Purification, gene cloning, and reconstitution of the heterotrimeric single-stranded DNA-binding protein from Schizosaccharomyces pombe.

We have purified a single-stranded DNA-binding protein (SSB) from Schizosaccharomyces pombe (Sp) and have shown that it is composed of three subunits of 68, 30, and 12 kDa. The SpSSB supports T antigen-dependent unwinding of SV40 ori containing DNA, but is not functional in the SV40 in vitro replication reaction. All three genes that encode the SpSSB subunit have been isolated. The cloned cDNA of the ssb1+, encoding the p68 subunit, contains 609 amino acids (68.3 kDa), while that of the ssb2+, encoding the p30 subunit, contains a 279 amino acids (30.3 kDa). The genomic DNA clone of the p12 subunit gene (ssb3+) has 2 introns and an open reading frame of 104 amino acids (11.8 kDa). Significant homology is observed among the largest and middle subunits of eukaryotic SSBs, but there is poor homology among the smallest subunits. In addition, we have reconstituted the SpSSB complex by coexpression of all three subunits in Escherichia coli. The reconstituted complex is active in single-stranded DNA binding and the T antigen-dependent unwinding of SV40 ori DNA. Finally, we observed a cell cycle-dependent phosphorylation pattern of the p30 subunit of SpSSB, which is similar to that observed for the human and Saccharomyces cerevisiae SSB.

The mechanism by which eukaryotes initiate DNA synthesis from origins of DNA replication on chromosomes is not understood. Insight into this process has come from detailed biochemical studies using the simian virus 40 (SV40) 1 in vitro replication system which demonstrated a pivotal role of the heterotrimeric single-stranded (ss) DNA binding protein (SSB) in DNA synthesis. SV40 DNA replication is initiated when the viral-encoded large T antigen binds as a double hexamer to the core origin, and recruits the human DNA polymerase ␣-primase complex and human (H) SSB (also called RP-A) to catalyze primer synthesis. Subsequently, T antigen functions as a bidirectional helicase to unwind the DNA in the presence of HSSB, and the coordinate actions of DNA polymerase ␣-primase and DNA polymerase ␦ holoenzyme at the replication forks catalyze leading and lagging strand synthesis. HSSB, in addition to its role in the initiation reaction, is also required for the elongation reaction as shown by its stimulatory effects on polymerase activity with primed templates (reviewed in Refs. [1][2][3][4]. Biochemical studies on the initiation of DNA replication on metazoan chromosomes have not been possible because origins of replication have not been defined to specific sequences. Recent studies in the budding yeast Saccharomyces cerevisiae (Sc), have led to the biochemical characterization of an origin recognition complex (ORC) that binds to the well defined origin of DNA replication, ARS1 (5), from this organism. The ORC is a complex of six different subunits that binds to ARS1 and other ARS sequences in an ATP-dependent manner (6 -8). Many factors appear to interact with the yeast ORC, and these include CDC6 and CDC46/MCM5 (9 -11). Undoubtedly, other factors required for the initiation reaction would include the polymerase ␣-primase complex and ScSSB. Recently, some of the homologous ORC subunits have been isolated from Schizosaccharomyces pombe, Drosophila, Xenopus, and human sources, suggesting a conserved mechanism for the initiation of DNA replication in lower and higher eukaryotes (12)(13)(14)(15)(16).
Our primary focus is to understand the mechanisms of initiation of DNA replication in eukaryotic cells. We are using the fission yeast S. pombe (Sp) as a model system because it has certain features that are similar to higher eukaryotes, and offers both biochemical and genetical analyses of DNA replication mechanisms. Origins of replication, as defined by an ARS assay and two-dimensional mapping techniques, have been isolated from S. pombe and characterized (17)(18)(19). These origins are in the range of Ͼ500 bp, larger than ARS elements from S. cerevisiae (ϳ100 bp) (18,19). The availability of these ARS elements will facilitate a biochemical study of an initiation complex from S. pombe. To this end, our objective is to isolate the ORC complex from S. pombe, along with SpSSB and the DNA polymerase ␣-primase complex, as a step toward analyzing the initiation of DNA synthesis from S. pombe origins. In this study, we report the purification of the single-stranded DNA-binding protein complex from S. pombe. The protein is heterotrimeric and shows regions of conservation to SSBs isolated from other eukaryotes. We demonstrate the overexpression and reconstitution of the active complex from Escherichia coli.

MATERIALS AND METHODS
Yeast and E. coli Strains S. pombe mutants used in this study are all isogenic derivatives of 972h Ϫ strain (kindly provided by Dr. T. Wang, Stanford University). The genotypes of cdc10-129, cdc22-m45, cdc25-22, and cdc2-33 were described previously (20). E. coli strains DH5␣, DH10, Y1080r Ϫ , XL-1 blue, and BL21(DE3) were used for plasmid construction, cDNA library maintenance, hosts for genomic library, and sources for making singlestranded DNA for sequencing and expression of recombinant protein, respectively.

