Novel Role for the C Terminus of Saccharomyces cerevisiae Rev1 in Mediating Protein-Protein Interactions^▿

Yeast strains.

Strains are listed in Table 1. The strains used in this study are derivatives of W1588-4C (MATa leu2-3,112 ade2-1 can1-100 his3-11,15 ura3-1 trp1-1 RAD5) and W1588-4A (MATα leu2-3,112 ade2-1 can1-100 his3-11,15 ura3-1 trp1-1 RAD5), which are W303 strains corrected for RAD5 (47). Strains YLW20, YSD3, and YSD6 were constructed by gene replacement by PCR amplifying the rev1::kanMX4, rev7::kanMX4, and rev3::kanMX4 cassettes from yeast deletion strains YOR346W, YIL139C, and YPL167C, respectively (gifts from L. Samson, MIT), transforming the PCR products into the strain W1588-4A, W1588-4C, and YSD5, respectively, and selecting the transformants on YPD plates (38) containing 200 μg/ml G418 disulfate (Sigma). To construct strain YLW35, REV1 was PCR amplified along with its native promoter and 3′ untranslated region from the strain W1588-4C and subcloned into the low-copy pRS416 plasmid (39). The BRCT region was mutated using the QuikChange site-directed mutagenesis kit (Stratagene). The mutated rev1-1 fragment was subsequently subcloned into the integrating plasmid, pRS306 (39), digested with the restriction enzyme, SexAI, and transformed into strain YLW20 using standard transformation techniques (8). Transformants were selected on SD medium (38) lacking uracil.

Strains YSD1 and YSD2 were constructed as follows. The KpnI and NotI fragment from the pAS311 construct (see below), containing either the GAL1 promoter alone or the GAL1 promoter together with the fragment encoding the Rev1 C-terminal 239-amino-acid region was ligated into the pRS404 plasmid (39). The resulting plasmid was digested with SnaB1 and transformed into strain W1588-4A harboring REV1 tagged at its endogenous locus with protein A. Transformants were selected on SD medium lacking tryptophan.

Strains YSD4 and YSD5 were constructed as follows. Cassettes containing the 13-Myc or 9-Myc sequence were amplified from the pFA6a-13Myc-His3MX6 (27) and pYM6-KlTRP (17) plasmids, respectively, using the forward primer 5′-GAGTGTATTCTCAATATGAAGAGGGCGAGAGCATTTTTGGATCTTTGTTTCGGATCCCCGGGTTAATTAA-3′ and reverse primer 5′-ACTTAGAGACATTTAATTTTAATTCCATTCTTCAAATTTCATTTTTGCACGAATTCGAGCTCGTTTAAAC-3′ (13-Myc) and the forward primer 5′-GTGTATTCTCAATATGAAGAGGGCGAGAGCATTTTTGGATCTTTGTTTCGTACGCTGCAGGTCGAC-3′ and reverse primer 5′-TAGAGACATTTAATTTTAATTCCATTCTTCAAATTTCATTTTTGCACATCGATGAATTCGAGCTCG-3′ (9-Myc). The 13-Myc cassette was transformed into strain YLW70, and transformants were selected on SD medium lacking histidine. Strains transformed with the 9-Myc-KlTRP cassette were selected on SD medium lacking tryptophan. For all fusion constructs, integrations at the appropriate genomic locus were confirmed by PCR and expression of the fusion proteins of expected molecular weights by immunoblotting with anti-Myc (Upstate) or PAP antibodies (Sigma).

Plasmids.

For overexpression studies, PCR-amplified REV1 fragments encoding the Rev1 C-terminal 426-amino-acid fragment (amino acids 560 to 985), the Rev1 C-terminal 239-amino-acid region (amino acids 747 to 985), the Rev1 PAD (amino acids 567 to 767), and the Rev1-BRCT region (amino acids from 155 to 268) were subcloned into the 2μm pAS311 plasmid that contains the GAL1 promoter and an HA-10HIS epitope tag (a gift from S. P. Bell, MIT). For all PCR amplifications of REV1, the pJN60-GST REV1 template was used (a gift from D. Hinkle, University of Rochester). REV7 was PCR amplified using the pMR7 template (from D. Hinkle, University of Rochester).

