Schistosomes are parasitic worms that infect and cause chronic disease in hundreds of millions of people in the developing world. A major reason these parasites are so damaging is that they are capable of living and reproducing in the human bloodstream for decades.

Previous research had shown that schistosomes have a population of stem cells that are proposed to promote the parasite’s survival inside the host’s bloodstream. However, it was not clear what role these cells play in the worms.

Collins et al. have now found that, in a parasitic worm called Schistosoma mansoni, a large number of these stem cells are destined to become cells that generate the parasite’s skin. This unique tissue is known as the tegument, and had long been thought to have evolved in parasitic flatworms to help them survive in their host and evade its immune defenses. Therefore, Collins et al.’s findings suggest a new mechanism by which stem cells can promote the survival of a parasite inside its host.

In the long-term, these findings could lead to new treatments for parasitic infections and may shed light on the evolution of parasitic flatworms. An important future challenge will be to determine if disrupting these parasites’ stem cells, and their ability to generate new tegumental cells, has any effect on the parasite inside its host.

Neoblasts are pluripotent stem cells essential for regeneration and tissue homeostasis in a variety of free-living flatworms, most notably freshwater planarians (Newmark and Sánchez Alvarado, 2002; Wagner et al., 2011). Previously, it was shown that schistosomes, like their free-living relatives, also possess neoblasts, capable of self-renewal and differentiation into tissues such as the intestine and muscle (Collins et al., 2013). However, the role these cells play in the biology of the parasites in their mammalian host was unexplored. To decipher the cellular functions of schistosome neoblasts, we compared the short-term and long-term transcriptional consequences for the parasite following neoblast depletion (Figure 1a).

Figure 1 with 2 supplements see all

Download asset Open asset

To examine the transcriptional effects of neoblast ablation, we exploited the observation that expression of genes in differentiated tissues (e.g., the intestine) is unaffected at 48 hr following irradiation, whereas the neoblasts are irreversibly killed (Collins et al., 2013). Previously, we demonstrated that many genes down-regulated at 48 hr following irradiation were associated with the schistosome neoblasts (Collins et al., 2013). Thus, we reasoned that by comparing the gene expression profiles of parasites shortly after neoblast ablation (48 hr) to parasites two weeks after their neoblasts had been killed, we could characterize the long-term consequences of neoblast depletion. Specifically, we expected genes down-regulated at both early and late time points to be neoblast-enriched factors, whereas genes only down-regulated at later time points would be genes that require neoblasts for maintaining their expression. To add specificity to this dataset, removing genes whose expression could be influenced non-specifically by irradiation (Solana et al., 2012), we also profiled parasite transcriptomes after long-term RNA interference (RNAi) targeting either of two genes required for the maintenance of proliferating neoblasts: fgfrA (Collins et al., 2013) or histone H2B (Figure 1—figure supplement 1).

From our transcriptional profiling experiments of male somatic tissues we identified 135 genes that were down regulated (≥1.25x, p<0.05) in both our irradiation and RNAi datasets (Figure 1b, Supplementary file 1). As anticipated, this gene set included a number of known stem cell- (e.g., nanos2 and ago2-1[Collins et al., 2013]) and cell cycle-specific (cyclinB and mcm2) genes (Figure 1b,c and Supplementary file 1). More importantly, we identified 105 genes that were not down-regulated at D2 post-irradiation but were significantly down regulated (≥1.25x, p<0.05) at D14 post-irradiation and following RNAi of either fgfrA or histone H2B (Figure 1b,c and Supplementary file 1). For brevity, we will refer to these 105 genes as delayed irradiation-sensitivity (DIS) genes. We also noted a small class of genes that were modestly down regulated at early time points and highly down regulated after long-term stem cell depletion (Supplementary file 1). The most striking example of this class was the schistosome orthologue of the planarian p53 (Pearson and Sánchez Alvarado, 2010), which was down-regulated ~2 fold at 48 hr and nearly 150 fold at D14 post-irradiation (Supplementary file 1).

