BYPASS OF N²-ETHYLGUANINE BY HUMAN DNA POLYMERASE κ

Abstract

The efficiency and fidelity of nucleotide incorporation and next-base extension by DNA polymerase (pol) κ past N²-ethyl-Gua were measured using steady-state and rapid kinetic analyses. DNA pol κ incorporated nucleotides and extended 3′ termini opposite N²-ethyl-Gua with measured efficiencies and fidelities similar to that opposite Gua indicating a role for DNA pol κ at the insertion and extension steps of N²-ethyl-Gua bypass. The DNA pol κ was maximally activated to similar levels by a twenty-fold lower concentration of Mn²⁺ compared to Mg²⁺. In addition, the steady state analysis indicated that high fidelity DNA pol κ-catalyzed N²-ethyl-Gua bypass is Mg²⁺-dependent. Strikingly, Mn²⁺ activation of DNA pol κ resulted in a dramatically lower efficiency of correct nucleotide incorporation opposite both N²-ethyl-Gua and Gua compared to that detected upon Mg²⁺ activation. This effect is largely governed by diminished correct nucleotide binding as indicated by the high K_m values for dCTP insertion opposite N²-ethyl-Gua and Gua with Mn²⁺ activation. A rapid kinetic analysis showed diminished burst amplitudes in the presence of Mn²⁺ compared to Mg²⁺ indicating that DNA pol κ preferentially utilizes Mg²⁺ activation. These kinetic data support a DNA pol κ wobble base pairing mechanism for dCTP incorporation opposite N²-ethyl-Gua. Furthermore, the dramatically different polymerization efficiencies of the Y-family DNA pols κ and ι in the presence of Mn²⁺ suggest a metal ion-dependent regulation in coordinating the activities of these DNA pols during translesion synthesis.

1. Introduction

The minor groove DNA adduct N²-ethyl-Guanine (N²-ethyl-Gua)³ is generated from the reduction of acetaldehyde with 2′-deoxyguanosine 3′-monophosphate [1]. Humans are exposed to acetaldehyde from both environmental and endogenous sources. Endogenous acetaldehyde pools arise from the oxidation of ethanol to acetaldehyde during ethanol metabolism [2]. Ethanol is a classified human carcinogen and acetaldehyde has been shown to contribute to tumor formation [3]. The N²-ethyl-Gua adduct has been shown to have a high mis-coding potential during nucleotide incorporation with Klenow fragment of Escherichia coli DNA pol I resulting in G→C transversion mutations [4]. Thus, the mutagenic properties of N²-ethyl-Gua could contribute to the formation of ethanol related cancers [5].

N²-ethyl-Gua is blocking to DNA replication by the replicative DNA pols T7⁻ DNA pol, HIV-1 reverse transcriptase, and the human B family DNA pol α [6,7]. The blocking effect of N²-ethyl-Gua on DNA pols has been observed in both in vitro and cell-based assays [6–8]. In contrast to the blocking effects on replicative DNA pols the N²-ethyl-Gua adduct is bypassed by the Y-family DNA pols η, ι, and κ [6,9–12]. The differences in lesion bypass properties result from the overall active site organization and fidelity checks mechanisms that are dramatically different between Y-family DNA pols and the higher-fidelity A- and B-family DNA pols. Replicative DNA pols have a “minor groove checking” mechanism [13,14]. A conserved K-K-K/R-Y motif identified in several B-family DNA pols and observed in the crystal structure of the bacteriophage RB69 DNA pol scans the DNA minor groove for misincorporation events and for potentially mutagenic base damage [14]. In contrast, the Y-family DNA pols lack a minor groove checking mechanism. DNA pol ι-mediated bypass of N²-ethyl-Gua occurs by a Hoogsteen base-pairing mechanism where the ethyl moiety of N²-ethyl-Gua is repositioned to the major groove by rotation to the syn conformation from the usual for B-DNA anti conformation [9]. The active sites of DNA pol η and DNA pol κ appear sufficiently open to accommodate the N²-ethyl-Gua adduct by utilizing a wobble base pairing mechanism to avoid steric interference when pairing with the correct dCTP [6,9,11].

DNA pol κ is a member of the DinB polymerase family and is the only human Y-family DNA pol with homologues in bacteria (DinB) and archaea (Dpo4) [15–17]. DNA pol κ exhibits the highest fidelity of the Y-family DNA pols when copying unadducted DNA templates but maintains a low level of fidelity compared to replicative DNA pols [15,18,19]. The higher fidelity of DNA pol κ likely relates to its higher level of selectivity of DNA adduct bypass compared to other Y-family DNA pols. DNA pol κ proficiently extends mis-matched primer termini leading to the proposal that DNA pol κ may act mainly as an extender polymerase of adducted 3′ termini during lesion bypass [20–22]. However, DNA pol κ efficiently inserts and extends nucleotides during bypass of some DNA adducts including several minor groove adducts at the N² position of Gua [11,23–25]. Structural studies of DNA pol κ in complex with DNA show that this polymerase encircles the DNA with a unique N-terminal domain, the “N-clasp”, which is not found in other Y-family DNA pols [26]. The “N-clasp” lies along the DNA major groove potentially disrupting polymerase binding of major groove DNA adducts but accommodating adducts that position into the DNA minor groove [25].

