Tolerance for 8-oxoguanine but not thymine glycol in alignment-based gap filling of partially complementary double-strand break ends by DNA polymerase lambda in human nuclear extracts

Abstract

Ionizing radiation induces various clustered DNA lesions, including double-strand breaks (DSBs) accompanied by nearby oxidative base damage. Previous work showed that, in HeLa nuclear extracts, DSBs with partially complementary 3' overhangs and a one-base gap in each strand are accurately rejoined, with the gaps being filled by DNA polymerase lambda. To determine the possible effect of oxidative base damage on this process, plasmid substrates were constructed containing overhangs with 8-oxoguanine or thymine glycol in base-pairing positions of 3-base (-ACG or -GTA) 3' overhangs. In this context, 8-oxoguanine was well tolerated by the end-joining machinery when present at one end of the break, but not when present at both ends. Thymine glycol was less well tolerated than 8-oxoguanine, reducing gap filling and accurate rejoining by at least 10-fold. The results suggest that complex DSBs can be accurately rejoined despite the presence of accompanying base damage, but that nonplanar bases constitute a major barrier to this process and promote error-prone joining. A chimeric DNA polymerase, in which the catalytic domain of polymerase lambda was replaced with that of polymerase beta, could not substitute for polymerase lambda in these assays, suggesting that this domain is specifically adapted for gap filling on aligned DSB ends.

Figure 1.

Catalytic activity of a chimeric DNA polymerase and interaction with core end joining proteins. (A) A 5′-³²P-end-labeled 60-base duplex (90 nM) was incubated with Ku (5 nM), X4L4 (25 nM) and 25, 50 or 100 polλ or a chimeric DNA polymerase (pol-ch) wherein the catalytic core of polλ was replaced with that of polβ (‘+’ = 100 nM). Complexes were resolved on a nondenaturing gel. The uppermost band (asterisk) is formed only when all three proteins are present, and presumably represents a stable complex of the three on DNA (11). This complex forms equally well with either polλ or pol-ch. (B) A partial duplex consisting of a 15-base primer annealed to a 33-base template (3 nM) was treated with 3.5, 7, 14, 35, 70 or 140 nM polλ or pol-ch. Following incubation for 1 h at 37°C, samples were analyzed on a sequencing gel.

Figure 2.

Gap filling and accurate end joining of a DSB substrate bearing partially complementary overhangs. (A) Schematic of the site-specifically labeled (asterisk) DSB substrate showing accurate end joining by gap filling and ligation, resulting in a 43-base BstXI/AvaI fragment. All sequences read 5′→3′ in the top strand and 3′→5′ in the bottom strand. Bolded T indicates the filled-in base. (B) The substrate shown, bearing -ACA and -ATG 3′ overhangs, was incubated for 6 h with X4L4-supplemented HeLa extracts that had been immunodepleted of polλ or mock depleted. Polλ, polμ or pol-ch (70 ng) was also added as indicated. After incubation, samples were cut with BstXI and AvaI, and analyzed on denaturing gels. (C) Same as (B), except the reaction was performed in the presence of ddTTP instead of dTTP, and only some samples contained X4L4, as indicated. ‘ch*’ indicates 280 ng of pol-ch was added. (D) Quantitation of data from (B) and from a replicate experiment; error bars show the range of values obtained in the two experiments.

Figure 3.

Effect of a single-base mismatch on end joining. (A) The 3-base 3′ overhang substrates shown were incubated in polλ-depleted extracts supplemented with various polymerases as in Figure 2, cut with BstXI and AvaI and analyzed on sequencing gels. (B) Proposed mechanisms of formation of 43- and 42-base products from the mismatched substrate. (C) Sequencing of repair joints in recircularized products. Sequences are expressed in terms of the radiolabeled (top) strand. Black and blue letters indicate sequences originating from the left and right ends of the break, respectively. Underlined nucleotides could have originated from either end and may have resulted from microhomology annealing and splicing. Overlines indicate apparent preservation of 3′ overhangs. The green letters in the last sequence begin 28 bases from the 5′ terminus of the right end. The first two sequences correspond to the two pathways shown in (B).

Figure 4.

End joining of a substrate containing an 8-oxoguanine base. (A) Schematic of the substrate and mechanism of formation of the accurate end-joining product (X = guanine or 8-oxoguanine). Bolded T indicates the filled-in base. (B) The substrates were incubated in polλ-depleted or mock-depleted nuclear extracts supplemented with polλ or polμ as indicated, cut with BstXI and AvaI, and analyzed on denaturing gels. The 43-base product corresponds to gap filling and accurate head-to-tail joining of the plasmid, as shown in (A). The 24-base product (formed only with the unmodified substrate) corresponds to accurate head-to-head joining of two plasmid molecules. The band migrating as a ∼14-mer is an apparent spontaneous degradation product of 8-oxoguanine (see text). (C) Same as (B), except that extracts were not immunodepleted, and some samples were treated with NH₄OH to cleave 8-oxoguanine residues. The lane marked ‘M’ contains a labeled 43-base marker of the expected sequence. (D) Same as (C), except that the end-joining buffer contained ddTTP instead of dTTP, in order to trap the filled-in but unligated intermediate (16-mer band), and some extracts were immunodepleted as indicated. (E) Quantitative gap-filling data from (D) and a replicate experiment. The abundance of the 16-mer band was calculated for each sample and normalized to the value obtained for the unmodified substrate in the polλ-supplemented extract. The value shown for the 8-oxoguanine substrate (asterisk) includes an apparent degradation product of the initial substrate that comigrates with the 16-mer.

Figure 5.

Generation of a thymine glycol-containing oligomer. (A) HPLC absorbance (260 nm) trace showing elution of products of oxidation of the oligomer CGAGGAACGCGGTA by OsO₄. (B) Deconvoluted electrospray mass spectrum of the earliest-eluting peak (36 min), consistent with oxidation of the single T in the sequence to thymine glycol. (C) Mass spectrum of the second-eluting peak (37.5 min), apparently containing unmodified oligomer.

Figure 6.

End joining of a substrate containing a thymine glycol base. (A) Schematic of the internally labeled (asterisk) substrate, showing proposed mechanisms of formation of products giving 43-, 42- and 35-base BstXI/AvaI fragments. The italicized T is the site of thymine glycol substitution. Bolded nucleotides indicate gap filling. The 35-base product is shown as arising by 3′→5′ resection, but it could in principle be generated by 5′→3′ resection as well. (B) Unmodified or thymine glycol-substituted substrate was incubated for 6 h with X4L4-supplemented HeLa nuclear extracts that had been immunodepleted of polλ or mock depleted. Polλ or polμ (70 ng) was added as indicated. (C) Same as (B), except extracts were not immunodepleted, polλ was added to all samples and samples contained either ddCTP or ddTTP in place of normal dNTPs as indicated. (D) Quantitative data from (B) and similar experiments; the abundance of each product is normalized to the total of all head-to-tail end-joining products in the sample with the unmodified substrate (‘T’) and added polλ; ‘T-glycol’ = thymine glycol-substituted substrate. Error bars show mean and SE of 3–4 experiments. (E) Similar quantitation of data from (C) and a replicate experiment. Error bars show the range of values obtained in the two experiments.