Specificity of DNA Lesion Bypass by the Yeast DNA Polymerase h *

DNA polymerase h (Pol h , xeroderma pigmentosum variant, or Rad30) plays an important role in an error-free response to unrepaired UV damage during replication. It faithfully synthesizes DNA opposite a thymine-thymine cis-syn -cyclobutane dimer. We have purified the yeast Pol h and studied its lesion bypass activity in vitro with various types of DNA damage. The yeast Pol h lacked a nuclease or a proofreading activity. It efficiently bypassed 8-oxoguanine, incorporating C, A, and G opposite the lesion with a relative efficiency of ; 100: 56:14, respectively. The yeast Pol h efficiently incorporated a C opposite an acetylaminofluorene-modified G, and efficiently inserte d a G orless frequently an A opposite an apurinic/apyrimidinic (AP) site but was unable to extend the DNA synthesis further in both cases. However, some continued DNA synthesis was observed in the presence of the yeast Pol z following the Pol h action opposite an AP site, achieving true lesion bypass. In contrast, the yeast Pol a was able to bypass efficiently a template AP site, predominantly incorporating an A residue opposite the lesion. These results suggest that other than UV damage, Pol h may also play a role in bypassing additional DNA lesions, some of which can be error-prone. repli-cative Cells contain support DNA from the achieving

DNA polymerase (Pol, xeroderma pigmentosum variant, or Rad30) plays an important role in an errorfree response to unrepaired UV damage during replication. It faithfully synthesizes DNA opposite a thyminethymine cis-syn-cyclobutane dimer. We have purified the yeast Pol and studied its lesion bypass activity in vitro with various types of DNA damage. The yeast Pol lacked a nuclease or a proofreading activity. It efficiently bypassed 8-oxoguanine, incorporating C, A, and G opposite the lesion with a relative efficiency of ϳ100: 56:14, respectively. The yeast Pol efficiently incorporated a C opposite an acetylaminofluorene-modified G, and efficiently inserted a G or less frequently an A opposite an apurinic/apyrimidinic (AP) site but was unable to extend the DNA synthesis further in both cases. However, some continued DNA synthesis was observed in the presence of the yeast Pol following the Pol action opposite an AP site, achieving true lesion bypass. In contrast, the yeast Pol␣ was able to bypass efficiently a template AP site, predominantly incorporating an A residue opposite the lesion. These results suggest that other than UV damage, Pol may also play a role in bypassing additional DNA lesions, some of which can be error-prone.
During replication, a variety of DNA lesions can block replicative DNA polymerases. Cells contain specialized proteins to overcome such replication blockage. These proteins can support DNA synthesis across from the damaged template, achieving trans-lesion synthesis or lesion bypass, thus allowing normal replication to continue downstream of the damage. Since DNA lesions may alter the coding property of the modified base or render the base noncoding, trans-lesion synthesis often results in mutations. One mechanism of lesion bypass is the damageinduced mutagenesis. In Escherichia coli, the damage-induced mutagenesis pathway is under the control of the SOS response and requires at least RecA, UmuC, and UmuD proteins (1)(2)(3). In the eukaryotic model organism Saccharomyces cerevisiae, the damage-induced mutagenesis pathway involves at least the following proteins: Rad6, Rad18, Rev1, Rev3, Rev6, Rev7, and Ngm2 (4 -11).
Yeast Rev3 forms a protein complex with Rev7 and is known as DNA polymerase (Pol) 1 (12). This polymerase is capable of limited trans-lesion synthesis opposite a template TT dimer in vitro (12). The Rev1 protein is a dCMP transferase that is able to efficiently insert a C residue opposite a template G or an apurinic/apyrimidinic (AP) site (13). In vitro, the combined activities of the Rev1 transferase and Pol can effectively bypass a template AP site (13). Genetic analyses have demonstrated the importance of this mutagenesis pathway in errorprone lesion bypass of UV-induced damage and AP sites in DNA (7,9,14).
