As detailed below, we found that Sh1 is somewhat variably positioned in wild-type young adult hermaphrodites. We confirmed the existence of bare regions between Sh1 and the DTC, characterized by germ cells that do not contact Sh1 or the DTC.

To better characterize the natural variability in the position of the distal border of Sh1 in the wild type, and to facilitate comparisons among different genotypes and markers, we binned gonads into one of three phenotypic classes based on the position of Sh1 (Figure 1C–F; Videos 1–3; Figure 1—figure supplement 1). To avoid ambiguity that results from the analysis of maximum projection images, we analyzed all gonads in 3 dimensions (based on 0.5 µm optical Z-series slices). We designated as Class 1, those gonads in which > 8 µm (a distance covered by 2 germ cell diameters) separated the DTC and Sh1. Class 2 includes those gonads in which Sh1 and DTC processes come within 8 µm or in which Sh1 extensions intercalate with DTC processes, up to the position of the mean length of DTC processes. Class 3 gonads are those in which the border of Sh1 is markedly shifted distally such that it lies between the distal end of the gonad and the mean DTC process length.

Video 1

Download asset

Video 2

Download asset

Video 3

Download asset

Using bcIs39[lim-7p::ced-1::GFP] to visualize Sh1, a transgene that produces a CED-1::GFP fusion protein that fully rescues the sheath cell function of CED-1 to engulf germ cell corpses (Zhou et al., 2001), we found that a majority of gonads (70%) are in Class 1, confirming the existence of bare regions. Indeed, in many cases, the distance between the proximal-most DTC measurement and the distal-most Sh1 measurement exceeds 25 µm (Figure 1G; Figure 1—figure supplement 1D). The remaining gonads are in Class 2.

Using three additional Sh1 markers (and one alternate DTC marker), we found that depending on the markers used, 50–90% of gonads fall into Class 1, while Class 2 make up the remainder (Figure 1—figure supplement 1). We note that both Class 1 and Class 2 gonads contain bare regions with spacing of germ cell nuclei (as seen by differential interference contrast microscopy) such that their surrounding cell boundaries are not plausibly in contact with the DTC or Sh1. We did not observe Class 3 gonads in any wild-type worms examined.

We also analyzed serial distal-to-proximal transmission electron micrographs (TEM) of a gonad from a wild-type, unmarked young adult worm fixed using high-pressure freezing and freeze-substitution freezing methods (Figure 1—figure supplement 2; Videos 4 and 5; Materials and methods). Focusing on an 80 µm region from the distal end to the distal extensions of Sh1, we followed DTC processes (and their fragments) and Sh1 filopodial extensions among the tightly packed germ cells (see Materials and methods). We observed germ cells that contact neither a DTC process (nor fragment thereof) nor filopodial extensions of Sh1, and thus represent bare regions. Therefore, this TEM analysis is consistent with the confocal analysis above using cell-type-specific markers in live worms, and it corroborates prior ultrastructural observations made on chemically fixed TEM preparations (Hall et al., 1999).

Video 4

Download asset

Video 5

Download asset

Previously, in fixed preparations, we observed that the distal edge of Sh1 was shifted nearly to the distal end of the gonad in worms bearing a hypomorphic mutation in the germline innexin inx-14(ag17) (Starich et al., 2014). We further investigated the position of Sh1 in inx-14(ag17) using live imaging of intact young adult hermaphrodites bearing contrasting markers for Sh1 and the DTC (see Materials and methods for details on markers used).

In contrast to inx-14(+), the distal border of Sh1 in inx-14(ag17) extends to the distal end of the gonad, well distal to the average position of the DTC processes (Figure 1C–E). The vast majority of inx-14(ag17) gonads fall into Class 3 (Figure 1), with consistent results from three different markers of Sh1 (Figure 1—figure supplement 1B–E). Although our classification scheme references Sh1 relative to DTC position, we do not attribute this gross positional alteration of Sh1 to a change in the DTC. Instead, within strains with similar markers, the change in the inx-14 genotype does not significantly change average DTC mean process length or the length of the longest DTC process, but it does have a significant effect on Sh1 distal border position (Figure 1G; Figure 1—figure supplement 1E).

To determine whether the position of Sh1 is shifted distally upon reduction of the somatic innexins, we investigated the position of Sh1 relative to the distal end of the gonad in a well-characterized compound reduction-of-function (rf) inx-8 mutant (Starich et al., 2020; Figure 1B–G). We found that Sh1 in worms bearing one partially functional somatic gonad innexin encoded by inx-8(tn1513tn1555) in an otherwise null inx-9(ok1502) background (hereafter referred to as ‘inx-8(rf)’; Starich et al., 2020) is also distally positioned, similar to inx-14(ag17) (Figure 1C–E). Thus, Sh1 is positioned distally as a result of reducing the function of either germ line or somatic innexins in the distal gonad.

We extended our analysis to inx-8(qy78[mKate2::inx-8]), an allele that encodes an N-terminal fusion protein of mKate2 and INX-8 that was used to mark Sh1 in a prior study (Gordon et al., 2020; Figure 2A). As detailed below, we found that—like inx-14(ag17) and inx-8(rf)—inx-8(qy78[mKate2::inx-8]) caused a distal shift in the position of Sh1 (Figure 2B–F). We documented this shift both by anti-INX-8 antibody (Starich et al., 2014) detection in otherwise unmarked strains (Figure 2B–C), as well as in live imaging of marked strains (Figure 2D–F).

