The Saccharomyces cerevisiae CDC25 gene encodes a guanine nucleotide exchange factor (GEF) for Ras proteins. Its catalytic domain is highly homologous to Ras-GEFs from all eukaryotes. Even though Cdc25 is the first Ras-GEF identified in any organism, we still know very little about how its function is regulated in yeast. In this work we provide evidence for the involvement of the N terminus of Cdc25 in the regulation of its activity. A truncated CDC25 lacking the noncatalytic C-terminal coding sequence was identified in a screen of high-copy suppressors of the heat-shock-sensitive phenotype of strains in which the Ras pathway is hyper-activated. The truncated gene acts as a dominant-negative mutant because it only suppresses the heat-shock sensitivity of strains that require the function of CDC25. Our two-hybrid assays and immunoprecipitation analyses show interactions between the N terminus of Cdc25 and itself, the C terminus, and the full-length protein. These results suggest that the dominant-negative effect may be a result of oligomerization with endogenous Cdc25. Further evidence of the role of the N terminus of Cdc25 in the regulation of its activity is provided by the mapping of the activating mutation of CDC25HS20 to the serine residue at position 365 in the noncatalytic N-terminal domain. This mutation induces a phenotype similar to activating mutants of other genes in the Ras pathway in yeast. Hence, the N terminus may exert a negative control on the catalytic activity of the protein. Taken together these results suggest that the N terminus plays a crucial role in regulating Cdc25 and consequently Ras activity, which in S. cerevisiae is essential for cell cycle progression.
IN the yeast Saccharomyces cerevisiae, the Ras1 and Ras2 proteins regulate adenylyl cyclase, which produces cAMP. Increase in cAMP levels activates the cAMP-dependent protein kinases, which have an essential role in progression from the G1 to S phase of the cell cycle (Broach 1991; Thevelein 1994). The Ras proteins are, in addition, necessary for completing mitosis (Morishitaet al. 1995) and regulating filamentous growth (Gimenoet al. 1992).
Ras belongs to a superfamily of small G proteins that bind guanine nucleotides and have an intrinsically slow GTPase activity. Ras is active when bound to GTP and inactive when bound to GDP (Lowy and Willmusen 1993). These biochemical properties are essential for its biological function. In mammalian cells mutations that increase the amount of GTP-bound Ras increase its oncogenic potential and are among the most frequent mutations in human tumors (Lowy and Willmusen 1993). Similar mutations in S. cerevisiae RAS cause an inability to respond properly to the nutritional environment, stimulate pseudophyphal growth, and induce heat-shock sensitivity (Broach 1991; Gimenoet al. 1992). Therefore, regulation of the nucleotide bound state of Ras is a critical step in modulating Ras function in intact cells.
The activation state of Ras is controlled by at least two classes of proteins: GTPase-activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs) (Boguski and McCormick 1993). GAP proteins inhibit Ras by specifically binding to the GTP-bound form and stimulating its GTPase activity (Boguski and McCormick 1993; Lowy and Willmusen 1993). GAP proteins in S. cerevisiae are encoded by the IRA1 and IRA2 genes and negatively regulate RAS function in intact cells (Broach 1991; Boguski and McCormick 1993). Positive regulation of Ras requires the action of GEFs that facilitate the exchange of GDP bound to Ras for GTP (Boguski and McCormick 1993). Ras-GEFs in yeast include Cdc25 and Sdc25 (Broach 1991; Boy-Marcotteet al. 1996). CDC25 is essential for cell growth, whereas SCD25 is not. SCD25 is transcribed only when nutrients are depleted or when cells are grown in non-fermentable carbon sources.
While genetic evidence shows that in yeast CDC25 is required for RAS functioning, the signal that activates Cdc25 is still unknown. Feeding glucose or a related fermentable sugar to starved S. cerevisiae cells or inducing intracellular acidification increases the activity of adenylyl cyclase that triggers a rapid and transient increase in cAMP levels. (Broach 1991; Thevelein 1994). CDC25 is not required for the increase in cAMP by intracellular acidification (Colomboet al. 1998). Although the catalytic C terminus of Cdc25 has previously been reported to be required for the glucose-induced rise in cAMP (Munder and Küntzel 1989; Van Aelstet al. 1990), recent reports suggest that Cdc25 is not the signal receiver (Goldberget al. 1994; Colomboet al. 1998).
