Dominant Mutations in Three Different Subunits of Replication Factor C Suppress Replication Defects in Yeast PCNA Mutants

Corresponding author: Connie Holm, Department of Pharmacology, Division of Cellular and Molecular Medicine, University of California, 9500 Gilman Dr., San Diego, CA 92093-0651., cholm{at}ucsd.edu (E-mail)

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To identify proteins that interact with the yeast proliferating cell nuclear antigen (PCNA), we used a genetic approach to isolate mutations that compensate for the defects in cold-sensitive (Cs^-) mutants of yeast PCNA (POL30). Because the cocrystal structure of human PCNA and a p21^WAF1/CIP1 peptide shows that the interdomain region of PCNA is a site of p21 interaction, we specifically looked for new mutations that suppress mutations in the equivalent region of yeast PCNA. In independent screens using three different Cs^- mutants, we identified spontaneously arising dominant suppressor mutations in the RFC3 gene. In addition, dominant suppressor mutations were identified in the RFC1 and RFC2 genes using a single pol30 mutant. An intimate association between PCNA and RFC1p, RFC2p, and RFC3p is suggested by the allele-restricted suppression of 10 different pol30 alleles by the RFC suppressors. RFC1, RFC2, and RFC3 encode three of the five subunits of the replication factor C complex, which is required to load PCNA onto DNA in reconstituted DNA replication reactions. Genomic sequencing reveals a common region in RFC1p, RFC2p, and RFC3p that is important for the functional interaction with PCNA. Biochemical analysis of the wild type and mutant PCNA and RFC3 proteins shows that mutant RFC3p enhances the production of long DNA products in pol -dependent DNA synthesis, which is consistent with an increase in processivity.

THE proliferating cell nuclear antigen (PCNA) is an essential factor in eukaryotic DNA replication and repair processes, and it may also be an important target for coordinating cell-cycle regulation with DNA replication (reviewed in JONSSON and HUBSCHER 1997 and KELMAN 1997 ). During DNA replication, PCNA markedly enhances the efficiency of incorporation of nucleotides into DNA by securely attaching the DNA synthesis machinery onto the DNA (reviewed in SO and DOWNEY 1992 ; WYMAN and BOTCHAN 1995 ; JONSSON and HUBSCHER 1997 ; KELMAN 1997 ). In vitro DNA replication reactions show that PCNA is first loaded onto the DNA by a complex of five proteins called replication factor C (RFC) in an ATP-dependent reaction (LEE et al. 1988 ; LEE and HURWITZ 1990 ; TSURIMOTO and STILLMAN 1991A ; YODER and BURGERS 1991 ; FIEN and STILLMAN 1992 ; PODUST et al. 1995 ). Next, PCNA binds to the DNA polymerase ( or ), and DNA elongation occurs processively (LEE and HURWITZ 1990 ; BURGERS 1991 ; LEE et al. 1991B ; TSURIMOTO and STILLMAN 1991B ; PODUST et al. 1992 ). Because RFC is also able to unload PCNA from the DNA, PCNA may be shuttled onto and off of Okazaki DNA fragments during lagging-strand DNA synthesis (YAO et al. 1996 ; CAI et al. 1997 ). Therefore, the interaction of PCNA and proteins of the RFC complex may play a crucial role in the execution of DNA synthesis. PCNA has been shown to function in many DNA repair processes, such as base and nucleotide excision repair (NICHOLS and SANCAR 1992 ; SHIVJI et al. 1992 ; MATSUMOTO et al. 1994 ; FROSINA et al. 1996 ), RAD6-dependent error prone repair (TORRES-RAMOS et al. 1996 ), DNA methylation (CHUANG et al. 1997 ), and DNA mismatch repair (JOHNSON et al. 1996 ; UMAR et al. 1996 ). That PCNA might play a role in DNA mismatch repair is particularly intriguing because mutations in known human mismatch repair genes, such as hMSH2, hMLH1, and hPMS2, have been linked to colorectal cancers (LIU et al. 1996 ).

Apart from its role in DNA replication and DNA repair, mammalian PCNA is also the target for the binding of the cyclin-dependent kinase (CDK) inhibitor p21^WAF1/CIP1 (XIONG et al. 1992 , XIONG et al. 1993 ; ZHANG et al. 1993 ; FLORES-ROZAS et al. 1994 ; WAGA et al. 1994 ), which is induced under conditions of DNA damage in a p53-dependent manner (EL-DEIRY et al. 1993 ). Although a p21 homolog has not yet been identified in yeast, the striking similarities in the crystal structures of human and yeast PCNA suggest that there may be similarities in their regulation as well. Biochemical studies show that p21 blocks DNA synthesis by binding to PCNA (FLORES-ROZAS et al. 1994 ; WAGA et al. 1994 ) and that the C-terminal 22 amino acids of p21 are sufficient to inhibit PCNA activity (WARBRICK et al. 1995 ). The determination of the cocrystal structure of human PCNA and the C-terminal p21 peptide has identified the interdomain and interconnector loop regions of the PCNA monomer as the site of interaction of the p21 peptide (GULBIS et al. 1996 ). However, this structure leaves open the question of the mechanism of inhibition of DNA synthesis. While it is possible that p21 binding affects the monomer-trimer ratio of PCNA in the cell, it appears more likely that p21 interferes with the binding of essential DNA replication proteins to PCNA. Candidates for these essential proteins include DNA polymerases or , or subunits of RFC.

We have previously used the Saccharomyces cerevisiae PCNA gene to identify cold-sensitive (Cs^-) mutations that affect the interdomain region of the yeast PCNA protein structure (AMIN and HOLM 1996 ). In vivo analysis of the Cs^- pol30 mutants suggests that the interdomain region of the PCNA monomer is important for both efficient DNA replication and for repair of methyl methanesulfonate (MMS) and UV-induced DNA damage (AMIN and HOLM 1996 ). A comparison of the yeast and human PCNA structures reveals that they are virtually superimposable (KRISHNA et al. 1994 ; GULBIS et al. 1996 ). Thus, it is striking that the p21 peptide makes contacts with the interconnector loop and the interdomain region of human PCNA, the same region in yeast PCNA where our Cs^- mutations are located (KRISHNA et al. 1994 ; GULBIS et al. 1996 ). Furthermore, because cold sensitivity often affects protein-protein interactions (CANTOR and SCHIMMEL 1980 ; STRAUSS and GUTHRIE 1991 ; MCALEAR et al. 1994 ), it is likely that the interdomain region of PCNA is a key region for protein-protein interactions.

To identify proteins that interact with the interdomain region of yeast PCNA in vivo, we looked for suppression of the defects of three Cs^- pol30 mutants by spontaneously arising mutations in other genes. We obtained both intragenic and extragenic suppressors that are dominant for suppression. In independent screens with three different pol30 alleles we obtained extragenic suppressors that affect the RFC3 protein. Additionally, we obtained mutations in RFC1 and RFC2 that suppress the defects of one of the pol30 alleles. Our results suggest that RFC1p, RFC2p, and RFC3p are important for a functional interaction with the interdomain region of yeast PCNA. In particular, we have identified an evolutionarily conserved region in RFC1p, RFC2p, and RFC3p that is important for the functional interaction. Furthermore, biochemical analysis of the wild-type and mutant PCNA and RFC3 proteins shows that the suppressor RFC3p enhances the production of long DNA products in polymerase -dependent DNA synthesis.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and media:
All yeast strains mentioned in this study have an S288c strain background, and they were constructed by using genetic methods described in SHERMAN et al. 1986 . For flow cytometry experiments, rho⁰ strains were created by growing rho⁺ strains CH2165 (POL30), CH2161 (pol30-104), and CH2392 (pol30-104 RFC3-3) in YEPD (see below) containing 25 µg/ml ethidium bromide as outlined in SHERMAN et al. 1986 . A list of all strains used in this study is presented in Table 1.