Recently, it became clear that the Rev1 protein belongs to a large protein family known as the UmuC family (15,16). This family also includes the E. coli DinB protein (15,16) and the yeast RAD30 gene product (17,18). Genetic analyses of rad30 mutant cells indicate that its wild-type protein is involved in an error-free response to UV radiation but is independent of the Rad5 error-free mechanism (17). Biochemical studies of the Rad30 protein revealed that it is a DNA polymerase capable of error-free trans-lesion synthesis opposite a template cis-syn-TT dimer (19). This seventh eukaryotic DNA polymerase is thus referred to as DNA Pol (19). In E. coli, DinB is DNA polymerase IV (20), and UmuDЈ 2 C complex is DNA polymerase V (21).
Most recently, it was found that the human xeroderma pigmentosum variant (XPV) is both a structural and functional homologue of the yeast Pol (22)(23)(24). Mutations in the XPV gene are responsible for the rare human hereditary disease xeroderma pigmentosum variant (1,25). But unlike the other xeroderma pigmentosum genes, it has long been recognized that XPV is not involved in nucleotide excision repair (1,25). UV sensitivity and elevated UV mutagenesis of the human XPV cells are now clearly explained by the activities of Pol in response to UV radiation. Intriguingly, humans contain another Rad30 homologue, whose gene is designated RAD30B (16). Additionally, a third human homologue more related to DinB has been identified (26). Apparently, proteins of the UmuC family are involved in different mechanisms of lesion bypass in response to unrepaired DNA damage during replication. The importance of lesion bypass is underscored by the fact that the fundamental mechanisms have been conserved from E. coli to humans.
Pol is able to bypass a template TT dimer in an error-free manner (19,23). However, it is not known whether Pol is also capable of bypassing other DNA lesions. Furthermore, it is not known whether the error-free lesion bypass by Pol is a unique feature confined to TT dimer or is more general to other DNA lesions. To address these questions, we have examined the response of the yeast Pol to several other DNA lesions in vitro. In this report, we (i) show the properties of the yeast Pol in response to an 8-oxoguanine, an AAF-modified guanine, and an AP site in the DNA template; and (ii) provide evidence that Pol can be error-prone during DNA synthesis opposite some lesions.

EXPERIMENTAL PROCEDURES
Materials-A mouse monoclonal antibody against the His 6 tag was purchased from Qiagen. Alkaline phosphatase-conjugated anti-mouse IgG was from Sigma. The Pfu DNA polymerase was from Stratagene. The E. coli uracil-DNA glycosylase was from New England Biolabs. The yeast rad30 deletion mutant strain (MATa his3 leu2 met15 ura3) and its isogenic wild-type strain BY4741 were purchased from Research Genetics. N-Acetoxy-N-2-acetylaminofluorene (AAAF) was obtained from the Midwest Research Institute. Purified yeast DNA polymerase ␣ (Pol␣) with associated primase activity was generously provided by David Hinkle, Department of Biology, University of Rochester. One unit of Pol␣ incorporates 1 nmol of total nucleotide per 30 min at 30°C, using activated salmon sperm DNA as the substrate.
To construct the Rad30 overexpression plasmid, the RAD30 gene was amplified by polymerase chain reaction from yeast genomic DNA using Pfu DNA polymerase and two primers, CATGCCATGGCATCTAGAAT-GTCAAAATTTACTTGGAAGGAG and ACGCGTCGACAAGCTTGTT-GCTGAAGCCATATAATTGTC. The resulting polymerase chain reaction product was cloned into the XbaI and SalI sites of the vector pEGUh6 (27) to yield pEGUh6-RAD30. Its expression in yeast was under the control of the inducible GAL1/GAL10 promoter and produced the Rad30 protein (Pol) containing a His 6 tag at its N terminus.
DNA Substrates with Site-specific Lesions-A 49-mer DNA template containing a site-specific cis-syn-TT dimer was prepared as described previously (28). Its sequence is 5Ј-AGCTACCATGCCTGCACGAATTA-AGCAATTCGTAATCATGGTCATAGCT-3Ј, where the TT dimer is underlined. A 30-mer DNA template containing a site-specific 8-oxoguanine (8-oxoG) was synthesized from a DNA synthesizer by Operon. Its sequence is 5Ј-GGATGGACTGCAGGATCCGGAGGCCGCGCG-3Ј, where the position of the 8-oxoG is underlined.
