Human DNA polymerase κ forms nonproductive complexes with matched primer termini but not with mismatched primer termini
- *Department of Biochemistry, University of Iowa College of Medicine, 51 Newton Road, Iowa City, IA 52242-1109; and
- †Sealy Center for Molecular Science, University of Texas Medical Branch, 6.104 Blocker Medical Research Building, 11th and Mechanic Streets, Galveston, TX 77555-1061
-
Edited by Jerard Hurwitz, Memorial Sloan–Kettering Cancer Center, New York, NY, and approved September 5, 2006 (received for review July 10, 2006)
Abstract
Human DNA polymerase kappa (pol κ) is a member of the Y family of DNA polymerases that function in translesion synthesis. It synthesizes DNA with moderate fidelity and does not efficiently incorporate nucleotides opposite DNA lesions. Pol κ has the unusual ability to efficiently extend from mismatched primer termini, and it extends readily from nucleotides inserted by other DNA polymerases opposite a variety of DNA lesions. All of this has suggested that pol κ functions during the extension step of translesion synthesis. Here, we have carried out pre-steady-state kinetic studies of pol κ using DNA with matched and mismatched primer termini. Interestingly, we find that mismatches present only a modest kinetic barrier to nucleotide incorporation by pol κ. Moreover, and quite surprisingly, active-site titrations revealed that the concentration of active pol κ is very low with matched DNA, and from DNA trapping experiments we determined that this was due to the formation of nonproductive protein·DNA complexes. In marked contrast, we found that the concentration of active pol κ was six-fold greater with mismatched DNA than with matched DNA. Thus, pol κ forms nonproductive complexes with matched but not with mismatched DNA. From these observations, we conclude that pol κ has evolved to specifically function on DNA substrates with aberrant primer-terminal base pairs, such as the ones it would encounter during the extension step of translesion synthesis.
Members of the Y family of DNA polymerases (including pols η, ι, κ, and Rev1) promote translesion synthesis (1–3). These enzymes use diverse mechanisms for incorporating nucleotides opposite damaged template residues. Pol η, for instance, catalyzes the efficient and accurate replication through cyclobutane pyrimidine dimers (4–6) and 8-oxoguanine lesions (7), and the lack of functional pol η in humans causes the variant form of xeroderma pigmentosum (8, 9). Structural studies have suggested that the active site of pol η is unusually large and can accommodate both the bases of a pyrimidine dimer (10), and from biochemical studies, it has been inferred that pol η uses the intrinsic Watson–Crick base pairing ability of the thymine-thymine dimer and of 8-oxoguanine to incorporate the correct incoming nucleotide opposite these lesions (11–14).
Unlike pol η and every other DNA polymerase examined so far, neither pol ι nor the Rev1 protein utilizes Watson–Crick base pairing to incorporate nucleotides. Pol ι catalyzes nucleotide incorporation opposite a variety of DNA lesions, including abasic sites (15, 16), the 3′ T of (6-4) photoproducts (15–17), and the N2-adducted G residues (18, 19). The structures of pol ι in ternary complexes with DNA and dNTPs show that this enzyme uses Hoogsteen base pairing between the template residue and the incoming nucleotide (20, 21). This allows pol ι to replicate through template lesions not capable of forming Watson–Crick base pairs or template lesions containing substantial minor groove modifications (18, 19, 22). The Rev1 protein is required for DNA damage-induced mutagenesis (23), and this enzyme only incorporates nucleotides opposite a narrow range of templates that include nondamaged G residues (24), abasic sites (24, 25), and N 2-adducted G residues (26). The structure of the Rev1 protein in a ternary complex with DNA and an incoming dCTP has shown that an arginine side chain of the protein acts as the template for nucleotide incorporation by forming hydrogen bonds with the Watson–Crick base pairing edge of the dCTP (27).
Human pol κ is a member of the DinB subfamily of the Y family of DNA polymerases. Pol κ synthesizes DNA with higher fidelity than other Y family members, and its error frequency (f inc) ranges from 10−3 to 10−4 (28, 29). Moreover, unlike other Y family polymerases, pol κ is unable to efficiently incorporate nucleotides opposite DNA lesions (28). The high fidelity of pol κ and its inability to incorporate nucleotides efficiently opposite DNA lesions can both be ascribed to the active site being more constrained with respect to the incoming dNTP and the template residue than the active site of pol η (30).
