SACY-1 DEAD-Box helicase links the somatic control of oocyte meiotic maturation to the sperm-to-oocyte switch and gamete maintenance in Caenorhabditis elegans.

In sexually reproducing animals, oocytes arrest at diplotene or diakinesis and resume meiosis (meiotic maturation) in response to hormones. In Caenorhabditis elegans, major sperm protein triggers meiotic resumption through a mechanism involving somatic Gα(s)-adenylate cyclase signaling and soma-to-germline gap-junctional communication. Using genetic mosaic analysis, we show that the major effector of Gα(s)-adenylate cyclase signaling, protein kinase A (PKA), is required in gonadal sheath cells for oocyte meiotic maturation and dispensable in the germ line. This result rules out a model in which cyclic nucleotides must transit through sheath-oocyte gap junctions to activate PKA in the germ line, as proposed in vertebrate systems. We conducted a genetic screen to identify regulators of oocyte meiotic maturation functioning downstream of Gα(s)-adenylate cyclase-PKA signaling. We molecularly identified 10 regulatory loci, which include essential and nonessential factors. sacy-1, which encodes a highly conserved DEAD-box helicase, is an essential germline factor that negatively regulates meiotic maturation. SACY-1 is a multifunctional protein that establishes a mechanistic link connecting the somatic control of meiotic maturation to germline sex determination and gamete maintenance. Modulatory factors include multiple subunits of a CoREST-like complex and the TWK-1 two-pore potassium channel. These factors are not absolutely required for meiotic maturation or its negative regulation in the absence of sperm, but function cumulatively to enable somatic control of meiotic maturation. This work provides insights into the genetic control of meiotic maturation signaling in C. elegans, and the conserved factors identified here might inform analysis in other systems through either homology or analogy.

IN sexually reproducing animals, cells of the germ line form gametes, which unite at fertilization and establish a heritable link between generations. Meiosis halves the number of chromosomes contributed by each gamete thereby ensuring the embryo inherits two full sets of chromosomes. Meiosis and germline sex determination are closely coordinated to ensure fertility (reviewed by Kimble and Crittenden 2007; Ewen and Koopman 2010; Murray ). During development germ cells must adopt a sexual fate so as to differentiate either as sperm or oocytes. Sex-specific timing in the meiotic process is commonly observed: spermatocytes proceed through the meiotic divisions in an uninterrupted fashion, whereas oocytes almost invariably arrest once, and sometimes twice, depending on the species. Oocyte meiotic maturation is defined by the transition between diakinesis and metaphase of meiosis I and is accompanied by nuclear envelope breakdown, rearrangement of the cortical cytoskeleton, and meiotic spindle assembly (Masui and Clarke 1979). While the timing of the meiotic divisions with respect to fertilization varies among species, maturation promoting factor (MPF), CDK1/cyclin B, is a universal regulator of oocyte meiotic cell cycle progression (reviewed by Ferrell ; Von Stetina and Orr-Weaver 2011). By contrast, the mechanisms by which the meiotic maturation process is regulated and integrated within the oogenic program are comparatively less well understood. Maternal age-related defects in the oocyte meiotic maturation process represent the largest single source of birth defects and infertility in developed countries, which motivates studies in both mammalian and invertebrate model systems (reviewed by Hassold and Hunt 2009).

The nematode Caenorhabditis elegans has emerged as a useful model for studying the regulation of oocyte meiotic maturation by intercellular signaling (reviewed by Han ; Von Stetina and Orr-Weaver 2011; Kim ). C. elegans hermaphrodites possess two U-shaped gonad arms that produce sperm during the last larval stage and oocytes during adulthood (Figure 1A). Oocyte meiotic maturation, ovulation, and fertilization occur iteratively in an assembly line-like fashion approximately every 23 min (McCarter ). By contrast, strong loss-of-function (lf) mutations in sex-determination genes that feminize the germ line such that exclusively oocytes are produced, cause meiotic maturation to occur infrequently (approximately every 8 hr) and thus oocytes stack up in the gonad arm (McCarter ). Mating of these females to males restores normal rates of meiotic maturation, ovulation, and fertilization (McCarter ). These experiments (McCarter ) and related observations (Ward and Carrel 1979) provided evidence that sperm produce a diffusible signal that promotes oocyte meiotic maturation and oocyte production. Subsequently, a biochemical purification of the maturation-inducing signal from sperm revealed that the major sperm proteins (MSPs) are sufficient to promote oocyte meiotic maturation and gonadal sheath cell contraction (Miller ). MSPs appear to be released from sperm using an unconventional vesicle budding mechanism (Kosinski ). MSPs were originally shown to function as the chief cytoskeletal element in the actin-independent amoeboid locomotion of nematode sperm (Italiano ; Miao ; reviewed by Roberts and Stewart 2012). MSPs were proposed to function as central elements of a sperm-sensing mechanism that couples meiotic maturation and fertilization rates to sperm availability, thereby ensuring efficient progeny production and utilization of resources (Miller ).

Figure 1 

kin-1 is required in the gonadal sheath cells for oocyte meiotic maturation. (A) An adult hermaphrodite gonad arm. Germline stem cells proliferate near the distal tip cell (DTC) and then enter meiosis as they move proximally. Oocytes grow by receiving cytoplasmic flow (arrows) and progress to the diakinesis stage of meiotic prophase I prior to undergoing meiotic maturation (at the −1 position) in response to MSP secreted from sperm in the spermatheca (sp). Five pairs of gonadal sheath cells surround the germ line in each gonad arm. (B) DIC images showing that loss of kin-1(+) function in the somatic gonad prevents meiotic maturation, causing oocytes to stack up in the gonad arm despite the presence of sperm. This genetic mosaic resulted from a complex loss within the Z1 lineage, such that the anterior sheath cells and the anterior DTC were mutant for kin-1, but a few spermathecal cells were kin-1(+) (the mosaic is displayed with anterior to the right). By contrast, wild-type and kin-1(ok338); tnEx109 gonad arms are fertile; embryos are observed in the uterus (ut). Bar, 50 μm. (C) Genetic mosaic analysis of kin-1 in meiotic maturation. Derivation of the somatic gonad and the germ line, the points in the lineage where the kin-1(+) array was lost and the resulting phenotypes are indicated (3378 animals were screened). Circles represent single mosaic animals, with array losses at the indicated position. Squares and stars indicate animals with complex losses affecting the somatic gonad (Figure S2). We did not recover losses of the kin-1(+) array in EMS or MS, similar to what was observed in the genetic mosaic analysis of gsa-1 (Govindan ).

MSP triggers multiple molecular readouts of meiotic maturation, including activation of the MPK-1 mitogen-activated protein kinase (MAPK) (Miller ), which is required for normal meiotic maturation (Lee ). MSP also promotes the reorganization of the oocyte microtubule cytoskeleton (Harris ), the localization of the AIR-2 Aurora B protein kinase to chromatin (Govindan ), and the remodeling of oocyte ribonucleoprotein particles (Jud ). In addition, MSP promotes the actomyosin-dependent cytoplasmic flows that drive oocyte growth and require the continued presence of sperm (Wolke ; Nadarajan ). In this regard, MSP might function in part by promoting the phosphorylation of the MLC-4 regulatory light chain of NMY-2 nonmuscle myosin (Nadarajan ). Thus far, all described outcomes of MSP signaling in the germ line require Gαs–adenylate cyclase activity in the gonadal sheath cells that surround oocytes (Govindan ). Genetic mosaic analysis established that genotypically wild-type oocytes, which are surrounded by gonadal sheath cells that lack Gαs or adenylate cyclase activity, behave as if they do not receive the MSP signal (Govindan ). Further, activation of Gαs–adenylate cyclase signaling in the gonadal sheath cells is sufficient to drive meiotic maturation at robust rates in the absence of sperm (Govindan , 2009).

A key question is how Gαs–adenylate cyclase signaling in gonadal sheath cells promotes oocyte meiotic maturation in response to the MSP signal. Part of the answer to this question is that the gonadal sheath cells form gap junctions with oocytes (Hall ), and these gap junctions function to inhibit meiotic maturation when sperm are absent (Govindan , 2009; Whitten and Miller 2007; T. Starich and D. Greenstein, unpublished results). A loss-of-function mutation in the gene, which encodes the gap junction protein INX-22, suppresses sterility ( encodes Gαs; Govindan ). Here we use genetic analysis to delineate the molecular mechanisms by which somatic Gαs–adenylate cyclase signaling promotes oocyte meiotic maturation. We use genetic mosaic analysis to show that the protein kinase A (PKA) target of Gαs–adenylate cyclase signaling is required in gonadal sheath cells for oocyte meiotic maturation. Not only is PKA activity in the germ line dispensable for meiotic maturation, but PKA does not function in the germ line as a negative regulator of MPF activation, as observed in vertebrate systems (Maller and Krebs 1977; Lincoln ; Han ; Pirino ; Oh ). This genetic result rules out a model in which cyclic nucleotides must move through sheath-oocyte gap junctions to regulate meiotic maturation via PKA activity in the C. elegans germ line, as has been proposed in vertebrate systems (Anderson and Albertini 1976; Sela-Abramovich ; Norris , 2009).

