Background: Oocyte maturation is an important prerequisite for the production of progeny. Although several germ-line mutations have been reported, the precise mechanism by which the last step of oocyte maturation is controlled remains unclear. In Caenorhabditis elegans, CCCH-type zinc-finger proteins have been shown to be involved in germ cell formation, although their involvement in oocyte maturation has not been fully investigated.
Results: Using a multiple RNAi technique, we have identified three novel redundant CCCH-type zinc-finger genes, named by us moe-1, -2 (oma-1, -2) and moe-3, as a group related by functions and nucleotide sequence. Although a single RNAi of each moe gene was not effective, double or triple RNAi induced defects in oocyte maturation. We found that each moe transcript was expressed from the distal to proximal region of the gonad, while their corresponding proteins are accumulated exclusively in proximal oocytes, with a close association to germ granules. Although MOE-2 protein is rapidly removed from germ granules after fertilization, we found that MOE-2 associates with the centrosome-peripheral structure in dividing blastomeres.
Conclusions: Our results suggest that moe gene products are unique multifunctional proteins in terms of their redundancy and characteristic behaviour during the course of oocyte maturation. These gene products participate in processes in the final step of the meiotic cell cycle control, a novel function for CCCH-type zinc-finger family proteins thus far discovered.
The establishment of a germ cell line is one of the most important events for all living organisms to ensure descendants. In Caenorhabditis elegans gonads, primordial germ cells increase in number by rounds of mitotic division on the distal sides of the gonads and then enter the meiotic cell cycle (Hirsh et al. 1976; Austin & Kimble 1987). On the proximal sides of the gonads, germ cells continue to grow and form large female cells called oocytes (McCarter et al. 1999; Hall et al. 1999). During the course of meiosis, oocytes of various species accumulate large amounts of various RNAs and proteins in their cytoplasm in preparation for subsequent developmental processes. The final step of oocyte maturation is stimulated by a variety of meiosis-inducing substances (MIS), depending on the species, such as progesterone in the frog (Masui 1967) and 1-methyladenine in the starfish (Kanatani 1969), and oocytes finally acquire full competence for fertilization. It has recently been reported that a sperm cytoskeletal protein, called MSP, stimulates the final step of oocyte maturation in C. elegans (Miller et al. 2001). The oocyte is arrested at the late meiosis stage, and MSP promotes the resumption of meiosis and gonadial sheath contraction to induce ovulation. The fully matured oocytes complete fertilization in the spermatheca and progress to the embryogenesis stage in the uterus. Although several germ-line mutations have been reported, the precise mechanism by which the last step of oocyte maturation is controlled remains unclear.
During the course of C. elegans oogenesis, germ cell-specific granules, called P granules, migrate from the circumferential region of the nucleus to the cytoplasm (Pitt et al. 2000). P granules contain various proteins and RNA species and are thought to be crucial for germ-line specification (Kawasaki et al. 1998; Seydoux & Strome 1999; Schisa et al. 2001). In fertilized embryos, P granules migrate to the posterior region of the eggs and are specifically inherited by cells of the germ-line lineage (Strome & Wood 1983; Hird et al. 1996). It is known that P granules contain several CCCH-type zinc-finger proteins that posses the characteristic Cystein-X8–10-Cystein-X5-Cystein-X3-Histidine sequence of the CCCH-type zinc-finger domain (Bai & Tolias 1996; Clarke & Berg 1998). Although some CCCH-type zinc-finger proteins posses RNA-binding activity (Bai & Tolias 1996), the precise role of this motif is currently unknown.
Several CCCH-type zinc-finger proteins, including PIE-1, MEX-1, MEX-5/6 and POS-1, have been identified in genetic studies as maternal factors that are important for the differentiation of germ cells in C. elegans (Mello et al. 1996; Guedes & Priess 1997; Tabara et al. 1999; Reese et al. 2000; Schubert et al. 2000; Tenenhaus et al. 2001). They are essentially cytoplasmic proteins and are thought to regulate RNA metabolism. It has been reported that mouse CCCH-type zinc-finger proteins bind directly to an AU-rich sequence in the 3′UTR region, resulting in destabilization of the targeted mRNA (Lai et al. 1999; Lai & Blackshear 2001), although the exact roles of these motifs are not yet known. Besides these genetically identified proteins, there are many as-yet uncharacterized CCCH-type zinc-finger proteins in the C. elegans genome. The functions and significance of these remaining CCCH-type zinc-finger proteins have not been elucidated.
