The C. elegans NR4A nuclear receptor gene nhr-6 promotes cell cycle progression in the spermatheca lineage

Abstract

Background

NR4A nuclear receptors are a conserved, functionally diverse group of nuclear receptors that regulate multiple cellular processes including proliferation and differentiation. The gene nhr-6 encodes the sole C. elegans NR4A nuclear receptor homolog with an essential role in reproduction by regulating morphogenesis of the spermatheca, a somatic gonad organ involved in ovulation and fertilization.

Results

Here, we identify the spermatheca cell lineage defects that occur in nhr-6 mutants. Utilizing cell marker analysis, we find that nhr-6 is required for cell cycle progression and that the cell proliferation phenotype is not due to premature cell cycle exit. We also show that loss of the negative cell cycle regulators fzr-1 and lin-35 suppresses the cell proliferation defects. We further demonstrate that NHR-6 activity intersects with Eph receptor signaling during spermatheca cell proliferation.

Conclusion

NHR-6 has an essential function in promoting cell cycle progression during G1 phase in a specific spermatheca cell lineage. Genetic suppression of the proliferation phenotype does not affect the differentiation phenotypes observed in nhr-6 mutants, indicating a dualistic role for nhr-6 in in regulating cell proliferation and cell differentiation during spermatheca organogenesis.

Introduction

The proper development of an organism is dependent upon the strict regulation and coordination of both cell proliferation and differentiation. How and when a cell makes the decision to divide or differentiate is dependent upon various developmental cues that lead to the activation or repression of specific genes (Sears and Nevins, 2002). The nematode Caenorhabditis elegans serves as an excellent model for the study of developmental regulation of both cell proliferation and differentiation as its invariant somatic cell lineage allows for detailed analysis of these processes at the single-cell level (van den Heuvel and Kipreos, 2012).

Key to understanding the cellular mechanisms of proliferation and differentiation during development is the determination of the role of epigenetic regulators that control these processes. During proliferation, mitogens stimulate the expression of members of the cyclin family of proteins along with their catalytic binding partners, the cyclin dependent kinases (CDKs) (Morgan, 1997; Kerkhoff and Rapp, 1998). Expression of cyclins and CDKs drives progression of the cell cycle by phosphorylating key substrates. Cyclin D forms a complex with CDK-4/6 and phosphorylates the retinoblastoma protein (pRB), which releases E2F transcription factors and leads to expression of genes required for progression into S-phase including cyclin E and A (Lees et al., 1993; Ohtani et al., 1995; Mittnacht, 1998). Exit from the cell cycle during terminal differentiation requires repression of cyclin-CDK complexes by cyclin dependent kinase inhibitors (CKIs) as well as the activity of pRB and other negative regulators (Sherr and Roberts, 1999; Buttitta and Edgar, 2007). Additionally, terminal differentiation signals a different set of cell-type specific genes to be expressed that determine the size, shape and function of the cell (Levine and Tjian, 2003). During development, proliferation and differentiation are intimately linked and under transcriptional control, therefore, understanding the function of different transcription factors involved is of vital importance.

One interesting group of transcription factors with critical roles in regulating cell proliferation and differentiation is the NR4A subgroup of the nuclear receptor (NR) superfamily (Mohan et al., 2012). NR4A NRs act as early response genes that are induced by an array of signals including fatty acids, growth factors, cytokines and neurotransmitters (Maxwell and Muscat, 2006). Unlike canonical NRs, they are said to be orphan receptors as they have no known natural ligand and contain an unusually hydrophobic ligand binding pocket that is not conducive to classical ligand-mediated NR regulation (Wang et al., 2003; Wansa et al., 2003). There are three NR4A mammalian paralogs; Nur77 (NR4A1), Nurr1 (NR4A2) and NOR1 (NR4A3), that are expressed in a wide variety of tissues and display cell and tissue specific functions. It has been shown that mammalian NR4A NRs are induced by mitogenic stimuli and regulate genes involved in cell cycle progression as well as cell cycle exit and differentiation (Kolluri et al., 2003; Nomiyama et al., 2006; Wingate et al., 2006). For example, Nur77 and NOR1 play a role in proliferation of islet β-cells by regulating cyclin levels and the E2F1 transcription factor (Tessem et al., 2014) while NOR1 also transcriptionally regulates the S-phase kinase associated protein, SKP2, during proliferation of vascular smooth muscle cells (Gizard et al., 2011); conversely Nur77 induces cell cycle exit and differentiation in dopamine cells (Castro et al., 2001). Furthermore, NR4A NRs can exhibit opposing functions in regulating these cellular processes, as observed during vascular remodeling and disease (Zhao and Bruemmer, 2010).

The ability to rapidly respond to various physiological stimuli suggests that NR4A NRs likely function as critical regulators of many developmental processes. This is supported by gene knockdown studies in mice that that lead to defective development of different tissue and organ types (Ponnio et al., 2002; Mullican et al., 2007; Sekiya et al., 2013; Vacca et al., 2013). Nonetheless, much remains to be learned regarding NR4A NR regulation of organ formation and tissue differentiation. In C. elegans, the gene nhr-6 encodes the sole homolog of the NR4A NR (Kostrouch et al., 1995; Gissendanner et al., 2004). nhr-6 is a lineage-specific regulator of the spermatheca, a functionally complex somatic gonad organ that functions in oocyte fertilization and ovulation (Gissendanner et al., 2008). nhr-6 mutants exhibit severe morphogenetic defects in the spermatheca, including a decreased number of cells that is likely due to a failure of cells to undergo the correct number of divisions in the spermatheca cell lineage. This indicates that the cellular functions of NR4A NRs are conserved in C. elegans. NHR-6 also exhibits DNA binding activities similar to NR4A NRs and a functional DNA binding domain is required for NHR-6 activity (Heard et al., 2010), suggesting a primary function for NHR-6 in regulating transcriptional programs that promote cell proliferation and cell differentiation.

Here, we further define the role of nhr-6 in regulating spermatheca development. We show that nhr-6 is required at a specific point in the spermatheca lineage to promote cell cycle progression. We also show that the loss of negative G1/S regulators can suppress the proliferation phenotype seen in nhr-6 mutants, indicating a specific cell cycle function for nhr-6. We also report that NHR-6 functions with Eph receptor signaling in regulating spermatheca development. These results provide a mechanistic framework for elucidating the transcriptional regulation of organ formation by NR4A NRs.