Steady-state kinetics studies have indicated that human pol κ is quite proficient at extending from mismatched primer-terminal base pairs; its mismatch extension frequencies (f ext o) range from 10−1 to 10−2 (31). Pol κ is unique among Y family polymerases in this respect. In addition, pol κ can efficiently extend from aberrant primer-terminal base pairs containing certain template lesions. For example, it extends readily from nucleotides incorporated opposite the 3′ T of a thymine dimer (31), O 6-methylguanine and 8-oxoguanine (32), and N 2-adducted G residues (18, 19).
To better understand how pol κ tolerates mismatched primer-terminal base pairs, we carried out pre-steady-state kinetic studies with this polymerase. Although steady-state kinetic studies are often used to quantify the efficiency of nucleotide incorporation on matched and mismatched DNA substrates, they provide no information regarding the mechanism of nucleotide incorporation. This is because in steady-state kinetics, one only measures the rate of the slowest step of the overall reaction, which in the case of DNA polymerases is very often the uninformative DNA dissociation step. Thus, to understand how the DNA mismatch impacts on the DNA-binding step, the nucleotide-binding step, and the nucleotide-incorporation step, we performed pre-steady-state kinetic studies.
Interestingly, we find that most of the pol κ·DNA complexes are nonproductive when the DNA contains a matched primer-terminal base pair. Moreover, and quite surprisingly, when pol κ is bound to DNA with a mismatched primer-terminal base pair, this nonproductive complex is no longer formed and nearly all of the pol κ molecules are active. Although nonproductive complex formation is rather common when DNA polymerases bind to abnormal DNA substrates, to our knowledge, these observations with pol κ illustrate the first example where nonproductive complex formation occurs with normal but not with abnormal DNA substrates. These findings imply that pol κ has evolved to specifically function on DNA substrates with aberrant primer terminal base pairs.
Results
Previous steady-state kinetic studies have shown that human pol κ is an efficient extender of primer-terminal mismatches (31). This is highly unusual, because all classical polymerases and nearly all translesion synthesis polymerases examined to date have significantly reduced abilities to extend from such mismatches. To better understand the mechanistic basis of the ability of pol κ to extend from mismatched primer termini, we carried out the pre-steady-state kinetic studies describe below.
Kinetics of Nucleotide Incorporation on Matched DNA.
We first examined the mechanism of nucleotide incorporation on a DNA substrate with a matched T·A primer-terminal pair. To observe incorporation in the first enzyme turnover, we used a rapid chemical quench flow instrument, which allows one to initiate and quench reactions on the millisecond time scale. We preincubated 300 nM pol κ (final concentration) with 200 nM 32P-labeled DNA in one syringe and with 100 μM dATP, the correct incoming nucleotide, in the other syringe. Reactions were initiated and quenched after various incubation times ranging from 0.2 to 15 s, and the amount of product formed was graphed as a function of time (Fig. 1 A). As the result shows, the nucleotide incorporation following a matched primer-terminal base pair displays biphasic kinetics (i.e., “burst kinetics”). The best fit of the data to the full burst equation (see Materials and Methods) yielded an observed first-order rate constant for the fast pre-steady-state phase (k 1) equal to 4.5 s−1, an observed rate constant for the slow steady-state phase (k 2) equal to 0.11 s−1, and an apparent enzyme concentration ([E]app) equal to 34 nM.
Active-site titration with a DNA substrate containing a matched primer-terminal base pair. (A) Preincubated pol κ (300 nM) and the DNA substrate with a T·A primer-terminal base pair (200 nM) were mixed with dATP (100 μM) by using a rapid chemical quench flow instrument for various reaction times. The data (filled circles) were fit to the full burst equation with an [E]app equal to 34 ± 2 nM and rate constants equal to 4.5 ± 1.3 s−1 and 0.11 ± 0.02 s−1. (B) Preincubated pol κ (300 nM) and the matched DNA substrate (filled circles, 10 nM; open circles, 20 nM; filled squares, 50 nM; open squares, 100 nM; filled triangles, 150 nM; open triangles, 200 nM) were mixed with dATP (100 μM) for various reaction times. The solid lines represent the best fits to the full burst equation. (C) The [E]app values (filled circles) were graphed as a function of total DNA concentration. The solid line represents the best fit to the quadratic equation, with an active-site concentration equal to 44 ± 8 nM and a K d DNA equal to 70 ± 33 nM.