Comprehensive gene knockout projects have been carried out in C. elegans (Fraser et al. 2000; Gonczy et al. 2000; Maeda et al. 2001). It has been found that 147 genes in chromosome I are essential for proper development (Zipperlen et al. 2001). Interestingly, some of these genes have been reported to have defects at early developmental stages. However, the functions of the remaining ‘silent’ genes have not been well characterized, probably as a consequence of gene redundancy.
In the present study we identified three novel redundant CCCH-type zinc-finger genes as a group related by functions and nucleotide sequence. We found that they have overlapping functions that are crucial for oocyte maturation and germ granule distribution. Thus, our approach has revealed multiple and redundant functions of a new class of CCCH-type zinc-finger proteins, call MOE family proteins. MOE family proteins are unique CCCH-type zinc-finger proteins in terms of their redundancy, multiple functions and characteristic behaviour during the course of oogenesis.
Screening novel CCCH-type genes redundancy for in generating progeny
Using C. elegans genome sequence information, we have cloned 17 CCCH-type zinc-finger protein genes from a C. elegans early embryonic total RNA and cDNA library. We constructed a phylogenetic tree from their amino acid sequences (Fig. 1A) and classified them into seven subgroups. It has been shown that several CCCH-type zinc-finger proteins, including PIE-1, MEX-1, POS-1 and MEX-5/MEX-6, are crucial proteins for proper germ-line determination in C. elegans (Mello et al. 1996; Guedes & Priess 1997; Tenenhaus et al. 1998; Tabara et al. 1999; Schubert et al. 2000). We confirmed this in RNAi experiments, one of the most powerful methods for reverse genetic analysis in C. elegans (Fire et al. 1998). We also found that inhibition of the remaining CCCH-type zinc-finger gene products does not cause any obvious defects in germ-line formation and subsequent developmental processes (data not shown).
Because some genes encode highly related sequences (Fig. 1A), we speculated that these genes may posses overlapping functions. To classify the redundancy of the CCCH-type zinc-finger protein family, we then attempted to simultaneously inhibit several CCCH-type zinc-finger genes, based on the classification shown in Fig. 1A. Using a unique sequence as a target for RNAi, we aimed to identify the function of each CCCH-type zinc-finger protein subfamily. We found that the inhibition of one related group of CCCH-type zinc-finger proteins, C09G9.6, ZC513.6 and F32A11.6 (Fig. 1A), caused a large reduction in the number of laid eggs (Fig. 1B). Double RNAi treatments of these genes (C09G9.6 and ZC513.6, designated as 1+2 in Fig. 1B) resulted in a more than 60% reduction in the number of laid eggs. Intriguingly, we found that triple RNAi treatment (C09G9.6, ZC513.6 and F32A11.6, designated as 1+2+3 in Fig. 1B) resulted in an even greater reduction in the number of eggs laid, i.e. to more than 85% reduction in the number of laid eggs compared to the control. These results suggested that F32A11.6 also collaborates additionally with C09G9.6 and ZC513.6 function. We obtained the same results using a method of bacteria-mediated RNAi (feeding method; Timmons et al. 2001), indicating that the phenotypes were not derived from microinjection artefacts. RNAi of other combinations of CCCH-type zinc-finger genes (such as the combination of F38C2.5 and Y57G11c.25, which are themselves nonessential) had no obvious effects (data not shown). These results indicate that the newly identified genes, C09G9.6, ZC513.6 and F32A11.6, possess largely redundant but collectively essential functions for generating progeny in C. elegans. For reasons described below, we refer to C09G9.6, ZC513.6 and F32A11.6 as moe-1, moe-2 and moe-3.
