Hydrogen Bonding of 7,8-Dihydro-8-oxodeoxyguanosine with a Charged Residue in the Little Finger Domain Determines Miscoding Events in Sulfolobus solfataricus DNA Polymerase Dpo4*♦

Sulfolobus solfataricus P2 DNA polymerase IV (Dpo4) has been shown to catalyze bypass of 7,8-dihydro-8-oxodeoxyguanosine (8-oxoG) in a highly efficient and relatively accurate manner. Crystal structures have revealed a potential role for Arg332 in stabilizing the anti conformation of the 8-oxoG template base by means of a hydrogen bond or ion-dipole pair, which results in an increased enzymatic efficiency for dCTP insertion and makes formation of a Hoogsteen pair between 8-oxoG and dATP less favorable. Site-directed mutagenesis was used to replace Arg332 with Ala, Glu, Leu, or His in order to probe the importance of Arg332 in accurate and efficient bypass of 8-oxoG. The double mutant Ala331Ala332 was also prepared to address the contribution of Arg331. Transientstate kinetic results suggest that Glu332 retains fidelity against bypass of 8-oxoG that is similar to wild type Dpo4, a result that was confirmed by tandem mass spectrometric analysis of full-length extension products. A crystal structure of the Dpo4 Glu332 mutant and 8-oxoG:C pair revealed water-mediated hydrogen bonds between Glu332 and the O-8 atom of 8-oxoG. The space normally occupied by Arg332 side chain is empty in the crystal structures of the Ala332 mutant. Two other crystal structures show that a Hoogsteen base pair is formed between 8-oxoG and A in the active site of both Glu332 and Ala332 mutants. These results support the view that a bond between Arg332 and 8-oxoG plays a role in determining the fidelity and efficiency of Dpo4-catalyzed bypass of the lesion.

Enzyme-catalyzed reactions are fundamental in biology (1)(2)(3). Certain elements of an enzymatic mechanism may be obvious once elucidated (i.e. required catalytic residues, prosthetic groups, the presence or absence of cofactors), and the relevance for function is readily discerned. In other instances, features of interest may be difficult to identify in the multitude of non-essential elements that often comprise enzyme-substrate interactions (4,5). For example, nucleotide selection and subsequent incorporation by DNA polymerases have been described in great detail (6,7), and the chemical components necessary for phosphoryl transfer are conserved across all domains of life and across all polymerase subfamilies. Yet, in the course of analyzing different polymerases it has become apparent that not all of these enzymes use the same mechanism(s) to define substrate specificity, a vital aspect of polymerase function (8).
The Y-family DNA polymerases appear to present a case of relaxed substrate selection, but the less accurate mode of copying template DNA is supplanted by the greater propensity of Y-family polymerases to effectively utilize damaged DNA as a substrate (9 -12). Indeed, the ubiquitous nature of DNA damage, both endogenous and exogenous in origin, makes it necessary for the replisome to have some means of bypassing covalently modified DNA (13). Some general features unique to Y-family polymerases that are relevant to translesion DNA synthesis include an active site that leaves the newly formed base pair relatively unconstrained by protein-DNA interactions and employment of an additional domain termed the "little finger" or palm-associated domain (PAD), which has important contacts with the template DNA near the active site (14,15). Other mechanistic features of the Y-family that likely contribute to effective partitioning between low and high-fidelity polymerases include low processivity, relatively slow forward rates of polymerization (k pol ), and the requirement for high dNTP concentrations to achieve maximum catalytic rates (i.e. "high" K D ,dNTP) (16 -19). Previous work from our group has shown that a model Y-family polymerase, Dpo4 from Sulfolobus solfataricus P2, is able to bypass 8-oxoG, 3 a major lesion arising from oxidative stress, in a highly accurate and efficient manner (20). The kinetic parameters and LC-MS analysis of extension products indicated that Dpo4 was ϳ20-fold more efficient at insertion of dCTP opposite 8-oxoG relative to dATP insertion. X-ray crystal structures revealed that the 8-oxoG:A pair was in the syn:anti configuration, which allows A to form a Hoogsteen pair with 8-oxoG. Conversely, the 8-oxoG:C pair was in the Watson-Crick geometry and such a configuration appeared to be stabilized by either a hydrogen bond or an ion-dipole pair between the O-8 atom and the side chain of Arg 332 . The importance of such an interaction was further confirmed by separate studies that observed a water-mediated hydrogen bond between Arg 332 and 8-oxoG (21).
