . Such solutions are single-rate parameterizations of the DRM . Results of the current study thus suggest that the adaptation processes engaged when subjects are exposed to novel dynamics differ from the processes engaged during the manipulation of everyday objects . Fast adaptation to the dynamics of familiar objects is consistent with the idea that there are two components to learning: structural learning and parameterization [41] , [42] . In structural learning , the motor system extracts the structural form of the sensorimotor transformation that underlies a particular task . Once the structure has been learned , adaptation to different tasks that share the same structure can proceed rapidly through parameterization of the existing structure . When considering a set of similar objects , the structure can be represented by the form of the equations which describe the dynamics , whereas the parameters represent the values that vary across objects ( such as mass and moment of inertia ) . As such , when manipulating familiar objects with dynamics which are captured by an existing structure , adaptation can be rapid because it primarily involves parameterization . For example , when lifting an object of unknown mass , subjects adapt their predictive load and grip forces rapidly within a few trials [5] , [23] , [24] , [25] . Similarly , rapid adaptation of digit forces occurs when subjects lift objects with an asymmetrical centre of mass [26] , [27] , [28] , [29] . Rapid adaptation was also observed in the current study , in which the time constant for adaptation was 3 . 3 trials . In contrast , when subjects are presented with objects which have novel and unusual dynamics they adapt much more slowly . For example , objects with internal degrees of freedom can take hundreds of trials to learn [43] , [44] . Similarly , adaptation to novel state-dependent dynamics has a long time course [22] . For example , the time constant for adaptation to velocity-dependent curl fields is around 40 trials ( see Figure S3-B of Text S1 ) . We suggest that this longer time course reflects the requirement to learn the unfamiliar structure of the dynamics . The formulation of the current model draws upon various state-space models that have been previously applied to adaptation in the motor system [15] , [16] , [17] , [18] , [19] , [20] . For example , dynamic perturbation studies which examined velocity-dependent force-fields showed that context-specific adaptation could be modeled using a multiple-state model that included a generalization function tuned to the current context [15] , [16] . In this case , the context was the target direction which thus involved a change in the kinematics of the movement . In contrast , in our experiments the movement kinematics were constant because the arm remained in the same configuration and subjects made the same movement on all trials . What varied in our experiments was the visual orientation of the object . Results from the current study show that , in addition to kinematic contexts and novel state-dependent dynamics , multiple-context state-space models can also be applied to visual contexts and the familiar dynamics of objects . The role of visual context was particularly striking in Experiment 2 ( see Figure 3 ) . In this experiment , subjects experienced the dynamics of the object at a single orientation only . The context-dependent behavior observed in this experiment was in response to changes in the visual orientation of the object; the dynamics and kinematics associated with probe trials remained constant . Moreover , based on fitting the model to this experiment , we were able to predict how subjects would behave when they experienced the dynamics of the object at multiple orientations . Our results suggest that the dynamics of familiar objects are not represented globally , but rather as a set of local representations that are selectively engaged by the current context of the object ( its visual orientation ) . Previous studies have also examined the ability of contextual cues to engage separate representations in the motor system . For example , a number of studies have focused on the ability of subjects to learn opposing dynamic or kinematic perturbations , in which the direction of the perturbation alternates over successive blocks . In the absence of an appropriate contextual cue , the opposing perturbations compete for a single representation and concurrent adaptation is not possible [45] , [46] , [47] , [48] , [49] , [50] , [51] , [52] . However , if a suitable contextual cue is associated with each perturbation , separate representations are engaged and concurrent adaptation can be achieved [19] , [20] , [53] , [54] , [55] , [56] , [57] . In the current study , the dual-context experiment similarly required subjects to adapt to opposing dynamic perturbations because the forces produced by the object reversed direction as it alternated between 180° and 0° ( Figure 5 ) . In this case , however , subjects had perfect contextual information ( the visual orientation of the object ) which allowed them to produce forces in the appropriate direction from the very first trail . Rather , the observed changes in performance when the context switched ( Figure 5E ) were associated with adaptation ( and de-adaptation ) of force magnitude . Of relevance to the current study , two previous studies of object manipulation have also characterized the rate of adaptation and de-adaptation [30] , and the pattern of generalization [31] of familiar object dynamics . In these previous studies , grip force adaptation was examined during a bimanual object manipulation task . Importantly , the current model successfully reproduced key features from these previous results ( Figure S6 in Text S1 ) . Specifically , results from four experiments ( two experiments from each previous study ) were concurrently fit by the model , which reproduced the time-course of adaptation and de-adaptation reported in the first study and the local context-dependent pattern of generalization reported in the second study . Interestingly , the rate of de-adaptation of grip force in the first study was found to be slower than the rate of adaptation [30] . This result could be reproduced by the current model by assuming that grip force de-adaptation is not an active error-driven process , but rather occurs through passive decay ( as captured by the retention constant ) . This assumption seemed justified because , in the task examined by these previous studies , producing a grip force response on the unlinked de-adaptation trials did not cause the virtual object to displace . In the absence of a kinematic error , de-adaptation in the task may thus rely on passive trial-by-trial decay . The relatively low value for the retention constant ( α = 0 . 8122 ) obtained when fitting the model is consistent with this suggestion , which would allow grip force to rapidly decay . Similarly , the relatively high learning-rate constant ( β = 0 . 2568 ) would allow grip force to adapt quickly despite this high rate of passive decay . In the current study , passive decay was also found to be an important process , responsible for the progressive de-adaptation of non-active contexts despite ongoing exposure to the object . We have suggested that because the structure of dynamics which are familiar is already represented by the sensorimotor system , adaptation is rapid and engages a single-rate process . This is in contrast to adaptation to novel force-fields , which proceeds more slowly and is mediated by a dual-rate process . However , this view is not the only interpretation of our results . For example , adaptation to the dynamics of the object in our task may be inherently easier . Alternatively , there may be something inherently difficult about adapting to velocity-dependent force-fields . This argument is weakened by the variety of previous studies reviewed above . Specifically , in studies of familiar dynamics , adaptation is always rapid . Moreover , for those cases of familiar dynamics examined in the current study ( the current task , and the bimanual object manipulation and lifting tasks in Figure S6 and Figure S7 in Text S1 , respectively ) such adaptation appears to be mediated by a single-rate process . In contrast , in previous studies of novel dynamics reviewed above , adaptation is always slow , and in the case of velocity-dependent force-fields , is mediated by a dual-rate process . In further support of our view , the rate of adaptation to a sensorimotor task can increase dramatically when subjects become familiar with the structure of the task [41] . Thus , we hypothesize that familiarity plays a key role in determining the rate of adaptation and may explain the observed differences in the processes which mediate adaptation to familiar versus novel dynamics . In summary , by using state-space models , the current study has highlighted similarities and differences in the processes which mediate adaptation to novel and familiar dynamics . In both cases , adaptation is mediated by multiple context-specific representations . In the case of novel dynamics , these representations are selected based on the kinematic context of the movement . In the case of familiar object dynamics , the representations can be selected based on the visual context of the object . And whereas the relatively slow adaptation to novel dynamics is mediated by a dual-rate process , the rapid adaptation observed when subjects manipulate objects with familiar dynamics appears to be mediated by a single-rate process . Thus , the human ability to skillfully manipulate objects appears to be mediated by multiple representations . These representations , which capture the local dynamics associated with specific contexts of the object , can be selectively engaged by visual information alone , and are updated based on the dynamics of specific objects via a single-rate adaptation process .
|
Skillful object manipulation is an essential feature of human behavior . How humans process and represent information associated with objects is thus a fundamental question in neuroscience . Here , we examine the representation of the mechanical properties of objects which define the mapping between the forces applied to an object and the motion that results . Knowledge of this mapping , which can change depending on the orientation with which an object is grasped , is essential for skillful manipulation . Subjects performed a virtual object manipulation task by grasping the handle of a novel robotic interface which simulated the dynamics of a familiar object which could be presented at different orientations . Using this task , we show that adaptation to the properties of a particular object is extremely rapid , and that such adaptation is confined to the specific orientation at which the object is experienced . Moreover , the pattern of adaptation observed when the orientation of the object and its mechanical properties were changed from trial-to-trial was reproduced by a model which included multiple representations and a generalization function tuned for object orientation . These results suggest that the skillful manipulation of objects with familiar dynamics is mediated by multiple context-specific representations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"motor",
"systems",
"computational",
"neuroscience",
"biology",
"computational",
"biology",
"neuroscience"
] |
2011
|
A Single-Rate Context-Dependent Learning Process Underlies Rapid Adaptation to Familiar Object Dynamics
|
Wounded leaves of Arabidopsis thaliana show transient immunity to Botrytis cinerea , the causal agent of grey mould . Using a fluorescent probe , histological staining and a luminol assay , we now show that reactive oxygen species ( ROS ) , including H2O2 and O2− , are produced within minutes after wounding . ROS are formed in the absence of the enzymes Atrboh D and F and can be prevented by diphenylene iodonium ( DPI ) or catalase . H2O2 was shown to protect plants upon exogenous application . ROS accumulation and resistance to B . cinerea were abolished when wounded leaves were incubated under dry conditions , an effect that was found to depend on abscisic acid ( ABA ) . Accordingly , ABA biosynthesis mutants ( aba2 and aba3 ) were still fully resistant under dry conditions even without wounding . Under dry conditions , wounded plants contained higher ABA levels and displayed enhanced expression of ABA-dependent and ABA-reporter genes . Mutants impaired in cutin synthesis such as bdg and lacs2 . 3 are already known to display a high level of resistance to B . cinerea and were found to produce ROS even when leaves were not wounded . An increased permeability of the cuticle and enhanced ROS production were detected in aba2 and aba3 mutants as described for bdg and lacs2 . 3 . Moreover , leaf surfaces treated with cutinase produced ROS and became more protected to B . cinerea . Thus , increased permeability of the cuticle is strongly linked with ROS formation and resistance to B . cinerea . The amount of oxalic acid , an inhibitor of ROS secreted by B . cinerea could be reduced using plants over expressing a fungal oxalate decarboxylase of Trametes versicolor . Infection of such plants resulted in a faster ROS accumulation and resistance to B . cinerea than that observed in untransformed controls , demonstrating the importance of fungal suppression of ROS formation by oxalic acid . Thus , changes in the diffusive properties of the cuticle are linked with the induction ROS and attending innate defenses .
The cuticle is mainly considered as a constitutive barrier against water loss , irradiation , xenobiotics or pathogens [1] , [2] . The structure of this lipid boundary layer covering aerial parts of plants is made of waxes covering and interspersed in cutin , a polymer layer formed by a network of esterified Ω-hydroxylated fatty acids that are produced and secreted by the epidermis cells [3] . Waxes comprise a mixture of very long-chain fatty acids ( 24–36 carbon atoms ) that seem ubiquitously present in most plant species . In addition , triterpenes , β-diketones as well as phenylpropanoids are associated with the wax fraction [4] , [5] . The enzymatic machinery for wax biosynthesis is in the endoplasmic reticulum and members of a subfamily of ATP binding cassette ( ABC ) transporters export the resulting products through the membrane to the cell wall [5] . In many plant species the cutin polyester contains C16 or C18 , fatty acids as well as glycerol [6] , [7] . The fatty acids in the cutin can be hydroxylated at midchains ( C8 , C9 , or C10 ) or in Ω-positions and are linked together or to glycerol by ester bonds . It is still unclear if the cutin polymers exist as free polymers or if they are anchored in some ways to the cell wall [8] . The polymerization and the transport of the cutin precursors are likely to occur in the cell and conveyed to the cell wall via oleophilic droplets , secretion vesicles , lipid transfer proteins ( LTP ) or ABC transporters [7] , [8] . While considerable knowledge is available on single components , the detailed chemical structure of the entire cuticle is still not known and the relation between the structure and the biological function of the individual components remains to be defined . The cuticle has been proposed to be a physical barrier to the penetration by pathogens . This intuitive view is supported by the fact that most pathogens that penetrate directly through the cell wall produce cutinase . However , the biological relevance of cutinase or cutinolytic lipase could not be documented in all cases . Studies using antibodies directed against cutinase support a role for this enzyme in pathogenicity [9] , [10] . Moreover , the effects of cutinase disruption were studied in various fungal pathogens , but only few studies found supportive evidence [11] , [12] , while most other results question the function of cutinase as a breaching enzyme [13]–[16] . These contradictory results might be the consequence of the functional redundancy of these enzymes , making gene disruption experiments difficult to interpret . Components of the cuticle might function as important developmental cues perceived by invading microorganisms [6] . For instance , cutin monomers can induce the expression of cutinase [17] , [18] or act as a plant signal for the induction of germination and appressorium in fungal pathogens [17]–[20] . Similarly , surface waxes can also affect the development of fungi at the plant surface [21] . The plant itself might also perceive degradation products of its own cuticle . Synthetic cutin monomers of the C18 fatty acid family applied on leaves of barley or rice increased resistance to Erysiphe graminis and Magnaporthe grisea respectively , without detectable direct fungicidal activity [22] , [23] . Cell cultures of Solanum tuberosum respond to cutin monomers by medium alkalinization , ethylene ( ET ) production and accumulation of defense-related genes [23] . Abraded cucumber hypocotyls respond to cutin hydrolysates of cucumber , apple and tomato by producing H2O2 [24] that has been repeatedly associated with defense either as a signal , as an executer of cell death or as cofactor in the strengthening of the cell wall [25]–[27] . A decrease in the lesion size caused by Rhizoctonia solani was observed when bean leaves were inoculated with spore droplets amended with a fully active cutinase compared to droplets with an inactive cutinase or without cutinase [28] . These observations support the notion that plants have the potential to recognize breakdown products of the cuticle and activate defense-related mechanisms . Experiments with transgenic plants over expressing an active fungal cutinase in the apoplasm ( CUTE plants ) also support these observations [29] . An alternative explanation was provided by observations in CUTE plants as well as a series of mutants with defects in the formation of the cuticle such as lacerata ( lcr; affected in the cytochrome P450-dependent enzyme CYP86A8 , likely to be involved in fatty acid hydroxylation of cutin monomers ) [30] , bodyguard ( bdg; impaired in a member of the a/b hydrolase family associated with the organization of the cutin polyester ) [31] or bre1/lacs1 ( a mutant of the long-chain acyl-CoA synthetase2 , LACS2 , involved in the development of the cuticle and essential for its biosynthesis ) [32] . All those mutants display a strong resistance to B . cinerea; this was always associated with an increased cuticular permeability and production of a diffusate endowed with growth-inhibiting activity against B . cinerea [32] . The increased cuticular permeability was proposed to allow the diffusion of toxic compound ( s ) from the cell to the surface or facilitate the passage of MAMPs or DAMPs ( microbe or damage-associated molecular patterns [33] ) from the surface to the inside of the cell , resulting in increased resistance [32] , [34] . Since biochemical alterations of the cuticle were found to increase the resistance of plants to pathogens , physical alterations of the cuticle , such as wounding were tested to see if they could also increase the defense potential of the plant to virulent necrotrophic pathogens . Wounding of Arabidopsis thaliana leaves leads to strong and transient immunity to the virulent pathogen B . cinerea . This resistance is strictly limited to the wound site and is independent of the major plant defense signaling pathways involving salicylic acid ( SA ) , jasmonic acid ( JA ) , and ET [34] . In this work , we have attempted to further our understanding of the molecular events taking place after wounding . A rapid formation of ROS has been observed after wounding and ROS can act as a signal for innate immunity but can also serve as an oxidant for lignification [35] , [36] . This prompted us to carry out observations on the formation of ROS under our conditions of wounding . A strong correlation was observed between ROS formation and resistance to B . cinerea . We discovered that ABA is involved in the regulation of ROS production most likely causing changes in the permeability of the cellular envelope . This led to the finding that an increase in ROS also takes place in plants where cuticular permeability was affected by mutations or simply by a digestive treatment with cutinase . We propose that the cuticle acts as a sensor for pathogens that invade directly through the cell wall , leading to ROS formation whenever the cuticle is degraded . B . cinerea is known to form oxalic acid that can potentially prevent ROS formation . In line with this observation we have also shown that transgenic plants constitutively expressing an oxalic acid-degrading enzyme recovered their ability to produce ROS in response to B . cinerea infection and were resistant to this fungus .
Wounding A . thaliana leaves with forceps as previously described [34] lead to an increase of fluorescence when leaves infiltrated with the 5- ( and-6 ) -carboxy-2 , 7-dichlorodihydrofluorescein diacetate ( DCF-DA ) probe were viewed under the microscope ( Fig . 1B ) . This dye detects a broad range of oxidizing reagents including H2O2 and O2− [37] . A detailed time-course of ROS production was determined in a fluorimeter using wounded leaf discs . Fluorescence appeared within the first minutes after wounding and increased steadily thereafter ( Fig . 1A ) ; unwounded controls also showed a detectable increase , likely the result of wounding inflicted during the preparation of leaf discs . In whole excised leaves , fluorescence started to be detected 2 min after wounding ( Fig . S1 ) . After inoculation with B . cinerea , fluorescence started to accumulate 12 h after inoculation ( Fig . 1B ) and wounded leaves inoculated with B . cinerea showed the same behavior as mock-treated wounded leaves ( Fig . 1B ) . Staining of wounded leaves with diaminobenzidine ( DAB ) or with nitroblue tetrazolium ( NBT ) identified production of H2O2 and O2− respectively , at the wounding site ( Fig . 1C ) . ROS production was seriously diminished with a concomitant increase in fungal growth on leaves that were treated with DPI , an inhibitor of superoxide formation ( Fig . 1D ) . Infiltration of leaves with catalase prevented the coloration of wounding sites with DAB confirming the production of H2O2 upon wounding ( Fig . 1E ) . The luminol assay [38] was used to detect the formation of H2O2 . A strong luminescence became visible at the wound sites ( Fig . 1F ) . Wound-induced ROS and wound-induced resistance ( WIR ) to B . cinerea were still detected in mutants of NADPH oxidase D ( atrboh D ) and F ( atrboh F ) as well as in the double mutant ( atrboh D/F ) meaning that other genes are possibly implicated in ROS production ( Fig . S2 ) . ROS accumulation was still detected after wounding of the triple mutant ( nia1nia2noa1-2 ) that is impaired in the biosynthesis of NO [39] ( Fig . 1G ) . Laser confocal microscopy ( LCM ) was used to determine the localization of ROS in leaves treated with DCF-DA . In wounded leaves the fluorescence mainly localized at chloroplasts in mesophyll cells and to some extent to membranes in mesophyll and epidermal cells ( Fig . 2 ) . We have also tested the effects of H2O2 applied on leaves together with B . cinerea . H2O2 at concentrations as low as 1 nM caused enhanced resistance to fungal infection ( Fig . S3A ) . This concentration is lower than the IC50 value of H2O2 ( determined at 8 . 3 mM ) for inhibition of hyphal growth in vitro ( Fig . S3B ) . Taken together , these results show that wounding leads to a rapid burst of ROS , including H2O2 , which could potentially take part in the resistance of A . thaliana to B . cinerea . We observed that both ROS production and WIR could be strongly impaired when the plants were maintained uncovered for 1 . 5 h at ambient humidity after wounding before the measurements of ROS and resistance ( Fig . 3A and B ) . Thereafter , we will refer to these conditions as dry in contrast to the humid incubation environment in covered trays . In contrast , ROS production and WIR to B . cinerea remained unaffected when , after wounding , plants were kept for 1 . 5 h under high humidity in covered plastic trays ( Fig . 3A and B ) . Maintaining plants under dry conditions after wounding strongly reduced wound-induced callose formation , a typical defense reaction to wounding ( Fig . 3C ) . The strong effects of a dry environment on the suppression of H2O2 production and WIR to B . cinerea made us suspect a possible involvement of ABA in the suppression of ROS . Indeed , mutants blocked in the late steps of ABA biosynthesis , such as aba2 and aba3 , were not blocked in ROS production after wounding under dry conditions and induced WIR in response to B . cinerea ( Fig . 4 ) . Both unwounded aba2 and aba3 mutants showed a marked resistance to B . cinerea accompanied by a faster and more intense ROS production after B . cinerea inoculation compared to WT plants ( Figs . 4B and 5 ) . The kinetics of ROS production was then also tested in unwounded aba mutants and showed that ROS were already released 3 h after exposure to water , PDB medium ( mock treatment ) or B . cinerea infection without any wounding ( Fig . 5 ) . The level of ABA was increased in wounded plants incubated in dry conditions ( Fig . 6A ) . To confirm an increase in the level of ABA , the expression of ABA-dependent genes RAB18 , RDB29 and NCED23 was tested . These genes showed an increase in expression after wounding and mock treatment as well as wounding and B . cinerea in dry conditions that was clearly detectable at 15 min after treatment ( except NCED23 in mock-treated plants ) ( Fig . S4 ) . Changes in ABA levels were further tested using transgenic plants containing a LUC reporter gene under the control of the ABA-specific promoters LTI23 or HB6 . The activity of the reporter gene was mainly observed at the wound site and it was stronger in wounded plants incubated under dry compared to humid conditions ( Fig . 6B ) . Furthermore , exogenous applications of ABA at 100 mM led to a suppression of ROS and WIR in response to B . cinerea ( Fig . 6C ) . Thus , ABA is likely to be involved in the suppression of wound-induced ROS when plants are kept under dry conditions [25] . The resistance to B . cinerea displayed in aba mutants was comparable to that observed in mutants with permeable cuticles such as bdg or lacs2 . 3 [32] . We therefore tested and confirmed that bdg and lacs2 . 3 also produced ROS after wounding ( data not shown ) . Furthermore , both bdg and lacs2 . 3 strongly displayed DCF-DA fluorescence after treatment with water , mock or B . cinerea compared to WT plants ( Fig . 5 ) . The cuticle of bdg and lacs2 . 3 was previously shown to be more permeable and it was postulated that this feature would allow an easier diffusion of elicitors into the cell or an improved passage of antibiotic substances towards the surface [32] . Consequently , we tested if the aba mutants also display alterations in cuticle permeability . This was tested using the cell wall stain toluidine blue applied as droplets on the adaxial side of leaves as previously described [32] . Results showed that aba2 and aba3 mutants strongly stained in blue , as did bdg and lacs2 . 3 compared to WT plants , indicating altered cuticular properties ( Fig . 7A ) . We have ascertained that stomatal density did not interfere with the toluidine blue tests . The density of stomata in WT Col-0 , aba2 , aba3 , bdg and lacs2 . 3 mutants showed some differences that could however not account for the toluidine blue staining observed only in the mutants compared to WT plants ( Fig . S5 ) . Calcofluor staining was also used to visualize permeable cuticles [32] and marked differences were obtained between WT Col-0 and aba2 , aba3 , bdg and lacs2 . 3 mutants ( Fig . 7A ) . Increased efflux of chlorophyll is another measure of cuticular permeability [31] , [40] . When dipped in ethanol , aba2 and aba3 as well as lacs2 . 3 and to a lesser extent bdg mutants released chlorophyll faster than WT Col-0 plants ( Fig . 7B ) thus corroborating the results of the toluidine blue and Calcofluor tests . Cuticular permeability was also compared when wounded plants were incubated under humid or dry conditions . Calcofluor staining was more intense at wounded sites in plants incubated under humid conditions compared to dry conditions indicating a better access of Calcofluor to the cell wall glucans ( Fig . 7C ) . Similarly , chlorophyll leaching proceeded more rapidly when tested on wounded leaves incubated under humid compared to dry conditions ( Fig . 7D ) . Furthermore , the expression of the BDG and LACS genes involved in cuticular biosynthesis were compared in wounded plants incubated under humid or dry conditions . Wounding under dry conditions enhanced the expression of BDG or LACS genes , compared to wounded plants incubated under humid conditions . While BDG was already enhanced within 15 min after wounding and dry conditions , the expression of LACS2 . 3 was clearly higher 30 min after treatment under the same condition ( Fig . 8 ) . In all experiments , expression of both genes after wounding followed by B . cinerea inoculation was not as extensive as the expression after wounding and mock treatment . The composition of the cuticle was altered in wounded plants after incubation in dry conditions compared to humid conditions as was the composition of the cuticle in aba2 and aba3 mutants compared to WT plants ( Fig . S6 ) . Furthermore , exogenous treatment of A . thaliana with ABA decreased the cuticular permeability ( Fig . S7 ) . ABA might therefore be involved in the control of a wound repair mechanism that decreases the permeability of the cuticle . We have further tested the relation between cuticle , permeability , ROS and resistance to B . cinerea after a localized treatment with cutinase ( from Fusarium solani , prepared using heterologous expression in S . cerevisiae ) . Localized application of cutinase led to ROS production visualized with DAB and DCF-DA staining ( Fig . 9A ) as well as to an increase in resistance to B . cinerea ( Fig . 9B ) . These experiments support the hypothesis that plants can perceive and react to the degradation of the cuticle . The cutinase produced by B . cinerea during infection [16] might potentially cause ROS production during the early stages of infection . But the data presented in Figs . 1B and 5 show an increase in fluorescence beyond 12 h after inoculation . B . cinerea was reported to release oxalic acid during infection [41] . Oxalic acid inhibits ROS production in tobacco and soybean cells [42] and might potentially slow down ROS production during the infection . The importance of oxalic acid as a suppressor of ROS was tested using transformed A . thaliana over-expressing an oxalate decarboxylase gene from the basidiomycete Trametes versicolor [43] . By removing oxalic acid released by B . cinerea , we would predict an increase in both ROS formation and resistance . In line with our expectations , transgenic T3 lines that over expressed the OXALATE DECARBOXYLASE gene and exhibited increased oxalate decarboxylase activity showed an increase in resistance to B . cinerea ( Fig . 10A ) . ROS appeared as early as 3 hours post inoculation at sites inoculated with B . cinerea ( Fig . 10B ) in transgenic plants compared to controls . Thus , oxalic acid produced by B . cinerea might help the fungus to avoid the effects of ROS produced during the early steps of infection .
We have previously reported a very marked resistance in A . thaliana to B . cinerea in response to localized wounding . This resistance is based on priming of camalexin synthesis , of the expression GLUTATHIONE S TRANSFERASE 1 ( GST1 ) gene , and MAPK kinase activity [34] . Here , we have followed up these observations and described early events associated with WIR . We have described the production of ROS within 2 minutes at the site of wounding using the fluorescent dye DCF-DA [44] . To check the validity of the DCF-DA dye for ROS detection under our experimental conditions , we have also monitored ROS production using luminol , a method that mainly detects H2O2 [38] and could confirm ROS production after wounding ( Fig . 1F ) . The wound sites also reacted to DAB and NBT staining confirming the formation of H2O2 and O2− ( Fig . 1C ) . Treatment with the NADPH oxidoreductase inhibitor DPI ( Fig . 1D ) or infiltration of leaves with catalase ( Fig . 1E ) before wounding inhibited ROS development , as measured by DCF-DA fluorescence or DAB accumulation , respectively indicating that a substantial part of ROS is O2− and H2O2 . Mutants unable to produce NO still produced ROS after wounding ( Fig . 1G ) , making a contribution of NO to the initial wound-induced burst of ROS unlikely . Observations of plants with the LCM indicated strong fluorescence at the chloroplasts and a weaker one at the cell border after wounding ( Fig . 2 ) . The plastidic origin of some of the ROS might explain in part why ROS were still observed in atrbohD or atrbohF mutants since AtRBOHD or AtRBOHF are localized at the plasma membrane ( Fig . S2 ) . Taken together , our observations indicate a rapid ( within 2 min ) production of ROS after wounding ( Fig . S1 ) . Since we cannot exclude the presence of other ROS besides superoxide and H2O2 , we will use the term ROS to collectively refer to the oxidative species that can be detected after wounding . Our experiments with exogenous applications of H2O2 show that ROS can have both a direct effect against B . cinerea and an indirect effect possibly by activation of defenses ( Fig . S3 ) . These observations are in line with previous studies showing ROS production after wounding [24] , [36] , [45] , or in response to pathogens [26] including B . cinerea [46] . What is the biological importance of ROS production for WIR to B . cinerea ? Firstly , exogenously applied H2O2 can inhibit growth of B . cinerea ( Fig . S3 ) . Secondly treatments with DPI and catalase or incubation under dry conditions abolished ROS , WIR to B . cinerea as well as callose accumulation ( Figs . 1D , E and 3 ) . Thirdly , the absence of ROS formation and resistance to B . cinerea under dry conditions could be rescued in mutants impaired in ABA biosynthesis ( aba2 and aba3 ) and exogenous application of ABA suppressed both ROS and resistance to B . cinerea ( Figs . 4 and 6C ) . Finally , localized treatments with cutinase resulted both in ROS production and increased resistance to B . cinerea on treated sites ( Fig . 9 ) [47] . Taken together , these results support the biological importance of ROS produced in response to wounding for WIR to B . cinerea . Several reports have proposed that ROS and subsequent cell death formed in response to B . cinerea or other necrotrophs might facilitate the infection by the pathogen ( reviewed in [26] ) . This is clearly different from the situation described here where ROS produced after wounding are strongly induced prior to an inoculation and lead to an early induction of defenses such as rapid callose formation ( Figs . 1 and 3 ) . In a previous article , we have shown wound-induced priming of camalexin synthesis , expression of GLUTATHIONE S TRANSFERASE 1 ( GST1 ) and MAPK kinase activity [34] . How do our data agree with the conventional observation that wounding is associated with susceptibility ? Unless wounded plants are maintained under humid conditions WIR is lost . This resolves the apparent paradox , since most of the time plants wounded under natural conditions may not be under conditions of saturating humidity . What is exactly the contribution of ABA ? Our experiments have shown that ABA is implicated in the control of ROS formation in wounded plants that are incubated under dry conditions . Wounding followed by incubation in dry conditions lead to an increase in ABA levels ( Fig . 6A ) . Furthermore , both the expression of ABA-dependent genes ( RAB18 , RDB29 and NCED23 ) [48] and the expression of the ABA reporter gene constructs ATH6: LUC and ATLTI23:LUC [49] were induced ( Figs . S4 and 6B ) . ABA applied on leaves suppressed wound-induced ROS and subsequent resistance to B . cinerea ( Fig . 6C ) . Our data are in agreement with observations made on ABA-deficient sitiens mutants of tomato , where the accumulation of H2O2 was both earlier and stronger than in WT plants after inoculation with B . cinerea [50] . Our results and those of Asselbergh et al . ( 2007 ) [50] suggest a negative control of ABA on ROS formation and resistance . Several reports show a link between ABA and increased susceptibility to pathogens that was mostly explained by antagonistic interactions of ABA with defense signaling controlled by SA , JA or ET [51] . How ABA prevents wound-induced ROS accumulation in A . thaliana remains a study to be carried out on its own . This will be interesting , since ABA controls stomatal closure via ROS production [52] . But the action of ABA further unveiled when we observed that aba mutants had a resistant phenotype reminiscent of plants affected in cuticle integrity such as CUTE or the cuticle mutants bdg and lacs2 . 3 that are immune to B . cinerea [32] , [47] . For those reasons , ROS production was followed in bdg or lacs2 . 3 . Indeed , these mutants displayed a strong DCF-DA fluorescence even after water or mock treatments or inoculation with B . cinerea ( Fig . 5 ) . The bdg or lacs2 . 3 mutants were previously shown to have an increased cuticular permeability compared to WT plants ( Fig . 7A , B and [32] , [53] ) . Accordingly , we have characterized cuticular properties in aba mutants and observed a higher permeability in aba2 and aba3 mutants than in WT plants ( Fig . 7A and B ) . Cuticular permeability measured by Calcofluor white or chlorophyll efflux was also decreased after incubation of wounded plants under dry conditions compared to plants left at high humidity despite the opening caused by the wound ( Fig . 7C and D ) . In addition , a change was observed in the composition of aliphatic cuticle monomers [30] , [31] and in the expression of the LACS and BDG genes involved in cuticle biosynthesis in wounded plants incubated under dry compared to humid conditions ( Fig . 8 ) . Incubation of wounded plants for 1 . 5 h under dry conditions was sufficient to increase detectable changes in of 16:0 and 18:3 cuticle monomers ( Fig . S6 ) . Thus , the changes cuticle properties in response to the environmental conditions are accompanied by changes in aliphatic monomers of the cuticle . The aba2 and aba3 mutants also displayed a different composition in aliphatic cuticle monomers compared to Col-0 WT plants ( Fig . S6 ) . However , it is not yet feasible to link changes in structural components of the cuticle to functions such as cuticular permeability . Exogenous application of ABA to WT or aba2 , aba 3 lacs2 . 3 and bdg mutants decreased the cuticle permeability , further supporting a role for ABA in this process ( Fig . S7 ) . Curvers et al . ( 2010 ) [54] reported recently that tomato mutants impaired in ABA biosynthesis show enhanced cuticular permeability and resistance to B . cinerea . Our results in A . thaliana are in full agreement with an effect of ABA on the cuticle in tomato . It becomes now interesting to determine how ABA exerts its effects on the structure of the cuticle . How could an increase in cuticular permeability affect the formation of ROS ? The strong resistance of cuticular mutants to B . cinerea was explained by the facilitated diffusion of potentially antibiotic compounds produced by the plant and/or of elicitors from the medium/pathogen through the permeable cuticle surface [32] , [53] . The perception of elicitors , including breakdown products of plant cuticles , has been previously described to produce ROS [22]–[24] , [26] . The fact that cuticular mutants lacs2 . 3 , bdg and aba2 , aba3 produced ROS even when exposed to water or mock solution alone ( Fig . 5 ) might be explained by the perception of elicitors present at the surface of the non-sterile leaves that dissolve in the water or in the mock solution and diffuse through the cuticle to the cell where they are perceived . The hypothesis that diffusion is facilitated through a permeabilized cuticle is further supported by the effect of cutinase treatments . This enzyme was already shown to degrade cutin and increase the permeability of the cuticle and resistance to B . cinerea when expressed constitutively in A . thaliana [29] , [47] . By digesting the cuticle , this enzyme generates cutin monomers , increases the permeability of the cuticle and therefore improves diffusion of breakdown products that , together with other possible elicitors present at the leaf surface , would subsequently be recognized and lead to ROS formation and resistance ( Fig . 9 ) . Recognition of cutin monomers as well as wax components has also been described to induce the production of H2O2 in abraded epicotyls of cucumber [24] . How can these results be tied into a comprehensive model ? When the cuticle is intact , it functions in protection against water loss , irradiation and xenobiotics . When it is permeabilized upon degradation by either enzymes secreted during pathogenesis or acted upon by mechanical action , elicitors might have a facilitated access to the cell , will be recognized and eventually lead to a rapid release of ROS and subsequent defense reactions . The model in Fig . 11 explains why plants are not continuously in an induced state , despite the existence of an extensive microbial flora at the leaf surface . As long as the cuticle prevents passage of elicitors , no induction of defenses takes place , illustrating economic energy management by the plant . Why does a virulent necrotrophic pathogen like B . cinerea that degrades the cuticle not lead to a rapid burst of ROS ? Virulent pathogens produce suppressors or effectors that can interfere with plant defenses [55] . Oxalic acid produced by B . cinerea [41] can suppress the formation of ROS [42] and might thus protect the fungus from their effects . We have previously shown that removal of oxalic acid produced by B . cinerea enhances the protection of A . thaliana [56] . Results obtained here with plants over expressing an oxalate decarboxylase extend these findings . In oxdec plants ROS appeared early during infection and growth of B . cinerea was decreased ( Fig . 10 ) . Similar observations were published for the virulent oxalate-producing fungus Sclerotinia sclerotiorum and B . cinerea [43] , [57] . In addition , B . cinerea was shown to produce ABA [58] , [59] that might also be possibly involved in repressing ROS production during infection . B . cinerea can also induce ABA production in the plant [60] . The AP-1 transcription factor Bap1 plays a pivotal role in ROS detoxification of B . cinerea in vitro . But Bap1 was not found to be essential for pathogenesis and the role of an oxidative burst was questioned [61] . Here we show that removal of oxalic acid can indeed restore early ROS production by the plant and impair pathogenicity of B . cinerea , thus confirming previous results on the importance of oxalic acid for B . cinerea [56] , [57] . In conclusion , we propose a model whereby the cuticle is part of a sensing device besides its passive protective role of aerial plant surfaces . Without modification of the diffusive properties of the cuticle , defenses are not induced . Modifications of the surface with subsequent increased permeability will allow for a better passage for molecular determinants that can be recognized by the plant and lead to the activation of defenses . This mechanism is doubled up by another mechanism provided by the wall itself , that , when exposed to the appropriate enzymes will breakdown to damage associated determinants ( DAMPs ) that are also recognized and initiate defenses . Virulent pathogens have evolved mechanisms to interfere with such mechanisms and data presented here support these findings . Future work should now be directed at the molecular mechanisms that lead to a rapid generation of ROS . It will be interesting to determine if they overlap with a similar responses observed when plant react to other elicitors that lead to ROS formation such as flagellin .
Arabidopsis thaliana seeds were grown on a pasteurized soil mix of humus and Perlite ( 3∶1 ) . Seeds were kept at 4°C for two days and then transferred to the growth chamber . Plants were grown in a 12 h light/12 h dark cycle with 60–70% of relative humidity , with a day temperature of 20–22°C and a night temperature of 16–18°C . WT plants were obtained from the Nottingham Arabidopsis Stock Center ( Nottingham , UK ) . The Arabidopsis mutant referred to as aba2 was aba2–1 and aba3 was aba3–1 [62] . The lacs2–3 , bdg2 and the nia1nia2noa1–2 mutants were previously described [31] , [32] , [39] . B . cinerea strains BMM , provided by Brigitte Mauch-Mani ( University of Neuchâtel , Switzerland ) , were grown on Difco ( Becton Dickinson , http://www . bd . com ) potato dextrose agar 39 g l−1 . Spores were harvested in water and filtered through glass wool to remove hyphae . Spores were diluted in ¼ strength Difco potato dextrose broth ( PDB ) at 6 g l−1 for inoculation . Droplets of 6 µl spore suspension ( 5×104 spores ml−1 ) were deposited on leaves of 4-week-old plants for quantification of lesions size ( mm ) after 3 days . Spores ( 2×105 spores ml−1 ) were also sprayed on whole plants for RT-PCR experiments . The inoculated plants were kept under high humidity in covered trays . Control plants were mock inoculated with ¼ strength PDB solution . Leaves were wounded by gently pressing the lamina with a laboratory forceps . For wounding of entire leaves , the pressing was carried out on both sides of the main vein . Wounded leaves were incubated in covered trays at high humidity ( referred to as humid conditions ) ; in some cases the trays were left uncovered after wounding ( referred to as dry conditions ) under the same laboratory conditions . Inoculation with B . cinerea was performed within 10 min after wounding , by placing a droplet of spores on the wound site . Fungal structures and dead plant cells were stained by boiling inoculated leaves for 5 min in a solution of alcoholic lactophenol trypan blue . Stained leaves were extensively cleared in chloral hydrate ( 2 . 5 g ml−1 ) at room temperature by gentle shaking , and then observed using a Leica DMR microscope with bright-field settings . ROS were detected using the fluorescent probe 5- ( and 6 ) -carboxy-2′ , 7′-dichloro dihydrofluorescein diacetate ( DCF-DA ) ( Sigma-Aldrich , www . sigmaaldrich . com ) . Wounded or unwounded leaves were vacuum-infiltrated ( 3×3 min ) in 60 µM of DCF-DA in a standard medium ( 1 mM KCl , 1 mM MgCl2 , 1 mM CaCl2 , 5 mM 2-morpholinoethanesulfonic acid adjusted to pH 6 . 1 with NaOH ) [44] . Leaves were then rapidly rinsed in DCF-DA medium and observed using a Leica DMR epifluorescence microscope with a GFP filter set ( excitation 480/40 nm , emission 527/30 nm ) ( Leica , www . leica . com ) . Microscope images were saved as TIFF files and processed for densitometric quantification with Image J version 1 . 44 ( NIH ) . Software settings were kept the same for every image analyzed; the surface of each analyzed picture was the same ( 2 . 278 mm2 ) . One representative image of the fluorescent leaf surface was placed above each histogram as a visual illustration . Quantification of ROS using DCF-DA was also performed on wounded or unwounded leaf discs of 5 mm incubated in 60 µM of DCF-DA in a 96-well plate ( Sarstedt , www . sarstedt . com ) . After vacuum-infiltration ( 3×3 min ) , ROS were determined using a FL×800 microplate fluorescence reader with a excitation filter 485/20 nm and an emission filter 528/40 nm ( Bio-Tek instruments , www . biotek . com ) . Accumulation of O2− and H2O2 in leaves was determined using nitroblue tetrazolium ( NBT ) staining [63] and 3 , 3′-diaminobenzidine ( DAB ) staining [64] respectively . The destained leaves were observed using a Leica DMR microscope with bright-field settings . The H2O2 accumulation was also determined using the luminol test [65] . A solution containing 50 µl of 0 . 5 mM luminol ( 3-aminophtalhydrazide , Sigma-Aldrich ) in 0 . 2 N NH3 , pH 9 . 5 added to 0 . 8 ml of 0 . 2 N NH3 , pH 9 . 5 and 100 µl of 0 . 5 mM K3Fe ( CN ) 6 in 0 . 2 N NH3 , pH 9 . 5 was added to leaf discs of 8 mm in 24-well plates ( Corning incorporated , www . corning . com ) and luminescence was measured immediately using CCD camera ( Princeton Instrument Versarray system , www . princetoninstruments . com ) equipped with a Sigma Aspherical objective ( www . sigma-foto . de ) in a dark box . The pictures were analyzed with an Imaging System . Arabidopsis reporter lines consisting of either a pAtHB6 or pLTI65 promoter fragment fused to the LUC gene were generously given by Prof . Erwin Grill [66] . For imaging of LUC activity , plants were sprayed with a solution of 1 mM luciferin ( Applichem , www . applichem . com ) in 10 mM MES , pH 7 . 0 , 0 . 01% Tween 80 . Ten min after luciferin spraying , light emission was detected using an intensified CCD camera ( Princeton Instrument Versarray system , www . princetoninstruments . com ) equipped with a Sigma Aspherical objective ( www . sigma-foto . de ) in a dark box . The pictures were analyzed with the MetaVue Imaging System ( www . biovis . com/metavue . htm ) . Twenty-four hours post infiltration , leaves were harvested and distained in 3∶1 ethanol: lactic acid , previously diluted in 1:2 ethanol . The solution was changed several times until the chlorophyll had totally disappeared . Translucent leaves were progressively re-hydrated in 70% ethanol for about 2 hours and in 50% ethanol for 2 hours . Leaves were left in water and gently shaken overnight . Leaves were then incubated for 24 hours in 150 mM K2HPO4 ( pH 9 . 5 ) containing 0 . 01% aniline blue . Stained material was mounted on glass slides in 50% glycerol and examined under UV light with a LEICA DMR fluorescence microscope . The callose deposition was determined by counting the pixels using the Image J 1 . 44 software ( NIH ) . RNA was prepared using the Trizol reagent containing 38% saturated phenol , 0 . 8 M guanidine thiocyanate , 0 . 4 M ammonium thiocyanate , 0 . 1 M sodium acetate and 5% glycerol . RNA ( 1 µg ) was then retrotranscribed into cDNA ( Omniscript RT kit , Qiagen , www . qiagen . com ) . RT-PCR was performed using Sensimix SYBR Green Kit ( Bioline , www . bioline . com ) . Gene expression values were normalized to expression of the plant gene At4g26410 , previously described as a stable reference gene [67] . The primers used were rab18fw 5′- AACATGGCGTCTTACCAGAA; rab18rev 5′-AGTTCCAAAGCCTTCAGTCC; rd29bfw 5′-GAATCAAAAGCTGGGATGGA; rd29brev 5′-TGCTCTGTGTAGGTGCTTGG; nced23fw 5′-ATTGGCTATGTCGGAGGATG; nced23rev 5′-CGACGTCCGGTGATTTAGTT; lacs2-3fw 5′-GTGCCGAGAGGAGAGATTTG; lacs2-3rev 5′-CGAGGTTTTCAACAGCAACA; bdgfw 5′-TTCTTGGCTTTCCTCTTCCA; bdgrev 5′- CCATAACCCAACAGGTCCAC . Leaf discs of 8 mm were floated in 24-well plates ( Corning incorporated , www . corning . com ) filled with 1 . 5 ml in each well of either 50 µM DPI ( Sigma-Aldrich ) , 100 µM ABA ( Sigma-Aldrich ) in 0 . 05% EtOH and distilled water or EtOH 0 . 05% as controls for 24 h before wounding . Catalase ( 300 U ml−1 catalase; Sigma-Aldrich ) was infiltrated into the leaves prior to wounding . After wounding , either ROS were visualized on the discs with DCF-DA or DAB or discs were floated on distilled water for 24 h and inoculated with B . cinerea and infection was determined after 3 days . The effect of ABA on cuticle permeability was determined after spraying leaves with 100 µM ABA in 0 . 05% EtOH followed by a 24 h incubation period under humid conditions . Eight mL droplets of purified preparation of cutinase ( 5 mg l−1 ) from F . oxysporum [68] or 10 mM Na-acetate pH 5 . 2 in controls were applied on the leaf surface . After 72 h incubation under moist conditions , the droplets were removed and leaves were stained with DCF-DA or DAB to detect ROS . Alternatively , the droplet was rinsed off the leaf and replaced by a droplet of spore suspension of B . cinerea and incubated as described above . Leaf material ( ca 500 mg ) was frozen in liquid nitrogen and collected in 2 ml Eppendorf tubes , and then homogenized ( twice ) with 1000 µl of 70°C-warm extraction buffer ( water/propane-1-ol/HCl : 1/2/0 . 005 , v/v ) without thawing . The sample was transferred to a glass tube and 200 ng of the internal standard abscisic acid-d6 ( Santa Cruz Biotechnology , www . scbt . com ) and 2 ml of dichloromethane were added . The sample was then mixed 15 s with a vortex and centrifuged 1 min at 14 , 000 g . The lower organic phase was transferred to a new glass tube and dried by the addition of anhydrous Na2SO4 . Then , carboxylic acids including ABA were methylated to their corresponding methyl esters at room temperature for 30 min after the addition of 10 µl of 2 M bis-trimethylsilyldiazomethane ( Sigma-Aldrich ) and 100 µl of MeOH . Methylation was stopped by the addition of 10 µl of 2 M acetic acid during 30 min at room temperature . Extraction of the vapor phase was performed using a VOC column conditioned with 3×1 ml of dichlororomethane . The VOC column and a nitrogen needle were fixed on the screw cap of the tube . The solvent was evaporated under a nitrogen stream at 70°C and heated for 2 min at 200°C . The VOC column was eluted with 1 ml of dichloromethane in a new glass tube . The eluate was evaporated and then dissolved in 20 µl of hexane before injecting 3 µl on a capillary column HP1 ( 25 m×0 . 25 mm ) GC column ( Agilent , www . agilent . com ) fitted to a Hewlett Packard 5980 GC coupled to a 5970 mass specific detector . The methyl esters of ABA and ABA-d6 were detected and quantified by selective ion monitoring at m/z 190 and 194 respectively . The amount of ABA ( measured as methyl ABA ) was calculated by reference to the amount of internal standard . The results are expressed in ng mg−1 fresh weight of plant tissue . The surface of 15 to 20 leaves of 4 weeks-old A . thaliana Col-0 was first determined as described previously [69] . Then the leaves were extracted with chloroform:methanol ( 1∶1; v/v ) and dried before depolymerization . The samples were depolymerized using transesterification with 2 ml BF3 ( Fluka , Sigma-Aldrich ) for 12 h at 75°C . After addition of 2 ml saturated NaCl/H2O and 20 µg dotriacontane as internal standard , aliphatic monomers were extracted 3 times with 1 ml of chloroform . The combined organic phase was evaporated in a stream of nitrogen to a volume of ∼100 µl . All samples were treated with bis- ( N , N , -trimethylsilyl ) -tri-fluoroacetamide ( BSTFA; Macherey-Nagel , www . mn-net . com ) for 40 min at 70°C to convert free hydroxyl and carboxyl groups into their corresponding trimethylsilyl ( TMS ) derivatives . Remaining solvent and derivatization reagents were removed under a stream of N2 and the samples were resolubilized in 100 ml dichloromethane prior to vapour phase extraction . Monomers were identified on the basis of their electron-impact MS spectra ( 70 eV , m/z 50–700 ) on a HP 6890 GC system coupled to an HP 5973 mass-selective detector ( USA ) . The depolymerisation products were separated by on a capillary column ( ZB-AAA , 10 m , 0 . 25 mm , Zebron , Phenomenex , www . phenomenex . com ) by injection at 50°C , 2 min at 50°C , 5 °C min−1 to 225°C , 1 min at 225°C , 20°C min−1 to 310°C , 10 min at 310°C . The results are expressed in g cm−2 leaf tissue . Chlorophyll extraction and quantification was performed according to the protocol of Sieber et al . 2009 [29] . Leaves were cut at the petiole , weighed and immersed in 30 ml of 80% ethanol . Chlorophyll was extracted in the dark at room temperature with gentle agitation . Aliquots were removed at 2 , 5 , 10 , 20 , 30 and 40 min after immersion . After ABA treatment , aliquots were removed at 40 and 60 min after immersion in ethanol . The chlorophyll content was determined by measuring absorbance at 664 and 647 nm and the micromolar concentration of total chlorophyll per gram of fresh weight of tissue was calculated from the following equation: ( 7 . 93× ( A664 nm ) + 19 . 53× ( A647 nm ) ) g−1 fresh weight . The toluidine blue test was carried out by placing 6 ml droplets of a 0 . 025% toluidine blue solution in ¼ PDB placed on the leaf surface . After 2 h incubation leaves were washed gently with distilled water to remove excess of the toluidine blue solution from leaves . For staining with Calcofluor white , leaves were bleached in absolute ethanol overnight , equilibrated in 0 . 2 M NaPO4 ( pH 9 ) for 1 h , and incubated for 1 min in 0 . 5% Calcofluor white in 0 . 2 M NaPO4 ( pH 9 ) . Leaves were rinsed in NaPO4 buffer to remove excess of Calcofluor white and viewed under UV light on a GelDoc 2000 system ( Biorad , www . biorad . com ) . The gene used to transform A . thaliana plants is the OXALATE DECARBOXYLASE from Trametes versicolor ( TOXDEC , Genbank accession number AY370675 ) . The gene was kindly provided by Andreas Walz ( Institute for Phytomedicine , University of Hohenheim ) as cDNA cloned into the transformation vector pBI 101 ( p221-TOXDC ) . A . thaliana Col-0 plants were transformed using Agrobacterium tumefaciens and the flower dip method [70] . Resulting seeds were collected and grown on selective medium; the expression of the gene was determined respectively by crude PCR and Northern Blot . The activity of oxalate decarboxylase was measured in the T3 generation along with resistance to B . cinerea strain BMM . Oxalate decarboxylase activity was measured as described [71] with the following modifications . Leaves ( 100 mg ) were homogenized using a Polytron ( Kinematica , www . kinematica-inc . com ) in 1 ml extraction buffer containing 50 mM potassium phosphate buffer pH 7 . 5 , 1 mM EDTA , 1 mM phenylmethylsulphonyl fluoride and 5 mM sodium ascorbate; 200 µl of this extract was added to 10 µl oxalic acid ( 1 M , pH 6 . 2 ) and incubated 3 h at 37°C . After centrifugation during 5 min , 12 µl of a 30% sodium acetate solution ( w/v ) and 300 µl reagent solution containing 0 . 5 g citric acid monohydrate , 10 g acetic acetamide in 100 ml isopropanol were added to 150 µl extract . The solution added to 1 ml of acetic acid anhydride was incubated 40 min at 50°C; the intensity of the resulting pink color reflects oxalate decarboxylase activity . Kruskal-Wallis one way analysis of variance ( ANOVA ) on ranks followed by a Dunn's test was performed using SigmaPlot version 11 . 1 software ( Systat Software , San Jose , CA ) . Different letters above each bar represent statistically significant differences ( Dunn's test; P<0 . 05 ) .
|
This study provides an explanation for the strong resistance to B . cinerea observed in wounded plants or plants with cuticular defects . We have observed that a production of ROS and a permeable cuticle is common to all these situations . ROS , that include hydrogen peroxide , are known inducers of resistance and can also act directly against the invading fungus . Degradation of the cuticle by exposure to cutinase also results in the production of ROS and resistance . These observations lead to a model where the cuticle plays a central role as a barrier against water-soluble elicitors from the surface . Under normal circumstances , the cuticle does not allow the passage of elicitors and no responses are induced . Under conditions where the cuticular barrier is broken , ROS and resistance are induced . This illustrates why plants that are in fact permanently exposed to potential elicitors do not constantly induce immune responses: this only takes place once the cuticle has been permeabilized , for example after an infection with a pathogen . This study also demonstrates how a cuticle-degrading pathogen avoids the generation of ROS by producing an effector that interferes with ROS production . Removal of this effector restores both ROS and resistance .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"plant",
"science",
"plant",
"biology",
"plant",
"pathology",
"biology"
] |
2011
|
A Permeable Cuticle Is Associated with the Release of Reactive Oxygen Species and Induction of Innate Immunity
|
Global gene expression data combined with bioinformatic analysis provides strong evidence that mammalian miRNAs mediate repression of gene expression primarily through binding sites within the 3′ untranslated region ( UTR ) . Using RNA induced silencing complex immunoprecipitation ( RISC-IP ) techniques we have identified multiple cellular targets for a human cytomegalovirus ( HCMV ) miRNA , miR-US25-1 . Strikingly , this miRNA binds target sites primarily within 5′UTRs , mediating significant reduction in gene expression . Intriguingly , many of the genes targeted by miR-US25-1 are associated with cell cycle control , including cyclin E2 , BRCC3 , EID1 , MAPRE2 , and CD147 , suggesting that miR-US25-1 is targeting genes within a related pathway . Deletion of miR-US25-1 from HCMV results in over expression of cyclin E2 in the context of viral infection . Our studies demonstrate that a viral miRNA mediates translational repression of multiple cellular genes by targeting mRNA 5′UTRs .
The recent discovery of a new class of regulatory genes known as microRNAs ( miRNAs ) has resulted in a paradigm shift in gene regulation research . miRNAs are small single-stranded RNA species of approximately 20–24 bases in length that regulate gene expression through post transcriptional mechanisms [1] . Expression of miRNAs is thought to be ubiquitous among multicellular eukaryotes [2] . In addition to eukaryotic miRNAs , more than 100 viral miRNAs have been identified , almost all of which are expressed by herpesviruses [3] . Targets for the majority of viral miRNAs are currently unknown due to the difficulty involved in identifying novel target transcripts . This remains one of the major challenges in elucidating the function of miRNAs . However , recent reports have begun to elucidate the various roles of viral miRNAs . These include blocking apoptosis , immune evasion and regulation of viral replication through targeting of both cellular and viral gene expression [4] , [5] . HCMV , a member of the beta-herpesvirus sub family , encodes at least 11 miRNAs [6]–[8] . Previously , we demonstrated that the HCMV encoded miRNA , miR-UL112-1 , targets a number of the virus's own genes , including the immediate early transactivator IE72 which is essential for driving acute replication of HCMV [9] . In addition , miR-UL112-1 targets the cellular gene MICB , resulting in protection against recognition by natural killer cells [10] . miR-UL112-1 may therefore play an important role in establishing and maintaining viral latency and persistence through regulation of viral gene expression and subversion of host antiviral pathways . Indeed , consensus is beginning to emerge that herpesvirus miRNAs may , in general , be important in establishing and maintaining viral persistence [9] , [11]–[13] . Current studies indicate that miRNA targeting in mammalian cells occurs predominantly through binding to sequences within 3′UTRs [14]–[16] . The reason for this bias is unclear , although a recent study demonstrated that miRNA-mediated repression of a reporter construct was less efficient when the target site was placed in the ORF compared to the 3′UTR [17] . In contrast , inhibition of gene expression through targeting the 5′UTR has been demonstrated , at least in the context of an artificial reporter construct , indicating that miRNA targeting of 5′UTRs is possible [18] . However , statistical analysis of conserved miRNA target sequences and global biochemical screens have demonstrated that mammalian miRNA target sites rarely occur within 5′UTRs [14]–[16] , [19] . The role of 5′UTRs in miRNA regulation is further complicated by a study that found miR-10a induces , rather than inhibits , protein expression through binding to 5′UTRs of cellular transcripts [20] . In addition , binding of the liver specific miRNA , miR-122 to the 5′UTR of Hepatitis C ( HCV ) genome is required for virus replication [21] . These studies suggest a model in which binding to the 5′UTR results in mechanistic effects divergent from 3′UTR binding . Here , we identify cellular transcript targets of one of the most highly expressed HCMV encoded miRNAs , miR-US25-1 , using a recently developed biochemical approach called RISC immunoprecipitation . Strikingly , the majority of identified transcripts contained miR-US25-1 target sites within the 5′UTR rather than the 3′UTR . The target transcripts include a number of genes associated with cell cycle control , including cyclin E2 , as well as histone proteins , suggesting that miR-US25-1 is targeting functionally related genes . Crucially , we demonstrate that targeting of cyclin E2 by miR-US25-1 occurs in the context of HCMV infection and results in inhibition of cyclin E2 protein expression .
Identifying target transcripts is still one of the main challenges in determining the functional role of miRNAs . Although bioinformatic strategies have proven useful , they are limited by high false positive rates . Due to a lack of effective approaches , target transcripts for the majority of viral miRNAs , and miRNAs in general remain unknown . Alternative experimental approaches for the identification of miRNA targets that do not completely rely on bioinformatic predictions are therefore desirable . Recently , approaches for miRNA target identification have been devised that rely on immunoprecipitation of the RISC complex and the associated target transcripts [22] , [23] . As part of the RISC complex , miRNAs bind to target transcripts and form stable interactions . Using a HEK293 cell line that expresses a tagged component of RISC ( Argonaute 2 ( Ago2 ) tagged with myc epitope ) these complexes can be immunoprecipitated and targeted transcripts identified by microarray analysis ( Figure 1a ) . This approach was used to identify cellular targets of one of the most highly expressed HCMV miRNAs , miR-US25-1 . The Ago2 tagged cell line was transiently transfected with a construct encoding the full-length pre-miRNA of miR-US25-1 , under the control of the human U6 polymerase III promoter . Three days post transfection cells were harvested and RISC complexes were immunoprecipitated with a myc-epitope specific antibody . mRNA levels within the immunoprecipitated complexes as well in whole cell lysates were quantified by microarray analysis . As previously described [22] , [23] , association of a specific mRNA with the RISC complex is represented by quantitative enrichment of the mRNA in the immunoprecipitated fraction relative to the total ( whole cell ) fraction . In order to determine which targets are specifically associated with miR-US25-1 , fold enrichment of transcripts immunoprecipitated from miR-US25-1 transfected cells was compared to values from cells transfected with a vector expressing a negative control miRNA . miR-US25-1 specific targets are only expected to be enriched in cells expressing miR-US25-1 . Transcripts were then ranked according to the level of enrichment with the highest enriched transcripts considered potential targets of miR-US25-1 ( Table S1 ) . Over all , the results indicate that miR-US25-1 RISC complexes associated with a relatively small population of transcripts ( Figure 1b and c ) and fold changes were skewed towards enrichment as would be expected if miR-US25-1 were targeting a specific population of genes . The majority of transcripts were not enriched , showing enrichment levels close to 1 ( Figure 1b ) . Thirty-six transcripts showed greater than 2-fold enrichment , while only 1 transcript was reduced by more than 2-fold . The highest level of enrichment was 6 . 5 fold . To increase confidence in target identification , a second , parallel approach was used in which a biotinylated synthetic mimic of miR-US25-1 was transfected into HEK293 cells and miR-US25-1 specific miRNA protein complexes ( miRNPs ) were isolated using streptavidin bead pull downs ( Figure 2a and Table S2 ) . In contrast to the previous approach this should only pull down direct targets of miR-US25-1 . Again , most genes showed little or no enrichment with a relatively small population of transcripts showing exponential increase in enrichment ( Figure 2b and c ) . However , the levels of enrichment were much higher following biotin isolation , reaching a maximum of 23 . 6 fold . In this case fold changes did not skew towards enrichment as miR-US25-1 was compared to a second HCMV miRNA , miR-US5-2 , rather than a negative control . Transcripts showing a negative enrichment ratio , likely represent targets of miR-US5-2 . The two data sets were compared to determine whether the same genes were identified by both RISC pull down approaches . As the enrichment levels in the biotin approach were higher than those found with the myc-Ago2 approach , averaging the enrichment levels would result in bias towards the biotin data set . To avoid this bias , the data sets were combined using rank sum analysis . Transcripts were assigned a rank based on the comparative level of enrichment ( highest enriched = rank 1 , lowest = rank 24526 ) then the average rank between the myc-Ago2 approach and the biotin approach was calculated for each gene . Although differences existed in the rankings of the two data sets ( for example TRIM28 was ranked 1st in the c-myc approach , but 188th in the biotin approach ) a population of transcripts were enriched by both approaches . Fifteen of the top 20 genes showed greater than 2 fold enrichment by both approaches , giving high confidence that these transcripts were likely targets of miR-US25-1 . Table 1 shows the top 20 ranked genes by rank sum analysis including a summary description of their function and the enrichment levels by each approach . A number of these targets are involved in cell cycle control ( cyclin E2 , BRCC3 , PSMA4 and EID1 [24]–[27] ) and tumor progression ( ASRGL1 , CD147 , MAPRE2 and ASPSCR1 [28]–[31] ) , while three of the targets encode histone genes ( LOC440093 , H3F3B , HIST2H4A ) , indicating that miR-US25-1 targets functionally related genes . If the observed enrichment were due to direct targeting by miR-US25-1 , the identified population of transcripts would be expected to contain binding sites for the miRNA . Binding of the 5′ end of the miRNA , specifically nucleotides 1–8 , known as the seed sequence , are thought to be particularly important [1] , [32] , [33] . Therefore , the transcripts in the database were searched for seed sequence matches complimentary to nucleotides , 1–7 , 2–8 and 1–8 of miR-US25-1 . The number of seed matches within the top 50 enriched transcripts from Table S3 was then compared to the number of matches within the rest of the gene list , thereby determining whether genes highly enriched , were more likely to contain predicted miR-US25-1 target sites . Strikingly , miR-US25-1 seed matches were significantly overrepresented within the 5′UTRs of enriched genes ( Figure 3 and Table 2 ) . Twelve of the top 20 genes shown in Table 1 contained at least one 7 base seed match within the 5′UTR . In addition , a further 3 genes contain target sites within the 5′UTRs that do not strictly adhere to Watson-Crick base pairing within the seed region . One gene , ATP6V0C , contained a 7 base target within the coding region , 100 bases down stream of the 5′UTR ( Table 1 and Figure S1 ) . Crucially the number of target sites within the top 50 genes increased from 19 and 14 with the biotin and c-myc approach individually , to 24 target sites in the combined data set , providing additional evidence that the combined approach provides a more robust method of identifying target transcripts . To confirm that enrichment of identified transcripts was due to binding of miR-US25-1 to the 5′UTR , the 5′UTRs and ∼500 bases of upstream genomic sequence of two of the top target genes , cyclin E2 and H3F3B , were cloned in front of a luciferase reporter construct ( Figure 4a and b ) . These constructs were cotransfected into c-myc tagged Ago2 expressing 293 cells with plasmids expressing either miR-US25-1 or the negative control plasmid . RISC-IP analysis was conducted as described above , with levels of luciferase transcript measured using specific RT-PCR primers to the coding region of the reporter gene . Figure 4c shows miR-US25-1 expression resulted in enrichment of luciferase transcript , indicating that the 5′UTRs are indeed sufficient for miR-US25-1 binding . Deletion of the identified seed sequence targets from the 5′UTRs resulted in a loss of enrichment , confirming that the 5′UTR sequences are sufficient and that the target sites are necessary for miR-US25-1 binding . To determine the effect on gene expression of miR-US25-1 binding on the identified 5′UTR targets , luciferase assays were conducted using the 5′UTR constructs described above . In each case expression of miR-US25-1 resulted in a significant reduction in luciferase activity and protein levels ( Figure 4d ) . miR-US25-1 regulation was dependent on the cloned 5′UTRs and the seed target sites as deletion of these target sites rescued expression of luciferase . Expression of miR-US25-1 appears to have resulted in a greater reduction of luciferase protein as determined by western blot , compared to luciferase activity . We speculate that the reduction in luciferase activity is not linearly reflecting the reduction in actually protein levels , possible due to enzymatic nature of the luciferase assay . Direct measurement of luciferase protein may therefore be a more sensitive measure of miRNA regulation . Although these results confirm that miR-US25-1 can bind to a specific population of cellular transcripts , it is important to determine whether these genes are targeted in the context of a viral infection . HEK293 cells are not permissive to HCMV and have a different gene expression profile than cells that are HCMV permissive . To enable RISC-IP analysis of permissive cell lines , a direct antibody to Ago2 was generated using a peptide from the N-terminus of the protein . This antibody was shown to efficiently recognize endogenous Ago2 ( Figure S2a and b ) . RISC complexes were immunoprecipitated from either uninfected human primary fibroblast cells or cells infected with HCMV . The associated RNA was isolated and subjected to RT-PCR analysis using primers specific to the top target cyclin E2 and TRIM28 . Although TRIM28 was not in the top 20 targets shown in Table 2 , it was the top target following immunoprecipitation of tagged Ago2 from cells transfected with the pSIREN construct ( Table S1 ) . As shown in Figure 5a , cyclin E2 and TRIM28 were effectively enriched from cells infected with HCMV compared to the uninfected control cells . In addition , immunoprecipitation using a pre-bleed control serum , which is not expected to pull down Ago2 , did not result in enrichment , indicating that the effect was specifically due to association with RISC complexes . To determine the specific effects of miR-US25-1 on the expression of target proteins in the context of viral infection , miR-US25-1 pre-miRNA coding region was deleted from the virus . As shown in Figure 5b , successful disruption of miR-US25-1 expression was confirmed by RT-PCR analysis . Cells infected with wild type HCMV express high levels of both miR-US25-1 and miR-UL112-1 , whereas miR-US25-1 levels were below background in cells infected with the knockout virus . miR-UL112-1 levels were equivalent to wild type levels indicating efficient infection with the miR-US25-1 knockout virus . Low ( MOI 0 . 5 – Figure 5c ) and high ( MOI 10 – supplementary Figure S2c ) multiplicity growth curve analysis show the knockout virus was able to replicate with wild type kinetics in human primary fibroblast cells . The effects of virally expressed miR-US25-1 on two top targets , cyclin E2 and TRIM28 , were determined by western blot analysis . As cyclin E protein levels are regulated throughout the cell cycle , various serum conditions were used to look at the effects of virus infection during normal cycling populations , populations arrested by serum starvation and cells induced from a resting state using replacement of serum . Western blot analysis indicates that serum starvation effectively repressed cyclin E2 expression as expected and serum rescue resulted in resumption of cyclin E2 expression . Furthermore , as is the case with cyclin E1 , cyclin E2 levels were induced by HCMV infection in all serum conditions . Figure 5d also shows a clear increase in expression of cyclin E2 , and to a lesser extent TRIM28 , in cells infected with the miR-US25-1 knock-out virus compared to wild type infected cells , demonstrating that miR-US25-1 reduces the expression of these target genes . Time course experiments show that expression of cyclin E2 was equivalent between wild type and KO virus infection 24 hours post infection , and regulation by miR-US25-1 does not occur until approximately 48 hours post infection ( supplementary Figure S2d ) . This concurs with previous studies showing levels of miR-US25-1 increase during the progression of the viral infection and suggests that miR-US25-1 levels at 24 hours post infection are not high enough to produce measurable effects on cyclin E2 protein levels [7] . A slight increase ( approximately 1 . 5 fold ) was observed in RNA levels of cyclin E2 and TRIM28 , consistent with previous reports that miRNA targeting can cause moderate decreases in transcript levels ( Figure 5e and f ) . Following serum rescue , levels of cyclin E2 RNA increase in mock infected cells . The fact that cyclin E2 protein levels do not show a similar increase at this time likely reflects the delay between transcriptional activation and protein translation .
These observations provide the first comprehensive identification of multiple cellular targets of a viral miRNA using RISC-IP analysis . Strikingly , the study demonstrates that miR-US25-1 mediates inhibition of gene expression through the novel mechanism of targeting 5′UTR sequences . Furthermore , we demonstrate that miR-US25-1 targets multiple cellular genes related to cell cycle control . To our knowledge this is the first example of a viral miRNA targeting 5′UTRs and is the first demonstration of an endogenous miRNA repressing protein expression through targeting sequences within the 5′UTR . These results are in contrast to previous studies demonstrating miRNA targeting of 5′UTRs . Targeting of cellular 5′UTRs by miR-10a moderately increased gene expression while targeting of the HCV genome by miR-122 was shown to be required for viral replication [20] , [21] . Clearly , targeting of 5′UTRs by miRNAs can mediate distinct positive or negative regulatory effects depending on the context . It will be interesting to determine how miRNAs mediate distinct regulatory effects and whether inhibition of protein expression through miRNA targeting of 5′UTRs is more common than previously thought , or whether this mechanism is specific to miR-US25-1 or viral miRNAs . The functions of a number of cellular genes identified in this study have important implications for the biology of HCMV and viruses in general . Infection with HCMV has long been known to manipulate the cell cycle by altering the expression of cyclin dependent kinases ( CDKs ) and their associated cyclin subunits [27] . Cyclin E proteins are expressed early in G1 phase where they bind to and activate CDK2 , resulting in progression into S phase . Previous studies have demonstrated that HCMV induces resting G0 cells to enter the cell cycle whereupon the virus blocks further progression at the G1/S boundary [34] . By blocking the cell cycle at the G1/S phase the virus creates a cellular environment conducive for DNA replication . HCMV induced expression of cyclin E1 is thought to play an important role in driving cells into the G1/S phase [35] . Here , we show the virus also induces expression of cyclin E2 early in infection , then moderates cyclin E2 protein levels through targeting by miR-US25-1 . miR-US25-1 may therefore function as a rheostat regulator , modulating expression of cyclin E2 to generate the correct balance in protein induction . This may contribute to the viruses ability to block cell cycle progression at the G1/S phase , or to protect the infected cell against toxicity . Over-expression of cyclin E has been linked to sensitivity to apoptosis and unchecked induction of cyclin E2 may be detrimental to the virus [36] , [37] . Alternatively , miR-US25-1 function may be unrelated to cell cycle control . Recent studies have suggested that herpesvirus miRNAs may be important during persistent or latent infection [9] , [11]–[13] . HCMV is thought to reside within haematopoietic stem cell populations that give rise to latently infected monocyte and macrophage cells [38] , [39] . By targeting genes involved in cell cycle progression and differentiation , the virus could manipulate the production of cells generated by latently infected progenitors to favor certain cell types such as monocytes and macrophages . Although deletion of US25-1 did not result in a phenotypic effect on the replication following infection of primary human fibroblast cells , regulation of the target genes identified may be important in other cell types , such as endothelial cells or macrophage cells , or during the latent or persistent phase of the virus life cycle . Finally , the study of viral miRNAs may provide a powerful method for identifying novel cellular regulatory and antiviral pathways . This study suggests that viral miRNAs , like cellular miRNAs , may function through targeting multiple genes within related pathways . Many of target genes identified in this study are functionally related and contain the same 5′UTR sequence motif . It is likely that miR-US25-1 has evolved to target this 5′UTR motif , thereby subverting the regulatory pathway for the benefit of the virus . Investigation of viral miRNAs may therefore lead to discovery of additional novel cellular pathways .
RISC-IP analysis was carried as out previously described [22] , [23] . In brief HEK293 cells stably transfected with c-myc tagged Ago2 were transfected with the pSIREN expression plasmid or a synthetic biotinylated siRNA using Fugene ( Roche ) or RNAimax ( Invitrogen ) according to manufacturers specifications . Three days post infection cells were lysed , samples taken for total RNA levels and miRNP complexes immunoprecipitated using anti-c-myc antibody beads ( Sigma ) or streptavidin beads . RNA was isolated using Trizol and analyzed for quality using an Agilent Bioanalyzer and transcript levels determined on the Illumina HumanRef-8 platform . Microarray data was analyzed using Gene sifter software . Enrichment of specific transcripts , through association with miRNP complexes was determined by dividing the immunoprecipitated levels of transcripts by the total levels , thereby taking into consideration any direct effects of miR-US25-1 on transcript levels . This approach initially identifies any transcript associated with any miRNP complex . To specifically identify those transcripts targeted by miR-US25-1 , the enrichment profile was compared to cells transfected with a negative control vector , resulting in exclusion of transcripts enriched through association with cellular miRNAs . Transcripts were then ranked according to the level of enrichment with the highest enriched transcripts considered potential targets of miR-US25-1 . For example , in cells transfected with the negative control plasmid , cyclin E2 levels were 1958 in the total sample and 2219 in the IP sample , giving an enrichment value of 1 . 1 . In cells transfected with miR-US25-1 expression plasmid , cyclin E2 levels were 2718 in the total sample compared to 16744 in the IP sample , giving an enrichment value of 6 . 1 . By dividing the enrichment value from cells transfected with US25-1 compared to the control cells ( 6 . 1/1 . 1 ) the overall enrichment ratio is calculated as 5 . 4 . Argonaute specific antibody was generated by immunization of rabbits with a peptide corresponding to the N terminal region of Argonaute 2 ( 5-MYSGAGPALAPPAPPPPIQGYAFKPPPRPD3′ ) . For virus infections , primary human fibroblast cells ( Clontech ) were infected at high multiplicity ( MOI of 10 ) with the laboratory lab strain AD169 . RISC-IP analysis was conducted as above , except antibody to endogenous Ago2 was used to immunoprecipitated miRNP complexes and transcript levels determined using direct RT-PCR primer probe sets ( ABI ) for CCNE2 , TRIM28 , and GAPDH . Transcript sequences were down loaded from NCBI using RefSeq ID's . Transcript data sets were searched for seed sequence matches using a Java based script program . Statistical overrepresentation of seed matches within the top 50 transcripts from Table S3 was determined by Fisher exact test . Predicted binding between miR-US25-1 and target sites within the 10 most highly enriched transcripts were determined using the online RNA folding algorithm , mfold ( http://mfold . bioinfo . rpi . edu/cgi-bin/rna-form1 . cgi ) . 5′UTRs and approximately 500 bases upstream sequence of target genes were PCR amplified from human fibroblast DNA and cloned upstream of the pGL4 luc2 ( Promega ) firefly luciferase construct . 5′UTR luciferase constructs were cotransfected with a renillin control construct and miRNA mimics using Lipofectamine 2000 ( Invitrogen ) into HEK293 cells according to manufacturers instructions . Cells were harvested 18 hours post transfection and luciferase levels measured using Promega's dual reporter kit . Protein levels were determined using an anti-firefly luciferase antibody ( Sigma ) . Human primary fibroblast cells were grown in either 10% serum or 0 . 01% serum for 18 hours before infection at high multiplicity ( MOI of 10 ) with either wild type AD169 or miR-US25-1 knock out virus . Serum rescued cells were recovered with 10% serum 48 hours post infection . Seventy-two hours post infection , cells were harvested using RIPA buffer and total protein levels determined by BCA analysis . Thirty micrograms total protein was loaded and proteins detected using primary antibodies to cyclin E2 ( Abcam ) , TRIM28 ( Cellsignal ) , and GAPDH ( Abcam ) and secondary HRP-conjugated antibodies ( Jackson labs ) with ECL reagent ( GE bioscience ) . The predicted pre-miRNA region of miR-US25-1 plus approximately 100 additional bases were PCR amplified ( primers shown in supplementary Table S4 ) and cloned into the pSIREN expression plasmid ( Clontech ) . pSIREN NEG was supplied by Clontech . Synthetic miRNA mimics were purchased from Dharmacon . miR-US25-1 pre-miRNA coding region was deleted from AD169 BAC clone using BAC technology as previously described [40] . Briefly a PCR amplified cassette containing FRT flanked Kanamycin was recombined into AD169 BAC genome replacing the miR-US25-1 coding region using primers listed in Table S4 . The Kanamycin cassette was then removed by recombining the FRT sites through inducible FLIP recombinase . The resulting BAC was isolated and electroporated into human primary fibroblast cells to produce infectious virus . Total RNA was harvested using Trizol and concentrations determined on a nano-drop spectrophotometer . 100 ng of total RNA was then reverse transcribed using either random hexemers or specific RT primers for miRNA RT-PCR . Specific primer probe sets were then used for real time amplification using TAQMAN probes . All primers and probes shown in Table S4 . Gene specific primer probe sets were from ABI .
|
Regulation of gene expression is as important as the genes themselves in determining the diverse array of living creatures we see in nature . Recently , scientists have discovered a whole new level of gene regulation through the actions of small molecules called microRNAs ( miRNAs ) . It is currently thought that miRNAs regulate gene expression primarily through binding to target sites within the 3′UTR of mRNAs . Here we identify a population of cellular genes that are targeted by a virally encoded miRNA . Many of the genes are related to cell cycle control , suggesting that the viral miRNA is targeting genes within a related pathway . In contrast to most miRNAs , this miRNA inhibits gene expression through binding to target sites within the 5′UTRs , suggesting that viral miRNAs may target genes through mechanisms divergent from cellular miRNAs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/post-translational",
"regulation",
"of",
"gene",
"expression",
"molecular",
"biology/histone",
"modification",
"cell",
"biology/cell",
"growth",
"and",
"division",
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression",
"molecular",
"biology/translational",
"regulation",
"cell",
"biology/gene",
"expression"
] |
2010
|
A Viral microRNA Down-Regulates Multiple Cell Cycle Genes through mRNA 5′UTRs
|
Polygenic scores have recently been used to summarise genetic effects among an ensemble of markers that do not individually achieve significance in a large-scale association study . Markers are selected using an initial training sample and used to construct a score in an independent replication sample by forming the weighted sum of associated alleles within each subject . Association between a trait and this composite score implies that a genetic signal is present among the selected markers , and the score can then be used for prediction of individual trait values . This approach has been used to obtain evidence of a genetic effect when no single markers are significant , to establish a common genetic basis for related disorders , and to construct risk prediction models . In some cases , however , the desired association or prediction has not been achieved . Here , the power and predictive accuracy of a polygenic score are derived from a quantitative genetics model as a function of the sizes of the two samples , explained genetic variance , selection thresholds for including a marker in the score , and methods for weighting effect sizes in the score . Expressions are derived for quantitative and discrete traits , the latter allowing for case/control sampling . A novel approach to estimating the variance explained by a marker panel is also proposed . It is shown that published studies with significant association of polygenic scores have been well powered , whereas those with negative results can be explained by low sample size . It is also shown that useful levels of prediction may only be approached when predictors are estimated from very large samples , up to an order of magnitude greater than currently available . Therefore , polygenic scores currently have more utility for association testing than predicting complex traits , but prediction will become more feasible as sample sizes continue to grow .
Although individually significant markers in genome-wide association scans ( GWAS ) explain limited heritability of complex traits , evidence has been accruing that a considerable proportion of phenotypic variation can be explained by the ensemble of markers not achieving significance . Thus , while most of the specific genes underlying complex traits have yet to be identified , it is likely that many are represented on current genotyping products and specific identification is largely a matter of study size [1] . Polygenic score analysis has recently generated much interest for assessing the explanatory power of an ensemble of markers . A GWAS is conducted on an initial training sample , and the markers are ranked by their evidence for association , usually their P-values . An independent replication sample is then analysed by constructing , for each subject , a polygenic score consisting of the weighted sum of its trait-associated alleles , for some subset of top ranking markers . Two related but distinct applications of this score are then possible . Firstly , testing for association between the score and the trait in the replication sample can determine whether associated markers reside within those contributing to the score . Secondly and perhaps more usefully , the polygenic score can be used to predict individual trait values or risks of disease [2] , potentially giving a predictor with better discrimination properties than one based on established markers only . Different considerations apply for these two applications , as the size of the replication sample has a direct bearing on the power of association testing , whereas the accuracy of individual predictions depends only on the size of the training sample . The first successful application of polygenic score analysis to GWAS data was in schizophrenia [3] , in which few individual markers were significant and the common disease common variant hypothesis remained in question . It was shown that a large mass , up to half , of all markers in one GWAS could be jointly associated with disease in a second sample , implying a polygenic component to disease risk that justified larger study sizes [4] . Furthermore , markers from schizophrenia GWAS could together be associated with bipolar disorder , and vice versa , establishing a common polygenic basis to those conditions , whereas such cross-prediction was not achieved with clinically distinct conditions such as cardiovascular disease . This common basis has further been exploited to discriminate sub-types of bipolar disorder [5] . Similar results using a large mass of markers have been obtained for other complex traits including multiple sclerosis [6] , height [7] , cardiovascular risk [8] , rheumatoid arthritis [9] and body mass index [10] . In addition , several studies have demonstrated association of a score based on a limited number of top ranking markers [11]–[13] . In some cases , however , the polygenic association is less clear: studies of breast and prostate cancers have been inconclusive , owing in part to technical aspects in analysis but also , potentially , to their sample sizes [14] , [15] . An aim of the present work is to determine whether negative results from those studies could be explained by their sample size , or whether a true lack of polygenic effect is the more likely explanation . Applications of polygenic scores to individual disease prediction have so far been less successful , although proof of concept has been established through simulations [2] . Several studies have shown that a limited number of top ranking markers can discriminate disease cases from unaffected subjects , but the degree of discrimination falls short both of clinical utility and the maximum achievable from genetic data [16]–[18] . The use of a mass of markers across the whole genome has been explored , but to date has not yielded a noticeable improvement in discrimination [14] , [19] . Polygenic scores must be estimated from a finite training sample , and their effectiveness for association testing and risk prediction depends on the precision of this estimation as well as the proportion of variation explained by the polygenic score . The role of the sample size has not been thoroughly considered in this context . Several authors have expressed sensitivity and specificity in terms of the genetic variance of a predictor [17] , [20]–[22] , but they did not distinguish the variance explained by an estimated predictor from that of the true predictor , that is the one that would be estimated from an infinitely large sample . While large samples lead to small sampling variance on individual marker effects , the errors accumulate across multiple markers such that the effect of sampling variation on the polygenic score can be considerable . Wray et al [2] used simulations to study the predictive accuracy of scores estimated from finite case/control studies , but did not obtain an explicit relation between sample size and accuracy . Similarly , the International Schizophrenia Consortium ( ISC ) [3] used simulations to show empirical relations between sample size and accuracy under several genetic models . Daetwyler et al [23] considered the effect of sampling variation on the correlation between polygenic score and total genetic value . Their results can be adapted to prediction of phenotypes rather than genetic values , and also to other measures of power and accuracy , but their conclusions are limited by an assumption that all the markers have effects and are included in the score . In this work , statistical properties of polygenic score analyses are derived from a quantitative genetics model as a function of the explained genetic variance and sample sizes in discovery and replication samples . A range of options for constructing the score is considered , including estimation of the score from a different trait to the one predicted , selection of markers according to their P-values , and different methods for weighting markers in the score . The power is obtained for testing a polygenic score for association in a replication sample , and the correlation , mean square error , and area under the receiver-operator characteristic curve ( AUC ) are obtained for a predictor estimated from a finite training sample . These results are used to assess some recent studies and to discuss prospects for the future utility of polygenic score analyses for the prediction of complex traits .
In the framework considered here , a set of genetic markers is genotyped on an initial training sample and each marker is tested for association to a trait . Effect sizes are estimated for each marker and used to construct a polygenic score for each subject in an independent replication sample . The score is tested for association in the replication sample , in which the tested trait may differ from that in the training sample . The correlation and mean square error between the polygenic score and the tested trait are calculated . If the traits are binary , the AUC is obtained . More precisely , consider a pair of traits expressed as a linear combination of genetic effects and an error term that includes environmental and unmodelled genetic effects: ( 1 ) where is a matrix of coefficients , is a -vector of coded genetic markers , and is a pair of random errors that are independent of . Now suppose that the genetic effects on are estimated from a sample of size and used to construct a polygenic score to be tested for association to in an independent sample of size . Define the polygenic score to beSome important statistical properties of can be expressed in terms of and , expressions for which are derived in the Methods . The coefficient of determination for the polygenic score on the second trait is ( 2 ) which is the squared correlation between the score and the trait . The prediction mean square error is ( 3 ) The asymptotic non-centrality parameter of the test for association of with is ( 4 ) on 1df , and the power of the two-tailed test of association at significance level is ( 5 ) Binary traits are assumed to arise from a liability threshold model [24] leading to calculation of the AUC also in terms of and , with the expressions given in the Methods . For binary traits the coefficient of determination in equation ( 2 ) may be transformed to the liability scale for more satisfactory interpretation [25] , the details also given in the Methods . The expressions for power and accuracy are derived in terms of the parameters listed in Table 1 . Estimates of marker effects are either obtained from linear regression or set to a signed constant , which corresponds to the common approach of counting risk alleles across markers . A proportion of markers is assumed to have no effect , and markers may be selected by thresholding on their P-values . Equation 4 suggests an estimating equation for any parameter of the quantitative model , given the association test between and . Write explicitly as a function of some parameter in Table 1 , treating all other parameters as fixed and known . For example , might be the variance of marker effects in the training sample , from which is the explained genetic variance of the marker panel . Alternatively might be the covariance between marker effects in the two samples , assuming fixed values for the explained variances , and so on . Equation 4 is the squared coefficient of the linear regression of on , scaled by its sampling variance . The sampling distribution of that coefficient is normal , with mean the square root of equation 4 . Therefore applying normal theory an estimator is the solution to the equation ( 6 ) where is the observed association statistic . An approximate 95% confidence interval for is given by where is the solution ofand is the solution of The ISC was the first to demonstrate the utility of testing polygenic scores [3] . In their main result , odds ratios for 74062 nearly independent SNPs were estimated in 3322 cases and 3587 controls and used to construct a polygenic score that was tested in 2687 cases and 2656 controls of the Molecular Genetics of Schizophrenia study [26] . The score was more strongly associated as higher P-value thresholds were used for including SNPs , with the most significant reported association having with an inclusion threshold of . Assuming a prevalence of 1% , equation 5 gives a power of 80% at nominal significance if the explained genetic variance in liability is 7 . 2% , rising to 99% if the explained genetic variance is 11 . 7% . Assuming a heritability of 80% [3] this shows that the test was well powered if the marker panel explains about 10% of the heritability , which seems reasonable . The observed result of can be used in equation 6 to give an estimated explained genetic variance of 28 . 7% ( 95% CI: 23 . 6%–33 . 7% ) , which is 36% of the heritability , assuming that all SNPs have effects that are identical in the two samples . The estimate reduces only to 26 . 9% if 99% of the SNPs are assumed null . These results are similar to a recent estimate using mixed modelling of the same data [27] . In the ISC report , the P-value of the polygenic score decreases as the SNP inclusion threshold increases . This seems to suggest that a large number of associated markers lie within the mass of individually non-significant SNPs . In Figure 1 and Figure 2 , the expected P-value of the polygenic score is shown as a function of the inclusion threshold , with the explained genetic variance set to its estimated value of 28 . 7% and other parameters as stated above . The figures show that this trend could be observed when as many as 90% of SNPs have no effects , and for the linear regression estimator the significance of the score continues to improve until the whole marker panel is included . Only for a very high proportion of null SNPs is there an optimal inclusion threshold less than 1 . The allele count estimator has an optimum threshold less than 1 for all scenarios , but it is consistently less significant . Thus , in this dataset with high power , decreasing P-values are consistent with a range of polygenic models including those with a high proportion of null markers . In this analysis the replication sample was smaller than the training sample , and we may ask what balance of sample sizes is optimal . Given the total sample size of 3322+2687 = 6009 cases and 3587+2656 = 6243 controls , the non-centrality parameter can be numerically maximised over the proportion of subjects allocated to the training sample . It is found that the optimal split is close to one-half regardless of the proportion of null SNPs or the P-value threshold , and the non-centrality parameter is roughly symmetrical around one-half . This suggests that given two samples of different size , it matters little which is chosen for training and which for testing . Furthermore , given an initial sample to be split into training and replication subsets , an obvious rule of thumb is to make an even split . Similar properties are seen under different genetic models ( results not shown ) . Note that these results apply to association testing and not to individual prediction , which is discussed in the next subsection . For association testing there is a balance to be made between the precision of estimating the score in the training subset , and the power of testing the score in the replication subset . For prediction , however , the size of the replication subset does not affect the accuracy , only how precisely it is estimated; thus a larger training subset is more desirable in the prediction context . The ISC further tested the schizophrenia-derived score against bipolar disorder , to test for a common genetic basis to those conditions . Their strongest result was with the Wellcome Trust Case-Control Consortium ( WTCCC ) sample of 1829 cases and 2935 controls , obtaining with an inclusion threshold of . Assume similar heritability for bipolar disorder as for schizophrenia [28] and the same genetic variance explained by the markers , estimated above to be 28 . 7% . Then using equation 5 , the study had 80% power at nominal significance if the correlation is 28% between genetic effects on schizophrenia and bipolar disorder . Using equation 6 , the estimated correlation given the observed association statistic is 70 . 6% ( 95%CI: 51 . 3%–89 . 7% ) assuming that all SNPs have effects with explained variance 28 . 7% in both samples . If 99% of SNPs are assumed null , the estimated correlation reduces to 66 . 2% ( 95%CI: 48 . 1%–84 . 1% ) . The International Multiple Sclerosis Consortium performed a similar exercise using a training sample of 931 cases and 2431 controls , a replication sample of 876 cases and 2077 controls , and a marker panel of 59470 nearly independent SNPs [6] . They also observed decreasing P-values for association as more SNPs were included in the score , obtaining when all SNPs were included . Assuming prevalence of 0 . 1% this analysis has 80% power at nominal significance for explained genetic variance of 9 . 4% , and the observed result yields an estimate of 31 . 5% ( 95%CI: 24 . 9%–37 . 9% ) assuming all SNPs have effects . In applying these ideas to breast and prostate cancers , Machiela et al did not find significant associations of polygenic scores [14] . While this could be explained by the genetic architecture of the diseases , a possible explanation ( noted by the authors ) is the lower sample size together with the low heritability . Their breast cancer study used a total sample of 2287 subjects , approximately half of which were cases and half controls , which was split into training and testing subsets in a 9∶1 ratio for 10-fold cross-validation . The marker panel consisted of 161 , 702 nearly independent SNPs . Assuming a prevalence of 3 . 6% and sibling relative risk of 2 . 5 [29] , this design has only 17% power to detect an association of the polygenic score , even if the markers explain the full heritability . If the sample were split in a 1∶1 ratio , the power would increase to 37% . Their prostate cancer study had a total of 2277 subjects , approximately half of which were cases , again split in a 9∶1 ratio and a marker panel of 165 , 508 nearly independent SNPs . Assuming a prevalence of 2 . 4% and sibling relative risk of 2 . 8 [29] , this design has 19% power if the markers explain the full heritability . If the sample were split in a 1∶1 ratio , the power would be 42% . It is clear that even with the optimistic assumption that the markers explain the full heritability , this study was unlikely to detect an association of the polygenic score for either cancer . What sample size would have sufficient power to detect association of the polygenic score ? For breast cancer the heritability of liability is estimated as 44% [21] . If the marker panel explains half of this heritability , roughly as in the ISC study , then two samples each of 1978 cases and 1978 controls would have 80% power at nominal significance . For prostate cancer the heritability of liability is also 44% and 1766 cases and controls would be required in each sample . For the ISC study , assuming explained genetic variance of 28 . 7% , 735 cases and controls in each sample are sufficient . Thus it appears that association testing is well powered at current sample sizes if two independent studies are used for training and testing , but less well powered if a single sample is split into two subsets . As a final example of association testing , this time with a quantitative trait , Simonson et al studied the Framingham Risk Score for cardiovascular disease risk [8] . They also used 10-fold cross-validation of a single sample , giving training samples of 1575 subjects and testing samples of 175 subjects . They used a full set of 250 , 378 SNPs , which is here assumed to be similar to 100 , 000 independent SNPs . They first selected SNPs with P-values <0 . 1 into the score , then selected SNPs with 0 . 1<P<0 . 2 , then 0 . 2<P<0 . 3 and so on , giving ten analyses . Even if the trait is fully heritable and explained by these markers , this analysis has 20% power for the SNPs with P<0 . 1 , reducing with each P-value interval . For 0 . 4<P<0 . 5 , in which the authors found nominal significance of the score , the power is 6 . 7% . If all SNPs are included in the score , the power would be 38% if the trait is fully explained by the markers , but under a more conservative model in which the explained genetic variance is 30% , the power is just 8% and increases to 13% under an even split of training and testing samples . Again , splitting a single GWAS sample does not admit high power for testing a polygenic score . In their study of breast and prostate cancers Machiela et al also calculated the AUC for prediction of disease from the polygenic score . Here it is more important for the training sample to be large , ensuring accurate estimation of the score , justifying the 10-fold cross-validation design . Their AUC did not exceed 53% for breast cancer and 56 . 4% for prostate cancer . Under the same assumptions as above , the analytic AUC is 53 . 6% for breast cancer if the markers explain the full genetic variance , or 51 . 8% if they explain half . However if the sample were infinitely large , the AUCs would be 89% and 79% respectively . For prostate cancer , the analytic AUCs are 54 . 1% if the markers explain the full genetic variance , and 52% if they explain half; for a large sample they would be 90% and 80% . Thus , the low AUCs observed by Machiela et al are compatible with their study design , but they could be considerably higher if a larger training sample were available . Evans et al considered prediction for the seven diseases of the WTCCC [19] . Approximately 2000 cases were available for each disease , with a common set of 1480 controls , and a marker panel of all SNPs on the Affymetrix 500K chip after quality control and exclusion of previously known loci . As this is the same chip used in the ISC study , it is assumed here that the panel is equivalent to 74062 independent SNPs . Logistic regression and allele score estimators were both used to construct scores , and a series of P-value thresholds from 10−5 to 0 . 8 were considered . Table 2 compares the results of Evans et al to the analytic AUC for the diseases without strong MHC effects , using P<0 . 8 to select SNPs into the score , as that threshold generally gave the highest AUC . At that threshold , the choice of has little bearing on the results unless it is very close to 1 , so it is set to 0 . Also shown is the maximum AUC possible for each disease , obtained by letting the sample size grow to infinity . Bearing in mind that those authors noticed inflation in AUC for null SNPs , it is again clear that their modest results are compatible with the study design , and more encouraging results might be obtained from a larger sample . The calculations also confirm their observation that the allele count estimator is consistently less accurate than logistic regression; however while the two estimators give similar results at this sample size , more considerable differences emerge in the limit of large samples . Several other studies have reported pseudo-R2 from the regression of disease on the polygenic score [3] , [6] , [9] . Although prediction was not emphasised by those studies , they may still be evaluated for that purpose . Recently , Lee et al have argued that , for genetic predictors , R2 on the liability scale is a more interpretable measure of accuracy for binary traits [25] . In Table 3 , liability R2 derived from those reports are compared to analytic values assuming different levels of heritability explained by the markers . The choice of has little bearing on these results so it is set to 0 throughout . The reported values are consistent with the markers explaining around half the heritability , with variation above and below . This is in line with the estimates of explained variance that were reported by those studies , and those estimates also agree well with those obtained using the method proposed here ( equation 6 ) . The low reported values of R2 do not directly reflect the degree of missing heritability; rather they reflect the effect of sampling variation on the variance explained by an estimated score . Corresponding AUC values are also shown , and it is again clear that the currently modest utility of polygenic scores for discrimination is explained by limited training sample sizes , and much better results are possible through larger samples . What sample size would permit estimation of a score with AUC at a clinical useful level , or otherwise close to its maximum value ? The answer depends on , the proportion of null markers in the panel , because if this is high then the individual marker effects will also be high and a low P-value threshold will eliminate much sampling error from the estimated score . Figure 3 shows AUC as a function of sample size for Crohn's disease , which has a high heritability of 76% , and breast cancer , which has low heritability of 44% [21] , based on a panel of 100 , 000 independent markers . This is a similar number to current genotyping products , and results are given under a scenario in which the panel explains half the heritability [30] . For each sample size and , the P-value threshold is applied that leads to the highest AUC . An AUC of 0 . 75 is generally regarded as the minimum useful level for screening subjects already considered at risk , whereas AUC of 0 . 99 is sufficient for screening the population at large [31] . For these two diseases the latter cannot be achieved from genetic data alone , so Table 4 gives minimum sample sizes for AUC of 0 . 75 and for 90% , 95% and 99% of the maximum possible AUC given the heritability . The most favourable condition shown is , that is there are 1000 markers with effects on disease . Figure 3 and Table 4 show that a few thousand cases and controls could yield a clinically useful AUC , but under most conditions several tens of thousands are needed . Under less favourable conditions – low heritability , low proportion of null markers – several hundred thousand cases and controls are needed to obtain an AUC within 10% of the achievable level , and even an AUC of 0 . 75 requires some tens of thousands of subjects . In the worst case the order of magnitude is of the millions . Whole genome genotyping is now becoming feasible , under which the entire narrow-sense heritability would be represented . Assuming this is equivalent to about one million independent common SNPs [32] , the required sample sizes are shown in Figure 4 and Table 5 . Again , unless the heritability is explained by about 1000 markers , several tens to hundreds of thousands of subjects are needed to obtain a clinically useful AUC; for the genetic predictor to approach its potential , the order of magnitude is of the millions . The sample sizes to achieve AUC of 0 . 75 are larger than for 100 , 000 SNPs explaining half the heritability , but the latter scenario cannot achieve AUC of 0 . 99 , so the clinical context can influence the choice of marker panel used to derive the predictor . It is clear that at current sample sizes , polygenic scores are only going to approach useful levels of discrimination if the marker panels include a high proportion of associated loci and the number of such loci is relatively small . Furthermore , for highly polygenic conditions the sample sizes needed to approach this potential are an order of magnitude higher than are currently available . Finally , Table 6 gives similar calculations for the correlation between predicted and observed quantitative traits with high ( h2 = 0 . 8 ) and moderate ( h2 = 0 . 4 ) heritability . The prospects here appear more challenging in terms of the sample sizes needed to approach the achievable correlation . For example , height has heritability of about 0 . 8 , and the number of associated variants is known to be at least in the hundreds [7] . In the most optimistic scenario shown , 31 , 000 subjects would be required to derive a predictor with correlation 0 . 8 with the true height . In fact these sample sizes are now being approached by collaborative studies , and this result confirms that this is necessary for accurate prediction of quantitative traits in addition to the primary goal of identifying individually associated markers .
To date polygenic score analyses have been performed opportunistically . The results provided here allow a more informed appraisal of these analyses , characterisation of the statistical properties of the methods , and insights into the future prospects of polygenic modelling . R code to compute the formulae in this paper is available from the author ( sites . google . com/site/fdudbridge/software/ ) . Current sample sizes are clearly adequate for testing association of a polygenic score in a replication sample , as long as full size samples are used for both training and testing . This is already apparent from the extraordinary significance levels reported in the seminal studies [3] , [6] , but here it is shown that those results are compatible with realistic genetic models and are not necessarily explained by analytic biases that accumulate across SNPs . This had been previously shown by the ISC study , which simulated plausible genetic models and showed that they led to similar results to those observed in the data [3]; here , the result is shown directly for the common quantitative model , without recourse to simulations . Studies that split a single sample into cross-validation subsets have been less successful [8] , [14] , but here it is shown that this could be explained by their limited sample sizes , and more encouraging results for the same traits could be obtained with modestly increased samples . When a sample is to be split into two subsets , a roughly even split yields the greatest power for testing association of the score . However , for predicting individual trait values it is more important for the training set to be large , and standard procedures such as 10-fold cross-validation remain preferable . If both testing and prediction are intended from a single sample , a pragmatic approach is to ensure adequate power by allocating about 2000 cases and 2000 controls to the replication sample , providing this is less than half the total , and then to ensure high predictive accuracy by allocating the remainder to the training sample . The outlook for disease and trait prediction is more challenging . To date the severe shortfall in the accuracy of genetic predictors has generally been ascribed to incomplete coverage of marker panels or failure to identify sufficiently many associated markers . Here , however , no criteria for declaring individual significance are imposed , but neither does the calculation force the predictor to include markers that contribute no information . Under this pragmatic approach it results that tens of thousands of subjects , at least , are needed to derive predictors that are clinically useful . Furthermore , previous results on the potential accuracy of genetic prediction [17] , [20]–[22] only become relevant at very large sample sizes . Such numbers are now coming within reach of national biobank projects and international consortia , so the emergence of useful genetic predictors may not be too far off , although such large samples create issues of effect heterogeneity that are not addressed here . Recent estimates of the proportion of markers having effects also suggest that the more optimistic scenarios shown in Table 4 , Table 5 , Table 6 may apply [9] , [33] . Although the focus here is on AUC , various other measures of predictive accuracy are possible and can be computed within the same framework [24] , [25] . The expressions given here could be adapted to other measures without much difficulty . For some diseases , fairly high AUC has already been observed [17] , [19] . This does not conflict with the present work but reflects the presence of major gene effects , usually in the MHC , which depart from the quantitative model treated here . Similarly , some diseases have non-genetic risk factors that already admit clinically useful predictors . There the more relevant issue is the extent to which genetics improves established models [34] . Again the focus has tended to lie on identifying specific markers to improve prediction , rather than the sample size needed to accurately estimate their combined effects . The approach taken here could easily be extended to accommodate additional fixed effects . A fairly general construction of the polygenic score has been described , including weighted and unweighted methods from single marker analysis , and shrinkage methods used in multivariate analysis . There is little to choose between these estimators in terms of power , correlation or AUC , but the unweighted estimator will perform relatively worse as sample size increases since its sampling error does not reduce to zero . Shrinkage estimation leads to reduced mean square error for prediction and has some other advantages [35] , [36] , but in the main applications for polygenic scores to date , namely association testing and AUC , it does not improve over the linear regression estimate . However , some ideal conditions have been assumed including independence of markers and of study subjects . In reality markers will be in linkage disequilibrium and the approximation by an effective number of independent tests is heuristic . Similarly , subjects will be related , if distantly . Results from real data may depart from those presented here if proper account is taken of relationships between subjects . In particular , shrinkage estimation is likely to improve power and correlation , as well as mean square error , by analysing all markers simultaneously rather than each one marginally [37] . The assumption that effects are normally distributed is necessary when markers are selected by their P-values but not otherwise . Similarly , allowing a proportion of markers to have no effect only makes a difference when selecting markers by P-values . Thus the present results are relevant even if one does not entirely accept the polygenic model proposed . The normal distribution simplifies some calculations , but various heavy-tailed distributions have also been proposed for GWAS data [38] , [39] and would lead to improved prediction if such models held in truth . Furthermore the assumption of normality applies to effects on the standardised genotype scale , but there are plausible models for effect sizes as a function of allele frequency , leading to non-normal effects on the standardised scale . This may particularly affect the results for shrinkage estimation when the degree of shrinkage varies for markers with different allele frequencies . The numerical results presented are therefore not definitive but should be taken a guide to the likely magnitude of results in specific applications . A novel approach to estimating parameters of the polygenic model has been proposed , showing promise for inferring the explained genetic variance and/or proportion of null markers . The method yields estimates that are similar to those obtained by existing approaches [1] , [37] . A similar approach to estimation has been developed by Stahl et al [9] , based on simulating GWAS data from proposed models , and using rejection sampling to construct posterior distributions of their parameters . Apart from the accommodation of prior distributions ( which were uninformative ) , this is essentially the same approach as used here except that whole genome simulation is used to obtain a sampling distribution . The analytic results provided here should allow this approach to be implemented more efficiently , and this will be attempted in future work . The P-value thresholds that maximise the power and AUC are more permissive than the usual thresholds for individual markers . This means that polygenic analyses can be powerful while still including many non-significant markers , so they will continue to be useful as long as individually associated markers remain to be discovered . Although larger samples are needed for useful risk prediction , polygenic scores have an ongoing current role in assessing the variance explained by marker panels and the genetic correlation between related traits and populations .
Recall equation 1 in which a pair of traits is expressed as a linear combination of genetic effects and an error term that includes environmental and unmodelled genetic effects:where is a matrix of coefficients , is a -vector of coded genetic markers , and is a pair of random errors that are independent of . Assume that the markers are independent and standardised . In the usual case of single nucleotide polymorphism ( SNP ) genotypes under Hardy-Weinberg Equilibrium , where is the number of minor alleles and is the minor allele frequency at SNP . The genetic effects are regarded as fixed across samples but random over with , and . Then the variance-covariance matrix of is written asFor continuous traits , assume without loss of generality that and are standardised so that and are the proportions of variation of each trait explained by . These quantities will be called the explained genetic variances , and are bounded above by the heritabilities . The genetic effects on are estimated from a sample of size and used to construct a polygenic score to be tested for association to in an independent sample of size . Define the polygenic score to beClearly . Furthermore if then ( 7 ) These expressions are equalities in the limit of large but are approximations for a finite number of markers because the true effects are a sample from their random effects distribution . Equations 2–6 in Results follow immediately , in which the key quantities are and . They in turn depend upon the form of the estimator , for which three alternatives are now discussed . A natural estimate of is the least squares estimate from the univariate linear regression of on . Then is asymptotically normally distributed with sampling mean and variance since is standardised by definition . Assuming that genetic effects are small , it is henceforth conservatively taken that as previously suggested by Daetwyler et al [23] . The total variance of this estimator over markers and samples is , and its correlation with the effects on iswhere are the sampling errors . Immediate power and accuracy calculations are then available by substituting and into equations 2–5 . When , as when the same trait is considered in both samples , equation 2 gives the formula previously derived by Daetwyler et al [23] , modified to allow for prediction of the phenotype rather than the genetic value . In the present notation , corresponds to equation 1 of those authors , with the additional factor being the genetic variance of the phenotype . This shows that the key determinants of the predictive accuracy are the variance explained by the markers and the ratio of the sample size to the number of markers . Now suppose markers are only selected into the polygenic score if they have two-tailed P-values between thresholds where . Asymptotically the equivalent constraint for is obtained from the Wald statistic as ( 8 ) where , . Suppose further that a proportion of the markers have no effect on ( ie . ) , and the remaining markers have effects drawn from . Then among the null markers the variance of , conditional on selection into the polygenic score , is obtained from properties of the truncated normal distribution as [40]Similarly , among the non-null markers the variance of , conditional on selection into the polygenic score , is ( 9 ) where , . The probability that a null marker is selected into the polygenic score is and the corresponding probability for a non-null marker is . Therefore the total variance of is ( 10 ) Note that when and , that is all markers are included in the score , then equation 10 reverts to , which is invariant to the proportion of null markers and does not assume a normal distribution for the non-null effects [23] . To obtain allowing for selection of markers , note that the regression of on has the same coefficient regardless of selection on . For non-null markers this coefficient isand the covariance is this coefficient times the conditional variance of given in equation 9 . For null markers the covariance is zero , so the total covariance isThis expression is substituted into equations 2–5 together with the variance in equation 10 to obtain the power and accuracy of the polygenic score when markers are selected into the score based on their P-values . It is well known that estimation and prediction for multivariate models can be improved , in terms of mean squared error , by assuming that their effects come from a common underlying distribution . A common approach in quantitative genetics is to fit a mixed model in which genetic markers have random effects for which best linear unbiased predictors ( BLUPs ) are obtained [35] . This is one of several closely related formulations of multilevel models [36] . As these approaches tend to give similar results when the number of markers is large , a basic Bayesian estimation scheme is outlined here and will be assumed to give typical results for a shrinkage estimator . Suppose has the prior distribution , and let the “data” consist of the univariate linear regression estimates , . Then the posterior for given is also normal , where [41] . A natural estimator for is therefore the posterior mean for which and . Since all effects are shrunk by the same factor it follows that this approach leads to the same power and correlation as the linear regression estimator , but the mean square error is reduced to . A currently common approach is to construct the polygenic score by summing the number of trait-increasing alleles across selected markers , without considering their effect sizes other than to identify the direction of association at each marker . This may be called an unweighted score , in contrast to the above approaches that estimate weights for each marker . The unweighted score may be more robust against errors in estimating the effect sizes arising from limited sample size , population heterogeneity , “winner's curse” bias , and confounding by population structure . Here a related approach is considered in which all markers are given the same absolute effect size on the standardised genotype scale . This is equivalent to the allele counting approach when all markers have the same allele frequency . When allele frequencies are heterogeneous , allele counting assumes that all markers have the same effect on the trait , whereas the present approach assumes that all markers contribute the same proportion of variance to the trait . Both models can be criticised but the present approach will allow the comparison of weighted to unweighted scores without considering the distribution of allele frequencies or their relation to the effect sizes . The polygenic score is now calculated aswhere and is the linear regression estimate as before . Clearly . The covariance is obtained by integrating over the distribution of . Allow again for selection of markers by their P-values as in equation 8 and denote the selection event by . Then using the symmetry of the distribution of the required covariance isThe probabilities in this expression are as follows . The selection probability is againThe probability density for nonzero isGiven some value of the probability that its estimator is also positive , and the marker is selected into the score , isSimilarly given the probability that its estimator is negative , and the marker is selected , isFinally the conditional mean of given is given by properties of the bivariate normal distribution as . The integral can be evaluated numerically , yielding values for power and accuracy from equations 2–5 . The forgoing is based on linear regression , which is the usual approach for quantitative traits . For binary traits the standard analysis is logistic regression , used both for estimating the coefficients in the polygenic score and for testing the association of the score in a replication sample . For small effects the log-odds are approximately linear in the predictors , so we may continue to work in a linear regression framework for estimating power and accuracy . That is , the binary trait is coded as 0/1 and treated as the response in ordinary linear regression . The variance of in equation 2 is now the binomial variance where is the proportion of study subjects with . In a prospective sample , is the population proportion of the trait , whereas in a case/control sample ( to be discussed further below ) , it is the sampling proportion of cases . The binary traits are now assumed to arise from a liability threshold model , under which all individuals have an underlying normally distributed trait , called the liability , and all those whose liability exceeds a fixed threshold will exhibit the trait . Although the liability is not directly observed , this model has several advantages for modelling polygenic effects , including independence of the genetic effects from the trait prevalence , and an elegant linear transformation between effects on liability to corresponding effects on the observed ( 0/1 ) trait . This model has recently been elucidated by several authors for studying the quantitative genetics of binary traits in humans , and the reader is referred to their papers for more detailed discussion [21] , [24] , [30] . Assuming the marginal liabilities are distributed as a standard normal , the threshold for exhibiting trait is where is the population prevalence . The genetic effects are now taken to act on liability , and for small effects a linear transformation to the corresponding effect on the observed trait may be obtained as [30] ( 11 ) Given the genetic variance-covariance matrix on the liability scale , the statistical properties of the polygenic score may now be calculated as before , but substituting for and for throughout , and using as the sampling variance of . Sensitivity and specificity are often of interest in the prediction of binary traits . In particular , the accuracy of a predictor can be assessed by the AUC constructed as follows . Subjects are classified such that those with a polygenic score above a fixed threshold are predicted to have the trait , those below the threshold to not have it . Sensitivity is the proportion of subjects with the trait who are correctly predicted as such , and specificity the proportion of subjects without the trait correctly predicted as such . Each possible threshold leads to a value of sensitivity and specificity , defining the receiver operator characteristic curve by plotting sensitivity against 1-specificity . The AUC can be defined as the probability that a pair of subjects , one with the trait and one without , is correctly classified by the predictor . Because the central limit theorem implies that the polygenic score is normally distributed , the expected AUC can be calculated as [21] ( 12 ) In this expression , is formed from effects on estimated on the observed scale whereas the conditional means and variances are conveniently calculated on the liability scale for . There is a linear transformation between effects on and those on , defined by their bivariate normal distribution , and equation 11 gives another linear transformation between effects on and those on . Equation 12 may therefore be equivalently written in terms of effects on with the corresponding score denoted . The conditional means and variances are functions of the variance in explained by [21] , [40] , which is , givingandSimilarlyand In case/control studies the increased ascertainment of cases leads to departure from the normal distribution of liability assumed in the previous subsection . To overcome this problem , it is again assumed that there is a linear transformation from an effect on liability to one on the observed trait in which the 0/1 response denotes ascertained case status . When there is no selection on or the regression of on has coefficient from equation 11 . The converse regression of on has coefficient . The latter will also apply when there is ascertainment on , but the regression of on ascertained can also be written as , where denotes ascertainment , so thatThe desired quantity is the coefficient for the regression of ascertained on which is thusIn general the variance of genetic markers will differ from 1 under ascertainment but it will henceforth be assumed that its expectation over markers is approximately 1 . A heuristic justification for this assumption is given in the Text S1 . Based on this assumption , an effect on liability is transformed by the factor ( 13 ) to the observed case/control scale . Similarly to before , given the genetic variance-covariance matrix on the liability scale , the properties of the polygenic score can be calculated on the observed scale , substituting for and for throughout , and using as the sampling variance of . To obtain the AUC , the same approach as before is used , but now usingas the variance in explained by . Therefore , andSimilarlyandThe transformation from liability to observed scales differs from that of Lee et al [30] , which is for the total genetic liability ( their equation 19 ) . Here the interest is in the individual marker effects on the observed scale , because they are what are estimated when constructing the polygenic score . The derived expressions involve which is the coefficient of determination on the observed scale . Lee et al have argued that , for a genetic predictor , R2 on the liability scale is more interpretable for binary traits as it is invariant to the population prevalence and sampling ratio [25] . An approximate transformation to the liability scale is obtained by transforming the genetic effects using equation 13 and rescaling the trait variance from the binomial variance on the observed scale to the unit variance on the liability scale . Therefore , An alternative to the liability threshold model is the log-risk model for binary traits , which is equivalent to the logistic model in the limit of low prevalence . Here the polygenic score estimates the log risk of disease , which is assumed to be normally distributed in the population with mean and variance , where is the sibling relative recurrence risk [17] , [20] . Under this model the log risk has the same variance in cases and controls , but the mean log risk among cases is increased by that same variance , becoming . This model allows a simpler calculation of AUC for rare disease , which is given here but not pursued further . Given and and denoting log-risk of trait by , the transformation from log-risk to observed scales iswith the same adjustment for case/control ascertainment ( equation 13 ) . The difference in polygenic scores between cases and controls is the variance of the score , Since the polygenic score has the same variance in cases and controls , equation 12 gives the AUC as The derived expressions were compared to simulations in which the major assumptions were examined under realistic scenarios . These assumptions include a large number of markers with effects , for equality in equation 7 , and small genetic effects , so that effects on the liability scale are approximately linear . In case/control designs the disease prevalence is assumed to be not too small , so that the variance of the ascertained genotypes remains near 1 as assumed in equation 13 and Text S1 . Effects are assumed to be normally distributed on the standardised genotype scale . Sample sizes are assumed large so that estimates of genetic effects are normally distributed . A baseline scenario was defined to reflect that seen in recent studies , as follows . Two normally distributed traits were simulated with explained genetic variances 0 . 4 and 0 . 3 and correlation of genetic effects of 0 . 65 . Genotypes from 100 , 000 independent SNPs were simulated , with minor allele frequencies uniformly distributed on ( 0 . 01 , 0 . 5 ) . This reflects current marker panels that directly explain about half the heritability [1] . The proportion of null SNPs was 0 . 95 or 0 . 99 [9] , with the same SNPs having effects for both traits . Their effect sizes were drawn from the bivariate normal distribution such that the desired variances and covariance were attained . The traits were then generated from the quantitative model in equation 1 . The polygenic score was estimated using the first trait in a sample of 4000 unrelated subjects . The score was constructed using P-value thresholds of 0 . 1 and 0 . 001 for π0 = 0 . 95 and 0 . 99 respectively; these thresholds yielded the highest R2 and AUC values . The score was then tested for association with the second trait in an independent sample of 4000 subjects . The correlation and mean square error between the score and the second trait were also estimated in the second sample . The association tests were used in equation 6 to estimate the explained genetic variances in the first and second samples in turn , and then the covariance between effects in the two samples , each time keeping other parameters fixed to their simulation values . Table S1 shows estimates from 1000 simulations compared to the analytic values , for the three estimators discussed . Mean square error for the allele count estimator is not meaningful without further scaling of the polygenic score , which is a further problem not of present interest . All simulations agree well with the analytic results . Because the variances and covariance are bounded in ( 0 , 1 ) , their median estimates are shown with the coverage , rather than their means . The proposed estimating equations are seen to be accurate , but the confidence intervals are anti-conservative when the number of markers with effects is low , here 1000 . This is because the realised variance and covariance in equation 7 depart from their large m expectation , with resulting over-dispersion in the estimating equation ( left hand side of equation 6 ) . However when the number of markers with effects is 5000 , the correct coverage is attained . The traits were then treated as liabilities for binary diseases with prevalence 0 . 2 . Disease status was simulated prospectively , as in a cohort study . The polygenic score was estimated and tested using both linear and logistic regression . Table S2 shows estimates of power and AUC compared to the analytic values . Results for the shrinkage estimator are identical to the regression estimator and are not shown . All simulations agree well with the analytic results , and the proposed estimating equations are accurate . The results for logistic regression agree well with those for linear regression , justifying the use of the latter to derive the analytic results . Then , a case/control design was simulated in which the disease prevalence was now 0 . 001 . The same total sample sizes were used but included equal numbers of cases and controls . A computationally efficient approach to this simulation is described in Text S2 . The results are given in Table S3 . Again all simulations are seen to agree with the analytic values , but when the number of markers with effects is low , there is a downward bias in the parameter estimates and the confidence intervals of the parameter estimates are anti-conservative . Again the logistic regression results agree well with those for linear regression . Taking Tables S1 , S2 , S3 together , the analytic methods are accurate for the strongest effects likely to be seen in current studies , but when the number of SNPs with effects is about 1000 , there is downward bias in the effect estimates and under-coverage of the confidence intervals , the degree of which appears to vary with the strength of the association . To assess robustness to normality of the marker effects , the simulations were repeated with the effects drawn from Laplace distributions and then rescaled to give the same explained variance and correlation as before . Instead of π0 = 0 . 95 and P<0 . 1 , simulations with π0 = 0 and P<1 were performed to verify that this situation does not assume normality . The results in Table S4 , Table S5 and Table S6 confirm this to be the case , whereas when π0 = 0 . 99 and P<0 . 001 the analytic expressions tend to underestimate the power and accuracy . This is due to the heavier tails of the Laplace distribution compared to the normal , and quantitatively different results would be seen for different generating models . Again , bias and under-coverage is seen when there are 1000 markers with effects .
|
Recently there has been much interest in combining multiple genetic markers into a single score for predicting disease risk . Even if many of the individual markers have no detected effect , the combined score could be a strong predictor of disease . This has allowed researchers to demonstrate that some diseases have a strong genetic basis , even if few actual genes have been identified , and it has also revealed a common genetic basis for distinct diseases . These analyses have so far been performed opportunistically , with mixed results . Here I derive formulae based on the heritability of disease and size of the study , allowing researchers to plan their analyses from a more informed position . I show that discouraging results in some previous studies were due to the low number of subjects studied , but a modest increase in study size would allow more successful analysis . However , I also show that , for genetics to become useful for predicting individual risk of disease , hundreds of thousands of subjects may be needed to estimate the gene effects . This is larger than most existing studies , but will become more common in the near future , so that gene scores will become more useful for predicting disease than has appeared to date .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"medicine",
"public",
"health",
"and",
"epidemiology",
"statistics",
"mathematics",
"biostatistics",
"genetic",
"epidemiology",
"personalized",
"medicine",
"epidemiology",
"biology",
"genetic",
"association",
"studies",
"genetics",
"human",
"genetics",
"genetics",
"of",
"disease",
"statistical",
"methods",
"genetics",
"and",
"genomics"
] |
2013
|
Power and Predictive Accuracy of Polygenic Risk Scores
|
Candida albicans is the leading fungal pathogen of humans , causing life-threatening disease in immunocompromised individuals . Treatment of candidiasis is hampered by the limited number of antifungal drugs whose efficacy is compromised by host toxicity , fungistatic activity , and the emergence of drug resistance . We previously established that the molecular chaperone Hsp90 , which regulates the form and function of diverse client proteins , potentiates resistance to the azoles in C . albicans and in the model yeast Saccharomyces cerevisiae . Genetic studies in S . cerevisiae revealed that Hsp90's role in azole resistance is to enable crucial cellular responses to the membrane stress exerted by azoles via the client protein calcineurin . Here , we demonstrate that Hsp90 governs cellular circuitry required for resistance to the only new class of antifungals to reach the clinic in decades , the echinocandins , which inhibit biosynthesis of a critical component of the fungal cell wall . Pharmacological or genetic impairment of Hsp90 function reduced tolerance of C . albicans laboratory strains and resistance of clinical isolates to the echinocandins and created a fungicidal combination . Compromising calcineurin function phenocopied compromising Hsp90 function . We established that calcineurin is an Hsp90 client protein in C . albicans: reciprocal co-immunoprecipitation validated physical interaction; Hsp90 inhibition blocked calcineurin activation; and calcineurin levels were depleted upon genetic reduction of Hsp90 . The downstream effector of calcineurin , Crz1 , played a partial role in mediating calcineurin-dependent stress responses activated by echinocandins . Hsp90's role in echinocandin resistance has therapeutic potential given that genetic compromise of C . albicans HSP90 expression enhanced the efficacy of an echinocandin in a murine model of disseminated candidiasis . Our results identify the first Hsp90 client protein in C . albicans , establish an entirely new role for Hsp90 in mediating resistance to echinocandins , and demonstrate that targeting Hsp90 provides a promising therapeutic strategy for the treatment of life-threatening fungal disease .
Candida species have intimate yet perilous connections with their human hosts . They are commensals of the human microbiota of the gastrointestinal tract , mucous membranes , and skin . They also rank as the most common causative agents of invasive fungal infections and are responsible for a broad spectrum of disease [1] , [2] . For the immunocompetent individual , Candida infections are most often superficial in nature including thrush and vaginitis . For the immunocompromised individual , these opportunists are far more menacing , as they can disseminate and cause life-threatening systemic disease . Candida albicans is the most frequently encountered Candida species in the clinic and is the fourth most common cause of hospital acquired infectious disease with mortality rates approaching 50% [2] , [3] . The frequency of fungal infections continues to increase in pace with the growing immunocompromised patient population , including individuals undergoing chemotherapy , transplantation of solid organs or hematopoietic stem cells , as well as those infected with HIV [4] , [5] . Treatment of invasive fungal infections remains notoriously challenging , due in large part to the limited availability of clinically useful antifungal drugs . Fungi are eukaryotes and share close evolutionary relationships with their human hosts [6] , [7] . This makes the identification of drug targets in fungi that do not have homologs of similar function and susceptibility to inhibition in humans a daunting task . Most antifungal drugs in clinical use target the biosynthesis or function of ergosterol , the predominant sterol of fungal membranes , or the biosynthesis of ( 1 , 3 ) -β-D-glucan , a critical component of the fungal cell wall [8] , [9] . The azoles are the largest class of antifungal drugs in clinical use and have been deployed for several decades . They inhibit lanosterol 14α-demethylase , blocking ergosterol biosynthesis and resulting in the accumulation of a toxic sterol intermediate that disrupts membrane integrity and results in cell membrane stress . The echinocandins are the only new class of antifungal drug to be approved for clinical use in decades and inhibit ( 1 , 3 ) -β-D-glucan synthase , disrupting cell wall integrity and resulting in cell wall stress . The efficacy of antifungal drugs can be hampered by fungistatic rather than fungicidal activity , by host toxicity , and by the emergence of drug resistance . The azoles are generally fungistatic against Candida species and many immunocompromised patients are on long-term treatment due to persistent infections or on prophylaxis to prevent future infections . This creates favorable conditions for the evolution of drug resistance . In experimental populations and clinical isolates , resistance often emerges by multiple mechanisms [8]–[10] . Resistance mechanisms that minimize the impact of the drug include overexpression of multidrug transporters or alterations of the target enzyme . Other mechanisms function to minimize drug toxicity , such as loss of function of Erg3 in the ergosterol biosynthesis pathway , which blocks the production of a toxic sterol that would otherwise accumulate when the azoles inhibit their target . Mechanisms that mitigate drug toxicity are often dependent upon cellular stress responses that are crucial for tolerance of the membrane stress exerted by azoles [8] , [9] . Far less is known about resistance to echinocandins , at least in part due to their more recent approval for clinical use . The most common mechanism of echinocandin resistance is mutation of the drug target [11] . The ( 1 , 3 ) -β-D-glucan synthase complex consists of a regulatory subunit , Rho1 , and a catalytic subunit encoded by FKS1 , FKS2 , and FKS3 . Resistance is most commonly associated with characteristic mutations in FKS1 that reduce sensitivity of the enzyme to inhibition by echinocandins [11]–[13] . While the echinocandins are thought to be fungicidal against C . albicans , this organism has the capacity for robust growth at high drug concentrations , known as the paradoxical effect [14] . C . albicans may utilize multiple cellular stress response pathways to tolerate cell wall stress induced by echinocandins including upregulation of other components of the cell wall as well as responses mediated by the cell wall integrity signaling pathway [15] , [16] . A key regulator of cellular stress responses crucial for resistance to the azoles is the molecular chaperone Hsp90 . Hsp90 is an essential chaperone that regulates the form and function of many key signal transducers [17]–[19] . Pharmacological inhibition of Hsp90 blocks the emergence of azole resistance in C . albicans and abrogates resistance of laboratory mutants and clinical isolates that evolved resistance in a human host [20] , [21] . Impairing Hsp90 function converts the fungistatic azoles into a fungicidal combination and enhances the therapeutic efficacy of azoles in two metazoan models of disseminated C . albicans infection [22] . Hsp90's role in the emergence and maintenance of azole resistance is conserved in the model yeast Saccharomyces cerevisiae [21] . The key mediator of Hsp90-dependent azole resistance is calcineurin , a protein phosphatase that regulates crucial responses to environmental stress , including the membrane stress exerted by exposure to azoles [20] , [21] . In both S . cerevisiae and C . albicans , compromising calcineurin phenocopies compromising Hsp90 , reducing azole resistance of diverse mutants . In S . cerevisiae , Hsp90 interacts physically with the catalytic subunit of calcineurin keeping it stable and poised for activation [23] . High-throughput genomic and proteomic studies have mapped Hsp90 physical interactors in S . cerevisiae [24] , while to date not a single Hsp90 client protein has been identified in C . albicans . Given Hsp90's role in azole resistance , we postulated that this chaperone might also govern crucial responses to the cell wall stress exerted by echinocandins in C . albicans . We recently discovered that Hsp90 is required for the basal tolerance of Aspergillus species to echinocandins , which are fungistatic against Aspergillus species , and that Hsp90 inhibitors enhance the efficacy of echinocandins in an invertebrate model of Aspergillus fumigatus infection [21] , [22] . A . fumigatus is the principal causal agent of invasive aspergillosis with alarming mortality rates up to 90% that still remain at 40% with the best current treatment options [25] , [26] . Compromising calcineurin tracks with compromising Hsp90 , enhancing the activity of echinocandins [27] , [28] . While initial studies did not detect a role for Hsp90 in echinocandin resistance in C . albicans [21] , there are two lines of evidence implicating the Hsp90 client protein calcineurin in mediating responses to cell wall stress in this pathogen . First , stimulation of chitin synthesis rescues C . albicans from echinocandins and this stimulation is mediated via calcineurin in concert with the cell wall integrity signaling pathway and the high osmolarity glycerol signaling pathway [15] , [16] . Second , inhibition of calcineurin can block the paradoxical growth of C . albicans observed at elevated echinocandin concentrations [29] . Whether calcineurin mediates basal tolerance to echinocandins is unclear given that in one study , deletion of calcineurin enhanced the killing activity of an echinocandin [30] , while in another study there was no effect [31] . Thus , if Hsp90 regulates calcineurin function , then it is poised to mediate crucial cellular responses to the echinocandins . Here , we investigated Hsp90's role in tolerance to echinocandins in C . albicans . We found that pharmacological or genetic compromise of Hsp90 function reduced tolerance of laboratory strains to the echinocandins and created a fungicidal combination . Inhibition of Hsp90 also reduced resistance acquired by mutation in FKS1 in both laboratory-derived mutants and clinical isolates that acquired resistance in a human host . Compromising calcineurin function phenocopied compromising Hsp90 function . Consistent with calcineurin being the key mediator of Hsp90-dependent echinocandin tolerance , we established that calcineurin is an Hsp90 client protein in C . albicans . The downstream effector of calcineurin , Crz1 , played a partial role in mediating calcineurin-dependent stress responses that are activated by echinocandins . Hsp90's key role in governing crucial responses to cell wall stress exerted by echinocandins was not conserved in S . cerevisiae , emphasizing the importance of performing molecular studies in the pathogen . In a murine model of disseminated candidiasis , genetic impairment of HSP90 expression enhanced the therapeutic efficacy of an echinocandin . Our findings identify the first Hsp90 client protein in C . albicans and establish an entirely new role for Hsp90 in mediating echinocandin resistance . Further , our results demonstrate that targeting Hsp90 provides a promising therapeutic strategy for the treatment of life-threatening disease .
To determine the impact of compromising Hsp90 function on tolerance to echinocandins , we first used two structurally unrelated inhibitors geldanamycin ( GdA ) or radicicol ( RAD ) that bind with high affinity to Hsp90's unusual adenosine triphosphate ( ATP ) binding pocket and inhibit ATP-dependent chaperone function [32] , [33] . We used concentrations that abrogate resistance to azoles , but have no impact on growth on their own [20]–[22] . The impact of Hsp90 inhibitors on tolerance to the widely used echinocandin micafungin ( MF ) was evaluated using an antifungal susceptibility test that measures growth across a gradient of MF concentrations relative to a MF-free growth control . Both strains tested showed robust tolerance to MF ( Figure 1A ) . Inhibition of Hsp90 with GdA or RAD dramatically enhanced sensitivity to MF in either synthetic defined medium ( Figure 1A ) or in rich medium ( Figure S1A ) . Comparable effects were observed with another widely used echinocandin , caspofungin ( CS , Figure S1B ) . The same trends were observed when a dilution series of cells was spotted on solid medium with a fixed concentration of MF; concentrations of Hsp90 inhibitors that had no impact on growth on their own enhanced susceptibility to MF ( Figure 1B ) . Notably , while synergy of Hsp90 inhibitors with MF was observed in both liquid and solid media ( Figure 1 ) , the synergy with CS was restricted to liquid medium ( data not shown ) . This explains why the synergy between Hsp90 inhibitors and echinocandins was not detected in a previous study , which used CS on solid medium [21] . The basis for the different responses with MF and CS on solid medium is unclear and the response with a third echinocandin , anidulafungin , remains to be determined . Strains were more sensitive to echinocandins in a medium used for clinical susceptibility testing ( RPMI , Figure S1C and Figure S1D ) , however , compromising Hsp90 or calcineurin function further enhanced the sensitivity ( Figure S1C ) . Next , we exploited genetic regulation of Hsp90 to validate the impact of compromising Hsp90 function on echinocandin tolerance . Deletion of one HSP90 allele had negligible effect on MF tolerance ( Figure 1C and D ) . Replacing the native HSP90 promoter of the heterozygote with a tetracycline-repressible promoter has no effect on basal Hsp90 levels in the absence of tetracycline at 30°C , but blocks induction of HSP90 in response to stress such as elevated temperature of 37°C or exposure to antifungal drugs [22] . Even in the absence of tetracycline , compromising HSP90 expression in the tetO-HSP90/hsp90Δ strain resulted in hypersensitivity to MF in both liquid and solid media ( Figure 1C and 1D ) . While the tetO-HSP90/hsp90Δ strain also had a reduced growth rate , Hsp90 inhibitors at concentrations that have no effect on growth on their own dramatically enhanced echinocandin sensitivity ruling out the possibility that the hypersensitivity is simply due to reduced growth rate ( Figure 1B ) . Restoring a wild-type HSP90 allele to the tetO-HSP90/hsp90Δ strain complemented both the reduced growth rate and the hypersensitivity to MF . Thus , pharmacological and genetic studies establish that Hsp90 enables tolerance to echinocandins . It is now well established that a key mediator of Hsp90-dependent azole resistance is calcineurin , a protein phosphatase that regulates numerous responses to membrane stress in C . albicans [8] , [20] , [21] . If Hsp90 governs crucial cellular responses to echinocandins via calcineurin , then inhibition of calcineurin should phenocopy Hsp90 inhibition . We initially compromised calcineurin function pharmacologically using two structurally unrelated inhibitors cyclosporin A ( CsA ) and FK506 that inhibit calcineurin by distinct mechanisms [34] . CsA binds to Cpr1 , a peptidyl-prolyl cis-trans isomerase ( cyclophilin A ) , forming a drug-protein complex that blocks calcineurin function . FK506 forms a different drug-protein complex that binds to the structurally unrelated peptidyl-prolyl cis-trans isomerase FKBP12 to block calcineurin function . We used concentrations of CsA and FK506 that had no impact on growth on their own but that abrogate azole resistance [20] , [21] . Inhibition of calcineurin with either CsA or FK506 abolished MF tolerance of C . albicans ( Figure 2 ) . Next , we abolished calcineurin function genetically by either deleting the gene encoding the catalytic subunit of calcineurin , CNA1 , or by deleting the gene encoding the regulatory subunit of calcineurin required for its activation , CNB1 . In both cases , loss of calcineurin function abrogated MF tolerance ( Figure 2 ) . Reconstituting a wild-type allele of CNB1 restored tolerance . Thus , impairing calcineurin function recapitulates the effects of impairing Hsp90 , reducing echinocandin tolerance of C . albicans . The echinocandins are generally fungicidal against yeast species such as C . albicans [11] . However , C . albicans is able to grow vigorously at intermediate echinocandin concentrations in laboratory growth conditions ( Figures 1 and 2 and Figure S1 ) . Our previous assays did not resolve whether inhibition of Hsp90 or calcineurin results in a complete block in fungal growth in the presence of echinocandins or whether it creates a fungicidal condition . To determine if compromising Hsp90 or calcineurin function is fungistatic or fungicidal in the presence of echinocandins , we used tandem assays with an antifungal susceptibility test followed by spotting onto rich medium without any inhibitors . The common approach to address cidality by measuring colony forming units ( CFU ) in a culture exposed to treatment over time worked well for azoles [22] , but was not accurate for echinocandins . Exposure of C . albicans to MF caused severe clumping such that large aggregates of cells were not separable , rendering CFU counts inaccurate ( data not shown ) . A strain with wild-type or heterozygous HSP90 levels was able to grow on rich medium following exposure to all concentrations of MF tested ( Figure 3 , left panel ) . Genetic compromise of HSP90 expression in the tetO-HSP90/hsp90Δ strain or pharmacological inhibition of Hsp90 with GdA was cidal in combination with any dose of MF tested; no cells were able to grow on the rich medium following exposure to the treatments ( Figure 3 ) . Comparable effects were seen with genetic or pharmacological compromise of calcineurin function ( Figure 3 ) . Thus , Hsp90 and calcineurin regulate crucial cellular responses for surviving the cell wall stress exerted by the echinocandins . Compromising calcineurin pharmacologically or genetically phenocopies compromising Hsp90 suggesting a functional relationship between these regulators . Genetic studies established that calcineurin is a key mediator of Hsp90-dependent azole resistance [20] , [21] . In S . cerevisiae , Hsp90 physically interacts with the catalytic subunit of calcineurin keeping it stable and poised for activation [23] . High-throughput studies have mapped Hsp90 physical interactors in S . cerevisiae [24] , while to date not a single Hsp90 client protein has been characterized in C . albicans . In order to determine if Hsp90 and calcineurin physically interact in C . albicans , we engineered strains harboring epitope-tagged proteins for co-immunoprecipitation . We tagged the catalytic subunit of calcineurin , Cna1 , at the C-terminus using a 6X-histidine and FLAG epitope tag that has been used successfully for purification of the C . albicans septin complex [35] . The Cna1-His-FLAG protein is functional and sufficient to mediate the canonical calcineurin-dependent response to calcium stress ( Figure S2A ) . Immunoprecipitation with anti-FLAG agarose co-purified both FLAG-tagged Cna1 and wild-type Hsp90 ( Figure 4A ) . For the control strain lacking the tagged CNA1 allele , Hsp90 was present in the input but was not immunoprecipitated . To further validate the physical interaction between Hsp90 and calcineurin , we performed the reciprocal co-immunoprecipitation using the same tagged allele of calcineurin in addition to an HSP90 allele tagged at the C-terminus with a tandem affinity purification ( TAP ) tag which consists of a calmodulin binding peptide , a TEV cleavage site and two IgG binding domains of Staphylococcus aureus protein A that has been used with great success in S . cerevisiae [36] . The Hsp90-TAP protein is functional and able to support growth and all essential Hsp90 functions ( Figure S2B ) . Immunoprecipitation with IgG sepharose for the TAP tag , co-purifies both Hsp90-TAP and Cna1-His-FLAG ( Figure 4B ) . For the control strain lacking Hsp90-TAP , the tagged allele of calcineurin was present in the input but was not immunoprecipitated . Thus , reciprocal co-immunoprecipitation demonstrates physical interaction between Hsp90 and calcineurin in C . albicans . If calcineurin is an Hsp90 client protein , then one would expect that inhibition of Hsp90 function would compromise calcineurin activation . To determine if this is indeed the case , we used a well-established reporter system that exploits the calcineurin downstream effector Crz1 . Crz1 is a transcription factor that is dephosphorylated by calcineurin in response to calcineurin activation by calcium [37]–[39] . Dephosphorylated Crz1 translocates to the nucleus and drives expression of genes containing calcineurin-dependent response elements ( CDREs ) in their promoters . We used a strain harboring a construct with the UTR2 promoter , which contains a CDRE element and is regulated by calcineurin [37] , fused to lacZ and integrated at the UTR2 locus [40] . In S . cerevisiae , a similar reporter that contains four tandem copies of CDRE and a CYC1 minimal promoter driving lacZ has been used extensively [39] . As expected , exposure of cells containing the UTR2-lacZ reporter to calcium chloride resulted in activation of calcineurin relative to the untreated control ( P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test Figure 4C ) . Inhibition of calcineurin with CsA caused a dramatic reduction of calcineurin activation ( P<0 . 001 ) . Inhibition of Hsp90 with GdA or RAD was as effective in blocking calcineurin activation as CsA ( Figure 4C ) . A hallmark of Hsp90 client proteins is that they are destabilized and degraded upon compromising Hsp90 function . To determine if calcineurin levels are reduced upon genetic reduction of Hsp90 , we turned to a strain with its only HSP90 allele regulated by the MAL2 repressible promoter . In this system , HSP90 expression is induced by maltose and repressed by glucose ( Figure 4D ) . The MAL2 promoter does not drive as strong expression as the native HSP90 promoter , thus even when fully induced in maltose , the MAL2p-HSP90/hsp90Δ strain had a modest reduction of Hsp90 levels relative to a heterozygote with its only HSP90 allele under the control of the native promoter ( Figure 4D ) . Growth of cells in an equal mixture of glucose and maltose as the carbon source resulted in a dramatic reduction of Hsp90 levels ( Figure 4D ) . Under these conditions , the MAL2p-HSP90/hsp90Δ strain has reduced growth rate and reaches approximately half the stationary phase cell density as a wild-type strain [40] . This genetic depletion of Hsp90 was accompanied by a dramatic reduction of calcineurin levels as measured by immunoblot hybridization with an anti-FLAG antibody to detect the Cna1-His-FLAG protein ( Figure 4D ) . Hybridization with an anti-H3 antibody confirmed comparable amounts of protein were loaded for all strains . Taken together , these results support the model that calcineurin is a client protein in C . albicans . Due to the important role of calcineurin in mediating crucial responses to the stress exerted by exposure to azoles and echinocandins [20] , [21] , [30] , [31] , we postulated that these drugs would cause activation of calcineurin . We used the UTR2p-lacZ reporter to monitor calcineurin activation in response to concentrations of the azole antifungal drug fluconazole ( FL ) and the echinocandin MF that each cause modest inhibition of growth . Preliminary studies revealed maximum activation of calcineurin occurred at different time points in response to the different drugs ( data not shown ) . Exposure to MF for 8 hours caused significant activation of calcineurin ( Figure 5A , P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test ) . Pharmacological inhibition of calcineurin or Hsp90 blocked MF-induced calcineurin activation ( P<0 . 001 ) . Treatment conditions were optimized such that all cultures underwent comparable growth with equivalent protein yields . Exposure to FL for 24 hours also led to significant calcineurin activation ( Figure 5B , P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test ) . Inhibition of calcineurin or Hsp90 blocked FL-induced calcineurin activation ( Figure 5B , P<0 . 001 ) . Thus , both echinocandins and azoles activate calcineurin-dependent stress responses mediated via the transcription factor Crz1 and inhibition of Hsp90 blocks these responses . Crz1 is the key mediator of calcineurin-dependent transcriptional responses [37] , [41] and is implicated in tolerance to azoles in both S . cerevisiae and C . albicans [20] , [42] . While deletion of calcineurin causes a complete loss of azole tolerance , deletion of CRZ1 causes only a partial reduction in both species . To determine if Crz1 is also an important effector of calcineurin-dependent echinocandin tolerance , we compared the phenotypic consequences of deletion of CRZ1 with deletion of the catalytic subunit of calcineurin , CNA1 . Mutants with homozygous deletion of CNA1 were hypersensitive to MF in both liquid and solid assays ( Figures 2 and 6 ) . Two independent crz1 null mutants demonstrated partial loss of MF tolerance , but were not as sensitive as the cna1 mutants ( Figure 6 ) . Reconstitution of a wild-type CRZ1 allele restored MF tolerance . Thus , Crz1 is a key mediator of calcineurin-dependent echinocandin tolerance , but other calcineurin downstream effectors affecting this trait remain to be identified . To determine if Hsp90 and calcineurin are involved in bona fide echinocandin resistance arising due to mutations in the target Fks1 we tested for synergy between inhibitors of Hsp90 ( GdA ) or calcineurin ( CsA ) and the echinocandin MF . We utilized a checkerboard format to explore a range of concentrations of each inhibitor to more accurately define the thresholds of synergy . For a standard laboratory strain , SC5314 , potent synergy was observed such that very low concentrations of either GdA or CsA were sufficient to abrogate MF tolerance ( Figure 7 ) . Next , we tested an echinocandin resistant mutant that was selected in vitro in the SC5314 background by plating on a high concentration of the echinocandin caspofungin ( CS ) and contained the common Fks1 mutation F641S [12] . For this laboratory derived Fks1 mutant , C42 , synergy was observed; GdA or CsA reduced MF resistance , though not to the same extent as for SC5314 ( Figure 7 ) . To determine if the synergy between GdA or CsA and MF was conserved in an isolate that evolved echinocandin resistance in a human host , we tested a clinical isolate harboring the same F641S Fks1 mutation ( DPL15 , generously provided by D . S . Perlin ) . Comparable synergy between GdA and MF was observed for both the clinical and laboratory-derived Fks1 mutants , however , the synergy between CsA and MF was more potent against the clinical isolate ( Figure 7 ) . Interestingly , these synergies were not observed for all echinocandin resistant clinical isolates tested , even those harboring the identical FKS1 mutation; of the 14 FKS1 mutants tested , synergy was observed for 8 ( data not shown ) . These results suggest that Hsp90 and calcineurin enable cellular stress responses required for clinically relevant echinocandin resistance . To determine if impairing Hsp90 function holds therapeutic potential in combination with an echinocandin , we turned to a well-established murine model in which fungal inoculum is delivered by tail vein injection and progresses from the bloodstream to deep-seated infection of major organs such as the kidney [22] , [40] . Due to toxicity of currently available Hsp90 inhibitors that do not distinguish pathogen from host in the context of an acute fungal infection [22] , we used genetic regulation of HSP90 to test this hypothesis in an in vivo system . We compared kidney fungal burden of mice infected with either a strain with wild-type HSP90 levels or a strain with its only HSP90 allele expressed under the tetO promoter . In the absence of tetracycline , the tetO-HSP90/hsp90Δ strain has HSP90 levels comparable to a heterozygote but HSP90 expression from the tetO promoter cannot be upregulated in response to host temperatures or drug stress [22] . Mice infected with the tetO-HSP90/hsp90Δ strain demonstrated significantly reduced kidney fungal burden relative to those infected with a strain expressing wild-type HSP90 levels ( P<0 . 05 , ANOVA , Bonferroni's Multiple Comparison Test , Figure 8 ) . Treatment of mice with a dose of MF that had negligible effect on mice infected with the strain with wild-type HSP90 levels resulted in a significant reduction in fungal burden for mice infected with the tetO-HSP90/hsp90Δ strain ( P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test , Figure 8 ) . Thus , genetic compromise of HSP90 expression enhances the efficacy of MF in a murine model . Given that calcineurin is the key mediator of Hsp90-dependent resistance to azoles in both S . cerevisiae and C . albicans , we postulated that these key regulators of cellular signaling might also mediate tolerance to echinocandins in both species . Consistent with previous findings [23] , we confirmed that Hsp90 and calcineurin physically interact in S . cerevisiae ( Figure 9A ) , as they do in C . albicans ( Figure 4 ) . To monitor calcineurin activation in S . cerevisiae , we used a reporter system similar to that used for C . albicans . Cells contained an integrated plasmid with four tandem copies of CDRE and a CYC1 minimal promoter driving lacZ [39] . As expected for an Hsp90 client protein , calcineurin activation was blocked upon pharmacological inhibition of Hsp90 ( Figure 9B , P<0 . 001 , ANOVA , Bonferroni's Multiple Comparison Test ) . FL activated calcineurin in S . cerevisiae ( Figure 9C , P<0 . 0001 , t-test ) , as it did with C . albicans ( Figure 5B ) , consistent with the key role for both regulators in azole tolerance . MF also activated calcineurin in S . cerevisiae ( Figure 9C , P<0 . 0001 ) as it did in C . albicans ( Figure 5A ) . Despite activation of calcineurin by MF in S . cerevisiae , compromise of calcineurin or Hsp90 function had negligible effect on MF tolerance . Neither deletion of the gene encoding the regulatory subunit Cnb1 nor deletion of the redundant genes encoding the catalytic subunit Cna1 and Cna2 reduced MF tolerance ( Figure 9D ) . Consistent with this result , pharmacological inhibition of calcineurin with CsA had no impact on MF tolerance . A strain with genetically reduced Hsp90 levels ( Lo90 [21] ) had a modest reduction in tolerance , however , a concentration the Hsp90 inhibitor GdA that abrogates azole resistance had no effect on MF tolerance ( Figure 9D ) . This suggests that the slight reduction in MF tolerance of the Lo90 strain may be due to a reduced growth rate rather than compromise of Hsp90 function . Thus , while the functional relationship between Hsp90 and calcineurin is conserved between C . albicans and S . cerevisiae , as is the activation of calcineurin in response to drug stress , these regulators play a crucial role in cellular responses to echinocandins in the pathogenic yeast but not in the model yeast .
Our results establish a new role for Hsp90 in echinocandin resistance in the pathogenic yeast C . albicans . Hsp90 regulates crucial cellular responses to the cell wall stress exerted by echinocandins such that compromising Hsp90 function reduces echinocandin tolerance of laboratory strains and resistance of clinical isolates ( Figures 1 and 7 ) . In a murine model of disseminated C . albicans infection , genetic compromise of HSP90 enhances the efficacy of an echinocandin ( Figure 8 ) . We demonstrate that calcineurin is an Hsp90 client protein ( Figure 4 ) : calcineurin physically interacts with Hsp90; calcineurin activation is blocked upon impairment of Hsp90 function; and calcineurin levels are depleted upon genetic reduction of Hsp90 . Our findings implicate calcineurin as the key mediator of Hsp90-dependent echinocandin resistance . Exposure to azoles and echinocandins activates calcineurin-dependent stress responses ( Figure 5 ) and the downstream effector Crz1 plays a partial role in echinocandin tolerance ( Figure 6 ) . In addition to defining a novel mechanism of resistance to the only new class of antifungal drugs to reach the clinic in decades , these results provide the first characterization of an Hsp90 client protein in C . albicans . The requirement for Hsp90 and calcineurin in mediating crucial cellular responses to the echinocandins in C . albicans but not in S . cerevisiae ( Figures 1 , 2 , and 9 ) stands in contrast to the conserved role for both regulators in cellular responses to azoles in both species . It is intriguing that calcineurin is activated in response to echinocandin stress in S . cerevisiae yet the functional consequence of deleting calcineurin is negligible for this trait ( Figure 9 ) . Activation of signaling molecules does not always predict functional consequences of their deletion under equivalent conditions . For example , Mkc1 , the mitogen activated protein kinase ( MAPK ) in the PKC pathway , is activated by hydrogen peroxide but is not required for survival under this condition [43] . Our results suggest that there may be other redundant pathways operating in parallel with Hsp90 and calcineurin in S . cerevisiae . The protein kinase C ( PKC ) cell wall integrity pathway has a well-established function in mediating tolerance to echinocandins in S . cerevisiae [44] , [45] . In C . albicans , the PKC pathway is activated under diverse stress conditions [43] and works in concert with calcineurin and the high osmolarity glycerol pathway to regulate chitin synthesis , which can enhance tolerance to echinocandins [15] , [16] . There may be considerable interaction between PKC signaling , calcineurin , and Hsp90 . In S . cerevisiae , expression of one of the two partially redundant genes encoding the essential ( 1 , 3 ) -β-D-glucan synthase activity , FKS2 , is regulated by both PKC signaling and calcineurin [46] , [47] . In S . cerevisiae , Hsp90 may also interact with PKC signaling by chaperoning PKC [48] and the MAPK Slt2 [49] , [50] . Stress response signaling and canonical resistance mechanisms are intimately connected in defining a resistance phenotype . Compromising Hsp90 or calcineurin blocks the stress responses crucial for basal tolerance of strains that were not previously exposed to echinocandins ( Figures 1 , 2 , and 3 ) . There is heterogeneity in the phenotypic consequences of compromising these cellular regulators in strains that acquired resistance by mutation in the drug target Fks1 ( Figure 7 ) . For some isolates , resistance is not affected ( data not shown ) , while for others resistance is reduced , though not to the extent of a sensitive strain ( Figure 7 ) . This suggests that Hsp90 is not required to enable the phenotypic consequences of the mutant Fks1 protein . Rather , in many of the Fks1 mutants , Hsp90 and calcineurin-dependent stress responses contribute to the overall resistance phenotypes . Notably , the calcineurin inhibitor was more effective than the Hsp90 inhibitor at reducing MF resistance of some clinical isolates ( Figure 7 and data not shown ) ; this may be due to additional effects of CsA on targets distinct from calcineurin . The accumulation of mutations that reduce the dependence of resistance on Hsp90 is reminiscent of the evolution of azole resistance from Hsp90-dependence towards Hsp90-independence observed in isolates that evolved azole resistance in a human host [21] . Hsp90 chaperones many cellular regulators in addition to calcineurin . High-throughput genomic and proteomic studies suggest that Hsp90 may interact with up to 10% of the S . cerevisiae proteome [24] . Thus , Hsp90 is poised to regulate responses to antifungal drugs via other signal transduction pathways governing cellular stress responses . That Hsp90 regulates cellular responses to antifungal drugs targeting both the cell membrane and the cell wall via calcineurin emphasizes the importance of calcineurin as regulator of cellular stress responses . Cases of discordance between the phenotypic effects of compromising Hsp90 versus compromising calcineurin may reflect the relative importance of other Hsp90 client proteins in a particular trait or may reflect specificity of the agents used to inhibit these regulators [51] . Our results suggest that targeting Hsp90 may provide a powerful therapeutic strategy in the treatment of fungal infectious disease . In vitro , compromising Hsp90 function enhances the efficacy of echinocandins against isolates that evolved resistance in a human host and against isolates not previously exposed to echinocandins ( Figure 1 , 3 , and 7 ) . In a murine model of disseminated candidiasis , genetic impairment of HSP90 expression enhances the efficacy of an echinocandin ( Figure 8 ) . These findings add a new dimension to combinatorial therapeutic strategies for the treatment of C . albicans infections . Our previous work established that genetic reduction of Hsp90 levels enhances the efficacy of fluconazole in a murine model of disseminated C . albicans infection [22] and that further genetic depletion of C . albicans Hsp90 results in complete clearance of an infection in the murine model [40] . These studies establish firm proof-of-principle of Hsp90 as a therapeutic target . Current Hsp90 inhibitors that are well-tolerated in humans as anti-cancer agents exhibit toxicity in the mouse model in the context of an acute fungal infection [22] . However , in an invertebrate model of fungal pathogenesis , these pharmacological inhibitors of Hsp90 function enhance the efficacy of the two most widely deployed classes of antifungal drugs , azoles and echinocandins , against the two leading fungal pathogens of humans , Candida albicans and Aspergillus fumigatus [22] . Thus , compromising Hsp90 has broad therapeutic potential in combinatorial therapeutic regimens against fungal infections . Further support for targeting Hsp90 in antifungal therapy emerges from a recombinant antibody against the C . albicans chaperone . This recombinant antibody had therapeutic benefits in a clinical trial in combination with amphotericin B , which targets ergosterol [52] . This antibody also demonstrated synergy with the echinocandin caspofungin in a murine model [53] . The mechanism by which this antibody works , however , is unclear as the antibody is unlikely to be able to cross the fungal cell wall and access the cytosol of intact fungal cells , where Hsp90 regulates calcineurin-dependent signaling governing drug resistance . The antibody may work by influencing host immune responses to the pathogen . Consistent with this thinking , heat-shock proteins are immunodominant antigens for the recognition of many pathogens and play a central role in mediating both innate and adaptive immune responses [54] , [55] . Hsp90 has taken center stage as a therapeutic target for diverse diseases including cancer and neurodegeneration . Our findings suggest that Hsp90 may provide a much-needed target for life-threatening fungal infectious disease . Inhibitors of Hsp90 and calcineurin both have potent anti-malarial activity , thus extending their impact to the protozoan parasite Plasmodium falciparum [56] . Compromising host Hsp90 function in the context of an acute fungal infection is not well tolerated [22] . Perhaps in a related manner , the utility of calcineurin inhibitors in antifungal therapy has been complicated by their immunosuppressive effects [57] . Thus , the challenge in successfully exploiting this strategy lies in developing fungal selective inhibitors of Hsp90 or in targeting fungal specific components of the Hsp90 chaperone machine . Our findings may point to broader paradigm of targeting fungal stress response pathways in the treatment of life-threatening fungal infectious disease .
Archives of C . albicans and S . cerevisiae strains were maintained in at −80°C in 25% glycerol . Strains were grown in either YPD ( 1% yeast extract , 2% bactopeptone , 2% glucose ) , YPM ( as YPD except with 2% maltose ) , or in synthetic defined media ( yeast nitrogen base , 2% glucose ) and supplemented with the required amino acids . 2% agar was added for solid media . Strains were transformed following standard protocols . Strains used in this study are listed in Table S1 . Strain construction is described in Text S1 . Recombinant DNA procedures were performed according to standard protocols . Plasmids used in this study are listed in Table S2 . Plasmid construction is described in the Text S1 . Plasmids were sequenced to verify the absence of any nonsense mutations . Primers used in this study are listed in Table S3 . Antifungal susceptibility was determined in flat bottom , 96-well microtiter plates ( Sarstedt ) using a modified broth microdilution protocol , as described [21] . Minimum inhibitory concentration ( MIC ) tests were set up in a total volume of 0 . 2 ml/well with 2-fold dilutions of micafungin ( MF , generously provided by Julia R . Köhler ) or caspofungin ( CS , generously provided by Rochelle Bagatell ) . Echinocandin gradients were typically from 2 µg/ml down to 0 with the following concentration steps in µg/ml: 1 , 0 . 5 , 0 . 25 , 0 . 125 , 0 . 0625 , 0 . 03125 , 0 . 015625 , 0 . 0078125 , 0 . 00390625 , 0 . 00195313 . For gradients from 16 µg/ml down to 0 , the concentration steps in µg/ml were: 8 , 4 , 2 , 1 , 0 . 5 , 0 . 25 , 0 . 125 , 0 . 0625 , 0 . 03125 , 0 . 015625 . Cell densities of overnight cultures were determined and dilutions were prepared such that ∼103 cells were inoculated into each well . Geldanamycin ( GdA , A . G . Scientific , Inc . ) and radicicol ( RAD , A . G . Scientific , Inc . ) were used to inhibit Hsp90 at the indicated concentrations , and cyclosporin A ( CsA , CalBiochem ) and FK506 ( A . G . Scientific , Inc . ) were used to inhibit calcineurin at the indicated concentrations . Checkerboard assays were set up in a total volume of 0 . 2 ml/well with 2-fold dilutions of MF across the x-axis of the plate and 2-fold dilutions of either GdA or CsA across the y-axis of the plate . Plates were inoculated as with MIC tests . Dimethyl sulfoxide ( DMSO , Sigma Aldrich Co . ) was the vehicle for GdA , RAD , CsA , and FK506 . Sterile water was the vehicle for MF and CS . Plates were incubated in the dark at 30°C for the time period indicated , at which point plates were sealed and re-suspended by agitation . Absorbance was determined at 600 nm using a spectrophotometer ( Molecular Devices ) and was corrected for background from the corresponding medium . Each strain was tested in duplicate on at least two occasions . MIC data was quantitatively displayed with color using the program Java TreeView 1 . 1 . 3 ( http://jtreeview . sourceforge . net ) . Strains were grown overnight to saturation in YPD and cell concentrations were standardized based on optical density . Five-fold dilutions ( from ∼1×106 cells/ml ) were spotted onto indicated media using a spotter ( Frogger , V&P Scientific , Inc ) . Plates were photographed after 2 days in the dark at 30°C . All spottings were done in duplicate on at least two separate occasions . C . albicans cultures were grown overnight in YPD at 30°C with or without 10 µM CsA , 5 µg/ml FK506 , 5 µM GdA , or 5 µM RAD . Cells were diluted to OD600 of 0 . 5 and grown at 25°C for 2 h , at which point they were treated with MF , FL , or CaCl2 , as indicated . S . cerevisiae cultures were grown overnight in synthetic defined medium containing ammonium chloride at 30°C with 1 µg/mL FK506 or 5 µM GdA , as indicated . Cells were diluted to OD600 of 0 . 3 and treated with 0 . 2 M CaCl2 , FK506 , or GdA , as indicated . Cells were grown for 3 hours at 25°C . Protein was extracted as described [46] , [58] , and protein concentrations were determined by Bradford analysis . β-galactosidase activity was measured using the substrate ONPG ( O-nitrophenyl-β-D-galactopyranosidase , Sigma Aldrich Co . ) , as described [46] . β-galactosidase activity is given in units of nanomoles ONPG converted per minute per milligram of protein ( Miller Units ) . Statistical significance was evaluated using GraphPad Prism 4 . 0 . Yeast cultures were grown overnight in YPD at 30°C . Cells were diluted to OD600 of 0 . 2 in 40 ml and grown to mid-log phase . Cells were washed with sterile H20 and resuspended in 500 µl of lysis buffer containing 20 mM Tris pH 7 . 5 , 100 mM KCl , 5 mM MgCl and 20% glycerol , with one protease inhibitor cocktail ( complete , EDTA-free tablet , Roche Diagnostics ) per 10 ml , 1 mM PMSF ( EMD Chemicals ) and 20 mM sodium molybdate ( Sigma Aldrich Co . ) added fresh before use . Cells were transferred to a 2 mL screw-cap tube and the tube was filled , alternating with glass beads and additional lysis buffer until the beads were just below the meniscus at the top of the tube to reduce foaming during bead beating . Cells were disrupted by bead beating twice for 4 minutes with a 10 minute break on ice between cycles . Lysates were recovered by piercing a hole in the bottom of each tube , placing each tube in a larger 14 ml tube , and centrifuging at 1308×g for three 5-minute cycles , recovering the lysates at each interval . Total collected lysates were cleared by centrifugation at 20817×g for 10 minutes at 4°C and protein concentrations were determined by Bradford analysis . Anti-FLAG immunoprecipitations were done by diluting protein samples to 1 mg/ml in tris-buffered saline with 20 mM sodium molybdate and incubating with anti-FLAG M2 affinity agarose ( Sigma Aldrich Co . ) that was washed twice with tris-buffered saline prior to use , as per the manufacturer's specifications , at 4°C overnight . Unbound material was removed by three washes with 1 ml tris-buffered saline and protein was eluted by boiling the sample in one volume of 2× sample buffer . Anti-IgG immunoprecipitations were done by diluting protein samples to 1 mg/ml in lysis buffer with 0 . 2% tween and incubating with rabbit IgG agarose ( Sigma Aldrich Co . ) that was washed three times with lysis buffer prior to use , at 4°C overnight . Unbound material was removed by washing six times with 1 ml lysis buffer with 0 . 1% tween and protein was eluted by boiling the sample in one volume of 2× sample buffer . Yeast cultures were grown to mid-log phase , protein was extracted as above , and protein concentrations were determined by Bradford analysis . Protein samples were mixed with one-fifth volume of 6× sample buffer , were boiled for 5 minutes , and then separated on a 10% SDS-PAGE gel . Protein was electrotransferred to PVDF membrane ( Bio-Rad Laboratories , Inc . ) and blocked with 5% skim milk in phosphate buffered saline with 0 . 1% tween . Blots were hybridized with antibody against CaHsp90 ( 1∶10000 dilution , [59] ) , histone H3 ( 1∶3000 dilution; Abcam ab1791 ) , FLAG ( 1∶10000 , Sigma Aldrich Co . ) , Hsc82/Hsp82 ( 1∶5000 , [60] ) , or TAP ( 1∶5000 , Open Biosystems ) . Inoculum was prepared as described for injection of 100 µL of a 2×106 CFU/mL suspension [22] . Inoculum concentrations were verified by cell counts and CFU measurements . Male CD1 mice ( Charles River Laboratories ) age 8 weeks ( weight 30–34 g ) were infected via the tail vein . For infection with the wild type , the sample sizes were n = 6 mice for the untreated group and n = 5 mice for the MF treatment group . For the tetO-HSP90/hsp90Δ strain the sample sizes were n = 7 mice for the untreated group and n = 8 for the MF treatment group . An initial dose finding experiment was performed to determine a concentration of MF that would have negligible effect on fungal burden of mice infected with the wild type; a dose of 2 mg/kg MF ( Astellas Pharma , Inc; Deerfield , IL ) delivered intraperitoneally at one-hour post infection and then daily resulted in clearance of the fungal burden ( data not shown ) , while a dose of 0 . 2 mg/kg had no significant effect and was chosen as the dose for this study . Mice were observed three times daily for signs of illness and weighed daily . At day 4 following injection , mice were sacrificed by CO2 asphyxiation and the left kidney was removed aseptically , homogenized in PBS and serial dilutions plated for determination of kidney fungal burden , as described [22] . CFU values were expressed as CFU/g of tissue , log-transformed and compared using an ANOVA with post-hoc testing of significance between groups ( GraphPad Prism 4 . 0 ) . Murine work was performed under a protocol approved by the Institutional Animal Use and Care Committee at Duke University Medical Center .
|
Fungal pathogens pose a serious threat to people with compromised immune systems . Chief among the opportunistic fungal pathogens is Candida albicans . Treatment of C . albicans infections remains challenging because there are very few effective drugs and the pathogen has evolved many strategies to survive drug exposure . The echinocandins are the only new class of antifungal drug to reach the clinic in decades and they block biosynthesis of an essential component of the fungal cell wall . We discovered that the molecular chaperone Hsp90 , which is required for its client proteins in the cell to fold and function , governs the ability of C . albicans to survive exposure to echinocandins . Compromising Hsp90 function renders the echinocandins more effective at killing C . albicans laboratory strains and clinical isolates . Hsp90 orchestrates the crucial responses to cell wall stress exerted by the echinocandins by enabling the function of its client protein calcineurin , which allows the fungus to survive otherwise lethal conditions . Our results suggest that compromising Hsp90 function provides a powerful and much-needed strategy to render existing antifungal drugs more effective in the treatment of life-threatening fungal infections .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/fungal",
"infections",
"microbiology/medical",
"microbiology",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance",
"microbiology"
] |
2009
|
Hsp90 Governs Echinocandin Resistance in the Pathogenic Yeast Candida albicans via Calcineurin
|
The ocular onchocercosis is caused by the zoonotic parasite Onchocerca lupi ( Spirurida: Onchocercidae ) . A major hindrance to scientific progress is the absence of a reliable diagnostic test in affected individuals . Microscopic examination of skin snip sediments and the identification of adults embedded in ocular nodules are seldom performed and labour-intensive . A quantitative real-time PCR ( qPCR ) assay was herein standardized for the detection of O . lupi DNA and the results compared with microscopic examination and conventional PCR ( cPCR ) . The specificity of qPCR and cPCR was assessed by processing the most common filarial nematodes infecting dogs , skin samples from O . lupi infected ( n = 35 dogs ) or uninfected animals ( n = 21 dogs; n = 152 cats ) and specimens of potential insect vector ( n = 93 blackflies; n = 59 mosquitoes/midges ) . The analytical sensitivity of both assays was assessed using 10-fold serial dilutions of DNA from adult specimen and from a pool of microfilariae . The qPCR on skin samples revealed an analytical specificity of 100% and a sensitivity up to 8 x 10−1 fg/2μl O . lupi adult-DNA and up to 3 . 6 x 10−1 pg/2μl of mfs-DNA ( corresponding to 1 x 10−2 mfs/2μl ) . Only 9 . 5% O . lupi-infected skin samples were positive for cPCR with a sensitivity of 8 x 10−1 pg/2μl of DNA . Out of 152 blackflies and mosquitoes/midges , eight specimens experimentally infected ( n = 1 S . erythrocephalum; n = 1 S . ornatum; n = 6 Simulium sp . ) were positive by qPCR . The qPCR assay herein standardized represents an important step forward in the diagnosis of zoonotic onchocercosis caused by O . lupi , especially for the detection and quantification of low number of mfs . This assay provides a fundamental contribution for the establishment of surveillance strategies aiming at assessing the presence of O . lupi in carnivores and in insect species acting as potential intermediate hosts . The O . lupi qPCR assay will enable disease progress monitoring as well as the diagnosis of apparently clinical healthy dogs and cats .
Within the genus Onchocerca ( Spirurida: Onchocercidae ) , Onchocerca volvulus and Onchocerca lupi parasitize humans and carnivores , respectively [1–5] , the latter being a zoonotic agent [6 , 7] . While O . volvulus is a well-known parasite of humans transmitted by blackflies ( Simulium spp . ) [8 , 9] , the epidemiology of O . lupi is far from being understood , particularly because the information about insect species acting as vectors is lacking . Only Simulium tribulatum was suggested as the putative vector of this filarial worm in California ( USA ) , but proof of its intermediate host competence is currently absent [10] . Onchocerca lupi belongs to the spirurids in the Nematode clade III [11] was first detected from a Caucasian wolf ( Canis lupus ) in Georgia [12] , and , only recently , diagnosed in dogs and cats from Europe ( Greece , Portugal , Spain , Germany , Hungary ) and USA [13–20] . The reports of O . lupi infection are mainly based on the presence of ocular nodules on the eyelids , conjunctiva , and sclera [3 , 21 , 22] , though the localization of adult worms in the retrobulbar area of the canine patients may impair the assessment of its distribution in endemic areas [23] . The detection of microfilariae ( mfs ) in skin snip sediments is the only available tool for the diagnosis of the infection when nodules are not apparent in the eyes . The retrieval and identification of mfs in skin snip samples is a rather invasive and time-consuming method , highly dependent on the anatomical location of skin biopsy and mfs density [24] . Again , the detection of mfs may depend upon the prepatent period , previous microfilaricidal treatments , and on the operator’s skills in examining skin sediments , as described for O . volvulus [25 , 26] . Conventional PCR ( cPCR ) amplification and sequencing of mitochondrial NADH dehydrogenase subunit 5 ( ND5 ) and cytochrome c oxidase subunit 1 ( cox1 ) genes are available for the molecular identification of O . lupi adults and mfs [7 , 27 , 28] . The cPCR , however , may be relatively labour-intensive and exhibit low sensitivity , mainly for mfs detection , limiting the establishment of large-scale epidemiological studies in vertebrate hosts and putative vectors . Here , we developed a quantitative real-time PCR ( qPCR ) assay based on the hybridization probe to detect O . lupi DNA in host and putative vector samples . The diagnostic validity of qPCR assay was compared with microscopic examination and cPCR methods .
All dogs’ and cats’ skin samples were collected in previous studies [17 , 29] and approved by the ethical committee of the Department of Veterinary Medicine of the University of Bari ( Prot . Uniba 1/16 ) and by the ethical committee of the Faculty of Veterinary Medicine , Universidade Lusófona de Humanidades e Tecnologias . Genomic DNA of adult specimens of O . lupi ( n = 3 ) , as well as DNA from single ( n = 7 ) or pooled mfs ( n = 10 ) , collected from dogs in different geographical locations ( Table 1 ) were used as control . All specimens were previously identified based on morphological and molecular analyses [18 , 30] . Primers ( O . l . F 5′-GGAGGTGGTCCTGGTAGTAG-3′; O . l . R 5′- GCAAACCCAAAACTATAGTATCC-3′ ) and a TaqMan-MGB hydrolysis probe ( FAM-5’-CTTAGAGTAGAGGGTCAGCC-3’-non-fluorescent quencher-MGB; Applied Biosystems; Foster City , CA , USA ) , targeting partial cox1 gene ( 90bp ) , were designed by alignment of sequences from a wide range of closely related filarial nematodes available from GenBank database ( Table 2 ) , using Primer Express 2 . 0 ( Applied Biosystems , Foster City , CA ) . Specificity of the primers and probe for O . lupi were confirmed in silico using the basic local alignment search tool ( BLAST , GenBank , NCBI ) . qPCR reactions were carried out in a final volume of 20μl , consisting of 10μl of IQ Supermix ( Bio-Rad Laboratories , Hercules CA , USA ) , 7 . 1μl of Di-Ethyl Pyro-Carbonate ( DEPC ) treated pyrogen-free DNase/RNase-free water ( Invitrogen , Carlsbad , CA , USA ) , 2μl of template DNA ( except no-template controls ) , 5 pmol and 0 . 5 pmol for primers and probe , respectively . The run protocol consisted of a hot-start at 95°C for 3 min , and 40 cycles of denaturation ( 95°C for 10 sec ) and annealing-extension ( 64°C for 30 sec ) . All assays were carried out in duplicate and a no-template control was included in each run . The qPCR was performed in a CFX96 Real-Time System ( Bio-Rad Laboratories , Inc . , Hercules CA , USA ) and the increase in the fluorescent signal was registered during the extension step of the reaction and analysed by the CFX Manager Software Version 3 . 1 ( Bio-Rad ) . To investigate the analytical specificity of the assay , genomic samples of Onchocerca spp . and of the most common filarial nematodes infesting dogs ( Table 1 ) were used . The specificity of the assay was tested by using DNA from skin samples of naturally infected dogs , which were positive for O . lupi ( n = 35 ) at microscopic examination [29] . Skin samples were divided in five groups ( G1-G5 ) according to their mfs load ( Table 3 ) , being 14 also co-infected with Cercopithifilaria bainae and Cercopithifilaria sp . II . Skin samples ( dogs n = 21; cats n = 152 ) , which did not test positive to any mfs [17 , 29] , were used as negative control . Specimens of blackflies ( n = 66 ) and mosquitoes/midges ( n = 39 ) collected from 2011 to 2014 in Greece [31] , and 27 blackflies and 20 Aedes albopictus ( colony specimens ) experimentally infected by intrathoracic microinjection with mfs of O . lupi ( parasitic load of 20mfs/μl ) were analyzed after death ( i . e . , from one to 10 days post infection ) ( Table 4 ) . The analytical sensitivity of the qPCR assay was assessed using 10-fold serial dilutions of DNA from adult specimen ( i . e . , ranging from 8 × 104 to 8 × 10−3 fg/2μl of reaction ) and from a pool of 10 mfs ( i . e . , ranging from 10 to 10 × 10−3 microfilariae/2μl of reaction , corresponding to 3 . 6 ×10−1 ng/2μl to 3 . 6 ×101 fg/2μl of DNA ) . Ten replicates of each serial dilution were submitted to the same run for assessment of intra-assay reproducibility . Genomic DNA was isolated from all skin samples and from O . lupi adults and mfs , blackflies , mosquitoes and midges specimens using the commercial kits DNeasy Blood & Tissue Kit ( Qiagen , GmbH , Hilden , Germany ) , respectively , following the manufacturers’ instructions . The amounts of purified DNA were determined spectrophotometrically using the Qubit ( Applied Biosystems , Foster City , CA , USA ) . The analytical specificity and sensitivity of the cPCR for the specific amplification of cox1 gene fragment ( ∼689bp; [32] ) was assessed by testing genomic DNA of: i ) skin samples with different parasitic load of O . lupi ( Table 3 ) , ii ) serial dilution of O . lupi mfs DNA ( i . e . , from 3 . 6 ×101 pg/2μl to 3 . 6 ×10−3 pg/2μl ) and iii ) DNA of adult specimens ( i . e . , from 8 ×101 ng/2μl to 8 x 10−3 fg/2μl ) . All cPCR products were resolved in 0 . 5x GelRed stained ( Biotium , CA , USA ) agarose gels ( 2% ) , purified using enzymatic purification ( Exo I-FastAP; Thermo Fisher Scientific , MA , USA ) and sequenced in an automated sequencer ( 3130 Genetic Analyzer ) . All sequences generated were compared with those available in GenBank using Basic Local Alignment Search Tool ( BLAST ) [33] .
All O . lupi naturally-infected dog positive at skin samples examination by microscopy , considered the gold standard method as true positives , were positive by the O . lupi qPCR herein assessed ( specificity of 100% ) . Out of 21 skin samples microscopically and qPCR positive for O . lupi , two were positive by cPCR ( parasite load of 8 and 25 mfs ) , revealing a low analytical cPCR specificity ( i . e . , 9 . 5% ) . None of cat’s skin samples were positive by qPCR . A specific fluorescent signal was recorded for all O . lupi adult and mfs positive controls tested ( Fig 1 ) . No fluorescence was obtained for all other Onchocerca species or filarial nematodes examined as well as for skin samples used as negative control . The analytical sensitivity of qPCR was confirmed by detection of up to 8 x 10−1 fg/2μl and 3 . 6 x 10−1 pg/2μl of DNA ( i . e . , corresponding to 1 x 10−2 mfs/2μl ) of O . lupi adult worm and mfs , respectively ( Fig 2A and 2B ) . qPCR efficiencies ranged from 108 . 7 to 115 . 3% with an R2 from 0 . 996 to 0 . 999 and Slope from -3 . 003 to -3 . 131 , for both adult and mfs ( Fig 2A and 2B ) . The mean parasite load detected for the positive skin samples , ranged from 1 . 9 to 96 mfs/2μl of reaction , corresponding to 6 . 1 x 10−2 ng/2μl ( mean cycle threshold of 33 . 49 ) and to 3 . 4 ng/2μl DNA ( mean cycle threshold of 27 . 52 ) , respectively ( Table 3 ) . The results of mfs detection by qPCR overlapped the values obtained by the microscopic examination . The detection limit registered for cPCR was up to 8 x 10−1 pg/2μl for adult worms and up to 3 . 6 x 101 pg/2μl for mfs DNA ( i . e . , corresponding to 1 mf/2μl ) , respectively ( Fig 3 ) . Out of 152 blackflies , mosquitoes and midges , eight Simulim spp . ( n = 1 S . erythrocephalum; n = 1 S . ornatum; n = 6 Simulium sp . ) , experimentally infected and died from 1 to three days post infection , returned positive signal for O . lupi DNA ( Table 4 ) . All field-collected blackflies and mosquitoes were negative for O . lupi DNA using qPCR ( Table 4 ) . All blackflies positive for qPCR scored positive also for cPCR . Sequences derived from all amplicons of cPCR matched with 99–100% nucleotide identity appropriate reference sequences of O . lupi available from GenBank ( accession numbers KC686702 , KC686701 ) .
A qPCR assay has been developed for the detection of O . lupi in animal skin snip samples and potential vectors and proved to be a sensitive and specific tool for the diagnosis of this parasite , with a mean detection limit as low as 1 . 9 mfs per reaction . In addition , the high sensitivity of the qPCR protocol has been demonstrated by detecting a small amount of DNA ( up to 8 x 10−1 fg/2μl for adult and up to 3 . 6 x 10−1 pg/2μl for mfs ) , by the slope value of standard curve ( −3 . 131 ) , the efficiency ( 115 . 3% ) and the coefficient of determination ( R2 = 0 . 999 ) . These features of the assay are due to the selection of a stable hydrolysis probe designed ( 100% specific for O . lupi DNA ) , as well as to the choice of the target gene used . Indeed , cox1 gene of the mitochondrial DNA has been well recognized as a “barcode” for filarial nematodes [34] , with a high amplification efficiency , also due to the large copy numbers enabling the detection of minimum amounts of DNA [35–37] . Though few Onchocerca species DNA were herein tested , which may represent a limitation of the qPCR assay , this new tool provides an alternative to the labor intensive microscopic examination of skin snip samples and to cPCR for the diagnosis of O . lupi [38] . The qPCR assay was highly specific in revealing O . lupi DNA both in co-infected samples from dogs as well as in potential vector species , avoiding the sequencing confirmation needed using cPCR with filarioid generic primers [32] . Overall , the positive fluorescent signal from samples of O . lupi , from different geographical areas ( i . e . , Europe and USA ) , which displayed genetic intraspecific variability [18] , indicates the usefulness of the qPCR also for the surveillance of O . lupi where the parasite has been reported [13 , 14 , 16 , 17 , 19 , 39–41] . Similarly , even if the qPCR cannot discriminate between viable and nonviable parasites or immature and infective larvae , the assay could be useful for detecting O . lupi in blackfly , mosquito and/or midge species , potentially involved in the transmission of this parasite . Indeed , the specificity of the qPCR to amplify exclusively the DNA of the pathogen in potential insect vectors herein tested , may ultimately assist in the quest to identify the elusive vector of O . lupi . The newly designed assay represents an improvement in the diagnosis of onchocercosis , by the detection and quantification of low mf densities from tissue samples and could provide a contribution to disease progress monitoring and to the surveillance of O . lupi-infected dogs , avoiding the introduction and/or spread of this life-threatening parasitic nematode , as well as to the identification of apparently healthy animals [29 , 42] . The qPCR may speed-up time of diagnosis and prompt treatments of infected animals , which may avoid the appearance of nodular lesions in the eyes or in other anatomical localizations [43] . A TaqMan-based specific and sensitive assay without sequencing is expected to assist high-throughput analysis of samples , eventually leading to improve disease monitoring under the frame of a Public Health perspective . This would be particularly relevant considering that , since its first description of its zoonotic potential [7] , cases of zoonotic onchocercosis are being detected increasingly in people from Europe , Iran and the USA [44–47] .
|
The diagnosis of zoonotic ocular onchocercosis caused by Onchocerca lupi ( Spirurida: Onchocercidae ) is currently based on microscopic examination of skin snip sediments and on the identification of adults embedded in ocular nodules . These methods are labour-intensive and require multiple steps to achieve the diagnosis . In this context , a novel quantitative real-time PCR assay ( qPCR ) has been herein standardized and analytical specificity and sensitivity assessed . The results indicate that the qPCR assay could represent an important step forward in the diagnosis of onchocercosis , in carnivores and in insect species acting as potential intermediate hosts .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"onchocerca",
"volvulus",
"helminths",
"wolves",
"vertebrates",
"parasitic",
"diseases",
"dogs",
"mammals",
"animals",
"onchocerca",
"nematode",
"infections",
"veterinary",
"diagnostics",
"dna",
"veterinary",
"science",
"veterinary",
"medicine",
"biochemistry",
"eukaryota",
"nucleic",
"acids",
"genetics",
"nematoda",
"biology",
"and",
"life",
"sciences",
"amniotes",
"organisms"
] |
2018
|
A real-time PCR tool for the surveillance of zoonotic Onchocerca lupi in dogs, cats and potential vectors
|
Hematopoietic stem cells in mammals are known to reside mostly in the bone marrow , but also transitively passage in small numbers in the blood . Experimental findings have suggested that they exist in a dynamic equilibrium , continuously migrating between these two compartments . Here we construct an individual-based mathematical model of this process , which is parametrised using existing empirical findings from mice . This approach allows us to quantify the amount of migration between the bone marrow niches and the peripheral blood . We use this model to investigate clonal hematopoiesis , which is a significant risk factor for hematologic cancers . We also analyse the engraftment of donor stem cells into non-conditioned and conditioned hosts , quantifying the impact of different treatment scenarios . The simplicity of the model permits a thorough mathematical analysis , providing deeper insights into the dynamics of both the model and of the real-world system . We predict the time taken for mutant clones to expand within a host , as well as chimerism levels that can be expected following transplantation therapy , and the probability that a preconditioned host is reconstituted by donor cells .
The hematopoietic system has evolved to satisfy the immune , respiratory , and coagulation demands of the host . A complex division tree provides both amplification of cell numbers and a variety of differentiated cells with distinct roles in the body [1–3] . In a typical adult human ∼1011 terminally differentiated blood cells are produced each day [3–5] . It has been argued that the division tree prevents the accumulation of mutations , which are inevitable given the huge number of cell divisions [6–8] . At the base of the tree are hematopoietic stem cells ( HSCs ) . These have the ability to differentiate into all hematopoietic cell lineages , as well as the capacity to self-renew [1 , 9] , although the exact role of HSCs in blood production is still debated [10 , 11] . With an aging population , hematopoietic malignancies are increasingly prevalent [12] . Clonal hematopoiesis—where a lineage derived from a single HSC is overrepresented—has been identified as a significant risk factor for hematologic cancers [13–15] . To assess the risks posed to the host we need an understanding of how fast clones are growing , when they initiate , and if they would subvert physiologic homeostatic control . The number of HSCs within a mouse is estimated at ∼0 . 01% of bone marrow cellularity [16 , 17] , which amounts to ∼10 , 000 HSCs per host [3 , 16 , 18 , 19] . In humans this number is subject to debate; limited data has lead to the hypothesis that HSC numbers are conserved across all mammals [18] , but the fraction of ‘active’ HSCs depends on the mass of the organism [20] ( see also Refs [5 , 21] for a discussion ) . Within an organism , the HSCs predominantly reside in so-called bone marrow niches: specialised micro-environments that provide optimal conditions for maintenance and regulation of the HSCs [22 , 23] . There are likely a finite number of niches within the bone marrow , and it is believed that they are not all occupied at the same time [16] . The number of niches is likely roughly equal to the number of HSCs , and through transplantation experiments in mice it has been shown that ∼1% of the niches are unoccupied at any time [16 , 24] . A similar number of HSCs are found in the peripheral blood of the host [16] . These free HSCs are phenotypically and functionally comparable to ( although distinguishable from ) bone marrow HSCs [19 , 25] . The HSCs have a residence time of minutes in the peripheral blood , and parabiosis experiments ( anatomical joining of two individuals ) have shown that circulating HSCs can engraft to the bone marrow [25] . It has also been shown that HSCs can detach from the niches without cell division taking place [19] . These findings paint a picture of HSCs migrating between the peripheral blood and the bone marrow niches , maintaining a dynamic equilibrium between the two compartments . In this manuscript we construct a model from the above described processes , and we use this to answer questions about clonally dominant hematopoiesis . We first consider this in mice , where we use previously reported values to parametrise our model . The model is general enough that it also captures scenarios of transplantation into both preconditioned ( host HSCs removed ) and non-preconditioned hosts: the free niches and the migration between compartments also allows for intravenously injected donor HSCs to attach to the bone marrow niches and to contribute to hematopoiesis in the host . In the discussion we comment on the implications of these results for human hematopoiesis .
A Wolfram Mathematica notebook containing the analytical details can be found at https://github . com/ashcroftp/clonal-hematopoiesis-2017 . This location also contains the Gillespie stochastic simulation code used to generate all data in this manuscript , along with the data files .
By considering just the cells of the host organism , we can compute the steady state of our system from Eq ( 2 ) , and hence express the model parameters δ , d , and a in terms of the known quantities displayed in Table 1 . These expressions are shown in Table 2 , where we also enumerate the possible values of these deduced model parameters . Even for the narrow range of values reported in the literature ( Table 1 ) , we find disparate dynamics in our model . At one extreme , the average time a cell spends in the BM compartment ( 1/d ) can be less than two hours ( for s* = 100 cells and ℓ = 1 minute ) . Thus under these parameters the HSCs migrate back-and-forth very frequently between the niches and blood , and the flux of cells between these compartments over a day ( s*/ℓ ) is significantly larger than the population size . In fact , under these conditions 144 , 000 HSCs per day leave the marrow and enter the blood . With slower turnover in the PB compartment ( ℓ = 5 minutes , but still s* = 100 ) , the average BM residency time of a single HSC is eight hours , and 28 , 800 HSCs leave the bone marrow per day . At the other extreme , if the PB compartment is as small as reported in Ref . [19] ( s* = 1 cell ) , then the residency time of each HSC in the bone marrow niche is between 8 and 290 days ( for ℓ = 1 and 5 minutes , respectively ) . Under these conditions the number of cells entering the PB compartment per day is 1 , 440 and 288 , respectively . For an intermediate PB size of s* = 10 , the BM residency time is between 17 and 90 hours ( for ℓ = 1 and 5 minutes , respectively ) , and the flux of cells leaving the BM is a factor ten greater than for s* = 1 . Clonal dominance occurs when a single HSC generates a mature lineage which outweighs the lineages of other HSCs , or where one clone of HSCs outnumbers the others . The definition of when a clone is dominant is not entirely conclusive . Previous studies of human malignancies have used a variant allele frequency of 2% , corresponding to a clone that represents 4% of the population [39 , 40] . For completeness we investigate clonality ranges from 0 . 1% to 100% . In the context of disease , this clone usually carries specific mutations which may confer a selective advantage over the wildtype cells in a defined cellular compartment . The de novo emergence of such a mutant occurs following a reproduction event . Therefore , in our model with ρ = 0 , after the mutant cell is generated it is located in the PB compartment , and for the clone to expand it must first migrate back to the BM . This initial phase of the dynamics is considered in general in the next section of transplant dynamics , where a positive number S of mutant/donor cells are placed in the PB . We find ( as shown in the S1 Supporting Information ) that the expected number of these cells that attach to the BM after this initial dynamical phase is n 2 = a ( N - n * ) / N δ + a ( N - n * ) / N S = ( 1 - β ℓ n * s * ) S . ( 3 ) We then apply a fast-variable elimination technique to calculate how long it takes for this clone to expand within the host [41 , 42] . This procedure reduces the dimensionality of our system , and makes it analytically tractable . A full description of the analysis can be found in the S1 Supporting Information , but we outline the main steps and results of this procedure below . We first move from the master equation—the exact probabilistic description of the stochastic dynamics—to a set of four stochastic differential equations ( SDEs ) for each of the variables via an expansion in powers of the large parameter N [43] . We then use the projection method of Constable et al . [41 , 42] to reduce this system to a single SDE describing the relative size of the clone . This projection relies on the weak-selection assumption , i . e . 0 ≤ ε ≪ 1 . The standard results of Brownian motion are then applied to obtain the statistics of the clone’s expansion . In particular , the probability that the mutant/donor HSCs reach a fraction 0 < σ ≤ 1 of the occupied BM niches is given by ϕ ( z 0 , σ ) = 1 - e - Λ z 0 1 - e - Λ σ ξ , ( 4 ) where z0 is the initial clone size can be found explicitly from Eq ( 3 ) , such that z0 = n2/N . We also have ξ = n*/N , and Λ is a constant describing the strength of deterministic drift relative to stochastic diffusion . Concretely , we have Λ = ε N d β + d δ + β δ ( d + β ) δ = ε N ( 1 + s * n * - β ℓ ) . ( 5 ) The mean time for the clone to expand to size σ ( i . e . the mean conditional time ) is written as Tξ ( z0 , σ ) = θ ( z0 , σ ) /ϕ ( z0 , σ ) , where θ ( z0 , σ ) is given by the solution of ∂ 2 θ ( z 0 , σ ) ∂ z 0 2 + Λ ∂ θ ( z 0 , σ ) ∂ z 0 = - N B ϕ ( z 0 , σ ) z 0 ( ξ - z 0 ) , θ ( 0 ) = θ ( σ ξ ) = 0 . ( 6 ) Here B is another constant describing the magnitude of the diffusion , and is given by B = d ( d + β ) β δ 2 ξ ( d β + d δ + β δ ) 2 = β N s * s * n * − β ℓ ( 1 + s * n * − β ℓ ) 2 . ( 7 ) Although a general closed-form solution to Eq ( 6 ) is possible , it is too long to display here . Instead we use an algebraic software package to solve the second-order differential equation . A similar expression to Eq ( 6 ) can be obtained for the second moment of the fixation time , as shown in [44] and repeated in the S1 Supporting Information . The first scenario we consider is the expansion of a neutral clone ( ε = 0 ) ; i . e . how likely is it that a single cell expands into a detectable clone in the absence of selection ? It is known that the time to fixation of a neutral clone in a fixed-size population grows linearly in the system size [45] . Interestingly and importantly , in intestinal crypts this fixation is seen frequently because N = O ( 10 ) [46] . In the hematopoietic system , however , it likely takes considerably longer than this due to the relatively large number of stem cells . Solving Eq ( 6 ) with ε = 0 gives the mean conditional expansion time as T ξ ( z 0 , σ ) = N B [ ξ - z 0 z 0 log ( ξ ξ - z 0 ) + 1 - σ σ log ( 1 - σ ) ] . ( 8 ) From this solution we find that it takes , on average , 5–45 years for a neutral clone to reach 1% clonality ( ∼100 HSCs ) . Expanding to larger sizes takes considerably longer , as highlighted in Fig 2 . Therefore , clonal hematopoiesis in mice is unlikely to result from neutral clonal expansion; for a clone to expand within the lifetime of a mouse it must have a selective advantage . Neutral results for human systems are considered in the discussion . When the mutant clone has an advantage , there is always some selective force promoting this cell type . Therefore the probability of such a clone expanding is higher than the neutral case , as seen from Eq ( 4 ) . In Fig 2 we illustrate the time taken for a single mutant HSC to reach specified levels of clonal dominance for different selective advantages . Advantageous clones ( β2/β > 1 ) initially grow exponentially in time [Fig 2 ( a ) ] , and are much faster than neutral expansion ( β2/β = 1 ) . These clones can reach levels of up to 90% relatively quickly , however replacing the final few host cells takes much longer . The advantage that a mutant clone must have if it is to represent a certain fraction of the population in a given period of time can be found from Fig 2 ( b ) . For a single mutant to completely take over in two years , it requires a fold reproductive advantage of β2/β ≈ 2 [dashed lines in Fig 2 ( b ) ] . This means that the cells in this clone are dividing at least twice as fast as the wildtype host cells . To achieve 1% clonality in this timeframe , the advantage only has to be β2/β ≈ 1 . 2 . For the clone to expand in shorter time intervals , a substantially larger selective advantage is required . For example , 100% clonality in six months from emergence of the mutant requires β2/β ≈ 5 . 5 , i . e . the dominant clone needs to divide more than five times faster than the wildtype counterparts . As shown in the S1 Supporting Information , Eqs ( 4 ) and ( 6 ) are equivalent to the results obtained from a two-species Moran process . This suggests the two-compartment structure is not necessary to capture the behaviour of clonal dominance . However , the consideration of multiple compartments is required to understand transplantation dynamics , as covered in the next section . We now turn our attention to the scenario of HSC transplantation . As previously mentioned this situation is analogous to the disease spread case , with the exception that the initial ‘dose’ of HSCs can be larger than one . We first consider the case of a non-preconditioned host . We then move onto transplantation in preconditioned hosts , where all host cells have been removed .
Our model has been kept to a minimal level of biological detail to allow for parametrisation from experimental results . This has the added benefit of analytic tractability . The model is constructed under steady-state conditions , which is the case for neutral clonal expansion . However , in the case of donor-cell transplantation following myeloablative preconditioning , we are no longer in a steady state . Here we expect some regulatory mechanisms to affect the HSC dynamics , including a faster reproductive rate and a reduced probability of cells detaching from the niche . There are also possibilities for mutants to exploit or evade the homeostatic mechanisms [63] . Different mechanisms of stem cell control have recenty been considered for hematopoietic cells [64] , as well as in colonic crypts [65] . The steady state assumption is also unable to capture the different dynamics associated with ageing . For example , in young individuals the hematopoietic system is undergoing expansion . In our model there is no distinction between young and old systems . In S3 Fig we demonstrate the impact of a ( logistically ) growing number of niches . Such growth means clonal hematopoiesis is likely to be detected earlier , and therefore would increase our lower bound estimate on the number of HSCs in man . Telomere-length distributions have been used to infer the HSC dynamics from adolescence to adulthood , and have suggested a slowing down of HSC divisions as life progresses [66] . Faster dynamics in early life would lead to a higher incidence among young people , which again increases our lower bound estimate . It is also not entirely clear how to extrapolate the parameters from the reported mouse data to a human system . Here we have taken the simplest approach and appropriately scaled the unknown parameters . However , hematopietic behaviour may differ between species . For example , results of HSC transplantation following myeloablative therapy in non-human primates have shown that clones of hematopoietic cells persist for many years [67 , 68] . This could be due to single HSCs remaining attached to the niche and over-contributing to the hematopoietic system , or due to clonal expansion of the HSCs to large enough numbers such that a contributing fraction will always be found in the BM . Both of these mechanisms are features of our model: the time a cell spends in the BM is much longer than the time in the PB and can be increased further by tuning the model parameters , namely by decreasing s* or increasing ℓ . Changes to these parameters seems to have little effect on our predictions of clonal expansion , as shown in S1 and S2 Figs . Clonal extinctions are also a feature of our work , and have been identified in non-human primates [68] . A more general point to discuss is the role of hematopoietic stem cells in blood production . In our model we are only considering HSC dynamics , however it has been proposed that downstream progenitor cells are responsible for maintaining hematopoiesis [11] in mice . Hence , myeloid clonality would also be determined by the behaviour of these progenitor cells . On the other hand , an independent study found that HSCs are driving multi-lineage hematopoiesis [10] , suggesting we are correct in our approach . Again we also expect there to be variation between species in this balance of HSC/progenitor activity . With little quantitative information available , we have assumed that HSCs are the driving force of steady-state hematopoiesis across mice and humans . In conclusion , this simple mathematical model encompasses multiple HSC-engraftment scenarios and qualitatively captures empirically observed effects . The mathematical calculations provide insight into how the dynamics of the model unfold . The analytical results , which we have verified against stochastic simulations , allow us to easily investigate how parameter variation affects the outcome . We now hope to extend this analysis , incorporating further effects of disease and combining this model with the differentiation tree of hematopoietic cells .
|
Clonal hematopoiesis—where mature myeloid cells in the blood deriving from a single stem cell are over-represented—is a major risk factor for overt hematologic malignancies . To quantify how likely this phenomena is , we combine existing observations with a novel stochastic model and extensive mathematical analysis . This approach allows us to observe the hidden dynamics of the hematopoietic system . We conclude that for a clone to be detectable within the lifetime of a mouse , it requires a selective advantage . I . e . the clonal expansion cannot be explained by neutral drift alone . Furthermore , we use our model to describe the dynamics of hematopoiesis after stem cell transplantation . In agreement with earlier findings , we observe that niche-space saturation decreases engraftment efficiency . We further discuss the implications of our findings for human hematopoiesis where the quantity and role of stem cells is frequently debated .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"viral",
"transmission",
"and",
"infection",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"microbiology",
"cloning",
"physiological",
"processes",
"stem",
"cells",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"stem",
"cell",
"niche",
"molecular",
"biology",
"hematopoietic",
"stem",
"cells",
"host",
"cells",
"blood",
"cell",
"biology",
"anatomy",
"virology",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"hematopoietic",
"system",
"hematopoiesis"
] |
2017
|
Clonal dominance and transplantation dynamics in hematopoietic stem cell compartments
|
Upon attachment to their respective receptor , human rhinoviruses ( HRVs ) are internalized into the host cell via different pathways but undergo similar structural changes . This ultimately results in the delivery of the viral RNA into the cytoplasm for replication . To improve our understanding of the conformational modifications associated with the release of the viral genome , we have determined the X-ray structure at 3 . 0 Å resolution of the end-stage of HRV2 uncoating , the empty capsid . The structure shows important conformational changes in the capsid protomer . In particular , a hinge movement around the hydrophobic pocket of VP1 allows a coordinated shift of VP2 and VP3 . This overall displacement forces a reorganization of the inter-protomer interfaces , resulting in a particle expansion and in the opening of new channels in the capsid core . These new breaches in the capsid , opening one at the base of the canyon and the second at the particle two-fold axes , might act as gates for the externalization of the VP1 N-terminus and the extrusion of the viral RNA , respectively . The structural comparison between native and empty HRV2 particles unveils a number of pH-sensitive amino acid residues , conserved in rhinoviruses , which participate in the structural rearrangements involved in the uncoating process .
A key step in the life cycle of viruses is the delivery of its genome into a compartment of the host cell appropriate for its replication . This involves the recognition of specific cell surface receptors by the viral capsid and the passage of the genome across at least one membrane barrier . Enveloped viruses achieve this by fusing with cellular membranes . In non-enveloped viruses , including picornaviruses , the capsid proteins and the viral capsid as a whole must provide the machinery for the translocation of the viral genome in a process that remains poorly understood . The discovery of short , membrane altering amphipathic or hydrophobic sequences in capsid proteins , such as the VP1-N-terminus of poliovirus ( PV ) [1] and the entire VP4 in human rhinovirus 16 [2] , suggested that these peptides were involved in breaching host membranes . While crystallographic structures of various picornaviruses showed that VP4 and the N-terminus of VP1 line the interior of the capsid , biochemical experiments revealed that these peptides can be transiently exposed , demonstrating that the virions are highly dynamic and temporarily externalize otherwise internal structures via ‘breathing’ [3] , [4] . Human rhinoviruses ( HRVs ) , members of the Picornaviridae family , are the cause of about 50% of all mild infections of the upper respiratory tract . Despite being rarely life threatening , prevalence and recurrent nature of the common cold make these viruses of paramount economic importance due to the huge expenditures for medication and working days lost . Based on the complete genome sequences , HRVs were classified as three species within the genus Enteroviruses that included 74 HRV-A , 25 HRV-B and 7 HRV-C [5] . Independent from phylogeny , HRVs are also classified on the basis of receptor usage: the minor receptor group , 12 HRV-A , bind low-density lipoprotein receptor ( LDLR ) , very-LDLR ( VLDLR ) and LDLR-related protein ( LRP ) [6] , while the remaining HRV-A and HRV-B , which belong to the major receptor group , use intercellular adhesion molecule 1 ( ICAM-1 ) for cell entry [7] . Some major group HRVs might also use heparan sulphate as an additional receptor either with or without adaptation in tissue culture [8]-[10] . The receptor ( s ) binding the recently identified HRV-Cs are still unknown [11] . A characteristic trait of all HRVs is their instability at low pH , which distinguishes them from the otherwise closely related enterovirus species A to D , including the 3 polioviruses . Like all members of the Picornaviridae family , HRVs consist of a T = 1 ( pseudo T = 3 ) icosahedral capsid of about 30 nm in diameter that is built from 60 copies of each of 4 coat proteins VP1 , VP2 , VP3 , and VP4 , protecting a plus-sense single-stranded RNA genome . A prominent feature of the HRV shell is a star-shaped mesa on each of the five-fold symmetry axes that , in the case of minor group HRVs , harbors the binding sites for the LDL family of receptors [12]–[14] . The five-fold vertex is surrounded by a large depression or canyon , containing the binding site for the ICAM-1 receptor in major group HRVs [15] , [16] . At the bottom of the canyon , buried between the two ß-sheets of the VP1 core , there is a hydrophobic pocket which in some picornavirus capsid structures is occupied by elongated electron densities that have been modeled as different fatty acid cofactors , known as pocket factors . These molecules are thought to stabilize the native conformation of the virions [17]–[23] . A number of drugs with antiviral activity have been shown to bind in this pocket , displacing the pocket factor because of their higher binding affinity . Antiviral drugs bound to the pocket rigidify the capsid , preventing the required structural changes [3] , [24]–[26] . The stabilizing role of the pocket factor has been challenged by more recent data on HRV14 , showing that mutations filling the VP1 pocket have no effect on viral replication . It has been suggested that the VP1 pocket itself regulates the structural dynamics required for viral infection [27] . Although much is known about the binding of rhinoviruses to their receptors and their uptake into the cell , the mechanism by which their genomic RNA leaves the capsid , crosses the endosomal membrane and arrives to the cytosol is still enigmatic . The different dependence on receptor function and pH for initiation of RNA release in the various HRV serotypes adds an additional dimension to this problem [28] . In major group HRVs and PVs , interaction with the receptor initiates irreversible structural changes and exit of VP4 . In contrast , in the minor group viruses the low endosomal pH alone triggers these changes [29] and the receptor might rather have a stabilizing function [30] , [31] . The subviral A-particles , remaining after loss of VP4 , sediment at 135S ( compared to 150S for the native virion ) . A-particles are the dominant form of the virus found in cells early in infection . They are believed to be a necessary intermediate in the entry process and , in the case of poliovirus , have been demonstrated to be infectious [32] , [33] . With respect to native virions , A-particles exhibit changes in antigenicity and sensitivity towards protease digestion , have increased hydrophobicity and readily attach to liposomes [34] , likely through the exposure of N-terminal residues of VP1 [4] . The N-terminal segment of VP1 , possibly in conjunction with VP4 , may facilitate RNA translocation into the cytoplasm by forming a pore in the endosomal membrane . This is supported by the observation that , at comparably higher concentrations , the derived peptides alone can permeabilize membranes [2] , [35] , [36] . After release of the RNA , empty capsids ( B-particles ) sedimenting at 80S appear and polyprotein synthesis commences [37] . Similar empty particles can be produced in vitro by exposure of native virions to pH≤5 . 6 , or by incubation at 50 to 56°C in low ionic strength buffer [37] , [38] . A body of experimental data from cryo EM studies [39]–[45] addressed the low and medium resolution structures of the subviral particles of PVs and HRVs . All of these studies showed significant alterations occurring concomitant with genome release: externalization of myristoyl-VP4 and the N-terminus of VP1 , expansion of the virus particle and an iris-like movement extending the pores at the five-fold axes . In a widely accepted model , it was proposed that receptor binding to the viral capsid ( and/or low pH ) induced changes widening the channel at the five-fold axes , allowing for VP4 , the N-terminus of VP1 and the viral RNA to exit the capsid [16] , [39] , [46] , [47] . A number of cryo-EM structures were also compatible with an alternative model in which VP4 and the N-termini of VP1 exit via pores at the base of the canyon and the RNA egresses via the channel at the five-fold axis channel [40] , [41] , [43] , [44] . A recent work in PV provided additional evidences of the site of egress of VP1 at the base of the canyon by locating the binding site for the Fab fragment of an antipeptide antibody directed against the VP1 N-terminus [48] . Furthermore , new cryoEM studies of the PV 80S particles , together with the cryo-electron tomography characterization of an additional state , in which the particles were caught in the act of RNA release , suggested that the viral RNA exits through holes close to the two-fold axes not far from the site at which N-terminus of VP1 exits in 135S particles [42] , [45] . However , the low and medium resolution of the cryo-EM reconstructions were insufficient to establish neither the detailed structural changes occurred during the formation of the subviral particles nor the identification of the specific interactions involved in the stabilization of these changes . Also , until now , crystallization of subviral particles had not been achieved . Here we report the 3 . 0 Å resolution crystal structure of the HRV2 empty particle . This structure and its comparison with the native virus shed light on the structural rearrangements produced in the viral capsid during RNA uncoating , unveiling the interactions involved in the pH-triggered conformational changes .
The crystal structure of the HRV2 80S empty particle was determined at 3 . 0 Å resolution by molecular replacement using 15-fold non-crystallographic symmetry averaging , starting with the phases corresponding to the native HRV2 structure [18] . The resulting averaged maps showed well-defined density and allowed rebuilding of most of the conformational changes that had occurred on transition from native HRV2 capsid to the 80S B-particle . Analysis of the electron density confirmed the absence of VP4 , as predicted from comparison with the poliovirus uncoating scenario [44] and the cryo-EM structure of HRV2 previously determined at 15 Å [40] . Externalization of N-terminal residues of VP1 in the empty capsid is compatible with the lack of ordered electron density up to position 1062 ( VP1 , VP2 and VP3 proteins are numbered starting with 1000 , 2000 and 3000 respectively ) . The overall icosahedral organization of the 80S particles is similar to that of the native virus . However , the 80S capsid has an average diameter of 326 Å ( calculated using VIPERdb [49] ) . When compared with the corresponding 314 Å diameter of the native virions , this implies an average expansion of 3 . 8% ( Figure 1A and Video S1 ) , which is in accordance with the observations derived from electron microscopy studies for this virus and other related picornavirus particles [39]–[41] , [43] , [45] , [50] . As the empty shell is derived from the native structure , this expansion implies a reduction in the thickness of the protein layer . Indeed , the average thickness of the HRV2 80S particle shell ( measured as the difference between the averaged outer and inner radius obtained from VIPERdb ) is 51 Å , in contrast to the 56 Å found in the native virion ( Figure 1A and Video S1 ) . Moreover , the 80S capsid appears smoother than the native capsid; on average , the canyon in the 80S particle is approximately 5 Å smaller in depth and width . The overall shape of the protomer and the disposition of the VP1-3 subunits are maintained in the native , the 135S A-particles ( Pickl-Herk et al . , in preparation ) and the 80S B-particles , despite of a number of changes have occurred . Structural comparisons of the native and the 80S protomers using SHP program [51] resulted in a root-mean-square deviation ( r . m . s . d . ) of 0 . 93 Å for the superimposition of 625 Cα atoms . The superimpositions of the individual VP1 , VP2 and VP3 proteins in the native and the 80S structures gave r . m . s . d . values of 1 . 17 Å , 0 . 55 Å and 0 . 81 Å for the superimposition of 199 , 219 and 212 equivalent Cα atoms , respectively . The biggest changes were concentrated in the N-terminal regions and in the loops connecting the strands of the ß-barrel . Thus , new superpositions were performed for each VP with program Lsqkab [52] , using the ß-barrel core as a guide ( Figures 1B–D ) . VP1 . Superposition of the ß-barrels of the two VP1 structures allows the definition of two regions with different levels of resemblance: i ) the ß-barrel and ii ) the region comprising the CD loop , the GH loop and the C-terminal end of the protein ( Figures 1B and 2 ) . Breaking the VP1 molecule into these independent rigid bodies allowed the superimposition of 206 Cα atoms with a r . m . s . deviation of 0 . 76 Å . Thus , the rearrangement of VP1 in the 80S empty capsid with respect to the native virion can be explained by a hinge-type movement , consisting in an approximately 5 . 6° outward rotation of one of these two regions with respect to the other ( Table S1 ) . The pivot of this rotary motion is located in the ß-strand I' region ( close to residue Lys1243 ) ( Figures 1B , 2A and 2B ) . The conformation of other loops has also changed to some extent: the EF-loop is displaced towards the base of the barrel and the short αB helix , contained within , is moved by about 3Å ( Figure 1B ) ; the BC loop at the five-fold axis appears highly flexible , as indicated by the highest B factors observed in the region ( 102 . 9 Å2 ) in comparison with the average ( Table 1 ) ; the DE loop residues , from Asp1135 to Gly1137 , also at the five-fold axis , are displaced upwards by 2 . 5 Å ( Figure 1B ) . Finally , the N-terminal 61 residues of VP1 were not seen in our density maps , indicating that this region is disordered in the crystal structure of the 80S empty capsid . The first amino acid residue seen in the density ( Arg1062 ) is located at the capsid interior , just below the canyon floor . VP2 . Superposition of the VP2 ß-barrels of the structures of native virus and 80S empty capsid also highlights the conservation of the overall folding , with only some local rearrangements ( Figure 1C ) . In particular , the N-terminal ß-hairpin , which participates in the inter-pentameric contacts , is displaced by about 3 Å and twisted by 20° around the axis given by the ß-strand A2 . The C-terminus also diverges from the native structure , especially from residue Ser2254 onwards , just after the strand I' . A number of rearrangements are also observed in loops connecting the ß-strands of the VP2 barrel; residues Leu2231-Thr2239 within the HI-loop , at the three-fold axis , are moved by more than 3 Å away from their position in the native structure . The loop preceding the B strand , including residues from Ile2050 to Ser2059 , at the two-fold axis , seems to be highly flexible , with averaged B-factors of 127 . 8 A2 . VP3 . Superposition of the VP3 barrels shows that , as in VP2 , the conformation of most of the VP3 loops is conserved between the native and the 80S structure . Only a rigid body displacement of the VP3 N-terminus with respect to the barrel is observed ( Figure 1D ) . The N-terminal extension of VP3 runs nearly parallel to its position in VP3 in the native virion , at a distance of 3Å ( measured as the Cα- Cα distance between Gly3001 residues of the superposed structures; Figure 1D ) . This deviation gradually diminishes until the chains converge at residue Asp3050 , just after helix αA . The smooth decrease in distance allows maintenance of the intactness of the VP3 five-stranded ß-tube at the inner face of the five-fold symmetry axes while the ß-barrels of VP2 and VP3 are displaced with respect to VP1 ( Figures 2E and 3 and Video S1 ) . The VP3 C-terminus also appears slightly shifted between strands ßI and ßI' ( residues from Gly3214 to Cys3219 ) and from Met3222 onwards . The VP3 GH loop is displaced by about 1–2 Å in its first half , adopting a new conformation between residues Ser3179 and Tyr3187 ( with a maximum distance of 5 . 2 Å at position Arg3182; Figure 1D ) . Finally , the BC and FG loops at the three-fold axis and the EF loop , near the two-fold axis , also present local rearrangements compared to the native structure ( Figure 1D ) . Comparison of the overall positions of each VP subunit in the 80S structure to those corresponding to the native particle reveals that the relative positions of VP2 and VP3 with respect to each other are maintained . However , in the 80S particle these subunits are displaced with regard to the VP1 ß-barrel by between 4 and 6 Å towards the icosahedral three-fold axis . This is easily seen in the superposition of both protomers when the ß-barrel of VP1 is used as a reference ( Figure 2E and Video S2 ) . In fact , the displacement of VP2 and VP3 is parallel to the displacement seen at the αA helix and the C-terminal region of VP1 ( Table S1 ) . Moreover , there is a disulfide bridge between residues Cys2229 and Cys3120 , at the VP2-VP3 interface ( Figure 2E ) . This S-S bond , which was also present in the native HRV2 structure ( PDB code: 1FPN ) , links VP2 and VP3 , assisting the concerted shift of these subunits . Thus , the hinge-type movement of VP1 and the parallel displacement of VP2 and VP3 facilitate the expansion of the protomer and , in consequence , of the outer limits of the pentamer ( Figure S1 ) . Together they led to a bigger , but thinner , empty capsid ( Figure 1 and Video S1 ) . As stated above , the hinge-type movement of VP1 has its pivotal center in the region around Lys1243 , which is located just below the ß-barrel , near the hydrophobic pocket in VP1 . In the native virion , this pocket was filled with a density which was interpreted by the presence of a 12-carbon atoms long fatty acid ( Figure 4B; [18] ) . In the structure of the 80S particle , the density within the pocket is absent and the cavity appears collapsed . The GH loop of VP1 has a conformation resembling the closed structure observed in HRV14 and HRV3 [15] , [53] , with the side chain of Met1213 in an extended conformation across the pocket factor binding site ( Figure 4A ) . Surprisingly , the 80S empty particle possesses pores or breaks crossing the capsid shell ( Figure 5A ) . These pores , absent in the HRV2 native virions ( Figure 5B ) , connect the canyon floor with the internal surface of the particle , close to residues Arg1062-Glu1068 . This region corresponds to the first visible amino acids at the VP1 N-terminus and includes the acidic residues Asp1063 and Glu1064 ( Figure 5C ) . The glutamic acid at position 1064 is strictly conserved in all HRVs except for the major group rhinoviruses HRV3 and HRV14 , in which Asp1063 is replaced by Ser . ( See alignment in [5] ) . The opening of these pores at the canyon floor is formed by residues Met1104 , Ala1105 and Glu1106 ( within the αA helix ) of one VP1 subunit and Tyr1159 and Gln1162 ( in the αB helix ) of the adjacent VP1 . The walls of the channel are formed by three different regions contributed from two neighboring VP1 molecules ( residues Ser1069 , Phe1070 , Leu1071 , Arg1073 at the N-terminal loop , residues Ile1107 and Lys1110 from the αA helix and residues Ser1163 , Gly1164 and Thr1165 from the αB helix ) and by the C-terminus of VP3 ( residues Ala3223 , Asp3225 ) . Capsid expansion implies significant rearrangements of the interactions at the interfaces between protomers , resulting in an overall loss of intra- and inter-protomer contacts in the HRV2 empty particle ( Table S2 and Figure S2 ) . A total of 197 interactions stabilized the intra-pentamer interfaces of the native HRV2 particles and only 95 of these contacts are conserved in the 80S particle . Twenty new contacts are formed , replacing some of the stabilizing interactions within the HRV2 80S pentamer . However , the overall contacting interfaces are weakened ( Table S2 ) . Some of the residues involved in the re-organization of the intra-pentamer contacts , in particular those located at the five-fold symmetry axes channel , are sensitive to a reduction in pH and conserved in the acid-labile rhinoviruses ( Figure 3 ) . The five-fold channel of the 80S particle appears slightly wider at its outer surface when compared to the equivalent structure in the virion . This is due to the structural rearrangements of loops BC , DE and HI in the 80S VP1 structure ( Figures 1B , 3 and S1 ) . The iris-type movement described in the cryo-EM structure of HRV2 80S particle [40] is generated by these loop changes ( Figure 3 and S1A ) . In native HRV2 virions , this channel was surrounded by a ring of five symmetry-related aspartate residues ( Asp1135 , within the DE-loop ) , interacting with five solvent molecules surrounding a peak of density at the five-fold axis that was interpreted as stemming from a Ca2+ ion with partial occupancy [18] . The involvement of the DE loop of VP1 in the interaction with metal ions on the five-fold axis was found in the structures of all other rhinovirus analyzed . These cations were predicted to play a role in regulation of rhinovirus stability , although no conformational changes were observed in EGTA-treated virus structures [54] . In the 80S structure , the side chain of Asp1135 appears mostly disordered and no extra density is found to position any ion and/or solvent molecule in the region . Below the Asp1135 ring , there are five symmetry-related histidines ( His1173 , at the FG-loop ) . In the structure of the native virions , these histidine residues were bridged to each other through poorly ordered metal ions or solvent molecules [18] . In the 80S particle the DE loop is shifted towards the FG loop , thereby approaching a new histidine ring ( built from His1138 ) to this area . In fact , this second group of histidines ( His1138 ) appears to interact with His1173 of the neighboring VP1 subunit around the five-fold axis , closing the ring at this level ( Figure 3A ) . The inner part of the five-fold channel is occupied by a ß-tube , of about 5 . 5 Å in diameter , formed by five symmetry-related N-termini of VP3 ( residues from Gly3001 to Ser3010 ) . In the 80S empty capsid , the VP3 ß-tube maintains not only its native conformation but also its exact position in the five-fold channel and most of the interactions with the surrounding residues as in the native virions ( Figures 3 and S1B and Video S1 ) . The intactness of the VP3 ß-tube between native and empty HRV2 capsids accounts for the displacement observed at the VP3 N-terminal extension when the ß-barrels are superimposed ( Figure 1D ) . In native HRV2 particles , a total of 99 interactions were established between pentamers . Only 36 of these contacts are conserved in the 80S particle , while 21 new contacts are formed , partially replacing the lost interactions ( Table S2 ) . About 30% of the 63 missing contacts involved the N-terminal end of VP1 , which in the 80S structure is disordered up to position 1062 ( Table S2 and Figure S2 ) . In native HRV2 virions , the VP1 N-terminus was involved in extensive interactions with VP2 and VP3 . Some of these contacts were mediated by acidic residues that are highly conserved in acid-labile rhinoviruses ( Figure 3C ) . In addition , this region , together with the ß-hairpin of VP2 ( strands A1–A2 ) , participated in a crucial element stabilizing the pentamer interface: an extended antiparallel ß-sheet , comprising the VP2 ß-hairpin of one pentamer , sandwiched between the four-stranded CHEF sheet of the VP3 ß-barrel and the VP1 N-terminus of the adjacent pentamer . This pentamer assembly feature is found in all picornaviruses [55] . In the structure of the 80S empty capsid , the VP1 N-terminus is disordered and the ß-hairpin of VP2 is considerably displaced , making this ß-sheet shorter and , thus , weakening the inter-pentamer contacts . In addition to the re-organization of the inter-pentamer network , the αA helices of the two adjacent VP2 subunits , which were in close contact in the native capsid ( Figure 6B and Table S2 ) , become separated by 10Å in the 80S capsid ( Figures 6A and C ) . Moreover , in the empty capsid , the AB loop of VP2 ( residues from Ile2050 to Ser2059 ) is also displaced and appears to be highly flexible ( see above ) . Altogether , these changes disrupt 13 interactions of the native capsid and produce a break in the interface between pentamers that crosses the capsid along the two-fold symmetry axes . The approximate size of this break in the HRV2 empty capsid is about 10×10 Å , having its length limited by the flexible most C-terminal residues of both VP2 subunits . Assuming the flexibility of the AB loop and VP2 C-terminus , the capsid break would then extend up to the ß-barrel of the adjacent VP3 subunits , allowing the expansion of the channel dimensions to approximately 25×10 Å ( Figure 6C ) . Indeed , it is conceivable that it temporally expands even more during egress of the viral RNA .
Early studies in major group rhinoviruses suggested an allosteric competition between the pocket factor , binding below the canyon floor , and ICAM-1 , binding to the outer face of the canyon floor [47] . According to this view , the expulsion of the pocket factor would allow stronger binding of ICAM-1 thus inducing a hinge movement near the pocket . This would lead to an opening at the five-fold axes , which was suggested to be the way for the RNA to exit the capsid [60] . In contrast , in the minor group virus HRV2 , the receptor binding site does not overlap the pocket , and so LDLR does not enter into competition with the pocket factor [12] , [14] . Instead , the viral particles travel with the lipoprotein receptor through clathrin-coated pits to the endosome , where viral uncoating is triggered by low pH [29] , [30] , [62]–[65] . Among other models , and by analogy with the major group rhinoviruses , it was thought that in minor group HRVs the RNA genome may also be externalized through the five-fold axis of the capsid , which would be connected to a pore on the endosomal membrane , formed by five N-termini of VP1 and five VP4 molecules [28] . The X-ray structure of the HRV2 80S empty capsid shows indeed a hinge movement around the VP1 pocket . However , and in contrast to previous predictions , this movement does not induce any widening of the icosahedral five-fold axis but facilitates the expansion of the capsid protomer , forcing a reorganization of the interfaces between protomers and resulting in the opening of new channels in the capsid core that enhance the permeabilization of the virion . These new breaches in the capsid could then act as gates for the externalization of the VP1 N-terminus and the extrusion of the viral RNA . The 3D-structures of two representative members of the HRV-A species ( HRV2 and HRV16 ) are available at resolutions higher than 2 . 6 Å [18] , [19] . In the structure of native HRV2 , the N-terminus of VP1 was disordered for the first twelve amino acids and weak density was observed for the positioning of Glu1013 and Val1014 [18] . The main chain amino group of Leu1015 is hydrogen-bonded to the side chain of Asp1063 , located at the beginning of the α-helix and highly conserved in HRV-A and -B species ( Figure 3C ) . In addition , the main chain of Asp1063 is hydrogen-bonded to the side chain of the strictly conserved Glu1013 residue ( Figure 3C ) . The electron density for Glu1013 residues was poorly defined in the structure of native HRV2 , but is clearly visible in HRV16 , the only rhinovirus whose VP1 N-terminus could be traced entirely . In the HRV16 structure , Glu1013 is located at the C-terminus of an amphipathic α-helix which stacks on the outside of a ten-stranded β-barrel formed by five symmetry-related VP4 proteins , lying on the interior of the virus particle along the icosahedral five-fold axis [19] ( Figure 3B ) . In the HRV16 virion the side chain of Glu1013 is hydrogen bonded with the main chain amino groups of Glu1063 and Glu1064 and with the side chain of the conserved His3029 of VP3 ( Figure 3C ) . Additional conserved acidic residues in the region , Glu1048 and Glu1052 , form polar interactions with Lys3217 and Lys2052 side chains of VP3 and VP2 , respectively . All these interactions , which help to stabilize the conformation of the VP1 N-terminus at the capsid interior of native virions , are missing in the 80S structure , where the first amino acid seen in the electron density is Arg1062 ( Figure 3A ) . The pore observed at the base of the canyon crossing the capsid appears to be a leftover of the exit site of the VP1 N-terminus ( Figure 5C ) . The structure of the 80S particle shows that residues Asp1063 and Glu1064 are located at the internal surface opening of this pore . However , no ordered density was seen , neither within the pore nor in its vicinity , to build additional residues extending from Arg1062 . The absence of ordered density could be explained because in our 80S particle , obtained by heating the virus at 56°C , the VP1 N-terminus was not extruded but it remains disordered in the capsid interior . In fact , the lack of hydrophobicity of the 80S particle is in favour of this [34] . Furthermore , the size of the pore , as seen in our 80S structure , appears too narrow for accommodating a polypeptide chain , although it might temporarily widen , in particular , on heating . It must be taken into account that the HRV2 80S empty capsid represents the relaxed state of the capsid after VP4 and the viral genome have been extruded . The outer surface of this pore is surrounded by residues Met1104 , Ala1105 , Glu1106 , Tyr1159 and Gln1162 , located at the canyon floor , along the line connecting the three-fold and five-fold axes . This site is not totally coincident with the putative exit site of the PV VP1 N-terminus , as hypothesized on the basis of the cryo-EM reconstructions of 135S-derived particles [43] , [48] . The five-fold axis has been hypothesized as a port of exit for the RNA during infection since the first structures of mature rhinovirus and poliovirus were determined [15] , [23] . The low-resolution cryo-EM structure of the HRV2 80S particle also suggested an iris-type of movement of VP1 that would open a passage of ∼10Å diameter through the five-fold axis [40] . In fact , the crystal structure of the 80S empty capsid shows a small widening at the outer surface of the five-fold channel ( Figures 5A , 5B and S1A ) . However , the constriction observed at the five-fold interior , due to the presence of the VP3 β-tube , is maintained in both native virions and in the empty 80S particles ( Figures 3A , 3B and S1B ) . The presence of this β-plug constriction makes RNA egress through at that channel very unlikely . In contrast to the intactness of the protein organization at the 5-fold interior of the empty 80S particle , the reorganization of the inter-pentamer interactions due to capsid expansion induces a shift by about 5 Å to the VP2 helix αA0 , located at the two-fold axis , contributing to the formation of the biggest holes that are seen in the structure of the 80S particle ( Figure 6A and Video S1 ) . The dimensions of these channels strongly suggest that they could serve as routes for the egress of the viral RNA out of the capsid . In fact , some recent observations also point to this conclusion: a structure of a poliovirus uncoating intermediate “caught in the act of RNA release” , determined by cryo-electron tomography [42] shows the externalization of the viral genome via one of the holes , close to the two-fold axis . Moreover , inspection of the electrostatic potential at the internal surface of the 80S two-fold axis ( Figure 6D ) showed that the channel is mostly electronegative but contains small electropositive regions at the inner and outer parts of the channel . The electronegative nature of the channel would facilitate the extrusion of the RNA molecule , traversing by floating away from the repulsive walls of the channel . Similar situations have been described in other proteins involved in translocation of nucleic acids [66] , [67] . This electronegative surface is mainly contributed by the main chain oxygens of different amino acids ( Figure 6D ) and would remain essentially unchanged upon acidification .
HRV2 was grown in HeLa-H1 cells in suspension cultures and purified as previously described [12] . To obtain empty HRV2 particles , a suspension of HRV2 ( 0 . 03 mg/ml ) in 50 mM Tris ( pH 8 . 0 ) was heated to 56°C for 12 min . The quality of the particles obtained was checked by negative staining electron microscopy , showing a highly homogeneous material . Samples were then concentrated using Centricon 100K tubes ( Millipore ) to a final concentration of 3 mg/ml . The initial low concentration was found to be necessary to avoid aggregation and minimize the formation of A-particles . Crystals were obtained by the vapor diffusion method in hanging drops at room temperature by mixing equal volumes of 80S particles ( 3 mg/ml ) and a reservoir solution containing 0 . 3 M to 0 . 6 M sodium acetate with a pH range between 6 . 5 and 8 . 0 and containing 5% glycerol ( vol/vol ) as an additive . Crystals were transferred to a cryo-protecting solution containing 20% glycerol in the crystallization buffer and incubated 1 min prior to cooling by immersion in liquid nitrogen . Ten data sets up to 3 . 0 Å resolution were collected from different crystals using synchrotron radiation at the ESRF , Grenoble , France ( beamline ID23-1 ) and the SLS , PSI , Villigen , Switzerland ( beamline X06DA ) , using a Mar-Mosaic 225 charge-coupled-device detector in both cases . Diffraction images were processed using MOSFLM [68] and internally scaled with SCALA [52] ( Table 1 ) . Crystals of HRV2 80S empty capsids , belonging to space group I222 and with unit cell parameters of a = 313 . 9Å , b = 357 . 8 Å c = 383 . 1 Å , were closely related with those of the native HRV2 [18] ( a = 308 . 7 Å , b = 352 . 2 Å c = 380 . 5 Å ) , showing only a small enlargement in all three axes . In these I222 crystals , three icosahedral two-fold axes of the viral particle coincide with the three crystallographic two-fold axes , leaving ¼ of a virus particle ( 15 protomers ) in the crystal asymmetric unit . Initial phases were obtained after a rigid body fitting of the coordinates of the native HRV2 structure ( PDB:1FPN ) into the new unit cell , using the program CNS [69] with 15-fold non-crystallographic symmetry constraints . After 10 cycles of rigid body refinement , considering first the whole protomer ( VP1 , VP2 , and VP3 proteins ) as a unique body , plus 10 new cycles , considering VP1 , VP2 and VP3 as independent bodies , the R factor decreased from 54 . 1% till 39 . 7% for data in the resolution shell 50 . 0−4 . 0 Å . Similar results were obtained by using as an initial model the coordinates of the native HRV2 structure previously fitted into the 15 Å cryo-electron microscopy maps of the HRV2 80S particles by using URO [70] and the positioning of the new model into the HRV2 80S unit cell , followed by a rigid body refinement with CNS , considering VP1 , VP2 and VP3 as independent bodies . At this point a 2Fo-Fc electron density map was calculated to 4 . 0 Å resolution . Inspection of this initial map showed that the first 60 residues of VP1 were totally disordered and these amino acids were removed from the model prior to the calculation of the initial phases . Cycles of 15-fold non-crystallographic averaging and solvent flattening with the program DM [71] were used to refine and extend the initial phases from 6 . 0 to 3 . 0 Å resolution . The averaging and solvent masks used covered the whole asymmetric unit . The resulting density allowed us to determine most of the structural differences between native HRV2 and the 80S empty capsid and model rebuilding was started by using the graphic program Coot [72] . Refinement was performed with CNS with non-crystallographic symmetry constraints and bulk solvent correction , using data in the resolution shell 50 . 0−3 . 0 Å . New maps were calculated and iteratively improved by density modification cycles , using updated averaging and solvent masks , to a final correlation coefficient of 0 . 93 . Iterative positional and temperature refinement using CNS was alternated with manual model rebuilding with Coot . The final refinement statistics are summarized in Table 1 . The coordinates and structure factors have been deposited in the Protein Data Bank ( code 3TN9 ) . Figures were drawn and rendered with PyMol [73] . Electrostatic potential surfaces were calculated with the program APBS [74] . Animations were produced by using the Chimera visualization software package [75] .
|
Human Rhinoviruses ( HRVs ) , members of the Picornaviridae family , are small non-enveloped viruses possessing an icosahedral capsid that protects the single-stranded RNA genome . Although much is known about their binding to cell receptors and their uptake into the host cell , the mechanism by which their genomic RNA leaves the capsid and arrives to the cytosol to initiate replication is poorly understood . In HRV2 , a member of the minor group HRVs , upon binding to lipoprotein receptors ( LDL-R ) on the cell surface virions are taken up into vesicles and directed to early endosomes . The low pH conditions found in the endosome , and not the binding to LDL-R , catalyze the delivery of the viral genome . The crystal structure of the HRV2 empty particle , representing the last stage of the uncoating process , unveils the structural rearrangements produced in the viral capsid during the externalization of the VP1 N-terminus and the delivery of the genomic RNA . We propose that RNA exit occurs through large capsid disruptions that are produced at the particle two-fold symmetry axes . Our data also suggests that the VP1 N-terminus would be externalized through a new pore , opening at the canyon floor .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viral",
"entry",
"viral",
"transmission",
"and",
"infection",
"virology",
"subviral",
"particles",
"biology",
"microbiology",
"viral",
"structure",
"viral",
"load"
] |
2012
|
Insights into Minor Group Rhinovirus Uncoating: The X-ray Structure of the HRV2 Empty Capsid
|
As in many species , gustatory pheromones regulate the mating behavior of Drosophila . Recently , several ppk genes , encoding ion channel subunits of the DEG/ENaC family , have been implicated in this process , leading to the identification of gustatory neurons that detect specific pheromones . In a subset of taste hairs on the legs of Drosophila , there are two ppk23-expressing , pheromone-sensing neurons with complementary response profiles; one neuron detects female pheromones that stimulate male courtship , the other detects male pheromones that inhibit male-male courtship . In contrast to ppk23 , ppk25 , is only expressed in a single gustatory neuron per taste hair , and males with impaired ppk25 function court females at reduced rates but do not display abnormal courtship of other males . These findings raised the possibility that ppk25 expression defines a subset of pheromone-sensing neurons . Here we show that ppk25 is expressed and functions in neurons that detect female-specific pheromones and mediates their stimulatory effect on male courtship . Furthermore , the role of ppk25 and ppk25-expressing neurons is not restricted to responses to female-specific pheromones . ppk25 is also required in the same subset of neurons for stimulation of male courtship by young males , males of the Tai2 strain , and by synthetic 7-pentacosene ( 7-P ) , a hydrocarbon normally found at low levels in both males and females . Finally , we unexpectedly find that , in females , ppk25 and ppk25-expressing cells regulate receptivity to mating . In the absence of the third antennal segment , which has both olfactory and auditory functions , mutations in ppk25 or silencing of ppk25-expressing neurons block female receptivity to males . Together these results indicate that ppk25 identifies a functionally specialized subset of pheromone-sensing neurons . While ppk25 neurons are required for the responses to multiple pheromones , in both males and females these neurons are specifically involved in stimulating courtship and mating .
Ever since the identification of the first pheromone , Bombykol , as the sexual attractant of the silkmoth more than fifty years ago [1] , the mechanisms underlying the detection of pheromones and their regulation of animal behavior have been an important area of inquiry . The expanding understanding of the molecular basis of pheromone detection has been aided by studies in Drosophila melanogaster , a species in which males perform a series of highly stereotyped behaviors toward females eventually culminating in mating [2] , [3] . A number of pheromones that modulate male courtship have been identified [4] , [5] . 7 , 11-heptacosadiene ( 7 , 11- HD ) and 7 , 11-nonacosadiene ( 7 , 11-ND ) are the major excitatory compounds selectively produced by mature females , while 7-Tricosene ( 7-T ) and the volatile cis-Vaccenyl acetate ( cVA ) are produced by mature males and inhibit male-male courtship . While olfaction is involved in the inhibition of courtship by cVA ( reviewed in [6] ) and stimulation of courtship by unknown fly odors [7] , as well as by food odors [8] , most known Drosophila pheromones are low volatility cuticular hydrocarbons believed to be detected by direct contact with gustatory organs [9] . Recently , several laboratories have independently reported that three members of the DEG/ENaC family of ion channel subunits , ppk25 , ppk23 and ppk29 , are required for the gustatory detection of pheromones that modulate male courtship behavior [10]–[14] . ppk23 expression marks a subset of gustatory neurons , two per chemosensory hair , both of which also express fruitless ( fru ) , a key transcription factor that regulates sexually dimorphic development of neurons involved in sex-specific behaviors ( reviewed in [15] ) . Importantly , ppk23-expressing cells respond to pheromones [13] , [14] , and the two cells present in a single chemosensory hair detect distinct compounds [13] . F cells ( female-sensing ) respond to female stimulatory pheromones , while M cells ( male-sensing ) respond to inhibitory male pheromones . ppk25 is also required for courtship behavior , but unlike ppk23 , ppk25 is only expressed in one fru-positive gustatory neuron per chemosensory bristle . Furthermore , while loss of ppk23 function decreases courtship of females and also increases courtship directed at other males [13] , [14] , mutations in ppk25 decrease male courtship of females , but do not increase courtship of other males [10] , [16] . Together , these results raised the possibility that ppk25 neurons represent a functionally specialized subset of pheromone-sensing neurons [17] . Here , we show that ppk25 specifically marks the F cell subset of ppk23-expressing cells and is required for their response to stimulatory female-specific pheromones . In contrast , ppk23-expressing M cells do not express ppk25 or require ppk25 function to detect inhibitory male pheromones . Furthermore , we show that the function of ppk25 is not restricted to gustatory detection of female-specific pheromones . ppk25 and ppk25-expressing neurons are also required for stimulation of courtship by pheromones present on immature males [18] , and by pheromones present on males of the Tai2 strain [19] . Finally , we show that in addition to regulating male courtship behavior , ppk25 and ppk25-expressing gustatory neurons also regulate female mating behavior , suggesting their involvement in detection of male pheromones that stimulate female receptivity .
To evaluate the ligand specificity of ppk25 cells , expression of the genetically-encoded calcium indicator , G-CaMP3 [20] , was targeted using the ppk25-Gal4 driver [10] . Single bristles on the front legs of both males and females were stimulated with two female pheromones that had been previously shown to stimulate the F ( female-sensing ) subset of ppk23-expressing cells ( 7 , 11-HD and 7 , 11-ND ) , and three compounds produced by males that stimulate M ( male-sensing ) cells ( 7-tricosene ( 7T ) , 7-Pentacosene ( 7P ) and cVA ) [13] . As shown previously for F cells [13] , ppk25-expressing cells in both males and females showed robust calcium responses to the female pheromones , 7 , 11-HD and 7 , 11-ND , but not to the male compounds , 7T , 7P or cVA ( Fig . 1A ) . Importantly , this response requires ppk25 , as ppk25 null mutants no longer responded to the female cues and targeted expression of ppk25 in mutants rescued this defect ( Fig . 1A ) . To confirm that ppk25 is required in cells that detect female pheromones but not in those that detect male pheromones , ppk23-Gal4 was used to drive expression of G-CaMP3 in all pheromone-sensing cells in a ppk25 null mutant background . As described previously for flies with normal ppk25 [13] , ppk23-Gal4 labeled two cells under each bristle in ppk25 mutants . However , while one of these cells responded specifically to male compounds as previously described for M cells [13] , the second cell , did not respond to female compounds as expected of F cells ( Fig . 1B ) . Thus , ppk25 is essential for the recognition of courtship-stimulating pheromones produced by females but not of courtship-inhibiting pheromones produced by males . To test whether ppk25 is essential for behavioral responses to individual pheromones , control and ppk25 mutant males were paired with oenocyte-lacking ( oe- ) flies painted with single cuticular hydrocarbons in courtship assays . In oe- flies , the pheromone-producing cells called oenocytes have been genetically ablated , such that only residual hydrocarbons are present on their cuticle . oe- individuals therefore serve as pheromone-blank flies to which a single synthetic pheromone can be added to test its effect on male courtship [21] . For better comparison with previous studies on the effects of ppk23 mutations [13] , the number of wing extensions in a 20-minute observation period was used as a measure of male courtship . As observed previously [13] , painting oe- females with 7 , 11 HD , the excitatory female-specific pheromone , increased the levels of courtship displayed by control males . In contrast , ppk25 mutant males were not affected by the presence of 7 , 11 HD , and this stimulatory effect was restored by targeted expression of ppk25 ( Fig . 2A ) . In contrast to the inability of ppk25 mutants to detect stimulatory female pheromones , painting oe- males with 7-T inhibited wing extensions of ppk25 mutant and normal males to similar extents ( Fig . 2B ) . Thus , ppk25 mutant males respond to male cuticular hydrocarbons but behave as though they are selectively blind to female cuticular hydrocarbons . These data demonstrate that ppk25 is specifically required for the stimulatory effect of 7 , 11 HD , the major female taste pheromone , but is not required for detection of the major male inhibitory taste pheromone , 7T . In addition to decreasing male courtship toward females , loss of ppk23 also increases male courtship toward other males and both phenotypes are efficiently rescued by expressing ppk23 in ppk23-Gal4 cells [13] , [14] . To test whether the ppk25-positive cells are involved in both courtship defects of ppk23 mutants , we rescued ppk23 function exclusively in ppk25-Gal4 cells in a ppk23 mutant background . Male flies expressing functional ppk23 only in ppk25 cells showed normal courtship toward females but , like ppk23 mutant males , displayed abnormally high levels of male-male courtship as measured using the courtship index ( CI , the percentage of a ten minute observation time during which the male is courting [16] ) ( Fig . 2C ) . These experiments show that ppk25 cells are required for normal recognition of female-specific pheromones and for their stimulatory effects on male courtship whereas the ppk23-positive , ppk25-negative cells are required for recognition of inhibitory male pheromones . Given the requirement of ppk25 for detection of female stimulatory pheromones but not male inhibitory pheromones , we investigated whether ppk25 might also be required for detection of stimulatory pheromones that are not specific to the female . Immature Drosophila of either sex , while lacking adult female-specific pheromones , efficiently elicit courtship from adult males , most likely as a result of the unusual long-chain hydrocarbons present on their cuticle [18] . We therefore tested whether ppk25 is needed for male courtship of immature males ( Fig . 3 ) . As expected , males with one functional copy of ppk25 showed strong courtship toward young males . In contrast , ppk25 homozygous mutant males performed very little courtship as measured by either CI or fraction of males initiating courtship , but courtship was restored in ppk25 mutants when ppk25 expression was driven with the ppk25-Gal4 driver ( Fig . 3A ) . As shown previously [10] , mutations in ppk25 do not cause a generalized defect in behavior since the Total Behavior Index ( TBI - the fraction of the observation time the male spends performing any observable behavior: courtship , walking or preening ) was not affected in ppk25 mutants ( Fig . 3A ) . To further test the role of ppk25 in courtship of immature males and to identify the type of cell involved , we knocked-down ppk25 mRNA using targeted expression of a UAS-ppk25-RNAi transgene [10] with Poxn-Gal4 , a driver expressed specifically and at high levels in all gustatory neurons [22] . In control males , Poxn-Gal4 drove expression of either GFP , or of an RNAi that targets an unrelated gene ( CG13895 ) . Compared to control males , males with gustatory neuron-specific knockdown of ppk25 courted young males at significantly reduced levels , as reflected by a reduction in both CI and fraction of initiating males , while the TBI remained unchanged ( Fig . 3B ) . Together , these results indicate that ppk25 function in gustatory neurons marked by expression of ppk25-Gal4 , is required for activation of courtship not only by female-specific pheromones [10] , but also by pheromones present on immature males . In addition to females and immature males , males of the naturally occurring Tai2 strain of Drosophila melanogaster also stimulate male courtship , likely as a result of their unusual pheromone profile [19] , [23] . As expected , control males displayed high levels of male-male courtship when paired with Tai2 males ( Fig . 4A ) . In contrast , courtship of Tai2 males was largely eliminated in ppk25 mutant males and rescued by targeted expression of ppk25 using ppk25-Gal4 ( Fig . 4A ) . Finally , knockdown of ppk25 in all gustatory neurons using Poxn-Gal4-driven RNAi results in a similarly severe decrease in male courtship of Tai2 males ( Fig . 4B ) . Together , these results show that courtship stimulation by Tai2 males requires ppk25 in the subset of gustatory neurons defined by expression of ppk25-Gal4 . One of the significant differences in the pheromone profile of Tai2 males is the elevated level of 7-Pentacosene ( 7P ) [19] , [23] , which can stimulate male courtship behavior [24] and may therefore underlie courtship stimulation by Tai2 males . We therefore tested whether ppk25 is required for the stimulation of courtship by 7-P . 7-P perfuming of oe- females resulted in a significant increase in the wing extensions for control males , but not for males lacking functional ppk25 , and the stimulatory effect of 7-P was restored in ppk25 mutants by targeted expression of ppk25 with ppk25-Gal4 ( Fig . 4C ) . These results show that , as is the case for 7 , 11-HD ( Fig . 2 ) , the stimulatory effect of 7-P on male courtship requires ppk25 function in ppk25-Gal4 cells . Thus , in addition to functioning in the detection of female-specific pheromones that stimulate male courtship , ppk25 and ppk25-expressing neurons are important for detecting at least two other types of excitatory pheromonal cues that promote courtship: pheromones emitted by immature Drosophila , and 7-P , likely accounting for its requirement in the courtship of Tai2 males . In contrast , and unlike ppk23 , ppk25 is not required for detection of the major male inhibitory pheromone , 7T , indicating that ppk25-expressing neurons represent a subset of pheromone-sensing neurons specialized in detecting pheromones that stimulate male courtship behavior . Since ppk25 is required for the detection of a variety of stimulatory pheromones in males , we considered the possibility that it may also be involved in regulating female behavior . Indeed , expression of reporters under the control of ppk25-Gal4 [10] , as well as ppk23-Gal4 and ppk29-Gal4 [11]–[14] is seen in chemosensory neurons of females . Furthermore , chemical senses , in addition to vision and hearing , likely control female receptivity to mating [25]–[27] . To test the role of DEG/ENaC channels in regulating female receptivity , we paired wild-type ( Canton-S ) males with control females or females that were homozygous mutant for ppk23 , ppk25 or ppk29 , and used the fraction of females that mated within thirty minutes as a measure of female receptivity [28] . Females with homozygous mutations in ppk25 , ppk23 or ppk29 displayed showed similar levels of receptivity as control females ( Fig . 5A ) . However , we considered the possibility that the apparent lack of requirement for ppk25 , ppk23 and ppk29 in female receptivity observed under our experimental conditions could result from the redundant action of multiple sensory cues . Indeed , olfactory and/or acoustic signals detected by the antennae play a crucial role in controlling female receptivity [25] , [26] , [29] , [30] . We therefore tested the effect of ppk mutations on the receptivity of females whose antennae had been inactivated by surgical removal of the third antennal segment . Even after antennal inactivation , more than 60% of control females mated within 30 minutes . In sharp contrast , females that were homozygous mutant for ppk25 , ppk29 or ppk23 displayed little , if any , receptivity ( Fig . 5B ) . In order to further explore the redundant antennal cues driving receptivity , we looked at the effect of ppk25 and ppk23 mutations on the receptivity of females that were only lacking the aristae rather than the whole antennae , thereby impairing auditory function but leaving olfactory function intact [31] . Here again , we find that while the receptivity of control females remains high after removal of the arista , mutations in ppk23 and ppk25 significantly decrease receptivity in aristaless females ( Fig . 5C ) . These data suggest that acoustic stimuli are sufficient to promote high levels of female receptivity in the absence of functional ppk25 or ppk23 . The effect of ppk25 inactivation on mating does not result from a reduced sexual attractiveness of the mutant females , as levels of male courtship during these female receptivity assays are similar in the presence of control and ppk25 mutant females ( Fig . 5D , left panel ) . To further rule out any possible effect of the female genotype on male behavior , we conducted male courtship assays using decapitated ppk25 mutant females and controls , thereby reducing behavioral feedback from females . In these assays , males courted both sets of females with equal intensity ( Fig . 5D , right panel ) . Surprisingly , given that the UAS-ppk25 transgene by itself does not rescue the courtship defect of ppk25 mutant males [10] , we find that the presence of UAS-ppk25 significantly increases the receptivity of ppk25 mutant females ( Fig . 5E ) . This observation suggests that leaky expression from UAS-ppk25 , as seen with other UAS transgenes [11] , [32] , [33] , is sufficient to partially rescue ppk25 function in females , further supporting the role of ppk25 in female receptivity . Indeed , a similar transgene with a truncated version of ppk25 , UAS-ppk25Δ [10] , or a UAS-ppk29 rescue transgene [11] do not increase the receptivity of ppk25 mutant females ( Fig . 5E ) , and the UAS-ppk25 transgene does not increase the receptivity of ppk29 mutant females ( data not shown ) . Taken together , these results suggest that all three DEG/ENaC genes play critical roles in female receptivity that can be obscured by the presence of redundant sensory inputs from the antennae . Since in males , ppk25 and ppk23 function within defined subsets of gustatory neurons [10]–[14] , we sought to clarify the relationships between the expression of these ppks in female gustatory neurons , and to test their function in regulating female receptivity . Expression of ppk25 and ppk23 in females was visualized by driving expression of GFP using ppk25-Gal4 and ppk23-Gal4 . Consistent with the sexual dimorphism in the total number of taste hairs on the front legs [34] as well as in the number of ppk23-expressing taste hairs [12] , [13] , there are fewer ppk25-expressing hairs on the front legs of females compared to males ( data not shown ) . As seen in males [12]–[14] , ppk23-positive hairs have two fru-positive chemosensory neurons labeled by fru-LexA-driven expression of Red Fluorescent Protein ( RFP ) , both of which also express GFP under the control of ppk23-Gal4 ( Fig . 6A ) . In contrast , ppk25-Gal4-driven expression of GFP co-localizes with only one of the two fru-positive cells within any particular taste hair ( Fig . 6B ) . Therefore , as in males , ppk23-Gal4 is expressed in each of the two fru-positive taste neurons present in a subset of chemosensory bristles , one of which also expresses ppk25-Gal4 . To test the function of neurons expressing ppk25 or ppk23 , we used the same drivers to target expression of tetanus toxin ( TNT ) , a synaptic transmission blocker [35] . For both ppk25-Gal4 and ppk23-Gal4 , targeted expression of TNT , but not of an inactive variant ( IN-TNT ) , resulted in a severe loss of receptivity . ( Fig . 6C ) . In contrast , expression of TNT within Or22b-expressing olfactory neurons that have no known role in taste or pheromone reception [36] , had no effect . To discriminate between developmental and acute requirements of neuronal function , we expressed the temperature-sensitive version of Drosophila dynamin , shibire ( Shits ) [37] in ppk25-Gal4 neurons and shifted females to the restrictive temperature of 30°C a few minutes before and during the receptivity assay . When assayed at 30°C , females expressing Shits in ppk25-Gal4 neurons were significantly less receptive than females expressing GFP in the same neurons or females carrying UAS-Shits but lacking the Gal4 driver ( Fig . 6D ) . Furthermore , ppk25-Gal4>UAS-Shits females placed at the permissive temperature of 23°C had significantly higher receptivity compared to genetically identical females maintained at the non-permissive temperature ( Fig . 6D ) . Together , these data suggest that the cells identified by ppk25- and ppk23-Gal4 expression are directly involved in female receptivity to mating , most likely in response to male pheromones .
GCaMP-based calcium imaging of ppk23-expressing , pheromone-sensing gustatory neurons identified two types of neurons with different response specificities: F cells respond to female-specific pheromones that stimulate male courtship , while M cells respond to pheromones enriched in males that inhibit male courtship [13] . Here , we present several lines of evidence indicating that ppk25 expression specifically labels F cells . First , GCaMP imaging shows that ppk25-expressing cells respond to female-specific stimulatory pheromones but not to male inhibitory pheromones . Second , in ppk25 mutants , F cells lose their response to pheromones while the response of M cells is unchanged . Finally , behavioral responses to 7 , 11HD , a female-specific pheromone detected by F cells , requires ppk25 function in ppk25-Gal4 neurons , while behavioral response to 7T , a male inhibitory pheromone detected by M cells , does not . These findings are also consistent with previous results showing that ppk25 is required for stimulation of male courtship by females but not for inhibition of male-male courtship [10] . Together , these observations identify a functionally specialized group of gustatory neurons defined by ppk25 function and expression that detects female-specific pheromones and mediates their stimulatory effect on male courtship . Furthermore , mutations in ppk23 or silencing of ppk23 neurons result in loss of both the physiological and behavioral effects of male inhibitory pheromones [13] , [14] , while mutations in ppk25 or silencing of ppk25 neurons do not ( [10] and this report ) . Consistent with those observations , we find that ppk23 mutant males in which ppk23 function is rescued specifically in ppk25 neurons have normal responses to females but still display increased courtship of other males . Together , these data suggest that pheromone-sensing neurons fall into two functionally complementary types . F cells are defined by expression of ppk25 , respond to female-specific pheromones , and mediate the stimulatory effects of those pheromones on courtship . In contrast , M cells express ppk23 but not ppk25 , respond to male pheromones , and mediate the inhibitory effects of those pheromones on courtship . In addition to female-specific pheromones , pheromones found on immature Drosophila [18] as well as 7-Pentacosene [24] can activate male courtship . Both targeted rescue and knockdown experiments demonstrate that the activation of courtship by these pheromones or pheromone blends requires functional ppk25 within the subset of gustatory neurons that also mediate responses to adult female pheromones ( [10] and this report ) . The requirement of ppk25 cells for courtship activation by 7P was surprising given that , in both males or females , while 7P elicits a response from cells that respond to male inhibitory pheromones ( [13] and this report ) , ppk25 cells display no detectable response to this hydrocarbon ( Fig . 1 ) . Compared to their responses to female-specific pheromones , ppk25 cells may respond to 7P with either lower intensity or slower kinetics , preventing detection in our GCaMP assay . Indeed , the physiological responses of olfactory neurons to their cognate olfactory neurons vary in both magnitude and kinetics , neither of which correlates with the strength of the behavioral response [38] . Furthermore , the volatile pheromone cVA was originally shown to strongly activate Or67d neurons [39] but an improved odor testing paradigm revealed that cVA also activates Or65a neurons weakly [40] and subsequent work showed that Or65a neurons are critical for cVA's chronic effects on Drosophila aggression [41] . Finally , while we have tested responses from ppk25-expressing cells at different positions on the front legs , we cannot rule out the possibility that a subset of ppk25 cells , perhaps in a less accessible part of the front leg , or on the second or third pairs of legs , display a detectable GCaMP response to 7P . Alternatively , the lack of GCaMP signal may accurately reflect the fact that ppk25 neurons do not directly detect 7P . By analogy with the integration of olfactory information resulting from non-synaptic interactions between neighboring olfactory neurons [42] , 7-P activation of male courtship may require non-synaptic interactions between ppk25 cells and the neighboring M cells which detect 7P [13] . Whether it is through the direct detection of pheromones , or in a more indirect regulatory role , ppk25 neurons are required for the stimulation of courtship by several different pheromones . In contrast , ppk25 neurons are not required for the responses to at least two different courtship-inhibiting pheromones , suggesting that they are functionally specialized in courtship activation . This specialized function is analogous but opposed to that of the previously described but non-overlapping subsets of neurons on the front legs of males that express Gr32a [7] , [43] , [44] or Gr66a [45] , [46] and detect pheromones that inhibit male courtship . In addition to their roles in pheromonal control of male courtship , we show here that ppk25 , ppk23 and ppk29 also play critical roles in regulating female mating behavior . During the Drosophila courtship ritual , females actively assess the courting male by detecting a variety of cues , including male pheromones and appropriate sensory stimulation of the female is required for mating to occur [47] . Here , we show that ppk25 , ppk23 and ppk29 are all required for antenna-less Drosophila females to become receptive to mating . Furthermore , as in males , ppk23-Gal4 and ppk25-Gal4 are strongly expressed in gustatory neurons of female legs ( [10] , [13] and present work ) . While in addition to its expression in gustatory neurons , ppk25 is also expressed at lower levels in the olfactory system [10] , ppk23 and ppk29 are only detectably expressed in gustatory neurons [11]–[14] . Therefore , the dramatic loss of receptivity observed for antenna-less females carrying mutations in any of the three ppks , or whose ppk23- or ppk25-expressing neurons have been silenced suggests that all three ppks function in a common subset of pheromone-sensing gustatory neurons that regulate female receptivity to mating . Finally , while the identity of the pheromone ( s ) involved remains to be determined , the role of ppk25 in promoting female receptivity is consistent with the involvement of this DEG/ENaC subunit in the responses to multiple pheromones . These data provide , to the best of our knowledge , the first evidence that female receptivity to mating in Drosophila is regulated by gustatory detection of pheromones . Furthermore , gustatory pheromone detection is at least partially redundant with other sensory stimuli , in particular auditory stimuli , as mutations in ppk genes only affect the receptivity of females whose aristae have been removed . Similar redundancy exists in sensory stimulation of male courtship behavior in Drosophila [9] , [16] , [44] , [48] and , more generally in sensory detection of signals that drive sexual behavior in many other species , with potential evolutionary and ecological advantages [49] . While male pheromones , in particular cVA and 7T , have been shown to regulate female mating receptivity , detection of these two pheromones has been reported to involve olfactory rather than gustatory organs; for cVA through the antennal Or67d olfactory receptor [26] and for 7T through unknown receptors on the antennae [25] . The male pheromone that stimulates female receptivity and is detected by ppk25 gustatory neurons therefore remains to be identified . In conclusion , we show that a subset of pheromone-sensing neurons , identified by the expression and function of ppk25 , have a specialized role in stimulating male courtship and female receptivity in Drosophila . The identification and manipulation of these neurons may lead to a better understanding of how gustatory neural circuits drive not only courtship and mating , but also other Drosophila behaviors that depend on the detection of conspecific pheromones [50] . Finally , these results are relevant to the consideration of two mutually exclusive models for the molecular role of DEG/ENaC channels in pheromone detection [17] . Channel gating may result from direct interaction with pheromones or pheromones-protein complexes . Alternatively , these channels may have an indirect role , modulating the excitability of specific subsets of pheromone-sensing neurons . The existence of multiple pheromones requiring ppk25 function is surprising , given the high specificity of known pheromone receptors [51] , [52] . However , the three identified ppk25-dependent pheromones are unsaturated linear hydrocarbons with chain-lengths varying from C25 to C29 with a double-bond at position C7 , while ppk25 function is not required for 7-T , another hydrocarbon with a double-bond at C7 , but with a shorter chain of 23 carbons . The presence of ppk25 in a heterotrimeric DEG/ENaC could therefore modify channel specificity by requiring a ligand with a longer hydrocarbon chain . However , our results are also compatible with the possibility that ppk25 , ppk23 and ppk29 function less directly by modulating the excitability of pheromone-sensing neurons whose ligand specificity is dictated by as yet unidentified pheromone receptors . The specific function of ppk25 in pheromone detection should prove invaluable for dissecting the molecular mechanisms underlying the function of ppk25 , ppk23 and ppk29 in pheromone response , with implications for the roles of DEG/ENaCs in a number of other sensory processes [53] , [54] .
Mutations in ppk25 , ppk23 and ppk29 were described previously [11] , [13] , [16] . ppk25 null mutant flies were heterozygous for two different deletions of the ppk25 gene; an imprecise excision of a P-element that removes the regulatory regions and the first half of the ppk25 gene ( Δ5–22 ) and a deletion spanning 20 genes in the ppk25 regions ( Δ42E ) [16] . ppk29 mutants were homozygous for a transposable element insertion in exon 5 of the ppk29 gene [11] . ppk23 mutants were generated by deletion of an 8 . 3 kb ppk23-containing region through FLP-FRT-mediated recombination of two piggybac transposons [13] . Targeted expression of ppk23 or ppk25 was achieved using the ppk25-Gal4 , UAS-ppk23 and UAS-ppk25 transgenes [10] , [13] . The UAS-ppk25 RNAi line was obtained from the Transgenic RNAi Project at Harvard Medical School ( stock number JF02434 ) . Poxn-Gal4 , fru-LexA , and lexAop-FRT-tdTomato::nls; UAS-stinger lines were gifts from David Mellert and UAS-Shi ( ts ) was a gift from Kathy Siwicki . All other lines were obtained from the Bloomington Stock Center at Indiana University . G-CaMP imaging was performed on single leg chemosensory bristles as described previously [13] , except for the use of the 20×UAS-GCAMP3 transgene [55] . Briefly , single bristles were stimulated by bringing them in contact with a custom-built glass capillary filled with ∼5 µL of a solution of synthetic pheromones ( Cayman Chemical , Ann Harbor , MI ) . Pheromones were diluted to 100 ng/µl in 10% hexane∶90% water solution . Calcium-induced fluorescent increases of one or two cells under a single bristle were monitored by spinning disk confocal microscopy . Errors bars represent the SEM and t-tests were used for statistical significance . Flies were raised at 25°C , 50% relative humidity in a 12 h∶12 h light∶dark cycle . For testing male courtship , males were raised individually for 4–8 days post-eclosion unless noted otherwise , while females and males used as courtship targets were raised in groups . In perfuming experiments , synthetic pheromones ( Cayman Chemical , Ann Harbor , MI ) were applied to oe- males or females lacking cuticular hydrocarbons [21] as described [13] . All assays were conducted under infrared lights and monitored using infrared-sensitive cameras , except perfuming assays with oe- targets and male-male courtship assays , which were performed as previously described [13] , [14] . Courtship directed at decapitated females or males with normal cuticular hydrocarbons was measured for 10 minutes and quantitated using the Courtship Index ( CI , the fraction of time the male spends performing any courtship behavior ×100 [56] ) . As a measure of total behavioral activity , we used the Total Behavioral Index ( TBI , the fraction of time during which the male courts , walks , or preens ×100 [10] ) . For better comparison with previous work [13] , courtship behavior toward perfumed oe- females and males was measured as the number of times the male extends its wings in a 20-minute observation period . For measures of the number of wing extensions , CI and TBI , error bars indicate the SEM , and the Kruskal-Wallis test followed by Dunn's post hoc test was used to determine statistical significance . For the fraction of males initiating courtship , error bars indicate the SEM , and statistical significance was calculated using Fisher's exact test . For female receptivity assays , females were aged for 4–8 days before removal of the third antennal segment with fine forceps at least two days before testing . Canton-S males aged 3–6 days were used for all female receptivity assays . Males and females were aspirated into plexiglass chambers , their behavior recorded for 30 minutes in the light and the number of females copulating within 30 minutes was determined . Neuronal inactivation with Shibirets was achieved by placing females in a 30° Celsius room for 20 minutes prior to , and during the 30-minute receptivity assay . Error bars for female receptivity indicate the SEM , and statistical significance was calculated using Fisher's exact test . All behaviors were scored blind and analyzed either manually or using the LIFESONG X software ( version 0 . 8 ) [57] .
|
Drosophila mating behaviors serve as an attractive model to understand how external sensory cues are detected and used to generate appropriate behavioral responses . Pheromones present on the cuticle of Drosophila have important roles in stimulating male courtship toward females and inhibiting male courtship directed at other males . Recently , stimulatory pheromones emitted by females and inhibitory pheromones emitted by males have been shown to stimulate distinct subsets of gustatory neurons on the legs . We have previously shown that a DEG/ENaC ion channel subunit , ppk25 , is involved in male courtship toward females but not in inhibition of male-male courtship . Here we show that ppk25 is specifically expressed and functions in a subset of gustatory neurons that mediate physiological and behavioral responses to female-specific stimulatory pheromones . Furthermore , ppk25 is also required for the function of those neurons to activate male courtship in response to other pheromones that are not female-specific . In addition to their roles in males , we find that ppk25 , and the related DEG/ENaC subunits ppk23 and ppk29 , also stimulate female mating behavior . In conclusion , these results show that , in both sexes , ppk25 functions in a group of neurons with a specialized role in stimulating mating behaviors .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"molecular",
"neuroscience",
"animal",
"genetics",
"neuroscience",
"animals",
"gene",
"function",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"genetics",
"drosophila",
"research",
"and",
"analysis",
"methods",
"insects",
"arthropoda",
"genetics",
"biology",
"and",
"life",
"sciences",
"sensory",
"systems",
"sensory",
"perception",
"behavioral",
"neuroscience",
"organisms"
] |
2014
|
Drosophila Pheromone-Sensing Neurons Expressing the ppk25 Ion Channel Subunit Stimulate Male Courtship and Female Receptivity
|
The imprint of natural selection on protein coding genes is often difficult to identify because selection is frequently transient or episodic , i . e . it affects only a subset of lineages . Existing computational techniques , which are designed to identify sites subject to pervasive selection , may fail to recognize sites where selection is episodic: a large proportion of positively selected sites . We present a mixed effects model of evolution ( MEME ) that is capable of identifying instances of both episodic and pervasive positive selection at the level of an individual site . Using empirical and simulated data , we demonstrate the superior performance of MEME over older models under a broad range of scenarios . We find that episodic selection is widespread and conclude that the number of sites experiencing positive selection may have been vastly underestimated .
Following the introduction of computationally tractable codon-substitution models [1] , [2] nearly two decades ago , there has been sustained interest in using these models to study the past action of natural selection on protein coding genes . Positive selection can be inferred whenever the estimated ratio ( ) of non-synonymous ( ) to synonymous ( ) substitution rates significantly exceeds one ( reviewed in [3] and [4] ) . In the original models , the ratio was shared by all sites in an alignment , providing little power to detect the signature of positive selection . Indeed , even among classical examples of positively selected genes [5] , [6] , [7] , most substitutions are expected to be neutral or deleterious [8] . Consequently , relatively few genes in which mean estimates are significantly greater than one are expected to exist , e . g . only were found in a human - chimpanzee genome-wide comparison [9] . Random effects codon-substitution models [10] permitted to vary from site to site , which made it possible to identify instances when positive selection had acted only upon a small proportion of sites . Such site-level models can detect which positions in a sequence alignment may have been influenced by diversifying positive selection , e . g . [11] , [12] . However , these models posit that diversifying selective pressure at each site remains constant throughout time , i . e . affects most lineages in the phylogenetic tree , ( Figure 1A ) , and there are very few cases where this assumption is biologically justified ( see [13] , [14] , [15] , [16] for examples of models that allow selection to vary throughout the tree ) . When a site evolves under purifying selection on most lineages , site methods which assume is constant over time may be unable to identify any episodic positive selection , since they will likely infer [17] . It has been noted that positive selection is more readily identified in smaller alignments: counterintuitively , including additional sequences may cause sites to no longer be detected [18] , [19] . This phenomenon could be readily explained by purifying selection on some lineages masking the signal of positive selection on others . We present a mixed effects model of evolution ( MEME ) , based on the broad class of branch-site random effects phylogenetic methods recently developed by our group [20] . MEME allows the distribution of to vary from site to site ( the fixed effect ) and also from branch to branch at a site ( the random effect , Figure 1B ) . Our approach provides a qualitative methodological advance over existing approaches which integrate site-to-site and lineage-to-lineage rate variation , e . g . the branch-site methods [17] or codon-based covarion models [13] . MEME can reliably capture the molecular footprints of both episodic and pervasive positive selection , a task for which current models are not well suited . Using empirical sequence data sets spanning diverse taxonomic categories and gene functions , along with comprehensive simulations , we demonstrate that MEME matches the performance of traditional site methods when natural selection is pervasive , and that MEME reliably identifies episodes of diversifying evolution affecting a small subset of branches at individual sites , where site methods often report purifying selection at the same site . For most empirical data sets analyzed here , episodic selection appears to be the dominant form of adaptive evolution . The biological implications of this type of selection are discussed for each specific data set . We conclude by providing practical guidelines for applying MEME to biological data , and argue that while it is possible to reliably identify sites or branches subject to episodic diversifying selection , statistical power to detect individual branch-site pairs evolving adaptively is inherently limited by a small sample size available for such inference .
The fitting of MEME to an alignment of coding sequences proceeds in three stages: First , the codon model with an alignment-wide is fitted to the data using parameter estimates under a GTR nucleotide model as initial values . Although in some cases nucleotide branch lengths may be a good approximation to codon branch lengths [23] , [24] , recent results indicate that in other instances , nucleotide models can significantly underestimate branch lengths and possibly bias downstream inference [25] . The resulting maximum likelihood estimates , and , for each branch , are used in the site-by-site analyses in the next two steps . Thus we are assuming that the relative branch length and mutational bias parameters are shared across sites and are well approximated by those estimated under a simpler codon model . However , the absolute branch lengths also depend on the site- and model-specific rate parameters below . Second , at each site , we first fit the alternative random effects model of lineage-specific selective pressure with two categories of : and ( unrestricted ) . The probability ( in equation 1 ) that branch is evolving with , is , and the complementary probability that it is evolving with is . By equation 1 , the phylogenetic likelihood at a site , marginalized over all possible joint assignments of , is equivalent to computing the standard likelihood function with the following mixture transition matrix for each branch : ( 2 ) Consequently , the alternative substitution model includes four parameters for each site , inferred jointly from all branches of the tree: and . These form the fixed effects component of the model . Estimating separately for each site accounts for the site-to-site variability in synonymous substitution rates [26] . Lastly , at every site , we fit the model from the previous step , but with : our null model . Using simulated data , we determined that an appropriate asymptotic test statistic for testing most worst-case null of of is a mixture of and ( see Text S1 ) . Mixture statistics of this form often arise in hypothesis testing where model parameters take values on the boundaries of the parameter space , and closed-form expressions for mixing coefficients are difficult to obtain [27] . Throughout the manuscript , we compare MEME to the fixed effects likelihood approach , introduced in [24] ( see Text S1 for motivation ) . The procedure used by FEL differs from MEME in that a single pair of rates are fitted at each site ( no variation over branches ) in Step 2 , and the test in Step 3 is to determine if . Positive selection is inferred by FEL when and the p-value derived from the LRT is significant , based on the asymptotic distribution . If the LRT indicates that a particular site ( ) is subject to episodic diversifying selection , it may be of interest to explore which branches at that site have undergone diversification . The empirical Bayes ( EB ) procedure originally used to identify individual sites subject to diversifying selection in random effects models [28] , can be readily adapted here . To compute the empirical posterior probability at branch that , we apply Bayes' theorem , using to denote the data at site and to denote all the maximum likelihood parameter estimates from the alternative MEME model fitted to site :To compute the two likelihood terms and , we use and , respectively , for the model assigned to branch in equation 2 . The rest of the branches employ the matrices fitted under the alternative model of MEME . Having computed for each branch , we evaluate the empirical Bayes factor for the event of observing positive selection at each branch:When , sequence data increase the prior odds of observing selection at the branch . We do not recommend using this type of inference other than for the purposes of data exploration , even for large values of ( e . g . 100 ) . Intuitively , all the information contributing to the estimate of is derived from observing the evolution along a single branch at a single site ( i . e . from a sample with size ) . To quantify this supposition , we simulated sequence data using the vertebrate rhodopsin phylogeny and branch lengths , applied positive selection of varying strength to five branches in the tree selected a priori ( see Text S1 ) , and applied the EB procedure to infer the identity of selected branches .
