Human DNA Polymerase ν Catalyzes Correct and Incorrect DNA Synthesis with High Catalytic Efficiency*

Background: DNA polymerase ν is a low fidelity polymerase with an unknown function. Results: The kinetic mechanism of correct and incorrect base pair formation was determined. Conclusion: Phosphodiester bond formation is rapid for both correct and mispair formation. Significance: The presence of a mispair does not induce the polymerase to adopt a low catalytic conformation. DNA polymerase ν (pol ν) is a low fidelity A-family polymerase with a putative role in interstrand cross-link repair and homologous recombination. We carried out pre-steady-state kinetic analysis to elucidate the kinetic mechanism of this enzyme. We found that the mechanism consists of seven steps, similar that of other A-family polymerases. pol ν binds to DNA with a Kd for DNA of 9.2 nm, with an off-rate constant of 0.013 s−1and an on-rate constant of 14 μm−1 s−1. dNTP binding is rapid with Kd values of 20 and 476 μm for the correct and incorrect dNTP, respectively. Pyrophosphorylation occurs with a Kd value for PPi of 3.7 mm and a maximal rate constant of 11 s−1. Pre-steady-state kinetics, examination of the elemental effect using dNTPαS, and pulse-chase experiments indicate that a rapid phosphodiester bond formation step is flanked by slow conformational changes for both correct and incorrect base pair formation. These experiments in combination with computer simulations indicate that the first conformational change occurs with rate constants of 75 and 20 s−1; rapid phosphodiester bond formation occurs with a Keq of 2.2 and 1.7, and the second conformational change occurs with rate constants of 2.1 and 0.5 s−1, for correct and incorrect base pair formation, respectively. The presence of a mispair does not induce the polymerase to adopt a low catalytic conformation. pol ν catalyzes both correct and mispair formation with high catalytic efficiency.

DNA polymerase (pol 2 ) is one of the 17 DNA polymerases encoded by the human genome (1). Although the roles of many of these polymerases are known (2)(3)(4), the function of pol has yet to be elucidated. There is evidence that pol functions in cross-link repair and/or homologous recombination. For instance, mutations in mus308, the Drosophila ortholog of pol , cause hypersensitivity to cross-linking agents (nitrogen mustard and cisplatin) but not methyl methanesulfonate (5,6). In human cells, siRNA knockdown of pol also causes sensitivity to cross-linking agents and impairs homologous recombination (7,8). Consistent with a direct role in the repair of these crosslinking agents, pol can bypass a psoralen-DNA intrastrand cross-link and DNA-DNA cross-links with linkages in the major groove of the DNA (8,9).
DNA polymerases have been categorized into families based upon sequence homology. pol as well as pols ␥ and are human A-family polymerases. Other members of this family that have been extensively studied are DNA polymerase I of Escherichia coli (Klenow fragment, KF) (10 -13), Thermus aquaticus (Taq pol) (14,15), Bacillus stearothermophilus (BF) (16,17), and bacteriophage T7 (T7 pol) (18,19), as well as pol ␥ (20 -22). Although the A-family polymerases have sequence similarities, differences in sequence provide specialized functions. For instance, E. coli pol I is a high fidelity polymerase with a 3Ј-to 5Ј-exonuclease proofreading activity as well as a 5Ј-to 3Ј-exonuclease activity. These activities let E. coli pol I function in Okazaki fragment synthesis as well as DNA repair in E. coli. pol ␥ is a multisubunit protein with high fidelity polymerase, 3Јto 5Ј-exonuclease proofreading, and 5Ј-deoxyribose phosphate lyase activities that enable pol ␥ to be the sole polymerase responsible for the replication, recombination, and repair of mitochondrial DNA (23).
pol is a 900-amino acid protein, in which the C-terminal 419 residues are highly related to the polymerase domains of A-family polymerases (1). pol lacks the helicase, exonuclease, and lyase activities of other A-family polymerases. It exhibits low fidelity, with base pair substitution and deletion rates similar to Y-family polymerases (24). The polymerase is especially unfaithful with a high frequency for the formation of dG/dT mismatch formation (24,25). Pre-steady-state analysis showed that the lower fidelity of pol compared with that of KF(exoϪ) is due to lower k pol parameters for correct dNTP incorporation (26).