The pESC-Split TAP 2μm plasmid containing the oppositely oriented GAL1 and GAL10 promoters (used for experiments for which results are shown in Fig. 4 and 5), a kind gift from S. P. Bell (MIT), was constructed by PCR amplifying the coding sequence of the calmodulin-binding peptide (CBP) tag, a thrombin cleavage site, and the restriction sites NotI and AscI and inserting them into the NotI and PacI sites of the pESC-TRP plasmid driven by the GAL1 promoter. Similarly, the PCR-amplified coding sequence of the Tobacco etch virus cleavage site, protein A, and the restriction sites FseI and AsiSI were inserted into the BamHI and XhoI sites in front of the GAL10 promoter. A plasmid map is available on request. Appropriate REV1 fragments and REV7 were then inserted into this plasmid in frame with the protein A (ProA) and CBP tags, respectively.

MMS sensitivity and UV mutagenesis assays.

For methyl methanesulfonate (MMS) sensitivity assays, yeast strains transformed with plasmid constructs were grown in SD medium lacking tryptophan and supplemented with 2% raffinose for 2 days at 30°C. They were then subcultured into SD medium lacking tryptophan, supplemented with 2% galactose, and grown to 1 × 10⁷ cells/ml. Appropriate dilutions were plated on SD medium plates lacking tryptophan supplemented with 2% galactose and the indicated amounts of MMS. Colonies were counted after 4 days at 30°C.

For UV mutagenesis assays, three independent colonies of strains YSD1 and YSD2 were grown for 2 days in SD medium lacking tryptophan and supplemented with 2% raffinose. They were then inoculated into synthetic complete (SC) medium with 2% galactose and grown to ∼1 × 10⁸ cell/ml. Appropriate dilutions were plated on SC or SD lacking adenine medium plates with 2% galactose. Plates were irradiated at 1 J/m²/s using a G15T8 UV lamp (General Electric) at 254 nm. ADE2 revertants were counted after 7 days of growth at 30°C in the dark. The frequency of ADE2 reversion was calculated by subtracting the spontaneous value from the frequency at the 10-J/m² UV dose.

Immunoprecipitations and Immunoblot analysis.

For the Rev1-Rev7 interaction studies, the W1588-4C strain transformed with the plasmids described in the legend to Fig. 4, were initially grown for 2 days in SD medium lacking tryptophan supplemented with 2% raffinose at 30°C. The cells were then subcultured and grown overnight at 30°C in SD medium lacking tryptophan and supplemented with 2% galactose to 1 × 10⁷ cells/ml. The cells were centrifuged (4,000 × g, 5 min at 4^° C) and washed once in water. All further steps were carried out at 4°C. Lysis and immunoprecipitations were carried out essentially as described previously (6), with some modifications. Briefly, cell pellets were resuspended in 0.5 to 1 ml ice-cold lysis buffer (50 mM HEPES, pH 7.7, 0.15 M NaCl, 1 mM Na-EDTA, 1% Triton X-100, 0.1% sodium deoxycholate supplemented with 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and a protease inhibitor cocktail [Roche]) and lysed using a bead beater with 0.5 g of glass beads per sample. The lysate was centrifuged at 13,500 × g for 7 min. The supernatant was mixed with 50 μl immunoglobulin G (IgG)-Sepharose beads (Pharmacia) that were preequilibrated in lysis buffer. The supernatant and IgG beads were incubated on a rotating wheel for 2 h. The beads were washed three times with lysis buffer, and bound proteins were eluted using sodium dodecyl sulfate (SDS) loading buffer. For the immunoprecipitations described in the legend to Fig. 6, cells were lysed as described above and anti-Myc or anti-hemagglutinin (anti-HA) antibodies were used for immunoprecipitations as described previously (13).

To prepare cell extracts for the experiments described in the legends to Fig. 1D, 2D, and 2F, protein was extracted by trichloroacetic acid precipitation (17).

For immunoblotting, the protein samples were separated on SDS-polyacrylamide gel electrophoresis (PAGE), transferred to a polyvinylidene difluoride membrane (Immobilon-P; Millipore), and probed with appropriate antibodies. Myc and HA-tagged proteins were detected using the mouse monoclonal antibody clone 4A6 (Upstate) and mono HA.11, clone 16B12 antibody (Covance), respectively. Rev7-CBP and ProA-tagged proteins were detected using the rabbit anti-CBP antibody (Upstate) and the rabbit PAP antibody (Sigma), respectively. All antibodies were used at a 1:1,000 dilution.

