The structure and duplex context of DNA interstrand crosslinks affects the activity of DNA polymerase η

Abstract

Several important anti-tumor agents form DNA interstrand crosslinks (ICLs), but their clinical efficiency is counteracted by multiple complex DNA repair pathways. All of these pathways require unhooking of the ICL from one strand of a DNA duplex by nucleases, followed by bypass of the unhooked ICL by translesion synthesis (TLS) polymerases. The structures of the unhooked ICLs remain unknown, yet the position of incisions and processing of the unhooked ICLs significantly influence the efficiency and fidelity of bypass by TLS polymerases. We have synthesized a panel of model unhooked nitrogen mustard ICLs to systematically investigate how the state of an unhooked ICL affects pol η activity. We find that duplex distortion induced by a crosslink plays a crucial role in translesion synthesis, and length of the duplex surrounding an unhooked ICL critically affects polymerase efficiency. We report the synthesis of a putative ICL repair intermediate that mimics the complete processing of an unhooked ICL to a single crosslinked nucleotide, and find that it provides only a minimal obstacle for DNA polymerases. Our results raise the possibility that, depending on the structure and extent of processing of an ICL, its bypass may not absolutely require TLS polymerases.

Figure 1.

ICL substrates used in this study. (A) Structure of a nitrogen mustard (NM) ICL linking two guanine bases. (B) Structure of the 5 atom ((i), 5a), 6 atom ((ii), 6a) and 8 atom ((iii), 8a) NM ICL. (C) Substrates used for polymerase assay substrates reactions. 5′-FAM labeled primer P15 was annealed to various templates. (i) single stranded DNA undamaged control (ii) single stranded substrate with ICL precursor ‘diol’ (Supplementary Figure S3B, i, C2) (iii) double stranded substrate with ICL precursor ‘diol’, (iv) ICL substrate within a 20 bp duplex, (v) ICL substrate within a 6 bp duplex (vi) single nucleotide ICL substrate. Crosslinked or adducted bases are highlighted in red.

Figure 2.

Synthesis of the single nucleotide ICL. (A) Reaction conditions: (a) OsO₄, NMM, THF, 0°C, 72%; (b) (i) NaIO₄, MeOH, THF; (ii) NH₂OMe, 86%; (c) Zn/HCl, MeOH, HOAc, 85%; (d) NH3, MeOH. (B) (e) (i) NaIO₄, H₂O (ii) NaBH₃CN. (C) 20% denaturing PAGE analysis of reaction of the 11mer (GAAAGAAG4AC) with amine 3. DNA was visualized with SYBR gold.

Figure 3.

Shortening of duplex around the ICL facilitates bypass. (A) 20 bp, 6 bp and 1 nt ICL substrates used. The crosslinked base in the template strand was designated ‘0’, and all primer extension products up to −1 were evaluated as ‘approach’, from 0 to +3 as ‘insertion’ and beyond +4 as ‘extension’. (B–E) Translesion synthesis assay of 5a ICL templates with Klenow and pol η. Unmodified (lanes 1 and 6), diol (monoadduct, lanes 2 and 7) and 5a ICL-containing templates (lanes 3–5, 8–10) were annealed to the FAM labeled primer P15 and incubated with (B) 1 nM Klenow, (D) 40 nM pol η for 10 min at 37°C. Products were resolved by 10% denaturing PAGE. Quantification and analysis of primer extension products with (C) Klenow and (E) pol η. Each lane was divided into approach, insertion and extension segments and corresponding band intensities expressed as a percentage of the total products combined. Data represent the mean of three experiments and error bars indicate S.D.

Figure 4.

Duplex distortion facilitates approach and insertion by pol η at the ICL. Control and monoadduct (lanes 1–2, 6–7) and 5a, 6a and 8a ICL templates in a (A) 20 bp duplex (lanes 3–5) or (C) 6 bp duplex (lanes 8–10) were annealed to primer P15 and incubated with 40 nM pol η for 10 mins at 37°C. The products were resolved by 10% denaturing PAGE. Quantification and analysis of primer extension products with 20 bp ICLs (B) and 6 bp ICLs (D). Each lane was divided into approach, insertion and extension segments and corresponding band intensities expressed as a percentage of all the products combined. Data represent the mean of three experiments and error bars indicate S.D.

Figure 5.

pol η is more accurate across NM ICLs than undamaged DNA. (A) Sequence of ICL substrate for single base insertion assays. (B) Templates were annealed to FAM labeled P0 primer and incubated with 20 nM pol η at 37°C for 5 min with 1, 10 or 100 μM of individual dNTPs (A/T/C/G) or all four dNTPs combined (N). The products were resolved by 10% denaturing PAGE and band intensities quantified using ImageQuant. Quantification and analysis of single nucleotide incorporation with (C) 1 μM dNTPs (D) 10 μM dNTPs and (E) 100 μM dNTPs. The ratio of band intensity for each individual nucleotide (lane A/T/C/G) to that of the four dNTPs combined (lane N) was calculated and expressed as the ‘normalized intensity’. Data are represented as the mean of three experiments and error bars indicate S.D. P-values were calculated by a one-way ANOVA with the Bonferroni correction for multiple comparisons.