Biphasic or burst kinetics occurs because the nucleotide-incorporation step precedes the much slower rate-limiting step, which is likely DNA dissociation. Thus, nucleotide incorporation in the first enzyme turnover is faster than in subsequent turnovers. The presence of a pre-steady-state burst makes it possible to determine the mechanism of nucleotide incorporation, which entails measuring the dissociation constant for the nucleotide-binding step (K d dNTP) and the maximum first-order rate constant for the nucleotide-incorporation step (k pol). However, before determining this mechanism, we next carried out an active-site titration to determine the concentration of active pol κ·DNA complexes as well as the dissociation constant for the DNA-binding step (K d DNA).
Active-Site Titration with Matched DNA.
Because [E]app, which is obtained by fitting to the full burst equation, is equal to the concentration of active pol κ·DNA complexes formed during the preincubation period, it was possible to carry out an active-site titration. To do this, we performed the experiments described above using various concentrations of DNA ranging from 10 to 200 nM and graphed the amount of product formed as a function of time for each DNA concentration (Fig. 1 B). From the best fits of these data, [E]app values were obtained for each experiment. These values were then graphed as a function of DNA concentration (Fig. 1 C), and the best fit of these data to the quadratic equation (see Materials and Methods) yielded a K d DNA equal to 70 nM (Table 1) and a concentration of active pol κ·DNA complexes equal to 44 nM.
Comparison of mechanisms of extension from matched and mismatched DNA substrates for human pol κ and T7 DNA polymerase
This result was surprising because the active-site titration provided a concentration of active pol κ·DNA complexes representing only 15% of the total pol κ added to the reaction (300 nM). There are several reasons why an active-site titration might yield a value significantly less than the concentration of total protein. One possibility is that only a small fraction of molecules in the pol κ protein preparation are functional. Another is that the pol κ protein for some reason forms nonproductive complexes with the DNA, which have been observed previously with other DNA polymerases under specific circumstances (see Discussion).
Mechanism of Nucleotide Incorporation on Matched DNA.
Despite the low concentration of active pol κ·DNA complexes observed in the active-site titration, we examined the mechanism of nucleotide incorporation following a matched primer-terminal base pair. We carried out experiments as described above except that we used various concentrations of the incoming nucleotide dATP ranging from 2 to 100 μM. The amount of product was graphed as a function of time for each dATP concentration (Fig. 2 A), and the observed first-order rate constants for the fast pre-steady-state phase (k obs) were graphed as a function of dATP concentration (Fig. 2 B). The best fit of this data to the hyperbolic equation (see Materials and Methods) yielded a K d dNTP equal to 43 μM and a rate constant for the nucleotide-incorporation step (k pol) equal to 5.6 s−1 (Table 1).
Pre-steady-state kinetics of extension from a matched primer-terminal base pair. (A) Preincubated pol κ (300 nM) and the DNA substrate with a T·A primer-terminal base pair (200 nM) were mixed with various concentrations of dATP (filled circles, 2 μM; open circles, 5 μM; filled squares, 10 μM; open squares, 20 μM; filled triangles, 50 μM; open triangles, 100 μM) for various reaction times. The solid lines represent the best fits to the full burst equation. (B) The observed rate constants of the burst phases (filled circles) were graphed as a function of dATP concentration. The solid line represents the best fit to the hyperbolic equation, with a k pol equal to 5.6 ± 0.7 s−1 and a K d dNTP equal to 43 ± 11 μM.
Kinetics of Nucleotide Incorporation on Mismatched DNA.
To determine the impact of a primer-terminal A·A mismatch on the mechanism of nucleotide incorporation by pol κ, we next examined whether or not pol κ displayed biphasic kinetics in the presence of the mismatch. We preincubated 150 nM pol κ protein with 500 nM DNA and initiated the reaction by adding 200 μM dATP (Fig. 3 A). Biphasic kinetics was observed with the mismatch-containing DNA, and the best fit of these data yielded an [E]app equal to 130 nM, k 1 equal to 0.055 s−1, and a k 2 equal to 0.013 s−1.
