Structural basis of homologous recombination

Abstract

Homologous recombination (HR) is a pathway to faithfully repair DNA double-strand breaks (DSBs). At the core of this pathway is a DNA recombinase, which, as a nucleoprotein filament on ssDNA, pairs with homologous DNA as a template to repair the damaged site. In eukaryotes Rad51 is the recombinase capable of carrying out essential steps including strand invasion, homology search on the sister chromatid and strand exchange. Importantly, a tightly regulated process involving many protein factors has evolved to ensure proper localisation of this DNA repair machinery and its correct timing within the cell cycle. Dysregulation of any of the proteins involved can result in unchecked DNA damage, leading to uncontrolled cell division and cancer. Indeed, many are tumour suppressors and are key targets in the development of new cancer therapies. Over the past 40 years, our structural and mechanistic understanding of homologous recombination has steadily increased with notable recent advancements due to the advances in single particle cryo electron microscopy. These have resulted in higher resolution structural models of the signalling proteins ATM (ataxia telangiectasia mutated), and ATR (ataxia telangiectasia and Rad3-related protein), along with various structures of Rad51. However, structural information of the other major players involved, such as BRCA1 (breast cancer type 1 susceptibility protein) and BRCA2 (breast cancer type 2 susceptibility protein), has been limited to crystal structures of isolated domains and low-resolution electron microscopy reconstructions of the full-length proteins. Here we summarise the current structural understanding of homologous recombination, focusing on key proteins in recruitment and signalling events as well as the mediators for the Rad51 recombinase.

Keywords: Cryo electron microscopy; DNA damage signalling and repair; Double-strand break repair; Homologous recombination; X-ray crystallography.

Fig. 1

Brief overview of DDR signalling. Canonical double-strand break (DSB) DNA damage response signalling pathway (see text for details)

Fig. 2

Structures of the MRN complex. a Domain structure of human Rad50, along with crystal structure of Pyrococcus furiosus RAD50 ATPase dimer (left, PDB ID: 3QKR) and human RAD50 coiled-coil domain (right, PDB ID: 5GOX) The N-terminal Walker A domain is shown in pink, C-terminal Walker B domain in orange, coiled-coil domain in green and Zn-hook in blue. Zn²⁺ is indicated as a red sphere. Bound Mre11 helix-loop-helix is shown in violet. b Domain structure of human Mre11 and crystal structure of human Mre11 N-terminal domain (PDB ID: 3T1I). Nuclease domain is shown in cyan, capping domain in red. Mn²⁺ is indicated as grey spheres. The latching loops that bind Nbs1 are disordered in this structure and are annotated as grey dash line. c Domain structure of human Nbs1 and crystal structure of Schizosaccharomyces pombe Nbs1 (PDB ID: 3HUF) with FHA domain shown in blue, BRCT 1 shown in yellow and BRCT 2 in orange. d Surface representation of ATP-hydrolysis driven RAD50 dimerisation (PDB ID: 5F3 W for the close conformation and 4FBW for the open conformation). e Intramolecular and intermolecular complex forms of MRN

Fig. 3

Structure of ATM and ATR-ATRIP. a Domain structure of human ATM (HsATM) and the Saccharomyces cerevisiae orthologue Tel1^ATM (ScTel1^ATM) and b human ATR (HsATR) with its binding partner ATR-interacting protein (HsATRIP) with Saccharomyces cerevisiae Mec1^ATR (ScMec1^ATR) and the ATRIP orthologue, Ddc2 (ScDdc2). The N-terminal HEAT-repeats are shown in cyan, the FAT (FRAP, ATR, TRAPP) in purple, the kinase domain in yellow, the PRD (PIKK regulatory domain) in orange and the FATC (FAT C-terminal domain) in dark blue. ATRIP/Ddc2, the specific ATR-interacting protein is also shown coloured green. Residue numbers are also labelled above. c The cryo-EM structure of human ATM and d the cryoEM structure of yeast Mec1^ATR-Ddc2^ATRIP, which is the most complete among the yeast and human structures, both coloured as in the schematic above, showing the dimeric architecture of these kinases along with approximate dimensions. e, f Features important for activity (see text), including the active site loop, are shown in a close up view of the atomic model, 90° rotated with respect to c, d

Fig. 4

Known structures of RPA and Rad51. a Structure and domain organisation of RPA. RPA consist of three subunits: RPA70, RPA32, and RPA14. The structure shown is RPA from of U. maydis (PDB ID: 4GNX) showing secondary structure elements of the major OB folds involved in ssDNA binding. b Structure of human Rad51 in the presence of ATP showing secondary structure elements and the right-handed helical formation (PDB ID: 5NWL). Rad51 consists of a N-terminal DNA binding domain (residues 1–84), a small linker domain (residues 85–97), and a C-terminal ATPase domain (residues 98–339). Important residues involved in ATP binding are highlighted and are located at the dimer interface of two Rad51 monomers within the filamentous crystal structure. c CryoEM structure of the presynaptic Rad51 filament in the presence of AMP-PNP bound to ssDNA (EMDB ID: 9566, PDB ID: 5H1B). Due to the intercalation of Arg235 and Val273, the Rad51 filament engages DNA as triplet clusters extending the DNA length by nearly 1.5 times in comparison to B-DNA. This mode of binding may help in the formation of the synaptic filament during the search for homologous regions

Fig. 5

Known structures of the main Rad51 mediators BRCA2, BRCA1, and PALB2. a Domain structure of BRCA2. BRC4 motif is colored in orange in the BRC4-Rad51 complex (upper left, PDB ID: 1N0 W). Helical domain in BRCA2 C-terminal Domain (BRCA2DBD) complex with DSS1 and ssDNA (upper right, PDB ID: 1MJE) is colored in sky blue, three OB folds in cyan and tower domain in pink. Magentas coil near helical and OB1 represents DSS1. ssDNA is shown in orange. EM map of full length BRCA2 (bottom, EMDB ID: 2779) is colored in yellow with its orthogonal view. b Domain structure of PALB2 and crystal structure of its WD40 domain (PDB ID: 3EU7) bound to BRCA2 N-terminal motif is colored (ruby). c Domain structure of BRCA1. NMR structure of RING heterodimer of BRCA1 and BARD1, BRCA1 is colored in red and BARD1 in light blue, bound Zn²⁺ ions are shown in sphere representation (PDB ID: 1JM7). Crystal structure of tandem BRCT repeats (purple and blue, PDB ID: 1T29). EM map of BRCA1-BARD1 is shown in violet (EMDB ID: 8833)