To validate our transcriptional profiling experiments, we examined a subset of these DIS genes by whole-mount in situ hybridization at D2 and D7 following irradiation. As anticipated, expression of cathepsin B, a gene expressed in differentiated intestinal cells, was unaffected at either time point (Figure 2a). Conversely, the expression of genes associated with the neoblasts (fgfrA and nanos2) was substantially reduced at D2 and the expression of these genes did not return by D7 (Figure 2a), confirming that stem cells are irreversibly depleted by irradiation. Consistent with our RNAseq data, the number of cells expressing p53 is modestly reduced at D2 post-irradiation and dramatically reduced by D7 (Figure 2a). In contrast to the neoblast-expressed genes and p53, at D2 the number of cells expressing the DIS genes tsp-2, sm13, sm29, and val-8 was unaffected (Figure 2a). However, by D7 post-irradiation the expression of these genes was severely depleted (Figure 2a). We did note in our RNAseq experiments, and in independent qPCR experiments, a modest increase in val-8 mRNA levels 48 hr post-irradiation (Figure 1 and Figure 1—figure supplement 2). Since, the number of val-8⁺ cells did not appear to dramatically change at 48 hr post-irradiation (Figure 2a), it is possible that some val-8⁺ cells had elevated levels of the val-8 mRNA. To directly examine the relationship between genes expressed in neoblasts and the DIS genes, we performed double fluorescence in situ hybridization (FISH) experiments with histone H2B, p53, and tsp-2. We observed no co-expression of the DIS gene tsp-2 with histone H2B Figure 2—figure supplement 1, suggesting that DIS genes are expressed in a population of cells other than neoblasts. Consistent with our observations following irradiation, we observed that p53 was expressed in both the histone H2B⁺ neoblasts and tsp-2⁺ cells (Figure 2—figure supplement 1). Together, these data strongly support the model that the DIS genes are expressed in an irradiation-sensitive population of cells that is molecularly distinct from the neoblasts.

Figure 2 with 1 supplement see all

Download asset Open asset

Upon closer examination we noted that a number of the DIS genes encoded proteins previously shown by immunological and/or proteomic approaches to be associated with the parasite's surface (e.g., tsp-2 (Tran et al., 2006; Pearson et al., 2012; Wilson, 2012), sm13 (Abath et al., 2000; Wilson, 2012), sm29 (Braschi and Wilson, 2006; Cardoso et al., 2008; Wilson, 2012), sm25 (Abath et al., 1999; Castro-Borges et al., 2011; Wilson, 2012)). The schistosome surface is covered by a continuous syncytial structure, called the tegument (Figure 2b), which serves as the primary barrier between the parasite and its host. This unique tissue is connected by cytoplasmic bridges to nucleated cell bodies that sit in the mesenchyme, beneath the parasite’s body-wall muscles (Morris and Threadgold, 1968; Wilson and Barnes, 1974) (Figure 2b). To determine if these DIS genes are expressed in a tegument-associated cell population, we performed double FISH experiments. We first examined the distribution of the mRNA for a Tetraspanin, TSP-2, that encodes a well-characterized tegument-specific factor (Braschi and Wilson, 2006; Tran et al., 2006; Pearson et al., 2012). TSP-2 is currently being explored as an anti-schistosome vaccine candidate due to its presence on the parasite surface (Hotez et al., 2010). Consistent with tsp-2 being expressed in a tegument-associated cell population, we found that a majority of tsp-2⁺ cells are located immediately beneath the parasite’s body-wall muscle layer (Figure 2b). To further examine this tsp-2⁺ cell population, we performed double FISH with other DIS genes known to encode proteins expressed in the tegument. We observed that DIS genes encoding a panel of known tegumental factors, including sm13 (Smp_195190), sm29 (Smp_072190), sm25 (Smp_195180), an amino acid transporter (Smp_176940) (Wilson, 2012), a dysferlin protein (Smp_141010) (Braschi and Wilson, 2006; Wilson, 2012), an endophillin B1 (Castro-Borges et al., 2011; Wilson, 2012), and a cd59-like molecule (Smp_081920) (Wilson, 2012) were expressed in a largely overlapping population of cells with tsp-2 immediately beneath the dorsal body-wall muscles (Figure 2c). Given their position in the parasite, and their expression of many known tegumental genes, our data indicate that tsp-2⁺ cells represent a population of tegument-associated cells.

