Fidelity and Damage Bypass Ability of Schizosaccharomyces pombe Eso1 Protein, Comprised of DNA Polymerase η and Sister Chromatid Cohesion Protein Ctf7*

DNA polymerase η (Polη) functions in error-free bypass of ultraviolet light-induced DNA lesions, and mutational inactivation of Polη in humans causes the cancer prone syndrome, the variant form of xeroderma pigmentosum (XPV). BothSaccharomyces cerevisiae and human Polη efficiently insert two adenines opposite the two thymines of a cyclobutane pyrimidine dimer. Interestingly, in the fission yeastSchizosaccharomyces pombe, theeso1 + encoded protein is comprised of two domains, wherein the NH2 terminus is highly homologous to Polη, and the COOH terminus is highly homologous to the S. cerevisiae Ctf7 protein which is essential for the establishment of sister chromatid cohesion during S phase. Here we characterize the DNA polymerase activity of S. pombe GST-Eso1 fusion protein and a truncated version containing only the Polη domain. Both proteins exhibit a similar DNA polymerase activity with a low processivity, and steady-state kinetic analyses show that on undamaged DNA, both proteins misincorporate nucleotides with frequencies of ∼10−2 to 10−3. We also examine the two proteins for their ability to replicate a cyclobutane pyrimidine dimer-containing DNA template and find that both proteins replicate through the lesion equally well. Thus, fusion with Ctf7 has no significant effect on the DNA replication or damage bypass properties of Polη. The possible role of Ctf7 fusion with Polη in the replication of Cohesin-bound DNA sequences is discussed.

DNA polymerase (Pol) functions in error-free bypass of ultraviolet light-induced DNA lesions, and mutational inactivation of Pol in humans causes the cancer prone syndrome, the variant form of xeroderma pigmentosum (XPV). Both Saccharomyces cerevisiae and human Pol efficiently insert two adenines opposite the two thymines of a cyclobutane pyrimidine dimer. Interestingly, in the fission yeast Schizosaccharomyces pombe, the eso1 ؉ encoded protein is comprised of two domains, wherein the NH 2 terminus is highly homologous to Pol, and the COOH terminus is highly homologous to the S. cerevisiae Ctf7 protein which is essential for the establishment of sister chromatid cohesion during S phase. Here we characterize the DNA polymerase activity of S. pombe GST-Eso1 fusion protein and a truncated version containing only the Pol domain. Both proteins exhibit a similar DNA polymerase activity with a low processivity, and steady-state kinetic analyses show that on undamaged DNA, both proteins misincorporate nucleotides with frequencies of ϳ10 ؊2 to 10 ؊3 . We also examine the two proteins for their ability to replicate a cyclobutane pyrimidine dimer-containing DNA template and find that both proteins replicate through the lesion equally well. Thus, fusion with Ctf7 has no significant effect on the DNA replication or damage bypass properties of Pol. The possible role of Ctf7 fusion with Pol in the replication of Cohesin-bound DNA sequences is discussed.
In Saccharomyces cerevisiae as well as humans, DNA polymerase functions in error-free replication of UV-damaged DNA. In S. cerevisiae, inactivation of Pol confers enhanced UV sensitivity and leads to an increase in UV-induced mutation frequencies (1)(2)(3). In humans, inactivation of Pol results in the cancer prone syndrome, the variant form of xeroderma pigmentosum (4,5). Cells from variant form of xeroderma pigmentosum patients display a deficiency in the replication of UVdamaged DNA (6 -8), and they are hypermutable with UV light (9,10).
Pol is unique among eukaryotic DNA polymerases in its proficient ability to replicate through DNA lesions which distort the DNA helix (5,(11)(12)(13)(14). Both yeast and human Pol replicate through a cis-syn thymine-thymine (T-T) dimer with the same efficiency and accuracy as they replicate through the two undamaged Ts (12,15). The ability of Pol to insert nucleotides opposite distorting DNA lesions and to carry out extension of the nascent DNA strand has suggested that Pol is more tolerant of geometric distortions in DNA than are other DNA polymerases which cannot bypass DNA lesions. Accordingly, both S. cerevisiae and human Pol are low fidelity enzymes, misincorporating nucleotides with a frequency of 10 Ϫ2 to 10 Ϫ3 (12,16,17).