To assess the performance of MEME on both simulated and empirical data , we selected the fixed effects likelihood method ( FEL [24] ) as the most appropriate reference test for pervasive diversifying selection , because FEL most closely matches the assumptions made by MEME ( see Text S1 ) . We simulated data sets under a number of scenarios: refer to Text S1 for details of simulation strategies . To gauge the comparative performance of MEME and FEL when identifying sites subject to pervasive diversifying selection , we used a collection of 16 protein-coding alignments , representing a diverse array of taxa , genes subject to differing levels of conservation , and a range of data set sizes ( Table 2 ) . In alignments analyzed , MEME identified all the sites inferred by FEL to be under diversifying positive selection and found between ( e . g . West Nile virus NS3 ) and ( Diatom SIT ) additional sites that were subject to episodic diversifying selection ( Table 2 ) . In four data sets , sites identified by FEL with p-values close to were missed by MEME . Note that MEME p-values for these sites remained in the range ( Table 2 ) , i . e . marginally significant . Sites identified by both methods tended to have a greater average proportion of lineages under selection ( , measured by the mean of MLE estimates of ) ; sites found only by MEME experienced more episodic selection ( ) . In data sets ( Table 2 ) , sites that FEL inferred to be under purifying selection are instead identified by MEME as likely to have been subjected to episodic diversifying selection . Almost universally ( Tables S4 , S5 , S6 , S7 , S8 , S9 , S10 , S11 , S12 , S13 , S14 , S15 , S16 , S17 , S18 , S19 ) , such sites had a smaller estimated proportion of positively selected lineages ( ) . This behavior is consistent with the relative performance of the two tests on simulated data and corroborates the expectation that MEME has greater power to identify sites when only a proportion of lineages evolved under positive selection . Vertebrate rhodopsin , Japanese encephalitis virus env , and Camelid VHH are investigated in detail below; for a discussion other genes , see Text S1 . The vertebrate rhodopsin ( a low-light vision protein ) data set was previously experimentally investigated for the substitutions that modulate the wavelength of the light absorbed by the molecule ( , [18] ) . The authors asserted that , because none of the sites that they had determined as affecting by site-directed mutagenesis were detected by site-level computational methods , “statistical tests of positive selection can be misleading without experimental support . ” Other authors reanalyzed the same data set more comprehensively and went even further , questioning the utility of -based methods for detecting experimentally validated sites , because “most of the current statistical methods are designed to identify codon sites with high values , which may not have anything to do with functional changes . The codon sites showing functional changes generally do not show a high value” [29] . The validity of this generalization has been correctly questioned with a simple counter-argument that the sites detected by computational methods may also be functionally important , because the change in is unlikely to be the sole determinant of adaptation [17] . The MEME analysis of this gene suggests another obvious alternative , also expounded by previous studies [17]: the failure of the original computational analysis [18] to identify functionally important sites results from the fact that these sites have been subjected to episodic selection , which is masked by predominantly purifying selection elsewhere in the tree . Indeed , among three sites that alter found by MEME ( 96 , 183 and 195 , versus none found by FEL ) , no more than of the branches exhibited ( Table S17 ) ; at these sites , the average is less than 1 . We note that , because adaptive evolution will not always adhere to a single , simple scenario of episodic diversifying selection , we do not expect MEME to find all sites experimentally confirmed to alter . For example , three of the nine missed sites ( ) appear to have been subjected to partial selective sweeps and have been detected using a specialized model of directional evolution [29] . Three sites from this alignment can be used to illustrate how the inclusion of lineage variability modifies inference of selection ( Figure 2 ) . Site 54 was inferred to have experienced pervasive non-synonymous substitutions throughout its evolutionary history . Both FEL and MEME detect this site as positively selected ( ) . Sixty three percent of the lineages at this site evolved with , whereas the remainder were conserved ( ) , according to MEME . The log-likelihood of the site is only marginally higher for MEME , which suggests that MEME behaves like FEL at sites with “canonical” patterns of diversifying selection , corroborating the simulation results . At codon 273 , FEL obtained a maximum likelihood estimate of , but failed to infer positive selection , as the signal was not statistically significant ( ) . MEME , on the other hand , allocated ( 0 . 013–0 . 10: 95% confidence interval obtained by latin hypercube sampling importance resampling [30] ) of branches to a rate class with ( 2 . 94–6726 ) and inferred positive selection ( ) . The difference in log-likelihoods between MEME and FEL is points: MEME fits significantly better , based on a 2-degrees of freedom likelihood ratio test ( ) . The maximum likelihood estimates of individual model parameters have large associated errors ( although in all posterior samples we obtained ) , as is expected for inference based on a single site . This has also been noted by Yang and dos Reis [17] . The point estimates themselves , however , are immaterial for inferring whether or not a site is positively selected , since the likelihood ratio test is used for that purpose . Perhaps the most dramatic example of the added power of MEME is illustrated by site 210 . At this site , the evolutionary history is replete with non-synonymous substitutions along deep lineages followed by extensive synonymous evolution , indicative of purifying selection . There is also a small clade with repeated synonymous and nonsynonymous substitutions . Averaging over all branches , FEL determined that the site , overall , is under negative selection ( ) . MEME reported that of the branches were under a very strong selective constraint ( ) , but that the remaining were under strong diversifying selection ( ) . The log-likelihood improvement is now at the cost of two parameters , which is highly significant ( ) . Site 210 is the ideal illustration of why it is undesirable to average over all lineages: bursts of diversification followed by conservation will most likely be missed by traditional site methods . No evidence for selection was found in this envelope gene in previous analyses [28] , and FEL found only one site under positive selection . Despite the low levels of divergence among a relatively small number of taxa ( 23 isolates ) , MEME found episodic selection at sites called negatively selected by FEL ( Table S12 ) . Two of these sites fall within a beta-barrel epitope known to be involved in escape from neutralizing antibodies [31] . Sites 33 and 242 showed evidence of repeated toggling at terminal lineages . Remarkably , site 33 – likely a part of a neutralizing antibody epitope [32] – changed from isoleucine to leucine on 6 terminal lineages; site 242 changed from phenylalanine to serine on 5 terminal lineages . These substitutions co-occur on three terminal lineages . Evidence of recombination was detected in this alignment , and corrected for using a partitioning approach ( details on how MEME can correct for recombination are in Text S1 ) . The camelid VHH data set comprises partial variable domain sequences ( germline alleles ) of llama and dromedary heavy chain only antibodies ( Table S3 ) . 11 of 16 sites in the variable complementarity determining regions ( CDR ) 1 ( sites 26–33 ) and 2 ( sites 51–58 ) were found to be under diversifying selection by MEME ( 2/16 were detected by FEL and 2 more were marginally significant ) . Because CDR regions are driven to diversify in order to provide a broad basis of antigen recognition , positive selection is expected to be commonplace in the CDRs [33] . MEME was able to uncover selective signatures at a majority of those sites . Of the remaining sites classified by MEME as positively selected , six were associated with VHH family differentiation [34] . Unlike standard antibodies , which must maintain relatively conserved framework regions ( FR ) involved in binding heavy and light chains to form functional tetramers , VHH antibodies are free of such functional constraints . A previous analysis of camelid VHH for evidence of positive selection using counting methods [35] reported evidence for positive selection at a single site ( 14 ) in FR1 ( sites 1–25 in Table S3 ) , but this analysis could find no clear evidence of positive or negative selection at FR sites . In contrast , MEME inferred episodic selection at six sites in FR1 , six sites in FR2 ( sites 34–50 ) , and sites in FR3 ( sites ) . The well-known lack of power of counting methods to detect even pervasive selection [17] likely hampered the previous study . Although a previous analysis of vertebrate rhodopsin sequences found no sites under selection at posterior probability [18] , the same authors found 7 selected sites in the subset of squirrelfish sequences , and 2 selected sites when the subset of fish sequences was analyzed . These results run counter to the expectation that more data should provide greater power to detect selection . MEME , on the other hand , detects more selected sites when more sequences are included . One site is identified in the squirrelfish alignment , in the fish alignment , and in the complete rhodopsin alignment . All but sites detected in the subset alignments are also identified in the full alignment ( Table S20 ) . Allowing to vary over branches at least partially mitigates the pathology of constant- models which effectively rely on an average for inferring selection . A similar pattern is seen in the analysis of the influenza A virus H3N2 hemagglutinin sequences , where site-level methods also appear to be sensitive to sequence sampling ( [19] , see Text S1 and Table 23 ) .
We have presented a mixed effects model of evolution , MEME , and a statistical test for detecting the signal of past episodic positive selection from molecular sequence data . Our model corrects the biologically unrealistic assumption that selective pressure , as measured by the ratio , remains constant over lineages . Based on comprehensive simulations and empirical analysis of an array of taxonomically diverse genes , MEME can be recommended as a replacement for existing site models . MEME matches the performance of older approaches when natural selection is pervasive , but possesses greater power to identify sites where episodes of positive selection are confined to a small subset of branches in a phylogenetic tree . Our results suggest that it may be necessary to revise previous estimates of the proportion of sites under positive selection in many genes . Using the FEL method , which assumes constant selective pressure at a site , we are able to detect sites across all empirical alignments . MEME identifies of these sites ( the remaining are borderline significant ) and additional sites – nearly times as many as FEL . For individual data sets ( e . g . Drosophila adh and Diatom SIT , Table 2 ) , the differences may be even more dramatic . The greater power of MEME indicates that selection acting at individual sites is considerably more widespread than constant models would suggest . It also suggests that natural selection is predominantly episodic , with transient periods of adaptive evolution masked by the prevalence of purifying or neutral selection on other branches . We emphasize that MEME is not just a quantitative improvement over existing models: for sites in our empirical analyses , we obtain qualitatively different conclusions . FEL asserts that these sites evolved under significant purifying selection , but MEME is able to identify the signature of positive selection on some branches . Furthermore , MEME is less sensitive to sampling effects that plague existing positive selection detection tools [18] , [19] . Variable levels of purifying selection pressure across different lineages prevented these older methods from detecting instances of episodic positive selection; MEME is able to peer through the fog of purifying selection . It is important to bear in mind that the mixture statistic used to calculate the p-values reported here is based on a null model under which all sites are evolving neutrally . This , however , is not biologically realistic: the null hypothesis against which sites ideally ought to be screened is one under which sites are evolving either neutrally or under purifying selection . But the proportion of sites evolving under negative selection and the strength of this selection are unknown and vary from case to case , which means that such a null hypothesis would be very sensitive to modeling assumptions that cannot be justified in general . Instead , the neutral null hypothesis represents a worst case scenario for our inference , so that the p-values we obtain are upper bounds of the true p-values . This ensures that our inference is conservative . Even in the worst ( and biologically unrealistic ) case for MEME , namely when selective pressures are constant throughout the phylogeny , the loss of power compared to FEL is minimal: a site with FEL p-values between and will be missed by MEME , since its p-value will be for the same ranges of the likelihood ratio test statistic ( LRT ) . In our simulation scenarios under the assumption of constant , this translates to no more a loss in power ( Table S3 ) . Our inference is performed in a site-wise rather than an alignment-wide manner , and we therefore control the site-wise rather than the family-wise error rate . We do not recommend combining the results of multiple site-wise inferences to perform alignment-wide inference . To aid interpretation of the results while taking account of multiple testing , we calculate the false discovery rate [36]; the resulting q-value upper bounds are reported alongside their corresponding p-value upper bounds in Tables S4 , S5 , S6 , S7 , S8 , S9 , S10 , S11 , S12 , S13 , S14 , S15 , S16 , S17 , S18 , S19 . This gives an upper bound on how many of the reported sites can be expected to be false discoveries: for instance , of the 30 sites reported in Table S5 we expect no more than ( 14% ) to be false , and probably far fewer because of the conservative choice of null model . We emphasize that q-values are usually much larger than their corresponding p-values and caution that p-values ( regardless of whether they have been corrected for multiple testing ) cannot be used to estimate an expected number of false discoveries in the same way . MEME is a conceptual advance over the first generation of random effects models designed to detect episodic selection ( called “branch-site models” in the literature [17] ) . MEME does not require a priori designation of , or an exhaustive search for , the branches under selection , and it allows each site to have its own selective history . Whereas branch-site models make restrictive a priori assumptions about how values are distributed across the tree – sometimes leading to very poor statistical performance [20] – MEME treats the selective class on each branch as a random effect that is marginalized over in the likelihood calculation . For computational tractability , MEME assumes that the value taken by on each branch is independent of that on any other branch , i . e . selective pressures between branches are uncorrelated . This assumption could potentially be violated: for example , if changes very slowly across the phylogeny , then values on neighboring branches will be correlated . Further research is needed to understand how inference of selection would be affected if these correlations were directly accounted for , and whether the additional model and computational complexity would be justified . In practice , MEME could be combined with models of directional selection to improve power , e . g . [15] , [16] . Unlike covarion models [37] , [13] , MEME does not allow to change in the middle of a tree branch . The effect of this restriction is unclear , but it could be tested by implementing a mixed effects covarion model , where switching rates and proportion of time spent under are estimated at an individual site . The ability of MEME , or similar substitution model-based methods , to accurately infer the identity of individual branches subject to diversifying selection at a given site seems unavoidably limited . Most of the information that such inference might be based on is limited to character substitutions along a single branch at a single site , i . e . one realization of the Markov substitution process . Selection along terminal branches in the context of negatively selected background can be detected more reliably than selection along interior branches among neutrally evolving background lineages . However , we caution that despite obvious interest in identifying specific branch-site combinations subject to diversifying selection , such inference is based on very limited data ( the evolution of one codon along one branch ) , and cannot be recommended for purposes other than data exploration and result visualization . This observation could be codified as the “selection inference uncertainty principle” – one cannot simultaneously infer both the site and the branch subject to diversifying selection . In this manuscript , we describe how to infer the location of sites , pooling information over branches; previously [20] we have outlined a complementary approach to find selected branches by pooling information over sites . Finally , although MEME is considerably more powerful than existing methods at detecting bursts of selection , it still requires that a measurable proportion of lineages ( ) experience non-synonymous evolution at a site . When a single substitution modifies an adaptive trait and is subsequently fixed , we expect based methods to have very little power . Specialized methods which make use of change in allele frequencies [15] , [16] , or between and within-population diversification patterns [38] , will be required in such cases .
|
Identifying regions of protein coding genes that have undergone adaptive evolution is important to answering many questions in evolutionary biology and genetics . In order to tease out genetic evidence for natural selection , genes from a diverse array of taxa must be analyzed , only a subset of which may have undergone adaptive evolution; the same gene region may be under stabilizing or relaxed selection in lineages leading to other taxa . Most current computational methods designed to detect the imprint of natural selection at a site in a protein coding gene assume the strength and direction of natural selection is constant across all lineages . Here , we present a method to detect adaptive evolution , even when the selective forces are not constant across taxa . Using a variety of well-characterized genes , we find evidence suggesting that natural selection is generally episodic and that modeling it as such reveals that many more sites are subject to episodic positive selection than previously appreciated .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"adaptation",
"mathematics",
"natural",
"selection",
"statistics",
"statistical",
"methods",
"biology",
"evolutionary",
"biology",
"comparative",
"genomics",
"evolutionary",
"processes"
] |
2012
|
Detecting Individual Sites Subject to Episodic Diversifying Selection
|
Over 20 years , from October 1989 , the Darwin prospective melioidosis study has documented 540 cases from tropical Australia , providing new insights into epidemiology and the clinical spectrum . The principal presentation was pneumonia in 278 ( 51% ) , genitourinary infection in 76 ( 14% ) , skin infection in 68 ( 13% ) , bacteremia without evident focus in 59 ( 11% ) , septic arthritis/osteomyelitis in 20 ( 4% ) and neurological melioidosis in 14 ( 3% ) . 298 ( 55% ) were bacteremic and 116 ( 21% ) developed septic shock ( 58 fatal ) . Internal organ abscesses and secondary foci in lungs and/or joints were common . Prostatic abscesses occurred in 76 ( 20% of 372 males ) . 96 ( 18% ) had occupational exposure to Burkholderia pseudomallei . 118 ( 22% ) had a specific recreational or occupational incident considered the likely infecting event . 436 ( 81% ) presented during the monsoonal wet season . The higher proportion with pneumonia in December to February supports the hypothesis of infection by inhalation during severe weather events . Recurrent melioidosis occurred in 29 , mostly attributed to poor adherence to therapy . Mortality decreased from 30% in the first 5 years to 9% in the last five years ( p<0 . 001 ) . Risk factors for melioidosis included diabetes ( 39% ) , hazardous alcohol use ( 39% ) , chronic lung disease ( 26% ) and chronic renal disease ( 12% ) . There was no identifiable risk factor in 20% . Of the 77 fatal cases ( 14% ) , 75 had at least one risk factor; the other 2 were elderly . On multivariate analysis of risk factors , age , location and season , the only independent predictors of mortality were the presence of at least one risk factor ( OR 9 . 4; 95% CI 2 . 3–39 ) and age ≥50 years ( OR 2 . 0; 95% CI 1 . 2–2 . 3 ) . Melioidosis should be seen as an opportunistic infection that is unlikely to kill a healthy person , provided infection is diagnosed early and resources are available to provide appropriate antibiotics and critical care .
Melioidosis is the clinical disease following infection with the soil and water bacterium Burkholderia pseudomallei [1] , [2] . It occurs in humans and a wide variety of animals and is thought to usually follow percutaneous inoculation . In addition , inhalation of aerosolized bacteria probably occurs during severe weather events such as tropical storms , aspiration is documented with near drowning and ingestion can occur , especially in grazing animals but also from mastitis-associated infected breast milk [3] , [4] . Zoonotic transmission is described but is exceedingly uncommon , as are person-to-person transmission , nosocomial transmission and laboratory-acquired infection . While melioidosis can present as a rapidly fatal septicemic illness and B . pseudomallei is now considered a potential biothreat agent , there remain major gaps in understanding the global distribution , epidemiology and pathogenesis of this infection . The known endemic distribution of B . pseudomallei is expanding well beyond the traditional melioidosis-endemic regions of Southeast Asia and northern Australia , with recent case reports of melioidosis from the Americas , Madagascar , Mauritius , India and elsewhere in south Asia , China and Taiwan [5] , [6] . It remains unclear to what extent this reflects true expansion of endemicity rather than unmasking of the long-standing environmental presence of the bacterium . Since October 1989 we have prospectively documented all cases of melioidosis in the tropical “Top End” of the Northern Territory of Australia . We described the presentations of the first 252 cases after 10 years of the Darwin prospective melioidosis study [7] and we now present the findings from 540 cases over 20 years .
This study was approved by the Human Research Ethics Committee of the Northern Territory Department of Health and Families and the Menzies School of Health Research ( HREC 02/38 ) and data were analysed anonymously . The Top End has a population of around 150 , 000 in an area of 516 , 945 km2 , with almost 125 , 000 living in the Northern Territory capital city of Darwin ( 12°S ) . All patients with culture-confirmed melioidosis in the Top End from October 1st 1989 until September 30th 2009 were included . Investigation , treatment and follow-up were supervised in all cases in consultation with the Infectious Disease Department at Royal Darwin Hospital , the 350 bed referral hospital for the Top End . We followed all patients until death or after completion of therapy . Hazardous alcohol use was defined as greater than an average daily consumption of six standard drinks ( 60 g alcohol total ) for males and four ( 40g alcohol total ) for females . Chronic lung disease was defined as a documented diagnosis of chronic obstructive airways disease . Chronic renal disease was defined as a creatinine of >150 umol/L ( N . R . <90 umol/L ) before the melioidosis illness or after completion of therapy if not previously documented . Septic shock was defined as the presence of hypotension not responsive to fluid replacement together with hypoperfusion abnormalities manifest as end organ dysfunction [8] . Patient details were stored in a database and analysed using Stata version 10 ( Stata Corporation , Texas ) . Chi-squared or Fisher exact tests were used to assess categorical variables; p<0 . 05 was considered significant and risk ratios and 95% confidence intervals were then calculated . To identify associations with a fatal outcome and with presentation with pneumonia and with bacteremia we conducted multivariable logistic regression analyses with stepwise backwards elimination of patient demographic and risk factor variables , with odds ratios and 95% confidence intervals calculated .
There were 540 cases and 77 deaths ( 14% ) attributable to melioidosis over the 20 years . Ages ranged from 8 months to 91 years ( median 49 years ) . There were 26 children ≤15 years old ( 5% ) and two of these died , one with congenital heart disease and one with severe rheumatic heart disease . 372 patients ( 69% ) were male and 281 Indigenous Australians ( 52% ) . 262 patients ( 49% ) lived in the suburbs of Darwin , 65 ( 12% ) on rural properties ( “blocks” ) outside Darwin city , 37 ( 7% ) in the regional towns of Katherine and Nhulunbuy and 169 ( 31% ) in remote Indigenous communities . Four infections were considered acquired in sub-tropical central Australia and three were acquired elsewhere in tropical northern Australia outside the Northern Territory . Table 1 shows patient risk factors and outcomes by risk factor . There were only 2 patients with confirmed HIV infection , although a small number were not tested and 1 patient was seropositive for HTLV-I . Mortality was significantly higher in those with chronic respiratory disease ( 19% vs 13%; risk ratio = 1 . 5 ( 95% CI 1 . 1–2 . 4 ) ; p = 0 . 048 ) . Although no other individual risk factor , including diabetes , was predictive of mortality , the absence of any risk factors was strongly predictive of survival; of the 106 ( 20% ) with no identified risk factor for melioidosis , only two died ( 2% ) ; both elderly , aged 75 and 82 years , respectively . 407 patients ( 75% ) were considered to have exposure to environmental B . pseudomallei through their recreational activities and 96 ( 18% ) had direct exposure through occupational activities including gardening and outdoor maintenance , plumbing , building construction , plant machine operation and military exercises . Only 103 ( 19% ) had no evident environmental or recreational exposure . In 118 ( 22% ) cases there was a specific exposure scenario that was considered the likely infecting event . These included skin wounds sustained whilst working outdoors or gardening , or while playing sports such as soccer and rugby on muddy playing fields , or while fishing in fresh water rivers , hunting such as chasing feral pigs through tropical savannah swamps and motor vehicle accidents involving wet soil exposure . Several cases were in disabled people who rarely ventured outside their accommodation but were potentially exposed to aerosolised bacteria during storms . Regional clusters of cases occurred following severe weather events such as the Katherine river flood in January 1998 , and Category 5 tropical cyclone Thelma , which hit the Tiwi Islands in December 1998 [9] . One cluster with nine cases of melioidosis and four deaths was attributed to confirmed B . pseudomallei contamination of the un-chlorinated water supply in a remote Aboriginal community [10] . Overall 436 ( 81%; 95%CI 77%–84%; p<0 . 005 ) presented during the wet season ( November 1st–April 30th ) and mortality was higher in cases presenting in January ( 23/102 ( 23% ) died; p = 0 . 007 ) than in other months . Pneumonia was a significantly more common presentation in the peak monsoonal months of December to February ( 172/280; 61% ) than in the other 9 months ( 106/260; 41%; p<0 . 001 ) . Of all presentations , 461 ( 85% ) were considered acute ( defined as symptoms present for less than 2 months ) and from recent infection . 60 ( 11% ) were chronic in nature ( defined as symptoms present for over 2 months; 25 pneumonia , 23 skin ulcer ( s ) , 12 others ) . These chronic infections were considered to be mostly acquired during the current or preceding wet season , with the delay until presentation explaining some of the cases diagnosed during the dry season . 17 of the 53 ( 32% ) cases who presented during the mid dry season months of June 1st to September 30th fulfilling the definition for chronic melioidosis . Patients with chronic melioidosis were less likely to be diabetic than those with acute melioidosis ( 20% vs 42%; p<0 . 001 ) , with 42% having no identified risk factor in comparison to 17% of those with acute disease ( p<0 . 001 ) . Only 1 of 60 patients ( 2% ) with chronic melioidosis died ( p<0 . 001 ) . The remaining 19 ( 4% ) patients were thought to have reactivation of disease from a latent focus of B . pseudomallei infection , based on long-standing prior radiological abnormalities and/or known long-standing positive melioidosis serology ( 13 pneumonia , 2 bacteremia no focus , 2 genitourinary infection , 1 each soft tissue infection and skin abscess ) . Those with presumptive reactivated melioidosis were more likely to have underlying chronic lung disease ( 47% compared with 25% for all others; p = 0 . 03 ) and rheumatic heart disease and/or congestive cardiac failure ( 32% compared with 6% for all others; p<0 . 001 ) and 5/19 of these died ( 26% compared with 14% for all others; p = 0 . 13 ) . The clinical presentations and outcomes are shown in Table 2 . Overall 298 ( 55% ) patients were bacteremic . Pneumonia was the commonest principal clinical presentation on admission ( 278 cases; 51% ) , followed by genitourinary infection ( 76 cases; 14% ) and skin infection ( 68 cases; 13% ) . There were 20 ( 4% ) patients presenting with septic arthritis and/or osteomyelitis and 14 ( 3% ) with neurological melioidosis , of whom 10 presented with meningo-encephalitis , 2 with myelitis and 2 with cerebral abscesses . Bacteremia without an evident clinical focus was also a common presentation ( 59 cases; 11% ) , with severity of illness ranging from rapidly fatal septic shock to a clinically very mild febrile illness . When septic shock occurred it was usually present on or within 24 hours of admission . Of the 116 patients ( 21% ) with septic shock , 58 ( 50% ) died from acute fulminant melioidosis . In contrast , for those without septic shock on presentation , mortality was 4% overall ( 19/424 ) ; even in the 195 of those without septic shock who were bacteremic , only 13 ( 7% ) died . Of the 106 patients with no identified risk factor , 23 ( 22% ) were bacteremic and 6 ( 6% ) had septic shock and the only deaths in this group were the 2 elderly patients as already noted . Table 3 shows significant risk factors in the 278 melioidosis patients with a primary presentation of pneumonia . On univariate analysis age ≥50 years , diabetes , excessive alcohol consumption and rheumatic heart disease and/or congestive cardiac failure were each associated with a propensity for presentation with pneumonia in comparison to other presentations . However on multivariable analysis diabetes and age were not independent predictors of a presentation with pneumonia , while chronic lung disease , excessive alcohol consumption and rheumatic heart disease and/or congestive cardiac failure were . Table 4 lists internal organ abscesses and other foci of infection . Following findings early in the study of frequent internal collections , CT scanning of abdomen and pelvis has been routinely performed on all patients with melioidosis since around 1995 . Prostatic abscesses were present in 76 males ( 20% ) , the majority of which required drainage [11] . In comparison to case series from Thailand , hepatic abscesses were uncommon and as with splenic and renal abscesses rarely required drainage . Three women had mastitis and three men had epididymo-orchitis . Lymphadenitis ( sometimes suppurating ) , muscle abscesses , diffuse myositis and cellulitis were all seen but were uncommon . Four patients had para-intestinal masses which were considered possible primary infection following ingestion of B . pseudomallei , as was a presentation with a ruptured large gastric ulcer with subphrenic abscess and suppurative peritonitis . Mediastinal widening on chest X-ray and CT scan was seen , sometimes with clearly enlarged mediastinal lymph nodes and usually in association with pneumonia ( 12/17 cases ) . Four patients had suppurative pericarditis , three with contiguous pulmonary infection and one without evident pulmonary infection who developed acute pericardial tamponade requiring emergency thoracotomy and a pericardial window . Two had mycotic pseudo-aneurysms and one woman presented with a ruptured uterus from a massive uterine wall abscess . In addition to the initial principal clinical presentation , subsequent clinically-evident secondary foci were not uncommon and examples are shown in Table 5 . Secondary pneumonia was especially common in those presenting with genitourinary infection , septic arthritis/osteomyelitis and bacteremia without an apparent clinical focus , but was unusual in those presenting with skin infection . Secondary foci were also less common in those presenting with pneumonia , although brain abscesses and septic arthritis requiring surgery occurred in this group . Of note , the pattern of secondary neurological melioidosis was different from the encephalomyelitis seen as a primary presentation . Of the eight patients with secondary neurological disease , all were blood culture positive ( in comparison to 3/14 of those with primary neurological melioidosis; p = 0 . 001 ) and 5 had abscesses ( 4 intracranial , 1 spinal cord ) . 121 patients ( 22% ) were admitted to the Royal Darwin Hospital Intensive Care Unit ( ICU ) and of these 40 ( 33% ) died . In the ICU 97 were ventilated ( 41 died; 42% ) ( Table 6 ) and 60 received granulocyte colony-stimulating factor ( G-CSF ) ( 15 died; 25% ) . Three patients also received activated protein C therapy ( 1 died ) . Of the 77 deaths overall , 75 were during the initial hospital admission , with the time from admission to death in these ranging from 0 to 111 days ( median 3 days ) . Two patients were dead on arrival at hospital , 8 died on the day of admission , and 7 died the day after admission . Of the 465 patients surviving the initial infection , 30 ( 6% ) re-presented with culture-confirmed recurrent melioidosis subsequent to completion of antibiotic therapy , with 2 deaths in this group . Of these 30 , 25 were considered to have relapse of an unsuccessfully eradicated infection , usually resulting from poor adherence to antimicrobial therapy . In these cases , the time from initial admission to first relapse was 3 . 6–28 months ( median 8 months ) , with one of these patients dying . Two of these patients had a second relapses ( 25 and 27 months after their first relapse ) ; one of these patients died during the second relapse and the other patient had a third relapse 5 years after the second relapse . For those surviving the initial admission , diabetes was more common in those who relapsed ( 16/25 ( 64% ) vs 166/440 ( 38% ) ; p = 0 . 016 ) , as was bacteremia on admission ( 18/25 ( 72% ) vs 221/440 ( 50% ) ; p = 0 . 034 ) . There were 5 patients with recurrent melioidosis where the B . pseudomallei isolates were different from the original isolate by pulsed-field gel electrophoresis or multilocus sequence typing ( MLST ) . One patient with cystic fibrosis had three separate presentations with melioidosis at ages 10 , 14 and 18 years . There was a good response to therapy each episode , with each B . pseudomallei isolate being a different sequence type ( ST ) and frequent sputum cultures between episodes being consistently culture negative for B . pseudomallei . He was considered to have been re-infected on three separate occasions . Three other patients with recurrent melioidosis but disparate isolates on typing were also considered likely to have new infections , occurring 14 , 58 and 72 months after the initial infection , respectively . One further patient had disparate isolates on MLST from presentations 8 months apart and was thought to have relapse of a probable initial infection with multiple B . pseudomallei strains , with the osteomyelitis of the second presentation being evident clinically during the first presentation . Clinically apparent re-infection with B . pseudomallei is therefore thought to have occurred in only 4/465 ( 1% ) patients surviving the initial admission , despite most survivors remaining in the melioidosis-endemic location , with many having persisting risk factors and continuing environmental exposure to B . pseudomallei . Of the 463 patients who did not die from melioidosis , all except one have eventually cleared their infection with antibiotic therapy . This is a patient with moderately severe bronchiectasis who presented with a productive cough at age 61 years , with B . pseudomallei cultured from sputum . Her sputum has remained consistently B . pseudomallei culture positive for 8 years , despite multiple courses of intravenous and prolonged oral antibiotics and also a lobectomy of the most severely bronchiectatic lung lobe . She nevertheless remains generally well . Table 6 shows decreasing mortality over the 20 years of the study that was not explained by either increasing recruitment of less sick patients or fewer risk factors in patients . To further assess associations with mortality we included the following categorical variables in the initial logistic regression model; age ( ≥50 years ) , indigenous ethnicity , each of the risk factors from Table 1 , location , and presentation in December , January or February . No individual risk factor was a significant independent predictor of mortality ( data not shown ) , although in a similar logistic regression model with bacteremia as the outcome , independent predictors of bacteremia were indigenous ethnicity , age ≥50 years , diabetes , hazardous alcohol use , chronic renal disease , malignancy and immunosuppression ( Table 7 ) . Therefore our final model for mortality from melioidosis incorporated presence or absence of any of the defined risk factors as a dichotomous variable . In this model the independent factors associated with mortality were age ≥50 years ( OR 2 . 0; 95% CI 1 . 2–3 . 3 ) and presence of any risk factor ( OR 9 . 4; 95% CI 2 . 3–39 ) , but not indigenous ethnicity , geographical location or season ( Table 8 ) .
Serological surveys suggest that most infections with B . pseudomallei are asymptomatic , with over half of teenagers seropositive in the highly endemic region of northeast Thailand [12] . It was estimated that for children in northeast Thailand approximately 1 in 4600 antibody-producing exposures results in clinical infection [13] . The Darwin prospective melioidosis study provides strong support for the vast majority of melioidosis cases being from recent infection , with 81% of cases presenting during the monsoonal wet season , similar to a figure of 75% in Thailand [14] . Nevertheless latency of B . pseudomallei with subsequent reactivation is well recognised , being described as long as 62 years after infection in a returned World War II prisoner of war infected in southeast Asia [15] . It was estimated from serology studies that following the Vietnam War around 225 , 000 US service personnel may have been infected with B . pseudomallei [16] . This was called the “Vietnamese time bomb” , but the subsequent number of melioidosis cases following return to the USA has been comparatively small . Reactivation from a latent focus was considered to have occurred in only 19/540 ( 4% ) cases in the Darwin study . We previously estimated the average annual incidence rate of melioidosis in the Top End of the Northern Territory to be 19 . 6 cases per 100 , 000 population , with an estimated rate in diabetics of 260 cases per 100 , 000/year [17] . Yearly rates between 1990 and 2002 ranged from a low of 5 . 4/100 , 000 in 1993 to a high of 41 . 7/100 , 000 in 1998 , a year with two severe tropical cyclones with intense rainfall and winds . This compares to northeast Thailand , with an average annual melioidosis incidence rate between 1997 and 2006 of 12 . 7/100 , 000 and with a highest rate of 21 . 3/100 , 000 in 2006 [18] . In the Darwin study 75% of melioidosis cases reported recreational activities that would result in exposure to environmental B . pseudomallei and 18% had clear occupational exposure . Both males ( 69% ) and indigenous Australians ( 52% ) were over-represented , most likely reflecting increased environmental exposure . There were 118 cases ( 22% ) where history revealed a likely specific infecting event . B . pseudomallei is common in the urban environment of Darwin and most of the 49% of patients in the study who lived in the city of Darwin were infected in the city environs , including domestic gardens and yards . Mortality in the Darwin study was not linked to geographical location , being actually higher ( although not statistically significantly so ) in the urban population than in the rural and remote population ( Table 8 ) . In contrast , in northeast Thailand 81% of cases of melioidosis were in rural rice farmers and their children [14] . In Singapore melioidosis has occurred in construction workers , gardeners and military personnel , but in that tropical island city state , where over 80% of people live in high-rise apartments , the reasons for infection often remain unclear [19] . Earlier in the Darwin prospective melioidosis study we established that the incubation period for acute melioidosis following specific infecting events was 1–21 days ( mean , 9 days ) [20] . The incubation period , clinical presentations of melioidosis and outcomes are thought to be determined by a combination of bacterial load infecting the individual , putative B . pseudomallei strain differences in virulence , mode of infection and , most importantly host risk factors for disease [21] . For instance , less severe disease with symptoms present for over 2 months before presentation ( chronic melioidosis ) was significantly less common in diabetics and was more commonly seen in those without underlying risk factors . The association between inhalation as a route of acquisition and increased severity of disease with higher mortality than percutaneous exposure is well recognised for anthrax , plague and tularaemia , but appears to have been under-appreciated in melioidosis . While the association between melioidosis and rainfall is well established [14] and there is epidemiological support for inhalation of aerosolised B . pseudomallei during severe weather events resulting in a pneumonic presentation with higher mortality [9] , [19] , [22] , the overall contribution of inhalation of B . pseudomallei in comparison to percutaneous inoculation remains entirely unclear . Support for inhalation of B . pseudomallei from this study includes that 61% of admissions during the peak monsoonal months of December to February were with pneumonia , in comparison to only 41% in the other 9 months , plus the recognition that mediastinal lymphadenopathy is not uncommon . Diabetes is the most important risk factor for melioidosis , followed by hazardous alcohol use , chronic lung disease and chronic renal disease [7] , [14] , [17] , [19] , [23] , [24] , [25] . Malignancy , immunosuppression and thalassemia are also recognised risk factors [24] . In the Darwin study 39% of patients were diabetic , with nearly all having adult onset type 2 diabetes . Rates of diabetes from other endemic locations were 57% in the largest series from Thailand [24] , 48% in Singapore [19] , 60% in Taiwan [26] , 38% in bacteremic melioidosis patients in Malaysia [27] and 42% in north Queensland , Australia [25] . When considering the estimated prevalence of diabetes in the whole population , we previously calculated the risk of melioidosis in diabetics in the tropical Top End of the Northern Territory to be 21 . 2 ( 95% CI 17 . 1–26 . 3 ) times the risk in non-diabetics [17] , which is similar to data from Thailand [14] . A lack of association of melioidosis with HIV infection [28] , [29] supports a limited role for adaptive immunity in protection against acquisition of and mortality from melioidosis , despite evidence for a cell-mediated immune response to B . pseudomallei [30] , [31] . We proposed that a unifying hypothesis for the predominance of diabetes , excessive alcohol consumption and chronic renal disease in melioidosis patients was the critical role of innate immunity and especially robust neutrophil function in controlling infection with B . pseudomallei [7] . The specific defects in neutrophil function in diabetes , alcohol excess and renal disease have been well described and were the basis for trialling therapy with granulocyte-colony stimulating factor ( G-CSF ) in melioidosis [32] , [33] . The dysfunctional neutrophil hypothesis is supported by a study in a mouse model showing a critical role for neutrophils in resistance to melioidosis [34] and a recent study from Thailand showing that , in comparison to non-diabetics , otherwise healthy diabetics had neutrophils displaying impaired phagocytosis of B . pseudomallei , reduced migration in response to interleukin-8 and an inability to delay apoptosis [35] . The occurrence of melioidosis in chronic granulomatous disease also supports a key role for neutrophils [36] . Our clinical impression is that the risk for melioidosis in those with hazardous alcohol use may often be directly related to binge drinking rather than chronic liver disease , with high blood alcohol levels at the time of exposure to B . pseudomallei inhibiting protection against bacterial propagation and dissemination . This is consistent with earlier studies on neutrophil function in alcohol intoxication [37] , [38] . An additional potential pathogenetic mechanism for more severe disease in those with hazardous alcohol intake is the induction by alcohol of bacterial genes encoding various potential virulence mechanisms , as recently shown in transcriptional profiling studies of Acinetobacter baumannii grown in the presence of alcohol [39] . In the Darwin study hazardous alcohol use but not diabetes was an independent predictor of presentation with melioidosis pneumonia ( Table 3 ) . In addition to neutrophil dysfunction , alcohol excess also adversely affects many other components of innate pulmonary host defences , from decreased ciliary beat frequency to impaired alveolar macrophage phagocytosis and inhibited cytokine responses [40] . Various aspects of adaptive pulmonary immunity are also affected by alcohol , involving both cellular and humoral responses . We have noted two comorbidities previously unrecognized as potential risk factors for melioidosis . Chronic lung disease was an independent predictor of pneumonic melioidosis , which may reflect defective innate immunity such as impaired alveolar macrophage function [41] . Rheumatic heart disease and cardiac failure may predispose to melioidosis by similar mechanisms [42] . It is being increasingly recognised that patients with cystic fibrosis are at substantial risk of infection with B . pseudomallei if they live in or travel to endemic regions [43] . Chronic infection can occur , with acute flares of pneumonia and progressive deterioration of lung function , as also seen with B . cepacia infection in cystic fibrosis [44] , [45] , [46] . Patients with cystic fibrosis should consider avoiding travel to locations where melioidosis is common . One Darwin patient with cystic fibrosis has had three separate infections with different genotypes of B . pseudomallei . In addition , there is only 1/463 survivors in the Darwin study in whom clearance of B . pseudomallei has not been possible . This patient has severe bronchiectasis and has had persisting pulmonary infection for 8 years; such inability to eradicate B . pseudomallei from sputum has only been previously documented in cystic fibrosis [43] . Around half of melioidosis cases present with pneumonia , which can be part of a fatal septicaemia , a less severe unilateral infection indistinguishable from other community-acquired pneumonias or a chronic illness mimicking tuberculosis [2] , [47] . In the Darwin study mortality was 49% in those with pneumonia who also had septic shock , in comparison to 6% in those with pneumonia without septic shock and 4% in those with chronic pneumonia . Early clinical descriptions and animal studies showed that melioidosis pneumonia can follow percutaneous infection [48] , but the proportions of our pneumonia cases which were from percutaneous exposure , inhalation or aspiration are unknown . Nevertheless the finding of mediastinal widening on chest X-ray and CT scan in some melioidosis patients is analogous to inhalational anthrax . Other presentations in the Darwin study range from skin lesions without systemic illness [49] , to overwhelming sepsis with abscesses disseminated in multiple internal organs . Genitourinary [11] , bone , joint and neurological infections [50] , [51] , [52] are all well recognised . One manifestation of melioidosis commonly seen in Thailand [53] , but not seen over the 20 years of the Darwin study is children presenting with parotid abscesses . The reasons for this difference remain unclear . The dramatic presentation of melioidosis brainstem encephalitis or myelitis has been noted to be more commonly seen in Australia than in Thailand [1] , [3] . Recent mouse studies have suggested that such neurological presentations may result from direct entry of B . pseudomallei to the brain from the nasal mucosa via the olfactory nerve or similar pathways [54] . Genetic differences between B . pseudomallei strains may account for regional clinical variations . It was recently demonstrated that the global B . pseudomallei population probably evolved from an ancestral Australian population which subsequently spread to Southeast Asia [55] . One possible explanation for the neurological disease being more common in Australia is differences in propensity between B . pseudomallei populations for actin-based motility of bacteria along nerve pathways , conferred by variants of the BimA gene which have been found to be geographically restricted [56] . The concept of direct brain invasion by B . pseudomallei in the Darwin cases of primary melioidosis meningo-encephalomyelitis is supported by the low bacteremia rate in comparison to all 8 of those with secondary brain infections being bacteremic . Most of the less common presentations of melioidosis seen in the Darwin study have also been described from other locations . These include mycotic aneurysms [57] , epididymo-orchitis [58] , pericarditis [59] and mastitis with maternal to child transmission of melioidosis [4] . The common presence of diverse internal organ abscesses necessitating routine imaging is also well recognised [1] , [60] , [61] . In contrast to Thai studies , where spleen and liver abscesses predominate [60] , prostate abscesses were extremely common in the Darwin series , being present in 76/372 ( 20% ) males . While liver , spleen and renal abscesses responded to prolonged antibiotic therapy , prostatic abscesses usually required drainage , whether primary or secondary [11] . Although unusual , the presence of para-intestinal masses supports that ingestion can occasionally be the primary route of infection in humans , as is more commonly seen in grazing animals [62] . Whatever the initial clinical presentation , secondary foci of infection are common in melioidosis ( Table 5 ) , presumably from bacteremic spread and reflecting the high rate of bacteremia overall ( 55% ) . Of those 59 patients presenting with bacteremia without an apparent focus , 25 ( 42% ) subsequently developed an evident secondary focus of infection . Secondary pneumonia and septic arthritis were especially common . Therapy of melioidosis requires prolonged antibiotics to cure infection and prevent relapse [63] . In the Darwin study 25/465 ( 5 . 4% ) patients who survived the initial infection relapsed after treatment , with a median time to relapse of 8 months from initial admission , in comparison to 86/889 ( 9 . 7% ) and 6 months from commencement of oral therapy in Thailand [24] . Choice and duration of and compliance with antibiotic therapy were the strongest indicators of risk for relapse in both locations . Diabetes was significantly associated with risk for relapse in the Darwin series but not in Thailand , while in both locations bacteremia on initial admission was associated with relapse , although only significantly so in Thailand . Genotyping of B . pseudomallei from recurrent melioidosis has shown that reinfection can also occur but is less common than relapse [24] , [64] . In northeast Thailand reinfection occurred in 30/899 ( 3 . 4% ) patients , making the incidence of melioidosis reinfection substantially higher than that of primary infection [24] . This is in contrast to the Darwin patients , where reinfection was documented in only 4/465 ( 1% ) , despite most having persisting risk factors and continuing exposure and even raising the possibility of some acquired immunity to reinfection following melioidosis . Simultaneous infection with more than one strain of B . pseudomallei has been shown to very uncommon ( 2/133 cases in Thailand ) [65] , but was thought likely in one of the patients in this study . Mortality from melioidosis in the Darwin study was 14% , with 75 of the 77 deaths occurring during the initial hospital admission and only 2 deaths from relapsed melioidosis . Mortality during the first 5 years of the study ( from October 1989 ) was 30% and during the last 5 years ( until October 2009 ) was 9% ( p<0 . 001 ) . These rates compare with 49% mortality in the large Thai study from 1986–2004 [24] , with mortality now decreasing in that region [18] , 65% in bacteremic patients in Malaysia during 1976–1991 [27] , 16% in Singapore between 1998–2007 [19] , 22% in Taiwan between 2000 and 2005 [26] and 25% in north Queensland between 1996 and 2004 [25] . The decreasing mortality over the 20 years of the Darwin study cannot be attributed to ascertainment bias from improved diagnosis of less severe cases . Indeed , the bacteremia rate was higher in the last 5 years ( 66% ) , possibly reflecting both more frequent repeat culturing in suspected cases and improved laboratory detection of low level bacteremia . The median age , rates of septic shock , percentages with various risk factors and the proportion with no risk factors did not change significantly over the 20 years ( Table 6 ) . The overall bacteremia rate of 55% compares with up to 65% in Thailand [24][66] , 50% in Singapore [19] and 60% in north Queensland [25] . We attribute the improved survival over time to a combination of earlier diagnosis of melioidosis through increased community and health staff awareness of the possibility , earlier treatment with ceftazidime or meropenem [67] and probably most importantly , access to and improvements in intensive care management of the septic patient . In many melioidosis-endemic regions renal replacement therapy and other resources for managing the metabolic abnormalities and organ dysfunction seen in severe sepsis are limited and without these mortality in septicemic melioidosis will remain high [2] , [24] , [68] . A greater proportion of patients were ventilated in the second half of the Darwin study and mortality in patients with septic shock was 100% in the first five years and decreased to 27% in the last 5 years ( Table 6 ) . Our initial optimism of potential benefit from G-CSF therapy in septicemic melioidosis [32] has been tempered by a randomized controlled trial in Thailand which showed that G-CSF conferred no mortality benefit in severe melioidosis in that setting [33] . Nevertheless those treated with G-CSF in that study had a longer duration of survival , suggesting that if state-of-the-art ICU therapy is available G-CSF may be beneficial . A decrease in mortality in melioidosis similar to that seen in Darwin has also occurred in Singapore , where mortality in 1989–1996 was 40% in comparison to the more recent 16% [19] . It is notable that during the more recent series from Singapore there was a period in March–April 2004 with case numbers , proportion with pneumonia ( 83% ) and mortality ( 53% ) all higher than at other times [19] . This cluster of cases followed heavy rainfall and strong winds . Genotyping showed a diversity of strains , excluding a point source outbreak and the severe disease in this cluster was attributed to a possible shift to inhalation of aerosolized B . pseudomallei [19] , [69] . This is analogous to clusters seen in the Darwin study following severe weather events [9] . This study provides strong support for the presence of specific host risk factors being the most important determinant of mortality from melioidosis . Older age is also recognised as a risk factor for melioidosis [14] , [17] , [19] and in the Darwin study age ≥50 years was an independent predictor of death from melioidosis ( Table 8 ) . Of the 77 deaths from melioidosis over the 20 years , 2 were in elderly patients without other evident risk factors and all of the other 75 fatal cases had at least one of the specific recognised risk factors listed in Table 1 . That severe disease is very uncommon in melioidosis in patients without risk factors is evident from the much lower rates of bacteremia ( 22% ) and septic shock ( 6% ) in these patients in the Darwin study . The association of diabetes with bacteremia in patients with melioidosis has been noted in Thailand [23] . Nevertheless , while all the listed risk factors including diabetes were independently associated with bacteremia in the Darwin study ( Table 7 ) , no individual risk factor apart from age was an independent predictor of mortality . This reflects that 80% of melioidosis cases in the Darwin study had at least one risk factor irrespective of age and it was the presence of any of these risk factors that was highly predictive of mortality ( OR 9 . 4; 95% CI 2 . 3–39 ) . Although the higher proportion of presentations with pneumonia during the peak monsoonal months of December to February supports a role for inhalation and although mortality was higher on univariate analysis during these 3 months , multivariate analysis showed that seasonality was not itself a significant independent predictor of mortality ( Table 8 ) . Furthermore , while severe melioidosis is associated with an array of pathogen induced immune dysregulation [70] , that no death occurred in a patient without risk factors does not support an important role for cytokine-related human genetic polymorphisms in determining outcomes in melioidosis [71] . Therefore , although disease may be more severe following inhalation and/or higher infecting load of B . pseudomallei , the only predictor of mortality from melioidosis is the presence of defined risk factors such as diabetes , hazardous alcohol use , chronic lung or renal disease and older age . In conclusion , melioidosis should be more seen as an opportunistic pathogen that is very unlikely to kill a healthy person , provided the infection is diagnosed early and resources are available to provide appropriate antibiotics and critical care where required .
|
Melioidosis is an occupationally and recreationally acquired infection important in Southeast Asia and northern Australia . Recently cases have been reported from more diverse locations globally . The responsible bacterium , Burkholderia pseudomallei , is considered a potential biothreat agent . Risk factors predisposing to melioidosis are well recognised , most notably diabetes . The Darwin prospective melioidosis study has identified 540 cases of melioidosis over 20 years and analysis of the epidemiology and clinical findings provides important new insights into this disease . Risk factors identified in addition to diabetes , hazardous alcohol use and chronic renal disease include chronic lung disease , malignancies , rheumatic heart disease , cardiac failure and age ≥50 years . Half of patients presented with pneumonia and septic shock was common ( 21% ) . The decrease in mortality from 30% in the first 5 years of the study to 9% in the last five years is attributed to earlier diagnosis and improvements in intensive care management . Of the 77 fatal cases ( 14% ) , all had known risk factors for melioidosis . This supports the most important conclusion of the study , which is that melioidosis is very unlikely to kill a healthy person , provided the infection is diagnosed early and resources are available to provide appropriate antibiotics and critical care where required .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"microbiology/environmental",
"microbiology",
"public",
"health",
"and",
"epidemiology/environmental",
"health",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"critical",
"care",
"and",
"emergency",
"medicine/sepsis",
"and",
"multiple",
"organ",
"failure",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"infectious",
"diseases/bacterial",
"infections",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"microbiology/medical",
"microbiology",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] |
2010
|
The Epidemiology and Clinical Spectrum of Melioidosis: 540 Cases from the 20 Year Darwin Prospective Study
|
Case management in children with cutaneous leishmaniasis ( CL ) is mainly based on studies performed in adults . We aimed to determine the efficacy and harms of interventions to treat CL in children . We conducted a systematic review of clinical trials and cohort studies , assessing treatments of CL in children ( ≤12 years old ) . We performed structured searches in PubMed , CENTRAL , LILACS , SciELO , Scopus , the International Clinical Trials Registry Platform ( ICTRP ) , clinicaltrials . gov and Google Scholar . No restrictions regarding ethnicity , country , sex or year of publication were applied . Languages were limited to English , Spanish and Portuguese . Two reviewers screened articles , completed the data extraction and assessment of risk of bias . A qualitative summary of the included studies was performed . We identified 1092 records , and included 8 manuscripts ( 6 Randomized Clinical Trials [RCT] and 2 non-randomized studies ) . Most of the articles excluded in full-text review did not report outcomes separately for children . In American CL ( ACL ) , 5 studies evaluated miltefosine and/or meglumine antimoniate ( MA ) . Their efficacy varied from 68–83% and 17–69% , respectively . In Old-World CL ( OWCL ) , two studies evaluated systemic therapies: rifampicin and MA; and one study assessed efficacy of cryotherapy ( 42% , Per Protocol [PP] ) vs intralesional MA ( 72% , PP ) . Few studies ( 4 ) provided information on adverse events ( AEs ) for children , and no serious AEs were reported in participants . Risk of bias was generally low to unclear in ACL studies , and unclear to high in OWCL studies . Information on efficacy of treatment for CL in children is scarce . There is an unmet need to develop specific formulations , surveillance of AEs , and guidelines both for the management of CL and clinical trials involving the pediatric population . The protocol of this review was registered in the PROSPERO International register of systematic reviews , number CRD42017062164 .
Cutaneous leishmaniasis is the most common presentation of Leishmaniasis , with global estimates of 0 . 7 to 1 . 2 million cases per year[1] . In contexts of peri-domestic and anthroponotic transmission , children represent an important number of cases . In 2016 for the Americas , among 48 , 915 reported cases of CL , 15 . 5% were children ≤10 years of age[2] . In the Eastern-Mediterranean region , 100 , 000 new cases of CL are reported annually[3] , and previous reports from Iran note that pediatric patients comprise 7–10% of the cases[4] . Children are a special population of CL cases . Observational studies and clinical trials have shown higher proportions of failure of treatment with first line drugs for CL in children[5–8] compared with adults , especially in younger age groups ( <12 years of age ) [9] . Differences in immune response[10] , drug clearance[11] and overall drug exposure [5 , 12] in pediatric patients have contributed to this disparity in therapeutic response . Furthermore , the anatomical , physiological and biochemical changes that occur from birth to adolescence affect the pharmacokinetics/pharmacodynamics ( PK/PD ) of drugs[13] and the frequency and type of adverse events[14] . There are no specific guidelines for the treatment of CL in the pediatric population , and children are generally treated using interventions developed and tested for adults . Considering this information and policy gap , we sought to summarize the literature in terms of efficacy and safety of antileishmanial drugs in children ( ≤12 years of age ) .
Study design and literature search: We conducted a systematic review of the literature following a protocol registered in the PROSPERO International register of systematic reviews , number CRD42017062164 . The PICO question[15] covered a population of pediatric patients with confirmed diagnosis of cutaneous leishmaniasis without mucosal or visceral involvement; intervention with systemic and local treatments for cutaneous leishmaniasis; comparison group with placebo , active drug or local treatments . The outcomes of interest for this review were proportion of cured patients in each arm and treatment safety ( frequency of adverse events ) . Study designs included randomized clinical trials and non-randomized ( cohort ) studies . Inclusion criteria: Original articles that assess treatments of cutaneous leishmaniasis including pediatric patients ( ≤12 years of age ) , with no restrictions regarding ethnicity , country , sex , or year of publication . Exclusion criteria were: 1 ) articles assessing a disease other than cutaneous leishmaniasis ( e . g . mucosal or muco-cutaneous leishmaniasis , visceral leishmaniasis ) ; 2 ) publication language other than English , Spanish or Portuguese; 3 ) case reports , case series , case-control studies and systematic reviews/meta-analyses; 4 ) in vitro or animal studies; 5 ) full-text not available ( after two requests to corresponding authors ) ; 6 ) reviews , books , and articles without available full texts ( conferences , editorials , author responses ) ; 7 ) articles that did not report outcomes separately for children; and 8 ) duplicated reports . Search strategy and references/data management: From December 2016 to February 2017 , two reviewers ( MdMC and AUR ) performed an independent literature search . Structured searches were conducted in electronic bibliographic databases: MEDLINE ( via PubMed ) , the Cochrane Central Register of Controlled Trials ( CENTRAL ) , LILACS , SciELO , Scopus , the WHO- International Clinical Trials Registry Platform ( ICTRP ) , clinicaltrials . gov ( U . S . National Institutes of Health ) and Google Scholar for gray literature . The search terms used were “cutaneous leishmaniasis”; “treatment” , “therapy” , “management”; “outcome” , “cure” or “failure”; and “infant” , “child” or “adolescent” . The search strategy used in PubMed is provided as an example: ( Cutaneous AND leishman* ) AND ( treatment OR management OR Therapy ) AND ( outcome* OR cure OR failure ) AND ( infant[MeSH] OR child[MeSH] OR adolescent[MeSH] ) . See supplementary information ( S1 File ) for a complete list of terms adapted to each search engine’s requirement . We limited the search to articles with human participants and to three languages: English , Spanish , and Portuguese . No restrictions regarding ethnicity , country , sex , or year of publication were applied in this search strategy . If full-text articles were unavailable , authors were contacted by e-mail ( e . g . for unpublished studies ) up to a maximum of two attempts within a period of two weeks . Manuscripts whose authors did not answer after the second request were excluded . Additional search of references included in previously published major systematic reviews of interventions for cutaneous leishmaniasis [16–18] was performed ( by AUR ) in order to find relevant references that were not captured in the initial search . Two programs were used to manage references and conduct the review process . The lists of references from different databases and reviewers were imported into Mendeley Reference manager to merge the records and remove duplicates . Thereafter , an overall list of references was imported to the online software Covidence[19] to perform screening by title/abstract , full text review , data extraction and risk of bias assessment . Screening and inclusion of studies: First , the two reviewers performed an independent screening of titles and abstracts identified in the literature searches . Subsequently , the reviewers performed a full text review , applying the a priori inclusion and exclusion criteria . Disagreements in inclusion/exclusion of abstracts and articles were solved by consensus or by a third reviewer ( AC ) . Data extraction and summary: Two reviewers ( MdMC and AUR ) independently extracted the relevant data on the pediatric population , using a predefined template , based on the PICO question and adapted to the software Covidence[19] . Disagreements between reviewers were resolved by consensus or by a third researcher ( AC ) . Data extracted included: 1 ) Primary outcome: proportion of patients with therapeutic cure ( or therapeutic failure if it was the only outcome reported ) in each arm of study , and the definitions of these variables used by the study authors . 2 ) The secondary outcomes were drug safety ( frequency of adverse events ) and time to cure ( if available ) . 3 ) Sociodemographic and clinical data: age range ( of pediatric population ) ; total enrolled patients and number of pediatric participants; diagnostic methods and Leishmania species . 4 ) Setting and methodology: country where the study was conducted , aim , design , inclusion/exclusion criteria , intervention/comparator , and follow-up . We performed an assessment of the risk of bias for each individual study following the Cochrane Collaboration tool[15] and the Newcastle Ottawa Scale[20] for non-randomized studies . A qualitative summary of the evidence was conducted , since a meta-analysis was not planned for this study , considering the expected high heterogeneity of design and outcome definitions , which has been evident in previous systematic reviews[16–18] .
The database searches resulted in 1290 records; 222 duplicates were removed and another 24 reports were added as part of the review of references of other studies ( Fig 1 ) . Of the remaining 1092 records , 758 were excluded based on evaluation of title and abstract; 334 full-text articles were evaluated , and nine ( 9 ) articles were selected for data extraction . Excluded studies were mostly those that included children and adults but failed to report outcomes separately for pediatric patients ( n = 125 ) or were performed only in adults or adolescents ( >12 years old ) . While reviewing data from the 9 selected studies , one of these[21] was excluded as it tested a therapeutic approach ( diminazene aceturate ) not currently used in clinical practice . This intervention was previously reported as having insufficient evidence to support its use[16] . The eight included studies were published between 2001 and 2017 , principally during the last ten years ( 6/8 articles ) . Six were randomized controlled trials , and two were non-randomized studies ( Table 1 ) . Five studies were conducted in the Americas[5 , 7 , 22–24] , and all of them assessed systemic treatments . The 3 studies conducted in Old-World CL endemic countries ( Saudi Arabia , Sudan and Iran ) assessed both local and systemic therapies[25–27] . The principal inclusion criterion for study participants in the eight articles was parasitological confirmation of CL , and all excluded patients with previous anti-leishmanial treatment , mucosal or muco-cutaneous disease . The number of lesions varied among studies . Length of follow-up was also variable , but in general , it was longer for the studies in the American region ( 6–12 months ) than for those conducted in other regions ( 45 days to 6 months ) . Regarding definition of therapeutic response , in the ACL studies cure was defined as complete re-epithelization of lesions and absence of inflammatory signs such as induration , crust or raised borders . Final response was defined at 6–7 months after initiating treatment in four studies ( two studies counted 6 months from the end of treatment[23 , 24] ) , and at 12 months in one study[7] . Initial response was determined at 3 months ( 13 weeks or 2 months after treatment ) for most of the studies ( Table 2 ) . For OWCL , definitions of therapeutic outcome varied among authors . In the studies by Layegh et al , there is a clear description of criteria of cure ( Table 2 ) , including re-epithelization of ulcerated lesions and complete resolution of induration for other type of lesions[26 , 27] . Jaffar[25] describes the outcome as “complete healing of lesions at the end of three months” . The timing for measuring the outcome also varied in OWCL studies , from 45 days to 6 months , being different for each OWCL study included in the review ( Table 1 ) . A combined population of 461 children with CL was included in this review . Reported ages ranged between 2 and 15 years , although two studies did not report the lowest age of enrolled patients[7 , 27] ( Table 2 ) . Distribution by sex was described in five studies [5 , 22 , 23 , 26 , 27] , with the proportion of male patients ranging from 41% to 61% , except for the study by Chrusciak-Talhari et al[23] , with a higher proportion of males ( 65% [13/20] and 70% [7/10] ) in both treatment groups . Information of Leishmania species was available only for American CL studies and the most frequently isolated species corresponded to L . ( V . ) panamensis , L . ( V . ) guyanensis and L . ( V . ) braziliensis ( Table 2 ) . In contrast , OWCL studies did not report isolated Leishmania species , and interpretation of efficacy data relies on the known endemic species in the area ( Table 1 ) . Overall , the risk of bias for the main outcome ( therapeutic response ) was lower for the studies conducted in the Americas compared to OWCL . Most of the ACL studies reported clearly the random generation of allocation sequence , blinding of main outcome and had a low risk of incomplete outcome reporting . Three of these studies were classified as unclear risk for allocation concealment . In addition , all the clinical trials assessed for ACL were open-label trials , mainly due to the different route of administration of the drugs ( oral vs parenteral ) . Regarding the randomized trials in OWCL , most of the reports were classified as unclear or high risk of bias for all categories within the assessment tool , except for the incomplete outcome data for all outcomes . Detailed risk assessment is available in supplementary Tables 1 and 2 ( S1 Table and S2 Table ) . One common limitation of the evidence generated in the reviewed studies is the small sample size of each study , and consequent modest generalizability of the results . Only two studies were performed exclusively in children [22 , 26] . In the remaining articles , efficacy of treatments in pediatric patients was determined by subgroup analysis or as a secondary outcome ( as in the study of PK of miltefosine in Colombian patients ) . These features of the included studies qualify the validity of the findings . Considering the large number of articles excluded , the characteristics of the pediatric population and the age categories used for reporting efficacy data in the excluded articles were described . Seven articles did not report the age range of enrolled participants ( Fig 1 ) , and most of them did not report outcomes separately for children <12 years ( n = 125 ) . The median sample size of these articles was 85 ( IQR: 52–131 ) and the majority assessed interventions for OWCL ( 75% , n = 94 ) . Median of the lower age limit of enrolled patients was 5 years old ( IQR: 2–7 ) . Children represented over 50% of the sample in at least 11% of these articles , being as high as 87% in the study of Ben-Salah et al[28] ( 87% of patients were <18 years old ) . However , the proportion of children in the sample was not reported in 78% of the articles ( n = 97 ) , and the age categories used to describe the population were variable ( <18 , <15 , <12 or <10 years old ) .
This systematic review presents a summary of evidence on efficacy and safety for treatments in pediatric populations with cutaneous leishmaniasis . It also provides evidence of the gaps in reporting of treatment outcomes for pediatric patients , even when included in clinical trials . Most articles presented overall results without stratifying outcomes for adult and pediatric populations , and cut-offs for reporting age distribution of participants were variable . Eight reports including 461 pediatric patients aged 2–15 years were reviewed , and no studies enrolled children less than 2 years of age . Miltefosine and meglumine antimoniate were identified as interventions for pediatric ACL , while rifampicin , cryotherapy , systemic and intralesional meglumine antimonate were evaluated for OWCL . Efficacy of miltefosine and meglumine antimoniate in children varied from 68–83% and 17–69% , respectively . In general , miltefosine shows a higher response rate in the study populations . None of the studies assessed superiority , but the largest reviewed study[22] showed that this drug was non-inferior to meglumine antimoniate in patients with L . panamensis and L . guyanensis infection . Thus , for pediatric populations , miltefosine offers a good and achievable therapeutic option , with its oral route of administration , facilitating adherence and enabling home-based supervision of treatment , and access to therapy . In contrast , for OWCL , evidence was more limited , and the results are difficult to compare due to the variable definitions of therapeutic response , including different length of follow-up . One clinical trial [26] showed that in Iranian children with cutaneous leishmaniasis , intralesional meglumine antimoniate had greater efficacy than cryotherapy ( 72 . 3% vs 41 . 7% , respectively ) . None of the included studies evaluated the efficacy of miltefosine . One of the excluded studies that did not report outcomes for patients aged <12 years , showed higher efficacy of paromomycin in patients <18 years of age[28] . This intervention was found to have good evidence in previous systematic reviews[16] . Only six studies in this wide literature search assessed efficacy in children and adults separately [5 , 7 , 23–25 , 27] . In general , drug efficacy was lower in pediatric patients compared with adults , independent of the treatment . This is similar to previous studies reporting age as a risk factor for therapeutic failure in leishmaniasis[6 , 8 , 29] , explained at least partly by differences in pharmacokinetics [5 , 11 , 12] and other host factors[10] . However , most studies in CL do not stratify outcomes by age groups or for adults vs . children ( 125 manuscripts excluded for this reason ) . In addition , varying age cut-offs for pediatric cases and inconsistent reporting of age categories limit the interpretation of findings regarding treatment outcomes in different age groups and their generalizability . The role of ontogeny in the disposition and actions of drugs is important to understanding age-related differences in therapeutic response[30] , especially in younger children . Efforts to provide a rationale for age subgroups in pediatric trials are ongoing[31] and regulatory guidelines provide some , though arbitrary , reference age categories for pediatric studies[32–34] . Nevertheless , high variability in the boundaries of age categories is common[35] , and this is more evident in studies including both children and adults , as described in this review . Inclusion of children and consideration of the factors influencing drug distribution , metabolism and pharmacokinetics in the design of clinical trials for interventions of CL , and indeed all Neglected Tropical Diseases ( NTDs ) , would improve the quality of evidence supporting treatment recommendations , as well as to provide additional insights on the PK/PD of currently available drugs for CL and other NTDs . Some examples include the allometric dosing of miltefosine for visceral leishmaniasis ( NCT02431143 ) , and the dosing of benznidazole for Chagas disease[36] . Importantly , this review also highlights the lack of data regarding treatments in patients less than two years of age , whose treatment options are often considered off-label . Among the current treatment options for leishmaniasis , amphotericin B has evidence of safety in neonates and infants , obtained from clinical trials for systemic fungal infections[37] . However , there is no efficacy data for CL treatment in children under 2 years of age for this or any other recommended treatment . Notably , in a cross-sectional study conducted in a reference center , 9% of pediatric CL cases were less than 2 years old[38] . Off-label use of drugs can be a source of adverse events in children , and other concerns , such as unavailability of appropriate formulations for young children , can affect compliance and effectiveness of drugs[31] . In addition , use of “in house” preparations when no pediatric formulation is available constitutes a risk of dosage error[39] . Patient acceptability , including mode of administration to the child and any related pain or discomfort , are aspects relevant for pharmaceutical development of medicines for pediatric use[32] . These aspects may not be fully met by available drugs for treating CL , where parenteral drugs have been used for decades and oral drugs are only available in capsules , which may be difficult to swallow for young children . Novel designs in pediatric trials to collect efficacy and PK data , such as opportunistic and scavenged sampling ( use of residual blood/plasma from laboratory testing obtained during routine clinical care ) , are an alternative to address this lack of information and may overcome some of the constraints involved in conducting clinical trials in children[40] . In addition , considering that at least 11% of the articles excluded for not reporting outcomes in children have a sample size formed by more than 50% of pediatric patients , valuable efficacy and safety data could be obtained . Implementation of data sharing platforms and re-analysis of individual-patient data from these studies may overcome the limitations for conducting additional clinical trials in this vulnerable population . Safety data in children were limited , as they were commonly incorporated with adverse events ( AEs ) of adults . Reported AEs in the included studies were similar in nature and frequency to those described in adults , although the available data do not allow differences to be fully assessed . Children may be more or less susceptible than adults to the adverse effects of different drugs[14] and some AEs are unique to this population . Monitoring of drug safety in children is critical , because during the process of drug development , clinical trials generate only limited data on AEs in children [41] . Therefore , development and implementation of tools and strategies for surveillance of adverse reactions to antileishmanial treatments , including intralesional and other local therapies in children , warrants attention from the health research community and public health professionals . Comparison of efficacy data between studies was difficult , in particular for OWCL , due to the differences in outcome definition and duration of follow-up . This variability did not allow a quantitative synthesis of evidence . Previous systematic reviews[16–18] have identified this issue and initiatives to standardize protocols for clinical trials in CL are ongoing[42] . Another limitation of our study is the restriction of the search to three languages ( English , Spanish and Portuguese ) , which might be relevant for OWCL , since articles in other languages that may have valuable information from Africa , the Middle East and Asia were excluded . In our study we defined 12 years of age as the upper limit for inclusion of articles , considering patients >12 years old as adolescents . These age categories are similar to the ICH ( The International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use ) guidelines for pediatric clinical trials[34] , but studies using different classifications for pediatric patients ( e . g . ≤18 years old ) , might have been excluded from this analysis . In conclusion , this study documented the absence of guidelines and scarcity of evidence supporting case management of CL in children . Data sharing platforms to allow individual-patient data analysis , high-quality studies and clinical trials are needed to provide robust data on drug efficacy and safety to support the development of guidelines and implementation of interventions for children with CL .
|
Cutaneous leishmaniasis ( CL ) is a parasitic disease that causes chronic , often ulcerated , skin lesions that leave lifelong scars on the face or other visible areas . In some regions of the world , children represent a high proportion of cases . Treatment options for children are limited , and may require administration of poorly tolerated drugs . Despite the differences in responses to these drugs , treatment regimens for children are based on extrapolation of efficacy data in adults . We systematically reviewed the medical literature , searching for controlled studies assessing treatments for CL in children . Eight articles ( 461 patients aged 2–15 years ) were included . None of the studies enrolled children <2 years of age . Identified treatments were miltefosine , systemic and intralesional meglumine antimoniate ( MA ) , cryotherapy , and rifampicin . Sub-optimal quality and small sample sizes limit the generalizability of results from some of these studies . In general , for the Americas , oral miltefosine showed high efficacy , and in an Iranian study , intralesional MA showed higher efficacy than cryotherapy . This study provides evidence of the scarcity of data available to support treatment recommendations in children and of the unmet need to develop and test better treatment options for this vulnerable population .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"clinical",
"research",
"design",
"tropical",
"diseases",
"database",
"searching",
"randomized",
"controlled",
"trials",
"parasitic",
"diseases",
"pediatrics",
"research",
"design",
"clinical",
"medicine",
"signs",
"and",
"symptoms",
"neglected",
"tropical",
"diseases",
"pharmacology",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"zoonoses",
"lesions",
"adverse",
"events",
"protozoan",
"infections",
"diagnostic",
"medicine",
"drug",
"research",
"and",
"development",
"clinical",
"trials",
"database",
"and",
"informatics",
"methods",
"leishmaniasis"
] |
2018
|
Interventions to treat cutaneous leishmaniasis in children: A systematic review
|
Deleterious alleles have long been proposed to play an important role in patterning phenotypic variation and are central to commonly held ideas explaining the hybrid vigor observed in the offspring of a cross between two inbred parents . We test these ideas using evolutionary measures of sequence conservation to ask whether incorporating information about putatively deleterious alleles can inform genomic selection ( GS ) models and improve phenotypic prediction . We measured a number of agronomic traits in both the inbred parents and hybrids of an elite maize partial diallel population and re-sequenced the parents of the population . Inbred elite maize lines vary for more than 350 , 000 putatively deleterious sites , but show a lower burden of such sites than a comparable set of traditional landraces . Our modeling reveals widespread evidence for incomplete dominance at these loci , and supports theoretical models that more damaging variants are usually more recessive . We identify haplotype blocks using an identity-by-decent ( IBD ) analysis and perform genomic prediction analyses in which we weigh blocks on the basis of complementation for segregating putatively deleterious variants . Cross-validation results show that incorporating sequence conservation in genomic selection improves prediction accuracy for grain yield and other fitness-related traits as well as heterosis for those traits . Our results provide empirical support for an important role for incomplete dominance of deleterious alleles in explaining heterosis and demonstrate the utility of incorporating functional annotation in phenotypic prediction and plant breeding .
Understanding the genetic basis of phenotypic variation is critical to many biological endeavors from human health to conservation and agriculture . Although most new mutations are likely deleterious [1] , their importance in patterning phenotypic variation is controversial and not well understood [2] . Empirical work suggests that , although the long-term burden of deleterious variants is relatively insensitive to demography [3] , population bottlenecks and expansion may lead to an increased abundance of deleterious alleles over shorter time scales such as those associated with domestication [4] , postglacial colonization [5] or recent human migration [6] . Even when the impacts on total load are minimal , demographic change may have important consequences for the contribution of deleterious variants to phenotypic variation [3 , 7–9] . Together , these considerations point to a potentially important role for deleterious variants in determining patterns of phenotypic variation , especially for traits closely related to fitness . In addition to its global agricultural importance , maize has long been an important genetic model system [10] and central to debates about the basis of hybrid vigor and the role of deleterious alleles [11 , 12] . The maize domestication bottleneck has lead to an increased burden of deleterious alleles in maize compared to its wild ancestor teosinte [13] , and rapid expansion following domestication likely lead to an increase in new mutations and stronger purifying selection [4] . More recently , modern maize breeding has lead to dramatic reductions in effective population size [14] , but inbreeding during the development of modern inbred lines may have decreased load by purging recessive deleterious alleles [15] . Nonetheless , substantial evidence suggests an abundance of deleterious alleles present in modern germplasm , from the observed maintenance of heterozygosity during the processes of inbreeding [16 , 17] and selection [18] to genome-wide association results that reveal an excess of associations with genes segregating for damaging protein-coding variants [19] . Modern maize agriculture takes advantage of hybrid maize plants that result from the cross between two parental inbred lines [12] . These crosses result in a phenomenon known as hybrid vigor or heterosis , in which the hybrid plant shows improved agronomic qualities compared to its parents . Heterosis cannot be easily predicted from parental phenotype alone , and the genetic underpinnings of heterosis remain largely unknown . The most straightforward explanation for heterosis has been simple complementation of recessive deleterious alleles that are homozygous in one of the inbred parents [20 , 21] . While this model is supported by considerable empirical evidence [22 , 23] , it fails in its simplest form to explain a number of observations such as heterosis and inbreeding depression in polyploid plants [11 , 24 , 25] . Other explanations , such as single-gene heterozygote advantage , clearly may play an important role in some cases [26] , [27] , but mapping studies suggest such models are not easily generalizable [28] . In this study , we set out to investigate the contribution of deleterious alleles to phenotypic variation and hybrid vigor in maize . We created a partial diallel population from 12 maize inbred lines which together represent much of the ancestry of present-day commercial U . S . corn hybrids [29 , 30] . We measured a number of agronomically relevant phenotypes in both parents and hybrids , including flowering time ( days to 50% pollen shed , DTP; days to 50% silking , DTS; anthesis-silking interval , ASI ) , plant size ( plant height , PHT; height of primary ear , EHT ) , grain quality ( test weight which is a measure of grain density , TW ) , and grain yield ( GY ) . We conducted whole genome sequencing of the parental lines and characterized genome-wide deleterious variants using genomic evolutionary rate profiling ( GERP ) [31] . We then test models of additivity and dominance for each phenotype using putatively deleterious variants and investigate the relationship between dominance and phenotypic effect size and the long-term fitness consequences of a mutation as measured by GERP . Finally , we take advantage of a Bayesian genomic selection framework [32] approach to explicitly test the utility of including GERP scores in phenotypic prediction for hybrid traits and heterosis .
We formed a partial diallel population from the F1 progeny of 12 inbred maize lines ( S1 Table , S1 Fig ) . Field performance of the 66 F1 hybrids and 12 inbred parents were evaluated along with two current commercial check hybrids in Urbana , IL over three years ( 2009-2011 ) in a resolvable incomplete block design with three replicates . To avoid competition effects , inbreds and hybrids were grown in different blocks within the field . Plots consisted of four rows ( 5 . 3 m long with row spacing of 0 . 76 m at a plant density of 74 , 000 plants ha−1 ) , with all observations taken from the inside two rows to minimize effects of shading and maturity differences from adjacent plots . We measured plant height ( PHT , in cm ) , height of primary ear ( EHT , in cm ) , days to 50% pollen shed ( DTP ) , days to 50% silking ( DTS ) , anthesis-silking interval ( ASI , in days ) , grain yield adjusted to 15 . 5% moisture ( GY , in bu/A ) , and test weight ( TW , weight of 1 bushel of grain in pounds ) . We estimated Best Linear Unbiased Estimates ( BLUEs ) of the genetic effects in ASReml-R ( VSN International ) with the following linear mixed model: Y i j k l = μ + ς i + δ i j + β k i j + α l + ς i · α l + ε where Yijkl is the phenotypic value of the lth genotype evaluated in the kth block of the jth replicate within the ith year; μ , the overall mean; ςi , the fixed effect of the ith year; δij , the random effect of the jth replicate nested within the ith year; βkij , the random effect of the kth block nested within the ith year and jth replicate; αl , the fixed genetic effect of the lth individual; ςi · αl , the random interaction effect of the lth individual with the ith year; and ε , the model residuals . We calculated the broad sense heritability ( H2 ) of traits based on the analysis of all individuals ( inbred parents , hybrid progeny , and checks ) following the equation: H 2 = V G / ( V G + V G × E / i + V E / ( i × j ) ) where i = 3 ( number of years ) and j = 3 ( number of replicates per year ) . The BLUE values for each cross can be found in S1 Table; values across all hybrids were relatively normally distributed for all traits ( Shapiro-Wilk normality tests P values >0 . 05 , S1 Fig ) , though some traits were highly correlated ( e . g . Spearman correlation r = 0 . 98 for DTS and DTP , S2 Fig ) . We estimated mid-parent heterosis ( MPH ) as: M P H i j = G ^ i j − m e a n ( G ^ i , G ^ j ) where G ^ i j , G ^ i and G ^ j are the BLUE values of the hybrid and its two parents i and j . Note that for ASI , lower trait values are considered superior . General combining ability ( GCA ) was estimated following Falconer and Mackay [33] , and the estimated values can be found in S2 Table . We extracted DNA from the 12 inbred lines following [34] and sheared the DNA on a Covaris ( Woburn , Massachusetts ) for library preparation . Libraries were prepared using an Illumina paired-end protocol with 180 bp fragments and sequenced using 100 bp paired-end reads on a HiSeq 2000 . Raw sequencing data are available at NCBI SRA ( PRJNA381642 ) . We trimmed raw sequence reads for adapter contamination with Scythe ( https://github . com/vsbuffalo/scythe ) and for quality and sequence length ( ≥20 nucleotides ) with Sickle ( https://github . com/najoshi/sickle ) . We mapped filtered reads to the maize B73 reference genome ( AGPv2 ) with bwa-mem [35] , keeping reads with mapping quality higher than 10 and with a best alignment score higher than the second best one for further analyses . We called single nucleotide polymorphisms ( SNPs ) using the mpileup function from samtools [36] . To deal with known issues with paralogy in maize [15] , SNPs were filtered to be heterozygous in fewer than 3 inbred lines , have a mean minor allele depth of at least 4 , have a mean depth across all individuals less than 30 and have missing alleles in fewer than 6 inbred lines . Data on the total number of SNPs called and the rate of missing data per line are shown in S3 Table . We estimated the allelic error rate using three independent data sets: for all individuals using 41 , 292 overlapping SNPs from the maize SNP50k bead chip [14]; for all individuals using 180 , 313 overlapping SNPs identified through genotyping-by-sequencing ( GBS ) [37]; and for B73 and Mo17 using 10 , 426 , 715 SNP from the HapMap2 project [15] . Alignments and genotypes for each of the 12 inbreds are available at CyVerse ( https://doi . org/10 . 7946/P2WS60 ) . Because these parents are highly inbred , knowing their homozygous genotype also allows us to know the genotype of the F1 derived from any two of the parents . To test whether alignment to the B73 reference introduces a bias in relatedness estimation , we computed kinship matrices using both our SNP data as well as genotyping-by-sequencing data ( version AllZeaGBSv2 . 7 downloaded from ( www . panzea . org ) ) obtained from alignments to a set of sequencing reads ascertained from a broad germplasm base [38] . The two matrices were nearly identical ( Pearson’s correlation coefficient r = 0 . 995 ) , suggesting the degree of relatedness among lines is not sensitive to using B73 as the reference genome . We used genomic evolutionary rate profiling ( GERP ) [39] estimated from a multi-species whole-genome alignment of 13 plant genomes [40] including Zea mays , Coelorachis tuberculosa , Vossia cuspidata , Sorghum bicolor , Oryza sativa , Setaria italica , Brachypodium distachyon , Hordeum vulgare , Musa acuminata , Populus trichocarpa , Vitis vinifera , Arabidopsis thaliana , and Panicum virgatum; the alignment and estimated GERP scores are available at CyVerse ( https://doi . org/10 . 7946/P2WS60 ) . We define “GERP-SNPs” as the subset of SNPs with GERP score >0 , and at each SNP we assign the minor allele in the multi-species alignment as the likely deleterious allele . Finally , we predicted the functional consequences of GERP-SNPs based on genome annotation information obtained from SnpEff [41] . The multi-species alignment made use of the B73 AGPv3 assembly , and to ensure consistent coordinates , we ported our SNP coordinates from AGPv2 to AGPv3 using the Gramene assembly converter ( http://ensembl . gramene . org/Zea_mays/Tools/AssemblyConverter ? db=core ) . To compare GERP scores ( for all SNPs with GERP > 0 ) to recombination rate and allele frequencies , we obtained the NAM genetic map [42] from the Panzea website ( http://www . panzea . org/ ) and allele frequencies from the > 1 , 200 maize lines sequenced as part of HapMap3 . 2 [43] . To compare the burden of deleterious alleles in modern inbred lines to landraces , we extracted genotypic data of 23 specially-inbred traditional landrace cultivars ( see [15] for more details ) from HapMap3 . 2 . For each line , we calculated burden as the count of minor alleles present across all GERP-SNPs divided by the total number of non-missing sites . We separated sites into fixed ( present in all individuals of a group ) and segregating for landrace and modern maize samples separately . We estimated the additive and dominant effects of individual GERP-SNPs using a GBLUP model [44] implemented in GVCBLUP [45]: Y i = μ + ∑ j = 1 n X i j α j + ∑ j = 1 n W i j d j + ε where Yi is the BLUE value of the ith hybrid , μ is the average genotypic value , αj is the allele substitution effect of the jth GERP-SNP , dj is the dominant effect of the jth GERP-SNP , Xij = {2p , 2p − 1 , 2p − 2} , ε is the model residuals , and Wij = {−2p2 , 2p ( 1 − p ) , −2 ( 1 − p ) 2} are the genotype encodings for genotypes A1 A1 , A1 A2 , and A2 A2 in the ith hybrid for the jth GERP-SNP with p of the A1 allele . The additive and dominance SNP encoding ensures that the effects are independent for a given GERP-SNP . We extracted additive ( a = α − 2p ( 1 − p ) d ) and dominant ( d ) effects from the GVCBLUP output file ( see supplemantary file of Da et al . , [44] for more details ) . We first estimated the total variance explained under models of complete additivity ( d = 0 ) or complete dominance ( α = 0 ) . Then , to assess correlations between SNP effects and GERP scores , we calculated the degree of dominance ( k = d/a ) [46] for SNPs that each explained greater than the genome-wide mean per-SNP variance ( total variance explained divided by total number of GERP-SNPs ) . Because this approach can lead to very large absolute values of k , we truncated GERP-SNPs with |k = d/a|>2 for all further analyses . To compare the variance explained by our model to that explained by random SNPs , we used a 2-dimensional sampling approach to create 10 equal-sized datasets of randomly sampled SNPs ( including SNPs with GERP score < = 0 ) matched for allele frequency ( in bins of 10% ) and recombination rate ( in quartiles of cM/Mb ) . For each dataset we fit the above model separately and estimated SNP effects and phenotypic variance explained by each SNP . To test the relationship between GERP score and dominance under a simple model of mutation-selection equilibrium , we estimated the selection coefficient s by assuming that yield is a measure of fitness . We assigned the yield-increasing allele at each GERP-SNP a random dominance value in the range of 0 ≥ k ≥ 1 and calculated its equilibrium allele frequency p under mutation-selection balance using p=μs for values of k > 0 . 98 and p = 2 μ k + 1 for k ≤ 0 . 98 . We then simulated datasets using binomial sampling to choose SNPs in a sample of size n = 12 inbreds . We imputed missing data and identified regions of identity by descent ( IBD ) between the 12 inbred lines using the fastIBD method implemented in BEAGLE [47] . We then defined haplotype blocks as contiguous regions within which there were no IBD break points across all pairwise comparisons of the parental lines ( S3 Fig ) . Haplotype blocks at least 1 Kb in size were kept for further analyses . Because there is no recombination in an inbred parent , this allows us to project the diploid genotype of each F1 based on the haplotypes of the two parents . In the projected diploid genotype of each F1 , haplotype blocks were weighted by the summed GERP scores of all GERP-SNPs ( python script ‘gerpIBD . py’ available at https://github . com/yangjl/zmSNPtools ) ; blocks with no SNPs with positive GERP scores were excluded from further analysis . For a particular SNP with a GERP score g , the homozygote for the conserved ( major ) allele was assigned a value of 0 , the homozygote for the putatively deleterious allele a value of 2g , and the heterozygote a value of ( 1 + k ) × g , where k is the dominance estimated from the GBLUP model above . We used the BayesC option from GenSel4 [32] for genomic selection model training with 41 , 000 iterations . We removed the first 1 , 000 iterations as burn-in . We used the model Y i = μ + ∑ j = 1 n r j I i j + ε where Yi is the BLUE value of the ith hybrid , rj is the regression coefficient for the jth haplotype block , and Iij is the sum of GERP scores under an additive , dominance or incomplete dominance model for the ith hybrid in the jth haplotype block . We used a 5-fold cross-validation method to conduct prediction , dividing the diallel population randomly into training ( 80% ) and validation sets ( 20% ) 100 times . After model training , we obtained prediction accuracies by comparing the predicted breeding values with the observed BLUE values in the corresponding validation sets . For comparison , we permuted GERP scores using 50k SNP ( ≈ 100Mb or larger ) windows which were circularly shuffled 10 times to estimate a null conservation score for each IBD block . We conducted permutations on all GERP-SNPs as well as on a restricted set of GERP-SNPs only in genic regions to control for GERP differences between genic ( N = 221 , 960 ) and intergenic regions ( N = 123 , 216 ) . We conducted permutation cross-validation experiments using the same training and validation sets . We estimated the posterior phenotypic variance explained using all of the data to derive correlations between breeding values estimated from the prediction model and observed BLUE values . Note that the correlation used here is different from the prediction accuracy ( r ) used for the cross-validation experiments , where the latter is defined as the correlation between real and estimated values; the two statistics will converge to the same value when there is no error in SNP/haplotype effect estimation [48] . Finally , to compare our genomic prediction model to a classical model of general combining ability , we used the following equations: Y i j = μ + G C A i + G C A j + ε Y i j = μ + G C A i + G C A j + G i j + ε where Yij is the BLUE value of the hybrid of the ith and jth inbreds , μ is the overall mean , GCAi and GCAj are the general combining abilities of the ith and jth inbreds , Gij is the breeding value of the hybrid of the ith and jth inbreds as estimated by our genomic prediction model , and ε the model residuals . Sequencing data have been deposited in NCBI SRA ( SRP103329 ) database , and code for all analyses are available in the public GitHub repository ( https://github . com/yangjl/GERP-diallel ) .
We created a partial diallel population from 12 maize inbred lines which together represent much of the ancestry of present-day commercial U . S . corn hybrids ( S1 Table ) [29 , 30] . We measured a number of agronomically relevant phenotypes in both parents and hybrids , including flowering time ( days to 50% pollen shed , DTP; days to 50% silking , DTS; anthesis-silking interval , ASI ) , plant size ( plant height , PHT; height of primary ear , EHT ) , test weight ( TW; a measure of quality based on grain density ) , and grain yield ( GY ) . In an agronomic setting GY—a measure of seed production per unit area—is the primary trait selected by breeders and thus analogous to fitness . Plant height and ear height , both common measures of plant health or viability , were significantly correlated to GY ( S2 Fig ) . For each genotype we derived best linear unbiased estimators ( BLUEs ) of its phenotype from mixed linear models ( S1 Table ) to control for spatial and environmental variation ( see Methods ) . We estimated mid-parent heterosis ( MPH , Fig 1a ) for each trait as the percent difference between the hybrid compared to the mean value of its two parents ( see Methods , S1 Table ) . Consistent with previous work [28] , we find that grain yield ( GY ) showed the highest level of heterosis ( MPH of 182% ± 60% ) . While flowering time ( DTS and DTP ) is an important adaptive phenotype globally [49] , it showed relatively little heterosis in this study , likely due to the relatively narrow geographic range represented by the parental lines . We resequenced the 12 inbred parents to an average depth of ≈ 10× , resulting in a filtered set of 13 . 8M SNPs . Compared to corresponding SNPs identified by previous studies ( see Methods ) , we observed a mean genotypic concordance rate of 99 . 1% . In order to quantify the deleterious consequences of variants a priori , we made use of Genomic Evolutionary Rate Profiling ( GERP ) [39] scores of the maize genome [50] . GERP scores provide a quantitative measure of the evolutionary conservation of a site across a phylogeny that allows characterization of the long-term fitness consequences of both coding and noncoding positions in the genome [51] . Sites with more positive GERP scores are inferred to be under stronger purifying selection , and SNPs observed at such sites are thus inferred to be more deleterious . At each site with GERP scores > 0 ( hereafter called GERP-SNPs ) , we designated the minor allele from the multispecies alignment as putatively deleterious . Of the 350k total segregating GERP-SNPs in our parental lines , 14% are detected in coding regions , equally split between synonymous ( N = 64 , 439 ) and non-synonymous ( N = 65 , 376 ) sites ( S4 Table ) . Each line carries , on average , 139k potential deleterious SNPs , 19k of which are in coding regions ( S5 Table ) . The reference genome B73 contains only ≈ 1/3 of the deleterious SNPs of the other parents , likely due to reference bias in identifying deleterious variants . The F1 hybrids of the diallel each contain an average of ≈ 56 , 000 homozygous deleterious SNPs , ranging from 47 , 219 ( PH207 x PHG35 ) to 77 , 210 ( PHG84 x PHZ51 ) ( S6 Table ) . To compare the burden of deleterious variants between our elite maize lines and traditionally cultivated landraces , we used genotypes from the maize HapMap3 . 2 [43] for our diallel parents and 23 specially-inbred landrace lines [15] ( S5 Table ) . Compared to landraces , the parents of our diallel exhibited a greater burden of fixed ( allele frequency of 1 ) deleterious variants but a much smaller burden of segregating SNPs , resulting in a slightly lower overall proportion of deleterious sites ( mean of 1 . 3M deleterious alleles out of 6 . 5M total sites vs . 0 . 6/3 . 3M; Fig 1b ) . Population genetic theory predicts that deleterious variants should be at low overall frequencies , and that such variants should be enriched in regions of the genome with extremely low recombination [52] . Using data from more than 1 , 200 lines in maize HapMap3 . 2 [43] , we find that allele frequency of the minor alleles in the multi-species alignment shows a strong negative correlation with GERP score ( Fig 1c ) . This negative correlation holds using allele frequency derived from our 12 parental lines ( S4 Fig ) , though as expected is less significant given the smaller sample size . SNPs found in regions of the genome with low recombination also show higher overall GERP scores ( Fig 1d ) , a trend particularly noticeable around centromeres ( S5 Fig ) . These results match previous empirical findings in maize that deleterious alleles are rare [19] and most abundant in the lowest recombination regions [17 , 40 , 53] , and support the use of GERP scores as a quantitative measure of the long-term fitness effects of an observed variant . We first investigate the impacts of deleterious variants on phenotype using simple linear regressions . Across all hybrids , the number of homozygote GERP-SNPs was negatively correlated with grain yield , plant height , and ear-height per se ( see S6 Table for complementation data and S7 Table for correlations with all traits ) . We next applied a genomic best linear unbiased prediction ( GBLUP ) [44] modeling approach to estimate the effect sizes and variance explained by GERP-SNPs for each of the phenotypes per se across our diallel ( see Methods ) . GERP-SNPs had larger average effects and explained more phenotypic variance than the same number of randomly sampled SNPs ( including SNPs with GERP score < = 0 ) matched for allele frequency and recombination ( Fig 2a ) . We found the cumulative proportion of dominance variance explained by GERP-SNPs was higher for traits showing high heterosis ( Spearman correlation P value < 0 . 01 , r = 0 . 9 ) , from ≈ 0 for flowering time traits to as much as 24% for grain yield ( S6 Fig ) . Distributions of per-SNP dominance k = d a ( see Methods ) across traits were consistent with the cumulative partitioning of variance components ( Fig 2b ) and matched well with expectations from previous studies showing a predominantly additive basis for flowering time [54] and plant height [55] but meaningful contributions of dominance to test weight and grain yield [28 , 30] . Although our diallel population is relatively small , our estimated values explain as much ( for traits with low dominance variance like flowering time ) or more variance ( for traits with substantial dominance variance like grain yield ) than sets of data with randomly shuffled values of dominance ( n = 10 randomizations of k per trait; S7 Fig ) . We then evaluated the relationship between GERP score and SNP effect size , dominance , and contribution to phenotypic variance . We found weak or negligible correlations between effect size and GERP score for flowering time and grain quality , but a strong positive correlation for fitness-related traits ( Fig 2c and 2d ) . The variance explained by individual SNPs , however , was largely independent of GERP score ( S8 Fig ) , likely due to the observed negative correlation between allele frequency and GERP score ( Fig 1c ) . Finally , we observed a positive relationship between GERP score and the degree of dominance ( k ) for grain yield ( Fig 2e ) , such that the putatively deleterious allele at SNPs with higher GERP scores are also estimated to be more recessive for their phenotypic effects on grain yield ( larger k for the major allele ) . We investigated a number of possible caveats to the results presented in Fig 2 . First , to control for the potential inflation of SNP effect sizes in regions of high linkage disequilibrium , we removed SNPs from regions of the genome in the lowest quartile of recombination . While some individual correlations changed significance , our overall results appear robust to the removal of low recombination regions ( S9 Fig ) . Second , we tested the impact of reference bias caused by inclusion of the B73 genome in the multi-species alignment used to estimate GERP scores . To do so , we removed the 11 hybrids which include as one parent the reference genome line B73 and repeated the above analyses . Doing so dramatically reduces the size of our dataset , but we nonetheless find significant correlations between complementation and phenotype ( S7 Table ) , that GERP-SNPs explain a greater proportion of overall variation than randomly sampled SNPs ( S10a Fig ) , and that the relative pattern of dominance among traits remains the same ( S10b Fig ) . While most of the correlations between effect size and GERP score lose significance ( S10c and S10d Fig ) , likely due to the decreased sample size , the positive correlation between dominance and GERP score remains significant even in the absence of B73-derived hybrids ( S10e Fig ) . Finally , because natural selection will maintain dominant deleterious alleles at lower frequencies than their recessive counterparts , we investigated whether the ascertainment bias against rare alleles present in our small sample would lead to the observed correlation between GERP and dominance . Simulations of SNPs with random dominance at mutation-selection balance ( see Methods ) , however , failed to find any relationship between dominance and GERP score ( S11 Fig ) , though we caution that the dramatic demographic shifts involved in the recent history of maize [4] make such a simulation approximate at best . To explicitly test the informativeness of alleles identified a priori as putatively deleterious , we implemented a genomic prediction model that evaluates complementation at the haplotype level by incorporating GERP scores of individual SNPs as weights ( see Methods ) . We explored the explanatory power with several different models and found that a model which incorporates both GERP scores and dominance ( k ) estimated from our GBLUP model explained a greater amount of the posterior phenotypic variance for most traits per se ( Fig 3a ) and heterosis ( MPH ) ( Fig 3b ) . A simple additive model showed superior explanatory power for flowering time , however , consistent with previous association mapping results that flowering time traits are predominantly controlled by a large number of additive effect loci [54] . To explicitly test the utility of incorporating GERP information in prediction models , we compared cross-validation prediction accuracies of the observed GERP scores to those from datasets in which GERP scores were circularly shuffled along the genome ( see Methods ) . Models incorporating our observed GERP scores out-performed permutations ( Fig 3c and 3d ) , even when considering only SNPs in genes ( S12 Fig ) . Our model improved prediction accuracy of grain yield by more than 4 . 3% , and improvements were also seen for plant height ( 0 . 8% ) and testing weight ( 3 . 3% ) . While our model showed no improvement in predicting heterosis for traits showing low levels of heterosis ( Fig 1a ) , including GERP scores significantly improved prediction accuracy for heterosis of grain yield ( by 1% ) . Finally , our approach also significantly improved model fit for phenotypes of all traits per se as well as heterosis for GY and PHT compared to traditional models of genomic selection that use general combining ability ( see Methods , S2 Table ) calculated directly from the pedigree of the hybrid population [56] ( ANOVA FDR <0 . 01 and difference in AIC < 0 , S8 Table ) .
In this study , we use genomic and phenotypic data from a partial diallel population of maize to show that an incomplete dominance model of deleterious mutation both fits predictions of population genetic theory and explains phenotypic variation for fitness-related phenotypes and hybrid vigor . We find genome-wide support for hypotheses predicting that more damaging variants are more recessive . Finally , we show that leveraging evolutionary annotation information in silico enables us to predict grain yield and other traits , including heterosis , with greater accuracy . Together , these results help reconcile alternative explanations for hybrid vigor and point to the utility of leveraging evolutionary history to facilitate breeding for crop improvement .
|
A key long-term goal of biology is understanding the genetic basis of phenotypic variation . Although most new mutations are likely disadvantageous , their prevalence and importance in explaining patterns of phenotypic variation is controversial and not well understood . In this study we combine whole genome-sequencing and field evaluation of a maize mapping population to investigate the contribution of deleterious mutations to phenotype . We show that a priori prediction of deleterious alleles correlates well with effect sizes for grain yield and that variants predicted to be more damaging are on average more recessive . We develop a simple model allowing for variation in the heterozygous effects of deleterious mutations and demonstrate its improved ability to predict both phenotypes and hybrid vigor . Our results help reconcile alternative explanations for hybrid vigor and highlight the use of leveraging evolutionary history to facilitate breeding for crop improvement .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"biotechnology",
"alleles",
"plant",
"science",
"model",
"organisms",
"mathematics",
"forecasting",
"statistics",
"(mathematics)",
"experimental",
"organism",
"systems",
"plant",
"genomics",
"molecular",
"genetics",
"plants",
"research",
"and",
"analysis",
"methods",
"grasses",
"mathematical",
"and",
"statistical",
"techniques",
"maize",
"plant",
"genetics",
"molecular",
"biology",
"genetic",
"loci",
"inbreeding",
"heterosis",
"eukaryota",
"plant",
"and",
"algal",
"models",
"phenotypes",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"plant",
"biotechnology",
"statistical",
"methods",
"organisms"
] |
2017
|
Incomplete dominance of deleterious alleles contributes substantially to trait variation and heterosis in maize
|
To improve our knowledge on the epidemiological status of African trypanosomiasis , better tools are required to monitor Trypanosome genotypes circulating in both mammalian hosts and tsetse fly vectors . This is important in determining the diversity of Trypanosomes and understanding how environmental factors and control efforts affect Trypanosome evolution . We present a single test approach for molecular detection of different Trypanosome species and subspecies using newly designed primers to amplify the Internal Transcribed Spacer 1 region of ribosomal RNA genes , coupled to Illumina sequencing of the amplicons . The protocol is based on Illumina’s widely used 16s bacterial metagenomic analysis procedure that makes use of multiplex PCR and dual indexing . Results from analysis of wild tsetse flies collected from Zambia and Zimbabwe show that conventional methods for Trypanosome species detection based on band size comparisons on gels is not always able to accurately distinguish between T . vivax and T . godfreyi . Additionally , this approach shows increased sensitivity in the detection of Trypanosomes at species level with the exception of the Trypanozoon subgenus . We identified subspecies of T . congolense , T . simiae , T . vivax , and T . godfreyi without the need for additional tests . Results show T . congolense Kilifi subspecies is more closely related to T . simiae than to other T . congolense subspecies . This agrees with previous studies using satellite DNA and 18s RNA analysis . While current classification does not list any subspecies for T . godfreyi , we observed two distinct clusters for these species . Interestingly , sequences matching T . congolense Tsavo ( now classified as T . simiae Tsavo ) clusters distinctly from other T . simiae Tsavo sequences suggesting the Nannomonas group is more divergent than currently thought thus the need for better classification criteria . This method presents a simple but comprehensive way of identification of Trypanosome species and subspecies-specific using one PCR assay for molecular epidemiology of trypanosomes .
Human African trypanosomiasis ( HAT ) or sleeping sickness is classified as a neglected tropical disease by WHO , that is endemic in sub-Sahara Africa . HAT affects impoverished rural areas of sub-Saharan Africa , where it coexists with animal trypanosomiasis constituting a major health and economic burden [1] . The disease is caused by protozoan parasites of the genus Trypanosoma , it is transmitted by the bite of blood-sucking tsetse flies ( Diptera , genus Glossina ) . The human disease is caused by Trypanosoma brucei rhodesiense and Trypanosoma brucei gambiense , causing an acute and chronic disease in humans respectively [2] . T . b . rhodesiense is found in East Africa and transmitted by Glossina morsitans , while T . b gambiense is distributed in West Africa and is mainly transmitted by Glossina pallidipes [3–5] . Uganda is the only country that both forms of the disease occur with the potential for overlapping infections [6] . According to WHO , the incidence of sleeping sickness has fallen over the years , from 10 , 388 cases reported in 2008 to 2 , 804 cases reported in 2015 [7] . However , WHO estimates the number of actual cases to be below 20 , 000 [8] . This decrease is attributed to improved case detection and treatment and vector management [9] . Despite this decreased incidence , it is estimated that up to 70 million people distributed over 1 . 5 million km2 remain at risk of contracting the disease [10] . Besides , African animal trypanosomiasis ( AAT ) is one of the biggest constraints to livestock production and a threat to food security in sub-Saharan Africa . The parasites T . congolense ( Savannah ) and T . vivax are considered the most important animal Trypanosomes due to their predominant distribution in sub-Saharan Africa and their economic impact [11] . They cause pathogenic infections in cattle ( Nagana ) and also infect sheep , goats , pigs , horses , and dogs , while T . brucei brucei ( and T . brucei rhodesiense ) is pathogenic to camels , horses , and dogs , but causes mild or no clinical disease cattle , sheep , goats and pigs [12–14] . T . simiae causes a fatal disease in pigs and mild disease in sheep and goats . T . godfreyi shows a chronic , occasionally fatal disease in pigs experimentally [15 , 16] . T . evansi was originally found to infect camels but it is present in dromedaries , horses , and other equines as well as in a wide range of animals causing Surra disease , while T . equiperdum causes dourine in equines [17] . Three species ( T , evansi , T . vivax , and T . equiperdum ) are independent of the tsetse fly vector and thus distributed outside Africa [18 , 19] . Their transmission is either mechanically , for T . evansi and T . vivax , or sexually for , T . equiperdum . T . vivax can be transmitted cyclically by Glossina spp . and mechanically and therefore can found in both tsetse-infested and tsetse-free areas [20] . Given that Trypanosome parasites are maintained in wild and domestic animals as reservoirs , this complicates control measures . Morphological methods have limited ability to distinguish between Trypanosome species due to the existence of trypanosomes sharing developmental sites , and mixed and immature infections . Thus , molecular methods are used for species identification . Identification of Trypanosome species and subspecies is important to interrogate aspects such as what contribution different species/subspecies make to livestock disease and , are species/subspecies differences responsible for assumed “strain” differences in drug response among others . The ribosomal RNA sequence region harboring internal transcribed spacer sequences have been used to identify Trypanosome species in hosts and vectors . Epidemiological and screening studies rely on polymerase chain reaction ( PCR ) to amplify the internal transcribed spacer 1 ( ITS1 ) region of ribosomal genes to analyze Trypanosome species diversity [16–19] . This locus located between the 18s and 5 . 8s ribosomal subunit genes which are about 100–200 copies [21] and is widely used to identify Trypanosome species based on amplicon size in [22] a gel . However , ITS1 PCR coupled with viewing products on agarose gels fails to distinguish some species/genotypes such as T . simiae and T . simiae Tsavo . Another limitation with ITS1 PCR is the sensitivity of detection , showing bias in detection of some Trypanosome species over others [23 , 24] . Some are prone to non-specific amplification particularly in bovine blood samples [25] . To address some of the problems that ITS PCR method poses , fluorescent fragment length barcoding ( FFLB ) method has been developed for Trypanosome species detection [26] . FFLB is based on length variation in regions of the 18s and 28s ribosomal RNA gene region . Fluorescently tagged primers , designed in conserved regions of the 18s and 28s ribosomal RNA genes , are used to amplify fragments with inter-species size variation , and sizes are determined accurately using an automated DNA sequencer . FFLB has been shown to be more sensitive in the identification of Trypanosome species and subspecies and has the capacity to detect new species through identification of unique barcodes [27 , 28] . However , the method requires the use of four different PCR reactions per sample . A major problem with identification of Trypanosome species with the use of ribosomal RNA genes is that they cannot be used to distinguish between Trypanozoon species ( T . brucei brucei , T . brucei rhodesiense , T . brucei gambiense , T . evansi , and T . equiperdum ) [22 , 26 , 29 , 30] . Currently , Trypanozoon subspecies are identified by specific PCR [31–34] and microsatellites markers [32 , 35 , 36] . When dealing with a large number of samples either for tsetse fly or animal infection prevalence studies , undertaking multiple PCRs for each sample is an expensive and a laborious undertaking . Most often PCR amplicons are sequenced to confirm species identification usually through capillary sequencing . Recently , next-generation sequencing ( NGS ) has been established as a well-established method for profiling bacterial and fungal , communities . Among the many advantages , NGS provides a higher sensitivity to detect low-frequency variants , the lower limit of detection of DNA , higher throughput with sample multiplexing and comprehensive coverage among others . With the exception of Plasmodium in mosquitoes , relatively few studies have applied this technology in the diagnostics of protozoal infections [37 , 38] . It is therefore suited in the analysis of the genetic diversity of Trypanosome genotypes which is a composite aspect of understanding anthropogenic disturbance that may change repertoires of trypanosomes infecting human and livestock [39] .
For this study , we analyzed tsetse fly samples from three different groups collected at three different locations ( Fig 1 ) at different times . The first group was used for the initial analysis and to validate our method and consisted of 188 tsetse flies collected from the area around Hurungwe Game reserve in Zimbabwe between March and April 2014 . The second group was included in our final analysis to expand Trypanosome species spectrum and diversity and consisted of 200 tsetse flies from Rufunsa area ( Zambia ) near Lower Zambezi National park ( surrounding farms and villages ) collected in November and December 2013 ) . For these samples , information on tsetse fly species and sex was not available . The third group comprised of 85 flies caught in Zambia; on the border between Kafue National park and public settlement area , collected in June 2017 . For this group , flies were sorted according to sex and their species identity determined morphologically . Flies from all three groups had been collected using either custom-made mobile traps attached on a slow-moving vehicle ( Kafue and Rufunsa groups ) or Epsilon traps ( Hurungwe group ) . Individual flies were preserved in separate tubes containing silica gel ready for crushing and DNA extraction . All flies analyzed in this study were caught on public land . DNA extraction from all tsetse fly samples analyzed in this study was done by following a protocol adopted for extraction of DNA from crushed tsetse fly samples . Briefly , dried flies in tubes containing stainless beads were transferred to a smashing machine and crushed at 3 , 000 rpm for 45 sec . DNA from crushed flies was isolated using the DNA Isolation kit for mammalian blood ( Roche USA ) as per the manufacturer’s protocol with the slight modification suggested for extraction of DNA from Buffy coat , where Red blood cell lysis step is bypassed . This allows lysis of all cells in the solution at once including trypanosomes using the white cell lysis buffer . The DNA sample was stored at -80°C until analysis . The following sequences were retrieved from NCBI , Trypanosoma brucei ( JX910378 , JX910373 , JN673391 , FJ712717 , AF306777 , AF306774 , AF306771 and AB742530 ) , Trypanosoma vivax ( JN673394 , KC196703 and TVU22316 ) , Trypanosoma congolense ( JN673389 , TCU22319 , TCU22318 , TCU22317 and TCU22315 ) , Trypanosoma simiae ( JN673387 and TSU22320 ) , Trypanosoma godfreyi ( JN673385 ) Trypanosoma evansi ( D89527 ) , Trypanosoma otospermophili ( AB175625 ) , and Trypanosoma grosi ( AB175624 ) . They were aligned in Geneious 9 . 1 . 5 software ( Biomatters Ltd , Auckland , New Zealand ) using MAFFT multiple aligner with default settings and ITS1 region identified by comparing annotations and terminal regions of 18s and 1 . 5s ribosomal RNA regions . Pairs of primers flanking the ITS1 region were picked manually based on the consensus of bases in the alignment flanking the ITS1 region . Manual editing was done on the final primer pair that was chosen , to improve the range of Trypanosome species and subspecies . We used Primer-BLAST ( https://www . ncbi . nlm . nih . gov/tools/primer-blast ) to confirm that the primers would amplify the target species , check the species range and the melting temperature . The final pair comprised our new primers named Amplification of ITS ( AITS ) forward ( AITSF ) and reverse ( AITSR ) .
In silico evaluation of the primers showed that our newly designed primer pair ( AITSF/AITSR ) had a broad range similar to previously developed ITS1/ITS2 primer set [41] while the range of the CF/BR primer set , previously developed to detect pathogenic Trypanosomes [24] was confined to the pathogenic ( S1 Table ) . We evaluated the sensitivity of newly designed AITSF/AITSR primers to amplify ITS1 region of different Trypanosome species in comparison to commonly used ITS1 primers; CF/BR primers . PCR was performed on pGEMT-easy plasmid DNA containing ITS1 inserts from different Trypanosome species at different dilutions . Our evaluation was based on the visual sight of bands in a gel ( the conventional method of analysis ) . Our results showed that AITSF/AITSR primers were slightly more sensitive in the detection of T . brucei , T . simiae and T . congolense ( S1 Fig ) . AITSF/AITSR primers could detect 103 T . brucei , T . simiae , T . vivax and T . congolense and T . godfreyi ITS1 copies while CF/BR primers could detect 103 T . godfreyi and T . vivax ITS1 copies , 104 T . simiae and T . congolense ITS1 copies and 105 T . brucei ITS1 copies . Trypanosomes have about 115 copies of ribosomal RNA genes [21] . Reads generated from amplicon sequencing were of relatively good quality . Apart from those from Zimbabwe , more than 90% of the reads passed quality filtering in all samples ( Table 2 ) . The no . of ASVs generated in replicate runs was slightly different indicating slightly different detection sensitivities in the replicate PCR runs . Only the forward read was retained for downstream analysis in reads that did not merge due to either amplicon being longer than 600 b . p or due to low-quality bases in the overlap bases . This did not affect the final identification of reads as shown by the simulated data results described later . We analyzed the Rufunsa samples in replicates and compared the results . Both replicates had similar results in regard to individual Trypanosome species detection per sample seen in the gel image analysis ( Fig 3A ) as well as amplicon read analysis ( Fig 3B ) . The outcome of detection for each of the Trypanosome species and subspecies in replicate runs was comparable and the Fischer’s exact test confirmed that there was no significant difference ( P<0 . 05 ) in the number of positive detections in replicate runs ( S2 Table ) . Simulation of data generated from Trypanosome sequences downloaded from NCBI and analyzed using the AMPtk ( amplicon toolkit ) pipeline ( version 1 . 2 . 4 ) ( https://github . com/nextgenusfs/amptk ) showed that amplicon sequence variants ( ASVs ) generated by the pipeline as primary units of representing sequence diversity , were more accurate in correctly inferring the diversity sequences compared to operational taxonomic units ( OTUs ) derived from clustering sequences at 97% identity ( S3 Table ) . The specificity and precision of distinguishing between individual sequences of the same Trypanosome species are reflected by the number of ASVs or OTUs representing each of the different species . For example , only one OTU was generated for all three Trypanosoma theileri sequences , and three OTUs were generated for seven Trypanosoma simiae sequences , while the number of ASVs generated in each case represented each sequence accurately . The simulated data results indicated that read analysis using the AMPtk pipeline and ASVs instead of OTUs was suitable for sensitive identification of Trypanosome reads . By comparing gel images after PCR and sequence data , it was observed that the sensitivity of detection of Trypanosome DNA was increased by sequencing . Samples with bands that were barely visible after the 1st PCR became visible after the 2nd PCR and were confirmed as positive after sequencing ( Fig 4A ) . It was also observed that some T . godfreyi and T . vivax amplicon bands were of a relatively similar size and it was difficult to distinguish the two by gel analysis alone ( Fig 4B ) . From this example , sample no . 10 has an ITS amplicon size of about 400 b . p similar to that of sample no . 6 and 8 . Sequence analysis showed that the band in sample no . 10 was identified as T . vivax while bands observed in sample no . 6 and 8 were identified as T . godfreyi despite their similar sizes . Mixed and single infections with multiple and single bands respectively were observed and confirmed by amplicon sequence analysis . Results for the second PCR using dual-index primers showed consistency with those of the first PCR . There were no bands visible outside the expected range indicating the absence of non-specific amplification in both PCR steps . The 1st PCR amplicons were slightly longer than expected sizes due to the adapter sequences ( approx . 80 bp ) added to the primer , therefore the bands observed corresponded to T . congolense ( Kilifi/Forest and Savannah ) ; 650–800 b . p , T . brucei; 520–540 bp , T . simiae; 440–500 bp , T . godfreyi; 320–400 bp , and T . vivax; 290–400 bp . The accuracy in distinguishing between Trypanosome species and subspecies was analyzed by phylogenetic analysis of ASV sequences and their species identity allocated by BLAST . ASVs were named after the area of collection of the sample they originated from , ASV number allocated during analysis , accession number and the taxonomic name of their respective top hit BLAST subject sequence . Phylogenetic analysis of all ASVs obtained from this study showed that ASVs named after same Trypanosome species clustered together regardless of sample collection location . Sub-clustering into different subspecies of the same species was also observed ( Fig 5 ) . The Nannomonas subgenus showed the highest diversity of sub-clustering where T . simiae clustered into two main subspecies; T . simiae and T . simiae Tsavo . Two T . simiae Tsavo II ASVs from Kafue , with 91% and 97% identity to T . congolense Tsavo ( Accession number U22318 ) recently reviewed and classified as T . simiae Tsavo [49 , 50] clustered distinctly from the rest of the T . simiae Tsavo I ASVs . T . congolense ASVs showed the highest diversity and clustered into three main subspecies; Kilifi , Riverine/Forest , and Savannah . T . congolense Savannah represented the most diversity in all the ASVs analyzed from all the samples . T . congolense Kilifi clustered separately and far from T . congolense Savannah and Riverine/Forest subspecies . T . godfreyi showed sub-clustering into two main subspecies while T . vivax ( belonging to the Dutonella subgenus ) also clustered into two subspecies . It was expected that the Trypanozoon subgenus ( T . brucei/T . evansi ) did not show any distinct sub-clustering . The prevalence of Trypanosome infection in tsetse flies caught in the Rufunsa area , Zambia , was 25 . 6% , that of in the Kafue area , also Zambia , 28 . 2% , while that of the Hurungwe area , Zimbabwe , was 47 . 3% . Flies caught in Rufunsa had the highest prevalence of T . congolense while those from Kafue had the highest prevalence of T . godfreyi ( Table 3 ) . The highest prevalence of T . brucei/ T . evansi was recorded in flies caught in Hurungwe . We did not detect any T . brucei/ T . evansi from flies collected in Kafue . Mixed infections were predominant in flies caught in Rufunsa and Hurungwe while flies caught in Kafue were predominantly infected with T . godfreyi ( Fig 6 ) . Only tsetse flies from the Kafue region were sorted by sex during collection and we observed that the infection rate in female flies ( 38 . 6% ) was more than twice that of male flies ( 17 . 1% ) . Additionally , we did not detect T . congolense and T . vivax infections in male flies . Flies caught in Hurungwe did not have single infections with T . congolense or T . godfreyi .
This study reports a new and versatile approach for detection of Trypanosome DNA in multiple samples with high sensitivity and precision than conventional PCR-gel approach . We have established that conventional ITS PCR gel analysis is not an accurate way of determining the prevalence of Trypanosome species infections since identification of species by band size is inaccurate and may lead to misidentification of some Trypanosome species . Our new approach is sensitive at the subspecies level and has a high capacity to process large amounts of samples in one run ( approximately a 700 samples mixed library ) owing to the high repertoire of Illumina dual indexing primers . However , we did not see any unique clusters that could distinguish between the Trypanozoon subspecies which are of high priority because 1 ) they cause HAT ( T . b rhodesiense and T . gambiense ) and 2 ) their distribution is not restricted to Africa ( T . evansi and T . equiperdum ) . However , we did identify two clusters of T . vivax . This is important since T . vivax is distributed outside Africa since it can be transmitted both cyclically by tsetse flies and also mechanically . Failure to distinguish between Trypanozoon subspecies was expected since the ribosomal RNA genes are highly conserved in this subgenus and cannot be able to tell apart the subspecies [29 , 30] . Moreover , a study based on genome-wide SNP analysis of 56 Trypanozoon genomes , including eight T . evansi and four T . equiperdum has revealed extensively similar genomes [51] . A single molecular test able to distinguish between members of the Trypanozoon subspecies is yet to be developed thus , subspecies specific based tests remain obligatory for their identification . As part of this work , we have also developed new primers that show high sensitivity to T . brucei compared to conventional primers and cover a wider range of the Trypanosoma genus . With our approach , it is now possible to identify species and subspecies of Trypanosomes by sequence analysis on individual samples as opposed to pooled samples for a large dataset which allows for the detection of new isolates . It is also possible to make a better inference of the Trypanosome species circulating in an area . This approach is practical and , with the decreasing cost of next-generation sequencing , cost-effective way to monitor large field samples of all kinds . They can , therefore , be utilized in a wide range of samples from vectors and hosts and the analysis of new Trypanosome species . The results obtained in this study indicate that T . vivax and T . godfreyi have very similarly sized ITS1 amplicons making it difficult to identify one from the other based solely on gel band sizes . Sequencing and clustering of the reads effectively address this issue . Phylogenetic analysis shows several interesting population substructures in the cases of T . simiae and T . congolense . Within the T . congolense clade , Savannah and Riverine/Forest subspecies show more sequence similarity while the Kilifi type shows more divergence . This agrees with a previous study that found T . congolense Savannah and Riverine/Forest had 71% similarity in satellite DNA sequence [52] and that the Kilifi subspecies was as divergent from other T . congolense subspecies [53] . The clustering of T . congolense Kilifi close to T . simiae species than other T . congolense subspecies is quite interesting in that an earlier study had identified a new T . congolense Tsavo strain ( Accession number U22318 ) [54] which has been classified as T . simiae Tsavo [55] . We identified two ASVs from Kafue area ( classified as T . simiae Tsavo II in this study ) that had 91% and 97% identity to the U22318 T . congolense Tsavo sequence and that clustered with T . simiae Tsavo rather than other T . congolense species sequences supporting the T . simiae Tsavo classification . However , they cluster separately from the other T . simiae Tsavo ASVs , suggesting that they may have a divergent genotype . Perhaps there is a complex relationship between T . congolense and T . simiae species yet to be identified . Prevalence of Trypanosome infection in caught tsetse flies differed in the sampled areas with single and mixed infection being detected in flies caught agreeing with previous studies [37 , 56 , 57] . This may be an important factor in the exchange of information between species . We also observed that the infection rate of female tsetse flies was more than twice that of male flies . This result is in contrast to dissection data from the Tinde experiment where male Glossina morsitans centralis had a salivary gland infection rate ( 5 . 4% ) more than twice that of females ( 2 . 1% ) [58] . However , our results agree with other studies on Glossina morsitans , reporting high infection rates in female flies compared to males [59 , 60] . More research is needed to find out the role of sex and infection rate differences between the different Glossina species in both laboratory and wild caught flies . To conclude , our results imply that with this approach , it is possible to detect and distinguish between different Trypanosome species and subspecies accurately ( with the exception of Trypanozoon subgenus ) and therefore infer prevalence of infection more precisely using a single test without having to undertake satellite DNA analysis that requires species-specific primers . This is made possible by deep sequencing which enables resolution at a single nucleotide level . This high resolution at sub-cluster level utilizing only the ITS1 region has not been shown before thus a practical and sensitive barcoding of African trypanosomes . Using our approach , it is thus possible to distinguish T . godfreyi from T . vivax , as well as highlight finer subpopulation structures within the T . simiae and T . congolense clades that raise interesting questions regarding their classification . It is highly likely that there are genomic and taxonomic differences between T . vivax , T . godfreyi and T . congolense subspecies that need to be studied . This could provide answers on the evolution of Trypanosomes such as; what contribution do these Trypanosome subspecies make to livestock disease ? Are these genotypes responsible for assumed “strain” differences in drug response ? Can these new genotypes be correlated with the old morphological criteria and species designations ? Do these “strains” have the potential of evolving to new subspecies that could pose new risks ? There is a need for more studies to catch up with the molecular taxonomy to answer these questions .
|
Tsetse flies are central actors in the transmission of Trypanosomes to vertebrate hosts . Therefore , detection of Trypanosomes in the tsetse flies is important for understanding the epidemiology of African trypanosomiasis as a component of new control or surveillance strategies . We have developed a method that combines multiplex PCR and next-generation sequencing for the detection of different Trypanosome species and subspecies . Similar to the widely used bacterial metagenomic analysis protocol , this method uses a modular , two-step PCR process followed by sequencing of all amplicons in a single run , making sequencing of amplicons more efficient and cost-effective when dealing with large sample sizes . As part of this approach , we designed novel Internal Transcribed Spacer 1 primers optimized for short read sequencing and have slightly better sensitivity than conventional primers . Taxonomic identification of amplicons is based on BLAST searches against the constantly updated NCBI’s nt database . Our approach is more accurate than traditional gel-based analyses which are prone to misidentification of species . It is also able to discriminate between subspecies of T . congolense , T . simiae , T . vivax , and T . godfreyi species . This method has the potential to provide new insights into the epidemiology of different Trypanosome genotypes and the discovery of new ones .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"trypanosoma",
"congolense",
"parasitic",
"protozoans",
"animals",
"glossina",
"organisms",
"trypanosoma",
"brucei",
"protozoans",
"trypanosoma",
"vivax",
"molecular",
"biology",
"techniques",
"tsetse",
"fly",
"insect",
"vectors",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"infectious",
"diseases",
"artificial",
"gene",
"amplification",
"and",
"extension",
"bioinformatics",
"blast",
"algorithm",
"molecular",
"biology",
"insects",
"disease",
"vectors",
"arthropoda",
"trypanosoma",
"eukaryota",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"trypanosoma",
"brucei",
"gambiense",
"polymerase",
"chain",
"reaction"
] |
2019
|
A single test approach for accurate and sensitive detection and taxonomic characterization of Trypanosomes by comprehensive analysis of internal transcribed spacer 1 amplicons
|
Phylogenetic inference is an attractive means to reconstruct transmission histories and epidemics . However , there is not a perfect correspondence between transmission history and virus phylogeny . Both node height and topological differences may occur , depending on the interaction between within-host evolutionary dynamics and between-host transmission patterns . To investigate these interactions , we added a within-host evolutionary model in epidemiological simulations and examined if the resulting phylogeny could recover different types of contact networks . To further improve realism , we also introduced patient-specific differences in infectivity across disease stages , and on the epidemic level we considered incomplete sampling and the age of the epidemic . Second , we implemented an inference method based on approximate Bayesian computation ( ABC ) to discriminate among three well-studied network models and jointly estimate both network parameters and key epidemiological quantities such as the infection rate . Our ABC framework used both topological and distance-based tree statistics for comparison between simulated and observed trees . Overall , our simulations showed that a virus time-scaled phylogeny ( genealogy ) may be substantially different from the between-host transmission tree . This has important implications for the interpretation of what a phylogeny reveals about the underlying epidemic contact network . In particular , we found that while the within-host evolutionary process obscures the transmission tree , the diversification process and infectivity dynamics also add discriminatory power to differentiate between different types of contact networks . We also found that the possibility to differentiate contact networks depends on how far an epidemic has progressed , where distance-based tree statistics have more power early in an epidemic . Finally , we applied our ABC inference on two different outbreaks from the Swedish HIV-1 epidemic .
Infectious diseases that are directly transmitted spread over contact networks , where each individual host can be represented by a node with a finite set of contacts ( edges ) via which they can transmit the pathogen . The structure of these networks is a major determinant of the pathogen transmission dynamics and possible control strategies [1] . For example , it has been suggested that human sexual contact networks are characterized by a power-law-like degree distribution [2–4] which , in a specific range of the scaling exponent , results in an infinite variance of the network’s degree distribution . This implies the absence of an epidemic threshold , making prophylactic strategies for sexually transmitted diseases very challenging . The main issue with contact network-based epidemiology has been the difficulty of collecting individual- and population-level data needed to develop an accurate representation of the underlying host population’s contact structure . This has led to an interest in methods to infer information about host contact networks from epidemiological data . Previously , Britton and O’Neill [5] estimated the parameters of an Erdős-Rényi network and a stochastic epidemic process on it using recovery times of infected hosts , and Groendyke et al . extended the approach to exponential-family random graph models [6] using covariate information [7] . The use of other common epidemiological measures such as the basic reproduction number ( R0 ) , epidemic peak size , duration and final size , has been shown to be effective in classifying the degree of heterogeneity in a population’s unobserved contact structure [8] . During the course of an epidemic , the pathogen spreads over a subset of edges in the social network forming a subgraph that is the realized transmission history . Keeping track of who transmits to whom and assuming that every individual may be infected only once and by only one other individual , such a transmission history can be represented as a rooted tree ( transmission tree ) [9] . However , full transmission histories are rarely observed and commonly available epidemiological data such as diagnosis-recovery times of infected people may provide information on who was infected , when , and for how long , but it provides limited information on who acquired infection from whom [10] . Since pathogens evolve over a transmission history , the analysis of pathogen genetic sequences taken from different hosts provides a way to infer the most likely donor and recipient [11] introducing constraints on the space of possible transmission trees , which are a trace of the underlying contact network . Phylodynamics [12] focuses on linking methods of phylogenetic analysis with epidemiological models under the assumption that if the evolution of a pathogen occurs sufficiently fast , transmission histories become “recorded” in the between-host pathogen phylogeny ( phylogenetic tree ) . Phylodynamic analyses of HIV-1 have shown that asymmetry in viral phylogenies may be indicative of heterogeneity in transmissions [13]; networks with more heterogeneous degree distributions yield transmission trees with smaller mean cluster sizes , shorter mean branch lengths , and somewhat higher tree imbalance than networks with relatively homogeneous degree distributions . However , it has been argued that these direct effects are relatively modest for dynamic networks [14] or if only a small fraction of infected individuals are sampled [15] . Also , factors other than contact rate , such as high infectiousness during acute infection , may have a more dramatic impact on asymmetry [15] . However , previous studies as well as more recent papers [16–18] , assume that the unobserved transmission tree is identical to the reconstructed time-scaled phylogeny ( virus genealogy ) , i . e . the internal nodes of the genealogy correspond to transmission events between hosts over time and within-host diversity is fundamentally ignorable . This is unrealistic since all coalescent events in a pathogen phylogeny occur within hosts , pushing the genealogy node heights further back in time than the nodes of the transmission tree , known as the pre-transmission interval [19] . In addition , the order of coalescent events may not correspond to the order of transmission events but rather reflect within-host dynamics [20–22] . The objective of this study was to include within-host evolution , disease stage , and individual specific transmission rates to improve the realism of social network reconstruction . We simulated epidemic spread on three prototypic network types and investigated the behavior of several tree statistics , including both topological imbalance measures and tree-based distance measures . In addition , we investigated the effect of varying epidemic size , varying sampling proportion , as well as heterochronous sampling on the tree statistics . Finally , we analyzed data from two different epidemiological scenarios of spread among injecting drug users ( IDU ) in the Swedish HIV-1 epidemic using approximate Bayesian computation ( ABC ) for network model choice and parameter inference following the algorithm defined in [23] . We found that virus geneaologies can differ from the underlying transmission tree in both topology and branch length and , therefore , that meaningful inference of social networks needs to take within-host evolution into account .
The phylogeny of pathogens such as HIV-1 collected from infected persons in an epidemic reveals a considerable amount of information about the underlying transmission history since mutations are typically accumulated faster than transmission occurs . The common assumption that the internal nodes of a phylogeny correspond to transmission events between hosts over time is , however , unrealistic because transmitted lineages must already exist in the donor at the time of transmission . Thus , neglecting the time difference between the common ancestor and the transmission event ( i . e . the pre-transmission interval , [19] ) will bias the estimated time of transmission backwards in time . Furthermore , new infections may come from HIV-1 variants derived from a latent reservoir ( lineages can persist for long time in the host [35 , 36] ) , and the order of coalescent events may not correspond to the order of transmission events but reflect instead within-host dynamics [21] . To address these issues , we used a two-phase coalescent model described by a linear growth from a single transmitted variant ( transmission bottleneck ) to a maximum population size followed by either stabilization or decline of the effective population size [21] . Let N ( t ) denote the viral population size at time t since infection ( expressed in days ) , such that N ( t ) = α 1 + β 1 t , t ≤ t x α 1 + β 1 t x + β 2 ( t - t x ) , t > t x where α1 is the population size ( i . e . the number of virus variants in a given host ) at the moment of infection , β1 is the rate of population size increase until tx ( time at maximum diversity ) , and β2 the rate of decline after the maximum . We assume α1 = 1 , β1 = 3 , tx uniformly distributed between ta = 2 and tb = 8 ( years ) and β2 ∼ U ( ϕ , 0 ) where ϕ = ( Nmin − α1 − β1ta ) / ( tM − ta ) with Nmin being the minimum population size , ( assumed to be 100 ) and tM the maximum sampling time ( 20 years ) ( S2 Fig ) . Virus genealogies conditional on a transmission history are simulated by generating random coalescence times for each person in the tree . Random coalescence times are generated from the inverse cumulative density function ( derivation in [21] ) F−1 ( u ) =1− ( 1−u ) b ( k2 ) ( a+bt ) b−1 where u is a uniform random variate on ( 0 , 1 ) , t is the current time along the forward time axis , b is the linear rate of change ( β1 or β2 depending on the phase ) , a is the starting population size ( 1 in the first phase and β1tx in the second phase ) and k is the number of extant sampled lineages in a given host . For each host we draw random values of tx and β2 from the prior distributions . Starting at the last transmission or sampling event , we first move to the next event along the reverse time axis , which is either a transmission event , a rate change , or the time at which the current host was infected . If the event is a transmission event , then k is incremented and a random coalescence time is generated . If that time occurs before ( along the reverse time axis ) the next event , then two random extant lineages in the sample are selected to coalesce; or if not , then time is moved to the time of the next event . At tx , when the rate changes , the parameters of the inverse cumulative density function are changed to correspond to the first model-phase and the process continues until the transmission time of the current host is reached . In the rare instance where more than one sampled lineage exists at the time of infection , the existing lineages are randomly coalesced with zero length branches . Finally , each individual sub-tree is joined into a single viral genealogy according to the transmission history . The 4 model specifications introduced in the previous section were used for simulations until the “end” of each outbreak , i . e . when there are no infectives left . We compared outbreaks of similar final size and multiple realizations of virus genealogies for each transmission history . All simulations were implemented using the statistical software R [37] . Transmission trees and virus genealogies are complex objects . Therefore , in order to evaluate and compare them we used a number of summary statistics ( Table 1 ) . These tree statistics include topology measures , branch length summaries , and lineages through time progression . The Sackin’s index can be normalized according to a reference model ( we used the Yule model ) in order to obtain a statistic that does not depend on the tree size . Both Sackin’s index and Colless index depend only on the topology of the tree , and they are invariant under isomorphisms and relabeling of leaves . They reach their maximum value at caterpillars ( ladder-like trees ) , and their minimum on the maximally balanced trees . A binary tree is considered to be perfectly balanced if each internal node in the tree divides the leaves descending from it into two equally sized groups . The expected number of cherries in a tree with n taxa under a Yule model is n/3 . In an asymmetric tree ( more ladder-like tree ) , tips tend to coalesce with branches deeper in the tree , and there are fewer cherries than expected . The number of cherries and Sackin’s index complement each other well , as the number of cherries captures asymmetry in the recent evolutionary past , while Sackin’s index captures asymmetry over the entire evolutionary history of the sample . These two measures are only weakly correlated [15] . A high ratio of internal branch to external branch length occurs in ‘star-like’ trees . The tree height in a virus genealogy represents the time from the first infection to the last sampling event . Since epidemics progress at different speed on different networks , heterogeneities in tree heights are expected . The topological distance was obtained as twice the number of internal branches defining different bipartitions of the tips . A topological distance that takes branch lengths into account was also considered ( the sum of the branch lengths that need be erased to have two similar trees . ) The number of lineages through time was normalized in time and by the maximum number of lineages [40] . We used the R package ape ( Analyses of Phylogenetics and Evolution ) [41] to create and plot the phylogenies and the package apTreeshape [42] for the evaluation of some tree statistics . To further investigate how well time-scaled phylogenies can estimate the epidemic process and identify the underlying contact network , we applied an inference framework for model selection and parameter estimation based on approximate Bayesian computation ( ABC ) . ABC is a methodology to estimate model parameters replacing the likelihood function with a simulation-based procedure and a distance function to measure the similarity between simulated and observed data . Various ABC algorithms have been proposed , from the simple ABC-rejection [43] to ABC Markov chain Monte Carlo ( MCMC ) [44] and ABC based on sequential Monte Carlo ( SMC ) methods [23 , 45] . Here , we use ABC-SMC as proposed by Toni et al . [23] because it addresses some of the potential drawbacks of previous ABC algorithms , such as slow convergence rate , by sampling from a sequence of intermediate distributions . The SMC sampler introduces a number of intermediate steps decreasing iteratively the tolerance threshold ϵ for samples acceptance . At the first iteration , N particles θ′ ( representing the parameters of interest ) are generated from the prior distribution and data are simulated from the model based on θ′ . The proposed parameters are accepted if the difference between the summary statistics of the simulated data D′ and the observed data D is below the threshold ϵ1 . At iteration t > 1 , the particles are drawn from the previous population of the accepted samples at the iteration t − 1 ( with threshold ϵt−1 ) with slight perturbations . In our work , data ( observed virus genealogy ) and simulated trees are compared through the use of summary statistics which correspond to the above listed tree statistics ( Table 1 ) . The three network models M = {WS , ER , BA} were used to simulate outbreaks using the stage-varying infectivity profile with ratio 10:1 acute:chronic and patient infectivity variation ( σ = 3 ) . We assumed that network model and one network parameter were unknown . For ER , the network parameter of interest was the probability of drawing an edge between two arbitrary vertices; for BA it was the number of edges to add in each time step of the generating algorithm , and for WS it was the neighborhood within which the vertices of the lattice are connected . We also estimated the removal rate γ and the infection rate in the acute phase λ1 ( infection rates in the chronic and immuno-compromised stage can be obtained deterministically from the acute phase infection rate ) . Therefore , θ consists of 3 parameters for each type of network and they are model specific . All remaining parameters characterizing both the network structure and the epidemic process were considered known . The output of the algorithm were the approximations of the model M marginal posterior distributions P ( M|D ) which is the proportion of times that each model is selected in N samples , and the marginal posterior distributions of parameters P ( θ|D , M ) for the candidate models . We used a discrete uniform distribution from 1 to 3 as model prior π ( M ) . We chose to decrease the tolerance values following an exponential decay such that ϵt = ϵ0 exp ( −0 . 5t ) where t is the current sequential step , as proposed in [46] . A pilot run of 100 simulations for each model in M was used to define the initial thresholds . Convergence was assumed when the acceptance rate of newly proposed particles had dropped below 1 in 100 , and visual inspection of the posterior distribution showed no change in the last iterations . We found that convergence was achieved with T = 10 iterations and N = 1000 particles per iteration . The prior distributions on the parameters λ1 and γ were Uniform ( 0 . 0001 , 0 . 1 ) and ( 0 . 00025 , 0 . 1 ) , respectively . The computation time of the algorithm depended on the tree size and sampling fraction and it took between 1 and 2 hours in a parallel implementation on 8 processors . Further details of the algorithm and its computational cost can be found in S2 Text . We applied the ABC inference method to the analysis of two HIV-1 DNA sequence sets sampled from different IDU transmission epidemics in Sweden [47 , 48] . To reconstruct the time-scaled virus phylogenies from the DNA sequences we used a Bayesian Skyline coalescent model in BEAST 1 . 8 [49] . The general time reversible nucleotide substitution model was used with an uncorrelated log-normal relaxed clock and a discretized gamma distribution with four categories was used to model rate heterogeneity across the sequence . For the log-normal relaxed clock parameters , a uniform prior on the positive axis was assumed for the mean , and an exponential with mean 1/3 for the standard deviation . A Uniform prior on ( 0 , 1 ) was used for the nucleotide frequencies . The MCMC was run for 10 million iterations , with a 10% burn-in period and samples saved every 10000 iterations . We selected the maximum credibility tree and the negative branches were set equal to zero .
The within-host model generates virus genealogies that are consistent with a given transmission history , but not necessarily identical to it . An example of the impact of the within-host evolutionary process in a small size network/epidemic is shown in Fig 2 . This figure shows a transmission history ( A ) , its transmission tree representation ( B ) , and four compatible virus genealogies ( C-F ) . The genealogies display branch elongation/compression as compared to the transmission tree but also changes in topology . For instance , lineage 5 , sampled in individual 5 infected by 2 soon after 2’s own infection , appears consistently on the top part of the simulated virus genealogies and its branch can only be elongated by a small amount ( because the pretransmission interval is small ) . On the other hand , lineage 10 , infected later by 2 , is located in different parts of the possible genealogies , thus indicating changes in the virus genealogy topology versus the underlying transmission tree . This happens because longer time implies an increase in the virus diversity in 2 , i . e . more lineages are available in the donor . Therefore , as many virus trees are possible under any transmission history , it is important to evaluate the additional variation within-host diversity inflicts on the epidemiological inference . An epidemic can spread faster on ER and BA networks , thus the resulting transmission tree from a WS network includes longer times resulting in taller trees ( Fig 3A ) . This is mainly because WS has higher clustering than ER or BA . Both the unobservable transmission tree and the observable virus genealogy show the same tree height information . Other tree statistics , however , show different patterns of network discrimination based on transmission tree or virus genealogy . The proportion of cherries per taxa is slightly less informative on virus genealogies than on transmission trees ( Fig 3A ) . In particular , while there is a decrease for ER and WS ( less balanced in virus genealogies than transmission trees ) , it increases for BA ( more balanced in virus genealogies than transmission trees ) . A similar pattern is seen using Sackin’s Index or Colless’ Index ( ER and WS less balanced in virus genealogy , BA more balanced ( Fig 3D and 3E ) ) . Overall , differences between BA and WS become more evident in virus genealogies . Because Sackin’s Index and Colless’ Index are highly correlated we will only report Sackin’s Index from now on . An epidemic spreading on an ER network is similar to a population random mixing model . Therefore , it is expected to generate a balanced transmission tree [13 , 15] . The average number of people infected by an individual in the ER network show little variance and therefore the within host evolution model will add some heterogeneity producing small changes in the tree topology leading to an increase in the unbalancedness in the resulting virus genealogy . In BA models , instead , the presence of superspreaders generate transmission histories that are very unbalanced [13] . When a donor infects two or more recipients within a short interval , the order of transmissions along with infection times become impossible to accurately reconstruct; all splits are within the donor , describing within-host evolution in the donor ( also shown in [21] ) . Overall , the pretransmission interval associated with each and every transmission is a random draw from the possible coalescence times in the donor’s viral population . Therefore , in BA networks the virus genealogy will show larger changes in the tree topology with respect to the underlying transmission trees and result in more balanced trees . The ratio of the mean internal to external branch lengths is informative about the type of network ( smallest for BA , higher for ER , highest for WS ) . While the trends were similar in transmission trees and virus genealogies , the expected ranges overlapped for ER and WS in transmission trees , and virus genealogies showed generally smaller ratios ( Fig 3C ) . Branch lengths increase linearly as a function of tree height during epidemic spread on both ER and BA networks . Deviations from linearity are observed for epidemic spread on WS ( S3 Fig ) . At the end of an epidemic , the mean branch length is constant among networks but longer in virus genealogies rather than in transmission trees ( Fig 3F ) . Overall , trees from ER and WS networks are more imbalanced based on virus genealogies . However , as an epidemic spreads much faster in a BA network , the resulting virus genealogy will instead become more balanced because the virus does not have time to evolve time-structure between transmission events . Infectivity is known to vary across HIV-1 pathogenesis [50 , 51] . Thus , rather than assuming a constant transmission rate throughout an infected person’s disease stages , we tested 7 different infectivity profiles varying the ratio between the acute and chronic transmission rates and measured how they affected network model discrimination . The transmission rate in the pre-AIDS stage was held constant . Tree height becomes much less informative of network type the bigger the difference is between acute and chronic stage infectivity ( Fig 4A ) . This is because higher acute stage infectivity causes more infections in the acute phase and consequently the epidemic spread is faster . Since infection happens so rapidly , the external branches become very long compared to the internal branches , as all internal nodes are pushed to the root the higher the ratio between acute and chronic stage infectivity . Therefore , the total tree height is dominated by the external branch lengths ( which are on average 1/γ ) . Similarly , mean internal over external branch lengths , which was an important index when constant infectivity was assumed , is less informative if we assume high acute/chronic stage infectivity ratios . Differences observed between ER and BA assuming a constant infectivity profile diminish ( Fig 4D ) . Hence , branch length and tree height measures are less informative of network type when differences in acute-chronic infectivity are considered . Topological tree measures , i . e . , cherries per taxa , and Sackin’s Index , were less affected by differences in acute-chronic infectivities ( Fig 4B and 4C ) . Both these measures , calculated on the possible virus genealogies , still informed about the underlying contact network structure that HIV spread upon . The next stage of introducing realistic host evolutionary dynamics is to model patient specific differences . We did that by introducing variability in the overall infectivity level while keeping the acute-chronic ratio at 10:1 ( σ = 0 , 3 , 10 ) . While it was clear that introducing a non-constant infectivity profile diminished genealogical differences between underlying contact network structures , it was less obvious what effect introducing between-patient variation had ( Fig 5 ) . While tree height remained with no power to discriminate between networks , internal to external branch length ratios became more discriminative ( BA had lowest ratio , ER intermediary , and WS high ) . Individual variability seems to affect tree symmetry near the root more than towards the tips , since the Sackin’s Index shows much more variation than the number of cherries per taxa . However , they both improved their power to discriminate between contact network structures , and Sackin’s Index could differentiate WS from BA and ER networks . Thus , these simulations showed that the complex interactions that affect the resulting tree statistics when multiple levels of variability interact ( within-host coalescence process , timing of infections relative to disease- and epidemic-stage , disease-stage infectivity differences , and patient individual variation ) , may induce non-trivial dynamic patterns . If no influx of susceptibles occurs , the mean branch length increases as trees grow taller because it takes longer time to find uninfected hosts later in an epidemic ( Fig 6A and 6B , see also S4A Fig ) . At 100% sampling of infecteds at any time during an epidemic , the mean branch length increases as a function of total number of sampled infecteds ( number of taxa , Fig 6B ) . BA typically produces shorter tree branches than ER and WS as more individuals are sampled . Thus , if it were possible to sample everyone at time of infection , then the trend of adding longer tips towards the end of the epidemic becomes more pronounced ( Fig 6A ) . The internal to external branch length ratios typically decrease as the epidemic progresses ( S4B Fig ) . This is mainly explained by the depletion of susceptible neighbors for individuals infected late in an epidemic , thus generating very long external branches . In addition , branches added later in the epidemic , resulting from chronic donors , divide already existing branches into shorter segments . BA trees show lower ratios than ER and WS throughout the epidemic , but WS and ER are less distinguishable during an epidemic . On the topological level , the Sackin’s Index typically decreases as an epidemic matures ( Fig 6C ) . At the end of an epidemic ( Fig 3 ) , BA and ER show more unbalanced trees throughout an epidemic and the most imbalanced trees come from WS networks ( Fig 6C and 6D ) . Simulations on networks of size 5000 show similar results: for comparison , see Fig 6 with S5 Fig and S4 Fig with S6 Fig . Thus , while these statistics are indicative of the underlying contact network , they are confounded by epidemic stage and the size of the susceptible population . Consequently , to be able to infer the underlying contact network from genealogies we must also know what stage an epidemic has reached and the number of susceptibles . While a genealogical tree grows as an epidemic matures , the sampling fraction has no real effect on mean branch length , albeit with smaller sample fractions the estimation becomes somewhat more uncertain due to stochastic effects ( S7 Fig ) . Interestingly , lower sampling fraction increases mean branch lengths derived from any underlying contact network ( S7 Fig ) . This happens because the remaining branches in the genealogy represent increased numbers of infected hosts . However , this effect does not cause additional confusion over that caused by epidemic stage , as the differences between BA , ER , and WS networks are distinct at all epidemic stages and number of infected . On the other hand , we do not usually know at what stage an epidemic is ( i . e . , number of actually infected ) but only the number of sampled hosts . The mean branch length as a function of number of taxa ( Fig 7 ) could mislead the inference of underlying contact network , especially for small sample fractions . In fact , any branch length or tree height index would be affected by mistaking number of sampled hosts with stage of the epidemic because the number of infected grows faster than the number of sampled early in an epidemic . The topological indices were also affected by sampling fraction . While general trends ( Fig 3 ) remain constant through the cumulative number of samples over an epidemic , it is again important to know at what stage an epidemic is at time of sampling . Similar to branch length indices , topological indices can be misleading if sampling faction and stage of the epidemic are unknown . We illustrate the performance of the ABC inference on 100 simulated viral genealogies for each network type of size 1000 . The parameters were chosen so that the mean degree of each network type was 8 , the diagnosis rate was 2 . 8 years−1 ( derived from the average time from seroconversion to diagnosis in Sweden estimated in [52] ) , and infection rate in the acute phase was λ1 = 0 . 005 . The sampling probability p was set at 0 . 5 because the data from the general HIV Swedish epidemic have coverage of around 50% [52–54] . To investigate model selection performance of the ABC algorithm , we record the number of times that the true model has the highest posterior model probability P ( M|D ) among the three models for the 100 simulated datasets . The algorithm was able to discriminate among the network models quite well . For the first network model ( ER ) , in 78 out of the 100 simulated datasets , the true model had the highest posterior probability among the 3 different network types . For the second model ( BA ) , similar results were obtained; 76 out of the 100 simulated datasets identified BA . Outbreaks on the WS network were misclassified only 1 time out of 100 . The corresponding network parameters were estimated reasonably well in most cases ( Table 2 ) . For illustration , we report the results of a randomly chosen experiment where the observed data were obtained from an epidemic spread on an ER network ( Table 2 ) . The parameters to estimate are the mean degree , diagnosis and infection rate for an outbreak on an ER network . The estimation of the removal ( i . e . diagnosis ) rate was sometimes skewed towards the upper bound , which probably is due to branch elongation induced by the within-host evolution model . We only estimate three parameters per network/epidemic model . In principle , it would be possible to add the rewiring probability , ρ , of the WS network in the ABC inference and estimate it . However , the rewiring probability of the WS networks turns out to be quite difficult to identify in the model choice setting . This is because for large values of ρ a WS network becomes indistinguishable from a ER network S1 Table . At ρ < 0 . 1 there was typically still a good chance ( P ( M = m|D ) = 0 . 70 − 0 . 95 ) to identify the correct network structure m . Inference of epidemic parameters as well as network type becomes more complex in real outbreaks . We consider two genealogies from separate IDU-associated HIV-1 CRF01 and subtype B epidemics in Sweden , respectively . The CRF01 tree was sampled from a rapid outbreak that was imported from Finland [48] around 2003 , which was quiescent until the outbreak started in 2006 . The subtype B tree was sampled from the more slowly , and typical , spreading IDU epidemic in Sweden [47] . While tree indices were different between the trees from the Swedish HIV-epidemic ( Fig 8 ) , and superficially in line with what one might expect comparing an outbreak scenario to a more endemic situation , e . g . , mean branch lengths were 279 and 913 , and tree height 4176 and 10527 , respectively , they cannot be directly compared because these trees represent different stages in the respective epidemic . Furthermore , real data is rarely sampled at 100% of all infected or even diagnosed , so comparisons to our simulated overall network differences are difficult to evaluate . Thus , to evaluate genealogies from real epidemics we must consider epidemic stage and sampling fraction ( Figs 6 and 7 ) . In the ABC analysis of the two IDU HIV-1 transmission chains among IDU we have assumed the same epidemiological model as in the simulations ( stage-varying infectivity profile with ratio 10:1 acute:chronic , patient infectivity variation ( σ = 3 ) ) and a sampling fraction of 50% . For the CRF01 IDU outbreak , the susceptible population was assumed to be 200 and in the Swedish subtype B ongoing epidemic it was set to 3000 . Results are shown in Table 3 . Overall , convergence was more difficult to achieve in the analysis of the real data and the tolerance levels ϵ had to be set to higher values than in the simulation studies . However , the posterior model probabilities seem to indicate that the two outbreaks display different associations to the three network models considered even though there is no single model ( among the three network models considered ) that can be used to appropriately describe each outbreak . For the CRF01 outbreak , there is weak evidence ( 45% ) in favor of the BA network type although the ER was also supported with a posterior probability of 39% , with small differences in the parameter estimates . The infection rate in the acute phase was λ1 = 0 . 018 ( 0 . 009 , 0 . 031 ) vs λ1 = 0 . 023 ( 0 . 0015 , 0 . 036 ) , γ = 0 . 001 ( 0 . 0006 , 0 . 002 ) vs γ = 0 . 001 ( 0 . 0005 , 0 . 002 ) and the mean degree was 3 . 2 ( 2 . 4 , 3 . 6 ) vs 3 . 7 ( 2 . 2 , 4 . 1 ) in the BA and ER respectively . The HIV-1 subtype B outbreak was mostly ( 57% ) associated with an ER network type . However , there was considerable uncertainty in the parameter estimates: λ1 = 0 . 0025 ( 0 . 001 , 0 . 004 ) , γ = 0 . 0003 ( 0 . 0003 , 0 . 001 ) and the mean degree 1 . 4 ( 1 . 3 , 2 . 1 ) .
In this study we addressed several outstanding factors that could affect HIV phylogenetic tree shape in addition to the underlying contact network upon which HIV spreads . While previous studies have evaluated how contact networks affect the resulting tree [13 , 14 , 16] , they ignored differences between transmission trees and virus phylogenies , varying infectivity over disease progression , among patient infectivity variation , sampling fraction , and epidemic stage . Here , we show that all these factors put further restrictions on what type of phylogeny one can expect , but also that these additional factors may confound the inference of contact network . Transmission histories are not perfectly reconstructed by virus phylogenies . In fact , it has been previously shown that virus phylogenies have a time bias that elongates external branches , shifts internal nodes backwards , and may cause lineage disordering relative to the transmission history [19] . In this study we account for these factors by sampling ( many ) possible virus genealogies from a transmission history using a recently developed within-host coalescent model [21] . Because many virus genealogies may be consistent with one single transmission history , one would expect this factor to add uncertainty to the network inference . However , there is also added signal about transmission times because the within-host diversity changes over the time of infection . Thus , because the degree distributions are different for each network type ( Fig 1 ) , transmissions happen after different lengths of infection time , which affects the phylogenetic tree shape . Indeed , we show that the contact network inference from virus genealogies can be quite different than that from transmission trees , and that tree balance differences in fact may be more informative using virus phylogenies . Besides , transmission histories or trees can never be observed , or only partially and then with great uncertainty , which is the main reason for turning to phylogenetic reconstruction in the first place . It is well known that infectivity is not constant over disease progression , albeit the literature is uncertain about how big the difference is between acute and chronic stage infectivity [51 , 55] . Indeed , we find that varying infectivity affects the expected phylogeny under different contact networks . In fact , this factor alone seems to diminish phylogenetic differences between contact networks . Somewhat surprisingly , patient variation in infectivity works in the opposite direction , i . e . , it seemed to amplify differences in the contact network structures . The result is that virus genealogies do carry a signal of what type of contact network HIV spread upon , but the expectations are different than what one would expect from a naïve model where no virus diversity exists and all hosts are described by an identical constant infectivity over their pathogenesis . We do not investigate systematically if there is one factor that explains the expected genealogies simulated under the different network assumptions . Rather , the complex dynamic interaction of the heterogeneity factors included implies that a single cause for the differences observed may not exist . However , we compare only epidemic spread on three network types with the same mean degree and we used tree statistics to assess the differences . In certain settings , ER and BA behaved similarly ( and differently from WS ) . In this case , it is likely that the difference is due to clustering , since there is no large difference between the degree distributions in the ER and WS , whilst WS is the only network ( among the three considered ) allowing clustering . In some other scenarios , WS and ER behaved more similarly ( and differently from BA ) . This is most likely due to the different degree distributions . We show that any tree index that one would measure is affected by sampling fraction and the stage of the epidemic . We show that phylogenies cannot be meaningfully interpreted without this additional knowledge , as tree statistics otherwise may mislead the inference of contact network . While our results relate to epidemic situations relevant to HIV epidemics , they may also be relevant to other measurably evolving pathogens such as hepatitis C and influenza . The developed ABC inference framework for network identification and parameter estimation showed discriminatory power and ability to recover epidemiological parameters when applied to simulated data . The model used for validating the ABC algorithm included stage varying infectivity , individual , and within-host variability . For complex models such as epidemic spreads on networks the likelihood function is computationally costly to evaluate and ABC offered a way to perform likelihood-free statistical inference . Furthermore , the use of summary statistics allowed us to study the relationship between readily measurable tree statistics and complex transmission dynamics . The analysis of the two outbreaks from the Swedish HIV-1 epidemic showed that inference on real datasets is typically much harder . As is to be expected , real world networks do not match perfectly with the simplified models considered in this study , that were chosen for comparability with previous studies [13 , 14 , 16] . In fact , in the ABC algorithm , the proposed parameters values are accepted if the simulated data based on them are close enough to the observed data . If the observed data were generated from a rather different or more complex model , then the simulated data from the candidate model probably will be far away from the observed data . Hence , very few proposed parameter values will be accepted . More realistic models , such as dynamic networks , may be able to better capture the features of the outbreaks , especially those occurring over a long period of time . Another class of network models that could be suitable in modelling these outbreaks is the configuration model which is flexible and has been studied extensively in the literature , or exponential random graph models ( although in general slower to fit to data ) which include a broader spectrum of degree distributions and clustering levels rather than the three simple networks considered here . The ABC inference scheme can also be extended to take into account uncertainty in the phylogenetic reconstruction , as shown in [56] . Each summary statistics calculated on the simulated trees would then be compared to the distribution of the same statistics calculated on the posterior distribution of virus genealogies . Romero et al . [56] apply this procedure investigating transmission between a heterosexual couple . However , the outbreaks we are analyzing in this paper are bigger in size and the whole algorithm would be more computationally expensive . Lastly , this work could be further extended to integrate other sources of network data coming from social surveys and/or public health intervention studies , as recently outlined in a review paper by [57] to improve network analysis in HIV epidemiology .
|
Over the past few years , epidemiological models for infectious diseases have incorporated network structure: each individual has a set of contacts to whom they can pass the infection . However , collecting data to develop a good representation of the network is very challenging . The increasing availability of sequence data provide additional information: previous work has shown that it is possible to recover social network features from viral phylogenies . Until now , however , it was assumed that within-host evolution is negligible in the reconstruction . Here , we propose an approach based on approximate Bayesian computation to infer network structure from a virus phylogeny , explicitly including a model for within-host evolution . In addition , we incorporate other important heterogeneity factors such as individual-based transmission rates and infectivity varying by disease-stage . We show that in some situations the within-host virus diversity adds valuable signal to identify network structure , but in other situations it muddles the underlying contact structure .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"taxonomy",
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"infectious",
"disease",
"epidemiology",
"pathogens",
"spatial",
"epidemiology",
"microbiology",
"epidemiological",
"methods",
"and",
"statistics",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"phylogenetics",
"data",
"management",
"rna",
"viruses",
"phylogenetic",
"analysis",
"network",
"analysis",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"lentivirus",
"medical",
"microbiology",
"epidemiology",
"hiv",
"microbial",
"pathogens",
"hiv-1",
"evolutionary",
"systematics",
"molecular",
"biology",
"viral",
"evolution",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"epidemiological",
"statistics",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"organisms"
] |
2017
|
Inference of Transmission Network Structure from HIV Phylogenetic Trees
|
Insect seminal fluid proteins are powerful modulators of many aspects of female physiology and behaviour including longevity , egg production , sperm storage , and remating . The crucial role of these proteins in reproduction makes them promising targets for developing tools aimed at reducing the population sizes of vectors of disease . In the malaria mosquito Anopheles gambiae , seminal secretions produced by the male accessory glands ( MAGs ) are transferred to females in the form of a coagulated mass called the mating plug . The potential of seminal fluid proteins as tools for mosquito control demands that we improve our limited understanding of the composition and function of the plug . Here , we show that the plug is a key determinant of An . gambiae reproductive success . We uncover the composition of the plug and demonstrate it is formed through the cross-linking of seminal proteins mediated by a MAG-specific transglutaminase ( TGase ) , a mechanism remarkably similar to mammalian semen coagulation . Interfering with TGase expression in males inhibits plug formation and transfer , and prevents females from storing sperm with obvious consequences for fertility . Moreover , we show that the MAG-specific TGase is restricted to the anopheline lineage , where it functions to promote sperm storage rather than as a mechanical barrier to re-insemination . Taken together , these data represent a major advance in our understanding of the factors shaping Anopheles reproductive biology .
Anopheles gambiae mosquitoes are the principal vectors of human malaria , a disease with devastating consequences for public health and the economic development of disease-endemic countries . The creation of new tools to control vector populations is a focal point of intensive efforts to eradicate the burden of malaria . As mosquitoes generally copulate only once during their lives [1] , interfering with the mating process is a promising avenue for research into vector control . Unfortunately , very little is known about the molecular or physiological basis of mating and insemination in malaria vectors . Of particular concern is our lack of knowledge about factors and pathways ensuring male reproductive success , such as those that result in sperm storage , oviposition , and the inhibition of remating in females . Improving our understanding of mating biology might not only inform currently proposed strategies for vector control [2] , but could potentially allow the development of entirely novel tools for combating malaria . In sharp contrast with this scenario , a wealth of information is available on the mating biology of some non-vector insect species , particularly the fruit fly Drosophila melanogaster . Seminal fluid proteins ( generally named Acps ) derived from the male accessory glands ( MAGs ) and transferred to females during copulation have been demonstrated to be crucial regulators of sperm storage and viability and to be the triggers of oviposition and the reduced receptivity to remating experienced by D . melanogaster females after mating ( reviewed in [3] ) . Recently , large numbers of MAG-expressed proteins have been identified in numerous insects ( e . g . , honeybees [4] , butterflies [5] , crickets [6] , medflies [7] ) including two mosquito vectors: Aedes aegypti [8] and An . gambiae [9] . However , assigning specific functions to these seminal fluid proteins has proved to be difficult . Even in D . melanogaster , where genetic tools are well developed , only a handful of the >100 secreted Acps identified in the MAGs have been functionally characterized [3] . Our current knowledge of the importance of seminal fluid proteins to mating biology in An . gambiae is limited , and their role in reproduction is inferred mainly by the presence of similar functional classes amongst Anopheles and Drosophila Acps [9] . For instance , in contrast to our understanding in Drosophila [10] and Aedes aegypti [8] where many Acps have been identified in mated females , not a single MAG-expressed protein has been demonstrated to be transferred to females in An . gambiae ( but see [11] ) . MAG secretions in An . gambiae are deposited into the atrium of the female reproductive tract in the form of a gelatinous mating plug ( Figure 1 ) [12] , [13] . Mating plugs are a common feature in the reproduction of many organisms including invertebrates , reptiles , and mammals [14] , however , among mosquitoes , they are exclusive to anopheline species [15] . The An . gambiae plug is formed entirely within the male and is digested by the female over a period of 24 h . Nothing is known about its composition or how the liquid contents of the MAGs coagulate to form a solid mass during mating . Even the function of the plug is unclear . One prominent hypothesis is that the mating plug of An . gambiae serves as a physical barrier to re-insemination by blocking access to the spermatheca [15] , [16] . Indeed , Gillies [13] observed rare instances of females with two plugs and sperm trapped between them . Alternatively , the mating plug might act to prevent loss of sperm from the female storage organ , ensuring male reproductive success [17] . However , early researchers ( e . g . , [12] , [18] ) dismissed both of these possibilities and proposed that the mating plug serves no function and is simply a vestige of the ancestral dipteran spermatophore . Here , we show that the plug is a crucial determinant of An . gambiae reproductive success . By studying its composition , we were able to identify the mechanism of plug formation , which is based on the cross-linking activity of a MAG-specific transglutaminase ( TGase ) on seminal proteins . RNAi-mediated depletion of this TGase in males prevents plug formation and transfer , and severely impairs fertility . Females that do not receive a mating plug cannot retain sperm in their sperm storage organ , the spermatheca , and therefore do not become inseminated . Moreover , we show that the plug provides little defence against re-insemination .
To identify MAG proteins that are transferred to females during copulation , we examined the composition of mating plugs dissected from the reproductive tracts of recently mated An . gambiae females by mass spectrometric ( MS ) proteomic analyses [19]–[23] . To determine the source of the proteins found in the plug , we also analyzed the composition of the MAGs , and the atria of virgin females ( Figure 1 and Figure S1 ) , followed by reverse transcription PCR ( RT-PCR ) . These analyses identified 27 plug proteins: 15 derived from the male , six derived from the female , and six found in both male and female reproductive tissues ( Figure 1 ) . Five of the male-derived proteins ( Acps ) were previously shown to be exclusively expressed in the accessory glands , and included four proteins located within a “male fertilization island” on chromosome arm 3R [9] . The 10 remaining male-derived proteins in the plug were not previously known to play a role in reproduction and included five proteases . Even though proteases have been shown to be important components of the seminal fluid of other Diptera [3] , [8] , [10] , a previous study had failed to identify these enzymes in the MAGs of An . gambiae [9] . The six plug proteins derived from the female reproductive tract included two secreted atrium-specific serine proteases ( AGAP005194 and AGAP005195 ) whose transcripts were previously shown to be strongly downregulated 24 h after mating ( Figure 1 ) [24] . Multiple gel bands in particular from 50 to 140 kDa in the MS proteomic analysis of mating plug samples contained peptides derived from one particular protein , AGAP009368 , which we have named Plugin ( Figure S1 ) . Plugin was found by MS in both MAG and mating plug samples ( Figure 1 ) , and quantitative RT-PCR revealed that it is expressed exclusively in the MAGs ( Figure 2A ) . Western blot analysis confirmed this tissue specificity and showed the presence of high molecular weight bands in plug extracts ( Figure 2B ) . Within the MAGs , Plugin was detected by immunofluorescence primarily in the anterior region of a secretory epithelium and in the channels formed by an actin-rich muscle network lining the outside of the glands ( Figure 2C , 2D ) . Plugin lacks any recognizable protein domains , but is glutamine-rich ( 134/557 residues , Figure S1 ) . Many of these glutamine residues are excellent candidates for TGase-mediated cross-linking sites , as they often occur in tandem with a lysine at the +2 position , and are located in a region of the protein predicted to be intrinsically disordered [25] . This observation , combined with the MS identification of both Plugin tryptic peptides in the digests of high molecular weight gel bands and a MAG-derived TGase ( AGAP009099 ) in the mating plug , suggested that plug formation may be mediated by cross-linking of Plugin by this TGase . We tested for TGase activity in males using a monodansylcadaverine ( MDC ) incorporation assay [26] , which allows the incorporation of the fluorescent amine MDC into TGase substrates to be detected under UV illumination . High levels of TGase activity were detected in homogenized MAGs , but not in the mating plug ( Figure 3A ) , nor other male and female tissues ( unpublished data ) . MDC was incorporated into proteins that perfectly matched the observed sizes of Plugin , strongly indicating that this protein is the primary substrate for TGase in the MAGs . This incorporation was blocked by the addition of EDTA and GTP , which suggests that the TGase activity in the MAGs is calcium-dependent . These inhibitors also greatly reduced the formation of the higher molecular weight Plugin-immunoreactive bands in the MAG samples ( Figure 3A ) . The high levels of TGase activity observed in the MAGs prompted a closer investigation of An . gambiae TGases . Unusually for insects , which are believed to possess only a single TGase [27] , An . gambiae mosquitoes have three genes ( AGAP009098 , AGAP009099 , and AGAP009100 ) , clustered on chromosomal arm 3R . AGAP009099 was expressed exclusively in the MAGs ( Figure 3B ) , as confirmed by Western blot ( Figure 3C ) , while the other two genes were ubiquitously transcribed at much lower levels ( Figure 3B ) . These results suggested that AGAP009099 is principally responsible for the TGase activity detected in the MAGs . Within the MAGs , the AGAP009099 protein was localized in a similar pattern to Plugin , however it was primarily concentrated in the posterior part of the glands ( Figure 3D ) . The role of AGAP009099 in plug formation was then tested in vivo by RNA interference-mediated knockdown . Injections of male adults with double stranded RNA ( dsRNA ) targeting AGAP009099 ( ds9099 ) induced a significant reduction in both transcript ( mean = 67 . 0% , paired t test: t11 = −3 . 60 , p = 0 . 0042 ) and protein ( mean = 58 . 1% , t test assuming unequal variances: t47 . 2 = −4 . 25 , p = 0 . 0001 ) levels relative to males injected with control dsRNA ( dsLacZ ) . When injected males were allowed to mate with virgin females , 55 out of 367 females ( 15 . 0% ) mated to ds9099-injected males failed to receive a mating plug , compared to 4 out of 228 ( 1 . 8% ) females mated to control males . This large and statistically significant difference ( contingency test , χ2 = 27 . 55 , p<0 . 0001 ) demonstrates that AGAP009099 is crucial for the formation of the mating plug . We next assessed the function of the mating plug in Anopheles reproduction . In the large majority of cases where ds9099-injected males failed to transfer a mating plug , no sperm was found in the female spermatheca by microscopic analysis ( 41/55 = 74 . 5% ) . The absence of sperm was confirmed by our inability to amplify a Y-chromosome specific sequence by quantitative PCR in these spermathecal samples ( Table S1 ) . In these females , sperm were observed in the atria , indicating successful transfer , but were not appropriately stored and therefore would not be available for fertilization . In contrast , when a mating plug was found in the atrium , the spermatheca always contained sperm . Thus , the mating plug is important for sperm storage and for ensuring successful insemination . Only 2 . 5% of field-caught female An . gambiae store sperm from more than one male in their spermathecae [1] . One possible explanation for the low numbers of multiple inseminations is that , prior to the establishment of long-term mating refractoriness , if females mate again within a few hours of the first copulation , the presence of a plug might effectively block sperm from the second male from entering the spermatheca . We directly tested this hypothesis by mating females with wild-type males followed in rapid succession by transgenic males to be able to distinguish alleles from this second mating . Quantitative PCR of relative quantities of the two sperm types showed that 24 of the 38 twice-mated females tested ( 66% ) had sperm from both males in their spermathecae ( mean % sperm from 2nd male = 38% , range = 7%–56% ) , demonstrating that the mating plug is an inefficient physical barrier to re-insemination . Anophelines are the only mosquitoes that produce mating plugs [15] . If TGase activity underlies the ability to produce plugs , we would not expect to find a MAG-specific TGase in species that transfer uncoagulated seminal fluid , such as culicine mosquitoes . To test this hypothesis , we searched for TGase genes in the complete genomes of two culicines , Aedes aegypti and Culex quinquefasciatus , the only other mosquito species sequenced to date . We identified two culicine TGases retaining partial synteny with the three genes identified in An . gambiae and the single one present in Drosophila melanogaster ( Figure 4A , Table S2 ) . Phylogenetic analysis of TGases from these and other insects revealed that AGAP009100 , Aedes 1 , and Culex 1 cluster with the single TGase from Drosophila ( Figure 4B ) , suggesting that these genes may retain the ancestral function . AGAP009098 clusters with the second culicine TGase ( Aedes 2 and Culex 2 ) in a mosquito-specific group . No culicine TGase clusters with AGAP009099 , consistent with the lack of seminal coagulation in these mosquitoes . Importantly , neither of the Aedes or Culex TGase genes showed enriched expression in the MAGs ( Figure 4C ) , and we found no evidence of TGase activity in the glands of either species using the MDC incorporation assay ( Figure 4D ) . These two findings support the conclusion from the phylogenetic analysis that culicines lack an orthologue of the plug forming TGase and strengthen the correlation between the presence of AGAP009099 in An . gambiae and plug formation in this species .
We have identified the molecular composition , mechanism of formation , and function of the mating plug of An . gambiae . Our MS analysis identified 15 MAG-expressed proteins that are transferred to females as part of the mating plug . Two of these proteins , the MAG-specific TGase AGAP009099 and its glutamine-rich substrate Plugin , are responsible for the coagulation of the liquid secretions of the MAGs into a solid mass . Some of the other MAG-proteins from the plug , particularly the three small Acp-like proteins AGAP009362 , AGAP009370 , and AGAP012830 , could represent important modulators of female behavioural responses to copulation , such as a reduced receptivity to further mating and induced oviposition [9] . To our knowledge , these are the first proteins transferred to females during mating that have been identified in Anopheles . Further studies will clarify the role of these proteins in modulating female reproductive biology and possibly in sperm function . The identification of a number of female proteins , mainly proteases , associated with the mating plug suggests a direct interaction between male and female proteins that may be important for plug processing . Indeed two of the female proteases identified on the plug ( AGAP005194 and AGAP005195 ) were shown previously to be expressed exclusively in the atrium of virgin females and were considerably downregulated at 24 h after mating [24] . This transcriptional modulation is entirely compatible with a role of these enzymes in plug digestion , which is mostly completed in the female atrium by 24 h post-mating . Mating plugs are found in a wide assortment of vertebrate and invertebrate species , and many hypotheses have been advanced to explain their function . However , in the vast majority of taxa , empirical evidence for a specific role of the plug in mating remains elusive [14] . Perhaps the most common presumption is that plugs act as barriers to re-insemination . The high levels of monogamy observed in wild mosquito populations are thought to be enforced , at least over the short term , by the presence of a mating plug [15] . We have shown that the plug provides little defence against the storage of sperm from subsequent males . This is consistent with the observed mating behaviour of An . gambiae . In this generally monoandrous species , virgin females enter a swarm of males , mate , and leave the swarm while still in copula [28] , [29] . It is unlikely that a female would re-enter the swarm ( and indeed double plugs are almost never observed in the field [12] , [30] ) , and therefore there would be very little selective pressure for a plug that acts as a physical block . Instead , we have demonstrated that the plug plays an important role in the reproductive biology of An . gambiae . By manipulating the expression of the MAG-specific TGase , we have prevented plug formation and transfer , resulting in the incomplete storage of sperm by the female . The presence of a plug in the early post-copulatory hours may be needed to facilitate sperm retention in the sperm storage organ until motility is acquired . Indeed the sperm of An . gambiae are deposited in a immotile state directly into the spermatheca , immediately followed by plug transfer [12] , [31] , and become motile only >17 h after copulation [32] . In support of this hypothesis , we observed sperm in the atria of females that did not receive a plug , strongly suggesting that they had been transferred but had leaked out of the spermatheca . Interestingly , a role for the mating plug in sperm storage was dismissed by some earlier researchers [12] . Using the forced copulation technique for mating , females could be inseminated even when they did not receive a plug [33] . However , the forced copulation required the female to be anaesthetized during mating . One possible explanation for these earlier results is that female activity is required for sperm loss . Sperm backflow from the spermatheca into the atrium occurring in plug-less matings may be a passive consequence of female movement , or—more intriguingly—females may use the mating plug as a male “quality check , ” actively ejecting sperm from males that fail to transfer it . As the spermathecae of a small number of females that did not receive a plug contained observable sperm , future investigations of the relative numbers of sperm stored by these females could shed light on this issue . An alternative or additional explanation for the observed lack of sperm in the storage organs of females mated with plug-less males is that the formation of a complete mating plug in the male reproductive tract may be important for the correct delivery of sperm to the female . Although sperm are directly deposited in the spermatheca prior to plug transfer , it cannot be ruled out at this stage that coagulation of seminal fluids within the MAGs may play a role in the successful completion of the transfer process . The expansion of the TGase family in mosquitoes , and the acquisition of a function in seminal coagulation , underlies the ability of An . gambiae to form mating plugs . The presence of multiple TGases has not been previously reported in insects , but is common in vertebrates , and nine have been characterized in mammals [27] . These proteins fulfil numerous functions including seminal coagulation [34] , which is achieved by the cross-linking of glutamine-rich substrates such as semenogelins and seminal vesicle secretory proteins by the prostate-specific TGase TGM4 [34]–[36] . Thus , mosquitoes and mammals have independently evolved highly similar systems of semen coagulation . The convergent evolution of similar systems for plug formation in mosquitoes and in mammals is made all the more remarkable by the fact that other organisms have developed very different TGase-independent mechanisms to achieve similar results [37]–[39] . TGase-mediated cross-linking of plug proteins seems to be finely controlled as the plug-forming TGase and its major substrate ( i . e . , Plugin ) are expressed in two different compartments of the MAGs ( Figure 2C , 2D , Figure 3D ) . These two proteins may be brought together in the aedegus during copulation by contraction of the muscle cells surrounding the MAGs , and a concomitant release of calcium from intracellular stores could cause the activation of the secreted TGase . Our findings reveal a crucial role of the mating plug in mosquito reproductive biology and identify this important structure as a potential target for the manipulation of mosquito fertility . This discovery was only possible because we first identified the molecular composition and mechanism of formation of the plug . Understanding the basic genetics underlying mating biology is an essential starting point for developing new tools that target mosquito reproduction and may influence the design of novel vector control strategies currently under development . The proteins identified in this study will not only provide a powerful basis for understanding other processes that regulate mosquito fertility , but will also allow comparative studies of reproduction in other organisms . Indeed , given the remarkable similarity between mechanisms of seminal coagulation in mosquitoes and mammals , our results can inform studies of mammalian , including human , reproduction .
Mosquitoes from a laboratory colony of the G3 strain were separated by sex as pupae and raised in cages supplied with sucrose ad libitum . Matings were performed as described previously [24] . Reproductive tissues were dissected from virgin males ( MAGs ) , virgin females ( atria ) , or recently mated females ( mating plugs ) ; washed in fresh PBS; and stored on ice in 20 µl of a 5% ( v/v ) acetic acid solution . The overall digest , chromatographic , and MS strategies used have been described previously [19]–[23] . Briefly , the supernatant was applied to SDS precast NuPAGE gels and following electrophoresis , Coomassie-stained . Bands were excised , destained , reduced , and alkylated with iodoacetic acid prior to proteolytic digestion with trypsin . After extraction the peptide mixtures were analysed by on-line nanoLC-MS and MS/MS using Q-TOF technology on Q-TOF and Q-Star instruments and by Mascot search of the MSDB/NCBI and An . gambiae database initially , then using An . gambiae predicted proteome ReAnoXcel [40] supplemented with the latest Ensembl and SNAP protein predictions . Identified peptides were individually blasted against the translated genome , and gene models corresponding to the identified genomic regions were developed using ab initio predictions informed by available ESTs , microarray data , and manual models submitted to Vectorbase . The genomic location of each gene model is provided in Table S3 . The source of proteins identified in the plug was confirmed by RT-PCR performed using cDNAs from MAGs , testes , the rest of the male body , and virgin non-bloodfed females . MAGs were dissected from 4-d-old virgin males , homogenized with a micropestle in either TGase “+” buffer ( 50 mM Tris pH 7 . 6 , 1 mM DTT , 5 mM CaCl2 ) or TGase “−” buffer ( TGase “+” buffer with 250 mM EDTA and 0 . 3 mM dGTP ) and frozen/thawed in dry ice three times before the addition of 5 mM MDC . Samples were incubated at 37°C for 60 min , vortexed briefly , and spun down 10 min at 13 , 000 rpm . Proteins in the supernatant were separated by SDS-PAGE and visualized under UV illumination using an LAS-3000 imaging system ( FujiFilm ) . Plugin localization was subsequently tested by Western blot . Total protein loaded on gels was visualized using SimplyBlue SafeStain ( Invitrogen ) . Affinity-purified polyclonal antibodies against Plugin and AGAP009099 were raised in rabbits against peptide epitopes ( Plugin: NEHRDPQNHQLPSSC; AGAP009099: CGSRYTDPMEKKYES ) by a commercial supplier ( GenScript Corp . , Piscataway , NJ ) . Tissues were homogenized in 20 µl PBS containing a protease inhibitor cocktail ( Complete Mini , Roche ) and frozen/thawed three times on dry ice . Samples were centrifuged at 13 , 000 rpm for 15 min at 4°C . The supernatant was heated at 70°C for 10 min and applied to precast NuPAGE ( Invitrogen ) gels under reducing conditions according to the manufacturer's instructions . Proteins were transferred to a nitrocellulose membrane ( under reducing conditions ) using the XCell II Blot module ( Invitrogen ) . Blots were immunostained using standard protocols using the following primary antibody titres: anti-Plugin: 0 . 59 µg/ml; anti-9099: 0 . 96 µg/ml; and anti-β-actin ( 1∶1000 dilution of ab8229; Abcam , Cambridge , MA ) . HRP-conjugated secondary antibodies ( Santa Cruz Biotechnologies: sc-2030 and sc-2314 ) were used at a dilution of 1∶10 , 000 . Bands were visualized using ECL Western blotting detection reagents ( GE Heatlhcare ) on an LAS-3000 imaging system ( FujiFilm ) . Individual MAGs from males injected with ds9099 or dsLacZ , were placed in a 110 µl PBS containing a protease inhibitor cocktail ( Complete Mini , Roche ) , homogenized in an ultrasonic bath for 10 min , frozen/thawed on dry ice three times , and centrifuged at 13 , 000 rpm for 15 min at 4°C . Duplicate 50 µl aliquots of the supernatant were loaded into separate wells of a flat-bottomed 96-well plate and incubated overnight at 4°C . A standard curve was prepared from the MAGs of uninjected males with a series of six 2-fold dilutions . ELISAs were carried out essentially as described previously [41] . Anti-9099 was used at a concentration of 0 . 96 µg/ml and the secondary antibody , sc-2314 , at a dilution of 1∶2 , 000 . MAGs from 3–4-d-old males were dissected on ice , fixed in 4% formaldehyde , washed in PBS , bleached with 2% hydrogen peroxide to minimize autofluorescence , washed in PBS , then blocked and permeabilized in PBS with 1% BSA and 0 . 03% Triton X-100 . Samples were incubated in either 2 µg/ml anti-Plugin or 3 µg/ml anti-9099 in blocking buffer , then a 1∶1 , 000 dilution of anti-rabbit Cy3 followed by a 1∶250 dilution of Alexa Fluor 488 phalloidin ( Invitrogen ) to stain actin . Tissues were then mounted in DAPI-containing Vectashield medium ( Vector Laboratories , Inc . ) and visualized using a Leica SP5 inverted confocal microscope . Stacks were generated using 19 consecutive 0 . 5 µm optical sections . A 481 bp region of AGAP009099 was amplified from MAG cDNA using the primers ( FWD: 5′-GAGCGGTCGTGGTCGATAGTAAG-3′ and REV: 5′-CCCTCGTAGTTGTTGCTCCAGTT-3′ ) and cloned into pLL10 [42] . This purified linearized plasmid was used to make ds9099 , and pLL100 for the synthesis of dsLacZ ( dsRNA targeting the bacterial LacZ transcript , which is not present in mosquitoes ) , following established protocols [42] , [43] . Males were sexed as pupae and injected with 69 nl of dsRNA ( 3 µg/µl ) within 24 h of eclosion . Surviving males were allowed to mate with 5–6-d-old virgin females 4–5 d after injection . Mated females were immediately dissected to visually ascertain the presence of a mating plug in the atrium and/or sperm in the spermatheca . Males were dissected and their MAGs used for qRT-PCR analysis of RNAi-induced knockdown , or for ELISA . RNA extraction , cDNA synthesis , and SYBR-green based qRT-PCR was performed as described previously [24] using the primers listed in Table S1 . The ribosomal protein gene RpL19 was used for normalization in An . gambiae ( AGAP004422 ) , using previously described primers [24] . Wild-type 4-d-old virgin females were placed in a cage containing approximately 250 wild-type males . Copulating pairs were captured as described previously [24] , the males removed , and the females introduced into a cage containing approximately 250 males homozygous for the transgene dsRed ( FC , unpublished data ) . Females mating for a second time were recaptured and placed in a cage without males for 48 h . After this period , females were dissected in PBS , and individual spermathecae were placed in 23 µl of grinding buffer ( 80 mM NaCl , 8 . 5 mM EDTA , 24 mM Tris [pH 7 . 5] , 0 . 5% SDS , and 5 . 5% sucrose ) . Samples were placed in an ultrasonic water bath for 10 min or until each spermatheca was ruptured . Three µl of 0 . 01 M Proteinase K was added to each tube , and samples were heated at 37°C for 15 min , then 95°C for 10 min . Samples were analyzed by SYBR green-based qPCR using 5 µl of undiluted spermathecal DNA . Y-specific primers ( Table S1 ) were designed within the Y-specific region of the chimeric An . gambiae scaffold AAAB01008227 amplified by Kryzywinski et al . [44] using the primer pair 128125I . These primers were tested on multiple male and female genomic DNA extractions and only produced a product in males ( two copies per Y-chromosome , unpublished data ) . In both cases , matings were completed in the space of 60 min . Females were rested for 45 min between the two matings . The selected amino acid sequences were subjected to multiple alignments using the Clustal W ( http://www . ebi . ac . uk/Tools/clustalw/ ) and Clustal X ( 1 . 83 ) algorithms . A phylogenetic tree was constructed by the neighbour joining method using p-distance estimates , tested by the interior-branch test , and displayed using TreeView ( 1 . 6 . 6 ) software . Reliability of each node was assessed with 1 , 000 bootstrap replications . The genomic locations of the TGase genes encoding the proteins used in the tree , as well as Plugin , are reported in Table S2 .
|
Male seminal fluid proteins trigger a wide range of behavioural and physiological changes in females and can have important effects on reproductive success . In many animals , seminal fluid is transferred to females as a gelatinous mass termed a mating plug . Although many hypotheses have been put forward to explain the function of mating plugs , their precise role in most organisms remains unclear . We have studied the composition , mechanism of formation , and function of the mating plug in the mosquito Anopheles gambiae , the principal vector of human malaria . We show that the plug is formed through the action of a transglutaminase enzyme that links seminal fluid proteins together resulting in semen coagulation . This process is similar to the way seminal fluid is coagulated in mammals . Interfering with the production of this transglutaminase prevented plug formation . Females that did not receive a plug failed to store sperm correctly , with important consequences for fertility . Our data show that the mating plug is an important feature of An . gambiae reproduction , and reinforce the notion that a deeper understanding of mosquito reproductive biology can aid efforts to eradicate these disease vectors .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/animal",
"genetics",
"evolutionary",
"biology/sexual",
"behavior",
"molecular",
"biology/molecular",
"evolution",
"genetics",
"and",
"genomics/gene",
"function",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] |
2009
|
Transglutaminase-Mediated Semen Coagulation Controls Sperm Storage in the Malaria Mosquito
|
TRIM5 proteins can restrict retroviral infection soon after delivery of the viral core into the cytoplasm . However , the molecular mechanisms by which TRIM5α inhibits infection have been elusive , in part due to the difficulty of developing and executing biochemical assays that examine this stage of the retroviral life cycle . Prevailing models suggest that TRIM5α causes premature disassembly of retroviral capsids and/or degradation of capsids by proteasomes , but whether one of these events leads to the other is unclear . Furthermore , how TRIM5α affects the essential components of the viral core , other than capsid , is unknown . To address these questions , we devised a biochemical assay in which the fate of multiple components of retroviral cores during infection can be determined . We utilized cells that can be efficiently infected by VSV-G-pseudotyped retroviruses , and fractionated the cytosolic proteins on linear gradients following synchronized infection . The fates of capsid and integrase proteins , as well as viral genomic RNA and reverse transcription products were then monitored . We found that components of MLV and HIV-1 cores formed a large complex under non-restrictive conditions . In contrast , when MLV infection was restricted by human TRIM5α , the integrase protein and reverse transcription products were lost from infected cells , while capsid and viral RNA were both solubilized . Similarly , when HIV-1 infection was restricted by rhesus TRIM5α or owl monkey TRIMCyp , the integrase protein and reverse transcription products were lost . However , viral RNA was also lost , and high levels of preexisting soluble CA prevented the determination of whether CA was solubilized . Notably , proteasome inhibition blocked all of the aforementioned biochemical consequences of TRIM5α-mediated restriction but had no effect on its antiviral potency . Together , our results show how TRIM5α affects various retroviral core components and indicate that proteasomes are required for TRIM5α-induced core disruption but not for TRIM5α-induced restriction .
Primates express a range of restriction factors that inhibit retroviral infection , and variation in restriction factors is an important determinant of retroviral tropism [1]–[3] . TRIM5α is one such factor [4] , and is a member of the large family of tripartite motif ( TRIM ) proteins that share a common N-terminus composed of a RING domain that functions as an E3 ubiquitin ligase , one or two B-box domains required for higher-order assembly and a coiled-coil dimerization domain ( RBCC ) [5]–[8] . TRIM5α also encodes a variable C-terminal B30 . 2/SPRY domain that recognizes incoming retroviruses [4] , [9]–[13] and the consequence of this recognition is that infection is inhibited soon after viral entry [14] , before reverse-transcription is completed . The viral capsid ( CA ) protein is the direct target of TRIM5α proteins [15]–[17] , and is recognized by TRIM5α multimers only in the context of assembled viral cores , but not as monomers [15] , [16] , [18] , [19] . The RING domain of TRIM5α exhibits E3 ubiquitin ligase activity , and its removal , or mutation of key cysteine residues that are required for this activity reduces the potency of TRIM5α-mediated restriction [4] , [10] , [20] , [21] . TRIM5α proteins with distinct spectra of antiretroviral activity are present in most , perhaps all , primate species . For example , the prototypic rhesus macaque TRIM5α ( rhTRIM5α ) is a potent inhibitor of HIV-1 infection but does not efficiently restrict simian immunodeficiency viruses of rhesus macaques ( SIVmac ) [4] . Human TRIM5α ( huTRIM5α ) and African green monkey TRIM5α ( AGM TRIM5α ) also exhibit antiretroviral activity [22]–[24] and although AGM TRIM5α restricts a broad range of retroviruses , huTRIM5α is known to restrict only equine infectious anemia virus ( EIAV ) , and N-tropic MLV ( N-MLV ) [22]–[25] . Thus , the antiretroviral activity of TRIM5α appears to be quite plastic . Underscoring this point , in two different primate lineages ( macaques and owl monkeys ) , independent retrotransposition events have placed a cyclophilin A ( CypA ) cDNA into the TRIM5 locus , generating a fusion gene with utterly different antiretroviral specificity , wherein the B30 . 2/SPRY domain is replaced by CypA [26]–[29] . Although various domains of TRIM5α that are required for restriction have been well defined [7] , [8] , [30] , the precise mechanism by which TRIM5α acts on the incoming viral cores to disrupt infection has been enigmatic . The presence of a restricting TRIM5α protein causes a decrease in the yield of pelletable CA protein following infection and , in the case of huTRIM5α restriction of N-MLV , the loss of particulate CA protein is accompanied by an increase of soluble CA [31] , [32] . These experiments prompted a model whereby TRIM5α accelerates the uncoating of retroviral cores . Consistent with this model , a chimeric rhTRIM5α protein , containing the RING domain of TRIM21 , lead to the shortening of capsid-nucleocapsid tubes assembled in vitro [33] , [34] . A second aspect of TRIM5α-induced restriction is the role played by proteasomes . While inhibition of proteasomes does not rescue infection of restricting cells , it does rescue the formation of an integration-competent reverse-transcription complex , and appears to stabilize capsids in the cytoplasm of restricting cells [14] , [25] , [35]–[39] . One interpretation of these data is that TRIM5α causes a two-phase block to infection , in which passage of viral DNA to the nucleus is blocked , and then TRIM5 induces the viral core is disassembled by proteasomes . In other studies , however , inhibition of proteasomes was shown to cause a general increase in cytosolic particulate capsid independent of TRIM5 restriction [40] , [41] . It has also been proposed that TRIM5α accelerates degradation of CA , by a proteasome-independent pathway [36] . In addition to acting directly on the viral core , TRIM5α has been recently shown to promote innate immune signaling , an activity that is stimulated by and may contribute to restriction of retroviral infection [42] . Overall , it is unclear what the sequence of events is during restriction , and which events are necessary or superfluous for antiviral activity . As most studies of TRIM5-mediated restriction have focused on CA , the fate of other components of the viral core during restriction is unknown . Given that inhibition of proteasomes in restricting cells can rescue the formation of an integration competent reverse transcription complex [35] , [37] , one idea is that degradation of core-associated CA leads to the liberation of viral RNA and other core proteins e . g . enzymes . Thus , the physical separation of viral genomes and enzymes could lead to a block in reverse transcription . Alternatively , proteasomes may directly be involved in degradation of other core components . The lack of clarity in current pictures of how TRIM5α works is at least partly due to the difficulty of analyzing retroviral cores in infected cells using biochemical assays . This problem was partly overcome by the development of a “fate-of-capsid” assay , in which viral cores in cytosolic extracts prepared from infected cells are pelleted through a sucrose cushion [31] , [32] . This approach has been utilized in a number or studies of retroviral restriction by TRIM proteins [31] , [32] , [39]–[41] , [43]–[46] and capsid stability in infected cells [46]–[48] . Although this assay is very informative and essentially the only widely used assay for the biochemical analysis of post-entry events [31] , [32] , it does have limitations . First , only a fraction of the input material is actually analyzed - the endocytosed CA , which is thought to constitute the majority of the internalized material , is excluded . In addition , this approach has been applied only for the analysis of CA . Moreover , although restriction by TRIM5α likely occurs at early times after infection ( i . e . 1–2 hours ) [14] , most studies employing this assay analyze events that take place at later stages in infection . Finally , it has been debated whether the CA analyzed during TRIM5α restriction represents viral cores in the infectious pathway [32] , [36] , [38] , [40] , as a large fraction of internalized retroviral particles are thought to be nonproductively trapped and degraded in endosomes and lysosomes [49] . In order to overcome these problems , we developed a biochemical assay by which we can monitor the effects of TRIM5α on various components of retroviral cores at early times in infected cells . The approach we took was , essentially , to elaborate existing “fate of capsid” assays . Specifically , we utilized Chinese hamster ovary K1 ( CHO-K1 ) -derived pgsA745 cells ( pgsA ) which lack surface glycosaminoglycans and , perhaps as a consequence , can be very efficiently infected by VSV-G-pseudotyped viruses . Cytosolic proteins isolated from infected pgsA cells and its derivatives stably expressing various TRIM5α proteins were fractionated on linear sucrose gradients . This approach enabled the fates of CA , integrase ( IN ) , viral genomic RNA and reverse transcription products to be monitored . Using this assay we could show that the aforementioned viral components cosediment in a dense fraction . Moreover , we found that various components of retroviral cores have different fates during TRIM5α-mediated restriction , and can be degraded or disassembled . All of these effects on retroviral cores could be at least partially blocked by proteasome inhibition , but this manipulation did not rescue infectivity . These findings suggest that events that occur prior to core disassembly , rather than core disassembly itself or the action of proteasomes , is crucial for TRIM5α-mediated restriction .
To facilitate analyses of TRIM5α-mediated restriction , we developed a biochemical assay in which we monitored various components of retroviral cores in newly infected cells . We used a CHO-derived cell line ( pgsA ) , because it can be very efficiently infected by VSV-G pseudotyped retroviruses and does not express a TRIM5 protein that restricts MLV or HIV-1 infection [50] . Virions were bound to cells at 4°C , the inoculum was removed , and cells were either harvested immediately ( T = 0 hr ) or incubated at 37°C to allow infection to proceed for two hours ( T = 2 hr ) . Our previous observations indicate that events critical for TRIM5α restriction take place during this time [14] . Extracts from infected cells were separated on linear sucrose gradients and the presence of various core components in gradient fractions assessed . Initially , we focused on N-MLV infections and characterized the effects of huTRIM5α restriction on the CA protein . As indicated in Fig . 1A , N-MLV was efficiently restricted in the pgsA-huTRIM5α cell line , as compared to unmodified pgsA cells . When cells were harvested immediately after the virion-binding step ( T = 0 hr ) , CA was present throughout the gradient but enriched in fractions 5 to 8 . This distribution is likely a consequence of virions being bound to plasma membrane fragments of varying sizes ( Fig . 1B ) . As expected , the amount of virions bound to pgsA and pgsA-huTRIM5α cells was similar ( Fig . 1B ) . When cells were harvested after a 2 hour incubation at 37°C following virion binding ( T = 2 hr ) two distinct populations of CA molecules were present in unmodified pgsA cells . One concentration of CA molecules was present at the very top of the gradient ( fractions 1 and 2 ) , and presumably represented non-particulate material . A second concentration of CA molecules that were presumably part of a larger complex was evident in fractions 6 to 8 , towards the bottom of the gradient ( Fig . 1C ) . Strikingly , the dense peak of CA protein was absent when pgsA-huTRIM5α cells were used as targets ( Fig . 1C ) . Moreover , a clear increase of CA concentration in soluble fractions was observed ( Fig . 1C ) . These results are consistent with prior findings using the established “fate of capsid” assay [31] , [32] and imply that TRIM5α may lead to the disassembly of the capsid during the time at which TRIM5 proteins are known to exert their effects . Although prior data [31] , [32] , and the findings in Fig . 1 illuminate what happens to CA protein during TRIM5α-induced restriction , the fate of other core components under restrictive conditions was unknown . Therefore , we next asked how the behavior of other components of the N-MLV cores , namely IN , viral RNA and reverse transcription products , are affected by huTRIM5α . To determine the distribution of MLV IN , we inserted a 3×HA epitope tag at its C-terminus in the context of a Gag-Pol expression plasmid . MLV particles generated using this modified Gag-Pol expression plasmid were highly infectious . Notably , when cells were harvested and subjected to analysis immediately after virion binding ( T = 0 hr ) , IN protein was detected primarily in fractions 4 to 8 ( Fig . 2A , Fig . S1A ) , the same fractions in which CA protein was enriched after virion binding ( Fig . 1B ) . As expected , there was no major difference in the amount and migration pattern of IN when pgsA or pgsA-huTRIM5α cells were used . At two hours after infection ( T = 2 hr ) ( Fig . 2B ) , a dense complex containing IN was detected in unmodified pgsA cells , and virtually all of the IN protein co-migrated in the gradient with the large CA containing complex identified in Fig . 1 . Strikingly , the presence of huTRIM5α appeared to induce complete loss of IN at 2 h after infection ( Fig . 2B ) . Note that a protein band detected in fractions 1–3 migrates slightly more slowly than IN and is also detected in uninfected cells ( Fig . 2B ) as well as when the T = 0 hr blots subjected to a longer exposure ( Fig . S1A ) , indicating that it is a nonspecifically cross-reacting species . In contrast to the CA protein ( Fig . 1C ) , IN was not enriched in soluble fractions under restrictive conditions ( Fig . 2B ) , rather it appeared to be removed from cells . We next determined the fate of viral genomic RNA during TRIM5α mediated restriction by performing quantitative RT-PCR on the gradient fractions . As was the case with the IN protein , the viral RNA was found primarily in fraction 4 to 8 of the gradient after virion binding to pgsA cells ( Fig . 2C ) . After 2 h of infection in pgsA cells , viral RNA was found mostly in a large complex that comigrated with IN and the large CA containing complex ( Fig . 2D ) . The co-migration of CA , IN and viral RNA suggested that they were part of the same complex , perhaps representing intact , or nearly intact viral cores . Consistent with this notion , the migration of viral RNA through the gradient ( peaking at fraction 7 ) was very different to the migration of a cellular RNA encoding GAPDH , which was localized to fraction 3 ( Fig . 2E ) . Notably , the presence of huTRIM5α in target cells caused a loss of viral RNA from the large complex ( Fig . 2D ) . This huTRIM5α-induced loss of viral RNA from the large complex was accompanied by the appearance of a peak of viral RNA in fraction 3 where cellular GAPDH RNA is present ( Fig . 2D , E ) . In other words , huTRIM5α appeared to liberate viral RNA from a sub-viral complex , causing it to adopt the behavior of a generic cellular mRNA . Although the bulk of reverse transcription is likely not completed at T = 2 hr [14] , we could easily detect reverse-transcription products in infected cells at this time point . These viral DNA species co-migrated with other components of the viral core under non-restrictive conditions ( Fig . 2F ) . However , as expected , reverse transcription was blocked in cells expressing huTRIM5α and little viral DNA was detected anywhere on the gradient ( Fig . 2F ) . Given that all components analyzed that are predicted to be components of the viral core , co-fractionated with each other , it is likely that the sub-viral complexes detected herein represent functional complexes in which reverse transcription is taking place . The fact that TRIM5α clearly affected the fates of each of the viral components that were present in the dense fraction indicates that they were present in the cytoplasm , as they should not be affected by TRIM5α if they were in any other cellular location ( e . g . endosomes ) . Moreover , these results suggest that TRIM5α-mediated restriction involves both disassembly and degradation , with the differing ultimate fates of various core components . We next performed three control experiments to verify that the effects that we observed in Figs . 1 and 2 are truly relevant to restriction . Because it is thought that a significant fraction of internalized virions remain trapped in endosomes , we first asked whether the different gradient-migration patterns of core components under restrictive and nonrestrictive conditions was dependent on viral entry into the cytoplasm . To that end , cells were infected with VSV-G pseudotyped N-MLV for two hours in the presence of ammonium chloride ( NH4Cl ) , which prevents endosome acidification and blocks VSV-G mediated entry . After 2 h of infection in the presence of NH4Cl , CA was distributed throughout the gradient ( Fig . 3A ) , while IN ( Fig . 3B ) and viral RNA ( Fig . 3C ) were found primarily in fraction 4–8 . This pattern was similar to that observed when cells were harvested immediately after virion binding , and quite different to that observed when infection was allowed to proceed for 2 h in the absence of NH4Cl ( Fig . 1 , 2 ) . Importantly , the migration profile of the core components in the presence of NH4Cl was not affected by huTRIM5α . As an additional control experiment , we infected the non-restricting pgsA cells with either VSV-G-pseudotyped N- MLV ( Env ( + ) ) or N-MLV VLPs without VSV-G ( Env ( − ) ) for two hours . We could not detect CA ( Fig . S2A ) or IN ( Fig . S2B ) in the gradients prepared from cells incubated with Env ( − ) VLPs , This suggested that the Env ( − ) particles either did not efficiently bind to the target cells , or were degraded in endosomes without entering the cytoplasm . As expected , Env ( − ) particles were completely non-infectious ( Fig . S2C ) and , importantly , the initial virus inoculum contained equal amounts of Env ( + ) and Env ( − ) particles ( Fig . S2D ) . Thus , the changes in the behavior of core components induced by huTRIM5α was dependent on VSV-G-mediated binding and entry . Next , to determine whether the effects of huTRIM5α on N-MLV cores is a result of restriction activity , we performed similar experiments to those described above with B-MLV , which is insensitive to huTRIM5α restriction [22]–[25] . When viral cores were harvested after synchronization ( T = 0 hr ) , B-MLV CA , IN ( Fig . 4A , Fig . S1B ) and viral RNA ( Fig . 4B ) co-fractionated , primarily in fractions 5 to 8 , although CA was also detectable in other fractions ( as was observed with N-MLV ( Fig . 1–3 ) ) . All core components migrated in a similar pattern irrespective of the presence of huTRIM5α ( Fig . 4A , 4B ) . At two hours post-infection , CA , IN ( Fig . 4C , Fig . S1C ) and viral RNA ( Fig . 4D ) were all observed as components of large complexes regardless of the presence of huTRIM5α . Moreover , unlike N-MLV , huTRIM5α did not lead to any observable increase of B-MLV CA ( Fig . 4C ) or viral RNA ( Fig . 4D ) in soluble fractions ( 1 to 3 ) . As expected , the level of reverse transcription at this time point was not affected by huTRIM5α and viral cDNA co-fractionated with other core components ( Fig . 4E ) . Collectively these results validate our assay and support the notion that changes in the behavior of viral core components are induced by a restricting TRIM5α protein . The role of proteasomes during TRIM5α restriction has been a matter of debate , with several studies showing that inhibition of proteasomes does not restore infectivity that is restricted by TRIM5 proteins [14] , [31] , [32] , [35] , [37] . Moreover , in studies that employed the fate of capsid assay , huTRIM5α was reported to retain a significant ability to induce solubilization of MLV CA in the presence of proteasome inhibitor MG115 [41] . Some experiments have shown that proteasome inhibition causes a general increase in particulate HIV-1 and MLV capsids in both restricting and non-restricting cells [40] , [41] . However , it has been demonstrated that proteasome inhibition can restore reverse transcription , and the formation of a functional preintegration complex in the presence of TRIM5α [35] , [37] . As such , it is somewhat unclear whether proteasome inhibitor-restored reverse transcription complexes lack other core components or whether they are indistinguishable from unrestricted cores . Indeed , previous studies have indicated that viral DNA that is synthesized under TRIM5α-restricted , but proteasome inhibitor-restored conditions cannot enter the nucleus and become integrated [35] , [37] . Therefore , we asked whether inhibition of proteasomes in cells expressing huTRIM5α could restore the presence of large N-MLV sub-viral complexes containing CA , IN and viral RNA . To that end , we infected pgsA-huTRIM5α cells in the presence of MG132 , a proteasome inhibitor . As observed above in Fig . 1 and Fig . 2 , when huTRIM5α expressing cells were infected in the absence of MG132 , large complexes containing CA ( Fig . 5A ) and viral RNA ( Fig . 5B ) were lost and there was a concomitant increase in the levels of CA and viral RNA in soluble fractions . Strikingly , MG132 treatment restored large subviral complexes containing CA ( Fig . 5C ) and viral RNA ( Fig . 5D ) as well as the formation of reverse transcription products ( Fig . 5E ) . Importantly , in contrast to a previous study [41] , we did not observe a non-specific increase in dense N-MLV capsid in cells in the presence of MG132 . This may be either due to the fact that a different proteasome inhibitor was used in the study by Diaz-Griffero et al . [41] or that the indirect effects of proteasome inhibition our assays is minimized , because we analyzed an earlier time point in infection . Nonetheless , as previously reported , proteasome inhibition did not restore N-MLV infectivity in huTRIM5α cells ( Fig . 5F ) . These results suggest that although proteasomes play an important role in mediating the observed biochemical changes on viral cores induced by TRIM5α in our assays , they are not central to TRIM5α restriction . Recent findings have suggested the possibility that the uncoating ( loss of CA protein ) of HIV-1 viral cores early after infection is stimulated by reverse transcription [46] , [51] . In addition , reverse transcription was suggested to be required for rhTRIM5α-mediated disassembly of the HIV-1 core using the fate of capsid assay [46] . It was possible therefore that the initiation of reverse transcription might facilitate , or even be required for , the apparent disassembly and destruction of core components that we observed . Therefore , we repeated the above experiments in the presence of the reverse transcriptase inhibitor , AZT . Importantly the doses of AZT used were sufficient to block infection ( Fig . S3A ) , and the synthesis of reverse transcripts ( Fig . S3B ) under non-restricting conditions . Notably , treatment of pgsA cells with AZT during the 2 h infection assay did not affect the distribution of CA and IN in sucrose gradients: CA was present in both sets of fractions containing soluble proteins and large complexes while IN localized primarily to fractions containing large complexes ( Fig . 6A ) . The presence of huTRIM5α in target cells led to complete disappearance of both CA and IN from large complexes , with an accompanying increase of CA in soluble fractions under these conditions ( Fig . 6A ) . Similar to CA , even in the presence of AZT , huTRIM5α lead to the release of viral genomic RNA from the large complex ( Fig . 6B ) . Notably , the peak of viral RNA in the presence of huTRIM5α was lower than that in its absence ( Fig . 6B ) , suggesting that the viral RNA that is released from the core may be targeted for degradation ( discussed in detail below ) . Inhibition of proteasomes under restricting conditions , when reverse transcription was blocked substantially restored the presence of CA , IN ( Fig . 6A ) and viral RNA ( Fig . 6C ) in large complexes . These results confirm our previous findings and suggest that huTRIM5α action involves both disassembly and proteasome-mediated degradation of viral core components , and that these events occur independently of reverse transcription . We next sought to extend these observations and asked whether HIV-1 cores are similarly affected by TRIM5α restriction . To this end , we generated pgsA cells that stably express rhTRIM5α , which potently restricts HIV-1 infection ( Fig . S4A ) . When cells expressing hu- or rhTRIM5α were harvested immediately after synchronization , CA was detected primarily in the top two fractions and in fractions 5 to 7 ( Fig . 7A ) , whereas IN ( Fig . 7A ) and viral RNA ( Fig . 7B ) were more distinctly localized in fractions 5 to 7 . As expected , there was no difference in the behavior and amounts of HIV-1 core components harvested from huTRIM5α and rhTRIM5α cells at T = 0 h ( Fig . 7A , 7B ) . At T = 2 h post-infection , the CA protein in huTRIM5α cells was present as two distinct species with distinct migration properties in the gradient . A predominant species was present at the top of the gradient , likely corresponding to soluble proteins while a second species was present in denser sucrose fractions , likely representing viral reverse transcription complexes ( Fig . 7C ) . The overall profile of the behavior of HIV-1 CA molecules in the sucrose gradient was quite similar to that of MLV , but the relative abundance of the soluble CA protein was greater in the case of HIV-1 , suggesting the possibility that HIV-1 cores are either uncoated more rapidly following infection , or are inherently less stable in sucrose gradients . As was the case with MLV , the larger CA containing complex was lost in the presence of a restrictive TRIM5α protein ( in this case rhTRIM5α , Fig . 7C ) . However , a corresponding increase of CA in soluble fractions was not observed , perhaps because soluble CA was already quite abundant under non-restricting conditions ( Fig . 7B ) . Similarly , as was the case with MLV , rhTRIM5α restriction also led to the disappearance of HIV-1 IN from dense fractions , without any concurrent increase in soluble protein containing fractions ( Fig . 7C ) . Although rhTRIM5α restriction appeared to induce a decrease in the levels of viral RNA in dense fractions , this was not accompanied by an increase in the absolute levels of soluble RNA , although the relative amounts of soluble RNA vs . large-complex-associated RNA were increased in the presence of rhTRIM5α ( Fig . 7D ) . This was unlike our observations with N-MLV , and makes the analysis of HIV-1 RNA profiles difficult to interpret ( see discussion ) . Of note , under non-restricting conditions , the peak of viral RNA in dense fractions did not perfectly overlap with that of CA and IN ( Fig . 7D ) . This could possibly be a consequence of instability of HIV-1 cores in cells or on sucrose gradients , as was suggested by the relative abundance of soluble CA versus complex-associated CA ( Fig . 7C ) . In contrast , the products of reverse transcription co-fractionated nearly precisely with CA and IN under non-restricting condition and , as expected , were substantially reduced in rhTRIM5α cells ( Fig . 7E ) suggesting that the large complexes containing CA , IN and viral DNA are , or are derived from , functional HIV-1 reverse transcription complexes . To overcome any potential impact of reverse transcription on uncoating [46] , [51] , we repeated the above experiments in the presence of the reverse transcriptase inhibitor nevirapine . Importantly the doses of nevirapine used were sufficient to block infection ( Fig . S4B ) , and reverse transcription ( Fig . S4C ) under non-restricting conditions . As was found with MLV , inhibition of reverse transcription in restrictive or non-restrictive cells did not affect the behavior of viral CA and IN proteins , neither of which were present in large complexes in the presence of rhTRIM5α ( Fig . 8A ) . These results contrast with recent findings which suggest that inhibition of reverse transcription blocks the rhTRIM5α-mediated disassembly of the HIV-1 cores [46] . Notably however , nevirapine treatment of huTRIM5α cells substantially increased the level of viral RNA in dense fractions , and caused it to co-fractionate with CA and IN ( Fig . 8B ) . This finding suggests that the poor co-fractionation of CA , IN and viral RNA observed in Fig . 7D is a consequence of reverse transcription , or RNaseH activity , rather than misbehavior of HIV-1 cores on sucrose gradients . Notably , under restricting conditions , in the presence of nevirapine , viral RNA was lost from the large complexes , with no accompanying increase in soluble fractions ( Fig . 8B ) . This contrasts with our findings with MLV , where restriction led to an increase in the levels of soluble viral RNA . Finally , we determined whether proteasome inhibition restored the presence of HIV-1 cores under restricting conditions . When rhTRIM5α-expressing , HIV-1 infected cells were treated with MG132 and nevirapine , CA , IN ( Fig . 8A ) , viral RNA ( Fig . 8C ) and reverse transcription products ( Fig . 8D ) were all significantly restored in dense fractions . However , as was the case with restricted MLV infection , proteasome inhibition did not restore virus infectivity ( Fig . 8E ) . It is important to note that , unlike a previous study [40] , we did not observe a non-specific increase in particulate capsid in cells in the presence of MG132 . This may be either due to the fact that our assays are performed at much earlier time points post-infection , which may minimize indirect effects of proteasome inhibition , or that a different proteasome inhibitor was used in the study by Diaz-Griffero et al . [40] . Nonetheless , these results suggest that rhTRIM5α modifies the HIV-1 cores in a way that likely leads to the degradation of both IN and viral RNA . Although this process is sensitive to proteasome inhibition , but is not required for the antiviral activity of TRIM5α to be manifested . We then sought to confirm there observations by asking whether similar changes on HIV-1 cores can be induced by a different restrictive TRIM protein , namely owl monkey TRIMCyp ( omkTRIMCyp ) [27] , which was shown previously to reduce the amount of pelletable capsid in a fate of capsid assay [32] , [40] , [43] . This experimental system is better internally controlled , as restriction by omkTRIMCyp protein can be overcome by treatment of cells by cyclosporin A ( CsA ) , which prevents its binding to viral CA protein [27] . Since restricted and non-restricted HIV-1 core components were more reliably compared in the presence of reverse transcriptase inhibitor nevirapine ( Fig . 8 ) , we performed similar experiments in pgsA-omkTRIMCyp cells in its presence . As expected , although the treatment of pgsA-omkTRIMCyp cells with CsA restored infectivity ( Fig . 9A ) and reverse transcription ( Fig . 9B ) , the doses of nevirapine used in these experiments were sufficient to block both of these processes ( Fig . 9A , B ) . In omkTRIMCyp-expressing cells treated with nevirapine alone , viral CA ( Fig . 9C ) , IN ( Fig . 9D ) and viral RNA ( Fig . 9E ) were absent in dense fractions , without any notable increase in soluble fractions , similar to our observations with rhTRIM5α . As expected , large sub-viral complexes containing CA ( Fig . 9C ) , IN ( Fig . 9D ) and viral RNA ( Fig . 9E ) were restored in the presence of CsA . Importantly , when pgsA-omkTRIMCyp cells were treated with MG132 and nevirapine , CA ( Fig . 9C ) , IN ( Fig . 9D ) and viral genomic RNA ( Fig . 9E ) were all restored in dense fractions to almost the same level as observed under CsA treatment . However , as it was the case with restricted MLV and HIV-1 infections , proteasome inhibition did not restore virus infectivity ( Fig . 9A ) . These results together show that omkTRIMCyp disrupts HIV-1 cores in a similar way to that of rhTRIM5α and leads to degradation of at least some core components . Likewise , although this process is sensitive to proteasome inhibition , proteasomes are not required for the antiviral activity of omkTRIMCyp .
We formulated an experimental approach in which the fates of multiple viral core components can be tracked in infected cells , with the aim of understanding how TRIM5α restricts retroviral infection . The approach is similar in principle to the “fate of capsid” assay [31] , [32] , in which the putative separation of viral cores from infected cell lysates on sucrose gradients enables the analysis of their composition . However , our assay is more elaborate , and perhaps more effective , in several aspects . First , we monitored TRIM5α- and TRIMCyp-induced changes not only for CA , but also for IN , viral RNA and reverse transcription products in the same fractionation experiment . Second , in our assay , all of the input cellular material is analyzed , without the need for exclusion of putatively endocytosed virions . Although it is generally held that the majority of retroviral particles become trapped in endosomes of target cells , complicating analysis of early events in infection , this did not seem to be a major problem in our experiments . Indeed , the nearly complete disappearance of IN at T = 2 h specifically from restricting cells , argues that there is very little virus associated with the cells that had not reached the cytoplasm by this time point . Although the reasons for this are not clear , possibilities include highly efficient VSV-G-mediated entry in pgsA cells , particular instability of endocytosed virions in pgsA cells , or the fairly low MOIs used in these experiments . Third , infections are fully synchronized and the unbound input virus is removed before infection , which could limit the number of virions that are nonspecifically endocytosed . Fourth , analysis is carried out at an early time ( 2 h ) after infection when events relevant to TRIM5 restriction occur [14] . Fifth , we have incorporated quantitative aspects in our experimental system: Q-PCR analysis of viral RNA has proven to be an accurate and quantitative indicator of the fate of the viral core undergoing TRIM5α restriction . Overall our findings suggest that of all of the above components are present in a large complex comprising all or part of the virion core that is a functional intermediate in the infection pathway . Our findings provide insight into events that take place during TRIM5α restriction ( Fig . 10 ) . In parallel with previous findings [31] , [32] , we observed that N-MLV CA was redistributed from large complexes to soluble fractions in cells expressing huTRIM5α ( Fig . 1C ) . We expanded these observations and show that viral RNA was released from large complexes as a result of huTRIM5α restriction ( Fig . 2D ) . In contrast , MLV IN was not retained in a soluble form following its loss from dense fractions , and appeared to be degraded ( Fig . 2B , Fig . 10 ) . HIV-1 differed from MLV in that neither HIV-1 CA nor viral RNA was apparently increased in soluble fractions concurrent with their loss from large complexes ( Fig . 7B , 7D , 8C , 8E ) . However , the comparative pre-existing abundance of CA in soluble fractions may have masked any redistribution of CA protein to those fractions . Possible reasons for the discrepant fate of MLV and HIV-1 RNA under restricting conditions are discussed below . In the case of HIV-1 IN , the protein was lost from cells under restricting conditions in much the same way as was observed for MLV . Collectively these results indicate that TRIM5α causes both disassembly and degradation of viral components with similarities and differences in the fates of individual core components across retroviral genera ( Fig . 10 ) . Recent findings have suggested the possibility that the uncoating of retroviral cores early after infection is stimulated by reverse transcription [46] , [51] and that rhTRIM5α-mediated disassembly of HIV-1 cores requires reverse transcription activity [46] . Although in some experiments reverse transcriptase inhibitors modestly increased the amount of capsid detected by western blotting , we did not observe any effect of RT inhibitors on TRIM5-mediated disassembly/degradation of cores in this study . The reasons underlying the discrepancy between our results and the study by Yang et al . [46] are not clear . However , one would predict that reverse transcription is not required for restriction by TRIM5 , based on the fact that TRIM5 acts rapidly after entry [14] , before majority of reverse transcription is completed . The precise role of proteasomes in TRIM5-mediated restriction has been difficult to unambiguously determine . As previously demonstrated [35] , [37] , inhibition of proteasomes in restricting cells restored MLV and HIV-1 reverse transcription ( Fig . 5E , 8D ) . Importantly , we found that proteasome inhibition restored a core complex that is biochemically indistinguishable from unrestricted viral cores , and contained CA , IN and viral RNA ( Fig . 5 , 6 , 8 ) . As such , it is unlikely that TRIM5α mediates the complete disassembly of cores without the aid of proteasomes . Nevertheless , it is clear that proteasomes are not required for restriction by TRIM5α , as MG132 treatment of restricting cells does not restore virus infectivity ( [14] , [31] , [32] , [35] , [37] and Fig . 5F , 8E , 9A ) . Recent findings suggest that TRIM21/TRIM5α chimeras have the propensity to form hexameric lattices on the HIV-1 core , and it is possible that this phenomenon , in itself , constitutes the underlying mechanistic basis for restriction [52] . The assembly of such a lattice on the core may block the targeting of viral reverse-transcription or pre-integration complexes to the nucleus , because circular viral DNA forms are not generated during restricted HIV-1 infection under conditions of proteasome inhibition [35] , [37] . However , because HIV-1 and MLV apparently have different underlying mechanisms of entering the nucleus , it is possible that the other mechanisms that sequester viral DNA ( e . g . failure to uncoat ) may underlie the inability of HIV-1 or MLV to productively infect restrictive cells under conditions of proteasome inhibition . It is intriguing that some N-MLV and HIV-1 core components , notably viral RNA ( and perhaps CA ) , have somewhat different fates under restrictive conditions ( Fig . 10 ) . A possible explanation for this difference is that N-MLV core components are intrinsically more stable and as such , are degraded at a slower rate after TRIM5α-induced disassembly . Alternatively , rhTRIM5α and omkTRIMCyp may either specifically recruit a cofactor that more efficiently degrades the core components or simply disassemble HIV-1 cores at a faster rate . The loss of both N-MLV and HIV-1 IN in dense fractions without any apparent increase in soluble fractions may reflect the previously reported intrinsic instability of these proteins [53] . We did not detect obvious ubiquitinylation of any core proteins undergoing restriction in our assays . It is conceivable that ubiquitin-independent degradation or disassembly by proteasomes may be important for the observed effects on the cores [54]–[59] . Alternatively , if TRIM5 is responsible for ubiquitin modification of only a small fraction of core-associated proteins ( e . g . CA ) , we would not be able to detect this modification yet it could be responsible for core disassembly . The most striking difference between HIV-1 and MLV restriction is the fate of the viral RNA following its release from the core . It appears that MLV RNA is largely preserved , in a soluble form , whereas HIV-1 RNA is lost . We speculate that the mechanism by which HIV-1 viral RNA is lost during restriction is related to its nucleotide composition . It has long been known that the high AU content destabilizes the HIV-1 genome [60]–[66] . It is therefore conceivable that once the HIV-1 genome is exposed in the cytosol as a result of restriction , AU-rich elements may lead to the degradation of the genome , in the same way as has been observed with several RNAs coding for oncoproteins and growth factors [67] . Alternatively , proteasomes themselves , which have been suggested to comprise RNase activity , or other putative TRIM5α associated RNase activities may lead to selective degradation of AU-rich viral RNA molecules [68] . Nevertheless , it is unlikely that RNA degradation is critical for TRIM5 restriction as TRIMCypA chimeras containing the RBCC domain from other TRIM proteins , certain RING domain mutants of TRIM5α and squirrel monkey TRIM5α can restrict HIV-1 and SIVmac infection , respectively , after reverse transcription is completed [39] , [69] , [70] . TRIM5α mediated restriction serves as a useful model on which to base investigations of post-entry events . As such , the assay developed here could also be utilized to study restriction-independent events in newly infected cells . For example , it has been suggested that retroviral cores are optimally stable , and changes in CA stability in vitro can lead to defects in reverse transcription [71] . The assay developed here could identify the effects of such changes on multiple viral core components in infected cells . However , a caveat of our assay is that the precise nature of the ‘large complexes’ to which we refer is not known . For instance , it is not known whether the large complexes containing CA and the cofractionating core components actually represent intact conical viral cores . Previous investigations of cores isolated from extracellular virions and infected cells revealed notable differences in the density of N-MLV , B-MLV and HIV-1 ‘cores’ [31] , [71]–[74] . We did not observe such differences in our assays , as separation of cytosolic extracts in our experimental system is based on size , rather than density . Therefore , it is plausible that MLV and HIV-1 cores of different densities migrate almost identically on the sucrose gradients as they have similar sizes . Notably , even under non-restrictive conditions , a significant fraction of CA is present in soluble fractions . A similar phenomenon has been previously observed by others during isolation and biochemical characterization of HIV-1 , and to a lesser extent MLV , reverse transcription complexes in infected cells [75] , [76] This could be a consequence of the disassembly of some fraction of CA immediately upon infection or of the fact that only a proportion of the virion CA protein is actually assembled into cores in mature virions [77] . It is unlikely that the soluble CA represents complete disintegration of a fraction of viral cores in dense sucrose gradients [71] , [73] , as neither viral RNA nor IN is solubilized under non-restrictive conditions . Overall , we devised a novel experimental approach in which events that take place during TRIM5α restriction can be analyzed , and that can be applied generically to the study of early events in retrovirus replication cycle . Our results indicate that viral core components have distinct fates during TRIM5α restriction and are either disassembled or degraded . Importantly , in line with the two-step mechanism previously proposed [35] , [37] , [39] , [69] , although the TRIM5α-induced biochemical changes on the viral cores in our assays are sensitive to proteasome inhibition , proteasomal degradation is clearly not required for restriction . Future studies will address by which mechanism TRIM5α can restrict retrovirus infection as well as the mechanistic details of how different core components are affected by restriction .
CHO K1-derived pgsA-745 cells ( CRL-2242 , ATCC ) and all of its derivatives were maintained in Ham's F12 media ( Life technologies , 11765-054 ) supplemented with 10% fetal bovine serum and 1 mM L-glutamine . HEK 293T cells were obtained from ATCC ( CRL-11268 ) and maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum . VSV-G pseudotyped viruses were produced by transfection of 293T cells with plasmids expressing HIV-1 or MLV Gag-Pol , a packagable vector genome ( see below ) carrying GFP [15] , [78] and VSV-G at a ratio of 5∶5∶1 , respectively , using polyethyleneimine ( PolySciences , Warrington , Pennsylvania , United States ) as described previously [79] . Sequences encoding huTRIM5α , rhTRIM5α and omkTRIMCyp were inserted into LNCX retroviral vector plasmid ( Clontech ) , which were subsequently used to generate cloned pgsA-745 cell lines stably expressing huTRIM5α and rhTRIM5α . MLV and HIV-1 vector genome plasmids , CNCG and CCGW , respectively , encode a GFP reporter under the control of CMV promoter [15] , [78] . NL4-3 derived HIV-1 Gag-Pol sequence were inserted into the pCRV-1 plasmid [80] and carry a hemagglutinin ( HA ) tag at the C-terminus of integrase ( pNL-GP IN-HA ) . Sequences encoding B-MLV and N-MLV Gag-Pol inserts carrying a single copy or three copies of HA-tag at the C-terminus of integrase were inserted into pCAGGS plasmid [81] . Further details of plasmids are available upon request . PgsA745 cells , or derivatives thereof , ( 4×106 ) were plated on 10-cm cell culture dishes one-day before infection . For each treatment and time point , two such 10-cm dishes were used . In parallel , 2 . 5×104 PgsA745 cells were plated in 24-well plates to determine virus infectivity in each experiment . The corresponding MOI on 10-cm dishes was ∼0 . 025 for MLV infections and ∼0 . 01 for HIV-1 infections . Cell culture supernatants containing VSV-G pseudotyped viruses were filtered and treated with RNase free DNaseI ( Roche ) at a concentration of 1 unit/ml for 1 hour at 37°C in the presence of 6 mM MgCl2 . Cells were washed with ice-cold phosphate-buffered saline ( PBS ) and 6–7 ml of chilled virus ( adjusted to contain 20 mM HEPES ) was added to the cells . After allowing virus binding to cells at 4°C for 30 minutes , the inoculum was removed and cells were washed three times with PBS . Parallel infections were carried to determine the infectious titer in a given experiment . Cells were either harvested immediately ( T = 0 hr ) or incubated at 37°C for 2 hours ( T = 2 hr ) in complete cell culture media . In some experiments , cyclosporine A , proteasome and reverse transcriptase inhibitors were included during virion binding and during incubation at 37°C . Cells were collected in 1X PBS-EDTA , pelleted and resuspended in 1 ml of hypotonic buffer ( 10 mM Tris-Cl pH 8 . 0 , 10 mM KCl , 1 mM EDTA supplemented with complete protease inhibitors ( Roche ) and SuperaseIN ( Life technologies ) ) . After incubation on ice for 15 minutes , cell suspension was dounce homogenized by 50 strokes , using pestle B . The disruption of cells and the integrity of nuclei were monitored by Trypan blue staining of cells and nuclei ( Fig . S5 ) . Nuclear material was pelleted by centrifugation at 1000×g for 5 minutes and post-nuclear supernatant was layered on top of a 10–50% ( w/v ) linear sucrose gradient prepared in 1X STE buffer ( 100 mM NaCl , 10 mM Tris-Cl ( pH 8 . 0 ) , 1 mM EDTA ) . Samples were ultracentrifuged on a SW50 . 1 rotor at 30000 rpm for 1 hour . Ten 500 µl fractions from top of the gradient were collected , and proteins , RNA and DNA in each fraction was analyzed as described below . Proteins in each sucrose fraction were precipitated by trichloroacetic acid as described previously [82] . Protein pellets were resuspended in 50 µl of 1X protein sample buffer and analyzed by western blotting . The primary antibodies used were: mouse monoclonal anti-HA ( HA . 11 Covance ) , mouse monoclonal anti-HIV-1 p24CA ( 183-H12-5C NIH ) , mouse monoclonal anti-HIV-1 IN ( a gift from Michael Malim ) and goat polyclonal anti-MLV p30 ( a gift from Stephen Goff ) . For analysis of DNA and RNA , 50 µl of each fraction of the sucrose gradient was digested with proteinase K , phenol∶chloroform extracted and precipitated using sodium acetate/ethanol as described previously [82] . For analysis of RNA , samples were further treated with DNase I , extracted again and reverse-transcribed using the ImProm-II reverse transcription kit ( Promega ) . The resulting cDNA and DNA samples were used as template for quantitative real-time PCR ( qPCR ) using FastStart Universal SYBR Green Master Mix ( Roche ) and ABI 7500 Fast PCR system . PCR primers were designed within the GFP region of the vector genome . The primer pairs used in this study are as follows: GFP: Forward: 5′ AAGTTCATCTGCACCACCGGCAA Reverse: 5′ TGCACGCCGTAGGTCAGG; GAPDH: Forward: 5′ AGG TGA AGG TCG GAG TCA ACG , Reverse: 5′ GGT CAT TGA TGG CAA CAA TAT CCA CTT TAC .
|
The TRIM5 proteins found in primates are inhibitors of retroviral infection that act soon after delivery of the viral core into the cytoplasm . It has been difficult to elucidate how TRIM5 proteins work , because techniques that can be applied to this step of the viral life cycle are cumbersome . We developed an experimental approach in which we can monitor TRIM5-induced changes in the viral core at early times after infection , when TRIM5 exerts its effects . Specifically , we monitored the fate of the viral capsid protein , the integrase enzyme and the viral genome . We show that TRIM5 induces disassembly of each of these core components , and while some core components simply dissociate , others are degraded . These dissociation and degradation events all appear to be dependent on the activity of the proteasome . However , we also find that each of these TRIM5-induced effects events are not necessary for inhibition . The assay developed herein provides important insight into the mechanism of TRIM5α restriction and can , in principle , be applied to other important processes that occur at this point in the retrovirus life cycle .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology",
"biology",
"microbiology"
] |
2013
|
Fates of Retroviral Core Components during Unrestricted and TRIM5-Restricted Infection
|
Adaptation is a common feature of many sensory systems . But its occurrence to pain sensation has remained elusive . Here we address the problem at the receptor level and show that the capsaicin ion channel TRPV1 , which mediates nociception at the peripheral nerve terminals , possesses properties essential to the adaptation of sensory responses . Ca2+ influx following the channel opening caused a profound shift ( ∼14-fold ) of the agonist sensitivity , but did not alter the maximum attainable current . The shift was adequate to render the channel irresponsive to normally saturating concentrations , leaving the notion that the channel became no longer functional after desensitization . By simultaneous patch-clamp recording and total internal reflection fluorescence ( TIRF ) imaging , it was shown that the depletion of phosphatidylinositol 4 , 5-bisphosphate ( PIP2 ) induced by Ca2+ influx had a rapid time course synchronous to the desensitization of the current . The extent of the depletion was comparable to that by rapamycin-induced activation of a PIP2 5-phosphatase , which also caused a significant reduction of the agonist sensitivity without affecting the maximum response . These results support a prominent contribution of PIP2 depletion to the desensitization of TRPV1 and suggest the adaptation as a possible physiological function for the Ca2+ influx through the channel .
Adaptation is an essential feature of many sensory receptors , allowing them to continuously respond to varying stimuli . For example , photoreceptors can adjust their performance to an illumination level varying over orders of magnitude [1] . Hair cells can detect bundle deflections <1 nm in the presence of large static stimuli [2] . In contrast , whether adaptation also occurs to pain receptors has not been established , though neuronal plasticity is known to exist in pain sensation . The desensitization of pain receptors , on the other hand , has been extensively investigated . ( Here , “desensitization” refers to a loss of activity of the receptor after stimulation , whereas “adaptation” means that the receptor , after a complete desensitization , remains fully responsive to stimuli over a shifted intensity range . ) Capsaicin sensitivity is a hallmark of peripheral nociceptors [3] , and is mediated by TRPV1 in the C and Aδ-fibers [4] . Topical application of capsaicin to skin causes desensitization of these neurons , rendering them subsequently irresponsive to noxious stimuli [5] . The desensitization of capsaicin responses is historically divided into acute desensitization ( i . e . , a diminution of current during stimulation ) and tachyphylaxis ( i . e . , a reduction of current over repeated stimulation ) [6 , 7] . The tachyphylaxis appears to arise from a failure of recovery from desensitization [6] . Several mechanisms have been proposed for TRPV1 desensitization . These involve , for example , calcineurin [8–11] and calmodulin [12–14] . Ca2+ influx through TRPV1 also causes depletion of phosphatidylinositol 4 , 5-bisphosphate ( PIP2 ) during desensitization , and the recovery of the channel from desensitization requires the resynthesis of the lipid [15] . PIP2 was first reported as a tonic inhibitor for TRPV1 [16] . But exogenous PIP2 applied in excised patches was later shown to have a stimulatory effect as well , thus supporting its role in desensitization [17–20] . In addition , PIP2 has been implicated in the desensitization of TRPM8 [21 , 22] and TRPM4/TRPM5 [23 , 24] . A common problem in studying PIP2 regulation is that , while multiple ways exist for depleting membrane PIP2 , there are few tools for inhibiting the depletion , making it difficult to assess the causal relation between the depletion and the functional observation . There is only one pharmacological inhibitor for PLC β/γ , which has various side effects . In the case of TRPV1 , questions remain on whether endogenous PIP2 has indeed a stimulatory effect in living cells , whether its depletion by Ca2+ influx is adequate to desensitize the channel , and if so , to what extent it contributes . These questions are difficult to resolve by conventional pharmacological experiments . As an alternative , we have combined patch-clamp recording with total internal reflection fluorescence ( TIRF ) microscopy to simultaneously detect PIP2 depletion and current desensitization and to quantify the contribution of PIP2 depletion . Our data support a prominent role for PIP2 depletion in TRPV1 desensitization . In addition , it was revealed that Ca2+ influx caused more than an order of magnitude shift in the agonist sensitivity without compromising the channel maximum response , suggesting that the desensitization of TRPV1 may have a physiological function conferring adaptation of nociceptors .
Rat TRPV1 was provided by David Julius [4] . Pleckstrin homology domain tagged with red fluorescent protein ( mRFP ) mRFP-PH-PLC-δ was from Christopher Kearn ( University of Washington , Seattle , Washington , USA ) . Lyn-FRB and FKBP-Inp54p were from Tobias Meyer [25] . HEK293 cells were maintained in DMEM plus 10% fetal bovine serum ( Hyclone Laboratories , Inc . ) with 1% penicillin/streptomycin , incubated at 37 °C in 5% CO2 . Transfection was made at a confluence of approximately 80% using the standard calcium phosphate precipitation method [26] . Experiments took place usually 10–28 h after transfection . For primary culture of dorsal root ganglia ( DRG ) neurons , 6- to 8-wk-old adult male mice were used ( SJL/J strain , The Jackson Laboratory ) . Mice were deeply anaesthetized and decapitated , and the spinal cord was removed . Approximately 10–14 DRGs from thoracic and lumbar segments of spinal cords were rapidly dissected and cleaned in Ca2+/Mg2+-free HBSS . Ganglia were dissociated by enzymatic treatment with papain and collagenase/dispase and mechanical trituration through fire-polished glass pipettes until solution became cloudy [27] . The resulting suspension of single cells was plated on poly-D-lysine–coated coverslips , maintained in DMEM ( Gibco , Invitrogen ) containing 5% fetal bovine serum ( Hyclone ) , 50 U/ml penicillin/streptomycin , and 50 ng/ml NGF ( Sigma ) and incubated at 37 °C in a humidified incubator gassed with 5% CO2 . Patch-clamp recording was performed 12–20 h after plating . Conventional whole-cell patch-clamp recording was used . Patch pipettes were fabricated from borosilicate glass ( Sutter Instrument ) , coated with Sylgard ( Dow-Corning ) , and fire-polished to a resistance between 0 . 5–2 . 5 MΩ . Currents were amplified using an Axopatch 200B ( Axon Instruments ) amplifier , filtered at 1 kHz , and digitized at 5 kHz with multifunctional data acquisition cards ( National Instruments ) driven by custom-designed software using NIDAQmx library . Pipette series resistance and capacitance were compensated using the built-in circuitry of the amplifier , and the liquid junction potential between the pipette and bath solutions was zeroed priori to seal formation . Currents were normally evoked from a holding potential of −60 mV . All experiments were conducted at room temperature ( 22–25 °C ) . The control bath solution contained ( in mM ) : 140 NaCl , 5 KCl , 1 . 8 CaCl2 , 10 HEPES , 10–30 glucose , ( pH 7 . 4 ) ( adjusted with NaOH ) . The standard pipette solution consisted of ( in mM ) : 140 CsCl , 10 HEPES , 1 EGTA , ( pH 7 . 4 ) ( adjusted with CsOH ) . In a subset of experiments , the pipette solution was supplemented with 2 mM Mg2+ , and the bath solution was Ca2+-free with 5 mM EGTA . The removal of the bath Ca2+ did not affect the desensitization . The perfusion solutions were the same as the bath solutions except for appropriate agonists and/or Ca2+ . Exchange of external solutions was controlled by a gravity-driven local perfusion system with manually controlled solenoid valves ( ALA Scientific Instruments ) . The recording apparatus and perfusion lines were always thoroughly washed with ethanol after experiments . Capsaicin was purchased from Fluka ( Sigma ) . Capsazepine was from Precision Biochemicals . Rapamycin was from Calbiochem . All other chemicals were from Sigma . Capsaicin , capsazepine , and other water-insoluble reagents were dissolved in 100% ethanol or DMSO to make a stock solution , and were diluted into the recording solution at appropriate final concentrations before experiment ( <0 . 1% final ethanol and 0 . 3% DMSO ) . Fluorescence detection was integrated with patch-clamp recording on an inverted epifluorescence microscope ( Olympus IX 71 ) . Wide-field illumination was delivered by a 100 W Xenon short arc lamp ( USHIO Inc . ) using a commercial epi-illuminator attached to the real port of the microscope . Light was gated with an electronic shutter mounted on a filter wheel ( Lambda 10–3 , Sutter Instruments ) . TIRF excitation was provided through a high NA objective ( 60× , NA 1 . 45 , Olympus ) , and the illumination beam was introduced from the side of the microscope through a side-facing filter cube and brought to focus at the rear focal plane of the objective . mRFP was excited at 543 nm with a helium neon laser ( Melles Griot ) . Laser light was cleaned with a spatial filter made of an aspheric lens and a 10 μm pinhole and expanded using a pair of plano lenses ( final beam size ∼9 mm ) . The expanded beam was steered into the excitation port of the microscope with free space mirrors and a periscope mounted on the air table . The angle of the incident light was adjusted with one of the kinematic mirror mounted on the periscope . Whenever necessary , the total internal reflection was verified with a dilute aqueous suspension of 100 nm fluorescent beads ( Invitrogen-Molecular Probes ) , which were visible only when adhered to the coverslip under TIRF mode , and with a half-ball lens placed on the coverslip to redirect the reflected light into air . Fluorescent images were acquired using a cool CCD camera ( ORCA-ER , Hamamatsu ) controlled with a custom-written program using a public 1 , 394 digital camera driver ( Carnegie Mellon University ) . Real-time detection of the fluorescence was accomplished with a PMT mounted on the side port of the microscope . The signal from PMT was digitized together with the patch-clamp current .
TRPV1 desensitizes in the presence of external Ca2+ . Figure 1A shows a representative trace of a whole-cell current recorded from a HEK293 cell expressing the channel . In the absence of Ca2+ , 1 μM capsaicin evoked a large and sustained response . The presence of 1 . 8 mM Ca2+ caused the current to rapidly inactivate . The desensitization was nearly complete with prolonged stimulation . The rate of the desensitization varied from cell to cell , but typically ranged on the order of seconds . Once desensitized , the channel exhibited little response to subsequent application of the same stimulus , irrespective of the presence or absence of Ca2+ . When the current was partially desensitized , for example , due to a brief stimulation period , a small activity could be reactivated , but it was generally no larger than the residual current from the previous stimulation . This behavior is contrary to other ligand-gated channels , which often return to the resting state upon the removal of agonist . Because 1 μM capsaicin was saturating ( Figure 1D ) , the desensitized TRPV1 had been thought to be nonfunctional . However , when tested with a supramaximal concentration , we observed the reactivation of a significant current ( Figure 1B ) , though its amplitude was somewhat variable and generally less than the current before desensitization ( ∼59% for 10 μM capsaicin , Figure 1G ) . The response was not due to the recovery of the channel from desensitization since ATP was explicitly excluded from the pipette solution . The observation was also in agreement with a recent report on the possible reactivation of the channel by 30 μM capsaicin after desensitization [28] . The responsiveness of the desensitized channel suggests that the desensitization alters the apparent agonist sensitivity of the channel , which may arise from two possible mechanisms , either a change in the apparent agonist affinity or a compromise in the capability of conformational change for gating . To separate them , we measured the maximum attainable response of the channel after desensitization . Cautious to the possible side effects of a high concentration of capsaicin [29] , we combined it with low pH to lower the required capsaicin concentration . Figure 1C illustrates such a response before and after desensitization . Similar to 10 μM capsaicin alone , 1 μM capsaicin plus ( pH 6 ) produced a partial activity of ∼60% of the initial one . However , 10 μM capsaicin at ( pH 6 ) was able to evoke a nearly full response ( ∼97% , Figure 1G ) . The activation was robust and repeatable , suggesting that the dosage was in the saturating range . Retention of this full maximum response indicates that the desensitization had a specific impact on the ( apparent ) agonist sensitivity without altering the ability of conformational changes for gate opening . Several other lines of evidence further support that the channel retained normal gating after desensitization . First , the channel remained activated by depolarization in the absence of agonist , as illustrated in Figure 2A ( control cells transfected with empty vectors did not show a detectable response ) . The resulting current-voltage relations were superimposable ( Figure 2B ) , and the half-activation voltages were comparable ( V1/2 = 145 ± 2 mV before and V1/2 = 152 ± 2 mV after desensitization , n = 12 ) . This similar voltage-dependence of the channel before and after desensitization was in contrast to the large shift observed with capsaicin sensitivity and suggests that the desensitization did not impact the intrinsic gating machinery . Consistently , in the presence of agonist ( 1 μM before and 10 μM after ) , the channel also exhibited a similar current-voltage relationship before and after desensitization ( Figure 2C and 2D ) . At these saturating capsaicin concentrations , the I-V curve was less rectifying than in the absence of agonist . This occurred presumably because capsaicin binding makes a dominant contribution to the free energy of channel activation . Fluctuation analysis further revealed a static number of functional channels on cell surfaces . Figure 2E summarizes the variance-mean relationship of currents evoked by a voltage step from 0 to −60 mV in the presence of a low dose of capsaicin ( 0 . 3 μM before and 100 μM after desensitization ) . Parabolic fitting of the curves gave a relative change of N/N0 = 1 . 1 ± 0 . 1 for the number of channels and i/i0 = 0 . 98 ± 0 . 02 for unitary current ( n = 5 ) . Direct measurement of single-channel currents confirmed that the unitary current after desensitization was similar to that before desensitization ( Figure 2F ) . Collectively , these data support that the large shift of agonist sensitivity after desensitization arose from alteration in the agonist binding properties and was not due to changes in voltage-gating machinery . To quantify the change of the agonist sensitivity , we measured the full dose-response curves of capsaicin ( Figure 1E and 1F ) . Figure 1H summarizes the dose-response relations of the channel before and after desensitization and also the channel after recovery from desensitization . The data were normalized to the initial 1 μM capsaicin response . The recovery of the channel was prompted by the addition of 4 mM ATP into the pipette solution . No protein phosphorylation kinase activator was needed [30] . The recovered channel had a dose-response relation similar to the resting channel ( EC50 = 0 . 37 ± 0 . 04 versus 0 . 40 ± 0 . 02 μM ) , though its peak response was slightly larger as also previously observed [15] . The channel after desensitization , on the other hand , gave rise to a profound shift in the half maximal effective concentration ( EC50 = 5 . 29 ± 0 . 48 ) , which was ∼14-fold of the control . Contrary to the change in the ( apparent ) capsaicin sensitivity , the apparent cooperativity of the gating remained comparable under all three conditions ( nH = 2 . 3 ± 0 . 3 for control , 1 . 9 ± 0 . 4 for desensitized , and 2 . 0 ± 0 . 3 for recovered ) , suggesting that the gating of the channel stayed largely intact . The channel after desensitization appeared to also reach a full maximum response with capsaicin alone at ∼100 μM . The plasticity of the responsiveness of TRPV1 endowed by Ca2+ influx suggests that the desensitization may function as a feedback for the channel to adapt to different environments . As a further test , we examined the phenomenon with native TRPV1 in sensory neurons . Figure 3A illustrates recordings of whole-cell currents from cultured DRG neurons . The channel exhibited similar characteristics as observed in the heterologous expression system . Prolonged application of 1 μM capsaicin plus 1 . 8 mM Ca2+ resulted in nearly complete desensitization of the current . Repeat of the same stimulus after desensitization evoked virtually no activity ( Figure 3B ) . However , increasing capsaicin to 10 μM activated a partial response , and a full maximum response was obtained by further increasing capsaicin to 100 μM or combining 10 μM capsaicin with pH 6 ( Figure 3B ) . The result confirmed that the native TRPV1 was capable of adaptation in sensory neurons . In another experiment , we examined whether the adaptation phenomenon was specific to capsaicin . Figure 3C shows the response of the channel to low pH in transiently transfected HEK 293 cells . In the presence of 1 . 8 mM Ca2+ , application of pH 5 . 5 fully desensitized the channel . Following desensitization , the channel became irresponsive to the same stimulus ( pH 5 . 5 ) , but was activated by further acidifying the extracellular solution . Finally , combined application of capsaicin ( 10 μM ) and low pH pH 6 produced a maximum response similar to that before desensitization . These characteristics resemble those of capsaicin responses and suggest that the proton activation of the channel was also adaptable . As TRPV1 is involved in hyperalgesia , we tested whether channels that are sensitized under conditions mimicking the effect of inflammation ( favoring TRPV1 phosphorylation ) are also subject to similar adaptation regulation . The sensitization was induced by local perfusion of phorhol 12 , 13-dibutyrate ( PDBu ) to phosphorylate the channel ( Figure 3E ) . A low concentration of capsaicin ( 0 . 04 μM ) was applied before and after the application of PDBu , and the two responses were compared to ensure the occurrence of sensitization . The treatment with PDBu caused on average ∼7-fold increase in the current . Subsequent to the phosphorylation , the channel was desensitized by application of 1 μM capsaicin and 1 . 8 mM Ca2+ , and was then further tested with various stimuli for its responsiveness . Similar to channels at the resting conditions , the sensitized channels were also fully desensitized , and subsequently became irresponsive to 1 μM capsaicin ( Figure 3E and 3F ) . Furthermore , 10 μM capsaicin or 1 μM capsaicin plus pH 6 evoked a partial response after desensitization , which was about half of the initial maximum current prior to desensitization ( Figure 3F ) . These changes mirrored those of the nonsensitized channels , suggesting that the adaptation can occur under both physiological conditions and conditions that simulate pathological activation . The data also indicated that the Ca2+ influx provides a mechanism for the channel to recover from the sensitized state back to normal . We previously observed that the depletion of PIP2 occurred concomitantly with the desensitization of TRPV1 and that the subsequent replenishment of PIP2 was required for the recovery of the channel function from desensitization [15] . However , the precise contribution of the depletion of PIP2 has been difficult to determine , in part because the desensitized channel was thought no longer functional . The finding that the desensitization only alters the agonist sensitivity now provides a possible way to tackle the problem . In our first set of experiments , we examined whether the depletion of PIP2 has the right temporal relation to the desensitization of the current and to what extent PIP2 is depleted by the Ca2+ influx through TRPV1 . To monitor the concentration of PIP2 on the plasma membrane , we coexpressed the channel with the PIP2 binding construct , the PH domain of PLC-γ tagged with mRFP . The fluorescence on the plasma membrane was detected using TIRF microscopy combined with patch-clamp recording in real time . Figure 4A shows a representative trace of the recording , where the top is the current evoked by 1 μM capsaicin plus 1 . 8 mM Ca2+ , and the bottom corresponds to the change of the fluorescence from the footprint of the cell in contact with the coverslip . As evident from the recording , the activation of the channel caused an immediate decay of the fluorescence intensity . The decay was rapid and occurred most profoundly during the onset of the desensitization . It also reached a plateau after the current became fully desensitized . The two processes exhibited a close temporal correlation . Figure 4B shows both the TIRF image of the footprint and the whole-cell fluorescence in the far-field . While the TIRF intensity was reduced visibly after desensitization , the difference was less clear for the far-field images . Like the desensitization of the current , the decay of the fluorescence also showed considerably variable kinetics . However , the fluorescence covaried with the current , as evident from the correlation plot of their half-decay times ( Figure 4C , right ) . The overall averages of their half-decay times were comparable ( t1/2 = 5 . 4 ± 0 . 8 s for current and 3 . 5 ± 0 . 6 s for fluorescence ) . The depletion of PIP2 thus had adequate rapidity to potentially regulate the channel activity . At the steady state , the reduction of the fluorescence intensity reached ∼59% ( after subtracting the background response ) , indicating a considerable depletion of PIP2 by the Ca2+ influx . In some patches , we observed an onset of the fluorescence decay slightly lagging behind the inactivation of the current . Conceivably , it may arise because of the different depletion rates of PIP2 localized to the channel and in the bulk membrane . The contribution of mRFP bleaching to the fluorescence change was estimated negligible on the desensitization time scale under our illumination condition ( unpublished data ) . If the depletion of PIP2 contributes to the desensitization of TRPV1 , it should have a functional effect similar to the Ca2+ influx , which involves specific modulation of the agonist sensitivity while preserving the maximum response . To test the hypothesis , we exploited a rapamycin-inducible PIP2 depletion assay as recently reported [25 , 31] , which uses a constitutively active yeast lipid phosphatase that , when bound to rapamycin , can translocate rapidly from the cytoplasm to the plasma membrane to cleave the phosphate at the 5 position of PIP2 . Figure 5A shows a whole-cell recording from a HEK293 cell cotransfected with TRPV1 along with the phosphatase fusion protein Inp54p-FKBP and the membrane-anchored , FKBP-binding chimera Lyn-FRB . Rapamycin at 0 . 1 μM was continuously applied through local perfusion with a brief interruption for perfusion of 1 μM capsaicin ( in the absence of Ca2+ ) to monitor the channel activity . The treatment with rapamycin caused a progressive suppression of the channel activity . After ∼5 min , the current was diminished to ∼12% of the initial response ( Figure 5B ) . The channel was subsequently tested with supramaximal stimuli to assess its maximum attainable response . Application of 10 μM capsaicin evoked a current on average ∼56% of the initial 1 μM capsaicin response . When combined with low pH 6 , the response became comparable to the initial value ( 95% ) . Also reminiscent of the desensitized channel , 1 μM capsaicin plus pH 6 produced a partial response similar to that of 10 μM capsaicin alone . Without coexpressing the phosphatase , rapamycin itself had no detectable effect on the channel function ( unpublished data ) . Since the rapamycin assay depletes PIP2 without generating secondary signaling products ( e . g . , DAG and IP3 ) , our data support that the depletion of PIP2 on the plasma membrane indeed has an inhibitory role on TRPV1 . The effect of the depletion was consistent with that of the desensitization by Ca2+ influx , involving changes only on the apparent agonist sensitivity , but not the maximum attainable response of the channel . We further examined the voltage sensitivity of the channel after depletion of PIP2 using the rapamycin inducible system . Figure 5C illustrates current families recorded from same cells before and after application of rapamycin . No capsaicin was applied throughout experiments . Currents were elicited by brief depolarization ( 50 ms ) in 20-mV increments . Similar voltage responses were observed after depleting PIP2 . The resultant current-voltage relationships were also similar for depolarization up to +280 mV , indicating that the voltage dependence of the channel was largely unchanged ( Figure 5D ) . The results thus were parallel to those obtained with desensitization by capsaicin and Ca2+ influx . As a complementary experiment , we also examined the effect of the products of PIP2 hydrolysis that might occur during desensitization . Figure 5E–5H summarizes the results of OAG , an analog of DAG and IP3 . OAG ( 50 μM ) was applied to cells by local perfusion , while IP3 ( 20 μM ) was dialyzed into cells through the patch pipette . In both cases , 1 μM capsaicin evoked approximately the same current before and after application of each compound ( ∼5 min for IP3 and ∼1 min for OAG ) . The results thus further support a direct effect of PIP2 depletion on the channel function . To establish a role of PIP2 depletion in the desensitization of TRPV1 , it remained to be shown whether the extent of the depletion of PIP2 occurring during desensitization is strong enough to alter the channel activity . To address the issue , we detected the extent of the depletion of PIP2 induced by rapamycin , which has a known effect on the channel , and compared it to the depletion induced by Ca2+ influx . Figure 6A shows the simultaneous recording of the whole-cell current and the fluorescence of mRFP from a cell cotransfected with both the PIP2 probe ( mRFP-PH-PLC-γ ) and the constructs for rapamycin-induble depletion of PIP2 ( FKBP-Inp54p and Lyn-FRB ) . As expected , the application of rapamycin ( 0 . 1 μM ) caused a simultaneous decay of both the current and the fluorescence intensity . Notably , at the beginning of the treatment , the current appeared to reduce more slowly than the fluorescence intensity . It was also observed in the previous experiment that most of the reduction of the current occurred during the second half period of the treatment ( Figure 5A ) . The difference could arise if the depletion of PIP2 localized to the channel lagged behind the bulk PIP2 in the membrane , or the channel required only a relatively low amount of PIP2 to maintain its function . In either case , it is consistent with a high affinity of the channel to bind the lipid . At the end of ∼5-min rapamycin treatment , the fluorescence intensity of mRFP was reduced by ∼63% , while the corresponding whole-cell current changed by ∼80% relative to the initial 1 μM capsaicin response . This change of the fluorescence intensity was similar to that occurring during the desensitization by Ca2+ influx through the channel . Thus , the depletion of PIP2 during the desensitization could reach an extent that would diminish the response of the channel to 1 μM capsaicin . In our final experiment , we examined to what extent the depletion of PIP2 contributes to the desensitization of the channel . The contribution was quantified in terms of the shift of the dose-response relation following the depletion of PIP2 . Figure 7A shows the protocol of the experiment , where rapamycin ( 0 . 1 μM ) was continuously applied for ∼5 min until the response of the channel to 1 μM capsaicin was adequately reduced ( ∼10%–20% of the initial current ) . According to the previous measurement , such a treatment would give rise to a depletion level of PIP2 similar to that by the Ca2+ influx during the desensitization . Figure 7B shows the resulting dose-response curve measured at the end of the rapamycin application ( normalized to the initial 1 μM capsaicin response ) . Noticeably , after the depletion of PIP2 , the channel became only responsive at a concentration above 1 μM capsaicin . The treatment also mostly caused a shift of the EC50 while the peak response was retained . Fitting by the Hill's equation resulted in EC50 = 3 ± 0 . 3 μM and nH = 1 . 8 ± 0 . 1 . The resting channel had EC50 ≈ 0 . 4 μM ( Figure 1 ) . The rapamycin treatment caused an approximately 8-fold increase in the half-maximum effective concentration . This change accounted for ∼57% of the overall shift resulting from the desensitization by Ca2+ influx , indicating that the depletion of PIP2 constitutes a prominent component of the adaptation of the channel .
Both the desensitization mechanisms and the function of PIP2 have been extensively studied for TRPV1 , but a consensus has not been reached . The purpose of this study is to determine whether the depletion of PIP2 by Ca2+ influx through the channel has a causal relation to the desensitization . The problem has been difficult to resolve partly because the channel was thought nonfunctional after desensitization and partly because of lack of tools for reliably inhibiting the depletion of PIP2 . There is only one pharmacological inhibitor for PLC β and γ , which is known to have various side effects . Evidences supporting PIP2 depletion for TRPV1 desensitization mainly include the observation of PIP2 depletion occurring during desensitization and the requirement of PIP2 replenishment for recovery from desensitization [15] and the stimulatory effect of exogenous PIP2 applied in excised patches or dialyzed into cells [17–20] . Questions remain on whether the endogenous PIP2 has a similar effect in living cells , whether its depletion by Ca2+ influx is adequate for the induction of desensitization , and if so , to what extent it contributes . To address these issues , we directly measured the time course and the extent of depletion by combining patch-clamp recording with TIRF microscopy , an approach which allows the current and PIP2 fluorescence to be monitored simultaneously . To assess the function of PIP2 , we employed a rapamycin-inducible assay that can deplete PIP2 without activating secondary signaling cascades . Our data support that ( 1 ) the endogenous PIP2 can upregulate TRPV1 , ( 2 ) the depletion of PIP2 by Ca2+ influx is fast enough to regulate the channel , ( 3 ) the extent of depletion is adequate to alter channel function , and ( 4 ) the depletion of PIP2 accounts for ∼60% of the sensitivity shift induced by desensitization ( with 1 μM capsaicin and 1 . 8 mM Ca2+ ) . The actual contribution could be higher since PIP , the product of the 5-phosphatase , may partially substitute for PIP2 if the interaction is electrostatic . In a recent report , Lukacs et al . [18] showed that the rapamycin-induced PIP2 depletion had instead a stimulatory effect on TRPV1 at low capsaicin concentration ( 1 nM ) and was ineffective at the micromolar range , though their other evidence such as the effect of PIP2 in excised patches supports that the depletion of PIP2 inhibited the channel . We had not been able to observe this stimulatory effect in our experiments where cells generally showed indiscernible activity at concentrations below 30 nM ( Figure 1 ) . On the other hand , our rapamycin experiment at higher capsaicin concentrations produced a robust shift ( ∼8-fold ) in the capsaicin dose-response curve . The reason for this discrepancy is uncertain , but we noticed that we had used different FKBP agonist ( rapamycin versus rapalog ) and also different sources of constructs ( see Methods ) , which appeared to differ in several places including the membrane-anchoring protein domain and the wild-type FRB versus a mutant form . More recently , Klein et al . [20] also reported an inhibitory effect of rapamycin-induced PIP2 depletion on capsaicin-activated responses of TRPV1 , a result consistent with ours . The functional effect of PIP2 depletion on TRPV1 was in concordance with that of Ca2+ influx , both of which altered the agonist sensitivity while preserving the maximum response of the channel . The latter is consistent with a recent report that the channel could be reactivated after desensitization by capsaicin at extreme concentrations [28] . In retrospect , the quantitative measurement on the effect of PIP2 depletion explains some seemingly paradoxical observations . For example , while multiple mechanisms have been proposed for desensitization , the inhibition of PIP2 resynthesis was able to fully prevent channel recovery [15] . According to the present data , the depletion of PIP2 alone could render the channel unresponsive to normally saturating capsaicin . Thus , without replenishment of PIP2 , the channel would remain as if nonfunctional ( up to 1 μM ) , even though its sensitivity was partially recovered . The desensitization of TRPV1 has been difficult to study . The finding that the desensitized channel remained responsive to supramaximal stimuli may help alleviate the problem . The adaptive response of the channel provides insight about gating mechanism . Allosteric models are commonly applied to the gating of ion channels including TRPV1 [26 , 32–35] . In their simplest form such as the Monod-Wyman-Changeux ( MWC ) model , gating involves three processes: binding , intrinsic gating , and coupling between them . A change in any of these processes may give rise to a shift in the apparent agonist sensitivity as we observed after desensitization . However , changes in the intrinsic gating or the coupling would also affect the maximum open probability ( Po ) . Such changes may go undetected if the coupling strength of the agonist is so strong that the reduced maximum open probability remains close to unity . But this is unlikely for capsaicin , which , in the absence of low pH , activates a sub-maximal Po [26 , 35 , 36] . Furthermore , voltage is a much weaker activator of TRPV1 [33 , 34] . If the intrinsic gating was altered , the effect on voltage response would be more profound than on capsaicin or low pH . Our calculations show that the voltage response would be diminished by ∼90% if the intrinsic gating were altered so as to produce a 10-fold shift in the capsaicin sensitivity ( Protocol S1 ) . In contrast , experiments showed virtually no change in voltage responses before or after desensitization . Thus , both results indicate that desensitization is unlikely to alter the intrinsic gating of the channel . Retaining a constant maximum capsaicin response after desensitization also argues against changes in the coupling strength . In the context of the MWC model , a straightforward explanation for our observations is that desensitization altered the agonist binding . It appears that the binding of agonists is not a rigid docking process , but involves local structural changes . In other receptors the binding of agonists has been reported to involve residues far away from the binding site [37] . If capsaicin or proton binding in TRPV1 involves regions beyond their physical binding sites , PIP2 and other regulators may influence local structural changes in such regions to mediate binding . In retrospect , the result is consistent with the previous observation that protonation of the channel alters capsaicin binding [32] , suggesting possible interaction between regions mediating capsaicin and proton binding . Desensitization does not affect voltage responses if the voltage sensitivity resides in other regions of the channel . A recent study showed that the binding of ATP at the ankyrin repeats in TRPV1 altered capsaicin , but not voltage , responses [19] , further arguing for separation of agonist and voltage sensitivity . Implicit to this explanation of our data is that endogenous PIP2 must act as a modulator rather than an activator ( i . e . , weak coupling strength ) ; otherwise , depletion of PIP2 would have affected the maximum response to either capsaicin or voltage . Exogenous PIP2 or its analogs has been shown to activate TRPV1 in excised patches [17] . This increased sensitivity could arise from additional effects of these compounds at high concentrations . Modulation of the channel in the patch by the resting tension may also increase the sensitivity of the channel so that the normal “modulator” now adds enough energy causing the channel to open to a measurable degree [38] . We demonstrated the feasibility to integrate patch-clamp recording with TIRF imaging for the study of TRPV1 . The technique provides several advantages particularly useful for quantitative measurements . First , it avoids the fluorescence from cytoplasm where TRPV1 is predominantly expressed . Second , it allows for real-time detection of both electrical and optical signals . Third , the footprint in the evanescent field is relatively resistant to cell deformations arising from solution perfusion . One disadvantage is that the footprint membrane may reside in an environment different from the far field . The cleft between a poly-L-lysine-coated surface and the cell membrane is ∼12 nm [39] . Diffusion of agonist into this space may be slow , thereby creating a concentration profile different from the bulk . However , the problem should be minimized for the present study where capsaicin is hydrophobic and can diffuse in the membrane . Furthermore , the binding site for capsaicin lies intracellular [40] , so the agonist in the bulk phase also needs to cross the bilayer , which is likely rate-limiting . Another issue is the uncertainty regarding the [Ca2+] in the cleft . Assuming a diameter of 10 μM , the number of Ca2+ ions in the cleft , before activation of the channel , is ∼106 . With a whole-cell current of 10 nA , the bottom half surface of the cell will conduct ∼1 . 5 × 1010 Ca2+ ions/s . The Ca2+ ions initially trapped in the cleft would only sustain for an activation period of ∼1 ms . The replenishment of Ca2+ in the cleft takes >100 ms ( Figure 8 ) . Thus Ca2+ entry through channels in the footprint may be impeded , which will cause a relatively slower depletion of PIP2 than in the far field . This could be a reason why we sometimes observed a slight lag between the onset of the fluorescence decay and the inactivation of the current . The desensitization of membrane receptors is often considered as a mechanism to shape electrical responses and protect cells from toxicity . This is likely true for TRPV1 given its high permeability to Ca2+ . However , such a function could not explain some of the unconventional desensitization features . In particular , the desensitization of TRPV1 is not immediately reversible following the removal of agonist . Although a full recovery from desensitization is possible , it takes several minutes and requires a high concentration of ATP [15] . The present observation that the desensitization specifically affects the agonist sensitivity further alludes to other physiological functions . A conceivable candidate is adaptation . Adaptation has been demonstrated to other sensory systems , but its occurrence in nociception has remained elusive . Limited studies were mostly conducted long ago based on unit recording , and the results were controversial ( see , for example , [41] ) . On the other hand , it is known that capsaicin has dual effects , which at low concentration acts as an algesic while at high concentration as an analgesic . This analgesic effect is consistent with a strong shift of the agonist sensitivity and may be understood as a result of adaptation . That the receptor remains responsive to supramaximal concentrations predicts that further increasing the stimulus intensity would be able to continuously evoke pain with an intensity as strong as in the beginning . It will be interesting to test if this is indeed the case and more generally whether the adaptation of nociception occurs in vivo .
|
Sensory receptors can adjust their sensitivity to continuously varying stimuli , a process known as adaptation . Adaptation has been extensively studied in vision , hearing , and olfactory systems , but whether it also occurs to pain receptors has not been established . TRPV1 is an ion channel expressed in peripheral nerve terminals and is responsible for detection of pain-producing stimuli such as heat , acids , and irritant chemicals ( e . g . , capsaicin , the hot ingredient of chili peppers ) . We showed here that the channel has essential properties for adaptation since prolonged activation and calcium influx through the channel resulted in a dramatic decrease in sensitivity to further activation without reducing the maximal possible response of the channel . To address the mechanisms we simultaneously measured channel responsiveness and a component of the plasma membrane called PIP2 whose depletion may underlie desensitization . We showed that the depletion of PIP2 both had a time course synchronous to current desensitization and reached an extent adequate for significantly altering channel responsiveness , suggesting this process mediates the adaptation of TRPV1 channels . We postulate that adaptation is an important feature of pain receptors and may contribute to plasticity of pain sensation .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"physiology",
"biophysics",
"neuroscience"
] |
2009
|
Interaction with Phosphoinositides Confers Adaptation onto the TRPV1 Pain Receptor
|
Snakebite envenoming kills more than more than 20 , 000 people in Sub-Saharan Africa every year . Poorly regulated markets have been inundated with low-price , low-quality antivenoms . This review aimed to systematically collect and analyse the clinical data on all antivenom products now available in markets of sub-Saharan Africa . Our market analysis identified 12 polyspecific and 4 monospecific antivenom products in African markets . Our search strategy was first based on a systematic search of publication databases , followed by manual searches and discussions with experts . All types of data , including programmatic data , were eligible . All types of publications were eligible , including grey literature . Cohorts of less than 10 patients were excluded . 26 publications met the inclusion criteria . Many publications had to be excluded because clinical outcomes were not clearly linked to a specific product . Our narrative summaries present product-specific clinical data in terms of safety and effectiveness against the different species and envenoming syndromes . Three products ( EchiTabPlus , EchiTabG , SAIMR-Echis-monovalent ) were found to have been tested in robust clinical studies and found effective against envenoming caused by the West African carpet viper ( Echis ocellatus ) . Four products ( Inoserp-Panafricain , Fav-Afrique , SAIMR-Polyvalent , Antivipmyn-Africa ) were found to have been evaluated only in observational single-arm studies , with varying results . For nine other products , there are either no data in the public domain , or only negative data suggesting a lack of effectiveness . Clinical data vary among the different antivenom products currently in African markets . Some products are available commercially although they have been found to lack effectiveness . The World Health Organization should strengthen its capacity to assess antivenom products , support antivenom manufacturers , and assist African countries and international aid organizations in selecting appropriate quality antivenoms .