Thermodynamically and geometrically the 8-oxoG:A pair is similar to a T:A pair. In isolated oligonucleotides the 8-oxoG:A pair decreases the T m by 5.5°C relative to T:A, and the T m for the corresponding 8-oxoG:C pair is 16.9°C lower than a G:C pair (22). The question then arises as to why Dpo4 retains high efficiency and fidelity during bypass of 8-oxoG. The role of Arg 332 in nucleotide selection during Dpo4-catalyzed bypass of 8-oxoG was investigated using a series of mutant enzymes. In an effort to determine what factors contribute to the stabilization of the 8-oxoG:C pair, site-directed mutagenesis was used to alter different chemical aspects of the Arg 332 side chain, including steric occupancy and hydrogen bonding potential. Transient-state kinetic and LC-MS/MS analyses were then combined with x-ray crystallographic studies to compare mutant-catalyzed bypass of 8-oxoG with wild type Dpo4.

EXPERIMENTAL PROCEDURES
Materials-Site-directed mutagenesis was performed using synthetic oligonucleotide primers and a QuikChange mutagenesis kit (Stratagene, La Jolla, CA). Wild type Dpo4 and all of the mutant enzymes were expressed in Escherichia coli and purified to electrophoretic homogeneity as described previously (23). All unlabeled dNTPs were obtained from Amersham Biosciences, and [␥-32 P]ATP was purchased from PerkinElmer Life Sciences. All oligonucleotides used in this work were synthesized by Midland Certified Reagent Co. (Midland, TX) and purified using high performance liquid chromatography by the manufacturer, with analysis by matrix-assisted laser desorption time-of-flight MS. The 13-base primer sequence used in the kinetic and mass spectral analyses was 5Ј-GGGGGAAG-GATTC-3Ј. The 14-base primer sequences used in the indicated kinetic assays and the crystal structures were 5Ј-GGGG-GAAGGATTCC-3Ј for the 8-oxoG:C structure and 5Ј-GGGGGAAGGATTCA-3Ј for the 8-oxoG:A structure. The template DNA sequence used in the kinetic and mass spectral assays and in the 8-oxoG:C and 8-oxoG:A structures was 5Ј-TCACXGAATCCTTCCCCC-3Ј, where X ϭ G or 8-oxoG, as indicated. The DNA control template sequence used in the full-length extension assay (Fig. 1B) was 5Ј-TCATGGAATC-CTTCCCCC-3Ј.
Full-length Extension Assay-A 32 P-labeled primer was annealed to either an unmodified or adducted template oligonucleotide. Each reaction was initiated by adding dNTP⅐Mg 2ϩ (each dNTP at 250 M and 5 mM MgCl 2 ) solution to a preincubated Dpo4⅐DNA complex (100 nM Dpo4 and 200 nM DNA).
The reaction was carried out at 37°C in 50 mM Tris-HCl (pH 7.4) buffer containing 50 mM NaCl, 5 mM DTT, 100 g l Ϫ1 bovine serum albumin, and 5% (v/v) glycerol. At the indicated time, 5-l aliquots were quenched with 50 l of 500 mM EDTA (pH 9.0). The samples were then mixed with 100 l of a 95% formamide/20 mM EDTA solution and separated on a 20% polyacrylamide (w/v)/7 M urea gel. Products were visualized and quantified using a phosphorimaging screen and Quantity One TM software, respectively (Bio-Rad). Formation of an 18-base extension product from a 13-base primer was quantified by fitting the data to Equation 1, where A ϭ amount of product formed during the first binding event between Dpo4 and DNA, k obs ϭ an observed rate constant defining nucleotide incorporation, n ϭ number of incorporation events required to observe product formation, and t ϭ time. All statistical values given indicate S.E.
Transient-state Kinetics-All pre-steady-state experiments were performed using a KinTek RQF-3 model chemical quench-flow apparatus (KinTek Corp., Austin, TX) with 50 mM Tris-HCl (pH 7.4) buffer in the drive syringes. All RQF experiments were carried out at 37°C in a buffer containing 50 mM Tris-HCl (pH 7.4) buffer containing 50 mM NaCl, 5 mM DTT, 100 g l Ϫ1 bovine serum albumin, and 5% (v/v) glycerol. Polymerase catalysis was stopped by the addition of 500 mM EDTA (pH 9.0). Substrate and product DNA was separated by electrophoresis on a 20% polyacrylamide (w/v)/7 M urea gel. The products were then visualized using a phosphorimaging device and quantitated using Quantity One TM software. Results obtained under single-turnover conditions were fit to Equation 2, where A ϭ product formed in first binding event, k obs ϭ rate constant defining polymerization under the conditions used for the experiment being analyzed, and t ϭ time. Results obtained under conditions that allowed a second round of Dpo4⅐DNA binding and polymerase action were fit to Equation 3, where k ss represents a steady-state velocity of nucleotide incorporation. To obtain an estimate of the nucleotide binding affinity for each mutant, the concentration of dNTP in the reaction mixture was varied, and pre-steady-state experiments were performed under excess enzyme conditions. The resulting rate constants, k obs , were then plotted as a function of dNTP concentration, and the data were fit to the hyperbolic expression k obs ϭ (k pol ⅐[dNTP])/([dNTP] ϩ K D,dNTP ) using GraphPad Prism.
