The Gly-952 residue of Saccharomyces cerevisiae DNA polymerase alpha is important in discriminating correct deoxyribonucleotides from incorrect ones.

Gly-952 is a conserved residue in Saccharomyces cerevisiae DNA polymerase alpha (pol alpha) that is strictly required for catalytic activity and for genetic complementation of a pol alpha-deficient yeast strain. This study analyzes the role of Gly-952 by characterizing the biochemical properties of Gly-952 mutants. Analysis of the nucleotide incorporation specificity of pol alpha G952A showed that this mutant incorporates nucleotides with extraordinarily low fidelity. In a steady-state kinetic assay to measure nucleotide misincorporation, pol alpha G952A incorporated incorrect nucleotides more efficiently than correct nucleotides opposite template C, G, and T. The fidelity of the G952A mutant polymerase was highest at template A, where the ratio of incorporation of dCMP to dTMP was as high as 0.37. Correct nucleotide insertion was 500- to 3500-fold lower for G952A than for wild type pol alpha, with up to 22-fold increase in pyrimidine misincorporation. The Km for G952A pol alpha bound to mismatched termini T:T, T:C, C:A, and A:C was 71- to 460-fold lower than to a matched terminus. Furthermore, pol alpha G952A preferentially incorporated pyrimidine instead of dAMP opposite an abasic site, cis-syn cyclobutane di-thymine, or (6-4) di-thymine photoproduct. These data demonstrate that Gly-952 is a critical residue for catalytic efficiency and error prevention in S. cerevisiae pol alpha.

DNA polymerases belong to at least five classes, whose members share a high degree of amino acid sequence similarity (1)(2)(3). Polymerase classes include class A (e.g. Escherichia coli pol 1 I, Taq pol I, pol ␥), class B (e.g. pol ␣, RB69 pol), class C (e.g. pol III), class X (e.g. pol ␤), and class Y (e.g. pol ). Despite the fact that the consensus sequence of each polymerase class is significantly different from the consensus sequences of other polymerase classes, the tertiary structures of all polymerases are highly similar. The crystal structures of many polymerases have a common architecture, which has been described as a right hand composed of three domains corresponding to the palm, fingers, and thumb (3). These subdomains form a DNA binding track and the catalytic pocket.
When nucleotides are incorporated into DNA during catalysis, the fingers subdomain rotates from an open to a closed position that faces the incoming dNTP (4 -6). This movement may be the rate-limiting step of the polymerase reaction and may also contribute to the fidelity of polymerization (6 -9). Class A DNA polymerases have a conserved sequence called motif B with the amino acid sequence, R---K---F---YG, which lies in the fingers subdomain and faces toward the catalytic site. In the open complex, Tyr is positioned exactly where template base is expected for maintaining the stacking interactions with the incoming dNTP, whereas in the closed complex, the substrate base is held in position by the Phe residue. Basic amino acids, Arg and Lys, interact with the triphosphate moiety of the incoming nucleotide (4 -6, 10 -12).
Class A DNA polymerases have several catalytically essential residues including a Lys residue in motif B of Taq pol I (Lys-663) and Klenow fragment of E. coli pol I (Lys-758) (10 -12). Phe-667 in Taq pol I (and the equivalent residue in Klenow and T7 pol) is also critical for polymerase activity and for discrimination of 2Ј,3Ј-dideoxynucleotides (11,13,14). Tyr-766 in Klenow fragment (Tyr-671 in Taq pol I) plays a role in fidelity of DNA synthesis (15,16). Recently, Ponamarev et al. (17) reported that a human pol ␥ mutant in which a Tyr equivalent to Tyr-671 is changed to Cys has a higher K m and lower fidelity than wild type (17). Interestingly, these aromatic resides seem to interact differently with template (Tyr-671) or substrate  in the open and closed structures, which suggests that they might function as a molecular chaperon (4 -6, 18). Lys-663, Phe-667, and Tyr-671 in Taq pol I (and the corresponding residues in other class A DNA polymerases) and Arg-659 in Taq pol I are essential for genetic complementation of a pol I ts strain of E. coli (10). These motif B residues may constitute the wall of the catalytic pocket together with residues located in other motifs (for example, see Refs. 19 -23). Moreover, other motif B mutants have altered DNA replication fidelity (18, 24 -26). Thus motif B is important for catalytic activity and fidelity of DNA synthesis in class A DNA polymerases.