The complete kinetic profiles of E. coli pol I (10 -13, 27) and bacteriophage T7 polymerase (18,19) have been determined. The mechanism consists of seven steps shown in Fig. 1A, in which phosphodiester bond formation, step 4, is flanked by conformational changes that are rate-limiting. Upon binding the correct dNTP, the polymerase is able to form an active conformation that catalyzes phosphodiester bond formation rapidly. When an incorrect dNTP binds, the enzyme is unable to adopt the highly catalytic conformation, and phosphodiester bond formation becomes rate-limiting (11,12,18,28). The kinetic schemes of the low fidelity Dpo4 and pol also consist of those seven steps, with the difference that the rate phosphodiester bond formation is similar to those of the conformational changes (29,30). In this study, we evaluated the kinetic mechanism of pol and found that similar to KF and T7 pol, pol catalyzes correct base pair formation with a rapid phosphodiester bond formation step. In contrast to its high fidelity cousins, pol catalyzes mispair formation via the same mechanism. Despite the binding of the wrong dNTP, the enzyme is able to form a high catalytic conformation.

Experimental Procedures
General-[␥-32 P]ATP, [␣-32 P]dCTP, and [␣-32 P]dATP were purchased from PerkinElmer Life Sciences, and T4 polynucleotide kinase was from USB/Affymetrix. The dNTPs (ultrapure grade) were purchased from GE Healthcare, and the concentrations were determined by UV absorbance (31). Oligodeoxynucleotides were synthesized at the Macromolecular Core Facility at the Pennsylvania State University College of Medicine. The sequences of the oligodeoxynucleotides are shown in Table 1. The primer was 32 P-labeled with [␥-32 P]ATP and annealed with a 20% excess of the template as described previously (32). The P15/T24 substrate was used in all except the pyrophosphorylation experiments and is referred to as the DNA substrate.
pol was purified by a modification of the method described by Takata et al. (24). The protein has a deletion of a C-terminal polyproline segment, an N-terminal His tag, and C-terminal FLAG tag. After lysing the cells, the resulting supernatant was loaded onto a nickel-nitrilotriacetic acid column (3 ml), and the enzyme was eluted with an imidazole gradient. Purity was evaluated by SDS-PAGE, and the protein concentration was determined by the Qubit protein assay kit (Life Technologies, Inc.).
Kinetic Analyses-All assays with pol were performed in 40 mM Tris-HCl (pH 8.0), 5 mM MgCl 2 at 37°C. The concentrations reported for each species are those that occurred during the reaction. The reactions were initiated by the addition of dNTP containing MgCl 2 to preincubated protein and DNA in buffer. Steady-state reactions were quenched with equal volumes of STOP solution containing 10% 0.5 M Na 2 EDTA, 90% formamide, and 0.025% (w/v) xylene cyanol and bromphenol blue. Fast reactions were carried out with an RQF-3 quench flow apparatus (KinTek) and were quenched with 0.3 M Na 2 EDTA, which was diluted with STOP solution to load onto the gel. The progress of the reaction was analyzed by denaturing PAGE, and the radioactivity on the gel was visualized with a Typhoon 9200 Variable Mode Imager (GE Healthcare). The reactions were quantitated by dividing the total radioactivity in the product band(s) by the radioactivity in the product and reactant bands.
Equilibrium Polymerase-DNA Binding Experiments-The binding affinity of the DNA to the polymerase was evaluated with fluorescence anisotropy. The DNA (1 nM) consisted of the primer strand modified on the 5Ј-end with 6-carboxyfluorescein and the template strand containing dG. The fluorescence anisotropy was measured with 0 -100 nM pol at 37°C using a Tecan Safire 2 plate reader with half-volume 96-well plates. The excitation and emission wavelengths were 471 and 525 nm, respectively, with an emission bandwidth of 20 nm.
Data Analysis-The kinetic constants were fit using GraphPad Prism Version 6. Computer simulations were performed with DynaFit Version 4 (BioKin Ltd.) (33).