Active-site titration with a DNA substrate containing a mismatched primer-terminal base pair. (A) Preincubated pol κ (150 nM) and the DNA substrate with an A·A primer-terminal base pair (500 nM) were mixed with dATP (200 μM) by using a rapid chemical quench flow instrument for various reaction times. The data were fit to the full burst equation with [E]app equal to 130 ± 32 nM and rate constants equal to 0.055 ± 0.020 s−1 and 0.013 ± 0.009 s−1. (B) Preincubated pol κ (150 nM) and the mismatched DNA substrate (filled circles, 25 nM; open circles, 50 nM; filled squares, 100 nM; open squares, 200 nM; filled triangles, 300 nM; open triangles, 500 nM) were mixed with dATP (200 μM) for various reaction times. The solid lines represent the best fits to the full burst equation. (C) The [E]app values (filled circles) were graphed as a function of total DNA concentration. The solid line represents the best fit to the quadratic equation, with an active-site concentration equal to 140 ± 10 nM and a K d DNA equal to 46 ± 16 nM.
A direct comparison of the kinetics for the matched DNA substrate (Fig. 1 A) and the mismatched DNA substrate (Fig. 3 A) show two obvious differences. First, the rate of the pre-steady-state burst was 82-fold slower with the mismatched DNA substrate than with the matched DNA. This was not unexpected, because a primer-terminal mismatch should provide some kinetic barrier, however small, to nucleotide incorporation. Second, and quite unexpectedly, [E]app was substantially greater with the mismatched DNA than with the matched DNA. One cannot directly compare the values from these two experiments, because we used half as much pol κ protein with the mismatched DNA substrate. However, by comparing the [E]app relative to the total pol κ concentration in these experiments, one finds that with the mismatched DNA it was 87% of the total protein concentration, whereas with the matched DNA it was only 11% of the total protein concentration.
Active-Site Titration with Mismatched DNA.
To further investigate this discrepancy in [E]app, we carried out an active-site titration with the DNA substrate containing the primer-terminal mismatch (Fig. 3 B and C). We obtained a K d DNA equal to 46 nM (Table 1) and a concentration of active pol κ·DNA complexes equal to 140 nM. This concentration of active pol κ·DNA complexes represents 93% of the total pol κ protein used in the assay. Because the protein stocks used to examine the kinetics of nucleotide incorporation with matched and mismatched DNA substrates were the same, this rules out the possibility mentioned above that most of the pol κ is inactive. These results strongly suggest that pol κ forms nonproductive complexes with DNA substrates containing matched primer-terminal base pairs, but not with substrates containing mismatches.
Mechanism of Nucleotide Incorporation on Mismatched DNA.
We next examined the mechanism of nucleotide incorporation following a primer-terminal mismatch by carrying out experiments as described above except that various concentrations of dATP were used (Fig. 4). From these data, we obtained a K d dNTP equal to 110 μM and a k pol for the nucleotide-incorporation step equal to 0.097 s−1 (Table 1).
Pre-steady-state kinetics of extension from a mismatched primer-terminal base pair. (A) Preincubated pol κ (150 nM) and the DNA substrate with a A·A primer-terminal mismatch (500 nM) were mixed with various concentrations of dATP (filled circles, 10 μM; open circles, 20 μM; filled squares, 50 μM; open squares, 100 μM; filled triangles, 150 μM; open triangles, 200 μM) for various reaction times. The solid lines represent the best fits to the full burst equation. (B) The observed rate constants of the burst phases (filled circles) were graphed as a function of dATP concentration. The solid line represents the best fit to the hyperbolic equation, with a k pol equal to 0.097 ± 0.029 s−1 and a K d dNTP equal to 110 ± 70 μM.
DNA Trapping Experiment.