encodes the adenylate cyclase that is required in gonadal sheath cells for oocyte meiotic maturation (Govindan ). To identify new regulators of oocyte meiotic maturation that function downstream of somatic Gαs–ACY-4–PKA signaling, we conducted a genetic screen for suppressor of -4(lf) sterility (Sacy) mutations. We characterized 66 Sacy mutations in at least 17 genes. By using whole-genome sequencing and other positional cloning tools, together with an analysis of previously isolated mutations, we molecularly identified 10 Sacy loci. The centerpiece of our analysis is , which encodes a highly conserved DEAD-box helicase that functions in the germ line downstream of PKA signaling. Genetic analysis reveals that SACY-1 mediates multiple functions necessary for C. elegans reproduction: it is a strong negative regulator of oocyte meiotic maturation; it is a component of the germline sex-determination system, functioning in the hermaphrodite sperm-to-oocyte switch; and it is required to prevent necrotic cell death of gametes. Thus, links the somatic control of meiotic maturation to germline sex determination and the maintenance of oocyte quality. In addition, our genetic screen identified multiple components of a CoREST-like complex and the TWK-1 two-pore domain potassium channel, which we show function in the germline and somatic gonad, respectively. Genetic evidence suggests that CoREST and TWK-1 likely function cumulatively to regulate meiotic maturation. This work provides a foundation for unraveling the genetic control of meiotic maturation signaling in C. elegans. The insights gained may prove informative in the analysis of systems less amenable to forward genetic approaches.

Materials and Methods

Strains

C. elegans were cultured using standard methods at 20° (Brenner 1974), except as otherwise noted. OP50-1, a streptomycin-resistant OP50 derivative, was used for routine strain maintenance and nematode growth medium (NGM) contained 200 μg/ml streptomycin sulfate added before autoclaving. Streptomycin was omitted for all experiments using HT115(DE3), and their respective controls. Alleles generated in this study are described in Table 1, and the molecular changes identified are listed in the Supporting Information, Table S1. In addition, the following mutations were used: LGI: , , , , , , , , , , , , , , , , , , , , , , and ; LG II: and ; LGIII: , , and ; LGIV: , , , , , , , , , , and ; LGV: , , , , , , , and ; LGX: and . The following rearrangements, deficiencies, duplications, and extrachromosomal arrays were used: ; (I; III), (I; III), (IV; V), (I; f), , tnEx31[::gfp], ::gfp], ::gfp], tnEx131[::gfp , tnEx133[::gfp , tnEx134[::gfp , ::, tnEx175[::gfp , tnEx180[::gfp], tnEx181[::gfp ::gfp], tnEx188[ΔC284)::gfp ::gfp]. The genotypes of strains used in this study are listed in Table S2.

Table 1 

Complementation groups of Sacy mutations

Linkage group	Gene	Allelesa
I	pde-6	tn1237, tn1242, tn1336, ok3410
	tom-1b	tn1454, tn1463, ok2437
	sacy-1	tn1385, tn1391, tn1440
	twk-1c	tn1397, tn1398, tn1403
	spr-4	tn1383, tn1402, tn1404, tn1438, tn1444, tn1467, by105
	spr-5c	tn1378, tn1379, tn1394, ar197, by134
	uev-1d	tn1381, tn1382, ok2610
	Unassigned	tn1389, tn1390, tn1392, tn1393, tn1395, tn1415, tn1416, tn1418, tn1419, tn1434, tn1441, tn1442, tn1445, tn1446, tn1464, tn1471
II	sacy-2	tn1401, tn1410, tn1421
	Unassigned	tn1424, tn1428, tn1432, tn1451, tn1452, tn1469
III	sacy-3	tn1396, tn1408, tn1412, tn1414, tn1422, tn1437
	Unassigned	tn1449, tn1456
IV	spr-2e	tn1380, tn1436, ar211, tm4802
	sacy-4	tn1387, tn1413, tn1431, tn1468
	Unassigned	tn1386, tn1455
V	Unassigned	tn1409
X	Unassigned	tn1384, tn1388

tn alleles were isolated in this study; the independently isolated alleles listed suppress acy-4(lf) sterility and fail to complement at least one tn allele. In addition, previously isolated mutations in spr-1 and spr-3 suppress acy-4(lf) sterility (see text). inx-22(tm1661), ceh-18(mg57), and vab-1(dx31) do not suppress acy-4(lf) sterility (Govindan ; S. Kim, J. A. Govindan, and D. Greenstein, unpublished results).

tom-1 encodes the C. elegans tomosyn ortholog (Dybbs ; Gracheva ). In addition to the alleles listed, other tom-1 alleles were tested (ok188, ok285, and tm4724) but found not to suppress acy-4(lf) sterility (Figure S1).

Oocytes stack in the gonad arms of twk-1(tn1397); fog-2(oz40) and spr-5(by134); fog-2(oz40), as they do in fog-2(oz40) females, indicating that neither twk-1 nor spr-5 is a negative regulator of meiotic maturation in the absence of sperm.

uev-1 encodes a ubiquitin-conjugating enzyme variant (Jones ; Kramer ).

spr-2(ar199) mutation does not suppress acy-4(lf) sterility.

Isolation of suppressor of acy-4(lf) sterility (Sacy) mutations

L4-stage ; animals were mutagenized with 50 mM ethyl methanesulfonate (EMS) (Brenner 1974). GFP+ F1 animals were cultured individually and fertile GFP− animals were sought in the F2 generation. Approximately 20,000 haploid genomes were screened and 63 suppressors had sufficient brood sizes to be analyzed further. Brood sizes are expressed as mean ± SD. All Sacy mutations were outcrossed to the parental strain and were recessive. Mutations analyzed in detail were outcrossed at least five times, or as otherwise noted. The polymerase chain reaction verified that all suppressor strains retained the deletion, and did not contain a wild-type copy of the gene or gfp sequences (oligonucleotides used in this study are listed in Table S3). alleles were identified as EMS-induced mutations in N2 that suppressed sterility following . The three alleles failed to complement for this property, but exhibited normal RNAi responses with and triggers.

Genetic mapping and molecular identification of Sacy mutations

Assignment to linkage groups used SNP mapping (Davis ) with crosses to DG2574, which was generated by introgressing the mutation, balanced by , into the CB4856 Hawaiian background using 10 backcrosses. It proved difficult to fine map many of the Sacy mutations in the Hawaiian background, possibly because complex genetic interactions between Bristol and Hawaiian loci modified the penetrance of (S. Kim, J. A. Govindan, and D. Greenstein, unpublished results). In addition, more than half of Sacy mutations localize to LGI and thus the documented incompatibility between Bristol and Hawaiian strains caused by the system (Seidel ) might have distorted the mapping results. Therefore, we utilized a strategy combining complementation testing, whole-genome sequencing, and transgenic rescue. Complementation tests were conducted between Sacy mutations mapping to the same linkage group. Briefly, trans-heterozygotes sacy(a)/sacy(b); ; were constructed and the fraction of GFP− fertile progeny was measured and compared to the parental strains. Because Sacy mutations might exhibit nonallelic noncomplementation, these assignments are viewed as provisional unless validated by sequencing of multiple alleles, transgenic rescue, or suppression of sterility by other available alleles.

Whole-genome sequencing for mutant identification was conducted on Illumina GAIIx and HiSeq2000 instruments according to the manufacturer’s instructions. The average depth of coverage was ∼49-fold. Data were analyzed using MAQGene (Bigelow ). Candidate Sacy mutations in independently isolated alleles were identified and confirmed by Sanger sequencing. Phylogenetic analysis was conducted as described (Dereeper ).

Transgenic rescue and expression studies

Transgenic animals expressing translational gfp fusions were generated using recombineering (Warming ; Tursun ) and either microinjection (Stinchcomb ) or biolistic transformation (Praitis ). To create C-terminal TWK-1::GFP fusions, fosmid WRM0616aE06 was used. To generate a C-terminal truncation of TWK-1::GFP (TWK-1ΔC284::GFP), Escherichia coli GalK was first inserted before Thr284, and then GalK and the C terminus of TWK-1 (residues 284–451) were deleted using recombineering. The C-terminal truncated ::gfp and controls were directly injected into ; animals, and several transgenic lines expressing the ::gfp co-injection marker were established. The transgene arrays were tested for rescuing function in ; animals by restoring sterility. Fosmid WRM0640aH10 was used to generate an N-terminal GFP::SACY-1 fusion. Cre-mediated recombination was used to introduce the gene (Zhang ) into the fosmid for biolistic transformation. A gfp::–expressing extrachromosomal array () was crossed into mutant backgrounds and tested for rescue. To generate C-terminal GFP fusions to ACY-1, ACY-2, and ACY-3, we used fosmids WRM067dG12, WRM0638bH07, and WRM0618cF11, respectively. Fusion constructs were injected into the wild type with as co-injection marker. Transgenes were then crossed into the mutant background.

