INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Protein kinases play crucial roles in the regulation of a wide variety of cellular functions. A novel family of protein kinases, bearing a phosphospecific protein-protein interaction motif, the forkhead-associated (FHA) domain (11), has been identified and shown to be involved in checkpoint regulation and DNA repair induced by DNA damage (16). Protein kinases belonging to this family include Saccharomyces cerevisiae Rad53, Dun1, and Mek1p (MRE4), Schizosaccharomyces pombe Cds1 and Mek1, Drosophila melanogaster Dmnk, and mammalian Chk2 (5, 7, 19, 20, 23, 26, 36, 43). It has been reported that the yeast Rad53, Cds1, and Dun1 protein kinases are required for the S-phase checkpoint and for the activation of the DNA repair machinery upon DNA damage, although Dun1 is involved only in the latter process (2, 12, 23, 25, 36, 42, 43). In contrast, S. cerevisiae Mek1p is involved in the regulation of meiotic recombination (19).

It has recently been reported that, in response to DNA damage and DNA replicational stress, Chk2 (mammalian Cds1) is activated and phosphorylates Cdc25C, thereby inactivating Cdc25 phosphatase activity and preventing entry of cells into mitosis (5, 7, 8, 20, 39). More recently, Chk2 has been shown to stabilize p53 by phosphorylating p53 on serine 20, which interferes with Mdm2 binding (9, 14, 35). In contrast, D. melanogaster Dmnk, which is most closely related to Chk2, is highly expressed in ovaries and in germ cell nuclei during early embryogenesis, suggesting its possible function(s) in characteristic features of germ cells such as meiosis and/or germline establishment (26).

Caenorhabditis elegans is an excellent model organism in which to study meiosis, the cell cycle, and development. With the determination of the entire genomic sequence of C. elegans, this multicellular organism further provides a unique opportunity to study the role of this entire gene family during development. Specific and functional disruption of gene expression by utilizing double-stranded RNA (dsRNA)-mediated genetic interference (RNAi) enables us to address the biological consequence of reduction or elimination of the activity of a particular gene (4, 13, 34). To examine the function of Cds1- and Dmnk-related kinase in C. elegans and a possible functional conservation of this kinase family among different species, we identified cDNA clones encoding C. elegans homologs of Cds1, Chk2, and Dmnk, designated Ce-CDS-1 and Ce-CDS-2, and tested their function in this organism by RNAi. We show here that reduction or elimination of the function of Ce-CDS-1 and Ce-CDS-2 results in the inhibition of meiotic crossing over and the loss of chiasmata, apparently without affecting synapsis. Our observations indicate that Ce-CDS-1 and Ce-CDS-2 play a crucial role(s) in meiotic recombination in C. elegans. The functional comparison of this family of protein kinases among different species is discussed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Worm strains. Worms were cultured as described previously (6). Experiments were performed at 20°C. The strains utilized in this study are as follows: N2 (Bristol), wild-type strain; LG I, glp-4(bn2ts); LG III, unc-25(e156) dpy-18(e364); and LG X, unc-3(e151) dpy-6(e14). These strains are described in detail in the Caenorhabditis Genetic Center data releases.

cDNA cloning. The existence of C. elegans homologs of Dmnk was first inferred from the results of TBLASTN searches of the C. elegans genome database (Sanger Center). In this database, the gene designated Y60A3A.12, encoding a protein kinase (which we call Ce-CDS-1) that is quite similar to Dmnk, contains a single FHA domain and a canonical protein kinase domain. This gene also contains C. elegans expressed sequence tag clone yk523b3 (Yuji Kohara, National Institute of Genetics, Mishima, Japan). In the genome database, another gene, designated T08D2.7 (localized to T08D2, a floating cosmid without a physical map), may also encode a C. elegans homolog of Dmnk (which we call Ce-CDS-2) and exhibits >95% identity with Y60A3A.12. Thus, there may be two putative orthologs of Dmnk in C. elegans, although an expressed sequence tag clone for T08D2.7 has not been reported. To isolate a cDNA clone corresponding to Y60A3A.12 (which we call Ce-cds-1), total RNA was extracted from C. elegans (mixed stages of development), and single-stranded cDNA was synthesized and PCR amplified as described previously (26). The sequences of the Ce-cds-1-specific primers were as follows: 5' TTGAATTCCGGTCACCCGAGACGACA 3' and 5' CCGAATTCAAAACGCAATAAAATGGGGGGCT 3' (restriction sites for EcoRI are underlined). Amplified DNA fragments of the expected size were digested with EcoRI and cloned into the EcoRI sites of the Bluescript vector (pBS; Stratagene). Both strands of Ce-cds-1 cDNA, subcloned into pBS (pBS-Ce-cds-1), were completely sequenced. The 5' end of the Ce-cds-1 transcript was determined by reverse transcriptase-PCR using a specific primer to Ce-cds-1 (5' CCGAATTCAAAACGCAATAAAATGGGGGGCT 3') and the SL1 primer (5' GGTTTAATTACCCAAGTTTGA 3') containing the SL1 spliced leader sequence. The nucleotide sequences of PCR-amplified products were determined by using a Ce-cds-1 internal primer (5' CGCACACGAAATGATCGTCTG 3').