Figure 2 with 3 supplements see all

Download asset Open asset

Figure 2—source data 1

Source data for Figure 2B and C.

https://cdn.elifesciences.org/articles/74955/elife-74955-fig2-data1-v1.xlsx

Download elife-74955-fig2-data1-v1.xlsx

Figure 2—source data 2

Source data for Figure 2D–F.

https://cdn.elifesciences.org/articles/74955/elife-74955-fig2-data2-v1.xlsx

Download elife-74955-fig2-data2-v1.xlsx

We used anti-INX-8 antibody staining to compare the distal border of Sh1 in the unmarked wild type and in inx-8(qy78[mKate2::inx-8]) gonads (N2 and DG5063, respectively; see Supplementary file 1; Figure 2—figure supplement 1). The mean distal boundary of INX-8 staining in the wild type lies at ~80 µm from the distal end of the gonad, while the mean distal boundary of INX-8 staining in worms carrying inx-8(qy78[mKate2::inx-8]) lies at ~50 µm from the distal end, and the difference between these means is significant (Figure 2B–C). Moreover, in 89% of wild-type gonads (n=28), the distal border of Sh1 is greater than 60 µm (~15–16 cell diameters) from the distal end of the gonad, while in 79% of inx-8(qy78[mKate2::inx-8]) gonads (n=32) the distal border of Sh1 is less than 60 µm from the distal end (Figure 2B–C).

Similar results were obtained using live imaging. Using bcIs39 to visualize Sh1, no Class 3 gonads (n=27) are observed in the wild type, compared to 65% (n=26) in inx-8(qy78[mKate2::inx-8)] (Figure 2D–E). The same mKate2::INX-8 fusion, independently generated in an inx-9(ok1502) null mutant background (i.e. inx-8(qy102[mKate2::inx-8]), see Materials and methods Gordon et al., 2020), was also observed to shift Sh1 distally in inx-8(qy102[mKate2::inx-8]) inx-9(ok1502) double mutants. In contrast, loss of inx-9 alone does not significantly affect Sh1 position (Figure 2—figure supplement 2). If mKate2::INX-8 were innocuous, the inx-8(qy102[mKate2::inx-8)] allele in the absence of inx-9 (inx-9(1502)) should have had no effect on Sh1 position.

To explain the observed distal Sh1 border shift of mKate2::INX-8 compared to previous descriptions of distal gonad architecture, Gordon et al., 2020 postulated that widely used markers of Sh1 driven by the lim-7 promoter may not mark the entirety of the cell, excluding the distal-most areas from detection. Although anti-INX-8 antibody staining clearly indicates a more proximal position of Sh1 in otherwise unmarked wild-type strains and an evident distal Sh1 shift in mKate2::INX-8, we wished to further determine whether the distal shift of Sh1 we observed in inx-8(qy78[mKate2::inx-8]) and inx-14(ag17) backgrounds reflected a disparity between the expression patterns of mKate2::INX-8 and lim-7p-driven markers. First, we consistently observe a dramatic distal shift of Sh1 in inx-14(ag17) using lim-7 promoter-driven transgenes (bcIs39 and tnIs5) as well as using a functional endogenously tagged allele of the fatty acid synthase gene fasn-1, [fasn-1(av138[fasn-1::GFP])] (Figure 1D; Figure 1—figure supplement 1). We next examined the overlap between the mKate2::INX-8 and lim-7p-driven markers in strains expressing both inx-8(qy78[mKate2::inx-8]) and GFP markers encoded by tnIs6 [lim-7p::gfp] or bcIs39 [lim-7p::ced-1::gfp]. In over 85% of gonad arms examined (n=20 and 26, respectively), the overlap between mKate2::INX-8 and GFP was complete (Figure 2—figure supplement 3). In both tnIs6 [lim-7p::gfp] or bcIs39 [lim-7p::ced-1::gfp], the remaining 15% of gonads displayed either reduced Sh1 expression or expression in one, but not both, cells of the Sh1 pair, which may result from stochastic transgene downregulation. Finally, we binned these gonads into phenotypic Classes 1–3 based on measurements taken while observing either mKate2 or GFP expression (Figure 2—figure supplement 3D). The number of Class 3 gonads is somewhat underestimated with the GFP markers. Nevertheless, we observe a striking distal shift of Sh1 in the inx-14 and inx-8 mutants bearing these markers. In sum, if the GFP markers were not marking the distal-most part of Sh1, we should not have been able to so readily detect the abnormal distal displacement of Sh1 in inx-14(ag17), inx-8(rf) or inx-8(qy78[mKate2::inx-8]) strains bearing these markers.

We conclude that despite variability in the exact position of the Sh1 border, and despite minor differences in transgene expression, bare regions between the Sh1 and the DTC are present in the wild type, and they are greatly diminished due to a distal shift in the position of Sh1 in worms bearing reduced inx-14 or inx-8 function or bearing the mKate2::INX-8 fusion.