To directly demonstrate the presence of a nonproductive pol κ·DNA complex with the matched DNA substrate, we carried out a DNA trapping experiment (Fig. 5). In this experiment, we preincubated 150 nM pol κ with 200 nM 32P-labeled DNA, and we initiated the reaction by adding 200 μM dATP and a 10-fold excess of unlabeled DNA to trap any pol κ protein that dissociated from the 32P-labeled DNA. Reactions were quenched after various time intervals ranging from 1 to 120 s. The data were fit to an equation describing a double exponential function (see Materials and Methods). The fast phase (with an amplitude equal to 29 nM and a rate constant equal to 0.91 s−1) represents the incorporation of nucleotide by pol κ molecules bound to DNA in the productive mode. The slow phase (with an amplitude equal to 35 nM and a rate constant equal to 0.017 s−1) represents the eventual incorporation of nucleotides by pol κ molecules that bound to the DNA in the nonproductive complex. The presence of this slow phase demonstrates that the nonproductive complex does indeed exist and that some of these nonproductively bound pol κ molecules can slowly enter the productive binding mode without first dissociating from the DNA.
DNA trapping experiment with a DNA substrate containing a matched primer-terminal base pair. Preincubated pol κ (150 nM) and 32P-labeled DNA substrate containing a matched T·A primer-terminal base pair (200 nM) were rapidly mixed with dATP (200 μM) and an excess of unlabeled DNA substrate (2 μM) for various reaction times (filled circles). The solid line represents the best fit to an equation describing a double exponential function with A 1 equal to 29 ± 3 nM, A 2 equal to 35 ± 6 nM, k 1 equal to 0.91 ± 0.28 s−1, and k 2 equal to 0.017 ± 0.009 s−1.
Discussion
Human pol κ has the unusual ability to extend from primer-terminal mismatches with relatively high efficiencies. Steady-state kinetics was used previously to measure the relative extension frequencies from matched versus mismatched primer-termini (f ext o). For pol κ, f ext o ranged from 10−1 to 10−2 (31). Here, we have used pre-steady-state kinetics to examine the mechanistic basis of its mismatch extension ability. Using pre-steady-state kinetics, one can in principle measure the equilibrium and rate constants associated with all of the individual steps of the nucleotide incorporation reaction. Of particular interest are the dissociation constant for the DNA-binding step (K d DNA), the dissociation constant for the incoming nucleotide-binding step (K d dNTP), and the rate constant for the nucleotide-incorporation step (k pol), because these are the steps that contribute the most to the specificity of DNA polymerases. When these enzymes discriminate against incorporating incorrect incoming nucleotides or against extending from mismatched primer-termini, this discrimination generally occurs at one or more of these three steps.
From pre-steady-state kinetic analyses, we have found strong evidence for the formation of a nonproductive complex when pol κ extends from normal, matched primer-terminal base pairs. We observed a very low apparent enzyme concentration ([E]app) when using matched DNA substrates (Fig. 1). This is significant because [E]app corresponds to the amounts of active pol κ·DNA complex present at the beginning of the reaction. Active-site titrations showed that there is little pol κ active (15% of the total protein) under these conditions. This low active-site titration cannot be attributed to inactive protein in the pol κ preparation, because protein from the same preparation was almost completely active (93% of the total protein) when used with mismatch-containing DNA substrates (Fig. 3). Moreover, the DNA trapping experiment (Fig. 5) directly demonstrated the presence of nonproductive pol κ molecules bound to the DNA that can convert to productive complexes at a rate significantly slower than the rate of nucleotide incorporation. In addition, the observations of large and small [E]app values with the A·A mismatched substrate and the T·A matched substrate, respectively, are not limited to only these sequence contexts. We have also examined this parameter with A·T, G·C, and C·G matched substrates as well as T·C, T·G, A·C, and A·G mismatched substrates (data not shown). In these cases also, we observed significantly larger [E]app values with the mismatched DNA than with the matched DNA. Moreover, we have carried out further experiments using different preparations of pol κ and obtained the same results. Thus, this trend is consistent through multiple protein preparations.
The precise nature of this nonproductive protein·DNA complex is not clear. There are two possible general models that could explain our observations. In the first model, the nonproductive complex would form upon DNA binding, and this nonproductive enzyme·DNA binary complex would not be capable of dNTP binding or nucleotide incorporation. This model was initially put forward to explain the low [E]app observed with HIV-1 reverse transcriptase at RNA hairpin-induced and DNA hairpin-induced pause sites (33, 34). In the second model, the nonproductive complex would form upon dNTP binding, and this nonproductive enzyme·DNA·dNTP ternary complex would not be capable of nucleotide incorporation. This model has been proposed to explain the low [E]app observed with T7 DNA polymerase and HIV-1 reverse transcriptase at sites of DNA damage (35, 36). The data presented here are not sufficient to discriminate between these models. Kinetic simulations have shown that the differences between these two models are subtle (35), and far more sophisticated experiments will be necessary to address this complicated issue.