In class B DNA polymerases, mutants at the Lys residues corresponding to Taq pol I Lys-663 demonstrate large increases in K m for the incoming nucleotide (27)(28)(29). However, compared with class A DNA polymerases, less is known about the function of other motif B residues in class B DNA polymerases. Class B DNA polymerases share a motif in the P helix that is similar to motif B in the O helix of class A DNA polymerases and has the sequence Q---K---N--YG. This motif in the P helix and residues in the N helix may together perform a comparable function to motif B (30,31). In the P helix of class B DNA polymerases, the distance between the Lys and Tyr is one residue shorter than in the O helix of class A DNA polymerases (32). As a result, the conserved residues in the putative ␣ helices do not have the same spatial relationships among class A and B polymerases and may be displaced by an approximately [1/4] helical turn. In the accompanying article (32), data are presented suggesting that conserved Gly-952 in Saccharomyces cerevisiae pol ␣ may play a similar role to motif B residue Tyr-951 in class A DNA polymerases. This idea is consistent with the hypothesis that the active sites of class A and B DNA polymerases are structurally distinct (32).
This manuscript presents steady-state kinetic analysis of pol ␣ G952A and G952Y and shows that Gly-952 is a critical residue for correct nucleotide incorporation by S. cerevisiae pol ␣. Parameters that influence polymerase fidelity include template-substrate interaction (33), stability of the catalytic complex (18,26,34), and efficiency of correct nucleotide incorporation (35,36). This study demonstrates that correct nucleotide incorporation efficiency is altered in pol ␣ G952A producing a dramatic loss of DNA polymerase fidelity. Thus, this mutant demonstrates that efficiency of correct nucleotide incorporation is directly related to polymerase fidelity.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis and Enzyme Preparation-Expression, purification, and specific activity determination of recombinant pol ␣ are described (32).
Primer Extension Assay-Primer extension assay was performed as described previously (24) with slight modifications. The 14-mer DNA primer, 5Ј-CGC GCC GAA TTC CC-3Ј was 5Ј-32 P-labeled, annealed to the 46-mer DNA template strand, 5Ј-GCG CGG AAG CTT GGC TGC AGA ATA TTG CTA GCG GGA ATT CGG CGC G-3Ј. The reaction was carried out in a volume of 25 l containing 10 nM template-primer, 100 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 50 mM KCl, 100 ng/ml BSA, 2 mM DTT, 1250 M dNTP, and 10 nM wild type pol ␣, 30 nM G952Y, or 70 nM G952A. Reactions were incubated at 37°C for 10 min. The amount of enzyme in the reaction was adjusted such that 80 -90% of the primer is utilized in the presence of correct nucleotide dGMP.
Single Nucleotide Incorporation Kinetics-Steady-state kinetic constants were determined for incorporation of correct and incorrect nucleotides (Ultrapure dNTP; Amersham Biosciences). 32 P-Labeled 14mer (5Ј-CGC GCC GAA TTC CC-3Ј), 15-mer (5Ј-CGC GCC GAA TTC CCG-3Ј), 16-mer (5Ј-CGC GCC GAA TTC CCG C-3Ј), or 17-mer (5Ј-CGC GCC GAA TTC CCG CT-3Ј) primer was annealed to 46-mer of oligonucleotide, 5Ј-GCG CGG AAG CTT GGC TGC AGA ATA TTG CTA GCG GGA ATT CGG CGC G-3Ј, where underlined nucleotides were the target template sites. Reaction conditions were used such that 20% of template-primer was utilized (37). The reaction mixture contained 100 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 50 mM KCl, 100 ng/l BSA, 2 mM DTT, 10 nM template-primer, various concentrations of dNTP, and the optimum concentration of DNA polymerase. Reaction was incubated at 37°C for 5 (wild type) or 10 min (mutants) in a final volume of 20 l or with appropriate changes to obtain the proper reaction efficiency and substrate utilization. Reactions were terminated and analyzed as described above, using data from at least three experiments. Kinetic parameters (K m and k cat ) for incorporation of each dNMP were determined by Hanes-Woolf plots.