Based on our observations that inx-8(qy78[mKate2::inx-8)] gonads display a distally shifted border of Sh1, we hypothesized that this allele might encode an INX-8 protein that poisons innexin complexes thereby interfering with the channel and/or adhesion functions important for Sh1 positioning. If this were due to a toxic effect of the mKate2::INX-8 fusion protein on gap junctions, we would predict that the distal shift of Sh1 would be dependent on the presence of the INX-8 coding region. To test this hypothesis, we used CRISPR-Cas9 genome editing to generate inx-8 null alleles in both the inx-8(qy78[mKate2::inx-8]) and wild-type genetic backgrounds. We generated deletions with identical breakpoints in the inx-8 locus in both genetic backgrounds [i.e. inx-8(qy78tn2031) and inx-8(tn2034)] starting 136 bp upstream of the wild-type inx-8 ATG start codon and extending 221 bp into inx-8 exon 3 (Figure 2 and Figure 2—figure supplement 2). In the inx-8(qy78[mKate2::inx-8]) context, this deletion also removes the mKate2 moiety. These deletions are expected to constitute inx-8 null alleles because, in addition to removing the start codon, they delete amino acids 1–349 (out of 382 amino acids), including virtually all residues essential for spanning the plasma membrane and forming a channel (Starich and Greenstein, 2020). Consistent with our hypothesis, the deletion mutant inx-8(qy78tn2031) displays a significantly lower frequency of Class 3 gonads than the original inx-8(qy78[mKate2::inx-8]) allele. Sh1 in the inx-8(qy78tn2031) deletion mutant is positioned more proximally, and therefore, like the wild type, exhibits bare regions that are lacking in inx-8(qy78[mKate2::inx-8]) mutants (Figure 2D–F). In addition, inx-8(qy78tn2031) exhibits a nearly normal brood size and suppresses the low-level embryonic lethality observed in the inx-8(qy78) starting strain (Table 1). Likewise, the identical deletion generated in the wild-type genetic background, inx-8(tn2034), also exhibits the wild-type Sh1 position (Figure 2—figure supplement 2), with a normal brood size and negligible embryonic lethality (Table 1). However, the observation of a small percentage of Class 3 gonads in inx-8(qy78tn2031) and inx-8(tn2034) animals but not in the wild type (Figure 2—figure supplement 2), suggests that the position of Sh1 is sensitive to the dosage of somatic gap junction hemichannels. In fact, the inx-8(qy78tn2031)/+ heterozygote reveals apparent haploinsufficiency for this phenotype (Figure 2—figure supplement 2).

We made several additional observations from a brood-size and embryonic-viability analysis (Table 1). First, the marker transgenes do not have an appreciable effect on brood size or embryonic lethality in our hands. Second, genetic interactions exist between inx-8(qy78[mKate2::inx-8]), inx-14(ag17) and the inx-9(ok1502) null mutation. Specifically, inx-14(ag17) enhances both the reduced brood size and embryonic lethality of inx-8(qy78[mKate2::inx-8]); whereas inx-9(ok1502) weakly suppresses the embryonic lethality defect (Table 1). Third, in the course of this analysis, we noted that strains bearing inx-8(qy78[mKate2::inx-8]) displayed highly variable population growth dynamics that we attribute to stress. Although the relevant stressor has yet to be determined, we observed this effect over multiple generations after starvation or transport. Given our previous finding that key metabolites such as malonyl-CoA are transported through these junctions to support embryonic growth, this finding deserves future scrutiny.

Finally, to assess more fully the nature of the inx-8(qy78[mKate2::inx-8]) allele with respect to Sh1 position, we generated and imaged worms bearing this allele in heterozygous and hemizygous configurations (Figure 2—figure supplement 2; we abbreviate inx-8(qy78[mKate2::inx-8]) to ‘qy78’ and null to (0), below, to aid comparison between genotypes). We infer loss-of-function behavior since other loss-of-function mutations affecting soma-germline gap junctions, such as inx-14(ag17), also cause a distal shift in the Sh1 border. Nevertheless, inx-8(qy78) is dominant: a substantial proportion of gonads in qy78/+ are in Class 3, similar to qy78/qy78, but distinct from deletion heterozygotes [inx-8(qy78tn2031/+)]. However, in a hemizygous configuration (qy78/0), the Sh1 distal displacement phenotype is largely suppressed. Thus, qy78 acts as a dosage-sensitive antimorph. We speculate that the fusion protein produced by inx-8(qy78[mKate2::inx-8]) can form dysfunctional junctions on its own or with INX-8(+), but not when it is present below a certain abundance threshold. As also mentioned above, the inx-8 deletion allele qy78tn2031, while largely suppressing the Sh1 distal mispositioning defect, displays a weak dominant haploinsufficient behavior consistent with dosage sensitivity. Alternatively, this behavior could result from genetic positional effects that impinge on nearby inx-9 function and for which Sh1 position is particularly sensitive.

A recent model proposed that the position of the Sh1 border influences the stem/non-stem decision in underlying germ cells (Gordon et al., 2020). However, because the previous study did not examine the position of stem or progenitor cells, and because the model was based on results using the antimorphic inx-8(qy78[mKate2::inx-8]) allele, we investigated this relationship.

