Report

Efficient Bypass of a Thymine-Thymine Dimer by Yeast DNA Polymerase, Polη

See allHide authors and affiliations

Science  12 Feb 1999:
Vol. 283, Issue 5404, pp. 1001-1004
DOI: 10.1126/science.283.5404.1001

Abstract

The RAD30 gene of the yeast Saccharomyces cerevisiae is required for the error-free postreplicational repair of DNA that has been damaged by ultraviolet irradiation. Here,RAD30 is shown to encode a DNA polymerase that can replicate efficiently past a thymine-thymine cis-syn cyclobutane dimer, a lesion that normally blocks DNA polymerases. When incubated in vitro with all four nucleotides, Rad30 incorporates two adenines opposite the thymine-thymine dimer. Rad30 is the seventh eukaryotic DNA polymerase to be described and hence is named DNA polymerase η.

Cells possess a variety of mechanisms to repair damaged DNA. If left unrepaired, DNA lesions in the template strand can block the replication machinery (1). In S. cerevisiae, genes in the RAD6 epistasis group function in the replication of DNA that has been damaged by ultraviolet (UV) light and other agents (2). Of the genes in this group, REV1, REV3, and REV7 are required for UV mutagenesis, whereas RAD5 andRAD30 function in alternate pathways of error-free bypass of UV-induced DNA damage (3, 4). Rev3 and Rev7 associate to form a DNA polymerase, Polζ, that can weakly bypasscis-syn thymine-thymine (T-T) dimers (5). Rev1 has a deoxycytidyl transferase activity that transfers a deoxycytidine 5′-monophosphate residue to the 3′ end of a DNA primer in a template-dependent reaction (6). The addition of Rev1 has no effect on the Polζ bypass of cis-syn T-T dimers (6). Rad5 is a DNA-dependent adenosine triphosphatase (7) and Rad30 is homologous withEscherichia coli DinB and UmuC and with S. cerevisiae Rev1 (4). Here, we elucidate the function of yeast Rad30.

The RAD30 gene was fused in-frame downstream of the glutathione S-transferase (GST) gene, and the resulting fusion protein (∼95 kD) was purified to near homogeneity (8) from a protease-deficient yeast strain harboring the GST-Rad30 expression plasmid pBJ590. Plasmid pBJ590 fully complements the UV sensitivity of the rad30Δ mutation, indicating that the Rad30 fusion protein functions normally in vivo. We first examined whether Rad30 has deoxynucleotidyl transferase activity (9). DNA substrates containing a different template nucleotide at the primer-template junction were incubated with Rad30 in the presence of just one deoxynucleoside triphosphate (dNTP). Rad30 predominantly incorporated the correct nucleotide across from each of the four template nucleotides, but incorrect nucleotides were also inserted across from some template residues (Fig. 1A). For instance, a C residue and a T residue were weakly incorporated opposite template C (Fig. 1A, lane 6) and opposite template G (Fig. 1A, lane 11), respectively. Rad30 exhibited the lowest specificity when the template nucleotide was T, as in substrate S-3. Whereas the A residue was incorporated with the highest efficiency (Fig. 1A, lane 16), G, T, and C residues were also incorporated with varying efficiencies (Fig. 1A, lanes 15, 17, and 18, respectively). The pattern of misincorporations seemed to depend on the sequence context of the template. For example, in substrate S-4, which has A as the template nucleotide, a T residue was incorporated opposite the A template, and a T residue was inserted across from the subsequent C residue in the template (Fig. 1A, lane 23). However, a T residue was not incorporated across from the C template in substrate S-1 (Fig. 1A, lane 5). Oligo deoxyriboadenylate or oligo deoxyribothymidylate was not extended by Rad30 in the absence of template DNA, and no incorporation was observed when ribonucleoside triphosphates replaced dNTPs.

Figure 1

Deoxynucleotidyl transferase and DNA polymerase activities of Rad30. (A) Specificity of nucleotide incorporation by Rad30. Nucleotide sequence adjacent to the primer:template junction is shown for each of the DNA substrates (S-1 through S-4). DNA substrates (10 nM) were incubated with GST-Rad30 (2.5 nM) for 5 min at 30°C in the absence (–) of dNTP or in the presence of a single dNTP (G, A, T, or C). As a control, substrates were incubated with all four dNTPs (N) and without the enzyme. The amount of product formed (in femtomoles) was determined by phosphorimaging; dashes indicate that no detectable product was formed. (B) DNA polymerase activity of GST-Rad30. GST-Rad30 (2.5 nM) was incubated for 5 min at 30°C with DNA substrates (10 nM) S-1, S-2, S-3, and S-4 (lanes 1 through 4, respectively) in the presence of all four dNTPs. Lanes 5 through 12 show the effect of inhibitors on Rad30 DNA polymerase activity. GST-Rad30 (2.5 nM) was incubated with DNA substrate S-4 (10 nM) for 5 min at 30°C in the presence of all four dNTPs, and inhibitors were added as indicated. In lane 5, aphidicolin (50 μg/ml) was added in the regular assay, and deoxycytidine 5′-triphosphate concentration was lowered to 10 μM (10). Lanes 6 through 9 show 0.5-, 2-, 5-, and 10-min preincubation, respectively, with 10 mM NEM (10). Lanes 10 through 12 show 1:1, 5:1, and 10:1 ddNTP:dNTP, respectively. The reactions contained all four ddNTPs. The positions of the 40-, 41-, 42-, and 43-nt primers and the 75-nt full-length product are indicated.