Genetic mosaic analysis

Genetic mosaic analysis for was performed using a rescuing extrachromosomal array, ::gfp], carrying the cell autonomous ::gfp marker (Yochem ). , , , and animals bearing were used for the analysis. To identify genetic mosaics with array losses in the somatic gonad, young adult hermaphrodites were examined on a Zeiss Axioskop using DIC and fluorescence microscopy with a 100× Plan-Apochromat (numerical aperature, N.A. 1.4) objective lens. To determine the point of array loss in animals exhibiting mosaic expression of the ::gfp marker, the following cells were routinely examined: distal tip cell (DTC), gonadal sheath, spermatheca, coelomocytes, the head mesodermal cell (HMC), body wall muscles, hyp11, intestine, excretory cell, B, F, K, DVA, DVC, and the germ line (the presence of GFP+ progeny could only be scored if at least one gonad arm was fertile). To identify germline-loss mosaics, L4 hermaphrodites were cultured individually. Animals producing entire broods of GFP− progeny were further examined by fluorescence microscopy. Unfertilized oocytes laid by germline mosaics were counted over the first 6 days of adulthood. For , genetic mosaics were identified in similar fashion using ; ; tnEx180[::gfp] as the parent strain. somatic gonad mosaics were fertile in the affected gonad arm, but germline mosaics could not be sought. For genetic mosaic analysis of the gamete degeneration phenotype, we used ; ; ::. Coordinated (non-Unc) animals showing the gamete degeneration phenotype were sought and analyzed by DIC and fluorescence microscopy.

RNA interference

All RNA interference (RNAi) experiments were conducted at 22° using injection (Fire ) or feeding (Timmons and Fire 1998), as modified (Govindan ).

Immunohistochemistry and microscopy

TWK-1::GFP was detected in dissected and fixed (Finney and Ruvkun 1990) gonads using mouse monoclonal anti-GFP antibodies (Abcam; 1:500). Rhodamine phalloidin (Sigma; 1:200) was used to detect actin. MSP and RME-2 were detected in dissected gonads as described (Kosinski ), using mouse monoclonal anti-MSP antibody 4A5 (1:300) and rabbit anti–RME-2 antibody (Grant and Hirsh 1999; kindly provided by B. Grant, Rutgers University, 1:50). Secondary antibodies were Alexa 488-conjugated goat antimouse (Life Technologies, 1:500), Cy3-conjugated goat antimouse (Jackson ImmunoResearch, 1:500), and Alexa 488-conjugated goat antirabbit (Life Technologies, 1:500). Acridine orange was used to stain apoptotic germ cells (Gumienny ) and necrotic gametes in and mutant backgrounds. Adult hermaphrodites (18–24 hr post-L4) were cultured for 1 hr on 8 ml of OP50-1–seeded NGM to which 0.5 ml of M9 containing acridine orange (20 μg/ml) was added. After staining, worms were transferred to fresh medium for 1 hr and then analyzed by fluorescence microscopy. Prior to microscopy, worms were kept in the dark. DIC and fluorescent images were acquired on a Zeiss motorized Axioplan 2 microscope with either 40× Plan-Neofluar (N.A. 1.3) or 63× Plan-Apochromat (N.A. 1.4) objective lenses using a AxioCam MRm camera and AxioVision software.

Results

kin-1 is required in the gonadal sheath cells for oocyte meiotic maturation

Gαs–ACY-4 signaling is required in the gonadal sheath cells for oocyte meiotic maturation (Govindan ). cAMP-dependent PKA is a canonical downstream effector of Gαs–adenylate cyclase signaling. In C. elegans, the catalytic and regulatory subunits of PKA are encoded by and , respectively (Gross ; Lu ). The KIN-2 regulatory subunit functions by binding and inactivating the catalytic subunit in the absence of cAMP; binding of cAMP to KIN-2 alleviates its inhibition of KIN-1 activity (reviewed by Taylor ). In a female genetic background, a reduction-of-function (rf) mutation or derepress oocyte meiotic maturation and MAPK activation in oocytes in the absence of sperm (Govindan ). Further, suppresses the sterility caused by a strong loss-of-function mutation in (Govindan ). experiments using the genetic background, which is sensitive to RNAi in the germ line but resistant in the somatic gonad (Sijen ; Kumsta and Hansen 2012), suggested that functions in the soma to inhibit meiotic maturation in the absence of sperm (Govindan ). Because cyclic nucleotides have been suggested to move through gap junctions to regulate PKA activity in oocytes in mammalian systems (Sela-Abramovich ; Norris ), we sought to assess directly the involvement and focus of action of the kinase in meiotic maturation in C. elegans using genetic mosaic analysis.

We conducted genetic mosaic analysis using the deletion allele. The allele deletes conserved subregions III–VI of the catalytic domain (Hanks ) and introduces a frameshift before subregion VII, which generates multiple stop codons in all known isoforms. exhibits a larval lethal phenotype apparently identical to that caused by an absence of Gαs in the strong loss-of-function allele (Korswagen ). We rescued lethality using an extrachromosomal array bearing a wild-type copy of the gene linked to a cell-autonomous nuclear GFP marker. Approximately 98% of ; ::gfp] animals reaching the L4 stage are fertile (n = 595) and segregate GFP-expressing fertile animals, non–GFP-expressing arrested larvae, and genetic mosaics. Loss of function in germline lineages (P3 and P4; Figure 1C) did not affect viability or fertility. Thus, is not required in the germ line for meiotic maturation. The progeny of germline mosaic animals (n = 26) arrested as L1 larvae, recapitulating the zygotic phenotype.

To assess whether might function in the germ line to inhibit meiotic maturation, as in vertebrate systems (Maller and Krebs 1977; Mehlmann 2005), we asked whether oocytes continue to undergo meiotic maturation and ovulation upon the depletion of sperm through self-fertilization in germline mosaics. We observed that germline mosaics produced a total of 10 ± 14 unfertilized oocytes (n = 15) as compared to 14 ± 21 unfertilized oocytes (n = 18) for nonmosaic siblings (P > 0.5, Student’s t-test). Thus, function is dispensable in the germ line both for meiotic maturation and also for its inhibition when sperm are absent or limiting.

Next, we sought array losses in the MS lineage, which gives rise to the somatic cells of the gonad. Array losses within the MS lineage cause sterility within a gonad arm (Figure 1B). The fertility of a specific gonad arm (anterior or posterior) depends on the genotype of that gonad arm. For the animal to be completely sterile, independent losses were needed affecting both gonad arms (Figure 1C). somatic gonad-loss mosaics exhibit sterility because oocytes fail to undergo meiotic maturation, ovulation, and fertilization, phenocopying mutants (Govindan ). Complex losses found within the sheath-spermathecal lineages suggest that is needed in the gonadal sheath cells, not in the spermatheca, for meiotic maturation (Figure S2).

Genetic and molecular identification of suppressor of acy-4(lf) sterility mutations

To identify new regulators of oocyte meiotic maturation that might function downstream of somatic Gαs–ACY-4–PKA signaling, we conducted a forward genetic screen for mutations that suppress the sterility of mutants (Figure 2A; Table 1). We identified 63 suppressors from ∼20,000 mutagenized haploid genomes and an additional three suppressors from a related screen (see Materials and Methods). We refer to these suppressors as Sacy mutants (suppressor of -4(lf) mutant sterility). All isolated Sacy mutations are recessive. The brood sizes of the Sacy mutants in the background are variable and smaller than those of the wild type (∼10–90 progeny vs. 339 ± 31 (n = 37); Figure 2B). Linkage analysis and complementation testing indicate that the 66 suppressors represent at least 17 genes (Table 1).

Figure 2 

Genetic screen for acy-4(lf) suppressor mutations. (A) acy-4(ok1806) animals possessing an extrachromosomal acy-4(+) array (green) were mutagenized, and F1 progeny were cultured individually. Cultures containing fertile animals not possessing the extrachromosomal array (nongreen animals) were sought in the F2 generation. (B) Brood sizes of Sacy mutants, measured in the acy-4(ok1806) background. Brood sizes of pde-6 alleles were measured in a unc-46(e177) acy-4(ok1806) background. At least 10 hermaphrodites were scored for each genotype. Error bars represent one standard deviation.

Molecular identification of Sacy genes, an overview

We used a combination of positional cloning and whole-genome sequencing to identify 8 of 17 Sacy loci identified in our screen (Figures 3 and 4; Table 1). Based on their molecular identities, strong loss-of-function phenotypes, and likely modes of action, there appear to be several pathways that function cumulatively to affect the regulation of meiotic maturation downstream of PKA. Among these pathways, acts as a strong negative regulator of meiotic maturation and provides a mechanistic connection to the fundamental germline processes of sex determination and gamete maintenance. We will first describe , followed by the nonessential Sacy loci.

Figure 3 

Molecular identification of Sacy mutations in pde-6, twk-1, and CoREST genes. Newly identified Sacy mutations (tn alleles) and independently isolated mutant alleles that suppress acy-4(lf) sterility are shown (asterisks indicate premature stop codons).