DNA pols require divalent metal ions for activation [27,28]. Mg²⁺ is the preferred metal for most DNA pols [29]. Activation of DNA pols using divalent metal ions other than Mg²⁺ results in diminished effects on polymerization efficiency and fidelity [30,31]. However, it was recently demonstrated that DNA pol ι exhibits increased activity and fidelity when copying opposite a template Thy in the presence of Mn²⁺ [32]. We subsequently demonstrated that Mn²⁺ activation of DNA pol ι results in increased polymerization efficiency during nucleotide incorporation opposite the N²-ethyl-Gua adduct [9].

The blocking and potentially mutagenic effects of the N²-ethyl-Gua adduct on replicative DNA pols with the potential for N²-ethyl-Gua-mediated carcinogenesis indicate cellular bypass of this adduct by a specialized Y-family DNA pol. In our continuing efforts to better understand translesion synthesis past N²-ethyl-Gua a series of kinetic studies was performed using human DNA pol κ. The data indicate that DNA pol κ is Mg²⁺ activated and is efficient at both the insertion and extension steps when bypassing N²-ethyl-Gua. These results indicate that DNA pol κ catalyzes efficient nucleotide insertion and extension during bypass of this minor groove N²-alkyl-Gua DNA adduct. Furthermore, the data support a wobble base pairing mechanism for DNA pol κ during dCTP incorporation opposite the N²-ethyl-Gua lesion. The preference for Mg²⁺ activation in this DNA pol κ bypass process suggests that metal ion availability at the replication fork could regulate the activity of Y-family DNA pols during translesion synthesis. These experiments demonstrate the contrasting abilities of DNA pol κ and DNA pol ι to utilize Mn²⁺ for efficient N²-ethyl-Gua bypass providing new insights into the molecular mechanisms of N²-Gua adduct translesion synthesis.

2. Material and methods

2.1 Oligonucleotides

N²-ethyl-Gua phosphoramidites and template oligonucleotides were prepared as described previously [6]. The DNA primer oligonucleotides, 5′-(6-FAM)-GCTCCGGAACCC-3′, 5′-(6-FAM)-GCTCCGGAACCCTT-3′, 5′-(6-FAM)-GCTCCGGAACCCTTC-3′, and 5′-(6-FAM)-GCTCCGGAACCCTTT-3′ were purchased from Operon Biotechnologies, Inc. (Huntsville, AL).

2.2 Expression and Purification of Human DNA Pol κ

The recombinant catalytic fragment of human DNA pol κ (amino acids 19-526) was made as an MBP-DNA pol κ fusion protein with a PreScission Protease (GE Healthcare, Piscataway, NJ) cleavage site seven residues from the DNA pol κ N-terminal methionine. The PreScission Protease recognition sequence and DNA pol κ coding sequence were verified by DNA sequencing. The plasmid constructs were transformed into E. coli BL21(DE3) Rosetta 2 cells (Stratagene, Santa Clara, CA) for overexpression. Cells were grown to an O.D.₆₀₀ = 0.5 at 37 °C and quickly cooled on ice to 17 °C. After induction with 1 mM isopropyl-β-D-thiogalactopyranoside, the cells were allowed to grow for 15 h at 17 °C. Cell extracts were prepared and the MBP-DNA pol κ fusion protein was bound to an amylose resin in 20 mM Tris-HCL buffer (pH 7.5) containing 1 mM EDTA and 200 mM NaCl. The fusion protein was cleaved overnight by on-column incubation with PreScission Protease at 4 °C. The recovered DNA pol κ was purified to homogeneity using phosphocellulose chromatography.

2.3 Assays

For primer extension assays the DNA primer (12-mer) was hybridized to the 32-mer DNA template and added to reactions containing 20 mM Tris-HCl buffer (pH 7.5) and 2 mM DTT, 0.1 mg/ml BSA, 100 μM dNTP’s, 10 nM DNA pol κ, and the amount of MgCl₂ or MnCl₂ indicated in the figure legends. Incubations were 15 min at 37 °C and reactions were quenched with EtOH. Samples were dried and resuspended in 5 μl of a 95% formamide/dye solution. Extension products were separated on 8 M urea/23% polyacrylamide gels, imaged using a PhosphoImager (GE Healthcare, Piscataway, NJ), and quantified using ImageQuant software (GE Healthcare, Piscataway, NJ).

For steady-state kinetic assays, the site specific insertion procedure of Boosalis et al. [33] was used. The DNA primers (14-mer for insertion and 15-mer for extension) were hybridized to the 32-mer DNA template and added to reactions containing 20 mM Tris-HCl buffer (pH 7.5) and 2 mM DTT, 0.1 mg/ml BSA, 2 mM MgCl₂ or 0.075 mM MnCl₂, 50 nM template-primer, and 0.625–3.75 nM DNA pol κ. The amounts of DNA pol κ in reactions yielded less than 20% extended product maximally. Incubations were 5–10 min at 37 °C and reactions were processed as described above. All extended product bands were used to determine kinetic parameters (K_m and k_cat values) by non-linear regression using SigmaPlot 8.02 software (Systat Software, Inc., San Jose, CA) and GraphPad Prism 5.0 software (GraphPad, San Diego, CA). Relative insertion frequencies were calculated as 1/[(k_cat/K_m)_correct/(k_cat/K_m)_incorrect].