MOE-1 and MOE-2 play essential roles in oocyte maturation
To determine what kind of defect leads to a reduction in the number of laid eggs, we first examined the injected parental P0 phenotypes in detail. In the wild-type, germ cells proliferated by repeated rounds of mitosis on the distal side of the gonad and then progressed into meiosis and moved proximally, producing maturing oocytes on the proximal side. We found that the proximal gonad arm expands greatly in C09G9.6/ZC513.6-inhibited individuals (Fig. 2Aa) compared with the proximal arms in wild-type and control-injected gonads (Fig. 2Ag,h). Double RNAi against ZC513.6/F32A11.6 (Fig. 2Ac) and C09G9.6/F32A11.6 (Fig. 2Ae) and single RNAi against these three genes (Fig. 2Ab,d,f) induced no abnormalities in the gonad. In the expanded gonad of C09G9.6/ZC513.6-inhibited individuals, proximal oocytes were greatly enlarged (Fig. 2Ba). The enlarged oocytes contained abnormally expanded germinal vesicles (Fig. 2Ba, indicated by an arrow) compared to those in normal oocytes (Fig. 2Bb,c,d, indicated by an arrowhead). The triple RNAi of C09G9.6, ZC513.6 and F32A11.6 show more penetration but essentially the same phenotype as that of C09G9.6/ZC513.6-double-compromised individuals (data not shown). These results suggest that abnormality in progeny in C09G9.6, ZC513.6 and F32A11.6-compromised individuals is derived from inappropriate oocyte expansion. From these characteristic phenotypes, we named the C09G9.6, ZC513.6 and F32A11.6 genes maturating oocyte expansion (moe) family genes: moe-1, moe-2 and moe-3, respectively. Similar results have recently been reported by Detwiler et al. (2001), who named C09G9.6 and ZC513.6 as oma-1 and oma-2, respectively, but there was no mention at all in their report about moe-3.
MOE-1 and MOE-2 have about a 63% amino acid similarity, and MOE-2 and MOE-3 have about a 45% amino acid similarity (Fig. 3A). Each of them contains two copies of the conserved CCCH-type zinc-finger domain (Fig. 3B).
As mentioned above, moe-1/moe-2-inhibited gonads contain greatly expanded oocytes, and the germinal vesicles of proximal oocytes in the gonads are also enlarged, while distal oocytes in the pachytene stage and mitotic germ cells appeared normal. In moe-1/moe-2-suppressed oocytes, germinal vesicle breakdown was never observed. DNA staining of RNAi-treated gonads revealed that chromosome condensation and progression of meiosis I to the diakinesis stage occurs normally (Fig. 4A). However, after the diakinesis stage, the distribution of the six condensed chromosomes appears highly dispersed in enlarged germinal vesicles (Fig. 4Ab, arrowhead) compared to that in buffer-injected control oocytes (Fig. 4Aa). We next examined the tubulin distribution in moe-1/moe-2-inhibited oocytes. About 40% (n = 18) of the moe-1/moe-2-inhibited oocytes showed substantially reduced α-tubulin staining (Fig. 4Bb) compared to the controls (Fig. 4Ba), suggesting that the defect in oocyte maturation is mediated, at least in part, by an abnormal microtuble network during the maturation step.
MOE-1 and MOE-2 are specifically co-expressed in proximal oocytes
In the C. elegans gonad, it is known that MSP (derived from sperm) influences the final step of oocyte maturation (Miller et al. 2001) and that a defect in the spermatheca leads to an abnormality in maturing oocytes. Although the phenotype of moe-1/moe-2 inhibition is apparent in proximal oocytes, it is not clear whether the defect is due to an abnormality in the oocyte itself or not.
Therefore, to determine whether the maturation defects in moe gene-compromised individuals are attributed to an abnormality in the oocyte itself, we performed in situ hybridization analysis using moe-1-, moe-2- and moe-3-specific probes to verify the expression patterns of moe gene products. The results of the in situ experiments revealed that all three genes are specifically and exclusively expressed in the female gonad (Fig. 5A,C,E). It is important to note that there is essentially no signal in the spermatheca. These observations indicate that moe gene products function at the female germ cells. Although gonadal staining is not absolutely quantitative, it was evident that moe-1 and moe-2 gene transcripts are widely distributed throughout gonadal oocytes from the mitotic stage to the developing diakinesis stage (Fig. 5A,C), whereas the expression of moe-3 is relatively weak and is restricted to the distal region of the gonad (Fig. 5E).