More than 100 , 000 people die from snakebite envenoming every year , associating this neglected tropical medical condition with one of the highest burdens of mortality of all neglected tropical diseases ( NTDs ) . In sub-Saharan Africa alone , snakebites are estimated to cause between 435 , 000 and 580 , 000 envenomings , and between 20 , 000 and 32 , 000 deaths every year [1] . 30 different species have been found to cause life-threatening envenomings [2] . While six different clinical syndromes are described by the World Health Organization ( WHO ) [2] , field organisations in sub-Saharan Africa , such as Médecins Sans Frontières / Doctors without Borders ( MSF ) , distinguish three major syndromes requiring antivenom therapy: neurotoxic , haematotoxic and cytotoxic ( see Table 1 ) . Prompt administration of a safe , effective and geographically appropriate antivenom is the cornerstone of effective snakebite management , although supportive care is crucial too , including assisted ventilation in case of neurotoxic envenoming . Despite the medical need for antivenom treatment , a decades-long antivenom supply crisis continues to affect sub-Saharan Africa [3] . Some locally inappropriate antivenom products , which are not prepared using the venom of snake species found in the sub-region have nevertheless been commercialised in sub-Saharan Africa [4] . In addition , several historical suppliers ceased production of their African antivenom products , citing limited profit . Notably , in 2014 Sanofi-Pasteur produced the last batch of Fav-Afrique , a polyspecific antivenom intended for use in sub-Saharan Africa and marketed for nearly two decades . The uncertain quality and specificity of certain antivenom products in many African countries has eroded the confidence of healthcare workers in antivenom therapy . In order to restore it , better regulation of the antivenom market , and the phase-out of ineffective and locally inappropriate products , is necessary . The manufacture of snake antivenom products follows a three-step process: immunisation of animals ( most often horses ) with a mixture of venoms; collection and fractionation of animal plasma , followed by refinement of immunoglobulins . In practice , products differ by the composition of the immunising venom mixtures , the animal immunisation protocols , the fractionation and purification processes , and the fragmentation ( or otherwise ) and concentration of immunoglobulins . Due to these major differences , the clinical safety and efficacy of an individual product cannot be extrapolated to another product . While the WHO recommends the evaluation of snake antivenom products in pre-clinical and clinical studies before their commercialisation [5] , the number of robustly-designed clinical trials on snakebite and antivenom therapy is extremely limited [6] . In sub-Saharan Africa , half of the clinical studies listed in a recent review were observational single-arm studies , which preclude any direct comparison between antivenom products [7] . Acknowledging this limitation , we nonetheless proceeded to review the available clinical data related to each individual antivenom product currently available for use in sub-Saharan Africa . The product-by-product clinical data summaries in this review do not allow for head-to-head comparisons between products . However , they may be helpful in determining the risk-benefit of each individual product in a specific region , dependent on the local distribution of snake species .
In 2014 we established a list of antivenom products commercially available in sub-Saharan Africa . The list was regularly updated and is presented in ( Table 2 ) . A number of different sources were used: the WHO database of venomous snakes , direct enquiries to suppliers , discussions with experts , and reporting from MSF teams at the project level . In comparison to a list published by Brown during a similar study [8] , a number of major changes are evident , reflecting the fluctuant nature of the African antivenom market: the production of several products has ceased , while new products have reached markets in sub-Saharan Africa . Although the manufacture of Fav-Afrique was ceased by Sanofi-Pasteur , we included the product in our list because another company , MicroPharm , has announced its intention to re-launch the product [9] . In order to improve readability , an abridged name was attributed to every individual product and used hereafter in the manuscript . For each individual product , Table 2 lists the different species that are mentioned on the product insert . This must be viewed with much caution . According to good practice , the list of species venoms used in the immunising mixture ( and therefore specifically neutralised by the antivenom ) should be clearly indicated on the product insert . This should be distinguished from the list of species venoms that can be neutralised by paraspecific activity . However , some product inserts simply provided a list of species venoms , without necessarily clarifying whether they had been used in the immunising mixture . Likewise , some manufacturers have yet to adopt the most recent taxonomic changes , for example the identification of new species within the Echis genus . Some of the antivenoms commercialized in sub-Saharan Africa are actually raised against venom of Echis carinatus , a species endemic in Asia , but not in Africa . While all carpet vipers of Africa and South Asia used to be classified as Echis carinatus , distinctive Echis species have now been identified ( i . e . Echis ocellatus , E . pyramidum , E . leucogaster ) and the composition of the immunizing mixtures of the different antivenoms intended for us in sub-Saharan Africa should take into consideration those changes . Finally , only a small minority of antivenom manufacturers report the geographical origins of the venoms used in the immunizing mixture , a good practice that should be generalized as there may be major intra-species venom variations between specimens of a given species coming from different geographical locations . The antivenom products that are included in this review have very different profiles . Most of them contain F ( ab ) ’2 fragments of equine immunoglobulins , but one product contains Fab fragments of ovine immunoglobulins , and another contains intact equine immunoglobulins . Most importantly , very different venom mixtures are used for the preparation of the different products . A few products are monospecific antivenom products , which are raised against the venom of only one snake species . Many products are polyspecific “pan-African” products that are raised against the venoms of the medically most important snakes across different sub-regions of sub-Saharan Africa . Between these two models , some antivenoms have a narrower polyspecificity; they are raised against a limited number of venoms of medically important snakes of a specific sub-region of sub-Saharan Africa . The first phase of our search strategy was a database search . We searched for publications related to any of the above antivenom products on PubMed ( Medline ) , Cochrane , Embase , Web of Science , Scopus , as well as on regional databases ( AJOL , Scielo ) . The keyword “antiven*” was used in association with: a ) geographical keywords ( “Africa” or regional country names ) ; b ) the names of the companies and products listed in ( Table 2 ) ; or c ) the taxonomic names of the medically most important African snake species . The search was performed on July 24th 2015 . The second phase of our search strategy employed additional search methods: the references of some specific articles and reviews were manually reviewed; experts were asked to review their own personal archives for additional studies; and conference abstract books were searched manually , including the most recent conference abstract books of the International Society of Toxinology and all conference abstract books of the African Society of Venimology . A complementary search was performed in January 2016 on PubMed alone with the keywords “VACSERA” and “Premium” as these manufacturers had not been included in the list of suppliers at the time of the initial database search . Finally , on February 5th 2018 , an additional PubMed search with the unique keyword “antiven*” was performed for the period since July 24th 2015 , in order to capture any additional papers featured in peer-reviewed journals since the first database search . All types of clinical data were eligible for inclusion: randomized controlled trials , case-control studies , observational cohort studies , case series , and programmatic data . All patient populations of all ages were included . Studies reporting less than 10 patients per antivenom product were excluded . No date restrictions were applied . All forms of publication were eligible for inclusion: peer-reviewed journal articles , university theses , conference abstracts , and posters . Only publications in French and English were included . In the event of duplicate publications , defined as different publications related to the same group of patients , the chosen study was either the most recent , or more often , the study with the largest published dataset . The standardised Newcastle-Ottawa Scale was initially proposed to evaluate the quality of the included studies . However , following capture of the relevant papers it was deemed not worthwhile , as the scale was not well adapted to the overall very low quality of selected studies . Instead we classified studies according to four categories that were adapted from Chippaux’s categorisation of clinical studies on snakebite envenomings [7]: While randomised clinical trials provided clinical evidence of the highest quality , anecdotal clinical studies provided the evidence of the lowest quality . Observational cohort studies and non-randomised comparative clinical studies provided evidence of moderate quality . However , heterogeneity was noted within each grouping . A wealth of clinical data could be extracted from some retrospective reports , while some prospective cohort studies were reported so poorly that only minimal data could be extracted .
ET-Plus was trialed in a large-scale RCT against ET-G in northern Nigeria to treat carpet viper envenomings ( Echis ocellatus ) [10] . 194 victims received ET-Plus . Results were good: blood coagulability , which is typically altered in carpet viper envenomings , was restored in 83% of patients within 6 hours; no fatalities were recorded . Of concern , adverse events were recorded in more than one-fourth of patients , including severe adverse events in one-tenth . An initial dose of three vials of ET-Plus was found to be a little more effective than an initial dose of one vial of ET-G , and a little less safe . A prospective study in Paoua , Central African Republic [11] , where Echis ocellatus is believed to be the medically most important species , was suggestive of ET-Plus effectiveness: there was only one death among the 306 victims of cytotoxic or hematotoxic envenomings who received ET-Plus . An immediate hypersensitivity reaction was seen in 21 patients ( 6 . 9% ) . A retrospective survey in Nigeria found that case fatality due to snakebite after introduction of ET-Plus was low [12] , but these results should be interpreted with caution given the low quality of programmatic data . Summary: ET-Plus was found to be satisfactorily clinically effective against Echis ocellatus envenomings; there is no clinical evidence on its effectiveness for envenomings caused by other species; the rate of adverse events appears moderate . ET-G was trialed in a large-scale RCT against ET-Plus in northern Nigeria to treat carpet viper evenomings ( Echis ocellatus ) [10] . 206 victims received ET-G . Efficacy was found to be good with an initial dose of one vial , although not as good as with an initial dose of three vials of ET-Plus: blood coagulability was restored in 76% of patients within 6 hours; no fatalities were recorded; adverse events were recorded in less than 20% of patients , of which severe adverse events were recorded in fewer than 5% of patients . Further to the above , an initial dose of one vial of ET-G was found to be a little less effective than an initial dose of three vials of ET-Plus , but a little safer . ThusET-G would seem to be more dose-effective and safe than ET-Plus to treat envenoming by Echis ocellatus . A low-quality retrospective survey was conducted in Nigeria following the introduction of ET-G [12] , which seems to indicate that mortality due to snakebite was low in patients treated with this antivenom . These results should be interpreted with much caution . Summary: ET-G was found to be very clinically effective against Echis ocellatus envenomings , the only species against which it is indicated , as clotting was restored in the majority of patients treated with just one vial . The rate of adverse events seems low-moderate . FAV-A was tested in four good quality prospective cohort studies in West Africa . In northern Cameroon [13] , in a region where Echis ocellatus envenomings are common , FAV-A was used successfully in all 41 patients , and only two minor adverse events were attributed to the antivenom . In a similar setting with 278 patients in Central Ghana [14] , FAV-A was associated with a low mortality rate of 1 . 8%; only 22% of patients required repeat antivenom doses . In southern Chad [15] , 4 of 60 patients treated by FAV-A died; of note , no repeat antivenom doses could be given to patients who would need them , due to limited resources . In a prospective study in Paoua in CAR [11] , a region where envenomings are caused predominantly by Echis ocellatus and occasionally by other species including neurotoxic elapids , there were two deaths among the 27 patients treated with FAV-A , both with features of neurotoxic syndrome . This raises questions of the effectiveness of FAV-A against elapid neurotoxic envenomings . In addition , 78% of snakebite victims in this cohort had to receive repeat doses of FAV-A . This contrasts with a previous retrospective analysis of 644 patients in Paoua [16] , which found that FAV-A was associated with a low mortality rate of 0 . 5% . In three cohorts in Djibouti [17 , 18 , 19] , FAV-A was found to restore blood coagulability following bites by Echis pyramidum . There were no fatalities among the total of 74 patients treated . In all of these studies , few adverse events were reported . But this apparent good safety profile should be interpreted with caution , as the quality of monitoring of adverse reactions varied across the studies . Summary: FAV-A was found to be clinically effective against African carpet viper envenomings , notably Echis ocellatus and Echis pyramidum . There are few clinical indications related to its effect on envenomings caused by other species . The rate of adverse events seems low . In a good quality post-marketing surveillance study in central Ghana [14] , a region where envenomings are often caused by Echis ocellatus , ASNA-C was associated with a very high mortality rate of 22% , with a mean number of 11 . 7 vials used per patient . More than half of the 66 treated patients had to be administered repeat antivenom doses . Five cases of anaphylactic shock were reported amongst 66 patients . In a programmatic setting in Nigeria , from a low quality report [20] , ASNA-C was believed to be ineffective in restoring blood coagulopathy , and to cause many cases of allergic reactions . Summary: ASNA-C was found ineffective at neutralising envenomings caused by Echis ocellatus . The rate of severe adverse events appears high . Two observational cohort studies were conducted in a setting where snakebite caused by the carpet viper Echis ocellatus are a major cause of envenomings . In northern Benin [21] , Antivip-A was found effective at stopping bleeding within 2 hours in 60% of patients . Nine of 289 persons treated with Antivip-A died , including one girl who did not receive a full dose due to a shortage , six severe cases admitted with complications , and two cases believed to be bitten by a snake of the genus Atractaspis spp . In Paoua , Central Africa Republic [16] , results with Antivip-A appeared less beneficial in a smaller cohort treated by MSF; five of 50 persons treated with Antivip-A died . Of note , in four of five cases , the second dose of two vials of Antivip-A was given with much delay , more than 12 hours after the initial dose . In 10 of 13 patients with visible bleeding , bleeding was not stopped within two hours following antivenom administration . Antivip-A was also tested in Kindia , lower Guinea . All 118 patients treated with Antivip-A for what seems to be mild cytotoxic or mild haematotoxic envenomings survived [22] . However 4 of 22 patients treated for neurotoxic envenomation died . A susbsequent study in the same setting reviewed the efficacy of Antivip-A in an overlapping cohort of patients with neurotoxic syndrome caused by Elapidae [23] . The case-fatality rate in the groups treated with a low dose or high dose of Antivip-A was 15 . 4% and 17 . 6% respectively . An absence of clinical benefit was observed . Of note , assisted ventilation , a critical component of neurotoxic envenoming treatment , was not available . Across these studies , a low rate of adverse events of between 10% and 15% was reported . Summary: Conflicting results exist in relation to the effectiveness of Antivip-A for the neutralization of viperid bites in West Africa . In neurotoxic envenomings , Antivip-A showed poor results . Its safety profile appears good . One multicentre observational clinical study in northern Benin and in lower Guinea evaluated Inoserp-P in 209 patients [24] . A low case fatality rate , with one death among 109 patients was reported in lower Guinea , where neurotoxic envenomings represented 12% of admitted cases . In northern Benin , where many cases are caused by the carpet viper Echis ocellatus , four of 100 treated patients died . Blood coagulability was found to be restored within 24 hours in 98% of patients . Adverse events were reported in only 8% of patients . Inoserp-P was also evaluated in Senegal in 63 patients [25] . It appeared to be effective and well tolerated , in spite of protocol deviations , including a lower initial dose than what is recommended . Blood coagulability was restored within 24 hours in 87 . 5% of patients . There were two deaths , including one neurotoxic pediatric case and one hematotoxic case in an adult presenting five days after the bite . Summary: There is scant evidence available related to Inoserp-P . Inoserp-P seemed relatively effective in West African settings , and well tolerated . SAIMR-Poly is one of the most clinically-trusted antivenoms in sub-Saharan Africa , but there is paradoxically very little published material providing robust evidence of its efficacy . Most studies are of a poor quality , based on retrospective reports involving a small number of patients . Of the six studies included in this review [26 , 27 , 28 , 29 , 30 , 31] SAIMR-Poly was used in 144 envenomed patients , and fatality was reported in only five cases . More information is available on adverse events associated with SAIMR-Poly . An observational study in South Africa found that 13 of 17 patients who received SAIMR-Poly had severe early anaphylactoid reactions [27] . Varying rates of adverse events were reported in other publications of lower quality , with one publication expressing concerns over the high rates of adverse events [31] . Summary: There is limited evidence on the effectiveness of SAIMR-Poly . The rate of adverse events seems high . Three studies in the 1970s evaluated SAIMR-Echis in northern Nigeria [32 , 33 , 34] , where Echis ocellatus is a frequent cause of envenoming . In a randomised controlled trial [34] , the antivenom was found more effective in the treatment of carpet viper envenomings than a polyspecific antivenom then manufactured by Behringwerke . While no fatality was recorded in the 46 patients who received either antivenom , SAIMR-Echis reversed haematological abnormalities more rapidly and at a lower dosage than did Behringwerke’s antivenom . Similar observations were made in two less robust studies in the same region [32 , 33] . No fatality was recorded in the groups of 48 and 16 patients respectively who were treated for Echis ocellatus envenoming with SAIMR-Echis . In the latter group , bleeding was found to stop for the majority within 24 hours , although recurrent bleeding was observed in four patients a few days later . Adverse side effects were reported in two of the three above studies . In one case series [33] , 14 of 48 patients who received SAIMR-Echis had reactions . In the randomised controlled trial [34] , immediate hypersensitivity was observed in four of 23 patients . Summary: There is fairly robust evidence that SAIMR-Echis is able to treat the typical haemotoxic syndromes caused by Echis ocellatus envenoming . The rate of adverse events seems between moderate and high . One retrospective study reported mortality outcomes in 23 patients with prolonged clotting time in Gondar , north-west Ethiopia [35] . There were four deaths ( 17% ) in this group . However , the number of vials that were given to the snakebite victims in this cohort was sub-optimal due to a short supply; most victims received between one and three vials , while between three and six vials were required according to clinicians . Summary: There is one anecdotal report that suggests limited effectiveness of VACSERA-Poly in north-west Ethiopia . We could not find any publicly available clinical evidence that met our inclusion criteria related to the remaining antivenom products listed in ( Table 1 ) .
A minority of the antivenoms included in this review were supported by robust clinical data prior to their registration and commercialization in African countries . The absence of good quality , clinical effectiveness and safety data for the majority of these products is a major concern , as is the absence of publicly available pre-clinical data for some products . This inacceptable situation prompted the WHO to commission in December 2015 the preclinical testing of the different products intended for use in Africa in a systematic and blinded manner . The results of the assessment will be crucial to determine which products should be phased out , and which should be rolled out in the different sub-regions of Africa . The overall dearth of clinical data on antivenoms intended for use in sub-Saharan Africa must be addressed . Comparative clinical trials should be implemented to compare the safety and effectiveness of products . Clinical trials must adopt a multi-centre methodology , in order to provide evidence of the effectiveness of products in different sub-regions and against different snake species . Many of the studies included in this review took place in rural hospitals and clinics with a long experience of snakebite management . With additional support , these sites could potentially host clinical trials . For as long as anti-venom treatment is distributed in sub-Saharan Africa without adequate supporting clinical data , the safety and effectiveness of such treatment cannot be ascertained . Urgent investments in research are required to more accurately determine the regional specificity of existing forms of antivenom treatment . Additional financial and structural investments are required to ensure the sustained production and supply of antivenom as a priority intervention to reduce snakebite-associated morbidity and mortality across sub-Saharan Africa .
|
Snakebite envenomation represents one of the most neglected tropical medical conditions worldwide . Despite high levels of morbidity and mortality associated with snakebite , its neglected nature has compromised the availability and evaluation of antivenom treatment . This review was initiated by Médecins Sans Frontières’ / Doctors Without Borders ( MSF ) Access Campaign , as the existing antivenom access crisis in sub-Saharan Africa was deepening . This study sought to review the clinical data pertaining to each antivenom product currently available for use in sub-Saharan Africa . 16 different antivenoms were identified . A total of 26 studies met the inclusion criteria . Given the heterogeneity of study methodology and quality , data for individual antivenoms are presented in the form of a narrative analysis . Only two studies reported clinical data collected from randomized controlled trials . Consistent monitoring of the side effects of antivenom treatment was lacking . The absence of good quality data for the majority of antivenoms in sub-Saharan Africa is a major concern . Further robust data collection is required , while urgent investments are needed at the global level to ensure a sustained production of safe and effective antivenom treatment , and its affordable access across sub-Saharan Africa .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"toxins",
"pathology",
"and",
"laboratory",
"medicine",
"clinical",
"research",
"design",
"tropical",
"diseases",
"geographical",
"locations",
"vertebrates",
"database",
"searching",
"animals",
"toxic",
"agents",
"toxicology",
"research",
"design",
"signs",
"and",
"symptoms",
"reptiles",
"neglected",
"tropical",
"diseases",
"africa",
"snakebite",
"research",
"and",
"analysis",
"methods",
"venoms",
"vipers",
"adverse",
"events",
"snakes",
"people",
"and",
"places",
"eukaryota",
"diagnostic",
"medicine",
"squamates",
"database",
"and",
"informatics",
"methods",
"hemorrhage",
"biology",
"and",
"life",
"sciences",
"vascular",
"medicine",
"amniotes",
"organisms"
] |
2019
|
Reviewing evidence of the clinical effectiveness of commercially available antivenoms in sub-Saharan Africa identifies the need for a multi-centre, multi-antivenom clinical trial
|
In Human African Trypanosomiasis , neurological symptoms dominate and cardiac involvement has been suggested . Because of increasing resistance to the available drugs for HAT , new compounds are desperately needed . Evaluation of cardiotoxicity is one parameter of drug safety , but without knowledge of the baseline heart involvement in HAT , cardiologic findings and drug-induced alterations will be difficult to interpret . The aims of the study were to assess the frequency and characteristics of electrocardiographic findings in the first stage of HAT , to compare these findings to those of second stage patients and healthy controls and to assess any potential effects of different therapeutic antiparasitic compounds with respect to ECG changes after treatment . Four hundred and six patients with first stage HAT were recruited in the Democratic Republic of Congo , Angola and Sudan between 2002 and 2007 in a series of clinical trials comparing the efficacy and safety of the experimental treatment DB289 to the standard first stage treatment , pentamidine . These ECGs were compared to the ECGs of healthy volunteers ( n = 61 ) and to those of second stage HAT patients ( n = 56 ) . In first and second stage HAT , a prolonged QTc interval , repolarization changes and low voltage were significantly more frequent than in healthy controls . Treatment in first stage was associated with repolarization changes in both the DB289 and the pentamidine group to a similar extent . The QTc interval did not change during treatment . Cardiac involvement in HAT , as demonstrated by ECG alterations , appears early in the evolution of the disease . The prolongation of the QTC interval comprises a risk of fatal arrhythmias if new drugs with an additional potential of QTC prolongation will be used . During treatment ECG abnormalities such as repolarization changes consistent with peri-myocarditis occur frequently and appear to be associated with the disease stage , but not with a specific drug .
Human African Trypanosomiasis ( HAT ) or sleeping sickness evolves in two stages , the first or early ( hemo-lymphatic ) stage and the second or late ( meningo-encephalitic ) stage which is characterized by invasion of the central nervous system ( CNS ) by trypanosomes . Neuropsychiatric disturbances are the most prominent and best documented features of the disease [1] . Cardiac involvement plays an important role in American trypanosomiasis ( Chagas' disease ) ; however , in the African form , cardiac involvement has been suggested but has never been studied systematically in the first stage of the disease . Cardiac involvement has been observed in up to 73% of HAT patients in post mortem histological studies [2] , [3] . Those findings are supported by the recent study of Blum et al [4] that showed cardiac alterations in 71% of second stage HAT patients , but are in contrast to previous studies , where ECG findings were reported in only 35–48% of the patients [5]–[7] . The latter studies included both first and second-stage HAT patients . A low prevalence of ECG findings in first stage disease could explain this discrepancy . Thus , cardiac involvement , as documented by ECG findings , may parallel CNS involvement and ECG findings could be used as additional tool for assessing the advancement of the disease . Because of increasing resistance to the available drugs for HAT , new compounds or drug combinations are desperately needed . Evaluation of cardiotoxicity and the risk of cardiac arrhythmia is one parameter of drug safety , but without knowledge of heart involvement in HAT , cardiologic findings and drug-induced ECG alterations will be difficult to interpret . Pentamidine administered intramuscularly is currently the primary treatment for first stage HAT . A large number of diamidine compounds have been synthesized in an attempt to develop an oral agent for this disease . DB289 ( pafuramidine maleate ) is one of these diamidine compounds . It can be orally administered and showed good efficacy against first stage HAT in Phase II trials . Since diamidines such as pentamidine have been shown to have arrhythmic potential [8] , knowledge of HAT cardiopathy and scrupulous analysis of the potential cardiac effects of new antiparasitic drugs is essential . The overall aim of the study was to assess the cardiac involvement in first stage HAT by ECG examination and to study the effect of different antiparasitic drugs on ECG findings .
The objectives of this study were to assess the frequency and character of ECG findings in patients with first stage HAT and to compare them to healthy control subjects and to second stage HAT patients . Secondary objectives were to assess differences between administered HAT therapies , including ECG changes during and after treatment , and to discuss the findings with respect to clinical relevance and tolerability of medical therapy . The objectives were to study and characterize ECG alterations in T . b . gambiense patients with respect to the stage of the disease ( first versus second stage ) and treatment induced alterations ( baseline versus after treatment ) . Only studies with clear definition of ECG criteria and complete ECG description ( including QTc intervals ) , description of the stage of the disease and ECG before and after treatment were included . Using these criteria the following studies were not included: Electrocardiograms ( ECG ) were performed prior to and following treatment and analyzed in a total of 523 participants; 406 were patients with first stage HAT , 56 with second stage HAT and 61 were healthy controls . Patients and controls from different clinical trials are included in the present analysis . All HAT patients and healthy controls underwent a clinical assessment , including medical history , baseline physical examination , blood sampling for hematology and chemistry , and ECG . To estimate the normal intra-individual fluctuation of ECG parameters , two ECG recordings per subject were obtained at baseline . ECG changes appearing after treatment were compared to these intra-individual ECG changes . ECG data were also obtained during treatment in first stage HAT patients in all patients treated with DB 289 and in 40 patients treated with pentamidine . A clinical assessment was performed and the ECG was repeated after completion of treatment of HAT . All ECG tracings were interpreted by a single reader using standardized criteria as described below . The ECG data for the first stage HAT patients were compared to the ECG data from the healthy controls and the second stage HAT patients . The PQ , QRS and QT intervals were measured manually by the principal investigator in three consecutive cycles and mean values were calculated . Measures of the intervals were performed in lead II , when feasible with lead V2 or I as second choice . QTc was calculated by the Bazett formula ( QTc = QT/SQR ( RR ) . QTc shorter than 440 ms and shorter than 460 ms were considered normal for men and women , respectively [18] , [19] and because a QTc longer than 500 ms is known as predictor of torsades de pointes [19] both limits were used for the analysis . For overall ECG interpretation , the following criteria were used . Right atrial hypertrophy ( RAH ) : p>2 . 5 mV; left atrial hypertrophy ( LAH ) : p>120 ms; right ventricular hypertrophy ( RVH ) : Sokolow index right: RV1 and S V5>1 . 05 mV , left ventricular hypertrophy ( LVH ) : Cornell voltage: R aVL and SV3 men >2 . 8; women >2 . 0; peripheral low voltage: R I and RII and R III<1 . 6 mV; PR depression: >0 . 8 mV; ST elevation: >0 . 1 mV without notch , concave , from deep S; ST depression: >1 mV; repolarization changes: limb leads: discordant in at least one lead; precordial leads: negative in either V3 , V4 , V5 or V6 . Early repolarization type: ST elevation concave , notch at the J point , positive T waves . Normal ECG included axis deviation , early repolarization type [20] , ST elevation >0 . 1 mV without notch , concave , from deep S in precordial leads [20] and partial right bundle brunch block ( RBBB ) . Minor ECG changes included intraventricular conduction delay , left or right atrial hypertrophy , isolated premature atrial or ventricular captures and left anterior hemiblock ( LAHB ) . Major changes included: AV block I–III , low voltage , left and right ventricular hypertrophy , complete bundle branch block , PR depression , ST depression and repolarization changes . Written informed consent ( illiterates signed by fingerprint ) was obtained from all study participants . Ethical approval was granted by the Ethics Committees of the DRC , Angola , South Sudan and the Ethics Committee for the two cantons of Basel , Switzerland ( Ethikkomission beider Basel ) . Data from the various studies were pooled into one single database that contained the clinical examination , demographic data and the ECG details . Analysis was done using the statistical software package STATA 9 . 0 ( www . stata . com ) . All continuous variables are reported as mean±SD . A p-value<0 . 05 was considered statistically significant . Nominal variables were compared using the Χ2- test . Comparisons between the patient groups and treatments were performed using the t-test , ANOVA plus Bonferroni correction , the Kruskal Wallis or Mann Whitney U tests where appropriate .
Patient baseline characteristics are summarized in Table 1 . Age and gender distribution were similar among the disease stages , the HAT treatment groups and the healthy controls , respectively . Pyrimethamine-sulfadoxine ( SP ) was given as first line malaria treatment prior to HAT treatment in all centers with the exception of CNPP Kinshasa , where quinine was used as standard malaria treatment due to the high level of SP resistance in Kinshasa . The ECG baseline intervals and characteristics are listed by disease stage in Table 2 . At baseline , QTc prolongation , which was defined as >440 ms in men and , >460 ms in women , was observed in 11–13% of all HAT patients . These QTc values are considered to represent an increased risk for arrhythmia . Only one patient had a QTc interval over 500 ms , which is associated with an elevated risk for torsade de pointes . The proportion of major ECG findings indicating heart involvement was significantly lower ( p-value = 0 . 0001 ) in first stage ( 53 . 5% ) than in second stage HAT ( 69 . 5% ) . The QTc interval of HAT patients treated with melarsoprol following malaria treatment was 412 msec in the sulfadoxin/pyrimethamin group ( SD 19 ) and 431 msec ( SD 24 ) in the quinine group . Changes of ECG intervals and findings according to the different treatment groups are shown in the Table 3 and 4 . During treatment , no patient in the DB289 or the pentamidine groups developed a QTc longer than 500 ms . One patient with a QTc longer than 500 ms at baseline had a normal QTc after treatment . In the group of second stage patients , one patient developed a significantly prolonged QTc interval during melarsoprol treatment . The development or disappearance of AV block I consisted mostly of increases or decreases of a few milliseconds , usually from just below to just above the upper limit of normal ( 200 ms ) or vise versa . During the treatment period in the DB289 group no relevant conduction problems such as AV block II or III or ventricular arrhythmias were seen . In the pentamidine group , two patients developed an AV block II ( Type Wenckebach ) , which resolved spontaneously and was asymptomatic . One patient with second stage HAT developed a bigeminal rhythm during treatment with melarsoprol , which subsided after administration of corticosteroids . There were no further significant changes in rhythm or conduction , such as ventricular arrhythmia , appearance of AV block III or formation of bundle branch block , observed in ECG recordings during treatment compared to baseline or during treatment compared to after treatment .
Cardiac involvement , as demonstrated by ECG alterations , appears early in the evolution of HAT and precedes CNS involvement . As HAT itself was associated with QTc prolongation in more than 10% of patients , the additional risk of a drug with potential QTc prolongation properties has to be considered because of the risk of fatal arrhythmias . DB289 and pentamidine treatment were not associated with prolongation of the QTc intervals and they had no obvious cardiotoxic effect . During treatment , ECG changes such as repolarization alterations occurred frequently , were not associated with one specific drug , and were more common in the second stage of the disease .
|
In Human African Trypanosomiasis ( HAT ) , neurological symptoms dominate and cardiac involvement has been suggested . Because of increasing resistance to the available drugs for HAT , new compounds are desperately needed . Evaluation of cardiotoxicity is one parameter of drug safety , but without knowledge of the baseline heart involvement in HAT , cardiologic findings and drug-induced alterations will be difficult to interpret . The electrocardiogram ( ECG ) is a tool to evaluate cardiac involvement and the risk of arrythmias . We analysed the ECG of 465 HAT patients and compared them with the ECG of 61 healthy volunteers . In HAT patients the QTc interval was prolonged . This comprises a risk of fatal arrhythmias if new drugs with antiarrhythmic potential will be used . Further , repolarization changes and low voltage were more frequent than in healthy controls . This could be explained by an inflammation of the heart . Treatment of HAT was associated with appearance of repolarization changes but not with a QTc prolongation . These changes appear to be associated with the disease , but not with a specific drug . The main conclusion of this study is that heart involvement is frequent in HAT and mostly well tolerated . However , it can become relevant , if new compounds with antiarrhythmic potential will be used .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"cardiovascular",
"disorders"
] |
2009
|
Cardiac Alterations in Human African Trypanosomiasis (T.b. gambiense) with Respect to the Disease Stage and Antiparasitic Treatment
|
Mammalian genomes contain several dozens of large ( >0 . 5 Mbp ) lineage-specific gene loci harbouring functionally related genes . However , spatial chromatin folding , organization of the enhancer-promoter networks and their relevance to Topologically Associating Domains ( TADs ) in these loci remain poorly understood . TADs are principle units of the genome folding and represents the DNA regions within which DNA interacts more frequently and less frequently across the TAD boundary . Here , we used Chromatin Conformation Capture Carbon Copy ( 5C ) technology to characterize spatial chromatin interaction network in the 3 . 1 Mb Epidermal Differentiation Complex ( EDC ) locus harbouring 61 functionally related genes that show lineage-specific activation during terminal keratinocyte differentiation in the epidermis . 5C data validated by 3D-FISH demonstrate that the EDC locus is organized into several TADs showing distinct lineage-specific chromatin interaction networks based on their transcription activity and the gene-rich or gene-poor status . Correlation of the 5C results with genome-wide studies for enhancer-specific histone modifications ( H3K4me1 and H3K27ac ) revealed that the majority of spatial chromatin interactions that involves the gene-rich TADs at the EDC locus in keratinocytes include both intra- and inter-TAD interaction networks , connecting gene promoters and enhancers . Compared to thymocytes in which the EDC locus is mostly transcriptionally inactive , these interactions were found to be keratinocyte-specific . In keratinocytes , the promoter-enhancer anchoring regions in the gene-rich transcriptionally active TADs are enriched for the binding of chromatin architectural proteins CTCF , Rad21 and chromatin remodeler Brg1 . In contrast to gene-rich TADs , gene-poor TADs show preferential spatial contacts with each other , do not contain active enhancers and show decreased binding of CTCF , Rad21 and Brg1 in keratinocytes . Thus , spatial interactions between gene promoters and enhancers at the multi-TAD EDC locus in skin epithelial cells are cell type-specific and involve extensive contacts within TADs as well as between different gene-rich TADs , forming the framework for lineage-specific transcription .
Metazoan development requires the concerted specification of divergent lineages among a genetically homogenous cell population and the tightly controlled , coordinate genesis of cellular structural and functional diversity driven by proper spatial and temporal regulation of transcription . Genome topology in the nucleus plays an important role in regulation of gene transcription by facilitating or restricting spatial interactions between gene promoters and distal gene regulatory elements [1–7] . In the interphase nucleus , chromosomes occupy distinct positions called chromosome territories with some intermingling between the borders of the neighboring chromosomes [8] . Each chromosome is organized into Topologically Associating Domains ( TADs ) , principal units of the chromatin folding that might be further divided into sub-TADs [9–11] . TADs range in size from several hundred Kb up to about 1 . 5 Mb in mice and humans . TADs are defined as chromatin domains with higher frequency of spatial contacts within the domains compared to the regions across TAD borders [9 , 10 , 12] . TAD borders are mostly conserved between the different cell types and mammalian species [9 , 12 , 13] , although lineage-specific differences in the TAD borders have been described [9 , 12] . The spatial chromatin contacts involve interactions between proximal gene promoters and distal gene regulatory regions , such as enhancers , silencers , insulators and locus control regions . These interactions vary substantially between different cell types and change during cell differentiation [6 , 11 , 12 , 14] . The spatial interactions between gene promoters and enhancers mostly occur within TADs [15–17] . However , less frequent inter-TAD contacts occur between the transcriptionally active loci , and these often represent enhancer-promoter contacts that are largely cell-type specific [17–19] . The functional significance of the inter-TAD contacts remains to be further determined . Spatial genome organization is controlled , at least in part , by a number of chromatin architectural proteins including CCCTC- binding factor ( CTCF ) , Cohesin , condensin together with the Mediator co-activator complex [11 , 20–22] . CTCF binding is often detected at the TAD borders , although most CTCF bound regions are found inside the TADs [9 , 10 , 12] . Cohesin is frequently , but not always , binds together with CTCF at the bases of chromatin loops [14 , 20 , 21 , 23] . Cohesin controls spatial contacts between gene promoters and enhancers together with or independently of CTCF [11 , 20 , 22] . The Mediator complex is also frequently involved in the promoter-enhancer interactions together with cohesin [11 , 22 , 24] . Functionally-related and co-regulated genes frequently form conserved clusters or loci in the mammalian genomes , which size varies from several kilobases to several megabases [25 , 26] . In mouse genome , there are several dozens of the large ( more than 0 . 5 Mbp ) gene loci , in which gene transcription is frequently regulated in a lineage-specific manner [27 , 28] . Large lineage-specific gene loci are present on a vast majority of chromosomes and harbour the olfactory receptor family genes ( chromosomes 2 , 7 , 9 , 10 , 14 , 16 , 17 , 19 ) , immunoglobulin kappa and heavy chain genes ( chromosomes 6 and 12 , respectively ) , keratinocyte-specific genes ( chromosomes 3 , 11 , 15 and 16 ) , as well as some other gene families [29] . The detailed chromatin conformation capture analysis of the several lineage-specific gene loci , including the Hox , beta–globin and protocadherin genes , revealed the importance of their proper spatial organization in executing lineage-specific gene expression programs by restricting the promoter-enhancer contacts to individual TADs [21 , 30–32] . However , high resolution mapping of the chromatin interaction networks in the large lineage-specific loci and their relevance to the distinct TADs remain largely unexplored . Epidermal Differentiation Complex ( EDC ) is a unique large locus in the mouse genome containing 61 functionally-related genes occupying 3 . 1 Mb domain in the gene-rich region of mouse chromosome 3 or 1 . 6 Mb domain on human chromosome 1 [33–35] . Central part of the EDC locus contains functionally-related genes involved in the control of epidermal differentiation and barrier acquisition , while two flanking EDC regions harbour the S-100 family genes involved not only in epidermal differentiation , but also playing various functions in other tissues [33–35] . In mouse genome , the central part of the EDC is separated from its 5’- ( centromere proximal ) domain by a gene desert , while another gene-poor domain separates the 3’-flank of the EDC from the neighbouring gene-rich domain on chromosome 3 [33–35] . During epidermal morphogenesis and transition of the single-layered surface epithelium ( E11 . 5 ) to stratified epidermis ( E16 . 5 ) , higher-order chromatin folding of the EDC harbouring region on the mouse chromosome 3 show remarkable plasticity resulting in relocation of the EDC from the nuclear periphery towards nuclear interior [36] . These changes are associated with remodelling of chromatin compaction in the central EDC domain and increased transcription of many EDC genes involved in terminal keratinocyte differentiation [29] . Developmentally-regulated higher-order chromatin remodelling of the EDC locus in keratinocytes is orchestrated by the epidermal master transcription regulator p63 , which directly regulates expression of the ATP-dependent remodeller Brg1 and genome organizer Satb1 in the epidermal progenitor cells [29 , 36] . In turn , Brg1 controls the developmentally-regulated relocation of the EDC towards the nuclear interior , while Satb1 promotes establishing the proper level of chromatin compaction in the central EDC domain required to maintain or balance gene transcription in the locus in terminally differentiating keratinocytes [29 , 36] . In spite of the essential role of the EDC locus in epidermal differentiation and barrier acquisition , surprisingly little is known about the distal gene regulatory elements in this region and their interactions with the target gene promoters . Several non-coding regulatory elements showing the enhancer or silencer activities were identified in this locus based on the non-coding sequence homology in mammalian species [33] . 3C studies demonstrated the long-range spatial contacts between a conserved AP-1 dependent gene enhancer with the selected gene promoters in this locus in cultured epidermal keratinocytes [37] . However , the comprehensive pattern of spatial chromatin contacts in the EDC locus , including its organization into distinct TADs , promoter-enhancer regulatory network and the factors involved in its establishment and maintenance remain unexplored . Here , we map the spatial chromatin contacts at the EDC and neighbouring genome region in murine basal epidermal keratinocytes and thymocytes ( used as a control in which the keratinocyte-specific genes at the EDC are inactive ) , at high resolution employing the Chromosome Conformation Capture Carbon Copy ( 5C technology ) . We demonstrate that in keratinocytes , the EDC locus is organized into several gene-rich and gene-poor TADs and forms lineage-specific spatial contact networks . Furthermore , in keratinocytes , in addition to the intra-TAD contacts , a substantial number of keratinocyte-specific spatial interactions connecting putative gene enhancers with promoters were detected between different gene-rich TADs . We also show enrichment for binding of CTCF , Rad21 , and ATP-dependent chromatin remodeller Brg1 in the spatial enhancer-promoter contacts within and between gene-rich TADs , suggesting their roles in the establishment of the unique spatial chromatin organization and control of gene expression in the large multi-TAD EDC locus in skin epithelial cells .
In the mouse genome , there are 33 large ( more than 0 . 5 Mbp ) lineage-specific gene loci containing at least 10 functionally related genes [27–29] ( Fig 1 ) . Large lineage-specific gene loci are present on a vast majority of chromosomes and harbour the olfactory receptor family genes ( chromosomes 2 , 7 , 9 , 10 , 14 , 16 , 17 , 19 ) , immunoglobulin kappa and heavy chain genes ( chromosomes 6 and 12 , respectively ) , keratinocyte-specific genes ( chromosomes 3 , 11 , 15 and 16 ) , as well as some other gene families ( Fig 1A and 1B ) . To correlate the genomic location of such loci to the distinct TADs genome-wide , we used the TAD maps generated using Hi-C technology for mouse embryonic stem cells [9] . Interestingly , this analysis revealed that among 24 lineage-specific loci occupying between 0 . 5–1 . 6 Mbs in the genome , 21 loci were localized within single TADs on the corresponding chromosomes , while only 3 loci were spread between two neighbouring TADs ( Fig 1A and 1B ) . Epithelial-specific gene loci , such as Keratin type I and type II [KtyI/II] loci , Keratin-associated protein [KAP] locus having size between 0 . 75–1 Mbs , were localized within individual TADs on mouse chromosomes 11 , 15 and 16 , respectively ( Fig 1B ) . However , 100% lineage-specific loci of larger size ( >1 . 6 Mb ) including Epidermal Differentiation Complex [EDC] locus [34] were occupying several ( from two to four ) TADs on the corresponding chromosomes ( Fig 1A and 1B ) . Because TADs were defined as genomic regions with a higher frequency of spatial contacts within the domains compared to inter-domain interactions [9 , 10 , 12] , these data raised the questions whether large multi-TAD lineage-specific gene loci display any unique features in the chromatin interaction patterns between functionally related genes localized in different TADs and how such interactions are regulated . To address this question , we focused on the EDC locus occupying ~3 . 1 Mb in one of the most gene-dense regions of mouse chromosome 3 [29] . Its central domain consists of the co-regulated genes involved in the control of terminal keratinocyte differentiation and epidermal barrier acquisition , including Loricrin ( Lor ) , the Small proline-rich ( Sprr ) gene family , Involucrin ( Ivl ) , Late cornified envelope ( Lce ) gene family and Fillagrin-like ( Flg-like ) gene family [33 , 36 , 38] ( Fig 2A ) . The S100 family genes flank the 5’- and 3’-ends of the EDC [38] ( Fig 2A ) . In addition to the gene-rich domains , the EDC locus in mice contains a gene-poor region ( “desert” ) separating the part of S100 family genes at 5’ of the EDC and the Lor gene at central EDC domain , while another gene-poor domain separates the 3’-flank of the EDC from the neighbouring gene-rich domain on chromosome 3 ( Fig 2A , S1 Table ) . To study the potential connection between gene activity and spatial chromatin folding in the EDC locus at higher resolution , we correlated gene expression determined by microarray profiling with data obtained with Chromosome Conformation Capture Carbon Copy ( 5C ) technology in freshly plated neonatal epidermal keratinocytes . Consistently with the data demonstrating specific roles for many genes that constitute the central EDC domain in the control of epidermal barrier formation [39–42] , microarray showed that in epidermal keratinocytes most of these genes were expressed at much higher levels compared to thymocytes , used as the control in which those genes were inactive ( S1A Fig , S1 Table ) . 5C is well-suited for analyses of the spatial genome folding , as it allows the simultaneous detection of the spatial chromatin looping contacts and identification of TADs [17 , 43] . For the 5C analysis , 381 forward and 382 reverse 5C probes were designed in an alternating scheme using the tools from my5C software suite [44] to interrogate HindIII sites with the unique anchoring regions at the EDC and its flanking regions ( mm9 , chromosome 3: 89 . 9–95 . 2 Mbp ) ( Fig 2B , S2 Table ) . The designed probe pool interrogated 145 , 542 potential pair-wise spatial chromatin contacts within this 5 . 3 Mbp genomic region . Two biological 5C library replicates were generated and analyzed for each cell type . Consistently with previous studies using 5C and Hi-C technologies [9–11 , 45] , the raw 5C data for both replicates ( shown as heatmaps with all reverse probes plotted as columns and forward probes as rows ( Fig 2C , S1B Fig ) , clearly demonstrated that neighbouring chromatin regions interact to each other frequently , creating a black “diagonal” in the middle part of the heatmaps . Raw 5C data showed high similarity between the biological replicates for both cell lineages ( Fig 2C , S1B Fig ) , and the raw 5C counts highly correlated between the replicates ( for keratinocyte libraries—Pearson correlation coefficient 0 . 88; for thymocyte libraries—Pearson correlation coefficient 0 . 94 ) , indicating a high quality of our 5C data ( Fig 2C , S1B Fig ) . However , the correlations between the keratinocyte and thymocyte libraries were much lower ( Pearson correlation coefficient 0 . 61 ) , indicating lineage-specific differences in folding of the locus between both cell types . Correction of the 5C data for non-biological biases associated with this technology was performed as described previously [17 , 43] ( see Materials and Methods for details ) ( S2–S5 Figs ) . The corrected data were binned ( bin size 150kb with the step size of 15kb ) to account for the differences in the 5C probe coverage in the different parts of the 5 . 3 Mbp genomic region ( Fig 2D , S2E Fig , S3E Fig , S4E Fig , S5E Fig and S6A Fig ) . The heatmaps representing 5C data clearly showed several consecutive chromatin regions with high spatial self-associations ( visible as darker “triangles” above a black “diagonal” ) corresponding to the distinct TADs in keratinocytes and thymocytes ( Fig 2D , S6A Fig ) [9 , 10 , 17] . To define the positions of the TAD boundaries , we performed the insulation index analysis on each replicate data set separately , as described elsewhere [17 , 46] ( see Materials and Methods for details ) ( Fig 2D , S6A Fig , S3 Table ) . This analysis identified the boundaries separating TADs in the 5 . 3 Mbp region in keratinocytes and thymocytes ( Fig 2D , S3 Table ) . The accuracy of the TAD boundary calculations performed by comparing the determined boundary midpoint positions between the replicates ( S3 Table ) indicated that TAD boundary positions were determined with the accuracy of about +/- 100 kb . 5C data revealed that 5 . 3 Mbp chromatin domain containing EDC locus on mouse chromosome 3 is organized into seven distinct TADs ( Fig 2D , S3 Table ) . We calculated density of the protein-coding genes in these TADs and correlated the results to the average gene density in the mouse genome ( 75 kb per gene ) . Based on these analyses , we defined the gene-rich ( 21–36 Kb per gene ) and gene-poor ( 166–400 Kb per gene ) TADs in the EDC locus ( S1 and S3 Tables ) . 5’-flanking region of the EDC locus containing the S100 family genes constituted the part of the gene-rich TAD1 , which also harbours neighbouring non-EDC genes including the house-keeping Rps27 gene . Gene-poor domain of the EDC separating S100 family genes from its central domain constituted the TAD2 . In turn , the central EDC domain and its 3’ flank region were organized into two distinct TADs: Lor gene was located at the boundary between TAD2 and TAD3 , which also contained the Ivl gene , Sprr gene family and the major part of the Lce gene family , while TAD4 encompassing the remaining part of the Lce gene family , Flg-like gene family , 3’-flanking part of S100 gene family , and the part of Tdpoz gene family . The chromatin domain located further outside of the 3’-end of the EDC locus was organized into gene-poor TAD5 containing the remaining part of the Tdpoz gene family , as well as into gene-rich TAD6 and part of the TAD7 , respectively ( Fig 2D , S3 Table ) . These data were quite consistent with the Hi-C data obtained from mouse embryonic stem cells [9] ( Fig 1B , S1 Table ) , as well as with 5C data obtained from thymocytes . In thymocytes , the border between TAD T1 and TAD T2 ( 90 . 8 Mb ) , as well as between TAD T2 and TAD T3 ( 92 . 1 Mb ) were only slightly shifted compared to keratinocytes ( 90 . 7 Mb and 90 . 9 Mb , respectively ) ( S6A and S3 Tables ) . Similar to keratinocytes , the central EDC domain in thymocytes was organized into TAD3 ( 92 . 1–92 . 7 Mb ) and TAD4 ( 92 . 7–93 . 9 Mb ) ( S6A Fig , S3 Table ) . However , TAD T4 and TAD5 in thymocytes did not show clear separation and the border between them was rather softened ( Fig 2D , S6A Fig , S3 Table ) . The borders between the TAD 4/5 and TAD 6 ( 93 . 9 Mb ) , as well as between TAD6 and TAD7 ( 94 . 8 Mb ) in thymocytes were quite similar compared to the corresponding borders in keratinocytes ( Fig 2D , S6A Fig , S3 Table ) . Importantly , the TAD borders were weaker in thymocytes versus keratinocytes , while the frequency of the spatial inter-chromatin contacts both within and between different TADs in keratinocytes was substantially higher in comparison to thymocytes ( Fig 2D , S6A Fig ) . Interestingly , we observed high frequency of the spatial chromatin contacts between the gene-poor TAD2 and TAD5 , flanking the gene-rich TAD3 and TAD4 in keratinocytes , while such interactions were not seen in thymocytes ( Fig 2D , S6A Fig ) . Such high frequency of contacts was not observed on the heat map between TAD1 and TAD3 , separated by the gene-poor TAD2 in keratinocytes ( Fig 2D ) . These data suggested that the gene-poor TADs at the 5 . 3 Mbp chromatin domain on mouse chromosome 3 appears to be segregated into transcriptionally-inactive compartment spatially separated from the transcriptionally active gene-rich TADs in keratinocytes , which is quite consistent with the model proposing the existence of the compartments A and B topologically separated in the nucleus based on the differences in their transcription activity [9 , 45] . Importantly , such separation was not observed for the transcriptionally inactive gene-rich and gene-poor TADs in thymocytes ( S6A Fig ) , presumably incorporated into compartment B in these cells . To validate the 5C data , we performed 3D-FISH analysis of the distances between loci located in the distinct EDC domains in the freshly plated primary epidermal keratinocytes and thymocytes , as well as in cryo-sections of P0 . 5 mouse skin in situ . First , we checked if the central part of the EDC is indeed organized into two adjacent gene-rich TAD3 and TAD4 in both cell types ( Fig 2D , S6A Fig ) . For the 3D FISH analysis , we have chosen the BAC probes depicting the regions within the TAD3 near its 5’ and 3’ borders ( probes A and B , respectively ) , or located within the adjacent TAD4 ( probe C ) ( S4 Table , Fig 2D–2F , S6B Fig ) . We expected that the spatial distances between the regions located within the same TAD should be shorter in comparison to the distances between the regions located in the different TADs , when the similar linear genomic distances separate such regions [9 , 10] . Indeed , 3D-FISH analyses demonstrated that despite the fact that the genomic distances between the centers of the regions covered by the probes A and B located within the same TAD3 were slightly longer ( 716 , 849 bp ) compared to the distances between the regions covered by the probes B and C ( 638 , 779 bp ) located in the TAD3 and TAD4 , respectively ( S4 Table ) , spatial distances between the centers of the 3D FISH signals generated by the probes A and B were significantly shorter compared to the distances between the probes B and C in all cell populations ( Fig 2D–2F , S6A Fig , S5 Table ) . Thus , this analysis confirmed the folding of the EDC central domain into two separate TADs both in keratinocytes and thymocytes . Importantly , 3D-FISH data also showed that the spatial distances between the 3D-signals were rather similar in basal epidermal keratinocytes in situ and in the freshly isolated epidermal keratinocytes ( Fig 2F , S6B Fig , S5 Table ) , thus confirming that the cell isolation procedure for 5C does not significantly alter the spatial organization of the EDC locus . However , 3D FISH analysis also revealed that the distances between the probes A—B and B—C were significantly larger in thymocytes compared to keratinocytes ( p<0 . 0001 , Mann-Whitney U-test ) ( S6B Fig , S5 Table ) , demonstrating that chromatin in the transcriptionally inactive domains of the EDC locus in thymocytes is less condensed and likely to be more randomly folded compared to the active locus in keratinocytes . Next , we checked whether gene-poor TAD2 and TAD5 are indeed located closely to each other in keratinocytes , as this has been suggested by the 5C data ( Fig 2D ) . We performed the 3D-FISH analysis of the basal epidermal KCs in situ using the probes covering the centre of the TAD2 ( probe D ) and TAD5 ( probe E ) ( Fig 2D , S4 Table ) . TAD2 and TAD5 are separated from each other in the genome by the gene-rich TAD3 and TAD4 ( Fig 2D ) . 3D-FISH data showed that despite the genomic distances between the regions covered by the probes D and E were much longer ( 2 , 610 , 522 bp ) compared to the distances between the probes B and C that depict TAD3 and TAD4 ( 638 , 779 bp ) , the spatial distances between the probes D and E , as well as between the probes B and C , were quite similar ( Fig 2F , S5 Table ) . These data demonstrated close association of the gene-poor TAD2 and TAD5 in keratinocytes , thus demonstrating the consistence with the 5C results . Thus , 3D-FISH analyses confirmed the organization of the central and 3’-flanking regions of the EDC into two separate gene-rich TADs in both cell lineages , as well as the compartmentalization of the gene-poor TAD2 and TAD5 in keratinocytes . Furthermore , this analysis also confirmed the less condensed and potentially more randomly spatially organized the transcriptionally inactive locus in thymocytes in comparison to the active locus in keratinocytes . The concordance between the 5C and 3D-FISH data , as well as between 3D-FISH data obtained from isolated keratinocytes and basal epidermal keratinocytes in situ suggested that the gene-rich and gene-poor TADs in the EDC locus and its neighbouring regions indeed form a unique and relatively stable spatial composition that might serve as a platform for the control of lineage-specific transcription . To further characterise the spatial chromatin interaction network at the EDC locus and distinguish “true” chromatin interactions in the EDC locus from the random background interactions , we used an approach described previously [17 , 43] , which is based on the establishment of the background baseline defining the expected frequency of the random chromatin contacts normalized to the genomic distances separating the interacting fragments . This approach allowed identifying interactions reproducible in both 5C replicates with significantly higher interaction frequency compared to the background: 1139 “true” interactions in keratinocytes and 1033 interactions in thymocytes; q-value<0 . 05; ( Fig 3A , S6 Table , S6C Fig , S7 Table ) . The reproducibility of the called 5C interactions between both replicates was similar to the previously published 5C datasets [11 , 17 , 43] . To compare the common and cell-type specific 5C interactions between keratinocytes and thymocytes , we also identified a subset of the interactions that were interrogated in all four 5C libraries after the 5C dataset normalization . This approach revealed 338 keratinocyte-specific 5C interactions , 747 thymocyte-specific interactions , while only 136 interactions were common between both cell types ( S6D Fig , S8–S10 Tables ) . Thus identification of the “true” 5C interactions in keratinocytes and thymocytes further demonstrated that spatial organization of the EDC locus is largely lineage-specific . To further characterize the patterns of the 5C interactions in the transcriptionally active EDC locus in keratinocytes and to check if there are differences in the frequency of the “true” 5C spatial contacts within and between different TADs , we used the 5C contact sets reproducible in both keratinocyte libraries ( S3A Fig , S6 Table ) . Interestingly , we identified substantially more 5C interactions between different TADs ( 799 or 70 . 15% ) , than within the individual TADs ( 340 or 29 . 85% ) ( Fig 3A ) . Analyses of the 5C interactions between different TADs revealed that gene–rich TAD3 and TAD4 harbouring the majority of genes activated during terminal keratinocyte differentiation interact equally extensively with the gene-rich TAD1 and gene-poor TAD2 ( S6E Fig ) . However , TAD1 harbouring a part of the S100 family genes showed a markedly decreased number of interactions with neighbouring gene-poor TAD2 compared to more distantly located TAD3 and TAD4 ( S6E Fig ) . Remarkably , gene-poor TAD5 showed preferential interactions with gene-poor TAD2 , which , in turn , interacted quite extensively with the gene-rich TAD3 and TAD4 ( S6E Fig ) . Gene-rich TAD6 that does not contain keratinocyte-specific genes also interacted quite extensively with TAD1 , TAD2 , TAD3 and TAD4 , while showed only very limited number of interactions with neighbouring gene-poor TAD5 ( S6E Fig ) . Next , we checked the frequency of the inter-TAD and intra-TAD 5C interactions at the EDC locus as a function of the genomic distances separating contacting fragments in keratinocytes . Surprisingly , we found that the frequency of all detected spatial contacts within the TADs were generally only slightly higher in comparison to the contacts between the TADs ( S6F Fig ) Such extensive chromatin interaction network between different neighbouring gene-rich TADs harbouring the lineage-specific genes , as well as lineage-specific folding of the EDC locus suggests the functional relevance of these contacts for coordination of the gene expression in keratinocytes during execution of epidermal differentiation program . 5C analysis demonstrated that majority of all 1139 “true” 5C interactions in keratinocytes ( 47 . 3% ) involve the contacts between the gene promoters and non-promoter chromatin domains , while considerably lower number of interactions were involving either two promoters ( 26 . 3% ) or two non-promoter chromatin domains ( 26 . 4% ) , respectively ( Fig 3B ) . Thus , vast majority of the 5C contacts ( 73 . 6% ) involve the non-promoter elements ( possibly including gene enhancers ) at the EDC locus in keratinocytes . To further characterize the 5C interactions between gene promoters and enhancers , we identified putative gene enhancers in the EDC locus and its neighbouring regions by performing ChIP-seq analysis for enhancer-specific histone modifications with anti-H3K4me1 and anti-H3K27ac antibodies on the freshly isolated FACS sorted basal ( Integrin 6 alpha high , Sca1 high ) mouse epidermal keratinocytes . ChIP-seq analyses revealed 16 regions in the EDC locus and its neighbouring regions with the high levels of both H3K4me1 and H3K27ac modifications , serving as the signatures of active enhancers [3 , 47 , 48] ( Fig 3C , S11 Table ) . Interestingly , the putative active enhancers were identified exclusively in the gene-rich TADs: TAD1 ( E1-E7 ) , TAD3 ( E8 ) , TAD4 ( E9-E11 ) , TAD6 ( E12-E14 ) and TAD7 ( E15-E16 ) ( Fig 3C ) . Among these enhancers , two groups of closely located enhancers ( within less than 10 kb distance from end to end for each enhancer: E2/E3 and E4-E7 ) formed two clusters ( potential super-enhancers ) within the TAD1 , while the enhancers within other TADs were quite distantly located from each other and did not show clustering ( serving probably as typical enhancers ) ( Fig 3C , S11 Table ) . Moreover , lack of any enhancers was seen in the gene-poor TAD2 and TAD5 . To identify spatial interaction network between the enhancers and gene promoters , we assigned the 5C interactions involving the restriction fragments within 10 kb of each enhancer to either corresponding individual enhancers ( E1 , E8-E16 ) or to the clusters of closely located enhancers ( E2/E3 and E4-E7 ) . We found that 22% ( 252 out of 1139 ) of all 5C looping interactions at the EDC were anchored to the fragments bearing the gene enhancers ( S12 Table ) . Then we assigned the closest gene transcription start sites ( TSSs ) located not further than 10 kb away from the restriction fragments anchoring the 5C interactions on the opposite side of the enhancers ( Fig 4A , S13 Table ) . Our analysis revealed that about 52% ( 144 out of 273 ) of the 5C interactions involving gene enhancers were the interactions between gene enhancers and promoters , consistently with the data showing the involvement of the gene enhancers in long-range spatial contacts with the target promoters [43 , 49 , 50] . All enhancer-bearing regions , except the one for E16 , were engaged in multiple spatial chromatin interactions with the regions anchoring gene promoters , revealing the potential enhancer-promoter regulatory network ( Fig 4A , S12 Table ) . All enhancers or enhancer clusters except E8 were involved in the long-range contacts with multiple gene promoters , consistently with data obtained from other cell types [16 , 43 , 51] ( Fig 4A , S13 Table ) . In turn , some gene promoters in the EDC locus were involved in the long-range spatial contacts with several enhancers . For instance , Ivl gene was involved in contacts with the enhancer clusters E2/E3 and E4-E7 , while the S100a11 gene was interacting with enhancers E9 and E11 ( Fig 4B , S13 Table ) , consistently with observations that gene promoters might interacts with several enhancers [17 , 43 , 49] . Enhancers were frequently involved in the spatial interactions not only with the gene promoters located in the same TADs , but also with the gene promoters located in the different TADs . For instance , in addition to the interactions with multiple gene promoters in the TAD1 , a cluster of the enhancers E2/E3 ( located in TAD1 ) were interacting with the regions containing Sprr3 and Ivl gene promoters in the TAD3 , the Crct1 , Lce3d , S100a10 and S100a11 gene promoters in the TAD4 , the Pi4kb , Pogz , Them5 and Tuft1 gene promoters in TAD6 , as well as with the Bnipl gene promoter in the TAD7 ( Fig 4C , S14 Table ) . Enhancer E9 , located in the TAD4 , spatially contacted the promoter regions of Flg , Rptn , S100a10 , S100a11 , Tchh and Tchhl1 genes in the same TAD , as well as to the promoter regions of Ints3 , Npr1 and Pglyrp4 genes in the TAD1 , the promoter of Sprr2h gene in the TAD3 , and promoters of Cgn , Lingo4 and Them5 genes in the TAD6 ( Fig 4D , S14 Table ) . These data were quite intriguing , as many recent studies demonstrated that the contacts between promoters and enhancers are mostly constrained by the same TADs [15–17] . Interestingly , we also found a relatively low number of interactions between the enhancers located in gene-rich TADs with distinct chromatin domains located in gene-poor TAD2 and TAD5 ( Fig 4A , S12 Table ) . The vast majority of such interactions involved distal elements not associated with any gene promoters in the TAD2 and TAD5 , although interaction between the Tdpoz3 gene promoter ( TAD5 ) and the E14 enhancer located in TAD6 was also seen ( S14 Table ) . Next , we compared the frequency of promoter-enhancer contacts between and within TADs as a function of genomic distances separating interacting regions [17 , 52] . We found that all promoter-enhancer spatial interactions within the TADs were connecting the regions separated by genomic distances of up to 0 . 6 Mb , while the inter-TAD interactions were much longer connecting the regions separated from each other by 0 . 5 Mb-5 . 1 Mb distances ( Fig 4E ) . As expected , the frequencies of short-range intra-domain contacts were higher compared to the long-range inter-TAD contacts . However , several inter- and intra- domain contacts found between the promoters and enhancers separated by similar genomic distances had comparable frequencies ( Fig 4E ) . Thus , our data revealed the organization of the enhancer-promoter network in the EDC locus with a high frequency of short-range contacts within gene-rich TADs and less frequent , but extensive long-range promoter-enhancer interactions between gene-rich TADs , while gene-poor TADs were lacking of any enhancers . To gain further insights about the proteins that could be potentially involved in the control of higher-order chromatin folding and promoter-enhancer interactions at the EDC locus in keratinocytes , we correlated the 5C interaction data with the ChIP-seq data for the binding of the chromatin architectural proteins CTCF and cohesin subunit Rad21 , known to control the higher-order chromatin folding in all studied cell types [4 , 53] . We also correlated 5C data with ChIP-seq data for the binding of ATP-dependent chromatin remodeler Brg1/Smarca4 , known to regulate nuclear positioning of the EDC locus in keratinocytes during epidermal development [36] . We found a heterogeneous distribution in the binding patterns for these proteins at the EDC locus ( Fig 5A and 5B ) . CTCF and Rad21 showed high frequency of the binding in the gene-rich TAD1 , TAD4 , TAD6 and TAD7 , while lower frequency of binding was seen in the TAD3 and TAD2 and lack of binding was detected in TAD5 ( Fig 5A and 5B ) . We found CTCF binding within 100kb of all the TAD border midpoints , except the border between TADs 5 and 6 , where it was found within 160kb ( Fig 5A ) , consistent with recently established role for CTCF in the TAD organization [9 , 15 , 21] . Similarly to CTCF and Rad21 , Brg1 binding was abundant in all gene-rich , but not gene-poor TADs ( Fig 5A and 5B ) . A substantial fraction of the 5C interactions showing CTCF , Rad21 and Brg1 binding ( between 38% and 50% of all interactions for the individual proteins ) ( Fig 5C ) , suggested that they might be involved in the control of the higher-order chromatin folding at the EDC locus in keratinocytes . Exact Fisher statistical test showed the enrichment for the regions bound by CTCF in all significant 5C looping interactions in comparison to all background 5C interactions ( Fig 5D ) . This was consistent with a well-established role of CTCF in the control of higher-order chromatin folding in different cell types [4 , 53 , 54] . We further analyzed the pair-wise combinations of the chromatin architectural protein binding in the regions anchoring the 5C interactions at the EDC locus in keratinocytes . Consistently with the previously published data , our analysis revealed most frequent presence of the cohesin subunit Rad21 in the regions anchoring the 5C interactions that were also anchored to the CTCF binding regions ( 76 . 6% ) ( Fig 5E ) . Brg1 was also frequently seen in the regions involved in the 5C interactions anchoring CTCF-bound regions ( 63 . 4% ) ( Fig 5E ) . CTCF , and Brg1 were present in the regions anchoring 59 . 4% , and 56 . 5% of the 5C interactions anchored to the Rad21 binding regions respectively ( Fig 5E ) . These data demonstrates that CTCF , Rad21 and Brg1 frequently present in the regions anchoring the same 5C interactions , suggesting that they might functionally cooperate in the control of establishment of the spatial interacting network within the EDC locus and its genomic neighbourhood in keratinocytes . Next we check if CTCF , Rad21 , and Brg1 are involved in spatial contact between gene promoters and enhancers . We found that CTCF , Rad21 and Brg1 were even more frequently bound to the bases of the 5C loops involving gene enhancers than in the bases of all significant 5C loops ( Fig 5C and 5F ) . Exact Fisher statistical test demonstrated highly significant enrichment of this protein binding in the regions anchored to the enhancer spatial interactome ( Fig 5G ) , supporting their involvement in establishing promoter-enhancer contacts in keratinocytes . This is consistent with the role of Rad21 together with or independently from CTCF that has been well documented in several cell types [4 , 53] . Moreover , Brg1 binding has also been reported to be frequently associated with active enhancers [55 , 56] , and promoter-enhancer spatial interactions [57] . Thus , our data suggest the important role for CTCF , Rad21 and Brg1 in organization of the 5C interactome within and between gene-rich TADs in the EDC locus in keratinocytes and in establishing promoter-enhancer spatial network in this locus .
Mouse genome contains 11 large multi-TAD gene loci , occupying >1 . 6 Mb each on the corresponding chromosomes , show a clustering of functionally related genes whose transcription is regulated in a lineage-specific manner [9 , 27 , 28] . In this manuscript , we demonstrate that in skin epithelial cells , EDC is organized into four TADs with the distinct chromatin interaction patterns within and between these and neighbouring TADs involving gene promoters and enhancers . We also show the promoter-enhancer anchoring regions in the gene-rich transcriptionally active TADs are enriched for the binding of chromatin architectural proteins CTCF , Rad21 and chromatin remodeler Brg1 . In contrast to gene-rich TADs , gene-poor TADs show preferential spatial contacts with each other , do not contain active enhancers and show decreased binding of CTCF , Rad21 and Brg1 in keratinocytes . The validation of the 5C data by 3D-FISH analyses performed according to the recommendations published previously [32] confirm that in epidermal keratinocytes , the central gene-rich EDC region , harbouring the majority of the genes activated during terminal keratinocyte differentiation , has two adjacent gene-rich TAD3 and TAD4 , which are flanked by two gene-poor TAD2 and TAD5 further surrounded by the gene-rich TAD1 , TAD6 and TAD7 ( Fig 2D ) . Our 5C data at the EDC locus in keratinocytes are concordant with the data on the TAD organization identified by Hi-C approach in mouse embryonic stem cells and our 5C data in thymocytes used as a control in which EDC locus is largely inactive ( Figs 1B and 2D ) [9] . Some differences in the positions of the TAD borders between these datasets might reflect the differences in the resolution depth depicted by the 5C and Hi-C technologies , or real differences in the TAD borders between pluripotent ( ground state of TAD organization in embryonic stem cells ) versus differentiated cells . It remains to be determined whether these differences might also be linked to the distinct chromatin compartmentalization patterns in keratinocytes and thymocytes associated with striking differences in the EDC gene transcription between two cell lineages . Combination of the 5C , 3D-FISH and ChIP-seq approaches reveal several differences between gene-rich and gene-poor TADs that constitute EDC locus and its neighbouring regions in epidermal keratinocytes . Gene-rich and gene-poor TADs within the locus show distinct inter-TAD spatial chromatin contact patterns . Gene-poor TADs ( TAD2 and TAD5 ) and gene-rich TADs ( TAD3 and TAD4 ) are compartmentalized in the nucleus as distinct topological domains , the transcriptionally inactive chromatin domains ( compartment B ) and active transcription ( compartment A ) [12 , 45] . However , TAD2 and TAD5 show heterogeneity in their chromatin interaction patterns–TAD5 show preferential interactions with TAD2 , while TAD2 also interacts with neighbouring gene-rich TAD3 and TAD4 . In contrast to gene-rich TADs , gene-poor TADs do not contain active enhancers and show markedly decreased binding of CTCF , Rad21 and Brg1 proteins . Interestingly , the network of spatial interactions involving gene promoters and enhancers at the EDC locus in keratinocytes are not restricted to intra-TAD interactions , but the interactions extend to different gene-rich transcriptionally active TADs . Our 5C data demonstrate that that majority ( 73 . 6% ) of the “true” 5C contacts in 5 . 3 Mb chromatin domain in keratinocytes analysed in this study are mapped at sites near gene promoters and their interactions connect to non-promoter chromatin domains ( 47 . 3% ) or to other promoters ( 26 . 3% ) . The promoter-promoter interactions are recently demonstrated using high-resolution capture Hi-C [49] , and they are frequently identified by the ChIA-PET approach using anti-RNA polymerase II antibody [58] . The role of the promoter-promoter contacts in gene expression control is not well understood , however , promoters can share common transcription factories ( foci enriched in RNA polymerase II ) [59 , 60] , while some promoters can function as enhancers for their interacting promoter partners [58 , 61] . Correlation of the 5C data with ChIPseq analyses for enhancer-specific histone modifications in KCs ( high level of H3K4me1 and H3K27ac ) reveal 16 putative active gene enhancers at the EDC locus in keratinocytes . Two of these enhancers ( E9 and E11 ) ( Fig 3C ) were previously identified based on the non-coding region homology between several mammalian species and were shown to possess the enhancer activity in enhancer-reporter assay in cultured mouse keratinocytes [33] . About 52% of the significant 5C contacts involving gene enhancers show their interaction with the gene promoters , thus supporting a view on functional importance of such contacts identified in this study . However , further analyses are required to demonstrate functional relevance of these spatial contacts to the control of gene transcription in the epidermal progenitor cells and differentiating keratinocytes . Intriguingly , in addition to the intra-TAD contacts , we demonstrate the extensive enhancer-promoter interactions across the TADs borders . Although less frequent , these contacts were longer-ranged ( from 500 kb to 5 . 1 Mb ) compared to the intra-TAD contacts ( up to 600Kb ) . These data are consistent with recent reports demonstrating the presence of promoter-enhancer contacts across TAD boundaries in different cell types [16 , 17 , 19] . In cultured mouse keratinocytes , the recent 3C data identified interactions between the AP-1 dependent enhancer located in the TAD3 with several promoters within TAD3 and TAD4 , as well as with S100a6 promoter in TAD1 [37] . However , it is still unclear whether adjacent closely associated TADs can share regulatory elements by forming meta-TADs at large loci harbouring multiple co-regulated genes , similarly to the meta-TAD domains described in differentiating neuronal progenitor cells [18] . Interestingly , the enhancers found in TAD1 form two closely located clusters ( E2/E3 and E4-E7 ) , embedded into the genes of S100 family . These enhancer clusters showed extensive long-range intra-TAD chromatin contacts with multiple genes in the central part of the EDC ( TAD3 and TAD4 ) activated during terminal keratinocyte differentiation , suggesting that they might serve as the locus-control regions or super-enhancers for the EDC genes . In addition , we identified the gene enhancer ( E9 ) spatially interacting with Flg gene promoter ( Fig 4D ) . These enhancers have been previously identified among the highly-conserved non-coding regions in several mammalian genomes and showed the activity in the reporter assay in cultured keratinocytes [33] . It will be important to determine if this conserved enhancer controls Flg gene expression in normal and diseased epidermis , as the defects in Flg gene and changes in its expression are associated with ichthyosis vulgaris , the most common disorder of epidermal differentiation , and also serve as strong risk factors for atopic eczema [62] . The binding studies for chromatin architectural proteins CTCF , Rad21 , and ATP-dependent chromatin remodeler Brg1 revealed the enrichment in the CTCF binding in the regions anchored to all significant 5C contacts . In particular , binding of the CTCF , cohesin complex subunit Rad21 and ATP-remodeller Brg1 was enriched in the regions anchoring the 5C interactions involving gene enhancers within gene-rich TADs . These findings are consistent with the well-established roles of CTCF and cohesin complex in the control of spatial genome topology [11 , 13 , 20 , 22] . Recent Hi-C data from Khavari’s lab on human keratinocytes also revealed a role for Rad21 in the control of enhancer-promoter contacts in both progenitors and differentiated cells ( J Invest Dermatol , 2017 , 137 , 5S , S80 , abstract ) . In addition , Brg1 is frequently found at the gene enhancers [55 , 56] and it was reported to be involved in the enhancer-promoter looping interactions [57 , 63] . However , since CTCF and cohesin are ubiquitously expressed across the broad range of cell types , suggesting that additional proteins with more restricted expression patterns might be involved in shaping lineage-specific spatial genome organization . Taken together , our findings provide new insights into the spatial chromatin organization at the large multi-TAD EDC locus with extensive spatial contacts involving gene promoters and enhancers within and between different gene-rich TADs . Such interactions might contribute to the coordinated gene regulation in the EDC locus during terminal keratinocyte differentiation in the epidermis . These data serve as an important platform for future studies to reveal the intricate interplay between the chromatin architectural protein , chromatin remodelers , transcription factors and gene regulatory elements in the control of spatial genome organization and gene expression programmes in basal and differentiating epidermal keratinocytes during normal skin development and homeostasis , as well as during skin responses to environmental stressors and in disorders of epidermal differentiation , such as atopic dermatitis , psoriasis and cancers .
All animal studies were performed under protocol approved by the University of California Berkley Institutional Animal Care and Use Committee and the UK Home Office Project Licence . C57Bl/6 mice were purchased from Charles River . The skin tissue samples were collected from P1 . 5-P3 . 5 C57Bl/6 animals as previously described [64 , 65] . Keratinocytes were isolated for micro-array , 5C , 3D FISH and ChIP-seq analysis from the skin of the new born C57BL/6 animals . Primary thymocytes were isolated from the C57Bl/6 animals . For the FISH analysis of 3D preserved nuclei , skin samples were processed as previously described [66 , 67]] Primary epidermal keratinocytes was isolated from the skin of the new-born C57Bl/6 mice as previously described [68 , 69] . Briefly , the skin was removed from the neonatal mice and incubated with 0 . 25% trypsin in Hanks Balanced Salt solution overnight at 4 C , followed by separation of dermis from epidermis . Epidermis was placed into pre-chilled low calcium primary keratinocyte culture ( EMEM , 4% chelated FBS , 0 . 05mM CaCl2 , 0 . 4ug/ml hydrocortison , 5ug/ml insulin , 10mg/ml EGF , 10−10 M cholera toxin , 2x10-9 T3 , 2mM L-glutamin , 100U/ml penicillin , 100ug/ml streptomycin ) and triturated to obtain the single cell suspension . The cells were filtered through a 70 μm silicon strainer and were either seeded at high density at the low calcium primary keratinocyte medium onto collagen solution ( 0 . 97X Hanks Balanced Salt Solution ( HBSS ) , 9 . 70μg/mL Bovine Serum Albumin ( BSA ) , 19 . 40 mM 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ( HEPES ) , 0 . 97 X Vitrogen-100 Collagen ) coated culture dishes for 15 hours at 32°C in the atmosphere of 8% carbon dioxide and 90% humidity , or were used for FACS to isolated viable basal keratinocyte population . Primary thymocytes were isolated from C57Bl/6 mouse thymi as described in [70] . The thymi were transferred into pre-chilled T cell medium ( RPMI medium 1640 ( ATCC modification ) , 10% foetal bovine serum , 0 . 1x 2-mercaptoethanol ) and crushed to release total thymus T cell population . The cells were filtered through a 70 μm cell strainer , pelleted by centrifugation and re-suspended in Red Blood Cell lysis buffer ( Sigma ) for 3 min . Cell were then washed with the pre-chilled T-cell medium , re-suspended in the medium , filtered through a 70 μm cell strainer and counted using haemocytometer . RNA was isolated from the primary keratinocytes plated on the collagen solution coated dishes for 15 hours at 32°C and 8% CO2 or primary thymocytes using TRI Reagent solution and TURBO-DNA-free kit ( Invitrogen ) . Total RNA was amplified with Arcturus Ribo-Amp PLUS system ( Applied Biosystems ) as previously described [36] . RNA was converted into labelled cDNA and micro-array analysis was performed by MoGene ( St Louis , MO , USA ) using 41K Whole Mouse Genome 60-mer oligo micro-arrays ( Aglinent Technologies ) . Micro-array datasets were analysed using the distribution of background intensity and signal intensity values ( Agilent Feature Extraction software version 7 . 5 ) . Two 3C templates were constructed for freshly plated epidermal keratinocytes and primary thymocytes according to [71] with modifications . Briefly , epidermal keratinocytes isolated from mouse epidermis were seeded in the low calcium primary keratinocyte medium at high density on the collagen coated plate for 15 hours at 32°C , 8% CO2 and 90% humidity . The primary thymocytes were isolated as described above . Cells were washed twice with the growth medium and fixed with 1% formaldehyde ( Electron Microscopy Systems ) in the growth medium for 10 minutes at room temperature with gentle mixing every 2 minutes . The glycine was added to a final concentration of 125 mM . Quenching was initiated at room temperature and the cells were placed on ice for 5 min . The medium was removed and cells were washed ones with ice cold PBS and then fresh ice cold PBS was added . Cross-linked cells were collected , counted , pelleted by centrifugation in aliquots and quick-frozen . Cells were stored at -80°C . Per a 5C library , the frozen pallet of 6x107 cells 1 . 2 ml of lysis buffer ( 10 mM Tris-HCl , pH 8 . 0 , 10 mM sodium chloride , 0 . 2% ( vol/vol ) Igepal C-630 ( Sigma ) ) supplemented with 120 ul of protease inhibitor cocktail ( Sigma ) was added and cells were incubated on ice for 30 minutes . Cells were lysed using a 5 ml dounce homogenizer , washed twice with ice cold 1x NEBuffer2 buffer ( 10 mM Tris-HCl , 50 mM NaCl , 10 mM MgCl2 , 1 mM DTT ) , and re-suspended in 630 ul of 1x NEBuffer2 . Nuclear suspension was divided into 50 ul aliquots . To the nuclear suspension 312 ul of 1xNEBuffer2 was added . SDS was then added to a final concentration of 0 . 1% and lysates were incubated at 65°C for 10 min . Triton X-100 was then added to a final concentration of 1% to quench SDS . To each aliquot of solubilized chromatin 800 U of HindIII enzyme ( New England Biolabs ) was added and the digestion was performed overnight at 37°C with shaking . HindIII was inactivated by incubating lysates at 65°C for 30 min after addition of SDS to a final concentration of 1 . 56% . Ligation was performed under diluted conditions that promote intra-molecular ligation at 16°C for 4 hr in ligation buffer ( 1% Triton X-100 , 0 . 1 mg/ml BSA , 1 mM ATP , 50 mM Tris-HCl ( pH 7 . 5] ) 10 mM MgCl2 , 10 mM DTT ) with 10 ul of T4 DNA ligase ( Invitrogen ) . To reverse crosslinks , samples were then treated with 63 . 5 mg/ml Proteinase K ( Invitrogen ) at 65°C . Four hours later , Proteinase K was added again to 127 mg/ml and then incubated overnight at 65°C . DNA was purified by subjecting samples to a series of phenol and phenol-chloroform extractions before precipitation with ethanol . Pellets were re-suspended in 1–2 ml TE Buffer , pH8 . 0 and precipitation with ethanol . Pellets were re-suspended in 500 ul of TE buffer and treated with DNase-free RNase at final concentration of 100 ng/ul for 1 hour at 37°C . 3C templates were further purified using Amicon Ultra Centrifugal 30K Filter for DNA Purification and Concentration ( Millipore ) . Using the Millipore columns , samples were washed twice with 1X TE buffer . Following sample recovery from Millipore columns , initial sample volume was then restored with 1X TE buffer , pH 8 . 0 . The concentration of the 3C template was assessed by gel electrophoresis with high molecular weight DNA ladder as a standard ( Invitrogen ) using TotalLab Quant gel densitometry software . Controls for DNA integrity ( undigested chromatin control ) and restriction digestion ( no ligase control ) were also checked and passed the quality control . The quality of the 3C templates were further assessed by running the PCR with series of 2-fold dilutions of the templates with forward ( ATGGAGACCTGCCGCCGGCTCATCACAC ) and reverse ( CGTGCTGTGACTTCGCACTTTTCTGATC ) primers amplifying the product of head to head ligation of two HindIII sites located 1164bp apart as described in [71] using Quant gel densitometry software ( Total lab ) . Two independent 5C libraries were constructed for each cell type as described in [71] with modifications . 5C probes were designed at HindIII restriction sites using the my5Csuite primer design tools [44] . An alternating scheme was pursued in which reverse and forward probes were designed against every other fragment . Probes were excluded if unique mapping could not be achieved for fragments spanning highly repetitive sequences . Probe setting were as follows: U-BLAST , 3; S-BLAST , 50; MER , 800; MIN , FRAGSIZE , 100; MAX FRAGSIZE , 50000; OPT_TM , 65: and OPT_PSIZE , 40 . The universal T7 sequence was tethered to all forward primers ( TAATACGACTCACTATAGCC ) and the reverse complement to the universal T3 sequence was tethered to all reverse probes ( TATTAACCCTCACTAAAGGGA ) . In total , 381 forward probes and 382 reverse probes were designed , spanning 5 . 3 Mb EDC containing locus ( Fig 1B , S2 Table ) . To construct 5C libraries , first probes were annealed to the 3C templates at 48°C for 16 hours . Each multiplex annealing reaction contained 1xNEBuffer4 ( New England Biolabs ) , 560 ng of 3C template and 0 . 4 fmole of each 5C probe . The annealed probes were nick ligated with 10 U of Taq ligase in 1x Taq ligase buffer ( New England Biolabs ) at 48°C for 1 hour . The resulting 5C library was amplified by PCR with 25 cycles using universal T7 ( TAATACGACTCACTATAGCC ) and T3 ( TATTAACCCTCACTAAAGGGA ) primers . 15 ligation reactions amplified in 6 PCR reactions each were performed to generate each 5C library . The PCR reactions for each 5C library were pooled before further processing . 5C library amplification reactions produced the products of expected size ( 101 bp ) , while the negative control PCR reactions ( included no 5C template control , no ligation control or no 5C probe control ) did not yield any PCR product . 5C libraries were size fractioned ( 101 bp ) and purified from the agarose gel using QIAquick gel purification kit ( QIagene ) . 3’ A-tails were added using dATP and Taq polymerase , followed by subsequent ligation to bar-coded custom designed adaptor oligonucleotide [72] for Illumina pair-end sequencing . Adaptor-modified 5C libraries were purified after separation in the agarose gel using QIAquick gel purification kit ( Qiagene ) . The purified libraries were amplified by 18 cycles of PCR with PE1 . 0 and PE2 . 0 primers ( Illumina ) . The amplified libraries ( 233 bp ) were purified from the agarose gel , quantified using Nanodrop 1000 spectrophotometer ( Thermo Fisher ) and send for the sequencing on the HiSeq 2000 system at the EMBL Genome Core Facility ( Heidelberg , Germany ) . The 5C library sequencing data sets were de-multiplexed using Novobarcode ( Novocraft ) . The reads were aligned to the pseudo-genome consisting of all 5C probes ( S2 Table ) using Bowtie [73] . To account for poor quality reads , sequences were required to have only one unique alignment . After mapping , interactions were counted when both paired end reads could be uniquely mapped to the 5C probe pseudo-genome . Only interactions between forward-reverse probe pairs were considered as true counts . Next , we performed the data correction to remove the technical biases associated with the 5C technology as described in [17 , 43] with some modifications . First , we removed the probes that performed significantly differently in comparison to the overall probe sets . A global average relationship between interaction frequency and genomic distance was calculated using Loess smoothing for each replicate dataset . Contact profile for each probe across the interrogated region was compared to this average . We removed the probes with the individual Loess of more or less than 0 . 85 of the scaled Z score distance from the global Loess . We removed 38 probes for the replicate 1 and 37 probes for the replicate 2 for the downstream analysis ( S15 Table , S2B , S3B , S4B and S5B Figs ) . After this step , we removed the signal interaction with very high contact frequency in comparison to their neighbors . We removed such interactions if they have a Z score of 25 or more ( S16 Table , S2C , S3C , S4C and S5C Figs ) . Z score was calculated as described in [17] . Finally , we normalized the profile of each probe so they could be quantitatively compared to each other as described in [17] , but we calculated a global average relations between interaction frequency and genomic distance with Loess smoothing for each replicate separately ( S2D , S3D , S4D and S5D Figs ) . TAD boundary positions were identified by calculating an insulation score along the locus as described in [17 , 46] . The normalized 5C data were binned at 150 kb with 15 kb step size . Next , we calculated the combined number of interactions across each bin by summing all interactions up to 500 kb upstream of the bin and up to 500 kb downstream of the bin . The sum for each bin was divided by the average sum for all bins to yield insulation score . The insulation score was plotted along the whole locus to obtain an insulation profiles ( Fig 2D , S6A Fig ) . Local minima in these profiles indicate the position of the TAD boundaries . The local minima in the insulation profile were detected by identifying the bins with the lowest insulation score in a local 435 kb window . The mid-points of these bins were set as the TAD boundaries . The average position of the midpoint between the replicate was used as the TAD boundaries in the manuscript ( S3 Table ) . To detect the “true” statistically significant chromatin looping interactions between the individual restriction fragments , we applied a “5 C peak calling approach as described before [17 , 43] . We called the significant 5C peaks for the 5C libraries separately . Peaks were defined as normalized ligation frequencies ( signals ) that are significantly higher than expected for the genomic distances separating the interacting fragments . Expected values were calculated as the average interaction frequency for each genomic distance by using Loess smoothing ( alpha value 0 . 01 ) . This provides a weighted average and a weighted standard deviation for each genomic distance . We assumed that the large majority of interactions were not significant looping contacts , and we interpreted these weighted averages as the expected interaction frequencies for given genomic distances . We then transformed observed 5C interaction frequencies into a Z score by calculating the ( observed value-expected value ) /standard deviation . The calculated Z score distribution was fit to a Weibull distribution . p values were calculated for each Z score and transformed into q values for false discovery rate analysis . We used q-value threshold of 0 . 05 for the 5C peak calling . Only 5C peaks reproducible in both replicates in KCs or TCs were used for subsequent analysis ( S6 and S7 Tables ) . 3D FISH analysis of the spatially preserved nuclei in the mouse skin tissue and freshly isolated primary keratinocytes [74] and thymocytes was performed as previously described with modifications [36 , 67 , 74 , 75] . Primary keratinocytes were seeded overnight on the collagen coated cover slips . The adherent cells were fixed with formaldehyde and prepared for 3D FISH as described in [75] . Primary thymocytes were seeded on the slides coated with 1 mg/ml of Poly-L-lysine hydrobromide per [75] . 20 μm sections of the frozen skin sample with structurally preserved nuclei were used for the 3D analysis . BAC based probes were prepared for the selected regions ( S4 Table ) by nick-translation using in house synthesized Bio-dUTP , FITC-dUTP , Cy3-dUTP or Dig-dUTP as described in [76] . After hybridization the samples were stained with Cy5-streptavidin or anti-Dig-Cy3 antibodies ( S17 Table ) when needed . DNA was stained with DAPI ( Sigma ) . 3D images were collected using a Zeiss LSM510 confocal microscope . Nuclei were scanned with a z-axial distance of 200 nm , yielding separate stacks of 8-bit grey scale images , with pixel size 100–200 nm , for each fluorescent channel . For each optical section , images were collected sequentially for all fluorophores and the axial chromatic shift corrected for in each channel as described in [77] . Images were processed and analyzed using ImageJ ( NIH ) . Inter-locus distances were calculated after correction for chromatic aberration , as previously described [36] . The differences between the inter-locus distances in different samples were analyzed using Mann-Whitney U-test . For ChIP analysis new born C57Bl6 total keratinocyte single cell suspension was prepared as described above . To ensure analysis of viable cells with intact chromatin , keratinocytes were stained with UV Live/Dead Fixable Dye ( Life Technologies ) for 30 min on ice prior fixation with 1% PFA for 10 min at RT . Fixed cells were labeled with CD49f-PE and Sca-1-FITC antibodies ( S17 Table ) for 1 hour on ice . CD49f+/Sca-1+ basal keratinocytes were gated after exclusion of dead ( UV Live/Dead Fixable Dye-positive , Life Technologies ) cells and sorted on a MoFlo XDP cell sorter ( Beckman Coulter ) , as described in [78] . Sorted cells were pelleted at 2 . 000 g and stored at -80°C . ChIP was performed using FACS sorted epidermal keratinocytes isolated from newborn mouse skin anti-H3K27ac , anti-Rad21 and anti-CTCF antibodies ( S17 Table ) using ChIP-IT High Sensitivity kit ( Active Motif ) as described in [36 , 79] . ChIP with anti-H3K4me1 antibodies ( S17 Table ) was performed using Micrococcal nuclease ( MNase ) digestion epidermal keratinocyte chromatin as per [80] . 1x106 cells were used per MNase digestion and 1μg of the antibodies per IP comprising pre-cleared chromatin corresponding to 5x105 cells . Indexed ChIP-Seq libraries from immune-precipitated and control input chromatin were generated using NEBNext ChIP-Seq Library Prep Master Mix Set ( New England BioLabs ) for Illumina and NEBNext Multiplex Oligos for Illumina ( New England BioLabs ) . The libraries were sequenced on the HiSeq 2500 system ( Illumina ) , producing 30–70 million reads per library . Sequencing reads were aligned to the mm9 mouse genome assembly [73] . Specific areas of protein binding or histone modification presence were identified with MACS using default parameters [81] . The normalized ChIP-seq signals together with the previously published ChIP-seq signals for Brg1 were visualized using UCSC genome browser ( http://genome . ucsc . edu ) [82 , 83] . High confidence H3K4Me1 ChIP-seq peaks were merged if they were located within 5 kb end-to-end distances from each other and the same operation was applied to the H3K27ac ChIP-seq peaks . Enhancers were defined as merged H3K4me1 and H3K27ac peaks located within 2 kb end-to-end distance . We did not exclude putative enhancers located near gene promoters , as recent studies indicate that gene promoters could poses gene enhancer activity and enhancers could be located close to gene promoters [58 , 61] . The positions of the gene enhancers were visualized using UCSC genome browser ( http://genome . ucsc . edu ) [82 , 83] . An enrichment of selected non-histone protein binding in the regions anchoring 5C interactions at EDC locus or interactions involving gene enhancers at the locus were calculated for the extended 5C fragments to include nearest adjacent fragments interrogated by the 5C probes on the opposite strand as described in [11] .
|
Gene activity programmes in different cell types control development and homeostasis of multi-cellular organisms . Spatial genome organization controls gene activity by facilitating or restricting contacts between gene promoters and remote gene enhancers . Functionally related co-regulated genes are often located together in genomes loci . The spatial organization of very large co-regulated gene loci remains poorly understood . We analyse the spatial contact network in the Epidermal Differentiation Complex ( EDC ) locus that contains 61 co-regulated genes activated during epidermal differentiation in epidermal cells and thymocytes , where the locus is mostly inactive . Our analysis demonstrated that the gene-rich and gene-poor regions in the EDC are organized in separate Topologically Associating Domains ( TADs ) . We further found that spatial contact in the EDC locus is mostly cell type specific . In keratinocytes such contacts connect gene promoters with gene enhancers both within and between gene-rich TADs . Chromatin architectural proteins CTCF and Rad21 together with chromatin remodeller Brg1 were often bound near the spatially contacting gene promoters and enhancers in keratinocytes . In contrast to gene-rich TADs , gene-poor TADs show preferential spatial contacts with each other , do not contain active enhancers and show decreased binding of CTCF , Rad21 and Brg1 . These data illustrate how the chromatin networks required for lineage-specific transcription are organized in skin epithelial cells and demonstrate that spatial interactions involving gene promoters and enhancers at the EDC locus are not restricted by the TAD boundaries and involve , together with intra-TAD interactions , the extensive contacts between the different TADs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"keratinocytes",
"medicine",
"and",
"health",
"sciences",
"dna-binding",
"proteins",
"epithelial",
"cells",
"dna",
"transcription",
"stem",
"cells",
"genome",
"analysis",
"epigenetics",
"mammalian",
"genomics",
"chromatin",
"genomic",
"libraries",
"animal",
"cells",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"biological",
"tissue",
"hematopoietic",
"progenitor",
"cells",
"genetic",
"loci",
"animal",
"genomics",
"thymocytes",
"biochemistry",
"cell",
"biology",
"anatomy",
"genetics",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"genomics",
"computational",
"biology"
] |
2017
|
5C analysis of the Epidermal Differentiation Complex locus reveals distinct chromatin interaction networks between gene-rich and gene-poor TADs in skin epithelial cells
|
Inter-individual variation in regulatory circuits controlling gene expression is a powerful source of functional information . The study of associations among genetic variants and gene expression provides important insights about cell circuitry but cannot specify whether and when potential variants dynamically alter their genetic effect during the course of response . Here we develop a computational procedure that captures temporal changes in genetic effects , and apply it to analyze transcription during inhibition of the TOR signaling pathway in segregating yeast cells . We found a high-order coordination of gene modules: sets of genes co-associated with the same genetic variant and sharing a common temporal genetic effect pattern . The temporal genetic effects of some modules represented a single state-transitioning pattern; for example , at 10–30 minutes following stimulation , genetic effects in the phosphate utilization module attained a characteristic transition to a new steady state . In contrast , another module showed an impulse pattern of genetic effects; for example , in the poor nitrogen sources utilization module , a spike up of a genetic effect at 10–20 minutes following stimulation reflected inter-individual variation in the timing ( rather than magnitude ) of response . Our analysis suggests that the same mechanism typically leads to both inter-individual variation and the temporal genetic effect pattern in a module . Our methodology provides a quantitative genetic approach to studying the molecular mechanisms that shape dynamic changes in transcriptional responses .
Inherited variation in gene expression is likely to have a major effect on cellular and disease phenotypes , and may allow the underlying DNA polymorphisms ( genetic variants ) to be identified [1] . The genetic effect of a particular variant on a certain RNA is the quantitative change in gene expression that is associated with changing the variant's genotype ( allele ) . Two recent studies have demonstrated that genetic effects on longitudinal gene expression data might be either stable – where the genetic effect is similar at all time points ( a non-dynamic effect pattern; Fig . 1A ) – or flexible , changing the magnitude of effect during time points ( a dynamic effect pattern; Fig . 1B , C ) [2] , [3] . Dynamic effect patterns may be described in terms of the shape of changes in genetic effects over time . A linear-like genetic effect pattern ( Fig . 1B ) reflects a gradual change in the magnitude of genetic effects , whereas in a non-linear genetic effect pattern ( Fig . 1C ) , the level of genetic effect is sustained in some time periods and spikes up or down in others ( Fig . 1C ) . In most studies , transcription responses across individuals have been monitored only in two time points ( before and after stimulation ) and therefore the dynamics of changes in genetic effects over time could not be characterized [4]–[9] . Understanding non-linear genetic effects can , in principle , allow the timing of influence of certain regulatory mechanisms to be revealed . For example , a single state-transitioning in genetic effects may uncover the timing of alteration in a regulatory mechanism interacting with a genetic variant ( e . g . , transition to a new steady state at t3 , Fig . 1C , left ) . Such a mechanism can be revealed even when additional mechanisms are acting in parallel ( e . g . , up-regulation during the entire time course; Fig . 1C , left ) . The linear genetic effect pattern , in contrast , lacks sharp alterations and therefore does not specify finely-timed information about regulatory mechanisms ( Fig . 1B ) . This study is focused on mapping temporal patterns of non-linear genetic effects and using this information to address major questions about dynamic transcription responses . Which dynamic genetic effect patterns are prevalent in global gene responses ? Are there any general principles - either functional or mechanistic - shared among genes carrying the same temporal genetic effect patterns ? Can we derive insights about the mechanisms underlying such dynamic genetic effect patterns ? Here we developed DyVER ( Dynamic Variant Effect on Response ) , a statistical framework to predict genetic variants and study their dynamic changes in genetic effect sizes . DyVER was mainly designed to achieve an accurate detection of non-linear genetic effects ( Fig . 1C ) during time points . The methodology is based on the notion of a two-state digital model that pinpoints the particular time point at which a rapid change in genetic effects occurs; it is therefore suitable for revealing the timing of state transitions in genetic effects . DyVER takes as input synchronous data in several time points and across a population , and is tailored for recombinant inbred strains that are commonly utilized in genetic studies [2] , [10]–[14] . DyVER differs from extant genetic approaches in several aspects . First , some existing methods construct a full model of the response curve across individuals . Their number of parameters is therefore increasing with the number of time points ( e . g . , [15] ) . DyVER , in contrast , is primarily designed for the specific task of identifying the time points of alterations in effect sizes . This partial modeling allows the use of only a small number of parameters regardless the number of time points and the shape of the temporal pattern . Secondly , DyVER is focused on modeling the dynamics in genetic effects while eliminating the confounding gene expression variables . This is unlike extant approaches , which commonly fit both gene expression and genetic effects to a certain function over time [11] , [15]–[20] . Finally , if desired , DyVER can exploit the order in the input time course data , unlike several approaches that are based on unordered correlated traits ( e . g . , multivariate methods [21] , [22] or dimension reduction methods [23] ) . Notably , DyVER is a practical translation of differential expression approaches ( with or without time-series data [24]–[26] ) for the case of statistical genetic studies . Here we report on the use of DyVER to investigate temporal gene responses at six time points after stimulation with the TOR inhibitor rapamycin and across genotyped yeast segregants [27] . The results depict a complex map of non-linear changes in genetic effects . We identify a causal variant that affects the timing of spike up in transcript levels . Importantly , our findings suggest a previously unknown high-order temporal coordination of genetic effects: modules of genes influenced by a common dynamic genetic variant not only participated in the same biological pathway , but also shared orchestrated dynamics of genetic effects . Based on this modularity , we hypothesize that in some cases dynamic effect patterns are a property of the regulatory mechanism within which a genetic variant resides ( rather than a property of the target responding transcript ) . We demonstrate that using this notion it is possible to enhance the identification of underlying causal genes based on their characteristic temporal effect pattern . Our results indicate the utility of studying dynamic genetic effects acting on global gene transcription .
We devised a new method , DyVER , to identify genetic variants that underlie the expression of genes and their particular dynamic effect patterns . DyVER takes as input the measured transcription response of a gene over several consecutive time points following stimulation and across a cohort , as well as a set of potential genetic variants and their genotyping ( Fig . 2A ) . Given a candidate genetic variant with two alternative alleles , DyVER proceeds in three steps ( Methods ) : ( 1 ) It first calculates the observed effect of the variant , namely the difference in gene response between strains carrying the two distinct alleles ( Fig . 2B ) . The observed genetic effects are used as data in the subsequent steps . ( 2 ) To identify non-linear dynamic shapes of genetic effects , DyVER assumes a ‘digital’ regulatory model that distinguishes two possible states of genetic effects: first , a strong effect of genetic variant on the gene response ( denoted the high-effect state ) ; and second , a lower ( such as zero ) effect , or possibly an opposite effect ( denoted the low-effect state ) . Several previous methods have employed a two-state model , although not in a dynamic or a genetic effect context [28] . Based on a maximum likelihood approach , DyVER seeks a genetic variant and a sequence of states that best describe the dynamic changes in the size of the genetic effect . For example , if a gene is affected mainly by a variant v during a late time interval , DyVER successfully infers the correct effect pattern low→low→high→high for the correct variant v as it attains the highest likelihood score ( Fig . 2B and C , right panel ) . For incorrect variants , the likelihood scores are typically lower ( Fig . 2B and C , left panel ) . DyVER's predicted sequence of states is referred to as the temporal two-state model . Finally , ( 3 ) DyVER calculates the statistical significance of association for each genetic variant based on a likelihood ratio score that takes as input the inferred temporal two-state model ( Fig . 2D ) . We refer to this score as the DyVER score . Notably , although DyVER requires synchronous observations in particular time points , it is still possible to apply DyVER on partial observations in each of the time points ( Methods ) . Overall , step 1 allows DyVER to focus on dynamics in genetic effects regardless of the magnitude of transcription response , whereas the discrete modeling in step 2 allows detecting any sequence of spikes up or down in genetic effects . The two-state model from step 2 enhances the performance of the DyVER score ( step 3 ) by allowing a separate parameterization for each of the states . Specifically , to infer an optimal temporal two-state model , DyVER uses a two-state hidden Markov process where the observed effects are treated as the outcome of a sequence of hidden high-effect and low-effect states ( step 2; Fig . 2C ) . The corresponding likelihood function consists of two components: ( i ) the probability of observed effects given a certain temporal two-state model; and ( ii ) the probability of a temporal two-state model , which may use a penalty factor to prioritize two-state models with a lower number of transitions between states , assuming dependencies among consecutive time points . In the absence of penalty , the order of time points is irrelevant and therefore the predicted two-state model can be viewed as a partition of an unordered group of time points into two sub-groups . The DyVER score exploits this partition for a different parameterization of the ( unordered ) time points in each of the two states . The addition of the penalty factor makes it possible to avoid an overfitted two-state model that is then given as input to the next step , hence further improving the DyVER score's performance . We compared DyVER's performance to that of five alternative methods . In the first method , the most naïve approach , an ANOVA test is applied at each time point independently and the predicted genetic variant is the one with the most significant ( minimal ) ANOVA P value score . The second method builds on dimension reduction using principal component analysis ( PCA ) : Given T time points for each strain as input , it first reduces the T-dimensionality of the data into a single dimension by projecting each strain onto the first principal component . Next , it applies an ANOVA test on this one-dimensional data [23] . The third method models dynamics in gene expression as well as dynamics in genetic effect sizes [15] . For comparison , in the fourth method , a linear time term is included as a covariate in the ANOVA test to model dynamic changes in gene expression ( without direct modeling of dynamics in genetic effects ) . Finally , we compared DyVER to a random prediction of association relationships . We called these approaches ‘naïve’ , ‘PCA’ , ‘detailed dynamics’ , ‘expression dynamics’ and ‘random’ , respectively . In both DyVER and all compared methods , for each simulated gene , the resulting P values were Bonferroni-corrected for the testing of multiple genetic variants . The quality of predicted variants were evaluated using the accuracy metric , defined as the tradeoff between the sensitivity and specificity of revealing genetic variants across different significance cutoffs . The accuracy metric ranges between 0 and 1 for poor and excellent performance , respectively ( Methods ) . To characterize DyVER's ability to reveal dynamic genetic variants and distinguish their effect patterns , we generated synthetic collections of genes that are associated with genetic variants over time . A single synthetic ‘collection’ consisted of 500 genes , 300 of them associated with a genetic variant over time , with two characteristic parameters: ( i ) the number of time points , and ( ii ) the effect size ( in all cases we used 50 strains and 100 genetic variants ) . In a complete synthetic ‘dataset’ we generated 72 collections for various numbers of time points and effect size values . Overall , four synthetic datasets were generated in this study , each consisting of a different key class of dynamic effect patterns ( see Methods ) : a linear-like pattern ( Fig . 1B ) , a single state-transitioning based on a sigmoid function ( Fig . 1C , left ) , and impulse and multiple-pulse ( complex ) patterns based on the product of two sigmoid functions ( Fig . 1C , middle and right , respectively ) [24] . In the following , we first analyze the performance of the DyVER's predicted associations ( based on the DyVER score ) in the absence of penalty and then present the contribution of the penalty factor . DyVER showed good accuracy in all non-linear dynamic effect patterns ( 0 . 5 penalty; Fig . 3 ) . Fig . 3A presents the accuracy metric for synthetic datasets of varying numbers of time points . Accuracy values are averaged across the eight collections of distinct effect size . In all non-linear dynamic effect patterns , DyVER displayed the best accuracy in all tested time points ranging between 3 and 27 , with improved accuracy for a larger number of time points . Importantly , although DyVER was not designed for linear-like effect patterns , it still attains the second-best performance for this case . The ‘expression dynamics’ approach yielded the most accurate predictions for the linear case , but attained poor results in the non-linear case . The tradeoff between sensitivity and specificity in the accuracy measure across the different methods is further demonstrated in Figure S1A and B . Results were similar for varying effect sizes ( Fig . 3B and Figure S1C and S1D ) and for an additional synthetic dataset that is based on prototypical effects in C . elegans ( Methods; Figure S2 ) . Furthermore , although DyVER's accuracy is reduced in the case of missing data , it is still notably high in comparison to alternative methods ( Figure S3 ) . Taken together , our results indicated that DyVER performs well on a broad range of genetic effect patterns . We next aimed to characterize DyVER's applicability to short-term steady state of high genetic effects . To tackle this goal we compared two synthetic impulse datasets , both consisting of 27 time points across various effect sizes . For all genes , the short-impulse dataset consisted of a high-effect steady state of short duration ( five time points ) , whereas the long-impulse dataset consisted of a high-effect steady state of long duration ( fifteen time points ) . Figure S4A records the performance of DyVER compared to the five alternative methods on the short-impulse and the long-impulse datasets , and clearly shows that DyVER outperformed the alternative procedures when genetic influences were acting in short impulses , even with low-effect sizes . The performance of both DyVER and the alternative methods declined when applied on a short impulse compared to a long impulse of genetic effects , but notably , the performance reduction was lowest with DyVER ( Figure S4B ) . For example , for high-effect sizes ( 0 . 625 ) , the sensitivity of DyVER is 0 . 7 and 1 with short and long impulses , respectively . The sensitivity of PCA , in contrast , is respectively 0 . 47 and 1 with short and long impulses for the same effect size . Thus , even when genetic variants acted during short time intervals , DyVER still performed relatively well . This was unlike the alternative methods , whose performances were drastically reduced even for relatively high-effect sizes . DyVER predicts a temporal two-state model , which may provide insights concerning the timing of changes in genetic effects ( Fig . 2C ) . To evaluate the quality of this prediction , we compared the ‘ground truth’ ( simulated ) models against the inferred two-state models . We chose to work with the established error rate statistics , defined as the number of erroneous two-state models expressed as a fraction of the total number of significant correctly predicted variants . We called this metric a two-state pattern error rate ( in short , error rate ) , and calculated it both for the case of stringent ( exact ) matching or flexible ( non-exact ) matching between the true and inferred models ( Methods ) . In both cases , we found that DyVER performs well in predicting two-state models , where the flexible case outperforms the stringent case , as expected . For example , using single state-transitioning patterns with nine time points , effect size 0 . 75 , significance cutoff 0 . 001 and the absence of penalty ( probability of transition 0 . 5 ) , the stringent and flexible error rates are 0 . 41 and 0 . 33 , respectively ( Figure S5 ) . The error rate increased with decreasing penalty ( e . g . , for transition probabilities of 0 . 01 ( high penalty ) and 0 . 5 ( no penalty ) , stringent error rates are 0 . 32 and 0 . 41 , respectively ) . As expected , error rates rose when a higher statistical significance cutoff ( 0 . 05 ) was used , whereas the gap between the error rates for different significance cutoffs remained relatively constant when the penalty increased . Results obtained for other effect sizes were similar . Collectively , our results indicated that DyVER outperforms extant methods even in the absence of penalty and the presence of missing data ( Fig . 3 , Figures S1–S4 ) , and that these performance can be even enhanced by the addition of a penalty component ( Figure S5 ) . These results hold when the complexity of dynamic effect patterns is relatively low , as in the case of genetic effects in biological data ( e . g . , Figure S6 ) . We applied DyVER in an unbiased manner ( without penalty ) to the available dataset of 95 yeast segregants that were stimulated by rapamycin and profiled at six time points ( Methods ) [27] . DyVER predicted 351 associations to 145 distinct variants ( false discovery rate [FDR] 6% ) . Of these 351 associations , 145 had highly significant dynamic associations ( 15% FDR , Table S1 , Methods ) and 105 of them showed non-linear genetic effect patterns ( Fig . 4 ) . In agreement with previous findings [2] , [11] , our results suggest that non-linear associations are prevalent: of the eight previously known causal genes , six were found to have an association with at least one target gene exhibiting a non-linear genetic effect pattern ( Table S2 ) . Correlations among genetic effects of consecutive time points were much larger than correlations between non-consecutive time points [P value <10−15 ( Wilcoxon test ) ] , justifying our ‘memoryless’ Markov assumption that the next time point is mainly dependent on the current time point ( Figure S7 ) . The 105 genes carrying non-linear effect patterns were partitioned into groups based on their predicted two-state pattern ( Table S1 ) ; seven two-state pattern groups ( C1–C7 ) were created , each including at least two genes ( Fig . 4A and B ) . The partition revealed three prototypical non-linear genetic effect patterns ( Fig . 4A ) , including ( i ) a single upward spike followed by a sustained high level of genetic effect ( 70 genes in C1–C4 ) . These different groups were characterized by distinct timing of a state-transitioning , including an abrupt change in early time points ( 0–10 min , C1 ) , as well as an intermediate-early ( 0–20 min , C2 ) and intermediate-late ( 20–30 min , C3 ) single state-transitioning . For example , SFA1 and ESF1 ( in groups C1 , C2 ) demonstrate a sustained genetic effect with a state transition at 0–10 and 0–20 minutes after rapamycin stimulation , respectively ( Fig . 4B ) . In the case of the four genes exhibiting a late state-transitioning ( at 30–50 min , C4 ) , a sustained new level of genetic effects might occur at later time points that were not measured in the current dataset [27] . ( ii ) A single downward spike of genetic effect ( C5–C6 , 22 genes ) . In group C5 , we observe an abrupt downward spike in 10–20 minutes followed by a sustained low level of genetic effect ( for example , PHM6 , Fig . 4B ) . Group C6 represents a delayed gradual single state-transitioning during 20–50 minutes . ( iii ) An impulse of high genetic effect at 10–30 minutes after treatment ( 9 genes in C7 , e . g . , UGA4 , Fig . 4B ) . Overall , the single state-transitioning patterns were over-represented , whereas complex patterns of genetic effects were rare ( 1 gene , YER053C-A ) and were under-represented [cis: P value <10−19 , trans: P value <10−50 ( t-test ) , ( Figure S6A ) ] . Our findings of rare complex patterns in yeast parallel similar observations in the mouse ( Methods , Figure S6B ) ; Yet , the particular shape of effect patterns may differ between biological systems ( Figure S8 ) . We next explored the pleiotropic trans-acting variants that arise from this analysis . Using DyVER's predictions we organized the genes into six co-association modules , each containing a group of ( at least two ) genes with the same trans-associated variant ( Fig . 5A and B ) . Functional enrichment strongly related all six modules with specific biochemical pathways . For example , the entire module no . 3 consists of genes that play a role in uptake of phosphate ( Pi ) from extracellular sources and its accumulation in vacuoles ( 5 of 5 genes; Fig . 5A and B , Figure S9A ) . The module's validated causal gene is PHO84 , a high-affinity phosphate transporter that carries a missense mutation in one of the parental strains ( Figure S9B ) [29] , [30] . The nine genes in module no . 5 carry two distinct functionalities and are therefore treated as two distinct sub-modules , no . 5-I and no . 5-II ( three daughter cell-specific genes and six poor nitrogen source degradation genes , respectively , Fig . 5A ) . Next we examined whether module genes show characteristic temporal effect patterns . On analyzing the modules we found that modules nos . 1 , 3 , 4 , 5-I and 5-II relate to a specific prototypic temporal genetic effect pattern , whereas the remaining two modules ( nos . 2 and 6 ) are more general and show several distinct patterns ( Fig . 5A ) . For example , module no . 1 contains 34 genes , 32 of which have an upward spike ( a single state transition ) of genetic effect at 10–30 minutes after rapamycin stimulation [FDR 0 . 01 ( hyper-geometric test ) ] . As another example , module no . 3 contains five genes , all showing a downward spike of genetic effects at 10–30 minutes after stimulation . Specifically the downward spike occurs either 20–30 minutes after stimulation [4 genes , FDR 0 . 01 ( hyper-geometric test ) ] or 10–20 minutes after stimulation ( 1 gene , Fig . 5A–C , Figure S9C ) . Overall , we found four modules with over-represented patterns of single state-transitioning at specific time points ( nos . 1 , 3 , 4 and 5-I ) and one sub-module of an impulse effect pattern ( no . 5-II ) . The observed coordination of temporal genetic effects does not necessarily reflect a coordination of transcription responses ( Figure S10 ) . In previous reports , baseline expression levels were used to identify eight genetic variants underlying similar modules ( Table S2 ) , but the coordinated temporal genetic effects and the timing of upward or downward spikes of genetic effects were not characterized . A plausible explanation for the ‘shared variant , shared temporal genetic effect pattern’ hypothesis is that the same molecular mechanism underlies both inter-individual variation and the dynamics of genetic effects . In such cases , the dynamic pattern of effect is an attribute of the underlying regulatory mechanism ( rather than of the target genes ) , probably owing to temporal changes in the influence or activity of the regulatory mechanism . This hypothesis is further supported by the consistency in the timing of state transitions in module genes and their underlying ( known ) causal genes ( Figs . 5D versus 5E ) : The trans-associated causal gene of module no . 1 ( IRA2 ) attains a sustained-like pattern of gene expression that resembles the temporal genetic effect pattern of its target genes ( Fig . 5D and E , left ) . The cis-associated causal genes in modules nos . 3 and 4 ( PHO84 and GPA1 ) exhibit drastic changes in their transcription response at the same time point at which there is a ( downward or upward ) spike in the genetic effect of their target genes ( 20–30 and 30–40 min; Fig . 5D and E , middle and right , respectively ) . The poor nitrogen source degradation system ( module no . 5-II ) demonstrates the ability of our method to reveal novel associations acting on the timing of response and affecting an entire cellular pathway ( Figs . 5 , 6 ) . During growth on relatively poor nitrogen sources ( allantoate , allantoin , and GABA ) , yeast cells activate premeases responsible for uptake of nitrogen sources and further increase the expression of enzymes that participate in degradation of poor nitrogen sources for the generation of ammonia . Exposure to the TOR inhibitor rapamycin also leads to the same nitrogen-regulated response [31] . Module no . 5-II consists of six of the twelve genes in the allantoin , allantoate and GABA degradation pathways , with all six genes having a significant impulse effect pattern ( DAL1 , 2 , 4 , 7 , 80 and UGA4; Fig . 6A–C ) . An additional gene in these pathways , DAL5 , is weakly associated using the same impulse pattern at the same genomic position ( Fig . 6A–C ) . The impulse pattern reflects a difference in the timing of initiation of response among the strains carrying the RM and BY alleles in Chr2: 533–562 kb . For example , strains carrying the BY allele showed early up-regulation of DAL80 in response to rapamycin , which was already detected at 10 minutes after stimulation . The RM-carrying strains , in contrast , showed a clear delay in response to rapamycin , but all strains reached a similar expression level by 30 minutes after stimulation ( Fig . 6D ) . The underlying genetic variant acting on the timing rather than on the magnitude of response has not been previously documented . In the genomic interval ( Chr2: 533–562 kb ) , two genes ( RPB5 , CNS1 ) have temporal transcription profiles that match the expected early impulse of high genetic effect , the promoter of five genes ( RPB5 , CNS1 , ADH5 , RTC2 , YBR144C ) is bound by nitrogen-related transcription factors [32] , and four genes ( RPB5 , CNS1 , ADH5 , RTC2 ) were previously reported in nitrogen-related cellular processes ( Figure S11 ) . These criteria therefore suggest that RPB5 or CNS1 are two leading candidates in module 5-II .
In this work we present the DyVER computational algorithm for identifying genetic variants that lead to dynamic changes in genetic effects . DyVER was tailored to identify abrupt changes in the levels of genetic effects , which may provide valuable information about the timing of alterations in the particular regulatory mechanisms interacting with the underlying genetic variant . In comparison with other approaches , DyVER attained the most accurate identification of non-linear genetic effect patterns , even in the absence of penalty ( Fig . 3 , Figures S1–S4 ) , likely due to ( i ) a focus on genetic effects rather than on modeling the original phenotype values , and ( ii ) the prior knowledge about the separation of the time points into two distinct groups that differ in their observed effects ( encoded in the temporal two-state model ) , thus allowing a different parameterization for each of these groups . DyVER is using an HMM-based model for revealing genetic variants acting on time-series gene expression data . HMM modeling has been applied in various contexts , but not for the case of direct identification of underlying genetic variants . For example , HMM has been utilized for the identification of CNVs or haplotypes [33] , [34] . Alternatively , an existing method was mainly focused on revealing differential expression between conditions using an HMM approach [26] . DyVER extends this method by providing a statistical genetics P value score and by allowing a number of parameters that is not increasing with the number of strains . Our method opens multiple directions for future investigations . First , it is important to extend DyVER for the case of outbred heterozygous population , including human . In the current study , DyVER was designed for the case of a inbred ( homozygous ) strains that are common in genetic studies ( e . g . , in yeast , nematode , fly , mouse and rat ) due to several major advantages: first , inbred strain enable controlled stimulations , and second , they avoid major challenges that are common in human genetics , including haplotype analysis , rare variants and uncontrolled variables . Future extensions may generalize the method for the heterozygous case , possibly by calculating genetic effects between each pair of genotypes ( rather than between the only two possible genotypes as in the homozygous case ) , requiring to add additional one or two Gaussians within each of the model states . Second , the usage of a our probabilistic model leads to several limitations: the number of states should be specified in advance; we only capture correlations between sequential time points but cannot capture higher-order correlations among time points; and we generally assume that the probability of a time point is independent of the probabilities of its neighboring time points . Future improvements that handle more than two states and a more sophisticated probabilistic graphical model [35] may therefore enhance DyVER's performance . Third , DyVER relies on at least a few synchronized strains in each of the time points . Although DyVER allows missing data and possibly different strains in different time points ( Figure S3 ) , it still cannot be applied on non-synchronous data ( as in [11] ) . Data imputation methods can potentially enhance the DyVER analysis beyond this synchronization requirement . Building on the DyVER approach , we analyzed temporal gene expression patterns following rapamycin treatment in yeast segregants . Our analysis identified 105 genes exhibiting significant non-linear genetic effects over time , 56 of them are well-established associations ( in modules 1 , 2 , 3 , 4 , 5-I and 6 ) , and the remaining genes are new candidates for future experimental investigations ( e . g . , Fig . 4B ) . For example , our study suggests a novel genetic variant residing in chr2: 533–562 kb as the underlying regulator of the timing of upward spikes in gene expression after rapamycin treatment . Reassuringly , this regulator acts primarily on genes that play a role in poor nitrogen source degradation ( 6 of 6 genes , module 5-II , Fig . 6 ) . The application of DyVER in yeast provided several novel insights that were mainly attained due to the unique capability of DyVER to classify associations based on their optimized temporal effect patterns . First , we use the temporal effect pattern to automatically organize the genes into clusters based on their predicted patterns ( Fig . 4A and Figure S12 ) . This organization is substantially different from previous studies [2] , [11] that have grouped time-series associations only manually . Based on this clustering , we found that abrupt single state-transitioning and impulse patterns occur in certain prototypical time points . In particular , DyVER identified an upward spike of genetic effect at 0–10 , 0–20 , 20–30 and 30–50 minutes ( 22 , 34 , 10 and 4 genes , groups C1 , C2 , C3 and C4 , respectively ) ; a downward spike followed by a new sustained low level of genetic effect ( 6 and 16 genes at 10–20 and 20–50 minutes , groups C5 and C6 , respectively ) , and a single pulse of high genetic effects ( 9 genes , group C7 , Fig . 4 ) . Second , many studies have shown that groups of co-associated genes also share similar functionalities . Interestingly , our results indicate that such co-associated genes typically share not only a similar functionality , but also a similar predicted pattern of temporal genetic effect ( Fig . 5 ) . One plausible explanation is that a causal regulator typically alters its functionality during its response to stimulation; therefore , a genetic variant interacting with such a regulator is likely to affect its target only during those time intervals in which the regulator is functional . Based on this rationale , the temporal effect patterns in target genes may uncover the temporal dynamics of their causal regulatory mechanisms . Thus , DyVER's characterization of temporal effect patterns , which are probably a property of the causal regulatory mechanisms , may provide a starting point for improved identification of causal genes . For example , it might be possible to pinpoint a causal gene in a genomic interval based on its predicted dynamics over time ( as demonstrated in Fig . 5D and E and Figure S11C ) . Furthermore , it may be possible to discriminate between two genetic variants differing in their dynamic over time , even when these variants are co-localized at a nearby genomic position ( as in module nos . 5-I and 5-II , Fig . 5A ) . Taken together , our results highlight the utility of studying temporal genetic effect patterns to discover and characterize dynamic causal regulators . The next step is to extend and apply our approach to map genetic effects in transcriptome of a wide range of mammalian cell types .
To generate synthetic data we first generated 50 strains carrying 100 genetic variants , sampling one of the two alleles with equal probabilities . A single synthetic collection consists of 500 genes , of which 300 are associated with a certain variant over T time points . Overall , for a single dataset we generated 72 collections , constructed for all combinations of eight possible ‘effect sizes’ ( defined below , ranging between 0 . 125 and 1 ) and nine different numbers of time points ( ranging between 3 and 27 ) . In all cases , the low-effect state represents the absence of effect ( ) and the high-effect state represents the presence of an effect ( ) , where is the effect size . is the simulated observed effect , which is generated by sampling from a Gaussian distribution . A dataset was constructed for each class of temporal effect patterns . For a single state transition effect pattern ( here , sustained ) we used a sigmoid function: Where , q = v = 0 . 5 and . For an impulse effect pattern we used the product of two sigmoid function with five parameters [24] , where , is the length of the impulse effect ( here , ) : To generate the complex pattern for T time points we concatenated two impulse patterns , each for T/2 time points . For the dataset of linear effect patterns , observed effects are sampled from a linear function: For the purpose of comparing predicted to gold-standard temporal two-state models ( Figure S5 ) we generated a different collection of synthetic sustained dataset as follows: we first generated the temporal two-state model by sampling from the corresponding distribution ( from equation 3 ) with . The observed effects were then generated by sampling from the corresponding Gaussian distribution , , where and are the mean of the high- and low-effect state , and . To generate an input with a percentage of k% missing data , in each time point , we omitted the information for k% randomly selected strains ( thus , each time point consists of a different list of strains ) . An additional synthetic dataset was created similarly to the above datasets , but using previously published functions in C . elegans [11] . For each of the 300 associated genes in this synthetic dataset , we first randomly chose a function out of the 18 functions that were published in C . elegans; the observed effects were then sampled from this selected function . The compared methods were implemented as follows . In the ‘naïve’ method we assumed a simple fixed effect model on each time point independently , , where is the observed expression level for strain j carrying genotype i; is fixed effect of genotype i and . The most significant ( minimal ) ANOVA P value score is taken as the resulting P value . In the ‘PCA’ method , we project the T-dimensionality of each strain into the first principal component and then applies an ANOVA test assuming a fixed effect model where is the first principal component for strain j carrying genotype i; is the fixed effect of genotype i and ( the first principle component was chosen since it performs better than the consecutive components , see Figure S15 ) . For the ‘expression dynamics’ method , we used the model where is the observed expression level in time point t for strain j carrying genotype i , and are two fixed effects for genotype i and . The formulation was implemented using the lme4 R package . In all cases above , an F-test was used to test the model . For the more sophisticated ‘detailed dynamics’ method , we use the longGWAS R package that is part of its original publication [15] . For each synthetic dataset , DyVER was applied to predict a genetic variant using the DyVER score ( P values were Bonferroni-corrected for multiple variants ) . To quantify the ability to correctly predict such genetic variants , we define the accuracy measure . Genes are split into two groups: one contains genes that are associated with a genetic variant , and the other contains the remaining , non-associated genes . A mapping method may provide a negative prediction ( i . e . , a non-significant P value for all candidate variants ) , or alternatively , a positive prediction of either the correct variant or an incorrect variant . We define true positives as associated genes whose correct genetic variant is predicted with a significant P value . True negatives are non-associated genes that were not significantly associated with any variant . False negatives are associated genes that were not significantly associated with any variant . Finally , false positives are defined as erroneous significant predictions as a result of two possible scenarios , either a non-associated gene that is wrongly predicted to be associated with a certain variant , or alternatively , an associated gene whose predicted variant is incorrect . We adopt the standard formulations for sensitivity ( number of true positives out of the total number of positives ) and specificity ( number of true negatives out of the total number of negatives ) . Similarly to a standard ‘Receiver Operating Characteristic’ ( ROC ) analysis , we can plot the sensitivity against the 1-speificity across different P values , providing an overall view of the performance of the method: the higher the curve , the better the accuracy ( defined as the area under the curve ) . Notably , using a standard sensitivity definition , sensitivity should increase with higher P value thresholds . In contrast , using our definition of sensitivity , it is dependent on the particular predicted variant . Thus , even with a very high P value threshold and many affected genes , the sensitivity of a random algorithm might remain close to zero . The accuracy therefore ranges between 0 ( for a random prediction ) and 1 ( for a perfect prediction ) . Finally , to quantify the ability of DyVER to correctly predict the temporal two-state model , we define the two-state pattern error rate ( shortened to error rate ) as the number of wrongly predicted temporal two-state models expressed as a proportion of the total number of ( significant ) correctly identified variants . We test two different rules for matching between the simulated and predicted model . In the stringent case , we require a fully correct two-state model , and in the flexible case , we require correct transitions between states but allow incorrect timing of transition . We applied DyVER to genotyping data and gene expression data that were monitored during six time points following exposure to rapamycin in 95 yeast segregants and their two parental yeast strains: BY4716 ( BY ) and RM11-1a ( RM ) [27] . DyVER was applied to the log expression of 2700 genes with the highest difference between the BY and RM parental strains . To ensure that the biological results are unbiased , DyVER was applied with penalty 0 . 5 . Multiple testing was controlled as follows: DyVER score P values were first Bonferroni-corrected for multiple variants; the corrected DyVER score P values were then controlled for multiple testing of genes ( FDR 6% ) . We then further filtered the genes based on the dynamic association score ( FDR 15% ) . In total out of 2700 genes , we obtained 351 ( 13% ) predicted associations ( based on the corrected DyVER score P value ) and 145 ( 5 . 3% ) predicted dynamic associations ( based on the dynamic association score ) . Next , we further removed 40 genes carrying linear-like patterns , based on strong correlation with a linear model ( r>0 . 95 ) and more than 5% change in genetic effect in any two consecutive time points ( Table S1 ) . The partition into groups was generated automatically according to DyVER's predicated two-state model ( Figure S12 ) . In addition , we applied DyVER to genotyping data and log gene expression data of 403 genes that were monitored using a meso-scaled technology during three time points following exposure to lipopolysaccaride in 45 mouse BXD strains [2] . Of the 403 genes , 14 genes ( 3 . 4% ) were identified as significant dynamic associations ( FDR 10%; Figure S6B ) .
|
Genetic variation is postulated to play a major role in transcriptional responses to stimulation . Such process involves two inter-related dynamic processes: first , the time-dependent changes in gene expression , and second , the time-dependent changes in genetic effects . Although the dynamics of gene expression has been extensively investigated , the dynamics of genetic effects yet remain poorly understood . Here we develop DyVER , a method that combines genotyping with time-series gene expression data to uncover the timing of transitions in the magnitude of genetic effects . We examine gene expression in yeast segregants during rapamycin response , finding several distinct ways of change in the magnitude of genetic effects over time . These include impulse-like and sustained transitions in genetic effects , acting both in cis and trans . Our findings suggest that associations of genes with the same genetic variant often occur via the same timing of state transition in genetic effects . Furthermore , the results uncover a previously unknown variant whose impulse-like temporal genetic effect suggests a novel molecular function for determining the timing rather than the magnitude of response . Our results show that steady-state association studies miss important genetic information , and demonstrate the power of DyVER to render a comprehensive map of dynamic changes in genetic effects .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"quantitative",
"trait",
"association",
"studies",
"genome",
"analysis",
"gene",
"expression",
"genetics",
"gene",
"regulation",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2014
|
Dissecting Dynamic Genetic Variation That Controls Temporal Gene Response in Yeast
|
We present a machine learning-based methodology capable of providing real-time ( “nowcast” ) and forecast estimates of influenza activity in the US by leveraging data from multiple data sources including: Google searches , Twitter microblogs , nearly real-time hospital visit records , and data from a participatory surveillance system . Our main contribution consists of combining multiple influenza-like illnesses ( ILI ) activity estimates , generated independently with each data source , into a single prediction of ILI utilizing machine learning ensemble approaches . Our methodology exploits the information in each data source and produces accurate weekly ILI predictions for up to four weeks ahead of the release of CDC’s ILI reports . We evaluate the predictive ability of our ensemble approach during the 2013–2014 ( retrospective ) and 2014–2015 ( live ) flu seasons for each of the four weekly time horizons . Our ensemble approach demonstrates several advantages: ( 1 ) our ensemble method’s predictions outperform every prediction using each data source independently , ( 2 ) our methodology can produce predictions one week ahead of GFT’s real-time estimates with comparable accuracy , and ( 3 ) our two and three week forecast estimates have comparable accuracy to real-time predictions using an autoregressive model . Moreover , our results show that considerable insight is gained from incorporating disparate data streams , in the form of social media and crowd sourced data , into influenza predictions in all time horizons .
We collected CDC-reported ILI , considered the ground truth for this study , from the ILINet website ( http://gis . cdc . gov/grasp/fluview/fluportaldashboard . html ) . We used five independent data sets to develop our ILI weak predictors: ( a ) near real-time hospital visit records from athenahealth , a medical practices management company; ( b ) Google Trends , a Google service that provides approximate search volumes for specific queries ( www . google . com/trends ) , ( c ) influenza-related Twitter microblogging posts , ( d ) FluNearYou , a participatory surveillance system to self-report ILI; and ( e ) Google Flu Trends . All datasets were accessed and downloaded on March 16 , 2015 .
Stacked linear regression is a machine learning methodology commonly used in finance to combine weak predictors of stock prices [30 , 31] . The goal of this methodology is not to identify which ( so called ) “weak predictor” , vk ( t ) , is the best one to predict the quantity y ( t ) ( in our case flu activity ) , but to linearly combine the information contained in all the “weak predictors” to obtain a more accurate and robust single predictor of a quantity y ( t ) . A multivariate approach is used to determine the best linear combination of weak predictors capable of producing the best prediction of the quantity y ( t ) over a training period . Since the weak predictors are , by construction , highly correlated ( indeed , each individual predictor was designed to minimize the square error between the predictions and flu activity ) , a way to discard redundant information is needed . Regularized approaches that penalize the size of the multiplying coefficients , αk , in the multivariate regression , such as Ridge or LASSO regularizations ( L2 and L1 , respectively ) , are good candidates to handle this . We chose LASSO regularization for our ensemble approach since we are interested in identifying models with the smallest number of independent variables ( vk ( t ) ) . Additionally , a non-negative constraint for each multiplicative coefficient αk is imposed . This linear combination is then used to predict the value of y ( t ) for values of t outside of the training period . Support Vector Machine ( SVM ) models [32] are similar to multivariate linear regression models with the important difference that non-linear functions can be chosen as the best relationship between the variables . This is achieved by introducing transformations ( called kernels ) that map the independent variables to higher dimensional feature spaces . The independent variables can even be mapped to an infinite dimensional feature space with the use of a radial basis function ( RBF ) kernel . SVM models are fitted by minimizing an epsilon-insensitive cost function where errors ( between the predictions and the observed values ) of magnitude less than epsilon are ignored in the cost function . This approach typically leads to better generalization of the chosen model on out-of-sample data . The SVM kernel type , margin width , and regularization hyper parameters were chosen via cross-validation on the training data . Decision Tree models are created by recursively splitting the input space , creating local models in each region of the input space . Decision trees , however , have been shown to be unstable as small changes in the data can lead to drastically different tree structures . Boosting methods , such as Adaptive Boosting ( AdaBoost ) , are often employed to fix this problem . Adaptive Boosting ( AdaBoost ) regression [33] fits a sequence of weak learners ( in this case decision trees ) on sequentially reweighted versions of the training data . At each iteration , the weights are individually modified so that the training examples incorrectly predicted by the previous decision tree are given more importance when training the next decision tree . The final prediction is obtained by taking the weighted median of the predictions outputted by the ensemble of weak learners ( AdaBoost . R2 algorithm: [33] ) . In all of the aforementioned regression approaches the goal was to use all available information , in a given point in time , to produce accurate predictions of CDC’s %ILI one , two , three , and four weeks ahead of the release of CDC reports , effectively predicting ILI three weeks into the future . At a given point in time , historical values up to two weeks prior to current date were available for all data sources ( CDC , FNY , ATH , GT , GFT , and TWT ) . In addition real-time ILI estimates were available , with one-week lag , for ATH , GT , GFT , TWT . With this information , we produced predictions for every week starting on July 06 , 2013 and up to February 21 , 2015 . For our first prediction , on the week of July 06 , 2013 , the first training set included 31 weeks worth of historical data from all data sources . For subsequent weeks , we dynamically increased the training set to include all available information at the given date , from all data sources . As a reference , we produced ILI predictions using only historical CDC reported ILI . We achieved this via an autoregressive model with three weekly lagged components as independent variables ( equation 1 in Paul et al 2014 [11] ) . We trained this model for the time period 11/06/11–2/08/15 , and produced out-of-sample predictions for the four weekly time horizons during the time period of our study . We used the same procedure as the ARX model for Twitter , training on the 2011–2012 and 2012–2013 flu seasons , and producing predictions on the 2013–2014 and 2014–2015 flu seasons . These predictions were used to assess the added value provided by our digital disease detection systems’ information . We report 5 evaluation metrics to compare the performance of the five independent predictors and the multiple ensemble methods: Pearson correlation , root mean squared error ( RMSE ) , maximum absolute percent error ( MAPE ) , Root Mean Square Percent error ( RMSPE ) , and hit rate . The definitions of all evaluation metrics are given below . Our notation is as follows: yi denotes the observed value of the CDC’s ILI at time ti , xi denotes the predicted value by any model at time ti , y¯ denotes the mean or average of the values {yi} and similarly x¯ denotes the mean or average of the values {xi} . Pearson Correlation , a measure of the linear dependence between two variables during a time period [t1 , tn] , is defined as: r=∑i=1n ( yi−y¯ ) ( xi−x¯ ) ∑i=1n ( yi−y¯ ) 2∑i=1n ( xi−x¯ ) 2 Root Mean Squared Error ( RMSE ) , a measure of the difference between predicted and true values is defined as: RMSE=1n∑i=1n ( yi−xi ) 2 Root Mean Squared Percent Error ( RMSPE ) , a measure of the percent difference between predicted and true values is defined as: RMSPE=1n∑i=1n ( yi−xiyi ) 2×100 Maximum Absolute Percent Error ( MAPE ) , a measure of the magnitude of the maximum percent difference between predicted and true values , is defined as MAPE= ( maxi|yi−xi|yi ) ×100 Hit Rate , a measure of how well the algorithm predicts the direction of change in the signal ( independently of the magnitude of the change ) , is defined as: HitRate=∑i=2n ( sign ( yi−yi−1 ) ==sign ( xi−xi−1 ) ) n−1×100 where the symbol = = denotes an if statement that returns the value 1 , if the signs of predicted and observed changes are the same , and 0 otherwise . These metrics were calculated for the time period: July 06 , 2013 to February 21 , 2015 .
Table 1 presents the performance of the 5 real-time ( nowcast ) weak predictors as measured by each individual evaluation metric . This table is labeled “last week” since at a given point in time the revised version of all these estimates is only available on the Sunday of the reported week ( or Monday of the subsequent week ) and thus the information effectively predicts the %ILI of last week . For context , we included the metrics of three additional real-time predictions: ( 1 ) the baseline autoregressive predictions described in the previous section; ( 2 ) the CDC’s Virology data , and ( 3 ) the best real-time ensemble method predictions , produced with a support vector machine ( with RBF kernel ) . As Table 1 shows , the real-time ensemble predictions outperform any individual weak predictor in all but one metric ( the hit rate ) . A 0 . 989 Pearson correlation and an average error of about 0 . 176%ILI ( RMSE ) make the ensemble approach a very accurate predictor . The ensemble predictions are very robust as indicated by the size of the MAPE , which measures how much the ensemble method is off-target with respect to the revised CDC ILI estimates . The worst performance was 23 . 6% , which is comparable to the LASSO’s 20 . 2% MAPE . See Table 2 . This error is smaller than two thirds of the smallest MAPE of any of the individual weak predictors . In terms of hit rate , which reflects the ability of the method to predict the upward or downward tendency of the CDC’s ILI ( in addition to the Pearson correlation and independently of producing an accurate point estimate , as captured by RMSE ) , athenaheath data ( ATH ) offers the best results . Furthermore , Table 1 quantitatively shows the added value of using real-time digital disease detection information over a simple historical autoregressive approach . This can be seen by the improvement of the Pearson correlation from 0 . 930 to 0 . 989 , the near three-fold reduction on the RMSE , and the maximum absolute error cut in half . The top panel of Fig 1 graphically shows the revised CDC’s ILI along with the predictions of: the 5 data sources , the baseline , and the best ensemble approach ( SVM RBF ) , as a function of time . The errors for each predictor are displayed in the bottom panel of Fig 1 . The real-time estimates produced with our ensemble method are capable of predicting the timing and magnitude of the two peaks of the 2014–2015 season exactly , whereas they predict the peak of the 2013–2014 season with a one-week lag . Overall predictions track very accurately the CDC’s revised %ILI . This can also be seen in the top left panel of Fig 2 . Since none of the five weak predictors produce predictions into the future ( forecasts ) , we do not have the equivalent of Table 1 for the three forecast time horizons ( labeled “this week” , “next week” , and “in two weeks” ) . Table 2 presents the performance of 4 different machine learning ensemble approaches and the baseline autoregressive predictions for the four time horizons . Figs 2 , 3 and 4 show these results graphically . Ensemble predictions produced with the AdaBoost method show the best accuracy ( lowest RMSE ) and robustness ( lowest MAPE ) , for the three forecast time horizons . Correlation is also highest with AdaBoost in all three horizons . While the hit rate seems to be highest for different methods in different time horizons , Adaboost has an overall best performance as observed in Figs 3 and 4 . We highlight the fact that our ensemble predictions one week into the future , labeled “this week” , have comparable accuracy to real-time GFT predictions , as measured by RMSE . As shown in Table 2 , our ensemble approach produces better results than the baseline AR3 autoregressive model in all similarity metrics and all time horizons . This fact shows quantitatively the value of using social media and crowd-sourced data in improving influenza predictions in future %ILI predictions . Specifically , the average error ( RMSE ) of our ensemble predictions nearly halves the errors of the autoregressive predictions in all time horizons . Pearson correlations of our ensemble approach predictions improve their autoregressive counterparts , from 0 . 845 to 0 . 960 , in the one week forecast; from 0 . 759 to 0 . 927 , in the two-week forecast , and from 0 . 683 to 0 . 904 , in the three week forecast . Note also that our forecast estimates in all time horizons ( up to four weeks ahead of the release of CDC’s reports ) show at least comparable accuracy to “real-time” estimates obtained with a purely autoregressive model . The ability of the ensemble approach forecasts to capture the timing and magnitude of the peaks in the flu seasons decays as the time horizon increases , as observed in Fig 2 . Indeed , one-week forecasts predict the 2013–2014 peak with a one-week lag and with a percent error of about 10% , and they predict the two 2014–2015 peaks with a one-week lag and with percentage errors less than 2% . The two-week forecasts capture the 2013–2014 peak with a one-week lag and show percentage errors of about 10% , and they predict the two 2014–2015 peaks with a two-week lag and percentage errors up to 20% . Finally , the three-week forecasts capture the 2013–2014 peak with a two week lag and show percentage errors of about 20% , and they predict the two 2014–2015 peaks with a two-three week lag and with percentage errors up to 25–30% .
Using weekly information from reports published by the CDC as our gold standard for national flu activity may not necessarily be ideal . Indeed , two data sources considered in this study , athenahealth and Flu Near You , aim at tracking the percentage of the general population with ILI symptoms independently . While athenahealth can be thought of as a subsample of the CDC-reported %ILI ( since it calculates the %ILI in a similar fashion to the CDC , except with the information from those patients seeking medical attention in facilities managed by athenahealth ) , Flu Near You aims at providing an estimate of flu activity from a potentially distinct population ( people willing to report their health status in weekly surveys via a mobile phone app ) . Interestingly , while the sectors of the population sampled by the CDC and FNY maybe distinct ( they may overlap when people report their symptoms using the FNY app and they seek medical attention ) , Fig 1 and a recent study [16] show that their ILI estimates track one another quite well ( Pearson correlation of . 948 ) suggesting that both FNY and CDC datasets may be good proxies of ILI activity in the population . Finally , the best ensemble methodology may change for future flu seasons , and thus , continuous monitoring of the multiple methodologies’ performances should be conducted as new predictions are produced . We presented a methodology that optimally combines the information from multiple real-time flu predictors to produce more accurate and robust real-time flu predictions than any other existing system . Moreover , our ensemble approach is capable of using real-time and historical information to accurately forecast flu estimates one , two , and three weeks into the future .
|
The aggregated activity patterns of Internet users have enabled the detection and tracking of multiple population-wide events such as disease outbreaks , financial markets performance , and preferences in online movie selections . As a consequence , a collection of mathematical models aiming at monitoring and predicting these events in real-time have been proposed in the past decade . As we discover new methods and data sources suitable to track these events , it is not clear whether more information will lead to improved predictions . In the context of digital disease detection at the population level , we show that it is advantageous to combine the information from multiple flu activity predictors in the US instead of simply choosing the best performing flu predictor . Our findings suggest that the information from multiple data sources such as Google searches , Twitter microblogs , nearly real-time hospital visit records , and data from a participatory surveillance system , complement one another and produce the most accurate and robust set of flu predictions when combined optimally .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Combining Search, Social Media, and Traditional Data Sources to Improve Influenza Surveillance
|
The heart exhibits the highest basal oxygen ( O2 ) consumption per tissue mass of any organ in the body and is uniquely dependent on aerobic metabolism to sustain contractile function . During acute hypoxic states , the body responds with a compensatory increase in cardiac output that further increases myocardial O2 demand , predisposing the heart to ischemic stress and myocardial dysfunction . Here , we test the utility of a novel engineered protein derived from the heme-based nitric oxide ( NO ) /oxygen ( H-NOX ) family of bacterial proteins as an O2 delivery biotherapeutic ( Omniox-cardiovascular [OMX-CV] ) for the hypoxic myocardium . Because of their unique binding characteristics , H-NOX–based variants effectively deliver O2 to hypoxic tissues , but not those at physiologic O2 tension . Additionally , H-NOX–based variants exhibit tunable binding that is specific for O2 with subphysiologic reactivity towards NO , circumventing a significant toxicity exhibited by hemoglobin ( Hb ) -based O2 carriers ( HBOCs ) . Juvenile lambs were sedated , mechanically ventilated , and instrumented to measure cardiovascular parameters . Biventricular admittance catheters were inserted to perform pressure-volume ( PV ) analyses . Systemic hypoxia was induced by ventilation with 10% O2 . Following 15 minutes of hypoxia , the lambs were treated with OMX-CV ( 200 mg/kg IV ) or vehicle . Acute hypoxia induced significant increases in heart rate ( HR ) , pulmonary blood flow ( PBF ) , and pulmonary vascular resistance ( PVR ) ( p < 0 . 05 ) . At 1 hour , vehicle-treated lambs exhibited severe hypoxia and a significant decrease in biventricular contractile function . However , in OMX-CV–treated animals , myocardial oxygenation was improved without negatively impacting systemic or PVR , and both right ventricle ( RV ) and left ventricle ( LV ) contractile function were maintained at pre-hypoxic baseline levels . These data suggest that OMX-CV is a promising and safe O2 delivery biotherapeutic for the preservation of myocardial contractility in the setting of acute hypoxia .
Inadequate oxygen ( O2 ) delivery relative to metabolic demand leads to progressive bioenergetic collapse and cellular dysfunction . When systemic , this defines the clinical entity of shock , a major cause of morbidity and mortality in both adults and children [1 , 2 , 3 , 4] . Rather than a specific disease state , shock is a shared pathologic end point arising from disorders such as respiratory failure , hemorrhage , or sepsis that ultimately impair cardiovascular function . For this reason , maintaining a balance between myocardial O2 supply and demand underlies a central therapeutic framework of critical care medicine . Of all organs , the heart is metabolically unique both in regard to its energetic demands as well as its O2 utilization and extraction characteristics . Given its primary physiologic function as a continuous generator of mechanical force , the heart requires an extraordinary supply of biochemical energy and exhibits a far greater rate of ATP turnover than any other organ [5] . Furthermore , the heart is exquisitely dependent on aerobic metabolism to meet these high bioenergetic needs , without the ability to derive any meaningful contribution from anaerobic pathways such as glycolysis [6] . This is reflected in the large myocardial volume devoted to mitochondria and the heart’s status as the highest O2 consumer per gram tissue mass of any organ [5 , 6] . Importantly , its high O2 extraction ratio results in lower venous O2 contents than other tissues , with a significant fraction of cardiomyocytes being exposed to physiologically hypoxic environments at baseline [7 , 8] . When myocardial O2 supply becomes limited in the face of increased demand , dramatic increases in coronary blood flow as well as cardiomyocyte O2 extraction attempt to compensate [7 , 8 , 9] . When inadequate , biochemical signs of a switch to anaerobic metabolism are accompanied by an immediate impairment of contractile function [10] . O2 consumption is thus vital to provide the biochemical energy required to maintain cardiac mechanical function . In this study , we describe , for the first time , the use of a novel O2 delivery biotherapeutic to alleviate hypoxia-induced tissue dysfunction in the heart . Derived from the heme-based nitric oxide ( NO ) /oxygen ( H-NOX ) sensing proteins found in the thermostable bacterium Thermoanaerobacter tengcongensis ( Tt ) [11] , the protein component of Omniox-cardiovascular ( OMX-CV ) is engineered via trimerization and polyethylene glycol ( PEG ) -ylation ( as illustrated in Fig 1A ) to increase circulation half-life , and alterations to the heme-binding pocket to fine-tune both selectivity and avidity of interaction with the diatomic gases NO and molecular O2 [12 , 13] . Unlike hemoglobin ( Hb ) -based O2 delivery biotherapeutics that scavenged NO and therefore triggered significant vascular sequelae , including hypertension , renal dysfunction , and increased risk of myocardial infarction and death [14 , 15 , 16] , the protein component of OMX-CV is uniquely tuned to bind molecular O2 in a way that reduces NO reactivity 50-fold compared with Hb [13] , alleviating the potential risk of vasoconstriction . Additionally , relative to Hb , the protein component of OMX-CV binds to O2 with a very high affinity , exhibiting a dissociation constant ( KD ) of about 2 . 4 μM [13] . Fig 1B shows a schematic comparing the O2 affinities of wild-type Tt H-NOX and OMX-CV with that of Hb , and illustrates how OMX-CV can effectively deliver O2 only to tissues that are significantly hypoxic while bypassing those at physiologic O2 tensions . Following O2 delivery within the hypoxic capillary environment , the unbound OMX-CV molecules enter the systemic venous and pulmonary vascular beds . In this manner , OMX-CV circulates and can be predicted to sustain an ongoing , targeted O2 delivery to the most hypoxic organs and tissues without unnecessary and potentially harmful [18] oxygenation of tissues at physiologic O2 tensions . We hypothesized that in the setting of severe myocardial hypoxia , OMX-CV administration would increase O2 delivery to the heart and improve cardiac mechanical function . In order to test this hypothesis , we utilized a juvenile lamb model of severe acute alveolar hypoxia . The lamb is a robust large animal model that has been extensively utilized because of its close approximation of human cardiovascular function [19] . Here , we present data regarding the safety and efficacy of OMX-CV administration in the setting of systemic hypoxia supporting the use of OMX-CV as a promising novel O2 delivery biotherapeutic .
Previous studies have described the acute cardiovascular response to progressive alveolar hypoxia in large animal models [20 , 21] . Here , we established a model of acute alveolar hypoxia in juvenile lambs triggered via inhalation of a gas mixture containing 10% O2 ( Fig 2A ) . Physiologic data were compared at pre-hypoxic baseline and at 15 minutes of hypoxia ( prior to experimental intervention ) for all animals included in the analysis ( n = 13 ) As expected , we witnessed a precipitous fall in arterial O2 tension ( PaO2 ) with the onset of alveolar hypoxia that was then sustained for the duration of the study ( Fig 2B ) . Fig 2C–2F demonstrate the dramatic changes in physiologic parameters that accompany this severe hypoxemia at the 15-minute time point . The animals all exhibit acute increases in heart rate ( HR ) , systemic blood pressure ( systolic and mean ) , pulmonary blood pressure ( systolic , diastolic , and mean ) , and left and right atrial pressures . As expected [22] , there is a significant increase in pulmonary vascular resistance ( PVR ) attributable to hypoxic pulmonary vasoconstriction . However , there is not a significant alteration in either systemic diastolic blood pressure or systemic vascular resistance ( SVR ) . Additionally , there is an overall increase in cardiac output of approximately 15% ( Fig 2G ) . Although this just fails to reach statistical significance when evaluated at the 15-minute time point ( p = . 063 ) , there is a significant increase in cardiac output amongst all animals ( but no between-group difference ) when evaluated over the duration of the hypoxic exposure ( Fig 3 ) . Table 1 provides additional cardiovascular physiologic parameters comparing OMX-CV and vehicle groups at their respective hypoxic baselines ( before drug ) and study conclusion ( 60 minutes ) . Taking into consideration the historical challenges related to NO scavenging encountered in the use of hemoglobin-based oxygen carriers ( HBOCs ) , we evaluated the physiologic impact of OMX-CV administration on systemic and pulmonary vascular reactivity . Importantly , the total amount of OMX-CV administered relative to circulating Hb is quite low . In an average 10-kg lamb with a serum Hb concentration of 10 g/dL and a circulating blood volume of 70 mL/kg , Hb O2 carrying capacity is approximately 4 . 8 mM . For these experiments , we provided approximately 54 mL total of OMX-CV infusion , representing an infused OMX-CV O2-binding capacity of approximately 0 . 1 mM , or 2% that of circulating Hb . As noted in Table 1 , this does not result in appreciable differences in circulating PaO2 values but is readily available for oxygenating severely hypoxic tissues . Given the substantial physiologic changes induced by the hypoxic stimulus , we specifically evaluated effects on SVR and PVR in the setting of systemic hypoxia prior to and immediately following drug or vehicle administration ( n = 7 control and n = 6 OMX-CV ) . As seen in Fig 4 , we observed no significant increase in either the indexed PVR ( Fig 4A ) or indexed SVR ( Fig 4B ) with administration of OMX-CV when compared with vehicle control under hypoxic conditions . Furthermore , there was no difference in the absolute value or percent change between the OMX-CV–treated and vehicle-treated groups . While hypoxia clearly results in a pre-constricted pulmonary vasculature , this occurs through a NO-independent mechanism , and PVR would be expected to remain quite sensitive to abrupt changes in NO signaling [23 , 24] . Additionally , SVR is also increased during hypoxia , as evidenced by increased mean systemic pressure , and was similarly unaffected by OMX-CV administration ( Fig 4B ) , affirming a lack of direct vasoreactivity . To directly assess the effect of OMX-CV on myocardial tissue oxygenation , following the final assessment of physiologic parameters , pimonidazole ( Hypoxyprobe , 85 mg/kg ) , a well-established marker of tissue hypoxia [25] , was administered intravenously to a subset of animals ( n = 3 per treatment group ) . Thirty minutes after administration of pimonidazole , the animals were humanely killed and tissues collected for processing and measurement of pimonidazole adduct levels in the ventricular myocardium . Pimonidazole freely diffuses into cells and is competitively metabolized via oxidative or reductive chemical reactions , depending on the tissue O2 content . In severely hypoxic environments ( below 10 mm Hg ) , reductive metabolism is favored and in its reduced state , pimonidazole forms covalent adducts with sulfhydryl groups of proteins and glutathione , leading to accumulation of pimonidazole adducts inside the cell [25] . Pimonidazole adducts can be recognized using pimonidazole-targeted primary antibodies and quantified using standard ELISA and immunofluorescent ( IF ) methods . As seen in Fig 5A & 5B , the OMX-CV–treated animals exhibited a significant reduction in myocardial hypoxia compared with controls , as evidenced by lower levels of bound pimonidazole observed via IF microscopy and quantified by ELISA . To verify that the improved myocardial tissue oxygenation in the OMX-CV group was mediated by transcapillary O2 diffusion , rather than vascular extravasation , IF microcopy was performed with antibodies directed against OMX-CV . As seen in Fig 5C , OMX-CV localized within the capillary vascular spaces throughout the heart and not the extracellular spaces surrounding the cardiomyocytes . Thus , at tested doses , a high-affinity O2 delivery biotherapeutic can relieve tissue hypoxia in the heart . To determine whether this improvement in myocardial O2 delivery translates into a physiologic benefit , we utilized cardiac pressure volume loop analysis to evaluate contractile function of the bilateral ventricles . As noted previously by other groups , evaluation of cardiac function in intact animal studies is often obscured by compensatory physiologic alterations to ventricular loading conditions and sympathetic tone [10 , 21] . Indeed , we observed in our own data that from the onset of hypoxia , both the OMX-CV and control groups exhibited similar elevations in cardiac output ( about 15% ) above the normoxic baseline , and that this was sustained throughout our study ( Fig 3 ) . This suggests a full mobilization of compensatory mechanisms that may account for the lack of a significant difference in cardiac output between the OMX-CV and control groups at early time points . Initially advanced by Suga and Sagawa in the 1970s [26] , evaluation of two-dimensional ventricular pressure-volume ( PV ) loops with a focus on the end systolic pressure-volume relationship ( ESPVR ) is now the widely adopted standard used to assess the load-independent contractile state of the ventricles [27] . This method has previously been used to validate the hypoxic depression of myocardial contractile function in dogs and shown to correlate closely with myocardial O2 deficiency and the onset of anaerobic metabolism [8 , 10] . In order to delineate the ESPVR , a family of loops was generated ( as seen in Fig 6A & 6B ) through transient preload suppression induced by graduated occlusion of the inferior vena cava ( IVC ) . The slope of the tangent connecting the end systolic points of these loops gives the most precise representation of intrinsic contractility of the ventricle . As seen in Fig 6A , which shows a representative set of loops and their ESPVR from the LV of a control animal , the decline in slope from baseline ( black ) to hypoxia ( green ) demonstrates a decrease in contractility . In contrast , the LV loops of an OMX-CV–treated animal ( Fig 6B ) exhibit an increasing slope , indicating an improvement in contractile function . By normalizing the slope of the ESPVR at 60 minutes to the baseline for each animal ( n = 7 control and n = 6 OMX-CV ) , we observed that OMX-CV–treated animals maintained an average contractility up to 2-fold above their own baseline under hypoxic conditions ( Fig 6C ) , while RV ( Fig 6C ) and LV ( Fig 6D ) contractility were both reduced in vehicle controls . These data indicate that OMX-CV treatment was able to reverse the effects of myocardial hypoxia and preserve cardiac contractility . We finally explored the role of sympathetic activation in the cardiovascular response to acute alveolar hypoxia by measuring plasma levels of the sympathetic hormones epinephrine and norepinephrine at baseline and after 60 minutes of hypoxia . Released by the adrenal medulla in response to increased stimulation of the sympathetic nervous system , these hormones exhibit potent cardiovascular effects mediated through binding of alpha- and beta-adrenergic receptors in the heart and vasculature . Similar to what has been described [21] , we noted a significant increase in the levels of these catecholamines under hypoxic stress , marking an activated sympathetic response . Interestingly , we found a significant difference in the levels of epinephrine and norepinephrine between the OMX-CV– and vehicle-treated animals ( n = 7 control and n = 6 OMX-CV ) , with hypoxia inducing an approximately 3-fold higher increase in both hormones in the vehicle group compared with OMX-CV ( Fig 6E and 6F ) . Thus , increased adrenergic signaling was not responsible for the improved myocardial contractility of OMX-CV–treated animals compared with the control group , although the improved performance in the presence of the lower induction of catecholamines suggests a greater capacity of the OMX-CV–treated myocardium to respond to adrenergic signaling under hypoxic stress . We therefore conclude that while cardiac output can be maintained during severe acute alveolar hypoxia through diverse adaptive mechanisms , OMX-CV directly improves the intrinsic contractile function of the heart by virtue of its ability to increase myocardial O2 content .
Here , we have provided preclinical data highlighting the therapeutic efficacy of the OMX-CV biotherapeutic in relieving hypoxic myocardial dysfunction in a large animal model . H-NOX–based variants are ideally suited for O2 delivery to hypoxic tissues , such as the myocardium , because of their O2 affinity as well as pharmacokinetic and safety profiles [13] . OMX-CV’s O2 affinity aligns extremely well with the unique O2 demands and microenvironments encountered within the stressed heart , its half-life enables long-term efficacy following single intravenous infusion , and its O2 specificity minimizes the vasoactive side effects encountered with HBOCs . The cardiovascular system responds to acute hypoxia by attempting to augment and enhance systemic O2 delivery . Cardiac output increases with accompanying elevations in both HR and contractile state , which further escalate myocardial O2 demand . In response to the high and variable demand for O2 during states of acute stress , as well as the tight interrelationship between myocardial function and O2 supply , the heart has evolved robust adaptive mechanisms to augment myocardial O2 delivery and extraction [28] . For example , during exercise-induced elevations in cardiac output , O2 utilization may increase by greater than 5-fold , supported by substantial increases in coronary blood flow , capillary recruitment , and increased O2 extraction [7] . Even under unstressed conditions , the heart exhibits a high O2 extraction ratio with a correspondingly low venous saturation . When demand increases , the heart has a unique capacity to increase extraction to a greater extent than other tissues [8] . Cain and colleagues initially demonstrated that global hypoxic hypoxia and anemic hypoxia induced global anaerobic metabolism at greatly differing values of mixed venous partial pressure of oxygen ( PO2 ) [9] . These differences in tissue responses to the same level of hypoxia in the blood implied that simple diffusion forces are not the limiting factor to tissue O2 extraction , and Schumacker and colleagues subsequently confirmed that a constant critical O2 extraction ratio exists in dogs [29] . Although the exact mechanisms underlying these differences are unclear , the physiologic consequence is that most tissues will start to experience O2 deficiency despite a relatively high average O2 saturation of the blood exiting their capillaries . In contrast to the other tissues and organs , the myocardium can achieve a substantially higher O2 extraction ratio , only exhibiting signs of anaerobic metabolism at a critically low coronary venous saturation [8] . This markedly hypoxic venous and end capillary blood reflects a correspondingly hypoxic tissue bed , creating the ideal cellular microenvironment to facilitate O2 dissociation and delivery by OMX-CV . Consistent with this prediction , we have shown here that in the stressed , hypoxic lamb heart , myocardial oxygenation and contractile function can be preserved with the administration of OMX-CV . This is particularly remarkable given that the total amount of OMX-CV used in our studies equates to only approximately 2% of the total O2 carrying capacity of the circulating Hb . Importantly , the small amount of OMX-CV administered relative to total circulating Hb serves to limit any potential negative impact on total O2 bioavailability . Furthermore , the high O2 affinity of OMX-CV precludes O2 delivery under non-hypoxic conditions . This is in marked contrast to the less avid delivery profile of Hb and most HBOCs , which have been shown to contribute to pathologic hyperoxygenation of tissue and circulatory microenvironments [30] . This excessive O2 release has been shown to cause oxidative stress to the tissues through the production of toxic reactive oxygen species ( ROS ) and to induce detrimental microvascular shunting mechanisms that may inappropriately impair tissue perfusion . Delivery of excess O2 in the setting of shock is a frequent contributor to microcirculatory shunting with significant clinical consequences [31] . While vascular indices can frequently be normalized within the macrocirculation in the setting of shock , tissue perfusion can nevertheless be compromised because of shunting at the microcirculatory level . Importantly , in adult patients with severe sepsis and traumatic hemorrhagic shock , for example , the loss of coherence between the resuscitated macrocirculation and the microcirculation is one of the most sensitive and specific hemodynamic indicators associated with increased multi-organ failure and mortality [32 , 33 , 34 , 35] . Similarly , in critically ill children with sepsis , a persistently altered microcirculation has been associated with increased mortality [36] . OMX-CV allows a more targeted delivery of O2 to only the most hypoxic tissue beds and may help alleviate the underappreciated but significant morbidities associated with excessive tissue oxygenation in this setting . Interestingly , we noted in our study that OMX-CV administration was associated with a smaller increase in circulating catecholamine levels in the setting of systemic hypoxia . While it is unclear what exactly underlies this difference in catecholamine production and release , it does suggest potential implications related to cardiac function . Hypoxia is a well-established stimulus for catecholamine secretion both in vitro and in vivo [37 , 38 , 39] , and adrenergic responses to hypoxic stress are important for the maintenance of cardiorespiratory homeostasis [40 , 41] . In the perinatal period , catecholamine production by adrenomedullary chromaffin cells is directly stimulated by cellular hypoxia [42 , 43] . However , as mammals age , this primary cellular response to O2 is blunted and cholinergic innervation becomes the predominant regulatory mechanism [44] . The sympathetic response to hypoxia therefore matures to reflect the integrated input from peripheral and central chemoreceptors . In our juvenile lamb model of systemic hypoxia , OMX-CV administration appears to blunt hypoxia-driven catecholamine production . It is not clear if this reflects augmented O2 delivery to chemoreceptors or the chromaffin cells themselves , or perhaps represents some secondary mechanism related to more favorable hemodynamics associated with improved myocardial oxygenation . Importantly , in the control animals , diminished cardiac contractility is observed despite dramatically elevated levels of circulating catecholamines , while the OMX-CV–treated animals exhibit preserved contractility despite smaller increases in catecholamine levels . Epinephrine and norepinephrine are potent inotropes , vital to the regulation of cardiac contractility and hemodynamic function in response to physiologic stress . Here , we show that OMX-CV supports preservation of the cardiac response to these key regulators , which are important not only as endogenous hormones but also as exogenous agents heavily utilized for cardiovascular support in critical care medicine . With respect to its safety profile , OMX-CV exhibits significant advantages over previously developed HBOCs [45] . As the protein responsible for storage and transport of O2 in red blood cells ( RBCs ) [46] , Hb has been the precursor for the synthesis and formulation of HBOCs previously developed as RBC substitutes [47 , 48 , 49 , 50] . The first HBOC to be developed in this capacity consisted of partially purified “stroma-free” Hb [51] . However , transfusion of acellular Hb led to several major side effects [52 , 53 , 54 , 55 , 56] . Extracellular tetrameric Hb readily dissociates into two pairs of dimers [53 , 54] , which are extremely prone to oxidation [56] and enhanced renal excretion [53 , 57] . Hb oxidation to methemoglobin ( metHb ) promotes unfolding of the globin chains and releases cytotoxic heme into the circulation , leading to kidney tubule damage and eventual renal failure [53 , 54] . Furthermore , metHb can no longer carry O2 and can also contribute to the generation of harmful ROS [52 , 55] . Additionally , extracellular Hb can trigger vasoconstriction and systemic hypertension by various mechanisms [30 , 58 , 59] . Foremost amongst these is the indiscriminate scavenging of NO , an important intrinsic vasodilator that is locally produced by endothelial cells to relax vascular smooth muscle [58 , 60] . Also , potentially important is the hyperoxygenation of local vasculature that can elicit inappropriate vasoconstriction within the microcirculation , compared to more tempered O2 delivery into the vessel lumen from physiologic RBC-encapsulated Hb [30 , 45] . Overall , the presence of extracellular Hb in the circulation may lead to direct tissue toxicity via heme release and ROS generation , while simultaneously impairing blood flow because of pathologic alterations in vasomotor tone . With its unique structure and O2-binding characteristics , OMX-CV averts the potential for many of these deleterious side effects . In this study , we have shown a lack of direct vasoreactivity in both the systemic and pulmonary vascular beds , providing strong evidence for selective O2 delivery in severely hypoxic microenvironments and lack of vasoactivity . In summary , we present preclinical data from a large animal model highlighting the therapeutic efficacy of a novel O2 delivery biotherapeutic agent , OMX-CV , in relieving hypoxic myocardial dysfunction . OMX-CV is ideally suited for myocardial O2 delivery because of its unique O2-binding characteristics and safety profile . Its high O2 affinity complements the unique O2 demands and microenvironments encountered within the stressed heart , while its low reactivity with NO minimizes the vasoactive side effects encountered with HBOCs . Additionally , while exogenous O2 administration can increase systemic arterial O2 content , it can also result in microvascular shunting mechanisms that limit deep tissue oxygenation [61 , 62] . OMX-CV therefore has the potential to improve oxygenation in a wide range of tissues and clinical scenarios in which O2 delivery may be compromised .
All protocols and procedures for this work were approved by the Institutional Animal Care and Use Committee of the University of California , San Francisco . AN155428 . In this study , 13 juvenile lambs ( 4–6 weeks of age ) were anesthetized with fentanyl , ketamine , and diazepam and paralyzed with vecuronium to facilitate intubation and mechanical ventilation . Ongoing sedation and neuromuscular blockade were administered as a continuous infusion of ketamine , fentanyl , diazepam , and vecuronium . The sedative mixture was titrated to maintain age-appropriate HR . Femoral venous and arterial access were obtained via cutdown of the hind limbs , and arterial pressure was continuously transduced and recorded . The animals were ventilated with 21% FiO2 initially , with a positive end expiratory pressure of 5 cm H2O , tidal volumes of 10 mL/kg , and respiratory rate titrated to maintain pCO2 of 35–45 millimeters mercury ( mmHg ) by arterial blood gas measurements . Thoracotomy was performed and Sorenson Neonatal Transducers ( Abbott Critical Care Systems , N . Chicago , IL ) were introduced into the left and right atria and main pulmonary artery ( MPA ) to continually transduce and record pressures . An ultrasonic flow probe ( Transonics Sytems , Ithaca , NY ) was placed on the left pulmonary artery ( LPA ) to continuously monitor and record blood flow . Admittance PV catheters ( Transonics Systems , Ithaca , NY ) were introduced into the RV and LV via ventriculostomy to perform ventricular pressure volume analysis . These catheters consist of a solid-state sensor that directly measures pressure with high precision and excitation and recording electrodes that measure volume based on electrical admittance . Alternating current applied to the excitation electrodes generates an electrical field within the ventricle and the recording electrodes measure voltage changes , allowing calculation of resistance and conductance . With input of a measured blood resistivity and baseline stroke volume ( as assessed by total cardiac output estimate from LPA flow/HR ) , time varying conductance can be used to solve for ventricular blood volume in real time [63] . Animals with Hb levels of less than 7 . 5 g/dL following surgical instrumentation were transfused with fresh whole maternal blood in increments of 5 mL/kg to achieve this minimum threshold . Following instrumentation , the animals were allowed to recover to steady state until they required no further adjustment to sedatives and exhibited stable hemodynamic parameters . This time was designated as the normoxic baseline and blood gas analysis was performed . Baseline ventricular ESPVR was assessed by transient IVC occlusion . Following baseline assessment , the animals were subjected to sustained alveolar hypoxia by ventilation with an admixture of atmospheric gas and nitrogen to achieve a FiO2 of 10% . Arterial blood gas analysis was performed every 15 minutes with blood withdrawn from the femoral artery and analyzed using a Radiometer ABL5 pH/blood gas analyzer ( Radiometer , Copenhagen , Denmark ) . Ventilatory rate was adjusted to maintain PCO2 35–45 mmHg and metabolic acidosis was corrected with NaHCO3 boluses to maintain pH >7 . 30 . All protocols and procedures for this work were approved by the Institutional Animal Care and Use Committee of the University of California , San Francisco . Animals’ vital signs , including core temperature , were monitored throughout the study , and they were given intravenous fluids and prophylactic antibiotics per protocol . At the end of each protocol , all lambs were euthanized with a lethal injection of sodium pentobarbital followed by bilateral thoracotomy , as described in the NIH Guidelines for the Care and Use of Laboratory Animals . The engineered Tt H-NOX protein described in this study was produced by QuikChange Site-Directed Mutagenesis ( Agilent ) , subcloned into an expression plasmid , transformed into Escherichia coli , and expressed essentially as described [11] . Cells were harvested by hollow-fiber tangential-flow filtration and processed immediately . The His-tagged Tt H-NOX protein was purified from cell lysate using Ni-affinity chromatography and further polished by passage over an anion-exchange column to remove remaining host cell DNA , host cell proteins , and endotoxins . The purified protein was formulated to produce OMX-CV , and frozen at −80 °C until use . Protein concentrations were determined using UV-Vis spectrophotometry as described [11] . Prior to use in animal studies , OMX-CV was subjected to purity testing by SDS-PAGE ( Invitrogen ) and SEC-HPLC ( Agilent ) and safety testing by kinetic chromogenic LAL test for endotoxin ( Charles River Laboratories ) . For use in animal studies , proteins lots were required to be greater than 95% pure and have endotoxin levels less than 0 . 1 EU/mg . After 15 minutes of alveolar hypoxia , the animals received either 200 mg/kg of OMX-CV ( about 4 mL/kg by volume ) as a bolus over 10 minutes , followed by continuous infusion at 70 mg/kg/hour ( OMX-CV group n = 6 ) , or an equivalent volume of the OMX-CV vehicle solution administered in the same manner ( control group n = 7 ) . At 60 minutes of alveolar hypoxia , repeat evaluation of the ESPVR was assessed by IVC occlusion . Physiologic data were continuously recorded and analyzed using the Ponemah Physiology Platform ( Data Sciences International , New Brighton , MN ) with Acquisition Interface , ACQ-7700 ( Data Sciences International , St . Paul , MN ) . For calculation of total cardiac output , LPA blood flow was assumed to represent 45% of total output , as previously established in juvenile lambs by Rudolph . This was indexed to animal size by dividing by the animal’s body weight in kilograms . PVR was calculated as the difference of mean pulmonary arterial pressure and left atrial pressure divided by the indexed cardiac output . SVR was calculated as the difference of mean systemic arterial pressure and right atrial pressure divided by the indexed cardiac output . Pressure volume loop recording and analysis were performed using Labscribe software ( iWorx , Dover , NH ) . At baseline and again at 60 minutes of hypoxia , plasma and serum samples were collected from all animals ( n = 7 control and n = 6 OMX-CV ) for additional analysis , including measurement of circulating catecholamines . Determination of epinephrine and norepinephrine levels in plasma was performed using a colorimetric ELISA kit ( ABNOVA ) according to the manufacturer’s instructions . In a subset of animals ( n = 3 control and n = 3 OMX-CV ) , following the final physiologic assessment , pimonidazole ( 85 mg/kg ) was administered intravenously over 10–15 minutes , as tolerated . Thirty minutes following the pimonidazole infusion , the animals were euthanized for tissue collection . Myocardial tissues were snap-frozen and proteins were then extracted and processed for competitive pimonidazole ELISA , as described [64] . Standard curves for the pimonidazole ELISA were fit using a five-parameter logistic equation and used to determine IC50 values . Values were normalized to the protein concentration in each sample and then expressed relative to the vehicle control . Myocardial tissues were frozen in OCT and processed for cryosectioning , followed by immunohistochemical analysis . Sections were fixed with 100% methanol for 20 minutes at −20 °C , then blocked and permeabilized with 5% BSA , 5% goat serum , and 0 . 1% Tween 20 for 1–2 hours at room temperature . Sections were then incubated with anti-pimonidazole ( Hypoxyprobe , 1:100 ) , anti-OMX-CV ( 1:200 , Mouse monoclonal ) antibodies overnight at 4 °C , followed by anti-rabbit or anti-mouse secondary antibodies ( 1:1 , 000 , Jackson Immunoresearch Laboratories , West Grove , PA ) for 2 hours at room temperature . The sections were mounted in SlowFade DAPI ( Invitrogen ) and imaged at the UCSF Laboratory for Cell Analysis Core with an HD AxioImager Zeiss microscope equipped with a CCD digital camera . Comparison of physiologic data comparing pre-hypoxic baseline to the first hypoxic physiologic time point was performed using a paired Student t test . Evaluation of cardiac output over the duration of the study between groups was performed using two-way ANOVA analysis . Evaluation of PVR and SVR before and after treatment between groups was performed using two-way ANOVA analysis . Pimonidazole levels were compared between groups using an unpaired Student t test . For ESPVR data , the slope of the ESPVR at 60 minutes of hypoxia for each ventricle of each animal was normalized to its own baseline ESPVR . These normalized values were then compared between groups using an unpaired Student t test . Epinephrine and norepinephrine levels at 60 minutes of hypoxia were compared between groups using unpaired Student t test . For all statistical tests performed , p ≤ 0 . 05 was considered to be significant . All analyses were performed using GraphPad Prism version 6 . 04 for Macintosh , GraphPad Software , La Jolla , CA .
|
While hemoglobin is the primary oxygen delivery molecule used to maintain tissue oxygenation in metazoans , many organisms have other heme-containing proteins that can bind oxygen and other diatomic gases . Here , we tested whether a member of the H-NOX family of heme-containing proteins found in the thermostable bacterium Thermoanaerobacter tengcongensis can be engineered to deliver oxygen to severely hypoxic tissues in large mammals . This class of molecules has the advantage of high oxygen affinity and minimal nitric oxide reactivity . We demonstrate that these molecules can effectively deliver oxygen to a lamb heart with induced severe hypoxia , without overexposing the animal to oxygen or triggering systemic vascular reactivity . These molecules thus represent a novel class of oxygen delivery biotherapeutics to specifically target hypoxic tissue beds without the toxicity concerns of hemoglobin-based oxygen carriers . As tissue hypoxia is a central feature of many disease processes , this therapeutic approach may have broad clinical applicability .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"cardiovascular",
"anatomy",
"oxygen",
"cardiac",
"output",
"pulmonology",
"arteries",
"hypoxia",
"pulmonary",
"arteries",
"immunologic",
"techniques",
"cardiology",
"medical",
"hypoxia",
"research",
"and",
"analysis",
"methods",
"blood",
"vessels",
"immunoassays",
"chemistry",
"blood",
"pressure",
"short",
"reports",
"chemical",
"elements",
"cell",
"biology",
"anatomy",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"vascular",
"medicine",
"heart"
] |
2018
|
Preservation of myocardial contractility during acute hypoxia with OMX-CV, a novel oxygen delivery biotherapeutic
|
Discrimination between self and non-self is a prerequisite for any defence mechanism; in innate defence , this discrimination is often mediated by lectins recognizing non-self carbohydrate structures and so relies on an arsenal of host lectins with different specificities towards target organism carbohydrate structures . Recently , cytoplasmic lectins isolated from fungal fruiting bodies have been shown to play a role in the defence of multicellular fungi against predators and parasites . Here , we present a novel fruiting body lectin , CCL2 , from the ink cap mushroom Coprinopsis cinerea . We demonstrate the toxicity of the lectin towards Caenorhabditis elegans and Drosophila melanogaster and present its NMR solution structure in complex with the trisaccharide , GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc , to which it binds with high specificity and affinity in vitro . The structure reveals that the monomeric CCL2 adopts a β-trefoil fold and recognizes the trisaccharide by a single , topologically novel carbohydrate-binding site . Site-directed mutagenesis of CCL2 and identification of C . elegans mutants resistant to this lectin show that its nematotoxicity is mediated by binding to α1 , 3-fucosylated N-glycan core structures of nematode glycoproteins; feeding with fluorescently labeled CCL2 demonstrates that these target glycoproteins localize to the C . elegans intestine . Since the identified glycoepitope is characteristic for invertebrates but absent from fungi , our data show that the defence function of fruiting body lectins is based on the specific recognition of non-self carbohydrate structures . The trisaccharide specifically recognized by CCL2 is a key carbohydrate determinant of pollen and insect venom allergens implying this particular glycoepitope is targeted by both fungal defence and mammalian immune systems . In summary , our results demonstrate how the plasticity of a common protein fold can contribute to the recognition and control of antagonists by an innate defence mechanism , whereby the monovalency of the lectin for its ligand implies a novel mechanism of lectin-mediated toxicity .
Adequate and efficient defence mechanisms to protect an organism's integrity and survival have been essential for the evolution of multicellularity since loss of individual cells may be detrimental for a multicellular organism . Any defence mechanism thereby critically relies on the ability to discriminate between self and non-self . Since all living cells display specific carbohydrate structures on their surface [1] , glycans have been used for the recognition of non-self since the beginning of multicellular life [2] . Accordingly , many of the proteins that are able bind to specific carbohydrate structures , commonly referred to as lectins , have been implicated in defence , mainly in the innate immune systems of animals which is considered an ancestral defence mechanism and a first and immediate line of defence against potentially harmful microorganisms [3] . These lectins are either membrane-bound or secreted and localize to the interface between the host and the environment where they bind to microorganism-associated carbohydrates and function either as receptors triggering the expression of host immune effectors , by opsonizing the microorganisms for host immune effectors or immune cells ( reviewed in [4] ) or as direct immune effectors by killing the microorganism upon binding [5]–[8] . In analogy to latter function of combining non-self recognition and killing , plants use insecticidal lectins to defend themselves against herbivorous insects [9] . Recently , a group of fungal lectins , commonly referred to as fruiting body lectins , has been shown to play a role in the defence of multicellular fungi against predators and parasites based on their toxicity to various model organisms [10]–[15] . According to the above role of lectins in defence , most defence lectins should be specific for carbohydrate structures that do not exist in the host ( are non-self ) and are characteristic for the target organism . To date , only very few target carbohydrate structures or glycoconjugates of such lectins involved in innate defence mechanisms have been identified and their recognition by the lectin investigated at molecular level [5] , [8] , [10] , [16] , [17] . In organisms lacking an antibody-based adaptive immunity , such a lectin-based defence strategy critically relies on a large diversity in carbohydrate specificities . This diversity can be achieved either by diversification on the level of lectin folds and/or by the plasticity of a common lectin fold . The known fruiting body lectins belong to six structural families [14] of which the β-propeller-fold lectins , actinoporin-like lectins , galectins and β-trefoil ( ricin B or R-type ) lectins [18] are the most prominent ones . Some of these lectins are multidomain proteins harbouring in addition a cysteine protease/dimerization domain ( R-type Marasmius oreades agglutinin [MOA] and Polyporus squamosus lectin [PSL] ) [19] , [20] or a pore-forming module ( R-type Laetiporus sulphureus lectin [LSL] ) [21] . In the first case , it was demonstrated that both domains are required for toxicity [10] suggesting that the lectin domain guides the catalytic domain to specific target structures . However , most lectins implicated in the defence of plants and fungi are composed just of lectin domains and contain multiple binding sites for either the same or different carbohydrate structures . For some of these lectins it has been demonstrated that this multivalency is essential for their toxicity [22] . These results suggest that lectin-mediated toxicity involves crosslinking of glycoconjugates but the exact mechanism remains unclear . We describe the identification and characterization of a novel , monovalent lectin , CCL2 , from fruiting bodies of the ink cap mushroom Coprinopsis cinerea and present the NMR structure of CCL2 in ligand-free form and in complex with its in vivo ligand . The lectin was found to bind specifically and with an atypical high affinity to Fucα1 , 3-modified core N-glycans in vitro , using a single , topologically novel binding site on its β-trefoil fold . N-glycans carrying such a modification are characteristic for invertebrates but absent from fungi . We applied biotoxicity assays to demonstrate toxicity towards two model invertebrates . In accordance with the in vitro binding data , the nematotoxicity of CCL2 was dependent on core α1 , 3-fucosylation of C . elegans N-glycans on intestinal proteins of the nematode . These results show how multicellular organisms exploit the plasticity of a common protein fold to create a novel lectin specificity and an alternative mechanism of lectin-mediated toxicity for defence .
We detected a soluble 15 kDa protein from fruiting bodies of the model mushroom C . cinerea by virtue of its binding to horseradish peroxidase ( HRP ) in immunoblots . The protein was present in extracts from fruiting bodies but not from vegetative mycelium , indicating a fruiting body-specific expression . We isolated the protein using HRP-affinity chromatography ( Figure 1A ) and identified it as hypothetical protein CC1G_11781 of C . cinerea strain Okayama7 by MALDI-MS/MS . Since the protein , termed CCL2 ( Coprinopsis cinerea lectin 2 ) , was extracted from fruiting bodies of the C . cinerea strain AmutBmut ( Swamy et al 1984 ) , the respective genomic locus of strain AmutBmut was cloned and sequenced . This sequence served as a basis for the cloning of the respective cDNA from total RNA isolated from AmutBmut fruiting bodies . A second cDNA , coding for an isoprotein ( 52% identity; Table S1 ) , termed CCL1 ( Coprinopsis cinerea lectin 1 ) ( CC1G_11778 ) , was cloned and sequenced accordingly . The two proteins are predicted to contain neither a signal sequence for classical secretion nor N-glycosylation sites . The cDNAs coding for CCL1 and CCL2 were cloned in pET expression vectors and the proteins were expressed in the cytoplasm of E . coli BL21 ( DE3 ) . The recombinant proteins were highly expressed and soluble ( Figure S1 ) and versions containing eight N-terminal His-residues were purified using metal-affinity chromatography . Size exclusion chromatography of the purified CCL2 showed that the protein exists as a monomer in solution ( Figure S2 ) . Immunoblots using a CCL2-specific antiserum confirmed that CCL2 is abundant in fruiting bodies and absent from vegetative mycelium ( Figure 1B ) . The differential expression of both CCL2 and CCL1 was quantified at the transcript level by qRT-PCR ( Figure 1C ) . The results indicate that the mRNA levels of CCL1 and CCL2 are more than 1000-fold and 60 , 000-fold , respectively , higher in fruiting bodies than in the vegetative mycelium . Based on the binding to the plant glycoprotein HRP and a similar expression pattern as previously characterized lectins from this organism [23] , [24] , we hypothesized that CCL2 is a lectin . Fluorescently labeled CCL2 was used to probe a glycan array offered by the Consortium of Functional Glycomics ( CFG ) ( Figure 2 and Table S2 ) , confirming that CCL2 is a lectin that binds specifically to carbohydrate structures containing the Fucα1 , 3GlcNAc motif e . g . the LewisX antigen ( Galβ1 , 4[Fucα1 , 3]GlcNAc; Glycan structure #133/134 on the array ) . The disaccharide Fucα1 , 3GlcNAc alone , however , showed a very low fluorescence , suggesting that at least a trisaccharide was required for efficient binding . Glycan array analysis with purified CCL1 ( Figure S3 and Table S3 ) yielded almost the same results as with CCL2 . The binding specificity of CCL2 was further studied with several carbohydrates in vitro by NMR spectroscopy and isothermal titration calorimetry ( ITC ) as summarized in Table 1 . The trisaccharide LewisX bound with a moderate KD of 456 µM and the NMR spectra displayed intermediate to slow exchange behavior during the titration , whereas the binding of sialylated LewisX , was slightly better by a factor of ∼3 . However , fucosylated chitobiose ( GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc-spacer; Figure 3A ) , absent on the glycan array , had by far the highest affinity among the tested oligosaccharides with a KD of 1 . 4 µM ( Table 1 and Figure S4 ) . Monitoring the binding by NMR spectroscopy revealed large chemical shift changes under the slow exchange regime ( Figures 3B and C ) . Binding occurs with a stoichiometry of 1∶1 and no further changes were observed by adding an excess of ligand ( 1∶50 ) . The largest chemical shift deviations occurred at residues W78 , N90-T95 , G108 and K109 ( Figure 3D ) . Since CCL2 did not show sequence similarity to any known structure we determined the 3D structure of CCL2 by NMR spectroscopy ( Figure 4 ) . CCL2 adopts a β-trefoil fold consisting of three β-β-β-β repeats with a pseudo C3 symmetry . β1 and β4 of each repeat form together a β-barrel whereas β2 and β3 adopt a β-hairpin that usually harbors the carbohydrate-binding site [25] . The β-trefoil structure can be compared to a tree [26] in which the trunk is represented by the β-barrel ( β1 and β4 , β5 and β8 , β9 and β12 ) , the roots are formed by the N- and C-terminus together with the two loops β4–β5 and β8–β9 , the upper crown is formed by the three β-hairpins ( β2 and β3 , β6 and β7 , β10 and β11 ) and the lower crown by the loops connecting the β-barrel with the β-hairpin loops . As can be seen from Figures 4B and D , the loops β6–β7 and β7–β8 in subdomain β are shorter than in the other subdomains . In addition , subdomain β shows a deviation from the most characteristic feature of β-trefoil proteins , the QxW motif in each subdomain [25] . Subdomain β contains a YxW instead . A search for structurally similar proteins revealed a large number of bacterial , fungal and plant toxins displaying high structural similarity but low sequence identity ( Table S4 ) . The 3D structure was used to visualize the largest chemical shift deviations from the titration experiment with GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc ( from Figure 3D ) in Figure 4C . The largest deviations occur at the interface between subdomain β and γ , mainly on strand β8 and its unusually short preceding loop β7–β8 ( β subdomain ) and in the β9–β10 loop ( γ subdomain ) . This arrangement does not correspond to the typical binding interface of β-trefoil lectins and therefore we decided to investigate this new binding mode . We solved the 3D structure of the complex between CCL2 and fucosylated chitobiose ( GlcNAcβ1 , 4[Fucα1 , 3]GlcNAcβ–sp ) by NMR spectroscopy . 82 intermolecular distance restraints that are well distributed over the binding interface ( Figure 5A ) were derived from a 3D 13C F1-edited F3-filtered HSQC-NOESY [27] spectrum ( Figure S5 ) . A precise structural ensemble of the complex was obtained ( Figure 5B and Table 2 ) . The carbohydrate is bound at the interface of the subdomains β and γ in the lower crown ( Figure 5C ) , in particular between the β-strands β6 and β8 and the linker β7–β8 of the β subdomain and the loop between β9–β10 of the γ subdomain . Compared to the canonical binding sites ( Figure 5G and Figure S6 ) this is a very unusual binding location for ricin B type lectins . The well-defined trisaccharide is oriented such that GlcNAc2 ( see Figure 3A for nomenclature of the individual sugars in the trisaccharide ) stacks on top of Fuc2′ thereby locking the conformational freedom of the glycan resulting in a narrow clustering of the glycosidic angles ( Figure S7 ) . The hydrophobic B-face of Fuc2′ is oriented towards the protein ( bottom ) and the hydrophobic B-face of GlcNAc2 towards the solution ( top ) . In this orientation GlcNAc1 is tilted horizontally such that its B-face is located on the back contacting the protein . Contacts to all three sugar units are mediated by a large number of potential H-bonds and hydrophobic interactions ( Figures 5D–F and Table 3 ) . The specific recognition of each sugar unit can be described as follows: Fuc2′ approaches the edge of β-strand β8 and the tip of loop β9–β10 with its b-face and bridges subdomain β and γ in this way ( Figures 5C–F ) . In this orientation O4 and O5 face down and are specifically recognized by H-bonds to the main chain ( V93 HN and O ) of the unusually short loop between strands β7 and β8 ( Figure 5E ) . The equatorial hydroxyl groups of O3 and O2 form H-bonds to G108 HN ( second largest chemical shift deviation , Figure 3D ) and Lys109 NH3+ . In addition the hydrophobic methyl group and the axial H2 , both facing downwards , form hydrophobic contacts with Trp94/Trp95 and Val93 , respectively . The methyl group is located above the ring of W94 enabling favorable Me-π interactions that are supported by an upfield shift of the H6 resonance ( −0 . 18 ppm; Table S5 ) . In total all characteristic groups of Fuc2′ are specifically recognized by 4 H-bonds , hydrophobic and π interactions . Both the location of Fuc2′ at the subdomain interface and the recognition by three H-bonds to the main-chain are unprecedented in all ricin B type lectin complex structures . GlcNAc1 is specifically recognized at its equatorial acetamido group by a H-bond of its HN to Asn91 O , and at O6 by an H-bond to the side chain of Asn90 . The acetamido group forms hydrophobic interactions to Val93 and Me-π interactions with Tyr57 which is supported by an upfield shift of the methyl 1H resonance ( −0 . 24 ppm ) . Its hydrophobic b-face packs to the Tyr92 side chain . Only a GlcNAc would be recognized at this position since the equatorial orientation of the acetamido and the CH2OH group are necessary for their recognition by H-bonds and the equatorial positioning of O3 and O4 is required for the stacking between Fuc2′ and GlcNAc2 . GlcNAc2 is mainly recognized via its acetamido group by an H-bond to Trp78 HN ( supported by the largest HN chemical shift deviation , Figure 3D ) , hydrophobic interactions of the methyl with Leu87 and a stacking of the entire acetamido group to the ring of Tyr92 ( Figure 5D ) . Me-π interactions to Y92 are supported by an upfield shift ( −0 . 24 ppm ) . GlcNAc2 that stacks on top of Fuc2′ is slightly laterally shifted exposing the hydrophobic H4 facing downwards . H4 is located on top of the Trp94 ring and favorable H-π interactions are supported by an upfield shift of its resonance ( −0 . 39 ppm; Table S5 ) . Two additional potential H-bonds are observed in some structures of the ensemble: between the carbonyl of W78 and O3 of GlcNAc2 and between K109 NH3+ and GlcNAc2 O6 . In summary , GlcNAc1 and Fuc2′ are specifically recognized by interactions to almost all of their functional groups whereas the recognition of GlcNAc2 is more relaxed . It is mainly recognized at its equatorial acetamido group attached to C2 . This residue must be able to stack to Fuc2′ in order to properly position the acetamido group; both GlcNAc and GalNAc fulfill this requirement and will be recognized in this position . Accordingly , CCL2 binds to fucosylated LacdiNAc ( GalNAcβ1 , 4[Fucα1 , 3]GlcNAc; Glycan structure #89 ) on the array . The large number of H-bonds to the main chain is remarkable . The unusually short β7–β8 loop contributes three and the β9–β10 loop one such H-bonds . Since the protein main chain does not change upon binding , part of the recognition pattern on the protein is preformed . However , the lengths and conformations of these loops are a special feature of CCL2 homologues as illustrated on a structure-based alignment ( Figure 6 ) and are not conserved in the β-trefoil fold . Note also that the short β7–β8 loop lacks the typical 310 helix segment as seen for example in the structurally most closely related R-type lectin MOA ( Figure 5G ) which would clash with the carbohydrate . The interaction of CCL2 with GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc is governed by a large ΔH gain of −50 kJ mol−1 at the expense of 16 kJ mol−1 for −TΔS ( Figure 7A ) . The thermodynamicbinding parameters are comparable to those of other high affinity lectins in Figure 7B ( Table S6 ) . In contrast to typical lectin interactions with medium affinity CCL2 uses an unusually large number of H-bonds ( 5–7 to backbone , 5 to side chain ) and hydrophobic contacts ( Trp78 , Tyr92 and Trp94 ) for recognition of its target . A comparable number and kind of contacts is only found for few high affinity lectin interactions with a comparable KD∼1 µM . Interestingly , the calreticulin interaction with Glcα1 , 3Manα1 , 2Manα1 , 2Man with a KD of 0 . 77 µM is governed by almost identical thermodynamic values [28] , whereas the structurally closely related R-type lectin PSL [20] that binds to 6'sialyl lactose with a KD of 1 . 3 µM [29] displays a moderately favored enthalpy but almost no entropic penalty . Both lectins use a similar number of direct H-bonds for their target recognition as CCL2 does: 10 ( 2 to backbone , 8 to side chain ) and 9 ( 4 to the backbone , 5 to side chains ) , respectively , and a comparable amount of hydrophobic interactions . We tested the toxicity of CCL2 against four model organisms: the nematode Caenorhabditis elegans , the insects Aedes aegypti and Drosophila melanogaster , and the amoeba Acanthamoeba castellanii . The biotoxicity assays were performed either by feeding the test organisms with E . coli expressing the recombinant lectin as described previously [30] , or by adding the purified lectin to the food source of the organisms . These experiments showed a toxicity of CCL2 for C . elegans and D . melanogaster ( Figure 8 ) but not for A . aegypti or A . castellanii ( Figure S8 ) . In the case of C . elegans , feeding on CCL2-expressing E . coli stopped the development of all wildtype ( N2 ) L1 larvae in the assay ( Figure 8B ) . This toxicity was dose-dependent and the presence of 30% of CCL2-expressing E . coli among the fed bacteria was sufficient to reduce the development of more than 95% of the L1 larvae ( Figure S9 ) . In the case of D . melanogaster , CCL2 caused a significant delay in development of both pupae and flies by 4- and 10-fold , respectively , relative to the control ( Figure 8D ) . The toxicity of CCL1 towards C . elegans ( Figure S10 ) was found to be similar to that of CCL2 . The observed toxicity was likely to be mediated by binding of CCL2 to the N-glycan cores of glycoproteins in the susceptible organisms since α1 , 3-fucosylation of N-glycan cores was demonstrated both for C . elegans and D . melanogaster and caused cross-reactivity of anti-HRP antisera with these organisms [31] . Therefore , C . elegans mutants impaired in either fucose biosynthesis ( bre-1 ) or a variety of fucosyltransferases were tested for their susceptibility to CCL2-mediated toxicity ( see scheme in Figures 8A and B; Figure S11 ) . In agreement with our predictions , the bre-1 ( ye4 ) mutant that is unable to synthesize GDP-fucose was completely resistant to CCL2 intoxication . In addition , the fut-1 ( ok892 ) mutant lacking Fucα1 , 3 at the proximal GlcNAc of the chitobiose core [32] was partially resistant , as most of the worms survived and developed , but just half of the larvae reached L4 stage after 48 hours . On the other hand , a deletion in the fut-6 gene , which results in loss of tetrafucosylated N-glycans in C . elegans , as does a deletion in the fut-1 gene [32] , was as sensitive as N2 ( wildtype ) to CCL2 . In order to further explore these results , a fut-6 ( ok475 ) fut-1 ( ok892 ) double mutant was constructed and found to be completely resistant to CCL2 ( Figure 8B ) . As nematodes are able to α1 , 3-fucosylate both GlcNAc residues of the core region of some N-glycans [33] and both fut-1 and fut-6 are required for the full fucosylation of this core region ( see scheme in Figure 8A; Yan , Paschinger and Wilson , personal communication ) , our results suggest that the α1 , 3-fucosylated chitobiose core of N-glycans is the ligand of CCL2 in C . elegans . The partial resistance of the fut-1 mutant can be explained by binding of CCL1/2 to N-glycan cores carrying a single fucose on the distal GlcNAc ( Manβ1 , 4[Fucα1 , 3]GlcNAc ) . We hypothesize that this is a less favorable ligand due to the lack of an acetamido group on the mannose . To study the phenotype of CCL2-mediated intoxication and to follow the fate of the toxic lectin in the worms , different C . elegans strains were fed with E . coli cells producing an N-terminal fusion of CCL2 to the red fluorescent dTomato protein [34] . As can be observed in Figure 8C , a strong fluorescence was observed in the upper intestine of the completely susceptible worms N2 and fut-6 ( ok475 ) as a result of CCL2-binding to the intestinal epithelium . This fluorescence was accompanied by an evident damage of intestinal cells which resulted in a massive expansion of the intestinal lumen . In agreement with the effects on larval development ( Figure 8B ) , the fut-6 ( ok475 ) fut-1 ( ok892 ) double mutant that is resistant to CCL2-mediated intoxication , showed neither red fluorescence nor cell damage or expansion of the intestinal lumen . These results suggest that , in the absence of binding to the intestinal epithelium , the ingested lectin is completely cleared from the lumen after 1 hour . Accordingly , an intermediate phenotype , with some staining and cell damage , mostly in the upper part of the intestinal epithelium , was observed in the partially resistant fut-1 ( ok892 ) mutant . We evaluated the contribution of individual amino acid side chains on the carbohydrate-binding affinity by introducing several point mutations at the binding interface followed by ITC measurements . All variants expressed well ( except N91A ) and folded properly as judged from 15N-HSQC spectra ( Figure S12 ) . Significant decreases in affinity were observed for all mutants except N91A ( Table 1 and Figure S13 ) . The Y92A mutation decreased the affinity beyond the detection limit . The second largest affinity decreases are observed for W94A and W78A , indicating that the aromatic side chains provide the largest contribution to carbohydrate-binding affinity . A significant decrease in affinity was also observed for Y57A , L87A , N90A and V93A point mutants ( 4- to 17-fold ) . CCL2 variants were also tested in vivo for toxicity towards C . elegans . Remarkably , those mutants that retained carbohydrate binding with high affinity ( KD<30 µM ) in vitro were as toxic as wild type CCL2 . Mutants with lower in vitro affinity , however , showed a decreased toxicity towards C . elegans ( Figure 8E ) . In summary , these results confirm the carbohydrate-coordinating residues of CCL2 that were identified by NMR spectroscopy and suggest that high carbohydrate-binding affinity of the lectin is required for toxicity .
Our results strongly suggest that the newly identified lectins play a role in fungal defence . The lack of motility and the resulting inability of multicellular fungi and plants to escape from predators and parasites has led to the development of very similar defence strategies . In the absence of adaptive immune mechanisms and circulating immune cells , both types of organisms solely rely on innate defence . Whereas plant defence has already been intensively studied [9] , [35]–[37] , fungal defence has only recently been explored . It appears that , similar to plants , in addition to small molecules [38] , proteins play a key role in the defence of multicellular fungi , in particular against predators and parasites [39] . Among the different types of potential fungal defence proteins identified [13] , [26] , [40] , [41] the number and diversity of lectins is remarkably high , in accordance with the suitability of glycoepitopes for non-self recognition in innate defence mechanisms . Most fungal defence lectins are highly abundant in reproductive and long-term survival structures such as fruiting bodies and sclerotia , respectively , which require special protection [13] . This expression pattern , also found for CCL1 and CCL2 ( Figures 1B–C ) , is analogous to plants where the expression of many lectins is confined to seeds . The strong and specific toxicity of CCL1 and CCL2 towards D . melanogaster and C . elegans is in accordance with the prevalence of the phyla Arthropoda and Nematoda as predators of mushrooms both in nature [42] , [43] and in mushroom farms [44] , [45] . In addition , this specificity of CCL1/CCL2-mediated toxicity correlates with the identification of α1 , 3 fucosylated N-glycan cores as target structures of these lectins in vivo ( Figure 8 ) , since this epitope is present exclusively in plant and invertebrate N-glycans [31] . The NMR structure revealed that CCL2 recognizes the fucose-containing trisaccharide , GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc , as part of this epitope with high specificity . Within this trisaccharide , almost all functional groups of Fucα1 , 3GlcNAc and the acetamido group of the distal GlcNAc2 are recognized . The recognition of the distal saccharide is more relaxed , a GalNAc with an acetamido at the same position will be equally well recognized . Accordingly , among the glycans of the mammalian glycan array , GalNAcβ1 , 4[Fucα1 , 3]GlcNAc ( fucosylated LacdiNAc = LDN-F ) was one of the best binders . Since there is space for extensions at O6 of GlcNAc1 and O4 of GlcNAc2 ( Figure 5E ) we can derive the following recognition sequence: X-1 , 4GalNAc/GlcNAcβ1 , 4[Fucα1 , 3][Y-1 , 6]GlcNAc in which X and Y are tolerated extensions . In addition , binding of substituted LewisX structures on the glycan array ( Figure 2 ) suggests that substitutions at O3 and O6 of the galactose ( corresponding to the distal GlcNAc in α1 , 3 fucosylated chitobiose ) and at O6 of GlcNAc ( corresponding to the proximal GlcNAc in α1 , 3 fucosylated chitobiose ) are allowed . Accordingly , we would expect specific binding of CCL2 to paucimannose-type N-glycans carrying both α1 , 6 and α1 , 3-linked fucose on the proximal and possibly α1 , 3-linked fucose on the distal GlcNAc ( Figure 8A ) . The GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc motif is also a central part of the anti-HRP epitope that it is recognized by antisera raised against HRP in agreement with the isolation of CCL2 as HRP-binding lectin ( Figure 1A ) . Since this epitope is also a key carbohydrate determinant of pollen and insect venom allergens [46] , it appears that the same glycoepitope has been selected as target by the antibody-mediated mammalian adaptive immune system and a lectin-mediated fungal defence system . The high affinity of CCL2 to the recognized trisaccharide determined by ITC is remarkable . Typically , individual carbohydrate binding sites of lectins have a rather low affinity to their ligands and this low affinity is usually compensated by multivalency achieved either by multiple binding sites on the same polypeptide chain or by oligomerization of polypeptide chains with one or few binding sites which leads to a high avidity towards multivalent ligands [47] . However , high affinity carbohydrate binding sites of lectins have been described and they differ from low affinity binding sites by their degree of specificity [48]: whereas low affinity binding sites often have a broad specificity towards terminal mono- or disaccharides present on many different glycans , high affinity sites recognize distinct oligosaccharides that are characteristic for specific glycans and glycoconjugates . The high affinity and specificity of the carbohydrate binding site in CCL2 towards the recognized trisaccharide is achieved by H-bonds and key hydrophobic contacts to almost all functional groups of Fucα1 , 3GlcNAc as well as the acetamido group of the distal GlcNAc2 . The ladder interaction is central for the high affinity , the absence of the distal acetamido group as in Galβ1 , 4[Fucα1 , 3]GlcNAc ( LewisX ) leads to a drop in affinity by ∼300 fold ( in Table 1 ) . To our knowledge , CCL2 is the only lectin that binds GlcNAcβ1 , 4[Fucα1 , 3]GlcNAc with such a high specificity and affinity , making CCL2 superior to anti-HRP for detection of this glycoepitope . Since this and the other recognized glycoepitope , GalNAcβ1 , 4[Fucα1 , 3]GlcNAc ( LDN-F ) , are present in parasitic helminths [49]–[51] , CCL2 may be used for the diagnostics of parasitic infections in animals and humans . The toxicity of CCL2-binding to at least one of these epitopes in vivo , may be exploited to develop novel approaches for the prevention or therapy of these infections . Another application could be the use of CCL2 on lectin microarrays for differential glycan profiling [52] or cellular glycomics [53] . The NMR solution structure of CCL2 in complex with its ligand demonstrates the versatility and plasticity of the β-trefoil fold with regard to carbohydrate binding . First , the carbohydrate specificity of CCL2 is very different from other β-trefoil lectins which recognize terminal galactose epitopes like Galα1 , 3Gal [19] , Galβ1 , 3GalNAc [54] or Galβ1 , 3GlcNAc [21] , rather than an epitope with a terminal fucose . Second , unlike most β-trefoil lectins which utilize three almost identical binding sites per monomer , CCL2 recognizes the identified carbohydrate ligand via a single binding site . This binding site of CCL2 is located at a very unusual site of the β-trefoil fold , the interface between subdomains β and γ . This stands in contrast to the fungal β-trefoil lectin SSA that also uses a single but canonical binding site [54] . None of the typical carbohydrate binding residues present in other β-trefoil lectins are found in CCL2 emphasizing the uniqueness of this non canonical binding site ( Figure 6 ) . Based on few β-trefoil complexes in which the binding site is slightly shifted from the canonical towards the CCL2 location [55]–[58] we speculate that this non-canonical binding site might have arisen from a previous recognition of other parts of the invertebrate N-glycan by the canonical binding site β ( Figure 5G ) and then have changed to recognize another epitope of the same glycan by the non-canonical binding site . The key residues of the CCL2 binding site are highly conserved in CCL2 homologs of other fungi ( Figure 6 and Table S1 ) , but highly variable in other β-trefoil lectins . The unusual carbohydrate specificity is mainly based on H-bonds from the protein main chain which requires the proper arrangement of three main chain sections: most importantly the characteristically short β7–β8 loop , strand β6 and the β9–β10 loop . In particular , the short β7–β8 loop is conserved in all CCL2 homologues with a consensus sequence LPxxYVW , a signature we propose for the identification of lectins with a similar target specificity . In summary , based on sequence alignment we predict that the homologous CCL2 like genes of basidiomycetes have the same unusual binding location and the same target specificity as CCL2 ( except LB_L2 that lacks the crucial Y93 ) . As we do not have any evidence for a difference in regulation , specificity or function between the different paralogs , e . g . CCL1 and CCL2 , we speculate that this redundancy is a strategy to avoid loss of specific defense effectors by individual gene mutations . The strong toxicity of CCL2 towards C . elegans and D . melanogaster is surprising in the light of the monomeric state of the lectin in solution and the consequential lack of multivalency for the identified ligand since clustering of glycoconjugates on cell surfaces is generally regarded as a prerequisite for lectin-mediated toxicity [59] . CCL2 mutant proteins unable to bind the HRP epitope are not able to bind anymore to the C . elegans intestinal epithelium which rules out the presence of an additional binding site on CCL2 with different specificity for this tissue ( A . Butschi , unpublished results ) . Thus , we hypothesize that the high affinity of the single carbohydrate-binding site of CCL2 compensates for the lack of multivalency and that CCL2 acts by a novel toxicity mechanism that does not seem to involve clustering . Accordingly , CCL2 variants with a lower affinity in vitro showed a reduced toxicity in C . elegans . Remarkably , the consequences of intoxication of C . elegans by CCL1/2 and the multivalent fruiting body lectins MOA and CGL2 are very similar , all of them leading to disintegration of the intestinal epithelium and a substantial enlargement of the intestinal lumen ( Figure 8C ) [10] , [16] . In addition , experiments aiming at the localization of the target glycoconjugates using fluorescently labeled CCL2 and CGL2 gave very similar results ( Figure 8C ) [16] . Interestingly , disintegration of the intestinal epithelium and enlargement of the intestinal lumen were also observed with the nematode-specific Cry toxins from Bacillus thuringiensis where carbohydrate-dependent binding to the intestinal epithelium appears to trigger expulsion of microvilli from the apical side of the intestinal epithelial cells [60] . In any case , interference with carbohydrate binding by the lectin , either by mutating genes involved in the biosynthesis of the identified target glycans in C . elegans or altering the identified carbohydrate binding sites in the lectin , abolished toxicity and binding of the fluorescently labeled lectin to the intestinal epithelium ( Figure 8C ) [10] , [16] . It should be noted , however , that not all variants of CCL2 were tested for toxicity towards C . elegans and none was tested for toxicity towards D . melanogaster . Thus , although we can show that the recognition of specific glycans is a crucial part of lectin-mediated defence mechanisms , the exact mechanisms of toxicity remain to be elucidated . Possible mechanisms are direct membrane damage or the interference with cellular signaling pathways , recycling of cell surface receptors , cell-cell or cell-matrix interactions . In order to distinguish between these possibilities and to find potential targets of novel antihelminthics , we are currently in the process of identifying the glycoprotein ( s ) targeted by CCL2 and CGL2 in C . elegans .
LewisX trisaccharide methyl glycoside , 3′-Sialyl-LewisX tetrasaccharide methyl glycoside and Fucα1 , 3GlcNAc-OMe were purchased from Carbosynth , UK . The chemically synthesized fucosylated chitobiose GlcNAcβ1 , 4[Fucα1 , 3]GlcNAcβ-O ( CH2 ) 5COONa [61] was a kind gift of Mayeul Collot , ENS , France . LewisX tetrasaccharide and 3′-Sialyl-lactose were a kind gift of Eric Samain , CERMAV , France . The identity and purity of the carbohydrates was checked using 2D NMR spectroscopy . Detailed information of the strains used in this study can be found in Table S7 . Escherichia coli strain DH5α was used for cloning and amplification of plasmids , strains BL21 ( DE3 ) and BL21 ( DE3 ) /pLysS were used for bacterial expression of proteins and biotoxicity assays and strain OP50 was used to feed C . elegans during regular breeding . Cultivation conditions of the various organisms are described in Text S1 . CCL2 was isolated and purified from C . cinerea as described in Text S1 . Purified CCL2 was separated by SDS-PAGE , excised from the gel and identified by MALDI-MS/MS . Details of the procedure are described in Text S1 . Details of the quantification are described in Text S1 . The PCR-based cloning strategies for the various CCL1- and CCL2-encoding genes are described in Text S1 . Protein expression of CCL2 was evaluated by immunoblotting . Soluble protein extracts of vegetative mycelium and fruiting bodies from C . cinerea were obtained as described above and separated on a 12% SDS-PAGE and probed with specific antiserum raised in rabbits against purified recombinant CCL2 ( Pineda Antikörper-Sevice , Berlin , Germany ) and detected with HRP-conjugated secondary antibodies . Transcription levels of both genes were assessed by quantitative real-time PCR ( qRT-PCR ) as described in Text S1 . Purified CCL1 and CCL2 were fluorescently labeled with Alexa Fluor 488 ( Invitrogen ) according to the manufacturer's protocol and used ( at a final concentration of 200 µg/ml ) to probe versions 4 . 2 and 3 . 1 , respectively , of the mammalian glycan array offered by Core H of the Consortium for Functional Glycomics ( CFG ) . Unlabelled and uniformly 15N or 13C/15N labeled proteins were overexpressed in E . coli as His8-fusions and purified with affinity chromatography ( see Text S1 ) . Samples were dialyzed against NMR buffer ( 50 mM KH2PO4 , pH 5 . 7 , 150 mM NaCl ) . Complexes of CCL2 with GlcNAcβ1 , 4[Fucα1 , 3]GlcNAcβ-O ( CH2 ) 5COONa were prepared by titrating the concentrated carbohydrate solution of typically 10 mM into a ∼1 mM solution of CCL2 in NMR buffer until a 1∶1 stoichiometry was reached . Subsequently , the pH was lowered to 4 . 7 using 10% deuterated acetic acid to avoid precipitation . NMR spectra were acquired on Avance III 500 , 600 , 700 , 750 and Avance 900 Bruker spectrometers at 310 K . NMR data were processed using Topspin 2 . 1 ( Bruker ) and analyzed with Sparky ( Goddard , T . D . & Kneller , D . G . SPARKY 3 . University of California , San Francisco ) . The 1H , 13C , 15N chemical shifts of the protein , free and in complex , were assigned by standard methods [62] . Assignment of carbohydrate resonances of the complex was achieved using NOE correlations and exchange peaks with signals of the free carbohydrate since neither TOCSY based spectra nor a natural abundance 13C-HSQC showed bound carbohydrate signals . The following spectra were used for this purpose 2D 1H-1H NOESY , 2D 13C/15N F1-filtered NOESY and 2D 13C F1-filtered F2-filtered NOESY [63] . The assignments of intermolecular NOEs were derived from 3D 13C F1-edited , F3-filtered NOESY-HSQC [27] spectra of the protein-carbohydrate complex . More details are found in the Text S1 . The AtnosCandid software package [64] , [65] was used to generate initial CCL2 structures ( free and bound ) using three 3D NOESY spectra ( 13Cali-edited , 13Caro-edited and 15N-edited ) and one 2D NOESY spectrum . The automatically generated upper limit restraints file was used as a starting point for the first level of manually refining the protein structures by a simulated annealing protocol using the Cyana package [64] . Preliminary structures of the CCL2-carbohydrate complex were generated using the Cyana package with the above mentioned restraints and manually assigned intermolecular and intra-carbohydrate NOE distance constraints . To create the topology of the carbohydrate for the Cyana library file an initial model was generated by SWEET [66] . The carbohydrate spacer was truncated to a methyl group . 300 structures were generated by CYANA starting from random carbohydrate and protein starting structures . Ensemble of 30 structures of CCL2 free and in complex were refined with AMBER 9 . 0 [67] . in implicit solvent using NOE-derived distances , torsion angles and hydrogen bond restraints as summarized in Table 2 . For more details see Text S1 . The Ramachandran statistics of CCL2 free and in complex , respectively , show 79 . 9% and 80 . 2% in the most favored regions , 18 . 0% and 18 . 7% in the additionally allowed regions , 1 . 5% and 1 . 0% in the generously allowed regions and 0 . 6% and 0 . 2% in the disallowed regions . Biotoxicity assays for A . aegypti and A . castellanii were performed with recombinant E . coli as previously described [30] . For C . elegans , a liquid toxicity assay was performed as follows: a synchronous population of L1 larvae as well as a bacterial culture of recombinant E . coli expressing CCL2 or containing a vector control were obtained as described [22] . E . coli cells were pelleted and re-suspended in sterile PBS to an OD600 = 2 . The assay was set up in 96-well plates ( TPP ) by mixing 80 µl of the bacterial suspension and 20 µl of L1 larvae containing approximately 30 individuals . Each treatment ( different bacterial and/or worm strain combinations ) was done in 5 replicates . The worms were allowed to feed on the suspended bacteria at 20°C in the dark . The total number of animals and the percentage of individuals reaching L4 stage were quantified after 48 h . The biotoxicity assay with D . melanogaster was performed adding purified protein to the rearing medium as previously described [68] using 20 eggs . For the statistical analysis of the toxicity assays , pairwise comparisons were done using the non-parametric Kolmogorov-Smirnov test in the case of C . elegans , A . castellanii and D . melanogaster and the parametric T-student test for A . castellanii . The response variables ( development , survival and clearing area ) were compared between the tested lectin and the control or between mutant and wildtype . Details are described in Text S1 . More information is found in Text S1 . ITC experiments were performed on a VP-ITC instrument ( MicroCal ) . The calorimeter was calibrated according to the manufacturer's instructions . Protein and carbohydrate samples were dialyzed against NMR buffer at room temperature using a 3 . 5 kDa membrane ( Spectra/Por ) and Micro DispoDialyzer ( 100 Da cutoff; Harvard Apparatus ) , respectively . The disaccharide Fucα1 , 3GlcNAc-OMe was not dialyzed but dissolved in NMR buffer . The sample cell ( 1 . 4 mL ) was loaded with 70 µM protein; carbohydrate concentration in the syringe was 2–4 mM . A titration experiment typically consisted of 30–50 injections , each of 3 µL volume and 6 s duration , with a 6 . 7 min interval between additions . Stirring rate was 307 rpm . Raw data were integrated , corrected for nonspecific heats , normalized for the molar concentration , and analyzed according to a 1∶1 binding model . The atomic coordinates of the structures of CCL2 free and in complex with the fucosylated chitobiose ( GlcNAcβ1 , 4[Fucα1 , 3]GlcNAcβ-OMe ) have been deposited in the Protein Data Bank with accession codes 2LIE and 2LIQ , respectively . The chemical shifts of the free protein and in complex were deposited in the BioMagResBank ( BMRB ) under the accession numbers 17890 and 17902 , respectively . The cDNA sequences of CCL1 and CCL2 from C . cinerea strain AmutBmut were deposited in GenBank under accession number ADO87036 and ACD88750 , respectively .
|
All multicellular organisms have developed mechanisms to defend themselves against predators , parasites and pathogens . As a common mechanism , animals , plants and fungi use a large arsenal of carbohydrate-binding proteins ( lectins ) to protect themselves from predation and parasitism . The success of this type of innate defence mechanism critically depends on the diversity of specific recognition of foreign carbohydrate structures by the host lectins . In this study , we use NMR structure determination to show that part of this diversity is created by the plasticity of common protein folds . The identified fungal lectin that is toxic to nematodes and insects , adopts a common lectin fold but is remarkable in terms of its specificity and affinity for the recognized foreign carbohydrate structure , the number and location of the carbohydrate binding sites on the protein and the degree of oligomerization . Since the identified in vivo target of the fungal lectin is characteristic for invertebrates , our results may be exploited to develop novel approaches for the control of animal and human parasites .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biomacromolecule-ligand",
"interactions",
"microbiology",
"host-pathogen",
"interaction",
"carbohydrates",
"animal",
"models",
"fungi",
"caenorhabditis",
"elegans",
"model",
"organisms",
"drosophila",
"melanogaster",
"protein",
"structure",
"mycology",
"proteins",
"biology",
"biophysics",
"fungal",
"biochemistry",
"biochemistry",
"computational",
"biology",
"glycobiology",
"defense",
"proteins",
"macromolecular",
"structure",
"analysis"
] |
2012
|
Plasticity of the β-Trefoil Protein Fold in the Recognition and Control of Invertebrate Predators and Parasites by a Fungal Defence System
|
Evidence has accumulated in recent decades on the drastic impact of climate change on biodiversity . Warming temperatures have induced changes in species physiology , phenology , and have decreased body size . Such modifications can impact population dynamics and could lead to changes in life cycle and demography . More specifically , conceptual frameworks predict that global warming will severely threaten tropical ectotherms while temperate ectotherms should resist or even benefit from higher temperatures . However , experimental studies measuring the impacts of future warming trends on temperate ectotherms' life cycle and population persistence are lacking . Here we investigate the impacts of future climates on a model vertebrate ectotherm species using a large-scale warming experiment . We manipulated climatic conditions in 18 seminatural populations over two years to obtain a present climate treatment and a warm climate treatment matching IPCC predictions for future climate . Warmer temperatures caused a faster body growth , an earlier reproductive onset , and an increased voltinism , leading to a highly accelerated life cycle but also to a decrease in adult survival . A matrix population model predicts that warm climate populations in our experiment should go extinct in around 20 y . Comparing our experimental climatic conditions to conditions encountered by populations across Europe , we suggest that warming climates should threaten a significant number of populations at the southern range of the distribution . Our findings stress the importance of experimental approaches on the entire life cycle to more accurately predict population and species persistence in future climates .
Over the last decades , consequences of global warming on biodiversity have become obvious [1–3] , with many species likely to be committed to extinction by 2050 [4] . Climate warming has already led to changes in species phenology [1] , physiology ( increased metabolic rates [5] ) , morphology ( shrinking body size [6] ) , life cycle demography [7] , and distribution [1] , and , as a consequence , in community structure [8] . Because their body temperature , and hence their basic physiological functions , directly depend on environmental conditions , ectotherms are particularly at risk with climate change [5] , while the number of studies assessing their response to changing climate is far lower than for endotherms [9] . The evaluation of their vulnerability is therefore urgent . For instance , a recent study predicted local extinctions of populations from various lizard families worldwide to reach 39% by 2080 due to climate change [10] . Theoretical studies predict that climate change will principally threaten tropical ectotherms [11–14] , while temperate ectotherms should resist or even benefit from the warmer temperatures [13 , 15–17] . However , most evidence on the impacts of climate change on species comes from long-term field survey data [1 , 8] , or on the contrary , on short term laboratory experiments lacking ecological realism and complexity [18–20] . Despite the growing evidence on the strong impact of ecological context on species adaptation to temperature [21] , there is little large scale realistic experimental evidence on animals , especially on vertebrates [20 , 22–25] . More importantly , to our knowledge , the impact of climate change on a species’ entire life cycle and population persistence has never been experimentally tested on a vertebrate [26] . This information gap hinders the prediction of future impacts , because unraveling the impact of predicted climate on different demographic parameters is essential for the precise estimation of extinction probability [27 , 28] . The Intergovernmental Panel on Climate Change ( IPCC ) predicts a global temperature increase between +0 . 3 and +4 . 8°C over the next century , depending on the CO2 emission scenarios [29] . Experimental studies should thus implement realistic IPCC climate change projections relying on several greenhouse gas emission scenarios and describe population responses to said scenarios in large field experiments [24 , 25] . Here , we studied the effect of a warmer climate on the life cycle and demography of a lizard species with large-scale experimental mesocosms ( Fig 1 ) . Using common lizards ( Zootoca vivipara ) as model species , we aimed to determine whether predicted temperature increases will be detrimental or beneficial to temperate lizards and to identify the key parameters involved in potential declines of populations , especially in populations at the southern margin of the distribution area . To that end , we took advantage of an innovative experimental facility , the Metatron , a system with large seminatural enclosures in which climatic conditions can be manipulated ( Fig 1 ) [30] . We created 18 lizard populations in the Metatron over two years of experiment ( 2012: 8 populations , 2013: 10 populations ) and allocated them to one of two climatic treatments throughout the summer: “present climate” ( existing local area climate ) and “warm climate” ( ~2°C warmer than ambient temperatures ) , coherent with IPCC climate change projections for the end of the century ( global temperature increase projections for a midrange emission scenario , Representative Concentration Pathway ( RCP ) 4 . 5: +1 . 8 ± 0 . 5°C [29] ) . We investigated adult and juvenile survival , body growth , and female reproduction to estimate the effect of warmer climatic conditions on lizard life history and population growth rate . We further compared our results to climatic conditions across Europe to inform predictions about more general fate of European lizard populations .
Warm climatic conditions had a strong positive impact on juvenile body growth ( Table 1 , Fig 2a ) but had no effect on body condition ( Table 1 ) . Warm climates also led to an earlier reproductive onset in these juveniles . Indeed , female juveniles from the “warm summer climate” populations were more likely to reproduce the following spring ( Table 1 , Fig 2b ) . This accelerated reproductive onset was likely due to the higher individual body growth rate , as female body size in May had a significant impact on probability of gravidity ( Likelihood ratio test , χ² = 24 . 9 , p < 0 . 001 ) . There was no overall effect of climate treatment on annual survival ( Table 1 ) , although juveniles from the “warm climate” treatment tended to survive less during the summer ( S3 Table ) . Warmer climate was detrimental for the survival of older individuals . The annual survival of adults and yearlings was lower in “warm climate” environments ( Table 2 , Fig 2c ) , and this effect was mainly due to a difference of survival during the summer ( S3 Table ) . Warmer climatic conditions had , however , a positive impact on the body condition of adults that survived ( Table 2 ) , while there was no impact of climatic conditions on individual growth rate . We further found a tendency for an earlier laying date in adult females from the “warm climate” enclosures ( Table 2 ) . Moreover , we found out that some females had produced a second clutch during the summer 2012 . Twelve neonates , from five females , hatched in the “warm climate” enclosures during the summer ( Fig 2d ) , while we did not find such neonates in “present climate . ” These neonates were born from a second clutch of these females . We modeled the impact of our climatic treatment on lizard population dynamics with an age-structured Leslie matrix fitted with the survival and reproduction parameters obtained from our field experiment ( S3 Text , S3 Fig , S4 Table ) . Population growth rate in “warm climate” environments was very low ( λ = 0 . 75 [0 . 72 , 0 . 77] , mean [95% CI] , results for a deterministic model ) , while populations in “present climate” environments were maintaining themselves ( λ = 0 . 98 [0 . 95 , 1 . 01] , confidence interval crossing 1 ) . As a consequence , populations in warm climates should go extinct rapidly ( years to extinction , mean [95% CI] , warm climate = 22 y [20 , 24] , present climate = 298 y [118 , no extinction] ) . Using a stochastic model yielded very similar results ( S3 Text ) . We compared maximum daily temperatures in common lizard populations across Europe to maximum daily temperatures experienced by lizards in our experimental setup to categorize populations into “risk profiles” ( S4 Text , S6 Table ) . We showed that under a 2°C temperature increase scenario , a significant number of European populations , mostly at the southern margin of the distribution , may be at risk from warming climates . Fourteen percent of European populations may be threatened in the future if temperature increases by 2°C ( Fig 3 , S4 Text , risk levels A to C ) . Moreover , if temperature rises by 3°C , 21% of the populations might be at risk in the future ( Fig 3 , S4 Text , risk levels A to D ) . Additionally , comparing with a survey done by Sinervo et al . [10] on European populations of common lizards , we found that populations classified by the authors as nearly extinct or extinct fell significantly more within our “at risk” profiles than populations classified as maintaining themselves ( χ² = 7 . 8 , p = 0 . 005 , S4 Text ) . Risk profile projections depend on the demographic parameters obtained from our experiment , and as such should be sensitive to differences in demographic parameter estimates in the natural populations , particularly on changes in adult and juvenile survival rates ( S3 Text , S5 Table ) , as well as on uncertainty in climatic data observations .
We found that warmer climatic conditions strongly modified lizard's life history . On one hand , warm climatic conditions had a strong positive impact on juvenile body growth . In ectotherms , a difference of 2°C , as generated in our experiment , can largely increase metabolic rate [5] and hence energetic needs . When juveniles can compensate for this increased metabolism by foraging more , it should lead to a faster body growth rate . Invertebrate diversity and abundance were high in enclosures , and there was no difference between climatic conditions ( p > 0 . 55 , S1 Text , Material and Methods ) . Juveniles could thus compensate by foraging more , resulting into a faster growth rate with subsequent consequences on their entire life history . For instance , reproduction is size-dependent in reptiles [31] and should be favored by an accelerated individual growth . Indeed , female juveniles from the “warm climate” populations were more likely to reproduce the following spring because of the fast summer individual growth rate . Such results are consistent with patterns observed in natural populations , as body size and individual growth rate were shown to increase with temperature in common lizard populations [32–35] , while age at first reproduction depended on body size [36] . Warm climatic conditions were therefore mostly beneficial at juvenile stages as juvenile survival during the summer was only slightly decreased in our experimentally warmer climates , with overall 30% of juveniles from all populations surviving their first year , as in natural populations [33 , 37] . On the other hand , a warmer climate was mostly detrimental for older individuals . Only 42% of adults and yearlings from the “warm climate” treatment survived after one year , while 52% survived in present climates , comparable to survival rates found in natural populations from France , Belgium , and the Netherlands [33 , 37–39] . One explanation for this difference could be a summer heat stress , daily temperatures surpassing lizard critical thermal maximum . However , this view was not supported , as temperature only rarely surpassed critical thermal maximum ( CTmax = 40°C [40] ) , and as a large temporal and spatial thermal heterogeneity within enclosures allowed lizards to find cool conditions during warm hours in both climates ( S2 Text , S1 Table ) . In addition , climatic conditions had no effect on juvenile survival , while juvenile individuals likely have lower CTmax , as in other lizard species [41 , 42] . A second , more likely hypothesis could be linked to metabolic costs [43] . In ectotherms , metabolic rate scales positively with body size and temperature [44] . Warmer temperatures should increase energetic needs that cannot be fully compensated by an increase in foraging , in particular when warming induces restriction of lizard activity period , as suggested by a recent study [10] ( but see [45] ) . This explanation may also explain the discrepancy of effects between ages as the rise in energetic needs in smaller individuals ( i . e . , juveniles ) may be more easily compensated by foraging . However , adult body condition did not decrease in warmer conditions over the summer ( S3 Table ) and even increased after the winter in surviving individuals from the warm climate ( Table 2 ) . As the better spring body condition can be explained by a lower lizard density and thus competition for food during the spring ( impact of lizard density on adult body condition , Likelihood ratio test , χ² = 5 . 91 , p = 0 . 02 ) , our energetic needs hypothesis may still explain our results and would concur with previous results on fish [46 , 47] and marine invertebrates [48] . In these studies , juveniles and smaller individuals survived better in higher temperatures than larger ones , which were failing to meet overall energy demands [46] . On top of energy demands , a warming-accelerated metabolism and foraging could change various physiological parameters ( e . g . , increased oxidative stress [49 , 50] ) leading to physiological exhaustion and mortality in adults only [50 , 51] . A last possibility is that our climatic treatments , mainly set in the summer , generated a mismatch between summer and winter temperatures , hence increasing mortality during the winter . Because adult mortality during the summer was already affected by climatic treatment ( S3 Table ) , it seems however unlikely that a temperature mismatch between summer and winter temperature could be the sole cause of the observed mortality increase . These negative impacts of a warmer summer climate on adult life expectancy could be balanced by a higher investment in reproduction . In this species , reproduction occurs once a year in the spring , but summer climatic conditions could change reproductive investment during the following spring . Although this change was not observed , we found out some females produced a second clutch during the summer of climate manipulation . Twelve juveniles , from five females , hatched in the “warm climate” enclosures during the summer 2012 , while we did not find any neonates in “present climate . ” These findings are surprising , as in natural populations viviparous common lizards have never been observed to reproduce twice a year [37 , 39] , although oviparous common lizards can produce second clutches [36] . Increased voltinism due to climate warming has been recently demonstrated in butterflies [52] , and , in multivoltine lizards ( Uta stansburiana ) , bivoltinism frequency was shown to increase with nocturnal temperature [53] . However , this is the first study to our knowledge showing that a univoltine vertebrate can shift to multivoltinism due to environmental conditions . Nevertheless , second clutches were too rare to balance the drop in survival rate ( S3 Text ) . Together with an earlier onset of reproduction and a decrease in adult survival , these results suggest an acceleration of common lizard population turnover as a response to climate warming . Theoretical studies demonstrate that warming can accelerate metabolic and demographic rates in ectotherms [3] . Our work provides the first experimental evidence of such demographic acceleration , which should in turn change population dynamics and persistence [54] . Indeed , the earlier onset of reproduction of young females in warmer conditions was not sufficient to compensate for the drop in adult and yearling survival at these temperatures . As population growth rate was more sensitive to survival rates than to yearling fecundity ( S3 Text , S5 Table ) , populations in a warm climate were predicted to go extinct in around 20 y , while populations in a present climate maintained themselves ( λ = 0 . 98 , 95% CI for λ crossing 1 , [0 . 95 , 1 . 01] ) . These predictions are made even worse by the absence of warming enhanced dispersal movements ( S3 Table ) , which could allow individuals to track their climatic niche [55] , but see [56] . When we compared climatic conditions in our experiment to conditions encountered by common lizard populations across Europe , we found that several populations at the southern margin of the distribution should be at risk from climate warming in the near future , while populations at the northern margin should not be threatened ( Fig 3 , S4 Text ) . Considering a scenario of around +2°C temperature increase by the end of the century ( consistent with RCP 4 . 5 greenhouse gas emission scenario [29] ) , we showed that 14% of populations surveyed should be threatened by the climate change in the next century , 11% in the very near future ( around 2050 , S4 Text ) . If we consider a higher temperature increase of 3°C , which could be attained with RCP 6 . 0 high emission scenario , the percentage of threatened populations went up to 21% , and with a very high temperature increase scenario of 4°C , possible under RCP 8 . 5 emission scenario , it attained 30% of the populations ( S4 Text ) . Moreover , we showed that two European populations , located at the extreme southern margin of the distribution , might already be threatened under the current levels of temperature ( S4 Text , “imminent risk” profiles ) . Finally , we found that nearly extinct or extinct populations from Sinervo et al . survey [10] were more likely to fall within our “at risk with 2°C increase” profiles than populations found to maintain themselves , confirming that temperature was probably one of the main drivers of the observed extinctions . Further modeling on range dynamics and extinction risks of Z . vivipara and other lizard species should use spatially explicit demographic models ( e . g . , [28 , 57 , 58] , but see [59] for a review of available methods ) informed by our experimental results as well as by data from field surveys ( e . g . , density and demographic parameters ) , to draw a better picture of the impacts of climate change on lizard population and range dynamics under several greenhouse gas emission scenarios . Overall , we showed that lizard populations at the southern margin of their distribution should be particularly sensitive to a warming climate , leading to potential population extirpations and a shrinking of lizard’s range , while populations at higher latitudes should not be threatened . The limitation of a species range has been attributed to two interacting factors , abiotic conditions such as temperature and hygrometry and biotic conditions such as competitive interactions [60 , 61] . Our results support the idea that common lizards range is limited in the south by abiotic conditions due to the climate-dependent species demography . Our study demonstrates for the first time a change in life history tactics due to a 2°C climate warming , with an acceleration of the pace of life and generation turnover . This acceleration was associated with a decrease in population density , which could lead to the extinction of common lizard populations at the southern margin of their distribution . Previous studies on natural populations of common lizards showed that the current rate of warming had rather positive effects on populations [32 , 35] , mostly because they found either no effect or positive effects of warmer spring temperature over the past 20 y on body growth rate and/or survival , with the exception of one study showing slightly negative relationships between temperature and survival in some populations [38] . However , the effect of temperature is unlikely to be linear , and thermal physiology of ectotherm species suggests a threshold of temperature above which performance decreases steeply [62] . Our simulated warming matches the summer temperatures predicted for the end of the century and could exceed a threshold where thermal conditions shift from beneficial to detrimental for adult survival . If the trend of temperature increase follows IPCC predictions , we can predict demographic accelerations in ectotherm species . The functioning of communities strongly depends on the fine tuning of species interactions , and changes in species pace of life can destabilize community assemblages and hence induce their extinction [8 , 63] . Using a model species , our findings emphasize that climate change is not only a problem for tropical ectotherms [11–14] but , contrary to more optimistic predictions [13 , 15–17] , it could endanger temperate ectotherms with population extirpations and a shrinking of their range of distribution by the disappearance of southern populations . In species with a restricted range distribution , such population extirpations could ultimately lead to species extinctions if these species are unable to adapt to warmer climates . Now , we should therefore study how species can adapt to future climatic conditions through phenotypic and phenological modifications . For instance , a selection for an earlier onset of reproduction and an increased voltinism might allow species to shift towards a faster life history strategy and populations to be rescued by compensating lower adult survival rates [64–66] . However , such acceleration of population turnover might , on the contrary , be detrimental . For instance , in a European butterfly ( Lasiommata megera ) , an increased voltinism led populations into a developmental trap where individuals attempted third generations , resulting in higher mortality and the loss of the third generation [67] . Future experiments should therefore simulate future warmer climates on several generations to study species adaptiveness and persistence .
The common lizard ( Z . vivipara; Jacquin 1787 ) is a small ( adult snout–vent length 50–70 mm ) viviparous lacertid lizard inhabiting dense vegetation patches across Europe and Asia . Common lizards have been extensively studied for their biology and population dynamics ( e . g . , [32 , 35–37 , 68–70] , S6 Table ) , making them a good model species to study the consequences of climate change on temperate lizards . Lizards hibernate from October to March in our study site ( Ariège , France ) , and mating occurs right after emergence . After approximately two months of gestation , females lay on average five ( range 1–12 ) soft shelled eggs . Juveniles emerge within one hour after laying and are immediately independent [37] . The lizards used in this study were captured in 2010 from natural populations of the Cévennes mountains ( Lozère , France , 44°27' N , 3°44' E , Licence no . 2010-189-16 DREAL ) , marked by toe clipping , and translocated to the Metatron , an infrastructure composed of seminatural caged enclosures located at the Station of Experimental Ecology in Moulis ( Ariège , France , 43°01' N , 1°05' E ) . This unique structure offers 48 interconnected enclosures , each measuring 10 x 10 m , containing natural lizard habitat ( dense vegetation , hiding places , and rocks [30 , 56 , 71–76] , Fig 1 ) . Each enclosure is delimited by tarpaulins buried 30 cm into the ground , preventing escape and terrestrial predation [30] , and are fully enclosed with a net preventing avian predation and allowing isolation of each enclosure ( Fig 1 ) . Each enclosure acts as a mini ecosystem , with natural vegetation and insect communities and a relatively wide variety of thermal microhabitats ( shaded , dense , and diverse vegetation , sun-battered rocks and logs , and ponds , Fig 1 ) . Diversity within these caged habitats is relatively high , with more than 140 vegetal species found within the enclosures for 134 species found in the nearby outside habitat ( estimated in May 2014 ) . Considering invertebrate communities , a monitoring allowed to determine more than 123 invertebrate families present in the enclosures against only 106 in the nearby outside habitat ( S1 Text , estimated in May 2014 ) . Enclosures can be connected to a 19-m-long one-way corridor with a pitfall trap at the end ( Fig 1 ) . This distance corresponds to the minimum dispersal distance of the common lizard [77] . Finally , temperature , illuminance , and hygrometry within each enclosure are monitored every 30 min and can be manipulated through the use of motor-driven shutters and a sprinkler system . Lizards were maintained in the Metatron for two years prior to the experiment in “present climate” conditions ( see next section ) using similar population densities and structures than in this study . Between May 2012 and May 2014 , we performed two studies manipulating summer climatic conditions and monitoring consequences on lizard populations . We used data from these two years of experiment altogether . The same experimental procedure was used for the two years . From mid-May , at the end of female gestation period , we captured all surviving lizards maintained in the Metatron during multiple successive capture sessions . Each lizard was measured for snout–vent length and total length and weighted . A tail tip was taken for routine genetic sampling . Yearlings ( 1-year-old lizards ) and adult males were kept only for the amount of time necessary to ensure that we had captured all surviving individuals from the enclosures and were released into the Metatron on average one month after capture , whereas females were maintained in the laboratory until parturition . In the laboratory , lizards were kept in 25 x 15 . 5 x 15 cm individual glass terraria with a 3 cm litter layer , a piece of cardboard and a plastic tube for shelter and a piece of absorbent paper . A light bulb ( 25 W ) and an ultraviolet lamp ( Zoomed Reptisun 5 . 0 UVB 36 W ) provided heat for thermoregulation and light 6 h per d ( from 9:00 to 12:00 and from 14:00 to 17:00 ) . Lizards were lightly sprayed with water three times a day ( in the morning , at mid-day , and in the evening ) and offered one cricket ( Acheta domestica ) daily . Between early June and mid-July , females laid eggs in the terraria . Offspring were marked and measured for body length ( snout–vent length and total length to the nearest mm ) and mass ( to the nearest 0 . 001 g ) immediately after birth; their sex was determined by counting ventral scales [78] , and a tail tip was taken for genetic sampling . Families were then released into the Metatron . Lizards were released into the Metatron controlling for body size and source population . From June to the end of September ( 2012 and 2013 ) , we applied several climatic treatments to the enclosures . In 2012 , we created nine populations from three climatic treatments ( three populations in each treatment ) , while in 2013 we created ten populations from the two extreme climatic treatments ( five populations in each treatment ) . Enclosures were chosen to be the most homogeneous respective to the vegetal cover ( F2 , 6 = 0 . 80 , p = 0 . 49 and F1 , 8 = 0 . 54 , p = 0 . 48 , respectively for 2012 and 2013 ) , vegetal height ( F2 , 6 = 2 . 26 , p = 0 . 18 and F1 , 8 = 0 . 04 , p = 0 . 85 , respectively for 2012 and 2013 ) , vegetal composition ( F2 , 6 = 0 . 01 , p = 0 . 99 and F1 , 8 = 3 . 16 , p = 0 . 11 , respectively for 2012 and 2013 ) , and invertebrate prey diversity ( F2 , 6 = 0 . 91 , p = 0 . 45 and F1 , 8 = 2 . 60 , p = 0 . 15 , respectively for 2012 and 2013 , see S1 Text ) . In 2012 , we had a “present climate” ( PC ) in which automatic shutters were allowed to close when temperature exceeded 28°C , an “intermediate climate” level , in which shutters closed when temperature surpassed 34°C and a “warm climate” ( WC ) in which shutters were only allowed to close when temperature rose above 38°C . In 2013 , we only kept the present and warm climate treatments because the intermediate treatment had similar temperatures and gave similar results to the warm climate treatment . Enclosed habitats are warmer than outside habitats . Closing the shutters both stopped temperature from rising and caused temperatures to drop , evening out temperature peaks . As a result , “present climate” enclosures showed similar summer temperatures to ambient temperatures outside of the Metatron ( temperatures in the nearby meteorological station of Saint-Girons Antichan , S2 Text ) , while “warm climate” enclosures were on average 2°C warmer ( e . g . , mean daily temperatures between mid-June and mid-September 2012 and 2013 , PC: 26 . 4 ± 0 . 3°C , WC: 28 . 3 ± 0 . 3°C , mean ± SE , F1 , 282 = 23 . 1 , p-value < 0 . 001; maximum daily temperatures: PC: 29 . 2 ± 0 . 3°C , WC: 32 . 1 ± 0 . 3°C mean ± SE , F1 , 282 = 50 . 6 , p-value < 0 . 001 , see S2 Text , S1 Fig , S2 Fig ) . Our treatments generated significant differences over the summer in temperature and illuminance , but not in hygrometry ( S2 Text , S1 Table ) , while the treatment effects were negligible during the winter and the spring ( S2 Table ) . Such temperature differences are coherent with IPCC climate change projections for southern Europe [29] , which predicts a 3°C temperature increase by 2080 , with the largest warming during the summer . Indeed , projections from RCP 4 . 5 scenario ( an emission stabilization scenario ) in southern Europe predict a temperature increase of between 1 . 2 and 5 . 5°C between June and August against −0 . 2 and 3 . 0°C between December and February [29] . Thanks to a dense and diverse vegetation , there was a large temporal and spatial variation within enclosures of warm and present climate allowing cooler refuges despite an overall warmer environment ( S2 Text ) . In 2012 , we only had three enclosures in the intermediate climate level , and in one of them , a technical problem ( important disturbance in the enclosure related to maintenance issues of the Metatron ) caused a quasiextinction of a population . Moreover , when we compared summer temperatures between “intermediate” and “warm” climate treatments , we did not find significant differences in mean , maximum , or minimum temperatures ( S2 Text ) . Hence , we decided to exclude the data from the quasiextinct enclosure and merge the data from the two remaining intermediate climate enclosures to the warm climate enclosures for the analyses . Each year , populations were composed of 11 ± 1 adult females , 6 ± 1 adult males , 9 ± 2 yearlings and 38 ± 4 juveniles . These population densities conform with local densities observed in natural populations [37 , 79] and in other seminatural experiments on common lizards [68 , 71 , 77 , 80–82] . There was no difference between treatments in juvenile birthdate , in individual snout–vent length , or mass at release ( p > 0 . 36 for all ) . In mid-July , one-way corridors between enclosures were opened to allow lizard dispersal from enclosures . A pitfall trap at the end of each corridor allowed the capture of dispersing individuals . Dispersing individuals were measured , weighed , and released into another enclosure at random . In mid-September , we performed three capture–recapture sessions to measure lizard body growth and survival in each enclosure . In these three sessions , we were able to capture 93% of survivors ( capture probability estimated by MARK version 6 . 1 [83] ) . All surviving lizards were measured for snout–vent length and total length , weighed , and released into their enclosure to hibernate in the Metatron . During these capture sessions in 2012 , we caught 12 neonate juveniles born in the enclosures during the summer . A tail tip was taken from these individuals to assess maternity and paternity . Finally , the following spring , we recaptured all surviving lizards from each enclosure during multiple capture sessions ( >10 ) without release and brought them into the laboratory . All surviving lizards were measured and weighed again and kept in similar conditions as described above until female parturition , allowing assessment of female reproductive success . Genomic DNA of females and neonate juveniles was extracted from tail tips using the QIAquick 96 Purification Kit ( QIAGEN ) according to the manufacturer’s instructions after a digestion of tissue samples with proteinase K . Individuals were genotyped using eight microsatellite markers [78] . We checked for perfect match between juveniles and their assessed maternities ( no mismatch between female and juvenile ) using CERVUS software , v . 3 . 0 ( see [78] for details on methodology ) . The 12 neonate juveniles born in the Metatron during the summer 2012 were assigned to five females . These females had already produced a first clutch during their stay in the laboratory in June , and neonates found were born from a second clutch during the summer . We modeled the effect of climatic treatment on individual dispersal probability , survival probability , body growth ( difference between snout–vent length at release at the beginning of the experiment and snout–vent length at capture ) , body condition ( residuals from a linear model of body mass by body length ) , and finally on female probability of gravidity ( probability that a female will lay eggs ) , clutch size ( number of viable offspring laid by a gravid female ) , and laying date ( treated as a continuous variable ) . We analyzed juvenile data separately from adult and yearling data , since this allowed us to include a family effect in the analysis concerning juveniles , as siblings cannot be considered as independent . We first analyzed dispersal propensity , then we excluded dispersing individuals from the latter analyses , as dispersing individuals could not be assigned to a unique temperature treatment for the whole summer period . For survival probability , body growth rate , and body condition , we analyzed effects of climatic treatments over a year . However , we also provide in S3 Table the effects of treatment by the end of summer in order to better understand paths of effects . To estimate the effect of temperature treatment on juvenile , yearling , and adult demography , we performed generalized mixed models and linear mixed models with lmer procedure [84] in R , version 3 . 1 . 1 [85] . Dispersal , survival , and probability of gravidity were modeled using a generalized mixed model with a binomial distribution and a logit link . Body growth , body condition , and laying date were modeled as linear variables . Finally , clutch size was modeled using a Poisson distribution , except for clutch size in September 2012 where we used a zero-inflated Poisson GLM because of the low number of neonate juveniles recovered in September 2012 . Models included temperature treatment as a categorical variable and several covariates plus random intercepts . For juveniles , we included birthdate modeled as a continuous covariate , and for adult and yearlings , we included age modeled as a two-level factorial variable ( yearling or adult ) and sex . Finally , mixed modeling allowed adding random intercepts to the models: 1 ) a family effect in juvenile analyses , as juveniles from a family are not independent , 2 ) enclosure identity to account for variation due to potential differences among enclosures , and 3 ) the year of experiment to account for the block design . Following Zuur et al . [86] , we fitted full models with all fixed variables and every combination of random intercepts with a restricted maximum likelihood approach . We compared models using the respective AIC and chose the best structure of the random component for each dependent variable . We compared a full model with temperature treatment , necessary covariates , and random intercepts to a model including only the covariates and random intercepts through their ΔAIC . We then performed likelihood ratio tests to evaluate the impact of the temperature treatment . We provided estimates and standard errors of the effect of each fixed variable . We further calculated both the marginal ( effect of the fixed variables ) and the conditional ( effect of the fixed and random variables ) R² , as well as the PCV for each random variable following Nakagawa and Schielzeth [87] . Adults survived less in warmer conditions; hence we tested for the impact on adult density in September on juvenile survival and body growth . Similarly , juveniles grew more in warmer conditions , thus we also tested for the impact of their body growth in September on winter survival . Finally , we checked that shifts in invertebrate communities due to warming climates could not explain the lower adult survival in warm climate enclosures . There were no differences between warm and present enclosures in the number of insect families ( F1 , 17 = 0 . 37 , p = 0 . 55 ) , or in the density of insects ( F1 , 17 = 0 . 17 , p = 0 . 69 ) or arachnids ( F1 , 17 = 0 . 02 , p = 0 . 89 ) after one year; therefore , it was unlikely that differences in prey availability could lead to differences in survival . Nevertheless , we tested the impact of insect density the following year on adult survival and on juvenile survival and body growth .
|
Ongoing climate change has potentially drastic impacts on biodiversity . Because their body temperature depends on their external environment , ectotherm ( “cold-blooded” ) species are thought to be more at risk from warming climates than endotherm ( “warm-blooded” ) species that regulate their temperature internally . Tropical ectotherms should be particularly threatened by climate change , while temperate ectotherms should resist or even benefit from higher temperatures . While most of the evidence on the impacts of climate change comes from long-term field studies , experimental evidence of the impact of future climatic conditions is still lacking . Here we investigate the impacts of future climates on a temperate lizard using a seminatural warming experiment . We find that warmer temperatures led to a highly accelerated life cycle and a decrease in adult survival . As a result , we postulate that populations in such warm climates would be expected to go extinct in around 20 y . Comparing our experimental conditions to climatic conditions in European populations of common lizards , we show that many populations should be threatened in the next century , particularly in Southern Europe . Our findings challenge the optimistic view that climate change is only a threat for tropical ectotherms and stress the importance of experimental approaches to predicting the consequences of future warming trends .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Live Fast, Die Young: Experimental Evidence of Population Extinction Risk due to Climate Change
|
As a deubiquitinating enzyme ( DUB ) , the physiological substrates of ataxin-3 ( ATX-3 ) remain elusive , which limits our understanding of its normal cellular function and that of pathogenic mechanism of spinocerebellar ataxia type 3 ( SCA3 ) . Here , we identify p53 to be a novel substrate of ATX-3 . ATX-3 binds to native and polyubiquitinated p53 and deubiquitinates and stabilizes p53 by repressing its degradation through the ubiquitin ( Ub ) -proteasome pathway . ATX-3 deletion destabilizes p53 , resulting in deficiency of p53 activity and functions , whereas ectopic expression of ATX-3 induces selective transcription/expression of p53 target genes and promotes p53-dependent apoptosis in both mammalian cells and the central nervous system of zebrafish . Furthermore , the polyglutamine ( polyQ ) -expanded ATX-3 retains enhanced interaction and deubiquitination catalytic activity to p53 and causes more severe p53-dependent neurodegeneration in zebrafish brains and in the substantia nigra pars compacta ( SNpc ) or striatum of a transgenic SCA3 mouse model . Our findings identify a novel molecular link between ATX-3 and p53-mediated cell death and provide an explanation for the direct involvement of p53 in SCA3 disease pathogenesis .
Spinocerebellar ataxia type 3 ( SCA3 ) , also known as Machado–Joseph disease ( MJD ) , is an autosomal-dominantly inherited ataxia and one of at least nine polyglutamine ( polyQ ) neurodegenerative disorders described so far [1–3] . SCA3 is caused by an unstable cytosine-adenine-guanine ( CAG ) trinucleotide expansion mutation in the ATXN3 gene leading to an expanded polyQ tract within the ataxin-3 ( ATX-3 ) protein [4] . As a deubiquitylase , ATX-3 is highly conserved and ubiquitously expressed in cells throughout the body [5] . ATX-3 knockout ( KO ) mice have no major abnormalities [6] . It is possible that , besides ATX-3 , three other members of the MJD family of cysteine proteases , including ATX-3 Like , JosD1 , and JosD2 [7] , may exert similar functions to ATX-3 and compensate for its absence in KO models . ATX-3 has a structured N-terminal Josephin domain comprising the catalytic site , two ubiquitin ( Ub ) -binding sites , and an unstructured C-terminal , which contains two or three Ub-interacting motifs ( UIMs ) flanking a polyQ tract [8 , 9] . The expansion of the polyQ tract is thought to trigger a pathogenic cascade , leading to cellular dysfunction and selective neuronal cell death [10] . Expansion length is inversely correlated with age of disease onset and directly with disease severity . However , the precise pathogenic mechanism triggered by polyQ-expanded ATX-3 in SCA3 patients has remained elusive [11–16] . A number of works have been carried out to explore ATX-3’s biological and potential cellular roles , and identification of molecular partners interacting with ATX-3 is hoped to facilitate identification of its physiological functions . For example , as a highly specialized deubiquitinating enzyme ( DUB ) , a function of ATX-3 has been shown to be involved in the cellular protein quality control system by interacting with p97/valosin-containing protein ( VCP ) [9 , 17–22] and several E3 Ub ligases [15 , 20 , 23–28] . Moreover , several lines of evidence have shown that ATX-3 can bind DNA and interact with transcription regulators , thus being involved in transcriptional regulation [29–31] . Thus , ATX-3 has been associated with a wide range of biological activities . The absence of ATX-3 leads to an increase of total ubiquitinated protein levels in ATX-3 KO mice [6] , whereas overexpression of ATX-3 results in significantly reduced cellular protein ubiquitination in HEK293 cells [32] , suggesting that ATX-3 may regulate the ubiquitination status of many proteins . However , the substrates targeted by ATX-3 in the physiological context remain unclear , thus limiting our understanding of its cellular functions . Whether the polyQ expansion in ATX-3 may contribute to the neuropathology by affecting its molecular interactions with other proteins or endogenous functions of normal ATX-3 is unknown . To develop effective therapies for this incurable disorder , it is important to identify ATX-3’s preferred substrates and to determine how the polyQ expansion causes the protein’s dysfunction . In the present study , we used immunoprecipitation coupled with mass spectrometry to search for the proteins that associate with ATX-3 . We have found that ATX-3 interacts with p53 and functions as a novel p53 DUB . ATX-3 deubiquitinates , stabilizes p53 , and further regulates the functions of p53 in transactivation and apoptosis both in vitro and in vivo . Whether and how the polyQ expansion in ATX-3 affects its functional regulation of p53 and further neurodegeneration have also been investigated .