Single Nucleotide Extension Kinetics-Kinetics were examined from mismatched termini as follows. 32 P-Labeled 15-mer (5Ј-CGC GCC GAA TTC CCN-3Ј), 16-mer (5Ј-CGC GCC GAA TTC CCG N-3Ј), 17-mer (5Ј-CGC GCC GAA TTC CCG CN-3Ј), or 18-mer (5Ј-CGC GCC GAA TTC CCG CTN-3Ј) primers were annealed to the 46-mer oligonucleotide, 5Ј-GCG CGG AAG CTT GGC TGC AGA ATA TTG CTA GCG GGA ATT CGG CGC G-3Ј, where N represents either one of the four nucleotides, and the underlined nucleotides were the target template sites. The reaction mixture (20 l) contained 100 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 50 mM KCl, 100 ng/l BSA, 2 mM DTT, 10 nM template-primer, various concentrations of dNTP, and an optimum concentration of each enzyme. The mixture was incubated at 37°C for 20 min (wild type) or 60 min (G952A). Reaction products were analyzed as described above, using data from at least three experiments. Under the conditions employed, reaction product increased linearly for up to an hour for both wild type and G952A pol ␣.
Enzyme Dissociation Constants-Apparent K m and k cat were determined as described above except that the DNA concentration was varied from 10 to 200 nM. K D (DNA) values were determined with DNA concentration of 40 and 200 nM using the following equation, K D ϭ [D low ](k cat / K m ) high /(k cat /K m ) low (38), in which k cat /K m , is an average of data from three independent experiments. Relative dissociation constant, K D (rel), was determined using the equilibrium binding method (32,38) as follows: The 5Ј-32 P-labeled 15-/46-mer primer-template (sequences described above) was competed for extension with each unlabeled templateprimer or challenge DNA (single-or double-stranded 46-mer DNA). Equimolar (10 nM) 5Ј-32 P-labeled template-primer and unlabeled template-primer/the challenge DNA were incubated on ice for 5 min with wild type (26 nM) or G952A (140 nM) pol ␣ in 25 l containing 100 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 50 mM KCl, 100 ng/ml BSA, 2 mM DTT. The reaction was initiated by adding 1250 M dCTP and 418 nM trap DNA (double-stranded M13 mp2 DNA), incubated at 37°C for 10 min and analyzed as described above. K D (rel) was determined using the average % extension values, thus S.D. values are not given, although three independent experiments were carried out.
Kinetic parameters for single nucleotide insertion opposite the abasic site were performed at 37°C for 30 min in a 5-l mixture containing 100 mM Tris HCl, pH 8.0, 5 mM MgCl 2 , 50 mM KCl, 100 ng/l BSA, 2 mM DTT, 40 nM template-primer, and various amounts of enzyme and dNTP. For wild type, 312-4500 M of each nucleotide and 5.8 -87.7 nM enzyme were used. In reactions with G952A, substrate concentration and enzyme were 39 -4500 M and 7.4 -21 nM, respectively.

RESULTS
Primer Extension Analysis-Gly-952 is one of the critical residues for the functions of S. cerevisiae DNA pol ␣ (32). Mutants of Gly-952 do not complement the temperature sensitivity of pol ␣-deficient yeast, and Ala, Tyr, Arg, and Glu substitution mutants of Gly-952 have severely impaired catalytic function. The catalytic activity of Glu and Arg substitution mutants was not measurable, and specific activity of the Tyr and Ala mutants was reduced 260-and 1500-fold, respectively.
The crystal structure of RB69 DNA polymerase suggests a putative function for Gly-952 in S. cerevisiae DNA pol ␣. In this structure, the residue analogous to Gly-952 lies in close proximity with the template base (30) (Fig. 1), suggesting that Gly-952 may interact with the template. This interaction may be altered in Gly-952 mutants leading to loss of fidelity. This possibility was tested by measuring misincorporation using a primer extension assay in the presence of the correct or incorrect nucleotide substrate. Fig. 2A shows a primer extension assay with wild type, G952A, and G952Y pol ␣. Wild type pol ␣ incorporated primarily dGMP opposite template C with some misincorporation of dTMP and dAMP. In contrast, G952Y incorporated dAMP and dGMP with equal efficiency ( Fig. 2A, lanes [11][12], and G952A incorporated the three incorrect nucleotides and dGMP in similar amounts ( Fig. 2A, lanes 15-18). The low fidelity extension was also observed by using other three template bases; templates T, A, and G (data not shown). No incorporation was observed when the template was omitted from reactions with wild type or mutant pol ␣ ( Fig. 2A, lanes 7, 13, and 19), showing that the misincorporations are template-dependent. These results demonstrate that G952Y and G952A pol ␣ are low fidelity DNA polymerases.
