CACN-1 is required in the Caenorhabditis elegans somatic gonad for proper oocyte development

Abstract

CACN-1/Cactin is a conserved protein identified in a genome-wide screen for genes that regulate distal tip cell migration in the nematode Caenorhabditis elegans. In addition to possessing distal tip cells that migrate past their correct stopping point, animals depleted of cacn-1 are sterile. In this study, we show that CACN-1 is needed in the soma for proper germ line development and maturation. When CACN-1 is depleted, sheath cells are absent and/or abnormal. When sheath cells are absent, hermaphrodites produce sperm, but do not switch appropriately to oocyte production. When sheath cells are abnormal, some oocytes develop but are not successfully ovulated and undergo endomitotic reduplication (Emo). Our previous proteomic studies show that CACN-1 interacts with a network of splicing factors. Here, these interactors were screened using RNAi. Depletion of many of these factors led to missing or abnormal sheath cells and germ line defects, particularly absent and/or Emo oocytes. These results suggest CACN-1 is part of a protein network that influences somatic gonad development and function through alternative splicing or post-transcriptional gene regulation.

1. Introduction

Cell-cell interactions and signaling are crucial for proper development and differentiation. Germ line development, differentiation, and maturation across species is regulated through soma-germ line interactions (Lehmann, 2012; Hanna and Hennebold, 2014; Killian and Hubbard, 2005; McCarter et al., 1997; Eppig, 1991). Somatic cells provide a niche for proper germ line differentiation and development (Jemc, 2011). Defects in development of the soma and disruption of somatically expressed genes can cause reproductive defects including ovarian failure, adeno-carcinomas, granulosa tumors and infertility (Matzuk and Lamb, 2008; Singh and Schimenti, 2015; Pangas et al., 2008; Wu et al., 2007). However, the molecular mechanisms that control somatic gonadal cell specification and development are not well understood.

The reproductive system of the nematode Caenorhabditis elegans is an ideal model for the study of somatic gonad development and soma-germ line interactions. The C. elegans reproductive system consists of two symmetrical U-shaped gonad arms connected to a shared uterus. Each somatic gonad arm is enveloped by an outer basal lamina and consists of the distal tip cell (DTC), 5 pairs of gonadal sheath cells, the spermatheca, and the spermatheca-uterine valve (sp-ut) (McCarter et al., 1997; Hubbard and Greenstein, 2000; Strome, 1986). The distal mitotic germ cell pool is the source of meiotic cells, which move proximally as they undergo gametogenesis. Sperm are produced in the last stage of larval development and stored in the spermatheca. The hermaphrodites then undergo the sperm-ooycte switch and produce oocytes throughout adulthood. Oocytes progress through meiosis as they move proximally, arresting at diakinesis of meiosis I in the proximal arm. Upon receiving cues from sperm and sheath cells, the most proximal oocyte adjacent to the spermatheca (the −1 oocyte) matures, is ovulated into the spermatheca, fertilized, and finally expelled into the uterus (Hubbard and Greenstein, 2005).

Germ cells and somatic gonadal cells arise from Z2/Z3 and Z1/Z4 precursor cells respectively (Lints and Hall, 2013). These cells and their daughters are not spatially segregated until the L2/L3 molt when the gonad begins to establish organization and structure (Kimble and Hirsh, 1979). During early L3, guided by the distal tip cells (DTC), the two gonad arms begin to elongate away from the midpoint of the animal. Germ cells proliferate throughout the arm. Midway through L3 the most proximal germ cells begin to enter meiosis (Killian and Hubbard, 2005; Hansen et al., 2004; Kimble and White, 1981). The epithelial sheath cells that surround the gonad play key roles in germ line patterning during development (Killian and Hubbard, 2005).

Sheath cells originate as five lateral pairs adjacent to the proximal germ line. As the animal develops, sheath cells migrate and expand, forming a single layer of cells covering most of the gonad arm (Hall et al., 1999; Hirsh et al., 1976). The distal sheath cells (Sh1-2) promote germ line proliferation and exit from meiotic pachytene (Killian and Hubbard, 2005; McCarter et al., 1997), whereas the 3 pairs of proximal sheath cells (Sh 3-5) are important for oocyte maturation and for the smooth-muscle-like contractions that push oocytes into the spermatheca for fertilization (Strome, 1986; Miller et al., 2003, 2001; McCarter et al., 1999). Defects in sheath cells, including mutations in sheath cell expressed genes like ceh-18, a POU class homeoprotein (Rose et al., 1997), pro-1, a protein involved in rRNA processing (Voutev et al., 2006; Killian and Hubbard, 2004), and mup-2, a troponin homolog (Myers et al., 1996), or ablation of sheath cell precursor cells (Sh15) (Killian and Hubbard, 2005; McCarter et al., 1997) produce an array of germ line phenotypes including tumors (Killian and Hubbard, 2005; McCarter et al., 1997; Killian and Hubbard, 2004), sperm filled proximal gonad arms that lack oocytes (Killian and Hubbard, 2005) and endo-mitotically duplicating oocytes (Killian and Hubbard, 2005; McCarter et al., 1997; Rose et al., 1997; Myers et al., 1996) (the Emo phenotype (Iwasaki et al., 1996)). Even though it is clear that sheath cell development has a crucial role in proper germ line formation, little is known about the molecular mechanisms that control sheath cell specification, development and patterning in C. elegans. For example, the protein LIN-9 functions in an RB-related pathway, and is necessary for the development of the correct number of somatic sheath cells, however the mechanism of this regulation remains unknown (Beitel et al., 2000).