Figure 4 

sacy-1 mutations suppress acy-4(lf) sterility. (A) sacy-1 alleles isolated as acy-4(lf) suppressor mutations are shown. The sacy-1(tm5503) deletion is underlined. (B) C. elegans SACY-1 is highly conserved. ClustalW alignment of SACY-1, Drosophila Abstrakt (Irion and Leptin 1999; Schmucker ), and human DDX41. SACY-1 and Abstrakt share 54% (323/603) identity and 70% (424/603) similarity; SACY-1 and DDX41 share 60% (318/533) identity and 75% (401/533) similarity. sacy-1 mutant alleles (triangles) and the DEAD box (boxed in red) are indicated. Ce_SACY-1 (NP_491962.1), Dm_Abstrakt (NP_524220.1), and Hs_DDX41 (NP_057306.2) were used for the analysis. Conserved domains [DEAD-box domain (DEADc), helicase domain (HELICc), and zinc finger domain (ZnF)] and motifs (Q, I, Ia, Ib, II, III, IV, V, and VI) are indicated (Henn ). (C) Rescuing GFP::SACY-1 fusion (tnEx159) is broadly expressed in the nuclei and cytoplasm of most or all cells. Embryos (e), spermatheca (sp), oocytes (−1, −2, and −3). Bar, 50 μm.

SACY-1 DEAD-box helicase functions in the germ line downstream of KIN-1 and is a negative regulator of meiotic maturation in the absence of sperm

Whole-genome sequencing identified three noncomplementing alleles (, , and ) mapping to the center of LGI as missense mutations in , which encodes a DEAD-box helicase related to Drosophila Abstrakt and human DDX41 (Figure 4). We confirmed the missense mutations in using Sanger sequencing (Figure 4) and named . Whole-genome sequencing also identified independent tightly linked missense mutations in in and . encodes a polybromo protein that likely regulates chromatin structures and transcription during development of the gonad (Shibata ). Several lines of evidence indicate that mutations in , not , are responsible for suppression. We separated the two mutations ( and ) from the mutations ( and , respectively) by recombination with closely linked markers. The mutations were able to suppress sterility individually, but the mutations exhibited no activity as suppressors. Further, we examined a deletion allele and found that ; double mutants are sterile and exhibit the meiotic maturation defect (Table S4). Increased suppression of ; by mutations was also not consistently observed (Table S4). Moreover, suppresses sterility (Figure 5A). To assess whether RNAi depletion of function in the germ line or somatic gonad mediates the suppression of sterility, we conducted RNAi experiments in the somatic gonad RNAi-deficient genetic background. using ; ; suppresses sterility in the F1 generation (Figure 5A). Thus, reduction of function in the germ line suppresses sterility. We used recombineering and biolistic transformation to generate an N-terminal GFP::SACY-1 fusion within the fosmid context (Figure 4C) and assessed its rescuing activity (Figure 5, C and D). The GFP::SACY-1 fusion rescues suppression (Figure 5C). We observed that GFP::SACY-1 is expressed in most or all germline and somatic cells and localizes to the nucleus and cytoplasm (Figure 4C).

Figure 5 

acy-4(lf) suppressor mutations in sacy-1 reduce but do not eliminate gene function. (A) sacy-1 RNAi suppresses acy-4(lf) sterility. sacy-1 dsRNAs were injected into the acy-4(ok1806) hermaphrodites bearing an acy-4–rescuing array and non–array-bearing F1 progeny were scored by DIC microscopy. sacy-1 RNAi also suppresses acy-4(lf) sterility in the rrf-1(pk1417) background, suggesting that sacy-1 likely functions in the germ line. The numbers of gonad arms scored are indicated. (B) sacy-1–feeding RNAi induces embryonic lethality. L1-stage larvae were fed bacterial food expressing the indicated dsRNAs and embryonic lethality was scored in the F1 generation. The embryonic lethality is enhanced in sacy-1 mutants, as compared to the wild type. Three independent experiments were conducted and the error bars represent one standard deviation. At least 400 embryos were analyzed for each experimental condition. *P < 0.05, **P < 0.01, ***P < 0.001 compared to the empty vector RNAi controls using Student’s t-test. (C) GFP::SACY-1 (tnEx159) partially rescues sacy-1(tn1391) for suppression of acy-4(lf) sterility. tnEx159[gfp::sacy-1 unc-119(+)] itself does not reduce fertility because sacy-1(tn1391); acy-4(ok1806); tnEx159; tnEx37[acy-4(+) sur-5::gfp] hermaphrodites have a brood size of 105 ± 67 (n = 27). The numbers of animals scored are indicated. (D) GFP::SACY-1 rescues the sterility of the sacy-1(tm5503) deletion allele.

Structural and bioinformatic studies show that DEAD-box RNA helicases contain a highly conserved core helicase domain containing ATP and RNA binding sites (Linder and Jankowsky 2011). Each of the three missense alleles changes a highly conserved glycine residue to an arginine residue in either the DEAD-box or helicase domains (Figure 4B). results in a G533R substitution in motif VI, which contributes to ATP binding and hydrolysis (Linder 2006). The R534 residue adjacent to the mutation site interacts with the γ-phosphate of ATP (Schutz ). results in G331R in motif Ib, which contributes to RNA binding (Schutz ). results in G473R in a region between motifs IV and V, which is surface exposed in the crystal structure of the human ortholog DDX41 (Schutz ).

When separated from the mutation, each missense allele was comparably viable and fertile as the wild type. For example, has a brood size of 349 ± 77 (n = 35), compared to a wild-type brood size of 339 ± 31 (n = 37, P value = 0.49). When placed in trans to the deficiency , which deletes , each missense allele was viable and fertile. Interestingly, when the missense alleles were treated with by feeding, we observed increased levels of embryonic lethality compared to the wild type (Figure 5B). Since the wild type is relatively impervious to (Figure 5B; ∼10% lethality with the full-length dsRNA trigger), an off-target RNAi effect seems an insufficient explanation for the experimental observation. Further, dsRNA triggers that target different portions of the cDNA also cause high levels of embryonic lethality specifically in the missense alleles (Figure 5B). BLAST searches indicate that these dsRNA triggers are highly specific to . It seems unlikely that the missense alleles have an enhanced sensitivity to RNAi because even mutants, which exhibit enhanced RNAi responses (Sijen ), only exhibit a moderate increase in embryonic lethality after (Figure 5B). While we are unable to completely exclude the possibility that off-target effects might contribute to the embryonic lethality observed following , an analysis of the strong loss-of-function allele (Figure 4A) shows it to be an essential gene with a zygotic sterile phenotype (see below). Thus, we conclude that the mutant alleles isolated as suppressors are reduction-of-function alleles.

To determine whether is a negative regulator of oocyte meiotic maturation in the absence of sperm, we feminized the strongest of the missense alleles, , by making double mutants with mutations affecting germline sex determination [e.g., , , , and ]. In all cases we observed increased numbers of oocytes in the uterus (Figure S3B), indicating that is a negative regulator of meiotic maturation. To test whether functions downstream of Gαs–ACY-4–PKA signaling, we removed function from the somatic gonad in genetic mosaics generated in a genetic background. We observed that all genetic mosaics that lost function in somatic gonadal lineages were fertile (Table 2). Thus, functions in the germ line downstream of Gαs–ACY-4–PKA signaling. Additional mutant phenotypes are now described.

Table 2 

sacy-1, spr-5, and twk-1 are epistatic to kin-1

Genotypea	Fertile/total gonad arms	Number of animals screened
kin-1(ok338)	0/13	1822
sacy-1(tn1385) kin-1(ok338)	10/10	1105
spr-5(by134) kin-1(ok338)	12/12	1600
twk-1(tn1397) kin-1(ok338)	11/11	1297

Epistasis tests were conducted in the respective double mutant backgrounds by analyzing genetic mosaics with losses of the kin-1(+)-rescuing array in the somatic gonad lineage. Genetic mosaics were sought in animals bearing the tnEx109[kin-1(+) sur-5::gfp] extrachromosomal array. The genotype refers to the somatic cells of a gonad arm in the genetic mosaics.

sacy-1 functions in the hermaphrodite sperm-to-oocyte switch

We serendipitously found that strongly suppresses the self-sterility caused by (Table 3), a strong loss-of-function mutation (Y29stop) in , which encodes an F-box protein required for spermatogenesis in C. elegans hermaphrodites but not males (Schedl and Kimble 1988; Clifford ; Nayak ). Suppression of self-sterility required that the reduction-of-function mutation be present both maternally and zygotically (Table 3; mrfzrf). also strongly suppressed the self-sterility of (Table 3), another strong loss-of-function mutation (W148stop; Clifford ). The two weaker missense alleles, and , also suppressed the self-sterility of and to varying degrees (Table 3). The suppression of sterility appears to involve a reinstatement of the hermaphrodite sperm-to-oocyte switch in the absence of function because the suppressed animals appear indistinguishable from wild-type hermaphrodites, and they exclusively produce oocytes in the adult stage. The suppression of self-sterility by the missense alleles appears to reflect a loss of function because also suppresses sterility (Table 3). in the ; background, in which the RNAi response is compromised in the somatic gonad but not the germ line, results in efficient suppression of self-sterility (Table 3), suggesting that functions in the germ line as a component of the sperm-to-oocyte switch.