Processivity Analysis-Results of the primer extension assay show that wild type pol ␣ elongates the primer to nearly the full-length of the 46-mer template ( Fig. 2A, lane 2), but Gly-952 mutants are much less efficient, adding one or two nucleotides to the primer and producing a 15-or 16-mer ( Fig. 2A, lanes 8  and 14). This suggests that the mutant polymerases may have reduced processivity. The processivity was measured directly using a poly(dA) template and correct nucleotide dTTP. With limiting enzyme, wild type pol ␣ incorporated up to 10 nucleotides, but mutant polymerases G952Y and G952A incorporated one to four nucleotides (Fig. 2B). These results suggest that G952Y and G952A have intrinsic low processivity, which contributes to low efficiency during the primer extension assay ( Fig. 2A).
However, short primer elongation might also result if the mutant polymerases generate mismatched primer termini that are extended poorly. The following observations support this idea. 1) In the presence of four dNTPs, primer extension products were shorter for G952A than for G952Y ( Fig. 2A, lanes 8  and 14); however, these enzymes have similar processivity (Fig.  2B, lanes 9 and 17; measured on poly(dA) in the presence of dTTP such that the enzyme can not form mismatches). 2) Primer extension products tended to form a doublet ( Fig. 2A,  lanes 8 and 14) 2 or to migrate at illegitimate positions (compare Fig. 2A, lanes 14 and 18). These results suggest that mutant polymerases generated the mismatched primer termini, which forced to cease the primer extension. Therefore we examined single nucleotide insertion and misextension kinetics of Gly-952 pol ␣ mutants.
Determination of Kinetic Values-Steady-state kinetic parameters for single nucleotide incorporation were determined using a gel-based assay (Table I). This assay measures the discrimination factor (DF), which is the ratio k cat /K m (i.e. incorporation efficiency) for the incorrect nucleotide to k cat /K m for the correct nucleotide. Wild type pol ␣ incorporated the incorrect nucleotide with discrimination factors between 3.3 ϫ 10 Ϫ5 and 8.4 ϫ 10 Ϫ4 (Table I). Kinetic parameters for wild type S. cerevisiae pol ␣ were similar to kinetic parameters of calf thymus, Drosophila, and human pol ␣ (41)(42)(43). However, at some template bases, the kinetics of S. cerevisiae pol ␣ differed by Ͼ10-fold from some other ␣ polymerases. For example, S. cerevisiae pol ␣ incorporated A:dGMP, C:dAMP, and T:dGMP more efficiently than calf thymus pol ␣ (43). However, this result may reflect different nucleotide sequences instead of reflecting different biochemical characteristics of different enzymes (for example see Ref. 42).
The results with Gly-952 pol ␣ mutants were unique and very different from the results with wild type enzyme. At template G, G952Y pol ␣ incorporated dAMP more efficiently than the correct dCMP (1.2), and the other incorrect substrates, dTMP and dGMP, were incorporated with similar efficiency as dCMP (0.82 and 0.32, respectively). At templates C and A, G952Y frequently misincorporated dAMP with a DF of 0.39 and 0.64, respectively. G952A preferentially incorporated the correct nucleotide only at template A, but frequently misinserted dCMP opposite A with a DF of 0.37. At other template bases, misincorporation was more frequent than incorporation of the correct nucleotide with preferential misincorporation of pyrimidines at all template bases (1.3 to 5.6).