CDC25 encodes a 1589-amino-acid protein expressed as a polypeptide of ~180-kD (Joneset al. 1991; Grosset al. 1992b) with a C-terminal catalytic domain (amino acids 1121–1573) that has high homology to the catalytic domain of Ras-GEFs from all organisms (Quilliamet al. 1995). The C-terminal catalytic domain of CDC25 (amino acids 1102–1589) is sufficient for full biological activity in yeast (Laiet al. 1993). The noncatalytic N-terminal domain of Cdc25 (residues 1–1121) has no homology to other Ras-GEFs (Quilliamet al. 1995). This domain of Cdc25 is characterized by the presence of an SH3 domain (residues ~60–130). SH3 domains are present in the regulatory region of many signaling molecules and they bind proline-rich sequences. In Cdc25 the SH3 domain binds adenylyl cyclase and seems to enhance its responsiveness to activation by Ras in vitro (Freemanet al. 1996; Mintzer and Field 1999).
A contribution from the N terminus of Cdc25 to its biological function has been suggested by the finding that upon glucose stimulation in yeast, residues within amino acids 114–348 become phosphorylated leading to decreased association of Cdc25 to the membranes and accessibility to Ras (Grosset al. 1992a). The putative phosphorylated residues are predicted to be substrates of the cAMP-dependent protein kinase. This phosphorylation is believed to be part of a negative feedback loop because in yeast the activity of the cAMP-dependent protein kinase is regulated by Ras. Deletion of amino acids 114–348 in Cdc25 results in a constitutively high level of cAMP presumably because of increased accessibility to Ras. Deletion of this fragment may also affect biological activity by modulating stability of the protein through a cyclin destruction box at position 148 (Kaplon and Jacquet 1995).
Even though Cdc25 is the first Ras-GEF identified in any organism, still very little is known about how its function is regulated in yeast. In this article we provide evidence of the involvement of the N-terminal domain of Cdc25 in regulation of its activity. We show that the N terminus of Cdc25 acts as a dominant-negative mutant inhibiting the function of Cdc25 in vivo possibly by interaction with the endogenous Cdc25, thus interfering with its ability to activate Ras. In addition we show that the activating mutation of the heat-shock-sensitive CDC25 mutant CDC25HS20 (Broeket al. 1987) maps to the noncatalytic N terminus of the protein at position 365. These data suggest that the N terminus plays a crucial role in regulating Cdc25 and consequently Ras activity, which in S. cerevisiae is essential for cell cycle progression.
MATERIALS AND METHODS
Yeast strains and media: The S. cerevisiae strains used in this study are the following: SP1 (MATa his3 leu2 ura3 trp1 ade8) (Todaet al. 1985); IR-1 (MATa his3 leu2 ura3 trp1 ade8 ira1Δ::HIS3) (Ballesteret al. 1989); IR-2.5 (MATa his3 leu2 ura3 trp1 ade8 ira2Δ::ADE8) (Andersenet al. 1993); IR-2.53 (MATa his3 leu2 ura3 trp1 ade8 ira1Δ::HIS3 ira2Δ::ADE8) (Andersenet al. 1993); TK161-R2V (MATa his3 leu2 ura3 trp1 ade8 RAS2val19) (Todaet al. 1985); TMRV-25 (MATa his3 leu2 ura3 trp1 ade8 cdc25Δ::URA3 RAS2val19) (Broeket al. 1987); TT1A-1 [MATα his3 leu2 ura3 trp1 ade8 cdc25Δ::URA3 pCDC25(TRP1)-1] (Broeket al. 1987); LVA25-5 (MATa his3 leu2 trp1 ura3 ade8 cdc25-5ts), STS1 (MATa his3 leu2 trp1 ura3 ade8 ras1Δ::URA3 ras2ts) (Powerset al. 1989); and HF7c (MATa ura3-52 his3-200 lys2-801 ade2-101 trp1-901 leu2-3112 gal4-542 gal80-538 LYS2::GAL1-HIS3 URA3::GAL4-LacZ) (Van Aelst 1997). The strains RB32C MATα his3 leu2 ura3 trp1 ade8 cdc25Δ::URA3 IRA1Δ::HIS3, RB5A, MATα his3 leu2 ura3 trp1 ade8 cdc25Δ::URA3 ira2Δ::ADE8, and RB5B MATα his3 leu2 ura3 trp1 ade8 cdc25Δ::URA3 ira1Δ::ADE8 were generated by crossing the isogenic strain IR2.53 with the strain TT1A-1. The resulting diploid was sporulated and tetrads were dissected to obtain individual spores.
Standard methods for yeast transformation were as described previously (Roseet al. 1990). Yeast was grown in YPD (1% yeast extract, 2% peptone, 2% dextrose) or synthetic complete (SC) medium containing yeast nitrogen base at 0.67 g/liter, 2% dextrose, and amino acid supplements as described (Roseet al. 1990).