Because Rad30 added each of the four nucleotides onto the primer in a template-specific manner in the presence of a single dNTP, we next examined whether the enzyme synthesized DNA in the presence of all four nucleotides (9). Extensive synthesis was observed for all four DNA substrates examined, and polymerization proceeded to the end of the template (Fig. 1B, lanes 1 through 4). Under the conditions used, the DNA polymerase activity was proportional to enzyme concentration and was linear with time.

We examined the sensitivity of Rad30 activity to known inhibitors of DNA polymerases. Rad30 had 90% activity at 150 nM aphidicolin [50 μg/ml (Fig. 1B, lane 5), which is a concentration that completely inhibits yeast polymerases α, δ, and ɛ (10)]. Yeast DNA polymerases α, δ, ɛ, and γ are 50% inhibited when preincubated with 1 mM N-ethylmaleimide (NEM) for 4.7, 3.6, 2.8, and 0.5 min, respectively (10). Rad30 retained 80% activity when preincubated for 10 min in 10 mM NEM (Fig. 1B, lane 9). DNA polymerase β is sensitive to dideoxyribonucleoside triphosphates (ddNTPs), being completely inhibited at a ratio of 5:1 of ddNTP:dNTP (11). Rad30 exhibits a modest sensitivity to ddNTPs, showing 30% activity at a ddNTP:dNTP ratio of 10:1 (Fig. 1B, lane 12). The insensitivity of Rad30 to inhibitors of yeast DNA polymerases α, δ, ɛ, and γ indicated that Rad30 was distinct from these other enzymes. To exclude the possibility that Polζ and Polβ contaminated our Rad30 preparation, we showed that GST-Rad30 purified from a rev3Δ yeast strain, deleted for the Polζ catalytic subunit, and purified from apolβΔ yeast strain had the same activity as GST-Rad30 from the wild-type strain.

To confirm that the DNA polymerase activity was intrinsic to Rad30, we showed that the activity copurified with the GST-Rad30 protein (Fig. 2A) and that a mutant Rad30, which was missing the carboxyl terminal 404 amino acids (12), had no activity with any of the four DNA substrates (Fig. 2B). The corresponding rad30 mutant gene did not complement the UV sensitivity of the rad30Δ strain. The SIDEVF domain (13) present in Rad30 at residues 153 to 158 is highly conserved among Rad30, DinB, UmuC, and Rev1. A rad30mutation in which Asp155 and Glu156 have been changed to Ala did not complement the UV sensitivity of therad30Δ strain, and the mutant protein lacked DNA polymerase activity (14). From these observations, we conclude that Rad30 is a DNA polymerase.

Figure 2

DNA polymerase activity is intrinsic to Rad30. (A) Coelution of DNA polymerase activity with GST-Rad30. Fractions from the final Mini-Q chromatography step were assayed for DNA polymerase activity. Two microliters of each fraction were separated on a 9% denaturing polyacrylamide gel and stained with Coomassie blue (top). Fractions were diluted 50-fold and assayed for DNA polymerase activity with substrate S-3 (bottom). (B) Carboxyl-terminal truncation of Rad30 abolishes DNA polymerase activity. Approximately 200 ng of carboxyl-terminally truncated GST-Rad30 (lane 2) was subjected to electrophoresis on a 9% denaturing polyacrylamide gel. Lane 1 contains size markers (in kilodaltons). Lanes 3 through 6 show the activity of 10 nM of carboxyl-terminus deleted GST-Rad30 with 10 nM of substrates S-1, S-2, S-3, and S-4, respectively. Reactions were carried out in the presence of all four dNTPs for 5 min at 30°C.

To test the possibility that Rad30 performs error-free translesion synthesis, we examined whether the enzyme incorporated a nucleotide opposite a cis-syn T-T dimer (9) by carrying out assays in the presence of only one of the four dNTPs. Similar to previous results with template T (Fig. 1A, lanes 13 through 18), Rad30 incorporated nucleotides across from the nondamaged TT template (Fig. 3A, lanes 3 through 5). Again, an A residue was preferentially incorporated, but G and T residues were also incorporated. Two A residues were also incorporated opposite the dimer, and this was followed by the addition of two more A residues (Fig. 3A, lane 10). As on the nondamaged template, G and T residues were also incorporated across from the dimer (Fig. 3A, lanes 9 and 11).

