Two CDC25 homologues are differentially expressed during mouse development

Genetic and biochemical analysis of species as divergent as yeast and humans has facilitated the identification of many cell division cycle (cdc) genes that universally control the eukaryotic cell cycle. cdc25 plays an important role in the eukaryotic cell cycle by acting as a mitotic inducer. Initially, cdc25 was identified in Schizosaccharomyces pombe (S. pombe; Russell and Nurse, 1986) and was shown to be the tyrosine phosphatase which dephosphorylates the tyrosine-15 residue on the p34^cdc2 kinase and activates it (Gould and Nurse, 1989; Dunphy and Kumagai, 1991; Gautier et al., 1991; Millar et al., 1991a; Strausfeld et al., 1991; Lee et al., 1992). p34^cdc2, initially identified in S. pombe, is a highly conserved serinethreonine kinase found in all eukaryotes examined. In frogs, p34^cdc2 in concert with cyclin B forms an active kinase, termed maturation promoting factor (MPF) (Gautier et al., 1988) which was first identified in oocytes as an activity that caused resumption of meiosis (Masui and Markert, 1971; Smith and Ecker, 1971). The oscillation of MPF activity was also observed in mitotic cell cycles in Xenopus embryos (Gerhart et al., 1984). Wee1 and Mik1 have been characterized as the tyrosine kinases responsible for inactivation of MPF by phosphorylating p34^cdc2 on tyrosine 15 (Featherstone and Russell, 1991; Parker et al., 1991; Lundgren et al., 1991; Lee et al., 1992). Therefore, the tyrosine phosphatase cdc25 acts as a positive regulator of MPF at entry into M-phase.

Multiple cdc25 genes have been identified in flies, humans, rats, and mice. All of the genes isolated are capable of rescuing a temperature-sensitive cdc25 mutant of S. pombe, while sequence identity defines a highly conserved region common to all tyrosine phosphatases (Moreno and Nurse, 1991; Millar and Russell, 1992). In Drosophila, two cdc25 homologs, twine and string (twn and stg), have been identified (Edgar and O’Farrell, 1989; Alphey et al., 1992). Twn is expressed in developing male and female germ cells and both male and female mutants of twn are sterile, demonstrating that twn plays a role in gametogenesis (Alphey et al., 1992; Courtot et al., 1992; White-Cooper et al., 1993). A different role has been suggested for stg. Embryos mutant for stg arrest at the G₂ phase of the fourteenth cell cycle and do not divide or progress further in development (Edgar and O’Farrell, 1989). Maternally derived stg mRNA supports the first 13 cell cycles of the syncitial embryo and a switch to zygotic transcription occurs at cycle 14. These observations suggest that stg is required for mitosis during Drosophila embryogenesis (Edgar and O’Farrell, 1989, 1990). Thus, distinct functions and patterns of expression of the different cdc25 genes are observed during Drosophila development.

In mouse, little is known about the functions and expression patterns of cdc25 genes. Only two mouse genes, homologous to human CDC25B and CDC25C, have been identified (Kakizuka et al., 1992; Nargi and Woodford-Thomas, 1994). Expression studies, limited to late postimplantation stages and some adult tissues, hint at differential expression of these genes. Here, we identify a new mouse cdc25 gene, homologous to human CDC25A, which we designate cdc25a. A highly conserved phosphatase domain is shared by cdc25a and the other mammalian cdc25 gene products described previously. Functional characterization of cdc25a demonstrates phosphatase activity in a standard para-nitrophenylphosphate (PNPP) assay. Mapping studies indicate that cdc25a is genetically distinct and unlinked to cdc25b (manuscript in preparation). To extend the expression analysis in pre-and postimplantation stages of development, in situ hybridization was carried out in addition to northern and RNase protection analysis. The side-by-side comparison of the distribution of transcripts of the two cdc25 genes, cdc25a and cdc25b, illustrates unique expression patterns and suggest that these genes play distinct roles during mouse development.