YEPD (rich) and SD (synthetic dextrose) media were used to grow yeast cells. YEPD medium contains 1% yeast extract, 2% bactopeptone, 2% dextrose, with the presence or absence of 2% bactoagar. SD medium contains 0.67% yeast nitrogen base, 2% dextrose, and 2% bactoagar. SC (synthetic complete) medium contains 60 mg of leucine, 30 mg of lysine, and 20 mg of uracil, adenine, histidine, and tryptophan in 1 liter of SD medium. For MMS-containing plates, MMS (Sigma, St. Louis) was added to autoclaved YEPD at a final concentration of 0.01, 0.015, or 0.02% prior to pouring plates. Sporulation medium contains 1% potassium acetate, 0.1% yeast extract, 2% bactoagar, and 0.05% dextrose.

Pseudoreversion screen:
To obtain pseudorevertants of pol30 mutants, we selected spontaneously arising suppressors of three different cold-sensitive PCNA mutations. Specifically, 10–33 independent cultures of strain CH2159 (pol30-100), CH2161 (pol30-104), or CH2171 (pol30-108) were grown overnight in YEPD at 30°. Approximately 10⁷ cells derived from each of the pol30 mutant cultures were spread on each YEPD plate and incubated at the restrictive temperature of 14° or the semipermissive temperature of 20°. The use of two different temperatures was initially intended to enhance the specificity of the screening process. Spontaneously arising revertant colonies were selected and retested for loss of their Cs^- and Mms^s phenotypes. In addition, the revertants were also examined for the appearance of new phenotypes, including heat sensitivity or sensitivity to hydroxyurea.

Genetic analysis of intragenic and extragenic suppressors:
To determine if suppression of pol30 mutant phenotypes is conferred by a single gene, the suppressed pol30 strains were crossed to a Cs^- strain (CH2159, CH2161, or CH2171) bearing the same pol30 allele as present in the pseudorevertant. Examination of the phenotype of the diploid strain revealed whether suppression is dominant or recessive. Because the diploid strains were homozygous for the pol30 mutant allele, 2:2 segregation of the Cs⁺ phenotype was indicative of a single gene conferring suppression upon tetrad analysis. Further analysis of 24 strains carrying single-gene suppressors was carried out to determine whether suppression was due to intragenic or extragenic mutations. The revertant strains were crossed with a POL30 strain (CH2237), and tetrad analysis revealed that 12 of the 24 revertants were likely to contain intragenic mutations; genomic sequencing was used to confirm that the mutations were in the POL30 gene (mutations and amino acid changes described below) and not the neighboring RFC5 gene on chromosome II.

To determine the number of different genes represented among the remaining 12 extragenic suppressors, we crossed each of the MATa suppressor strains with one or more of the MAT suppressor strains and performed tetrad analysis. The suppressor genes were initially referred to as SOP1, SOP2, SOP3, etc., for suppressor of pol30. To identify the gene product encoded by the SOP genes, we tested whether any of the sop genes were candidates suggested from biochemical studies. Because all of the extragenic suppressors of the pol30-104 mutation affected a single gene (SOP1), we crossed a single extragenic suppressor strain CH2386 (pol30-104 SOP1-1) with strains containing markers to each of the genes encoding subunits of the RFC complex. These strains are CH2237 (RFC1.URA3), CH1785 (ira2::HIS3, which is linked to RFC4), CH756 (cdc8-1, which is linked to RFC2), CH2368 (sec21-1, which is linked to RFC3), and CH595 (POL30, to identify suppressor mutations in pol30 or RFC5). In a cross between strain CH2368 (sec21-1) and CH2386 (pol30-104 SOP1^sup), we found that of the 29 spores that contained an unsuppressed cold-sensitive pol30-104 mutation, 26 of the spores contained the sec21-1 mutation. This result indicated that the SOP1 gene is linked to the SEC21 gene, which is 6 kb away from RFC3 on chromosome XIV. Similarly, two of the extragenic suppressors of the pol30-100 mutation were identified as RFC1-19 in a cross between strains CH2237 (RFC1.URA3) and CH2405 (pol30-100 RFC1-19) or CH2406 (pol30-100 RFC1-19). The remaining extragenic suppressors of pol30-100 and the single extragenic suppressor of pol30-108 were identified as RFC2-10 or RFC3-3 by directly sequencing the genomic copy of these genes.

Flow cytometry:
The extent of suppression of the cold-sensitive defect of pol30 mutant cells was examined by flow cytometry. Briefly, rho⁰ strains CH2253 (POL30), CH2252 (pol30-104), and CH2431 (pol30-104 RFC3-3) were grown to log phase in YEPD at 35°. The asynchronously growing cultures were then divided into two portions; one was incubated at 35° for 3 hr, and the other was incubated at 14° for 24 hr. Samples were collected, sonicated, and then prepared for flow cytometry by staining with propidium iodide (HUTTER and EIPEL 1979 ). For each sample, 10,000 cells were counted to assess the DNA content by a Becton Dickinson (San Jose, CA) cell fluorescence cell sorter.

Allele-restricted suppression of pol30 alleles:
To determine whether suppression was allele-restricted, each SOP allele was tested to determine whether it could suppress various Cs^- and Mms^s mutant alleles of pol30. For this purpose, suppressor strains CH2405 (pol30-100 RFC1-19), CH2385 (pol30-100 RFC2-1), CH2386 (pol30-104 RFC3-1), or CH2387 (pol30-104 RFC3-2) were crossed with strains CH2165 (POL30), CH2158 (pol30-100), CH2179 (pol30-100), CH2159 (pol30-101), CH2180 (pol30-101), CH2162 (pol30-102), CH2161(pol30-104), CH2181 (pol30-104), CH2170 (pol30-106), CH2183 (pol30-106), CH2171 (pol30-108), CH2184 (pol30-108), or CH2369 (pol30-114). Tetrad analysis was performed and the spores were tested for sensitivity to cold (20° or 14°) and sensitivity to different concentrations of MMS (0.01, 0.015, or 0.02%). To test whether the RFC3-3 mutation suppresses pol30 alleles other than pol30-100, pol30-104, or pol30-108, the RFC3 gene or the RFC3-3 gene was amplified from the genome of strain CH2171 (RFC3) or CH2409 (RFC3-3), respectively. The 2.3-kb PCR fragments containing either RFC3 or RFC3-3 were digested with SspI and ligated into a SmaI-digested vector pCH1099 (CEN6 ARS1 URA3). The resulting plasmids pCH1662 (RFC3 URA3) and pCH1663 (RFC3-3 URA3) were each transformed into strains CH2165 (POL30), CH2158 (pol30-100), CH2159 (pol30-101), CH2161 (pol30-104), CH2162 (pol30-102), CH2170 (pol30-106), CH2171 (pol30-108), or CH2369 (pol30-114). The transformants were tested for sensitivity to cold (20° or 14°) and sensitivity to different concentrations of MMS (0.01, 0.015, or 0.02%).