For rapid kinetic assays a Bio-Logic SFM-400 (Bio-Logic USA, LLC, Knoxville, TN) configured for quench-flow or a KinTek RQF-3 quench-flow apparatus (KinTek Corp., Austin, TX) was used. Reactions were as described for the steady-state experiments and were initiated by rapid mixing of a pre-incubated template-primer and DNA pol κ mixture from one syringe with dNTP’s and Mg²⁺ or Mn²⁺ chloride salts from a second syringe. Reactions were at 37 °C for time intervals ranging from 5 to 5000 msec and quenched in 0.3 M EDTA. A portion of the quenched reaction mixture (10 μl) was mixed with 30 μl of 95% formamide/dye solution and processed as described above. The amount of product formed was plotted versus time and fit by non-linear regression to the burst equation y = A(1 − e^−kobs·t) + k_ss·t, where A = burst amplitude and indicates the amount of product formed in the first binding event, k_obs = pre-steady-state rate of nucleotide incorporation, k_ss = steady-state rate of nucleotide incorporation and t = time.

3. Results

3.1 Primer Extension Reactions

DNA pol κ bypasses N²-ethyl-Gua the same as unadducted Gua using Mg²⁺ or Mn²⁺ as the activating divalent metal ion. Different DNA pols utilize Mg²⁺ and Mn²⁺ for activation with varying effects on catalytic efficiency and fidelity [31,34]. For example, the human Y-family DNA pol ι exhibits increased catalytic efficiencies with increased fidelities during nucleotide incorporation opposite unadducted DNA and increased bypass efficiency of some DNA adducts [9,32]. Our results have demonstrated that DNA pol ι exhibits increased efficiency but reduced fidelity opposite N²-ethyl-Gua in the presence of Mn²⁺ compared to Mg²⁺ [9,32]. The replicative DNA pols exhibit decreased efficiency and fidelity using Mn²⁺ as the activating metal [30,31]. The metal-dependant effects on DNA pol κ activity during N²-ethyl-Gua bypass were determined in primer extension reactions. A 12-mer primer annealed to a 32-mer DNA template with the 3′-end positioned three nucleotides before the target N²-ethyl-Gua or Gua was used for a substrate (Fig. 1A). Incubation of the template-primer with DNA pol κ in the presence of increased concentrations of MgCl₂ (Fig. 1B) or MnCl₂ (Fig. 1C) generated extension products ranging from 13 to 25 nucleotides in length. DNA pol κ extended the 12-mer primer past the target N²-ethyl-Gua- and Gua-containing DNA templates without detectable polymerase pausing before, at, or after the target site indicating that both nucleotide incorporation and extension past N²-ethyl-Gua occur similarly to that observed opposite Gua.

Figure 1.

DNA pol κ catalyzed primer extension opposite N²-ethyl-Gua- or Gua-containing DNA templates. Primer extension assays were carried out as described in Materials and Methods. The primer (12-mer) was designed with the 3′-end three nucleotides from the target Gua or N²-ethyl-Gua (A). DNA pol κ (10 nM) catalyzed primer extension opposite N²-ethyl-Gua or Gua with increasing concentrations of MgCl₂ (B). The maximum level of primer extension was observed at MgCl₂ concentrations ranging from 2 mM to 20 mM for both templates. DNA pol κ (10 nM) catalyzed primer extension opposite N²-ethyl-Gua or Gua with increasing concentrations of MnCl₂ (C). Full-length primer extension products were observed in the presence of MnCl₂ at concentrations of 0.075 mM and higher.

The DNA pol κ activities in these primer extension analyses demonstrate different sensitivities to Mg²⁺ and Mn²⁺ concentrations. The DNA pol κ extends the DNA primer to completion upon addition of increased levels of each activating metal ion with maximal Mn²⁺-activated DNA pol κ activity detected at levels ~26-fold lower than those for MgCl₂ (Fig. 1B and C). The metal ion concentration dependent activities of DNA pol κ are similar to those previously described with other Y-family DNA pols [9]. The maximal Mg²⁺-activated DNA pol κ activity is detected in the presence of 2–20 mM MgCl₂ (Fig. 1B, lanes 6–10 and 19–23). At MgCl₂ concentrations above 20 mM, DNA pol κ generates reduced levels of full-length extension products and less-than-full-length primer extension products are detected on the gel as bands 13–14 nucleotides in length (Fig. 1B, lanes 10–13 and 23–26). DNA pol κ catalyzes full-length primer extension in the presence of 0.075–0.25 mM MnCl₂ as products are observed on the gel as bands 25 nucleotides in length (Fig. 1C, lanes 6–9 and 19–22). At concentrations of MnCl₂ above 0.25 mM, less-than-full-length primer extension products are detected on the gel ranging from 17–22 nucleotides in length (Fig. 1C, lanes 10–13 and 23–26). The accumulation of shorter primer extension products at these higher metal ion concentrations indicates decreased levels of DNA pol κ activity.

3.2 Efficiency and Fidelity of N²-Ethyl-Gua Bypass by DNA Pol κ

A steady-state kinetic assay was performed to precisely quantify the efficiency and fidelity of nucleotide incorporation by DNA pol κ opposite N²-ethyl-Gua and Gua in the presence of MgCl₂ or MnCl₂. Single nucleotide incorporation experiments were performed with DNA pol κ using 2 mM MgCl₂ or 0.075 mM MnCl₂ and steady-state kinetic parameters (K_m and k_cat) were determined. The metal ion concentrations in these steady state experiments are the lowest levels that result in the maximal amount of primer extension observed by DNA pol κ (Fig. 1). For single nucleotide incorporation reactions, a 14-mer primer was annealed to a 32-mer DNA template to position the 3′ terminus one nucleotide before the target N²-ethyl-Gua or Gua. Separate reactions were initiated by addition of increased concentrations of each of the four dNTP’s. The data were quantified and summarized in Table 1 (Supplemental Fig. S1).

