The Caenorhabditis elegans hermaphrodite gonad is a fruitful system in which to study organogenesis, meiosis, and stem cell niche biology. Recent work from our group (Gordon et al., 2020) used two endogenously tagged alleles of genetically redundant innexin genes inx-8 and inx-9 to visualize the somatic gonadal sheath of the C. elegans hermaphrodite. We discovered that the distal most pair of sheath cells, called Sh1, lies immediately adjacent to the distal tip cell (DTC), which is the stem cell niche of the germline stem cells. Previously (based on electron microscopy and on cytoplasmic GFP overexpression from transgenes active in the sheath [lim-7p::GFP] [Hall et al., 1999] or its progenitor cells [lag-2p::GFP] [Killian and Hubbard, 2005]), Sh1 cells were thought to associate only with germ cells well into the meiotic cell cycle, so our finding required a reimagining of the anatomy of the distal gonad.

Here, we confirm that the Sh1 cells fall at the boundary of a population of germ cells in a stem-like state, report other markers that label the Sh1 cells, and verify that these markers can be used to assess gonad anatomy without unduly impacting the gonad itself. We also discuss reagents that are not suitable markers of Sh1 cells, including an overexpressed, functional cell death receptor that is used to mark Sh1 in a recent study (Tolkin et al., 2022). Finally, we consider best practices for using endogenously tagged proteins for cell and developmental studies.

We discovered that the distal position of Sh1 is much closer to the distal end of the young adult hermaphrodite gonad than than was previously observed, where it forms an interface with the DTC’s proximal projections and overlaps substantially with the proliferative zone of the germline where mitotic cell divisions occur (Gordon et al., 2020). Importantly, that study did not claim that Sh1 is necessary or sufficient for what we term ‘niche exit’; we simply observed that Sh1 associates with germ cells as they exit the niche by division. We have now confirmed this finding with functional manipulations of germ cell cycling and cell fate. We observed a distal Sh1 position in other strains with endogenously tagged sheath cell membrane proteins that act in molecular pathways outside of gap junctional coupling, and in a substantial fraction of traditional transgenic animals expressing lim-7 promoter-driven CED-1::GFP, GFP::CAAX and cytoplasmic GFP (though these strains have high variability in fluorescence intensity and localization). Therefore, we consider the results presented here to be confirmatory of the foundational finding of Gordon et al., 2020, which is that almost all mitotic germ cells in the adult hermaphrodite contact the DTC or Sh1, with a noteworthy population in contact with both. Other recent work provides further evidence of a role for sheath cell contact in promoting adult germ cell proliferation, specifically through modulation of Notch receptor glp-1 expression (Gopal et al., 2020). We focus especially on young adults in these studies (less than 24 hr post mid-L4, see Methods). An important caveat to the work is that the gonad is dynamic, and cell shapes and positions change over time. Indeed, dynamic processes could lead to the surprising difference in position often seen between the two Sh1 cells in a single gonad arm, if one Sh1 cell grows more actively over germ cells as they leave the niche. The high variability of expression levels of an overexpressed lim-7p::ced-1::GFP transgene has allowed for this surprising discovery, though that variability makes it a poor marker of the absolute position of the Sh1 cells, it sometimes causes gonad defects, and it interacts synthetically with qy78 to cause germline defects.

This work was inspired by a recent preprint (Tolkin et al., 2021; biorxiv, version 1). This preprint initially reported a severe brood size and embryonic lethality defect in strains bearing the qy78 allele (just over 100 offspring per animal, with day 1 embryonic lethality of nearly 90%) and hypothesized that abnormal innexin function caused by endogenously tagging INX-8 with mKate could be responsible for both the fertility defect and the novel finding of a distal position of Sh1 abutting the stem cell zone. This indeed would be a serious concern, and we are grateful that other researchers in the community are vigilant and interested in the strains we generated.