Table 3 

sacy-1 functions in the hermaphrodite sperm-to-oocyte switch

Genotype	sacy-1		Fertilitya (%)	Number scoredb
	Maternal	Zygotic
fog-2(oz40)	WT	WT	0	241
fog-2(q71)	WT	WT	0	175
sacy-1(tn1385); fog-2(oz40)	rf	rf	76c	206
sacy-1(tn1385); fog-2(oz40)	WT	rf	0	182
sacy-1(tn1385)/+; fog-2(oz40)	rf	WT	0d	72
sacy-1(tn1391); fog-2(oz40)	rf	rf	16	244
sacy-1(tn1440); fog-2(oz40)	rf	rf	30	236
sacy-1(tn1385); fog-2(q71)	rf	rf	49	162
sacy-1(tn1391); fog-2(q71)	rf	rf	12	192
sacy-1(tn1440); fog-2(q71)	rf	rf	10	164
fog-2(oz40); control RNAie	WT	WT	0	1116
fog-2(oz40); sacy-1 RNAi	RNAi	RNAi	49f	792
rrf-1(pk1417); fog-2(oz40);	RNAi	RNAi	71g	753
sacy-1 RNAi
tra-2(e2020gf); sacy-1 RNAi	RNAi	RNAi	0	74
sacy-1(tn1385); fem-3(e1996)	rf	rf	0	290
fog-1(e2121) sacy-1(tn1385)	rf	rf	0	342
sacy-1(tn1385) fog-3(q470)	rf	rf	0	348

Fertility was scored on a per-animal basis, except for the RNAi experiments, in which case fertility was scored on a per-gonad arm basis because sacy-1(RNAi) in fog-2(oz40) or rrf-1(pk1417); fog-2(oz40) backgrounds caused a gamete degeneration phenotype in 3–4% of gonad arms. This gamete degeneration phenotype was the same as that caused by sacy-1(tm5503).

Number scored refers to gonad arms for RNAi experiments and animals for the remainder.

Brood size of fertile animals was 194 ± 88 (n = 40).

sacy-1(tn1385); fog-2(oz40) females were crossed to gsa-1(pk75)/hT2[qIs48]; fog-2(oz40) males and GFP+ XX progeny were analyzed. The hT2 balancer contains an insertion of a myo-2::gfp.

Cbr-lin-12 dsRNA were used as a negative control RNAi (Felix 2007).

Brood size of fertile animals was 172 ± 94 (n = 42).

Brood size of fertile animals was 148 ± 71 (n = 40).

Genetic analysis has identified many of the key genes that control sex determination in C. elegans (Figure 6; reviewed by Ellis and Schedl 2007; Kimble and Crittenden 2007). Therefore we assessed whether suppresses the self-sterility of strong loss-of-function mutations in , , and . In these experiments we ensured that the reduction-of-function mutation was both maternally and zygotically homozygous, but in no case did we find evidence for suppression (Table 3). Biochemical studies established that FOG-2 can form a ternary complex with the KH-domain protein GLD-1 and the 3′-UTR of (Clifford ), consistent with the proposal that GLD-1 binds the 3′-UTR to mediate translational repression as a key element of the hermaphrodite sperm-to-oocyte switch (Jan ). Therefore, we examined whether could suppress germline feminization and self-sterility caused by the dominant mutation that deletes GLD-1 binding sites within the 3′-UTR (Jan ; Clifford ). We found that was unable to suppress the self-sterility of mutants (Table 3). Taken together, these results suggest that functions in the hermaphrodite sperm-to-oocyte switch upstream of .

Figure 6 

The C. elegans germline sex-determination pathway (genes promoting the male and female fate are shown in blue and black, respectively). The data in Table 3 suggest that sacy-1 promotes the oocyte fate antagonistically to fog-2, which promotes spermatogenesis.

sacy-1 prevents necrotic cell death of gametes

The analysis thus far relied on weak reduction-of-function alleles recovered as suppressors. To address whether plays essential roles during oogenesis, we analyzed the deletion allele generated by S. Mitani. results from a 619-bp deletion that removes the entire second and third exons and a portion of the fourth (Figure 4A). Potential unspliced or alternatively spliced messages are either predicted to be out of frame or to lack conserved regions of the DEAD-box domain. homozygous hermaphrodites produced from heterozygous parents develop to adulthood but are sterile. The gfp:: transgene fully rescues the sterility (Figure 5D). adult hermaphrodites do not produce fertilized embryos (Table 4); instead they contain oocytes and sperm that become vacuolated and appear to degenerate (Figure 7). We conducted a time-course analysis to examine the onset and progression of gamete degeneration. By DIC microscopy, we observed a mixture of small and large vacuoles in sperm and oocytes on day 1 of adulthood. With time these vacuoles appeared to grow in size or fuse (Figure 8B and Figure S4). Ultimately, we observed the gonad arms to contain gamete remnants in which Brownian motion was observed to occur in residual cytoplasm. We also observed a similar sperm degeneration phenotype in adult males (Figure 7A), which were never observed to sire cross-progeny. The somatic gonad appeared to develop normally in hermaphrodites and males. We sought genetic mosaics using the ; ; :: strain in which the rescuing GFP::SACY-1 fusion is expressed in most or all germline and somatic cells and serves as a cell-autonomous marker for mosaic analysis. The genetic background was utilized as a marker for the AB lineage so as to identify rare non-Unc mosaics showing the gamete degeneration phenotype. We found four genetic mosaics with P1 losses, a single mosaic with a P3 loss (this animal had an independent loss within the C lineage), and two mosaics with P4 losses (Figure 7B; n > 40,000). This result indicates that function is required in the germ line to prevent gamete degeneration.

Table 4 

unc-68(e540) partially suppresses the sacy-1(tm5503) gamete necrosis phenotype

Genotype	Fertilitya (%)	Number of gonad arms scored
sacy-1(tm5503)	0	266
sacy-1(tm5503); unc-68(e540)	20	258
sacy-1(tm5503); unc-24(e138)	0	208
sacy-1(tm5503); unc-32(e189)	3	232
sacy-1(tm5503); unc-33(mn407)	8	136

Day-1 adults were examined by DIC microscopy. Gonad arms producing fertilized embryos were scored as fertile although the embryos failed to hatch.

Figure 7 

sacy-1 is required for gamete maintenance. (A) sacy-1(tm5503) adult hermaphrodites and males produce gametes that degenerate. Embryos (e), spermatheca (sp), oocytes (−1, −2, and −3), vulva (vu), sperm (s). (B) sacy-1 functions in the germ line to prevent gamete necrosis. GFP::SACY-1 fusion rescues sacy-1(tm5503) sterility (top). A genetic mosaic that lost GFP::SACY-1 in the primordial germ cell P4 exhibits gamete necrosis and is sterile (bottom). (C) sacy-1(tm5503) hermaphrodites produce male and female gametes that ultimately degenerate. The yolk receptor RME-2 and MSP were used for markers of oocyte and sperm fates, respectively. Proximal is to the left. Bars, 50 μm.

Figure 8 

Germline feminization delays the onset of oocyte necrosis in sacy-1(tm5503) mutants. (A) Oocytes in fog-3(q470) females stack within the gonad arm, and the uterus and spermatheca (sp) are empty (top). In sacy-1(tm5503) fog-3(q470) females, oocytes undergo meiotic maturation despite the absence of sperm, and the uterus fills with unfertilized oocytes (bottom). Bar, 50 μm. (B) Feminization of the gonad delays the onset of oocyte necrosis. A time-course analysis of gamete degeneration conducted over the first 4 days of adulthood. The numbers of gonad arms scored are indicated. The severity of degeneration was scored using a qualitative scale. Representative images illustrative of the scoring criteria are shown in Figure S4.

To address whether gamete degeneration in is dependent on the apoptotic pathway, we examined double mutants between and or . Both and are required for apoptosis (Ellis and Horvitz 1986); encodes a caspase (Yuan ) and encodes an Apaf-1–like protein (Yuan and Horvitz 1992). Using DIC microscopy, we observed gamete degeneration in 99% of ; adult hermaphrodite gonad arms (n = 150) and in all ; gonad arms (n = 108). Further, we used acridine orange staining to examine early degenerating gametes. We observed that acridine orange stains early degenerating gametes in the proximal gonad of ; double mutants, as well as single mutant hermaphrodites (Figure 9). Apparently, the sheath cells might engulf and acidify some of these degenerating gametes. These results indicate that gamete degeneration in is independent of the chief apoptotic effectors.

Figure 9 

The gamete degeneration phenotype in sacy-1(tm5503) is independent of apoptosis. Acridine orange was used to identify germ cells dying by apoptosis or necrosis. Wild-type and sacy-1(tm5503) hermaphrodites exhibit apoptotic germ cells in the gonadal loop region (arrowheads), but ced-3(n717) hermaphrodites do not. sacy-1(tm5503) and sacy-1(tm5503); ced-3(n717) hermaphrodites exhibit acridine orange staining in the proximal gonad arm (arrows) that appears to coincide with degenerating gametes. This proximal acridine orange staining is not observed in wild-type hermaphrodites. Proximal is to the left. Bar, 50 μm.