To further examine the expression patterns of these gene products, we produced specific antisera to MOE-1 and MOE-2. Immunocytochemical staining using anti-MOE-1- and anti-MOE-2-specific antibodies revealed cytoplasmic expression in oocytes in the proximal region of the gonads (Fig. 6A,C). The distribution of MOE-1 protein in proximal oocytes corresponds to that of the MOE-2 protein (Fig. 6Ac, merged pattern), supporting the speculation that these proteins share overlapping functions. Both the MOE-1 and MOE-2 proteins accumulated heavily in the most-proximal oocyte (Fig. 6A,C). The oocyte-specific expression indicates that the reduction in the number of laid eggs by moe-1/moe-2 double RNAi is caused by a defect in the oocyte itself and that the cytoplasmic accumulation of MOE-1 and MOE-2 proteins in maturing oocytes is required for proper oocyte maturation. An interesting finding is the characteristic timing of MOE protein appearance during the course of oogenesis. Despite the fact that mRNAs of the MOE family are expressed from the most-distal region of the female gonads (Fig. 5), the corresponding proteins start to accumulate only after the diakinesis stage (Fig. 6A,C). This finding indicates the existence of post-transcriptional regulation of moe mRNAs translation.
The immunological signals of MOE-1 (Fig. 6Ca) and MOE-2 (not shown) were completely suppressed by absorption using corresponding antigens (Fig. 6Cb). Antigen competition of the anti-MOE-1 antibody reduced the intensity of the MOE-1 signal, but it did not influence the MOE-2 signal, and vice versa for the anti-MOE-2 antibody. This observation eliminates the possibility of a cross-reaction between secondary antibodies in double-staining experiments. It was also confirmed that the RNAi treatment of MOE-1 and MOE-2 specifically eliminates the corresponding immunological signals from oocytes, clearly demonstrating that the signals we observed were derived from MOE gene products.
MOE-1 and MOE-2 are novel components of P granules and regulate P granules distribution in growing oocytes
Since MOE genes encode typical CCCH-type zinc-finger proteins similar to PIE-1, MEX-1 and POS-1 (Fig. 3B), we were interested in determining whether MOE proteins also associate with P granules in C. elegans oocytes. Double-staining experiments using anti-MOE-1 and anti-P granules (Strome & Wood 1983) antibodies revealed that MOE-1 was also localized on P granules in growing oocytes (Fig. 6Ba,b,c), as well as in the cytoplasm. Similar results were obtained using an anti-MOE-2 antibody (Fig. 6Bd,e,f).
The MOE-1 and MOE-2 signals on P granules were eliminated by absorption using their corresponding antigens, indicating that there was no cross-reaction of the secondary antibodies (Fig. 6C). Furthermore, RNAi treatment of moe-1 and moe-2 completely eliminated the fluorescence signal on P granules from anti-MOE-1/2 antibodies. The results of these control experiments rule out the possibility that the staining of P granules by anti-MOE antibodies is merely due to a cross-reaction or non-specific reactions of the antibodies used. Interestingly, MOE-1 and MOE-2 were found to prominently associate with P granules in oocytes in the middle to proximal region of the gonad, although P granules themselves are present from the most primordial stage of female germ cell formation (Fig. 6C).
Because MOE-1 and MOE-2 proteins have been shown to be associated with P granules in the developing oocytes, we next investigated the effect of MOE protein removal on the distribution of P granules. In wild-type gonads, P granules are closely associated with the periphery of the nucleus and germinal vesicle in female germ cells from the distal to proximal region of the gonad, except for some of the most proximal maturing oocytes (Fig. 7B). In contrast, in the case of moe-1/moe-2-suppressed gonads, we found that P granules (K76 antigen) in oocytes in the middle to proximal region of the gonad rarely associated with germinal vesicles and were distributed obscurely (Fig. 7C). It should be noted that the distribution of P granules in mitotic germ cells (in the distal region of the gonad), even in moe-1/moe-2 suppressed gonads, appears normal. An abnormality in the distribution of P granules in moe-1/moe-2-suppressed gonads became apparent from the post-mitotic stage. This timing seems to be correlated with the timing of MOE-1 and MOE-2 protein expression in wild-type gonads (Fig. 6C). Thus, it seems that the expression of MOE-1 and MOE-2 proteins are required for the proper assembly of P granules and/or control of their distribution in developing meiotic-stage oocytes.