We analyzed proteins co-immunoprecipitated with 3×Flag-tagged ATX-3 from H2O2-treated 293T cells using mass spectrometric analysis ( Fig 1A ) . p53 was found to associate with ATX-3 . The interaction between ATX-3 and p53 was confirmed under physiological condition ( Fig 1B ) and with in vitro purified forms ( Fig 1C ) , indicating a direct association between these two proteins . Furthermore , the amino terminus of ATX-3 was found to be necessary for binding with p53 ( Fig 1D ) . This result was corroborated in cells by immunoprecipitation ( S1A Fig ) . In addition , glutathione S-transferase ( GST ) –pull-down assay showed that both the DNA-binding domain and the C-terminal regulatory domain of p53 were sufficient for its interaction with ATX-3 ( Fig 1E ) . The two or three UIMs of ATX-3 , depending on the splice isoform , mediate its binding to poly-ubiquitinated substrates . We observed that full-length ( FL ) ATX-3 bound robustly to both the native and ubiquitinated form of p53 in vitro ( S1B Fig ) and in cells ( Fig 1F ) . Mutating the active site cysteine 14 did not affect the Ub chain binding activity of ATX-3 , whereas ΔC and ΔN deletion as well as the UIM mutations ( S236/256A and A232/252G ) resulted in either abolished or impaired Ub binding activity ( S1D Fig ) . Consistently , as shown in Fig 1G , the cysteine 14 mutation did not affect the binding of ATX-3 to either native or ubiquitinated p53 , whereas the ΔN mutant lost its binding to both forms of p53 . The ΔC mutant was found to bind to the native p53 with decreased affinity , and the two UIM mutants showed significantly compromised binding to ubiquitinated p53 . We constructed a catalytic inactive ΔC mutant to exclude the possibility that the N-terminal domain might be able to interact with ubiquitinated p53 but further be deubiquitinated before detection . Catalytic inactive ΔC mutant showed similar binding affinity to native p53 as the ΔC mutant did , confirming that N-terminal domain only bound to native p53 ( S1C Fig ) . Taken together , these results indicated that the binding of ATX-3 to p53 was synergistically regulated by the Josephin domain and the UIMs , with the former being primarily responsible for the binding of ATX-3 to the native p53 and further facilitating the latter to bind to ubiquitinated p53 . The interaction between ATX-3 and p53 suggested that p53 might be a substrate of ATX-3 . Therefore , we tested whether ATX-3 affected the levels of p53 ubiquitination in vivo . As the p53 levels may differ among different cell lines , to show the generality of the effect of ATX-3 on p53 , we generated ATX-3 stably knockdown cell lines in HeLa , HCT116 , and RKO cells using two non-overlapping short hairpin RNA ( shRNA ) constructs . For the following experiments using knockdown cells in the study , either two clones of ATX-3 shRNA stably knockdown cells were used or one clone of knockdown cells was used but with rescue experiments performed in parallel . As indicated in Fig 2A , knockdown of ATX-3 in HeLa cells significantly increased the level of ubiquitinated-p53 ( lane 3 ) , and transient expression of ATX-3 effectively eliminated the increased ubiquitination that resulted from ATX-3 knockdown ( lane 4 ) . Moreover , ATX-3 overexpression resulted in a suppression of p53 ubiquitination ( lane 2 ) . Similar results were observed in ATX-3+/+ and ATX-3-/- mouse embryonic fibroblast ( MEF ) cells ( Fig 2B left and right panel ) . Furthermore , our results showed that ATX-3 affected p53 ubiquitination in cells , and both the N-terminal Josephin domain and the C-terminal UIM domains of ATX-3 are required for this activity ( Fig 2C ) . These results suggest that ATX-3 may act as a p53-directed DUB . To determine whether ATX-3 can deubiquitinate p53 directly in vitro , we performed in vitro deubiquitination assays . Our results showed that FL ATX-3 deubiquitinated the ubiquitinated-p53 directly in a time- and dose-dependent manner ( Fig 2D ) , which could be repressed by a nonspecific deubiquitinating inhibitor N-ethylmaleimide ( NEMi ) ( Fig 2E left and right panel ) , and the effective deubiquitination of p53 required both the DUB activity and the poly-Ub binding ability of ATX-3 ( Fig 2F ) . As ATX-3 interacts with p53 under physiological conditions and regulates the ubiquitination of p53 in cells , it is possible that ATX-3 may regulate the turnover of p53 via the Ub-proteasome pathway . We found that co-transfection of ATX-3 and p53 led to the accumulation of p53 protein compared with p53 alone-transfected cells ( Fig 3A ) . In contrast , deletion of ATX-3 resulted in an overt reduction in p53 protein level ( Fig 3B , left ) , with no appreciable change at the p53 mRNA level ( Fig 3B , right ) , indicating that the regulation of p53 by ATX-3 is unlikely at the transcriptional level . p53 levels were also compared in ATX-3 wild-type ( WT ) and KO mice primary cultures by western blot ( S1E Fig ) . The basal p53 levels in ATX-3 KO primary MEFs were significantly lower than those in WT primary MEF cells . ATX-3 regulates p53 posttranslationally , because the half-life of p53 was significantly shortened in the ATX-3 stably knocked-down HCT116 cells ( Fig 3C upper ) . This result was further validated in another clone of ATX-3 shRNA-stable knockdown cell line ( Fig 3C lower ) as well as in ATX-3-/- MEF cells ( Fig 3D ) . In HCT116 p53-/- cells , the ectopically expressed p53 showed significantly prolonged half-life in ATX-3 and p53 co-transfected cells compared to that in p53 alone-transfected cells ( Fig 3E ) , thus confirming the positive regulatory effect of ATX-3 on p53 stability . Furthermore , we found that MG132 , a proteasome inhibitor , but not NH4Cl , a lysosome inhibitor , or 3-methyladenine ( 3-MA ) , a well-characterized inhibitor of autophagy , could suppress p53 degradation after cycloheximide ( CHX ) treatment ( Fig 3F ) . Pretreatment of cells with MG132 blocked the degradation of p53 in both ATX-3+/+ and ATX-3-/- MEF cells ( Fig 3G ) . Taken together , these results indicate that ATX-3 stabilizes p53 in cells via the Ub-proteasome pathway . To explore the functional consequences of ATX-3-modulated p53 stability , the normal functions of ATX-3 in regulating p53-dependent biological activities were tested . In dual luciferase reporter assay , the DNA binding ability of p53 , which is measured by the fluorescence intensity of PG13-Luc reporter , was significantly decreased in both ATX-3 knockdown HeLa cells ( Fig 4A ) and ATX-3 KO MEF cell lines ( Fig 4B ) . Furthermore , acetylation of p53 at K373/K382 is reported to be a marker for the stimulation of the p53 transactivation activity . We treated the MEF cells with doxycycline ( DOX ) and found that the levels of acetylated p53 and one of its target gene products , the cyclin-dependent kinase inhibitor p21cip1/waf proteins , were remarkably decreased in ATX-3-/- MEF cells compared to ATX-3+/+ controls , and this could be restored to normal levels by transient expression of ATX-3 ( Fig 4C ) . These results suggest that ATX-3 deletion inhibits the stimulation of p53 transactivation activity . The expression of several p53 target genes , for example , CDNK1A , CCNB1 , and BBC3 , was affected at both mRNA ( Fig 4D and S2A Fig ) and protein levels ( Fig 4E and S2B Fig ) associating with ATX-3 levels in HCT116 and MEF cells . The mRNA levels of cyclin B1 were higher in ATX-3 knockdown cells than that in control cells . This is because p53 acts as a repressor for cyclin B1 . ATX-3 knockdown leads to an inhibition of p53 transcriptional activities , thus relieving its suppression on cyclin B1 , resulting in its increased expression . Notably , overexpression of ATX-3 induced up-regulations of CDNK1A and BBC3 but a down-regulation of CCNB1 , which required both the DUB activity and the poly-Ub binding ability of ATX-3 ( S2C–S2F Fig ) . In contrast , knockdown of ATX-3 significantly blocked their induction , and this could be restored upon ATX-3 ectopic expression ( Fig 4D and 4E ) . The inductions of all three p53 target genes were not changed in HCT116 p53-/- cells , indicating that these inductions were p53 dependent ( Fig 4D and 4E ) . In addition , fluorescence-activated cell sorting ( FACS ) analysis of cell cycle showed that ATX-3 deletion resulted in an increased proportion of cells in G2/M phase ( Fig 4F ) . Without p53 and ATX-3 double KO MEF cells in hand , we examined the p53-dependence of this effect by using ATX-3 stably knocked down HCT116 p53+/+ and HCT116 p53-/- cells . An increase of G2/M phase cells was observed in ATX-3 knockdown HCT116 p53+/+ cells but not in HCT116 p53-/- cells , suggesting that ATX-3 was involved in the regulation of cell cycle arrest in G2/M phase , which was also p53 dependent ( S2G Fig ) . Therefore , our data demonstrated that ATX-3 was able to regulate p53-dependent gene expressions and cell cycle arrest . As p53 is a well-established apoptosis-regulator , we next examined whether ATX-3 affected p53-dependent apoptosis . We found that hallmarks of apoptosis , including the cleaved caspase-3 and poly ( ADP-ribose ) polymerase ( PARP1 ) , were less in ATX-3-/- MEF cells , while overexpression of ATX-3 resulted in significant caspase-3 and PARP1 cleavage ( Fig 5A ) , indicating that ATX-3 was involved in the regulation of apoptosis in cells . Flow cytometry analysis using Annexin V-FITC/propidium iodide ( PI ) staining in HCT116 cells showed that knockdown of ATX-3 led to a decrease of camptothecin ( CPT ) -induced apoptosis , while ectopic expression of ATX-3 but not the catalytic inactive mutant ATX-3-C14A resulted in a significant increase of apoptosis in HCT116 p53+/+ but not HCT116 p53-/- cells ( Fig 5B ) , indicating that ATX-3 promoted p53-mediated apoptosis , which required its deubiquitinating enzymatic activity . Using the zebrafish model system , we further examined whether ATX-3 induces p53-dependent apoptosis in vivo . As our cellular results showed that knockdown of ATX-3 led to a significant decrease of CPT-induced apoptosis , it is quite possible no apoptosis signal would be detected under unperturbed conditions when ATX-3 is knocked down or knocked out . Therefore , the apoptosis as well as neurodegeneration in zebrafish were performed under ectopic expression conditions instead of knockdown or KO conditions . p53 WT and mutant zebrafish embryos were injected with mRNA of FL and various ATX-3 mutants . Twenty-four h post injections , the embryos were harvested and TUNEL-positive cells were analyzed . When the FL ATX-3 mRNA was injected into WT but not the p53 mutant zebrafish embryos , significantly more apoptotic cells were observed in TUNEL-staining assays , indicating that ATX-3 caused p53-dependent apoptosis in vivo . In addition , injections of mRNA of the catalytic inactive mutant ATX-3-C14A ( C/A ) , the C-terminal deletion ( ΔC ) , and the two UIM mutants ( S/A and A/G ) ATX-3 in WT p53 zebrafish embryos exhibited significant reduction in apoptosis compared with that of FL ATX-3 , and no apoptosis was observed in p53 mutant zebrafish , indicating that the induction of p53-dependent apoptosis was critically dependent on the catalytic activity and the UIM domain of ATX-3 ( Fig 5C and 5D ) . Interestingly , we observed that most TUNEL-positive cells localized in the head area of the zebrafish , suggesting that the apoptosis may occur mainly in the nervous system . To confirm this , TUNEL assay was performed by using the Tg ( HuC:EGFP ) transgenic zebrafish embryos , in which GFP-positive cells represent the expression of zebrafish neuronal marker elavl3 ( formerly known as HuC ) [33] . We observed that the TUNEL-positive cells induced by the FL ATX-3 mRNA injection were localized in GFP-positive brain regions ( telencephalon , S3B Fig; diencephalon/hindbrain , S3C Fig ) of the zebrafish , indicating that the ATX-3 mRNA injection-induced apoptosis occurred mainly in the nervous system of zebrafish . PolyQ-expanded ATX-3 ( ATX-3exp ( 80Q ) ) is thought to undergo conformational changes and acquire toxic properties , leading to altered molecular interactions . We wonder whether the polyQ expansion affects the ATX-3/p53 interaction . GST—pull-down assay ( Fig 6A ) and coimmunoprecipitation experiments ( Fig 6B ) showed that p53 , both native and ubiquitinated form , bound ATX-3exp ( 80Q ) stronger than the normal ATX-3 . Consistently , ATX-3exp ( 80Q ) exhibited stronger DUB activity than the normal ATX-3 in vitro ( Fig 6C ) and in cells ( Fig 6D ) . Besides , we found that the degradation of p53 in the ATX-3exp ( 80Q ) -expressing cells was slower than that of normal ATX-3-expressing cells ( S4C and S4D Fig ) , and ectopic expression of polyQ-expanded ATX-3 induced higher levels of p53 protein than the normal ATX-3 in RKO , 293T , and MEF cells ( S4E Fig ) , indicating that polyQ-expanded ATX-3 possessed enhanced capability to stabilize p53 . The expression levels of p53-responsive genes ( such as p21 and Puma ) were also higher in ATX-3exp ( 80Q ) expressing RKO cells ( S4F Fig ) , suggesting that p53 was functionally enhanced by polyQ expansion in ATX-3 . For unknown reasons , HCT116 cells do not behave as significantly as other cell lines ( such as RKO , 293T , and MEFs ) in terms of p53 induction ( Fig 6F and S4E Fig ) . To determine the p53 dependence issue , we used the HCT116 cell lines with both p53+/+ and p53-/- genotypes . In HCT116 p53+/+ , but not HCT116 p53-/- cells , both the normal and polyQ-expanded ATX-3 increased the induction of p53-responsive genes at both mRNA ( Fig 6E and S4A Fig ) and protein levels ( Fig 6F and S4B Fig ) when compared to the empty vector control group , but the difference did not reach the statistical significance between the normal ATX-3 group and polyQ-expanded ATX-3 group . All together , these data indicated that the polyQ expansion did not disturb the binding and the DUB activity of the normal ATX-3 to p53 and appeared to augment p53 stabilization . CPT-induced apoptosis was analyzed by flow cytometry using Annexin V/PI . Apoptotic cells are those positive for Annexin V staining ( either positive or negative for PI staining ) ( S3A Fig ) . Our results showed that expression of the FL ATX-3exp ( 80Q ) led to a similar increase of p53-dependent apoptosis ( including early and late apoptosis/necrosis ) in HCT116 p53+/+ ( Fig 7A ) and zebrafish ( Fig 7C and 7D ) comparing to the normal ATX-3 . Early apoptotic cells are Annexin V+/PI- staining , whereas late apoptotic/necrotic cells are Annexin V+/PI+ staining ( PI-positive staining is due to a loss of plasma membrane integrity ) . Interestingly , we found that the normal ATX-3 induced a significantly higher percentage of early apoptotic cells , whereas the ATX-3exp ( 80Q ) led to more late apoptotic/necrotic cells in HCT116 p53+/+ but not HCT116 p53-/- cells ( Fig 7A and S3A Fig ) . In support , we found that CPT treatment resulted in clearly different morphological nuclear changes in the ATX-3exp ( 80Q ) - and the normal ATX-3-expressing HCT116 p53+/+ but not HCT116 p53-/- cells after staining the nuclear DNA by Hoechst-33342 ( S5A Fig ) . Nuclei of the normal ATX-3-expressing cells were round shaped but without condensed and fragmented chromatin , which represent an early apoptotic event . In contrast , nuclei of ATX-3exp ( 80Q ) -expressing cells looked more amorphous , without any defined surface outline , and the nuclear heterochromatin was extensively packaged , indicating necrotic cell death ( S5A Fig ) . Furthermore , as shown in Fig 7B , the p53-dependent apoptosis program marker such as cleaved PARP-1 in the ATX-3exp ( 80Q ) -expressing HCT116 p53+/+ cells was found to be higher than the empty vector control group but weaker when compared to the normal ATX-3 group , indicating that the mitochondrial apoptotic p53 program played a role in this process , but the extent of this influence may not be as profound as that of the normal ATX-3 . More importantly , High Mobility Group Box 1 ( HMGB1 ) protein that released into the culture medium ( a classical biochemical hallmark specific for necrosis [34 , 35] ) as well as the level of receptor-interacting serine/threonine protein kinases 1 ( RIP1 , an important mediator of necrosis ) were found to be significantly enhanced in ATX-3exp ( 80Q ) -expressing HCT116 p53+/+ cells compared to those of the empty vector control and the normal ATX-3 group after CPT treatment , indicating that CPT apparently triggered a necrotic program in addition to the mitochondrion-dependent apoptotic program in polyQ-expanded ATX-3-expressing cells . These markers were not significantly changed in HCT116 p53-/- cells ( Fig 7B ) , demonstrating the changes of these markers were p53 dependent . Together , these results suggested that a continuum of apoptosis and necrosis existed in response to CPT insult in ATX-3exp ( 80Q ) -expressing cells , both of which were mediated by p53 . Next , we set out to determine if polyQ-expanded ATX-3 causes more neuronal cell death in vivo by using zebrafish as a model vertebrate . We injected mRNA of the normal ATX-3 and ATX-3exp ( 80Q ) into wild-type and p53 mutant fish one-cell embryos . In zebrafish , otx2 is expressed in the prospective forebrain/midbrain in mid/gastrulae [36–38] , and neurogenin 1 ( ngn1 ) is reported to be a determinant of zebrafish basal forebrain dopaminergic neurons [39 , 40] . We performed the whole-mount in situ hybridization using otx2 and ngn1 as two neural markers to evaluate the neural loss upon the normal ATX-3 and ATX-3exp ( 80Q ) expression in zebrafish . We observed that expression of normal ATX-3 or ATX-3exp ( 80Q ) resulted in decreased signals of otx2 ( blue staining mainly in MB area , Fig 7E and S5B Fig ) and ngn1 ( blue staining including telencephalon ( TE ) , midbrain ( MB ) , and hindbrain ( HB ) areas , S5C Fig ) in WT but not p53 mutant zebrafishes at 24 h post fertilization ( hpf ) , with more profound reduced signals of otx2 and ngn1 in ATX-3exp ( 80Q ) mRNA injection groups ( Fig 7E , S5B and S5C Fig ) . By 48 and 72 hpf , brains of ATX-3exp ( 80Q ) -injected zebrafishes showed even weaker levels of both otx2 and ngn1 ( Fig 7E , S5B and S5C Fig ) . These results provided clear evidences that expression of ATX-3exp ( 80Q ) led to more neuronal loss in brains of WT but not p53 mutant zebrafishes , suggesting that ATX-3exp ( 80Q ) caused more severe neuronal degeneration in SCA3 in a p53-dependent manner . Previous studies have shown that neurodegeneration affects particular brain regions in MJD pathology [41] . To further study the neurodegeneration in specific affected brain regions in MJD , we generated an in vivo MJD genetic mouse model by ectopic expression of WT and mutant ATX-3 ( 80Q and C14A ) with an enhanced green fluorescent protein ( EGFP ) tag in the substantia nigra pars compacta ( SNpc ) or striatum of p53+/+ and p53-/- mouse brain using lentiviral vectors ( LV ) . The expression of these LV was first validated in 293T cells using fluorescence microscopy ( S6A Fig ) and also by western blot with anti-ATX-3 antibody ( S6B Fig ) . Immunostaining analysis for tyrosine hydroxylase ( TH ) , a marker for dopaminergic neurons in the SNpc , and for dopamine- and cyclic AMP-regulated neuronal phosphoprotein ( DARPP-32 ) , a regulator of dopamine receptor signaling , was performed to evaluate the neurodegeneration induced by lentiviral transduction . Our results showed that , in p53+/+ mice , WT ATX-3 caused a nonsignificant reduction of 15% TH-positive neurons in the SNpc ( Fig 8A and 8B ) and a reduction of 31% DARPP-32–positive neurons in the striatum ( S6C and S6D Fig ) , whereas mutant ATX-3exp ( 80Q ) induced a more significant loss of neurons in both of these two brain areas when compared to the empty vector expressed brain section , with a reduction of 57% for TH-positive neurons in the SNpc ( Fig 8A and 8B ) and of 51% for DARPP-32–positive neurons in the striatum ( S6C and S6D Fig ) . The loss of TH or DARPP-32 immunointensity mainly occurred in EGFP-positive neurons . No significant neuronal loss was detected in either the SNpc or the striatum of p53+/+ mice injected with the catalytic inactive mutant ATX-3-C14A ( Fig 8A and 8B , S6C and S6D Fig ) . The evidences for p53 involvement in the mutant ATX-3exp ( 80Q ) -induced neurodegeneration were provided by the fact that no significant decreases of both neuronal markers were detected in p53-/- mice after LV transduction ( Fig 8C and 8D , S6E and S6F Fig ) . Further supports for p53 involvement were evidenced by the observation that WT and mutant ATX-3exp ( 80Q ) , but not the catalytic inactive mutant C14A , caused marked increases in p53 immunoreactivity in the SNpc ( Fig 8A and 8B ) and striatum ( S6C and S6D Fig ) of p53+/+ mice , whereas no p53 staining was detected in the p53-/- mice ( Fig 8C and 8D , S6E and S6F Fig ) . These results provided consistent evidences for a role of p53 in the WT and mutant ATX-3exp ( 80Q ) expression induced neuronal degeneration . To determine whether loss of TH staining in the SNpc as well as DARPP-32 staining in the striatum , associated with increased expression of WT and mutant ATX-3exp ( 80Q ) , was due to neuronal death , we performed TUNEL analysis and monitored expression of activated caspase-3 . The numbers of TUNEL-positive and activated caspase-3-positive cells were significantly increased in the EGFP-positive neurons in both the SNpc ( Fig 8E and 8F ) and striatum ( S7A and S7B Fig ) of p53+/+ but not p53-/- mice . These results confirmed the association of cell death with WT and mutant ATX-3exp ( 80Q ) expression . It should be noted that WT ATX-3 induced significant higher TUNEL and activated caspase-3 staining compared to the mutant ATX-3exp ( 80Q ) . As signs of non-apoptotic cell death , including extracellular release of HMGB1 and higher expression of RIP1 , were observed in mutant ATX-3exp ( 80Q ) -expressed cells ( Fig 7B ) , and it is generally accepted that activated caspase-3 are rarely detected in the case of necrosis [42] , we hypothesized that necrosis might occur in these brain areas . To gain further insights into mechanism of the neuronal death , we evaluated the expression of RIP1 , which is an important molecule mediating necrosis when caspases are inhibited [43] . We found that mutant ATX-3exp ( 80Q ) -expressed brain sections of p53+/+ but not p53-/- mice had more intense staining for RIP1 than did the WT ATX-3 sections ( Fig 8E and 8F , S7A–S7C Fig ) . A high magnification view of the anti-RIP1 and anti-TH immunostaining in the SNpc of p53+/+ mice showed that most TH-positive neurons showed intact nuclear membrane , whereas those neurons that had condensed chromatin structures and nuclei were highly RIP1 positive but negative for TH staining ( S7C Fig ) . No obvious RIP1 immunostaining was observed in p53-/- mice ( S7C Fig ) . These results indicated the involvement of RIP1 in polyQ ATX-3exp ( 80Q ) -induced neuronal death in mouse brains . Together , our in vivo data have demonstrated that the polyQ ATX-3 caused p53-dependent neuronal death in both apoptotic and necrosis manner in mouse brains .
p53 activity is crucial in determining the cellular fate , keeping a delicate balance between cancer-suppressive and age-promoting functions [44–48] . Therefore , tight regulation of p53 is essential for maintaining normal cellular functions . It has been shown that p53 is mainly regulated at the level of protein stability , which occurs predominantly through the Ub-mediated proteasomal degradation . On the flip side of the regulation , deubiquitination , which is mediated by DUBs , provides a parallel important regulatory control of p53 stability . Previously , several DUBs from the ubiquitin-specific protease ( USP ) [49–53] and otubain ( OTU ) family members [54] have been shown to regulate the Mdm2-p53 pathway , each with different detailed mechanisms of action . For example , USP7 ( also named HAUSP ) was the first identified USP that stabilizes p53 [55] . Later , it was found to deubiquitinate Mdm2 and Mdmx as well [56] , thus showing selective deubiquitination to regulate the homeostatic levels of p53 , Mdm2 , and Mdmx under both normal and stress conditions [55] . Unlike USP7 , USP10 [52] , a cytoplasmic DUB , had recently been shown to directly deubiquitinate p53 , but not Mdm2 and Mdmx , and to regulate the subcellular localization and stability of p53 by opposing the effects of Mdm2 . In the present study , we report that p53 is a novel substrate of ATX-3 under physiological conditions . Previous studies have suggested a possible role of the tumor suppressor protein p53 in neurodegenerative diseases , although the evidences are indirect . p53 is mutated in approximately half of all human cancers , and accumulating evidences also supported a role of p53 in neurodegeneration [57 , 58] . For example , p53 was found to be highly elevated in brains affected by several neurodegenerative diseases , including Alzheimer’s disease ( AD ) , Parkinson’s disease ( PD ) , Huntington’s disease ( HD ) , amyotrophic lateral sclerosis ( ALS ) , HIV-associated neurocognitive disorders ( HAND ) , etc . [59] . Furthermore , several epidemiological studies have found an inverse correlation between the risk of developing neurodegenerative disorders and cancer [60–62] , suggesting that some common protein effectors might likely be involved between these two multifactorial chronic pathologies . Given the important role of p53 in neurodegenerative diseases and cancer , it is thus a likely possible candidate . Recent studies have reported that ATX-3 and ATX-3 like are involved in gastric cancer [63] and breast cancer [64] , supporting the association of the Josephin family of DUBs with cancer . Importantly , aberrant activation of the p53 pathway has previously been reported in both MJD patient brain tissues and transgenic animal disease models [65–69] , and elevated p53 level was observed in MJD transgenic mice [65] . Therefore , the overall above-cited data concur to suggest the possibility of a functional link between ATX-3 and p53 . p53 has not come out as a potential issue in the ATX-3 KO mice from the literature . As we mentioned in the Introduction section , this may be because other members of the MJD DUBs may compensate for its absence in ATX-3 KO models under unstressed conditions . Here in our study , we discovered that ATX-3 interacts with p53 and functions as a DUB for p53 . We have observed that the Josephin domain of ATX-3 is sufficient for the direct binding of ATX-3 to native p53 , whereas the ubiquitinated p53 interacts with ATX-3 primarily through the UIMs , indicating that UIMs function to help recruit and bind the ubiquitinated p53 ( Fig 9A and 9B ) . Therefore , both the Josephin and UIM domain coordinately regulate the interaction between ATX-3 and p53 ( Fig 9A and 9B ) . During the DUB process , in addition to the catalytic cysteine 14 site and the N-terminal Josephin domain , the UIMs of ATX-3 are also required for its DUB activity towards p53 , which may function to position the polyubiquitinated p53 correctly relative to the catalytic site for subsequent cleavage ( Fig 9B ) . Thus , ATX-3 deubiquitinates and stabilizes p53 ( Fig 9C ) , which is an essential step for p53 function in cell cycle arrest and apoptosis ( Fig 9D ) . The direct interaction between ATX-3 and p53 is primarily mediated by the Josephin domain , and the first two UIM domains function to enhance the interaction by trapping the Ub-chains on p53 . The polyQ tract between UIM2 and UIM3 is expanded in MJD . Our results suggest that polyQ length enhances rather than disturbs the binding and deubiquitination of ATX-3exp to p53 ( Fig 6A–6D ) , which , in turn , causes more p53-dependent apoptosis/necrosis ( Figs 7A , 7B , 7E , 8E and 8F , S5B , S5C , S7A and S7B Figs ) . Previously , mutant huntingtin with expanded polyQ was found to bind to p53 and cause more cell death than WT huntingtin in neuronal cultures [70] . Here , polyQ-expanded ATX-3 was found to cause an increased percentage of cells undergoing p53-dependent late apoptotic/necrotic cell death than the normal ATX-3 did in HCT116 cells ( Fig 7A and 7B , S3A Fig ) and in neurons ( Fig 8E and 8F ) . In consistence with our result , Evert’s group also observed an increased necrotic cell death in a cellular model for SCA3 upon polyQ ATX-3 expression [71] , and p53 have been recently reported to play important roles in activating necrotic cell death [35 , 72 , 73] . Whether polyQ-expanded ATX-3 has additional targets that work together with p53 in inducing necrosis is not known yet , but we did find that the level of RIP1 , an important mediator in necrosis , increased significantly in p53+/+ cells upon polyQ-expanded ATX-3 expression when compared to the empty vector control and the normal ATX-3 group , but remained at a basal low level in p53-/- cells ( Figs 7B , 8E and 8F , S7A and S7B Fig ) . However , we cannot rule out the possibility that some other signaling pathways or p53 status ( such as modification or subcellular distribution ) [74] might also been affected upon polyQ-expanded ATX-3 expression , which led to the change of the percentage of cells undergoing apoptosis and necrosis . We first used zebrafish as a model vertebrate to test the influence of polyQ expansion on the neuronal cell death in vivo . It is intriguing to note that injection of ATX-3 mRNA into the zebrafish embryos led to p53-dependent apoptosis , which occurred mainly in the central nervous system of zebrafish at early development stage ( 24 hpf ) . However , the polyQ-expanded ATX-3 induced progressive severe p53-dependent neurodegeneration in the central nervous system of zebrafish , suggesting that it caused other kinds of p53-mediated neural cell death besides apoptosis , too . By generating a lentiviral-based in vivo MJD genetic model in p53+/+ and p53-/- mice , our in situ detection of two apoptotic markers ( TUNEL and active caspase-3 ) and one necrosis marker ( RIP1 ) data have provided convincing evidences that both enhanced apoptotic-like and non-apoptotic cell death are observed in the SNpc and striatum of ATX-3exp ( 80Q ) -expressed neurons in p53+/+ mice brains but not in p53-/- mice brains ( Fig 8E and 8F , S7A–S7C Fig ) . Meanwhile , significantly more p53-positive cells were detected in ATX-3exp ( 80Q ) -expressed mouse brains sections ( Fig 8A and 8B , S6C and S6D Fig ) . These data supported an idea that , due to enhanced interaction to p53 and up-regulation of p53 , polyQ-expanded ATX-3 led to an increased p53-dependent neuronal cell death ( including both early apoptotic and late apoptotic/necrotic manner ) . All together , the aforementioned studies and our results provide consistent evidences for the involvement of p53 in SCA3 pathogenesis , and the activation of the p53 pathway likely triggers neuronal dysfunction and eventually neuronal cell death in SCA3 . In conclusion , by identifying p53 as a new substrate of ATX-3 , our study not only reveals a physiological function of ATX-3 and a new mechanism of p53 regulation but also establishes a novel molecular link between disease mutant ATX-3 and p53-mediated neurodegeneration , which sheds light on the molecular pathogenic mechanisms in SCA3 .
Our work involving zebrafish and mouse experiments was in full compliance with the Regulations for the Care and Use of Laboratory Animals by the Ministry of Science and Technology of China and with the Institute of Zoology's Guidelines for the Care and Use of Laboratory Animals . The experimental protocols was approved by the Animal Care and Use Committee at the Institute of Zoology , Chinese Academy of Sciences ( Permission Number: IOZ-13048 ) . ATX-3+/+ and ATX-3-/- MEFs , HEK293T , A549 , RKO , and HeLa cells were cultured in DMEM supplemented with 10% FBS . U2OS , HCT116 p53+/+ , and HCT116 p53-/- cells were cultured in McCoy’s 5A supplemented with 10% FBS . WT zebrafish embryos were obtained from natural matings of zebrafish Tuebingen strain . Tg ( HuC:EGFP ) transgenic fish embryos express EGFP in the post-mitotic thalamic neurons . Homozygous p53 ( M214K ) mutant fish line carrying a loss-of-function p53 point mutation was kindly provided by Prof . Jinrong Peng at College of Life Sciences , Zhejiang University . Embryos were raised in Holtfreter’s solution at 28 . 5°C and staged by morphology as described [75] . P53 +/- mice were obtained from Jackson Laboratory . P53+/- mice were intercrossed to get P53-/- mice and P53+/- mice . For genotyping the p53 locus , primers X7 ( 5′-TAT ACT CAG AGC CGG CCT-3′ ) , NEO19 ( 5′-CAT TCA GGA CAT AGC GTT GG-3′ ) , and X6 . 5 ( 5′-ACA GCG TGG TGG TAC CTT AT-3′ ) were used as described previously [76] . Four-wk-old mice were used . ATX-3 was cloned into pCS2-Flag , p3×Flag-CMV26-Myc , pGEX-4T-1 , and pET-28a vectors . p53 was cloned into pCMV-HA , pCMV-Flag , pGEX-4T-1 , and pET-28a vectors . ATX-3 mutants were generated by site-directed mutagenesis ( Stratagene ) . Anti-ATX-3 ( 1H9 ) was purchased from Merck . Anti-p53 ( DO-1 , PAb1620 , and pAb421 ) and anti-Bax were purchased from Calbiochem . Anti-p53 ( FL393 ) , anti-Ub , anti-p21 , anti-Cyclin B1 , and anti-PARP1 were purchased from SantaCruz . Anti-Flag ( M2 ) was purchased from Sigma . Anti-His was purchased from Abmart . Anti-HA was purchased from Covance . Anti-Puma , anti-p17 specific caspase 3 , anti-RIP1 , and anti-HMGB1 were purchased from ProteinTech . Anti-GFP was purchased from Thermo Fisher Scientific Inc . Anti-DARPP-32 and anti-TH antibodies were purchased from Cell Signaling . 3×Flag-ATX-3 transiently expressed in HEK293T cells was immunoprecipitated with anti-Flag M2 affinity gel and eluted with 3×Flag peptide ( Sigma ) . Eluted proteins were identified with a gel-based liquid chromatography-tandem mass spectrometry ( Gel-LC-MS/MS ) approach and ion trap mass spectrometry ( LTQ; Thermo Electron ) . A Mascot database search and the Scaffold program ( Proteome Software ) were used to visualize and validate results . Cells were lysed with NETN buffer ( 20 mM Tris-HCl , pH 8 . 0 , 100 mM NaCl , 1 mM EDTA , 0 . 5% Nonidet P-40 ) containing 1 mM Na3VO4 , 10 mM NaF , 10 mM NEMi , and a cocktail of protease inhibitors . Whole cell lysates obtained by centrifugation were incubated with primary antibody overnight at 4°C . Protein A/G PLUS-Agarose beads ( Santa Cruz ) were then added and incubated for 2 h at 4°C . The immunocomplexes were then washed with NETN buffer for six times and separated by SDS-PAGE . Immunoblotting was performed following standard procedures . GST and His fusion proteins were expressed in Escherichia coli strain BL21 ( DE3 ) and affinity-purified using Glutathione-Sepharose 4B beads ( GE Healthcare ) or Ni-NTA Agarose ( Qiagen ) , respectively , according to the manufacturer's instructions . For in vitro binding assays , bead-immobilized GST proteins were incubated with purified His proteins or with cell lysates in assay buffer at 4°C for 3 h followed by extensive washing . The bound proteins were separated by SDS-PAGE and analyzed by western blot with indicated antibodies . For the in vitro deubiquitination assays , the ubiquitinated p53 protein was incubated with recombinant GST-ATX-3 ( 100 ng ) or the same amount of other indicated proteins in a deubiquitination buffer ( 50 mM Tris-HCl pH 8 . 0 , 50 mM NaCl , 1 mM EDTA , 10 mM DTT , 5% glycerol ) for 2 h at 37°C . Western blot was performed to detect p53 ubiquitination . GST controls were included in all DUB assays . The in vitro p53 ubiquitination assays were conducted in a total of 20 μl reaction buffer containing recombinant p53 ( 20 ng ) , Mdm2 ( 100 ng ) , UbE1 ( 0 . 025 μM , Boston Biochem ) , UbcH5 ( 0 . 4 μM , Biomol ) , Ub ( 40 μM , Boston Biochem ) , 50 mM Tris-HCl ( pH 7 . 5 ) , 5 mM MgCl2 , 2 mM ATP , and 2 mM DTT in the absence or presence of varying amount of bacterial His-ATX-3 at 37°C for 2 h . The reactions were stopped by adding SDS sample buffer followed by immunoblotting with anti-p53 antibodies . The cells were treated with proteasome inhibitor MG132 ( 20 μM ) for 4 h and then lysed in NETN lysis buffer with mild sonication . p53 was immunoprecipitated from the cell extract and subsequently resolved by SDS-PAGE and analyzed by western blot . For the preparation of a large amount of ubiquitinated p53 as the substrate for the deubiquitination assay in vitro , HEK293T cells were transfected together with the Flag-p53 , HA-Ub , and Mdm2 expression vectors . After treatment as described above , the ubiquitinated p53 was purified from the cell extracts with anti-Flag M2 beads . After extensive washing , the proteins were eluted with Flag peptides ( Sigma ) . For protein half-life assays , 20 μg/ml CHX was added to cell cultures to block protein synthesis . Cells were collected at indicated time points , and protein levels were measured by western blot . The relative intensities of the bands were determined by densitometry analyses using Photoshop 7 . 0 software ( Adobe ) . The half-lives of proteins were calculated from three independent experiments . To determine which degradation pathway was involved , 20 μg/ml CHX was added for the indicated intervals in the presence of MG132 ( 20 μM ) , NH4Cl ( 20 mM ) , or 3-MA ( 5 mM ) . ATX-3+/+ and ATX-3-/- MEF cells were co-transfected with the indicated reporter constructs PG13 and the internal control Renilla luciferase pRL-null ( pRL-CMV , Promega ) at a ratio of 8:1 using PEI . Luciferase assays were performed using a dual-luciferase reporter assay system ( E1910 , Promega ) according to the instructions of the manufacturer . Data were normalized for activity of Renilla luciferase to account for transfection efficiency . The assays were performed in duplicate , and data represent the average of five independent experiments . For the flow cytometric analysis of cell cycle with PI DNA staining , the cells were harvested and washed once with PBS , followed by fixation in cold 70% ethanol at 4°C overnight . Then the cells were washed twice with PBS and treated with ribonuclease . Two hundred μl of PI was added before flow cytometry analysis . Apoptosis was assessed by using Becton—Dickinson FACScan flow cytometer according to the manufacturer’s instructions . Cells were treated with or without 1 μM CPT ( a topoisomerase I inhibitor ) for 24 h . Cells were collected and washed once with PBS , followed by incubation in annexin V ( A13201; Invitrogen ) solution in dark at room temperature for 15 min and 10 μl of PI in annexin V binding buffer . Flow cytometry analysis was carried out within 1 h . Data analysis was performed with CellQuest software . The numbers of apoptotic cells that are positive for annexin V staining ( positive and negative for PI staining ) were counted as a proportion to the total number of gated cells and expressed as percent of apoptotic cells in a histogram . Early apoptotic cells are positive for annexin V staining and negative for PI staining , whereas late apoptotic/necrotic cells are positive for annexin V and PI staining due to a loss of plasma membrane integrity . Total RNA was extracted from HCT116 or MEF cells using TRIzol ( Invitrogen ) , and 1 . 5 μg of total RNA was used to prepare the first-strand cDNA using the SuperScript II polymerase ( Invitrogen ) . Quantitative real-time PCR reactions were carried out in triplicate on a Thermal Cycler using SYBR Green dye to measure amplification . Relative mRNA levels of each gene shown were normalized to the expression of the housekeeping genes GAPDH . Normal , necrotic , and apoptotic cells were observed under fluorescence microscopy . Cells were fixed by 4% paraformaldehyde in PBS , and their nuclear DNA was stained with Hoechst-33342 for detection of necrosis and apoptosis by morphological features , according to Gschwind and Huber [77] . mRNA of FL and various ATX-3 mutants were in vitro synthesized from corresponding linearized plasmids using mMESSAGE mMACHINE Kit ( Ambion ) . Digoxigenin-UTP-labeled antisense RNA probes were transcribed in vitro using MEGAscript Kit ( Ambion ) according to the manufacturer’s instructions . Microinjection and whole-mount in situ hybridization were performed as before [78–81] . Apoptosis were determined by TUNEL assay [82] . WT and p53 mutant zebrafish embryos were injected with object mRNA at one-cell stage and then harvested at 24 hpf for TUNEL labeling using In Situ Cell Death Detection Kit , TMR red ( 12156792910 , Roche ) according to the manufacturer’s instruction . Tg ( HuC:EGFP ) zebrafish embryos were fixed and stained with anti-GFP antibody . TUNEL-positive cells were imaged by confocal microscopy . Images were analyzed with ImageJ software . Significance was analyzed using the unpaired t test . Lentiviral-EGFP vectors encoding human WT ATX-3 or mutant ATX-3 , including ATX-3-80Q and ATX-3-C14A , were produced in 293T cells with a three-plasmid system . The lentiviral particles were re-suspended in artificial cerebrospinal fluid . Viral stocks were stored at -80°C . Mice were anesthetized with 1% pentobarbital sodium and then placed into a stereotactic frame . Concentrated viral stocks were thawed on ice . L were stereotaxically injected into the striatum at the following coordinates: anterior-posterior: + 0 . 6 mm from the bregma; mediolateral: -2 . 5 mm from midline; and dorsoventral: -3 . 2 mm below surface of the dura; tooth bar: 0 . LV were stereotaxically injected into the SNpc at the following coordinates: anterior-posterior: -3 . 8 mm from the bregma; mediolateral: -2 . 1 mm from midline; and dorsoventral: -4 . 5 mm below surface of the dura; tooth bar: 0 . P53+/+ and P53-/- mice received a single injection of lentivirus in each side: left hemisphere ( LV-GFP ) and right hemisphere ( ATX-3-WT , ATX-3-80Q , or ATX-3-C14A ) . Mice were kept in their home cages for 8 wk before being killed for immunostaining analysis . Mice were perfused transcardially with 0 . 1 M phosphate buffered saline followed by fixation with 4% paraformaldehyde . Serial coronal sections were cut through the entire striatum and SNpc at 25 μm . Free-floating cryosections from injected mice were blocked in PBS/0 . 3% TritonX-100 containing 10% normal donkey serum ( Gibco ) and then incubated overnight at 4°C with the following primary antibodies: a rabbit polyclonal anti- DARPP-32 antibody ( Cell Signaling , 1:200 ) ; a rabbit polyclonal anti-TH antibody ( Cell Signaling , 1:200 ) ; a mouse monoclonal anti-p53 antibody , clone PAb1620 ( 1:50; Millipore ) ; a rabbit polyclonal anti-RIP1-specific antibody ( 1:50 , Proteintech ) ; and a rabbit polyclonal anti-caspase 3 , p17-specific antibody ( 1:50 , Proteintech ) . Sections were washed in PBS and then incubated with the corresponding secondary antibodies coupled to fluorophores ( Molecular Probes ) for 1 h at 37°C . TUNEL analysis was performed using cryosections with the ApopTag-fluorescein in situ apoptosis detection kit ( Chemicon ) according the manufacturer's instruction . Apoptotic cells were quantified , and apoptotic indices were calculated by computer-assisted image analysis following identification of apoptotic cells by morphological analysis , the TUNEL assay , or activated caspase-3 immunostaining . Imaging was done by total internal reflection fluorescence microscope and three-dimensional structured illumination microscopy . Cell counts were determined from anatomically matched sections from each of the animals , and three animals were used for cell counts .
|
Ataxin-3 ( ATX-3 ) is a ubiquitously expressed protein that mutated in a neurodegenerative disease called spinocerebellar ataxia type 3 ( SCA3 ) . It contains a polyglutamine ( polyQ ) tract near its C-terminus , the expansion of which is known to be the causative factor for SCA3 . It has been known for a long time that ATX-3 is a deubiquitinating enzyme ( DUB ) . However , the substrates targeted by ATX-3 in the physiological context remain elusive , thus largely limiting our understanding of its cellular function and that of the pathogenic mechanism of SCA3 . This study has identified p53 to be a novel substrate of ATX-3 , and its function is tightly regulated by ATX-3 . PolyQ expansion augments ATX-3’s cellular function in p53 regulation . Due to enhanced interaction to p53 and up-regulation of p53 , polyQ-expanded ATX-3 led to an increased p53-dependent neuronal cell death in zebrafish and mouse models , thus providing clear in vivo evidences for the direct involvement of p53 in SCA3 pathology . This study not only establishes a basic function of ATX-3 but also provides an explanation of how the interplays between ATX-3 and p53 contribute to the SCA3 pathogenesis; thus , it is an important contribution for the future development of therapeutic approaches for this currently untreatable neurodegenerative disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"molecular",
"probe",
"techniques",
"cell",
"processes",
"immunoblotting",
"vertebrates",
"neuroscience",
"animals",
"animal",
"models",
"osteichthyes",
"developmental",
"biology",
"model",
"organisms",
"immunoprecipitation",
"neuronal",
"death",
"molecular",
"biology",
"techniques",
"embryos",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"embryology",
"staining",
"fishes",
"animal",
"cells",
"molecular",
"biology",
"precipitation",
"techniques",
"zebrafish",
"cell",
"staining",
"cellular",
"neuroscience",
"cell",
"biology",
"apoptosis",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2016
|
The Machado–Joseph Disease Deubiquitinase Ataxin-3 Regulates the Stability and Apoptotic Function of p53
|
Quantitative genetic analysis has long been used to study how natural variation of genotype can influence an organism's phenotype . While most studies have focused on genetic determinants of phenotypic average , it is rapidly becoming understood that stochastic noise is genetically determined . However , it is not known how many traits display genetic control of stochastic noise nor how broadly these stochastic loci are distributed within the genome . Understanding these questions is critical to our understanding of quantitative traits and how they relate to the underlying causal loci , especially since stochastic noise may be directly influenced by underlying changes in the wiring of regulatory networks . We identified QTLs controlling natural variation in stochastic noise of glucosinolates , plant defense metabolites , as well as QTLs for stochastic noise of related transcripts . These loci included stochastic noise QTLs unique for either transcript or metabolite variation . Validation of these loci showed that genetic polymorphism within the regulatory network alters stochastic noise independent of effects on corresponding average levels . We examined this phenomenon more globally , using transcriptomic datasets , and found that the Arabidopsis transcriptome exhibits significant , heritable differences in stochastic noise . Further analysis allowed us to identify QTLs that control genomic stochastic noise . Some genomic QTL were in common with those altering average transcript abundance , while others were unique to stochastic noise . Using a single isogenic population , we confirmed that natural variation at ELF3 alters stochastic noise in the circadian clock and metabolism . Since polymorphisms controlling stochastic noise in genomic phenotypes exist within wild germplasm for naturally selected phenotypes , this suggests that analysis of Arabidopsis evolution should account for genetic control of stochastic variance and average phenotypes . It remains to be determined if natural genetic variation controlling stochasticity is equally distributed across the genomes of other multi-cellular eukaryotes .
Almost all phenotypes are not fixed within species but instead exhibit significant levels of variation among individuals that is controlled by quantitative genetic loci . The study of such quantitative genetic variation has long been fundamental to evolution and ecology and is rapidly becoming a central focus of numerous other research fields , including breeding for improved crops and individualized medicine for humans . An ultimate goal of research on the basis of quantitative genetic variation is to generate a sufficient level of understanding to be able to predict phenotypic range of a species based on knowledge of that species' genetic variation . These efforts are complicated because phenotypic diversity is typically under polygenic control and can involve complex interactions with numerous factors including , but not limited to , the environment , development , epistatic interactions between genes , and potential higher-order interaction among these factors [1] , [2] . Yet even in systems where these are understood to a significant degree , it has been difficult to develop predictive frameworks linking genotype to phenotype . Some of this difficulty has been ascribed to concepts such as epigenetic variance and difficulties in detecting small-effect loci [3] , [4] . In this report , we propose that an additional explanation is the presence of numerous polymorphic loci that specify the amount of stochastic noise . If these polymorphisms are frequent in number , heritable and discrete from loci altering mean phenotpyes they can lead to an inability to fully describe the variance within any phenotype using current statistical approaches that focus solely upon the mean phenotype . The idea that phenotypic variance is genetically determined is supported by a significant amount of research on how cells can limit stochastic noise/variance in genetic , metabolic , and signaling networks through network topology , a characteristic that is known as network robustness [5]–[10] . The specific topology of a network can increase or decrease the robustness of the output , wherein robustness is defined as the inverse of variance . Therefore , the genetic variation for loci within these networks could lead to allele specific changes in robustness/variance of the phenotype . Typically , robustness is thought to be under directional selection pressure to reduce the variance of an output and correspondingly increase network robustness . In evolutionary theory , this is predominantly described as canalization wherein genes function to minimize the variance ( maximize the robustness ) of a phenotype [11] , [12] . In yeast , phenotypic and genetic robustness ( i . e . canalization ) were shown to correlate using genomic knockout datasets [13] . In plants , loci that control natural variation in canalization of critical developmental processes such as cotyledon opening and leaf formation have been mapped and cloned revealing that canalization genes can be known members of regulatory networks controlling these processes [14] , [15] . Additionally , it has been shown that heat-shock protein 90 plays a major role in canalizing existing natural variation possibly as a pool of hidden evolutionary potential [16]-[18] . However it should be noted that the genomic level of , distribution of and importance of naturally variable loci controlling within-genotype variance is currently not fully described in most eukaryotes [19] . While canalization and robustness research focuses on the benefits of decreasing within-genotype variance , there is evidence that increases in per genotype variance can also be beneficial . This is occasionally called the portfolio effect wherein the fitness of a genotype is determined by the portfolio of phenotypes that it can obtain [20] . In some bacterial settings rapid environmental fluctuations have been shown to favor the development of stochastic switching as the optimal means of response [21]–[25] . Similarly in eukaryotes , it has been shown that natural variation can alter stochastic noise of gene expression [19] , [26] and that stochastic noise in defense phenotypes could help to delay the evolution of counter-resistance in biotic pests [27] , [28] . As such , it is possible that there is wide-spread genetic variation controlling stochastic noise in eukaryotic phenotypes that may play a beneficial role in the evolution of these organisms [16]–[18] , [25] . However , little is understood about the genomic distribution of natural quantitative genetic variation for stochastic noise in eukaryotes or about the direction of selection on that natural variation . The concept that stochastic noise is genetically determined in a quantitative , polygenic manner is supported by the analysis of stochastic variation in expression of a MET17 reporter fusion construct in Saccharomyces cerevisiae [29] . This study identified significant genetic diversity regulating stochastic noise of gene expression and showed that stochastic noise was a complex trait controlled by at least three quantitative trait loci ( QTLs ) [29] . However , given the nature of these alleles , it is not known if these polymorphisms are present in wild populations or are laboratory derived . Additional evidence comes from the study of the S . cerevisiae galactose regulon where it was found that genetic manipulation of the regulatory feedback loop could lead to increased stochastic noise in the network's output [30] , [31] . Genetic control of stochastic noise has also been identified using QTLs for yield stability in crops [32] and gene expression in 18 isogenic mouse lines [19] . Further , it has been shown that HSP90 likely buffers genetic variation which could appear as stochastic noise in fluctuating environments , but little is known about the genomic distribution of natural variation in stochastic noise within a constrained environment [16]–[18] . These studies indicate that there is the genetic variation to regulate stochastic noise in physiology and gene expression suggesting that stochastic noise itself is a phenotype subject to natural selection with potential for pressure in both positive and negative directions . To begin testing the genomic extent of natural genetic variation in stochastic noise we used the model plant , Arabidopsis thaliana . Arabidopsis is quickly becoming a key organism in the study of complex traits through the use of systems biology and quantitative genomics approaches [33]–[40] . This is due to large repositories of transcriptomic and metabolomic data for homozygous QTL and association mapping populations that , when combined with whole genome sequence of natural accessions , provides the ability to rapidly develop and test hypotheses as well as find causal genes underlying specific loci of interest [41]–[44] . This has enabled the identification and validation of numerous genes and defense pathways under natural selection [45]–[50] . Among these defense mechanisms with known selective consequences are the glucosinolate metabolites , thioglucosides that provide defense against numerous biotic pests and whose accumulation is heritable and under balancing or fluctuating selection in the field [51]–[62] . This makes Arabidopsis an ideal system to search for the genetic and molecular basis of complex phenotypes , such as stochastic noise , in higher organisms . Using previous datasets , we identified QTLs that control natural variation in stochastic noise of glucosinolate metabolites and related transcripts within a single controlled environment . There were QTLs unique for the different phenotypic levels and we showed that known genes underlying these glucosinolate loci led to altered glucosinolate stochastic noise . We then extended this analysis to show that the Arabidopsis transcriptome shows significantly heritable stochastic noise for expression levels . Further , we were able to identify QTLs that control global stochastic noise in gene expression . Some loci were in common with those altering the average transcript abundance while others appeared unique to controlling transcriptomic stochastic noise . Using an existing single isogenic population , we confirmed that natural variation at the ELF3 locus alters stochastic noise in both physiological and metabolic phenotypes . Given the wide spread genomic variation controlling natural variation in stochastic noise in a single environment that we found within the wild Arabidopsis germplasm , our results suggest that any analysis of Arabidopsis evolution needs to account not only for genetic control of average phenotype value but also for genetic control of stochasticity . It remains to be determined how widely distributed this level of genomic natural variation exists for stochasticity within a wider range of multi-cellular eukaryotes .
To test if there is genetic variation affecting stochastic noise in the higher plant Arabidopsis thaliana we used a previous analysis of quantitative variation in glucosinolate defense metabolites [63] . The glucosinolate biosynthetic , transport and regulatory networks have been highly characterized [64]–[69] , providing extensive information about the loci responsible for differences in mean glucosinolates within Arabidopsis thaliana accessions [60] , [70]–[72] . Given the extensive knowledge it is possible to use existing glucosinolate data to search for QTLs controlling stochastic noise in glucosinolate accumulation . If stochastic noise QTLs are found they can be compared to existing analyses to determine if the same QTLs control phenotypic mean . A previous analysis of glucosinolate variation in the Arabidopsis thaliana recombinant inbred line population ( RIL ) derived from the Bayreuth ( Bay ) and Shahdara ( Sha; syn:Shakdara ) accessions [42] reported both the mean glucosinolate accumulation and standard deviation per line for three replicated experiments quantifying concentrations of 62 different glucosinolate phenotypes in 392 RILs [63] . We used this information to obtain the coefficient of variation ( CV ) of each glucosinolate phenotype for each RIL by dividing the standard deviation of the phenotype by its mean , and used this dimensionless measure of stochastic noise in glucosinolate accumulation to perform QTL analysis [21] , [26] . This identified five QTL hotspots controlling differences in glucosinolate CV ( Figure 1 ) . The pattern of CV QTL was similar to that found for QTL affecting differences in mean phenotype where GSL . ELONG and GSL . AOP are the major loci followed by two additional hotspots on chromosome 2 that had also been found to affect mean glucosinolates but were less significant than for glucosinolate CV ( Figure 1 ) [63] . Further , we found a new QTL for CV that was not found for the mean phenotype within this population but had previously been found as a glucosinolate QTL in other populations , GSL . MYB2976 ( Figure 1 ) [63] . There were also several QTLs that affected the mean phenotype but did not cause significant differences in glucosinolate CV ( Figure 1 ) [63] . Thus , it is possible to find QTLs controlling CV differences and these are not necessarily the same loci as those that affect the phenotypic mean . Fortunately , several of the identified QTLs have already been cloned and previously published single gene validation lines exist with published glucosinolate data to allow rapid validation of the CV phenotypes [63] , [67] , [73] . We have previously shown that the GSL . AOP locus is controlled by differential expression of two enzymes , AOP2 and AOP3 , which evolved from a tandem duplication event to control different reactions with the same precursor [63] , [74] . Using the same data set which previously showed that the QTL allele for increased glucosinolate accumulation and glucosinolate network transcript abundance was caused by expression of the AOP2 gene [63] , we showed that introducing the AOP2 gene into a natural knockout background ( Col-0 ) also significantly increased glucosinolate CV ( Figure 1B and C ) . This increase correlates with the elevated CV found in Sha , which contains the functional AOP2 allele at the GSL . AOP locus ( data not shown ) . The GSL . MYB2976 locus co-localizes with a previously cloned QTL from a different RIL population ( Ler x Cvi ) that is controlled by two glucosinolate transcription factors , MYB29 and MYB76 [67] , [72] , [73] . We used data from previous single gene manipulations of MYB29 and MYB76 as well as the related MYB28 , also linked to glucosinolate QTLs in other populations , to test if these genes can influence natural variation in glucosinolate CV [62] , [67]–[69] , [73] , [75] . Interestingly , increasing or decreasing MYB28 expression significantly increases CV for all glucosinolates ( Figure 1B and C ) . This is in contrast to previously published data showing that increasing MYB28 expression increased glucosinolate content while decreasing MYB28 expression correspondingly diminished glucosinolate content . Together this suggests that the effect of genetic variants on CV and mean is not always correlated [67]–[69] , [73] , [75] . In contrast to MYB28 , only increases in MYB29 and MYB76 expression altered metabolite CV while decreased expression at either gene had no impact on glucosinolate CV ( Figure 1B and C ) . This differs from their impact on mean glucosinolate accumulation where increases and decreases in all three gene expression lead to correlated increases and decreases in glucosinolate metabolites [67]–[69] , [73] , [75] . Interestingly , the natural variation in gene expression of MYB29 and 76 in the Bay-0 x Sha population appears to be a shift from a Col-0 like level in the Bay-0 genotype to elevated expression in the Sha genotype [40] , [76] agreeing with the observed introduction of a CV QTL in this position . It is possible absence of a MYB2976 QTL altering the mean phenotypes may be an issue of not having sufficient RILs to identify this locus in the background of the other QTLs showing significant epistatic interactions [63] . To test if the use of CV may be biasing our analysis , we used Levene's F-test to compare variances between the various mutants and WT and obtained similar results ( Figure 1 ) . In summary , MYB28 , MYB29 , MYB76 and AOP2 alter glucosinolate CV , mean and unadjusted variance ( Figure 1 ) [63] , [67]–[69] , [73] , [75] . Since MYB29 , MYB76 and AOP2 underlie CV QTLs , they are good candidates to control natural variation in glucosinolate CV within Arabidopsis thaliana . The observation that MYB28 and MYB29 perturbations have similar consequences upon mean glucosinolate accumulation but different influences on glucosinolate CV shows that the CV is not being driven by underlying changes in mean and is a valid approach for this analysis . The GSL . AOP and GSL . MYB2976 QTLs control differences in both the mean accumulation of glucosinolate metabolites and the relevant biochemical pathway transcripts [63] , [67] , [73] . Having found that these QTL controlled differences in CV for glucosinolate metabolites , we next tested whether these QTL also control differences in CV for transcripts involved in glucosinolate production . We used pre-existing microarray data [40] , [77] and found little evidence for impacts of the GSL . AOP and GSL . MYB2976 loci on the CV of transcript accumulation for individual transcripts in the GLS pathway ( Figure S1 ) , in contrast to their effect on CV for glucosinolate metabolites . Hereafter these loci are referred to as CV eQTL ( CV eQTL = a QTL altering the coefficient of variation in transcript accumulation ) to delineate them from standard eQTL ( eQTL = a QTL altering the mean transcript accumulation ) . Similarly , there was no evidence that these loci impact the GLS related biosynthetic networks ( Figure S2 ) . This is in contrast to previous observations showing that AOP2 , MYB29 and MYB76 can cause changes in glucosinolate pathway transcription and are known eQTL ( expression QTL ) hotspots for mean glucosinolate transcript abundance[63] . This is not entirely surprising as glucosinolate regulation shows extensive hallmarks of incoherent feed-forward loops [65] , [67] , [73] which can cause non-linear relationships in variance at different output levels . Thus the difference in CV partitioning between metabolites and transcripts at these loci is not entirely unexpected . Together , these data suggest that although the GSL . AOP and GSL . MYB2976 QTLs and the underlying causal loci ( AOP2 , MYB29 and MYB76 ) affect the mean transcript and metabolite abundance in the GLS pathway , and the CV of metabolite abundance , they don't alter the CV of transcript accumulation in this pathway . Interestingly , a hotspot on chromosome 2 , controls the per transcript CV abundance of most genes in the GLS pathway and CV in glucosinolate content ( GSL . ELF3 , Figure 1 ) . This locus fits the definition of a network CV eQTL as it alters the CV of the glucosinolate transcript network . While there was no network CV eQTL at the GSL . AOP locus , the AOP2 and AOP3 genes showed evidence for a cis positioned eQTL controlling the CV for transcript accumulation for only these two genes and not the entire pathway ( Figures S1 and S2 ) . Interestingly , not all glucosinolate associated transcripts known to have a large effect cis-eQTL also had a cis-CV eQTL . For instance , the GSL . MAM locus contains cis-eQTL for the MAM genes yet there was no corresponding cis-CV eQTL ( Figure S1 ) [63] . If our use of CV was solely tracking changes in mean abundance , the large effect cis-eQTL by default should have large effect cis-CV eQTL . The lack of this absolute relationship suggests that changes in mean are not driving changes in CV and supports the use of CV for mapping stochastic noise QTLs . Additionally supporting this is the fact that we utilized the same threshold estimation approaches for both CV-eQTL and cis-eQTL detection arguing against this being different statistical power issues [40] . The above analysis of existing glucosinolate quantifications suggests that there is significant genetic control of the CV for these defense metabolites . The CV itself may be under selective pressure to generate differences in stochastic variability between different natural populations of Arabidopsis [25] , [28] . To query if genetic control of phenotypic CV is a global phenomenon within Arabidopsis , we used a pre-existing dataset consisting of replicated microarray experiments conducted on 211 lines of the Bay x Sha RIL population and the RIL parents [40] , [77] . The distribution of CV across the transcripts was similar between Bay and Sha with a statistically significant difference of Bay showing a slight shift of the peak towards a higher CV ( Figure 2A ) . Interestingly , the distribution of CV across the transcripts was more distinctly bimodal within the RILs suggesting significant transgressive segregation in the population only impacting a specific subset of transcripts ( Figure 2A ) . The replicated nature of this experiment allowed us to directly assess the heritability of per line CV differences in both the parents and the RILs across 22 , 746 different transcripts representing the majority of the genome . The per transcript CV were correlated between Bay and Sha with an average heritability of 17% ( Figure 2B and Figure 3A ) . The average heritability of per transcript CV was much higher in the RILs than the parental genotypes with an average heritability of 57% ( Figure 2B ) . This is similar to the average heritability reported for the mean transcript abundance for the same experiment ( ∼68% ) with the majority of this difference being due to the lack of a high heritability tail for transcript CV as compared to heritability for mean transcript abundance [40] , [77] . As found previously for the mean transcript values , there was very little relationship between the heritability as measured in the Bay/Sha parents versus the RIL ( Figure 3B ) . For the mean transcript abundance , this discrepancy was explainable by transgressive segregation due to QTLs of opposing effect and is likely true for CV-eQTLs as well , suggesting that similar levels of robustness in the two parents are obtained via different genetic networks [40] , [77] . Supporting this is the observation that the standard deviation of transcript CV across the RILs is significantly greater than would be expected by modeling the expected CV using the parental values . In 1000 models , the maximal standard deviation of CV averaged across the transcripts in the RILs was 0 . 09 with a mean of 0 . 08 . In contrast , the actual biological values showed an average standard deviation of CV per transcript across the RILs of 0 . 17 indicating that the RILs show a significantly larger distribution of CVs per transcript per RIL than would be expected given the parental value . One concern with CV and any other estimate of variance is the potential for a correlation between variance and mean . The above analysis with glucosinolate accumulation did not suggest that this was a concern within Arabidopsis natural variation because we could identify instances where there were QTLs with large effect on the mean but no effect on the CV , even when identical approaches were used to determine significance thresholds . Within the RIL transcriptomic data , we did observe a statistically negative correlation ( P<0 . 001 ) whereby transcripts with the lowest average abundance had the highest CV and vice versa however this correlation explained only 0 . 8% of the total variation in CV leaving 99 . 2% of the variation to be available for genetic control of CV independent of the mean ( Figure 3C ) . This significant but minimal negative correlation likely derives from technical issues in microarrays surrounding the detection of lower expressed transcripts using Affymetrix microarray technology . To test if this technical issue constrains our ability to identify biologically controlled transcript CV , we compared the average per transcript expression to the heritability of per transcript CV within the RILs . This analysis showed that higher expressed genes had only a slightly more reproducible transcript CVs , therefore the technical issues surrounding low expressed genes does not impact our ability to identify biologically controlled CV ( Figure 3E ) . Additionally , the magnitude of per transcript CVs in the RILs showed very little relationship to the heritability of per transcript CV suggesting that any CV/expression level correlation is not creating relationships at higher levels ( Figure 3D ) . Thus , the use of CV to map QTLs for the transcripts appears to be valid . Interestingly , there was a strong negative correlation between the heritability of transcript abundance and the transcript CV , such that transcripts with the lowest CV had the highest heritability ( P<0 . 0001 , R2 = 0 . 59; Figure 3F ) . To ensure that the relationship between mean transcript abundance and transcript CV was not driving this correlation we repeated the analysis as a partial correlation while controlling for mean transcript abundance , this still showed a highly significant negative relationship between the heritability of transcript abundance and the transcript CV ( P<0 . 0001 , R2 = 0 . 42 ) . Together , this suggests that quantitative genetic control of CV is a genome wide phenomenon within Arabidopsis thaliana that is not limited to defense metabolites and is at least partially independent from genetic variation controlling the mean phenotype . The high estimated heritability of per transcript CV within the Bay x Sha RIL population suggests that it is possible to map CV eQTL for all transcripts . We used composite interval mapping to map CV eQTL for all 22 , 746 transcripts within 211 lines of the Bay x Sha population previously used to map eQTL [40] , [77] . This identified 98 , 014 significant CV eQTLs that altered the stochastic noise for 21 , 974 transcripts for an average of nearly 4 CV eQTL per transcript ( Figure S3 ) . This is nearly twice the number of eQTLs per transcript found using the average transcript abundance as a phenotype [40] . This difference may be due to the use of two different experiments in the CV eQTL analysis , whereas the eQTL analysis used just one experiment , reducing its statistical power [40] . Given that we used identical methods to identify global permutation thresholds for both datasets , we do not feel that a higher false positive rate can explain the elevated number of CV eQTLs [40] , [78]-[80] . In addition , the elevated number of CV eQTLs is not universal as the glucosinolate transcript measurements actually identified more eQTLs than CV eQTL ( Figure S1 ) [63] . Thus , the elevated CV eQTL level may be more indicative of the specific biological process within which that transcript functions . An analysis of the distribution of additive effects for the CV eQTL showed a slight bias towards Bay alleles having a negative impact on CV ( 50429 CV eQTLs with Bay additive effect <0 versus 47585 with Bay additive effect >0 ) ( Figure 4A ) . The vast majority of CV eQTLs had absolute effects less than 0 . 1 CV and these were almost entirely acting in trans ( Figure 4A and B ) . In contrast , CV eQTL with absolute effects greater than 0 . 1 were predominantly acting in cis ( Figure 4B ) . This is similar to eQTL controlling the mean accumulation of a transcript where on average trans-eQTLs have smaller additive effects than cis-eQTLs [38] , [40] . This analysis identified 3 , 720 transcripts as having a cis-CV eQTL , in contrast with the 5 , 127 transcripts having a cis-eQTL for mean expression level ( Figure 4 ) [40] . While about ¼ of all eQTLs detected were cis , only 1/26th of all CV eQTL were cis , showing that natural variation at trans positions is dramatically more prevalent in controlling transcript CV than average expression ( Figure 4 ) [40] . As expected by the decreased ratio of cis-CV eQTL relative to that found for eQTL , the cis diagonal , while present , was very faint ( Figure 5B ) . Only 1 , 660 transcripts had both a cis-eQTL and cis-CV eQTL and these included nearly all of the large effect CV eQTLs ( Figure 4 ) [40] . Thus , while a cis-eQTL can be associated with a cis-CV eQTL , it is not a necessity ( Figure 4 ) . These results show that stochastic noise measured as CV in transcript abundance is a highly heritable trait suitable for genome wide QTL analysis in multi-cellular eukaryotes . As in eQTL analyses of mean transcript abundance , differences in the CV of transcript accumulation seem to be broadly caused by abundant loci acting in trans , while substantial changes are less frequent and usually associated with variation in cis . We counted the number of loci per chromosomal position controlling stochastic noise within the Arabidopsis transcriptome to better understand the genomic distribution of CV eQTLs ( Figure 5A ) . This identified a number of locations within the genome that contain trans-hotspots for CV eQTL . Several of these were in common with eQTL trans-hotpots that had previously been identified such as the locations on Chromosome II . However , the relative impact of the trans-hotspots upon the transcriptome was different for the two traits ( Figure 5A ) [40] . For instance , the trans-hotspots at 12 and 42 cM on chromosome II caused similar numbers of eQTL , yet the hotspot at 42 cM affected many more CV eQTLs than the hotspot at 12cM . Additional hotspots were detected with CV eQTL that were not detected using mean transcript accumulation , most notable is the locus at the bottom of chromosome III that is the highest trans-hotspot for CV eQTL but barely registered for eQTL ( Figure 5A ) [40] . Two other apparent CV eQTL specific trans-hotspots were peaks over the permutation threshold near the GSL . AOP and GSL . MYB2976 loci on chromosomes IV and V ( Figure 5A ) [40] . However , none of the glucosinolate transcripts' CVs were regulated by the trans-CV eQTL hotspots near GSL . AOP and GSL . MYB2976 ( Figure S1 ) . This raises the question of whether these CV loci near GSL . AOP and GSL . MYB2976 are due to pleiotropic consequences of the metabolic CV controlled by GSL . AOP and GSL . MYB2976 ( Figure 1 ) or if there are additional genes in these regions that alter transcriptomic CV . The detected CV eQTL hotspots have additive effect biases , with most of the CV eQTLs in one hotspot increasing the CV in the same direction , as noticed before for eQTL hotspots ( Figure 5B ) [40] . The two major hotspots had opposite effects; with the Sha allele causing increased stochastic noise at the hotspot in chromosome III and decreasing stochastic noise in all hotspots on chromosome II ( Figure 5B ) . = This observation further shows that increased mean abundance does not inherently cause increased CV . Thus , transcript mean abundance and CV are not measures of a single phenotype and instead can involve different genetic mechanisms even when investigating the same locus . The global effect of trans-CV eQTL hotspots led us to test if we could directly map QTL controlling genome-wide transcriptomic CV ( as opposed to per transcript CV ) . Taking the average CV across all transcripts showed that Bay and Sha have different CV and that the main source of this is the previously identified loci on Chromosome II and III ( Figure 6 ) . Thus , these loci appear to have genome wide effects upon stochastic noise of gene expression and likely other traits . The chromosome II locus found using the global CVs of transcript abundance , glucosinolate accumulation and glucosinolate network expression maps close to the previously identified ELF3 QTL ( Figure 1 , Figures S1 and S2 ) [81] . Allelic variation in ELF3 between Bay and Sha has been shown to affect circadian rhythms and shade avoidance responses but not the wave form of the circadian oscillation [81] . We next wanted to test if the ELF3 locus could be the same as the global CV eQTL hotspot . Because of ELF3's involvement in the circadian clock , we first asked whether we could identify stochastic noise QTL for circadian rhythms in the Bay x Sha population and whether these QTL would overlap the ELF3 region . Circadian rhythms in transcript abundance have been measured in this population [82]; we used this same approach to map CV for the expression of circadian clock regulated genes . Briefly , transcripts previously identified as being regulated by the circadian clock were grouped into 24 CT phase groups based upon each transcript's time of peak expression ( CT ) during the 24 hour photoperiod [82]-[84] . Transcript expression values were then Z normalized and a single expression estimate was independently obtained for each CT phase group for each microarray . These were then used to estimate the variance of the CT phase groups expression as described . Both the ELF3 locus and the chromosome III hotspot were found to alter CV for gene expression across the circadian clock output networks with opposing effects as had been found for general gene expression ( Figure 5 and Figure 7 ) . In contrast , the other identified trans-CV eQTL hotspots ( Figure 5 ) , do not appear to influence the CV of transcripts regulated by the circadian clock ( Figure 7 versus Figure 5 ) . To test if ELF3 is the causative gene controlling stochastic noise in this region we utilized previously generated Col-0 elf3 . 1 knockout mutants lines containing a CCR2:luc reporter gene that were rescued with the genomic Bay and Sha ELF3 alleles ( elf3:Bay-0 and elf3:Sha ) [81] . Since Bay and Sha ELF3 genomic alleles have been shown to affect the period of CCR2:luc oscillations in free running conditions under different light environments [81] , we monitored the CV in period in at least 650 T1 plants per transgene distributed in 10 independent experiments performed in constant red or in constant red plus far red light . Independent of light conditions we found that the Sha ELF3 allele reduced stochastic noise in the circadian oscillation period , in agreement with the direction of the global CV eQTL and circadian CT phase group QTL at the ELF3 position ( Figure 7 and Figure 8 ) . Although plants in both red and red plus far red light presented lower CV ( ( P = 0 . 002 in red light versus P = 0 . 043 in red plus far red light , via ANOVA ) , the difference in CV between the two alleles was not significantly affected by the light treatment ( Figure 8 , P = 0 . 35 via ANOVA ) . The Bay and Sha alleles of ELF3 did not affect CV for amplitude , phase or quality of the rhythms ( measured as the relative amplitude error ) in the transgenic plants ( P = 0 . 10 , P = 0 . 18 and P = 0 . 50 respectively , data not shown ) . To further test if ELF3 , could be the gene underlying other the CV QTL identified for other phenotypes at this locus , we tested if the transgenic lines differed in the level of stochastic noise for glucosinolate metabolites ( Figure 1 ) . Different alleles of ELF3 led to changes in glucosinolate stochastic noise with the Sha ELF3 allele increasing stochastic noise for the short chain aliphatic glucosinolate 4-methylsulfinylbutyl ( 4-MSOB ) but decreasing stochastic noise for the long chain aliphatic glucosinolate 7-methylsulfinylheptyl ( 7-MSOH ) ( Figure 1 and Figure 8 ) . Since the different alleles of ELF3 ( Bay v Sha ) have also been shown to affect flowering time [81] , we measured two traits related to this character in the transgenic lines and found that variation between the ELF3 alleles led to differences in stochastic noise for flowering ( Figure 8 , Figure S4 ) . The observation that the Sha allele of ELF3 led to higher stochastic noise in flowering time and 4-MSOB accumulation , whereas it was also associated with lower stochastic noise in circadian periodicity and 7-MSOH accumulation suggests that ELF3 is not simply making the plant more or less robust but instead is partitioning noise between specific phenotypes ( Figure 8 ) . Interestingly , this differential effect of ELF3 upon stochastic noise agrees with the observed CV eQTL at this locus . : The Sha allele at the ELF3 QTL was associated with decreased stochastic noise of transcriptional networks for circadian genes and most glucosinolate networks but the Sha allele had increased stochastic noise in the FLC ( At5G10140 –Flowering locus C ) and GS-OX2 ( At1g62540 ) transcripts ( Table S1 ) [85]–[88] . The increased noise in FLC nicely correlates with the observed flowering time noise . Furthermore , GS-OX2 is required to synthesize 4-MSOB and concordantly links to the increased noise in this metabolite . Interestingly , YUCCA3 ( At1g04610 ) transcript accumulation also shows a CV eQTL at the ELF3 locus suggesting a potential impact on auxin by this locus [89] . In summary , our results show that natural variation in ELF3 leads to changes in stochastic noise in both plant and molecular phenotypes and that the direction of effect depends upon the specific phenotype . Therefore , ELF3 is not a gene leading to plants displaying a general increase in phenotypic noise but instead affects noise in a network specific manner . Finally , it should be noted that there is no measurable difference in gene expression between the Bay and Sha alleles at ELF3 showing that these altered stochastic noise phenotypes in metabolism , transcription and physiology are dependent upon the biochemical differences in the two alleles [40] , [90] .
Stochastic noise is frequently divided into that which comes from sources internal to the organism ( intrinsic ) and environmental sources external to the organism ( extrinsic ) [26] . In unicellular organisms it is possible to use internal reporters and massive population sizes to begin to partition the two sources . This is much more difficult for multi-cellular organisms . However , in this study , a number of findings support that we are likely measuring largely intrinsic sources of noise rather than purely extrinsic sources . The first is that our measures of noise are correlated with the genotype of the organism , which would not have been true if the variance we were measuring was purely extrinsic/environmental noise . It could be argued that we are mapping genetic variation that leads to differential sensitivity to extrinsic noise . However , each experiment is highly replicated so to be mapping differential sensitivity to extrinsic noise would have required the sources of extrinsic noise to be the same in each experiment and to show similar variations across the experiments . While we can not entirely rule out this possibility , it is much more likely that we have identified loci controlling natural variation in intrinsic stochastic noise within a multi-cellular organism . An interesting observation in this data is that there is an unexpectedly high genomic level of natural genetic variation controlling stochastic noise in transcripts , metabolites and physiology . The high frequency of trans-CV eQTL rules out the possibility that this is simply the mapping of large effect indel polymorphisms that would be expected to alter transcript CV in cis . Additionally , the finding that the genes underlying trans-CV eQTL also control stochastic noise in metabolites and complex physiology such as the circadian clock shows genetic control of stochastic noise impacts all levels of the plant . Interestingly previous reports have shown that HSP90 could be expected to control stochastic noise in numerous Arabidopsis phenotypes but we did not identify any trans-CV-eQTL hotspots linked to any of the known HSP90 genes [16]–[18] , [92] . This suggests that natural variation in HSP90 is not a major driver of stochastic variation within this Arabidopsis population for this environment . It is possible that if we had used multiple environments that natural variation in HSP90 may have been identified but this was not the case . Our findings raise the question of what genetically variable control of noise means in an ecological and evolutionary context . One possible answer would be that this genetic control of noise is meaningless because stochastic noise may not be under selection . However , this answer runs up against two impediments . The first is that in bacteria , natural variation in phenotypic stochastic noise has been shown to be adaptive under situations where the environment is highly unpredictable [25] , [91] , similar to that found in plant/herbivore interactions [28] . Additionally , several of the glucosinolate loci , including the GSL . AOP2 locus that we show controls stochastic noise in glucosinolates , have been shown to be under selection in Arabidopsis and other related species [52] , [56] , [60] , [93]–[100] . While these findings do not show that the stochastic noise variation is directly under selection pressure , it is clearly controlled by genetic loci that are themselves likely under selection pressure . Further , this suggests that selection is not solely focused upon decreasing stochastic noise within non-stressful environments especially for defense related traits . The next question then becomes how natural variation in stochastic noise within environments that are not overtly stressful could benefit a multi-cellular organism . The answer to this might come in the form of a question that is related to the interest in identifying the genetic basis of local adaptation . However , the term local adaptation always engenders the response “what is local ? ” . It is possible that altering the stochastic noise of a system could alter the range of environments where it can successfully function . For instance , increasing the stochastic noise of the circadian period may enable that particular genotype to occupy more longitudinal niches , albeit at the likely cost of never being the optimal genotype in any specific niche . In contrast , decreasing the stochastic noise of the circadian period would optimize the fitness in a specific niche but likely at the loss of fitness across other niches . In this instance , natural variation in stochastic noise could lead to genetic control over what constitutes local for a specific genotype . As such , it may not be the variance itself that is adaptive , but instead the ability of variance to produce a more flexible network . In contrast , stochastic noise in defense metabolites , such as glucosinolates , could represent a different benefit of natural variation in CV . Glucosinolates are a major anti-herbivore and anti-pathogen defense of Arabidopsis and relatives [53] , [54] , [58] , [59] , [98] , [99] and as such could impart a pressure upon these herbivores and pathogens to counter adapt [101] . One mechanism that has been suggested as effective in slowing counter-adaptation is to increase the unpredictability of the defense compound ( i . e . stochastic noise ) [27] . As such , genetic control on the stochastic noise of defense compounds could in and of itself provide direct benefits to the efficaciousness of the defense . However , the observation that there is natural variation in stochastic noise of defense metabolites would suggest that high levels of noise are not always beneficial , possibly depending upon the ratio of generalist and specialist herbivores in a given genotype's normal locale [98] . Testing these different potential benefits of stochastic noise will require the development of genotypes that differ solely in stochastic noise to allow this effect to be partitioned away from any influence upon the mean phenotype . A major difficulty in systems biology is the presence of massive datasets that are largely correlative when comparing different transcripts . This has lead to numerous attempts to derive causal information from these correlative datasets . However , even the best approaches are susceptible to a number of systemic errors that deal with predicting regulatory loop structure as well as combinatorial regulation [102] . For regulatory loops , correlative approaches using average responses generate a number of possible network topologies that are similar with respect to regulation of phenotypic average , but that make very different predictions about how perturbations will control the stochastic noise of the system [5] , [6] , [103]–[105] . Given this , it may be possible to use the presence of genetically controlled stochastic noise to help better refine systems biology models . The mean transcript , protein , or metabolite levels could be used to generate multiple initial models that could then be analyzed by using the stochastic noise in the system to determine which model most accurately predicts the observed stochastic noise . Future work on this approach could be useful but would require true independent replication in systems biology experiments to allow accurate estimations of stochastic noise for each measured phenotype . The identification that stochastic noise of phenotypes has a level of genetic control that appears to be on par with that observed for the phenotype average suggests that there is a fount of phenotypic information that has largely not been studied in most modern genetic , genomic or systems biology studies . For instance , numerous natural and induced mutant screens and surveys have been conducted in Arabidopsis to determine the genes controlling the phenotypic average [106]–[108] . Similar large scale approaches have been conducted in numerous other organisms focused on phenotypic averages [109]–[111] . While these have provided great advances in our understanding of biology , it raises the question of what would happen if we repeat these screens and surveys to identify genetic variation controlling stochastic noise in phenotypes . Would we identify the same genes or would we begin to identify a large suite of previously unknown genes that control stochastic variation rather than phenotypic average ? Experiments focused on the stochastic nature of a phenotype require independent replication but could yield a new view of organismal biology that is currently specified by our focus upon phenotypic averages .
To directly estimate the CV for each individual genes transcript accumulation ( 22 , 746 transcripts in total ) as a separate phenotype within the Bay x Sha RIL population [42] , we obtained two independent microarray experiments ( TABM-224 and TABM-518 ) wherein 211 RILs were each measured in duplicate within each experiment providing four replications [40] . Raw image data from the RIL GeneChips were converted to numeric data via Bioconductor software ( www . bioconductor . org ) . We utilized quantile normalization across all arrays to reduce non-biological variation coming from the technology itself , and when applied at the probe level it has been shown to outperform other normalization methods that are based on what is referred to as a “base-line array” [112] . After quantile normalization , we utilized the absolute expression values to measure the CV for each gene separately for each experiment using σ/µ [21] , [26] , [113] , thus providing two independent biological replicate measures of CV for each gene . The use of CV as a direct phenotype has previously been used in a number of instances . By measuring the within line CV as a phenotype for the Bay x Sha population allows us to then utilize CV as a direct measurement of stochastic variation as a phenotype . The level of per line replication for the array data does not support the use of Levene's variance tests or measures . Additionally , all lines were planted and harvested within a randomized complete block design at all stages thus limiting any potential technical bias to generate these observations [40] , [77] . To estimate the CV for specific transcript networks , we utilized a previously published approach whereby we average the expression across a group of genes to provide an estimate of the gene network's expression value [38] , [114] . Briefly , this network approach uses any a priori defined group of genes as a network . Every transcript that is defined within a network is z transformed to place them all on the same scale . For every microarray within the dataset , the network expression value is obtained by averaging across the z values for all transcripts within the network . This provides a single network value that can then be utilized for downstream applications . This approach has previously been used to map network QTL controlling the difference in average expression [63] and can be extended to identify differences in network stochasticity using the CV value instead of the average expression . Gene membership within specific circadian networks were defined as previously described [83] . Gene membership within glucosinolate pathways were defined as previously described [63] , [67] . This approach was also used to generate a global CV average by averaging the CV across all 22 , 746 transcripts measured on the ATH1 Affymetrix microarray . To estimate the CV for specific defense metabolites , we utilized previously published data wherein the µ and σ for a large set of glucosinolates within a Bay x Sha RIL population consisting of 403 lines had been measured [63] . The glucosinolates were measured in a similar growth stage and growth chamber as that for the transcriptomics analysis allowing for better comparison between the datasets [40] , [63] . For measuring altered glucosinolate metabolite CV in the independent transgenic lines , we compiled data from multiple independent experiments that had previously been published in separate papers . We analyzed the same lines in at least four independent experiments with replication allowing us to test if the CV differed across these genotypes [58] , [63] , [67] , [73] . Glucosinolate genotype analysis: To test if variation at specific glucosinolate genes could alter the CV of either metabolite or transcripts , we obtained previously published data involving multiple independent biological replicates for the following genotypes all of which are generated within the Arabidopsis Col-0 accession background . To elevate MYB gene expression , we used previous lines where the Arabidopsis Col-0 versions of MYB28 , 29 and 76 were separately introduced back into Arabidopsis Col-0 using a 35S promoter to induce their expression – 35S:MYB28 , 35S:MYB29 or 35S:MYB76 [73] . To mimic natural variants that have low to no expression of MYB28 , 29 or 76 , we used previously obtained insertional T-DNA mutants within each of these genes obtained from the Arabidopsis Col-0 accession; myb28-1 , myb29-2 and myb76-1 [67] , [73] . All insertional T-DNA mutants underwent at least one backcross and had previously been shown to abolish or dramatically diminish MYB gene expression [67] , [73] . To mimic the natural variation at the AOP2 locus , we utilized the Arabidopsis Col-0 accessions that contains a natural knockout of AOP2 and introduced the functional enzyme encoding gene back into this natural null background [63] , [115] . Thus , all of these lines are single gene manipulations of major glucosinolate loci within a common genomic background , Col-0 . For estimating broad-sense heritability , we utilized the independent measures of CV directly as a phenotypic measure . This allowed us to estimate broad-sense heritability ( H ) for each CV phenotype as H = σ2g/σ2p , where σ2g is the estimated CV phenotypes genetic variance among different genotypes in this sample of 211 RILs , and σ2p is the CV phenotypic variance for each phenotype [116] . Heritability was estimated for all expression phenotypes . The metabolite phenotypes did not have the individual values from each independent experiment , and therefore , heritability was not measurable . To map QTL for the CV phenotypes , metabolic , network and individual gene expression , we measured the average CV for each phenotype across all experiments and used the average CV in conjunction with a previously generated map for 211 Bay x Sha RILs ( [42] , [77]; see also the file “Average CV per transcript per RIL” at http://plantsciences . ucdavis . edu/kliebenstein/TableS1Plosgenetics . txt [note: this file is ∼28 MB] ) . For glucosinolates , we utilized a larger collection of 400 Bay x Sha RILs [42] . Composite interval mapping ( CIM ) analysis [117] was employed in conjunction with the 5 cM framework map . The “zmapqtl” CIM module of QTL-Cartographer Version 1 . 17 [118] with a walking speed of 1 cM and a window size of 10 cM was employed to analyze each phenotype . To obtain a threshold criterion for declaring statistically significant eQTL , a global permutation threshold was obtained by permuting the e-traits while maintaining the genetic information [40] . For each of 100 randomly selected phenotypes , the null distribution of the maximum likelihood ratio test ( LRT ) statistic was empirically estimated using permutation thresholds based on 1 , 000 permutations [40] , [78]–[80] . We then utilized the 95th percentile permutation threshold across the 100 null distributions [40] . We utilized the resulting output to localize , summarize and count CV QTL using the Eqtl module of QTL-Cartographer in conjunction with the previously optimized 5 cM exclusionary window where no CV QTL can be closer than 5 cM to the nearest QTL ( Table S1 ) [40] , [118] . This is an identical approach at all stages to that used to previously map the eQTL for this dataset and as such should increase the direct comparison between datasets [40] . Additionally , we have been able to clone and biologically validate causal loci controlling several of the trans effect loci controlling subtle shifts in physiological networks as identified from the eQTL analysis [90] , [119] . We have previously shown that single-feature polymorphisms are not a significant difficulty in this population for this array data when estimating expression values [40] , [77] . As such , we did not control for potential single-feature polymorphism issues . The low level of cis-CV-eQTL within our results further supports this observation . To determine whether a genetic location associated with multiple CV QTLs was a significant cluster or ‘hotspot’ , we estimated a significance threshold using permutation as previously described for transcriptomic data [40] , [90] , [109] , [120]–[123] . The positions of the 98 , 014 CV eQTLs ( Table S1 ) at the marker intervals were permuted across the genome 1 , 000 times , and the maximal number of CV eQTLs per genetic position per permutation was obtained . Using the distribution of the maximum number of CV eQTLs , the criterion for declaration of a significant eQTL hotspot was 422 CV eQTLs per genetic position at alpha = 0 . 05 . The permutated hotspot approach has been used to identify genes that cause the transcriptional difference for a number of hotspots showing that this approach is identifying biologically validatable effects [39] , [63] , [120] , [124] , [125] . elf3-1 null mutants carrying the CCR2::luc reporter gene were obtained from Dr . Stacey Harmer ( University of California , Davis ) . Full genomic clones of ELF3 from Bay and Sha including 1 . 5 kb of upstream promoter were cloned in pJIHOON212 . elf3-1-CCR2::luc plants were transformed with these constructs using Agrobacterium tumefaciens [81] , [126] . To account for differences between elf3 . 1: Bay and elf3 . 1:Sha due to the transformation protocol , transgenic plants obtained from two independent batches were used , but no effect of the Agrobacterium inoculate was detected ( P = 0 . 31 via ANOVA , data not shown ) . elf3:Bay and elf3:Sha transgenic T1 seeds from two different Agrobacterium transformation batches were placed on MS medium with the appropriate antibiotic and stratified for 4 days ( 4°C , dark ) . After entrainment under white light in 12∶12 photoperiods for 7 days , resistant plants were transferred to new MS plates and moved to continuous red light or red + far-red light conditions , where luminescence was recorded for 6 to 7 days . Five independent experiments were conducted in continuous red light ( R , total PAR of 64 uE ) and 5 experiments in continuous red plus far red light ( R+FR , total PAR of 64uE with a R:FR ratio of 0 . 5 ) conditions created with LED lights . Plants were monitored using a CCD camera taking pictures every 2 hours . The data collected was analyzed for rhythmicity using the luciferase activity method described in [127] . Only plants showing stable rhythms ( Relative Amplitude Error below 0 . 5 ) were considered for the analysis . Between 12 and 150 T1 plants ( average 75 . 2 , median 86 ) for each transgene were included in each experiment . Coefficient of variance was calculated as the standard deviation divided by the mean period estimate for each transgenic line in each experiment . elf3:Bay and elf3:Sha transgenic T1 seeds from three different Agrobacterium tumefaciens transformation batches were planted on soil including elf3 . 1 mutants and WT Col-0 as a control . The extreme hypocotyl length , flowering time and cotyledon color phenotypes of the elf3 . 1 mutants were assessed to distinguish transformed from untransformed plants [128] . Transformed plants were grown for 25 days in a 10 hour photoperiod . At 25 days , leaf tissue was harvested from each plant and individually extracted and assayed via HPLC for glucosinolate identity and concentration as previously described [72] , [115] . The experiment was replicated 9 times for a total of 106 elf3:Bay and 108 elf3:Sha independent T1 plants . Levene's F-tests were used to compare variance between the two T1 genotype classes .
|
Understanding how genetic variation controls phenotypic variation is a fundamental goal of biology in both modern medicine and agriculture . Yet , frequently , even a large set of genetic polymorphisms do not fully explain variance of a phenotype within a discrete set of individuals . Numerous mechanistic theories have been proposed , e . g . epigenetics , but we postulated that there may be genome-wide polymorphism controlling phenotype stochastic noise among genotypes . This is similar to what is being found in studies of bet-hedging theory in prokaryotic or single-celled organisms or stability in eukaryotes . Utilizing Arabidopsis , we tested this hypothesis at a genomic level by mapping quantitative trait loci for stochastic noise in global transcriptomics , plant defense metabolism , circadian clock oscillation , and flowering time within a single non-stressful environment . We cloned and validated a set of genes including transcription factors and enzymes that control natural variation in phenotypic noise . These genes provided evidence that stochastic noise can vary independently of average phenotypes . Since population genetic models and quantitative genetic studies focus on natural genetic variations impact upon average phenotypes , these observations suggest that stochastic noise needs to be incorporated to better explain the genotype-to-phenotype link .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome",
"expression",
"analysis",
"plant",
"biology",
"gene",
"regulation",
"population",
"genetics",
"quantitative",
"traits",
"epistasis",
"genome",
"analysis",
"tools",
"model",
"organisms",
"trait",
"locus",
"analysis",
"molecular",
"genetics",
"arabidopsis",
"thaliana",
"genome",
"complexity",
"genetic",
"polymorphism",
"plant",
"genetics",
"biology",
"systems",
"biology",
"plant",
"and",
"algal",
"models",
"phenotypes",
"heredity",
"genetics",
"genomics",
"gene",
"networks",
"genetics",
"and",
"genomics",
"complex",
"traits"
] |
2011
|
Genomic Analysis of QTLs and Genes Altering Natural Variation in Stochastic Noise
|
Histone H3 lysine 9 ( H3K9 ) methylation is associated with gene repression and heterochromatin formation . In Drosophila , SU ( VAR ) 3–9 is responsible for H3K9 methylation mainly at pericentric heterochromatin . However , the histone methyltransferases responsible for H3K9 methylation at euchromatic sites , telomeres , and at the peculiar Chromosome 4 have not yet been identified . Here , we show that DmSETDB1 is involved in nonpericentric H3K9 methylation . Analysis of two DmSetdb1 alleles generated by homologous recombination , a deletion , and an allele where the 3HA tag is fused to the endogenous DmSetdb1 , reveals that this gene is essential for fly viability and that DmSETDB1 localizes mainly at Chromosome 4 . It also shows that DmSETDB1 is responsible for some of the H3K9 mono- and dimethyl marks in euchromatin and for H3K9 dimethylation on Chromosome 4 . Moreover , DmSETDB1 is required for variegated repression of transgenes inserted on Chromosome 4 . This study defines DmSETDB1 as a H3K9 methyltransferase that specifically targets euchromatin and the autosomal Chromosome 4 and shows that it is an essential factor for Chromosome 4 silencing .
Methylation of conserved lysine residues on histone H3 and H4 tails plays a key role in gene regulation , chromatin structure , and establishment and maintenance of epigenetic memory ( reviewed in [1] ) . As proposed by the “histone code” hypothesis [2] , these marks , in association with other modifications , are interpreted by chromatin-specific regulatory complexes that in turn influence chromatin structure and its accessibility to transcription factors . Euchromatin is characterized by histone H3 methylated at lysine 4 ( K4 ) , K36 , and K79 , while heterochromatin is characterized by histone H3 methylated at K9 and K27 and histone H4 methylated at K20 [1] . Moreover , histone methylation can be present in mono- , di- , or trimethylation states [3 , 4] . All but one enzyme responsible for histone lysine methylation share an evolutionary conserved domain of about 130 amino acids , called the SET domain [5 , 6] . Numerous SET domain-containing proteins responsible for methylation of specific residues have been described in all eukaryotic organisms ( reviewed in [7] ) . Enzymes with histone demethylase activity were only recently characterized [8] . In Drosophila , similarly as in other organisms , histone H3 lysine 9 ( H3K9 ) methylation plays a crucial role for heterochromatin formation and maintenance and for gene silencing . Methylated H3K9 is a docking site for the recruitment of the heterochromatin protein 1 ( HP1 ) through its chromodomain [9–11] . Su ( var ) 3–9 was the first H3K9 methyltransferase characterized in Drosophila [12] . It was historically identified in genetic screens , together with Su ( var ) 2–5 encoding HP1 and Su ( var ) 3–7 , as a haplo-suppressor and triplo-enhancer of position effect variegation [13] , a phenomenon that reflects the mosaic heterochromatin-induced silencing of genes . SU ( VAR ) 3–9 is responsible for H3K9 dimethylation at the chromocenter and trimethylation at the core of the chromocenter , but not for H3K9 monomethyl marks at the chromocenter and along the chromosome arms , nor for the dimethyl marks at the chromosome arms , telomeres , and Chromosome 4 [14 , 15] . Recently , Drosophila dG9a was shown to display H3K9- as well as H3K27- and H4-methyltransferase activity , to localize at discrete bands in euchromatin , and to be excluded from Chromosome 4 [16 , 17] , suggesting that it methylates H3K9 at euchromatic sites . But the histone methyltransferases ( HMTases ) that methylate H3K9 outside the chromocenter have not been formally characterized . Drosophila melanogaster's Chromosome 4 is the smallest autosome and displays a peculiar chromatin organization ( for a review on Chromosome 4 see [18] ) . It is mostly heterochromatic , composed of a highly condensed ~3–4-Mb centromeric region that is under-replicated and a 1 . 2-Mb polytenized arm exhibiting a banded pattern . The banded region displays characteristics typical of heterochromatin based on a number of criteria: transposable and repetitive elements are represented at a high density [19 , 20] , P elements often display a variegated expression [21–23] , H3K9 dimethyl marks are present [14 , 15] , and HP1 is distributed in a banded pattern [24] . Surprisingly , in opposition to these heterochromatic characteristics , the banded portion shows a gene density comparable to euchromatin; many of these genes are essential , therefore expressed during development [25 , 26] . In addition , the H3K9 dimethyl mark is not deposited by the heterochromatic SU ( VAR ) 3–9 [14 , 15] . These features converge to the conclusion that chromatin of the Chromosome 4 banded region is different from centromeric heterochromatin . Human SETDB1 ( mouse ESET ) is an essential H3K9 methyltransferase involved in silencing in euchromatin [27–30] . It is composed of a Tudor- , a methyl CpG binding- ( MBD ) , and a bifurcated SET-domain that is surrounded by pre- and post-SET domains [31] . Recently , the D . melanogaster homologue gene of SETDB1 ( named dsetb1 , eggless , or dEset ) was identified; the domains characteristic of mammalian SETDB1 are well conserved , reaching 76% identity in the SET-C terminus and post-SET domains [16 , 32 , 33] . In addition , an histone deacetylase-interacting domain was identified [33] . This gene was shown to be involved in H3K9 trimethylation both in germ and somatic cells of the germarium and to be required for oogenesis at early stages of egg chamber formation [32] . Here , we show that DmSETDB1 is the missing euchromatin- and Chromosome 4-specific H3K9 HMTase . We generated a DmSetdb1 mutant allele and a 3HA-tagged DmSetdb1 allele by homologous recombination and show that this gene is essential for fly viability and that the endogenous DmSETDB1 protein localizes mainly at Chromosome 4 . In addition , we evidence that DmSETDB1 is responsible for some H3K9 mono- and dimethyl marks in euchromatin , as well as for Chromosome 4 H3K9 dimethylation . Moreover , DmSETDB1 turned out to be required for repression of variegating transgenes inserted on Chromosome 4 , a function that is consistent with the role of DmSETDB1 in Chromosome 4 H3K9 dimethylation . Therefore , DmSETDB1 is a key H3K9 methyltransferase in Drosophila involved in repression of the peculiar Chromosome 4 .
The open reading frame ( ORF ) of CG30426 was identified by protein BLAST search ( National Center for Biotechnology Information [NCBI] , http://www . ncbi . nlm . nih . gov ) as the closest Dm homologue of the human H3K9 methyltransferase SETDB1 . Others also identified CG30426 by protein BLAST or in a SET-domain phylogenic tree as the Dm homologue of SETDB1 [16 , 34] . Subsequently , CG30426 and the neighboring CG30422 ( see representation Figure 1A ) were shown to produce a single 3 . 9-kb mRNA transcript in ovaries [32] and constitute a single transcript in females [33] , suggesting that the DmSetdb1 gene is composed of both CG30422 and CG30426 . As a full insert cDNA corresponding to CG30426 alone ( AT13877 ) is present in public databases , we addressed whether the DmSetdb1 gene was transcribed from several transcription start sites , subjected to alternative splicing in a tissue-specific manner , or if the 3 . 9-kb transcript was the unique product . Northern blot analysis shows that in embryos , third instar larvae , male and female adults , a single 3 . 9-kb transcript is detected with a probe specific for CG30426 , with a stronger signal in embryos ( Figure 1B ) . The same profile is obtained with a probe spanning CG30422 ( unpublished data ) . Therefore , DmSetdb1 is expressed as a 3 . 9-kb transcript encompassing both CG30422 and CG30426 , which is present at all developmental stages and encodes a 1 , 261-amino acid protein . To study DmSETDB1 function in vivo , we generated the DmSetdb110 . 1a mutant allele by homologous recombination [35 , 36] . In this allele , amino acids 421 to 1 , 261 comprising the Tudor , MBD , pre-SET , SET-N , SET-C , and post-SET domains are deleted . The entire ORF was not removed because when the present study was designed , CG30422 was not considered part of the DmSetdb1 gene . The 5′-end of the DmSetdb1 gene is transcribed in the DmSetdb110 . 1a allele ( unpublished data ) , therefore the 420 first amino acids of DmSETDB1 are potentially translated , followed by 16 unrelated amino acids and a stop codon ( Figure 1A ) . This mutation is recessive lethal , in that homozygotes die at late pupal stage , with no escapers . The same phenotype is observed in individuals transheterozygous for DmSetdb110 . 1a and the chromosomal deficiency Df ( 2R ) ED4065 deleting the DmSetdb1 gene ( deleted segment: 60C8-60E7 ) . The polytene chromosomes of homozygote DmSetdb110 . 1a larvae appear normal ( Figure S1 ) . The DmSetdb110 . 1a homozygous mutant flies can be rescued into the adult stage by expression of DmSETDB1421–1 , 261 or 3HA-DmSETDB1421–1 , 261 transgenes ( UAS- DmSetdb1421–1 , 261 daGal4 heterozygotes ) . The rescued females are sterile , while the males are fertile , leading to the conclusion that DmSETDB1421–1 , 261 is partially functional . Collectively , phenotypic analysis of DmSetdb110 . 1a homozygotes and of transheterozygotes for DmSetdb110 . 1a and the chromosomal deficiency Df ( 2R ) ED4065 shows that DmSetdb1 is an essential gene , and that DmSetdb110 . 1a behaves as a null allele . We next investigated the biological function of endogenous DmSETDB1 . We first looked at the localization of the endogenous DmSETDB1 on polytene chromosomes . Therefore we generated the DmSetdb13HA allele by homologous recombination , which produces the endogenous DmSETDB1 protein tagged internally with a 3HA ( Figure 1A ) . DmSetdb13HA is expressed at a similar level compared with the wild-type allele ( Figure 1B ) . The transcript of DmSetdb13HA is slightly longer than that of the wild-type allele due to the 3HA sequence ( Figure 1B ) ; it was amplified by reverse transcriptase-PCR and sequenced , and shows no aberrant splicing ( unpublished data ) . DmSetdb13HA homozygous flies are viable and can be maintained as a stock , showing that the 3HA tag does not impair DmSETB1 function . Staining of homozygous DmSetdb13HA larvae polytene chromosomes with anti-HA shows a strong signal on Chromosome 4 ( Figure 2A ) . DmSETDB1 is also present over the whole length of the euchromatic arms , with some spots being more occupied . The chromocenter is weakly stained ( Figure 2A ) , a feature whose significance needs to be studied further , as DmSETDB1 is not methylating the chromocenter ( see below ) . As a negative control , polytene chromosomes of wild-type larvae stained with anti-HA show no signal ( unpublished data ) . Thus , endogenous DmSETDB1 localizes at Chromosome 4 and chromosome arms . By analogy to mammalian SETDB1 , which is a H3K9 mono- , di- , and tri-HMTase [27 , 28] , we asked whether DmSETDB1 is responsible for some of the H3K9 methyl marks present in chromatin . To address this , H3K9 mono- , di- , and trimethyl marks of wild-type and homozygous DmSetdb110 . 1a mutant larvae on polytene chromosomes were compared . Similarly as described in the literature , in wild-type conditions , the H3K9 monomethyl antibody stains the chromocenter and the euchromatic arms , although faintly . In the DmSetdb110 . 1a mutant , the monomethyl H3K9 signal is less intense on euchromatin , but does not completely disappear . However , the signal at the chromocenter remains unchanged ( Figure 2B ) . Therefore , DmSETDB1 is involved in some but not in all of the euchromatic H3K9 monomethylation , and displays no activity at the chromocenter . The H3K9 dimethyl antibody stains the chromocenter and Chromosome 4 in wild-type larvae , while the telomeres and the few euchromatic bands that were shown to bear H3K9 dimethyl marks [14 , 15] are not easily detectable . In the DmSetdb110 . 1a mutant background the mark is strongly reduced at the arm of Chromosome 4 , while the telomere and chromocenter are not affected ( Figure 2C ) . As a consequence , HP1 is present at the chromocenter and at the telomere , but it is not recruited to the Chromosome 4 arm , except for a few signals visualized as faint bands ( Figure 3B ) . Loss of HP1 at Chromosome 4 reinforces the conclusions made with the H3K9 dimethyl staining , namely that DmSETDB1 is the H3K9 dimethyl HMTase of the Chromosome 4 arm . We wanted to analyze the euchromatic and telomeric H3K9 dimethyl marks of the other chromosomes in DmSetdb1 mutant larvae , but the currently available antibodies do not allow detection of these marks . To circumvent this technical problem , stainings were performed with an antibody recognizing HP1 that produces significant signals . In the DmSetdb110 . 1a mutant background , HP1 is present on telomere , but disappears from some bands known to be strongly enriched in H3K9 dimethyl and HP1 , as for instance region 31 of Chromosome 2 [37] ( Figure 2D ) . Telomeres of the other chromosome arms are also bound by HP1 in the DmSetdb110 . 1a mutant background . Taken together , these results show that DmSETDB1 has an H3K9 dimethyl HMTase activity at some sites on the euchromatic arms , at Chromosome 4 , but not at telomeres . In terms of the H3K9 trimethyl modification present at the core of the chromocenter and few sites on the chromosome arms , we could not detect any difference between wild-type and DmSetdb110 . 1a homozygous mutant polytene chromosomes ( unpublished data ) . We also examined other methylation marks associated with repression , namely mono- and dimethyl H3K27 , and could not show any change in the DmSetdb110 . 1a mutant background ( unpublished data ) , arguing in favor of the specificity of DmSETDB1 for H3K9 . From these data we conclude that DmSETDB1 is responsible for some of the H3K9 mono- and dimethyl marks in euchromatin , and for most of Chromosome 4 H3K9 dimethylation . Others HMTases must be responsible for persistent H3K9 mono- and dimethylation in euchromatin , for H3K9 monomethylation at the chromocenter , and for H3K9 dimethylation at the telomeres . We next asked if overexpression of DmSETDB1 induces an increase of H3K9 methylation . This would confirm the ability of DmSETDB1 to mono- and dimethylate H3K9 and address if it can trimethylate H3K9 , as described for its mammalian homologue [27 , 28] . In addition , this would show whether DmSETDB1 is a limiting factor for the H3K9 methylation level or not . We overexpressed DmSETDB1421–1 , 261 , a less than full-length protein that nonetheless contains the Tudor , MBD , pre-SET , SET , and post-SET domains and can rescue the DmSetdb110 . 1a homozygotes ( see above ) . In addition , the 3HA-tagged version of DmSETDB1421–1 , 261 localizes similarly to the full-length protein , namely at Chromosome 4 and at euchromatin , although the signal is stronger at euchromatin most likely because of its higher expression ( Figure 3A ) . It is not possible to assess whether H3K9 methylation is present at the chromocenter , since it becomes disorganized upon DmSETDB1 overexpression ( Figure S1 ) . Thus , we consider that DmSETDB1421–126 is suitable to study the HMTase activity of DmSETDB1 . Increased expression of DmSETDB1421–1 , 261 is lethal , as ubiquitously overexpressing flies ( UAS-DmSetdb1421–1 , 261 daGal4 homozygotes ) die during the pupal stage , while heterozygous individuals survive and are fertile . Polytene chromosomes of larvae overexpressing DmSETDB1421–1 , 261 show an aberrant morphology . They appear thickened with unusual constrictions , and the chromocenter looks disorganized and decondensed ( Figure S1 ) . Such chromatin defects could be the cause of lethality . Upon DmSETDB1421–1 , 261 overexpression , there is a strong increase in H3K9 mono- , di- , and trimethylation on all chromosome arms , including Chromosome 4 ( Figure 3B ) . As a control , H3K27 mono- and dimethyl marks do not change when DmSETDB1 is overexpressed ( unpublished data ) . The same stainings were repeated under conditions where the DmSETDB1421–1 , 261 ( H1195K ) protein is overexpressed . The histidine 1 , 195 position is invariant among the SET proteins and is part of the cofactor AdoMet-binding pocket . The corresponding point mutation in human SETDB1 abolishes HMTase activity [27] . 3HA-DmSETDB1421–1 , 261 ( H1195K ) localizes similarly as 3HA-DmSETDB1421–1 , 261 ( Figure 3A ) , showing that the enzymatic activity is not required for chromatin localization of DmSETDB1421–1 , 261 . Overexpression of the mutant protein does not induce any increase or change in the H3K9 mono- , di- , or trimethylation patterns ( Figure 3B ) . HP1 recognizes H3K9 di- and trimethylated histones [9–11] and localizes at the chromocenter , the telomeres , Chromosome 4 , and at approximately 200 euchromatic sites of wild-type polytene chromosomes [24] . We wondered whether the profile of HP1 would be altered under DmSETDB1-overexpressing conditions . When DmSETDB1421–1 , 261 is overexpressed , HP1 is absent from the loose chromocenter , remains on Chromosome 4 , and is recruited to the euchromatic arms , more intensely at some sites ( Figure 2B ) . Western blot analysis shows that the total amount of HP1 is similar in DmSETDB1421–1 , 261 overexpressing and in wild-type larvae ( Figure 3C ) . These results indicate that HP1 is not expressed in larger amounts nor stabilized . Recruitment of HP1 to the chromosome arms does not occur upon overexpression of the DmSETDB1421–1 , 261 ( H1195K ) mutated protein ( Figure 2B ) , showing that DmSETDB1421–1 , 261 alone cannot recruit HP1 . Taken together , these results show that overexpressed DmSETDB1421–1 , 261 is located at and has an H3K9 mono- , di- , and tri-HMTase activity on the euchromatic arms and on Chromosome 4 , leading to the recruitment of HP1 . Global levels of H3K9 mono- , di- , and trimethylation were also measured by western blot analysis in tissue extracts from wild type , overexpressing DmSETDB1421–1 , 261 and DmSetdb110 . 1a homozygote mutant third instar larvae . Overexpression of DmSETDB1421–1 , 261 markedly increases mono- , di- , and trimethyl H3K9 levels , whereas absence of DmSETDB1 results in a modest decrease of these three modifications ( Figure 3C , first and second panels ) . The reduction observed in the DmSetdb110 . 1a mutant background is subtle but reproducible . Total H3 and HP1 levels ( unpublished data and Figure 3C , second panel ) are not influenced by the overexpression or the absence of DmSETDB1 . As expected , overexpression of the DmSETDB1421–1 , 261 ( H1195K ) mutant protein has no effect on H3K9 dimethylation ( Figure 3C , third panel ) or trimethylation ( unpublished data ) levels , except for a subtle increase in signal strength . Note that the increase in H3K9 dimethylation upon DmSETDB1421–1 , 261 overexpression is stronger in the third compared to the second panel , because the larvae are homozygous for the transgene . We conclude that DmSETDB1 is an H3K9 mono- , di- , and tri-HMTase and that increased expression positively influences the H3K9 methylation level . Given that DmSETDB1 strongly localizes to and methylates H3K9 on Chromosome 4 , we next assessed its role in gene regulation on that peculiar chromosome . To do this , we analyzed whether DmSETDB1 level would affect expression of white transgenes when placed on Chromosome 4 . Therefore , we used previously characterized lines where the white gene is expressed from P elements inserted in or at the edge of Chromosome 4 heterochromatic interspersed domains [21 , 22 , 38–40] . These lines display a variegated phenotype , indicating that the white gene is stochastically silenced . This pattern is reminiscent of heterochromatic position-effect variegation on other chromosomes , and mutations in HP1 or Su ( var ) 3–7 result in re-expression of the white gene , in all but one line ( 39C5 ) [21 , 38–40] . On the other hand , these variegating reporters do not respond to an additional or missing dose of SU ( VAR ) 3–9 ( mentioned in [18] as personal communication from K . Haynes [41] ) . If DmSETDB1 were implicated in repression via its HMTase activity , its absence would lead to reactivation of white expression . In parallel , four variegating lines were tested , two that have P elements inserted near centromeric heterochromatin of Chromosome 2 , one that has a P element inserted in the subtelomeric region of 2L , and the other being the In ( 1 ) wm4h line , in which an inversion relocates the endogenous white gene next to centromeric heterochromatin . White expression was analyzed in wild-type DmSetdb110 . 1a heterozygous and homozygous mutant late pupae . In the heterozygous mutant background , none of these lines differs from the wild type ( unpublished data ) . In the DmSetdb110 . 1a homozygous mutant context , however , the lines with transgene on Chromosome 4 show a robust expression of the white reporter ( Figure 4 , compare DmSetdb1 +/− and −/− in panels A–E ) . This is neither the case for the three transgenes on Chromosome 2 ( H , F , and I ) , nor for the white gene on the X Chromosome ( G ) . The expression in the In ( 1 ) wm4h line ( G ) is even reproducibly lower in the absence of DmSETDB1 , for as yet not understood reasons . These results show that DmSetdb1 is a recessive suppressor of variegation of Chromosome 4 . In the absence of DmSETDB1 , repression of transgenes located in the vicinity of Chromosome 4 heterochromatic domains is abolished .
Whereas Su ( var ) 3–9 and dG9a are not essential ( [42] , C . Seum , unpublished data ) , DmSetdb1 is the first gene described encoding a H3K9 methyltransferase that is required for fly viability . DmSetdb1 transcript can be detected at every stage of development . Our analysis by Northern blot confirms that the only transcript is 3 . 9 kb long , encompassing both CG30422 and CG30426 . Early embryos show relative high mRNA levels , suggesting deposition of the transcript in the embryo . Others conclude that DmSetdb1 transcript is not present in 0–3-h embryos when tested by reverse transcriptase-PCR [33] , a result that is not easily reconciled with our observations . DmSetdb110 . 1a homozygotes are rescued with the UAS- DmSetdb1421–1 , 261 daGal4 transgene; the males are fertile , while the females are sterile . Thus , the rescue is not complete in females , because of either nonappropriate expression of the transgene or because DmSETDB1421–1 , 261 is not full-length . This observation is consistent with the fact that DmSetdb1 ( eggless ) was shown to be required for oogenesis [32] . Preliminary data suggest that sterility in rescued females and in eggless mutant alleles [32] is due , at least in part , to defects in germline development . Indeed , using the FLP-ovoD1 system [43] , we could not generate any DmSetdb110 . 1a homozygous mutant germline clone ( unpublished data ) . This suggests that germline-specific expression of DmSetdb1 is required before stage 5 of oogenesis . This does not exclude , however , that a maternal contribution is required for proper oogenesis . The polyclonal antibody directed against a DmSETDB1 peptide we generated does not recognize DmSETDB1 on polytene chromosomes . Therefore , we generated the DmSetdb13HA allele that results into the expression of the endogenous DmSETDB1 protein tagged with 3HA ( Figure 1A ) . Such an approach has the advantage that the endogenously expressed protein can be detected with highly specific monocolonal antibodies . This allowed us to show that DmSETDB1 localizes at a high level on Chromosome 4 and over the chromosome arms ( Figure 2A ) . DmSETDB1 is also present at the chromocenter . We do not know if this feature has any biological significance as DmSETDB1 does not methylate the chromocenter . The association of DmSETDB1 with chromatin is not dependent on its own catalytic activity , since the DmSETDB1421–1 , 261 ( H1195K ) mutant protein localizes similarly to DmSETDB1421–1 , 261 ( Figure 3A ) . The mode of DmSETDB1 recruitment thus differs from that of SU ( VAR ) 3–9 , since the latter appears to require its HMTase activity for binding to heterochromatin [44] . It is currently not known how DmSETDB1 is recruited to chromatin . Mammalian SETDB1 is recruited to DNA together with HP1 , either via the KRAB-zinc-finger protein KAP1 corepressor [27 , 45] or by the ERG transcription factor [46] , or as a component of the MBD1-mAM/MCAF1-SETDB1 complex [30 , 47 , 48] . It is tempting to speculate that in Drosophila transcriptional repressors also recruit DmSETDB1 onto euchromatin or at Chromosome 4 . Comparative analysis of H3K9 methylation and HP1 profile on polytene chromosomes of wild-type and DmSetdb110 . 1a homozygous mutant larvae shows that DmSETDB1 is involved in some of the H3K9 mono- and dimethyl marks in euchromatin and in dimethyl marks on Chromosome 4 ( Figure 2B and 2C ) . Loss of methylation at Chromosome 4 and euchromatin is coherent with the localization profile of the DmSETDB1 protein itself . Western blot analysis of the H3K9 methylation level in mixed salivary glands , brain , and imaginal discs tissue in DmSetdb1 mutant background shows a decrease in all three H3K9 methyl marks ( Figure 3C ) . We could not evidence any change of trimethylation in polytene chromosomes of DmSetdb110 . 1a mutant larvae . This suggests a distinct H3K9 trimethylation profile in the tissues analyzed by Western blot and in polytene chromosomes . This hypothesis is corroborated by the recent finding that DmSETDB1 trimethylates H3K9 in germ and somatic cells of the germarium [32] . The overexpression data provide a mirror image , in that they show the ability of DmSETDB1 to mono- , di- , and trimethylate H3K9 ( Figure 3 ) . Thus , Drosophila DmSETDB1 and mammalian SETDB1 are conserved with respect to their HMTase activity , as both Drosophila DmSETDB1 and mammalian SETDB1 are H3K9 mono- , di- , and tri-HMTases [27 , 28] . Although such a mechanism has not yet been described , we cannot exclude that DmSETDB1 is exclusively a H3K9 monomethyltransferase providing monomethyl substrates for other enzymes; but in that case , the partner enzyme would not be SU ( VAR ) 3–9 , since its absence does not impair Chromosome 4 or euchromatic dimethylation . In mammals , conversion of the H3K9 dimethyl- to the trimethyl-state by SETDB1 is strongly facilitated by the mAM cofactor [28] . Such a mechanism can also be envisaged for DmSETDB1 , and CG12340 is a candidate Drosophila homologue of mAM . We could not detect any HMTase activity of DmSETDB1 in cell-free conditions . Immunopurified DmSETDB1 , regardless of whether expressed in mammalian or in Drosophila S2 embryo cell lines , did not show any activity when tested on GST-H3 , GST-H4 , core histones , or oligonucleosomes , while mammalian SETDB1 produced under identical conditions showed robust H3 specific activity ( unpublished data ) . We hypothesize that another protein or a post-translational modification is necessary for HMTase function of DmSETDB1 . This activity would not be present in S2 cell line; this is consistent with the fact that overexpression of DmSETDB1 in S2 cells does not induce any increase in H3K9 mono- , di- , or trimethylation ( unpublished data ) . DmSETDB1 functions in association with HP1; HP1 is recruited when DmSETDB1421–1 , 261 is overexpressed and lost from some euchromatic bands and Chromosome 4 in the DmSetdb110 . 1a mutant . In addition , HP1 is required for DmSETDB1-dependent repression of Chromosome 4 variegating transgenes [21 , 38–40] . We speculate that HP1 is recruited to chromatin by both the DmSETDB1 protein and the H3K9 methyl mark . Indeed , the DmSETDB1 protein is not able to recruit HP1 , because the DmSETDB1421–1 , 261 ( H1195K ) mutant protein does not influence HP1 localization . On the other hand , the H3K9 methyl mark alone is not sufficient to recruit HP1 [49] . Therefore , we hypothesize that HP1 recognizes the H3K9 methyl mark in association with DmSETDB1 , or with another factor . The situation is similar for Suv39H1 , where the protein itself does not recruit HP1 , despite a direct interaction that is necessary for HP1 binding in collaboration with the H3K9 methyl mark [49] . We do not know if a direct DmSETDB1-HP1 interaction occurs , but two arguments in mammals argue in favor of this . First , KAP1 directly binds HP1 [50] and SETDB1 [27] , and in such a complex , contacts between HP1 and SETDB1 are probable . Second , heterochromatin targeted HP1 recruits SETDB1 [51 , 52] , although an intermediate factor cannot be excluded . Although both DmSETDB1 and SU ( VAR ) 3–9 methylate H3K9 , one cannot substitute for the other . Indeed , in a mutant background for one enzyme , the other will not compensate for its absence . In addition , we can conclude that both enzymes function independently; SU ( VAR ) 3–9-mediated H3K9 di- and trimethylation and HP1 deposition at the chromocenter are not affected in the DmSetdb110 . 1a mutant context , and conversely , H3K9 mono- and dimethyl marks at euchromatic arms , dimethyl marks on Chromosome 4 , and the associated HP1 , are not affected in a Su ( var ) 3–9 mutant background [14 , 15] . Surprisingly , SU ( VAR ) 3–9 is present on Chromosome 4; it is most probably recruited by HP1 , but it does not induce any H3K9 methylation [14 , 15] . Thus , DmSETDB1 and SU ( VAR ) 3–9 exert nonoverlapping and independent functions , suggesting that they accomplish distinct biological roles . We anticipate that at least one additional HMTase is involved in H3K9 methylation in Drosophila . H3K9 monomethylation at the chromocenter , H3K9 dimethylation at the telomeres , and some of the H3K9 mono- and dimethylation marks at euchromatic bands are not deposited by SU ( VAR ) 3–9 nor DmSETDB1 . One candidate , dG9a , was recently shown to methylate H3K9 and to localize to euchromatin [16 , 17] . The repressive function of DmSETDB1 demonstrated for Chromosome 4 is consistent with the fact that H3K9 methylation is generally found in association with transcriptional silencing [53 , 54] . Indeed , the mammalian SETDB1 homologue fulfills such a function [27 , 45 , 47 , 48] . DmSETDB1 could also be implicated positively in gene expression , since H3K9 di- and trimethylation , as well as HP1γ were recently found in the coding region of active genes [55 , 56] . One task will be to identify endogenous genes that are regulated by DmSETDB1 in euchromatin and at Chromosome 4 . Genes located in the region 31 are potential candidates , given that the HP1 signal is lost in the DmSetdb110 . 1a mutant . The second set of candidate genes are those physically associated with HP1 but not with SU ( VAR ) 3–9 . Greil et al . [57] performed large-scale mapping of HP1 and SU ( VAR ) 3–9 targeted loci in embryonic Kc cells and showed that whereas HP1 and SU ( VAR ) 3–9 bind together to transposable elements and pericentric genes , HP1 binds to many genes on Chromosome 4 , mostly independently of SU ( VAR ) 3–9 . The latter , together with a class of euchromatic genes showing the same protein-factor occupation profile , possibly depend on DmSETDB1 for H3K9 methylation and regulation . DmSETDB1 is the H3K9 HMTase responsible for heterochromatin silencing on Chromosome 4 , because variegating transgenes are derepressed in a DmSetdb110 . 1a mutant background . As both alleles have to be mutated in order to obtain an effect , the DmSetdb1 gene is a recessive suppressor of variegation on Chromosome 4 . Conversely , loss of a single dose of HP1 or SU ( VAR ) 3–7 results in loss of silencing [21 , 38–40] . This difference could be explained by the fact that DmSETDB1 is an enzyme , whereas HP1 and SU ( VAR ) 3–7 are dosage-sensitive structural components . Alternatively , DmSETDB1 might be present in excess . Heterochromatic variegating reporters are responding to an additional or missing dose of SU ( VAR ) 3–9 when inserted on Chromosomes 2 , 3 , or X , but not on Chromosome 4 ( mentioned in [18] as personal communication from K . Haynes [41] ) . This observation is henceforth explained by the fact that DmSETDB1 mediates H3K9 dimethylation on Chromosome 4 . Conversely , and as expected , variegating expression responding to the SU ( VAR ) 3–9 dosage is not under the control of DmSETDB1 ( Figure 4F–4I ) . This corroborates once again that SU ( VAR ) 3–9 and DmSETDB1 function independently . Mammalian SETDB1 is involved in epigenetic maintenance , since silencing is stably maintained for more than 40 population doublings , once it is established on an integrated reporter by a short transient pulse of the corepressor KAP1 that subsequently recruits SETDB1 and HP1 [58] . DmSETDB1 could also be involved in epigenetic maintenance; in that case , transient expression would suffice for long-term repression of Chromosome 4 variegating transgenes . The arm of Chromosome 4 is composed of a minimum of three euchromatic domains interspersed with heterochromatic domains [21 , 38] . The variegating P elements that we tested were inserted within the banded region , in or at the edge of heterochromatic domains [38] . Chromosome 4 heterochromatic bands are qualitatively different from centromeric heterochromatin , as they are H3K9 dimethylated and regulated by DmSETDB1 , not by SU ( VAR ) 3–9 . Two possibilities can be envisaged for the Chromosome 4 domains that are methylated by DmSETDB1 . First , they could be representative of equivalent bands at euchromatic arms , which would be smaller and/or more dispersed , and therefore would not yet have been identified functionally . Alternatively , D . melanogaster Chromosome 4 could make use of specific machinery dedicated to gene regulation and/or epigenetic maintenance . The other well-known example of chromosome-specific regulation is the dosage compensation of sex chromosomes [59] . In that case , DmSETDB1 function would depend on partners or DNA sequences specific for Chromosome 4 , such as for instance the Chromosome 4-specific factor POF [60 , 61] , or the Hoppel element , also known as 1360 , which is over-represented on the D . melanogaster Chromosome 4 [62] , and which could be an initiation site for heterochromatin formation [21] . In conclusion we have characterized DmSETDB1 as a major nonheterochromatic H3K9 methyltransferase in Drosophila . We also demonstrated that DmSetdb1 is an essential gene and that its loss has functional consequences on gene expression on Chromosome 4 . This work represents an important step toward the understanding of the differential specificity and mode of action of distinct H3K9 HMTases and underlines a specific mode of regulation of Chromosome 4 in Drosophila .
The 39C12 , 39C72 , 118E10 , 118 E15 , 6M193 , 39C3 , and 39C5 lines contain the P[hsp26pt , hsp70-w] element and are gifts from Sarah Elgin [21 , 22 , 40] . Heidi was described in [63] . The stocks y w ( v ) ; P[ry+ , 70FLP]4 P[v+ , 70I-SceI]2B Sco/S[2] CyO and w1118; P[ry+ , 70FLP]10 were provided by Y . Rong and K . Golic . Description of other stocks can be found at FlyBase ( http://flybase . bio . indiana . edu ) . DmSetdb1421–1 , 261 ( CG30426 ) ORF was cloned by RT-PCR . 3HA- DmSetdb1421–1 , 261 carries in the N terminus a 3HA epitope derived from pBSKS-3HA [64] . DmSetdb1421–1 , 261 ( H1195K ) point mutation was generated by PCR . All constructs were verified by sequencing . 3HA- DmSetdb1421–1 , 261 was cloned into pCaSpeR . DmSetdb1421–1 , 261 , DmSetdb1421–1 , 261 ( H1195K ) , 3HA- DmSetdb1421–1 , 261 , and 3HA- DmSetdb1421–1 , 261 ( H1195K ) were cloned into pUASP vector [65] . Cloning details are available upon request . Constructs were injected into w1118 embryos with the pUChsπdelta2–3 plasmid at a 3:1 ratio . Transformant flies were selected with the white marker . DmSETDB1421–1 , 261 versions cloned in the pUASP vector and located on Chromosome 3 were recombined with the daGal4 driver located on Chromosome 3 . Homozygous DmSetdb110 . 1a larvae were selected from the stock w; Setdb110 . 1a /CyO GFP , where nonfluorescent homozygous mutant larvae were selected . DmSetdb110 . 1a were generated as follows . Cloning 4 . 1-kb genomic DNA located 5′ from CG30426 as well as 3 . 9-kb located 3′ from CG30426 , ( corresponding respectively to positions 95154–91021 and 88189–84215 [NCBI] ) , were amplified with high fidelity Taq DNA polymerase ( Roche , http://www . roche . com ) . PCR products were sequenced to ensure integrity of genes present in those regions . The 5′ amplified region was cloned into the NotI site of pW25 ( a gift from K . Golic ) , and the 3′ region was cloned into the AscI site . The procedure for the targeting screen was performed as described previously [35 , 36] . Briefly the targeting construct was injected into the w1 , 118 strain with the pUChsπdelta2–3 plasmid at a 3:1 ratio to obtain “donor” lines . A total of four independent donor lines on Chromosomes 3 or X were obtained . A total of 200 females of each donor line were crossed with yw; 70FLP , 70I-SceI , Sco/CyO males . We carried out two heat shocks on first- and second-instar larvae for one hour at 37 °C . From the progeny , 800 mosaic females carrying the 70FLP , 70 I-SceI chromosome were crossed with yw homozygous males expressing 70FLP constitutively . From the progeny , nonmosaic white positive flies were selected and further analyzed , to confirm that the wHs marker replaced the coding region of DmSetdb1 . The reduction step eliminates the wHs marker flanked by two loxP sites . The homologous recombinants were crossed to the yw;CyO P[w+ 70CreI/Sco] ( FlyBase ) line expressing the Cre recombinase . From the progeny , white negative flies were further characterized , and deletion of DmSetdb1 was confirmed by sequencing the region where homologous recombination occurred . DmSetdb13HA was generated using the following procedure . A 4 . 1-kb Xba/Not DNA fragment containing sequences 5′ of CG30426 ( positions 95154–91021[NCBI] ) with a I-SceI site inserted at position 93069 ( EagI ) , a 3 . 0-kb Xba/EagI fragment containing CG30426 ( positions 91020–88310 ) , and a 3HA tag in at position 91020 , were cloned into the pTV2 ( NotI ) vector [66] . In this clone , the ORF is conserved from CG30422 to CG30426 , and the I-CreI site faces position 88312 . All PCR products were sequenced . The targeting screen procedure is similar to the DmSetdb110 . 1a allele . The reduction step involves a recombination that replaces the endogenous CG30426 with the 3HA-tagged CG30426 and deletes the wHs marker . Females recombinant/SM5 were crossed with males CyO/+;70-I Cre 1A/ TM3 . Heat shocks were made on first instar larvae 30 min at 37 °C , and variegated males were balanced with w1 , 118; CyO;TM3/ T2-3ApXa females . w−/CyO flies were crossed with each other . Homozygote-reduced recombinant flies were analyzed by PCR . The region where the 3HA is inserted was sequenced . Brains , salivary glands , and imaginal discs from third instar larvae were dissected in PBS , resuspended in 50 mM Tris ( pH 7 . 8 ) , 150 mM NaCl , 5 mM EDTA , 1% SDS , 1 mM PMSF , and protease inhibitors ( Complete , Roche ) , boiled 10 min , and cleared by centrifugation . We separated 20-μg or 5-μg extract on 15% SDS-PAGE , and proteins were transferred on PVDF membrane ( Millipore , http://www . millipore . com ) by semi-dry transfer . Membranes were blocked in TBS , 0 . 1% tween , 5% non fat milk , hybridized in TBS , 0 . 1% tween , 1% non fat milk , with α-H3K9me1 ( 1/1000 ) ( a gift from T . Jenuwein ) , α-H3K9me2 ( 1/1000 ) ( a gift from T . Jenuwein ) , α-H3K9me3 ( 1/1000 ) ( Upstate Biotechnology 07–523 , http://www . upstate . com ) , α-H3 ( 1/5000 ) ( Abcam 1791 , http://www . abcam . com ) , α-HP1 ( 1/4000 ) ( a gift from L . Wallrath ) , or α-α-tubulin ( 1/5000 ) ( Sigma T 9026 , http://www . sigmaaldrich . com ) . Membranes were washed with TBS , 0 . 1% tween , hybridized with HRP-coupled secondary antibody , washed , and revealed by chemoluminescence . Where indicated , the membranes were stripped and reprobed . Total RNA from 0–4-h and 0–18-h embryos , third instar larvae , and male and female adults was extracted using Trizol reagent ( Invitrogen , http://www . invitrogen . com ) . We separated 20 μg RNA from each sample on 1% agarose-formaldehyde gel , transferred to Hybond-N+ membrane ( Amersham , http://www . amersham . com ) , and UV crosslinked . Membrane was hybridized in Rapid-Hyb buffer ( Amersham ) with a probe covering nucleotides 2 , 399–3 , 789 of DmSetdb1 ORF and subsequently with an RNA-loading control probe recognizing rp49 . Probes were radioactively 32P labeled using the Redi-prime labeling kit ( Amersham ) as described by the manufacturer . Polytene chromosomes were performed as described previously [67] . Briefly , salivary glands were dissected in Cohen's buffer , fixed for 2 min in 2% formaldehyde , 2% Triton X-100 , and then squashed in 2% formaldehyde , 45% acetic acid . Slides were hybridized with the following primary antibodies: α-HA ( 1/200 ) ( Covance MMS-101R , http://www . covance . com ) , α-HP1 ( 1/400 ) ( a gift from L . Wallrath ) , α-H3K9me1 ( 1/200 ) ( a gift from T . Jenuwein ) , α-H3K9me2 ( 1/100 ) ( Upstate Biotechnology 07–441 ) , and α-H3K9me3 ( 1/200 ) ( Upstate Biotechnology 07–523 ) . Picture of flies at pupal stage with the following genotype were taken: Figure 4A: w/w; DmSetdb110 . 1a/+; 39C12/+ and w/w; DmSetdb110 . 1a/ DmSetdb110 . 1a; 39C12/+ . Figure 4B: w/w; DmSetdb110 . 1a/+; 39C72/+ and w/w; DmSetdb110 . 1a/DmSetdb110 . 1a; 39C72/+ . Figure 4C: w/w; DmSetdb110 . 1a/+; 118E10/+ and w/w; DmSetdb110 . 1a/DmSetdb110 . 1a; 118E10/+ . Figure 4D: w/w; DmSetdb110 . 1a/+; 6M193/+ and w/w; DmSetdb110 . 1a/DmSetdb110 . 1a; 6M193/+ . Figure 4E: w/w; DmSetdb110 . 1a/+; 118E15/+ and w/w; DmSetdb110 . 1a/DmSetdb110 . 1a; 118E15/+ . Figure 4F: w/w; DmSetdb110 . 1a/39C3 and w/w; DmSetdb110 . 1a 39C3/ DmSetdb110 . 1a . Figure 4G: In ( 1 ) wm4h /w; DmSetdb110 . 1a/+ and In ( 1 ) wm4h/w; DmSetdb110 . 1a/ DmSetdb110 . 1a . Figure 4H: w/w; DmSetdb110 . 1a/Heidi and w/w; DmSetdb110 . 1a Heidi/DmSetdb110 . 1a . Figure 4I: w/w; DmSetdb110 . 1a/39C5 and w/w; DmSetdb110 . 1a 39C5/ DmSetdb110 . 1a . Lines 39C12 , 39C72 , 118E10 , 118E15 , and 6M193 [22] are on Chromosome 4 , and lines 39C3 , 39C5 [22] , and Heidi [63] are on Chromosome 2 .
The Flybase ( http://www . flybase . org ) accession numbers for the Drosophila DmSetdb1 gene are CG30422 and CG30426 . The National Center for Biotechnology Information ( NCBI ) ( http://www . ncbi . nlm . nih . gov ) accession number for DmSetdb1 full insert cDNA is BT023947 . The NCBI accession number for 4 . 1-kb genomic DNA located 5′ from CG30426 as well as 3 . 9-kb located 3′ from CG30426 ( positions 95154–91021 and 88189–84215 , respectively ) is AE003465 .
|
DNA is the basic unit carrying genetic information . Within the nucleus , DNA is wrapped around an eight-histone complex to form the nucleosome . The nucleosomes and other associated proteins assemble to a higher order structure called chromatin . The histones are mainly globular , excepted for their tails that protrude from the nucleosome core . The amino acids of the histone tails are often modified . For example , several conserved lysine residues can be methylated . Methylation of lysine 9 on histone H3 ( H3K9 ) is important for proper chromatin structure and gene regulation . Here , we characterize Drosophila DmSETDB1 as a histone methyltransferase responsible for H3K9 methylation of the chromosome arms and Chromosome 4 . In addition , we show that in the absence of DmSETDB1 , silencing of Chromosome 4 is abolished . This study is an important step towards the understanding of the differential chromatin domain specificity and mode of action of H3K9 methyltransferases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"drosophila",
"molecular",
"biology",
"genetics",
"and",
"genomics"
] |
2007
|
Drosophila SETDB1 Is Required for Chromosome 4 Silencing
|
The environmental bacterium Burkholderia pseudomallei causes the infectious disease melioidosis with a high case-fatality rate in tropical and subtropical regions . Direct pathogen detection can be difficult , and therefore an indirect serological test which might aid early diagnosis is desirable . However , current tests for antibodies against B . pseudomallei , including the reference indirect haemagglutination assay ( IHA ) , lack sensitivity , specificity and standardization . Consequently , serological tests currently do not play a role in the diagnosis of melioidosis in endemic areas . Recently , a number of promising diagnostic antigens have been identified , but a standardized , easy-to-perform clinical laboratory test for sensitive multiplex detection of antibodies against B . pseudomallei is still lacking . In this study , we developed and validated a protein microarray which can be used in a standard 96-well format . Our array contains 20 recombinant and purified B . pseudomallei proteins , previously identified as serodiagnostic candidates in melioidosis . In total , we analyzed 196 sera and plasmas from melioidosis patients from northeast Thailand and 210 negative controls from melioidosis-endemic and non-endemic regions . Our protein array clearly discriminated between sera from melioidosis patients and controls with a specificity of 97% . Importantly , the array showed a higher sensitivity than did the IHA in melioidosis patients upon admission ( cut-off IHA titer ≥1:160: IHA 57 . 3% , protein array: 86 . 7%; p = 0 . 0001 ) . Testing of sera from single patients at 0 , 12 and 52 weeks post-admission revealed that protein antigens induce either a short- or long-term antibody response . Our protein array provides a standardized , rapid , easy-to-perform test for the detection of B . pseudomallei-specific antibody patterns . Thus , this system has the potential to improve the serodiagnosis of melioidosis in clinical settings . Moreover , our high-throughput assay might be useful for the detection of anti-B . pseudomallei antibodies in epidemiological studies . Further studies are needed to elucidate the clinical and diagnostic significance of the different antibody kinetics observed during melioidosis .
Melioidosis is an often fatal tropical infectious disease caused by the Gram-negative environmental bacterium Burkholderia pseudomallei [1 , 2] . The disease is known to be highly endemic in Southeast Asia and northern Australia . However , an increasing number of melioidosis case reports or environmental isolation of B . pseudomallei from other parts of Asia , Africa , the Caribbean , and Central and South America suggest a worldwide , but grossly underreported distribution of B . pseudomallei between latitudes 20° N and 20° S [3–9] . Recently , Limmathurotsakul and coworkers predicted about 165 , 000 cases of human melioidosis per year worldwide , from which 89 , 000 people die [10] . Farmers and indigenous inhabitants of rural tropical areas are population groups at greatest risk of infection , especially in times of heavy rains [1 , 2 , 5] . Melioidosis usually has an incubation period of 1 to 21 days ( mean: 9 days ) and causes a wide range of acute or chronic clinical manifestations , including pneumonia , abscesses in various organs , neurological manifestations , or severe septicemia [1 , 2 , 11–13] . Since B . pseudomallei is intrinsically resistant to many antibiotics , it requires an immediate diagnosis followed by specific and prolonged antibiotic therapy . Melioidosis has a case fatality rate of around 40% in northeast Thailand [14] . In acute forms , death can occur within 24–48 hours of the onset of symptoms [15 , 16] . The rapid diagnosis of melioidosis is still a major obstacle in many potentially endemic parts of the world . Cultural identification of B . pseudomallei can be difficult , especially in non-endemic areas where clinical suspicion and awareness in the laboratory is low [1 , 13 , 17] . Even in endemic areas , the culture method has a low sensitivity and might take several days until results are available [18] . In addition , laboratory facilities for microbiological culture are unavailable in many countries of the world where melioidosis is endemic or suspected to be present . Serodiagnostic methods might have the potential to complement direct pathogen detection . The indirect hemagglutination assay ( IHA ) is the known standard serology test for melioidosis [1 , 13 , 19 , 20] . This assay , based on sheep red blood cells sensitized with crude B . pseudomallei antigen is simple to perform and inexpensive . However , the diagnostic sensitivity of this approach upon admission is about 56% and a high seropositive background in endemic areas reduces the specificity [21 , 22] . The crude preparations are difficult to standardize , and different strains have been used for antigen preparations in different laboratories . Protein microarrays are an effective approach to perform large scale serological studies and enable a fast , parallel analysis of a multitude of possible antigens [23 , 24] . They can be produced and probed in a high-throughput manner and are hence highly standardized [23] . In a previous study , Felgner et al . ( 2009 ) identified 49 B . pseudomallei proteinogenic antigens that were significantly more reactive in melioidosis patients than in controls [25] . Based on a selection of 20 of those antigens , we constructed a protein microarray using a robust , commercially available technology that can be used for high-throughput testing in a clinical laboratory [23 , 26 , 27] . The results of probing 196 melioidosis positive and 210 negative control samples from endemic and non-endemic areas as well as samples from patients with other bacteremia or fungemia demonstrated a high sensitivity and specificity . Moreover , for the first time to the authors’ knowledge , the multiplex detection of short- and long-term antibodies against various protein antigens in melioidosis patients is described .
This retrospective study was approved by the ethics committees of Faculty of Tropical Medicine , Mahidol University ( Submission number TMEC 12–014 ) ; of Sappasithiprasong Hospital , Ubon Ratchathani ( reference 018/2555 ) ; and the Oxford Tropical Research Ethics Committee ( reference 64–11 ) . The study was conducted according to the principles of the Declaration of Helsinki ( 2008 ) and the International Conference on Harmonization ( ICH ) Good Clinical Practice ( GCP ) guidelines . Written informed consent was obtained for all patients enrolled in the study . The bacterial strains and plasmids used are listed in S1 Table . Escherichia coli strains DH5α and expression strain BL21DE3pLysS as well as the B . pseudomallei strain K96243 were cultured in Luria-Bertani ( LB ) medium or LB agar at 37°C . Unless stated otherwise , the concentrations of antibiotics added to LB medium for E . coli were as follows: ampicillin ( Ap , Sigma-Aldrich , Germany ) , 100 μg/ml and/or chloramphenicol ( Cm , Sigma-Aldrich , Germany ) , 25 μg/ml . Twenty B . pseudomallei antigens with serodiagnostic potential were chosen as targets from studies by Felgner et al . ( 2009 ) and Suwannasaen et al . ( 2011 ) , and are listed in Table 1 . Proteins were selected based on their diagnostic sensitivity and specificity as determined by Felgner et al . ( 2009 ) , their genomic location ( Chromosome 1 or 2 ) , their bacterial location ( cytoplasm , extracellular , periplasm , membrane/outer membrane ) , their predicted function ( protein folding and stabilization , metabolism , virulence , unknown function etc . ) and first of all their solubility in phosphate buffered saline after the freezing and storage process . All protein antigens were analyzed by PSORTb version 3 . 0 . 2 ( http://www . psort . org/psortb/ ) , and any signal sequences or transmembrane domains were excluded for further cloning . The respective protein encoding DNA fragments were amplified by PCR using specific oligonucleotides ( S2 Table ) and genomic DNA from B . pseudomallei K96243 strain as the template . Oligonucleotides were created using primer design software Primer`D`Signer 1 . 1 ( IBA GmbH , Göttingen , Germany ) . PCR products were digested and cloned using appropriate restriction enzymes and protein expression plasmids ( S2 Table ) . The correctness of all cloned genes was confirmed by DNA sequencing . For protein expression , plasmids were transformed in E . coli expression strain BL21 ( DE3 ) pLysS by heat shock and were grown in LB medium with permanent agitation at 37°C to an optical density ( OD540nm ) of 0 . 5 . Protein expression was induced by adding isopropyl-β-D-thiogalactopyranoside ( IPTG , 1 mM final concentration ) ( Carl Roth GmbH , Germany ) , and after 3 hours , cells were harvested by centrifugation at 8000 x g and 4°C for 10 minutes . Afterwards , cells were disrupted by six cycles ( 3 min at 4°C ) of ultrasonic homogenizer UP50H ( Hielscher Ultrasonics GmbH , Germany ) , and the lysates were centrifuged at 4°C and 12000 x g for 30 minutes . Supernatants were stored at -20°C until use . The protein purification of Strep-tag or His-tag recombinant proteins was performed by using Gravity flow Strep Tactin-Sepharose Columns ( IBA GmbH , Göttingen , Germany ) or Ni-NTH Agarose ( Qiagen , Germany ) according to the manufacturers’ instructions . Afterwards , purified proteins were dialyzed against Dulbecco’s Phosphate Buffered Saline ( DPBS ) ( Gibco-life technologies , USA ) , and their purity was confirmed by SDS page ( S1 Fig ) . Recombinant proteins were stored at -20°C until use for protein array construction . Patients included in the study formed a consecutive series . Sera and plasma from culture-confirmed melioidosis patients were collected from September 2012 to November 2014 in the highly endemic area of Ubon Ratchathani , Thailand . ( Table 2 ) as described previously [29] . Negative control sera ( n = 100 ) consisted of 50 sera of healthy individuals from Ubon Ratchathani ( endemic ) , 25 sera from healthy individuals with diabetes from the same region , and 25 sera from healthy individuals in Bangkok . Further negative controls were drawn from healthy individuals and patients with other bacteremia or fungemia in the non-endemic area of Greifswald ( Germany ) . Sera from melioidosis patients were taken within the first week ( ´week 0´ , n = 75 ) post-admission ( p . a . ) , 12 weeks p . a . ( ´week 12´ , n = 50 ) and 52 weeks p . a . ( ´week 52´ , n = 46 ) . Endemic samples ( week 0 , 12 and 52 ) were considered melioidosis positive if B . pseudomallei was isolated from blood , pus , or any other body fluid . The majority of patients were male ( week 0 , 12 or 52: 68% , 72% and 71 . 7% , respectively ) with a median age of 55 years . IHA titers were performed on all sera drawn in Thailand as described previously [30 , 31] ( Table 2 ) . A serum was classified as positive if the cut-off for the IHA titer was equal to or higher than 160 . This cut off has been widely used in studies in Thailand [32 , 33] , although lower cut offs were used in other endemic regions [22 , 34] , possibly to methodological variations and/or less background seropositivity . The IHA titers of the analyzed plasma samples were not determined . The melioidosis negative sera ( n = 85 ) drawn in the non-endemic region of Germany consisted of sera from patients with other bacteremia or fungemia ( n = 60 ) and healthy blood donors ( n = 25 ) ( Table 2 ) . IHA titers of these sera were also not determined . All sera or plasmas were stored at -80°C . All purified proteins were spotted on a 4 . 2 x 4 . 2-mm glass microarray surface with a spotted area of 3 . 6 x 3 . 6 mm and incorporated in the ArrayStrip system provided by Alere Technologies GmbH ( Germany ) , resulting in the first-generation B . pseudom . 01-Array . Recombinant proteins were covalently immobilized as triplicates at five different concentrations ( 0 . 01 to 0 . 45 mg/ml ) ; subsequently bovine serum albumin was immobilized to a concentration of 0 . 5 mg/ml on the array . Horseradish peroxidase ( HRP ) and purified IgG and IgM antibodies from different species ( humans , mice , pigs , sheep , goats and cattle ) served as positive controls , and spotted bovine serum albumin ( BSA ) functioned as the negative control . After manufacturing , each single ArrayStrip was sealed under a noble gas ( argon ) atmosphere into nontransparent bags and stored at 4°C until use . Antibody detection using the B . pseudom . 01-Array was performed according to a previously optimized manufacturer’s protocol . Briefly , protein arrays were first incubated with washing buffer ( 1xPBS/0 . 05% Tween 20/0 . 25% TritonX100 ) at 37°C and 400 rpm for 5 minutes . Afterwards , protein arrays were incubated with blocking buffer ( 1xPBS/0 . 05% Tween 20/0 . 25% TritonX100 and 2% Blocking Reagent ( No 11 096 176 001; Roche , Switzerland ) ) at 37°C and 300 rpm for 5 min in order to block unspecific binding sides . Subsequently , diluted sera and plasmas ( 10−3 ) were incubated for 30 min at 37°C and 300 rpm . After a washing step as described above ( 37°C , 400 rpm , 5 min ) , the protein arrays were incubated with a diluted ( 10−3 ) HRP coupled anti-human IgG antibody ( Sigma-Aldrich , USA ) at 37°C and 300 rpm min for 30 min . To avoid strong background signals , protein arrays were washed again twice with washing buffer ( 37°C , 400 rpm , 4 min ) and finally incubated with the specific substrate D1 ( Alere Technologies GmbH , Jena , Germany ) for exactly 10 min without shaking at 25°C . Finally , the protein arrays were read out by the ArrayMate and data were analyzed using IconoClust software according to the manufacturer’s specifications ( both by Alere Technologies GmbH , Germany ) . The following parameters for evaluating the arrays were used: The normalized intensities ( NI ) of the spots were determined as NI = 1- ( M/BG ) , where M is the average intensity of the spot and BG is the intensity of the local background . Hence , results range between 0 ( no signal ) and 1 ( maximal intensity ) . Spot intensities of at least 0 . 3 were defined as a specific antibody response to the respective antigens . The recognition of at least two different antigens per serum or plasma with signal intensities above 0 . 3 was considered melioidosis positive . Sensitivities and specificities of the IHA and the protein array were calculated using following equations: sensitivity = ∑ true melioidosis positive tested individuals / ∑ total melioidosis positive individuals; specificity = ∑ true melioidosis negative tested individuals / ∑ total melioidosis negative individuals . Readers were blind to clinical outcome and to results of other tests at the time of reading . The two-sided Fisher's exact test was used to show whether the proportions of positive and negative signals differ between individual groups , i . e , melioidosis positive and negative samples . [35] . Fisher's exact test was carried out for the signals of each spotted substance in the microarray , using R as the language for statistical computing ( R Core Team , 2015 . R: A language and environment for statistical computing . R Foundation for Statistical Computing , Vienna , Austria . URL http://www . R-project . org/ ) . P < 0 . 01 was considered statistically significant . In this study , the software programs GraphPadPrism 5 . 0 ( GraphPad software , Inc . , USA ) , Excel 2010 ( Microsoft Corporation , USA ) and Multi experiment Viewer 4 . 9 . 0 ( TM4 suite , USA ) were used for visualization of the data .
In this study , a protein microarray was developed containing 20 B . pseudomallei proteins , previously identified by Felgner et al . ( 2009 ) to have serodiagnostic potential in melioidosis [25] . All proteins were expressed in E . coli , purified ( S1 Fig ) , and spotted at five increasing concentrations ( 0 . 01 to 0 . 45 mg/ml ) on the glass microarray surface ( Fig 1 ) . Among those antigens are cytoplasmic proteins ( n = 9 ) , extracellular proteins ( n = 9 ) , outer membrane/membrane proteins ( n = 2 ) , and periplasmic proteins ( n = 1 ) ( Table 1 ) . The antigens are predicted to be involved in protein folding and stabilization , cell motility , detoxification , virulence , and transport , and may have yet unknown functions . In contrast to the protein microarray platform used by Felgner et al . , recombinant proteins were exempted from signal sequences or transmembrane domains to maintain them in a soluble state . The whole protocol starting from sera or plasma incubation to final data analysis takes about two hours ( S2 Fig ) [23 , 26 , 27] . The protocol uses pure and highly standardized chemicals and antibodies that are available worldwide ( see Material and Methods ) . Our protein array was validated using different groups of sera of melioidosis patients ( n = 171 ) or negative control individuals ( n = 185 ) ( Table 2 ) . In total , three melioidosis positive groups ( week 0 , 12 and 52 p . a . ) and two melioidosis negative groups ( healthy individuals from Thailand/Germany and patients with other types of bacteremia/fungaemia from Germany ) were used for a parallel and comprehensive analysis of human IgG reactivity . The positive sera of patients upon admission ( week 0 ) were clearly distinguishable from all negative control sera ( endemic and non-endemic regions ) and from sera of patients with other types of bacteremia or fungaemia ( Fig 2 ) . Furthermore , all melioidosis-positive sera taken at weeks 12 and 52 p . a . were also found to be highly distinguishable from all negative control groups ( S3 Fig ) . We observed strong signal intensities even for the low antigen concentrations ( 0 . 01 or 0 . 05 mg/ml antigen ) , and most antigens showed signal intensities greater than 0 . 3 at a concentration of 0 . 45 mg/ml , with a median of 4 recognized antigens at this antigen concentration ( S4 Fig ) . Therefore , we used this concentration for all further analyses . In total , 17 of 20 antigens showed signal intensities above 0 . 3 . The strongest average signal intensities were found for antigens BPSL2697 and BPSL2096 , followed by BPSS0477 , BPSL2522 , BPSL2698 , BPSS0476 and BPSS1532 ( Fig 3 ) . Lower signal intensities were found for BPSL3319 , BPSS1722 , BPSL2030 , BPSS1525 , BPSL2520 , BPSS1516 , BPSL0280 , BPSS2141 , BPSL1445 and BPSS0530 ( Fig 3 ) . Importantly , no signals could be measured for antigens BPSS1385 , BPSL1661-1001 and -1002 , although these proteins have been previously described as serodiagnostic marker proteins [25] . No influence could be observed for the nature of protein tags ( His- or Strep-tag ) for BPSL2697 , BPSL2096 , BPSS0477 , BPSS1532 ( S2 File ) . Hence , the results for antigens purified with His-tag are not further discussed . In order to elucidate the discriminatory power of the single antigens further , all different groups of melioidosis-positive sera ( S5 Fig ) were compared with the various controls using the two-sided Fisher's exact test as described by Glantz [35] . Comparisons of sera from healthy donors from both endemic and non-endemic regions with that of melioidosis-positive sera of week 0 p . a . showed thirteen significantly recognized antigens , twelve antigens from sera of week 12 p . a . and six antigens from week 52 p . a . ( comparisons 01 , 02 and 03 shown in S1 File ) . Comparisons using sera of patients with other bacteremias or fungaemias revealed eleven significantly recognized antigens from melioidosis-positive sera of week 0 and 12 and four antigens from sera of week 52 ( comparisons 05 , 06 and 07 shown in S1 File ) . Testing a higher number of sera will likely increase the number of significantly recognized B . pseudomallei antigens . Thirteen of the serodiagnostic marker proteins found by Felgner et al . ( 2009 ) were confirmed here . Since plasma is routinely drawn in clinical practice , we additionally examined blood plasmas ( week 0 p . a . ) of melioidosis-positive ( n = 25 ) and -negative ( n = 25 ) individuals . As shown for sera , melioidosis-positive plasmas were also highly distinguishable from negative control plasmas ( S6 Fig and S7 Fig ) . Compared to the respective sera , almost identical numbers of antigens per plasma were recognized ( S8 Fig ) . In addition , no significant differences in signal intensities per antigen could be observed when sera and plasma samples were used from the same patient ( S9 Fig ) . In one melioidosis plasma sample , we found positive signals for BPSL1661-1002 , which were not observed for any positive serum sample . Unfortunately , the corresponding serum sample was not available . However , as shown for sera , a total of 17 of 20 B . pseudomallei antigens were recognized by at least one melioidosis-positive plasma sample . Our results indicate that in addition to blood sera , also blood plasmas can be used to detect antibodies against B . pseudomallei in our protein microarray system . Depending on many different parameters , antigens can elicit antibody responses of different durations . Here , we used melioidosis-positive sera drawn at weeks 0 , 12 and 52 p . a . from individual patients ( n = 36 ) to investigate the antibody responses to the different protein antigens over a prolonged period of time . In general , signal intensities of almost all antigens and the number of antigens detected declined over time ( Figs 3 and 4 ) . Two groups of differentially recognized antigens could be described . Antigens of the first group ( BPSL2030 , BPSL2096 , BPSL2522 , BPSL2697 , BPSL2698 , BPSS0476 and BPSS0477 ) induced a relatively strong , constant antibody response over a prolonged period of time . Even 52 weeks after patient admission , an antibody response against these antigens could be detected in at least 50% of sera ( Fig 5A ) . Recognition of those antigens at weeks 12 and 52 p . a . was observed in sera which were positive for these antibodies at week 0 p . a . but also in sera which were negative for those antibodies at week 0 p . a . ( Fig 5A ) . In contrast , antigens of group 2 ( BPSS1532 , BPSL3319 , BPSS1722 , BPSL2030 , BPSS1525 , BPSL2520 ) did not show significant recognition in sera from 52 weeks p . a . ( Fig 5B ) . Interestingly , three antigens of group 1 ( BPSL2030 , BPSL2096 and BPSS0476 ) showed the same or a higher number of significant signals ( signal intensity ≥ 0 . 3 ) if incubated with sera of week 12 p . a . compared to sera of week 0 ( Fig 5A ) . The same was observed for four antigens ( BPSL2520 , BPSS1525 , BPSS1532 and BPSS1722 ) of group two antigens ( Fig 5B ) . Among group two members , particularly the antigen BPSL0280 induced only a very short antibody response . After only 12 weeks p . a . , the average signal intensity and number of significant signals was similar to the signals observed for sera obtained at 52 weeks p . a . ( Fig 5B ) . Importantly , the categorization of antigens into the two groups was confirmed for the complete set of melioidosis sera , including patients where only sera from single time points were available ( Table 2 and S3 File ) . By using a multiplex detection approach , we revealed for the first time that different B . pseudomallei protein antigens induce long- and short-term antibody responses . Thus , the identified groups of antigens might have the potential to distinguish between more recent B . pseudomallei infections and infections which occurred further in the past . We further compared sensitivity and specificity of IHA titer values with those results obtained from protein array experiments . In general , mean and median IHA titer measured in sera from patients at admission week 0 p . a . were higher compared to IHA titer in sera of week 52 p . a . and negative control sera from Thailand , whereas sera of week 12 p . a . showed the highest IHA titer measured ( Table 2 and Fig 6A ) . Average signal intensities obtained from protein array experiments showed the same tendencies , but only sera of week 0 p . a . correlated with the IHA titer values ( week 0: rsp = 0 . 3470 , p = 0 . 023; week 12: rsp = 0 . 2743 , p = 0 . 054; week 52: rsp = 0 . 1602 , p = 0 . 2877; healthy: rsp = 0 . 1107 , p = 0 . 2728 ) ( Fig 6B ) . However , from 75 tested sera of patients upon admission ( week 0 p . a . ) , 32 sera had an IHA titer lower than 160 and were classified as melioidosis negative . When these 75 sera were analyzed using the protein arrays , only 10 sera gave less than two significant signals to two different antigens and had to be classified as melioidosis negative . Six sera were solely classified as melioidosis negative by the protein array , as opposed to 28 sera classified as negative by IHA . Four sera were classified as melioidosis negative by both methods ( Fig 7 ) . Thus , the sensitivity of the protein array ( 86 . 7% ) was clearly higher than that of the IHA ( 57 . 3% ) , as shown in Table 3 . No significant differences in sensitivities could be observed using sera drawn at 12 or 52 weeks p . a . ( Table 3 ) . The specificities of the IHA test and the protein array were barely distinguishable , with 96% and 97% , respectively .
The alarmingly large number of predicted melioidosis cases worldwide with a high mortality [10] emphasizes the need to improve the current diagnostic tools to detect B . pseudomallei infections . The methods for the analysis of immune responses during infection have recently been expanded by including protein microarrays that target pathogen-specific antigens [36–38] . Protein microarrays have the potential advantage of overcoming the limitations of a more or less monoplex antibody detection when single antigens are used [20 , 39] . The application of single antigens and thus the restriction to certain epitopes might limit the sensitivity and specificity of diagnostic serological assays . Recently , a protein array approach was used as an antigen discovery platform , and a significant number of serodiagnostic marker proteins of B . pseudomallei were identified that were more reactive in melioidosis patients compared to controls [25 , 40] . Based on the work by Felgner et al . [25] , we selected 20 antigens to develop a B . pseudomallei protein microarray using a miniaturized technical platform , which can be automated and is applicable in a routine setting . The validation of our microarray with sera taken from patients at defined time points after admission demonstrates a significantly higher sensitivity of our protein array to detect melioidosis upon admission compared to the standard IHA ( 86 . 5% vs 57 . 3% , respectively ) . In our study , 13 of the 20 proteins identified by Felgner et al . were confirmed as specific serodiagnostic markers . Seven proteins were not statistically significantly recognized by melioidosis-positive sera compared to control sera . Two proteins were not recognized by any serum or plasma tested . This discrepancy to the results found by Felgner et al . ( 2009 ) may be explained by the following factors: i . All proteins used in this study were free of any transmembrane domains and/or any signal sequences . ii . In contrast to the in vitro translation system used by Felgner et al . , proteins in our study were expressed in E . coli and purified . iii . Our miniaturized array system used very different protocols for sera incubation , washing steps , and the detection system . In summary , all these factors might have led to the discrepant recognition of diagnostic proteins . Based on ten of their identified serodiagnostic proteins , Felgner et al . developed an immunostrip assay , reporting a sensitivity of 95% and a specificity of 83% , which represented a major advantage over current standard diagnostic tests [25] . Although the sensitivity of 86 . 5% with our miniaturized protein array seems to be slightly lower , we observed a higher specificity of 97% compared to the immunostrip assay . The results of our protein array are promising , since our pool of negative control sera also contained 60 patients’ sera with proven positive blood cultures for other bacterial pathogens and fungi . The high specificity ( 96% ) of the IHA test in our healthy Thai control cohort is surprising in the context of the previous literature . It seems possible that this cohort from Ubon Ratchathani experienced a lower exposure to Burkholderia spp . than previously published cohorts . Most signals and the highest signal intensities in positive sera were obtained from proteins involved in protein folding and stabilization or detoxification , like the two GroEL homologs ( BPSL2697 and BPSS0477 ) , two GroES homologs ( BPSL2698 and BPSS0476 ) , and the alkyl hydroperoxide reductase BPSL2096 , respectively [25 , 41] . Heat-shock proteins such as GroEL are generally not considered to be good serodiagnostic candidates , because cross-reactivities between different bacterial species have been described . However , the results described by Felgner et al . ( 2009 ) were confirmed and showed that the heat-shock protein GroEL ( BPSL2697 ) is the most significantly differentially reactive antigen [25] . With respect to the GroEL homologs BPSL2697 and BPSS0477 , and the GroES homologs BPSL2698 and BPSS0476 , we cannot exclude the induction of cross reacting IgG with specificity for common epitopes , since both homolog pairs show high identities with each other ( 84 . 3% and 79 . 2% , respectively ) . Many melioidosis sera reacted with both homologs , but we also found sera recognizing only one of the paralogs , implying an induction of specific antibodies against one of these antigens . Both GroEL and GroES are known to induce strong humoral and cellular immune responses in a variety of bacterial infections [42–45] , and have been proposed as universal vaccine candidates [46] . The alkyl hydroperoxide reductase BPSL2096 was another protein found to be highly antigenic . An upregulation of AhpC homologs was shown for other intracellular species as part of the response to host oxidative stress , and homologs of AhpC were previously shown to be highly immunogenic [47–49] . The immune response to the facultatively intracellular B . pseudomallei is the subject of intensive research , with many questions still unanswered [50–52] . In fact , both cell-mediated and humoral immune responses play important roles in protection against melioidosis [29 , 53] . The validation of our microarray with sera taken from patients at defined time points after admission ( week 0 , 12 and 52 p . a . ) revealed significant differences between antibodies with different specificities . To the best of our knowledge , this is the first report of short- and long-term human antibodies to B . pseudomallei protein antigens . A significant , strong antibody response was demonstrated for seven B . pseudomallei proteins ( group 1 antigens ) , even 52 weeks after infection . Among them are GroEL and GroES homologs , the alkyl hydroperoxide reductase BPSL2096 , and the outer membrane protein BPSL2522 . The outer membrane protein BPSL2522 was shown to be protective in a murine model of disease and to induce a cellular and humoral immune response [54] . Future investigations will show , if group 1 antigens are also useful as markers to detect previous infections in e . g . epidemiological studies . Significant array signals to group 2 antigens could be an indication of a more recent B . pseudomallei infection , since these antigens were mainly recognized in sera of weeks 0 and 12 p . a . , but rarely seen in sera drawn 52 weeks after infection . Interestingly , among this group are two effector proteins ( BPSS1525 and BPSS1532 ) of the type III secretion system cluster 3 ( TTSS3 ) , which are essential for full virulence in murine models [55–57] . A previous study demonstrated that BPSS1525 ( BopE ) could induce specific CD4+ T cells but not CD8+ cells [58] . Further short-term antibody responses were found for flagellin ( BPSL3319 ) and FlgK ( BPSL0280 ) , two proteins that are part of the flagellar apparatus . Flagellin is important for full virulence in mice and is likely to evoke a T cell response [40 , 59 , 60] . The flagellar hook-associated protein FlgK was found to elicit a very short-term antibody response of less than 12 weeks . This is in contrast to the study by Suwannasaen et al . , who describe that FlgK was mainly recognized in sera of recovered melioidosis patients by using the protein array technology of Felgner et al . [40] . But more data are needed to validate the proteins identified as possible early and late antigens in melioidosis diagnostics . In summary , the protein array technology used in this study enables a comprehensive recognition of B . pseudomallei-specific antibody responses in melioidosis . It allows multiplex antigen detection in a miniaturized and automated fashion replacing the traditional “one antigen at a time” method [23 , 61–64] . Generally , our technical approach allows non-proteinogenic antigens , such as the various polysaccharide antigens described , to be included in the protein microarray format . Our method can also be applied for the analysis of B . pseudomallei protein expression in vivo using experimental animal models and can be used to elucidate the exposure to B . pseudomallei in humans and animals in epidemiological studies . Future multicenter studies are needed to determine the true sensitivity and specificity of this protein array as a diagnostic tool in different parts of the world . We are aware that the protein array technology presented might not be affordable in remote rural endemic areas . However , results of future multicenter protein array studies might finally translate into multiple antigen based point of care ( POC ) devices such as lateral flow assays , which should be applicable in low-resource tropical settings .
|
Melioidosis is a potentially fatal infectious disease caused by the Gram-negative environmental bacterium Burkholderia pseudomallei . Since the clinical presentations of melioidosis are extremely variable and no specific signs or symptoms exist , early laboratory-based diagnosis is highly desirable to start appropriate antibiotics . Routine methods for culture detection of B . pseudomallei are highly specific but take several days for a result , and depending on the clinical sample and other factors , sensitivity can be low . The standard serology test for melioidosis is an indirect hemagglutination assay ( IHA ) based on crude B . pseudomallei antigen preparations . Due to the variable prevalence of background seropositivity in endemic areas and the low diagnostic sensitivity of the IHA upon admission , the test is currently not recommended for the diagnosis of melioidosis , but still widely used . Thus , we generated a protein array containing 20 B . pseudomallei antigens previously shown to have serodiagnostic potential . Our array allows highly specific and sensitive antibody recognition in blood sera and plasmas from patients with melioidosis . The standardized microarray device is simple to use and fast , and is thus applicable in a routine diagnostic laboratory . In this study , the multiplex testing of antibodies in melioidosis sera from different time points after admission allowed the detection of short- and long-term antibodies to various antigens . Further studies will examine the potential role of those antibodies to discriminate different stages of the disease . Furthermore , the protein microarray might be useful in studies aimed at elucidating the exposure of humans and animals to B . pseudomallei in different parts of the world .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"melioidosis",
"immunology",
"bacterial",
"diseases",
"chaperone",
"proteins",
"bioassays",
"and",
"physiological",
"analysis",
"antibody",
"response",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"proteins",
"recombinant",
"proteins",
"blood",
"plasma",
"hematology",
"immune",
"response",
"microarrays",
"biochemistry",
"blood",
"anatomy",
"antigen",
"processing",
"and",
"recognition",
"physiology",
"biology",
"and",
"life",
"sciences",
"serum",
"proteins"
] |
2016
|
Rapid and Sensitive Multiplex Detection of Burkholderia pseudomallei-Specific Antibodies in Melioidosis Patients Based on a Protein Microarray Approach
|
The properties of the human immunodeficiency virus ( HIV ) pose serious difficulties for the development of an effective prophylactic vaccine . Here we describe the construction and characterization of recombinant ( r ) , replication-competent forms of rhesus monkey rhadinovirus ( RRV ) , a gamma-2 herpesvirus , containing a near-full-length ( nfl ) genome of the simian immunodeficiency virus ( SIV ) . A 306-nucleotide deletion in the pol gene rendered this nfl genome replication-incompetent as a consequence of deletion of the active site of the essential reverse transcriptase enzyme . Three variations were constructed to drive expression of the SIV proteins: one with SIV’s own promoter region , one with a cytomegalovirus ( cmv ) immediate-early promoter/enhancer region , and one with an RRV dual promoter ( p26 plus PAN ) . Following infection of rhesus fibroblasts in culture with these rRRV vectors , synthesis of the early protein Nef and the late structural proteins Gag and Env could be demonstrated . Expression levels of the SIV proteins were highest with the rRRV-SIVcmv-nfl construct . Electron microscopic examination of rhesus fibroblasts infected with rRRV-SIVcmv-nfl revealed numerous budding and mature SIV particles and these infected cells released impressive levels of p27 Gag protein ( >150 ng/ml ) into the cell-free supernatant . The released SIV particles were shown to be incompetent for replication . Monkeys inoculated with rRRV-SIVcmv-nfl became persistently infected , made readily-detectable antibodies against SIV , and developed T-cell responses against all nine SIV gene products . Thus , rRRV expressing a near-full-length SIV genome mimics live-attenuated strains of SIV in several important respects: the infection is persistent; >95% of the SIV proteome is naturally expressed; SIV particles are formed; and CD8+ T-cell responses are maintained indefinitely in an effector-differentiated state . Although the magnitude of anti-SIV immune responses in monkeys infected with rRRV-SIVcmv-nfl falls short of what is seen with live-attenuated SIV infection , further experimentation seems warranted .
There are good reasons for believing that development of an effective preventive vaccine against HIV-1 is going to be a very difficult task [1–3] . HIV is able to replicate continuously without relent despite apparently strong humoral and cellular immune responses to the virus . The HIV envelope glycoprotein is shielded with a large amount of carbohydrate and the trimer spike as it exists of the surface of virions is difficult for antibodies to access and difficult for antibodies to block infectivity . HIV-1 is highly variable from one individual to another and even within a single individual evolves to evade ongoing immune responses . The virus encodes a number of gene products that function at least in part to evade intrinsic , innate and adaptive immune responses . And during the course of an infection , HIV-1 gradually destroys CD4+ T lymphocytes , a key orchestrator of adaptive immune responses . The inability of infection by one HIV-1 strain to routinely provide protection against superinfection by a different HIV-1 strain supports this perception of great difficulty in development of a protective vaccine [4] . Investigation of a variety of creative , non-standard approaches to a vaccine seem justified given this expected difficulty . Two particular vaccine approaches have shown the greatest protective effects in monkey studies to date using virulent strains of simian immunodeficiency virus ( SIV ) for challenge of Indian-origin rhesus monkeys . The first one consists of live-attenuated strains of SIV , such as those deleted of the nef gene , which have far and away provided the greatest degree of protection against challenge [5–8] . However , even live attenuated SIV has not provided very good protection against challenge with SIV strains not closely matched in sequence to that of the vaccine strain [9–11] . This last point seems consistent with the inability of infection by one HIV-1 strain to routinely provide protection against superinfection as described in the previous paragraph . The second approach consists of live recombinant forms of a fibroblast-adapted strain of the beta-herpesvirus rhesus cytomegalovirus ( CMV ) . Approximately 50% of macaques vaccinated with these CMV-based vectors manifested complete control of viral replication shortly after SIVmac239 infection [12–14] . The remaining monkeys not protected by this CMV-based vaccine exhibited persisting SIV levels in plasma indistinguishable from those in control , unvaccinated monkeys . Independent recombinant CMV vectors expressing Gag , or Pol , or Env , or a Rev-Tat-Nef fusion protein ( RTN ) were combined , but Env-specific antibodies were not elicited . There are a number of potential advantages to use of a recombinant herpesvirus as a vaccine vector . Herpesviruses possess large genomes and can accommodate a large amount of inserted genetic information . Importantly , herpesviruses persist for the lifetime of the infected host and immune responses to their encoded proteins persist in an up , on , active fashion for life . Being a DNA virus , any inserted genetic information can be expected to remain relatively stable for prolonged periods . Furthermore , there are eight distinct human herpesviruses from which to choose , each with distinct target cells for replication , sites of persistence , and composition of genes . A live-attenuated strain of the human alpha herpesvirus varicella zoster virus is part of childhood immunization programs in many countries [15] . Here we describe the construction and properties of a recombinant gamma-herpesvirus ( the rhesus monkey rhadinovirus , RRV ) containing a near-full-length genome of SIV capable of expressing 96 . 7% of its protein products .
The complete SIVmac239 proviral genome including both LTRs is 10 , 279 base pairs [16] . For the construction of near-full-length ( nfl ) recombinants using a different promoter/enhancer region , we eliminated nucleotides 1–521 from the left LTR in the numbering system of Regier and Desrosiers ( Fig 1 ) [16] . We also eliminated nucleotides 9 , 864–10 , 279 from the right LTR ( Fig 1 ) . The remaining sequences retain the RNA start site in the left LTR and the overlap region with nef in the right LTR ( Fig 1 ) . A V5 tag was added to the end of nef , followed by the BGH poly A addition site ( Fig 1 ) . In order to definitively obviate replication competence , a 306-nucleotide in-frame deletion was introduced into pol to remove the active site of the reverse transcriptase enzyme ( Fig 1 ) . These changes retain 96 . 7% of the coding capacity of the SIV genome . In one recombinant ( r ) construct ( rRRV-SIVnfl-cmv ) , a CMV promoter/enhancer region was placed just upstream of the SIV sequences . In a second construct , the SIV promoter/enhancer region was used by restoring the 1–521 nucleotides of the left LTR ( rRRV-SIVnfl-ltr ) . In a third construct , a promoter construct consisting of the promoter for RRV ORF26 ( p26 ) and the promoter for the RRV Poly Adenylated Nuclear RNA ( PAN ) was inserted just upstream of the SIV sequences ( rRRV-SIVnfl-dual ) . ORF26 encodes a RRV capsid protein and is made late during lytic RRV replication . Conversely , PAN encodes one of the most abundant non-coding RNA transcripts of RRV and is made at the onset of lytic replication [17–20] . These constructions were inserted between the left terminal repeats and the first open reading frame ( R1 ) of the RRV genome using procedures previously described [21–24] . To evaluate the expression of SIV proteins , rhesus fibroblasts ( RF ) permissive for lytic RRV replication were infected with rRRVs expressing SIV-nfl under the control of either the CMV promoter , the SIVmac239 LTR region , or a promoter construct denoted RRV dual promoter , consisting of both p26 and PAN promoters . Cell lysates were prepared and analyzed by immunoblotting for the presence of the particular SIV gene products . All three rRRV vectors produced the early protein Nef and the late structural proteins Gag and Env ( Fig 2 ) . The Nef protein was the first detected with all three vectors , suggesting appropriate time-ordered synthesis . Expression levels of the SIV proteins were highest in RF cells infected with the rRRV-SIVcmv-nfl construct . Additionally , the production of the SIV late antigens was first seen at day three post infection in the rRRV-SIVcmv-nfl-infected RF cells . The progression of cytophatic effect was similar with all three viruses , with rRRV-SIVdual-nfl being very slightly slower than the other two , which is why this one goes out to day 6 rather than just day 5 in Fig 2 . The observed intracellular expression levels of the SIV proteins were consistent with levels of SIVmac239 Gag p27 detected in RF cell culture supernatants post rRRV infection . RF cells infected with construct rRRV-SIVcmv-nfl released over 150 ng/ml of Gag p27 into the cell-free supernatant ( Fig 3A ) . These impressive levels of p27 are similar to what is observed following replication of SIV in permissive cells [25 , 26] . In order to confirm the production and integrity of rRRV-derived SIVnfl virions , transmission electron microscopy ( TEM ) analysis of rRRV-SIVcmv-nfl-infected RF cells was performed . The TEM images revealed not only RRV particles but also numerous SIV particles of appropriate size and morphology . Free SIV virions released from the cell and virions budding from the surface of the plasma membrane were visualized ( Fig 3C & 3D ) . Of note , some of the extracellular released SIV particles exhibited the cylindrical or rod-shaped nucleoids of mature lentiviral virions . As expected , replication assays performed in CEMx174 cells confirmed that the rRRV-derived SIVnfl particles were replication-incompetent ( Fig 3B ) . Since the rRRV-SIVcmv-nfl construct resulted in high levels of SIV protein expression in vitro , we set out to evaluate its ability to infect , persist , and elicit anti-SIV immune responses in vivo . Six RRV seronegative rhesus macaques were inoculated intravenously with 109 genome copies of rRRV-SIVcmv-nfl ( Table 1 ) . Enzyme-linked immunosorbet assays ( ELISAs ) were performed to monitor the development of anti-RRV and anti-SIV Env antibodies following the inoculation . Anti-RRV antibodies emerged in all animals by week 6 post rRRV-SIVcmv-nfl inoculation and steadily increased in the ensuing weeks ( Fig 4A ) . Anti-Env antibodies were characterized based on their ability to bind gp140 and gp120 . Vaccine-induced gp140-binding antibodies were first detected at week 6 post vaccination and sharply increased afterward ( Fig 4B ) . Except for monkey r11099 , all animals maintained high , stable levels of gp140-specific antibodies for the 20 weeks of measurement post rRRV-SIVcmv-nfl inoculation ( Fig 4B ) . Vaccine-induced gp120-binding antibodies were also first detected at week 6 post vaccination , although only two macaques ( r11005 and r10094 ) exhibited a similar surge in antibody levels as was observed for gp140-binding antibodies ( Fig 4C ) . Vaccine-elicited gp120-binding antibodies were still increasing in r10094 and r11015 at week 20 post vaccination , the last follow up ( Fig 4C ) . By comparison , these responses remained at low levels in the remaining rRRV-SIVcmv-nfl vaccinated monkeys . Serial dilutions of week 20 sera were also used to assess vaccine-induced gp140- and gp120-binding antibodies . As references , these analyses included sera from two macaques that had been infected with SIVmac239Δnef for 18 weeks and two monkeys that had been inoculated with a rRRV encoding codon-modified SIVmac239 gp160 ( rRRV-SIVc . m . -env ) for 19 weeks [24] . The SIVmac239Δnef vaccinees exhibited the highest levels of both gp140- and gp120-binding antibodies ( Fig 5A & 5B ) . While vaccination with rRRV-SIVc . m . -env resulted in greater levels of gp120-binding antibodies than those achieved with rRRV-SIVcmv-nfl ( Fig 5A ) , reactivity to gp140 was comparable among recipients of both rRRV constructs ( Fig 5B ) . We also evaluated neutralizing activity against SIVmac316 and SIVmac239 using sera obtained at week 20 post rRRV-SIVcmv-nfl administration ( Fig 6 and Table 2 ) . SIVmac316 is an easy-to-neutralize derivative of SIVmac239 [27 , 28] . Sera from all monkeys were capable of neutralizing SIVmac316 with ID50 titers ranging from 1:16 to 1:794 ( Fig 6; Table 2 ) . Little or no neutralizing activity was detected against the difficult-to-neutralize SIVmac239 clone . Vaccine-induced SIV-specific cellular immune responses in peripheral blood mononuclear cells ( PBMC ) were also evaluated . Two of the rRRV-SIVcmv-nfl-vaccinated monkeys ( r11089 and r11099 ) expressed the major histocompatibility complex class I ( MHC-I ) allele Mamu-A*01 , which allowed us to track vaccine-induced SIV-specific CD8+ T-cells by fluorochrome-labeled MHC-I tetramer staining [29] . Mamu-A*01 tetramers folded with peptides corresponding to the immunodominant Gag CM9 ( amino acids 181–189 ) and Tat SL8 ( amino acids 28–35 ) were chosen for this analysis . Vaccine-induced Gag CM9-specific CD8+ T-cells were first detected in both animals at week 4 post vaccination , when they reached a peak of 1 . 3% of peripheral CD8+ T-cells in monkey r11089 ( Fig 7A ) . In animal r11099 , by comparison , these Gag-specific CD8+ T-cell responses reached their highest frequency ( 0 . 3% ) at week 12 post vaccination ( Fig 7A ) . Low levels of vaccine-elicited Tat SL8-specific CD8+ T-cells were detected in both animals at week 4 post vaccination ( Fig 7B ) . While these responses reached a peak of 1 . 8% of peripheral CD8+ T-cells in r11089 at week 12 post rRRV-SIVcmv-nfl inoculation , they remained at a low frequency in r11099 until week 20 ( Fig 7B ) . By that time , CD8+ T-cell responses against both Gag CM9 and Tat SL8 had decayed considerably in the two Mamu-A*01+ vaccinees ( Fig 7A & 7B ) . We also combined MHC-I tetramer staining with multi-color flow cytometry to evaluate the memory phenotype of vaccine-elicited SIV-specific CD8+ T-cells at week 10 post SIVcmv-nfl inoculation . This analysis relied on the differential expression of CD28 and CCR7 by central memory ( TCM; CD28+CCR7+ ) , transitional memory ( TEM1; CD28+CCR7- ) , and terminally differentiated effector memory ( TEM2; CD28-CCR7- ) CD8+ T-cells in rhesus macaques [30] . The vast majority of Gag CM9- and Tat SL8-specific CD8+ T-cells in r11089 exhibited either the TEM2 or TEM1 signature , consistent with the persistent nature of the rRRV-SIVcmv-nfl vector ( Fig 8A ) . The low frequencies of tetramer+ CD8+ T-cells in r11099 at week 10 post vaccination made it difficult to accurately delineate the memory phenotype of vaccine-elicited CD8+ T-cells in this animal ( Fig 8B ) . However , based on the few tetramer+ CD8+ T-cells analyzed , TEM2 and TEM1 subsets appeared to predominate ( Fig 8B ) . Most tetramer+ CD8+ T-cells in r11089 also expressed high levels of the cytotoxicity-associated molecule Granzyme B ( Fig 9A ) . A smaller fraction of tetramer+ CD8+ T-cells in r11099 also appeared to express Granzyme B , albeit at low levels ( Fig 8B ) . To further evaluate the breadth of vaccine-elicited SIV-specific T-cell responses in the rRRV-SIVcmv-nfl-inoculated monkeys , we carried out intracellular cytokine staining ( ICS ) assays in PBMC at weeks 10 and 18 post rRRV-SIVcmv-nfl inoculation . Peptide pools corresponding to eight of the nine SIV proteins were used in the first assay–Pol was the only protein absent from that analysis . Vaccine-elicited CD8+ T-cell responses against all eight SIV proteins were detected , especially against Nef ( Fig 9A ) . SIV-specific CD4+ T-cells were also detected at this time point , although they focused on Gag and were present at much lower frequencies ( Fig 9B ) . To further characterize the breadth of vaccine-elicited T-cell responses , we repeated the ICS assay at week 18 using separate pools of peptides corresponding to each of the nine SIV gene products . Nef remained the most frequently targeted viral protein by vaccine-elicited CD8+ T-cells , although abundant Pol- and Tat-specific CD8+ T-cells were also present at this time point ( Fig 9C ) . CD8+ T-cell reactivity against the Vif and Env pools was either negative or at borderline levels in most animals , even though responses against these two proteins were measured in at least one animal at week 10 post rRRV-SIVcmv-nfl inoculation ( Fig 9C ) . The frequency of vaccine-elicited SIV-specific CD4+ T-cells at week 18 was even lower than that measured at week 10 ( Fig 9B & 9D ) . In sum , despite the high animal-to-animal variability in these ICS assays , these data illustrate the capacity of rRRV-SIVcmv-nfl to elicit T-cell responses against all nine SIV gene products .
Herpesviruses have a number of potential advantages when being considered as vectors for vaccine delivery . Herpesviruses have large DNA genomes and can potentially accommodate significant amounts of inserted genetic information . That potential advantage has been borne out in our study described here in that 9 , 343 base pairs of genetic information have been successfully inserted . Being a DNA virus , inserts can be expected to be reasonably stable in the absence of direct repeats in the insert and with the absence of severe selective disadvantage . Furthermore , herpesviruses persist for the life of the infected individual and immune responses to their proteins persist in an up , on , active fashion for life . This is important when considering vaccine approaches for HIV/AIDS since immunological memory will probably never be enough to contain HIV-1; once a memory response kicks in for a previously-vaccinated , subsequently-infected individual , HIV-1 will employ its vast array of immune evasion and other strategies to allow continuous viral replication . Our results indeed indicate that rRRV-SIVnfl established persistent infection in RRV seronegative monkeys . Antibody responses increased to high levels and persisted through the 20 weeks of analysis . Similarly , cellular responses persisted to one extent or another in the 20 weeks of follow-up analyses . Also , the responding CD8+ T cells exhibited an effector memory phenotype consistent with recurrent antigen production . In contrast to the unconventionally MHC-restricted CD8+ TEM responses elicited by the 68 . 1 rhesus CMV vaccine developed by Hansen et al . [12–14] , rRRV-SIVnfl-vaccinated macaques developed CD8+ T-cells capable of recognizing immunodominant SIV epitopes restricted by classical MHC class I molecules . We are not aware of any live vector system that has attempted what we have described here , i . e . insertion of a nearly complete SIV or HIV genome capable of expression of all nine viral gene products . Ourmanov et al . have described insertion of gag-pol and env genes of SIV into a single recombinant modified vaccinia Ankara ( rMVA ) vector and the ability of the proteins expressed from this construct to assemble into SIV virion particles [31] . However , six of the SIV genes are not present in this rMVA vector and the infection of monkeys is not persistent . While live attenuated strains of HIV-1 are not likely to be tried in people anytime soon , finding other ways to mimic them would seem to be a worthwhile goal . The characteristics of the rRRV-SIVnfl strains described here parallel what many consider to be important features for the relative success of live attenuated SIV in monkeys . The infection appears to be persistent , probably lifelong; >95% of the SIV proteome is naturally expressed; virion particles are formed in abundance; and CD8+ T cell responses are maintained indefinitely in an effector-differentiated state . Can we reasonably expect rRRV-SIVnfl to perform as well as live attenuated SIV in monkey vaccine/challenge experiments , particularly since the magnitude of anti-SIV immune responses with rRRV-SIVcmv-nfl fell considerably short of what is seen with live-attenuated SIV infection ? It is important to remember in this regard that we have so far tested only one strain of rRRV-SIVnfl in monkeys and that the level of transgene expression in cultured cells is not always predictive of the magnitude of the immune response to that transgene product . For example , a rRRV vector with a CMV promoter/enhancer region driving expression of an SIV env gene with an expression-optimized codon usage expressed high levels of Env protein in cultured cells but failed to elicit detectable anti-Env antibodies in vivo [22] . In contrast , rRRV expressing a version of SIV env with a sub-optimal codon usage capable of being induced by the RRV transinducer ORF57 elicited readily-detectable anti-Env antibodies in infected monkeys [24] . Consequently , no matter how rRRV-SIVcmv-nfl performs in monkey vaccine/challenge experiments , a variety of rRRV-SIVnfl vector designs will need to be compared for the magnitude , persistence and nature of the immune responses to the SIV products and their ability to protect against SIV challenge . Furthermore , there may be ways of enhancing the protective effects of rRRV-SIVnfl , for example by priming or boosting regimens .
The details regarding animal welfare described herein are either similar or identical to those published recently [32] . “The Indian rhesus macaques ( Macaca mulatta ) utilized in this study were housed at the Wisconsin National Primate Research Center ( WNPRC ) . All animals were cared for in accordance with the guidelines of the Weatherall report and the principles described in the National Research Council’s Guide for the Care and Use of Laboratory Animals under a protocol approved by the University of Wisconsin Graduate School Animal Care and Use Committee” ( animal welfare assurance no . A3368-01; protocol no . G005022 ) [33] . “Furthermore , the macaques in this study were managed according to the animal husbandry program of the WNPRC , which aims at providing consistent and excellent care to nonhuman primates at the center . This program is employed by the Colony Management Unit and is based on the laws , regulations , and guidelines promulgated by the United States Department of Agriculture ( e . g . , the Animal Welfare Act and its regulations , and the Animal Care Policy Manual ) , Institute for Laboratory Animal Research ( e . g . , Guide for the Care and Use of Laboratory Animals , 8th edition ) , Public Health Service , National Research Council , Centers for Disease Control , and the Association for Assessment and Accreditation of Laboratory Animal Care International . The nutritional plan utilized by the WNPRC is based on recommendations published by the National Research Council . Specifically , macaques were fed twice daily with 2050 Teklad Global 20% Protein Primate Diet and food intake was closely monitored by Animal Research Technicians . This diet was also supplemented with a variety of fruits , vegetables , and other edible objects as part of the environmental enrichment program established by the Behavioral Management Unit . Paired/grouped animals exhibiting stereotypical and/or incompatible behaviors were reported to the Behavioral Management staff and managed accordingly . All primary enclosures ( i . e . , stationary cages , mobile racks , and pens ) and animal rooms were cleaned daily with water and sanitized at least once every two weeks . ” Lights were on a 12:12 diurnal schedule . Vaccinations were performed under anesthesia ( Ketamine administered at 5–12 mg/kg depending on the animal ) and all efforts were made to minimize suffering . Euthanasia was performed at the end of the study or whenever an animal experienced conditions deemed distressful by one of the veterinarians at the WNPRC . All euthanasia were performed in accordance with the recommendations of the Panel on Euthanasia of the American Veterinary Medical Association and consisted of an IV overdose ( greater than or equal to 50 mg/kg or to effect ) of sodium pentobarbital or equivalent , as approved by a clinical veterinarian , preceded by ketamine ( at least 15 mg/kg body weight ) given by the intramuscular ( IM ) route . Additional animal information , including MHC-I , age at the beginning of study , and sex , is shown in Table 1 . Early passage rhesus fibroblasts ( RFs ) were cultured and maintained in Dulbecco’s modified Eagle medium ( DMEM , Thermo Fisher Scientific ) , supplemented with 20% fetal calf serum ( Thermo Fisher Scientific ) , and primocin ( InvivoGen ) . TZM-bl cells ( ATCC ) were cultivated in DMEM DMEM medium ( Thermo Fisher Scientific ) , supplemented with 10% fetal calf serum ( Thermo Fisher Scientific ) , and primocin ( InvivoGen ) . CEMx174 cells ( NIH AIDS Reagent Program ) were cultivated in RPMI 1640 medium ( Thermo Fisher Scientific ) , supplemented with 10% fetal calf serum ( Thermo Fisher Scientific ) , and primocin ( InvivoGen ) . A plasmid containing proviral SIVmac239 DNA served as a template for the generation of a near full-length genome ( nfl ) sequence of SIVmac239 . Two polymerase chain reactions ( PCR ) led to two SIVmac239 sequence fragments . The primers used were forward primer; ACTTAAGCTTGGTACCGAGCTCGGATCCTCGCTCTGCGGAGAGGCTGGC and reverse primer; GAGTTCCTTTGACTGTAAAACTCCTGCAGGGTGTGGTATTCC , as well as forward primer; GGAATACCACACCCTGCAGGAGTTTTACAGTCAAAGGAACTC and reverse primer; CCACTGTGCTGGATATCTGCAGAATTCGCGAGTTTCCTTCTTGTCAGC . Subsequently , using the PCR-derived BamHI and EcoRI overlaps the SIV-nfl sequence was Gibson cloned ( New England BioLabs ) into expression plasmid pcDNA6/V5 His A ( Thermo Scientific ) , hereby generating a SIVnfl sequence containing a 520 bp deletion in the 5’ Long Terminal Repeat ( LTR ) region , a deletion in the pol region spanning 306 bp , corresponding to the active site of the reverse transcriptase , and a 414 bp deletion in the 3’ LTR region . Furthermore , the stop-codon after the nef open reading frame ( ORF ) was deleted resulting in a pcDNA6-derived V5-tag following nef . A subsequent PCR using the pcDNA6-SIV-nfl plasmid as a template , the forward primer; AGGTACTAGTCCGGCGCCCCGTTTAAACTGACACCTACTCAGACAATGCGAT and reverse primer; ACTATGTGTTACTACTAGTTGTTTAAACTGCTTCGCGATGTACGGGCCAGAT yielded a construct comprised of the cytomegalovirus immediate-early enhancer and promoter ( pCMV ) , the SIV-nfl sequence containing a C-terminal V5-tagged nef ORF , and the bovine growth hormone ( BGH ) polyA signal . Subsequently , utilizing PmeI restriction site overlaps , the SIVnfl construct was cloned into cosmid ah28dA/H between the left terminal repeats ( TR ) and the first ORF R1 of RRV via Gibson assembly ( New England BioLabs ) . Full length Recombinant RRV was made via co-transfection of five overlapping cosmids as previously described [21] . Recombinant RRV-infected RF culture supernatants were harvested , spun down twice at 2000 rcf for 5 min to remove any cell debris , and resulting virus titers were measured via quantitative real-time PCR using a RRV latency-associated nuclear antigen ( LANA ) specific primer set . The reaction was completed using the TaqMan Fast Virus 1-Step Master Mix ( Thermo Fischer Scientific ) in a Real-time PCR thermocycler ( Thermo Fisher Scientific ) ; Forward primer; ACCGCCTGTTGCGTGTTA , reverse primer; CAATCGCCAACGCCTCAA , reporter; FAM- CAGGCCCCATCCCC . FAM- CAGGCCCCATCCCC . High-titered recombinant RRV stocks were aliquoted and stored frozen . For immunoblotting experiments , 2 . 0×105 RFs were seeded into wells of a 6-well plate . The next day , cells were infected with 50 μL of a stock of recombinant RRV containing 109 genome copies/ml expressing either SIVcmv-nfl , SIVlrt-nfl , or SIVdual-nfl . Cells were kept in culture up to six days . Every day cells of one well infected with the respective rRRV were harvested up until the cells were exhibiting advanced cytopathic effect . Cells were harvested , resuspended , and lysed with an NP40-based lysis buffer including a protease inhibitor ( Roche ) . Cell lysates were spun down to remove any cell debris . Subsequently , supernatants were transferred into new tubes and their protein levels were measured and normalized using a bicinchoninic acid protein assay kit ( Pierce ) . Each lysate was mixed with an equal volume of 2x SDS Laemmli sample buffer ( Sigma-Aldrich ) containing 2-mercaptoethanol . Then , samples were incubated at 97 °C for 10 min and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) . The proteins from the gels were transferred to a polyvinylidene difluoride ( PVDF ) membrane ( Bio-Rad ) and subjected to immunoblotting . Furthermore , the PVDF membranes were blocked with 1 x PBS containing 5% skim milk briefly and incubated with the appropriate primary antibodies directed against specific SIV antigens overnight at 4 °C . Three wash steps in PBS containing tween were followed by incubation with the appropriate HRP-conjugated secondary antibodies . After a final washing procedure , specific signals were detected with a LAS4000 mini system ( GE Healthcare Systems ) using SuperSignal West Pico chemiluminescent substrate ( Pierce ) . Two hundred and fifty thousand RF cells were seeded into 6-well culture plates and subsequently infected with a total of 3 . 5×109 genome copies of SIVcmv-nfl or SIVdual-nfl . Cell culture supernatants were collected daily and all samples were subjected to a commercial antigen capture assay ( ABL ) to measure the levels of SIVmac239 Gag p27 . Supernatant of uninfected RF cells served as a negative control . Cell culture supernatant from rRRV-SIV-nfl-infected , early passage RFs were taken . Subsequently , CEMx174 cells were infected with these supernatants containing a total of 20 ng of SIVmac239 Gag p27 . Gag p27 levels were measured by a commercial antigen capture assay ( ABL ) . Additionally , 20ng of Gag p27 of HEK293T cell-produced wild type SIVmac239 served as a positive control . The supernatants from the CEMx174 cell cultures were collected daily post infection and were subsequently used in an antigen capture assays measuring Gag p27 levels ( ABL ) . Supernatant taken from cultures of uninfected RF cell served as a negative control . One million early passage RF cells were seeded into T75 culture flasks . The following day , cells were infected with 200 μL containing 2 . 0×108 genome copies of rRRV-SIVcmv-nfl . cells were then harvested at days four to six post infection . Subsequently , the cells were fixed with 2 . 5% glutaraldehyde in sodium cacodylate buffer . Subsequent transmission electron microscopic ( TEM ) images were taken at the Core Electron Microscopy Facility , University of Massachusetts , according to published procedures [34] . One milliliter of PBS containing 109 genome copies of rRRV-SIVcmv-nfl was administered to six rhesus macaques intravenously . Sera from all research animals were tested for the presence of gp120-binding antibodies . ELISA plates were coated with recombinant gp120 of SIVmac239 ( Immune Technology ) for 1 h at 37 °C and the plates were then washed with PBS Tween20 ( Sigma-Aldrich ) . Subsequently , the ELISA plates were blocked with PBS containing 5% of skim milk ( Bio-Rad ) . The monkey sera were diluted 1:20 in blocking buffer and applied to the ELISA plate . Then , ELISA plates were incubated for 1 h at 37 °C and the plates washed again . Subsequently , an HRP-conjugated goat anti-rhesus IgG H+L antibody ( SouthernBiotech ) was added and the ELISA plates incubated for one h at 37 °C . Sequentially , TMB substrate ( SouthernBiotech ) and stop solution ( SouthernBiotech ) were added . Lastly , the absorbance at 450 nm per well was read in a microplate reader ( PerkinElmer ) . Similarly , sera from all six research animals were tested for the presence of antibodies against SIVmac239 gp140 by coating the ELISA plates with homologous purified protein and probing the samples with a horseradish-peroxidase-conjugated anti-rhesus IgG antibody ( SouthernBiotech ) . All sera were also screened for the presence of anti-RRV antibodies by coating ELISA plates with purified RRV lysate and subsequently probing of samples ( diluted 1:20 ) with a horseradish-peroxidase-conjugated anti-rhesus IgG antibody ( SouthernBiotech ) . Sera from all six research animals were screened for neutralization of SIVmac316 utilizing the TZM-bl assay , as described previously [35] . The ID50 titer is defined as the reciprocal of the highest dilution of serum that reduced SIVmac316 infectivity by 50% . Values were calculated using the Sigmoidal , 4PL , X is log ( concentration ) equation in Prism7 ( GraphPad Software ) . The tetramer staining assays performed as part of the time course analysis of vaccine-induced SIV-specific CD8+ T-cells in Mamu-A*01+ macaques ( Fig 7 ) was done by labeling PBMC with titrated amounts of fluorochrome-conjugated Mamu-A*01/Gag CM9 ( MBL International Inc . ) or Mamu-A*01/Tat SL8 ( Tetramer Core Facility ) tetramers [29] . Up to 800 , 000 PBMC were incubated with the respective tetramers at 37 °C for 1 hr and then stained with fluorochrome-labeled monoclonal antibodies ( mAbs ) directed against the surface molecules CD3 ( clone SP34-2 ) and CD8α ( clone RPA-T8 ) . After a 25-min incubation at RT , the cells were washed and then fixed with PBS containing 2% of paraformaldehyde . Data were acquired with a SORP BD LSR II ( BD BioSciences , San Jose , CA ) flow cytometer and analyzed with FlowJo software version 9 . 9 . 3 ( Tree Star , Inc . Ashland , OR ) . CD3+ T cells were gated within the lymphocyte gate defined by the forward and side-scatter properties . Tetramer positive cell frequencies were determined within the CD3+CD8+ cell population . The following description of how the memory phenotype and granzyme B content of tetramer+ CD8+ T-cells were determined is either identical or similar to that used in one of our recent publications [36] . “Up to 8 . 0×105 cells were incubated in the presence of the appropriate fluorochrome-labeled tetramer at 37 °C for 1 h and then stained with monoclonal antibodies ( mAbs ) directed against the surface molecules CD3 ( clone SP34-2 ) , CD8α ( clone RPA-T8 ) , CD28 ( clone 28 . 2 ) , CCR7 ( clone 150503 ) , CD14 ( clone M5E2 ) , CD16 ( clone 3G8 ) , and CD20 ( clone 2H7 ) . Amine-reactive dye ( ARD; Live/DEAD Fixable Aqua Dead Cell Stain; Life Technologies ) was also added to this mAb cocktail . After a 25-min incubation at room temperature , we treated the cells with BD FACS Lysing Solution ( BD Biosciences ) for 10 min and subsequently washed them with “Wash Buffer” ( Dulbecco’s PBS with 0 . 1% BSA and 0 . 45 g/L NaN3 ) . Cells were permeabilized by treatment with “Perm buffer” [1X BD FACS Lysing Solution 2 ( Beckton Dickinson ) and 0 . 05% of Tween-20 ( Sigma-Aldrich ) ] for 10 min . Cells were then washed once and stained with a Granzyme B-specific mAb ( clone GB12 ) . After a 30-min incubation in the dark at room temperature , cells were washed and stored at 4 °C until acquisition . Samples were acquired using FACS DIVA version 6 on a Special Order Research Product BD LSR II apparatus equipped with a 50-mW 405-nm violet , a 100-mW 488-nm blue , and a 30-mW 635-nm red laser . We used FlowJo 9 . 6 ( Treestar , Inc . ) to analyze data . First , we gated on diagonally clustered singlets by plotting forward scatter height ( FSC-H ) versus FSC area ( FSC-A ) and then side scatter height ( SSC-H ) versus SSC area ( SSC-A ) . Next , we created a time gate that included only those events that were recorded within the 5th and 90th percentiles and then gated on “dump channel” negative , CD3+ cells . At this stage , we delineated the lymphocyte population based on its FSC-A and SSC-A properties and subsequently gated on CD8+ cells . After outlining tetramer+ cells , we conducted our memory phenotyping analysis within this gate . Cells stained with fluorochrome-labeled mAbs of the same isotypes as the anti-Granzyme B , anti-CD28 , and anti-CCR7 mAbs guided the identification of the memory subsets within the tetramer+ population . Based on this gating strategy , the tetramer frequencies shown in Fig 9 correspond to percentages of live CD3+ CD8+ tetramer+ lymphocytes . ” Pools of peptides ( 15mers overlapping by 11 amino acids ) spanning all nine SIVmac239 gene products were used for T-cell stimulation in the week 18 assay . Given the large size of Gag , Pol , and Env , the peptides covering these polyproteins were divided in 2 , 3 , and 2 pools , respectively . Peptides spanning each of the remaining accessory ( Vpr , Vpx , Vif , and Nef ) and regulatory ( Rev and Tat ) proteins were grouped in individual pools . For the week 10 assay , Vpx and Vpr peptides were grouped in a single pool , as were the Rev and Tat peptides . Pol peptides were not used in the week 10 assay . The final assay concentration of each 15mer was 1 . 0 μM . The following description on how the ICS assays were set up is nearly identical to that used in our recent publications [32] . “PBMC obtained from the research animals were stimulated with the appropriate pools of SIVmac239 peptides in RPMI 1640 medium supplemented with GlutaMax ( Life Technologies ) , 10% FBS ( VWR ) , and 1% antibiotic/antimycotic ( VWR ) containing co-stimulatory mAbs against CD28 and CD49d for 9 h at 37 °C in an incubator with a 5 . 0% CO2 concentration . Moreover , a phycoerythrin-conjugated mAb specific for CD107a was included in the assay . Brefeldin A ( Biolegend , Inc . ) and GolgiStop ( BD Biosciences ) were added to all tubes 1 h into the incubation time to inhibit any protein transport . Surface molecules of cells were stained as mentioned above and cells were fixed with a 2% paraformaldehyde solution . In addition to the same mAbs against CD14 , CD16 , and CD20 and the ARD reagent described above , the surface staining master mix also included mAbs against CD4 ( clone OKT4; Biolegend , Inc . ) and CD8 ( clone RPA-T8; Biolegend , Inc . ) . Cells were permeabilized by resuspending them in “Perm Buffer” ( 1× BD FACS lysing solution 2 ( Beckton Dickinson ) and 0 . 05% Tween-20 [Sigma-Aldrich] ) for 10 min and subsequently washed with Wash Buffer . Cells were then incubated with mAbs against CD3 ( clone SP34-2 ) , IFN-γ ( clone 4S . B3 ) , TNF-α ( clone Mab11 ) , and CD69 ( clone FN50 ) for 1 h in the dark at RT . After this incubation was completed , the cells were washed and subsequently stored at 4 °C until acquisition . The data were analyzed by gating first on live CD14–CD16–CD20–CD3+ lymphocytes and then on cells expressing either CD4 or CD8 but not both markers . Functional analyses were conducted within these two compartments . Cells were considered positive for IFN-γ , TNF-α , or CD107a only if they co-expressed these molecules with CD69 , a marker of recent activation . Once the appropriate gates were created , we employed the Boolean gate platform to generate a full array of possible combinations , equating to 8 response patterns when testing three functions ( 23 = 8 ) . Leukocyte activation cocktail ( LAC; BD Pharmingen ) -stimulated cells stained with fluorochrome-labeled mAbs of the same isotypes as those against IFN-γ , TNF-α , and CD107a guided the identification of positive populations . We used two criteria to determine if responses were positive . First , the frequency of events in each Boolean gate had to be at least two-fold higher than their corresponding values in background-subtracted negative-control tests . Second , the Boolean gates for each response had to contain ≥10 events . The magnitude of responding CD4+ or CD8+ T-cells was calculated by adding the frequencies of positive responses producing any combination of IFN-γ , TNF-α , and CD107a . All calculations , including background subtraction and evaluation of the frequencies of responding cells , were performed with Microsoft Excel . ”
|
Given the magnitude and impact of the HIV/AIDS pandemic , development of a safe , effective vaccine against HIV remains a top priority for biomedical research . While live-attenuated strains of the simian immunodeficiency virus ( SIV ) have shown promise in monkey studies , concern for safety has limited efforts along these lines . In an attempt to mimic the epitope presentation , epitope coverage , and persistence of live attenuated SIV , we have generated recombinant strains of rhesus monkey rhadinovirus ( RRV; a gamma-2 herpesvirus ) containing a near-full-length genome of SIV . The near-full-length genome retains 96 . 7% of the coding capacity of SIV yet is incompetent for replication . Such recombinant RRV produces abundant SIV particles in infected cells in culture . Monkeys inoculated with one of these recombinant RRV strains became persistently infected , made readily detectable antibodies against the SIV envelope protein , and developed cellular immune responses to all nine SIV gene products .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"physiology",
"pathogens",
"immunology",
"microbiology",
"vertebrates",
"animals",
"mammals",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"primates",
"rna",
"viruses",
"cytotoxic",
"t",
"cells",
"antibodies",
"immunologic",
"techniques",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"white",
"blood",
"cells",
"monkeys",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"t",
"cells",
"immunoassays",
"immune",
"response",
"siv",
"macaque",
"biochemistry",
"eukaryota",
"cell",
"biology",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"lentivirus",
"amniotes",
"organisms"
] |
2018
|
A recombinant herpesviral vector containing a near-full-length SIVmac239 genome produces SIV particles and elicits immune responses to all nine SIV gene products
|
Lepromatous leprosy caused by Mycobacterium leprae is associated with antigen specific T cell unresponsiveness/anergy whose underlying mechanisms are not fully defined . We investigated the role of CD25+FOXP3+ regulatory T cells in both skin lesions and M . leprae stimulated PBMC cultures of 28 each of freshly diagnosed patients with borderline tuberculoid ( BT ) and lepromatous leprosy ( LL ) as well as 7 healthy household contacts of leprosy patients and 4 normal skin samples . Quantitative reverse transcribed PCR ( qPCR ) , immuno-histochemistry/flowcytometry and ELISA were used respectively for gene expression , phenotype characterization and cytokine levels in PBMC culture supernatants . Both skin lesions as well as in vitro antigen stimulated PBMC showed increased percentage/mean fluorescence intensity of cells and higher gene expression for FOXP3+ , TGF-β in lepromatous ( p<0 . 01 ) as compared to tuberculoid leprosy patients . CD4+CD25+FOXP3+ T cells ( Tregs ) were increased in unstimulated basal cultures ( p<0 . 0003 ) and showed further increase in in vitro antigen but not mitogen ( phytohemaglutinin ) stimulated PBMC ( iTreg ) in lepromatous as compared to tuberculoid leprosy patients ( p<0 . 002 ) . iTregs of lepromatous patients showed intracellular TGF-β which was further confirmed by increase in TGF-β in culture supernatants ( p<0 . 003 ) . Furthermore , TGF-β in iTreg cells was associated with phosphorylation of STAT5A . TGF-β was seen in CD25+ cells of the CD4+ but not that of CD8+ T cell lineage in leprosy patients . iTregs did not show intracellular IFN-γ or IL-17 in lepromatous leprosy patients . Our results indicate that FOXP3+ iTregs with TGF-β may down regulate T cell responses leading to the antigen specific anergy associated with lepromatous leprosy .
The hall mark of leprosy caused by Mycobacterium leprae is involvement of the skin and peripheral nerves of man . Leprosy patients present with varied clinic-pathological manifestations and bacterial load which are influenced by the host immune response . Tuberculoid leprosy , both polar ( TT ) and borderline forms ( BT ) show localized paucibacillary , hypo pigmented , hypo anesthetic patches and early nerve damage . Borderline ( BL ) and polar forms of lepromatous leprosy ( LL ) present as generalized disease with multiple , multibacillary skin patches along with involvement of other organs [1] . Whereas tuberculoid leprosy patients show good recall T cell mediated immune and poor antibody responses to the M . leprae antigens , lepromatous patients show a reverse pattern . Moreover , LL patients show specific T cell unresponsiveness to the causative organism though capable of mounting T cell responses to other antigens . The mechanisms underlying the antigen specific anergy are being intensely investigated . In the 70 s a subset of suppressor T cells were first described as a distinct population that inhibited responses through soluble factors [2] . Such cells with CD8 phenotype were thought to be responsible for the T cell anergy in lepromatous leprosy [3] . Others found that monocytes/macrophages from LL patients , either alone [4] or through soluble factors including PGE2 and IL-10 were able to suppress in vitro T cell responses [5] , [6] . Suppressor T cells became a subject of controversy amongst immunologists as molecular or biochemical markers could not be found to identify this subset . With the discovery of Th1 and Th2 subsets having mutually exclusive cytokine patterns , suppression was thought to be mediated by regulatory cytokines [7] . Th1 and Th2 were reported to be associated with tuberculoid and lepromatous leprosy respectively and became a popular notion to explain the leprosy spectrum [8] . However , the finding that some patients of both clinical types of leprosy also showed non polarized Th0 subset with production of both IFN-γ and IL-4 was intriguing and made it difficult to reconcile the leprosy spectrum and anergy based solely on the Th1 and Th2 paradigm [9] . Recent studies from our laboratory showed that leprosy patients with the non polarized Th0 subset had increased percentage of Th17 cells which may constitute the third subset of Th types in leprosy patients who failed to show Th1 and Th2 polarization [10] . Nevertheless , the nature of antigen specific anergy in leprosy continues to evade consensus . A seminal discovery was made in 1995 by Sakaguchi et al [11] who showed that cells responsible for inhibition of organ specific autoimmunity were CD4+ T cells which expressed CD25 and later shown to also express transcription factor forkhead box P3 ( FoxP3 ) in mice [12] and man [13] . The discovery of CD4+CD25+FOXP3+ cells as suppressors of autoimmune responses and their presence in man [14] has revived the concept of a distinct lineage of T cells that negatively regulate the immune responses in order to maintain homeostasis and were designated as regulatory T cells ( Tregs ) . Though various subsets of Tregs have been described , in general they can be divided into thymus derived naturally occurring Tregs and peripherally derived adaptive/inducible Tregs ( Th3 , Tr1 ) [15] . Whereas the former act by direct cell contact , the latter induce suppression through cytokines TGF-β ( Th3 ) and IL-10 ( Tr1 ) [15] , . FOXP3 is thought to be a primary requirement for suppressive function . However , in the humans , low expression has been noted transiently in CD4+CD25neg non suppressive T cells and activated T cells with and without [17] , [18] suppressive function . Moreover , in the periphery , CD4+CD25+ Tregs may be induced by antigen from CD4+CD25neg naïve T cells . Though mice express CD25 constitutively , in the human only Tregs that highly express it ( CD25hi ) show suppressive activity [16] . Other human subtypes that have been suggested include CD45RA+FOXP3low , CD45RA−FOXP3+ resting and activated Tregs respectively as well as CD45RAneg FOXP3low non suppressive , cytokine secreting Tregs [19] . Recently , recommendations have been made to simplify the nomenclature of Treg cells to include ‘thymus derived’ ( nTreg ) instead of ‘natural’ and ‘peripheral derived’ instead of ‘induced or adaptive’ and that ‘in vitro-induced Tregs ( iTregs ) should be distinguished from populations generated in vivo [20] . . With a view to understanding the antigen specific anergy associated with the generalized form of leprosy , we revisited the concept of suppressor T cells by investigating the role of Tregs in skin lesions and in vitro stimulated PBMC cultures of lepromatous and tuberculoid leprosy patients as well as healthy household contacts using qPCR for gene expression , flowcytometry for phenotype characterization and ELISA for cytokine levels in culture supernatants . Though we await the acceptance of the new nomenclature , we thought it appropriate to use iTreg to denote in vitro stimulated PBMC in this study . Taken together , our studies show an increase in TGF-β+ CD4+ CD25+ FOXP3+ T cells both in dermal lesions and in in vitro induced antigen but not mitogen stimulated PBMC ( iTreg ) of lepromatous leprosy patients .
Amongst the leprosy types , lepromatous leprosy is associated with antigen specific T cell unresponsiveness/anergy whose underlying mechanisms have not been fully characterized . In view of recent reports of T regulatory ( Treg ) cells that dampen immune responses [16] , we investigated the role of CD25+FOXP3+ T cells in patients with anergic generalized form of lepromatous leprosy and compared with patients of the more limited form of borderline tuberculoid leprosy ( BT ) . Healthy house hold subjects with long time exposure to infected lepromatous leprosy patients and skin samples from healthy patients undergoing cosmetic surgery were included as controls . Skin lesions and mitogen/antigen stimulated PBMC were studied using quantitative reverse transcribed PCR ( qPCR ) for expression of genes along the FOXP3 pathway , cytokines , transcription factors and signaling molecules; flowcytometry was used for identification of cell types and fluorescence intensity of markers in PBMC and ELISA for measuring cytokines TGF-β , IL-10 , IFN-γ and IL-17 in the culture supernatants of stimulated PBMC . A total of 56 newly diagnosed untreated leprosy patients ( 43 males , 13 females aged between 19–60 years ) from Leprosy Clinics of the Department of Dermatology , Safdarjung Hospital , New Delhi were included in the study ( Table 1 ) and classified on the basis of Ridley-Jopling classification [1] . Study group included 28 borderline tuberculoid ( BT ) , 28 lepromatous ( LL ) , 7 healthy household contacts ( HC ) exposed to leprosy patients and skin samples from 4 healthy subjects undergoing cosmetic surgery . Exclusion criteria included patients below 15 years of age , pregnant women , clinical evidence of anemia and other infections such as tuberculosis , HIV and helminthic infestation . Skin biopsies from 10 each of BT , LL patients and 4 normal subjects were investigated for immunohistochemistry and gene expression studies by quantitative RT-PCR ( qPCR ) . PBMC were investigated on additional 10 each of BT , and LL patients for gene expression; other 8 each of BT and LL subjects , for flow cytometry analysis . The study was approved by Institutional Ethical Committee [08-09-EC ( 3/7 ) ] of Safdarjung Hospital , New Delhi , India . Informed written consent was obtained from the patients after counseling and prior to obtaining blood and tissue samples . Fresh PBMC were isolated in <2 h after obtaining the sample , from 10 ml of sterile heparinized ( Brawn laboratories , Haryana India ) blood by Ficoll-Hypaque density gradient method ( Histopaque , Sigma Aldrich , USA ) after diluting with 1∶1 in RPMI 1640 ( Sigma Aldrich , MO , USA ) as described earlier [10] . In brief mononuclear cells were isolated by centrifugation at 800 g for 20 minutes , washed three times in sterile 1× HBSS ( GIBCO , NY , USA ) and re-suspended in RPMI 1640 with 10% pooled human AB serum , 2 mM L-glutamine , 100 units of penicillin ( Alembic Chemicals , India ) and 100 µg streptomycin ( Sarabhai Chemicals , India ) ) . Cell yield ranged from 1 . 3 to 1 . 5×106 per ml and cell viability ranged from 95–98% as estimated by 0 . 2% trypan blue staining ( Sigma Aldrich , MO , USA ) 2×106 cells/ml were cultured for 48 h in sterile flat bottom 24- well plates ( Falcon , NJ , USA ) with and without 25 µl of T cell mitogen PHA ( 5 µg/ml of phytohemagglutinin , Sigma ) and of heat killed armadillo derived M leprae sonicated antigen ( 10 µg/ml ) kindly provided by P J Brennan of Colorado State University and incubated at 37°C in humidified 5% CO2+air . After harvest cells were washed as above and stored in RNA later ( Sigma ) for gene expression studies or immediately processed for flow cytometry analysis as given below . Paired culture supernatants were collected , centrifuged to remove cell debris and stored at −80°C for estimation of cytokines by ELISA . Skin biopsies were obtained from typical lesions by anesthetizing the area with 1% lignocaine ( Kremoint Pharma , Mumbai Maharashtra , India ) and applying sterile 4 mm punch ( Cardiograph Co , Satara , Maharashtra , India ) . Normal skin was obtained from 4 subjects undergoing cosmetic surgery . Part of the biopsy was processed in buffered formalin for routine histopathology and immunohistochemistry . The remainder was placed in 1 ml of RNA later ( Sigma ) and stored at −80°C till further use . RNA was isolated from: i ) stored skin biopsies after thawing and crushing the tissue with liquid nitrogen in pestle and mortar , ii ) PBMC were homogenized in 1 ml syringes using RNeasy Mini Kit ( Qiagen , Maryland , USA ) according to the manufacturer's instructions . The isolated RNA was quantified using Nanodrop spectrophotometer ( Nanodrop Technologies , Wilmington , USA ) . Only samples with OD of 1 . 8 to 2 . 0 at 260/280 nm were used . The quality of RNA was also checked for 28 s and 18 s RNA by electropherogram using Bio analyzer ( Agilent Technologies , Inc , Singapore ) . RNA Integration Number value of ≥7 was considered to be optimum . For cDNA synthesis 1 µg total RNA was transcribed with RT First strand kit ( SA Biosciences , MD , USA ) . Reactions were performed according to the manufacturer's instructions and the cDNA stored at −20°C till further use . Gene expression was measured in real-time using customized real time PCR arrays ( SA Biosciences , Quiagen Co . CA , USA ) as per the manufacturer's instructions . Duplicate samples of cDNA from antigen stimulated PBMC from each subject was amplified in 96 well plates containing primers for the genes of interest , cytokines IL-2 , TGFβ , IL-10 , IL-27 and IL-25 , CD marker CD28 , transcription factors FOXP3 , STAT5A , GATA3 , NFkB1 , STAT3 , STAT4 and chemokine IL-8 as well as 5 housekeeping genes β2M , HPRT1 , RPL13A , GAPDH , ACTB . 1 µg of cDNA was used per reaction in wells containing the ready to use PCR master mix and appropriate primers . These were then subjected to qPCR ( ABI 7000 , Applied Biosystems Singapore ) for 2 h . Threshold cycle values were normalized and expressed as ΔCt: mean Ct of gene of interest - mean Ct of set of 5 housekeeping genes . For intracellular staining , in vitro antigen and PHA stimulated cells were incubated with monensin ( BD GolgiStop ) for 8 h prior to harvest to block secretion of cytokine . All reagents were obtained from BD Biosciences , San Diego , CA . and used as per manufacturer's instructions . Staining was undertaken within 1 h after harvest and washing three times as above and determining cell viability which ranged from 91–95% . In brief , for cell surface staining , 0 . 5×106cells/50 µl in staining buffer were incubated with a cocktail containing anti human CD3 ( Per cpcy-5 . 5 , clone:UCHT1 ) , CD4 ( APC-H7 , clone:SK3 ) , CD8 ( PE-Cy7 , clone:RPA-T8 ) and CD25 ( FITC , clone:M-A251 ) for 45 min at 4°C . After cell surface staining , cells were incubated with 1× FOXP3 buffer A for 10 min at room temperature; cells were washed two times and permeabilized with buffer C for 30 min at room temperature . The cells were washed two times , resuspended in stain buffer and incubated with anti human FOXP3 ( APC , clone:259D/C7 ) and TGF-β ( PE , clone:TW4-9E7 ) at room temperature for 30 min in the dark , followed by two washes as before and resuspended in 500 µl . For evaluating phosphorylation of STAT5A ( Alexa Flour-647 , clone:47/stat5 ( PY694 ) , cells were first fixed for 10 min at room temperature , permeabilized as before with appropriate buffer and stained with a cocktail of anti human STAT5A anti human CD25 , CD3 , CD4 and CD8 antibodies . . CD3+CD4+ and CD3+CD8+ T cells were gated following forward angle and side scatter characteristics of lymphocytes . Results were analyzed using BD FACS aria flow cytometry along with isotype controls of phycoerythrin ( PE mouse IgG1 ) , Alexa Fluor 488 ( mouse IgG1 ) , Alexa Flour 647 ( mouseIgG1 ) . Supplementary figure show strategy and standardization used for validating the results on multi color flowcytometry . Cytokines were estimated by ELISA ( Ready Set Go , e-Bioscience , San Diego , CA , USA ) as per manufacturer's instructions . In brief , 100 µl/well of cell free supernatants from antigen stimulated PBMC cultures were tested in duplicate in 96-well plates ( Nunc , Rochester , NY , USA ) pre-coated with biotin conjugated anti human antibodies for TGF-β IL-10 , IFN-γ and IL-17 . Plates were incubated overnight at 4°C , washed 5 times , blotted and wells blocked with assay diluents for 1 h at room temperature . After washing with buffer , appropriate avidin-horseradish peroxidase-conjugated anti-mouse antibody was added and the plates incubated at room temperature for 30 min . After washing as before , color development was undertaken using peroxidase color substrate TMB ( Tetramethylbanzedine ) and the reaction stopped by the addition of 1 N H2SO4 . The optical density ( OD ) of each well was read at 450 nm . 4–5 µm thick formalin fixed paraffin embedded ( FFPE ) tissues were cut by rotary microtome ( Leica Biosystems Nussloch , Germany ) , sections picked up on poly L-lysine ( Sigma Aldrich , MO , USA ) coated slides and stored at room temperature . Antibodies ( dilution 1∶50 ) used in this study were mouse anti human FOXP3 ( forkhead box protein3 , e-Biosciences , San Diego , USA ) , rabbit polyclonal anti human TGF-β1 and IL-10 , ( Santa Cruz Biotechnology CA , USA ) . IHC was performed using enhancer HRP-polymer detection method ( BioGenex , USA ) . In brief after deparaffinization , rehydration and blockade of endogenous peroxidase activity by 30% H2O2 and antigen-retrieval with Tris-EDTA ( pH-9 . 0 ) buffer , sections were incubated with 1% albumin , bovine , pH 7 . 0 heat- shock fractionated protein block ( USB co , Cleveland , OH USA ) for 1 h , followed by incubation with anti-human FOXP3 , TGF-β and IL-10 for 1 h . Color was developed using diaminobenzidine ( DAB1 ) chromogen system . The staining protocols were all performed at room temperature except for the primary antibody incubation at 4°C in humidified chamber . Positive and negative stained cells were counted under the microscope using Image Pro express 6 . 0 software ( Media cybernetics , USA ) and percentage calculated after examining 1000 cells from multiple fields . Nonparametric statistics was performed using Graph Pad Prism version 5 ( Graph Pad Software , Inc . , San Diego , CA , USA ) . Data were analyzed using two tailed Mann-Whitney test . p<0 . 05 was considered as statistically significant .
Skin lesions of both tuberculoid and lepromatous leprosy patients showed the presence of nuclear FOXP3+ staining using immunohistochemistry ( Figure 1 ) . Whereas they were present in a circumscribed pattern around as well as amongst the epitheloid cells of the tuberculoid granulomas , FOXP3+ cells were scattered amongst the foamy macrophages of the lepromatous granulomas in the dermis . TGF-β and IL-10 showed diffuse cytoplasmic staining with the latter showing lower intensity in our hands . The distribution of positive cells for all markers was not uniform and 1000 cells were enumerated to obtain percentage of positive cells . As may be seen from Figure 1B there was significant increase in FOXP3+ cells ( p<0 . 006 , two tailed Mann Whitney test ) in lepromatous ( LL ) with Mean% ± SD of 7 . 3±3 . 8 as compared to 3 . 6±2 . 0 in tuberculoid leprosy ( BT ) . TGF-β and IL-10 reported to be associated with FOXP3 cells [16] were seen in both leprosy types . The percentage of TGF-β and IL-10+ cells were also significantly higher in lepromatous ( p<0 . 003 , p<0 . 002 respectively ) with Mean% ± SD being 16 . 55±3 . 2 , as compared to 10 . 2±2 . 9 in tuberculoid leprosy granulomas . In conformity with the above , using qPCR ( Figure 2A ) significant increase in gene expression was observed in lepromatous as compared to tuberculoid lesions and normal skin for FOXP3 ( p<0 . 04 and p<0 . 03 respectively ) TGF-β ( p<0 . 02 ) and IL-10 ( p<0 . 01 and p<0 . 002 respectively ) . In conformity with the findings observed in the skin , antigen stimulated PMBC ( Figure 2B ) showed higher expression of FOXP3 , TGF-β and IL-10 in LL as compared to tuberculoid subjects ( p<0 . 02 ) . This was further confirmed by the increase of the cytokines in the PBMC culture supernatants by ELISA ( Table 2 p<0 . 0003 and p<0 . 02 respectively for TGF-β and IL-10 ) . Healthy contacts with long time exposure to leprosy patients also showed expression of FOXP3 and inhibitory cytokines to a lesser extent . In contrast , IFN-γ and IL-17 showed significantly higher levels in tuberculoid as compared to lepromatous leprosy ( Table 2 , p<0 . 01 and p<0 . 001 respectively ) . To further characterize the nature of FOXP3+ cells we undertook flowcytometry analysis both in leprosy and house hold contact subjects . Figure S2 shows the strategy used for validating the antibodies and manual gating used in the study . Figure 3 shows both representative and group data on CD3+ gated cells in antigen stimulated PBMC cultures . Basal unstimulated PBMC gated for CD3+CD4+ showed low but significantly higher percentage but not mean fluorescence intensity ( MFI ) of CD25+FOXP3+ cells in lepromatous ( p<0 . 003 ) as compared to other clinical groups ( Figure 4 ) with Mean % ± SD being 4 . 0%±0 . 7 and 2 . 4%±0 . 6 respectively with contacts showing 2 . 0%±0 . 5 . Antigen stimulated PBMC showed further increase in CD25+FOXP3+ cells in lepromatous subjects as compared to tuberculoid ( p<0 . 0002 ) and healthy contacts ( p<0 . 0003 ) indicative of increase in iTregs in the anergic form of leprosy . Moreover , CD25+FOXP3+ cells of basal cultures showing intracellular TGF-β was significantly higher in lepromatous as compared to tuberculoid leprosy ( p<0 . 002 ) . On antigen stimulation TGF-β+ cells increased further with lepromatous subjects , as expected showing significant increase in comparison to tuberculoid leprosy ( p<0 . 01 ) and healthy subjects ( p<0 . 0003 ) . The Mean% ± SD of TGF-β producing cells in tuberculoid , lepromatous and healthy contacts was 67 . 4±26 . 1 , 96 . 1±2 . 5 , and 41 . 2±6 . 6 respectively ( Figure 4A ) . We further analyzed the data using Mean Fluorescence Intensity ( MFI ) of FOXP3 and TGFβ in the CD25+ T cell populations . Figure 4B confirms the above observations and shows the increase in MFI in lepromatous as compared to tuberculoid subjects for both FOXP3 ( p<0 . 002 ) and TGF-β ( p<0 . 01 ) in the CD4+CD25+ T cells in antigen stimulated PBMC cultures . That the iTreg discrimination noted in the two leprosy types was driven by M . leprae antigens was indicated by PHA stimulated PBMC cultures which showed general increase but not statistically significant differences between the two leprosy types either in percentage of cells with lineage specific markers or in MFI for FOXP3 and TGF-β ( Figure S3 ) . CD8+ population of T cells with CD25+FOXP3+ were lower than the CD4+ T cells in both leprosy types and healthy contacts ( Figures 3C , and 4 ) . They were also significantly higher ( Figure 4 , p<0 . 004 ) in lepromatous ( Mean% ± SD: 3 . 6±1 . 8 ) as compared to the tuberculoid ( Mean% ± SD: 0 . 87±0 . 76 ) and healthy contact groups ( Mean% ± SD: 0 . 43±0 . 24 ) . Importantly , there was negligible intracellular TGF-β in CD8+ CD25+ , FOXP3 cells in all 3 clinical groups which may reflect on their functional state ( Figure 3C ) . Taken together the data provides evidence for increase in antigen induced iTregs in lepromatous leprosy which bear the signature markers of CD25 and FOXP3 in the CD4 lineage of T cells . We further graded the CD25 as high ( hi ) , low and negative ( neg ) in FOXP3+ CD4+ T cells of antigen stimulated PBMC cultures . Figure 5A shows representative data of one subject each of the three clinical groups . As may be seen from Figure 5B , the percentage of CD25hi was lower than the CD25low population in both leprosy types . Importantly , significant increase in lepromatous as compared to tuberculoid and healthy subjects was observed with both the CD25hi ( p<0 . 03 , p<0 . 01 respectively ) and CD25low ( p<0 . 03 , p<0 . 001 respectively ) populations of FOXP3+ cells ( Figures 5B , ) . Though CD25neg FOXP3+ cells were present in higher percentages , they did not show discrimination between lepromatous and tuberculoid leprosy patients ( Figure 5B ) . The MFI of FOXP3 also showed significant increase in lepromatous as compared to tuberculoid leprosy in both CD25hi ( p<0 . 02 ) and CD25low ( p<0 . 04 ) population . However , in general MFI of FOXP3 was lower in the CD25low populations in the leprosy groups as compared to the CD25hi population . . Moreover , CD25neg cells showed the lowest MFI in all three clinical groups ( Figure 5C ) . Significant differences in MFI were observed between the clinical groups ( p<0 . 05 , p<0 . 04 ) ) , which needs further investigation as CD25neg cells were reported to transiently express FOXP3 and be non suppressive in nature [17] , [18] .
Antigen specific T cell unresponsiveness is the hall mark of lepromatous leprosy and is thought to contribute to chronic disease and the persistence of the leprosy bacillus in the host . The discovery of FOXP3 as a molecular marker of Treg cells has renewed the interest in T cell based mechanisms which dampen effector functions . Experimental models have provided valuable insight into the role of Tregs for tolerance and mucosal immunity including gut infections [16] but their function in other human diseases is not well defined . The impact of FOXP3+ cells in host defence appears to vary with the pathogen as well as the cell manipulations used in in vivo and in vitro experimental systems [23] Thus , in some viral infections they provided protection [24] in some bacterial infections they are detrimental [25] and in some parasitic infections including malaria , they had no effect [26] Therefore , we investigated the role of CD25+FOXP3+ Treg cells in the anergic form of lepromatous leprosy . Patients and healthy contacts that had been exposed to leprosy patients were studied due to non availability of animal models that mimic the human clinical types . The present investigation provides evidence for the increase of CD4+ CD25+ FOXP3+ iTreg cells in antigen stimulated PBMC cultures of anergic lepromatous leprosy patients . Immunostaining of skin lesions showed localization of FOXP3 cells in the granulomas of both types of leprosy with a significant increase in bacilli laden lepromatous lesions . This was supported by the increase in TGF-β and IL-10 producing cells which have been associated with suppression mediated by Treg cells in experimental models [16] and human PBMC [18] . Gene expression studies by qPCR further confirmed the above findings . Our studies in leprosy are consistent with dermal leishmaniasis where T regs were related to the dynamic status of immune responses , appearing in early lesions , decreasing thereafter and reappearing in chronic lesions [27] . Differing patterns were observed in post kala azar dermal leishmaniasis lesions where patients from India [28] but not from Sudan [29] showed increased Tregs cells in the skin . To understand the nature and development of FOXP3+cells in leprosy , we combined gene expression for a wide variety of markers and transcription markers associated with Treg cells with flow cytometry for phenotypic characterization and ELISA for relevant cytokines in in vitro antigen stimulated PBMC cultures Increase in gene expression of Treg signatures , FOXP3 , TGF-β and IL-10 was observed in lepromatous as compared to tuberculoid patients . Flow cytometry analysis also showed increase in CD4+CD25+FOXP3+ cells in unstimulated basal cultures suggesting either presence of ‘natural tT reg’ or in vivo induced Tregs during the course of the disease . Further increase in in vitro cultures of antigen but not PHA stimulated PBMC of lepromatous leprosy patients indicated induction of iTregs . Both percentages of these cells as well as MFI values showed increase in lepromatous as compared to tuberculoid subjects for FOXP3 and TGFβ in the CD4+CD25+ T cell populations . Moreover ELISA showed increased levels of TGF-β ( p<0 . 0003 ) in the culture supernatants of stimulated PBMC from lepromatous as compared to tuberculoid subjects . IL-10 was also increased in the culture supernatants of lepromatous as compared to tuberculoid leprosy patients ( p<0 . 02 ) In general , increase in FOXP3+ cells in leprosy is in agreement with other studies in leprosy; however there are differences in the nature of cytokine associated with them . Increased association of TGF-β after antigen stimulation [30] as well as increase in a subset of Treg cells with IL-10 in anti-CD3 and anti-CD28 stimulated PBMC was observed in lepromatous patients [31] . Our results are in conformity with those of Palermo et al [32] wherein PBMC of Brazilian lepromatous patients were reported to show increase in CD25+ FOXP3+cells but differ in the associated inhibitory cytokine . Whereas association of IL-10 was observed by them we found iTreg association with TGFβ [32] . Varied results were also reported where increase in FOXP3+ cells was associated with tuberculoid leprosy using flow cytometry analysis [33] . Treg increase both in the blood and in pleural fluid was shown in patients with active tuberculosis , whose causative pathogen is similar to that of leprosy [34] . That TGF-β was produced by CD4+CD25+ FOXP3+ population was confirmed by flowcytometry on antigen stimulated PBMC . Intracellular TGF-β required phosphorylation of STAT5A . Though high gene expression of STAT5A was observed both in the skin and in PBMC , no significant differences were observed in the clinical groups . Importantly , phosphorylation appeared to be a pre-requisite as p-STAT5A was higher in CD25hi FOXP3+ cells of lepromatous patients and >90% of TGF-β producing cells was associated with phosphorylated STAT5A . This transcription factor after activation by IL-2 has been shown to bind to FOXP3 gene and cooperate with STAT3A for FOXP3 induction [35] . Activation of STAT5 has divergent effects on T cell subsets , leading to expansion of CD8+ memory cells and development of CD4+CD25+ regulatory T cells [36] . CD8+ lineage with CD25+FOXP3+ were also increased in lepromatous leprosy subjects but they did not show intracellular TGF-β . Differences between the two clinical types of leprosy were only observed with TGF-β producing CD4+T cells which had CD25+FOXP3+ phenotype ( p<0 . 002 ) and not with FOXP3+ cells alone or where CD25 was absent . It is of interest that FOXP3+ iTreg cells were higher in healthy contacts ( p<0 . 001 ) as compared to tuberculoid leprosy suggesting that dampening of immune responses during early stages of infection may help in protection from clinical disease akin to the dynamic state noted in dermal leishmaniasis [25] . In most experimental models and in T cell clones CD25hi cells have been incriminated for suppressive ability [16] . Though our studies have not formally established that CD25+ FOXP3+cells exerted suppression , nevertheless , they seem to be associated with the T cell anergy in leprosy as both CD25hi ( p<0 . 03 ) and CD25low ( p<0 . 004 ) FOXP3+ cells showed significant increases in the anergic lepromatous as compared to the limited form of tuberculoid leprosy . Furthermore , whereas ELISA on culture supernatants of antigen stimulated PBMC showed IFN-γ and IL-17 to be increased in tuberculoid and not in lepromatous leprosy , Treg cells did not show intracellular IFN-γ in either type of leprosy . The latter is in agreement with the down regulation of this cytokine observed by Palermo et al [32] . However , flow cytometry on antigen stimulated PBMC detected a small percentage of IL-17A producing cells in tuberculoid patients . The latter feature is of interest as we had shown recently that Th17 cells were associated more with tuberculoid leprosy and the non polarized Th0 phenotypes in both types of leprosy [10] . Further studies are required to define the relationship of Th17 and iTreg cells . Kumar et al [31] showed that TGF-β led to increased phosphorylation of SMAD3 , NFATC and facilitation of FOXP3 expression with low ubiquitination adding to the stability and suppressive potential of the Treg cells in leprosy . The peripheral population of Tregs studied by us has features of both natural ( n/tTreg ) and iTreg populations . Unstimulated basal PBMC showing CD4+CD25+FOXP3+ cells may indicate nTreg . Alternatively they may belong to the in vivo generated Tregs of an ongoing natural immune response in the untreated leprosy patients . Unstimulated ex vivo PBMC of tuberculoid subjects showed low numbers ( 1–15% ) whereas the generalized lepromatous patients showed <30% of the Tregs with intracellular TGF-β in the unstimulated PBMC . Though there is an overlap between nTreg and iTreg in certain situations with regard to this cytokine , consensus exists for its association with iTregs . On antigen stimulation this population of cells increased further in the lepromatous patients , which may be related to the expansion of pre-existing iTreg population . The stimulated PBMC in both leprosy types showed many fold increase in TGF-β . Our studies also showed some features that need further investigation . Of interest was the sequential decrease in MFI of FOXP3 in CD25hi , CD25low and CD25neg cells . It has been reported that CD25neg cells show transient expression of low FOXP3 [17] , [18] which is in agreement with our study in leprosy . Such cells were reported to be functionally non regulatory/suppressive [17] , [18] . Further studies are required to formally establish the functional nature of such FOXP3 cells in leprosy . IL-2 considered to be critical for both types of Treg cells , showed decreased expression in both types of leprosy as compared to healthy subjects in PBMC and between the two leprosy types in the skin . Furthermore , earlier studies from several groups , including ours had shown marked reduction of this cytokine in lepromatous patients [22] . It has been suggested that T cells do not suppress the initial activation of CD4+CD25neg T cells but influence inhibition by the production of IL-2 by the effector cells which results in expansion of the Tregs and subsequent suppressor function [37] . Thus the discrepancy noted in our studies may be related to time kinetics of early antigen interaction which is difficult to capture in a disease which has a long incubation period . CD28 co-stimulatory signals considered to be essential for differentiation of nTreg/iTregs and expression of FOXP3 independently of IL-2 [16] . Expression of CD28 did not discriminate the leprosy types . GATA3 shown to control FOXP3+ regulatory function in dermal and gut inflammation in murine models [38] showed a lack of association with the clinical groups in antigen stimulated PBMC . This was puzzling as GATA3 is a Th2 transcription factor and many lepromatous patients show a Th2 polarization state . In our study , GATA3 showed decrease in the dermal lesions of lepromatous leprosy as compared to normal skin . In M . tuberculosis infection , T-bet expressing Treg cells and effector cells have been shown to expand under Th1 conditions [39] whereas Tregs with GATA-3 were seen under Th2 conditions [40] . These differences may be related as indicated earlier [23] to differences between pathogens , experimental and human models of disease , sites of inflammation , as well as differences between stimulation of naïve T cells as compared to recall responses studied by us . Importantly , our studies show that the antigen specific T cell anergy and cytokine dysregulation associated with lepromatous leprosy may be linked to the increase in TGF-β producing iTreg population belonging to a suppressive lineage of T cells . The role of bacillary load ( BI ) on the evolution and maintenance of this population requires studies on lepromatous leprosy patients after negligible BI is achieved . This proves to be a logistic problem in public health as in the current regimen patients are released from treatment after 1 year when their BI is still positive since the BI reduction is at the rate of 1 log per year . Our earlier studies on lepromatous patients had incriminated monocytes/macrophage lineage and their soluble factors containing prostaglandin E2 , leukotrines and throboxanes in the inhibition of in vitro T cell proliferation in tuberculoid subjects [6] . Recent studies have indicated that prostaglandin E2 induces FOXP3 gene expression and iTreg function in human CD4+ T cells [41] , [42] which is compatible with our findings . Thus soluble factors released by bacilli laden monocytes/macrophage may play a role in inducing Treg cell function in lepromatous leprosy thereby resulting in the iTreg mediated antigen specific unresponsiveness associated with this disease .
|
Lepromatous leprosy is a generalized infectious disease caused by Mycobacterium leprae with the patients showing T cell mediated unresponsiveness to the pathogen and chronicity of lesions . The causation of unresponsiveness and anergy in this form of leprosy is not fully understood . The recent discovery of CD25+FOXP3+ cells with regulatory functions ( Tregs ) in mice and man have made it possible to study their role in the dampening of T cell responses in lepromatous leprosy . We investigated both skin and PBMC from leprosy patients for lineage specific molecular , and phenotypic markers of Tregs as well as cytokines in situ and in in vitro M . leprae stimulated PBMC cultures ( iTreg ) . Our studies find an increase in lineage specific CD4+ iTregs in lepromatous leprosy as compared to the limited form of borderline tuberculoid leprosy . Such cells secrete TGF-β , an inhibitory cytokine and may play a role in negatively regulating the T cell immune responses in lepromatous disease .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"immunopathology",
"immune",
"cells",
"t",
"cells",
"immunology",
"biology",
"microbiology",
"immune",
"response"
] |
2014
|
Increase in TGF-β Secreting CD4+CD25+ FOXP3+ T Regulatory Cells in Anergic Lepromatous Leprosy Patients
|
Wolbachia has been deployed in several countries to reduce transmission of dengue , Zika and chikungunya viruses . During releases , Wolbachia-infected females are likely to lay their eggs in local available breeding sites , which might already be colonized by local Aedes sp . mosquitoes . Therefore , there is an urgent need to estimate the deleterious effects of intra and interspecific larval competition on mosquito life history traits , especially on the duration of larval development time , larval mortality and adult size . Three different mosquito populations were used: Ae . aegypti infected with Wolbachia ( wMelBr strain ) , wild Ae . aegypti and wild Ae . albopictus . A total of 21 treatments explored intra and interspecific larval competition with varying larval densities , species proportions and food levels . Each treatment had eight replicates with two distinct food levels: 0 . 25 or 0 . 50 g of Chitosan and fallen avocado leaves . Overall , overcrowding reduced fitness correlates of the three populations . Ae . albopictus larvae presented lower larval mortality , shorter development time to adult and smaller wing sizes than Ae . aegypti . The presence of Wolbachia had a slight positive effect on larval biology , since infected individuals had higher survivorship than uninfected Ae . aegypti larvae . In all treatments , Ae . albopictus outperformed both wild Ae . aegypti and the Wolbachia-infected group in larval competition , irrespective of larval density and the amount of food resources . The major force that can slow down Wolbachia invasion is the population density of wild mosquitoes . Given that Ae . aegypti currently dominates in Rio , in comparison with Ae . albopictus frequency , additional attention must be given to the population density of Ae . aegypti during releases to increase the likelihood of Wolbachia invasion .
Infectious diseases caused by arboviruses are a growing global health concern . Among the disease vectors , mosquitoes from the genus Aedes and mostly important Ae . aegypti ( Linnaeus , 1762 ) and Ae . albopictus ( Skuse , 1894 ) have a prominent role in transmitting several arboviruses to humans . In the last 50 years , dengue virus ( DENV ) has shown a 30-fold increase in global incidence , with around 400 million new infections yearly [1–3] . In the last decade , chikungunya became pandemic after spreading from limited regions of Africa and Asia and arriving into the Americas . Two CHIKV genotypes were detected in Brazil: The Asian genotype has probably invaded the country through the Caribbean and the East-Central-South African ( ECSA ) genotype was first detected in the Bahia State [4 , 5] . Most recently , Zika virus ( ZIKV ) emerged in the Pacific and later in the Americas , causing a public health emergency due to its association with microcephaly in newborns [6–8] . The Ae . aegypti mosquito is more frequently observed in highly urbanized areas . It is extremely well adapted to live in close association with human dwellings . Females blood feed preferentially on humans and lay eggs on man-made containers often located on the surroundings of residences [9–12] . Conversely , Ae . albopictus is more frequently collected in wooded areas next to humans and tends to bite on a variety of vertebrates outdoors [13] . Both species overlap their spatial distribution in suburban areas , especially in those districts with high vegetation coverage [11 , 14–16] . Thus , eventually , Ae . aegypti and Ae . albopictus share the use of the same breeding sites , which triggers a series of ecological interactions due to the limited resources available . Many studies have investigated the negative outcomes of competing environment on adult life history traits . Inadequate nutrition during the larval stage of mosquitoes can be associated with reduced wing size , shorter longevity and flight performance , higher susceptibility to arboviral infections and replication under laboratory conditions [17–21] . Density-dependent competition in larval stages causes mortality and thus reduced recruitment to the adult stage , showing that Ae . aegypti vectorial capacity is strongly dependent on the larval habitat quality . Effective vector control activities are used as the primary approach to mitigate arbovirus transmission , especially in the absence of vaccines . Ae . aegypti control still relies massively on source reduction and on using chemicals such as insecticides . However , maintaining infestation level below a theoretical threshold to avoid outbreaks requires a constant and somehow utopic military discipline of field health agents over time [22] . Moreover , the overuse of insecticides favors the dissemination of alleles that confer resistance among wild Ae . aegypti populations , jeopardizing insecticide efficiency as tools for vector control [23 , 24] . Due to the low capacity of traditional control measures to reduce mosquito populations , new approaches to mitigate transmission must be tested . One of the innovative approaches currently being tested is the deployment of the maternally inherited endosymbiont Wolbachia pipientis into wild mosquito populations [25] . The use of Wolbachia as a natural control agent is supported by findings showing that Ae . aegypti females infected with the wMel strain are able to block DENV , CHIKV and ZIKV [26–28] . Accordingly , Wolbachia releases may be used to mitigate arbovirus transmission by two different strategies: suppression of mosquito population by massive male-releases or the substitution of a highly susceptible population by one that blocks arbovirus transmission [29 , 30] . Wolbachia deployments are taking place in five countries , including Brazil ( www . eliminatedengue . com ) . When Wolbachia-infected mosquitoes are released , females tend to behave as their wild counterparts , i . e . , will blood feed on local householders and lay eggs in the available breeding sites . Considering that in Rio de Janeiro city there is a strong co-occurrence of Ae . aegypti and Ae . albopictus [11 , 16] , females carrying Wolbachia will eventually lay eggs in containers already colonized by wild Aedes sp . mosquitoes . Therefore , our main objectives were to estimate the deleterious effects of larval competition on mosquito life history traits , but also to determine to what extent larval competition of wMel-Ae . aegypti mosquitoes with wild Ae . aegypti and Ae . albopictus may jeopardize Wolbachia invasion .
We used three different mosquito populations: Ae . aegypti infected with Wolbachia ( wMelBr strain ) , wild Ae . aegypti and wild Ae . albopictus . The lineage of Ae . aegypti with wMel was imported from Australia to Brazil ( IBAMA license 11BR005873/DF ) . Briefly , a backcrossing with 250 virgin females ( wMel ) and 200 wild males was conducted for nine consecutive generations , producing wMelBr [31] . After that period , the wMelBr colony was outcrossed every five generations with 10% wild males from a pool of four districts ( Jurujuba , Tubiacanga , Urca and Vila Valqueire ) with high nuclear genome homogeneity across Rio to refresh the genetic pool [32] . We used the F19 of wMelBr generation . Wild Ae . aegypti and Ae . albopictus mosquitoes were obtained from four districts in Rio de Janeiro city ( the same districts with which the wMelBr colony was outcrossed ) by collecting eggs laid on the wooden paddle of ovitraps . A total of thirty ovitraps were installed uniformly in each area , of approximately 1 km2 , to represent the genetic variation of the wild population . Wooden paddles were brought to Fiocruz , eggs were hatched and larvae were classified using taxonomic keys [33] . Larvae of each species were pooled and reared in dechlorinated water and fed with TetraMin ( fish food ) , maintained in a climate controlled insectary , at 26 ± 1°C and 70 ± 10% relative humidity . Adult females were kept under a 12:12 hour light:dark cycle , ad libitum access to sugar solution ( 10% ) and blood fed twice a week using anesthetized mice ( CEUA L-0007/09 ) . Eggs were stored under insectary conditions until the experiment . We used the F1 of wild Ae . aegypti and F2 of Ae . albopictus . The larval competition experiments were performed in a semi-field setting , an open building located at the Army Institute of Biology in Rio de Janeiro , Brazil ( 22°53’34”S , 43°14’33”W ) , but with limited control access to unauthorized personnel . The experiment was subject to the influences of climate variation , such as humidity and air temperature , as well as rainfall . These conditions were continuously recorded by means of a weather station ( Instrutemp , ITWH model 1080 ) installed on site . The intraspecific larval competition of Ae . aegypti ( with and without wMelBr ) and interspecific competition with Ae . albopictus were investigated by monitoring the development of larvae at different densities , species proportions , and food levels in containers . Twenty-one treatments were set and used different proportions of wild Ae . aegypti: wMelBr-Ae . aegypti: Ae . albopictus ( 20:0:0 , 40:0:0 , 60:0:0 , 0:20:0 , 0:40:0 , 0:60:0 , 0:0:20; 0:0:40 , 0:0:60 , 20:0:20 , 30:0:30 , 20:0:40 , 40:0:20 , 0:20:20 , 0:30:30 , 0:20:40 , 0:40:20 , 20:20:0 , 30:30:0 , 20:40:0 , 40:20:0 ) . The densities evaluated herein were based on that of Braks et al . ( 2004 ) [19] and represent larval crowding in nature [12 , 34] . Larvae were placed as L1 in 400 ml of tap water into black plastic containers ( 9 . 5 cm in height , 8 . 5 cm base diameter ) . Each treatment had eight containers with two distinct food levels . The food consisted of 0 . 25 or 0 . 50 g of Chitosan ( an analogue of insect chitin used to simulate the remains of arthropods ) , and fallen avocado leaves ( extra source of natural nutrition commonly used in Aedes competition assays ) , in the same proportion , that were collected , washed , dried , broken into small pieces and weighed . Therefore , our experimental design consisted of 200 plastic containers ( 12 cm in diameter x 15 cm in height ) . Each container was identified and received the appropriate quantity of Chitosan and leaf litter , with 400 ml of tap water , three days before the addition of larvae . Containers were covered with black tulle to prevent oviposition by wild mosquitoes . One hour after eggs were hatched , larvae were counted with the help of a stereo microscope and then added to their appropriate containers . Each container was monitored daily for the presence of pupae , which were collected and placed in small covered vials ( 6 . 5 cm height x 2 . 5 cm diameter ) and kept until adult emergence . On the day of emergence , adults were killed with acetyl acetate and , after being sexed , one wing was removed . Wing length was defined as the distance from the axillary incision to the apical margin excluding the fringe [35] . The experiment ended when the last pupa became adult . In treatments with the presence of wMelBr and wild Ae . aegypti simultaneously , all adults and dead pupae were screened for Wolbachia . Screening was performed using the Taqman multiplex Real Time—Polymerase Chain Reaction . Adult mosquitoes and dead pupae were individually screened on ViiA7 Real Time PCR machine ( Life Technologies ) . The genomic DNA was extracted using a squash buffer ( 0 . 1 M NaCl; 10 mM Tris Base; 1 mM EDTA; pH 8 . 2 ) supplemented with 9 μg of Proteinase K per mosquito ( Qiagen ) . After macerating the mosquitoes with a 2mm glass-bead on a Mini-beadbeater ( Biospec Products ) , samples were placed on a thermocycler following the thermal cycle: 56° C for 5 minutes and 98° C for 15 minutes . Genomic DNA was diluted 1:10 in ultra-pure water and then used as the template for Wolbachia screening . We used the WD0513 gene that amplifies a fragment of 110 bp with the following primers: TM513-Forw: CAA ATT GCT CTT GTC CTG TGG and TM513-Rev: GGG TGT TAA GCA GAG TTA CGG and TM513-probe 5’-/FAM Cy5/ TGA AAT GGA AAA ATT GGC GAG GTG TAG G -–BHQ-1/-3’ . In the same reaction , a ribosomal gene from Ae . aegypti that amplifies a fragment of 68 bp was analyzed with the primers: RPS17-Forw: RPS17-Forw: 5’- TCC GTG GTA TCT CCA TCA AGC T- 3’ and RPS17-Rev: 5’- CAC TTC CGG CAC GTA GTT GTC- 3’ , and RPS17-probe: 5’-/FAM/CAG GAG GAG GAA CGT GAG CGC AG/3BHQ_1/-3’ . Negative and positive controls of Ae . aegypti ( with and without wMelBr ) and Ae . albopictus were used in all reactions . Reagents used in the qPCR were: 5 μL of TaqMan Universal PCR Master Mix ( Thermo Fisher ) , 0 . 5mM of RPS17 primers , 0 . 6mM of TM513 primers , 0 . 1mM of RPS17 probe , 0 . 25mM of TM513 probe and 1μL of diluted DNA . Water was added to complete a final volume of 10 μL . Three biological aspects were observed throughout the experiment: larval survivorship , developmental time and wing length . Survivorship was calculated , for each container , by the frequency of larvae that reached the adult stage . Developmental time per container was calculated as the average number of days from hatching until the emergence of the adult was observed in the plastic vial . An important parameter in population ecology is the performance index λ’ , related to the growth rate r’ by λ’ = exp ( r´ ) . We calculate λ’ using values of observed biological aspects , such as survival of immature , development time and adult size of cohorts of mosquitoes , for each replicate . An estimate of the performance index has been adapted by [36] from the equation established by [37] using r’ as a measurement of population growth . According to this index , the condition λ′ > 1 . 0 represents an increase in the population , whereas condition λ′ < 1 . 0 points to a population decrease . The λ’ index was calculated for each replicate as follows: r′=ln ( 1N0∑xAxf ( w¯x ) ) D+∑xxAxf ( w¯x ) ∑xAxf ( w¯x ) , where N0 is the initial number of females in a cohort , which we assumed to be 50% of the added larvae , since the sex ratio of the species studied here is generally 1:1 [38 , 39]; Ax is the number of adult females on day x; wx is the average size of the female wing on day x; fecundity of females is modeled by a function ƒ ( wx ) of the wing size , as proposed for Ae . aegypti [40] and Ae . albopictus [41] . No significant differences in fecundity have been found due to Wolbachia infection [31] , thus we assumed the same relationship between mosquito size and fecundity for infected and uninfected mosquitoes . D is the time required ( in days ) for a newly hatched female to mate , blood feed and lay eggs . In our experiments , D is typically equal to the number of days that a female takes to reach the adult stage plus four days , the length of the first gonotrophic cycle [42] . The effects of competition conditions on the performance of Ae . aegypti infected with wMelBr were analyzed using a Generalized Linear Model ( GLM ) . Development time , wing length , survival proportion and the performance index were analyzed each separately as outcomes using as explanatory variables: species , nutrients , competing numbers of wild Ae . aegypti , Ae . albopictus , and Ae . aegypti with wMelBr . For development time , performance index , and wing length we used a normal distribution and logarithmic link function . For survivorship we used logistic regression models with a binomial family/logit link function . For each of the outcomes we selected the model with lowest Akaike Information Criterion ( AIC ) . P-values lower than 0 . 05 were considered significant . We used R 3 . 0 . 1 software for these analyzes . The index values of λ' were used to make a model to simulate the impact of different levels of infestation of Ae . aegypti wild type and Ae . albopictus in the performance of Ae . aegypti with wMelBr in larval competition . Three nonlinear regressions were applied to each of the indices λ' for the three populations: wild Ae . aegypti , Ae . albopictus and Ae . aegypti with wMelBr , with the number ( nx ) of individuals in each cohort x ( aeg for wild Ae . aegypti , albo for Ae . albopictus , wmel for Ae . aegypti with wMelBr ) in competition , according to the following model: log ( λ’ ) ~ naeg + nalbo + nwmel . These analyses allowed us to evaluate the effect on the performance index when increasing both interspecific and intraspecific competition . The values obtained in the regressions were used to simulate the interspecific competition among the three populations . Once coefficients for interspecific competition were obtained we evaluated the intensity of interspecific competition that makes the growth rate negative , i . e . , r´ = log ( λ’ ) < 0 . For instance , if population of Ae . aegypti with wMelBr suppresses the wild Ae . aegypti population , this permitted us to evaluate the frequency of Ae . albopictus that causes a severe interspecific competition that might compromise sustained growth of Ae . aegypti with wMel . In this case , we find value for r´wMel = log ( λ’wMel ) = βwmel + αwmel nalbo + γwmel nwmel < 0 , where β , αwmel , and γwmel are coefficients obtained in the regression analysis . The use of anesthetized mice to blood feed mosquitoes was authorized by Fiocruz Ethical Committee for Animal Use ( CEUA L-0007/09 ) , which follows the National guidelines for the scientific use of animals disposed on the Law 11 . 794/2008 .
Under intraspecific competition , survival was inversely proportional to larval density in the three tested populations , as expected ( Ae . albopictus: t = -27 . 2 , P<0 . 05 , Ae . aegypti: t = -28 . 3 , P<0 . 05 , Ae . aegypti with wMelBr: t = -26 . 3 , P<0 . 05 ) . Ae . albopictus presented higher tolerance for increasing competition than wild Ae . aegypti and Ae . aegypti with wMelBr . On the other hand , Ae . aegypti presented a significant decrease in survivorship when larval density per container doubled . This pattern was observed independently of Wolbachia presence ( Fig 1 , Table 1 ) . Under interspecific competition , the survival of Ae . albopictus and Ae . aegypti with wMelBr larvae was significantly higher than survival of wild Ae . aegypti , irrespective of the amount of food resources ( Table 1 ) . Nonetheless , competitive advantage of Ae . albopictus over wild Ae . aegypti seemed slightly more evident in the most stressful and competitive treatments . Ae . aegypti with wMelBr also survived less than Ae . albopictus , although its survival is marginally higher than that observed for wild Ae . aegypti . In some particular treatments , Ae . aegypti larvae infected with wMelBr presented better survival than wild Ae . aegypti . Under intraspecific competition , overcrowding was directly related to the increase in developmental time ( DT ) ( P<0 . 05 ) . Wild Ae . aegypti and those infected with wMelBr have a longer DT starting at 40 larvae per container , while Ae . albopictus DT was notably affected only at a higher density , i . e . , 60 larvae per container ( Fig 2 , Table 2 ) . Wild Ae . aegypti presented the longest DT at high densities with an average duration of 57 . 17 and 47 . 99 days in low and high food resources , respectively . Under interspecific competition , Ae . aegypti with wMelBr had a similar DT to Ae . albopictus at high food resources , but was outcompeted when food resources were scarce . Interestingly , Ae . albopictus presented a shorter DT than wild Ae . aegypti ( t = 4 . 33 , P<0 . 05 ) , but the presence of Wolbachia did not alter the DT in Ae . aegypti . The three populations had a significant decrease in wing size due to overcrowding ( Table 3 , Fig 3 ) . Ae . albopictus had a sharper decrease than wild Ae . aegypti and Ae . aegypti with wMelBr . The presence of Wolbachia did not seem to influence mosquito wing size . The amount of nutrients had a positive effect for wild Ae . aegypti and a negative effect for Ae . albopictus . Overall , overcrowding had a significant effect on the performance of the three populations ( Fig 4 , Table 4 ) . The value of λ' for wild Ae . aegypti and Ae . aegypti with wMelBr suffered a reduction from 1 . 2 to 1 . 0 when the larval density was doubled ( treatment 2 ) . On the other hand , the λ' for Ae . albopictus was only reduced to 1 . 0 when the larval density was tripled ( treatment 3 ) . Remarkably , all populations tested were able to maintain λ' above 1 under low densities , meaning that they could be successfully sustained in the wild . Under these experimental settings , Ae . albopictus showed superior performance to wild Ae . aegypti and the presence of Wolbachia did not seem to affect Ae . aegypti performance ( Fig 4 , Table 4 ) . We simulated interspecific competition among the three different populations applying the results from the nonlinear regression analyzes for each of the Ae . aegypti , Ae . albopictus , Ae . aegypti with wMel performance indices . As expected , increasing either intraspecific or interspecific competition makes the performance index smaller for the three populations . In Fig 5 , when performance indices reach values below lines for λ = 1 , the interspecific competition does not allow a positive growth rate . We generally observe that the values under which Ae . albopictus can sustain a positive growth rate are larger than values for both Ae . aegypti and Ae . aegypti with wMel populations . We also studied the frequency of Ae . albopictus that could make interspecific competition intense enough to make Ae . aegypti/wMel performance index λwmel<1 , i . e . impacting severely on the sustained growth assuming only larval competition ( Fig 6 ) . As the number of Ae . aegypti/wMel larvae increases ( intraspecific competition ) , the frequency of Ae . albopictus that causes the performance index to reach an unsustainable level λwmel<1 decreases .
The deployment of Wolbachia to reduce dengue transmission is currently being undertaken in several regions of the world . During releases , Ae . aegypti mosquitoes infected with Wolbachia will lay eggs in breeding sites in which wild mosquitoes previously colonized , i . e . , intra-specific competition with local Ae . aegypti and other native mosquitoes such as Ae . albopictus might be an important issue to determine the pace of Wolbachia invasion . Due to the co-occurrence of Ae . aegypti and Ae . albopictus in several countries of Southeast Asia and Latin America , we investigated how the intra- and interspecific competition with Ae . albopictus might undermine wMel invasion . We explored three critical aspects of mosquito biology under different competition scenarios: larval survivorship , developmental time and wing length . Using these estimates , we calculated a performance index that is related to growth rate for wild Ae . aegypti , Ae . albopictus , and Ae . aegypti with wMelBr . Overcrowding significantly reduced larval survivorship in the three populations tested , as expected [18 , 19 , 43–45] . Our data show that under intraspecific competition settings , larval survivorship decreased more intensely for Ae . aegypti than for Ae . albopictus . Also , the presence of Wolbachia did not affect this pattern ( Fig 1 ) . The effects of Ae . albopictus density on Ae . aegypti mortality and vice-versa has been evaluated elsewhere [19 , 34 , 46 , 47] . In summary , one of these populations gets more severely affected when the density of the other increases , and this increase in mortality might be seen in the larval or adult stage [48 , 49] . Ae . albopictus is frequently pointed as a better competitor than Ae . aegypti [19 , 34 , 36 , 46 , 50 , 51] as well as other species such as Ae . japonicus [52] and Culex pipiens [53] . Larvae survival under interspecific competition conditions may vary due to the difference in efficacy with which larvae exploit food resources [54] . Ae . albopictus showed a larval survivorship higher than that observed for both wild and infected Ae . aegypti . This finding strongly suggests Wolbachia has limited role in affecting larvae mortality under competitive scenarios . The wMel strain has relatively mild effects on mosquito fitness [25 , 55] , but interestingly , a superior survivorship of infected larvae was reported in a competitive environment when compared with uninfected larvae [39] . On the other hand , Wolbachia infection reduced the tolerance of Ae . aegypti larvae to starvation , probably due to an increasing rate of depletion of accumulated energy reserves [45] . Our data support an overall beneficial impact of Wolbachia infection on Ae . aegypti larval survivorship , since infected larvae present a superior survival rate than their uninfected counterparts ( the exception being observed in starvation scenarios , in which Wolbachia reduced larval survivorship ) . The development time from egg hatch to adult is a critical fitness aspect of mosquito biology under field conditions . Delayed larvae are more subject to external factors such as predation [56] , water evaporation and breeding site treatment or removal . Our results show that nutrient levels caused longer development time for all three populations when food resources were scarce [19 , 48 , 57] . Ae . albopictus had a rapid development time when compared with Ae . aegypti [11] even in the more competitive treatments . The presence of Wolbachia did not accelerate Ae . aegypti developmental time . Differences in larvae development time were highest in treatments with interspecific competition with 0 . 25 g of litter . The exception was the Ae . albopictus/Ae . aegypti with wMel , in which larvae of both populations have distinct development time differences only at 0 . 50 g litter , with the former developing faster than the latter . Overall , results regarding the influence of Wolbachia on larval development time are conflicting . The experimental design and settings established in our competitive assays were unable to detect any changes on Ae . aegypti development time due to Wolbachia presence . At intermediate ( 50 larvae/tray ) and high densities ( 250 larvae/tray ) , wMel infection led to more rapid larval development for both males and females , with no effect under a less crowded and more stressful condition [28] . Opposing results to our data , a slight delay was observed in wMel infected larvae related to their uninfected counterparts [39] . Despite these findings in disagreement , few strains of Wolbachia are known to modify adult feeding behavior , and might interfere with larval foraging capability as well [26 , 56 , 58 , 59] . Potential explanations for the effects of Wolbachia on mosquito larval development time involve immune up-regulation or increased metabolism observed in the adults , which may also influence larval development rate [26 , 60] . Other aspects still need an evaluation to better understand the effect of Wolbachia on immature development time , such as the effects of the population genetic background , Wolbachia strain and experimental design . Mosquito body size is ultimately a manifestation of larval habitat quality and can produce significant effects on an insect’s fitness and then alter mosquito vectorial capacity [12 , 61 , 62] . Physiological stress in juvenile stages produces negative effects that may pass into adulthood [63] . For instance , highly competitive environments produce mosquitoes with a small wing length , which are less likely to promote Wolbachia invasion since they should blood feed more often , possess shorter longevity and lower flight performance than bigger mosquitoes [18–21] . Hence , Ae . aegypti vectorial capacity is strongly dependent on the larval habitat quality [11 , 17–21] . Previous reports have shown an inversely proportional correlation between wing size and larval density in Ae . aegypti , as we observed [28 , 45] . Our results indicate reduction in mosquito size due to overcrowding in all three populations , which is highly expected [19 , 49] . Ae . albopictus wing size was consistently smaller than Ae . aegypti in almost all treatments , with visible differences when competition was intra or interspecific . The interaction between nutrients and population produced unexpected results . Ae . albopictus body size decreased at 0 . 5 g litter when compared with the 0 . 25 g treatment . Interestingly , body size of Ae . aegypti with wMel was not affected by availability of food resources . Hence , from the perspective of Wolbachia deployment , the infection with wMel strain does not pose a significant disadvantage during competition against wild mosquitoes [39] . We used three population growth correlates , i . e . larval survivorship , time to adulthood and adult wing size to estimate a composite index of mosquito performance ( λ' ) for each container [36 , 37] . Overall , larval density negatively affected the performance index λ' of the three populations , but remarkably only Ae . albopictus population growth was positive in all treatments . In fact , population growth of Ae . albopictus was significantly superior from the observed for wild Ae . aegypti , while the presence of Wolbachia provided no advantage to infected Ae . aegypti . Interspecific assays using Ae . aegypti and Ae . albopictus at different densities have shown a superior competitive ability of the the latter [19 , 34 , 36] . Despite being frequently described as a superior larval competitor to Ae . aegypti , these two species coexist in much of Brazil and in southeast US and Southeast Asia [16 , 64] . Part of the explanation for coexistence may rely on life-history trade-offs and abiotic factors [40 , 65–67] Coexistence between Ae . aegypti and Ae . albopictus may be possible due to dry and warm climates that would favor the former and mitigate effects of larval competition via differential mortality of Ae . albopictus [67] . This hypothesis was reinforced by Camara et al . ( 2016 ) [34] , observing that intensity of competition at the larval stage may vary seasonally , with harsh effects on development time during warmer Summer . Abiotic factors may also contribute to habitat segregation since urbanized areas tend to be warmer than arborized surrounding areas [68] . Additionally , one force that can impact Ae . aegypti/wMelBr invasion is the asymmetric reproductive interference among mosquitoes , in which male Ae . albopictus can inseminate and thus sterilize Ae . aegypti females . The act of reducing the reproductive success of a different species by mating a female of an incompatible species is called satyrization [69–71] . Evidence of satyrization of Ae . aegypti females seems to be more likely than on Ae . albopictus females , although still low ( less than 5% ) , biasing the asymmetric nature of cross matings in favor of the latter [72–74] . Therefore , although still not observed in Brazilian sites where Wolbachia deployment is ongoing , additional concern would be required if invasion is lagging . The major force that can affect Wolbachia invasion is the population density of wild mosquitoes [75 , 76] . This concern is even more important if we consider that mosquitoes from other species can lay eggs in the same breeding sites of Ae . aegypti . Therefore , during Wolbachia deployment , infected mosquitoes will lay their eggs in breeding sites already colonized by local mosquitoes , such as uninfected Ae . aegypti and Ae . albopictus . Assuming Ae . albopictus is a better competitor and the presence of Wolbachia does not increase mosquito performance at the larval stage , the natural density of Ae . albopictus may become an additional obstacle to slow invasion . However , we observed a negative growth rate of Ae . aegypti/wMelBr only when Ae . albopictus frequency was high . In Rio de Janeiro , we selected four neighborhoods with different landscapes and performed adult mosquito collections with BG-Sentinel Traps installed at the peridomestic area of local householders on a weekly basis for 104 consecutive weeks [31] . We observed that the frequency of Ae . albopictus was lower than 5% in the four sites during the 104 weeks of trapping . In fact , during approximately four consecutive months , no Ae . albopictus mosquitoes were collected in any trap from any field site ( Eliminate Dengue Program ) . Therefore , Ae . albopictus is more likely to slow down Wolbachia invasion , rather than to stop it . Density-dependent traits can promote strong effects on Wolbachia dynamics in Ae . aegypti field populations [77] . Therefore , an estimation of the population sizes of Ae . aegypti and other mosquito populations that can occasionally lay eggs in the same breeding sites , such as Ae . albopictus , Culex quinquefasciatus and Limmatus durhami , might provide important information on the Wolbachia invasion pattern in highly infested field sites [31 , 76–80] .
|
Several countries are seeking new vector control tools to reduce the transmission of arboviruses such as dengue , chikungunya and Zika . One of these innovative approaches relies on the release of Aedes aegypti mosquitoes infected with the endosymbiont Wolbachia , since this bacterium can block the aforementioned viruses and interrupt transmission . Countries in Latin America and Southeast Asia have a strong co-occurrence of Ae . aegypti and Ae . albopictus in their urban landscapes . Therefore , Wolbachia-infected Ae . aegypti mosquitoes are likely to lay their eggs in local breeding sites already colonized by wild uninfected conspecifics and/or Ae . albopictus larvae . We conducted experiments to study larval competition with varying larval densities , species proportions and food levels . Interestingly , Ae . albopictus proved to be a superior competitor under different scenarios: its larvae had superior survivorship , faster development rate and a higher performance index than Ae . aegypti ( both infected and uninfected groups ) . The presence of Wolbachia increased larval survivorship of Ae . aegypti . Our data show that the population density of wild mosquitoes , especially interspecific competition , can become an additional force to reduce the pace of Wolbachia invasion in endemic regions .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"animals",
"wolbachia",
"viruses",
"developmental",
"biology",
"animal",
"anatomy",
"population",
"biology",
"insect",
"vectors",
"zoology",
"bacteria",
"population",
"density",
"infectious",
"diseases",
"aedes",
"aegypti",
"life",
"cycles",
"arboviral",
"infections",
"wings",
"disease",
"vectors",
"insects",
"arthropoda",
"population",
"metrics",
"mosquitoes",
"eukaryota",
"arboviruses",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"viral",
"diseases",
"larvae",
"organisms"
] |
2017
|
How does competition among wild type mosquitoes influence the performance of Aedes aegypti and dissemination of Wolbachia pipientis?
|
Given the complexity of developmental networks , it is often difficult to predict the effect of genetic perturbations , even within coding genes . Regulatory factors generally have pleiotropic effects , exhibit partially redundant roles , and regulate highly interconnected pathways with ample cross-talk . Here , we delineate a logical model encompassing 48 components and 82 regulatory interactions involved in mesoderm specification during Drosophila development , thereby providing a formal integration of all available genetic information from the literature . The four main tissues derived from mesoderm correspond to alternative stable states . We demonstrate that the model can predict known mutant phenotypes and use it to systematically predict the effects of over 300 new , often non-intuitive , loss- and gain-of-function mutations , and combinations thereof . We further validated several novel predictions experimentally , thereby demonstrating the robustness of model . Logical modelling can thus contribute to formally explain and predict regulatory outcomes underlying cell fate decisions .
Functional genomic approaches ( based on microarrays and next-generation sequencing ) provide a powerful means to decipher the molecular mechanisms underlying the control of development and cell differentiation , as well as deregulations thereof associated with diseases such as cancer . Together with low-throughput experimental data , these high-throughput methods enable the delineation of large and sophisticate regulatory networks . Understanding and predicting the behaviour of such complex networks require the use of proper mathematical modelling frameworks . Various dynamical models have been proposed for a handful of relatively well known developmental processes , many using differential equations and referring to Drosophila development ( see e . g . [1–5] and references therein ) . However , these modelling studies consider relatively limited numbers of regulatory components ( at most a dozen ) and require the quantitative determination of poorly documented parameters . In this context , formal qualitative modelling approaches constitute an interesting alternative , at least as a first step towards more quantitative modelling . In particular , logical ( Boolean or multilevel ) modelling has been applied to various regulatory and signalling networks of increasing sizes over the past decade ( see e . g . [6–19] and references therein ) . But only few attempts were made to predict novel phenotypes , and therefore the full predictive value of the network and its usefulness to test hypotheses regarding novel genetic perturbations remain unclear . Here , we set out to decipher the network controlling the specification of mesoderm , one of the three germ-layers , into its four main derivatives , namely visceral muscle , heart , somatic muscle , and fat body , the primordia of which are iterated in segmentally repeated units along the anterior-posterior axis of the Drosophila embryo ( Fig 1 ) . Mesoderm specification is induced by ectodermal signals such as Decapentaplegic ( Dpp ) , which controls dorsal-ventral differentiation [20–25] , Wingless ( Wg ) , which is essential for dorsally located heart cell precursors and to the majority of somatic muscles that develop from more ventrally located cells [26–28] , and Hedgehog ( Hh ) , which specifies the visceral mesoderm dorsally and the fat body ventrally , itself characterised by the expression of Serpent ( Srp ) [29–31] . During embryonic stages 8–10 , the mesoderm is thereby progressively specified into four different tissue primordia , each of which is characterised by the expression of specific lineage transcription factors ( Fig 2 ) [27 , 32] . Collating all phenotypic data from the literature into a mathematical model allows to formally assess the coherence between the current view of the network with individual published results on single or multiple mutant phenotypes . More specifically , we aim to further characterise the crucial regulatory components and interactions driving mesoderm specification . As we mostly rely on published qualitative molecular and genetic data , we use a flexible logical modelling framework and the software GINsim ( cf . Material and methods ) , which enables the use of multilevel variables whenever justified , along with fully asynchronous updating . Systematic simulations of the resulting logical model were then performed to ( i ) assess the coherence and comprehensiveness of our representation of the underlying network , ( ii ) identify gaps in the current understanding and characterisation of mesoderm specification , and ( iii ) ultimately predict phenotypic outcomes of novel genetic perturbations . We demonstrate that the resulting logical model can recapitulate all known mutant phenotypes , therefore indicating that this formal representation of the network is sufficient and coherent to explain mesoderm cell fate decisions . By running simulations on over 300 genetic mutation combinations ( many of which are double mutants with non-intuitive outcomes ) , the model could predict the phenotypic outcome for each novel mutant background , at least in terms of gene expression patterns , thereby providing new testable hypothesis that we experimentally confirmed . This approach thus provides developmental biologists with a very useful tool kit to test novel hypotheses , which are often very difficult to carry out experimentally . Moreover , the model provides novel insights into the underlying regulatory network driving these cell fate decisions .