To prepare AAF-damaged DNA, 2 nmol of the oligonucleotide 5Ј-CCTTCTTCTTCATACAAGCTTACTTCTTCC-3Ј was incubated with 200 nmol of AAAF in 100 l of TE buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) containing 20% ethanol at 37°C in the dark for 3 h. AAAF predominantly modified the unique G in the DNA. After extracting 5 times with water-saturated ether to remove free AAAF, damaged DNA was separated from undamaged oligonucleotide by electrophoresis on a 20% denaturing polyacrylamide gel. Modified DNA migrated slower on the gel and was sliced out of the gel. The gel slices were soaked in 150 l water at room temperature for 4 h. The AAF-damaged 30-mer DNA was recovered using GenElute DNA spin column (Supelco).
To prepare a 30-mer DNA template containing a site-specific AP site, the following oligonucleotide with a uracil residue at position 13, 5Ј-GGATGGCATGCAUTAACCGGAGGCCGCGCG-3Ј, was synthesized first (Operon). The primer 5Ј-CGCGCGGCCTCCGGTTA-3Ј was then labeled with 32 P at its 5Ј end and annealed to the uracil-containing template. Finally, 10 pmol of the uracil-containing substrate was treated with 4 units of E. coli uracil-DNA glycosylase at 37°C for 60 min. The resulting substrate was examined for the presence of the site-specific AP site by cleavage with the E. coli endonuclease III at 37°C for 30 min, followed by electrophoresis on a 15% non-denaturing polyacrylamide gel as described previously (29).
Purification of the Yeast DNA Pol-Yeast cells harboring pEGUh6-RAD30 were grown in minimal medium containing 2% sucrose for 2 days. Expression of Rad30 was induced by diluting the culture 10-fold in 16 liters of YPG (2% Bacto-peptone, 1% yeast extract, 2% galactose) medium supplemented with 0.5% sucrose and incubation for 15 h at 30°C with shaking. Cells were collected by centrifugation in a Beckman JA10 rotor at 6,000 rpm for 10 min at 4°C and washed in water. After resuspending in an extraction buffer containing 50 mM Tris-HCl, pH 7.5, 600 mM KCl, 10% sucrose, 5 mM ␤-mercaptoethanol, and protease inhibitors (30), cells were homogenized by Zirconium beads in a beadbeater (Biospec Products) for 15 pulses of 30 s each on ice. The clarified extract (ϳ100 ml) was loaded onto two connected HiTrap chelating columns charged with NiSO 4 (Amersham Pharmacia Biotech, 2 ϫ 5 ml) followed by washing the column with 100 ml of Ni buffer A (50 mM Tris-HCl, pH 7.5, 0.5 M NaCl, 10% glycerol, 5 mM ␤-mercaptoethanol, 10 mM imidazole, and protease inhibitors) at 5 ml/min. Bound proteins were eluted with a linear gradient of 25-100 mM imidazole at 5 ml/min for 1 h. Fractions containing the Rad30 protein were identified by Western blots using a mouse monoclonal anti-His antibody. The pooled Rad30 fractions were concentrated by polyethylene glycol 10,000 overnight and desalted through 5 ϫ 5 ml HiTrap desalting columns (Amersham Pharmacia Biotech) in T buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 10% glycerol, 5 mM ␤-mercaptoethanol, and protease inhibitors). The resulting sample (ϳ40 ml) was loaded onto an FPLC Mono S HR5/5 column (Amersham Pharmacia Biotech) pre-equilibrated with the T buffer. After washing the column with 10 ml of the T buffer, bound proteins were eluted with a 30-ml linear gradient of 0 -500 mM KCl in the T buffer. The Rad30 protein was identified by Western blot analyses and DNA polymerase activity assays. It was eluted at ϳ220 mM KCl. The purity of the isolated Rad30 was examined by electrophoresis on a 10% SDS-polyacrylamide gel followed by silver staining of the gel.