Table 1.

Nucleotide insertion opposite Gua and N²-ethyl-Gua by DNA pol κ^a

Metal Ion	dNTP	Km (μM)	kcat (min−1)	kcat/Km (min−1 μM−1)	relative insertion frequencyb
At Template N2-ethyl-Gua
0.075mM Mn2+
	Cyt	1000 ± 300	270 ± 50	2.7 × 10−1	1
	Gua	70 ± 40	1.6 ± 0.4	2.2 × 10−2	1/12
	Ade	3.0 ± 2.3	0.29 ± .05	9.8 × 10−2	1/2.7
	Thy	130 ± 60	12.1 ± 2.9	9.3 × 10−2	1/3
2mM Mg2+
	Cyt	1.5 ± 0.1	150 ± 3	1.0 × 102	1
	Gua	120 ± 15	1.9 ± 0.06	1.5 × 10−2	1/6 700
	Ade	1300 ± 800	2.4 ± 0.82	2.0 × 10−3	1/50 000
	Thy	630 ± 140	6.4 ± 0.82	1.0 × 10−2	1/10 000
At Template Gua
0.075mM Mn2+
	Cyt	470 ± 130	240 ± 30	5.1 × 10−1	1
	Gua	35 ± 18	1.8 ± 0.3	5.1 × 10−2	1/10
	Ade	5.5 ± 1.6	0.58 ± 0.05	1.0 × 10−1	1/5
	Thy	40 ± 15	12.2 ± 2.0	3.0 × 10−1	1/2
2mM Mg2+
	Cyt	1.6 ± 0.17	150 ± 5	0.9 × 102	1
	Gua	160 ± 30	1.3 ± 0.06	8.1 × 10−3	1/12 000
	Ade	660 ± 150	1.6 ± 0.16	2.4 × 10−3	1/47 000
	Thy	490 ± 260	3.8 ± 0.92	7.7 × 10−3	1/12 000

^aK_m and k_cat values were determined by quantifying gel band intensities using ImageQuant software and non-linear regression analysis of product vs [dNTP] curves using SigmaPlot 8.0.2.

^bRelative insertion frequency was calculated as 1/([k_cat/K_m]_correct/[k_cat/K_m]_incorrect).

The steady-state kinetic analysis indicates that DNA pol κ bypasses N²-ethyl-Gua with the same efficiency as Gua. The DNA pol κ incorporates the correct dCTP opposite N²-ethyl-Gua with an efficiency approximately 400-fold higher when Mg²⁺ is the activating divalent metal ion compared to when Mn²⁺ is the activating metal. In the presence of MgCl₂, the k_cat/K_m value for dCTP incorporation opposite N²-ethyl-Gua is 1.0 × 10² min⁻¹ μM ⁻¹ compared to 2.7 × 10⁻¹ min⁻¹ μM⁻¹ in the presence of MnCl₂. Similarly, the metal ion effects on the efficiencies of DNA pol κ polymerization opposite Gua parallel those opposite N²-ethyl-Gua. In the presence of MgCl₂, the efficiency of DNA pol κ-catalyzed dCTP incorporation opposite Gua, at 0.9 × 10² min⁻¹ μM⁻¹, is ~200-fold higher than that observed in the presence of MnCl₂ at 5.1 × 10⁻¹ min⁻¹ μM⁻¹. The higher efficiencies for correct nucleotide incorporation by DNA pol κ in the presence of MgCl₂ can be attributed to increased nucleotide binding as is reflected in the lower K_m values for dCTP insertion (Table 1). Interestingly, the lower efficiency of DNA pol κ-catalyzed nucleotide incorporation in the presence of MnCl₂ contrasts the increased efficiency observed with DNA pol ι using Mn²⁺ as the activating divalent metal ion [9,32].

The steady-state analysis further indicates that high fidelity DNA pol κ-catalyzed N²-ethyl-Gua bypass is Mg²⁺-dependent. In the presence of MgCl₂, the relative insertion frequencies by DNA pol κ for correct dCTP compared to incorrect dGTP (1/6,700), dATP (1/50,000), and dTTP (1/10,000) opposite N²-ethyl-Gua indicate a relatively high level of fidelity for this Y-family DNA pol during adduct bypass. This level of fidelity is very similar to that measured during bypass of the unadducted template Gua (Table 1). The fidelity of DNA pol κ in the presence of MgCl₂ can be attributed to the high K_m values and low k_cat values for incorporation of the incorrect nucleotides compared to those for dCTP insertion. In contrast, Mn²⁺ activation of DNA pol κ showed low insertion frequencies for correct dCTP compared to incorrect dGTP (1/12), dATP (1/2.7), and dTTP (1/3) opposite N²-ethyl-Gua indicating a low level of fidelity during adduct bypass and similarly low levels of fidelity during Gua bypass (Table 1). The dramatically lower levels of correct versus incorrect nucleotide incorporation by DNA pol κ in the presence of MnCl₂ compared to MgCl₂ were the result of a ~600-fold change in the K_m value for correct dCTP when Mn²⁺ is the activating metal compared to when Mg²⁺ is used (Table 1). Surprisingly, the K_m values for incorrect nucleotides measured in the presence of MnCl₂ are ~10- to 300-fold lower than those for the correct nucleotide opposite N²-ethyl-Gua or Gua. These data indicate a lack of nucleotide discrimination by DNA pol κ during polymerization opposite N²-ethyl-Gua or Gua upon Mn²⁺ activation. These Mn²⁺-related effects on polymerization fidelity by DNA pol κ are similar to those observed for replicative DNA pols [31,34–36]. In particular, this Mn²⁺-dependent decrease in the K_m values for mismatched deoxynucleotides in the polymerase-template complex has been described in fidelity measurements of the T4 DNA pol [37].

