Promotion of presynaptic filament assembly by the ensemble of S. cerevisiae Rad51 paralogues with Rad52

Abstract

The conserved budding yeast Rad51 paralogues, including Rad55, Rad57, Csm2 and Psy3 are indispensable for homologous recombination (HR)-mediated chromosome damage repair. Rad55 and Rad57 are associated in a heterodimer, while Csm2 and Psy3 form the Shu complex with Shu1 and Shu2. Here we show that Rad55 bridges an interaction between Csm2 with Rad51 and Rad52 and, using a fully reconstituted system, demonstrate that the Shu complex synergizes with Rad55-Rad57 and Rad52 to promote nucleation of Rad51 on single-stranded DNA pre-occupied by replication protein A (RPA). The csm2-F46A allele is unable to interact with Rad55, ablating the ability of the Shu complex to enhance Rad51 presynaptic filament assembly in vitro and impairing HR in vivo. Our results reveal that Rad55-Rad57, the Shu complex and Rad52 act as a functional ensemble to promote Rad51-filament assembly, which has important implications for understanding the role of the human RAD51 paralogues in Fanconi anaemia and cancer predisposition.

Figure 1. Interactions of Rad55–Rad57 with the Shu complex and Rad52.

(a) Rad55–Rad57 was incubated with Csm2–Psy3 and the protein complex was captured through the (His)₆ tag on Rad55 using Ni²⁺ resin. E, eluate from the resin; S, supernatant containing unbound proteins; W, wash of the resin. (b) Rad52 was tested for Y2H interaction with the indicated proteins. (c) The Y2H interaction between Rad52 and Csm2 was examined in wild-type or rad55Δ cells. (d) The Y2H interaction between Rad52 and Rad55 was examined in wild-type or csm2Δ cells. (e) GST-tagged Rad52 was mixed with Rad55–Rad57, Shu complex, or both, and protein complexes were captured on glutathione resin. Immunoblotting for the FLAG tag on Rad57 and Psy3 was used to identify Rad55–Rad57 and Shu complex retained on the resin.

Figure 2. Enhancement of Rad51 loading onto RPA-coated ssDNA by the Shu complex and Rad55–Rad57.

(a) (i) Schematic of the Rad51 loading assay. (ii) Magnetic bead resin with immobilized ssDNA was pre-incubated with RPA, then Rad51 was added along with combinations of Rad52, Shu complex and Rad55–Rad57. The amount of Rad51 retained on the ssDNA was determined by immunoblotting. (b) (i) Schematic of the DNA strand exchange assay. (ii) RPA-coated ssDNA was incubated with Rad51 and combinations of Rad52, Shu complex and Rad55–Rad57, mixed with radiolabelled homologous double-stranded DNA (dsDNA) and Rad54 and then analysed. P, DNA strand exchange product; S, dsDNA substrate. S.d. are plotted as error bars (n=3) and ‘*' indicates significance.

Figure 3. Impairment of physical interaction and functional synergy of Shu complex with Rad55–Rad57 by the csm2–F46A mutation.

(a) Cartoon view of the Csm2–Psy3 heterodimer, with Csm2 in blue and Psy3 in green. Csm2–F46 is highlighted in red. (b) Y2H analysis for interaction of Csm2 and csm2–F46A with Rad55, Rad51, Rad52 and Psy3. (c) Pull-down assay (as in Fig. 1a) to examine interaction of (His)₆-Rad55–Rad57 with Csm2–Psy3 harbouring csm2–F46A. (d) Analysis of the DNA-binding activity of Shu complex harbouring csm2–F46A. (e) Rad51 loading and (f) DNA strand exchange assays were carried out as in Fig. 2 to evaluate the efficacy of Shu complex harbouring csm2–F46A in the promotion of presynaptic filament assembly. S.d. are plotted as error bars (n=3) and ‘*' indicates significance.

Figure 4. Impairment of homologous recombination by the csm2–F46A mutation.

(a) csm2–F46A cells are as sensitive to MMS as a csm2Δ mutant. Cultures were fivefold serially diluted onto YPD medium with the indicated dose of MMS and incubated for 2 days at 30 °C. (b) Similar to csm2Δ, csm2–F46A alleviates the MMS and HU sensitivity of an sgs1Δ mutant. Cells of the indicated genotypes were fivefold serially diluted and tested for sensitivity to 0.012% MMS or 100 mM HU. (c) WT, csm2Δ or csm2-F64A cells harbouring a direct repeat HR reporter (leu2-ΔEcoRI::URA3::leu2-ΔBstEII) were tested for spontaneous rates of Rad51-dependent gene conversion (GC) and Rad51-independent single-strand annealing (SSA) as described. The rates of GC are significantly decreased in both csm2–F46A and csm2Δ strains (P<0.01 by student's t-test) with a corresponding increase in SSA relative to WT strains (P<0.02 by student's t-test). S.d. are plotted as the error bars (n=3) and ‘*' indicates significance. (d) Like csm2Δ cells, csm2–F46A cells exhibit an elevated mutation rate in a canavanine mutagenesis assay. S.d. are plotted as error bars (n=5) and ‘*' indicates significance.