LC-MS/MS Analysis of Oligonucleotide Products from Dpo4 Reactions-Dpo4 (5 M) was preincubated with primer-template DNA (10 M), and the reaction was initiated by the addition of dNTP (1 mM each) and MgCl 2 (5 mM) in a final volume of 100 l. Dpo4 catalysis was allowed to proceed at 37°C for 4 h in NaCl, 1 mM DTT, 50 g l Ϫ1 bovine serum albumin, and 5% glycerol (v/v). The reaction was terminated by extraction of the remaining dNTPs using a size-exclusion chromatography column (Bio-Spin 6 chromatography column, Bio-Rad). Concentrated stocks of Tris-HCl, DTT, and EDTA were added to restore the concentrations to 50, 5, and 1 mM, respectively. Next, E. coli uracil-DNA glycosylase (20 units, Sigma-Aldrich) was added, and the solution was incubated at 37°C for 6 h to hydrolyze the uracil residue on the extended primer. The reaction mixture was then heated at 95°C for 1 h in the presence of 0.25 M piperidine followed by removal of the solvent by centrifugation under vacuum. The dried sample was resuspended in 100 l of H 2 O for MS analysis.
LC-MS/MS analysis (23,24) was performed on a Waters Aquity ultraperformance liquid chromatography system (UPLC; Waters, Milford, MA) connected to a Finnigan LTQ mass spectrometer (Thermo Fisher Scientific, Waltham, MA) operating in the ESI negative ion mode. An Aquity UPLC BEH octadecylsilane (C 18 ) column (1.7 m, 1.0 ϫ 100 mm) was used with the following LC conditions: buffer A contained 10 mM NH 4 CH 3 CO 2 plus 2% CH 3 CN (v/v), and buffer B contained 10 mM NH 4 CH 3 CO 2 plus 95% CH 3 CN (v/v). The following gradient program was used with a flow rate of 150 l min Ϫ1 : 0-3 min, linear gradient from 100% A to 97%A, 3% B (v/v); 3-4.5 min, linear gradient to 80% A, 20% B (v/v); 4 -5.5 min, linear gradient to 100% B; 5-5.5 min, hold at 100% B; 5.5-6.5 min, linear gradient to 100% A; 6.5-9.5 min, hold at 100% A. The temperature of the column was maintained at 50°C. Samples were injected with an autosampler system. ESI conditions were as follow: source voltage 4 kV, source current 100 A, auxiliary gas flow rate setting 20, sweep gas flow rate setting 5, sheath gas flow setting 34, capillary voltage Ϫ49 V, capillary temperature 350°C, tube lens voltage Ϫ90 V. MS/MS conditions were as follows: normalized collision energy 35%, activation Q 0.250, activation time 30 ms. The doubly (negatively) charged species were generally used for CID analysis. The calculations of the CID fragmentations of oligonucleotide sequences were done using a program linked to the Mass Spectrometry Group of Medicinal Chemistry at the University of Utah (www.medlib.med.utah.edu/massspec). The nomenclature used in supplemental Tables S1-S5 has been described previously (25).
Structure Determination and Refinement-The refined wild type Dpo4⅐dG complex (Protein Data Bank accession code 2c22 (20)) minus solvent molecules and dGTP was used as the starting model for R332A (8-oxoG:C). The initial position of the model was optimized by several rounds of rigid body refinement while gradually increasing the resolution of the diffraction data. The refined structure of the R332A(8-oxoG:C) complex served as the starting model for the other three crystals, and the locations of the individual models were optimized by rigid body refinement as described above.
Manual model rebuilding was done with the program TUR-BO-FROODO. 4 The maps were computed using the A-modified coefficients (30). Clear positive density for the Ca 2ϩ ions and the dGTP was observed in the initial difference Fourier electron density maps of all four complexes. Also, unambiguous negative or positive density (regarding the initial model selected; see above) appeared for the mutated residues, which were then replaced with the correct residue type. The CNS package (crystallography NMR software) (31) was used for the refinement of the models by performing simulated annealing, gradient minimization, and refinement of individual isotropic temperature factors. The statistics of the refined models for all structures are summarized in Table 5, and representative electron density maps for the final models are depicted in the supplemental data section. The crystallographic figures were prepared using PyMOL. 5

Primer Extension Past G and 8-OxoG Using All Four dNTPs-
As a general measure of how Dpo4 catalysis opposite 8-oxoG is affected by mutating the Arg 332 residue, a time course was performed for each mutant in the presence of all four dNTPs ( Fig.  1). An observed rate constant defining five incorporation events can be measured by following the appearance of the fully extended 18-mer primer and fitting the data to Equation 1, where n ϭ 5 ( Table 1). The Ala 332 and Glu 332 mutants were each ϳ2-fold faster at full-length extension opposite unmodified DNA than WT Dpo4 or the other mutants. However, WT Dpo4 was ϳ2-6-fold faster than any of the mutants at full-length extension opposite 8-oxoG. To address the potential contribution of Arg 331 to bypass of 8-oxoG, a double alanine mutant (Ala 331 Ala 332 ) was prepared. The Ala 331 Ala 332 double mutant failed to extend the primer during the first binding event, but full-length extension by the double mutant did occur under conditions that allowed multiple binding events (supplemental Fig.  S1). It is important to note that in the absence of any other evidence, the exact identity of the fully extended products is unknown.