As a test of whether gamete degeneration in mutant hermaphrodites involves necrotic cell death, we examined a ; double mutant. encodes the ryanodine receptor (Maryon ) and an mutation was found to reduce the penetrance of necrotic cell death (Xu ). We found that the mutation decreased the penetrance of gamete degeneration; whereas all gonad arms failed to produce fertilized embryos, 20% of ; gonad arms produced fertilized embryos (Table 4). The fertilized embryos in ; animals failed to hatch. As controls, we tested mutations in , , and , but found that none were as effective as in ameliorating gamete degeneration (Table 4). We did observe that 8% of gonad arms were fertile in ; animals (Table 4). encodes a microtubule-binding CRMP protein that is exclusively expressed in neurons and is required for normal axon guidance and elongation (Maniar ). The slight reduction of gamete necrosis in ; animals might be a secondary physiological consequence of their slow growth.

Germline feminization delays the onset of gamete degeneration and reveals sacy-1 as a strong negative regulator of meiotic maturation

We next investigated the genetic requirements for gamete degeneration in hermaphrodites. Feminization with strong loss-of-function mutations in the sex determination pathway, , , , or , delayed the time of onset and the severity of oocyte degeneration in females (Figure 8 and Figure S3; S. Kim and D. Greenstein, unpublished results). By contrast, masculinization of the germ line using a gain-of-function mutation in (Barton ) did not suppress gamete degeneration; we observed vacuolated and morphologically abnormal sperm in ; animals (Figure S5). Upon mating, females produce embryos that arrest without properly undergoing morphogenesis and fail to hatch. We did not explore the basis for this embryonic lethality further. We did not observe mating to wild-type males to overtly increase the penetrance or severity of oocyte degeneration. Possibly, the presence of mutant sperm in the gonad arm might potentiate oocyte degeneration. While the physiological basis for the delayed onset of oocyte degeneration upon germline feminization is unclear, this phenomenon proved useful in that it enabled us to examine oocytes and embryos produced by females. In all female backgrounds tested, we observed oocytes in females to undergo meiotic maturation and ovulation at apparently high rates; the uterus filled with unfertilized oocytes (Figure 8A and Figure S3C). We also observed apparently defective ovulation in females such that the gonad arms often contained endomitotic oocytes. We confirmed that MSP was undetectable in females as expected (Figure S6). Thus, is a strong negative regulator of oocyte meiotic maturation in the absence of sperm.

Since only reduction-of-function alleles were recovered as suppressors, we wished to determine the genetic behavior of a loss-of-function allele. Thus, we conducted genetic epistasis analysis between and using strong loss-of-function alleles of both genes. In a hermaphrodite background, we observed gamete degeneration in all ; animals examined (n = 94). Therefore gamete degeneration is independent of signaling. We therefore employed germline-feminizing mutations to overcome the gamete degeneration phenotype in conducting epistasis experiments between and . In a ; background, we observed oocyte meiotic maturation to occur constitutively; however, ovulation typically failed and endomitotic oocytes accumulated in the gonad arm (Figure S7). All examined ; gonad arms contained endomitotic oocytes (n = 42), in contrast to ; gonad arms, which exclusively contained oocytes arrested in diakinesis (n = 43). Thus, is epistatic to for meiotic maturation, as was also observed for the reduction-of-function alleles.

Since is a strong negative regulator of oocyte meiotic maturation, we investigated epistasis with and , which encode TIS-11 zinc-finger proteins that are redundantly required for oocyte meiotic maturation (Detwiler ). In a hermaphrodite background, we observed gamete degeneration in ; ; animals. By contrast, in a ; ; females, we observed that oocyte degeneration was markedly delayed. We observed that diakinesis-stage oocytes failed to undergo meiotic maturation and ovulation and accumulated in the gonad arms of all quadruple mutant female animals examined (n = 30). Thus, and appear to function downstream or in parallel to in the regulation of meiotic maturation.

Multiple nonessential Sacy genes mediate the somatic control of oocyte meiotic maturation

In contrast to , which is an essential gene, we now describe seven nonessential Sacy loci (, , and ) that affect the regulation of meiotic maturation by somatic Gαs–adenylate cyclase–PKA signaling.

PDE-6 phosphodiesterase is a negative regulator of meiotic progression:

We isolated mutations in a related screen for mutations that suppress sterility following yet exhibit normal RNAi responses and are viable and fertile. From this screen, only mutations from a single complementation group, represented by , , and , suppress sterility. One of these mutations, , was tested and also found to suppress the sterility caused by mosaic loss of activity in the somatic gonad, as expected (Table S5A). A combination of whole-genome and Sanger sequencing identified independent nonsense mutations in each allele (Figure 3). We also found that the deletion allele suppresses sterility and fails to complement for suppression of sterility. encodes a phosphodiesterase and the mutant alleles introduce stop codons prior to or within the coding sequence for the PDE domain (Figure 3), suggesting they represent strong loss-of-function alleles. The likely human ortholog of PDE-6 is PDE8 (Figure S8), the high-affinity cAMP-specific phosphodiesterase, which specifically lowers cAMP levels via phosphodiester bond hydrolysis (Fisher ; Soderling ). The chromatin localization of the AIR-2 Aurora B kinase in proximal oocytes is a marker for graded MSP responses (Schumacher ; Govindan ). In an background, the mutation extends AIR-2::GFP chromatin localization distally (Table S5B), suggesting that an enhancement in Gαs–ACY-4–PKA signaling in gonadal sheath cells results in a heightened MSP response in oocytes.

Mutations of multiple CoREST components suppress acy-4(lf) sterility:

Three noncomplementing alleles, , , and , map to the right end of LGI and define the gene (Figure 3). Using , we genetically mapped the suppression to the interval between SNPs and , which contains two genes, and . Previously isolated mutations, the strong loss-of-function allele, (Y284stop) (Eimer ; Nottke ), and (A665T) (Jarriault and Greenwald 2002), also suppress sterility and fail to complement for suppression. Consistent with this gene identification, injection of dsRNA into ; ::gfp] hermaphrodites suppresses sterility in the F1 generation (Figure S9). experiments conducted in the somatic gonad RNAi-deficient ; background indicate that depletion in the germ line results in meiotic maturation in the absence of (Figure S9). by itself does not appear to function as a negative regulator of meiotic maturation in the absence of sperm because oocytes stack in the gonad arm in ; females (Table 1).

and other spr genes were originally identified as suppressors of the presenilin mutant egg-laying defect (Wen ; Eimer ; Jarriault and Greenwald 2002; Lakowski ; reviewed by Lakowski ). The spr genes encode multiple chromatin-modifying components that might constitute a C. elegans CoREST-like complex, similar to mammalian CoREST (corepressor for element-1–silencing transcription factor; reviewed by Lakowski ). encodes an H3K4me2 demethylase that is thought to contribute to transcriptional repression by remodeling chromatin structure (Eimer ; Jarriault and Greenwald 2002; Katz ). mutations have been shown to confer a mortal germline phenotype after more than ∼20 generations (Katz ) and to exhibit modest defects in meiotic DNA double-strand break repair (Nottke ). Suppression of sterility is observed in the first generation in which mutations become homozygous and is efficient in subsequent generations. We have not examined germline mortality in these strains in detail, but did observe declines in fecundity after many generations, consistent with prior findings. mutations bypass the presenilin requirement for egg laying by derepressing presenilin expression (Eimer ). Therefore, we considered the possibility that mutations might bypass the requirement of for meiotic maturation via derepression of other adenylate cyclase(s) in the gonadal sheath cells. We generated C-terminal GFP fusions for ACY-1, ACY-2, and ACY-3 using fosmid recombineering and compared their expression in the wild type and mutants. We observed apparently identical expression patterns in the wild type and mutants and in no case did we observe expression of an adenylate cyclase other than in the gonadal sheath cells (S. Kim and D. Greenstein, unpublished results). As a more direct test, we conducted genetic mosaic analysis of in an mutant background. We observed that loss of activity in the somatic gonad in the genetic background resulted in fertility, in contrast to the wild-type background (Table 2). Thus, functions downstream or in parallel to , consistent with a function in the germ line. Prior work identified >100 genes upregulated in mutant gonads (Nottke ). We tested whether RNAi of any of these candidate genes might restore sterility to ; mutants, but in no case was such a result obtained. Possibly might regulate genes not identified by the microarray analysis or multiple pathways might contribute to the suppression of sterility.

During this analysis, we observed that unlike the wild type, worms exhibit a growth defect on standard nematode growth medium with the bacterial strain HT115(DE3) as a food source, as opposed to OP50-1. Under these conditions, animals exhibit larval lethality or arrest, or they grow slowly and appear grossly unhealthy (Figure S10). This phenotype is rescued by an extrachromosomal array (), indicating that an absence of activity prevents normal growth on the HT115(DE3) food source. This is an unexpected result because HT115(DE3) is a high-quality food source for C. elegans (Brooks ; Y. You, personal communication). We found that also suppresses the HT115 growth defect: double mutant animals are viable and fertile with HT115 as a food source, as they are on standard OP50 bacterial strains. While the basis for this unusual phenotype remains to be determined, we nonetheless screened all Sacy mutants and found two additional loci that suppress the HT115 growth defect. Genetic mapping and DNA sequencing established that these two Sacy genes are and (Figure 3; Table 1). We therefore tested independently isolated alleles of spr loci and found that and suppress sterility and the HT115 growth defect. We did, however, observe that failed to suppress sterility. Interestingly, specifically eliminates the expression of SPR-2 in somatic cells after the 50-cell stage (A. Gontijo and B. Lakowski, personal communication), consistent with a role for CoREST function in the germ line downstream of . Similarly, we observed that suppresses sterility and the HT115 growth defect. We did not isolate mutant alleles in our screen, yet both and suppress sterility. Neither of these alleles suppresses the HT115 growth defect. We constructed double mutants and observed that they were fertile and suppressed for the HT115 growth defect (Figure S10). These results suggest that a CoREST-like complex has a function in the germ line that participates in the regulation of oocyte meiotic maturation by the somatic gonad and sperm signaling. Further, the observation that mutations in , , , and suppress the HT115 growth defect suggests that CoREST-like complex genes and might coordinately function in additional processes besides oocyte meiotic maturation.