MOE-2 protein decrease from P granules after fertilization
As mentioned previously, MOE-2 protein associates with P granules in maturing oocytes. To determine the consequence of the expression of MOE-2 protein in post-maturation stages, we examined MOE-2 protein expression after fertilization. Interestingly, we found that the amount of MOE-2 protein in oocytes decreases to undetectable levels through several cell divisions. Although a strong immunosignal of anti-MOE-2 antibody was observed in fertilized eggs (Fig. 8Aa, arrowhead), signals of greatly reduced intensity were observed at the cleavage stage (Fig. 8Aa, arrows). The reduction in the intensity of MOE-2 immunostaining was not due to a change in the antibody permeability, because a simultaneous staining of anti-P granules antibody (Fig. 8Ab) equally stained all embryos without any reduction. MOE-2 signals on P granules were observed just after fertilization (Fig. 8Ba,b,c), but the intensity of the signals decreased to an undetectable level in cleaving-stage embryos (Fig. 8Bd,e,f, and our unpublished observations). There were essentially no MOE-2 signals on P granules in embryos after the two-cell stage (Fig. 8Aa,b, arrows). These observations suggest that the association of MOE-2 with P granules is a transient event, only detectable from diakinesis-stage oocytes to fertilized eggs.
We found that the reduction in the amount of MOE-2 protein after fertilization is dependent on the 26S proteasome pathway. 26S proteasome is a cytoplasmic proteinase complex that plays a pivotal role in ubiquitin-dependent proteolysis (Kawahara et al. 2000a,b). 26S proteasome is composed of about 30 independent subunits, and one of these subunits, Rpn1, has been shown to be essential for proteasome function. As shown in Fig. 9C, suppression of proteasome activity by the RNAi of the rpn1 subunit gene completely abolished the reduction in the amount of MOE-2 protein after fertilization. In proteasome-suppressed embryos, cell division was eventually blocked, resulting in the formation of several large and polyploid nuclei (Fig. 9D) (Kawahara et al. 2000a), and there was no sign of reduction in the amount of MOE-2 protein (Fig. 9C). This is in contrast to the results for normal fertilized embryos, showing that the amount of MOE-2 protein was reduced in cleaving-stage embryos (Fig. 9A,B; right embryo) compared to that in fertilized eggs (Fig. 9A,B; left embryo). These results indicate that the reduction in the amount of MOE-2 protein after fertilization is mediated directly or indirectly by a 26S proteasome-dependent protein degradation pathway.
MOE-2 protein associates with the centrosome-peripheral structure in dividing blastomeres
Although MOE-2 protein is removed from P granules after fertilization, we found that MOE-2 closely associates with the centrosome-peripheral structure after the two-cell stage (Fig. 10Aa,d). This association was observed not only at the early cleavage stage (Fig. 10Aa,d) but also at later blastula stages (Fig. 10Ba). Disruption of the mitotic apparatus by nocodazole treatment resulted in the disappearance of specific staining (Fig. 10Aj), indicating that MOE-2 localization is dependent on mitotic spindle formation. The MOE-2 signal on the spindle was completely absorbed by the corresponding antigenic peptides and was not detected by pre-immune serum (Fig. 10Ag, and our unpublished observation). Furthermore, RNAi treatment of moe-2 largely eliminated the anti-MOE-2 signal on the centrosome-peripheral structure (Fig. 10Bc) compared to that in the wild-type (Fig. 10Ba). The results of these control experiments further support the unique distribution of MOE-2 protein in dividing blastomeres. Since the mitotic spindle is important for the control of cell division (Keating & White 1998), this finding suggests that MOE-2 is partly involved in control of mitotic spindle and microtuble organization in C. elegans embryogenesis, although suppression of moe-2 alone did not produce severe defects in embryogenesis (our unpublished observations).
