Insights into the mechanism of Rad51 recombinase from the structure and properties of a filament interface mutant

Abstract

Rad51 protein promotes homologous recombination in eukaryotes. Recombination activities are activated by Rad51 filament assembly on ssDNA. Previous studies of yeast Rad51 showed that His352 occupies an important position at the filament interface, where it could relay signals between subunits and active sites. To investigate, we characterized yeast Rad51 H352A and H352Y mutants, and solved the structure of H352Y. H352A forms catalytically competent but salt-labile complexes on ssDNA. In contrast, H352Y forms salt-resistant complexes on ssDNA, but is defective in nucleotide exchange, RPA displacement and strand exchange with full-length DNA substrates. The 2.5 A crystal structure of H352Y reveals a right-handed helical filament in a high-pitch (130 A) conformation with P6(1) symmetry. The catalytic core and dimer interface regions of H352Y closely resemble those of DNA-bound Escherichia coli RecA protein. The H352Y mutation stabilizes Phe187 from the adjacent subunit in a position that interferes with the gamma-phosphate-binding site of the Walker A motif/P-loop, potentially explaining the limited catalysis observed. Comparison of Rad51 H352Y, RecA-DNA and related structures reveals that the presence of bound DNA correlates with the isomerization of a conserved cis peptide near Walker B to the trans configuration, which appears to prime the catalytic glutamate residue for ATP hydrolysis.

Figure 1.

DNA strand exchange activities of wild-type and mutant Rad51 proteins. DNA strand exchange assays were carried out using full-length φX174 DNA substrates as described under ‘Materials and Methods’ section. (A) Schematic of in vitro DNA strand exchange reaction, showing interconversion of single- and double-stranded DNA species. (B) Timecourses of DNA strand exchange reactions visualized by separating DNA substrates and products via agarose gel electrophoresis and staining with ethidium bromide. All reactions contained 33 µM each of φX174 circular ssDNA and φX174 linear dsDNA, 11 µM Rad51 enzyme, 1.65 µM RPA and 2.5 mM ATP (final concentrations). Other reaction components and assay conditions are described under ‘Materials and Methods’ section. Lanes 1–5, reaction with wild-type Rad51. Lanes 2–10, reaction with H352A Rad51 mutant. Lanes 11–15, reaction with H352Y Rad51 mutant. (C) Results of panel B quantified by fluorescence imaging. Data are averages from three separate experiments. Filled circles, wild-type Rad51. Open squares, H352A Rad51 mutant. Filled triangles, H352Y Rad51 mutant.

Figure 2.

ssDNA-binding site size determination for wild-type and mutant Rad51 proteins. Etheno-modified M13mp19 ssDNA (εDNA) was titrated with Rad51 wild-type (open circles), H352A (filled squares), or H352Y (crosses) under low-salt conditions while monitoring the enhancement of εDNA fluorescence as described under ‘Materials and Methods’ section. Solutions contained 3 µM εDNA and 2 mM ATP in buffer consisting of 30 mM Tris–acetate (pH 7.5), 10 mM magnesium acetate and 0.1 mM DTT. The apparent binding site sizes of proteins on this single-stranded lattice are estimated from the titration endpoints. The average endpoint is indicated in this figure by the intersection of the solid asymptotic lines, with the vertical line showing the protein concentration at the average endpoint, corresponding to one Rad51 protomer per three εDNA nucleotide residues.

Figure 3.

Electrophoretic mobility shift assays for ssDNA-binding activities of wild-type and mutant Rad51 proteins—effects of ATP and salt. Protein–ssDNA complexes were assembled at a constant concentration of 30 µM φX174 ssDNA and at variable concentrations of 0–30 µM wild-type or mutant Rad51 enzyme as described under ‘Materials and Methods’ section. Complexes were assembled in buffer containing 30 mM Tris–acetate, pH 7.5, 5 mM MgCl₂, with or without ATP and NaCl as indicated. Samples were electrophoresed on agarose gels and visualized by staining with ethidium bromide. Top row—experiments with wild-type Rad51. Nucleotide and salt conditions were as follows: (a) 0 mM ATP (apo conditions), 0 mM NaCl; (b) 2 mM ATP, 0 mM NaCl; (c) 2 mM ATP, 150 mM NaCl; (d) 2 mM ATP, 500 mM NaCl; (e) 2 mM ATP, 1000 mM NaCl. Middle row—experiments with Rad51 H352Y mutant. Nucleotide and salt conditions were as follows: (f) 0 mM ATP (apo conditions), 0 mM NaCl; (g) 2 mM ATP, 0 mM NaCl; (h) 2 mM ATP, 150 mM NaCl; (i) 2 mM ATP, 500 mM NaCl; (j) 2 mM ATP, 1000 mM NaCl. Bottom row—experiments with Rad51 H352A mutant. Nucleotide and salt conditions were as follows: (k) 0 mM ATP (apo conditions), 0 mM NaCl; (l) 2 mM ATP, 0 mM NaCl; (m) 2 mM ATP, 20 mM NaCl; (n) 2 mM ATP, 150 mM NaCl; (o) 2 mM ATP, 500 mM NaCl. Note the lower NaCl concentrations used with H352A compared with H352Y and wild-type Rad51.

Figure 4.

Steady-state rates of ATP hydrolysis by wild-type and mutant Rad51 proteins. ATP hydrolysis was measured by TLC radiometric assay as described under ‘Materials and Methods’ section. Reaction mixtures contained 3 mM ATP and 2.0 µM of either Rad51 wild-type, H352A, or H352Y as indicated below the figure. Reactions were performed either in the absence (gray bars) or in the presence (white bars) of 12 µM φX174 ssDNA. Monovalent salt was added to reactions at the concentrations indicated below the figure. All other buffer components and assay conditions were as described under ‘Materials and Methods’ section.

Figure 5.

Binding and hydrolysis of stoichiometric ATP by Rad51 and mutants. ATP binding and hydrolysis was measured at a 1:1 concentration ratio of protein to ATP. Nucleotide binding was measured by polyacrylamide gel EMSA, and hydrolysis was measured by TLC, as described under ‘Materials and Methods’ section. (A) 10 µM wild-type or mutant Rad51, as indicated, was incubated with 10 µM α-[³²P]-ATP (10 µC_i/ml) for 2 h either in the absence (lanes 3–5) or presence (lanes 6–8) of 30 µM φX174 ssDNA. Control experiments lacking enzyme or lacking both enzyme and ssDNA are shown in lanes 1 and 2, respectively. Samples were electrophoresed on native 12% polyacrylamide gels to separate free from bound nucleotide, which were detected and quantified by phosphorimaging. Other buffer components and assay conditions are described under ‘Materials and Methods’ section. (B) The percentage of bound nucleotide in (A), was quantified for each form of Rad51 in the absence (gray bars) and presence (white bars) of ssDNA. (C) The percentage of ATP hydrolyzed in samples identical to those analyzed in (A), was measured for each form of Rad51 protein in the absence (gray bars) and presence (white bars) of ssDNA.

Figure 6.

RPA displacement assays. (A) The displacement of RPA protein from ssDNA by Rad51 wild-type and mutant proteins was monitored by the increase in RPA tryptophan fluorescence as described under ‘Materials and Methods’ section. (B) Continuous fluorescence assays for RPA displacement. All reactions contained 2.5 mM ATP, 0.25 µM RPA and 5 µM φX174 ssDNA, except for the ‘RPA alone’ control reaction (red) which lacked ssDNA. Reactions contained 2 µM of either wild-type (cyan), H352A (blue, orange), or H352Y (green) Rad51, except for the ‘RPA alone’ and ‘RPA–ssDNA’ (black) controls which lacked Rad51. All reactions contained 50 mM monovalent salt except for one reaction with H352A that contained 100 mM monovalent salt (orange). Other buffer components and reaction conditions are described under ‘Materials and Methods’ section.

Figure 7.

The H352Y interface mutant of yeast Rad51 (center, blue) forms a 6₁ helical filament with a pitch of 130 Å. In comparison, the I345T mutant of this protein (left, green and gold, PDB code 1SZP) forms a filament of the same pitch but as a 3₁ helix of dimers. Despite the difference in helical symmetry, the filaments superimpose remarkably well (right.) Both proteins crystallized in the nucleotide-free state.