Overproduction of the PAD-containing C-terminal 426-amino-acid fragment of yeast Rev1 confers a strong dominant-negative phenotype on survival after DNA damage.

Recent studies in higher eukaryotes have demonstrated interactions between the C terminus of Rev1 and numerous translesion DNA polymerases (9, 29, 34, 40). These include Polκ, Polι, Polλ, Polη, and the Polζ accessory subunit, Rev7. We reasoned that, if the C terminus of yeast Rev1 was involved in mediating protein-protein interactions in a manner analogous to its higher eukaryotic counterparts, its overproduction would interfere with the activity of critical DNA damage resistance proteins. We expected that this would manifest as a severe defect in the survival of wild-type cells exposed to DNA damage. We therefore ectopically overproduced the C-terminal 426-amino-acid fragment of Rev1, including a portion of the thumb, linker, and PAD (30), tagged with HA, from a high-copy plasmid under the control of a GAL-inducible promoter. After induction in galactose, we assessed survival on selective medium plates containing various amounts of the DNA alkylating agent MMS. As shown in Fig. 1A, overproduction of the C-terminal 426-amino-acid fragment of Rev1 causes a decrease in the survival of otherwise wild-type cells exposed to MMS relative to the empty expression plasmid. This result strongly indicates that this region of Rev1 interferes with the activity of proteins involved in resistance to MMS. The decrease in survival after MMS treatment is due to a dominant-negative effect of the Rev1 C terminus, as overexpression in the rev1Δ strain (Fig. 1B) does not further sensitize the strain to MMS. Interestingly, overproduction of the Rev1 C terminus in the rev1-1 strain, which harbors a G193R point mutation in the BRCT motif, shows a modest exacerbation of the sensitivity of this strain to MMS (Fig. 1C), suggesting that the rev1-1 strain, already substantially defective in Rev1 activity, is further inhibited by overexpression of the C-terminal Rev1 fragment. Immunoblot analysis of lysates from the wild-type, rev1Δ, and rev1-1 strains using an anti-HA antibody confirmed the overexpression of the Rev1 C-terminal 426-amino-acid fragment (Fig. 1D).

The Rev1 PAD is not required for the dominant-negative phenotype of Rev1 C terminus-overproducing strains.

Recent experiments using purified proteins has demonstrated a requirement for the PAD of Rev1 to mediate a direct physical interaction with Rev7, the accessory subunit of yeast Polζ (1). To test whether overproduction of the extreme C terminus of yeast Rev1 (excluding the PAD) also confers a dominant-negative effect, we similarly expressed the last 239-amino-acid fragment of Rev1 in the wild-type, rev1Δ, and rev1-1 strains and plated the cells on selective medium plates containing MMS. As shown in Fig. 2A to C, overproduction of the C-terminal 239-amino-acid region of Rev1 causes a similar sensitization of wild-type yeast cells, a slight suppression of the MMS sensitivity of the rev1Δ strain, and a moderate exacerbation of the sensitivity of the rev1-1 strain. Immunoblot analysis of lysates from these strains with an anti-HA antibody revealed that the Rev1 C-terminal 239-amino-acid fragment was equally expressed in the wild-type, rev1Δ, and rev1-1 strains (Fig. 2D).

To rule out potential artifacts arising from the presence of the multicopy plasmid, we integrated a construct encoding the HA-tagged Rev1 C-terminal 239-amino-acid fragment under the control of the GAL promoter into the chromosomal trp1-1 locus of wild-type cells. To confirm that this strain still exhibited the dominant-negative phenotype after DNA damage, we induced expression of the Rev1 C-terminal 239-amino-acid region in galactose and spotted the cells on selective medium plates that contained various amounts of MMS. As shown in Fig. 2E (upper panel), overexpression of the fragment encoding the Rev1 C-terminal 239-amino-acid region causes a marked sensitization of the wild-type strain to MMS relative to a strain integrated with the GAL promoter alone. In contrast, repression of the GAL promoter by growing the cells in the presence of glucose abrogates the sensitivity of these cells to MMS treatment, suggesting that the dominant-negative effect on survival is due to overexpression of the Rev1 C terminus (Fig. 2E, lower panel). We verified the induction of the Rev1 C terminus in galactose by immunoblot analysis of lysates from these strains grown in glucose or galactose using an anti-HA antibody (Fig. 2F).