Genomic sequencing:
To identify a nucleotide change(s) between the original pol30, RFC1, RFC2, or RFC3 alleles and their suppressing counterparts, the suppressor genes were sequenced using a PCR-based genomic sequencing technique (Promega, Madison, WI). Briefly, the genomic suppressor gene was amplified using genomic DNA isolated (SHERMAN et al. 1986 ) from each suppressor strain by PCR using Taq polymerase (AmpliTaq; Perkin Elmer, Norwalk, CT). The wild-type POL30, RFC1, RFC2, or RFC3 genes were also amplified from unsuppressed pol30 strains, and they were used as controls in the entire sequencing process. In 20 sequencing runs, we never observed a sequencing gel that showed an apparent mutation in the wild-type sequence. For the PCR amplification of genes, multiple independent PCR reactions were pooled together to prevent an accumulation of single "jackpot" mutations caused from errors that could be introduced by the Taq polymerase. The amplified wild-type and suppressor genes were then sequenced using the cycle sequencing kit from Promega.

Purification of wild-type and mutant RFC:
To purify wild-type and mutant RFC complexes we used a procedure that was similar to one previously described in LEE et al. 1991A . To purify wild-type RFC, protease-deficient strain CH2401 (BJ2168; GERIK et al. 1997 ) was transformed with plasmid pCH1656 (pBL420 containing wild-type RFC1, RFC2, RFC3, RFC4, and RFC5 under the control of GAL1-10 UAS; GERIK et al. 1997 ) and plasmid pCH1655 (pMTL4 containing GAL4 under the control of GAL1 UAS; GERIK et al. 1997 ) to produce strain CH2404. To purify RFC^sup, we constructed strain CH2586 by transforming strain CH2401 with plasmids pCH1669 (RFC3-3; created by replacing RFC3 in pBL413 with a fragment containing the RFC3-3 mutation from plasmid pCH1663 using the NcoI and StuI restriction enzymes; plasmid pBL413 was obtained from Peter Burgers) and pCH1667 (pBL425 containing RFC1, RFC2, RFC4, and RFC5 under the control of GAL1-10 UAS; GERIK et al. 1997 ). The yeast strains were grown in selective media containing 3% glycerol, 2% lactate, and 0.1% glucose for 2 days, followed by a 3-hr incubation in rich media containing 2% lactate, 0.2% glucose, and 2% galactose. Cell noodles were made with 600 g of cells using liquid nitrogen, and they were ground in a motorized grinder under liquid nitrogen. The ground cells were resuspended in buffer A (25 mM Tris-HCl pH 7.5, 10% glycerol, 1 mM DTT, 1 mM EDTA, 0.01% Nonidet P-40, 0.1 mM phenylmethylsulfonyl fluoride, 5 µM pepstatin A, and 5 µM leupeptin) containing 500 mM NaCl, and they were pelleted. The extracts were fractionated first by using ammonium sulfate, which was added to 5% saturation, followed by fractionation with 65% saturation of ammonium sulfate. The pellet was resuspended in buffer A containing 200 mM NaCl to obtain a protein concentration of 5–10 mg/ml. The solution was dialyzed overnight with buffer A and the extracts were subjected to a phosphocellulose P11 column. The column was equilibrated and washed with buffer A containing 200 mM NaCl before elution with buffer A containing 400 mM NaCl followed by buffer A containing 800 mM NaCl. The fractions eluted with buffer A containing 800 mM NaCl were collected and assayed for protein amount using the Bradford assay. The peak fractions were pooled (615 mg protein) and chromatographed further using a hydroxyapatite column, a single-stranded DNA column, and a Q Sepharose column as described in LEE et al. 1991A . We monitored the purification of wild type and mutant RFC using the DNA synthesis assay described below. We obtained 6 mg of protein after the final step.

PCNA and DNA polymerase purification:
To purify wild-type and mutant PCNA proteins we followed the protocol outlined in FIEN and STILLMAN 1992 . The hexahistidine-tagged (his-tagged) expression plasmid containing wild-type POL30, plasmid pCH1695, was obtained from Mike McAlear [pMM115, which has the POL30 gene cloned into a pRSETA vector from Invitrogen (San Diego); BECKWITH et al. 1998 ]. To purify his-tagged Pol30-104p, we replaced a ClaI-HindIII fragment in plasmid pCH1695 with a ClaI-HindIII fragment containing the pol30-104 mutation from plasmid pCH1598 (AMIN and HOLM 1996 ) to produce plasmid pCH1700. To purify S. cerevisiae DNA polymerase (pol ) we used the protocols outlined in BAUER et al. 1988 and ZUO et al. 1997 . Briefly, we obtained a protease-deficient yeast strain RDKY1293 from Richard Kolodner, and we prepared cell extracts to purify native pol . The extracts were subjected to ammonium sulfate fractionation and column chromatography using phosphocellulose P11, Q Sepharose, SP Sepharose, a single-stranded DNA cellulose, and hydroxyapatite columns in a sequential manner. The fractions obtained were assayed for DNA synthesis activity (described below) during the purification process.

DNA synthesis assays:
To carry out the DNA synthesis assay we prepared an end-labeled singly primed DNA template using X174 viral DNA as template (5386 nucleotides; New England Biolabs, Beverly, MA) and a 30-mer primer that anneals to it from nucleotide 5127–5156 as described previously (LEE et al. 1991A ). The DNA synthesis reaction (30 µl) was carried out using 250 fmol of pol , 240 fmol of wild-type RFC or 50 fmol of RFC^sup protein, 1000 fmol of wild-type or mutant PCNA protein, 300 ng of Escherichia coli single-strand binding protein (Pharmacia, Piscataway, NJ), and 300 fmol of primer/template. In addition, the reaction contained 100 µM dNTP, 2 mM ATP, 7 mM MgCl₂, 1 µg BSA, 0.5 mM DTT, and 40 mM Tris pH 7.8. The reaction was incubated at 37° for 30 min and then stopped with 1 µl of 0.1 M EDTA and 2 µl of 6x dye (0.25% bromophenol blue and 30% glycerol). The reactions were loaded onto a 1% agarose gel containing 50 mM NaOH and 3 mM EDTA. The gel was run for 16 hr at 35 V, and it was dried and exposed to film.