Analysis of Primer Incorporation/ Extension Products Using LC-MS/ MS with Dpo4 Extension Products-
To identify the sequence of the fulllength extension products ( Fig. 1), LC-MS/MS analysis of the oligonucleotide products was used, as described previously (Fig. 2) (23,24). Control reactions with unmodified DNA were performed. Accurate copying of the unmodified DNA template was reflected in the only product observed for each of the mutants, indicating that fidelity is unperturbed (Ͼ99% C incorporation, data not shown). The reactions with 8-oxoG were performed in triplicate to obtain a statistical evaluation of the mutant insertion spectrum ( Table 2, supplemental Tables  S1-S5, and supplemental Figs. S2-S30). All of the mutants catalyzed incorporation of C and A opposite 8-oxoG ( Table 2). Some of the products had an additional A or C added to the 3Ј terminus of the primer in what is assumed to be a template-independent manner (blunt-end addition). The range of misincorporation was small (11.1 Ϯ 0.4% A incorporation for Ala 331 Ala 332 and 16.0 Ϯ 0.8% A incorporation for His 332 ). The trend in fidelity of 8-oxoG bypass predicted from the LC-MS/MS results was: Ala 331 Ala 332 Ӎ WT Dpo4 Ӎ Glu 332 (highest C incorporation) Ͼ Leu 332 Ͼ Ala 332 Ӎ His 332 (lowest C incorporation).
Transient-state Kinetic Analysis of Dpo4 Mutants and dCTP Incorporation Opposite 8-OxoG-Single-turnover experiments were performed with the concentration of nucleotide varied in order to provide an estimate of k pol and K D,dNTP for each mutant ( Fig. 3 and Table 3). The estimates obtained in the kinetic approaches highlight the first insertion event opposite 8-oxoG, as opposed to LC-MS/MS analysis, which was conducted under conditions that allow multiple rounds of binding and dissociation.
Dpo4 Ala 332 displays several interesting characteristics. The maximum forward rate of polymerization for Ala 332 (k pol ) was 4.4-and 4.9-fold faster than observed previously with WT Dpo4 for unmodified G and 8-oxoG-modified substrates, respectively (20). The Ala 332 mutant also displayed a higher affinity (lower K D,dCTP ) for dCTP when bound to the unmodi-  (Table 1).

TABLE 1 Transient-state kinetic parameters describing full-length extension of 13/18-mer primer-template DNA
The experimental conditions were: ͓Dpo4͔ ϭ 100 nM, ͓DNA͔ ϭ 100 nM, ͓dNTP mix͔ ϭ 1 mM, ͓13/18-mer trap͔ ϭ 1 mM. ND, not determined due to lack of sufficient product formation.   In kinetic terms, the Ala 332 mutant was slightly more efficient at accurate bypass of G compared with 8-oxoG. The Glu 332 mutant also exhibited slightly different kinetics from WT Dpo4. For instance, the k pol value for dCTP incorporation opposite G was 4.8-fold faster than for WT Dpo4, similar to Ala 332 . The k pol value for dCTP incorporation opposite 8-oxoG was 3.5-fold faster than WT Dpo4 insertion of dCTP opposite 8-oxoG. The lower K D,dCTP observed with WT Dpo4 for the 8-oxoG substrate relative to G was also apparent with the Glu 332 mutant, but the absolute affinity of the Glu 332 mutant for dCTP was diminished for both G and 8-oxoG. The catalytic efficiency of the Glu 332 mutant was 2.3-fold greater than WT Dpo4 for dCTP incorporation opposite G but 1.7-fold less efficient than WT Dpo4 for incorporation opposite 8-oxoG. Overall the Glu 332 mutant was ϳ5-fold more efficient at incorporating C opposite 8-oxoG compared with insertion opposite unmodified G.