Results
Purification and Concentration of pol -The His tag pol was purified with an imidazole gradient on an nickel-nitrilotriacetic acid column. From a 2-liter incubation, we obtained 2 ml of 990 nM protein. The active site concentration of pol was determined by measuring the burst amplitude of a reaction containing an excess of DNA and a saturating concentration of dNTP. As shown in Fig. 2, the reaction could be fit to Equation 1, where P is the concentration of the products; A is the burst amplitude; k f is the rate constant of the fast phase, and k s the steady-state rate constant. The burst amplitude of 6 Ϯ 2 nM agreed with the protein concentration of 6.6 nM determined spectrophotometrically. The burst rate constant was Ͼ100 s Ϫ1 indicating that the polymerase has high catalytic activity.   Table 2. The DNA sequence was chosen because we previously used this sequence to examine the reactivity of KF (32, 34 -37). dCTP was incorporated opposite dG with k cat /K m values 5-fold greater than dTTP. The purine triphosphates were incorporated about 10-fold less readily. The differences in k cat /K m values are driven by both the k cat and K m parameters. These numbers are consistent with the reports that pol is a low fidelity enzyme that is especially poor at discrimination against the incorporation of dT opposite dG in select sequences.
Pre-steady-state Incorporation of dNTPs-Steady-state kinetic analysis does not provide insights into intermediates in the reaction pathway of enzyme-catalyzed reactions. This is especially true of DNA polymerases because under conditions in which the DNA is in excess over the polymerase, the rate-limiting step is the dissociation of the DNA from the polymerase. Thus, to more fully explore the mechanisms by which pol catalyzes DNA replication, we performed pre-steady-state kinetic experiments. We carried out single-turnover reactions by preincubating 15 nM DNA with 150 nM polymerase and initiating the reaction by the addition of dNTP and 5 mM MgCl 2 . The resulting time courses at different dNTP concentrations are shown in Fig. 3. The incorporation of dCTP ( Fig. 3B) is much more rapid than the incorrect dNTPs. For example, at a dNTP concentration of 100 M, the half-life for the reaction of dCTP was 0.013 s, whereas that for incorrect dNTPs was ϳ6.5 s, a 500-fold dif-   Table 5 for A and B. For C and D, the lines are based upon the parameters of dATP with the following changes: C, K d for dGTP ϭ 1.5 mM; ference. This discrimination differs from the relative k cat /K m ratios that range from 5 to 50. In these single-turnover reactions, we do not observed a preference for dTTP incorporation over dATP or dGTP. The time courses exhibited biphasic kinetics in which a rapid reaction was followed by a slower reaction. The time courses were initially fit to Equation 1. The amplitude (A) and burst rate constant k were dependent on the dNTP concentrations, whereas the steady-state rate constant (k s ) was not. The resulting amplitudes and burst rate constants are plotted against dNTP concentration in Fig. 4. These values were fitted to a hyperbola (Equation 2) in which V is the value of the parameter at a specific [dNTP]. From these plots, we obtained the following: A max , the maximum amplitude; K A , the [dNTP] at 50% of A max ; k pol , the maximum burst rate constant and the apparent dissociation constant of the dNTP, and K d app for dNTP. These parameters as well as the k s from the initial fitting are presented in Table 3. The k pol for the incorporation of the correct dCTP is 100 Ϯ 8 s Ϫ1 . This value is comparable with the values obtained from KF and T7 pol (11,18) and is very similar to the k pol of 155 s Ϫ1 for KF(exoϪ) with the same DNA substrate (37). This value is larger than the k pol found for the Y-family polymerases and but similar to that of pol (30,38,39).
The k pol values for the incorrect dNTPs were 5-fold lower, ranging from 14 to 22 s Ϫ1 . These values are much larger than those for KF and T7 pol (12,18,28,37). The K d app value for dNTP was lower for the correct dCTP (4.3 M) than for the incorrect dNTPs (140 -290 M) by a factor of 30 -70. The apparent K d app values for dNTP for the correct and incorrect dNTPs are similar to those found for KF and T7 pol. The burst amplitude was also dependent on the dNTP concentration. The maximum amplitude (A max ) for all four dNTPs were similar, ranging from 10 to 13 nM. The K A value is lower for the correct dCTP (20 M) that the incorrect dNTPs (140 -440 M). In summary, dCTP is favored over the incorrect dNTPs by a higher k pol , and lower K d app for dNTP and K A values. In contrast to the steady-state experiments, we do not observe a preference of dTTP over dATP and dGTP.