The DNA synthesis assay conditions used during the purification process were similar to the assay described above with the exception that 33.3 µM of [³H]dCTP and unlabeled X174 primer/template were used during the reaction instead of ³²P-end-labeled X174 primer/template. The reactions were incubated for 30 min at 37° or 25°. Next, 10 µl of 10 mg/mL salmon sperm DNA, 100 µl of 0.1 M sodium pyrophosphate, and 5 ml of 5% TCA were added to each reaction. The reactions were incubated on ice for 10 min, after which they were poured over 25-mm glass fiber filters (ENZO Diagnostics) to filter out the unincorporated ³H nucleotides. The filters were washed once with 5% TCA, once with 1% TCA, and then with ethanol. The filters were dried and then counted using a scintillation counter.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We screened for proteins that interact functionally with yeast PCNA using the powerful genetic technique of pseudoreversion, or second-site suppression. This approach has been used successfully to identify many gene products, such as -tubulin, which interacts with ß-tubulin, and the actin-binding protein fimbrin (ADAMS and BOTSTEIN 1989 ; OAKLEY and OAKLEY 1989 ). We expected to recover intragenic suppressors affecting the pol30 gene itself using this method. More important, if an interaction with another protein is affected by the original pol30 mutation, we hoped to recover extragenic suppressors that were dominant and allele-restricted for suppression (ADAMS and BOTSTEIN 1989 ; MCALEAR et al. 1994 ; SANDROCK et al. 1997 ).

We selected spontaneously arising mutations that can compensate for the DNA replication and DNA repair defects of PCNA mutants. To isolate suppressors of Cs^- pol30 mutants, we incubated strains CH2159 (pol30-100), CH2161 (pol30-104), and CH2171 (pol30-108) at the semipermissive (20°) or nonpermissive temperature (14°). Spontaneously arising Cs⁺ colonies (revertants) were selected and analyzed genetically. All of the suppressor mutations suppress the Cs^- phenotype of the original pol30 mutations equally well at both temperatures. Apart from the suppression phenotype, the suppressors conferred no new phenotypes, such as heat-sensitivity or sensitivity to the DNA synthesis inhibitor hydroxyurea. Genetic analysis revealed that suppression was linked to a single gene [suppressor of pol30 (SOP)1, 2, 3, etc.] in each of the strains. Both intragenic and extragenic suppressors were obtained for each of the three pol30 alleles.

Several intragenic mutations suppress the defects caused by the original pol30 mutations:
Each pol30 mutant gave rise to at least one new pol30 mutation that could compensate for the defects caused by the original mutation, and we recovered a total of 12 such intragenic suppressor mutations. To demonstrate that each of the suppressor mutations was linked to the pol30 gene, we crossed the pol30^sup strains with a wild-type POL30 strain (CH2237). We observed no Cs^- spores in any of the crosses upon tetrad analysis (at least seven tetrads were analyzed for each strain). Using genomic sequencing we confirmed that all the suppressor mutations that were linked to the pol30 locus were indeed within the pol30 gene. Specifically, we observed a single second pol30 mutation in every intragenic suppressor strain; no back mutations restoring the wild-type POL30 sequence were observed. The amino acid changes identified in the case of the 12 intragenic suppressor strains affect five amino acids of PCNA: pol30-120 (D41N), pol30-121 (D71Y), pol30-122 (D41G), pol30-123 (L205S), pol30-124 (G218C), and pol30-125 (S219P). It is striking that amino acid changes affecting aspartic acid 41 are seen in suppressors of both pol30-100 and pol30-104. Similarly, the same amino acid change affecting aspartic acid 71 was observed in a pol30-104 suppressor and a pol30-108 suppressor. Mapping the pol30^sup mutations on the crystal structure of PCNA revealed that only one of the suppressor mutations altered an amino acid close to the one changed by the original pol30 mutation (Figure 1). One obvious possibility is that all of the suppressor mutations alter the 3-D structure in such a way as to compensate for the original alteration.

View larger version (32K): In this window In a new window Download PPT slide   Figure 1. Locations of amino acids altered in strains carrying intragenic suppressors of pol30 mutations. The structure of a PCNA monomer is shown using the Rasmol program. A second mutation within the pol30 gene itself was found in 12 of 24 pseudorevertant strains that were isolated using three cold-sensitive pol30 mutants. The original amino acids changed in cold-sensitive pol30 mutants [pol30-100 (K253E), pol30-104 (A251V), and pol30-108 (A251T)] are shown in green in the yeast PCNA monomer structure (KRISHNA et al. 1994 ). The residues affected in the 12 intragenic suppressor strains map to five different amino acids, which are shown in red [D41N (pol30-120); D41G (pol30-122); D71Y (pol30-121); L205S (pol30-123); G218C (pol30-124); and S219P (pol30-125)].

Defects in pol30-104 (A251V) and pol30-108 (A251T) mutants are suppressed by dominant mutations in RFC3:
To learn about proteins that interact with PCNA, we analyzed the extragenic suppressors of pol30 mutations. Because alanine 252 of human PCNA is known to interact directly with the C-terminal region of p21, we first isolated second-site suppressors of a Cs^- mutant of yeast PCNA (pol30-104) that affects the equivalent residue, alanine 251. Genetic analysis of five revertant strains showed not only that suppression of the mutant phenotypes was due to extragenic mutations affecting single genes, but that all five of the suppressors mapped to the same locus (SOP1). Furthermore, diploid strains that are homozygous for pol30-104 and heterozygous for SOP1^sup exhibit dominant suppression in all cases. Thus, even in the presence of the wild-type copy of SOP1, the SOP1^sup allele is able to restore function to the mutant PCNA. Because all of the suppressors confer dominant suppression and affect the same gene product, our results suggested that the SOP1 gene product is required for proper function of PCNA.

To determine the extent of suppression of pol30-104 by SOP1^sup, flow cytometry was used to examine the DNA content of asynchronously dividing cultures of POL30, pol30-104, and pol30-104 SOP1^sup strains at the permissive (35°) and nonpermissive temperatures (14°). While the DNA profiles of all strains were similar at the permissive temperature, the DNA profiles at the restrictive temperature differed significantly (Figure 2). Whereas the POL30 culture contains cells in both the G1 and G2 phases of the cell cycle, the majority of cells in the pol30-104 strain are arrested in the G2 phase of the cell cycle. In contrast, the suppressor strains show an intermediate phenotype. Because we have previously shown that the G2 arrest of pol30-104 cells is due to a defect in progression through S phase (AMIN and HOLM 1996 ), these results suggest that the DNA replication defect of the pol30-104 mutation is partially suppressed in the pol30-104 SOP1^sup double mutant strain.

View larger version (17K): In this window In a new window Download PPT slide   Figure 2. DNA content of POL30, pol30, and pol30 SOP1 strains. Early log phase cultures of the rho0 strains CH2253 (POL30), CH2252 (pol30-104), and CH2431 (pol30-104 SOP1-3) were shifted to the permissive temperature (35°) for 3 hr or the restrictive temperature (14°) for 24 hr. Propidium iodide staining and flow cytometry were subsequently used to monitor the DNA content of the cells (see MATERIALS AND METHODS). The x-axis represents the relative DNA content and the y-axis represents cell number.