Tolkin et al., 2022 (revised) downscales concerns about brood size (over 200 offspring per animal) and embryonic lethality (under 10% total) but still proposes that the distal Sh1 position revealed by qy78 is an artifact of protein tagging. The study uses an overexpressed, GFP-tagged, functional cell death receptor protein as the preferred marker of the sheath (the bcIs39 allele encoding lim-7p::ced-1::GFP), but does not justify why such an element should be assumed to be a robust and non-phenotypic marker of the distal sheath. This genetic element—which we demonstrated to interact synthetically with qy78 to cause gonad and germline defects (Figure 5)—is present in the genetic background of the worms analyzed in the vast majority of the experiments reported in the four data figures of Tolkin et al., 2022: Figure 1 (all, no true wild-type analyzed), Figure 2 (half of panels D–F with another marker in the other half of those panels, see next), Figure 3 (all, no wild-type analyzed), Figure 4 (all, no wild-type analyzed). In all of these experiments, we conclude the synthetic interaction is driving the phenotypes attributed to qy78, including the dosage-dependent and deletion-mediated suppression of the phenotypes attributed to qy78 shown in Tolkin et al., 2022 (Figure 2—figure supplement 2).

In the rest of Figure 2D-F, Tolkin et al. use a fasn-1::GFP sheath marker coexpressed with qy78. Only six samples were observed, and only 2/6 had the ‘Class 3’ phenotype with a dramatically distal Sh1 border. This differs notably from what is observed for the innexin-defective inx-14(ag17) hypomorphic allele shown in Figure 1—figure supplement 1C for which 21/21 fasn-1::GFP coexpressing samples had ‘Class 3’ gonads. We do not find this experiment to be decisive because of low sample size for this genotype and results in Figure 2F that appear to fall within the distribution of the control.

Next, we consider the other experiments that do not include the bcIs39 marker. In Figure 2B–C and Figure 2–figure supplements 1 and 2, Tolkin et al. use an anti-INX-8 antibody in an immunofluorescence experiment that reports a more proximal Sh1 boundary in the wild type than in worms carrying the qy78 allele. We note the conspicuous absence of innexin detected in the DTC, where it had been reported previously (see the wild-type gonad image in both Figure 3 of Starich et al., 2014, and Figure 2A of Starich et al., 2020). We also note that the anti-INX-8 sheath localization is like a honeycomb, not sparse and punctate. Starich et al., 2020 present evidence that a honeycomb localization pattern for innexins indicates that gap junctions are not forming properly, so this pattern in the wild-type sample is unexpected. Different imaging modalities often yield different patterns, and the previously published control image was made on a compound microscope while the image in Tolkin et al., 2022 (Figure 2B) was made with a confocal microscope. However, it does not seem likely that a switch in imaging approaches would detect more abundant signal in one region (Sh1) and yet lose signal in another (the DTC) in the same field of view. Setting aside questions of reproducibility of this antibody staining experiment, this type of data could shed light on the position of the sheath in a genetically wild-type animal. Of note, Figure 3C of Starich et al., 2014 shows antibody staining of gap junctions forming in what appears to be the Sh1 region at a distance of ~10 germ cell diameters from the distal end.

Transmission electron microscopy was also included in the revised version of Tolkin et al., 2022. The 3D reconstruction in Figure 1—figure supplement 2 shows a sheath that has a honeycomb pattern, presumably due to the thinness of Sh1 as it overlies germ cell bodies (annotation was based on the presence of mitochondria), so we know this technique under-annotates the thinnest regions of the Sh1 cell. It also appears to show what Tolkin et al., 2022, would call a ‘Class 2’ gonad—the DTC and Sh1 cells terminate within ~2 germ cell diameters of one another. It is not representative of a ‘Class 1’ gonad with a large gap (that ‘exceeds 25 µm’). According to the scaling in Figure 1—figure supplement 2, the distal Sh1 annotations fall ~40 µm from the distal end. Taken at face value, this TEM data is equivocal on the question of whether there is a ‘bare region’ between the DTC and Sh1.

When we inspect the TEM stack in Video 4, we observe unannotated structures surrounding the germ cells. They have low complexity, making us wonder if they are an artifact of tissue shrinkage during fixation. Some appear to be contiguous with the annotated somatic cells (DTC and Sh1). Perhaps a more generous annotation would reveal more of the thin somatic cell structures that we know from live imaging are there (minimally, continuous Sh1 cover across germ cell bodies in the more proximal region). With that level of annotation, what distal Sh1 structures would appear? This is a very helpful type of data to include, and hopefully future studies of more TEM samples will help decide the issue.