To initiate this study , we performed an extensive analysis of all reported genetic and molecular data in the literature to identify the main regulatory components involved in Drosophila mesoderm specification , along with the known interactions between them . Indeed , dozens of articles extensively cover the genetic bases of the sub-division of Drosophila mesoderm ( this is evident by the bibliographical entries linked to key regulatory components in the model file and model documentation provided as S1 Text ) . Cis-regulatory information is sometimes available , enabling us to infer direct interactions and epistatic relations . In particular , we relied on recent ChIP data reporting the in vivo occupancy of six key mesoderm transcription factors ( Bagpipe ( Bap ) , Biniou ( Bin ) , Dorsocross 1 , 2 and 3 ( Doc ) , Myocyte Enhancer Factor 2 ( Mef2 ) , Tinman ( Tin ) and Twist ( Twi ) ) [33–35] to assess direct interactions inferred from genetic experiments . Encoded using the software GINsim ( Computational and experimental procedures ) , the resulting regulatory graph ( Fig 3 ) is provided in a computer readable format , along with extensive annotations ( text and links to relevant literature and database entries , see S1 File ) . This regulatory graph encompasses 48 nodes ( including 12 input components , representing mainly ectodermal signals ) and 82 regulatory interactions . In many cases , the definition of the logical rule associated with each node is straightforward ( e . g . when a node is the target of a unique regulator ) . However , for more complex regulatory relationships , i . e . when multiple interactions converge on the same component , we examine the following scenarios: ( i ) Is the presence of an inhibitor sufficient to completely or partially block gene expression ? ( ii ) Which activator ( s ) is ( are ) sufficient to drive the expression of the target gene ? ( iii ) Can the activators do so in the presence of repressor ( s ) ? After several iterations , we obtained a set of logical rules consistent with all available knowledge on the regulation of each gene in the network , which further enabled the recapitulation of all published phenotypes ( see below ) , demonstrating the robustness of the model . Before attempting to simulate the specification of the mesoderm into its four main presumptive tissues ( visceral muscle ( VM ) , heart ( H ) , somatic muscle ( SM ) , and the fat body ( FB ) , we needed to specify the patterns of gene expression expected as a result of wild type development . Based on published data ( mainly in-situ hybridization or immunostaining assays ) , we have derived the qualitative levels of expression of the 48 network components in each of the four presumptive territories ( VM , H , SM and FB ) from the literature ( S1 Fig ) . Only subsets of these components are crucial in the specification of each of the four tissue subtypes . These tissue markers can be readily identified based on the phenotypes reported in loss-of-function mutant embryos , leading to severe defects in tissue formation , or following ectopic expression , often leading to specific tissue expansion . Embryos lacking Tin , Bap or Bin , for example , do not develop VM . Moreover , tin mutant embryos fail to develop H cells , and have severe defects in all tissues derived from the dorsal mesoderm [32 , 36 , 37] . Overall , ten network components play such dramatic roles in specific tissues ( emphasised by bold contours for the corresponding coloured cells in the S1 Fig ) . Note that the ventral mesoderm territory that gives rise to both SM and FB is subdivided into regions that have low ( yet significant , hence the use of the value 1 ) and high Twi expression . Indeed , the inhibition of Notch ( N ) combined with the presence of Wg and Daughterless ( DA ) activates Twi at a higher level ( maximal level , i . e . value 2 ) , thereby delimiting a region of high Twi expression [38 , 39] . We systematically searched for relevant information and refined the logical rules until model behaviour was found fully consistent with all published data . To ease simulations , our regulatory graph was reduced by hiding intermediate signalling components ( components in grey in Fig 3 , see also Material and methods ) . Provided that we do not delete any regulatory circuit , the resulting reduced model preserves the stable states of the system , which represent the different specification states ( i . e . mesoderm derivatives for wild-type or mutant situations ) . To perform simulations , the initial values for each component must be specified , in particular for the signalling input components coming from the ectoderm . For each of the four presumptive tissue territories , we thus have a specific input combination ( S1 Fig , left ) . For the sake of simplicity , we set all internal nodes to zero for each wild-type initial state , with the notable exception of Twi , which was set to the value 1 . For each territory , the target values at stage 10 were evaluated based on published data ( S1 Fig , right ) . For example , in the region that will form VM , the initial state ( stage 8 ) is characterised by the presence of Twi , which activates Tin [40] and Mef2 [41] expression . Bap is then activated by Tin at stage 9 [37] , which is followed by the activation of Bin by Bap [42 , 43] in late stage 9 embryos . Finally ( stage 10 ) , Bap is activated at its maximum level ( value 3 ) by Cubitus Interruptus ( Ci ) and Engrailed ( En ) [27] . To recapitulate the formation of each mesodermal tissue derivative in the wild-type situation , we thus ran four different simulations using an asynchronous updating policy ( Material and methods ) . A detailed comparison of simulation results with experimental data led us to refine the logical rules , and sometimes even to consider additional regulators , until we converged on the regulatory graph shown in Fig 3 , along with the rules listed in S1 Table ( see also the S2 Text for more information about the delineation of the logical rules associated with Tin and Bap ) . Our final model qualitatively recapitulates all aspects of the major events in the specification of the four main domains of the mesoderm , from stage 8 to stage 10 ( S2 Fig ) . In parallel , we also simulated the effects of genetic perturbations reported in the literature , the results of which led to some model adjustments . Iterating this procedure for all known mesodermal mutants led to a model that is robust and consistent with all relevant published data . The simulated phenotypes resulting from seven selected genetic perturbations are illustrated in Fig 4 . For example , the simulation of a wg loss-of- function ( lof ) gives rise to a loss of cardiac tissue , as observed experimentally [27 , 44] , while dpp lof gives rise to an extension of FB at the expenses of VM , mirroring previously reported experimental data [30 , 32] . We can also simulate more complex genetic backgrounds . For example , a double gof of dpp and hh combined with a lof of wg leads to an expansion of VM in the entire mesoderm [27] . To date , our simulations recapitulate all mutants reported in the literature . Although expected , since this literature information was used as input to generate the model , these results demonstrate the coherence of the model , which was based on disjoint information generated from published studies from different labs , mostly based on single mutant phenotypes , with only a small number of documented multiple genetic perturbations . A study focused on heart or VM development , for example , often will not have examined markers for FB , yet the model can simulate the phenotypes in all four mesodermal domains . Given the accuracy of our model to recapitulate all known published phenotypes , we reasoned that the model provides a very useful tool to perform systematic novel in silico perturbations at large scale . In fact , some of the results obtained with the simulations described above already correspond to new predictions , as biologists typically check only subsets of markers for each mutant studied ( see in particular the tissue domains shown in yellow in Fig 4 , which correspond to situations that have not been fully analysed experimentally , and for which we obtain combinations of markers associated with different tissues ) . Nevertheless , our aim here is to go beyond this and perform a more systematic assessment of the effects of combinations of two perturbations affecting different pathways and/or tissue markers . Single and multiple mutants can be readily defined using GINsim ( cf . Material and methods ) , while they can often be very difficult , and sometimes impossible , to generate experimentally . To this end , we simulated the effects of single or double perturbations in each of the four tissue domains . The interpretation of the results generated is not trivial . To assess these results more efficiently , we used the expression of the key lineage transcription factors for each tissue as defining signature: for VM , expression of Tin ( level 1 ) , Bap ( level 2 ) , and Bin; for H , expression of Tin ( level 2 ) , Doc , and Pannier ( Pnr ) ; for FB , expression of Srp; and for SM , expression of Twi ( level 2 ) , Pox meso ( Poxm ) , DSix4 and Zfh1 . These tissue signatures were then matched against the stable states reached during model simulations , thereby automating the interpretation of the resulting phenotypes . Practical considerations ( mutant strain availability ) led us to consider fourteen components ( Twi , Tin , Bap , Bin , Ci , Doc , Pan , Pnr , Slp , Srp , Mef2 , Mad , Med , Nicd ) for systematic single and pairwise combinations of loss- and gain-of function in silico perturbations . The results of the 338 mutant simulations performed are displayed in a matrix form ( Fig 5 ) and can be browsed in a convenient searchable web archive ( S2 File ) . This format enables an easy comparison of the effects of different perturbations , which facilitates the detection of dominant or synergic effects of different perturbations . For example , slp lof generally shows a loss of H tissue , a result similar to that obtained for wg lof [26 , 27] . Although some of these mutants have been partly documented experimentally , most of the double perturbations listed in Fig 5 have not been fully experimentally assessed in all four-tissue domains . To demonstrate the usefulness and accuracy of these predictions , we experimentally tested six genetic perturbations ( two double mutants and the associated four single mutants ) , examining the effects within all four tissue domains ( Fig 6 and S3 Fig ) . Model predictions for each of these mutants are highlighted in Fig 5 . To examine the phenotype of each tissue , Tin , Srp , Glycogen phosphorylase ( GlyP ) and Bin were used as markers for the development of H , FB , SM and VM , respectively . We first assessed our predictions for lof mutants of Medea ( Med ) and Sloppy-paired ( Slp ) , and the double mutant . Medea is directly required for the induction of tinman ( Tin ) by Dpp via the tin-D dorsal mesoderm enhancer [23] . Once expressed , Tin and Med have a direct protein-protein interaction that is required for dorsal mesoderm specification [28] . As the heart specification requires both activation by Med-Tin and repression of the VM within the heart domain by Slp , we were interested to examine if loss of Med and Slp would completely abolished cardiac mesoderm specification and be sufficient to extend the VM territory . For Med lof , our model simulations predict a loss of H , which is indeed what we observed experimentally ( Fig 6A ) . Although the expression status of some genes within the VM region is changed in the mutant , the VM develops largely unperturbed , as predicted . The simulation of slp lof also results in a loss of H , and two stable states within the SM , one leading to normal SM development , while the other state lacks some marker expression , and therefore should perturb SM development . When we examine slp lof mutant experimentally , we observe the predicted loss of H , while SM appears largely normal , indicating that the corresponding stable state is the correct outcome . Simulations of the double lof mutant give the combined phenotype of the two single mutants , which again qualitatively fits with experimental data , with the H even more severely affected in the double mutant ( Fig 6A , in situ for tin expression ) . In contrast , the VM develops largely unperturbed , indicating that loss of heart , even the severe disruption seen in the double mutant , is not sufficient to lead to expansion of VM , in this genetic background . We next tested a combination of two gof conditions , where it is not a priori obvious what the phenotypic consequence would be within the FB or SM domains . Slp is normally expressed in the H region , where it inhibits VM development through the direct inhibition of Bap expression [32 , 45] . Doc is expressed within the Heart domain ( segmentally repeated patches of cells within the dorsal mesoderm ) at stage 10 , where it is essential for heart development [46 , 47] . For a gof of Doc , our simulations predicts normal H development , with minor perturbations of VM , FB and SM . Our experimental results largely confirm these predictions , with very minor perturbations on the development of each tissue ( based on the expression of the corresponding tissue markers ) , despite the ubiquitous expression of Doc ( Fig 6B ) . The simulation of a gof of Slp predicts a severe perturbation of VM ( yellow cell ) , characterised by the lack of expression of the key cell markers Bap ( level 1 instead of 3 ) and Bin , as expected [32 , 45] , along with a potential perturbation of FB ( obtention of two stable states , both with Srp expression ) . When the two gof genotypes are combined , our model predicts normal H and SM specification , but a loss of VM and FB , which is exactly what we observe experimentally , as seen in the in situ shown in Fig 6B . These results therefore demonstrate that our qualitative model can correctly predict the interaction between two gof causing a severe loss of FB . The expansion of heart cells can be further explained by the ectopic expression of heart markers in our simulations .
The logical model presented here integrates all major genetic processes underlying the formation of four tissues during Drosophila mesoderm specification . The model is based on the integration of extensive analysis of in vivo experimental data , especially genetic data ( patterns of gene expression and mutant phenotypes ) , partly confirmed by functional genomic data ( ChIP data for transcription factor occupancy ) . These data were translated mathematically in terms of a regulatory graph and logical rules . The simulation of our model qualitatively recapitulates the expression of the main lineage markers of each region from developmental stage 8 to 10 , for the wild type case , as well as for over twenty reported mutant genotypes . This study is the first attempt to model the regulatory network controlling the specification of mesoderm during Drosophila development , and more broadly represents one of the most comprehensive developmental networks that have been modelled to date . Mesoderm specification has been extensively studied in many species , including the sea urchin [48] . Recently , the Davidson group developed a Boolean model that recapitulates the specification of the sea urchin endo-mesoderm in the wild-type case , as well as experimental data for three genetic perturbations [49] . The approach of Davidson's group converge with ours in the delineation of a reference network with reliable annotations , which then serve as a scaffold to define logical rules and perform simulations . Both approaches implement the crucial components and interactions , along with the dynamical unfolding of the corresponding developmental network in an intuitive manner . Importantly , we demonstrate that we can not only recapitulate the known mutant phenotypes , but also predict various novel phenotypes . In the case of our study , several regulatory mechanisms were simplified , in particular regarding the signalling pathways involved . We have developed more complete models of most Drosophila signalling pathways [50] , but we retained simpler implementations of these pathways to keep our mesoderm specification model computationally tractable . A limitation of this study resides in the poor documentation of specific markers associated with each type of embryonic domain . In particular , our marker set is limited to Srp in the case of FB . Presumably , others regulatory factors must be implicated in the specification of this tissue , which remain to be discovered . This lack of information complicates the interpretation of mutant phenotypes . For example , it is known that Bap lof leads to the loss of VM , but we miss information about effects on other tissues . Although Bap is crucial for VM development , it is also expressed at later stages in H . At this point , we assume that H , SM and FB develop normally in Bap lof mutant , as no other experimental defect has been reported . Finally , Boolean models of embryonic processes generally rely on qualitative expression data from in-situ hybridisation . Our discrete model ( as the sea urchin model [49] ) is therefore limited to qualitative results , such as the presence or absence of a given tissue in a given presumptive territory . Although we cannot reproduce quantitative data , such as an increase or a decrease of specific cell numbers , we can still recapitulate the presence of different cell types . Our logical model could further serve as a scaffold to build more quantitative models when more quantitative and systematic experimental datasets will become available . For now , the advantage of logical modelling is that models can be easily abstracted at a level subsuming missing data , which is less straightforward for more quantitative modelling frameworks , such as differential or stochastic equations . Given the complexity of embryonic development , the shear number of parameters involved and the high inter-connected nature of regulatory networks , logical modelling offers an accurate solution that can be applied to many systems with the amount of data that is available today .
We use the multilevel logical formalism , originally proposed by René Thomas [51] , which has already been used to model various networks involved in the control of cell differentiation or proliferation ( see e . g . [6 , 8 , 9 , 11 , 12 , 19] . In short , both the structure of a logical model and its dynamics are represented in terms of graphs ( in the sense of the graph theory ) , called regulatory graphs and state transition graphs , which are briefly described hereafter . In a regulatory graph , the vertices ( or nodes ) represent regulatory genes or products ( transcription factors , kinases , etc . ) . In many cases , these regulatory components can be satisfactorily represented by Boolean variables , which can take only two values , 0 or 1 , corresponding to the absence or presence of the component , respectively . However , in some situations ( e . g . the consideration of a morphogen ) , more qualitatively different levels may be required . The arcs ( or arrows ) connecting pairs of vertices represent regulatory interactions between components ( e . g . transcriptional activations or inhibitions , phosphorylation , etc . ) . These arcs are usually associated with a plus ( + ) or minus ( - ) sign , denoting an activation or inhibition effect of the source node onto the target node , respectively . When the source of an arc is associated with a multilevel variable , a threshold ( i . e . minimal level ) must be specified . To complete this model description , logical rules ( or logical parameters ) are further defined to indicate how each component reacts to different combinations of regulatory interactions ( S2 Text and S1 Table ) . The simulation of a logical model can be represented by a state transition graph ( STG ) , whose vertices represent logical states ( i . e . a vector encompassing values for all components ) , whereas arcs represent transitions between states enabled by the corresponding regulatory graph and logical rules . In this work , we use an asynchronous updating mode , meaning that we consider all possible unitary transitions ( affecting only one variable at a time ) whenever there is a call to change some component value ( s ) at a given state . One recurrent problem with logical simulations ( in particular when using asynchronous updating ) is the potential combinatory explosion of the STG when dealing with large regulatory graphs . Consequently , it is often difficult to generate and analyse the STG for complex networks encompassing several dozens of components . However , using proper algorithms and software tools ( see below ) , it is possible to characterise the asymptotical behaviour of the systems , which is of special interest for us here . Indeed , attractors , especially stable states ( states with no successor ) , are usually associated with specific differentiated states . Logical models provide a realistic description of cellular events , as they are capable of reproducing time dependent processes in a qualitative manner ( i . e . focusing on the sequential order of transitions ) . The software GINsim ( for Gene Interaction Network simulation ) implements the logical formalism [52] . It allows the edition , analysis and simulation of regulatory graphs . Freely available ( http://ginsim . org ) , GINsim supports the annotation of components and interactions with free text and URLs . Once a model is defined , the user can select a simulation mode and define a set of initial states . GINsim can then be used to compute state transition graphs and report the stable states . GINsim also enables the definition and the simulation of different types of mutants ( loss-of-function , ectopic gene expression , and combinations thereof ) by blocking the levels of expression of the corresponding variables in defined intervals . To further ease the analysis of multiple perturbations , we have written a set of scripts in python , which iteratively compute the behaviour of our mesoderm specification model for each region and mutant considered , process the results and generate a synthetic web page ( cf . Results and S2 File ) . To enable the dynamical analysis of comprehensive regulatory graphs , we take advantage of a novel reduction method implemented in GINsim . This functionality allows the user to select components of a regulatory graph to be made implicit . The software verifies that the proposed reduction does not fundamentally change the network topology ( elimination of regulatory circuits ) and update the logical rules for the components targeted by reduced nodes . The original and reduced networks have the same stable states ( in terms of levels of common variables ) , while differences may appear as to their reachability [53] . The following Drosophila lines were used: UAS-Slp and UAS-Doc lines were kindly provided by M . Frasch ( Doc line C2 [46] ) . We crossed both stocks with a marked double balancer to generate the homozygous stock x/y;UAS-Doc;UAS-Slp . Males from the UAS-Doc , UAS-Slp and UAS-Doc;UAS-Slp lines were crossed with females carrying a homozygous twist-GAL4 driver , kindly provided by Maria Leptin . Slp1 and Med1e loss-of-function mutations were obtained from the Bloomington stock centre ( stock numbers 5349 and 9033 ) , and crossed together to make the double loss-of-function stock , which were placed over lacZ-marked balancers . Embryos were collected using standard procedures . Fluorescent in situ hybridisation was performed as described previously [54] . The following ESTs were used to generate anti-sense probes: RE01329 ( tin ) , SD07261 ( srp ) , and LD24485 ( Glyp ) , while a full length cDNA was used for bin ( gift from M . Frasch ) and lacZ . The probes were detected with peroxidase-conjugated antibodies ( Roche ) and developed using the TSA system ( Perkin Elmer ) . slp and Med mutant embryos were unambiguously identified based on the absence of lacZ expression from the balancer chromosome .
|
We delineate a logical model encompassing 48 components and 82 regulatory interactions controlling mesoderm specification during Drosophila development , thereby integrating all major genetic processes underlying the formation of four mesodermal tissues . The model is based on in vivo genetic data , partly confirmed by functional genomic data . Model simulations qualitatively recapitulate the expression of the main lineage markers of each mesodermal derivative , from developmental stage 8 to 10 , for the wild type case , as well as for over twenty reported mutant genotypes . We further use this model to systematically predict the effects of over 300 loss- and gain-of-function mutations , and combinations thereof . By generating specific mutant combinations , we validated several novel predictions experimentally demonstrating the robustness of model . This modelling study is the first to tackle the regulatory network controlling the specification of mesoderm during Drosophila development , and more broadly deals with one of the most comprehensive developmental networks that have been modelled to date .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infographics",
"invertebrates",
"gene",
"regulation",
"animals",
"simulation",
"and",
"modeling",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"embryos",
"drosophila",
"research",
"and",
"analysis",
"methods",
"embryology",
"computer",
"and",
"information",
"sciences",
"gene",
"expression",
"mesoderm",
"insects",
"arthropoda",
"data",
"visualization",
"phenotypes",
"graphs",
"genetics",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2016
|
Qualitative Dynamical Modelling Can Formally Explain Mesoderm Specification and Predict Novel Developmental Phenotypes
|
One view of adaptation is that it proceeds by the slow and steady accumulation of beneficial mutations with small effects . It is difficult to test this model , since in most cases the genetic basis of adaptation can only be studied a posteriori with traits that have evolved for a long period of time through an unknown sequence of steps . In this paper , we show how ace-1 , a gene involved in resistance to organophosphorous insecticide in the mosquito Culex pipiens , has evolved during 40 years of an insecticide control program . Initially , a major resistance allele with strong deleterious side effects spread through the population . Later , a duplication combining a susceptible and a resistance ace-1 allele began to spread but did not replace the original resistance allele , as it is sublethal when homozygous . Last , a second duplication , ( also sublethal when homozygous ) began to spread because heterozygotes for the two duplications do not exhibit deleterious pleiotropic effects . Double overdominance now maintains these four alleles across treated and nontreated areas . Thus , ace-1 evolution does not proceed via the steady accumulation of beneficial mutations . Instead , resistance evolution has been an erratic combination of mutation , positive selection , and the rearrangement of existing variation leading to complex genetic architecture .
Adaptation is often envisioned as a slow and regular improvement , a view embodied by Fisher's geometrical model of adaptation , whereby mutations fix if they bring the current phenotype closer to an optimum [1 , 2] . However , the trajectory towards the optimum can take a truly tortuous path [3] , as adaptation uses almost any allele that brings it closer to the optimum , regardless of its negative side effects . “Evolution is a tinkerer , ” as emphasized by Jacob [4] , may indeed be more than a pretty metaphor . That beneficial mutations often have deleterious pleiotropic effects is well established ( i . e . , they generate a fitness cost [1 , 5–8] ) . Pleiotropy causes an “evolutionary inertia” [9] whereby beneficial mutations often only ameliorate the side effects of the last beneficial mutation . This process of “amelioration” [10] can follow a scenario a la Fisher [11] with modifiers or compensatory mutations occurring at different loci , or with allele replacement at the same locus ( Haldane [12] ) . Both cases have been reported ( e . g . , [10 , 13–16] ) . However , it is perhaps less often appreciated that pleiotropy can have more dramatic consequences . First , the pleiotropic effects of beneficial mutations may be more complex than simply deleterious . Second , they may trigger the evolution of the genetic architecture and gene number . This paper illustrates these two aspects with the tortuous path taken during the evolution of insecticide resistance in Culex pipiens mosquitoes in southern France . In natural populations of the mosquito Culex pipiens , various resistance genes have been selected over the course of ∼40 y of control using organophosphorous ( OP ) insecticides ( see [17] for a review ) . OP insecticides kill by inhibiting acetylcholinesterase ( AChE1 ) in the central nervous system . The genetic basis of OP resistance involves two main loci , the super-locus Ester and the locus ace-1 , both of which have major resistance alleles . Strong direct and indirect evidence indicate that the resistance alleles have a “fitness cost”; i . e . , they are selected against in the nontreated areas ( see reviews in [17 , 18] ) . At the Ester locus , the different resistance alleles are not identical and one has slowly replaced the other [15 , 16] . At the ace-1 locus , the resistance allele present worldwide , ace-1R , displays a single amino acid mutation ( G119S ) , changing the glycine at position 119 into a serine . This large effect point mutation confers high resistance towards OP insecticides due to lower affinity and has arisen independently in several mosquito species [19–21] . However , this mutation also exhibits a strong deleterious side effect: G119S is less efficient at degrading acetylcholine ( ACh ) than the susceptible variant of AChE1 . Overall , G119S causes a more than 60% reduction in enzymatic activity in the absence of insecticide [22] , which alters the optimal functioning of cholinergic synapses of the central nervous system and probably causes the various developmental and behavioral problems that have been identified in individuals carrying ace-1R [23–25] . Field surveys also quantified the overall fitness cost of this allele in the wild . Fitness values are 1 and 0 . 89 for a susceptible and resistant homozygote , respectively , in the absence of insecticide ( i . e . , when only the cost applies [26 , 27] ) . However , in a treated habitat , only ace-1R-carrying individuals survive . A few years after the appearance and spread of ace-1R , several duplications of the ace-1 gene appeared in wild populations , each involving a resistant and a susceptible copy of the ace-1 gene on the same chromosome [28–30] . We refer to these new duplicated haplotypes as ace-1D . The ace-1 copies involved in these duplicated haplotypes are barely distinguishable from local single-copy alleles . In particular , within a population , the sequence of the local ace-1D resistant copy is always identical to the sequence of the single-copy ace-1R allele [30] . In the Montpellier area of Southern France , the appearance of a duplicated haplotype was indirectly inferred and traced back to at least 1993 [29] . More recently , sequence data of ace-1 confirmed the presence of duplication , but surprisingly also revealed the presence of two distinct duplicated haplotypes in Montpellier area ( ace-1D2 and ace-1D3 ) [30] . Other independent duplications were also found in the Caribbean ( Martinique , ace-1D1 ) and Philippines ( Palawan , ace-1D4 ) [30] . Gene duplication is an important type of mutation because it fosters the evolution of new functions [12 , 31–33] . In the case of ace-1 , a strong genetic constraint drives resistance evolution , as the degree of resistance and the ability to degrade ACh trade off; OP and ACh molecules compete for the same active site of AChE1 . Duplication may be a way to disentangle the two functions , i . e . , by improving synapse signalling and mosquito's fitness while maintaining resistance . Thus , our working hypothesis is that the spread of ace-1 duplications is driven by the increased AChE1 activity in individuals carrying ace-1D compared to those carrying ace-1R . In the last survey eight y ago [29] , indirect estimation indicated a rapid initial increase of ace-1D , suggesting a strong selective advantage of ace-1D over ace-1R ( in the range of 3%–6% of ace-1R fitness [29] ) . Based on these previous findings , we expected the duplicated haplotype to quickly replace ace-1R . The aim of this study was to test this prediction and to understand more precisely the fitness relationship between the various genotypes ( single resistance or susceptible alleles , duplicated haplotypes and their heterozygotes ) in treated and nontreated areas . Surprisingly , we show that ace-1D did not replace ace-1R . Several experiments indicated that duplications , while favored when heterozygous , exhibit very strong deleterious effects when homozygous . Our work thus highlights a novel scenario of adaptive evolution and trade-offs that hinge on the genetic architecture underlying the expression of resistance variation .
In the following , and according to the nomenclature used in Labbé et al [30] , the duplicated haplotypes ace-1D2 and ace-1D3 will be denoted D2 and D3 , susceptible copies being denoted D2 ( S ) and D3 ( S ) , and resistant copies , D2 ( R ) and D3 ( R ) , respectively . D refers to either D2 or D3 . The single copy resistant allele ( ace-1R ) will be denoted as R . Finally , single susceptible alleles ( ace-1S ) will be denoted S . The only S allele present in the strains used in laboratory experiments and originating from the susceptible reference strain SLAB [34] will be denoted SSLAB . We first analyzed the frequency variation of R and D ( = D2 + D3 ) across treated and nontreated areas around Montpellier from 1986 to 2002 , using a purely descriptive model ( see Methods ) . As in previous analyses [26 , 27 , 29] , we found that their frequency showed clinal variation: R and D are more frequent in the treated than the nontreated area ( Table 1 ) . The straightforward interpretation of these clines is that R and , to a lesser extent , D have a selective advantage in the presence of insecticide over S , but that they have a fitness cost in its absence . Migration leads to the smooth decline of frequency in the nontreated area and prevents fixation of resistance in the treated area . We also confirm that D was rare until around 1993 , when it started to rapidly increase in frequency . Surprisingly , D frequency stopped increasing around 1996 and reached a plateau at about 22% ( this holds when only summers clines are considered , analysis not shown ) ( Figure 1; Table 2 ) ; . However , this analysis does not distinguish D2 and D3 duplications . To assess the frequency and the distribution of D2 and D3 along the cline , five populations were sampled in 2005 , on the same transect as for cline analysis , two in the treated area ( Maurin and Lattes ) , one at the limit between the two areas ( Distill ) , and two in the nontreated area ( Viols and Ganges , Figure 2 ) . Molecular tests available allow detection of D2 and D3 only when SSLAB is the only S allele present . Resistant females from these samples were thus mated with SLAB-TC males ( S/S ) and their offspring analyzed ( see Methods ) . , Despite the large crossing ( 700 females from each sample ) , the number of egg rafts was very low ( this is usually observed for the C . p . pipiens subspecies , P . Labbé and M . Weill , personal observation ) , and some females were able to lay eggs several times , so that no reliable frequencies could be estimated . Nevertheless , molecular analyses revealed that both duplications are now segregating in these populations ( Table 3 ) . This plateau frequency pattern strongly suggests that equilibrium has been reached among the four alleles segregating in these populations , S , R , D2 , and D3 . The next step in the analysis was therefore to infer , from the clinal patterns observed in the last years ( 1999–2002 ) , the different genotypic fitnesses ( in treated and nontreated areas ) that could correspond to this equilibrium . With four alleles , there are ten diploid genotypes and therefore nine relative fitnesses to estimate in each habitat . Since we do not have access to the respective frequencies of D2 and D3 , some of these parameters cannot be estimated separately ( see Methods ) . We report fitness estimates that assume that the two duplications have identical fitness effects . We developed a maximum-likelihood analysis by combining exact simulations with an optimization routine . We used previous estimation for migration: C . pipiens displays a migration of 6 . 6 km . generation−1/2 [26] . This analysis indicated that the stable cline we observe could be expected at equilibrium between migration and selection . To achieve this equilibrium , the different genotypes should display relative fitness , as reported in Table 4 . Since the equilibrium situation seems to have been reached around 1995–1996 , it therefore suggests that both duplications spread before that date . Importantly , this equilibrium requires a situation of double overdominance: heterozygotes involving duplication , i . e . , ( D2/S ) , ( D3/S ) , ( D2/D3 ) , ( D3/R ) , and ( D2/R ) , must have a higher fitness than R , D2 , and D3 homozygotes in absence of insecticide . It also requires that R and S homozygotes are the best homozygote genotypes in the treated and nontreated area , respectively . With such a fitness scheme , each duplication could be maintained independently even if they have identical fitness effects . In addition , the presence of both allows them to reach a higher total frequency . More precisely , the overall D frequency when both are present should be intermediate between the frequency they could reach alone and twice this frequency , due to the overdominance of the ( D2/D3 ) genotype . The last step in the analysis was to confirm experimentally this possible pattern of double overdominance . To do so , we followed the relative success , over a single generation in the laboratory , of different pairs of genotypes in the absence of insecticide . In a first set of experiments , we showed that ( D2/S ) and ( D3/S ) largely outperformed ( D2/D2 ) and ( D3/D3 ) in terms of survival , development time , and fertility ( Figure 3; Tables 5 , 6 and 7 ) . Homozygotes for both duplications indeed show an extremely low fitness with high mortality at pupation and emergence along with low fertility , such that it was nearly impossible to fix a strain for these duplications . By contrast , we showed that heterozygotes involving the two duplications ( D2/D3 ) were as fit as ( D3/S ) ( Figure 3; Table 5 ) . Finally , we compared the fitness of ( D3/S ) and ( D3/R ) with R homozygotes ( Figure 4; Table 5 ) . Here again , we found that the heterozygotes outperformed the homozygotes , although less strikingly . We were not able to perform this last comparison with the D2 duplication because the survival and fertility of ( D2/D2 ) homozygotes are too low to maintain a laboratory strain fixed for this duplication . Overall , these experiments directly confirm the fitness relationship deduced from the clinal pattern observed in the field . Homozygotes for either D2 or D3 have a severely reduced fitness , but heterozygotes for the two duplications , ( D2/S ) , ( D3/S ) , ( D2/D3 ) , and ( D3/R ) perform well .
This study shows how ace-1 , one of the two major genes involved in resistance to OP insecticide in the mosquito Culex pipiens , evolved in the last 40 y of control using insecticides . The evolution of resistance to insecticide in C . pipiens does not follow a classical scenario whereby a beneficial mutation with deleterious side effects spreads and is followed by a steady “amelioration” process correcting for these side effects . About 10 y after the beginning of OP treatments ( 1977 ) , the major resistance allele ace-1R , which is beneficial in the treated area but has strong deleterious pleiotropic effects in absence of insecticide , appeared and spread [17 , 18] . Because of this fitness cost and incoming gene flow from the nontreated area favored by the absence of insecticide treatment in winter , it did not fix but remained polymorphic with a clinal pattern across treated and nontreated areas [26 , 27 , 35] . In the early nineties , at least one and probably two duplications involving a resistance and susceptible ace-1 copy , started to spread and replace ace-1R [29 , 30] . Our lab experiments indicate that these two duplications are both severely deleterious when homozygous , but that heterozygotes involving either ace-1S or ace-1R alleles do not exhibit deleterious side effects . After ∼1999 , the pooled frequency of the duplications does not vary significantly , suggesting that the four “alleles” ( ace-1S , ace-1R , ace-1D2 , and ace-1D3 ) reached a stable equilibrium . The stability of the duplicated haplotype frequencies since 1996 suggests that both duplications occurred and spread before 1996 . This latter conclusion is tentative , since our data may not be accurate enough to detect the small perturbation of the frequency equilibrium caused by the spread of one of the duplications after 1996 , due to the indirect estimation of duplicated haplotype frequency . Clearly , however , considerable polymorphism in this system is maintained by both overdominance and migration . What is the mechanism by which overdominance operates ? First , a duplicated haplotype restores AChE1 activity while maintaining resistance [30] . It therefore combines resistance with no deleterious pleiotropic effect caused by a deficit of AChE1 activity . An excess of AChE1 activity may be deleterious as well , but the one caused by the duplication is mild . For this reason , duplications certainly exhibit at least marginal overdominance over treated and nontreated areas . Duplications could be described as “generalist” haplotypes that perform well in both habitats , compared to the “specialist” alleles ace-1S and ace-1R . Second , through lab experiments , we found , surprisingly , that both ace-1D2 and ace-1D3 duplications cause very strong deleterious effects when homozygous . However , these effects disappear in ace-1D2/ace-1D3 heterozygotes . The simplest explanation for this observation is that each duplication occurred independently rather than being generated by recombination ( which may not be surprising , given the high duplication rate at this locus [30] ) , and that they carry distinct recessive sublethal mutations not necessarily related to ace-1–mediated resistance . Distinct unequal crossing over could have generated each duplication and in the same time disrupted different genes close to ace-1 . Alternatively , different sublethal recessive mutations could have hitchhiked with the initial spread of each duplication . In all cases , these explanations require that the duplications cannot get rid of these deleterious mutations by recombining . It therefore suggests that recombination is very low around or within the duplicated haploypes or that they are situated on chromosomal inversions ( and thus behave like a “balancer region” in analogy with the balancer chromosomes used in laboratory Drosophila ) . The fact that we never observed in laboratory crosses a recombinant separating the two ace-1 duplicated copies [30] is consistent with this last hypothesis . We consider this explanation plausible since duplication events involve an important modification of the genome , which can easily disrupt other genes or regulatory regions [31 , 33 , 36] . This example of adaptation involves three successful steps ( the formation of ace-1R , ace-1D2 and ace-1D3 ) , each of them being driven by natural selection . However , each of these three steps presents severe deleterious pleiotropic effects . The deleterious pleiotropic effect of ace-1R may be unavoidable if changing the AChE1 active site to decrease affinity towards OP insecticide necessarily also leads to a lower affinity towards ACh . However , the occurrence and spread of duplications that are sublethal when homozygotic is more perplexing . For instance , another ace-1 duplication , ace-1D1 seems to have spread and be almost fixed on the island of Martinique [37] , indicating that ace-1 duplications do not necessarily involve strong deleterious pleiotropic effects when homozygous ( confirmed by laboratory analyses , P . Labbé and M . Weill , unpublished data ) . However , once a duplication with a recessive cost has spread , even partially , in a population , selection is likely to be less effective at replacing it with a better one because the fate of a beneficial mutation is mostly determined when it is rare and therefore by its heterozygous effect [38] . Since ace-1D2 or ace-1D3 duplications enjoy almost no fitness cost when heterozygotic , any new duplication has a low chance to spread in these populations ( although migration bringing fitter haplotypes could help escaping this apparent dead end ) . Thus , selection has taken the mosquito populations on a difficult path indeed . More generally , we may wonder if such a tortuous adaptive trajectory is frequent in nature . Clearly , gene duplications may solve many genetic trade-offs and chromosome rearrangements such as inversions may strongly perturb the genetic architecture . This type of situation may be more common with strong selective pressure and strong pleiotropy , whereas less intense selection may select for more subtle variation ( e . g . , see [39] ) . Nevertheless , insecticide resistance may not be an exception , since other selective pressures appear to be intense as well ( e . g . , parasitism [40 , 41] ) and thus could favor similar complex genetic responses . The new molecular tools available will allow deeper investigation of adaptation genetics evolution , and thus will help to settle the issue of the frequency of the kind of complex patterns uncovered by our study . In the long term , selection may produce exquisite adaptations , but this study lays bare that even impressive adaptations are likely to have begun with a process of trial and error that seems to be anything but optimal . It appears that natural selection is forced to tinker with available variability , despite the costs , rather than build impressive and cost-free adaptations that are wholly novel .
Data used in cline analysis were collected in the Montpellier area since 1986 along a transect across the treated and the nontreated areas ( Figure 2 ) . Published data from the summers of 1986 , 1991 , 1995 , and 1996 , spring of 1993 , and winters of 1995 and 1996 [26 , 29 , 35] were used to perform the overall analysis . This was complemented with unpublished samples: summer of 1999 , 2001 , and 2002 , spring of 1996 and 2000 , and winter of 1999 and 2002 ( complete dataset in Table S1 ) . For illustration , we also indicate the frequency observed in a single population in 1984 near the coast , but this population was not included in the cline analysis . Five populations were also sampled on the same transect in 2005 ( Maurin , Lattes , Distill , Viols , and Ganges ) to assess the distribution of the two duplicated haplotypes in the Montpellier area . In laboratory experiments , different strains were used to compare life history traits . Two reference strains were used: SLAB , the susceptible reference strain ( homozygote for the allele SSlab [34] ) , and SR , homozygote for ace-1R ( R ) , but with the same genetic background as SLAB [23] . Two other strains , named MAURIN-D and BIFACE-D and respectively harboring ace-1D2 ( D2 ) and ace-1D3 ( D3 ) , were also used . The duplicated strains originate from MAURIN and BIFACE strains [30] backcrossed for more than 15 generations with SLAB ( method in [23] ) . These strains are not homozygous but contain three genotypes: ( SSLAB/SSLAB ) , ( D/SSLAB ) , and ( D/D ) ( D2 and D3 for MAURIN-D and BIFACE-D , respectively ) . Egg rafts resulting from the cross of homozygous males and females originating from BIFACE-D were used to constitute a strain homozygous for each duplicated haplotype , BIFACE-DFix . SLAB-TC strain ( SLAB strain cured from Wolbachia bacteria [25] ) was used for crosses with field samples to avoid incompatibility phenomena . Identification of ace-1 phenotype . For each mosquito , the head was used to establish the phenotype at the ace-1 locus , using the TPP test [42] , based on enzymatic activity of AChE1 in the presence or absence of insecticide . Single copy allele homozygotes ( S/S ) and ( R/R ) are easily detectable using this test . However , the heterozygotes ( R/S ) and the genotypes involving a duplicated haplotype ( heterozygotes: ( D/S ) and ( D/R ) and homozygotes: ( D/D ) ) can not be distinguished ( they all display a [RS] phenotype ) . Moreover , this test does not allow distinction between the two duplicated haplotypes D2 and D3 . The pooled frequency of D2 and D3 can be estimated from the apparent excess of [RS] phenotypes caused by the presence of the duplications [26] . This method assumes Hardy–Weinberg proportions for the different genotypes and is therefore not as accurate as direct observation of the different genotypes . However , it has been shown to correctly estimate D frequency in field samples where duplication frequency was independently estimated using crosses [26] . Note that no molecular test is currently available to directly detect the duplicated haplotypes in natural populations , as their D ( R ) and D ( S ) copies are not different from single copy alleles ( R and S ) present in the same sampling sites . Clines description . We first analyzed the spatial variation in allele frequency across the treated and nontreated areas for each sample independently ( i . e . , for a given year and season ) using a purely descriptive model . We estimated the pooled frequency of both duplications . More specifically , we assumed that the frequency ( denoted p ) of each resistance allele ( indicated by the subscript i = R or D ) at time j followed a scaled negative exponential as follows: where x is the distance from the coast and hij , bij and aij are the estimated parameters . hij measures the frequency of resistance allele i on the coast ( i . e . , at x = 0 ) at time j . The parameters bij and aij describe rates of decline of the frequency of allele i ( at time j ) with distance and with the square of distance from the coast , respectively . We allow for a flexible clinal shape because it tends to vary with season [35] . The second step in the analysis was to compare frequency patterns across years and seasons . For this purpose , we fitted all samples simultaneously to measure the variation of clines trough time using the method developed in Labbé et al . [16] . More specifically , we assumed that hij values changed smoothly as a logistic function of time ( measured in months ) where t1j is the number of months after January 1986 when the date of sampling is before January 1986 + t* and is t* otherwise . t2j is the number of months after January 1986 + t* . αi , βi , γi , and t* are estimated parameters . The overall change in frequency over the 1986–2002 period is measured for each resistance allele i by αi and βi , which measure the rate of frequency change between 1986 and 1986 + t* and between 1986 + t* and 2002 , respectively . t* was introduced to allow for changes in the rate of allele replacement . Parameter γi is related to the initial frequency hi0 of each allele i ( hi0 = Exp ( γi ) /[1 + Exp ( γi ) ] ) . Expected phenotypic distributions were computed using allelic frequency and assuming Hardy–Weinberg proportions in each location ( see [16] ) . The phenotype was considered to be a three-state random variable ( [RR] , [RS] , and [SS] ) . The log-likelihood of a sample was computed from the phenotypic multinomial distribution and maximized using the Metropolis algorithm ( see [26 , 27 , 35] ) . Models were compared using F-tests in order to correct for overdispersion . Deviance was also corrected for overdispersion to find the support limits of each parameter [43 , 44] . Fitness estimation of the different genotypes . In order to determine the fitness of the different genotypes that would yield the stable clines observed over the period 1999–2002 , we also used a maximum-likelihood approach . For a given set of fitness values , we obtained by simulation the distribution of genotypes at equilibrium at different distances from the coast . From this distribution , we computed the expected frequency of [RR] , [RS] , and [SS] for each population in our dataset . The likelihood was then computed and maximized in the same way as with the descriptive models above . The simulation was performed using a stepping stone across treated and nontreated areas using the dispersion kernel that has been previously estimated for C . pipiens as described in Lenormand et al . [26] but with a treated area of 16 km that reflects the treatment practices over this period ( P . Labbé , unpublished data ) . Because we do not have estimates for the relative frequency of the two duplications , we could not estimate their relative fitness . We therefore report estimates assuming that the two duplications have the same fitness effects . The fitness of a given genotype was modelled with two components: a fitness cost c ( expressed in both treated and non treated areas ) and a reduction in survival due to the presence of insecticide in the treated area s . The full parameter range was constrained to reflect that s should be lower ( or equal ) for genotypes with an increasing number R copies . At first it may be surprising that polymorphism with more than two alleles could be maintained at migration—selection equilibrium in a situation with only two habitats ( treated and nontreated areas ) . First , even with haploid selection , the number of alleles that can be maintained is larger than the number of habitats if dispersal is localized ( as confirmed by our simulations , see also [45] ) . Second , particular overdominance relationships among alleles ( such as the ones we find ) can also increase the number of alleles that can be maintained at equilibrium . Crosses . Larvae were sampled in five populations along the sampling transect ( Maurin , Lattes , Distill , Viols , and Ganges; Figure 2 ) and reared in the laboratory . They were exposed to a dose of insecticide that kills all [SS] individuals ( 25 × 10−6 M of Propoxur ) . About 700 resistant females from each population were crossed with about 800 males of SLAB-TC strain ( SSLAB/SSLAB ) . They were repeatedly blood fed each week until they died . Egg rafts were collected every day , isolated , and reared to the third instar . They were then exposed to a dose of insecticide that kills all [SS] individuals in order to discard all susceptible field alleles resulting from ( D/S ) or ( R/S ) mothers . DNA was extracted from a pool of ∼20 survivors of each egg raft . Molecular tests described below were used to detect the presence of each resistance haplotype ( R , D2 , and D3 ) . Four types could be identified: genotypes ( D2/D3 ) and ( R/R ) and phenotypes [D2] ( i . e . , ( D2/D2 ) , ( D2/S ) , or ( D2/R ) ) , and [D3] ( i . e . , ( D3/D3 ) , ( D3/S ) , or ( D3/R ) ) . Molecular tests . Using partial ace-1 sequences of each haplotype [30] , we designed RFLP tests to discriminate D2 , D3 , R , and SSLAB . PCR amplification of a 458 bp fragment of exon 3 using the primers CxEx3dir 5′-CGA CTC GGA CCC ACT CGT-3′ and CpEx3rev 5′-GAC TTG CGA CAC GGT ACT GCA-3′ was performed ( 30 cycles , 93 °C for 30s , 55 °C for 30s , and 72 °C for 1min ) . The amplified fragment was then digested in parallel by different restriction enzymes . First , the fragment was cut twice by the enzyme BsrBI only when SSLAB is present , generating three fragments ( 127 bp , 141 bp , and 190 bp; Figures S1 and S2 ) , all the other alleles being cut only once ( two fragments of 127 bp and 331 bp; Figures S1 and S2 ) . Second , the fragment is cut by the enzyme EagI only when D2 ( S ) is present , generating two fragments ( 150 bp and 308 bp; Figures S1 and S2 ) . Third , the fragment is cut by the enzyme HinfI only when D3 ( S ) is present , generating two fragments ( 102 bp and 354 bp; Figures S1 and S2; note that there is a HinfI site in the primer CxEx3dir , subtracting 2 bp in each fragment; Figure S1 ) . Discriminating between the resistance and susceptible copies is possible using the test provided by Weill et al . [20]: the G119S mutation providing resistance creates a site for the enzyme AluI . Larval mortality: ( D/D ) versus ( D/S ) . Trials between ( D/D ) and ( D/S ) individuals were performed in triplicate . Larvae of different genotypes were reared in competition under the same environmental conditions ( food , temperature , etc . ) . They were selected at the first instar stage using Propoxur at a concentration of 25 × 10−6 M , which eliminates only ( S/S ) individuals . Adults were collected during the first and the second wk after the first adult emergence ( ∼30 individuals each wk ) . They will be respectively referred as early ( first wk ) and late ( second wk ) emerging adults . Genotype frequency was measured at second larvae instar and both adulthood stages . Three trials were conducted: ( i ) ( D3/S ) versus ( D3/D3 ) , ( ii ) ( D2/S ) versus ( D2/D2 ) , and ( iii ) ( D3/S ) versus ( D2/D3 ) . The different genotypes were obtained respectively from ( i ) a cross between males and females from BIFACE-D ( progeny genotypes ( D3/D3 ) , ( D3/SSLAB ) , and ( SSLAB/SSLAB ) ) , ( ii ) a cross between males and females from MAURIN-D ( progeny genotypes ( D2/D2 ) , ( D2/SSLAB ) , and ( SSLAB/SSLAB ) ) and ( iii ) a cross between males from BIFACE-DFix and females from MAURIN-D and the reverse cross ( progeny genotypes ( D2/D3 ) and ( D3/SSLAB ) ) . Each sample was analyzed using the BsrBI-based RFLP test to determine the proportion of individuals of genotype ( D3/SSLAB ) ( in the first and third trials ) or ( D2/SSLAB ) ( in the second trial ) . Larval mortality: ( D/S ) or ( D/R ) versus ( R/R ) . Trials were conducted between ( D/R ) or ( D/S ) individuals and individuals homozygote for the single resistance allele ( R/R ) . Trials were performed in triplicates with 500 first instar larvae of each genotype reared under the same environmental conditions ( food , temperature , etc . ) . In each replicate , early and late emerging adults were collected as indicated above . Two trials were carried out: ( i ) ( D3/R ) versus ( R/R ) and ( ii ) ( D3/S ) versus ( R/R ) . The different genotypes were obtained from ( i ) a cross between females from BIFACE-DFix and males from strain SR to obtain ( D3/R ) individuals , and ( ii ) a cross between females from BIFACE-DFix and males from SLAB to obtain ( D3/SSLAB ) individuals . ( R/R ) individuals were directly obtained from strain SR . Each sample was analyzed using the AluI-based RFLP test to determine the proportion of individuals of genotype ( D3/R ) ( first trial ) or ( D3/SSLAB ) ( second trial ) . Fertility . Different crosses were realized in order to determine the fertility of individuals carrying a duplicated allele at the homozygous or heterozygous state ( ( D/D ) or ( D/S ) ) . In each case , the proportion of females laying eggs and the proportion of hatching eggs rafts was recorded . For each series of cross , about 50 males were mated independently with five females each , all from the same strain . Five days after mating , the genotype of the males was determined using the BsrBI-based RFLP test . Females were then grouped according to the male genotype , blood fed , and kept without access to laying substrate . Six days later , they were allowed to lay eggs individually . The females were genotyped either after they laid eggs or less than 6 h after their death . Two series of crosses were realized ( i ) between males and females from BIFACE-D ( genotype ( D3/D3 ) or ( D3/SSLAB ) after selection ) , and ( ii ) between males and females from MAURIN-D ( genotype ( D2/D2 ) or ( D2/SSLAB ) after selection ) , to assess the fertility of individuals carrying ace-1D3 and ace-1D2 , respectively . Statistical analysis . For the larval mortality analysis , the following generalized linear model ( GLM ) was fitted , with binomial error: DD = Time + Replicate + Time . Replicate , where DD represents the proportion of the ( D/D ) genotype in the population ( i . e . , ( D3/D3 ) , ( D2/D2 ) , and ( D2/D3 ) for the first , second , and third crosses , respectively ) , Time is a factor indicating when the sample was taken , Replicate is a factor indicating the three containers in which the experiment was replicated , and Time . Replicate is the interaction between the two factors . Two analyses were performed: ( i ) one to test for a difference in mortality between ( D/D ) and ( D/S ) individuals ( in that case , Time was 2nd instar or emerging adult ) , and ( ii ) the second to test for a difference in development time between ( D/D ) and ( D/S ) individuals ( in that case , Time was early emerging or late emerging adults ) . A similar model was used to analyze the proportion of ( D3/S ) and ( D3/R ) in trials versus ( R/R ) . These models were simplified according to Crawley [46]: significance of the different terms was tested starting from the higher-order terms using F-test . Nonsignificant terms ( p > 0 . 05 ) were removed . Factor levels of qualitative variables that were not different in their estimates ( using F-test ) were grouped as described by Crawley [46] . This process yielded a minimal adequate model . The fertility of males and females of different genotypes was analyzed by comparing the proportion of females laying eggs and the proportion of hatching egg rafts among the different types of cross . The number of females laying eggs ( Ne ) and the number of hatching egg rafts ( Nh ) were analyzed using GLM with binomial error: Male + Female + Male . Female , where Male and Female are factors indicating male ( or female ) genotype . These models were simplified as above . All analyses were performed using R software ( v 2 . 0 . 1 . , http://www . r-project . org ) .
The National Center for Biotechnology Information ( NCBI ) GenBank database ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=Nucleotide ) accession numbers for ace-1 are AJ489456 and AJ515147 .
|
Adaptation is not always a straightforward process , and often results from natural selection tinkering with available variation . We present in this study just such a tortuous natural selection pathway , which allows the mosquito Culex pipiens to resist organophosphorous insecticides . In the Montpellier area , following the use of insecticide to control mosquito populations , a high-resistance allele of the insecticide target enzyme appeared . But this allele also displayed strong deleterious side effects . Recently , several duplicated haplotypes began to spread in natural population that put in tandem a susceptible and a resistant allele . We show that the duplicated haplotypes actually display reduced side effects compared to the resistant allele when in the heterozygous state , but also new and strong costs in the homozygote . This pattern leads to an unexpected equilibrium between four different alleles across treated and nontreated areas . The story of resistance in C . pipiens is indeed far from a slow progression toward a “perfect” adaptation . Rather , selection for resistance to insecticide is a long process of trial and error leading to an uncommon genetic architecture .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"Supporting",
"Information"
] |
[
"molecular",
"biology",
"arthropods",
"eukaryotes",
"culex",
"evolutionary",
"biology",
"animals",
"genetics",
"and",
"genomics",
"insects"
] |
2007
|
Forty Years of Erratic Insecticide Resistance Evolution in the Mosquito Culex pipiens
|
As international travel increases , there is rising exposure to many pathogens not traditionally encountered in the resource-rich countries of the world . Filarial infections , a great problem throughout the tropics and subtropics , are relatively rare among travelers even to filaria-endemic regions of the world . The GeoSentinel Surveillance Network , a global network of medicine/travel clinics , was established in 1995 to detect morbidity trends among travelers . We examined data from the GeoSentinel database to determine demographic and travel characteristics associated with filaria acquisition and to understand the differences in clinical presentation between nonendemic visitors and those born in filaria-endemic regions of the world . Filarial infections comprised 0 . 62% ( n = 271 ) of all medical conditions reported to the GeoSentinel Network from travelers; 37% of patients were diagnosed with Onchocerca volvulus , 25% were infected with Loa loa , and another 25% were diagnosed with Wuchereria bancrofti . Most infections were reported from immigrants and from those immigrants returning to their county of origin ( those visiting friends and relatives ) ; the majority of filarial infections were acquired in sub-Saharan Africa . Among the patients who were natives of filaria-nonendemic regions , 70 . 6% acquired their filarial infection with exposure greater than 1 month . Moreover , nonendemic visitors to filaria-endemic regions were more likely to present to GeoSentinel sites with clinically symptomatic conditions compared with those who had lifelong exposure . Codifying the filarial infections presenting to the GeoSentinel Surveillance Network has provided insights into the clinical differences seen among filaria-infected expatriates and those from endemic regions and demonstrated that O . volvulus infection can be acquired with short-term travel .
Parasitic diseases are widespread throughout the developing world and are associated with a heavy burden of morbidity and mortality . Human filariae , nematodes transmitted by arthropod vectors , are endemic in tropical and subtropical regions of the world . With an estimated 80 million people who travel to developing countries each year [1] , exposure to filarial parasites is likely to become more common . It has been suggested that infection with filariae requires prolonged and intense exposure to the vectors that transmit them [2] . Moreover , when comparing nonendemic visitors who have acquired filarial infections with those born in endemic regions , the nonendemic visitors appear to have greater numbers of objective clinical symptoms and fewer clinically asymptomatic ( or subclinical ) infections [3]–[7] . The GeoSentinel Surveillance Network , a global network of specialized travel/tropical medicine clinics on six continents , was established in 1995 to contribute clinician-based sentinel surveillance on all travelers seen [8] . We examined data from the GeoSentinel database to identify demographic and travel characteristics associated with filaria acquisition in addition to species distribution of filarial acquisition and patient symptoms . Because there have been no comprehensive studies that have addressed the acquisition of filarial infections among nonendemic travelers , the present study was performed to understand travel-related filarial infections from a global viewpoint that could inform physicians and travelers alike .
Demographic , travel , and clinical data were collected from all patients seen at each GeoSentinel site . Travel information was also collected , including trip start and end dates for travel within 6 months and countries visited in the previous 5 years . Countries listed included birth country , country lived in prior to age 10 , country of residence , and country of citizenship . Patient classification , the reason for recent travel , symptoms , and final diagnosis were reported by health care providers at GeoSentinel site clinics . Patient information was entered without identifiers into an Access database ( Microsoft ) . Each individual record with a diagnosis of filarial infection was examined manually to verify that the place of exposure was in a filaria-endemic country and that the data provided were accurate and complete . The GeoSentinel data-collection protocol was reviewed by the institutional review board officer at the National Center for Infectious Diseases at the Centers for Disease Control and Prevention and classified as public health surveillance and not as human-subjects research requiring submission to institutional review boards . Data entered into the GeoSentinel database from patients seen from August 1997 through December 2004 were used . This analysis focused on data extracted from persons who were assigned codes corresponding to infection with Onchocerca volvulus , Wuchereria bancrofti , Loa loa , other filarial species , or unknown filarial species . Prior to analysis , a survey of all GeoSentinel sites was performed to ensure that the definition of infection was uniform among the reporting sites . Data were managed in Microsoft Access and were analyzed using SAS v . 9 . 1 ( SAS Institute ) . Crude odds ratios were calculated from a bivariate analysis , and statistical significance was determined by χ2 tests .
From a total of 43 , 722 individual patient encounters , filarial infections were diagnosed for 271 ( 0 . 62% ) persons who presented to GeoSentinel sites from August 1997 through July 2004 . The reporting of cases to GeoSentinel was lowest in 1997 and 1998 ( 3 . 7% and 8 . 9% respectively ) ; from 1999 through 2004 , filariasis as a proportion of morbidity ( ill patients reporting to the clinics ) fluctuated between 11% ( n = 30 ) and 17 . 5% ( n = 47 ) . Of the 271 patients with filarial infections , 37% were diagnosed with O . volvulus , 25% were infected with L . loa , and another 25% were diagnosed with W . bancrofti . Among all filarial infections , 5 . 5% were identified as other filarial species , ( e . g . , Mansonella , Brugia spp . ) , and 5 . 5% of all filarial infections reported in the database were unspecified . Three patients were coinfected with L . loa and other filarial species; one patient presented with O . volvulus and L . loa coinfection ( Figure 1 ) . Overall , 122 ( 45% ) patients were female; gender was not recorded for 17 ( 6 . 3% ) patients . Patient mean age was 34 . 9 years ( range 0–84 ) . The region of acquisition among filaria-infected individuals was assigned when possible ( n = 230 ) . The majority ( 75% ) of infections were acquired in Africa ( both Northern Africa and Sub-Saharan Africa ) and 10% in South America ( see Table 1 ) . The remaining individuals were exposed in , Oceania , the Caribbean , South Central Asia , and Central America . Of all filarial infections reported to the GeoSentinel ntwork ( n = 271 ) , the majority were reported by the North American sites ( 76 . 4% ) ; 18 . 5% were reported from European sites , and the remainder were reported from GeoSentinel sites in the Middle East , Australia/New Zealand , and South Central Asia . Among the 271 patients diagnosed with filarial infections , the majority ( 62% ) occurred among immigrants . Non-urban expatriates and travelers represented the second largest group of patients with filarial infections . Foreign visitors , urban expatriates , and students ( Figure 2 ) comprised the groups in which there were relatively few filarial infections . As an overall proportion of GeoSentinel reports , filarial infections were found to occur in 1 . 6% of immigrants , 2 . 4% of non-urban expatriates , 1 . 5% of students , 0 . 2% of foreign visitors , 0 . 2% of urban expatriates , and 0 . 2% of travelers . The ‘reasons for travel’ were predominantly for immigration or for immigrants who were VFR in endemic regions ( 63% ) . An additional 16% of patients traveled for missionary or volunteer activities , and the remainder traveled for tourism , research/education , or business-related purposes ( Figure 3 ) . When grouped by type of parasite , immigrants and VFR had the greatest proportion of diagnosed onchocerciasis ( 48% ) compared with nonendemic visitors ( 20% ) . Twenty-nine percent of VFR and immigrants with filarial infections were infected with W . bancrofti , while only 18% of nonendemic visitors had W . bancrofti infection . The diagnosis of L . loa was greatest among nonendemic visitors ( 43% ) , compared with 15% of VFR and immigrants with loiasis ( Figure 4 ) . Travel duration was known definitively for 108 of the 271 individuals with filarial infection . Among these 108 , 48 persons originated from nonendemic regions but only 34 had recorded travel data definitively related to the place of exposure . Trip duration ranged from 7 days to 17 . 7 years ( geometric mean duration: 125 days; median duration: 87 days ) . The majority of patients with O . volvulus infections had trip durations of up to 1 month ( Table 2 ) . The majority of those with L . loa infections had traveled between 1 and 6 months , while the highest percentage of patients with W . bancrofti infection occurred after more than 6 months of travel ( and presumed exposure ) . The time to presentation to a GeoSentinel site after arrival in a filaria-nonendemic country was calculated to identify the possible incubation period between exposure and clinical presentation in only nonendemic visitors ( VFR and immigrants excluded from this analysis ) . For O . volvulus infections , 67% presented to a GeoSentinel site within 1 month of return , and 100% of those with W . bancrofti presented between 1 and 6 months . Among those with L . loa , 12% presented within the first month of return , 77% within 1 to 6 months of return , and the remainder at least 6 months after return . Among patients infected with other filarial species , the majority presented within 1 month of return . These data suggest that Onchocerca infections are more likely to be symptomatic early in the infection compared to either Loa loa or Wuchereria bancrofti ( data not shown ) . In studies done previously in loiasis [3] and onchocerciasis [4] among limited numbers of expatriates , the data suggested that the clinical symptoms were more pronounced ( and less likely to be asymptomatic ) in travelers ( temporary visitors ) to filaria-endemic regions of the world compared with those with lifelong exposure and chronic infections [3] . To examine this issue more closely , a comparison was made between those infections that were clinically symptomatic and those that were clinically asymptomatic ( Table 3 ) . Characterization of symptoms included those associated with the following organ systems: skin , cardiac , respiratory , gastrointestinal , genitourinary , neurologic , musculoskeletal , ophthalmologic , and otolaryngologic , in addition to complaints of fatigue , fever , and psychological problems . If the patient had no complaints or symptoms or in which filarial infection was identified incidentally following evaluation for another condition , then asymptomatic was recorded . As seen , those individuals in the GeoSentinel database identified to have filarial infection who were born and raised in endemic regions were 2 . 5 times as likely to be clinically asymptomatic ( CI 1 . 07 , –5 . 93 ) compared with those who traveled from filaria-nonendemic to filaria-endemic regions ( P< . 03 )
While filarial infection and disease are most frequently diagnosed among native residents of endemic regions , the risk of infection acquisition among travelers from nonendemic regions is sizeable . Filarial species are found in tropical and sub-tropical regions of the world and , as travel to these regions becomes more popular , filarial infection among nonendemic visitors becomes increasingly common as well . We describe here important epidemiologic characteristics of filarial infections acquired by world travelers from nonendemic regions as reported to the GeoSentinel network . While clinical presentation of filarial disease is known to differ between visitors to and natives of endemic regions [3] , our analysis also provides a quantitative assessment of filarial acquisition among travelers and helps describe the differences in clinical presentation between those native to filaria-endemic regions and those traveling to those regions . Filarial infections comprised 271 cases ( 0 . 62% ) of all medical conditions reported to the GeoSentinel network . O . volvulus was responsible for the greatest number of filarial infections ( n = 101 ) , followed by equal numbers ( n = 68 ) of L . loa and W . bancrofti ( Figure 1 ) . Because the GeoSentinel database includes immigrants/refugees who undergo laboratory screening that includes filarial serologies when eosinophilia or clinical signs or symptoms of filarial disease are present , it is not surprising that the majority of filaria-infected patients in the GeoSentinel network were immigrants ( 62% ) . Due to lifelong chronic exposure , the prevalence of filarial infections among immigrants can be significant . It has , however , typically been said that infection acquisition is low for short-term , nonendemic travelers . Although travel information was only available for a subset of the total number of filaria-nonendemic visitors , it was still unexpected to find that almost one-third ( 30% ) of travelers from nonendemic regions acquired their filarial infections during trips of 31 days or less ( the majority of O . volvulus infections ) , and only 38% of filarial infections occurred from trip durations exceeding 180 days ( Table 2 ) . There are numerous case reports and case series that describe durations of exposure as short as 10 days among filaria-infected patients from nonendemic regions [5] , [9]–[12] . It is possible that the lack of preventive measures such as insect repellent and bednets , as well as individuals close proximity to vector habitats , played a role in infection acquisition regardless of short durations of exposure . Further , development of symptoms may also be dependent on the density of filarial larval inoculation as well as individual innate immune responses [13] . Because almost all of the major filarial infections ( O . volvulus , W . bancrofti , L . loa , M . perstans , M . streptocerca ) are endemic in sub-Saharan Africa , it is not surprising that 72% of filarial infections reported to GeoSentinel were acquired in this region: 95 . 5% of those with onchocerciasis were acquired in sub-Saharan Africa; three were acquired elsewhere . Thirty-two percent of the W . bancrofti infections were acquired in South America , compared with only 12% of W . bancrofti infections reported from sub-Saharan African regions , 22% from South Central Asia , and 14% from the Caribbean . As expected , 100% of loiasis cases were acquired in West and Central Africa , as the parasite is endemic only in this region . While short-term nonendemic visitors appear less likely to acquire filarial infections , among those with relatively long-term exposure there have been many case reports of travel-related filarial infections and associated clinical symptoms [3]–[5] , [11] , [13]–[15] . Presentation of clinical disease among patients with L . loa , O . volvulus , and W . bancrofti differs considerably between expatriates ( or long-term temporary residents ) and those born in filaria-endemic regions of the world . Among those infected with L . loa , infected expatriates typically have a greater frequency of Calabar swellings , higher grade levels of filaria-specific antibody and peripheral eosinophil counts , and more nonspecific complaints , while those born and raised in endemic regions are more likely to have asymptomatic infections associated with microfilaremia . Those born in regions with endemic O . volvulus infection generally have higher levels of skin microfilariae and more ocular disease than do nonendemic visitors to these regions [16] . Those living in regions with endemic lymphatic filariasis most commonly have asymptomatic ( or subclinical ) infections , although significant proportions of infected individuals develop hydrocele , lymphedema elephantiasis , or chyluria . Nonendemic visitors ( and short-term visitors ) rarely have asymptomatic microfilaremic condition , but rather are more likely to develop lymphadenitis , hepatomegaly , splenomegaly and reversible lymphedema [17] . This study corroborates many of the anecdotal reports about the differences between the clinical presentations among travelers compared with those with chronic ( and often lifelong ) exposure to filarial parasites . Case report findings describe the clinical manifestations of filarial disease to be greater among expatriates , while those from filaria-endemic regions present commonly without symptoms . Indeed , our results from the GeoSentinel network indicate that filaria-infected patients with long-term exposure to filariae were more commonly asymptomatic ( or subclinical ) compared with those expatriates with filarial infections . With the collection of surveillance data on travel-related medical conditions by the GeoSentinel network , epidemiologic data can describe morbidity and mortality trends among travelers [18] . While these networks are generally used to follow acute infections among nonendemic visitors , we have demonstrated here the utility of surveillance for chronic infections , as well . Diagnoses of filarial infections in industrialized countries will likely continue to rise as increasing numbers of people travel to endemic regions and as increasing numbers of refugees and immigrants arrive from endemic areas . The majority of nonendemic filaria-infected visitors ( 64 . 7% ) presented to a GeoSentinel site clinic between 1 and 6 months after return of travel , underscoring the need for surveillance of chronic infections to ensure safety and treatment of returning travelers from developing regions . In conclusion , analysis of data on filarial infections available from the GeoSentinel network enabled us to describe characteristics of patients presenting with filarial infection and to determine that filarial infections can be acquired with relatively short-term exposure . Our study not only corroborates but expands the understanding of the differences in filarial disease manifestation between those traveling to and those born in filaria-endemic regions of the world by providing a quantitative analysis of filarial acquisition among nonendemic visitors . Moreover , our data demonstrate that globally acquired travel data can be used to follow not only acute but also chronic infections and can ultimately provide a more comprehensive backdrop to pre-travel advice and to post-travel treatment for those at risk of acquiring a filarial infection .
In addition to the authors , members contributing data include: Graham Brown and Joseph Torresi , Royal Melbourne Hospital , Melbourne , Australia; Giampiero Carosi and Francesco Castelli , University of Brescia , Brescia , Italy; Lin Chen , Mount Auburn Hospital , Harvard University , Cambridge , Massachusetts , USA; Bradley Connor , Cornell University , New York , New York , USA; Jean Delmont and Philippe Parola , Hôpital Nord , Marseille , France; Carlos Franco and Phyllis Kozarsky , Emory University , Atlanta , Georgia , USA; David Freedman , University of Alabama , Birmingham , Alabama , USA; Stefanie Gelman and Devon Hale , University of Utah , Salt Lake City , Utah , USA; Alejandra Gurtman , Mount Sinai Medical Center , New York City , New York , USA; Jean Haulman and Elaine Jong , University of Washington , Seattle , Washington , USA; Kevin Kain , University of Toronto , Toronto , Canada; Carmelo Licitra , Orlando Regional Health Center , Orlando , Florida , USA; Prativa Pandey , CIWEC Clinic Travel Medicine Center , Kathmandu , Nepal; Patricia Schlagenhauf and Robert Steffen , University of Zurich , Zurich , Switzerland; Eli Schwartz , Sheba Medical Center , Tel Hashomer , Israel; Marc Shaw , Travellers Health and Vaccination Centre , Auckland , New Zealand; Mary Wilson , Harvard University , Cambridge , Massachusetts , USA; Murray Wittner , Albert Einstein School of Medicine , Bronx , New York , USA .
|
As international travel increases , there is rising exposure to many pathogens not traditionally encountered in the resource-rich countries of the world . The GeoSentinel Surveillance Network , a global network of medicine/travel clinics , was established in 1995 to detect morbidity trends among travelers . Filarial infections ( parasitic worm infections that cause , among others , onchocerciasis [river blindness] , lymphatic filariasis [e . g . elephantiasis , lymphedema , hydrocele] and loiasis [African eyeworm] ) comprised 0 . 62% ( n = 271 ) of the 43 , 722 medical conditions reported to the GeoSentinel Network between 1995 and 2004 . Immigrants from filarial-endemic regions comprised the group most likely to have acquired a filarial infection; sub-Saharan Africa was the region of the world where the majority of filarial infections were acquired . Long-term travel ( greater than 1 month ) was more likely to be associated with acquisition of one of the filarial infections than shorter-term travel .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Members",
"of",
"the",
"Geosentinel",
"Surveillance",
"Network"
] |
[
"infectious",
"diseases/helminth",
"infections",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"infectious",
"diseases"
] |
2007
|
Filariasis in Travelers Presenting to the GeoSentinel Surveillance Network
|
DNA double-strand breaks trigger the production of locus-derived siRNAs in fruit flies , human cells and plants . At least in flies , their biogenesis depends on active transcription running towards the break . Since siRNAs derive from a double-stranded RNA precursor , a major question is how broken DNA ends can generate matching sense and antisense transcripts . We performed a genome-wide RNAi-screen in cultured Drosophila cells , which revealed that in addition to DNA repair factors , many spliceosome components are required for efficient siRNA generation . We validated this observation through site-specific DNA cleavage with CRISPR-cas9 followed by deep sequencing of small RNAs . DNA breaks in intron-less genes or upstream of a gene’s first intron did not efficiently trigger siRNA production . When DNA double-strand breaks were induced downstream of an intron , however , this led to robust siRNA generation . Furthermore , a downstream break slowed down splicing of the upstream intron and a detailed analysis of siRNA coverage at the targeted locus revealed that unspliced pre-mRNA contributes the sense strand to the siRNA precursor . Since splicing factors are stimulating the response but unspliced transcripts are entering the siRNA biogenesis , the spliceosome is apparently stalled in a pre-catalytic state and serves as a signaling hub . We conclude that convergent transcription at DNA breaks is stimulated by a splicing dependent control process . The resulting double-stranded RNA is converted into siRNAs that instruct the degradation of cognate mRNAs . In addition to a potential role in DNA repair , the break-induced transcription may thus be a means to cull improper RNAs from the transcriptome of Drosophila melanogaster . Since the splicing factors identified in our screen also stimulated siRNA production from high copy transgenes , it is possible that this surveillance mechanism serves in genome defense beyond DNA double-strand breaks .
DNA is constantly challenged by mutagenic processes of extrinsic and intrinsic origin . Of these damages , DNA double-strand breaks are particularly problematic lesions because they disrupt the continuity of genetic information . Their repair can either proceed via end-joining activities or through homology-directed repair [1] . A detailed mechanistic understanding of these repair processes is not only important for the prevention and treatment of diseases , such as cancer , but also to help researchers direct the outcome of genome editing experiments with precision [2 , 3] . The information stored in DNA is read-out by the process of transcription into RNA . DNA damage is therefore not only an impediment for replication , but also a hindrance for RNA biosynthesis . If DNA damage has occurred in an actively transcribed region , concomitant action of DNA repair factors and the transcription machinery is not possible; access to the DNA must thus be regulated [4] . Stalling of an RNA polymerase upstream of the damaged site may lead to extensive RNA-DNA hybrids , called R-loops , which can themselves cause genomic instability [5–7] . Consequently , a domain of specific chromatin states is assembled around a DNA double-strand break and transcriptional silencing can occur [8] . On the other hand , RNA polymerases stalled by certain types of base damage serve as sensors and thus promote repair of the lesion during transcription-couple repair [9 , 10] . It is also established that some RNA binding proteins are recruited to sites of DNA damage [5] and non-coding transcription may play an important role in DNA repair [11 , 12] . Finally , the Prp19 component of the spliceosome can interact with RPA bound to single-stranded DNA and reinforces activation of the protein kinase ATR in a manner that is independent of its function during the splicing reaction[13] . Recently , formation of locus-specific siRNAs has been observed at DNA double-strand breaks [14–17] . They may promote repair via homologous recombination in mammalian cells and plants [15 , 18] , but the molecular mechanisms through which siRNAs promote homologous recombination are not fully established; in mammalian cells they may involve targeting of Rad51 to the damaged site via protein-protein interactions with Ago2 [18] , but it is challenging to exclude indirect effects via perturbed miRNA biogenesis in these experiments . In Neurospora crassa , their biogenesis is even dependent on , rather than important for , homologous recombination [19–21] . In Drosophila the generation of DNA damage-induced siRNAs depends on transcription and is limited to only one side of the broken DNA , the region between a transcription start site and the DNA end . It was thus proposed that the DNA ends serve as transcription initiation sites to generate corresponding antisense transcripts to generate dsRNA [16] . Consistently , transcription initiation at DNA breaks has now been directly observed in S . pombe [12] . However , no DNA repair defects could be observed in dcr2 mutant Drosophila melanogaster flies where the siRNA pathway is completely inactivated but the miRNA pathway remains mostly unperturbed [22] . Although the significance of the siRNAs for DNA repair is thus a matter of debate , their presence demonstrates that transcripts running towards a DNA DSB are subject to some sort of surveillance and , as a consequence , at least partially converted into double-stranded RNA ( the precursor of siRNAs ) . To shed light on the mechanistic details of this process , we conducted a genome-wide RNA interference screen in cultured Drosophila cells using siRNAs generated from a linearized plasmid . These siRNAs control expression of a reporter gene . In addition to DNA double-strand break repair proteins , we discovered that many splicing factors , in particular components from the Prp19 and Prp19-related spliceosome sub-complexes are important for siRNA generation . Consistently , the presence of upstream introns greatly stimulated siRNA generation at chromosomal DNA DSBs induced by CRISPR-cas9 . An intron-less gene , on the other hand , only generated few siRNAs upon cleavage . We propose that the perturbed transcript maturation that ensues when RNA polymerase II encounters a DNA double-strand break is sensed with participation of the spliceosome . As a consequence , double-stranded RNA is generated and processed into siRNAs . Intriguingly , the splicing factors identified in our screen for break-induced siRNA generation were also important for siRNA generation from high-copy transgenes . It is thus conceivable that the transcript maturation surveillance mechanism serves in genome defense beyond DNA double-strand breaks .
We had previously described a reporter system that allows for a dual luciferase-based readout of DNA double-strand break derived siRNA activity [16] . In short , a linearized plasmid with either a truncated or an inverted coding sequence of Renilla luciferase is co-transfected with a mix of circular expression vectors for Renilla and firefly luciferase to control for transfection efficiency . The siRNAs generated from the linearized plasmid can then repress full-length Renilla luciferase expression in trans . When combined with prior experimental RNAi , this assay system has a high signal-to-noise ratio and can easily be scaled up . Note that the promoters in all reporter constructs contain an intron in the 5’-UTR , which precedes the Renilla or firefly luciferase CDS . We thus performed a genome-wide RNAi screen in Drosophila S2-cells to identify factors that are required for the generation of DNA double-strand break derived siRNAs ( see S1 Fig and S2 Fig for details ) . Two independent biological replicates of the entire screen were performed and averaged . After removing likely false positives , such as retracted gene models or genes that are not expressed in S2-cells , we selected a total of 142 positive and 66 negative candidates from the screen for further validation ( S1 Fig and S1 Table ) . We re-screened the original dsRNA trigger of our candidates for a third biological replicate and then generated two independent , non-overlapping dsRNAs for each candidate to identify false positives due to off-target RNAi . Only those candidates that scored positive for at least two out of the three distinct RNAi triggers ( = screened dsRNA and two validation constructs ) were retained . We also counter-screened the entire set of candidates with a cell line where a GFP reporter is repressed by two perfect matches to miR-277 in its 3’-UTR; expression of this reporter is driven by the same promoter as the Renilla luciferase in the screen . Since miR-277 is processed by Dcr-1 , then loaded into Ago2 via the Dcr-2/R2D2 complex , we could distinguish between core RNAi pathway components or factors that non-specifically activate transcription of our reporter and those factors that are specifically required for DNA damage-induced siRNA biogenesis . After this stringent validation process ( summarized in S1 Table ) , we retained a set of 89 genes that promote DSB-derived siRNA generation or function and 36 candidates that are potential repressors of DBS-derived siRNA production . To obtain an initial overview of the biological processes involved in siRNA generation at the DNA break we performed a gene ontology analysis of the validated candidates ( Fig 1A ) and calculated significances using g:profiler [23] . As expected for a screen with linearized DNA , we identified a series of DNA replication/repair factors among the positive candidates ( GO-term enrichment of “DNA metabolic process” with a p-value of 9 . 0x10-5 ) . DNA double-strand breaks are recognized by the Mre11-Rad50-Nbs1 ( MRN ) complex; all components of this complex were among the initial candidates and 2 out of 3 passed our validation experiments . In addition to DNA damage signaling , the MRN complex initiates 5’ to 3’resection of the break and thus commits the site for homology-directed repair . The 3’ single-stranded end is subsequently extended further by CtIP/Sae2 , promoting homologous recombination [24] . Bioinformatic analysis has suggested that the Drosophila CtIP homolog is CG5872 [25] and this gene was also required for correct re-localization of a heterochromatic DSB [26] . We identified CG5872 as a strong candidate in our screen; CG5872 is thus most likely the Drosophila homolog of CtIP/Sae2 and we propose that it should be called dCtIP ( Fig 1B ) . For further repair , DNA synthesis carried out by the replicative polymerases DNA-polδ and DNA-polε after Rad51-mediated annealing of the exposed 3’ single-stranded regions with a homologous template . We identified DNA-polδ and all subunits of replication factor C ( RfC ) as stimulators of break-induced siRNA generation . RfC normally loads the processivity clamp proliferating cell nuclear antigen ( PCNA , called mus209 in Drosophila ) to ensure long-range DNA synthesis . However , the PCNA-homolog mus209 was initially among the negative candidates but did not pass our validation criteria ( Fig 1C ) . NHEJ factors , such as the Ku70/Ku80 complex , were not identified in the screen . Taken together , our screening efforts demonstrate that recognition and processing of the DNA double-strand break for homology-directed repair promotes and thus precedes siRNA generation . Much more striking than the recovery of the DNA repair factors , however , was the enrichment of splicing factors ( see Fig 1A , positive candidates ) . For example , the GO-term “RNA splicing” was found enriched among the candidates with a p-value of 2x10-29 . Among the potential repressors of DSB-derived siRNAs , we found that mRNA 3’-end processing activities were enriched ( e . g . GO-term “mRNA cleavage” p-value 4 . 7x10-14 ) . Although GO-term analyses must be interpreted with caution , this overview is consistent with the hypothesis that DSB-derived siRNAs are generated with a contribution of the mRNA splicing reaction . On the other hand , canonical transcript termination via mRNA cleavage appears to remove the trigger for siRNA generation . The candidates with the GO-term association “splicing” showed a validation success rate comparable to other candidates ( Fig 2A ) . Since the involvement of splicing is reminiscent of a recently proposed model for transposon recognition in the fungus Cryptococcus neoformans [27] , we tested our candidates for their requirement to repress a genomically integrated , endo-siRNA generating high-copy transgene analogous to a previously published system [28] . In the current study , we measured a cell line where the high-copy integrated Renilla luciferase responds more strongly to an impaired RNAi pathway than the firefly luciferase integrated at low copy-number; since identical plasmid constructs were used this reporter system allows for direct comparison with the screening data ( Fig 2B ) . This comparison defined two groups of candidates: The first is required for the generation of DNA break induced as well as high-copy transgene induced siRNAs and comprises , among others , the splicing factors . The second group is specific for DNA double-strand break induced siRNAs and comprises the homologous recombination factors discussed above ( including dCtIP ) as well as RfC . Perturbed mRNA splicing may thus be a common trigger for siRNA biogenesis at DNA double-strand breaks as well as high-copy transgenes . On the other hand , RNAi-mediated depletion of splicing factors could indirectly affect siRNA biogenesis; for example , altered splicing efficiencies could result in lower protein levels of core RNAi factors . Our analysis with the miR-277 perfect match reporter cell line suggested that upon knockdown of the splicing factors , the core RNAi pathway is unaffected and reporter expression is unchanged ( S1 Table ) . In addition , we tested several candidates involved in the splicing reaction for de-regulation of RNAi factors via Western blot and saw no consistent protein level changes for core RNAi components ( Fig 2C ) . We constructed a map of our candidates on the various spliceosome complexes that assemble along the path of a splicing reaction ( Fig 2D ) . Although we identified factors from all complexes , the recovery was particularly prominent among members of the Prp19- and Prp19-related complexes ( 8 out of 16 vs . 26 out of 138 for all spliceosome components , p<0 , 04 χ2-test , see also S3 Fig ) . This complex has a pivotal role in enabling the transitions from the pre-catalytic spliceosome into the catalytic phases . We wanted to confirm the importance of splicing for break-induced siRNA generation by creating DNA breaks at specific positions relative to splice sites . To this end , we employed the cas9-CRISPR nuclease technology and programmed the enzyme to cleave chromosomal sites before or after an intron , then deep sequenced the small RNAs and mapped them back to the cleaved locus . To obtain a quantitative measure of siRNA generation at a given site , we calculated the normalized read density per base pair of the affected transcript upstream and downstream of the cleavage site . If break-derived siRNAs are efficiently generated , then the average reads/bp values are higher between the promoter and the DSB site than in the region following the break [16] . We first targeted a strongly expressed , intron-less gene ( tctp ) for cleavage at three different positions . Although cleavage at each site was detectable ( as judged by a T7 endonuclease assay , S4 Fig ) , we observed siRNA generation that was only slightly above background ( average ratios of before vs . after the cut of 1 . 6–2 . 9 , Fig 3A ) . Analogously , when we targeted the spliced gene CG15098 ( with a similar expression level ) before the first intron , we observed only low levels of siRNA generation ( Fig 3B ) . A cut close to the first intron-exon junction ( 82 nt downstream ) resulted in rather moderate siRNA induction . Cleavage further downstream , however , led to a strong generation of siRNAs with an increasing ratio of reads upstream vs . downstream of the break ( up to 58 . 3 ) as the cut was moved downstream along the gene ( Fig 3B , see also S5 Fig and S6 Fig for detailed traces ) . Thus , splicing of the transcript affected by the DNA break greatly stimulates siRNA generation . We extended our analysis to the CG18273 gene , which is only moderately expressed in S2-cells ( S7 Fig ) . Upon cas9-mediated DSB induction , an siRNA response could be induced here as well . Interestingly , this gene showed moderate , cleavage-independent siRNA coverage in its 3’-portion . Prior to this zone , the DNA-break induced coverage was about 10-fold lower than the coverages we observed in the cleaved CG15098 gene . The strength of the DSB-induced siRNA response correlates thus with the host gene expression level , consistent with the notion that the mRNA transcript contributes the sense strand to dsRNA formation . CG18273 has a short ( 60 nt ) first intron followed by a rather long second exon ( 2375 nt ) . Cleavage within this second exon resulted in siRNA formation , indicating that a single short intron can suffice to trigger the response . Furthermore , when we induced DNA cleavage close to the end of the CG18273 gene ( 4686 nt downstream of the transcription start site ) , we observed siRNA coverage all the way to the start of the transcriptional unit . The DSB-induced small RNA response can thus cover a window of several kbp even in moderately expressed regions . Since splicing is required to trigger the DNA-damage dependent siRNA response , we tested whether a knockdown of splicing factors identified in our screen simply reduces mRNA splicing and thereby diminishes the trigger for siRNA production . We thus determined splicing efficiencies at our CG15098 model locus and the tsr gene ( both show strong expression in our S2 cells ) following depletion of candidates recovered in our screen . After knockdown of the SR protein kinase Doa , the hnRNP protein hrb27C and the spliceosome component l ( 1 ) 10Bb ( the Bud31 homolog from the Prp19-related complex ) , we isolated total RNA and used qRT-PCR ( random hexamer primed ) to quantify the levels of unspliced pre-mRNA and spliced mRNA ( Fig 4 ) . The values were normalized to an amplicon that was internal to one of the exons and thus reported on the total amount of transcript from each locus ( i . e . spliced and unspliced message ) . Overall , we found that the unspliced pre-mRNA did not increase relative to a control knockdown ( Renilla luciferase ) . There was one exception , however: At the short intron in tsr the pre-mRNA became more abundant after RNAi against the spliceosome component l ( 1 ) 10Bb . Even in this case , though , the amount of spliced message was comparable to the total amount of transcript produced , indicating that the majority of transcripts are correctly spliced . This was also seen for all other cases where we compared the level of spliced exon-exon junctions to the total amount of message , indicating that the canonically spliced mRNA accounts for essentially all of the transcripts detected at steady-state . In summary , we conclude that during the time of our knockdown-experiments there is no major change in general splicing efficiency . This argues that upon our experimental RNAi , the splicing factors and spliceosome components we identified did not yet induce major changes in mature mRNA levels . Rather than influencing the general cellular protein content , they may thus limit the signaling events that emanate from splicing reactions / spliceosomes perturbed by a nearby DNA break . We used qRT-PCR to directly test whether a downstream DNA break perturbs progression of the splicing reaction . We induced cas9-mediated DNA cleavage downstream of the third intron of the CG15098 gene , then isolated total RNA and reverse transcribed both nascent and mature mRNA with random hexamer primers . We then interrogated nascent RNA with primers covering an exon-intron junction and mature mRNA with primers spanning an exon-exon junction . The samples were normalized to total CG15098 levels with an amplicon located inside of exon 3 as described above . Control samples were analyzed analogously and we calculated the cut/uncut ratio of each of the amplicons ( Fig 4C ) . Indeed , we found significantly more unspliced RNA at the exon3-intron3 junction when the DNA was cut ( student’s t-test p<0 . 01 , 3 biological replicates ) . A downstream DNA break thus has the potential to stall progression of the splicing reaction at least transiently . Since siRNA generation at DNA breaks depends on transcription levels in Drosophila , the sense strand of the siRNA precursor is most likely the normal transcript originating at the locus . Because upstream introns stimulate siRNA generation ( Fig 3 ) , we asked whether the splicing reaction takes place on the transcript molecule that contributes to the siRNA precursor . To this end , we analyzed the small RNA sequencing data from our cuts in the CG15098 gene in detail . The siRNA coverage started essentially adjacent to the induced breaks and continued relatively uniformly until the beginning of the CG15098 transcription start site ( see e . g . S6 Fig ) . The pattern of siRNA coverage was astonishingly reproducible ( Fig 5A ) , but there was no strong correlation between gene structure and local siRNA coverage ( Fig 5B , calculated based on four biological replicates ) . Intron 2 , but not intron 1 or 3 , showed a somewhat lower coverage than the exons 1–3 ( p = 0 . 016 , two-sided Student’s t-test , n = 4 biological replicates ) . Exon-exon junction spanning siRNAs were essentially absent , consistent with the notion that the siRNAs are not generated by RdRP-activities acting on mature mRNA ( p<10−4 , S8 Fig ) . Furthermore , we found that the coverage of the 3’ intron-exon junctions–the site of the second transesterification reaction—was not diminished; rather , there was a trend towards slightly enhanced coverage relative to exonic reads ( 3’-junction 1: p = 0 . 11 , 3’-junction 2: p = 0 . 06 ) . Clearly , the sense transcript is not fully spliced prior to siRNA generation . If the splicing reaction is stalled after the first catalytic step , then siRNAs covering the 5’ exon-intron junction should be diminished . There was no change for intron 1 and 3 of CG15098 . For intron 2 we observed a reduction of 5’ exon-intron spanning reads when compared with the exonic coverage ( p = 2x10-4 ) , but not when compared with the adjacent intron 2 coverage ( p = 0 , 3 ) . The reduced coverage for intron 2 as well as its 5’ splice junction remained when the DNA break was located at different positions relative to intron 2 ( S9 Fig ) . We thus favor the interpretation that it is an inherent property of this intron ( e . g . sequence-dependent ) . Because upstream splicing stimulates the siRNA response but the reaction is not completed , we conclude that the spliceosome is stalled most likely in a pre-catalytic state when dsRNA formation is triggered . This is also consistent with our qRT-PCR analysis ( Fig 4C ) where we saw an increase of nascent RNA as detected by an amplicon spanning the exon-intron junction , i . e . the site of the first catalytic step , after DNA cleavage .
The precursor of siRNAs in Drosophila is double-stranded RNA ( dsRNA ) , which must be generated through convergent transcription since flies lack an RNA-dependent RNA polymerase . It was previously proposed that the DNA end serves as an initiation site for transcription that produces antisense RNA , followed by pairing with the normal transcript to generate dsRNA [16] . Our screening and validation experiments demonstrate that DNA end processing by the MRN-complex stimulates siRNA generation . Thus , recognition of the damage is independent of , and can precede , siRNA biogenesis . This is consistent with the observation that DNA damage signaling occurs normally in the absence of DNA damage induced siRNAs [14] and with the finding that DNA repair is unaffected when these siRNAs can no longer be made [22] . Our identification of the Drosophila CtIP homolog CG5872 , a nuclease , in the screen further demonstrate that the generation of a 3’-single stranded DNA overhang facilitates the initiation of antisense transcription . Yet , siRNA coverage started adjacent to the break site , arguing that nucleolytic processing of the 5’-strand is not overly extensive . Based on our screening efforts we conclude that spliceosome components are required to trigger an siRNA response and cas9/CRISPR mediated cleavage in the genome revealed that an intron upstream of the break stimulates siRNA generation . The simplest interpretation is that spliceosomes participate in triggering this response independently of the splicing reaction . Consistently , we did not observe major splicing defects during the course of the knock-down experiments ( Fig 4A and 4B ) , we could demonstrate that a downstream cut does slow down transcript maturation ( Fig 4C ) and the siRNA coverage analysis argued for a pre-catalytic stalling event ( Fig 5 ) . What is the mechanism of dsRNA generation at the break ? A straightforward hypothesis is that upon reaching the broken DNA end , the transcribing RNA polymerase II simply turns around and continues transcription of the other DNA template strand , thus forming a hairpin transcript ( “U-turn move” ) . This phenomenon is well known when DNA templates with protruding 3’-ends are transcribed by bacteriophage T7 RNA polymerase [29] . Similarly , free RNA polymerases could spontaneously initiate transcription at the newly formed DNA end; this is also readily observed in vitro with DNA templates that bear a 3’ single-stranded extension and was proposed to occur at DNA double-strand breaks in S . pombe [12] . In both cases , however , one would not predict that the presence of an intron should stimulate the generation of dsRNA; rather , DNA breaks in the intronless gene should have led to a comparable extent of siRNA generation as the ones in the intron-containing gene . Consistently , we had observed that transfected linear PCR products do not trigger an siRNA response [16] . Thus , while we cannot exclude that U-turn transcripts are formed , it is unlikely that they are the source of the majority of the siRNA precursors . It remains possible , however , that association of the spliceosome with the nascent mRNP leads to a remodeling or modification of the RNA polymerase complex , which favors the execution of a U-turn at the DNA end ( see Fig 6 ) . A formal possibility is that non-canonical enzymes are recruited to serve as RNA-dependent RNA polymerases ( RdRP ) acting on the transcript affected by the DNA break . For example , it has been demonstrated that RNA polymerase II can use an RNA template to create a corresponding RNA transcript in the case of human hepatitis delta virus or plant viroid replication [30–33] and that bacteriophage T7 RNA polymerase can replicate short RNA templates [34] . However , this has only been observed for RNAs with a particular secondary structure , while the DNA damage-induced siRNA response appears to be generic . Since we did not detect any exon-exon junction spanning reads in our siRNA coverage analysis , any RdRP-like activity would be limited to the unspliced , nascent transcript . Taken together , we do not consider this a likely scenario . The most parismonious hypothesis is that RNA polymerase stalls at the break , co-transcriptional mRNA maturation is concomitantly delayed and that this induces a signaling event with participation of the spliceosome ( Fig 6 ) . Such a stalled transcript probably leads to a persistent R-loop with a corresponding displaced , single-stranded DNA region . This single-stranded DNA , together with a signal from the spliceosome , could serve as an initiation site for antisense transcription . It is also conceivable that such R-loops may be larger or more persistent if the stalling occurs in the vicinity of an engaged spliceosome without the need for a specific signaling step . This mechanism may also act when R-loops occur independently of a DNA break , consistent with the observation that the splicing factors we identified were also required for the small RNA response triggered by high-copy transgenes . Several previous publications have reported a requirement of splicing factors , but not splicing in general , for small RNA-mediated transcriptional silencing in fission yeast [35 , 36] . Consistent with this , centromeric non-coding transcripts are indeed spliced in S . pombe [36] . However , messages coding for essential fission yeast silencing factors appear to be particularly sensitive to diminished splicing activity , thus leading to reduced silencing efficiency; a similar phenomenon can affect the expression of the Drosophila melanogaster piRNA factor piwi [37 , 38] . It remains a matter of debate whether intron-less , cDNA-based rescue constructs can bypass the silencing defect induced by perturbed splicing in S . pombe [36 , 37] . Other splicing factors , such as smD1 , appear to function independently of their splicing activity during miRNA and siRNA RISC formation [39 , 40] . We now demonstrate that for DNA double-strand break triggered siRNAs , an intron is required upstream ( with respect to transcription ) of the lesion to trigger siRNA formation ( Fig 3 ) . This strongly supports the notion that the splicing process acts in cis during siRNA generation rather than in trans via perturbed splicing of silencing factor mRNAs . Based on our siRNA coverage and qRT-PCR analysis , we propose that the spliceosome is stalled at a pre-catalytic stage ( see the model in Fig 6 ) . The Prp19 complex promotes the transition into the catalytic splicing phases and members of this complex appeared enriched among the spliceosome components we identified ( Fig 2D and S3 Fig ) . Prp19 is of central importance for the splicing-mediated identification of transposon transcripts in C . neoformans [27] , but in this case the reaction was stalled after the first catalytic step . The Drosophila response we describe is conceptually similar to the recently discovered spliceosome-mediated decay in the budding yeast Saccharomyces cerevisiae . There , nucleases are recruited to intron-less genes as a consequence of non-productive association with the spliceosome [41] . Splicing controls piRNA biogenesis in fruit flies; this is a class of small RNAs that represses–together with endo-siRNAs–transposable elements in the germline . Here , spliced transcripts from the so-called master control loci are prevented from entering the piRNA biogenesis pathway [42] . On the other hand , the Tho/TREX complex , normally deposited on RNA as a consequence of splicing , is essential for piRNA biogenesis and must associate through an alternative route with unspliced piRNA precursors [43] . Interestingly , we identified the Tho complex component tho2 as a potential inhibitor of both , DNA damage-induced siRNAs as well as high-copy transgene induced siRNAs ( S1 Table ) . Induction of DNA damage by UV-light revealed that in human cells , splicing is both a sensor for as well as a target of the DNA damage response . This depends on the formation of R-loops due to stalled polymerases and results in pleiotropic splicing changes . The phenomenon bears many parallels to our analysis , but it was limited to transcription-blocking lesions and the authors specifically excluded DNA double-strand breaks as triggers [44] . A consequence of R-loop formation is the generation of a corresponding stretch of displaced , single-stranded DNA . Potentially , this DNA is covered by replication protein A ( RPA ) , [45] , a situation that can trigger DNA damage signaling via direct interaction with Prp19 and accumulation of ATR-interacting protein ( ATRIP ) in mammalian cells [13] . This role of Prp19 appears to be independent of its function during the splicing reaction . Future experiments should therefore address the question whether multiple recruitment platforms rely on the conversion of the Prp19 complex from a regulator of the splicing reaction into a trigger for DNA damage associated signaling events . Due to the focus on siRNA biogenesis in our screen , rather than their downstream function , we cannot directly conclude on the benefits of the break-derived siRNAs for the organism . We previously demonstrated that dcr-2 and ago2 play–at best–an accessory role during DNA repair [22] , but this only addresses the importance of the siRNAs and not their precursor molecules . It is possible that conversion of the stalled transcript into double-stranded RNA or the process of antisense transcription per se are important for DNA repair . For example , this could limit the extent of R-loop formation behind a stalled RNA polymerase ( both temporally and spatially ) or set the optimal length of RPA-covered single-stranded DNA , as was recently proposed for Schizosaccharomyces pombe [12] . Control of R-Loop size during transcription has also been described: In the centromeric regions of fission yeast , a specific arrangement of replication origins and repetitive elements leads to frequent collisions of an RNA polymerase and the replication machinery , followed by the generation of siRNAs in an RdRP-dependent process . In this case , RNA interference prevents excessive R-loop formation by releasing RNA polymerase II and thus fosters genome integrity [46] . Other publications have described a role for RNAi during DNA repair in other organisms [14 , 15 , 18 , 19 , 47] . In the case of the damage-induced Neurospora crassa qiRNAs , DNA repair was even proposed to be the trigger for small RNA production [19] . We note that in the case of our model Gene CG15098 , no repetitive and thus recombination-favoring sequence arrangement was required . Clearly , more results are needed to delineate common and divergent features between these experimental systems . Since the splicing-dependent surveillance mechanism is also important for siRNA generation from high-copy transgenes ( Fig 2B ) , it is possible that the DNA double-strand breaks fortuitously trigger the splicing-dependent mRNA maturation check-point in Drosophila melanogaster , while its primary importance might be to act as a surveillance mechanism against foreign DNA sequences . The list of candidates we identified in our screen provides a valuable resource to further our understanding of the molecular mechanisms and the biological significance of this surveillance mechanism .
Cells were cultured and transfected as described [48] . DNA double-strand breaks were induced in an S2-cell clone that stably expresses the cas9 nuclease [49] with linear U6-sgRNA fusion fragments generated by overlap extension PCR as described [50] . The primer sequences for sgRNA expression are provided in S2 Table . The cell line with stable high-copy integration of lucifcerase reproters was generated by transfecting S2-cells with a mix containing pRB1 ( Renilla luc . ) in a 10-fold excess over pRB2 ( firefly luc . ) and phsHygro . After selection of stable resistance the cells were diluted , clones were picked and clone 4 was used for the experiments . A suspension of DMel cells ( an S2-cell isolate ) in Express5 SFM was prepared ( final concentration of 25 . 000 cells/well ) and 30 μL cell suspension were added to each well of a Greiner 384-well plate containing 5 μL of dsRNA ( conc . 50ng/μL ) . The plates were sealed and incubated at 25 C for 48 h . On each plate of the screen , we included a set of positive controls ( knock-down of Ago2 , Dcr-2 and Ago1 ) as well as a set of negative controls ( knock-down of GFP and DsRed ) . Furthermore , we included a knock-down of Renilla luciferase as a positive control for potential repressors of DSB-induced siRNA generation . Finally , we knocked down the apoptosis inhibitor thread to set thresholds for exclusion of data from predominantly dead cells . For transfection , 0 . 15 μl/well Fugene ( Promega ) and 6 ng/well of pRB1 , 20 ng/well of pRB2 and 30 ng/well of linearized pRB3 were pre-diluted in 14 . 85 μl/well Express5 SFM and incubated at room temperature for 30 min . 15 μL of the thus prepared transfection solution were then dispensed into each well of the assay 384-well plates containing 35 μL of cell suspension with dsRNA . The plates were again sealed and incubated at 25°C . 96 h after knockdown , cells were lysed and luminescence was measured . Fluc and Rluc buffers were added subsequently to the cell lysate and the signals were measured 5 min after the addition of the according substrate with no filter ( Fluc ) and a F485 Coelenterazin filter ( Rluc ) . The primary analysis of the FLuc and RLuc data that was obtained from luminescence measurements was done in R using the cellHTS2 package [51] . The raw data was log-transformed and normalized on the plate median , each channel was analyzed separately and no variance adjustment was applied . The mean RLuc values were plotted against the mean FLuc values and a smoothened curve was fitted to the data using locally weighted polynomial regression ( LOESS ) with a smoothing parameter of 0 . 9 . The thus predicted LOESS values were subtracted from the corresponding mean RLuc values to obtain the LOESS-residuals ( resi ) . The residuals were then used to calculate a z-score . A comprehensive overview of the screening data and controls is given in S2 Fig , the original screening data has been deposited at http://www . genomernai . org . For validation , S2 cells ( lab stock ) were grown in FBS-supplemented Schneider’s medium and the validation screens were conducted in 96-well plates with linearized pRB4 as siRNA inducer . The knockdown constructs used for validation are provided in S1 Table . The oligonucleotide sequences for the generation of the independent dsRNA constructs for validation are provided in S1 Table . In vitro transcription and treatment of the cells was as previously described [16] . A total volume of 3 ml of stable cas9-expressing cell culture was transfected with 1500 ng of U6-sgRNA PCR product as described [50] . Five days after transfection , total RNA was isolated using Trizol and deep sequencing libraries were generated [52] . In certain cases , two sgRNAs were introduced , leading to cas9-mediated cleavage in distinct genes . Up to four libraries were barcoded , combined and analyzed on one lane of an Illumina HiSeq instrument . Subsequently , the obtained reads were sorted , trimmed and mapped to the regions of interest using bowtie [53] . Further analysis was performed with in-house Perl scripts ( available on request ) . Cells were treated with dsRNA to induce knock-down of the corresponding genes under identical conditions as for screening and validation . Five days after induction of knock-down , total RNA was isolated using Trizol and the RNA was reverse transcribed with random hexamer primers and the Superscript-III enzyme . For the analysis of splicing upstream of a DNA break , we transfected the corresponding sgRNA expression PCR products into our stable cas9-expressing cell line , then isolated total RNA 3 days after transfection . Quantitative PCR was carried out on a Biometra TOptical real-time PCR cycler with the Dynamo Flash SybrGreen PCR kit ( Finnzymes ) . Data analysis was done according to the 2-ΔΔCt method [54] , primer sequences for qPCR are given in S2 Table .
|
DNA damage ultimately threatens the integrity of our genome; in addition , it is an impediment for transcription and thus affects cells independently of the mutagenic effects . DNA double-strand breaks in transcribed regions can induce the production of corresponding small interfering RNAs , a class of regulators that is well known for controlling gene expression . The reason why they are generated at sites of DNA damage is not understood , putative roles in DNA repair remain controversial . Thus , the function of DNA damage-induced siRNAs is a major open question . In an unbiased approach , we screened Drosophila cells genome-wide to detect those factors that are required to produce damage-induced siRNAs . In addition to DNA repair proteins , we found that many splicing factors and spliceosome components were necessary . The need for an intron upstream of the break was confirmed by cas9-CRISPR mediated cleavage , followed by small RNA sequencing . A detailed look at the siRNA coverage revealed nonetheless that unspliced transcripts give rise to siRNAs . Hence , the spliceosome must be stalled and signal to the transcription machinery . The new role of damage-induced siRNAs in local RNA surveillance should be considered on par with their putative function during DNA repair .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"gene",
"regulation",
"animals",
"dna",
"transcription",
"animal",
"models",
"dna",
"damage",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"dna",
"drosophila",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"genome",
"complexity",
"gene",
"expression",
"rna",
"splicing",
"insects",
"arthropoda",
"biochemistry",
"rna",
"double",
"stranded",
"rna",
"spliceosomes",
"rna",
"processing",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"non-coding",
"rna",
"computational",
"biology",
"introns",
"organisms"
] |
2017
|
Splicing stimulates siRNA formation at Drosophila DNA double-strand breaks
|
Bacterial signaling systems are prime drug targets for combating the global health threat of antibiotic resistant bacterial infections including those caused by Staphylococcus aureus . S . aureus is the primary cause of acute bacterial skin and soft tissue infections ( SSTIs ) and the quorum sensing operon agr is causally associated with these . Whether efficacious chemical inhibitors of agr signaling can be developed that promote host defense against SSTIs while sparing the normal microbiota of the skin is unknown . In a high throughput screen , we identified a small molecule inhibitor ( SMI ) , savirin ( S . aureus virulence inhibitor ) that disrupted agr-mediated quorum sensing in this pathogen but not in the important skin commensal Staphylococcus epidermidis . Mechanistic studies employing electrophoretic mobility shift assays and a novel AgrA activation reporter strain revealed the transcriptional regulator AgrA as the target of inhibition within the pathogen , preventing virulence gene upregulation . Consistent with its minimal impact on exponential phase growth , including skin microbiota members , savirin did not provoke stress responses or membrane dysfunction induced by conventional antibiotics as determined by transcriptional profiling and membrane potential and integrity studies . Importantly , savirin was efficacious in two murine skin infection models , abating tissue injury and selectively promoting clearance of agr+ but not Δagr bacteria when administered at the time of infection or delayed until maximal abscess development . The mechanism of enhanced host defense involved in part enhanced intracellular killing of agr+ but not Δagr in macrophages and by low pH . Notably , resistance or tolerance to savirin inhibition of agr was not observed after multiple passages either in vivo or in vitro where under the same conditions resistance to growth inhibition was induced after passage with conventional antibiotics . Therefore , chemical inhibitors can selectively target AgrA in S . aureus to promote host defense while sparing agr signaling in S . epidermidis and limiting resistance development .
The global health threat of antibiotic resistant bacterial infections mandates rethinking of how antibiotics are used , how targets for new antibiotics are identified , and how mechanisms for promoting host defense can be enhanced [1] , [2] . In this regard , there is much interest in chemical inhibition of bacterial signaling systems , particularly quorum sensing , because of its regulation of virulence in many medically relevant pathogens where antibiotic resistance is problematic [3] , [4] . While chemical inhibitors of quorum sensing ( QSIs ) have been described in vitro , few have demonstrated in vivo efficacy [5] . Moreover , concerns have been raised about the specificity and selectivity of these compounds [6] as well as the potential for resistance development to quorum sensing inhibition [7] . Therefore , the future of quorum sensing inhibition as a medical strategy to replace or augment conventional antibiotics is uncertain . Of the quorum sensing systems in Gram positive pathogens being targeted for chemical inhibition , the agr operon of Staphylococcus aureus has received noteworthy attention [3] , [8] . This interest derives from its significant medical burden [9] , its known propensity for developing resistance to newly introduced antibiotics [10] , and the failure of all vaccines to date to prevent infection [11] . While chemical inhibitors of agr have been identified [8] , none have demonstrated efficacy in mammalian models of infection . Moreover , none have demonstrated selectivity towards agr signaling in the pathogen S . aureus while sparing agr signaling in the skin commensal Staphylococcus epidermidis , an important contributor to host defense against skin infection [12] . Approximately 90% of S . aureus infections involve skin and soft tissues ( SSTIs ) [9] , [13] and agr is positively associated with human SSTIs [14] , [15] . Moreover , competitive interference with agr signaling is sufficient to abrogate experimental skin abscesses [16] , and we have shown that innate immunity against experimental S . aureus skin infection requires active suppression of agr signaling [17]–[19] . Therefore , we postulated that selective chemical inhibition of agr signaling in S . aureus could promote host defense against SSTIs , providing evidence for limiting conventional antibiotic use in the majority of S . aureus infections . Here we describe a QSI identified in a high throughput screen that selectively inhibited agr signaling in S . aureus , but not in S . epidermidis , by blocking the function of the transcriptional regulator of the operon , AgrA , preventing upregulation of the agr-regulated genes essential for skin infection . It was efficacious in murine models of agr-dependent skin infection without apparent induction of resistance or tolerance after passage in vivo . These data provide proof-of-principle that AgrA transcriptional function in S . aureus can be selectively inhibited to attenuate quorum sensing with minimal toxicity to the bacterium or induction of stress responses observed with conventional antibiotics . Thus , selective AgrA blockade could enhance agr-dependent host defense in the skin while potentially preserving the normal microbiota , limiting resistance induction , and sparing conventional antibiotics for treatment of invasive systemic infections .
The agr quorum sensing operon encodes two promoters [3] , [20]; P2 that drives production of a two component sensor-regulator , AgrC and AgrA , and its autoinducing peptide pheromone ligand , and P3 that drives production of a regulatory molecule RNAIII that together with AgrA is responsible for transcriptional control of approximately 200 genes including multiple virulence factors and metabolic pathways involved in stationary phase growth [15] . P3 also drives P2 providing positive feedback to the production of the receptor ( AgrC ) , the transcriptional regulator ( AgrA ) , and the cyclic thiolactone peptide pheromone ( AIP ) . Critically , the virulence factors most closely associated with human SSTIs , alpha hemolysin ( hla ) , phenol soluble modulins ( PSMs ) , and Panton-Valentine Leukocidin ( PVL ) are agr regulated [14] , [15] . We screened 24 , 087 compounds selected for diversity for inhibition of AIP-induced agr::P3 activation using a reporter strain where P3 drives production of GFP ( ALC1743 ) ( http://pubchem . ncbi . nlm . nih . gov/assay/assay . cgi ? aid=1206&loc=ea_ras ) . We pursued one compound where dose-response experiments using an additional reporter strain indicated that it had minimal impact on exponential phase growth during the 3 hr assay starting at a CFU of 2×107/ml and ending at ∼1×108/ml . It inhibited optimally at 5 µg ml−1 ( 13 . 5 µM ) ( Fig . S1 ) . Termed savirin ( Fig . 1A ) , for Staphylcoccus aureus virulence inhibitor , its molecular weight ( 368 ) and lipophilicity ( XLogP3-3 . 5 ) meet standards for drug development [21] . S . aureus isolates belong to one of four agr alleles depending on variations in AIP ( amino acid sequence and length ) and the cognate receptor , AgrC [3] , [20] . While agr I alleles predominate in human disease , all four can contribute to SSTIs [9] . Therefore , an optimal chemical for agr disruption should work against all agr alleles . Savirin ( 5 µg ml−1 ) inhibited agr::P3 activation in reporter strains of each agr type ( Fig . S2 ) . Therefore , we pursued its efficacy in vitro and in vivo using a strain ( LAC ) of the epidemic methicillin-resistant USA300 clone and the predominant agr group I [15] , [18] , [19] , [22] . We demonstrated by qRT-PCR that savirin ( 5 µg ml−1 ) inhibited both AIP1-induced RNAIII ( Fig . 1B ) and RNAIII produced at a longer time point without addition of exogenous AIP1 ( Fig . 1 C ) with no effect on exponential phase growth ( Fig . 1 D ) . Stationary phase growth was negatively affected by both the genetic deletion of agr ( Δagr ) and by savirin treatment ( Fig . 1D ) consistent with the known role of agr in regulating metabolic pathways of this growth phase in LAC [15] . Importantly , savirin did not significantly affect AIP1-induced RNAIII levels ( Fig . 1E ) or agr-dependent stationary phase growth ( Fig . 1F ) in the related Gram positive member of the skin microbiota , S . epidermidis . Nor did it affect growth of a Gram negative member of the skin microbiota , Pseudomonas aeruginosa [23] ( data not shown ) . Because savirin could have different effects on growth in larger bulk cultures , we evaluated the effects of savirin on both exponential and stationary phase growth in 5 ml cultures diluted to measure OD600 under 0 . 8 . The results were qualitatively similar ( Fig . S3 ) . In addition , savirin did not disrupt membrane integrity ( Fig . S4A ) or membrane potential ( Fig . S4B ) , properties that are altered by antibiotic compounds that could affect agr signaling [24] , [25] and that could be impaired by agr-independent , non-specific toxic effects [6] . To pursue the molecular mechanism by which savirin inhibits agr signaling in S . aureus but not in S . epidermidis , we examined the differences in histidine kinase function and transcriptional control between the two . Because residues within the histidine kinase domain of AgrC that are critical for agr activation are conserved between S . aureus and S . epidermidis [26] , we pursued AgrA function as the molecular target of savirin . We used in silico docking of savirin to the C-terminal DNA binding domain ( AgrAc ) [27] of both S . aureus and S . epidermidis using the online server Swissdock [28] . Savirin docked to AgrAc of S . aureus between Tyr229 , which is adjacent to a residue critical for AgrA folding [29] ( Cys228 ) , and Arg218 near the DNA binding interface with a calculated binding energy of −6 . 1 kcal/mol ( Fig . 2A ) . Notably , mutation of Arg218 to His has been described in clinical isolates with defective agr function [30] . At this position , savirin is within hydrogen bonding distance of the backbone carbonyl of Glu217 and within π-stacking distance of Tyr229 ( Fig . 2A , enlarged view ) . Importantly , this site differed in S . epidermidis where the key Tyr229 is a Phe and His227 is an Asn . Consistent with this , attempts to dock savirin to this site in S . epidermidis were unsuccessful , demonstrating that the DNA binding domain of AgrA is the likely target of savirin . We performed electrophoretic mobility shift assays to prove that savirin blocked the DNA binding function of AgrA . Incubation of purified AgrAc ( 2 µM ) with the high affinity site in P2 and P3 ( 0 . 1 µM ) ( Fig . 2B ) shifted electrophoretic mobility of the FAM labeled nucleotide and increasing concentrations of savirin ( 5–160 µg ml−1 or 13 . 5–432 µM ) vs . vehicle inhibited this shift with an IC50 of 83 µM or 30 . 3 µg ml−1 ( Fig . 2B ) . To prove that AgrA was the target within the pathogen , we constructed a novel reporter strain ( AH3048 ) where plasmid-encoded AgrA constitutively produced without induction drives activation of agr::P3 lux in the absence of the rest of the agr operon , including agrB , agrC , and agrD [31] . As positive controls , we evaluated the ability of diflunisal and 4-phenoxyphenol , compounds published by others as inhibitors of AgrAc DNA binding ability [27] , [32] and both inhibited dose-dependently ( Fig . S5 ) . Additionally , increasing concentrations of savirin ( 0 . 4–6 . 3 µM or 0 . 29–2 . 33 µg ml−1 ) suppressed constitutive luminescence without affecting viability where AIP2 , an inhibitor of non-agrII AgrC signaling [16] , [20] , had no effect on luminescence demonstrating that savirin specifically suppressed AgrA-dependent activation of P3 within the microorganism ( Fig . 2C ) . In comparison to the positive control compounds ( Fig . S5 ) , savirin inhibited luminescence at 6 . 3 µM equivalent to the inhibition of the controls at 100 µM . Moreover , the concentration of savirin required for optimal inhibition of the agrA reporter was equivalent to the concentration that optimally inhibited agr::P3 activation within the pathogen in strain LAC ( 1–5 µg ml−1 ) ( Fig . 1 ) . However , the concentration of savirin required to inhibit in the EMSA assay was much higher due to the excess of AgrAc required for the optimal shift in electrophoretic mobility of the labeled nucleotide . Together , these mechanistic studies indicate that AgrA within S . aureus is savirin's molecular target . We investigated the transcriptional impact of savirin on agr virulence by microarray analysis [15] , [33] and confirmed the results by qRT-PCR and direct measurement of virulence factor function in LAC and in multiple clinical isolates . All of these were performed with the same concentration of savirin , 5 µg ml−1 . The effect of savirin vs . vehicle on AIP1 induced transcription in LAC was compared to the differences between LAC and Δagr LAC . Two hundred and five non-redundant transcripts were different and changed by greater than two fold between LAC and Δagr LAC ( Table S1 ) . Of these , savirin affected 122 or 60% of agr-regulated transcripts by a similar magnitude and direction including downregulation of agr secreted virulence factors ( the majority of transcripts affected ) , transcriptional regulators , and metabolic pathways important for SSTIs [14] , [15] ( Fig . 3A ) . Of the remainder of the potentially agr-regulated transcripts not affected by savirin , the majority were hypothetical or involved in metabolism . In contrast , savirin affected only 5% of the non agr-regulated transcriptome ( Table S1 ) demonstrating selectivity towards agr-dependent transcription . The transcripts upregulated by savirin in both LAC and Δagr LAC could reflect a stress response or be implicated in resistance or tolerance induced by savirin exposure . Of the 19 transcripts upregulated , only 5 with potential roles in drug efflux or resistance were significantly affected ( Table S2 ) . However , this did not include the two most closely implicated with antibiotic resistance , norA ( SAUSA300 0680 ) or mecA ( SAUSA300 0032 ) . Importantly , transcripts were not affected for known stress response genes and the anti-inflammatory exotoxins induced by bactericidal agents [24] , [34]–[36] or agr ablation [25] , [37] ( Table S1 ) . We confirmed by qRT-PCR that savirin inhibited AIP1 induced transcripts for RNAIII and AgrA regulated genes including hla , psm alpha , pvl ( lukS ) , agrA , and agrC , ( Fig . 3B ) . We also confirmed by qRT-PCR that the anti-inflammatory exotoxin set7 was not affected by savirin ( fold increase of vehicle 4 . 66±1 . 47 SEM vs 4 . 66±0 . 7 SEM for savirin , n = 3 ) . Alpha hemolysin activity ( Fig . 3C ) and PMN lysis capacity ( Fig . 3D ) in savirin-treated bacterial supernatants were inhibited as well as lipase and protease activity ( data not shown ) . Moreover , savirin inhibited psm alpha transcripts in clinical isolates of all four agr alleles ( Fig . 4 ) . Additionally , savirin reduced alpha hemolysin activity in supernatants from numerous MRSA and MSSA clinical isolates from multiple sites of infection ( Fig . S6 ) . Given this SMI's selective effect on virulence factor production by multiple isolates , we pursued savirin's in vivo efficacy in two murine models of skin and soft tissue infection . To confirm that savirin inhibited agr signaling in vivo and that it did not affect infection with LAC Δagr , we used an airpouch skin infection model . Mice genetically deficient in the NADPH oxidase ( Nox2−/− ) lack control of agr::P3 activation in this model causing maximal in vivo quorum sensing [17]–[19] . The airpouch in the skin was infected with LAC expressing a fluorescent reporter of agr::P3 activation ( AH1677 ) [19] and savirin ( 10 µg ) was co-administered at the time of infection . Savirin treatment significantly inhibited agr::P3 activation in bacteria from a lavage of the pouch as well as consequential weight loss ( as a measure of morbidity ) and bacterial burden in the pouch lavage and systemically in the kidney ( Fig . 5A ) . Moreover , when C57BL/6 mice were infected with LAC Δagr using the same model , savirin ( 10 µg ) did not affect weight loss or bacterial burden in the pouch lavage or the kidney ( Fig . 5B ) . These in vivo data are consistent with our in vitro data demonstrating that savirin selectively inhibits agr activation and that it has minimal impact on bacteria lacking agr . In addition , we evaluated savirin in an established model of agr-dependent dermonecrotic skin infection in hairless immunocompetent mice [15] , [38] . In this model , clearance of Δagr LAC vs . LAC was enhanced by day 7 ( Fig . S7 ) demonstrating that agr contributes not only to early tissue injury [15] , [38] but to persistence in the skin . Subcutaneous injection of savirin ( 5 µg ) vs . vehicle at the time of infection abrogated abscesses and dermonecrosis ( measured as area of ulceration ) ( Day 1–3 ) ( Fig . 5C ) similarly to the genetic deletion of agr ( Fig . 5C , images ) and prevented early morbidity ( measured as weight loss ) . At day 3 the bacterial burden in the skin abscess was unaffected by savirin treatment ( Fig . 5C ) , indicating that savirin inhibited toxin-induced tissue injury and not bacterial viability at this time point . In contrast , at day 7 savirin treatment promoted bacterial clearance from abscesses and systemically from the spleen ( Fig . 5C ) , replicating the phenotype of agr deletion ( Fig . S7 ) . Because ongoing quorum sensing is likely as the pathogen reaches the required density in discrete locales to accumulate AIP and activate AgrC , we examined the effect of delayed delivery of savirin both in vitro and in vivo . Delayed delivery inhibited RNAIII production in vitro , dermonecrosis in vivo , and promoted bacterial clearance from the skin and systemically from the spleen at day 7 ( Fig . 5D ) . These data indicate that savirin promoted bacterial clearance not by inducing non-specific , agr-independent toxicity in the bacteria , because it did not lead to a reduction in CFU of Δagr at 24 hr ( Fig . 5B ) or of LAC agr+ at 3 days ( Fig . 5C ) , but by rendering LAC less able to survive within the skin leading to clearance by skin host defense mechanisms during the resolution of the infection ( Fig . 5 C , D ) ( Figs . S7 ) . Skin host defense mechanisms are comprised in part of phagocytes , antimicrobial peptides , lytic lipids , and an acidic environment [39]–[41] . Given the time frame that clearance was enhanced , we postulated that savirin treatment of LAC ( 5 µg ml−1 ) but not Δagr LAC would augment killing of the bacteria in vitro by macrophages . As predicted , survival of vehicle treated LAC intracellularly from 1–5 hrs was significantly greater than savirin treated LAC ( Fig . 6 A ) . In contrast , savirin had no effect on the intracellular survival of LAC Δagr ( Fig . 6A ) indicating that savirin's effect on intracellular viability was agr-specific . Because optimal killing of S . aureus within macrophage phagolysosomes requires acidification [42] and because agr regulates transcripts involved in acid resistance [43] ( urease , kdpDE , Fig . 3A ) , we incubated savirin- and vehicle-treated LAC and LAC Δagr at pH 2 . 5 and evaluated viability . As with survival inside macrophages , savirin treatment promoted killing of agr+ but not Δagr bacteria ( Fig . 6B ) . Of interest in both of these assays , the vehicle treated Δagr bacteria were more easily killed compared to the vehicle treated agr+ bacteria indicating that agr contributes to survival inside macrophages and to acid resistance ( Fig . 6A , B ) . However , savirin treatment did not enhance killing by the antimicrobial peptide beta defensin 3 , reactive oxidants , or lytic lipids ( data not shown ) indicating that savirin enhanced killing by some but not all skin defense mechanisms . These data suggest that enhanced killing by macrophages or the acidic environment of the skin may contribute in part to the ability of savirin to promote clearance of agr+ bacteria from the skin . S . aureus has a remarkable propensity for developing resistance or tolerance to antibiotics [10] but whether it would become resistant to inhibition of quorum sensing , as has been postulated for Gram negative bacteria [7] , is unknown . Resistance or tolerance to savirin suppression of quorum sensing could occur by either selecting for the survival of spontaneously arising agr dysfunctional mutants or by stimulating drug efflux necessitating higher concentrations of savirin for efficacy . To be clinically significant , resistance or tolerance induced by repeated exposure should occur in vivo . To address this , we serially passaged LAC with savirin ( 5 µg ) vs . vehicle sequentially through the skin of ten individual mice 24 hrs after infection . We compared this to in vivo passage with sub-inhibitory concentrations of antibiotics known to induce resistance in USA300 strains , erythromycin and clindamycin , because of the genetic expression of ermC [44] . We chose clindamycin as a control because it is used clinically for the treatment of SSTI's and emergence of resistance to clindamycin is clinically important [44] . Passage in vivo with conventional antibiotics induced resistance to killing by clindamycin ( Fig . 7A ) but passage with savirin did not affect its ability to inhibit agr signaling in the savirin passaged bacteria , as measured by AIP1 induction of RNAIII by qRT-PCR , or the dose response of savirin optimal for inhibition of RNAIII production ( 1–5 µg ml−1 ) ( Fig . 7B ) . Equivalent data were obtained with in vitro passage every day for ten days ( Fig . 7C , D ) . To address resistance at the colony level , we plated the in vivo passaged bacteria on milk agar plates where proteolysis is agr-dependent and contributes to colony growth ( Fig . 7E ) . While passage of S . aureus in vitro leads to the production of agr dysfunctional colonies [45] , whether this happens with in vivo passage is unknown . The passaged bacteria were diluted to give 15–20 colonies per plate , spread on plates containing either vehicle or 10 µg ml−1 savirin , and proteolytic and non-proteolytic colonies enumerated at 72 hr . Both antibiotic- and savirin- passaged bacteria plated on vehicle had equally large colonies with clear zones of proteolysis ( ≥1 . 0 mm ) and neither had small non-proteolytic colonies indicative of agr dysfunction ( Fig . 7E , F ) . In contrast , when both the savirin and antibiotic passaged bacteria were plated on savirin containing plates , the majority of the colonies converted to a non-proteolytic phenotype however a small number had zones of proteolysis ≥1 . 0 mm ( Fig . 7E , F ) . These data demonstrated that plating on savirin was able to suppress agr-dependent protease production and that there was no difference between antibiotic- and savirin- in vivo passaged bacteria in their sensitivity to savirin inhibition . In total , these data indicate that under conditions where resistance to growth inhibition can be induced in vivo with a conventional antibiotic used for treatment of SSTI's , savirin exposure did not lead to loss of agr function or tolerance to savirin inhibition of agr function at both the population and colony level .
In this work we have used an SMI as a tool to address many of the concerns raised about the use of quorum sensing inhibitors as therapies or adjuncts for the prevention or treatment of antibiotic resistant bacterial infections [6] , [7] . We identified an SMI in a high throughput screen that inhibited signaling of the agr quorum sensing operon in the medically significant pathogen , S . aureus ( Fig . 8 model ) . We addressed the specificity of the inhibitor for agr signaling in this pathogen , its lack of generalized non-specific , agr-independent toxic effects on the bacterium , its molecular mechanism of action , its selective efficacy in vivo , and the potential for resistance development . If QSIs are to be efficacious for treating bacterial infections , they must work by enhancing host defense against the pathogen rendered either avirulent by the inhibitor or less fit for survival within the host . The evidence that our SMI works this way rather than by some non-specific , agr-independent toxicity on the bacterium in vivo includes: 1 ) its lack of effect on the number of Δagr bacteria 24 hr after infection , 2 ) its lack of effect on the number of agr+ bacteria early ( day 3 ) at the site of skin infection , and 3 ) its lack of effect on macrophage or low pH killing of Δagr bacteria . Moreover , the reduction in CFU observed in our 2 models of SSTIs both at the site of infection and systemically ( 1 . 5–2 . 0 logs ) was similar to that seen with conventional antibiotics tested in a murine surgical wound infection model [46] suggesting that if drugs were developed as QSIs with adequate bioavailability and pharmacokinetic properties that the eventual reduction in bacterial number could approach that seen with currently used antibiotics . Because the majority of S . aureus infections involve skin and skin structures and are dependent on agr signaling in humans and animals [9] , [14] , [15] , limiting antibiotic use in these infections could have a major impact on preserving conventional antibiotics for systemic , life-threatening infections [47] . In this regard , a clinical trial is ongoing which is testing whether treatment of uncomplicated skin abscesses could be limited to incision and drainage without systemic antibiotic use ( NCT00730028 , Uncomplicated Skin and Soft Tissue Infections Caused by Community-Associated Methicillin-Resistant Staphylococcus aureus ) . Our data are consistent with this approach and suggest that a QSI could either substitute for or be used as an adjunct to conventional antibiotics in this setting . Additionally , a QSI could be substituted for antibiotics used prophylactically to prevent wound infections until clinical signs of infection were apparent . Whether a QSI like savirin could be an adjunct with conventional antibiotics for treating systemic infections with or without a biofilm component is a matter of speculation and was not addressed by our studies . In fact , QSI's may have very different clinical utility in Gram negative and positive infections . Because even appropriate antibiotic use drives resistance [1] , any strategy that spares conventional antibiotic use could positively impact resistance development . However , more work is needed in understanding the host defense status of patients presenting with acute bacterial skin infections because the effective use of a QSI is dependent on patients having adequate host defense systems to clear the QSI-treated , less virulent and/or less fit pathogen . Intriguingly , our compound was efficacious in mice lacking the Nox2 phagocyte oxidase , an important component of host defense against S . aureus in humans [17]–[19] , suggesting that agr inhibitors may have efficacy in some patients with impaired host defense systems . More experimental work is required to determine which host defense elements are essential for agr inhibitor efficacy . The potential for resistance development to QSI's has been addressed primarily in Gram negative bacteria ( particularly Pseudomonas aeruginosa ) where QS mutants ( cheaters ) arise during infection by taking advantage of the metabolic effort exerted by QS enabled bacteria for survival [7] . Whether this happens even experimentally in vivo with S . aureus infection is uncertain . Based on studies in Gram negative bacteria [7] , the use of a QSI like savirin could give rise to mutants with a selective advantage over wild-type organisms . However , given the mechanism of action of savirin and its potential binding site in AgrAc , mutants resistant to savirin are most likely to be agr dysfunctional . Mutations in either agrA or agrC do arise in human infection [30] and savirin's potential binding site includes a known mutation in agrA ( Arg 218 to His ) [30] . However , elegant epidemiologic investigation has determined that these arise primarily from colonizing strains prior to the initiation of infection and not spontaneously from agr enabled bacteria during the course of infection [48] . Moreover , these mutants are less fit for transmission between patients [30] , [48] suggesting that even if agr mutants arise with savirin exposure , they are unlikely to have a selective advantage over wild-type bacteria . Importantly , infection with agr mutants is primarily associated with bacteremia in hospitalized patients with impaired host defense systems and not with acute skin infection in immunocompetent individuals [14] , [48] . This information along with our experimental data with in vivo passage in mice suggests that agr inhibitors may not drive the selection of agr mutants in skin infection . However , resistance or tolerance to agr inhibitors could arise by inducing a survival response in the bacteria that leads to upregulation of efflux mechanisms . Our microarray data suggest this as a possibility but neither in vivo nor in vitro passage with savirin resulted in resistance or tolerance to agr inhibition at either the population or colony level under the conditions we used . Currently , it is impossible to predict whether these issues would arise in human infection and whether our method for chronic exposure with in vivo passage in mice actually reflects how skin bacteria would be exposed to a QSI during human infection . The mechanism of action of our SMI suggests that focusing on a site for targeted drug development within the DNA binding domain of the transcriptional regulator AgrA that is different between S . aureus and S . epidermidis , would be optimal for creating an agr inhibitor that spares the important contribution of S . epidermidis to host defense against skin infection [12] , [23] . However , additional work is required to prove that savirin binds directly to the proposed site in AgrAc and to prove that savirin does not affect skin colonization by S . epidermidis . Other investigators have reported compounds that inhibit AgrA DNA binding but whether these compounds would also inhibit in S . epidermidis was not addressed [27] , [32] . Our novel AgrA activation reporter assay could be duplicated using AgrA from S . epidermidis for dual screening of compound libraries for inhibitors of S . aureus but not S . epidermidis AgrA DNA binding function . Using this strategy a drug selective for agr inhibition in S . aureus could be developed with appropriate bioavailability and pharmacokinetic properties to enhance host defense against skin and soft tissue infections while minimizing the impact on normal microbiota and on antibiotic resistance .
All animal experiments were conducted at the AAALAAC accredited Veterinary Medical Unit of the New Mexico Veteran's Affairs Health Care Service in accordance with the applicable portions of the USA Animal Welfare Act as regulated by USDA , the Eighth Edition of The Guide for the Care and Use of Laboratory Animals , and the rules and regulations of the USA Department of Veterans Affairs governing experimental vertebrate animal use . These studies were approved by the NMVAHCS Institutional Animal Care and Use Committee ( Protocol #10-HG-41 ) . Human neutrophils were purchased from AllCells and the source of the neutrophils was anonymous . AIPs1-4 were synthesized by Biopeptide Co . , Inc and stored in DMSO at −80°C . Savirin ( 3- ( 4-propan-2-ylphenyl ) sulfonyl-1H-triazolo [1 , 5-a] quinazolin-5-one , CID#3243271 ) was synthesized by ChemDiv , confirmed purified by HPLC , and stored in DMSO at −80°C . The S . aureus strains used in this study were as follows: USA 300 strain LAC and its agr deletion mutant as described [15] , [37]; ALC1743 ( agr I [agr::P3-gfp] ) and ALC3253 ( Newman [agr::P3-gfp] ) as described [17] , [18]; AH1677 ( agr I LAC [agr::P3-yfp] ) ; AH430 ( agrII 502a [agr::P3-yfp] ) , AH1747 ( agr III MW2 [agr::P3-yfp] ) , AH1872 ( agr IV MN TG [agr::P3-yfp] ) as described [19]; and agr IV clinical isolates ( NRS165 and NRS166 ) were obtained through the Network on Antimicrobial Resistance in Staphylcoccus aureus ( NARSA ) supported under NIAID , NIH contract No . HHSN272200700055C . MRSA and MSSA clinical isolates were provided by Dr . Larry Massie , Pathology Service , NMVAHCS and agr typed by PCR as described [19] . Staphylococcus epidermidis strain #12228 was obtained from ATCC and a Pseudomonas aeruginosa isolate was provided by Dr . Graham Timmins , College of Pharmacy , University of New Mexico . To generate early exponential phase , non-quorum sensing bacteria , frozen stocks were cultured in trypticase soy broth ( TSB ) ( Becton Dickinson ) as described [17] . CFU were determined after washing in PBS/0 . 1% Triton X-100 and sonication to disrupt clumps by plating serial dilutions on blood agar plates . Growth in TSB was measured at OD600 in 96 well plates using a plate reader ( Molecular Devices ) at 37°C with shaking , reading at 30 min intervals for 16 hr . The initial cultures were sufficiently diluted such that the maximal OD600 was confirmed to be within the linear range of the plate reader ( <1 . 25 OD600 ) . Additionally , growth was measured in 5 ml cultures in 50 ml sterile conical tubes with shaking and the OD600 determined on 1∶2 and 1∶4 dilutions of the bacterial cultures to ensure that maximal growth was adequately detected and the OD600 of the diluted samples was under 0 . 8 and clearly within the linear range of the spectrophotometer . A fluorescence-based , high throughput assay was developed to screen 24 , 087 compounds selected for diversity from the Molecular Libraries Small Molecule Repository of the NIH Molecular Libraries Screening Center Network ( summary available at http://pubchem . ncbi . nlm . nih . gov/assay/assay . cgi ? aid=1206&loc=ea_ras ) . Using the Hypercyt flow cytometry sampling platform [49] , a 384 well plate format was used that contained per well 2 . 5×107 early exponential phase ALC1743 containing agr::P3 driving expression of GFP . After incubation for 3 hrs with 100 nM synthetic AIP1 , the induced fluorescence of the bacteria was compared between vehicle controls and compounds in 0 . 2% DMSO . Erythrosin B generated singlet oxygen was used as a positive control to inactivate AIP1 [17] , [19] . Secondary assays included evaluation with a separate reporter strain ALC3253 in 1 ml assays and analysis of viability at 3 hr by CFU determination . Early exponential phase non-fluorescent agr:: P3 reporter strains ( 2×107/ml TSB ) were incubated ( 200 rpm at 37°C ) in polystyrene tubes with broth , 50 nM synthetic AIP , or indicated concentrations of savirin for the indicated time . After incubation , bacteria were centrifuged ( 3000 rpm , 4 minutes , 4°C ) , supernatants decanted , and the pellet washed with PBS/0 . 1% Triton X-100 , fixed with 1% paraformaldehyde containing 25 mM CaCl2 , sonicated , and then evaluated for fluorescence by flow cytometry ( Accuri C6 , Accuri Cytometers , Inc . , Ann Arbor , MI ) . Promoter activation was measured as induction of fluorescence . Quantitative RT-PCR was carried out for transcripts of interest relative to 16S RNA using a probe-based assay as described with minor modifications [18] , [19] . Early exponential phase S . aureus strains and clinical isolates were cultured as indicated in the figure legends . For S . epidermidis , overnight culture supernatant was used as a source of AIP . It was Millipore filtered and diluted 1∶2 with TSB . RNA was isolated and purified using the Qiagen RNA Protect Bacteria Reagent and RNeasy Mini Kit ( Qiagen ) using both mechanical and enzymatic disruption . RNA was purified with silica columns and subjected to DNase treatment to remove contaminating DNA . cDNA was generated using a high capacity cDNA RT kit with an RNAse inhibitor ( Applied Biosystems ) and a Bio-Rad thermocycler . Thermal cycling conditions were as follows: 10 minutes at 25°C , 120 minutes at 37°C , 5 minutes at 85°C , hold at 4°C . Quantitative PCR was performed using an ABI7500 Real-Time PCR system with Taqman Gene Expression master mix , ROX probe/quencher , and appropriate primer sequences ( Applied Biosystems ) . Samples were assayed in triplicate . The data are represented as the fold increase of the transcript relative to 16S compared to the inoculum bacteria . The primer-probe sequences used were as follows: For S . aureus: RNAIII forward primer AATTAGCAAGTGAGTAACATTTGCTAGT , RNAIII reverse primer GATGTTGTTTACGATAGCTTACATGC , RNAIII probe FAM-AGTTAGTTTCCTTGGACTCAGT-GCTATGTATTTTTCTT-BHQ; psmα forward primer TAAG-CTTAATCGAACAATTC , psmα reverse primer CCCCTTCAAATA-AGATGTTCATATC , psmα probe FAM-AAAGACCTCCTTTGTTTGTTA-TGAAATCTTATTTACCAG-BHQ; hla forward primer ACAATTTTAGAGAGCCCAACTGAT , hla reverse primer TCCCCAATTTTGATTCACCAT , hla probe FAM-AAAAAGTAGGCTGGAAAGTGATA-BHQ; pvl-lukS forward primer CACAAAATGCCAGTGTTATCCA , pvl-lukS reverse primer TTTGCAGCGTTTTGTTTTCG , pvl-lukS probe FAM-AGGTAACTTCAATCCAGAATT-TATTGGTGTCCTATC-BHQ-2; 16S forward primer TGATCCTGGCTCAGGATGA , 16S reverse primer TTCGCTCGACTTGCATGTA , 16S probe FAM-CGCTGGCGGCGTGCCTA-BHQ; agrA forward primer CTACAAAGTTGCAGCGATGGA , agrA reverse primer TGGGCAATGAGTCTGTGAGA , agrA probe FAM-AGAAACTGCACATACACGCT-BHQ; agrC forward primer AAGATGACATGCCTGGCCTA , agrC reverse primer TGTGCACGTAAAATTTTCGCAG , agrC probe FAM- TGGTATCGAGAATCTTAAAGTACGTG-BHQ; and set7 forward primer ACGGAAAAACCAGTTCATGC , set7 reverse primer GCTTATCTTTGCCAATTAAAGCA , set7 probe FAM-CAGGTTATATCAGTTTCATTCAACCA-BHQ . For S . epidermidis: 16S forward primer TACACACCGCCCGTCACA , 16S reverse primer CTTCGACGGCTAGCTCCAAAT , 16S probe FAM-CACCCGAAGCCGGTGGAGTAACC-BHQ; and RNAIII forward primer ACTAAATCACCGATTGTAGAAATGATATCT , RNAIII reverse primer ATTTGCTTAATCTAGTCGAGTGAATGTTA , RNAIII probe FAM-ATTTGCTTAATCTAGTCGAGTGAATGTTA-BHQ . Membrane integrity was measured as described using propidium iodide [25] . LAC was cultured overnight ( 18 hr ) in RPMI supplemented with 1% casamino acids in the presence of savirin ( 5 µg ml−1 ) or vehicle control . The cultures were washed by centrifugation and the pellet resuspended in PBS supplemented with 1% BSA . The samples were set to an OD600 of 0 . 4 and an aliquot was heat killed ( 90°C for 10 minutes ) to serve as a positive control . Samples ( 50 µl ) were mixed with 1 ml PBS/1% BSA containing propidium iodide . Membrane damage was determined by measuring bacterial fluorescence by flow cytometry ( Accuri C6 ) . Membrane potential was measured using the BacLight Membrane Potential Kit ( Molecular Probes ) following the manufacturer's recommendations . Membrane potential in this assay is based on the shift between the green fluorescence of DiOC2 to red in the cytosol of bacteria with higher membrane potential . The proton ionophore CCCP was used as a positive control for disrupting membrane potential . LAC was cultured with 50 nM AIP1 for 5 hr in TSB in the presence of savirin ( 5 µg ml−1 ) or vehicle control . After diluting into TSB , the bacteria were incubated with 30 µM DiOC2 in the dark for 16 min prior to analyzing by flow cytometry ( Accuri , C6 ) . Measurements from both the red and green channels were taken and data presented as a ratio of red channel divided by the green channel to reflect the shift to greater change in membrane potential . Savirin ( PubChem ID SMR000016143 ) was docked onto the C-terminal domain of AgrA of S . aureus AgrAc ( residues 137–238 with an initiator methionine ) deposited in the Protein Data Bank ( PDB ) accession number 3BS1 [50] using the online server SwissDock ( http://www . swissdock . ch ) [28] . The docking origin was set near Val235 with a search area of 10 Å in all directions and allowing for flexible side chains within 3 Å of the ligand . A model of the S . epidermidis AgrA DNA binding domain was prepared by threading the amino acid sequence ( UniProt database accession number Q84FX9 ) onto the structural coordinates of the S . aureus protein ( PDB 3BS1 ) using the I-TASSER server ( http://zhanglab . ccmb . med . umich . edu/I-TASSER/ ) . Savirin docking to S . epidermidis AgrAc was performed as described above for S . aureus AgrAc with the origin set to the Cα atom of Phe229 . Structural images were generated using PyMOL ( PyMOL Molecular Graphics System , v . 1 . 5 . 04 , Schrödinger , LLC ) . E . coli expressing the 6X-histidine tagged C-terminal DNA binding domain of AgrA ( AgrAC ) from S . aureus isolate Newman was provided by Dr . Chuan He ( University of Chicago , Chicago , IL , USA ) . Expression and purification of AgrAC was carried out as previously described with minor modifications [29] . Briefly , AgrAC expressing E . coli were grown in Terrific broth to an OD600 of 0 . 6 and induced with 1 mM isopropyl β-D-1 thiogalactopyranoside overnight at room temperature . Harvested cells were flash frozen in liquid nitrogen , thawed and lysed using lysozyme and sonication . Soluble AgrAC was affinity purified using Talon Superflow Metal Affinity Resin ( Clonetech ) followed by gel filtration on a Superdex S200 column ( GE Healthcare ) . Tris ( 2-carboxyethyl ) phosphine ( TCEP ) at 1 mM was used as a reducing agent throughout purification . Protein was stored at −80°C in PBS , 20% glyercol , 5 mM DTT , 1 mM TCEP and 1 mM MgCl2 . Electrophoretic mobility shift assays ( EMSA ) using purified AgrAC ( 2 µM ) were performed as described [27] using a 16 base pair DNA duplex probe ( 0 . 1 µM ) containing the high affinity LytTR binding site present in both agr P2 and P3 [27] . It was synthesized with a 3′ 6-fluorescein ( FAM ) to facilitate detection ( Integrated DNA Technologies , USA ) . Samples including AgrAC , DNA probe , vehicle and/or savirin ( 5–160 µg ml−1 or 13 . 5–432 µM ) were loaded in Tris-acetate-EDTA ( TAE ) buffer containing 10 mM dithiothreitol . Assays including the 16 bp probe were run with 10% native polyacrylamide gels . An AgrA-dependent lux reporter strain , AH3048 , was generated by transforming S . aureus Δagr strain ROJ48 [31] with pCM63 [51] . Plasmid pCM63 consists of the agrA gene cloned into plasmid pEPSA5 , which placed transcription of agrA under the control of the xylose-inducible Tx5 promoter . To construct plasmid pCM63 , the agrA gene was PCR amplified from AH1263 genomic DNA using primers CML609 ( 5′-GTTGTTGAATTCCCATAAGGATGTGAATG-3′ ) and CLM610 ( 5′-GTTGTTTCTAGACTTATTATATTTTTTTAACGTTTCTCACCG-3′ ) , the PCR product was digested with EcoRI and XbaI , and ligated into similarly digested pEPSA5 . Preliminary experiments demonstrated that light production by AH3048 increased in a xylose dose-dependent fashion , without impacting growth , up to a xylose concentration of 0 . 25% . For testing the impact of savirin on light production , AH3048 cultures were not induced with xylose because the constitutive level of agrA transcription from pCM63 was sufficient for luminescence induction . An overnight culture of AH3048 grown in TSB with 10 µg ml−1 chloramphenicol ( for plasmid maintenance ) was used to inoculate ( at 1∶500 dilution ) TSB containing antibiotic in 96-well microtiter plates ( Costar 3603 ) at 200 µl per well . A 2-fold serial dilution series of savirin ( 0 . 4–6 . 3 µM or 0 . 29–2 . 33 µg ml−1 ) was used and the concentrations were tested in quadruplicate . Microtiter plates were incubated at 37°C with shaking ( 1000 rpm ) in a Stuart SI505 incubator ( Bibby Scientific , Burlington , NJ ) with a humidified chamber . Luminescence and OD600 readings were recorded at 30 min increments using a Tecan Systems ( San Jose , CA ) Infinite M200 plate reader . Maximal light production occurred after 6 hrs of growth . As a specificity control , a 2-fold dilution series ( 0 . 5 nM to 1000 nM ) of AIP-2 ( Anaspec , Fremont , CA ) was tested in quadruplicate , as well as 12 control wells containing vehicle ( DMSO ) . As positive controls , two compounds demonstrated by others to inhibit AgrAc in EMSA assays , diflunisal and 4-phenoxyphenol ( Sigma ) [27] , [32] , were evaluated for luminescence inhibition in the same assay at concentrations from 1 . 56–100 µM . To compare the transcript levels of LAC and the Δagr mutant in the presence or absence of savirin ( 5 µg ml−1 ) , the bacteria were grown for 5 hr in TSB with 50 nM AIP1 or an equivalent amount of DMSO as the vehicle control and processed for microarray analysis as described [33] . The comparisons were LAC vehicle vs . LAC savirin , Δagr vehicle vs . Δagr savirin , and LAC vehicle vs . Δagr vehicle , n = 3 . The bacterial RNA was purified as described [32] . Samples were hybridized to a custom Affymetrix GeneChip ( RMLchip7 ) that contains all open reading frames of the USA300 genome . Samples were scanned using Affymetrix 7Gplus GeneChip scanner according to standard GeneChip protocols with the image files converted using GeneChip Operating Software ( GCOS v1 . 4 ) . The data were quantile-normalized and a 3-way ANOVA with multiple test correction using the false discovery rate ( p<0 . 05 ) was performed using Partek Genomics Suite software ( Partek , inc . v6 . 5 ) . These data were combined with fold change values , signal confidence ( above background ) , and call consistency ( as a percent ) as calculated using custom Excel templates to generate final gene lists for each comparison . The microarray data were deposited in the Gene Expression Omnibus ( GEO ) database ( http://www . ncbi . nlm . nih . gov/projects/geo/ ) under the accession number GSE52978 . All microarray data are MIAME compliant . Alpha hemolysin activity was measured in 0 . 45 µm filtered cultured supernatant standardized by OD600 after bacterial strains were grown overnight in TSB in the presence or absence of savirin ( 5 µg ml−1 ) . The assay was performed using rabbit erythrocyte lysis as described [18] . One unit of hemolytic activity was defined as the amount of bacterial supernatant able to liberate half of the total hemoglobin from the erythrocytes and expressed as HA50 . The ability of secreted toxins to lyse human neutrophils was determined by LDH release . Overnight supernatant from MRSA agr group I clinical isolate #32 generated with either savirin ( 5 ug ml−1 ) or DMSO vehicle control was 0 . 45 µm filtered , stored at −80°C , and thawed on ice . Human neutrophils ( AllCells ) were washed twice in saline to remove EDTA , suspended in RPMI with 10 mM HEPES and 1% HSA , and assessed for viability by Trypan blue staining ( >97% ) . The experiment was run in triplicate and each tube contained 3×106 neutrophils in 100 µl RPMI to which was added 100 µl of either RPMI , TSB diluted 1∶5 or 1∶10 in RPMI , or treated supernatants diluted 1∶5 or 1∶10 in RPMI . PBS with 0 . 1% Triton-X100 ( 100 µl ) was used for 100% lysis . Tubes were incubated at 37°C in a 5% CO2 incubator for 1 and 2 hours . At each time , the tubes were centrifuged at 13 , 000 rpm , at 4°C , for 5 minutes . Cell free supernatant ( 100 µl ) was transferred to a micotiter plate and immediately processed for LDH according to the Cytotoxicity Detection Kit ( Roche ) . A blank was created for each plate with 10%TSB in RPMI . The data are depicted as the percentage of total lysis after correction for LDH release stimulated by media alone . For all in vivo experiments , savirin was solubilized at 1 mg ml−1 in 0 . 5% hydroxypropyl methylcellulose ( Sigma ) in endotoxin-free sterile water made 3 . 0 mM in NaOH with cell culture tested 1 N NaOH ( Sigma ) , and put through a 0 . 22 µM filter ( Millex-GV ) . The vehicle control was the HPMC used to solubilize the savirin . Sample sizes were determined by preliminary experiments to determine the number of mice required to observe significance . Dermonecrosis model: SKH1 hairless immunocompetent mice ( ≈8–16 wk , ≈26–34 g , male ) were obtained from Charles Rivers ( Wilmington , MA ) . At Day 0 , early exponential phase bacteria ( 4×107 ) washed in sterile normal saline were injected concurrently with savirin ( 5 µg ) or vehicle in 50 µl subcutaneously into the flank using a 3/10 cc insulin syringe with a 28 ½ gauge needle ( Becton Dickinson ) . For delayed delivery , 10 µg savirin was administered 24 and 48 hr after infection in 50 µl . The animals were divided into two groups to have equivalent mean abscess sizes prior to administering drug or vehicle . Abscess area ( maximal on Day 1 ) and ulcer area ( necrosis optimal on Days 3–4 ) were measured with calipers as described [15] , [38] and recorded daily in addition to weight loss . The slightly raised abscess area ( mm2 ) was calculated from the equation ( π/2 ) [ ( length of the abscess ) × ( width of the abscess ) ] . The flat ulcer area ( mm2 ) was calculated from the equation ( length of the ulcer ) × ( width of the ulcer ) or alternatively from digital images using Adobe Photoshop standardized to a micrometer with equivalent results . On Day 3 or Day 7 , the mice were euthanized using isoflurane inhalation . The abscess/ulcer area was excised ( 1 . 5 cm2 ) and the spleens removed . Tissues and spleens were placed in 1 ml of HBSS/0 . 1% HSA in a bead-beating tube containing sterile 2 . 3 mm beads ( Biospec ) and were processed for bacterial CFU by homogenizing the spleens in a bead beater , diluting all samples 1∶10 in 1 ml PBS/0 . 1%Triton , sonicating , and plating serial dilutions on blood agar as described [17]–[19] . Airpouch model: age matched Nox2−/− male mice ( Jackson Labs ) or C57BL/6 male mice ( Charles Rivers ) were infected with either 2×107 early exponential phase non-fluorescent AH1677 bacteria ( Nox2−/− ) or 5×107 LAC Δagr ( C57BL/6 ) into an air pouch generated by the injection of 5 ml of air subcutaneously as described [17]–[19] . Savirin ( 10 µg ) vs . vehicle in 50 µl was injected into the pouch at time 0 . After 24 hours , weight loss was determined , the air pouch was lavaged with HBSS/0 . 1% HSA and the kidneys removed . The bacteria in the lavage were analyzed by flow cytometry for promoter activation ( AH1677 ) and both the lavage and kidneys processed as above for CFU determination . Early exponential phase LAC+50 nM synthetic AIP1 or Δagr LAC ( 2×107/ml TSB ) were incubated for 5 hr at 37°C with shaking ( 200 rpm ) in the presence of savirin ( 5 µg ml−1 ) or vehicle control . Bacteria were opsonized ( 1×108/ml ) with rabbit IgG anti-Staphylcoccus aureus ( Accurate Antibody YVS6881 ) ( 100 µg/ml ) in phenol red-free Dulbecco's Modified Eagle Media , DMEM , containing 4 . 5 g/L D-glucose/2%Hepes+1% FCS ) . The experiment was performed in triplicate . Murine macrophage RAW264 cells ( 5×106 ) in 250 µl of DMEM+2% FCS were combined with 5×106 opsonized bacteria in 250 µl of DMEM+1% FCS ( MOI 1∶1 ) in sterile polystyrene 12×75 mm tubes , centrifuged briefly to initiate contact , and incubated for 1 hr at 37°C in 5% CO2 . The infected cells were treated with lysostaphin ( Sigma ) ( 2 µg/ml for 15 min ) to kill extracellular bacteria and then washed and suspended in fresh media . Half of the samples were incubated for an additional 4 hrs . To determine the intracellular CFU at 1 and 5 hr , the relevant cells were centrifuged , suspended in PBS/0 . 1% Triton-X-100 and sonicated to disrupt cells and dilutions plated on blood agar . The cell line was tested for Mycoplasma sp . contamination by PCR ( Life Technologies ) . Early exponential phase LAC+50 nM synthetic AIP1 or Δagr LAC ( 1×108/ml DMEM , 4 . 5 g/L D-glucose/2%Hepes ) were incubated for 5 hr at 37°C with shaking ( 200 rpm ) in the presence of savirin ( 5 µg ml−1 ) or vehicle control . Bacteria were centrifuged , washed , resuspended in DMEM/2%Hepes acidified with either HCl to pH 2 . 5 or 10 µg ml−1 linoleic acid ( Sigma ) and incubated for the indicated times . Dilutions were plated on sheep blood agar to determine the residual viability . In vitro data were analyzed by the two-tailed Student's t-test or two way measures ANOVA as indicated in figure legends . In vivo data were analyzed by the two-tailed Mann-Whitney U test for non-parametrics . All evaluations were conducted using GraphPad Prism v . 5 . o and results were considered significantly different with p<0 . 05 .
|
New approaches are needed to lessen the burden of antibiotic resistant bacterial infections . One strategy is to develop therapies that target virulence which rely on host defense elements to clear the bacteria rather than direct antimicrobial killing . Quorum sensing is a bacterial signaling mechanism that often regulates virulence in medically relevant bacterial pathogens . Therefore , drugs that inhibit quorum sensing can promote host defense by rendering the pathogenic bacteria avirulent and/or less fit for survival within the host . Our work addressed this strategy in the pathogen Staphylococcus aureus which is the major cause of acute bacterial skin and soft tissue infections . We conducted a high throughput screen to identify compounds that could inhibit signaling by the quorum sensing operon , agr . We found a compound that we termed savirin ( S . aureus virulence inhibitor ) that could inhibit signaling by this operon . The drug helped the innate immune system in animals to clear bacteria that express this operon without affecting clearance of bacteria that do not have this operon . We addressed the mechanism of action of this compound and whether resistance or tolerance to this compound would likely develop . Our data indicate for the first time that host defense against S . aureus skin infections can be enhanced by chemical inhibition of agr-mediated quorum sensing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"skin",
"infections",
"dermatology",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences"
] |
2014
|
Selective Chemical Inhibition of agr Quorum Sensing in Staphylococcus aureus Promotes Host Defense with Minimal Impact on Resistance
|
The Gram-negative bacterium Xanthomonas euvesicatoria ( Xe ) is the causal agent of bacterial spot disease of pepper and tomato . Xe delivers effector proteins into host cells through the type III secretion system to promote disease . Here , we show that the Xe effector XopAU , which is conserved in numerous Xanthomonas species , is a catalytically active protein kinase and contributes to the development of disease symptoms in pepper plants . Agrobacterium-mediated expression of XopAU in host and non-host plants activated typical defense responses , including MAP kinase phosphorylation , accumulation of pathogenesis-related ( PR ) proteins and elicitation of cell death , that were dependent on the kinase activity of the effector . XopAU-mediated cell death was not dependent on early signaling components of effector-triggered immunity and was also observed when the effector was delivered into pepper leaves by Xanthomonas campestris pv . campestris , but not by Xe . Protein-protein interaction studies in yeast and in planta revealed that XopAU physically interacts with components of plant immunity-associated MAP kinase cascades . Remarkably , XopAU directly phosphorylated MKK2 in vitro and enhanced its phosphorylation at multiple sites in planta . Consistent with the notion that MKK2 is a target of XopAU , silencing of the MKK2 homolog or overexpression of the catalytically inactive mutant MKK2K99R in N . benthamiana plants reduced XopAU-mediated cell death and MAPK phosphorylation . Furthermore , yeast co-expressing XopAU and MKK2 displayed reduced growth and this phenotype was dependent on the kinase activity of both proteins . Together , our results support the conclusion that XopAU contributes to Xe disease symptoms in pepper plants and manipulates host MAPK signaling through phosphorylation and activation of MKK2 .
Plant immunity against microbial pathogens relies on a complex detection and signaling network [1] . A first line of plant immune responses is activated by cell surface-exposed pattern recognition receptors ( PRRs ) that detect broadly conserved pathogen molecules ( pathogen/microbe-associated molecular patterns , PAMP/MAMPs ) [2] . Activation of PRRs initiates downstream signaling events that lead to the production of reactive oxygen species , stimulation of mitogen-activated protein kinase ( MAPK ) cascades , defense gene induction , release of ethylene , and callose deposition at the plant cell wall [3 , 4] . These host responses limit the growth of a large number of potential pathogens and are referred to as pattern-triggered immunity ( PTI ) . Host-adapted pathogens overcome PTI through the activity of effector proteins that are targeted to the plant apoplast or delivered into the host cytoplasm [5] . To cope with these pathogens , plants have evolved other types of receptors known as resistance ( R ) proteins that specifically recognize effectors or their activity [6] . R proteins activate effector-triggered immunity ( ETI ) that consists of defense responses similar to PTI , but more robust and often accompanied by a localized cell death known as the hypersensitive response ( HR ) [7] . Mitogen-activated protein kinases ( MAPKs ) cascades play a fundamental role in plant immunity and are involved in both PTI and ETI signaling [8] . The tomato MAPKKK MAP3Kα and MAP3Kε were found to participate in signaling pathways that mediate elicitation of the ETI-associated HR in N . benthamiana plants , and to be required for disease resistance to bacterial pathogens in tomato [9 , 10] . The MEK2 MAPKK was identified as a central regulator of the HR elicited upon detection of effectors by several R proteins in N . benthamiana plants , and as required for tomato disease resistance to Pseudomonas and Xanthomonas bacteria [11] . Epistasis analysis revealed that MEK2 acts downstream of both MAP3Kα and MAP3Kε and upstream of the SIPK and WIPK MAP kinases [9 , 10 , 12] . Notably , SIPK and WIPK , and their respective Arabidopsis homologs MPK6 and MPK3 , are also important regulators of PTI [13 , 14] . In line with these findings , the Arabidopsis MKK4 and MKK5 , which are the MAPKKs upstream of MPK6 and MPK3 , were also shown to participate in PTI signaling [13] . Many Gram-negative plant pathogenic bacteria utilize a type III secretion system to deliver effector proteins into the host cells [15] . Type III effector proteins contribute to bacterial virulence by subverting plant signaling pathways , suppressing immune responses , and modulating host metabolism and hormone signaling [16 , 17] . MAPK cascades have emerged as important targets of type III effectors of plant and mammalian bacterial pathogens [18] . For example , Yersinia pestis YopJ interferes with the activation of immune responses in mammalian cells by inhibiting phosphorylation of MAPKK6 through acetylation of Ser and Thr residues in the activation loop of the kinase [19] . The Salmonella phosphothreonine lyase effector SpvC irreversibly removes a phosphate from ERK1/2 MAPK to downregulate cytokine release from infected cells [20] . Several Pseudomonas syringae effectors were found to suppress immunity in Arabidopsis by interfering with the activity of components of MAPK cascades: HopAI1 encodes a phosphothreonine lyase that irreversibly removes a phosphate from MPK3 and MPK6 thereby suppressing PTI activation [21] . HopF2 inhibits PTI through inactivation of MKK5 , the upstream MAPKK of MPK3 and MPK6 , by ADP-ribosylation [22] . Finally , AvrB enhances plant susceptibility by promoting phosphorylation and activation of MPK4 , which perturbs hormone signaling to the benefit of the bacterium [23] . The Gram-negative bacterium Xanthomonas euvesicatoria ( Xe ) is the causal agent of bacterial spot disease in pepper and tomato plants [24] . Xe bacteria penetrate into plant tissues through wounds and stomata , proliferate and colonize the apoplast of the aerial parts of the plants , and cause the appearance of water soaked lesions that develop into necrotic black spots . The ability of Xe to cause disease largely depends on the type III secretion system . To date , the pool of known Xe effectors includes approximately 35 proteins mostly identified in the 85–10 strain [25–27] . Biochemical activity and cellular targets have been elucidated only for a few Xe effectors . The XopD effector is a SUMO protease that alters host transcription to suppress hormone signaling [28] . The XopN and XopQ effectors target host 14-3-3 proteins to suppress PTI and ETI signaling , respectively [29 , 30] . The XopJ effector causes degradation of a proteasome subunit to suppress salicylic acid-mediated defense and protein secretion [31–33] . By a machine learning approach applied to the Xe strain 85–10 , we have recently identified XopAU as a type III secreted effector and demonstrated its translocation into cells of pepper leaves [26] . XopAU is conserved in multiple Xanthomonas spp . and in a few Acidovorax spp . , and encodes a putative serine/threonine protein kinase [26] . The xopAU gene from Xe 85–10 contains a plant-inducible promoter ( PIP ) box and expression of its homolog from Xanthomonas citri was found to be regulated by the HrpG/HrpX-regulon , which controls transcription of genes encoding structural components of the type III secretion system and some effector genes [34] . Here , we investigated XopAU molecular properties and virulence function . We found that XopAU is a catalytically active protein kinase that contributes to the development of disease symptoms in susceptible plants . In addition , we identified the MAPKK MKK2 as a binding partner and direct substrate of XopAU phosphorylation . Moreover , by genetic and functional analysis we provide evidence that MKK2 is required for the XopAU molecular function .
XopAU is a type III secreted effector originally identified in the Xanthomonas euvesicatoria strain 85–10 ( Xe ) [26] . Homologs of the effector are present in multiple species of the Xanthomonas genus and in Acidovorax spp . Promoter regions of all xopAU homologs contain a PIP box motif ( S1 Table ) indicating that their expression is controlled by the HrpG/HrpX regulon [35] . To determine evolutionary relationships between xopAU homologous genes , a representative of each Xanthomonas species encoding xopAU and an Acidovorax avenae homolog were used to construct a phylogenetic tree ( Fig 1A , S1 Fig and S1 Table ) . xopAU homologs classified into two allelic groups and their phylogenetic relationships correlated to the relationships among the corresponding Xanthomonas species that were deduced by a sequence comparison of the gyrB phylogenetic marker gene [36] ( Fig 1 and S1 Table ) . This correlation suggests that the two alleles were transmitted vertically after their acquisition in parental strains . xopAU homologs of group 1 share a low degree of sequence similarity to group 2 homologs , have a different GC content ( group 1 , 62 . 9%-64 . 3%; group 2 , 55 . 2%-55 . 8% ) , and a distinct genomic location ( S1 Table; S2 Fig ) implying that the two xopAU alleles were independently acquired by Xanthomonas spp . The Xanthomonas species containing the group 1 xopAU allele correspond to a complete clade in the Xanthomonas genus ( Fig 1A ) [37] . Conversely , Xanthomonas species containing the group 2 xopAU allele ( X . fragariae and X . gardneri ) are members of a clade that also includes the X . arboricola species [37] , which does not encode a xopAU allele . The borders of the xopAU deletion in the genome of a X . arboricola strain are shown in S2B Fig . Analysis of the NCBI conserved domains database revealed that Xe XopAU contains a putative protein kinase domain at the C-terminus ( 203–493 amino acids ) , which includes kinase subdomains I-XI and the majority of the residues that are nearly invariant throughout the kinase superfamily [38] ( S3 Fig ) . Notably , the kinase nearly invariant residues identified in Xe XopAU are also conserved in its homologs from other Xanthomonas strains . To test whether XopAU is a catalytically active protein kinase , it was expressed in E . coli as a glutathione S-transferase ( GST ) fusion , purified and assayed for kinase activity in the presence of [γ-32P]ATP . The GST-XopAU fusion was able to autophosphorylate in vitro and its activity was abolished by the introduction of an alanine substitution at the conserved lysine ( K240 ) of the putative ATP binding site ( Fig 1C ) . Together , this analysis revealed that the xopAU gene is conserved in the genome of numerous Xanthomonas species and XopAU of the Xe strain 85–10 is a catalytically active protein kinase . To examine whether XopAU from Xe causes detectable phenotypes in planta , the effector was fused to a His tag ( His-XopAU ) and transiently expressed via Agrobacterium under the control of an estradiol inducible system in leaves of the Xe non-host plant N . benthamiana . At 24–48 hours after estradiol application , cell death was visible in leaf areas expressing His-XopAU and confirmed by a higher ion leakage than in leaf areas infiltrated with Agrobacterium carrying an empty vector ( Fig 2A and 2B ) . No cell death or ion leakage was observed without estradiol application . Furthermore , the His-XopAUK240A kinase deficient variant did not induce a visible cell death and ion leakage ( Fig 2A and 2B ) indicating that XopAU kinase activity was required for this phenotype . Expression of wild-type and kinase deficient His-XopAU variants in the infiltrated areas after estradiol application was confirmed by Western blot analysis ( Fig 2C ) . Because cell death is a typical immune response triggered by pathogens in resistant plants , we hypothesized that XopAU activates immune signaling . To test this hypothesis , we transiently expressed His-XopAU in N . benthamiana leaves and monitored phosphorylation of MAP kinases and accumulation of pathogenesis-related ( PR ) proteins , which are additional phenotypes associated with plant immunity [8 , 39] . Phosphorylation of MAP kinases was assessed by using antibodies against the phosphorylated form of mammalian MAP kinases of the ERK family that recognize also phosphorylated plant MAP kinases . Accumulation of PR proteins was monitored with antibodies against the tobacco PR-2 and PR-3 isoforms . As shown in Fig 2C , MAP kinase phosphorylation and PR protein accumulation were induced in leaves of N . benthamiana plants expressing His-XopAU , but no induction was observed in leaves expressing the kinase deficient variant His-XopAUK240A . To examine whether XopAU induces immune responses in Xe host plants , wild-type and kinase deficient His-XopAU variants were transiently expressed via Agrobacterium in leaves of the pepper line ECW30R and tomato line Hawaii 7981 . Similar as in N . benthamiana leaves , expression of wild-type but not kinase deficient His-XopAU in these plants induced a cell death at 48–72 hours after estradiol application ( Fig 2D ) . In pepper leaves , cell death was accompanied by enhanced MAP kinase phosphorylation and higher accumulation of the PR-3 protein , but not of PR-2 ( Fig 2E ) . Together , these results suggest that expression of XopAU via Agrobacterium activated immune signaling in Xe host and non-host plants . To assess whether activation of immune signaling by XopAU is caused by recognition of the effector by an R protein , we tested if silencing of early components of ETI signaling affects XopAU-mediated cell death . The genes silenced in these experiments were EDS1 and NDR1 , which are required for ETI mediated by multiple R genes of the TIR-NBS-LRR and CC-NBS-LRR class , respectively , and RAR1 , which is required for ETI mediated by multiple R genes of different structural classes [40] . Virus-induced gene silencing ( VIGS ) techniques based on the tobacco rattle virus ( TRV ) vector were used to silence the genes in N . benthamiana plants [41] . Four weeks after infection of plants with the TRV vector carrying fragments of the genes to be silenced , transcript levels of NDR1 , EDS1 and RAR1 were reduced by about 60% to 80% in silenced plants as compared to plants infected with the TRV empty vector ( S4 Fig ) . At this time , Agrobacterium strains expressing His-XopAU were used to inoculate silenced and control plants , and cell death was monitored visually and quantified by measuring ion leakage . As a control , silenced leaves were also inoculated with Agrobacterium expressing the R gene/effector gene pair Cf4/avr4 , which elicits a hypersensitive response in N . benthamiana leaves that is dependent on expression of EDS1 and RAR1 , but not of NDR1 [42] . As expected , silencing of EDS1 and RAR1 , but not of NDR1 severely reduced Cf4/avr4-mediated cell death and ion leakage ( Fig 2F and 2G ) . Conversely , cell death mediated by His-XopAU was not affected by silencing of any of the tested ETI signaling components ( Fig 2F and 2G ) . These results suggest that it is unlikely that the cell death observed upon XopAU expression in leaf tissues is triggered by recognition of the effector by an R protein . To assess the contribution of XopAU to bacterial virulence , the corresponding gene was inactivated in Xe bacteria by insertion mutagenesis . The mutant strain Xe xopAU:GnR and wild-type Xe were used to infect ECW30R pepper plants that were then monitored for bacterial growth and development of disease symptoms . Disease symptoms were estimated visually and quantified by measuring chlorophyll content and ion leakage , as parameters of chlorosis and necrosis that are typically observed in pepper leaves infected by Xe bacteria . Leaves infected with the mutant strain Xe xopAU:GnR displayed a similar chlorophyll content , ion leakage and bacterial growth as leaves infected with wild-type Xe bacteria ( Fig 3 and S5A Fig ) . We hypothesized that a weak contribution of XopAU to Xe pathogenicity is more likely to be revealed in an attenuated bacterial strain . To test this possibility , we generated the double mutant strain Xe xopAU:GnR/avrBs2:KnR , mutated in both the xopAU and avrBs2 effector genes . The AvrBs2 effector was chosen for this analysis because it was previously shown to contribute to Xe virulence activity and its deletion allowed to reveal the virulence activity of other effectors [43 , 44] . The double mutant along with wild-type Xe and the single mutants Xe xopAU:GnR and Xe avrBs2:KnR were used to infect ECW30R pepper plants that were then monitored for bacterial growth , chlorophyll content and ion leakage . In infected leaves , the Xe avrBs2:KnR mutant displayed reduced bacterial growth and ion leakage , but similar chlorophyll content , as compared to Xe wild-type and Xe xopAU:GnR ( Fig 3 and S5A Fig ) . The Xe xopAU:GnR/avrBs2:KnR double mutant was similar to Xe avrBs2:GnR in bacterial growth and ion leakage , but caused less chlorosis as indicated by a higher chlorophyll content ( Fig 3 and S5 Fig ) . To confirm that the reduction in chlorotic symptoms was the result of a mutation in xopAU , the gene was re-introduced into the Xe xopAU:GnR/avrBs2:KnR strain driven by its native promoter . When inoculated into pepper leaves the complemented Xe xopAU:GnR/avrBs2:KnR ( xopAU ) strain caused similar disease symptoms as the Xe avrBs2:KnR strain ( Fig 3 and S5A Fig ) . Because expression of His-XopAU via Agrobacterium activated immune signaling ( see above ) , we tested whether deletion of the xopAU gene negatively affects the activation of defense responses . We monitored accumulation of the PR proteins PR-2 and PR-3 and mRNA levels of the PR-1 gene in infected pepper leaves at 3 dpi . Western blot analysis revealed that PR-2 and PR-3 accumulation was lower in leaves inoculated with the Xe xopAU:GnR/avrBs2:KnR double mutant strain than in leaves inoculated with the Xe avrBs2:KnR strain ( Fig 3C ) . Leaves inoculated with the Xe xopAU:GnR/avrBs2:KnR double mutant complemented with XopAU-HA driven by its native promoter , but not with the kinase deficient XopAU-HAK240A , accumulated similar PR-2 and PR-3 protein levels as leaves infected with the Xe avrBs2:KnR strain . Similarly , qRT-PCR analysis revealed that transcript levels of the PR-1 gene were about 8 . 6-fold lower in leaves inoculated with the Xe xopAU:GnR/avrBs2:KnR double mutant or with this strain complemented with the kinase deficient XopAU-HAK240A than in leaves infected with the Xe avrBs2:KnR strain or with the double mutant strain complemented with XopAU-HA ( Fig 3D ) . These results indicate that expression of a catalytically active XopAU in Xe strains promotes the activation of defense responses in infected pepper leaves . To further assess the contribution of XopAU to development of disease symptoms caused by Xe , we engineered a Xe strain carrying a HA-tagged XopAU variant ( XopAU-HA ) driven by a constitutive lac promoter in a broad-host plasmid . Overexpression of XopAU-HA in this strain was validated by Western blot analysis ( S6A Fig ) . A mock solution , Xe bacteria overexpressing XopAU-HA or carrying an empty vector were infiltrated into ECW30R pepper leaves that were then monitored for the development of chlorosis and necrosis ( i . e . chlorophyll content and ion leakage , respectively ) at 1 , 3 and 5 days post-inoculation ( dpi ) . Xe overexpressing XopAU-HA displayed lower chlorophyll content and reduced ion leakage , compared to Xe containing an empty vector ( Fig 4 ) . To confirm that the observed phenotype is due to the biochemical activity of XopAU , we generated an Xe strain overexpressing the catalytically inactive XopAUK240A-HA variant . No difference was observed in chlorosis and ion leakage between pepper plants inoculated with Xe overexpressing XopAUK240A-HA and Xe carrying an empty vector ( Fig 4 ) . Together , observations obtained by using Xe xopAU:GnR/avrBs2:KnR double mutant and Xe bacteria overexpressing XopAU-HA suggest that the XopAU effector participates in the development of disease symptoms . Next , we tested whether infection of ECW30R pepper leaves with Xe overexpressing XopAU-HA caused activation of defense responses as observed when the effector was expressed via Agrobacterium . First , we monitored accumulation of PR proteins in infected leaf tissues by Western blot analysis . Both PR-2 and PR-3 accumulated at higher levels in plants inoculated with Xe overexpressing XopAU-HA than in plants inoculated with Xe carrying an empty vector or overexpressing the kinase deficient variant XopAUK240A-HA at 3 and 5 dpi ( Fig 5A ) . We then assessed the mRNA levels of four genes ( PTI5 , ACO1 , OPR3 and PR-1 ) , whose expression reflects the activation of different defense and stress pathways , at the early stages of infection ( 16 hours after inoculation ) . qRT-PCR analysis revealed that transcript levels of the PR-1 gene , which is known to be induced by salicylic acid and pathogen attack [45] , were about 30 fold higher in pepper leaves inoculated with Xe overexpressing XopAU-HA than in plants inoculated with Xe carrying an empty vector or overexpressing the kinase deficient variant XopAUK240A-HA ( Fig 5B ) . The mRNA levels of the PTI5 and ACO1 genes , which are involved in ethylene signaling and biosynthesis , respectively [46] , were not significantly altered by overexpression of XopAU-HA ( Fig 5B ) . Finally , transcripts of the OPR3 gene , which encodes a component of the jasmonic acid biosynthesis pathway and is induced by wounding [47] , displayed only a slight induction ( about 2 fold ) when leaves were infected with Xe overexpressing XopAU-HA ( Fig 5B ) . This analysis suggests that XopAU overexpression activated plant defense signaling . Because accumulation of PR proteins might affect the ability of bacteria to colonize the plant , we examined whether overexpression of XopAU-HA affects bacterial growth in pepper leaves . As shown in Fig 5C , Xe bacteria overexpressing XopAU-HA displayed a similar growth as wild-type and bacteria overexpressing XopAUK240A-HA up to 6 dpi , and a reduced growth only at the late stages of infection ( 8 dpi ) , which may be ascribed to high accumulation of PR proteins . Activation of defense responses by XopAU is accompanied by cell death when XopAU is expressed through Agrobacterium but not when the effector is delivered in plant cells by Xe bacteria . This discrepancy may be related to the interplay between XopAU and other species-specific virulence determinants . To explore this possibility , a plasmid for overexpression of XopAU-HA or XopAUK240A-HA was mobilized into the crucifer pathogen Xanthomonas campestris pv . campestris strain 8004 ( Xcc ) , which does not encode a XopAU homolog . ECW30R pepper leaves were inoculated with Xe or Xcc strains overexpressing XopAU-HA , the kinase deficient variant XopAUK240A-HA or an empty vector , and cell death was monitored visually and quantified by ion leakage . Leaf areas inoculated with Xcc overexpressing XopAU-HA displayed cell death and a concomitant increase in ion leakage at 2 to 3 dpi , while Xcc containing an empty vector or overexpressing XopAUK240A-HA did not induce any visible phenotype ( Fig 6A and 6B ) . At the same time , overexpression of XopAU-HA through Xe induced strong chlorosis ( Fig 6A ) . To examine if the cell death induced by the expression of XopAU from Xcc is host specific , Xcc bacteria overexpressing XopAU-HA , XopAUK240A-HA or an empty vector were infiltrated into the leaves of N . benthamiana . Leaves inoculated with Xcc overexpressing XopAU-HA , but not XopAUK240A-HA or an empty vector , displayed cell death , which was confirmed by increased ion leakage at 24–48 h after inoculation ( S7A and S7B Fig ) , thus indicating that this phenotype was not host specific . Overexpression of XopAU in Xcc did not affect bacterial growth in infected pepper and N . benthamiana leaves ( Fig 6C and S7C Fig ) . It should be pointed out that expression of XopAU-HA and XopAUK240A-HA was higher in Xcc as compared to Xe bacteria as detected by Western blot analysis ( S6A Fig ) . Together , these observations suggest that the phenotype caused by XopAU when delivered by Xe may be tuned by bacterial determinants absent in Xcc strains . Alternatively , the different phenotypes may derive from the higher expression levels of XopAU-HA in Xcc compared to Xe . MAP kinase cascades were previously shown to be involved in cell death signaling associated with plant immunity [12] . We hypothesized that XopAU induces immune responses by manipulating and activating components of MAP kinase cascades . To test this hypothesis , we examined the ability of His-XopAU to elicit cell death in N . benthamiana plants that were silenced either for the MEK2 gene , which encodes a positive regulator of cell death , or for the MAP3Kα and MAP3Kε genes , which encode MAPKKKs acting upstream of MEK2 [9 , 10] . For gene silencing by VIGS , N . benthamiana plants were inoculated with Agrobacterium strains containing plasmids for the expression of TRV either empty or carrying a fragment of the gene to be silenced . Four weeks later , MEK2 , MAP3Kα and MAP3Kε transcript levels were reduced by at least 75% in silenced plants as compared to TRV-infected plants ( S4 Fig ) . At this time , Agrobacterium strains expressing His-XopAU were used to inoculate silenced and control plants , and cell death was monitored visually and quantified by measuring ion leakage . Cell death and ion leakage induced by His-XopAU were significantly reduced in MEK2-silenced plants compared to MAP3Kα- and MAP3Kε-silenced plants , and to TRV-infected plants ( Fig 7A and 7B ) . In addition , Western blot analysis revealed that phosphorylation of MAPKs induced by His-XopAU was reduced in the MEK2-silenced plants ( Fig 7C ) . Similarly , a reduction in cell death and ion leakage was also observed in MEK2-silenced plants challenged with Xcc overexpressing XopAU ( Fig 7D and 7E ) . To provide additional evidence that cell death induced by XopAU requires a functional MEK2 , we examined whether expression of the catalytically inactive variant of the tomato MEK2 homolog MKK2 ( MKK2K99R ) causes a dominant negative effect on XopAU-mediated cell death . His-XopAU was co-expressed via Agrobacterium in N . benthamiana leaves with MKK2-HA , MKK2K99R-HA , or an unrelated protein ( GFP ) driven by an estradiol inducible system . Cell death was visually monitored in the inoculated leaves and quantified by measuring ion leakage at 48 h after estradiol application . Expression of MKK2K99R-HA , but not that of MKK2-HA or GFP , significantly reduced the cell death and ion leakage induced by His-XopAU ( Fig 8 ) indicating that a catalytically active MEK2/MKK2 is required for XopAU-mediated cell death . Expression of certain type III effectors in yeast has been shown to cause phenotypes that can be exploited to elucidate effector function , biochemical activity and host targets [48] . To test whether XopAU causes a detectable phenotype in the yeast Saccharomyces cerevisiae , the effector was fused to a c-myc tag , expressed in the yeast strain W303 driven by the GAL1 promoter , and protein accumulation was confirmed by Western blot analysis ( S6B Fig ) . The effect of XopAU on yeast growth was examined by serially diluting yeast cultures that carry a vector either empty or for expression of the effector and plating them onto repressing ( glucose ) or inducing ( galactose ) media . The strain expressing XopAU exhibited similar growth as the control strain containing an empty vector both in repressing and inducing media ( Fig 9 ) . Because MEK2/MKK2 was required for XopAU-mediated phenotypes in planta , we hypothesized a similar requirement in yeast . To test this hypothesis , yeast were engineered to express under the control of the GAL1 promoter either XopAU , XopAU kinase deficient ( XopAUK240A ) , MKK2 and MKK2 kinase deficient ( MKK2K99R ) with an empty vector or the following protein combinations: XopAU with MKK2 , XopAU with MKK2K99R , and XopAUK240A with MKK2 . All the proteins were fused to a c-myc tag and their expression was validated by Western blot analysis ( S6B Fig ) . Yeast co-expressing XopAU with MKK2 displayed a significant reduced growth when plated on inducing medium , but not on repressing medium , as compared to yeast strains expressing each protein alone or protein combinations that included a kinase deficient variant of either XopAU or MKK2 ( Fig 9 ) . These results indicate that XopAU required MKK2 to cause growth inhibition in yeast and this phenotype was dependent on the kinase activity of both proteins . Because XopAU-mediated phenotypes were dependent on MEK2/MKK2 in planta and yeast , we hypothesized that MKK2 is a direct plant target manipulated by XopAU virulence activity . To explore this hypothesis , we tested whether XopAU physically interacts with MKK2 in a yeast two-hybrid system . To this aim , XopAU and its catalytically deficient form XopAUK240A were used as bait in yeast cells that expressed the tomato MAPKKs MKK1 , MKK2 , MKK3 or MKK4 as preys . All bait and prey proteins were expressed in the yeast cells as confirmed by Western blot analysis ( S6C and S6D Fig ) . While XopAU did not interact with any of the tomato MAPKKs , XopAUK240A specifically interacted with MKK2 , as evident by activation of the reporter genes LEU2 and lacZ ( Fig 10A ) . XopAUK240A also interacted with N . benthamiana MEK2 and pepper MKK2 ( S8 Fig ) . The lack of interaction between the catalytically active XopAU and MKK2 could be the consequence of the growth inhibition phenotype observed when both proteins were co-expressed in yeast ( Fig 9 ) . In parallel investigation aimed at the identification of additional candidate plant targets of the effector , XopAU was used as bait in a yeast two-hybrid screen of a tomato cDNA library [49] . This screen identified three MAPKs ( MPK1 , MPK3 , and MPK9 ) that consistently interacted with the kinase active and inactive forms of XopAU resulting in the activation of both reporter genes ( Fig 10A and S6D Fig ) . Next , we used split luciferase complementation assays to validate in planta protein-protein interactions that were observed in yeast . Wild-type XopAU could not be used in these experiments because it caused cell death when fused to the C-terminus of the firefly luciferase protein ( C-LUC ) and failed to accumulate in leaves . Instead , we used XopAUK240A that was fused to C-LUC and co-expressed in N . benthamiana leaves through Agrobacterium along with MPK1 ( representative of the XopAU-interacting MAP kinases ) or MKK2 fused to the N-terminus of the firefly luciferase ( N-LUC ) . As negative controls , C-LUC-XopAUK240A was co-expressed with N-LUC fused to the tomato receptor-like cytoplasmic kinase BSK830 , while N-LUC-MPK1 and N-LUC-MKK2 were co-expressed with C-LUC fused to the kinase domain of the tomato receptor-like kinase BTI9 . Expression of all the fusion proteins was validated by Western blot analysis ( S6E Fig ) . Protein-protein interactions in planta were quantified by measurements of luminescence at 48 h after agro-infiltration . Co-expression of C-LUC-XopAUK240A and N-LUC-MPK1 or N-LUC-MKK2 resulted in emission of significantly higher luminescence compared to the negative controls indicating a physical interaction in planta between these two pairs of fusion proteins ( Fig 10B ) . Together , these results indicate that XopAU physically interacts with multiple components of MAP kinase cascades at the MAPK and MAPKK levels . In vitro kinase assays were performed to test whether proteins that physically interacted with XopAU are substrates of XopAU phosphorylation . Kinase deficient variants of MPK1 ( MPK1K92R ) , MPK3 ( MPK3K70R ) , MKK2 ( MKK2K99R ) and MKK1 ( MKK1K99R ) , which did not interact in yeast with XopAU and thus served as a negative control , were expressed as GST fusions in E . coli , purified and incubated with GST-XopAU in the presence of [γ-32P]ATP . As shown in Fig 11A , GST-XopAU phosphorylated GST-MKK2K99R , but not GST-MPK1K92R , GST-MPK3K70R or GST-MKK1K99R . The kinase deficient GST-XopAUK240A was not able to phosphorylate GST-MKK2K99R confirming that labeling of GST-MKK2K99R was dependent on the GST-XopAU catalytic activity ( Fig 11B ) . The effect of XopAU on the phosphorylation state of MKK2 was then examined in planta . To this aim , MKK2 tagged with a HA epitope tag ( MKK2-HA ) was co-expressed via Agrobacterium in leaves of N . benthamiana plants along with His-XopAU in the wild-type or the kinase deficient form ( His-XopAUK240A ) . Expression of MKK2-HA , His-XopAU and His-XopAUK240A was driven by an estradiol inducible system . MKK2-HA was immunoprecipitated from leaf samples , which were collected at 12 h after estradiol application , digested with trypsin , and analyzed by quantitative mass spectrometry of phosphopeptides . This analysis identified six residues ( Thr33 , Ser73 , Tyr176 , Thr215 , Ser221 and Ser269 ) that were differentially phosphorylated in the presence of His-XopAU compared to His-XopAUK240A ( Table 1 and S4 Table ) . Remarkably , expression of His-XopAU resulted in an average increase of 200 fold in the phosphorylation of both Thr215 and Ser221 ( Table 1 and S4 Table ) , which are part of the S/TxxxS/T activation motif of MKK2 [12] . Phosphorylation of Thr33 , Ser73 , Tyr176 and Ser269 was also enhanced upon expression of XopAU by about 36 , 28 , 51 and 73 folds , respectively ( Table 1 ) . Together , these observations demonstrate that XopAU phosphorylated MKK2 in vitro and either directly or indirectly promoted phosphorylation of multiple MKK2 sites in planta , possibly resulting in its activation .
In this study , we uncovered biochemical properties of the Xanthomonas euvesicatoria type III effector XopAU that encodes a protein kinase and contributes to the development of disease symptoms in pepper plants . In addition , we show that XopAU manipulates MAP kinase signaling by activating the immunity-associated MAPKK MKK2 . This is the first report of a type III effector of phytopathogenic bacteria that encodes a catalytically active serine/threonine protein kinase representing a novel enzymatic activity for type III effectors acting within plant cells . Type III effectors with protein kinase activity were previously identified in bacterial pathogens that infect mammalian cells and they include the YpkA/YopO effector from Yersinia and OspG from Shigella [50] . YpkA/YopO phosphorylates the heterotrimeric G-protein Gαq in the GTP binding loop inhibiting Gαq activation and signal transduction [51] . While OspG substrates are yet to be identified , its kinase activity is required to inhibit degradation of phosphorylated IkBα and NF-κB activation induced by TNF-α stimulation , resulting in the interference of host innate immune responses [52] . Interestingly , kinase activity of both YpkA/YopO and OspG requires binding of a host factor for activation ( i . e . actin and ubiquitin , respectively ) possibly to prevent undesired activity while in the bacterium [53–55] . It will be interesting to test whether XopAU is a constitutively active kinase or it is activated in the plant cell by binding of a host factor or by posttranslational modifications . Protein-protein interaction studies revealed that XopAU interacts in yeast with multiple tomato MAPKs and with the immunity-associated MAPKK MKK2 . The interactions of XopAU with the MPK1 MAP kinase and MKK2 were also confirmed in planta . Moreover , MKK2 , but not the MAPKs , was a substrate of XopAU phosphorylation in vitro . In several instances components of MAP kinase cascades were found to be targeted by type III effectors . With the exception of activation of the MAP kinase MPK4 by the P . syringae effector AvrB [23] , other effector-MAPK interactions results in inactivation of the host MAPKs and interference with MAPK signaling . For example , members of the HopAI1 effector family of phosphothreonine lyases from plant and animal bacterial pathogens interact with MAP kinases and suppress their activities by irreversibly removing a phosphate to inhibit host immune responses [56] . Similarly , Yersinia YopJ and P . syringae HopF2 inhibit the signaling ability of MAPKKs by acetylation and ADP-ribosylation , respectively [19 , 22] . Our results demonstrate that this is not the case for XopAU . In fact , expression of XopAU in planta promoted phosphorylation of MAP kinases and MKK2 in their activation domains , and induced plant defense responses that are typically observed upon MKK2 activation [12] . The possibility that XopAU-mediated activation of defense responses is the result of recognition of the effector by a plant R protein is unlikely because silencing of early components of ETI signaling in N . benthamiana plants did not affect XopAU-mediated cell death . Expression of XopAU along with MKK2 in planta enhanced phosphorylation of MKK2 at six residues , including Thr215 and Ser221 that are part of the S/TxxxS/T MKK2 activation motif . It remains to be established whether XopAU directly phosphorylates these residues in planta or activates a mechanism resulting in their phosphorylation by another kinase ( s ) . Additional host proteins , such as the XopAU-interacting protein MPK1 , MPK3 and MPK9 may be involved in the activation of MKK2 by XopAU . MPK1 might modulate XopAU activity and substrate affinity by phosphorylation , and promote phosphorylation and activation of MKK2 by XopAU . Alternatively , XopAU might enhance MPK1 phosphorylation by MKK2 by acting as a scaffold protein that interacts and bridges between MKK2 and MPK1 . Functional evidence provides further support to the biochemical data for a role of MKK2 as a target of XopAU . In planta , silencing of NbMEK2 , the N . benthamiana ortholog of MKK2 , or overexpression of a kinase dead variant of MKK2 suppressed XopAU-mediated cell death and MAPK phosphorylation indicating that MKK2 is required for XopAU molecular function . In yeast , expression of both XopAU and MKK2 in a catalytic active form , but not that of each protein alone , resulted in growth inhibition suggesting molecular cooperation between the two proteins . MKK2 does not have a closely related homolog in yeast , while its downstream MAPKs , MPK1 and MPK3 , share 53% and 48% identity to FUS3 ( NCBI acc . num . AAA34613 . 1 ) and HOG1 ( NCBI acc . num . AJV50684 . 1 ) , respectively . Interestingly , activation of either FUS3 or HOG1 pathways was reported to promote cell cycle arrest [57 , 58] . Based on these observations and on the finding that MKK2 is a substrate of XopAU phosphorylation in vitro , we hypothesize that the growth inhibition phenotype caused by co-expression of MKK2 and XopAU in yeast is a result of MKK2 activation by XopAU and subsequent MKK2 initiation of yeast MAPK cascades involved in cell cycle arrest . Gene inactivation analysis in an attenuated Xe strain revealed that XopAU contributes to the appearance of disease symptoms in susceptible pepper plants , but not to bacterial growth . In addition , molecular analysis demonstrated that XopAU overexpression activates plant defense responses in Xe host and non-host plants . Accumulation of defense proteins reaches high levels at late stages of infection in pepper leaves infected with Xe overexpressing XopAU and may be the source of the decreased bacterial growth observed for this strain at 8 dpi . Remarkably , defense responses induced by XopAU were accompanied by the appearance of chlorosis when the effector was expressed by Xe , or by cell death when XopAU was delivered/expressed by Agrobacterium and Xcc . These different phenotypes may be related to different XopAU expression levels in the various experimental systems , as observed in Xe and Xcc bacteria overexpressing XopAU . Alternatively , the differential response observed when XopAU is delivered/expressed by different bacteria may result from inhibition of XopAU-mediated cell death by other Xe determinants . Possible XopAU antagonists are Xe 85–10 type III effectors , such as XopE1 and XopM , that are absent in Xcc 8004 and were found to suppress cell death induced by a constitutively active form of the immunity-associated MAPKK MEK2 [59] , the tobacco ortholog of MKK2 , which we report here to be activated by XopAU . Additional candidates are effectors that were shown to suppress ETI-dependent or independent cell death . For example , XopB inhibits ETI-related cell death triggered by recognition of the AvrBsT effector in pepper plants , as well as cell death induced by several other effectors in tobacco [60] . AvrBsT suppresses the ETI-related cell death induced by AvrBs1 in pepper [61] , while XopJ delays the appearance of necrotic disease symptoms interfering with host salicylic acid signaling [32] . Despite the fact that MKK2 was identified as a target of XopAU , it is yet to be established how activation of this immunity-associated MKK by XopAU may contribute to Xe pathogenicity . MAP kinase cascades activated by tomato MKK2 and its orthologs in other plant species have been implicated as key signaling modules not only in plant immunity [8] , but also in other physiological processes , such as the response to abiotic stress ( e . g . wounding , osmotic and oxidative stress ) , stomata development and floral senescence [62–66] . It is possible that XopAU-mediated activation of MKK2 selectively induces a subset of cellular responses that are beneficial to the pathogen . Alternatively , activation of defense responses through MKK2 could be connected to the contribution of XopAU to the development of disease symptoms . In support of this hypothesis , a correlation was observed between the appearance of chlorotic symptoms and accumulation of PR proteins in pepper leaves infected with Xe strains expressing catalytically active and inactive variants of XopAU . In summary , we provide evidence that XopAU is a functional protein kinase that manipulates host MAPK signaling by activating the immunity-associated MAPKK MKK2 . In addition , based on the different phenotypes observed when the effector is expressed by different bacteria , we propose a functional interaction between XopAU and other bacterial determinant ( s ) . This study provides new insights about a possible role for activation of host immunity-associated MAPK cascades in disease development .
The genomic region of Xe 85–10 ( NZ_CP017190 . 1 ) from position 4 , 861 , 200 to 4 , 862 , 753 ( base pairs ) , which contains the ORF of the xopAU gene [67] , was used to search homologous sequences in bacterial genomes of the non-redundant NCBI database . The xopAU and gyrB genes from a representative strain for each Xanthomonas species were selected for the phylogenetic analysis ( S1 Table ) . Phylogenetic analysis was performed by using the neighbor joining method based on the xopAU and gyrB sequence alignments that were obtained by using Clustal X [68] . The bootstrap consensus tree inferred from 100 replicates is taken to represent the evolutionary history of the taxa analyzed . Bacterial and yeast strains used in this study are listed in S2 Table and were grown as follows: Escherichia coli in Lysogeny Broth ( LB ) medium at 37°C; Xanthomonas euvesicatoria ( Xe ) , Xanthomonas campestris pv . campestris ( Xcc ) , and Agrobacterium tumefaciens in LB medium at 28°C; yeast ( Saccharomyces cerevisiae ) at 30°C in selective synthetic complete medium supplemented with 2% glucose , or 2% galactose and 1% raffinose [69] . Plant cultivars used in this study are: pepper ( Capsicum annuum ) ECW20R [43] and ECW30R [70] , Nicotiana benthamiana [71] , and tomato ( Solanum lycopersicum ) Hawaii 7981 [72] . Plasmid constructs used in this study are described in S3 Table . For cloning , DNA fragments were amplified from the Xanthomonas genome or cDNA of pepper , tomato or N . benthamiana plants , using Phusion DNA Polymerases ( Thermo Fisher Scientific , Inc . Waltham MA , USA ) or PrimeSTAR HS DNA Polymerase ( Clontech Laboratories , Inc . Mountain View CA , USA ) . Site-directed mutagenesis was carried out using the QuikChange II kit ( Agilent technologies , Inc . Santa Clara CA , USA ) . Sequences of oligonucleotides used in this study are available upon request . To generate an Xe insertion mutant in the avrBs2 gene ( XCV0052 ) by single crossover , an avrBs2 DNA fragment ( 187–827 bp ) was cloned into the pVIK165 plasmid . The obtained plasmid was mobilized into the Xe or Xe xopAU:GnR [67] strains and bacteria were plated on LB media with kanamycin selection . Gene disruption was verified by PCR and loss of Xe avirulence in resistant ECW20R pepper plants . For overexpression of the xopAU gene , the xopAU coding region was fused to a HA epitope tag and cloned into the pBBR1MCS2 broad host vector driven by the lac promoter . For complementation of the Xe xopAU:GnR/avrBs2:KnR strain , the xopAU gene was cloned with its native promoter ( 646 bp upstream to the start codon ) into the pBBR1MCS-3 plasmid in reverse orientation to the lac promoter . Plasmids were mobilized into Xanthomonas strains by triparental mating [73] . Binary vectors were transformed into Agrobacterium GV2260 by electroporation . For transient expression , Agrobacterium overnight cultures were pelleted , resuspended in induction medium ( 10 mM MgCl2 , 10 mM MES pH 5 . 6 , 200 mM acetosyringone ) , and incubated at 25°C with shaking for 4 h . Bacterial cultures were diluted to OD600 = 0 . 1 and infiltrated into leaves of N . benthamiana , pepper or tomato plants using a needleless syringe . When using the XVE estradiol-inducible system [74] , plants were sprayed with an induction solution ( 5 μM 17β-estradiol , 1% Tween-20 ) at 24 h after agro-infiltration . For inoculation , 7-week-old pepper or 4-weeks-old N . benthamiana and tomato plants were infiltrated with bacterial suspensions ( 105 CFU/mL when monitoring bacterial growth; 107 CFU/mL when measuring ion leakage and chlorophyll content ) in 10 mM MgCl2 by using a needleless syringe . For measurement of bacterial growth , three 1-cm-diameter leaf discs were sampled from at least three plants and ground in 1 mL of 10 mM MgCl2 . Bacterial numbers were determined by plating 10 μL from 10-fold serial dilutions and counting the resulting colonies . For measurements of chlorophyll content , 10–20 1-cm2 leaf disks were sampled for each treatment , placed in a tube containing 2 ml of acetone , and incubated overnight in the dark . Absorption was determined at OD660 and OD642 . Total chlorophyll content was quantified with the equation: 7 . 12 × OD660 + 16 . 8 × OD642 [75] . Chlorophyll content was calculated for each inoculated leaf area relative to a mock-infiltrated area of the same leaf . For the measurements of ion leakage , two 1 . 5-cm-diameter leaf disks were sampled from inoculated areas of at least five plants , and floated in 10-mL tubes containing 5 mL of double-distilled water for 4 h at 25 °C with shaking . Conductivity was measured using a DDS-12DW conductivity meter ( BANTE Instruments , Shanghai , China ) . Yeast two-hybrid ( Y2H ) interactions and library screening were conducted as described [76] . To enable the use of 17β-estradiol to activate the GAL1 promoter , the yeast EGY48 strain was integrated with the Gal4-ER-VP16 transactivator [77] and renamed EGY48ES . The xopAU and xopAUK240A genes were cloned into the bait plasmid pEG202 and plasmids were transformed into EGY48ES by lithium acetate transformation . Baits were tested for interactions with either a tomato cDNA library [49] or tomato proteins ( S3 Table ) that were fused to the pB42 transcriptional activation domain in the prey plasmid pJG4-5 . Expression of prey constructs was induced by growing yeast on media supplemented with 2% glucose and 0 . 5 μM of 17β-estradiol . Y2H interactions were tested using the LEU2 and lacZ reporter genes by plating yeast on selective media plates lacking leucine or containing x-gal , respectively . The xopAU and MKK2 genes were cloned into the yeast galactose inducible expression vectors pGML10 and pGMU10 , respectively . Plasmids were co-transformed into the yeast strain W303 by lithium acetate transformation . For monitoring growth , yeast cultures were grown overnight at 30°C in liquid selective media containing 2% glucose , washed twice in 10 mM MgCl2 , and normalized to OD600 = 0 . 1 . Ten-fold serial dilutions were spotted ( 10 μl ) onto repressing ( 2% glucose ) or inducing ( 2% galactose and 1% raffinose ) solid selective media , plates were incubated in 30°C for 72–96 h , and monitored visually for yeast growth inhibition . The xopAU , xopAUK240A , MKK2 , MPK1 , BSK830 ( GenBank acc . num . XP_004252882 . 1 ) and BTI9 [78] genes were cloned in frame to firefly luciferase fragments in the binary vectors pCAMBIA:N-LUC and pCAMBIA:C-LUC [79] . The obtained vectors were transformed into Agrobacterium and co-expressed in N . benthamiana leaves . Luciferase activity was measured at 48 h after infiltration: 3 mm diameter leaf disks were harvested and floated in 100 μL water in a white 96-well plate . Samples were supplemented with 0 . 5 mM D-luciferin ( Sigma-Aldrich , St . Louis MO , USA ) and incubated in the dark for 10 min . Luminescence was measured using a Veritas Microplate Luminometer ( Promega Corporation , Madison WI , USA ) . MPK1K92R , MPK3K70R , MKK2 , MKK2K99R , MKK1K99R , xopAU and xopAUK240A were cloned into the pGEX-4T-1 GST fusion expression vector ( GE Healthcare , Little Chalfont , UK ) . Plasmids were transformed into E . coli Rosetta strain ( MERCK , Kenilworth NJ , USA ) . Bacterial cultures were grown at 37°C while shaking to OD600 = 0 . 4–0 . 6 , supplemented with 0 . 1 mM Isopropyl β-D-1-thiogalactopyranoside ( IPTG ) , and incubated overnight at 16°C with shaking . Bacteria were pelleted , resuspended in binding buffer ( Tris pH 7 . 4 containing 1 mM PMSF , 5 μg/mL leupeptin and 5 μg/mL aprotinin ) , lysed using a French press and centrifuged . Supernatants were incubated with glutathione agarose ( Sigma-Aldrich ) and proteins were purified according to manufacturer’s instructions . GST fusion proteins ( 0 . 1–0 . 5 μg ) were incubated in a kinase assay solution [50 mM Tris-HCl , pH 7 . 0 , 1 mM dithiothreitol , 10 mM MgCl2 , 20 μM ATP , 10 μCi [γ-32P]ATP ( 3 , 000 Ci/mmol; PerkinElmer , Inc . Waltham MA , USA ) ] at 25°C for 30 min . Reactions were stopped by the addition of SDS-sample buffer . Half of the reaction volume was fractionated on SDS-PAGE and stained with Coomassie blue . The second half was fractionated on SDS-PAGE , transferred onto a PVDF membrane , and the membrane was exposed to autoradiography . For TRV infection , cotyledons of one-week-old N . benthamiana plants were co-infiltrated with Agrobacterium containing pTRV1 and pTRV2 in 1:1 ratio as described [10] . TRV2 plasmids used for silencing are described in S3 Table . Plants were grown in a growth chamber at 20°C in long day conditions ( 16 h light , 8 h dark ) . Total RNA was isolated from leaves ( 50 mg ) using the SV total RNA isolation system ( Promega Corporation ) . RNA samples ( 2 μg ) were reverse-transcribed using qScript cDNA Synthesis Kit ( Quanta BioSciences , Inc . Gaithersburg MD , USA ) and subjected to quantitative RT-PCR using gene specific primers ( available upon request ) . cDNAs were amplified using the SYBR Premix Ex Taq II ( Clontech Laboratories ) and the Mx3000P qPCR System ( Agilent technologies , Inc . Santa Clara CA , USA ) . The GAPDH gene was used for normalization , and gene expression was calculated by the comparative Ct method [80] . For protein extraction from yeast and bacteria , overnight cultures were pelleted , resuspended in lysis buffer ( 4% SDS , 100 mM Tris/HCl pH 7 . 6 , 0 . 1 M dithiothreitol ) and incubated in 95°C with SDS sample buffer for 10 min . For protein extraction from plant tissues , 3–6 leaf disks of 1 cm diameter were frozen in liquid nitrogen , homogenized in extraction buffer ( 100 mM Tris pH 7 . 4 , 1% Triton X-100 , 1 mM PMSF , 5 μg/mL leupeptin , 5 μg/mL aprotinin , 50 mM NaF and 1 mM Na3VO4 ) , and centrifuged . MKK2-HA was transiently co-expressed in N . benthamiana leaves with either His-XopAU or His-XopAUK240A driven by the XVE estradiol inducible system [74] . Ten gram of leaf tissues was harvested at12 h after estradiol application and ground in liquid nitrogen . Protein extraction buffer was added to the powder and samples were centrifuged . The supernatant was collected , centrifuged again , filtered through Miracloth and incubated overnight at 4°C with Monoclonal α:HA-agarose ( Sigma-Aldrich ) on a tube roller . HA-agarose beads were washed twice in Tris pH 7 . 4 and submitted to phosphopeptide mass-spectrometry analysis . Mass spectrometry analysis was performed at The Nancy & Stephen Grand Israel National Center for Personalized Medicine , Weizmann Institute of Science . For the identification of MKK2 sites phosphorylated in planta in the presence XopAU or XopAUK240A , immunoprecipitated samples were analyzed by LC-MS/MS as described [81] . Briefly , samples were subjected to in-solution , on-bead , trypsin digestion , and separated by using Split-less Nano Ultra Performance Liquid Chromatography ( nanoUPLC; 10 kpsi NanoAcquity , Waters ) . The nanoUPLC was coupled online through a nanoESI emitter ( 10 μm tip; New Objective; Woburn , MA , USA ) to a quadrupole orbitrap mass spectrometer ( Q Exactive Plus , Thermo Scientific ) using a FlexIon nanospray apparatus ( Proxeon ) . For “Discovery” analysis and calculation of total ion current data was acquired in Data Dependent Acquisition ( DDA ) mode , using a Top 12 method [82] . For “Targeted” analysis , data was acquired in parallel reaction monitoring ( PRM ) mode [83] , using an inclusion list containing all relevant peptides in the phosphorylated or un-modified form , as well as MKK2 peptide DVDNPNVVR . For DDA data analysis , raw data was imported into the Expressionist software ( GeneData ) . The software was used for retention time alignment and peak detection of precursor peptides . A master peak list was generated from all MS/MS events and sent for database searching using Mascot v2 . 5 ( Matrix Sciences ) . Data was searched against a protein database containing all available Nicotianoideae protein sequences from UniprotKB [84] , Agrobacterium tumefaciens protein sequences , tagged MKK2 and XopAU , and 125 common laboratory contaminant proteins . Peptide identifications were filtered such that the global false discovery rate was maximum of 1% , and were then imported back to Expressionist to annotate identified peaks and calculate peptide intensities . For “Targeted” analysis [85] of MKK2 peptides , raw data were imported into the Skyline software [86] . Each peptide was manually curated to select the three most intense and reliable transitions , as well as to determine exact peak boundaries . Spectral libraries from the DDA experiments were also constructed and used in Skyline to evaluate the confidence in peak peptide assignment . Peak areas were then exported to a Microsoft Excel file , where the ratio of phosphorylated versus unmodified species of the peptide was calculated for each experiment . This ratio was then used to determine differential phosphorylation of MKK2 between samples that expressed XopAU or XopAUK240A ( S4 Table ) . The intensity of the DVDNPNVVR peptide of MKK2 , which does not contain any phosphorylation site , was determined in each sample and along with total ion current of the analyzed samples was used to compare MKK2 and total protein levels .
|
Many bacterial pathogens inject effector proteins into their eukaryotic host cells through the type III secretion system to modulate host cellular processes . Elucidating the function of bacterial effectors and identification of their host targets is important for understanding the molecular mechanisms of pathogenicity and host immunity . In this study , we analyzed the mode of action of XopAU , a type III effector from the pepper and tomato pathogen Xanthomonas euvesicatoria . We found that XopAU is a catalytically active protein kinase providing the first report of an effector from a plant bacterial pathogen that displays such an enzymatic activity . We show that expression of XopAU activates immune responses and contributes to the development of disease symptoms . Interestingly , XopAU-mediated phenotypes are altered when the effector is expressed by different species of Xanthomonas , suggesting an interplay between this effector and other species-specific virulence determinants . Furthermore , we provide biochemical and genetic evidence that XopAU interferes with host immune signaling by activating the MAPKK MKK2 . Together , our results provide new insights into the interaction between the plant immune system and bacterial type III effector proteins .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"cell",
"death",
"plant",
"anatomy",
"cell",
"processes",
"microbiology",
"fungi",
"plant",
"science",
"plants",
"bacteria",
"mapk",
"signaling",
"cascades",
"agrobacteria",
"plant",
"microbiology",
"proteins",
"tomatoes",
"leaves",
"fruits",
"yeast",
"biochemistry",
"signal",
"transduction",
"eukaryota",
"cell",
"biology",
"post-translational",
"modification",
"biology",
"and",
"life",
"sciences",
"xanthomonas",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
] |
2018
|
The Xanthomonas euvesicatoria type III effector XopAU is an active protein kinase that manipulates plant MAP kinase signaling
|
Gcn4 is a master transcriptional regulator of amino acid and vitamin biosynthetic enzymes subject to the general amino acid control ( GAAC ) , whose expression is upregulated in response to amino acid starvation in Saccharomyces cerevisiae . We found that accumulation of the threonine pathway intermediate β-aspartate semialdehyde ( ASA ) , substrate of homoserine dehydrogenase ( Hom6 ) , attenuates the GAAC transcriptional response by accelerating degradation of Gcn4 , already an exceedingly unstable protein , in cells starved for isoleucine and valine . The reduction in Gcn4 abundance on ASA accumulation requires Cdk8/Srb10 and Pho85 , cyclin-dependent kinases ( CDKs ) known to mediate rapid turnover of Gcn4 by the proteasome via phosphorylation of the Gcn4 activation domain under nonstarvation conditions . Interestingly , rescue of Gcn4 abundance in hom6 cells by elimination of SRB10 is not accompanied by recovery of transcriptional activation , while equivalent rescue of UAS-bound Gcn4 in hom6 pho85 cells restores greater than wild-type activation of Gcn4 target genes . These and other findings suggest that the two CDKs target different populations of Gcn4 on ASA accumulation , with Srb10 clearing mostly inactive Gcn4 molecules at the promoter that are enriched for sumoylation of the activation domain , and Pho85 clearing molecules unbound to the UAS that include both fully functional and inactive Gcn4 species .
Cells undergo rapid transcriptional reprogramming in response to environmental changes by mobilizing transcriptional activators and repressors . Transcriptional activators function by binding to specific DNA sequences ( UAS elements in yeast ) and recruiting transcriptional cofactor proteins/complexes that remove repressive chromatin structure and directly recruit the transcriptional machinery to the promoters of genes under their control . Various mechanisms have been elucidated for stimulating activator function in response to environmental signals , including dissociation from a repressor , as in the case of yeast Gal4 [1] , or increased entry into the nucleus as described for Pho4 and Gln3 [2] . The yeast activator Gcn4 is regulated by a unique translational control mechanism that rapidly increases the rate of Gcn4 synthesis in response to limitation for any amino acid—the conditions where increased transcription of amino acid biosynthetic genes under Gcn4 control is essential to maintaining cell growth . Gcn4 is also negatively regulated by a pathway that evokes its phosphorylation , ubiquitylation , and degradation by the proteasome , such that continued high-level translation of GCN4 mRNA is required to sustain induction of Gcn4 protein and its target genes . Together , these systems provide for reversible , short-lived induction of Gcn4 , except under conditions of extreme starvation—in which protein synthesis is strongly impaired—where Gcn4 turnover is attenuated ( reviewed in [3] ) . In addition to stimulating the transcription of genes encoding enzymes representing all of the amino acid biosynthetic pathways—the regulatory response dubbed general amino acid control ( GAAC ) — one-tenth or more of the yeast genome is induced by Gcn4 , including genes involved in producing amino acid precursors , mitochondrial carrier proteins , vitamins and cofactors , amino acid transporters , autophagy , or the metabolism of purine , glycogen , and trehalose [4] , [5] . The induction of Gcn4 expression at the translational level in amino acid-starved cells requires the protein kinase Gcn2 , which is activated by uncharged tRNAs cognate to the limiting amino acid . Gcn2's sole substrate in yeast is the α subunit of general translation initiation factor 2 ( eIF2 ) . In its GTP-bound form , eIF2 delivers charged methionyl initiator tRNA ( Met-tRNAiMet ) to the small ( 40S ) ribosomal subunit in the first step of translation initiation . The inactive eIF2-GDP complex is released at the end of the process and must be recycled to eIF2-GTP by the guanine nucleotide exchange factor eIF2B . Phosphorylation of eIF2α on serine-51 by Gcn2 converts eIF2-GDP from substrate to inhibitor of eIF2B , impeding the formation of the eIF2-GTP-Met-tRNAiMet ternary complex ( TC ) . While this reduces the rate of bulk protein synthesis and limits amino acid consumption , it specifically induces translation of GCN4 mRNA owing to specialized regulatory sequences ( upstream ORFs ) present in the mRNA leader that couple reduced TC concentration to increased initiation at the GCN4 AUG start codon [3] . The newly synthesized Gcn4 enters the nucleus—a constitutive process for this activator [6] , binds to the UAS elements of its target genes and recruits multiple cofactors to the promoter . The recruited cofactors include the nucleosome remodeling complexes SWI/SNF and RSC; and the SAGA and Mediator complexes , which carry out histone acetylation and/or function as adaptors to recruit general transcription factors and RNA polymerase II ( PolII ) , culminating in increased assembly of transcription initiation complexes and elevated transcription of the coding sequences ( CDS ) [7]–[10] . In nutrient-replete yeast cells , and under conditions of moderate amino acid limitation , where Gcn2 is activated and translation of GCN4 mRNA induced , Gcn4 is a highly unstable protein owing to its ubiquitylation by ubiquitin ligase SCFCDC4 and attendant degradation by the proteasome [11]–[13] . This process helps to maintain Gcn4 at a low , basal level in nonstarved cells , and allows rapid restoration of the basal level when the translation rate of GCN4 mRNA is repressed by replenishing amino acids in starved cells . Rapid degradation of Gcn4 in sated or moderately starved cells requires its phosphorylation by the CDKs Cdk8/Srb10 and Pho85 , with Pho85 making the greater contribution [12] , [14] . In severely starved cells , Pho85's contribution to Gcn4 turnover is essentially eliminated , owing to the destabilization and consequent disappearance of its cyclin Pcl5 [15] , [16] , which accounts in large part for the stabilization of Gcn4 under these conditions . By contrast , Srb10 contributes to Gcn4 turnover under all conditions examined , making the minor contribution in sated or moderately starved cells but the major contribution in severely starved cells ( where Pho85 is inactive ) [14] . Despite its lesser importance in Gcn4 turnover , Srb10 appears to be responsible for clearing the fraction of Gcn4 that is sumoylated on Lys residues 50 and 58 . It appears that sumoylation of Gcn4 on K50/K58 reduces its occupancy at the UAS elements of target genes in the early stages of GAAC induction during moderate starvation for Ile/Val imposed with the inhibitor sulfometuron ( SM ) . However , the higher levels of UAS-bound unsumoylated Gcn4 that result from Arg substitutions of K50/K58 do not evoke increased PolII occupancy or higher transcription rates under these conditions in otherwise WT cells [17] . As sumoylation of transcription factors can inhibit transcriptional activation by impairing their ability to recruit RNA polymerase [18] , sumoylation of Gcn4 might impair its activator function to dampen the general control response . There is evidence that phosphorylation of Gcn4 by Srb10 or Pho85 reduces its activation function and that the phosphorylated species must be ubiquitylated and degraded by the proteasome to maintain WT basal expression of Gcn4 target genes under nonstarvation conditions . Thus , blocking proteasomal degradation of Gcn4 reduces target gene transcription in non-starved cells in a manner suppressed by eliminating CDK phosphorylation sites in Gcn4 or deleting both Srb10 and Pho85 [19] . It is unclear whether the accumulation of phosphorylated Gcn4 also impairs transcriptional activation under inducing conditions of amino acid starvation . The phosphorylation of Gcn4 by both kinases appears to be nucleus-localized , as Gcn4 mutants lacking nuclear localization signals are stabilized [6] . Srb10 is associated with the Mediator coactivator complex [20] , which phosphorylates the CTD of the largest subunit of RNA polymerase II , Rpb1 [21] . The fact that Gcn4 recruits Mediator to the promoter [7] , [8] is consistent with the possibility that Gcn4 participates in down-regulating its own function and stability by recruiting at least one of its inactivating CDKs [14] . Biosynthesis of threonine in yeast , as in other microorganisms and several plants , is a five-step pathway initiated with L-aspartic acid as the primary substrate [22] , [23] ( Fig . 1A ) . The absence of the threonine biosynthetic pathway in humans makes it a valuable target for drug development against fungal pathogens [24] . Transcription of at least 4 genes of the threonine pathway , HOM3 , HOM2 , THR1 and THR4 is under Gcn4 control ( Fig . 1A ) [4] , [25] . Threonine biosynthesis is also subject to feedback inhibition by threonine , which inhibits the activity of the first enzyme in the pathway , aspartate kinase ( Hom3 ) , and also partially inhibits homoserine kinase ( Thr1 ) ( Fig . 1A ) [26] . Peptide prolyl isomerase FKBP12 participates in the feedback inhibition of Hom3 by physical interaction between these two proteins [27] , [28] . Besides threonine auxotrophy , thr1Δ and thr4Δ mutants exhibit myriad phenotypes that result from accumulation of the pathway intermediate homoserine ( HS ) , as they are mitigated in thr1Δ hom3Δ double mutants that cannot produce HS ( Fig . 1A ) [29] . Accumulation of the substrate of Hom6 , β-aspartate semialdehyde ( ASA ) , also is toxic , as releasing feedback inhibition of Hom3 is lethal in hom6Δ cells ( that accumulate ASA ) in a manner rescued by simultaneously blocking ASA synthesis by eliminating Hom2 or Hom3 [27] , [29] . However , hom6Δ mutants do not share all phenotypes of thr1Δ and thr4Δ mutants , and hom6Δ suppresses those unique to the latter mutants , implicating HS and ruling out a role for ASA in conferring many defects displayed by thr1Δ and thr4Δ cells . There is circumstantial evidence that HS toxicity results from its incorporation into proteins in place of threonine , which might evoke increased degradation of the HS-substituted proteins by the proteasome [29] . Previously , we screened the entire library of viable haploid deletion mutants of Saccharomyces cerevisiae for sensitivity to SM ( SMS phenotype ) to identify genes required for a robust GAAC , which allowed us to implicate various cofactors in the mechanism of transcriptional activation by Gcn4 [7] , [30] , and certain vacuolar sorting proteins ( Vps ) in maintaining high-level Gcn4 activation function in cells starved for Ile/Val [13] . In the course of that work , we also discovered that hom6Δ , thr1Δ , and thr4Δ mutants are also SMS , and undertook here to elucidate the mechanisms underlying this phenotype . In fact , it had been shown previously that thr1Δ mutants are SMS , and that this phenotype is suppressed by deleting HOM3 . As SM evokes derepression of threonine pathway enzymes by Gcn4 [4] ( Fig . 1A ) , the SMS phenotype of thr1Δ mutants was attributed to Hom3-dependent accumulation of HS , and its attendant toxicity to cellular processes , when HOM3 and HOM2 transcription is induced by Gcn4 [29] . This explanation would not apply to hom6Δ cells , however , which cannot produce HS , leading us to examine whether the SMS phenotype in this instance results from ASA accumulation and impairment of the GAAC response . The results of our analysis indicate that ASA accumulation indeed attenuates GAAC , by accelerating further the already rapid degradation of Gcn4 triggered by the CDKs Pho85 and Srb10 . They further suggest that Srb10 functions primarily in efficient clearance of inactive Gcn4 molecules , enriched for sumoylated species , whereas Pho85 clears unsumoylated , highly functional Gcn4 in addition to defective species .
As noted above and displayed in Fig . 1B , yeast deletion mutants lacking HOM6 , THR1 , or THR4 are sensitive to sulfometuron methyl ( SM ) , which evokes starvation for isoleucine and valine ( Ile/Val ) by inhibition of the ILV2-encoded biosynthetic enzyme [4] . At the SM concentration employed , growth of the hom6Δ , thr1Δ and thr4Δ strains is impaired to an extent similar to that of the gcn4Δ strain , lacking the activator of GAAC . Unlike these mutants , the hom3Δ and hom2Δ mutants grow like the wild-type ( WT ) strain on SM-containing medium ( Fig . 1B , SC+SM ) . These findings indicate that thr1Δ , thr4Δ , and hom6Δ strains , but not hom3Δ and hom2Δ mutants , are sensitive to Ile/Val starvation imposed by SM . Moreover the hom6Δ mutant grows more slowly than WT ( Slg- phenotype ) even on medium lacking SM ( Fig . 1B , SC ) . To determine whether the SMS phenotypes of the thr1Δ , thr4Δ , and hom6Δ mutants reflect defective transcriptional activation by Gcn4 , we measured induction of a UASGCRE-CYC1-lacZ reporter , driven by the CYC1 promoter and tandem Gcn4 binding sites from HIS4 ( the UASGCRE ) replacing the endogenous CYC1 UAS; and of a HIS4-lacZ reporter containing the native HIS4 5′-noncoding region . ( HIS4 is a known Gcn4 target gene [5] , [31] . ) As expected , treatment with SM for 6 h evokes a strong increase in UASGCRE-CYC1-lacZ reporter expression in WT , but not in gcn4Δ cells ( Fig . 1C ) . Smaller induction ratios were observed for all five mutants of the threonine pathway , with the largest defect seen for the hom6Δ strain ( ∼75% reduction of induced UASGCRE-CYC1-lacZ expression ) and the smallest defects observed for the hom3Δ and hom2Δ mutants ( ∼25% reductions ) ( Fig . 1C , left ) . In the case of the HIS4-lacZ reporter , the hom6Δ mutant , but not the hom3Δ or thr1Δ strains , displayed a marked ( ∼75% ) reduction in induction by SM ( Fig . 1C , right ) . To confirm these findings , we measured induction of native mRNAs for HIS4 and ARG1 ( another known Gcn4 target gene ) . Consistent with the HIS4-lacZ data , we observed induction defects for the hom6Δ mutant , but not the hom3Δ or thr1Δ strains , for both mRNAs ( Fig . 1D ) . The magnitude of the induction defect in the hom6Δ mutant was considerably greater after 120 min versus 30 min of SM treatment , displaying ∼60% and ∼67% reductions for HIS4 and ARG1 mRNAs , respectively , at the longer incubation time , even though full induction of both mRNAs was achieved by 30 min of SM treatment in WT cells ( Fig . 1D ) . The foregoing results indicate that the SM-sensitivity of the hom6Δ mutant reflects a substantial defect in GAAC resulting from reduced transcriptional activation by Gcn4 , which becomes more severe as starvation proceeds . By contrast , the other four threonine pathway mutants exhibit smaller defects in transcriptional activation , and the hom3Δ and thr1Δ strains actually display no detectable impairment of HIS4 and ARG1 induction by SM . The strong SMS phenotypes of the thr1Δ and thr4Δ mutants ( Fig . 1B ) can be reconciled with their moderate GAAC defects ( Figs . 1C–D ) by recalling that they accumulate the toxic intermediate HS , and that Gcn4-mediated induction of HOM2 and HOM3 under SM-induced starvation conditions is expected to elevate HS production in these strains ( Fig . 1A ) , in the manner proposed previously for thr1Δ cells [29] . By contrast , induction of the threonine pathway during SM treatment in hom3Δ or hom2Δ mutants should have no effect on cell growth ( as observed in Fig . 1B ) because they cannot produce HS . The HOM6 product , homoserine dehydrogenase ( HSD ) , converts β-aspartate semialdehyde ( ASA ) into HS . If the GAAC defect in hom6Δ cells results from the absence of this reaction , then hom6 mutants that produce catalytically defective HSD should display a strong GAAC defect . Based on a crystal structure of yeast HSD , 4 active site substitutions were generated that were previously characterized for their effects on HSD catalytic activity in vitro [32] . We introduced the corresponding mutations into plasmid-borne HOM6 and examined the ability of the mutant alleles to complement the transcriptional activation defects of hom6Δ cells . As expected , introduction of WT HOM6 complemented the Slg- and SMS phenotypes on media containing threonine , and the failure to grow on medium lacking threonine , of the hom6Δ strain ( Fig . 2A , SC , SC+SM and SC-Thr , respectively ) . Except for the E208D allele , the plasmid-borne hom6 alleles encoding HSD active site substitutions abolished complementation of the threonine auxotrophy and SM-sensitivity of the hom6Δ strain ( Fig . 2A ) . Consistent with this , the three defective alleles failed to restore SM-induction of the UASGCRE-CYC1-lacZ reporter , whereas E208D restored a WT level of induction ( Fig . 2B ) . Interestingly , the previously determined kinetic parameters of the hom6-E208D product indicated a reduced substrate affinity , but high-level catalytic activity , in comparison to WT HSD [32] . Accordingly , our results demonstrate that HSD catalytic activity is required for a robust GAAC response . We presume that the diminished substrate affinity of the hom6-E208D mutant does not significantly reduce the rate of converting ASA to HS in living cells . We asked next whether the requirement for HSD activity for the GAAC response reflects a requirement for HS synthesis or , rather , the need to prevent accumulation of ASA . If the inability to produce HS is the salient defect , then supplementing hom6Δ cells with HS should restore their GAAC response . We found that a supplement of 1 mM HS restores growth on SC-Thr medium for the hom3Δ , hom2Δ and hom6Δ strains , but not for the thr1Δ or thr4Δ strains ( Fig . 3A ) , consistent with the position of Thr1 and Thr4 downstream of HS production in the Thr pathway ( Fig . 1A ) . A supplement of 5 mM HS was required to confer growth of the hom3Δ , hom2Δ and hom6Δ strains indistinguishable from that of WT; although this elevated HS concentration retards the growth of WT cells ( Fig . 3A ) , presumably reflecting HS toxicity [29] . Importantly , HS supplementation did not rescue the defective SM-induction of the UASGCRE-CYC1-lacZ reporter in the hom6Δ mutant ( Fig . 3B ) , indicating that its GAAC defect does not result from the inability to produce HS . If accumulation of the Hom6/HSD substrate ( ASA ) in hom6Δ cells is responsible for the GAAC defect , then the GAAC response should be restored by preventing ASA production by eliminating Hom2 or Hom3; moreover , the GAAC defect should be exacerbated by eliminating feedback inhibition of the Hom3 product ( Fig . 1A ) . Indeed , deleting HOM2 or HOM3 in the hom6Δ mutant restored SM-induction of the UASGCRE-CYC1-lacZ reporter essentially to the same levels observed in the hom2Δ or hom3Δ single mutants ( Fig . 3C ) . Introducing WT HOM3 into the hom3Δ hom6Δ strain reinstated a defect in SM-induction of UASGCRE-CYC1-lacZ similar to that seen in the hom6Δ single mutant ( Fig . 3D ) . Importantly , a relatively greater induction defect was observed when the feedback-resistant allele hom3-E282D ( dubbed HOM3fbr ) was introduced instead into the hom3Δ hom6Δ strain ( Fig . 3D , cf . last two columns ) . As expected , introduction of HOM3fbr into the hom3Δ hom6Δ strain confers a strong Slg- phenotype ( Fig . S1 ) , owing to accumulation of ASA and its toxic effects on cell growth [27] . These findings demonstrate that the GAAC defect in hom6Δ cells results from ASA accumulation . We sought next to determine whether ASA accumulation impairs the GAAC by reducing Gcn4 abundance . Starvation for Ile/Val by SM rapidly increases Gcn4 synthesis by inducing the translation of GCN4 mRNA [3] . Western analysis of WT cells reveals the expected rapid induction of Gcn4 after only 30 min of SM treatment , with a gradual decline in abundance as starvation continues up to 120 min [9] ( Fig . 4A and B ) . Gcn4 abundance was decidedly reduced over much of the time course of SM treatment in vector transformants of the hom6Δ strain , again reaching its lowest level at 120 min of SM treatment . This reduction in Gcn4 abundance was mitigated by the absence of HOM3 in the hom6Δ hom3Δ double mutant , and exacerbated in transformants of the double mutant harboring feedback-resistant HOM3fbr , in which ASA accumulation is eliminated or exacerbated , respectively ( Fig . 4A and B ) . Even after only 30 min of SM treatment , the hom6Δ HOM3fbr strain displayed low-level Gcn4 similar to that observed in hom6Δ/vector transformants after prolonged SM treatment for 120 min . The gradual decrease in Gcn4 abundance in hom6Δ cells ( Fig . 4B ) is consistent with the greater reduction in Gcn4 target gene transcription seen at 120 min versus 30 min of SM treatment ( Fig . 1D ) . Moreover low-level HIS4 mRNA was observed in the hom6Δ HOM3fbr strain even after only 30 min of SM treatment ( Fig . S2 ) . To determine whether the reduced Gcn4 abundance on ASA accumulation reflects decreased translation of GCN4 mRNA , we assayed a GCN4-lacZ fusion shown to be a faithful reporter of GCN4 transcription and the translational efficiency of GCN4 mRNA [33] , [34] . Expression of this reporter shows the expected ∼10-fold induction in WT cells after 6 h of SM treatment , which is dampened somewhat both in the hom6Δ strain and in HOM3 transformants of the hom6Δ hom3Δ double mutant , but not in the vector transformants of the same strain ( Fig . 4C ) . However , the HOM3fbr and HOM3 transformants of the double mutant exhibit indistinguishable levels of GCN4-lacZ expression . Similar results were obtained after only 1 h or 2 h of SM treatment ( Fig . 4D ) , with the hom6Δ strain and HOM3fbr transformants of the hom6Δ hom3Δ double mutant both exhibiting similar reductions in GCN4-lacZ expression of ∼33% compared to the WT strain . As expected , a gcn2Δ mutant , lacking the key activator of GCN4 mRNA translation [3] , is completely defective for GCN4-lacZ expression ( Fig . 4D ) . While these findings suggest a reduction in Gcn4 synthesis on ASA accumulation in cells lacking HOM6 , the ∼33% reductions in GCN4-lacZ expression observed in the hom6Δ and hom6Δ HOM3fbr strains do not account for the 60–70% reductions in Gcn4 abundance observed after 2 h of SM treatment in the same strains . These findings suggest that Gcn4 is also degraded more rapidly than usual in response to ASA accumulation . To provide direct evidence supporting this last conclusion , we measured the turnover of newly synthesized Gcn4 by a pulse-chase experiment . Cells were cultured with SM for 30 min and pulse-labeled with [35S]-methionine/cysteine for the last 15 min of the starvation period , and then chased with excess nonradioactive methionine/cysteine . Consistent with previous reports , Gcn4 is normally a highly unstable protein and decays with a half-life of ∼10–12 min in SM-treated WT cells ( Fig . 4E–F ) [13] . Importantly , Gcn4 decay was markedly accelerated in the hom6Δ HOM3fbr strain , with the Gcn4 half-life dropping below 5 min , thus confirming that Gcn4 is degraded more rapidly in response to ASA accumulation ( Fig . 4E–F ) . As noted above , rapid degradation of Gcn4 is dependent on its phosphorylation by Pho85 and Srb10 in the nucleus , leading to its ubiquitylation and degradation by the proteasome [3] . It was also shown that the DNA-binding activity of Gcn4 is required for its sumoylation [17] . While it has been assumed that phosphorylation of Gcn4 by Srb10 likewise requires its binding to the UASGCRE [35] , this has not been directly demonstrated . We hypothesized that the increased rate of Gcn4 turnover on ASA accumulation results from its increased phosphorylation by Pho85 or Srb10 and attendant degradation by the proteasome; and wished to determine whether , like sumoylation , the increased phosphorylation occurs when Gcn4 is bound to the UASGCRE . To this end , we asked whether inactivating the DNA-binding ability of Gcn4 would suppress the effect of ASA accumulation on its abundance by conducting Western analysis of Gcn4 variants described previously [36] lacking the C-terminal basic region or leucine zipper , which are both required for DNA binding by Gcn4 [37] . The variant lacking the DNA binding domain , gcn4-Δ235-250 , also lacks one of two nuclear localization sequences ( NLS2 ) identified in Gcn4 , whereas the variant lacking the leucine zipper , gcn4-Δ251-281 , retains both NLSs , and it was shown that the leucine zipper is dispensable for nuclear localization of GFP-tagged Gcn4 [6] . We verified that both gcn4 alleles are indistinguishable from deletion of the entire GCN4 coding sequence in the inability to permit growth on SM-containing medium ( Fig . S3 ) . Western analysis of SM-treated cells revealed that both variants differ dramatically from WT Gcn4 and display no detectable reduction in abundance in hom6Δ cells treated for 2 h with SM ( Fig . 5A ) . ( Note that both truncated variants are well expressed and show the expected increased electrophoretic mobility compared to WT Gcn4 ( Fig . 5A ) ) . Furthermore , introducing HOM3fbr into hom6Δ cells , which severely diminishes WT Gcn4 after only 30 min of SM treatment ( Fig . 4A ) , has no effect on abundance of the gcn4-Δ235-250 and gcn4-Δ251-281 mutant proteins ( Fig . S4A ) . While these results suggested that ASA evokes accelerated degradation only when Gcn4 is capable of UASGCRE-binding , it was possible that the greater stability of the mutant variants results from a failure to accumulate ASA on SM treatment owing to the absence of Gcn4-mediated derepression of threonine biosynthetic enzymes required for ASA production . To address this last possibility , we repeated the experiment with the gcn4-Δ251-281 strains containing or lacking HOM6 after introducing WT GCN4 to reinstate the GAAC . Now we observed that the abundance of the truncated gcn4-Δ251-281 product was strongly reduced in hom6Δ cells , mirroring the behavior of full-length WT Gcn4 present in the same cells ( Fig . 5B ) . The same results were observed in the corresponding strains also containing HOM3fbr ( Fig . S4B ) . These results indicate that UAS-binding by Gcn4 is not required for its rapid degradation on ASA accumulation . Considering that Pho85 is responsible , whereas UAS-binding is dispensable , for the bulk of Gcn4 turnover under normal growth conditions [6] , [12] , [14] , these findings are consistent with the possibility that Pho85 plays a prominent role in the accelerated degradation of Gcn4 evoked by excess ASA . We sought next to determine the contributions of Srb10 and Pho85 to the enhanced degradation of Gcn4 in response to excess ASA . Consistent with previous findings [12] , [14] , deletion of either SRB10 or PHO85 increases the abundance of Gcn4 in otherwise WT cells treated with SM , with pho85Δ evoking a somewhat greater increase than srb10Δ ( Fig . 5C , lanes 1 , 5 , 9; Fig . 5D , vector transformants ) . As already shown above , Gcn4 abundance is severely diminished after 120 min of SM treatment in hom6Δ or hom6Δ HOM3fbr cells compared to the isogenic HOM6 cells ( Fig . 5C , lanes 3–4 vs . 1–2 ) . Importantly , eliminating SRB10 almost completely eliminates this reduction in Gcn4 abundance in both hom6Δ and hom6Δ HOM3fbr cells treated with SM ( Fig . 5C , lanes 7–8 vs . 3–4; & Fig . 5D ) . The slower migrating Gcn4 species evident in WT cells is considerably reduced in srb10Δ cells , suggesting that it represents a phosphorylated form of Gcn4 that depends on Srb10 , which is generally consistent with previous results [14] . Deletion of PHO85 also suppresses the reduction in Gcn4 abundance evoked by SM treatment of hom6Δ or hom6Δ HOM3fbr cells ( Fig . 5C , cf . lanes 3–4 vs . 11–12 ) . In fact in pho85Δ strains , Gcn4 abundance is higher in hom6Δ and hom6Δ HOM3fbr cells ( where ASA accumulates ) compared to the isogenic HOM6 cells ( Fig . 5C , cf . lanes 11–12 vs . 9–10; and Fig . 5D ) . Interestingly , deleting PHO85 seems to increase the relative abundance of the slower migrating Gcn4 species , which presumably represent products of Srb10 phosphorylation that are not efficiently cleared in pho85Δ cells ( Fig . 5C , cf . lanes 9–12 vs . 1–4 ) . The results in Figs . 5C–D suggest that both Srb10 and Pho85 are required for the strong depletion of Gcn4 that occurs on ASA accumulation . The stronger effect of pho85Δ versus srb10Δ on Gcn4 abundance observed on ASA accumulation in these experiments is consistent with previous results indicating a relatively greater contribution of Pho85 to Gcn4 degradation under normal growth conditions [12] , [14] . Gcn4 was found to be phosphorylated in vitro by Srb10 on multiple CDK consensus sites , including Ser17 , Ser210 , Thr61 , Thr105 and possibly Thr165 [14] , and by Pho85 both in vivo and in vitro on Thr165 [12] . Moreover , the T165A substitution alone was sufficient to confer marked stabilization of Gcn4 in vivo [12] . Importantly , we found that the Gcn4-T165A variant showed no reduction in abundance in SM-induced hom6Δ cells compared to HOM6 cells ( Fig . 5E–F ) . Moreover , replacing WT GCN4 with the GCN4-T165A allele in hom6Δ cells restored UASGCRE-CYC1-lacZ reporter ( Fig . 5G ) and ARG1 mRNA expression ( Fig . 5H ) to levels essentially equivalent to those seen in HOM6 GCN4 cells . Expression of HIS4 mRNA also was boosted by GCN4-T165A in hom6Δ cells , although expression remained below that seen in HOM6 GCN4 cells ( Fig . 5H ) , suggesting either that the Gcn4-T165A variant is not functionally equivalent to WT Gcn4 or that a fraction of Gcn4-T165A rescued in hom6Δ cells has a lower than WT specific activity . In any event , these findings provide strong evidence that phosphorylation of T165 by Pho85 and/or Srb10 is required for the pronounced depletion of Gcn4 evoked by ASA accumulation in hom6Δ cells . Having found that removing either Srb10 or Pho85 restores high-level Gcn4 abundance during ASA accumulation in hom6Δ cells , we expected to find that transcriptional activation by Gcn4 would likewise be restored in both hom6Δ srb10Δ and hom6Δ pho85Δ strains , particularly since these CDKs have been implicated in reducing Gcn4 activation function via phosphorylation of Gcn4 [19] . However , we observed distinct differences in the activation function of Gcn4 in cells lacking Srb10 versus Pho85 . First , we found that eliminating PHO85 restores the ability of hom6Δ cells to grow on SM containing plates ( Fig 6A ) . By contrast , hom6Δ srb10Δ cells cannot grow on SM medium , even though HOM6 srb10Δ cells grow at the WT rate on SM medium ( Fig . 6A ) . These findings suggest that deletion of SRB10 does not rescue the defective GAAC response to SM treatment in hom6Δ cells , whereas deletion of PHO85 does . Consistent with the growth assays , we found that eliminating PHO85 fully restores transcriptional activation of HIS4 and ARG1 in hom6Δ cells , conferring even higher than WT levels of both transcripts in the hom6Δ pho85Δ double mutants ( Fig . 6B ) . It is noteworthy that deleting HOM6 provokes no reduction in HIS4 or ARG1 mRNAs , and even seems to elevate HIS4 mRNA , in pho85Δ cells ( Fig . 6B ) . By contrast , deleting SRB10 evokes little or no increase in HIS4 or ARG1 mRNA levels in SM-treated hom6Δ cells ( Fig . 6B ) . The failure of srb10Δ to rescue activation of these genes in hom6Δ cells cannot be attributed simply to the loss of a coactivator function of Srb10 [10] , as srb10Δ had little or no effect on levels of HIS4 or ARG1 mRNAs in otherwise WT HOM6 cells ( Fig . 6B , srb10Δ vs . WT ) , nor on the ability to grow in SM medium ( Fig . 6A , srb10Δ vs . WT ) . These findings suggest that the Gcn4 molecules rescued from accelerated degradation on ASA accumulation by elimination of Srb10 are relatively nonfunctional in transcriptional activation . By contrast , the Gcn4 molecules rescued from degradation by elimination of Pho85 from hom6Δ cells appear to include highly functional species capable of evoking a greater than WT level of transcriptional activation . Having found above that deleting PHO85 restores a higher level of Gcn4 in hom6Δ cells than does deleting SRB10 ( Fig . 5D ) , it was important to determine whether the higher level of transcriptional activation seen in hom6Δ pho85Δ versus hom6Δ srb10Δ cells ( Figs . 6A–B ) arises simply from relatively greater UAS occupancy by Gcn4 in hom6Δ pho85Δ cells . To address this possibility , we conducted ChIP analysis to measure the occupancy of Gcn4 at the ARG1 UAS and the occupancies of Rpb3 ( a PolII subunit ) at the promoter ( TATA element ) and the 5′ or 3′ ends of the CDS at ARG1 after 2 h of SM treatment . It was shown previously that SM treatment of WT cells evokes large increases in occupancies of Gcn4 and Rpb3 at ARG1 that are completely absent in gcn4Δ cells [9] , [10] . Importantly , these increases in occupancy are strongly diminished in SM-treated hom6Δ cells ( Fig . 6C ) , providing direct evidence that the GAAC defect in the hom6Δ mutant results from low-level Gcn4 occupancy of the UAS with attendant reduced recruitment of PolII to the promoter . As expected from the ability of srb10Δ to restore cellular Gcn4 abundance in hom6Δ cells ( Fig . 5C–D ) , Gcn4 occupancy of the ARG1 UAS is substantially higher in hom6Δ srb10Δ versus hom6Δ SRB10 cells ( Fig . 6C , blue bars ) . However , this increase in Gcn4 occupancy is associated with much smaller increases in Rpb3 occupancies at all three locations at ARG1 ( Fig . 6C , orange , green , purple bars ) , consistent with the idea that the Gcn4 recovered in hom6Δ srb10Δ cells is relatively inactive . Deletion of PHO85 had strikingly different consequences on Gcn4 activity . In HOM6 cells , the pho85Δ mutation evokes a reduction in UAS occupancy of Gcn4 , but actually increases Rpb3 occupancies compared to the WT strain ( Fig . 6C , pho85Δ vs . WT ) , which is consistent with the higher than WT levels of ARG1 mRNA in pho85Δ cells shown above ( Fig . 6B ) . This effect of pho85Δ was noted previously [17] , and is not understood mechanistically; however , it might indicate that Pho85 clears fully functional Gcn4 molecules as a homoeostatic mechanism to prevent hyperinduction of the GAAC response , such that UAS-bound Gcn4 has a greater than WT specific activity in pho85Δ cells . Despite the much higher total cellular abundance of Gcn4 observed in hom6Δ pho85Δ versus hom6Δ srb10Δ cells ( Fig . 5D ) , Gcn4 occupancy of the ARG1 UAS is comparable in these two strains ( Fig . 6C , blue bars ) . In contrast , the Rpb3 occupancies at all three locations at ARG1 are substantially higher in the hom6Δ pho85Δ versus hom6Δ srb10Δ cells ( Fig . 6C , orange , green , purple bars ) . In fact , the Rpb3 occupancies observed in hom6Δ pho85Δ cells exceed those in WT cells despite a lower than WT level of Gcn4 UAS occupancy in the mutant cells ( Fig . 6C ) . These findings suggest that the Gcn4 molecules rescued in hom6Δ pho85Δ cells that are capable of UAS binding have a greater than WT specific activity . As elaborated in the Discussion , the reduced ability of UAS-bound Gcn4 to activate transcription in the hom6Δ srb10Δ double mutant could be explained by proposing that Gcn4 is rendered less functional in response to ASA accumulation and that Srb10 is required to clear the inactive Gcn4 molecules from the promoter by targeting them for degradation . The apparent hyperactivity of UAS-bound Gcn4 in hom6Δ pho85Δ cells could be explained by proposing that Pho85 targets both fully functional Gcn4 and defective species rendered incapable of UAS-binding on ASA accumulation in hom6Δ cells . It was shown recently that Gcn4 is sumoylated at target gene promoters , and Srb10 was implicated in clearing these sumoylated Gcn4 molecules [17] . We considered the possibility that sumoylation of Gcn4 bound to promoters increases on ASA accumulation and enhances the clearance of inactive Gcn4 by Srb10 . If so , we would expect to find elevated sumoylation of Gcn4 in srb10Δ hom6Δ strains , but not in pho85Δ hom6Δ strains . To examine this possibility , we immunoprecipitated Gcn4 from whole cell extracts ( WCEs ) and probed the immune complexes with antibodies against Smt3 ( yeast SUMO ) . After normalizing the Smt3 signal for Gcn4 abundance in the immune complexes , we observed that the Gcn4 present in hom6Δ srb10Δ cells after 2 h of SM treatment has an ∼2-fold higher level of sumoylation than observed in WT or hom6Δ cells under the same conditions , whereas sumoylation of Gcn4 is ∼2-fold lower in hom6Δ pho85Δ compared to WT or hom6Δ cells ( Fig . 7A–B ) . The finding that sumoylated Gcn4 is elevated specifically in the srb10Δ hom6Δ strain supports the idea that sumoylation of Gcn4 increases during SM-starvation of hom6Δ cells and that Srb10 targets the sumoylated Gcn4 for clearance from target promoters . As a result of Srb10 function , the proportion of Gcn4 that is sumoylated should not increase in response to ASA accumulation in hom6Δ SRB10 cells , as we observed ( Fig . 7B ) . By the same token , the fact that elimination of PHO85 from hom6Δ cells does not significantly alter the sumoylation of Gcn4 following SM treatment implies that Pho85 plays little role in clearing sumoylated Gcn4 and is therefore restricted primarily to clearing unsumoylated Gcn4 , which would include Gcn4 molecules not bound to the UASGCRE . It was shown that sumoylation of Gcn4 at Lys50 and Lys58 contributes to clearing Gcn4 from the UAS via Srb10 phosphorylation in the early stages of SM induction; and this process is eliminated by arginine substitutions at both Lys residues [17] . To determine whether sumoylation of Gcn4 stimulates the clearing of Gcn4 from promoters on ASA accumulation , we examined the effects of the K50R and K58R substitutions on Gcn4 abundance in SM-treated hom6Δ cells . We found that the Gcn4-K50R , K58R mutant displayed a reduction in abundance on SM-treatment of hom6Δ cells very similar to that observed for WT Gcn4 ( Fig . 7C–D ) . The K50R , K58R substitution also had no effect on Gcn4 abundance in SM-treated pho85Δ and pho85Δ hom6Δ cells ( Fig . S6 ) , where turnover of Gcn4 is dependent on Srb10 , thus suggesting that Srb10-dependent degradation of Gcn4 on ASA accumulation is not enhanced by sumoylation of Lys50/Lys58 . We also found that , in otherwise WT cells , the Gcn4-K50R , K58R variant confers essentially WT SM-induction of the UASGCRE-CYC1-lacZ reporter and HIS4 and ARG1 mRNAs , in accordance with previous findings [17]; and that the Gcn4-K50R , K58R variant resembles WT Gcn4 in being unable to sustain efficient SM-activation of UASGCRE-CYC1-lacZ , HIS4 and ARG1 expression in hom6Δ cells ( Fig . 7E–F ) . Thus , although our data suggest that sumoylation of promoter-bound Gcn4 increases on ASA accumulation , and that the sumoylated Gcn4 molecules are cleared from the promoter primarily by Srb10 , as concluded previously [17] , the sumoylation of Lys50/Lys58 is not critically required for the enhanced degradation of Gcn4 that occurs under conditions of ASA excess .
In this report we have shown that accumulation of ASA in hom6 mutants lacking functional homoserine dehydrogenase ( HSD ) impairs the GAAC response to starvation for Ile/Val by accelerating degradation of the activator Gcn4 . It is remarkable that ASA accumulation increases the rate of Gcn4 turnover considering that Gcn4 is already exceedingly short-lived under normal growth conditions [11] , [12] , [14] . The effect of ASA accumulation in reducing Gcn4 abundance and occupancy of the ARG1 UAS was mitigated in mutants lacking either of the CDKs , Srb10 and Pho85 , known to phosphorylate Gcn4 and target it for ubiquitylation and rapid degradation by the proteasome . Deletion of SRB10 restored an essentially WT level of cellular Gcn4 in Ile/Val-starved hom6Δ cells , whereas deletion of PH085 conferred an even greater than WT level of cellular Gcn4 in starved hom6Δ cells . Similarly , mutating a key phosphorylation site of Pho85 and possibly Srb10 , Thr-165 , also rescued WT Gcn4 abundance in Ile/Val-starved hom6Δ cells . These findings are consistent with the model that ASA accumulation evokes an increased rate of phosphorylation-dependent degradation of Gcn4 by the proteasome . While both Srb10 and Pho85 are required for the accelerated Gcn4 turnover , it appears that Pho85 plays the larger role—just as observed under normal growth conditions [12] . This last conclusion is consistent with our finding that UASGCRE-binding by Gcn4 is dispensable for its rapid turnover on ASA accumulation , which is also the case under normal growth conditions [6] . Interestingly , the outcome on the GAAC response differed significantly depending on which of the two CDKs was eliminated in hom6Δ cells . On removal of Srb10 from hom6Δ cells , the recovery of UAS-bound Gcn4 was accompanied by only a small increase in transcriptional activation of ARG1 , such that srb10Δ hom6Δ cells cannot grow on SM medium . By contrast , hom6Δ cells lacking Pho85 can grow on SM medium , and we observed an even greater than WT activation of ARG1 transcription conferred by essentially the same level of UAS-bound Gcn4 seen in hom6Δ srb10Δ cells , which is actually less than the UAS occupancy of Gcn4 found in fully WT cells . It could be argued that the low-level activation of ARG1 transcription seen in the srb10Δ hom6Δ double mutant reflects the requirement for Srb10 for efficient activation by Gcn4 observed previously [7] , [10] . However , here we observed no effect of deleting SRB10 on cell growth , and little or no effect on the induction of ARG1 and HIS4 mRNAs or PolII occupancy of ARG1 CDS in otherwise WT SM-treated cells; and the small defects we observed seem inadequate to explain the nearly complete absence of increased ARG1 and HIS4 transcription and PolII occupancy at ARG1 occurring in SM-treated hom6Δ srb10Δ cells . Hence , we favor the alternative explanation that the specific activity of the UAS-bound Gcn4 rescued by eliminating Srb10 in hom6Δ cells is lower than that rescued by eliminating Pho85 in hom6Δ cells . This in turn suggests that these CDKs target different populations of Gcn4 . The notion that Srb10 and Pho85 recognize different populations of Gcn4 also fits with our demonstration that Gcn4 is more highly sumoylated in Ile/Val-starved srb10Δ hom6Δ cells than in starved WT or hom6Δ cells , whereas Gcn4 is hypo-sumoylated in Ile/Val-starved pho85Δ hom6Δ cells . This finding is consistent with the previous conclusion that Srb10 is required to clear sumoylated Gcn4 from promoters [17] . Hence , we suggest that the putative population of defective Gcn4 molecules that are phosphorylated by Srb10 and subsequently degraded also tend to be hyper-sumoylated . However , we found that sumoylation of the known sites of this modification in Gcn4 , Lys50/Lys58 , was unimportant for the accelerated degradation of Gcn4 in Ile/Val-starved hom6Δ cells . Thus , while sumoylation appears to be a characteristic of Gcn4 molecules that are phosphorylated by Srb10 and subsequently cleared from the promoter , we have no evidence that sumoylation enhances the unusually rapid degradation of these Gcn4 molecules that occurs during ASA accumulation . To explain in greater detail our proposal that Srb10 and Pho85 target distinct populations of Gcn4 , we begin by positing that phosphorylation of the Gcn4 activation domain ( AD ) by Srb10 and Pho85 occurs most rapidly when the AD is not engaged with coactivators at the promoter . Hence , both functional and non-functional Gcn4 molecules not bound to the UAS would be susceptible to rapid turnover , whereas UAS-bound Gcn4 would turn over more slowly unless it harbors a damaged or modified AD that cannot engage with coactivators . Pho85 is located in the nucleus [6]; however , we have observed only low-level recruitment of Pho85 to the ARG1 UASGCRE by ChIP analysis , at a level decidedly smaller than that seen for Srb10 ( Fig . S7 ) or other Mediator subunits [38] . Moreover , Pho85 is responsible for the majority of Gcn4 degradation under both nonstarvation conditions and moderate-starvation conditions where the Pho85/Pcl5 complex is abundant [12] , which includes our SM-induction conditions . This can explain the previous finding that Gcn4 DNA binding activity is dispensable for rapid Gcn4 turnover under such conditions [6] . Thus , we envision that Pho85 primarily targets Gcn4 molecules when they are not bound to a UASGCRE . In contrast , Srb10 is recruited by Gcn4 to the ARG1 promoter and , hence , likely plays a prominent role in the degradation of UAS-bound Gcn4 molecules that become disengaged from coactivators either stochastically or because of damage or modification of the AD ( Fig . 8A ) . Again , this proposal is consistent with the previous finding that Srb10 is required to clear sumoylated Gcn4 , as sumoylation is impaired by mutations that impair UAS-binding by Gcn4 [17] . There is evidence that Gcn4 is deactivated under normal growth conditions by Srb10 and Pho85 , and that the phosphorylated , inactive protein must be degraded by the proteasome to prevent a reduction in the specific activity of UAS-bound Gcn4 [19] . We suggest that ASA accumulation in hom6Δ cells provokes damage or modification of Gcn4 that increases its rate of phosphorylation and subsequent turnover by the proteasome . The inability of the putative damaged or modified Gcn4 molecules to bind to the UAS or engage with coactivators could be responsible for their enhanced phosphorylation . Because eliminating the DNA binding activity of Gcn4 did not abolish its rapid turnover in hom6Δ cells harboring an intact GAAC , and DNA binding is not required for the Pho85-dominated turnover of Gcn4 under normal conditions [6] , [12] , we propose that Pho85 plays a predominant role in targeting the putative defective Gcn4 molecules generated under ASA excess , presumably when they are dissociated from the promoter; whereas Srb10 would make a lesser contribution and mediate the rapid degradation of defective , UAS-bound Gcn4 species ( Fig . 8B ) . Accordingly , eliminating Srb10 will spare from degradation defective Gcn4 molecules that are capable of UAS binding , and because Pho85 will continue to target functional molecules when they become disengaged from the UAS , the specific activity of UAS-bound Gcn4 should decline in srb10Δ cells , as we observed . By contrast , eliminating Pho85 will rescue both damaged molecules incapable of UAS binding as well as functional Gcn4 molecules that are phosphorylated by Pho85 when they disengage from the UAS; and because Srb10 will continue to clear activation-defective Gcn4 species capable of UAS binding , the specific activity of UAS-bound Gcn4 should increase in pho85Δ cells , as we observed ( Fig . 8B ) . Even in HOM6 pho85Δ cells , where no ASA accumulation occurs , the specific activity of UAS-bound Gcn4 exceeds that in WT cells , as seen both here and previously [17] , and this phenomenon is also explained by our model ( Fig . 8A ) . The proposal that Pho85 is responsible for clearing defective molecules incapable of binding to the UAS can also explain why a sizeable fraction of the Gcn4 spared from degradation in pho85Δ cells appears to be incapable of UAS binding , as indicated by the lower than WT UAS occupancy despite higher than WT cellular abundance of Gcn4 seen in pho85Δ and pho85Δ hom6Δ cells ( Fig . 6C vs . Fig . 5C–D ) . As noted above , it is also possible that the lower than WT UAS-occupancy of Gcn4 in pho85Δ cells reflects an unknown feedback regulatory mechanism that limits Gcn4 binding to the UAS as a way to prevent hyperactivation of Gcn4 target genes beyond the elevated levels seen in pho85Δ cells . An intriguing observation not anticipated by the model in Fig . 8 is that ASA accumulation provoked by SM-treatment of hom6Δ pho85Δ cells leads to a higher level of Gcn4 than occurs in HOM6 pho85Δ cells where ASA does not accumulate ( Fig . 5C–D ) . We recently obtained evidence that most of this effect can be accounted for by an unexpected increase in GCN4 transcription or translation , as expression of the GCN4-lacZ reporter was found to be ∼2-fold higher in SM-treated hom6Δ pho85Δ versus HOM6 pho85Δ cells ( Fig . S8 ) . A final interesting question is whether attenuation of the GAAC evoked by ASA accumulation is adaptive in WT yeast in the wild . Perhaps the enzyme HSD is frequently targeted for inhibition by plants , animals , or other microorganisms as a means inhibiting yeast growth . Indeed , the threonine pathway does not exist in mammals and has been identified as a valuable target for developing new antifungal therapeutics [24] . Moreover , it was shown that a strain of Streptomyces produces a natural antibiotic that targets HSD [39] . Reducing threonine biosynthesis by inhibiting HSD should activate eIF2α phosphorylation by Gcn2 and thereby reduce general protein synthesis , which is an appropriate response to limitation for threonine as a means of reducing the rate of threonine consumption . However , the concurrent transcriptional induction of Gcn4 target genes , including threonine biosynthetic pathway genes , evoked by translational upregulation of GCN4 mRNA might not be adaptive in this instance , owing to the toxic effects of ASA on cell physiology , including cytokinesis [27] . This toxicity of ASA provides a plausible rationale for the ability of this intermediate to suppress GAAC by accelerating Gcn4 turnover in the manner discovered here .
Yeast strains were grown at 30°C in rich YPD medium ( 1% yeast extract , 2% peptone and 2% glucose ) or defined synthetic complete ( SC ) medium ( 1 . 45g yeast nitrogen base , 5g ammonium sulfate , 2% glucose and 2g amino acid mix per liter ) lacking leucine , uracil or histidine wherever appropriate for selection of plasmids; and lacking isoleucine and valine ( Ile/Val ) for treatment with sulfometuron ( SM ) at 0 . 5 µg/ml . Increasing threonine in SC medium from approximately 1 mM to 2 . 5 mM diminished the slow growth phenotype of hom6Δ cells , and eliminated that of thr1Δ and thr4Δ cells , in SC medium lacking Ile/Val . Therefore , overnight growth to saturation was achieved in SC medium supplemented with 2 . 5 mM threonine and thereafter yeast strains were cultured in SC with 1 mM threonine . Accordingly , a moderate threonine limitation was imposed in our experiments and , as threonine is a precursor in Ile/Val biosynthesis ( Fig . 1A ) , this should intensify the limitation for Ile/Val provoked by SM treatment . Yeast strains used in this study are listed in Table 1 . Yeast strains purchased from Research Genetics or previously reported were verified for all auxotrophic requirements indicated in the genotype; and gene deletions were confirmed by PCR amplification of predicted deletion junctions using primers described in Table S1 . To generate HOM6 deletion strains , the appropriate hphMX4 gene deletion cassette conferring hygromycin B resistance [40] was PCR-amplified from plasmid pAG32 using primers HOM6-MX4-F and HOM6-MX4-R , thus introducing homologous flanking sequences upstream and downstream of HOM6 coding sequences , and used to delete HOM6 by transforming the appropriate strains to hygromycin B resistance on YPD agar plates . HOM6 deletion was further confirmed by demonstrating acquisition of threonine auxotrophy , except when deleted in hom2Δ or hom3Δ strains F2057 and F1929 , respectively; and by PCR-amplification of predicted junction fragments containing hphMX4 and sequences upstream or downstream of HOM6 coding sequences using primer pairs HOM6-A/HphMX-R1 and HOM6-DN-R/HphMX-F1 respectively . To generate SRB10-myc13 and PHO85-myc13 strains , a myc13::HIS3MX6 cassette was PCR-amplified from plasmid pFA6a-13myc-HIS3MX6 using primer pairs SRB10-MYC13-F/SRB10-MYC13-R or PHO85-MYC13-F/PHO85-MYC13-R , respectively , and used to transform strains BY4741 and YR001to His+ . Cassette insertions were confirmed by PCR analysis of genomic DNA using the appropriate primers specific for the myc13::HIS3MX6 cassette and SRB10 or PHO85 , and by Western analysis of whole cell extracts ( WCEs ) using anti-Myc antibodies ( Roche ) . To generate GCN4 deletion strains , plasmid pHQ1240 containing a gcn4Δ::hisG::URA3::hisG cassette was digested with SspI and used to transform strains F947 and YR006 to Ura+ . Deletion of GCN4 was indicated by acquisition of SM-sensitivity and verified by PCR amplification of gcn4Δ::hisG::URA3::hisG from chromosomal DNA using primer pairs specific for sequences upstream and downstream of the GCN4 CDS . The URA3 gene was subsequently evicted by selecting for growth on medium containing 5-fluoroorotic acid . All plasmids used in this study are listed in Table 2 , and primers used in plasmid constructions are listed in Table S1 . To construct pYPR010 , HOM6 ( chrX:689 , 322 . 690 , 749 ) was PCR-amplified from chromosomal yeast DNA of strain BY4741 , using primers HOM6-HindIII-F and HOM6-BamHI-R , and inserted between the HindIII and BamHI sites in YCplac111 . To construct pYPR018 , pYPR020 , pYPR022 and PYPR024 , HOM6 in pYPR010 was mutagenized using the QuikChange II XL Site-Directed Mutagenesis Kit ( Stratagene ) to produce hom6 mutant alleles encoding the K117A , E208D , E208L and D219L substitutions , respectively , using sets of complementary primer pairs harboring the corresponding mutations ( Table S1 ) . To construct pYPR028 , HOM3 ( chrV:256 , 132 . 258 , 737 ) was PCR-amplified from chromosomal yeast DNA of BY4741 using primers HOM3-F1 and HOM3-R1 and inserted between the SpeI and EcoRI sites in pRS313 . pYPR030 , containing HOM3-E282D ( HOM3fbr ) , was generated by fusion-PCR using HOM3-F1 and HOM3-R1 as outside primers and complementary primers HOM3-E282D-F and HOM3-E282D-R encoding the appropriate mutation , and inserted between the SpeI and EcoRI sites of pRS313 . To construct pYPR013 , the ApaI-SpeI fragment containing GCN4 was isolated from plasmid p164 , polishing the ApaI end using Klenow polymerase exonuclease activity , and inserted between the SmaI and SpeI sites of YCplac111 . pYPR038 and pYPR047 were constructed by fusion-PCR using Gcn4c-SphI-F and GCN4c-SpeI-R as outside primers in combination with primers GCN4-K50 , 58R-F and GCN4-K50 , 58R-R or primers GCN4-T165A-F and GCN4-T165A-R , respectively , and pYPR013 as PCR template . The PCR products were inserted between the SphI and SpeI sites in YCplac111 . Yeast strains transformed with plasmids pHYC2 ( UASGCRE-CYC1-lacZ ) , p367 ( HIS4-lacZ ) , or p180 ( GCN4-lacZ ) were grown to saturation and diluted in two identical cultures in SC-Ura/Ile/Val at A600 = 0 . 5 , and after 2 . 5 h of growth , 0 . 5 µg/ml SM was added to one set of cultures . Cells were harvested from untreated ( unstarved ) cultures after a total 6 h of growth and SM-treated cultures grown for 6 h in the presence of SM [7] . Whole cell extracts ( WCEs ) were prepared and assayed for β-galactosidase activity as previously described [41] . Mean specific activities were calculated from results obtained from three independent transformants . Yeast strains were cultured to an A600 of 0 . 4–0 . 6 in SC-Ile/Val , achieving at least two cell doublings , and treated with 0 . 5 µg/ml SM for the indicated times or left untreated . Total RNA was isolated by hot phenol extraction as previously described [42] . RNA concentration was quantified by Nanodrop spectroscopy and analyzed for integrity by agarose gel electrophoresis and ethidium bromide staining . An aliquot of 1 µg total RNA was used for cDNA synthesis using SuperScript III First-strand Synthesis Supermix for qRT-PCR ( Invitrogen ) and the resulting cDNA was diluted 10-fold . qRT-PCR was performed using Brilliant III Ultra-Fast qPCR Master Mix ( Agilient Technologies ) using the diluted cDNA in multiplex PCR and the appropriate TaqMan probes ( Table S1 ) to quantify ACT1 ( labelled with FAM ) , ARG1 , or HIS4 ( both labelled with HEX ) . qRT-PCR reactions were performed in triplicate using cDNA synthesized from RNA extracted from at least two independent cultures . ARG1 or HIS4 cDNA abundance was normalized to that of ACT1 by calculating 2 ( −ΔCt ) , where ΔCt is ( Ct ( Target ) - Ct ( ACT1 ) ) . Fold changes in mRNA abundance were normalized to those measured in uninduced WT cells , or as indicated , and plotted . ChIP assays were conducted essentially as described previously [7] , [43] . Yeast strains were cultured in 100 ml SC-Ile/Val as described above for RNA isolation , treated with 0 . 5 µg/ml SM for 2 h or as indicated , cross-linked for 15 min with 10 ml formaldehyde solution ( 50 mM HEPES KOH , pH 7 . 5 , 1 mM EDTA , 100 mM NaCl and 11% formaldehyde ) and quenched with 15 ml of 2 . 5 M glycine . WCEs were prepared by glass beads lysis in 400 µl FA lysis buffer ( 50 mM HEPES KOH , pH 7 . 5 , 1 mM EDTA , 150 mM NaCl , 1% TritonX-100 and 0 . 1% Na-deoxycholate ) with protease inhibitors for 45 min at 4°C and the supernatant collected after removing the beads was pooled with 600 µl FA lysis buffer used for washing the beads . The resulting lysate was sonicated to yield DNA fragments of 300–500 bp and cleared by centrifugation . 50 µl aliquots of lysates were immunoprecipitated for 2 h at 4°C with α-Gcn4 , ( Rabbit ) [13] or α-Rpb3 antibodies ( Mouse , Neoclone ) coupled with α-rabbit IgG or α-mouse IgG conjugated magnetic beads ( Dynabeads , Invitrogen ) , respectively , or with α-c-Myc ( Rabbit , Roche ) coupled with α-rabbit IgG conjugated magnetic beads . Recovered immune complexes were washed and eluted as described [43] . For matched input and IP samples , the crosslinks were reversed by incubation at 65°C overnight , treated with proteinase K , extracted twice with phenol:chloroform:isoamyl alcohol ( 25∶24∶1 ) and once with chloroform:isoamyl alcohol ( 24∶1 ) , and ethanol precipitated , resuspending the resulting pellets in 30–40 µl TE containing RNAase as described earlier [43] . Quantitative PCRs were performed in the presence of [33P]-dATP with undiluted IP DNA and 500-fold diluted input DNA and further analyzed as previously described [7] , [43] . The primers employed for ChIP analysis are listed in Table S1 . WCEs were prepared in denaturing conditions with trichloroacetic acid , as described previously [44] and analyzed by immunoblotting with α-Gcd6 [45] and affinity purified α-Gcn4 antibodies [13] . Western signals were quantified by ImageJ software . The analysis was performed essentially as previously described [13] , [46] . Yeast cells collected from a 10 ml culture at A600 = 0 . 4–0 . 6 were washed with SC-Met/Ile/Val , inoculated into 0 . 5 ml SC-Met/Ile/Val containing 1 µg/ml SM and incubated for 15 min in a shaking water bath at 30°C; after which 1 . 0 mCi [35S]methionine/cysteine labelling mix was added and incubation continued for an additional 15 min . Cells were collected , transferred to 5 ml of pre-warmed SC-Ile/Val containing 10 mM methionine and 10 mM cysteine , and an 1 ml aliquots were removed immediately or after appropriate times of chase . Aliquots were denatured with 170 µl of 1 . 85 M NaOH , 7 . 4% 2-marcaptoethanol and precipitated with 70 µl 100% TCA on ice , washed with chilled acetone and dried under vacuum in a SpeedVac . The dried pellets were resuspended in 120 µl of 2 . 5% SDS , 5 mM EDTA , 1 mM PMSF by vortexing , boiled for 1 min , and cleared by centrifugation . Incorporation of label was measured by scintillation counting [13] and aliquots of extract containing equal amounts of radioactivity ( 5 . 7×105 cpm ) were combined with 1 ml of immunoprecipitation ( IP ) buffer ( 50 mM Na-HEPES [pH 7 . 5] , 150 mM NaCl , 5 mM EDTA , 1% Triton X-100 , 1 mM PMSF ) containing 1 mg/ml BSA and 1 µl affinity-purified α-Gcn4 antibodies and mixed by rotating at 4°C for 2 h . Twenty µl of a 50% slurry of protein A-agarose beads pretreated with IP buffer containing BSA ( 1 mg/ml ) was added , and mixing continued for 2 h . The beads were washed thrice with 500 µl cold IP buffer containing 0 . 1% SDS , resuspended in loading buffer , boiled , and resolved by SDS-PAGE using 4 to 20% gels . The gel was dried and subjected to autoradiography , and the [35S]-labeled Gcn4 was quantified by phosphorimaging analysis . A modification of a previously described protocol was employed [47] , as follows . Yeast strains were cultured and treated with SM as described above for ChIP analysis . 40–60 A600 units of cells were lysed at 4°C with glass beads by 10 cycles of vortexing , 30s-on and 30s-off , in 500 µL of chilled lysis buffer ( 50 mM Tris-HCl [pH 8 . 0] , 5 mM EDTA , 150 mM NaCl , 0 . 2% Triton ×100 and 1 mM PMSF ) containing 10 mM sodium ethyl maleimide ( NEM ) and protease inhibitors . The resulting lysate was cleared by centrifugation at 13 , 000 rpm for 30 min at 4°C and soluble protein concentration was determined by the Bio-Rad protein assay . For each sample , a 40 µl suspension of magnetic beads conjugated with α-Rabbit IgG ( Dynabeads , Invitrogen ) was washed twice with lysis buffer containing 5 mg/mL BSA and rotated with 1 µl affinity purified α-Gcn4 antibody [13] in 200 µl lysis buffer/BSA for 3 h at 4°C . The magnetic beads coupled with α-Gcn4 antibody were washed twice with lysis buffer/BSA to remove unbound antibody and resuspended in 200 µl lysis buffer/BSA . Aliquots containing 1 mg of protein were added to the magnetic beads suspension , adjusting the final volume to 500 µl with lysis buffer/BSA , and further rotated for 2 h at 4°C . IP samples were washed thrice with lysis buffer containing 0 . 1% SDS and resuspended in 30 µl 1× Novex tris-glycine SDS sample buffer ( Invitrogen ) and boiled for 3 min . Aliquots of 5 µl and 25 µl were subjected to Western analysis with α-Gcn4 antibodies [13] and α-SUMO ( α-Smt3 ) polyclonal antibodies [47] .
|
Transcriptional activator Gcn4 maintains amino acid homeostasis in budding yeast by inducing multiple amino acid biosynthetic pathways in response to starvation for any amino acid—the general amino acid control . Gcn4 abundance is tightly regulated by the interplay between an intricate translational control mechanism , which induces Gcn4 synthesis in starved cells , and a pathway of phosphorylation and ubiquitylation that mediates its rapid degradation by the proteasome . Here , we discovered that accumulation of a threonine biosynthetic pathway intermediate , β-aspartate semialdehyde ( ASA ) , in hom6Δ mutant cells impairs general amino acid control in cells starved for isoleucine and valine by accelerating the already rapid degradation of Gcn4 , in a manner requiring its phosphorylation by cyclin-dependent kinases Cdk8/Srb10 and Pho85 . Interestingly , our results unveil a division of labor between these two kinases wherein Srb10 primarily targets inactive Gcn4 molecules—presumably damaged under conditions of ASA excess—while Pho85 clears a greater proportion of functional Gcn4 species from the cell . The ability of ASA to inhibit transcriptional induction of threonine pathway enzymes by Gcn4 , dampening ASA accumulation and its toxic effects on cell physiology , should be adaptive in the wild when yeast encounters natural antibiotics that target Hom6 enzymatic activity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"and",
"life",
"sciences"
] |
2014
|
Accumulation of a Threonine Biosynthetic Intermediate Attenuates General Amino Acid Control by Accelerating Degradation of Gcn4 via Pho85 and Cdk8
|
Biological systems are inherently variable , with their dynamics influenced by intrinsic and extrinsic sources . These systems are often only partially characterized , with large uncertainties about specific sources of extrinsic variability and biochemical properties . Moreover , it is not yet well understood how different sources of variability combine and affect biological systems in concert . To successfully design biomedical therapies or synthetic circuits with robust performance , it is crucial to account for uncertainty and effects of variability . Here we introduce an efficient modeling and simulation framework to study systems that are simultaneously subject to multiple sources of variability , and apply it to make design decisions on small genetic networks that play a role of basic design elements of synthetic circuits . Specifically , the framework was used to explore the effect of transcriptional and post-transcriptional autoregulation on fluctuations in protein expression in simple genetic networks . We found that autoregulation could either suppress or increase the output variability , depending on specific noise sources and network parameters . We showed that transcriptional autoregulation was more successful than post-transcriptional in suppressing variability across a wide range of intrinsic and extrinsic magnitudes and sources . We derived the following design principles to guide the design of circuits that best suppress variability: ( i ) high protein cooperativity and low miRNA cooperativity , ( ii ) imperfect complementarity between miRNA and mRNA was preferred to perfect complementarity , and ( iii ) correlated expression of mRNA and miRNA – for example , on the same transcript – was best for suppression of protein variability . Results further showed that correlations in kinetic parameters between cells affected the ability to suppress variability , and that variability in transient states did not necessarily follow the same principles as variability in the steady state . Our model and findings provide a general framework to guide design principles in synthetic biology .