3.3 Efficiency of N²-Ethyl-Gua:Cyt extension by DNA Pol κ

The ability of DNA pol κ to extend from the N²-ethyl-Gua-containing template-primer after incorporation of dCTP or dTTP opposite the adduct was measured using a steady-state kinetic assay. The K_m and k_cat values were determined from extension reactions using two different 15-mer DNA primers containing either Cyt or Thy positioned directly opposite the target N²-ethyl-Gua or Gua. Reactions contained increasing concentrations of dGTP in the presence of MgCl₂ or MnCl₂ to measure incorporation opposite the next template Cyt positioned 5′ to the target Gua and N2-ethyl-Gua. The data were quantified and the results summarized in Table 2 (Supplemental Fig. S2).

Table 2.

Extension from Gua:Cyt, Gua:Thy and N²-ethyl-Gua:Cyt, N²-ethyl-Gua:Thy base pairs by DNA pol κ^a

Metal ion	Base pair	Km (μM)	kcat (min−1)	kcat/Km (min−1 μM−1)	relative extension frequencyb
At Template N2 -ethyl-Gua
0.075mM Mn2+
	N2etGua:Cyt	1500 ± 1200	430 ± 250	2.8 × 10−1	1
	N2etGua:Thy	100 ± 20	20 ± 1	2.0 × 10−1	1/1.4
2mM Mg2+
	N2etGua:Cyt	90 ± 60	390 ± 110	4.3	1
	N2etGua:Thy	12 ± 2	40 ± 0.7	3.3	1/1.3
At Template Gua
0.075mM Mn2+
	Gua:Cyt	630 ± 300	200 ± 60	3.2 × 10−1	1
	Gua:Thy	70 ± 20	30 ± 2	4.3 × 10−1	1/0.7
2mM Mg2+
	Gua:Cyt	50 ± 30	530 ± 130	11	1
	Gua:Thy	6.9 ± 1.3	60 ± 2	8.7	1/1.2

^aK_m and k_cat values were determined by quantifying gel band intensities using ImageQuant software and non-linear regression analysis of product vs [dNTP] curves using SigmaPlot 8.0.2.

^bRelative extension frequency was calculated as 1/([k_cat/K_m]_correct/[k_cat/K_m]_incorrect).

The steady-state analysis indicates that DNA pol κ extends N²-ethyl-Gua:Cyt 3′-termini with approximately the same efficiency as Gua:Cyt 3′-termini. The k_cat/K_m value of 4.3 min⁻¹ μM⁻¹ for extension from the N²-ethyl-Gua:Cyt base pair (measured as incorporation of the next correct dGTP nucleotide) is similar to the 11 min⁻¹ μM⁻¹ measured for extension from the Gua:Cyt base pair (Table 2). However, the measured efficiency of extension from the N²-ethyl-Gua:Cyt 3′-termini of 4.3 min⁻¹ μM⁻¹ is approximately 15-fold higher when Mg²⁺ is the activating divalent metal ion compared to 2.8 × 10⁻¹ min⁻¹ μM⁻¹ when Mn²⁺ is the activating metal. Furthermore, the extension efficiency from the mismatched N²-ethyl-Gua:Thy 3′ terminus of 3.3 min⁻¹ μM⁻¹ is similar to the efficiency from correct 3′ Gua:Cyt termini (Table 2). These data indicate that DNA pol κ extends the N²-ethyl-Gua:Cyt template-primer substrate to effect highly efficient bypass of the N²-ethyl-Gua lesion.

3.4 Rapid Kinetic Analysis of DNA pol κ-catalyzed nucleotide insertion opposite N²-ethyl-Gua

A rapid kinetic analysis was performed to further investigate the metal-dependent effects on nucleotide binding and polymerization rates of DNA pol κ using Mg²⁺ and Mn²⁺ as the activating metals. The results show that DNA pol κ-catalyzed DNA polymerization occurs in two phases of nucleotide incorporation. First, a rapid rate or “burst” (k_obs) of product formation is observed. This is followed by a second, slower phase of steady-state (k_ss) product formation [38]. The burst rate is an observed rate of the chemical reaction at the active site, while the burst amplitude is a stoichiometric measurement of the amount of active enzyme in the reaction [38,39].

Formation of the optimal DNA pol κ–DNA template:primer-dNTP ternary complex requires Mg²⁺ as the activating metal as demonstrated in a rapid kinetic “burst” analysis (Fig. 2). The level of DNA pol κ activity measured in the first turnover of substrate to product is dependent upon the metal ion concentration for Mg²⁺ (Fig. 2A) and for Mn²⁺ (Fig. 2B). The measured burst of DNA pol κ–single nucleotide incorporation activity increased as the amount of metal ion was increased from 0.25 mM to 8 mM in the reaction (Fig. 2A and B). A maximum burst amplitude of ~70 ± 5 nM is observed at 5 and 8 mM MgCl₂ (Fig. 2A). In contrast, a lower burst amplitude of ~40 ± 6 nM is observed using MnCl₂ as the activating metal (Fig. 2B). The greater burst amplitude observed using MgCl₂ compared to MnCl₂ indicates that Mg²⁺ has greater capacity to generate the active DNA pol κ–DNA template:primer-dNTP ternary complex. The burst rates were similar for DNA pol κ (~2.4 ± 0.5 s⁻¹ in the presence of MgCl₂ and ~3.1 ± 0.7 s⁻¹ in the presence of MnCl₂) at all metal ion concentrations tested. The similar burst rates observed for DNA pol κ in the presence of MgCl₂ and MnCl₂ indicate that the Mg²⁺ or Mn²⁺ ions may equally facilitate chemistry at the active site.