The dNTP concentration dependence on the burst amplitude was not observed for KF and T7 pol. In this respect, pol is unique in A-family polymerases and indicates a difference in mechanism. In subsequent experiments, we considered two possible mechanisms illustrated in Fig. 1, A and B, to explain the [dNTP]-dependent decrease in burst amplitude. Fig. 1A is the seven-step mechanism that has been found for KF and T7 pol. It consists of conformational changes prior to and after phosphodiester bond formation. This mechanism would produce a dNTP-dependent burst if step 5 was rate-limiting. The mechanism in Fig. 1B contains an inactive complex and would produce a decreased burst if the inactive complex is stable such that k Ϫ8 or k Ϫ9 values are slower than k pol. The presence of inactive complexes has been observed with DNA adducts (40,41) and with pol with correct base pairs (42).
pol is unique in its propensity to incorporate dT opposite dG in some sequences (1,25,26,43). Our DNA sequence, which showed a preference for dTTP incorporation under steadystate conditions, did not show any preference under singleturnover conditions.
DNA Binding Affinity-The binding affinity of the DNA to pol was measured by two methods, fluorescence anisotropy and active site titration. The fluorescence intensity did not change when polymerase was added (data not shown), consistent with the assumption that the fluorescein does not interact with the polymerase and change the binding affinity. The increase in anisotropy was plotted against polymerase concentration in Fig. 5A. The data were fitted to Equation 2, in which V is the increase in anisotropy, and [N] is polymerase concentration. The K d for DNA was determined to be 6.5 Ϯ 1.8 nM.  The binding affinity was also measured by active site titration. In these experiments, pol was reacted with increasing amounts of DNA. The reaction was quenched after 100 ms, enough time for the burst phase of the reaction to be complete. The amount of product was plotted against DNA concentration, and the K d for DNA and polymerase concentration (E) determined by fitting the data to Equation 3, where A is the concentration of product; D is the DNA, and K is the K d for DNA. By this procedure, the K d for DNA was determined 9.2 Ϯ 1.3 nM, consistent with the fluorescence aniso tropy method. The enzyme concentration (E) was determined to be 22 Ϯ 2 nM, consistent with the protein concentration determined in Fig. 1.
Dissociation Rate of the DNA-The dissociation rate of DNA and pol was determined by the procedure illustrated in Fig.   6A. pol (25 nM) was preincubated with 32 P-labeled P15/T24 (50 nM) in which the primer strand is 5Ј-32 P-labeled. The reaction was initiated by the addition of an excess of unlabeled trap DNA (P15/T24). The amount of 32 P-DNA still bound to the polymerase was assayed at various times by the addition of dCTP, which will react to form the 16-mer product. At time increases, the amount of 16-mer becomes less, as the 32 P-labeled P15/T24 substrate dissociates from the enzyme. The dissociation of the [ 32 P]DNA is essentially irreversible because of the large excess of unlabeled DNA. The data are presented in Fig. 6B. The solid line in Fig. 6B is the best fit to the single exponential in Equation 4, where P is the 16-mer concentration; A is the product concentration in the absence of the DNA trap; C is the product concentration at infinite time, and k the rate constant for the dissociation of the DNA from the protein, which was found to equal 0.013 Ϯ 0.001 s Ϫ1 . This value, in combination with a K d for DNA of 9.2 nM, leads to a DNA on rate constant of 1.4 M Ϫ1 s Ϫ1 .
The mechanism in Fig. 1A predicts a first-order dissociation, whereas that in Fig. 1B would be biphasic, depending on the relative rate constants. Fig. 1B can account for the biphasic nature of the time course if conversion of the inactive complexes E X ⅐D n and E X ⅐D n ⅐dNTP to E⅐D n and E⅐D n ⅐dNTP were slow. Under this scenario the initial preincubation of enzyme and DNA would form both E⅐D n and E X ⅐D n . Upon addition of dNTP, the fast phase would be the reaction of ED to product, whereas the slow phase would be the conversion of E X ⅐D n to E⅐D n . However, the dissociation rate of 0.013 s Ϫ1 is slower than the slow phase in the reaction, and thus we cannot test this mechanism with this DNA dissociation mechanism.