In its simplest form, the model proposed by Gordon et al., 2020 predicts that when the distal edge of Sh1 is positioned distally, the stem/non-stem border should similarly shift distally. The SYGL-1 protein serves as a stem cell marker as sygl-1 is a direct transcriptional target of GLP-1/Notch in the germ line (Chen et al., 2020; Kershner et al., 2014; Lee et al., 2019; Lee et al., 2016; Shin et al., 2017). We analyzed the proximal extent of the pool of SYGL-1-positive cells bearing a well-characterized OLLAS epitope tag on SYGL-1 (Shin et al., 2017) and compared that boundary relative to the distal Sh1 border (Figure 3 and Materials and methods). In the case of inx-14(ag17), although the distal border of Sh1 was shifted drastically and significantly, there was no significant change in the position of the border of the SYGL-1-positive stem cell pool. In the case of the inx-8(qy78[mKate2::inx-8]) allele, the border of the SYGL-1-positive pool was marginally shifted distally relative to the wild type, though not commensurate with the extent to which Sh1 shifted distally in this background (Figure 3C). Furthermore, the marginal shift of the border of the stem cell pool in inx-8(qy78[mKate2::inx-8]) was suppressed when inx-8 was deleted in both inx-8(qy78tn2031) and inx-8(tn2034). The marginal shift in the border of the stem cell pool is not observed in inx-14(ag17), suggesting that such a defect in inx-8(qy78[mKate2::inx-8]) is due to the altered function of mKate2::INX-8, rather than the position of Sh1 per se (Figure 3A–C). To detect any subtle correlation between the proximal border of the SYGL-1 pool and the distal border of Sh1, we plotted these against one another and computed an R value (Figure 3D). By Pearson correlation, there is no significant relationship in any genotype examined between the position of the Sh1 border and the position of the SYGL-1-positive stem cell pool border. There is no significant relationship regardless of whether measurements are made in microns (Figure 3) or cell diameters (CD; Figure 3—figure supplement 1).

Figure 3 with 2 supplements see all

Download asset Open asset

Figure 3—source data 1

Source data for Figure 3 and Figure 3—figure supplement 1.

https://cdn.elifesciences.org/articles/74955/elife-74955-fig3-data1-v1.xlsx

Download elife-74955-fig3-data1-v1.xlsx

The recent model also proposed that Sh1 controls spindle orientation at the stem/progenitor border. However, we found that in the wild-type (marker-only) strain, the distal border of Sh1 was proximal to the proximal SYGL-1-positive border in 86% of the gonads (n=320), with a distance between them of ≥5 CD in 67% of the gonads (Figure 3 and Figure 3—figure supplement 1). As mentioned above, this 5 CD distance is inconsistent with the hypothesis that Sh1 is controlling spindle orientation at the stem cell border as such control would be expected to occur over a distance of only 1–2 cell diameters.

We conclude that (1) there is no correlation between the position of the distal border of Sh1 and the proximal border of the SYGL-1-positive stem cell pool, and (2) if spindle-oriented divisions occur at the Sh1 border, they are not influencing stem cell fate.

Although we found no correlation between the stem cell pool border and Sh1 position, we wondered whether altered Sh1 position might nevertheless influence the position of the border between the progenitor zone (PZ) and the transition zone that marks overt meiotic entry. In the wild type, consistent with our previous analysis in live worms, we found that the distal position of Sh1 is somewhat variable and can be either distal or proximal of the PZ border, using the length of the CYE-1-positive region to define the PZ border, following CYE-1 and pSUN-1(S8) co-staining (Figure 4 and Materials and methods; Mohammad et al., 2018). We found that although there is a subtle shift in the PZ border in inx-8(qy78[mKate2::inx-8]) and inx-14(ag17), it does not correlate with the dramatic shift in Sh1 position seen in these mutants (Figure 4). Again, we saw no significant correlation between Sh1 position and the position of the PZ border (Figure 4D). As with the SYGL-1 analysis, the relationships are similar when measurements are made in microns (Figure 4) or in CD (Figure 4—figure supplement 1), noting that differences between genotypes are more apparent in the micron measurements since cells are slightly larger more proximally in the PZ.

Figure 4 with 1 supplement see all

Download asset Open asset

Figure 4—source data 1

Source data for Figure 4 and Figure 4—figure supplement 1.

https://cdn.elifesciences.org/articles/74955/elife-74955-fig4-data1-v1.xlsx

Download elife-74955-fig4-data1-v1.xlsx

In sum, the border between the PZ and the transition zone is not markedly altered relative to the dramatic distal displacement of the Sh1 border in inx-14(ag17) or in inx-8(qy78[mKate2::inx-8]), suggesting that the position of Sh1 has no bearing on the position of meiotic entry.

Our studies show that impaired innexin function distally mispositions Sh1, but that this displacement does not similarly shift the proximal border of the adult stem cell pool (Figure 5). We show that bare regions of variable size normally exist between the DTC and Sh1, in which germ cells lack contact with the DTC or Sh1, but that these bare regions are diminished or lost with reduced innexin activity either in the soma or the germ line. While our studies rule out a role for Sh1 in stem cell fate, they demonstrate a role for innexins in somatic gonad architecture.

Figure 5

Download asset Open asset

A recent challenge (Gordon et al., 2020) to previous literature (primarily Hall et al., 1999; Starich et al., 2014) was made regarding the anatomical position of Sh1. Results presented here regarding Sh1 position concur with all these published observations in a variety of genetic scenarios and provide an explanation for apparent discrepancies, one that is consistent with all of the observations to date. Our results also refute the model proposed in Gordon et al., 2020 concerning the role of Sh1 in cell fate decisions, but reveal a novel role for innexins in cell positioning in this system. As requested by eLife, we include in the discussion below comments on the Li et al., 2022 manuscript as it appeared in bioRxiv July 18, 2022. The major points of resolution and remaining discrepancies concern (1) the normal anatomical position of Sh1, (2) the utility of a CED-1::GFP transgene to mark Sh1, (3) the implications of Sh1 position for the stem cell fate decision, and (4) brood size analyses.