A two-pore domain potassium (TWIK) channel functions in sheath cells downstream of KIN-1 to regulate meiotic maturation:

Whole-genome sequencing identified three noncomplementing alleles (, , and ) mapping to the center of LGI as nonsense mutations in , which encodes a TWIK channel (Figure 3). and introduce stop codons after the fourth transmembrane domain (W330stop and E272stop, respectively) and introduces a stop codon after the second transmembrane domain (Figure 3; W177stop). The crystal structures of TWIK channels suggest that all four transmembrane helices likely participate in the formation of a functional channel complex (Brohawn ; Miller and Long 2012). is therefore a strong loss-of-function allele. , but not or , exhibits an adult-onset paralyzed uncoordinated (Unc) phenotype in a wild-type genetic background. Thus, and must reduce but not eliminate function. Interestingly, ; animals do not exhibit an adult-onset Unc phenotype. Thus, mutant alleles are recessive suppressors of sterility, and is a recessive suppressor of the movement defect. To test whether mutations are causal for the suppression of sterility and the adult-onset Unc phenotype, we conducted rescue tests by introducing a fosmid clone, WRM0616aE06, into ; + and genetic backgrounds, respectively. ; animals bearing the extrachromosomal array were sterile and exhibited the defect in meiotic maturation (Table S6A). Further, we also observed rescue of the Unc phenotype (Table S6B). The suppression of the Unc phenotype by is partially affected by function because 23% (n = 95) of ; animals exhibit the adult-onset Unc phenotype as compared to 0% of ; adults (n = 93). Thus, , , and genetically interact in a separate biological context.

In C. elegans, high-copy extrachromosomal arrays generated by microinjection, as done for , are typically silenced in the germ line (Kelly ). To test more definitively whether sterility requires function in the somatic gonad, we conducted genetic mosaic analysis of suppression. We sought fertile genetic mosaics among ; ; tnEx180[::gfp] animals, produced by heterozygous parents using as a dominantly marked balancer chromosome for . Two genetic mosaics with losses in the EMS founder cell, which is a precursor to the somatic gonad (Figure 1C), were fertile in both gonad arms. These genetic mosaics produced GFP-containing progeny, indicating that their germ lines were . A third genetic mosaic animal resulted from complex losses within the Z1 lineage (Figure S11). This animal was fertile in the anterior gonad arm, but was sterile in the posterior gonad arm. All gonadal sheath cells in the anterior gonad arm of this mosaic were mutant, but the germline and several anterior spermathecal cells were . A fourth genetic mosaic, resulting from a loss in the Z4 somatic gonadal precursor cell (and some cells within the C lineage; see Figure 1C), was fertile in the posterior gonad arm, but not the anterior gonad arm. The germ line of this animal was also . We conclude that functions downstream of in the somatic gonad to regulate meiotic maturation.

To examine TWK-1 expression, we used recombineering to fuse GFP to the C terminus of TWK-1 within the fosmid context and generated transgenic lines in wild-type and mutant backgrounds. The ::gfp extrachromosomal arrays rescued both suppression and Unc phenotypes (Table S6A), indicating that the TWK-1::GFP fusion protein is functional in vivo and might represent the endogenous expression pattern. TWK-1::GFP is expressed in the gonadal sheath cells, the distal tip cell, and a few unidentified neurons, but is not expressed in spermathecal cells (Figure 10). TWK-1 is not itself a strong negative regulator of meiotic maturation as oocytes in unmated ; females stack in the gonad arm and do not exhibit elevated rates of meiotic maturation (Table 1). To test whether functions downstream or in parallel to somatic PKA signaling needed for oocyte meiotic maturation, we conducted genetic mosaic analysis in ; ::gfp] animals, similar to what was done for and . All genetic mosaics that lost function in somatic gonadal lineages were fertile (Table 2). Thus, is epistatic to for oocyte meiotic maturation. Taken together, these data suggest that TWK-1 has a function in gonadal sheath cells that contributes to the regulation of meiotic maturation by Gαs–ACY-4–PKA signaling.

Figure 10 

Expression of TWK-1::GFP in the somatic gonad. Immunostaining of TWK-1::GFP in dissected and fixed gonads using anti-GFP antibodies. TWK-1::GFP is expressed in the distal tip cell (DTC) (A) and the gonadal sheath cells (sh), but not spermathecal cells (sp) (B). Phalloidin was used to detect actin in the proximal gonadal sheath cells and the spermatheca. Identical exposure times were used to acquire GFP images. Proximal oocytes (−1, −2, and −3). Bar, 50 μm.

TWIK channels conduct potassium ions across the plasma membrane to control the negative resting potential of excitable cells. The finding that TWK-1 functions downstream or in parallel to PKA signaling provides additional evidence that the sheath cells play a critical role in regulating C. elegans oocyte meiotic maturation. TWK-1 has several human homologs, including TREK-1 and TREK-2 (Figure S12A). Application of intracellular cAMP or stimulation of a Gαs-coupled receptor block TREK-1 and TREK-2 channel activity (Patel ; Lesage ). Interestingly, electrophysiological studies found that C-terminal truncation of TREK-1 shortly after the fourth transmembrane domain renders the channel insensitive to cAMP inhibition and thus constitutively open (Patel ). Multiple PKA phosphorylation sites in the C terminus were shown to mediate cAMP regulation of the channel (Patel ; Murbartian ; Kang ). We therefore considered the possibility that TWK-1 C terminus might play an essential regulatory function. However, and reduce but do not eliminate activity and are predicted to truncate the protein at positions 272 and 330, respectively (Figure S12). Thus, the TWK-1 channel must retain some function in the absence of the C terminus. As a further test of whether the C terminus of TWK-1 is essential for function, we asked whether a TWK-1 C-terminal truncation (ΔC284) could rescue the strongest loss-of-function allele, . We chose the position of the C-terminal TWK-1 deletion to correspond to the TREK-1 deletion that renders the channel constitutively open. We observed that TWK-1ΔC284::GFP could mediate function for both oocyte maturation and motility (Figure S12B). Another possibility is that the regulation of the channel via the C terminus might be important for closing the channel to facilitate meiotic maturation. If so, then a constitutively open channel, resulting from a C-terminal truncation, might confer a dominant disruptive effect on the meiotic maturation and ovulation process. Such an effect is predicted to be most evident with increased dosage achieved through expression from high-copy arrays. Since TWK-1ΔC284::GFP is functional in rescue assays, it seems unlikely that regulation of TWK-1 by PKA, if it exists in this system, plays an essential role. Alternatively, the C-terminal deletion might produce offsetting effects by concomitantly reducing function and removing inhibitory regulation. Further studies will be needed to assess whether PKA plays a direct role in regulating TWK-1 in the meiotic maturation process.

Sacy genes might act cumulatively downstream of somatic cAMP signaling

Sacy mutations enable partial fertility in the absence of somatic adenylate cyclase signaling, but in no case is fertility restored to wild-type levels. Possibly, Sacy genes might define cumulatively acting pathways that enable somatic control of meiotic maturation to ensure optimal rates of progeny production and sperm utilization. To begin to address this issue, we asked whether mutations in or might increase brood sizes in combination with a strong-loss-of-function mutation in . This analysis is complicated by the fact that the strongest loss-of-function mutations in both and confer pleiotropic phenotypes that can negatively impact fertility. Nonetheless, we observed that the weak mutations, and , further increased brood sizes in the ; genetic background (Table 5). While not definitive, these results are consistent with the idea that brood size is a complex function of multiple pathways downstream of . In the case of and , genetic and molecular analyses are consistent with the idea the two pathways might act in combination, but that neither is essential for reproduction. By contrast, genetic analysis of suggests that it functions as a major regulator of oogenesis and a strong negative regulator of meiotic maturation, possibly representing a downstream integrator of the upstream signaling pathways.

Table 5 

Evidence for cumulative action of Sacy mutations in the suppression of acy-4(lf) sterility

Genotype	Brood size (± SD)	Number of animals scored
acy-4(ok1806)	1 (± 2)	30
spr-5(by134); acy-4(ok1806)	81 (± 35)	40
twk-1(tn1403); acy-4(ok1806)	32 (± 25)	36
sacy-1(tn1440); acy-4(ok1806)	20 (± 19)	39
twk-1(tn1403) spr-5(by134); acy-4(ok1806)	128* (± 46)	40
sacy-1(tn1440) spr-5(by134); acy-4(ok1806)	108* (± 52)	40

P < 0.01 compared to spr-5(by134); acy-4(ok1806) using Student’s t-test.