Biological systems are complex , inherently noisy and only partially understood [1]–[5] . Systems and synthetic biologists are striving to better understand these systems , as well as to discover generally applicable principles for controlling them in biomedical and biotechnological applications . For example , a branch of systems biology studies how to best interfere with variable and under-characterized signaling pathways to identify novel drug targets [6]–[8] . In clinical pharmacology , decisions need to be made by using uncertain models of drug effects on the human body across populations and should ideally be robust to patient-to-patient differences [9] , [10] . Synthetic biologists strive to build synthetic circuits that perform desired functions across a population of cells , despite their noisy nature , cell-to-cell variability , and changing environments [11] , [12] . We are faced with a challenge of how to best represent , simulate , analyze , and carry out design for noisy systems with under-characterized biochemical properties , which are often represented as uncertain parameters . We propose a modeling and simulation framework as a tool to aid in meeting these challenges . In this manuscript we specifically focus on making design decisions for synthetic genetic networks , but the modeling technique is sufficiently general to be applicable to a wide range of problems in biology , biological engineering , and medicine . The framework accounts for different sources of variability and is computationally efficient , so that it allows screening across broad parameter ranges . Intrinsic variability ( or intrinsic stochasticity ) is a relatively well understood aspect of biological models . It arises from the probabilistic nature of the timing of collision events between reacting biological molecules , and its effect is most pronounced when the number of molecules in the system is small . Traditionally intrinsic variability is modeled by a stochastic master equation , which is the foundation for modeling stochastic dynamics in most physical , chemical , and biological phenomena [13] . Unfortunately , its analytic solution can only be found for a few trivial models , and a good alternative for studying stochastic models is the exact simulation framework of Gillespie [14] . However , to obtain a distribution resulting from the intrinsic variability , many trajectories of the Gillespie algorithm need to be simulated , which can be computationally expensive . A practical alternative is to use analytically tractable approximation schemes [15] , [16] . In this manuscript we approximate stochastic dynamics by van Kampen's -expansion [13] ( also called the linear noise approximation or perturbation expansion; see Methods ) , for its computational efficiency and analytic form . The -expansion separates the macroscopic dynamics from the fluctuations around it , describing each of these parts by a set of ordinary differential equations ( ODEs ) . This model allows for efficient propagation of the first two moments of the intrinsic noise distribution through time using deterministic equations for mean , variances , and covariances . The model very accurately approximates stochastic dynamics for medium and large molecular numbers and when variability is small compared to the mean number of molecules , but can lose on accuracy when numbers of molecules are very small and the relative fluctuation size increases [17]–[19]; here we check using Gillespie simulations that the -expansion distributions are accurate . Analytic studies of the accuracy of the -expansion and comparison with other traditional approaches for modeling intrinsic noise such as Fokker Planck equation and the chemical Langevin equation can be found in [17] , [18] , [20] and a summary on validity of the -expansion in the Supplementary material of [19] . Despite this caveat , and mainly due to its efficiency and possibility of analytic study , the -expansion has played an important role in advancing the understanding of intrinsic variability [3] , [18] , [21]–[24] . The -expansion can most conservatively be applied to models with a single steady state — in this manuscript we will only consider such models — but with certain limitations or modifications it can also be applied to multimodal and oscillatory models [22] , [25]–[27] . Significant effort has been invested in modeling intrinsic variability in systems and synthetic biology , although it has been shown that extrinsic variability generally dominates , especially in eukaryotic systems [2] , [28] . Extrinsic variability arises due to varying components upstream of the system of interest; these components affect the system , varying stochastically in time themselves , and might be present in different amounts in cells due to differences such as size and stage of the cell cycle [2] , [3] , [29] . For example , the numbers of ribosomes and the numbers of RNA polymerases vary in time and between cells . Another source of extrinsic variability is cell-to-cell variability of the gene copy number , which is common in synthetic biology applications when genes are delivered into cells by plasmid transfection , after which different numbers of plasmids are taken up by different cells . As a result of such sources of variability , single cells within a population possess distinct quantitative dynamic behaviors . As yet , there is no commonly accepted framework for modeling extrinsic variability . Despite several strong mathematical and theoretical studies of intrinsic and extrinsic variability [2] , [3] , [30] , [31] , computational modeling efforts that combine intrinsic and extrinsic variability are still rare . Shahrezaei et al . [32] proposed an extension to the Gillespie approach that includes kinetic parameter perturbations representing extrinsic variability; the downside of this method is that it is extremely costly . Scott et al . [26] proposed a more efficient , approximate model for steady-state extrinsic variability that can account for variations of one parameter at a time . Zechner et al . [33] used low-order moments through the moment closure approach to approximate intrinsic and extrinsic distributions; this approach requires analytic derivation of a new model structure for each additional extrinsic factor . Hallen et al . [34] proposed a non-mechanistic method of modeling extrinsic variability by perturbing the steady-state intrinsic noise distribution , but without any mechanistic assumption regarding the sources of extrinsic variability . Singh et al . [35] model extrinsic variability by adding noisy exogenous signals to an intrinsic stochastic model . Here we model extrinsic variability by introducing variability in model parameters and initial conditions; rather than considering them as point values , we consider them as distributions ( Figures 1 , S1 ) . We propagate these distributions through a model to simulate model output distributions resulting from extrinsic variability . For computational convenience we work with normally distributed parameters , ( for simplicity denotes a vector of all parameters and initial conditions ) . To simulate propagation of extrinsic variability through the model , we use the Unscented Transform ( UT ) . The UT efficiently maps the first two moments of the variability distribution in the parameter space onto the first two moments of variability distribution in the output . Estimates of the mean and covariance matrix obtained by the UT are accurate to second order in the Taylor series expansion for any nonlinear function , which makes the algorithm very appealing for propagating distributions through nonlinear functions [36] . Nonlinearity is propagated through simulating the nonlinear function for a chosen set of parameters ( called sigma-points , see Methods ) and reconstructing the output distribution from these individual simulations . This can capture nonlinearity such as a shifts in a mean , for example , when extrinsic variability increases . Despite knowing that extrinsic variability contributes significantly to the total variability , little is known about its sources [37] . This creates a major caveat that arises when attempting to make informed design decisions . The second ubiquitous caveat on the way to building predictive models of stochastic systems is that kinetic parameters are often unknown . Here we are motivated by a question of how to make robust design choices , given these uncertainties and limited knowledge of extrinsic variability . In this paper we use our method to study variability in simple gene regulatory networks; such networks are a basis for transcription-based synthetic circuits . We first introduce a combined intrinsic and extrinsic modeling approach and derive the expression for the total variability . The framework is then used to explore the effect of self-repression on protein variability in simple genetic networks that are simultaneously under different sources of noise . We are interested in details related to how self-repression can be used as an element in synthetic circuits to reduce variability . We consider two types of self-repression , on a transcriptional and on a post-transcriptional level , and ask which is more successful in reducing protein variability . We are further interested in specific design principles that help optimally achieve the aim of noise suppression .
The central methodological advance of this manuscript is a method and efficient framework to model total variability as a combination of intrinsic and extrinsic sources . Here we formalize the framework ( overview in Figure 1 ) and illustrate it on a simple model of transcription and translation of a single gene , with species mRNA , , and protein , ( Figure 2A ) [38] . The parameters of the model are transcription rate , , translation rate , , and mRNA and protein degradation rates , and , respectively . is the initial condition representing the copy number of genes encoding protein . We denote the vector of parameters and initial conditions by . Here we represent intrinsic variability through the -expansion , which provides a set of ordinary differential equations for the concentration means that are identical to a traditional ODE model without variability , plus an additional set of differential equations that describe time derivatives of the individual variances and covariances ( Figure 2B ) . All equations together approximate the time evolution of the intrinsic noise distribution . Note that the ODEs for the means do not depend on the variances and covariances but that the ODEs for the variances and covariances depend on means , variances , and covariances . Note also that ODEs for propagating the variance and covariance do not introduce any additional rate constant parameters beyond those needed to propagate the concentration means , but that initial values for variance and covariance need to be introduced in order to integrate their differential equations . We express this as a system of ordinary differential equations ( 1 ) ( 2 ) where is a vector representing the mean numbers of mRNA and protein and is the symmetric covariance matrix We simulated intrinsic variability in single gene expression using the -expansion model . Time-course trajectories of mean number of mRNA and protein ( red ) and their variances ( green ) are shown in Figure 2C . The Figure also shows individual trajectories from stochastic simulation runs ( blue ) using the Gillespie algorithm ( Methods ) [14] , which are consistent with the -expansion simulations . A more detailed comparison shows that the distributions from the -expansion and the stochastic Gillespie simulations are nearly identical ( Figure S2 ) . The multivariate Gaussian distribution is depicted by an ellipse with mean whose axes' directions are determined by eigenvectors and their sizes by eigenvalues of the covariance matrix ( Figure 2D , see Methods for further details ) . This represents variability in mRNA and protein counts due to intrinsic noise only . We next introduced extrinsic variability into the model , by introducing parameters that follow a distribution with a probability density function , rather than parameters fixed to a particular value . Operationally , the combined distribution is a superposition of intrinsic distributions for different parameter realizations sampled from the underlying parameter probability distribution , ( Figure 2E ) . The resulting output distribution represents the total variability ( i . e . , variability resulting from intrinsic as well as extrinsic sources ) . Mathematically we represent this distribution with a mixture model with probability density function ( 3 ) where is a probability density of the intrinsic noise distribution , in our specific case a normal distribution with mean and a covariance matrix resulting from the -expansion model with parameters . For single gene expression , the mixture model readsand its steady state is schematically depicted in Figure 2E . This framework allows us to calculate the mean and variance of the total variability distribution; the mean of a random variable drawn from the mixture model ( 3 ) was calculated as ( 4 ) and by using the law of total variance [39] we obtained the covariance matrix [30] ( 5 ) This equation shows that the total variability is decomposable into two parts: variability due to intrinsic stochasticity , , and variability due to extrinsic sources , ( Figure 2F ) . So far we have not restricted ourselves to any specific form of the parameter distribution ; in order to obtain a combined model and calculate the total variability , we generally need to exhaustively draw samples from and simulate the -expansion model for each sampled parameter . However , for normally distributed parameters , , we can take a considerably more efficient approach , the Unscented Transform ( UT , see Methods ) . The UT selects a small collection of representative control points in the parameter space ( called sigma points ) , propagates them through the -expansion simulation model to obtain the output instance for each , and then reconstructs the first two moments of the output distribution ( which is equivalent to the above mixture model distribution ) by appropriate weighting of the sigma points . From the output distribution , the total mean and the total variance are easily computed . See Text S1 for further details and an example for the single gene expression model . We checked that our novel combined simulation framework accurately approximates the exact total protein and mRNA distributions resulting from Gillespie simulation . We first introduced extrinsic variability into one parameter only , and then into four parameters . In the first case , approximate distributions resulting from the novel framework fit very well to exact distributions in both steady and transient states ( Figure S3 ) . In the second example with extrinsic variability in four parameters , the approximation was slightly worse; this is becasue the total variability distribution was not Gaussian , but skewed , and perhaps also to inexactness in the -expansion ( Figure S4 ) . However , our approximation approximated well the distribution up to the second moment , i . e . the mean and the variance . Figure 3 shows examples of interplay between intrinsic and extrinsic variability for distinct amounts of extrinsic variability . In this example we considered parameters to be independent ( i . e . , the covariance matrix had zero for all off-diagonal entries ) , and all parameters were varied with the same coefficient of variation ( see Methods ) . When the variability in parameters was low , intrinsic variability was dominant in the system ( Figure 3A ) ; this can be seen from the sizes of the green ellipses ( the “average” of the green ellipses corresponds to in Eq . 5 ) , which are much larger than the size of the red ellipse ( corresponding to ) . On the other hand , for high parameter variability the extrinsic variability became dominant ( Figure 3F; red ellipse is larger than the green ones ) . For intermediate parameter variability , both intrinsic and extrinsic sources contributed comparably to the overall variability . Notice also that with increasing extrinsic variability , the intrinsic components changed their size and shape; this corresponds to the observation that intrinsic variability can change with kinetic parameters [40] . The total variability is the combination of red and averaged green ellipses; with increasing extrinsic variability , the variability of the output increased . High extrinsic variability has also shifted the mean; for example protein mean has increased when extrinsic variability was increased ( this can be seen by comparing the x-component of the red ellipse mean on Figures 3A and 3F , where the mean of the red ellipse represents the mean of the output – combined intrinsic and extrinsic noise – distribution ) . In the remainder of this paper we report results obtained by using this framework to study how different topologies of self-repressing genetic networks affect the total variability in proteins . Negative feedback and self-repression are recurrent motifs observed in gene regulatory networks . Their potential role in suppressing variability has identified them as promising design elements in synthetic biology [32] , [41]–[48] . Here we studied the effects of self-repression on variability of the output protein in a single gene model , while taking into account both intrinsic and extrinsic sources of variability . We compared a case of negative autoregulation on a transcriptional level with a case of post-transcriptional regulation . In the transcriptional autoregulation case , the output protein acts as a repressor that multimerizes and binds to its own promoter region to repress its own transcription ( Figure 4A ) . In the post-transcriptional case , we model synthetic microRNA ( miRNA ) that binds to its cognate mRNA , simultaneously preventing translation and promoting degradation ( Figure 5 ) . In the remainder of this section we explore the effectiveness of these control mechanisms . The results provide specific guidelines for designing transcriptional and post-transcriptional autoregulatory networks for achieving best ( or worst ) suppression of protein variability .
Variability is ubiquitous to biological systems , and yet it is not well understood . In order to engineer and intervene in such systems successfully , variability must be taken into account in planning and interpreting experiments and in designing synthetic networks . Here we have introduced a framework for modeling and simulation of dynamical systems under both intrinsic and extrinsic sources of variability . One reason that understanding and treating variability is so especially important for synthetic biology is that one seeks to design circuits and devices that can perform successfully across a population of cells and variety of experimental conditions; without understanding variability and fluctuations , one risks building devices or designing therapies that only perform well in a fraction of cells ( or patients ) or only under limited environmental conditions . Here we addressed the problem of constructing circuits that exhibit low variability across cells . We investigated transcriptional and post-transcriptional designs of self repression , and studied their relative promise for noise suppression . We explored their behavior for a range of feedback strengths and other design parameters , and for various hypothetical amounts and types of extrinsic variability . We derived a set of design principles for best noise suppression in a protein concentration of interest . We showed that transcriptional autoregulation is more successful than post-transcriptional in suppressing noise under a wide range of intrinsic and extrinsic variability levels and conditions . The following design principles were shown to best suppress protein concentration variability: ( i ) high cooperativity in protein binding to DNA and low cooperativity in miRNA binding to mRNA , ( ii ) imperfect complementarity between miRNA and mRNA was preferred to perfect complementarity , and ( iii ) correlated expression of mRNA and miRNA — for example , on the same transcript . We further showed that correlations in kinetic parameters between cells affected the ability to suppress variability , and that variability in transient states did not necessarily follow the same principles as variability in the steady state . Also , we note that biochemical design can not avoid fundamental limits on the lower bound of stochastic fluctuations [57] . In this work we have focused on one objective , which is to decrease variability of protein expression . In synthetic biology applications , one might be interested in additional objectives , such as maximizing the mean protein expression , or keeping the mean fixed while suppressing variability etc . The same analysis can be performed for such objectives , and could potentially lead to different guidelines . Our novel modeling and simulation framework combines intrinsic and extrinsic sources of variability by combining the -expansion with the unscented transform . In contrast to few previous methods , our framework is based on a deterministic simulation with ordinary differential equations . One advantage of our method is that it incorporates extrinsic variability by repeated simulation of the same intrinsic model for carefully selected parameter combinations where the number of simulations increases linearly with the number of extrinsic factors ( see Methods ) , rather than generating a new model and exponentially increasing the model size for each new extrinsic factor [33] . This not only increases computational efficiency , but makes possible analysis with a wide variety of available tools . It also allows for screening across a large range of parameters ( such as feedback strength , and rates , and unknown sources and quantities of extrinsic variability ) . The new approach is well suited to support the design of synthetic circuits from building blocks with partially unknown kinetic properties but that will perform desired functions robustly in variable intracellular and extracellular environments and across a population of cells . An important aspect of the framework is that it can be applied to study variability and dynamics at transient as well as steady states; we have shown that variability might behave differently in both regimes , which reiterates the importance of developing and using tools that are applicable beyond steady-state regimes . The computational efficiency of the framework , however , comes at a cost of approximation; macroscopic dynamics remain non-linear , but intrinsic and extrinsic variability distributions are both approximated by the first two moments . We have shown that transient and steady-state distributions resulting from our approximation framework closely fit to exact distributions simulated by a Gillespie algorithm and parameter sampling . We note once more that the -expansion is derived under the assumption of large numbers of molecules and we recommend the results to be tested by more accurate methods , especially for low molecule numbers . For example , a sufficient number of Gillespie runs provides the exact distribution [14]; another powerful technique , the Langevin approximation , can be used to obtain more accurate than -expansion [20] , although still approximate estimates of the variance at a steady state , and its disadvantage is that it requires monitoring of the convergence . An important point , however , is that we have focused less on calculating exact quantitative values of variability , and more on what changes in the network drive variability increases or decreases . Our approximation framework is sufficient for addressing this question , and due to computational efficiency in both steady and transient states it is even advantageous compared to exact simulation . We have applied the framework to systems with a single steady state , but there is scope for extending the framework to oscillatory and multimodal systems [25]–[27] . Derivation of improved versions of the -expansion is an area of active research , for example accounting for slow and fast variables [58] or increasing accuracy for small number of molecules [59] . By developing better approximation frameworks for intrinsic noise modeling , our combined intrinsic-extrinsic method will increase in accuracy , too . Furthermore , while the dynamics originating from intrinsic variability was modelled by the -expansion approximation framework in this manuscript , a similar framework could be constructed based on other approximation modeling approaches for intrinsic variability , for example the moment closure method [60] or mass fluctuation kinetics [16] . The extrinsic variability modeled through parameter distributions in this manuscript was assumed to be constant in time ( i . e . , static ) , and a natural extension would be to introduce time-varying parameter distributions ( i . e . , dynamic ) [31] , [61] . Because of this , for example , our method in its current form can not be used to quantify the dependence of the total output variation on the lifetime of extrinsic fluctuations [32] . There are different sources of biological variability and uncertainty; a careful interpretation and appropriate inclusion in a mathematical model is crucial to elucidate the dynamics of biological variability and understanding of these different levels of biological complexity . Firstly , there are intrinsic and extrinsic sources of variability ( discussed at length above ) . Secondly , there is parameter uncertainty , which encompasses our limited knowledge of kinetic rate values . This uncertainty is commonly included in biological models and simulations as a distribution of kinetic values , however , it is important to treat it separately from extrinsic variability , which is also modeled by treating parameters as distributions . An example of appropriate mathematical treatment is a hierarchical Bayesian model [33] , [62] . A further type of uncertainty is model uncertainty , by which we refer to our uncertainty in model topology; the present manuscript does not address this type of uncertainty . And thirdly , when modeling or analyzing experimental biological data , one also needs to account for measurement noise ( or measurement error ) . In order to infer sources of variability and reduce uncertainty in parameters , there is a need for inference techniques for models that account for these different levels of biological variability , uncertainty , and measurement noise . The presented modeling framework is only the first step in the systems biology cycle of modeling , designing experiments , and updating the model by inference from collected data . Based on this modeling framework , we are now developing tools that account for variability also in the experimental design stage , and for updating the models from noisy data through parameter inference and model selection algorithms . These techniques will allow us to better explore and account for biological variability in synthetic biology , a ubiquitous aspect of biological systems that is in practice often neglected .
The -expansion is an approximation of the master equation . It separates the macroscopic part of dynamics from the fluctuations around it , describing each of these parts by a set of ODEs [13] . The resulting -expansion model approximates the first two moments of the intrinsic noise distribution in the form of ODEs for each mean , variance , and covariance; for species in the model , the -expansion generates equations for the means , equations for variances , and equations for covariances between all pairs of species . Readers interested in a theoretical derivation of the -expansion modeling framework are referred to van Kampen [13] . As a brief summary we note that the -expansion is derived assuming that the macroscopic part of dynamics can be separated from the fluctuations centred around it , and that the fluctuations scale as the square root of the number of molecules [63] . This is captured by the following ansatz: , where is the number of molecules , is the macroscopic concentration and the fluctuations . Parameter represents the system size and can be thought of as proportional to the volume . This ansatz is introduced into the master equation governing the evolution of probability distribution and then the master equation is Taylor-expanded around . The terms of order collected together determine equations governing macroscopic dynamics , and the terms of order the equations for dynamics of fluctuations . The resulting set of ordinary differential equations describing macroscopic ( or average ) behavior is ( 8 ) where is the stoichiometry matrix and are the reaction propensities of reactions . The ODE for the covariance matrix of the fluctuations , , centred around is ( 9 ) where is a Jacobian of a deterministic system given by equation ( 8 ) , evaluated along the macroscopic trajectory , and is called the drift . Matrix is the diffusion matrix , . Equations ( 8 ) and ( 9 ) together form the -expansion model , which approximates the first two moments ( mean and covariance matrix ) of the distribution that solves the master equation . The -expansion model is a deterministic and approximate description of intrinsic noise distribution dynamics in time . We visualised the multivariate normal distributions as ellipses . The ellipse is one of the contours that represent equal probability mass of the distribution . In general , these ellipses are constructed from eigenvalues and eigenvectors of the covariance matrix . They are centred at mean , their axis directions are determined by the eigenvectors and axis sizes by the eigenvalues , . Different values of specify different contours . In our figures we choose unless otherwise specified . In particular , the ellipse determined by equationrepresents the multivariate analogue of the “confidence interval” containing of the probability mass of the distribution . For example , a bivariate ( ) normal distribution represented by an ellipse with axes lengths , , contains of distribution , and an ellipse containing of a distribution has axes of lengths , . The unscented transform ( UT ) is an algorithm that efficiently propagates normally distributed inputs through a nonlinear function to obtain a distribution on the function outputs . Figure 13 provides an overview of the UT . Here was the multivariate normal distribution on the space of parameters and initial conditions of dimension . We deterministically sampled from a set of sigma points ( 10a ) ( 10b ) ( 10c ) and assigned weights ( 11a ) ( 11b ) ( 11c ) where . See ref . [36] for guidelines on how to choose parameters , and ; we chose , , . The above weights define the scaled unscented transform , which is the generalized version of the unscented transform algorithm [64] . Function was evaluated for every sigma point , The model outputs obtained were then reconstructed into a multivariate normal distribution with mean and covariance matrix determined by the following equations: ( 12a ) ( 12b ) ( 12c ) We note that the UT is computationally much more for efficient compared to random sampling; UT only requires 2L+1 simulations ( one for each sigma point ) , while this number needs to be considerably higher for exhaustive sampling , especially for high dimensional parameter spaces . In terms of accuracy , the UT estimates the first two moments accurately to second order in the Taylor series expansion for any nonlinear function [65] , [66] . However , should one sample 2L+1 points randomly and propagate them , one will not achieve this accuracy and therefore exhaustive sampling is needed to achieve the same level of accuracy as the UT ( the exact number of simulations needed depends on the specific system and space dimension of parameters and initial conditions ) . The UT can also be used to propagate log-normal distributions ( see Text S1 for details ) , and mixtures of ( log- ) normal distributions . The variance of protein numbers is not by itself a good measure of variability , as it does not account for the protein mean; for example , with increasing feedback strength in an autoregulatory network , the fluctuations in protein numbers decrease , but the mean number decreases , too . If absolute numbers of proteins are small , then protein variance becomes relatively large compared to its mean . For most applications , both mean and variance of the protein numbers should be included in the measure of variability . Here we use the squared coefficient of variation , [4] , normalised by , the measure of variability without self-repression ( stands for no self-repression ) , to obtain the relative coefficient of variation , . Therefore , will have value when no self-repression is present in the model . Whenever variability according to this measure is less than , self-repression is effective in suppressing variability . Extrinsic variability is described by the covariance matrix of parameters and initial conditions . Extrinsic variability can enter the system though variability in a single parameter or a combination of parameters , and parameters can be arbitrarily ( anti- ) correlated . A diagonal covariance matrix assumes independence between all pairs of parameters . ( Anti- ) correlations between parameters are modeled by including non-zero off-diagonal terms . The covariance matrix is symmetric by definition . We defined the amount of extrinsic variability by specifying a coefficient of variation for each parameter , ( 13 ) and set the diagonal entries of the matrix to . The parameter dimension is denoted by . We chose different amounts of extrinsic variability , as noted in specific figure captions . When independence between parameters was assumed , the off-diagonal terms were set to zero . Additionally , we also created arbitrarily correlated parameter distributions by generating random covariance matrices in the following manner . We first generated a random correlation matrix of size , by generating a symmetric matrix with diagonal elements and off-diagonal elements sampled from a uniform distribution , and then finding the nearest valid correlation matrix by using Higham's algorithm [67] . This correlation matrix was transformed to a covariance matrix by multiplying each element of the correlation matric by the standard deviations corresponding to its row and column , , where , were the variances ( i . e . , diagonal elements of the covariance matrix ) . The units are in the form of number of molecules ( or number of molecules/sec ) rather than concentration ( or concentration/sec ) . This allows us to more easily to compare the exact Gillespie simulation with the output of our approximation framework . Parameters used in examples in figures were set to and . Parameters that determine the feedback strength , and , were varied as indicated in the figures . The Hill coefficient took integer values between 1 and 4 . The units of zeroth-order parameters were number of molecules per second ( or molecules per second ) , first-order parameters per second and second-order parameters per molecule per second . The conclusions were checked to hold for other parameter combinations , in particular , , , , and .
|
Variability is inherent in biological systems , and in order to understand them , we need to be able to model different sources of variability . Systems have evolved to harness and control the variability , and more recently , synthetic biologists are trying to learn how to control variability in engineered biological systems . Several sources of variability exist; they arise due to stochastic expression of genes , which is most pronounced when numbers of mRNA and protein molecules are low , as well as due to differences between individual cells . Here we propose a modeling framework that combines different sources of biological variability . Furthermore , current research seeks to control biological variability though robust design of synthetic biological circuits , for example for use in therapies and other biomedical or biotechnological applications . Here we apply our framework to guide design of synthetic circuits that use transcriptional and post-transcriptional regulation to suppress variability in the output protein of interest . We find that certain properties and network designs are better than others in their ability to control variability , and here we report on the design guidelines to aid synthetic circuit design to suppress variability , in spite of our uncertain knowledge of parameters .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biotechnology",
"bioengineering",
"systems",
"biology",
"biochemical",
"simulations",
"biological",
"systems",
"engineering",
"synthetic",
"biology",
"biology",
"computational",
"biology",
"molecular",
"biology",
"genetics",
"and",
"genomics",
"engineering"
] |
2013
|
Combined Model of Intrinsic and Extrinsic Variability for Computational Network Design with Application to Synthetic Biology
|
Computational modeling of neuronal morphology is a powerful tool for understanding developmental processes and structure-function relationships . We present a multifaceted approach based on stochastic sampling of morphological measures from digital reconstructions of real cells . We examined how dendritic elongation , branching , and taper are controlled by three morphometric determinants: Branch Order , Radius , and Path Distance from the soma . Virtual dendrites were simulated starting from 3 , 715 neuronal trees reconstructed in 16 different laboratories , including morphological classes as diverse as spinal motoneurons and dentate granule cells . Several emergent morphometrics were used to compare real and virtual trees . Relating model parameters to Branch Order best constrained the number of terminations for most morphological classes , except pyramidal cell apical trees , which were better described by a dependence on Path Distance . In contrast , bifurcation asymmetry was best constrained by Radius for apical , but Path Distance for basal trees . All determinants showed similar performance in capturing total surface area , while surface area asymmetry was best determined by Path Distance . Grouping by other characteristics , such as size , asymmetry , arborizations , or animal species , showed smaller differences than observed between apical and basal , pointing to the biological importance of this separation . Hybrid models using combinations of the determinants confirmed these trends and allowed a detailed characterization of morphological relations . The differential findings between morphological groups suggest different underlying developmental mechanisms . By comparing the effects of several morphometric determinants on the simulation of different neuronal classes , this approach sheds light on possible growth mechanism variations responsible for the observed neuronal diversity .
Dendritic morphology underlies many aspects of nervous system structure and function . Dendrites , along with axons , define the connectivity of the brain [1] , [2] , and play a large role in information processing at the single cell level [3] , [4] . Many studies have highlighted the importance of dendritic branching pattern in neuronal behavior . Mainen and Sejnowski [5] have shown that the full range of firing patterns for a wide variety of cortical cell types can be accounted for by branching morphology alone . Others have shown that the backpropagation of action potentials into the dendrites is strongly affected by branching pattern [6] . These results , among others , have contributed to a now widespread acceptance that dendritic morphology is an essential substrate of brain activity and function . Despite its importance , dendritic branching remains poorly understood [7] . Dendritic branching is driven by a complex interaction of intracellular and extracellular signaling cascades which are proving difficult to completely unravel by molecular biology alone . The same chemical can have different effects in different cells [8] and even different parts of the same cells [9] . Much of the molecular work is carried out on cultured cells where separating apical and basal trees , and even dendrite from axons , is difficult ( for example see [10] ) . Computational modeling offers a complementary approach to traditional molecular means of uncovering fundamental properties of dendritic branching ( e . g . , [11] , [12] ) . Here we focus on data driven simulations , where the parameters controlling branching behavior are measured from real cells , reduced to statistical distributions , and resampled to form virtual trees ( e . g . . [13]–[16] ) . One advantage of this approach is the insights it gives into dendritic development . Many attempts have been made to model mechanistic aspects of dendritic development directly , such as MAP2 phosphorylation states [17] , or growth cone navigation [18]–[20] . Other models , while not aiming to represent developmental processes explicitly , can yield insights into general principles or specific mechanisms at play . For example , the 3D modeling approach used by Samsonovich and Ascoli [21] demonstrates the importance of somatic repulsive forces for the shaping of principal cells in the rat hippocampus . While data driven simulations have increased our understanding of dendritic development , they are difficult to compare directly . Different studies often focus on separate structural levels or details , and are rarely based on the same cell classes . Here we expand on previous approaches by testing a suite of three closely related models , both individually and in hybrid combinations . Also , because data driven modeling generally requires quality neuronal reconstructions , they tend to be limited to one or two dendritic tree types . From these studies , it is often difficult to determine how general the results are , and to discern biological insights from data or model peculiarities . With a large digital database of neuromorphological reconstructions now , online ( NeuroMorpho . Org ) , we were able to apply our models to a wide variety of dendritic trees ( Figure 1 ) from 16 different labs . This allows the separation of general trends from more specific model-morphology interactions . The core of our modeling approach is a recursive branching process as described in Figure 2A ( detailed in Materials and Methods and [22] ) . All of the basic parameters of the model ( defined in the five text boxes in Figure 2B ) are measured from each real cell and resampled to create virtual trees . Every branch in the real trees has an associated taper rate and pathlength , every bifurcation has a daughter diameter ratio , etc . With every basic parameter extracted from real cells , the accompanying fundamental determinant ( Figure 2C ) is also measured . For example , when measuring the taper rate of a real branch , the thickness ( radius ) , the number of bifurcations from the soma ( branch order ) , and the somatic path distance of that same branch are also recorded . Within each tree group ( e . g . , Martone's Purkinje ) , and for each of the three fundamental determinants , series of distributions are then generated which best describe each basic parameter for different bins of the fundamental determinant . For example , one distribution will describe all of the taper rate values which occur at Branch Order four . It is this distribution that will be sampled to select the taper rate every time a branch of order four is added to a virtual tree of this group based on this fundamental determinant ( as described in Figure 2A ) . This process is repeated for each of the five basic parameters , 68 groups of real cells , and three fundamental determinants . The term “fundamental determinants” is meant to describe the parameters which are primary in the model and drive the selection of other values , but should not be taken to imply that they are the only or most crucial developmental factors underlying branching behavior . The comparative approach constrains the choice of fundamental determinants to those compatible with the common mechanics of the model . Nevertheless , the chosen determinants are biologically important and have all been implicated by earlier studies ( reviewed in [7] ) in the control of bifurcation probability ( one of our basic parameters ) . Radius correlates with microtubule density [23] and has previously been shown to capture some , but not all , aspects of dendritic branching in several neuronal classes [13] , [22] . Branch order takes into account the division of resources from the soma and has been used to control the distribution of bifurcations in several computational models [11] , [14] . Path distance affects the time of subcellular transport and signaling to and from the soma , and has been useful in constraining motoneuron and pyramidal cell virtual growth [13] , [24] . In earlier efforts ( e . g . [16] , [25] ) , basic parameters were assumed to be uniformly distributed throughout the dendritic tree ( Figure 3 inset ) . While some cell types were well captured in this way , others resulted in virtual trees which continued to bifurcate indefinitely . Later studies [22] determined that this was due to basic parameter values being applied in the virtual trees where they did not occur in the real trees . For example , in the apical trees of one group of pyramidal cells ( Figure 3 ) , the daughter diameter ratio tends to be larger near the soma than farther distally . Most importantly , the proportion of bifurcations with two equally sized daughters ( unitary values of the diameter ratio ) is smaller close to the soma , where most of the bifurcations occurred in this case . Without grouping by fundamental determinant , these dependencies are not captured in the virtual trees . Using radius as a fundamental determinant for all basic parameters in CA1 pyramidal cells prevented the explosive virtual growth , but the resulting trees were still excessively varied in size [22] . The model also proved to be very sensitive to radius , a notoriously noise prone measurement in neuronal reconstructions . Here we expand on this work by applying three different fundamental determinants to a wider variety of tree types . The creation of three individual models with the same underlying mechanics also allows the implementation of hybrid variations . This step overcomes some limitations of the simpler models by introducing more freedom , but complicates biological interpretability . Most importantly , the details of how the mix models improve upon ( or do not ) the individual models provides information on the individual models themselves . We explored two alternatives ( detailed in Materials and Methods ) . In the first , each basic parameter was under the control of a separate fundamental determinant ( leading to 243 possible combinations ) . The second “Mix” strategy varied the proportional influence ( in %10 steps ) of each fundamental determinant in controlling all of the basic parameters . The comparative application of different but related models to extremely diverse morphological classes enables us to look both within and across cellular/subcellular features for parameter interactions . These interactions may then point to important developmental principles . Four biologically important morphometrics which are emergent to the model are used to compare the real and virtual trees ( Figure 2D , Materials and Methods ) . These morphometrics capture features related to both tree size and branch patterns , giving a relative measure of model behavior . A distance metric is used which takes into account both the differences between the means of the real and virtual trees , and the variability in model behavior ( see Materials and Methods for details ) . We find that the apical and basal arborizations of pyramidal cells differ more than groups of dendrites divided by other criteria ( such as tree size ) . We propose , based on the parameter interactions , that extracellular environment and intracellular competition for resources may be particularly important in the development of apical and basal tree types .
The three individual models were evaluated in terms of their ability to produce virtual trees with values of the emergent morphometrics that best matched the corresponding real trees . Strong trends were shown when considering all of the tree classes together ( Figure 4 ) . In terms of the ability of the three fundamental determinants to reproduce the number of bifurcations across the whole set of morphologies , Branch Order was the clear winner ( Figure 4A ) . The Branch Order model variant created trees which were significantly closer in number of bifurcations to the real trees than either Radius or Path Distance ( Figure 4A upper ) . In particular , the mean number of bifurcations of virtual trees differed by an average of only 10% from the measured ( real ) value . This relative difference was over twice and nearly three times as large for the models based on Path Distance and Radius , respectively . Still looking at number of bifurcations , Branch Order was also the best model ( assessed by the distance metric ) , for well over half of the 68 tree groups ( Figure 4A lower ) . While a model based on branch order may be expected to best control the number of branches , apical trees of pyramidal cells offer a striking exception to this general trend , which is discussed in depth below . In this sense , the comparative approach is particularly powerful by naturally providing biologically relevant “mutual” controls among the different morphological groups and model variants . Overall , bifurcation asymmetry was best determined by both Path Distance and Radius ( Figure 4B ) . Both Path Distance and Radius were significantly better than Branch Order ( Figure 4B upper ) , and were each determined to be the best for roughly twice as many tree groups as Branch Order ( Figure 4B lower ) . No fundamental determinant was significantly better than the others at determining surface area ( Figure 4C upper ) . Likewise , Path Distance , Radius , and Branch Order each best determined surface area for roughly one third of the tree groups ( Figure 4C lower ) . On the other hand , surface area asymmetry was overwhelmingly best determined by Path Distance ( Figure 4D ) . The relative difference for this emergent morphometric was on average half for the Path Distance model than for either of the other fundamental determinants . Moreover , 84% of the tree groups had their surface area asymmetry best reproduced by the Path Distance model ( Figure 4D lower ) . These trends were generally robust throughout individual tree groups . However , a finer analysis organized by morphological classes revealed additional insights . The tree groups were first divided into apical ( n = 18 ) , basal ( n = 18 ) , and non-pyramidal ( n = 32 ) . The Branch Order model was significantly better than either Radius or Path Distance at determining the number of bifurcations in both basal and non-pyramidal tree types ( Figure 5B and 5C ) . In particular , Branch Order “won” more than three quarters of the basal groups . This was definitely not the case for apical trees , where over half of the 18 groups had their number of bifurcations best determined by Path Distance ( Figure 5A ) . Figure 5D shows a more detailed analysis for a representative apical tree group . In this example , Path Distance better captures not only the mean , but also the pattern of bifurcations as a function of branch order ( “Sholl-like” plots ) . In contrast , when looking at basal trees from the same cells ( Figure 5E ) the Branch Order model provides a much better match to the real data . The situation is almost reversed if models are evaluated based on another emergent morphometric , namely bifurcation asymmetry instead of the number of bifurcations ( Figure 6 ) . Path Distance is the worst model at capturing apical asymmetry ( Figure 6A ) but the best at capturing basal asymmetry ( Figure 6B ) , both in terms of average distance ( top panels ) and numbers of groups ( bottom ) . Non-pyramidal cells fall in between apical and basal with both Radius and Path Distance producing the best results more often than Branch Order ( Figure 6C ) . Another example Sholl-like analysis carried out on an single group of pyramidal cells is consistent with the trends observed across the corresponding sets of tree types , and opposite to the patterns observed for number of bifurcations ( Figure 6D ) . In particular , the distribution of apical bifurcation asymmetry values as a function of branch order is better reflected by the Radius model than by the Path Distance model . Figure 6E shows that the converse is true for the basal trees from the same cells . While Figures 5 and 6 show that the interaction between fundamental determinants and emergent morphometrics is different for apical and basal trees , it is important to notice that the overall quality of the simulations is different as well , as becomes apparent when the units are on the same scale ( Figure 7 ) . Both Branch Order and Radius are better able to capture the number of bifurcations in basal than in apical arbors ( Figure 7A ) , but the inverse relation holds for bifurcation asymmetry ( Figure 7B ) . In both cases , non-pyramidal cells fall in between . This differential performance can be quantified for a given fundamental determinant and emergent morphometric as the ratio of the larger over the smaller of the mean differences between real and virtual trees for the two arbor types . In particular , we formalize the performance ratio as the absolute value of the logarithm of this value ( this definition yields a positive value that is independent of the numerator vs . denominator ) . This value is larger for Branch Order and number of bifurcations and smaller for Radius and asymmetry , i . e . the contrast between apical and basal trees is greatest when testing the Branch Order model for number of bifurcations . Such a measure also allows the comparison of different criteria to divide neuronal groups besides basal and apical , such as other cellular classifications ( e . g . pyramidal and non-pyramidal ) , developmental stage ( young and adult ) , animal species ( rat and others ) , or median-based metrics ( with respect to e . g . size and symmetry ) . The ability of the different models to differentiate between apical and basal trees is much greater than for other divisions tested ( Figure 7C ) . In fact , at least part of the effect observed in other division may simply reflect the apical/basal divide . For example , basal trees tend to be among the smallest and most symmetric , while apical trees tend to be relatively large and asymmetric ( Table 1 ) . The contrast between the basal-apical distinction and all others is particularly prominent considering the logarithmic relation in the performance definition . Attempts to investigate further distinctions by cluster analysis ( not shown ) confirmed these observations . When clustering the 68 groups on the ability of the models to capture the emergent morphometrics , the more distant clusters break along the apical-basal-non pyramidal lines as opposed to other morphometrics ( e . g . tree size , asymmetry ) or metadata ( e . g . animal age or strain ) . After comparing the ability of the “pure” fundamental determinants to control virtual growth and the emergence of various morphometrics in different cell classes , we examined the effect of mixing the influences of Branch Order , Radius , and Path Distance in the hybrid models . The “% Mix” model combines the three fundamental determinants in each of 66 fixed proportions , and samples the basic parameters according to the respective weights . In the “243 Mix” model , every basic parameter can be controlled by a different fundamental determinant . For any tree group and emergent morphometric , the best individual variants of each of these two hybrid models are singled out . Even if all variants were statistically equivalent in their ability to reproduce the morphology of real trees , better quality can be expected because of the sheer number of repetitions ( and the selection of the winner ) . Thus , in order to compare the two hybrids and the best individual models fairly , each of the three approaches was “normalized” to the same number of 243 iterations ( with varying random seeds ) , and the best result was chosen in each case . The general trend across all 68 cell groups is that the 243 Mix clearly outperforms the best individual model , with the % Mix yielding somewhat intermediate results depending on the emergent morphometric ( Figure 8 ) . In particular , the 243 Mix is significantly better at capturing bifurcation asymmetry , surface area , and surface area asymmetry than the individual models ( Figure 8A ) . The percent Mix paradigm constitutes an improvement relative to the best individuals with respect to bifurcation and surface asymmetry , but only for the latter significantly . In all cases , the difference between real and virtual trees was considerably larger for the surface area morphometric than for the number of bifurcations . Visual and qualitative inspection of corresponding virtual and real dendrogram confirmed these findings . In particular , the 243 Mix model demonstrated a striking ability to capture the peculiarities of dendritic branching for each of the examined tree types ( Figure 8B ) . The relative weights of the fundamental determinants in the winning combination of the two hybrid models for each emergent morphometric reflects the trends observed when examining the performance of the pure models . Specifically , we compared the fraction of tree groups “won” by each individual determinant with the proportions of the winning % Mix model and the composition of the 243 Mix . Averaging the results over all tree types reveals similar values of the three determinants from the three protocols within any one morphometric property ( Figure 9 ) . Similarly , the separate examination of basal and apical arbors consistently reproduces the findings of Figures 4 and 5 ( not shown ) . Sampling each basic parameter using a separate fundamental determinant , the 243 Mix model provides an opportunity to gain additional insights into how specific aspects of dendritic structure and development can interact to produce mature morphologies . In particular , it is instructive to analyze how the makeup of the 243 hybrid breaks down for the five basic parameters across the emergent morphometrics throughout all cell types ( Figure 9 , bottom panels ) . For example , Branch Order controls over two thirds of the bifurcation probability in the winning variant selected by the number of bifurcations , but less than one sixth in the model that wins according to bifurcation asymmetry ( Figure 9A and 9B ) . When capturing bifurcation asymmetry , Branch Order contributes above average to taper rate and branch path length , Radius to daughter ratio and parent-daughter ratio , and Path Distance to bifurcation probability ( Figure 9B bottom ) . Interestingly , Surface Area requires a finely balanced contribution of the three determinants in all five basic parameters ( Figure 9C ) , and this emergent morphometric is particularly challenging for the other models ( Figure 8A ) . Even though the Radius model is very rarely the best at capturing surface asymmetry , Radius is the best driver of bifurcation probability in the 243 Mix nearly half of the time ( Figure 9D ) . These findings help to explain the success of the 243 Mix model while giving insights into which fundamental parameter/basic parameter interactions are driving the best individual model choices . For example , the best individual model with regards to the number of bifurcations seems to be highly influenced by bifurcation probability ( Figure 9A ) . In contrast , the large percentage of tree groups which have their surface area asymmetry best captured by Path Distance may be due to the inability of Radius to determine parent-daughter ratio and of Branch Order to determine bifurcation probability with regards to this emergent parameter ( Figure 9D ) .
Dendritic development is a complicated process ( reviewed in [7] ) . Intracellular transport [26] , [27] , extracellularly initiated signaling cascades ( e . g . [10] , [28] ) , synaptic activity [29] , membrane tension [30] , and electrical activity [31] all interact to influence dendritic branching . Morphological modeling constitutes a powerful tool to try and tease out the relative influence of different mechanisms in determining the shapes of different types of dendritic trees . Theories and hypotheses about developmental principles , such as directly relating branch behavior to microtubule density [23] , can be tested quantitatively and rigorously with data driven models ( e . g . , [22] ) . This is an iterative process whereby model failures can point to specific gaps in our understanding , driving new theories , experiments , hypotheses , and computational simulations . Most previous modeling attempts varied widely in both their core methodology ( i . e . the specifics of the algorithm and the choice of variables ) and in the cell classes they attempted to recreate ( see [7] for review ) . This has made direct comparison of results , and the definition of universal modeling “rules , ” particularly difficult . Additionally , when only one model and a single dataset are used , it is impossible to differentiate which results are a function of biology and which are a function of the model details . We have addressed these challenges by applying several closely related models to a large database of different cell classes . Such an approach enabled the abstraction of broad tendencies as to which fundamental determinants best capture different aspects of morphology . In turn , examining the deviations from these general findings in specific cases may point to important developmental differences between tree types . This investigation led to the discovery of striking differences between apical and basal arbors of pyramidal cells . The general results link individual fundamental determinants to the specific emergent morphometric they each best capture , and provide a baseline for comparing particular tree types . The number of bifurcations is best described by Branch Order and worse by Radius . Biologically , the cell may have the ability to “count” branch order locally when determining whether to bifurcate again , possibly detecting the partition of available downstream resources at each bifurcation . The poor performance of Radius suggests that a constant taper rate relating to steady microtubule loss is not a primary mechanism to limit or arrest branching . However , Radius is a better performer than Branch Order with regards to bifurcation asymmetry . Radius may modulate asymmetry by allowing larger branches to bifurcate while their smaller sisters terminate . Interstitial branching , the formation of side branches off of existing branches , constitutes a potential biological underpinning , as it typically produces a larger diameter disparity than terminal branching ( the splitting of an extending growth cone ) . Path Distance can also regulate asymmetry if all branches terminate equidistant from the soma ( symmetric trees ) , or form a distal tuft of bifurcation ( asymmetric trees ) . This may relate to the transport of intracellular messengers or reaction to localized extracellular signals . Since only Path Distance fully succeeds in capturing surface area asymmetry , Radius may be missing vital length or position dependence . Finally , the equal contribution of all fundamental determinants to surface area suggests that this emergent morphometric is not specifically constrained by any individual corresponding biological correlate . A limitation in regards to the interpretation of results is inherent in the restricted amount of data available in each individual group of cells . This scarcity prevents the practical or statistically meaningful investigation of the branching behavior of all neuronal classes separately . Therefore our analysis concentrated on sub-groupings of the 68 unique datasets . The groups were divided based on a wide variety of criteria , including emergent parameter values , laboratory of origin , animal species and age , brain region , and arbor type ( apical , basal , or non-pyramidal ) . In addition to investigating the relative model performance of many of these divisions by hand , the ability of all of the model variants to capture emergent morphometrics was subjected to cluster analysis . The resulting groups were systematically compared to the above divisions as well as visually inspected for other meaningful classification criteria . Of all the various tree groupings consistent with the available collection of real morphologies , the model performance was only statistically differentiated between apical and basal dendrites ( Figure 7C ) . Apical and basal arborizations differed in the pattern ( Figures 4 and 5 ) , and the direction of their responses ( Figure 7A and 7B ) . Several potential biological explanations merit further investigation . One important aspect to note is that pyramidal cells , as opposed to many of the other modeled tree types , grow in a very layer specific manner ( as seen graphically in Figure 8B ) . Both the real and virtual CA1 apical trees show a distal increase in bifurcations , corresponding to the tuft in stratum lacunosum-moleculare . In contrast , basal trees have the majority of their terminations in a relatively small window relative to the soma ( see also Figure 5D and 5E ) . The fact that these trees are exposed to different inputs and extracellular chemicals gradients as they cross ( or do not , in the case of basal dendrites ) histological layers could largely explain their contrasting branching behavior . There is some indirect experimental evidence which supports this hypothesis . Baker et al . [32] have shown differential responses of pyramidal and non-pyramidal cortical cells to neurotrophin-3 . Other studies have shown that basal and apical dendrites respond differently to neurotrophins ( NTs ) , with basal response being layer specific , while apical responses are more general , perhaps due to their crossing several cellular layers [33] , [34] . These previous studies , however , have applied NTs in a bath fashion and have not looked directly at the morphology of apical trees in different layers . In order to test apical layer specific responses directly , it would be necessary to vary the NTs in a layer specific manner , perhaps through genetic manipulation of different incoming pathways , and perform layer specific analysis of apical tree morphology . The morphological response of dendrites to NTs and other chemicals is very complex ( reviewed in [7] ) , making the generation of specific hypotheses difficult . NTs and their receptor patterns can vary with developmental time [35] , [36] and activity [37] . Other studies [9] have shown uniform sub-cellular distributions for some receptors , but rapid mobility of these receptors [38] . This with problems maintaining morphological details in certain culture preparations [39] leaves open the possibility of layer specificity , at least for some cell types or developmental periods . It is also possible that while NTs are obviously important to neuronal morphology , layer specific responses to them may be mediated through other pathways . However , some intriguing results from bath application of NTs provide possible testable hypotheses . For example , supposes it is the layer specific responses to NTs that is limiting basal dendrites to particular cortical layers . Then our results would suggest that by increasing expression of BDNF in layer 5 , basal dendrites from cells in layers 4 , which respond very strongly to BDNF [33] , [34] , may grow into that deeper layer . Also , layer 6 basal dendrites are inhibited by NGF and BDNF while layer 4 and 5 dendrites have the opposite response . Likewise apical trees in layer 6 have the weakest response to these two NTs . If pyramidal dendritic NT response is layer as well as cell type-specific , as our data suggests , the expression pattern of these NTs may be similar , and different in layer 6 than in 4 and 5 . The strongest responses to NT-4 are seen in basal dendrites from neurons in layers 5 and 6 , and apical dendrites from layer 4 [34] . As these three structures have no overlap in the layers they innervate , it is possible that NT-4 may provide a general growth control in these structures without disrupting layer specific responses . An alternative or additional mechanism that could underlie the differential performance of various models in the simulation of apical and basal trees involves shifting competition for an intracellular signal or cytoskeletal metabolite . Previous statistical analyses have provided convincing indication that dendritic branching may be homeostatically regulated by global and local competition for limited intracellular resources [40] . Such an explanation could account for the sudden termination often observed in basal arbors , and the burst of bifurcations in apical tufts . More time-lapse studies of growing pyramidal cells could help clarify this possibility . As flexibility is added to the models by allowing the different fundamental parameters to contribute to a single virtual tree through model mixing one would expect an improvement in the virtual emergent morphometrics . Both bifurcation asymmetry and surface area were significantly better reproduced by the 243 Mix paradigm than by either the % Mix or individual models ( Figure 8A ) . However , neither mix paradigm was better than the best individual model in capturing number of bifurcations ( Figure 8A ) , suggesting that the total branch count may be under relatively simple biological control relative to the other emergent morphometrics . There are several dimensions in which this work could be expanded . While we are trying to gain developmental insights , digital reconstructions of real cells in publicly available databases are currently limited to adult ( or at least relatively mature ) neurons [41] . Based on early proposals based on electron microscopy [23] , several studies , including the present one , have attempted to correlate branching behavior with local diameter ( e . g . [13] , [22] ) . However , the thickness of dendrites changes during development , and the “final” diameter measures ( as reported in the digital reconstructions of real neurons ) only indirectly reflect the values at the actual time of growth . With developmental time series of reconstruction data , we could model the development of dendrites more directly . This study raised the possibility that apical and basal dendrites differ from each other due to the histological environment through which they extend , while the morphologies of non-pyramidal cells might be more intrinsically driven . By expanding the suite of fundamental determinants to include planar and radial distance from the soma , this hypothesis could be tested more directly . Such an extension would require 3D embedding of the virtual cells ( see e . g . , [21] ) . Additionally , while we have concentrated here on “normal” cells , this comparative method could also be used to detect differences between experimental preparations or disease states , possibly hinting at the underlying developmental processes . As they occur in different parts of the same cells , the striking contrast between apical and basal trees may be costly to control and achieve , and is likely to be relevant from the information processing standpoint . This puts renewed emphasis on the question of what this divide could facilitate in the brain . Due largely to methodological considerations , the relatively thin basal branches are seldom investigated in electrophysiological experiments . Even modeling studies tend to concentrate on different divisions of the apical tree ( e . g . [6] , [12] ) . This study emphasizes the unique aspects of pyramidal cell morphology and provides motivation for a closer look at the functional consequences of its distinct arborizations .
In this study , morphometric parameters that control dendritic branching are measured from groups of real cells and resampled stochastically to create virtual trees of the corresponding class . The real neurons consist of 736 digital reconstructions from 16 different labs . The apical and basal trees of pyramidal cells are treated separately , summing up to a total of 68 individual groups ( Table 1 ) . These 3D reconstructions were downloaded from the NeuroMorpho . Org inventory [42] in their “standardized form . ” In particular , all cells are checked for format uniformity and data integrity through a combination of automated , semi-automated , and manual methods , addressing common reconstruction issues . Every morphological file ( in “SWC” format ) contains one numbered line for each tracing point in the neuronal structure , described by the three coordinates of its spatial position , local dendritic radius , and the number of the line representing the parent point towards the soma [43] . Virtual trees in the form of dendrograms are generated with a simple recursive algorithm ( Figure 2 ) . Starting from an initial diameter , a branch grows for a certain path length and tapers its thickness . Then it either stops or bifurcates into two daughters whose initial diameters are determined based on the parent's . Each daughter iterates independently through the same process , until all branches are terminated ( Figure 2A ) . Thus there are five “basic” parameters controlling growth in addition to the initial diameter: branch path length , taper rate , bifurcation probability , parent-daughter ratio ( between the parent diameter and the larger daughter diameter ) , and daughter ratio ( between the larger and smaller daughter diameters ) . Each of these basic parameters is sampled stochastically from statistical distributions derived from the values measured in the real trees ( Figure 2B ) . Except for the “unique” case of the initial diameter , the basic parameters extracted from different portions of real trees vary considerably [22] . To obtain distributions faithful to the observed data , basic parameters are thus sampled according to the local value of a “fundamental” determinant . Three variants of this model are based on distinct fundamental determinants ( Figure 2A ) , namely branch radius , path distance from the soma , and branch order ( i . e . , the number of bifurcations towards the soma ) . Thus , basic parameters measured from a homogeneous group of real dendritic trees are binned by the corresponding local value of the fundamental determinant . Branch pathlength , daughter ratio , and taper rate are based on the fundamental determinants value at the beginning of a branch , while bifurcation probability and parent-daughter ratio are based on the values at the end of branches . Aside from the bifurcation probability ( a scalar fraction ) , each bin is then fitted by least square error to the best of three 2-parameter functions: gamma , Gaussian , and uniform . In a previous study [22] , a variety of functional distribution and fitting methodologies were tested , including reproducing all discrete values in a large lookup table for each basic parameter . As long as the basic parameter varied with the fundamental determinant , the model proved to be very robust to binning and distribution fitting particulars . Thus , the selection of parametric functions in the present work optimally combined accuracy and simplicity . For the parameters controlling diameter change , the proportion of measures assuming a unitary value ( i . e . reflecting a lack of diameter change ) , referred to as “Unity Fraction” in previous work [22] are sampled separately according to their occurrence in each bin . Two types of hybrid models were also tested by “mixing” the fundamental determinants . In the “% Mix” model , each fundamental determinant contributes a percentage of influence over the sampling of the basic parameters . These percentages are varied for each fundamental determinant from 0% to 100% at 10% increments . For example , Branch Order may contribute 10% , Path Length 70% , and Radius the remaining 20% . This sums up to 66 distinct variants of the % Mix model including the “pure” ( unmixed ) models . For the basic parameters controlling diameter , the probability of sampling a value of one is first computed as the weighted average of the three individual probabilities . For all basic parameters not determined to be one , values are sampled from all three fundamental determinant distributions and averaged together based on their relative weights . In the second mixing method , each basic parameter depends on a different fundamental determinant . For example , taper rate could be based on Radius , parent-daughter ratio on Path Distance , and bifurcation probability , branch path length , and daughter ratio all on Branch Order . With five basic parameters and three fundamental determinants , this creates an additional 35 ( minus the three “pure” cases ) variants of this model ( hence the name “243 Mix” ) . When comparing the individual and % Mix results to the more numerous 243 mix results , both the individual models and the % Mix models were run a total of 243 times with different random seeds . Any morphometrics not directly used in the algorithm are “emergent” to the model . We chose four emergent morphometrics to compare virtual and real cells , selected for their biological and electrophysiological significance ( Figure 2D ) . The total number of bifurcations provides a measure of branching complexity . Since all considered trees are binary , this count equals the number of terminations plus one . Bifurcation asymmetry characterizes how evenly those terminations are distributed throughout the tree . It is the average over all bifurcation of ( n1−n2 ) / ( n1+n2−2 ) , where n1 and n2 are the number of terminal tips of the larger and smaller daughter subtrees , respectively . The total surface area is a size metric , while surface area asymmetry is defined by the same expression as above , but with n1 and n2 representing the surface areas of the daughter subtrees . Mean emergent morphometric values for each group of real trees are reported in the last four columns of Table 1 . A custom java program ( LNded2 . 0 ) , running on a Pentium M under Windows XP , extracts the basic parameters from the real cells , fits them according to the appropriate fundamental determinants , and samples the resulting statistical distributions to create virtual dendrograms . The program then outputs the emergent morphometrics from real and virtual trees to Microsoft Excel for comparison and analysis . The code and necessary documentation for all model variants is available for public download under the ModelDB section [44] , [45] of the Senselab database ( http://senselab . yale . med . edu ) . For every model and each cell group , ten virtual trees were created for each real tree . The virtual trees were then divided into ten groups , each having the number of trees matching the real groups . The mean and standard deviation for the emergent morphometrics was computed for each group and the mean of those means and standard deviations were compared to the corresponding ( single ) values for the group of real cells . Both for the three individual models , and for the two mixing paradigms , a “best” model was chosen for each tree group as that with the smallest “distance” between real and virtual trees . The distance metric was defined for each emergent morphometric as to account for both the gap between the real and virtual mean measures , and the stochastic variability of the simulation repeats . In particular , this metric was computed as the absolute difference between the mean of means of the ten groups of virtual cells and the mean of the single group of real cells , or as the standard error of the mean of the ten groups of virtual cells , whichever was greater . Error bars in all figures represent standard error unless otherwise noted . An asterisk directly above a column signifies a significant difference ( P< . 05 ) from the other two columns while an asterisk between two columns signifies a significant difference only between those two columns as determined by the Mann-Whitney U non-parametric comparison using http://udel . edu/̃ mcdonald/statkruskalwallis . xls by Dr . John H . McDonald . All statistics were computed using Microsoft Excel .
|
Neurons in the brain have a variety of complex arbor shapes that help determine both their interconnectivity and functional roles . Molecular biology is beginning to uncover important details on the development of these tree-like structures , but how and why vastly different shapes arise is still largely unknown . We developed a novel set of computer models of branching in which measurements of real nerve cell structures digitally traced from microscopic imaging are resampled to create virtual trees . The different rules that the models use to create the most similar virtual trees to the real data support specific hypotheses regarding development . Surprisingly , the arborizations that differed most in the optimal rules were found on opposite sides of the same type of neuron , namely apical and basal trees of pyramidal cells . The details of the rules suggest that pyramidal cell trees may respond in unique and complex ways to their external environment . By better understanding how these trees are formed in the brain , we can learn more about their normal function and why they are often malformed in neurological diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience/neurodevelopment",
"neuroscience/theoretical",
"neuroscience"
] |
2008
|
A Comparative Computer Simulation of Dendritic Morphology
|
While the apicomplexan parasites Plasmodium falciparum and Toxoplasma gondii are thought to primarily depend on glycolysis for ATP synthesis , recent studies have shown that they can fully catabolize glucose in a canonical TCA cycle . However , these parasites lack a mitochondrial isoform of pyruvate dehydrogenase and the identity of the enzyme that catalyses the conversion of pyruvate to acetyl-CoA remains enigmatic . Here we demonstrate that the mitochondrial branched chain ketoacid dehydrogenase ( BCKDH ) complex is the missing link , functionally replacing mitochondrial PDH in both T . gondii and P . berghei . Deletion of the E1a subunit of T . gondii and P . berghei BCKDH significantly impacted on intracellular growth and virulence of both parasites . Interestingly , disruption of the P . berghei E1a restricted parasite development to reticulocytes only and completely prevented maturation of oocysts during mosquito transmission . Overall this study highlights the importance of the molecular adaptation of BCKDH in this important class of pathogens .
The phylum of Apicomplexa comprises a large number of obligate intracellular parasites that infect organisms across the whole animal kingdom . Two important members of this phylum , Plasmodium spp . and Toxoplasma gondii , are the etiological agents of malaria and toxoplasmosis , respectively . Malaria remains one of the most significant global public health challenges ( World Malaria Report 2012 , www . who . int ) , while toxoplasmosis causes severe disease and death in immunocompromised individuals and can lead to complications in development of the foetus if contracted during pregnancy [1] . Both Plasmodium spp and T . gondii invade a range of mammalian cells and replicate within a membrane-enclosed compartment called the parasitophorous vacuole ( PV ) . Residence within the PV provides protection from host cell defence mechanisms , while allowing the rapidly developing parasite stages to access small molecules that can diffuse freely across the PV membrane ( PVM ) [2] , [3] . Both T . gondii replicative forms and Plasmodium blood stages were thought to rely primarily on glucose uptake and glycolysis for generation of ATP and other intermediates required for energy generation and replication [4]–[8] , and to lack a canonical , pyruvate-fuelled TCA cycle . In particular , Plasmodium-infected erythrocytes exhibit an extraordinarily high rate of glucose uptake [9] and selective inhibitors of the Plasmodium hexose transporter are cytotoxic [10] , [11] . Moreover , genomic and biochemical studies have shown that apicomplexan parasites target their single canonical pyruvate dehydrogenase complex ( PDH ) to the apicoplast , a non-photosynthetic plastid organelle involved in fatty acid biosynthesis , rather than to the mitochondrion [12]–[14] . The absence of a mitochondrial PDH complex in these parasites suggested that glycolytic pyruvate was not converted to acetyl-CoA in the mitochondrion and further catabolised through the TCA cycle [13]–[15] . In other organisms , lipids and branched chain amino acids ( BCAA ) can be catabolised in the mitochondrion to generate acetyl-CoA via pathways not dependent on PDH ( Fig . 1A ) . However , Plasmodium spp . lack the enzymes needed for the β-oxidation of fatty acids and BCAA degradation . While T . gondii retained the enzymatic machinery necessary for β-oxidation , these parasites appear to lack a typical mitochondrial acyl-carnitine/carnitine carrier [16] , [17] . Moreover , the genes coding for β-oxidation enzymes are apparently not expressed in tachyzoites , although they may be active in oocysts [18] . The possibility that T . gondii tachyzoites rely on host BCAA to generate mitochondrial acetyl-CoA was recently investigated , but disruption of the gene encoding the first enzyme involved in BCAA degradation , branched chain amino acid transferase ( BCAT ) in tachyzoites presented no phenotypic defect [19] . Together , these studies suggested that there was minimal synthesis and catabolism of acetyl-CoA in the mitochondrion . Several studies have recently led to a reappraisal of this model of carbon metabolism in Apicomplexa . Firstly , detailed 13C-glucose and 13C-glutamine tracer experiments on T . gondii tachyzoite stages showed that carbon skeletons derived from both carbon sources were actively catabolised in a canonical TCA cycle , with the majority of pyruvate entering via acetyl-CoA . Chemical disruption of the aconitase enzyme activity , catalysing an early step in the TCA cycle , completely ablated parasite growth and infectivity in mammalian cells , indicating that the conversion of citrate to isocitrate is important for parasite growth and pathogenesis and that dysregulation of glucose catabolism in the mitochondrion is likely to be lethal [20] . Second , similar studies undertaken in P . falciparum indicate that glucose is further catabolised in the TCA cycle in asexual blood stages [21]–[23] , and at a dramatically increased rate in sexual gametocyte stages [21] . A functional , canonical TCA cycle capable of generating reducing equivalents is likely necessary for maintenance of mitochondrial protein transport and the re-oxidation of inner membrane dehydrogenases required for pyrimidine biosynthesis [24]–[28] . The importance of an active respiratory chain in P . falciparum blood stages is highlighted by the sensitivity of this stage to the antimalarial drug atovaquone , which targets respiratory chain complexes [29] , and by the increased expression of TCA cycle enzymes in parasites isolated from patients in endemic areas [30] , [31] . Atovaquone is also known to kill the rapidly dividing tachyzoite and cyst-forming bradyzoite stages of T . gondii [32] . These studies suggest that the translocation of the conventional mitochondrial PDH to the apicoplast was associated with a new enzyme activity that functionally replaced PDH in regulating TCA cycle metabolism , although the identity of this enzyme remains unknown . The possibility that other mitochondrial dehydrogenases may individually or collectively fill this missing link was raised by the finding that the P . falciparum α-ketoglutarate dehydrogenase ( α-KDH ) can catalyze the conversion of pyruvate to acetyl-CoA in vitro , although slightly less efficiently than PDH [33] . On the other hand , Cobbold et al . , proposed that the P . falciparum branched chain ketoacid dehydrogenase ( BCKDH ) , the only enzyme implicated in BCAA degradation retained in the Plasmodium spp . , may substitute for PDH based on the finding that catabolism of glucose in the TCA cycle in a P . falciparum PDH mutant was inhibited by oxythiamine , an inhibitor of thiamine pyrophosphate ( TPP ) -dependent dehydrogenases [22] . However , oxythiamine also inhibits α-KDH ( and all other TPP-dependent enzymes ) , and the enzymatic activity of P . falciparum BCKDH was not tested . The identity of the enzyme ( s ) that link glycolysis with mitochondrial metabolism , and their functional significance in the normal growth and virulence of these parasites therefore remains an open question . The genomes of the apicomplexan parasites that contain a functional mitochondrion encode all of the subunits of the BCKDH complex , which include the branched chain α-keto acid dehydrogenase E1 subunits ( EC 1 . 2 . 4 . 4 ) , the dihydrolipoyl transacylase E2 subunit ( EC 2 . 3 . 1 . 168 ) , and the lipoamide dehydrogenase E3 subunit ( EC 1 . 8 . 1 . 4 ) – ( Table S1 ) . The eukaryotic BCKDH and PDH complexes share many structural and enzymatic properties , catalysing analogous reactions in central carbon metabolism where the initial α-ketoacid is decarboxylated by the E1 subunit - a thiamine diphosphate ( TPP ) -dependent heterotetramer consisting of two α subunits ( E1a ) and two β subunits ( E1b ) [34] . Given the functional similarity between these complexes , we have previously postulated that the BCKDH complex could have assumed the function of the mitochondrial PDH in the Apicomplexa [16] . In this study , we provide unequivocal evidence that BCKDH primarily fulfils the function of mitochondrial PDH in both T . gondii and Plasmodium berghei , a rodent model for malaria . P . berghei allows phenotypic evaluation under physiological conditions in vivo and offers the potential to interrogate the whole parasite life cycle from mosquito to mouse . We find that genetic disruption of the BCKDH-E1a subunit in these parasites leads to a block in the conversion of pyruvate to acetyl-CoA , and global changes in metabolic fluxes , as shown by metabolite profiling and comprehensive 13C-stable isotope labelling approaches . More importantly , the functional disruption of the BCKDH multi-enzyme complex was associated with a growth defect and reduced virulence of T . gondii in mice , while in P . berghei it resulted in strong alteration of intraerythrocytic development and severely diminished virulence in mice . In addition , the absence of BCKDH affects all the vector stages and blocks oocyst development in the mosquito , indicating that this pathway is essential for transmission of the disease .
Point mutations in human BCKDH-E1a are associated with complete loss of catalytic activity [34] , indicating that genetic depletion of this subunit should be sufficient to abrogate BCKDH function . Deletion of the gene coding for the E1a subunit of TgBCKDH was achieved by double homologous recombination ( Tge1a_ko ) in the RHku80_ko ( hereafter termed ‘RH’ ) background strain , which favours homologous recombination over random integration ( Fig . S1A ) [35] , [36] . Transgenic parasites were cloned and loss of TgBCKDHE1a was demonstrated by genomic PCR ( Fig . S1B ) , while absence of the protein was confirmed by Western blot using cross-reacting anti-P . falciparum E1a antibodies ( Fig . 1B ) . The E1a subunit was detected as a ∼45 kDa and a ∼90 kDa band in Western blots of wild type parasites . The 90 kDa band likely corresponds to the E1a/E1b heterodimer , as the intensity of this band was severely diminished under strong denaturating conditions ( Fig . 1B ) . Neither band was detected in the knockout . The Tge1a_ko formed smaller plaques in a human foreskin fibroblast ( HFF ) lytic plaque assay compared to the parental RH strain ( Fig . 1C ) indicating a reduced ability to infect and/or grow in host cells . Further phenotypic analyses revealed that neither Tge1a_ko tachyzoite invasion nor egress from infected host cells were affected ( data not shown ) and that the reduction in fitness was due to significantly reduced intracellular growth compared to RH parasites , as monitored by the reduced number of parasites per vacuole established by Tge1a_ko after 24 h ( Fig . 1D ) . This phenotype was exacerbated when infected HFF were cultivated in the absence of glucose in a reversible fashion , while removal of glutamine did not aggravate this defect ( Fig . 1D ) . To validate that the phenotypes observed in Tge1a_ko are only due to the loss of the E1a subunit of BCKDH , we targeted a second copy of the E1a subunit where the N-terminal mitochondrion targeting signal was replaced by the mitochondrial transit signal of TgSOD3 ( SOD3mycBCKDH-E1a , Fig . 1E ) in Tge1a_ko parasites . In addition , we attempted to complement Tge1a_ko parasites by targeting the product of a second copy of the TgPDH-E1a subunit to the mitochondrion via replacement of its bipartite targeting signal with the mitochondrial transit signal of TgSOD3 ( SOD3mycPDH-E1a , Fig . 1E ) . TgPDH-E1a and TgBCKDH-E1a show significant similarity by sequence alignment ( ∼25% ) and the catalytic residues are clearly conserved between the two subunits ( Fig . S3A ) . The mitochondrial SOD3mycPDH-E1a was unable to rescue the intracellular growth defect of Tge1a_ko parasites while complementation with SOD3mycBCKDH-E1a restored the growth of Tge1a_ko ( Fig . 1F ) . This highlights a lack of permissiveness to interchange subunits between the different α-ketoacid dehydrogenase complexes but moreover confirmed that the phenotypes observed with Tge1a_ko are solely due to the absence of BCKDH activity . To further examine whether BCKDH is required for virulence , mice were injected intraperitoneally with ∼15 RH or Tge1a_ko tachyzoites . The inoculation of virulent RH parasites resulted in acute toxoplasmosis in all mice after 8 days leading to their culling . In contrast , 3 out of 5 mice infected with Tge1a_ko parasites remained alive after 21 days ( Fig . 1G ) . The surviving mice had seroconverted and were resistant to subsequent challenge with ∼1 , 000 RH parasites ( Fig . 1G ) . Taken together , these results establish that the BCKDH complex is implicated in glucose catabolism and is important for both parasite fitness in vitro and virulence in vivo . To investigate the underlying basis of the intracellular growth defect in the BCKDH mutant , T . gondii RH and Tge1a_ko tachyzoites were cultivated in HFF and metabolite levels in egressed tachyzoites determined by both GC-MS and LC-MS ( Fig . 2A ) . Significant differences were observed in the levels of several glycolytic and early TCA cycle intermediates in Tge1a_ko tachyzoites , compared to the RH control strain . This included a 4- to 10-fold decrease in 2-hydroxyethyl-TPP ( the intermediate in synthesis of acetyl-CoA from pyruvate ) , acetyl-CoA , and citrate , and a 2- to 4-fold increase in 3-phosphoglycerate ( 3-PGA ) , pyruvate and lactate ( Fig . 2A ) . These data are consistent with a defect in the conversion of pyruvate to acetyl-CoA and citrate as well as an increased flux to lactate production . To confirm that Tge1a_ko parasites have a defect in acetyl-CoA synthesis , freshly egressed RH and Tge1a_ko tachyzoites were metabolically labelled with 13C-U-glucose . The intermediates in glycolysis , the pentose phosphate pathway ( PPP ) and the TCA cycle were strongly labelled in RH parasites ( Fig . 2B ) . Citrate isotopomers were generated containing +2 , +3 and +4 13C carbons , indicative of entry of both 13C2-acetyl-CoA and 13C3-oxaloacetate derived from pyruvate into the TCA cycle of RH parasites ( Fig . 2D ) . In contrast , the labelling of acetyl-CoA and TCA cycle intermediates , including citrate and the C4 dicarboxylic acids , were dramatically reduced in Tge1a_ko ( Fig . S2A , 2B and 2D ) , suggesting that loss of BCKDH is associated with a block in entry of glucose-derived pyruvate into the TCA cycle . This was supported by complementary labelling with 13C-U-glutamine , which revealed equivalent or elevated enrichment of label in all TCA cycle intermediates in Tge1a_ko compared to RH parasites ( Fig . 2C ) . The predominant citrate isotopologue generated in 13C-glutamine-fed Tge1a_ko had +4 13C atoms ( Fig . 2D ) indicating that glutamine-derived 13C4-oxaloacetate combines with a residual source of unlabelled acetyl-CoA to allow citrate synthesis . This could reflect low level capacity of the α-ketoglutarate dehydrogenase to convert pyruvate to acetyl-CoA , or more likely , the conversion of mitochondrial-produced 13C4-oxaloacetate to citrate in the cytosol via the ATP-citrate lyase or the second putative citrate lyase ( TGME49_203110 , www . toxodb . org ) present in the genome of T . gondii . Interestingly , significant labelling of glycolytic intermediates and hexose-phosphate was detected in 13C-glutamine-fed Tge1a_ko tachyzoites , which was absent in wild type parasites ( RH ) ( Fig . 2C ) . Collectively , these findings show that the BCKDH is required for the conversion of pyruvate to mitochondrial acetyl-CoA and operation of a cyclical TCA cycle . In the absence of BCKDH , the continued production of C4 dicarboxylic acids derived from glutamine by the oxidative TCA cycle leads to increased gluconeogenic fluxes despite the fact that these parasites continue to utilize glucose and have a high glycolytic flux . To investigate whether BCKDH is also required for catabolism of branched chain amino acids ( BCAA ) , RH and Tge1a_ko were labelled with 13C-U-leucine , 13C-U-isoleucine and 13C-U-valine . An untargeted metabolome-wide isotope analysis detected no significant 13C-enrichment in TCA cycle intermediates , despite detecting efficient uptake of branched chain amino acids and conversion to the respective 13C-labeled branched chain keto acids ( data not shown ) . No significant differences in the steady state levels of leucine , isoleucine or valine were detected between the parental and knock out strains ( Fig . 2A ) . Together these data strongly suggest that mitochondrial acetyl-CoA is not derived from BCAAs under normal growth conditions ( Fig . S2B ) . To determine whether the role of BCKDH in acetyl-CoA production could be by-passed by addition of exogenous acetate , fibroblasts infected with Tge1a_ko were cultivated in media with or without acetate . Supplementation of the medium with acetate led to a partial but significant rescue of the severe growth defect observed in the absence of glucose ( Fig . 2E ) . This result is consistent with BCKDH having a role in acetyl-CoA production and suggests some redundancy in the functions of BCKDH and acetyl-CoA synthetase in generating mitochondrial and/or cytoplasmic pools of acetyl-CoA . In vitro enzyme assays were performed to investigate the substrate selectivity of T . gondii BCKDH . As attempts to express an active , recombinant BCKDH complex were unsuccessful , enzyme assays were performed on whole cell lysates from RH and Tge1a_ko parasites . PDH activity was detected when cell lysates were incubated with 0 . 5 mM pyruvate in the presence of cofactors , with over two-fold higher concentration of acetyl-CoA production observed in RH ( 276 µM ) than Tge1a_ko ( 124 µM ) extracts , confirming a role of BCKDH in acetyl-CoA production ( Fig . 3B ) ( p<0 . 05 ) . The significant level of background PDH-like activity observed in Tge1a_ko extracts is likely mediated by the apicoplast PDH or mitochondrial α-KDH complexes . Minimal production of the branched chain acyl-CoAs , 3-methylpropanoyl-CoA ( 128 nM ) and 3-methylbutanoyl-CoA ( below limit of quantitation; LOQ = 5 nM ) , was detected following incubations with their respective substrates , 4-methyl-2-oxopentanoate and 3-methyl-2-oxobutanoate . Branched chain acyl-CoA formation was significantly lower in Tge1a_ko compared to RH extracts ( Fig . 3C–D ) , suggesting BCKDH does indeed possess classical BCKDH-like activity . However , accurate quantification of acyl-CoA products from assays with higher substrate concentrations ( 2 mM ) confirmed that the BCKDH-like activity was 1000- to 10 , 000-fold lower than the PDH-like activity ( Fig . 3A ) , suggesting that this enzyme functions primarily as a PDH in vivo . Interestingly , the hydroxyalkyl-TPP intermediates for all three substrates were detected in a BCKDH-dependent manner ( Fig . 3E–G ) . To determine whether BCKDH has a similar role in malaria parasites , a P . berghei mutant lacking the BCKDH E1a subunit was generated by double homologous recombination ( Pbe1a_ko ) ( Fig . S4A ) . Several independent positive transgenic pools were obtained after drug cycling . However , their slow growth hampered the cloning of the mutants by limiting dilution in wild type immunocompetent CD1 mice . Since the parasites seemed to exhibit a severe fitness defect that might lead to clearance of the infection by the mouse immune system , we switched to immunodeficient RAG-1 -/- mice for cloning purposes [37] . Loss of expression of Pbe1a in the clonal Pbe1a_ko line from these mice , was demonstrated by genomic PCR and Western blot analysis ( Fig . S4D and 4A ) . Strikingly , mice infected with 15×106 Pbe1a_ko parasites had constant low parasitaemias ( 5–15% over 10 days ) , while infection with WT parasites led to an exponential rise in parasitaemia ( up to 45% after 4 days ) , leading to culling due to illness ( Fig . 4B ) . Despite having much lower parasitaemias , mice infected with Pbe1a_ko parasites developed symptoms of severe anaemia and were culled 10–12 days post-infection ( Fig . 4B ) . The haematocrit level between the mice infected with WT or Pbe1a_ko was comparable during the first five days of infection ( Fig . 4C ) . Haematocrit levels continuously decreased over the subsequent 7 days in the Pbe1a_ko-infected mice ( Fig . 4C ) , consistent with the observed anaemia in these animals . To understand why the haematocrit level decreased despite relatively low parasitaemia , parasite distribution was further investigated in the different red blood cell types . In wild type parasite infected mice , parasites could be found both within reticulocytes as well as in normocytes . In contrast , the majority of Pbe1a_ko parasites were present within reticulocytes throughout the course of infection . To assess the importance of this apparent cell tropism , we induced reticulocytosis in mice with phenylhydrazine prior to infection . In mice infected with wild type , parasitaemia increased as expected and phenylhydrazine pre-treatment slightly accentuated the growth of the parasites . Pre-treatment of mice with phenylhydrazine rescued significantly the growth defect observed with Pbe1a_ko ( although not to wild type levels as reticulocytes maturate into normocytes over the course of infection ) while in mice not pre-treated the parasitaemia levels remained low throughout the 5 days of infection supporting the observation that Pbe1a_ko seemed to preferentially infect reticulocytes ( Fig . 4D ) . This differential distribution is not explained by an invasion defect of the Pbe1a_ko parasites for normocytes as shown in ( Fig . 4E ) . Indeed , we observed no difference in invasion efficiency between WT and Pbe1a_ko parasites when using purified normocytes and reticulocytes as target host cells . In vitro maturation was examined following in vitro invasion of purified normocytes or reticulocytes . This selective distribution appears to be due to rapid loss of viability of Pbe1a_ko in the normocytes . Specifically , WT parasites developed to the schizont stage in both reticulocytes and normocytes ( Fig . 4F and 4G , respectively ) , whereas Pbe1a_ko parasites developed normally in reticulocytes ( Fig . 4F ) , but rapidly degenerated within normocytes ( Fig . 4G ) . Taken together , these findings demonstrate that a functional BCKDH complex is required for the development of P . berghei in mature erythrocytes . The abortive infections of normocytes are most likely eliminated from the circulation by the spleen and liver , resulting in the lower parasitaemia and protracted course of infection of the mutant . To confirm that the observed attenuation phenotype is solely attributable to the deletion of the PbBCKDH-E1a gene , Pbe1a_ko parasites were complemented with a copy of the P . falciparum BCKDH-E1a gene , PfE1a ( Fig . S4C ) . Pbe1a_ko+PfE1a parasites were obtained after several passages in mice with intermittent drug selection . Integration of the complementation plasmid and PfE1a protein expression were confirmed by genomic PCR ( Fig . S4D ) and Western blot analyses ( Fig . S4E ) . Complementation with PfE1a restored the ability of these parasites to complete their development in mature erythrocytes ( Fig . 4F and 4G ) and their growth rate in vivo was comparable to the parental WT strain ( Fig . S4F ) . These results also indicate that the E1a subunit from P . falciparum can assemble with the P . berghei subunits to form a functional BCKDH multi-enzyme complex . To determine whether the BCKDH complex fulfils the function of a mitochondrial PDH in the P . berghei asexual blood stages , ring/early trophozoite stages of WT and Pbe1a_ko parasite-infected RBCs ( iRBCs ) were matured to schizonts in vitro and labelled with 13C-U-glucose or 13C-U-glutamine over the final 5 h of maturation . Schizont-iRBCs were purified and the labelling of intracellular metabolite pools determined by GC-MS . Intermediates in glycolysis , the PPP and TCA cycle were highly enriched when RBCs infected with WT parasites were labelled with 13C-glucose ( Fig . 5A and S5A ) while in uninfected RBCs , 13C-labelling from glucose was only detected in the glycolysis and PPP metabolites ( Fig . S5A ) . After 5 h labelling , the major isotopologues of citrate in WT iRBCs contained two 13C-carbons , reflecting the incorporation of 13C2-acetyl-CoA into citrate after one round of the TCA cycle ( Fig . 5B ) . While a similar labelling of glycolytic and PPP intermediates was observed in Pbe1a_ko-iRBC , incorporation into TCA intermediates and associated amino acids was greatly diminished in the Pbe1a_ko compared to WT iRBCs ( Fig . 5A , 5B and S5A ) . It is notable that +3 isotopologues of malate and aspartate ( a proxy of oxaloacetate ) were still detected in 13C-glucose-fed parasites , reflecting continued conversion of pyruvate to malate ( via oxaloacetate ) by the action of PEPC ( Fig . 5B ) , as recently observed [23] . These findings demonstrate that BCKDH is required for the mitochondrial catabolism of glucose by conversion of pyruvate to acetyl-CoA in P . berghei blood stage parasites . Importantly , both wild type and Pbe1a_ko-infected RBC catabolized 13C-U-glutamine in a canonical TCA cycle ( Fig . 5C and 5D ) , as has recently been shown to occur in P . falciparum [21]–[23] . However , in contrast to the situation in T . gondii , no evidence for increased gluconeogenesis in the Pbe1a_ko-infected RBC was observed , consistent with the absence of key gluconeogenic enzymes in these parasites [10] . As anticipated , label derived from 13C-U-leucine was not incorporated into TCA cycle intermediates in either WT , Pbe1a_ko-infected RBC , or uninfected RBCs ( Fig . S5B ) , indicating that PbBCKDH does not catabolise α-ketoacids generated from BCAA and consistent with the absence of BCAA-specific aminotransferase ( BCAT ) in the malaria parasite genomes ( Table S1 ) . To assess whether differences in the acetyl-CoA levels of Pbe1a_ko , via conversion of acetate to acetyl-CoA from the acetyl-CoA synthetase , could be responsible for the reticulocyte tropism that is observed , we followed in vitro maturation of wild type and Pbe1a_ko parasites within normocytes in presence or not of exogenous acetate . Pbe1a_ko parasite growth in normocytes was partially restored by supplementation of the medium with acetate ( Fig . 6A–B and S6A ) , indicating that the presence of acetate in reticulocytes could be one of the factors contributing to the ability of Pbe1a_ko parasites to survive and develop within this cell type but not within normocytes . We next assessed the effect of deleting the BCKDH-E1a gene on sexual development and mosquito transmission of P . berghei . To minimise a potential indirect effect of the attenuated growth of the Pbe1a_ko mutant on gametocyte numbers , mice were injected with phenylhydrazine two days before infection , inducing mild anaemia and increased reticulocytosis . As a result , a similar parasitaemia was obtained for all strains on day 3 post-infection ( Fig . 7A ) . The numbers of morphologically mature female ( macro- ) and male ( micro- ) gametocytes were lower in Pbe1a_ko parasites than in either wild type or PfE1a complemented clones ( Fig . 7A ) . The small number of microgametocytes in the Pbe1a_ko clone had a considerably reduced capacity to differentiate ( exflagellate ) into microgametes when stimulated by xanthurenic acid in vitro . Possibly as a result , the ability of macrogametocytes to convert into ookinetes upon activation in vitro was also reduced ( Fig . 7B ) . Absence of BCKDH thus reduced both the numbers of morphologically mature gametocytes and their ability to develop in vitro . To assess the role of BCKDH in transmission , Anopheles stephensi mosquitoes were allowed to feed on infected mice . Consistent with the in vitro data , the Pbe1a_ko mutant gave rise to a considerably reduced number of oocysts per infected mosquito midgut , an effect that was fully reversed by complementation with the PfE1a gene ( Fig . 7C ) . Importantly , after day 7 of infection , Pbe1a_ko showed a marked inability to increase in size ( Fig . 7D and 7E ) , and the mutant cysts invariably failed to undergo sporogony ( Fig . 7F ) . Pbe1a_ko-infected mosquitoes contained no sporozoites in their salivary glands on day 21 post infection , neither were they able to re-infect mice , either by bite or intravenous injection of disrupted salivary glands ( Fig . 7G ) . Taken together , these data demonstrate that in addition to its role for blood stage development in normocytes , BCKDH is necessary for normal gametocyte production and fitness , and that BCKDH is essential for oocysts to mature and undergo sporogony .
Using a combination of genetic and metabolomic approaches , we show that the BCKDH complex not only substitutes for the loss of mitochondrial PDH in Apicomplexa , but is also required for normal growth and virulence of T . gondii and P . berghei . While T . gondii and Plasmodium spp . were thought to depend on glycolysis for energy , recent studies have highlighted the potential importance of oxidative phosphorylation and mitochondrial metabolism [27] , [28] , [30] , [38] . Indeed 13C-glucose and 13C-glutamine labelling studies have recently confirmed the operation of a canonical TCA cycle in both T . gondii and P . falciparum [20] , [21] . A critical question raised by these studies concerns the identity of the enzyme ( s ) that feed carbon skeletons into the TCA cycle . In most organisms , the operation of the TCA cycle is dependent on production acetyl-CoA from pyruvate , via the catalytic activity of a mitochondrial PDH complex . Apicomplexa have a canonical PDH , but this complex is targeted to the apicoplast , suggesting that other enzymes fulfil the function of the PDH in the mitochondrion [13]–[15] . Recent studies have raised the possibility that mitochondrial acetyl-CoA could be generated from glycolytic pyruvate via one of the other TPP-dependent mitochondrial dehydrogenases [22] , [33] . However , direct evidence for involvement of either BCKDH or α-KDH , or another uncharacterized dehydrogenase has not been obtained . Here , we have generated T . gondii and P . berghei BCKDH null mutants by reverse genetics and demonstrated loss of mitochondrial glucose metabolism . Deletion of the BCKDH-E1a gene from T . gondii and P . berghei was non-lethal , but in each case led to a marked reduction in growth rate and impacted on parasite virulence in mice . Metabolite profiling coupled with 13C-U-glucose labelling established that BCKDH catalyses the conversion of mitochondrial pyruvate to 13C2-acetyl-CoA and fuels a canonical TCA cycle both in T . gondii and P . berghei . Specifically , deletion of BCKDH-E1a was associated with loss of labelling of 13C2-acetyl-CoA and +2 13C-isotopologues of other TCA cycle intermediates . In T . gondii , loss of the BCKDH-E1a gene also led to significant decreases in 2-hydroxyethyl-TPP , the intermediate in acetyl-CoA production from pyruvate , as well as acetyl-CoA and citrate . Conversely , pyruvate levels increased in a manner consistent with this substrate no longer being consumed by the enzyme . In vitro analysis of α-ketoacid dehydrogenase activity in T . gondii extracts demonstrated some affinity for both pyruvate and branched-chain keto-acids , but with a high selectivity for the conversion of pyruvate to acetyl-CoA compared to the minimal conversion of branched chain keto-acids to their respective branched chain acyl-CoA products . Taken together , our data demonstrate that the apicomplexan BCKDH complex has been repurposed to function as a PDH and allow the further oxidation of pyruvate in a canonical TCA cycle . Sequence alignment analysis of the T . gondii and P . berghei E1a with other BCKDH-E1a ( Fig . S3B and S3C ) and E1b subunits failed to ascertain whether active site residue substitutions can account for substrate versatility [39] , [40] and further structural characterization will be required to refine our understanding of the substrate specificity of this class of enzyme . However , these results account for the retention of BCKDH genes in Plasmodium spp . , despite the loss of other genes involved in BCAA degradation . They are also consistent with the presence of the two genes coding for the subunits of the mitochondrial pyruvate carrier recently described in yeast [41] , [42] ( Table S1 ) . Interestingly , more distantly related members of the Alveolata , such as dinoflagellates also lack a conventional mitochondrial PDH complex , but retain a BCKDH [17] , suggesting that the repurposing of BCKDH evolved early in the evolution of this group and is likely to be conserved in all members of this group that contain a functional TCA cycle . Intriguingly , the disruption of the BCKDH enzyme complex was associated with significant remodelling in the central carbon metabolism in T . gondii ( Fig . 8 ) . Metabolic profiling of Tge1a_ko parasites revealed that despite a decrease in citrate , other TCA cycle metabolites including the C4-dicarboxylic acids succinate , fumarate and malate were all unaltered or even increased in abundance , indicating that an alternative carbon source enters the TCA cycle below α-ketoglutarate . T . gondii tachyzoites co-utilize glutamine in the presence of glucose and appear to use this amino acid as an alternative substrate in the absence of glucose [43] . Unexpectedly , we found that carbon skeletons derived from 13C-glutamine were channelled into the gluconeogenic pathway in Tge1a_ko parasites under glucose-replete conditions ( Fig . 2C and 8 ) . This metabolic perturbation is not due to interruption of the early steps in the TCA cycle or increased glutaminolysis , since increased gluconeogenic flux is not observed in parasites treated with sodium fluoroacetate , an inhibitor of the TCA cycle enzyme aconitase [20] . Increased gluconeogenesis in Tge1a_ko might result from allosteric activation of key enzymes , as a result of the accumulation of pyruvate or other metabolites . Alternatively , the decreased production of acetyl-CoA in the mutant could lead to global changes in the acetylation state of multiple enzymes in these pathways [44]–[46] . The gluconeogenic enzyme , phosphenolpyruvate carboxykinase ( PEPCK ) is activated by deacetylation [47] , and this enzyme has been shown to be acetylated in T . gondii tachyzoites , as is cytosolic glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) [48] . Bacterial acetylated GAPDH has been reported to drive the forward reaction during glycolysis , while it is more active in catalysing the reverse reaction used in gluconeogenesis when deacetylated [45] . A global decrease in protein acetylation as a consequence of blocked acetyl-CoA synthesis could therefore underlie the inappropriate activation of gluconeogenesis in Tge1a_ko ( Fig . 8 ) . The phenotypic analyses of the BCKDH null mutant revealed an important impact on the physiology of the blood stages of the rodent malaria parasite . In contrast to other P . berghei mutants with defects in central carbon metabolism , such as the mitochondrial complex II ( succinate-ubiquinone reductase ) [28] and the type II NADH:ubiquinone oxidoreductase ( NDH2 ) [27] , the Pbe1a_ko mutant exhibited an obvious phenotypic defect during asexual development in mature red blood cells . Pbe1a_ko parasites readily invaded reticulocytes and normocytes , but failed to develop to schizonts in the latter . While the molecular basis for the selective growth in reticulocytes compared to normocytes is unknown , it is likely that differences in the metabolism of these two host cell types contribute to the observed tropism . In particular , reticulocytes are known to be metabolically more active than normocytes , and may contain essential metabolites that could compensate for the loss of BCKDH function [49] , [50] . In support of this conclusion , Pbe1a_ko parasite growth in normocytes was partially restored by supplementation of the medium with acetate ( Fig . 6 and 8 ) , indicating that the mutant is able to scavenge acetate or other essential nutrients from reticulocytes in order to survive in the absence of BCKDH . Alternatively , intrinsic- and parasite-induced differences in the permeability of the reticulocyte/normocyte plasma membrane [51]–[53] could lead to a differential accumulation of toxic compounds , such as lactate and pyruvate , and lethality in the Pbe1a_ko strain . These findings also raise the question of the importance of BCKDH for each of the different human malaria parasite species , given that P . falciparum can develop in both immature and mature erythrocytes , whereas P . vivax exhibits a strict preference for reticulocytes [54] and establishes a chronic infection in the liver [55] . We have recently shown that flux of pyruvate derived from glucose into the TCA cycle increases markedly as P . falciparum asexual RBC stages differentiate to gametocyte stages , implying that the steps catalysed by BCKDH may be regulated in a stage-specific manner [21] . Consistent with an increased role for P . falciparum BCKDH in sexual development , the P . berghei BCKDH-E1a mutant exhibited a reduction in both numbers and fitness of gametocytes . Both phenotypes were reversed by complementing the mutant with an orthologous gene from P . falciparum . While a reduction in gametocyte numbers alone would be difficult to interpret given the different infection courses of wild type and mutant parasites , a reduction in the ability of each microgametocyte to release microgametes and of macrogametes to convert to ookinetes is highly significant . Our phenotype is clearly less severe , but otherwise resembles that described by [28] in P . berghei for a mutant in mitochondrial complex II , predicted to have a disrupted TCA cycle downstream of both glucose and glutamine . These data are entirely consistent with a key role for BCKDH in a glucose-fuelled TCA cycle that in P . berghei increases in importance during sexual development . The P . berghei BCKDH-E1a is essential for transmission through mosquitoes as it is required for sporogony , a period when the young oocysts grow rapidly in size and differentiate into thousands of sporozoites . Oocyst growth was also blocked in an NDH2-deficient P . berghei mutant [27] , further highlighting the importance of oxidative phosphorylation during these stages of parasite development . Together , our data suggest that BCKDH plays a key role in the stage-specific changes in flux of glycolytic pyruvate into the TCA cycle in mosquito stages . This study highlights the metabolic adaptations that have occurred during the evolution of this diverse group of protists and provides new insights into the central carbon metabolism of key pathogenic stages . Since perturbation of mitochondrial metabolism results in significant decrease in parasite fitness in the mammalian host and inhibition of vector transmission , components of the BCKDH complex may be suitable targets for drug development .
All animal experiments were approved and performed in accordance with a project licence issued by the UK Home Office and by the Direction générale de la santé , Domaine de l’expérimentation animale ( Avenue de Beau-Séjour 24 , 1206 Genève ) with the authorization Number ( 1026/3604/2 , GE30/13 ) according to the guidelines set by the cantonal and international guidelines and regulations issued by the Swiss Federal Veterinary Office . No human samples were used in these experiments . Human foreskin fibroblasts ( HFF ) were obtained from ATCC . The antibodies used in this study were described before as follows: mouse monoclonal anti-myc ( 9E10 ) , polyclonal rabbit anti-GAP45 [56] , rabbit anti-PfProfilin [56] . For the production of specific polyclonal antibodies , PfBCKDH-E1a sequence ( PF13_0070 , aa 277 to 425 ) was amplified using primers 2391 and 2392 ( Table S2 ) and cloned into pETHb in frame with 6 histidine residues between the NcoI and SalI restriction sites . The corresponding recombinant protein was expressed in Escherichia coli BL21 strain and purified by affinity chromatography on Ni-NTA agarose ( Qiagen ) according to the manufacturer’s protocol under denaturing conditions . Antibodies against PfBCKDH-E1a were raised in rabbits by Eurogentec S . A . ( Seraing , Belgium ) according to their standard protocol . Immunofluorescence assay ( IFA ) and Western blots analysis with these antibodies were performed as previously described [33] . T . gondii tachyzoites ( RHku80_ko ( RH ) , RHku80_ko/bckdhE1a_ko ( Tge1a_ko ) ) were maintained in HFF using Dulbecco’s Modified Eagle’s Medium ( DMEM , GIBCO , Invitrogen ) supplemented with 5% foetal calf serum , 2 mM glutamine and 25 µg/ml gentamicin at 37°C and 5% CO2 . 2 kb genomic flanking sequences ( FS ) of TgBCKDH-E1a ( TGME49_239490 , www . toxodb . org ) ORF were amplified using LATaq polymerase ( TaKaRa ) and primers 2821 and 2822 for the 5’FS and primers 2823 and 2824 for the 3’FS ( Table S2 ) . Amplified regions were then cloned into the KpnI , XhoI sites for the 5’FS and BamHI , NotI sites for the 3’FS of the pTub5HXGPRT vector . T . gondii RHku80_ko tachyzoites were transfected by electroporation as previously described [57] and stable transfectants were selected for by hypoxanthine-xanthine-guanine-phosphoribosyltransferase ( HXGPRT ) expression in the presence of mycophenolic acid and xanthine as described earlier [58] . Parasites were cloned by limiting dilution in 96 well plates and clones were assessed by genomic PCR and Western blot analysis . Plaque assays: HFF monolayers were infected with parasites and let to develop for 7 days before fixation with PFA/GA and Giemsa staining ( Sigma-Aldrich GS500 ) mounted with Fluoromount G and visualized using ZEISS MIRAX imaging system equipped with a Plan-Apochromat 20×/0 . 8 objective at the bioimaging facility of the Faculty of Medicine , University of Geneva . Intracellular growth assays: Prior to infection , the HFF monolayers were washed and pre-incubated for 24 h with medium containing the relevant carbon source and kept in this medium for the rest of the experiment . Complete medium is DMEM 41966 ( Gibco , Life Technologies ) supplemented with 5% FCS , up to 6 mM glutamine , 25 µg/ml gentamicine . Medium depleted in glutamine is DMEM 11960 supplemented with 25 µg/ml gentamicine ( Gibco , Life Technologies ) and medium depleted in glucose is DMEM 11966 supplemented with up to 6 mM glutamine and 25 µg/ml gentamicine . HFF were inoculated with parasites and coverslips were fixed 24 h post-infection with 4% PFA and stained by IFA with rabbit anti-TgGAP45 , mouse anti-myc 9E10 . Number of parasites per vacuole was counted in triplicates for each condition ( n = 3 ) . More than 200 vacuoles were counted per replicate . On day 0 , mice were infected by intraperitoneal injection with either wild type RH or Tge1a_ko parasites ( ∼15 parasites per mouse ) . 5 female CD1 mice were infected per group . The health of the mice was monitored daily until they presented severe symptoms of acute toxoplasmosis ( bristled hair and complete prostration with incapacity to drink or eat ) and were sacrificed on that day in accordance to the Swiss regulations of animal welfare . 21 days post-infection , sera from mice that survived primary infection were assessed for seroconversion by Western blot for the presence of anti-T . gondii antibodies . Mice were then challenged with ∼1000 wild-type RH parasites to assess immunization and survival . The P . berghei ANKA GFP-con clone 2 . 3 . 4 was used to generate transgenic parasites and maintained in female CD1 or Theiler's Original outbred mice as described previously [59] . The course of infection was monitored on Giemsa-stained tail blood smears . PbBCKDH-E1a sequence was retrieved from the online Plasmodium genome database , ( PBANKA_141110 , www . plasmodb . org ) . To generate the PbBCKDH-E1a knockout ( Pbe1a_ko ) , primers 3835 and 2482 were used to amplify a 2 kb region of homology at the 5’end of the PbBCKDH-E1a locus ( Table S2 ) . The PCR fragment was cloned between KpnI and ApaI restriction sites of the pBS-DHFR vector containing the T . gondii dhfr conferring pyrimethamine resistance [60] . The 3’ flanking region 2 kb was amplified using primers 3836 and 2483 and cloned between EcoRV and BamHI restriction sites of pBS-DHFR . The final construct was linearized with NotI prior to transfection . Complementation of Pbe1a_ko by P . falciparum BCKDH-E1a was performed using a knock-in strategy in the promoter region of PbBCKDH-E1a still present in Pbe1a_ko parasites . Primers 4067 and 4068 were used to amplify this promoter region and the PCR ( pPbE1a ) fragment was cloned between the NotI and ApaI sites of pARL-GFP-Ty-hDHFR . Primers 4070 and 4256 were used to amplify the PfBCKDH-E1a cDNA ( PF3D7_1312600 , www . plasmodb . org ) and cloned between the ApaI and NcoI of the pPbE1a-GFP-Ty-hDHFR . Finally , the 3’UTR of PbBCKDH-E1a was amplified using primers 4257 and 4258 and the PCR fragment cloned between the NcoI and EcoRV sites of pPbE1a-PfE1a-Ty-hDFR in order to generate the final plasmid , pPbE1a-PfE1a-3’PbE1a-hDHFR . This plasmid was linearized by MfeI prior transfection in Pbe1a_ko strain . Transfections were carried out as previously described [61] . Briefly , after overnight culture of infected red blood cells ( 37°C , 90 rpm ) , mature schizont-infected RBC were purified by Nycodenz gradient and collected . 100 µL of Human T Cell Nucleofector Kit ( Amaxa ) and 15 µg of digested DNA . Electroporation was performed using the U33 program of the Nucleofector electroporator ( Amaxa/Lonza ) . Electroporated parasites were mixed with blood enriched in reticulocytes from phenylhydrazine-treated mice to allow re-invasion and immediately injected ( intraperitoneal ) into CD1 female mice . Mice were treated with pyrimethamine in drinking water ( conc . Final = 0 . 07 mg/ml ) , 24 hr after infection . After 3 drug-cycling , infected blood was collected and P . berghei genomic DNA was extracted with SV Wizard Genomic DNA kit ( Promega ) . Pbe1a_ko pools were cloned in RAG-1 -/- mice by limiting dilution . The integration of the different constructs was confirmed by genomic PCR , using primers listed in table S2 and loss or presence of the protein was validated by Western blot analysis . After separation of the plasma from red blood cells and washes in RPMI by centrifugation ( 300× g for 10 min ) , mouse reticulocytes were separated from normocytes by Percoll/NaCl density gradient ( 1 . 096–1 . 058 g/mL ) and centrifuged at 250× g for 30 min , as previously described [62] , [63] . Reticulocytes were collected from the interface of the two Percoll layers and washed twice with RPMI before culturing . Invasion efficiency in normocytes or purified reticulocytes was assessed by flow cytometry as previously described [64] . Haematocrit levels were observed in mice over the course of infection with P . berghei wild type and Pbe1a_ko parasites . Briefly , every two days , haematocrit-capillaries ( Hirschmann Laborgeräte , 75 mm/18 µL , ammonia heparinized 0 , 9 IU/capillary ) were filled with tail blood and centrifuged at 10 , 000 rpm for 5 min to separate the blood layers . Percentage haematocrit was calculated by dividing the packed red blood cell volume length by the total blood volume length ( red blood cells and serum ) . The mice used for the phenotyping were pretreated with 150 µl of 6 mg/ml of phenylhydrazine and infected with 107 parasites two days later . Three days after infection parasitaemia and gametocytaemia were quantified on Giemsa-stained thin blood smears . To quantify exflagellation rates 10 µl blood collected from the tail vain was mixed with 500 µl of ookinete medium ( RPMI1640 containing 25 mM HEPES , 20% FCS , 100 µM xanthurenic acid , pH 7 . 4 ) and examined using an improved Neubauer hemocytometer 12 min later . The number of erythrocytes and exflagellating microgametocytes were counted and expressed as a percentage of all microgametocytes as independently determined from stained blood films . To assess ookinete formation 100 µl infected blood were cultured in 10 ml ookinete medium at 19°C for 20 h . Live staining with a Cy3-congugated mouse monoclonal antibody against the P28 protein labeled banana shaped ookinetes and undifferentiated macrogamete derived parasites , whose shape was assessed by microscopy and recorded for at least 100 cells per sample . For transmission studies ∼200 female A . stephensi mosquitos were allowed to feed on the same infected mice . Exflagellation assays , ookinete cultures and mosquito feeds were performed using the same animals ( three per strain ) on day three of the infection . Midguts from 20 fed mosquitoes from each cage were dissected in PBS 7 and 14 days after feeding , examined for oocysts , photographed and analysed as described [65] . On day 21 post feeding salivary glands from 20 mosquitoes per group were dissected , homogenized and sporozoites quantified in a hemocytometer . Finally , on day 23 the remaining mosquitoes were allowed to feed on uninfected mice that had been pre-treated with phenylhydrazine two days before . In parallel the salivary glands from 20 mosquitoes were homogenized and injected intravenously into different mice . All animals were observed for 14 days and the time of parasite appearance in the blood was noted . Metabolite extraction and analysis of T . gondii tachyzoites was performed as previously described [20] . Intracellular and egressed tachyzoites ( 2×108 cell equivalents ) were extracted in chloroform/methanol/water ( 1∶3∶1 v/v containing 1 nmol scyllo-inositol or 5 nmol 13C-valine as internal standard for GC-MS and LC-MS , respectively ) for 1 h at 4°C , with periodic sonication . For GC-MS analysis polar and apolar metabolites were separated by phase partitioning . Polar metabolite extracts were derivitised by methoximation and trimethylsilylation ( TMS ) and analysed by GC-MS as previously described [66] . For LC-MS analysis , monophasic metabolite extracts were analysed directly with a ZIC-pHILIC - Orbitrap platform and data analysed with IDEOM as previously described [19] , [67] . All metabolites were identified by comparison of retention time and exact mass ( LC-MS ) or mass spectra ( GC-MS ) with authentic chemical standards , with the exception of 2-hydroxyethyl-TPP putatively identified by exact mass ( <1 ppm ) and predicted retention time based on a quantitative structure-retention relationship model for this chromatographic method [68] . T . gondii: stable isotope labelling , metabolite extraction , and GC-MS analysis was performed for polar metabolites , as previously described [20] . Briefly , infected HFF or freshly egressed tachyzoites were incubated in low-glucose , glutamine-free DMEM , supplemented with either 13C-U-glucose , 13C-U-glutamine , 13C-U-leucine or 13C-U-leucine/13C-U-isoleucine/13C-U-valine ( final concentration 8 mM , Spectra Stable Isotopes ) . Parasites were harvested after 4 h and metabolites extracted as above . Changes in the mass isotopologue abundances of key intermediates in central carbon metabolism were assessed by GC-MS or LC-MS analysis . The level of labelling of individual metabolites was calculated as the percent of the metabolite pool containing one or more 13C atoms after correction for natural abundance and amount of 13C-carbon source compared to 12C-carbon source in the culture medium ( as determined by GC-MS analysis ) . The mass isotopologue distributions ( MIDs ) of individual metabolites were corrected for the occurrence of natural isotopes in both the metabolite and , in GC-MS experiments , the derivatisation reagent [69] . An untargeted metabolome-wide isotope analysis was performed on high resolution LC-MS data to detect all labelled metabolic products of 13C-U-glucose or 13C-U-leucine/13C-U-isoleucine/13C-U-valine according to the method previously described [70] . P . berghei: stable isotope labelling , metabolite extraction , and GC-MS analysis was adapted from the above protocol . Blood ( 1 mL ) from individual mice infected ( iRBC ) with WT or Pbe1a_ko parasites ( blood at equivalent parasitemia of ∼4% ) was collected and all white cells removed by passage of the blood over cellulose CF11 columns . Parasites were cultured for maturation in vitro for 5 h ( RPMI1640 containing 25 mM HEPES [Sigma] , 0 . 5% Albumax , 0 . 2 mM hypoxanthine [pH 7 . 5] , 25 µg/ml Gentamycin ) , before addition of 8 mM 13C-U-glucose , 13C-U-glutamine , or 13C-U-leucine as required . After 5 h of labelling , cultures were rapidly transferred to a 50 mL centrifuge tube and cellular metabolism quenched as above and schizont-infected RBCs ( iRBCs ) were purified from uninfected and ring-infected RBCs on Nycodenz gradient , at 4°C . iRBCs and uninfected RBCs ( control ) were pelleted by centrifugation ( 800× g , 10 min , 4°C ) , washed three times with ice-cold PBS , extracted and analysed as above . Cell extracts were prepared from 108 freshly egressed T . gondii tachyzoites by hypotonic lysis with ice-cold lysis buffer ( 1 mM NaHEPES ( pH 7 . 4 , Sigma ) , 2 mM ethylene glycol tetraacetic acid ( Sigma ) , 2 mM dithiothreitol ( Biovectra ) ) for 10 min on ice , in the presence of EDTA-free protease inhibitors ( complete , Roche ) . Extracts were incubated with 0 . 5 mM or 2 mM of the test keto-acid ( 4-methyl-2-oxopentanoate , 3-methyl-2-oxobutanoate or pyruvate ) in the presence of cofactors ( 5 mM NAD+ , 1 mM CoA , 0 . 1 mM Thiamine-PP , 0 . 1 mM FAD and 0 . 05 mM lipoic acid ) in 200 µL MgCl2/KH2PO4 ( 1 . 3 mM/33 mM ) buffer ( pH 7 ) at 37°C for 2 h ( n = 2 ) . Samples ( 20 µL ) were extracted in 80 µL methanol/acetonitrile ( 1∶3 ) and analysed by LC-MS ( pHILIC-QTOF ) as described in the metabolomics methods . 3-methylbutanoyl-CoA , 2-methylpropanoyl-CoA and acetyl-CoA were quantified by reference to calibration curves of authentic standards in sample matrix ( extracted lysates ) . Additional TPP intermediates were not quantified due to lack of authentic reference standards , and data are expressed as LC-MS peak areas . The raw high-resolution MS data were interrogated for the occurrence of other possible metabolic products derived from the branched chain keto-acids or acyl-CoAs , and none were detected . Additional controls were analysed to validate the assay including substrate-free and cofactor-free incubations , technical replicates of pooled samples for LC-MS quality control and matrix controls for RH and Tge1a_ko samples ( data not shown ) .
|
The mitochondrial tricarboxylic acid ( TCA ) cycle is one of the core metabolic pathways of eukaryotic cells , which contributes to cellular energy generation and provision of essential intermediates for macromolecule synthesis . Apicomplexan parasites possess the complete sets of genes coding for the TCA cycle . However , they lack a key mitochondrial enzyme complex that is normally required for production of acetyl-CoA from pyruvate , allowing further oxidation of glycolytic intermediates in the TCA cycle . This study unequivocally resolves how acetyl-CoA is generated in the mitochondrion using a combination of genetic , biochemical and metabolomic approaches . Specifically , we show that T . gondii and P . bergei utilize a second mitochondrial dehydrogenase complex , BCKDH , that is normally involved in branched amino acid catabolism , to convert pyruvate to acetyl-CoA and further catabolize glucose in the TCA cycle . In T . gondii , loss of BCKDH leads to global defects in glucose metabolism , increased gluconeogenesis and a marked attenuation of growth in host cells and virulence in animals . In P . bergei , loss of BCKDH leads to a defect in parasite proliferation in mature red blood cells , although the mutant retains the capacity to proliferate within 'immature' reticulocytes , highlighting the role of host metabolism/physiology on the development of Plasmodium asexual stages .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"plasmodium",
"vivax",
"protozoans",
"malarial",
"parasites",
"biology",
"and",
"life",
"sciences",
"malaria",
"protozoan",
"infections",
"plasmodium",
"falciparum",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"organisms",
"toxoplasmosis"
] |
2014
|
BCKDH: The Missing Link in Apicomplexan Mitochondrial Metabolism Is Required for Full Virulence of Toxoplasma gondii and Plasmodium berghei
|
Cytomegaloviruses ( CMVs ) persistently and systemically infect the myeloid cells of immunocompetent hosts . Persistence implies immune evasion , and CMVs evade CD8+ T cells by inhibiting MHC class I-restricted antigen presentation . Myeloid cells can also interact with CD4+ T cells via MHC class II ( MHC II ) . Human CMV ( HCMV ) attacks the MHC II presentation pathway in vitro , but what role this evasion might play in host colonization is unknown . We show that Murine CMV ( MCMV ) down-regulates MHC II via M78 , a multi-membrane spanning viral protein that captured MHC II from the cell surface and was necessary although not sufficient for its degradation in low pH endosomes . M78-deficient MCMV down-regulated MHC I but not MHC II . After intranasal inoculation , it showed a severe defect in salivary gland colonization that was associated with increased MHC II expression on infected cells , and was significantly rescued by CD4+ T cell loss . Therefore MCMV requires CD4+ T cell evasion by M78 to colonize the salivary glands , its main site of long-term shedding .
Herpesviruses establish persistent , productive infections , with high prevalence in most populations and significant disease burdens . Vaccine development against Human cytomegalovirus ( HCMV ) is motivated by a high incidence of congenital infection . However vaccines to date have not prevented viral persistence or transmission . CMVs have extensive arsenals of immune evasion genes , so a key task in developing better vaccines is to understand how viral evasion limits immune defence . This requires animal models . Murine CMV ( MCMV ) provides an accessible model for HCMV . Both viruses evade CD8+ T cells by attacking MHC class I ( MHC I ) -restricted antigen presentation . For MCMV this promotes salivary gland ( SG ) infection [1] . Similar evasion by Rhesus CMV helps it to re-infect immune hosts [2] . HCMV also attacks MHC II-restricted antigen presentation , limiting MHC II induction by interferon-γ ( IFNγ ) [3] and triggering MHC II degradation via US2 [4] , US3 [5] and pp65 [6] . CD4+ T cell evasion is implied . Nonetheless clinical deficiencies show that CD4+ T cells are a key defence against HCMV [7] . They also help to control MCMV in the salivary glands [8] via an IFNγ-dependent mechanism [9] . Thus , CMV infections feature both CD4+ T cell evasion and CD4+ T cell-dependent control . How these fit together is unclear . CD4+ T cell evasion through attack on MHC II has a less clear rationale than for CD8+ T cells and MHC I . MHC II can present cell-endogenous antigens after autophagy [10] , but presents mainly cell-exogenous antigens that are recovered from endosomes . Thus the presenting cell need not be infected , making cytotoxicity an inappropriate response to antigen recognition . Moreover , while CD4+ T cells can induce target cell apoptosis [11] they are much less effective killers than CD8+ T cells . Mainly they activate antigen+ presenting cells . γ-herpesviruses even exploit CD4+ T cell recognition to drive infected B cell proliferation [12] . CD4+ T cell evasion by CMVs suggests that the delivery of activation signals to infected presenting cells would hinder a key aspect of host colonization . Herpesviruses infect multiple cell types , and the outcome when immunity meets viral evasion can be cell type-dependent . For example CD8+ T cell evasion by Murid Herpesvirus-4 ( MuHV-4 ) results in CD8+ T cells controlling epithelial but not myeloid cell infection and gives CD4+ T cells an essential role in host defence [13] . CMVs are shed from the SG , where electron microscopy indicates that MCMV is produced mainly by acinar cells [14] . These lack MHC II , SG MHC II expression being limited to interdigitating myeloid cells [15] . However MCMV reaches the acinar cells via infected migratory dendritic cells [16] . Thus , dendritic cell recognition by CD4+ T cells could reduce acinar cell infection indirectly . Even if the dendritic cells were not killed , CD4+ T cell engagement might deliver anti-viral signals and arrest their migration , thereby reducing their capacity to spread infection . Analysis of how MCMV might evade CD4+ T cells has focussed on cytokines: low MHC II expression on infected cells is attributed to an inhibition of IFNγ signalling [17] and an induction of IL-10 [18] reducing MHC II transcription . The viral genes responsible have not been identified . M27 reduces IFNγ signaling through STAT-2 , but from day ( d ) 7 of infection M27- MCMV is no less attenuated in IFNγ receptor-deficient than in wild-type mice [19] , so its main effect seems to be on type I IFN . We show that MCMV , like HCMV , degrades MHC II in infected cells , and that this requires M78 , a multi-membrane spanning viral protein with homology to chemokine receptors but without canonical signalling motifs . M78- MCMV shows reduced virus production from infected macrophages , and poorly colonizes the SG [20–22] . Myeloid cells infected by M78- but not wild-type ( WT ) MCMV expressed MHC II in vivo , and CD4+ T cell loss significantly reversed the M78-dependent defect in SG colonization . Therefore CD4+ T cell evasion is an important M78 function that acts in myeloid cells and is necessary for MCMV to colonize its main site of long-term shedding .
Most myeloid cells express MHC II inducibly rather than constitutively . IFNγ induces MHC II expression but also inhibits MCMV replication [9 , 23] . To track viral effects on MHC II without this complication , we induced MHC II in RAW-264 monocytes ( normally MHC II- ) by expressing the MHC II transactivator ( C2TA ) , which acts down-stream of IFNγ [24] . RAW-C2TA cells were constitutively MHC II+ . When they were exposed to MCMV-GFP , GFP+ cells lost MHC II but not CD44 or CD71 ( Fig 1a ) . Immunostaining cells exposed to β-galactosidase ( βgal ) + MCMV ( Fig 1b ) similarly showed normal or increased MHC II expression in βgal- ( uninfected ) cells and MHC II loss in strongly βgal+ cells . In weakly βgal+ cells MHC II was clumped and internalized , suggesting that this was an intermediate stage in down-regulation . Antibody IBL5/22 , which recognizes a conformation-independent MHC II epitope , gave equivalent results to antibody M5/114 ( Fig 1c ) , so MCMV caused MHC II degradation rather than just denaturation . MCMV also reduced MHC II on thioglycollate-induced peritoneal macrophages of BALB/c and C57BL/6 mice ( Fig 1d ) . Thus the degradation did not depend on constitutive C2TA expression , and applied across at least 2 MHC II haplotypes . Unlike RAW-C2TA cells , MCMV-infected fibroblasts expressing C2TA preserved MHC II ( Fig 2a ) . This result suggested that MHC II degradation might require macrophage-specific , acidic endosomes . To test the need for low pH , we exposed infected RAW-C2TA cells to ammonium chloride or bafilomycin , which inhibit endosomal acidification . Both treatments significantly rescued MHC II expression ( Fig 2b and 2c ) . By contrast the proteasome inhibitor MG-132 had no effect ( Fig 2d and 2e ) . Complete MHC II loss from infected cells suggested that in addition to any effect on nascent protein , MCMV removed mature MHC II from the plasma membrane . To test this idea , we incubated RAW-C2TA cells with an MHC II-specific antibody ( 1h , 4°C ) , using excess antibody to minimize cross-linking and using rat antibodies to avoid binding by MCMV Fc receptors . We washed off unbound antibody , infected the cells or not overnight ( 18h , 37°C ) , then added a fluorescently labelled secondary antibody ( 1h , 4°C ) and assayed its binding by flow cytometry ( Fig 2f ) . Thus , we assayed MHC II that was on the plasma membrane before infection ( for primary antibody binding ) then retained there ( for secondary antibody binding ) . We also stained MHC II on uninfected cells without a 37°C incubation ( 1h , 4°C ) ( t = 0h ) . The surface MHC II of uninfected RAW-C2TA cells was similar at t = 0h and t = 18h , indicating that its turnover is normally slow . By contrast MCMV-infected ( GFP+ ) cells showed a marked loss of surface-tagged MHC II over 18h . Therefore MCMV removed MHC II from the plasma membrane of infected cells . As RAW-C2TA cells maintained surface MHC II for at least 18h , reduced synthesis—for example through transcriptional suppression—could not explain the down-regulation . Minimal MHC II down-regulation by MCMV in BALB/c-3T3-C2TA cells ( Fig 2a ) and rescue by bafilomycin ( Fig 2b–2e ) also argued against transcriptional suppression making a significant contribution . MHC II loss from the cell surface , and the intracellular distribution of the MHC II rescued by ammonium chloride or bafilomycin ( Fig 2 ) , argued that MCMV initiates MHC II degradation by relocalizing it to endosomes . The MCMV M78 is highly endocytic [21] , so we tested its contribution , infecting RAW-C2TA cells with WT or M78- MCMV ( Fig 3 ) . M78- MCMV is reported to grow normally in fibroblasts and poorly in macrophages [20 , 22] . We found impaired infectious virus production in RAW-264 cells after low multiplicity infection , with less defect after higher multiplicity infection ( Fig 3a ) . There was no difference between MHC II+ and MHC II- RAW-264 cells . There was no obvious defect in RAW-264 cell infection as measured by viral β-gal expression ( Fig 3b ) , or by flow cytometry of viral GFP expression ( S1a Fig ) , and RAW-C2TA cells infected by WT or M78- MCMV showed no obvious difference in viral or MHC II transcription ( S1b and S1c Fig ) . Flow cytometry of WT MCMV-infected cells showed a marked loss of both surface ( non-permeabilized ) , and total ( permeabilized ) MHC II ( Fig 3b ) . M78- MCMV infection little affected either . MHC II loss from peritoneal macrophages also required M78 ( Fig 3c ) . Identifying infected cells by immunofluorescent staining with a polyclonal anti-MCMV antibody ( Fig 3d ) or with an IE1-specific antibody ( Fig 3e ) confirmed that MHC II loss was M78-dependent . We also tested bone marrow-derived macrophages ( Fig 3f ) . When uninfected , approximately 30% of these were MHC II+; when infected with WT or M78+ REV MCMV , expression decreased to <5%; and when infected with M78- MCMV there was no decrease . Flow cytometry of cells infected with an independent M78- mutant ( Fig 3g ) confirmed that MHC II loss was M78-dependent . MHC I loss was contrastingly M78-independent . In H2d cells this depends primarily on M06 , which degrades MHC I in lysosomes [25] . Therefore M78 was not required for all MCMV-driven , endosomal protein degradation . As M78+ MCMV did not down-regulate CD71 ( Fig 1a ) , which cycles through early endosomes , or CD44 , M78 did not generally promote glycoprotein endocytosis and loss . RAW-C2TA cells expressing M78 via retroviral transduction showed some MHC II loss but much less than that of infected cells ( Fig 4a ) . Both MCMV-infected RAW-C2TA cells ( Fig 4b ) and RAW-C2TA cells transduced with an M78+ retrovirus ( Fig 4c and 4d ) expressed M78 in internal vesicles—previous studies have identified it in early endosomes [21] . In transduced cells , CD44 remained on the plasma membrane whereas MHC II relocalized to M78+ vesicles . MCMV-infected bone marrow-derived macrophages ( Fig 4e ) and RAW-C2TA cells transiently transfected with an M78 expression plasmid ( Fig 4f ) similarly showed MHC II / M78 co-localization in vesicles . Surprisingly , despite M78-transduced cells being readily obtained and despite retroviral transduction generally giving uniformly strong recombinant gene expression [26] , M78-transduced RAW-C2TA cells showed variable M78 expression ( Fig 4d ) , as did RAW-C2TA-M78 cell clones and an independently-derived RAW-C2TA-M78 line ( Fig 4g ) . Cells with strong M78 expression redistributed MHC II , but puromycin-resistant cells that expressed less M78 had less effect on MHC II . To understand why , we tracked M78 from the cell surface . We infected RAW-C2TA cells with MCMV expressing HA-tagged M78 ( 4h , 37°C ) , labelled them with HA-specific rabbit IgG and MHC II-specific rat IgG ( 1h , 4°C ) , washed off excess antibody , then incubated the cells at 37°C for different times before fixation , permeabilization and staining with anti-rabbit IgG and anti-rat IgG labelled secondary antibodies . M78 tagged by antibody at the infected cell surface was rapidly endocytosed ( Fig 4h ) . It co-localized with MHC II in internal vesicles , then became undetectable , as did surface-tagged MHC II ( Fig 4i ) . Transfected HA-M78 was also lost rapidly , whereas transfected HA-CCR5 was preserved ( Fig 4j ) . Therefore rapid M78 and MHC II endocytosis appeared to be accompanied by M78 degradation , and in infected cells also by MHC II degradation . In transduced cells , even high level M78 expression did not lead to MHC II loss ( Fig 4d ) . As M78 was necessary for MCMV-driven degradation , M78- MCMV provided an opportunity to understand what CD4+ T cell evasion contributes to host colonization in vivo . We gave mice intranasal ( i . n . ) WT or M78- MCMV ( Fig 5 ) . In the lungs MCMV infects myeloid cells and type II alveolar epithelial cells [27] . Both can express MHC II [13] . After 1-3d few WT-infected lung cells expressed MHC II , whereas many of those infected by M78- MCMV did so ( Fig 5a and 5b ) . Therefore M78 also down-regulated MHC II in vivo . M78 has been studied as an MCMV gene of unknown function: M78- MCMV given intraperitoneally ( i . p . ) shows reduced liver , spleen and SG infections [20]; given i . n . it shows reduced lung and SG infections [22] . Rat CMV lacking its M78 homolog ( R78 ) is also attenuated in vivo [28] . Plaque assays of infectious virus and QPCR of viral DNA showed normal acute lung infection . This reflected presumably that myeloid cells are not a major source of acute virus production in the lungs [27] . However M78- MCMV was cleared faster from the lungs , and showed a marked defect in SG infection ( Fig 5c ) . Antibody responses to M78- MCMV were significantly lower than those to WT infection ( Fig 6a ) , consistent with M78- viral loads being lower . ELIspot assays ( Fig 6b and 6c ) showed no obvious difference in CD4+ T cell response between M78- and WT MCMV . We assessed the functional contribution of CD4+ T cells to M78- MCMV attenuation by infecting BALB/c mice depleted of T cell subsets ( Fig 6d ) . CD8+ T cell depletion increased M78- MCMV titers in the lungs at d10 . However it increased WT titers by a similar amount ( p>0 . 5 ) . It did not significantly affect SG infection . Therefore M78- MCMV attenuation was not due to better control by CD8+ T cells . CD4+ T cell depletion did not alter lung infection . However it increased SG infection by M78- MCMV . Some M78-dependent defect remained relative to WT , but unlike with CD8+ T cell depletion there was a significant relative increase in M78- virus titers ( p<0 . 04 ) . CD4+ T cell depletion also increased M78- viral genome loads relative to WT ( Fig 6e ) . Therefore CD4+ T cell evasion by M78 was important for salivary gland colonization by MCMV . To confirm this result , we infected MHC II-deficient C57BL/6 mice ( IA-/- ) with WT , M78- mutant or M78+ revertant ( REV ) viruses ( Fig 6f and 6g ) . At d10 , virus titers in the lungs and SG of immunocompetent mice ( IA+/- ) were significantly lower for M78- MCMV than for WT or REV . IA-/- lungs showed higher M78- plaque titers , and the increase in M78- titer with CD4+ T cell loss ( IA-/- mice ) was significantly greater than for WT or REV MCMV ( p<0 . 05 ) . CD4+ T cell loss also increased M78- titers in SG , relative to both WT and REV ( p<0 . 004 ) , as well as M78- viral DNA loads ( p<0 . 01 ) . To exclude a generally greater effect of CD4+ T cell depletion on low level SG colonization , we compared our untagged M78 deletion mutant ( M78-I ) with M131- MCMV [29] , which also poorly colonizes SG ( Fig 6h ) . Again M78- titers in SG showed a greater defect relative to WT in IA+/- than in IA-/- mice ( p<0 . 01 ) , indicating significant rescue by CD4+ T cell loss . By contrast M131- titers in SG showed a greater defect relative to WT in IA-/- mice . Nor did CD4+ T cell depletion rescue SG infection by i . n . M33- MCMV ( Fig 6i ) . Therefore CD4+ T cell loss specifically rescued SG infection by M78- MCMV . Salivary acinar cells are MHC II- [14] ( Fig 6j , uninfected ) . However dendritic cells disseminate i . n . MCMV , appear in the SG before acinar cells become infected [16] , and can express MHC II . For at least 2 weeks after i . n . MCMV >80% of infected cells in the SG are CD11c+ ( myeloid ) rather than Aquaporin V+ ( acinar ) . At d10 after WT infection of CD4+ T cell-depleted mice , all infected cells ( n>100 , counting samples from 6 mice ) were MHC II- ( Fig 6j , WT ) . By contrast those infected by M78- MCMV were MHC II+ ( Fig 6j , M78- ) . Without CD4+ T cell depletion , no M78- infected cells were seen . Therefore M78 promoted SG colonization before acinar cell infection , by guarding infected , disseminating myeloid cells against CD4+ T cell engagement .
M33 of MCMV [30] , and UL33 and US28 of HCMV [31] encode chemokine receptor homologs that signal constitutively . M78 and its equivalents in HCMV ( UL78 ) and rat CMV ( R78 ) have not been shown to signal , nor to bind chemokines , implying that they have other functions . M78 relocalized MHC II to endosomes and was required for MCMV-driven MHC II degradation . M78- MCMV also produced less infectious virus than did WT MCMV from MHC II+ and MHC II- RAW-264 cells , so MHC II degradation is unlikely to be the only M78 function . However CD4+ T cell loss significantly rescued poor SG infection by M78- MCMV . Therefore M78-dependent CD4+ T cell evasion made a demonstrably important contribution to host colonization . MCMV reaches the SG via infected , migratory dendritic cells [16 , 32 , 33] . With WT MCMV , these cells lacked detectable MHC II . With M78- MCMV , no infected cells were seen in SG of immunocompetent mice , and CD4+ T cell loss led to the appearance of MHC II+ infected cells . Thus M78 promoted SG infection by increasing virus transport to the SG , presumably protecting infected dendritic cells against CD4+ T cell engagement . Impaired virus production by M78- infected myeloid cells might also impair infection transfer to SG acinar cells , but the primary defect was in the initial arrival of infected cells . Infected dendritic cells drive the systemic spread of i . n . MCMV [16] , but type 2 alveolar epithelial cells appear to produce most of the infectious virus in lungs [27] . Thus , normal acute lung infection by M78- MCMV was consistent with a myeloid cell-focussed defect . M78 could be considered a systemic spread-specific infection module that couples CD4+ T cell evasion to virus production in myeloid cells . Despite MHC II degradation by CMVs , CD4+ T cells play an important role in infection control [7 , 8] . We hypothesize that protective CD4+ T cells normally respond to antigen on uninfected presenting cells and act indirectly . This idea is supported by salivary acinar cells lacking MHC II [14] , by most MHC II presenting cell-exogenous rather than cell-endogenous antigens , and by MCMV disrupting other presenting functions in infected myeloid cells [34 , 35] . The failure of M78 disruption to increase MCMV-specific CD4+ T cell responses was consistent with uninfected presenting cells priming most CD4+ T cells . It follows that protective CD4+ T cells may not directly recognize CMV-infected cells , but rather recruit and activate other anti-viral effectors with independent modes of recognition , for example NK cells . What then does MHC II degradation in infected cells achieve ? Because MCMV exploits normal dendritic cell migration to spread [16] , infected cells are likely to encounter CD4+ T cells in lymph nodes . CD4+ T cell recognition of MHC II plus antigen on infected dendritic cells would not necessarily lead to killing , as CD4+ T cells primarily activate rather than kill engaged presenting cells . ( CD8+ T cell evasion is probably more important for infected cell survival . ) However CD4+ T cell-derived cytokines can have anti-viral effects [9] , and CD4+ T cell engagement would reduce the migration of infected myeloid cells and so their capacity to spread infection . Increased CD4+ T cell engagement would also promote local antibody responses and innate effector recruitment . Thus MHC II expression on infected myeloid cells , even if presenting non-viral antigens , would open up a new front of host defence , with more precise targeting of recruited effectors to infected cells . By removing MHC II from infected myeloid cells , M78 isolated them from any CD4+ T cell engagement , promoting systemic infection spread and making MCMV-infected cells harder for CD4+ T cell-dependent defences to find .
BALB/c , C57BL/6 and IA-/- mice [36] were maintained at University of Queensland animal units , and infected i . n . at 6–12 weeks old ( 3x104 p . f . u . in 30μl , under anesthesia ) . IA-/- mice for breeding were kindly provided by Prof . Geoff Hill ( Queensland Institute for Medical Research ) . We depleted CD4+ / CD8+ T cells with mAbs GK1 . 5 / 2 . 43 ( Bio X Cell , 200μg/mouse/48h , from 96h pre-infection ) . Flow cytometry of spleen cells showed that the depletions were >95% complete ( S2 Fig ) . Embryonic fibroblasts were obtained from 15–17d old embryos harvested from pregnant out-bred CD1 mice . Experiments were approved by the University of Queensland Animal Ethics Committee ( projects 301/13 , 391/15 and 479/15 ) in accordance with the Australian code for the care and use of animals for scientific purposes , from the Australian National Health and Medical Research Council . M78 was amplified from K181 MCMV with Q5 polymerase ( New England Biolabs ) , adding MfeI and SalI sites to its 5' and 3' ends , ligated into pMSCV-IRES-PURO [37] and verified by DNA sequencing . Expression plasmids for HA epitope-tagged-M78 and CCR5 are described [21] . Each HA tag was N-terminal and so extracellular when the protein spanned the plasma membrane . Peritoneal macrophages were obtained by peritoneal lavage 48h after i . p . 3% Brewer's thioglycollate . B cells were removed by adherence to plastic and washing with PBS . Recovered cells were >90% F4/80+CD19- . Macrophages were grown from bone marrow by culture with M-CSF-1 ( 10ng/ml , Peprotech ) . These cells , mouse embryo-derived fibroblasts , NIH-3T3 ( American Type Culture Collection ( ATCC ) CRL-1658 ) , 293T ( ATCC CRL-3216 ) , RAW-264 cells ( ATCC TIB-71 ) , RAW-264 cells transduced with the human MHC II transactivator to induce MHC II ( RAW-C2TA ) [38] , BALB/c-3T3 and BALB/c-3T3 cells transduced with C2TA , were grown in Dulbecco’s modified Eagle’s medium with 2mM glutamine , 100IU/ml penicillin , 100μg/ml streptomycin and 10% fetal calf serum ( complete medium ) . Retroviral transduction was by transfecting 293T cells with pMSCV-M78-PURO and a packaging plasmid [37] , adding supernatants to cells with hexadimethrine bromide ( 10μg/ml ) , then selecting with puromycin ( 10μg/ml ) . RAW-C2TA cells were transfected by electroporation . We used MCMV strain K181 . Variants expressing GFP from the M131 intron ( MCMV-GFP ) [27] or βgal from the M33 intron ( MCMV-βgal ) [39]; with a premature stop codon in M131 ( M131- ) [29]; with a βgal cassette replacing M33 ( M33- ) [39]; with a βgal expression cassette at genomic coordinate 111681 ( Genbank GU305914 ) disrupting M78 ( M78- ) ; and a revertant with an N-terminal HA epitope tag on M78 ( REV ) [21] are described . An independent M78 mutant ( M78-I ) was made by homologous recombination , deleting the ORF ( coordinates 111084–112499 ) . This mutation was also recombined into MCMV-GFP . Viruses were grown in NIH-3T3 cells . Viruses were plaque assayed on embryonic fibroblasts [27] . Statistical comparison was by Student's 2-tailed unpaired t test unless otherwise stated . Organs were fixed in 1% formaldehyde-10mM sodium periodate-75mM L-lysine ( 24h , 4°C ) , equilibrated in 30% sucrose ( 18h 4°C ) , then frozen . 6μM sections were blocked with 0 . 3% Triton X-100 / 5% normal goat serum , then incubated ( 18h , 4°C ) with mAbs to MHC II ( rat , M5/114 ) and MCMV IE1 ( mouse IgG1 , CROMA101 ) [40] . Sections were washed x3 in PBS , incubated ( 1h , 23°C ) with Alexa 488-goat anti-mouse IgG1 and Alexa 568-goat anti-rat IgG pAb plus DAPI ( 1μg/ml ) , then washed x3 in PBS , and mounted in ProLong Gold ( Life Technologies ) . Cultured cells were adhered to coverslips , then fixed ( 2% formaldehyde , 30min ) , blocked in PBS / 0 . 1% Triton-X100 / 1% bovine serum albumin , then stained with antibodies to MHC II ( M5/114 or IBL5/22 ) , CD44 ( rat mAb IM7 ) , MCMV IE1 ( CROMA101 ) , βgal ( chicken pAb , AbCam ) , MCMV ( rabbit pAb ) , and M78 ( rabbit pAb ) [21] . Cells were washed x3 in PBS / 0 . 1% Tween-20 , incubated with combinations of Alexa488-goat anti-chicken IgG pAb , Alexa488-goat anti-mouse IgG1 , Alexa488- or Alex568-goat anti-rabbit pAb and Alexa568-goat anti-rat IgG pAb , plus DAPI ( 1μg/ml ) , then washed x3 in PBS / 0 . 01% Tween-20 , x2 in PBS and mounted ProLong Gold . GFP fluorescence was visualized directly . Images were acquired with a Zeiss LSM510 microscope and analyzed with ImageJ . Cells were detached from plates , blocked with 1% BSA / 1μg/ml anti-CD16/32 ( 2 . 4G2 ) , incubated ( 1h , 4°C ) with antibodies to MHC II ( APC-mAb M5/114 ) , CD44 ( biotin-IM7 ) , MHC class I ( biotin-34-5-8S ) , CD71 ( biotin-C2 , BD Biosciences ) , washed x2 in PBS , then incubated with fluorescein-streptavidin , washed x2 in PBS and analysed on an Accuri flow cytometer . GFP fluorescence was measured directly . To detect βgal , cells were fixed in 2% PFA after surface staining , then permeabilized in 70% ethanol , washed x2 and incubated with chicken anti-βgal followed by Alexa488-goat anti-chicken IgG pAb ( AbCam ) . To test CD4+ and CD8+ T cell depletions , spleen cells were stained with antibodies to CD4 ( fluorescein-RM4-4 ) , and CD8β ( phycoerythrin-H35-17 . 2 ) ( BD Biosciences ) . DNA was extracted from organs or blood ( Wizard Genomic DNA Purification , Promega ) . MCMV coordinates 111218–111461 were amplified ( LightCycler 480 SYBR green , Roche Diagnostics ) and converted to genome copies by comparison with plasmid DNA amplified in parallel . Cellular DNA was quantified in the same samples by amplification of a β-actin gene segment . Viral DNA loads were normalized by cellular DNA loads . RNA was extracted from cells ( Ambion ) and reverse-transcribed with an oligo-dT primer ( New England Biolabs ) . MHC II , β2M cellular cDNAs and IE1 and M33 viral cDNAs were then amplified by PCR . To distinguish cDNA from genomic DNA , each primer pair spanned an intron . The primers were ( 5' to 3' ) : MHC II—GATGCCGCTCAACATCTTGCTC and CATCCACACAGCTTATTAGGAATG; β2M—TAGACCAAAGATGAGTAACTGCATC and GAGACTGATACATACGCCTGCAG; IE1—AACCGTCCGCTGTGACCTGAC and CGATGCGCTCGAAGATATCATTG; M33—TCAGGATGATCACCGTGTTGATG and GAAACTTCTTAACCTTTCCAACGG . PCR products were separated by electrophoresis on 2% agarose gels and visualized by staining with ethidium bromide .
|
Human cytomegalovirus is the commonest infectious cause of harm to unborn children . Vaccines have not stopped it establishing chronic , systemic infections . Murine cytomegalovirus ( MCMV ) provides an accessible model to understand why . We show that MCMV evades CD4+ T cells via its M78 protein , and that this helps infection to spread despite the immune response . Thus while CD4+ T cells are important for host defence , viral evasion limits their capacity to act alone in controlling infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"nuclear",
"staining",
"immunology",
"bone",
"marrow",
"cells",
"clinical",
"medicine",
"cytotoxic",
"t",
"cells",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"white",
"blood",
"cells",
"major",
"histocompatibility",
"complex",
"animal",
"cells",
"t",
"cells",
"spectrophotometry",
"cytophotometry",
"dapi",
"staining",
"cell",
"staining",
"cell",
"biology",
"clinical",
"immunology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages",
"spectrum",
"analysis",
"techniques"
] |
2018
|
Murine cytomegalovirus degrades MHC class II to colonize the salivary glands
|
A hallmark of diseases of protein conformation and aging is the appearance of protein aggregates associated with cellular toxicity . We posit that the functional properties of the proteostasis network ( PN ) protect the proteome from misfolding and combat the proteotoxic events leading to cellular pathology . In this study , we have identified new components of the proteostasis network that can suppress aggregation and proteotoxicity , by performing RNA interference ( RNAi ) genetic screens for multiple unrelated conformationally challenged cytoplasmic proteins expressed in Caenorhabditis elegans . We identified 88 suppressors of polyglutamine ( polyQ ) aggregation , of which 63 modifiers also suppressed aggregation of mutant SOD1G93A . Of these , only 23 gene-modifiers suppressed aggregation and restored animal motility , revealing that aggregation and toxicity can be genetically uncoupled . Nine of these modifiers were shown to be effective in restoring the folding and function of multiple endogenous temperature-sensitive ( TS ) mutant proteins , of which five improved folding in a HSF-1–dependent manner , by inducing cytoplasmic chaperones . This triage screening strategy also identified a novel set of PN regulatory components that , by altering metabolic and RNA processing functions , establish alternate cellular environments not generally dependent on stress response activation and that are broadly protective against misfolded and aggregation-prone proteins .
Protein misfolding is an intrinsic aspect of protein biogenesis that , under optimal conditions , is kept in check by the properties of the proteostasis network ( PN ) , and upon stress , aging , and expression of aggregation-prone proteins causes cellular dysfunction that places the organism at risk for diseases of protein conformation . These are common and prominent features in Alzheimer's disease ( AD ) , Parkinson's disease ( PD ) , Amyotrophic Lateral Sclerosis ( ALS ) , polyglutamine disorders such as Huntington's disease ( HD ) , muscular dystrophies , metabolic disorders , and certain cancers [1]–[5] . Each disease is associated with its own characteristic set of aggregation-prone proteins that differ in sequence , function and expression patterns . Nevertheless , protein misfolding and aggregation have similar consequences to the cell with deleterious consequences on gene expression , protein synthesis , folding , trafficking , clearance , and cell signaling . The PN of molecular pathways coordinates protein synthesis , folding , trafficking and clearance [6]–[8] and determines the fate of proteins that do not acquire a native conformation . The chronic expression of aggregation-prone proteins , as occurs in conformational disorders , not only affects the function of proteins harboring mutations , but also challenges the stability of the PN , leading to the amplification of protein damage and persistent proteotoxicity [4] , [9]–[12] . While protein aggregates and inclusions represent a prominent feature of many human diseases , it remains unanswered whether a cell responds identically to different protein aggregates . Elucidating the mechanism ( s ) by which misfolded proteins , oligomeric species , and/or aggregates interfere with cellular function represents a prominent challenge , given the complexity of the aggregation process , and the large number of cellular processes affected as proteostasis decline is propagated across tissues [6] , [9] , [13] , [14] . Many studies have indicated that large inclusions correlate poorly with onset and severity of neurodegeneration and support a role for the soluble oligomeric species in toxicity [13]–[19] . Although intermediate species formed by distinct proteins have been suggested to display common structural motifs , it has been difficult to evaluate the contribution of different types of oligomers to toxicity [20] . Furthermore , despite the common theme of protein aggregation , growing evidence suggests that the cause of toxicity for each disease may , in part , be specific to the subset of molecular processes affected by the aggregated protein [21] . An alternative approach to understanding the origin of toxicity in each disease is to identify genetic modifiers that suppress aggregation and prevent the accumulation of metastable and misfolded proteins by enhancing global folding capacity [22] . Multiple in vitro , cell-based , and animal model systems have been developed to investigate the molecular events underlying aggregation-driven toxicity and identify modifiers of disease phenotypes [23]–[31] . While mammalian model systems are notoriously challenging to perform genome-wide screens due to the differences in genetic background and environment , screens performed in Saccharomyces cerevisiae , Caenorhabditis elegans and Drosophila melanogaster have identified processes that maintain proteome stability , promote folding and clearance . These include among others , molecular chaperones , proteasome subunits , components of the autophagy machinery , and the stress-induced transcriptional regulators DAF-16/FOXO and HSF-1 [32]–[39] . Here , we established a screening strategy in C . elegans to identify novel genetic modifiers of proteostasis that reshape the network to increase the cellular capacity for folding , prevent protein aggregation and suppress toxicity . Our approach was to identify components of the PN , that when down-regulated , enhance the functional properties of the PN to restore folding of the various folding sensors employed in the screen . This approach was designed to complement our previous efforts to identify enhancers of misfolding by screening for genes that when down-regulated caused the premature appearance of protein aggregates [35] . We identified 63 genetic modifiers that suppressed both polyQ and mutant SOD1 aggregation in muscle cells , of which 23 also suppressed associated toxicity . Of these , 9 modifiers systematically reduced the misfolding phenotypes of endogenous temperature-sensitive proteins . These modifiers were then characterized for dependence on HSF-1 activation and expression of cytosolic chaperones to enhance folding . This study introduces new proteostasis modifiers with a global effect on the stability of the muscle cell proteome , with likely broader relevance for conformational disorders .
We sought to determine whether modifier genes identified in a genome-wide RNAi screen for suppression of polyQ aggregation and toxicity in C . elegans would be efficacious on other disease-associated aggregation-prone proteins and endogenous metastable proteins . With this strategy , we tested the hypothesis of conserved modifier genes and pathways within the PN for protein misfolding and aggregation . We initiated the screening strategy with a genome-wide RNAi screen to identify genes in C . elegans that , when knocked down , suppress aggregation of expanded polyQ::YFP fusion proteins expressed in body wall muscle ( BWM ) cells [26] . For this screen , animals expressing Q35 were used , as this is a threshold length for polyQ that exhibits adult onset protein aggregation and toxicity [26] . This screen was accomplished using a semi-robotic assay developed for feeding RNAi bacteria [40] to larval 1 ( L1 , day 1 ) stage Q35 animals [35] . During early development , Q35 protein is soluble in muscle cells until animals reach day 3 of age , when aggregation is first detected , and thereafter aggregation and toxicity increase with aging ( Figure S1B , S1D ) [26] , [35] . Therefore , we selected day 6 , corresponding to three days after the onset of Q35 aggregation ( Figure S1B ) , to perform the RNAi screen to identify gene knockdowns that led to suppression of polyQ foci , relative to the empty vector ( EV ) control . RNAi against yfp was used as a control for the efficiency of RNAi gene-knockdown ( Figure 1A: VIII , XVI , XXIV; Figure 1C ) . The screen was highly robust and identified 151 genetic modifiers that suppressed Q35 aggregation ( Figure 1A , Table S1 ) . Of these modifier genes , 91 exhibited a strong suppressor effect on aggregation by reducing the number of Q35 foci by 60 to 80% in ≥75% of the RNAi-treated animals ( Table S1 , Figure 1C ) . The remaining 60 modifiers gave more variable results and were less effective as suppressors ( i . e . observed in ∼50% of the RNAi-treated animals with ∼50% reduction in foci; Table S1 ) . We next used a counter-screen with animals expressing soluble Q24 , as these animals do not exhibit aggregation or toxicity [26] , to control for phenotypic changes caused by RNAi that are not related to aggregation , such as changes in YFP fluorescence , body morphology and size , egg-laying and sterility ( Table S1 ) . We observed that none of the Q35 aggregation modifiers had any effect on Q24 protein ( Figure 1B ) , suggesting that the RNAi suppressor effect was not due to transgene silencing . The modifiers that did not meet these criteria or had deleterious consequences on animal development and viability were not studied further . Moreover , to assess whether suppression of aggregation was due to changes in polyQ expression , we examined mRNA and steady-state protein levels for a representative group of modifiers . We quantified the levels of q35::yfp mRNA by rt-PCR ( Figure S2 ) and the levels of Q35::YFP protein by SDS-PAGE and western blot analysis ( Figure 2A ) . The results show that , for the RNAi modifiers tested , suppression of aggregation occurred without affecting the polyQ mRNA or protein levels . To obtain evidence that the suppression of visible Q35 aggregates corresponds to the appearance of soluble Q35 protein ( Figure 1A ) , we used the dynamic imaging method of Fluorescence Recovery After Photobleaching ( FRAP ) . Inclusion-localized Q35 corresponds to an immobile state with very limited fluorescence recovery following photobleaching ( Figure 2B and 2C: Q35 control ) [23] , [35] , whereas the fluorescence of diffuse-looking Q35 in animals fed with modifier RNAi recovered immediately , consistent with a diffuse and soluble state analogous to soluble Q24 ( Figure 2B hmg-3 and Figure 2C ) . These results provide biophysical evidence for Q35 solubility identified by the visual screening . We further examined the biochemical properties of Q35 in total protein extracts , for a representative group of modifiers . We found that the amount of aggregated polyQ protein detected using native PAGE analysis was reduced , and that the levels of soluble and diffuse species were increased ( Figure 2D ) . Taken together , these results reveal that RNAi knockdown of specific modifier genes suppressed polyQ aggregation by maintaining the protein in a mobile soluble state . The identity of the RNAi-targeted genes was verified by sequencing of the dsRNA plasmids , followed by Blast analysis in NCBI and Wormbase databases . The Q35 aggregation modifier genes are 88% conserved , with predicted human orthologs , and can be grouped into seven functional categories of cell cycle , DNA synthesis and repair; RNA synthesis and processing; protein synthesis; protein folding and turnover; cell structure and protein trafficking; signaling; and energy and metabolism ( Table S1 , Figure S3A , S3B ) . The fraction of modifiers represented in each functional class is significantly distinct from their representation in the C . elegans RNAi library ( Figure S3A , S3B ) [41] , indicating enrichment for cellular processes important for proteostasis . We next asked whether the Q35 modifiers were effective on another polyQ model as a way to distinguish the most robust polyQ aggregation suppressors . This was done by screening a transgenic line expressing Q37::YFP . These animals exhibit a more rapid onset of aggregation relative to Q35 animals , between day 2 and 3 of age , together with a more rapid decline in motility ( Figure S1A , S1B , S1D ) . These phenotypes are dependent solely upon the CAG-repeat length as the levels of Q35 and Q37 are identical ( Figure S1C ) . Q37 animals were fed RNAi from L1 stage ( day 1 ) and aggregation was examined on day 5 , corresponding to three days post-aggregation onset ( Figure 3A , 3B ) . Of the initial 151 modifiers of Q35 aggregation , only 88 of these also suppressed Q37 aggregation , of which 81 corresponded to the strongest Q35 suppressors ( Table S1 ) . We designated the set of common modifiers of Q35 and Q37 aggregation as Class A strong modifiers ( Table 1 ) , and the remaining 63 genes as Class B weak modifiers ( Figure 3C ) . We next tested whether the genetic modifiers of polyQ aggregation would be effective on yet another model that expresses the mutant human SOD1G93A . This model shows aggregation onset during embryonic development and a distinctive pattern of foci that persists throughout adulthood ( Figure S1E , S1F ) [12] . Of the 88 polyQ aggregation modifiers , 63 also suppressed mutant SOD1G93A aggregation in 5 day old animals ( Figure 4A , Table S1 ) , without causing phenotypic changes in SOD1wt animals ( not shown ) . A subset of these common aggregation suppressors were also examined by SDS-PAGE and western blot analysis , and shown not to reduce steady-state protein levels of SOD1G93A ( Figure 4B ) . 95% of the mutant SODG93A suppressors belong to the polyQ Class A modifiers ( Table S1 ) . Unlike Q35 aggregation that is only detected in young adult animals , mutant SODG93A aggregation occurs in embryos [12] , and yet many Class A modifiers were effective suppressors of SOD1 aggregation providing additional support that Class A modifiers are robust modifiers of protein folding . Moreover , these modifiers , that are common to polyQ and SOD1 , exhibit a similar overall distribution into functional classes ( Figure S3C ) as was described for modifiers of polyQ aggregation ( Figure S3B ) , thus identifying new modifier pathways that are common to protein aggregation . We propose that these new modifiers can either suppress aggregation directly by affecting cellular processes that mediate aggregate formation , or indirectly by altering some aspect of the PN that confers a protective action that increases folding . To distinguish between these two possibilities , we employed genetic tests to determine which modifiers reflect an improvement of the folding environment , by reducing misfolding and associated toxicity . Protein aggregation is a common feature of many diseases; however , the relationship between aggregation and cellular toxicity remains controversial . The appearance of aggregates and inclusions has been linked both to cellular dysgenesis and toxicity , as well as protection from toxicity [13]–[19] . Therefore , we took advantage of an unbiased genetic approach to test the relationship between suppression of aggregation and toxicity . Because the initial genetic screens were based solely on aggregation phenotypes , we were able to subsequently perform cellular toxicity assays to assess this relationship . Relative to wt animals or animals expressing soluble polyQ ( Q24 ) , Q35 animals exhibit muscle dysfunction resulting in a 40% loss of motility at 6 days of age ( Figure S1D ) [26] . Therefore , we quantified the motility of RNAi-treated Q35 animals as a measure of polyQ-associated cellular toxicity , using an automated worm tracker system analysis , validated by manual methods . As a reference positive control for toxicity suppression , we show that motility was restored to near wt levels by knockdown of the Q35 transgene expression with yfp-RNAi ( Figure 5A ) . All 88 Class A modifiers were tested for effects on the motility of Q35 animals and wt control animals ( Table S2 ) . Because we sought to identify improvement of motility directly associated to suppression of Q35 aggregation , we excluded any modifier that , alone , had effects on the motility of wt animals . Of the 88 modifiers tested in wt animals , 33 gene knockdowns ( 37% ) affected the motility of wt animals ( Figure 5B ) , and were excluded from further analysis . For the remaining 55 modifiers , 42% improved the motility of Q35 animals to wt levels , 36% had no effect , and 22% enhanced the toxicity of Q35 ( Figure 5A ) . These results revealed that suppression of aggregation , as detected by visual , biophysical , and biochemical measures , does not necessarily predict that the physiological health of the cell will be restored . The genetic uncoupling between aggregation and toxicity further reinforces previous similar observations [13] , [14] , [19] . Taken together , the toxicity in diseases of protein conformation is the outcome of a complex series of misfolding events , involving multiple species and aberrant interactions within the cell . Among the challenges with studies of protein misfolding and aggregation have been the concerns with the physiological imbalance associated with overexpression of heterologous proteins in the respective model systems . To circumvent this concern , we asked whether the PN modifiers that suppressed polyQ and mutant SOD1 aggregation ( Figure 5C ) would also restore the folding of endogenous metastable proteins harboring temperature sensitive ( TS ) mutations . TS-mutations represent an important class of highly sensitive folding sensors that are expressed at normal endogenous levels and have quantifiable phenotypes when properly folded at the permissive condition or misfolded at the restrictive temperature [10] , [42] , [43] . This strategy was also used to distinguish between modifiers that directly perturb and suppress the formation of protein aggregates , from the modifiers that reshape the PN to improve the protein-folding environment . We examined the properties of four TS mutant proteins corresponding to the paramyosin ortholog UNC-15 , the basement-membrane protein perlecan UNC-52 , the myosin-assembly protein UNC-45 , and the myosin heavy chain UNC-54 [10] . At the permissive temperature ( 15°C ) , each of these TS-proteins is known to be fully functional and animals harboring these mutants exhibit a wt phenotype , whereas at the restrictive temperature ( 23°C or 25°C ) these TS-proteins misfold and cause muscle dysfunction that can be measured as slow movement and paralysis ( UNC-15 , UNC-54 , Figure 6A , 6C ) , egg-laying defects leading to swelling and paralysis ( UNC-45 , Figure 6B ) , and stiff-paralysis ( UNC-52 , Figure 6D ) ( see Materials and Methods ) [10] , [42] , [44] . These phenotypes are specific to animals expressing the TS mutations , and are not observed in wt animals . We tested all 23 RNAi modifiers that suppressed both Q35 aggregation and toxicity ( Figure 5C ) , on each of the TS strains , and found that a total of nine modifiers reduced the number of animals displaying TS phenotypes by 40% to 90% at 23°C , a slightly lower restrictive temperature at which the efficiency of the RNAi protocol was maintained ( Figure 6A–6D , Table 2 ) . These results suggest that while protein misfolding is common to all three classes of folding sensors ( polyQ , mutant SOD1 , and TS-mutant proteins ) , the cellular environment and the PN are influenced differentially by the modifiers , as reflected by the effects on these sensors . We propose that these final nine modifiers ( Figure 7A ) are core PN modulators that confer improvement of cellular folding capacity in C . elegans muscle cells . Upstream of the core components of the PN is the master cytosolic stress-responsive pathway that leads to HSF-1 activation and expression of molecular chaperones for stability of the proteome . To examine whether the nine PN modifiers ( Figure 7A ) lead to HSF-1 activation as a general mechanism for proteostasis improvement , we introduced a hypomorphic mutation of hsf-1 ( sy441 ) into the background of the polyQ strain Q37 and knocked-down each modifier gene ( Figure 7B ) . Our results show that: suppression of aggregation by ucr-2 . 3 , gei-11 and C45G3 . 4 was completely dependent on HSF-1; whereas T22D11 . 5 , ZK430 . 7 , Y110A7A . 8 and R05D11 . 4 exhibited a weaker dependence on HSF-1; and F43G9 . 1 and let-607 were independent of HSF-1 . We next examined whether chaperone gene expression was affected downstream of the nine PN modifiers , by monitoring the expression of Hsp70 ( C12C8 . 1 , F44E5 . 4 ) and small Hsp ( hsp-16 . 1 ) , upon knockdown of each genetic modifier ( Figure 7C ) . We show that five of nine PN modifiers induced expression of cytoplasmic chaperones . Knockdown of let-607 ( ER-UPR , Table 2 ) had the strongest effect , suggesting an important regulatory crosstalk between cytoplasmic and ER stress response pathways . Knockdown of the TCA cycle component T22B11 . 5 led to upregulation of chaperones and establishes a link between folding and metabolic state , whereas knockdown of gei-11 ( putative negative regulator of cholinergic signal , Table 2 ) induction of hsp suggests an effect of cholinergic signaling on muscle homeostasis ( Figure 7C ) . Reduction of R05D11 . 4 ( translation ) leading to induction of hsp-70 was consistent with an effect of protein synthesis on folding machinery . For the remaining four PN modifiers , knockdown of ZK430 . 7 ( RNA processing ) had a modest effect on hsp expression , while knockdown of Y110A7A . 8 ( splicing ) , ucr-2 . 3 ( ETC ) and F43G9 . 1 ( TCA cycle ) had no effects on hsp levels ( Figure 7C , Table 2 ) . This suggests that proteostasis was restored through other pathways involving reduced metabolism and energy production . Taken together , these results demonstrate that improvement of the cellular folding environment by these novel proteostasis modulators is not simply a consequence of a generalized induction of the heat shock response and molecular chaperones but also involves other PN pathways , not previously linked to proteome surveillance .
The genome-wide screen for suppression of polyQ aggregation identified a collection of modifier genes from distinct cellular functional classes ( Table S1 , Figure S3B ) . Modifiers in the category cell structure and protein transport , include cytoskeleton components ( filamin ) and matrix proteins ( ppn-1 , mua-3 , gon-1 ) , supporting observations that aggregation can be affected by disturbing the integrity of cell structure [10] . Other genes in this group encode motor proteins involved in vesicular trafficking ( klp-15 , nex-1 ) , consistent with a role for protein movement and transport in aggregation . Cell growth and replication modifiers are involved in progression through the cell cycle ( cyclin-dependent kinases ) , and DNA replication and recombination ( transposases ) , with a likely general effect on growth rate and development . Energy and metabolism modifier genes are involved in energy production and mitochondrial electron transport chain ( ETC ) function . Restriction in energy levels not only affects overall protein biogenesis , as it is a highly ATP-dependent process , but also , metabolic enzymes can influence protein folding in the cell by altering the levels of organic/inorganic solutes with effects in polypeptide chain solvation [6] . Notably , reduced ATP synthesis , ETC activity or mitochondrial function have been shown to enhance lifespan , possibly by delaying age-dependent decline in protein folding capacity and by upregulation of stress-response pathways that promote proteostasis and survival [45]–[48] . Gene expression and protein synthesis related modifiers function in RNA metabolism , ribosome biogenesis , and protein synthesis . This is consistent with reduced translation increasing C . elegans lifespan , perhaps by activating a physiological state with increased stress resistance and folding capacities [49]–[51] . Post-translational control modifier genes are involved in chaperone-assisted folding , such as HSP70 superfamily members , DNAJ co-chaperones and cyclophilins , and post-translation modifying enzymes such as SUMO and E3-ubiquitin ligases . The role of chaperones on protein solubility , misfolding and aggregation has been well established [52] , [53] , and an imbalance in certain co-chaperones has been suggested to alter chaperone activity in the cell and folding [54] . Signaling RNAi-targeted genes included nuclear hormone receptors , G-protein-coupled receptors , C-type lectins ( endocytic receptors ) , or calcium export and channel-transport activity . These regulators affect reproduction , growth , morphogenesis ( development ) , and locomotion by altering signaling pathways involved in neuronal and muscle function . Information retrieved from comparative analysis of genetic screens of different misfolded proteins can provide important insights to identify both common and protein-specific pathways for conformational disorders ( Table S3 ) . Highly relevant to this point is our ability to compare the modifier genes identified in this study with a previous complementary screen using the same threshold Q-length properties of the Q35 model to identify genes that when knocked-down by RNAi led to premature onset of polyQ aggregation [35] . Together , these genome-wide screens identified 341 genetic modifiers that cluster into the same functional classes and pathways , but correspond to distinct genes within these pathways . Specific modifier genes may interfere with the misfolded species at different stages of the aggregation process with opposite outcomes , consistent with the functional properties of a network , where different components within a process can shift the equilibrium in opposing directions to alter the stability of the proteome , to intensify or suppress the polyQ phenotype . While it might seem counter-intuitive that molecular chaperones could suppress polyQ aggregation when knocked-down , this is consistent with observations that proteome stability can be enhanced or suppressed by changing the composition of the cellular chaperome [52]–[55] . For example , reducing the expression of cyn-11 and cyn-12 ( cyclophilin D isoforms ) and dnj-5 , that function primarily as co-chaperones to regulate Hsp70 and Hsp90 activities [56] , promotes a polyQ soluble state . These co-chaperones could function as negative regulators of chaperone function , and their down-regulation results in enhanced chaperone activities leading to suppression of misfolding [52] , [57] , [58] . This would be consistent with evidence that Hsp70 folding activity is negatively regulated by co-chaperones and co-factors such as CHIP [59] and BAG-1 [60] . Therefore , enhancement of folding can be achieved by both positive and negative regulation of chaperones or by a compensatory response that up-regulates other chaperones . Another functional class common to both Q35 screens is the protein trafficking and cell matrix . Suppression of aggregation by knockdown of cell cytoskeleton proteins , such as intermediate filaments ( MUA-3 and MUA-6 ) and filamin , is supported by experimental evidence that the dynamics of aggregation rely in part on translocation of proteins into inclusions . For example , the active transport of Htt-exon1 along microtubules has been shown to be required for inclusion body formation [61] . In contrast , premature aggregation was observed when the expression of vesicle proteins involved in protein trafficking was knocked-down [35] , including TFG-1 COP-II complex , APT-3 and APT-1 , and cell membrane assembly proteins ( SNAP-25 ) . Interference with these processes can disturb essential steps of the folding and secretory pathways , increase the load of misfolded proteins in the cell and lead to premature polyQ aggregation . An important observation from these studies is that the Class A PN modifiers ( Figure 3C , Table 1 ) were highly effective to suppress polyQ aggregation , and yet only 42% of these modifiers also reduced toxicity , with approximately equal numbers of modifier genes with either no effect on toxicity or even enhancing toxicity ( Figure 5A ) . These results provide independent evidence that suppression of aggregation alone does not predict that the physiological health of the cell will be restored . From a mechanistic perspective , it is increasingly clear that a series of events are associated with the conversion of the nascent polyQ protein into different oligomeric states , immobile aggregate species and inclusion bodies [62]–[65] . We conclude that the genetic suppression of aggregation can occur via a wide range of mechanisms that are dissociated from the effect on toxicity , consistent with previous observations that interference with the aggregation process could in some cases enhance the formation of “toxic oligomeric species” [13]–[19] , [66]–[68] . Moreover , each modifier gene is certain to function within its own network of interacting partners ( Table 2 ) , revealing an expanding network through which each modifier can suppress aggregation , but with differential effects on toxicity depending on the affected cellular function . The demonstration that protein aggregates are uncoupled from cellular toxicity has implications for the understanding of the PN and for development of therapeutics . We took advantage of C . elegans models of protein aggregation-toxicity in addition to folding sensor strains harboring TS mutations in endogenous proteins to identify a core group of modulators that improve folding in muscle cells ( Figure 7A ) . Our results highlight important aspects of the PN , relevant for both ‘gain-of-toxic function’ by aggregation-prone proteins , and ‘loss-of-function’-derived toxicity due to protein misfolding . The nine PN modifiers that remained at the end of the screening tree function in the mitochondrial respiratory chain and TCA cycle , that regulate metabolism , energy balance and prevention of oxidative stress , in addition to rRNA processing and transcription , that determine gene expression and proteome load ( Table 2 ) . While these modifier genes , upon initial inspection , seem not to be directly involved in folding , perturbation of their specific functions and networks of interactions re-adjusts the PN to enhance its capacity , as suggested by activation of the heat shock response and chaperone expression ( Figure 7B , 7C ) . Moreover , comparison to other genetic screens performed in Drosophila [32] , [33] , [39] , C . elegans [34]– and yeast [38] , [71] ( Table S3 ) provides additional insights into the regulation of the PN by these modifiers . ucr-2 . 3 encodes an ubiquinol-cytochrome c reductase subunit of the mitochondrial respiratory chain , and F43G9 . 1 and T22B11 . 5 encode TCA-cycle enzymes ( isocitrate dehydrogenase and 2-oxoglutarate dehydrogenase , respectively ) ( Table 2 ) . Disruption of the respiratory chain and energy production has been suggested to have consequences on cellular homeostasis [47] , [48] . Intriguingly , knockdown of ucr-2 . 3 was shown to enhance the toxicity of human Tau expressed in C . elegans neuronal cells ( Table S3 ) [34] . This discrepancy may be related to Tau-specific proteotoxicity , not derived from aggregation or misfolding , but associated with microtubule binding and disruption . F43G9 . 1 was also identified as a enhancer of lifespan [70] , which is consistent with a role in proteostasis . ZK430 . 7 encodes an rRNA processing factor , Y110A7A . 8 is a putative mRNA splicing factor , and R05D11 . 4 encodes an RNA helicase required for translation ( Table 2 ) . Thus , perturbing components of the gene expression machinery can enhance proteostasis , likely by altering the expression load of unstable proteins and activating stress responses to restore proteostasis . In particular , knockdown of Y110A7A . 8 activates the osmotic stress response [69] and causes premature onset of polyQ aggregation on 3 day old animals [35] , consistent with an increase in misfolding ( Table S3 ) . However , on 6 day old animals we show that the number of aggregates is suppressed , which suggests a time-dependent response by the PN to enhance the folding machinery and restore folding . gei-11 encodes a Myb-family transcription factor proposed to regulate cholinergic receptor function at the BWM cells [72] , which affects muscle function and homeostasis [43] , [73]; and let-607 encodes a CREBH ortholog transcription factor predicted to be a component of the C . elegans ER stress response . The role of let-607 is particularly intriguing as it reveals a genetic crosstalk between the cytoplasmic and ER lumen stress pathways . Knockdown of let-607 induces chaperone expression that is not dependent upon HSF-1 ( Figure 7B , 7C ) , suggesting that other stress responses such as the ER unfolded protein response may be involved in the suppression of cytosolic protein aggregation . Taken together , these results emphasize that diverse genetic and cellular mechanisms can restore cellular proteostasis beyond the traditional heat shock response . These nine gene modifiers of BWM protein homeostasis represent core components of the PN that evoke a robust and effective improvement of disease-related and endogenous metastable protein folding . Identification of these processes is a fundamental step towards identifying new components that constitute the network , and the cellular and organismal mechanisms by which they contribute to protein homeostasis and protect against chronic expression of misfolded toxic proteins .
Animals were maintained according to standard methods , at 20°C on nematode growth media ( NGM ) with OP50 E . coli [74] . The strains utilized in this work , and previously described , are the following: wild-type ( wt ) Bristol strain N2; polyQ strains Q0 AM134 ( rmIs126[Punc-54::yfp] ) , Q24 AM138 ( rmIs130[Punc-54::q24::yfp]II ) , Q35 AM140 ( rmIs132[Punc-54::q35::yfp]I ) , Q37 AM470 ( rmIs225[Punc-54::q37::yfp]II ) ( Text S1 ) [26] , [35]; human SOD1 strains SOD1G93A AM265 ( rmIs177[Punc-54::sod1G93A::yfp] ) and SOD1WT AM263 ( rmIs175[Punc-54::sod1wt::yfp] ) [12]; temperature sensitive ( TS ) mutant strains CB1402 [unc-15 ( e1402 ) ] , CB1157 [unc-54 ( e1157 ) ] , HE250 [unc-52 ( e669su250 ) ] and CB286 [unc-45 ( e286 ) ] [10] . The transgenic polyQ and SOD1 strains had been integrated by gamma-irradiation , 5 times backcrossed , and were previously described [12] , [26] . The strain Q37;hsf-1 ( sy441 ) was generated by genetic cross of the original strains AM470 and PS3551[hsf-1 ( sy441 ) ]I . The genome-wide RNAi screen for suppression of aggregation in C . elegans muscle cells was performed using the commercial RNAi library , with bacteria expressing dsRNA for 87% of the predicted C . elegans genes ( GeneService , USA ) [35] , [40] . A semi-automated high throughput setup system was used , consisting of a robotic device ( Biomek FX Liquid Handler , Beckman Coulter , USA ) programmed to add bacteria and age-synchronized animals in liquid culture to 96-well plates . RNAi bacterial cultures were grown for approximately 8 h in LB-ampicillin 50 µg/ml ( 65 µl ) , at 37°C with continuous shaking at 315 rpm ( Orbital shaker , GeneMachines HiGro , Genomic Solutions , USA ) , and induced with 0 . 5 mM isopropyl β-D-thiogalatoside ( IPTG , Sigma ) for 3 h at 37°C . To obtain an age synchronized population of L1 larvae ( first larval state post egg hatching ) , Q35 gravid adults were bleached with a NaOCl solution [250 mM NaOH and 1∶4 ( v/v ) dilution of commercial bleach] and the eggs hatched in M9 buffer overnight at 20°C . Day 1 is defined as 18 h following NaOCl age-synchronization and animals are said to be 1 day old ( L1 stage ) . 10 to 15 animals were added to each well in the 96-well plate in a volume of 50 µl of M9 plus [M9 , 1 µg/ml cholesterol , 50 µg/ml ampicillin , 10 µg/ml tetracycline , 0 . 1 µg/ml fungizone and 170 µg/ml IPTG] and incubated at 20°C with continuous shaking at 200 rpm ( Innova 4430 Incubator Shaker , New Brunswick , USA ) . Animals were scored 5 days later ( 6 days old ) for reduction in the number of fluorescent foci using the stereomicroscope Leica MZ16FA equipped for epifluorescence ( Leica Microsystems , Switzerland ) . As a negative control , animals were fed bacteria carrying the L4440 empty vector ( EV ) . Suppression of aggregation was scored positive when more than 50% of the animals had a 50% or higher reduction in foci number relative to the EV control , without loss of YFP fluorescence , changes in growth rate or development of the animals . The candidate positive hits were re-screened ( n≥3 ) , then tested in the Q24 soluble control strain , and counter screened in Q37 animals ( 5 days old ) and SOD1G93A animals ( 5 days old ) . In Q37 and SOD1G93A animals , suppression of aggregation was scored positive when more than 50% of the animals showed a reduction in foci number ( >25% ) . RNAi was always added on day 1 . The identity of the RNAi-targeted genes was verified by sequencing of the dsRNA plasmids , followed by Blast analysis in the NCBI and Wormbase databases revealing high specificity of genomic sequence targeting . Gene-knockdown by the respective RNAi was also confirmed for a representative group of hits by rtPCR ( data not shown ) . For RNAi assays on plates ( for foci scoring , FRAP and motility analysis , to collect animals for western blot and real-time qPCR , and for TS assays ) , NGM media was supplemented with 100 µg/ml ampicillin , 1 mM IPTG and 12 µg/ml tetracycline ( Sigma ) , and seeded with overnight ( 16 h ) RNAi bacteria cultures , pre-induced with IPTG ( 1 mM , 3 h ) . One day old ( L1 ) animals ( 15 to 20 animals ) were transferred onto NGM-RNAi bacteria seeded plates and grown at 20°C , and at the time indicated aggregation was scored in at least 50 animals , for each condition ( n = 3 ) . Aggregates were defined as discrete , bright foci that can be distinguished from their surrounding fluorescence by increased brightness intensity . The detection limit for these foci , measured with the higher resolution Zeiss Axiovert 200 microscope , is in the order of 3 µm in length ( for elongated foci in Q35 ) and ∼7 µm2 in area ( for round foci ) , with the microscopy tools and fluorescence exposure utilized in the genetic screen ( Leica MZ16FA ) . Data collected from different experiments was compiled to calculate aggregate number averages relative to the control in EV RNAi . Fluorescent microscopy images were taken using an Axiovert 200 microscope with a Hamamatsu digital camera C4742-98 ( Carl Zeiss , Germany ) . All assays were performed blind as to the identity of the RNAi by attributing to each modifier a number corresponding to a well with the dsRNA bacterial stock , in a 96-well plate . To examine the biophysical properties of polyQ protein , animals were subjected to FRAP analysis . Animals were mounted on a 3% ( w/v ) agar pad on a glass slide and immobilized in 2 mM levamisole . FRAP was measured using the Zeiss LSM510 confocal microscope ( Carl Zeiss , Germany ) , and the 63× objective lens at 5× zoom power , with the 514 nm line for excitation . An area of 0 . 623 µm2 was bleached for 35 iterations at 100% transmission , after which time an image was collected every 123 . 35 ms . Relative fluorescence intensity ( RFI ) was determined as previously described [43] , [75] . For SDS-PAGE analysis , 6 day old animals grown on RNAi-seeded NGM plates were collected and resuspended in PELE buffer [20 mM Tris pH7 . 4 , 10% glycerol , 2% Triton X-100 , 0 . 5 mM PMSF , 1 µg/ml leupeptin , 1 µg/ml pepstatin , 1 mM EDTA , 1 mM DTT , protease inhibitor cocktail tablet ( Roche Diagnostics #11836170001 ) ] . Lysis of ∼100 animals was accomplished by a combination of 4 cycles of freeze-thaw , grinding with a motorized pestle ( Kontes #749541-000 and #749520-0000 ) , followed by 8 min sonication ( Sonicator Bath Branson 1510 , Branson ) . To dissolve the polyQ aggregates , SDS was added to a final concentration of 5 . 5% ( v/v ) and samples were boiled for a total of 10 min . Total protein concentration was determined using the Bradford assay ( Bio-Rad #500-0006 ) . 15 µg ( for Q35 ) or 20 µg ( for SOD1 ) of total protein , in the linear range for YFP detection [43] , were analyzed on a 10% SDS-PAGE followed by Western blotting . For YFP ( polyQ and SOD1 ) detection , blots were probed with the anti-GFP IR800 conjugated antibody ( 1∶5 , 000 dilution; Rockland Immunochemicals #600-132-215 ) . For α-tubulin detection , blots were probed with the anti-α-tubulin primary antibody ( 1;4 , 000 dilution; Sigma #T-5168 ) followed by the secondary antibody Alexa Fluor 680 goat anti-mouse IgG ( 1∶10 , 000 dilution; Molecular probes #A-21057 ) . Antibody binding was detected with the Odyssey Infrared Imaging System ( LI-COR Biosciences , USA ) . The ratio between band intensities YFP/α-tubulin was calculated for each sample ( Adobe Photoshop 7 . 0 , arbitrary units ) and compared to the EV control ( relative % ) . A representative group of modifiers was tested ( 3 biological replicates ) . Statistically significant changes in protein amounts were considered if p<0 . 05 ( Student's T-test ) . For native PAGE analysis , animals ( ∼100 ) were collected with M9 buffer and resuspended in native-lysis buffer [50 mM Tris pH7 . 4 , 5 mM MgCl2 , 0 . 5% Triton X-100 , 0 . 2 mM PMSF , 1 µg/ml leupeptin , protease inhibitor cocktail tablet ( Roche Diagnostics #11836170001 ) ] . Lysis was achieved with 4 cycles of freeze-thaw , and homogenization by grinding with the motorized pestle , always maintaining the tubes on ice . Total protein concentration was determined as before and 40 µg were analyzed on a 5% native PAGE ( at 4°C ) , followed by gel scan ( STORM 860 , #91393 , GE Healthcare , UK ) . This experiment was done in triplicate . Animals ( 6 days old ) grown on RNAi NGM plates at 20°C were picked ( 20–25 animals ) onto the center of a NGM OP50-seeded plate ( full surface area covered with OP50 ) , equilibrated at 20°C . Animals' movements were digitally recorded using a Leica M205 FA microscope with a Hamamatsu digital camera C10600-10B ( Orca-R2 , Leica Microsystems , Switzerland ) , and the Hamamatsu Simple PCI Imaging software . Videos of 45 s were recorded at 2×2 binning and 5 frames per second , and captured frames were merged into * . avi format and imported directly into ImageJ . Using the LOCI bio-formats plugin and a custom stack de-flicker plugin ( http://www . loci . wisc . edu/bio-formats/imagej ) , light average intensity was normalized for each frame . To enhance the definition of the animals in the movies , the difference between each frame and the constant background was calculated , using the ‘Maximum Z-stack’ projection . The resulting movie was converted to binary format using Otsu Thresholding 2 . Binary objects representing the animals were tracked using custom ImageJ plugin , wrMTrck ( based on “MTrack2” by Nico Stuurman [76] ) . The average speed of each animal was calculated by dividing the length of each track ( corrected for animal body length ) by the duration of the track ( body length per second , or BLPS ) . The wrMTrck plugin and scripts for automated analysis are open-source and publicly available at http://www . phage . dk/plugins . Videos were recorded for a minimum of 75 animals per experiment ( n≥3 ) and motility measurements are given as a percentage of wt motility ( % wt in EV RNAi ) . RNAi modifiers that affected the motility of wt animals were removed from further analysis . All motility assays were also performed blind as to the identity of the RNAi gene-target . Measurements of motility were validated by other read-outs that included manual-based motility assays [10] , [77] . The first manual assay measured how fast it took animals placed in the center of a ring ( circumference only ) of OP50 bacteria to reach the food , and the second manual assay monitored the number of worms that traveled 1 cm in 1 minute on OP50 bacteria-seeded NGM plates . All results shown were obtained with the automated worm tracker , which provides reproducible and unbiased results . Temperature sensitive ( TS ) mutant animals were age-synchronized to L1 stage by NaOCl bleaching , grown on RNAi-seeded NGM plates ( 15–20 animals per plate ) from day 1 at a sensitized temperature of 23°C ( to maintain the RNAi suppressor effect on aggregation , which was used as a control ) and scored for phenotypes on day 5 . For the 25°C restrictive temperature control experiment , L1 nematodes were grown on EV RNAi at 15°C until L4 stage to avoid embryonic and developmental phenotypes , then transferred to 25°C and scored 2 days later for the same phenotypes . For the 15°C permissive temperature control experiment , animals were synchronized to L1 , added to EV RNAi plates , grown at 15°C and scored for phenotypes on day 6 ( to account for slower but normal development at this temperature ) . At least 50 animals were scored for each specific phenotype , per experiment ( n = 3 ) , as described previously [10] , [42] , [43] , and all assays were performed blind . For the slow movement/paralysis assay [unc-15 ( e1402 ) and unc-54 ( e1157 ) ] , 15–20 animals were placed on a OP50-NGM plate at room temperature in the center of a 1 cm circle ( drawn on the bottom of the plate ) . Animals remaining in the 1 cm circle after 5 min were considered to possess a slow movement or paralyzed phenotype . To score for stiff paralysis [unc-52 ( e669su250 ) ] , partially paralyzed animals with moving heads and stick-like bodies were scored . For the egg-laying phenotype [unc-45 ( e286 ) ] partially paralyzed animals with a large belly of accumulated eggs were scored . Wt animals ( 5 days old and ∼50 ) were collected from RNAi-NGM plates and RNA was extracted with the Trizol reagent ( Invitrogen ) , followed by DNase treatment ( Applied Biosystems #AM1906 ) . mRNA was then reverse transcribed using the iScript cDNA Synthesis Kit ( Bio-Rad #170-8891 ) . 10 ng of cDNA were used for real-time PCR amplification using the iQ SYBR Green Supermix ( Bio-Rad #170-8880 ) and the iCycler system ( Bio-Rad ) ( see Text S1 ) . The relative expression levels of each gene were determined using the Comparative CT Method ( Real-Time PCR Applications Guide , Bio-Rad ) . Gene expression levels were normalized relative to actin ( act-1 ) in the same sample ( internal control ) , and then relative to the levels of the same gene in EV control sample . Measurements were performed for ≥3 biological samples for each condition .
|
A common characteristic of protein conformational diseases is the appearance of protein aggregates associated with late-onset symptoms . Here , we have taken an unbiased genetic approach to test the hypothesis that protein aggregation and toxicity are co-linked genetic traits that are regulated by a common proteostasis network . To address this , we took advantage of the tractable genetic model Caenorhabditis elegans expressing expanded polyglutamines ( polyQ ) and performed a genome-wide RNA interference ( RNAi ) screen to identify genes that altered the proteostasis environment and suppressed aggregation and toxicity . These modifiers were subsequently tested on animals expressing mutant SOD1 and on animals expressing endogenous proteins with temperature-sensitive mutations . This screening triage resulted in the identification of nine genes with effects on protein folding , corresponding to new proteostasis pathways involved in metabolism and RNA processing functions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"cellular",
"stress",
"responses",
"molecular",
"cell",
"biology",
"cell",
"biology",
"genetic",
"screens",
"genetics",
"gene",
"classes",
"molecular",
"genetics",
"biology",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics"
] |
2011
|
A Genetic Screening Strategy Identifies Novel Regulators of the Proteostasis Network
|
Skin pigmentation is one of the most variable phenotypic traits in humans . A non-synonymous substitution ( rs1426654 ) in the third exon of SLC24A5 accounts for lighter skin in Europeans but not in East Asians . A previous genome-wide association study carried out in a heterogeneous sample of UK immigrants of South Asian descent suggested that this gene also contributes significantly to skin pigmentation variation among South Asians . In the present study , we have quantitatively assessed skin pigmentation for a largely homogeneous cohort of 1228 individuals from the Southern region of the Indian subcontinent . Our data confirm significant association of rs1426654 SNP with skin pigmentation , explaining about 27% of total phenotypic variation in the cohort studied . Our extensive survey of the polymorphism in 1573 individuals from 54 ethnic populations across the Indian subcontinent reveals wide presence of the derived-A allele , although the frequencies vary substantially among populations . We also show that the geospatial pattern of this allele is complex , but most importantly , reflects strong influence of language , geography and demographic history of the populations . Sequencing 11 . 74 kb of SLC24A5 in 95 individuals worldwide reveals that the rs1426654-A alleles in South Asian and West Eurasian populations are monophyletic and occur on the background of a common haplotype that is characterized by low genetic diversity . We date the coalescence of the light skin associated allele at 22–28 KYA . Both our sequence and genome-wide genotype data confirm that this gene has been a target for positive selection among Europeans . However , the latter also shows additional evidence of selection in populations of the Middle East , Central Asia , Pakistan and North India but not in South India .
Human skin color varies widely among and within populations and is a classic example of adaptive evolution . Skin pigmentation in humans is largely determined by the quantity and distribution of the pigment melanin , which is packed in melanosomes and then transferred from melanocytes ( melanin-forming cells ) to the surrounding epidermal keratinocytes [1] . Human melanin is primarily composed of two distinct polymers: eumelanin ( brown/black ) and pheomelanin ( yellow/red ) , which differ in their physical properties and chemical composition [2] . In addition to the amount and type of melanin , other factors such as the size , shape , number , and cellular distribution of melanosomes also contribute to the variation in skin color [3] . Comparative studies of model organisms , pigmentation disorders and genome-wide studies have played a key role in the identification of human pigmentation genes [4]–[7] . A total of 378 candidate loci , including 171 cloned genes , are currently recorded in the Color Genes database ( http://www . espcr . org/micemut/ ) , yet only a few of them have been confirmed to have potentially function-altering polymorphisms in humans . A significant correlation between skin color and ultraviolet radiation ( UVR ) levels observed at the global scale suggests that natural selection plays an important role in determining the distribution of this phenotypic trait [8] . The evolution of dark skin at low latitudes has been mainly accredited to the requirement of photo-protection against UVR which causes sunburn and skin cancer , whereas the evolution of light skin has been most commonly associated with vitamin D deficiency [9] , [10] . It has been proposed that as humans started to colonize higher latitudes , where UVR levels were lower , dark skin could not absorb sufficient UVR for efficient vitamin D synthesis , hence natural selection favored the evolution of light skin [8] , [11] . This is indirectly supported by the observation that candidate pigmentation genes are collectively enriched by high-FST single-nucleotide polymorphisms ( SNP ) [12]–[14] . Furthermore , data mining of publicly available datasets , such as HapMap , Perlegen and Human Genome Diversity Project ( HGDP ) , has provided evidence of selection signals in pigmentation-related genes in one or more populations ( see [15] and references therein ) , [16] thus elucidating the history of human adaptation to local environments for this complex trait . One of the key pigmentation genes in humans is SLC24A5 ( OMIM 609802 ) . It is located on chromosome 15q21 . 1 and encodes a protein called NCKX5 . The association of this gene with lighter pigmentation was initially discovered in zebrafish [4] . Using admixed populations , it was further demonstrated in this study [4] that a non-synonymous variant ( ref SNP ID: rs1426654 ) in the third exon of this gene explains 25–38% of the skin color variation between Europeans and West Africans . The ancestral ( G ) allele of the SNP predominates in African and East Asian populations ( 93–100% ) , whereas the derived ( A ) allele is almost fixed in Europe ( 98 . 7–100% ) [4] . Functional assays of this gene suggested its direct involvement in human melanogenesis through cation-exchange activity [17] , [18] . However , the fact that the ancestral ( G ) allele is virtually fixed not only in Africans but also in East Asians suggests that light skin at high latitudes evolved independently in East and West Eurasia [19] . Genome-wide scans have also identified SLC24A5 as one of the most important “hot spots” for positive selection in Europeans , thereby supporting the role of natural selection acting on this gene [4] , [20] , [21] . Populations of South Asia live at lower latitudes than would be expected to require selection for lighter skin color on the basis of improved vitamin D synthesis [8] . Nevertheless , South Asians do exhibit a wide variation in skin color [22] . Two previous studies have assessed the genetics of skin pigmentation variation in expatriates from South Asia . The first of these [6] concluded that non-synonymous variants at three genes , SLC24A5 , SLC45A2 ( OMIM 606202 ) , and TYR ( OMIM 606933 ) , collectively contribute to variation in skin pigmentation in South Asians , with SLC24A5 showing the largest effect . The second study on common disease variants suggested high prevalence of the light skin associated allele of SLC24A5 in Asian Indians [23] . Nevertheless , both the studies involved populations that were structured and represented only a small range of the vast ethnic and genetic landscape of South Asia . Hence , comprehensive assessment of this phenotypic trait in native populations of South Asia has been lacking so far . Therefore , in the present study , we sought to address the following objectives . First , we aimed to quantify the amount of skin pigmentation variation that can be explained by the rs1426654 SNP of SLC24A5 in a homogeneous cohort of 1228 individuals from South Asia . Second , we studied the geospatial pattern of rs1426654-A allele in the Indian subcontinent using 1573 individuals from 54 populations and investigated how various factors influence its distribution . Third , we aimed to uncover the fine-scale genetic variation of SLC24A5 and determined the coalescence age of rs1426654 by resequencing 11 . 74 kb in a diverse set of 95 individuals . Lastly , we assessed whether SLC24A5 resequencing data and genome-wide genotype data were in concordance with the earlier reported evidence of positive selection in Europeans , and tested for any further evidence of selection among the studied populations . Our results confirm that rs1426654 plays a key role in pigmentation variation , while in-depth study of the light skin associated allele ( rs1426654-A ) among Indian populations reveals that the genetic architecture of skin pigmentation in South Asia is quite complex . The present study also provides important insights on evidence of positive selection and the evolutionary history of this light skin associated allele .
Phenotypic assessment of melanin index ( MI ) across 1674 individuals from two distinct cohorts , Cohort A and Cohort B ( see Materials and Methods; Tables 1 , S1 and S2 ) demonstrated a wide variation in skin color ( MI 28–79 ) in South Asia . Comparison with published datasets for the regions of the world revealed that the observed range in South Asians was three times greater than that in East Asians and Europeans and comparable to that of Southeast Asians ( Table 1 ) . Notably , Cohort A ( n = 1228 ) which included individuals from three closely related agricultural castes of Andhra Pradesh in South India , shows remarkable variation in skin color ( MI 30–64 ) , similar to heterogeneous pool of samples in Cohort B ( MI 28–79 ) . We tested the association of the rs1426654 SNP with pigmentation differences between the low ( MI<38 ) and high ( MI>50 ) MI groups of Cohort A ( Figure 1A ) , using a logistic regression model . A likelihood-ratio test to discern the association of the rs1426654 SNP to skin pigmentation , in addition to the influence of sex and population ( caste ) , showed a highly significant effect of rs1426654 genotype on skin pigmentation ( p = 2 . 4×10−31 ) with an odds ratio of 26 . 2 ( 95% CI 12–67 . 5 ) for the A allele . Furthermore , the cross-validated Area Under the Curve ( AUC ) score of 0 . 83 suggested that this model has a high discrimination power between the low and high MI groups . In summary , most of the pigmentation differences observed between the low and high MI groups could be explained by the rs1426654 SNP . We further aimed to estimate the effect size of the SNP . However , direct estimation of the effect size based on the samples genotyped from high or low MI group of Cohort A would only allow us to assess the effect of genotype for the extremes of pigmentation phenotype rather than for the whole distribution . Therefore , to estimate how much variation in MI could be explained by the rs1426654 SNP if all 1228 individuals in Cohort A had been genotyped , we used a multiple imputation approach based on simulations . The distribution of estimated mean MI across the genotypes , as obtained separately for males and females from the imputed dataset , is presented in Figure 1B and Table S3A . We observed that the estimated mean MI for each genotype in females was lower than that of males ( Table S3A ) . Analysis of the imputed datasets using a General Linear Model ( GLM ) revealed that the effect of genotype was highly significant ( p<1×10−16 ) . Notably , the total variation in pigmentation ( R2 ) that can be explained by the full model ( including sex and genotype ) was calculated to be 29% ( 95% CI , 24–34 ) , while that by the SNP alone was 27% ( 95% CI , 22–32 ) . Besides the quantitative assessment of the effect size , we found that the effect of the SNP was not exactly additive . Individuals with GG genotypes were darker than expected under the additive model ( Table S3B ) . This result is consistent with the similar mode of inheritance observed in SLC24A5 by Lamason [4] and in other pigmentation genes , such as KITLG ( OMIM 184745 ) and SLC45A2 [7] , [19] . Similar to Cohort A , our genotype-phenotype association tests on heterogeneous populations of Cohort B ( Table S2 ) , using a GLM after adjusting for sex and population , revealed that the effect of genotype was significant ( p = 3 . 24×10−8 ) . However , unlike Cohort A , where we did not observe any significant difference in mean MI of three castes ( p = 0 . 65 ) , the effect of population in Cohort B was highly significant ( p<2 . 2×10−16 ) . In an attempt to map the geospatial pattern of rs1426654-A allele frequencies across South Asia , we genotyped 1054 individuals across 43 ethnic groups including major language groups and geographic regions ( see Materials and Methods , Cohort C ) from the Indian subcontinent . In summary , 1573 individuals from 54 distinct tribal and caste populations from all the three cohorts ( A , B and C ) were assessed for this polymorphism ( Table S4; Figure S1 ) . We found that the rs1426654-A allele is widely present throughout the subcontinent , although its frequency varies substantially among populations ( 0 . 03 to 1 ) with an average frequency of 0 . 53±0 . 32 ( Table S4 ) . To explain how the various genetic and non-genetic factors affect the geospatial distribution of the rs1426654-A allele in the Indian subcontinent , we assessed the correlation of rs1426654-A allele frequency with major geographical divisions , language families and the ancestry component detected in previous studies of Indian populations [24] , [25] . However , to avoid bias due to low sample sizes in some of the populations , only data from 1446 individuals representing 40 populations were used ( Table S5 ) . Although we observe a considerable local heterogeneity , there is a general trend of rs1426654-A allele frequency being higher in the Northern ( 0 . 70±0 . 18 ) and Northwestern regions ( 0 . 87±0 . 13 ) , moderate in the Southern ( 0 . 55±0 . 22 ) , and very low or virtually absent in Northeastern populations of the Indian subcontinent ( Figure 2 , Table S6 ) . Notably , the Onge and the Great Andamanese populations of Andaman Islands also showed absence of the derived-A allele . Given the fact that one can observe a pronounced latitudinal cline for skin pigmentation across world populations , we also sought to test the observed derived-A allele frequencies in terms of absolute latitude and longitude in South Asia . We found that the rs1426654-A allele frequency in South Asia does not significantly correlate with latitude ( r = 0 . 23 , p = 0 . 15 ) . However , a significant negative correlation with longitude ( r = −0 . 49; p = 0 . 002 ) was observed . We found that the Tibeto-Burman and the Austroasiatic language families have the lowest frequencies of the A allele ( and Table S6 ) . The rs1426654-A allele frequency was significantly higher in Indo-European speakers than in other language groups ( Table S6 ) . In particular , there was a significant difference ( p<0 . 001 ) between the A allele frequencies of the Indo-European and the Dravidian speaking groups . We found that both language and geography have a significant influence on rs1426654-A allele frequency , as revealed by Mantel tests ( p<0 . 001 ) . We also studied the geospatial pattern of rs1426654-A allele frequencies at the global level using 2763 subjects from previously published data ( Table S7 ) and 1446 individuals from the present study ( Table S5 ) . The isofrequency map illustrates high frequencies of the rs1426654-A allele in Europe , Middle East , Pakistan , moderate to high frequencies in Northwest and Central Asia , while being almost absent in East Asians and Africans with notable exceptions in Bantu ( Southwest ) , San , Mandeka , and Ethiopians ( Table S7 , Figure 2 ) . As rs1426654-A allele frequency was found to be higher in West Eurasian populations that are known to share one of the genome-wide ancestry components of South Asia [24] , [25] , we sought to test the correlation between the derived-A allele frequency and the proportion of the West Eurasian ancestry component ( as depicted by the “light green component” in [24] ) for the studied populations . For this , we used the genome-wide information available on Indian populations from literature [24]–[28] ( Table S8 ) and relevant global reference populations to perform the ADMIXTURE run . Population structure as inferred by ADMIXTURE analysis at K = 7 is shown in Figure S2A . The proportions of k5 light green ancestry component obtained at K = 7 for the populations studied were plotted against the rs1426654-A allele frequency available for all populations and South Asia in particular ( Figure S2B ) . As shown in Figure S2B , we obtained a significant positive correlation for South Asian populations ( r = 0 . 90 , p<0 . 0001 ) but a weak , although significant correlation when all populations sharing the k5 component ( r = 0 . 64 , p = 0 . 04 ) were considered . We resequenced 11 . 74 kb of SLC24A5 ( Figure 3 ) , covering all the nine exons ( 1617 bp ) , introns ( 5797 bp ) , 5′ flanking ( 4150 bp ) , and 3′ flanking ( 177 bp ) regions ( Figure 3 ) in a global sample set of 95 individuals ( see Materials and Methods ) grouped into 8 broad geographic regions . A total of 60 variable sites ( including 23 singletons ) , one insertion , and one tetranucleotide repeat were identified with derived allele frequencies ranging from 0 . 005 to 0 . 39 . Results of the resequencing study for these variable sites are presented in Table S9 . According to dbSNP ( http://www . ncbi . nlm . nih . gov/projects/SNP/ ) build 137 ( June 2012 ) , 21 of these 62 identified variants were novel . The insertion present in the 5′ flanking region ( position 48411803 ) was confined to two San individuals ( San 15 and San 17 ) . Comparison of polymorphic sites across different regions revealed that the exons of SLC24A5 are highly conserved in humans . We detected only two variable positions within exons , with rs1426654 being the only non-synonymous SNP . The other variant , a synonymous ( Ser-Ser ) mutation identified at exon 7 at position 48431227 , was shared by four Africans . In contrast to low variation in the exonic region , a highly polymorphic tetranucleotide repeat ( GAAA ) was observed in the 5′ flanking region ( GAAA-GA-GAAA-GAAAAA- ( GAAA ) n-GAAAAA-GAAAA ) at position 48412029 . These repeats varied from 3 to 12 copies . A detailed analysis of the repeats did not reveal any correlation with the geographical origin of the samples or the haplogroups studied , in general ( Table S10 ) . However , chromosomes belonging to haplogroup H ( Figure S3 ) , defined by the rs1426654-A allele , were associated with larger repeat lengths ( 7–13 ) , albeit this association was not restricted only to them ( Table S10 ) . The nucleotide diversity estimated for the consensus resequenced region ( 11741 bp ) was observed to be 0 . 00042±0 . 00004 ( with Jukes-Cantor correction ) , which is low compared to the average of 0 . 00071±0 . 00042 for 647 genes resequenced in the NIEHS SNP database ( http://egp . gs . washington . edu/ ) . A sliding window approach based on similar measures ( window size = 100 bp , step size = 25 bp ) for the 5′ flanking region ( 4150 bp ) sequenced revealed that the 2726–2875 region demonstrates the highest nucleotide diversity of 0 . 00651 ( Figure S4 ) . Various molecular diversity indices studied for the eight geographical groups are presented in Table S11 and Figure S5 . Average pairwise differences observed among and within 8 different geographical regions using 11741 bp sequence data are summarized in Figure 4 . Populations from regions previously reported to exhibit a high frequency of the rs1426654-A allele ( North Africa and Middle East , Central Asia , South Asia and Europe; see Figure 2 ) show low levels of intra- and inter-population diversity in the resequenced region ( Figure 4 , Table S11 ) . We tested if our sequence data supports the well-documented evidence of positive selection for SLC24A5 in previous studies [4] , [13] , [20] , [21] , [29] , [30] and whether it provides any additional evidence of selection . None of the populations tested showed significant departure from neutrality , except for Europeans , who had negative Tajima's D ( p = 0 . 02 ) and Fu and Li's F* ( p = 0 . 04 ) as estimated from calibrated population genetic models using COSI ( Table S12 ) . Hence , these observations confirm that SLC24A5 has been under strong selective pressure in Europeans . In addition to this , we also performed haplotype-based selection tests based on genome-wide data ( see Materials and Methods ) of 1035 individuals including 145 Indians . XP-EHH scores demonstrated that SLC24A5 ranks among the top 10 candidate genes for positive selection in Europe , Middle East and Pakistan , and among the top 1% in Central Asia , Iran and North India ( Table S13 ) . Likewise , scores from our iHS analysis had significant empirical p-values for Central Asia and North India ( Table S13 ) . It is interesting to note that both of our haplotype-based selection tests demonstrated evidence of positive selection in North Indians , but no such evidence of positive selection was found in South Indians ( Table S13 ) . The difference in detecting selection signals from genotype and sequence data has also been pinpointed in a previous study [31] . Firstly , a phylogenetic tree was drawn on the basis of common variants observed in our worldwide resequencing data ( 11 . 74 kb ) of 95 individuals . The schematic tree representing the 8 most common haplogroups is shown in Figure S3 . Haplogroup G was the most common and geographically widely spread clade , being found in 7 of the 8 geographical groups examined . Haplogroup C was confined to sub-Saharan Africans only , while the rest of the observed haplogroups were shared between African and non-African populations . We conclude that all of the 73 phased chromosomes ( from Europe , sub-Saharan Africa , Middle East , South Asia , North and Central Asia ) with the rs1426654-A allele form a monophyletic group because they share the same haplotype background regardless of their geographic origin . In other words , all carriers of the mutation in our global sample share it by descent . The presence of the derived A allele in sub-Saharan Africa , although in low frequencies ( 2/73 - one heterozygous Mandeka and one heterozygous San individual ) ( Figure S3 ) is consistent with earlier findings [32] . We estimated the coalescence time of the rs1426654 mutation at 28 , 100 years ( 95% CI - 4 , 900 to 58 , 400 years ) using BEAST . Using the same mutation rate , the coalescent age estimated by rho statistics was 21 , 702 years ±10 , 282 years . Despite the different assumptions used in the two coalescent age estimation methods , both the age estimates show substantial overlap .
A number of previous studies have focused on admixed populations in the search for genes that determine skin pigmentation variation in humans [4] , [5] , [33]–[35] . Our formal tests for association , using a large homogenous population from South India ( Cohort A ) as well as a heterogeneous pool of samples across India ( Cohort B ) , demonstrated a highly significant effect of SLC24A5 on skin pigmentation . Further analysis of Cohort A revealed that this SNP determines most of the variation between the pigmentation extremes and contributes about 22–32% of the total skin color variation , thus suggesting that SLC24A5 plays a key role in the pigmentation diversity observed among South Asians . Furthermore , confounding effect of population structure on the genetics of skin pigmentation , evident in Cohort B suggests that the marked population substructure of South Asians must be taken into account when genetic association studies are conducted in these populations . Our extensive survey of rs1426654-A allele frequency in the Indian subcontinent reveals an average frequency of 0 . 53 with a substantial variation among populations , ranging from 0 . 03 to 1 ( Table S4 ) . This finding stands in contrast to the previous understanding of the spread of this allele , where a study [23] based on a cohort of 15 Indian ethnic groups sampled in the US ( n = 576 ) , estimated the average A allele frequency at 0 . 86 , with a relatively low level of variation among populations ( observed range 0 . 70 to 1 ) . The most plausible cause of this discordance might be that fewer populations were included in the former study and the groups were defined by their generic linguistic affiliation in major branches of the Dravidian and Indo-European languages , rather than by finer resolution of the endogamous units . Notably , in the subset of 8 populations that could be characterized on a similar basis in both studies , the estimates of A allele frequencies did not diverge significantly in their combined averages ( Table S14 ) . Therefore , these comparisons suggest that sampling strategies are pivotal in determining the extent of genetic diversity observed in Indian populations and that sampling of expatriates may have a homogenizing effect . Moreover , the expatriates are known to represent mainly urban populations of India , which constitute only 30% ( Census 2011; http://censusindia . gov . in/ ) of the total population of the subcontinent , and therefore are unlikely to be representative of the wealth of genetic variation harbored within the subcontinent . Our quest to determine whether and to what extent the distribution of the rs1426654 derived- A allele frequency in South Asian populations correlates with language and/or geography revealed that both of these variables have a significant predictive value on allele frequencies . In particular , we found that although frequencies among populations studied vary considerably , this polymorphism has an evident geographic structure with higher frequencies of the derived allele in North and Northwest regions and a declining pattern as one moves further South and East ( Table S5 , Figure 2 ) . However , when we plotted the rs1426654-A allele frequency against the geographical coordinates of our sampled populations , we found a significant correlation with longitude but not with latitude . The lack of a clear latitudinal ( North-South ) cline in the A allele frequency , which would have been expected under the model of natural selection , could be partly explained by the complexity of the South Asian genetic landscape , influenced by differences in population histories shaped by various micro-level migrations within the subcontinent , strict endogamy and social barriers . For example , Saurashtrians , who migrated from “Saurashtra” region of Gujarat to South India ( Madurai ) for work , have a relatively high rs1426654-A allele frequency of 0 . 70 . It is believed that those Saurashtrians presently dwelling in Madurai were invited by Nayak kings for their expertise in silk-weaving [36] . Similarly , Toda have higher A allele frequency ( 0 . 86 ) compared to Kurumba ( 0 . 20 ) , their geographical neighbors , most likely due to their higher proportion of West Eurasian ancestry which is supported by Y chromosome evidence [37] . Notably , Brahmins , irrespective of their geographic source ( North , Central or South India ) have higher A allele frequency ( Table S5 ) . Conversely , the higher longitudinal correlation could be due to the fact that Tibeto-Burman and Austroasiatic speakers are characterized by very low A allele frequency ( Table S6 ) because of their East Asian ancestry [26] , [38] . Therefore , their inclusion in our sampling might have resulted in the inflation of the longitudinal correlation coefficient . Although the last decade has witnessed significant improvement in the understanding of the genetic basis of skin pigmentation , our knowledge about the exact mechanisms behind the evolution of light skin in humans is still incomplete . The genetic evidence that has accumulated till date suggests a complex evolutionary history for skin pigmentation . It has been argued that natural selection in response to UVR had a causative role in the evolution of light skin color at high latitudes [8] , [39] , [40] . Evidence of population-specific signatures of selection of pigmentation genes at different timescales suggests that the evolution of light skin was not a one-step process [41] , [42] but a consequence of multiple events or episodes during human evolution . It appears that some of the mutations which have been associated with light skin started to accumulate relatively early in modern human history in the proto-Eurasian populations following the Out-of-Africa expansion , whereas other mutations arose after the divergence of East and West Eurasian populations [4] , [19] , [29] , [41] . Hence , studies focusing on the time-scale of genetic changes in pigmentation genes are vital for understanding the complex evolutionary history of human skin pigmentation . Therefore , in this study , we focused on providing an age estimate of the rs1426654 mutation , which has a major effect on skin pigmentation in West Eurasian and South Asian populations . Notably , previous studies providing age estimates for this locus have been mostly confined to the estimation of onset of selective sweep rather than the coalescence time of the mutation . A study of extended haplotype homozygosity in HapMap populations estimated that the most intense signals of selection detected in European and East Asian populations are found in haplotypes which extend 0 . 52 cM on average in length [20] . Assuming a star-shaped genealogy and a generation time of 25 years , the authors dated the peak of these signals to ∼6 . 6 KYA [20] . They also observed that the second-longest haplotype ( 1 . 15 cM ) in Europe includes SLC24A5 , where rs1426654-A was found to be fixed . Using the same formula used by Voight [20] to date the average peaks of selection signals in Europe and East Asia , the selective sweep specifically at SLC24A5 in the HapMap European sample can be dated to ∼3 KYA . Besides this , a recent study by Beleza [42] , focusing on analyses of diversity in microsatellite loci , estimated that the selective sweep at SLC24A5 occurred around 11 . 3 KYA ( 95% CI , 1–55 . 8 KYA ) and 18 . 7 KYA ( 5 . 8–38 . 3 KYA ) under additive and dominant models , respectively [42] . Our Bayesian coalescent age estimate of the rs1426654-A allele at ∼28 KYA ( 95% HPD , 5–58 KYA ) , as well as the rho-based estimate at 21 . 7 ( ±10 . 3 ) KYA , are older in their point estimates than both of the above selective sweep date estimates , although these age estimates have broad and overlapping error margins . This finding is not surprising because sweeps can also operate on standing variation . Besides this , both our rho-based point estimate and Bayesian mean age estimate postdate the estimated time of the split between Europeans and Asians calculated by Scally [43] using a similar mutation rate . Although our confidence intervals cannot rule out entirely the possibility of older dates ( >28 KYA ) , our findings are broadly consistent with the evolutionary model of skin pigmentation proposed in earlier studies [41] , [42] , [44] . It appears that the most plausible scenario is that light skin evolved as an adaptation to local environmental conditions as humans started moving to northerly latitudes , with the initial phase of skin lightening occurring in proto Eurasian populations , while genetic variation in SLC24A5 formed the later phase which led to lighter skin in Europeans and South Asians , but not East Asians . This was followed by a European-specific selective sweep , which favored the rapid spread of this mutation in these populations . Our coalescence age estimates of 28 KYA ( 95% HPD 5–58 KYA ) show wide margins , also evident in the earlier sweep date estimates for the gene [42] . This can be due to the fact that the power of our analysis was limited by the need to reduce our sequence range to a subset of sites from a region with sufficiently high LD around the rs1426654-A allele and very low level of sequence variation . Therefore , we speculate that narrowing down the coalescence age estimates and specifying the geographic source of the rs1426654-A allele will depend rather on the success of ancient DNA studies than on more extensive sequencing . Earlier studies have highlighted SLC24A5 as one of the top candidate genes demonstrating evidence for positive selection in Europeans [4] , [13] , [20] , [21] , [29] , [30] and in Middle Eastern and Pakistani populations from South Asia [13] , [29] on the basis of either FST or extended haplotype homozygosity from genotype data . Here , relying on our previous scans of extended haplotype homozygosity on Indian populations [24] , we note that both XP-EHH and iHS suggest that positive selection has occurred in North Indian ( within top 5% and top 1% respectively ) but not in South Indian populations . One possible explanation for the regional differences in empirical ranks of the SLC24A5 in India could be the “melanin threshold” hypothesis [45] . According to this hypothesis , natural selection affects the variation in pigmentation phenotype only up to a certain adaptive optimum , beyond which individuals may show variation that is subject to other factors such as admixture , genetic drift etc . However , differently from the expectations of this hypothesis , we do observe high range of melanin indices both in North and South Indian populations of Cohort B ( Table S2 ) . Furthermore , the high positive correlation of rs1426654-A allele with the light-green South Asian ancestry component ( Figure S2A ) advocates that the rs1422654-A allele frequency patterns in India could be also explained by demographic history of the populations in addition to selection . It is also possible that while XP-EHH and iHS tests have increased power to detect selection signatures associated with high allele frequencies , the low ranking position of SLC24A5 in selection scans of South Indians is due to the overall lower frequency of the rs1422654-A allele . Therefore , the complex patterning of light skin allele in India and its correlation with geography , language , and ancestry component observed in the present study , portrays an interesting interplay between selection and demographic history of the populations . This stands in contrast to Europe where the frequency of the light skin associated allele of SLC24A5 has almost reached to fixation and seems to be attributable solely to natural selection . This aspect of skin pigmentation variation observed in South Asians is pivotal in understanding the different mechanisms that contribute to the global skin pigmentation variation and in further understanding of this complex phenotypic trait . To summarize , we have provided evidence using a homogeneous cohort that the rs1426654 SNP plays a key role in skin pigmentation variation in South Asia . We have shown that the rs1426654-A allele is widespread in the Indian subcontinent and its complex pattern is a result of combination of processes involving selection and demographic history of populations , influenced by their linguistic and geographic affiliations . Phylogenetic analyses of resequencing data confirm that the rs1426654-A allele in West Eurasian and South Asian populations occurs on the same haplotype background . Both sequence and genome-wide genotype data confirm evidence of positive selection in Europeans , while the latter supports further evidence of selection in populations of Middle East , Pakistan , Central Asia and North India but not in South India . We date the coalescence of the light skin allele ( rs1426654-A ) to 22–28 KYA ( 95% CI , 5–58 KYA ) . However , since this allele has become fixed in many populations across its current distribution , we propose that ancient DNA research might have greater potential to improve our understanding of when and where it first appeared .
This study was approved by the Research Ethics Committee of the Estonian Biocentre , Tartu , Estonia and the Institutional Ethical Committee ( IEC ) of the Centre for Cellular and Molecular Biology , Hyderabad , India . All recruited individuals were >18 years of age and their ethnic origin was determined via personal interviews . Written informed consent was obtained from all participants . Skin pigmentation was measured using DermaSpectrometer ( Cortex Technology , Hadsund , Denmark ) . Erythema ( E ) and melanin index ( MI ) readings were taken from the upper inner arm ( medial aspect ) [46] . For a subset of individuals , additional measurements were taken from the forehead , representing the most tanned or sun-exposed region of the skin . However , only MI readings from the upper inner arm were used for association analyses . DNA was isolated either from blood or saliva ( using Oragene DNA kits , Canada ) . The study involved three distinct cohorts , A , B and C . Sampling locations of these cohorts are shown in Figure S1 . Cohort A included 1228 randomly recruited individuals from three major agricultural castes ( Kapu , Naidu and Reddy ) of Andhra Pradesh , India . For all the above individuals , MI readings were taken from the right and left upper inner arm and their mean was calculated to determine each individual's MI . Following the phenotypic screening , thresholds were set for the “low” ( MI<38 ) and “high” ( MI>50 ) MI groups respectively , representing approximately the top and bottom 10% of the MI distribution , for collection of DNA samples ( Figure 1A ) . Eighty-four out of 120 individuals from the low MI group and 102 out of 127 individuals from the high MI group were genotyped successfully . The 10% threshold was implemented after an initial pilot study , following which the values were continuously redefined as the sample collection progressed . Consequently , during the fieldwork , DNA from 56 individuals was collected outside the determined thresholds ( MI 38–50 ) . Therefore , in summary , 242 individuals ( 189 males , 53 females ) from this cohort were genotyped for the rs1426654 SNP ( Table S1 ) . Cohort B comprised of 446 individuals , including 10 caste and tribal populations of Tamil Nadu , Maharashtra and Haryana states of India . For each individual , three readings of MI were taken from the right upper inner arm and the values were averaged . Out of these , 277 individuals ( 246 males and 31 females ) were genotyped ( Table S2 ) . Cohort C included 1054 individuals , representing 43 endogamous populations from different ethnic backgrounds , language families ( Dravidian , Indo-European , Austroasiatic , Tibeto-Burman speakers ) , castes , tribes , with their geographical locations covering most of the states . No records for MI were available for this cohort . In summary , 1573 individuals from 54 distinct tribal and caste populations including all the three cohorts ( A , B and C ) were assessed for the rs1426654 polymorphism ( Table S4 and Figure S1 ) . A detailed description of the geographic location , linguistic affiliation and socio-cultural background of each cohort is given in Tables S1 , S2 and S4S4 . Populations from Cohort A and Cohort B with MI readings were used for genotype-phenotype analyses and genotyping results from all three cohorts ( A , B and C ) were used to map the spread of rs1426654-A allele and test its correlation with language , geography and ancestry component . For the resequencing study , we designed a global panel comprising of 95 individuals . This included 70 subjects from HGDP-Centre d'Étude du Polymorphisme Humain ( HGDP-CEPH ) worldwide panel [47] , and additionally 3 Europeans , 18 Indians , and 4 Central Asians to cover the underrepresented regions of the CEPH panel . For population-level analyses , these 95 individuals were broadly classified into 8 major groups based on their geography and ethnicity: sub-Saharan Africa ( n = 22 ) , North Africa/Middle East ( n = 7 ) , Europe ( n = 11 ) , North and Central Asia ( n = 7 ) , South Asia ( n = 23 ) , East Asia ( n = 14 ) , Native Americans ( n = 4 ) and Melanesia ( n = 7 ) . List of the populations included in the resequencing project , representing these regions is given in Table S9 . A 443 bp region of SLC24A5 flanking the rs1426654 SNP was amplified by PCR using s . E3 , 4F and s . E3 , 4R primers ( Table S15 ) . The cycling protocol consisted of 96°C for 3 min , 32 cycles of 96°C for 30 s , 57°C for 30 s , 72°C for 1 min and final extension at 72°C for 5 min . The PCR product was then either directly sequenced or digested overnight at 37°C using Hin6I restriction endonuclease enzyme . All digested products were run on a 3% agarose gel . The products for sequencing were run on 3730XL DNA Analyzer ( Applied Biosystems , Foster City , CA ) using Big Dye Terminator sequencing kit ( v3 . 1 Applied Biosystems ) . The effect of the functional SLC24A5 SNP ( rs1426654 ) on skin pigmentation differences between low ( <38 MI ) and high ( >50 MI ) MI groups of Cohort A was tested using a logistic regression model . For this , we compared a model that included sex and population ( caste ) as predictors to a model in which the genotype was added as an independent variable . An association between SNP and melanin index was tested using a likelihood-ratio test after adjusting for sex and population and , assuming additivity , odds ratio was calculated for the rs1426654-A allele . Furthermore , we calculated the cross-validated Area Under the Curve ( AUC ) value to quantify how accurately this polymorphism predicts the occurrence of an individual in the low or high MI group , using the R package caret [48] . To estimate the effect size of the SNP , we used a simulation-based approach known as multiple imputation [49] . This method uses regression models and Bayesian sampling to impute missing values conditional on other predictors . Using random imputations , 1000 complete datasets were generated . The desired analysis was performed on each dataset using methods based on complete data . Results were pooled to derive corrected point estimates and inference [49] , [50] . Using this methodology , we estimated the mean MI for each genotype separately for males and females . We also estimated the coefficient of determination ( R2 ) for the full model which included sex and genotype , and the variation of melanin index that can be explained by rs1426654 SNP alone . We tested the effect of genotypes on melanin index using a generalized linear model ( GLM ) . All the above stated analyses were performed using the R package MICE 2 . 9 [50] . For randomly collected samples ( Cohort B ) , similarly , the effect of rs1426654 genotypes on melanin index was assessed using a GLM . Furthermore , the effect of the genotype in the cohort studied was tested using an additive model . All statistical analyses were performed using the R computing package ( version 2 . 15 . 2 . 1 ) ( http://www . r-project . org/ ) . To visualize the geospatial pattern of the rs1426654 SNP in South Asia and to compare it with other populations across the world , an isofrequency map was generated using 1446 individuals genotyped across all three cohorts ( Table S5 ) and 2763 subjects from previously published datasets ( Table S7 ) . The isofrequency map was drawn using Surfer 8 . 0 ( Golden Software Inc , Golden , Colorado ) . To test the distribution of the rs1426654-A allele across different language families and geographical coordinates , all of the individuals genotyped under the three cohorts were grouped into 7 geographical zones and 4 major language families pertinent to India ( Table S6 ) . Some populations were regrouped with their geographical neighbors of same ethnicity ( Table S5 ) . Populations that could not be grouped and had low sample size ( n<15 ) were excluded . Therefore , data from 1446 individuals representing 40 populations were used for the linguistic and geographical analyses ( Table S5 and S7 ) . The rs1426654-A allele frequency was also assayed across the geographical coordinates ( absolute latitude and longitude ) using Pearson's correlation test . A Mantel test was used to examine the interaction of the allele frequencies with geography and language . For this , the genetic distance matrix ( based on FST ) was generated in Arlequin 3 . 5 . 1 . 3 [51] and the geographical matrix was calculated from geographic coordinates . For the language matrix , we used the binary approach by coding populations speaking a language from the same language family as 0 and different language family as 1 . A Mantel test was performed using Arlequin with 10 , 000 permutations . We tested the correlation between the derived rs1426654-A allele frequency and the proportion of the ancestry component that South Asian populations share with West Eurasians ( as depicted by the “light green component” in [24] ) . For this , genome-wide datasets on Indian populations available from literature [24]–[28] and relevant global reference populations were combined and subjected to structure-like analysis using ADMIXTURE [52] to determine the proportions of the hypothetical ancestral populations using the methods described by Metspalu [24] . A list of the populations included in the run and their source from the literature is given in Table S8 . We ran ADMIXTURE 100 times from K = 2 to K = 9 to monitor convergence between individual runs . Log-likelihood scores suggested that the global maximum was reached at K = 7 . Population structure of the studied populations as inferred by ADMIXTURE analysis at K = 7 using 98 , 189 SNPs is shown in Figure S2A . The proportions of the k5 light green ancestry component ( Figure S2A ) at K = 7 were then extracted and compared with rs1426654-A allele frequency , for those world and South Asian populations for which the rs1426654 frequency was available , using Pearson's correlation test . A total of 11 . 74 kb region of SLC24A5 comprising exons ( 1617 bp ) , introns ( 5797 bp ) , 5′ flanking ( 4150 bp ) , and 3′flanking ( 177 bp ) regions spanning over 25 . 6 kb ( 48409019–48434692 ) was resequenced ( Figure 3 ) in 95 multiethnic individuals using 31 pairs of validated primers ( Table S15 ) . PCR products were purified with Exo-SAP prior to sequencing . Bidirectional sequencing for each fragment was performed using Big Dye Terminator sequencing kit ( v3 . 1 Applied Biosystems ) and run on 3730XL DNA Analyzer ( Applied Biosystems , Foster City , CA ) . The sequences were then assembled and analyzed by Seqscape ver 2 . 5 ( Applied Biosystems ) . BIOEDIT 7 . 1 . 3 was used to align the sequences to the NCBI Reference Sequence ( NG_011500 . 1; 28421 bp ) . Variants were annotated with SNPs included in dbSNP build 137 , June 2012 . All of the variants were confirmed by manual inspection . The sequences were phased using PHASE 2 . 1 . 1 implemented in DnaSP 5 . 10 . 01 [53] . Sequence diversity measures ( π and θ ) were computed using DnaSP [53] and Arlequin 3 . 5 [51] was used to perform the interpopulation and intrapopulation analyses . For resequenced data , we tested for the effects of selection using Tajima's D [54] , Fu and Li's D* and F* [55] statistics , calculated in DnaSP 5 . 10 . 01 [53] . All the tests were performed under the standard assumption of constant population size . However , since these tests are known to be strongly influenced by population history , the significance of the results was also estimated by means of coalescent simulations using the COSI 1 . 2 . 1 software with the best-fit population model [56] . We performed 10 , 000 replicates . Coalescent simulations were conditioned on a specific mutation and recombination rate . We used a mutation rate of 5×10−10 substitutions/site/year , as reported by Scally and Durbin [43] . Estimates for the local SLC24A5 recombination rate were obtained from HapMap Build 37 [57] and the length of simulated sequence matched that of the resequenced region ( 11741 bp ) . In the absence of an appropriate demographic model and empirical distribution , we have used the evolutionarily closest population implemented in COSI to assess the significance . For selection analyses based on genome-wide genotype data , we used a merged data set of Illumina Infinium 650K , 610K and 660K available for 145 Indians and worldwide samples including Bantu ( n = 19 ) , Middle East ( n = 133 ) , Europe ( n = 100 ) , Central Asia ( n = 77 ) , Iran ( n = 20 ) , Pakistan ( n = 165 ) , East Asia ( n = 211 ) , Oceania ( n = 27 ) from published datasets . Two haplotype-based selection tests , Cross-Population Extended Haplotype Homozygosity ( XP-EHH ) and Integrated Haplotype Scores ( iHS ) , were used to assess the empirical rank of the SLC24A5 in the haplotype homozygosity scans performed across the genome in each of the 8 world regions . iHS and XP-EHH statistics were calculated using code by Joseph Pickrell , available at hgdp selection browser ( http://hgdp . uchicago . edu/ ) . The analyses were based on a genome scan of 13 , 274 windows of size 200 kb each . Unphased SNP data were retrieved for the genomic window containing SLC24A5 ( chromosome 15:46 . 2–46 . 4 Mb ( Build 36/hg18 ) and compared to the empirical distribution of other windows across the genome . Yoruba was used as the reference population in XP-EHH analyses . Data were phased using Beagle 3 . 1 . [58] . We estimated the phylogeny of SLC24A5 haplotypes based on sequences of 11 . 74 kb for our diverse set of 95 individuals . For this , haplotypes were inferred from the genotype data using PHASE v . 2 . 1 . 1 [59] . A neighbor-joining phylogenetic tree was constructed from these data using MEGA 5 [60] . A schematic tree representing the eight most common branches of the haplotype tree is shown in Figure S3 . We estimated the age of the rs1426654 mutation using 8837 bp of the SLC24A5 gene . This region was determined by the largest linkage-disequilibrium block identified by the four-gamete rule algorithm , using a minimum D' value of 0 . 8 , as implemented in Haploview 4 . 2 [61] . Coalescence times were estimated using Bayesian phylogenetic analysis in BEAST 1 . 7 . 0 [62] . The analysis was conducted on a dataset of 73 sequences carrying the rs1426654-A allele . We further restricted our dataset to 7837 bp comprising of third codon sites , introns and flanking regions . The F81 [63] nucleotide substitution model was selected as the best-fit model using the Bayesian information criterion in Modelgenerator [64] . The analysis was performed using a strict molecular clock and the Bayesian skyride coalescent model [65] . The molecular clock was calibrated using the mutation rate reported by Scally and Durbin [43] , with a mean of 5×10−10 mutations/site/year and a standard deviation of 5 . 1×10−11 . Posterior distributions of parameters were estimated by Markov chain Monte Carlo simulation , with samples drawn every 1000 steps over a total of 10 , 000 , 000 steps . Three independent runs were conducted to check for convergence to the stationary distribution and the first 1000 samples were discarded as burn-in . Sufficient sampling of parameters was evaluated using Tracer 1 . 5 [66] and samples from independent runs were combined . Sampled posterior trees were summarized to generate a maximum-clade-credibility tree . Statistical uncertainty in age estimates is reflected by the 95% credibility intervals . We also estimated the coalescent times using the rho statistics [67] in Network 4 . 6 ( http://www . fluxus-engineering . com/sharenet . htm ) assuming a rate of 5×10−10 substitutions/site/year [43] and using sequence length of 8837 bp . The standard deviation was calculated according to Saillard [68] .
|
Human skin color is one of the most visible aspects of human diversity . The genetic basis of pigmentation in Europeans has been understood to some extent , but our knowledge about South Asians has been restricted to a handful of studies . It has been suggested that a single nucleotide difference in SLC24A5 accounts for 25–38% European-African pigmentation differences and correlates with lighter skin . This genetic variant has also been associated with skin color variation among South Asians living in the UK . Here , we report a study based on a homogenous cohort of South India . Our results confirm that SLC24A5 plays a key role in pigmentation diversity of South Asians . Country-wide screening of the variant reveals that the light skin associated allele is widespread in the Indian subcontinent and its complex patterning is shaped by a combination of processes involving selection and demographic history of the populations . By studying the variation of SLC24A5 sequences among a diverse set of individuals , we show that the light skin associated allele in South Asians is identical by descent to that found in Europeans . Our study also provides new insights into positive selection acting on the gene and the evolutionary history of light skin in humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2013
|
The Light Skin Allele of SLC24A5 in South Asians and Europeans Shares Identity by Descent
|
For many organisms the ability to transduce light into cellular signals is crucial for survival . Light stimulates DNA repair and metabolism changes in bacteria , avoidance responses in single-cell organisms , attraction responses in plants , and both visual and nonvisual perception in animals . Despite these widely differing responses , in all of nature there are only six known families of proteins that can transduce light . Although the roundworm Caenorhabditis elegans has none of the known light transduction systems , we show here that C . elegans strongly accelerates its locomotion in response to blue or shorter wavelengths of light , with maximal responsiveness to ultraviolet light . Our data suggest that C . elegans uses this light response to escape the lethal doses of sunlight that permeate its habitat . Short-wavelength light drives locomotion by bypassing two critical signals , cyclic adenosine monophosphate ( cAMP ) and diacylglycerol ( DAG ) , that neurons use to shape and control behaviors . C . elegans mutants lacking these signals are paralyzed and unresponsive to harsh physical stimuli in ambient light , but short-wavelength light rapidly rescues their paralysis and restores normal levels of coordinated locomotion . This light response is mediated by LITE-1 , a novel ultraviolet light receptor that acts in neurons and is a member of the invertebrate Gustatory receptor ( Gr ) family . Heterologous expression of the receptor in muscle cells is sufficient to confer light responsiveness on cells that are normally unresponsive to light . Our results reveal a novel molecular solution for ultraviolet light detection and an unusual sensory modality in C . elegans that is unlike any previously described light response in any organism .
An animal's complement of sensory abilities reflects its unique evolutionary and natural history . Despite its status as a major model organism , little is known about the natural history of Caenorhabditis elegans . Past studies have revealed four major modalities through which C . elegans senses its environment: chemosensation , mechanosensation , osmosensation , and thermosensation [1–4] . These four modalities would seem sufficient to meet the sensory needs of what is often referred to as a subterranean animal; however , recent studies suggest that C . elegans may spend much of its time above ground , living on small surface-dwelling animals or their carcasses [5 , 6] . C . elegans may therefore be frequently exposed to direct sunlight , which can damage or kill cells by photo-oxidative reactions [7] . If C . elegans spends significant time above ground , it would need a sensory mechanism for detecting and avoiding lethal doses of direct sunlight . However , despite extensive observation under blue , blue-violet , and even ultraviolet ( UV ) light , there are no published reports of a behavioral response to high energy light ( blue wavelengths or shorter ) in C . elegans . Here we show that C . elegans does in fact have a strong response to short wavelength light that takes the form of a robust acceleration of locomotion . Our data suggest that C . elegans uses this light response to escape the ultraviolet light in direct sunlight . This light response can restore normal or hyperactive locomotion to certain kinds of paralyzed synaptic signaling mutants . To identify the molecular basis for this form of light reception , we performed a forward genetic screen and identified mutants defective in the response . The mutations disrupted LITE-1 , which is a member of the Gustatory receptor ( Gr ) family that we show functions as a short wavelength light detector when expressed in a heterologous tissue .
Cyclic AMP ( cAMP ) and diacylglycerol ( DAG ) are important universal signals that neurons use to shape and drive behaviors , learning , and memory . Neurons use Gα proteins to tightly control the production of cAMP and DAG at synapses . In C . elegans , convergent Gαq and Gαs pathways make DAG and cAMP , respectively , which control synaptic activity to generate the locomotion behavior ( Figure S1 ) . Mutations that eliminate either of these two major pathways result in animals that are nearly paralyzed and are essentially unresponsive to harsh physical stimulation ( Videos S1 and S3 ) . However , blue-violet light projected onto paralyzed unc-31 null mutants , which have a nonfunctional Gαs pathway [8] , restored coordinated locomotion ( Figure 1A and Video S2 ) and increased their mean locomotion rate over a 6-min period to a level that was 65-fold higher than the basal rate of the mutant and 2-fold higher than the basal rate of wild type ( Figure 1C ) . Blue-violet light caused a similar increase in the locomotion rate of an acy-1 ( adenylyl cyclase ) null mutant that lacks all Gαs-driven locomotion specifically in neurons [9 , 10] ( Figure 1C ) . Thus , the light response pathway does not require the production of cAMP through the neuronal Gαs–adenylyl cyclase pathway . Blue-violet light also restored coordinated locomotion to a strong reduction-of-function egl-30 ( Gαq ) mutant ( Video S4 ) and increased its locomotion rate 67-fold , to a level that was not significantly different from the basal rate of wild-type animals . Thus , the light response pathway , at the most , requires only very low levels of the DAG and activated RhoA produced by the Gαq pathway . In contrast , blue-violet light did not affect the movement of a similarly paralyzed unc-13 mutant , which has defects in the late stages of neurotransmitter release ( Figure 1C ) . Thus , the light response requires neurotransmitter release and does not induce nervous system–independent muscle activity . We determined the wavelength specificity of the response by projecting light through a series of filters in a stereomicroscope nosepiece ( Figure 1B ) and measuring the locomotion response of the unc-31 mutant at various wavelengths and a constant light power of 720 μW/mm2 . At this power , the mutant only responded to wavelengths of ∼500 nm ( blue-green ) or shorter ( Figure 1D ) . The fold-increase over the basal response increased from zero at 545 nm ( green ) to ∼65-fold at 441 nm ( blue-violet ) . Although the maximum ultraviolet ( UV ) power produced by this light system was 52 μW/mm2 , UV light of this power caused a response equal to 720 μW/mm2 of blue light ( Figure 1D ) . Thus , the light response is most sensitive to UV light , but higher levels of violet and blue light also activate it . Wild-type animals had the same wavelength sensitivity as the unc-31 mutant , although wild type's peak locomotion rate in 720-μW/mm2 blue violet light was ∼1 . 5-fold higher than the unc-31 null mutant ( Figure S2 ) . A comparison of the light responses at various powers of blue-violet , blue , and green light again highlighted the strong specificity of the response for short wavelengths , especially ultraviolet light . Increasing the power of green light to 5 , 500 μW/mm2 only improved the locomotion rate of the unc-31 null by 6 . 7-fold ( Figure 2A ) and did not affect the locomotion rate of wild type ( Figure 2B ) . In contrast , the blue and blue-violet responses of the unc-31 mutant increased sharply between 50 and 700 μW/mm2 , peaking at 1 , 400 and 2 , 800 μW/mm2 for blue-violet and blue light , respectively . At these powers , both blue and blue-violet light increased the locomotion rates of wild-type and unc-31 mutants 3 . 5- and 67-fold , respectively . At blue-violet powers greater than 1 , 400 μW/mm2 the locomotion rate decreased significantly from its peak level ( Figure 2A ) , possibly due to overstimulation of the response or light damage over the 6-min assay period ( more on this below ) . Wild-type animals required about half as much light power as the unc-31 null mutants to maximize their responses to blue and blue-violet light ( Figure 2B ) . The dose-response data also showed that a UV power of 50 μW/mm2 produces a response that is about equal to 350 μW/mm2 blue-violet light . Thus , UV light is about 7-fold more potent than blue-violet light in producing the response . The white light power of ambient room lighting is about 0 . 5 μW/mm2 , which is 100-fold weaker than the minimum power necessary to induce the light response ( Table S1 ) . Thus , this light response is optimized for high powers of ultraviolet light . The high powers of light required to elicit this response raise the possibility that the animals are responding to temperature changes induced by the light . However , when we inserted a temperature probe into a pellet of adult worms , such that the probe could only be heated by the interaction of light with the worms , the temperature changes induced by blue-violet and green light were not statistically different , and amounted to less than 0 . 7 °C ( Figure 3A ) . Moreover , increasing the power of green light to a level that raised the temperature of the worms by 1 . 9 °C had no effect on their movement when we reproduced this power in a locomotion rate assay . In contrast , a much lower power of blue-violet light increased wild type's movement almost 3 . 5-fold ( Figure 3A ) . Thus , there is no correlation between the small temperature changes induced by the light and the behavioral response . Furthermore , when we transferred paralyzed unc-31 mutants from room temperature plates to culture plates preheated to various temperatures between 24 °C and 50 °C , and assayed them for locomotion rate in the first minute after transfer , we found no temperature that increased their locomotion rate ( Figure 3B ) . Thus , abrupt temperature increases do not make paralyzed unc-31 mutants move . To determine the time course of the light response , we measured locomotion rates at 5-s intervals during the first 10 s of light exposure , and then at 10-s intervals thereafter , and we blocked illumination at 1 min to observe the time course of decay . In blue-violet light , wild-type worms showed a rapid response initiation ( 1–2 s ) and a slow response decay after turning off the light ( ∼2 min ) . Their average locomotion rate in the first 5 s of exposure was 2-fold higher than the basal rate ( Figure 4A ) . Their response peaked at 20 s , and , after turning off the light at 1 min , their locomotion slowly decayed to the basal rate over the next 140 s ( Figure 4A ) . In contrast to wild type , the unc-31 null showed no response during the first two 5-s intervals but rapidly accelerated thereafter , until after 1 min , its locomotion rate was 100-fold higher than its basal rate ( Figure 4B ) . After turning off the light , the unc-31 null mutant took about twice as long as wild type for its response to decay to basal levels ( Figure 4B ) . The delayed response of the unc-31 null mutant , as well as the long decay times of both wild type and the mutant , suggests that light induces the build-up of a signal , and that the signal must reach higher levels in the mutant to induce the response . The unc-31 mutant's peak locomotion rate after only 1 min of illumination was about 50% higher than its average rate over 6 min of continuous light exposure of the same power . This suggests that worms slow down over longer periods of illumination , and extended time course assays showed that to be true . Wild-type worms exposed to 1 , 500 μW/mm2 blue-violet light ( i . e . , twice the power that maximizes the locomotion response ) over a 30-min illumination period steadily slowed down after their responses peaked ( Figure 4C ) . This is because the animal is acutely injured by light exposure; it is not an adaptive response , because extended illumination with 2 , 800-μW/mm2 blue-violet light caused death in 25 min ( Figure 4D ) . However , green light of the same power , or even double this power , did not kill worms and did not even affect their locomotion rate over a 1-h illumination period ( Figure 4D ) . This suggests that these high powers of green light do not damage worms and that there is a sharp energy cutoff between blue and green wavelengths for heat-independent biological light damage . The light-induced death is not caused by overactivation of the light response , because mutants lacking the response ( described below ) die at the same time as wild-type animals during blue-violet light exposure ( Figure S3 ) . Short wavelength–light kills worms , which suggests that the locomotion response is an escape strategy . To test this , we illuminated crowded plates containing defined numbers of animals and quantified the number of animals that remained in the illuminated area after 45 s of light , and the number of animals that fully entered the illuminated area over the next 5 min . Our results clearly showed that the response to light is a photophobic response , because animals actively avoided the blue-violet illuminated area ( Figure 5A ) . To test the hypothesis that avoidance of direct sunlight is the ecological reason that C . elegans has a light response , we used blue , blue-violet , and UV excitation filters to measure the power of these colors in direct sunlight at solar noon at an altitude of 1 , 276 feet ( 389 m ) above sea level . Figure 5B shows the sunlight values for each color after correcting for the percent transmission of each filter . We reproduced these values on the stereomicroscope , and then tested the response of wild-type animals to these different wavelengths at these powers . This experiment could only reproduce the long-wavelength UV light in sunlight , because the glass in the microscope objective blocks all UV light below 350 nm . Since our wavelength-sensitivity curves showed potency increasing dramatically as the wavelength shortens , it is likely that the experiment underestimates the animal's true response to sunlight , since it cannot measure the response to UV wavelengths shorter than 350 nm , and sunlight at Earth's surface contains UV wavelengths down to 291 nm [7] . Despite this limitation , when we reproduced the power for direct sunlight long wavelength UV light ( >350 nm ) , we observed a significant locomotion response , whereas the sunlight powers of blue and blue-violet light produced no response ( Figure 5B ) . We therefore hypothesize that the C . elegans light response has evolved as a photophobic response to the ultraviolet light in direct sunlight , but that higher powers of blue-violet and blue light can also evoke the response . To identify the ultraviolet light receptor , we performed a forward genetic screen to look for mutants that are not paralyzed , but are unresponsive to short wavelength light . We tested ∼250 , 000 grandprogeny of ethyl methane sulfonate ( EMS ) –mutagenized animals for their light responses , which represents 24-fold knockout coverage for an average protein [11] and found 20 light-unresponsive ( Lite ) mutants . These mutants represent one major gene target ( 18 alleles ) , which we named lite-1 , and two very rare targets ( 1 allele each ) , which we named lite-2 and lite-3 . The current study focuses on lite-1 . lite-1 null mutants illuminated with optimal blue-violet light often showed no response to the light , but they responded normally to physical stimulation with a platinum wire ( Videos S5 and S6 ) . However , although some of the mutations should completely eliminate LITE-1′s function , even the strongest lite-1 mutants still showed a residual response to light ( Figure 6A ) . Although these data show that worms have a LITE-1–independent mechanism for responding to light , LITE-1 is clearly part of the major light response pathway . LITE-1 also has a major role in the light responses of the paralyzed synaptic signaling mutants . Without LITE-1 , the unc-31 mutant responds only weakly to light , and the Gαq mutant shows no detectable response ( Figure 6B ) . To find the molecular basis of the light response , we mapped the lite-1 mutations to a 146-Kb interval containing 28 genes on the X chromosome ( Figure 7A ) and identified lite-1 using candidate gene sequencing . Among the 18 lite-1 alleles , our genetic screen produced seven splice site mutations , five early stop codons , five amino acid substitutions , and a single base insertion ( Figure 7A and Table S2 ) . LITE-1 ( NCBI Protein Database ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=protein ) accession number NP_509043 . 3 ) is a 439–amino acid protein related to the insect Gustatory receptor ( Gr ) family . Two other homologs complete the C . elegans Gr family: GUR-3 ( NP_509743 . 2; 39% identical ) and EGL-47 ( NP_001023728 . 1; 22% identical in the C-terminal 105 residues ) . In flies , the Gr family has at least 68 members [12–14] . Sequence conservation within this family is low ( 15%–25% ) ; however , a region near the C terminus including the last transmembrane domain is more highly conserved [15–17] . When compared to this family , LITE-1 shows the highest homology to Drosophila Gr28b ( NP_995640 . 1 ) , for which no function has been reported . Although LITE-1 and Gr28b are paralogs rather than orthologs , they are of similar lengths and have similar spacing of their transmembrane domains ( Figure 7B and 7C ) . They are most homologous near their C termini ( 26% identical and 47% similar over a 68–amino acid stretch; Figure S4 ) . Interestingly , transmembrane topology algorithms predict 8-transmembrane domains with extracellular N and C termini for both proteins ( Figure 7C ) . Four of the five LITE-1 missense mutations disrupt residues within the transmembrane domains ( Figure S4 ) . To determine if LITE-1 functions in the nervous system to control light-induced locomotion , we produced a transgenic strain containing the lite-1 cDNA under control of a pan-neuronal promoter in a lite-1 null mutant . Illuminating this strain with a power of blue-violet light that maximizes the wild-type response caused a brief acceleration followed by rapid paralysis ( Video S8 ) . However , reducing the light potency by switching to blue light , and decreasing the power to 16% of the optimal power for wild type , produced a light response in the transgenic strain that did not significantly differ from the wild-type response in initial robustness ( ∼4 . 5-fold increase in locomotion rate for both strains ) ( Figure 8A and Video S8 ) . However , unlike wild type , the transgenic strain could not maintain its high locomotion rate for more than 2–3 min ( NKC , KGM , unpublished data ) , and the data in Figure 8A represent the locomotion rate during the interval from 2–3 min after the start of light exposure . Expressing the lite-1 cDNA only in the cholinergic motor neurons of a lite-1 null mutant conferred light-induced coiling and paralysis at blue light powers optimal for wild type , but reducing the light power 10-fold increased its light-induced locomotion rate ∼3-fold ( Figure 8B and Video S9 ) during the interval from 1 . 5–2 . 5 min after the start of illumination . This suggests that LITE-1 can strongly stimulate the activity of cholinergic neurons; however , it is unclear whether cholinergic neurons overlap with LITE-1′s native site-of-action , or whether LITE-1 is simply sufficient to confer light-induced activation of cholinergic neurons . To investigate LITE-1′s site-of-action , we used a low-power blue-violet laser to specifically illuminate the head or tail of wild-type animals on culture plates ( Figure 9A ) . The initial direction of movement was backward in 48/50 trials in response to head illumination and forward in 50/50 trials in response to tail illumination . The dominant response is forward movement , because the initial direction of movement during whole body illumination was forward in 48/50 trials . Unexpectedly , either whole-body or tail-only illumination rescued the paralyzed synaptic signaling mutants equally well , whereas head-only illumination produced only a weak response in the unc-31 mutant and no response in the egl-30 mutant ( Figure 9B ) . Together with our transgenic rescue data showing a site-of-action in neurons , these data show that LITE-1′s dominant site-of-action with respect to light-induced forward locomotion is one or more tail neurons or tail neuronal processes . Because lite-1 is part of a very large and complex operon with widely dispersed regulatory elements ( unpublished data ) , we were unable to define a rescuing promoter for use in driving a transcriptional GFP reporter to determine where LITE-1 is expressed . We therefore recombineered GFP onto the N terminus of the lite-1 gene in the context of a large fosmid containing all of the presumptive regulatory sequences for lite-1 expression ( Figure 9C ) . The GFP-LITE-1 fosmid transgene is sufficient to rescue the light response of a lite-1 null mutant to near wild-type levels ( Figure S5 ) . In multiple integrated lines made from this tagged fosmid , we detected GFP-LITE-1 in only two neurons , one of which we identified as PVT based on its position , its large elongated cell soma , and its process morphology . PVT produces GFP-LITE-1 in its cell body in the posterior of the animal and exports it via a ventral nerve cord process to a terminal region with large swellings in the nerve ring ( in the head ) where GFP-LITE-1 concentrates ( Figure 9D ) . The other neuron has a process with large swellings of concentrated GFP-LITE-1 in the tail ( Figure 9D ) . Although we were not able to detect the other neuron's cell body , its process belongs to AVG based on the unusual swellings , its unique route through the dorsorectal commissure , and its wavy path and abrupt termination in the mid-tail region ( Zeynep Altun and David Hall , personal communication; Figure S6 ) . Previous studies showed that AVG and PVT both contribute to establishing or maintaining ventral cord axonal tracks in embryos or newly hatched larvae [18–21]; however , their role in older larvae and adults is unknown . Since we can only detect the rescuing LITE-1 transgene in these two neurons , our data suggest that the distal processes of PVT and AVG may be sufficient to function as tail and head light sensors for light-induced forward and reverse locomotion , respectively . However , we cautiously note that laser ablation experiments , in which we individually eliminated each of these neurons , did not affect the forward or reverse light response ( unpublished data ) . We therefore conclude that LITE-1 must function in other neurons as well , where it must be below our level of detection using the GFP-LITE-1 transgene . Given that LITE-1 is concentrated in regions that sense light , and that it has homology to receptors , we hypothesized that it could be a novel ultraviolet light receptor , and we obtained compelling evidence for this . Expressing lite-1 heterologously in body wall and egg laying muscles conferred light responsiveness to a tissue that is normally unresponsive to light . In these transgenic animals , blue-violet light caused a rapid and powerful muscle contraction and shortening of body length . This , combined with activation of the egg laying muscles , caused egg ejection as the increased internal pressure caused by the body contraction forced eggs out of the now open vulva ( Video S7 ) . The light-induced egg ejection only occurred in strains containing the lite-1 cDNA in muscle cells and not in any of the control strains ( Figure 10B ) . By imaging animals as they contracted in the light , and measuring their lengths at each time point , we found that light exposure induces body contraction within 330 ms , and the contraction is complete by 3–4 s ( Figure 10C ) . Interestingly , the light-induced contraction was undiminished in lite-2 or lite-3 mutant backgrounds . Since the lite-2 and lite-3 mutants have Lite phenotypes as strong as lite-1 nulls ( NKC , KGM , unpublished data ) , these data suggest that LITE-1 can function independently in heterologous cells as a light receptor .
Here we report a novel sensory modality in C . elegans: photosensation of ultraviolet light . The fact that C . elegans has a robust UV light response suggests that it often lives on surfaces that could be exposed to direct sunlight . In support of this , a recent study failed to find C . elegans in soil , but instead found it on snails [5] . Other studies have found C . elegans on terrestrial isopods , millipedes and other arthropods , and slugs [6] . According to one theory , dauer juveniles embark on an animal and wait for it to die . They then resume development and propagate on the decomposing body [6] . Interestingly , in quantitative assays , dauer larvae have an exceptionally strong light response . For example , the levels of blue-violet light that maximize the responses of wild-type and unc-31 mutant adults , which result in 3 . 5-fold and 65-fold locomotion rate increases , respectively , cause 20-fold and 1 , 400-fold locomotion rate increases in dauer larvae from the same strains ( NKC , KGM , unpublished data ) . Having such a robust mechanism for escaping damaging doses of short wavelength light would allow C . elegans to avoid the potentially lethal doses of sunlight that may often permeate its above ground habitat . By using forward genetic screens , we discovered a novel molecular solution for ultraviolet light detection that is evolutionarily tailored to activate neurons . LITE-1 has no homology to any of the six known photoreceptor families: rhodopsins , phytochromes , xanthopsins , cryptochromes , phototropins , and BLUF proteins [22] . Each of these families associates with one or more small-molecule chromophores , such as retinal or flavin-based molecules that interact with photons to activate the protein [23–26] . Since LITE-1 has none of the known chromophore interacting domains or residues , it is not clear whether it has a permanently bound chromophore , or whether it binds to a photo-oxidation product produced by short wavelength light . LITE-1′s homology to gustatory receptors that bind small , water-soluble molecules is consistent with either possibility . LITE-1 is one of only several Gr family members for which a loss-of-function phenotype has been described . In Drosophila , recent studies have shown that Gr64a and Gr5a mediate the sensation of most or all types of sugars [27] , and that Gr21a and Gr63a together mediate CO2 ( or bicarbonate ion ) reception [28–30] . Our addition of ultraviolet light to this emerging list is unexpected . LITE-1′s membership in the Gr family provides no clues about how the receptor exerts its effects . The mechanism of action of Gr receptors remains unknown , largely because they have been notoriously difficult to express in heterologous systems . LITE-1 is no exception to this ( unpublished data ) . Although similar in membrane topology to G protein–coupled seven transmembrane receptors , the Gr family is unrelated to G protein–coupled receptors at the sequence level , and there is no direct evidence that Gr receptors exert their effects through G proteins . Perhaps the most interesting aspect of the C . elegans light response is the light-induced rescue of near-paralyzed synaptic signaling mutants , and our finding that tail illumination is sufficient for this rescue . These findings show that a small subset of neurons , possibly as few as one based on our rescuing transgene analysis , can drive a robust , coordinated locomotion response that largely bypasses the synaptic signals that are required for locomotion under normal lighting . A recent study found that mice neurons lacking both CAPS-1 and CAPS-2 ( the mouse orthologs of UNC-31 ) are strongly defective in synaptic vesicle priming and neurotransmitter release when electrically stimulated for short periods of time , but that this defect can be rescued by the large Ca++ increases that accompany extended trains of electrical stimuli [31] . Our finding that LITE-1 induces a powerful light-dependent contraction of muscle cells is consistent with LITE-1 activity ultimately increasing calcium levels in excitable cells . We hypothesize that chronic firing of light circuit neurons during exposure to light could raise internal Ca++ to the level necessary to bypass the need for the Gα pathways . The ∼10-s delay in the response of the unc-31 and Gαq mutants to light stimulation versus 1–2 s for wild type seems consistent with the time that would be needed for the light-induced buildup of a stimulatory signal at each sequential synapse in a polysynaptic circuit stretching from sensory neurons through interneurons and , ultimately , to the motor neurons and the muscle cell . Future studies may be able to harness LITE-1 as a tool for the photoactivation of neurons in living animals and cells or for the investigation of cAMP- and DAG-based synaptic signaling pathways in both invertebrate and vertebrate systems .
See Text S1 for specialized worm culture methods . Locomotion plates were made as described [32] . Other worm culture and manipulation essentially followed previously described methods [33 , 34] . We prepared 24-well culture plates for genetic screens as previously described [9] . Wild-type strains were N2 ( Bristol ) [33] and CB4856 ( Hawaiian ) [35] as indicated . During outcrossing of tax-2 ( p691 ) , we discovered that the original strain contained a second site mutation in lite-3 , which we designated ce360 . Text S1 gives a complete strain list for this study . We used a 405-nm , 5-mW laser with 0 . 9-mm-diameter beam size ( CrystaLaser #BCL-005–405 ) with CL2005 adjustable power supply . The laser head was mounted to a standard laboratory stand using a standard laboratory clamp . We positioned the laser to project onto the culture plate as near to vertical as possible in the center of the field of view while viewing animals through an Olympus SZX-12 stereomicroscope equipped with a 1 . 2× , 0 . 13 numerical aperture plan apochromatic objective . We used the equipment shown and described in Figure S7 to project light of the desired wavelength and power through the stereomicroscope objective onto the culture plate surface . We used the standard Chroma filters shown in Figure 1B to restrict wavelengths to various ranges from UV to red . To measure the light power per mm2 at the culture plate surface , we always placed the detector in a defined location/orientation on the microscope stage due to inherent variation of different regions of the detector surface . We then zoomed to 108× to concentrate a light beam of 9 . 62 mm2 onto a subregion of the detector . Text S1 describe specialized procedures for light-dark locomotion assays , regional illumination locomotion assays , and short– and long–time course locomotion assays . Text S1 provides detailed descriptions of methods for measuring the temperature changes induced by direct illumination of a Checktemp 1 digital temperature probe ( Hanna Instruments ) , methods for measuring the light-induced temperature changes of worm pellets , and methods for testing the effects of temperature on the locomotion rate of the unc-31 null mutant . We produced synchronous populations of unstarved young adults on spread plates ( 4 , 500 per plate × 10 plates ) and adjusted the light source color and power as described above and in Figure S7 . After putting the first plate on the scope stage under low intensity white illumination , we focused on a random area near the center of the plate , and zoomed up to 108× . We then switched the light path to green or blue-violet light ( 1 , 460 μW/mm2 ) and simultaneously started both channels of a timer counting down at 45 s and 5 min 45 s . At 45 s , we counted the number of adults that were fully in the field of view through eyepieces . Over the next 5 min , we then counted the number of animals that subsequently fully entered the field of view . If an animal that was in the field of view partially left the field of view and then fully re-entered , we counted it . We obtained 32-mm-diameter excitation filters from Chroma ( D350/50x for ultraviolet; D436/20x for blue-violet; and HQ470/40x for blue ) . These excitation filters have identical properties to the excitation filters used in the nosepiece of the stereomicroscope to produce colored light on the culture plates . Each filter has a metal frame that fits precisely over the Newport 818-UV detector with the 1 , 000× attenuator attached ( Figure S7 ) . We sealed around the interface of the detector and filter with opaque tape to prevent light leakage . We used the Newport 830-C power meter ( see Figure S7 ) , set on the center wavelength of each filter , to take the readings . The readings were taken at solar noon ± 10 min on February 4 , 2006 , in Edmond , Oklahoma , in full sunlight outdoors . To take readings , we pointed the detector with its filter attached directly at the sun and noted the maximal reading that occurred when the angle was optimal . We then adjusted this reading for the percent transmittance of each filter ( 55% for D350 , 70% for D436 , and 73% for HQ470 ) and normalized the readings to power per mm2 , based on the detector surface area as stated by the manufacturer ( 100 mm2 ) . To isolate Lite mutants , we first produced and plated ∼30 adult F2 grandprogeny of EMS-mutagenized N2 ( wild type ) in each well of 24-well culture plates . We used two methods to screen wells for Lite mutants . For Method 1 , we used a 405-nm , 5-mW laser with 0 . 9-mm-diameter beam size ( CrystaLaser; #BCL-005–405 ) with a CL2005 adjustable power supply . We mounted the laser head to a standard laboratory stand using a standard laboratory clamp . We positioned the laser to project onto the culture plate as near to vertical as possible in the center of the field of view while viewing animals through an Olympus SZX-12 stereomicroscope equipped with a 1 . 2× , 0 . 13 numerical aperture ( NA ) plan apochromatic objective . We set the laser power on 5 . 0 mW . We kept the laser in a fixed position and moved the plate to illuminate the desired region . To screen for Lite mutants , we simply positioned a 24-well plate on the scope stage and , starting with the first well , moved the plate such that an adult animal was illuminated over most of its body by the laser light . If the animal didn't move away within 3 s or so ( usually less ) , we marked its position ( using a pick mark in the bacterial paste ) , picked it to a streak plate for further observation , returned to the marked position , and then moved on to test the next animal , etc . and proceeded systematically to test each adult animal in the well before moving on to the next well . For Method 2 , we screened the wells using a mercury light source and the CFP filter . The light power was ∼1 , 500 μW/mm2 at the screening magnification of 108× . We moved animals into the field of view by moving the plate and noted their light responses ( single or multiple animals at a time ) . We used Methods 1 and 2 about equally . We picked candidate Lite mutants all to the same streak plate for further testing of their light responses . We discarded paralyzed or very sluggish animals , and cloned putative Lite mutants to individual streak plates . After growing 4–5 d at 20 °C to produce adult progeny of the original mutants , we discarded obvious non-Lite mutants , subjectively scored real Lite mutants for the strength of their light responses , and confirmed the homozygosity of the strain or cloned candidate homozygotes if , as in several cases , the original mutant was heterozygous . We carried out the screen in 13 consecutive weekly cycles of screening 4 d/wk with 3–5 people screening ∼2 h/d to achieve the screening goal of 250 , 000 F2s . We outcrossed most Lite mutants at least twice by crossing lite/+ males with dpy-5 ( e61 ) hermaphrodites and then re-isolating Lite animals in the F2 generation . To make the 5× outcrossed lite-1 null mutant reference allele ce314 , we first outcrossed it once through N2 and then repeated the above dpy-5 crossing procedure twice . For some mutants , including ce314 , we also isolated the dpy-5 ( e61 ) ; lite-1 double mutant from the progeny of this cross for use as a marked strain for complementation testing . We complement tested all Lite mutants by crossing Lite/+ males with dpy-5; lite-1 ( ce302 ) or dpy-5; lite-1 ( ce314 ) and scoring adult non-Dpy cross progeny for their Lite responses . During outcrossing , we found that all of the Lite mutants are X-linked . To map lite-1 ( ce302 ) and lite-1 ( ce314 ) to a subregion of the X chromosome , we crossed CB4856 males to lite-1 mutant hermaphrodites and re-isolated putative lite-1 homozygotes in the F2 generation of this cross . We then checked the adult progeny of these animals for homozygosity ( absence of wild-type animals ) , and , upon starvation , we checked the homozygous cultures for various X-linked CB4856 SNPs as described [9] . We used snip-SNPs identified by [35] to map the mutations to a subregion in a manner similar to that previously described [9] . We combined the mapping data for ce302 and ce314 based on our noncomplementation data . After using snip-SNPs to identify recombinants that break left and right of the lite-1 locus ( between ceP173 and ceP171 ) , we tested the recombinants for the presence of other SNPs in the ceP173 – ceP171 interval by restriction analysis or sequencing as described [9] . The SNPs that we used to narrow ce302 and ce314 to the final interval have been previously reported as locations on specific genomic DNA clones [35 , 36] . The specific locations are as follows ( genomic DNA clone/ location on clone; in order from left to right on chromosome ) : ceP173 ( Y23B4A/ 12 , 491 ) ; ceP43 ( T13C5/ 12 , 745 ) ; ceP176 ( T10E10/ 6753 ) ; ceP179 ( T22E5/ 27 , 633 ) ; ceP171 ( F38B6/ 13 , 786 ) . After identifying the ce302 and ce314 mutations from the above analysis , we sequenced genomic DNA from the other 16 lite-1 mutants by making crude plate lysates from a freshly starved streak plate of each strain . We then amplified the lite-1 gene from the lysates using Expand 20 Kb+ and sequenced the resulting products . We constructed the egl-30 ( ad805 ) ; lite-1 ( ce314 ) and unc-31 ( e928 ) ; lite-1 ( ce314 ) double mutants using standard genetic crossing methods . Text S1 describes all of the DNA constructs and transgenes used in this study . In all constructs involving the cloning of PCR fragments , we sequenced the inserts and used clones containing no mutations in the fragment of interest to establish the final plasmid stock . We produced transgenic strains bearing extrachromosomal arrays by the method of Mello et al . [38] . We used pBluescript carrier DNA to bring the final concentration of DNA in all injection mixtures to 175 ng/μl . All injection mixtures in this study included the co-transformation marker plasmids containing the same promoter as the experimental DNA , but hooked to GFP instead . The injection mixtures , plasmid concentrations , and host strains of all of the transgenic strains used in this study are listed in Text S1 . We produced integrated arrays using previously a described method [10] , except we screened cultures for 100% transmittance of GFP , and we used 7 , 200 Rads of γ irradiation . We outcrossed ceIs37 twice to wild type , keeping versions with and without lite-1 ( ce314 ) . We confirmed the homozygous presence and absence of the ce314 mutation by PCR and sequencing . We recombineered GFP onto the N terminus of the lite-1 gene on the fosmid WRM062dF04 ( Geneservice ) to make the new fosmid KG#319 using a modification of a previously described method [37] . Briefly , we substituted the plasmid pRedFlp4 ( gift of Mihail Sarov ) for pRedFlp as a source of the Red/ ET recombination proteins and Flp recombinase . pRedFlp4 substitutes the HgrR hygromycin resistance gene for the AmpR gene in pRedFlp . Whereas the original protocol required Amp/ Trimethoprim to select for pRedFlp transformation into the fosmid host , the modified method uses only Hygromycin ( we used HygroGold ( InvivoGen at 200 μg/ ml ) ) . Trimethoprim cannot be used for transforming into fosmid hosts since fosmids contain the DHFR gene that confers Trimethoprim resistance . In addition , we modified the pR6KGFP plasmid to allow N-terminal or internal fusions to the gene of interest . The new plasmid ( pR6KGFPX ) lacks a GFP stop codon , contains a two-nucleotide “GG” insertion immediately after the 34-bp FRT and before the reverse primer homology region to maintain reading frame with the downstream protein , and changes an in-frame TGA stop codon in the reverse primer homology region to GGA . Finally , we did not subclone the GFP-tagged lite-1 gene from the fosmid into the pPUB vector but instead left it as a GFP-tagged fosmid . The final tagged-product structure is as follows: ATG of lite-1 gene → 6 forward primer codons → ATG of GFP → GFP coding sequences minus stop codon → 1 copy of FRT → GG nucleotides → reverse primer codons → the rest of the lite-1 gene and genomic sequences on the fosmid . We confirmed the final fosmid structure by PCR to detect insertion of GFP and to confirm that no fosmid lacking GFP was present in the clone , and we sequenced the insertion region . For the egg ejection assay , we picked 30 gravid adults from growing cultures to a standard spread plate , spacing them in different regions of the plate ( see Text S1 for “growing cultures” and “spread plate” definitions ) . Each adult carried ≥ 6 eggs . If the strain contained a GFP-marked extrachromosomal array , we required that the adults show uniform green in their muscle cells . To avoid making the GFP-positive animals eject their eggs while choosing them , we zoomed out as far as possible when the GFP light was on ( to reduce its intensity ) and worked quickly . As soon as we identified a uniform green , GFP-positive animal , we switched to standard white light to examine egg number and to make the transfer . With no plate on the stage , we adjusted the total CFP light power to 14 . 6 mW at 108× magnification , and then switched back to normal white light . We removed the lid from the plate containing the 30 animals to be assayed , focused on the first animal , zoomed to 108× magnification , and centered the field of view on the animal's vulva . We then turned up the white stage light such that we would be able to clearly see the animal after switching to CFP light , and then simultaneously started the timer and slid the CFP filter into place . At the end of the 20 s , we switched the light back to white stage light and noted whether any eggs were dumped during the 20-s stimulus . We picked off the animal we had just assayed , re-checked/ adjusted the light power , and repeated the assay on the other 29 animals for each strain . To take the time course of light-induced muscle contraction , we produced growing cultures of each strain on spread plates . Under normal white stage light illumination , we transferred five gravid adults to a standard spread plate for each strain . We required that each adult be carrying ≥ 6 eggs . With no plate on the stage , we adjusted the CFP light power to 32 . 5 mW total power at 108× magnification . We then increased the white stage light as far as possible to shorten exposure time which , under these conditions , was ∼50 ms . After choosing the first animal to assay on the culture plate ( plate lid removed ) , we zoomed to 75× and centered the animal , which took up ∼75% of the camera screen field of view . We then simultaneously switched to the CFP light path and clicked the “Acquire” button to collect a 10-s time course of images , spaced 330 ms apart , using an ORCA-AG camera and Metamorph Premier software ( Version 6 . 3 r1 ) . We used the multi-line region tool and clicked along the midline of each animal from the tip of the nose to a defined point near the end of the tail . We logged length measurements in units of pixels and used an Excel spreadsheet to convert data to the final units of “percent unstimulated length” for each time point . We collected fluorescent images using a Nikon Eclipse TE2000-E inverted microscope equipped with a 60× 1 . 4 N . A . oil planapochromat objective ( CF160-type ) , a 1 . 5× tube lens , a motorized linear-encoded z-drive , and a motorized filter turret containing a Semrock GFP filter cubes . Our illumination source was an X-Cite 120 illuminator ( EXFO ) , and we captured 12-bit images with an ORCA-AG camera ( Hamamatsu ) controlled by Metamorph Premier software ( Version 6 . 3 r1 ) . We further processed z-series stacks of images by the Adaptive PSF Blind Deconvolution method ( 10 iterations with low noise level setting ) using AutoDeblur Gold CWF software ( ImageQuant ) . We then used AutoDeblur to resize the stacks 2-fold in the X and Y dimensions , and then used Metamorph to produce maximum projections of the stacks , adjust the scaling and generate 8-bit images for display .
|
In all of nature , scientists have discovered only six different mechanisms by which organisms sense light , and only one of these mechanisms can detect ultraviolet light ( the rhodopsins that sense ultraviolet light in non-mammalian vertebrates ) . The widely studied model organism Caenorhabditis elegans has none of the known light transduction systems , but we discovered that C . elegans has a robust locomotory response to ultraviolet light . C . elegans may use this light response to escape damaging or lethal doses of sunlight . Ultraviolet and other shortwave light , such as violet and blue wavelengths , drive locomotion by bypassing two critical signals , cyclic adenosine monophosphate ( cAMP ) and diacylglycerol ( DAG ) , that neurons use to shape and control behaviors . C . elegans mutants lacking these signals are paralyzed and unresponsive to harsh physical stimuli in ambient light , but short-wavelength light rapidly rescues their paralysis and restores greater-than-normal levels of coordinated locomotion . This astonishing light response is mediated by a novel ultraviolet light receptor that acts in neurons . Our results reveal a novel molecular solution for ultraviolet light detection and an unusual sensory modality in C . elegans that is unlike any previously described light response in any organism .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience"
] |
2008
|
A Novel Molecular Solution for Ultraviolet Light Detection in Caenorhabditis elegans
|
Although prostate cancer typically runs an indolent course , a subset of men develop aggressive , fatal forms of this disease . We hypothesize that germline variation modulates susceptibility to aggressive prostate cancer . The goal of this work is to identify susceptibility genes using the C57BL/6-Tg ( TRAMP ) 8247Ng/J ( TRAMP ) mouse model of neuroendocrine prostate cancer . Quantitative trait locus ( QTL ) mapping was performed in transgene-positive ( TRAMPxNOD/ShiLtJ ) F2 intercross males ( n = 228 ) , which facilitated identification of 11 loci associated with aggressive disease development . Microarray data derived from 126 ( TRAMPxNOD/ShiLtJ ) F2 primary tumors were used to prioritize candidate genes within QTLs , with candidate genes deemed as being high priority when possessing both high levels of expression-trait correlation and a proximal expression QTL . This process enabled the identification of 35 aggressive prostate tumorigenesis candidate genes . The role of these genes in aggressive forms of human prostate cancer was investigated using two concurrent approaches . First , logistic regression analysis in two human prostate gene expression datasets revealed that expression levels of five genes ( CXCL14 , ITGAX , LPCAT2 , RNASEH2A , and ZNF322 ) were positively correlated with aggressive prostate cancer and two genes ( CCL19 and HIST1H1A ) were protective for aggressive prostate cancer . Higher than average levels of expression of the five genes that were positively correlated with aggressive disease were consistently associated with patient outcome in both human prostate cancer tumor gene expression datasets . Second , three of these five genes ( CXCL14 , ITGAX , and LPCAT2 ) harbored polymorphisms associated with aggressive disease development in a human GWAS cohort consisting of 1 , 172 prostate cancer patients . This study is the first example of using a systems genetics approach to successfully identify novel susceptibility genes for aggressive prostate cancer . Such approaches will facilitate the identification of novel germline factors driving aggressive disease susceptibility and allow for new insights into these deadly forms of prostate cancer .