We identified CACN-1, a protein of unknown function, conserved from yeast to humans, in a screen for regulators of DTC migration (Cram et al., 2006; Tannoury et al., 2010). CACN-1 is the homolog of Cactin, which functions with the Rel/NFKB pathway to control dorsal-ventral patterning in Drosophila (Lin et al., 2000), in innate immunity in Litopenaeus vannamei (Zhang et al., 2014), and with TRIM39 to negatively regulate the NF_KB pathway in human cell lines (Suzuki et al., 2015). CACN-1 lacks enzymatic activity but contains a nuclear localization signal at its N-terminus, two coiled-coiled domains and a conserved C-terminus (Schultz et al., 2000; Kosugi et al., 2009). Recently, a functional role for CACN-1 has begun to emerge. Our proteomics work suggests CACN-1 is part of a novel network containing many spliceosomal components (Doherty et al., 2014). This result is consistent with the observation that both human and Arabidopsis thaliania (Baldwin et al., 2013) CACTIN proteins co-purify with spliceosomal proteins (Jurica et al., 2002; Bessonov et al., 2008; Ilagan et al., 2009; Ashton-Beaucage et al., 2014), and that in Schizosaccharomyces pombe, Cay1/cactin promotes proper splicing and protein stability of the telomeric protein Rap1 (Lorenzi et al., 2015).

The spliceosome is responsible for the removal of introns and ligation of exons (Zahler, 2012) and for alternative splicing, which allows for variety in mRNA products and is a mechanism to regulate gene expression (Wollerton et al., 2001). The spliceosome consists of over 100 different factors including small ribonuclear-protein particles (snRNPs), accessory proteins and an array of RNA binding proteins (RBPs) (Zahler, 2012; Chen and Cheng, 2012). In the C. elegans genome 2562 annotated genes, or 13% of the genome, are alternatively spliced (Zahler, 2005). In addition, regulation of gene expression in the C. elegans germ line is primarily translational, most commonly through binding of RBPs to the 3′ UTR of target messages (Merritt et al., 2008). Even though splicing and post-transcriptional gene regulation play an important role in C. elegans biology, many C. elegans splicing factors and RBPs remain uncharacterized.

CACN-1 has been previously shown to cause germ line over-proliferation and issues in germ cell differentiation when depleted solely in the germ line of C. elegans (Kerins et al., 2010). However, CACN-1's role in the somatic gonad and the mechanism by which it exerts these germ line effects remains largely unknown. In this study, we demonstrate that CACN-1 is necessary primarily in the soma for the presence and proper morphology of the gonadal sheath cells that regulate the sperm-oocyte switch as well as germ line maturation and ovulation in C. elegans. Depletion of a set of previously identified CACN-1 interacting proteins (Doherty et al., 2014) similarly results in sheath cell and germ line defects. The results of this study reveal that CACN-1 and its network of interacting proteins are important for proper development of the gonad and suggest that CACN-1 may be involved in pre-mRNA splicing or post-transcriptional regulation to mediate these decisions.

2. Materials and methods

2.1. Nematode strains

Nematodes were grown on nematode growth media (NGM) (0.107 M NaCl, 0.25% wt/vol Peptone (Fischer Science Education), 1.7% wt/vol BD Bacto-Agar (Fisher Scientific), 0.5% Nyastatin (Sigma), 0.1 mM CaCl₂, 0.1 mM MgSO_4, 0.5% wt/vol Cholesterol, 2.5 mM KPO₄), seeded with Escherichia coli OP50 using standard techniques (Myers et al., 1996). Nematodes were cultured at 23°C unless specified otherwise. The strains used in this study are as follows: N2 (wild type reference strain from Bristol), NL2550 ppw-1(pk2505) and NL2098 rrf-1(pk1417) and the GFP expression lines DG1575 lim-7::GFP and OD27 AIR-2::GFP.

2.2. RNA interference

Starved dauer nematodes were allowed to recover for 48 h on NGM plates newly seeded with OP50. This procedure produces young gravid adults for egg collection. Eggs were released using an alkaline hypochlorite solution as described in Hope (1999), and washed 3 × with filter sterilized M9 buffer (22 mM KH₂PO₄, 42 mM NaHPO₄, 86 mM NaCl, and 1 mM MgSO₄) (‘egg prep’). Clean eggs were then transferred to NGM previously seeded with HT115(DE3) bacteria that express dsRNA for RNAi. Strains utilized in each RNAi experiment are indicated.

The RNAi feeding protocol was performed essentially as described in Timmons et al. (2001). HT115(DE3) bacteria transformed with the dsRNA construct of interest were grown over night at 37°C in Luria broth (LB) supplemented with 40 μg/ml ampicillin. The following day, 150 μl of the culture was seeded on NGM agar supplemented with 25 μg/ml carbenicillin and isopropylthio-β-galactoside (IPTG) and incubated at room temperature for 24–72 h to induce dsRNA expression. Eggs collected from the alkaline lysis procedure were transferred onto these plates and incubated at 23°C for the times specified.

The cacn-1 ORF RNAi clone is a full-length cDNA matching Wormbase (WS2000) predictions (Open biosystems; Huntsville, AL, USA). All CACN-1 interactor RNAi clones are described in Doherty et al. (2014). Empty pPD129.36 vector (“empty RNAi”) was used as a negative control in all RNAi experiments.

2.3. RNA in situ hybridization

RNA in situ hybridization was performed in accordance with Lee and Schedl (2006). Nematodes subjected to both cacn-1 and control empty RNAi were collected at adulthood (54 h post ‘egg prep’). Gonad arms were extracted in 1 × Phosphate Buffered Saline (PBS) and fixed in 3% paraformaldehyde/0.25% glutaraldehyde/0.1 M K₂HPO₄ (pH 7.2). Samples were washed with in 1 × Phosphate Buffered Saline and 0.1% Tween 20 (PBST) and stored in cold methanol overnight, then washed with PBST and digested with Proteinase K, 50 μg/ml in PBST, for 30 min. Gonad arms were then washed with PBST, fixed with 3% paraformalde-hyde/0.25% glutaraldehyde/0. 1 M K₂HPO₄ (pH 7.2) for 15 min, and washed with PBST containing 2 mg/ml glycine. Gonad arms were then hybridized in a 1:2 ratio of RNA cacn-1 DIG labeled probe to hybridization buffer (5 × sodium chloride sodium citrate solution, 50% formamide, 100 μg/ml salmon sperm DNA, 50 μg/ml heparin sodium salt and 0.1% Tween 20), in a 48°C water bath for 24 h.

Anti-cacn-1 probes were generated using 580 base pairs of the cacn-1 gene. PCR was run with DIG labeled nucleotides (Roche Diagnostics Corporation, Indianapolis, IN, USA). Sense and anti-sense probes were created using asymmetric one-way PCR. Probes were then diluted with sterile water and dehydrated with 1 M NaCl overnight at −20°C. The reaction was centrifuged and washed with ethanol. The remaining DNA probe was then re-suspended in hybridization buffer and stored at −20°C.