Finally, each study furnishes additional strains with Sh1 fluorescence expression terminating in variously distal or more proximal domains: Tolkin et al. use fasn-1::gfp (endogenously tagged, cytoplasmic), acy-4::gfp (extrachromosomal, membrane localized), and lim-7p::gfp (integrated array, cytoplasmic); we report ina-1::mNG (endogenously tagged, membrane localized), cam-1::mNG (endogenously tagged, membrane localized), and lim-7p::GFP::CAAX (integrated array, membrane localized). The challenge with definitively proving a more proximal boundary of Sh1—after seeing images like those described in our Figures 2H and 3 and its Figure 3—figure supplement 1, Gordon et al., 2020 (Figure 1—figure supplement 1C) , and Tolkin et al., 2022 (Figure 2—figure supplement 3) —is the challenge of proving a negative. When distal expression is not observed, how can one be certain that the whole sheath is labeled? Absence of evidence is not evidence of absence, and the aforementioned images make it clear that some fluorescent markers fail to capture distal Sh1 structures even when the structures are detectable by other means, and even when they brightly label more proximal cells.

Taken together, we suspect that we are seeing more or less the same things but describing them differently. Tolkin et al., 2022 observe a DTC-Sh1 interface in many worms (~30%) expressing their favored sheath marker (the highly variable bcIs39). These ‘Class 2’ gonads have the pattern that we first reported for endogenously tagged sheath proteins (see Gordon et al., 2020, Figure 1B), including ina-1::mNG (an integrin subunit) and cam-1::mNG (a Wnt pathway member) that are not obvious candidates to have dominant antimorphic effects on innexin signaling. Tolkin et al., 2022 indeed identify several genetic backgrounds in which the spatial relationship between the DTC and Sh1 is perturbed, though how the mechanism acts through changes in innexin function, CED-1::GFP overexpression, or both, remains to be seen. We strongly agree that innexins in the gonadal sheath are important for gonad and germline development.

In physics, the observer effect states that it is impossible to observe a system without changing it. In biological imaging in C. elegans, this means that we can either observe wild-type animals that are dead, dissected and/or fixed and coated or stained, or we can observe genetically modified animals that are alive. Some fine, membranous cellular structures do not survive fixation (Gerdes et al., 2013; Kornberg and Roy, 2014). On the other hand, any genomic modification runs the risk of altering an animal’s physiology.

We feel most confident examining endogenously tagged gene products in Sh1 for several reasons. First, proteins expressed at physiological levels are less likely to directly damage a cell vs. overexpressed fluorescent proteins (Kintaka et al., 2016). Second, the ability to cross-reference among strains with different tagged proteins that act in different molecular pathways allows us to use concordant results in reconstructing cell positions; any single marker may or may not localize to the region of interest, but concordant results among independent experiments help construct an accurate picture of the cell. One factor to consider, however, is that not every endogenously expressed protein is likely to localize evenly across all regions of a cell. We would expect in a large cell like Sh1 that interacts with germ cells in many stages of maturation that some cell-surface proteins would be regionalized. Along those lines, it seems possible that the Sh1 cells might have mechanisms to exclude the cell death receptor CED-1 from the cell membrane domain that contacts proliferating germ cells. The bcIs39 transgene is typically used to study engulfment of apoptotic germ cell corpses at the bend of the gonad and rescues ced-1 loss-of-function mutants for apoptotic germ cell corpse engulfment (Zhou et al., 2001). We find this marker to be unreliable in the distal region of the cell, and to cause gonad defects especially but not only when combined with endogenously tagged inx-8(qy78). A recent study (Tolkin et al., 2022) uses this transgene in most of the backgrounds analyzed (sometimes detecting the CED-1::GFP by anti-GFP antibody staining, which appears to amplify the variability of the marker), so we find this problematic reagent to undermine that study’s conclusions.