The Leu 332 mutant can be hypothesized to serve as an intermediate between the Ala 332 and Glu 332 mutations in that the potential for hydrogen bonding between residue 332 and 8-oxoG is lost, but the steric occupancy of the Leu aliphatic side chain is retained, similar to the Glu 332 mutant. Pre-steadystate kinetic analysis revealed that the k pol values for the Leu 332 mutant incorporation opposite G and 8-oxoG were 4.1-and 1.9-fold faster than those for WT Dpo4. The lower K D,dCTP when the Leu 332 mutant incorporates C opposite 8-oxoG (compared with K D,dCTP for insertion opposite G) was similar to both WT Dpo4 and the Glu 332 mutant, but the absolute value for Leu 332 -catalyzed bypass of 8-oxoG was increased relative to WT Dpo4. The Leu 332 mutant was still ϳ2-fold more efficient at incorporation of dCTP opposite 8-oxoG compared with G, but that difference was smaller than what is observed with WT Dpo4 and Glu 332 .
The His 332 mutant exhibited faster forward rate constants relative to WT Dpo4. The k pol values for the His 332 mutant incorporation opposite G and 8-oxoG were 3.6-and 4.6-fold faster than for WT Dpo4. The nucleotide binding affinity trend was opposite that of WT Dpo4, Glu 332 , and Leu 332 , with tighter dCTP binding during bypass of G. As in the case of Ala 332 , the kinetic analysis indicated that His 332 inserted dCTP opposite G with slightly greater efficiency than opposite 8-oxoG.
The Ala 331 Ala 332 double mutant had slower forward rate constants relative to WT Dpo4 for both G and 8-oxoG. The Ala 331 Ala 332 double mutant-catalyzed insertion of dCTP opposite 8-oxoG was ϳ4-fold faster than dCTP insertion opposite G. The measured binding affinity of dCTP was tighter than that of WT Dpo4 for unmodified DNA, but the binding affinity of dCTP opposite 8-oxoG was similar to that observed for WT Dpo4. The catalytic efficiency for dCTP incorporation was increased ϳ4-fold for unmodified DNA and decreased ϳ2-fold for 8-oxoG-modified DNA.  (Table 3). Similar results were obtained for the other mutants.
a Determined previously (20). b Not determined because of lack of sufficient product formation.

Role of Arg 332 in Dpo4 Bypass of 8-OxoG
Transient-state Kinetic Analysis of Dpo4 Mutants and dATP Incorporation Opposite 8-OxoG-Previous steady-state analysis indicated that Dpo4 is ϳ90-fold more efficient at dCTP incorporation opposite 8-oxoG compared with dATP incorporation opposite the lesion (20). Previous LC-MS/MS analysis of the full-length extension products was consistent with the steady-state results (ϳ95% C and ϳ5% A incorporation). Presteady-state analysis of WT Dpo4 and mutant-catalyzed insertion of dATP opposite 8-oxoG was performed ( Fig. 3 and Table  3). A useful comparison can be made by dividing the catalytic efficiency of dCTP incorporation by the efficiency of dATP incorporation ([k pol /K D,dCTP ]/[k pol /K D,dATP ]). This ratio effectively measures the kinetic preference of dCTP over dATP in the concentration range of nucleotides used here ( Table 4). The kinetic parameters indicate that WT Dpo4 was the most accurate enzyme tested, although Glu 332 also maintains a near wild type preference for dCTP (17-fold; Table 4). The remaining mutants all exhibited decreased substrate selectivity opposite 8-oxoG, indicating some disruption to the enzymatic properties that define high-fidelity bypass of 8-oxoG. Neither the Ala 332 nor the Leu 332 mutant is capable of forming a hydrogen bond with 8-oxoG. The His 332 residue is apparently ineffective at forming a hydrogen bond with 8-oxoG at pH 7.4 (the pK a of the N-3 atom on the imidazole ring is presumably ϳ6.0), consistent with the view that hydrogen bonding between Arg 332 and the O-8 atom of 8-oxoG is important for accurate and efficient bypass of the lesion. The loss of fidelity observed with Ala 332 is driven by a much tighter binding of dATP relative to that observed with WT Dpo4 (Table 3), indicating that the Hoogsteen pair is better accommodated by the mutant. Likewise, the Leu 332 and His 332 mutants had lower K D,dATP values than WT Dpo4 but not as low as Ala 332 .
Of all of the enzymes tested, only the Ala 331 Ala 332 double mutant failed to incorporate dATP opposite 8-oxoG in the presteady-state experiments. Steadystate experiments revealed that the Ala 331 Ala 332 double mutant inserted dCTP opposite 8-oxoG with an ϳ200-fold greater efficiency than it did dATP (supplemental Table S6). The overall steady-state efficiency for dCTP insertion opposite 8-oxoG was decreased ϳ12-fold relative to WT Dpo4, and the steady-state efficiency of dATP incorporation was decreased ϳ27fold relative to WT Dpo4.