Replication Proteins and Other Reagents
SV40 T antigen and human proteins used for the SV40 DNA replication assay were described previously (21). Enzymes used in DNA manipulations were from New England Biolabs, Perkin-Elmer, and Boehringer Mannheim. gt11 based S. pombe genomic library was obtained from Clonetech, while the pDB20 based S. pombe cDNA library was kindly provided by Dr. L. Guarente, Massachussetts Institute of Technology (22). pBluescript SKϩ, used for cloning and sequencing, was from Stratagene. SV40-origin containing plasmid pSV01⌬EP (23) and pET19b-GST, used for SpSSB expression in E. coli (24), were described previously.

Purification of SSB from S. pombe Cells
SSB from S. pombe was purified according to the procedure described for the isolation of the S. cerevisiae SSB with some modifications (25). All steps were carried out at 4°C. S. pombe 972h Ϫ wild type cells (500 g, wet weight) were resuspended in 1 liter of buffer A (25 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10% glycerol, 1 mM DTT, 0.1 mM phenylmethanesulfonyl fluoride, 10 mM benzamidine, 0.2 g/ml aprotinin, 0.2 g/ml leupeptin, 0.1 g/ml antipain, and 50 M EGTA) with 1 M NaCl and 10 mM sodium bisulfite. Disruption of cells was carried out using a ratio of 1 volume of glass beads to 2 volumes of cell suspension with a Bead-Beater (Biospec Products: total volume, 360 ml). Disruption was carried out using 8 cycles of vortexing for 30 s followed by 1 min of cooling in ice. The lysate was centrifuged at 16, 200 ϫ g for 35 min. The supernatant (800 ml, 20 g protein) was adjusted to 0.5 M NaCl with buffer A and applied to a 40 ml (2.3 ϫ 8 cm) Affi-Gel blue column (Bio-Rad), preequilibrated with buffer A ϩ 0.5 M NaCl. This column was washed with 20 column volumes of buffer A ϩ 0.8 M NaCl and the Sp SSB eluted with 2 column volumes of buffer A ϩ 2 M NaCl with 40% ethylene glycol. The eluted protein peak was pooled (76 ml, 170 mg protein), diluted to 0.5 M NaCl with buffer A, and applied to a 20-ml (2.3 ϫ 4.5 cm) ssDNA cellulose column (Sigma), preequilibrated with buffer A ϩ 0.5 M NaCl. This column was washed with 20 column volumes of buffer A ϩ 0.8 M NaCl and 2 column volumes of buffer A ϩ 2 M NaCl with 40% ethylene glycol. The peak of protein obtained in the latter wash was pooled (5.6 ml, 1 mg protein) and dialyzed against 2 liters of buffer A ϩ 0.2 M NaCl containing 20% sucrose. After clarification by centrifugation, this material was diluted to 0.1 M NaCl with buffer A and applied to a 1-ml (0.6 ϫ 3.5 cm) Q-Sepharose column (Pharmacia Biotech Inc.) preequilibrated with buffer A ϩ 0.1 M NaCl. This column was washed with 10 ml of buffer A ϩ 0.1 M NaCl and eluted with a 10-ml linear gradient from 0.1 to 0.4 M NaCl in buffer A. Fractions containing SpSSB, eluting at 0.22 M NaCl, were pooled (0.7 ml) and dialyzed against 500 ml of buffer A ϩ 0.2 M NaCl. This procedure yielded 0.4 mg of SpSSB. SpSSB was subjected to glycerol gradient centrifugation as follows: 110 g of the protein was sedimented at 300,000 ϫ g for 21 h in a 5-ml 15-35% glycerol gradient in buffer A ϩ 0.2 M NaCl in a SW50.1 rotor (Beckman) at 4°C; about 95% of protein was recovered as a single peak.

SpSSB Antibody
Purified SpSSB was separated by SDS-PAGE and individual subunits excised and used as antigens for the preparation of antisera against the SpSSB p68 and p30 subunits in rabbits. Affinity purification of antibodies from these sera was performed using the purified proteins.

SV40 T Antigen-dependent Unwinding Assay
T antigen (150 ng) was incubated with the indicated amounts of SpSSB for 90 min at 37°C. The reaction mixture (15 l) contained 40 mM creatine phosphate (di-Tris salt, pH 7.7), 8 mM MgCl 2 , 0.5 mM DTT, 4 mM ATP, 2.5 g of BSA, 0.5 g of creatine kinase, 5 fmol of 32 P-labeled substrate DNA (the 311-bp EcoRI fragment from pSV01⌬EP containing the SV40 origin) and 0.4 g of a 1-kb DNA ladder (Life Technologies, Inc.) as a competitor. Reactions were terminated by the addition of 4 l of a mixture containing 5% SDS, 100 mM EDTA, and 0.5 g proteinase K followed by incubation at 37°C for 15 min. The reactions were then subjected to 4% PAGE.

Gel Mobility Shift Assay
A ssDNA of 45 nucleotides was 5Ј-end labeled with ␥-[ 32 P]ATP using T4 polynucleotide kinase. The sequence of the substrate DNA used was as follows: 5Ј-GTTCCGCGTGGATCCACTCCTGTTGCATCAGCCAC-CCAAAGTGTG-3Ј. 32 P-labeled ssDNA (5 fmol) was incubated with SSB (as indicated) at 37°C for 10 min, and the reactions were then subjected to 5% PAGE. The reaction mixture (15 l) was similar to that described for the unwinding assay. For competition analysis, various amounts of ssDNA (42 nucleotides; 5Ј-AGATCTGGTACCTTAATGATTTGCTCTT-TCACTGACAACATC-3Ј) or dsDNA (1-kb ladder) were used as competitors.

