Somatic cAMP signaling regulates MSP-dependent oocyte growth and meiotic maturation in C. elegans

The C. elegans hermaphrodite gonad(Fig. 1A) is a paradigm for studying the role of soma-germline interactions. The somatic distal tip cell(DTC) caps the distal end of each gonad arm and maintains a population of proliferating germline stem cells; laser ablation of the DTC causes all germ cells to enter meiosis (Kimble and White,1981). Laser ablation studies also show that cells of the somatic gonadal sheath and spermathecal lineages play multiple roles in germline development (McCarter et al.,1997). DTC signaling, which promotes the proliferation of the germline stem cell population and inhibits entry into the meiotic pathway, is among the best understood soma-germline interactions(Hansen and Schedl, 2006; Kimble and Crittenden,2007).

Following exit from pachytene, germ cells differentiate as spermatocytes in the L4 stage and as oocytes in the adult stage. Actomyosin-dependent cytoplasmic streaming drives oocyte growth in the loop region of the gonad(Fig. 1A)(Wolke et al., 2007). Diakinesis-stage oocytes develop in the proximal gonad arm, with meiotic maturation occurring in an assembly-line fashion. Oocyte meiotic maturation is defined by the transition between diakinesis and metaphase of meiosis I and is accompanied by nuclear envelope breakdown (NEBD), rearrangement of the cortical cytoskeleton and meiotic spindle assembly. In most sexually reproducing animals, oocytes arrest during meiotic prophase I and then resume meiosis in response to hormonal signaling: for example, luteinizing hormone(LH) in mammals and MSP in C. elegans. When sperm are absent, as in mutant hermaphrodites that do not produce sperm (e.g. fog mutant females), oocytes arrest for prolonged periods until insemination(McCarter et al., 1999). The presence of sperm in the gonad also stimulates progression through pachytene and actomyosin-dependent streaming into growing oocytes(Jaramillo-Lambert et al.,2007; Wolke et al.,2007). Sperm secrete MSP by an unconventional vesicle budding mechanism to generate an extracellular MSP gradient(Kosinski et al., 2005). MSP is sufficient to trigger activation of mitogen activated protein kinase (MAPK)in proximal oocytes (Miller et al.,2001), which is required to initiate meiotic maturation(Lee et al., 2007; Arur et al., 2009). Oocytes form gap junctions with smooth muscle-like gonadal sheath cells that regulate meiotic maturation and contract to drive ovulation(Hall et al., 1999; Miller et al., 2003; Govindan et al., 2006).

Oocytes and sheath cells sense MSP through an oocyte MSP/EPH receptor(Miller et al., 2003) and unidentified receptors that are proposed to be G-protein-coupled receptors(GPCRs) (Govindan et al.,2006; Cheng et al.,2008). The VAB-1 MSP/EPH receptor negatively modulates oocyte meiotic maturation in the absence of sperm, and MSP counteracts this modulation (Miller et al.,2003; Cheng et al.,2008). Antagonistic G-protein pathways function in MSP signaling(Govindan et al., 2006; Harris et al., 2006) and regulate the trafficking of VAB-1 in the oocyte(Cheng et al., 2008).

Here we show that Gα_s-adenylate cyclase signaling in the gonadal sheath cells is required for all MSP meiotic maturation responses in the germline, including promoting the cytoplasmic streaming that drives oocyte growth. Thus, the gonadal sheath cells ensure that oocyte production and growth, as well as meiotic maturation, occur efficiently when sperm are present. In the accompanying paper(Nadarajan et al., 2009), we show that GLP-1/Notch signaling functions in the distal germline to regulate MSP-dependent cytoplasmic streaming and oocyte growth. Together, these results suggest that MSP signaling organizes key steps by which germ cells generate zygotes when sperm are available for fertilization.

C. elegans strains used are available on request. Oocyte meiotic maturation rates (McCarter et al.,1999) and cytoplasmic streaming in the loop region(Wolke et al., 2007) were examined. MSP-142 (Baker et al.,2002) was injected (200 nM) into the uterus of unmated females(Miller et al., 2001). Statistical analyses used Student's t-test.