Dissociation Rate of the dNTP-The off-rate of the dNTP is rapid for all polymerases examined thus far. To test whether this is also true for pol , we carried out trapping experiments described in Fig. 7. The ratio of the product of the trapped versus nontrap experiment should equal k pol /(k off ϩ k pol ), in which k pol is the forward rate to product formation, and k off is the dNTP dissociation rate constant as shown in Fig. 1C. The nontrapped experiment was conducted as shown in Fig. 7A (F) in which the reaction of preincubated pol (60 nM), DNA (300 nM), and [␣-32 P]dCTP (20 M) was initiated by the addition of Mg 2ϩ . The data in Fig. 7B was fit to the burst equation  The difference between the trapped experiment (Ⅺ) and its control (OE) is shown in Fig. 7D, in which the error bars represent the 95% confidence intervals of the difference. Thus, the amplitude of the product resulting from the ternary complex is 0.8 nM versus the untrapped amplitude of 32 nM. This result is consistent with a rapid dissociation of dCTP from the ternary complex. The ratio of the products at the various time points is shown in Fig. 7E. During the first 0.25 s, the ratio is ϳ20. Because we found k pol ϭ 100 s Ϫ1 , the off-rate of the dNTP (k off ) is very fast, with a rate constant of ϳ2,000 s Ϫ1 . With a K d app for dNTP of 20 M, we estimate the on-rate constant to be 10 M Ϫ1 s Ϫ1 .
Pyrophosphorylation-The DNA polymerase-catalyzed reaction is reversible. Addition of pyrophosphate results in the successive removal of the terminal dNMP. The time course for pyrophosphorylation of the P16C/T24 DNA substrate at different concentrations of pyrophosphate is shown in Fig. 8A. The product in the y axis is the total of the 13-, 14-, and 15-mer that is formed during the reaction. The time courses can be fit to a first-order equation. The resulting rate constant can be fit to a hyperbola with a maximum rate of 11 Ϯ 2 s Ϫ1 , and a K d for PP i ϭ 3.7 Ϯ 1.9 mM (Fig. 8B). We also examined the rate of phosphorylation with a terminal mispair using P16T/T24. The reaction, containing 280 nM pol , 40 nM DNA, 10 mM PP i , and 20 mM MgCl 2 , shown in Fig. 8C demonstrates that the presence of a mispair does not affect pyrophosphorylation.
Elemental Sulfur Effect-The rate-limiting step was investigated using deoxynucleotide triphosphates in which a nonbridging oxygen on the ␣-phosphate is replaced by sulfur (dNTP(␣S)). As shown in Fig. 9A, the reactions with 5 M dCTP or dCTP␣S exhibited similar kinetics. The reactions were fit to the burst equation with the parameters shown in Table 4. The kinetic parameters were virtually identical. This result is consistent with those for KF and T7 pol in which phosphodiester bond formation is rapid and such that the ϳ100-fold reduction in rate of this step did not affect the rate of phosphodiester bond formation. The elemental sulfur effect was also examined for the misincorporation of dATP opposite dG. The time course profiles shown in Fig. 9B and parameters in Table 4 again demonstrate that the sulfur substitution had no effect on the reaction kinetics. This result suggests that phosphodiester bond formation is more rapid than the conformational change steps, a result very different from those obtained for KF and T7 pol, which had elemental effects of 65 for KF (12) and 13-37 for T7 pol (19) during incorrect base pair formation. These results indicate that phosphodiester bond formation is rapid for both correct and incorrect base pair formation.