Peptide Sequencing
The three subunits of purified SpSSB were separated by SDS-PAGE, and the individual subunits were excised, cleaved by trypsin, and sequenced (Microchemistry, Sloan-Kettering Institute). Seven distinct peptide sequences for the p68 subunits, two distinct sequences for p30 subunits, and four distinct sequences for the p12 subunit of the SpSSB were obtained. These peptide sequences are underlined in Fig. 3 (A, B, and D).

Polymerase Chain Reaction (PCR)
All PCR reactions were performed in 50-l reaction mixtures containing 100 pmol each of primer, 2 units of Ultma DNA polymerase (Perkin-Elmer) or Taq DNA polymerase (Boehringer), and substrate DNA using conditions specified by the manufacturer.

Cloning of the ssb Genes
p68 Subunit Gene (ssb1 ϩ )-Two degenerate oligonucleotides were used to amplify the region of the gene by PCR. The primer sequences used were as follows: T29a (32 mer; 5Ј-AAYGTICARAAYGARTAYG-ARYTIATGTTYGA-3Ј) and T35b (30 mer; 5Ј-CATRTARCAIGCYT-CIGCCATRCARTTCAT-3Ј); the abbreviations used in the description of the primers are: I (inosine), R(G/A), Y(C/T), D(G/A/T), W(A/T), S(G/C). A PCR product of 0.9 kb was cloned and found to encode the predicted amino acid sequence. This fragment was then used to probe a S. pombe cDNA library for the isolation of ssb1 ϩ gene. From 2 ϫ 10 5 colonies screened, two positive clones were identified. These clones were subcloned into pBluescript SKϩ vector and both strands were completely sequenced.
p30 Subunit Gene (ssb2 ϩ )-A 0.4-kb product was obtained after PCR with two degenerate oligonucleotides (26 mer each): T13a (5Ј-AAYACI-ACITAYCARATHGARGAYGG-3Ј) and T24b (5Ј-RTGDATIGCYTGCA-TIACIGTCATYTG-3Ј). After sequencing of this fragment confirmed that it contained the predicted amino acid sequence, it was used as a probe to isolate cDNA clones. A cDNA library (1 ϫ 10 5 colonies) was screened and three positive clones were identified. The complete sequence of the clone containing the longest insert was determined using the same procedure as described above.
p12 Subunit Gene (ssb3 ϩ )-The primer sequences used for PCR were as follows: T8a (26 mer; 5Ј-GAYATGYTICCIGARTGYWSIGGIAA-3Ј) and T34b (29 mer; 5Ј-GTRTTRTCIACIGTIARRTGCATRTCRAA-3Ј). The PCR product of 132 bp was cloned and found to encode the predicted amino acid sequence. This fragment was then used as a probe to screen a genomic library; 2 ϫ 10 5 plaques were screened and a single clone was identified which was sequenced completely on both strands, as described above. A cDNA clone of the p12 subunit was obtained from S. pombe cDNA library by PCR using two primers (30 mers each); 12A (5Ј-TAAGGATCCATGGAACGCCCTACACCTCGA-3Ј) and 12B (5Ј-CCATCCCGGGTACCTTATTCAAAAAAAAGT-3Ј). The PCR amplified fragment was digested with NcoI and KpnI and ligated to a NcoI and KpnI-digested pET19b-GST vector and sequenced on both strands. The resulting plasmid was designated pET19b-SpSSB p12.

Construction of Plasmids Expressing the SSB Genes
The three SpSSB genes were coexpressed in E. coli using the pET system developed by Studier et al. (26). A description of this procedure is summarized in Fig. 5.
The plasmid that expressed the p68 subunit was made as follows. The SpSSB p68 cDNA was amplified by PCR using two primers. One primer 68A (28 mer: 5Ј-TTTGGATCCATGGCTGAGCGATTATCCG-3Ј) has an NcoI site that contains the start codon, and the other, 68D (29 mer: 5Ј-TAAACCCGGGTACCTTATTGAGCAGACTC-3Ј), has a KpnI site after the stop codon (both sites are underlined). The PCR amplified 1.8-kb fragment was digested with both NcoI and KpnI and ligated to the NcoI-KpnI-digested pET19b-GST vector. The resulting plasmid is devoid of the glutathione S-transferase gene from the parental pET19b-GST vector. The cloned SpSSB p68 subunit cDNA was expressed from the authentic start codon of the cDNA. This plasmid is referred to as pET19b-SpSSB p68. The construction of p12 subunit gene expression plasmid, pET19b-SpSSB p12, was described above.
The plasmid that expressed the p30 subunit gene was made as follows. Two primers (30 mers each): 30A (5Ј-TACGGATCCATGGCT-TATGATGCTTTTGGC-3Ј) containing a BamHI site and 30D (5Ј-CTT-GAATTCTGCATCAGCGTGAACTTGAGA-3Ј) containing an EcoRI site were used to generate a 0.2-kb product by PCR of the N-terminal region of the p30 gene. This fragment was digested with both BamHI and EcoRI and cloned into the same sites of pBluescript SKϩ. The 0.7-kb EcoRI-HindIII fragment containing the adjoining C-terminal region of the gene of SpSSB p30 cDNA was ligated to the EcoRI and HindIII digested plasmid to construct pBS-SpSSB p30. For formation of pET19b-SpSSB p30, the 0.9 Kb NcoI-HindIII fragment from pBS-SpSSB p30 containing the entire region of SpSSB p30 gene was ligated to the NcoI-HindIII site of pET19b-GST.