Figure 2.

A rapid kinetic experiment was used to measure the effects of increasing Mg²⁺ or Mn²⁺ concentrations on single nucleotide incorporation by DNA pol κ. The experiment was performed under enzyme limiting conditions (50 nM DNA pol κ, 100 nM template-primer DNA, and 0.5 mM dCTP) with MgCl₂ (A) or MnCl₂ (B) concentrations of 0.25 mM, 0.5 mM, 2 mM, 5 mM, and 8 mM. Burst amplitudes using MgCl₂ were: A = 15 ± 4, 36 ± 4, 51 ± 4, 65 ± 5, and 62 ± 4 nM at each concentration respectively; and using MnCl₂ were: A= 19 ± 4, 20 ± 1, 22 ± 2, 27 ± 2, and 43 ± 6 nM, respectively.

The rapid reaction analysis demonstrates that DNA pol κ catalyzes dCTP incorporation opposite the adduct N²-ethyl-Gua similarly to that opposite the unadducted Gua. The DNA pol κ rapidly incorporates dCTP opposite N²-ethyl-Gua (Fig. 3A) and Gua (Fig. 3B) with burst amplitudes of ~60 ± 4 nM using MgCl₂ and ~40 ± 4 nM using MnCl₂. Furthermore, DNA pol κ exhibits similar burst rates for dCTP incorporation opposite N²-ethyl-Gua (k_obs = 4.0 ± 0.6 s⁻¹) and Gua (k_obs = 3.5 ± 0.3 s⁻¹) in the presence of MgCl₂ and just slightly lower rates of dCTP incorporation opposite N²-ethyl-Gua (k_obs = 2.1 ± 0.3 s⁻¹) and Gua (k_obs = 1.4 ± 0.2 s⁻¹) with MnCl₂. These data indicate that the DNA pol κ active site accommodates the ethyl moiety of the N²-ethyl-Gua lesion with little or no effect on correct nucleotide incorporation.

Figure 3.

The kinetics of dCTP incorporation opposite Gua and N²-ethyl-Gua were measured in a rapid kinetic experiment under enzyme limiting conditions (50 nM DNA pol κ, 100 nM template-primer DNA, and 8 mM MgCl₂ or MnCl₂). The data were fit to the burst equation (see Materials and Methods) and the following kinetic parameters were determined: opposite template Gua (A) the k_obs = 3.5 ± 0.3 s⁻¹; k_ss = 4.4 ± 1.0 s⁻¹; A = 63 ± 3 nM in the presence of MgCl₂ and using MnCl₂ the k_obs = 1.4 ± 0.2 s⁻¹; k_ss = 3.8 ± 1.0 s⁻¹; A = 47 ± 5 nM; opposite template N²-ethyl-Gua (B) the k_obs = 4.0 ± 0.6 s⁻¹; k_ss = 4.7 ± 1.4 s⁻¹; A = 62 ± 4 nM in the presence of MgCl₂ and using MnCl₂ the k_obs = 2.1 ± 0.3 s⁻¹; k_ss = 7 ± 1 s⁻¹; A = 40 ± 4 nM.

3.5 Activating metal-dependent effects on DNA and dCTP binding in DNA Pol κ

A rapid kinetic experiment was performed to directly measure the metal-dependent effects in DNA Pol κ on DNA binding (K_d^DNA), nucleotide binding (K_d^dCTP) and on the first-order rate of polymerization (k_pol). An active site titration analysis shows that Mg²⁺ is required to generate maximal levels of active DNA pol κ (Fig. 4). The burst amplitudes were measured using increased concentrations of DNA substrate in the presence of MgCl₂ (Fig. 4A) and MnCl₂ (Fig. 4C) as the activating metals. A plot of the burst amplitudes (E:P/T) relative to DNA concentrations (P/T) reveals about a 3-fold higher level of active DNA pol κ is generated in the presence of MgCl₂ (Fig. 4B) compared to MnCl₂ (Fig. 4D). These results show that the DNA pol κ preparation contains approximately 100% active enzyme (~116 ± 7 nM) in the presence of MgCl₂ and approximately 40% active enzyme (~40 ± 5 nM) in the presence of MnCl₂. The active site titration showed that metal ion activation affects the amount of catalytically competent DNA pol κ ternary complex and further indicates that Mg²⁺ is the preferred metal for maximal DNA pol κ activity. Interestingly, there was only a slight difference in the measured K_d^DNA for DNA pol κ whether Mg²⁺ (12 ± 5 nM) or Mn²⁺ (20 ± 10 nM) was used.

Figure 4.

Active site titration and determination of K_d^DNA for DNA Pol κ in the presence of MgCl₂ or MnCl₂. DNA pol κ (100 nM) was incubated with increasing concentrations of template-primer DNA (6–300 nM) and 500 μM dCTP at different reaction times in the presence of MgCl₂ (A–B) or MnCl₂ (C–D). The data were fit to the burst equation (A and C). The burst amplitudes ([E:P/T]) were plotted versus template-primer concentrations ([P/T]) (B and D) and fit to the quadratic equation (y = (0.5(K_d+E_t+D_t))−((0.25(K_d+E_t+D_t)²)−(E_tD_t))^1/2). The active enzyme concentration was determined to be approximately 116 ± 7 nM.