Pulse-Chase Experiments-The pulse-chase experiment was performed to investigate the possibility of kinetic steps that precede and/or come after phosphodiester bond formation. In these experiments, pol /DNA was reacted with [␣-32 P]dCTP for various times. Fig. 10A shows the 16-mer product formed after quenching with 1 N HCl (E) versus the pulse-chase (F) in which the reaction was quenched with 1 mM unlabeled dCTP for 30 s followed by 1 N HCl. The acid quenches all the enzymebound species, resulting in only the D n ϩ 1 species being recognized as product. When the reaction is chased with cold dCTP, an increase in product is observed, due to the forward reaction of D n -containing species. These species cannot be E⅐D n , because it will bind to unlabeled dCTP, or E⅐D n ⅐dNTP because the off-rate of the dNTP, (k Ϫ2 ) is much greater than k pol . Thus, these experiments support the existence of E*⅐D n ⅐dNTP. In addition, the observation that the pulse-chase continuously   forms a greater amount of product is consistent with a mechanism in which a kinetic roadblock follows the product formation step that was observed for KF (13). If k 5 was large, we would observe a transient increase in pulse-chased product as was observed with T7 pol (18). This observation supports the exis tence of E*⅐D n ϩ 1 ⅐PP i . For dCTP, the amplitude of the pulsequench is 69% that of the pulse-chase. Because step 5 is slow, we can estimate that E*⅐D n ϩ 1 ⅐PP i comprises the bulk of the D n ϩ 1 species. Thus, the ratio of (E*⅐D n ϩ 1 ⅐PP i ) to (E*⅐D n ⅐dNTP) equals 0.69/0.31 and thus K 4 ϭ 2.2. For dATP, the burst amplitude of the quench-chase experiment is 57% that of the pulsechase experiment. By a similar analysis the K 4 ϭ 1.3.
Computer Simulation-The kinetic scheme for the pol -catalyzed incorporation of dCTP opposite dG with the DNA substrate P15/T24 is consistent with the mechanism in Fig. 1A and the parameters in Table 5. For the simulation, several of the rate constants were set based upon the experiments described above. The rate constants for the association (1.4 M Ϫ1 s Ϫ1 ) and dissociation (0.013 s Ϫ1 ) of the DNA and polymerase were set by the DNA affinity experiments in Fig. 5 and the DNA dissociation rate experiments in Fig. 6. These values were obtained for the P15/T24, and we assumed that the same values held for the product P16/T24. The off-rate constant for the dNTP (k Ϫ2 ) was set to 2000 s Ϫ1 based upon the experiments shown in Fig. 7. The equilibrium constant for phosphodiester bond formation (k 4 /k Ϫ4 ) was set to 2.2 and 1.3 for dCTP and dATP, respectively, based upon the pulse-chase experiments in Fig. 10. The phosphodiester bond formation step is rapid, compared with step 3, because the incorporation of dCTP or dATP did not exhibit an elemental sulfur effect as shown in Fig. 9. The value for k Ϫ5 was set at 11 s Ϫ1 because this is the k pyro obtained in Fig. 8. The K d for PP i was set to be 3.7 mM based upon the experiments in Fig. 8. With these constants, the parameters k Ϫ2, k 3 , k Ϫ3 , and k 5 were fit by computer simulation. Initial estimates of k Ϫ2 was based upon the initial estimates for the K d app for dNTP parameters in Table 3. The initial estimates for k 3 and k 5 was based upon the maximal k pol and k s values from Table 3. Based upon these values, the parameters converged to those listed in Table 5. These calculated lines are shown in Fig. 3.

Discussion
pol catalyzes correct base pair formation with high efficiency, similar to KF and T7 pol. The k pol , K d for dNTP, and K d for DNA values are close to those found for T7 pol and KF (11,18). The low fidelity of pol is due to its ability to rapidly catalyze mispair formation; the k pol value decreased only 5-fold with a mispair. These results differs from those obtained with a 77-kDa construct of pol (pol -77) (26). The k pol of pol -77 was found to be Ͻ1 s Ϫ1 , although we found a k pol of 100 s Ϫ1 . This is also the construct that was recently crystallized (44). Our construct is that described by Takata et al. (24) in which a short C-terminal polyproline stretch is removed and FLAG and His tags were added. Interestingly, the k cat and K m values that we obtained are similar to those found for the pol -77 (44). We speculate that the excised N-terminal domain impacts the activity of the catalytic core.
pol catalyzes phosphodiester bond formation via the sevenstep kinetic scheme that was formulated for the A-family polymerase KF (11-13, 27, 45, 46) and T7 pol (18,19,28,47). The kinetic scheme for correct base pair formation by pol is very similar to that of KF and T7 pol, in that phosphodiester bond formation is much faster than the surrounding conformational changes. Although the major aspects of the kinetic mechanism are similar between pol and KF and T7 pol, there are small differences. pol is more like KF, with a rate-limiting step after phosphodiester bond formation and an equilibrium constant for phosphodiester bond formation (K 4 ) of 2.3 for pol and 4 for KF. In contrast, the rate-limiting step for T7 pol is the conformational change preceding bond formation (step 3), whereas K 4 ϭ 0.5. pol differs from KF and T7 pol in that K 3 ϭ 0.75, although it is 10 for KF (11,12) and 3 for T7 pol (18).