Although Pol is a low fidelity enzyme, it does not contribute to spontaneous mutagenesis, since the rate of spontaneous mutations at several loci examined remains the same in the presence or absence of Pol in S. cerevisiae (18). 1 Thus, Pol may have little or no effect on normal replication, and its function may be primarily restricted to promoting replication through DNA lesions. The Rad6-Rad18 complex, comprised of the ubiquitin conjugating and DNA binding activities (19,20), may be a key factor in limiting Pol action to damage bypass. Although the mechanism of the Rad6-Rad18 enzyme complex remains unknown, it is possible that ubiquitin conjugation by the Rad6-Rad18 complex leads to dissociation of some protein(s) from the replication machinery stalled at a lesion site, and that, in turn, promotes the assembly of a trans-lesion DNA synthesis polymerase such as Pol into the stalled replication complex. Alternatively, Pol could be kept in an inactive state by its binding to other protein(s), and upon the infliction of damage to DNA, the Rad6-Rad18 mediated protein ubiquitination may stimulate the dissolution of the inhibitory protein(s), thereby activating Pol.
Interestingly, in the fission yeast Schizosaccharomyces pombe, the eso1 ϩ -encoded protein is comprised of two domains, of which the NH 2 -terminal two-thirds is highly homologous to S. cerevisiae and human Pol, and the COOH-terminal onethird is highly homologous to the S. cerevisiae Ctf7 protein (also called Eco1) (21), an essential protein required for the establishment of sister chromatid cohesion during S phase (21)(22)(23). Deletion analyses have indicated that the COOH-terminal Ctf7 portion of Eso1 is sufficient and necessary for sister chromatid cohesion, whereas deletion of the NH 2 -terminal Pol portion increases the sensitivity to UV irradiation but has no effect on sister chromatid cohesion (21). Thus, although the two proteins are encoded by the same gene, they retain their respective functions in sister chromatid cohesion and damage bypass. The fusion of Pol with Ctf7 in S. pombe raised the possibility that although the two proteins are encoded by separate genes in other species, even there these proteins may associate in vivo, and that association may modulate the function of one or both the proteins. For instance, it could be that association with Ctf7 * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  inactivates Pol and its activation requires that the two proteins dissociate following the Rad6-Rad18-dependent ubiquitination of one or both proteins; alternatively, association with Ctf7 could improve the fidelity and processivity of Pol. Here, we purify the S. pombe eso1 ϩ -encoded protein containing both the Pol and Ctf7 domains, and compare the DNA synthesis properties of Pol alone and of Pol fused to Ctf7 in Eso1. Unexpectedly, we find that fusion with Ctf7 has no effect on Pol's DNA synthesis or damage bypass ability, or on its fidelity or processivity. We discuss these observations in relation to the possible role of Pol's association with Ctf7.

