DNA-Protein Cross-link Formation in Nucleosome Core Particles Treated with Methyl Methanesulfonate

Abstract

N7-Methyl-2′-deoxyguanosine (MdG) is the major damage product in DNA produced by methylating agents, but it often thought to be nontoxic and nonmutagenic. MdG is chemically unstable. An abasic site (AP) is the major product produced from MdG under physiologically relevant conditions. AP formation is frequently considered to be responsible for the cytotoxic effects of MdG, but the reaction is suppressed in nucleosome core particles (NCPs). Recently, it was discovered that histone proteins form reversible DNA-protein cross-links (DPCs) with MdG in reconstituted NCPs, as well as in methyl methanesulfonate (MMS) treated cells. In this study, the formation and reactivity of MdG in MMS treated NCPs was examined at single nucleotide resolution. Sequences consisting of three or more consecutive dGs are more reactive with MMS. The efficiency and selectivity of MdG formation by MMS is largely unaffected within a NCP, although reactivity at several dGs is ~1.5–2.5-fold higher in NCPs. DPC formation from MdG (DPC_MdG) predominates over AP at all positions within the NCP. With few exceptions, DPC_MdG yield is strongly dependent upon the accessibility of the major groove containing MdG to lysine-rich histone N-terminal tails. These data indicate that histone-MdG DPC formation will depend upon DNA sequence and translational position within a NCP.

Graphical Abstract

Introduction.

N7-Methyl-2′-deoxyguanosine (MdG, Scheme 1) is the major DNA modification produced by methylating agents, including the commonly used research tool methyl methanesulfonate (MMS), and the anti-cancer agent, temozolomide (TMZ), that is used to treat glioblastoma.^1–4 MdG has been considered to be less biologically significant than N3-methyl-2′-deoxyadenosine (MdA) and O6-methyl-2′-deoxyguanosine (OmG), which are formed in lower yields, because it had not been shown to block replication or reduce polymerase fidelity.^{5, 6} Recent reports may lead to a reevaluation of MdG′s biological effects. For instance, β−2′-fluoro-7-methyl-2′-deoxyguanosine, a stable analogue of MdG, has been shown to block replication by human DNA polymerase κ and to induce human DNA polymerase η to misincorporate dA or T opposite it.⁷ MdG was also shown to form reversible DNA-protein cross-links (DPC_MdG) with histone proteins in nucleosome core particles (NCPs) and in cells treated with MMS.^{8, 9} Herein, we explore the generality of DNA-protein cross-link (DPC) formation from MdG at single nucleotide resolution in a MMS treated NCP.

Scheme 1.

Formation and consequences of methylated DNA.

MdA and OmG are formed in lower yields than MdG (Scheme 1), but are considered to be more biologically significant.^{1, 4} Thymidine is misincorporated opposite OmG during replication, resulting in G to A transitions. OmG formation is reversed by O6-methylguanine-DNA methyltransferase (MGMT), and the resistance of cells containing higher MGMT levels to TMZ is an indication of the lesion’s cytotoxicity.^10–12 MdA cytotoxicity is attributed to the lesion’s ability to block replication. In addition, MdA rapidly depurinates to produce cytotoxic abasic sites (AP). AP formation in DNA is believed to be the predominant reaction of MdG but the reaction is slow in free DNA and suppressed further in NCPs.^{3, 8} The half-life for MdG hydrolysis within a NCP is dependent upon its proximity to highly basic histone N-terminal tails, and can be >11 weeks.⁸ The slow rates for AP formation and the fact that MdG persists in cells for multiple days suggest that the recently reported promutagenic effects of MdG and DPC_MdG formation involving it could be biologically significant.^13–15 Although DPC_MdG has not been shown to induce biological effects, sensitivity to their potential significance is heightened by the recent discoveries of proteases that recognize and digest DPCs in a manner that is independent of DNA sequence.^16–20

Previous studies were carried out using NCPs containing MdG at specific positions or within cells treated with MMS.^{8, 9} Experiments in cells did not provide information concerning the effect of MdG location within the NCP on DPC formation. However, the limited number of positions within a NCP at which MdG was selectively introduced clearly showed that the juxtaposition of the histone tails and alkylated nucleotide was important for controlling DPC formation. In this study we used MMS to more globally determine MdG and DPC formation efficiency as a function of position as a means for providing insight into the effect of NCP structure on the generation of these products.

Experimental Procedures.

General Methods.

Oligonucleotides used for polymerase chain reaction were purchased from Integrated DNA Technologies. Oligonucleotides for preparing 145-mer 601 DNA were synthesized on an Applied Biosystems Incorporated 394 oligonucleotide synthesizer. The oligonucleotide synthesis reagents were purchased from Glen-Research. T4 polynucleotide kinase (PNK), terminal transferase and proteinase K were from New England Biolabs. γ−³²P-ATP and α−³²P-cordycepin triphosphate were purchased from Perkin Elmer. MMS was from Sigma. All experiments were performed in clear siliconized tubes from Bio Plas Incorporated. Quantification of radio-labeled oligonucleotides was done using a Molecular Dynamics Phosphorimager 840 equipped with ImageQuant TL software.

Preparation of 145-mer 601 DNA.

The chemically synthesized oligonucleotides were enzymatically phosphorylated at their 5′-ter-mini. The reaction mixture (30 μL) containing each oligonucleotide (Figure S1, 1.5 nmol), 1 × T4 DNA ligase buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, 10 mM DTT, 1 mM ATP) and T4 PNK (50 units) was incubated at 37 °C for 4 h, followed by incubating at 65 °C for 30 min to inactivate the T4 PNK. The phosphorylated oligonucleotides were combined with the unphosphorylated 5′-terminal oligonucleotide (Figure S1, 1.5 nmol) and the scaffolds (Figure S1, 2 nmol). The DNA mixture was heated at 95 °C for 2 min, followed by slowly cooling to room temperature (~3–4 h). T4 DNA ligase (3600 units) and additional ATP (final concentration = 2 mM) were added and the mixture (~150 μL) was incubated at 16 °C overnight. The reaction was extracted with phenol chloroform (equal volume). The aqueous solution containing DNA was carefully removed and mixed with 95% formamide (50 μL), heated at 95 °C for 3 min, and purified by 8% denaturing PAGE. The gel (20 × 16 × 0.1 cm) was run under limiting power (15 W) at room temperature until the xylene cyanol migrated to the bottom. The product band was excised, crushed, and the DNA was eluted overnight at room temperature in 3 mL elution buffer (0.2 M NaCl and 1 mM EDTA). The slurry was filtered using a polyprep-column (BioRad) and the DNA was concentrated and the buffer was exchanged with H₂O (6 rounds) using a 10K Amicon Ultra membrane at 4 °C. The DNA (~500 pmol) was stored in −80 °C until further use.

Preparation of 5′-and 3′−³²P-labeled 601 NCPs.

The 145-mer 601 DNA was ³²P-labeled at the 5′-terminus or 3′-terminus. For 5′−³²P-labeling, the reaction mixture (10 μL) containing DNA (11.5 pmol), 1 x PNK buffer (70 mM Tris-HCl, pH 7.6, 10 mM MgCl2, 5 mM DTT), γ−³²P-ATP (30 μCi) and T4 PNK (15 units) was incubated at 37 °C for 3 h. For 3′−³²P-labeling, the reaction mixture (10 μL) containing DNA (11.5 pmol), 1 x terminal transferase buffer (20 mM Tris-acetate, pH 7.9, 50 mM KOAc, 10 mM MgOAc₂), CoCl₂ (0.25 mM), β−³²P-cordycepin triphosphate (30 μCi) and terminal transferase (30 units) was incubated at 37 °C for 3 h. Free γ−³²P-ATP and α−³²P-cordycepin triphosphate were removed using Sephadex G-50 resin (1 mL). To generate the 145-bp DNA, a mixture (15 μL) containing the ³²P-labeled strand (11.5 pmol), the complementary 145-mer strand (16.1 pmol), 10 mM phosphate buffer (pH 7.2) and 100 mM NaCl was heated at 95 °C for 1 min and then incubated at room temperature for 2 h. To reconstitute NCPs, salmon sperm DNA (20 μg, ~168 pmol of 185-bp DNA) and the hybridized DNA (2 pmol, ~4.5 million cpm) were combined to a final volume of 20 μL containing 2 M NaCl and 0.1 mg/mL BSA. Histone octamer (1.2 equivalents of total DNA) was added and the mixture (~23 μL) was incubated at room temperature for 30 min before beginning a series of dilutions using the nucleosome reconstitution buffer (10 mM HEPES, pH 7.5, 1 mM EDTA, 0.1 mg/mL BSA) at room temperature. Dilution #: volume of buffer added in μL, incubation time in minutes: 1: 24, 60; 2: 12, 60; 3: 12, 60; 4: 20, 30; 5: 20, 30; 6: 40, 30; 7: 100, 30; 8: 200, 30. After the final dilution (total volume ~450 μL), precipitate (if any) was pelleted via a brief spin (1 min, 13,000 g) at 4 °C. A small aliquot (5 μL, ~30,000 cpm) was removed, mixed with 40 % sucrose (3 μL) and analyzed by 6% native PAGE. The gel (10 × 8 × 0.15 cm) was run at 4 °C under limiting power (3 W) until the bromophenol blue band migrated to the bottom (Figure S2).

Determining the methylation efficiency by MMS at N7-dG in free DNA and NCPs.

A mixture (20 μL) containing ³²P-labeled 145-bp DNA (2 pmol, ~4.5 million cpm), salmon sperm DNA (20 μg, ~168 pmol of 185-bp DNA), 2 M NaCl and 0.1 mg/mL BSA was diluted following the same procedures describe above. Free DNA (containing salmon sperm DNA) or NCP (3 × 0.1 mL, ~3 million cpm) was incubated at 37 °C for 30 min. MMS was added to a final concentration of 5 mM. The mixture was incubated at 37 °C for 1 h, followed by phenol extraction (equal volume). The aqueous solution containing the DNA was collected and ethanol precipitated (twice). The precipitated DNA was resuspended in piperidine (50 μL, 1 M), heated (90 °C, 30 min), followed by evaporating to dryness in a speed vacuum. To remove the residual piperidine, the DNA was dissolved in water (20 μL) and evaporated to dryness in a speed vacuum. Drying was repeated in 2–3 times. The DNA was dissolved in 95% formamide (5 μL) and analyzed by 10% denaturing PAGE. The gel (42 × 36 × 0.04 cm) was run at room temperature under the limiting power (50 W) until the bromophenol blue migrated to the bottom of the gel. The methylation efficiency at an individual dG (dG_n) was calculated using Eqn. 1.

$$\text{Methylation Eff}{\text{. at dG}}^{\text{n}}=\frac{{\text{Amount of cleaved DNA at dG}}^{\text{n}}}{\text{Amount of total DNA}}\times 100\%$$ (Eqn. 1)

Determining the DPC_MdG and DPC_AP yields in MMS-treated NCPs.

Native 601 NCPs with ³²P labeling at the 5′- or 3′-terminus of the top (1–145) or bottom (146–290) strand, respectively, were reconstituted using the salt dilution method as described above. NCPs (3 × 90 μL, ~3 million cpm) were incubated at 37 °C for 30 min. MMS was added to a final concentration of 5 mM. The mixture was incubated at 37 °C for 1 h. The unreacted MMS was removed by extensive buffer (10 mM HEPES, pH 7.5, 1 mM EDTA, 0.1 mg/mL BSA and 90 mM NaCl) exchange (6 rounds) using a 10 K Amicon Ultra membrane at 4 °C. The NCPs were concentrated to ~35 μL. A small aliquot (1 μL) was removed for determining the DNA methylation efficiency as described above. The remainder of the NCP sample was incubated at 37 °C for 24 h, followed by analysis by 10% SDS PAGE. The gel (20 × 16 × 0.1 cm) was run at 4 °C under the limiting power (3 W) until the bromophenol blue migrated to the bottom of the gel. The bands containing all DPCs (DPC_Tot) were excised and the DNA within this material was eluted using 600–800 μL buffer (10 mM HEPES, pH 7.5, 200 mM NaCl, 1 mM EDTA, 0.1% SDS) containing 4 units of proteinase K at 4 °C overnight. The slurry was filtered using a polyprep-column (BioRad). Salmon sperm DNA (5 μg) was added to the solution, followed by ethanol precipitation. The DNA was resuspended in water (90 μL) and divided into two halves. To determine the DPC_Tot yield from the labeled strand, consisting of the DPCs from AP (DPC_AP) and MdG (DPC_MdG), half of the DNA was subjected to hot piperidine treatment (1 M, 90 °C, 30 min). The DNA samples were dissolved in 95% formamide (5 μL) and a portion of the sample (~30,000 cpm) was analyzed by 10% denaturing PAGE. The gel (42 × 36 × 0.04 cm) was run at room temperature under the limiting power (50 W) until the bromophenol blue migrated to the bottom of the gel. The total yield of DPCs (DPC_Tot) at an individual dG (dGⁿ) was calculated using Eqn. 2.

$${\text{Yield of DPC}}_{\text{Tot}}{\text{ at dG}}^{\text{n}}=\%\text{ DPC yield detected from SDS-PAGE }\times \frac{{\text{Amount of cleaved DNA at dG}}^{\text{n}}\text{ in piperidine treatment}}{\text{Amount of total DNA in piperidine treatment}}$$ (Eqn. 2)

To determine the amount of DPC_AP, the other half of the DNA was treated with NaOH (100 mM, 37 °C, 30 min), followed by neutralization and ethanol precipitation. The precipitated DNA was resuspended in 95% formamide (5 μL) and a portion of the sample (~30,000 cpm) was analyzed by 10% denaturing PAGE. The gel (42 × 36 × 0.04 cm) was run at room temperature under the limiting power (50 W) until the bromophenol blue migrated to the bottom of the gel. The yield of DPC_AP at dGⁿ was calculated using Eqn. 3.

$${\text{Yield of DPC}}_{\text{AP}}{\text{ at dG}}^{\text{n}}=\%\text{ DPC yield detected from SDS}-\text{PAGE }\times \frac{{\text{Amount of cleaved DNA at dG}}^{\text{n}}\text{ in NaOH treatment}}{\text{Amount of total DNA in NaOH treatment}}$$ (Eqn. 3)

The yield of DPC_MdG at dGⁿ is calculated to be the difference between DPC_Tot measured using piperidine treatment and DPC_AP measured using NaOH (Eqn. 4). The normalized yield of DPC_MdG at dGⁿ was calculated using Eqn. 5.

$${\text{Yield of DPC}}_{\text{MDG}}{\text{ at dG}}^{\text{n}}{\text{ = Yield of DPC}}_{\text{Tot}}{\text{ at dG}}_{{}^{\text{n}}}{-\text{ Yield of DPC}}_{\text{AP}}{\text{ at dG}}^{\text{n}}$$ (Eqn. 4)

$${\text{Normalized DPC}}_{\text{MdG}}{\text{ yield at dG}}^{\text{n}}=\frac{{\text{Yield of DPC}}_{\text{MdG}}{\text{ at dG}}^{\text{n}}}{{\text{Methylation efficiency at dG}}^{\text{n}}}\times 100\%$$ (Eqn. 5)

Results and Discussion.

Nucleosome Core Particle Design and Preparation.

NCPs containing the strong positioning 601 sequence and Xenopus laevis histone proteins that were expressed in E. coli were used.^21–24 The G•C content in the 601 sequence is slightly greater than 50%. The G•C base pairs are distributed throughout the core particle, including in regions that are proximal and distal to lysine rich N-terminal histone tails (Figure 1). For instance, dG⁷³, dG⁷⁵, dG²¹⁷ and dG²²⁰ are at or near the dyad axis (superhelical location (SHL) 0), and are distal from the N-terminal histone tails. In contrast, dG^36–38 (SHL −3.5), dG⁶⁵⁻⁶⁹ (SHL −0.5) and dG^243−246 (SHL −2.5) are proximal to the histone H2B, H3 and H4 tails respectively. Overall, the NCP containing 601 DNA provides a heterogeneous substrate that is desirable for probing MdG generation and subsequent DPC formation.

Figure 1.

Nucleosome core particle structure. A. Single gyre containing DNA from SHL 0 to −7. B. Single gyre containing DNA from SHL 0 to 7. The red and cyan spheres indicate the nucleotides in the top and bottom strands, respectively. C. Nucleosomal DNA sequence. The structure was generated by superimposing two NCP structures (pdb: 1kx5 and 3lz0).

Methylation by MMS in free DNA and NCPs.

Methylation was examined at single nucleotide resolution in free DNA and NCP under otherwise identical conditions. Of particular note, is that salmon sperm DNA was added to free DNA samples to maintain the same overall DNA concentration as is present in NCP samples. Overall, the absolute amount of MdG at each nucleotide and selectivity were remarkably similar in free DNA and the NCP (Figure 2, Figures S3,4). At several positions, including dG⁵², dG⁶⁵, dG⁶⁶, dG¹⁷⁷, dG²⁰⁷, dG₂₄₃ and dG₂₄₄, the methylation efficiencies are ~1.5–2.5-fold higher in NCPs than that in free DNA. Some DNA damaging molecules, such as the minor groove binding duocarmycin, exhibit decreased reactivity at DNA within the nucleosome core region.^{25, 26} However, a related report on duocarmycin reactivity in a NCP (lacking linker DNA) showed that there was no difference in reactivity compared to with free DNA.²⁷ Similar reactivity in free DNA and the NCP was observed even when the N7-position of dG²² (SHL −5) was pointed directly towards the octameric core (Figure 2, Figure S3 and Figure 5A). Perhaps the absence of specific DNA-protein interactions enables the small MMS molecule to diffuse between the major groove and the octameric core. However, this is a very different pattern than the much more reactive hydroxyl radical, which very clearly favors reacting with nucleotides whose minor grooves face away from the octameric core as opposed to towards the core.²⁸

Figure 2.

Methyl methanesulfonate alkylation of free DNA and a NCP. Data are the average ± std. dev. of 3 replicates. See Figure S10 for reaction at all dGs.

Figure 5.

Inaccessibility of the C8-position of various dGs in a NCP to histone protein N-terminal tails (A-C). Red arrows point towards the C8-position. The structures were generated by superimposing two NCP structures (pdb: 1kx5 and 3lz0).

MMS reacted preferentially at runs of dGs (≥ 3 consecutive dGs, e.g. dG³⁶⁻³⁸, dG⁶⁵⁻⁶⁹, dG^243−246), a phenomenon that has been observed when examining the reactivity of other alkylating agents with free DNA.^{29, 30} Within the runs of dGs, methylation is greater at the 5′-portions. For instance, methylation at dG⁶⁹ was almost 5-fold lower than at dG⁶⁵ in free DNA and almost 15-times less in the NCP (Figure S10). Interruption of consecutive dGs by a single nucleotide (e.g. 5′-dG⁵¹GCGG⁵⁵, SHL −2) was sufficient to reduce methylation efficiency (Figure 2).

MdG-histone DNA-protein cross-links.

DPCs in MMS treated NCPs were isolated from SDS PAGE at 4 °C (Figure S5–6). Reversible histone DPCs are formed with MdG and AP in NCPs (Scheme 2). DPC_MdG is distinguishable from DPC_AP by exploiting their differences in sensitivity to alkaline conditions.^{8, 9, 31–33} The DNA within DPC_AP is cleaved over time by the lysine residues within the proximal N-terminal histone tails.^31–35 Complete strand scission at the cross-linked AP sites is achieved upon treatment with NaOH (0.1 M, 37 °C, 30 min). The DNA within a DPC_MdG is uncleaved under these conditions. The DNA in both types of DPCs are cleaved upon hot piperidine treatment (1 M, 90 °C, 30 min). Denaturing PAGE analysis of a mixture of isolated DPCs subjected to these treatments enabled quantification of the yields of each DPC type at nucleotide resolution. The relative amounts of DPCs derived from MdG and AP vary with respect to time because each lesion undergoes an irreversible transformation, MdG to AP and AP to a strand break, respectively. In this study MMS treated NCPs were incubated for 24 h (37 °C). DPC_MdG accounted for the majority of the cross-links at all positions (Figure 3A). DPC_AP accounted for ~1/3 of the total DPCs at dG⁹³ and dG⁹⁴ (SHL 2). However, at most positions, DPC_AP accounted for <10% of the total DPCs. These data are consistent with the slow rates of MdG depurination, particularly in NCPs, resulting in low AP levels.^{3, 8}

Scheme 2.

Distinguishing between DNA-protein cross-links involving histones and MdG or AP.

Figure 3.

DNA-protein cross-link formation in MMS treated NCPs. A. Yields of DPC_MdG and DPC_AP. B. Comparison of MdG and DPC_MdG formation. C. Nucleotide resolution normalized DPC_MdG yield. Data are the average ± std. dev. of 3 replicates. See Figures S11–13 for reaction at all dGs.