To determine the identity of the SOP1 gene, we tested whether SOP1 encodes a protein that has been previously suggested from biochemical studies to interact with PCNA. Obvious candidates include both the proteins of the RFC complex that loads PCNA onto the DNA and the genes encoding the catalytic subunits of DNA polymerases and . Thus, we performed crosses to test whether the SOP1 gene was RFC1, RFC2, RFC3, RFC4, RFC5, CDC2 (encodes the catalytic subunit of DNA polymerase ), or POL2 (encodes the catalytic subunit of DNA polymerase ). Tetrad analysis revealed that the SOP1 gene is linked to the SEC21 locus, suggesting that the neighboring RFC3 gene was a likely candidate. We confirmed that SOP1 is indeed RFC3 by genomic sequencing (Table 2) and by showing that an RFC3 suppressor allele (RFC3-3) is sufficient to suppress a pol30-104 mutation when carried on a plasmid (data not shown). This in vivo result gives biological support to biochemical results that show that both the RFC complex (TSURIMOTO and STILLMAN 1991A ; GERIK et al. 1997 ) and RFC3p in particular interact physically with PCNA (MOSSI et al. 1997 ).

The generality of suppression of pol30 defects by RFC3 was demonstrated when a mutation in RFC3 was isolated as a suppressor of a second allele of pol30, pol30-108. Like pol30-104, pol30-108 causes an amino acid change in alanine 251 of yeast PCNA, but with pol30-108 the change is to a threonine instead of valine. Only one pseudorevertant strain of pol30-108 contained an extragenic suppressor. Genetic analysis revealed that this suppressor maps to a single gene that confers dominant suppression. Using genomic sequencing we determined once again that the suppressor mutation was in the RFC3 gene (Table 2). This result confirms that the RFC3 gene product is important for the functional interaction with the interdomain region of PCNA.

The pol30-100 (K253E) mutation is suppressed by mutations affecting not only RFC3 but also RFC1 and RFC2:
To determine whether proteins other than RFC3p might also interact with the interdomain region of PCNA, we chose a different Cs^- mutation to isolate additional second-site suppressors. The pol30-100 mutation causes an amino acid change from lysine 253 to glutamic acid, and it can be distinguished from the pol30-104 and pol30-108 mutants by its more severe DNA replication defect (B. J. MERRILL and C. HOLM, unpublished data). Six pseudorevertants of the pol30-100 mutant were analyzed genetically to identify the gene products affecting suppression. Genetic mapping and genomic sequencing (Table 2) revealed that the six extragenic suppressors of pol30-100 affect three different genes. Two of the six suppressor genes mapped to the RFC3 locus in this third independent screen. In addition, two independently derived pseudorevertant strains proved to carry dominant suppressor mutations in the RFC1 gene, which encodes the large subunit of replication factor C. The remaining two pseudorevertant strains carry dominant mutations affecting yet another member of the RFC gene family, the RFC2 gene. In summary, although the importance of the RFC complex as the loader of the PCNA clamp has been shown biochemically, using genetic studies we have identified three specific RFC members, RFC1p, RFC2p, and RFC3p that are important for a functional interaction with the interdomain region of yeast PCNA.

Allele-restricted suppression of different pol30 alleles by RFC3 suppressor alleles:
To determine whether the suppressor mutations suppress specific defects in PCNA, we next examined the allele specificity of suppression of pol30 alleles by RFC mutations. We used tetrad analysis or plasmid transformation to determine whether any of the RFC suppressors can suppress a collection of pol30 alleles that have previously been isolated (Table 3). All of the RFC suppressor mutations suppress the phenotypes of some, but not all, Cs^- pol30 alleles. For example, we observed that the RFC3-1 mutation, which was recovered as a suppressor of pol30-104 (A251V), fails to suppress the Cs^- phenotype of the pol30-101 mutation (R44G), which is located farthest away from the pol30-104 mutation. In addition, the Cs^- pol30-106 mutation, which is suppressed by all three RFC3 suppressors, is not suppressed by the RFC1 or RFC2 mutations. Suppression of the Mms^s phenotype was even more specific; sensitivity to MMS was usually suppressed in at most two of the pol30 alleles. The observation that the RFC3-2 mutation suppresses the phenotype of pol30-104 and not pol30-108 is particularly striking because both of these pol30 mutations affect alanine 251, although with different amino acid changes. The differences in suppression of pol30 alleles by the RFC suppressors suggest that the requirement of PCNA in MMS-induced DNA repair is likely to involve a different mechanism of action or different levels of activity compared to its role in DNA replication. Overall, these results demonstrate that suppression by RFC mutations is allele-restricted. Because dominant allele-restricted suppression suggests an interaction between proteins (ADAMS and BOTSTEIN 1989 ; SANDROCK et al. 1997 ), these results are consistent with a functional interaction between RFC1p, RFC2p, and RFC3p and PCNA.

Genomic sequencing of the RFC suppressor genes reveals a common region that is important for suppression of pol30 mutations:
To evaluate the conservation of the regions of RFC1p, RFC2p, and RFC3p that are important for interaction with PCNA in vivo, we sequenced the genomic copy of RFC1, RFC2, and RFC3 in the pseudorevertant strains (Table 2). We saw a striking clustering of amino acid changes in asparagine 77 or methionine 78 in RFC3p (8 of 12 extragenic suppressors; Table 2 and Figure 3). Furthermore, regardless of which pol30 mutation was used in the pseudoreversion screen, we obtained the same amino acid change in RFC3p from methionine 78 to isoleucine in five different suppressor strains (Table 2). Our results suggest that asparagine 77 and methionine 78 of RFC3p are critical for RFC3p function because they either affect protein folding or they are important points of contact with PCNA; this methionine is present in RFC3p of both yeast and humans (CULLMAN et al. 1995 ). Mapping the RFC1, RFC2, and RFC3 mutations onto the linear amino acid sequence of the respective genes revealed that all of the extragenic suppressors have amino acid changes in the region of RFC box IV (Figure 3), which is one of eight regions conserved among the RFC genes (CULLMAN et al. 1995 ). Although the function of RFC box IV is not apparent from its sequence, RFC boxes III and V have sequence similarity to the ATP-binding region of known ATPases. Our results suggest that RFC box IV may play an additional role in functionally interacting with PCNA.

View larger version (10K): In this window In a new window Download PPT slide   Figure 3. RFC suppressor mutations map near RFC box IV. Although RFC1p is the only RFC protein to have RFC box I (DNA ligase homology box), all RFC proteins share the conserved RFC boxes II–VIII (CULLMAN et al. 1995 ). Mapping the RFC1, RFC2, and RFC3 suppressor mutations (arrows) onto the linear sequence of the respective genes reveals a striking clustering of the amino acid changes in the region of RFC box IV (hatched box). Although the function of RFC box IV is not apparent from its sequence, the clustering of suppressor mutations nearby suggests that it may be important for interactions with PCNA. (This is modeled after Figure 8 in CULLMAN et al. 1995 .)

RFC3 suppressor mutations suppress the in vitro DNA synthesis defect in Pol30-104p:
To begin to understand the defect in pol30 mutants that is suppressed by RFC mutations, we examined the characteristics of the mutant proteins biochemically. Because pol30 mutants synthesize DNA very slowly at the restrictive temperature, it seemed likely that the biochemical defect in the pol30 mutant proteins would be a processivity defect. To determine whether Pol30-104p causes DNA synthesis defects in vitro, we used a singly primed X174 DNA template to examine the ability of wild-type and mutant PCNA and RFC proteins to stimulate Pol-dependent DNA synthesis. Although wild-type RFC with wild-type PCNA promotes a significant amount of DNA synthesis of full-length X174 DNA (5386 nucleotides) within 30 min at 37° (Figure 4, lanes 1 and 2), wild-type RFC with the Pol30-104 protein promotes very little full-length synthesis under the same conditions (Figure 4, lanes 5 and 6). Similar results were obtained when the incubation times were reduced to 3, 5, or 10 min (data not shown). These results are consistent with a defect in processivity in the Pol30-104 protein.