Results

nhr-6 is required for spermathecal cell proliferation in early L4

The C. elegans spermatheca is a tube-like somatic gonad organ that stores sperm and is the site of oocyte fertilization (McCarter et al., 1997). During somatic gonad development two SS (sheath/spermathecal precursor) cells undergo several rounds of division in the mid L3 and early L4 larval stages to give rise to 18 spermathecal cells as well as 5 pairs of sheath cells; 6 more spermatheca cells are derived from the uterine lineage giving the adult spermatheca 24 total cells (Kimble and Hirsh, 1979). The adult spermatheca is characterized by a narrow distal constriction made of 8 cells while the other 16 cells make up the wider bag-like chamber that connects to the proximal spermatheca-uterine (SP-UT) valve (Kimble and Hirsh, 1979). Animals homozygous for the nhr-6 deletion null allele, lg6001, display an adult spermatheca with severe morphological defects, including loss of the spermatheca-uterine valve and distal constriction, as well as having ∼1/2 the total number of cells (Gissendanner et al., 2008).

To better understand the nhr-6(lg6001) mutant spermatheca phenotype, we examined spermatheca development in cohorts of wild-type and mutant animals from mid L3 until mid L4 when the spermatheca cell divisions are taking place. By mid L3 in wild-type animals, two SS daughter cells are generated that will contribute 9 cells each to the spermatheca (Fig. 1A). We gave the designation “SPC” (spermatheca precursor cell) to these SS daughter cells. The SPC corresponds to the lineage designations Z1.paapp and Z1.appp for the anterior spermatheca (Fig. 1A) and Z4.paaa and Z4.appaa for the posterior spermatheca. By the end of L3 (during the L3 lethargus) in wild-type animals, each SPC had undergone an asymmetric division giving rise to a large distal daughter cell (Z1.apppa/paappa and Z4.appaap/Z4.paaap) and a smaller more proximal daughter cell (Z1.apppp/paappp and Z4.appaaa/Z4.paaaa) (Fig. 1A, B). In early L4 animals, both SPC daughter cells had divided again (the final division for Z1.apppp/paappp and Z4.appaaa/Z4.paaaa) (Fig. 1A, B) and by early/mid L4 another round of divisions had completed yielding 12 SPC-derived spermatheca cells (Fig. 2A). By mid L4 all SPC lineage divisions were completed and the 18 SPC-derived cells, along with the 6 uterine lineage-derived cells, formed an elongated organ separated from the gonad arm by gonadal sheath pair 5 (Fig. 2A).

Figure 1. Schematic representations and DIC micrographs of developing wild-type and nhr-6 mutant spermathecae in late L3 and early L4 animals.

(A) SS cell lineage. Adapted from (Kimble and Hirsh, 1979) (B) Illustrations of the anterior somatic gonad nuclei configurations and corresponding micrographs for WT and nhr-6 mutant animals from mid L3 to early L4. SPC-derived spermatheca nuclei are dark green; uterine lineage-derived spermatheca nuclei and precursor nuclei are shaded dark gray; sheath cell nuclei and precursor nuclei are shaded light gray; sujn and sujc, which will form the SP-UT valve, are colored blue and red, respectively. Germline nuclei are light green. See color key in Fig. 2. In mid L3, Z1.papaa and Z4.apaaa (not shown) each generates the DE1 and DE2 precursor cells (Newman et al., 1996). DE1 divides to generate two spermatheca cells and DE2 divides to generate a spermatheca cell and a sujc. Only Z1.papaa-derived DE1 and DE2 nuclei are indicated. Z1.appa and Z1.paapa are sheath cell precursors that will generate sheath cell pairs 2-5. In the WT animal shown in early L4, the proximal SPC daughter cell has not yet divided. Bar = 5 μm.

Figure 2. Spermatheca cell lineage defects at early/mid and mid L4 in nhr-6 mutant animals.

(A) Illustrations of the anterior somatic gonad nuclei configurations and corresponding micrographs for WT and nhr-6 mutant animals from early/mid L4 to mid L4. Bar = 5 μm. (B) SPC lineage phenotypes in nhr-6 mutants. The nhr-6 mutant animals shown in (A) are representative of the “3-cell” phenotype.

In nhr-6 mutant animals, a normal spermatheca developmental pattern was observed from mid L3 through L3 lethargus giving rise to the four SPC daughter cells (Fig. 1B). Following the rapid division phase of early L4 through early/mid L4, it appeared in most animals that the two larger more distal SPC daughter cells had divided normally while the smaller proximal SPC daughter cells had not divided (Fig. 2A).No additional spermathecal cell divisions were apparent as the animals continued development (see Fig. 2A at mid L4). In contrast to the SPC lineage, the uterine lineage-derived spermatheca cells exhibited a normal developmental pattern in all mutant animals observed.

Since the spermatheca cell number phenotype in nhr-6 mutants is variable, we sought to more precisely characterize and quantify the SPC lineage defects in mutant animals. We analyzed the cell divisions of both anterior SPCs beginning at the L3 lethargus in single nhr-6(lg6001) animals. In all nhr-6 mutant animals observed (n=12) the asymmetric SPC division in L3 lethargus occurred normally. In the majority of animals (8/12) the larger distal SPC daughter cell (Z1.apppa and Z1.paappa) underwent an additional division but further cell divisions ceased after this point. The smaller proximal SPC daughter cell (Z1.apppp and Z1.paappp) in these animals did not divide (see Fig. 2B, 3 cell SPC phenotype) yielding a mid L4 spermatheca nuclei count of 12 (including the uterine-derived spermatheca cells). In a smaller number of animals (3/12), the Z1.apppa and Z1.paappa cells as well as Z1.apppp and Z1.paappp cells divided once and ceased divisions to give a mid L4 spermatheca nuclei count of 14 (see Fig. 2B, 4 cell SPC phenotype). Only one of the 12 animals followed failed to undergo a division in both SPC daughter cells and displayed a mid L4 spermatheca nuclei number of 10 (see Fig. 2B, 2 cell SPC phenotype). This observation explains the variable number of spermatheca nuclei in nhr-6 mutants and we conclude that nhr-6 is required for both Z1.apppa and Z1.paappa SPC daughter cells to undergo the correct number of divisions in early L4.

