BRCA2 stabilises RAD51 and DMC1 nucleoprotein filaments through a conserved interaction mode

Abstract

BRCA2 is essential for DNA repair by homologous recombination in mitosis and meiosis. It interacts with recombinases RAD51 and DMC1 to facilitate the formation of nucleoprotein filaments on resected DNA ends that catalyse recombination-mediated repair. BRCA2's BRC repeats bind and disrupt RAD51 and DMC1 filaments, whereas its PhePP motifs bind recombinases and stabilise their nucleoprotein filaments. However, the mechanism of filament stabilisation has hitherto remained unknown. Here, we report the crystal structure of a BRCA2-DMC1 complex, revealing how core interaction sites of PhePP motifs bind to recombinases. The interaction mode is conserved for RAD51 and DMC1, which selectively bind to BRCA2's two distinct PhePP motifs via subtly divergent binding pockets. PhePP motif sequences surrounding their core interaction sites protect nucleoprotein filaments from BRC-mediated disruption. Hence, we report the structural basis of how BRCA2's PhePP motifs stabilise RAD51 and DMC1 nucleoprotein filaments for their essential roles in mitotic and meiotic recombination.

Fig. 1. Crystal structure of a BRCA2-DMC1 complex.

a Schematic of the human BRCA2 sequence (top) with multiple sequence alignments of the exon 14 (middle) and exon 27 (bottom) regions, highlighting their PhePP (FxPP) motifs. The constructs used in this manuscript, corresponding to Ex14 (amino-acids 2387-2420), Ex14-Tr (amino-acids 2401-2414) and Ex27 (amino-acids 3270-3315), are indicated. b Crystal structure of the complex between BRCA2 Ex14-Tr peptides and a DMC1 ΔN octameric ring. c Interaction between the PhePP region of the BRCA2 Ex14 peptide (purple) and a DMC1 protomer (yellow), highlighting FxPP amino-acids F2406, P2408 and P2409, which are bound in a pocket between two loops on the DMC1 surface (labelled as loops A and B). d, e Superposition of the BRCA2-DMC1 structure (purple and yellow) with the RAD51 core (white). d PhePP-binding loop A interacts with the N-terminal end of the BRCA2 Ex14 peptide, and diverges between recombinases, consisting of 147-GAGGYPG-153 in DMC1 and 148-IDRGGGE-154 in RAD51. e PhePP-binding loop B interacts with the C-terminal end of the BRCA2 Ex14 peptide, and also diverges between recombinases, consisting of 177-FNVD-180 in DMC1 and 178-YGLS-181 in RAD51.

Fig. 2. BRCA2 Ex14 interacts with DMC1’s oligomeric core via its PhePP-binding site.

a Microscale thermophoresis (MST) of interactions between BRCA2 peptides and DMC1 ΔN; data are presented as mean values, with error bars indicating standard error, n  =  3 independent experiments. BRCA2 Ex14 and Ex14-Tr bind to DMC1 core with affinities of 30 μM and 36 μM, respectively, whereas interactions are not detectable for the BRCA2 Ex14-AAA mutant (F2406A, P2408A, P2409A), or for the DMC1 ΔN loop mutant (replacing 147-GAGGYPG-153 and 178-NVDHDA-183 with RAD51 amino-acids IDRGGGE and GLSGSD, respectively). BRCA2 Ex27 binds to DMC1 ΔN with an affinity of >96 μM. b, c Amylose pull-downs of (b) DMC1 ΔN, (c) DMC1 ΔN wild-type and loop mutant, following recombinant co-expression with MBP-BRCA2 Ex14 wild-type, Ex14-Tr, Ex14-AAA mutant, Ex27, BRC4 and free MBP (empty). d, e Amylose pull-downs of (d) His-DMC1 ΔN V179E mutant and (e) His-DMC1 ΔN F85E mutant, following recombinant expression and mixing with recombinantly expressed MBP-BRCA2 Ex14 wild-type, Ex14-Tr, Ex27, BRC4 and free MBP (empty). The MBP-BRCA2 fusion proteins exhibited some degradation to free MBP and intermediate species (MBP fused to a partially degraded peptide), which was more pronounced for Ex14 and Ex27 than for Ex14-Tr and BRC4. b–e Gel images are representative of at least three replicates. Source data are provided as a Source Data file.

Fig. 3. BRCA2 Ex14 stabilises DMC1-ssDNA complexes.