Sheath cell contraction rates were measured as previously described(McCarter et al., 1999),except 0.1% levamisole was used as an anesthetic and measurements were taken immediately. Identical results were obtained using 0.1% tricaine/0.01%tetramisole. Because we found that egl-30 (Gα_q)function is necessary and sufficient to promote sheath contractions, we re-examined the role of vab-1 in sheath contractions(Miller et al., 2003; Yin et al., 2004; Corrigan et al., 2005). Under our current mounting conditions, vab-1(dx31) and vab-1(e2027) null mutants exhibit normal sheath cell contractions. We determined that two factors probably contribute to previous measurements of low sheath contractions in vab-1(dx31). First, vab-1 mutants are more sensitive to the prolonged anesthetic mounting conditions previously used [6.0±3.3 contractions per minute (n=14) in vab-1(dx31) measured immediately after mounting, versus 2.7±2.5 contractions per minute (n=17) when measured 30 minutes after mounting, P<0.001]. Second, vab-1 mutants that had a severe notched head phenotype exhibited lower sheath cell contraction rates even when measurements were made immediately after mounting(3.25±0.90 contractions per minute; n=8).

gsa-1 genetic mosaics were sought using DG1922; pRP1505 (1μg/ml) rescued gsa-1(pk75)(Korswagen et al., 1997) and pTG96 (sur-5::gfp; 75 μg/ml) provided a cell-autonomous marker(Yochem et al., 1998). gsa-1 double mutant backgrounds were similarly analyzed. acy-4 genetic mosaics were sought using a sur-5::gfp array(tnEx37) in which WRM061bE10 (75 μg/ml) rescued acy-4(ok1806). Mosaics were characterized using DIC and GFP fluorescence microscopy; ∼5000 and ∼3000 animals were scored for gsa-1 and acy-4, respectively. We assessed the genotype of the germline by progeny testing, except if both gonad arms were sterile. Somatic gonads mutant for gsa-1 or acy-4 invariably resulted in sterility of the affected arms, suggesting that perdurance is not an issue for these mosaics.

A Zeiss motorized Axioplan 2 microscope with a 63× PlanApo (NA 1.4)objective lens and an apotome adaptor was employed for fluorescence microscopy. Fluorescence images were acquired with an AxioCam MRm camera and AxioVision acquisition software. MAPK activation(Miller et al., 2001), oocyte microtubule rearrangement (Harris et al.,2006), MSP binding (Miller et al., 2003) and extracellular MSP staining(Kosinski et al., 2005) were examined in dissected gonads.

Construction of ACY-4::GFP and GSA-1::GFP fusions used recombineering(Warming et al., 2005). GFP(amplified from pPD95.77) was fused to the ACY-4 C-terminus in the WRM061bE10 fosmid (primer sequences used are available on request). Wild-type hermaphrodites were injected with the acy-4::gfp fosmid at 25μg/ml and pRF4 plasmid at 50 μg/ml. Two extrachromosomal arrays(tnEx42 and tnEx43) were crossed into acy-4(ok1806)and found to restore fertility. In a fog-2(oz40) female background, tnEx42 and tnEx43 did not exhibit elevated maturation rates. GFP was introduced into the first loop of GSA-1 using the WRM0623aH05 fosmid. gsa-1(pk75)/hT2 hermaphrodites were injected with the gsa-1::gfp fosmid at 10 μg/ml and the pRF4 plasmid at 50μg/ml. Homozygous gsa-1(pk75) animals expressing GSA-1::GFP were found to be rescued for fertility.

Antibodies to INX-14 and INX-22 were generated to the C-terminal 119 and 162 residues of INX-14 and INX-22, respectively, using fusions with maltose binding protein (MBP). MBP fusions were purified by SDS-PAGE and used to immunize rabbits and guinea pigs. Affinity purification used glutathione-S-transferase fusions. Antibodies were used to stain dissected and fixed gonads (Finney and Ruvkun,1990). Specificity of the INX-14 and INX-22 antibodies was established by loss of staining following inx-14(RNAi) or inx-22(RNAi) by injection; markedly reduced staining was observed in inx-22(tm1661) gonads using the INX-22 antibodies.

C. elegans possesses a single stimulatory G alpha(Gα_s) subunit, encoded by gsa-1, which is required for viability (Korswagen et al.,1997) and to promote meiotic maturation(Govindan et al., 2006). gsa-1(RNAi) in a wild-type background causes sterility due to a failure of oocytes to undergo meiotic maturation; however, fertility is observed following gsa-1(RNAi) in the rrf-1(pk1417) genetic background (Govindan et al.,2006) that is defective in somatic RNAi responses(Sijen et al., 2001). We conducted genetic mosaic analysis to address the cellular focus of action for Gα_s signaling in oocyte meiotic maturation(Fig. 1B,C). The gsa-1(pk75) null allele was rescued by an extrachromosomal array carrying a cell-autonomous nuclear GFP marker. Mitotic instability of the array generates a clone of gsa-1(0) mutant cells within the lineage that can be recognized by the pattern of GFP expression. Loss of gsa-1 function in the germline lineages (P2-P4; Fig. 1C) did not affect viability or fertility. The progeny of germline mosaic animals arrested as L1 larvae, recapitulating the gsa-1(0) zygotic phenotype. Thus, gsa-1 is not required in the germline for oocyte meiotic maturation,and there is no maternal contribution of gsa-1 to development.