The polymerase activity of wild type and G952A pol ␣ was 2 Doublet bands were observed with shorter exposure time and on the digital image from the Fuji Image Analyzer (data not shown).  2. A, nucleotide selectivity of wild type, G952Y, and G952A incorporation. 10 nM 32 P-labeled 14-mer was annealed with 46-mer template and used as a template-primer. Template was omitted in some experiments (lanes 7, 13, and 19). Reactions were adjusted such that 80 -90% of primer was utilized in the presence of the correct nucleotide. Reactions contained 1250 M dATP, dCTP, dGTP, or dTTP and no enzyme, wild type (10 nM), G952Y (30 nM), or G952A pol ␣ (70 nM). Reactions were incubated at 37°C for 20 min. The primer was positioned so that the first nucleotide was the target. Template sequence (right side), primer position (triangle, left side), and doublet bands (filled triangle) are indicated. B, processivity of wild type, G952Y, and G952A pol ␣. Polymerase processivity was measured using 32 P-labeled oligo(dT) 16 annealed to poly(dA) as described under "Experimental Procedures." Wild type (lanes 1-8), G952Y (lanes 9 -16), and G952A (lanes 17-24) pol ␣ was added at a concentration of 320 to 2.5 ng per reaction. Open and closed triangles indicate the position of primer and the gel top, respectively.  also measured using DNA substrates with terminal mismatches (Table II). The misextension efficiencies were similar to those of pol ␣ from other species (41,43,44), although extension from A:G (template:primer) was Ͼ10-fold more efficient than calf thymus pol ␣, and extension from C:C, G:T, T:G, and A:G was Ͼ10-fold less efficient than Drosophila pol ␣. Again, G952A pol ␣ was error-prone during primer extension with mismatched substrates with terminal T:C (template:primer), T:T, C:A, and A:C mismatches. Discrimination factors were 0.25-0.34, and the apparent K m values were 71to 460-fold lower than from matched termini. It is noteworthy that the K m from a matched terminus was always higher than from a mismatched terminus.  3. DNA synthesis and nucleotide selectivity of wild type, G952A, and G952Y pol ␣ on damaged DNA templates. Nucleotide incorporation was carried out using the following template lesions/ enzyme conditions: abasic site with 6 nM polymerase (A), CPD photoproduct with 48 nM polymerase (B), or (6 -4) photoproduct with 24 nM polymerase (C). Each enzyme was incubated with undamaged template (lanes 1, 2, 4, and 6) or damaged template (other lanes) at 37°C for 60 min. dNTPs were included in lanes 1-9, 14, and 19; enzyme additions were as follows: no enzyme (lane 1), wild type (lanes 2 and 3), G952Y (lanes 4 and 5) or G952A (lanes 6 and 7). Reactions contained dTTP (lanes 10, 15, and 20), dCTP (lanes 11, 16, and 21), dATP (lanes 12, 17,  and 22), or dGTP (lanes 13, 18, and 23). Enzyme additions were as follows: no enzyme (lane 8), wild type (lanes 9 -13), G952Y (lanes 14 -18), and G952A (lanes 19 -23). Sequences of the templates (right side) and primer position (triangle, both sides) are indicated. The primer on each template was positioned so that the first or first two nucleotides were opposite the lesion. Note that two gels are shown in panel B (lanes  1-7 and 8 -23).
Nucleotide Incorporation at Damaged Template Sites-Gly-952 mutant DNA polymerases showed very poor discrimination of incorrect substrates. These data suggest that the structure of the catalytic pocket may be severely altered in these mutants. If this is the case, the mutants might utilize damaged DNA templates differently from wild type pol ␣. This possibility was tested by measuring nucleotide incorporation specificity opposite an abasic, cis-syn cyclobutane di-thymine (CPD), and (6 -4) di-thymine photoproduct in the template strand of the DNA substrate.
These results suggest that the Gly-952 mutants incorporate a single nucleotide opposite abasic, CPD, and (6 -4) photoproduct lesions more efficiently than wild type pol ␣. As observed with undamaged DNA substrates (see Table I and Fig. 2A), G952A tends to incorporate a pyrimidine opposite the target, and G952Y prefers to incorporate dAMP opposite the damaged template sites.
Steady-state kinetic analyses for polymerase activity with damaged DNA substrates are summarized in Table III. Wild type pol ␣ incorporated dAMP opposite the abasic site with 10to 50-fold higher efficiency than the other three nucleotides, although the abasic site was utilized 1000-fold less efficiently than the undamaged template T. In contrast, G952A incorporated dTMP, dCMP, and dAMP opposite the abasic site 12-, 4.3-, and 1.3-fold more efficiently than opposite T on the control template, respectively.