MATERIALS AND METHODS
Expression of S. pombe Eso1 and Pol Proteins-To obtain the fulllength Eso1-(1-872) protein and the truncated Eso1-(1-609) protein which lacks the Ctf7 domain but contains the entire Pol domain, herein referred to as Pol, the eso1 ϩ gene was amplified from the S. pombe strain JFP41 (h ϩ ura4-D18 ade6-M216 leu1-32) total genomic DNA by polymerase chain reaction using oligonucleotide N7119 (5Ј-CAGGGGTACCG-GATCCACATATGGAATTAGGCAAAAGCAAATTCTC-3Ј) and the oligonucleotide N7118 (5Ј-GGTCGTCGACGGATCCTCAACTTTCATA-AACAGCATATCGAAG-3Ј) or the oligonucleotide N7117 (5Ј-CAGGGTC-GACGGATCCTCATCTTTTGTTGTTTGTTTCATCGG-3Ј), respectively. The amplified DNAs were then digested with Asp-718 and SalI restriction endonucleases and cloned into YIplac211, generating plasmids pBJ796 and pBJ800, respectively. The cloned polymerase chain reaction fragments in pBJ796 and pBJ800 were sequenced and found not to contain any mutations. Subsequently, the wild type and truncated eso1 genes were cloned in-frame with the glutathione S-transferase (GST) 2 gene under the control of a galactose-inducible phosphoglycerate kinase promoter in plasmid pBJ760 (24), generating the expression vectors pBJ808 and pBJ811, respectively. The GST-tagged proteins were expressed in the S. cerevisiae strain BJ5464 harboring either the plasmid pBJ808 or pBJ811. S. cerevisiae cells were grown overnight in synthetic complete medium lacking leucine (SC-leu) and containing 2% dextrose, 2.5% lactate, and 3% glycerol before diluting 30-fold in SC-leu as above but lacking dextrose. After 16 h, 2% galactose was added and cells were grown for an additional 7 h before harvesting by centrifugation.
Purification of S. pombe Pol and Eso1 Proteins-To purify GST-Eso1 and GST-Pol proteins, cells were resuspended in 2 volumes of ice-cold cell breakage buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 300 mM NaCl, 10% sucrose, 0.5 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, and one Complete Protease Inhibitor Mixture Tablet per 50 ml of extract (Roche Molecular Biochemicals)), and lysed at 4°C in a French Press at 120,000 kilopascal. Cell debris was removed by centrifugation at 100,000 ϫ g, and cell extract was passed over a 100 l of glutathione-Sepharose 4B column (Amersham Pharmacia Biotech) at 4°C. The beads were washed with 10 volumes ice-cold cell breakage buffer containing 1 M NaCl, followed by 5 volumes ice-cold low salt buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 50 mM NaCl, 10% glycerol). The GST-Eso1 and the GST-Pol proteins were each batch eluted twice at 4°C with 100 l of elution buffer (100 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.01% Nonidet P-40, 10% glycerol, 25 mM glutathione). Aliquots of each purified protein were stored at Ϫ80°C.
For examining replication in the presence of a cyclobutane-pyrimidine dimer (CPD), we used the following 75-nucleotide oligomer as the template: 5Ј-AGCTACCTAGCCTGCACGAAGAGTTCGTATTATGCCT-ACACTGGAGTACCGGAGCATCGTCGTGACTGGGAAAAC. The CPD template contained a cis-syn thymine-thymine (T-T) dimer at the underlined region, whereas the nondamaged control template contained two undamaged thymines. We used the following oligodeoxynucleotide primer for CPD bypass analysis: 44-mer, 5Ј-GTTTTCCCAGTCACG-ACGATGCTCCGGTACTCCAGTGTAGGCAT.
Primers were 5Ј-32 P-end-labeled using polynucleotide kinase (Roche Molecular Biochemicals) and [␥-32 P]ATP (Amersham Pharmacia Biotech). Labeled primers (0.25 or 0.05 M) were annealed to templates (0.5 or 0.1 M, respectively) in the presence of 50 mM Tris-HCl, pH 7.5, and 100 mM NaCl by heating the mixture to 90°C for 2 min and cooling to 25°C over several hours.
DNA Polymerase Activity Assays-The standard DNA polymerase assay (5 l) contained 25 mM Tris-HCl, pH 7.5, 100 ng/ml bovine serum albumin, 5 mM dithiothreitol, 5 mM MgCl 2 , and 10% glycerol, and was carried out for 5 min at 25°C. For DNA synthesis, a 50 nM 5Ј-32 P-endlabeled primer-template G substrate and 100 M of each of the four deoxynucleotides were used, and reactions were initiated by the addition of various concentrations (0 to 25 nM) of purified GST-Eso1 or GST-Pol proteins. Reactions were quenched in 10 volumes of formamide-loading buffer (80% deionized formamide, 10 mM EDTA, pH 8.0, 1 mg/ml xylene cyanol, 1 mg/ml bromphenol blue), heated at 90°C for 2 min, cooled on ice, and resolved on 10% polyacrylamide sequencing gels containing 5.2 M urea.