SPC-derived cells in early L4 fail to express an S-phase marker in nhr-6 mutants

Cell cycle progression is tightly regulated in somatic cells during C. elegans larval development, particularly at the transition from G1 into S phase (Euling and Ambros, 1996; Koreth and van den Heuvel, 2005). The lineage analysis suggested that the decreased cell number phenotype of nhr-6 mutants was due to the permanent failure of specific SPC daughter cells to divide and not a delay in cell cycle progression. To further test this, we examined cell cycle progression using a ribonucleotide reductase reporter gene (rnr∷GFP) as an S-phase marker. rnr expression closely correlates with the timing of S-phase entry in the developing worm (Hong et al., 1998). We compared expression patterns of the rnr∷GFP transcriptional reporter between wild type and nhr-6(lg6001) animals in the SPC cell lineage. Wild type animals in late L3 exhibited bright rnr∷GFP expression in the SPC and in uterine and gonadal sheath precursor cells (Z1.appa/papa) that are dividing at this stage (Fig. 3A, A′, B, B′; N=77). At the time point shown, the SPC is very close to undergoing a division in late L3. After SPC division, rnr∷GFP expression is off during L3 lethargus but resumes in early L4 when SPC daughter cell divisions begin. Expression of rnr∷GFP continues as cells in the SPC lineage undergo rapid divisions during early and early/mid L4 development (Fig. 3E,E′,F,F′; N=78). At the latter stage, compared to the SPCs, uterine cells showed weaker expression and no expression was detected in the gonadal sheath cells which ceased dividing.

Figure 3. rnr∷GFP expression in the developing spermatheca of nhr-6 mutants.

Only the anterior spermatheca is shown. Late L3 DIC (A-D) and rnr∷GFP epifluorescence (A′-D′) micrographs of the developing spermatheca in rnr∷GFP (A,A′,B,B′) and nhr-6(lg6001);rnr∷GFP (C,C′,D,D′) animals. Both SPCs (arrows) and both gonadal sheath precursors (arrowheads) are indicated. The fainter signal in the sheath precursor is likely due to the cell being slightly earlier or slightly later in the cell cycle relative to the SPC at the point the image was taken. rnr∷GFP expression in both the SPC and gonadal sheath precursor has been observed in all animals scored at this time point. Early L4 DIC (E-H) and rnr∷GFP epiflourescence (E′-H′) micrographs of the developing spermatheca in rnr∷GFP (E,E′,F,F′) and nhr-6(lg6001);rnr∷GFP (G,G′,H,H′) animals. SPC daughter cells (asterisks) and gonadal sheath (arrowheads) daughter cells in the plane of focus are indicated.

The expression pattern of nhr-6 mutants during late L3 (Fig. 3C,C′D,D′; N=83) was identical to that seen in wild type animals. Expression was seen in the SPC as well as in the uterine tissue and both gonadal sheath precursor cells. In early L4 expression of rnr∷GFP was not observed and expression remained absent for the remainder of L4 development (Fig. 3G, G′, H, H′; N=84). The cells at this stage are the Z1.apppa/paappa daughter cells (SPC granddaughter cells) that fail to divide in the lineage analysis, thus the lack of rnr∷GFP expression in these cells indicates a failure in cell cycle progression and the continued absence of expression indicates that these cells do not enter S-phase prior to onset of terminal differentiation. rnr∷GFP expression in other somatic gonad cells is similar to wild-type.

Loss of nhr-6 function does not cause early cell cycle exit

One explanation for the nhr-6(lg6001) cell cycle block in the SPC daughter cells is that they have exited the cell cycle early. Members of the cyclin-dependent kinase inhibitor (CKI) family have a well-documented role in promoting cell cycle exit by inhibiting the action of G1 CDKs and are highly expressed during terminal differentiation (Sherr and Roberts, 1995; Harper and Elledge, 1996). The C. elegans CKI family member, cki-1, encodes a CIP/KIP protein that also functions in regulating cell cycle progression. cki-1 has been shown to act as a negative G1/S regulator which can prematurely arrest cells in G1 when ectopically expressed and displays a reciprocal expression pattern to rnr with expression observed at the onset of terminal differentiation (Hong et al., 1998; Buck et al., 2009). To test if the SPC daughter cells exit the cell cycle early we examined cki-1∷GFP expression in wild-type and nhr-6 mutant animals during L4.

In early/mid L4, cki-1∷GFP animals show expression in the uterine and vulva tissue as those cells have completed divisions (Fig. 4A,B; N=59). However, the SPC daughter cells lack cki-1∷GFP expression as they are undergoing several rounds of divisions at this time. The same expression pattern is seen in nhr-6(lg6001);cki-1∷GFP animals (Fig. 4C,D; N=57), indicating they have not exited the cell cycle at this stage. To ensure that cki-1 was being expressed during differentiation of the spermatheca we also looked at mid L4 animals. During mid L4 all cell divisions have ceased and the adult spermatheca structure is beginning to form. Wild-type cki-1∷GFP animals show strong expression at this developmental stage throughout the spermatheca (Fig. 4E,F; N=40) as well as the developing uterine tissue. nhr-6(lg6001);cki-1∷GFP animals also showed expression in the spermatheca at this developmental stage (Fig. 4G,H; N=52), indicating cell cycle exit and onset of terminal differentiation. These results show that the SPC daughter cells of nhr-6 mutants do not appear to exit the cell cycle early and that terminal differentiation is signaled at the same time point as it occurs in wild-type animals.

Figure 4. cki-1∷GFP expression in the spermatheca of nhr-6 mutant animals during the L4 stage.

Only the anterior spermatheca is shown. Early/mid L4 DIC (A,C) and cki-1∷GFP epifluorescence (B,D) micrographs of cki-1∷GFP (A,B) and nhr-6(lg6001);cki-1∷GFP (C,D) animals. Spermatheca nuclei are indicated by arrows, gonadal sheath nuclei are indicated by arrowheads. Mid L4 DIC (E,G) and cki-1∷GFP epiflourescence (F,H) micrographs of cki-1∷GFP (E,F) and nhr-6(lg6001);cki-1∷GFP (G,H) animals. The mid L4 spermatheca is outlined in white (E,G). Bar = 10 μm.