A PCR-based strategy was employed to isolate mouse cdc25 genes. Degenerate oligonucleiotides, designed to identify human CDC25C, were used for PCR amplification (Sadhu et al., 1990). NIH 3T3 fibroblast RNA was isolated and reverse transcribed as described by Sambrook et al. (1989) for use as a template. The PCR products were cloned into ClaI and XhoI sites in Bluescript as described previously (Sadhu et al., 1990). Two distinct, but related, clones pME10 and pME11 (200 bp and 225 bp inserts, respectively) were identified and sequenced. These two clones were used to probe oligo dT-primed lambda zap cDNA libraries of 8.5 dpc mouse embryos (a gift from Dr J. Gerhart) or NIH 3T3 cells (Stratagene). The longest clones identified with each probe were sequenced completely. Two cDNA fragments of 1.1 kb and 1.9 kb were obtained by screening an 8.5 dpc embryo library using the pME10 probe, and were found to be homologous to human CDC25A. Open reading frames were identified in the 1.1 kb and the 1.9 kb fragments coding for 192 amino acids and 436 amino acids respectively. To obtain the 5 ′ end of the CDC25A homolog a 5 ′ RACE was carried out (Frohman and Martin, 1989) using RACE-ready kidney cDNA (Clonetech Lab) according to the manufacturer’s published protocol. The gene-specific primer used for the first PCR reaction was 5 ′-TACTCTCCTGTAGATGTGACG-3 ′. A degenerate primer based on the sequence of the first six amino acids from human CDC25A (5 ′-ATGG^ACTIGGICCIAGICC-3 ′) and gene specific primer (5 ′-TCTAGAGAATCAGAGTGGCTC-3 ′) were used for the second PCR reaction. A 3 ′ overlap region spanning bases 222 to 336 was identified with the 1.9 kb cDNA clone. The 5 ′ sequence obtained by PCR is shown in Fig. 1 and represents bases 1 to 336. In addition, a 1.0 kb cDNA fragment was obtained from a NIH 3T3 fibroblast library with pME11, which was identical to cdc25b (cdc25M2; the mouse homolog of human CDC25B) (Kakizuka et al., 1992).

The 1.1 kb EcoRI fragment of cdc25a, isolated from the oligo dT-primed 8.5 dpc lambda Zap cDNA library (see above), was used for expression studies. The EcoRI fragment containing the C-terminal region of cdc25a, including the catalytic region and phosphatase box, was subcloned into pGEX1 (Pharmacia). Transformants in the correct orientation were identified and grown in L Broth overnight. Cells were grown to mid-logarithmic phase, IPTG-treated and harvested as described (Lee et al., 1992). The glutathione-S-transferase-cdc25a (GST-cdc25a) fusion protein, purified according to the protocol of Lee et al. (1992), was used for PNPP assays, where hydrolysis of the phosphate group of the substrate PNPP is used to demonstrate phosphatase activity (Guan et al., 1990). The PNPP assays were carried out using 1, 2, 3, or 5 μg of the fusion protein for 60 minutes or 0.6 μg of fusion protein for time intervals ranging from 15 minutes to 60 minutes. Absorbance was measured at 410 nm on a spectrophotometer. GST protein alone was used as a control in these assays.

A multiple tissue northern blot (Clonetech Labs, Lot no. 36344) was used to analyze cdc25a mRNA levels in different mouse tissues. 2 μg of poly(A)⁺ RNA were electrophoresed, transferred to a nylon membrane, prehybridized and hybridized according to the published protocol (Clonetech Labs, Multiple Tissue Northern Blot Protocol). Prehybridization and hybridization were carried out at 50°C in 5× SSPE, 10× Denhardt’s solution, 100 μg/ml salmon sperm DNA and 2% SDS. A 1.1 kb EcoRI fragment of the cdc25a cDNA was labeled with [α-³²P] dCTP using the multi-prime method (Amersham) and used as probe. Blots were rinsed in 2× SSC, 0.5% SDS and washed in 0.1× SSC, 0.1% SDS at 50°C according to the manufacturer’s protocol and exposed to film with intensifying screens.

