Mlh1-Mlh3, a meiotic crossover and DNA mismatch repair factor, is a Msh2-Msh3-stimulated endonuclease

Abstract

Crossing over between homologous chromosomes is initiated in meiotic prophase in most sexually reproducing organisms by the appearance of programmed double strand breaks throughout the genome. In Saccharomyces cerevisiae the double-strand breaks are resected to form three prime single-strand tails that primarily invade complementary sequences in unbroken homologs. These invasion intermediates are converted into double Holliday junctions and then resolved into crossovers that facilitate homolog segregation during Meiosis I. Work in yeast suggests that Msh4-Msh5 stabilizes invasion intermediates and double Holliday junctions, which are resolved into crossovers in steps requiring Sgs1 helicase, Exo1, and a putative endonuclease activity encoded by the DNA mismatch repair factor Mlh1-Mlh3. We purified Mlh1-Mlh3 and showed that it is a metal-dependent and Msh2-Msh3-stimulated endonuclease that makes single-strand breaks in supercoiled DNA. These observations support a direct role for an Mlh1-Mlh3 endonuclease activity in resolving recombination intermediates and in DNA mismatch repair.

Keywords: Crossing Over; DNA Enzymes; DNA Mismatch Repair; DNA Recombination; DNA Repair; Endonuclease; Meiosis; Mlh1-Mlh3.

FIGURE 1.

Purification of the Mlh1-Mlh3 complex. A, organization of Mlh3 including the conserved ATP binding domain (dark blue boxes) in the highly conserved N-terminal domain (cyan), the 117-amino acid predicted linker domain (gray box), the Mlh1 interaction domain in the less well conserved C-terminal domain of Mlh3 (blue box), and residues within interaction domain that form part of the endonuclease active site (red box). The aspartic acid residue at amino acid 523 that was mutated to asparagine (D523N) is indicated with an arrow. B, left panel, schematic of the Mlh1-Mlh3 complex based on crystal structures (8, 9, 57, 69, 70), including predicted lengths of linker arms (144 amino acids for Mlh1, 117 amino acids for Mlh3) and location of His₁₀, HA epitope, and FLAG epitope insertions. In this model Mlh1 and Mlh3 are each composed of two globular domains connected by a flexible linker, and Mlh1 and Mlh3 interact though their C-terminal domains. The N-terminal ATP binding domains of both proteins are highly conserved among the MutL homologs and undergo ATP-dependent conformational changes. Mlh1 is in green, and Mlh3 is in blue. Right panel, SDS-PAGE analysis of purified Mlh1-Mlh3 (WT) and Mlh1-mlh3-D523N (DN). Coomassie Blue R250-stained 8% Tris-glycine gel: 0.5 μg of Mlh1-mlh3-D523N; 0.5 μg of Mlh1-Mlh3; mw = molecular mass standards from top to bottom: 200, 116, 97, 66, 45, 31 kDa. C, mass spectrometry analysis of bands detected after Q-Sepharose chromatography.

FIGURE 2.

Mlh1-Mlh3 displays an endonuclease activity on supercoiled DNA substrates. See “Experimental Procedures” for details. A, Mlh1-Mlh3 (100, 150, 200, 250, and 300 nm) was incubated with linearized pUC19 DNA, resolved, and analyzed as described under “Experimental Procedures.” Ladder, 1 kb of DNA ladder (New England Biolabs). B, comparison of the nicking activity of Mlh1-Mlh3 (wt) and Mlh1-mlh3-D523N (D523N) in the presence of 1 mm Mn²⁺ on supercoiled pUC19. Mlh1-Mlh3 is at 150 nm in lanes 3 and 4. In lanes 5–9, Mlh1-Mlh3 is at 50, 150, 200, 250, and 300 nm, respectively. In lanes 10 and 11 Mlh1-Mlh3 D523N is at 150 and 300 nm, respectively. Migration of supercoiled (sc) and nicked (n) DNA is indicated. The linear product is indicated by a black triangle. C, immunodepletion analysis of Mlh1(FLAG)-Mlh3(HA). Mlh1(FLAG)-Mlh3(HA) was incubated with either Protein A-linked mouse anti-FLAG antibodies (anti-FLAG) or control mouse IgG (Control). The immunodepleted supernatants were then assayed for endonuclease activity. D, Mlh1-Mlh3 endonuclease activity is stimulated by a variety of divalent cations. Nicking activity of Mlh1-Mlh3 (150 nm) in the presence of 1 mm Mg²⁺, Mn²⁺, Zn²⁺, Cd²⁺, Co²⁺, Ca²⁺, or Ni²⁺ is indicated (left). Stimulation of the Mlh1-Mlh3 endonuclease activity in the presence of 1 mm Mn²⁺ by a second divalent metal ion (1 mm) (right) is shown.