DNA Binding Affinity of Wild Type and Mutant pol ␣-K D (DNA) was measured using wild type or mutant pol ␣ and three template primers as DNA substrates. K D (DNA) for each DNA substrate differed by Ͻ2-fold for wild type and G952A pol ␣ (Table IV). Relative binding affinity for each DNA substrates (single-stranded oligonucleotide, double-stranded oligonucleotide, or template-primer DNA) was also similar. These results indicate that the DNA binding affinity of G952A pol ␣ is near wild type, despite the fact that G952A has low catalytic activity and low processivity.

DISCUSSION
Incorporation of Correct Nucleotides-In the amino acid sequence, Gly-952 of S. cerevisiae pol ␣ is one of the most conserved residues among the classes of DNA polymerases. In the accompanying article (32), we showed that Gly-952 is one of the essential amino acids in the catalysis of S. cerevisiae pol ␣. In this work, this residue is further studied and discussed in the viewpoint of fidelity DNA synthesis.
DNA polymerases vary widely in their DNA replication fidelity. Nevertheless, it is unusual and surprising that G952A pol ␣ incorporates incorrect nucleotides more efficiently than correct nucleotides. Few previous studies report DNA polymerases with similarly low fidelity; these include pol and a mutant derivative of pol ␤. DNA pol preferentially inserts dGMP opposite template T (45), and mutant R283A pol ␤ incorporates dGMP and dAMP with equal efficiency opposite template T (46). In both cases, however, the correct nucleotides are the best substrates at the other three template bases. In contrast, G952A pol ␣ incorporates incorrect nucleotides more efficiently than correct nucleotides at templates C, G, and T (Table I). Fidelity is highest at template A, where the ratio of misincorporation of dCMP to dTMP (discrimination factor) was as high as 0.37. This result is even more surprising if one considers the fact that Watson-Crick base pairs form 15-to 150-fold more frequently than non Watson-Crick base pairs (47,48); thus, G952A pol ␣ selectively catalyzes nucleotide incorporation from a minor pool of non-Watson-Crick base pairs.
The unique misincorporation specificity of G952A pol ␣ is not because of an increase in misincorporation efficiency but because of a decrease in correct nucleotide incorporation efficiency. In particular, correct nucleotide incorporation is 500-to 3500-fold less efficient for G952A than for wild type, but the efficiency of pyrimidine misincorporation was similar to or moderately higher than wild type (i.e. 1.2-to 22-fold higher). This suggests that G952A has a defect in correct but not incorrect nucleotide insertion and that Gly-952 in S. cerevisiae pol ␣ plays a key role in the incorporation of correct nucleotides. Recently Beard et al. (36) showed that the error-prone DNA polymerases with low intrinsic fidelity insert correct nucleotides with low efficiency. Therefore, the basic discrimination mechanism revealed by G952A pol ␣ may also play a role in determining the fidelity of polymerases active in normal DNA metabolism.
Structural Implication-In Taq pol I, a class A DNA polymerase, four residues (Arg-659, Lys-663, Phe-667, and Tyr-671) in the R---K---F---YG motif are evolutionally conserved and functionally important (10). In S. cerevisiae pol ␣, (a class B DNA polymerase), the conserved residues do not have the same spatial relationship and may differ by an approximately [1/4] helical turn (32). This is consistent with the observation that C␤ of Tyr-671 (Taq) and C␣ of Gly-568 (analogous to Gly-952 in S. cerevisiae pol ␣) come within a 1.5-Å proximity when the nascent base pairs in the catalytic complex of Taq pol I and RB69 polymerase are superimposed (Fig. 1). This result suggests that the functional equivalent of Tyr-671 in Taq pol I may be Gly-952 (not Tyr-951) in S. cerevisiae pol ␣.
In the structure of RB69 polymerase, the C␣ of Gly-568 and the adenine N3 are within 3 Å, which does not leave room to accommodate an amino acid side chain without dislocating the template adenine. G952A pol ␣ forms pyrimidine-pyrimidine base pairs more efficiently than other base pairs, which might indicate that these small mispairs fit the catalytic pocket better than purine-containing base pairs. Therefore, one possible explanation for the low fidelity of G952A is that the Ala side chain narrows the catalytic pocket leading to altered catalytic efficiency and substrate specificity.