Total RNA was made from postimplantation mouse embryos (7.5 to 15.5 dpc), from isolated organs of 13.5 dpc and 16.5 dpc embryos and from adult testis and ovary using RNAzol B (Cinna/Biotex) by the method of Chomczynski and Sacchi (1987). An antisense cdc25a probe was generated from a linearized PvuII fragment of digested pME10 containing the 200 bp insert of cdc25a using T7RNA polymerase and [α-³²P]CTP (800 Ci/mM). The RNA probe was subsequently treated with DNAse, gel purified and approximately 10,000 cpm hybridized overnight with each sample of embryo RNA. RNase digestion and analysis was carried out according to the protocol of Zinn et al. (1983). tRNA (Sigma) was run in parallel as a control, while a β-actin probe was used on the same samples as a loading control.

In situ hybridization was carried out on postimplantation embryos (10.5 days post-coitus (dpc) to 17.5 dpc) according to the protocol of Tessarollo et al. (1993). RNA probes were synthesized using S³⁵-labeled UTP and T7 polymerase. Probes synthesized for cdc25a spanned bases 222 to 2106 while the cdc25b probe spanned 1454 to 2560 bases (Kakizuka et al., 1992). Control sense probes were synthesized using T3 polymerase and were used in comparison.

Female B6C3F1 mice (Jackson Labs) were stimulated to ovulate with pregnant mare’s serum gonadotropin (Sigma). 24 hours later the mice were injected with human chorionic gonadotropin (Sigma) and housed with male B6C3F1 mice overnight. The presence of a vaginal plug indicated a successful mating. Fertilized eggs at the one-cell stage were obtained from these mice by dissection and flushing of the oviduct. Embryos were fixed immediately or cultured further for 1-4 days to obtain 2-cell, 4-cell, 16-cell, morula and blastocyst stage embryos. These embryos of different ages were introduced into oviducts isolated from recipient females, fixed overnight in 4% paraformaldehyde at 4°C and dehydrated through an ethanol series.

Oviducts containing embryos were embedded in paraffin and 6 μm sections were cut and collected on siliconized slides. Sections were dewaxed in xylene, rehydrated through an ethanol series and refixed in 4% paraformaldehyde at 4°C (20 minutes). Sections were then treated with proteinase K (20 μg/ml for 7.5 minutes), fixed again in 4% paraformaldehyde (5 minutes), incubated in 0.25% acetic anhydride in 0.1 M triethanolamine (10 minutes) and again dehydrated. Slides were hybridized overnight under coverslips in a moist chamber at 50-55°C with 1-5 ng/μl of digoxigenin-labeled RNA probe in 75 μl hybridization solution. Probes were generated according to the protocol in the Genius Kit (Boehringer Mannheim Biochemicals). The probes for cdc25a and cdc25b were derived from pME10 and pME11, respectively using the T7 promoter and each was approximately 220 bases in length. Control (sense) probes were also synthesized similarly for pME10 and pME11 using the T3 promoter.

Following hybridization, slides were washed at high stringency (20-30 minutes at 65°C in 50% formamide, 2× SSC), RNase treated for 30 minutes at 37°C, and again washed at high stringency as before. After equilibrating in Buffer 1 from the Genius Kit, the slides were incubated in a 1:1000 dilution of alkaline phosphatase-conjugated anti-digoxigenin antibody (Boehringer Mannheim Biochemicals) and washed extensively in several changes of Buffer 1, before overnight incubation in substrate (Buffer 3 and the staining solution from the Genius Kit). The reaction was stopped by incubation for 10 minutes with TE, followed by 10 minutes in water, and finally gelvatol and coverslips were placed on the slides for photography and storage. Slides were analyzed using a confocal laser scanning microscope (CLSM; Zeiss, Germany). A ×40 oil immersion, Nomarski objective (N.A. 1.3, acrostigmatic lens) was used to identify embryos using a scan rate of 8 seconds. The Nomarski images are represented by the blue channel and the reflected peroxidase-digoxigenin labeled probe is represented by the red channel. All manipulations on the CLSM were according to the protocol described by Tsarfaty et al. (1992, 1994).