Purification of the Yeast DNA Pol-DNA Pol was expressed in yeast AMY32 (9) cells from the Rev3 and Rev7 overexpression plasmids pEGTh6-REV3 and pEGUh6-REV7, respectively. Their expressions in yeast were under the control of the inducible GAL1/GAL10 promoter and produced both the Rev3 and the Rev7 proteins containing a His 6 tag at their N termini. Overexpression of the plasmids was achieved as described above for the yeast Pol. Cell extracts were processed and purified through a HiTrap chelating column charged with NiSO 4 , followed by liquid chromatography on an FPLC Mono S HR5/5 column essentially as described above for the yeast Pol, except the buffers used. Ni-P buffer A (20 mM phosphate buffer, pH 7.4, 0.5 M NaCl, 10% glycerol, 5 mM ␤-mercaptoethanol, 10 mM imidazole, and protease inhibitors) and P buffer (20 mM phosphate buffer, pH 7.4, 1 mM EDTA, 10% glycerol, 5 mM ␤-mercaptoethanol, and protease inhibitors) were used for the Ni column and the Mono S column, respectively. The desalted Mono S sample (ϳ20 ml) was loaded onto an FPLC Mono Q HR5/5 column (Amersham Pharmacia Biotech). Bound proteins were eluted with a 30-ml linear gradient of 0 -500 mM KCl in the P buffer. The Rev3 and Rev7 proteins copurified and were identified by Western blot analyses using a mouse monoclonal antibody against the N-terminal His 6 tag.
DNA Lesion Bypass Assays-DNA polymerase assays were performed essentially as described by Johnson et al. (19). The reaction mixture (10 l) contained 25 mM potassium phosphate, pH 7.0, 5 mM MgCl 2 , 5 mM dithiothreitol, 100 g/l bovine serum albumin, 100 M dNTPs (dATP, dCTP, dTTP, and dGTP individually or together as indicated), 100 nM of the 32 P-labeled DNA substrate as indicated, and ϳ10 ng of the purified Pol. After incubation at 30°C for 10 min, reactions were terminated with 7 l of a stop solution (20 mM EDTA, 95% formamide, 0.05% bromphenol blue, and 0.05% xylene cyanol). The reaction products were resolved on a 20% polyacrylamide gel containing 8 M urea and visualized by autoradiography. Primer extension was quantitated by scanning densitometry of the autoradiogram using the SigmaGel software (Sigma) for analysis.

RESULTS
Purification of the Yeast Pol-To facilitate protein purification, we tagged the yeast Pol with 6 histidine residues at its N terminus. The His-tagged Pol (coded by the RAD30 gene) fully complemented the rad30 mutant for UV sensitivity (data not shown), indicating that the His tag did not affect its function. The most pure fraction of the His-tagged yeast Pol was analyzed by electrophoresis on a 10% SDS-polyacrylamide gel followed by silver staining of the gel (Fig. 1A). Two major bands and two faster migrating bands were evident on the SDSpolyacrylamide gel (Fig. 1A). Western blot analyses using a monoclonal antibody against the His tag confirmed that these bands are Pol (Fig. 1B). At present, it is not known whether they are modified Pol or partially degraded Pol or both. Thus, our Pol preparation is pure without detectable contamination by other proteins. The purified Pol migrated at ϳ66 kDa on a 10% SDS-polyacrylamide gel (Fig. 1A), consistent with its calculated molecular mass of 71 kDa. As expected, the pure Pol possesses a DNA polymerase activity on a primed 30-mer DNA template (Fig. 1C). Furthermore, Pol was able to bypass a cis-syn-TT dimer on a 50-mer DNA template (data not shown). Thus, our purified Pol is biochemically active. DNA polymerase activity of the yeast Pol requires Mg 2ϩ , and the polymerase was active with 1-20 mM MgCl 2 and 0 -80 mM KCl. Higher concentrations of MgCl 2 or KCl and over 1 mM EDTA significantly inhibited the yeast Pol activity.