View larger version (130K): In this window In a new window Download PPT slide   Figure 4. Effect of wild-type and mutant PCNA and RFC proteins in pol -dependent DNA synthesis. Combinations of mutant and wild-type RFC and PCNA were assayed for their ability to promote pol-dependent DNA synthesis at 37° (lanes 1, 2, 5, 6, 9, 10, 13, and 14) or 25° (lanes 3, 4, 7, 8, 11, 12, 15, and 16) in 30 min. Each reaction contained identical amounts of singly primed X174 DNA, pol, and E. coli single-strand binding protein (see MATERIALS AND METHODS). In addition, each reaction contained either wild-type (wt Pol30) or mutant (Pol30-104) PCNA, and either wild-type (wt RFC) or suppressor (sup RFC) replication factor C, as indicated at the top. Full-length X174 is 5.4 kb long. DNA product formation was assayed by using alkaline agarose gel electrophoresis followed by autoradiography. Details of the reaction conditions are described in MATERIALS AND METHODS.

It is interesting to note that the production of full-length product is dramatically affected by the temperature of the assay in all samples. For example, at 25° even DNA polymerase complexes containing wild-type RFC and PCNA proteins appear to stall when the DNA product reaches one-fifth of its full length (Figure 4, lanes 3 and 4). One explanation for this phenomenon is that the secondary structure of the template DNA may be substantially more stable at the lower temperature. This problem is circumvented by the presence of the RFC^sup complex in reactions containing wild-type PCNA; this combination of proteins pauses less on the template DNA, and it is able to promote the synthesis of full-length X174 DNA even at 25° (Figure 4, lanes 11 and 12).

If the lack of full-length X174 DNA synthesis accurately reflects the in vivo defect of the pol30-104 mutant, then this defect should be suppressed in vitro by the addition of the mutant RFC^sup complex, because RFC^sup suppresses the in vivo defect. Indeed, in lanes 13–16 of Figure 4, it is clear that suppressor RFC, which contains the RFC3-3 protein, alleviates the DNA synthesis defect; even though there is fivefold less RFC^sup protein than wild-type RFC in this experiment, the Pol30-104 protein promotes the efficient production of full-length X174 DNA (Figure 4, lanes 13 and 14; similar results were obtained in reactions in which the same amount of RFC^sup or wild-type RFC was added). To determine whether the in vitro phenotype also holds true for another PCNA mutant whose Cs^- phenotype is also suppressed by RFC3-3, we examined in vitro DNA synthesis with the Pol30-106 protein in reactions with wild-type or suppressor RFC. Consistent with the in vivo results (Table 3), the DNA synthesis defect in Pol30-106p is also suppressed by an RFC^sup complex containing RFC3-3p (data not shown).

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have identified spontaneously arising mutations in the RFC1, RFC2, and RFC3 genes that suppress the DNA replication and DNA repair defects of Cs^- pol30 mutants. All of the suppressor mutations confer a dominant suppression phenotype and show allele-restricted suppression of pol30 alleles. To elucidate the regions of RFC1, RFC2, and RFC3 that are important for suppression of pol30 mutant defects, we performed genomic sequencing. Our results suggest that the region around RFC box IV, which is conserved in all of the RFC genes, is important for a functional interaction of RFC1p, RFC2p, and RFC3p with the interdomain region of PCNA. In vitro analysis of the wild-type and mutant RFC and PCNA proteins reveals that mutant PCNA proteins during pol -dependent DNA synthesis are defective in enhancing elongation and that this defect is suppressed by a mutant RFC complex containing suppressor RFC3-3p.

Our second-site suppression screens were successful in identifying specific proteins that interact with PCNA in vivo. Many biochemical studies have previously shown that PCNA can interact with a large number of proteins, such as the RFC complex (TSURIMOTO and STILLMAN 1991A ; GERIK et al. 1997 ), DNA polymerase (BAUER and BURGERS 1988 ; ZHANG et al. 1995 ), FEN-1 (LI et al. 1995 ), Msh2/Msh3 (JOHNSON et al. 1996 ), Gadd45 (CHEN et al. 1995 ), a DNA methylation protein (CHUANG et al. 1997 ), and p21^WAF1/CIP1 (XIONG et al. 1992 , XIONG et al. 1993 ; ZHANG et al. 1993 ; FLORES-ROZAS et al. 1994 ; WAGA et al. 1994 ). Our studies suggest that out of these possible interactions, only the RFC interactions can compensate for the cold sensitivity caused by pol30 mutations in the interdomain region of yeast PCNA. It may be the case that our pol30 alleles are cold sensitive because of the lack of a specific interaction of PCNA with the RFC complex, which could result in a DNA synthesis defect. Of course, conclusions about the specificity of this interaction are somewhat limited because our screens were not fully saturated; although RFC1-19 and RFC2-10 were identified as suppressors of pol30-100, these mutations were not recovered from pseudoreversion screens with the pol30-104 and pol30-108 mutants. Nonetheless, it is striking that among 12 extragenic suppressors, all 12 affected subunits of the RFC complex. This result suggests that the most critical interaction with the interdomain region of PCNA is with subunits of the RFC complex.

Although it is clear that RFC requires five individual subunits to function in vitro in yeast (GERIK et al. 1997 ), the in vivo function of each of the RFC proteins remains unclear. One possibility is that each of the RFC proteins performs a unique cellular function. RFC2p and RFC5p, for example, may perform a checkpoint function (SUGIMOTO et al. 1996 ; NOSKOV et al. 1998 ; SHIMOMURA et al. 1998 ). However, the existence of only one mutant allele of each gene makes this result difficult to generalize. A second and more likely explanation for the essential nature of each RFC gene is that the specific structure created by the assembly of all five RFC subunits is absolutely essential for the function of the whole complex. This hypothesis is supported by increasing biochemical evidence resulting from purification and analysis of the different subunits of human RFC using recombinant baculoviruses (CAI et al. 1997 ; PODUST and FANNING 1997 ; UHLMANN et al. 1997A , UHLMANN et al. 1997B ). However, it is unclear how each of the subunits of RFC within a fully formed complex affects the interaction of the complex with PCNA.

Our in vivo observation that mutations in three different RFC genes can suppress the single PCNA defect caused by the pol30-100 mutation suggests that these three RFC genes must, at least to a certain extent, play overlapping roles in the cell. One possibility is that all three subunits (and perhaps RFC4p and RFC5p as well) make contacts with PCNA in the process of loading it onto the DNA. This idea is consistent with the observation that the human homologs of both RFC1p and RFC3p interact with the C-terminal face of human PCNA in vitro (MOSSI et al. 1997 ) and that there is amino acid sequence conservation among RFC proteins (CULLMAN et al. 1995 ). Thus, it is possible that changes in each of the RFC subunits can cause similar overall functional changes in the RFC complex. Alternatively, unique changes in each of the RFC1, RFC2, or RFC3 proteins may compensate for the defect in mutant PCNA proteins by acquiring an increased binding capability.