JOURNAL OF BIOLOGICAL CHEMISTRY 19837
Mutant Dpo4-catalyzed Extension beyond 8-OxoG Paired with C or A-The next-base extension of 8-oxoG:C and 8-oxoG:A pairs was measured by performing pre-steady-state experiments at a high concentration of the incoming dGTP (Fig. 4). All of the single mutants extended the 8-oxoG:C and 8-oxoG:A pairs in a similar manner, suggesting that next-base extension was relatively unaffected by the identity of the pair being extended. The rate of extension of the 8-oxoG:C pair was decreased for all four mutants relative to the k pol value for dCTP incorporation opposite 8-oxoG ( Fig. 4 compared with Table 3), indicating that some inhibition occurred when 8-oxoG:C entered the post-insertion site. Mutant-catalyzed extension of the 8-oxoG:A pair proceeded at rates similar to those measured for insertion of dATP opposite 8-oxoG ( Fig. 4 compared with Table 3). The Ala 331 Ala 332 double mutant extended the C:8-oxoG pair, but the rate and the amplitude of product formation were both decreased. The double mutant extended the A:8-oxoG pair, but the rate was very slow.
R332A Mutant⅐DNA Complex Structures-The structures of the Ala 332 mutant were determined in complex with a DNA duplex (18-mer template and 14-mer primer) containing either an 8-oxoG:A or an 8-oxoG:C base pair and an incoming dGTP at the active site (termed R332A(8-oxoG:A) and R332A(8-oxoG: C), respectively; Table 5 and supplemental Fig. S31). Both the R332A(8-oxoG:A) and the R332A(8-oxoG:C) structure represent type I complexes (11). The template C located 5Ј to 8-oxoG is accommodated inside the active site of both the R33A(8-oxoG:A) and the R332A(8-oxoG:C) structures and pairs with dGTP at the replication site (Fig. 5, A and B). The 8-oxoG pairs with the 3Ј-terminal base of the primer at the post-insertion (Ϫ1) site. The short side chain of Ala 332 allows considerable space for accommodating 8-oxo-G:C or 8-oxoG:A pairs (Fig. 6,  A and B). 8-oxoG is in a syn conformation in the R332A(8-oxoG:A) complex and forms a Hoogsteen pair with A. In this complex the O-8 atom forms a water-mediated hydrogen bond with Tyr 12 (Fig. 5A). The R332A(8-oxoG:C) complex shows the 8-oxoG residue in an anti conformation and forming a Watson-Crick paired with C. Lys 78 forms a water-mediated hydrogen bond to the N-2 of 8-oxoG (Fig. 5B). Such water-mediated hydrogen bonding interactions involving Tyr 12 or Lys 78 were TABLE 5 Crystal data and refinement parameters for the ternary (protein⅐DNA⅐dGTP) complexes of Dpo4 mutants APS, Advanced Photon Source; DND-CAT, DuPont-Northwestern-Dow Collaborative Access Team; r.m.s.d., root mean square deviation; c-v, cross-validation.
not observed in the WT Dpo4 complexes (Fig. 5, E and F) (20). There, the 8-oxoG pairs with the dNTP at the replicative position. The Tyr 12 and Lys 78 side chains are directed toward the base pair in the Ϫ1 position both in the WT Dpo4 and mutant structures (Fig. 5). Therefore, the presence of the water-mediated hydrogen bonds observed in case of the R332A mutant structures suggests that these residues may play a role in translocation of the primer-template duplex and may not be critical with regard to the insertion event.
R332E Mutant⅐DNA Complex Structures-The Glu 332 mutant protein was complexed with an 18-mer template-14mer primer DNA duplex containing either 8-oxoG:A or 8-oxoG:C pair at the active site and an incoming dGTP (termed R332E(8-oxoG:A) and R332E(8-oxoG:C), respectively; supplemental Fig. S31). As with the complexes of Dpo4 alanine mutants above, the template C located 5Ј to 8-oxoG pairs with dGTP and 8-oxoG pairs with the 3Ј-terminal base of the primer at the Ϫ1 site. 8-oxoG pairs in the Hoogsteen mode with A in the R332E(8-oxoG:A) structure (Fig. 5C). Here, Glu 332 engages in a direct but relatively long hydrogen bond (3.46 Å) as well as a water-mediated interaction with the exocyclic amino group N-2 of 8-oxoG. The R332E(8-oxoG:C) structure reveals a 8-oxoG:C pair in a Watson-Crick configuration with the O-8 oxygen linked to Glu 332 via two water molecules (Fig. 5D). The side chain of Glu 332 fills most of the space near the 8-oxoG base (Fig. 6, C and D). The available space is thus clearly reduced compared with the Ala 332 mutant structures but still is more open than in the WT Dpo4 structures. The presence of the long side chain of Arg 332 in WT Dpo4 (Fig. 6, E and F) may influence the choice of the inserted nucleotide. Steric hindrance and resulting repulsive interactions may take place in the case of the wild type 8-oxoG:dATP complex, leading to a destabilization of the Hoogsteen pair between 8-oxoG and dATP. Conversely, formation of a hydrogen bond between Arg 332 and O-8 can be expected to stabilize the Watson-Crick 8-oxoG: dCTP pair in the wild type 8-oxoG: dCTP complex (20), thus providing a rationalization of the preferred incorporation of C opposite 8-oxoG by Dpo4.