Loss of lin-35 and fzr-1 suppresses the nhr-6 cell number phenotype

The previous data suggests that nhr-6 functions to promote G1/S progression in the SPC lineage. If so, we would expect the loss of function of negative G1/S regulators to suppress the decreased cell number phenotype. Two important negative regulators of the G1/S transition in C.elegans are lin-35, which encodes the homolog of retinoblastoma protein (pRB) (Lu and Horvitz, 1998), and fzr-1, which encodes the homolog of Cdh1/Hct/FZR, a G1/S-specific subunit of the anaphase-promoting complex (Fay et al., 2002; Pesin and Orr-Weaver, 2008). lin-35 and fzr-1 regulate G1/S transition by negatively regulating G1 cyclin levels (Boxem and van den Heuvel, 2001; Fay et al., 2002). While single mutants do not display obvious cell proliferation phenotypes, lin-35;fzr-1 double loss-of-function mutants display hyperproliferation in several tissues indicating redundant functions in the regulation of G1 cyclins (Fay et al., 2002). To test for suppression we assessed spermatheca nuclei numbers in young adults with fzr-1;nhr-6 and lin-35;nhr-6 double loss of function.

We first examined spermatheca nuclei numbers in fzr-1;nhr-6 double loss-of function mutants using two different mutant alleles of fzr-1, ku298 and ok380, which on their own display a normal spermatheca with a normal number of nuclei (Fig. 5). When we observed the nhr-6(lg6001);fzr-1(ku298) and nhr-6(lg6001);fzr-1(ok380) double loss-of-function mutants there was a significant increase in the total number of nuclei when compared to nhr-6(lg6001) single mutants. WT animals had an average of 23.9 spermatheca nuclei, nhr-6 mutant animals had an average of 11.9 spermatheca nuclei, nhr-6(lg6001);fzr-1(ku298) double loss-of-function mutants had on average 18.9 spermatheca nuclei while the nhr-6(lg6001);fzr-1(ok380) double loss-of-function mutants displayed an average of 16.3 (Fig. 5A). Interestingly, a small number of the nhr-6(lg6001); fzr-1(ku298) animals (2/32) observed had the normal wild type number of spermatheca nuclei. Despite the improvement in cell number, the number of eggs laid by the nhr-6;fzr-1 double mutants was extremely low, similar to nhr-6 mutants (Fig. 5B). The increase in nuclei number in the double mutants corresponded to an increase in spermatheca size in comparison to nhr-6 single mutants (Fig. 5C-F). The SP-UT valve did not differentiate normally and the normal morphology of the distal constriction was also absent, even when spermatheca nuclei numbers were normal or near normal. The low numbers of eggs laid in the double mutants was likely due to the malformed spermatheca. In addition, a large proportion of the eggs laid exhibited an abnormal morphology that arises from defective ovulation, as has been previously observed in nhr-6 mutant animals (Gissendanner et al., 2008).

Figure 5. Suppression of the nhr-6 mutant spermatheca cell number phenotype by fzr-1 and lin-35 loss of function.

(A) Graph showing the average number of spermatheca nuclei in WT (N2) (N=20), fzr-1(ku298) (N=15), fzr-1(ok380) (N=15), nhr-6*lg6001)(N=20), fzr-1(ku298);nhr-6(lg6001) (N=32), and fzr-1(ok380);nhr-6(lg6001)(N=16) animals (p<.01(*);Student's t test). (B) Graph showing the average number of eggs laid for WT (N2) (N=10), fzr-1(ku298) (N=10), fzr-1(ok380) (N=10) and nhr-6(lg6001) (N=15), fzr-1(ku298);nhr-6(lg6001) (N=10) and fzr-1(ok380);nhr-6(lg6001) (N=15) animals (inset). (C,D,E,F) DIC micrographs of young adult spermatheca morphology in WT(C), fzr-1(ku298) (D), nhr-6(lg6001) (E), and nhr-6(lg6001);fzr-1(ku298) (F) animals. The double mutant shown in (F) had a WT number (24) of spermatheca nuclei. The spermatheca (white brackets), spermatheca-uterine valve (arrow), distal region of the spermatheca (asterisk) and gonadal sheath cells (arrowheads) in the plane of focus are indicated for each. In nhr-6(lg6001) and nhr-6(lg6001);fzr-1(ku298) animals the differentiation of the SP-UT is abnormal and the distal region of the spermatheca lacks the constricted morphology observed in WT and fzr-1(ku298) single mutants. Bar = 10 μm (G) Graph showing the average number of spermatheca nuclei in eri-1(mg366); GFP (RNAi) (N=25), lin-35(n745);eri-1(mg366); GFP (RNAi) (N=15), eri-1(mg366);nhr-6(RNAi) (N=25) and lin-35(n745);eri-1(mg366); nhr-6(RNAi) (N=30) animals (p<.01(*);Student's t test). (H) Graph showing the average number of eggs laid for eri-1(mg366); GFP (RNAi) (N=10), lin-35(n745);eri-1(mg366); GFP (RNAi) (N=10), eri-1(mg366);nhr-6(RNAi) (N=10) and lin-35(n745);eri-1(mg366); nhr-6(RNAi) (N=10) animals. Error bars, ± standard deviation.

Next, we assessed whether loss of lin-35 could restore SPC daughter cell numbers. Unfortunately, we were unable to construct an nhr-6(lg6001);lin-35(n745) double loss-of-function strain due to fully penetrant sterility (data not shown). Therefore, we utilized an RNAi approach. C. elegans RNAi has been shown to be substantially enhanced by loss of lin-35 and eri-1 gene function (Kennedy et al., 2004; Lehner et al., 2006a); therefore we assessed spermatheca cell numbers following nhr-6 RNAi in eri-1(mg366) single and lin-35(n745); eri-1(mg366) double loss-of-function genetic backgrounds. Control RNAi animals for eri-1 mutant and lin-35;eri-1 double mutant animals did not display any marked abnormalities in spermatheca nuclei number or morphogenesis (Fig. 5G). nhr-6 RNAi on eri-1 mutant animals gave a strong nhr-6 loss of function phenotype with an average nuclei number of 12.9, similar to nhr-6 mutants. When nhr-6 RNAi was performed on lin-35;eri-1 double mutant animals, the average nuclei number significantly increased to 14.8 with a small number of animals (2/30) displaying almost wild type nuclei numbers (Fig. 5G). Similar to fzr-1;nhr-6 double mutants, the animals in these experiments exhibited defective spermatheca development and severely reduced brood sizes despite the increase in cell number (Fig. 5G). Since fzr-1 and lin-35 function in parallel, and fzr-1; lin-35 double mutants display hyperproliferation in multiple tissues, we performed lin-35 RNAi on fzr-1;nhr-6 double mutant animals. However, we failed to observe any changes in cell number in these experiments (data not shown).

Taken together the independent loss of two major G1/S transition regulators was able to significantly increase the number of spermatheca nuclei in nhr-6 mutants, supporting a role for nhr-6 in G1/S cell cycle progression in the spermatheca lineage.