Our data suggest that tsp-2⁺ cells co-express many known tegumental factors and are lost within a few days following stem cell depletion. We envision two models to explain these observations. First, tsp-2⁺ cells could represent a relatively long-lived population that requires the continual presence of the somatic neoblasts for their survival. Alternatively, the tsp-2⁺ cells could be a short-lived cell population that requires a pool of stem cells for its continuous renewal. In the absence of this renewal, the tsp-2⁺ cells are rapidly depleted. To distinguish between these possibilities, we performed pulse-chase experiments with the thymidine analogue EdU (Salic and Mitchison, 2008). This approach allows us to specifically label neoblasts at S-phase and monitor their differentiation over time (Collins et al., 2013). In these experiments, parasite-infected mice were injected with EdU and the distribution of EdU⁺ cells relative to the tsp-2⁺ cells was monitored every other day for 11 days (Figure 3a). If the tsp-2⁺cells are long-lived and turn over slowly, we would anticipate that few tsp-2⁺cells would become EdU⁺ over the chase period. However, if these cells were renewed rapidly by the neoblasts, we would expect a large fraction of tsp-2⁺ cells to become EdU⁺. Furthermore, since EdU levels are diluted following cell division, over time differentiating neoblasts would contain less EdU, resulting in a reduction in the EdU levels in tsp-2⁺ cells.

Figure 3

Download asset Open asset

At D1 following an EdU pulse, <0.25% of tsp-2⁺ cells were EdU⁺ (Figure 3b,c), indicating that the tsp-2⁺ cells are not proliferative. After a 3-day chase period, however, we noted that over 40% of tsp-2⁺cells were newly born EdU⁺ cells (Figure 3b,c). This result suggests that stem cells initially incorporating EdU were capable of replenishing nearly half of the tsp-2⁺ cells within three days. Beyond D5, we noted a rapid reduction in the number of tsp-2⁺ EdU⁺ cells and an overall reduction in the EdU levels per cell (Figure 3b,c). These data suggest that tsp-2⁺ cells are a short-lived cell population that is continuously renewed by the neoblasts during the parasite’s life in its definitive host. To determine if this rapid rate of neoblast-driven renewal was unique to tsp-2⁺ cells, we also examined the kinetics of EdU labeling of the schistosome intestine. In contrast to the tsp-2⁺ cells, only 2.5% of cathespin B⁺ intestinal cells were EdU⁺ at D3, and this level remained fairly constant throughout the 11D time course (Figure 3b,d). Thus, the kinetics of tegumental cell birth differs from that of intestinal cells. Taken together, our data suggest that on a population level neoblasts are 'biased' toward the rejuvenation of tsp-2⁺ cells over other lineages.

In planarians, a population of postmitotic neoblast progeny displays similar sensitivity to irradiation as this tsp-2⁺ tegument-associated cell population (Eisenhoffer et al., 2008). These planarian neoblast progeny similarly express a p53-like protein as well as a large collection of planarian-specific molecules (Eisenhoffer et al., 2008; Zhu et al., 2015). Most importantly, these planarian cells serve as progenitors to terminally differentiated epidermal cells (van Wolfswinkel et al., 2014; Tu et al., 2015). Thus, it appears that free-living and parasitic flatworms utilize similar developmental strategies for epidermal maintenance. Presently, electron microscopy is the only methodology to unambiguously identify tegumental cell bodies in schistosomes. Therefore, with current technology, it is not possible to determine which of the irradiation-sensitive cells expressing DIS genes are terminally differentiated tegumental cell bodies. In light of this limitation, our data are consistent with two models (Figure 3e). In the first model, proliferating neoblasts differentiate to produce a short-lived population of terminally differentiated tegumental cell bodies expressing tsp-2 and other DIS genes (i.e., sm13, sm25, sm29, etc.). In the alternative model, cells expressing tsp-2 (and other DIS genes) represent a population of short-lived progenitors to terminally differentiated tegumental cell bodies. Regardless of which model (or combination of these models) is correct, our data suggest that a primary function of the schistosome neoblasts is to generate cells that contribute to the tegument.