Prostate cancer is a common disease , and it is estimated that approximately 233 , 000 new cases will be diagnosed in in the United States alone in 2014 [1] . However , it typically runs an indolent course , with most men succumbing to unrelated diseases . This is reflected in the low prostate cancer-specific mortality , with ∼29 , 000 men dying from this disease in the same period in the US . Currently , the assessment of prognosis relies heavily upon the evaluation of traditional clinical and pathological variables , and is fraught with inaccuracies . These inaccuracies lead to over-treatment of prostate cancer , which causes unnecessary suffering resulting from aggressive therapeutic interventions , and represents a significant public health burden . Accordingly , there is a pressing need to improve the molecular characterization of prostate cancer , in order to facilitate an improved prognostic accuracy and to detect men at increased risk of developing aggressive , fatal forms of this disease . One such feature that is garnering increased attention is the emergence of prostate tumors with a neuroendocrine ( NE ) phenotype [2] . Small cell NE prostate carcinoma is a rare histological subtype , which comprises 0 . 3% to 1 . 0% of all prostate malignancies [3] . Compared to prostate adenocarcinoma , which is the most common histological subtype , it typically runs a more aggressive course and is associated with visceral metastasis and poor outcomes ( median survival = 10 . 0 months vs . 125 . 0 months for adenocarcinoma ) [4]–[6] . However , it is becoming increasingly apparent that prostate adenocarcinomas , which comprise 90–95% of all prostatic neoplasms [3] , with extensive NE characteristics are associated with a particularly poor prognosis . Specifically , autopsy studies have demonstrated that at least 20–30% of end-stage prostate adenocarcinomas exhibit a significant degree of NE differentiation ( NED ) [7] , [8] . Furthermore , this NE phenotype is particularly prevalent in patients treated with androgen deprivation therapy ( ADT ) , and the appearance of recurrent tumors with NE characteristics following ADT is associated with castrate resistance , visceral metastasis , and death [8] , [9] . In addition , the incidence of prostate carcinomas with a prominent NE phenotype is expected to increase as use of second generation ADTs ( e . g . , enzalutamide , abiraterone ) becomes more widespread , such that NED will likely represent a new mechanism of therapeutic resistance [10] . The pathogenesis of NED remains unclear . Recent studies have demonstrated that RB1 loss is a crucial element of the pathogenesis of NE prostate cancer [10] . Additionally , these tumors are often associated with loss of androgen receptor expression , activation of the PI3K pathway , and amplification of N-MYC and AURKA [2] . However , like all forms of prostate cancer , the initiation and progression of NED will be influenced by host germline variation . Genome-wide association studies ( GWAS ) have revolutionized our understanding of how heritable factors influence prostate cancer development , and have facilitated the identification of multiple loci associated with aggressive disease ( e . g . , [11] ) . Yet GWAS have not been able to explain the complete influence of heritability on disease susceptibility . Therefore , alternative approaches for defining susceptibility will be required to augment GWAS and to fully understand how the germline modifies susceptibility to aggressive phenotypes like NED . The work presented here utilizes a systems genetics approach , which involves the integration of lines of evidence from a mouse model of aggressive prostate cancer and several human prostate cancer datasets to identify novel genes associated with aggressive disease ( Figure 1 ) . Candidate genes are initially identified using the C57BL/6-Tg ( TRAMP ) 8247Ng/J ( TRAMP ) mouse model of neuroendocrine prostate cancer , which develops extensive tumorigenesis and metastasis by 30 weeks of age [12]–[14] . Our earlier work demonstrated that disease aggressiveness in the TRAMP mouse is substantially modified by host genetic background [15] . This earlier study involved performing a ‘strain survey’ experiment where wildtype TRAMP mice were bred to one of eight inbred strains of mice . Characterization of disease aggressiveness traits in the eight resulting F1 strains revealed substantial strain-specific differences in prostate tumorigenesis and metastasis . Since the SV40 T antigen was expressed at equal levels and at the same developmental time point in each of the eight F1 strains , we concluded that the observed phenotypic differences in disease aggressiveness were a consequence of germline variation [15] . To explore the origins of this , an F2 mapping panel involving TRAMP and NOD/ShiLtJ , which is a strain that is highly susceptible to developing aggressive tumorigenesis , was generated . These mice were used to map quantitative trait loci ( QTLs ) associated with aggressive NE prostate cancer . Following this , QTL candidate genes were nominated from microarray gene expression data derived from ( TRAMP × NOD/ShiLtJ ) F2 tumors through a combination of expression QTL ( eQTL ) mapping and gene expression-trait correlation analysis . The relevance of these QTL candidate genes to aggressive forms of human prostate cancer were explored through two concurrent approaches: first , by correlating their expression levels with disease free survival ( DFS ) in two prostate tumor gene expression cohorts; and second , by analyzing a human GWAS dataset to correlate the frequencies of QTL candidate gene single nucleotide polymorphisms ( SNPs ) with clinical markers of disease aggressiveness . This approach , which is novel to the field of prostate cancer to the best of our knowledge , facilitated the identification of three novel aggressive prostate cancer susceptibility genes: CXCL14 , ITGAX , and LPCAT2 .
Earlier work demonstrated that germline variation present in the NOD/ShiLtJ strain renders ( TRAMP × NOD/ShiLtJ ) F1 male mice significantly more susceptible to aggressive prostate tumorigenesis [15] . Specifically , ( TRAMP × NOD/ShiLtJ ) F1 males displayed significantly increased primary tumor burden , local metastasis to regional lymph nodes , and distant metastasis to visceral organs including the lung , liver and kidneys compared to wildtype TRAMP C57BL/6J mice . Therefore , we hypothesized that the introduction of germline polymorphism through breeding will allow for the mapping of QTLs associated with aggressive tumorigenesis in the TRAMP mouse . To investigate this hypothesis , a ( TRAMP × NOD/ShiLtJ ) F2 intercross population consisting of 228 transgene-positive males was developed . Mice were aged until 30 weeks of age or until humane endpoints were achieved . As expected , substantial variation in aggressive prostate cancer phenotypes was observed in these F2 mice ( Table S1 ) . Of particular note , it was clear that there was a strong level of interdependency between tumor related-traits ( primary tumor burden , seminal vesicle tumor burden ) and traits commonly associated with survival in human prostate cancer ( reviewed in [16]; e . g . , age at death , distant metastasis free survival [DMFS] , presence or absence of lymph node metastasis , lymph node metastasis burden; Figure S1 ) in the F2 mapping population . As would be expected in humans , larger primary tumors were positively correlated with a younger age of death , a substantially reduced DMFS , an increased risk of lymph node metastasis , and an increased lymph node metastasis burden ( Figure S1A–D ) . The converse , however , was true for seminal vesicle tumor burden , which was negatively correlated with the same traits ( Figure S1E–H ) . Accordingly , there was a significant negative correlation between primary tumor burden and seminal vesicle tumor burden ( Pearson's r = −0 . 41 , P = 7 . 40×10−11; Figure S2 ) . Earlier work with the TRAMP model has demonstrated that these seminal vesicle tumors represent a form of epithelial-stromal tumor that resemble phyllodes tumors [17] , [18] , which are an uncommon neoplasm of uncertain malignant potential in humans [19] . However , our data clearly demonstrate that mice with greater seminal vesicle tumor burden , and thus a lower primary tumor burden , are less prone to more aggressive disease forms . Therefore , the germline polymorphisms driving lower seminal vesicle tumor burden and higher primary tumor burden may be associated with a predisposition for more aggressive disease . QTLs were mapped in ( TRAMP × NOD/ShiLtJ ) F2 males by performing a genome scan using 666 informative SNPs . Analyses were performed in J/qtl [20] using a single-locus model of inheritance . QTLs were considered statistically significant when genome-wide α<0 . 05 . For metastasis-related traits , a total of four QTLs were observed: two for DMFS ( chromosome 1 [logarithm of odds score ( LOD ) = 3 . 93] and chromosome 11 [LOD = 3 . 97] ) ; one for lymph node metastasis burden on chromosome 13 ( LOD = 4 . 69 ) ; and one on chromosome 11 for liver surface metastasis count ( LOD = 4 . 01 ) . A total of five QTLs were observed for tumor-related traits: one for primary tumor burden on chromosome 13 ( LOD = 4 . 86 ) ; and four for seminal vesicle tumor burden ( chromosome 2 [LOD = 5 . 01]; chromosome 4 [LOD = 5 . 24]; chromosome 8 [LOD = 4 . 22]; and chromosome 17 [LOD = 5 . 20] ) . Finally , two QTLs were evident for age of death , on chromosome 7 ( LOD = 4 . 35 ) and chromosome 8 ( LOD = 4 . 65 ) . QTL data are summarized in Table 1 and Figure S3 . As is typical with F2 intercross populations , the confidence intervals of QTLs , as defined by the 2-LOD drop beyond the peak region of linkage , are broad , and each QTL will encompass many hundreds of genes . Additionally , it should be noted that these eleven QTLs in fact represent nine genomic regions with overlap of age of death and seminal vesicle tumor burden QTLs on chromosome 8 , and nodal metastasis burden and primary tumor burden loci on chromosome 13 . Integration of germline variation and transcriptome data is a well-established means of nominating QTL candidate genes that influence a given trait through expression-related mechanisms [21] , [22] . Specifically , QTL candidate gene transcripts identified through this approach will possess both of the following: 1 ) they will exhibit a proximal expression QTL ( eQTL ) , which we define as an eQTL mapping ≤1 megabase ( Mb ) upstream or downstream of the transcription start site since 95% of enhancers are predicted to target transcripts within this range [23]; and 2 ) their expression levels will be correlated with the trait of interest . Only the expression of transcripts physically located within QTLs were considered in these analyses . We hypothesize that QTL candidate genes modifying susceptibility to aggressive prostate tumorigenesis through transcriptional-related mechanisms in ( TRAMP × NOD/ShiLtJ ) F2 males will possess both of these characteristics . To identify QTL candidate genes in this manner , microarray analysis was performed to analyze patterns of global gene expression in all available F2 prostate tumors ( n = 126 ) . Expression QTL mapping was performed using Matrix eQTL [24] . Benjamini-Hochberg false discovery rates ( FDR ) were calculated to correct for multiple testing [25] , with an FDR <0 . 05 used as the threshold for significant eQTLs . A total of 9 , 510 eQTLs were evident in TRAMP × NOD F2 tumors , of which 854 were defined as proximal eQTLs ( Table S2 ) and 8 , 656 as distal and/or trans-eQTLs ( Table S3 ) . However , of the 8 , 656 distal and/or trans-eQTLs , only 1 , 560 associations were between a SNP and transcript on different chromosomes ( i . e . , a true trans-eQTL ) . The high number of distal eQTLs , which reside on the same chromosome as the cognate transcript but outside of the 1 Mb window for mapping proximal eQTLs most likely reflects the low level of recombination typically observed in F2 populations . The genomic distributions of eQTLs relative to their cognate transcript are illustrated in Figure 2 . Of the 854 proximal eQTLs identified , 147 resided within the 2-LOD confidence intervals of the eleven aggressive disease QTLs described in Table 1 . To further increase the stringency of QTL candidate gene identification , the expression levels of all transcripts within the boundaries of each of 11 aggressive disease QTLs were correlated with the QTL trait ( Tables S4–S14 ) . Using the Benjamini-Hochberg FDR method [25] to correct for multiple testing ( FDR <0 . 05 ) , 35 high-confidence QTL candidate genes were identified , each of which exhibited a statistically significant proximal eQTL and correlation between transcript expression and the trait of interest ( Table 2 ) . Having used a highly stringent analytical approach to identify 35 aggressive tumorigenesis susceptibility genes in ( TRAMP × NOD/ShiLtJ ) F2 males , we aimed to determine whether the human orthologs of these genes play a similar role in human prostate cancer . Of the 35 QTL candidate genes identified in ( TRAMP × NOD/ShiLtJ ) F2 males , 29 had a human ortholog ( Table 2 ) . The 6 transcripts with no direct ortholog were omitted from further analyses owing to their probable irrelevance to human prostate cancer . We hypothesized that if the human orthologs of the remaining 29 QTL candidate genes play a similar role in aggressive prostate cancer susceptibility , they should exhibit the same characteristics that facilitated their identification in ( TRAMP × NOD/ShiLtJ ) F2 males . Specifically , their expression levels in primary tumors should be associated with aggressive prostate cancer , and they should be in linkage disequilibrium ( LD ) with germline SNPs associated with susceptibility to aggressive prostate cancer development . To address the first of these , the expression levels of QTL candidate genes were examined in two publicly-accessible prostate cancer gene expression datasets using cBioPortal for Cancer Genomics ( http://www . cbioportal . org/ [26] , [27] ) , which is a web-based resource that comprises multi-dimensional cancer genomics data for numerous cancer subtypes . We initially focused on a prostate cancer dataset provided by The Cancer Genome Atlas ( TCGA ) , which is comprised of a sufficient number of subjects to facilitate adequately powered survival analyses ( TCGA [Provisional] ) . Here , cBioPortal reports static levels of gene expression in individual prostate tumors from this RNA-seq based dataset . Findings in TCGA ( Provisional ) cohort were confirmed in a second microarray-based dataset ( Prostate Oncogenome Project [GSE21032]; [28] ) . Stepwise logistic regression analysis was performed to test the association between the expression levels of each of the 29 QTL candidate genes in the two datasets and dichotomized clinical variables , based on the common disease features reported for both cohorts ( Figure S4 ) . The comparisons of aggressive prostate cancer clinical variables used in logistic regression analyses , as well as the results of these tests are shown in Table 3 . In TCGA ( Provisional ) cohort , the expression levels of three genes were positively correlated with aggressive disease characteristics . Specifically , the expression levels of CXCL14 ( odds ratio [OR] = 1 . 62 [95% confidence interval 1 . 10–2 . 38] ) and RNASEH2A ( OR = 2 . 17 [1 . 04–4 . 52] ) were associated with disease recurrence; and LPCAT2 ( OR = 1 . 44 [1 . 02–2 . 03] ) with a higher pathological stage . In the GSE21032 cohort , the expression levels of three genes were associated with an increased risk of aggressive disease and two genes identified as having a protective effect . Specifically , the expression levels of CXCL14 were associated with a higher pathological stage ( OR = 1 . 75 [1 . 19–2 . 59] ) . Divergent effects were observed for tumor Gleason score , with two genes being associated with a higher Gleason score ( ITGAX; OR = 3 . 87 [1 . 88–7 . 56] and ZNF322; OR = 2 . 26 [1 . 27–4 . 02] ) and two genes with a Gleason score <7 ( CCL19; OR = 0 . 46 [0 . 24–0 . 88] and HIST1H1A; ( OR = 0 . 45 [0 . 24–0 . 86] ) . To test the correlation between candidate gene expression and disease recurrence , genes implicated in aggressive disease development in logistic regression analyses performed in TCGA ( Provisional ) and GSE21032 cohorts were combined to create two gene sets: 1 ) a set of five genes associated with an increased propensity for aggressive disease development ( CXCL14 , ITGAX , LPCAT2 , RNASEH2A , and ZNF322 ) ; and 2 ) a set of two genes with a protective effect ( CCL19 and HIST1H1A ) . The expression levels of transcripts within these two gene sets were correlated with disease free survival ( DFS ) using Kaplan-Meyer survival analysis in the TCGA ( Provisional ) cohort . Specifically , DFS was compared between cases with higher or lower levels of expression of one or more gene in either of the two gene sets to those cases with normal levels of expression of the same genes . Eighteen percent ( 45/246 ) of cases in TCGA ( Provisional ) dataset exhibited divergent levels of one or more of the five genes positively correlated with aggressive disease development ( Figure 3A ) . Strikingly , the directionality of expression was significantly higher than average for each of the five genes in all 45 cases . Accordingly , Kaplan-Meyer survival analysis demonstrated that higher than average expression levels of one or more of these genes was associated with a poorer DFS ( log-rank P = 0 . 025; Figure 3B ) . To confirm the findings from this cohort , survival analysis was performed in the GSE21032 dataset by comparing DFS in patients with higher than average levels of the five candidate genes compared to all other cases . In the GSE21032 dataset , 12% of cases ( 16/131 ) exhibited exclusively higher than average levels of expression of one or more candidate genes ( Figure 3C ) . As was the case in TCGA ( Provisional ) dataset , higher than average levels of expression of these genes was associated with a poorer DFS ( log-rank P = 0 . 010; Figure 3D ) . Analysis of both datasets in cBioPortal showed that none of the cases with higher levels of candidate gene expression in either cohort exhibited either copy number alteration or somatic mutations of these genes . This implies that candidate gene copy number alteration and/or somatic mutations likely have no influence upon DFS in these datasets . Similar analysis in larger prostate cancer datasets will be required to confirm or refute whether the observed associations in the TCGA ( Provisional ) and GSE21032 are correlated with primary tumor copy number variation . Finally , the expression levels of the two genes that were negatively correlated with aggressive disease on logistic regression were not correlated with DFS in either cohort . Our QTL mapping strategy demonstrates that QTL candidate gene germline variation is associated with aggressive tumorigenesis in the TRAMP mouse . To evaluate whether this is the case for the human orthologs of these genes , SNP allele frequencies were evaluated in a publicly available human prostate cancer GWAS dataset . Specifically , these analyses were performed using the Cancer Genetic Markers of Susceptibility ( CGEMS ) GWAS , which consists of 1 , 172 prostate cancer patients and 1 , 157 controls of European ancestry from the Prostate , Lung , Colon and Ovarian ( PLCO ) Cancer Screening Trial [29] , [30] . This relatively well-studied resource has facilitated the identification of novel loci associated with prostate cancer , including a second prostate cancer risk locus at 8q24 [31] . Given that we hypothesized that QTL candidates modulate prostate cancer aggressiveness but not prostate cancer initiation , controls were omitted from analyses . The CGEMS cohort is well suited for this purpose , with the case cohort subdivided into non-aggressive ( Gleason score <7 and stage <III; n = 484 ) and aggressive ( Gleason score ≥7 or stage ≥III; n = 688 ) cases . In addition to these clinical characteristics , case-case analyses were performed for the additional aggressive disease variables shown in Table 4 . These variables related to the size or direct extent of the primary tumor ( pros_stage_t ) , local metastasis to lymph nodes ( pros_stage_n ) and distant metastasis ( pros_stage_m ) . We elected to include these variables to more closely reflect the phenotypes used to identify QTL candidate genes in ( TRAMP × NOD/ShiLtJ ) F2 mice . In the study , 1 , 317 SNPs were mapped within a 100 kb radius of the 29 QTL candidate genes were tested in the CGEMS cohort . Analysis of aggressive vs . non-aggressive disease phenotypes were performed as per the comparisons described in Table 4 . Correction for multiple testing was performed using permutation testing ( n = 10 , 000 permutations ) . Fourteen of the 29 candidate genes exhibited evidence for association with clinical characteristics of aggressive prostate cancer ( Table 5 ) . Most notably , SNPs in three of the five genes associated with poor clinical outcomes in TCGA ( Provisional ) and GSE21032 prostate cancer gene expression datasets ( CXCL14 , ITGAX , and LPCAT2 ) were all associated with aggressive prostate cancer: for CXCL14 , associations were evident between rs801564 and metastasis to regional lymph nodes ( permutation P = 0 . 011; OR = 1 . 05 [1 . 01–1 . 09] ) , and between rs10515473 and Gleason score at prostatectomy ( permutation P = 0 . 001; OR = 0 . 72 [0 . 59–0 . 88] ) ; for ITGAX , an association was apparent between rs8047538 and Gleason score at prostatectomy ( permutation P = 0 . 007; OR = 1 . 33 [1 . 08–1 . 62] ) ; and for LPCAT2 , associations were evident between rs3764263 and primary tumor stage ( permutation P = 0 . 009; OR = 1 . 61 [1 . 12–2 . 31] ) , between rs289707 and biopsy Gleason score ( permutation P = 0 . 002; OR = 1 . 22 [1 . 07–1 . 38] ) , between rs2289119 and metastasis to regional lymph nodes ( permutation P = 0 . 009; OR = 1 . 06 [1 . 01–1 . 10] ) , and between rs17369578 and best Gleason score available ( permutation P = 0 . 003; OR = 1 . 41 [1 . 13–1 . 77] ) . Manhattan plots for all relevant genomic regions are shown in Figure S5 . Additionally , rare haplotypes ( <1% frequency ) in LD with three QTL candidate genes were associated with clinical markers of prostate cancer aggressiveness ( Table S15 ) .
A systems genetics approach has been employed in this study to identify three novel susceptibility genes for aggressive prostate cancer , and to the best of our knowledge , this is the first study of its type to use this approach in this form of cancer . The three high priority candidate genes identified in QTL mapping studies using the TRAMP mouse model have diverse cellular functions ( Table 6 ) , and have not been previously implicated as germline susceptibility genes for aggressive prostate cancer . Functional characterization of these genes to clarify their role in aggressive prostate cancer is therefore of much importance . However , given the strength of the genetic and genomic data implicating each of these genes in aggressive tumorigenesis , we argue that the required depth of such functional characterization is beyond the scope of the current study . Nevertheless , other studies support the role of some of these genes in aggressive tumorigenesis . For example , higher levels of expression of LPCAT2 are observed in a diverse range of tumors , notably breast and cervical carcinomas [32] . Additionally , a linkage study demonstrated that CXCL14 resides in a risk locus for aggressive prostate cancer in the 5q31 region [33] , and higher levels of this gene have been observed in tumors with a higher Gleason score [34] . Concomitantly , over-expression of CXCL14 in fibroblasts stimulates tumor angiogenesis and growth of prostate cancer cells [35] through activation of NOS1-derived nitric oxide signaling pathways [36] . These findings are in keeping with the results of our survival analyses of TCGA ( Provisional ) and the GSE21032 cohorts , which demonstrated that higher than average levels of expression of CXCL14 in bulk tumor tissue is associated with an increased risk of recurrence . Identification of these novel aggressive prostate cancer susceptibility genes has been facilitated through use of the TRAMP mouse model . However , the NE histological phenotype of tumors and the use of the non-physiological SV40 T-antigen to induce tumorigenesis have led to criticism of TRAMP [37] . The validity of these criticisms is , however , being increasingly questioned , particularly in light of the probable increase in incidence of human NE prostate tumors induced by increasingly efficacious ADTs [38] . The TRAMP model can therefore be viewed as a powerful tool to study the pathogenesis of NE forms of aggressive , castrate-resistant disease . Additionally , the SV40 T-antigen directly inactivates Rb and p53 [39] , and the aggressive disease seen in TRAMP mice therefore mimics somatic mutation of these potent tumor suppressors . We do , however , acknowledge that observations from the TRAMP model are sometimes not directly comparable to human prostate cancer . An example from the current study would be the association of higher levels of Cxcl14/CXCL14 being negatively associated with primary tumor burden in ( TRAMP × NOD/ShiLtJ ) F2 mice but positively correlated with disease recurrence in humans . Additionally , the traits used to nominate candidate genes in ( TRAMP × NOD/ShiLtJ ) F2 mice frequently differ from the associated aggressive disease traits observed in human populations , as illustrated in Table 6 . We therefore regard the TRAMP model as a powerful tool for nominating aggressive disease modifiers in a generalized sense , and the integration of different lines of evidence from human prostate cancer populations is of critical importance for deciphering the relevance of observations derived from mice . The integration of these different lines of evidence from human prostate cancer datasets to validate findings from our genetic screen in the TRAMP mouse has proven a pivotal element of this study . There are , however , a number of aspects of our analysis of the CGEMS GWAS data that warrant further discussion . First , we acknowledge that our use of a permutation test does not fully resolve the issue of correcting for type I errors . Rather , permutation testing has allowed us to report P-values that are both more stable and accurate than uncorrected values . Second , we also recognize that a genome-wide level of significance was not achieved with any of the SNPs characterized in the CGEMS GWAS dataset . One probable reason for this is the limited statistical power of the case-case analysis performed here , which reflects the relatively small study population . Validation of these findings in additional prostate cancer cohorts is therefore vital . However , this lack of genome-wide significance may reflect one of the few limitations of GWAS . Specifically , although GWAS have revolutionized our understanding of complex trait susceptibility , they have not yet been able to explain the complete influence of heritability on disease susceptibility . This is true of prostate cancer , where all of the variants thus far identified by GWAS are estimated to explain less than one third of familial disease risk [11] , [40] . It has been postulated that a possible reason for this is that biologically relevant modifiers that achieve the P<0 . 05 nominal level of significance are being missed since they do not reach the necessarily stringent level of genome-wide significance [41] . Therefore , alternative methodologies to augment GWAS , including the types of approaches described here , may facilitate characterization of some of this ‘missing heritability’ . Thus , the evidence for association between QTL candidate gene SNPs and aggressive disease development from these GWAS data in this study is insufficient in isolation . However , the power of these GWAS analyses is derived from consideration in unison with the mouse and human gene expression data . In summary , we have identified CXCL14 , ITGAX and , LPCAT2 as novel susceptibility genes for aggressive prostate cancer development . This is the first study of its type to address the influence of germline polymorphism on tumor progression and metastasis in prostate cancer using systems genetics approach . Additionally , this approach has identified novel modifiers of aggressive prostate cancer that might not be readily apparent through human association studies . Knowledge of these variants will allow for more accurate determination of a patient's risk of metastasis , thus improving prognostic accuracy and facilitating more personalized treatments .
C57BL/6J-Tg ( TRAMP ) 824Ng/J ( TRAMP ) and NOD/ShiLtJ mice were obtained from The Jackson Laboratory ( Bar Harbor , ME ) . F1 mice were generated by crossing TRAMP females , which were hemizygous for PB-TAg transgene ( Tg ) , to NOD/ShiLtJ males . F2 mice were generated by crossing Tg+ F1 females with Tg- F1 males . All animals were handled , housed and used in the experiments humanely in accordance with the NHGRI Animal Care and Use Committee guidelines . All work was performed under Animal Study Protocol G-09-2 . Mouse tail genomic DNA was extracted from F1 progeny with the HotSHOT method [42] for genotyping analysis . PCR screening was performed as described [14] to identify the hemizygous PB-TAg transgene positive F1 and F2 mice . As described previously in [15] , ( TRAMP × NOD/ShiLtJ ) F2 male mice were sacrificed by pentobarbital overdose at 30 weeks of age or humane endpoint , whichever was achieved first . Humane experimental endpoints for this study were rapid weight loss , hunched posture , labored breathing , trauma , impaired mobility , dysuria , or difficulty in obtaining food or water . Prostate tumor , seminal vesicles , lungs , liver , and lymph nodes were harvested from ( TRAMP × NOD/ShiLtJ ) F2 males . Prostate tumor and seminal vesicles were weighed to quantify tumor burden . Visible , enlarged lymph nodes in para-aortic region were weighed to quantify metastatic lymph node burden . Lungs were collected to determine isolated tumor cell infiltrates in lung parenchyma and microscopic metastatic lesions . Other organs displaying macroscopic metastatic lesions through gross observation were also collected for histology . These collected tissues were fixed in buffered formalin ( 10% w/v phosphate buffered formaldehyde , Fisher Scientific ) overnight and then transferred to 70% ethanol . Fixed tissues were embedded in paraffin , sectioned to a thickness of 4 µm and stained with hematoxylin and eosin ( H&E ) . Histology slides were scanned with Scanscope Digital microscope ( Aperio , Vista , CA ) . Genomic DNA was extracted from F2 tail biopsies using a Gentra Puregene DNA Extraction Kit ( Qiagen , Valencia , CA ) , per the manufacturers protocol . Five microliters of DNA at 75 ng/µl was used for SNP genotyping using the 1536 plex assay kit and GoldenGate Assay Mouse Medium Density Linkage Array following the manufacturers protocol ( Illumina , San Diego , CA ) . The intensity data for each SNP for 228 samples were normalized and the genotypes assigned using Illumina GenomeStudio Genotyping Analysis Module version 1 . 9 . 4 . SNPs with a GC score <0 . 7 and non-informative ( homozygous ) SNPs were excluded from further analysis . SNP Hardy–Weinberg equilibrium ( HWE ) P-values were estimated with PLINK . SNPs were omitted if the HWE P<0 . 001 . As described previously in [43] , total RNA extractions from ( TRAMP × NOD/ShiLtJ ) F2 tumor samples were carried out using TRIzol Reagent ( Life Technologies , Inc . ) according to the standard protocol . RNA quality and quantity was ensured using the Bioanalyzer ( Agilent , Inc . , Santa Clara , CA ) and NanoDrop ( Thermo Scientific , Inc . , Waltham , MA ) , respectively . Per RNA labeling , 200 ng of total RNA was used in conjunction with the Affymetrix ( Santa Clara , CA ) recommended protocol for the GeneChip 2 . 0 ST chips . Hybridization cocktails containing the fragmented and labeled cDNAs were hybridized to Affymetrix Mouse Genome 2 . 0 ST GeneChip . Chips were washed and stained by the Affymetrix Fluidics Station using the standard format and protocols as described by Affymetrix . Probe arrays were stained with streptavidin phycoerythrin solution ( Molecular Probes , Carlsbad , CA ) and enhanced by using an antibody solution containing 0 . 5 mg/mL of biotinylated anti-streptavidin ( Vector Laboratories , Burlingame , CA ) . An Affymetrix Gene Chip Scanner 3000 was used to scan the probe arrays . Gene expression intensities were calculated using Affymetrix AGCC software . Partek Genomic Suite was used to RMA normalize ( Robust Multichip Analysis ) , summarize , log2 transform the data , run ANOVA analysis and unsupervised hierarchical clustering . To account for genes expressed below the threshold of detection , average levels of gene expression across all samples were calculated and genes expressed in the lower 10th percentile excluded . This encompassed the average experiment-wide background intensity of 3 . 04±0 . 12 . Microarray data are available through Gene Expression Omnibus ( accession no . GSE58829 ) . QTL analysis was performed using J/qtl [20] . Mapping of QTLs was performed for all traits using a single-QTL analysis , using a binary model for binary traits ( e . g . , distant metastasis free survival [DMFS] ) and a non-parametric model for all other traits . Significance levels were computed using permutation testing [44] , using 10 , 000 permutations . Age of death was used as an additive covariate for tumor-related traits ( primary tumor burden , seminal vesicle tumor burden ) . Age and primary tumor burden were used as additive covariates for all metastasis-related traits . Confidence intervals for QTLs identified were estimated using 2-LOD support intervals , which is on the chromosome where the LOD score did not fall below 2 . 0 of its maximum [45] . Only those QTLs reaching a genome-wide α<0 . 05 were considered to be of interest . eQTL analysis was performed using Matrix-eQTL in R [24] . A linear model was used to test for association between gene expression and SNPs , with age and primary tumor burden used as covariates . A SNP that mapped ≤1 Mb upstream or downstream of the transcription start site was used to define proximal eQTLs . Correction for multiple testing was performed using the Benjamini-Hochberg FDR method . An FDR <0 . 05 was used as the threshold for significant eQTLs . Pearson correlation coefficients and associated P-values were calculated for all traits other than those with a binary distribution by correlating the log2 transformed expression intensities of all probes mapped to a given QTL with the relevant QTL trait using MedCalc ( Ostend , Belgium ) . For the latter , student's t-tests were performed to test the significance of transcript-trait correlations . Correction for multiple testing was performed using the Benjamini-Hochberg FDR method using the QVALUE module in R [46] . An FDR <0 . 05 was used as the threshold for significant correlations . QTL candidate gene expression levels were analyzed in the cBioPortal for Cancer Genomics database ( http://www . cbioportal . org; [27] ) . Two human prostate cancer datasets possessed sufficient gene expression and clinical data to facilitate assessment of candidate genes: a ) TCGA ( Provisional ) – the Cancer Genome Atlas provisional data ( https://tcga-data . nci . nih . gov/tcga/tcgaCancerDetails . jsp ? diseaseType=PRAD&diseaseName=Prostate%20adenocarcinoma ) ; and b ) GSE21032 - Prostate Oncogenome Project , Taylor et al . [28] . The gene expression levels in TCGA ( Provisional ) dataset available on the cBioPortal website are provided by The Cancer Genome Atlas . Here , level 3 expression data were generated from RNA-seq data by first generating ‘Reads per Kilobase per Million mapped reads’ ( RPKM; [47] ) counts . This is followed by utilization of MapSplice [48] to align sequence reads and ‘RNA-Seq by Expectation Maximization’ ( RSEM ) values [49] to perform gene quantitation . cBioPortal reports higher or lower levels of gene expression by a z-score of ≥2 or ≤−2 , respectively , where the z-score is the standard deviation of static levels of transcript expression in a given case compared to the mean transcript expression in diploid tumors . Diploid tumors were used for the purposes of normalization since candidate gene ploidy could presumably impact average expression levels of candidate genes . In the GSE21032 cohort , gene up- or down-regulation in a given case is again provided by cBioPortal as a z-score of ≥2 or ≤−2 , respectively . However , here a z-score of 2 was defined as an array probe-set intensity that is two standard deviations greater than the mean of the probe set intensity in the matched normal tissue , with the opposite being true for down-regulated genes . Therefore , to make candidate gene expression levels more comparable to those reported for TCGA ( Provisional ) cohort , raw gene expression data for GSE21032 were downloaded from cBioPortal ( http://cbio . mskcc . org/cancergenomics/prostate/data/MSKCC_PCa_mRNA_data . zip ) . The expression levels of the 29 QTL candidate genes were subsequently extracted of all primary tumors with mRNA data ( n = 131 ) , average expression levels and standard deviations calculated , and z-scores for candidate gene expression in individual tumors calculated using the following formula: ( [gene expression in individual tumor – average population gene expression]/population expression standard deviation ) . Logistic regression and Kaplan-Meyer survival analyses were performed using MedCalc ( Ostend , Belgium ) . Logistic regression was performed using the stepwise method , with individual dichotomized clinical variables ( Table 3; Figure S4 ) as dependent variables and z-scores for all 29 candidate genes as independent variables . Kaplan-Meyer survival curves were constructed by comparing the time to recurrence in cases from either cohort with higher levels of tumor candidate gene expression versus all other cases . The clinical characteristics of the CGEMS GWAS cohort have been described extensively elsewhere ( dbGaP Study Accession: phs000207 . v1 . p1; [31] ) . All SNPs analyzed were either located within a given QTL candidate gene or no more than 100 , 000 bp upstream or downstream . SNP HWE P-values were estimated with PLINK . SNPs were omitted if the HWE P<0 . 001 . Association analysis between aggressive prostate cancer phenotype and SNP or haplotype was performed using a generalized linear model ( glm ) . Age and PC1 , PC2 and PC3 were included as covariates in the glm . Analysis of aggressive vs . non-aggressive disease phenotypes were performed as per the comparisons described in Table 3 . Correction for multiple testing was performed using permutation testing ( n = 10 , 000 permutations ) using the glm on NIH biowulf super cluster computer system ( http://biowulf . nih . gov ) . Specifically , permutation testing was performed for each phenotype against one SNP under rearrangements of the labels on all individuals with 10 , 000 times . Permutation tests were performed only in instances where the uncorrected P<0 . 01 . Manhattan plots were constructed in R . For haplotype analysis , genome-wide LD blocks were estimated by using the Solid Spine algorithm of Haploview software with the default parameters , and fastPHASE was performed to generate haplotypes for each individual based on the LD blocks on NIH biowulf super cluster computer system ( http://biowulf . nih . gov ) . FDR P-values were calculated by the MULTITEST package of R . All analyses were performed by using R .
|
Prostate cancer is a remarkably common disease , and in 2014 it is estimated that it will account for 27% of new cancer cases in men in the US . However , less than 13% those diagnosed will succumb to prostate cancer , with most men dying from unrelated causes . The tests used to identify men at risk of fatal prostate cancer are inaccurate , which leads to overtreatment , unnecessary patient suffering , and represents a significant public health burden . Many studies have shown that hereditary genetic variation significantly alters susceptibility to fatal prostate cancer , although the identities of genes responsible for this are mostly unknown . Here , we used a mouse model of prostate cancer to identify such genes . We introduced hereditary genetic variation into this mouse model through breeding , and used a genetic mapping technique to identify 35 genes associated with aggressive disease . The levels of three of these genes were consistently abnormal in human prostate cancers with a more aggressive disease course . Additionally , hereditary differences in these same three genes were associated with markers of fatal prostate cancer in men . This approach has given us unique insights into how hereditary variation influences fatal forms of prostate cancer .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"animal",
"genetics",
"genetic",
"predisposition",
"cancer",
"genetics",
"quantitative",
"trait",
"loci",
"genetic",
"association",
"studies",
"heredity",
"gene",
"expression",
"genetics",
"biology",
"and",
"life",
"sciences",
"quantitative",
"traits",
"human",
"genetics",
"mammalian",
"genetics",
"genetics",
"of",
"disease",
"genetic",
"loci",
"complex",
"traits"
] |
2014
|
A Systems Genetics Approach Identifies CXCL14, ITGAX, and LPCAT2 as Novel Aggressive Prostate Cancer Susceptibility Genes
|
Analyses of large-scale population structure of pathogens enable the identification of migration patterns , diversity reservoirs or longevity of populations , the understanding of current evolutionary trajectories and the anticipation of future ones . This is particularly important for long-distance migrating fungal pathogens such as Puccinia striiformis f . sp . tritici ( PST ) , capable of rapid spread to new regions and crop varieties . Although a range of recent PST invasions at continental scales are well documented , the worldwide population structure and the center of origin of the pathogen were still unknown . In this study , we used multilocus microsatellite genotyping to infer worldwide population structure of PST and the origin of new invasions based on 409 isolates representative of distribution of the fungus on six continents . Bayesian and multivariate clustering methods partitioned the set of multilocus genotypes into six distinct genetic groups associated with their geographical origin . Analyses of linkage disequilibrium and genotypic diversity indicated a strong regional heterogeneity in levels of recombination , with clear signatures of recombination in the Himalayan ( Nepal and Pakistan ) and near-Himalayan regions ( China ) and a predominant clonal population structure in other regions . The higher genotypic diversity , recombinant population structure and high sexual reproduction ability in the Himalayan and neighboring regions suggests this area as the putative center of origin of PST . We used clustering methods and approximate Bayesian computation ( ABC ) to compare different competing scenarios describing ancestral relationship among ancestral populations and more recently founded populations . Our analyses confirmed the Middle East-East Africa as the most likely source of newly spreading , high-temperature-adapted strains; Europe as the source of South American , North American and Australian populations; and Mediterranean-Central Asian populations as the origin of South African populations . Although most geographic populations are not markedly affected by recent dispersal events , this study emphasizes the influence of human activities on recent long-distance spread of the pathogen .
Domestication of ecosystems , climate change and the expanding global trade have accelerated the pace of pathogen emergence and spread [1] , [2] . Widely distributed and genetically homogenous crop genotypes are conducive for rapid pathogen emergence and subsequent propagation across large areas . Even when novel pathogens are initially endemic to restricted geographical areas , they can rapidly emerge in new regions , provided they encounter a farmland with susceptible hosts and favorable environmental conditions . In widely grown food crops , many pathogens were introduced long ago , now being geographically widespread , and therefore they do not come to mind when considering invasive pathogens [3] . Although introductions could have occurred centuries ago , the evolution of such ubiquitous pathogens remains a significant cause of concern due to the risk of re-emergence caused by accidental dissemination of new , multi-virulent races [4] , [5] or new , highly aggressive strains [6] . An accurate understanding of the origin , distribution of diversity reservoirs and past and recent migration routes of these pathogens is crucial to understand current epidemics , develop risk-assessment models and alleviate the potential adverse effects of disease emergence [7] , [8] . This is particularly true for pathogens capable of long-distance migration , for which any newly advantageous mutant ( increased virulence , aggressiveness or resistance to fungicides ) has the potential to spread over large geographical area [9] . Yellow ( stripe ) rust on wheat , which is caused by Puccinia striiformis f . sp . tritici ( PST ) , is present in most wheat-growing regions of the world [5] , [10] , [11] , [12] , [13] , [14] . The pathogen has major negative impacts on wheat production due to re-emergences and invasions [5] , [6] , [15] . As for most ubiquitous pathogens of major crops , the origin , introduction pathways and current population structure of wheat yellow rust remains largely unknown . Although an origin in Transcaucasia has been hypothesized based on disease prevalence and geographical barriers [16] , [17] , it has never been assessed in light of new knowledge on the population structure of PST . Long-distance dispersal by wind is thought to play a key role in the dissemination of the disease . The fungus is capable of long-distance migration , with well-documented cases of recurrent re-establishment of pathogen populations in areas where there are no host plants during summer/winter to allow the pathogen survival , as for the main wheat-growing provinces of north-eastern China [9] . Such spread can be due to successive jumps from field to field by this polycyclic disease throughout the season ( as in the USA [18] ) , as well as direct long-distance migration caused by winds , as documented between England and Denmark [19] . Accidental spore transport via human travel may also contribute to the intercontinental dispersal of the pathogen , as exemplified by the introduction of PST into Australia in 1979 from Europe through contaminated clothing or goods [15] . Despite the capacity for long-distance migration , the worldwide spread of PST is relatively recent , with most emergences reported only within the last decades . The pathogen reached South Africa in 1996 from an unknown source , but the first pathotypes detected were similar to those present in the Middle East and Mediterranean regions [6] , [20] . PST was first reported in South America in the early 20th century , with an unknown origin [16] , [21] . More recently , an expansion of the geographic range of PST into the warm climate of south-eastern USA [22] was shown to be due to the emergence of an aggressive strain adapted to higher temperatures than usually reported to be optimal for PST [23] , [24] . The same genotype was found in Australia two years later , while another closely related one was observed in Europe , Central and West Asia and East Africa [6] . In addition to recently colonized areas , the disease is known to periodically re-emerge through the acquisition of new virulences . These events are well documented for pathotypes carrying virulence against resistance gene Yr9 , with a first report in 1986 in Eastern Africa ( Ethiopia ) and subsequent invasions of the Middle East , Pakistan and India , reaching Bangladesh in only 12 years [5] . The geographical origins of most of the emerging strains are unknown . The population structure of PST is therefore likely to display the hallmarks of a complex mixture of re-emergences over continuous wheat-growing areas and rare founder events due to long-distance migration . Recent spreads of the disease are likely to induce marked changes in patterns of population differentiation among regions , potentially erasing the signature of more ancient colonization events . Several recent studies , using different genetic markers , investigated the population structure of PST over relatively large geographic scales . Analyses revealed an overall clonal population structure with low genetic diversity worldwide [15] , [25] , [26] , [27] , except China and Pakistan [28] , [29] . Continental dispersal of the pathogen was well evidenced [30] , leading to the conclusion that a newly emerged strain/pathotype can sweep all the worldwide populations [9] , [26] . However , despite the importance of this pathogen and increased research efforts , aspects of the worldwide population structure of the pathogen , both in terms of population subdivision and diversity , remain undetermined . In the present study , we assembled a representative set of isolates from a larger collection of PST populations from the worldwide geographical range of the fungus and analyzed their genetic variability using a single set of highly variable genetic markers . Our objectives were: ( i ) to identify the main genetic groups in modern PST populations; ( ii ) to test the Himalayan region as putative centre of diversity; and ( iii ) to identify the geographical origin of recently emerged populations and investigate the ancestral relationships among geographically spaced populations .
A set of 409 isolates was selected to represent 11 geographical regions on six continents ( Africa , Asia , Australia , Europe , North America and South America ) from a collection of more than 4 , 000 isolates available at Institut National de la Recherche Agronomique ( INRA ) , France and the Global Rust Reference Centre , Aarhus University , Denmark . The selection was made to maximize the representation of each population ( partially assessed previously by AFLP , microsatellites and/or virulence profiles [6] , [10] , [25] , [28] , [29] , [31] , [32] ) , while balancing the number of isolates for each previously described genetic group , and trying to cover the geographic distribution of each . Isolates representative of aggressive strains were selected from the two recently emerged aggressive strains , PstS1 ( associated with the post-2000 epidemics in the USA and Australia ) , and the Euro-Asian strain , PstS2 , as well as a set of aggressive isolates frequently reported in Southern Europe , PstS3 , which were less aggressive than PstS1 and PstS2 [24] . Isolates from the recently invading populations ( e . g . South Africa ) were included to infer on their source using non-parametric and Structure analyses , and were therefore not included in population-based analyses . Details regarding the number of isolates are shown in Table 1 . For most isolates , DNA was already available , having been previously extracted through modified CTAB protocols [19] , [33] . For isolates from Pakistan ( 2008 ) , Nepal ( 2008 ) and China ( 2005 ) , DNA was extracted from 5 mg of spores following Ali et al . [34] . All of the isolates were multiplied from single pustule lesions to avoid a mixture of genotypes . Molecular genotyping was carried out using a set of 20 microsatellite loci in three multiplex reactions , with subsequent separation of the PCR products using a Beckman Coulter CEQ-8000 DNA Analyzer . Electrophorograms were processed using the CEQ-8000 Genetic Analysis System Software ( Beckman Coulter [34] ) . We investigated the existence of different genetic pools of PST using both multivariate and model-based Bayesian clustering approaches . These methods avoid the clustering of individuals on a priori knowledge such as geographical locations that may artificially group different genetic lineages introduced in the same area , hindering the detection of admixture events among them [35] . Multivariate analyses were carried out using discriminant analyses of principal components , implemented in the Adegenet package in R environment [36] . The number of clusters was identified based on the Bayesian Information Criterion ( BIC ) , as suggested by Jombart et al . [36] . The model-based Bayesian method implemented in Structure 2 . 2 [37] was mostly used to confirm results of multivariate analyses , bearing in mind that this method makes the assumption of linkage equilibrium and that violations of this hypothesis for instance due to asexual reproduction can lead to spurious assignments and overestimate the number of clusters [38] . The rationale of this method is to assign multilocus genotypes to different clusters while minimizing the Hardy-Weinberg disequilibrium and the gametic phase disequilibrium between loci within clusters ( where the number of clusters may be unknown ) . The Monte Carlo Markov Chain ( MCMC ) sampling scheme was run for 200 , 000 iterations with a 100 , 000 burn-in period , with K ranging from 1 to 10 and 20 independent replications for each K . The Structure outputs were processed with Clumpp [39]; a G′-statistic greater than 80% was used to assign groups of runs to a common clustering pattern . The relatedness among geographically spaced populations was plotted using a neighbor-joining population tree based on the genetic distance DA [40] , as implemented in the Population program [41] . Significance was assessed using 1 , 000 bootstraps . The level of population differentiation was assessed using pairwise FST statistics among pairs of populations ( GENETIX 4 . 05 . 2 [42] ) . The quality of the set of markers , with respect to the inference of population structure , was tested by assessing the ability of the set of microsatellite loci to detect multilocus genotypes ( MLGs ) under panmixia , using Genclone [43] . The redundancy of the set of loci was tested by estimating the linkage disequilibrium among different loci and generating 1 , 000 random permutations with Genetix 4 . 05 . 2 [42] . Within-population variability was assessed using allele richness and gene diversity , calculated with Fstat 2 . 9 . 3 [44] . Genotypic diversity was estimated with MultiLocus 1 . 3 [45] . Private allelic richness was estimated using a rarefaction approach , implemented in Adze [46] . Observed ( Ho ) and unbiased expected heterozygosity ( He ) were computed using Genetix 4 . 05 . 2 [42] . The null hypothesis of the Hardy-Weinberg equilibrium within each population was tested using the exact test implemented in Genepop 4 . 0 [47] . Calculations were performed both on the whole dataset and on the clone-corrected data ( i . e . , a dataset in which only one representative of each repeated MLG is kept ) . Only the clone-corrected data are reported in cases where the two datasets yielded different results because the sampling during epidemics would result in over-representation of certain clones due to the recent clonal burst at local and seasonal scale , which may bias the population genetic analyses [48] . We used approximate Bayesian computation ( ABC ) to compare different competing scenarios describing ancestral relationship among populations . The approach bypasses the calculation of exact likelihoods rendering it efficient for complex population genetic models such as those underlying biological invasions [49] , [50] . The rationale is to simulate datasets assuming different parameter values under different scenarios to estimate posterior probabilities of competing scenarios and the posterior distributions of the demographic parameters under a given scenario using comparisons between simulated and observed data sets based on summary statistics ( e . g . genetic distance between populations ) . As different scenarios compared were defined based on the results of population structure analyses , they are therefore described in the results section . These scenarios globally contrasted various hypotheses concerning the source and sink relationships among the different genetic groups identified using clustering methods . Simulations were performed using Diyabc [51] , model selection and parameter estimation was carried out using the abc package in R [52] . A total of 5×105 simulated data sets was generated for each scenario under the generalized stepwise mutation model , with two parameters: the mean mutation rate ( I ) and the mean parameter ( P ) of the geometric distribution used to model the length of mutation events ( in number of repeats ) . The mean mutation rate was drawn from a uniform distribution of 10−4 to 10−3 , while the mutation rate of each locus was drawn from a gamma distribution ( mean = μ , shape = 2 ) . The parameter P was kept in the range of 0 . 1 to 0 . 3 . A range of 40 contiguous allelic states was kept for each locus , characterized by individual value of mutation rate ( lL ) and the parameter of the geometric distribution ( PL ) , which were obtained from Gamma distribution ( with mean = 1 , range 5×10−5 to 5×10−2 for IL; and mean = P , shape = 2 , shape 0 . 01–0 . 09 for PL ) . Mean number of alleles per locus , mean genetic diversity [53] , mean variance in allele size , genetic differentiation between pairwise groups , FST [54] , and genetic distance ( δμ2 [55] were used as summary statistics . Relative posterior probabilities of different scenarios were estimated by fitting a multinomial logistic regression between the summary statistics and a polychotomous variable corresponding to the different model indexes [56] , using 1% of simulated dataset closest to the observed data . The posterior distributions of parameters were estimated for the most likely scenario using a local linear regression [57] , [58] on 1% simulated datasets closest to the observed data . Confidence in model choice was assessed using a leave-one-out method [52] for each model we drew 100 of the 5×105 simulated datasets used for model selection and treated them as observed datasets ( i . e . , pseudo-observed datasets ) . Posterior probabilities of competing models were evaluated for each pseudo-observed dataset , using all remaining simulated datasets and the same methodology as described for the observed dataset . Confidence in model choice was then estimated using the number of pseudo-observed dataset that gave higher posterior probability to the model they had been simulated with . In tests of goodness-of-fit ( i . e . , model checking ) , we simulated 100 datasets of similar number of markers as observed datasets and calculated for each dataset the average across loci of several test quantities . The set of test quantities included the summary statistics used in analyses of the observed dataset . However , because using the same statistics in parameter inference and model checking can overestimate the quality of the fit [51] , we selected additional summary statistics that had not been used in parameter inferences: mean allele size variance across loci , mean index of classification , and for each population pair the mean gene diversity , mean allelic richness , and mean allele size variance across loci . Test statistics computed from observed data were then ranked against the distributions obtained from simulated datasets [58] .
We performed multilocus genotyping of 409 PST isolates , representatives of a worldwide collection , using a set of 20 microsatellite markers . Plotting the multilocus genotypes detected against the number of loci re-sampled showed that the full set of SSRs was sufficient to discriminate clonal lineages ( supporting information; Fig . S1 ) . No significant linkage disequilibrium was found among SSR loci ( data not shown ) , suggesting a lack of redundancy among markers . Some of the loci were monomorphic in certain geographical areas , except that China had no such locus and Pakistan had only one ( RJN-12; supporting information; Table S1 ) . Genotypes clearly clustered according to their geographical origin in the non-parametric discriminant analysis of principal components ( DAPC ) analysis . The Bayesian information criteria ( BIC ) curve in the DAPC analyses supported K = 6 with a clear discrimination of genotypes from China , Pakistan , Nepal , Middle East-East Africa , North western ( NW ) Europe and Central Asia-Mediterranean region ( Fig . 1 and Fig . 2 ) . Analyses using the model-based clustering method implemented in Structure also identified an optimal number of clusters ( K ) equal to 6 , based on the rate of change in the log probability of data across successive K values [59] , and patterns of subdivision were largely consistent with the results of non-parametric DAPC ( Fig . 2 and Fig . S2 ) . At K = 2 , Middle Eastern , Mediterranean and Central Asian populations were assigned to one group; the Chinese population was assigned to the other group; and Nepalese , Pakistani and NW European populations had a mixed assignment of the two groups ( Fig . S2 ) . Increasing K to 3 individualized a Pakistan-specific group , while increasing K to 4 split the cluster of Middle East , Central Asia and Mediterranean region into two groups , one specific to the Middle East and East Africa and the other specific to the Central-Asia and Mediterranean region , with substantial admixture from the Middle East . The Middle Eastern and East African populations had no differentiation from each other and are termed as Middle East-East Africa , onward . At K = 5 , the NW European populations were separated from the Chinese population , and at K = 6 , the Nepalese group individualized ( Fig . S2 ) . Increasing K above 6 did not reveal any further subdivisions ( Fig . S2 ) . Population differentiation among the different geographically spaced populations was estimated by means of pairwise Fst . Populations showed a strong differentiation , with high and significant Fst values for all pairs except for PST from the Middle Eastern , Central Asian and Mediterranean regions ( Table 2 ) , indicative of a relatively recent shared ancestry or significant gene flow among these three latter populations . Chinese , Pakistani and Nepalese populations were differentiated from one another and from the Middle Eastern and Mediterranean populations . These two latter populations were not highly differentiated from one another ( Fig . S2; Table 2 ) . The NW European population showed a strong differentiation from Mediterranean and Middle Eastern populations but was closer to the Chinese population ( Fig . S2 and Fig . S3 ) . Populations from NW Europe , North America , South America , Australia , South Africa , East Africa , the Middle East and the Mediterranean region displayed low genotypic diversity as well as an excess of heterozygosity compared to expectations under HWE , consistent with long-term clonality . Asia appeared as the zone of the highest diversity of the pathogen , with Himalayan ( Nepalese and Pakistani ) and near Himalayan ( Chinese ) populations not departing from HWE , suggesting the occurrence of recombination within the populations ( Fig . 3 ) , and showing higher variability in terms of genotypic diversity and allele richness ( Fig . 4 ) . Populations from Middle East and East Africa were more diverse than the European and Mediterranean populations , where the maximum clonal resampling was observed . Pakistan displayed the highest number of private alleles ( Fig . 4 ) . Isolates representing NW Europe also had high private allele richness , probably due to their strict clonality [19] , [25] and isolation from other populations . We detected only a few recent migrants , admixed and unassigned isolates in each geographical region in clustering analyses ( Fig . 2 and Fig . S2 ) . The existence of such genotypes in the Himalayan and neighboring regions may reflect back migrations or shared ancestral variation , as both phenomena give the same signal with clustering algorithms [60] , [61] . Clear migration footprints were only found when focusing on recently colonized areas . Analyses confirmed NW Europe as the source of the North American and Australian populations , and the Mediterranean region and Central Asia appeared to be the source of the South African population ( Fig . 2 and Fig . S2 ) . Additionally , the South American isolates were assigned to NW European isolates and displayed very low diversity , revealing another incursion from NW Europe . Isolates of the recently emerged aggressive strain PstS1 , associated with the post-2000 epidemics in the USA and Australia , consisted of only a single multilocus genotype ( MLG-99 ) resampled in other geographical regions as well . PstS1 was closely related to the other recently emerged aggressive strain reported in Asia , Africa and Europe , the PstS2 , which consisted of different multilocus genotypes , including this MLG-99 . Both PstS1 and PstS2 were assigned to the Middle Eastern-East Africa group , suggesting a source in Middle East-East Africa for these strains ( Fig . 2 and Table 2 ) . An older set of aggressive isolates frequently reported in Southern Europe [25] , although with lesser number of virulences than PstS1 and PstS2 [24] , were assigned to the Central Asian-Mediterranean genetic group ( represented as PstS3 in Fig . 2 ) . We used approximate Bayesian computation ( ABC ) analyses to infer on the ancestral relationship among populations . To limit the number and complexity of the scenarios to be compared , we used a sequential approach: we defined nested subsets of competing scenarios based on our understanding of the population structure in different regions , and analyzed these subsets sequentially . The origin of recently emerged populations was investigated using clustering and differentiation analyses and it has been presented in the previous section . We started by analyzing the historical relationships among the three populations displaying recombining population structures and located in the region of highest diversity ( i . e . , Pakistan , Nepal and China ) . The four competing scenarios assumed three different population-trees , and admixture in the ancestry of the Nepalese population ( Fig . S4 ) as this population appeared admixed in population subdivision analyses ( Fig . S2 and Fig . S3 ) . Leave-one-out cross-validation for model selection [52] confirmed that our methodology was able to distinguish between the four different scenarios ( Table S5 ) . The emergence of the Nepalese populations following admixture between Pakistani and Chinese populations appeared the most likely [Scenario 4a , posterior probability ( PP ) : 0 . 9911; ( Table S6 ) ] . This scenario ( 4a ) was then used as a backbone to design scenarios investigating the origin of other populations . Because a relatively recent origin of the Nepalese population was indicated by parameter estimates [Maximum Posterior Probability ( MPP ) estimate: 103 generations; 95% CI: 14–332; ( Table S3 ) ] , this population was therefore dismissed as a possible source of other populations and not considered in subsequent analyses . The origin of Central Asian , Middle Eastern and Mediterranean population was investigated jointly , based on the relatedness of these populations in population structure analyses ( Fig . 1 , S2 and Table 2 ) , and their geographical proximity from the centre of diversity ( the Himalayan and neighboring areas ) . Central Asian and the Mediterranean populations were pooled , based on their genetic relatedness to limit the number of competing scenarios . Twelve scenarios were compared , assuming two different native populations ( China and Pakistan ) , sequential or independent introductions and admixture ( Fig . S5 ) . Leave-one-out analyses indicated a good ability to distinguish between the twelve scenarios ( Table S5 ) . Scenario 11b , assuming that the Central Asian , Middle Eastern and Mediterranean populations split following an admixture event between Chinese and Pakistani populations appeared the most likely ( PP = 0 . 717; Table S6; Fig . 5 ) . The Pakistani population had a slightly higher contribution to the admixture event [1-r5 = 0 . 57 , 95% Credibility interval ( CI ) : 0 . 096–0 . 954] than the Chinese population , which might account for the assignment of simulated dataset under this scenario to scenario 10b ( both populations originated from Pakistan ) in leave-one-out analyses . We also considered the possibility that populations not represented in our dataset could have contributed to the genetic makeup of extant worldwide PST populations . All scenarios including a hypothetical un-sampled source population had lower posterior probabilities than the most supported scenario ( 11b ) described above ( scenarios set d; Fig . S7; Table S6 ) . Leave-one-out analyses , however , indicated limited power to distinguish among the scenarios ( Table S5 ) . The most likely scenario identified in analyses of Central Asian , Middle Eastern and Mediterranean populations was then used as a backbone to investigate the origin of the NW European population . Six scenarios were compared , assuming three different populations of origin ( China , Pakistan and the Central-Asian-Middle-Eastern-Mediterranean lineage identified above ) and admixture ( Fig . S6 ) . The scenario ( 2c ) assuming that the NW European emerged following an admixture event between Chinese and Pakistani populations appeared the most likely ( PP = 0 . 499; Table S6; Fig . 5 ) . Parameter estimates for scenario 2c suggested a higher contribution of the Chinese population to the admixture event ( 1-r7 = 0 . 756 , 95% CI: 0 . 187–0 . 987 ) , which may account for the relatively high posterior probabilities for scenarios 1c and 4c [PP = 0 . 173 and PP = 0 . 270 , respectively] . Inclusion of a hypothetical un-sampled source population revealed a slightly higher posterior probability of scenario 2c described above ( PP = 0 . 303 ) than the posterior probability of scenario 4e ( PP = 0 . 287 ) assuming that the NW European population resulted from an admixture event between the Chinese and un-sampled population ( scenarios set e; Fig . S7; Table S6 ) . Like scenario 2c , parameter estimates for scenario 4e suggested a higher contribution of the Chinese population to the admixture event ( r8 = 0 . 797 , 95%CI: 0 . 069–0 . 982 ) .
We report the existence of a strong population subdivision within PST , with a clustering of isolates according to their geographical origin despite a capacity for long-distance migration [9] , [11] . This pattern stands in stark contrast with the previous understanding of the worldwide population structure of PST , which considers the potential replacement of local populations by new invasions [5] , [9] . On the basis of pathological survey that monitor the occurrence of strains with newly acquired virulences that defeat recently deployed Yr ( resistance ) genes , the population structure of PST was considered to be shaped by a continual replacement of pre-existing populations by the newly emerged and spread pathotypes , or aggressive strains . This process is well known as the boom and bust cycle [5] , [10] . However , such surveys were designed to track the spread of a new , virulent race and , therefore , were potentially biased due to sampling only from varieties with the defeated resistance gene ( s ) in question but not from local landraces or other varieties . These observations lead to Asia ( except China ) being considered as a single epidemiological zone , with rapid and recurrent spread of new virulences over the whole zone , as in the case of virulence matching the Yr9 resistance gene ( Fig . 6 [5] ) and the recent virulence matching the stem rust resistance gene Sr31 [4] . Indeed , such geographic migrations are also documented in our study , but recently spread genotypes appear to coexist with and are dominated by older populations specific to the main geographic areas , suggesting that migrants do not replace local populations in recombinant Asian populations despite the capacity for recurrent and long-distance dispersal . In contrast , the invasion of new genotypes in clonal populations would result in a population sweep and would replace the original population . This was observed in the USA , where the post-2000 PST population is dominated by the pathotypes characteristic of the aggressive strain , PstS1 , or its derivatives , shown above to have originated from the East African-Middle Eastern region . The outcome of new pathotype introductions may depend on the relative competitiveness of local populations in their region of origin , and the selective advantage of migrants ( e . g . their virulence towards a new resistance gene widely deployed or an increased tolerance to prevalent abiotic stresses e . g . , high temperatures ) . PST has long been considered a strictly asexual pathogen on wheat due to the lack of knowledge about the existence of alternate host to complete sexual reproduction [16] , [62] . Population genetic surveys that revealed clonal populations with very low diversity in USA [27] , Europe [25] and Australia [15] were consistent with a hypothesis of asexual reproduction . Recently , populations with higher diversity were reported in the Middle East [31] and Pakistan [29] , and a recombinant population structure was evidenced in China [28] . Herein , we identified a recombinant population structure and high diversity in Nepal and Pakistan and confirmed previous findings in China , suggesting the existence of possible sexual reproduction in PST populations from a broad area ranging from the Himalayan region to the Mongolian plateau . This possibility also recently gained indirect experimental support , with the demonstration of Berberis spp . as alternate host for PST under laboratory conditions [63] and a high ability for sexual reproduction ( in the form of telial production ) reported in the Asian populations of PST [32] . Although the role of Berberis spp . for the life cycle of PST under natural conditions remains to be further investigated , the presence of Berberis spp . in Pakistan , Nepal and China [64] , [65] , [66] is consistent with the existence of a sexual cycle of the pathogen in Asia . Clustering and differentiation analyses allowed us to identify the source of new incursions and emergence ( Fig . 2 and Fig . 6 ) . The source of the Australian and North American populations was confirmed to be NW Europe , in accordance with previous findings , suggesting the migration of PST from NW Europe to Australia in 1979 [15] and probably earlier to North America [11] , [67] . We also identified the NW European source of the South American PST population , which was reported earlier in the 20th century with no inference on its source [16] , [21] . This suggests that PST incursions into both North and South America originated from NW Europe , probably through human intervention . We also identified the Mediterranean-Central Asian population as the source of South African populations , first reported in 1996 [20] , which might have resulted from wind dispersal or human intervention ( Fig . 6 ) . Two closely related strains of PST , distinct from local populations , were recently reported in North America , Australia and Europe [6] . These strains were shown to be highly aggressive and adapted to high temperature [24] . One of the strains ( PstS1 ) was responsible for PST epidemics in south-central USA , a region previously considered too warm for yellow rust epidemics [in 2000]; [ 22 , 68] , and in Western Australia [in 2002; 69] . Another strain ( PstS2 ) , closely related to the first one , was reported in NW Europe with similar aggressiveness and strong differentiation from local PST populations . PstS2 was also shown to be present in the Mediterranean region and the Middle East-East Africa [6] . Our analyses revealed that PstS1 representative isolates had a single multilocus genotype ( MLG-99 ) , while PstS2 consisted of different , but closely related , MLGs . Assignment analyses revealed that both strains originated from the Middle East-East Africa , and such global patterns of dispersal would involve accidental spore transport linked with human activities . Thus , the incursion of PST into the Americas and the spread of aggressive strains are most probably the direct consequence of human-associated dispersal , as suggested for the initial introduction of PST in Australia [15] . These results suggests that PST may spread rapidly through winds or human activities , with former playing a greater role at regional scales [26] , the latter could be involved in inter-continental spread . Transcaucasia had previously been suggested as the centre of origin for the pathogen , mainly based on its diversity of virulence and distribution of pathotypes [16] , [17] . However , the diversity of virulence and the distribution of pathotypes are strongly influenced by the resistance in host populations , and this might lead to biased inferences of the location of the centre of origin of the pathogen . In our analyses , the representative isolates from Transcaucasia were less diverse , appeared to be clonal and did not exhibit strong divergence from the rest of the Middle-Eastern , Central Asian and Mediterranean populations . In contrast , the existence of high levels of diversity , private alleles , a recombinant population structure , the ability to produce sex-related structures [32] and the independent maintenance of PST populations in the Himalaya [66] suggest this latter region as a more plausible centre of origin for PST . The analysis of ancestral relationships among worldwide populations further confirmed the Himalayan populations to be the ancestral populations and the differentiation between Pakistan and China to be the most ancestral split ( Fig . 5; Table S2 , S3 , S4 , S5 , S6 ) . If one considers that the centre of origin of PST is in the Himalayan region , then PST would have adapted to wheat through host-range expansion or host shift and not followed a host-tracking co-evolution with early wheat domestication in the Fertile Crescent . This would add to the increasingly adopted view that infection of a novel host is a major route of disease emergence in fungal pathogens [70] . Additional sampling campaigns and full-genome re-sequencing of new isolates from wild and domesticated hosts should provide further insight into the history of PST and advance our understanding of the evolutionary response of pathogens to plant domestication and the development of agro-systems . Once adapted to wheat in its centre of origin in the Himalayas or the nearby regions , PST could have spread to the rest of the world while evolving independently in different parts of the Himalayan region , resulting in population subdivision within the native area [66] . PST could have spread northward from the Himalayas to the Mongolian plateau in China , where it maintained sexual reproduction and a high diversity in some parts , with an acquisition of virulences to the wheat varieties grown in China . Eastbound , ABC analyses indicated that relatively recent admixture event between Pakistani and Chinese populations resulted in the Nepali population . On the westward side of the Himalayan region , the three Middle-Eastern , Mediterranean and Central Asian populations could be the result of an admixture between Pakistan and China ( Fig . 5 ) with relatively higher contribution of the Pakistani population to the admixture event according to ABC inferences . The Mediterranean , Middle Eastern and Central Asian populations were less differentiated , despite their distribution across a large geographical area . The off-season maintenance in some regions , its subsequent spread to other Middle-Eastern , Mediterranean and Central Asian regions and the lack of local off-season survival through volunteers or sexual reproduction in all regions could result in a source and sink relationship at the scale of this whole region , as previously suggested [31] . The most recent incursion from this Middle-Eastern , Mediterranean and Central Asian population was into South Africa , where PST was absent until 1996 [20] . The NW European population was shown in ABC analyses to be the result of admixture between China and Pakistan or between China an un-sampled source population with a higher contribution of the Chinese population to the admixture event . This is consistent with the lower differentiation between the NW European population and the Chinese population ( Table 2 ) , compared to the Pakistani or the Middle Eastern , Central Asian and Mediterranean populations . These results suggest that PST has spread from China to NW Europe , and this spread more likely occurred through some human intervention rather than an airborne incursion from the Middle Eastern , Central Asian and/or Mediterranean regions . This NW European population succeeded in terms of off-season survival on volunteers in coastal areas with mild winter and resulted in a reduced ability for sexual reproduction [32] . The clonal evolution within the NW European population resulted in a strong negative FIS value , owing to a clonal divergence of the dikaryotic genomes [25] , in line with what is expected according to the Meselson Effect [71] . From NW Europe , PST has subsequently been introduced to North and South America and even more recently to Australia ( Fig . 6 ) . Our study suggests the Himalayan region as the likely centre of origin of PST , it confirms the previous migration hypotheses ( invasion of USA and Australia from NW Europe and the source of aggressive strains ) , and it also provides an integrative scenario for the worldwide spread of PST including new findings regarding the origins of South American and European populations ( Fig . 6 ) . More extended surveys , particularly in northern parts of Asia and Eastern part of Himalayan region along with a multigene phylogeny approach based on several gene sequences will provide a more detailed resolution in the analysis of the colonization history of the pathogen and its population structure in Asia , taking advantage of the historical/epidemiological records of PST emergences . The existence of a high genotypic diversity , a high ability for sexual reproduction as well as the independent maintenance of strongly differentiated populations in the Himalayan region suggests this region as the putative centre of origin of PST . Differences in the levels of diversity and mode of reproduction among geographically distant populations are particularly relevant in the context of risk-assessment for disease emergence: Asian populations ( China , Nepal and Pakistan ) with a high level of recombination , diversity and ability for sexual reproduction could serve as possible sources for the emergence of new , virulent and aggressive strains . The maintenance of populations specific to geographical regions in Asia suggests a survival of local populations in these regions , potentially through sexual reproduction . For integrated disease management , it would be important to quantify the relative contribution of sexual vs . asexual reproduction to the diversity in different world populations and identify sexual host ( s ) or clonal over-summering/-wintering pathways . Finally , this study emphasizes the potential role of human travel and commerce as a major driver in the emergence of PST . The intensification of business and tourism activities in regions known as major sources of pathogen diversity should be considered in the context of risks associated with the emergence and worldwide spread of plant disease .
|
Domestication of ecosystems , climate change and expanding global trade have accelerated the pace of disease emergence , caused by their introduction into new areas with susceptible hosts or the spread of new damaging pathogen genotypes . The wheat yellow rust pathogen ( PST ) is a pathogen with recent reports of invasions , significantly affecting worldwide wheat production . However , its origin and ancient migration routes remained unclear and the source areas of newly spreading strains were largely unknown . This information is important to understand the trajectories of current invasions and forecast the future spread , and more generally develop risk-assessment models of pathogen emergence . We analyzed a set of worldwide representative isolates of PST , which enabled to identify six different area-specific populations . Using population genetics tools , we identified its centre of diversity in the Himalayan and near Himalayan region . We also identified the source of the recently emerged populations; Middle East-East Africa as the source of high-temperature-adapted strains spreading worldwide; Europe as the source of American and Australian populations; and Mediterranean-Central Asian populations as the origin of South African populations . We demonstrate the influence of human activities on the recent long-distance spread of the disease , though most geographic populations are not markedly affected by recent dispersal events .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"evolutionary",
"ecology",
"ecology",
"evolutionary",
"biology",
"cereals",
"crops",
"genetics",
"population",
"genetics",
"biology",
"crop",
"diseases",
"population",
"ecology",
"microbial",
"ecology",
"agriculture"
] |
2014
|
Origin, Migration Routes and Worldwide Population Genetic Structure of the Wheat Yellow Rust Pathogen Puccinia striiformis f.sp. tritici
|
Elevated uric acid ( UA ) is a key risk factor for many disorders , including metabolic syndrome , gout and kidney stones . Despite frequent occurrence of these disorders , the genetic pathways influencing UA metabolism and the association with disease remain poorly understood . In humans , elevated UA levels resulted from the loss of the of the urate oxidase ( Uro ) gene around 15 million years ago . Therefore , we established a Drosophila melanogaster model with reduced expression of the orthologous Uro gene to study the pathogenesis arising from elevated UA . Reduced Uro expression in Drosophila resulted in elevated UA levels , accumulation of concretions in the excretory system , and shortening of lifespan when reared on diets containing high levels of yeast extract . Furthermore , high levels of dietary purines , but not protein or sugar , were sufficient to produce the same effects of shortened lifespan and concretion formation in the Drosophila model . The insulin-like signaling ( ILS ) pathway has been shown to respond to changes in nutrient status in several species . We observed that genetic suppression of ILS genes reduced both UA levels and concretion load in flies fed high levels of yeast extract . Further support for the role of the ILS pathway in modulating UA metabolism stems from a human candidate gene study identifying SNPs in the ILS genes AKT2 and FOXO3 being associated with serum UA levels or gout . Additionally , inhibition of the NADPH oxidase ( NOX ) gene rescued the reduced lifespan and concretion phenotypes in Uro knockdown flies . Thus , components of the ILS pathway and the downstream protein NOX represent potential therapeutic targets for treating UA associated pathologies , including gout and kidney stones , as well as extending human healthspan .
Purine homeostasis represents a conserved metabolic pathway that is sustained by multiple enzymes orchestrating de novo synthesis , salvage , and degradation of purine intermediates . Urate oxidase , which is conserved across species , catalyzes one of the last steps of purine degradation converting uric acid ( UA ) to allantoin . However , due to multiple point mutations in the urate oxidase gene ( Uro ) human ancestors lost the ability to synthesize a functional urate oxidase , thus , increasing serum and urinary UA levels [1–5] . While most mammals show serum UA levels of 1 mg/dl and lower , healthy humans generally are in the range of 4–6 mg/dl , close to the UA solubility limit of 6 . 8 mg/dl at physiological pH and body temperature [6 , 7] . Despite a regulatory network balancing UA production and excretion , UA levels increase due to both age and a nutrient-rich diet [8] . Several genetic risk factors are associated with increased UA levels including genes of purine homeostasis , glucose metabolism , or UA transporters . Dietary risk factors are sugar-sweetened beverages , alcohol , red meat , and seafood , all found in over-abundance in the Western diet [8 , 9] . In humans , elevated UA levels in the blood ( hyperuricemia ) or urine ( hyperuricosuria ) are key risk factors for crystalopathies such as gout and kidney stones , metabolic syndrome , as well as premature death and a higher all-cause mortality risk [10–16] . Alarmingly , the prevalence of hyperuricemia in the US population is over 20% [17 , 18] . Different treatment options are commonly used to lower UA levels . As first-line therapy , an inhibitor of xanthine dehydrogenase such as allopurinol is prescribed to prevent the xanthine dehydrogenase mediated oxidation of xanthine to UA . Further options include the use of uricosuric drugs such as probenecid to increase renal UA excretion or administration of a recombinant urate oxidase enzyme called pegloticase degrading extracellular UA to the water-soluble allantoin [19 , 20] . Common to the UA lowering therapies are the well-documented issues of adverse drug reactions , contraindications , relevant drug-drug interactions , and the need for anti-inflammatory prophylaxis [9] . Interestingly , cohort studies also link exceptional longevity and a reduced prevalence of age-related diseases with UA levels at the lower end of the human serum UA spectrum [21 , 22] . In sum , the data correlate elevated UA levels with a higher prevalence of crystalopathies as well as a shortened healthspan and lifespan . However , the molecular mechanisms by which UA mediates negative health outcomes are still a matter of debate . Drosophila melanogaster provides an excellent model system to study both crystalopathies and lifespan [23–27] . Crystalopathies in flies can be induced either by feeding lithogenic agents such as ethylene glycol [28–32] or via genetic inactivation of xanthine dehydrogenase [33] . The resulting crystals usually accumulate in the Malpighian tubules , the invertebrate homolog of the human kidney convoluted tubules , and therefore recapitulate the clinical pictures of calcium oxalate urolithiasis and xanthinuria , respectively [34] . In addition , aging studies employing Drosophila take advantage of its short lifespan and fast development . Many reports highlight the relevance of nutrient-sensing pathways such as the insulin-like signaling ( ILS ) cascade in extending the organism lifespan [35–37] . Yet , there is currently no Drosophila model addressing the impact of elevated UA levels on crystalopathies and lifespan . Therefore , we designed a fly model to study the effects of elevated UA levels by recapitulating both aspects encountered in hominoid evolution: depletion of urate oxidase activity combined with changing dietary habits . Our model showed a diet-dependent accumulation of UA and shortening of lifespan . Using this model , we identified the evolutionarily conserved ILS pathway and the downstream factors FoxO and NADPH oxidase to reduce UA and its associated pathologies which might provide new targets for UA lowering therapies .
Unlike humans , Drosophila melanogaster uses the enzyme urate oxidase to convert UA into allantoin ( Fig 1A ) . Therefore , we “humanized” Drosophila to recapitulate the lack of functional urate oxidase making UA the end-product of purine catabolism . We used the heterologous , ligand ( RU486 ) -induced gene switch system to generate spatially and temporally precise RNAi-mediated knockdown of Uro in Drosophila [38] . Thus , flies carrying the ubiquitous daughterless gene switch GAL4 driver ( DaGS ) were crossed to flies with one of the two different UAS-Uro-RNAi responder transgenes ( Uro-RNAi #1 or #2 ) . 2–3 day-old progeny ( DaGS>Uro-RNAi ) of such crosses were then fed diets supplemented with the Gal4 activating ligand RU486 ( + RU ) to induce the Uro knockdown or the corresponding vehicle control ( - RU ) . Compared to the isogenic control siblings , flies consuming the + RU diet showed for both RNAi constructs a 70% knockdown of the Uro mRNA ( Fig 1B ) . To mimic the change in dietary habits of Western civilization flies were fed diets with different concentrations of yeast extract . In comparison to their isogenic control siblings , i . e . DaGS>Uro-RNAi #1 flies reared on a diet without RU486 , the knockdown of Uro , i . e . DaGS>Uro-RNAi #1 flies fed a diet supplemented with RU486 , did not affect the median or total lifespan of flies consuming a low yeast ( LY ) diet containing 0 . 5% yeast extract ( Fig 1C and 1D , S1 Fig ) . However , when Uro knockdown flies were fed a high yeast ( HY ) diet containing 5% yeast extract their lifespan was significantly shorter compared to the control siblings receiving the same food , but without RU486 supplementation ( Fig 1C and 1D ) . We also compared the median lifespan of DaGS>Uro-RNAi #1 flies on different yeast concentrations , showing that the impact of urate oxidase depletion on survival becomes apparent at a dietary yeast concentration of 3% and above ( S1 Fig ) . At 3% yeast concentration the median lifespan of control flies ( DaGS>Uro-RNAi #1 flies reared without RU486 ) was 48 days , whereas the median lifespan of their Uro knockdown siblings ( DaGS>Uro-RNAi #1 flies reared with RU486 ) was 31 . 5 days , a reduction of 34% . Similarly , at 4% yeast concentration the median lifespans of control and knockdown flies were 45 and 27 days , respectively , a decrease of 40% . At 5% yeast concentration the difference in median lifespan was even more pronounced comparing 39 . 4 days for control flies and 20 . 5 days for the Uro knockdown flies , which represented a shortening of 48% . The latter value represented the plateau , given that a higher dietary yeast concentration of 8% did not further augment the difference in median lifespan between control ( 38 days ) and Uro knockdown ( 20 days ) flies ( S1 Fig ) . In other words , we observed an inverse correlation between the dietary yeast content and the median lifespan with the Uro knockdown flies being more susceptible to higher yeast concentrations than controls . To confirm our findings and demonstrate tissue-specificity as well as independence of the lifespan effect from the RU486 ligand , we used the ubiquitous ( Da-GAL4 ) or Malpighian tubule-specific ( c42-GAL4 , Uro-GAL4 ) driver lines not relying on RU486 for activation . Considering that Malpighian tubules are the main tissue expressing Uro the neuronal Elav-GAL4 driver was used as a negative control [39] . For the four non-gene switch driver lines crosses to the Uro-RNAi and w1118 , the genetic background flies used to outcross all driver and other transgenic lines , were performed in parallel . As expected , compared to the individual w1118 control crosses efficient Uro knockdown was achieved only when using the Da-GAL4 , c42-GAL4 and Uro-GAL4 drivers ( Fig 1E ) . Coinciding with the Uro knockdown , the shortened lifespan of flies consuming the HY diet was only observed with the driver lines actively reducing Uro expression but was absent in case of the Elav-GAL4 driver ( Fig 1F ) . Importantly , flies with reduced Uro expression were significantly longer lived on low nutrient diets with 0 . 5% and 1% yeast extract compared to diets with a higher yeast extract content ( Fig 1C and 1D , S1 Fig ) , demonstrating the importance of diet in UA mediated effects on lifespan . Next , using Uro knockdown flies ( DaGS>Uro-RNAi #1 + RU486 ) in comparison to their isogenic control siblings ( DaGS>Uro-RNAi #1 - RU486 ) metabolomic profiling verified the elevation of UA levels after feeding the HY diets for 14 days . The fivefold increase in UA levels in the Uro knockdown flies was accompanied by an almost complete loss of allantoin production ( Fig 2A ) . In accordance with the qRT-PCR ( Fig 1B ) , these data indicate on the metabolite level the efficiency of urate oxidase depletion ( Fig 2A ) . An unbiased micro-CT scan revealed radiopaque masses ( putative ‘concretions' ) specifically in the abdominal area of Uro knockdown flies , but not in the control population ( Fig 2B ) . This observation was further validated by optical microscopy of dissected animals revealing the presence of concretions in the Malpighian tubule and hindgut of Uro knockdown flies ( Fig 2C ) . Due to the restricted expression pattern of the Uro gene to the excretory system both the Malpighian tubule and the hindgut represent the anatomical sites where UA is most likely expected to accumulate [40] . Furthermore , metabolomic analyses of isolated concretions ( Fig 2D ) confirmed UA as the predominant component , accounting for 95% of the metabolites measured ( Fig 2E ) . We found enhanced UA aggregation with increasing age ( Fig 2F ) . While DaGS>Uro-RNAi flies without RU486 supplementation formed almost no concretions , in Uro knockdown flies their appearance was observed as early as 4 days of feeding the HY diet . The proportion of flies forming UA deposits increased gradually over time with almost 70% of Uro knockdown flies forming concretions after 14 days ( Fig 2F ) . In contrast , DaGS>Uro-RNAi #1 flies fed the LY diet , with or without RU486 , hardly formed any concretion ( Fig 2F ) . Very much like the lifespan phenotype , we investigated the concretion formation with the set of non-gene switch drivers used before . Again , in comparison to the w1118 control crosses performed in parallel a high rate of concretion formation was only detectable on a HY diet when the Uro-RNAi construct was expressed by the Da-GAL4 , c42-GAL4 or Uro-GAL4 driver , thus causing efficient urate oxidase depletion and UA accumulation ( Fig 2G ) . The w1118 or Uro-RNAi crosses with the Elav-GAL4 driver did not result in concretion formation underscoring the tissue-specificity of the effect . Using the high proportion ( ~ 70% ) of Uro knockdown flies developing concretions after feeding the HY diet for 14 days as a reference , we analyzed the dietary yeast concentration as factor in promoting concretion formation . Like the lifespan attenuation , concretion formation started to increase significantly in Uro knockdown flies when the dietary yeast concentration reached 3% or more ( S2A Fig ) . Concretion formation was also seen when comparing the control siblings and knockdown flies of the DaGS>Uro-RNAi #2 population ( S2B Fig ) . However , DaGS>Uro-RNAi #2 flies reared on the HY diet without RU486 supplementation showed a much higher background in terms of concretion formation ( cf . S2A and S2B Fig ) . In analogy to human pathophysiology the shortened lifespan and enhanced concretion formation of Uro knockdown flies match with the enhanced mortality in patients with gout as well as the age- and diet-dependent occurrence of crystalopathies in humans [16 , 41–43] . To determine what component of the HY diet triggered increased UA production and associated phenotypes , we systematically supplemented the ‘benign’ LY diet with various basic nutrient groups: purines , pyrimidines , proteins , or sugar . Of the components tested , we found that purines ( metabolic precursors of UA ) , but not pyrimidines , proteins , or sugar led to a dose-dependent increase in concretion formation when added to the LY food ( Fig 3A ) . Dietary purine supplementation above 10 mM caused a significantly higher concretion formation in Uro knockdown flies ( DaGS>Uro-RNAi #1 with RU486 ) compared to their isogenic controls ( DaGS>Uro-RNAi #1 without RU486 ) . Interestingly , purine supplementation of 20 and 40 mM even caused concretion formation to rise in the DaGS>Uro-RNAi #1 flies without RU486 ( Fig 3A ) , thereby showing the harmful and lithogenic potential of dietary purines . Given that the purine supplemented LY diet phenocopied the concretion formation seen with the HY diet , we examined the impact of purines on lifespan . Intriguingly , addition of 40 mM purines to the LY diet ( LY+Pu ) recapitulated the lifespan effects seen with the HY diet ( cf . Figs 3B and 1C ) . I . e . , purine addition caused a ) a shortening of the lifespan of control flies and b ) this lifespan attenuation was more pronounced in Uro knockdown flies , which are unable to convert the increased UA load to allantoin ( Fig 3B ) . Thus , altering the ability to utilize purines could be an effective way to prevent UA associated pathologies in flies and strengthen the relevance of the Uro knockdown model as translational research tool . Purine homeostasis is controlled through orchestrated steps involving de novo synthesis , salvage , and degradation of purine intermediates ( S3A Fig ) . Two drugs , allopurinol and methotrexate , are well-known to effectively perturb purine metabolism . Allopurinol is commonly used in the treatment of gout in humans and prevents purine degradation by inhibiting xanthine dehydrogenase ( XDH ) , an enzyme that converts xanthine and hypoxanthine to UA . Methotrexate interferes with de novo purine synthesis by inhibiting dihydrofolate reductase ( DHFR ) , an enzyme required for folate production ( S3A Fig ) . Adding increasing amounts of allopurinol ( Fig 4A ) or methotrexate ( Fig 4B ) to the HY diet consumed by Uro knockdown flies caused a dose-dependent reduction of the concretion formation . At concentrations of 10 mM allopurinol or 50 μM methotrexate concretion formation was almost completely suppressed . We also determined if the effects of drug supplementation on food palatability and its consumption confounded our results . We compared food consumption of flies by a dye-colored food intake assay after them being fed a certain diet for 14 days . We did not observe a statistically significant change in food intake of Uro knockdown flies in case the HY diet was supplemented with allopurinol or methotrexate , when analyzed by ANOVA and Tukey’s multiple comparison post-test ( S3B Fig ) . Next , we used a genetic strategy to verify known and identify novel targets that influence UA pathologies . Therefore , we generated a recombinant fly strain that harbors the DaGS driver and the Uro-RNAi #1 transgene , hereafter denoted by DaGS::Uro-RNAi . We confirmed that the recombinant DaGS::Uro-RNAi line when crossed to either w1118 or b35785 ( a strain that carries a no target UAS-mCherry-RNAi construct ) still forms concretions when fed the RU486 supplemented HY diet . Indeed , 60–65% of the progeny of both crosses ( DaGS::Uro-RNAi x w1118 and DaGS::Uro-RNAi x b35785 ) form concretions after being fed the RU486 supplemented HY diet for 14 days ( Fig 4C ) . Using a GAL4-RNAi line targeting expression of the GAL4 transcription factor itself as a positive control reduced concretion formation of DaGS::Uro-RNAi x GAL4-RNAi flies due to the suppression of urate oxidase depletion ( Fig 4C ) . Hence , the recombinant DaGS::Uro-RNAi line represents a useful tool to study UA associated pathologies . In agreement with the allopurinol and methotrexate treatments , knocking down the corresponding target genes , XDH and DHFR , by RNAi also reduced concretion formation in the recombinant DaGS::Uro-RNAi flies ( Fig 4C ) . Those data were complemented by knockdown of another crucial gene involved in purine de novo synthesis , 5'-phosphoribosyl-1'-pyrophosphate synthetase ( PRPS ) , whose knockdown by PRPS-RNAi reduced concretion formation of the DaGS::Uro-RNAi flies as efficiently as the DHFR-RNAi ( Fig 4C , S3A Fig ) . Additionally , allopurinol and methotrexate were able to revert the short lifespan of DaGS>Uro-RNAi #1 flies when added to the RU486 supplemented HY diet ( Fig 4D and 4E ) . Unlike allopurinol , methotrexate supplementation also triggered a lifespan extension of the control DaGS>Uro-RNAi #1 flies consuming the HY diet without RU486 ( Fig 4E ) . Thus , the beneficial effect of methotrexate on lifespan is likely independent of the Uro expression level , yet , accentuates the importance of the purine metabolism in aging ( Fig 4E ) . Overall , over-expression of purine homeostasis rescued the phenotypes of Uro knockdown flies , thereby supporting the validity of the fly model for UA based pathologies and suggesting an important role for purine metabolism in UA homeostasis . Common genetic variations identified to date explain only 6–7% of the variance encountered in serum UA levels [44–47] . Furthermore , commonly prescribed medications such as allopurinol can result in serious adverse drug reactions or provide insufficient efficacy in a high percentage of users [48–50] . Thus , other genes remain to be identified that may represent better targets to lower the metabolic UA load . To identify previously unrecognized regulators of UA metabolism , we started by examining the ILS pathway for two reasons . Firstly , ILS is a conserved signaling cascade activated when flies are fed a HY diet [25 , 51–54] . Secondly , a recent genome-wide association study identified polymorphisms in the human ILS gene IGF1R that were associated with serum UA concentration [47] . Thus , we examined components of the ILS pathway as putative targets controlling UA production and utilization . We used the recombinant DaGS::Uro-RNAi line to address the impact of perturbing ILS signaling . Both strategies the RNAi mediated knockdown of the insulin-like receptor ( InR ) gene or expression of a dominant-negative InR gene variant ( InR-DN ) markedly reduced Drosophila concretion formation when fed the HY diet ( Fig 5A ) . Similarly , over-expression of the phosphatase and tensin homolog ( PTEN ) , which acts as negative regulator of ILS signaling downstream of the InR , significantly reduced concretion formation compared to the DaGS::Uro-RNAi x w1118 flies ( Fig 5A ) . Downstream of PTEN the kinase AKT integrates incoming ILS cues into metabolism by phosphorylating different effector molecules . Like inhibition of the InR , blocking ILS signaling by RNAi mediated depletion of AKT in Uro knockdown flies also reduced concretion formation ( Fig 5A ) . One of the effector molecules inactivated by AKT phosphorylation is the evolutionarily conserved transcription factor FoxO [55] . Transgenic over-expression of FOXO in the background of the DaGS::Uro-RNAi flies showed a significant rescue of concretion formation with two different FOXO expression strains ( Fig 5B ) . In contrast , knockdown of FOXO by two different RNAi constructs did not affect concretion formation ( Fig 5B ) . In addition to over-expression and knockdown of FOXO , we examined the direct influence of UA levels on ( endogenous ) FoxO activity using the Uro knockdown flies . Based on the study by Alic et al . , we picked seven target genes that are up- or down-regulated by binding of FoxO and are part of either purine metabolism ( Ade2 , Ade5 , Aprt , Veil ) or the ILS pathway ( AKT , InR , Thor ) [56] . No significant change in the expression level of the seven genes was observed when comparing control and Uro knockdown flies ( S4A Fig ) . Thus , activity of FoxO was not altered in response to an increased UA level . However , as FOXO manipulation has a strong effect on UA concretion , we propose it is upstream of UA formation pathways . To further validate the physiological relevance of FOXO over-expression we compared the metabolite levels of relevant purine intermediates ( including UA ) in DaGS::Uro-RNAi x w1118 and DaGS::Uro-RNAi x FOXO flies . While early intermediates of purine synthesis such as IMP were not altered , purine degradation products were reduced in flies with the FOXO over-expression ( Fig 5C ) . In particular , UA levels were significantly reduced threefold , thus explaining the reduced deposit formation in flies over-expressing FOXO . To determine whether the ILS pathway influences UA levels in a conserved manner between flies and humans , we assessed whether single nucleotide polymorphisms ( SNPs ) in different genes of the ILS pathway are associated with either serum UA ( SUA ) levels or gout in humans . Using data from 80 , 795 adult subjects from the Kaiser Permanente Research Program on Genes , Environment , and Health ( S1 Table ) [50] , we found that SNPs in the AKT2 and FOXO3 genes were associated with either SUA levels or the incidence for gout ( Table 1 , S2 Table ) . This association was significant after adjusting for multiple testing of phenotypes and SNPs within each gene . In light of our fly studies identifying AKT inhibition ( Fig 5A ) and FOXO over-expression ( Fig 5B ) as suppressors of UA concretions the human SNP associations shown in Table 1 support the idea that modulation of the ILS signaling network could play a critical and eventually conserved role in affecting UA levels and associated pathology in both flies and humans . Next , we addressed the underlying mechanism of the link between the ILS pathway and UA levels . UA has been proposed to play several physiological roles , depending on where it is acting . For instance , UA can act as an antioxidant in the blood , where it accounts for up to 60% of the antioxidative capacity [57] . In turn , in an intracellular setting such as in adipocytes , UA is considered a prooxidant and proinflammatory molecule [12 , 58 , 59] . Similarly , in Drosophila UA is considered a damage-associated molecular pattern ( DAMP ) able to trigger a so-called sterile inflammation , which typically results in the production of antimicrobial peptides [60] . Compared to their isogenic control siblings Uro knockdown flies depicted a two- to eightfold increase in the expression of the three antimicrobial peptides attacin A ( Att A ) , diptericin ( Dipt ) , and metchnikowin ( Metch ) after 14 days on the HY diet ( S4B Fig ) . Increased expression of the antimicrobial peptides was seen for both crosses DaGS>Uro-RNAi #1 / #2 on the HY plus RU486 diet . Using the recombinant DaGS::Uro-RNAi line over-expression of FOXO diminished the high levels of antimicrobial peptides ( S4C Fig ) . However , knockdown of the known NFκB paralogs Relish , Dorsal or the Dorsal-related immunity factor ( Dif ) , whose gene products act as dimeric transcription factors activating expression of antimicrobial peptides , did not alleviate UA concretion formation ( S4D Fig ) [61 , 62] . Work by Zhao et al . indicates that over-expression of individual antimicrobial peptides like diptericin increases the tolerance of flies resisting oxidative stress [63] . Thus , the high expression of antimicrobial peptides ( S4B Fig ) could represent a defense mechanism of Uro knockdown flies against oxidative stress rather than being a driver of UA accumulation . Consistent with this idea of intracellular UA generating a prooxidative milieu , Uro knockdown flies showed significantly elevated ROS levels in the cells of the hindgut as revealed by the doubling of the ROS sensitive DHE staining intensity ( Fig 5D ) . Considering the well-documented role of FoxO proteins in cellular responses to combat oxidative stress [55 , 64 , 65] , we speculated if FOXO over-expression could reduce the build-up of ROS found in Uro knockdown flies . Comparing ROS levels of DaGS::Uro-RNAi x w1118 flies to DaGS::Uro-RNAi x FOXO flies showed that an increased FoxO abundance indeed caused a reduction of ROS in the hindgut as well as Malpighian tubules ( Fig 5E and 5F ) . The increase of ROS levels upon Uro knockdown ( Fig 5D–5F ) could stem from two sources: ( 1 ) inefficient expression of ROS combating genes , or ( 2 ) an overabundance of ROS-producing enzymes . From the eleven oxidative stress-related genes whose expression was examined in the Uro knockdown flies , only the ROS-producing NADPH Oxidase ( NOX ) showed a twofold higher mRNA content ( Fig 6A , S4E Fig ) . Elevated NOX expression was reversed completely by either FOXO over-expression or by introducing a NOX-RNAi element into DaGS::Uro-RNAi flies ( Fig 6A , NOX mRNA ) . Of note , neither the FOXO over-expression nor the presence of the NOX-RNAi element altered efficiency of the Uro knockdown ( Fig 6A , Uro mRNA ) . Next , we asked if reduction of NOX expression via RNAi could alleviate the concretion formation characteristically observed in Uro knockdown flies . As with FOXO over-expression , knocking down NOX expression in Uro knockdown flies significantly reduced UA accumulation , which was demonstrated with two different NOX-RNAi lines ( Fig 6B ) . Further comparison of purine metabolite levels between Uro knockdown flies ( DaGS::Uro-RNAi x w1118 ) and the Uro plus NOX double knockdown flies ( DaGS::Uro-RNAi x NOX-RNAi ) identified a significant drop of UA levels and depicted the interwoven nature of the NADPH oxidase with purine catabolism ( Fig 6C ) . We than speculated whether the effect of NOX depletion extended to an improvement in lifespan . Like concretion formation , the recombinant DaGS::Uro-RNAi line also recapitulated the lifespan attenuation on a HY diet when crossed to w1118 ( Fig 6D , S5A Fig ) or the mCherry-RNAi strain b35785 ( S5B Fig ) . The median lifespan of DaGS::Uro-RNAi x w1118 flies was reduced by 35% from 31 days without RU486 to 20 days in presence of RU486 ( Fig 6D ) . This lifespan attenuation on the HY diet was partially rescued in DaGS::Uro-RNAi x NOX-RNAi flies using NOX-RNAi #1 or #2 ( Fig 6D , S5F Fig ) . To corroborate the NOX related findings , we used a pharmacological approach by inhibiting NOX activity and the associated ROS production using either apocynin ( ACY; a NOX inhibitor ) or vitamin C ( Vit C; a ROS scavenger ) , respectively . Dietary ACY supplementation reduced the concretion formation of DaGS>Uro-RNAi flies on the HY plus RU486 diet in a dose-dependent manner . Concretion formation was reduced for both Uro-RNAi fly populations from about 70% without ACY supplementation down to 15% in presence of 5 mM ACY ( Fig 7A ) . In terms of lifespan , a 5 mM ACY supplementation to the HY diet was able to extend the short lifespan of both Uro knockdown fly lines whereas the corresponding control siblings ( DaGS>Uro-RNAi #1 / #2 without RU486 ) showed no change of survival in presence of ACY ( Fig 7B and 7C ) . Thus , the ACY treatment very much mirrored the results obtained with the NOX-RNAi ( cf . Figs 6B , 6D and 7A–7C ) . Last , we speculated if a water-soluble , i . e . membrane-permeable , ROS scavenger like Vit C could reduce the elevated ROS level found in Uro knockdown flies . Supplementation of the HY diet with increasing amounts of Vit C ranging from 10 to 100 mM gradually reduced concretion formation of Uro knockdown flies ( Fig 7D ) . 50 mM Vit C addition to the HY diet also rescued the lifespan effect observed in flies with reduced urate oxidase levels ( Fig 7E ) . Like the MTX treatment ( Fig 4E ) , dietary Vit C supplementation also extended the median survival of control flies ( Fig 7E ) . Worth mentioning , foods rich in Vit C are associated with a reduced risk for elevated UA and gout , due to its uricosuric effect [66 , 67] . While Vit C seemed to extend the median lifespan of both the control ( DaGS>Uro-RNAi #1 without RU486 ) and Uro knockdown flies ( DaGS>Uro-RNAi #1 with RU486 ) by 16% and 53% , respectively , the impact of apocynin on lifespan was specific for the urate oxidase depleted animals ( Fig 7B , 7C and 7E ) . In presence of RU486 , ACY extended the median lifespan of DaGS>Uro-RNAi #1 flies by 47% ( Fig 7B ) and that of DaGS>Uro-RNAi #2 flies by 30% ( Fig 7C ) . In summary , our study argues the importance of purines as a dietary component that limits lifespan . Generally , mutants in genes that influence lifespan upon dietary restriction either extend lifespan upon rich nutrient conditions while failing to extend lifespan upon dietary restriction conditions or attenuate the maximum lifespan upon dietary restriction . However , Uro mutants belong to a “novel” class of genes that limit lifespan only upon rich conditions but have little or no influence upon dietary restriction . We hypothesize that mutants that display such a phenotype encode a gene that amplifies the cellular damage that takes place under rich nutrient conditions compared to dietary restriction . We speculate that the increase of UA is deleterious , and a lower UA level is partially responsible for the lifespan extension upon dietary restriction . The fly model established here will also be useful in dissecting the pathological effects of elevated UA on crystal formation as well as the underlying genetic factors driving crystal formation in the first place . The latter point is of particular interest , considering that only 5% of hyperuricemic people with SUA above 9 mg/dl develop gout [8 , 9] . The genetic predisposition influencing UA accumulation is currently subject of genome-wide association studies and is complemented by our approach combining in vivo fly work and human candidate gene studies . Our data showing that genes such as AKT , FOXO and NOX are associated with UA levels and the incidence for deposit formation start shedding light on this important issue . In addition , the fly model will be beneficial to study the larger role of UA and UA deposits on influencing mortality as was underpinned by recent studies linking exceptional longevity and a reduced prevalence of age-related diseases with UA levels at the lower end of the human SUA spectrum [21 , 22] . Our work demonstrates that inhibition of Uro in flies can serve as a useful model to study UA induced pathologies related to lifespan and crystal formation . Our results also highlight the importance of ILS signaling , and its downstream effector NOX , as an important mediator of UA metabolism and related pathologies ( Fig 7F ) . Future research using this model can help determine the role of the several candidate genes in humans that have been implicated in altering UA levels . Given the over-nutrition encountered in developed countries , the causal nexus uncovered here identifies potential new drug targets that could help regulate UA levels , ameliorating pathologies associated with hyperuricemia and extending human healthspan .
Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact , Pankaj Kapahi ( pkapahi@buckinstitute . org ) . Bar graphs and scatter blots represent the mean ± standard error ( SE ) , when indicated . Lifespan curves , and other graphs were visualized using GraphPad Prism 5 software . For statistical comparisons between the control and single treatment group a two-tailed student’s t-test was used; except for lifespan analysis where the log-rank test was performed . When multiple conditions were compared ANOVA was used followed by Tukey’s multiple comparison post-test comparing all pairs of conditions . Significant differences are indicated according to the p-value by asterisks with p < 0 . 001 ( *** ) < 0 . 01 ( ** ) < 0 . 05 ( * ) . Non significant differences are indicated by n . s .
|
Enzymatic purine degradation in humans ends with uric acid ( UA ) . Multiple genetic and dietary factors raise UA levels above the norm , which is called hyperuricemia or hyperuricosuria when detected in the serum or urine , respectively . Clinical studies report a correlation between elevated UA and a plethora of chronic diseases including crystalopathies like UA kidney stones and gout , or metabolic and vascular disorders such as diabetes , obesity , and coronary artery disease . Here , we identified a regulatory role for the insulin-like signaling cascade affecting UA metabolism using a Drosophila melanogaster model . In the process we determined previously unrecognized potential drug targets to treat elevated UA levels and associated pathologies such as gout or UA kidney stones , with the potential additional benefit of extending human healthspan . Our work also establishes the fly as a model system to characterize the influence of genetic and dietary factors in gout or UA kidney stone development in a manner readily amenable for small-scale screening of drug interventions . The novelty of our findings , their broad impact , and relevance for multiple diseases opens up an important area of research to define mechanisms of UA accumulation .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Material",
"and",
"methods"
] |
[
"inflammatory",
"diseases",
"invertebrates",
"rheumatology",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"drugs",
"diet",
"organic",
"compounds",
"animals",
"purines",
"animal",
"models",
"fungi",
"drosophila",
"melanogaster",
"model",
"organisms",
"nutrition",
"experimental",
"organism",
"systems",
"pharmacology",
"drug",
"metabolism",
"drosophila",
"research",
"and",
"analysis",
"methods",
"purine",
"metabolism",
"animal",
"studies",
"chemistry",
"insects",
"gout",
"pharmacokinetics",
"yeast",
"arthropoda",
"biochemistry",
"methotrexate",
"eukaryota",
"organic",
"chemistry",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"metabolism",
"organisms"
] |
2019
|
A conserved role of the insulin-like signaling pathway in diet-dependent uric acid pathologies in Drosophila melanogaster
|
Rift Valley fever virus ( RVFV ) ( genus Phlebovirus , family Bunyaviridae ) is a negative-stranded RNA virus with a tripartite genome . RVFV is transmitted by mosquitoes and causes fever and severe hemorrhagic illness among humans , and fever and high rates of abortions in livestock . A nonstructural RVFV NSs protein inhibits the transcription of host mRNAs , including interferon-β mRNA , and is a major virulence factor . The present study explored a novel function of the RVFV NSs protein by testing the replication of RVFV lacking the NSs gene in the presence of actinomycin D ( ActD ) or α-amanitin , both of which served as a surrogate of the host mRNA synthesis suppression function of the NSs . In the presence of the host-transcriptional inhibitors , the replication of RVFV lacking the NSs protein , but not that carrying NSs , induced double-stranded RNA-dependent protein kinase ( PKR ) –mediated eukaryotic initiation factor ( eIF ) 2α phosphorylation , leading to the suppression of host and viral protein translation . RVFV NSs promoted post-transcriptional downregulation of PKR early in the course of the infection and suppressed the phosphorylated eIF2α accumulation . These data suggested that a combination of RVFV replication and NSs-induced host transcriptional suppression induces PKR-mediated eIF2α phosphorylation , while the NSs facilitates efficient viral translation by downregulating PKR and inhibiting PKR-mediated eIF2α phosphorylation . Thus , the two distinct functions of the NSs , i . e . , the suppression of host transcription , including that of type I interferon mRNAs , and the downregulation of PKR , work together to prevent host innate antiviral functions , allowing efficient replication and survival of RVFV in infected mammalian hosts .
Rift Valley fever virus ( RVFV ) is a mosquito-borne zoonotic pathogen , which is distributed in sub-Saharan Africa [1] and has also caused large outbreaks in Madagascar [2] , Egypt [3] , [4] , [5] , Saudi Arabia [6] , and Yemen [6] . In endemic areas , RVFV naturally circulates among mosquitoes and ruminants , such as sheep , goat and cattle [7] . RVFV infection in adult ruminants causes febrile illness and a high rate of abortions , while some newborn animals less than 1–2 weeks of age develop an acute infection which results in higher mortality rates than those in adults [8] . Humans infected with RVFV usually develop an acute febrile myalgic syndrome; however , a small percentage of patients have a lethal illness that results in hepatic damage , hemorrhagic fever-like illness , encephalitis and/or retinal vasculitis [8] . Due to the exotic origin of the virus , the potential for the aerosol transmission [9] , [10] , [11] , [12] and serious consequences for humans and livestocks , wild-type ( wt ) RVFV is classified as a Risk Group 3 pathogen , that needs to be handled in a high-containment facility , e . g . , a biosafety level ( BSL ) -4 laboratory , whereas a highly attenuated MP-12 strain of RVFV , produced after 12 serial passages of wt RVFV ZH548 in MRC-5 cells in the presence of 5-fluorouracil [13] , is a Risk Group 2 pathogen . RVFV , which belongs to the genus Phlebovirus , family Bunyaviridae , is a negative-stranded RNA virus carrying a single-stranded , tripartite RNA genome composed of S , M and L segments [14] . The S segment encodes N and NSs genes and uses an ambi-sense strategy to express the N and NSs proteins in infected cells; N mRNA encoding N protein is transcribed from the viral-sense ( negative-sense ) S segment , while NSs mRNA encoding NSs protein is transcribed from the antiviral-sense ( positive-sense ) S segment . Monocistronic M mRNA and L mRNA are transcribed from the viral-sense M and L segments , respectively . M mRNA has one large open-reading frame ( ORF ) which encodes the nonstructural NSm protein , a 78-kDa glycoprotein and two major viral structural glycoproteins , Gn and Gc [15] , [16] . L mRNA encodes L protein , a viral RNA-dependent RNA polymerase . Both N and L proteins are required for viral RNA synthesis , while Gn and Gc function as envelope proteins [14] . NSm and NSs , both nonstructural proteins , are dispensable for viral replication in cell cultures [17] , [18] , [19] , [20] , but are involved in viral pathogenesis [21] , [22] , [23] . RVFV NSs protein is not essential for virus replication in cell cultures [19] , [20] , yet plays a critical role in viral virulence [21] . A naturally occurring RVFV mutant Clone 13 , which lacks approximately 70% of the NSs gene [20] , is highly attenuated in mouse , and , when reassortant viruses between wt RVFV and Clone 13 were characterized , the NSs was revealed as a major determinant of viral virulence in the mouse model [21] . The NSs localizes in the nucleus and cytoplasm in both RVFV-infected cells and NSs-expressing cells; further , the nuclear NSs , but not the cytoplasmic NSs , forms a unique filamentous structure [24] . The NSs suppresses the transcription of host mRNAs by interacting with the p44 subunit of TFIIH , an essential transcriptional factor for cellular RNA polymerase II [25] . Furthermore , the RVFV NSs binds to Sin3A-Associated Protein 30 ( SAP30 ) , which is important for maintaining the repressor complex containing histone deacetylase 3 on the interferon ( IFN ) -β promoter , and suppresses the IFN-β promoter activation early in infection [26] . Accordingly , the NSs protein in the nucleus of infected cells most probably exerts these host transcriptional-suppressive activities , including that of IFN production inhibition , and contributes to viral virulence [21] . In contrast , the biological function of cytoplasmic NSs is largely unknown . RVFV NSs expression promotes RVFV minigenome RNA synthesis driven by N and L protein in expression studies [27] . Because RNA synthesis of bunyaviruses occurs in the cytoplasm , NSs protein in cytoplasm may promote the minigenome RNA synthesis by unknown mechanism . To establish that RVFV NSs exhibits a novel function , we hypothesized that , when combined , RVFV replication and NSs-induced host transcription suppression likely induces a cellular environment that is unsuitable for viral replication . Thus , to secure efficient RVFV replication , the NSs protein , in turn , alters this putative , virally unfriendly cellular environment to one that supports efficient viral replication . To test this possibility , we examined the replication of RVFV lacking the NSs gene , a mutant which was generated by employing a reverse genetics system [19] , in the presence of a host transcription inhibitor , e . g . , actinomycin D ( ActD ) or α-amanitin; these drugs were selected because they mimicked the host transcriptional suppressive activities of the NSs . Consistent with our supposition , RVFV lacking the NSs , but not RVFV , failed to replicate efficiently in ActD-treated cells . We noted that double-stranded RNA ( dsRNA ) -dependent protein kinase ( PKR ) -mediated eukaryotic initiation factor ( eIF ) 2α phosphorylation suppressed the translation of RVFV lacking NSs in the presence of ActD . Further studies uncovered that RVFV NSs promoted PKR downregulation as early as 4 hours post-infection , and prevented eIF2α phosphorylation , which secured efficient viral translation . We speculate that this novel function of RVFV is important for counteracting the antiviral activities of PKR and allowing efficient virus replication and survival in infected hosts .
To explore a novel biological function of RVFV NSs protein , in addition to its host transcriptional shutoff activity , we investigated the effect of ActD , a host transcriptional inhibitor , on the replication of MP-12 lacking the NSs gene in IFN-incompetent VeroE6 cells . VeroE6 cells were mock-infected or infected with MP-12 or rMP12-rLuc , which expressed Renilla luciferase ( rLuc ) in place of the NSs ( Figure 1A ) at a multiplicity of infection ( moi ) of 3 . After 1 h virus adsorption , cells were incubated in the absence or presence of 5 µg/ml of ActD . Supernatants were harvested at 16 hours post-infection ( h . p . i . ) , and virus titers were measured by plaque assay . ActD treatment had little effect on MP-12 titers , yet it significantly reduced the titer of rMP12-rLuc ( Figure 1B ) , which suggested that the RVFV NSs was important for an efficient virus replication in the presence of ActD . To understand how the NSs protein exerted an efficient viral replication in the presence of ActD , we analyzed the status of host and viral translation . VeroE6 cells were mock-infected or independently infected with MP-12 , rMP12-rLuc and rMP12-C13type , the latter of which lacks approximately 70% of the NSs ORF ( Figure 1A ) , at an moi of 3 . Cells were radiolabelled with [35S] Methionine/Cysteine from 15 to 16 h . p . i . Cell extracts were prepared at 16 h . p . i . , and the samples were applied to SDS-PAGE ( Figure 1C , top panel ) . In the absence of ActD , the host protein synthesis of rMP12-rLuc-infected cells ( Figure 1C , lane 5 ) and rMP12-C13type-infected cells ( Figure 1C , lane 7 ) was more efficient than that of MP-12-infected cells ( Figure 1C , lane 3 ) . It is possible that NSs-mediated transcriptional suppression that occurred only in MP-12-infected cells , but not in cells infected with viruses lacking NSs , caused a reduction of host mRNA abundance , leading to the less efficient translation of host mRNAs in MP-12-infected cells . Efficient N protein synthesis occurred in cells infected with all three viruses in the absence of ActD , a finding which was consistent with our previous studies [19] , while for some unknown reasons the accumulation of N proteins of rMP12-rLuc and rMP12-C13type was slightly higher than that of MP-12 . ActD treatment resulted in a strong inhibition of both host proteins and N protein synthesis in rMP12-rLuc-infected cells ( Figure 1C , lane 6 ) and rMP12-C13type-infected cells ( Figure 1C , lane 8 ) . In contrast , ActD treatment did not inhibit N protein translation in MP-12-infected cells; rather , it moderately inhibited host protein synthesis in MP-12-infected cells ( Figure 1C , lane 4 ) . A similar ActD-induced , moderate host protein synthesis inhibition also occurred in mock-infected cells ( Figure 1C , lane 2 ) . Western blot analysis of cell extracts at 16 h . p . i . clearly showed that ActD treatment strongly inhibited N protein accumulation in rMP12-rLuc-infected cells and rMP12-C13type-infected cells , but not in MP-12-infected cells ( Figure 1C , bottom panels ) . In summary , ActD treatment had little effect on MP-12 replication , whereas it strongly inhibited the expression of both host and viral proteins in the cells infected with MP-12 lacking the NSs gene , which resulted in poor virus production . To further confirm that the NSs exerted an efficient viral replication in the cells that underwent ActD-induced host transcriptional suppression , 293 cells , which showed higher RNA transfection efficiencies than did VeroE6 cells ( data not shown ) , were infected with rMP12-rLuc at an moi of 2 . After virus adsorption , infected cells were mock-transfected or independently transfected with in vitro-synthesized RNA transcripts encoding chloramphenicol acetyltransferase ( CAT ) , MP-12 NSs , or wt RVFV ZH501 NSs . Then , the cells were mock-treated or treated with ActD . Analysis of rLuc activities at 16 h . p . i . demonstrated that NSs expression had little effect on rLuc activities in the absence of ActD ( Figure 2A ) . In the presence of ActD , rLuc activities were clearly higher in the cells transfected with MP-12 NSs RNA transcripts or ZH501 NSs RNA transcripts than in those transfected with CAT RNA transcripts or in the mock-transfected samples ( Figure 2A ) . In NSs RNA-transfected cells , similar levels of rMP-12-rLuc titers were observed in both the ActD-treated and mock-treated samples , whereas the rMP12-rLuc virus titers in mock-transfected cells and CAT RNA-transfected cells were significantly lower in the presence of ActD compared to the ActD-untreated cells ( Figure 2B ) . Western blot analysis showed that NSs proteins were indeed expressed in the cells transfected with RNA transcripts encoding MP-12 NSs or wt RVFV ZH501 NSs ( Figure 2C , lanes 3 , 4 , 7 , and 8 ) . Also NSs protein expression increased the accumulation of rMP12-rLuc N protein in the presence of ActD ( Figure 2C , lanes 7 and 8 ) . These data demonstrated that NSs exerted efficient rMP12-rLuc replication in the presence of ActD . We noted that ActD treatment modestly reduced the accumulation of the CAT protein ( Figure 2C , lane 6 ) . Probably large amounts of CAT RNA transcripts that were introduced into the cells partly overcame the translational suppressive effects that were induced by the combination of rMP-12-rLuc replication and ActD treatment . To test the possibility that ActD treatment alone suppressed translation and RVFV NSs counteracted it , 293 cells were mock-treated or treated with 5 µg/ml of ActD . We examined the resulting polysome profiles at 16 h post–ActD treatment ( Figure S1A ) . Because ActD treatment at 5 µg/ml inhibits the transcription mediated by RNA polymerases I , II and III [28] , we expected a reduction in the abundance of cellular mRNAs , tRNAs , and ribosomal RNAs , leading to reduced abundances of polysomes . ActD treatment indeed resulted in a reduced abundance of polysomes , whereas it did not substantially alter the polysome pattern , a finding which suggested to us that ActD treatment did not abolish translational activities . To test the translational competence of the cells treated with transcriptional inhibitors , we transfected 293 cells with in vitro-synthesized RNA transcripts encoding rLuc gene . Cells were mock-treated or treated with ActD or α-amanitin , an RNA polymerase II inhibitor [29] . The rLuc activities at 16 h post-transfection were slightly increased in the cells treated with ActD or α-amanitin ( Figure S1B ) , demonstrating active host translation activities in the presence of either ActD or α-amanitin . These data led to the suggestion that by combining the replication of RVFV lacking the NSs and treatment of ActD , a cellular condition that is unfavorable for translation could be induced and that NSs expression somehow altered the cellular environment from one that was translationally inactive to one translationally active . To establish that NSs exerts an efficient viral translation in the presence of a transcription inhibitor , we tested whether coinfection of MP-12 and rMP12-rLuc increases the translation of rLuc mRNA of rMP12-rLuc in the presence of ActD . VeroE6 cells were mock-infected or co-infected with rMP12-rLuc and MP-12; rMP12-rLuc was infected at an moi . of 2 , while MP-12 was infected at moi . of 0 . 1 or 1 , as indicated in Figure 3 . The cell extracts were harvested at 16 h . p . i . , and the rLuc activities ( Figure 3A ) and accumulation of viral RNAs ( Figure 3B ) were examined . As expected , ActD-treatment reduced the rLuc activities in cells infected with rMP12-rLuc alone ( Figure 3A ) . Co-infection of MP-12 resulted in the reduction of both rLuc activities ( Figure 3A ) and the amounts of rLuc mRNA ( Figure 3B , lane 4 ) in the absence of ActD . In the presence of ActD , MP-12 co-infection also reduced the rLuc mRNA abundance ( Figure 3B , lane 9 ) , whereas it increased rLuc activities ( Figure 3A ) . The results suggested that NSs protein expressed from MP-12 S-segment promoted an efficient translation of the rLuc mRNA of rMP12-rLuc in the presence of ActD . We also noted that MP-12 co-infection did not reduce the abundance of the viral-sense rMP-12-rLuc S segment in the presence of ActD , and yet it caused the reduction of rLuc mRNA abundance ( Figure 3B , lane 9 ) , which we believe implies that NSs expression and ActD treatment generated a cellular environment that was more favorable for rMP-12-rLuc RNA replication than for transcription . The dsRNAs generated during viral replication activate PKR , which in turn phosphorylates eIF2α [30] . Also 5′-triphosphated single-stranded RNAs activate PKR [31] . The eIF2α is a component of eIF2 , which binds to GTP and Met-tRNA to deliver the Met-tRNA to the start codon in capped mRNA , forming a 43S pre-initiation complex [32] . Upon the binding of the 60S ribosomal subunits to the 43S preinitiation complex , eIF2-GDP is released from the ribosome and undergoes a GTP exchange reaction catalyzed by binding with eIF2B , and the resultant eIF2-GTP is recycled for the next round of translation initiation . The phosphorylated eIF2α binds to eIF2B with a high affinity and prevents eIF2B to be used for the subsequent eIF2-GDP-to-eIF2-GTP exchange reaction , leading to the suppression of translation initiation . Hence , the phosphorylation status of eIF2α plays a critical role in translational control [32] . We suspected that rMP12-rLuc replication in the presence of ActD may generate dsRNAs and/or 5′-triphosphated single-stranded RNAs , which activate PKR , resulting in the phosphorylation of eIF2α . When we analyzed the level of eIF2α phosphorylation in VeroE6 cells infected with rMP12-rLuc in the absence of transcription inhibitor , we found low levels of eIF2α phosphorylation from 8 to 24 h . p . i . and an efficient accumulation of N protein at 8 h . p . i . and onward ( Figure 4A–4C ) . In contrast , when we treated rMP12-rLuc-infected cells with ActD or α-amanitin , either compound induced an efficient accumulation of phosphorylated eIF2α from 8 to 16 h . p . i . , with a concomitant poor N protein accumulation ( Figure 4A–4C ) and virus replication suppression ( Figure 4D ) . The mechanism of reduction in the abundance of the phosphorylated eIF2α at 24 h . p . i . in the presence of ActD or α-amanitin ( Figure 4 , lanes 13 and 19 ) is unknown . When we treated rMP12-rLuc-infected cells with different concentrations of ActD or α-amanitin , we found that the inhibition of phosphorylated eIF2α accumulation was dependent on the concentrations used ( Figure S2 ) . Also a strong correlation was seen between an increase in eIF2α phosphorylation and the decrease in both N protein accumulation and infectious virus production ( Figure S2 ) . In contrast , no significant accumulation of phosphorylated eIF2α occurred in MP-12-infected VeroE6 cells in the presence of ActD ( Figure 4A–4C ) . As expected , ActD treatment had little effect on MP-12 replication ( Figure 4D ) . These data strongly suggested that the presence of highly phosphorylated eIF2α levels suppressed the translation of viral mRNAs , leading to inefficient virus production , and that the NSs protein somehow suppressed eIF2α phosphorylation , and facilitated efficient viral mRNA translation under the conditions of host transcriptional shutoff . Because activated caspases 3 , 7 and 8 can cleave PKR , releasing the biologically active C-terminus kinase domain from the N-terminus inhibitory domain , resulting in eIF2α phosphorylation [33] , we subsequently tested whether the accumulation of phosphorylated eIF2α following the combined activity of viral replication and ActD-treatment was due to induction of apoptosis [34] . VeroE6 cells , either infected with rMP12-rLuc or mock-infected , received the pan-caspase-inhibitor , benzyloxycarbonyl-Val-Ala-DL-Asp ( OMe ) fluoromethylketone ( Z-VADfmk ) , in the presence of ActD or α-amanitin ( Figure S3 ) . Judging from the resulting inhibition of cleaved caspase-3 accumulation in these cells , the Z-VADfmk treatment indeed inhibited apoptosis , whereas it had little effect on the eIF2α phosphorylation status and infectious virus yield . These data suggested to us that the accumulation of phosphorylated eIF2α in cells supporting rMP12-rLuc replication in the presence of ActD or α-amanitin was caspase-independent . Four different kinases , including PKR , PKR-like ER kinase ( PERK ) , heme-regulated inhibitor and the general control , non-depressible-2 ( GCN2 ) , are known to phosphorylate eIF2α [35] . To determine the role of PKR in the accumulation of phosphorylated eIF2α in rMP12-rLuc-infected cells treated with ActD or α-amanitin , we generated a recombinant MP-12 , rMP12-PKRΔE7 , which expresses a dominant-negative form of PKR , PKRΔE7 [36] carrying the N-terminal Flag epitope tag , in place of the NSs ( Figure 5A ) . If the replication of MP-12 lacking the NSs gene in cells subjected to host transcriptional suppression activates PKR , which in turn phosphorylates eIF2α , then the virally-encoded PKRΔE7 in rMP12-PKRΔE7-infected cells would interfere with the PKR function , resulting in the inhibition of PKR-mediated eIF2α phosphorylation and thereby leading to efficient viral translation and virus production . VeroE6 cells were mock-infected or independently infected with rMP12-rLuc , rMP12-PKRΔE7 and MP-12 at an moi of 3 . After the removal of the inocula , cells were treated with ActD or α-amanitin , and cell extracts were harvested at 16 h . p . i . As expected , rMP12-rLuc replication in the presence of ActD or α-amanitin induced eIF2α phosphorylation , resulted in reduced virus replication ( Figure 5B–5D ) . In contrast , efficient N protein accumulation and efficient virus replication , with no significant accumulation of phosphorylated eIF2α occurred in both MP-12-infected cells and rMP12-PKRΔE7-infected cells in the presence of ActD- or α-amanitin ( Figure 5B–5D ) . This finding strongly suggested that PKR is important for eIF2α phosphorylation in cells infected with the RVFV lacking the NSs in the presence of transcription inhibitors . To further confirm these data , viral protein accumulation and replication were analyzed in wt mouse embryonic fibroblast ( MEF ) cells and in Pkr0/0 MEF cells lacking a functional PKR expression [37] . MP-12 efficiently replicated in both wt MEF and Pkr0/0 MEF cells , and ActD treatment had little effect on N protein accumulation and virus replication ( Figure 5E , top and middle panels ) . rMP12-rLuc replication was not as efficient as MP-12 in wt MEF cells in the absence of ActD for an as yet unidentified reason ( Figure 5E , middle panel ) . In ActD-treated wt MEF cells , both rMP-12-C13type and rMP12-rLuc failed to efficiently accumulate N proteins , and rMP-12-rLuc replicated poorly , whereas rMP-12-rLuc underwent efficient N protein accumulation and viral replication in ActD-treated Pkr0/0 MEF cells ( Figure 5E , top and middle panels ) . Furthermore , accumulation of phosphorylated eIF2α did not occur in Pkr0/0 MEF cells that were infected with rMP12-rLuc in the presence of transcriptional inhibitors ( Figure 5E , bottom panel ) . These data were consistent with a notion that PKR triggered the accumulation of phosphorylated eIF2α in cells infected with the MP-12 lacking the NSs , under the conditions of cellular transcriptional suppression , and the NSs protein interfered with the PKR-mediated eIF2α phosphorylation ( Figure 5B–5D ) . To know how the NSs suppressed the eIF2α phosphorylation activity of the PKR function , a dsRNA-binding assay was performed to test the possibility that the NSs binds to dsRNA , sequesters dsRNA from PKR , and interferes with the dsRNA-mediated PKR activation . 293 cells were mock-infected or infected with rMP12-NSs-Flag carrying Flag-tagged NSs , rMP12-rLuc-Flag carrying Flag-tagged rLuc ( Figure 6A ) or rMP12-PKRΔE7 ( Figure 5A ) . In a separate experiment , 293 cells were transfected with in vitro-synthesized RNA transcripts encoding NSs . Lysates were prepared at 16 h . p . i . or 16 h post-transfection , and incubated with poly I∶C beads ( dsRNA ) or poly ( C ) beads ( ssRNA ) . Then the dsRNA-bound complexes were analyzed by a Western blot in which we used an anti-Flag antibody or anti-NSs antibody ( Figure 6B ) . As expected , dsRNA bound to PKRΔE7 [36] , whereas it poorly bound to the NSs from rMP12-NSs-Flag-infected cells and that from the NSs-expressing cells ( Figure 6B ) , which suggested that NSs did not suppress PKR activation by its binding to dsRNA . Because activated PKR undergoes a structural alteration and autophosphorylation [30] , we tested whether NSs prevented PKR autophosphorylation . 293 cells were infected with rMP12-NSs-Flag , rMP12-rLuc-Flag or rMP12-PKRΔE7 , and mock-treated or immediately treated with ActD . Cell lysates were harvested at 16 h . p . i . , and PKR was immunoprecipitated by anti-human PKR antibody , and the PKR bound to the protein A beads was subjected to an immunoprecipitation ( IP ) -kinase assay by using [γ-32P]ATP . We used 293 cells for this assay , because anti-human PKR antibody efficiently immunoprecipitated human PKR in 293 cells , but not non-human primate PKR in VeroE6 cells ( data not shown ) . PKR is induced by IFN-α/β treatment [38] and the abundance of PKR could affect the results of the IP-kinase assay . We suspected that replication of MP-12 mutants lacking NSs may induce IFN-β production , leading to PKR induction [38] , whereas ActD treatment prevented the IFN-β production [39] . Indeed , IFN-β mRNA accumulation occurred at 8 h . p . i . in rMP-12-rLuc-infected cells in the absence of ActD , but not in the presence of ActD ( Figure 6C ) , a finding which suggested that ActD treatment inhibited the transcriptional induction of PKR which was induced by the type I IFN in infected 293 cells . We noted an efficient accumulation of rMP12-rLuc N mRNA at 8 h . p . i . in the presence of ActD ( Figure 6C ) . Because rMP-12-rLuc replication in the presence of ActD did not induce an accumulation of phosphorylated eIF2α early in the course of the infection ( Figure 4A ) , this efficient rMP12-rLuc replication probably occurred prior to 8 h . p . i . in the presence of ActD . As shown in Figure 6D , immunoprecipitated PKR from rMP12-rLuc-Flag-infected cells was phosphorylated both in the presence and absence of ActD , which led us to suggest that rMP12-rLuc-Flag replication activated PKR . In contrast , PKR phosphorylation did not occur in mock-infected cells or cells infected with rMP12-NSs-Flag or rMP12-PKRΔE7 ( Figure 6D ) . Most probably , a dominant-negative PKRΔE7 suppressed PKR activation and prevented PKR phosphorylation [36] . To determine why we failed to detect the presence of PKR autophosporylation in rMP12-NSs-Flag-infected cells , we examined the amounts of the immunoprecipitated PKR in these samples ( Figure 6D , bottom ) . Strikingly , substantial reductions in the abundance of cytoplasmic PKR occurred only in the cells supporting rMP12-NSs-Flag replication both in the presence and absence of ActD . We subsequently tested the possibility that the NSs downregulated PKR expression or sequestered PKR into the nuclear compartment , leading to the suppression of eIF2α phosphorylation . 293 cells were mock-infected or infected with rMP12-NSs-Flag , rMP12-rLuc-Flag or rMP12-PKRΔE7 , treated with ActD , and cytoplasmic and nuclear fractions of cell extracts were prepared at 16 h . p . i . Western blot analyses showed the presence of RVFV NSs both in the cytoplasmic and nuclear fractions , while rLuc and PKRΔE7 signals were observed only in the cytoplasmic fraction . A substantial reduction in PKR abundance occurred in both the cytoplasmic and nuclear fractions of cells infected with rMP-12-NSs-Flag , but not in mock-infected cells and in those infected with rMP12-rLuc-Flag or rMP12-PKRΔE7 ( Figure 7A ) , demonstrating that NSs induced the PKR downregulation in the infected cells . The PKR downregulation also occurred in MP-12-infected VeroE6 cells and in MRC-5 cells that were infected with the wt ZH501 strain of RVFV ( data not shown ) . To determine whether NSs expression alone induces PKR downregulation , 293 cells were mock-infected or infected with rMP12-rLuc , immediately transfected with in vitro-synthesized RNA transcripts encoding rLuc or NSs , and treated with ActD ( Figure 7B , left panel ) . Cells were harvested at 16 h post-transfection . The clear reduction in the PKR abundance occurred in mock-infected cells expressing NSs , but not in those expressing rLuc ( Figure 7B , left panel ) , demonstrating that NSs protein alone exerted the PKR downregulation . We subsequently examined the requirement of the ActD treatment for the PKR downregulation . 293 cells were transfected with in vitro-synthesized RNA transcripts encoding NSs or rLuc in the absence of ActD . Analysis of cell extracts harvested at 8 h post-transfection showed the PKR downregulation in the NSs-expressing cells ( Figure 7B , right panel ) , demonstrating that the NSs-mediated PKR downregulation occurred in the absence of ActD . To further understand the mechanism of the NSs-mediated PKR downregulation , we examined whether NSs expression promoted degradation of PKR mRNA . 293 cells were mock-transfected or transfected with in vitro-synthesized RNA transcripts encoding NSs or rLuc . Then the cells were mock-treated or treated with ActD . Total RNAs were harvested at 8 h post-transfection and the expression levels of PKR mRNA were examined by quantitative real-time reverse transcription polymerase chain reaction ( RT-PCR ) analysis ( Figure 7C ) . The relative expression levels of PKR mRNA were significantly increased in cells that were transfected with rLuc RNA transcripts or NSs RNA transcripts both in the absence and presence of ActD , which demonstrated to us that NSs expression did not promote the degradation of PKR mRNA . Efficient ActD-mediated suppression of IFN-β mRNA accumulation in rMP-12-rLuc-infected 293 cells ( Figure 6C ) led us to suggest that unexpected increases in the abundance of PKR mRNA in the RNA-transfected 293 cells in the presence of ActD were IFN-β-independent . We suspect that the RNA transcripts that were taken up by the cells induced robust PKR mRNA synthesis prior to ActD- or NSs-mediated general transcription suppression . Since NSs expression did not decrease the abundance of PKR mRNA ( Figure 7C ) , we next tested whether putative NSs-mediated translational inhibition leads to a reduction in the abundance of PKR . To this end , 293 cells were mock-infected , infected with MP-12 , or transfected with in vitro-synthesized RNA transcripts encoding NSs . Then cells were incubated with 100 µg/ml of puromycin to completely shut off cellular translation or puromycin untreated , and cell extracts were harvested at 16 h . p . i . or post-transfection . As expected , puromycin treatment completely abolished the synthesis of N and NSs proteins in MP-12-infected cells ( Figure 7D , lane 4 ) and that of NSs in cells transfected with the RNA transcripts encoding NSs ( Figure 7D , lane 6 ) . We found that treatment of 293 cells with puromycin for 16 h decreased the abundance of PKR only slightly ( Figure 7D , lane 2 ) . Accordingly , it is highly unlikely that putative NSs-induced translational inhibition was the main reason for the reduction in PKR abundance . We next performed pulse-chase experiments to know whether NSs expression promoted PKR degradation . Because immunoprecipitation experiments using various anti-PKR antibodies failed to convincingly demonstrate a radiolabelled endogenous PKR signal from extracts of 293 cells ( data not shown ) , we examined the effect of NSs expression on the stability of an expressed mutant PKR , PKRK296R , lacking kinase activity [40] and carrying a N-terminal myc tag; expression of wild-type PKR was not used due to its strong host translation suppression effects ( data not shown ) . 293 cells were mock-transfected or transfected with pcDNA3 . 1-Myc-PKRK296R , a plasmid encoding PKRK296R under cytomegalovirus promoter and radiolabelled with [35S] methionine/cycteine between 14 and 16 h post–DNA transfection . After pulse-radiolabelling cell extracts were prepared from some samples . In other samples , cells were transfected with in vitro-synthesized RNA transcripts encoding rLuc or NSs , and cell extracts were harvested at 8 h post–RNA transfection . Then the cell extracts were subjected to radioimmunoprecipitation analysis using anti-myc antibody , which immunoprecipitated the pulse-radiolabelled , myc-tagged PKR ( Figure 7E , lane 2 ) . After 8 h chase , the amount of the radiolabelled myc-tagged PKR was clearly reduced in the cells expressing NSs ( Figure 7E , lane 4 ) , but not those expressing rLuc ( Figure 7E , lane 3 ) . Instead , the amount of radiolabelled myc-tagged PKR was increased slightly in rLuc-expressing cells , presumably due to radiolabelling of continuously synthesized myc-tagged PKR by residual [35S] methionine/cysteine . The complete shutoff of cellular translation by puromycin for 16 h did not result in the loss of endogenous PKR ( Figure 7D ) , whereas as early as 8 h post-transfection of RNA transcripts encoding NSs , the abundance of PKR decreased substantially ( Figure 7B ) without reducing the abundance of PKR mRNA ( Figure 7C ) . Furthermore , NSs expression reduced the abundance of myc-tagged PKR that had been radiolabelled prior to NSs expression ( Figure 7E ) . These data strongly suggested that NSs induced the downregulation of PKR at a post-transcriptional level and pointed to a possibility that the NSs promoted PKR degradation . We next tested whether RVFV NSs downregulated PKR by promoting PKR degradation through a ubiquitin-proteasome pathway . 293 cells were infected with rMP12-rLuc or MP-12 , and were immediately treated with proteasome inhibitor MG132 or lactacystin . Cells were harvested at 8 h . p . i . and analyzed in Western blotting . As expected , rMP12-rLuc replication did not induce PKR downregulation ( Figure 7F , lane 1 ) , while MP-12 replication induced the reduction of PKR abundance ( Figure 7F , lanes 4 ) . It was evident that the treatment of MP-12-infected cells with those proteasome inhibitors suppressed NSs-induced PKR downregulation ( Figure 7F , lanes 5 and 6 ) , suggesting that NSs promoted PKR downregulation by the degradation through the proteasome pathway . Phorbol 12-myristate 13-acetate ( PMA ) , a potent activator of protein kinase C ( PKC ) , induces PKR degradation [41] . Because the general PKC inhibitor , GÖ6983 , suppresses PMA-mediated PKR degradation [41] , we examined whether the NSs promoted PKR downregulation through PKC by treating MP-12-infected cells with GÖ6983 . Treatment of GÖ6983 did not inhibit the NSs-mediated PKR downregulation ( Figure 7F , lane 8 ) , suggesting that PKC was not involved in NSs-induced PKR downregulation . To know how quickly PKR downregulation occurred in MP-12-infected cells , whole-cell extracts were prepared at 2 , 4 , 6 and 8 h . p . i . from MP-12-infected 293 cells and the amounts of NSs and PKR were determined ( Figure 7G ) . In the absence of proteasome inhibitor , a substantial reduction of PKR abundance occurred as early as 4 h . p . i . , where the abundance of NSs ( Figure 7G , lane 2 ) was not as great as that at 8 h . p . i . ( Figure 7G , lane 4 ) . In the presence of MG132 , the abundance of PKR decreased only slightly , and NSs accumulation was also somewhat less efficient . A similar trend for less efficient NSs accumulation in the presence of proteasome inhibitors was also shown in the data for Figure 7F . Our observation of less efficient NSs accumulation in MG132-treated cells was consistent with the report that MG132 induces eIF2α phosphorylation through GCN2 activation and translational suppression [42] . The abundance of NSs in MG132-treated cells at 8 h . p . i . and that in untreated cells at 4 h . p . i . was similar , and yet PKR abundance in the former cells was clearly higher than that in the latter cells . Taken together , these data strongly suggested that RVFV NSs promoted PKR degradation through the proteasome-dependent pathway .
The present study explored a novel function of the RVFV NSs protein by testing the replication of RVFV lacking the NSs gene initially in type I IFN-incompetent VeroE6 cells [43] , [44] in the presence of ActD or α-amanitin , which served as a surrogate of the host mRNA synthesis suppression function of the NSs . The NSs protein was essential for efficient virus replication in the presence of ActD or α-amanitin . We found that the replication of RVFV lacking the NSs gene in the presence of a transcription inhibitor induced an accumulation of phosphorylated eIF2α , The accumulation of phosphorylated eIF2α in the presence of transcriptional inhibitors , did not occur in VeroE6 cells that were infected with RVFV expressing a dominant-negative form of PKR ( PKRΔE7 ) and in MEF cells lacking functional PKR ( Figure 5 ) . These findings suggested to us that PKR played a major role in increasing phosphorylated eIF2α; however , it is currently unclear whether other kinases , such as PERK or GCN2 , have a possible role in eIF2α phosphorylation in rMP-12-rLuc-infected cells in the presence of transcriptional inhibitors . The endogenous PKR-mediated eIF2α phosphorylation suppressed translation of the RVFV lacking the NSs gene , resulting in poor virus replication . Our data further suggested that the NSs promoted PKR degradation most probably through a proteasome-dependent pathway and prevented eIF2α phosphorylation , leading to efficient viral translation . The past studies [25] , [26] , [45] and the data shown in this study illustrate that two distinct biological activities of the RVFV NSs protein worked together to secure efficient RVFV replication . Namely , nuclear NSs protein inhibits transcription of host mRNAs , including the IFN mRNAs; this activity is critical for efficient RVFV replication in IFN-competent systems [25] , [45] . However , a combination of RVFV replication and NSs-mediated host mRNA transcriptional suppression possibly induces PKR activation and subsequent eIF2α phosphorylation as a combination of RVFV replication and ActD or α-amanitin treatment induced it . The NSs protein , in turn , promotes PKR downregulation as early as 4 h . p . i . , and prevents eIF2α phosphorylation to secure the translation of viral mRNAs and efficient virus replication . We think that both of these two NSs functions are tightly related and protect efficient viral replication by suppressing host antiviral responses . In RVFV-infected cells , the NSs establishes general transcriptional suppression at a later stage of infection ( after 8 h . p . i . ) [25] , while NSs also suppresses specific IFN-β mRNA transcription at early stages of infection ( about 3 h . p . i . ) by maintaining repressor complex including SAP30 on IFN-β promoter [26] , [45] . PKR downregulation early in infection is probably important for maintaining efficient viral translation in combination with the suppression of host IFN responses . We suspect that the NSs-mediated PKR downregulation activity is important for RVFV replication and survival in infected mammalian hosts . RVFV-infection in rhesus monkeys showed that type I IFN is detectable around 1 day post-infection in both clinically ill surviving monkeys and lethally infected monkeys , and one dead monkey even kept high titer of IFN ( 120 to 480 U/ml ) from 1 day post-infection [46] . In fact , the best correlation with outcome was early detection of IFN and not necessarily the height of the response . NSs suppresses IFN-β mRNA transcription early in the course of infection in cultured cells [26] , [45] . Our present data suggest that the early downregulation of PKR in RVFV-infected cells might contribute to the inhibition of type I IFN induction , because PKR is known to serve as a pathogen-recognition receptor [47] . Furthermore , some of the host pathogen-recognition receptors , e . g . , toll-like receptors 3 and 7 , in uninfected cells located near the RVFV-infected cells , recognize virus-specific signals , e . g . , viral RNAs in the virus particles or viral dsRNAs in the cell debris from infected cells , leading to type I IFN production . Because the transcriptional promoter of the PKR gene contains an IFN-stimulated response element , and IFN stimulation induces PKR mRNA transcription [38] , PKR abundance is likely increased in many uninfected cells of infected animals; RVFV needs to replicate in such cells to survive in infected hosts . The reduction in PKR abundance occurred as early as 4 h . p . i . ( Figure 7G ) and the NSs protein is produced very early in RVFV infection [48] . Immediate NSs synthesis and subsequent NSs-induced PKR downregulation in the cells , some of which have increased PKR abundance , could rapidly disarm the PKR-mediated , antiviral functions and contribute to RVFV replication and survival in animal hosts . ActD treatment of uninfected cells resulted in a significant reduction of the polysome fraction , but it did not abolish the translational activities ( Figure S1 ) . rMP12-rLuc replication in the cells treated with ActD or α-amanitin promoted the accumulation of the phosphorylated eIF2α , resulting in translational suppression of viral proteins ( Figure 4 ) . The increased amounts of phosphorylated eIF2α clearly correlated with the concentration of ActD or α-amanitin in rMP12-rLuc-infected cells ( Figure S2 ) . In contrast , rMP12-rLuc replication in the absence of transcriptional suppressor induced only low levels of eIF2α phosphorylation ( Figure 4 ) . Yet , the phosphorylation of PKR occurred in cells supporting the replication of RVFV lacking NSs both in the absence and presence of ActD ( Figure 6D ) . Thus , a significant accumulation of phosphorylated eIF2α occurred only in cells in which replication of RVFV lacking NSs was combined with the treatment with transcriptional inhibitors . Several different mechanisms are conceivable for the accumulation of phosphorylated eIF2α in rMP12-rLuc-infected in cells treated with ActD or α-amanitin ( Figure 4 ) . One possible mechanism relates to the eIF2α dephosphorylation step . Phosphorylation of eIF2α at Serine 51 induces a rapid synthesis of activating transcription factor ( ATF ) -4 mRNA , which can be translated in the presence of phosphorylated eIF2α [49] . Expressed ATF4 then induces GADD34 mRNA transcription [50] and GADD34 protein interacts with type 1 protein serine/threonine phosphatase , PP1 , and this complex dephosphorylates eIF2α to resume the cellular translation [51] . rMP12-rLuc replication in transcriptionally active cells probably induced PKR activation and eIF2α phosphorylation , the latter of which then induced GADD34 upregulation and subsequent eIF2α dephosphorylation allowing efficient viral translation . rMP12-rLuc replication in cells treated with ActD induced PKR activation , the extent of which was similar to that of infected , ActD-untreated cells ( Figure 6D ) , whereas the transcription inhibitors would prevent GADD34 upregulation and subsequent eIF2α dephosphorylation , causing an accumulation of phosphorylated eIF2α , which inhibited viral translation . Other possible mechanisms relate to the failure to suppress the PKR function . Cells infected with adenovirus , human immunodeficiency virus-1 , or herpes simplex virus undergo a dramatic increase in the abundance of Alu RNA , which carries a repetitive element of ∼300-nt in length that is transcribed by RNA polymerase III [52] , [53] , [54] , [55] . Alu RNA forms a stable complex with PKR , and antagonizes the PKR activation [56] . If rMP12-rLuc replication induces Alu RNA accumulation to prevent PKR activation , then the drug-induced transcriptional suppression would inhibit Alu RNA accumulation , preventing Alu RNA-mediated PKR inactivation . Another possibility is that rMP12-rLuc replication transcriptionally induces host mRNAs , and some of their gene products have PKR inhibition activities . An example of this possibility is that influenza virus replication induces P58IPK , an inhibitor of PKR [57] . ActD or α-amanitin treatment suppresses the expression of the putative PKR inhibitor , resulting in an accumulation of phosphorylated eIF2α . Alternatively , transcriptional suppression may result in reduced amounts of ribosomal proteins , such as L18 , which binds to PKR and inhibits PKR activation [58] , [59] . These possibilities are not mutually exclusive , and the combination of these possibilities may contribute to the accumulation of phosphorylated eIF2α in the cells supporting replication of RVFV lacking the NSs in the presence of ActD or α-amanitin . Although many viruses suppress PKR function using various strategies [32] , poliovirus is known to promote PKR degradation in infected cells [60] , [61]; poliovirus RNA and viral protein components are required for this activity [61] , and both eIF2α and PKR are highly activated in poliovirus-infected cells before PKR downregulation occurs [60] . In contrast to poliovirus , only RVFV NSs protein was probably required for PKR downregulation since the downregulation of PKR occurred as early as 4 h . p . i . in RVFV-infected cells ( Figure 7G ) , whereas phosphorylated eIF2α accumulation occurred around 8 h . p . i . in rMP-12-rLuc-infected cells in the presence of ActD ( Figure 4 ) . Furthermore , NSs downregulated PKRK296R , a non-phosphorylatable mutant PKR ( Figure 7E ) . These data suggest that the phosphorylation of PKR was not essential for the NSs-mediated PKR downregulation . A substantial reduction in the abundance of PKR occurred in both the cytoplasmic and nuclear fractions of MP-12-infected cells , but not in the mock-infected cells and in those infected with MP12 lacking NSs ( Figure 7 ) . Expression of NSs alone resulted in a reduction in the amount of PKR ( Figure 7B ) . These data established that NSs protein induced PKR downregulation . The NSs-induced PKR downregulation occurred as early as 4 h . p . i . of MP-12 ( Figure 7G ) . We also demonstrated the presence of PKR mRNA in NSs-expressing cells ( Figure 7C ) and that ActD treatment had little effect on the NSs-induced PKR downregulation ( Figure 7B ) . We failed to detect PKR at 16 h . p . i . of MP-12-infected 293 cells , and yet we easily detected the PKR after a 16 h-long incubation of 293 cells with puromycin ( Figure 7D ) . Accordingly , the putative NSs-induced translational inhibition was not a main reason for the reduction in the abundance of PKR . Furthermore , pulse-chase experiments showed that NSs expression reduced the abundance of radiolabelled myc-tagged PKR ( Figure 7E ) . All of these data strongly suggested that the NSs induced PKR downregulation at post-transcriptional levels . It was reported that treatment of a macrophage cell line with IFN-γ induced PKR degradation [62] . Because NSs induced PKR degradation in ActD-treated cells and ActD treatment completely inhibited the IFN-γ mRNA accumulation ( Figure 6C ) , it is highly unlikely that IFN-γ was involved in the NSs-induced PKR downregulation . PKC activation also potentially induces PKR downregulation [41] . Treatment of MP-12-infected cells with a PKC inhibitor , GÖ6983 , did not inhibit NSs-mediated PKR downregulation ( Figure 7D ) , which implied that PKC was not involved in it . Experiments using a proteasome inhibitor suggested that the NSs promoted PKR degradation through a proteasome pathway ( Figure 7F ) . MG132 can induce GCN activation and translational suppression [42] . In fact , treatment of MP-12-infected cells with MG132 and lactacystin moderately reduced the accumulation of NSs ( Figure 7F and 7G ) . Accordingly , a possibility still exists that the MG132-induced moderate translational inhibition affected the NSs-induced PKR degradation in MG132-treated cells . For example , if the NSs-induced PKR downregulation requires an unstable host protein , then the MG132-induced moderate translational suppression would prevent accumulation of this putative host protein , resulting in inhibition of the NSs-induced PKR downregulation . Obviously further studies are required to elucidate the mechanism of PKR downregulation mediated by RVFV NSs .
Vero E6 cells , wild-type mouse embryonic fibroblast ( MEF ) cells and Pkr0/0 MEF cells [37] were maintained in Dulbecco's modified minimum essential medium ( DMEM ) ( Invitrogen ) containing 10% fetal bovine serum . BHK/T7-9 cells [63] , which stably express T7 RNA polymerase , were grown in MEM-alpha ( Invitrogen ) containing 10% fetal bovine serum ( FBS ) . Penicillin ( 100 U/ml ) and streptomycin ( 100 µg/ml ) ( Invitrogen ) were added to the media . BHK/T7-9 cells were selected in medium containing 600 µg/ml hygromycin ( Cellgro ) . An RVFV vaccine candidate MP-12 [13] and recombinant MP-12 [19] were used for the experiments , and infectivity was assessed by a plaque assay in Vero E6 cells . Cells were treated with transcriptional inhibitors , ActD ( Sigma ) ( 5 µg/ml ) or α-amanitin ( Sigma ) ( 50 µg/ml ) immediately after infection or transfection . To induce the inhibition of proteasome function , cells were immediately treated with MG132 ( Sigma ) at 10 µM or lactacystin ( Sigma ) at 50 µM after infection or transfection . To suppress PKC activity , cells were treated with a general PKC inhibitor , GÖ6983 ( Calbiochem ) at 100 nM immediately after infection with rMP12-rLuc or MP-12 . Cells were treated with puromycin ( Cellgro ) at 100 µg/ml immediately after infection or transfection to inhibit cellular translation . Standard molecular biological techniques , including a PCR-based mutagenesis method , were used for plasmid constructions . PCR fragments encoding PKRΔE7 ORF with an N-terminal Flag-tag sequence were cloned between the Hpa I and Spe I sites of pProT7-S ( + ) plasmid [19] , designated as pProT7-S ( + ) -PKRΔE7 . PCR fragments encoding the rLuc ORF or NSs with a C-terminal Flag-tag sequence were cloned between the Hpa I and Spe I sites of the pProT7-S ( + ) plasmid , designated as pProT7-S ( + ) -rLuc-Flag or pProT7-S ( + ) -NSs-Flag , respectively . All of the constructs were confirmed to have the expected sequences . The PCR product of the entire human PKR ORF carrying a point mutation at K296R and the N-terminal myc tag was cloned between KpnI and XhoI of pcDNA3 . 1myc-His ( Invitrogen ) , resulted in pcDNA3 . 1-Myc-PKR296R . A recombinant MP-12 carrying the PKRΔE7 ORF , rLuc-Flag or the NSs-Flag in the place of the NSs ORF was recovered as described previously [19] . Briefly , subconfluent monolayers of BHK/T7-9 cells were co-transfected with an S-genome RNA expression plasmid , such as pProT7-S ( + ) -PKRΔE7 , pProT7-S ( + ) -rLuc-Flag or pProT7-S ( + ) -NSs-Flag , and a mixture of pPro-T7-M ( + ) , pPro-T7-L ( + ) , pT7-IRES-vN , pCAGGS-vG , and pT7-IRES-vL using TransIT-LT1 ( Mirus Bio Corporation ) . The culture medium was replaced with fresh medium 24 h later . At 5 days post-transfection , the culture supernatants were collected , clarified and then inoculated into VeroE6 cells . The supernatant of infected VeroE6 cells at 3 days post-infection was used for the experiment . RVFV MP-12 or ZH501 NSs ORF , or CAT ORF were cloned downstream of the T7 promoter between the Kpn I and Xho I sites of the pcDNA3 . 1-myc-HisA ( Invitrogen ) plasmid . For rLuc RNA transcripts , the rLuc ORF in pRL-SV40 plasmid ( Promega Corporation ) was inserted downstream of the T7 promoter . Capped and polyadenylated RNA transcripts were synthesized in vitro by using mMESSAGE mMACHINE T7 Ultra ( Ambion ) according to the manufacturer's instructions [64] . One microgram of in vitro-synthesized RNA transcripts was transfected into 293 cells in a 12-well plate with a TransIT-mRNA Transfection kit ( Mirus Bio Corporation ) according to the manufacturer's instructions . VeroE6 cells in 6-well plate were infected with a series of diluted virus samples in 400 µl . After 1 h adsorption at 37°C , we removed the inocula and added 2 ml of MEM containing 0 . 6% noble agar ( Difco Laboratories ) , 5% FBS and 5% tryptose phosphate broth . The cells were incubated at 37°C for 3 days . Then , 2 ml of MEM containing 0 . 6% agar , 100 µg of neutral red ( N2889 , Sigma ) , 5% FBS and 5% tryptose phosphate broth were added into the wells , and incubated for 16 h . The virus titers were determined in triplicate . Cells were lysed in sample buffer and boiled for 10 min . Equal amounts of samples were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) . Proteins were electroblotted onto polyvinylidene difluoride membranes ( immobilon P; Millipore ) . Western blot was performed as described previously [27] . The following primary antibodies were used: anti-RVFV [27]; anti-NSs [48]; anti-eIF2α ( #9722 , Cell Signaling Tech . ) ; anti–phospho-eIF2α ( S51 ) ( #9721 , Cell Signaling Tech . ) ; anti-Flag M2 ( F3165 , Sigma ) ; anti-PKR ( #3072 , Cell Signaling Tech . ) ; anti-Histone H1 ( sc-8030 , Santa Cruz Biotech . ) ; anti–c-Myc ( sc-40 , Santa Cruz Biotech . ) , and anti–β-actin ( sc-1616 , Santa Cruz Biotech . ) . The IP-kinase assay was performed as described previously [65] . Briefly , 293 cells were mock-infected or infected with recombinant MP-12 at an moi of 3 . Cells were dissolved in lysis buffer containing 10 mM Tris-HCl pH 7 . 6 , 50 mM KCl , 2 mM Magnesium acetate , 10 mM 2-mercaptoethanol , 1% Triton X-100 , 1 mM EDTA , phosphatase inhibitor cocktail ( Sigma ) and proteasome inhibitor cocktail ( Roche ) . After centrifugation at 10 , 000×g for 5 min , a cytoplasmic fraction was collected . Cytoplasmic lysates were subjected to immunoprecipitation with anti-PKR antibody ( Santa Cruz Biotech , K-17 ) . Protein A beads that bound to the immunoprecipitated PKR were washed twice with a buffer-A containing 20 mM Tris-HCl , pH 7 . 6; 50 mM KCl; 400 mM NaCl; 5 mM 2-mercaptoethanol; 1% Triton X-100; 1 mM EDTA; phosphatase inhibitor cocktail ( Sigma ) ; proteasome inhibitor cocktail ( Roche ) ; and 20% glycerol . The beads were further washed twice with buffer-B containing 20 mM Tris-HCl , pH 7 . 6; 100 mM KCl; 5 mM 2-mercaptoethanol; 1% Triton X-100; 0 . 1 mM EDTA; proteasome inhibitor cocktail ( Roche ) ; and 20% glycerol . Then , washed beads were resuspended in 2× kinase buffer containing 30 mM Hepes-KOH , pH 7 . 4; 2 mM dithiothreitol; 2 mM MgCl2; proteasome inhibitor cocktail ( Roche ) ; and 10 µCi of [γ-32P] ATP ( MP Biomedicals ) . The suspension ( 20 µl ) was incubated at 30°C for 20 min , and then 2× SDS sample buffer was added to terminate the reaction . The samples were separated on 10% SDS-PAGE gel and visualized on an autoradiograph . A portion of the samples were also used for Western blot analysis by employing anti-PKR monoclonal antibody ( BD biosciences ) to show the abundance of immunoprecipitated PKR . For the radiolabelling of host and viral proteins in infected cells , VeroE6 cells were incubated at 14 . 5 h post-infection for 30 min at 37°C with medium made up with MEM lacking methionine , cystine , and L-glutamine ( M2289 , Sigma ) ; 1% dialyzed FBS ( Invitrogen ) ; 20 mM L-glutamine; penicillin ( 100 U/ml ) and streptomycin ( 100 µg/ml ) . Then , Trans[35S]label metabolic reagent ( MP biomedicals ) was directly added into the medium ( 100 µCi/ml ) . After 1 h labelling , cells were washed with PBS once and lysed in sample buffer . Equal amounts of samples were subjected to SDS-PAGE in 10% polyacrylamide gel . The gel was dried and exposed to X-ray film ( KODAK BioMax XAR ) overnight at −80°C . For the radiolabelling of PKR , 293 cells were mock-treated or transfected with pcDNA3 . 1-Myc-PKRK296R , and labelled with Trans[35S]label metabolic reagent between 14 and 16 h post–DNA transfection . In some samples , cell extracts were prepared with chase using a lysis buffer . In other samples , cells were transfected with in vitro-synthesized RNA transcripts encoding rLuc or MP-12 NSs . Then , cells were washed , and incubated for 8 h in the presence of 2 mM methionine/cysteine . At 8 h post–RNA transfection , cell extracts prepared using lysis buffer , were employed for immunoprecipitation with anti-Myc antibody ( Santa Cruz: sc-40 ) , as described in an IP-kinase assay . Immunoprecipitated samples were separated on a 7 . 5% poryacrylamide gel and visualized on an autoradiograph . 293 cells were mock-infected or infected with rMP12-rLuc-Flag , rMP12-NSs-Flag or rMP12-PKRΔE7 at an moi of 3 . Alternatively , 293 cells were transfected with in vitro-synthesized RNA transcripts encoding MP-12 NSs . The cytoplasmic lysate was harvested at 16 h . p . i . or 16 h post-transfection , and incubated with poly C beads or poly I∶C beads on ice for 45 minutes . After washing beads with buffer-A for 4 times , the beads were mixed with 2× sample buffer , and bound proteins were analyzed with anti-Flag antibody ( Sigma ) on a Western blot . The luciferase assay was performed on a Renilla Luciferase Assay System ( E2810 , Promega Corporation ) according to the manufacturer's instructions . Total RNA was harvested by Trizol reagent ( Invitrogen ) , and Northern blot was performed as described previously [27] , [64] . Briefly , total RNA was denatured and separated on 1 . 2% denaturing agarose-formaldehyde gels and transferred onto a nylon membrane ( Nylon Membrane , positively charged , Roche ) . Northern blot analysis was performed by using strand-specific RNA probes for detecting IFN-β mRNA [64] , GAPDH mRNA [64] , rLuc mRNA [27] or RVFV antiviral-sense S-segment / N-mRNA [48] . 293 cells in 6-well plates were transfected with 2 µg of in vitro-synthesized RNA transcripts encoding rLuc or MP-12 NSs by TransIT mRNA transfection kit ( Mirus ) . Cells were mock-treated or treated with ActD ( 5 µg/ml ) immediately after RNA transfection . Total RNA were extracted by using an RNeasy Mini kit ( Qiagen ) at 16 h post-transfection . For each sample , we used 500 ng of RNA to synthesize 1st strand cDNA by High-Capacity cDNA Reverse Transcription Kits ( Applied Biosystems ) . Real-Time PCR was performed at the Real-Time PCR core facility , Sealy Center for Cancer Cell Biology , UTMB . We used an Applied Biosystems made-to-order 20× assay mix of primers and TaqMan MGB probes ( FAM dye-labled ) for our target gene PKR ( Applied Biosystems: assay ID#: Hs01091592_m1 ) and pre-developed an 18S rRNA ( FAM-dye labelled probe ) TaqMan assay reagent ( Applied Biosystems: 4352930E ) for endogenous control . Separate tubes ( singleplex ) real-time PCR was performed with 40 ng cDNA for both target gene and endogenous control by using a Taqman Gene expression master mix ( Applied Biosystems: 4370074 ) . The cycling parameters for real-time PCR were: UNG activation at 50°C for 2 min , AmpliTaq activation at 95°C for 10 min , denaturation at 95°C for 15 sec , and annealing/extension at 60°C for 1 min ( repeat 40 times ) on ABI7000 . Duplicate CT values were analyzed by the comparative CT ( ΔΔ CT ) method , as described by the manufacturer ( Applied Biosystems ) . The amount of target ( 2−ΔΔCT ) was obtained by normalized to endogenous reference ( 18S rRNA ) and relative to a calibrator ( one of the experimental samples ) .
|
The mosquito-borne bunyavirus Rift Valley fever virus ( RVFV ) devastates both humans and domestic animals; it causes abortions in ruminants and complications such as hemorrhage , encephalitis , or retinal vasculitis in humans . A major RVFV virulence factor , NSs , disables host cell mRNA synthesis . Here we describe our new evidence that showed NSs working in a second way; in addition to inhibiting host cell transcription , NSs kept translation active in infected cells . It is well-established that activated protein kinase PKR phosphorylates a translation factor , eIF2α , and then this phosphorylated eIF2α suppresses translation . We found that NSs decreased PKR abundance and prevented eIF2α phosphorylation in infected cells , allowing efficient viral translation and replication . In contrast , when cells were infected with an RVFV mutant lacking NSs in the presence of transcriptional inhibitors that mimic the transcription inhibition function of NSs , the PKR reduction did not occur and phoshorylated eIF2α was accumulated , resulting in the inhibition of virus gene expression and replication . Thus , NSs functions in two ways to help RVFV replicate in mammalian hosts: its newly identified PKR downregulation function secures efficient viral translation , and its host transcription inhibition function suppresses the expression of host innate antiviral functions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"virology",
"virology/immune",
"evasion",
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression",
"virology/host",
"antiviral",
"responses"
] |
2009
|
Rift Valley Fever Virus NSs Protein Promotes Post-Transcriptional Downregulation of Protein Kinase PKR and Inhibits eIF2α Phosphorylation
|
Epithelial tubes are the functional units of many organs , and proper tube geometry is crucial for organ function . Here , we characterize serrano ( sano ) , a novel cytoplasmic protein that is apically enriched in several tube-forming epithelia in Drosophila , including the tracheal system . Loss of sano results in elongated tracheae , whereas Sano overexpression causes shortened tracheae with reduced apical boundaries . Sano overexpression during larval and pupal stages causes planar cell polarity ( PCP ) defects in several adult tissues . In Sano-overexpressing pupal wing cells , core PCP proteins are mislocalized and prehairs are misoriented; sano loss or overexpression in the eye disrupts ommatidial polarity and rotation . Importantly , Sano binds the PCP regulator Dishevelled ( Dsh ) , and loss or ectopic expression of many known PCP proteins in the trachea gives rise to similar defects observed with loss or gain of sano , revealing a previously unrecognized role for PCP pathway components in tube size control .
Multicellular animals employ tubular structures in organs to transport vital fluids and gases that sustain life . Examples of organs with prominent tubular architecture include the circulatory system , the lung and kidney in mammals , the secretory and respiratory organs in flies , and the excretory organ in worms . Proper development of tubular networks is critical for the function of several organs , evidenced by disruption of these networks being an underlying cause of common human diseases including cardiovascular disease , polycystic kidney diseases , and asthma . The Drosophila trachea is a branched network of tubular epithelia that transports oxygen and other gases throughout tissues . The comparative simplicity and genetic tractability of this system has made it one of the most powerful model systems to dissect tubular epithelial morphogenesis . Tracheal formation begins as tracheal placodes invaginate from the epidermis during early embryogenesis . Through stereotypic cell migrations , cell shape changes , and rearrangements of cell-cell junctions , tracheal cells generate a tubular network that extends branches to all embryonic tissues [1]–[4] . Each tracheal branch assumes a specific size as a consequence of branch-specific signaling events [5]–[10] . Tube size control is mediated by changes in cell shape , cell arrangement , and possibly cell size , but does not involve changes in cell number [11] . One category of genes that affect tube size encodes components of septate junctions , as mutations cause overelongated trachea [12]–[17] . Defects in apical extracellular matrix ( ECM ) proteins - which modify the structure of the chitin matrix - also lead to overelongated trachea , indicating that a dynamic and highly patterned apical extracellular matrix ( ECM ) regulates epithelial cell shape and tube size [18]–[22] . In epithelia , cells are polarized along the apical/basal axis . In epithelial tubes , the apical surface of each cell faces the lumen , whereas the basal surface faces surrounding tissues and/or a basement membrane . In addition to apical/basal polarity , epithelial cells in most tissues require information about their orientation within the plane , orthogonal to the axis of apical/basal polarity , in order to generate polarized structures such as cilia , or to move or orient themselves in a directed fashion . This type of polarization is referred to as planar cell polarity ( PCP ) . In vertebrates , PCP is involved in diverse patterning events , including convergence extension during gastrulation , neural tube closure , inner ear sensory hair morphogenesis , and hair follicle orientation [23] . In Drosophila , PCP biases cell orientation in several adult epithelial tissues and has been implicated in ovarian border cell migration [24]–[27] . In many contexts , both in vertebrates and in Drosophila , a conserved PCP pathway – the Frizzled ( Fz ) pathway - mediates local cell-cell interactions that instruct neighboring cells to adopt appropriate polarity [24]–[27] . In Drosophila , loss or overexpression of PCP proteins causes disorganization of wing hairs and bristles on the thorax and/or alteration in the orientation of ommatidia in the compound eye . Analysis of such phenotypes revealed an evolutionarily conserved set of genes that control planar polarity – the “core” PCP factors . These factors include: Fz , a seven-pass transmembrane receptor [28]; Dishevelled ( Dsh ) , an adaptor protein that acts downstream of Fz [29]–[31]; Flamingo/Starry night ( Fmi/Stan ) , a cadherin-family member with a seven-pass transmembrane domain [32] , [33]; Strabismus/Vang Gogh ( Stbm/Vang ) , a four-pass transmembrane protein [34] , [35]; and Prickle ( Pk ) and Diego ( Dgo ) , each cytoplasmic proteins that are associated with the apical membrane during PCP signaling [36] , [37] . PCP pathway activity itself leads to polarized enrichment and distribution of core components in all Drosophila tissues analyzed to date . In pupal wing cells , core PCP proteins localize apico-laterally , partially overlapping with cellular junctions [17] , [38] , and each protein is enriched in a distal and/or proximal location in the cells during prehair formation [33] , [37] , [39]–[42] . The function of each core PCP protein is essential for the asymmetric accumulation of the other proteins . The PCP signal from Fz/Dsh directs asymmetric cytoskeletal reorganization and polarized cell morphology , in part by activating RhoA/Rho1 [43] and its downstream effector , Drosophila Rho-associated kinase , Drok [44] . In the wing , RhoA signals via Drok , which regulates myosin II activity via phosphorylation of Spaghetti squash ( Sqh ) , a Drosophila homolog of nonmuscle myosin II regulatory light chain ( MRLC ) [44] , [45] . Additional PCP regulators include Fat ( Ft ) and Dachsous ( Ds ) , two protocadherins that can interact in a heterophilic fashion across cell boundaries [46] , [47] , and the Golgi kinase Four jointed ( Fj ) [48] , [49] . Fj and Ds are expressed in a gradient in the eye and wing , making these proteins attractive candidates for providing upstream global polarity cues [46] , [47] . Alternatively , the Ft/Ds group may function in parallel to the core PCP proteins [50] . Recently , a role for PCP genes in regulating tube length and diameter by orienting cell divisions was demonstrated in vertebrate renal and gut epithelia [51] , [52] , but whether the PCP components affect tube geometry in Drosophila remains unknown . Here we identify serrano ( sano ) , a novel protein that affects tracheal tube length in Drosophila . sano mutant embryos have elongated tracheal dorsal trunks ( DTs ) , whereas overexpression of Sano results in shortened DTs . Sano directly binds the core PCP component Dsh , and tracheal morphology and geometry are similarly affected by alterations in Sano activity and PCP signaling . Our results implicate for the first time the PCP mutants in Drosophila tubular morphogenesis .
An enhancer trap screen for lines with expression in the developing salivary gland and trachea identified rp395 , a P-element insertion that expresses β-gal throughout the salivary gland , in trachea , and in several other embryonic tissues , including the hindgut , midgut endoderm , CNS midline , posterior spiracles , and epidermis ( Figure 1A–1D ) . Cloning and characterization of the flanking region revealed that the rp395 P element had inserted after nucleotide 14 of the RC and RD transcripts of serrano ( sano; CG12758 ) , two of five alternatively spliced transcripts , designated RA-RE ( Figure 1M ) . Northern analysis revealed a single size transcript of 4 . 6 kb , first detected in 4–8-hour embryos and reaching peak levels in 8–12-hour embryos ( Figure S1 ) . The transcript was detected at all subsequent developmental stages , but was not detected in RNA isolated from cultured Drosophila Schneider ( S2 ) cells . With minor exceptions , the endogenous sano transcripts recapitulate the pattern of rp395 β-gal expression ( Figure 1E–1H ) . Sano expression requires the transcription factors Sex combs reduced ( Scr ) in the salivary gland , Trachealess ( Trh ) in the trachea , and Single-minded ( Sim ) in the CNS midline ( Figure 1I–1L ) ; sano expression was not affected by loss of transcription factors including fork head , huckebein , or CrebA that are expressed early in salivary gland formation ( data not shown ) . Early transient tracheal expression of sano was observed in trh mutant cells also deficient for programmed cell death ( Df ( 3L ) H99 ) , suggesting that initial tracheal expression is in part trh-independent and complete loss of sano expression in trh mutants is due to tracheal cell death ( Figure S2 ) . Since other known regulators of tracheal development , including ventral veinless/drifter , trachea defective/apontic , breathless , branchless , and rhomboid , did not affect sano expression ( data not shown ) , initial sano expression could be regulated by factors that initiate trh expression . Since Trh and Sim bind the same consensus DNA sequence [53] , [54] , regulation of sano expression by these proteins could be direct . All predicted sano splice forms encode the same 778-residue ORF ( Figure 1M ) . Sano is highly conserved in arthropods ( Figure 2A ) , and is a member of a largely uncharacterized family of proteins with members from cnidarians to mammals that includes the recently identified Themis protein ( also known as Gasp ) . Themis/Gasp is a cytosolic thymocyte-adaptor protein that binds Grb2 and is required for positive selection of thymocytes [55]–[60] . Because we were unable to generate antiserum that detected endogenous Sano , we cloned and expressed both untagged and C-terminally tagged ( GFP or HA ) Sano in flies under Gal4/UAS control [61] . In both tracheal and salivary gland cells , each version of overexpressed Sano localized diffusely in the cytoplasm , with enrichment at apical membranes , colocalizing with the apical membrane markers Crumbs ( Crb ) and Stranded at Second ( SAS ) ( Figure 2B–2D; [62]–[64] ) . During late embryogenesis and in the 3rd instar larval salivary gland , however , Sano-GFP also localized to nuclei ( Figure 3F , 3F′ , and 3G and data not shown ) . Neither untagged ( detected with Sano antiserum ) nor HA-tagged Sano could be detected in nuclei at any stage . Taken altogether , these experiments suggest that Sano is an apically enriched cytoplasmic protein that may also sometimes localize to nuclei , a localization similar to that reported for the mammalian Themis/Gasp protein [55]–[60] . Three independent loss-of-function knock-out sano alleles , sanoKO1 , sanoKO2 and sanoKO3 , were generated by homologous recombination [65] . PCR analysis confirmed that exons common to all five splice forms were replaced with the mini-white+ gene ( Figure S3A and S3B ) . sano mRNA was not detected in sano homozygotes or in embryos transheterozygous for each sano allele over a deficiency that removes sano and nearby genes , indicating that the sano alleles are null ( Figure 1N and 1O; Figure S3C , S3D , S3E , S3F , S3G , and S3H; data not shown ) . Each sano allele is homozygous lethal , and lethal over the sano deficiency , with death occurring during the 2nd instar larval stage . The sano lethality was partially rescued by expression of the Sano ORF under the control of a heat-shock promoter ( HS-Sano ) induced during larval stages ( 11/45 viable adults when heat shocked at 58–70 hr AEL ) . Most features of salivary gland and tracheal development appear normal in sano mutant embryos ( data not shown ) . Interestingly , however , staining with 2A12 , a marker of tracheal lumen after stage 13 , revealed that the dorsal trunk ( DT ) in sano mutants is more elongated and convoluted than in wild type ( WT; Figure 3A–3D ) . Measurements of DT lengths from confocal projections of 2A12 staining from lateral views of stage 16 embryos revealed that sano mutant DTs are significantly ( ∼12% ) longer than wild type ( Figure 3E ) . Tracheal cell numbers in the dorsal trunk of sano mutants ( 14 . 7±0 . 6 , N = 5 , metamere 4 ) were comparable to those of WT ( 15 . 2±0 . 4 , N = 5 , metamere 4; p>0 . 5 , t-test ) , indicating that the elongated DT phenotype is not due to increased cell numbers . Conversely , Sano overexpression using btl-Gal4 caused shortened DTs with discontinuous staining with either 2A12 or SAS ( Figure 3F and 3F′ ) . The UAS-Sano-GFP distributions in tracheal cells revealed that cells connecting adjacent segments of the DT ( fusion cells ) contact each other basally , but that the tracheal lumens and apical membranes are discontinuous . Fusion cell markers including Dysfusion ( Dys ) , a bHLH-PAS transcription factor [66] , and Arf-like-3 ( Arl3 ) , a small GTPase [67] , were normally expressed in the discontinuous region of the DTs , indicating that fusion cells are not transformed to another fate ( Figure 3G; data not shown ) . No increase in apoptosis was detected in the Sano-overexpressing trachea ( Figure S5 ) , and tracheal cell numbers in the btl-Gal4:UAS-Sano trachea ( 16±0 . 7 , N = 7 , metamere 4 ) were comparable to WT ( 15 . 2±0 . 4 , N = 5 , metamere 4; p>0 . 1 , t-test ) , indicating that the shortened DT phenotype is not due to reduced numbers of tracheal cells . At 25°C , 100% of btl-Gal4:UAS-Sano embryos showed apical disconnection of DTs in more than one metamere , whereas neither btl-Gal4 nor UAS-Sano alone had the shortened apical DT phenotype ( Table 1 ) . Sano overexpression also caused mismigration and/or failure of other tracheal branches to connect ( data not shown ) . btl-Gal4-driven Sano expression in the trachea of sano null mutants rescued the elongated DT phenotypes observed in sano mutants ( Figure 3E ) and alleviated the gain-of-function phenotype of shortened DTs ( Table 1 ) , suggesting that an optimal dose of Sano is critical for proper tube length and that tube length is inversely related to Sano levels . Sano overexpression also reduced salivary gland lumenal length ( 88 . 7±2 . 0µm ( WT ) vs . 58 . 9±5 . 4µm ( Sano-overexpressing glands ) , N = 5 for each genotype; p<0 . 01 , t-test ) , suggesting that Sano has generalized effects on tube length ( Figure 3H and 3I ) . Tracheal tube size is controlled neither by the number nor the overall size of the individual cells [11] . Nonetheless , mutations in several genes have been discovered that , like loss of sano , lead to tracheal tube overelongation . Most of these known genes either regulate chitin synthesis or encode components of the septate junction , an invertebrate structure that has trans-epithelial barrier functions analogous to the vertebrate tight junction [16] , [18]–[22] , [68] . To test whether sano function is linked to either category of known genes affecting tube length , we analyzed luminal chitin using a fluorescent chitin binding protein ( CBP ) and a fluorescent chitin binding lectin ( Wheat Germ Agglutinin; WGA ) [21] . Both reagents revealed that the chitin cable , an extracellular scaffold upon which the tracheal branches elongate , is normal in sano mutants ( Figure S4A and S4B; data not shown ) . vermiform ( verm ) encodes an apically-secreted chitin-binding protein with predicted polysaccharide deacetylase activity [12] , [18] , [22] . Verm staining in sano mutant trachea was indistinguishable from WT ( Figure S4C and S4D ) . We conclude that tracheal length defects in sano mutants are not a consequence of detectable alterations in chitin biogenesis . Septate junction proteins , including Coracle ( Cor ) , Neurexin IV ( NrxIV ) , and Fasciclin 3 ( Fas3 ) , localized normally to the basolateral domain of sano mutant tracheal cells , suggesting that septate junctions are intact ( Figure S4E and S4F; data not shown ) . A 10 kDa dextran dye exclusion assay indicated that barrier function of septate junctions is intact in sano mutants ( Figure S4G , S4H , and S4I ) . Thus , neither septate junction function nor chitin cable assembly is disrupted in sano mutants , suggesting another mechanism for the elongated tracheal phenotype . sano is dynamically expressed in larval imaginal discs , structures that give rise to much of the adult during metamorphosis ( Figure S6 ) . Overexpression of Sano using several imaginal disc-specific Gal4 drivers caused planar polarity defects . For example , in the wild-type adult thorax , bristles point posteriorly , whereas in Sano-overexpressing adult thoraces , the bristles displayed altered orientations ( Figure 4A and 4B ) . In the WT wing , each cell produces a single distally-oriented , actin-rich protrusion ( a trichome , a . k . a . a “hair” ) . All Sano-overexpressing wing cells exhibited swirling hair patterns ( Figure 4C and 4D ) . Sano overexpression in the eye caused ommatidial polarity defects , including misoriented and symmetrical photoreceptor phenotypes , as well as abnormal photoreceptor numbers ( Figure 4E and 4F; data not shown ) , with about 14 . 5% of the ommatidia showing defects ( 137/916 , N = 5 ) . Polarity defects observed with Sano overexpression are similar to those observed when PCP genes are mutated or overexpressed [33] , [36] , [37] , [43] , [69]–[73] , suggesting that Sano perturbs PCP . Next we examined Sano-overexpressing pupal wing cells . Phalloidin staining of actin-rich prehairs at 32 hours after puparium formation ( APF ) revealed that hair formation is delayed in Sano-overexpressing cells ( Figure 5A ) , as observed in dsh mutant clones or in dgo pk double mutant clones [74] . Phalloidin staining of the slightly older pupal wings ( at 33–34 hours APF ) revealed Sano-overexpressing cells with prehairs the same size as surrounding wild-type hairs but with altered polarity ( Figure 5B ) . Sano overexpression sometimes produced multiple wing hairs , another PCP phenotype ( Figure S7B ) . Wild-type hairs near some sano-overexpressing clones exhibited non-cell-autonomous polarity defects ( Figure 5B , arrows; Figure S7C and S7D ) , distinct from those near fz or stbm/Vang mutant clones; wild-type cells proximal to fz clones or distal to stbm/Vang clones have reversed hair polarity [34] , [35] , [75] , which was not observed with Sano overexpression . All Sano overexpression clones that produced nonautonomous phenotypes mapped either between veins 3 and 4 , distal to the anterior crossvein , or between veins 4 and 5 , distal to the posterior crossvein ( Figure S7A; N>100 clones examined ) , both regions of which are sensitive to PCP alteration [49] . Core PCP proteins are asymmetrically localized in pupal wing cells during prehair formation and show typical “zigzag” localization patterns on the apical surfaces of the pupal wings [33] , [36] , [37] , [39] , [40] , [42] . When a PCP gene is mutated or overexpressed , other PCP proteins are typically mislocalized . Sano overexpression in pupal wings through either ptc-Gal4-driven expression or in Sano-overexpressing clones resulted in the mislocalization of all PCP proteins examined . Fmi , normally localized to both the proximal and distal sides of wing cells during prehair formation , was observed around the entire perimeter ( Figure 5C and 5D ) . A similar mislocalization was observed with Stbm and Pk ( Figure S8A; data not shown ) . Fz and Dsh , which normally localize to the distal side of the apical surface , exhibited reduced apical membrane distribution with Sano overexpression ( Figure 5E , Figure S8B ) . sano loss-of-function mutant clones in cells giving rise to adult tissues such as thorax and wing did not result in PCP phenotypes ( data not shown ) . Similarly , actin prehairs of sano mutant clones in pupal wing cells always pointed distally as in WT ( Figure S9A and S9B ) . Since mutations of some PCP genes , such as ft , show polarity defects only in very large clones [46] , we induced sano mutant clones at earlier time points to generate a range of sizes of clones missing sano function . Even very large clones did not exhibit PCP defects ( data not shown ) . However , although it was rare , when we induced clones relatively early ( 36–48 hours after egg laying ( AEL ) , we obtained only twin spots ( <5% , N = ∼70 ) , suggesting that the sano mutant cells either died or were eliminated from the wing epithelium ( Figure S9C and S9D ) . On the other hand , sano null eye clones had defects characteristic of loss of known PCP genes , including misoriented ommatidia and loss of asymmetry ( Figure 4G; Table 2 ) . sano null eye clones also often had abnormal numbers of photoreceptors ( Figure 4G; Table 2 ) . In 3rd instar eye discs , the expression of BarH1 , a marker for the R1 and R6 photoreceptors [76] , showed ommatidial misrotation , consistent with the adult phenotype ( Figure 4H and 4I ) , and the expression of mδ0 . 5-lacZ , a marker for the R4 photoreceptor , was absent or significantly reduced in sano null clones , consistent with a cell fate change of R4 to R3 , which has been observed with some PCP mutants , including fz and dsh ( Figure 4J and 4K; [77] ) . Our data suggest that although sano overexpression disrupts PCP signaling in multiple tissues , loss of sano results in a range of defects that are limited to fewer tissues . To determine if sano affects tube length by altering PCP signaling , we asked if other PCP mutants have tracheal length defects , including null mutants of the core PCP genes fz , dsh , fmi , dgo , stbm , and pk , the ft/ds group of PCP regulator genes fj , ft , and ds , and the PCP downstream effectors rhoA , Drok , zip and sqh . For dsh , a key hub in canonical Wingless ( Wg ) /Wnt signaling and in Fz-dependent PCP signaling , we used the dsh1 allele , which is defective for only its PCP function [69] , [70] . Interestingly , many PCP mutants had tracheal length defects , exhibiting similar elongated DT phenotypes as loss of sano ( Figure 6A–6F and 6I ) . Among the core PCP genes , fz , dsh and fmi had elongated DTs , whereas dgo , pk and stbm mutant embryos had normal DTs . Among the ft/ds PCP regulator group , fj and ds had elongated DTs . Among the PCP downstream effectors , rhoA and zip mutant embryos showed elongated DTs , revealing a potential role for the cytoskeleton in tracheal elongation . Drok mutant embryos also have convoluted trachea , but overall tracheal length was comparable to WT . ft and sqh mutant embryos had shorter DTs than WT , but the DTs were contiguous ( Figure 6I ) . Overexpression of Dsh or a constitutively-active form of RhoA in the trachea caused shortened DT defects with discontinuities similar to sano overexpression ( Figure 6G and 6G′; [78] ) , further implicating this pathway in apical cell surface elongation . Sano and Dsh are both cytoplasmic proteins , and Sano binds Dsh in yeast two-hybrid assays and co-immunoprecipitation ( co-IP ) ( Figure 6J and 6K ) , providing a physical link between Sano and PCP proteins that is consistent with genetic interactions between dsh and sano; double mutants of sano and dsh1 have elongation defects similar to those of sano or dsh1 alone , suggesting that Sano and Dsh act in a common pathway ( Figure 6I ) . Moreover , reduction of PCP function of Dsh ( dsh1/+ ) suppressed the Sano overexpression phenotype in the thorax , a finding also consistent with Sano acting through Dsh ( Figure S10 ) . The apical enrichment of Dsh in the late embryonic trachea and Fmi localization to the adherens junctions ( Figure S11 ) is consistent with PCP proteins acting at the apical membrane . These data suggest that Sano affects tube length by impinging on Dsh activity , likely through its role in PCP signaling . Also consistent with this model is our finding that sano;ft double mutant trachea have DT lengths that are intermediate between those of ft and sano mutants alone ( Figure 6I ) . Since wing epithelial cells become hexagonally packed prior to PCP proteins regulating hair formation [79] , we examined cell shape in Sano-overexpressing wing cells . As observed with other PCP mutants , Sano-overexpressing cells often assume a pentagonal shape instead of the typical hexagonal shape ( Figure 7A; [79] ) . Sano overexpressing cells also have smaller apical domains than surrounding wild-type cells ( Figure 7A; here , we define the apical domain as the area circumscribed by the zonula adherens , where E-Cad localizes ) . Using E-Cad staining , we measured the apical domain perimeters from several examples of single Sano-overexpressing cells and found a 29–41% decrease in the perimeters of the Sano-overexpressing cells compared to their wild-type neighbors ( Figure 7A and 7B ) . A decrease in apical domain size with Sano overexpression was also seen in adult tissues . For example , ptc-Gal4-driven Sano expression at the anterior-posterior wing margin resulted in a decreased distance between wing veins L3 and L4 compared to wild-type wings ( Figure 7C and 7D; 178 . 29±2 . 10 pixels vs . 210 . 91±4 . 58 pixels , N = 3 for each genotype; p<0 . 005 , t-test ) . Wing cell numbers in the area demarcated by veins L3 , L4 , the anterior crossvein , and an imaginary line starting from the tip of posterior crossvein and perpendicular to L3 , did not reveal a significant difference in cell number between WT and ptc-Gal4:UAS-Sano wings ( 354 . 3±16 . 9 vs . 360±17 . 3 , n = 3 for each genotype; p>0 . 5 , t-test ) , indicating that the decrease of the adult wing size is due to a decrease in apical domain size . Likewise , global wing expression of Sano , using the MS1096-Gal4 [80] , resulted in a decrease in overall wing size ( Figure 7E ) . To ask if sano affects tracheal tube length through changes in apical domain size , we examined E-Cadherin staining of WT , sano mutant , and Sano-overexpressing trachea . We also examined E-Cadherin staining in PCP mutants with altered tracheal tube length . Although it was difficult to ascertain differences in apical domain size of individual cells between WT and sano mutant trachea , which are expected to be at most ∼12% different , we observed a marked decrease in apical domain size in the Sano-overexpressing tracheal cells ( Figure 7G ) . A similar decrease in apical domain size was observed in Dsh-overexpressing tracheal cells ( data not shown ) . Moreover , the tracheal cells of rhoA , one of the PCP mutants , had larger apical domains than WT ( Figure 7H ) , indicating that the changes in tube length observed with Sano and other PCP genes are due to altered cell geometry and not altered cell arrangement . In escargot ( esg ) mutant trachea , where infrequent DT breaks occur , apical domain size was comparable to WT , suggesting that the smaller apical domain size observed with sano and dsh overexpression is not a due to a failure of adjacent DT segments to fuse ( Figure 7I and 7J ) .
Since the discovery that the Fz pathway controls PCP , many additional PCP components have been identified , including core factors , several PCP regulators , and general and tissue-specific downstream effectors [24] , [25] , [27] . Sano overexpression causes PCP defects in adult epithelial tissues as well as mislocalization of core PCP proteins . In wing cells , sano null cells appeared normal although we very occasionally obtained twin spot-only clones , suggesting a role for Sano in cell survival or in epithelial maintenance . It is unclear whether this function is related to PCP . On the other hand , sano loss in the eye gave rise to a range of defects , some of which are typical of PCP mutants , including loss of R4 cell specification , ommatidial misorientation , and loss of equatorial asymmetry . The direct physical interaction between Dsh and Sano ( Figure 6J and 6K ) provides potential mechanistic insight into Sano function . The interaction between Dsh and Sano appears quite different from that between Dsh and Naked cuticle ( Nkd ) , a Wingless ( Wg ) antagonist that also gives rise to PCP defects when overexpressed . Dsh participates both in canonical Wg/Wnt signaling and in Fz-dependent PCP signaling [29]–[31] . Like Sano , Nkd directly binds Dsh , and overproduced Nkd causes polarity defects and limits Wg signaling activity presumably by sequestering , degrading and/or modifying Dsh and thus blocking its participation in PCP [81] . Unlike Nkd , however , Sano overexpression does not cause defects typical of those seen when canonical Wg signaling is blocked . Moreover , whereas Nkd overexpression blocks Dsh activity , our studies of tube length control suggest that Dsh and Sano act in the same direction: gain or loss of Dsh mimics the gain or loss of Sano in the trachea . Similarly , dsh sano double mutants have the same tracheal length defects as each single mutant ( Figure 6I ) . Likewise , overexpression of either Dsh or Sano in the eye using sev-Gal4 causes similar changes in ommatidial polarity and rotation [82] ) . Finally , we showed that reduced dsh function suppresses the Sano overexpression PCP phenotypes in the thorax ( Figure S10 ) . Overall , the interaction and genetic data suggest that Sano and Dsh act together in a common pathway . It is intriguing that loss-of-function mutations in many , albeit not all , PCP genes result in similar tube elongation defects observed with loss of sano ( Figure 6 ) . PCP signaling can provide directional cues at the single cell level , such as directions on where to place the single hair within a Drosophila wing cell , or at the level of cell groups , such as controlling the organization of mechanosensory bristles in the Drosophila thorax and arrangement of photoreceptors in the Drosophila eye . PCP signaling also controls the behavior of cell populations undergoing extensive rearrangements , such as the dynamic morphogenetic changes that occur during body axis elongation in Drosophila and vertebrates and in ovarian border cell migration [23] , [27] , [83] . A recent study has implicated mammalian Fat4 , a vertebrate homologue of the Drosophila global PCP protein Fat , in promoting renal tubule elongation through its effects on oriented cell divisions [51] . In those studies , loss of Fat4 led to shorter renal tubules , a defect exacerbated by simultaneous loss of one copy of Vangl2 , a vertebrate homologue of the core PCP protein Stbm/Vang . Consistent with this finding , our studies reveal that , in the trachea , mutations in the proteins that negatively regulate Fat ( Ds and Fj ) and the Stbm/Vang complex ( Fz , Dsh ) have the opposite defect: longer tubes . In the case of the trachea , a tissue whose final cell divisions occur much earlier in development than when Sano affects tube length , the effects of the PCP pathway are on cell shape rather than on the orientation of cell division . Whether the subcellular mechanisms by which PCP genes regulate oriented cell divisions in vertebrates and apical membrane elongation in flies are similar or distinct is not clear , but the parallels in the two systems provide evidence for evolutionarily conserved functions for PCP genes in tubular architecture . The finding of a role for PCP genes in tube length control raises two crucial questions . ( 1 ) Are PCP proteins asymmetrically localized in tubular epithelia in the same way they are in wing , eye and border cells [23] , [26] , [27] ? ( 2 ) How do PCP genes regulate tube length ? We examined the subcellular localization of Dsh and Fmi in the tracheal cells , where Dsh localizes mainly in the cytoplasm and is enriched at the apical domain at later stages , and Fmi localizes to the adherens junctions ( Figure S11 ) . Unfortunately due to the irregular shape of tracheal cells and the three-dimensional structure of the tracheal tube , we could not determine with adequate resolution whether the PCP proteins are asymmetrically distributed . However , our data provides new insight into how PCP affects tracheal tube size . In the trachea , loss of sano or PCP function resulted in tubes that were 7–15% longer than WT based on apical domain measurements ( 2A12 staining ) . Since the sano trachea have the same number of cells as wild-type , each tracheal cell , on average , must have an apical domain that is approximately 12% longer than wild-type . Although an accurate measurement of apical dimensions in the trachea could not be obtained due to the shape and curvature of the tube , in rhoA mutant trachea , where the elongated DT defects were most obvious , the apical domains of the DT cells were consistently larger than WT ( Figure 7H ) . Similarly , E-Cad staining in dsh and ds mutant tracheal cells revealed slightly larger apical domains ( data not shown ) . Importantly , overexpression of Sano and Dsh resulted in tubes that were so much shorter than WT that the individual segments were often too short to anastomose . Examination of E-Cad staining in Sano- or Dsh-overexpressing trachea revealed markedly smaller apical domains in these cells , consistent with the decreased apical domain size observed in wing cells overexpressing Sano ( Figure 7G; data not shown ) . Thus , PCP components appear to control overall tube length by limiting the size of the apical domain . This activity could be mediated by increased linkage of the plasma membrane to the underlying cytoskeleton and/or by direct effects on plasma membrane growth by modulating relative levels of exoctyosis and endocytosis . In support of a link between PCP signaling and regulated vesicle trafficking , Rab11/Sec5-dependent recycling of E-cadherin has been implicated in junctional remodeling during hexagonal packing of wing cells , wherein the polarized recruitment of Sec5 is through the PCP protein Fmi [79] . Importantly , tracheal tube elongation has also been linked to regulated vesicle trafficking through Rab11 [64] , [84] and through core components of the secretory machinery [85] , [86] . There are two potential inconsistencies with a model that Sano functions to control tube length through its effects on PCP signaling: ( 1 ) Loss of sano does not give rise to overt PCP defects in all adult tissues and ( 2 ) not all of the components of the PCP signaling pathway disrupt tube length when their function is missing . Indeed , loss of ft , which is expressed early in the tracheal primordia , appears to have effects opposite those of loss of sano on tube length . Although the tissue-specificity of sano's role in PCP could reflect functional redundancy , it is also possible that sano and the large subset of the PCP genes that do have tracheal defects may function through novel , non-canonical , pathways to control tube length . In either case , it will be exciting to unravel the details of Sano's interactions with the cellular machinery to control apical domain size .
Fly strains used in this study were: fzP21 , fzK21 , pksple-13 , MS1096-Gal4 ( P . Adler ) ; Scr4 , trh1 , sim2 , megaEA97 , dsh1 , stan192 , stbm6 , pkeq , rhoA72F , rhoA72O , Drok2 , zip1 , zip2 , fjd1 , ds33K , ftGv-5 , esg35Ce-1 , Df ( 2L ) Exel8003 , Df ( 2R ) ED2076 , Df ( 2R ) ED3610 , Df ( 2R ) Bsc271 , sev-Gal4 , ( Bloomington stock center ) ; Df ( 2R ) Exel6068 ( Exelixis ) ; btl-Gal4 ( S . Hayashi ) ; sqhAX3 ( L . Luo ) ; ptc-Gal4 ( D . Pan ) ; HS-Scr ( M . Scott ) ; dgo380 , mδ0 . 5-lacZ ( D . Strutt ) ; fmiE59 ( T . Uemura ) ; sage-Gal4 ( A . Vaishnavi and D . J . A . , unpublished ) . The primary antibodies used were mouse α-β-gal ( Promega , 1∶500 ) , rabbit α-GFP ( Molecular Probes , 1∶10 , 000 ) , mouse α-HA ( Roche , 1∶500 ) , mouse 2A12 ( DSHB , 1∶10 ) , mouse α-Crb ( DSHB , 1∶10 ) , rabbit α-SAS ( D . Cavener , 1∶500 ) , mouse α-α-Spec ( DSHB , 1∶1 ) , rat α-DE-Cad ( DSHB , 1∶10 ) , mouse α-Fmi ( DSHB , 1∶10 ) , rat α-Dsh ( T . Uemura , 1∶1 , 000 ) , rabbit α-Stbm ( T . Wolff , 1∶200 ) , rabbit α-Pk ( J . Axelrod , 1∶2 , 000 ) , guinea pig α-Verm ( C . Samakovlis , 1∶500 ) , CBP-FITC ( New England BioLabs , 1∶500 ) , WGA-488 ( Molecular Probes , 1∶1000 ) , guinea pig α-Cor ( R . Fehon , 1∶2 , 000 ) , rabbit α-NrxIV ( H . Bellen , 1∶2 , 000 ) , mouse α-Fas3 ( DSHB , 1∶10 ) , rabbit α-Dys ( S . Crews , 1∶800 ) , rabbit α-Arl3 ( S . Hayashi , 1∶2 , 500 ) , mouse α-Elav ( DSHB , 1∶250 ) , and rat α-BarH1 ( H . McNeil , 1∶1000 ) . Fluorescence-labeled secondary antibodies were used at a 1∶500 dilution ( Molecular Probes ) . Embryo fixation and staining were performed as described [87] except for the α-E-Cad staining , for which embryos were fixed in 4% paraformaldehyde in PBS and devitellinized with ethanol . 3rd instar rp395 larval discs were dissected and fixed in 2% paraformaldehyde in PBS for 20 minutes , incubated with primary antibody overnight ( 4°C ) and then with the appropriate secondary antibody for two hours ( RT ) . In situ hybridizations were performed as described by [88] . The pPB3 cDNA , isolated by screening a cDNA library provided by L . Kauvar , was used to generate an anti-sense digoxigenin-labeled sano RNA probe . sano was identified in a P-element expression screen in Corey Goodman's laboratory . We obtained the rp395 line because of its salivary gland and tracheal expression . Sano was independently identified in an EP screen for genes that when misexpressed alter the eye phenotype of Dsh+Nkd overexpression ( S . Silva , G . Celik , C . -C . C . and K . A . W . , unpublished ) . Null sano mutants were generated by homologous recombination [65] . Genomic fragments upstream and downstream of the sano ORF were amplified by PCR and cloned into pW25 , which carries white+ , the recognition site for I-SceI endonuclease , and FRT sites . The construct was injected into embryos by Rainbow Transgenic Flies , Inc . Transformants were crossed to flies carrying hs-I-SceI and hs-Flp and progeny were heat shocked ( 37°C ) for 1 hour 48–72 hours AEL . The sano ORF was PCR-amplified and cloned into the pUAST [61] or pABAL expression vector [89] to create UAS-Sano and HS-Sano . UAS-Sano-GFP and UAS-Sano-HA were created using the Drosophila Gateway Vector system ( Carnegie Institution ) . A PCR fragment spanning the sano ORF was amplified and cloned into the pProEx expression vector ( Life Technologies , Inc . ) . The construct was transformed into BL21-DE3 cells , from which Sano inclusion body preparations were made . Recombinant full-length protein was further purified from an SDS-polyacrylamide gel slice as described [90] . Rat polyclonal antibodies were generated by Covance , Inc . and used at a dilution of 1∶50 . sano mutant or overexpressing clones were generated by the Flp-mediated recombination technique [91] , [92] . Clones were induced either 36–48 or 48–60 hours AEL by a one hour heat shock ( 37°C ) . The genotype for sano mutant clones was hs-FLP/+; ubi-GFP FRT42/sanoKO3 FRT42 . The genotype for Sano flip-out clones was either act>y+>Gal4/+ , UAS-Sano-GFP/hs-FLP or act>y+>Gal4/+ , arm-fz-GFP/+; UAS-Sano-HA/hs-FLP . Pupae were fixed 32–34 hours APF in 4% paraformaldehyde in PBS overnight ( 4°C ) . Pupal wings were dissected and washed several times in 0 . 5% PBST ( 0 . 5% Triton X-100 in PBS ) and incubated with phalloidin-568 ( Molecular Probes , 1∶1000 ) for one hour ( ice ) . For antibody staining , pupae were fixed at 28 hours APF in 4% paraformaldehyde in PBS for one hour ( 4°C ) . Pupal wings were dissected and washed in 0 . 1% PBST . Wings were incubated in primary antibodies overnight ( 4°C ) and then in secondary antibodies for two hours ( ice ) . sano mutant eye clones were generated using eyeless-FLP ( ey-FLP ) . sano mutant cells were distinguished by the absence of the GFP signal . The 3rd instar larvae were dissected in the PBS , fixed with fixation buffer ( 0 . 1M PIPES ( pH6 . 9 ) , 1mM EGTA ( pH6 . 9 ) , 1 . 0% Triton X-100 , 2mM MgSO4 , 1% formaldehyde ) , blocked in a solution ( 50mM Tris ( pH6 . 8 ) , 150mM NaCl , 0 . 1% Triton X-100 , 5mg/ml bovine serum albumin ( BSA ) ) . The discs were incubated in primary antibodies in a washing/incubation solution ( 50mM Tris ( pH6 . 8 ) , 150mM NaCl , 0 . 1% Triton X-100 , 1mg/ml BSA ) overnight at 4°C and then in secondary antibodies for two hours at RT . sevenless-Gal4 ( sev-Gal4 ) was used to overexpress Sano in the eye . sano mutant eye clones were generated using ey-FLP . sano mutant cells were w+/w+ , which can easily be distinguished from w+/w− heterozygous cells and from w−/w− twin spots in whole eyes . Since it is difficult to distinguish w+/w+ versus w+/w− in thin sections , however , we chose only eyes with large mutant clones and adjacent w−/w− twin spots for sectioning . Fixation and semi-thin sectioning of the adult eyes were slightly modified from [93] . Sections from at least five independent eyes were analyzed for each genotype . Embryos were stained with 2A12 and projections from lateral views of confocal sections of the DT lumen of st . 16 embryos ( at the four equal-compartment midgut stage ) were traced from the starting point of metamere one to the point where the last transverse connective ( TC ) meets the DT in metamere nine using the Image J program ( NIH ) . At least ten samples were measured and normalized to the length of the embryo for each genotype . An average length from three independent measurements of each sample was calculated . Pupal wings were stained for E-Cadherin and the perimeter of pupal wing cells overexpressing Sano-GFP and of their wild-type neighbors were measured by Image J . da-Gal4/UAS-Sano-HA; dsh-GFP/+ embryos were used for co-IP , and da-Gal4/UAS-Sano-HA and dsh-GFP embryos were used as controls . The embryos were collected and homogenized in radioimmunoprecipitation ( RIPA ) buffer ( Cell Signaling ) including protease inhibitor cocktail ( Roche ) . A small aliquot of the cleared supernatant was used for the Western to check the protein input with α-GFP and α-HA . Dynabeads Protein G ( Invitrogen ) was incubated with mouse α-HA ( Roche ) or rabbit α-GFP ( Molecular Probes ) for 10 minutes at RT . After several washes with PBTw ( 0 . 01% Tween-20 in 1× PBS ) , the remaining supernatant was incubated with antibody-bound Dynabeads Protein G for 20 minutes at RT . The beads were washed three times with RIPA buffer , and boiled in SDS sample buffer to elute the proteins . Bound antigen was detected by enhanced chemiluminescence ( GE Healthcare ) . The antibodies for Western blotting were used at the following concentrations: rat α-HA ( Roche , 1∶2 , 000 ) , mouse α-GFP ( Roche , 1∶2 , 000 ) . Co-Ips were repeated three times with the same results . The developmental Northern blot was prepared as described [94] and hybridized with a Bgl II/Not I fragment from pPB3 cDNA labeled by random priming . Fluorescence-labelled 10kDa dextran ( Molecular Probes ) injections were performed as described [22] , [95] , using wild-type embryos as a negative control and the mega mutant as a positive control . Embryos were dechorinated and fixed in the fixative for 20–30 min at RT . The fixative includes 800µl 5× buffer B ( 50mM KPO4 , pH6 . 8 , 225mM KCl , 75mM NaCl , 65mM MgCl2 ) , 800ml 37% formaldehyde , 2 . 5ml dH2O and 8ml heptane . Antibody staining was performed with α-Tgo antibody to mark the tracheal nuclei . After secondary antibody labelling , the embryos were treated with 10µg/ml proteinase K for 1 min and post-fixed with 3 . 7% formaldehyde in 0 . 1% Tween20 in 1× PBS ( PBT ) . ApopTag staining was performed using ApopTag Plus Peroxidase In Situ Apoptosis Kit ( Millipore , S7101 ) , and the cells undergoing apoptosis were labeled with rhodamine-conjugated α-Dig antibody ( 1∶10 , Roche ) .
|
Tubular organ formation is a ubiquitous process required to sustain life in multicellular organisms . In this study , we focused on the tracheal system of the fruit fly , Drosophila melanogaster , and identified Serrano ( Sano ) as a novel protein expressed in several embryonic tubular organs , including trachea . sano loss results in over-elongated trachea , whereas Sano overexpression causes shortened trachea , suggesting that sano is required for proper tracheal tube length . Interestingly , Sano overexpression results in typical planar cell polarity ( PCP ) defects in many adult tissues and pupal wing cells . The PCP pathway is highly conserved from flies to mammals and it has been known to control cell polarity within the plane of epithelial tissues . Importantly , we found that Sano binds Dishevelled ( Dsh ) , a key PCP regulator , and loss or ectopic expression of many known PCP proteins in the trachea give rise to similar defects observed with loss or gain of sano , suggesting a new role for the PCP genes in tube length control . Interestingly , the changes in tube length and PCP defects in the wing were linked to changes in apical domain size , suggesting that Sano and the PCP components affect either membrane recycling and/or the linkage of the membrane to the cytoskeleton .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology/organogenesis"
] |
2009
|
Serrano (Sano) Functions with the Planar Cell Polarity Genes to Control Tracheal Tube Length
|
Two interesting unanswered questions are the extent to which both the broad patterns and genetic details of adaptive divergence are repeatable across species , and the timescales over which parallel adaptation may be observed . Drosophila melanogaster is a key model system for population and evolutionary genomics . Findings from genetics and genomics suggest that recent adaptation to latitudinal environmental variation ( on the timescale of hundreds or thousands of years ) associated with Out-of-Africa colonization plays an important role in maintaining biological variation in the species . Additionally , studies of interspecific differences between D . melanogaster and its sister species D . simulans have revealed that a substantial proportion of proteins and amino acid residues exhibit adaptive divergence on a roughly few million years long timescale . Here we use population genomic approaches to attack the problem of parallelism between D . melanogaster and a highly diverged conger , D . hydei , on two timescales . D . hydei , a member of the repleta group of Drosophila , is similar to D . melanogaster , in that it too appears to be a recently cosmopolitan species and recent colonizer of high latitude environments . We observed parallelism both for genes exhibiting latitudinal allele frequency differentiation within species and for genes exhibiting recurrent adaptive protein divergence between species . Greater parallelism was observed for long-term adaptive protein evolution and this parallelism includes not only the specific genes/proteins that exhibit adaptive evolution , but extends even to the magnitudes of the selective effects on interspecific protein differences . Thus , despite the roughly 50 million years of time separating D . melanogaster and D . hydei , and despite their considerably divergent biology , they exhibit substantial parallelism , suggesting the existence of a fundamental predictability of adaptive evolution in the genus .
While parallel phenotypic evolution has long been recognized as one of the strongest pieces of evidence for adaptation [1] , the general repeatability of adaptive evolution in natural populations remains poorly understood . In large part this is because only recently has technology been available to facilitate the investigation of parallel evolution at various levels of biological organization , which have historically been hidden from view . For example , independently evolved , apparently similar phenotypes might in principle have completely different genetic explanations , suggesting a disconnection between genetic and phenotypic parallelism . Alternatively , the trajectory of adaptation may be severely constrained and highly repeatable across divergent taxa at the level of nucleotide or codon [2 , 3] . Parallel genetic evolution may occur at the level of nucleotide , gene , or pathway . For example , two lineages may have adapted to similar selection pressures through substitutions in largely non-overlapping genes which nevertheless belong to the same pathway . This would represent convergence at the level of pathway but not the level of gene . Parallel gene expression evolution may occur for transcript abundance , tissue expression , or alternative splicing . In addition to major gaps in our descriptions of the frequency with which parallelism occurs at different levels of biological organization ( from single nucleotides to complex phenotypes ) , we have little understanding of how lineage divergence in biological processes , ecology , or population genetics , may interact to influence the probability of parallelism at different levels of organization . For example , consider populations of two different species evolving in response to a shared , recently changed environment . Because the biology of the two species may differ substantially , the standing variation in the two species may interact with the environmental variation in different ways leading to directional selection on different phenotypes and genes , and this heterogeneity may vary dramatically based on the number of genes and their effect sizes contributing to variation in particular traits within species . Moreover , to the extent that adaptation to novel environments typically results from selection on standing variation , similarities and differences across species in the constellation of segregating variants due to variation in mutation , variation in patterns of mutation-selection balance , or to differences in the magnitude of drift may influence the degree of parallelism . Finally , we have little understanding of how the degree of parallelism scales with relatedness . While these problems are complex , we set out to begin attacking them in the Drosophila model , which has many benefits , including large numbers of species with diverse ecologies [4 , 5] that can be studied from comparative and population genetic perspectives . Multiple Drosophila species show phenotypic latitudinal clines [6–12] . As the central model species for Drosophila population genetics , D . melanogaster latitudinal variation has been subjected to considerable analysis , especially in North American and Australian populations ( reviewed in [13] ) . D . melanogaster evolved in Africa [14 , 15] . The species colonized Eurasia on the timescale of thousands of years and colonized the Americas and Australia on the timescale of hundreds of years [15–17] . D . melanogaster latitudinal clines , are robust , stable on decades long timescales ( e . g . , Voelker et al . 1978 [18] , Hoffmann and Weeks 2007 [11] , Eanes 2011 [19] ) , and often replicated on multiple continents [11 , 20 , 21] . More recently , population genomic analyses have been applied to gain a broader picture of the potential influence of spatially varying selection in the species [22–27] . The sibling species , D . simulans , which is broadly sympatric with D . melanogaster , has a roughly similar demographic history in that the species evolved in East Africa or Madagascar , and subsequently spread throughout Eurasia , the Americas and Australia [15 , 16 , 28 , 29] . This parallel history has naturally led to the question of whether recent colonization of similar , novel habitats in the two species has been accompanied by similar patterns of latitudinal differentiation . While relatively few studies exist on D . simulans latitudinal differentiation , the available data suggest that D . simulans shows weaker latitudinal differentiation at both the phenotypic and genomic levels [7 , 12 , 30–33] . This difference between the species has been speculated as being due to a more recent colonization history for D . simulans [16 , 29] ( so less time for selective differentiation to occur ) or due to differences in the ecology and demographics of the two species [34–37] . However , a recent paper on latitudinal gene expression differentiation in both species provided strong evidence for parallel latitudinal adaptation [27] . While additional work will be needed to understand the degree of similarities and differences in latitudinal adaptation in this pair of sister species , here we branch out to highly diverged lineage to continue studying parallel adaptation in Drosophila . Drosophila hydei is a member of the repleta group of Drosophila [38] , which shared a common ancestor with the melanogaster group roughly 50 million years ago ( 40–62 mya , [5 , 39] ) . The repleta group is roughly 20–30 million years old , likely originated in South America [40] , and generally exhibits a cactophilic ecology [4 , 40] . Compared to D . melanogaster , D . hydei produces relatively few , very large sperm and exhibits very high re-mating rates [41 , 42] . D . hydei is currently cosmopolitan in distribution . Indeed , D . hydei often appears in massive numbers in the same locations on rotting fruit where Drosophilists typically collect D . melanogaster and D . simulans ( e . g . , Patterson and Wagner 1943 [43] ) and is capable of exploiting a wide variety of resources [44] . Thus , while the species retains the ability to exploit cactus as a resource in its ancestral range [40] , it is clearly a generalist throughout most , if not all of its current distribution . While the temporal details of the geographic spread of D . hydei to achieve its current cosmopolitan distribution are currently unclear , Sturtevant in his species description [45] proposed that it first appeared in North America in the late 19th century . Thus , the temporal spread of D . hydei across North America may be roughly coincident with that of D . melanogaster [20] , suggesting that high temperate regions in North America have been colonized only recently [43] , similar to the situation with D . melanogaster . Given its history , we were interested in understanding whether patterns of latitudinal differentiation in D . hydei are similar to those in D . melanogaster . To address this question we produced a reference genome sequence and transcriptome and characterized patterns of sequence variation in high and low latitude populations of D . hydei . We then compared the properties of D . hydei genetic variation to the properties of genetic variation from D . melanogaster populations sampled from the same or similar locations . In addition to our interest in parallel latitudinal differentiation in these highly diverged species , we sought to address the question of parallel adaptation at longer timescales that encompass species divergence ( Fig 1 ) . A striking conclusion of recent Drosophila population genetic work is that a substantial proportion of protein divergence is the result of directional selection [46–49] . This finding , which is based on comparisons of synonymous and non-synonymous variation within and between species [46] , has come primarily from investigation of the melanogaster subgroup [48 , 49] , though it appears that similar conclusions are likely to hold for the obscura group as well [50] . Therefore , in addition to our investigation of geographic differentiation , we used our population genomics data to ask whether there is significant parallelism for the proteins evolving under recurrent directional selection in two highly diverged clades , the melanogaster subgroup and the repleta group .
We sequenced the D . hydei white female genome to a high coverage ( >170 fold , S1 Table ) . The genome size estimate based on k-mer frequencies from the short insert library was about 156 Mb ( million bp ) , which is consistent with , though slightly smaller than the species female genome size , 164 Mb , estimated by flow cytometry of ovary nuclei [51] . We used ALLPATHS-LG for the initial assembly and gap filling using Illumina short and long insert reads , and then further filled gaps by SSPACE using corrected PacBio data . After removing bacterial contamination , the final assembled genome was about 139 Mb , with a scaffold N50 of 754 kb and scaffold N90 of 163 kb . The GC ratio of the genome assembly was 39 . 63% . Analysis of gene content using BUSCO revealed that 95% percent ( S2 Table ) of the focal genes are included in the assembly , suggesting that the genome is sufficiently well assembled for most population and evolutionary analysis . In total we annotated 14 , 150 genes ( including ab initio genes ) and 12 , 380 protein-coding genes ( excluding ab initio genes ) . BUSCO and CEGMA analysis ( S2 Table ) showed that vast majority of conserved genes were well annotated . The assembled D . hydei genome repetitive sequence composition is comparable to that observed in other Drosophila species genome assemblies; 13 . 31% of the assembled genome is repetitive , including 1 . 85% retro-elements and 0 . 5% DNA transposons . Similar to other Drosophila species , LTR ( long terminal repeats ) have the highest abundance , followed by LINEs ( long interspersed nuclear elements ) [52] . In addition , Gypsy/DIRS1 has relatively high abundance in D . hydei , accounting for 0 . 79% of the genome . D . hydei retains the ancestral Drosophila karyotype , which is composed of five major acrocentric chromosome arms ( A-E ) plus a dot chromosome ( Muller F , chromosome 6 for D . hydei , and chromosome 4 for D . melanogaster ) [53] . Muller elements A-E correspond to chromosome arms X , 3 , 5 , 4 , 2 , for D . hydei and chromosome arms X , 2L , 2R , 3L , and 3R for D . melanogaster [54] . Using D . mojavensis and D . melanogaster synteny we assigned D . hydei scaffolds to Muller elements A-F ( See Methods , S3 Table ) based mostly on scaffold gene content . The scaffolds assigned to Muller elements encompass 98% of the assembly ( Table 1 ) . We used the scaffold assignments to assign genes to Muller elements/chromosomes ( S4 Table ) . As expected , Muller element assignment results are very similar using D . melanogaster and D . mojavensis , since Muller element gene content is generally highly conserved in Drosophila [55 , 56] . However , we observed small differences for Muller element F ( the dot chromosome ) assignment because a number of D . mojavensis dot chromosome sequences are assembled onto the Muller element E scaffolds ( chromosome 3R for D . melanogaster , and chromosome 2 for D . hydei ) [56 , 57] . Because of this , we used the alignment results with D . melanogaster for downstream analysis of the dot chromosome . A total of 9561 and 2301 genes were assigned to autosomes and X-chromosome , respectively . The GC content for the autosomes was 38 . 8% and the X-chromosome content was 40 . 2% , consistent with previous reports from D . melanogaster that GC content is greater for the X [58] . We first blasted annotated genes against the 20 Drosophila species genome annotations [52 , 59] . Of 12 , 380 genes included in the analysis , 11 , 483 had one reciprocal best hit in one of the genomes , which supports previous inferences that current Drosophila gene content generally reflects gene content of the Drosophila ancestral species [52 , 59] . We then defined orthologous genes of D . hydei , D . mojavensis and D . melanogaster by using synteny and sequence similarity ( reciprocal best hit ) . This yielded 10 , 000 putative orthologous genes between D . hydei and D . mojavensis , and 9401 such genes between D . hydei and D . melanogaster . All downstream orthologous gene related analysis and comparisons are focused on these gene sets . In addition to defining homologous genes and orthologous genes , we also studied gene family number gain and loss using OrthoMCL . Gene copy number appears to be relatively highly conserved with D . mojavensis; 9298 genes ( 9105 families ) share the same gene copy number as the D . mojavensis annotation . Genes showing large copy number increases in D . hydei relative to D . mojavensis tend to be retro-transposon proteins , such as Tc1-like gene and gag proteins . In total , we found 109 protein-coding genes for which copy number was greater in D . hydei relative to D . mojavensis , 41 of which have a homolog in D . melanogaster ( S5 Table ) . Interestingly , we found duplications of Ir54a , Ir56c , and Ir68b , which are ion-channel genes that are expressed in sensory cilia and may function in detection of chemical stimulus [60] . CG17387 ( testis specific expression , cilium movement ) and SPR ( sex peptide receptor ) exhibit species-specific duplications in D . hydei relative to other sequenced Drosophila species [52 , 59] . In addition , we found D . hydei duplications of Apc/Apc2 , fry , faf , ERR , ihog , Nox , Vps15 , and Didum . The overall level of nucleotide heterozygosity in D . hydei based on 1-kb window means was 0 . 0019 ( Table 2 ) , which is roughly half the nucleotide heterozygosity of North American D . melanogaster populations [49] , and even more severely reduced compared to African D . melanogaster populations [61 , 62] . There has been some speculation that in Drosophila , genome-wide levels of nucleotide heterozygosity may be determined primarily by the effects of selection on linked sites [48 , 49 , 63 , 64] . This conjecture would predict that all else being equal , species with higher recombination rates would have higher levels of average heterozygosity . D . hydei euchromatic recombination rates per physical distance are thought to be substantially greater than those of D . melanogaster [65 , 66] , a conclusion supported by our unpublished estimates of cM/Mb inferred by placing mutants of known genetic location [66] on the assembly . Nevertheless , D . hydei exhibits substantially lower mean heterozygosity than D . melanogaster . This difference could result from differences in demographic history or in the intensity of directional selection ( though our analysis of adaptive protein divergence below is consistent with roughly equal amounts of protein adaptation in the two species ) . In any case , the D . hydei heterozygosity estimates cast some doubt on the proposition that variation in mean heterozygosity across Drosophila species will be explained primarily as a consequence of interspecific differences in recombination rates and the interaction of recombination rate variation with selection . We used previously published estimates of D . melanogaster synonymous heterozygosity for 1-to-1 orthologs [49] and compared them to estimates of synonymous heterozygosity for D . hydei . The non-synonymous polymorphism and synonymous polymorphism were 0 . 0010 , and 0 . 0098 respectively , which is smaller than D . melanogaster homologous genes , 0 . 0012 , and 0 . 0152 respectively ( non-parametric t test , both p <2 . 2e-16 ) . Thus , the roughly 10-fold greater level of synonymous compared to non-synonymous variation in D . hydei is similar to that observed in other Drosophila species [48 , 49 , 67] . If levels of synonymous heterozygosity are determined primarily by selection at linked sites , the extensive chromosome rearrangements that have fixed since the D . melanogaster-D . hydei ancestor [68 , 69] implies that heterogeneous relative recombination rate variation experienced at the scale of genes ( or larger ) is probably poorly correlated between these species . We observed a very weak but highly significant correlation in synonymous heterozygosity between species ( Pearson’s r = 0 . 14 , p<2 . 2e-16 ) , consistent with some degree of conservation for genic parameters of mutation rates and/or selection at Drosophila synonymous sites [58 , 70–74] . Levels of variation on the X chromosome were nearly identical to those observed for the autosomes ( Table 2 ) , while the simple neutral equilibrium expectation under equal effective population sizes of males and females is that the X will exhibit three-fourths the heterozygosity of the autosomes [75] . Similar observations supporting roughly equal levels of nucleotide heterozygosity on the X vs . autosomes have also been made in African population samples of D . melanogaster and D . simulans [49 , 76–78] . In contrast , X-to-autosome heterozygosity ratios are substantially less than one in non-African populations of D . melanogaster ( ranging from 0 . 63 to 0 . 68 [79] and 0 . 64 to 0 . 69 [62] ) and D . simulans [48 , 80] . We observed a subtle but consistent pattern across windows and Muller elements that a greater proportion of sites were polymorphic in Panama than in Maine ( S6 Table ) . This result , which is robust to variation in quality and coverage , is consistent with the notion that Panama populations are closer to the ancestral geographic distribution of the species and that the recent expansion of D . hydei to high latitude North American populations [40] has been accompanied by a loss of low frequency variants . However , there is no evidence for a significant bottleneck or serial founder effects , as nucleotide diversity estimated for 1-kb non-overlapping windows was nearly identical in the two populations . Indeed , though a smaller proportion of sites are observed as polymorphic in the Maine sample ( S6 Table ) , Maine generally exhibits slightly greater nucleotide heterozygosity ( π ) compared to Panama ( Table 2 ) , presumably as a result of more intermediate frequency variants ( S1 Fig ) . Considering all 1-kb windows in the genome , 42 . 7% had greater π in Panama , 48 . 2% had greater π in Maine , and 9 . 1% had the same estimated π ( including windows with no segregating sites in either population ) in both populations . The X chromosome deviates from this general pattern , exhibiting slightly lower diversity in Maine than in Panama . In the Maine sample , the X-to-autosome ratio is 0 . 94 while in Panama sample the ratio is about 1 . 09 . We investigated the regions showing the greatest difference in π between the two populations ( 1-kb π difference > 0 . 002; 39 genes overlapped these windows ( S7 Table ) . These genes include nAChRalpha7 , mxc , fz4 , and X11Lbeta . Of the 39 genes , 32 ( 82% ) were X-linked . This enrichment of X-linked genes is not due to a large X-chromosome region of geographic differentiation , as these X-linked genes are not significantly closer to each other than expected . Using FST estimated in 1-kb non-overlapping windows we identified the windows in the 1% , 2 . 5% and 5% tails of the distribution . The top 1% , 2 . 5% , and 5% windows had mean 1-kb FST of 0 . 217 , 0 . 171 , and 0 . 139 , respectively ( Fig 2 , Table 3 ) . These estimates are slightly greater than those observed for outlier 1-kb windows from D . melanogaster sampled from the same locations ( Table 3 , 1-kb FST non parametric test , Wilcoxon test p < 2 . 2e-16 ) [81] . Similarly , median and mean 1-kb FST for the D . hydei genome ( Table 3 ) were 0 . 050 , and 0 . 061 , respectively , which are slightly greater than those of D . melanogaster populations sampled from the same locations [81] . Similar to observations from US populations of D . simulans [32 , 33] , levels of geographic differentiation were substantially higher on the X-chromosome ( 1-kb window mean FST = 0 . 077 ) than on the autosomes ( 1-kb window mean FST = 0 . 055 , Mann-Whitney U , p <2 . 2e-16 , S8 Table ) , and the pattern remains after coverage correction . Mean FST was homogeneous across autosomes ( S8 Table ) . We also characterized FST at the level of individual SNPs . As expected , based on the 1-kb window FST analysis , SNP FST was significantly elevated on the X chromosome ( p < 2 . 2e-16 , S9 Table ) . To determine whether genic DNA is over- or under-represented among the most differentiated genomic regions we determined the number of genes overlapping windows in the top 1% of the1-kb FST distribution . In total , 201 genes overlapped with these window FST outliers . Among them , 123 genes were located on X-chromosome . A comparable analysis but with the top 5% 1-kb FST outliers resulted in 953 genes , 427 of which were X-linked . For neither cutoff , however , is the number of genes spanned by outlier windows greater than expected based on the proportion of analyzed windows containing genic sequence ( p > 0 . 1 ) . However , because our minimum site and window coverage criteria were quite stringent , there were many more genic than non-genic windows in the analysis , potentially compromising our power to detect genic vs . non-genic enrichments . In addition , we asked if 1-kb windows that overlap genes show greater mean differentiation . We found the mean difference is small ( genic window FST = 0 . 059 vs . non-genic window FST = 0 . 063 ) . These results are consistent with previous analyses of latitudinal differentiation in Drosophila [23 , 25] suggesting no major differences in levels of differentiation for genic vs . non-genic regions . To identify SNPs and genes that may be more likely to experience spatially varying selection we focused on non-synonymous variants . We used a modified Fisher’s exact test to identify separately for each chromosome the outlier non-synonymous SNPs . We found 1070 protein coding genes ( S10 Table ) having at least one non-synonymous FST outlier SNP ( FDR 1e-5 and FST > 0 . 15 ) , 308 of which are on X chromosome ( p < 2 . 2e-16 ) . Genes carrying outlier nsSNPs are enriched in Gene Ontology categories such as receptor activity ( Benjamini corrected p = 3 . 18e-04 ) and molecular transducer activity ( Benjamini corrected p = 1 . 59e-04 ) ( S10 Table ) . The genes having at least one non-synonymous FST outlier SNP in D . melanogaster were enriched in taste receptor activity ( p = 2 . 90e-04 ) and related biological processes ( S10 Table ) , suggesting that receptor-related genes may experience spatially varying selection in both species . A number of genes harboring outlier nsSNPs were associated with functions such as regulation of transcription , chromatin modification , cell motion , ovarian follicle cell development , and several additional biological processes . We found that 259 of the 1070 genes overlapped with the top 5% 1-kb FST outliers , suggesting that , in agreement with other studies of Drosophila [23 , 25 , 33] , strongly differentiated SNPs tend to be associated with somewhat larger regions of latitudinal differentiation . Of the 259 genes , 122 are X-linked , further supporting greater geographic differentiation of the X chromosome . To identify genes carrying highly differentiated nsSNPs in both D . hydei and D . melanogaster we focused on the FST outliers identified from the set of “genic” SNPs ( Methods ) in both species using Fisher’s exact test with midp test correction and estimation of False Discovery Rate ( FDR ) , as described in Svetec et al . 2016 [81] . For both species each chromosome arm was analyzed independently . Among the 9401 one-to-one orthologous genes , 640 D . melanogaster genes [81] and 1031 D . hydei genes harbor at least one non-synonymous SNP FST outlier ( FDR 1e-5 and FST > 0 . 15 ) . Remarkably , we found 110 genes shared between D . hydei and D . melanogaster , which represents a 1 . 57-fold enrichment ( hypergeometric test , p = 6 . 22e-07 ) compared to the null hypothesis of independence ( Table 4 , S11 Table ) . While these shared genes show no major GO enrichment ( unsurprisingly given the relatively small number of genes ) , a number of shared genes were involved in functions such as sensory perception of smell ( scrib , Pino , Or2a , and Ir84a ) , detection of chemical stimulus ( Ir94a , Ir100a , Ir56a , Ir40a , Ir84a , and Ir94c ) , and sensory perception of taste ( Gr58b , Gr22e , and Ir100a ) . Several transcription factors , including pb , dpy , ush , Brf , and Elp2 also contain non-synonymous SNP FST outliers in both species . Notably , although the D . hydei X chromosome is enriched for genes carrying nsSNP outliers , there is no enrichment of shared outlier genes on the X chromosome ( hypergeometric test , p >0 . 05 ) , probably because there is no over-representation of X-linked genes carrying nsSNP outliers in D . melanogaster . Indeed , shared outliers genes are more likely to be autosomal than X-linked ( hypergeometric test , p = 1 . 76e-08 ) . Using the top 400 genes carrying the most differentiated nsSNPs ( ranked by FDR ) in each species also leads to an observed excess of shared outlier genes ( 27 shared genes , hypergeometric test , p = 0 . 01 ) ( Table 4 , S11 Table ) . While a component of the observed excess of shared genes could be attributable to variation in gene-size or SNP density ( larger genes are more likely to harbor more SNPs and thus share outliers just by chance ) , accounting for this source of variation ( Methods ) also revealed that the observed number of shared genes harboring outlier nsSNPs was significantly greater than expected ( 78 expected , 110 observed , 1 . 41-fold enrichment , p = 0 . 002 ) . We further investigated outlier gene sharing by considering only the 5004 genes that carry at least one nsSNP in both species; this constitutes the set of genes for which we could have , in principle , observed shared outlier genes given the constraints of our data . Of these genes , 513 and 892 carried an outlier nsSNP in D . melanogaster and D . hydei , respectively . As was the case in the aforementioned analysis on all genes , the same 110 genes were shared , and the probability that the sharing is due to chance is similar ( Table 4 , hypergeometric test , p = 0 . 015 ) . This excess of gene sharing was preserved after accounting for variation in SNP numbers across genes ( p = 0 . 014 ) . These results suggest that despite their long divergence time , distinct biology , and disparate biogeography , that there is a moderate predictability to the patterns of latitudinal differentiation in these two species , at least for genes/protein polymorphisms . We further investigated the evidence for parallel responses to spatially varying selection by determining whether outlier nsSNPs in the two species exhibit evidence of more systematic same-direction allele frequency differences ( e . g . , in both species the derived allele occurs at higher frequency in Maine ) relative to non-outlier nsSNPs . For the outliers nsSNPs located in the 110 shared outlier genes D . melanogaster showed a marginally significant bias ( 94 nsSNPs with higher frequency in Maine and 71 with higher frequency in Panama , χ2 test , p = 0 . 05 ) . The comparable D . hydei analysis revealed a similar trend that was not significant ( 34 nsSNPs with higher frequency in Maine and 26 with higher frequency in Panama , p = 0 . 22 ) , though it should be noted that the number of SNPs is small . The probability of the observed trends in both species was relatively small ( Fisher’s combined probability = 0 . 06 ) . We then considered all nsSNPs , not just those in the shared genes , and compared directionality for the top 1000 nsSNPs in each species to the remaining nsSNPs . In this analysis D . hydei exhibited 537 snSNPs with higher frequency in Maine and 463 nsSNPs with higher frequency in Panama ( p = 0 . 014 ) , consistent with the trend observed in the outlier genes . The comparable analysis for D . melanogaster revealed 606 nsSNPs with higher frequency in Maine and 394 nsSNPs with higher frequency in Panama ( p < 0 . 001 ) , supporting previous conclusions regarding recent selection in high latitude populations [24 , 79] . Overall then , there is some support for parallel directionally differentiated nsSNPs , but the effect is not large . However , it is worth noting that the power of these approaches to detect recent , parallel , population-specific allele frequency changes at a set of SNPs enriched for true targets of selection may be compromised by allele frequency changes of nearby neutral SNPs , as well as allele frequency distributions in the ancestral populations and the specific demographic histories of the populations . To investigate whether D . hydei , D . melanogaster , and D . simulans exhibit parallel patterns of latitudinal gene expression differentiation , we performed RNA-seq analysis of D . hydei Panama and Maine male flies raised at 21°C , and compared those data to comparable existing data from Maine and Panama populations of D . melanogaster and D . simulans reared at 21°C [27] . While the lack of D . hydei biological replicates precluded most formal statistical approaches , given the high coverage of the D . hydei RNA-seq data and the existing high quality D . melanogaster and D . simulans data , we thought the empirical pattern of expression fold-changes between populations was appropriate for generating broad , conservative inferences about gene expression parallelism . At total of 8760 orthologous genes were expressed in both D . hydei and D . melanogaster . We compared the top 300 most differentially expressed orthologous genes for both species and found 25 shared differentially expressed genes , which represents a highly significant excess of shared genes ( hypergeometric test , p < 4 . 4e-04 ) . Different cutoffs ( such as the top 500 rather than top 300 genes ) returned comparable results . The shared differentially expressed genes include trp , inaF-D , ImpL2 , Eip71CD , Cyp6d5 , Cyp12d1-p , Cpr92A , and Cpr30F . These 25 shared genes show no evidence of shared directionality ( for example , a gene showed higher or lower expression level in the Panama sample for both the species ) ( hypergeometric test , p > 0 . 1 ) . However , the small number of shared genes provides little power to detect such effects . To seek further evidence bearing on the question of shared expression directionality we compared for all genes expressed in both species , the observed fold changes between species for genes showing same direction differences vs . opposite direction differences . We observed that fold changes were slightly more correlated between species for same direction than for opposite direction genes ( Pearson’s r abs = 0 . 60 vs . abs = 0 . 56 , two-tailed test for Fisher’s z-transformations p = 0 . 03 , permutation test correcting gene number , p <0 . 001 ) . This provides weak , though significant further support for expression parallelism . A total of 5848 orthologous genes were expressed in D . hydei and D . simulans . Comparing the top 300 most differentially expressed genes in both species ( ranked by fold-change ) , revealed 25 shared genes ( hypergeometric test , p = 0 . 01 ) . One gene , Eip71CD , was differentially expressed in all three species . While these data support the idea that parallel latitudinal expression differentiation observed between two closely related species , D . melanogaster and D . simulans [27] extends to the very distantly related species , D . hydei , the limitations of our existing expression data leave open the question of the full extent of this form of parallelism . We used the McDonald–Kreitman test ( MK test ) to investigate patterns of adaptive protein divergence between D . hydei and its close relative , D . mojavensis . Of the 9315 one-to-one orthologs for which there was sufficient data to carry out an MK test , 807 ( 8 . 7% ) rejected neutrality at p < 0 . 05 ( S12 Table ) , while 316 genes ( 3 . 4% ) had p-values less than 0 . 01 . Of the 807 significant genes , 682 genes deviated from neutrality in the direction of adaptive protein divergence ( DoS , Direction of Selection ) and had estimated alpha ( α = proportion of amino acid fixations explained by directional selection ) greater than 0 . These results suggest that a substantial proportion of proteins have experienced recurrent directional selection in this clade [47] . In D . melanogaster , of 9328 genes tested , 1265 ( 13 . 56% ) had a significant MK test , 593 of which ( 6 . 35% ) had p < 0 . 01 ( S13 Table ) . Of these 1265 significant D . melanogaster genes , 638 ( S13 Table ) had DoS ( Direction of Selection ) and proportion of amino acid variants fixed by selection ( α ) greater than 0 , suggesting that a comparable number of genes ( 682 for D . hydei vs . 638 for D . melanogaster ) have experienced recurrent adaptive protein evolution in each clade . Thus , the whole genome evidence of pervasive , recurrent adaptive protein divergence in Drosophila now includes both the melanogaster subgroup [48 , 49] and the repleta group ( this report ) . This conclusion is likely to hold for the obscura group as well [50] . A significantly greater proportion of genes that reject the null hypothesis do so in the direction of adaptive divergence for D . hydei than for D . melanogaster ( χ2 test , p < 0 . 01 ) . The fact that D . melanogaster is substantially more polymorphic than D . hydei but exhibits a greater proportion of genes rejecting the null with α < 0 ( and comparable to polarized MK tests for D . melanogaster; Langley et al . 2012 [49] ) suggests that a simple explanation of population size variation interacting with slightly deleterious amino acid polymorphisms will not suffice . Note , however , that both D . hydei ( unpolarized MK vs . D . mojavensis ) and D . simulans ( polarized MK [48 , 49] ) exhibit a smaller proportion of genes with α < 0 than D . melanogaster , and both appear to have higher recombination rates compared to D . melanogaster [48] . This supports the idea that Hill-Robertson effects associated with recombination rate variation may contribute to the efficacy of selection on new amino acid polymorphisms [82] . However , any model of selection on protein variation must accommodate both estimates of adaptive and deleterious amino acid variation and its interaction with mean recombination rate differences between species and variance in recombination rates within species . For both the D . hydei-D . mojavensis and D . melanogaster-D . simulans clades , the genes showing evidence of recurrent adaptation were enriched on the X chromosome ( D . melanogaster-D . simulans X vs . autosome is 160 genes vs . 478 genes , D . hydei-D . mojavensis X vs . autosome is 133 genes vs . 549 genes ) , supporting faster-X adaptation ( χ2 test , p < 0 . 001 for both ) [48 , 83] . This is likely a conservative conclusion given that male-biased or male-specific genes , which appear to be more likely then most other classes of genes to experience recurrent protein adaptation ( below ) , are underrepresented on the X [47 , 84–86] . There was no evidence that genes having estimated α < 0 ( often interpreted as evidence of deleterious segregating protein variants ) are more likely than expected to be shared between clades ( hypergeometric test p = 0 . 18 ) . This is consistent with the idea that divergence in the local recombination rate between these highly diverged species due to extensive karyotype evolution and/or genome-wide differences in recombination rates alters the locus-specific efficacy of selection against deleterious amino acid variants . Alternatively , the distribution of selection coefficients for new amino acid variants may evolve at the gene level . For the 682 genes having significant MK test with evidence of directional selection in the D . hydei-D . mojavensis species pair , 296 showed male-biased or male–specific expression in our reference sequence whole male/whole female transcriptome data . Specifically , 194 genes showed male-specific expression and 102 genes showed male-biased expression . 119 genes showed female-biased gene expression , while no gene showed female-specific expression . Similar to D . melanogaster , male-biased and male-specific genes were significantly enriched among the genes with evidence of recurrent adaptive protein divergence ( ( χ2 test , p < 0 . 0001 ) , but female-biased and female-specific genes were not enriched ( χ2 test , p > 0 . 1 ) . These results support the idea that male reproduction is a “hotspot” of recurrent protein adaptation . There was no formal GO enrichment for the significant MK genes in D . hydei-D . mojavensis . For the 638 genes significant D . melanogaster genes , 249 showed male-biased or male–specific expression in our reference sequence whole male/whole female transcriptome data . Specifically , 141 genes showed male-specific expression and 108 genes showed male-biased expression . One significant gene showed female-specific expression , while 120 genes showed female-biased gene expression . Male-biased and male-specific genes were significantly enriched among the genes with evidence of recurrent adaptive divergence ( ( χ2 test , p <0 . 0001 ) , but female-biased and female-specific genes were not enriched ( ( χ2 test , p > 0 . 1 ) . GO analysis suggests that genes are enriched in ATP-binding ( Benjamini corrected p = 0 . 02 ) and ubiquitin-protein transferase activity ( Benjamini corrected p = 0 . 02 ) . We also found several GO terms including male gamete generation , spermatogenesis , dosage compensation , and regulation of RNA metabolic process that were enriched more than 2-fold but were not significant after multiple testing correction . In general , however , both clades show a strong enrichment of male-related functions for genes exhibiting recurrent adaptive protein divergence . For D . hydei we found several dynein proteins among the genes with strong evidence for directional selection ( MK test , p < 0 . 0003 , FDR < 0 . 05 ) . Seven of the top 50 most significant genes ( Dhc98D , Dhc16F , Dic61B , Dhc36c , nod , Vha100-3 and Dnah3 ) are involved in microtubule-based movement , motor activity and/or ATPase activity , among which , Dhc98D , Dhc16F , Dic61B , Dhc36c , and Dnah3 are components of the axonemal dynein complex . Six of the seven genes show male-biased gene expression or have mutant male fertility phenotypes in D . melanogaster ( FlyBase ) , suggesting that they may directly function in sperm development and motility . For example , Dic61B codes for an axonemal dynein intermediate chain exhibiting strong testis-biased expression ( FlyBase ) ; it is required for development and precise assembly of sperm axonemes and is essential for male fertility in D . melanogaster [87 , 88] . Given the rapid evolution of sperm length in D . hydei , along with its close relatives in the hydei group , D . bifurca and D . eohydei [89] , it is tempting to speculate that adaptive evolution of male-specific axonemal dyneins associated with sperm gigantism is related to this phenotype . Also notable among the significant MK genes are five ( aly , comr , tomb , can , and sa ) that are homologous to testis meiotic arrest genes in D . melanogaster ( reviewed in White Cooper and Davidson 2011 [90] ) . These genes are required for regulation of transcripts produced in the primary spermatocyte and whose products function during meiosis and spermatid development . It remains to be seen how the adaptive evolution of these proteins may functionally influence interspecific divergence of gene expression in the primary spermatocytes and how such expression evolution maps onto variation in sperm developmental processes or sperm morphology . GO analysis of the significant MK genes suggests significant enrichment for detection of chemical stimulus ( Benjamini corrected p = 2 . 67e-04 ) and genes involved in sensory perception of smell ( Benjamini corrected p = 1 . 44e-02 ) . The conclusion that recurrent adaptive protein divergence is common in two highly diverged Drosophila clades raises the interesting question of whether the specific proteins exhibiting evidence of recurrent selection in the two clades overlap to a greater degree than expected . For one-to-one orthologous genes , 6578 had sufficient data to perform MK tests in both species pairs . Of these , 467 ( 7 . 09% ) and 373 ( 5 . 67% ) genes showed evidence of recurrent adaptive protein divergence in D . melanogaster/D . simulans and D . hydei/D . mojavensis , respectively ( S13 and S14 Tables ) . The two species-pairs share evidence of recurrent adaptive protein divergence in 66 genes , which is highly significant ( 2 . 50 fold enrichment , hypergeometric test , p = 1 . 11e-12 , S15 Table ) . This pattern of excess sharing of significant MK genes holds even when we use a stricter MK test cutoff of p < 0 . 001 . The extensive parallelism supports the idea that there are strong tendencies in Drosophila for certain proteins to be frequent targets of recurrent directional selection . The 66 shared genes are dispersed in multiple functional pathways and show no obvious enrichment for particular biological process . However , of the 66 genes , 10 showed male-biased expression while 27 showed male-specific gene expression in D . hydei; 11 showed male-biased expression while 27 showed male-specific gene expression in D . melanogaster , supporting the idea that recurrent protein adaption for genes functioning in male reproduction will be a general pattern across the genus Drosophila . We used the D . melanogaster annotation in FlyBase to inspect the biology of these shared genes in slightly greater detail . First , it is worth noting that 35 of the 66 genes have CG numbers but no gene names , which reveals that fundamental biological attributes of many proteins experiencing chronic directional selection in Drosophila remain very poorly understood . Of these 35 genes , a large proportion ( 23 genes in both species and an additional one in D . hydei ) show testis-biased expression and for 12 there is experimental support from proteomics data that the gene product is a component of D . melanogaster sperm [91] . Turning to the named genes , three ( aly , comr , and can ) function as regulators of transcription during early spermatogenesis prior to the onset of meiosis . Also showing adaptive protein divergence in both clades is sneaky , a sperm acrosome protein required for breakdown of the sperm plasma membrane inside the oocyte [92] , and Dhc98D , a strongly male-biased axonemal dynein . Also notable is the shared significant gene qin , which plays a role in transposon silencing in the female germline [93 , 94] . The shared gene mof , which plays a role in male dosage compensation , supports previous work suggesting that some components of dosage compensation in Drosophila are likely to experience frequent directional selection [95] , though the possibility that other phenotypes are targets of selection is entirely plausible [96 , 97] . We examined the estimated α for the 66 shared significant MK genes and found , remarkably , that α was highly correlated across clades ( Spearman’s ρ = 0 . 50 , p = 1 . 6e-5 ) . This additional form of parallelism implies that beyond the sharing of proteins experiencing recurrent adaptation , for shared proteins the relative contribution of recurrent adaptation to protein divergence tends to be similar across highly diverged clades .
D . hydei and D . melanogaster shared a common ancestor several tens of million years ago [5 , 39] and have highly diverged ecologies , mating systems , and ancestral geographic ranges . While the recent spread of D . hydei to a cosmopolitan distribution is not as well understood as that of D . melanogaster , the colonization of high temperate regions in North America by D . hydei is likely to be recent , similar to the history inferred for D . melanogaster . Thus , the population genomic analysis of geographic differentiation and of recurrent directional selection on protein sequences in these two species provides some insight into the general repeatability of adaptive evolution on multiple timescales in the Drosophila model . We found , perhaps surprisingly , that parallel latitudinal differentiation at the population genomic level is sufficiently common to be detectable even in our relatively small datasets encompassing only two population samples for each species . Prior to the application of population genomic approaches , D . melanogaster latitudinal clines had been observed for many phenotypes and genetic variants , which suggested that highly differentiated genomic regions between lower and higher latitude population would be enriched for variants exhibiting clines [22 , 23] , a proposition supported by recent comparison of data from North American cline “endpoints” [81] with data from latitudinal sampling [26] . However , because the existence of latitudinal clines in D . hydei has not been systematically investigated , we are less confident that strongly differentiated genetic variants , in general , are highly enriched for targets of spatially varying selection in this species . Thus , we are limited in our confidence to speculate on the differences between these two species in latitudinal differentiation . Nevertheless , the genes showing high levels of latitudinal differentiation in both species provide a glimpse into the prevalence of parallelism and its underlying biological basis . Several of these genes function in detection of chemical stimulus or in taste . The appearance of strongly differentiated DNA repair genes in both species could be related to UV adaptation [81] . One of the major patterns emerging from our population genomic analysis of geographic differentiation is the large X-effect . The D . hydei large X-effect is not the result of a small portion of the chromosome showing extreme differentiation , but rather is a general chromosome wide effect . A similar pattern was observed in US D . simulans [32 , 33] , though not in D . melanogaster [25] . It remains to be seen through additional comparative work whether D . melanogaster is highly unusual in this regard and if so , whether selection on autosomal inversions in this species swamps any underlying signal of X chromosome dynamics broadly shared across species . A possible demographic explanation for greater X-linked differentiation is male-biased dispersal . Because male migrants carry only one X chromosome while females carry two , increased male relative to female migration results in a proportional decrease in the number of X chromosomes ( relative to autosomes ) moving from one population to another , which should increase X chromosome differentiation [98] . This hypothesis is amenable to both laboratory and field experiments [99–101] . Alternatively , recent models suggest that under a wide range of circumstances the X chromosome should show a disproportionate contribution to local adaptation [102] . One might suppose that a chromosome-wide effect should favor the demographic rather than the selective hypothesis . However , the inference from sequence divergence that much of the Drosophila genome , including non-coding sequence , is functionally important [103 , 104] suggests that the selective hypothesis should at least be seriously entertained . Further work will be required to clarify this issue . While the significant limitations of our population transcriptome data from D . hydei ( relative to our D . melanogaster and D . simulans data [27] ) weaken our power to detect parallel gene expression differentiation in these species , our results suggest that parallel expression differentiation play a general role in latitudinal adaptation in Drosophila [27] . Further quantification of latitudinal gene expression variation in better data from these three species would facilitate the analysis of parallel expression differentiation and permit a more quantitative test of the idea that parallel expression differentiation is significantly more common for closely related species than for more distantly related species , a trend that is consistent with our limited data . Our analysis of parallel ( at the level of the gene ) recurrent adaptive protein evolution in two distantly related clades revealed a number of salient results . First , both clades exhibit evidence of rampant adaptive evolution , supporting previous conclusions regarding the prevalence of adaptive protein divergence in Drosophila [46–49] . Second , our results suggest that the details of adaptive protein divergence are remarkably similar in these distantly related clades . The two species pairs share many more adaptively evolving proteins than expected under the simple null model . Indeed , it is tempting to speculate that our analysis of shared repleta group and melanogaster subgroup adaptively evolving proteins has identified a collection of proteins with relatively high probability of evolving adaptively in many Drosophila lineages . This conjecture is certainly testable . Third , for the proteins showing evidence of recurrent adaptation in both clades , the proportion of divergence explained by selection is highly correlated . Thus , it appears that there is a surprising level of parallelism in the degree to which protein divergence is determined by directional selection across broad phylogenetic distances in Drosophila . The biological patterns of genes with a history of recurrent protein adaption suggest that despite their highly diverged mating systems and reproductive biology , both clades have experienced recurrent protein adaptation at many orthologous genes that are testis-biased , testis-specific , or that are associated with spermatid development and differentiation . Understanding the ultimate cause of this rampant mode of Drosophila adaptation remains a substantial challenge . Finally , results from both clades support the notion that adaptive divergence is more common on the X chromosome . It is worth noting that the approach used here may substantially underestimate the prevalence of adaptive protein divergence , as MK tests are expected to be underpowered to detect adaptation in small proteins or adaptive protein divergence that occurs in relatively few residues of individual proteins . Whether this bias colors our conclusions about the prevalence of parallel protein adaptation in Drosophila remains unclear . One of the patterns observed here is that there appears to be greater parallelism for long-term adaptive protein divergence , often related to testis expression , than for shorter timescale latitudinal differentiation . This difference could have multiple explanations . First , because these two species are quite diverged they may interact with the environment or with environmental variation in different ways . A corollary of this hypothesis is that the more highly repeatable longer-term parallelism we observed is more likely to involve proteins and pathways experiencing selective processes that tend to be less linked to environmental variation . Male-male interactions , male-female interactions , or genomic conflicts ( such as those related to gametic selection or transposable elements ) are obvious candidates . Second , to the extent that evolution on short timescales in novel environments may often depend mostly on standing variation , the genetic details of the selection response may differ simply because the constellation of variation available to selection may only be weakly correlated in highly diverged species . Alternatively , if much of the selection response on short timescales depends on alleles ancestrally at mutation-selection equilibrium , then the predictability of differentiation may be reduced by stochastic effects that may dominate even strongly selected low frequency variants , or by evolutionary divergence of the genic parameters of mutation-selection balance . Finally , it is worth pointing out that we identified three genes , qin , Cht6 , and Msp-300 , that carry nsSNP latitudinal differentiation outliers in D . melanogaster and D . hydei and also show evidence of recurrent adaptive protein divergence between species in the two clades examined here . It remains to be seen whether such potential “hotspots” of adaptation result from agents of selection that tend to be shared on long and short timescales across highly diverged species , or instead , represents a chance occurrence .
We calculated k-mer frequencies ranging between 13-31mers using Jellyfish [105] , and then estimated the genome size using k-mer frequency and coverage [106 , 107] . In short , the formula is G = Kmer_num/Kmer_depth , where Kmer_num is the total number of k-mers of all the reads and the Kmer_depth is the average depth of k-mers . We performed quality control to Illumina short reads , with only high quality reads ( Q>30 for each base ) being kept for further analysis . PacBio clean reads were first generated from SMRT cell raw data and then further corrected by PacBioToCA [108] . We assembled the reference genome using high quality 190bp insert library reads and 2kb insert library reads by ALLPATHS-LG ( release#51298 ) [109] with standard parameters . We then used corrected PacBio reads to fill scaffold gaps by SSPACE-LongRead [110] . To remove possible microbial contamination we used tblastn to filter contaminated reads . Specifically , all the annotated proteins ( see below ) were used to blast Drosophila species ( Drosophila 12 species ( Clark et al . 2007 ) and 8 new modENCODE species [59] and Ensembl bacteria species by tblastn ( -e 1e-5 ) . If more than 1/3 of the total genes on a scaffold had a best-hit map to a bacterium the scaffold was discarded as contamination . Scaffolds that had no annotated genes were used to blast Drosophila species and bacteria species by blastn; if such a scaffold had no significant hit to a Drosophila species ( -e 1e-5 ) but had a hit ( -e 1e-10 ) to a bacterium , then the scaffold was considered a contaminant . We used only high quality reads ( Q >30 , length threshold >30 ) for transcriptome assembly . Before assembly , we normalized transcripts using normalize_by_kmer_coverage . pl provided by Trinity program ( version 2 . 0 . 6 ) using parameter—JM 40G —max_cov 40—pairs_together—PARALLEL_STATS JELLY_CPU 8 . Male and female white D . hydei RNA-seq reads , as well as reads pooled for the two sexes , were assembled using Trinity ( version 2 . 0 . 6 ) , using parameter—max_memory 40G —min_contig_length 200—CPU 10—inchworm_cpu 10—bflyCPU 10 . Alignment of reads . To assess assembly quality , high quality Illumina reads from the 190bp paired-end library were aligned to the assembly using BWA ( 0 . 7 . 13 , parameter bwa aln -n 0 . 01 -l 35 -o 1 -d 12 -e 12 -t 8 ) . 94 . 91% reads could be aligned to the assembled genome , which shows that most reads were incorporated into the assembly . The depth curve plotted based on the alignments showed a unimodal distribution ( S2 Fig ) , suggesting the reads were randomly distributed on the genome and which also suggests that the sequenced strain has very low heterozygosity . Core list of genes . We used two methods to estimate the proportion of highly conserved genes present in the assembly . First , we used BUSCO ( Benchmarking Universal Single-Copy Orthologs ) [111] to estimate the proportion of the 2765 arthropod orthologous genes that were completely or partially assembled . We also used CEGMA [112] to blast to the genome and identify CEGs ( Core Eukaryotic Genes ) in the assembly . The MAKER2 genome annotation pipeline was used for gene annotation ( maker version 2 . 31 . 8 , snap version 2013-11-29 , hmmer version 3 . 1b2 , TRF version 4 . 0 . 9-static , and RepeatMasker version 4 . 0 . 5 ) . To improve annotation accuracy we fed the de novo assembled transcriptomes , the best translated protein sequences generated by Trinity , and 20 Drosophila species protein sequences to help MAKER2 predict gene models , which were then used to train the HMM for D . hydei . After two rounds of HMM training , MAKER2 was used to predict gene models with ab-initio gene prediction algorithms SNAP and Augustus [113] . We generated two annotations , one of which allows ab-initio prediction . We used both annotations to estimate the genome quality , but only used the annotation without ab-initio prediction for downstream analysis . We aligned annotated D . hydei genes to the D . mojavensis and D . melanogaster genomes using tblastn ( -e 1e-10 ) . We assigned a D . hydei scaffold to a Muller element ( A through F ) if 55% of annotated genes on a scaffold had the best alignment to one , homologous Muller element based on the blast results to D . melanogaster and D . mojavensis [56] . For genes without gene annotation , we blasted sequences to D . mojavensis genome and used the criteria of minimum 50% alignment length with 30% sequence similarity to determine the Muller element . Using these methods , we assigned 136 of 139 Mb genome sequences to Muller elements A-F . We used TRF ( Tandem Repeats Finder , 4 . 0 . 9-static ) with default parameters to identify non-interspersed repetitive elements . Transposable elements ( TEs ) were first predicted by homology searches to RepBase TE libraries ( version 21 . 05 ) using RepeatProteinMask and RepeatMasker ( version 4 . 0 . 5 ) with default parameters . We then constructed a de novo repeat library using RepeatScout with default parameters and obtained consensus sequences and classification information for each repeat family . Using these RepeatScout consensus sequences as the input library we again searched repetitive elements in the assemblies using RepeatMasker with default parameters . After that , we merged the results from the above pipelines to generate the final classification . Reads from the Panama City and Maine pools were aligned to the D . hydei genome using Bowtie2 with the—very-sensitive setting . Variants were called using bcftools ( samtools . github . io/bcftools ) and PoPoolation2 [114] with a minimal quality score of 30 . Following Svetec et al . [81] , we required a minimum of 20× coverage at a site in both the Maine and Panama populations and at least two observations of an alternate base call in the entire dataset ( two populations ) to consider it in the population genetic analysis . We excluded triallelic sites . We calculated expected nucleotide diversity , π , following Kolaczkowski et al . [23] and FST following Svetec et al . [81] . For FST we performed the odds ratio test for independence using the ormidp . test function in the epitools package in R ( medipei . com/epitools/ ) and then used the p-values from midp tests to calculate the false discovery rate for each chromosome arm using the bioconductor package q-value ( http://github . com/jdstorey/qvalue ) . For scaffolds at least 1-kb long we calculated 1-kb non-overlapping FST windows for each chromosome for windows meeting the minimum 20× coverage per site . Windows at the end of a scaffold that were less than 1-kb long were discarded . In total , 99 . 22% of the assembly was analyzed using 1-kb windows . In addition , for most 1-kb window-based analyses we required that at least 50% of the sites in a window meet our minimum coverage criterion for a window to be included . For gene-based analyses we included SNPs in the gene region and 1-kb upstream and downstream of the transcript . Within these spans we categorized SNPs as synonymous , non-synonymous SNPs intronic , 3’UTR , 5’UTR , or flanking . To determine whether the number of shared genes with FST non-synonymous outliers in D . hydei and D . melanogaster was greater than expected , we performed 1000 independent bootstraps to obtain an empirical distribution of shared outlier genes considering the number of SNPs in each gene following Zhao et al . [27] , to account for the influence of gene size and SNP number on probability of outlier overlap . To do so , we estimated the number of outlier nsSNP numbers for each of the orthologous genes in D . hydei and D . melanogaster , and then randomly picked genes having equal or higher number of nsSNP outliers than the observed genes . We then calculated the number of shared 1-to-1 orthologous genes in D . hydei and D . melanogaster . The analysis was repeated 1000 times to generate an empirical distribution of p-values for shared genes harboring nsSNP outliers . We used D . mojavensis reference to infer the ancestor state of SNPs , and only consider biallelic SNPs , one of which is the same as D . mojavensis ancestor SNP , for downstream analysis . GO enrichment of each gene list was performed using DAVID v6 . 8 [115] or Gorilla ( http://cbl-gorilla . cs . technion . ac . il ) . To determine whether the number of shared genes with FST outliers in D . hydei and D . melanogaster is influenced by gene size and number of SNPs within genes , we carried out 1000 independent bootstraps to obtain an empirical distribution of shared outlier genes considering the number of SNPs in each gene . We first counted the numbers of outlier nsSNPs in the outlier genes used for comparisons . For example , one set of outlier genes of D . melanogaster included 369 genes having one nsSNP outlier , 84 genes having two SNP outliers , 27 genes having 3 outliers , etc . We also calculated nsSNP outlier numbers for each gene in the D . hydei list . We then randomly picked genes that had equal or greater numbers of nsSNPs than the observed nsSNPs in the outlier gene lists in each species , and then calculated the number of shared orthologous genes between D . melanogaster and D . hydei . After repeating 1000 times , we obtained the empirical distribution and P-values . Transcriptome sequencing for the samples described above was performed with Illumina RNA sequencing protocols . De novo and reference-guided assemblies of high quality clean reads were also performed using Trinity for downstream analysis . The reads were also mapped to the genome using tophat ( version 2 . 0 . 13 ) . FPKM and differential expression was calculated using Cufflinks and Cuffdiff2 , as well as DEseq2 following Zhao et al . 2015 [27] . After generating gene expression and differential expression estimates , we ranked the gene expression fold differences and identified the top 300 differentially expressed orthologous genes . To determine whether there was enrichment for shared latitudinal expression differentiation in D . melanogaster vs . D . hydei , as well as D . simulans vs . D . hydei , we compared the top 300 most differentially expressed genes in each species and applied the hypergeometric test for independence . We used a χ2 test to determine whether genes differentially expressed in both species tend to show greater transcript abundance in either the higher or lower latitude population . Protein coding genes from Drosophila 12 species [52] were downloaded from FlyBase . We used the longest protein sequence of each gene to perform an “all vs . all” alignment using BLASTP ( blast+ version 2 . 2 . 30+ ) with e-value cutoff 1e-5 . We then use OrthoMCL [116] to cluster genes from different species into gene orthologous groups , following manual check using the blast results . We used reciprocal best hit and synteny relationship ( between D . mojavensis and D . hydei ) to define one-to-one orthologous genes [27] . The reciprocal best hits between D . hydei and D . melanogaster as well as D . hydei and D . simulans were also used for investigating gene expression differentiation . High quality paired-end reads from the Panama City and Maine libraries were aligned to the genome . We called all bi-allelic SNPs that satisfied the following criteria: 1 ) minimum mapping quality ( Q-score ) of 30 [49] , 2 ) minimum coverage of 20 and 3 ) minor allele called at least 3 times to reduce the possibility that low-frequency slightly deleterious amino acid polymorphisms result in overly conservative conclusions regarding the prevalence of adaptive protein divergence [117 , 118] . We then used the SNP data to generate alternate reference genomes using an in-house Perl script . Specifically , using each bi-allelic SNP that passed the filtering criteria mentioned above we generated two genomes ( a . k . a . alternative references ) , with each one containing a set of SNPs . We then re-extracted the coding sequence of each gene from alternate references and performed multiple alignments using Genewise to remove insertions and deletions , then re-aligned using PRANK with –codon function for each D . hydei and D . mojavensis orthologous gene . To improve statistical power and make our analysis comparable to that from Langley et al . [49] , we only carried out MK tests for genes that showed at least one variant in each of four categories , polymorphic , fixed , synonymous , and nonsynonymous . For genes that passed the above criteria we carried out unpolarized McDonald–Kreitman tests using the MK . pl [48 , 49] , using Fisher’s exact test . Significant genes ( p < 0 . 05 ) were compared to significant genes from comparable unpolarized MK tests for D . melanogaster ( using D . simulans as outgroup ) . The D . melanogaster data included the Raleigh and Malawi samples reported in Langley et al . ( 2012 ) [49] . For each gene , we estimated the proportion of adaptive amino acid fixations ( α ) according to Smith and Eyre-Walker [47] , and the Direction of Selection ( DoS ) index according to Stoletzki and Eyre-Walker [119] .
|
Both local adaptation on short timescales and the long-term accumulation of adaptive differences between species have recently been investigated using comparative genomic and population genomic approaches in several species . However , the repeatability of adaptive evolution at the genetic level is poorly understood . Here we attack this problem by comparing patterns of long and short-term adaptation in Drosophila melanogaster to patterns of adaptation on two timescales in a highly diverged congener , Drosophila hydei . We found , despite the fact that these species diverged from a common ancestor roughly 50 million years ago , the population genomics of latitudinal allele frequency differentiation shows that there is a substantial shared set of genes likely playing a role in the short term adaptive divergence of populations in both species . Analyses of longer-term adaptive protein divergence for the D . hydei-D . mojavensis and D . melanogaster-D . simulans clades reveal a striking level of parallel adaptation . This parallelism includes not only the specific genes/proteins that exhibit adaptive evolution , but extends even to the magnitudes of the selective effects on interspecific protein differences .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"computational",
"biology",
"geographical",
"locations",
"panama",
"invertebrate",
"genomics",
"animals",
"animal",
"models",
"north",
"america",
"drosophila",
"melanogaster",
"model",
"organisms",
"experimental",
"organism",
"systems",
"genome",
"analysis",
"evolutionary",
"adaptation",
"drosophila",
"research",
"and",
"analysis",
"methods",
"central",
"america",
"genomic",
"libraries",
"sex",
"chromosomes",
"chromosome",
"biology",
"gene",
"expression",
"x",
"chromosomes",
"animal",
"genomics",
"insects",
"arthropoda",
"people",
"and",
"places",
"eukaryota",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"evolutionary",
"processes",
"organisms",
"chromosomes"
] |
2017
|
Genomics of parallel adaptation at two timescales in Drosophila
|
The bacterial cell cycle has been extensively studied under standard growth conditions . How it is modulated in response to environmental changes remains poorly understood . Here , we demonstrate that the freshwater bacterium Caulobacter crescentus blocks cell division and grows to filamentous cells in response to stress conditions affecting the cell membrane . Our data suggest that stress switches the membrane-bound cell cycle kinase CckA to its phosphatase mode , leading to the rapid dephosphorylation , inactivation and proteolysis of the master cell cycle regulator CtrA . The clearance of CtrA results in downregulation of division and morphogenesis genes and consequently a cell division block . Upon shift to non-stress conditions , cells quickly restart cell division and return to normal cell size . Our data indicate that the temporary inhibition of cell division through the regulated inactivation of CtrA constitutes a growth advantage under stress . Taken together , our work reveals a new mechanism that allows bacteria to alter their mode of proliferation in response to environmental cues by controlling the activity of a master cell cycle transcription factor . Furthermore , our results highlight the role of a bifunctional kinase in this process that integrates the cell cycle with environmental information .
