Requirement for XLF/Cernunnos in alignment-based gap filling by DNA polymerases lambda and mu for nonhomologous end joining in human whole-cell extracts

Abstract

XLF/Cernunnos is a core protein of the nonhomologous end-joining pathway of DNA double-strand break repair. To better define the role of Cernunnos in end joining, whole-cell extracts were prepared from Cernunnos-deficient human cells. These extracts effected little joining of DNA ends with cohesive 5' or 3' overhangs, and no joining at all of partially complementary 3' overhangs that required gap filling prior to ligation. Assays in which gap-filled but unligated intermediates were trapped using dideoxynucleotides revealed that there was no gap filling on aligned DSB ends in the Cernunnos-deficient extracts. Recombinant Cernunnos protein restored gap filling and end joining of partially complementary overhangs, and stimulated joining of cohesive ends more than twentyfold. XLF-dependent gap filling was nearly eliminated by immunodepletion of DNA polymerase lambda, but was restored by addition of either polymerase lambda or polymerase mu. Thus, Cernunnos is essential for gap filling by either polymerase during nonhomologous end joining, suggesting that it plays a major role in aligning the two DNA ends in the repair complex.

Figure 1.

Requirement for Cernunnos in joining of partially complementary DNA ends. (A) Internally labeled (*) substrate and head-to-tail repair product. Following alignment, single-base gap filling (bolded T's) and ligation, the repaired top strand yields a labeled 43-base AvaI/BstXI fragment. Similar intermolecular joining of two left-hand ends yields a 24-base product. (B) Complementation of end joining deficiency in Cernunnos-deficient BuS cell extracts by mixing with DNA-PK-deficient M059J extracts (1:2, 1:1 or 2:1 ratio) or by addition of Cernunnos protein (100 nM). The substrate (10 ng) was incubated in extracts for 6 hr, deproteinized and cut with BstXI and AvaI. (C) Partial cleavage of repair products with BsaHI, to yield a 13–mer. (D) Titration of end joining with various concentrations of Cernunnos, prepared in E. coli or in baculovirus-infected insect cells as indicated. (E) Quantitation of data from (D). (F) Coomassie-stained gels showing purity of His-tagged recombinant Cernunnos proteins (*) (expected M_r = 35 kDa). Cernunnos protein in (B and C) was from E. coli.

Figure 2.

Requirement for Cernunnos and polλ or polμ in alignment-based gap filling. (A) The substrate shown in Figure 1A was incubated in BuS cell extracts with or without recombinant Cernunnos for 6 h as in Figure 1, except that ddTTP was added instead of dTTP in some samples as indicated. The 16-mer band represents the gap-filled but unligated intermediate. (B) The same extracts were preincubated with beads coated with antibodies against polλ (+) or preimmune antibodies (−), then incubated with the substrate as in (A), but in the presence of all four dNTPs plus recombinant recombinant polλ (70 ng), polμ (70 ng) and/or Cernunnos as indicated. (C) Same as (B), except that all samples contained ddTTP to trap unligated intermediates. Cernunnos in these experiments was isolated from E. coli.

Figure 3.

Cernunnos dependence of joining of cohesive 5′ and 3′ overhangs. (A) A labeled plasmid (20 ng) bearing cohesive 3′ overhangs (−GTAC/−GTAC) was incubated for 6 h in BuS extracts supplemented with the indicated concentrations of Cernunnos. (B) Same as (A) except that the substrate (20 or 100 ng) had cohesive 5′ overhangs (CGGA−/TCCG−). (C) The 5′-overhang substrate (100 ng) was incubated for 6 h in extracts containing 0 or 100 nM Cernunnos and either the DNA-PK inhibitor KU-57788 (1 μM) or the ATM inhibitor KU-55933 (1 μM). Reactions without inhibitor contained an equivalent concentration of DMSO. End joining was analyzed by agarose gel electrophoresis.

Figure 4.

Effect of S→A mutations in Cernunnos on complementation of end joining. (A) Sequence context of the S245 and S251 phosphorylation sites. (B) The partially complementary substrate shown in Figure 1A was incubated for 6 h in BuS cell extracts supplemented with wild-type or mutant Cernunnos protein from E. coli, as indicated. (C) Quantitation of data from (A) and two replicate experiments. End joining for the mutant proteins was normalized to wild-type. (D) Same as (B), except that the substrate was KpnI-cut pUC19.