DNA polymerase fidelity assays were carried out under standard conditions but contained 50 nM 5Ј-32 P-end-labeled primer-template substrate (Template G, A, T, or C), and various concentrations of a single deoxynucleotide (0 to 2000 M). The reactions were initiated by the addition of 5 nM purified GST-Eso1 or GST-Pol, and were quenched after 2 to 20 min in 10 volumes of formamide-loading buffer. Gel band intensities of the substrates and products were quantitated using a PhosphorImager with ImageQuant software (Molecular Dynamics).
T-T dimer bypass assays were carried out under standard conditions with 10 nM 5Ј-32 P-end-labeled primer-CPD template substrate, and no deoxynucleotide, one of the four deoxynucleotides at 100 M each, or all four deoxynucleotides at 100 M each. The reactions were initiated by the addition of 5 nM purified GST-Eso1 or GST-Pol, and quenched after 5 min in 10 volumes of formamide-loading buffer.
Analysis of Fidelity-For each concentration of nucleotide, the relative amount of the extended product was divided by the reaction time, resulting in the linear observed rate of nucleotide incorporation, v obs .
The v obs was plotted as a function of deoxynucleotide concentration, and the data were fit by nonlinear regression using SigmaPlot 4.0 to the Michaelis-Menten equation (Equation 1), V max and K m steady-state kinetic parameters for the incorporation of the correct and incorrect deoxynucleotides were obtained from the bestfit curve. These parameters were used to calculate the frequency of deoxynucleotide misincorporation (f inc ) as described (25,26) using the following equation (Equation 2), Processivity Assays-Processivity was measured by preincubating 50 nM GST-Eso1 or GST-Pol with 50 nM 5Ј-32 P-end-labeled primer-template DNA substrate under standard conditions for 1 h. Reactions were initiated by the addition of all four deoxynucleotides (100 M of each dNTP), and 1 mg/ml sonicated herring sperm DNA as a trap. To demonstrate the effectiveness of the trap, we performed a control reaction in which GST-Eso1 or GST-Pol was preincubated with the trap DNA and the 5Ј-32 P-end-labeled primer-template substrate for 1 h before the addition of deoxynucleotides. After 15 or 30 s, reactions were quenched and resolved as described above for the deoxynucleotide incorporation assays.
By definition, the processivity, P, is the probability that for each nucleotide incorporation event, the polymerase moves ahead to incorporate at least one additional nucleotide (27). To determine the processivity of Eso1p and Pol, we first quantitated the gel band intensities of the nucleotide incorporation products at the 15-s reaction time point using the PhosphorImager. Next, we calculated the fraction of polymerase molecules that incorporated at least N nucleotides, N , which is the ratio of the intensity of all gel bands greater than or equal to the gel band corresponding to N nucleotide incorporations (gel band N) to the total intensity of all gel bands (excluding the unextended primer). Formally, this is expressed in the following equation (Equation 3), The abbreviations used are: GST, glutathione S-transferase; CPD, cyclobutane pyrimidine dimer.
where I x is the intensity of gel band x. In a single DNA binding event, it takes N-1 consecutive steps, each occurring with a probability P, starting at gel band 1 to reach any gel band N. Thus, the relationship between the fraction of polymerase molecules that incorporated at least N nucleotides, N , and the processivity, P, is expressed in the following equation (Equation 4), Consequently, we graphed N , the fraction of polymerase molecules that incorporated at least N nucleotides, versus N-1, and obtained a value for the processivity, P, from the best fit curve to Equation 4 using nonlinear regression (SigmaPlot 4.0). The average number of nucleotides incorporated per DNA binding event (1/[1 Ϫ P]) was then calculated.