Second site suppression can be attributed to a specific physical interaction between two proteins (JARVIK and BOTSTEIN 1973 ), or it can result from a more general physical or functional interaction (SANDROCK et al. 1997 ). In general these two forms of suppression can be distinguished as follows. If suppression derives from the restoration of a specific physical interaction, then suppression is generally strictly allele specific, and interaction between the suppressor proteins and a wild-type protein is frequently compromised. If suppression derives from the creation of a new functional or physical interaction, then suppression is more general (allele-restricted), and the suppressor protein may exhibit enhanced binding to the wild-type protein. As an example of this latter type of suppression, the phenotype of many actin mutant alleles is suppressed by any one of a number of mutations in the SAC6 gene, which encodes an actin-binding protein (ADAMS and BOTSTEIN 1989 ; SANDROCK et al. 1997 ). In this example, the mechanism of suppression by the mutant Sac6 proteins is the creation of a new binding site, which has an increased ability to bind to both the mutant and wild-type actin proteins (SANDROCK et al. 1997 ). In many ways, our results with mutant and wild-type RFC and PCNA proteins are similar to the results with actin and SAC6p. We observe a similar allele-restricted suppression of different pol30 alleles by the RFC suppressors as well as an increased ability of the RFC^sup complex to promote the synthesis of long DNA products in the presence of wild-type PCNA during pol -dependent DNA synthesis. Thus, the nature of the suppression is consistent with a model in which suppressor mutations in RFC1, RFC2, or RFC3 result in altering the overall affinity or function of RFC rather than altering only the specific site where the amino acid is changed.

While RFC can act as a clamp loader and unloader based on in vitro reactions, it is not clear from previous studies whether RFC remains attached to the DNA polymerase machinery during processive DNA elongation. If RFC exits after loading PCNA onto the DNA, then our DNA synthesis results with suppressor RFC would require that RFC^sup loads mutant or wild-type PCNA onto the DNA in a way that is both different from normal and stable. This idea is consistent with in vitro studies conducted with human RFC containing an N-terminal deletion in the largest subunit of RFC; these studies suggest that RFC is no longer required once PCNA is loaded onto the DNA (PODUST et al. 1998 ). From other biochemical data, however, it seems likely that RFC^sup remains on the DNA with PCNA in a stable manner. For example, Waga and Stillman have recently examined the effect of p21 on PCNA loading by RFC, and they show that RFC remains bound to PCNA after the loading reaction; while p21 does not inhibit the loading process, it causes RFC to dissociate from the PCNA-bound DNA under conditions allowing ATP hydrolysis (WAGA and STILLMAN 1998 ). If RFC continues to associate with PCNA after the binding of DNA polymerase, the increased synthesis of long DNA products observed with RFC^sup could be due to its ability to remain tightly associated with the DNA polymerase holoenzyme better than wild-type RFC does during DNA elongation. Additional studies will be necessary to firmly establish whether wild-type RFC remains associated with the DNA replication complex after the binding of DNA polymerase.

	ACKNOWLEDGMENTS

We thank Scott Oh for construction of the RFC3-3 and RFC3 plasmids and for assistance with genomic sequencing; Peter Chu and Tsahai Tafari for the initial genetic analysis on the pol30-100 and pol30-108 suppressors; Hernan Flores-Rozas for advice on purification of Pol and RFC; Peter Burgers for providing us with strain BJ2168 and plasmids pBL413, pBL420, pBL425, and MTL4 used in the purification of wild-type and mutant RFC; Mike McAlear for the his-tagged POL30 plasmid; Richard Kolodner for strain RKY1293 used for the purification of pol ; Brad Merrill for assistance with molecular graphics; Dr. Scott Emr for providing the sec21 strain; and members of the Holm laboratory for contributing ideas during the course of this project. This project was supported by a grant to C.H. from the National Institutes of Health (NIH; GM-36510), and N.S.A. was supported in part by an NIH training grant (GM-07552).

Manuscript received April 23, 1999; Accepted for publication August 25, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADAMS, A. D. and D. BOTSTEIN, 1989  Dominant suppressors of yeast actin mutations that are reciprocally suppressed. Genetics 121:675-683[Abstract/Free Full Text].

AMIN, N. S. and C. HOLM, 1996  In vivo analysis reveals that the interdomain region of the yeast proliferating cell nuclear antigen is important for DNA replication and DNA repair. Genetics 144:479-493[Abstract].

BAUER, G. A. and P. M. J. BURGERS, 1988  Protein-protein interactions of yeast DNA polymerase III with mammalian and yeast proliferating cell nuclear antigen (PCNA)/cyclin. Biochim. Biophys. Acta 951:274-279[Medline].

BAUER, G. A., H. M. HELLER, and P. M. J. BURGERS, 1988  DNA polymerase III from Saccharomyces cerevisiae: I. Purification and characterization. J. Biol. Chem. 263:917-924[Abstract/Free Full Text].

BECKWITH, W. H., S. QIANG, R. BOSSO, K. J. GERIK, and P. M. J. BURGERS et al., 1998  Destabilized PCNA trimers suppress defective RFC1 proteins in vivo and in vitro.. Biochemistry 37:3711-3722[Medline].

BURGERS, P. M. J., 1991  Saccharomyces cerevisiae replication factor C: II. Formation and activity of complexes with the proliferating cell nuclear antigen and with DNA polymerases and . J. Biol. Chem. 266:22698-22706[Abstract/Free Full Text].

CAI, J., E. GIBBS, F. UHLMANN, B. PHILLIPS, and N. YAO et al., 1997  A complex consisting of human replication factor C p40, p37, and p36 subunits is a DNA-dependent ATPase and an intermediate in the assembly of the holoenzyme. J. Biol. Chem. 272:18974-18981[Abstract/Free Full Text].

CANTOR, C. R., and P. R. SCHIMMEL, 1980 Biophysical Chemistry: The Conformation of Biological Macromolecules. W. H. Freeman, San Francisco.

CHEN, I.-T., M. SMITH, P. M. O'CONNOR, and A. J. FORNACE, JR., 1995  Direct interaction of Gadd45 with PCNA and evidence for competitive interaction of Gadd45 and p21^WAF1/CIP1 with PCNA. Oncogene 11:1931-1937[Medline].

CHUANG, L. S.-H., H.-I. IAN, T.-W. KOH, H.-H. NG, and G. XU et al., 1997  Human DNA-(cytosine-5) methyltransferase-PCNA complex as a target for p21^WAF1. Science 277:1996-2000[Abstract/Free Full Text].

CULLMAN, G., K. FIEN, R. KOBAYASHI, and B. STILLMAN, 1995  Characterization of the five RFC genes from S. cerevisiae.. Mol. Cell. Biol. 15:4661-4671[Abstract].

EL-DEIRY, W. S., T. TOKINO, V. E. VELCULESCU, D. B. LEVY, and R. PARSONS et al., 1993  WAF1, a potential mediator of p53 tumor suppression. Cell 75:817-825[Medline].

