Mismatch Extension Ability of Yeast and Human DNA Polymerase h *

DNA polymerase h (Polh) functions in error-free replication of UV-damaged DNA, and in vitro it efficiently bypasses a cis-syn T-T dimer by incorporating two adenines opposite the lesion. Steady state kinetic studies have shown that both yeast and human Polh are lowfidelity enzymes, and they misincorporate nucleotides with a frequency of 10–10 on both undamaged and T-T dimer-containing DNA templates. To better understand the role of Polh in error-free translesion DNA synthesis, here we examine the ability of Polh to extend from base mismatches. We find that both yeast and human Polh extend from mismatched base pairs with a frequency of ;10 relative to matched base pairs. In the absence of efficient extension of mismatched primer termini, the ensuing dissociation of Polh from DNA may favor the excision of mismatched nucleotides by a proofreading exonuclease. Thus, we expect DNA synthesis by Polh to be more accurate than that predicted from the fidelity of nucleotide incorporation alone.


M. Todd Washington, Robert E. Johnson, Satya Prakash, and Louise Prakash ‡
From the Sealy Center for Molecular Science, University of Texas Medical Branch, Galveston, Texas 77555-1061 DNA polymerase (Pol) functions in error-free replication of UV-damaged DNA, and in vitro it efficiently bypasses a cis-syn T-T dimer by incorporating two adenines opposite the lesion. Steady state kinetic studies have shown that both yeast and human Pol are lowfidelity enzymes, and they misincorporate nucleotides with a frequency of 10 ؊2 -10 ؊3 on both undamaged and T-T dimer-containing DNA templates. To better understand the role of Pol in error-free translesion DNA synthesis, here we examine the ability of Pol to extend from base mismatches. We find that both yeast and human Pol extend from mismatched base pairs with a frequency of ϳ10 ؊3 relative to matched base pairs. In the absence of efficient extension of mismatched primer termini, the ensuing dissociation of Pol from DNA may favor the excision of mismatched nucleotides by a proofreading exonuclease. Thus, we expect DNA synthesis by Pol to be more accurate than that predicted from the fidelity of nucleotide incorporation alone.
DNA polymerase (Pol) 1 functions in error-free replication of UV-damaged DNA, and mutations in the gene encoding this enzyme result in increased UV mutability in both yeast and humans (1). In humans, inactivation of Pol causes the variant form of the cancer prone syndrome xeroderma pigmentosum (2,3). Pol is unique among eukaryotic DNA polymerases in its ability to efficiently replicate DNA containing a cis-syn T-T dimer, and it does so by incorporating two adenines across from the two thymines of the dimer (3)(4)(5)(6).
The high fidelity of replicative DNA polymerases arises, in part, because their active sites are intolerant of the distorted geometry resulting from mispairs between the template residue and the incoming nucleotide (7). Steady state kinetic studies of yeast and human Pol have indicated that it is a lowfidelity enzyme, misincorporating nucleotides with a frequency of 10 Ϫ2 -10 Ϫ3 on undamaged DNA (5,8). Remarkably, however, Pol synthesizes DNA opposite a T-T dimer with the same efficiency and accuracy as opposite undamaged T residues (5,6). The low fidelity of Pol may reflect an unusual tolerance of its active site for deviant geometry arising from distorting template lesions such as a T-T dimer.
The accuracy of synthesis by DNA polymerases depends on the frequency of incorporation of incorrect nucleotides into DNA and on the frequency of extension of the mismatched primer termini. Extension of mismatched primers is a critical step in mutation fixation, because in the absence of efficient extension, the mismatched nucleotide can be excised by a proofreading exonuclease, or if the mismatch is not excised, cell death may ensue as a result of incomplete DNA synthesis. Thus, for a mutation to be expressed, extension from the misincorporated nucleotide must occur. To better understand how Pol, with a low nucleotide insertion fidelity, can function in an error-free pathway of translesion DNA synthesis in vivo, here we examine the ability of Pol to extend from base mispairs. We find that yeast and human Pol extend from mismatched primer-templates with a frequency of ϳ10 Ϫ3 relative to matched primer-templates. This implies that Pol, which has a low processivity, will have a greater likelihood of dissociating from the DNA template after the incorporation of an incorrect nucleotide than a correct one. That would lower the error rate of DNA synthesis in vivo, because the mismatched primer terminus could then be subjected to the proofreading 3Ј35Ј exonuclease activity of other protein factors.

MATERIALS AND METHODS
DNA Substrates-DNA substrates containing all possible correct base pairs or mispairs at the 3Ј primer terminus were generated using four different oligodeoxynucleotide primers and four oligodeoxynucleotide templates. The four 45-nucleotide primers have the following sequence: 5Ј-GTTTT CCCAG TCACG ACGAT GCTCC GGTAC TCCAG TGTAG GCATN, where N is G, A, T, or C. The four 52-nucleotide templates have the following sequence: 5Ј-TTCGT ATNAT GCCTA CACTG GAGTA CCGGA GCATC GTCGT GACTG GGAAA AC, where N is G, A, T, or C. The various combinations of primers and templates were annealed by mixing 1 M 32 P-end labeled primer with 1.5 M template in 50 mM Tris-HCl, pH 7.5, and 100 mM NaCl and heating to 90°C for 2 min before slowly cooling to room temperature over several hours.
Steady state Kinetics Assays-Yeast and human Pol were expressed in and purified from yeast strain BJ5464 as described (4,5). The steady state kinetics of single nucleotide incorporation were measured by incubating 1 nM yeast or human Pol with 20 nM DNA substrate in 25 mM sodium phosphate, pH 7.0, buffer containing 5 mM magnesium chloride, 5 mM dithiothreitol, 10 g/ml bovine serum albumin, and 10% glycerol for 10 min at 25°C. For nucleotide incorporation following a correctly base paired or mispaired primer terminus, the concentration of dATP was varied from 0 to 5 M or from 0 to 2000 M, respectively. Reactions were quenched after 10 min by adding 10 volume of loading buffer (95% formamide, 0.03% bromphenol blue, and 0.3% cyanol blue). Samples were then run on 10% polyacrylamide sequencing gels to separate the unextended and extended DNA primers. Gel band intensities were quantified using a PhosphorImager and ImageQuant software (Molecular Dynamics). The observed rate of nucleotide incorporation was calculated by dividing the amount of reaction product formed by the 10-min incubation time. The observed rate of nucleotide incorporation was then plotted as a function of nucleotide concentration, and the apparent K m and V max parameters were obtained from the best fit to the Michaelis-Menten equation using nonlinear regression (Sigma Plot 4.0). The intrinsic efficiency of mismatch extension, f ext o , which is a constant that represents the efficiency of extending mismatched termini in competition with matched termini at equal DNA concentra-* This work was supported by National Institutes of Health Grant GM19261. 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 U.S.C. Section 1734 solely to indicate this fact.