Searching for hidden functions of CCCH genes
It has been shown, using genetic approaches such as random mutagenesis and RNAi experiments, that several CCCH-type zinc-finger genes have pivotal functions for germ-line determination (Mello et al. 1996; Guedes & Priess 1997; Tenenhaus et al. 1998; Tabara et al. 1999; Reese et al. 2000; Schubert et al. 2000). On the other hand, many genes have been shown to be non-essential genes; loss of function of these genes does not induce any obvious defects throughout the course of development of C. elegans. This may be, at least in part, due to gene redundancy in the C. elegans genome. We have determined the functional significance of the closely related CCCH-type ‘zinc-finger’ gene family. In this study, we classified 17 CCCH-type zinc-finger genes into seven subgroups, and related genes were simultaneously inhibited using a multiple RNAi technique. As a result, we have successfully identified a novel gene family, designated the moe family, as key regulatory components of C. elegans oocyte maturation.
moe family genes are essential for oocyte maturation
Prior to fertilization and development, female germ cells in C. elegans undergo repeated rounds of mitosis and then enter the meiotic cell cycle. During meiotic maturation, oocytes grow progressively, accumulate maternal stocks of various kinds of proteins and RNAs, and acquire full competence for later fertilization and developmental processes. Despite the vital importance of the maturation step, oocyte maturation-specific mutants in C. elegans have not been found.
In this study we found that novel CCCH-type zinc-finger family gene products, the MOE family, are required for the final step of oocyte maturation in C. elegans oogenesis. MOE-1 and MOE-2-compromised individuals show the following characteristic sterile phenotypes: (i) drastic reduction in the number of laid eggs, (ii) characteristic expansion of the gonad and oocytes, (iii) expansion of the germinal nucleus and characteristic scattering of condensed chromosomes, (iv) abnormal structure of the microtuble network, and (v) abnormal distribution of P granules in the developing oocyte. In addition, triple RNAi of moe-1/-2/-3 showed further extended phenotypes. Although inhibition of moe-3 alone or double inhibition of moe-3 with either moe-1 or moe-2 genes showed a weak phenotype, if any, triple inhibition of moe-1, moe-2 and moe-3 induced a more extended phenotype than in the case of moe-1 and moe-2 double inhibition (Fig. 1B). This observation indicates that moe-3 also possesses a function that overlaps those of moe-1 and moe-2.
The morphology of early stage oocytes, from the mitotic stage to the diakinesis stages, appear completely normal in moe-1/moe-2-inhibited individuals, indicating that moe family genes are required for maturation completion but not for initiation of maturation. These phenotypes have obvious differences compared with previously described germ-line-defect mutants of C. elegans. For example, oocytes in the emo-1 mutant become endomitotic in the gonad arm (Iwasaki et al. 1996), while mutations in glp-1 give rise to germ cells that would normally divide mitotically to enter meiosis (Austin & Kimble 1987). Thus, MOE family proteins are novel regulators of the late maturation step and are specifically expressed in growing oocytes.
The only known oocyte maturation defect is sperm-defective mutation (Kimble et al. 1984; Schedl & Kimble 1988; McCarter et al. 1999; Miller et al. 2001) and the recently described oma genes mutation (Detwiler et al. 2001). The expression patterns of oma gene transcripts have not been reported previously. In this paper we have provided evidence that the distribution of moe transcripts and their translated products are not identical. The results of our in situ hybridization analysis clearly showed, for the first time, that the expression of MOE transcripts is specific to the female germ-line, from mitotic stage to maturation stage (Fig. 5). We found that the expression of both moe-1 and moe-2 transcripts preceded their corresponding protein expression. As shown in Fig. 5, strong staining of both moe-1 and moe-2 mRNA was observed from the distal to proximal portion of the gonad, while their proteins were found to be specifically accumulated in growing oocytes in the proximal portion of the gonad (Fig. 6A,C). In addition, MOE-1 protein began to accumulate prior to MOE-2 accumulation (Fig. 6A), despite the fact that moe-2 mRNA are expressed earlier to its protein expression. These observations indicate that there is post-transcriptional regulation of MOE-1 and MOE-2 protein expression during the course of oocyte growth. We speculate that mRNAs for moe-1, -2 and -3 begin to accumulated from the mitotic stage in preparation for later processes, but that their translations are completely suppressed until the oocytes reach the meiotic cell cycle stage. Alternatively, MOE proteins might be unstable in mitotic-stage female germ cells. In any case, it is highly probable that there is a post-transcriptional control of MOE protein expression, and that such a control mechanism might be necessary for the proper regulation of meiotic cell cycle progression in C. elegans oocytes.