The major difference between pol and its higher fidelity cousins occurs during mispair formation. The conformational change preceding bond formation and the bond formation step significantly decrease in rate with KF and T7 pol. However, with pol , phosphodiester bond formation remains rapid, although K 4 decreases from 2.3 to 1.3. Small decreases are also associated with the conformational changes; k 3 decreased from 75 to 13 s Ϫ1 , and k 5 decreased from 2.1 to 0.3 s Ϫ1 . Another difference between Kf and T7 pol and pol is the relative roles that polymerization and dNTP binding play in fidelity. For KF and T7 pol, the Ͼ1000-fold reduction in k pol has more of an effect on fidelity than a 10-fold increase in K d for DNA. In contrast, the 22-fold increase in K d for dNTP plays a larger role in fidelity than the 5-7-fold decrease in k pol . pol is more like pol than pol . pol catalyzes phosphodiester bond formation very rapidly, with k pol values reported from 50 to 190 s Ϫ1 (48,49). The elemental sulfur effect is small for both correct (1.6) and incorrect (2.5) base pair formation (48) indicating that the conformational changes are the primary rate-limiting steps. Thus, the low fidelity of pol is due to a high rate of mispair formation. In contrast, k pol for pol is lower, less than 10 s Ϫ1 (39,50). The low fidelity is due to decreased correct base pair formation, rather than increased mispair formation. A similarity between pol and is the decreased burst amplitudes observed for correct base pair formation. However, the mechanisms that cause the decreased burst are different. pol forms a stable inactive complex that is kinetically detected by a biphasic dissociation of the DNA from the polymerase. With pol , we were not able to detect such a species. We have concluded that the decreased burst amplitude is due to K 3 Ͻ 1 and a ratelimiting step 5.
The high catalytic activity of pol is easy to rationalize based upon its sequence and structural similarities with KF, T7 pol, Taq pol, and BF. Simple homology modeling of the pol structure on those of BF, T7, and Taq pol shows that pol could potentially have identical interactions with the DNA. The recent pol crystal structure shows the active site structure of pol is very similar to that of the high fidelity A-family polymerases (44). Similarities include the binding surface between the polymerase and the DNA, with tight contact with minor groove and less contact with the major groove of the DNA. The amino acid residues that contact the terminal base pair and nascent base pair are totally conserved, as are the acidic residues that bind the critical Mg 2ϩ ions. Thus, the positioning of the nucleophilic 3Ј-hydroxyl on the DNA and the triphosphate is optimal for high catalytic activity.
Based upon the structural similarities between pol and KF, T7 pol, Taq pol, and BF, the low fidelity of pol is more difficult to understand. High fidelity polymerases are able to insert the correct versus incorrect dNTP at a ratio of 10 4 -10 6 to 1. A current hypothesis is that polymerases amplify the differences in energy and structure of correct versus mispairs to both enhance the rate of correct base pair formation and decrease the rate of mispair formation (51)(52)(53). During T7 pol-catalyzed mispair formation, the energy of the E*⅐D , ⅐dNTP complex containing the incorrect dNTP is much higher in energy than that for the correct dNTP. In addition, the energy of the phosphoryltransfer transition state is also drastically increased (28,54). Thus, incorrect dNTPs are rejected in a two-step process because of the high energies of the closed complex as well as the transition state. The structures of these complexes as well as the interactions that account for the differences in energies between correct and incorrect base pairs are key for elucidating the mechanisms that polymerases use for base discrimination.