Genetic screen: The ira1Δ strain (IR-1) was transformed with a yeast genomic library cloned into a high-copy plasmid (Yep13M4) that contains the LEU2 gene as a selectable marker (Nikawaet al. 1987). Transformants growing on selective plates were replica plated and heat shocked for 15 min at 55°. Surviving colonies were subjected to segregation analysis (Rose and Broach 1991). Plasmids were recovered from yeast using standard procedures and amplified in Escherichia coli. Sequencing was performed using the Sequenase version 2.0 protocol from Amersham (Buckinghamshire, UK).
Plasmids: Plasmids for mapping the functional domain of CDC25tru were constructed as follows: the plasmid isolated in the screen (pIRIS21) was digested with SmaI and HindIII, blunt ended with Klenow polymerase and ligated to BamHI linkers, and then digested with BamHI and cloned into the BamHI site of pUC119. We then generated a PCR product containing a KpnI and a BamHI site immediately 5′ to the ATG of the truncated CDC25 up to the EcoRV site at position 1791 (Broeket al. 1987). This fragment was used to replace the KpnI/EcoRV fragment of the pUC119-IRIS21 to generate puc119-I21N. We then subcloned the BamHI fragment from pUC119-I21N into the BglII site of the pEMBLYe30/2 plasmid containing a phosphoglycerate kinase promoter-terminator cassette (Banroqueset al. 1986) to generate pEMBL-IR-IS21N (pEMBL-CDC25tru) that expresses amino acids 1–1087 (also referred to as CDC251–1087). To generate CDC25507–1050 we digested pIR-IS21 with PstI, blunt ended with Klenow polymerase and ligated to BamHI linkers. We then subcloned the BamHI fragment into pEMBLYe30/2, resulting in the plasmid named pEMBL-I21Pst/frag. This fragment contains an in frame ATG six nucleotides downstream of the PstI site. The plasmid expressing the hemagglutinin (HA)-tagged CDC251–875 was generated by PCR amplification using the 5′-primer 5′-ACACGGTACCGGATCCATGTCCGATACTAACACGTCTA that introduces a BamHI site next to the ATG and the 3′ primer 5′-TATAGTCCAGTTAATTCTCCAGTATCTCC that introduces a SalI site. The PCR product was cloned into the BamHI/SalI site of the pBGF1 (Chardinet al. 1993) plasmid in frame with two copies of the sequence coding for the HA epitope to generate pBFG1-2628. To generate CDC251–505 we digested the plasmid pBFG1-2628 with PstI, which has a site in CDC25 (position 1826) and a site in the multicloning sequence of the vector. The digested plasmid was religated to generate pBFG1-2628ΔPst. The plasmid expressing HA-tagged CDC25181–875 was generated by PCR amplification using the 5′ primer 5′-TACTCAGGATCCATGTTATCAAATGCCCAC that introduces a BamHI site next to the ATG and the 3′ primer described above. The PCR product was cloned into the BamHI/SalI site of the pBGF1 to generate pBGF1-PCR#3. The plasmid expressing HA-tagged CDC25181–364 was generated by digesting pBGF1-PCR#3 with SacI (in the coding sequence at position 2369) and ApaI (in the multicloning site). The blunt-ended plasmid was religated to generate pBFG1-PCR#3ΔSac. HA-tagged CDC25347–875 was generated by PCR using the 5′ primer, 5′-CATATGGATCCCTTGTTAACCTATATACTAGA introducing a BamHI site and the 3′ primer from above and cloning into pBFG1 to generate pBFG1-PCR#4.
Plasmids for the two-hybrid assay were generated as follows: we made a fusion between the Gal4 transcription-activating domain and Cdc251–875 by subcloning a NcoI/XhoI fragment from the plasmid pBFG1-2628 into the vector pACT2 (Durfeeet al. 1993) to generate pACT2628. The same strategy was used to clone CDC251–505 from pBFG1-2628ΔPst into pACT2 to generate pACT2628ΔPst. A Gal4-DNA-binding domain (DBD) fusion with Cdc251–875 or Cdc25181–875 was generated by subcloning the CDC25 fragments from their respective pBFG1 plasmids into the BamHI/SalI site of the vector pGBT10 (Van Aelst 1997) to generate pGBT10-IRIS21 and pGBT10-PCR#3. To generate a Gal4-DBD fusion with full-length Cdc25 we first generated a full-length clone in pUC119 that contains a BamHI site in frame with the GAL4 sequence. An EcoRV/SalI fragment isolated from the pCDC25(LEU2)-2 plasmid (Broeket al. 1987) was cloned into the EcoRV/SalI site of pUC119-I21N (see above) to generate pUC119/I21 full length. We partially digested puc119/I21 with BamHI and SalI and subcloned into a pGBT10 plasmid (pGBT10-CDC25). To generate a Gal4(DBD)-Cdc25877–1589 fusion we subcloned a BglII/SalI fragment from pGBT10-CDC25 into the vector pGBT11 (Van Aelst 1997) to generate pGBT11-CDC25. The BglII site in the full-length CDC25 is in frame with the GAL4 sequence.