Although we do not know the basis of nonproductive complex formation, our kinetic analyses allow us to directly compare the mechanisms of extension from matched versus mismatched DNA by pol κ in the productive binding mode. By comparing the K d DNA, we find that pol κ binds the A·A mismatch-containing DNA with a 1.5-fold higher affinity than the T·A matched DNA (Table 1). This difference is within the range of differences expected for different sequence contexts, and we thus conclude that pol κ binds to matched and mismatched DNA substrates with the same affinity. Incidentally, our previously published steady-state measurements of f ext o had assumed that the binding affinity for matched and mismatched DNA was the same (31), and this assumption turns out indeed to be correct for pol κ. From these K d DNA measurements, we find that the difference in free-energy changes of matched versus mismatched DNA (ΔΔG DNA) is very small (0.25 kcal/mol). Thus the ΔG DNA values for the ground states of the E·DNA binary complex with matched and mismatched DNA are virtually the same (Fig. 6). Similarly, a mismatched primer-terminal base pair has little impact on the DNA-binding affinity in the case of the T7 DNA polymerase (37, 38). It too has a very small ΔΔG DNA (0.34 kcal/mol), and this is reflected in essentially the same ΔG DNA values as for the E·DNA complex with matched and mismatched DNA.
Free-energy profiles comparing mechanisms of extension from matched and mismatched DNA substrates by human pol κ and T7 DNA polymerase. (Left) Portions of the free-energy profiles for the extension of matched (black) and mismatched (gray) DNA substrates by pol κ. The equilibrium constants and rate constants used to calculate these free energies are given in Table 1. (Right) Comparable portions of the free-energy profiles for T7 DNA polymerase. The equilibrium and rate constants are from Patel et al. (37) and Wong et al. (38) and are given in Table 1. To calculate these values, we have assumed arbitrary concentrations of DNA equal to 100 nM and dNTP equal to 100 μM.
Comparison of the K d dNTP values for pol κ show that the incoming dNTP binds 2.6-fold tighter to the E·DNA complex when the primer-terminal base pair is matched than when it is mismatched. This slight difference corresponds to a ΔΔG dNTP that is small (−0.55 kcal/mol), and thus the ΔG dNTP for the ground states for the E·DNA·dNTP ternary complexes do not differ. This is similar to what was observed in the T7 DNA polymerase, when a 4-fold difference in K d dNTP (corresponding to a ΔΔG dNTP of −0.81 kcal/mol) was reported (37, 38). Consequently, for both human pol κ and T7 DNA polymerase, almost no discrimination between matched and mismatched DNA substrates occurs at either the DNA-binding step or the subsequent dNTP-binding step (Fig. 6).
In the case of the T7 DNA polymerase, almost all of the discrimination against mismatch extension occurs at the nucleotide-incorporation step (Table 1). There is a 12,000-fold slower rate constant for this step when the polymerase is bound to a mismatch-containing DNA (37, 38). The change in free energy of activation (ΔΔG ‡) is large (−5.48 kcal/mol), and this is responsible for the substantial difference in the free energy of the transition states for the nucleotide incorporation step, (E·DNA·dNTP)‡ (Fig. 6). Here, we show that in the case of human pol κ, this difference is only 58-fold (Table 1). This reflects a much smaller ΔΔG ‡ equal to −2.38 kcal/mol and leads to transition states that are much closer than what was observed with T7 DNA polymerase (Fig. 6).
From the ratios of the K d DNA, K d dNTP, and k pol values for matched versus mismatched DNA, it is possible to calculate the relative frequency of extension for matched versus mismatched DNA (f rel). The f rel is a direct comparison of the efficiencies of extension by the productive complexes on the matched DNA and on the mismatched DNA. We find that in the case of pol κ, the f rel is equal to 100 (Table 1); thus, the productive complex bound to matched DNA is 100-fold more efficient at incorporating the next correct nucleotide than is the productive complex bound to mismatched DNA.