Our observation that loss of the bare regions correlates with only a relatively modest effect on brood size (Table 1) raises the question of what, if any, biological significance the bare regions might have. Because the innexins have multiple functions in germline development (Starich et al., 2014; Starich et al., 2020; Starich and Greenstein, 2020), we are unable to ascribe the observed brood-size reduction in strains lacking bare regions to the absence of the bare regions per se. Thus, any potential roles for the bare regions vis-à-vis brood size and germline development will require additional study.

Our studies also provide evidence that innexin gap junctions not only serve as communication and adhesion junctions, but that in the context of an organ system, they contribute to the positioning of cells relative to each other. How might gap junctions influence the relative position of the DTC and Sh1 in the distal gonad arm? The DTC forms gap junctions with germ cells, which must be disassembled as germ cells enter a bare region, only to be reassembled again when in contact with Sh1 (and then again with more proximally located sheath cells Sh2–5). A detailed TEM analysis of the gonad (Hall et al., 1999) led to the consideration that a constant interplay of association and dissociation likely occurs between Sh1 and the underlying germ cells that move proximally along the arm: as germ cell flux continually moves germ cells towards the proximal end of the gonad, the Sh1 cells presumably extend their filopodia distally and form new gap junction connections with incoming germ cells. Otherwise, the bare regions would increase in length along the distal-proximal axis. Under optimal growth conditions in the laboratory, this constant association and dissociation between Sh1 and germ cells may be pushed to its limit.

To complement the role of gap junctions in promoting robust proliferation revealed by the inx-8 inx-9 double loss-of-function mutant or loss of the germline-encoded channel components, the kinetics of gap junction coupling between the somatic gonad and germ cells may play a role in determining the strength of the interactions between the two cell types. Unlike sheath–oocyte junctions, which form large plaques containing many functional gap junctions, the gap junctions formed in the distal arm appear to represent looser associations of a few gap-junction channels (Starich et al., 2014). Nonetheless, these associations may be sufficient to maintain adhesion with the underlying germline, functioning like regularly spaced rivets, albeit dynamic and removeable ones.

Disentangling the adhesive and channel functions of gap junctions is a complex issue. The mutants used in this study are competent to form soma–germline gap junctions—without these, germ cells fail to proliferate. However, they may form such junctions less efficiently than their wild-type counterparts. For example, the pattern of localization of gap junction puncta in inx-8(rf) and inx-14(ag17) appears more diffuse than in the wild type (Starich et al., 2014; Starich et al., 2020). Alternatively or additionally, the mutant innexins in this study may assemble into hemichannels as readily as wild-type innexins, but the pairing or opening of gap junction channels may be compromised. Studies of connexin gap-junction channels in paired Xenopus oocytes strongly suggest that opening of hemichannels facilitates their assembly into gap junctions. That study proposed hemichannel opening collapses the intermembrane space between juxtaposed cells to allow the extracellular loops of connexins to dock into gap junctions (Beahm and Hall, 2004). If a similar model applies to innexin-containing gap junctions, then rivet and channel function would be coupled.

How could impaired innexin function cause Sh1 to creep distally? One hypothesis is that when fewer junctions are made, Sh1 cannot adhere as well to germ cells. This scenario would be consistent with the observation that Sh1 is distally mispositioned when inx-8 and inx-9 are missing from Sh1 but inx-8(+) is present in the DTC by virtue of heterologous expression of inx-8(+) in the DTC in an inx-8(0) inx-9(0) double mutant (Starich et al., 2014). It is also possible that the DTC and Sh1 engage in an active repellent or a passive space-excluding interaction that somehow involves gap junction function. Another possibility is that a deficit in gap-junction function might trigger Sh1 spreading as a means to increase the surface area over which junctions may form to supply sufficient active biomolecules that transit through these junctions. Nevertheless, our studies show that the position of the germline stem cell decision is not dictated by the position of Sh1.

Live specimens were grown at 20 °C, and staged by picking mid-L4 larvae, then allowing them to grow at 20 °C until imaging them 24 hr later. Animals were immobilized using 10 mM Levamisole (Sigma T1512) in M9 buffer. Imaging was carried out on a Nikon W1 spinning disk confocal microscope. We define a biological replicate as any set of individuals collected, processed and imaged together. Numbers of individuals included in each analysis is indicated in the figure panels or legends.

Image analysis was carried out on two-dimensional maximum-projection Z-stack images of 3D confocal data. The distance from the distal end of the gonad to the end of each DTC process was measured along a line drawn from the end of each process parallel to the distal-proximal axis to a line drawn tangent to the distal end, orthogonal to the distal-proximal axis line (Figure 1—figure supplement 1). The distance between the distal end of the gonad and the most distal extent of Sh1 was measured in the same way. All data points were recorded for each sample and used to calculate the mean values presented in Figures 1 and 2 and related figure supplements.