For the construction of pET19b-SpSSB p12/p30 (see Fig. 5), the pET19b-SpSSB p30 plasmid was digested with both XbaI and HindIII and both ends were blunt-ended with T4 DNA polymerase. This 0.9-kb XbaI-HindIII fragment containing the entire SpSSB p30 cDNA was ligated to the Acc65I-digested and T4 DNA polymerase-treated pET19b-SpSSB p12. We use the isoschizomer Acc65I instead of KpnI, since the constructed plasmid retained a KpnI site. pET19b-rSpSSB (containing all three subunits) was made in a similar way (also see Fig.  5). A 1.2-kb XbaI-HindIII fragment from the pET19b-SpSSB p12/p30, following treatment with T4 DNA polymerase, was ligated to the Acc65I site (filled in with T4 DNA polymerase) of the pET19b-SpSSB p68. All recombinant plasmids were sequenced to insure that no mutations had occurred during these manipulations.

Purification of Recombinant (r) SpSSB from E. coli
E. coli BL21(DE3) containing pET19b-rSpSSB was grown in 12 liters of L broth containing 100 g/ml ampicillin at 37°C to an A 600 of 0.5. Isopropyl ␤-D-thiogalactoside (0.4 mM) was added and the cells were incubated for 3 h. Induced cells were collected by centrifugation and stored frozen at Ϫ20°C until needed. Purification of rSpSSB was purified from E. coli following the purification procedure used for the isolation of the SpSSB from S. pombe cells, except that a phosphocellulose column was used instead of Q-Sepharose to remove a contaminating E. coli protein that had ssDNA binding activity. Induced cells (97 g in 300 ml of buffer A ϩ 1 M NaCl) were sonicated and centrifuged as described earlier. The supernatant was adjusted to 0.5 M NaCl with buffer A (660 ml, 1.1 g of protein) and applied to a 40-ml (2.3 ϫ 8 cm) Affi-Gel blue column and the eluate (230 ml, 117 mg) was loaded onto a 13-ml (1.3 ϫ 8.5 cm) ssDNA cellulose column. The peak of eluted protein was pooled (4 ml, 2 mg) and dialyzed against 1 liter of buffer A ϩ 0.2 M NaCl containing 20% sucrose. After centrifugation, this material was diluted to 50 mM NaCl with buffer A and applied to a 2-ml (0.6 ϫ 7 cm) phosphocellulose column. This column was washed with 20 ml of buffer A ϩ 50 mM NaCl and eluted with a 30-ml linear gradient from 50 to 500 mM NaCl in buffer A. This procedure yielded 0.8 mg of rSpSSB of 95% purity. Glycerol gradient centrifugation of 160 g of this preparation was carried out at 300,000 ϫ g for 21 h in a 5-ml, 15-35% glycerol gradient in buffer A containing 0.2 M NaCl in a SW50.1 rotor (Beckman) at 4°C. Approximately 95% of protein was obtained as a single protein peak. Fractions containing the peak of rSpSSB were used in the assays described below.

RESULTS
The SSB purified from S. pombe Is a Heterotrimeric Complex-We assumed that the chromatographic properties of the Sp SSB would be similar to those of HSSB and ScSSB. For this reason, the purification procedure used for the isolation of the ScSSB (25), as described under "Materials and Methods," was used to purify SpSSB. An analysis of the protein composition of fractions obtained after each chromatographic step is shown in Fig. 1A. To determine the subunit structure of the SpSSB, the Q-Sepharose fraction was subjected to glycerol gradient sedimentation followed by SDS-PAGE analysis (Fig. 1B). As shown, SpSSB contained three polypeptides of p68, p32, and p12 kDa that cosedimented at 5.6 S. This subunit structure and sedimentation coefficient are similar to those observed for HSSB and ScSSB (27)(28)(29). Multiple bands migrating slightly slower than the SpSSB p30 subunit were observed which corresponded to phosphorylated derivatives of the p30 subunit (see below). The 50-and 20-kDa protein bands (Fig. 1B, lanes [11][12][13][14] are breakdown products of the p68 and p30 subunits of SpSSB, respectively, since they were recognized by antibodies against SpSSB p68 and p30 subunits, respectively (data not shown).
SpSSB Supports Unwinding of SV40 ori DNA by T Antigen-Previous studies demonstrated that the HSSB and ScSSB support the unwinding of SV40 ori DNA by T antigen (25,30). Since SpSSB has a similar subunit structure as HSSB and ScSSB, we examined whether it would support the unwinding reaction. T antigen was incubated with a 32 P-labeled fragment containing the SV40 core origin and the glycerol gradient fractions. After incubation the reactions were deproteinized and subjected to PAGE analysis. As shown in Fig. 2A, SpSSB glycerol gradient fractions 11-13 ( Fig. 1B) supported efficient unwinding reactions, consistent with the notion that the SpSSB is a functional homolog of HSSB and ScSSB. The omission of SSB (lane 2), T antigen (lane 3), or ATP (lane 4) resulted in no single-stranded DNA formation. As described below, SpSSB binds to ssDNA in a manner analogous to HSSB.