After gonad arms were hybridized with anti-cacn-1 DIG labeled probes they were washed several times with hybridization buffer and PBST both with and without 0.5 mg/ml BSA (New England Biolabs, Ipswich, MA, USA). Probe was detected using a diluted 1:1000 alkaline-phosphatase-conjugated anti-DIG (Roche Diagnostics Corporation, Indianapolis, IN, USA) in PBST with BSA overnight at 4°C. Gonad arms were then washed with PBST both with and without BSA as well as a staining solution containing 100 mM NaCl, 5 mM MgCl₂, 100 mM Tris (pH 9.5), 0.1% Tween 20, and 1 mM levamisole. Gonad arms were stained with staining solution containing 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NDT) dissolved tablet (Sigma Aldrich, St. Louis, MO, USA) and 100 ng/ml DAPI in the dark for approximately four hours at room temperature, washed with PBST, re-suspended in 100 ng/ml DAPI, and washed a final time in PBS. Excess liquid was removed and gonad arms were mounted onto slides with 2% agar pads and observed using the DIC filter of the fluorescent microscope for cacn-1 staining within the gonad arms of all the treated animals.

2.4. DAPI and Texas Red-X phalloidin staining

N2 nematodes were stained with 4′,6-diamidino-2-phenylindolea (DAPI), a nuclear stain, to observe oocytes and sperm within gonad arms. Young adults were collected with 1 × Phosphate Buffered Saline and 0.1% Tween 20 (PBST) into a 15 ml glass conical tube and allowed to settle by gravity. The animals were then washed three times with PBST to remove excess bacteria and fixed with cold methanol for 20 min. After fixation, animals were washed once more with PBST. The whole animal was then stained with 100 ng/ml of DAPI (Sigma-Aldrich, St. Louis, MO) in PBS for at least 20 min. After washing with PBS, excess liquid was removed and animals were mounted on slides coated with 2% agarose in H₂O. Gonad arm content and phenotypes were observed for nuclear staining using a Nikon 80i epifluorescence microscope equipped with a DAPI fluorescent filter.

To stain gonad arms with Texas Red-X phalloidin, an actin stain, animals were washed with PBS and transferred to a watch glass. Using a 25 gauge syringe (BD PrecisionGlide Needle (Becton Dickinson, Franklin Lakes, NJ, USA)), animals were sliced at the level of the pharynx or vulva, causing gonad arms to extrude from the body. The dissected gonad arms were fixed in 4% formaldehyde and then washed several times with PBST. After washing, gonad arms were stained with Texas Red-X phalloidin (Molecular Probes, Invitrogen, Life Technologies, Grand Island, NY, USA) at a final concentration of 0.4 units/ml, and placed in the dark at room temperature for 2 h or 4°C overnight. Gonad arms were than washed with PBST and stained with 100 ng/ml DAPI in PBS for 20 min, washed in PBS and left to settle. Extra liquid was removed and gonad arms were transferred onto slides coated with 2% agarose in H₂O for observation as described above.

2.5. Fluorescence microscopy and analysis

To analyze gonad arm morphology, partially synchronized populations of animals were mounted in a ~5 μl drop of M9 and 0.08 M sodium azide (or 0.01% tetramisole and 0.1% Tricaine) onto a slide with a 2% agarose in H₂O pad. Animals were observed using a 20 × or 60 × oil-immersion objective with a Nikon Eclipse 80i epiflourescence microscope equipped with a Spot RT3 CCD camera (Diagnostic instruments; Sterling Heights, MI, USA). Images were captured using Spot advanced version 4.6 software (Diagnostic Instruments, Sterling Heights, MI, USA). All Differential Interference Contrast (DIC) pictures were taken at 37.44 ms exposure. Fluorescent images were taken using either a GFP, DAPI or RFP filter as appropriate. Animals were scored for the presence and absence of sperm, oocytes and/or embryos in both the proximal and distal gonad arm of each animal. Cells were identified as sperm or oocytes based on nuclear staining and morphology. Animals stained with Texas Red-X phalloidin were scored for the presence and absence of sheath cells based on the visible actin filaments. Due to variability in the dissection and phalloidin staining procedure, as well as the variation in the orientation of the dissected gonads, images from this experiment were scored blindly for the presence of actin in the proximal gonad. For all experiments a Fishers exact t-test (two dimensional χ² analysis) using GraphPad Prism statistical software was used to compare the percent of normal gonad arms between empty and all other RNAi treatments.

3. Results

3.1. CACN-1 is necessary for the production of normal oocytes

Depletion of CACN-1 via feeding RNAi leads to sterility in wild type animals. Sterility can be caused by abnormal germ line development or maturation, as well as by defects in the somatic gonad (Hubbard and Greenstein, 2000). To investigate the cause of sterility, we used DIC microscopy and DAPI, a nuclear stain, to observe gonad arms and germ line development of wildtype and cacn-1 RNAi treated animals. Gonad arms in wildtype (N2) young adults at 65 h contained aligned oocytes in the proximal gonad, adjacent to the sperm-filled spermatheca, with developing embryos in the uterus (Fig. 1). Animals treated with cacn-1 RNAi exhibited proximal gonad arms that either contained sperm, but lacked oocytes (44%) or contained abnormal oocytes in addition to sperm (51%; n=208) (Fig. 1). Occasionally, we observed gonads comprised entirely of undifferentiated germ cells (1%; n=200). C. elegans germ cells, sperm and oocytes exhibit characteristic DNA morphologies (Hansen and Schedl, 2013; Greenstein, 2005). To visualize the DNA in these cells, gonad arms were stained with DAPI, a nuclear stain. N2 gonad arms contained normally stained germ line cells, oocytes, and sperm. Consistent with the DIC imaging, DAPI staining of CACN-1 depleted animals revealed proximal gonad arms that contained sperm, but lacked oocytes (43%), or for those that contained oocytes, large, dense areas of nuclear staining indicative of endomitotically duplicating oocytes (Emo) (Iwasaki et al., 1996) (57%; n=134) (Fig. 1D). In C. elegans, when ovulation fails, oocytes typically go through multiple rounds of nuclear envelope breakdown and replication becoming highly polyploidy (Iwasaki et al., 1996). Disruption of many different genes can result in the Emo phenotype, including genes involved in oocyte maturation and maintenance (Killian and Hubbard, 2004; Iwasaki et al., 1996), as well as somatic gonad structure and development (McCarter et al., 1997; Myers et al., 1996; Wissmann et al., 1999; Ono and Ono, 2004; Greenstein et al., 1994).