The need for caution when observing and interpreting endogenously tagged fluorescent proteins is noted. Several steps can and should be taken to increase confidence that a tagged protein is not causing cryptic or unwanted phenotypes. First, multiple edited lines should be recovered and outcrossed, thereby reducing the likelihood that a phenotype is caused by off-target Cas9 cutting creating lesions in any individual edited genome. Second, brood size should be estimated either by timed food depletion (less rigorous) or formal brood size assays (more rigorous). Third, edited lines should be examined for known phenotypes caused by loss of function of the targeted genes. This can be done, in order of least to most rigorous, by consulting the literature, by comparing to RNAi treatments or known mutants, and finally by introducing AID tags and using the degron strategy to deplete the gene product under the lab’s exact experimental conditions of choice (Zhang et al., 2015), however this step will not work for extracellular tags (because extracellular AID tags are not accessible to TIR1 ubiquitin ligase). Finally, any ‘markers’ used should be assessed on their own for phenotypes. Even with these controls in place, synthetic interactions can emerge between ‘markers’ and alleles, including tagged proteins of interest. These interactions can themselves reveal biologically relevant phenomena, but only if they are recognized.

In the end, no transgenic or genome-edited strain is truly wild type, and it should be our expectation that such strains might be somewhat sensitized as a result. Indeed, the synthetic interaction we document between lim-7p::ced-1::gfp and inx-8(qy78) suggests that the qy78 is sensitized for gonad defects caused by other genetic elements affecting the gonadal sheath. However, the perfect reagent does not exit. We can only look for congruent results among a set of independent reagents with non-overlapping weaknesses. Finally, we can formulate questions narrowly enough that, despite their shortcomings, our imperfect reagents are adequate to help answer them. In the future, new endogenously tagged alleles that are expressed in the sheath, single-copy, membrane-localized transgenes that do not affect distal gonad patterning, and different imaging modalities like electron microscopy will shed more light on the complex relationship between the gonadal sheath and the germline. At the present time, however, we consider the existence of an interface between the DTC and Sh1 cells that coincides with the boundary of the distal-most stem-like germ cells to be supported by the preponderance of evidence.

Source data files contain the numerical data used to generate the figures.

1. 1. Cinquin O
   2. Crittenden SL
   3. Morgan DE
   4. Kimble J

   (2010) Progression from a stem cell-like state to early differentiation in the C. elegans germ line

   PNAS 107:2048–2053.

   https://doi.org/10.1073/pnas.0912704107

   • PubMed
   • Google Scholar
2. 1. Edelstein A
   2. Amodaj N
   3. Hoover K
   4. Vale R
   5. Stuurman N

   (2010) Computer control of microscopes using manager

   Current Protocols in Molecular Biology 92:Unit14.

   https://doi.org/10.1002/0471142727.mb1420s92

   • Google Scholar
3. 1. Evans TC

   (2006)

   The C. Elegans Research Community, WormBook

   Transformation and microinjection, elegans research community, wormbook, The C. Elegans Research Community, WormBook, Wormbook, 10.1895/wormbook.1.108.1.

   • Google Scholar
4. 1. Fox PM
   2. Schedl T

   (2015) Analysis of germline stem cell differentiation following loss of GLP-1 notch activity in Caenorhabditis elegans

   Genetics 201:167–184.

   https://doi.org/10.1534/genetics.115.178061

   • PubMed
   • Google Scholar
5. Website

   1. Francis R
   2. Nayack S

   (2000) Nayak Lab’s DAPI Staining Protocol

   Accessed September 17, 2020.

   https://owd.tcnj.edu/~nayaklab/protocols/DAPIstaining.htm
6. 1. Gerdes HH
   2. Rustom A
   3. Wang X

   (2013) Tunneling nanotubes, an emerging intercellular communication route in development

   Mechanisms of Development 130:381–387.

   https://doi.org/10.1016/j.mod.2012.11.006

   • PubMed
   • Google Scholar
7. Preprint

   1. Gopal S
   2. Amran A
   3. Elton A
   4. Ng L
   5. Pocock R

   (2020) Notch-Directed Germ Cell Proliferation Is Mediated by Proteoglycan-Dependent Transcription

   bioRxiv.

   https://doi.org/10.1101/2020.07.30.229997

   • Google Scholar
8. 1. Gordon KL
   2. Zussman JW
   3. Li X
   4. Miller C
   5. Sherwood DR

   (2020) Stem cell niche exit in C. elegans via orientation and segregation of daughter cells by a cryptic cell outside the niche

   eLife 9:e56383.

   https://doi.org/10.7554/eLife.56383

   • PubMed
   • Google Scholar
9. 1. Hall DH
   2. Winfrey VP
   3. Blaeuer G
   4. Hoffman LH
   5. Furuta T
   6. Rose KL
   7. Hobert O
   8. Greenstein D