SpSSB Does Not Support the SV40 Monopolymerase Reaction-It is established that of a number of SSBs examined, only the human, mouse, and Drosophila SSBs support SV40 DNA replication (31,32). We analyzed whether SpSSB would function in the SV40 monopolymerase system. The reaction was assembled as described previously (27) in the presence of in- creasing amounts of SpSSB or HSSB. After incubation, acidprecipitable radioactivity was determined. As shown in Fig. 2B, increasing concentrations of HSSB stimulated DNA synthesis in the presence of T antigen. Similar concentrations of SpSSB did not support DNA synthesis in this reaction, as demonstrated for the ScSSB (25).
Cloning of the Genes Encoding SpSSB-Tryptic peptides from each of the individual subunits of SpSSB were used to obtain amino acid sequences. Two peptide sequences from each subunit were used to design degenerate oligonucleotides for PCR amplification of each of the three genes. Other peptide sequences were used to confirm an open reading frame of each gene. The respective PCR products were used as probes to screen S. pombe cDNA and genomic libraries. We obtained complete cDNA clones for the p68 and p30 subunit genes, and a genomic DNA clone for the p12 gene (see "Materials and Methods"). ssb1 ϩ Gene-We determined the nucleotide sequence from a cDNA clone containing a 2-kb insert for the SpSSB p68 subunit (ssb1 ϩ ) (Fig. 3A). The nucleotide sequence of the 2056-bp cDNA between the pDB20 vector sequence revealed a 5Ј-untranslated region of 23 bp, 1830 bp of an open reading frame (ORF), encoding a 609-amino acid protein with a predicted molecular mass of 68.3 kDa, and 178 bp at the 3Ј-untranslated region with an adjoining poly(A) sequence of 25 nucleotides. This ORF contained all seven peptide sequences obtained from the tryptic-digested p68 subunit. ssb2 ϩ Gene-We also determined the sequence of a cDNA clone containing a 1287-bp insert for the SpSSB p30 subunit gene (ssb2 ϩ ) (Fig. 3B). It contained 161 bp at the 5Ј-untranslated region, 840 bp of ORF, encoding a 279-amino acid protein of 30.3 kDa and 268 bp of a 3Ј-untranslated region fused to a poly(A) sequence of 18 nucleotides. Both peptide sequences obtained from the p30 subunit were found in this ORF.
ssb3 ϩ Gene-We obtained a genomic clone containing a 3-kb insert of the SpSSB p12 subunit gene (ssb3 ϩ ) (Fig. 3C). Partial sequence of this clone showed that the middle 1-kb HindIII fragment included peptide sequences obtained from tryptic peptides of the p12 subunit. The complete nucleotide sequence of this fragment was determined and is shown in Fig. 3D. This sequence contains the entire coding region as well as 461 and 143 bp of upstream and downstream untranslated sequences, respectively. We found that the genomic DNA sequence, corresponding to the second peptide sequence in Fig. 3D, has a 63-bp gap that contained a perfect match to the consensus sequence for a 5Ј-splice site (GTANG), branch (CTRAY), and a 3Ј-splice site (AG) of S. pombe introns (33). We also found another 101-bp sequence with perfect match to intron consensus sequences. The authenticity of the introns was confirmed by direct isolation of the cDNA sequences spanning this region. The nucleotide sequences of cDNAs obtained by PCR using an S. pombe cDNA library were devoid of the regions corresponding to the predicted introns. The ORF, determined from the cDNA encoding the small subunit, contained 104 amino acids, was calculated to be an 11.8-kDa protein, and contained all four peptide sequences obtained from the p12 subunit.
Amino Acid Sequence Comparison among the SpSSB Subunits and the Homologous Subunits of Other Eukaryotes-The amino acid sequences of the SSB subunits from a variety of eukaryotes have been reported (34 -41). We have used the Megalign program to determine the regions of homology (identity and homology) among the respective SSB subunits and the SpSSB subunits.
The amino acid sequence of the p68 subunit of SpSSB showed significant identity and homology to the corresponding subunit of S. cerevisiae ( (Fig. 4B).
Coexpression of the Three SpSSB Genes in E. coli Results in the Formation of a Stable Heterotrimeric Complex-We constructed expression plasmids for each SpSSB subunit using the T7 RNA polymerase expression system in E. coli (26). Each cDNA was cloned into a plasmid containing the T7 RNA polymerase promoter (Fig. 5). DNA sequencing established that the ATG start codon was maintained for each subunit and that there were no amino acid changes made in any of the coding sequences (see "Materials and Methods"). Each subunit of SpSSB expressed individually in E. coli was found to be insoluble and not amenable for biochemical analysis. Attempts to solubilize the proteins using various refolding protocols were unsuccessful (data not shown).