RESULTS AND DISCUSSION
We examined the steady state kinetics of nucleotide incorporation by Pol following the correctly base paired and mis-paired termini in primer-template substrates (7,9,10). For example, the rate of incorporation of an A residue by yeast Pol opposite a template T residue following a G⅐C base pair or an A⅐C, T⅐C, or C⅐C mispair was measured over a broad range of dATP concentrations (Fig. 1A). Gel band intensities were eval-  Table I. uated, and the rate of nucleotide incorporation was plotted as a function of nucleotide concentration. As shown in Fig. 1B, these plots yield curves typical of Michaelis-Menten kinetics. The apparent values of V max and K m for extension of each primer terminus were obtained from the best fit to the Michaelis-Menten equation using nonlinear regression. The frequency of mispair extension (f ext o ), which is the ratio of the apparent V max /K m of extension from the mispair to the apparent V max /K m of extension from a correct base pair, was then calculated (7,9,10).
As shown in Table I, for the incorporation of an A residue following a G⅐C base pair, the apparent K m for yeast Pol is 0.20 M, and the V max is 0.28 nM/min, whereas for the incorporation of an A following an A⅐C mispair, the apparent K m is 21 M, and the V max is 0.25 nM/min, respectively. Thus, for the A⅐C mispair  (Table I).
For most DNA polymerases, the frequency of extension from a given mispair (f ext o ) is approximately the same as the frequency of incorporating that same mispair (f inc ; Refs. 7, 10). We also examined human Pol for its ability to extend from base mispairs. As was observed for yeast Pol, the f ext o values for human Pol range from 10 Ϫ2 to 10 Ϫ3 , with an average of 2.5 ϫ 10 Ϫ3 (Table II), and a comparison of f ext o values with the previously published f inc values (5) indicates that human Pol is also somewhat less efficient at extending from mispairs than at inserting mispaired bases (Fig. 2B).
Pol replicates through a cis-syn T-T dimer with the same efficiency and fidelity as through undamaged T nucleotides (5,6). Furthermore, our steady state kinetic analyses of base mispair extension across from the T-T dimer indicate that these mispairs are also inefficiently extended and with the same frequency as mispairs in undamaged DNA. 2 When compared with other DNA polymerases, the mispair extension ability of Pol is greater than that of the high-fidelity  DNA polymerase ␣, the f ext o of which ranges from 10 Ϫ3 to 10 Ϫ6 (10). However, its mispair extension ability is considerably lower than that of the most promiscuous extender of mispairs known, yeast Pol, which extends from mispaired template primer termini with a frequency of 10 Ϫ1 to 10 Ϫ2 (11). Pol plays an essential role in mutagenic bypass of DNA lesions, and it specifically functions in damage bypass by extending from nucleotides placed opposite DNA lesions by another DNA polymerase (11).
Pol has low processivity (5,8), and thus it has a modest probability (0.2-0.3) of dissociating from the DNA template after each nucleotide incorporation. Our observation that both yeast and human Pol extend from mismatched primer termini with a frequency of ϳ10 Ϫ3 relative to a matched primer terminus implies that Pol has a substantially higher probability of dissociating from the primer terminus after the incorporation of an incorrect nucleotide than a correct nucleotide. Dissociation of Pol would prevent mutation fixation, because any mispairs left in DNA would then be subject to removal by the proofreading exonucleolytic activity of Pol␦ or other proofreading exonucleases. Thus, DNA synthesis by Pol would be more accurate than is indicated from the fidelity of nucleotide incorporation (f inc ) values. Because Pol extends from mismatched bases opposite a T-T dimer with the same efficiency as from undamaged DNA, we predict that the error frequency during T-T dimer bypass will also be lower than that suggested from the f inc values for the incorporation of wrong nucleotides opposite the two T nucleotides of the T-T dimer (5,6).
We expect the activity of Pol to be restricted to DNA syn-thesis during damage bypass. The Rad6-Rad18 complex, which is essential for damage bypass and which contains ubiquitin conjugating and DNA binding activities (12), may be crucial for modulating the specific targeting of Pol to sites where replication has stalled at a DNA lesion and for ensuring the dissociation of Pol from DNA once the lesion has been bypassed. Furthermore, association with other protein factors may increase the fidelity of nucleotide incorporation by Pol. Thus, in vivo, damage bypass by Pol would be much more accurate than 10 Ϫ2 -10 Ϫ3 , the frequency of nucleotide misincorporation.