   (1999) Ultrastructural features of the adult hermaphrodite gonad of Caenorhabditis elegans: relations between the germ line and soma

   Developmental Biology 212:101–123.

   https://doi.org/10.1006/dbio.1999.9356

   • PubMed
   • Google Scholar
10. 1. Henderson ST
    2. Gao D
    3. Lambie EJ
    4. Kimble J

    (1994) Lag-2 may encode a signaling ligand for the GLP-1 and LIN-12 receptors of C. elegans

    Development 120:2913–2924.

    https://doi.org/10.1242/dev.120.10.2913

    • PubMed
    • Google Scholar
11. 1. Heppert JK
    2. Pani AM
    3. Roberts AM
    4. Dickinson DJ
    5. Goldstein B

    (2018) A CRISPR tagging-based screen reveals localized players in wnt-directed asymmetric cell division

    Genetics 208:1147–1164.

    https://doi.org/10.1534/genetics.117.300487

    • PubMed
    • Google Scholar
12. 1. Hubbard EJA

    (2007) Caenorhabditis elegans germ line: a model for stem cell biology

    Developmental Dynamics 236:3343–3357.

    https://doi.org/10.1002/dvdy.21335

    • PubMed
    • Google Scholar
13. 1. Jayadev R
    2. Qiuyi C
    3. Daniel P K
    4. Eric L H

    (2019) α-integrins dictate distinct modes of type IV collagen recruitment to basement membranes

    The Journal of Cell Biology 218:3098–3116.

    https://doi.org/10.1083/jcb.201903124

    • Google Scholar
14. 1. Killian DJ
    2. Hubbard EJA

    (2005) Caenorhabditis elegans germline patterning requires coordinated development of the somatic gonadal sheath and the germ line

    Developmental Biology 279:322–335.

    https://doi.org/10.1016/j.ydbio.2004.12.021

    • PubMed
    • Google Scholar
15. 1. Kintaka R
    2. Makanae K
    3. Moriya H

    (2016) Cellular growth defects triggered by an overload of protein localization processes

    Scientific Reports 6:1–11.

    https://doi.org/10.1038/srep31774

    • PubMed
    • Google Scholar
16. 1. Kodoyianni V
    2. Maine EM
    3. Kimble J

    (1992) Molecular basis of loss-of-function mutations in the glp-1 gene of Caenorhabditis elegans

    Molecular Biology of the Cell 3:1199–1213.

    https://doi.org/10.1091/mbc.3.11.1199

    • Google Scholar
17. 1. Kornberg TB
    2. Roy S

    (2014) Cytonemes as specialized signaling filopodia

    Development 141:729–736.

    https://doi.org/10.1242/dev.086223

    • PubMed
    • Google Scholar
18. 1. Lee CH
    2. Shin H
    3. Kimble J

    (2019) Dynamics of notch-dependent transcriptional bursting in its native context

    Developmental Cell 50:426–435.

    https://doi.org/10.1016/j.devcel.2019.07.001

    • PubMed
    • Google Scholar
19. 1. Oshima A
    2. Tani K
    3. Fujiyoshi Y

    (2016) Atomic structure of the innexin-6 gap junction channel determined by cryo-EM

    Nature Communications 7:13681.

    https://doi.org/10.1038/ncomms13681

    • PubMed
    • Google Scholar
20. 1. Pekar O
    2. Ow MC
    3. Hui KY
    4. Noyes MB
    5. Hall SE
    6. Hubbard EJA

    (2017) Linking the environment, DAF-7/TGFβ signaling and LAG-2/DSL ligand expression in the germline stem cell niche

    Development 144:2896–2906.

    https://doi.org/10.1242/dev.147660

    • PubMed
    • Google Scholar
21. Software

    1. R Development Core Team

    (2020) R: A language and environment for statistical computing

    R Foundation for Statistical Computing, Vienna, Austria.

    https://www.r-project.org
22. 1. Roy D
    2. Kahler DJ
    3. Yun C
    4. Hubbard EJA

    (2018) Functional interactions between rsks-1/S6K, glp-1/notch, and regulators of Caenorhabditis elegans fertility and germline stem cell maintenance

    G3: Genes, Genomes, Genetics 8:3293–3309.