RESULTS
The Eso1 Protein of S. pombe-The amino-terminal twothirds of the S. pombe eso1 ϩ -encoded protein shares 23% identical and 34% similar amino acid residues with the S. cerevisiae Pol, and the COOH-terminal one-third of the protein shares 24% identical and 44% similar residues with the S. cerevisiae Ctf7 protein (21), which is highly conserved among eukaryotes and plays an essential role in the establishment of sister chromatid cohesion during DNA replication (21-23). Fig. 1 depicts the conserved motifs shared between Eso1 and the S. cerevisiae Pol and Ctf7 proteins. Two C 2 H 2 zinc finger motifs are present in Eso1, one of which corresponds to the motif present toward the COOH terminus of Pol and the other corresponds to the motif present toward the NH 2 terminus of Ctf7. The five highly conserved motifs found in the Rad30/UmuC/DinB protein family (28), are present in the Pol portion of Eso1.
To test for the influence of the COOH-terminal Ctf7-like portion on the polymerase activity of Eso1 protein, we expressed in S. cerevisiae both the full-length Eso1 protein (872 amino acids) and a truncated Eso1 protein that contains only the NH 2 -terminal 609 amino acids, and which corresponds to the Pol domain, as fusions with the glutathione S-transferase protein. The GST-Eso1 and GST-Pol proteins were affinity purified to near homogeneity ( Fig. 2A) and their DNA synthesis and damage bypass properties were analyzed.
DNA Polymerase Activity of S. pombe Eso1 and Pol Proteins-To test for DNA polymerase activity of S. pombe Eso1 and Pol proteins, various concentrations of purified proteins were incubated with labeled primer-template DNA substrate in the presence of all four deoxynucleotides, and reaction products were resolved on a denaturing polyacrylamide gel. As shown in Fig. 2B, both proteins exhibit nearly equivalent DNA polymerizing ability.
Fidelity of S. pombe Eso1 and Pol Proteins-Fidelity measures the likelihood that a polymerase will incorporate the correct versus the incorrect deoxynucleotide opposite a template residue. We used steady-state kinetics to measure the fidelity of S. pombe Eso1 and Pol proteins opposite all four undamaged template nucleotides, as described under "Materials and Methods." The incorporation of either the correct or incorrect deoxynucleotide was quantitated and used to calculate the V max and K m values.
S. pombe Eso1 protein (5 nM) was incubated with the primertemplate DNA and various concentrations of one of the four deoxynucleotides in a standing-start reaction. As shown in Fig.  3A, to examine the insertion opposite template G, the concentration of incorrect and correct deoxynucleotide was varied from 0 to 1000 M, and from 0 to 20 M, respectively. The kinetics of single deoxynucleotide incorporation by S. pombe Eso1 protein opposite the template G are shown in Fig. 3B. These data were fit to the Michaelis-Menten equation (Equation 1), and used to calculate the apparent K m and V max parameters. As shown in Table I, the f inc values for the Eso1 protein range from 2.4 ϫ 10 Ϫ4 to 6.5 ϫ 10 Ϫ3 . The f inc values were similarly calculated for the Pol protein. As shown in Table II, Pol misincorporates nucleotides with nearly the same frequencies as the Eso1 protein. Fig. 4 compares the efficiencies (V max /K m ) of correct and incorrect deoxynucleotide incorporation opposite templates G, A, T, and C by S. pombe Eso1 and Pol proteins. This comparison shows that the two proteins incorporate nucleotides with nearly the same efficiencies. Thus, the presence of the Ctf7 protein in the COOH terminus of Eso1 has little, if any, effect on Pol's ability to incorporate the correct or wrong nucleotides opposite undamaged template bases.