MOE family products have not been detected in the spermatheca. It has been reported that sperm-defective strains spawn unfertilized oocytes (McCarter et al. 1999), but this phenomenon has never been observed in MOE-inhibited individuals. In addition, moe oocytes could not be cross-fertilized by wild-type males. These findings support the view that moe gene products are essential for oocyte function, but do not play a role in sperm-related function. We confirmed that the triple-suppression of myt-1 with moe-1 and moe-2 RNAi results in ovulation, as was reported by Detwiler et al. (2001). Because MYT-1 is a suppressor kinase for meiotic CDC2 kinase activation, this result implies that both MOE-1 and MOE-2 work, either directly or indirectly, as suppressors of the MYT-1 pathway and that moe-1/moe-2-inhibited oocytes are arrested at the meiotic prophase. Some of the CCCH-type zinc-finger proteins are known to regulate RNA metabolism (Bai & Tolias 1996; Lai et al. 1999; Batchelder et al. 1999). Although the significance of the presence of MOE proteins in the cytoplasm of oocytes is not known, it is possible that MOE family proteins are necessary to suppress the translation of some maternally encoded mRNAs (Evans et al. 1994), such as transcripts for suppressor protein(s) of the MYT-1 pathway, until fertilization occurs.
MOE-2 protein is a component of P granules that disappears after fertilization
The previously described CCCH-type zinc-finger proteins PIE-1, MEX-1 and POS-1, are required for proper germ-line formation and are localized in germ cell-specific granules, known as P granules (Mello et al. 1996; Guedes & Priess 1997; Tabara et al. 1999). We found that MOE-1 and MOE-2 are associated with P granules in developing oocytes (Fig. 6B). Since P granules are thought to play a role in germ-line determination, our results suggest that MOE proteins play a role in the function of P granules. However, we found that the MOE-2 signal on P granules does not seem to be inherited in the developing blastomere. MOE-2 continued to associates with P granules from diakinesis-stage oocytes to pronuclear-stage fertilized eggs (Fig. 8B), its expression level thereafter decreased to an undetectable level (Fig. 8B and our unpublished observations). These observations suggest that the association of MOE-2 with P granules is a transient event, and that MOE-2 might have a specialized function for P granules, such as control of proper assembly and/or distribution of P granules during the final maturation stage of oocytes. In good agreement with this hypothesis, we found the abnormal P granule distribution in moe-1/moe-2-suppressed oocytes. Our interesting observation is that both the appearance of MOE proteins in wild-type oogenesis (Fig. 6C) and the abnormality in P granule distribution in moe-1/moe-2-suppressed oocytes (Fig. 7) begin at post-mitotic stage female germ cells. This apparent correlation implies possibly important roles for MOE-proteins in P granule regulation at oogenesis stage.
Although P granule-associated MOE-2 protein is removed after fertilization, we found that MOE-2 closely associates with the centrosome-peripheral structure (Fig. 10Aa,d). This association became prominent in embryos from the two-cell stage to the late blastula stage. Since the mitotic spindle is important for control of cell division (Keating & White 1998), this finding suggests that MOE-2 is partly involved in control of mitotic spindle and microtuble organization in C. elegans embryogenesis. Furthermore, our observations indicate that moe-1/moe-2 suppression results in an improper formation of the microtuble network in maturing oocytes (Fig. 4B). These observations indicate that MOE family proteins regulate various aspect of the microtuble organization and thus influences events of proper oocyte maturation and possibly embryonic blastomere division.
We have identified three novel CCCH-type zinc-finger genes as a group related by function and by nucleotide sequence. We have also found that MOE products have redundant functions that are collectively crucial for oocyte maturation. MOE-2 protein was found to be accumulated in the cytoplasm of proximal oocytes and was removed after fertilization by the proteasome-mediated degradation pathway, suggesting that it participates in processes in the final step of the meiotic cell cycle, a novel function for CCCH-type zinc-finger family proteins thus far discovered. Furthermore, MOE-2 protein was found to be associated with P granules in maturing oocytes and with mitotic spindles in dividing embryos. These results suggest that MOE proteins play crucial roles in the regulation of P granules and in events that control the cell cycle. Determination of the relationships with known components of P granules, as well as with the factor(s) that regulate meiotic and mitotic spindle organization are challenges for future studies.