A rapid kinetic analysis further demonstrated that DNA pol κ binds nucleotide more efficiently in the presence of MgCl₂ compared to MnCl₂ as reflected in the lower nucleotide dissociation constant for dCTP (K_d^dCTP) (Fig. 5). The DNA pol κ (50 nM) was incubated with increased concentrations of dCTP (50–750 μM) in the presence of 8 mM MgCl₂ or MnCl₂ and polymerization was measured at varied times from 5 msec to 5 s. The observed burst rates were fit to a hyperbolic equation to determine K_d^dCTP and k_pol. These data show that the K_d^dCTP is ~70 ± 10 μM in the presence of Mg²⁺ (Fig. 5A) and ~160 ± 70 μM in the presence of Mn²⁺ (Fig. 5B). However, this 2-fold difference in the K_d^dCTP does not apparently affect the polymerization rate (k_pol) measured as ~6.0 ± 0.3 s⁻¹ with Mg²⁺ and ~6.0 ± 1.0 s⁻¹ with Mn²⁺. These rapid kinetic data indicating reduced levels of nucleotide binding by DNA pol κ upon Mn²⁺ activation relative to Mg²⁺ activation parrallel our findings of the inefficient incorporation of dCTP in the presence of MnCl₂ in the steady-state analysis as reflected in the increased K_m value for dCTP. The results of both steady-state and rapid kinetic analyses with DNA pol κ indicate that Mg²⁺ is preferentially utilized to achieve high efficiency and fidelity during DNA polymerization contrasting our previous findings with the DNA pol ι (9).

Figure 5.

Determination of K_d^dCTP for DNA pol κ in the presence of MgCl₂ (A) or MnCl₂ (B). DNA pol κ (80 nM) was incubated with 100 nM template-primer DNA and mixed with increasing dCTP (50–750 μM) concentrations. A plot of “burst” rates (s⁻¹) versus dCTP concentration was fit to a hyperbolic equation to determine the maximum rate (k_pol) and the dissociation constant (K_d) for dCTP incorporation.

4. Discussion

The results presented here demonstrate that DNA pol κ is efficient at both the insertion and extension steps during bypass of N²-ethyl-Gua and indicate a role for DNA pol κ in translesion synthesis past bulky N²-alkyl-Gua lesions in vivo. Previous studies have shown that DNA pol κ is efficient at both the incorporation and extension steps when bypassing bulky N²-alkyl-Gua lesions such as N²-CH₂(9-anthracenyl)-Gua, N²-furfuryl-Gua, and N²-BP-Gua [11,23–25]. Furthermore, mice that are deficient for DNA pol κ have been shown to have a spontaneous mutator phenotype with a broad mutation spectrum that includes Gua→Cyt transversion, and Gua→Ade and Gua→Thy transition mutations suggesting that in the absence of DNA pol κ error-prone translesion synthesis, past bulky Gua adducts, is carried out by other DNA pols. Also, the DNA pol κ ^−/− mouse cells are sensitive to treatment with benzo[a]pyrene-dihydrodiol epoxide suggesting that DNA pol κ functions to bypass N²-Gua lesions in vivo [40].

There is evidence that DNA pol κ is efficient at extending mis-matched primer termini and it has been hypothesized that DNA pol κ could act mainly at the extension step in a two-DNA polymerase-mechanism of lesion bypass [21,22,41]. Our data show the high efficiency of DNA pol κ-catalyzed incorporation of dCTP opposite N²-ethyl-Gua. Further, the DNA pol κ exhibits similar efficiencies extending from the Gua:Cyt or the Gua:Thy 3′-termini in the presence of MgCl₂ and only a ~2.5 fold lower extension efficiency from the N²-ethyl-Gua:Cyt or N²-ethyl-Gua:Thy 3′-termini. Together, these data support a model where DNA pol κ participates at both the incorporation and extension steps during translesion synthesis of minor groove DNA adducts.

Y-family DNA pols bypass bulky DNA adducts that would otherwise be blocking to replicative DNA pols. The x-ray crystal structures of several Y-family DNA pols reveal that decreased flexibility in the nucleotide binding domain and an open active site contribute to the accommodation of bulky lesions and their efficient bypass during replication [42]. The size and shape of the DNA polymerase active site is not, however, the only factor in determining facile bypass of DNA adducts by Y-family DNA pols. This is evidenced by the fact that not all adducts are efficiently bypassed by all members of the Y-family. The Y-family DNA pols η, ι, and κ are able to bypass the N²-ethyl-Gua adduct but each recognize and bypass the lesion by distinct catalytic mechanisms [6,9–12]. DNA pol ι bypasses N²-ethylGua by rotating the adducted template base from the anti to syn configuration and forming Hoogsteen base pairs with the incoming nucleotides [9,43]. Biochemical data indicate that the DNA pols η and κ require Watson-Crick base pairing opposite undamaged DNA substrates [44,45]. Structural modeling of the N²-ethyl-Gua lesion into the active sites of DNA pols η and κ suggest that these DNA pols utilize wobble base pairing as a mechanism to avoid steric interference in forming the N²-ethyl-Gua:Cyt base pair during bypass of the adduct [9].