    https://doi.org/10.1534/g3.118.200511

    • PubMed
    • Google Scholar
23. 1. Schindelin J
    2. Arganda-Carreras I
    3. Frise E
    4. Kaynig V
    5. Longair M
    6. Pietzsch T
    7. Preibisch S
    8. Rueden C
    9. Saalfeld S
    10. Schmid B
    11. Tinevez JY
    12. White DJ
    13. Hartenstein V
    14. Eliceiri K
    15. Tomancak P
    16. Cardona A

    (2012) Fiji: an open-source platform for biological-image analysis

    Nature Methods 9:676–682.

    https://doi.org/10.1038/nmeth.2019

    • PubMed
    • Google Scholar
24. 1. Shin H
    2. Haupt KA
    3. Kershner AM
    4. Kroll-Conner P
    5. Wickens M
    6. Kimble J

    (2017) SYGL-1 and LST-1 link niche signaling to PUF RNA repression for stem cell maintenance in Caenorhabditis elegans

    PLOS Genetics 13:e1007121.

    https://doi.org/10.1371/journal.pgen.1007121

    • PubMed
    • Google Scholar
25. 1. Starich TA
    2. Hall DH
    3. Greenstein D

    (2014) Two classes of gap junction channels mediate soma-germline interactions essential for germline proliferation and gametogenesis in Caenorhabditis elegans

    Genetics 198:1127–1153.

    https://doi.org/10.1534/genetics.114.168815

    • PubMed
    • Google Scholar
26. 1. Starich TA
    2. Bai X
    3. Greenstein D

    (2020) Gap junctions deliver malonyl-coa from soma to germline to support embryogenesis in Caenorhabditis elegans

    eLife 9:e58619.

    https://doi.org/10.7554/eLife.58619

    • PubMed
    • Google Scholar
27. Preprint

    1. Tolkin T
    2. Mohammad A
    3. Starich T
    4. Nguyen KCQ
    5. Hall DH
    6. Schedl T
    7. Hubbard EJA
    8. Greenstein D

    (2021) Innexin Function Dictates the Spatial Relationship between Distal Somatic Cells in the Caenorhabditis elegans Gonad without Impacting the Germline Stem Cell Pool

    bioRxiv.

    https://doi.org/10.1101/2021.10.22.465523

    • Google Scholar
28. 1. Tolkin T
    2. Mohammad A
    3. Starich TA
    4. Nguyen KCQ
    5. Hall DH
    6. Schedl T
    7. Hubbard EJA
    8. Greenstein D

    (2022) Innexin function dictates the spatial relationship between distal somatic cells in the Caenorhabditis elegans gonad without impacting the germline stem cell pool

    eLife 10:74955.

    https://doi.org/10.7554/eLife.74955

    • Google Scholar
29. 1. Zhang L
    2. Ward JD
    3. Cheng Z
    4. Dernburg AF

    (2015) The auxin-inducible degradation (AID) system enables versatile conditional protein depletion in C. elegans

    Development 142:4374–4384.

    https://doi.org/10.1242/dev.129635

    • PubMed
    • Google Scholar
30. 1. Zhou Z
    2. Hartwieg E
    3. Horvitz HR

    (2001) CED-1 is a transmembrane receptor that mediates cell corpse engulfment in C. elegans

    Cell 104:43–56.

    https://doi.org/10.1016/s0092-8674(01)00190-8

    • PubMed
    • Google Scholar

Author details

1. Xin Li

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Data curation, Formal analysis, Investigation, Visualization, Writing - review and editing

   Competing interests

   No competing interests declared

2. Noor Singh

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Data curation, Formal analysis, Investigation, Visualization, Writing - review and editing

   Competing interests

   No competing interests declared

3. Camille Miller

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Data curation, Supervision, Investigation, Writing - review and editing

   Competing interests

   No competing interests declared

4. India Washington

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Investigation, Writing - review and editing

   Competing interests

   No competing interests declared

5. Bintou Sosseh

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Investigation, Writing - review and editing

   Competing interests

   No competing interests declared

6. Kacy Lynn Gordon

   Department of Biology, The University of North Carolina at Chapel Hill, Chapel Hill, United States

   Contribution

   Conceptualization, Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing

   For correspondence

   kacy.gordon@unc.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0967-4020

• 1,482

  views

• 105

  downloads

• 5

  citations

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