Fig. 1.

CACN-1 is necessary for the sperm to oocyte switch and proper oocyte development. (A) DIC images of wildtype animals reared on empty control RNAi (n=200) or cacn-1 RNAi (n=208) for 65 h, and scored for defects using differential interference contrast (DIC) microscopy. The stacked bar graph indicates the percentages of animals exhibiting the indicated phenotypes. The two populations are significantly different (Fisher's exact t-test, ****p<0.0001). (B) Animals were treated with cacn-1 RNAi for the indicated times and germ line phenotypes were scored as stated above. No increase in the percentage of animals with oocytes was observed over the timecourse. n=the number of gonad arms observed. (C) Representative DIC images of animals treated with empty and cacn-1 RNAi. Control gonads exhibit aligned oocytes adjacent to the spermatheca, which normally contains sperm. cacn-1 depletion resulted in proximal gonad arms that contained sperm, but lacked oocytes or contained abnormal oocytes in addition to sperm. Asterisks mark oocytes and triangles mark sperm-filled proximal gonads. (D) Representative images of DAPI stained RNAi treated animals. Gonad arms from control treated animals contained normal oocytes adjacent to the spermatheca. CACN-1 depletion resulted in sperm-filled proximal gonad arms or endomitotic oocytes (Emo). Asterisks mark normal oocytes, triangles mark sperm nuclei and arrows indicate Emo oocytes. Scale bar is 25 μm.

To determine whether animals that lacked oocytes were simply developmentally delayed in the sperm to oocyte switch, animals were reared on cacn-1 RNAi and observed over a time course throughout early adulthood (54–72 h). All timepoints had a similar penetrance of proximal gonad arms that contained sperm, but lacked oocytes (42%; n=740) with most of the remainder of the proximal gonads containing abnormal oocytes in addition to sperm (55%; n=740) (Fig. 1B). These data suggest that the absence of oocytes is not the result of a delay in oocyte production, but instead represents a failure to switch to oocyte production. Therefore, CACN-1 may regulate the sperm to oocyte switch.

3.2. cacn-1 is expressed in the somatic gonad

A transgenic construct, expressing CACN-1::GFP under the control of 913 bp of upstream sequence, is expressed in the pharynx, intestine, vulva and somatic gonad (Tannoury et al., 2010). A mini-gene construct, containing this sequence upstream of cacn-1 cDNA, is sufficient to rescue larval development and fertility in cacn-1(tm3042) mutant animals (Tannoury et al., 2010). However, expression of GFP was not observed in the germ line, likely due to silencing of the multi-copy extrachromosomal array. To observe the endogenous cacn-1 mRNA expression pattern in the gonad, animals were reared on cacn-1 and empty control RNAi for 54 h, dissected, and subjected to RNA in situ hybridization. As expected, cacn-1 treated animals had substantially smaller gonad arms than wild type (Tannoury et al., 2010). In control animals, cacn-1 mRNA was expressed from the distal gonad arm through to the proximal arm, the region where oocyte maturation occurs (Fig. 2A and D). No staining was observed when CACN-1 was depleted via RNAi, confirming not only the specificity of our CACN-1 probe, but also that feeding RNAi efficiently depletes CACN-1 expression (Fig. 2A and D). To determine if cacn-1 mRNA is expressed in the somatic gonadal sheath cells and/or the underlying germ line, RNA in situ hybridization was repeated on animals resistant to RNAi in the soma, NL2098 rrf-1(pk1417) (Sijen et al., 2001), or in the germ line NL2550 ppw-1(pk2505) (Grishok, 2005). Both rrf-1(pk1417) and ppw-1(pk2505) animals depleted of cacn-1 exhibited reduced staining compared to control animals (p<0.0001) (Fig. 2B, C, and D). These results suggest that cacn-1 mRNA is expressed in both the somatic gonadal sheath cells and the germ line.

Fig. 2.

CACN-1 mRNA is expressed in the soma and germ line of gonads. N2, rrf-1(pk1417) and ppw-1(pk2505) animals were reared on either empty or cacn-1 RNAi, dissected, and subjected to RNA in situ hybridization using a cacn-1 anti-sense probe. (A–C) The stacked bar graphs indicate the percentage of gonad arms exhibiting the indicated level of staining for N2 (empty; n=42, cacn-1; n=26), rrf-1(pk1417) (empty; n=48, cacn-1; n=155) and ppw-1(pk2505) (empty; n=90, cacn-1; n=61) respectively. The two populations were statistically different for all three worm strains subjected to RNA in situ hybridization (Fisher's exact's t-test, ****p<0.0001) (D) Control animals reared on empty RNAi exhibited cacn-1 staining from the distal gonad arm through to the proximal gonad where oocytes arrest adjacent to the spermatheca. Animals treated with cacn-1 RNAi lacked detectable staining in a significant percentage of gonad arms observed. Triangles mark the distal tip of the gonad and brackets outline the proximal gonad. Scale bar is 25 μm.

3.3. CACN-1 loss results in defective or missing sheath cells

Five pairs of single layer epithelial sheath cells cover the gonad arms and play key roles in germ line development, oocyte maturation, and ovulation (Killian and Hubbard, 2005; McCarter et al., 1997; Hubbard and Greenstein, 2000; Strome, 1986; McCarter et al., 1999). The high penetrance of Emo and abnormal bulging of the proximal gonad arm when CACN-1 is depleted suggested a possible defect in sheath cell development or morphology. To observe the presence or absence of sheath cells in cacn-1 treated animals, we used a sheath cell reporter strain, DG1575 LIM-7::GFP. LIM-7 is a HOX transcription factor expressed in the cytoplasm and nuclei of sheath cells (sh1-5) (Hall et al., 1999; Voutev et al., 2009). Control young adult animals exhibited visible GFP expression in 72% of gonad arms (n=53). In contrast, animals treated with cacn-1 RNAi exhibited GFP expression in only 10% of gonad arms (n=50). Of the proximal gonad arms with no GFP expression, 76% contained sperm, but lacked oocytes (Fig. 3). Similar results were observed at the L4 stage; control animals exhibited visible GFP in 100% of L4 gonad arms observed, while CACN-1 depletion resulted in only 28% of gonad arms expressing GFP (n=25, p<0.0001). These data suggest that CACN-1 is necessary for LIM-7::GFP expression during or before the L4 stage of larval development.