NHR-6 interacts with Eph receptor signaling during spermatheca cell proliferation

Regulation of NR4A nuclear receptor activity is mediated by post-translational modifications through interactions with various signaling cascades (Maxwell and Muscat, 2006). In a previous study, an RNAi-based screen for modifiers of signaling pathways in C. elegans identified nhr-6 as a modifier of vab-1, which encodes the C. elegans homolog of the Eph receptor (George et al., 1998; Lehner et al., 2006b). One of the major functions of Eph receptor signaling is in the regulation of morphogenetic processes (Singh et al., 2012). While Eph receptor signaling has not been previously shown to function during spermatheca development, the reported interaction with nhr-6 resulted in sterility, suggesting that vab-1 could have such a role, possibly functioning together with NHR-6 to regulate one or more spermatheca developmental processes.

To confirm the interaction, we performed nhr-6 RNAi on four different vab-1 alleles (Table 1). vab-1(e2) and vab-1(e699) are considered partial loss of function alleles and are mutations that specifically affect the receptor kinase and extracellular domains, respectively (George et al., 1998). Homozygotes for vab-1(e2) and vab-1(e699), as well as the null alleles dx31 and e2027, exhibit partially penetrant embryonic and larval arrest phenotypes. The animals that escape early arrest can develop into reproductive adults with robust brood sizes (Table 1), indicating that Eph receptor signaling is not critically essential for reproduction. nhr-6 RNAi has previously been demonstrated to be hypomorphic in a wild-type background with affected animals exhibiting moderately reduced brood sizes; thus, any strong enhancement of nhr-6 RNAi should result in significantly reduced brood sizes (Gissendanner et al., 2014). As shown in Table1, all vab-1 mutations strongly enhanced the brood size phenotype of nhr-6(RNAi) animals. The enhancement was synergistic, exceeding an expected additive-only effect in brood size reduction by 3-4 fold. We also tested two alleles of vab-2, which encodes an ephrin ligand for the Eph receptor (Chin-Sang et al., 1999). Similar to vab-1, vab-2 mutations also synergistically enhanced the nhr-6 RNAi reproductive phenotype (Table 1). There was no statistically significant difference between the enhancement phenotypes among all vab-1 and vab-2 experiments. Enhancement was also similar for vab-1; vab-2 double mutants, indicating that vab-1 and vab-2 are not acting in parallel (see below). The results also show that the function of vab-1 in reproduction requires both kinase and extracellular domain functions.

Table 1. nhr-6 RNAi Reproductive Interactions.

Genotype	Avg. # Eggs Laid (N)		% Severe Reproductive Phenotype3
	Control RNAi1	nhr-6 RNAi2	Control RNAi	nhr-6 RNAi
WT	276.1 ± 45.5 (18)	99.4 ± 46 (85)	0	5.9
vab-1(dx31)	149.1 ± 47.3 (35)	13.5 ± 17.7 (49)	2.8	79.6
vab-1(e2027)	159.9 ± 70.8 (30)	13.2 ± 27.8 (59)	3.3	84.7
vab-1(e2)	201.6 ± 51.9 (30)	26.6 ± 28.2 (40)	0	50
vab-1 (e699)	164.3 ± 69.1 (40)	11.3 ± 13.6 (34)	0	85.3
vab-2 (ju1)	158.5 ± 63.6 (39)	18.1 ± 19.9 (68)	2.6	69.1
vab-2 (e96)	180.5 ± 80 (30)	27.4 ± 22 (56)	3.3	48.2
vab-1 (e2027); vab-2 (e96)	67.3 ± 59.4 (20)	14.6 ± 23.2 (58)	20	74.1
efn-2 (ev658); efn-3 (ev696)	91.8 ± 27 (29)	59.3 ± 17.9 (40)	0	0
efn-4(bx80) efn-2(ev658); efn-3 (ev696)	118.5 ± 30 (28)	88.9 ± 27.9 (30)	0	0

¹GFP was utilized for the control RNAi experiments

²nhr-6 RNAi data for vab-1 and vab-2 were statistically similar to each other and statistically different from nhr-6 RNAi in WT and efn mutant backgrounds (one-way ANOVA with Tukey-Kramer post-hoc analysis at p =.05)

³Severe Reproductive Phenotype defined as animal laying <20 eggs

vab-1 and vab-2 mutant brood sizes amongst the control and nhr-6 RNAi animals were variable, ranging from sterility to occasional brood sizes exceeding 50 in the RNAi experiments. However, percentages of animals exhibiting severe brood size reduction (< 20 eggs laid) was substantially increased in the nhr-6 RNAi experiments (Table 1). Additionally, RNAi animals laid high numbers of morphologically abnormal eggs (data not shown). Three other genes in C. elegans encode ephrin proteins: efn-2, efn-3, and efn-4 (Wang et al., 1999; Chin-Sang et al., 2002). Strains bearing mutations in these genes did not exhibit strong enhancement, indicating that the interaction is specific to the vab-2 encoded ephrin.

Since vab-1 and vab-2 function in multiple developmental processes that could impact reproduction, we assessed spermatheca development in the nhr-6 RNAi animals. nhr-6 RNAi in vab-1 and vab-2 mutants generated spermathecae with significantly reduced cell numbers (Fig. 6A). To assess spermatheca morphology, we constructed vab-1(dx31) and vab-2(ju1) strains bearing transgenes encoding the epithelial adherens junction marker AJM-1∷GFP (Mohler et al., 1998). Expression of AJM-1∷GFP in vab-1;nhr-6(RNAi) and vab-2;nhr-6(RNAi) animals revealed that spermatheca morphology is severely affected in the RNAi animals (Fig. 6B-G). The phenotypes are similar to what is observed in nhr-6 null mutants (Gissendanner et al., 2008). There is pronounced loss of distal morphology in the RNAi animals and the length of the spermatheca is dramatically reduced. Spermatheca lengths in the vab-1(dx31); nhr-6(RNAi) animals averaged 26.2 ± 7.4 μm (n= 21) compared to 34.2 ± 6.4 μm (n=20) for nhr-6 RNAi alone and 41.0 ± 7.9 μm (n=20) for vab-1(dx31) alone with control animal spermatheca lengths averaging 51.7 ± 4.9 μm (n=11). Morphology of the spermatheca and SP-UT valve in vab-1 and vab-2 control RNAi animals were only moderately affected (data not shown). Spermathecae in vab-1; vab-2 double mutant animals are morphologically similar to the single loss of function animals (cell number = 23.9 ± 0.5, n=15) with a normal proximal-distal morphology (not shown) indicating that vab-1 and vab-2 do not function in parallel during spermatheca development.