Northern blot analysis. Total RNAs from various stages of C. elegans development were prepared as previously described (37). For RNA blot analysis, 10 µg of total RNA was separated by 1% agarose formaldehyde gels and transferred onto nylon membranes. The probe DNA was prepared from pBS-Ce-cds-1 by digestion with HindIII and EcoRI, labeled with [-³²P]dCTP (Amersham; 3,000 Ci/mmol) using the Multiprime labeling kit (Amersham), and hybridized as described previously (27).

RNA interference experiments. The loss-of-function, possible null phenotype for the Ce-cds-1 and Ce-cds-2 genes was generated by using RNAi as described previously (13). pBluescript vectors containing various regions of Ce-cds-1 cDNA (as illustrated in Fig. 2A) were linearized by digestion with restriction endonucleases that were specific for sites within the multiple cloning site at either end of the cDNA. The DNA templates were phenol-chloroform extracted and ethanol precipitated. Sense and antisense RNAs were synthesized from the appropriate DNA template by using T7 and T3 in vitro transcription kits (Boehringer Mannheim) according to the manufacturer's instructions. Equal volumes of each single-stranded RNA were mixed, denatured at 65°C, and cooled down slowly at room temperature to allow annealing of the complementary strands. Young adult hermaphrodites were injected with 250 µg of dsRNA per ml in each gonad. Hermaphrodites were allowed to lay eggs for 24 h and then were transferred to new plates. The number of eggs was counted 12 h later, and the number and sex of surviving progeny were scored 3 days later (F₁). Young adult F₁ animals were transferred to new plates, and the same analyses were performed for F₂ eggs and animals.

Recombination analysis. Hermaphrodites homozygous for an X chromosome carrying dpy-6(e14) and unc-3(e151) or chromosome III carrying dpy-18(e364) and unc-25(e156) were injected with Ce-cds-1 or control dsRNA and were crossed with wild-type males for 24 h. The F₁ progeny (non-Dpy and non-Unc hermaphrodites) were individually separated into new plates until early L4 stage. The F₁ heterozygotic hermaphrodites were allowed to lay eggs for 18 h, and the F₂ progeny were scored for recombinant phenotypes.

DAPI staining. Worms or eggs were washed briefly in M9 buffer and then fixed and stained in 1 ml of 200-ng/ml DAPI (4', 6'-diamidino-2-phenylindole) in 95% ethanol. After 30 min at room temperature, the worms or eggs were washed in M9 buffer and mounted on 5% agar pads for microscopy.

Electron microscopy. Young gravid adult worms were staged and prepared for electron microscopy by conventional chemical fixation as described previously (10).

Nucleotide sequence accession number. The GenBank/EMBL/DDBJ accession number for Ce-cds-1 is AB041996.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Identification of Ce-CDS-1 and Ce-CDS-2 and sequence comparison to the Cds1 and Chk2 ortholog. By employing PCR and database analysis, we have identified the C. elegans homolog of Cds-1- and Dmnk-related kinase, Ce-CDS-1 (Fig. 1A). In the C. elegans genome database, this protein is designated Y60A3A.12. This protein, which we named Ce-CDS-1, is presumably identical to C. elegans Chk2, reported by Matsuoka et al. (20). In the genome database, another gene, designated T08D2.7, may also encode a C. elegans homolog of Cds-1- and Dmnk-related kinase, Ce-CDS-2 (Fig. 1B). The Ce-cds-1 and Ce-cds-2 cDNAs encode a 53-kDa translational product of 476 amino acids and a putative 46.5-kDa translational product of 414 amino acids, respectively. The Ce-CDS-1 and Ce-CDS-2 proteins have structural similarities with S. cerevisiae Rad53 (ScRad53), Dun1 (ScDun1), and Mek1p (ScMek1p), S. pombe Cds1 (SpCds1), Mek1 (SpMek1), D. melanogaster Dmnk, and Chk2 (mammalian Cds1) (Fig. 1B and C). Ce-CDS-1 protein is 94% identical to Ce-CDS-2, 34% identical to Dmnk, 32% identical to Chk2, 29% identical to ScRad53, 28% identical to SpCds1, and 26% identical to ScDun1, ScMek1p, and SpMek1. On the basis of sequence analysis, Ce-CDS-1 and Ce-CDS-2 appear to be more closely related to ScRad53 than to ScMek1p (see below for further discussion). Ce-CDS-1 and Ce-CDS-2 have a single amino-terminal FHA domain that was first identified in several transcription factors with a forkhead DNA-binding domain and was also found in the Dmnk, Chk2, Rad53, Mek1, and Cds1 family of protein kinases. Northern (RNA) blot analysis of embryos at a variety of developmental stages revealed the presence of the Ce-cds-1 and/or Ce-cds-2 (which we call Ce-cds-1/2) transcript (Fig. 1D, left panel). The expression of Ce-cds-1/2 was first detectable at the L3 stage of larval development and became stronger as the worms reached adulthood (Fig. 1D, left panel). As shown in Fig. 1D (right panel), the expression of Ce-cds-1/2 was detected in N2 wild-type adult animals but not in glp-4 mutants, indicating that Ce-cds-1/2 expression in adults is dependent on the presence of a germline. The temporal expression pattern of Ce-cds-1/2 is reminiscent of that of Dmnk (26), suggesting that Ce-CDS-1/2 may also be involved in meiosis and/or germline establishment.