Two potential mechanisms by which mispairs are formed include the following: 1) non-Watson-Crick geometry base pairs such as wobble or Hoogsteen base pairs, and 2) Watson-Crick geometry base pairs that occur via the minor tautomers or ionized bases. Evidence for both mechanisms is found in crystallography experiments. Johnson and Beese (17) captured mispaired oligodeoxynucleotide substrates bound to BF. These structures show that BF readily discriminates against mismatches, binding them in inactive conformations. Crystal structures also indicate that a dT/dG mispair forms first as a The kinetic constants represent the best fit of the kinetic and equilibrium data. The parameters in italics were determined by the experiments outlined in the text and set as constants. The other parameters were fit by DynaFit Version 4 using the time course parameters in Fig. 3. wobble base pair in the active site of a partially open BF (55), followed by formation of a closed polymerase conformation in which the base pair adopts Watson-Crick geometry in which one of the bases isomerizes to its minor tautomer (56). In contrast, structures of mispairs with bacteriophage RB69 DNA polymerase, a high fidelity B-family polymerase, show the binding of mispairs in non-Watson-Crick geometry (57).
Therefore, pol can potentially accelerate mispair formation in two ways as follows: 1) by providing interactions that stabilize the Watson-Crick geometry of mispairs via minor tauto mers or ionized bases, and/or 2) relaxing the geometric selection in the active site so that non-Watson-Crick structures can bind, although the polymerase adopts the catalytically active closed configuration. pol has several minor structural differences with KF, T7 pol, Taq pol, and BF. These include substitutions of two amino acid residues that contact the DNA as well as amino acid substitutions and insertions that are distally located throughout the polymerase domain. None of these changes lead to clear-cut mechanisms that explain why the incorporation of mispairs is not disfavored by pol . It is more likely that several of the differences contribute to the overall reactivity.
Although most of the amino acid substitutions are located distally from the DNA, two, Lys-679 and Glu-675, are located in the major groove of the DNA near the nascent base pair (44). Lys-679 is a Thr in T7 and Taq pol, and an Ala in KF and BF, and Glu-675 is an Arg in KF, T7 pol, Taq pol, and BF. The K679A mutation demonstrated that the lysine contributes to the propensity to incorporate dT opposite dG (43,44). The lysine or glutamate may contribute to the stability of mispairs by forming hydrogen bonds to the non-Watson-Crick base pairs. Alternatively, they could provide electrostatic stabilization to minor tautomers or ionized base pairs (56,58).
pol appears to have all the necessary amino acid residues to select for Watson-Crick base pairs. The residues that form the steric constraints for the nascent base pair and interact with the minor groove of the DNA are conserved (1,44). Based upon the structural similarities, one would expect that a mispaired nascent base pair would lead to structural changes in the binding site that are transmitted to the active site so that catalysis does not occur. However, this does not happen with pol , suggesting that the connection between the Watson-Crick checking changes in the catalytic site are disrupted with pol . The many differences between pol and the high fidelity A-family polymerases that are located away from the DNA can lessen the fidelity by promoting a more flexible protein that is not as responsive to mispairs. For instance, although Arg-571 and Gln-769 are in position to make minor groove interactions at or adjacent to the nascent base pair, further away from the active site, the hydrogen bonding network between amino acid residues, water molecules, and the minor groove of the DNA are not observed in pol . These differences may compromise the ability of pol from adopting unreactive conformations that prevent the formation of mispairs.
A unique finding by Lee et al. (44) is the unusual open state in which the O and Ob helices exclude the template from the active site while allowing the dNTP to bind. In our kinetic model, the dNTP-bound state could potentially correspond to E⅐D n ⅐dNTP in Fig. 1A. This model would predict that the binding affinity of the dNTP would not be dependent on the template base. We find, however, the binding affinity of the correct dNTP is 20-fold larger than the incorrect dNTPs, values similar to other A-family polymerases (37). Thus, we do not kinetically detect the unusual open conformation observed in the crystal structure. This conformation can be a transiently formed structure that exists during the binding of the dNTP to the DNA and polymerase that was stabilized by the crystallization procedure.
In summary, we have found that pol forms both correct and incorrect base pairs with high catalytic efficiency. Despite having all the amino acid residues that function to probe for Watson-Crick geometry in the high fidelity A-family polymerases, pol does not prevent mispair formation. Further studies with pol may reveal how interactions in the binding site of the polymerase are transmitted to the active site to modulate polymerase activity.