Taken together, the results presented in Fig. 1 and 2 suggest that the C terminus of yeast Rev1 is likely involved in protein-protein interactions that are crucial for the physiological role of Rev1. Intriguingly, we note that the dominant-negative phenotype associated with overproduction of the Rev1 C terminus also partially requires the Rev1 BRCT motif, as we observe a moderate exacerbation of the MMS sensitivity of the rev1-1 strain.

Wild-type yeast overproducing the Rev1 C-terminal 239-amino-acid fragment are defective in DNA damage-induced mutagenesis.

Rev1 function is required for the introduction of DNA damage-induced mutations in the genome, as rev1Δ strains are severely defective in DNA damage-induced mutagenesis (21). As REV1-dependent mutagenesis has been most thoroughly characterized in response to UV, we examined the UV sensitivity and UV-induced mutagenesis in the strain YSD2, which harbors the chromosomally integrated 239-amino-acid segment of Rev1 under the control of the GAL promoter. After induction in galactose, this strain is sensitive not only to MMS (Fig. 2E) but also to UV, relative to the strain YSD1, which carries an integration of the GAL promoter alone (data not shown).

UV irradiation introduces lesions in DNA that are repaired by both error-free and error-prone mechanisms. The efficiency of the error-prone pathway is conveniently studied by assaying the reversion frequency of the ade2-1 allele after UV treatment. To determine if strain YSD2 was also defective in UV-induced mutagenesis, we exposed the cells to a low UV dose and measured the number of revertants per surviving cells. Even though the survival of either strain YSD1 or YSD2 was not significantly diminished after irradiation with this modest UV dose (Fig. 3A), strain YSD2 overproducing the Rev1 C-terminal 239-amino-acid region exhibits a decrease in the frequency of adenine revertants after UV treatment (Fig. 3B). These results strongly suggest that the overproduction of the Rev1 C terminus interferes with the ability of the cells to function in the mutagenic bypass of UV-induced DNA lesions.

REV7 is required for the dominant-negative phenotype of the Rev1 C terminus on cell survival after DNA damage.

The extreme C-terminal 100 amino acids of mouse and human Rev1 interact with Rev7, the accessory subunit of the Rev7-Rev3 complex that together comprise Polζ (9, 29, 34). To determine if yeast REV7 is required for the ability of the overproduced Rev1 C terminus to sensitize cells to DNA damage, we expressed the Rev1 C-terminal 239-amino-acid fragment (tagged with protein A) in the rev7Δ strain and assessed the viability of the cells after MMS treatment. As shown in Fig. 4, expression of the Rev1 C-terminal 239-amino-acid fragment does not exacerbate the sensitivity of the rev7Δ strain after MMS treatment (note that the survival curve for the rev7Δ strain transformed with the plasmid encoding the Rev1 C-terminal fragment overlaps with that of the empty plasmid control and is thus hidden in Fig. 4). This result suggests that the REV1-dependent function that is inhibited by expressing the C terminus of Rev1 is also REV7 dependent.

If the yeast Rev1 C terminus is involved in a functional interaction with Rev7, we predicted that coexpression of the two proteins would interfere with the ability of Rev7 to complement the MMS-sensitive phenotype of the rev7Δ strain. We therefore constructed plasmids under the control of the oppositely oriented GAL1 and GAL10 promoters to coexpress CBP-tagged Rev7 and the Rev1 C-terminal 239-amino-acid fragment tagged with ProA. After induction in galactose, cells were spotted on plates containing various amounts of MMS. As expected, expression of Rev7 alone complements the lethality of the rev7Δ strain after MMS treatment relative to cells transformed with the empty vector control (Fig. 4). However, coexpression of the Rev1 C terminus and Rev7 completely blocks the ability of Rev7 to complement the MMS-induced lethality of the rev7Δ strain. In contrast, coexpression of Rev7 and the full-length Rev1 protein rescues the lethality of this strain in response to MMS. Intriguingly, expression of the full-length Rev1 alone partially rescues the MMS sensitivity of the rev7Δ strain (Fig. 4). These data strongly support our inference that the C terminus of Rev1 is involved not only in a functional interaction with Rev7 but in a physical interaction as well.