Plasmids for mapping the activating domain of Cdc25 were generated as follows: the plasmids pCDC25HS20-BamHI/SalI, pCDC25HS20-ApaI/BamHI, pCDC25HS20-bglII, and pCDC25HS20-PstI were constructed first by digestion of the plasmid pCDC25HS-(LEU2)-20 (Broeket al. 1987) with the respective enzymes and gel purifying the plasmids. These were then ligated to gel-purified fragments isolated from digests of pCDC25(LEU2)-1 (Broeket al. 1987) that contain the wild-type allele of the CDC25 gene.
Preparation of cell extracts for analysis of protein expression, fractionation, and immunoprecipitation: For protein expression, whole cell extracts were prepared from the IR-1 strain expressing HA-tagged truncated Cdc25 proteins. Cells were grown at 30° to mid-log phase (OD600 = 0.8 to 1.2) in 10 ml of selective media. The cell pellet was washed once in Buffer A (20 mm Tris-HCl, pH 7.5, 50 mm NaCl, 2 mm phenylmethylsulfonyl fluoride, 50 μg/ml aprotinin, 10 μg/ml leupeptin, 8 μg/ml pepstatin A) and then resuspended in Buffer A + 1% NP-40. Cells were then lysed by vortexing with glass beads. Samples were centrifuged in an Eppendorf centrifuge at 250 × g for 5 min, the supernatant was removed, transferred to a new tube, and centrifuged at 16,000 × g for 30 min. SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer was added to an equivalent amount of protein and subjected to electrophoresis. Protein concentration was determined using a bicinchoninic acid (BCA) protein assay reagent (Pierce). After electrophoresis in a 10% gel, the proteins were electroblotted onto nitrocellulose membranes (Amersham), probed with the α-HA mouse monoclonal antibody (12CA5) at a dilution of 1:5000 (Babco), and developed using sheep anti-mouse alkaline phosphatase-conjugated antibodies and ECL chemiluminescence detection kit (Amersham).
For fractionation experiments, cell pellets from SP1 or TMRV-25 strains expressing HA-tagged Cdc25 truncated proteins were treated as above except that they were resuspended in 400 μl of Buffer A without detergent. Cells were lysed with glass beads at 4° and then centrifuged at 250 × g for 5 min to remove unlysed cells. The supernatant was removed and split into two aliquots. The first aliquot represents the total cell lysate (T) while the second aliquot was subjected to ultracentrifugation at 100,000 × g for 1 hr at 4°. The supernatant (S) was removed and saved while the membrane pellet (P) was resuspended in Buffer M (20 mm HEPES, pH 7.4, 250 mm sucrose). Equivalent amounts of each sample were resolved by SDS-PAGE (7.5% gel) and analyzed by immunoblotting as described above.
For immunoprecipitations membrane extracts were prepared from SP1 cells expressing HA-tagged Cdc25tru (amino acids 1–875) or TT1A-1 expressing the C terminus of Cdc25 (amino acids 877–1589) (Broeket al. 1987). Membrane pellets were prepared as described above and then resuspended in 250 μl of 2 mm EDTA, pH 12 (Gross et al. 1992a,b; Garreauet al. 1996), followed by centrifugation at 16,000 × g for 30 min at 4° in an Eppendorf centrifuge. The supernatant was neutralized by addition of 75 μl of 1 m sodium phosphate, pH 6.5 (Garreauet al. 1996) and used immediately. Equivalent amounts of each sample were resolved by SDS-PAGE and analyzed by immunoblotting as described above to determine the total amount of Cdc25 subjected to immunoprecipitation. For immunoprecipitation 20 μl from each sample were combined and diluted in 200 μl of a modified HNT buffer (Gross et al. 1992a,b; 50 mm N-2-hydroxyethyl piperazine, N′-2-ethanesulphonic acid, pH 7.4, 0.5% Triton X-100, and 60 mm NaCl containing protease inhibitors). The sample was incubated at 4° for 30 min and then immunoprecipitated for 1 hr by addition of α-HA antibody or α-Cdc25 antibody followed by incubation with protein A-Sepharose for 1 hr at 4°. Immunoprecipitates were washed three times with HNT buffer and the protein was eluted and subjected to SDS-PAGE. After electrophoresis in a 7.5% gel, the proteins were electroblotted onto nitrocellulose membranes, probed with α-HA 12CA5 antibody as described above or with α-Cdc25 antibody developed using sheep anti-rabbit alkaline phosphatase-conjugated antibodies and the ECL chemiluminescence detection kit.