A comparison of known cdc25 protein sequences from different species reveals a conserved carboxy-terminal region spanning approximately 100 amino acids. To isolate the mouse cdc25 cDNAs, PCR products were obtained using degenerate primers designed from the conserved carboxy terminal of the Drosophila stg gene (Sadhu et al., 1990) using NIH 3T3 cDNA as template. Two distinct PCR-derived clones (pME10 and pME11) were used as probes to screen an 8.5 dpc oligo dT-primed mouse embryo or a NIH 3T3 cDNA library. In the first screen, using pME10 as probe, two cDNAs of 1.1 and 1.9 kb were obtained and were found upon sequence analysis to represent an identical gene (Fig. 1). The 1.9 kb fragment is shown spanning bases 220 to 2106, and the 1.1 kb fragment spans bases 961 to 2106 and is therefore entirely contained within the 1.9 kb fragment (Fig. 1). Sequence comparison reveals 84% identity with human CDC25A while 63% and 65% identity is shared with human CDC25B and CDC25C respectively. Amino acid sequence comparison demonstrates these clones share 77% identity with the human CDC25A protein (Fig. 2). This clone represents the third mouse cdc25 gene identified and can be designated cdc25M3 but to conform with the human gene nomenclature we propose designating it cdc25a. The cdc25a start site obtained by 5 ′ RACE-PCR corresponds to the Met at position 1 of the human CDC25A.

A second screen of a NIH 3T3 fibroblast cDNA library using pME11 as a probe yielded a 1.0 kb cDNA identical to cdc25M2 (Kakizuka et al., 1992) which is the mouse homolog of human CDC25B and which we will refer to as cdc25b. Both mouse clones, cdc25a and cdc25b, reveal a shared amino acid motif (XXDCRYPYEYXG-GHIXXAXNL and KRX-IXXFXCEFSSERG-PRMCR) with all other mammalian cdc25 genes (Fig. 2). The motif FHCXXXXXR, which flanks the active site of protein tyrosine phosphatases, is included in this conserved region of mouse cdc25a suggesting potential tyrosine phosphatase activity. Amino acid sequence comparison demonstrates that cdc25a and cdc25b share a low overall identity of 29% and 34% respectively with the Drosophila stg gene and 34% and 32% identity respectively with the Drosophila gene twn.

In order to determine if cdc25a acts as a phosphatase, a GST-cdc25a fusion protein was assayed for its ability to hydrolyse PNPP (Guan et al., 1990, 1991). The GST-cdc25a fusion, constructed using the C-terminal 198 amino acids of cdc25a, was used to demonstrate a conversion of PNPP to p-nitrophenol. GST-cdc25a fusion protein or GST protein alone were used in the assays. The kinetic and concentration-dependent hydrolysis of PNPP indicates that cdc25a-GST is a phosphatase (data not shown). GST protein alone did not hydrolyse PNPP. Thus, the prediction that cdc25a is a phosphatase can be confirmed and, furthermore, suggest that it is functionally similar to other members of the cdc25 family.

In order to analyze cdc25a expression in adult animals and postimplantation embryos, we carried out northern analysis, in situ hybridization and RNase protection assays. Northern analysis reveals that a 3.4 kb cdc25a transcript is expressed in some adult tissues (Fig. 3). High levels of cdc25a transcript were detected in adult testis while lower levels were detected in kidney, liver, heart, and skeletal muscle and none were detected in spleen and lung (Fig. 3). Interestingly, in situ hybridization analysis of cdc25a and cdc25b show varying patterns of RNA expression in the adult testes and ovary. cdc25a is highly expressed in the adult testes particularly in primary and secondary spermatocytes, but the probe does not appear to hybridize to spermatogonia or spermatids (Fig. 4A,B). In contrast, cdc25b is poorly expressed in primary and secondary spermatocytes and in somatic tissue (Fig. 4E,F). In the ovary, cdc25a is primarily expressed in the oocytes of early preantral and antral follicles (Fig. 4C,D; arrows), while a lower level of expression is observed in granulosa cells and in other somatic tissue. In contrast, lower levels of cdc25b transcripts are observed in oocytes (Fig. 4G,H; arrows) while some hybridization is also observed in the somatic cells of the ovary. To characterize further the developmental expression of cdc25a, RNase protections were carried out on RNA isolated from different stages of postimplantation development from 7.5 dpc to 15.5 dpc. cdc25a expression was observed throughout and appears to be essentially unchanged across this developmental spectrum in comparison to the β-actin controls (Fig. 5). These data show that cdc25a RNA expression does not fluctuate significantly between 7.5 dpc and 15.5 dpc of postimplantation mouse development while cdc25a transcript levels vary in adult tissue.