Fig. 3.

CACN-1 is necessary for normal sheath cells in adult animals. DG1575 LIM-7::GFP animals were treated with empty and cacn-1 RNAi. GFP expression was visualized using fluorescence microscopy and germ line morphology with DIC microscopy. (A) The stacked bar graph indicates the percentage of gonad arms exhibiting the indicated level of GFP expression (empty; n=56, cacn-1; n=52). The two populations are significantly different (Fisher's exact t-test, ****p<0.0001). (B) The stacked bar graph indicates the percentage of gonad arms that exhibit GFP expression across the observed germ line phenotypes for both empty and cacn-1 treated animals. (C) Control animals exhibited LIM-7::GFP expression in the nuclei of sheath cells 1–4. Depletion of cacn-1 resulted in faint GFP expression, as indicated by an arrow, or no GFP expression. Faint GFP expression correlated with presence of abnormal oocytes. Asterisks mark oocytes and triangles mark sperm in DIC images. Scale bar is 25 μm.

To confirm that the loss of LIM-7::GFP expression was due to the absence of sheath cells, rather than disruption of reporter gene expression by cacn-1 RNAi, dissected gonad arms of cacn-1 RNAi treated animals were stained with Texas Red-X phalloidin to visualize sheath cell actin, and with DAPI to reveal the DNA morphology of the germ cells, oocytes and sperm. Due to variability inherent in the staining procedure and orientation of the dissected gonads, images from this experiment were scored blindly. Sheath cell actin in control animals was scored as normal in 50% of gonad arms and mildly abnormal in 39% of gonad arms (n=54) (Fig. 4). All of these gonad arms contained oocytes arrested at diakinesis (Fig. 4B). In contrast, CACN-1 depleted animals had no actin sheath cell staining in 93% of gonad arms (n=27) (Fig. 4). Proximal gonad arms primarily contained sperm, but lacked oocytes (72%) (Fig. 4B). These results are significantly different from control data (p<0.0001), suggesting that when sheath cells are missing, oocytes do not develop. Although the absence of phalloidin staining could be explained by a complete disruption of actin polymerization, rather than an absence of sheath cells, taken together with the LIM-7::GFP data, it seems more likely that depletion of cacn-1 leads to a loss of the sheath cells themselves. This interpretation is supported by previous ablation studies of sh2-5, which resulted in gonad arms containing sperm and spermatocytes but lacking oocytes (Killian and Hubbard, 2005). Taken together, these observations support the conclusion that sheath cells are needed for proper germ line development.

Fig. 4.

CACN-1 is necessary for the proper organization of the actin cytoskeleton in somatic sheath cells. Wildtype animals reared on empty and cacn-1 RNAi were dissected and stained with Texas Red-X phalloidin and DAPI. Actin and nuclear staining were observed using fluorescent microscopy. Sheath cell staining was scored blindly. (A) The stacked bar graph indicates the percentage of gonad arms exhibiting the indicated phenotypes (empty; n=54, cacn-1; n=27). The two populations are significantly different (Fisher's exact t-test, ****p<0.0001). (B) Control animals exhibited gonad arms that contained normally arrested oocytes and both normal and mild defects in sheath cell staining. In contrast, cacn-1 depletion resulted most commonly in no sheath cell staining and sperm-filled proximal gonad arms or severe defects in sheath cell staining with Emo oocytes. White boxes outline the region scored for these phenotypes in the proximal gonad. Asterisks mark oocytes, arrows mark the Emo phenotype and triangles mark sperm nuclei. Scale bar is 25 μm.

Gonad arms that lack sheath cells, or have sheath cells with grossly abnormal actin cytoskeletal organization, should be unable to contract to facilitate successful ovulations. Animals were reared on control or cacn-1 RNAi and observed by time-lapse DIC microscopy to assess contraction rates prior to ovulation. Control animals exhibited major ovulatory contractions at a rate of 15.6 contractions/min (n=6). This is similar to previously reported sheath cell contraction rates prior to ovulation (Rose et al., 1997). The few animals depleted of CACN-1 that contained mature oocytes exhibited no observable sheath cell contractions (n=6) (Data not shown). Importantly, in those animals that did produce oocytes, the oocytes arrested at diakinesis during young adulthood. These oocytes did not enter the spermatheca and became Emo. Therefore, some oocytes differentiate properly when CACN-1 is depleted, but become abnormal and Emo over time due to the inability of sheath cells to contract and force them into the spermatheca.

In wild type animals, the −1 oocyte, adjacent to the spermatheca, is triggered to mature prior to ovulation. Oocyte maturation involves characteristic steps including the disappearance of the nucleolus, germinal vesicle breakdown (GVBD), and spindle rearrangement (McCarter et al., 1999; Ward and Carrel, 1979). These steps normally occur as sheath cell contractions increase to ensure proper timing of oocyte maturation and ovulation (McCarter et al., 1997; Miller et al., 2001; Ward and Carrel, 1979). Animals reared on control RNAi exhibited normal disappearance of the nucleolus of the −1 oocyte (98%; n=62) (Table 1). A small proportion of the −2 oocytes of control animals also lacked nucleoli (21%; n=62), however, this phenotype was most commonly observed when the −1 oocyte was in the process of being ovulated. In contrast, cacn-1 RNAi treated oocytes appeared to mature early. Nucleoli were absent at high percentages in the −1 (100%; n=46), −2 (89%; n=45) and the −3 oocyte (70%; n=40) (Table 1) Animals were even observed to lack nucleoli in the −6 oocyte (1%; n=15). This not surprising, as sheath cells are known to inhibit premature maturation of oocytes through VAB-1/ephrin signaling (Miller et al., 2003) and the POU-homeobox gene ceh-18 (Rose et al., 1997; Greenstein et al., 1994) in parallel with G-protein pathways and downstream production of cAMP (Govindan et al., 2009; Govindan et al., 2006; Kim et al., 2013). In addition gap junctions between sheath cells and oocytes are necessary to inhibit oocyte maturation (Whitten and Miller, 2007). If sheath cells are missing, or abnormal, this inhibition would be lost, and oocytes would precociously mature.