Figure 6. Enhancement of nhr-6 RNAi phenotypes in vab-1 and vab-2 mutantsfunction.

(A) Graph showing average number of spermatheca nuclei in nhr-6 and control RNAi animals. N=15 for the control RNAi experiments and N=20 for the nhr-6 RNAi experiments. Cell number for nhr-6 RNAi in vab-1(dx31) and vab-2 (ju1) mutants was statistically different than nhr-6 RNAi in WT animals (*) by one-way ANOVA with Tukey-Kramer post-hoc analysis at p =.05. Data for vab-1(dx31) and vab-2 (ju1) were statistically similar. (B-G) Epifluorescence micrographs of spermatheca AJM-1∷GFP expression in control (B, D, F) and nhr-6 RNAi (C, E, G) animals for WT (B, C), vab-1(dx31) (D, E), and vab-2(ju1) (F, G). Proximal end of the spermatheca (P) is shown left distal end (D) is show right and asterisks indicate SP-UT valve location in all micrographs. AJM-1∷GFP is not expressed in the SP-UT valve. Bar = 10 μm. (H, H′) Expression of VAB-1∷GFP in the mid L4 spermatheca and corresponding DIC image (H′). Arrow indicates proximal end of spermatheca. Arrowhead indicates VAB-1∷GFP expression in the developing vulva. (I) Image of vab-2 mRNA localization from the NEXTDB website (http://nematode.lab.nig.ac.jp/; with permission from Dr. Yuji Kohara). Arrows indicate vab-2 localization in the distal somatic gonad region of an early L4 animal.

Previously, we had showed that nhr-6 interacted similarly with jun-1, which encodes the C. elegans homolog of the c-Jun transcription factor (Gissendanner et al., 2014). Since Eph-ephrin signaling is known to function through the c-Jun N-terminal kinase in some contexts (Stein et al., 1998; Becker et al., 2000; Xu et al., 2003; Yue et al., 2008; Arthur et al., 2011; Nguyen et al., 2013; Cho et al., 2014; Krupke and Burke, 2014), we sought to determine the relationship between vab-1/vab-2 and jun-1 by constructing vab-1(dx31) jun-1(gk551) and vab-2 (ju1); jun-1 (gk551) double loss-of-function mutants. Reproduction and growth was severely impacted in these animals. We were unable to establish a vab-1; jun-1 double mutant strain due to a growth arrest and vab-2; jun-1 double mutants exhibited an average brood size of 31.5 ± 24.8 (n=10). Interestingly, the spermathecae in these animals were similar to vab-2 single mutants in cell number (23.7 ± 0.7 spermatheca cells; N=15) and morphology. nhr-6 RNAi enhancement of the vab-2; jun-1 double mutant spermatheca cell number phenotype was also similar to vab-2 single mutants (15.4 + 2.6 spermatheca cells; N=20). This indicates that Eph-ephrin signaling and JUN-1 have common functions in spermatheca development.

We analyzed existing transgenic strains that express full-length VAB-1∷GFP and VAB-2∷GFP fusions (Chin-Sang et al., 1999). VAB-1∷GFP is expressed in the mid L4 spermatheca with accumulations highest in the apical regions of the organ (Fig. 6H). In early L4, VAB-1∷GFP expression is extremely faint and observed only in the most proximal spermatheca cells, while no expression was seen in the SPC daughter cells. We did not observe any VAB-2∷GFP expression at any stage of spermatheca development (although we do observe expression in other tissues). However, in situ hybridization data at the Nematode Expression Pattern Database show a small strip of vab-2 expression in an area consistent with the distal somatic gonad of an early L4 stage animal (http://nematode.lab.nig.ac.jp/db2/ShowCloneInfo.php?clone=338g11) (Fig. 6I).

These results demonstrate that NHR-6 and Eph receptor signaling function together during spermatheca development and that the severe brood size defects in vab-1;nhr-6(RNAi) and vab-2;nhr-6(RNAi) animals are likely due to defective spermatheca morphology and function.

Discussion

Regulation of cell cycle progression by nhr-6

Here, we have demonstrated a role for nhr-6 in cell cycle progression during spermatheca development. We find that in most animals, the cell division defects occur in the SPC lineage. This is not surprising since this lineage gives rise to most of the spermathecal cells. The initial divisions in this lineage occur normally with division failures being first observed in SPC daughter or granddaughter cells. The failure to divide could be due to a block, or a delay, in cell cycle progression, the latter resulting in decreased numbers of cells being born prior to onset of terminal differentiation. The variable cell number phenotypes in the nhr-6 mutants would be consistent with delayed progression. However, we find that the variability in cell numbers is most likely due to at least three distinct cell lineage defects resulting in division failures that would produce between 10-14 spermathecal cells depending on the lineage defect and whether the same lineage defect occurs with both SPCs (in our lineage observations both SPCs followed had the same lineage defect although this is clearly not the case in every instance as total cell numbers between 10 and 14 have been observed in young adult nhr-6 mutants). This explanation for the cell number phenotype is further supported by the observation that the cells that do not divide do not express rnr∷GFP at any point prior to the onset of terminal differentiation, indicating that there is no delay in cell cycle progression prior to cell cycle exit at mid L4. A potential explanation for the absence of rnr∷GFP expression in nhr-6 mutants is that in early L4 the SPC daughter cells prematurely exit the cell cycle. However, no cki-1∷GFP expression is seen in early L4 to early-mid L4 in wild type or mutant animals, while strong cki-1∷GFP expression is seen in both wild-type and mutant spermathecae in mid L4, indicating that terminal differentiation occurs at the normal time in mutant animals. We also note that the SPC division is asymmetric in nhr-6 mutants giving rise to a smaller proximal daughter cell and a larger distal daughter cell. This indicates that there is no lineage transformation that could potentially result in subsequent cell division failure. We favor the model that expression of NHR-6 in the spermatheca lineage is necessary to drive the later divisions in that lineage at a time when those divisions are occurring relatively rapidly.