Previous steady-state analyses of pol κ relied on the implicit assumption that the same percentage of protein molecules is active with different DNA substrates. Our finding that a large portion of pol κ molecules forms a nonproductive complex with the matched DNA substrate, but not with the mismatched DNA substrate, violates this assumption. Because under steady-state conditions, some portion of pol κ will be in a nonproductive complex, steady-state kinetics will underreport the intrinsic differences between pol κ on matched versus mismatched substrates. It remains unclear precisely how much pol κ will be in nonproductive complexes under steady-state conditions because of the uncertainty regarding the rates of DNA dissociation for the active, inactive, and postincorporation complexes as well as the rate of conversion between active and inactive complex on DNA. However, it must be emphasized that the unexpected finding of this nonproductive complex highlights the potential problems one might encounter when relying solely on steady-state kinetics data.
So what does this mean for the biological function of pol κ? Nonproductive complexes have been observed previously with several other polymerases. For example, HIV-1 reverse transcriptase was shown to form nonproductive complexes on normal DNA substrates (≈2- to 3-fold reduction in [E]app) (35), and it forms additional nonproductive complexes at RNA-hairpin induced- and DNA-hairpin-induced pause sites (33, 34) (≈5- to 10-fold reduction in [E]app). Furthermore, both T7 DNA polymerase and HIV-1 RT form nonproductive complexes with 8-oxoguanine lesions (39), cisplatin adducts (40), as well as O 6-methylguanine and O 6-benzylguanine lesions (36). In general, it seems that nonproductive complex formation is a common occurrence among DNA polymerases when they function on nonoptimal or “abnormal” DNA substrates.
To our knowledge, our analyses with pol κ provide the first example of a nonproductive complex of a DNA polymerase forming under a “normal” circumstance but not under an “abnormal” circumstance. This would suggest that pol κ has evolved to view normal, matched primer-terminal base pairs as nonoptimal DNA substrates. These findings imply that pol κ has evolved to specifically function on DNA substrates with aberrant primer terminal base pairs, such as ones that it would encounter during the extension step of translesion synthesis.
Materials and Methods
Purification of Pol κ.
Human pol κ1–559 was purified from yeast strain BJ5464 carrying plasmid pBJ849, which encodes pol κ1–559 fused in-frame with glutathione S-transferase (GST). The GST fusion protein was purified, and the GST portion of the fusion protein was removed by treatment with PreScission protease (Amersham Pharmacia) as described in ref. 30. Concentration of the protein was determined by using the Bio-Rad protein assay with BSA as a standard. Purified protein was stored at −80°C in 10-μl aliquots. Human pol κ1–559 is missing 311 amino acid residues from its C terminus. Previous biochemical studies have shown that the deletion of these residues has no affect on the activity of pol κ (30). Thus, we will simply refer to the shortened form of this protein as pol κ.
DNA and Nucleotide Substrates.
DNA substrates were made by annealing two synthetic oligodeoxynucleotides: a 45-mer template and a 25-mer primer. The 45-mer template (5′-GGA CGG CAT TGG ATC GAC CTA GAG TTG GTT GGA CGG CTG CGA GGC) was used to generate both matched and mismatched DNA substrates. The italic A is in the position of the primer-terminal base pair. Two 25-mer primers (5′-GCC TCG CAG CCG TCC AAC CAA CTC X) were used to generate the matched substrate (X is a T) and mismatched substrate (X is an A). The primer strand was 5′-32P end-labeled with T4 polynucleotide kinase (New England Biolabs) and [γ-32P]ATP [3,000 Ci/mmol (1 Ci = 37 GBq); PerkinElmer] at 37°C for 1 h. Labeled primer strands were separated from unreacted [γ-32P]ATP with a Sephadex G-25 spin column (GE Healthcare) and were annealed to template strands in 50 mM Tris·Cl and 100 mM NaCl by heating at 90°C for 2 min and slow cooling to room temperature over several hours. Annealed strands were stored at 4°C for up to 2 weeks. Solutions of dATP (100 mM) were obtained from New England Biolabs and stored at −80°C.
Pre-Steady-State Kinetics.