A detailed comparative analysis of cdc25a and cdc25b expression was carried out on 10.5 to 17.5 dpc embryos by in situ hybridization. Previous studies have described the expression pattern of cdc25b in postimplantation embryos (Kakizuka et al., 1992). cdc25a is expressed in most developing tissue and is observed ubiquitously in the 10.5 dpc embryo (Fig. 6A). For example, the developing heart, primitive gut, liver primordia, the branchial arch, and the developing brain, in particular the proliferative-neuroepithelium, express cdc25a. Similarly, in the 13.5 dpc (Fig. 6B) and the 17.5 dpc embryo cdc25a is expressed in most tissues (Fig. 6C,D,E). The developing liver of the 13.5 dpc embryo expresses high levels of cdc25a while expression levels decrease by 17.5 dpc. A similar pattern of cdc25a expression is observed in the developing kidney and adrenal gland of the 13.5 dpc and 17.5 dpc embryo. cdc25a expression in the brain is particularly apparent in the developing neuroepithelium of all embryos examined. In contrast, cdc25b reveals a more restricted pattern of expression. cdc25b transcripts are detected primarily in developing liver and neuroepithelium in 13.5 dpc embryos while the signal decreases in the 17.5 dpc embryo (data not shown) in agreement with previous results (Kakizuka et al., 1992). Since widespread embryonic expression of cdc25a was observed consistently in several in situ hybridization experiments, RNase protections were carried out to independently confirm these results. RNase protection analysis was carried out on RNA obtained from organs isolated from 13.5 dpc and 16.5 dpc embryos. cdc25a expression is detected in embryonic brain, heart, kidney and adrenal gland, liver, lung and stomach (Fig. 7).

To examine the expression pattern of cdc25a and cdc25b during preimplantation development, in situ hybridization experiments were carried out on one-cell, two-cell, four-cell (Fig. 8), eight-cell (not shown), sixteen-cell, morula and blastocyst stage embryos (Fig. 9). Distinct patterns of expression were observed for cdc25a and cdc25b. cdc25a transcripts were not detectable in the early embryo stages (one-, two-and four-cell, Fig. 8A,C,E respectively; eight-cell (data not shown), 16-cell and morula, Fig. 9A,C) when compared to the sense controls for each experiment (data not shown). The expression of cdc25a is first detected as a very strong signal at the blastocyst stage (Fig. 9E), both in trophectoderm and inner-cell mass (ICM). In contrast, cdc25b is observed early in preimplantation embryogenesis. cdc25b expression is detected at the one-cell stage (Fig. 8B) followed by a decrease in signal intensity at the two-cell stage (Fig. 8D). However, the cdc25b RNA is detectable at a higher level in the four-cell stage (Fig. 8F) and is expressed through the eight-cell (data not shown), 16-cell and morula stages (Fig. 9B,D, respectively). Maximal expression of cdc25b is observed at the blastocyst stage (Fig. 9F) in the trophectoderm and ICM. Numerous embryos from two or more experiments were analyzed at each stage of development and consistently demonstrate the expression patterns described. Confocal images were generated under uniform conditions for all stages shown. Since all embryos processed were in interphase and none were in mitosis, cdc25a and cdc25b expression is described for embryos at interphase. In addition, embryos were also examined with regard to compartmentalization of the cdc25 signal within the cell. However, the signal is prevalent in both the nuclear and cytoplasmic compartments. Absence of signal in some regions appears to be due to side preparation and was not a consistent observation.

This study identifies a new mouse homolog of cdc25. Cloning and sequence comparison of cdc25a reveals conservation of a phosphatase domain which is common to all cdc25 family members. As predicted from sequence analysis, cdc25a demonstrates phosphatase activity. Furthermore, we describe a distinct pattern of expression of two mouse cdc25 genes which correlate with distinct functions for each gene.