The C terminus of Rev1 interacts with Rev7.

To examine whether the Rev1 C terminus and Rev7 interact as components of a complex, we used the plasmid system described above to coexpress Rev7 and the C-terminal 239-amino-acid region of Rev1 (excluding the PAD), the PAD-containing C-terminal 426-amino-acid fragment, or the full-length Rev1 in wild-type cells. Following galactose induction, we performed coimmunoprecipitations using IgG-Sepharose beads followed by immunoblot analysis using an anti-CBP antibody. As shown in Fig. 5A, IgG beads immunoprecipitate Rev7 from cells coexpressing the C-terminal 239-amino-acid region of Rev1 and Rev7 (lane 3), whereas Rev7 is not immunoprecipitated from cells transformed with either the empty vector (lane 1) or Rev7 alone (lane 2). These results thus provide the first demonstration that the extreme C-terminal 239-amino-acid fragment of S. cerevisiae Rev1 (excluding the PAD) can mediate an interaction with Rev7.

Since the dominant-negative phenotype resulting from overproduction of the Rev1 C terminus was also exhibited by the PAD-containing C-terminal 426-amino-acid fragment (Fig. 1), we tested if this region likewise interacts with Rev7. We therefore performed immunoprecipitations from cells coexpressing Rev7 and either the C-terminal 426-amino-acid fragment of Rev1 or the full-length Rev1 protein. Consistent with the observations of Acharya et al. (1), Rev7 is immunoprecipitated from cells coexpressing Rev7 and either the Rev1 C-terminal 426-amino-acid fragment (Fig. 5B) or the full-length Rev1 (lane 4) but not from cells transformed with either the empty vector or Rev7 alone (lanes 1 and 2, respectively). Thus, consistent with an earlier study (1), the C-terminal 426-amino-acid fragment of yeast Rev1 (including the PAD) interacts with Rev7, although our immunoprecipitation experiments do not show that the interaction is direct.

To rule out the possibility that the interaction between Rev7 and Rev1 is mediated indirectly via independent binding of the two proteins to DNA, extracts from cells coexpressing Rev1 and Rev7 were treated before immunoprecipitation with ethidium bromide (EtBr), a reagent that disrupts DNA-protein interactions by distorting DNA structure by intercalation (18). However, despite the presence of the DNA-binding Rev1 PAD, the efficiency of interaction between Rev1 and Rev7 is not influenced by the addition of EtBr (Fig. 5C, lane 3), indicating that the interaction between Rev1 and Rev7 is independent of DNA. MMS treatment of the strains expressing Rev7 and the C-terminal 426-amino-acid fragment of Rev1 does not significantly alter the interaction between the two proteins (Fig. 5C, lane 2), implying that the association between the two proteins occurs independently of DNA damage.

To confirm that the interaction between Rev1 and Rev7 also occurs at endogenously expressed levels, we tagged Rev1 and Rev7 at their C termini with the protein A and 9-Myc epitope tags, respectively, by homologous recombination at the REV1 and REV7 loci. Upon immunoblot analysis using an anti-Myc antibody, we observed two bands specific to Rev7: one of the expected molecular size, Rev7-9Myc (∼45 kDa), and another of a lower molecular size (∼35 kDa) (data not shown). Since we observe similarly sized proteins using a ProA-tagged Rev7 (data not shown) and Rev7-13Myc (Fig. 6B), we speculate that the lower-migrating band is an N-terminal truncation of the Rev7 protein. MMS treatment of the Rev7-9Myc strain does not change the levels of this smaller Rev7 fragment nor is its expression altered during the yeast cell cycle (data not shown). The significance and mechanism of regulation of this truncated Rev7 product is unclear and is currently under investigation.