It was reported that active HSSB complex is produced from coexpression of the three genes on the same plasmid in E. coli (42). For this reason, an expression plasmid was constructed that contained all three subunits of SpSSB (Fig. 5). The resulting plasmid (pET19b-rSpSSB) contained a single T7 RNA polymerase promoter and the cDNA for each subunit was preceded by a Shine-Delgano ribosome binding site. This plasmid was used to transform BL21(DE3) and the resulting transformant was examined for the expression of recombinant proteins following SDS-PAGE (Fig. 6, lanes 1 and 2). When the three FIG. 3. Sequences of ssb genes. A, nucleotide sequence of the ssb1 ϩ (p68 subunit) cDNA and the predicted amino acid sequence. Underlined regions correspond to sequences obtained by amino acid sequencing of tryptic peptides. A putative poly(A) addition signal is shown as a double underlined sequence. Numbers on the right indicate the nucleotide position of the cDNA, and numbers in italic indicate the amino acid residues. Asterisk denotes the stop codon. The start codon TTTG-CAAACATGG (initiation codon showed by ϭ) contains only one mismatch from the eukaryotic consensus sequence for translation initiation codons (glycine residues at positions Ϫ9, Ϫ6, and ϩ4 and a purine at Ϫ3) (77). B, nucleotide sequence of the ssb2 ϩ (p30 subunit) cDNA and predicted amino acid sequence. The symbols defined above that are highlighted are similarly defined as in A. The sequence context start codon TACGTGCTCATGG has a good match with the eukaryotic consensus sequences described above. C, restriction map of the 3-kb EcoRI insert fragment from the genomic DNA clone containing the ssb3 ϩ gene (upper) and the middle 1-kb HindIII fragment (bottom). Several restriction sites are indicated; open bars show the exon regions. D, nucleotide sequence of the middle 1-kb Hin-dIII fragment that contains the ssb3 ϩ (p12 subunit) gene and the predicted amino acid sequence derived from the cDNA. The consensus sequences for the 5Ј-splice sites, branch, and 3Ј-splice sites for introns (GTANG, CTRAY, and AG, respectively) (33), are underlined. The MluI site is indicated by the dotted line. Other noted regions are the same as those referred to in A. The sequence context start codon TAAAATAATATGG has a good match with the eukaryotic consensus sequences.
subunits were expressed together, approximately 50% of the recombinant protein was soluble (data not shown).
Purification of the recombinant SpSSB expressed in E. coli was carried out using the purification procedure used to isolate native SpSSB from S. pombe cells with modifications (see "Materials and Methods"). Fig. 6 shows the SDS-PAGE analysis of the material obtained during the isolation. The soluble protein fraction from induced cells carrying pET19b-rSpSSB was successfully chromatographed through Affi-Gel blue and ssDNA cellulose columns. To remove contaminating ssDNA binding activity presumably contributed by E. coli, the pooled fraction after elution from ssDNA cellulose was chromatographed through a phosphocellulose column. The recombinant SpSSB had the same sedimentation coefficient as the native SpSSB measured by glycerol gradient centrifugation. No partial assembled SpSSB subunits were observed in other fractions in this gradient (data not shown). We conclude that recombinant SpSSB is a stable complex with a structure similar to SpSSB isolated from S. pombe (Fig. 6, lanes 7 and 8).
The Recombinant and Native SpSSBs Show Equivalent Activities-We analyzed the binding of native and recombinant SpSSB, as well as HSSB, to ssDNA using a gel mobility shift assay. The protein preparation used for the following assays was that described in Fig. 6, lanes 7-9. A ssDNA of 45 nucleotides was used as the substrate for the binding reaction. With the addition of increasing amount of protein, two complexes (C1 and C2) were detected with the native and recombinant SpSSB and HSSB (Fig. 7A). The efficiency of binding for the three SSBs was essentially similar (Fig. 7B). From this, we estimated that a monomer of SpSSB and HSSB occupied approximately 16 -22 nucletotides of ssDNA, consistent with previous reports that HSSB occupied 20 -30 nucleotides of ssDNA (29,43,44).
Since HSSB and ScSSB bind preferentially to ssDNA (25,27), competition experiments were carried out to measure the binding efficiency of SpSSB to ssDNA and dsDNA. The levels of the slow migrating complexes C1 and C2 were quantitatively decreased by the addition of competitor ssDNA, whereas dsDNA had no effect (Fig. 8). SpSSB was more sensitive to competition by ssDNA than by dsDNA. Similar results were obtained for HSSB in this assay.
In addition, the native and recombinant SpSSB were equally active in the unwinding of SV40 origin containing DNA (data not shown). Based on these results, we conclude that the recombinant and native SpSSBs have identical biochemical properties.