Isolation of a truncated form of the CDC25 gene: We performed a screen to isolate genes that in high copy number suppress the heat-shock-sensitive phenotype of an ira1Δ strain (see materials and methods). This strain is heat-shock sensitive as a result of an increase in the levels of GTP-bound Ras caused by deletion of IRA1, a gene coding for a yeast GAP (Tanaka et al. 1989, 1990a). One of the genes isolated (pIRIS21) is a truncated form of CDC25 (CDC25tru). Sequence analysis determined that the isolated gene contains nucleotides from position −243 to 3254 according to the published sequence by Broek et al. (1987). CDC25tru encodes a protein containing amino acids 1–1087; the full-length protein is 1589 amino acids in length. The minimal region coding for the catalytic domain of Cdc25 that is necessary for function in intact cells has been mapped to amino acids 1102–1589 (Laiet al. 1993). Since the truncated form of Cdc25 does not contain the residues necessary for binding to Ras proteins or for catalytic activity we speculated that CDC25tru inhibits the function of Cdc25 and not the function of Ras.
To test this hypothesis we determined whether overexpression of CDC25tru could suppress the heat-shock sensitivity of strains whose phenotype depends on the presence of the CDC25 gene. The heat-shock-sensitive phenotype of an ira1Δ or an ira2Δ (a second GAP in yeast, Tanaka et al. 1990a,b, 1991) strain has been shown to be suppressed by deletion of CDC25 (Tanaka et al. 1989, 1990b; see Figure 1A). In contrast, deletion of CDC25 in a RAS2val19 (Broeket al. 1987) or an ira1Δ ira2Δ (Figure 1A and Junget al. 1994) background results in strains that are still heat-shock sensitive, indicating that CDC25 function is not required under these conditions. The RAS2val19 mutation is equivalent to the activating mammalian rasval12 mutation and renders the protein constitutively active (Broach 1991; Lowy and Willmusen 1993). Further support for the lack of a contribution from CDC25 in a RAS2val19 or an ira1Δ ira2Δ background comes from our studies using a dominant-negative form of ras, H-rasala15, which acts similarly to the N17H-ras mutant in mammalian cells (Powerset al. 1989). This mutant H-rasala15 protein binds and blocks the activity of Cdc25 and can suppress the heat-shock-sensitive phenotype of an ira1Δ or an ira2Δ strain but not of the RAS2val19 or the ira1Δ ira2Δ strains (Figure 1B). As positive control in this assay we used PDE2 (encoding a cAMP phosphodiesterase) that can suppress the heat-shock sensitivity in all strains (Sasset al. 1986) and IRA2 that suppresses the heat-shock sensitivity of strains expressing the wild-type RAS2, ira1Δ, ira2Δ, or ira1Δ ira2Δ deletion strains, but not the heat-shock sensitivity of strains expressing the mutant-activated RAS2val19.
Consistent with our hypothesis, CDC25tru suppresses the heat-shock-sensitive phenotype only in the strains in which deletion of CDC25 or inhibition of the activity of Cdc25 leads to heat-shock resistance. Figure 2 shows that CDC25tru can suppress the heat-shock-sensitive phenotype of an ira1Δ or ira2Δ but not of the double ira1Δ ira2Δ deletion mutant or a strain expressing the activated RAS2val19 mutant.
Furthermore, overexpression of the CDC25tru in a cdc25-5 temperature-sensitive strain exacerbates its defect, making it temperature sensitive for growth at 33° (Figure 3A). In contrast, it has no effect on the temperature sensitivity of a ras2 temperature-sensitive mutant strain (Figure 3B). In this strain, overexpression of mammalian NF1, a Ras-GAP, induces a growth defect at 33°. Taken together, these results suggest that CDC25tru interferes with the ability of Cdc25 to activate Ras.