View larger version (140K): [in this window] [in a new window]   FIG. 1.   Structure of Ce-CDS-1 protein and expression of Ce-cds-1 during development of C. elegans. (A) The amino acid sequence of Ce-CDS-1. The FHA domain and ATP binding motif (GKGGFG----AIK) are indicated by a bracket and an arrow, respectively. Underlined nucleotides indicate a predicted ATP binding motif. (B) Sequence comparison of Ce-CDS-1 with Ce-CDS-2, Dmnk, Chk2, SpCds1, ScMek1p, SpMek1, ScRad53, and ScDun1 family of protein kinases. Amino acids identified in two or more proteins are shaded. (C) The phylogenetic relationship between Ce-CDS-1 and the other Cds1 orthologs from different species. Alignments were loaded into ClustalW, which calculated an unrooted tree and all branch lengths by using the neighbor-joining method. The resultant tree was produced in Phylip format. Dendro Maker for Macintosh was used to convert the tree into graphical format. (D) Temporal expression of Ce-cds-1/2 during development of C. elegans. Total RNA was prepared from C. elegans wild-type hermaphrodites at various stages of development or prepared from adult hermaphrodites with a greatly reduced germline (glp-4). RNA was separated by 1% agarose formaldehyde gels, transferred onto nylon membranes, and hybridized with a radiolabeled probe for Ce-cds-1. The filters were stained with methylene blue to show 18S and 28S rRNAs (28S rRNA in bottom panel).

Ce-CDS-1/2 is required for bivalent formation. To examine the role of Ce-CDS-1/2, we have employed RNAi to reduce or eliminate the function of the endogenous gene (4, 13, 34). dsRNAs covering various regions of Ce-cds-1 cDNA were synthesized and injected into adult hermaphrodites (Fig. 2A). Because of the high level of sequence identity (>95%) of Ce-cds-1 and Ce-cds-2, both genes are likely to be affected by the Ce-cds-1 dsRNAs used. The Ce-cds-1 (RNAi) F₁ animals were morphologically normal, exhibited no apparent somatic defects, and laid fertilized eggs (Table 1). However, the majority (~90%) of F₂ progeny died at various stages during embryogenesis. Of the remaining progeny (~10%) that did hatch, about half survived up to adulthood. These results suggest that some kind of gamete abnormality occurred in the F₁ animals. Males normally arise among self-progeny of hermaphrodites at a low frequency (0.1 to 0.2%) as the result of spontaneous nondisjunction of the X chromosome. Interestingly, the frequency of males in the surviving F₂ progeny was drastically increased (15 to 18%) compared to that in control hermaphrodites (Table 1). It has been indicated that this "high incidence of males" phenotype is diagnostic of an abnormality in chromosome segregation (15). Consistent with this notion, DAPI staining of the Ce-cds-1 (RNAi) F₂ early embryos revealed a high frequency of chromosome aneuploidy (Fig. 2B). This result suggests that the extensive embryonic lethality seen in F₂ progeny may be a consequence of errors during meiotic chromosome segregation and that Ce-CDS-1/2 is required for proper chromosome segregation during meiosis.