The mammalian bloodstream would appear to be a rather inhospitable niche for a pathogen. In the case of schistosomes, there is little dispute about the importance of the tegument in defending the parasite from host immunity (McLaren, 1980; Skelly and Alan Wilson, 2006), yet the properties of this tissue that afford the parasite protection in blood are unclear. Indeed, numerous ideas have been proposed to explain this phenomenon, including tegumental absorption of host antigens (Smithers et al., 1969; Clegg et al., 1971) and the turnover of the unique tegumental outer membranes (Perez and Terry, 1973; Wilson and Barnes, 1977). Based on our data, we suggest that neoblast-driven tegumental regeneration may play a key role in the parasite’s ability to survive in the mammalian host. By undergoing continuous tegumental renewal, the parasite is likely capable of rapidly turning over the tegument and regenerating damage inflicted inside the host (e.g., via immune attack). Thus, an important goal for future studies is to address the role of neoblasts in parasite survival and tegumental function in vivo.

1. 1. Collins JJ
   2. Newmark PA

   (2015) Stem cells rejuvinate the schistosome host-parasite interface

   Publicly available at Gene Expression Omnibus (Accession no: GSE75861).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE75861

1. 1. Collins JJ
   2. Wang B
   3. Lambrus BG
   4. Tharp ME
   5. Iyer H
   6. Newmark PA

   (2013) Adult stem cells in the human parasite Schistosoma mansoni

   Publicly available at Gene Expression Omnibus (Accession no: GSE42757).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE42757

1. 1. Abath FG
   2. Hirst EM
   3. Hagan P
   4. Simpson AJ

   (1999)

   Ultrastructural localization of Sm15 and Sm25, two major tegumental adult worm antigens of Schistosoma mansoni

   Parasite 6:243–247.

   • Google Scholar
2. 1. Abath FGC
   2. Xavier EM
   3. Allen R
   4. Gomes YM
   5. Lucena-Silva N
   6. Baliza M
   7. Simpson AJG

   (2000) Characterization of Sm13, a tegumental antigen of Schistosoma mansoni

   Parasitology Research 86:745–752.

   https://doi.org/10.1007/PL00008562

   • Google Scholar
3. 1. Basch PF

   (1981)

   Cultivation of Schistosoma mansoni in vitro. I. Establishment of cultures from cercariae and development until pairing

   The Journal of Parasitology 67:179–185.

   • Google Scholar
4. 1. Braschi S
   2. Wilson RA

   (2006) Proteins exposed at the adult schistosome surface revealed by biotinylation

   Molecular & Cellular Proteomics 5:347–356.

   https://doi.org/10.1074/mcp.M500287-MCP200

   • Google Scholar
5. 1. Cardoso FC
   2. Macedo GC
   3. Gava E
   4. Kitten GT
   5. Mati VL
   6. de Melo AL
   7. Caliari MV
   8. Almeida GT
   9. Venancio TM
   10. Verjovski-Almeida S
   11. Oliveira SC

   (2008) Schistosoma mansoni tegument protein Sm29 Is able to induce a Th1-type of immune response and protection against parasite Infection

   PLoS Neglected Tropical Diseases 2:e308.

   https://doi.org/10.1371/journal.pntd.0000308

   • Google Scholar
6. 1. Castro-Borges W
   2. Dowle A
   3. Curwen RS
   4. Thomas-Oates J
   5. Wilson RA

   (2011) Enzymatic shaving of the tegument surface of live schistosomes for proteomic analysis: a rational approach to select vaccine candidates

   PLoS Neglected Tropical Diseases 5:e993.

   https://doi.org/10.1371/journal.pntd.0000993

   • Google Scholar
7. 1. Clegg JA
   2. Smithers SR
   3. Terry RJ

   (1971) Acquisition of human antigens by Schistosoma mansoni during cultivation in vitro

   Nature 232:653–654.

   https://doi.org/10.1038/232653a0

   • Google Scholar
8. 1. Collins JJ
   2. Hou X
   3. Romanova EV
   4. Lambrus BG
   5. Miller CM
   6. Saberi A
   7. Sweedler JV
   8. Newmark PA