The Yeast Pol Does Not Possess A Nuclease Activity-In addition to DNA polymerizing activity, some DNA polymerases also possess 5Ј 3 3Ј or 3Ј 3 5Ј nuclease activity or both. To examine whether the yeast Pol possesses a nuclease activity, we incubated the polymerase with a 32 P-labeled 18-mer singlestranded DNA (Fig. 2A, primer P2). As shown in Fig. 2B, this DNA remains intact after incubation with Pol. To examine nuclease activities on a double-stranded DNA, we labeled several primers ( Fig. 2A) at their 5Ј ends with 32 P and annealed them to the 30-mer DNA template. These DNA substrates contained either a TC mismatch ( Fig. 2A, primer P2), a TT mismatch ( Fig. 2A, primer P3), a TG mismatch ( Fig. 2A, primer  P4), or a complementary TA base pair ( Fig. 2A, primer P1) at the primer end. Then we incubated these DNA substrates with the yeast Pol without dNTPs and analyzed by electrophoresis on a 20% denaturing gel. As shown in Fig. 2C, the labeled primers were not degraded regardless of whether they contained a terminal mismatch or not. To determine whether the unlabeled DNA template strand was degraded by Pol, we analyzed the incubation products by electrophoresis on a 15% native polyacrylamide gel. Again, a nuclease activity was not detected (Fig. 2D). These results show that the yeast Pol does not possess a nuclease or a proofreading activity.
Trans-lesion Synthesis of Template 8-Oxoguanine by the Yeast Pol-8-Oxoguanine is a major form of oxidative damage in DNA. To examine whether Pol can bypass this DNA lession, we synthesized a 30-mer DNA template containing a site-specific 8-oxoguanine residue (Fig. 3A). A 32 P-labeled 17mer primer was annealed to the template, right before the 8-oxoguanine residue. The yeast Pol efficiently bypassed the template 8-oxoguanine and extended the primer to the end of the template (Fig. 3B, lanes 5 and 6). DNA synthesis by Pol is similarly efficient using either the undamaged DNA template or the 8-oxoguanine-containing template (Fig. 3B, compare  lanes 1 and 2 with lanes 4 and 5). These results indicate that 8-oxoguanine in DNA does not block the yeast Pol and does not significantly affect the polymerase activity of the enzyme.
8-Oxoguanine is a miscoding DNA lesion. It is capable of FIG. 2. Nuclease activity assays of the purified Pol. A, the DNA template and four primers used for nuclease activity assays. The primers were labeled with 32 P at their 5Ј ends as indicated by an asterisk. B, the labeled primer P2 (100 nM) was incubated with increasing amounts of the yeast Pol under the polymerase assay conditions without dNTPs. Reaction products were resolved by electrophoresis on a 20% denaturing polyacrylamide gel containing 8 M urea. C, the DNA template (100 nM) annealed with different primers as indicated was incubated with the yeast Pol for 30 min at 30°C in the DNA polymerase assay buffer without dNTPs. Reaction products were separated by electrophoresis on a 20% denaturing polyacrylamide gel. D, DNA templates annealed with the primers indicated were identically incubated with the yeast Pol as in C. Reaction products were separated by electrophoresis on a 15% non-denaturing polyacrylamide gel. This native gel system would allow the reaction products to be detected if the yeast Pol had degraded the unlabeled DNA template strand. DNA size markers in nt are indicated between B and C.
base pairing with either a C or an A residue (31). To identify the base incorporated opposite the 8-oxoguanine residue by the yeast Pol, we performed DNA synthesis assays with only one deoxyribonucleoside triphosphate: dATP, dCTP, dGTP, or dTTP individually. As shown in Fig. 3C, under identical reaction conditions, except different deoxyribonucleoside triphosphates, the yeast Pol extended near 100% of the primers using dCTP, 56% of the primers with dATP, and 14% of the primers with dGTP opposite the template 8-oxoguanine. Primer extension was not detected with dTTP (Fig. 3C). These results suggest that the yeast Pol predominantly incorporates a C residue opposite a template 8-oxoguanine but can also incorporate an A residue with a lower efficiency and a G residue with the least efficiency. Hence, we conclude that when encountered with 8-oxoguanine residues in DNA, the yeast Pol could cause mutagenic trans-lesion synthesis.