Sequence comparison reveals that cdc25a is the mouse homolog of human CDC25A sharing 83% similarity between the two proteins and also demonstrates that cdc25a shares highly conserved regions with other cdc25 homologs (Fig. 2). The ‘phosphatase motif’ FHCXXXXXR characteristic of all tyrosine phosphatases, including the cdc25 genes, is completely conserved in cdc25a, suggesting that it is a bonafide cdc25 family member. The phosphatase activity of cdc25a was confirmed by demonstrating that a GST-cdc25a fusion protein could dephosphorylate the substrate PNPP. A similar response is observed in PNPP assays carried out with other cdc25 homologs, suggesting that cdc25a could act to dephosphorylate p34^cdc2 kinase.

Expression studies of cdc25 genes in mouse were carried out on preimplantation embryos (this study), postimplantation embryos (this study; Kakizuka et al., 1992) and adult tissue (this study; Kakizuka et al., 1992; Nargi and Woodford Thomas, 1994). Analysis of adult tissue shows cdc25a is expressed at a high level in testis and at lower levels in ovary, which correlates with meiotic and mitotic activity in these organs. A more detailed in situ hybridization analysis of cdc25a and cdc25b expression in testis and ovary reveal varied patterns of expression in male and female germ cells. It is interesting to note that cdc25a is expressed in oocytes from most stages of follicular development and primary and secondary spermatocytes which represent stages of meiotic germ cells. In contrast, cdc25b is not highly expressed in the germline of the ovary and testis. These results suggest that cdc25a may play a role in meiosis, which is analogous to the role of twn in Drosophila gametogenesis. However, it is important to note that, unlike twn, cdc25a is also expressed in other somatic tissue such as brain, heart, liver, kidney, and skeletal muscle. These observations collectively suggest that cdc25a may have a dual function in mitosis and meiosis. The observation of cdc25a transcript expression in non-dividing tissue such as heart and skeletal muscle may reflect posttranscriptional and posttranslational regulation of cdc25a activity. Indeed, recent evidence suggests that cdc25 itself may be regulated by phosphorylation (Kumagai and Dunphy, 1992; Ogg et al., 1994). A similar observation of cdc25b expression in nondividing adult tissue has been reported (Kakizuka et al., 1992). cdc25b is expressed predominantly in spleen, lung, heart, brain, intestine, and muscle (Kakizuka et al., 1992) while cdc25c is expressed in spleen and thymus (Nargi and Woodford-Thomas, 1994). These results demonstrate that cdc25 genes display varied patterns of expression in adult tissue.

RNase protections and northern analysis of postimplantation embryos reveal that cdc25a is expressed from 7.5 dpc to 15.5 dpc of gestation and does not change significantly at these stages. In contrast, cdc25b is expressed at 10.5 dpc, peaks at 13.5 dpc, and decreases at 17.5 dpc (Kakizuka et al., 1992). In situ hybridization and RNase protection analysis of postimplantation tissue reveals that cdc25a is ubiquitously expressed in most developing tissue, while cdc25b has a more restricted pattern of expression. cdc25b expression is mostly confined to proliferative liver and neural tissue as observed in this study and by Kakizuka and co-workers (1992). In contrast, cdc25a is expressed not only in proliferative liver and neuroepithelium, but also in developing heart, kidney, adrenal gland, stomach, intestine, and lung. cdc25a transcripts are observed in the 13.5 dpc embryonic liver and kidney but have decreased by 17.5 dpc, which correlates with a loss of mitotic activity and increased differentiation of these tissues. Collectively, these observations show that cdc25a and cdc25b are expressed differentially during postimplantation development.