Mapping the functional domain of CDC25tru: To determine the minimal region of the truncated CDC25 necessary for function we constructed various plasmids that express CDC25tru (coding for amino acids 1–1087) or smaller fragments of the gene under the control of the strong PGK promoter (see materials and methods). The results summarized in Figure 4A show that deletion of amino acids 876–1087 has no effect on the ability of the truncated protein to suppress the heat shock sensitivity of the iraΔ strain, but that further deletion of amino acids 506–875 abolishes its activity.
To further map the contribution of the N terminus to the function of the truncated protein we made three additional constructs coding for amino acids 181–875, amino acids 347–875, or amino acids 181–684, all lacking the SH3 domain of the Cdc25 protein. Whereas the first two constructs are functional in the heat-shock assay, the latter failed to suppress the heat-shock sensitivity of the ira1Δ strain. Expression of the constructs was confirmed by Western blot analysis (Figure 4B). The lack of rescue with fragment 507–1050 may be due to lack of expression of the protein product as determined by Western blot analysis using antibodies raised against amino acids 877–1050 in the C terminus of Cdc25 (not shown). Taken together, the minimum functional domain in Cdc25 essential for suppressing the heat-shock-sensitive phenotype is comprised between amino acids 347 and 875.
Interaction of Cdc25tru with Cdc25: To establish if Cdc25tru exerts its effects by interaction with the endogenous Cdc25 protein we performed two-hybrid analyses using constructs expressing various domains of Cdc25 (see Figure 5). We tested whether there is interaction between the N terminus (1–875) and the C terminus (877–1589) or with the full-length protein (1–1589). We also tested for interaction between N terminus (1–875) and N terminus (1–875). As shown in Figure 5, the N terminus of Cdc25 showed interaction with full-length Cdc25, the C terminus of Cdc25, as well as the N terminus itself. Deletion of the SH3 domain (protein containing amino acids 181–875) does not abolish the interaction, as can be seen in Figure 5. These results suggest that Cdc25tru can interact with endogenous full-length Cdc25 and that this interaction may explain its inhibitory effects.
If Cdc25tru interacts and inhibits the activity of the endogenous protein we would expect it to co-localize with Cdc25. The Cdc25 protein has been shown to be localized to the membrane fraction (Joneset al. 1991; Grosset al. 1992) and this localization was dependent on the C terminus of the protein (Garreauet al. 1996), which does not contain amino acids present in Cdc25tru. We prepared total cell extracts from strains expressing various HA-tagged Cdc25tru proteins, as described in materials and methods, and centrifuged the extracts at 100,000 × g for 1 hr at 4°. Equivalent amounts of total cell extract, supernatant, and pellet fractions were resolved by electrophoresis followed by immunoblot analysis using the α-HA antibody as a probe. As shown in Figure 6, a substantial amount of the N terminus of Cdc25 (1–875) fractionates with the membrane pellet in yeast. Similar results are obtained with deletion constructs lacking the SH3 domain (amino acids 181–875) and also a construct expressing amino acids 1–505 (not shown). Unlike the catalytic domain of Cdc25, which can only be extracted from the membrane with EDTA pH 12 (Garreauet al. 1996), the N-terminal fragments are soluble in buffer containing detergent (see Figure 4B).
To confirm our two-hybrid data we also performed immunoprecipitation analyses using membrane extracts from cells expressing HA-tagged N terminus (1–875) or C terminus of Cdc25 (877–1589) (Broeket al. 1987). The C terminus can be detected by a polyclonal antibody (α-Cdc25) raised against a peptide containing amino acids 877–1050 (see Figure 7A). This antibody does not recognize the HA-tagged N terminus of Cdc25 (Figure 7B, compare α-Cdc25: total N terminus, lane 7 vs. total C terminus, lane 8). The N terminus can be detected with α-HA antibodies that cannot recognize the C terminus of Cdc25 (Figure 7B, compare α-HA: total N terminus, lane 9 vs. total C terminus, lane 10).
We prepared membrane extracts from each cell type as described in materials and methods and combined the extracts expressing: (1) N terminus and C terminus, (2) N terminus and control (vector), and (3) control (vector) and C terminus. We then immunoprecipitated with no antibody, α-HA, or α-Cdc25 followed by Western blot analyses with α-Cdc25 antibody. A faint band that migrates close to the C terminus of Cdc25 can be observed in most of the immunoprecipitates, including the control sample where no antibody has been added (Figure 7B, lanes 1–6; 7C, lanes 1 and 2). As shown in Figure 7B, the α-HA antibody immunoprecipitates a band that can be detected with antibody to the C terminus of Cdc25 in extracts containing HA-tagged N terminus and C terminus of Cdc25 (Figure 7B, lane 2). In contrast, this band is not detected in extracts containing the HA-tagged N terminus (Figure 7B, lane 5) or the C terminus of Cdc25 (Figure 7C, lane 2) alone.