It is interesting that G952A pol ␣ has wild type DNA binding affinity despite its putative structural difference from wild type a K D was kinetically determined as described under "Experimental Procedures." This table shows the data from C:G as the template:primer terminus. Wild type and G952A provided K D from G:C, A:T-matched termini with 82 and 160, and 120 and 100, respectively. b K D (rel) value is the ratio of the binding affinity of polymerase enzyme for template primer DNA (T-P) to challenged DNA (Singlestranded (ss) or double-stranded (ds) DNA). pol ␣. Replicative DNA polymerases interact electrostatically with ϳ10 nucleotides of a double-stranded DNA substrate (6,30). This ϳ10-nucleotide site of interaction may be long enough to compensate for the proposed local clash between the template base and substituted side chain; this would minimize the phenotypic effect on DNA binding but still lead to the observed inefficient single nucleotide incorporation and limited processivity of DNA synthesis in mutant G952A pol ␣.
Although this model explains the presented data, it may or may not be correct to assume that the crystal structure of wild type RB69 DNA polymerase can be used to accurately explain the characteristics of Gly-952 mutants of S. cerevisiae DNA pol ␣. For example, the motif B structure of phage DNA polymerases and S. cerevisiae DNA pol ␣ are not biochemically identical. Mutation of Tyr-567 in RB69 DNA polymerase is associated with low fidelity DNA synthesis, whereas Pro substitution of the corresponding Tyr-951 in S. cerevisiae DNA pol ␣ does not alter the mutation frequency (32). In addition, this study shows that G952A pol ␣ has wild type DNA affinity and K D (Table IV), whereas the corresponding G391D mutant of ⌽29 DNA polymerase is deficient in template-primer binding (49).
Function of YG Sequence in S. cerevisiae pol ␣-Measurement of the apparent kinetic values showed that the inefficient nucleotide incorporation is essentially attributed to K m increase. At some sequence context, k cat and K D were slightly impaired (see Tables I and IV).
Beside this, Gly-952 mutant DNA polymerase was impaired in processivity (Fig. 2B). Processivity could be reduced because of the combined effect of k cat and K D , depending on the sequence context. Alternatively Gly-952 mutant DNA polymerase may have a defect in translocation, which would affect the second nucleotide incorporation cycle. Because similar results were obtained for Y951P pol ␣ (32), the YG sequence in motif B may play a role in processive DNA synthesis by S. cerevisiae pol ␣.
Utilization of Damaged DNA Templates-DNA pol ␣ has no significant translesion polymerase activity. When it is forced to utilize an abasic site in the template strand, it preferentially inserts dAMP. G952A pol ␣ also did not read through damaged DNA templates but incorporated a single dTMP or dCMP opposite AP, CPD, and (6 -4) photoproduct lesions (see Fig. 3 and Table III). Thus, pyrimidines are preferentially inserted at both undamaged and damaged template sites (see Tables I and III), suggesting that G952A pol ␣ may intrinsically prefer to insert pyrimidines, regardless of the template base.
Interestingly, Gly-952 mutant pol ␣ incorporated nucleotides more efficiently than wild type at an AP site, CPD, or (6 -4) photoproduct in a primer extension assay. These results were confirmed by kinetic analysis at the AP site, where G952A incorporated dTMP 80-fold more efficiently than wild type pol ␣ (Table III). Recent structural studies of translesion DNA polymerases suggest that a relatively wide catalytic pocket in these enzymes may accommodate two nucleotides during bypass of thymine dimers (50). However, as discussed above, the catalytic pocket of the Gly-952 mutant pol ␣ may be narrower than wild type, suggesting that an unusual mechanism may contribute to the observed high efficiency incorporation with a damaged DNA template. This possibility will be discussed further elsewhere.
In conclusion, G952A pol ␣ is impaired in correct nucleotide incorporation with undamaged DNA and is more efficient in nucleotide incorporation at certain damaged template sites, despite the fact that this enzyme may have a narrow catalytic pocket. This study shows that Gly-952 is catalytically essential in S. cerevisiae DNA pol ␣ and plays an important role in error prevention during nucleotide incorporation, extension, and translesion bypass during DNA replication.