Cell Cycle-dependent Phosphorylation of SpSSB-We initially observed that the p30 subunit of the purified SpSSB migrated as multiple bands following SDS-PAGE (see Fig. 1, A  and B). The possibility existed that these were either degradation products of the largest subunit or alternatively, modified derivatives of intact p30. It is known that the middle subunit of HSSB and ScSSB are phosphorylated in a cell cycle-dependent manner, being phosphorylated at G 1 /S through to M, when dephosphorylation occurs (45). The SpSSB preparation, isolated from S. pombe cells, was treated with alkaline phosphatase and subjected to SDS-PAGE. As shown in Fig. 9A, modified species were no longer present, and only the faster migrating species was detected (Fig. 9A, compare lanes 7 and  8). This suggests that the p30 subunit of SpSSB exists in different phosphorylated states in vivo. To determine whether phosphorylation occurred in a cell cycle-dependent manner, we used cell division cycle (cdc) mutants of S. pombe that arrested at known stages of the cell cycle at the nonpermissive temperature (20). Cells were grown to early log phase and shifted to FIG. 3-continued the nonpermissive temperature for 4 h for cell cycle arrest, and then extracts were prepared and analyzed by SDS-PAGE and immunoblotting using affinity-purified p30 antibodies. At the same time, cells were processed for FACS analysis to monitor DNA content to confirm their duration in the cell cycle. With the cdc10-129 mutant, the DNA content was 1 N indicating a G 1 arrest. In the cdc22-m45 and cdc25-22 mutants, their DNA contents were consistent with a S and G 2 arrest, respectively. The data from the immunoblot experiments demonstrated that the predominant fast migrating species of p30 subunit was observed only in the G 1 extract, whereas multiple slow migrating species were observed in the S and G 2 extracts (Fig. 9A.  lanes 2-5). To confirm that phosphorylation of the p30 subunit occurred as cells traversed the G 1 to S boundary, cdc10-129 cells were synchronized by first arresting their growth at the nonpermissive temperature followed by release at the permissive temperature. Extracts were prepared at various times and FACS analysis was used to monitor the DNA content of cells.
As shown in Fig. 9B, the p30 subunit was poorly phosphorylated between 0 and 40 min; phosphorylated derivatives, however, accumulated after incubation from 40 to 240 min. The majority of cells after incubation at the permissive temperature for 220 min were in the M phase (data not shown). The FACS analysis indicated that cells progressed through the G 1 /S boundary between 40 and 80 min. We conclude that the p30 subunit of SpSSB is hypophosphorylated in G 1 and becomes phosphorylated at the G 1 /S boundary and is maintained in this state through the M phase. These data are similar to the cell cycle-dependent phosphorylation of the middle subunit of the human and S. cerevisiae SSBs (45). DISCUSSION We have purified to homogeneity the SSB from S. pombe and shown that it is a heterotrimeric complex of 68, 30, and 12 kDa. Homologous SSBs isolated from a variety of widely divergent eukaryotes, including human, mouse, S. cerevisiae, Drosophila, Xenopus, and the trypanosome C. fasciculata, have a similar heterotrimeric structure. All of these multimeric SSBs characterized have strong affinity for ssDNA and supports the SV40 T antigen-dependent unwinding of SV40 origin containing DNA (25, 46 -48). The binding of ssDNA appears to be an important factor governing the T antigen-mediated unwinding reaction. This notion is supported by the observation that E. coli SSB, adenovirus DNA-binding protein, and herpesvirus ICP8 also support the unwinding reaction. However, other SSBs, such as the T4 gene product 32 and the T7 gene product 2.5 do not (49 -51). These findings suggest that the sequestration of ssDNA generated by the T antigen may not be the complete explanation for the role of a DNA-binding protein in the unwinding reaction.
At present, the most stringent assay available for determining the biological activity of the multisubunit DNA binding protein is its ability to support the SV40 T antigen-dependent replication of SV40 ori ϩ DNA. This reaction involves the direct interaction of T antigen with HSSB and the interactions between these two proteins and the DNA polymerase ␣-primase complex (21,31,52,53). In light of the marked species specificity involved in the parvovirus replication systems, it is not surprising that the evolutionarily divergent ScSSB, SpSSB, and the C. fasciculata SSB are unable to support this replication system. However, the SSB from mouse and Drosophila can support the SV40 replication reaction (25,31,32,48). The data suggest that the mouse and Drosophila SSBs have functional features that are conserved with HSSB, whereas the others do not.
We have isolated and determined the nucleotide sequence of each of the genes encoding the p68, p30, and p12 subunit of SpSSB. The p12 gene contains two introns. Coexpression of these subunits in E. coli resulted in the formation of an active heterotrimeric complex, as assayed by ssDNA binding and the T antigen-dependent unwinding reaction of SV40 ori containing DNA following by glycerol gradient sedimentation analysis. An active heterotrimeric complex was similarly observed for the reconstituted HSSB (42). It is not clear how coexpression of these subunits lead to the formation of an active complex, but it is likely that a coordinated folding event of all three subunits is important.
The amino acid sequence of the largest subunit of SSB is highly conserved among the various species analyzed here. All have a C 4 -type zinc finger motif, as previously reported, which spans amino acids 477-498 of SpSSB. It is noteworthy that the C-terminal two-thirds region of the largest subunit (amino acids 179 -609 of SpSSB) shows the highest degree of conservation. Interestingly, the C. fasciculata largest subunit equivalent lacks a segment that corresponds to approximately onethird of the N-terminal region present in the p68 homologs of human, frog, nematoda, and the two yeasts (40). As described above, C. fasciculata SSB is heterotrimeric, has ssDNA binding activity, and supports the unwinding reaction, indicating that the well conserved C-terminal two-thirds region is ample for these functions. Gomes and Wold (54) concluded from a deletional analysis of the HSSB p70 subunit that amino acids 1-411 of HSSB define the ssDNA binding domain (core is 169 -249). They also concluded that the conserved zinc finger motif is not involved in ssDNA binding. In addition, the 507amino acid C-terminal end region of HSSB p70 is involved in complex formation with p34 and p14 (54). These results are consistent with idea that the conserved C-terminal regions are FIG. 4-continued required for DNA binding and for complex formation with the middle and/or smallest SSB subunits. Based on the homology, we anticipate that it may also be true for S. pombe SSB, but it remains to be determined by further analysis.