Table 1.

CACN-1 inhibits precocious oocyte maturation. N2 and OD27 AIR-2::GFP strains were treated with empty and cacn-1 RNAi, and scored for AIR-2::GFP localization and presence of oocyte nucleoli. The number of oocytes observed for each location is indicated in parentheses.

Treatment	% Oocytes without a visible nucleolus
	Oocyte position in the gonad arm
	−1	−2	−3	−4	−5	−6
empty	98%(62)	21%(62)	0%(62)	0%(49)	0%(21)	0%(10)
cacn-1	100%(46)	89%(45)	70%(40)	43%(35)	18%(28)	1%(15)
	% Oocytes with chromatin-localized AIR-2::GFP
empty	100%(23)	13%(23)	0%(23)	0%(16) 0%(6)		0%(2)
cacn-1	100%(6)	100%(6)	80%(5)	33.3%(3) 0%(1)		–

To further investigate the hypothesis that oocytes are maturing early in cacn-1 animals, we used the OD27 AIR-2::GFP strain. AIR-2 encodes an aurora/Ipl1-related serine/threonine protein kinase that localizes to meiotic chromatin in the presence of sperm and the MSP (Schumacher et al., 1998). AIR-2::GFP has been used previously as a marker for oocyte maturation (Govindan et al., 2009), and should be visible only in maturing oocytes. In control animals AIR-2::GFP was observed in the −1 oocyte(100%; n=23) and a few of the −2 oocytes (13%; n=23) (Table 1; Fig. 5). AIR-2:: GFP expression coincided with loss of the nucleolus in the −1 oocyte (98%; n=62) and −2 oocyte (21%; n=62) (Table 1; Fig. 5). AIR-2::GFP animals reared on cacn-1 RNAi grew very slowly and often arrested at L4 and early adulthood, making assessment of oocyte maturation difficult. However, in the few escapers, CACN-1 depleted animals exhibited AIR-2::GFP fluorescence in 100% of the −1 and −2 oocytes and in 80% of the −3 oocytes (n=6) (Table 1; Fig. 5). GFP was observed even in the −4 oocyte (33.3%; n=3). As expected, the oocytes that expressed GFP also lacked a nucleolus (Table 1). These results suggest CACN-1 normally acts to inhibit oocyte maturation.

Fig. 5.

CACN-1 inhibits precocious oocyte maturation. OD27 AIR-2::GFP animals were treated with empty and cacn-1 RNAi. AIR-2::GFP was observed in young adults using fluorescent microscopy. (A,B) Control animals exhibit normal oocytes adjacent to the spermatheca, with only the −1 oocyte missing a nucleolus and expressing AIR-2::GFP. (C,D) CACN-1 depletion resulted in oocytes lacking a nucleolus and expressing AIR-2::GFP in the −1, −2 and −3 oocytes. The number of the oocytes, the spermatheca (sp) and visible embryos (Em) are labeled in all images. Arrows mark AIR-2::GFP expression on the chromosomes of oocytes. Scale bar is 25 μm.

3.4. CACN-1 interacting proteins cause germ line and sheath cell defects

Recently, we identified CACN-1 as a component of the C. elegans spliceosome using tandem affinity purification and liquid chromatography/mass spectrometry (TAP LC/MS) (Doherty et al., 2014). Half of the 20 identified interactors are known C. elegans splicing factors or human homologs of splicing factors. Many of these interactors caused distal tip cell (DTC) migration defects similar to those observed in cacn-1 RNAi (Doherty et al., 2014). Two previously known regulators of the sperm-oocyte switch, MOG-5 (Gallegos et al., 1998; Graham et al., 1993) and PRP-17 (Kerins et al., 2010), were both identified as CACN-1 interacting proteins. To determine if the CACN-1 interactors also play a role in germ line development, we depleted each factor by RNAi in N2 animals and compared the results to both control and cacn-1 RNAi. The same seven of the 20 confirmed interactors (F53B7.3, mog-5, prp-17, T11G6.8, snr-6, M03F8.3, C07A9.2) that caused DTC migration defects (Doherty et al., 2014) also caused severe germ line defects in treated animals (Fig. 6). All seven of these interactors are putative splicing factors. Depletion of F53B7.3, mog-5, prp-17, T11G6.8, snr-6, and M03F8.3 resulted in an absence of oocytes or abnormal oocytes in addition to sperm in the proximal gonad. Depletion of C07A9.2 produced gonad arms that contained abnormal oocytes in addition to sperm. The seven CACN-1 interactors were also stained with DAPI and all exhibited the Emo phenotype in gonad arms that contained abnormal oocytes (data not shown).

Fig. 6.

CACN-1 interactors are necessary for proper germ line development. Wildtype animals were treated with RNAi of the indicated CACN-1 interactors. Germ line phenotypes were scored using Differential Contrast Microscopy (DIC). (A) The stacked bar graph indicates the percentage of gonad arms exhibiting the indicated phenotypes. The indicated populations are significantly different from control animals (Fisher's exact t-test, ****p<0.0001). n=the number of gonad arms observed. (B) Representative DIC images of the gonad arms of animals treated with empty, cacn-1, and the seven cacn-1 interactors that caused significant germ line defects. Asterisks mark oocytes and triangles mark sperm-filled proximal gonad arms. Scale bar is 25 μm.

Our cacn-1 network (Doherty et al., 2014) also includes 6 additional predicted interacting proteins selected because they interact with two or more of the proteins already in the CACN-1 network. Of the six additional predicted interactors, only one, CDC-25.1, caused a significant mitotic germ line defect. These animals contained no mitotic germ cell pool but rather exhibited arrested oocytes in both the distal gonad arm and the proximal gonad arm, a phenotype previously described by Ashcroft and Golden (2002). DAPI staining revealed that depletion of CDC-25.1 resulted in oocytes that were arrested at diakinesis in the distal gonad arm through to the proximal arm rather than Emo (Fig. 6B).