Error-free Nucleotide Insertion Opposite AAF-adducted Guanine by the Yeast Pol-In contrast to the 8-oxoguanine, an AAF-adducted guanine in the template DNA blocks many DNA polymerases (32). Since the yeast Pol can efficiently bypass a TT dimer that is considered to be a bulky lesion, we asked whether this polymerase is able to bypass the bulky AAFguanine lesion. A 32 P-labeled 17-mer primer was annealed to a 30-mer DNA template right before the site-specific template AAF-guanine (Fig. 4A). The yeast Pol effectively incorporated one nucleotide opposite the template AAF-guanine (one nucleotide extension) but failed to extend the primer further (Fig.  4B, lane 2). To reveal the identity of the nucleotide incorporated opposite the AAF-guanine, we did lesion bypass assays with only one deoxyribonucleoside triphosphate in the reaction mixture. As show in Fig. 4B (lanes 3-6), the yeast Pol exclusively incorporated a C residue opposite the template AAFguanine. In contrast to Pol, the yeast Pol␣ was unable to insert any nucleotide opposite the template AAF-guanine, even when excess amount of the polymerase was used (Fig. 4C, lane  3). In comparison, the purified Pol␣ was able to use the same DNA template without the AAF adduct for DNA synthesis in the presence of all four dNTPs (Fig. 4C, lane 1) or for one nucleotide extension of the primer in the presence of dCTP alone (Fig. 4C, lane 2). These results show that the yeast Pol is able to perform error-free nucleotide insertion opposite a template AAF-guanine but is unable to extend the DNA synthesis further.
Error-prone Nucleotide Insertion Opposite a Template AP Site by the Yeast Pol-For UV-induced cyclobutane pyrimidine dimers in DNA, Pol bypasses the lesions in an error-free mode (19,23). However, an AP site would pose a challenge for any DNA polymerase that can bypass it. Since an AP site can derive from A, T, C, or G, any enzyme that can bypass the AP site would have to be error-prone. Thus, it is of particular interest to determine how the yeast Pol would respond to an AP site in DNA. First, we synthesized a 30-mer DNA template containing a site-specific uracil residue. A 32 P-labeled 17-mer primer was annealed to the uracil-containing template. After treatment with the E. coli uracil-DNA glycosylase, the uracilcontaining templates were completely converted to the AP sitecontaining templates (29). The primer lies right before the template AP site (Fig. 5A). Then we incubated the yeast Pol with the AP site-containing DNA substrate for lesion bypass assays. As shown in Fig. 5B (lane 2), the yeast Pol effectively incorporated one nucleotide opposite the AP site (one nucleotide extension) but was unable to continue DNA synthesis downstream of the template AP site. Under identical conditions, the corresponding undamaged template supported efficient primer extension by the yeast Pol up to the end of the template (Fig. 5B, lane 7). Since this undamaged control tem-plate contained a uracil residue, we further identified that A was exclusively incorporated opposite the template U (data not shown). Thus, the yeast Pol efficiently recognized the template U as a coding base.
To identify the nucleotide incorporated opposite the template AP site, we performed similar AP site lesion bypass assays with only one deoxyribonucleoside triphosphate. Both G (Fig. 5B,  lane 6) and A (Fig. 5B, lane 3) were significantly incorporated opposite the template AP site, supporting 67 and 38% 1-nt primer extensions, respectively. To a much less extent, C and T were also incorporated opposite the template AP site by the yeast Pol, supporting 6 and 5% 1-nt primer extensions, respectively (Fig. 5B, lanes 4 and 5). These results show that the yeast Pol predominantly incorporates a G and less frequently FIG. 3. Bypass of a template 8-oxoguanine (8-oxoG) by the yeast Pol. A, the DNA substrate used for 8-oxoG lesion bypass assays. The 17-nt primer was labeled with 32 P at its 5Ј end as indicated by an asterisk. B, standard polymerase assays were performed with increasing amounts of the yeast Pol as indicated using the DNA template containing a site-specific 8-oxoG (lanes 4 -6) or identical template without the lesion (lanes 1-3). C, standard polymerase assays were performed with 10 ng of the yeast Pol using the 8-oxoG-containing substrate as shown in A in the presence of dATP (lane 2), dCTP (lane 3), dGTP (lane 4), or dTTP (lane 5) individually. Reaction products were separated by electrophoresis on a 12% sequencing gel. Lane 1, control experiment without any dNTPs. DNA size markers in nt are indicated on the right.