Discussion

Suppressor genetics and the somatic control of oocyte meiotic maturation

Oocyte meiotic maturation is a conserved biological process required for sexual reproduction of animals. For the most part, our understanding in this area largely derives from biochemical and cell biological studies. For example, classical biochemical studies in amphibian oocytes led to the discovery of maturation promoting factor (Masui and Markert 1971). By contrast, comparably fewer forward genetic approaches have been undertaken. Historically, research studies on animal oocytes have benefited greatly from a diversity of experimental systems, including organisms less amenable to forward genetic analyses. Here we have taken the approach of screening for mutations that impact the regulation of the meiotic maturation process in C. elegans.

A tried-and-true approach in developmental genetics is to screen for mutations in which the developmental process of interest is disrupted—here oocyte meiotic maturation. Thus far, comparably few single gene mutations have been described that result in a block in oocyte meiotic maturation. Any mutation that completely feminizes the adult hermaphrodite gonad significantly blocks the process because of the absence of the MSP signal for meiotic maturation (McCarter ; Miller ). By contrast, strong loss-of-function mutations in block meiotic maturation despite the presence of sperm (Govindan ). The relatively few single gene mutations found to date that completely block meiotic maturation likely stems from a combination of factors. Meiotic maturation occurs in the adult stage and depends on signaling and cell cycle factors, such as , , or , which play earlier developmental roles (Boxem ; Govindan , 2009). Further, genetic redundancy, as observed for and (Detwiler ), might contribute to the robustness of the genetic network. Here we have taken advantage of epistasis in genetic pathways to isolate sterility suppressor mutations. Two technological advances enabled this approach. The first is the wide availability of deletion alleles isolated by the C. elegans Knockout Consortia (Moerman and Barstead 2008; S. Mitani, unpublished results). Many deletion alleles remove functional or catalytic domains in proteins, as does the deletion we used for screening. This consideration makes it less likely that intragenic or informational suppressors might be isolated. Further, the availability of deletion alleles of many genes facilitates gene identification and the analysis of strong loss-of-function alleles. A second enabling technology is whole-genome sequencing for mutant identification (Hillier ; Sarin ). In the current instance, this technology made it possible to determine the molecular identities of Sacy genes defined by multiple alleles, in the face of an inability to fine map many of the mutations by conventional means.

A salient feature of oocyte meiotic maturation signaling pathways in many organisms is that the somatic gonad exerts a controlling influence on the germ line. For example, it has long been known that removal of fully grown mammalian oocytes from antral follicles triggers meiotic resumption (Pincus and Enzmann 1935; Edwards 1965). Further, luteinizing hormone promotes meiotic resumption through activation of Gαs–adenylate cyclase–PKA signaling in mural granulosa cells, thereby regulating a cascade of paracrine and gap-junctional signaling pathways, which ultimately results in lowering intra-oocyte cAMP levels to trigger MPF activation (reviewed by Sun ; Downs 2010). In mammals, somatic gonadal control of meiotic maturation, in part, provides a means to link reproduction to pituitary hormone control. By contrast, in C. elegans, control by the somatic gonad provides a means to link oocyte meiotic maturation to sperm availability. In C. elegans, somatic Gαs–adenylate cyclase–PKA signaling is required for all described germline responses to the MSP hormone (Govindan ; this work; Figure 11). Further, Gαs–adenylate cyclase–PKA pathway activation can drive meiotic maturation at high rates in the absence of sperm (Govindan , 2009), approaching ∼50% of the maximal hermaphrodite rate, depending on the specific method of pathway activation.

Figure 11 

Model for the control of oocyte meiotic maturation in C. elegans. MSP signaling for oocyte meiotic maturation requires Gαs–ACY-4–PKA activity in the gonadal sheath cells. PDE-6 and TWK-1 may function in the gonadal sheath cells as negative regulators of meiotic maturation. The gonadal sheath cells inhibit meiotic maturation in part via gap-junctional communication involving the innexins INX-8 and INX-9 in the gonadal sheath cells (T. Starich and D. Greenstein, unpublished data) and INX-14 and INX-22 in oocytes (Govindan ). SACY-1 is a strong negative regulator of meiotic maturation that functions in the germ line upstream of, or in parallel to, the positive regulators OMA-1 and OMA-2. CoREST-like complex has a function in the germ line that is needed for the dependence of meiotic maturation on the Gαs–ACY-4–PKA sheath cell pathway. For illustrative purposes, TWK-1 and sheath cell MSP binding activity are displayed in a localized fashion, though this is unlikely to be the case (Govidan et al. 2009; this work).

In this work, we used suppressor genetics to ask two questions: (1) What are the molecular pathways that impose and mediate somatic control of meiotic maturation? (2) What happens to reproduction when signaling is perturbed? We identified mutations in at least 17 Sacy genes that enable partial fertility in the absence of somatic adenylate cyclase signaling. Cumulatively, we molecularly characterized mutations in 10 Sacy genes (, , , , , and ). In no case was fertility restored to wild-type levels, however. Given the scale of the suppressor screen, it seems doubtful that any viable single gene mutation can fully restore the fertility of hermaphrodites to wild-type levels. We identified two classes of mutations. The first class of suppressor is exemplified by , which encodes a highly conserved DEAD-box helicase. While we recovered reduction-of-function alleles as suppressor mutations, our screens could not recover strong loss-of-function alleles because of its essential functions needed for gamete maintenance and embryogenesis. SACY-1 is a strong negative regulator of oocyte meiotic maturation in the absence of sperm, and it likely functions upstream of OMA-1 and OMA-2 (Figure 11), which are redundantly required for meiotic maturation (Detwiler ). In contrast to SACY-1, the second suppressor class, exemplified by mutations affecting a CoREST-like complex and the TWK-1 two-pore potassium channel, is not absolutely required for meiotic maturation or its negative regulation in the absence of sperm (Figure 11). These genes and perhaps other Sacy loci appear to function cumulatively to enable somatic control of meiotic maturation.

Control of transcription and translation in the germ line and the regulation of meiotic maturation

Our findings suggest that CoREST-like chromatin regulators function in the germ line to ensure the somatic control of oocyte meiotic maturation. One potential model is that a CoREST-like complex prevents the germline expression of proteins that might interfere with the circuitry that establishes or maintains the dependence of meiotic maturation on somatic Gαs–adenylate cyclase–PKA and MSP signaling. In this model, the role of CoREST in oocyte meiotic maturation would be reminiscent of the synthetic multivulval (SynMuv) genes in vulval development (Fay and Yochem 2007). The SynMuv genes establish a necessary precondition for vulval induction by repressing ectopic transcription of , which encodes the anchor cell signal (Cui ; Saffer ). SynMuvB genes define a chromatin regulatory pathway involving the Rb retinoblastoma ortholog LIN-35 (Lu and Horvitz 1998). Interestingly, and display a synthetic genetic interaction affecting gonadogenesis (Bender ), suggesting that CoREST may mediate multiple functions needed for optimal germline development and reproduction. These functions might involve the regulation of transcription in the germ line.

By contrast, SACY-1 might function in post-transcriptional gene regulation important for the control of oocyte meiotic maturation, the hermaphrodite sperm-to-oocyte switch, and gamete maintenance. Genetic epistasis analysis suggests that functions upstream of and . Since oocytes are transcriptionally quiescent (Starck 1977; Gibert ; Schisa ; Walker ), and OMA-1 and OMA-2 are cytoplasmically localized (Detwiler ), an attractive hypothesis is that these two TIS11 zinc-finger proteins regulate meiotic maturation at a post-transcriptional level. Indeed, OMA-1 and OMA-2 have been shown to repress the translation of several mRNAs in oocytes and embryos (Jadhav ; Li ; Guven-Ozkan ). Our finding that SACY-1 functions as a component of the hermaphrodite sperm-to-oocyte switch likely upstream of is consistent with its potential involvement in post-transcriptional gene regulation (reviewed by Thomas ). Both SACY-1 and its Drosophila ortholog Abstrakt are found in the nucleus and cytoplasm (Irion and Leptin 1999), and thus these factors might function at a variety of levels to impact gene expression. We suggest that SACY-1 is a positive factor for TRA-2 expression, perhaps playing a similar role as Abstrakt, which promotes expression of the Inscutable protein (Irion ). If SACY-1 functions in a similar manner to regulate meiotic maturation, then it might function in part by promoting the translation of an inhibitory factor that restrains cell cycle progression. Interestingly, DDX41, the human ortholog of SACY-1, was recently found to be one of five genes recurrently mutated in patients with relapsing acute myeloid leukemia (Ding ). DDX41 also functions in a signaling pathway that detects invading viral double-stranded DNA in the cytoplasm and initiates an antiviral response (Zhang ), but other components of this innate immunity pathway appear not to be conserved in nematodes. Intriguingly, is required to prevent the necrotic cell death of gametes. Therefore, establishes a mechanistic link among three developmental processes critical for sexual reproduction: germline sex determination, somatic control of meiotic maturation, and preservation of gamete quality. The molecular genetic tools described in this work will facilitate the dissection of these key reproductive processes.