The rapid kinetic experiments with DNA pol κ indicate that the burst rates and burst amplitudes are the same for dCTP incorporation opposite Gua and N²-ethyl-Gua. These data suggest that DNA pol κ forms a wobble base pair (Supplemental Fig. 3) between the incoming dCTP and template N²-ethyl-Gua because there is no apparent interference by the adduct with formation of the active DNA pol κ:dNTP complex. The ethyl lesion at the N2 position of Gua might be predicted to sterically inhibit the formation of a Watson-Crick hydrogen bond with the O2 atom of Cyt. The N²-ethyl-Gua would therefore inhibit dCTP binding at the DNA pol κ active site via Watson-Crick hydrogen bonding resulting in a predicted decreased burst amplitude relative to that observed using a Gua containing DNA template. Alternatively, if a Watson-Crick hydrogen bond were formed between N²-ethyl-Gua and dCTP with the ethyl moiety rotated around the N2 atom in an anti conformation with respect to the N1 of Gua, lower burst amplitudes would be expected opposite N²-ethyl-Gua given the potential for the ethyl lesion to sample different rotational conformations during the reaction. Our data show that the burst amplitudes are the same whether Gua or N²-ethyl-Gua is used as the template base suggesting that a wobble base pair forms between N²-ethyl-Gua and incoming dCTP relieving the potential for steric overlap that would inhibit positioning of the incoming nucleotide for catalysis.

The activating metal ion affects the catalytic mechanism of Y-family DNA pols and is an important factor for maintaining high catalytic efficiency and maximum fidelity during translesion synthesis. The primer extension data demonstrate that DNA pol κ is fully active at low concentrations of Mg²⁺ or Mn²⁺ ions and the activity is metal ion concentration dependent. Steady state and rapid kinetic analyses with DNA pol κ demonstrate that MnCl₂ activation results in decreased efficiency and lower fidelity during nucleotide incorporation opposite both adducted and unadducted DNA templates. These data contrast our findings with DNA pol ι that has higher activity in the presence of Mn²⁺ (9). The DNA pol κ preference for Mg²⁺ as the activating divalent metal ion for high fidelity DNA replication is more similar to replicative DNA pols than to DNA pol ι. This preference for Mg²⁺ by DNA pol κ may reflect the higher fidelity of DNA pol κ when compared to other Y-family DNA pols when replicating unadducted DNA templates [15,18,19].

A surprising finding in our work is detection of the increased K_d for dCTP binding at the DNA pol κ active site with Mn²⁺ ion relative to Mg²⁺ without a significant change to the rate of the chemical step. Together with our previous work these results suggest that differences in the nucleotide incorporation efficiencies of DNA pols κ and ι in the presence of MnCl₂ likely reflect differences in shape and size of the active sites and the postions of DNA and incoming nucleotides for catalysis. DNA pol ι has a narrow binding pocket that makes tight contacts with both the template base and incoming nucleotide resulting in a shortened C1′-C1′ distance [46]. In contrast, DNA pol κ coordinates the DNA template more loosely and may thus rely more heavily on metal ion binding for coordination and correct positioning of the incoming dCTP. Differences in the inter-atomic distances between Mg²⁺ (3.197 Å) and Mn²⁺ (less than 3.0 Å) may affect the geometry of the DNA pol κ active site more than that of DNA pol ι [29]. The distance between two Mg²⁺ ions (usually 4 Å) is greater than that compared to two Mn²⁺ ions and would result in a tighter coordination of the incoming dCTP in the DNA pol κ active site than would Mn²⁺. The smaller distance between two Mn²⁺ ions would help to coordinate the incoming dNTP in the more confined active site of DNA pol ι [29]. The structures of DNA pol ι solved in the presence of Mg²⁺ show poor electron density around the proposed metal ion-binding sites [9,43,46]. A partial exclusion of Mg²⁺ from the DNA pol ι active site could be due to a preference for Mn²⁺ and its smaller inter-atomic distances that would help facilitate productive binding of the tri-phosphate nucleotide [32]. The structure of DNA pol ι reveals that the cysteine residue at position 196 lies in close proximity to the active site and may contribute to the coordination of Mn²⁺ that has a higher affinity for sulfur than Mg²⁺ [47]. These observations could explain the dramatic activation of DNA pol ι and less than efficient polymerization by DNA pol κ in the presence of Mn²⁺ [9,32].

The different properties of the DNA pols κ and ι observed when Mn²⁺ is the activating divalent metal ion suggest distinct biological roles for these DNA pols dependent upon metal ion concentrations. It is possible that small changes to the nuclear environment that alter concentrations of divalent metal ions could contribute to the specific recruitment and activation of a particular Y-family DNA pol at a blocked replication fork to facilitate lesion bypass. The concentrations of metals in the cell are tightly regulated by efflux and influx of metal ions through membrane transporters [48,49]. A transient rise in the concentration of Mn²⁺ at the replication fork, for example, could act to inhibit DNA pol κ from participating in lesion bypass by decreasing its incorporation and extension efficiencies while stimulating the activity of DNA pol ι for efficient translesion synthesis. Transient regulation of DNA pol activity through changing metal ion concentrations at a blocked DNA replication fork would be beneficial for genomic integrity. A rapid and short-lived response would relieve a stalled replication fork quickly and eliminate unnecessary, prolonged DNA synthesis by the low-fidelity Y-family DNA pols thus permitting efficient bypass of the DNA adduct while preserving the overall genome integrity.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi: