Abstract
SMARCAL1, ZRANB3 and HLTF are required for the remodeling of replication forks upon stress to promote genome stability. RAD51, along with the RAD51 paralog complex, were also found to have recombination-independent functions in fork reversal, yet the underlying mechanisms remained unclear. Using reconstituted reactions, we build upon previous data to show that SMARCAL1, ZRANB3 and HLTF have unequal biochemical capacities, explaining why they have non-redundant functions. SMARCAL1 uniquely anneals RPA-coated ssDNA, which depends on its direct interaction with RPA, but not on ATP. SMARCAL1, along with ZRANB3, but not HLTF efficiently employ ATPase driven translocase activity to rezip RPA-covered bubbled DNA, which was proposed to mimic elements of fork reversal. In contrast, ZRANB3 and HLTF but not SMARCAL1 are efficient in branch migration that occurs downstream in fork remodeling. We also show that low concentrations of RAD51 and the RAD51 paralog complex, RAD51B-RAD51C-RAD51D-XRCC2 (BCDX2), directly stimulate the motor-driven activities of SMARCAL1 and ZRANB3 but not HLTF, and the interplay is underpinned by physical interactions. Our data provide a possible mechanism explaining previous cellular experiments implicating RAD51 and BCDX2 in fork reversal.