These results suggest that the ability of Cdc25tru to interfere with the in vivo function of wild-type Cdc25 may be a result of its interaction with endogenous Cdc25, thereby either preventing dimerization of wild-type Cdc25 or perturbing the interaction with Ras.
A dominant-activated Cdc25 protein has a mutation in the N terminus: We had previously found a mutationally activated allele of CDC25, CDC25HS20 (Broeket al. 1987). This dominant-acting mutant is capable of inducing phenotypes similar to those observed in strains that express an activated RAS2val19 mutant, namely, the ability to induce sensitivity to heat shock in a wild-type strain. Since our data suggest that the N terminus plays a role in regulating Cdc25 activity we sought to determine if activation of the CDC25HS20 allele results from a mutation in the same region. To first map the region that contains the mutation, we replaced various domains of the mutant CDC25HS20 allele with sequences isolated from the wild-type CDC25 gene. The different constructs are shown in Figure 8A. We first determined that the resulting CDC25 genes were functional by testing for the ability to suppress the temperature sensitivity of a cdc25-5 mutant. All constructs tested positive in this assay (not shown). We then tested the constructs for the ability to induce heat-shock sensitivity in the cdc25-5 strain. The results in Figure 8A show that the construct containing amino acids 353–877 from the wild-type gene (CDC25HS20-BglII/BglII) loses the ability to induce heat-shock sensitivity of the cdc25-5 strain. To determine the nature of the mutation in this domain of the CDC25 gene we sequenced the BglII fragment. Analysis of the sequence indicates the presence of only one change, a C to T substitution at position 1094 of the nucleotide sequence (Figure 8B). This mutation changes the codon from a TCT encoding a serine to TTT that encodes a phenylalanine at position 365 of the amino acid sequence. These results may imply that phosphorylation of serine 365 in wild-type Cdc25 negatively influences Cdc25 function; however, this remains to be tested.
To determine if the mechanism of activation of the CDC25 gene has any relation to the mechanism related to the function of the truncated form of the gene, we tested whether the activating mutation has any effect on the ability of CDC25tru to suppress the heat-shock sensitivity of an ira1Δ strain. We introduced the activating mutation in CDC25tru by replacing the BglII fragment with the equivalent fragment isolated from the CDC25HS20 allele and tested a total of four independent constructs. The result from one of the constructs is shown in Figure 9 and demonstrates that the activating mutation has no effect on the ability of the CDC25tru to suppress the heat-shock-sensitive phenotype of an ira1Δ strain.
Our results show that the N terminus of Cdc25 plays a crucial role in regulating its activity. When expressed in yeast the N terminus of Cdc25 acts as a dominant-negative mutant (Figures 1, 2, 3 and 4). Interestingly in this regard, overexpression of the N terminus of mammalian Ras-GEFs, mSos, and Ras-GRF1/CDC25Mm was also observed to cause dominant-negative effects even though they do not share sequence homology with the yeast Cdc25 protein (Byrne et al. 1996; Zippelet al. 1996; Chenet al. 1997; Qianet al. 1998).
What is the mechanism by which the truncated form of Cdc25 inhibits the function of the full-length protein? At least two models can be proposed. The truncated Cdc25 can bind the wild-type Cdc25 protein, hindering its normal function. Alternatively, the N-terminal domain of Cdc25 can interact with an upstream regulatory element that activates Cdc25, preventing its interaction with the full-length protein. Evidence for oligomerization comes from our observation that in the yeast two-hybrid assay the N terminus can interact with either the N terminus or the C terminus of Cdc25 (Figure 5). Furthermore, immunoprecipitation assays demonstrate that an antibody specific for the N terminus coprecipitates the C terminus (Figure 7). Taken together the data suggest that the ability of Cdc25tru to interfere with the in vivo function of wild-type Cdc25 is likely a result of its interaction with endogenous Cdc25, thereby preventing dimerization of wild-type Cdc25 molecules, folding of each Cdc25 molecule interfering with N terminus/C terminus interactions, or perturbing the interaction with Ras. At this point our data do not discriminate between these possibilities.