DISCUSSION
Mechanisms of cellular dysfunction, including carcinogenesis and aging, are caused, in part, by covalent modification of DNA (13,32). 8-oxoG is an important lesion to consider because it is widely prevalent and known to be mutagenic. Many of the DNA polymerases studied in vitro incorporate a large fraction of A opposite 8-oxoG relative to C (33)(34)(35)(36)(37)(38)(39). Some polymerases, such as yeast pol (40,41) and RB69 (42), preferentially insert dCTP opposite 8-oxoG, but the ability of Dpo4 to bypass 8-oxoG in a manner that is not only accurate but also more efficient than catalysis opposite unmodified DNA makes it unique among the DNA polymerases studied to date (20). Previous work suggested that a hydrogen bond between Arg 332 and the O-8 atom of 8-oxoG facilitates the increased efficiency of Dpo4 catalysis (20,21). The role of Arg 332 in facilitating bypass efficiency was examined by studying the structure and mechanism of four mutant enzymes.
There are two major points to consider when discussing Dpo4-catalyzed bypass of 8-oxoG. First, the accuracy (or fidelity) of the reaction is high. The role of Arg 332 may be to help stabilize the anti conformation of the purine ring system of 8-oxoG, which would favor the 8-oxoG:C Watson-Crick-like pair over the 8-oxoG:A Hoogsteen pair. The second major point to consider is the matter of enzyme efficiency. Does Arg 332 play a predominant role in increasing the efficiency of 8-oxoG bypass, or is efficiency determined by several interactions between Dpo4 and the template DNA?
On the issue of fidelity, changing the identity of the Arg 332 residue does not result in obvious changes to Dpo4 catalysis opposite unmodified DNA ( Fig. 1 and LC-MS data not shown). Relatively subtle changes in fidelity were observed in LC-MS/MS analysis of the full-length extension products for 8-oxoG-modified DNA ( Table 2). The Glu 332 mutant exhibits fidelity that is very similar to WT Dpo4 (ϳ11% A incorporation). With the exception of the Ala 331 Ala 332 double mutant, the other mutants incorporated more A, indicating that some interaction was lost that moved the equilibrium between accurate and mutagenic bypass of 8-oxoG toward incorporation of A. The kinetic parameters are consistent with the LC-MS/MS results. The predicted trend in fidelity of 8-oxoG bypass is as follows: WT Dpo4 Ӎ Ala 331 Ala 332 (highest C incorporation) Ͼ Glu 332 Ͼ Leu 332 Ͼ Ala 332 Ӎ His 332 (lowest C incorporation). One interesting comparison here is between the apparent preference, as determined by kinetic efficiency, and the products identified by LC-MS/MS. The kinetic parameters indicate that WT Dpo4 and the Glu 332 would insert ϳ5-6% A compared with ϳ15% A for Leu 332 and ϳ27% A for the Ala 332 and His 332 mutants. Yet, the full-length extension products differ by only ϳ4%. When considered alone, the kinetic parameters suggest an important role for hydrogen bonding capability at position 332, but when considered in the context of both kinetics and the LC-MS/MS results, a less than definitive role for a hydrogen bond between residue 332 and the O-8 atom of 8-oxoG emerges in the determination of Dpo4 fidelity opposite the lesion. The differences highlighted by the kinetic results may be diminished in the LC-MS/MS results, because the fulllength extension assays are performed under conditions that allow multiple catalytic turnovers, which may reflect events that are not measured during single nucleotide incorporation (i.e. product dissociation, translocation, and/or nextbase extension).
Regarding the fidelity of 8-oxoG bypass, the Ala 331 Ala 332 mutant presents some interesting features. Both kinetic and LC-MS/MS analyses indicate that the double mutant has even higher fidelity than WT Dpo4. However, it is also apparent that the processivity of the double mutant is severely impeded, most likely at the step following dNTP insertion opposite 8-oxoG. A crystal structure was solved of the double mutant in complex with the C:8-oxoG pair, but the little finger domain is disordered, which makes it difficult to discern why the double mutant retains such high fidelity opposite 8-oxoG (data not shown). On the other hand, the full-length extension results with the Ala 331 Ala 332 double mutant clearly show an effect upon Dpo4 processivity. Indeed, all of the enzymes tested here are slower at next-base extension of the 8-oxoG:C pair, indicating the importance of positively charged residues in the little finger domain during the Dpo4 translocation step. Such a view is consistent with previous structural work showing that Arg 332 maintained contact with 8-oxoG even when the template base is shifted into the post-insertion site of a type II structure (20).