Sh1 visualization in live worms: bcIs39 [lim-7p::CED-1::GFP] (Zhou et al., 2001) encodes a functional membrane-localized fusion to CED-1. tnIs5 and tnIs6 encode an identical non-functional fusion to the first 61 amino acids of LIM-7 (tnIs5 or tnIs6) denoted here as ‘lim-7p::GFP’ that includes 2.23 kb upstream, the first two exons, and the first intron of lim-7 fused to GFP (Hall et al., 1999). inx-8(qy78[mKate2::inx-8]) is an N-terminal fusion of mKate2 to INX-8 generated by CRISPR-Cas9 genome editing (Gordon et al., 2020). fasn-1(av138[fasn-1::gfp]) was generated by fusing GFP to the C-terminus of FASN-1 using CRISPR-Cas9 genome editing (Starich et al., 2020). tnEx42[acy-4::gfp +rol-6(su1006)] contains GFP fused to the ACY-4 C-terminus by recombineering in the fosmid WRM061bE10 (Govindan et al., 2009). ACY-4::GFP rescues the sterility of acy-4(ok1806) null mutants via its function in the gonadal sheath cells.

DTC visualization in live worms: naIs37[lag-2p::mCherry-PH] encodes mCherry fused to the PH domain of rat phospholipase C delta (Pekar et al., 2017) and qIs154[lag-2p::MYR-tdTomato] encodes a src kinase myristoylation sequence fused to tdTomato (Byrd et al., 2014).

3D rendering of confocal stacks for Videos 1–3 was done using Imaris (Oxford Instruments, plc. Abingdon, UK) according to the following algorithm for batch processing. For the Sh1 channel (green): Enable Smooth = true; Surface Grain Size = 0.433 µm; Enable Eliminate Background = false; Diameter Of Largest Sphere = 1.63 µm; Enable Automatic Threshold = false; Manual Threshold Value = 650; Active Threshold = true; Enable Automatic Threshold B=true; Manual Threshold Value B=31622.8. For the DTC channel (red): Enable Smooth = true; Surface Grain Size = 0.433 µm; Enable Eliminate Background = false; Diameter Of Largest Sphere = 1.63 µm; Enable Automatic Threshold = false; Manual Threshold Value = 710; Active Threshold = true; Enable Automatic Threshold B=true; Manual Threshold Value B=14988.5.

C. elegans strains (Supplementary file 1) were grown on standard NGM media with OP50 as a food source (Brenner, 1974) for all figures in this work. Strain constructions employed a modified NGM medium [containing 6.25 µg/ml Nystatin and 12.5 mM potassium phosphate pH 6.0 (added after autoclaving) and 200 mg/ml streptomycin sulphate (added before autoclaving)] with E. coli strain OP50-1 as food source. Strains were grown at 20 °C. In addition to the wild-type strain N2, the following alleles, described in WormBase (https://wormbase.org//#012-34-5) or in the cited references, were used:

Chr. I—dpy-5(e61) (Brenner, 1974), inx-14(ag17) (Miyata et al., 2008; Starich et al., 2014), unc-13(e51) (Brenner, 1974), fasn-1(av138[fasn-1::gfp]) (Starich et al., 2020), sygl-1(q983[3xOLLAS::sygl-1]) (Shin et al., 2017), unc-54(e190) (Brenner, 1974).

Chr. IV—inx-8(qy78[mKate2::inx-8]) (Gordon et al., 2020), inx-8(qy78 tn2031) (this work), inx-8(tn2034) (this work), inx-8(tn2075) (this allele was generated on the tmC5 balancer chromosome, this work), inx-9(ok1502), inx-8(qy102[mKate2::inx-8)] inx-9(ok1502) (Gordon et al., 2020), inx-8(tn1513tn1553) inx-9(ok1502) (Starich and Greenstein, 2020), inx-8(tn1513tn1555) inx-9(ok1502) (Starich et al., 2020), dpy-20(e1282) (Hosono et al., 1982).

Chr. V—acy-4(ok1806) (Govindan et al., 2009).

Balancer chromosomes Dejima et al., 2018 used included: tmC18 [dpy-5(tmIs1236)] I, tmC27[tmIs1239] I, tmC5[tmIs1220] IV.

Integrated transgenes included: naIs37[lag-2p::mCherry::plcdeltaPH +unc-119(+)] I (Pekar et al., 2017), mIs11[myo-2p::gfp +pes-10p::gfp +gut promoter::gfp] IV, bcIs39[lim-7p::ced-1::gfp +lin-15(+)] V (Zhou et al., 2001), qIs154[lag-2p::MYR::tdTomato +ttx-3p::gfp] V (Byrd et al., 2014), tnIs5[lim-7p::gfp +rol-6(su1006)] X, tnIs6[lim-7p::gfp +rol-6(su1006)] X (Hall et al., 1999), cpIs122[lag-2p::mNeonGreen::plcdeltaPH] (Gordon et al., 2020).

Extrachromosomal arrays used included: tnEx42[acy-4::gfp +rol-6(su1006)] (Govindan et al., 2009).