NCP structure has a more profound effect on DPC_MdG yield than it does on DNA methylation (Figure 2, Figure 3). There are guanines of comparatively high methylation efficiency that form DPCs commensurately with the histone proteins. For instance, the major groove in which the C8-positions of dG₅₁, dG₅₂, dG₅₄ and dG₅₅ (SHL −2) lie are readily accessible to the histone H4 tail (Figure 4A). In contrast, there are positions of relatively high methylation efficiency (e.g. dG₆₅ - dG₆₈, SHL −0.5) that give rise to comparably modest DPC_MdG yields when the DPC yield is normalized based on methylation efficiency (Figure 3C). Inspection of the NCP structure provides insight into why this is. The respective C8-positions of dG₆₅ - dG₆₈ are oriented away from the octameric core and should be accessible to the lysine rich histone tails (Figure 4B,C). However, the region where the 20 amino acid N-terminal histone H4 tail protrudes from the core is ~2-helical turns removed from these nucleotides. The histone H3 tail is in principle much closer but it must traverse the DNA backbone to interact with the major groove containing dG₆₅ - dG₆₈.

Figure 4.

Accessibility of the C8-position of various dGs in a NCP to histone protein N-terminal tails (A-E) and core loop (F). Red arrows point towards the C8-position. The structures were generated by superimposing two NCP structures (pdb: 1kx5 and 3lz0).

Relatively moderate methylation was a necessary but insufficient requirement for DPC formation. Examination of the NCP crystal structure indicated C8-accessibility to histone tails significantly affected DPC formation. For instance, methylated dG³⁶ and dG³⁷ (SHL −3.5) are well positioned to react with the H2B tail (Figure 4D). Similarly, the methylated nucleotides within the contiguous dG sequence, dG²⁴³-dG²⁴⁵ (SHL −2.5) is also well positioned to react with histone H4 (Figure S9A) and the histone H3 tail can access MdG²²⁹ (Figure S9B). Examination of the composite structure indicates that the H2A tail is well positioned to react with alkylated dG¹⁷⁷, dG¹⁷⁸ and dG¹⁸¹ (SHL 4) (Figure S9C).

The position of highest DPC_MdG formation efficiency is at dG⁴⁰ (Figure 3C). The absolute DPC_MdG yield at this nucleotide is also one of the highest within the NCP (Figure 3B). Neither of the closest histone N-terminal tails (H2B, H4) appear to be well-positioned to react with MdG at C8 in this position (Figure 4E). However, the crystal structure indicates that a histone H4 lysine residue (Lys77) in a flexible loop region of the core is well-positioned to react with MdG at this site (Figure 4F).

Examination of the NCP structure also helps explain sites that produce DPC_MdG relatively inefficiently as well. For instance, dG²², dG²⁸ and dG⁹⁴ are amongst the top 25 guanosines that are alkylated in NCPs (Figure 2). However, DPC_MdG efficiency is modest at those sites (Figure 3B,C). The histone H2B tail is closest to the C8-position of dG²² and dG²⁸ (Figure 5A,B). However, it would need to fold over the strand containing dG²² or dG²⁸ to access the C8-position of the methylated nucleotide. In addition, the C8-position of dG²⁸ faces directly toward the octameric core which also limits the accessibility to the histone tails. Similarly, the histone H4 tail would be required to traverse the minor groove and wrap around the duplex to reach MdG at position 94 (Figure 5C).

Summary.

Monofunctional alkylating agents have been used to damage DNA for more than half a century. Molecules that carry out such chemistry are utilized to treat cancer (e.g. TMZ). DNA-protein cross-links from a monofunctional alkylated nucleotide (MdG) were discovered only recently, but their biological effects are unknown.⁸ Recent discoveries of DNA-protein repair systems,^16–20 inactivation of a base excision repair enzyme^{36, 37} and of naturally occurring DPCs that protect against DNA damage are strong indicators that such lesions are biologically significant.^{38, 39} The observations reported here indicate that alkylation of dG by a small molecule is not affected by nucleosome structure; whether this is also true within higher order chromatin remains to be seen. However, the transformation of alkylated DNA, specifically MdG, into DNA-protein cross-links will strongly depend upon position within the nucleosome and interactions with histone nucleophilic residues. It is also anticipated that within chromatin, the interactions between histone tails of one nucleosome with the damaged DNA of an adjacent one may also contribute to DNA-protein cross-link formation.^40–42

ABBREVIATIONS

AP

abasic site

MdA

N3-methyl-2′-deoxyadenosine

MdG

N7-methyl-2′-deoxyguanosine

OmG

O6-methyl-2′-deoxyguanosine

DPC

DNA-protein-cross-link

DPC_MdG

DPC between MdG and histone protein

DPC_AP

DPC between AP and histone protein

DPC_Tot

total DPCs (DPC_MdG + DPC_AP)

NCP

nucleosome core particle

SHL

superhelical location