View larger version (60K): [in this window] [in a new window]   FIG. 2.   Chromosome aneuploidy in Ce-cds-1 (RNAi) embryos and lack of bivalents in Ce-cds-1 (RNAi) oocytes at diakinesis. (A) Schematic diagram of dsRNAs (1 to 4) covering various regions of Ce-cds-1 cDNA. It should be noted that there may be two distinct genes (Y60A3A.12 and T08D2.7) encoding two putative Dmnk orthologs (Ce-CDS-1 and Ce-CDS-2) in C. elegans. Considering a high degree of identity (>95%) between the two genes, Ce-cds-1 dsRNAs used may affect both genes. KD, kinase domain. (B) Aneuploidy in Ce-cds-1 (RNAi) embryos. DAPI staining of control (mRor1) (27) dsRNA-injected F2 embryos (panel i) and Ce-cds-1 dsRNA-injected F2 embryos (panels ii and iii). Bar, 10 µm. Uneven segregation in Ce-cds-1 dsRNA-injected cells indicates aneuploidy. (C) Absence of bivalent formation in oocytes of the Ce-cds-1 (RNAi) F1 animals. Each image represents DAPI-stained nuclei at diakinesis. Oocytes from control dsRNA-injected F1 worms (panel i) and from Ce-cds-1 dsRNA-injected F1 worms (panel ii). The average number of DAPI-stained bodies detected in oocytes from control animals was close to 6, while an average of 11.5 bodies could be detected in oocytes from Ce-cds-1 dsRNA-injected worms. In this experiment, 138 oocytes from 35 independent animals (Ce-cds-1 RNAi) were examined. Bar, 10 µm.

Ce-CDS-1/2 is important for meiotic recombination. Although a lack of cytologically detectable chiasmata is likely to reflect a failure in crossing over, it could also occur due to premature release of the physical linkages (22, 24). To clarify this issue, we examined the frequency of crossing over during oocyte meiosis. Genetic exchange was assayed in an interval spanning about two-fifths of the X chromosome and also about one-fourth of an autosome (chromosome III) (Table 3). Hermaphrodites homozygous for dpy-6 and unc-3 (markers on the X chromosome) or dpy-18 and unc-25 (markers on chromosome III) were injected with either Ce-cds-1 dsRNA or control dsRNA and were crossed to wild-type males. The resultant F₁ hermaphrodites heterozygous for dpy-6 and unc-3 or for dpy-18 and unc-25 reproduced by self-fertilization, and the frequency of Unc non-Dpy and Dpy non-Unc recombinants was assayed among the progeny. In control experiments, crossing over was detected on 13.9% of the X chromosomes analyzed (Table 3), in close agreement with previous measurements for this interval. In contrast, drastically reduced crossing over (1.2%) was detected in RNAi experiments with Ce-cds-1 dsRNA (Table 3). Similar reduction in crossing over was also observed in our recombination analysis on chromosome III (Table 3). In some experiments, wild-type hermaphrodites were injected with either Ce-cds-1 dsRNA or control dsRNA and were crossed to males heterozygous for dpy-18 and unc-25. The resultant F₁ hermaphrodites heterozygous for dpy-18 and unc-25 reproduced by self-fertilization, and the frequency of Unc non-Dpy and Dpy non-Unc recombinants was assayed among the progeny. Similar results to those shown in Table 3 were obtained (data not shown). These severe reductions in the frequency of exchange indicate that the absence of chiasmata in oocytes of the Ce-cds-1 (RNAi) F₁ animals is due to failure of crossing over.

Ce-CDS-1/2 appears to be dispensable for formation of the SC. In S. cerevisiae, the initiation of meiotic recombination appears to be functionally linked to the initiation of homologous synapsis (18, 28, 30, 32, 38), although it has been recently reported that meiotic synapsis in Drosophila and C. elegans occurs in the absence of meiotic recombination (10, 21). To test whether Ce-CDS-1/2 is required for meiotic synapsis, transmission electron microscopy of thin sections of worm gonads and surrounding tissue was used to evaluate the structure of the synaptonemal complex (SC) in pachytene oocytes of Ce-cds-1 (RNAi) F₁ animals (Fig. 3). Control pachytene meiocytes showed a typical tripartite SC structure (33), including a ladder-like central element with rungs corresponding to the transverse elements (Fig. 3A), although the longitudinal components were somewhat difficult to discern. Like control worms, pachytene meiocytes from Ce-cds-1 (RNAi) F₁ animals appeared to have normal SC structure (Fig. 3B and C). This result suggests that Ce-CDS-1/2 may be dispensable for synapsis at pachytene, although the results might reflect a hypomorphic phenotype due to RNAi.

View larger version (83K): [in this window] [in a new window]   FIG. 3.   SC in pachytene chromosomes of control and Ce-cds-1 (RNAi) meiocytes. Each photograph shows a portion of a section through a pachytene nucleus. Image A shows a result from control dsRNA-injected F1 animals, while images B and C are results for Ce-cds-1 dsRNA-injected independent F1 worms. The complete SC is visible in both control and Ce-cds-1 dsRNA-injected F1 animals. Bar, 500 nm.

/