Electrophoretic mobility shift assays (EMSAs) analysing the ability of BRCA2 Ex14 to bind, promote and protect DMC1-ssDNA filaments. a EMSAs using TAE pH 7.5 conditions in which (i) DMC1-ssDNA binding was incomplete at the canonical ratio of three nucleotides to one protomer (boxed). The canonical ratio (boxed) from subpanel (i) was used in subsequent EMSAs. (ii) BRCA2 Ex14 promoted the formation of DMC1-ssDNA filaments. (iii) The BRCA2 Ex14-AAA mutant largely failed to stimulate filament formation and (iv) truncated BRCA2 Ex14-Tr mostly stimulated DMC1-ssDNA filament formation. b EMSAs using TEA pH 7.5 + KCl conditions in which (i) DMC1-ssDNA binding was largely complete at the canonical ratio of three nucleotides to one protomer (boxed). The canonical ratio (boxed) from subpanel (i) was used in subsequent EMSAs. (ii) BRCA2 Ex14 induced a super-shift, demonstrating binding to DMC1-ssDNA filaments. The super-shift was largely eliminated by (iii) the BRCA2 Ex14-AAA mutant and (iv) truncated BRCA2 Ex14-Tr. The order of addition did not affect the ability of Ex14 to induce the formation and super-shift of DMC1-ssDNA filaments, as shown in Supplementary Fig. 8. c EMSAs using TEA pH 7.5 + KCl conditions in which (i) DMC1-ssDNA binding was disrupted by a stoichiometric excess of BRC4 (dashed, boxed). (ii) BRCA2 Ex14 protected against BRC4-mediated disruption (dashed, boxed). The protection was completely abrogated by (iii) the BRCA2 Ex14-AAA mutant, and was diminished in (iv) truncated BRCA2 Ex14-Tr. BRCA2 peptide concentrations are shown as molar ratios with respect to DMC1 protomers. Arrowheads, free ssDNA; yellow arrows, DMC1-ssDNA complexes; blue arrows, BRCA2-DMC1-ssDNA complexes. The ssDNA substrate is a 100-nucleotide random sequence (provided in Methods). a–c Gel images are representative of at least three replicates. Source data are provided as a Source Data file.

Fig. 4. BRCA2 Ex27 stabilises RAD51-ssDNA complexes.

a Model of the BRCA2 Ex27-RAD51 filament structure generated by docking BRCA2 Ex27-RAD51 1:2 complex AlphaFold2 models (Supplementary Fig. 11) onto a previously reported structure of the RAD51 filament (PDB accession: 8BSC). b The modelled BRCA2 Ex27 peptide includes the PhePP motif (boxed), and a long extension that runs along the interface between adjacent RAD51 promoters, shrouding the F86 self-association interaction. c The PhePP motif is predicted to bind in the same manner as the BRCA2 Ex14-DMC1 interaction, involving FxPP amino-acids F3298, P3300 and P3301. d Amylose pull-downs of RAD51 ΔN following recombinant co-expression with MBP-BRCA2 Ex14, Ex27 and its AAA mutant (F3298A, P3300A, P3301A), BRC4 and free MBP (empty). The MBP-BRCA2 fusion proteins exhibited some degradation to free MBP and intermediate species (MBP fused to a partially degraded peptide), which was more pronounced for Ex14 and Ex27 than for BRC4. e EMSAs in which (i) RAD51-ssDNA binding was disrupted by equimolar quantities of BRC4 (dashed, boxed). (ii) BRCA2 Ex27 but not (iii) BRCA2 Ex14-AAA mutant protected against BRC4-mediated disruption (dashed, boxed), and (iv) RAD51 L180E mutant underwent similar disruption, but (v) was not protected by BRCA2 Ex27. BRCA2 peptide concentrations are shown as molar ratios with respect to RAD51 protomers. Arrowheads, free ssDNA; yellow arrows, RAD51-ssDNA complexes; red arrows, BRCA2-RAD51-ssDNA complexes. The ssDNA substrate is a 100-nucleotide random sequence (provided in Methods). The same EMSAs performed using a 100-nucleotide polydT ssDNA substrate, in which complete protection is conferred by Ex27 and abrogated upon mutation, are shown in Supplementary Fig. 12. d, e Gel images are representative of at least three replicates. Source data are provided as a Source Data file.