We also showed that the loss of two negative G1/S transition regulators, lin-35 and fzr-1, suppressed the cell number phenotype of nhr-6 mutant animals. The C. elegans pRB homolog, LIN-35, is thought to have a G1/S function similar to its mammalian counterpart as it has been shown that loss of lin-35 suppresses the cell cycle phenotypes seen in cyclin D (cyd-1) and CDK4/6 (cdk-4) mutants (Boxem and van den Heuvel, 2001). The C. elegans Cdh1/Hct/FZR homolog, fzr-1, also functions in a manner similar to its mammalian counterpart. Cdh1 functions as a substrate recognition protein for the APC during late mitosis through the onset of S-phase (Huang et al., 2001). Cdh1 is required for exit from mitosis into G1 where it aides in the APC driven degradation of G1 cyclins. Phosphorylation of Cdh1 by cyclin dependent kinases causes its release from the APC allowing further accumulation of cyclin-CDK complexes required for G1/S progression (Zachariae et al., 1998; Kramer et al., 2000). In C. elegans it has been shown that fzr-1 acts in parallel with lin-35 by regulating G1 cyclins and inhibiting G1/S progression (Fay et al., 2002). Our data showed strong suppression of the nhr-6 mutant cell number phenotype using two separate alleles of fzr-1. The suppression by fzr-1 loss of function suggests that nhr-6 promotes G1/S cell cycle progression in the spermatheca lineage. The cell cycle-blocked spermatheca cells in nhr-6 mutants are likely arrested in G1 since they fail to express rnr∷GFP. The suppression by lin-35 loss of function also supports a G1/S promoting function for nhr-6. The suppression observed is not as robust as seen with fzr-1 loss of function. This could be a technical issue since we relied on an RNAi approach to remove nhr-6 function in a lin-35 mutant background. However, it is also possible that NHR-6 functions in parallel or downstream of lin-35. Based on the functions of lin-35 and fzr-1, one possible mechanism for the nhr-6 cell cycle regulation would be the positive regulation of G1 cyclin/CDK levels or activity. Cyclin/CDK activity may exist below a threshold required for G1/S progress during the rapid division phase in early L4 in nhr-6 mutants. Loss of fzr-1 could increase the activity of cyclin/CDKs above this threshold, accounting for the suppression that is observed in the double mutants.

The role of nhr-6 in cell differentiation

Our data further support the likelihood that nhr-6 has an additional, distinct role in cell differentiation during spermatheca development. In a previous study (and further confirmed in this study), it was observed that nhr-6 RNAi was hypomorphic in a wild-type background (Gissendanner et al., 2014). These animals exhibited defective spermatheca function due to loss of distal morphology; a phenotype that was observed even when the spermathecae in these animals contained the normal number of cells. This is also observed in a different context in the current study by the cell number suppression caused by fzr-1 loss of function. Spermathecae in nhr-6; fzr-1 double mutants containing a normal number of cells are severely morphologically abnormal and the animals exhibit brood sizes reduced to the degree seen in nhr-6 single mutants. The spermatheca is a functionally complex organ with two distinct cell morphologies (distal and proximal) and nhr-6 is required for the proper formation of both types. Additionally, nhr-6 is also required for differentiation of the SP-UT valve by promoting the proper differentiation of the sujc component of the SP-UT valve (Gissendanner et al., 2008). The hypomorphic RNAi and fzr-1 loss-of-function suppression data implicate NHR-6 in directing distinct cell differentiation events independent of its activity in cell cycle progression. This is similar to the various activities of mammalian NR4A NRs that are dependent on cell type and context. Therefore, the spermatheca represents a powerful in vivo model system to dissect the regulation of distinct cellular activities of NR4A NRs.

Interactions with Eph receptor signaling

We also report that NHR-6 functionally interacts with Eph receptor signaling during spermatheca development. This type of interaction has not previously been observed for NR4A NRs in any experimental system. Our data demonstrate that NHR-6 and Eph receptor signaling promote cell proliferation during spermatheca development. Since the nhr-6 RNAi approach utilized in the interaction experiments is hypomorphic, the data are consistent with NHR-6 acting with Eph receptor signaling in either a common pathway or in parallel. The former would be significant, as it would provide the possibility that Eph receptor signaling may be an upstream regulator of NR4A NR activity during morphogenesis. However, our data demonstrate that Eph receptor signaling is not critically essential for spermatheca development. Thus, if Eph receptor signaling does function upstream of NHR-6, it is likely one of several signaling pathways intersecting with NHR-6. This is supported by previous interaction studies that connected NHR-6 with Jun transcription factor activity and CDC42 signaling (Gissendanner et al., 2014). jun-1 loss-of-function seemed to dramatically enhance vab-1 and vab-2 loss of function animals with respect to growth and reproduction with the former being synthetically lethal with jun-1 loss-of-function. However, we were able to show a lack of enhancement with respect to spermatheca development, indicating that VAB-1/VAB-2 signaling and JUN-1 may function in a common pathway during spermatheca development, even though they appear to have parallel functions in other processes. Further investigations that link these interactions to the functions of NHR-6 during cell proliferation and cell differentiation will yield new insights into the mechanisms of NR4A NR regulation of organ formation.

Experimental Procedures

C. elegans strains

Strains were maintained and manipulated under standard conditions (Brenner, 1974). The following strains were obtained from the Caenorhabditis Genetics Center for use in this study: N2 (Bristol); MH1829 (fzr-1(ku398) unc-4(e120)); RB622 (fzr-1(ok380)); VT765 (unc-36(251);maIs103[rnr∷GFP + unc-36(+)]); MT10430 (lin-35(n745)); VT825(dpy-20(e1282)IV;maIs113[cki-1∷GFP + dpy-20(+)]); GR1379 (lin-35(n745)I; eri-1(mg366)IV); CZ337 (vab-1(dx31)), CZ4111 (vab-2(ju1)); CZ910 (vab-1(e2027)); CB96 (vab-2(e96)); CZ414 (vab-1(e699)); CB2 (vab-1(e2)); NW1549 (efn-2(ev658); efn-3(ev696)); CZ2274(efn-4(bx80) efn-2(ev658); efn-3(ev696)); (CZ996 (juIs52 [vab-2∷GFP + rol-6]); CZ793 (juIs24[vab-1∷GFP + lin-15(+)]); SU93 (jcIs1 [ajm-1∷GFP + unc-29(+) + rol-6(su1006)]; and PS3729 (syIs78 [ajm-1∷GFP + unc-119(+)]. DN73 (jun-1(gk551)) was previously generated by the Gissendanner lab (Gissendanner et al., 2014) and CZ918 (vab-1(e2027); vab-2(e96)) was kindly provided by Dr. Andrew Chisholm, University of California-San Diego.