Cloning of CCCH-type zinc-finger genes
All CCCH-type zinc-finger proteins were searched on the basis of the PIE-1 sequence using NCBI protein databases of the C. elegans genomic sequence. Primers were designed on the basis of genomic information on each hypothetical CCCH-type zinc-finger gene, and RT-PCR was performed using the early embryo mRNA of C. elegans as templates. Cloned products were sequenced, and it was confirmed that they actually encode CCCH-type zinc-finger proteins. The cloned moe-1 sequence was identical to C09G9.6 of the cosmid sequence in C. elegans databases. moe-2 and moe-3 sequences were identical to ZC513.6 and F32A11.6. The ‘moe’ genes were registered in the CGC gene name database (supervised by J. Hodgkin at Oxford University, CGC genetic map and nomenclature curator).
RNA interference (RNAi)
RNAi experiments were performed using the following methods. moe-1 (521–878 bp), moe-2 (40–457 bp) and moe-3 (74–401 bp) and other cloned genes were amplified by PCR. The amplified RCR products were subcloned into the pCRII vector (Invitrogen), between T7 and SP6 polymerase sites. RNA was transcribed from plasmid DNA using MEGAscript (Ambion). The dsRNA was injected into the body cavity or gonads of wild-type young adult hermaphrodites. After a recovery period of about 5.5 h, the injected animals were transferred to new plates and allowed to lay for about 20.5 h at 20 °C. Laid eggs were allowed to develop for 24 h at 20 °C, and the number of laid eggs were counted using a phase contrast microscope.
In situ hybridization
To determine localizations of moe mRNAs, we performed whole-mount in situ hybridization as described by Kohara at <http://watson.genes.nig.ac.jp:8080/db/method/>. In situ hybridization was performed using specific probes that were used for the RNAi analysis described above. Sense and anti-sense RNA probes were labelled with digoxigenin-11-dUTP (Boehringer Mannheim). Hybridized probes were detected with an alkaline phosphatase (AP)-conjugated anti-DIG antibody and an AP-mediated colour detection system (Boehringer Mannheim).
The anti-MOE-1 and anti-MOE-2 antibodies were affinity-purified with corresponding antigens and used for immunocytochemistry. A detailed characterization of these antibodies will be described elsewhere. For immunocytochemistry of gonads, adult hermaphrodites were cut open to release the gonads, and freeze/cracked samples in M9 buffer were fixed by incubating in cold methanol followed by cold acetone.
An MOE-1 and MOE-2 double-staining experiment was performed as follows. First, anti-MOE-1 antibody at 33 nm was incubated with a secondary polyclonal antibody, Alexa™ 594-conjugated anti-rabbit IgG (Molecular Probes) at 16 nm. In parallel, anti-MOE-2 antibody at 17 nm was incubated with an Alexa 488-conjugated anti-rabbit IgG at 15 nm. To avoid the possible cross-reaction of secondary antibodies with each other, we first pre-incubated each of secondary ‘polyclonal’ antibodies with corresponding primary antibodies independently for 2 h before applying to the fixed samples. Fluorescence images were obtained using a Delta Vision confocal microscope system (Applied Precision Inc.).
Double staining experiments of P granules and MOE-1 or MOE-2, respectively, were performed using anti-MOE-1 antibody at 33 nm or anti-MOE-2 antibody at 34.7 nm and anti-P granules K76 antibody at 1 : 100 dilution (Hybridoma Bank). Alexa 488-conjugated anti-rabbit IgG antibody and Alexa 594-conjugated anti-mouse IgG antibody were used as secondary antibodies at 1 : 800 dilution. For antigen absorption experiments, an excessive amount of antigen was added to the primary antibody reaction solution; antigens for each antibodies were used at 125.2 µm (for anti-MOE-1) and 69.4 µm (for anti-MOE-2), respectively.
We sincerely thank Mrs T. Kobayashi for her excellent technical assistance. We are also grateful to Dr I. Kawasaki for providing us with an anti-P granules antibody. We thank Dr A. Sugimoto, Dr K. Ogura, Dr S. Takagi, Prof. I. Mori, Prof. M.J. Whitaker and Prof. H. Yokosawa for helpful comments and suggestions. The K76 monoclonal antibody developed by S. Strome and W. B. Wood was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD. This work was supported in part by a grant from the Ministry of Education, Culture, Sports, Science and Technology of Japan to H.K. (13043001).