   (2010) Genome-wide analyses reveal a role for peptide hormones in planarian germline development

   PLoS Biology 8:e1000509.

   https://doi.org/10.1371/journal.pbio.1000509

   • Google Scholar
9. 1. Collins JJ
   2. King RS
   3. Cogswell A
   4. Williams DL
   5. Newmark PA

   (2011) An atlas for Schistosoma mansoni organs and life-cycle stages using cell type-specific markers and confocal microscopy

   PLoS Neglected Tropical Diseases 5:e1009.

   https://doi.org/10.1371/journal.pntd.0001009

   • Google Scholar
10. 1. Collins JJ
    2. Wang B
    3. Lambrus BG
    4. Tharp ME
    5. Iyer H
    6. Newmark PA

    (2013) Adult somatic stem cells in the human parasite Schistosoma mansoni

    Nature 494:476–479.

    https://doi.org/10.1038/nature11924

    • Google Scholar
11. 1. Eisenhoffer GT
    2. Kang H
    3. Sánchez Alvarado A

    (2008) Molecular analysis of stem cells and their descendants during cell turnover and regeneration in the planarian Schmidtea mediterranea

    Cell Stem Cell 3:327–339.

    https://doi.org/10.1016/j.stem.2008.07.002

    • Google Scholar
12. 1. Hotez PJ
    2. Bethony JM
    3. Diemert DJ
    4. Pearson M
    5. Loukas A

    (2010) Developing vaccines to combat hookworm infection and intestinal schistosomiasis

    Nature Reviews Microbiology 8:814–826.

    https://doi.org/10.1038/nrmicro2438

    • Google Scholar
13. 1. McLaren DJ

    (1980)

    Schistosoma mansoni: The Parasite Surface in Relation to Host Immunity, Letchworth, England, John Wiley and Sons, Ltd

    Schistosoma mansoni: The Parasite Surface in Relation to Host Immunity, Letchworth, England, John Wiley and Sons, Ltd.

    • Google Scholar
14. 1. Morris GP
    2. Threadgold LT

    (1968)

    Ultrastructure of the tegument of adult Schistosoma mansoni

    The Journal of Parasitology 54:15–27.

    • Google Scholar
15. 1. Newmark PA
    2. Sánchez Alvarado A

    (2002) Not your father's planarian: a classic model enters the era of functional genomics

    Nature Reviews. Genetics 3:210–219.

    https://doi.org/10.1038/nrg759

    • Google Scholar
16. 1. Pearson BJ
    2. Sánchez Alvarado A

    (2010) A planarian p53 homolog regulates proliferation and self-renewal in adult stem cell lineages

    Development 137:213–221.

    https://doi.org/10.1242/dev.044297

    • Google Scholar
17. 1. Pearson MS
    2. Pickering DA
    3. McSorley HJ
    4. Bethony JM
    5. Tribolet L
    6. Dougall AM
    7. Hotez PJ
    8. Loukas A

    (2012) Enhanced protective efficacy of a chimeric form of the schistosomiasis vaccine antigen Sm-TSP-2

    PLoS Neglected Tropical Diseases 6:e1564.

    https://doi.org/10.1371/journal.pntd.0001564

    • Google Scholar
18. 1. Perez H
    2. Terry RJ

    (1973) The killing of adult Schistosoma mansoni in vitro in the presence of antisera to host antigenic determinants and peritoneal cells

    International Journal for Parasitology 3:499–503.

    https://doi.org/10.1016/0020-7519(73)90046-5

    • Google Scholar
19. 1. Salic A
    2. Mitchison TJ

    (2008) A chemical method for fast and sensitive detection of DNA synthesis in vivo

    Proceedings of the National Academy of Sciences of the United States of America 105:2415–2420.

    https://doi.org/10.1073/pnas.0712168105

    • Google Scholar
20. 1. Skelly PJ
    2. Alan Wilson R

    (2006) Making sense of the schistosome surface

    Advances in Parasitology 63:185–284.

    https://doi.org/10.1016/S0065-308X(06)63003-0

    • Google Scholar
21. 1. Smithers SR
    2. Terry RJ
    3. Hockley DJ

    (1969) Host antigens in schistosomiasis

    Proceedings of the Royal Society B: Biological Sciences 171:483–494.