an A opposite the template AP site but by itself cannot continue DNA synthesis further after the template AP site.
AP Site Bypass by Combined Activities of the DNA Polymerases and -Rev1 is another protein that can efficiently insert a nucleotide opposite a template AP site (13). However, unlike Pol, the Rev1 activity specifically inserts a C residue opposite the AP site (13). Following the Rev1 action, DNA Pol (the Rev3-Rev7 protein complex) can then continue DNA synthesis, achieving error-prone bypass of the AP site (13). Thus, we asked if Pol could also continue DNA synthesis following nucleotide insertion opposite the AP site by Pol. By using the AP site-containing template (Fig. 5A), we performed lesion bypass assays. As shown in Fig. 6 (lane 2), the yeast Pol, contained in a partially purified fraction, was unable to extend the 17-mer primer annealed right before the template AP site, consistent with the observation of Nelson et al. (13). The purified yeast Pol incorporated one nucleotide opposite the template AP site, extending the 17-mer primer to an 18-mer (Fig.  6, lane 1). However, upon subsequent addition of the yeast Pol, ϳ13% of the 18-mer products were further extended to the end of the template (Fig. 6, lane 3). These results suggest that the combined actions of the yeast Pol and Pol can result in error-prone bypass of the template AP site.
AP Site Bypass by the Yeast DNA Pol␣-As a control for the yeast Pol trans-lesion synthesis experiments, we performed AP site bypass assays with the purified yeast Pol␣ using the AP site-containing DNA substrate (Fig. 5A). Surprisingly, we found that the yeast Pol␣ alone was able to bypass efficiently the template AP site (Fig. 7A, lanes 3-5), in contrast to the yeast Pol activity on this substrate (Fig. 7A, lane 2). To reveal  1 and 2). C, polymerase assays with the purified yeast Pol␣ (4.5 units, 90 ng) were performed using the AAF-G substrate (lane 3) or its identical substrate without the AAF adduct (lanes 1 and 2). The reaction buffer contained the four dNTPs (N 4 ) or dCTP alone (C) as indicated. DNA size markers in nt are indicated on the right. the base identity opposite the template AP site, we performed the bypass assay in the presence of only one deoxyribonucleoside triphosphate. As shown in Fig. 7B (compare lane 1 with  lanes 2-4), an A residue was predominantly incorporated opposite the template AP site. Hence, we conclude that the yeast Pol␣ itself can cause a specific mutagenic trans-lesion synthesis opposite a template AP site. DISCUSSION The recently described DNA Pol is encoded by the RAD30 gene in yeast and the XPV gene in humans (17,18,22,24). The role of DNA Pol in response to UV radiation has largely been defined. It functions as a DNA polymerase bypassing cyclobutane pyrimidine dimers in an error-free manner (19,23). Thus, inactivation of the Pol gene results in enhanced UV sensitivity and UV-induced mutagenesis (17,18,33). Such defect in humans can lead to the hereditary disease xeroderma pigmentosum characterized by photosensitivity and a predisposition to skin cancer.
