Inhibition of Excision of Oxidatively Generated Hydantoin DNA Lesions by NEIL1 by the Competitive Binding of the Nucleotide Excision Repair Factor XPC-RAD23B

Abstract

The interplay between nucleotide excision repair (NER) and base excision repair (BER) of non-bulky, oxidatively generated DNA lesion, has long been a subject of significant interest. The hydantoin oxidation products of 8-oxoguanine, spiroiminodihydantoin (Sp) and 5-guanidinohydantoin (Gh), are substrates of both BER and NER in HeLa cell extracts and human cells [Shafirovich et al. (2019) Chem. Res. Toxicol. 32, 753–761]. The primary factor that recognizes DNA lesions is the DNA damage-sensing factor XPC-RAD23B (XPC), while the glycosylase NEIL1 is known to remove Gh and Sp lesions from double-stranded DNA. It is shown here that in aqueous solutions containing nanomolar concentrations of proteins, XPC and NEIL1 compete for binding to 147-mer oligonucleotide duplexes that contain single Gh or Sp lesions under conditions of [protein] >> [DNA] concentrations, thus inhibiting the rate of BER catalysed by NEIL1. The non-covalently bound NEIL1 molecules can be displaced by XPC at concentration ratios R = [XPC]/[NEIL1] > 0.2, while full displacement of NEIL1 is observed at R ≥ 0.5. In the absence of XPC and under single turnover conditions, only the burst phase is observable. However, upon progressively increasing the XPC concentration, the amplitude of the burst phase decreases gradually, and a slower time-dependent phase of incision product formation manifests itself with rate constants of 3.0×10⁻³ (Gh) and 0.90×10⁻³ s⁻¹ (Sp). These slow kinetics are attributed to the dissociation of XPC-DNA complexes that allow for the rebinding of NEIL1 to the temporarily exposed Gh or Sp lesions, and the incisions observed under these steady-state conditions.

Graphical Abstract

INTRODUCTION

The repair of oxidatively generated DNA lesions is a crucial factor for maintaining genomic stability during oxidative stress.¹ In general, non-bulky oxidatively generated DNA lesions are removed by base excision repair (BER) mechanisms, while bulky DNA double helix-distorting lesions are excised by nucleotide excision repair (NER) machinery.^2–5 Oxidatively generated DNA damage targets guanine,⁶ the most easily oxidizable natural nucleic acid base.⁷ Sequential oxidation of guanine generates a series of guanine modifications including the well-known 8-oxo-7,8-dihydroguanine (8-oxoG),⁶ and further oxidation products of 8-oxoG, such as the diastereomeric spiroiminodihydantoin (Sp) and 5-guanidinohydantoin (Gh) (Figure 1).⁸

Figure 1.

Structure of the 147 bp ³²P-internally labeled (*p) double-stranded DNA substrates harboring site-specifically positioned guanine lesions (H = S-Sp, or Gh).

Although the cellular levels of hydantoin lesions are low,⁹ these lesions can contribute to the malignant transformations of cells because they are at least one order of magnitude more mutagenic than 8-oxoG.¹⁰

Hydantoin lesions are excellent substrates of BER enzymes that include glycosylases NEIL1 and NEIL3.^11–14 Human NEIL1 is a bifunctional DNA glycosylase that cleaves the N-glycosyl bond of the hydantoin lesions and excises the resulting abasic site by a lyase β,δ elimination mechanism that results in 3′ and 5′ phosphate ends at the cleavage site.^15–17

The possible interplay between different repair pathways such as BER and NER in removing oxidatively generated DNA lesions from double-stranded DNA has been of significant interest for some time.^18–20 Recently, we found that hydantoin lesions are excised by NER and BER pathways in human cell-free extracts²¹ and in intact human cells.²² However, it remained unclear how these two repair pathways compete with one another in the excision of the same DNA lesions in biological systems.

In this work, we demonstrate that the heterodimeric DNA damage-sensing NER factor XPC-RAD23B (XPC) competes with the BER enzyme NEIL1 for binding to the same Gh and Sp hydantoin lesions embedded 147 base pair (bp), ³²P-internally labeled double-stranded DNA substrates (Figure 1). The formation of NEIL1- and XPC-RAD23B-DNA complexes was monitored by electrophoretic mobility shift assays (EMSA).²³ The kinetics of the glycosylase/lyase activity of NEIL1 in the presence of XPC-RAD23B was monitored by the formation of the ³²P-labeled DNA fragments derived from the DNA strand cleavage at the lesion site. Numerical least-squares fitting routines of the BER-induced product yields as a function of reaction time using the DynaFit4 software package,²⁴ allowed us to estimate the dissociation rate constant of XPC from non-covalent XPC-DNA complexes by monitoring the binding and subsequent cleavage of the DNA strands by NEIL1.

MATERIALS AND METHODS

Construction of 147 bp DNA Substrates

The 2’-oligodeoxynucleotides were purchased from Integrated DNA Technologies (Coralville, IA, USA) and purified by denaturing polyacrylamide gel electrophoresis. The 11-mer oligonucleotide adducts containing the S-Sp and R-Sp diastereomers were generated by the site-selective oxidation of guanine in the 5’-CCATCGCTACC sequence using photochemically generated carbonate radical anions at pH 7.5 – 8.0 to oxidize the guanine residue as described earlier.^25–27 The Gh-modified oligonucleotides were prepared by the oxidation of 5’-CCATC[8-oxoG]CTACC sequences with (NH₄)₂IrCl₆ complex at pH 6.0.^{8, 28} The Sp or Gh modified oligonucleotides were isolated and purified by anion-exchange HPLC.^{28, 29} The Sp adduct containing the S-Sp diastereomer was the first-eluting fraction (Sp1), whereas the Sp oligonucleotide with the R-Sp diastereomer eluted in the second fraction (Sp2) on a DNAPac PA-100 column (Dionex, Sunnyvale, CA, USA).³⁰

The 147 bp DNA duplexes were equivalent to the Widom 601 DNA sequence,³¹ except for the 11-mer fragment containing the single site-specific hydantoin lesion (Figure 1). These 11-mers were inserted into 147-mer sequences by replacing an equal number of the parent nucleotides so that the lesions were positioned at the 66^th nucleotide counted from the 5′-end of the 147-mer (Figure 1). The 147-mer modified strands were assembled by ligating the 5′−³²P-labeled 11-mers 5′-CCATCHCTACC (H = S-Sp, Gh, or G) to the 5′-and the 3′-flanking 60- and 76-mer sequences, 5′-CACAGGATGTATATATCTGACACGTGCCTGGAGACTAGGGAGTAATCCCCTTGGCGGTTA, and 5′-ACAGCGCGTACGTGCGTTTAAGCGGTGCTAGAGCTGTCTACGACCAATTGAGCGGCCTCGGCACCGGGATTCTCCA, respectively.³² The 147-mer modified sequences obtained by this approach were purified by denaturing PAGE and annealed with their complementary strands (bearing C-bases opposite the lesion, Figure 1) by heating at 90 °C for 5 min, followed by a slow cooling to room temperature overnight to form the 147 bp DNA duplexes.

XPC-RAD23B and NEIL1

Human NEIL1 protein was expressed in Escherichia coli cells and purified as described elsewhere.³³ The NEIL1 samples were provided by R. Steven Lloyd (Oregon Health and Science University). Human His-XPC-RAD23B (full length) was overexpressed and purified from insect cells as described previously.³⁴ The purified proteins were stored at −78 °C. The concentrations of the proteins were determined by the Bradford assay using bovine serum albumin (BSA) as the standard. Active site titrations^{12, 35} of XPC-RAD23B or NEIL1 were performed using 20 nM 147bp SpDNA duplexes spiked with traces of ³²P-internally labeled SpDNA, and all concentrations of XPC-RAD23B or NEIL1 obtained from the analysis of either native (XPC-RAD23B) or denaturing (NEIL1) gels are given as active concentrations.

EMSA Assays

The binding of proteins to 147 bp double-stranded DNA substrates was investigated in 20 μL binding buffer (20 mM sodium phosphate, pH 7.4, 5 mM MgCl₂, 1 mM EDTA, 150 mM NaCl, 1 mM dithiothreitol, 5% glycerol, 0.01% Triton X-100, 100 ng/μL BSA, 5 ng/μL) after incubation for 5 – 10 min at 25 °C. These solutions contained 0.1 nM ³²P-labeled DNA substrate, and the indicated concentrations of NEIL1 and/or XPC-RAD23B proteins, 5 ng/μL covalently closed circular pBR322 plasmid) containing 0.1 nM ³²P-labeled DNA substrate.³⁶ The samples were then placed on ice and aliquots of the reaction mixtures were loaded onto 5% non-denaturing polyacrylamide gels (acrylamine: bisacrylamide, 37.5:1) containing 2.5% glycerol and 1×TBE buffer, and electrophoresed at 7V/cm in a cold room (4 °C).

Glycosylase/Lyase Assay

The glycosylase/lyase activity of NEIL1 protein was evaluated in 100 μL buffer (20 mM sodium phosphate, pH 7.4; 5 mM MgCl₂, 1 mM EDTA, 150 mM NaCl, 1 mM dithiothreitol, 5% glycerol, 0.01% Triton X-100, 100 ng/μL BSA) containing 0.1 nM ³²P-internally labeled DNA substrate, the indicated concentrations of NEIL1 and XPC-RAD23B. Aliquots of the reaction mixtures were remove at fixed periods of time (5 s to 30 min) and quenched by the addition of 5 μL of denaturing loading buffer (80% formamide, 0.025% xylene cyanol, 0.025% bromphenol blue in 1×TBE buffer). The samples were then subjected to gel electrophoresis using 12% denaturing polyacrylamide gels in 1×TBE buffer for 2 h.

Analysis of Gel Images

The gels were dried, exposed to Molecular Dynamics Storage Phosphor Screens, and then scanned with a Typhoon FLA 9000 laser scanner. The gel autoradiographs were analysed using Total Lab TL120 1D software. The kinetic parameters (rates constants and their ratios) were obtained from the least-squares fits of the kinetic equations to the experimental data points using the DynaFit4 software package.²⁴

RESULTS AND DISCUSSION

Competitive Binding of XPC-RAD23B and NEIL1 Proteins to Hydantoin Lesions

The binding of NEIL1 and XPC-RAD23B proteins to 147 bp ³²P-internally labeled DNA duplexes containing single site-specifically positioned Sp and Gh lesions was monitored by standard EMSA methods.²³ Complexes of XPC-RAD23B and NEIL1 with DNA substrates can be easily resolved since the molecular weights of XPC-RAD23B and NEIL1 are 151.1³⁴ and 43.7 kDa,¹⁷ respectively.

Titrations of the 147 bp DNA substrates with single Sp or Gh lesions by XPC-RAD23B and NEIL1 proteins were performed in the presence of circular covalently closed pBR322 plasmid used as nonspecific competitor DNA²³ in order to minimize the formation of multiple XPC-DNA³⁴ and NEIL1-DNA³⁷ complexes. Figure 2 shows the ranges of NEIL1 and XPC protein concentrations that exhibit competitive binding phenomena. Typical autoradiographs of native gels demonstrate that addition of XPC-RAD23B (Figure 2A) or NEIL1 (Figure 2B) to 147 bp ³²P-internally labeled SpDNA duplexes result in the appearance of bands due to protein-DNA complexes that migrate more slowly than free Sp-modified DNA molecules.

Figure 2.

Competitive binding of XPC-RAD23B and NEIL1 to single hydantoin lesions in double-stranded DNA to 147 bp ³²P-internally labeled SpDNA and GhDNA duplexes in the presence of nonspecific pBR322 DNA competitor. (A) Binding of XPC-RAD23B to SpDNA; (B) Binding of NEIL1 to SpDNA; competitive binding of the premixed NEIL1 and XPC-RAD23B to SpDNA (C) or GhDNA (D). [SpDNA] or [GhDNA] concentrations, 0.1 nM, [pBR322] concentration, 1.8 nM.

The highest mobility protein-DNA complexes correspond to either one XPC-RAD23B molecule (XPC-SpDNA, Figure 2A) or one NEIL1 molecule (NEIL1-SpDNA, Figure 2B) bound per DNA molecule that are observed at the lower protein-concentration ranges. At the higher XPC concentrations, slower migrating and weaker intensity bands are observed that are attributed to dimeric (XPC)₂-SpDNA complexes³⁴ (lanes 6, 7, Figure 2A). In the case of NEIL1, formation of the DNA complexes with two NEIL1 molecules (NEIL1)₂-SpDNA has also been previously reported³⁷ and occurs at lower protein concentration (Figure 2B, lanes 3 – 7).

In contrast to XPC, which reversibly binds to SpDNA, the binding of NEIL1 to its substrates is not reversible since NEIL1 promptly catalyzes the incision of the DNA strands at the site of the lesions via its glycosylase/lyase activity.¹² To verify this conclusion, the NEIL1-SpDNA and (NEIL 1)₂-SpDNA bands were excised from a native gel (Figure S1A in Supporting Information), and the ³²P-labeled DNA was recovered. Analysis of these oligonucleotides strands by denaturing PAGE demonstrated that the Sp-modified strands are cleaved by NEIL1, resulting in the formation of 65 nt cleavage fragments as shown in Figure S1B in Supporting Information. Thus, the NEIL1 ³²P-labeled complexes resolved by native gel EMSA methods contain DNA duplexes in which the Sp lesion-bearing strands are incised by the NEIL1 glycosylase/lyase activity.

Based on fundamental principles, the relationship between the electrophoretic mobilities μ_M of charged particles and their molar mass M, μ_M ∝ M^S, was derived,³⁸ where S is a theoretical parameter with a value between 1/3 and 2/3. The corresponding linearized form of this relationship, Log(μ_M) ∝ S × Log(M) can be successfully employed to estimate molar masses of certain protein-DNA complexes from gel electrophoresis experiments.²³ In order to verify that the (XPC)₂-SpDNA and (NEIL1)₂-SpDNA) complexes truly contain two protein molecules each, we plotted the Log(μ_M) values (deduced from Figures 2A and 2B) of all four protein duplexes (XPC-SpDNA, (XPC)₂-SpDNA, NEIL1-SpDNA, (NEIL1)₂-SpDNA) as a function of Log(M). As shown in Supporting Information (Figure S2), the excellent fit of all four data points to the same straight line is consistent with the assignments of the monomeric and dimeric protein - DNA complexes shown in Figure 2.

The formation of protein-DNA complexes at different proportions of NEIL1 and XPC in pre-formed solutions of these proteins is depicted in Figures 2C and 2D. At a constant [NEIL1] = 3.8 nM concentration and variable XPC concentrations (0.4 nM to 7.8 nM), NEIL1-SpDNA complexes are dominant in the lower range of XPC concentrations in the case of [XPC]/[NEIL1] = 0.1 (Figure 2C, lane 2). At [XPC]/[NEIL1] concentration ratios of 0.2 – 0.5, the NEIL1 protein is gradually displaced by XPC (lanes 3 – 5) and minor traces are observed at the higher ratios (lanes 7 and 8, Figure 2C). Similar results were obtained in the experiments with 147 bp ³²P-internally labeled GhDNA in pre-formed NEIL1 and XPC-RAD23B solutions at constant [XPC] = 4.0 nM and [XPC] = 0.2 – 6.3 nM (Figure 2D).

In summary, non-denaturing gel electrophoresis analysis reveals that XPC-RAD23B and NEIL1 proteins bind competitively to DNA duplexes that harbor hydantoin lesions to form non-covalent complexes either with intact damaged DNA (in the case of XPC-RAD23B) or with the product duplexes, in which the lesions were excised by NEIL1 glycosylase/lyase activities. In the next section, employing denaturing gel electrophoresis analysis, we demonstrate that non-covalent binding to XPC-RAD23B inhibit the BER activity of NEIL1.

Inhibition of NEIL1 BER activity by XPC-RAD23B

Aliquots of pre-formed solutions of NEIL1 and XPC-RAD23B were added to solutions of ³²P-internally labeled 147-mer DNA duplexes with single Sp lesions, or unmodified 147-mer control duplexes containing guanine instead of Sp. Since NEIL1 is a bifunctional glycosylase, it excises the hydantoin lesions and subsequently cleaves the incised strand via its lyase activity.¹² After incubation of solutions containing NEIL1 and hydantoin-DNA substrates, the reactions were terminated by adding a denaturating loading buffer.¹² The fractions of incisions were monitored by denaturing PAGE methods. Typical autoradiographs of denaturing gels showing the reaction products detected after incubation of 147 bp ³²P-internally GhDNA duplexes with NEIL1, are shown in Figure 3A.

Figure 3.

Representative autoradiographs of denaturing gel assays of cleavage of 147 bp ³²P-internally labeled GhDNA duplexes (0.1 nM) induced by NEIL1 (5.8 nM) in the absence (A) or in the presence (B) of 4.5 nM XPC-RAD23B. Analogous results were obtained with SpDNA duplexes (data not shown).

Under single-turnover conditions ([NEIL1] >> [GhDNA or SpDNA]), double incisions of the GhDNA strands and the formation of 65 nt oligonucleotide cleavage products, are observed within 5 – 10 s after adding NEIL1 (Figure 3A). The kinetics of NEIL1 – induced incisions of GhDNA and SpDNA strands catalysed by NEIL1 are shown in Figure 4A.

Figure 4.

Effect of XPC-RAD23B on the formation of incision cleavage products in 147 bp ³²P-internally labeled GhDNA duplexes (0.1 nM) at different ratios of [XPC]₀/[NEIL1]₀ proteins. (Averages of three independent experiments are shown). The data points (all error bars ≤ 10%) represent concentrations of total cleavage products [P] (arising from dissociated and undissociated NEIL1-Gh-DNA complexes) and intact DNA substrates, [GhDNA_int]. (A) [hNEIL1] = 5.8 nM, (B) 5.3 nM, and (C) 6.3 nM. All parameters are defined in Schemes 1 and 2 in the text. The solid lines represent the computed values of [P] and [GhDNA_int] = [GhDNA] + [XPC-GhDNA] from numerical least-squares fits of the kinetic parameters (Scheme 2) to the experimental data points. Fixed parameters: k₁ = 1×10⁸ M⁻¹s⁻¹; optimized parameters with standard errors representing 95% confidence intervals: k_X/k_N = 0.78±0.04 (B), and 0.94±0.5 (C); k_-X = (2.8±0.3)×10⁻³ s⁻¹ (B), and (3.2±0.3)×10⁻³ s⁻¹ (C).

As expected, only the burst phase is evident (yield of cleavage products > 95%) in the case of single turnover conditions ([NEIL1 >> [DNA])¹² (Figure 4A). In the presence of nanomolar concentrations of XPC, the amplitude of the burst phase is reduced to ~ 25 – 50%, depending on the XPC concentration (Figures 4B, C).

In the case of SpDNA duplexes, addition of 4.8 nM XPC reduces the burst phase amplitude to ~62% (Figure 5), while in the case of GhDNA the burst phase is reduced to ~53% under similar conditions (Figure 4B).

Figure 5.

Effect of XPC-RAD23B on BER incisions of 147 bp ³²P-internally labeled SpDNA duplexes (0.1 nM) induced by NEIL1 (5.2 nM). The data points represent the total concentrations of [P] and intact substrate, [SpDNA_int]; all error bars ≤ 10%). The solid lines represent the computed numerical least-squares fits of the kinetic parameters (Scheme 2) to the experimental data points. Fixed parameter k₁ = 1×10⁸ M⁻¹s⁻¹; optimized parameters (standard errors representing 95% confidence intervals): k_X/k_N = 0.66±0.02, k_−X = (8.8±0.8)×10⁻⁴ s⁻¹.

Overall, these results can be understood in terms of a competitive binding of XPC and NEIL1 to the same DNA lesions, and the inhibition of NEIL1-catalysed incisions by the non-covalent binding of XPC to the SpDNA and XPC-GhDNA molecules.

The general mechanism of excisions of hydantoin lesions catalysed by NEIL1 was extensively investigated by Krishnamurthy et al. based on Scheme 1:¹²

Scheme 1.

This mechanism includes the following steps: substrate binding to form non-covalent complexes (rate constants k₁, k₋₁), cleavage of the glycosidic bond with the formation of an apurinic site (k_G), cleavage of the apurinic site by the NEIL1 lyase activity (k_L), and release of NEIL1 from the product P (k₃) thus completing the enzymatic cycle.³⁵ The slow escape of the protein from the product complex is typical for BER enzymes, for instance, adenine glycosylase MutY.³⁵

In the case of Sp and Gh hydantoin lesions (H in HDNA, Scheme 1), the lyase activity of NEIL1 is very high and the formation of cleavage products can be considered as a one-step reaction with the observed catalytic rate constant, k₂.¹² Experiments under single-turnover conditions ([NEIL1] >> [HDNA]) and high [NEIL1] ~ 400 nM showed that the strand cleavage by the combined glycosylase/lyase activities becomes a rate-determining step; in DNA duplexes containing C opposite S-Sp and Gh, the values of k₂ are equal to 3.0 and 2.3 s⁻¹, respectively.¹²

In our single-turnover experiments at NEIL1 concentrations of ~ 5 nM and DNA substrate concentrations of 0.1 nM, the strand cleavage kinetics can be described by a pseudo-second order rate constant (frequently called a specificity constant), k_N = k₁k₂/(k_-1 + k₂) = k₂/K_m, where K_m is the Michaelis constant, K_m = (k₋₁ + k₂)/k₁.³⁵ However, only two parameters (k₂, K_m) among the three (k₁, k₋₁, and k₂) can be unambiguously determined from the experimental data.³⁹ The simplest solution of this problem is based on the irreversible Van Slyke-Cullen mechanism,⁴⁰ in which substrate binding occurs with an apparent bimolecular association constant that is described by the specificity constant, k_N = k₂/K_m. In other words, this approach provides the lowest value of k₁ of the Michaelis – Menten mechanism (Scheme 1).³⁹

The biphasic kinetics of excision of the Sp and Gh lesions by NEIL1, as well as the inhibitory effect of XPC, can be summarized by the following steps:

This scheme represents the minimum number of steps, that can account for the experimental observations in Figures 4 and 5. The ratios of the kinetic parameters k_N and k_X can be estimated by numerical analysis of the kinetics of the reactions.^{24, 39} According to this mechanism, the strand cleavage induced by NEIL1 is an irreversible bimolecular reaction that occurs with a rate proportional to the rate constant k_N. Our experiments showed that at NEIL1 concentrations of ~ 5 nM, the rate of cleavage occurs on a time scale faster than the experimental manual mixing time, τ_m~ 5 – 10 s. The lower limit of the value of k_N can be estimated as k_N = 1/(τ_m[NEIL1]₀) τ (3 – 5)×10⁷ M⁻¹s⁻¹. In turn, the upper limit is k_N = k₂/[NEIL1]₀ ~ 5×10⁸ M⁻¹s⁻¹ with k₂ ~ 2 – 3 s⁻¹, the experimentally measured values for Sp/Gh lesions.¹² If k_N ≥ 5×10⁸ M⁻¹s⁻¹, the reaction rate-limiting step will controlled by the first order rate constant, k₂. The recycling step (rate constant k₃) can be excluded from Scheme 1 because the denaturing gel assays provide a measure of the total concentrations of incision products (P = dissociated + undissociated NEIL1-P complexes).

The gradual reduction of the burst phase amplitude as a function of increasing XPC concentration in Figure 4 indicates that the formation of XPC-HDNA complexes competes with the binding of NEIL1 and the subsequent strand cleavage reaction (Scheme 2). Under single turnover conditions, the effect of XPC concentration on the reciprocal burst phase amplitude can be expressed by the following equation,

$$1/A_{\text{b}}=1+\left(k_{\text{X}}/k_{\text{N}}\right)\left({[\text{XPC}]}_0/{[\text{NEIL}1]}_0\right)$$ (1)

where A_b is the burst phase amplitude estimated by a linear extrapolation of the data points of the slow kinetic phase to the t = 0 time point (Figures 4 and 5); [NEIL1]₀ and [XPC]₀ are the total protein concentrations (Figure 4). Equation 1 predicts that the value of 1/A_b should increase linearly as a function of the [XPC]₀/[NEIL1]₀ ratio. The experimental results shown in Figure 6 are consistent with this prediction and the slope yields the ratio k_X/k_N = 0.81±0.04.

Scheme 2.

Figure 6.

Effect of XPC-RAD23B on the burst amplitude of the incisions induced by NEIL1 in 147 bp ³²P-internally labeled 5-guanidinohydantoin DNA duplexes (GhDNA, 0.1 nM). The solid line represents the numerical least-squares fit of the slope, k_X/k_N = 0.81±0.04 (eq 1) with standard errors representing 95% confidence intervals obtained in six independent experiments with the weighted standard errors shown by the error bars.

This value is in very good agreement with the results obtained by the analysis of the NEIL1 incision kinetics by the DynaFit4 program (Table 1).

Table 1.

Kinetic parameters for reactions of XPC-RAD23B and hNEIL1 proteins with 147 bp double-stranded DNA substrates containing the NEIL1 substrates Sp and Gh according to Scheme 2.

DNA substrate	kX/kN	k-X, S−1
GhDNAa	0.81±0.10	(2.9±0.3)×10−3
SpDNAb	0.66±0.02	(8.8±0.8)×10−4

Standard errors for data values obtained in six^a or three^b independent experiments represent 95% confidence intervals.

The slow kinetic phase of product formation is controlled by the dissociation of the XPC-HDNA complex (k_−X) followed by the prompt and irreversible strand cleavage reaction under single turnover conditions. The value of k_−X was obtained from the numerical analysis of the experimental kinetic data using the DynaFit4 program.^{24, 39} The numerical values of k_-X and k_X/k_N thus are consistent with the model (Scheme 2) and the experimental data. In this analysis, a constant value of k_N = 1×10⁸ M⁻¹s⁻¹ was assumed that is consistent with diffusion-controlled bimolecular association rate constants (10⁸ – 10⁹ M⁻¹s⁻¹).^{39, 41} The time course of the incision reactions calculated by the DynaFit4 program using the indicated kinetic parameters provide a good fit to the experimental data points (Figures 4B,C and 5). We note that variations of k_N within the range of 3×10⁷ to 3×10⁸ M⁻¹s⁻¹, reduces the characteristic rise time of the burst phase, but does not affect the value of k_X (Figure S3 in Supporting Information). Numeric analysis of the kinetic data suggests that complexes of XPC with SpDNA are somewhat more stable than complexes with GhDNA since the rate constant, k_−X for SpDNA is ~ 3 times smaller than in the case of GhDNA.

Effects of Nonspecific Competitor DNA on the Excision of Gh and Sp DNA Lesions by NEIL1

In cells, the repair of DNA lesions occurs in the context of genomic DNA. NEIL1 is known to bind non-specifically to unmodified DNA that reduces its efficiency to excise DNA lesion substrates.⁴² The effects of unmodified 147-mer DNA duplexes (umDNA) on NEIL1 – catalyzed excision of hydantoin lesions are shown in Figure 7.

Figure 7.

Effect of different concentrations of 147 bp unmodified DNA duplexes (umDNA) on the formation of incisions of 147 bp ³²P-internally labeled GhDNA duplexes (0.1 nM) initiated by NEIL1 (4.2 nM). The data points (average values derived from three independent experiments; the error bars (≤ 10%) are not shown). Concentrations of cleaved products [P], and intact DNA duplexes [GhDNA_int] are shown. The solid lines represent numerical least-squares fits of the kinetic parameters (Scheme 3) to the experimental data points using k₁ = 1×10⁸ M⁻¹s⁻¹ as a fixed parameter, and optimized parameters with standard errors representing 95% confidence intervals: (A) k_umD/k_N = 0.91±0.02, (B) 1.14±0.05, and (C) 0.83±0.06 and (A) k_-umD = (7.1±0.6)×10⁻⁴ s⁻¹, (B) (10.0±0.1)×10⁻⁴ s⁻¹, and (C) (8.8±0.2)×10⁻⁴ s⁻¹.

Figure 7 shows that in the presence of an excess of umDNA, the amplitude of the burst phase associated with the NEIL1-induced incision of GhDNA duplexes is gradually reduced to ~ 10% as the [umDNA]₀/[NEIL1]₀ ratio is increased. Furthermore, the burst phase is followed by a slower time-resolved kinetic component as in the case of the XPC competitor experiments (Figure 4).

The biphasic kinetics observed in the presence of umDNA can be described by Scheme 3.

Scheme 3.

In this scheme, H = Gh, and the inhibition effect of umDNA is described by three kinetic parameters (k_N, k_umD, and k_−umD) that is similar to Scheme 2 that describes the inhibition of NEIL1 incision activity by XPC-RAD23B.

Figure 8 shows that the reciprocal burst phase amplitude increases linearly with increasing [umDNA]₀/[NEIL1]₀ ratios ≥ 1.5 that is consistent with the equation

$$1/A_{\text{b}}=1+\left(k_{\text{um}}/k_{\text{N}}\right)\left({[\mathrm{umDN}\text{A}]}_0/{[\text{NEIL}1]}_0\right)$$ (2)

which is similar to eq. 1 for the inhibition effect of XPC-RAD23B. The ratio k_umD/k_N = 0.95±0.05 at [umDNA]₀/[NEIL1]₀ ≥ 1.5 was determined from the slope of the linear portion of the plot shown in Figure 8.

Figure 8.

Effect of umDNA on the burst amplitude of incisions induced by NEIL1 (4.2 nM) in 147 bp ³²P-internally labeled 5-guanidinohydantoin DNA duplexes (GhDNA, 0.1 nM). The solid line represents a numerical least-squares fit of eq. 2 to the experimental data points in the range 1.5 ≤ [umDNA]₀/[NEIL1]₀ ≤ 12 (averages of six independent experiments) with k_-umD/k_N = 0.95±0.05; in the range 0 ≤ [umDNA]₀/[NEIL1]₀ ≤ 1.5 the solid line was calculated using Dyanfit4 with [NEIL1] = 4.2 nM, k_N = 1×10⁸ M⁻¹s⁻¹, k_umD/k_N = 0.95 and k_-umD = 1×10⁻³ s⁻¹ (Scheme 3).