Processivity of S. pombe Pol and Eso1
Proteins-Processivity is a measure of the number of deoxynucleotides a polymerase incorporates in a single DNA binding event. Processivity is expressed quantitatively as the probability, P, that following each nucleotide incorporation, the polymerase will move ahead to incorporate at least one additional deoxynucleotide (27). To ensure that we were measuring the activity of a single DNA binding event, we included excess, nonradiolabeled sonicated herring sperm DNA to trap any polymerase molecules that dissociated from the DNA. The Eso1 (Fig. 5A, lanes 1-3) or Pol (Fig. 5A, lanes 7-9) proteins were preincubated with radiolabeled primer-template DNA substrate for 1 h before the addition of excess herring sperm DNA and all four deoxynucleotides. To determine that the excess herring sperm DNA was indeed sufficient to prevent the re-binding of Eso1 or Pol proteins to the radiolabeled DNA substrate, the Eso1 (Fig. 5A,  lanes 4 -6) or Pol (Fig. 5A, lanes 10 -12) proteins were prein- cubated with the radiolabeled primer-template DNA substrate and the excess DNA trap for 1 h before the addition of all four deoxynucleotides. The lack of any DNA synthesis in these reactions (Fig. 5A, lanes 4 -6 and 10 -12) confirmed the adequacy of excess herring sperm DNA to trap all Eso1 or Pol molecules.
The gel band intensities of the nucleotide incorporation products at the 15-s time point were used to calculate the processivity, P, and the average number of nucleotides incorporated per DNA binding event (see "Materials and Methods"). Fig. 5B is a plot of N , the fraction of polymerase molecules that incorporated at least N nucleotides, versus N-1 for the Eso1 protein.
From the best fit curve to Equation 4, we obtained a value for P equal to 0.53 Ϯ 0.01, which means that for each nucleotide incorporation event, the Eso1 protein has a 53% chance of moving ahead to incorporate at least one additional nucleotide. Thus, the average number of nucleotides incorporated by the Eso1 protein per DNA binding event is 2.1. Likewise, Fig. 5C is the analogous plot for Pol. From the best fit curve to Equation 4, we obtained a value for P equal to 0.61 Ϯ 0.01. Thus, the average number of nucleotides incorporated by Pol per DNA binding event is 2.6.
T-T Dimer Bypass by S. pombe Eso1 and Pol Proteins-The ability of Eso1 and Pol proteins to replicate through a cis-syn T-T dimer was assessed using a standing start assay. As shown in Fig. 6A, both proteins are able to incorporate a deoxynucleotide opposite the 3Ј T of the dimer and to extend from it, and the DNA synthesis activity on the damaged DNA is as robust   FIG. 3. Kinetics of nucleotide incorporation by S. pombe Eso1 protein. A, deoxynucleotide incorporation opposite a template G residue. Eso1 protein (5 nM) was incubated for 2-20 min at 25°C with the primer-template DNA substrate (50 nM) and increasing concentrations of nucleotide. The samples were quenched and analyzed by denaturing PAGE. The unextended primer (n ϭ 0) and the extended primers (N ‫؍‬ 1 and 2) are indicated. B, quantitation of deoxynucleotide incorporation reactions. For each deoxynucleotide, the observed rate of deoxynucleotide incorporation is graphed as a function of deoxynucleotide concentration. The data were fit using Equation 1, and the resulting V max and K m parameters are listed in Table I.  as on the nondamaged DNA. To identify the nucleotide incorporated by Eso1 and Pol proteins opposite the two Ts of the dimer, we incubated each enzyme for 5 min in the presence of dimer containing DNA substrate and containing a single deoxynucleotide G, A, T, or C, at 100 M concentration. As shown in Fig. 6B, both proteins primarily incorporate an A residue opposite each of the Ts of the T-T dimer.

DISCUSSION
The fusion of Pol to the sister chromatid cohesion protein Ctf7 in the S. pombe Eso1 protein has presented the opportunity to determine whether the fusion with Ctf7 affects Pol function. Our results indicate that the DNA synthesis activity of Pol is not affected by fusion to Ctf7, as the Eso1 and Pol proteins synthesize DNA with remarkably similar efficiencies and fidelities, and on undamaged DNA, they misincorporate nucleotides with frequencies ranging from 2 ϫ 10 Ϫ2 to 2 ϫ 10 Ϫ4 . Also, the two proteins replicate a cis-syn T-T dimer equally well and both predominantly insert two A residues opposite the two Ts of the dimer. In all these properties, the S. pombe Eso1 and Pol proteins resemble S. cerevisiae and human Pol.