Significant homology was found among the p30 subunit amino acid sequences of S. pombe and those of the other species. Lee and Kim (55) reported that N-terminal and C-terminal deletion derivatives of the p34 subunit of HSSB formed a heterotrimeric complex with the p70 and p14 subunits. The N-terminal deletion lacked amino acids 2-30, and the C-terminal deletion lacked 33 amino acids from the end of the protein.
The HSSB complex containing the p34 N-terminal deletion was as active as native HSSB in supporting the SV40 in vitro DNA replication reaction, whereas the C-terminal deletion did not.
The reason for the lack of function of the latter, is that it did not support the T antigen catalyzed unwinding of origin containing DNA (55). The alignment of the amino acid sequence of the middle subunit (Fig. 4B) reveals three conserved regions: region 1 (amino acid 52-113 of SpSSB p30 subunit), region 2 (126 -167), and region 3 (257-274). The nonfunctional C-terminal deletion of the HSSB p34 subunit lacks region 3, and the  N-terminal deletion lacks a region that is not conserved. From this, we suggest that conserved region 1 and/or region 2 are involved in complex formation with the p70 and p14 subunits of HSSB. However, the function of these conserved regions remains to be determined.
Our results show that the p30 subunit of the purified SpSSB migrated as four distinct bands following SDS-PAGE (Fig. 1B). Immunoblotting experiments using affinity-purified p30 polyclonal antibodies indicated that the bands were derivatives of the p30 protein. Furthermore, alkaline phosphatase treatment resulted in a change of the slower mobility species to the faster p30 migrating form following SDS-PAGE analysis. Our results indicate that the phosphorylation occurs at the G 1 /S boundary through the M phase, similar to the cell cycle-dependent phosphorylation observed for the p34 subunit of HSSB and ScSSB (45). The Cdk-cyclin A complex and the DNA-dependent protein kinase phosphorylate the HSSB p34 subunit in vitro (56,57). However, Pan et al. (58) reported that the phosphorylated and hypophosphorylated HSSB functioned similarly in supporting the SV40 in vitro replication reaction and nucleotide excision repair in vitro. In addition, HSSB containing p34 subunit mutants with Ser to Ala substitution at Ser-23 and Ser-29 (Cdk kinase consensus phosphorylation sites) had similar activity as the native HSSB in the in vitro replication reaction (55,59). The mutant complexes became phosphorylated under these conditions (55,59). However, it remains to be determined what kinases affect the modifications of SpSSB and the biological significance of this cell cycle-dependent phosphorylation.
In contrast to the conserved regions of the p68 and p30 subunits, the p12 subunit showed the least homology among the known SSBs. Gene disruption experiments show that each of the three SSB genes is essential for cell viability in S. cerevisiae (36). Antibodies against each of the three subunits of HSSB inhibited the SV40 replication reaction, suggesting that each subunit has a specific function(s) in DNA replication (35,39,60).
In addition to replication, the trimeric SSB is also required for DNA recombination (61) and nucleotide excision repair in cell-free systems (62)(63)(64). Recently, the human repair proteins XPA and XPG were reported to bind to HSSB (65)(66)(67)(68). Mutations in the ScSSB p70 subunit gene (rfa1) show defective phenotypes in cell growth, repair, and recombination in yeast cells (69 -71). SSB may be involved in transcription since Sc-SSB was identified as a factor that binds to sequence elements involved in regulating transcription (72,73), and HSSB also binds to the acidic domain of transcriptional factor VP16 and p53 (74 -76). Since DNA replication, repair, and recombination are highly conserved in eukaryotes, the conserved region among the trimeric SSBs may be essential for interaction with factors involved in these processes. FIG. 9. Influence of cell cycle on the phosphorylation of the SpSSB p30 subunit. A, S. pombe cdc temperature-sensitive mutants were grown to A 600 of 0.5 at 25°C in YE medium and then shifted to 36°C for 3 h. Cells were collected by centrifugation and lysed with glass-beads in buffer A ϩ 0.2 M NaCl with 1 mM sodium vanadate and 10 mM sodium fluoride. The lysate (25 g) was subjected to SDS-PAGE (12.5%) analysis (lanes 1-5) and the SpSSB p30 subunit was detected by immunoblotting using affinity-purified polyclonal antibodies against the SpSSB p30 subunit. Purified SpSSB (lane 7) treated with calf intestinal alkaline phosphatase (lane 8) was used as a control to show the difference in migration between the phosphorylated and unphosphorylated p30 protein. S. pombe strains used were: wt, wild type (972h Ϫ ); cdc10, cdc10-129; cdc22, cdc22-m45; cdc25, cdc25-22; and cdc2, cdc2-33, respectively. B, S. pombe cdc10-129 cells were grown in YE medium to A 600 of 0.2 at 25°C and then shifted to 37°C for 4 h. The cell culture was then shifted to 25°C, and samples were removed at the indicated time. The p30 subunit of SpSSB was detected by immunoblotting. FACS analyses were done to determine the DNA content at the indicated time. Lysate preparation and immunoblotting were done as described in A.