FIEN, K. and B. STILLMAN, 1992  Identification of replication factor C from Saccharomyces cerevisiae: a component of the leading-strand DNA replication complex. Mol. Cell. Biol. 12:155-163[Abstract/Free Full Text].

FLORES-ROZAS, H., Z. KELMAN, F. B. DEAN, Z. Q. PAN, and J. W. HARPER et al., 1994  Cdk-interacting protein 1 directly binds with proliferating cell nuclear antigen and inhibits DNA replication catalyzed by the DNA polymerase holoenzyme. Proc. Natl. Acad. Sci. USA 91:8655-8659[Abstract/Free Full Text].

FROSINA, G., P. FORTINI, O. ROSSI, F. CARROZZINO, and G. RASPAGLIO et al., 1996  Two pathways for base excision repair in mammalian cells. J. Biol. Chem. 271:9573-9578[Abstract/Free Full Text].

GERIK, K. J., S. L. GARY, and P. M. J. BURGERS, 1997  Overproduction and affinity purification of Saccharomyces cerevisiae replication factor C. J. Biol. Chem. 272:1256-1262[Abstract/Free Full Text].

GULBIS, J. M., Z. KELMAN, J. HURWITZ, M. O'DONNELL, and J. KURIYAN, 1996  Structure of the C-terminal region of p21^WAF1/CIP1 complexed with human PCNA. Cell 87:297-306[Medline].

HUTTER, K.-J. and H. E. EIPEL, 1979  Microbial determinations by flow cytometry. J. Gen. Microbiol. 113:367-375.

JARVIK, J. and D. BOTSTEIN, 1973  A genetic method for determining the order of events in a biological pathway. Proc. Natl. Acad. Sci. USA 70:2046-2050[Abstract/Free Full Text].

JOHNSON, R. E., G. K. KOVVALI, S. N. GUZDER, N. S. AMIN, and C. HOLM et al., 1996  Evidence for the involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair. J. Biol. Chem. 271:27987-27990[Abstract/Free Full Text].

JONSSON, Z. O. and U. HUBSCHER, 1997  Proliferating cell nuclear antigen: more than a clamp for DNA polymerases. Bioessays 19:967-975[Medline].

KELMAN, Z., 1997  PCNA: structure, functions and interactions. Oncogene 14:629-640[Medline].

KRISHNA, T. S. R., X. P. KONG, S. GARY, P. BURGERS, and J. KURIYAN, 1994  Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell 79:1233-1243[Medline].

LEE, S.-H. and J. HURWITZ, 1990  Mechanism of elongation of primed DNA by DNA polymerase , proliferating cell nuclear antigen and activator I. Proc. Natl. Acad. Sci. USA 87:5672-5676[Abstract/Free Full Text].

LEE, S.-H., Y. ISHIMI, M. K. KENNY, P. BULLOCK, and F. B. DEAN et al., 1988  An inhibitor of the in vitro elongation reaction of simian virus 40 DNA replication is overcome by proliferating-cell nuclear antigen. Proc. Natl. Acad. Sci. USA 85:9469-9473[Abstract/Free Full Text].

LEE, S.-H., A. D. KWONG, Z.-Q. PAN, and J. HURWITZ, 1991a  Studies on the activator 1 protein complex, an accessory factor for proliferating cell nuclear antigen-dependent DNA polymerase . J. Biol. Chem. 266:594-602[Abstract/Free Full Text].

LEE, S.-H., Z.-Q. PAN, A. D. KWONG, P. M. J. BURGERS, and J. HURWITZ, 1991b  Synthesis of DNA by DNA polymerase in vitro.. J. Biol. Chem. 266:22707-22717[Abstract/Free Full Text].

LI, X., J. LI, J. HARRINGTON, M. R. LIEBER, and P. M. BURGERS, 1995  Lagging strand DNA synthesis at the eukaryotic replication fork involves binding and stimulation of FEN-1 by proliferating cell nuclear antigen. J. Biol. Chem. 270:22109-22112[Abstract/Free Full Text].

LIU, B., R. PARSONS, N. PAPADOPOULOS, N. C. NICOLAIDES, and H. T. LYNCH et al., 1996  Analysis of mismatch repair genes in hereditary non-polyposis colorectal cancer patients. Nat. Med. 2:169-174[Medline].

MATSUMOTO, Y., K. KIM, and D. F. BOGENHAGEN, 1994  Proliferating cell nuclear antigen-dependent abasic site repair in Xenopus laevis oocytes: an alternative pathway of base excision DNA repair. Mol. Cell. Biol. 14:6187-6197[Abstract/Free Full Text].

MCALEAR, M. A., E. A. HOWELL, K. K. ESPENSHADE, and C. HOLM, 1994  PCNA mutations suppress the cell cycle defect conferred by cdc44 mutations. Mol. Cell. Biol. 14:4390-4397[Abstract/Free Full Text].

MOSSI, R., Z. O. JONSSON, B. L. ALLEN, S. H. HARDIN, and U. HUBSCHER, 1997  Replication factor C interacts with the C-terminal side of proliferating cell nuclear antigen. J. Biol. Chem. 272:1769-1776[Abstract/Free Full Text].

NICHOLS, A. F. and A. SANCAR, 1992  Purification of PCNA as a nucleotide excision repair protein. Nucleic Acids Res. 20:2441-2446.

NOSKOV, V. N., H. ARAKI, and A. SUGINO, 1998  The RFC2 gene, encoding the third-largest subunit of the replication factor C complex, is required for an S-phase checkpoint in Saccharomyces cerevisiae.. Mol. Cell. Biol. 18:4914-4923[Abstract/Free Full Text].

OAKLEY, C. E. and B. R. OAKLEY, 1989  Identification of -tubulin, a new member of the tubulin superfamily encoded by the mipA of Aspergillus nidulans.. Nature 338:662-664[Medline].

PODUST, L. M., V. N. PODUST, J. M. SOGO, and U. HUBSCHER, 1995  Mammalian DNA polymerase auxiliary proteins: analysis of replication factor C-catalyzed proliferating cell nuclear antigen loading onto circular double-stranded DNA. Mol. Cell. Biol. 15:3072-3081[Abstract].

PODUST, V. N. and E. FANNING, 1997  Assembly of functional replication factor C expressed using recombinant baculoviruses. J. Biol. Chem. 272:6303-6310[Abstract/Free Full Text].

PODUST, V. N., B. A. GEORGAKI, B. STRACK, and U. HUBSCHER, 1992  Calf thymus RF-C as an essential component for DNA polymerase and holoenzymes function. Nucleic Acids Res. 20:4159-4165[Abstract/Free Full Text].

PODUST, V. N., N. TIWARI, S. STEPHAN, and E. FANNING, 1998  Replication factor C disengages from proliferating cell nuclear antigen (PCNA) upon sliding clamp formation, and PCNA itself tethers DNA polymerase to DNA. J. Biol. Chem. 273:31992-31999[Abstract/Free Full Text].

SANDROCK, T. M., J. L. O'DELL, and A. E. M. ADAMS, 1997  Allele-specific suppression by formation of new protein-protein interactions in yeast. Genetics 147:1635-1642[Abstract].

SHERMAN, F., G. R. FINK and J. B. HICKS, 1986 Methods in Yeast Genetics. Cold Spring Harbor Labo