The inability of Pol to bypass (6-4) photoproducts in DNA (23) provided the first clue that properties of trans-lesion synthesis by this polymerase may be lesion-specific. In this study, we have examined the response of the yeast Pol to the following three additional DNA lesions in vitro: 8-oxoguanine, AAFmodified guanine, and AP sites. It is known that 8-oxoguanine is a miscoding lesion that can direct either a C or an A base into the newly synthesized DNA strand (31). Consistent with this miscoding property of the 8-oxoguanine lesion, the yeast Pol also incorporated a C or an A base opposite the lesion. However, unlike other DNA polymerases tested (31), the yeast Pol additionally can incorporate a G base opposite the template 8-oxoguanine, although at a lower frequency than C or A incorporation. Comparing DNA synthesis on undamaged and 8-oxoguanine-containing templates indicates that this lesion does not significantly block the movement of the yeast Pol on the DNA template. Hence, when Pol encounters an 8-oxoguanine in the DNA template in vivo, trans-lesion synthesis by this polymerase will most likely occur resulting in error-prone lesion bypass.
In contrast to 8-oxoguanine and a cis-syn-TT dimer, the AAF adduct on a guanine effectively blocks the movement of the yeast Pol on the DNA template. However, the Pol is able to incorporate a C residue opposite the AAF-guanine lesion before aborting DNA synthesis. This activity appears to be specific to Pol, since the yeast Pol␣ is completely blocked before the AAF damage. It remains to be determined if other DNA polymerases could utilize such a DNA primer to continue the DNA synthesis after the error-free nucleotide insertion by Pol. Our preliminary results indicate that the purified yeast Pol␣ or the partially purified yeast Pol alone is unable to do so. Nevertheless, it is conceivable that the combined actions of Pol and another DNA polymerase could lead to an effective bypass of AAFguanine, achieving an error-free response to this lesion in cells.
AP sites are significant spontaneous and induced DNA lesions. They are considered as noninstructional lesions due to  loss of the coding base. In yeast, the Rev1-Rev3 mutagenesis pathway is the major mechanism responsible for the errorprone bypass of AP sites (13,14). An important role of the Rev1 protein is to insert a dCMP opposite the template AP site such that the Pol activity encoded by the REV3 and REV7 genes could continue the primer extension further downstream of the AP site (13). This mechanism of AP site bypass is most likely operational in humans, since both the Rev1 and the Rev3 homologues have been identified in humans (29,34,35), and the human REV1 is also capable of inserting a dCMP opposite a template AP site (29). However, it is not clear whether other mechanisms may additionally contribute to the AP site bypass in eukaryotes. Our in vitro studies suggest such a possibility. The yeast Pol is able to insert efficiently a nucleotide opposite the template AP site. Once incorporated into the newly synthesized DNA, the nucleotide opposite the template AP site cannot be removed by Pol due to lack of a 3Ј 3 5Ј proofreading nuclease activity. Thus, this Pol DNA synthesis product may be subsequently extended by another DNA polymerase to achieve a complete AP site bypass. Indeed, we observed that a partially purified Pol preparation is able to do so. If Pol contributes to the bypass of some AP sites in vivo, such bypass would be error-prone due to the predominant incorporation of a G or less frequently an A base opposite the AP site by this polymerase. Since the Pol is DNA damage-inducible (17,18), the putative contribution of this polymerase to AP site bypass, if any, may only be important when cellular DNA is damaged by exogenous agents.
Unlike Pol and Pol, the yeast DNA Pol␣ is able to bypass efficiently a template AP site in our studies. Lack of AP site bypass by the yeast Pol␣ was reported by Nelson et al. (13) at a different DNA sequence context, although the purified Pol␣ preparations used in both studies were from the same laboratory. The apparent discrepancy is not clear at present. We further observed that an A residue is predominantly incorporated opposite the template AP site by the yeast Pol␣, consistent with the "A rule" (36). Thus, when encountering a template AP site during DNA replication, Pol␣ is very likely to bypass the lesion and generate a mutation opposite the AP site. On the other hand, the yeast Pol␣ is unable to bypass a TT dimer (12) or insert a nucleotide opposite a template AAF-guanine (Fig.  4C). Thus, the response of a DNA polymerase to DNA damage is lesion-specific and depends on the specific interaction between the lesion and the polymerase. The unique properties of Pol toward DNA lesions may be a result of the structural features of this polymerase, which can lead to an error-free or an error-prone consequence. Hence, Pol may not be generalized as an error-free DNA polymerase for trans-lesion DNA synthesis based only on its response to UV radiation.