Dimerization of Cdc25 has been shown previously (Camuset al. 1997). The C-terminal catalytic domain of Cdc25 interacts with the C-terminal catalytic domain of both Cdc25 and Sdc25. Our results suggest that in addition to the C-terminal catalytic domain, residues in the N terminus may contribute to the oligomerization state of the Cdc25 protein in yeast. Oligomerization has been recently demonstrated for the N terminus of mammalian Ras-GRF1. The N terminus of Ras-GRF1 is characterized by the presence of a pleckstrin homology domain 1 (PH1), a coiled-coiled motif, a calcium-binding ilimaquinone domain (IQ), a Dbl homology domain (DH), and a second PH domain (PH2) (Farnsworthet al. 1995; Quilliamet al. 1995). A yeast two-hybrid screen using the DH domain of Ras-GRF1 as a bait identified a cDNA coding for the DH domain with its adjacent IQ domain of the Ras-GRF1 homolog, Ras-GRF2 (Famet al. 1997; Anborghet al. 1999). The DH domain is sufficient for formation of both hetero- and homo-oligomers. A mutation in the DH domain of Ras-GRF1 that abolished oligomerization impairs biological activity, whereas a deletion mutant that forms oligomers is still biologically active (Anborghet al. 1999). Although this suggests a correlation between oligomer formation and biological function, it should be taken with caution because the effect may be due to a loss of activity of the DH domain. It has been recently shown that the DH domain of Ras-GRF1 acts as a Rac-GEF (Kiyonoet al. 1999).
Further evidence of the role of the N terminus of Cdc25 on the regulation of its activity is provided by our finding that a mutation in the serine residue at position 365 (Figure 8) activates Cdc25, inducing a phenotype similar to activating mutants of other genes in the RAS/adenylyl cyclase pathway (Broeket al. 1987). This would suggest that the serine at position 365 plays a role in negatively regulating Cdc25 activity in yeast. Interestingly, the serine at position 365 is a putative phosphorylation site for protein kinase C. In this regard, a yeast homolog of protein kinase C, Pkc1, has been isolated and shown to function in maintenance of cell wall integrity and in the stress response (Cidet al. 1995).
A negative effect of the N terminus on the catalytic activity of Ras-GEFs has been documented. In vivo, deletion of the DH and PH2 domain of Ras-GRF1 renders the protein constitutively active (Buchsbaumet al. 1996). In vitro, the catalytic activity of a truncated Ras-GRF1 lacking the N terminus is much higher than the full-length protein (Baouzet al. 1997). Similar results have been obtained with mSos1. The N terminus of mSos1 is characterized by the presence of a DH domain, with only 30% similarity to the Ras-GRF1 DH domain, followed by a PH domain (Quilliamet al. 1995). The noncatalytic C-terminal domain of mSos1 is characterized by a proline-rich sequence that interacts with the adapter protein Grb2 (Quilliamet al. 1995). In vivo, deletion of the amino terminus activates the transforming potential of mSos1 (Corbalan-Garciaet al. 1998). In vitro, mSos1 protein lacking either the amino or the carboxy noncatalytic terminus, or both, displays a GEF activity significantly higher compared to the full-length protein. This suggests that both the N and C terminus exert a negative control on the interaction of the Sos catalytic domain and Ras (Corbalan-Garciaet al. 1998). Similar in vitro experiments have not been performed with Cdc25, probably because of inability to express high levels of intact full-length protein (Laiet al. 1993; Camuset al. 1997).
Much work has been done to decipher the mechanism by which Ras-GEFs catalyze the exchange of GDP for GTP by biochemical, mutational approaches and, more recently, determination of the crystal structure of H-ras with the catalytic domain of mSos1 (Boriak-Sjodin et al. 1998). In contrast, we know little about the mechanism by which sequences N terminal to the catalytic domain modulate the function of Ras-GEFs. Studies of Ras-GEFs from all organisms, including yeast, are revealing similar characteristics for their role in vivo despite the lack of conservation at the amino acid level. Further studies of yeast Cdc25 will help us formulate new paradigms that may be relevant to the function of Ras-GEFs from all eukaryotes.
We thank S. Elledge for the pACT2 plasmid. This work was supported by funds from the National Science Foundation, the American Cancer Society (JRFA) to R.B., the Santa Barbara Cottage Hospital, and the California Cancer Research Program. L.V.A. is supported by funds from the National Institutes of Health and the V-Foundation and is a recipient of the Sidney Kimmel Foundation for Cancer Research Award. This work was initiated in the laboratory of M. Wigler, who is supported by funds from the National Institutes of Health. We thank him for his generosity.
Communicating editor: M. Carlson
- Received October 18, 1999.
- Accepted December 16, 1999.
- Copyright © 2000 by the Genetics Society of America