The nhr-6(lg6001); fzr-1(ku298) unc-4(e120) and nhr-6(lg6001); fzr-1(ok380) double mutants were constructed by crossing DN20 (nhr-6(lg6001)) heterozygous males to MH1829 (fzr-1(ku298) unc-4(e120)) and RB622 (fzr-1(ok380)) hermaphrodites. F1s heterozygous for the nhr-6(lg6001) deletion was confirmed by single worm PCR. For nhr-6(lg6001); fzr-1(ku298) unc-4(e120), Unc F2s were picked to single plates and screened for low brood sizes (indicative of nhr-6(lg6001) homozygosity which was confirmed by single worm PCR). For nhr-6(lg6001); fzr-1(ok380), 100-125 F2 animals were collected on individual plates and screened for low brood sizes. Single worm PCR was used to confirm nhr-6(lg6001) and fzr-1(ok380) homozygotes. The nhr-6(lg6001);cki-1∷GFP and nhr-6(lg6001);rnr∷GFP strains were constructed in a similar manner with homozgosity of the integrated transgene confirmed by microscopic analysis of GFP expression.

To examine AJM-1∷GFP expression in vab-1(dx31) and vab-2(ju1) mutants, vab-1(dx31); juIs1 and vab-2 (ju1); syIs78 strains were constructed. syIs78 was utilized for vab-2 since juIs1 is integrated on chromosome IV where vab-2 is located.

vab-1 (dx31) jun-1(gk551) and vab-2 (ju1); jun-1 (gk551) double mutants were constructed by crossing vab-1/+ and vab-2/+ heterozygous males to DN73 (jun-1(gk551) hermaphrodites. F2 vab-1 or vab-2 homozygotes were selecting by picking weakly or moderately Vab animals that were viable. The jun-1 (gk551) allele in these animals was genotyped using single worm PCR with primers flanking the gk551 deletion. vab-1 and jun-1 are linked on chromosome II. F2 vab-1(dx31) jun-1 (gk551)/vab-1(dx31) jun-1(+) recombinants were recovered but the vab-1(dx31) jun-1(gk551) double homozygotes segregated from these animals were larval lethal. jun-1(gk551) homozgosity in the arrested larvae was confirmed by single worm PCR.

Spermatheca lineage analysis and nuclei counts

For the observations of spermatheca development in wild-type animals and nhr-6(lg6001) homozygotes, >50 animals for each strain were examined at distinct developmental time points during the L3 and L4 stages to generate the nuclei maps shown in Fig. 1 and 2. For observations at late L3/L3 lethargus and subsequent developmental time points, alkaline hypochlorite preparations of synchronized larvae were cultured at 20°C and cohorts of animals at the L3 lethargus (31-34 hrs), early L4 (35-37 hrs.), and early/mid-L4 (37-39 hrs) and mid-4 (39-42 hrs.) were placed on 2% agarose pads and anesthetized for analysis. All animals were anesthetized with 10 mM sodium azide. It was previously shown that a fully functional and nuclear localized full-length NHR-6∷GFP is expressed robustly in the spermatheca lineage from late L3 until late L4 and weakly in developing uterine and sheath cells (Heard et al., 2010). We integrated this transgene and utilized expression of the reporter gene at multiple developmental stages in conjunction with published cell lineages (Kimble and Hirsh, 1979) and cell maps of the somatic gonad (Newman et al., 1996; Lints and Hall, 2009) to facilitate identification of spermatheca, uterine, and sheath precursor cells.

Lineage analysis for nhr-6(lg6001) was performed using single animals beginning at late L3/L3 lethargus. Each animal was anesthetized with a single treatment of 0.01% tetramisole, observed, and then placed back on an agar plate. Each animal was observed every 45-60 minutes. We could identify both SPCs in the anterior spermatheca in different focal planes (see Fig. 1) and were able to track the fate of each SPC through two rounds of cell division. After the second round of division we could not detect any newly generated cells and the analysis was stopped at the mid L4 stage.

For the rnr∷GFP and cki-1∷GFP experiments, observations were made at the late L3, early L4, early/mid L4, and mid-L4 stages for the former; and early L4, early/mid L4, and mid-L4 for the latter. Expression patterns observed at these stages were the same for all SPC or SPC daughter cells in the animals examined.

All spermatheca nuclei counts for Figs. 5 and 6 were performed on young adult animals. Gonadal sheath pair 5 and the spermatheca-uterine valve were used as a distal and proximal spermatheca boundary, respectively. All nuclei within this boundary were scored.

RNAi experiments

All RNAi experiments were performed by post-embryonic feeding using an nhr-6 RNAi plasmid-containing HT115 bacterial strain from the Open Biosystems RNAi library. Control RNAi experiments utilized HT115 bacteria transformed with a GFP RNAi plasmid (pPD128.110 (L4417, Addgene)) or a her-1 RNAi plasmid-containing HT115 bacterial strain from the Open Biosystems RNAi library (for the data shown in Fig. 6). Bacteria were freshly streaked out on to ampicillin (100 μg/ml)/tetracycline (12.5 μg/ml) plates and colonies were cultured in 2 ml of LB in a 15 ml conical tube for 6.5 hours at 37° C with shaking. 75 μl of bacteria were spread onto NGM Lite plates containing 200 mg/ml ampicillin and 1mM IPTG and induced overnight at 20° C. Synchronized L1 larvae from alkaline hypochlorite preparations were added to the freshly induced bacterial lawns and animals were allowed to grow until young adult.

Quantification of brood sizes

Single young adults with no visible eggs in the uterus were transferred to brood count plates that contained a thin narrow lawn of OP50 bacteria. Animals were moved to a new brood count plate every 8-12 hours a total of four times. After worms were moved eggs and larvae were counted. Eggs were initially scored as abnormal when they displayed a small, roundish shape, as previously described (Gissendanner et al., 2008). Plates displaying abnormal eggs were kept 24 hours after the initial analysis to verify they did not hatch, at which point they were scored as abnormal.

vab-1 and vab-2 interaction

For brood size and microscopic phenotype analysis in the nhr-6 RNAi experiments, we only selected healthy, active young adults that lacked a notched head or were only mildly notched. Any animals that appeared unhealthy after the start of the experiment were discarded from the data set. RNAi experiments were replicated at least twice for each mutant strain along with wild-type controls. The data was collected over multiple experiments with each experiment including a wild-type control. Wild-type data was highly similar across experiments and the data from all experiments were combined for Table 1.