Pol is a low processivity enzyme, dissociating from DNA quite frequently. DNA polymerases achieve processive synthesis by associating with a multimeric ring ␤ clamp in Escherichia coli or PCNA in eukaryotes. T7 polymerase increases its processivity by forming a one-to-one complex with E. coli thioredoxin (29,30). The processivity of Pol, however, is not affected by its fusion to Ctf7. Both the Eso1 and Pol proteins exhibit low processivity, inserting 2-3 nucleotides per DNA binding event.
Although Ctf7 does not activate or inactivate the DNA polymerizing activity or T-T dimer bypass ability of Pol, fusion with Ctf7 may enable Pol to function in sister chromatid cohesion. Genetic studies in S. cerevisiae have suggested that a cohesion complex consisting of the Scc1, Scc3, Smc1, and Smc3 subunits is loaded onto chromosomes at the end of G 1 . Another protein, Scc2, although not a stoichiometric Cohesin subunit, is required for the association of the cohesion complex with chromosomes (23). The sixth protein, Ctf7, is neither a subunit of the Cohesin complex nor is it required for the association of Cohesin with chromosomes (23). Ctf7, however, is essential for the establishment of cohesion during DNA replication, but it is not required for the maintenance of cohesion during G 2 and M phases (22,23). One possible role for the fusion of Ctf7 with Pol in the Eso1 protein is that it promotes replication through  1-3) or Pol (lanes 7-9) was preincubated with 50 nM primer-template DNA for 1 h at 25°C, and the reactions were initiated by the addition of 100 M of each of four dNTPs, 5 mM MgCl 2 , and 1 mg/ml sonicated herring sperm DNA trap. Reactions were quenched after 15 or 30 s, and the samples were resolved by denaturing PAGE. The positions of the unextended primer (n ϭ 0) and extended primers (n ϭ 1-11) are indicated. As a control, 1 mg/ml sonicated herring sperm DNA trap was added to the preincubation mixture containing Eso1 (lanes 4 -6) or Pol proteins (lanes 10 -12), and the reactions were initiated by the addition of 100 M of each of four dNTPs and 5 mM MgCl 2 . B and C, graphs of N , the fraction of polymerase molecules that incorporated at least N nucleotides, versus N-1 for the Eso1 and Pol proteins, respectively. The solid lines reflect the best fit curves to Equation 4 and were used to obtain values for the processivity, P, as described under "Materials and Methods." chromosomal sites where Cohesin has been deposited onto DNA, and replication through such sites may be a prerequisite for the formation of protein links between Cohesin-bound sister chromatids.
While the above model explains the requirement of Ctf7 in S. cerevisiae and S. pombe for the establishment of sister chromatid cohesion during S phase, it fails to account for the fact that deletion of Pol has no apparent effect on sister chromatid cohesion in either yeast species. One possible explanation for this discrepancy is the involvement of yet two other highly related proteins, Trf4 and Trf5, in sister chromatid cohesion. A trf4 ts trf5⌬ double mutant is unable to complete S phase, and results in failure of cohesion between the replicated sister chromatids (31). The Trf proteins are members of the ␤-polymerase superfamily, and accordingly, a DNA polymerase activity has been identified in Trf4 (31). A role for the Trf4 and Trf5 polymerases in the replication of Cohesin-bound DNA has been previously proposed (31). The fact that the Trf4 and Trf5 proteins are essential for the establishment of sister chromatid cohesion suggests that these polymerases are indispensable for the replication of Cohesin-bound DNA sequences, whereas the dispensability of Pol for sister chromatid cohesion would suggest that this protein plays a much less critical role in the replication of Cohesin-bound DNA. Pol may have an accessory role in the replication of Cohesin-bound DNA, where it promotes replication through some sites by the Trf4 or Trf5 polymerase. Pol may act at such sites in a manner analogous to its role in damage bypass where it salvages the replication fork stalled at a lesion site.