Multiply mutant strains were constructed in a straightforward manner (Huang and Sternberg, 1995). tmC18 was used as a balancer chromosome for inx-14(ag17). tmC27 was used as a balancer chromosome for sygl-1(q983). tmC5 (or the derivatives described) or mIs11 were used as balancer chromosomes for inx-8 and inx-9 mutant alleles. The presence of inx-14(ag17) in strains was verified by PCR and DNA sequencing. The ag17 allele was originally described as an Arg to His change in the second extracellular loop of INX-14, but the exact residue position was not specified (Miyata et al., 2008). A 1.2 kb PCR fragment covering this region was amplified with primers inx-14delF and inx-14delR (see Supplementary file 2 for the sequence of oligonucleotides used in this study). The PCR fragment was sequenced with the inx-14delR primer. No sequence changes were found in Arg residues predicted to occupy the second extracellular loop. However, a CGT to CAT (R326H) change was identified at a residue position predicted to lie near the cytoplasmic end of the fourth transmembrane domain, and we surmise that this change represents the original ag17 mutation. The presence of the sygl-1(q983[3xOLLAS::sygl-1]) mutation in strains was verified by PCR with primers sygl1-F and sygl1-R, which produce a 216 bp product in the wild type and a 348 bp product in sygl-1(q983) and by anti-OLLAS staining. The presence of inx-8(qy78tn2031) and inx-8(tn2034) in strains was verified by PCR with oligonucleotide primers inx8_delta.F and inx8_delta.R. To construct DG5367 inx-14(ag17) fasn-1(av138[fasn-1::gfp]) naIs37 I, fasn-1(av138[fasn-1::gfp]) naIs37/dpy-5(e61) unc-13(e51) heterozygotes were generated and Dpy non-Unc recombinants were sought, which resulted in the isolation of a dpy-5(e61) fasn-1(av138[fasn-1::gfp]) naIs37 I intermediate strain. From, inx-14(ag17)/dpy-5(e61) fasn-1(av138[fasn-1::gfp]) naIs37 heterozygotes, recombinants of genotype inx-14(ag17) fasn-1(av138[fasn-1::gfp]) naIs37/dpy-5(e61) fasn-1(av138[fasn-1::gfp]) naIs37 were identified by virtue of their increased GFP signal using a fluorescence dissecting microscope. The presence of inx-14(ag17) in the desired recombinants was verified by DNA sequencing. naIs37 was shown to be tightly linked to unc-54(e190) on the right end of LGI as follows. From dpy-5(e61) fasn-1(av138[fasn-1::gfp]) naIs37/unc-54(e190) heterozygotes, of 15 unc-54(e190) homozygotes, none contained the naIs37 marker (4 were recombinants expressing fasn-1::gfp). Of 38 wild-type segregants of the aforementioned parental heterozygous strain, all expressed the naIs37 marker (11 were recombinants lacking fasn-1::gfp expression). DG5380 was derived from DG5020 by crossing to N2 males and segregating away naIs37; the bcIs39 source was MD701, available from the CGC.

Wild-type N2 young adults were analyzed using high-pressure freezing/freeze-substitution freezing (HPF/FS; Hall et al., 2012). Four animals from chemical immersion and four animals from HPF/FS were collected in serial sections on slot grids, ranging from 80 nm to 100 nm thickness on an RMC PowerTome XL ultramicrotome (Eden Instruments, Valence, France). Sections were post-stained in 2% uranyl acetate in H₂O for 20 min and in Reynold’s lead for 3 min. Sections from each animal were viewed in the gonad region using Digital Micrograph software (Gatan, Pleasanton, CA) for the JEOL JEM-1400Plus M (Jeol USA, Peabody, MA), using a Gatan Orius SC1000B digital camera.

We selected the most optimally positioned and resolved HPF/FS-fixed animal (2% osmium tetroxide, 0.1% uranyl acetate, and 2% H₂O in acetone) to collect high-resolution montages of the gonad arm from every third section covering 80 µm, including the DTC and sheath filopodia of a posterior distal gonad arm. This series was collected from the tail to the vulva region to encompass the whole posterior gonad arm, but we did not collect full images closer to the gonad reflection in order to save effort and expense. The pixel size for the images was 3.23 nm. The images were aligned with TrakEM2 software (Cardona et al., 2012) and traced on a Wacom DTZ 2100D tablet (Wacom, Portland, OR) to build the 3D model.

We followed the DTC processes and their fragments and the filopodial extensions of Sh1 due to their increased translucency compared with the germ cells and by the fact that these somatic cellular extensions contained many small mitochondria (Video 4). We observed what we infer to be shed pieces of the DTC beginning at approximately 30 µm from the distal end, as has been observed using light microscopic methods (Byrd et al., 2014). However, some DTC and Sh1 processes may appear discontinuous due to the fact that micrographs of every third section were used to generate the reconstructed model (Video 5). Regardless of the effect that this sampling technique has on our ability to resolve contiguous versus shed bits of somatic gonadal cells, this TEM analysis is in agreement with our light microscopic observations that a large subset of distal germ cell progenitors are not in contact either with the DTC or Sh1.

Because inx-8(qy78[mKate2::inx-8]) brood counts and embryonic lethality were highly variable, we used a single batch of NGM media with OP50 food source for all brood count measurements reported in Table 1. To prevent mold formation, the medium contained 6.25 µg/ml Nystatin. L4-stage hermaphrodites were cultured individually on 35 mm Petri plates and transferred approximately every 24 hr until they stopped producing embryos (4–6 days). Worms that crawled off the media and died were redacted (varied from 0 to 10% depending on the experiment). Embryos that failed to hatch after 24–36 hr were counted and scored as dead. In the majority of cases, these embryos exhibited morphological abnormalities. Control experiments demonstrated that these embryos were not simply delayed and never hatched. Embryos that hatched were counted and scored as viable.