We next asked whether CACN-1 interactors are also necessary for the development of functional sheath cells. We dissected and stained gonads with Texas Red-X phalloidin as previously described. Depletion of the 7 CACN-1 interactors resulted in significant defects in sheath cell staining compared to controls (po0.01). Although cacn-1 RNAi resulted in the highest percentage of gonad arms with no sheath cell staining (93%; n=27) several of the interactors (mog-5, M03F8.3, T11G6.8, F53B7.3) also exhibited a high percentage of no sheath staining (>40%) (Fig. 7A). Germ lines were also stained with DAPI to compare sheath cell staining with germ cells present in the gonad. Gonad arms that contained oocytes arrested in diakinesis exhibited the highest percentage of normal and mild defects in sheath staining (40% and 36% respectively; n=106) (Fig. 7B). Proximal gonad arms that contained only sperm almost exclusively lacked sheath cell staining (93%; n=57). Gonad arms that were Emo primarily displayed abnormal sheath cell staining (69%), although some also lacked sheath cell staining (24%; n=108) (Fig. 7B). In summary, proximal gonad arms that lack sheath cells completely typically contain sperm, but lack oocytes, whereas gonad arms with abnormal sheath cells usually exhibit the Emo phenotype. These data suggest that CACN-1 and its interactors are important for proper sheath cell development and subsequent germ line differentiation, development and maturation.

Fig. 7.

CACN-1 interacting proteins support normal somatic gonad and germ line development. Wildtype animals were treated with the indicated RNAi, dissected and stained for actin using Texas Red-X phalloidin and germ line phenotypes using DAPI. Actin and nuclear staining were observed using fluorescent microscopy. Sheath cell staining was scored blindly. A) The stacked bar graph indicates the percentage of gonad arms exhibiting the indicated phenotypes for empty, cacn-1, and CACN-1 interactor RNAi. The indicated populations are significantly different (Fisher's exact t-test, ****p<0.0001, **p<0.01). n=The number of gonad arms observed. B) The stacked bar graph indicates the percentage of gonad arms that exhibit the indicated sheath cell staining phenotypes across the observed germ line phenotypes for all RNAi treatments, n=the number of gonad arms.

Several cacn-1 interactors (skp-1, prp-8,Y37E3.8, ubl-1, eftu-2) are necessary for proper larval development, therefore, we were unable to score germ line phenotypes in the N2 background because they failed to reach adulthood. Conversely, depletion of genes through feeding RNAi is not always sufficient to cause a phenotype, therefore our data likely represent an underestimation of the role CACN-1 interactors play in proper oocyte development and fertility (Simmer et al., 2002; Ahringer, 2006).

3.5. CACN-1 and interactors are important in the soma to regulate germ line development and oocyte maturation

Our data suggest that loss of CACN-1 and its interactors leads to severe sheath cell and germ line defects ultimately resulting in sterility. To determine if these genes function in the somatic sheath cells or in the germ line itself, we performed RNAi depletions in animals resistant to RNAi in somatic cells, NL2098 rrf-1 (pk1417) (Sijen et al., 2001), or the germ line NL2550 ppw-1 (pk2505) (Grishok, 2005). RNAi depletion of CACN-1, and a majority of the interactors (F53B7.3, T11G6.8, mog-5, snr-6, C07A9.2), in the soma using ppw-1 mutants resulted in germ line defects similar to those observed in N2 (Fig. 8A). In contrast, loss of these genes in the germ line using rrf-1 mutants resulted in muted germ line defects compared to those observed in N2 (Fig. 8B). In the rrf-1 background, depletion of cacn-1 resulted in proximal gonad arms that contained sperm, but lacked oocytes less frequently (18%; n=100) than in N2 (42%; n=740). These animals also exhibited a lower percentage of gonad arms that contained abnormal oocytes in addition to sperm (27%; n=100). Germ line depletion of F53B7.3, T11G6.8, mog-5, snr-6, and C07A9.2 in the rrf-1 background had similarly muted effects (Fig. 8B).

Fig. 8.

CACN-1 and most of the CACN-1 interacting proteins act in the somatic gonad. cacn-1 and its seven interactors were depleted via RNAi in the soma using ppw-1 (pk2505) and in the germ line using rrf-1(pk1417). The stacked bar graphs indicate the percentage of animals exhibiting the indicated phenotypes in the (A) ppw-1(pk2505) and (B) rrf-1(pk1417) backgrounds. The indicated populations are significantly different from control animals (Fisher's exact t-test, ****p<0.0001, **p<0.01). n=the number of gonad arms observed.

The remaining cacn-1 interactors, M03F8.3 and prp-17, appear to be required in both the germ line and soma. Somatic depletion of M03F8.3, a homolog of a human pre-mRNA splicing factor, caused developmental arrest in the ppw-1 animals and an array of germ line defects in the rrf-1 background. Therefore, M03F8.3 is likely necessary in both the germ line and soma to direct proper germ line development. As described previously (Kerins et al., 2010), depletion of prp-17, a known regulator of fem-3 expression, the master regulator of the sperm to oocyte switch (Barton et al., 1987), exhibited a masculinization of the germ line (Mog) (Graham and Kimble, 1993) phenotype in the rrf-1 background. Here we show that loss of prp-17 in the soma of ppw-1(pk2505) animals resulted in a high percentage of abnormal oocytes in addition to sperm, rather than the expected Mog phenotype (73%, n=100). These data support a role for prp-17 in the sperm to oocyte switch in the germ line as well as a possible previously unidentified role in the somatic sheath cells (Kerins et al., 2010). Because the rrf-1 (pk1417) soma is not entirely defective for RNAi, the small percentage of germ line defects observed in these experiments may be due to residual RNAi effects in the somatic cells (Sijen et al., 2001). These results suggest that the major locus of cacn-1 action is in the somatic gonad.

4. Discussion

In this study we characterize the role of CACN-1 in germ line and somatic gonad development. We demonstrate that CACN-1 is necessary for the presence and proper morphology of the somatic gonadal sheath cells that regulate both the sperm-oocyte switch as well as oocyte maturation and ovulation. Seven splicing factors that co-purified with cacn-1 are also necessary for proper germ line and somatic gonad development. The majority of these genes exert their effect primarily in the soma, whereas prp-17, as previously described (Kerins et al., 2010), has a more prominent role in the germ line, and MO3F8.3, an uncharacterized splicing factor, is necessary in both the soma and germ line for proper development.

Because these genes have varying roles across the soma and the germ line it is possible that CACN-1, which lacks enzymatic activity, is scaffolding tissue-specific complexes of splicing factors. It is likely that our CACN-1 proteomics experiment identified primarily somatic complexes due to the expression of tagged CACN-1 driven by a heat shock promoter, which is not expressed in the germ line (Stringham et al., 1992). This could explain the predominance of somatic effects in the CACN-1 interactors we tested. It is possible that CACN-1 scaffolds a different set of splicing factors in the germ line to regulate other biological processes, however, these proteins were not identified using our proteomic approach.

We suspect that CACN-1 is exerting its effect on sheath cell development during late larval stages. Studies have shown that ablation of more than one spermathecal/sheath precursor cell (SS cell) in a single gonad arm during the L2/L3 molt leads to defects in germ cell proliferation or meiosis and an inability to produce sperm nor oocytes (McCarter et al., 1997). Ablation of 1SS or 2SS cell during the L2/L3 molt occasionally resulted in the Emo phenotype (McCarter et al., 1997) or even gonad arms that lacked oocytes (Killian and Hubbard, 2005). However, ablation of proximal sheath cells during L4 commonly produced gonad arms that lacked oocytes or exhibited the Emo phenotype (Killian and Hubbard, 2005; McCarter et al., 1997), the phenotypes we observed in this study. More specifically, ablation of sh2-5 produces proximal gonad arms filled solely with sperm (Killian and Hubbard, 2005), whereas ablation of the most proximal sheath cells, sh4-5, results in a 95% penetrance of the Emo phenotype (McCarter et al., 1997). Our interpretation of these RNAi results is that CACN-1 is necessary for development or function of the proximal sheath cells during the last larval stage of development. Strong depletion of cacn-1 results in sperm filled proximal gonad arms, with less effective RNAi resulting in the milder Emo phenotype. It is possible that CACN-1 is necessary for all late somatic cell divisions, but that cacn-1 RNAi affects sheath cells preferentially as they are one of the latest cell dividing lineages in C. elegans development (Kimble and Hirsh, 1979). For example, our preliminary evidence suggests that CACN-1 plays an important role in spermatheca development. CACN-1 may also be playing a role earlier in development, perhaps affecting the lineage of select SS precursor cells, however RNAi knockdown is not sufficient during early development to observe overt defects.

Very little is known about the molecular mechanisms regulating somatic gonad differentiation and development. Disruption of the general splicing of key transcripts that direct development and transitions from larval stages to adulthood could lead to the absence of sheath cells and/or the disruption of actin polymerization we observed in these experiments. In addition to serving a general role in splicing, accessory splicing factors regulate alternative splicing. In C. elegans, 25% of the genome is alternatively spliced (Ramani et al., 2011). Stage-specific splice isoforms help ensure temporal regulation of developmental events. For example, let-2 expresses specific collagen isoforms in larvae and adults, (Sibley et al., 1993) and the expression of active xol-1, the driver of the male fate, is inhibited by the splicing factor fox-1 (Hodgkin et al., 1994), which prevents the excision of the xol-1 terminal intron (Nicoll et al., 1997). In addition, lin-9, which is necessary for proper sheath cell development, is alternatively spliced, producing two different isoforms, lin-9L and lin-9S (Beitel et al., 2000). Lastly, unc-60 is alternatively spliced into two isoforms of ADF/cofilin, unc-60A and unc-60B. The unc-60A isoform is necessary for proper assembly of the contractile actin network of the myoepithelial sheath cells and subsequent oocyte maturation (Ono et al., 2008). CACN-1 and its interacting proteins may be regulating the alternative splicing of lin-9, unc-60 or other key developmental genes that regulate somatic gonad development.

In C. elegans, splicing factors are known to commonly act as RNA binding proteins (RBPs) to regulate gene expression post-transcriptionally (Tamburino et al., 2013). RBPs often bind to the 3′ UTR of mRNA to prevent translation and store transcripts until they are needed (Wilkie et al., 2003). For example, MOG-5, a protein that co-purified with CACN-1, acts as an RBP to prevent the translation of fem-3, the master regulator of the sperm-oocyte switch (Barton et al., 1987), through direct and/or indirect binding of the fem-3 3′UTR (Gallegos et al., 1998; Graham et al., 1993). PRP-17 is also an RBP and may prevent fem-3 translation during the sperm-oocyte switch (Kerins et al., 2010). Recently the canonical eIF4E isoform in C. elegans, ife-3, which has a conserved role of initiating protein synthesis, has been shown to regulate the sperm-oocyte switch by acting as a negative regulator of fem-3 (Mangio et al., 2015). Therefore, it is possible that CACN-1 may be scaffolding splicing factors that act both in the spliceosome and as RBPs that regulate gene expression post-transcriptionally.

Cactin is a well-conserved protein that associates with spliceosomal proteins in multiple species (Baldwin et al., 2013; Jurica et al., 2002; Bessonov et al., 2008; Ilagan et al., 2009; Ashton-Beaucage et al., 2014). Here, we show that CACN-1 and its network of splicing factor homologs are necessary for the presence of sheath cells and subsequent germ line differentiation and maturation in C. elegans. C. elegans, like other organisms, relies on soma-germ line interactions to coordinate the development and differentiation of the germ line to ensure the production of viable offspring. This study provides evidence that splicing and/or post-transcriptional regulation of gene expression may be playing an important unidentified role in somatic gonad development, and in regulation of germ line development by somatic cells. Because of the central role played by splicing and translational regulation of gene expression in development and disease (Echeverria et al., 2015) a better understanding of the CACN-1-scaffolded complexes in C. elegans may provide clues to the regulation of these processes in other species.