Figure 8.

The protomers and the dimer interfaces of the yeast Rad51 H352Y, Rad51 I345T and E. coli RecA recombinases are remarkably similar. (A) The ATPase core of the ‘A’ protomer of Rad51 I345T (magenta, PDB code 1SZP) was superimposed onto a protomer of Rad51 H352Y (dark blue), carrying along the adjacent protomer ‘D’ (salmon for I345T, cyan for H352Y) but not including it in the superposition; (B) Likewise, superimposing only the ‘D’ protomer of I345T onto H352Y. (C) While the ATPase core of one protomer of the RecA–ssDNA-ADP.AlF₄ complex (yellow, PDB code 3CMW) superimposes well onto that of Rad51-H352Y (dark blue), the adjacent protomer (not included in the superposition, orange for RecA, light blue for Rad51) overlays well only in the region of the dimer interface. The ADP-AlF₄ (green) and ssDNA backbone trace (gray) of the RecA complex are shown for reference. (D) In contrast, the DNA-free form of E. coli RecA (PDB 3CMV) presents a substantially different dimer interface than the others.

Figure 9.

Tyrosine 352 of the Rad51 H352Y mutant is present in two predominant alternate conformations. Detailed view of the dimer interfaces from the perspective of the ssDNA, using the superpositions as described in Figure 8, with the protomers on the right half of each figure used in the superposition. (A) One of the alternate conformations of Tyr352 of H352Y (cyan) is similar to His352 of the ‘D’ protomer (salmon) in the ‘AD’ dimer of I345T. (B) Neither conformation of Tyr352 (cyan) is similar to His352 of the ‘A’ protomer (magenta) of the ‘DA’ dimer of I345T. (C) While there is a good match between the dimer interfaces of Rad51 H352Y (dark blue and cyan) and RecA–ssDNA–ADP–AlF4 (yellow and orange), Phe217, equivalent to residue 352 of yeast, is disposed differently than either alternate conformation of Tyr352.

Figure 10.

Superposition of the Walker A and B motifs of E. coli RecA–ssDNA–ADP–AlF4 onto the Rad51–H352Y and -I345T mutants suggests a key role for P-loop residue Phe187 (Glu68 in E. coli) in sensing the gamma-phosphate or preventing the binding of ATP, and for residue 352 in assisting in catalysis. RecA is shown in yellow with its adjacent protomer in orange and its bound ADP-AlF₄ in light green in these figures. (A) Superimposed Rad51-H352Y (purple) and its adacent protomer (cyan); (B) Superimposed ‘A’ protomer of Rad51–I345T (magenta) with its adjacent ‘D’ protomer (salmon); (C) Superimposed ‘D’ protomer of Rad51–I345T (salmon) with its adjacent protomer (magenta.) In (A) and (C), Phe187 of the P-loop is only 1.65 Å and 1.70 Å, respectively, from the AlF₄ mimic of the γ-phosphate of the ATP. In (B), the δ-nitrogen of His352 is seen to be located almost exactly in between the amine groups of lysines 248 and 250 of E. coli, residues which have been shown to be essential for nucleotide hydrolysis.

Figure 11.

A role for cis to trans isomerization of the peptide linkage at the end of the conserved Walker B motif in configuring the active site for nucleotide hydrolysis. (A) The only RecA-family structures to clearly possess a trans peptide linkage between residues 144–145 at the end of Walker B contain DNA, as represented here by the active site of E. coli RecA–ssDNA–ADP–AlF4 complex (PDB code 3CMW). (B) As in nearly all other RecA-family structures, the active site of Rad51–H352Y shows a cis configuration at this peptide (between residues 280–281). (C) Compared with the cis configuration, in the trans configuration Ser281 (145 in E. coli) and Ala284 (148 in E. coli) move toward the active site, and push the putative nucleophilic activator Glu221 (96 in E. coli) 1.25 Å closer to the γ-phosphate (mimicked by the AlF₄ moiety) of the ATP.