Interestingly, we note that Rev1-ProA immunoprecipitates only the higher-migrating band of Rev7-9Myc (Fig. 5D, lane 1). The presence of the lower-migrating Rev7 band notwithstanding, we observe an interaction between Rev1 and Rev7 at endogenous expression levels. Pretreatment of the cells with MMS (lane 2) or treatment of the lysates with ethidium bromide (lane 3) does not alter the interaction between the two proteins, suggesting that the interaction between Rev1 and Rev7 is independent of DNA damage or DNA. These data further extend our previous conclusions by demonstrating that the endogenous protein levels of Rev1 and Rev7 are sufficient to mediate the interaction.

The Rev1 BRCT motif also interacts with Rev7.

The yeast Rev1 N-terminal BRCT motif is required for mutagenesis and resistance to DNA damaging agents, as the G193R point mutation in the yeast Rev1 BRCT motif (the rev1-1 mutation) abrogates its function in DNA damage-induced mutagenesis (19). Our results concerning the interaction between full-length Rev1 and Rev7 did not exclude the possibility that the interaction between Rev1 and Rev7 might involve a region of Rev1 distinct from the C terminus. To determine if the N-terminal BRCT region of Rev1 might also be involved in the interaction with Rev7, we expressed either the HA epitope-tagged BRCT region of Rev1 or the Rev1-1 protein in a Rev7-13Myc strain. We performed immunoprecipitations using an anti-Myc antibody and immunoblot analysis using an anti-HA antibody. As shown in Fig. 6A, Rev7-13Myc immunoprecipitates the HA-tagged BRCT region of Rev1 (lane 4). Surprisingly, the Rev1-1 BRCT domain is also pulled down by Rev7-13Myc (lane 6). The Rev1 C-terminal 239-amino-acid fragment used as a positive control also interacts efficiently with Rev7-13Myc (lane 8), whereas cells transformed with the empty expression vector do not immunoprecipitate Rev1 (lane 1). Conversely, we immunoprecipitated the HA-tagged BRCT Rev1, BRCT Rev1-1, and the Rev1 C-terminal 239-amino-acid region with anti-Myc antibodies (Fig. 6B). Taken together, these results thus define the BRCT motif of Rev1 and the extreme C-terminal 239-amino-acid fragment as novel regions necessary for the Rev7 interaction.

The Rev1 PAD interacts with Rev7.

The Rev1 PAD was recently shown to directly interact with Rev7 (1). Using the purified Rev1 and Rev7 proteins and the yeast two-hybrid assay, Acharya et al. demonstrated the involvement of the Rev1 PAD in a direct interaction with Rev7 (1). Although deletion of the PAD supports an interaction with Rev7 (Fig. 5A), we wished to determine if this domain coimmunoprecipitates with Rev7 independent of the extreme C-terminal 239-amino-acid fragment. We therefore created a construct similar to that of Acharya et al. (1) that expresses the HA-tagged Rev1 PAD (amino acids from 567 to 767) in a Rev7-13Myc strain and performed immunoprecipitations from crude extracts using anti-HA and anti-Myc antibodies. As shown in Fig. 6A (lane 10), Rev7 efficiently immunoprecipitates the Rev1 PAD, whereas Rev7 is not immunoprecipitated from cells transformed with the empty vector (data not shown). Similarly, the Rev1 PAD pulls down Rev7-13Myc (Fig. 6B, lane 10), whereas cells transformed with the empty vector control do not (data not shown). This result therefore demonstrates that, consistent, with results using purified proteins, the Rev1 PAD forms a complex with Rev7 (1).

Rev1 interacts with Rev7 independently of Rev3.

To test whether the interaction between Rev1 and Rev7 is mediated by Rev3, the catalytic subunit of Polζ, we expressed the HA-tagged BRCT Rev1, BRCT Rev1-1, the Rev1 PAD, and the C-terminal 239-amino-acid fragment in the rev3Δ strain that also harbors a 13-Myc tag at the REV7 locus. Immunoprecipitations from crude extracts using anti-Myc and anti-HA antibodies reveal that the interaction between Rev7 and (i) the Rev1 BRCT and the Rev1-1 BRCT regions, (ii) the Rev1 PAD, and (iii) the Rev1 C-terminal 239-amino-acid fragment occur even in the absence of Rev3, indicating that Rev3 is not required for the interaction between Rev1 and Rev7 (Fig. 6C and D).