All polymerase reactions were measured in 25 mM Tris·Cl (pH 7.5)/5 mM MgCl2/5 mM DTT/10% glycerol at 22°C. The incorporation of dATP opposite template T following both matched and mismatched primer termini was measured on a rapid chemical quench flow instrument (KinTek). For extension from a T·A match, pol κ (300 nM final concentration) was preincubated with the labeled DNA substrate (10–200 nM) in one syringe and mixed rapidly with dATP (2–100 μM) from the other syringe. The reaction was quenched after various times (0.2–15 s) with 0.2 M EDTA. For extension from an A·A mismatch, pol κ (150 nM final concentration) was preincubated with the DNA substrate (25–500 nM) and mixed with dATP (10–200 μM). The reaction was quenched after various times (1–60 s). Extended primer strands (the product) were separated from unextended primer strands (the substrate) on a 15% polyacrylamide sequencing gel containing 8 M urea. The labeled gel bands were quantified by using an InstantImager (Packard). Each set of DNA and dNTP concentrations were repeated at least three times to ensure reproducibility, and the amplitudes and rates obtained from these multiple experiments were in close agreement.
DNA Trapping Experiment.
DNA trapping experiments were carried out with the matched DNA substrate in the rapid chemical quench flow instrument. Pol κ (150 nM final concentration) was preincubated with the labeled DNA substrate (200 nM) in one syringe and mixed rapidly with dATP (200 μM) and an excess of unlabeled DNA substrate (2 μM) from the other syringe. Reactions were quenched after time intervals ranging from 1 to 120 s. To verify the effectiveness of the unlabeled DNA trap, control experiments were carried out in which labeled DNA (200 nM) and unlabeled DNA (2 μM) were preincubated in one syringe and rapidly mixed with pol κ (150 nM final concentration) preincubated with dATP (200 μM) from the other syringe. No incorporation was observed in these control experiments. The DNA products were quantified as described above.
Data Analysis.
The amounts of product (P) formed were graphed as a function of time (t), and the data were fit by nonlinear regression (SigmaPlot 8.0) to the full burst equation:
where [E]app is the apparent enzyme concentration (which is proportional to the pre-steady-state burst amplitude), k
1 is the first-order rate constant for the burst phase, and k
2 is the first-order rate constant for the steady-state phase. The full burst equation differs slightly from the more commonly
used, simplified burst equation. We could not use the simplified burst equation in this study because the data we obtained
with the mismatched DNA substrate violated one of its chief assumptions, namely that k
2 ≪ k
1 (41). Thus, we carried out all analyses using the full burst equation, which does not make such an assumption.
To determine the dissociation constant for the nucleotide-binding step (K
d
dNTP) and the maximum first-order rate constant for nucleotide incorporation (k
pol), the observed rate constants for the pre-steady-state phases (k
1 or k
obs) were graphed as a function of dATP concentration ([dATP]) and the data were fit by nonlinear regression to the hyperbolic
equation:
To determine the active concentration of pol κ ([pol κ]) and the dissociation constant for the DNA-binding step (K
d
DNA), the [E]app obtained from fits to the full burst equation were graphed as a function of total DNA concentration ([DNA]), and the data
were fit to the quadratic equation:
For DNA trapping experiments, the amounts of product (P) formed were graphed as a function of time (t) and the data were fit to an equation describing a double exponential function:
where A
1 and A
2 are the amplitudes of the first and second phases, respectively, and k
1 and k
2 are the first-order rate constants of the first and second phases, respectively.
Acknowledgments
We thank Bret Freudenthal, Craig Howell, and Christine Kondratick for assistance in preparing the manuscript. This work was supported by American Cancer Society Research Scholar Grant RSG-05-139-01-GMC (to M.T.W.) and National Institute on Environmental Health Sciences Grant ES012411 (to L.P.).
Footnotes
- §To whom correspondence should be addressed at: Department of Biochemistry, 4-403 Bowen Science Building, University of Iowa, Iowa City, IA 52242-1109. E-mail: todd-washington{at}uiowa.edu
-
Author contributions: K.D.C. and M.T.W. designed research; K.D.C. performed research; R.E.J., L.P., and S.P. contributed new reagents/analytic tools; K.D.C. and M.T.W. analyzed data; and K.D.C., L.P., S.P., and M.T.W. wrote the paper.
-
The authors declare no conflict of interest.
-
This article is a PNAS direct submission.
- © 2006 by The National Academy of Sciences of the USA