    https://doi.org/10.1098/rspb.1969.0007

    • Google Scholar
22. 1. Solana J
    2. Kao D
    3. Mihaylova Y
    4. Jaber-Hijazi F
    5. Malla S
    6. Wilson R
    7. Aboobaker A

    (2012) Defining the molecular profile of planarian pluripotent stem cells using a combinatorial RNA-seq, RNA interference and irradiation approach

    Genome Biology 13:R19.

    https://doi.org/10.1186/gb-2012-13-3-r19

    • Google Scholar
23. 1. Tran MH
    2. Pearson MS
    3. Bethony JM
    4. Smyth DJ
    5. Jones MK
    6. Duke M
    7. Don TA
    8. McManus DP
    9. Correa-Oliveira R
    10. Loukas A

    (2006) Tetraspanins on the surface of Schistosoma mansoni are protective antigens against schistosomiasis

    Nature Medicine 12:835–840.

    https://doi.org/10.1038/nm1430

    • Google Scholar
24. 1. Tu KC
    2. Cheng L-C
    3. TK Vu H
    4. Lange JJ
    5. McKinney SA
    6. Seidel CW
    7. Sánchez Alvarado A

    (2015) Egr-5 is a post-mitotic regulator of planarian epidermal differentiation

    eLife 4:.

    https://doi.org/10.7554/eLife.10501

    • Google Scholar
25. 1. van Wolfswinkel JC
    2. Wagner DE
    3. Reddien PW

    (2014) Single-cell analysis reveals functionally distinct classes within the planarian stem cell compartment

    Cell Stem Cell 15:326–339.

    https://doi.org/10.1016/j.stem.2014.06.007

    • Google Scholar
26. 1. Wagner DE
    2. Wang IE
    3. Reddien PW

    (2011) Clonogenic neoblasts are pluripotent adult stem cells that underlie planarian regeneration

    Science 332:811–816.

    https://doi.org/10.1126/science.1203983

    • Google Scholar
27. 1. Wilson RA
    2. Barnes PE

    (1974)

    The tegument of Schistosoma mansoni: observations on the formation, structure and composition of cytoplasmic inclusions in relation to tegument function

    Parasitology 68:239–258.

    • Google Scholar
28. 1. Wilson RA
    2. Barnes PE

    (1977)

    The formation and turnover of the membranocalyx on the tegument of

    Parasitology 74:61–71.

    • Google Scholar
29. 1. Wilson RA

    (2012) Proteomics at the schistosome-mammalian host interface: any prospects for diagnostics or vaccines?

    Parasitology 139:1178–1194.

    https://doi.org/10.1017/S0031182012000339

    • Google Scholar
30. 1. Zhu SJ
    2. Hallows SE
    3. Currie KW
    4. Xu C
    5. Pearson BJ

    (2015) A mex3 homolog is required for differentiation during planarian stem cell lineage development

    eLife 4:.

    https://doi.org/10.7554/eLife.07025

    • Google Scholar

Author details

1. James J Collins III

   1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, United States
   2. Department of Pharmacology, UT Southwestern Medical Center, Dallas, United States
   3. Howard Hughes Medical Institute, University of Illinois at Urbana-Champaign, Urbana, United States

   Contribution

   JJC, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   JamesJ.Collins@UTSouthwestern.edu

   Competing interests

   The authors declare that no competing interests exist.

2. George R Wendt

   Department of Pharmacology, UT Southwestern Medical Center, Dallas, United States

   Contribution

   GRW, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

3. Harini Iyer

   1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, United States
   2. Howard Hughes Medical Institute, University of Illinois at Urbana-Champaign, Urbana, United States

   Contribution

   HI, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Phillip A Newmark

   1. Department of Cell and Developmental Biology, University of Illinois at Urbana-Champaign, Urbana, United States
   2. Howard Hughes Medical Institute, University of Illinois at Urbana-Champaign, Urbana, United States

   Contribution

   PAN, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   pnewmark@life.illinois.edu

   Competing interests

   The authors declare that no competing interests exist.

• 3,152

  views

• 639

  downloads

• 51

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.