The bacterial cell cycle has been studied extensively in the past . Genetics , biochemistry and more recently , advanced microscopy techniques have provided important insight into the processes of DNA replication , chromosome segregation and cell division , and numerous regulatory mechanisms have been identified that precisely coordinate these processes in time and space . Most of this research has focused on cell cycle regulation under standard and stable laboratory growth conditions . However , in nature bacteria are exposed to drastic environmental changes , where they have to constantly adjust their growth rate and mode of proliferation [1 , 2] . It has frequently been reported that various bacteria transform into multi-chromosome containing filaments in response to certain environmental conditions [2–4] , indicating that bacteria dynamically modulate cell division and the cell cycle in response to environmental cues . Nevertheless , the precise mechanisms transducing environmental information into the cell division machinery and how these mechanisms help cells to survive under adverse conditions are not well understood . Cell cycle regulation has been studied in several model bacteria . One prominent example is the asymmetrically dividing alphaproteobacterium Caulobacter crescentus , a freshwater bacterium that mainly occurs in oligotrophic aquatic environments , but also in organically rich environments such as wastewater [5] . The Caulobacter cell cycle is characterized by asymmetric cell division and well-defined , morphologically distinct cell cycle phases , offering the possibility to examine cell cycle progression with high spatial and temporal resolution . Past work has identified a suite of key regulatory proteins required for cell cycle progression and important progress has been made in understanding how these factors are wired in higher-ordered circuits to drive cell cycle progression under optimal conditions [6 , 7] . However , how the Caulobacter cell cycle is modulated in response to environmental changes is only at the beginning of being explored . One major cell cycle regulator is the conserved response regulator CtrA , which regulates the transcription of nearly 100 genes involved in cell division , cell cycle regulation and morphogenesis [8 , 9] . By binding to the origin of DNA replication CtrA also serves as a negative regulator of DNA replication initiation [10] . CtrA activity is strictly regulated and oscillates in a cell cycle-dependent manner [11] . In G1-phase CtrA is active and represses the origin [10] . At the G1-to-S transition it is inactivated and rapidly proteolysed allowing DNA replication to initiate [12 , 13] . During S-phase , active CtrA accumulates again to induce the expression of cell division and morphogenesis genes that are required to complete the cell cycle by cell division [9] . The oscillations of CtrA depend on its precise regulation by the CckA-ChpT phosphorelay , which is comprised of the polarly localized bifunctional histidine kinase CckA and the phosphotransferase ChpT [14 , 15] . In response to spatiotemporal cues , CckA phosphorylates CtrA via ChpT , resulting in CtrA activation . Reversal of the phosphorelay leads to CtrA dephosphorylation and hence its inactivation [15 , 16] . CtrA inactivation is tightly coupled to its degradation by the ClpXP protease that depends on CpdR , an adaptor protein , whose activity also depends on CckA-ChpT [17] . In contrast to CtrA , CpdR is only active when dephosphorylated [17 , 18] . Thus , under conditions when CtrA is inactivated , it is also targeted for degradation by ClpXP and CpdR . In addition to CpdR , the cell-cycle dependent degradation of CtrA involves the adaptors RcdA and PopA that promote CtrA proteolysis by ClpXP in a second messenger and phosphorylation dependent manner [19–21] . Recent work has provided important insight into the mechanisms regulating the CckA kinase in a cell cycle-dependent manner . It was shown that protein interactions with the upstream regulators DivL and DivK at the cell poles allow CckA to switch between kinase and phosphatase activity [22 , 23] . In addition , CckA activity is also modulated by the second messenger cyclic diguanylate ( c-di-GMP ) [24] . This molecule , which accumulates at the G1-S transition [25] , directly binds CckA and stimulates its phosphatase activity , and thus CtrA inactivation and proteolysis [24] . As a transmembrane protein CckA may also directly respond to external signals , as typical for most other histidine kinases that have important sensing functions and directly transduce environmental information into cellular responses . However , external conditions that influence CckA activity have not been identified until now . The present study started with a systematic analysis of the effects of different stress conditions on C . crescentus cell cycle progression . This analysis revealed that under certain conditions Caulobacter specifically blocks cell division and grows to filamentous cells . Interestingly , the stress-induced cell filamentation that we observed is not mediated by induction of small cell division inhibitors , as previously described for conditions inducing DNA damage [1 , 26 , 27] . Instead , we found that stress leads to rapid downregulation of CtrA . Our data suggest that the stress-induced regulation of CtrA stems from stimulation of CckA phosphatase activity and that it provides a growth advantage under stress .
To analyze the consequences of different stress conditions on cell cycle progression in C . crescentus , we challenged bacterial cultures with carbon starvation , stationary phase , heat shock , DNA damage , oxidative stress , salt stress , sucrose stress , ethanol stress and changes in pH while monitoring changes in cell morphology and chromosome content using phase-contrast microscopy and flow cytometry , respectively . As previously shown , carbon starvation , growth in stationary phase and acute heat shock lead to a block of DNA replication initiation and a G1-arrest while cell size is largely maintained ( Fig 1A ) [28–30] . Interestingly , several other stress conditions caused C . crescentus cells to respond in a clearly different manner . Most conspicuously , upon treatment with 100 mM NaCl or 4% ethanol ( EtOH ) cells transformed into long filaments and accumulated multiple chromosomes ( Fig 1A ) . Consistent with the flow cytometry data , we found by using a fluorescent repressor-operator system ( FROS ) , which fluorescently marks the origins of replication [31] , that the elongated cells contained multiple well-segregated origins ( S1 Fig ) , demonstrating that cells continue to undergo DNA replication , chromosome segregation and cellular growth under these conditions , but that they stop dividing . Cells treated with NaCl or ethanol increased to up to eight to ten-fold of their original length and often accumulated six to seven chromosomes within eight hours , suggesting that growth and the cell cycle continued for three to four doubling times in the absence of division . The effects on cell division were observed within two hours of NaCl or EtOH treatment , however the phenotype became more pronounced over time ( Fig 1B ) . We also observed that the filamentous phenotype only occurred in a narrow range of NaCl or EtOH concentrations ( Fig 1B ) , in which cells are still viable and able to grow ( S2 Fig ) . Similar to the phenotypic response observed upon NaCl or EtOH treatment , we observed that incubation at 40°C also caused cell elongation . At this temperature C . crescentus is still able to grow and to replicate its DNA , whereas temperatures above 40°C lead to a growth and DNA replication arrest ( Fig 1 , S2 Fig ) . The chromosome content was not as strongly increased at 40°C as under the EtOH and NaCl stress conditions . We did not observe significant changes in cell morphology or chromosome content when cultivating C . crescentus at different acidic or alkaline pHs ( Fig 1A ) . Treatment of cells with sublethal doses of H2O2 led to an increased proportion of cells ( 54 . 1% ) with a DNA content that equals a single chromosome compared to the non-stress condition ( 38 . 2% ) ( Fig 1A ) , suggesting that under oxidative stress many cells arrest in G1-phase , similar to starvation conditions , stationary phase and acute heat shock . To assess whether cell filamentation in response to salt , ethanol or mild heat shock stress is reversible , we followed the fate of stress-treated filamentous cells by time-lapse microscopy upon shifting them to non-stress conditions . Filamentous cells from each of the three conditions were able to resume growth and to initiate cell division at multiple sites shortly after release to fresh growth medium with first cell division events occurring within two hours ( Fig 2 ) . Within six hours most bacteria were able to propagate and divide normally yielding typically sized Caulobacter daughter cells . Notably , a fraction of daughter cells that arose from the filamentous cells maintained elongated until the end of our time-lapse study , indicating that these cells were severely damaged . Altogether our data demonstrate that C . crescentus changes its morphology and cell cycle status in response to changing external conditions . Different stress conditions affect the cell cycle at different stages , either by transiently blocking DNA replication initiation and cellular growth or by delaying cell division . The filamentous phenotype of cells treated with salt , ethanol and mild heat shock is similar to the phenotype of cells treated with DNA damaging agents such as mitomycin C ( Fig 1A ) [26] . Previously , it has been shown that in C . crescentus and other bacteria DNA damage induces the expression of small division inhibitors , which directly interfere with components of the cell division apparatus and thereby block the process of cell division in response to DNA damage [1 , 26 , 27] . To test whether the transient block of cell division upon treatment with NaCl , EtOH or increased temperature is due to induction of the SOS response we monitored the promoter activity of the sidA gene , which encodes the SOS induced division inhibitor in C . crescentus [26] . In contrast to mitomycin C treatment , which caused strong induction of PsidA-gfp within two hours , the reporter was not turned on upon exposure to NaCl or EtOH stress ( Fig 3A and 3B , S3 Fig ) . A mild induction in a subpopulation of cells was observed at 40°C , which was however clearly lower compared to mitomycin treatment . These data show that the filamentous phenotype induced by salt stress , ethanol stress or mild heat shock does not appear to depend on the SOS response and SidA . Another previous study reported that a subpopulation of C . crescentus cells form helical filaments during long-term growth in stationary phase [32] . It was shown that in these filaments the level of the major cell division protein FtsZ was strongly reduced [32] . Therefore we wanted to test if FtsZ abundance and localization was affected under NaCl stress , EtOH stress and mild heat shock . However , Western Blot analysis showed that FtsZ protein abundance was not significantly altered within eight hours of stress treatment ( Fig 3D ) . Furthermore , using a strain expressing ftsZ-eYFP showed that FtsZ still localized in distinct foci along the length of the cells ( Fig 3C ) , suggesting that the observed block of cell division is not caused by a failure of FtsZ to localize to division sites . To better understand the molecular basis underlying the observed stress-induced morphological changes , we analyzed global changes in gene expression by RNA-sequencing ( RNA-seq ) . We focused on EtOH and NaCl stress as the effect on cell division was most pronounced under these conditions . Treatment with 100 mM NaCl for 60 minutes resulted in a >2-fold induction of 472 genes ( 11 . 6% ) and a >2-fold downregulation of 282 genes ( 6 . 9% ) . In the case of EtOH stress , 570 genes were >2-fold upregulated ( 14 . 0% ) and 489 genes ( 12 . 0% ) were >2-fold downregulated after 60 minutes . Statistical analysis revealed that in each case , the gene expression profile after 60 minutes was highly similar to the profile obtained after 30 minutes of stress treatment with significant scores ( z-scores ) greater than 40 ( Fig 4A ) . We also observed a strong overlap when comparing the gene expression data sets from the NaCl and EtOH conditions to each other for both upregulated and downregulated genes ( z-scores >20 ) , suggesting that salt and ethanol stress result in similar changes in gene expression . Consistent with our data obtained with the PsidA-gfp reporter , the SOS regulon was not strongly induced under ethanol and salt stress ( Fig 4A ) . Interestingly , among the most downregulated genes were many involved in flagella biosynthesis , pili biogenesis , chemotaxis , and cell cycle progression ( Fig 4B ) . In C . crescentus , these genes are under the direct control of the master cell cycle regulator CtrA [8] . Consistently , we found that gene expression changes in a divLts mutant , which upon shift to 37°C results in loss of CtrA function [22 , 33] , were similar to the gene expression changes induced by salt and ethanol stress in wild type cells ( Fig 4A and 4B ) . Genes that were downregulated in divLts cells , for example flagella , pili and chemotaxis genes , cell cycle regulators ( ccrM , sciP , divK ) or cell division genes ( ftsA , ftsQ , ftsW ) also showed significant downregulation in response to NaCl or EtOH treatment . By contrast , CtrA repressed genes , which were upregulated in the divLts mutant also showed upregulation in the NaCl treated and EtOH treated cultures . These data demonstrate that the regulon of CtrA is strongly influenced in response to salt and ethanol stress , suggesting that CtrA activity is affected under these conditions . Loss of CtrA function results in strong chromosome accumulation and cell filamentation [13 , 22] , phenotypes similar to those observed during salt and ethanol stress ( Fig 4C ) . The observed effect on the expression of CtrA-regulated genes strongly points to altered CtrA function in response to certain stress conditions . Consistently , when we monitored steady-state protein levels of CtrA by Western blotting , we found that the level of CtrA strongly and rapidly decreased upon exposure to salt , ethanol stress or mild heat shock ( Fig 5A ) . Most conspicuously , treatment with 4% EtOH resulted in complete elimination of CtrA within only 15 minutes . Exposure to NaCl or 40°C for 60 minutes led to a drop in CtrA levels to 15% and 10% , respectively . By contrast , in response to mitomycin-induced DNA damage CtrA levels were not significantly affected . In a strain depleted of the protease subunit ClpX and in a ΔcpdR mutant CtrA was stabilized ( Fig 5B ) , indicating that the ClpXP protease along with the CpdR adaptor , which are responsible for the cell cycle-dependent degradation of CtrA under optimal growth conditions [12 , 17] , are also required to rapidly downregulate CtrA in response to environmental stress . To test if the stress-induced reduction of CtrA steady-state levels was caused by increased proteolysis , we measured CtrA degradation rates by in vivo stability assays under the different stress conditions . The degradation rate of CtrA depended strongly on the external condition under which cells were cultured ( Fig 5C ) . Under optimal growth conditions CtrA had a half-life of approximately 26 minutes , similar to previously published results [22] . Exposure to increased external salt concentrations , or incubation at 40°C led to a strong decrease in half-life to 12 . 8 minutes or five minutes , respectively . Most remarkably , upon incubation with 4% EtOH the half-life was as short as two minutes ( Fig 5C , S4 Fig ) . The decrease in CtrA stability under these conditions contrasts the changes in CtrA stability induced by carbon starvation [29]; under this condition the half-life was prolonged to 173 minutes ( Fig 5C ) . Taken together , our data demonstrate that salt , ethanol and mild heat shock induce rapid proteolysis of CtrA . More generally , our data indicate that the rate of CtrA proteolysis is subject to modulation by environmental signals allowing for the integration of external information into the cell cycle . How do environmental stress conditions affect the rate of CtrA proteolysis ? Because CtrA stability is tightly linked to its phosphorylation [15 , 17] , we thought that the observed stress-induced decrease in CtrA stability might result from its dephosphorylation and inactivation involving the CckA-ChpT phosphorelay . To test this possibility , we measured CtrA activity following stress treatment in the ΔcpdR strain , in which CtrA is stabilized but growth rate unchanged under non-stress conditions ( Fig 5B , S5 Fig ) . As a read-out for CtrA's transcription factor activity we measured the expression of CtrA-dependent genes by quantitative RT-PCR . Our data showed that as in wild type cells the mRNA abundance of the genes sciP , fljO , ccrM , ftsQ and pilA rapidly dropped in ΔcpdR cells following EtOH or NaCl addition ( Fig 5D ) , suggesting that despite being stabilized , CtrA is not able to activate the expression of these genes after stress exposure . In agreement with our qRT-PCR data , we found that the phenotype of the ΔcpdR strain was indistinguishable from the filamentous phenotype of wild type cells upon treatment with EtOH or NaCl ( Fig 5E , Fig 1A ) . These data let suggest that the increased proteolysis of CtrA under stress conditions follows its inactivation as a transcription factor . Under non-stress conditions CckA dynamically switches between its kinase and phosphatase activities [16 , 34] . We reasoned that the observed inactivation of CtrA could either be due to complete loss of CckA function or to a more active mechanism , in which CckA switches from a kinase into a phosphatase . To distinguish between these possibilities we investigated if CckA phosphatase activity is required to induce rapid downregulation of CtrA upon EtOH treatment , the stress condition that impacted CtrA stability most strongly . To this end , we used a point mutant of CckA , CckA ( V366P ) , that retains kinase activity , but lacks significant phosphatase activity [16] . Indeed , constitutive expression of this mutant in a cckA deletion background partially prevented CtrA downregulation upon EtOH addition , whereas expression of wild type CckA from the same construct led to rapid removal of CtrA similar to the wild type ( Fig 6A ) . Monitoring CtrA degradation rates in cckA ( V366P ) -expressing cells showed that the half-life of CtrA during EtOH exposure was markedly increased in the mutant ( 7 . 7 min ) compared to the wild type ( 2 . 2 min ) ( Fig 6B ) . Similar results were also obtained under NaCl stress ( S6 Fig ) . Together , these data indicate that CckA phosphatase activity is critical to ensure rapid degradation of CtrA in response to external stress . We next tested if increasing the kinase activity of CckA results in a similar effect by using a CckA variant containing the substitution G319E , which has been shown to render CckA hyperactive as a kinase [16] . Because constitutive expression of cckA ( G319E ) leads to severe cell cycle and growth defects , we expressed this variant from an inducible promoter in a wild type background . Western blots showed that CtrA levels were completely stabilized in this mutant following EtOH treatment ( Fig 6A ) . Moreover , the half-life of CtrA in cells expressing this mutant was strongly increased in the presence of stress ( 22 . 3 min ) compared to the half-life in cells expressing wild type CckA ( 4 . 5 min ) ( Fig 6B , S6 Fig ) . Consistent with the strong stabilization of CtrA in cckA ( G319E ) -expressing cells we observed that cells neither formed filaments nor accumulated chromosomes when treated with salt , EtOH or mild heat shock , but instead arrested with a single chromosome in G1-phase ( Fig 6C , S7 Fig ) . Expression of a stable and active ctrA allele , ctrA ( D51E ) Δ3Ω [13] led to a similar phenotype: cells no longer accumulated multiple chromosomes in response to EtOH and NaCl stress and instead arrested in G1-phase ( Fig 6D , S8 Fig ) . These data suggest that the replication phenotype of cckA ( G319E ) cells is due to increased CtrA phosphorylation and activity . We observed that stress-induced cell filamentation and chromosome accumulation was also less severe in cckA ( V366P ) cells than in the wild type ( Fig 6C , S7 Fig ) . However , compared to the cckA ( G319E ) and the ctrA ( D51E ) Δ3Ω mutants , the change of phenotype was not as strong , likely because CtrA was only partially stabilized in the cckA ( V366P ) mutant ( Fig 6A ) . Together our data are consistent with a model , in which environmental stress causes CckA to switch to phosphatase activity , leading to CtrA dephosphorylation and rapid proteolysis . As an additional verification of this model , we analyzed CtrA activity by measuring its occupancy at the origin of replication ( Cori ) upon stress treatment by using quantitative chromatin immunoprecipitation ( qChIP ) . We compared CtrA occupancy between the ΔcpdR mutant and the cckA ( G319E ) mutant , in both of which CtrA is completely stbilized upon stress exposure ( Fig 5B , Fig 6A ) . In the ΔcpdR mutant addition of EtOH led to a decrease in CtrA occupancy to 18% ( Fig 6E ) , indicating that stress treatment results in CtrA dephosphorylation and inactivation . By contrast , in cells expressing cckA ( G319E ) , CtrA occupancy at Cori remained high even after stress exposure ( Fig 6E ) . These results reinforce our conclusion that stress triggers the inactivation and proteolysis of CtrA by stimulating CckA phosphatase activity . How is CckA phosphatase activity stimulated under stress conditions ? Previous work has reported different mechanisms that allow CckA to switch between its kinase and its phosphatase mode under standard conditions [22 , 24] . One such mechanism depends on the upstream regulatory factors DivL and DivK . While DivL directly interacts with CckA at the swarmer pole in stalked and predivisional cells and activates its kinase activity [23] , DivK acts as an antagonist of DivL-CckA favoring the phosphatase mode of CckA at the stalked pole [22] . Consistent with DivL's role in activating CckA kinase activity , previous work showed that loss of DivL function leads to a strong reduction in the level of phosphorylated CtrA and rapid CtrA degradation [22] . In agreement with these data we found that shifting a divLts mutant strain to the restrictive temperature resulted in rapid proteolysis and downregulation of CtrA ( Fig 7A and 7B ) [22] . Because the rate of CtrA degradation in the divLts strain was similarly fast as in wild type cells upon EtOH treatment , we hypothesized that stress-dependent inactivation and proteolysis of CtrA might be caused by a failure to accumulate and localize DivL . However , our microscopy data showed that fluorescently tagged DivL ( DivL-GFP ) correctly localized in a cell cycle-dependent manner even in the presence of EtOH ( Fig 7C , S9 Fig ) . Similarly , we did not observe that EtOH treatment induced significant changes in the localization pattern of CckA that changes during the cell cycle ( Fig 7D , S9 Fig ) [35] . By contrast , loss of DivL function has previously been shown to cause mislocalization of CckA [22 , 23] . Together , these data suggest that CtrA inactivation in response to adverse conditions is not due to failure to accumulate or localize DivL . To investigate whether the switch of CckA into its phosphatase mode is caused by increased activity of DivK , we measured CtrA levels following EtOH treatment in a strain in which DivK was inactivated . When shifting a cold-temperature sensitive mutant of DivK ( divKcs ) to the restrictive temperature 20°C [22 , 36] , we observed the same rapid downregulation of CtrA upon EtOH treatment as in the wild type ( Fig 7E ) , suggesting that the stress-dependent regulation of CtrA activity is independent of DivK . In addition to the upstream regulators DivK and DivL , it has recently been shown that c-di-GMP can modulate CckA activity through a direct interaction . Binding of c-di-GMP to CckA inhibits the kinase and stimulates the phosphatase activity of CckA [24] . To test the possibility that EtOH affects CckA activity through c-di-GMP , we investigated if the EtOH-dependent changes in CtrA abundance and stability were abolished in a CckA ( Y514D ) mutant that is compromised for c-di-GMP binding [24] . However , just like in the wild type , addition of EtOH resulted in rapid CtrA downregulation in this mutant ( Fig 7E ) . Similarly , we found that in a strain lacking all diguanylate cyclases ( cdG0 ) that is devoid of c-di-GMP [25] , EtOH addition resulted in the same drop in CtrA abundance as in the wild type ( Fig 7E ) . Consistent with the downregulation of CtrA in divKcs , CckA ( Y514D ) and cdG0 cells , we observed that EtOH-treatment led to the similar phenotypic changes as observed in wild type cells; cells became filamentous and accumulated multiple chromosomes ( Fig 7F , Fig 1A ) . Altogether these data suggest that external EtOH stress affects CckA and CtrA activity neither via the DivK-DivL pathway nor through c-di-GMP signaling . Hence , the histidine kinase might directly respond to external changes . We wondered whether the stress-dependent downregulation of CtrA provides a selective advantage and helps the culture to survive and grow under adverse conditions . To investigate the physiological relevance of the rapid shut-down of CtrA function during ethanol stress , we compared the growth rate of wild type cells with that of cckA ( V366P ) mutant cells , in which CtrA is partially stabilized due to lower CckA phosphatase activity ( Fig 6A and 6B ) . Under optimal conditions , the wild type and the mutant had identical growth rates suggesting that CckA phosphatase activity is not critical for cell growth and cell cycle progression in the absence of stress ( Fig 8A ) . Interestingly however , in the presence of 4% EtOH the cckA ( V366P ) mutant strain showed a clearly reduced growth rate compared to the wild type . This result suggests that under stress conditions the shut-down of CtrA function due to increased CckA phosphatase activity indeed provides a growth advantage .
This work reports a new mechanism by which bacteria delay cell division and consequently transform into filamentous cells under stress . Unlike previously described mechanisms that transiently block cell division through the induction of small division inhibitors [26 , 27] , the mechanism that we describe depends on a phospho-signaling system that modulates the stability and activity of a master cell cycle regulator required for cell division . The bifunctional histidine kinase CckA plays a central role in this regulation as it makes the decision of whether to divide or not by integrating the cell cycle with environmental information . It is well established that under optimal conditions CckA drives oscillations of the master cell cycle regulator CtrA through dynamically switching between its kinase and phosphatase activities [16 , 22 , 34 , 37] . Our new data suggest that environmental stress locks CckA in its phosphatase mode leading to the rapid inactivation of CtrA , its elimination through the protease ClpXP and consequently a block of CtrA regulated functions including cell division ( Fig 8B ) . Although our data show that CckA phosphatase activity is critical for the stress-dependent inactivation of CtrA , the detailed molecular process by which stress signals modulate CckA function remains to be elucidated . Our results rule out the involvement of the small signaling molecule c-di-GMP ( Fig 7E and 7F ) , which promotes CckA phosphatase activity at the G1-to-S transition and promotes CtrA degradation by ClpXP under non-stress conditions [21 , 24 , 37] . Similarly , the stress-dependent regulation does not appear to be mediated by the upstream regulators DivL and DivK ( Fig 7 ) , which are critical for the cell cycle-dependent regulation of CckA in the absence of stress [22] . The result that none of the known CckA regulators were involved in the stress-dependent regulation of CckA together with the rapid response time that we observed ( Fig 5 ) argues for a direct sensing mechanism . CckA has two PAS ( Per Arnt Sim ) domains , PAS-A and PAS-B , which sense distinct spatiotemporal signals and thereby mediate the cell cycle-dependent regulation of CckA activity [34] . These PAS domains might also perceive information about the environment , for example by binding molecules that are present under certain conditions . Alternatively , stress-sensing could be mediated in a fashion independent of CckA's PAS domains and instead involve its periplasmic or transmembrane regions . Indeed , previous work on other kinases demonstrated that environmental information can be directly sensed through the membrane [38] . For example , the histidine kinase DesK from Bacillus subtilis was shown to respond to temperature changes by sensing membrane thickness [39] . It is well documented that salt , EtOH and increased temperature induce changes in membrane properties , for example membrane fluidity or lipid composition [40–42] . These changes might directly induce changes in CckA conformation and activity . A direct sensing mechanism by CckA would provide an efficient means to transduce environmental information into the cell cycle . Nevertheless , we do not rule out the involvement of unidentified regulatory proteins that may interact with CckA to control its activity in response to stress . Other studies have analyzed the response of C . crescentus to increased salt concentrations . One of them investigated gene expression and proteome changes upon treatment with 60 mM NaCl [43] . Under this condition growth rate and CtrA regulated genes were hardly affected [43] , which is consistent with our finding that the salt-induced filamentous phenotype occurred only in a narrow range of concentrations ( Fig 1B ) . Another recent study investigated the response of individual cells of C . crescentus to repeated salt exposure using a microfluidics system [44] . The authors observed that an initial exposure to moderate NaCl concentrations ( 80 mM ) led to a cell division delay and cell-cycle synchronization and that the response of individual cells to a subsequent exposure to a higher NaCl concentration ( 100 mM ) was dependent on the cell cycle state [44] . It is possible that the stress-induced changes in CtrA activity that we report here contribute to these behaviors . Noticeably , while salt stress , EtOH stress and mild heat shock lead to rapid elimination of CtrA , our data as well as previously published results show that carbon starvation causes an increase in CtrA stability by a mechanism involving the small signaling molecule ( p ) ppGpp [28 , 29 , 45] . Although the precise mechanism of the starvation-dependent increase in CtrA stability remains unclear , it likely ensures , in combination with the downregulation of the DNA replication initiator DnaA , a block of DNA replication initiation under this condition [1 , 28 , 29] . Besides elucidating the mechanisms of how cell division is environmentally controlled , another important question concerns why cells inhibit cell division under stress . Cell division is a vulnerable process involving extensive membrane and cell wall remodeling [46] . Initiating this process under stress conditions , in particular those impacting the cell membrane , potentially causes cell lysis and consequently death . Preventing cell division in the presence of stress might thus provide a mechanism to preserve cell integrity . Continued global macromolecule synthesis , which still can take place under the conditions that we described ( Fig 1A , S2 Fig ) , allows for the production of new cell mass , enabling the rapid generation of new daughter cells when conditions improve . It is also possible that the filamentous morphology of division-inhibited cells provides an adaptive advantage under certain conditions in nature . A filamentous cell shape is expected to influence various cell properties , including surface area , mobility and adhesive forces , and is thus expected to affect the interaction with other species and the attachment of cells to biotic and abiotic surfaces [3 , 4] . Finally , while we have focussed in this study on C . crescentus , other alphaproteobacteria might employ similar mechanisms to control CtrA and cell division in response to external cues . Previous work in the nitrogen-fixing plant symbiont Sinorhizobium meliloti demonstrated that CtrA is strongly downregulated during the early steps of symbiosis [47] . In the pathogen Brucella abortus the CckA-ChpT-CtrA-CpdR pathway was shown to be required for intracellular survival in human macrophages [48] . Therefore , precise environmental regulation of CtrA abundance and activity likely plays an important role for the diverse functions that different alphaproteobacteria perform in the environment .
Wild type C . crescentus NA1000 and its mutant derivatives were routinely grown in PYE ( rich medium ) or M2G medium ( minimal medium containing 0 . 2% glucose ) . When necessary , growth medium was supplemented with 0 . 3% xylose , 0 . 2% glucose or 1 mM IPTG . Cultures were grown at 30°C with 200 rpm , temperature sensitive mutants were cultivated at 30°C and sensitivity was induced either at 37°C or 20°C depending on the mutant allele . Antibiotics were added as previously described [30 , 49] . Rifampicin was used at concentrations of 2 . 5 μg/ml ( liquid media ) and 5 μg/ml ( solid media ) for C . crescentus and 25 μg/ml ( liquid media ) and 50 μg/ml ( solid media ) for E . coli . E . coli strains were routinely grown in LB medium at 37°C , supplemented with antibiotics as required . For induction of stress , mixed ( non-synchronized ) Caulobacter cultures grown overnight to exponential phase in PYE medium ( OD 0 . 1–0 . 4 ) were shifted to medium supplemented with NaCl , EtOH , mitomycin C , sucrose or H2O2 at the indicated concentrations . To induce heat shock , cultures grown at 30°C were diluted in pre-heated medium and cultivated at 37 , 39 , 40 , 42 or 45°C for the indicated time . Carbon starvation was induced by shifting cells from M2G to M2 medium supplemented with 0 . 02% glucose as previously described [28] . Low and high pH stress medium was prepared by adjusting the pH with HCl or Na2CO3 and NaHCO3 [32] to the pH values of 4 . 9 and 9 . 1 , respectively . If necessary , cultures were backdiluted during the stress treatment to keep them in exponential phase . Strains used in this study are listed in S1 Table . Strain ΔcpdR::rif , KJ798 , was created by using the two-step recombination procedure [50] with plasmid pKJ808 . Strain KJ799 was generated by introducing the plasmid pKJ809 into C . crescentus NA1000 by electroporation . To construct strain KJ800 , plasmid pKJ810 was introduced instead . To construct strain KJ811 the empty vector pJS14 was introduced into C . crescentus NA1000 by electroporation . Samples were prepared as earlier described [28] and analyzed using a BD LSRFortessa flow cytometer or the BD LSR II ( BD Biosciences ) . Data were collected for at least 30000 cells . Flow cytometry data were analyzed with FlowJo . Each experiment was repeated independently and representative results are shown . For living cell analysis and time-lapse microscopy , cells were transferred onto a PYE 1% agarose pad with supplementation as required . Otherwise cells were fixed with 1% formaldehyde , pelleted , resuspended in an appropriate volume of ddH2O and mounted onto 1% agarose pads . Phase contrast and fluorescence images were taken using a Ti eclipse inverted research microscope ( Nikon ) with a 100x/1 . 45 NA objective ( Nikon ) . Fiji ( ImageJ ) was used for image processing . One ml culture was harvested and resuspended in 1 ml ddH2O . The optical density and the GFP fluorescence of 100 μl cells was measured in a SpectraMax i3x ( Molecular Devices ) plate reader . The fluorescence / OD ratio was calculated after blanking and the WT auto fluorescence signal was subtracted from the GFP signal . Pelleted cells were resuspended in 1X SDS sample buffer , normalized to the optical density of the culture and heated to 95°C for 10 min . Protein extracts of cell lysates were then subject to SDS-PAGE for 90 min at 130 V at room temperature on 11% Tris-glycine gels and transferred to nitrocellulose membranes . To verify equal loading , total protein was visualized using the TCE in-gel method [52] prior to blotting . Proteins were detected using primary antibodies against CtrA ( kindly provided by M . Laub ) , DnaK or FtsZ ( kindly provided by M . Thanbichler ) in appropriate dilutions , and a 1:5000 dilution of secondary HRP-conjugated antibody . SuperSignal Femto West ( Thermo Scientific ) was used as detection reagent . Blots were scanned with a Chemidoc ( Bio-Rad ) system . Images were processed with Bio-Rad Image Lab , Adobe Photoshop , Image J and the relative band intensities quantified with Image Lab software . To measure protein degradation in vivo , cells were grown under the desired conditions . Protein synthesis was blocked by addition of 100 μg/ml chloramphenicol . Samples were taken as indicated , every 10 min ( for 1 hour ) or at the time points 0 , 3 , 6 , 9 , 12 , 20 , and 30 minutes , and pellets were snap frozen in liquid nitrogen before being analyzed by Western blotting . RNA was collected from bacteria that were grown under the appropriate conditions and extracted using the RNeasy mini kit ( Qiagen ) . RNA-sequencing was performed by GENEWIZ , South Plainfield , NJ . For statistical analysis , the transcriptome data of the EtOH and NaCl stress conditions were compared to each other , to divLts transcriptome data and to previously published DNA damage microarray [26] . In order to integrate RNA-seq and microarray transcriptomics data we applied a fold-change cut-off of two to identify up- and down-regulated genes . For each condition , genes with a fold-change > 2 in the stress / non-stress sample comparison were assigned to the up-regulated group and genes with a fold-change < 0 . 5 to the down-regulated group . The similarity between two conditions is reflected by the extent of intersection of the genes in the respective up-regulated and down-regulated gene groups . As the intersection depends strongly on the size of the groups we used Monte-Carlo simulations to statistically evaluate group intersections . The z-score indicates the number of standard deviations from a random intersection . Hence , a high z-score means a strong deviation from a random similarity value . Samples were prepared as previously described [53] with the following modification: Lysates from one independent experiment were sonicated simultaneously ( Bioruptor Plus , diagenode , Ougrée , BE ) using 40 cycles of 30 seconds sonication ( High ) and 30 seconds pause . Real-time PCR was performed with a StepOnePlus real-time PCR system ( AppliedBiosystems , Foster City , CA ) using 5% of each ChIP sample and the SYBR green PCR master mix ( Bio-Rad ) in a 20 μl volume and 10 pmol primers ( OKH79 and OKH80 ) , amplifying a 88 bp region spanning the c and d CtrA binding boxes within Cori . The cycle threshold ( Ct ) of the input DNA was adjusted to 100% . The percentage of the input DNA was calculated ( 100*2^ ( adjusted input DNA-Ct ( IP ) ) ) for every condition and mutant . Each qPCR reaction was performed in triplicate . RNA was collected from bacteria that were grown under the appropriate conditions as described above . Equal amounts of isolated RNA were reverse transcribed into cDNA using the iScript cDNA synthesis kit ( Bio-rad ) . The cDNA was used as template for the real-time PCR reaction using the iTaq universal SYBR Green Supermix ( Bio-rad ) and primers as listed in S2 Table . Analysis was performed with a qTower instrument ( Analytik Jena ) using the standard run mode . For detection of primer dimerization or other artifacts of amplification , a dissociation curve was run immediately after completion of the real-time PCR . Individual gene expression profiles were normalized against 16S RNA , serving as an endogenous control . Relative expression levels were determined with the comparative Ct method . Each qPCR reaction was performed in triplicate . Complete data sets for the RNA-seq experiment are provided in S3 Table in the supplemental material and are available through GEO ( accession number GSE90030 ) .
|
Free-living bacteria are frequently exposed to various environmental stress conditions . To survive under such adverse conditions , cells must induce pathways that prevent and alleviate cellular damages , but they must also adjust their cell cycle to guarantee cellular integrity . It has long been observed that various bacteria transform into filamentous cells under certain conditions in nature , indicating that they dynamically modulate cell division and the cell cycle in response to environmental cues . The molecular bases that allow bacteria to regulate cell division in response to fluctuating environmental conditions remain poorly understood . Here , we describe a new mechanism by which Caulobacter crescentus blocks division and transforms into filamentous cells under stress . We find that the observed cell division block depends on precise regulation of the key cell cycle regulator CtrA . Under optimal conditions , the membrane-bound cell cycle kinase CckA activates CtrA in response to spatiotemporal cues to induce expression of genes required for cell division . Our data suggest that external stress triggers CckA to dephosphorylate and inactivate CtrA , thus ensuring the downregulation of CtrA-regulated functions , including cell division . Given that CckA and CtrA are highly conserved among alphaproteobacteria , the mechanism found here , might operate in diverse bacteria , including those that are medically and agriculturally relevant .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cellular",
"stress",
"responses",
"chemical",
"compounds",
"caulobacter",
"enzymes",
"cell",
"cycle",
"and",
"cell",
"division",
"metabolic",
"processes",
"cell",
"processes",
"enzymology",
"organic",
"compounds",
"phosphatases",
"prokaryotic",
"models",
"dna",
"replication",
"proteolysis",
"experimental",
"organism",
"systems",
"dna",
"alcohols",
"bacteria",
"research",
"and",
"analysis",
"methods",
"proteins",
"caulobacter",
"crescentus",
"gene",
"expression",
"chemistry",
"ethanol",
"biochemistry",
"cell",
"biology",
"organic",
"chemistry",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"metabolism",
"organisms"
] |
2016
|
A Kinase-Phosphatase Switch Transduces Environmental Information into a Bacterial Cell Cycle Circuit
|
A reduction in number and an increase in size of inflorescences is a common aspect of plant domestication . When maize was domesticated from teosinte , the number and arrangement of ears changed dramatically . Teosinte has long lateral branches that bear multiple small ears at their nodes and tassels at their tips . Maize has much shorter lateral branches that are tipped by a single large ear with no additional ears at the branch nodes . To investigate the genetic basis of this difference in prolificacy ( the number of ears on a plant ) , we performed a genome-wide QTL scan . A large effect QTL for prolificacy ( prol1 . 1 ) was detected on the short arm of chromosome 1 in a location that has previously been shown to influence multiple domestication traits . We fine-mapped prol1 . 1 to a 2 . 7 kb “causative region” upstream of the grassy tillers1 ( gt1 ) gene , which encodes a homeodomain leucine zipper transcription factor . Tissue in situ hybridizations reveal that the maize allele of prol1 . 1 is associated with up-regulation of gt1 expression in the nodal plexus . Given that maize does not initiate secondary ear buds , the expression of gt1 in the nodal plexus in maize may suppress their initiation . Population genetic analyses indicate positive selection on the maize allele of prol1 . 1 , causing a partial sweep that fixed the maize allele throughout most of domesticated maize . This work shows how a subtle cis-regulatory change in tissue specific gene expression altered plant architecture in a way that improved the harvestability of maize .
The “domestication syndrome” of crop plants is a suite of adaptive traits that arose in response to direct and indirect selection pressures during the domestication process [1]–[3] . This suite of traits includes an increase seed or fruit size , larger inflorescences , an increase in apical dominance , more determinate growth and flowering , loss of natural seed dispersal , loss of seed dormancy , and , in some cases , the gain of self-compatibility . These traits make crop plants easier to cultivate and harvest , resulting in increased value for human use . Among the domestication syndrome traits , the increase in apical dominance improves agricultural performance by enhancing harvestability . Apical dominance confers a reduction in the number of branches and inflorescences per plant . The inflorescences that do form , however , have either more and/or larger fruits or seeds . Thus , increased apical dominance can afford easier harvestability by reducing the number of inflorescences to be harvested without a concomitant loss in yield per plant . Moreover , larger seeds allow for more vigorous growth after germination when seedlings can face intense competition from weedy species . Finally , the fewer but larger inflorescences mature in a narrower window of time , enabling all the fruit/seed of a plant to be harvested at the same time of optimal maturation . Maize was domesticated from Balsas teosinte ( Zea mays subsp . parviglumis ) through a single domestication event in Mexico about 9000 years ago [4] , [5] . During maize domestication , there was a profound increase in apical dominance such that the amount of branching and the number , size and arrangement of the female inflorescences ( ears ) changed dramatically [6] , [7] . The teosinte plant has multiple long lateral branches , each tipped with a tassel . At each node along these lateral branches , there are clusters of several small ears ( Figure 1A ) . Summed over all branches , a single teosinte plant can easily have more than 100 small ears . By comparison , the maize plant has relatively few lateral branches ( often just two ) , each tipped by a single large ear rather than a tassel as in teosinte ( Figure 1C ) . Modern commercial varieties of maize typically have only one or two ears per plant , and even traditional landraces of maize rarely have more than 6 ears per plant . In maize genetics and breeding , the number of ears on a plant is scored as prolificacy , teosinte having high and modern maize low prolificacy . Here , we report a genome-wide scan for prolificacy QTL using a maize-teosinte BC2S3 mapping population [8] . We also report the fine-mapped of one of the discovered QTL to a 2 . 7 kb “causative region” located 7 . 5 kb upstream of the coding sequence of the known maize gene grassy tillers1 ( gt1 ) , which encodes a homeodomain leucine zipper ( HD-ZIP ) transcription factor [9] . We characterize the change in expression of gt1 between the maize and teosinte alleles of our mapping population , and the relationship between this expression change and reduced prolificacy in maize . We also performed molecular population genetic analysis that suggests the causative region was the target of a partial selective sweep that brought a haplotype at low frequency in teosinte to a higher frequency over most of the range of maize landraces . Our results show that a subtle change in the tissue specific gene expression is associated with a reduction in prolificacy during domestication .
Whole genome QTL mapping for loci affecting prolificacy was performed using a set of 866 maize-teosinte BC2S3 recombinant inbred lines ( RILs ) . This analysis identified eight QTL , distributed across the first 5 chromosomes ( Figure 2 , Table 1 ) . Of the eight QTL , one has a much larger effect than the other seven . This QTL ( prol1 . 1 ) is located on the short arm of chromosome 1 and accounts for 36 . 7% of the phenotypic variance . Plants in the mapping population that are homozygous teosinte at prol1 . 1 typically produce multiple ears at each node like teosinte ( Figure 1B ) . The 1 . 5 LOD support interval surrounding prol1 . 1 defines a 0 . 79 Mb segment between 22 . 63 Mb and 23 . 42 Mb ( B73 Reference Genome v2 ) on chromosome 1 . This region contains just 25 annotated genes including gt1 . The other seven QTL have much smaller LOD scores and smaller effects . This disparity in QTL size suggests that although the seven smaller QTL contribute to prolificacy , the phenotype is primarily controlled by prol1 . 1 . We chose prol1 . 1 for fine-mapping to identify the underlying causative gene . Two markers ( umc2226 and bnlg1803 ) that flank the QTL interval were used to screen for recombinant chromosomes in one of the 866 BC2S3 RILs that is heterozygous in the prol1 . 1 QTL interval . After screening ∼4000 plants of this RIL , 23 plants with a cross-over between the two markers were identified and self-pollinated to create progeny lines homozygous for the 23 recombinant chromosomes . The physical position of each of the 23 recombination events was determined using a combination of gel-based markers and DNA sequencing ( Figure 3 , Figure S1; Table S1 ) . Progeny lines homozygous for the 23 recombinant chromosomes were grown in a randomized-block design and scored for prolificacy . We also included two lines derived from the same BC2S3 RIL as controls: one homozygous teosinte and the other homozygous maize in the QTL interval . This set of 25 progeny lines fell into two discrete classes for prolificacy ( Figure 3 ) . One class , which included the maize control line , had an average prolificacy score of 2 . 38±0 . 05 ears . The other class , which included the teosinte control line , had an average prolificacy score of 7 . 24±0 . 12 ears . Separately , to estimate dominance relationships , we compared the trait values of the maize , teosinte and heterozygous genotypic classes at prol1 . 1 The dominance/additivity ratio is 0 . 08 , indicating additive gene action ( Table S2 ) . Examination of the relationship between the two phenotypic classes and the recombination breakpoints revealed that all members of the maize class carry maize chromosome between markers SBM07 ( AGP v2: 23 , 232 , 048 ) and SBM08 ( AGP v2: 23 , 234 , 775 ) ( Figure 3 , Figure S1 ) . Correspondingly , all members of the teosinte phenotypic class carry teosinte chromosome between these two markers . No other chromosomal region shows this absolute correspondence with phenotype . Thus , substitution mapping based on the recombination breakpoints indicates that prol1 . 1 or the factor that governs prolificacy maps to this interval . This interval , which we will refer to as the “causative region , ” is approximately 7 . 5 kb upstream of gt1 and measures 2720 bp in W22 , 3142 bp in our teosinte parent , and 2736 bp in the B73 reference genome ( Figure 3 , Figure S1 ) . The sequence alignment of W22 and the teosinte parent expands to ∼4 . 2 kb because there are several large insertions unique to either W22 or teosinte ( see below ) . The maize allele of prol1 . 1 confers a reduction in ear number , which by itself would cause a reduction in yield . To test whether there is a compensatory increase in either the number of kernels per ear or kernel weight , we assayed plants of the BC2S3 family used for fine-mapping to determine if prol1 . 1 has associated effects on these traits . The prol1 . 1 maize allele is not associated with an increase in ear size as measured by the total number of spikelets ( kernel forming units ) produced in the primary ear ( maize = 418 , heterozygous = 423 , teosinte = 421 , p = 0 . 86; Table S2 ) . However , the maize allele is associated with an increase in kernel weight ( maize = 0 . 216 g , heterozygous = 0 . 208 g , teosinte = 0 . 187 g , p<0 . 0001; Table S2 ) . Other aspects of plant architecture such as tillering and the number of nodes along the maize culm that produce ears do not appear to be affected by prol1 . 1 ( Table S2 ) . Thus , these data suggest that the reduction in secondary ears caused by prol1 . 1 in maize was compensated for by an increase in kernel weight such that yield itself may not have changed . Confirm of this interpretation would require a formal yield trial comparing the maize and teosinte genotypes . The location of prol1 . 1 at ∼7 . 5 kb upstream of coding sequence of gt1 suggests that it may represent a cis-regulatory element of gt1 . To investigate this possibility , we used ESTs from Genbank and genomic sequence of our maize and teosinte parents to construct a gene model for gt1 ( Figure S2 ) . This model agrees with the gt1 gene model presented elsewhere [9] . gt1 possesses three exons with two small introns and a transcript of ∼1350 bp that encodes a protein of 239 amino acids . The homeodomain and a putative nuclear localization signal are located in Exon 2 . We performed RT-PCR with primers designed to amplify most of the predicted transcript ( 1203 bp of the predicted 1350 bp ) using cDNAs isolated from immature ear-forming axillary branches of isogenic lines derived from our mapping population possessing the maize and teosinte alleles . We observed three size classes of RT-PCR products , presumably corresponding to three splice variants or isoforms of gt1 ( Figure 4 ) . The three size classes are present with both maize and teosinte alleles . We cloned and sequenced all three size classes and aligned these with the genomic sequence ( Figure S3 ) . The largest class contains the entire predicted open reading frame , encoding a predicted protein of 239 amino acids . The middle-sized product is missing most of Exon 2 and part of Exon 3 . The smallest-sized product is missing all of Exon 2 and parts of Exons 1 and 3 . Critically , the middle and small-sized products are both missing the homeodomain and all or part of the putative nuclear localization signal . The relative band intensities of different sized RT-PCR products ( Figure 4 ) suggest that transcript abundance for the isoforms differs between the maize and teosinte alleles: teosinte having a greater abundance of the full length product and maize a greater abundance of the middle-sized product that lacks the homeodomain . To test whether these differences in band intensity for the different isoforms are independent of the causative region , we performed RT-PCR with two of our recombinant isogenic lines . One of these has the teosinte causative region linked to the maize coding sequence ( T:M ) , and the other has the maize causative region linked to the teosinte coding sequence ( M:T ) . RT-PCR assays with these recombinant lines confirm that the differential band intensity for the isoforms is determined by the coding sequence and not the causative region 7 . 5 kb upstream of the coding sequence ( Figure 4 ) . To investigate the effect of the causative region on transcript abundance for our maize and teosinte alleles , we used an allele specific expression assay [10] . cDNA was made from RNA from immature ear-forming axillary branches of plants heterozygous at prol1 . 1-gt1 . PCR primers were designed flanking a 2 bp indel in the 3′ non-translated region that distinguishes the maize and teosinte alleles ( Figure S2 ) . This indel is in all three isoforms , and thus PCR products measure the overall difference in the abundance of the maize and teosinte transcripts without regard to any differences in relative abundance of the isoforms between maize and teosinte . In a heterozygous plant , the maize and teosinte alleles are expressed in the same cells with a common set of trans-acting factors , therefore any difference in transcript abundance of the alleles in heterozygous plants must be due to cis-regulatory factors . This assay shows a ratio of 1 . 35 for teosinte:maize gt1 transcript , suggesting a modest but statistically significant excess of teosinte relative to maize transcript ( z-test , p<0 . 001 ) . As an additional test of the effects of the causative region on gt1 transcript abundance , we used quantitative PCR ( qPCR ) to compare overall gt1 transcript abundance in immature ear-forming axillary branches of isogenic lines that are homozygous for the maize vs . teosinte alleles at prol1 . 1-gt1 . For this assay , we used a primer pair in the 3′ UTR of all three isoforms . The abundance of gt1 transcript relative to actin transcript for the teosinte class ( 1 . 03 , n = 12 ) was slightly higher than the maize class ( 0 . 88 , n = 12 ) , however this difference is not statistically significant ( t-test , p = 0 . 077 ) . Both the allele specific expression assay and qPCR suggest that the teosinte transcript abundance might be slightly higher than that of maize , but any difference is modest . Although a substantial change in gt1 transcript levels was not detected between the maize and teosinte alleles of prol1 . 1 in immature ear-forming axillary branches , we hypothesized that the absence of secondary ears in maize could be caused by a more subtle change that does not drastically alter overall transcript level but instead impacts the domain of gt1 expression . In order to test for such a tissue-specific expression difference , we performed RNA in situ hybridization on immature primary ear-forming branches of lines containing all possible combinations of the maize and teosinte causative region ( prol1 . 1 ) and gt1 coding sequence ( M:M , T:T , M:T , and T:M ) . A previous study demonstrated that gt1 is strongly expressed in the leaves of dormant tiller-forming lateral buds [9] , thus we anticipated that gt1 expression might differ in the leaves ( husks ) surrounding secondary ear buds of maize and teosinte . Contrary to this expectation , our sections revealed that lines containing the maize allele of prol1 . 1 ( M:M and M:T ) rarely , if at all , initiate secondary ear buds ( Text S1 , Table S3 ) . Expression of gt1 was observed in young leaves surrounding secondary ears of lines containing the teosinte allele of prol1 . 1 ( T:T and T:M ) ( Figure S4 ) , but was weak compared to dormant buds [9] , and required an extended incubation for detection , suggesting that these secondary ears are not dormant . Interestingly , an up-regulation of gt1 expression was observed in the stem node or nodal plexus [11] of primary branches for lines containing the maize allele of prol1 . 1 ( M:M and M:T , Figure 5 A , B ) . This nodal gt1 expression was either absent or only weakly detectable above background in lines containing the teosinte allele of prol1 . 1 ( Figure 5 C , D ) . While the nodal stripe of gt1 was weak , the difference between the maize and teosinte prol1 . 1 lines was consistently observed in both late ( Figure 5 ) and early staged ( Figure S5 ) ear-forming axillary branches . Taken together , these observations suggest that the allelic differences at prol1 . 1 involve changes in a cis-regulatory element that causes increased gt1 expression in the nodal plexus of maize , which in turn inhibits the initiation of secondary ear buds . To investigate whether the causative region shows evidence of past selection during maize domestication , we sequenced the entire causative region ( ∼2 . 7 kb ) plus flanking sequence ( ∼1000 bp upstream and ∼700 bp downstream ) in 15 inbred maize landraces and 9 inbred teosinte ( Text S2 , Table S4 ) . Diversity statistics across the region in both teosinte ( S = 85 , π = 0 . 00844 and Tajima's D = −1 . 16 ) and maize ( S = 32 , π = 0 . 00307 and Tajima's D = −0 . 439 ) are within the previously estimated range of these statistics for neutral genes [12] , where S and π were the number of segregating sites and nucleotide diversity , respectively . Although these data would superficially appear to be consistent with a loss of diversity due to the domestication bottleneck alone , a neighbor-joining tree of the sequences separates most maize from most teosinte sequences in the causative region ( Figure S6 ) . This separation of the mostly maize and mostly teosinte clusters reflects differences at numerous SNPs and multiple putative transposon insertions ( Figure S7 ) . We will refer to these maize and teosinte clusters hereafter as the class-M and class-T haplotypes , respectively . Linkage disequilibrium ( LD ) analysis of maize sequences confirms this separation , identifying a 2 . 5 kb block of strong LD corresponding to SNPs that differentiate class-M from class-T maize sequences ( Figure 6A , Figure S8 ) . This high LD block lies completely within the 2 . 7 kb causative region . The maize class-M haplotype in this block exhibits extremely low levels of nucleotide diversity ( π = 0 . 000740 ) and a strongly negative Tajima's D value ( D = −1 . 966 ) . These values are extremely unlikely under neutrality ( p<0 . 01; Text S2 ) , leading us to investigate instead a partial sweep model to explain the observed sequence data . To investigate the unusual pattern of diversity for the maize class-M haplotypes , we applied a maximum likelihood method to estimate the selection coefficient ( s ) and the degree of dominance ( h ) using structured coalescent simulations ( Text S2 ) . We specified a partial sweep model ( Figure 6B ) , consistent with the observation of both class-M and class-T haplotypes in domesticated maize sequences , and performed structured coalescent simulations over a wide range of parameter settings similar to previous studies [12] , [13] . Our maximum likelihood estimates suggest that the class-M allele is dominant ( h = 1 . 0 ) and under reasonably strong selection ( s = 0 . 0015 ) ( Figure 6C ) . We also estimated the age of class-M haplotype to be ∼13 , 000 generation ago using Thomson's method [14] , [15] . Although the observed length ( 2 . 5 kb ) of the swept region may seem short , simple calculations show that this length falls within the ∼1–7 kb range expected given available estimates of recombination and the age of the haplotype ( Text S2 ) . We assayed a diverse sample of maize and teosinte to better estimate the frequencies of the class-M and class-T haplotypes ( Table S5 ) . We used an ∼250 bp insertion specific to the class-T haplotype as a marker . We observed that the class-M haplotype exists at a relatively low frequency in ssp . parviglumis ( 5% ) and ssp . mexicana ( 8% ) while the class-T haplotype exists at a moderate frequency in maize landraces ( 29% ) ( Table 2 ) . These frequencies are consistent with the partial selective sweep discussed above that brought the class-M haplotype from a low frequency ( 5% ) in the progenitor population to a relatively high frequency ( 71% ) in domesticated maize . An examination of the distribution of the class-T haplotype in maize shows a distinct geographic pattern ( Figure S9 ) . With only three exceptions , the class-T haplotype is limited to southern Mexico , the Caribbean Islands and the northern coast of South America . One exception is its occurrence in the landrace Tuxpeño Norteño in northern Mexico , but this is a landrace thought to be recently derived from the landrace Tuxpeño of southern Mexico [4] . The two other exceptions are found in southern Brazil in landraces thought to have been brought to Brazil in the 1800s from the southern USA [16] . In turn , the southern US landraces are thought to have been brought there from southern Mexico and the Caribbean in the 1600s by the Spanish [17] . Thus , the class-T haplotype in maize has a distribution centered on southern Mexico and the Caribbean with recent dispersals to other regions .
A critical challenge during the domestication of crop plants was to improve the harvestability of the crop as compared to its progenitor . Many wild species are adapted to “spread their bets” and thereby increase the probability of successful reproduction under diverse environments [2] . This is especially true of annual species , like the ancestors of many crops , that colonize disturbed habitats [2] . In unfavorable environments , such species can flower and mature rapidly , producing smaller numbers of branches , inflorescences , flowers and seeds but still complete their reproductive cycle . In favorable environments , such species can flower over a longer period , sequentially producing more branches , inflorescences , flowers and seeds over time , maximizing their reproductive output . The latter strategy is not optimal for a crop as greater efficiency of harvest is achieved by having all seed mature synchronously . Similarly , harvesting a single large inflorescence or fruit from a plant is easier than harvesting dozens of smaller ones [18] . Thus , diverse crops have been selected to produce smaller numbers of larger seeds , fruits or inflorescences as a means of improving harvestability [2] . In the terminology of modern day maize breeders , crops were selected to be less prolific . Our QTL mapping for prolificacy confirms the results of three prior studies that indicated this trait is controlled by a relative small number of QTL including one of large effect on the short arm of chromosome 1 . First , in an F2 cross of Chalco teosinte ( Zea mays ssp . mexicana ) with a Mexican maize landrace ( Chapalote ) , one of the four detected QTL was located on the short arm of chromosome 1 and accounted for upwards of 19% of the phenotypic variance in prolificacy [19] . Second , in an F2 cross of Balsas teosinte with a different Mexican maize landrace ( Reventador ) , one of the seven detected QTL was located on the short arm of chromosome 1 and accounted for 25% of the phenotypic variance [20] . Finally , in a maize-teosinte BC1 cross of Balsas teosinte by a US inbred line ( W22 ) , seven prolificacy QTL were detected [21] . All seven QTL had small effects , but the one that explained the greatest portion of the variance ( 4 . 5% averaged over two environments ) was on the short arm of chromosome 1 . As in these prior studies , the QTL mapping reported here indicates that prolificacy is under relatively simple genetic control , involving only 8 QTL but including one QTL ( prol1 . 1 ) of large effect . prol1 . 1 accounted for 36 . 7% of the variation in the number of ears and reduces the number of ears from 7 . 2 for teosinte homozygous class to 2 . 4 for the maize homozygous class . The genetic architecture of the change in prolificacy during domestication appears to be relatively simple in several other crops as well . In tomato , five QTL of roughly equal effects for the number of flowers per truss between wild and domesticated tomato were detected [22] , [23] . In the common bean , three QTL were detected for the reduction in the number of pods per plant in a cross of wild and domesticated bean [24] . The QTL of largest effect confers a reduction from 29 to 17 pods per plant and accounts for 32% of trait variation . In pearl millet , the reduction in the number of spikes per plant is governed by four QTL , including one that controls 37% of trait variation [25] . In sunflower , the reduction of number of heads per plant was governed by seven QTL , one of which had a much larger effect than the other six [26] . This large effect QTL accounts for a difference of 4 . 8 heads per plant between the cultivated and wild genotypes , and it co-localizes with the previously described Branching ( B ) locus , which is known to influence apical dominance [27] . Thus , simple genetic architecture including QTL of relatively large effect is common for this trait . One theory of crop domestication is that traits change is often the result of recessive , loss of function alleles [28] . Contrary to this expectation , prol1 . 1 acts in an additive fashion with a dominance/additivity ratio of 0 . 08 , suggesting that domestication did not involve selection for a simple loss of function . Moreover , our expression assays indicate that gt1 has roughly equal expression in maize and teosinte ear-forming axillary branches and the phenotypic change is caused by a relatively subtle gain/increase of expression in the nodal plexus of the ear-forming branches of maize . These results demonstrate that rather than a simple loss of function allele , the gene underlying this QTL experienced an increase or gain of expression in a specific tissue . While selection for loss of function alleles may be a common feature of domestication , none of the three positionally mapped maize domestication QTL ( teosinte branched1 , teosinte glume architechture1 , and gt1 ) involved a loss of function allele [29 , 30 , this paper] . Seventy-five years ago , the “teosinte hypothesis” that a small number of large effect genes substitutions could convert teosinte into a useful food crop was proposed [31] . The experimental basis for this model was that maize-like and teosinte-like segregants were recovered in a large maize-teosinte F2 population at frequencies , suggesting that as few as five loci might control the critical differences in ear architecture . Subsequent QTL mapping identified six regions of the genome that harbor QTL of large effect on plant and ear architecture , consistent with the teosinte hypothesis [32] . Fine-mapping of two of these QTL identified an underlying gene of large effect in both cases . One of these is teosinte glume architecture ( tga1 ) that controls the difference between covered vs . naked grain [30] , and the other is teosinte branched ( tb1 ) , which conferred increased apical dominance during domestication [29] . In this paper , we have shown that a gene of large effect ( gt1 ) also underlies a third of these six QTL of large effect . This result adds further support to the view that a small number of genes of large effect were key in the dramatic morphological changes that occurred during maize domestication . Nevertheless , it is also clear a larger number of QTL of smaller effect on morphology were also involved in converting teosinte into modern maize [8] , [32] , [33] . The role played by genes of large effect , like gt1 , is not limited to maize domestication , but seems to be a common feature of plant domestication [34] . Recently , a large effect gene in sorghum that encodes a YABBY transcription factor was shown to control shattering vs . non-shattering inflorescences [35] . Previously , two domestication genes controlling shattering had been identified in rice , one encoding a homeodomain and the other a myb-domain transcription factor [36] , [37] . In tomato , two domestication genes for increase in fruit size have been isolated , one encoding a YABBY transcription factor and the other a putative cell signaling gene [38] , [39] . A single gene ( PROG1 ) , which encodes a zinc finger transcription factor , controls differences in plant architecture and grain yield between wild and cultivated rice [40] , [41] . The fine-mapping of prol1 . 1 was initiated using a publically available set of maize-teosinte RILs [8] . These RILs allow some QTL to be mapped to relatively small intervals . We mapped prol1 . 1 to a 0 . 79 Mbp segment that included only 25 annotated genes and then fine-mapped it to a 2 . 7 kbp causative interval . These same maize-teosinte RILs were recently used fine-map a QTL ( dtp10 . 1 ) for photoperiod response that was involved in the adaptation of maize to northern latitudes [8] , [42] . The dtp10 . 1 QTL was mapped to a 7 . 6 Mbp interval containing 103 annotated genes , and then fine-mapped to a 202 kbp interval containing a single annotated gene ( ZmCCT ) . Features of prol1 . 1 and dtp10 . 1 that made them good candidates for fine-mapping were ( a ) having large effects with strong statistical support ( LOD>100 ) so that progeny lines with recombinant chromosomes possessing the maize vs . teosinte alleles of the QTL segregated into two distinct classes ( i . e . Mendelized ) and ( b ) being located in genomic regions with sufficient recombination to capture multiple cross-overs per gene in an F2 family of 2000 plants . For example , prol1 . 1 is located near the end of the short arm of chromosome 1 , where we observed a recombination rate 1 . 3×10−3 cM/kbp which is over twice the genome-wide rate reported for a maize-teosinte crosses [21] . The location of prol1 . 1 just 7 . 5 kb 5′ of grassy tillers1 ( gt1 ) suggested that it may act as a cis-regulatory element of gt1 . Whipple et al [9] identified gt1 as a HD-Zip transcription factor , a class of proteins that is unique to plants . The role of gt1 in maize development is complex . Although named for the excessive tillering caused by loss of function alleles , these alleles also cause the derepression of carpels in tassel florets , leading to the formation of sterile carpels [9] . Additional changes include an increased numbers of ear-forming nodes along the main culm , elongation of the lateral branches , and elongation of the blades on the husk leaves . The formation of secondary ears is occasionally ( but not typically ) seen with maize gt1 mutant allele consistent with the effect of prol1 . 1 on gt1 expression that we observed . The infrequency of this phenotype with the maize mutant alleles might be due to differences in genetic background between our lines , for which about 10% of the genome comes from teosinte , and the elite maize inbreds in which gt1 mutant alleles have been assayed . One curiosity is that the teosinte allele we studied does not confer an increase in tillering ( Table S2 ) , suggesting the role of gt1 in regulating tillering is conserved between maize and teosinte . Another HD-Zip transcription factor , six-rowed spike1 ( Vrs1 ) , has been identified as a domestication gene , controlling the change from two-rowed spikes in the wild progenitor of barley to six-rowed spikes found in domesticated barley [43] . Vrs1 is expressed in the lateral spikelet primordia of immature spikes of wild barley where it represses their development . Loss of function vrs1 alleles selected during domestication fail to repress the development of these lateral spikelets , resulting in two additional fully fertile spikelets per rachis node . A comparison of gt1 and vrs1 offers an interesting contrast . Loss of function of vrs1 alleles were selected in barley , producing a larger number of organs ( spikelets or grains ) per spike , while selection for an allele that confers the gain of nodal expression of gt1 in maize caused a reduction in the number of organs ( ears ) per plant . In maize , our data suggest the reduction in ear number may be compensated for by an increase in grain weight such that yield may not be affected . It would be of interest to know if the production of more grains per spike in barley is compensated for by a reduction in the number of spikes per plant such that yield is not affected although harvestability is improved . The nature of the causative polymorphism for prol1 . 1 that governs gt1 expression in the nodal plexus and represses secondary ear formation remains unknown . There are multiple polymorphisms that distinguish the class-M and class-T haplotypes for the causative region , all of which are potential candidates for the functional variant that controls expression in the nodal plexus ( Figure S7 ) . Among these polymorphisms are at least four transposable element insertions including Cinful , Pif/Harbinger , and hAT elements . Given the evidence that a Hopscotch transposon is the functional variant at tb1 [29] , the transposons in the causative interval of gt1 are good candidates for future functional assays . Transposon inserts have also been identified in alleles of genes involved in millet and tomato domestication or improvement [44] , [45] , suggesting that transposons may be important contributors to regulator variation in crop plants . DNA sequence analysis of the prol1 . 1 locus in diverse maize and teosinte accessions revealed two distinct haplotypes . Both haplotypes were present in maize and teosinte , but the class-M haplotype was common in maize and rare in teosinte . Neutral coalescent simulations revealed that patterns of diversity in the class-M haplotype in maize were unlikely in the absence of selection , and subsequent parameter estimation supported a partial sweep model in which selection acted to increase the frequency of the class-M haplotype during domestication . The estimated age of the class-M haplotype at 13 , 000 BP predates maize domestication and is consistent with its observed presence in about ∼5% of the teosinte sampled . This observation suggests that selection at prol1 . 1 acted on standing variation , similar to observations for tb1 [29] and barren stalk1 [46] . It is curious that the class-T haplotype is found at a frequency of nearly 30% in maize , although the multi-eared phenotype that this haplotype confers is rare in maize . Furthermore , none of the maize races ( Table S3 ) that carry the class-T haplotype are known to exhibit the multiple ears along a single shank . These observations suggest that these landraces may have other factors that suppress the formation of multiple ears on a single shank . Thus , there may have been two pathways to the switch from several to a single ear per node in maize , one governed by prol1 . 1 and a second controlled by unknown factors that suppress multiple ear formation in plants carrying the class-T haplotype at prol1 . 1 . The presence of such a second genetic pathway could also explain the incomplete selective sweep at prol1 . 1 . In some maize populations , fixation of low-prolificacy alleles at genes in this proposed second pathway could have reduced or eliminated selection on prol1 . 1 . Previous analysis of gt1 and surrounding sequence uncovered evidence of selection at the 3′ UTR of the gene [9] . We reanalyzed this sequence data ( Text S2 ) and identified two distinct haplotypes distinguished by a ∼40 bp indel . The class-M haplotype at this locus bears the signature of a partial sweep from standing variation similar to that seen at prol1 . 1 ( Text S2 ) . A PCR survey of a large panel of maize landraces reveals that the class-M haplotype at the 3′ UTR has an overall frequency of 78% . Combined with the small size of both sweeps and geographical differences in the abundance of each haplotype ( Figure S9 ) , these results suggest that the class-M haplotypes at prol1 . 1 and gt1 may represent independent selective events [47] , perhaps on different regulatory aspects of gt1 . Neither prol1 . 1 nor gt1 were identified in a recent whole-genome analysis of selection during domestication [48] , likely due to the short span of the selected region and the presence of the class-T allele in 30% of maize lines . This result highlights the difficulty in identifying small selected regions from genome-wide scans , especially in the case of soft sweeps [49] , [50] . The shade avoidance response in plants involves an increase plant height , a decrease in branching , reduction in the number of flowers , and early flowering [51] . During domestication , human preference for easier harvestability resulted in a form of plant architecture that mimics the shade avoidance in that crops are less branched and produce fewer reproductive structures . Two maize domestication genes , gt1 and tb1 , are members of the developmental network controlling the shade avoidance response [9] , suggesting that domestication acted to constitutively fix aspects of the shade avoidance syndrome in maize . As the shade avoidance network becomes better known , it will be of interest to see if additional genes within this network also play a role in domestication .
Whole genome QTL mapping for prolificacy in maize was performed using a set of 866 maize-teosinte BC2S3 RILs that were genotyped at 19 , 838 markers using a “genotype by sequence” ( GBS ) approach [8] , [52] . The 19 , 838 markers were selected from over 50 , 000 GBS markers as the subset that defines the end-points of all cross-overs in the 866 RILs . For the RILs , the maize inbred W22 was the recurrent parent and the teosinte parent was CIMMYT accession 8759 of Zea mays ssp . parviglumis . The 866 lines were grown in 2 blocks during summer 2009 and two additional blocks in summers 2010 and 2011 at the West Madison Agricultural Research Center in Madison , WI . All four blocks were randomized and contained 866 plots with 10 plants per plot . Prolificacy was scored on five plants per plot as either ( 1 ) having secondary ears on the primary lateral branch or ( 0 ) lacking secondary ears on the primary lateral branch . Least Squared Means ( LSMs ) were determined for each line using the following model with PROC GLM ( SAS Institute , Cary , NC ) :Line represents the RILs ( 1 through 866 ) and seedlot represents different seed productions for a single RIL . Year is 2009 , 2010 or 2011 , and for 2009 there were two blocks ( A and B ) . The position of each plot within a block was recorded along the x-axis and y-axis of the field . Only the x-axis and the interaction between the x and y axes had a statistically significant effect so the y-axis was dropped from the model . The LSMs showed a continuous range of values and were used as the phenotypic values for QTL mapping . QTL mapping was carried out using a modified version of R/qtl [53] that allows the program to take into account the BC2S3 pedigree of the lines [8] . Given that the LSM showed continuous variation , the QTL model was set to “normal” for a normal distribution in R/qtl . The percentage of variance explained by each QTL was estimated by a drop-one-ANOVA as implemented in R/qtl [53] . We used one of the BC2S3 RILs ( MR0091 ) for fine-mapping of prol1 . 1 . MR0091 is heterozygous for a 33 . 9 Mb region including this QTL and homozygous maize for all other prolificacy QTL . We screened ∼4 , 000 MR0091-derived plants for cross-overs in the QTL interval between markers umc2226 and bnlg1803 . Twenty-three individuals with cross-overs in the QTL interval were identified and selfed . Selfed progeny from these 23 individuals that are homozygous for the recombinant chromosome plus two control lines ( homozygous non-recombinant maize and teosinte ) were grown in randomized block design with four blocks of 25 entries each . Prolificacy was scored as the total number of ears observed on the top two lateral branches of each plant . Thus , for maize ( W22 ) , which has a single ear per lateral branch , the prolificacy score is 2 . LSMs with standard errors for prolificacy for each of the recombinant chromosome progeny lines and controls were determined by ANOVA with line and block effects using the software package JMP version 4 . 0 ( SAS Institute , Cary , NC ) . To determine if there are pleotropic effects on other traits associated with prol11 . 1 , we genotyped ∼200 plants of RIL MR0091 that segregates for this QTL and measured tillering , number of ear branches , spikelet ( kernel ) number on the top ear of the plant , and the weight of 100 kernels . Plants for these experiments were grown at the West Madison Agricultural Research Station in Madison , WI . For all expression assays , total cellular RNA was isolated using Trizol ( Invitrogen ) from immature ear-forming axillary branches . A 1 µg aliquot of each of RNA sample was DNase treated and reverse transcribed using a polyT primer and Superscript III reverse transcriptase ( Invitrogen ) . cDNA integrity was checked by using 0 . 5 µl of the RT reactions as the template for PCR ( Taq Core Kit , Qiagen ) with actin primers ( 5′-ccaaggccaacagagagaaa-3′ , 5′-ccaaacggagaatagcatgag-3′ ) . The same actin primers were used to check for genomic DNA contamination; none was detected . To confirm the intron-exon structure of gt1 , PCRs were performed with cDNAs with primers ( 5′-acaggctacagaggcagagc-3′ , 5′-gcgcacttgcatgataatccacac-3′ ) that amplify most of the predicted transcript ( Figure S2 ) . cDNAs derived from both the maize and teosinte alleles were used . PCR products were assayed on standard Tris-borate-EDTA agarose gels . These PCRs consistently revealed three size classes of products for both maize and teosinte alleles . These PCR products were cloned using TOPO TA Cloning Kit ( Invitrogen ) and the clones sequenced at the University of Wisconsin Biotechnology Center using Sanger sequencing . Since the relative abundance of the three PCR size classes differed between the maize and teosinte alleles , we also assayed cDNAs derived from two lines with recombinant alleles: one having teosinte “causative region” and maize coding region ( W22-QTL1S-IN0383 ) , the other having maize “causative region” and teosinte coding region ( W22-QTL1S-IN1043 ) ( Figure S1 ) . To compare gt1 transcript accumulation for the maize and teosinte alleles , we performed an allele specific expression assay [10] with cDNAs from ear-forming axillary branches of 20 plants that were heterozygous for the maize/teosinte alleles of our mapping population . One µl aliquots of the 20 RT reactions were used as the template for PCRs with a primer pair in the 3′ UTR of gt1 including one fluorescently labeled primer ( 5′-FAM-catgatggacctcgcgcccg-3′ , 5′-gcgcacttgcatgataatccacac-3′ ) . This primer pair flanks a 2 bp indel that distinguishes the maize and teosinte transcripts . PCR products were assayed on an ABI 3700 fragment analyzer ( Applied Biosystems ) and the areas under the peaks corresponding to the maize and teosinte transcripts were determined using Gene Marker version 1 . 70 ( Softgenetics , State College , PA ) . The relative message level associated with the maize vs . teosinte alleles in each of the twenty samples was calculated as the ratio of the area under teosinte/maize allele peaks . Two technical replicates were performed for each of the 20 biological replicates . The same assay was also performed with the DNA from each plant used for RNA extraction to assess any bias in allele amplification in the PCRs . The DNA analysis showed a slight bias towards the maize allele with maize/teosinte ratios of 1 . 05 . Thus , the area under the teosinte peak with the cDNAs was multiplied by 1 . 05 to correct this bias . We also compared transcript accumulation for the maize and teosinte alleles using quantitative real-time PCR ( qPCR ) with cDNA from immature ear-forming axillary branches of 12 homozygous maize and 12 homozygous teosinte plants as described above . For this assay , cDNA was first concentrated using RNAClean XP beads ( Beckman Coulter ) . qPCR was performed on ABI Prism 7000 sequence detection system ( Applied Biosystems ) with Power SYBR Green PCR Master Mix ( Applied Biosystems ) . Transcript abundance for gt1 was assayed using a set of primers in the 3′ UTR ( 5′-gcaatcaaggtcactagtatagtctg-3′; 5′-gcgcacttgcatgataatccacac-3′ ) . Actin primers ( see above ) were used as the control . The annealing temperature/time used were 52°C for 30 sec; the extension temperature/time were 72°C for 45 sec . Young ear-forming axillary buds ( 44–50 days after planting ) were collected from the top two nodes bearing lateral buds from field grown plants . These ears were fixed in 4% para-formaldehyde 1 X phosphate-buffered saline overnight at 4°C , then dehydrated with an ethanol series and embedded in paraffin wax . Embedded tissue was sectioned to 8 µM with a Leica RM2155 microtome . The full gt1 cDNA coding sequence was used as a probe as described previously [9] . In situ hybridization with digoxygenin-UTP labeled antisense probe was preformed as previously described [54] . Strong gt1 expression characteristic of dormant lateral bud leaves or tassel floret carpels requires a relatively short development of the color reaction ( 3–4 hrs ) , while weaker gt1 expression in leaves of non-dormant buds and shoot nodes requires a more extended development ( 15–20 hrs . ) . We sequenced the gt1 control region plus some flanking sequence ( AGP v2: 23 , 231 , 760 to 23 , 235 , 500 ) for a set of 15 diverse maize and 9 diverse teosinte lines ( Table S4; Genbank Accessions KC759702-KC759727 ) . Initial PCR primers were designed at either end of this interval based on the B73 reference genome . PCR products for each of the 24 diverse lines were sequenced using the Sanger method . A primer walk across the interval was performed for each of the 24 lines . In cases where B73 specific primers failed for one of the diverse lines because of sequence divergence or large insertions , we used consensus sequence data from the diverse lines that were successfully amplified to design primers in conserved regions . Sequences were aligned with Clustal X [55] , and checked manually . Alignment regions with gaps or ambiguous alignment were removed from further analysis . Because the teosinte and maize individuals sequenced were inbred lines , we treated the sequence as haploid data ( Table S4 ) . After removing all gapped and tri-allelic sites , 2 , 871 base pairs remained . We calculated the number of segregating sites ( S ) , nucleotide diversity ( π ) and Tajima's D for both maize and teosinte using custom perl scripts . We used MEGA5 [56] to infer a neighbor-joining ( NJ ) tree for the region ( Figure S4A ) , and STRUCTURE [57] to test for admixture ( Text S2 ) . We used structured coalescent simulations to estimate the maximum likelihood values of the selection coefficient ( s ) and degree of dominance ( h ) of the class-M haplotype . We simulated a simple domestication model including a demographic bottleneck and a partial selective sweep ( Text S2 ) . Coalescent simulations made use of a modified version of the mbs software [58] . To estimate population frequencies of the class-M and class-T haplotypes in the gt1 control region , we chose an ∼250 bp insertion in the teosinte haplotype at AGP v2: 23 , 232 , 564 in the B73 reference genome as a marker for the teosinte haplotype . This insertion was identified from the sequences of the 24 diversity lines discussed above . The insertion is present in all of the class-T haplotypes . Primers ( 5′-gagactggcgactggtcct-3′ , 5′-gacgtgcagacagcagacat-3′ ) were designed in conserved sequences flanking the insertion . PCRs with these primers yield an ∼600 bp product for the teosinte haplotype and an ∼350 bp product for the maize haplotype . PCR product size differences were scored on 2% agarose gels for a panel of 68 maize landraces , 90 Z . mays ssp . parviglumis and 96 Z . mays ssp . mexicana ( Table S5 ) .
|
Crop species underwent profound transformations in morphology during domestication . Among crops , maize experienced a more striking change in morphology than other crops . Among the changes in maize from its ancestor , teosinte , was a switch from 100 or more small ears per plant in teosinte to just one or two large ears in maize . We show that this change in ear number has a relatively simple genetic architecture involving a gene of large effect , called grassy tillers1 . Moreover , we show that grassy tillers1 experienced a tissue-specific gain in expression in maize that is associated with suppressing the initiation of multiple ears per plant such that only one or two large ears are formed . Our results show how simple changes in gene expression can lead to profound differences in form .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"plant",
"science",
"social",
"and",
"behavioral",
"sciences",
"crops",
"genetics",
"biology",
"evolutionary",
"biology",
"anthropology",
"agriculture"
] |
2013
|
From Many, One: Genetic Control of Prolificacy during Maize Domestication
|
It has been suggested that dopamine ( DA ) represents reward-prediction-error ( RPE ) defined in reinforcement learning and therefore DA responds to unpredicted but not predicted reward . However , recent studies have found DA response sustained towards predictable reward in tasks involving self-paced behavior , and suggested that this response represents a motivational signal . We have previously shown that RPE can sustain if there is decay/forgetting of learned-values , which can be implemented as decay of synaptic strengths storing learned-values . This account , however , did not explain the suggested link between tonic/sustained DA and motivation . In the present work , we explored the motivational effects of the value-decay in self-paced approach behavior , modeled as a series of ‘Go’ or ‘No-Go’ selections towards a goal . Through simulations , we found that the value-decay can enhance motivation , specifically , facilitate fast goal-reaching , albeit counterintuitively . Mathematical analyses revealed that underlying potential mechanisms are twofold: ( 1 ) decay-induced sustained RPE creates a gradient of ‘Go’ values towards a goal , and ( 2 ) value-contrasts between ‘Go’ and ‘No-Go’ are generated because while chosen values are continually updated , unchosen values simply decay . Our model provides potential explanations for the key experimental findings that suggest DA's roles in motivation: ( i ) slowdown of behavior by post-training blockade of DA signaling , ( ii ) observations that DA blockade severely impairs effortful actions to obtain rewards while largely sparing seeking of easily obtainable rewards , and ( iii ) relationships between the reward amount , the level of motivation reflected in the speed of behavior , and the average level of DA . These results indicate that reinforcement learning with value-decay , or forgetting , provides a parsimonious mechanistic account for the DA's roles in value-learning and motivation . Our results also suggest that when biological systems for value-learning are active even though learning has apparently converged , the systems might be in a state of dynamic equilibrium , where learning and forgetting are balanced .
Electrophysiological [1] and fast-scan cyclic voltammetry ( FSCV ) [2 , 3] studies have conventionally shown that dopamine ( DA ) neuronal activity and transmitter release respond to unpredicted but not predicted reward , consistent with the suggestion that DA represents reward-prediction-error ( RPE ) [1 , 4] . On the other hand , recent FSCV studies [5–8] have found DA response sustained towards presumably predictable reward , arguing that it may represent sustained motivational drive . DA's roles in motivation processes have long been suggested [9–13] primarily from pharmacological results . A key finding is that post-training blockade of DA signaling causes motivational impairments such as slowdown of behavior ( e . g . , [14] ) , and this is difficult to explain with respect to the known role of DA in RPE representation because post-training RPE should be negligible so that blockade of RPE should have little impact . Therefore it has been considered that DA has two distinct reward-related roles , ( 1 ) representing RPE and ( 2 ) providing motivational drive , and these are played by phasic and tonic/sustained DA , respectively . Normative theories have been proposed for both the role as RPE [4] and the role as motivational drive [15 , 16] in the framework of reinforcement learning ( RL ) . On the other hand , as for the underlying synaptic/circuit mechanisms , much progress has been made for the role as RPE but not for the role as motivational drive . Specifically , how RPE is calculated in the upstream of DA neurons and how released DA implements RPE-dependent update of state/action values through synaptic plasticity have now become clarified [17–20] . In contrast , both the upstream and downstream mechanisms for DA's motivational role remain more elusive . In fact , FSCV studies that found sustained DA signals [5 , 8] have shown that those DA signals exhibited features indicative of RPE . Moreover , sustained response towards presumably predictable reward has also been found in the activity of DA neurons [21 , 22] , and these studies have also argued that the DA activity represents RPE . Consistent with these views , we have recently shown [23] that RPE can actually sustain after training if decay/forgetting of learned values , which can presumably be implemented as decay of plastic changes of synaptic strengths , is assumed in RL . It was further indicated that whether RPE/DA sustains or not can be coherently understood as reflecting differences in how fast learned values decay in time: faster decay causes more sustained RPE/DA . However , this account did not explain the suggested link between sustained DA and motivation . Even on the contrary , decay of learned values is apparently wasteful and could be perceived as a loss of motivational drive . In several recent studies reporting sustained DA signals [5–8] , a common feature is that self-paced actions are required , as argued in [8] . We conjectured that this feature could be critical for the putative motivational functions of sustained DA signals . However , in our previous study [23] , such a feature was not incorporated: our previous model was extremely simple and assumed that the subject automatically moved to the next state at every time step . In the present work , we constructed a new model , which incorporated the requirement of self-paced approach towards a goal , represented as a series of ‘Go’ or ‘No-Go’ ( or ‘Stay’ ) selections , into RL with decay of learned values . Using this new model , we investigated: ( 1 ) if the model ( as well as the previous non-self-paced model ) generates both phasic and sustained RPE/DA signals so that their mechanisms can be coherently understood , ( 2 ) if the model demonstrates any association between sustained DA signals and motivation , and ( 3 ) if the model can mechanistically account for the key experimental findings that suggest DA's roles in motivation , specifically , the ( i ) slowdown of self-paced behavior by post-training blockade of DA signaling [14] , ( ii ) severe impairment of effortful actions to obtain rewards , but not of seeking of easily obtainable rewards , by DA blockade [11 , 24] , and ( iii ) relationships between the reward amount , the level of motivation reflected in the speed of behavior , and the average level of DA [7] . Through simulations and mathematical ( bifurcation ) analyses , we have successfully answered these questions .
We modeled a behavioral task requiring self-paced voluntary approach ( whether spatially or not ) towards a goal as a series of ‘Go’ or ‘Stay’ ( ‘No-Go’ ) selections as illustrated in Fig 1 . We then simulated subject's behavior by a temporal-difference ( TD ) RL model incorporating the decay of learned values ( referred to as the ‘value-decay’ below ) . Specifically , we assumed that at every time step the subject selects ‘Go’ or ‘Stay’ depending on their learned values , which are updated according to RPE ( TD error ) when the corresponding action is taken . In addition , we also assumed that the learned values of all the actions ( whether selected or not ) decay in time at a constant rate ( see the Materials and Methods for details ) . RPE at each time step was assumed to be represented by the level of DA at the time step , and the value decay was assumed to be implemented as a decay of plastic changes of synaptic strengths storing learned values . Fig 2A shows the number of time-steps needed for goal-reaching ( i . e . , from the start to the goal in a single trial; referred to as the ‘time needed for goal-reaching’ below ) averaged over 500 trials , with the rate of the value-decay ( referred to as the ‘decay rate’ below ) varied . As shown in the figure , the time needed for goal-reaching is minimized in the case with a certain degree of value-decay . In other words , introduction of the value-decay can facilitate fast goal-reaching . Fig 2B shows the trial-by-trial change of the time needed for goal-reaching . Without the value-decay ( Fig 2B , left ) , the subject initially learns to reach the goal quickly , but subsequently a significant slowdown occurs . In contrast , with the value-decay ( Fig 2B , middle and right ) , the time needed for goal-reaching is kept small , never showing slowdown . The observed facilitation of fast goal-reaching by introduction of the value-decay ( Fig 2A ) is thus accompanied with such a qualitative change in the long-term dynamics . In the same simulated task using the same model , we examined how post-training blockade of DA signaling affects the subject's speed ( i . e . , the time needed for goal-reaching ) , again varying the decay rate . Specifically , with the assumption that DA represents RPE , we simulated the post-training DA blockade by reducing the size of RPE-dependent increment of action values to zero ( complete blockade ) or to a quarter of the original size ( partial blockade ) after 250 trials were completed . Fig 2C shows the results . As shown in the left panels of Fig 2C , without the value-decay , DA blockade causes little effect on the subject's speed . In contrast , in the case with the value-decay ( Fig 2C , middle and right panels ) , the same DA blockade rapidly causes pronounced slowdown ( i . e . , increase in the time needed for goal-reaching ) . In order to explore mechanisms underlying the fast goal-reaching achieved with the value-decay and its impairment by DA blockade , we examined the action values of ‘Go’ and ‘Stay’ at each state . The black and gray lines in Fig 3A respectively show the action values of ‘Go’ and ‘Stay’ at the end of the 500th trial , and Fig 3B shows their trial-by-trial evolutions . Without the value-decay ( left panels of Fig 3A and 3B ) , all the action values are eventually almost saturated to the reward amount ( = 1 ) , so that there remains little difference between the action values of ‘Stay’ and ‘Go’ at any states . As a result , subject should choose ‘Stay’ as frequently as ‘Go’ . This explains the observed slowdown in the case without the value-decay ( Fig 2B , left panel ) . In contrast , with the value-decay ( Fig 3A and 3B , middle and right panels ) , the action values of ‘Go’ shape a sustained gradient from the start to the goal , while the actions values of ‘Stay’ remain relatively small . Why does the value-decay create such a gradient of ‘Go’ values ? Fig 3C shows examples of RPE generated during the task . In the case without the value-decay ( left panel ) , positive RPE is generated at the beginning of each trial , but RPE is mostly nearly zero in other epochs . This is what we usually expect from TD RL models after learning [4 , 25] . On the contrary , in the case with the value-decay ( Fig 3C , middle and right panels ) , RPE remains to be positive in most of the time , indicating that decrement of action values due to the value-decay is balanced with RPE-dependent increment . Such sustained positive RPE is then considered to create the start-to-goal gradient of ‘Go’ values . This is because RPE generated when taking ‘Go’ at state Si ( i = 1 , … , 6 ) is calculated ( see the Materials and Methods ) as RPE=γ⋅max{Q ( ‘Stay’ at Si+1 ) , Q ( ‘Go’ at Si+1 ) }−Q ( ‘Go’ at Si ) , ( γ: time discount factor , satisfying 0 ≤ γ ≤ 1 ) which is not greater than Q ( ‘Go’ at Si + 1 ) − Q ( ‘Go’ at Si ) provided Q ( ‘Stay’ ) ≤ Q ( ‘Go’ ) ( this would naturally be expected ) , and then "0 < RPE" ensures 0<Q ( ‘Go’ at Si+1 ) −Q ( ‘Go’ at Si ) ⇔Q ( ‘Go’ at Si ) <Q ( ‘Go’ at Si+1 ) , which indicates a gradient towards the goal . Looking at the pattern of RPE ( Fig 3C ) , in the case with a relatively larger value-decay , RPE exhibits a ramp towards the goal ( Fig 3C , right; notably , this decay rate does not achieve the fastest goal-reaching , but still realizes a faster goal-reaching than the case without value-decay: cf . Fig 2A ) . This resembles the experimentally observed ramp-like patterns of DA neuronal activity [21 , 22] or striatal DA concentration [5–8] as we have previously suggested using the non-self-paced model [23] . But with a milder value-decay , RPE peaks both at the start and towards the goal , with the former more prominent ( Fig 3C , middle ) . In this way , our model generates various patterns of RPE , from phasic to ramping , depending on the decay rate , or indeed the relative strength of the value-decay to the number of states . This could potentially be in line with the fact that the studies reporting DA ramping [5–8 , 21 , 22] used operant or navigation tasks in which several different states within a trial seem likely to be defined whereas the studies reporting clearly phasic DA response [1 , 3] used a simple classical conditioning task where a smaller number of states might be defined . It has been also found in other studies [5 , 8] that elevations in DA levels occurred earlier in later task sessions . According to our simulation results ( Fig 3C ) , such a change could potentially be explained in our model if the decay rate gradually decreases ( i . e . , from the right panel of Fig 3C to the middle panel ) . In our simulations , such a decrease in the decay rate is in the direction towards an optimal decay rate in terms of the time needed for goal-reaching averaged over 500 trials ( Fig 2A ) . This suggests that the experimentally observed changes in the DA response pattern across sessions [5 , 8] might be an indicative of meta-learning processes to adjust the decay rate to an optimal level . Despite these potentially successful explanations of the various DA response patterns , however , not all the patterns can be explained by our model . In particular , it has been shown that the DA concentration decreases during the reward delivery ( sucrose infusion for 6 sec ) [2] . Our model does not explain such a decrease of DA: to explain this , it would be necessary to extend the model to describe the actual process of reward delivery/consumption . The reason why the blockade of DA signaling causes slowdown in the cases with the value-decay but not in the cases without the value-decay in our model ( Fig 2C ) can also be understood by looking at RPE . Specifically , in the cases with the value-decay , positive RPE is continued to be generated at every state ( Fig 3C , middle and right ) , and each ‘Go’ value is kept around a certain value ( Fig 3B , middle and right ) because increment according to RPE and decrement due to the value-decay are balanced . Then , if DA signaling is blocked and the size of RPE-dependent increment is reduced , such a balance is perturbed and thereby ‘Go’ values decrease , resulting in the slowdown . In contrast , in the cases without the value-decay , sustained positive RPE is generated only at the beginning of each trial ( Fig 3C , left ) , and it does not increase the value of ‘Go’ taken later in the trial . Thus , after learning has settled down , ‘Go’ values are almost frozen , and therefore blockade of DA signaling has little impact on subject behavior . Fig 4 shows the trial-by-trial changes of the action values ( the top panels of Fig 4A and 4B ) and the action values at the end of the 500th trial ( the bottom panels ) in the simulations where the size of RPE-dependent increment of action values was reduced to zero ( A ) or to a quarter of the original size ( B ) after 250 trials were completed . As shown in these figures , the abovementioned conjectures about the effects of DA blockade on the action values were confirmed . Given that the action values are represented in the striatal neural activity , the parallel reduction in the action values and the speed for goal-reaching by DA blockade in our model can be broadly in line with a recent finding of the parallel impairment of the striatal neural representation of actions and the action vigor in DA-depleted mice [26] . Also , intriguingly , in the cases with the value-decay , after DA signaling is reduced to a quarter of the original ( Fig 4B , middle and right panels ) , whereas the values of ‘Go’ actions distant from the goal degrade quite prominently , the values of ‘Go’ actions near the goal ( i . e . , A12 and A10 ) remain relatively large , although they are also significantly decreased from the original values . This could potentially be in line with the experimental observations that DA blockade severely impairs costly or effortful actions to obtain rewards but seeking of easily obtainable rewards are largely spared [11 , 24] . In order to more directly address this issue , we simulated an experiment examining the effects of DA depletion in the nucleus accumbens in a cost-benefit decision making task in a T-maze reported in [24] . In one condition of the experiment , there was small reward in one of the two arms of the T-maze whereas there was large reward accompanied with a high cost ( physical barrier ) in the other arm . In the baseline period after training ( exploration ) of the maze , rats preferred the high-cost-high-return arm . However , DA depletion reversed the preference so that the rats switched to prefer the low-cost-low-return arm . DA depletion also increased the response latency ( opening of the start door at the end of the start arm ) , although the latency subsequently recovered . In another condition of the experiment , the two arms contained small and large rewards as before , but neither was accompanied with a high cost . In this condition , rats preferred the large-reward arm , and DA depletion did not reverse the preference . Meanwhile , DA depletion still increased the response latency , though the latency subsequently recovered as before . We simulated this experiment by representing a high cost as an extra state preceding the reward ( State 5 in Fig 5A , right ) . Fig 5B and 5C show the ratio of choosing the large-reward arm ( Arm 1 in Fig 5A ) and the average time needed for reaching the T-junction ( State 4 in Fig 5A , right ) , respectively , in the condition with a high cost in the large-reward arm ( Fig 5A ) . Fig 5F and 5G show the results in the condition without a high cost ( Fig 5E ) . As shown in these figures , the model successfully reproduces the experimental observations that DA depletion induced a preference reversal only in the condition with a high cost ( Fig 5B and 5F ) while increased the latency in both conditions ( Fig 5C and 5G ) , although the subsequent recovery of the latency is not reproduced . Looking at the action values in the case with a high-cost ( Fig 5D ) , the value of ‘Go’ to Arm 1 at the T-junction is fairly high before DA depletion . However , because this action is apart from reward , its value degrades quite prominently after DA depletion , becoming lower than the value of ‘Go’ to Arm 2 , which is adjacent to reward ( even though it is small reward ) . This explains the preference reversal ( Fig 5B ) . In contrast , in the case without a high-cost ( Fig 5H ) , the value of ‘Go’ to Arm 1 degrades only moderately after DA depletion , remaining higher than the value of ‘Go’ to Arm 2 . In the meantime , in both conditions , initially there are value-contrasts between ‘Go’ and ‘Stay’ at States 1–3 but they degrade after DA depletion , explaining the increase in the latency ( Fig 5C and 5G ) . As we have shown above , the value-decay creates a gradient of ‘Go’ values towards the goal . It is known that temporal discounting of rewards also makes a gradient of values ( c . f . , [7] ) . However , we assumed no temporal discounting ( i . e . , time discount factor γ = 1 ) in the above simulations and thus the value-gradient observed in the above was caused solely by the value-decay . In order to compare the effects of the value-decay and the effects of temporal discounting , we conducted simulations of the original unbranched self-paced task ( Fig 1 ) assuming no value-decay but instead temporal discounting ( time discount factor γ = 0 . 8 ) . Fig 6 shows the resulting action values ( Fig 6A and 6B ) , RPE ( Fig 6C ) , and the effect of DA blockade on the time needed for goal-reaching ( Fig 6D ) . As shown in Fig 6A and 6B , a value-gradient is shaped , as expected . Contrary to the case with the value-decay , however , sustained positive RPE is generated only at the beginning of each trial ( Fig 6C ) , and because of this , post-training blockade of DA signaling causes little effect on the subject speed ( Fig 6D ) . Comparing the value gradient caused by the value-decay ( Fig 3A and 3B , middle/right ) and the gradient caused by temporal discounting ( Fig 6A and 6B ) , the differences of the action values between ‘Stay’ and ‘Go’ are much larger in the case with the value-decay . This is considered to be because , in the case with the value-decay , the values of unchosen actions just decay whereas those of chosen actions are kept updated according to RPE . In order to mathematically confirm this conjecture , especially , the long-term stability of such a large contrast between ‘Stay’ and ‘Go’ values , we considered a reduced dynamical system model of our original model , focusing on the last state preceding the goal ( i . e . , S6 in Fig 1 ) , and conducted bifurcation analysis . Specifically , we derived a two-dimensional dynamical system that approximately describes the dynamics of the action values of A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) at S6 ( Fig 7A; see the Materials and Methods for details ) , and examined how the system's behavior qualitatively changes along with the change in the degree of the value-decay . Temporal discounting was not assumed ( i . e . , γ was assumed to be 1 ) in this reduced model so as to isolate the effect of the value-decay . Fig 7B is the resulting bifurcation diagram showing the equilibrium action values of A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) at S6 ( with approximations ) with the degree of the value-decay varied , and Fig 7C shows the probability of choosing A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) at the equilibrium point . As shown in Fig 7B , it was revealed that as the degree of the value-decay increases , qualitative changes occur twice ( in technical terms , arrangements of the nullclines shown in Fig 7E indicate that both of them are saddle-node bifurcations ( c . f . , [27] ) ) , and when the value-decay is larger than a critical degree ( ψ ≈ 0 . 0559 ) , there exists a unique stable equilibrium with a large contrast between the action values of A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) . It is therefore mathematically confirmed that the value-decay causes a large contrast between the steady-state action values of ‘Stay’ ( A11 ) and ‘Go’ ( A12 ) as conjectured in the above . Similar mechanism is considered to underlie the observed contrasts between ‘Stay’ and ‘Go’ values at the other states ( Fig 3A and 3B , middle/right ) . Notably , the bifurcation diagram ( Fig 7B ) suggests that there exists bistability when the degree of the value-decay is within a certain range . We conducted a simulation of the original model with the decay rate φ = 0 . 0045 , and found that there indeed appears a phenomenon indicative of bistability . Specifically , the value of ‘Stay’ ( A11 ) was shown to fluctuate between two levels in long time scales ( Fig 7D ) . Such bistability can potentially cause a hysteresis , in a way that learned values depend on the initial condition or the learning history , although the range of the degree of the value-decay for bistability is not large . Fig 8 shows the dependence of the bifurcation diagram on the RL parameters . As shown in the figure , the existence and the range of bistability critically depend on the inverse temperature ( β ) ( representing the sharpness of soft-max selection ) and the time discount factor ( γ ) . The figure also indicates , however , that whether bistability exists or not , as the degree of the value-decay increases , there emerges a prominent contrast between ‘Stay’ and ‘Go’ values . Importantly , it is considered that the facilitation of fast goal-reaching by the value-decay in the simulations shown so far is actually caused by the value-contrasts between ‘Stay’ and ‘Go’ rather than the gradient of ‘Go’ values explained before , because value-based choice is made between ‘Stay’ and ‘Go’ rather than between successive ‘Go’ actions . Nevertheless , the decay-induced value-gradient can indeed cause a facilitatory effect if selection of ‘Go’ or ‘Stay’ is based on the state values rather than the action values . Specifically , if our model is modified in the way that the probability of choosing ‘Go’ or ‘Stay’ depends on the value of the current and the next state ( while action values are not defined: see the Materials and Methods for details ) , introduction of the decay of learned ( state ) values can still cause facilitation of goal-reaching ( Fig 9A ) . Since the values of ‘Go’ and ‘Stay’ are not defined and thus the "value-contrast" appeared in the original model does not exist , this facilitation is considered to come from the gradient of state values ( Fig 9B ) . Facilitation appears to be in similar levels as the decay rate changes from 0 . 01 to 0 . 02 ( Fig 9A ) , and it is considered to be because , while the slope near the start becomes shallower , the slope near the goal becomes steeper ( Fig 9B ) . We examined how the effect of the value-decay on fast goal-reaching depends on the RL parameters , specifically , the learning rate , the inverse temperature , and the time discount factor . Fig 10A shows the time needed for goal-reaching averaged over 500 trials in conditions varying one of the RL parameters and the decay rate . As shown in the figure panels , although a large inverse temperature ( indicating an exploitative choice policy ) realizes fast goal-reaching without the value-decay ( middle panel of Fig 10A ) , facilitation of fast goal-reaching by introduction of the value-decay occurs within a wide range of RL parameters . Notably , the right panel of Fig 10A shows that the value-decay can realize faster goal-reaching than temporal discounting does , given that the other parameters are fixed to the values used here . This is considered to reflect that while both the value-decay and temporal discounting create a value-gradient from the start to the goal , only the value-decay additionally induces value-contrasts between ‘Stay’ and ‘Go’ as we have shown above . In the results presented so far , we assumed in the model that RPE is calculated according to a major RL algorithm called Q-learning [28] ( Eq ( 1 ) in the Materials and Methods ) , based on the empirical suggestions that DA neuronal activity in the rat ventral tegmental area ( VTA ) and DA concentration in the nucleus accumbens represent Q-learning-type RPE [21 , 29] . However , there is in fact also an empirical suggestion that DA neuronal activity represents RPE calculated according to another major RL algorithm called SARSA [30] ( Eq ( 2 ) in the Materials and Methods ) rather than Q-learning in the monkey substantia nigra pars compacta ( SNc ) [31 , 32] . It remains elusive whether such a difference comes from the differences in the species , regions , task paradigms or other conditions . We examined how the model's behavior changes if SARSA-type RPE is assumed instead of Q-learning type RPE . Fig 10B shows the time needed for goal-reaching averaged over 500 trials , with the RL parameters varied as before , and Fig 10C shows the learned values of each action at the end of 500 trials . As shown in the figures , it turned out that the effects of the value-decay , as well as the underlying value-gradient and value-contrast , are very similar to the cases with Q-learning type RPE . There is , however , a prominent difference between the cases of SARSA and Q-learning . Specifically , in the case of SARSA , RPE generated upon taking ‘Go’ was much larger than RPE generated upon taking ‘Stay’ ( Fig 10D , left ) , whereas there was no such difference in the case of Q-learning ( Fig 10D , right ) . The difference in RPE between ‘Go’ and ‘Stay’ in the SARSA case is considered to reflect the value-contrast between the learned values of ‘Go’ and ‘Stay’ ( Fig 10C ) . This is not the case with Q-learning because the Q-learning-type RPE calculation uses the value of the maximum-valued action candidates , which would be ‘Go’ in most cases , regardless of which action is actually selected . The SARSA-type RPE calculation , by contrast , uses the value of actually selected action ( compare Eqs ( 1 ) and ( 2 ) in the Materials and Methods ) . The difference in RPE between ‘Go’ and ‘Stay’ in the SARSA case could potentially be related to a recent finding [33] that DA in the rat nucleus accumbens responded to a reward-predicting cue when movement was initiated but not when animal had to stay . However , our present model would be too simple to accurately represent the task used in that study and the neural circuits that are involved , and elaboration of the model is desired in the future . We examined how the facilitatory effect of the value-decay depends on the amount of the reward obtained at the goal , which was fixed at r = 1 in the simulations so far presented ( we again consider Q-learning-type RPE in the following ) . Fig 11A , 11B , 11C and 11D show the time needed for goal-reaching averaged over 500 trials , with the RL parameters varied as before , in the cases with reward amount 0 . 5 , 0 . 75 , 1 . 25 , and 1 . 5 , respectively . As shown in the figures , the overall tendency of the effect of the value-decay does not largely change across this threefold range of reward amount . Meanwhile , the figures indicate that as the reward amount increases , the time needed for goal-reaching generally decreases , or in other words , the subject's speed increases . The black line in Fig 11E shows this relationship in the case with the standard RL parameters used so far and the decay rate of 0 . 01 . As shown in this figure , there is a clear negative relationship between the reward amount and the time needed for goal-reaching . We also examined how the average RPE per time-step during 500 trials depends on the reward amount . As shown in the black line in Fig 11F , we found that there is a positive relationship between the reward amount and the average RPE . These negative and positive reward-amount-dependences of the time needed for goal-reaching and the average RPE , respectively , are in line with the experimental findings [7] that the subject's latency and the minute-by-minute DA level in the nucleus accumbens were negatively and positively related with the reward rate , respectively , given that RPE in our model is represented by DA as we assumed . The commonality of the effect of the value-decay across the range of reward amount ( Fig 11A–11D ) and the positive reward-amount-dependence of the average RPE ( Fig 11F , black line ) are considered to appear because our model is largely scalable to ( i . e . , variables are scaled in proportion to ) the changes in the reward amount except for the effect of the inverse temperature . The negative reward-amount-dependence of the time needed for goal-reaching ( Fig 11E , black line ) is considered to appear because as the reward amount increases , the overall magnitudes of learned values , and thereby also the value-contrasts between ‘Stay’ and ‘Go’ , increase . The gray lines in Fig 11E and 11F show the relationship between the reward amount and the time needed for goal-reaching ( Fig 11E ) or the RPE per time-step ( Fig 11F ) in the case without the value-decay , averaged over 500 trials . The gray circles and crosses in these figures show the averages for 1–100 trials and 401–500 trials , respectively . As shown in these , in the case without the value-decay , there are negative and positive reward-amount-dependences of the time needed for goal-reaching and the RPE per time-step in the initial phase , but such dependences gradually degrade along with trials . This is considered to be because the values of ‘Stay’ actions gradually increase toward the saturation ( Fig 3B , left ) . In contrast , in the case with the value-decay ( φ = 0 . 01 ) , there are little differences in the time needed for goal-reaching and the RPE per time-step between 1–100 trials ( black circles in Fig 11E and 11F ) and 401–500 trials ( black crosses in Fig 11E and 11F ) . This is reasonable given that gradual saturation of ‘Stay’ values does not occur in the case with the value-decay ( Fig 3B , middle ) . We further examined how the facilitatory effect of the value-decay depends on the architectures of the model , in particular , the number of states and the number of action candidates . Regarding the number of states , in the results so far shown , we assumed seven states , including the start and the goal , as shown in Fig 1 . Fig 12A and 12B show the time needed for goal-reaching averaged over 500 trials in the cases with four or ten states , respectively . As shown in the figures , although the optimal decay rate that realizes fastest goal-reaching varies depending on the number of states , facilitation of fast goal-reaching by introduction of the value-decay can occur in either case . Regarding the number of the action candidates , we have so far assumed that either of the two actions , ‘Go’ or ‘Stay’ , can be taken at each state except for the goal ( or the T-junction in the case of the T-maze ) . This can be a good model of certain types of self-paced tasks that are intrinsically unidirectional , such as pressing a lever for a fixed amount of times to get reward . However , there are also self-paced tasks that are more like bidirectional , for instance , movements in an elongated space with reward given at one of the ends . Such tasks might be better represented by adding ‘Back’ action to the action candidates at each state except for the start and the goal . Fig 12C shows the time needed for goal-reaching averaged over 500 trials in the case where the ‘Back’ action was added . As shown in this figure , while the time needed for goal-reaching is generally larger than the cases without the ‘Back’ action as naturally expected , the value-decay can facilitate fast goal-reaching in this case too . It is also a question of how robust the effect of the value-decay is to perturbations in reward environments . In particular , given that the values of unchosen actions just decay , it is conceivable that , if small reward is given at a state between the start and the goal ( e . g . , S4: Fig 13A ) whenever subject is located there ( i . e . , repeatedly at every time step if subject stays at S4 ) , subject might learn to stay there persistently rather than to reach the goal . Denoting the size of the small reward by x ( < 1 , which is the amount of the reward given at the goal ) , if 7x < x + 1 ⇔ x < 0 . 166… , such a persistent stay is however inferior to the fastest repetition of goal-reaching in terms of the average reward obtained per time-step . We examined the behavior of modeled subject when small reward is given at S4 with its size x varied from 0 to 0 . 1 , in the case with the value-decay ( φ = 0 . 01 ) . Fig 13B shows the resulting percentage of simulation runs ( out of total 20 runs for each condition ) in which subject completed 500 trials within 35000 time steps ( i . e . , within 70 time steps per trial on average ) without settling at S4 . As shown in the figure , the percentage for the completion of 500 trials is 100% when the size of the reward at S4 is ≤ 0 . 04 , whereas the percentage then decreases as the size of the reward at S4 further increases . This indicates that a persistent stay at S4 actually occurs even if it is not advantageous: Fig 13C and 13D show such an example . Fig 13E shows the number of time steps needed for goal-reaching averaged over 500 trials , only for the simulation runs completing 500 trials in the cases where the completion rate is less than 100% . As shown in the figure , the speed of goal-reaching is kept fast , comparable to the case without reward at S4 ( i . e . , x = 0 ) . These results indicate that the facilitatory effect of the value-decay on fast goal-reaching has a certain degree of tolerance to this kind of perturbation in reward environments , although it eventually fails as the perturbation becomes larger . Nonetheless , when temporal discounting ( γ = 0 . 9 , 0 . 8 , … ) was also assumed in the model with the small reward x = 0 . 1 at S4 , persistent stay at S4 before completing 500 trials was not observed in 20 simulation runs for each of the tested decay rates , and the value-decay could have facilitatory effects ( Fig 13F ) . The absence of persistent stay at S4 is considered to be because the value of ‘Stay’ at S4 is bounded due to temporal discounting . For example , in the case with γ = 0 . 9 and no value-decay , if the subject keeps staying at S4 , the value of ‘Stay’ at S4 converges to 1 ( solution of the equation of V: 0 = 0 . 1 + 0 . 9V − V ) . This is still larger than the convergence value of ‘Go’ at S4 , which is 0 . 92 = 0 . 81 . However , since the growth of the ‘Stay’ value from the initial value 0 is likely to be slower than the growth of the ‘Go’ value , subject would rarely begin to settle at S4 . In contrast , in the case with no temporal discounting and no value-decay , if the subject keeps staying at S4 , the value of ‘Stay’ at S4 increases unboundedly , leading to a persistent stay . Actually , the value-decay also bounds the value of ‘Stay’ at S4 , but its effect is weak when the decay rate is small as we have so far assumed . For example , in the case with no temporal discounting , φ = 0 . 01 , and the learning rate α = 0 . 5 , if the subject keeps staying at S4 , the value of ‘Stay’ at S4 converges to 4 . 95 ( solution of the equation of V: V = ( 1 − 0 . 01 ) ( V + 0 . 5×0 . 1 ) ) , which is fairly large . In this way , temporal discounting effectively prevents the subject from settling at S4 . The value-decay can then facilitate fast goal-reaching by creating the value-contrast between ‘Go’ and ‘Stay’ . So far we have assumed that subject exists in one of the discrete set of states , and selects either ‘Go’ or ‘Stay’ , moving to the next state or staying at the same state . Given this simple structure , our model can potentially represent a variety of self-paced behavior , from spatial movement to more abstract Go/No-Go decision sequences . At the same time , however , our model is likely to be too simple to accurately model any specific behavior . In particular , in the case of spatial movement , subject does not really exist only in one of a small number of locations , and would not abruptly stop or literally ‘stay’ at a particular location . Meanwhile , subject should stop or slow down in the face of a physical constraint ( e . g . , the start , the junction , or the end of a maze ) or a salient event ( e . g . , reward ) as observed in experiments [6] . An emerging question is whether our model can be extended to reproduce these observations while preserving its main features . In order to examine this , we developed an elaborated model of self-paced spatial movement in the T-maze . In this model , the exact one-to-one correspondence between the subject's physical location and the internal state assumed in the original model was changed into a loose coupling , in which each state corresponds to a range of physical locations ( Fig 14A ) . Also , ‘Stay’ action in the original model was replaced with ‘Slow’ action unless there is a physical constraint ( i . e . , the start , the T-junction , or the end ) . By selecting ‘Slow’ , subject moves straightforward for a time step with the "velocity" halved from the previous time step ( or further decreased when there is a physical constraint ) . ‘Slow’ was introduced to eliminate the abrupt/complete stop appeared in the original model , and mechanistically , it can represent inertia in decision and/or motor processes [34 , 35] . With these modifications , state transitions can sometimes occur even when subject chooses ‘Slow’ rather than ‘Go’ ( Fig 14A , Case 2 ) , different from the original model . At the T-junction , subject was assumed to take ‘Go’ to either of the two arms or ‘Stay’ in the same manner as in the original model . At the reward location , subject was assumed to take the consummatory action for a time step ( indicated by the double-lined arrows in Fig 14B and 14F ) , and proceed to the end state . Using this elaborated model ( see the Materials and Methods for details ) , we simulated the T-maze cost-benefit decision making task with DA depletion [24] that was simulated by the original model before ( Fig 5 ) . Fig 14C and 14D show the simulation results about the ratio of choosing the large-reward arm ( Arm 1 ) and the average time needed for reaching the T-junction in the task conditions with high cost in the large-reward arm ( Fig 14B ) , respectively . Fig 14G and H show the results in the task conditions without high cost in the large-reward arm ( Fig 14F ) . As shown in the figures , the experimentally observed effects of DA depletion , i . e . , the severe impairment of high-cost-high-return choice but not low-cost-high-return choice ( Fig 14C and 14G ) and the slowdown in both conditions ( Fig 14D and 14H ) , can be reproduced by the elaborated model , as well as by the original model ( Fig 5 ) . Simultaneously , the elaborated model can also reproduce the velocity profiles observed in a ( different ) T-maze task [6] , specifically , the slowdown and stop at the T-junction and the end of the maze and the absence of complete stop in the other locations ( Fig 14E and 14I ) . This exemplifies the potential of our original model to be extended to accurately represent specific self-paced behavior .
The notion that DA represents RPE has been supported by electrophysiological [1 , 4] , FSCV [2 , 3 , 36] and neuroimaging [37–39] results . Recently , optogenetic manipulations of DA neurons causally demonstrated the DA's role in representing RPE [40 , 41] . On the other hand , pharmacological blockade of DA signaling has been shown to cause motivational impairments such as slowdown of behavior [14] . Crucially , such effects have been observed even when DA signaling was blocked after animals were well trained and RPE-based learning had presumably already been completed . These motivational effects have thus been difficult to explain by the notion that DA represents RPE , unless different function of DA was also assumed [42 , 43] . Given such situations , Niv and colleagues [15] proposed a hypothesis that while DA's phasic response encodes RPE , DA's tonic concentration represents the average reward rate per unit time . They argue that as the reward rate decreases , optimal action speed should also decrease because the opportunity cost for not acting becomes relatively smaller than the extra cost for quickly acting , explaining why DA blockade causes slowdown . Extending this hypothesis , Lloyd and Dayan [16] proposed that quasi-tonic DA represents the expected amount of time discount of the value of next state caused by postponing action to get to the next state . This can explain the experimentally observed ramping DA signals [5–8] as reflecting a gradient of state values created by temporal discounting ( as in our Fig 6A and 6B ) , also consistent with the arguments by [7] . These normative hypotheses , at the Marr's levels of computation and algorithm [44 , 45] , provide intriguing predictions that are desired to be experimentally tested . Meanwhile , it is also important to explore the Marr's level of implementation , namely , circuit/synaptic operations , which could potentially provide inspirations for the upper levels and vice versa [45] . The abovementioned normative hypotheses highlight essential issues at the circuit/synaptic level , including how the sustained DA signals are generated in the upstream and utilized in the downstream , how the selection of action timing is implemented , and how temporal discounting is implemented . In our model , sustained DA signals are assumed to represent RPE , and thus the upstream and downstream mechanisms of sustained DA signaling should be nothing more than the mechanisms of how RPE is calculated in the upstream of DA neurons and how RPE-dependent value-update occurs through DA-dependent synaptic plasticity . Both of these mechanisms for RPE have been extensively explored ( e . g . , [46 , 47] ) and have now become clarified [17–20] . Regarding the selection of action timing , we assumed that it consists of a series of selections of two actions , ‘Go’ and ‘Stay’ . We could thus assume general mechanisms of action selection , for which implementation has been explored [48–52] with empirical supports [50 , 53 , 54] , although this leaves an important issue regarding how time is represented . As for the implementation of temporal discounting , we will discuss it below , in relation to the value-decay that can be implemented as decay of the plastic changes of the synaptic strengths . There exists a different model that has also tried to give a bottom-up unified explanation of both the learning and motivation roles of DA , referring to circuit architectures of the basal ganglia [55] . However , although this model captures a wide range of phenomena , there are several potential issues or limitations . Firstly , this model assumes that phasic DA represents a simple form of RPE , called the Rescorla-Wagner prediction error [56] , which lacks the upcoming-value term . However , RL models of the DA system , including our present model , widely assume the more complex form of RPE called the temporal difference ( TD ) RPE or TD error [25] ( see [57] for detailed explanation ) because there is a wealth of empirical supports that DA signals represent TD-RPE [1 , 20 , 58] . Secondly , because this model assumes the Rescorla-Wagner , rather than TD- , RPE , this model cannot describe the learning of the values of a series of actions or states , nor the changes of RPE , within a trial . As a corollary to this , this model does not explain the experimentally observed sustained DA signals [5–8 , 21 , 22] . Lastly , this model assumes that the two major basal ganglia pathways , the direct and indirect pathways , are associated with positive and negative reinforcement , respectively . Although this assumption is based on several lines of empirical results , alternative possibilities [43 , 46 , 47 , 59 , 60] have also been proposed for the operations of these pathways . Decay , or forgetting , is apparently wasteful . However , recent work [61] has suggested that decay/forgetting is in fact necessary to maximize future rewards in dynamic environments . Even in a static environment , potential benefit of decay/forgetting has been pointed out [62] . There is also a study [63] that considered decay to explain features of extinction . Forgetting for capturing extinction effects was also assumed in the model that we have discussed right above [55] . However , the authors clearly mentioned that they "assumed some forgetting" "to capture overall extinction effects" and "none of the results are qualitatively dependent on" the parameter for forgetting . Therefore , their work should not have anything to do with the effects of forgetting explored in our present work . Along with these theoretical/modeling works , it has been suggested that RL models with decay could fit the experimental data of human [64–66] , monkey [67] , and rat [68] choice behavior potentially better than models without decay . Moreover , existence and benefits of decay/forgetting have also been suggested in other types of learning [69 , 70] . Nonetheless , decay of learned values ( value-decay ) is not usually considered in RL model-based accounts of the functions of DA and cortico-basal ganglia circuits . RL models typically have the time discount factor and the inverse temperature ( representing choice sharpness ) as major parameters [25] . Temporal discounting generates a value-gradient ( Fig 6A and 6B ) [7 , 16] , and is suggested [71] to ensure that maximizing rewards simultaneously minimizes deviations from physiologically desirable states . Gradually increasing the inverse temperature , i . e . , choice sharpness , is known to be good for global optimization [72] . Possible neural implementation of these parameters have been explored [46 , 73–75] . However , it is not sure whether these parameters are actually biologically implemented in their original forms . We have shown that the value-decay can generate a value-gradient , and also value-contrasts which lead to a sharp choice of ‘Go’ . Choice-sharpening effect of decay is implied also in previous studies [62 , 66] . These indicate a possibility that the value-decay , or its presumed biological substrate , synaptic decay , might in effect partially implement the parameters for temporal discounting and inverse temperature . In this sense , the suggestions that sustained DA represents/reflects time-discounted state values [7 , 16] and our value-decay-based account are not necessarily mutually exclusive . Apart from temporal discounting and the inverse temperature , there is an additional note . There have been suggestions [34 , 35] that animal's and human's decision making can be affected by the subject's own choice history , which is not included in standard RL models . The value-decay assumed in our model is expected to cause a dependency of decision making on choice history . Whether it can ( partly ) explain experimentally observed choice patterns would be an interesting issue to explore . If the rate of the value-decay is always constant , after subject interrupts performing the task for a long period , learned values eventually diminish almost completely . Therefore , in order for our model to be valid , some sort of context-dependence of the value-decay needs to be assumed . There are several empirical implications . At the synaptic level , conditional synaptic decay depending on NMDA receptor-channels [76] or DA ( in drosophila ) [77] has been found . Behaviorally , memory decay was found to be highly context-dependent in motor learning [78] . More generally , it is widely observed that reactivation of consolidated memories makes them transiently labile [79] . With these in mind , we assume that the value-decay occurs when and only when subject is actively engaged in the relevant task/behavior . However , this issue awaits future verification . There is also an important limitation of our present model regarding the explanatory power for the experimental observations . Specifically , as mentioned before , our model explains the increase in the latency caused by DA depletion in the cost-benefit decision making task in a T-maze [24] , but does not explain the subsequent recovery of the latency . This recovery could possibly be explained if some slow compensatory mechanisms are additionally assumed in the model . It is important in future work to elaborate the model to account for this issue , as well as a diverse array of experimental observations on the DA's roles in motivation that are not dealt with in the present work . There are also many open issues in the model , both the functional ones and the structural ones . The functional issues include how the states and the time are represented [80 , 81] and how ‘Go’ and ‘Stay’ ( or ‘No-Go’ or ‘Slow’ ) are represented . As for the latter , while ‘Go’ and ‘Stay’ might be represented as two distinct actions , ‘Stay’ could instead be represented as disengagement of working-memory/attention as proposed in a recent work [82] . The structural issues include , among others , how different parts of the cortico-basal ganglia circuits and different subpopulations of DA neurons cooperate or divide labor [83–90] . Regarding this , a recent study [91] has shown that DA axons conveying motor signals are largely different from those conveying reward signals and that the motor and reward signals are dominant in the dorsal and ventral striatum , respectively . DA in our model is assumed to represent RPE , and it should thus be released from the axons conveying reward signals that are dense in the ventral striatum . Even with this specification , the structure of our model is still quite simple , and exploring whether and to what extent the present results can be extended to models with rich dynamics at the levels of circuits ( in the cortex [48 , 50 , 92–96] , the striatum [97–103] , the DAergic nuclei [104] , and the entire cortico-basal ganglia system [49 , 51 , 105–114] ) , neurons [115 , 116] , and synapses [117–120] would be important future work . Our model provides predictions that can be tested by various methods . First , if sustained DA signals indeed represent value-decay-induced sustained RPE , rather than being caused by other reasons [16 , 121] , the rate of the value-decay estimated from fitting of measured DA signals by our model should match the decay-rate estimated behaviorally . Behavioral estimation of decay-rate would be possible by preparing two choice options that are initially indifferent , manipulating the frequencies of their presentations , and then examining whether , and to what degree , less-frequently-presented option will be chosen less frequently . On the other hand , if sustained DA signals represent time-discounted state values [7 , 16] , time discount factor estimated from model-fitting of measured DA signals is expected to match behavioral estimation , e . g . , from intertemporal choices . Note , however , that the value-decay and temporal discounting might not be completely distinct entities; the value-decay could be a partial implementation of temporal discounting ( and the inverse temperature ) as we discussed before . Second , our model predicts that the strengths of cortico-striatal synapses are subject to decay in a context-dependent manner . This could be tested by measuring structural plasticity [18] during learning tasks ( across several sessions and intervals ) . Our model further predicts that manipulations of synaptic decay affect DA dynamics and behavior in specific ways . It has been indicated that a protein kinase that is constitutively active , protein kinase Mζ ( PKMζ ) , is necessary for maintaining various kinds of memories , including drug reward memory in the nucleus accumbens [122] . Specifically , inhibition of PKMζ in the nucleus accumbens core by injecting a selective peptide inhibitor has been shown to impair long-term drug reward memory [122] . It has also been shown that overexpression of PKMζ in the neocortex enhances long-term memory [123] . We predict that overexpression of PKMζ in the nucleus accumbens ( ventral striatum ) enhances reward memory , or in other words , reduces the value-decay , and thereby diminishes sustained DA signals and impairs goal-approach through the mechanisms described in the present work . Apart from PKMζ , it has also been indicated that DA is required for transforming the early phase of long-term potentiation ( LTP ) , which generally declines , into the late phase of LTP in the hippocampus [124 , 125] . Similar DAergic regulation of the stability of LTP could potentially exist in the striatum that is the target of the present work , and if so , the decay rate could be manipulated by DA receptor agonists or antagonists . In the striatal synapses , however , DA signaling would be required for the induction of potentiation before its maintenance , as we have actually assumed in our model . Therefore , it would be necessary to explore ways to specifically manipulate maintenance ( decay rate ) of potentiation . The results of the present study suggest that when biological systems for value-learning are active ( i . e . , when subject is actively engaged in the relevant task/behavior ) even though learning has apparently converged , the systems might be in a state of dynamic , rather than static , equilibrium where decay and update are balanced . As we have shown , such dynamic operation can potentially facilitate self-paced goal-reaching behavior , and this effect could be seen as a simple biologically plausible , though partial , implementation of temporal discounting and simulated annealing . It is also tempting to speculate that value-decay-induced sustained RPE might be subjectively felt as sustained motivation , considering recently suggested relationship between RPE and subjective happiness [126 , 127] . This is in accordance with the suggestion that DA signals subjective reward value [128 , 129] , or more precisely , "utility prediction error" [130] . Despite that dynamic operation has these potential advantages , however , there can also be disadvantages . Specifically , continual decay and update of values must be costly , especially given that DA signaling is highly energy-consuming [131] . This could potentially be related to neuropsychiatric and neurological disorders , in particular , Parkinson's disease [131 , 132] , which is characterized by motor and motivational impairments that are suggested to be independently associated with DA [133] . Better understanding of the dynamic nature of biological value-learning systems will hopefully contribute to clinical strategies against these diseases .
We posited that behavioral task requiring self-paced voluntary approach ( whether spatially or not ) towards a goal can be represented as a series of ‘Go’ or ‘Stay’ ( ‘No-Go’ ) selections as illustrated in Fig 1 . Discrete states ( S1 ~ S7 ) and time steps were assumed . In each trial , subject starts from S1 . At each time step , subject can take one of two actions , specifically , ‘Go’: moving to the next state or ‘Stay’: staying at the same state . Subject was assumed to learn the value of each action ( ‘Go’ or ‘Stay’ ) by a temporal-difference ( TD ) reinforcement learning ( RL ) algorithm incorporating the decay of learned values ( referred to as the ‘value-decay’ below ) [23] , and select an action based on their learned values in a soft-max manner [134] . Specifically , at each time step ( t ) , TD reward prediction error ( RPE ) δ ( t ) was assumed to be calculated according to the algorithm called Q-learning [28] , which has been suggested to be implemented in the cortico-basal ganglia circuit [21 , 43 , 59] , as follows: δ ( t ) =R ( S ( t ) ) +γmaxAcand ( t ) {Q ( Acand ( t ) ) }−Q ( A ( t−1 ) ) , ( 1 ) where S ( t ) represents the state where subject exists at time step t . R ( S ( t ) ) represents reward obtained at S ( t ) , which is r ( > 0 ) when S ( t ) = S7 ( goal ) and 0 at the other states , unless otherwise described . "Q ( A ) " generally represents the learned value of action A . Acand ( t ) represents the candidate of action that can be taken at time step t: when S ( t ) = Si ( i = 1 , 2 , … , 6 ) , Acand ( t ) = A2i−1 ( ‘Stay’ ) or A2i ( ‘Go’ ) ; when S ( t ) = S7 ( goal ) , candidate of action was not defined and the term γmaxAcand ( t ) {Q ( Acand ( t ) ) } was replaced with 0 . A ( t − 1 ) represents the action taken at time step t − 1; at the beginning of each trial , A ( t − 1 ) was not defined and the term Q ( A ( t − 1 ) ) was replaced with 0 so as to represent that the beginning of trial is not predictable . γ is the time discount factor ( 0 ≤ γ ≤ 1 ) . In a separate set of simulations ( Fig 10B , 10C and 10D , left ) , we also examined the case in which TD-RPE is calculated according to another RL algorithm called SARSA [30] as follows: δ ( t ) =R ( S ( t ) ) +γQ ( A ( t ) ) −Q ( A ( t−1 ) ) , ( 2 ) where A ( t ) represents the action taken at time step t . At each time step other than the beginning of a trial , the learned value of A ( t − 1 ) was assumed to be updated as follows: Q ( A ( t−1 ) ) new=Q ( A ( t−1 ) ) old+αδ ( t ) , ( 3 ) where α is the learning rate ( 0 ≤ α ≤ 1 ) . It was further assumed that the learned value of arbitrary action A decays at every time step as follows: Q ( A ) new= ( 1−φ ) Q ( A ) old , ( 4 ) where φ ( 0 ≤ φ ≤ 1 ) is a parameter referred to as the decay rate: φ = 0 corresponds to the case without value-decay . This sort of value-decay was introduced in [43] to account for the ramp-like activity of DA neurons reported in [21] , and was analyzed in [23] . In the present study , the decay rate φ was varied from 0 to 0 . 02 by 0 . 002 , unless otherwise described . Note that because ( 1 − φ ) is multiplied at every time step , even if φ is very close to 0 , significant decay can occur during a trial . For example , when the decay rate φ is 0 . 01 , the action values decline to at least ( 1–0 . 01 ) 7 ( ≈ 0 . 932 ) -fold of the original values during a trial . It should also be noted that the value-decay defined as above is fundamentally different from the decay of eligibility trace , which is a popular notion in the RL theory [25]: in terms of the eligibility trace , we assumed that only the value of the immediately preceding action ( Q ( A ( t − 1 ) ) ) is eligible for RPE-dependent update ( Eq ( 3 ) ) , corresponding to the TD ( 0 ) algorithm . At each time step other than when the goal was reached , action ‘Go’ or ‘Stay’ was assumed to be selected according to the following probabilities: P ( AGo ) =exp ( βQ ( AGo ) ) exp ( βQ ( AGo ) ) +exp ( βQ ( AStay ) ) ( 5 ) P ( AStay ) =exp ( βQ ( AStay ) ) exp ( βQ ( AGo ) ) +exp ( βQ ( AStay ) ) , ( 6 ) where β is a parameter called the inverse temperature , which represents the sharpness of the soft-max selection [134] . A trial ended when subject reached the goal and got the reward . Subsequently the subject was assumed to be ( automatically ) returned to the start ( S1 ) , and the next trial began . The learning rate α , the inverse temperature β , and the time discount factor γ were set to α = 0 . 5 , β = 5 , and γ = 1 unless otherwise described . Initial values of all the action values were set to 0 . The amount of reward obtained at the goal , r , was set to 1 in most simulations and analyses , but we also examined the cases with r = 0 . 5 , 0 . 75 , 1 . 25 , or 1 . 5 ( Fig 11 ) . The magnitude of rewards can in reality vary even more drastically . However , it has been shown [135] that the gain of DA neuron's response adaptively changes according to actual reward sizes . It could thus be possible to assume that r does not vary too drastically by virtue of such adaptive mechanisms . In a separate set of simulations ( Fig 13 ) , in order to examine the robustness of the effect of the value-decay to perturbations in reward environments , we assumed that there is also small reward , with size x , at S4 , which is given whenever subject is located at S4 ( i . e . , repeatedly at every time step if subject stays at S4 ) . In order to examine the dependence of the effect of the value-decay on the number of states from the start to the goal , we also conducted simulations for models that were modified to have 4 or 10 states , including the start and the goal , instead of 7 states in the original model ( Fig 12A and 12B ) . We also examined the case where the subject is allowed to take not only ‘Go’ or ‘Stay’ but also ‘Back’ action at Si ( i = 2 , 3 , … , 6 ) ( for this , we again assumed 7 states ) , which causes a backward transition to Si−1 . In this case ( Fig 12C ) , selection of ‘Go’ , ‘Stay’ , and ‘Back’ at Si ( i = 2 , 3 , … , 6 ) was assumed to be according to the probabilities: P ( A* ) = exp ( βQ ( A* ) ) /Sum , where A* was either AGo , AStay , or ABack , and Sum was exp ( βQ ( AGo ) ) + exp ( βQ ( AStay ) ) + exp ( βQ ( ABack ) ) . Initial values of all the action values , including the values of ‘Back’ actions , were set to 0 . Further , in a separate set of simulations ( Fig 9 ) , we considered a different model in which selection of ‘Go’ or ‘Stay’ is based on the state values rather than the action values ( ‘Back’ was not considered in this model ) . Specifically , in this model , RPE is calculated as: δ ( t ) =R ( S ( t ) ) +γV ( S ( t+1 ) ) −V ( S ( t ) ) , ( 7 ) where V ( S ( t ) ) represents the state value of S ( t ) ; if S ( t ) = S7 , V ( S ( t + 1 ) ) is assumed to be 0 . The state values are updated as follows: V ( S ( t ) ) new=V ( S ( t ) ) old+αδ ( t ) . ( 8 ) The learned value of arbitrary state S was assumed to decay at every time step as follows: V ( S ) new= ( 1−φ ) V ( S ) old . ( 9 ) ‘Go’ is selected at Si ( i = 2 , 3 , … , 6 ) with the probability exp ( βV ( Si+1 ) ) /{exp ( βV ( Si ) ) + exp ( βV ( Si+1 ) ) } , and ‘Stay’ is selected otherwise . The parameters were set to α = 0 . 5 , β = 5 , γ = 1 , and φ = 0 . 01 , and initial values of all the state values were set to 0 . For each condition with different parameter values or model architectures , 20 simulations of 500 trials with different series of pseudorandom numbers were performed , unless otherwise described . The particular number 500 was chosen because it was considered to be largely in the range of the number of trials used in experiments: e . g . , in [6] , rats completed ~15 or more sessions with each session containing 40 trials . 20 simulations could be interpreted to represent 20 subjects . In the figures showing the number of time steps needed for goal-reaching , we presented the mean ± standard error ( SE ) of the 20 simulations except for Fig 13E , where the mean ± SE for the simulation runs completing 500 trials ( which could be less than 20 for several conditions ) were presented . We also presented the theoretical minimum ( in the model with 7 states , it is 7 , including the steps at the start and the goal ) and the chance level , which is calculated ( in the model with 7 states ) as: 7+{1⋅h ( 6 , 1 ) ⋅12+2⋅h ( 6 , 2 ) ⋅ ( 12 ) 2+3⋅h ( 6 , 3 ) ⋅ ( 12 ) 3+⋯}⋅ ( 12 ) 6=13 , ( 10 ) where h ( 6 , k ) represents the number of ways for a repeated ( overlapping ) combination of k out of 6 and is calculated as h ( 6 , k ) = ( k + 5 ) ! ( k ! · 5 ! ) . Simulations were performed using MATLAB ( MathWorks Inc . ) . Program files to run simulations and make figures are available from ModelDB ( https://senselab . med . yale . edu/modeldb/showModel . cshtml ? model=195890 ) after the publication of this article . To simulate post-training blockade of DA signaling , we replaced δ ( t ) in Eq ( 3 ) with 0 ( complete blockade ) or δ ( t ) /4 ( partial blockade ) after 250 trials ( Figs 2C , 4 and 6D ) or 500 trials ( Figs 5 and 14 ) were completed . δ ( t ) was non-negative in those simulations because of the structure of the simulated tasks and the assumed Q-leaning-type calculation of RPE , and so the replacement of δ ( t ) with 0 or δ ( t ) /4 corresponded to that the size of an increment of action values according to non-negative RPE was reduced to zero or to a quarter of the original size . Notably , at the cellular/synaptic level , DA is known to have two major functions: ( i ) induce/modulate plasticity of corticostriatal synapses , and ( ii ) modulate responsiveness of striatal neurons [136] . Function ( i ) has been suggested to implement RPE-dependent update of learned values ( Eq ( 3 ) ) ( e . g . , [18] ) , and in the present work we incorporated the effect of DA blockade on this function into the model as described above , although function ( ii ) can also affect reaction time and valuation ( e . g . , [43] ) and assuming both of ( i ) and ( ii ) might be necessary to account for a wider range of phenomena caused by DA manipulations , in particular , changes in the speed or response time of a single rapid movement ( e . g . , [137 , 138] ) rather than ( or in addition to ) of a series of actions . In order to obtain qualitative understandings of how the value-decay affects the time evolution and steady-state of action values , beyond observations of simulation results , we reduced the original model ( Fig 1 ) to a simpler model through approximations , and conducted bifurcation analysis . Specifically , we considered a reduced continuous-time dynamical system model that approximately describes the time evolution of the values of ‘Stay’ and ‘Go’ at the state preceding the goal ( i . e . , A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) at S6 in Fig 1 ) . The reduced model is as follows: dq ( A11 ) dt=yαδ˜A11−ψq ( A11 ) ( 11 ) dq ( A12 ) dt=αδ˜A12−ψq ( A12 ) , ( 12 ) where q ( A11 ) and q ( A12 ) are the continuous-time variables that approximately represent the action values of A11 ( ‘Stay’ ) and A12 ( ‘Go’ ) , respectively . y approximately represents the expected value of the number of repetitions of A11 ( ‘Stay’ ) choice ( i . e . , how many time steps subject chooses A11 ( ‘Stay’ ) at S6 ) in a single trial , and it is calculated as: y=1⋅p ( A11 ) ⋅ ( 1−p ( A11 ) ) +2⋅p ( A11 ) 2⋅ ( 1−p ( A11 ) ) +⋯=p ( A11 ) 1−p ( A11 ) , ( 13 ) where p ( A11 ) represents the probability that A11 is chosen out of A11 and A12 according to Eq ( 6 ) when the values of A11 and A12 are q ( A11 ) and q ( A12 ) , respectively: p ( A11 ) =exp ( βq ( A11 ) ) exp ( βq ( A11 ) ) +exp ( βq ( A12 ) ) , ( 14 ) and substituting Eq ( 14 ) into Eq ( 13 ) results in: y=exp ( β ( q ( A11 ) −q ( A12 ) ) ) . ( 15 ) δ˜A11 and δ˜A12 represent TD-RPE generated when A11 or A12 with the value q ( A11 ) or q ( A12 ) is chosen , respectively: δ˜A11=γmax{q ( A11 ) , q ( A12 ) }−q ( A11 ) ( 16 ) δ˜A12=r−q ( A12 ) , ( 17 ) where r is the reward amount ( = 1 ) . ψ is a parameter representing the degree of the value-decay in a trial , which roughly corresponds to the decay rate φ in the original model multiplied by the number of time steps needed for goal-reaching . Notably , the reduced model is a continuous-time approximation of an algorithm in which update and decay of learned values occur once per every trial in a batch-wise manner whereas the original model is described as an online algorithm where update and value-decay occur at every time step; this difference is contained in our expression "approximate" referring to the reduced model . We analyzed the two-dimensional dynamics of q ( A11 ) and q ( A12 ) ( Eqs ( 11 ) and ( 12 ) ) under the assumption that q ( A11 ) ≤ q ( A12 ) ( i . e . , max{q ( A11 ) , q ( A12 ) } = q ( A12 ) in Eq ( 16 ) ) . More specifically , we numerically solved the equations dq ( A11 ) dt=0 and dq ( A12 ) dt=0 to draw the nullclines ( Fig 7E ) , and also numerically found the equilibriums and examined their stabilities to draw the bifurcation diagram ( Fig 7B ) and calculate p ( A11 ) and p ( A12 ) ( Fig 7C ) by using MATLAB . The result of the bifurcation analysis in the case with α = 0 . 5 , β = 5 , and γ = 1 ( Fig 7B ) was further confirmed by using XPP-Aut ( http://www . math . pitt . edu/~bard/xpp/xpp . html ) . We simulated an experiment examining the effects of DA depletion in the nucleus accumbens in a T-maze task reported in [24] . There were two conditions in the task . In the first condition , there was small reward in one of the two arms of the T-maze whereas there was large reward accompanied with high cost ( physical barrier preceding the reward ) in the other arm . In the second condition , the two arms contained small and large rewards as before , but neither was accompanied with high cost . We simulated this experiment by representing the high cost as an extra state preceding the reward . Specifically , we assumed a state-action diagram as shown in Fig 5A and 5E ( right panels ) . There were two action candidates , ‘Go’ and ‘Stay’ , at every state , except for the state at the T-junction ( State 4 ) and the state at the trial end , which was reached if ‘Go’ was chosen at State 7 or 8 . In State 4 , there were three action candidates , ‘Choose , and Go to , one of the arm ( Arm 1 ) ’ , ‘Choose , and Go to , the other arm ( Arm 2 ) ’ , and ‘Stay’ . In the state at the trial end ( State 9 , which is not depicted in Fig 5A and 5E ) , there was no action candidate , and subject was assumed to be automatically moved to the start state ( State 1 ) at the next time step . In the first condition of the simulated task ( Fig 5A ) , small reward ( size 0 . 5 ) was given when subject reached State 6 for the first time ( i . e . , only once in a trial ) , whereas large reward ( size 1 ) was given when subject reached State 7 for the first time . One extra state , i . e . , State 5 , preceding the state associated with large reward ( State 7 ) was assumed to represent high cost accompanied with the large reward . In the second condition ( Fig 5E ) , small ( size 0 . 5 ) or large ( size 1 ) reward was given when subject reached State 6 or State 5 , respectively , for the first time , representing that neither reward was accompanied with high cost . Calculation of Q-learning-type RPE and RPE-dependent update of action values were assumed in the same manner as before , with the parameters α = 0 . 5 , β = 5 , and γ = 1 . The value-decay was also assumed similarly , with the decay rate φ = 0 . 01 . Initial values of all the action values were set to 0 . 20 simulations of 1000 trials were conducted for each condition , and post-training DA depletion was simulated in such a way that the size of RPE-dependent increment of action values was reduced to a quarter of the original size after 500 trials were completed . By modifying the original model described above , we developed an elaborated model of self-paced spatial movement , and simulated the cost-benefit decision making task in a T-maze mentioned above . In this elaborated model , the exact one-to-one correspondence between the subject's physical location and the internal state assumed in the original model was changed into a loose coupling , in which each state corresponds to a range of physical locations ( Fig 14A ) . Also , ‘Stay’ action in the original model was replaced with ‘Slow’ action unless there is a physical constraint ( i . e . , the start , the T-junction , or the end ) . Specifically , it was assumed that , at each time step t , subject at a given location chooses either ‘Go’ or ‘Slow’ , except that the subject is at the start , T-junction , or the reward location ( in the ends of the T-maze ) . By selecting ‘Go’ , subject moves straightforward for a time step with the "velocity" 1 , meaning that the subject's physical location is displaced by 1 , or moves to the T-junction or the reward location when it is within 1 from the current location . By selecting ‘Slow’ , subject moves straightforward for a time step with the "velocity" halved , meaning that the subject's physical location is displaced by the half of the displacement during the previous time interval ( between t − 1 and t ) , or moves to the T-junction or the reward location when it is within the calculated displacement from the current location . In these ways , the "velocity" in this model was defined as the displacement in a time step . At the start ( State 1 ) , subject was assumed to take ‘Go’ or ‘Stay’ as in the original model ( because at the start , the previous "velocity" was not defined ) . At the T-junction , subject was assumed to take ‘Choose , and Go to , one of the arm ( Arm 1 ) ’ , ‘Choose , and Go to , the other arm ( Arm 2 ) ’ , or ‘Stay’ . By selecting ‘Choose , and Go to , Arm 1 or 2’ , the subject's physical location is displaced by 1 on the selected arm . By selecting ‘Stay’ , subject stays at the same place ( T-junction ) . At the reward location , subject was assumed to take the consummatory action for a time step ( indicated by the double-lined arrows in Fig 14B and 14F ) , and proceed to the end state . Calculation of Q-learning-type TD-RPE , update of action values , and the value-decay were assumed in the same manner as in the original model .
|
Dopamine ( DA ) has been suggested to have two reward-related roles: ( 1 ) representing reward-prediction-error ( RPE ) , and ( 2 ) providing motivational drive . Role ( 1 ) is based on the physiological results that DA responds to unpredicted but not predicted reward , whereas role ( 2 ) is supported by the pharmacological results that blockade of DA signaling causes motivational impairments such as slowdown of self-paced behavior . So far , these two roles are considered to be played by two different temporal patterns of DA signals: role ( 1 ) by phasic signals and role ( 2 ) by tonic/sustained signals . However , recent studies have found sustained DA signals with features indicative of both roles ( 1 ) and ( 2 ) , complicating this picture . Meanwhile , whereas synaptic/circuit mechanisms for role ( 1 ) , i . e . , how RPE is calculated in the upstream of DA neurons and how RPE-dependent update of learned-values occurs through DA-dependent synaptic plasticity , have now become clarified , mechanisms for role ( 2 ) remain unclear . In this work , we modeled self-paced behavior by a series of ‘Go’ or ‘No-Go’ selections in the framework of reinforcement-learning assuming DA's role ( 1 ) , and demonstrated that incorporation of decay/forgetting of learned-values , which is presumably implemented as decay of synaptic strengths storing learned-values , provides a potential unified mechanistic account for the DA's two roles , together with its various temporal patterns .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"learning",
"medicine",
"and",
"health",
"sciences",
"decision",
"making",
"nucleus",
"accumbens",
"applied",
"mathematics",
"brain",
"social",
"sciences",
"neuroscience",
"learning",
"and",
"memory",
"simulation",
"and",
"modeling",
"algorithms",
"synaptic",
"plasticity",
"cognitive",
"psychology",
"mathematics",
"cognition",
"research",
"and",
"analysis",
"methods",
"developmental",
"neuroscience",
"animal",
"cells",
"behavior",
"psychology",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"motivation",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"cognitive",
"science"
] |
2016
|
Forgetting in Reinforcement Learning Links Sustained Dopamine Signals to Motivation
|
Thelazia callipaeda is the causative agent of thelaziasis in canids , felids and humans . However , the population genetic structure regarding this parasite remains unclear . In this study , we first explored the genetic variation of 32 T . callipaeda clinical isolates using the following multi-molecular markers: cox1 , cytb , 12S rDNA , ITS1 and 18S rDNA . The isolates were collected from 13 patients from 11 geographical locations in China . Next , the population structure of T . callipaeda from Europe and other Asian countries was analyzed using the cox1 sequences collected during this study and from the GenBank database . In general , the Chinese clinical isolates of T . callipaeda expressed high genetic diversity . Based on the cox1 gene , a total of 21 haplotypes were identified . One only circulated in European countries ( Hap1 ) , while the other 20 haplotypes were dispersed in Korea , Japan and China . There were five nucleotide positions in the cox1 sequences that were confirmed as invariable among individuals from Europe and Asia , but the sequences were distinct between these two regions . Population differences between Europe and Asian countries were greater than those among China , Korea and Japan . The T . callipaeda populations from Europe and Asia should be divided into two separate sub-populations . These two groups started to diverge during the middle Pleistocene . Neutrality tests , mismatch distribution and Bayesian skyline plot ( BSP ) analysis all rejected possible population expansion of T . callipaeda . The Asian population of T . callipaeda has a high level of genetic diversity , but further studies should be performed to explore the biology , ecology and epidemiology of T . callipaeda .
The spirurid nematode Thelazia callipaeda ( Spirurida: Thelaziidae ) is the major etiological agent of ocular thelaziosis [1] . This worm can parasitize the conjunctival sac of domestic and wild carnivores and humans , causing conjunctivitis , lacrimation and itchiness , and even blindness [2 , 3] . The parasite is transmitted by a drosophilid insect of the genus Phortica ( Diptera , Drosophilidae ) , that feeds on the lachrymal secretions of mammals [4 , 5] . T . callipaeda was previously known as the oriental eyeworm because of its original description in eastern Asia countries [1] . Since the first cases of canine thelaziosis reported in Italy in 1988 , the nematode has spread through many southern , central , western and eastern European countries [6] . A broad spectrum of wild carnivores , such as wolves , wildcats , red foxes , badgers , beech martens and even brown hares , plays an important role in maintaining and spreading eyeworm infections to domestic animals and humans [7] . The first human case was described in Beijing , China in 1917 [8] . Although sporadic human thelaziosis has been reported in several European countries , human infections are mainly documented in people living in China , Japan , Korea and India [9] . China is most likely to have the largest number of cases of thelaziasis in the world with more than 600 cases reported to date [10] . Recently , an increase in T . callipaeda infections has been reported in animals and humans living in European countries and in China [6 , 11] . Consequently , T . callipaeda poses a serious threat to public health and thelaziosis has even been termed as an emerging enzootic disease [1 , 6 , 9] . Understanding the host specificity , transmission pattern and population genetic characteristics of the parasite are valuable for the prevention and control of thelaziasis in animals and humans [12] . However , our knowledge regarding these issues is still fragmented , and insufficient studies on population genetics of T . callipaeda have been carried out . One possible reason is that as a neglected pathogen , T . callipaeda does not draw enough attention from most parasitologists . This may also be attributed to the difficulty in collecting T . callipaeda isolates from different hosts and distinct geographical locations . More than 10 years ago , Otranto and Traversa [13] performed the first genetic variance analysis among Thelazia species by using the first internal transcribed spacer ( ITS1 ) ribosomal DNA sequence , and concluded that the ITS1 sequence was a useful genetic marker for the molecular identification of Thelazia spp . In 2005 , Otranto et al . [12] investigated the genetic variability among 50 individual adult specimens of T . callipaeda from Europe and Asia based on the mitochondrial cytochrome c oxidase subunit 1 gene ( cox1 ) . Recently , although various publications of the T . callipaeda infections in animals and humans have been reported [3 , 14–18] , no subsequent related studies about its population genetics have been reported . In this study , we explored the genetic variability within human T . callipaeda isolates collected from different geographical locations in China by using three mitochondrial genes and two nuclear ribosomal DNA sequences as follows: mitochondrial cytochrome b ( cytb ) , the small subunit of ribosomal DNA gene ( 12S rDNA ) and cox1; ITS1 and the small subunit of nuclear ribosomal DNA ( 18S rDNA ) . These molecular markers were selected because they are suitable for inferring population differences and conducting phylogenetic analysis at different taxonomic levels [12 , 13 , 19–21] . Using cox1 , we also performed a genetic variability comparative analysis on clinical T . callipaeda isolates collected in China and from previous publications . Additionally , the presumed transmission pattern of T . callipaeda investigated here relied on phylogeny and molecular dating methods .
This study was approved by the Life Science Ethics Committee of Zhengzhou University ( No . 2017–0006 ) . The protocol and written informed consent form were approved by the Human Ethics Committees of the Zhengzhou University . All subjects older than eighteen years old provided written informed consent; in the case of children , they provided written informed assent , and their parents/guardians provided written consent for them . All worms were collected from patients to treat their thelaziasis and not expressly for the purpose of the present study . A total of 32 worms were harvested from 13 patients from 11 distinct geographical locations in China from September 2007 to July 2016 ( Table 1 ) . All nematodes were removed from the eyes of patients with intraocular forceps , while the patients were anesthetized with oxybuprocaine . The collected nematodes were transferred to Petri dishes containing physiological saline ( 0 . 9% NaCl ) . These eyeworms were identified as T . callipaeda according to the morphological characteristics ( e . g . , shape of the buccal capsule , presence of transversally striated cuticle and cloacal papillae , morphology of the spicules in males and the position of the vulva in females ) described in Otranto et al . [22] . Total genomic DNA was extracted from individual specimens using the EasyPure Genomic DNA Kit ( Transgen , China ) following the manufacturer’s protocol . Five molecular markers , viz . cytb , cox1 , 12S rDNA , ITS1 and 18S rDNA , were amplified to explore the genetic diversity of T . callipaeda . For 12S , ITS1 and cox1 , the amplifications were obtained using primer combinations described in Casiraghi et al . ( 2004 ) [23] , Otranto and Traversa ( 2004 ) [13] , and Otranto et al . ( 2005 ) [12] , respectively . The forward and reverse primers used for amplifying the cytb and 18S markers were designed as follows: cobF 5′-TGATTGGTGGTTTTGGTAA-3′; and cobR 5′- ATAAGTACGAGTATCAATATC-3′; and 18SF 5′- CTCATAAAATAATTGG TGAATCTGAATAGC-3′ and 18SR 5′-ATAACTTTTCAGCAATGGTTACAG-3′ , respectively . PCR products were purified using the EasyPure PCR Purification Kit ( Transgen , China ) and sequenced in both directions at the Genwiz Company ( Beijing , China ) . All sequences were deposited in the GenBank database ( S1 Table ) . The sequenced genes were initially aligned using the default settings in the program Clustal X v . 2 . 0 [24] and adjusted in MEGA v . 6 . 06 [25] according to the corresponding amino acid sequences of protein-coding genes and secondary structure of ribosomal DNA sequences . The nucleotide composition , conserved sites , variable sites , parsimony-informative sites , and singleton sites were estimated using MEGA v . 6 . 06 . The program DnaSP v5 . 10 [26] was employed to analyze the number of haplotypes , haplotype diversity ( Hd ) , and nucleotide diversity ( Pi ) of each molecular marker . Network v5 . 0 [27] was used to draw a median-joining network to analyze the relationships among the detected haplotypes . Analysis of molecular variance ( AMOVA ) was computed in Arlequin v . 3 . 5 [28] with non-parametric permutations of 1 , 000 times ( p = 0 . 05 ) to detect the partitions of genetic diversity within and among populations . Pairwise FST values between populations were performed for all datasets in Arlequin to explore levels of genetic differentiation among the populations . The significance of FST values evaluated was based on 1000 random permutations . Demographic changes were also estimated using mismatch distributions in Arlequin with 1000 simulations , under a scenario of no recombination . The validity of the expansion model was tested by using the sum of squared deviations ( SSD ) and raggedness index ( RI ) between observed and expected mismatches . The neutrality tests using Tajima’s D [29] and Fu’s FS [30] were also applied through Arlequin as an assessment of possible population expansion . To make a worldwide genetic variability comparative analysis of T . callipaeda , all available sequences of cox1 in the GenBank database were included ( S1 Table ) . In addition , we performed a Bayesian skyline plot analysis ( BSP ) implemented in BEAST v1 . 8 . 2 [31] to estimate the change in population size over time , and the time to the most recent common ancestor ( tMRCA ) for each T . callipaeda haplotype . A piecewise-constant skyline model was selected . The molecular evolutionary rate of cox1 was fixed at 0 . 01 substitutions per site per million years ago ( Mya ) according to the substitution rate for nematode mtDNA [32] . Tracer v1 . 5 [33] was used to reconstruct the demographic history over time . The phylogenetic pattern of all cox1 haplotypes was estimated through maximum parsimony ( MP ) and Bayesian inference ( BI ) . MP analysis was performed in MEGA v . 6 . 06 . Confidence in each node was assessed by boot-strapping ( 1000 pseudo-replicates ) . Bayesian inference was performed in MrBayes v . 3 . 2 [34] , after determining the appropriate substitution model by applying the Akaike information criterion ( AIC ) in jModelTest 2 [35] . The analysis consisted of two runs , each with four MCMC chains running for 5 , 000 , 000 generations , and sampling every 100th generation . Stationarity was assessed using a convergence diagnostic . An average standard deviation of the split frequencies ( ASDSF ) < 0 . 01 was used as criteria for convergence between both runs . The consensus tree was drawn after removing the first 10 000 trees ( 20% ) as the burn-in phase . The approximate divergence time was estimated using an uncorrelated log-normal relaxed molecular-clock model in the BEAST v1 . 8 . 2 program . The substitution model was assigned following model selection by jModelTest 2 . For the earlier tree , a basic coalescent model was chosen , assuming a constant population size over the given time period considered . Two replicate MCMC runs were performed , with the tree and parameter values sampled every 1 , 000 steps over a total of 1×108 steps .
The sequence alignments for cytb , cox1 , 12S , ITS1 and 18S were 1035 bp , 660 bp , 453 bp , 772 bp and 1201 bp , respectively . In 32 isolates from eleven localities , the genetic markers of cox1 , 12S , ITS1 and 18S were used to identify 9 , 8 , 8 and 7 new haplotypes ( Haps ) , . For cytb , 8 Haps were identified in only 25 isolates because the sequences of isolates from Hefei ( HF ) and Liuan ( LA ) of Anhui province were not amplified ( S2 Table ) . With the exception of isolates from Pingdingshan ( PDS ) and Zhengzhou ( ZZ ) of Henan province , which shared two haplotypes , each geographic population shared only a single haplotype . Analyses of the median-joining networks ( Fig 1 ) showed that samples from HF and LA shared a haplotype ( Hap1 ) when using cox1 , 12S and ITS1 , but under the analysis with 18S , these samples identified two Haps ( Hap1 and Hap2 ) . Clinical eyeworms from Dandong ( DD ) and Tongchuan ( TC ) shared geographical specific Haps using each of the selected markers . Isolates from Huanggang ( HG ) and Wuhan ( WH ) also revealed specific Haps when using cytb , cox1 , 12S and ITS1; however , these samples shared a single haplotype using 18S . Using cytb , 12S and ITS1 , a single geographical specific haplotype was identified in T . callipaeda from Shangluo ( SL ) . Interestingly , using cox1 ( Hap3 ) and 18S ( Hap5 ) , these isolates from SL shared the same haplotype with all samples from Luoyang ( LY ) , 4 samples from Pingdingshan ( PDS ) and one from Zhengzhou ( ZZ ) . The remaining isolates from PDS and ZZ shared the same haplotype using 12S , ITS1 and 18S . However , when using cytb and cox1 , the remaining isolates possessed two distinct Haps . Using cytb , 12S , ITS1 and 18S , worms from Jiaozuo ( JZ ) shared the same haplotype with the second sample from ZZ . Only with cox1 , did the JZ isolates identify as a distinct haplotype . The results from the analysis of molecular variance ( AMOVA ) showed that much more genetic variance lay among the populations than within the populations for all datasets , more specifically , cytb: 69 . 77% vs . 30 . 23%; cox1: 91 . 32% vs . 8 . 68%; 12S: 94 . 26% vs . 5 . 74%; ITS1: 91 . 69% vs . 8 . 31%; 18S: 83 . 71% vs . 16 . 29% ( Table 2 ) . The pairwise fixation index ( FST ) values between specified geographical regions were estimated for all molecular markers used to measure the population differentiation ( S3 Table ) . With the exception of FST values between PDS and the remaining populations , and ZZ and the remaining populations , most of the FST values reached 1 . 00 . No estimated pairwise FST values were statistically significant besides those between LA and the remaining populations , and PDS and the remaining populations . Based on all genes , the neutrality tests of Tajima's D and Fu's FS for the total population showed non-significant positive values , except for the negative value of Fu's FS ( -1 . 65282 , p = 0 . 164 ) , using the 18S gene ( Table 3 ) . Using all markers , mismatch distribution analyses revealed multi-modal frequency distributions for the total population , rejecting possible population expansion ( S1 Fig ) . In addition , low values were found for the sum of squared deviation ( SSD ) and raggedness index ( RI ) under the demographic expansion model ( Table 3 ) . A total 21 haplotypes were identified within T . callipaeda isolates from 25 localities in 12 countries ( Table 1 ) . Hap1 was only found in T . callipaeda from European countries ( Italy , Germany , The Netherlands , Portugal , Serbia , Romania , Slovakia , Hungary , etc ) . The remaining haplotypes were shared by samples from Eastern Asia countries ( Korea , Japan and China ) . Haps2–8 and haps10–12 were shared by samples isolated from domestic dogs ( Canis familiaris ) . Hap9 and haps13–21 were identified in eyeworms collected from humans ( Homo sapiens ) . The alignment of 21 haplotypes revealed nucleotide variations ( 36 transitions and 1 transversion ) at 37 alignment positions ( S2 Fig ) . The majority of the nucleotide variability was at the third codon position ( n = 31; 83 . 8% ) , whereas the remainder was at the first codon position ( n = 6; 16 . 2% ) . Of the 37 variable positions , only two sites generated nonsynonymous mutations . One was at alignment position 189 ( A—G ) of Hap4 , which changed the amino acid from Methionine to Valine . The other , also located in Hap4 ( position 252: C—T ) , changed Leucine to Phenylalanine . Five nucleotides ( i . e . , G—A at alignment positions 89 , 149 , 206 and 257; and C—T at position 539 ) were invariable among all individuals from Europe and among all individuals from Eastern Asia ( Korea , Japan and China ) , but they were different between Europe and Asia . The pairwise comparisons among 21 haplotypes ranged from 0 . 15 to 2 . 86% ( Table 4 ) . Within each country studied , the intraspecific divergences in 4 haplotypes from Korea , 4 haplotypes from Japan and 13 haplotypes from China were 0 . 31–1 . 09 , 0 . 15–2 . 55 and 0 . 15–1 . 73% , respectively . The AMOVA results showed that much more genetic variance lay among the populations ( 74 . 79% ) than within the populations ( 25 . 21% ) . Between geographical regions , the population differences ( FST value ) were 0 . 961 for Europe vs . Korea , 0 . 866 for Europe vs . Japan , 0 . 827 for Europe vs . China , 0 . 444 for Korea vs . Japan , 0 . 275 for Korea vs . China , and 0 . 430 for Japan vs . China ( S4 Table ) . The likelihood models identified by the jModelTest ( AIC ) suggested that the HKY+G model was most suitable for cox1 haplotypes . Both maximum parsimony and Bayesian analyses generated consistent tree topologies ( Fig 2A and S3 Fig ) . Among the tested haplotypes , Hap1 and Hap4 composed a single clade ( clade I ) , and the remaining haplotypes made up another clade ( clade II ) . Within clade II , the earliest diversifications gave rise to Hap13 , then to Hap9 , Hap11 and Hap12 ( Japanese haplotypes without Hap10 ) . The next diversification event separated the remaining haplotypes . The molecular dating analysis suggested that the two clades began to diverge during the middle Pleistocene ( Fig 2A ) . The time of origin of clade I is estimated to be approximately 0 . 58 Mya ( late Pleistocene ) with a 95% highest posterior density ( HPD ) of 0 . 23–1 . 01 Mya . Clade II started to develop in the middle Pleistocene ( 0 . 78 Mya ) with a 95% HPD of 0 . 47–1 . 17 Mya . The early branching of the Japanese haplotypes ( without Hap10 ) started in the late Pleistocene ( 0 . 14 Mya , with a 95% HPD of 0 . 02–0 . 32 Mya ) . Neutrality tests of Tajima's D and Fu's FS for the total population showed non-significant positive values , thereby rejecting possible population expansion ( S5 Table ) . Mismatch distribution analyses revealed multi-modal frequency distributions ( Fig 2B ) . In addition , low values for the sum of squared deviation and raggedness index under the demographic expansion model were found ( S5 Table ) . The result of Bayesian skyline plot analysis of the T . callipaeda population revealed a gradual expansion trend , but also rejected sudden population expansion ( Fig 2C ) .
Although more than 10 species of the spirurid genus Thelazia can cause veterinary or medical problems in many parts of the world [13] , only T . callipaeda affects humans in China , causing mild to severe clinical symptom [36] . It is still unknown whether there are other human infective species or genotypes of T . callipaeda in China . In this study , we collected 32 eye worms from 13 patients from 11 distinct geographical locations in China over a period of 10 years , and performed a genetic variation analysis of these isolates using a multi-gene approach to investigate the population structure of T . callipaeda . Haplotypes of the Chinese T . callipaeda population identified by each molecular marker were generally consistent . Of all the geographic populations , only PDS and ZZ isolates revealed two different haplotypes . Worms from ZZ were harvested in two different patients . However , interestingly , all PDS isolates that were collected from a single child also had two haplotypes , indicating the patient probably experienced two or more infections of drosophilid flies . Each geographic population and the total population demonstrated high haplotype diversity ( Hd ) values ( nearly 1 . 0 ) by all markers , however , the nucleotide diversity ( Pi ) values of each gene were below 0 . 01 , indicating that multiple haplotypes were differentiated by few nucleotide mutations . Correspondingly , the AMOVA results also showed that much more genetic variance exists among the populations than within the populations [19] . The pairwise fixation index ( FST ) was used to measure the population differentiation . Most of the FST values between specified geographical regions were far more than 0 . 25 , indicating very high genetic differentiation and a long-term interruption of gene flow among the geographical populations [37] . Based on the partial cox1 sequence , Otranto et al . [12] performed the first genetic variability investigation of the T . callipaeda population in 2005 . Since then , sequences of T . callipaeda from different geographical locations and different hosts have been published [3 , 16–18] . In this study , based on previous research by Otranto et al . [12] , we added new data from Chinese clinical isolates and other published data to generate a comprehensive genetic variation analysis of the T . callipaeda population . Some interesting discoveries were made: ( 1 ) Thirteen novel genotypes of T . callipaeda were identified; four from Japan ( Haps9–12 ) , and another nine from Chinese clinical isolates ( Haps13–21 ) ; within the 4 Japanese genotypes , one was from a human , and the other 3 were from Canis familiaris . ( 2 ) In cox1 , Otranto et al . [12] found six nucleotide positions ( positions 89 , 149 , 206 , 257 , 539 and 608 ) that were invariable among individuals from Europe and Asia , but they were distinct between Europe and Asia . We confirmed five of them ( positions 89 , 149 , 206 , 257 , 539 and 608 ) using an enlarged dataset in this study . ( 3 ) Population differences between Europe and Asian countries were larger than those among China , Korea and Japan . Within Asian countries , the genetic difference between China and Korea was the smallest , while the genetic difference between Korea and Japan was the largest . Only within human isolates were the genetic differences among Chinese samples smaller than those between China and Japan . This phenomenon may be attributed to the geographical distance’s influence on gene flow among species [38] , and possibly to the geographical origin , which needs to be analyzed further in the future . ( 4 ) The T . callipaeda population from Europe and Asia should be divided into the following two subgroups: one group ( clade I ) comprising worms from European countries ( Hap1 ) and a sample isolated from a dog in Anhui of China ( Hap4 ) ; the other group ( clade II ) comprising T . callipaeda from Korea , Japan and China . Individuals in clade II diverged earlier than those in clade I . Within the Asian haplotypes , Hap13 ( clinical isolates of Anhui ) was the ancestral haplotype; then , it was transmitted to Japan , Korea and other regions of China . ( 5 ) All of the neutrality tests , mismatch distribution and BSP analyses rejected possible population expansion of T . callipaeda , indicating this nematode population was in the stable phase [37] . Generally , the population structures of parasites are associated with large genetic differentiation among populations from different geographical regions and/or hosts and low intra-population genetic variability [39 , 40] . In this study , there was a relatively large genetic differentiation ( above 0 . 82 ) between the European and Asian populations; however , the genetic variability among Asian countries was small ( below 0 . 44 ) . In addition , phylogenetic analyses supported the conclusion that isolates from Europe and Asia ( excluding the Hap4 ) are separate populations . However , the genetic differentiation of T . callipaeda from distinct hosts ( dogs , foxes , cats and human ) was inconspicuous , suggesting a lack of connection between eyeworms and host species [12] . That being said , an important factor to consider is the abundance of genetic diversity in the Asian isolates . A possible explanation for this genetic diversity is that T . callipaeda populations are tightly linked to the intermediate host ( vector ) . The drosophilid fly Phortica variegata was the vector of T . callipaeda in Europe , whereas the T . callipaeda in Anhui of China was transmitted by P . okadai [9 , 12] . Although it is still unknown whether other vector species exist in Asian countries , considering the high genetic diversity of T . callipaeda in Korea , Japan and China , it is very likely that some novel species or genotypes of Phortica flies exist in these areas . Hence , further studies should be launched to explore the biology , population genetics , ecology and epidemiology of T . callipaeda in the future . In summary , the Chinese clinical isolates of T . callipaeda expressed a high level of genetic diversity . Using the cox1 gene , thirteen new genotypes were identified , four from Japan ( Haps9–12 ) , and nine from China ( Haps13–21 ) . There were five positions in the cox1 sequences ( 89 , 149 , 206 , 257 , 539 and 608 ) that were confirmed invariable among individuals from Europe and Asia , but the sequences were distinct between these two regions . Population differences between Europe and Asian countries were greater than those among China , Korea and Japan , and the T . callipaeda populations from Europe and Asia should be divided into two separate sub-populations . These two groups started to diverge during the middle Pleistocene .
|
Thelazia callipaeda is the causative agent of thelaziasis canids , felids and humans . Despite the existing threat of thelaziosis in China , the genetic diversity of T . callipaeda has not been investigated across its wide geographical distribution in China , yet such information may provide insight into the disease epidemiology and the development of specific control measures . In this study , the genetic variation of 32 T . callipaeda clinical isolates collected from 13 patients from 11 geographical locations in China were explored using the following multi-molecular markers: cox1 , cytb , 12S rDNA , ITS1 and 18S rDNA . In addition , the population structure of T . callipaeda from Europe and other Asian countries was analyzed using the cox1 sequences collected during this study and from the GenBank database . In general , the Chinese clinical isolates of T . callipaeda demonstrated high genetic diversity . Based on the cox1 gene , a total of 21 haplotypes were identified , one circulated in European countries ( Hap1 ) , while the other 20 haplotypes were dispersed in Korea , Japan and China . There were five nucleotide positions in the cox1 that were confirmed as invariable among individuals from Europe and Asia , but the sequences were distinct between these two regions . Population differences between Europe and Asian countries were greater than those among China , Korea and Japan , such that the T . callipaeda population from Europe and Asia should be divided into two separate sub-populations . These two groups started to diverge during the middle Pleistocene .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"biogeography",
"ecology",
"and",
"environmental",
"sciences",
"china",
"japan",
"population",
"genetics",
"geographical",
"locations",
"genetic",
"mapping",
"population",
"biology",
"europe",
"korea",
"geography",
"phylogeography",
"people",
"and",
"places",
"haplotypes",
"asia",
"heredity",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology"
] |
2018
|
Population structure analysis of the neglected parasite Thelazia callipaeda revealed high genetic diversity in Eastern Asia isolates
|
Viral replication relies on host metabolic machinery and precursors to produce large numbers of progeny - often very rapidly . A fundamental example is the infection of Escherichia coli by bacteriophage T7 . The resource draw imposed by viral replication represents a significant and complex perturbation to the extensive and interconnected network of host metabolic pathways . To better understand this system , we have integrated a set of structured ordinary differential equations quantifying T7 replication and an E . coli flux balance analysis metabolic model . Further , we present here an integrated simulation algorithm enforcing mutual constraint by the models across the entire duration of phage replication . This method enables quantitative dynamic prediction of virion production given only specification of host nutritional environment , and predictions compare favorably to experimental measurements of phage replication in multiple environments . The level of detail of our computational predictions facilitates exploration of the dynamic changes in host metabolic fluxes that result from viral resource consumption , as well as analysis of the limiting processes dictating maximum viral progeny production . For example , although it is commonly assumed that viral infection dynamics are predominantly limited by the amount of protein synthesis machinery in the host , our results suggest that in many cases metabolic limitation is at least as strict . Taken together , these results emphasize the importance of considering viral infections in the context of host metabolism .
Any virus is necessarily a metabolic product of its host , since viruses lack the macromolecule machinery and small molecule precursors required to replicate . This dependence has been underscored by recent screens to determine the host genes required for viral infection in a variety of species . The published sets of host-gene viral dependencies have consistently included metabolic genes - both enzymes and regulators - in systems ranging from phages T7 and lambda , to the human viruses HIV and influenza [1]–[8] . In complementary findings , some bacterial viruses have recently been shown to encode components as well as direct modifiers of host metabolic machinery [9] , [10] . Taken together , these studies emphasize the need to understand viral infection in the context of host metabolism [11] . Viral host dependency screens are useful for identifying individual host genes involved in the metabolic interplay of viral infection; however , studying any of these single points of connection is likely to reveal a complex network of host-viral interactions [12] . Understanding infection as a highly integrated system is therefore necessary to predict the outcome of viral infection following perturbations , such as changes to the host nutritional environment . Similarly , metabolism is a deeply interconnected network , and viral infection represents a dynamic perturbation of it . Achieving a systems-level understanding of host-viral metabolic interaction therefore requires , a strong set of computational tools coupled with quantitative dynamic measurements . Given the challenge presented by developing such modeling tools and making the needed measurements , bacteria and their viruses , particularly E . coli and certain of its bacteriophages , are favorable candidate model systems for building a systems-level understanding of infection . These systems have a long history of study , individually and together , and as a result are associated with a wealth of well-established observations and experimental protocols . Additionally , the host-viral dependency screens involving E . coli identified sets of genes whose products were far better characterized and annotated than in any other screen [1] , [2] . These systems also have industrial relevance: threatening large-scale cultures [13] , and alternately providing highly specific disinfection tools [14] . Critically , E . coli and its phage are sufficiently understood to enable the construction of predictive computational models . Phage T7 replication has been described with structured ordinary differential equations ( ODEs ) , that account for the dynamic production of molecular species that comprise the phage during infection [15] ( Figure 1A right ) . This model was used to computationally predict the infection outcome of phage genome modifications [16] , [17] . Separately , host E . coli metabolism has been most comprehensively modeled using Flux Balance Analysis ( FBA ) , which uses linear optimization of an objective function to solve a system of steady-state mass balance ODEs [18] . FBA-based models have expanded to account for essentially all of the known metabolic functionality in E . coli ( Figure 1A upper left ) [19]–[21]; these models capture growth rates and nutrient exhaustion as well as the impact of genome perturbation and evolutionary outcomes over time [22]–[24] . Two previous extensions of the E . coli FBA and T7 ODE models have attempted to encode some dependence of viral replication on host state . One effort was based on the E . coli FBA model , with metabolic reactions added to describe production of MS2 virions [25] , thus demonstrating the fundamental translation of viral composition to host metabolic terms ( the analogous translation for T7 is denoted in Figure 1A , lower left ) . The implemented FBA objective function assumed that the host optimized all of its resources toward viral production immediately upon infection , resulting in an overprediction of phage production . The other modeling effort added a set of correlations between the host growth rate and the availability of replication machinery for T7 processes [26] , improving the model's predictions ( Figure 1A upper right ) to the T7 ODE model . Both of these efforts strongly suggest that a comprehensive , detailed effort to integrate the host and virus into a single computational model will significantly advance our understanding of viral infection in its metabolic context . Ideally such an effort would build on previous work with this host-virus system , despite the different ODE and FBA modeling techniques . Integration of FBA and ODE-type models sets the flux values for a subset of reactions using available kinetic rate equations [27] , providing a conceptual framework for combining the host and viral models as depicted in Figure 1A . Here we present an integrated model that is based equally on E . coli FBA and the T7 ODEs . It includes a mathematical description of metabolic reactions and demand introduced by the virus , as well as a simulation algorithm that facilitates interaction between the two models throughout the entire course of infection . Our integrated modeling approach enables us to predict phage production changes as the host nutritional environment shifts , and provides insight into the underlying limiting factors in T7 infection .
Our integration of the T7 ODEs and E . coli FBA ( Figure 1A ) began with a set of additions to each of the individual models . The E . coli FBA model stoichiometric matrix required new reactions to describe the routing of host precursors and energy towards viral synthesis . One reaction was constructed for the synthesis of each viral species represented in the ODE model: mRNA and protein for each of 59 viral genes , viral genome synthesis , and a reaction enforcing the recharge of nucleotide monophosphates ( NMPs ) released from host genome degradation ( 123 total reactions; Figure 1B and Methods ) . The T7 ODEs required one ‘production only’ reaction rate equation for each of the 123 phage reactions that consume the host metabolites that were added to the host FBA; the net concentration change rate for each molecular species in the original T7 ODEs consisted of production minus consumption terms . However , only the production rate term constrained the stoichiometric reaction in the FBA . Furthermore , predictions based on the T7 ODEs are valid for a single infection cycle only , and lysis has not been modeled because knowledge of the proteins involved is still insufficient to inform a meaningful representation [28] . As a result we constrained the scope of the integrated model to one single infection cycle . Next , we expanded the integrated-FBA approach beyond its original capacity to handle the viral demand for resources when these resource demands outpaced the host production capacity . The original implementation of integrated-FBA [27] included ODEs based on central metabolism , which were informed by the environmental state and thus remained within the capacity of host metabolism without any direct communication of host limitations . In contrast , the T7 ODEs do not encode variation in the environmental conditions or the corresponding changes in the host network state's supply of metabolites . As a result , conflicts between the viral metabolite demands and host metabolite supply can arise during the simulation . We therefore encoded communication of information about host limits to the T7 ODEs . This strategy was complicated by the fact that the kinetic formulation of the T7 ODEs is largely independent of small molecule concentrations , except for the nucleotides required for T7 genome synthesis . Furthermore , FBA does not provide concentration information . Consequently , we devised a metabolite allocation-based approach to bounding reaction rates . Recognizing that the host-viral metabolic interface is the set of common metabolites used in macromolecule synthesis , we split the matrix formulation ( Figure 2A ) into a sum of metabolite rate vectors that represent the host supply and viral demand , where the former constrains the latter . Given a selected host flux distribution , we calculate a strict bound on viral metabolite use . Due to the lack of kinetic information about how the viral metabolic reactions contribute to the metabolite demand , we assume that all viral reactions have an equal and high affinity for precursor metabolites . After calculating rates for the viral reactions from the T7 ODEs to determine the demand for viral metabolites , we scale the rates of all reactions consuming a given metabolite by the same fraction such that total demand is brought within host supply . This method assures that while all reactions are limited evenly , no reaction is limited by a metabolite it does not consume; if amino acids are scarce but dNTPs are available , genome synthesis can proceed but translation cannot . In summary , this allocation method converts the information about the host metabolic network state into constraints on the T7 ODEs . We implemented this method as part of an algorithm for T7 ODE and E . coli FBA integration with bidirectional information exchange and mutual constraint at each time step ( Figure 2B ) . After initial specification of the host nutritional environment , the overall viral demand is calculated ( without consideration of host limits ) using the T7 ODEs , and the host capacity calculated using FBA . Host supply and viral demand are reconciled by calculating the upper bounds on viral production fluxes , after which the T7 ODEs are re-evaluated over the integration time step because metabolite limitation of one viral ODE may affect the ODE solution as a whole . Finally , the infected host flux distribution is calculated using optimization on the host metabolic network , with viral fluxes bounded exactly to constrained T7 ODE reaction rate values . To validate the ability of the model to predict infection phenotypes , we observed E . coli infection by T7 during growth on tryptone broth . We first measured the growth of E . coli cultures in the presence and absence of T7 ( Figure 3A ) . The culture is cleared within 35 minutes , representing approximately two infection cycles at . Unfortunately , with standard OD resolution , the infected and uninfected cultures were not distinguishable from one another within the single infection cycle ( Figure 3A ) simulated by the model . Thus , differences in host growth rate were not a useful metric to assess the prediction performance of our computational model . We therefore returned to the traditional plaque assay-based approach to determine the number of phage produced per infected host cell during a single initial infection cycle , consistent with previous work with the T7 ODEs [15] , [26] ( Figure 3B ) . We observed rapid increases in the number of phage beginning around 10 minutes . To compare model predictions to observations , we simulated phage production time courses under the same environmental conditions using our fully integrated model as well as the T7 ODEs alone . We found that the T7 ODEs alone substantially overpredicted the production of T7 phage over time ( Figure 3B ) . This overprediction has been reported previously [15] , [26] . The integrated model more accurately captured the phage production time course ( Figure 3B ) , suggesting that the integrated model is limiting the production of T7 virions ( detailed comparison across media given below ) . To determine the cause of this limitation , we considered the model's predictions of phage production and host metabolism in more detail . We compared simulated intracellular concentrations of selected phage components for the integrated simulation to those during simulation of the T7 ODEs alone ( Figure 3C ) . The model predicts that production of Gene Product ( GP ) 1 is limited at translation; GP 1 is the T7 RNA polymerase and is required to transcribe middle and late T7 genes . Despite reduced transcription capacity , sufficient mRNA for the major capsid protein ( Gene 10A ) is still produced . Major capsid protein production is metabolically limited at translation , and thus procapsid availability for phage assembly is decreased , resulting in fewer phage produced during late infection than predicted by the T7 ODEs alone . In the integrated simulation , although phage T7 genome is produced at the same rate as the T7 ODEs alone , it is not packaged as quickly , with a considerable fraction of the total genomes produced remaining unpackaged after assumed lysis . This excess phage T7 genome resulting from phage production limitation at the protein level is consistent with previous experimental observations [15] . The most prominent limitation by metabolism appears during the later steps of replication: mid and late gene product synthesis and genome production . In contrast , mRNA production is relatively unperturbed early in the simulation , suggesting that metabolic limitation varies in its impact over different periods during infection . After considering the phage reaction changes in the integrated simulation , we used the model to investigate the changes in host metabolism during infection . The flux-balance component of the integrated model calculates a predicted flux distribution for E . coli growth on tryptone in the presence and absence of phage . Essentially all of the non-zero fluxes change dynamically over time in the presence of T7; a subset of these changes are shown alongside the underlying metabolic map ( Figure 4 ) . Many metabolic reactions experienced prominent flux changes that were coordinated during infection . Dynamic coordination of fluxes in time is not particularly surprising considering the underlying network structure of constraints . However , these similarities in addition to the sheer number of total fluxes that require consideration render unaided visual inspection of infection dynamic information rather uninformative . We found it useful to cluster the flux dynamics into broad categories , which facilitate interpretation of the interesting flux patterns in central and peripheral metabolism during viral replication . The majority of the observed flux clusters are driven by viral flux requirements ( Figure 4B ) . The increase in amino acid synthesis and uptake corresponds in time to the synthesis of viral proteins ( Figure 4Bi–ii ) , and similarly flux through nucleotide phosphorylation is high during the period of host genome digestion to dNMPs and viral use of dNTPs ( Figure 4Biii–iv ) . Increased nucleotide recharge and pooling is known to occur during phage T7 replication , due at least in part to interactions between phage gene products and host metabolic enzymes [2] . Some complex host flux dynamics result from multiple viral resource interactions ( Figure 4Bv–vii ) ; flux towards nucleotides first increases during rapid early viral mRNA production , and then decreases as viral genome synthesis occurs , corresponding to the presence of large quantities of nucleotides . Flux towards host membrane components and cofactors decreases as the ability of the host to synthesize biomass is reduced by the viral draw on components ( biomass flux decrease before 5 min ) and energy ( biomass flux decrease between 5 and 10 mins during dNTP recycling ) ( Figure 4Bvii–ix; light blue ) . This cluster is the largest of the nonzero flux clusters across and within media , and the sharp decrease in flux within 5 min represents the shutdown in processes that are not required by the virus . Interestingly , this shutdown is not explicitly encoded by either model and therefore represents an emergent property of the integrated model system . The detailed flux maps therefore provide potential for a deeper biological insight regarding the underlying metabolic changes that occur during viral infection . The T7 ODEs were originally parameterized to fit data where E . coli grew on tryptone broth or other rich media [15] . Later work incorporated correlations between available host machinery ( e . g . , ribosomes ) and host growth rate into the ODEs in order to account for the effect of growth rate on infection dynamics [26] . Host metabolism is encoded explicitly in our integrated host-virus model , and so instead of a given growth rate parameter , the integrated model requires only the environmental conditions as inputs . Unlike either individual model , the integrated model is capable of predicting the viral infection dynamics for many different culture conditions . We tested model predictions for three previously unmodeled conditions: glucose , succinate , and acetate minimal media . In each case , we measured the phage production over time ( Figure 5 , bottom left and top panels ) . For glucose and succinate media , the models produced dynamics nearly identical to each other as well as similar to the experimental data . However , for infections on acetate minimal media , the integrated model was more accurate than the T7 ODEs alone . The two predicted time courses differ because the integrated model accounts for the slow growth and nutritional limitation of E . coli on acetate ( roughly half of the growth rate on succinate ) . In particular , small decreases in gene product synthesis result in delayed achievement of the thresholds necessary for phage genome replication initiation . Furthermore , all of the simulations , from both the integrated model and the ODEs alone , deviate from the typical one-step-growth phage production trajectory . This is due to the rigid description of host DNA degradation and incorporation into viral genomes in the ODEs , which was originally characterized under a single environmental condition . Quantitative comparison of our observations to the model predictions verified that tryptone simulations were the most indicative of experiment , and that the tryptone and acetate integrated model simulations outperformed those of the ODEs alone ( Figure 5 , bottom right panel ) . We next wanted to understand how the host and viral fluxes change under these different nutrient conditions . Detailed individual media flux maps analogous to 4 are provided for glucose , succinate , and acetate media in Figures S2 , S3 , and S4 respectively . To generate a global evaluation of the host flux response to infection on varying media , we analyzed the aggregate similarity of the total flux distribution between pairs of media ( Figure S5 ) . Generally this comparison indicated that the flux distribution for infection during growth on acetate was very similar to the distribution during growth on succinate , while there was more divergence between the tryptone and glucose flux distributions than for any other media pair . Figure 6 displays the dynamic metabolic flux distribution for all four infection simulations , normalized to facilitate comparison . Of the fluxes that are non-zero in any of the media conditions , a large fraction show highly similar dynamics . These fluxes include critical biomass-related reactions such as those that contribute to membrane ( Figure 6Bi–iii ) or ribonuclotide biosynthesis . In some regions of the metabolic network , flux dynamics depend more on the media conditions; for example , in central metabolism the flux direction is often reversed between glucose and the other media because glycolysis is occurring rather than gluconeogenesis ( Figure 6Biv ) . Reactions involved in amino acid synthesis also exhibit this phenomenon , as they increase in rate on all three minimal media , yet are zero on tryptone medium ( Figure 6Bv ) , which contains amino acids . Another interesting example involves citric acid cycle activity , which is especially increased during the high energy demands of nucleotide recycling ( Figure 6Bvi ) . One final subset , adjacent to key metabolites such as pyruvate ( PYR ) , oxaloacetate ( OAA ) , and succinate ( SUCC ) , displayed erratic and rapid jumps between their extreme values , which results from equivalent optimal flux distributions calculated by FBA in highly interconnected sections of the metabolic network . Finally , we used our model results to address the issue of host-based limitation of viral infection . Many studies assume that phage infection of E . coli is limited by “machinery” – the number of ribosomes , RNA polymerases , and similar factors . Another possibility is that in some cases the host metabolic rates are limiting factors; however , decoupling this limitation is difficult due to the regulation of E . coli protein synthesis capacity by the availability and type of nutrients [29] . We sought to compare the effects of E . coli machinery- or metabolic-based limitation on T7 infection , an exploration enabled by our integrated simulation which can be perturbed in ways not practical experimentally . The detailed simulation output presented in Figure 3C indicates that metabolic limitation may be more prominent for certain phage processes and during specific periods of infection . As a summary output for comparison across conditions , we chose the phage production at seventeen minutes post infection . This point is shortly after which all cultures had begun to lyse , releasing phage , and thus making the bulk quantity relevant to phage propagation across generations within a host population . The boundary representing machinery limitations is provided by evaluation of the T7 ODEs alone across varied input growth rates ( Figure 7 ) . The region that falls below the model prediction is feasible ( dark gray ) , and everything above is not ( light gray ) . To calculate the bounding metabolic phage production limitation , we simulated the integrated model with the modification that excess host replication machinery components were provided to the ODE model ( accomplished by passing a higher host growth rate to the ODEs than that predicted by FBA ) . This calculation was carried out for carbon- and oxygen-limited growth at each resulting growth rate , which resulted in uniform predictions of phage production at each growth rate . Metabolic feasibility here refers to the supply of small molecule metabolites needed to build phage virions; the metabolic limit increases smoothly with host growth rate because the phage is made of a subset of the metabolites included in the host biomass reaction that represents FBA growth , and a state of host growth maximization is assumed for host supply . This context reveals the integrated model to be slightly mechanistically limited over the range of growth rates between approximately 0 . 4/hour and 1/hour , and more severely metabolically limited at higher and lower growth rates; however , simulations at very low growth rates do produce empty capsids , reflecting the strong repression of virion DNA production encoded in the ODEs . Metabolic limitation at high and low growth rates explains the better performance of the integrated model than the T7 ODEs alone in predicting phage production on acetate and tryptone media ( Figure 3 and 5 respectively ) .
In summary , we investigated the role of host metabolism in viral infection . E . coli infection by T7 provided a unique opportunity to address this issue because each system had been modeled , parameterized , and tested independently . We integrated the host metabolic FBA and T7 ODE models and compared the resulting integrated model predictions with new experimental observations . We found that our integrated model was not only a better predictor of viral infection dynamics than either of the individual models for a range of experimental conditions , but also shed new insight on the interplay between virus and host during infection . Most of the active host metabolic pathways were highly impacted by the metabolic demand imposed by virion production . Moreover , we grouped and categorized these pathways by their dynamics; these groups were directly related to the timing of viral demand for key virion components . It is commonly assumed that viral infection dynamics are predominantly limited by the amount of protein synthesis machinery in the host [15] , [30] . In contrast , our results suggest that in many cases metabolic limitation is at least as severe as machinery limitation . This conclusion in turn implies that the wealth of available metabolic reconstructions may enable computational predictions on virion production even when detailed information about interaction with host macromolecules is lacking . More broadly , these results emphasize the importance of considering viral infections in the context of host metabolism . Finally , we anticipate that models such as this integrated model may be used to rationally perturb the viral infection process by manipulating the host . The modeling and integration approaches developed here are general to a host flux-balance model and a set of viral ODEs , and by integrating the two it may be possible to predict key host metabolic factors whose absence would hamper infection , even as these factors depend on environmental conditions .
The bacterial host strain used was E . coli K12 BW25113 , and WT T7 phage ( ATCC , BAA-1025-B2 ) was propagated according to established protocol [31] . Tryptone media contained 10 g/liter Tryptone ( BD Bionutrients Bacto Tryptone ) and 5 g/liter NaCl consistent with previous T7 work [15] , [31] . M9 minimal media contained 56 . 4 g/liter Difco M9 Minimal salts , with added and ; carbon sources glucose , succinate , and acetate were added at 10 mM , 15 mM , and 30 mM to media preparations , respectively . All culture experiments were conducted at in a circulating water bath at a volume of 30 ml culture in a 250 ml flask that was magnetically stirred . Infections were at an initial MOI of 0 . 1 to assure hosts would only be infected once , and replicates were taken from separately infected flasks . Host population was measured as the optical density ( OD ) using a spectrophotometer at a wavelength of 595 nm . Phage dilution and storage was in SM phage buffer [26] . Measurements of phage titer were made by plating phage sample with fresh bacterial culture at 1 OD from tryptone media in 3 ml tryptone broth with 0 . 7% agar atop tryptone broth 1% agar , and incubating the plate in an inverted position at for approximately 3 hr [31] . One-step phage growth experiments were conducted consistent with published protocols [15] , [26] , [28] . Prior to infection , bacterial hosts grew exponentially to a total density of 0 . 2 OD . Pilot experiments suggested that essentially all phage absorbed into the host cells within five minutes . Therefore , after 5 minutes of infection in the initial culture flask , a sample was diluted 1000-fold in warm shaken media into another flask of the same total culture volume ( 30 ml ) to minimize adsorption of produced phage to new hosts . At 6 and 7 minutes ( time points selected as just following complete phage absorption ) infected hosts were counted . To count infected hosts samples were transferred into ice-cold 900 ml aliquots of phage buffer , returned to ice , and plated less than 30 minutes later . At 6 and 7 minutes , as well as all other time points , samples were transferred into room-temperature 900 ml aliquots phage buffer with chloroform for host lysis . The chloroformed samples were incubated at room temperature for 30 minutes with periodic vortexing , then stored at until plating , usually within an hour . Phage from lysed samples at later time points are reported normalized to the infected host count obtained by the difference of unlysed and lysed samples at 6 and 7 minutes . We implemented the T7 ODEs in MATLAB ( R2011a The MathWorks Inc . ) , informed by the equations presented in the initial publication [15] as well as the code available for the most recent version [26] . The T7 ODEs were originally compared to phage production data at having been simulated using parameters measured at either or [15] , [26] . The published flux bounds and regulatory rules of FBA correspond to E . coli growth at , and therefore for consistency the T7 parameters were modified to where necessary ( Table S5 ) . This modification included kinetic parameters and promotor strengths to maintain prediction constancy with the proportion of phage gene products produced [32] , ( Tables S4 ) . A stiff solver ( ode15s ) was used for all solutions of T7 ODEs , as required by discontinuous rate definition equations . The regulatory-FBA model reaction equations and metabolites are iMC1010v2 [33] , with the minor change that a few reversible reactions were reversed for pathway direction consistency . Media definitions for simulated M9 minimal were consistent with past publications and tryptone media was approximated as amino acids ( Table S1 ) ; the short time of T7 infection meant that media components were in excess for all simulations with growth rate limitations resulted from flux bound constraints . Some regulatory rules were updated to permit growth on rich media ( Table S2 ) . Flux bounds were mostly consistent with previous publications , with the exception of the relevant set used during growth on tryptone amino acids that were fit using growth rates we collected ( Figure S1 ) . Phage stoichiometry reactions were included in the FBA system ( Figure 1B ) , one for each gene”s mRNA and each gene product , as well as for phage genome synthesis and a reaction accounting for degraded host genome dNMP recycling to dNTPs . Included in these reactions are the precursor small molecules that make up each final macromolecule , as well as the energy required for transcription or translation . The FBA host biomass reaction energy requirements are typically phrased in terms of ATP only; to be consistent , the GTP used for energy in phage production processes is included in the reaction stoichiometry as ATP , and the energy requirements for the T7 DNA helicase , which is known to use dTTP preferentially [34] for energy , were also converted to ATP . A full list of assumptions and references for generating phage stoichiometry reactions is in Table S3 . We added a production rate equation consisting of only the positive terms from the net rate equation for each molecular species in the original T7 ODE model , to bound the forward-only reaction fluxes in FBA . Furthermore , another ODE was added to account for the fraction of the host genome material remaining for degradation . A set of input arguments to the T7 ODEs was also introduced to pass limits on one or more of the production rates . If a production rate was limited , its value is accounted for in the net rate equation . Implicit in this implementation is the assumption that if an mRNA or gene product is degraded , the components are not available to metabolism during infection [35] . The code used in preparation of this article is available at . A simplified flowchart of the integrated simulation algorithm is shown in Figure 2B . The FBA and ODE numerical simulations interacted at every 10 seconds of simulation time . Since host lysis is not modeled by the T7 ODEs , there is not a single logical exit criterion for the simulation . Thus the simulation is run for a set time length slightly greater than what is expected to be the productive duration of infection . A text expansion of the integrated simulation algorithm flowchart shown in ( Figure 2B ) follows , with further detail presented in ( Text S1 and Figure S6 ) : Because the T7 ODE kinetic rates do not depend on small molecule concentrations , we bound the phage macromolecule production rates themselves to host production capacity . The method to determine rate limits relies first on an initial ‘viral demand’ which is based on an evaluation of the T7 ODEs without applied limits over the integration time step . Implementation of this strategy takes advantage of the divided matrix formulation of the problem shown in Figure 2A . We further split the matrix ( detail in supplement ) into summed terms representing the small metabolites provided by the host , and those consumed by the viral production fluxes . In the resulting relationship , shown in Eq . 1 ( consistent with convention of FBA intake to organism being negative flux ) , is the solution of the typical host FBA problem neglecting biomass exchange , taking advantage of the fact that host biomass is composed of a superset of the small metabolites consumed by viral reactions . The simplified form of this relationship is enabled by allowance of metabolite accumulation at the intersection of host and viral reactions . ( 1 ) Once a feasible host flux distribution is selected ( by solving for a ‘host supply’ flux distribution ) , Eq . 1 provides a simple relation that must be obeyed by viral production flux rates in order to assure a solution exists to the combined host viral metabolic problem . The method devised to select a vector of maximal viral fluxes or rates ( to pass to T7 ODEs ) is detailed in the supplement , but essentially allows the maximal evenly scaled flux through viral reactions consuming any given metabolic precursor . For example , allowing full production of viral DNA even if amino acid availability is limiting protein synthesis , yet restricting both if a shared reactant such as ATP is limiting .
|
Viral infection is a serious problem with relatively few known solutions . Much of the complexity of viral infection is contributed by the host's own resources that the virus commandeers . Viruses lack the machinery and precursors required to replicate , and thus may be considered metabolic products of their host . Our goal is a systems-level understanding of host-viral metabolic interaction via computational tools and quantitative dynamic measurements . Here we present an integrated model of T7 phage viral replication and host E . coli metabolism that predicts phage production changes across media conditions and provides insight into the underlying limiting factors in T7 replication . The model simulations , supported by our experimental measurements , highlight the role of host metabolism in determining the dynamics of viral infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"theoretical",
"biology",
"biology",
"computational",
"biology",
"microbiology",
"genetics",
"and",
"genomics"
] |
2012
|
Determining Host Metabolic Limitations on Viral Replication via Integrated Modeling and Experimental Perturbation
|
Dengue virus ( DENV ) , a global disease , is divided into four serotypes ( DENV1-4 ) . Cross-reactive and non-neutralizing antibodies against envelope ( E ) protein of DENV bind to the Fcγ receptors ( FcγR ) of cells , and thereby exacerbate viral infection by heterologous serotypes via antibody-dependent enhancement ( ADE ) . Identification and modification of enhancing epitopes may mitigate enhancement of DENV infection . In this study , we characterized the cross-reactive DB21-6 and DB39-2 monoclonal antibodies ( mAbs ) against domain I-II of DENV; these antibodies poorly neutralized and potently enhanced DENV infection both in vitro and in vivo . In addition , two enhancing mAbs , DB21-6 and DB39-2 , were observed to compete with sera antibodies from patients infected with dengue . The epitopes of these enhancing mAbs were identified using phage display , structural prediction , and mapping of virus-like particle ( VLP ) mutants . N8 , R9 , V12 , and E13 are the reactive residues of DB21-6 , while N8 , R9 , and E13 are the reactive residues of DB39-2 . N8 substitution tends to maintain VLP secretion , and decreases the binding activity of DB21-6 and DB39-2 . The immunized sera from N8 substitution ( N8R ) DNA vaccine exerted greater neutralizing and protective activity than wild-type ( WT ) -immunized sera , both in vitro and in vivo . Furthermore , treatment with N8R-immunized sera reduced the enhancement of mortality in AG129 mice . These results support identification and substitution of enhancing epitope as a novel strategy for developing safe dengue vaccines .
Dengue virus ( DENV ) is a mosquito-borne virus that causes prevalent , global disease . It is estimated to cause 390 million infections annually , which lead to a spectrum of clinical syndromes ranging from dengue fever ( DF ) to severe dengue hemorrhagic fever ( DHF ) and dengue shock syndrome ( DSS ) in tropical and subtropical countries [1 , 2] . Although primary infection with DENV provides immunity against the same serotype , subsequent secondary infection with different DENV serotypes has a higher risk for developing severe dengue disease [3–5] . The presence of cross-reactive and non-neutralizing antibodies bound to DENV helps viral entry into Fcγ receptor ( FcγR ) -bearing cells , resulting in increased virus load and/or production of certain cytokines [6] . This phenomenon is termed antibody-dependent enhancement ( ADE ) [3 , 7] . At the time of writing , there is no approved vaccine against DENV infection [8] . DENV , which exists as four related serotypes ( DENV1-4 ) , is a member of the Flavivirus genus within the Flaviviridae family [9] . It is a positive-stranded RNA virus that encodes three structural proteins ( the capsid ( C ) , pre-membrane ( prM ) , and envelope ( E ) proteins ) and seven non-structural proteins ( NS1 , NS2a , NS2b , NS3 , NS4a , NS4b and NS5 ) [10 , 11] . The prM protein seems to function as a chaperone for the assembly of the E protein [12] . During virus maturation , the viral particle is activated by the low pH of the trans-Golgi network ( TGN ) . The prM protein is subsequently cleaved by furin to generate M protein , resulting in mature and infective virions [13] . Co-expression of prM and E proteins can produce virus-like particles ( VLPs ) , which are similar in structure and antigenicity to infectious virus particles , and have been used broadly in epitope mapping , diagnosis , and vaccine development [14–16] . The E protein plays an important role in facilitating attachment of DENV to cell surface receptor ( s ) , fusion of virus with endosomal membranes , and subsequent entry into target cells . The E protein is regarded to be the antigen involved in mediating the immune response , and the principle target of neutralizing antibodies [17] . The E protein forms 90 homodimers on the surface of the mature virion [18] . The E monomer consists of three domains: domain I ( EDI ) , domain II ( EDII ) , and domain III ( EDIII ) [19]; EDI links EDII with EDIII . EDII is an elongated dimerization domain , which contains the conserved fusion peptide [20] . EDIII , an immunoglobulin ( Ig ) -like domain , is considered to be the binding site of the receptor on the target cell . Several studies have shown that serotype-specific and neutralizing mouse monoclonal antibodies bind to EDIII [21–23] , whereas in human , only a small fraction of antibodies react to this region [24 , 25] . Recent study has shown that human neutralizing antibodies bind to complex epitopes on dengue virions [24] . However , a large fraction of cross-reactive and weakly neutralizing human antibodies can be isolated from natural DENV infection [26 , 27] . In the context of dengue pathogenesis , these cross-reactive and non-neutralizing antibodies against E or prM proteins derived from primary infection can enhance viral infection through ADE during secondary infection [26] . Therefore , identification of B-cell epitopes of DENV E protein , which induce cross-reactive and non-neutralizing antibodies , may provide valuable information for vaccine development . Although various strategies have been employed in an attempt to develop dengue vaccine ( including the use of attenuated or inactive virus , and the development of subunit vaccines ) , a safe and effective vaccine against DENV is not yet available [28] . Thus , there is a need to identify and substitute the epitopes recognized by poorly neutralizing and highly enhancing antibodies to improve the dengue vaccine . In this study , we found that the cross-reactive mAbs DB21-6 and DB39-2 exhibit poor neutralizing activity and high capacity for enhancing DENV infection . We used competitive enzyme-linked immunosorbent assay ( ELISA ) to determine the relationship between mAbs and sera antibodies from dengue patients . We proceeded to use phage display , bioinformatic analysis , and VLP mutants to identify the epitopes recognized by DB21-6 and DB39-2 . To further improve the DNA vaccines against DENV2 , we substituted the N8 residue of wild-type ( WT ) DENV2 E protein with arginine ( N8R ) in a plasmid for immunization . N8R-immunized sera produced higher neutralizing and protective activity than WT-immunized sera . Moreover , treatment of AG129 mice with N8R-immunized sera reduced mortality , as compared with mice treated with WT-immunized sera . Taken together , we have identified a novel cross-reactive and infection-enhancing epitope in E protein . Our results demonstrate that substitution of this enhancing epitope is a promising strategy for development of a safe dengue vaccine .
Mouse experiments were carried out in accordance with strict guidelines from the Care and Use Manual of the National Laboratory Animal Center , Taiwan . The protocol was approved by the Committee on the Ethics of Animal Experiments of Academia Sinica ( Permit Number: 11-04-166 ) . The human serum samples were collected during an outbreak between 2002 and 2003 in Taiwan . The study protocol was approved by the National Taiwan University Institutional Review Board ( NTUH-REC No . 200903086R ) . The written informed consent was obtained , and all human serum samples were coded for anonymity . Four dengue virus serotypes , DENV1 Hawaii , DENV2 16681 , DENV3 H87 , and DENV4 H241 , were prepared as previously described [29] . C6/36 cells were grown in medium consisting of 50% Mitsumashi and Maramorsch insect medium ( Sigma-Aldrich ) plus 50% Dulbecco’s modified Eagle’s medium ( DMEM , Gibco ) containing 10% fetal bovine serum ( FBS , Gibco ) and 100 U/ml penicillin , 100 μg/ml streptomycin , and 0 . 25 μg/ml amphotericin B ( Antibiotic-Antimycotic , Gibco ) . The C6/36 cells were infected with DENV at a multiplicity of infection ( MOI ) of 0 . 1–1 , and incubated at 28°C for 7 to 9 days . The viruses were harvested from supernatant , and then titrated in a baby hamster kidney fibroblast cell line ( BHK-21 ) by plaque assay . The aliquots were stored at -80°C . BHK-21 cells were grown in minimal essential medium ( MEM , Gibco ) supplemented with 10% FBS , 100 U/ml penicillin , 100 μg/ml streptomycin , and 0 . 25 μg/ml amphotericin B ( Antibiotic-Antimycotic , Gibco ) . Human erythroleukaemic K562 and monocytic THP-1 cells were grown in RPMI medium ( Gibco ) containing 10% FBS . The mouse mAbs , including DB21-6 and DB39-2 , were generated by immunization of BALB/c mice with DENV2 , and were produced in hybridoma cells , as previously described [21] . DB21-6 and DB39-2 were isotyped as IgG1 ( SouthernBiotech ) and purified using protein G Sepharose 4B gels ( GE Healthcare ) . Serial dilutions of mAbs were incubated with DENV1 Hawaii ( MOI = 1 ) , DENV2 16681 ( MOI = 1 ) , DENV3 H87 ( MOI = 5 ) , and DENV4 H241 ( MOI = 1 ) for 1 hour at 4°C . The mixtures were then used to infect K562 cells for 2 hours at 37°C . After washing , the cells were incubated with 2% FBS in RPMI medium ( Gibco ) at 37°C for 3 days . The infected cells were collected and fixed with 3 . 7% formaldehyde for 10 minutes at 4°C . For staining , the cells were permeabilized with 2% FBS in PBS containing 0 . 1% saponin ( Sigma ) , followed by staining with 4 μg/ml 4G2 for 0 . 5 hours at 4°C . The cells were washed and incubated with R-phycoerythrin ( RPE ) -conjugated goat anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) for 1 hour at 4°C . The cells were washed , and the percentages of infected cells were determined by flow cytometry . For infection of THP-1 cells , DENV2 16681 ( MOI = 1 or 10 ) was incubated with diluted mAbs for 1 hour at 4°C , and then incubated with cells for 2 hours at 37°C . After 3 days , the cells were fixed , permeabilized , and stained with hDB32-6 [21] . After washing , the cells were incubated with an RPE-conjugated goat anti-human IgG ( Jackson ImmunoResearch Laboratories ) , and were subsequently analyzed by flow cytometry . Type I and II interferon receptor-deficient mice ( AG129; 5- to 6-weeks-old ) were purchased from B&K Universal . The AG129 mice were given intraperitoneal ( i . p . ) injections of 5 μg mAbs in 200 μl PBS on days -1 and 1 . The mouse IgG1 isotype antibody was used as a negative control . On day 0 of infection , mice were intravenously ( i . v . ) inoculated with 1 × 105 pfu of the mouse-adapted DENV2 S221 ( obtained from Sujan Shresta ) [30] , in 100 μl PBS . The survival rates of AG129 mice were recorded for 30 days . AG129 mice were infected with 1 × 105 pfu DENV2 S221 by i . v . inoculation on day 0 , and treated with 5 μg mAbs via i . p . injection on days -1 and 1 . Viral RNA was extracted from pooled and infected mice sera using the QIAamp viral RNA minikit ( Qiagen ) . Quantitative RT-PCR was performed based on previously published procedures [31] , using the LightCycler 480 system ( Roche ) . The standard curve was generated with DENV2 S221 ( at concentrations from 101 to 107 pfu/ml ) . Viremia measurements were expressed as pfu equivalents/ml , which was calculated based on the threshold cycle value ( Ct ) according to the standard curve for DENV2 S221 . A total of 21 DENV2-infected patient serum samples were collected from 11 DF patients and 10 DHF patients during an outbreak between 2002 and 2003 in Taiwan . Diagnosis of DENV infection was based on IgM antibody-capture ELISA ( MAC-ELISA ) , reverse-transcriptase PCR ( RT-PCR ) , or virus isolation in cell cultures , as previously described [15] . These serum samples were collected between days 4 and 22 from the onset of symptoms; such sera contained anti-dengue antibodies . All of these patients were determined to have classical DF or DHF based on the criteria published by the World Health Organization ( WHO ) in 2009 [32] . The characteristics of patient serum samples enrolled in this study are also provided ( S1 Table ) . Competitive ELISA was performed as previously described [33] . Briefly , the plates were coated with polyclonal rabbit anti-DENV hyper-immune sera at 4°C overnight . After blocking , the diluted DENV2 viral supernatants ( 1 × 106 pfu ) were added for 2 hours at room temperature ( RT ) . The diluted mAbs and patient sera ( 1:100 dilution ) were incubated for 2 hours at RT . After washing , horseradish peroxidase ( HRP ) -conjugated anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) was added for 1 hour at RT . The peroxidase substrate o-phenylenediamine dihydrochloride ( OPD , Sigma-Aldrich ) was then added , and the reaction was stopped with 3N HCl . The optical density ( OD ) was measured at 490 nm . Normal human serum ( NHS ) was used as a control . The percentage of competition was calculated as follows: competition ( % ) = [1− ( OD of patient serum-mAb mixture/OD of NHS-mAb mixture ) ] × 100 . Phage display biopanning was performed as previously described [21] . Briefly , the plate was coated with 100 μg/ml mAbs at 4°C for 6 hours . After washing and blocking , 4 × 1010 pfu of phage-displayed peptide library ( New England BioLabs , Inc . ) were incubated for 50 mins at RT . After washing , bound phage was eluted with 100 μl 0 . 2 M glycine/HCl ( pH 2 . 2 ) and neutralized with 15 μl 1 M Tris/HCl ( pH 9 . 1 ) . The eluted phage was then amplified in ER2738 for subsequent rounds of selection . The phage was titrated onto LB plates containing IPTG and X-Gal . The second and third rounds of selection were identical to the first round except for the addition of 2 × 1011 pfu of amplified phage . The plate was coated with 50 μg/ml mAbs . After washing and blocking , the amplified phages were added , and incubated for 1 hour at RT . After washing , diluted HRP-conjugated anti-M13 antibody ( GE Healthcare ) was added at RT for 1 hour . The plates were developed , and subsequently terminated by 3N HCl . The OD was measured at 490 nm . The pCBD2-2J-2-9-1 plasmid expressing prM-E proteins of DENV2 has been previously characterized and described [14–16] . Site-directed mutagenesis was performed to replace each of the selected amino acid residues , as described in the previous study [21] . After mutagenesis , the plasmids were sequenced to ensure the absence of any further mutations at non-target sites . BHK-21 cells were transfected with constructs expressing the wild-type ( WT ) or mutant DENV2 E protein using polyjet in vitro DNA transfection reagent ( SignGen Laboratories ) . After 2 days , the cells were fixed , and permeabilized with 2% FBS in PBS containing 0 . 1% saponin ( Sigma ) . For staining , cells were incubated with DB21-6 , DB39-2 , 4G2 , and mixed mAbs ( DB32-6 , 3H5 , and DB25-2 ) at a concentration of 1 , 1 , 1 , and 1 μg/ml , respectively , at 4°C for 0 . 5 hours . After washing , the cells were incubated with RPE-conjugated goat anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) , and analyzed by flow cytometry . The relative index of a mAb to a mutant E protein was measured using the formula: [intensity of the mutant E/intensity of WT E ( recognized by a mAb ) ]/[intensity of mutant E/intensity of WT E ( recognized by mixed mAbs ) ] . BHK-21 cells were transfected with vectors expressing the WT or mutant E protein of DENV2 , as described above . At 48 hours post-transfection , culture supernatants were collected . The plates were coated with polyclonal rabbit anti-DENV hyper-immune sera at 4°C overnight . After blocking , two-fold dilutions of supernatants containing WT or mutant VLPs were added for 2 hour at RT . The wells were then incubated with diluted DB32-6 and 4G2 at RT for 2 hour . After washing , a 1:2000 dilution of HRP-conjugated anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) was added for 1 hour at RT . Finally , the plates were developed , and the reaction was subsequently terminated with 3N HCl . The OD was measured at 490 nm . Plasmids expressing WT E protein of DENV2 or a mutant E protein in which the N8 residue was substituted with R ( N8R ) were used for immunization . For coating , 25 mg of 1 . 0 μm gold powder was resuspended with 50 mM spermidine ( Sigma-Aldrich , St . Louis , MO ) . Then , 50 μg of plasmid DNA was added , followed by the addition of 1M CaCl2 ( Sigma-Aldrich , St . Louis , MO ) ; the solution was mixed and precipitated for 10 mins at RT . After collection by centrifugation , the gold-DNA complex was washed with absolute ethanol and resuspended in 0 . 1 mg/ml of polyvinylpyrrolidone ( PVP ) ( 360 kDa; Sigma Chemicals , Inc . ) solution . The slurry was injected into a TefzelR tube ( McMaster-Carr , Chicago , IL ) , and then coated . After the ethanol had dried off , the tube was cut into 0 . 5-inch bullets and stored at -20°C . The gold in each bullet contained 1 μg of DNA . Before use , the bullets were loaded into the Helios gene gun device ( Bio-Rad , Hercules , CA ) for delivery of plasmids . The abdominal epidermis of 6 week-old female BALB/c mice was injected with a gene gun using a helium pressure setting of 400 lb/inch2 . Each mouse was immunized by administering 4 bullets containing 1 μg plasmid DNA . Mice were immunized at 0 , 3 , and 6 weeks . Serum samples were collected before immunization and 3 weeks after the third immunization ( pre- , 1st , 2nd , 3rd immunized sera ) . The serum samples were pooled from five to six mice for each immunized group and evaluated by ELISA , neutralization assay , and in vivo ADE assay . C6/36 cells infected with DENV2 16681 were used as antigens . C6/36 cells were seeded into each well ( 2 × 104 cells/well ) of 96-well ELISA plates . After one day , 2 × 103 pfu of DENV2 16681 ( MOI = 0 . 1 ) was added to infect the cells at 37°C for 2 hours . The wells were washed with PBS , and then cultured in 2% FBS culture medium at 28°C for 5 days . Next , the infected cells were fixed with 1:1 methanol/acetone at 4°C for 10 mins . The plates were blocked with 5% skimmed milk at 4°C for 24 hours . Diluted immunized sera were then added for incubation at RT for 2 hours . The plates were then washed three times with phosphate-buffered saline containing 0 . 1% ( w/v ) Tween 20 ( PBST0 . 1 ) , and subsequently incubated with HRP-conjugated anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) . Finally , the plates were developed , and the reaction terminated with 3N HCl . The OD was measured at 490 nm . DENV2 16681 ( MOI = 1 ) was incubated with 3rd immunized sera for 1 hour at 4°C . Next , the mixtures were used to infect BHK-21 cells for 2 hours at 37°C . After 3 days , the cells were fixed , permeabilized , and stained with 4 μg/ml 4G2 . After washing , the cells were incubated with RPE-conjugated goat anti-mouse IgG ( Jackson ImmunoResearch Laboratories ) , and analyzed by flow cytometry . Inhibition percentage ( % ) = [1− ( the percentage of infected cells incubated with immunized sera/without immunized sera ) ] × 100 . The ICR mice were purchased from the Laboratory Animal Center , National Taiwan University College of Medicine . Serially-diluted immunized sera were incubated with 1 × 104 pfu ( 25-fold lethal dose , 25-fold LD50 ) of DENV2 16681 for 0 . 5 hours at 4°C . Two-day-old suckling mice were inoculated with 20 μl of the mixtures through intracranial ( i . c . ) injection . After challenge , the survival rates were recorded for 28 days . AG129 mice were given i . p . injections of dilutions of immunized sera on days -1 and 1 , and were i . v . inoculated with 1 × 105 pfu of DENV2 S221 on day 0 . The survival rates were recorded for 30 days . Survival rate was expressed using Kaplan-Meier survival curves , and statistical analyses were performed using GraphPad Prism 5 . For competition assays of mAbs and patient sera , Student’s t tests were used to identify significant differences and calculate P values ( *P<0 . 05 , ***P<0 . 001 , NS not significant ) . For evaluation of immunized sera against DENV2 by ELISA , two-way ANOVA with Bonferroni post-hoc test was used to determine the significant differences and calculate P values ( **P<0 . 01 , NS not significant ) . GraphPad Prism 5 was used to analyze 50% inhibition titers against DENV2 , based on inhibition percentages from pooled immunized sera .
In our previous study , we generated seventeen mAbs against the E protein of DENV [21] . Of these mAbs , DB21-6 and DB39-2 could recognize cells infected with DENV1-4 ( S1A Fig ) . In addition , these mAbs recognized transfected BHK-21 cells expressing DENV2 E and EDI-II proteins ( S1B and S1C Fig ) . Thus , the cross-reactive DB21-6 and DB39-2 recognized DENV1-4 and domain I-II on E protein . To estimate the in vitro neutralizing activity , we infected BHK-21 cells with a mixture of individual mAbs and DENV1-4 . Previous studies have reported that 4G2 is an anti-flavivirus antibody with neutralizing and enhancing activity at certain concentrations [16] . We observed that 4G2 exerts higher neutralization activity than DB21-6 and DB39-2 against DENV2 ( S2 Fig ) . In addition , DB21-6 and DB39-2 exhibited non-neutralizing activity against DENV1-4 ( 50% inhibition concentration , >33 μg/ml ) ( S2B Fig ) . To investigate in vitro enhancement of DENV infection through ADE [34 , 35] , we performed in vitro ADE assays , and detected the increases in the percentage of dengue-infected cells by flow cytometry [36] . The FcγRIIA-bearing K562 cells , which do not express type 1 interferon ( IFN ) [37] , were used to measure the enhancement of infected cells through extrinsic ADE . The serially-diluted mAbs were incubated with DENV1-4 , and then used to infect K562 cells . The infection percentage was measured by flow cytometry , revealing infection enhancement over a broad range of mAb concentrations ( Fig 1A ) . As compared to the other mAbs , 4G2 caused enhancement of DENV1-4 infection in K562 cells at lower antibody concentrations . DB21-6 and DB39-2 enhanced DENV1-4 infection in K562 cells at high antibody concentrations ( Fig 1A ) . To further confirm enhancement of infection , we proceeded to examine the enhancement of DENV2 16681 infection by DB21-6 and DB39-2 in FcγRI- and FcγRIIA-bearing THP-1 cells . Infection in THP-1 cells was enhanced to a greater extent by DB21-6 and DB39-2 than by 4G2 ( Fig 1B ) . DENV2 S221 was previously used to study enhancement of mortality via ADE in AG129 mice [30] . To evaluate the in vitro enhancement of DENV2 S221 infection by mAbs , we performed ADE assays using K562 cells and THP-1 cells . As for DENV1-4 infection , high concentrations of DB21-6 and DB39-2 enhanced DENV2 S221 infection in K562 cells ( S3A Fig ) . In addition , DB21-6 and DB39-2 enhanced DENV2 S221 infection in THP-1 cells at high concentrations of antibody ( S3B Fig ) . These results suggest that DB21-6 and DB39-2 can enhance DENV2 S221 infection in vitro . Next , we confirmed the in vivo enhancing activities in AG129 mice . The AG129 mice treated with 5 μg DB21-6 and infected with DENV2 S221 exhibited increased mortality as compared to control infected mice ( Fig 2A ) . In addition , AG129 mice treated with 5 μg of DB39-2 also exhibited elevated mortality ( Fig 2B ) . In order to determine viremia in DENV2 S221-infected AG129 mice following treatment with DB21-6 or DB39-2 , the viral RNA levels were measured by quantitative RT-PCR . The results indicate that viral loads were significantly increased after DB21-6 or DB39-2 treatment of infected AG129 mice , as compared to isotype control Ab treatment ( Fig 2C ) . These results indicate that DB21-6 and DB39-2 have non-neutralizing activities , and enhance mortality in AG129 mice . We proceeded to perform competition assay to determine whether sera antibodies from dengue patients compete with mAbs for binding to DENV2 . The characteristics of patient serum samples enrolled in this study are provided ( S1 Table ) . The sera antibodies from infected patients were observed to compete with DB21-6 and DB39-2 . The competition percentages of DB21-6 and DB39-2 were significantly higher in serum samples from DHF patients than those from DF patients ( Fig 2D ) , while the competition percentage of neutralizing DB32-6 [21] was similar for sera from either DF or DHF patients ( Fig 2D ) . We also performed the same experiment with more concentrated serum ( 1:50 dilution ) or diluted serum ( 1:200 dilution ) , and obtained similar results ( S4 Fig ) . These results suggest that serum samples from DHF patients contain higher levels of antibodies , which compete for binding with DB21-6 and DB39-2 mAbs . In order to identify the enhancing epitopes of DB21-6 and DB39-2 , we used a phage-displayed peptide library to screen the reactive phage clones . After three biopanning rounds , the phage titers were increased to 12 , 871-fold ( DB21-6 ) and 5 , 000-fold ( DB39-2 ) , respectively , compared to that of the first round ( Fig 3A ) . The individual phage clones from the third round of biopanning were randomly selected . As shown by ELISA , most selected phage clones exhibited significant reactivity to the mAbs , but not to normal mouse IgG ( NMIgG ) . Of the 30 selected phage clones , 29 clones reacted with DB21-6 ( Fig 3B ) . The immunopositive phage clones were amplified , and their phage DNA was isolated for DNA sequencing . Eleven phage clones with individual peptide sequences were identified ( Table 1 ) . Similarly , of the 47 selected phage clones , 46 reacted with DB39-2 ( Fig 3B ) . Thirteen of the 46 immunopositive phage clones that reacted with DB39-2 possess individual peptide sequences ( Table 1 ) . Alignment of peptide sequences revealed the binding motif of DB21-6 and DB39-2 to be N-R-x-x-V-E ( Table 1 ) . In addition , modeling of the peptide sequences with the pepitope server ( http://pepitope . tau . ac . il/ ) predicted that the epitope residues on the E protein are N8 , R9 , V12 , and E13 ( Table 1 ) . To further verify the epitope of DB21-6 and DB39-2 , we performed site-directed mutagenesis of the phage-displayed epitope using pCBD2-2J-2-9-1 as template . After confirmation of variants by sequencing , we transfected cells with the mutant plasmids , and detected binding activity by flow cytometry . The binding percentages for each transfectant were normalized to those of anti-EDIII mAbs ( DB32-6 , 3H5 , and DB25-2 ) [21] , and relative indices were calculated ( Fig 4A ) . 4G2 , which binds to residues at the fusion loop of EDII [16] , was used as a control to verify the structural change of E proteins caused by mutations ( Fig 4A ) . Based on the relative indices , we found that mutations at N8 , R9 , V12 , and E13 prevented binding by DB21-6 . The same method was used to identify the epitope residues of DB39-2 as N8 , R9 , and E13 . Structural modeling was applied to show that the recognition residues are located in domain I of E protein ( Fig 4B ) . The distance between these residues from the same monomers was analyzed using a structure modeling program , and was found to be less than 30°A ( Fig 4C ) ; interestingly , this distance can be spanned by a single IgG molecule [16] . This suggests that the N8 , R9 , V12 , and E13 residues constitute the epitope of DB21-6 . In addition , the N8 , R9 , and E13 residues constitute the epitope of DB39-2 . Alignments revealed that the binding motif of DB21-6 and DB39-2 corresponds to the N8 , R9 , V12 , and E13 residues , which are conserved in DENV1-4 ( S2 and S3 Tables ) . Finally , we used VLP-capture ELISA to demonstrate that the mutations at R9 , V12 , and E13 affect DENV2 VLP secretion ( Fig 4D ) . The effects of these mutations on the ability to secrete VLPs might be due to a change in the structure of E protein . However , the N8R substitution did not affect DENV2 VLP secretion ( Fig 4D ) . N8 substitution tends to maintain VLP secretion and reduces the binding activity of DB21-6 and DB39-2 . The BALB/c mice were immunized with vector , WT , or N8R plasmids at 0 , 3 , and 6 weeks . After three rounds of immunization , the serum samples were collected and pooled within each immunized group . Next , the immunized sera were examined by ELISA . A remarkable increase of antibody titer against DENV2 was observed after immunization ( S5A Fig ) . The 3rd WT- and N8R-immunized sera against DENV2 exhibited significantly higher absorbance values than those of vector-immunized sera ( Fig 5A ) . Analysis of immunized sera with anti-IgG1 and IgG2a antibodies revealed that the IgG1/IgG2a ratios increased between the second and third immunization ( S5B and S5C Fig ) . In addition , the immunized mice maintained their anti-DENV2 responses after 15 weeks ( S5D Fig ) . The immunized sera were evaluated for their neutralizing activity against DENV2 . Both WT- and N8R-immunized sera exhibited high neutralizing activities , while vector-immunized sera did not ( Fig 5B ) . Interestingly , DENV2 infection was more effectively neutralized by N8R-immunized sera than by WT-immunized sera ( Fig 5C ) . To further evaluate whether immunized sera could broadly neutralize the diverse DENV2 strains , BHK-21 cells were infected with mixtures of immunized sera and four different DENV2 strains: 16681 , NGC , PL046 , and Malaysia 07587 . Remarkably , the WT- and N8R-immunized sera exhibited high neutralizing activities against various types of DENV2 strain ( S6 Fig ) . Next , we examined the protective effect of immunized sera against DENV2 16681 in vivo . The survival rates of mice treated with WT-immunized sera at dilutions of 1:100 and 1:200 were significantly higher than that of mice treated with vector-immunized sera at a dilution of 1:100 ( Fig 5D ) , while the survival rates of mice treated with N8R-immunized sera at dilutions of 1:100 , 1:200 , and 1:400 were significantly higher than that of mice treated with vector-immunized sera at a dilution of 1:100 ( Fig 5E ) . In addition , treatment with WT-immunized sera afforded 50% protection at a dilution of 1:200 , while N8R-immunized sera afforded 50% protection at a dilution of 1:400 ( Fig 5D and 5E ) . Hence , N8R-immunized sera possessed higher neutralizing and protective activity than WT-immunized sera both in vitro and in vivo . In order to study the in vivo enhancement of mortality , we passively transferred different dilutions of WT- , N8R- , or vector-immunized sera into AG129 mice . Following infection with DENV2 S221 , the survival rate of mice treated with WT- or N8R-immunized sera ( 1:25 dilution ) was higher than that of mice treated with vector-immunized sera ( Fig 6A ) . However , mice treated with WT-immunized sera at a dilution of 1:100 showed higher mortality than mice treated with vector-immunized sera ( Fig 6B ) . Notably , the survival rate of mice treated with N8R-immunized sera at a dilution of 1:100 was higher than that of mice treated with vector-immunized sera ( Fig 6B ) . In addition , no enhancement of mortality was observed in mice treated with N8R-immunized sera ( Fig 6B ) . Finally , treatment with WT- or N8R-immunized sera at a dilution of 1:400 did not have a neutralizing or enhancing effect on the survival rates of mice ( Fig 6C ) . These results indicate that the N8R substitution of E protein can reduce in vivo enhancement of mortality . To further characterize these enhancing antibodies are produced in immunized sera , we performed competitive ELISA to inhibit the binding of HRP-conjugated DB21-6 or DB39-2 mAbs by immunized sera ( Fig 6D ) . The competition percentages of HRP-conjugated DB21-6 and DB39-2 were significantly higher in WT-immunized sera than those in N8R-immunized sera ( Fig 6E ) . These results suggest that N8R substitution would redirect immunodominance by reducing the generation of enhancing antibodies .
DENV infections stimulate immune responses and elicit a small proportion of protective antibodies . However , a high proportion of non-protective antibodies are also generated , which might be associated with enhancement of viral infections . Here , we characterized the ability of DB21-6 and DB39-2 to increase the percentage of dengue virus-infected cells . Furthermore , we confirmed that these mAbs enhance mortality in AG129 mice . Through competition assay , we found that sera antibodies from infected patients compete for binding with these mAbs . Using phage-display , structure prediction , and VLP mutants , we mapped the epitopes of enhancing mAbs DB21-6 and DB39-2 on EDI protein . To investigate how to reduce the enhancing effects while maintaining neutralizing activity , we substituted the N8 residue of E protein , and immunized mice with WT or N8R plasmids with a gene gun delivery system . After three immunizations , N8R-immunized sera produced neutralizing activity against DENV2 , and reduced enhancement of mortality as compared to WT-immunized sera . Thus , substitution of enhancing epitope residues can increase the immune response against viral infection while reducing the potential for ADE . The antibodies induced by E protein of DENV play important roles in neutralizing effects and regulation of viral infection [21 , 38–40] . There are three structural domains ( domain I , II , and III ) in E protein . In previous reports , some mouse mAbs that bind to domain III of E protein were found to exhibit neutralizing activity and obstruct viral infection [21–23] . However , the anti-E or prM antibodies are cross-reactive and weakly neutralizing , which may enhance viral infection through ADE [26 , 36] . Here , we demonstrated that cross-reactive DB21-6 and DB39-2 against EDI-II have poor neutralizing activities against DENV1-4 ( S2 Fig ) . In addition , we found that DB21-6 and DB39-2 have strong ADE activities in vitro ( Fig 1 ) . Previous studies have shown that anti-fusion loop 4G2 enhances viral infections in both in vitro ADE assays and AG129 mice [41] . We also observed that 4G2 has partially neutralizing activity against DENV1-4 ( S2 Fig ) and enhances in vitro viral infections at low antibody concentrations ( Fig 1A ) . Notably , we also found that DB21-6 and DB39-2 enhanced DENV1-4 infection in K562 cells at high concentrations ( Fig 1A ) . Furthermore , infection of DENV2 was enhanced to a greater extent by DB21-6 and DB39-2 than by 4G2 in THP-1 cells ( Fig 1B ) . In addition , DB21-6 and DB39-2 enhanced mortality in AG129 mice ( Fig 2A and 2B ) and increases the viral loads in infected mice sera ( Fig 2C ) . These results indicate that DB21-6 and DB39-2 have strong enhancing activity both in vitro and in vivo . ADE is regarded as an important mechanism leading to the development of severe dengue disease , including DHF/DSS [5] . Cross-reactive and non-neutralizing antibodies binding to viruses can enhance infection of FcγR-bearing cells by ADE , resulting in increased viral load and/or production of cytokines [6] . High viral load is correlated with dengue disease severity and DHF [42 , 43] . Thus , there is a need to be able to confirm the presence of enhancing antibodies in dengue patient sera . Our results indicate that the competition percentages of DB21-6 and DB39-2 were significantly higher in DHF patient sera than those in DF patient sera ( Fig 2D ) , suggesting that the higher levels of enhancing antibodies , DB21-6 and DB39-2 , in serum samples of dengue patients are associated with severe dengue disease . We hypothesize that the DENV infected patients might suffer more severe symptoms , such as DHF , when the expression level of the enhancing antibodies is higher . Identification of binding domain and epitope residues in the E protein may provide helpful information for investigation of neutralizing and enhancing mechanisms of dengue infection . Phage display is a powerful method for developing epitope-based diagnostics and identifying B-cell epitopes [21 , 44] . After screening a phage-displayed peptide library , we found that the phage clones selected using DB21-6 and DB39-2 mAbs displayed peptide sequences containing a consensus motif , N-R-x-x-V-E ( Table 1 ) . These displayed peptide sequences may be suitable for detecting enhancing antibodies in serum samples from dengue patients , and for providing information on the pathogenesis of dengue . By alignment of displayed peptide sequences and structural modeling , the candidate epitopes were predicted and verified using VLP mutants ( Figs 3 and 4 ) . The epitope residues of enhancing mAb DB21-6 are N8 , R9 , V12 , and E13 in domain I of DENV2 E protein ( Fig 4A and 4B ) , and the epitope residues of enhancing mAb DB39-2 are N8 , R9 , and E13 in domain I of DENV2 E protein ( Fig 4A and 4B ) . We aligned the N8 , R9 , V12 , and E13 residues , and found that these residues were conserved in DENV1-4 ( S2 Table ) . Thus , cross-reactive DB21-6 and DB39-2 can bind to DENV1-4 . A previous report indicated that G106 and L107 are the epitope residues of enhancing mAb 4G2 [16] . In our studies , we also confirmed that W101 , G106 , L107 , and F108 in the fusion loop are the epitope residues of 4G2 ( Fig 4A ) . The epitope residues recognized by 4G2 are different from those recognized by DB21-6 and DB39-2 . These findings suggest that DB21-6/DB39-2 and 4G2 enhance DENV infection through different mechanisms . In addition , the enhancing epitopes of DB21-6 ( N8 , R9 , V12 , and E13 ) and DB39-2 ( N8 , R9 , and E13 ) are novel and have not previously been reported . Therefore , further verification of these enhancing epitopes and the detailed molecular mechanism ( s ) by which these enhancing antibodies propagate dengue infection are worth investigating through cryo-electron microscopy ( cryo-EM ) . The E protein is targeted by most reported dengue vaccines , and is thus regarded as an important target [28] . Sanofi Pasteur published data from a phase III study on tetravalent dengue vaccine , which conferred moderate protection ( 56% ) against dengue disease [45] . Furthermore , the vaccine provided low protection ( 35% ) against DENV2 , but more than 75% protection against DENV3 and 4 , and 50% against DENV1 . Improvements in vaccine efficacy and the effect of the substitution of the enhancing epitope on safety are yet to be examined . Previous studies have shown that DNA vaccine candidates against DENV1 or DENV2 with substitutions in the fusion loop ( at G106 and L107 ) and the cross-reactive epitopes of EDIII ( at K310 , E311 , and P364 ) confer protective immunity [46 , 47] . In addition , enhancement of mortality by enhancing antibodies against the fusion loop was reduced in mice immunized with such vaccines . In this study , we have identified new enhancing antibodies and a novel enhancing epitope that are different from those previously reported . Mutations at R9 , V12 , and E13 may change the structure of E protein and affect VLP secretion . However , our VLP-capture ELISA results suggest that the N8R substitution does not affect DENV2 VLP secretion ( Fig 4D ) , which is crucial for its use in immunization . Moreover , we used an in vitro neutralizing assay and in vivo protection assays to show that both WT- and N8R-immunized sera exerted protective activities against DENV2 ( Fig 5 ) . Interestingly , N8R-immunized sera had higher in vitro neutralizing activity and in vivo protective activity than the WT-immunized sera ( Fig 5B–5E ) . These results suggest that immunization with the N8R DNA vaccine may increase neutralizing and protective immunity against DENV2 . An earlier investigation used mouse-adapted DENV2 S221 to study severe dengue disease via ADE in AG129 mice [30] . Here , we passively transferred diluted vector- , WT- , or N8R-immunized sera , and then challenged AG129 mice with DENV2 S221 . WT- and N8R-immunized sera were protective at a 1:25 dilution , as compared to vector-immunized sera . However , the mortality of mice was enhanced by treatment with WT-immunized sera at a 1:100 dilution , as compared to treatment with vector-immunized sera . Importantly , the mortality of mice treated with N8R-immunized sera at a 1:100 dilution was not enhanced ( Fig 6B ) . When the dilution was increased to 1:400 , no enhanced mortality was observed ( Fig 6C ) . Our results indicate that substituting the enhancing epitope can reduce the ADE phenomenon and increase protective activity in vivo . In this study , substitution of enhancing epitope and preservation of neutralizing epitope in immunized mice provide protective immunity . Such an approach would redirect immunodominance ( Fig 6E ) and improve immunogenicity by satisfying the required neutralizing occupancy [48] . In summary , we have identified a novel enhancing epitope , enabling us to reduce the potential for ADE through N8R substitution in DENV2 E protein . This may be a viable approach for developing new dengue vaccines that can increase the anti-DENV immune response .
|
Dengue virus ( DENV ) infects 390 million humans annually , and is the cause of one of the most important arthropod-borne viral diseases in the world . Currently , there are no available licensed vaccines or antiviral drugs for dengue , so development of safe vaccine and effective therapy is urgently needed . Here , we identified two monoclonal antibodies , DB21-6 and DB39-2 , which can enhance DENV1-4 infection and increase virus-induced mortality in AG129 mice . We found that serum samples from patients with severe dengue disease contain higher levels of antibodies against enhancing epitope . We proceeeded to identify enhancing epitope on E protein , and developed DNA vaccines by substitution . The substituted DNA vaccine with mutation at the enhancing epitope demonstrated augmented neutralizing activity against DENV2 , and reduced enhancement of mortality as compared to wild type-immunized sera . Our results show that substitution of enhancing epitope can increase the immune response against viral infection , while reducing the potential for antibody-dependent enhancement ( ADE ) . These novel findings may be useful for developing safe and efficacious vaccines against dengue .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
An Epitope-Substituted DNA Vaccine Improves Safety and Immunogenicity against Dengue Virus Type 2
|
Nonessential tRNA modifications by methyltransferases are evolutionarily conserved and have been reported to stabilize mature tRNA molecules and prevent rapid tRNA decay ( RTD ) . The tRNA modifying enzymes , NSUN2 and METTL1 , are mammalian orthologs of yeast Trm4 and Trm8 , which are required for protecting tRNA against RTD . A simultaneous overexpression of NSUN2 and METTL1 is widely observed among human cancers suggesting that targeting of both proteins provides a novel powerful strategy for cancer chemotherapy . Here , we show that combined knockdown of NSUN2 and METTL1 in HeLa cells drastically potentiate sensitivity of cells to 5-fluorouracil ( 5-FU ) whereas heat stress of cells revealed no effects . Since NSUN2 and METTL1 are phosphorylated by Aurora-B and Akt , respectively , and their tRNA modifying activities are suppressed by phosphorylation , overexpression of constitutively dephosphorylated forms of both methyltransferases is able to suppress 5-FU sensitivity . Thus , NSUN2 and METTL1 are implicated in 5-FU sensitivity in HeLa cells . Interfering with methylation of tRNAs might provide a promising rationale to improve 5-FU chemotherapy of cancer .
5-Fluorouracil ( 5-FU ) is a pyrimidine analog and is the most widely used chemotherapeutic agent for the treatment of a variety of solid cancers . Its mechanism of action has been attributed to the production of cytotoxic metabolites incorporated into RNA and DNA and inhibiting thymidylate synthase , finally leading to cell cycle arrest and apoptosis in cancer cells [1] . 5-FU is used against cancer for about 40 years and it is known that systemic administration of 5-FU might result in drug resistance of tumor cells . Furthermore , treatment regimens with increased dosage of 5-FU have been reported to cause severe side effects such as myelosuppression , mucositis , dermatitis and diarrhea . , In order to address this dilemma , different strategies were pursued to improve outcomes for patients and to reduce side effects of 5-FU therapy [2]–[9] . However , also with current approaches , there is still a need to develop new compounds or novel strategies by which cancer cells are killed more effectively and more selectively [10]–[12] . Overexpression of tRNA-modifying enzymes NSUN2 and METTL1 is widely observed among human cancers [13]–[16] . NSUN2 ( NOP2/Sun domain family , member 2 ) , also known as SAKI ( Substrate of AIM-1/Aurora kinase B ) , is a NOL1/NOP2/SUN domain-containing tRNA ( cytosine-5- ) -methyltransferase . It is phosphorylated at Ser139 by Aurora-B to inhibit its enzymatic activity during mitosis [17] . Trm4 , a yeast Saccharomyces cerevisiae homologue of human NSUN2 , participates in the nonessential modification of tRNA [18] , [19] , and a yeast mutant deficient in Trm4 shows no defect in cell growth and has normal sensitivities to various stresses [18] , [19] . On the other hand , another tRNA modification enzyme Trm8 , which is also nonessential and catalyzes tRNA 7-methylguanosine modification [20] , acts together with Trm4 to stabilize tRNA under heat stress [21] . If tRNA modifications caused by Trm4 and Trm8 are defective , a rapid degradation of tRNA is induced under heat stress , resulting in the expression of heat-sensitive phenotype [21] . The tRNA surveillance system that monitors compromised tRNAs with no modification by Trm4 and Trm8 uses a rapid tRNA degradation ( RTD ) pathway to decay non-modified tRNAs , leading to cell death [21]–[23] . A human tRNA ( guanine-N7- ) -methyltransferase , a homologue of yeast Trm8 , is known as METTL1 ( methyltransferase like 1 ) [20] , [24] . Whereas NSUN2 has been initially identified as a substrate of protein kinase ( Aurora-B ) in HeLa cells [17] , METTL1 has been initially identified as a substrate of Akt/protein kinase Bα ( PKBα ) in HeLa cells [13] . Interestingly , phosphorylated METTL1 at Ser27 by Akt is also enzymatically inactive [13] . The fact that both tRNA methyltransferases are evolutionally conserved suggests a similar tRNA surveillance system including Trm4 and Trm8 in human cells . Furthermore , the observation that the cytotoxic effect of 5-FU in yeast is enhanced by heat stress in a trm8 mutant strain [25] leads us to the hypothesis that nonessential tRNA modifications catalyzed by NSUN2 and METTL1 impacts the efficiency of 5-FU treatment in human cancer cells . Here , we provide evidence that tRNA methyltransferases , NSUN2 and METTL1 , strongly influences 5-FU sensitivity in human cancer cells . Therefore , targeting these methyltransferases might represent a promising rationale to improve 5-FU-treatment of tumors and to reduce 5-FU-related side effects in patients .
NSUN2 ( SAKI ) has been reported to be overexpressed and with gain in gene copy-number in various of human cancers [15] . Furthermore , NSUN2 has been implicated in myc-induced proliferation [26] . In line with these observations , the siRNA-mediated knockdown of NSUN2 negatively affects cancer cell growth [14] and homozygous knockout of the NSUN2 gene locus causes delayed cell growth in bulge stem cells [27] . However , in our previous studies , NSUN2 expression was not altered during the cell cycle of HeLa cervix carcinoma cells [17] . When we investigated normal human diploid fibroblasts , NSUN2 expression was found to be very low compared with HeLa cells and again NSUN2 was not differentially expresses during the cell cycle [17] . In initial studies we sought to analyze the impact of increased or decreased NSUN2 expression on the growth properties of HeLa cells . We therefore utilized cell lines clonally derived from stable transfectants described previously [17] . These studies indicated that there was a difference in the growth properties that arise as a result of heterogeneity among clones although we found that NSUN2 did not alter the growth properties of HeLa cells both onto plastic dish culture and in semisolid agar culture ( Figure S1 ) . Subsequently , we pooled cells from five independent clones for further experiments and examined expression levels of NSUN2 and METTL1 . We then generated Xpress-NSUN2-overexpressing HeLa cells as well as NSUN2 knockdown cells , the latter by using an shRNA targeting the 5′-UTR of NSUN2 mRNA . Successively we tested cell growth both onto plastic dish culture and in semisolid agar culture . The data clearly indicated that NSUN2 is related to neither cell multiplication nor cancerous cell growth ( Figure S2 and S3 ) . To further elucidate NSUN2 function in mammalian cells , we focused on mechanisms involved in tRNA methylation . NSUN2 is a mammalian homolog of yeast Trm4 . In yeast system , Trm4-mediated tRNA modification is nonessential , but the additional knockout of Trm8 , which is tRNA ( guanine-N7- ) -methyltransferase , under Trm4 knockout background leads to an unstable tRNA situation , resulting in a temperature-sensitive growth . Based on cooperative functions of Trm4 and Trm8 in yeast , we sought to analyze the effects of overexpressed NSUN2 and METTL1 in HeLa cells suffering heat stress . For this we used HeLa cell lines engineered to express NSUN2 , METTL1 and both methyltransferases . The ectopic expression of the methyltransferases was confirmed by Western blot analysis as depicted in Figure 1A . Contrary to our expectations , overexpression of NSUN2 and METTL1 did not affect heat stress-induced cytotoxicity ( Figure 1A , 1B and 1C ) . Next , we sought to investigate whether overexpression of NSUN2 and METTL1 protects from 5-FU-induced cytotoxicity , since tRNA modifying enzymes have been implicated as in vivo targets for 5-FU in yeast [25] . Although we could not observe a protective effect after 5-FU-treatment in HeLa cells expressing NSUN2 or METTL1 alone , we revealed that combined overexpression of NSUN2 and METTL1 significantly protected HeLa cells from 5-FU-induced cell death ( Figure 1D and 1E ) . Noteworthy , co-overexpression of NSUN2 and METTL1 did not affect cell proliferation growth in soft agar ( Figure S4 ) . To examine the effects of double knockdown of NSUN2 and METTL1 on heat stress-induced cytotoxicity and 5-FU-induced cytotoxicity , we established stable HeLa cells with knockdown of NSUN2 , METTL1 , or NSUN2 and METTL1 . As depicted in Figure 2A we fully knocked down NSUN2 , METTL1 protein expression and were also able to achieve an effective NSUN2/METTL1 double knockdown . Interestingly , the knockdown of NSUN2 did not affect cell multiplication and cancerous cell growth ( Figure S5 ) . Also , we noted that knockdown of METTL1 and double knockdown of NSUN2 and METTL1 did not affect cell multiplication and cancerous cell growth ( Figure S5 ) . Since it was of special interest whether the knockdown of NSUN2 , METTL1 and double knockdown of NSUN2 and METTL1 results in a temperature-sensitive growth phenotype as reported in yeast we incubated the cells at 43°C for a time period of 1 h to 6 h . Contrary to our expectations , the double knockdown of NSUN2 and METTL1 did not enhance heat stress-induced cytotoxicity ( Figure 2B and 2C ) . Subsequently we examined whether the double knockdown of NSUN2 and METTL1 affects 5-FU-induced cytotoxicity . Consistent with our observation of NSUN2 and METTL1 overexpression experiments , which shows protective effects of tRNA methyltransferases , we revealed that double knockdown of NSUN2 and METTL1 significantly sensitized HeLa cells to 5-FU treatment ( Figure 2D and 2E ) . To examine the effects of double knockdown of NSUN2 and METTL1 on further chemotherapeutic agents other than 5-FU , we used cisplatin , a platinum-containing anti-cancer drug that binds to and causes crosslinking of DNA , and paclitaxel , a microtubule-stabilizing mitotic inhibitor . Contradictory to 5-FU treatment , both agents caused no detectable differences in induced cytotoxicity between control cells and double knockout ( Figure 3 ) . As shown in Figure 2 , we found that double knockdown of NSUN2 and METTL1 potentiated the cytotoxic effect of 5-FU , measured by MTT assay . To confirm our results , colony formation assay , a method capable to essentially test every cell in the population for its ability to cell multiplication escaping cell reproductive death , was used to determine the effectiveness of 5-FU . The data revealed that double knockdown of NSUN2 and METTL1 caused increased sensitivity to 5-FU-treatment resulting in a dose-dependent and significant decreased clonal survival when compared to scrambled shRNA control ( Figure 4 ) . NSUN2 and METTL1 knockdown cells treated with 5–20 µM 5-FU showed a 60 - 68% decreased values of the average IC50 ( 4 . 54±0 . 12 µM ) compared with those of the parent cells ( 12 . 85±0 . 25 µM ) and control vector-transfected cells ( 12 . 79±1 . 05 µM ) ( Figure 4C ) . In the yeast system , a double knockout of Trm4 and Trm8 , the orthologs of human NSUN2 and METTL1 , shows an increased sensitivity to heat stress , resulting in a rapid tRNA ( ValAAC ) degradation . A similar situation may therefore occur when exposing HeLa cells depleted of NSUN2 and METTL1 to 5-FU . To address this question we performed tRNA stability assays using HeLa cells with double knockdown of NSUN2 and METTL1 and as control HeLa cells transfected wit scramble shRNA . As shown in Figure 5 , tRNA ( ValAAC ) was unstable in HeLa cells with double knockdown of NSUN2 and METTL1 when exposed to 5-FU whereas in control cells the amount of tRNA ( ValAAC ) remained stable . Interestingly , we monitored a rapid degradation of initiator tRNAMet ( tRNA ( iMet ) ) in both control cells and knockdown cells when exposed to heat stress and even when exposed to 5-FU whereas the amounts of elongator tRNAMet ( tRNA ( eMet ) ) molecules were not affected . NSUN2 is phosphorylated at Ser139 by Aurora-B , and the phosphorylation of this site is critical for repression of its methyltransferase activity [17] . The same regulation of enzymatic activity has been reported for METTL1 , which is phosphorylated at Ser27 by Akt [13] . If decreased sensitivities to 5-FU by double knockdown of NSUN2 and METTL1 are due to decreased methyltransferase activities resulting from decreased expression levels of both proteins , co-expression of enzymatic active forms of both methyltransferases in NSUN2 and METTL1 knockdown cells may confer resistance to 5-FU . To test this hypothesis , we established stable clones originating from UTR-targeting shRNA-mediated NSUN2 and METTL1 knockdown cells with co-expression of both NSUN2 and METTL1 using expression plasmids containing only the wild type coding region of NSUN2 and METTL1 or genetically engineered to encode phosphorylation- or dephosphorylation-mimetic proteins ( Figure 6A ) . Then 5-FU sensitivities were examined in these stable clones . As shown in Figure 6B and 6C , co-expression of wild type proteins NSUN2 and METTL1 conferred protection against 5-FU-induced cytotoxicity on NSUN2 and METTL1 knockdown cells . Compared to wild type , dephosphorylation-mimic forms were more effective ( Figure 6C ) . On the other hand , phosphorylation-mimic forms had no effect ( Figure 6B and 6C ) . Thus , the combined activity of NSUN2 and METTL1 seems to be critical for 5-FU sensitivity of cancer cells , suggesting the presence of a conserved RTD-like pathway regulated by NSUN2 and METTL1 for participating in a surveillance system for tRNA quality control in human .
The upregulated expression of NSUN2 in cancer cells is notable because the overexpression is accompanied by gene copy number gain [14] , [15] . NSUN2 has an enzyme activity that can methylate tRNA [17] . Over the past several decades , many reports have shown elevations of tRNA methyltranferase activity and tRNA level in cancer cells [28] , [29] . These observations indicate tRNAs as potential biomarkers for tumor progression and malignancy . Yet , there have been no mechanistic explanation linking increased expression of modifier enzymes of tRNAs to growth advantage of cancer cells . In this study we revealed that NSUN2 is not a critical regulator of cancer cell growth . We conclusively show that forced overexpression and knockdown of NSUN2 do not impact proliferation and cancerous growth properties in soft agar . Indeed , forced overexpression of NSUN2 did not have any oncogenic activity per se and did not potentiate ras-induced in vitro neoplastic transformation in BALB/c 3T3 ( Figure S6 ) . Although NSUN2 overexpression does not act as oncogene , the likelihood that NSUN2 overexpression might confer a growth advantage or a cancer stem cell property in cancer cells remains . Through the generation and phenotypic analysis of knockout mice , murine NSUN2 as a target of proto-oncogene Myc [26] is required to stem cell growth and normal stem cell differentiation in skin and is also required to testis differentiation [27] . NSUN2 catalyzes the formation of cytosine-5-methylation in nucleic acids including tRNA , rRNA , coding and non-coding RNA , and possibly genome DNA [17] , [30]–[34] . The altered expression levels of NUSUN2 might affect cell properties of certain tissues in an epigenetic fashion . Notably , mutations in NSUN2 gene resulting in the gene expression defect are found in patients with intelligence deficiencies [35]–[37] . Previous studies suggest an alternative mechanism by which NSUN2 overexpression contributes to cancer development . NSUN2 is phosphorylated by Aurora-B , resulting in the repression of enzymatic activity , and distributed to the cytoplasm for participating in spindle stability during mitotic cells [17] , [38] . Although the elucidation of the roles of NSUN2 overexpression in cancer cells is still challenging , mitotic spindle-related NSUN2 functions involved in chromosome segregation processes might induce chromosome instability leading to cancer progression . In yeast trm4/trm8 double mutant , certain hypo-modified tRNAs , particularly tRNA ( ValAAC ) , are destabilized by RTD pathway [21] . Due to this tRNA quality check , the double mutant is temperature-sensitive . The RTD pathway in yeast can also be triggered without coincident temperature sensitivity or without loss of modifications in different tRNA variants [23] . In our present study , double knockdown of NSUN2 and METTL1 in HeLa cells did not affect heat stress-induced cytotoxicity . Thus , it is likely that RTD is not evolutionarily conserved with respect to heat stress-induced rapid tRNA decay involving Trm4 and Trm8 . Wild type yeast cells are temperature-sensitive in the presence of higher 5-FU concentrations . Moreover , synthetic interaction between trm8 and pus1 ( pseudouridylation defective ) mutations is observed under heat stress [25] . Reduced pseudouridylation is the most likely cause of 5-FU-induced tRNA damage ( s ) that mediate tRNA destabilization [25] . Similarly , HeLa cells is 5-FU-sensitive under heat stress . However , the 5-FU-sensitive phenotype of yeast trm8 mutant under heat stress is not evolutionarily conserved because heat stress potentiates the sensitivity of those cells to the cytotoxicity of 5-FU in the same way as control cells ( Figure S7 ) . At normal temperature in the presence of 5-FU , analysis of double mutants showed that yeast trm8 genetically interacts with other genes encoding tRNA modification activities such as trm10 , pus1 , and mod5 [25] . Only trm8 mutant , but not trm4 mutant , is found to be 5-FU sensitive [25] . Although it is not clear whether yeast trm4/trm8 double mutant displays hypersensitivity to 5-FU , tRNA modifying enzymes should be important factors for determining 5-FU sensitivity in yeast . Here , our experiments using HeLa cells demonstrated that these RTD-related tRNA modifying enzymes are associated with 5-FU sensitivity . Apparently the effect of 5-FU is different in yeast when compared to HeLa cells . Knockdown of NSUN2 and METTL1 in HeLa cells potentiates sensitivity of the cells to 5-FU , whereas heat stress of cells revealed no effects . In contrast , yeast tRNA modification mutants show similar synthetic interactions for temperature sensitivity and sensitivity to 5-FU . These differences are related to tRNA stability ( Figure 5 ) . This result supports the idea that 5-FU-substituted and hypomodified tRNA that is caused by knockdown of tRNA methylases ( NSUN2 and METTL1 ) after 5-FU exposure is possibly monitored and checked by RTD pathway in mammalian cells . Although the sites of specific tRNA modifications have not been yet determined in HeLa cells used here , METTL1 and NSUN2 target sites are mostly conserved between mammals and yeasts [32] , [39] . In our model ( Figure 7 ) , loss of tRNA modification causes tRNA destabilization , which is detected as temperature sensitivity in yeast but not in mammals , but may be detected as 5-FU sensitivity in mammals . Interestingly , in our study , degradation of tRNA ( iMet ) is heat-sensitive but is also 5-FU-sensitive ( Figure 5 ) . From these data , the tRNA ( iMet ) stability is not likely to be related to tRNA modifications generated by NSUN2 and METTL1 . This observation is of particular interest when considering stress responses that undergo a quality control check by RTD , because tRNA ( iMet ) stability is not affected by stresses other than heat or 5-FU , such as ethanol , hydrogen peroxide ( oxidative stresses ) , low-pH ( pH 5 . 0 ) , cycloheximide , and sodium chloride ( high salt concentrations ) [40] . Loss of tRNA ( iMet ) at high temperature was reported previously by Watanabe et al . who has shown that degradation of initiator tRNA , tRNA ( iMet ) , depends on Xrn1/2 nucleases [40] . Yeast RTD pathway that includes Rat1 and Xrn2 , monitors the structural integrity of the acceptor and T-stem of tRNA and degrades unstable tRNA species . Since the problem of the structural stabilization of the acceptor and T-stem in tRNA ( iMet ) in HeLa has been not clarified , Watanabe and colleagues suggest that the degradation pathway in human may resemble the RTD pathway in yeast . A role of Xrn1/2 in degradation of tRNA ( ValAAC ) in HeLa exposed to 5-FU is unknown . Yeast Trm4 and Trm8 are required for protecting yeast tRNA against RTD , but the functional analogy of their mammalian orthologs NSUN2 and METTL1 would need further study to determine more precisely . DNA and RNA species other than tRNA , such as rRNA , snRNA , mRNA and pre-mRNA , are targets for 5-FU [1] . Those are also targets for NSUN2 with broad substrate specificity [17] , [26] , [31]–[34] . However , the possibility that the nucleic acid species other than tRNA are involved in the NSUN2 and METTL1-dependent potentiation mechanism under 5-FU treatment is probably eliminated , because METTL1 is evolutionarily conserved and catalyzes the formation of N7-methylguanine at position 46 in tRNA in a substrate-specific manner [13] , [20] , [24] . It has long been suggested that deregulation of translation contributes to cancer development [41] , [42] . Forced overexpression of components of translation machinery can lead to cell transformation [43] , [44] . Elevations of tRNA methyltransferase activities and elevated expression levels of tRNA have also long been recognized in cancer cells [28] , [29] , [45] , [46] . Although such several circumstantial evidences suggests that deregulation of translation might be a common cause for human cancer , one viewpoint still argues that such tRNA deregulation might be only characteristics of cancer cells like a by-product , which is associated with increased proliferation and elevated levels of protein synthesis . The tRNA overexpression that leads to increase the translation efficiency of genes is indeed characteristic of cancer cells but does not necessarily need to be related to their growth advantage . In our view , based on the present study , tRNA modifications are likely to contribute to cancer cell survival under certain stresses that interferes with tRNA stability . We therefore envisage new approaches implementing RNA-methyltransferase inhibitors or tRNA-modifying molecules as amplifier for chemotherapy of cancer .
The human cervical cancer cell line HeLa was provided by the late Professor Masakatsu Horikawa , Faculty of Pharmaceutical Sciences , Kanazawa University [47] . The cells were cultured in Dulbecco's modified Eagle's minimum essential medium containing 10% fetal bovine serum at 37°C in a humidified atmosphere of 5% CO2 and 95% air . 5-FU ( Nakarai Tesque , Kyoto , Japan ) was dissolved in distilled water . Cisplatin ( Sigma-Aldrich , St . Louis , MO , USA ) was dissolved in phosphate buffered saline solution without Ca2+ and Mg2+ . Paclitaxel ( Sigma-Aldrich , St . Louis , MO , USA ) was dissolved in dimethyl sulfoxide . The chemical treatment was performed carefully under a protocol approved for evaluation in in vitro cultured cells [48] . Rabbit polyclonal anti-NSUN2 [17] and anti-METTL1 ( ProteinTech Group , Chicago , IL , USA ) antibodies were used to detect endogenous tRNA modifying enzymes . Monoclonal anti-Xpress ( Invitrogen , Carlsbad , CA , USA ) and anti-α-tubulin ( Cedarlane Laboratories , Burlington , Ontario , Canada ) antibodies were also used . Whole cell lysates were prepared using Laemmli SDS-sample buffer . Protein concentration of lysed cells was quantified via the Bradford assay ( Bio-Rad Protein Assay , Bio-Rad Laboratories , Hercules , CA , USA ) . Proteins were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( 20 µg/lane ) , electroblotted onto a polyvinylidene difluoride membrane , Immobilon P ( Millipore Corporation , Billerica , MA , USA ) , using an electroblotter ( ATTO , Tokyo , Japan ) , and incubated with an appropriate antibody . The immunoblots were developed using electrogenerated chemiluminescence reagents kit ( GE-Amersham , Princeton , NJ , USA ) . Expression plasmids for Xpress-tagged NSUN2 and its constitutively dephosphorylated and phosphorylated forms ( Xpress-tagged NSUN2-S139A and Xpress-tagged NSUN2-S139E ) have been described previously [17] . The coding region of METTL1 cDNA ( GenBank: BC000550 . 1 , Open Biosystems , Huntsville , AL , USA ) was amplified by polymerase chain reaction and subcloned into pcDNA3 . 1/His , generating plasmid pcDNA3 . 1-Xpress-METTL1 . The following mutant METTL1 expression plasmids were also constructed: pcDNA3 . 1-Xpress-METTL1-SA , producing dephosphorylation-mimic METTL1-S27A , in which Ser27 of METTL1 was replaced with Ala by site-directed mutation at T79G; and pcDNA3 . 1-Xpress-METTL1-SE , producing phosphorylation-mimic METTL1-S27E in which Ser27 of METTL1 was replaced with Glu by site-directed mutation at T79G , C80A and C81A . HeLa cells were transfected with constructs encoding tRNA modifying enzymes or control vector by using lipofectamine 2000 ( Invitrogen , Carlsbad , CA , USA ) . G418-resistant colonies ( over twenty colonies ) were cloned in each transfection , and the protein expression of each clone was checked by immunoblot analysis . Five independent clones that overexpressed Xpress-tagged proteins were pooled and used as a stable transfectant . Oligonucleotides targeted to the 5′-untranslated region ( UTR ) of NSUN2 and METTL1 were synthesized , and were ligated into pSUPERIOR . PURO ( OligoEngine , Seattle , WA , USA ) . Puromycin-resistant colonies ( thirty colonies ) were cloned in each transfection , and the protein expression of each clone was checked by immunoblot analysis as shown previously [17] . Five independent clones that repress expression of endogenous tRNA modifying enzymes were pooled and used as a knockdown clone . Doubling times were calculated from growth curves in logarithmic growth phase . Colony-forming abilities in semisolid medium were determined by the method described previously [49] . Anchorage-independent colonies grown in 0 . 2% washed agar medium were counted . Cells were inoculated into 96-well plates . After 4 h of culture , each drug solution was added . For hyperthermia experiments , cells were cultured at 43°C for various periods . The treated cells were cultured for additional 72 h . Cell viability was determined by WST-1 assay ( Roche Applied Science , Indianapolis , IN , USA ) . The IC50 value represents the drug concentration resulting in 50% viability . Colony formation assay [50] , also known as clonogenic assay , was used to confirm the cytotoxic effect of 5-FU . Cells were plated onto a plastic surface and allowed to attach for 24 h . Each drug solution was added directly into each dish , as mentioned previously[48] . Colonies formed were stained and counted . The IC50 value represents the drug concentration resulting in 50% survival . Total RNA was isolated from HeLa cells treated with heat stress or 5-FU by using RNAiso Plus ( Takara-Clontech , Tokyo , Japan ) . RNA gel electrophoresis and Northern blot analysis identified tRNA ( ValAAC ) , tRNA ( iMet ) , tRNA ( eMet ) , and 5S rRNA by using each specific labeled probes as described below . Quantification of band intensity was calculated by Multi Gauge ( Fujifilm , Tokyo , Japan ) . The probe sequences are tRNA ( ValAAC ) : GGACCTTTCGCGTGTTAGG , tRNA ( iMet ) : GCAGAGGATGGTTTCGATCCATC , tRNA ( eMet ) : CCCCGTGTGAGGATCGAACTCAC , and 5S rRNA: CAGGGTGGTATGGCCGTAGAC . All experiments to obtain quantitative data were repeated independently three times ( n = 3 ) . Differences between values were analyzed using a two-tailed Welch's t-test . P-values of <0 . 05 were considered significant . Data is presented as the mean +/− one standard deviation ( SD ) .
|
The cellular mechanisms for sensing and responding to stress on nucleic acid metabolism or to genotoxic stress are the fundamental and ancient evolutionary biological activities with conserved and diverse biological functions . In yeast , hypomodified mature tRNA species are rapidly decayed under heat stress by the RTD pathway . Yet , it has been shown that tRNA-specific methyltransferases Trm4 and Trm8 protect from tRNA decay . 5-FU , a pyrimidine analog used for cancer treatment , is generally known to act as a thymidylate synthase inhibitor although other ways for the mechanisms of action are suggested . We studied NSUN2 and METTL1 , the human orthologs of Trm4 and Trm8 in yeast , and demonstrated that these RTD-related tRNA modifying enzymes are involved in 5-FU sensitivity in cervical cancer HeLa cells . We conclude that the evolutionarily conserved regulation of tRNA modifications is a potential mechanism of chemotherapy resistance in cancer cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"death",
"cellular",
"stress",
"responses",
"cancer",
"genetics",
"medicine",
"and",
"health",
"sciences",
"rna",
"interference",
"cervical",
"cancer",
"cancer",
"treatment",
"cell",
"processes",
"drug",
"dependence",
"cancers",
"and",
"neoplasms",
"toxicology",
"rna",
"stability",
"gene",
"function",
"oncology",
"behavioral",
"pharmacology",
"clinical",
"medicine",
"cell",
"growth",
"epigenetics",
"pharmacology",
"gene",
"expression",
"gynecological",
"tumors",
"drug",
"discovery",
"biochemistry",
"rna",
"rna",
"processing",
"cell",
"biology",
"drug",
"research",
"and",
"development",
"genetics",
"biology",
"and",
"life",
"sciences",
"molecular",
"cell",
"biology",
"genetics",
"of",
"disease",
"genetic",
"toxicology"
] |
2014
|
tRNA Modifying Enzymes, NSUN2 and METTL1, Determine Sensitivity to 5-Fluorouracil in HeLa Cells
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.