FIGURE 3.

Mlh1-Mlh3 endonuclease activity is unaffected by ATP. A, endonuclease activity of Mlh1-Mlh3 (120 nm) on supercoiled (sc) pUC19 DNA in the presence of various concentrations of ATP. + indicates an ATP concentration of 500 μm, and in lanes 8–10 ATP concentrations were 30, 100, 500 μm, respectively. Mg²⁺ and Mn²⁺ are present at 1 mm as indicated. The % of supercoiled substrate that was cleaved is indicated. n, nicked. B, endonuclease assays were performed in the same reactions as in A but contained 1 mm Mg²⁺ (Mn²⁺ is not present) and 0.5 mm ATP or GTP. nt, nucleotide. C, ATPase activity of Mlh1-Mlh3 in the presence of a 10-fold excess of single-stranded (ssDNA) and double-stranded (dsDNA) oligonucleotides or supercoiled plasmid. See “Experimental Procedures” for details. D, endonuclease activity of yMlh1-Mlh3 was not stimulated by yeast RFC (yRFC) or yeast PCNA (yPCNA). Lanes are from 1–17, left to right. In lanes 1–8, hMlh1-Pms2 (60 nm) was incubated with RFC (100 nm) and PCNA (100 nm) with and without 500 μm ATP. In lanes 9–16, yMlh1-Mlh3 (120 nm) was incubated with RFC (200 nm) and PCNA (200 nm) as indicated with and without 500 μm ATP. Lane 17 contains 200 nm RFC, 200 nm PCNA, and 500 μm ATP in the absence of yMlh1-Mlh3.

FIGURE 4.

Mlh1-Mlh3 endonuclease activity on supercoiled DNA is stimulated by Msh2-Msh3. A, electromobility shift assays were performed as described under “Experimental Procedures.” Mlh1-Mlh3 (100 nm) and Msh2-Msh3 (80 nm) were incubated with the 50 nm 5′-³²P-end-labeled 40-bp +8 loop containing substrate in the presence of 1 mm ATP, 2 mm MgCl₂, or 2 mm MnSO₄ as indicated. A representative assay for five trials is shown. B, nicking (n) activities of yeast and human Mlh1-Mlh3 (120 nm) were measured in the presence of 1 mm Mn²⁺ and yeast Msh2-Msh3 (120 nm) as indicated. The % of supercoiled substrate (sc) that was cleaved is shown ± S.D. C, endonuclease assays were performed as in B but contained 0–500 μm ATP, 120 nm Mlh1-Mlh3, and 120 nm Msh2-Msh3 when indicated. Reactions were performed in triplicate, samples were resolved on agarose gels, and the percent of DNA that was nicked is plotted. D, endonuclease-deficient yeast Mlh1-Mlh3 is not stimulated by yeast Msh2-Msh3. Nicking activity of yMlh1-Mlh3 and yMlh1-mlh3-D523N (each at 120 nm) was measured in the presence of 1 mm Mn²⁺, 0.5 mm ATP, and yMsh2-Msh3 (120 nm) as indicated. The % of supercoiled substrate that was cleaved is shown.

FIGURE 5.

Mlh1-Mlh3 binds preferentially to the +8 loop substrate. A, Mlh1-Mlh3 binding to a ³²P-end-labeled 49-bp homoduplex or +8 loop substrate (15 nm, concentration in molecules) was determined in filter binding assays (“Experimental Procedures”). The left panel shows a titration, and the right panel shows binding of 15 nm ³²P-end-labeled +8 loop substrate in the presence of 80 nm Mlh1-Mlh3 and 75 nm unlabeled +8 loop or homoduplex substrate. In both panels error bars represent S.D. B, Mlh1-Mlh3 endonuclease activity was not detected on linear, +8 loop, or Holliday Junction substrates. A 12% denaturing gel containing 8 m urea is shown. C, Mlh1-Mlh3 nicks a nicked (n) circular substrate. 200 nm Mlh1-Mlh3 was added as indicated. Lanes 1–4 contain supercoiled (sc) substrate. Lanes 5–8 contain nicked substrate. Linear product is indicated by a black triangle.