Figure 3

Deoxynucleotidyl transferase and translesion DNA synthesis activities of Rad30 on a template containing acis-syn T-T dimer. (A) Deoxynucleotidyl transferase activity of GST-Rad30 on a template with or without acis-syn T-T dimer. Sequence adjacent to the primer:template junction is shown. The DNA substrate (10 nM) without (lanes 1 through 6) or with (lanes 7 through 12) a cis-syn T-T dimer was incubated for 5 min at 30°C with each of the four dNTPs and GST-Rad30 (2.5 nM). Symbols are the same as in Fig. 1A. (B) DNA synthesis past a cis-syn T-T dimer by Rad30. Sequences adjacent to the primer:template junctions are shown for “Primer (P-30),” “Template,” and “Primer (P-44).” Position of the nondamaged TT or the cis-syn T-T dimer on the template is indicated by asterisks. Yeast DNA polymerase δ (10 nM) or GST-Rad 30 (2.5 nM) was incubated for 5 min at 30°C with the DNA substrate (10 nM) in the presence of all four dNTPs. Lanes 1 and 2 show running start reactions with Polδ (lane 1, nondamaged DNA template; lane 2, DNA template containing a cis-syn T-T dimer). Lanes 3 through 6 show reactions with GST-Rad30 (lane 3, running start reaction on a nondamaged DNA template; lane 4, running start reaction on a DNA template containing a cis-syn T-T dimer). Lanes 5 and 6 show standing start reactions on a nondamaged DNA template and on a DNA template containing a cis-syn T-T dimer, respectively. The 30-nt oligomer primer (P-30) was used for the running start reaction, and the 44-nt oligomer primer (P-44) was used for the standing start reaction. A portion of the template sequence is shown on the right. TT, nondamaged template; TT, template containing a cis-syndimer. ∧

We next examined the ability of Rad30 to replicate the damaged DNA template in the presence of all four dNTPs. For this, we performed both “standing start” and “running start” experiments. In the standing start substrate, the primer is annealed to the template DNA so that the 3′ hydroxyl of the primer is located just before the T-T dimer (9). In the running start substrate, the 3′ hydroxyl of the primer is located 15 nucleotides upstream of the T-T dimer (9). In contrast to yeast DNA polymerase δ, which stopped just before the cis-syn T-T dimer and did not carry out any translesion synthesis (Fig. 3B, lane 2), Rad30 efficiently inserted nucleotides across from the T-T dimer and carried out efficient synthesis past the dimer (Fig. 3B, lanes 3 through 6). In the running start assay, 82% of the primers were extended by Rad30 past the TT residues in the nondamaged template (Fig. 3B, lane 3), and remarkably, 71% were extended past the dimer on the damaged template (Fig. 3B, lane 4). In the standing start assay, 87% of the primers were extended past the TT residues in the nondamaged template (Fig. 3B, lane 5), and 85% were extended past the T-T dimer on the damaged template (Fig. 3B, lane 6). Sequence analysis of the nascent DNA (15) revealed that two A residues were inserted opposite the undamaged TT site and also opposite the T-T dimer. There were no misincorporations elsewhere in the DNhhA.

Rad30 is homologous to E. coli UmuC and S. cerevisiae Rev1 (4). The umuC andrev1 mutants are largely nonmutable by UV irradiation and by a variety of chemical agents. Recent reconstitution studies withE. coli proteins (16) suggest that UmuD′ and UmuC can promote misincorporation opposite an abasic site in the DNA template and can promote bypass of this lesion by DNA polymerase III. In yeast, Rev1 inserts a C residue opposite an abasic site, but does not insert a C residue or any other nucleotide opposite the T-T dimer, and in vitro, it has no effect on T-T dimer bypass by Polζ (6). The carboxyl terminus of Rad30 (residues 290 through 632) bears no homology to UmuC and Rev1 and may contribute to the unique DNA polymerase activity of Rad30. Rad30 differs from the other known prokaryotic and eukaryotic DNA polymerases in both primary structure and biochemical properties; thus, it represents a new class of DNA polymerases.

Xeroderma pigmentosum variant (XP-V) patients exhibit an increased incidence of sunlight-induced skin cancers. Unlike classical XP cells, XP-V cells have normal nucleotide excision repair but are defective in the replication of UV-damaged DNA (1). XP-V cell-free extracts are deficient in bypass replication of a cis-synT-T dimer (17). XP-V cells are hypermutable with UV light, and they exhibit an unusual mutational spectrum (18). From these studies, it has been inferred that XP-V cells are less likely than normal cells to incorporate deoxyadenosine 5′-monophosphate opposite photoproducts involving thymine (18). Our observations raise the possibility that XP-V cells harbor a defect in the human counterpart of yeast Rad30.

  • * To whom correspondence should be addressed. E-mail: lprakash{at}scms.utmb.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article