In situ hybridization analysis of preimplantation embryos demonstrate that cdc25a is expressed only at the blastocyst stage. cdc25b is expressed at the one-cell stage, decreases to background levels at the two-cell stage and is observed once again from the four-cell to the blastocyst stage. Interestingly, we note that cdc25a transcripts are expressed in ovarian oocytes, but are not observed in the one-cell embryo. cdc25b transcripts, on the other hand, are expressed in oocytes at low levels and are present in one-cell embryos. Maximal expression of cdc25b is observed at the blastocyst stage. cdc25c expression has not been examined in embryos. Taken together, these results demonstrate a different pattern of expression for two cdc25 mouse homologs, suggesting a distinct role for each of these genes. It is interesting to note that the onset of cdc25a and cdc25b expression is coincident with two important developmental transitions in the early embryo. Expression of cdc25a is first detected in the ICM and the trophectoderm of the blastocyst stage which gives rise to the embryo proper and extra-embryonic membranes respectively at implantation. Proliferative activity occurs at this stage in preparation for implantation (for reviews see Rossant, 1986; Cruz and Pedersen, 1991). Expression of cdc25b decreases at the two-cell stage and is re-expressed at the four-cell stage, which correlates well with the transition from maternal to zygotic transcription control in early mouse development (for review see Pratt, 1989; Telford et al., 1990; Schultz, 1993). An analogous transition is observed in the Drosophila embryo and is controlled by the stg gene (Edgar and O’Farrell, 1989).

Cell cycle-specific expression and functions of the different CDC25 genes have been examined in mammalian cells. CDC25C protein is expressed throughout the cell cycle in HeLa cells (Millar et al., 1991b) while CDC25A protein is observed predominantly at G₁ (Jinno et al., 1994). Microinjection of CDC25C antibodies arrest cells at G₂ while CDC25A arrest cells at G₁ (Millar et al., 1992; Hoffman et al., 1994; Jinno et al., 1994). These results suggest that CDC25C and CDC25A perform different functions in the cell cycle. The G₁/S cell cycle transition in mammals is thought to be controlled by the cyclin dependent kinases (cdks) and is negatively regulated by Thr14/Tyr15 phosphorylation of p33^cdk2 (Gu et al., 1992; Sebastian et al., 1993). Thus, dephosphorylation of p33^cdk2 is necessary for its activation. Recently, CDC25A has been shown to dephosphorylate p33^cdk2 in complex with cyclin E, in vitro (Hoffman et al., 1994). In addition, CDC25A phosphatase activity is detectable in cells at G₁ and S phase of the cell cycles but not in G₂ and M-phases (Hoffman et al., 1994). These results suggest that a distinct G₁/S role is played by cdc25A. It is tempting to speculate that different mouse cdc25 genes may also have distinct cell cycle-regulatory functions. cdc25a is first expressed at the blastocyst stage which correlates with establishment of a more ‘typical’ G₁ phase in the mouse embryo (Chisholm, 1988) corroborating the observation of a G₁ role for CDC25A as reported recently (Jinno et al., 1994; Hoffman et al., 1994). In contrast to the emerging G₁/S phase role for CDC25A, it is interesting to note that cdc25a is also expressed in meiotic germ cells and may function during meiosis in addition to mitosis. This observation suggests that cdc25a may play an analogous role to twn in Drosophila game togenesis. However, since cdc25a is not exclusively expressed in meiotic cells, unlike twn in Drosophila, overlapping and compensatory roles for mammalian CDC25 genes may have evolved. The mouse embryo offers an excellent model for examining this proposition further.

In conclusion, we have identified two distinct cdc25 mouse genes. Furthermore, these genes demonstrate distinct developmentally regulated patterns of expression which coincide with major developmental transitions during early embryogenesis and gametogenesis. The data suggest that key cell cycle regulators such as cdc25 could play an important role in regulating mitosis in the developing mouse embryo and meiosis in the germline.

We thank Paul Russell, Tony Hunter and Byron Sebastian for sharing sequence information at the initial stages of the project. We also thank Helen Piwnica-Worms and Scott Ogg for providing vectors and for advice on purifying fusion proteins. Sue Hong and Lynn Bixler are acknowledged for technical help with sequencing and embryo collection respectively, while Jim Resau is thanked for generating confocal images. Leslie Lock is thanked for helpful comments and critical reading of the manuscript. We also thank Cathy Hahn and Madeline Wilson for excellent secretarial assistance. This research was sponsored in part by the National Cancer Institute, DHHS, under contract N01-CO-74101, with ABL.