CRISPR-Cas9 genome editing was used to generate inx-8 null alleles in both the inx-8(qy78[mKate2::inx-8]) and wild-type genetic backgrounds. The approach taken generated identical 1525 bp deletions within the inx-8 locus in both genetic backgrounds starting 136 bp upstream of the wild-type inx-8 ATG start codon and extending 221 bp into inx-8 exon 3. In the inx-8(qy78[mKate2::inx-8]) context, this edit removes both the mKate2 moiety and inx-8. The deletions are expected to constitute inx-8 null alleles because, in addition to removing the start codon, they delete amino acids 1–349 (out of 382 amino acids), including virtually all residues essential for spanning the plasma membrane and forming a channel (Starich and Greenstein, 2020). The approach used pRB1017 to express two single guide RNAs (sgRNAs) under control of the C. elegans U6 promoter (Arribere et al., 2014). Oligonucleotides inx8_us_sgRNA1.F and inx8_us_sgRNA1.R were annealed and used to generate the plasmid inx8_us_sgRNA1 to direct Cas9 cleavage 136 bp upstream of the ATG initiator codon (Supplementary file 2 lists all oligonucleotides used in this study). Oligonucleotides inx8_sgRNA1.F and inx8_sgRNA1.R were annealed and used to generate the plasmid inx8_sgRNA1 to direct Cas9 cleavage in exon 3. To generate sgRNA clones, annealed oligonucleotides were ligated to BsaI-digested pRB1017 plasmid vector, and the resulting plasmids were verified by Sanger sequencing. pDD162 served as the source of Cas9 expressed under control of the eef-1A.1/eft-3 promoter (Dickinson et al., 2013). The repair template oligonucleotide used was inx8_rpr. Genome editing employed the dpy-10 co-conversion method (Arribere et al., 2014). The injection mix contained pJA58 (7.5 ng/μl), AF-ZF-827 (500 nM), inx8_us sgRNA1 (25 ng/μl), inx8_sgRNA1 (25 ng/μl), inx8_rpr (500 nM), and pDD162 (50 ng/μl) and was injected into adult hermaphrodites from strains DG5131 naIs37[lag-2p::mCherry::PH +unc-119(+)] I; inx-8(qy78[mKate2::inx-8]) IV; bcIs39[lim-7p::ced-1::gfp +lin-15(+)] V and DG5020 naIs37[plag-2::mCherryPH +unc-119(+)] I; bcIs39[lim-7p::ced-1::gfp +lin-15(+)]V. Correct targeting was verified by conducting PCR with primer pairs inx8_delta.F and inx8_delta.R followed by DNA sequencing. Three deletion alleles were recovered from the injections into DG5131 (qy78tn2031, qy78tn2032, and qy78tn2033), and two deletion alleles were recovered from the injections into DG5020 (tn2034 and tn2035). The deletion alleles were outcrossed to naIs37[lag-2p::mCherryPH +unc-119(+)]/+I; tmC5(tmIs1220[myo-2p::Venus])/+IV; bcIs39[lim-7p::ced-1::gfp +lin-15(+)]/+V males to generate DG5229 and DG5232. Homozygous strains were analyzed by confocal microscopy. To introduce an inx-8 null allele onto the tmC5[tmIs1220] IV balancer chromosome, the same injection mix of CRISPR reagents as described above was injected into dpy-20(e1282)/tmC5[tmIs1220] IV heterozygotes. The inx-8(tn2075) null allele was generated on the balancer chromosome, which resulted from a 1526 bp deletion that differed from the other CRISPR-generated inx-8 null alleles by the removal of an extra base at the cut site.

Immunostaining was carried out as described (Mohammad et al., 2018). Briefly, synchronized adult hermaphrodites, 24 hr past mid-L4, were dissected in PBST (PBS with 0.1% Tween 20), with 0.2 mM levamisole to extrude the gonads. The gonads were fixed in 3% paraformaldehyde solution for 10 min and then post-fixed in −20° chilled methanol for 10 min. After 3x10 min washes with PBST, they were blocked in 30% goat serum for 30 min at RT. The gonads were then incubated with the desired primary antibodies diluted (see below) in 30% goat serum at 4° overnight. The next day, after 3x10 min PBST washes, the gonads were further incubated with appropriate secondary antibodies, diluted in 30% goat serum, at 4° overnight. The gonads were washed 3 times with PBST, then incubated with 0.1 g/ml DAPI in PBST for 30 min. After removal of excess liquid, the gonads were mixed with anti-fading agent (Vectashield) and transferred to an agarose pad on a slide. Hyperstack images were captured using a spinning disk confocal microscope (PerkinElmer-Cetus, Norwalk, CT). Two overlapping hyperstack images were captured for each gonad arm to obtain coverage of >50 cell diameters from the distal end of the gonad. Images were further processed in Fiji, and DAPI stained nuclei were used to mark the number of cell diameters from the distal end. Employing pixel to micron ratio, specific to the images captured, cell diameters were converted into microns where required.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We thank Gabriela Huelgas-Morales for discussions and constructive suggestions during the course of this work, Swathi Arur for the affinity purified anti-GFP antibody, Verena Jantsch for anti-pSUN-1 antibody, Edward Kipreos for anti-CYE-1 antibody. We thank Michael Cammer and the NYU Langone Microscopy Core for experimental and technical support; the NYU Langone Microscopy Laboratory is partially supported by the Cancer Center Support Grant P30CA016087 at the Laura and Isaac Perlmutter Cancer Center. We thank WormBase. We thank Kacy Gordon and the CGC (which is funded by NIH Office of Research Infrastructure Programs P40 OD010440), for strains. NIH P30HD071593 and 1S10OD016214-01A1 supports electron microscopy in DHH laboratory.

© 2022, Tolkin, Mohammad et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.