The second major point to consider is the efficiency of Dpo4catalyzed bypass of 8-oxoG. In this regard, the Glu 332 mutant is most similar to WT Dpo4. The Glu 332 mutant is 5-fold more efficient at inserting dCTP opposite 8-oxoG compared with G. The Leu 332 mutant is 2-fold more efficient inserting dCTP opposite 8-oxoG compared with G, but the Ala 332 and His 332 mutants are slightly more efficient at inserting dCTP opposite G. With the exception of Ala 332 , which inserts dCTP opposite G with ϳ20-fold greater efficiency than WT Dpo4 (Table 4), the gap between efficiency of G and 8-oxoG bypass is caused primarily by a decrease in bypass efficiency opposite 8-oxoG (Table 4). As with the issue of fidelity, the differences in efficiency are not large, but the Glu 332 mutant is the only one of these mutants capable of effectively forming a hydrogen bond with 8-oxoG, and it is the most similar to WT Dpo4 in kinetic terms.
The fact that the Glu 332 mutant is similar to WT Dpo4 contradicted our initial hypothesis that the negatively charged side chain of Glu 332 would move 8-oxoG into the Hoogsteen mode. The crystal structures reveal how the similarities between WT Dpo4 and Glu 332 are maintained. A water-mediated hydrogen bond is formed between the carboxylic acid moiety of Glu 332 and the O-8 atom of 8-oxoG (Fig. 5C). The Glu 332 mutant also forms a hydrogen bond with the N-2 exocyclic amino group of 8-oxoG when the 8-oxoG:A Hoogsteen pair is formed (Fig. 5D). No such interaction is observed in the structure of WT Dpo4 and the 8-oxoG:A Hoogsteen pair (20), but it is unclear what effect a hydrogen bond between Glu 332 and the N-2 amino group of 8-oxoG has upon dATP insertion, as the efficiency is not increased relative to WT Dpo4 (Table 4).
An interesting difference between WT Dpo4 and the Ala 332 mutant is the increased proclivity to insert dATP, which reduces the fidelity of the Ala 332 mutant. The Ala 332 mutant crystal structures (Fig. 6, A and B) reveal an empty space in the region normally occupied by Arg 332 in WT Dpo4. The Arg 331 residue maintains contact with the template DNA to the 5Ј-side of the 8-oxoG lesion and does not change conformation to replace Arg 332 . The increased ability of the Ala 332 mutant to insert dATP opposite 8-oxoG may be related to the syn/anti equilibrium in the Dpo4 active site. A larger space could be more accommodating to the 8-oxoG:A Hoogsteen pair, as evidenced by the relatively low K D,dATP measured for Ala 332 insertion of dATP opposite 8-oxoG (Table 3).
In comparing the mechanism for Dpo4-catalyzed bypass of 8-oxoG with other DNA polymerases, an important similarity to bacteriophage pol T7 Ϫ is observed. The crystal structure of pol T7 Ϫ in ternary complex with a 8-oxoG: ddCTP pair revealed that Lys 536 is in position to form a hydrogen bond with the O-8 atom of 8-oxoG (33). The side chain of Lys 536 moves 3 Å relative to the position observed in a superimposed structure of pol T7 Ϫ bound to unmodified DNA. The stabilization of 8-oxoG by a hydrogen bond bears obvious resemblance to Dpo4, but it is estimated that pol T7 inserts A opposite 8-oxoG in ϳ30% of incorporation events, even when exonuclease activity is present (33). Insertion of dCTP opposite 8-oxoG by pol T7 Ϫ is inhibited ϳ180-fold relative to insertion of dCTP opposite G (comparing presteady-state data) (36). Both the level of dATP incorporation and the catalytic inhibition of pol T7 Ϫ is in direct contrast to what has been observed with Dpo4. The stabilization of 8-oxoG by Lys 536 in the pol T7 Ϫ structure is apparently not substantial enough to overcome other factors, i.e. a geometrically intolerant active site and kinking of the template DNA backbone, to facilitate high efficiency and high fidelity during T7 Ϫ -catalyzed bypass of 8-oxoG.
A sequence alignment of Dpo4 with Saccharomyces cerevisiae pol based on secondary structure predicts that a histidine residue should be found in the region occupied by Arg 332 (11). However, a structure-based alignment of the little finger domain from the two proteins suggests that Arg 332 is replaced by a lysine (Lys 498 ) in S. cerevisiae pol (Fig. 7). The results presented here may be consistent with our structure-based alignment, because the His 332 mutant has the lowest fidelity of all of the mutants tested here, whereas yeast pol is known to bypass 8-oxoG with relatively high fidelity (40). If our structurebased alignment is correct then the electrostatic interaction between the little finger domain and 8-oxoG may also be important for high-fidelity translesion synthesis opposite 8-oxoG by pol .