At [umDNA]₀/[NEIL1]₀ ratios < 1, in agreement with experiment, the DynaFit4 analysis according to Scheme 3, predicts that the dependence of 1/A_b vs [umDNA]₀/[NEIL1]₀ is nonlinear at low umDNA concentrations because the small changes in NEIL1 concentraions are insufficient to substantially affect the magnitude of the burst phase (solid line in Figure 8). By contrast, the inhibition of NEIL1-catalyzed strand cleavage by XPC-RAD23B is linear at all values of [XPC] >>0 (Figure 6) as predicted by eq. 1. In turn, at [umDNA]₀/[NEIL1]₀ ratios >> 1 the slope of the 1/A_b vs [umDNA]₀/[NEIL1]₀ plots predicted by Scheme 3 is proportional to k_um/k_N ratio (Figure S4, Supporting Information).

Further analysis of the inhibitory effect of unmodified DNA on the BER activity of NEIL1 was performed using the DynaFit4 software. In summary, the averaged values of the parameters obtained in six independent experiments using DynaFit4 are the following: k_umD/k_N = 0.98±0.15, and the dissociation rate constant of NEIL1-umDNA complexes is k_−umD = (7.8±0.9)×10⁻⁴ s⁻. Thus, the binding of NEIL1 to duplexes containing hydantoins lesions or to unmodified DNA duplexes occurs with similar rate constants. A similar effect has been reported for EcoRV restriction endonucleases that bind with the identical association rate constants to DNA molecules that either contained or lacked the EcoRV recognition sites.^{43, 44}

CONCLUSIONS AND PERSPECTIVES

Monitoring the glycosylase/lyase activity of the bifunctional DNA glycosylase NEIL1 we have demonstrated that the DNA damage-sensing NER factor XPC reduces the rates of incisions of the double-stranded DNA substrates containing the hydantoin Gh or Sp lesions. Numerical analysis of the kinetic data indicates that both NEIL1 and XPC proteins bind rapidly to the GhDNA or SpDNA substrates with rate constants close to the diffusion limit for bimolecular association rate constants of other proteins.^{39, 41} Thus, the preliminary partitioning of binding of NEIL1 and XPC to Gh/SpDNA hydantoin DNA lesions is determined by free diffusion mechanisms. The lifetimes of XPC-RAD23B complexes calculated from the dissociation constants (Table 1) ~6 min (GhDNA) and ~19 min (SpDNA). In comparison, the lifetimes of XPC-RAD23B complexes with UV-damaged DNA obtained by EMSA methods are 20 – 30 min.⁴⁵ We conclude that, at cellular levels, similar competitive processes between the NER XPC and BER NEIL1 proteins may play a role in the 5 – 10 times more efficient repair of hydantoin lesions by BER than by NER mechanisms in intact cells.²² In addition, the competing non-specific binding of XPC and NEIL1 to DNA lesions in intact cells significantly diminishes their removal by both NER and BER mechanisms and explains why high concentrations of these proteins are found in human cells. In human mesothelial cells, the nuclear concentrations of NEIL1 are in the range of 250 – 800 nM,^{42, 46} while the concentrations of XPC in human fibroblasts were reported to be 140 nM.⁴⁷

In the chromatin environment of human cells, access of the DNA to repair proteins is hindered, and thus the efficiencies of NER and BER mechanisms are reduced.^{48, 49} It has been shown that in nucleosomes, the efficiency of NER in human cell extracts is significantly lower than in the case of free DNA.^{32, 50} In general, BER proteins are smaller in size than the NER-initiating factor XPC-RAD23B, but BER excision activities are also reduced in nucleosomes.⁴² Furthermore, chromatin remodelling factors can lead to an enhancement of NER and BER activities in intact cells that facilitate access of the DNA lesions to proteins of both pathways.⁵¹ Finally, it has been shown that centrin-2 binds to XPC and enhances NER dual incision activities of UV photolesions in cell-free extracts.^{52, 53} It remains to be seen whether the BER-NER protein binding competition reported here in vitro for oxidatively generated guanine lesions, will also be observable in the more complex environments of human cells.

ABBREVIATIONS

BER

base excision repair

NER

nucleotide excision repair

8-oxoG

8-oxo-7,8-dihydroguanine Gh, 5-guanidinohydantoin

Sp

spiroiminodihydantoin

bp

base pair

EMSA

electrophoretic mobility shift assay

PAGE

polyacrylamide gel electrophoresis