Next, we sought array losses in the MS lineage, which generates the somatic cells of both gonad arms. Losses within the MS lineage cause sterility within a gonad arm because oocytes fail to undergo meiotic maturation despite the presence of sperm (Fig. 1B,C). The fertility of a gonad arm depends on the gsa-1 genotype of the somatic cells within that particular gonad arm. For example, MSppp losses cause sterility in the anterior gonad arm but do not affect the fertility of the posterior arm. For both gonad arms to be sterile, two independent losses are needed, such as for example, losses affecting both precursor cells of the somatic gonad (Z1 and Z4). Consistent with these results, a rescuing gsa-1(+)::gfp transgene is expressed in the sheath and spermathecal cells (Fig. 1G).

The ten follicle-like sheath cells that encompass each gonad arm derive from two lineally distinct sheath-spermathecal (SS) precursor cells(Fig. 1C). Array losses that affect a single SS cell in a gonad arm resulted in fertility, suggesting that gsa-1(+) function in half of the gonadal sheath cells is sufficient. Sheath cells are connected by gap junctions(Hall et al., 1999) and might share second messengers, which could explain the sufficiency of gsa-1(+) function in a subset of sheath cells in a gonad arm. Alternatively, gsa-1(+) function might perdure within the SS lineage.

In the GTP-bound state, Gα_s activates adenylate cyclase(ACY), which catalyzes the formation of cAMP. cAMP binds to the regulatory subunit of protein kinase A (encoded by kin-2), activating the catalytic subunit. In our prior analysis, we concluded that it was unclear which of the four acy genes participates in the regulation of meiotic maturation; however, we suggested an involvement of acy-1 based on an analysis of a gain-of-function mutation(Govindan et al., 2006). Here,loss-of-function studies suggest that acy-4 is required for meiotic maturation, and that acy-1, acy-2 and acy-3 are dispensable(Fig. 1E,F; see Table S1 in the supplementary material). We examined deletions in acy-4(ok1806) and acy-4(tm2510) that remove domains required for catalytic activity,and are likely to constitute null alleles (see Fig. S1 in the supplementary material). acy-4(ok1806) and acy-4(tm2510) hermaphrodites are sterile due to a failure of oocytes to undergo meiotic maturation despite the presence of sperm (Fig. 1E,F). Because acy-4(ok1806) and acy-4(tm2510)show indistinguishable phenotypes and fail to complement, we refer to these loss-of-function (lf) mutations as acy-4(lf). The sterility of acy-4(lf) mutants is not due to defective sperm because mating with wild-type males does not restore fertility, and acy-4(lf)hermaphrodites produce normal numbers of sperm. Early germline development,including the proliferation of germline stem cells, and meiotic entry and progression, are apparently normal in acy-4(lf) mutant alleles; the development of the somatic gonad is also normal.

To investigate the cellular focus of action of acy-4, we conducted genetic mosaic analysis (Fig. 1D). We examined mosaics that had a wild-type germline and a mutant somatic gonad and vice versa. The mosaic analysis showed that acy-4(+) is required in the somatic gonad, but not in the germline,for meiotic maturation and fertility (Fig. 1D). Similar to the gsa-1 genetic mosaic analysis,germline acy-4 losses resulted in the production of broods of animals that recapitulated the acy-4(lf) phenotype and were sterile, showing that there is no essential maternal requirement for acy-4 in development. Remarkably, we found one mosaic animal that had undergone multiple consecutive losses, such that the only acy-4(+) somatic cells in the animal were the gonadal sheath cells of the anterior arm and half of the spermathecal cells of the posterior arm(Fig. 2). In this mosaic, the anterior gonad arm was fertile, whereas the posterior gonad arm was sterile. Thus, acy-4(+) activity is required in the gonadal sheath cells, but not the spermathecal cells, for meiotic maturation. Moreover, expression of acy-4(+) only in the gonadal sheath cells of a gonad arm is sufficient to ensure the fertility of that gonad arm. In agreement with these findings, a rescuing acy-4::gfp fusion is expressed in gonadal sheath cells (Fig. 1G).

Time-lapse microscopy indicated that the hallmarks of meiotic maturation,nuclear envelope breakdown (NEBD) and cortical cytoskeletal rearrangement, do not occur in acy-4(lf) mutants. Quantitative analysis showed that acy-4(lf) mutants have an oocyte meiotic maturation rate of 0.0±0.0 (n=23) maturations per gonad arm per hour, compared with 2.93±0.52 (n=18) in wild-type hermaphrodites. As a result, the oocytes stack up in the gonad arm(Fig. 1E; see Table S2 in the supplementary material). A high-copy acy-4(+) array (tnEx37)was sufficient to trigger meiotic maturation in a female background: fog-3(q443); acy-4(ok1806); tnEx37 females have a meiotic maturation rate of 1.47±0.38 maturations per gonad arm per hr (n=14),compared with 0±0 in array-minus siblings (n=15) and 2.93±0.52 (n=18) in wild-type hermaphrodites. This result suggests that when acy-4 activity is heightened in the gonadal sheath cells by overexpression, meiotic maturation occurs without MSP. Taken together with the genetic mosaic analyses described above, these results show that acy-4(+) activity in the gonadal sheath cells is necessary and sufficient for oocyte meiotic maturation. The meiotic maturation defect in gsa-1(RNAi) and acy-4(lf) mutants is due to a signaling defect rather than an earlier developmental defect, because treatment of adults with the nonspecific phosphodiesterase inhibitors caffeine and isobutylmethylxanthine weakly restores fertility (see Fig. S2 in the supplementary material).

Since oocytes in acy-4(lf) mutants do not undergo meiotic maturation, we examined whether extracellular MSP is present and can bind to sheath cells and oocytes in acy-4(lf) mutants. Disruption of Gα_s-ACY-4 signaling after gsa-1(RNAi) or in acy-4(lf) backgrounds did not affect MSP release from sperm(Fig. 3A). Similarly,interfering with Gα_s-ACY-4 signaling did not reduce or eliminate MSP binding to the gonadal sheath cells and oocytes, as detected using a specific binding assay (Fig. 3B). When the binding assay was carried out in the presence of a 25-fold excess of unlabeled MSP, no binding was observed, indicating that the binding is specific. There was, however, a qualitative difference between MSP-FITC binding to the gonads of wild-type hermaphrodites, and fog-2(oz40) females, or acy-4(lf) or gsa-1(RNAi)hermaphrodites (Fig. 3B; see Fig. S3 in the supplementary material). Whereas MSP-FITC binds to oocytes and sheath cells in a uniform pattern in the wild type(Fig. 3B)(Miller et al., 2003),MSP-FITC labeled ring-like structures in fog-2(oz40) females or acy-4(lf) and gsa-1(RNAi) hermaphrodites(Fig. 3B, insets; see Fig. S3 in the supplementary material). These ring-like structures might correspond to clusters of MSP receptors that form at the sheath-oocyte interface (see Fig. S3 in the supplementary material). We quantified MSP-FITC binding to these ring-like structures in fog-2(oz40) and acy-4(lf) gonads and found that the binding is saturable (see Fig. S3 in the supplementary material), as it is in wild-type gonads(Miller et al., 2003). Importantly, MSP binding to the ring-like structures is observed in vab-1(dx31); acy-4(lf) double mutants (see Fig. S3 in the supplementary material), and thus must represent specific and saturable binding to unidentified receptors. The observation that these apparent receptor clusters form both in females, in which MSP is absent, and in acy-4(lf) mutants, is consistent with the possibility that Gα_s-ACY-4 signaling is required for a normal response to MSP.

Next, we examined molecular markers for MSP-dependent meiotic maturation signaling. MSP signaling triggers activation of MPK-1 MAP kinase (dpMPK-1) in proximal oocytes (Miller et al.,2001), which is required to initiate meiotic maturation(Lee et al., 2007; Arur et al., 2009). Whereas wild-type hermaphrodites exhibited dpMPK-1 in proximal oocytes, acy-4(lf) mutant oocytes did not(Fig. 3C). MSP signaling also promotes reorganization of the oocyte microtubule cytoskeleton prior to NEBD by affecting microtubule localization and dynamics(Harris et al., 2006). Cortical microtubule reorganization is an early response to MSP and has been suggested to promote meiotic spindle assembly by facilitating the search and capture of microtubules by meiotic chromatin following NEBD(Harris et al., 2006). In the proximal oocytes of wild-type hermaphrodites, microtubules were dispersed throughout the cytoplasm in an interphase-like meshwork(Fig. 3D), whereas microtubules were cortically enriched in fog-2(oz40) females and acy-4(lf) mutant hermaphrodites(Fig. 3D). This result further shows that, in the absence of acy-4(+) activity in the gonadal sheath cells, oocytes arrest in diakinesis despite the presence of MSP.

Extracellular MSP displays a graded distribution in the gonads of hermaphrodites and mated females, such that the -1 oocyte receives the highest level of MSP and distally localized oocytes receive less(Kosinski et al., 2005). MAPK activation and microtubule reorganization in response to MSP are initiated in oocytes -1 to -4, depending on the amount of sperm present; however, NEBD only occurs in the -1 oocyte, which is in the correct position for ovulation and fertilization. Whereas an unknown mechanism spatially restricts meiotic maturation to the -1 oocyte (Yamamoto et al., 2006), early MSP-dependent meiotic maturation responses in more distal oocytes might serve to increase meiotic maturation rates when sperm are plentiful. We developed AIR-2::GFP as a marker for graded MSP responses. Using antibodies, it was previously shown that the chromatin localization of the AIR-2 Aurora B kinase is dependent on the presence of sperm; strong chromatin localization of AIR-2 was observed in the -1 oocyte and fainter staining was observed in the -2 oocyte(Schumacher et al., 1998). In a wild-type hermaphrodite background, AIR-2::GFP was localized to chromatin in a sperm-dependent manner; however, we also observed localization to chromosomes of the -2 and -3 oocytes at a reduced frequency(Fig. 3E; Fig. 4B). Specifically, whereas 96% of -1 oocytes had chromatin-localized AIR-2::GFP in wild-type hermaphrodites (n=72), the -2 and -3 oocytes exhibited chromatin-localized AIR-2::GFP in only 32% and 1% of cases, respectively(Fig. 4B). MSP was sufficient to induce chromatin localization of AIR-2::GFP to proximal oocytes of unmated females (Fig. 3E; Fig. 4B). Interestingly, all proximal oocytes (-1 to -4; n=15) exhibited chromatin-localized AIR-2::GFP following MSP injection into unmated females(Fig. 4B), unlike the proximal-to-distal graded distributions observed in hermaphrodites or mated females (Fig. 4B). By contrast,we found that AIR-2::GFP localized to the cytoplasm of all oocytes in gsa-1(RNAi) and acy-4(lf) mutant hermaphrodites(Fig. 3E; Fig. 4B), resembling the localization observed in females. This result suggests that oocytes at any position are not able to sense MSP in the absence of gsa-1(+) or acy-4(+) function. Taken together, these data indicate that Gα_s-ACY-4 signaling is required in sheath cells for multiple MSP signaling responses in oocytes.

Because Gα_s-adenylate cyclase signaling in gonadal sheath cells is required for MSP-dependent meiotic maturation responses in the germline, we asked whether acy-4 and gsa-1 are required for sheath cell contractions. MSP is sufficient to stimulate the basal contractions of the myoepithelial gonadal sheath cells(Miller et al., 2001); the basal sheath contractions occur in the interval between the preceding ovulation and the subsequent onset of NEBD and are less intense than ovulatory contractions (McCarter et al.,1999). Importantly, MSP was found to be a bipartite signal: the 21 highly conserved C-terminal amino acids of MSP was sufficient to stimulate sheath cell contraction, but not meiotic maturation(Miller et al., 2001). We observed normal sheath cell contractions in gsa-1(RNAi) and acy-4(lf) hermaphrodites (Fig. 5). This observation confirms that extracellular MSP is available to bind to sheath cells in acy-4(lf) mutants. Thus, the failure of oocytes to undergo meiotic maturation when the gonadal sheath cells lack gsa-1(+) or acy-4(+) function must be due to an inability to sense or respond to MSP.

Previous results suggested that plc-3, which encodes a phospholipase C, is required for sheath contractions(Yin et al., 2004). Because phospholipase C is a major effector of Gα_q signaling(Neer, 1995), we considered the possibility that the MSP-dependent sheath cell contractions involve the function of egl-30 (Gα_q) signaling. Consistent with this possibility, reduction-of-function (rf) egl-30 mutations and egl-30(RNAi) lowered the sheath cell contraction rate(Fig. 5). By contrast, the egl-30(tg26gf) mutation was sufficient to stimulate sheath cell contractions in the absence of MSP, and egl-30(tg26gf)also increased the contraction rate in hermaphrodites(Fig. 5). EGL-30(Gα_q) is likely to function in the gonadal sheath cells because normal sheath cell contraction rates were observed after egl-30(RNAi) in the somatic RNAi-defective rrf-1(pk1417)strain (Fig. 5). However, a high-copy acy-4(+) array (tnEx37) was sufficient to promote sheath cell contractions in females (Fig. 5). Since acy-4(+) is not required for sheath contractions in hermaphrodites, this finding might be indicative of a regulatory intersection between the egl-30 and gsa-1pathways, as found in studies of locomotion(Schade et al., 2005).

Sheath cells form gap junctions with oocytes, and inx-14 and inx-22, which encode innexin/pannexin gap junction proteins,negatively regulate meiotic maturation in the absence of MSP(Govindan et al., 2006; Whitten and Miller, 2007). INX-14 and INX-22 are expressed in the germline and colocalize at plaque-like structures (Fig. 4C),suggesting that they are components of sheath-oocyte gap junctions. Interestingly, we also observed INX-14 and INX-22 throughout the germline(Fig. 4C), beginning in the L2 stage (T.A.S. and D.G., unpublished). Consistent with this expression data,the inx-14 deletion alleles tm2593 and tm2864caused zygotic-sterile phenotypes owing to a defect in germline proliferation and gametogenesis (see Fig. S4 in the supplementary material). Genetic mosaic analysis of inx-14 confirmed a germline focus of action (see Fig. S5 in the supplementary material).

To test whether communication between sheath cells and oocytes negatively regulates meiotic progression in oocytes -2 to -4, we examined the effects of inx-22 and inx-14(RNAi) on AIR-2::GFP localization(Fig. 4A,B). inx-22(tm1661) and inx-14(RNAi) increased the frequency with which AIR-2::GFP localizes to the chromosomes of the most proximal four oocytes. For example, in wild-type hermaphrodites, the -2 oocyte exhibited AIR-2::GFP chromatin localization in only 32% of cases (n=72),whereas the -2 oocyte almost always (94-100%) exhibited chromatin-localized AIR-2::GFP in inx-22(tm1661) or inx-14(RNAi) hermaphrodites(Fig. 4A,B). We obtained a similar result with a null mutation in ceh-18(Fig. 4B), which encodes a POU-class homeodomain protein required for proper sheath cell differentiation and function (Rose et al.,1997). inx-22(RNAi) and inx-14(RNAi) also extended AIR-2::GFP localization to distal oocytes in a female background (see Table S3 in the supplementary material). Thus, inx-14 and inx-22, and by inference gap-junctional communication between sheath cells and oocytes, negatively regulate the MSP meiotic maturation response.

Since inx-14, inx-22 and ceh-18 are negative regulators of MSP-dependent meiotic maturation, and gsa-1 and acy-4 are positive regulators, we tested their epistatic interaction. Previously we showed that ceh-18(mg57) hermaphrodites are fertile after gsa-1(RNAi) despite exhibiting normal RNAi responses(Govindan et al., 2006). We also reported that depletion of both inx-14 and inx-22 was required to allow meiotic maturation after gsa-1(RNAi). After further backcrossing of the inx-22(tm1661) strain, we found that it exhibited meiotic maturation and fertility following gsa-1(RNAi), despite showing normal germline and somatic RNAi responses(Fig. 4D). To assess epistasis without relying on RNAi, we analyzed genetic mosaics. We reduced gsa-1(+) function in the somatic gonad by genetic mosaic analysis in double mutant backgrounds (Fig. 4D). We found that mutations in ceh-18 and inx-22 suppressed the sterility defect caused by loss of gsa-1(+) function in the somatic gonad(Fig. 4D), suggesting that gap junction proteins function downstream of Gα_s signaling. However, mutations in ceh-18 and inx-22 did not suppress the sterility caused by acy-4(lf). Additional targets of Gα_s-ACY-4 signaling might therefore regulate meiotic maturation in parallel with gap junction proteins. Consistent with this view,a reduction-of-function mutation in kin-2, which encodes the regulatory subunit of protein kinase A (PKA), suppressed the acy-4requirement for fertility (Fig. 4D). Alternatively, there might be two pools of ACY-4, one functioning downstream of GSA-1 and negatively regulating INX-22 [giving suppression of gsa-1(0) by inx-22(tm1661)], and a second pool with an unknown upstream activator and an unknown downstream target[resulting in a failure of suppression of acy-4(lf) by inx-22(tm1661)]. Although we cannot exclude this possibility, the observation that somatic gonad-loss mosaics of gsa-1(0) phenocopy acy-4(lf) mutations makes this explanation less likely.

Sperm stimulate oocyte production in addition to meiotic maturation. We observed that disrupting Gα_s-ACY-4 signaling causes hermaphrodites to produce markedly fewer oocytes compared with the wild type(see Table S4 in the supplementary material). Oocyte production and growth are driven by the transport of materials from the cytoplasmic core (see Fig. 1A). A link between sperm and actomyosin-dependent cytoplasmic streaming was previously suggested(Wolke et al., 2007). To investigate the dependence of cytoplasmic streaming on the presence of sperm,we conducted a time-course analysis (Table 1). In hermaphrodites, we observed constant flow rates in young adults between 16 and 24 hours post-L4. By contrast, unmated fog-2(oz40) females exhibited lower flow rates at 16 and 20 hours post-L4. Flows ceased by 24 hours post-L4, but were restored by mating. To determine whether MSP is sufficient to stimulate the flows, we injected MSP into the uterus of unmated fog-2(oz40) females at 24 hours post-L4 and measured the flows in the loop region an hour later. We found that MSP injection is sufficient to promote the flows(Table 1). Moreover, the response was rapid, occurring within 15 minutes of mating or injection(Table 1). This result further suggests that the flows are not a consequence of meiotic maturation because MSP injection causes flow to resume after 15 minutes, whereas meiotic maturation resumes more slowly (∼25-45 minutes) after injection.

Next we asked whether somatic Gα_s-ACY-4 signaling is required for sustained cytoplasmic streaming. As was the case with unmated females, in gsa-1(RNAi) or acy-4(lf) hermaphrodites the flows ceased by 24 hours post-L4 (Table 1). Somatic ACY-4 activity appears to be sufficient to promote sustained cytoplasmic streaming in the absence of MSP, because we observed flows in females bearing a high-copy acy-4(+) array (tnEx37; Table 1). Remarkably, the flow rates in acy-4(lf); tnEx37 females were significantly higher than in the wild type (12.9±3.0 compared with 5.5±1.8; P<0.0001), perhaps because of an increase in ACY-4 activity. These results suggest that Gα_s-ACY-4 signaling in sheath cells is necessary and sufficient for MSP-dependent cytoplasmic streaming in the germline.

Consistent with the idea that Gα_s-ACY-4 signaling promotes the flows by antagonizing innexin function, we observed that cytoplasmic streaming continues unabated in inx-22(tm1661); fog-2(oz40) females(Table 1). To test whether the cessation of cytoplasmic streaming in gsa-1(RNAi) and acy-4(lf) hermaphrodites is a consequence of the failure of oocytes to mature, we conducted a time-course analysis of the flow rates in oma-1(zu405te33); oma-2(te51) hermaphrodites, in which meiotic maturation does not occur (Detwiler et al.,2001). oma-1(zu405te33); oma-2(te51) hermaphrodites exhibited constant flow rates at all times analyzed, but these rates were lower than in wild type (Table 1). The sustained cytoplasmic flow in oma-1; oma-2hermaphrodites explains why their oocytes grow abnormally large. These observations suggest that meiotic maturation per se is not required for the flows to be sustained; however, these data do not address whether oma-1 and oma-2 are required for normal flow rates, or whether normal flow rates have some dependence on continued meiotic maturation, as previously discussed (Lee et al., 2007).

The results presented here show that the gonadal sheath cells function as the major determinant of all known germline responses to the MSP hormone. Our results show that Gα_s-ACY-4 signaling in the gonadal sheath cells is required for all described MSP responses in the germline occurring throughout the gonad tube, including the actomyosin-dependent cytoplasmic streaming that drives oocyte growth in the loop region of the gonad. In the accompanying article, we show that MSP rapidly induces the phosphorylation of the MLC-4 regulatory light chain of NMY-2 smooth muscle myosin as far as the proliferative zone (Nadarajan et al.,2009). Via MSP signaling, the gonad is rapidly converted into a reproductive mode when sperm are available for fertilization. Meiotic maturation, ovulation and fertilization occur repeatedly in an assembly-line fashion at a rate that matches the number of sperm present(Kosinski et al., 2005). In addition, oocyte production and the cytoplasmic streaming that drives oocyte growth and supplies maternal provisions for embryogenesis are stimulated. Thus, MSP signaling coordinates the growth and development of oocytes with meiotic progression and fertilization.

Based on the examination of the extracellular MSP gradient in the gonad of hermaphrodites and mated females (Kosinski et al., 2005), it seems unlikely that MSP can reach the loop region where oocytes grow. The gonadal sheath cells cover the entire proximal gonad arm and most of the distal gonad arm, and form gap junctions with each other and with the germline, which provides a mechanism by which the gonadal sheath cells could communicate the presence of the MSP signal throughout the entire germline. Transmission electron microscopy and freeze-fracture analyses demonstrated that proximal sheath cells form gap junctions with oocytes(Hall et al., 1999). The observation that INX-14 and INX-22 colocalize within plaque-like structures at the sheath-oocyte interface is consistent with the possibility that these innexins are components of gap junctions between sheath cells and oocytes. inx-14 and inx-22 are negative regulators of meiotic maturation (Govindan et al.,2006; Whitten and Miller,2007), oocyte microtubule reorganization(Harris et al., 2006) and recycling of the VAB-1 MSP/EPH receptor(Cheng et al., 2008) in the absence of sperm. Here we show that inx-14 and inx-22 also function as negative regulators of the MSP response in hermaphrodites and mated females. We also found that inx-22 function is required for actomyosin-dependent cytoplasmic streaming to cease when MSP is absent. Genetic epistasis data suggest that Gα_s-adenylate cyclase signaling promotes meiotic progression in part by antagonizing inx-14and inx-22 function. The genetic finding that inx-14 and inx-22 function in the germline is epistatic to gsa-1function in the gonadal sheath cells is consistent with the possibility that direct communication between these two cell types is an important aspect of the regulation of meiotic maturation. The mechanism by which Gα_s antagonizes innexin function is unclear, but might involve gating. Equally unclear is the nature of the molecules that might move through these junctions, or the direction in which they pass, as previously discussed (Hall et al.,1999).

We consider two models to explain the requirement of Gα_s-adenylate cyclase in the gonadal sheath cells for MSP responses in the germline. In the first model, Gα_s-adenylate cyclase signaling would affect the competence of oocytes to respond to MSP. In this model, Gα_s-adenylate cyclase signaling would not participate in the direct sensing of the MSP gradient, but would function in parallel to enable oocytes or sheath cells to respond by other pathways. This model is difficult to reconcile with multiple lines of experimental evidence. Most importantly, activation of Gα_s-adenylate cyclase signaling in the sheath cells by multiple means is sufficient to drive meiotic maturation in the absence of MSP. For example, overexpression of acy-4 using a high-copy array causes meiotic maturation and elevated rates of cytoplasmic streaming in the absence of MSP(Table 1). Similarly, weak gain-of-function mutations in gsa-1 or a reduction-of-function mutation in the kin-2 regulatory subunit of cAMP-dependent PKA cause meiotic maturation in a female background(Govindan et al., 2006). The possibility that Gα_s-adenylate cyclase signaling has an earlier developmental role in the gonadal sheath cell lineages can be excluded by the finding that treatment of acy-4(lf) adults with phosphodiesterase inhibitors suppresses sterility. Likewise, our data exclude the possibility that MSP is unavailable to bind MSP receptors or that the receptors themselves are not expressed. Rather, we observed extracellular MSP in gsa-1(RNAi) and acy-4(lf) gonads, and in both cases clustered MSP binding sites were observed on sheath cells with a pattern and affinity identical to that observed in unmated females lacking MSP. Crucially, gsa-1(RNAi) and acy-4(lf) hermaphrodites exhibit gonadal sheath cell contractions, which require the 21 C-terminal amino acids of MSP and EGL-30 (Gα_q) signaling.

Based on this reasoning, we favor a second model in which the unidentified MSP receptors on gonadal sheath cells are GPCRs: Gα_s-coupled receptors would trigger meiotic maturation, Gα_o/i-coupled receptors would inhibit meiotic maturation in the absence of MSP and Gα_q-coupled receptors would promote sheath cell contraction(Fig. 6). In this model, the VAB-1 MSP/EPH receptor on oocytes would have a modulatory role. This model is consistent with the finding that the oocyte VAB-1 MSP/EPH receptor accumulates in recycling endosomes, in the absence of MSP, where it functions to inhibit meiotic maturation (Cheng et al.,2008). Importantly, Gα_s-adenylate cyclase signaling was shown to be required for the trafficking of VAB-1 to the oocyte plasma membrane from recycling endosomes when MSP is present(Cheng et al., 2008). This result is consistent with the view that the sheath cells control the MSP response and function as the initial MSP sensor. The identification of the sheath cell MSP receptors will represent a critical test of this model.

The regulation of meiotic resumption in C. elegans and humans is remarkably similar. The somatic cells of the follicle maintain meiotic arrest of mammalian oocytes (Pincus and Enzmann,1935). In C. elegans, the gonadal sheath cells inhibit meiotic maturation when sperm are absent and promote meiotic resumption when sperm are present (Govindan et al.,2006; this work). In both systems, extracellular hormones (LH in mammals and MSP in C. elegans) trigger meiotic resumption through the conserved Gα_s-adenylate cyclase signaling pathway. Mural granulosa cells on the periphery of the follicle express the LH receptor,which is a GPCR. Cumulus granulosa cells form gap junctions with the oocyte using specialized extensions called transzonal projections, which penetrate the zona pellucida and reach the oocyte cell surface(Anderson and Albertini, 1976). Disruption of gap-junctional communication within ovarian follicles using inhibitors is sufficient to promote meiotic maturation(Piontkewitz and Dekel, 1993; Sela-Abramovich et al., 2006). LH signaling was recently shown to promote the closure of gap junctions between somatic cells of the follicle(Norris et al., 2008). Evidence was presented showing that closure of gap junctions within the follicle in response to LH, although sufficient for meiotic resumption, was not strictly necessary, suggesting the involvement of additional pathways(Norris et al., 2008). Similarly, the INX-14 and INX-22 gap junction proteins negatively regulate meiotic progression in both the presence and absence of MSP in C. elegans. Furthermore, an important part of the mechanism by which MSP promotes meiotic maturation involves the antagonistic action of Gα_s-adenylate cyclase signaling and the INX-14 and INX-22 innexins. A major difference between the systems is that Gα_s-adenylate cyclase signaling also has a function within the oocyte to maintain meiotic arrest in vertebrates and mammals(Maller and Krebs, 1977; Mehlmann et al., 2002; Mehlmann et al., 2004; Horner et al., 2003). In C. elegans, Gα_s-ACY-4 signaling functions exclusively in the gonadal sheath cells to promote meiotic maturation, as established by genetic mosaic analysis.