High-affinity DNA-binding domains of replication protein A (RPA) direct SMARCAL1-dependent replication fork remodeling

Abstract

SMARCAL1 catalyzes replication fork remodeling to maintain genome stability. It is recruited to replication forks via an interaction with replication protein A (RPA), the major ssDNA-binding protein in eukaryotic cells. In addition to directing its localization, RPA also activates SMARCAL1 on some fork substrates but inhibits it on others, thereby conferring substrate specificity to SMARCAL1 fork-remodeling reactions. We investigated the mechanism by which RPA regulates SMARCAL1. Our results indicate that although an interaction between SMARCAL1 and RPA is essential for SMARCAL1 activation, the location of the interacting surface on RPA is not. Counterintuitively, high-affinity DNA binding of RPA DNA-binding domain (DBD) A and DBD-B near the fork junction makes it easier for SMARCAL1 to remodel the fork, which requires removing RPA. We also found that RPA DBD-C and DBD-D are not required for SMARCAL1 regulation. Thus, the orientation of the high-affinity RPA DBDs at forks dictates SMARCAL1 substrate specificity.

Keywords: ATPase; DNA Damage Response; DNA Repair; DNA Replication; Fork Regression; Genomic Instability; Replication Protein A; SMARCAL1.

FIGURE 1.

SMARCAL1 is regulated by RPA. A, schematic of RPA bound to ssDNA showing the polarity of the four DBDs and location of the SMARCAL1-interacting domain. B, schematic of RPA orientation on the stimulatory and inhibitory substrates. The good SMARCAL1 substrates have the high-affinity DBDs bound close to the fork junction. Also note that the position of RPA32C (the SMARCAL1-interacting domain) may be different on the good and bad SMARCAL1 substrates. The actual spatial locations of the DNA and proteins will be different in the three-dimensional structure of an actual replication fork. C, model for repair of stalled replication forks by SMARCAL1. Panel i, RPA bound to the leading-strand template stimulates SMARCAL1-catalyzed fork regression. RPA should be present on the leading strand only when the polymerase is stalled. Panel ii, RPA bound to the lagging-strand template inhibits SMARCAL1-catalyzed fork regression, thereby preventing aberrant remodeling of an actively elongating fork. However, RPA bound to the nascent leading strand of a reversed fork stimulates SMARCAL1-mediated restoration to a normal fork structure.

FIGURE 2.

A SMARCAL1 mutant containing an RPA70N-interacting motif binds DNA and is recruited to RPA foci in cells after DNA damage. A, schematic of WT SMARCAL1 (with an RPA32C-interacting motif), SMARCAL1-ΔN (without an RPA-interacting domain), and SMARCAL1-RPA70BD (with an RPA70N-interacting motif). B, U2OS cells were transfected with expression vectors for GFP-tagged SMARCAL1, SMARCAL1-RPA70BD, and SMARCAL1-ΔN. The cells were treated with hydroxyurea and imaged for RPA and SMARCAL1. C, a model fork substrate was prebound with RPA (so that 100% of the substrate was bound) and incubated with increasing concentrations of the indicated SMARCAL1 protein for 30 min at room temperature. The products were then resolved on a 5% polyacrylamide gel. NPC, no-SMARCAL1 protein control.

FIGURE 3.

A SMARCAL1 mutant that binds to the RPA70N domain is regulated similarly to wild-type SMARCAL1. A and C, schematic of the fork regression and restoration assays, respectively. The ³²P-labeled strands are indicated by asterisks. The physiological reaction that is mimicked is indicated in parentheses. BM, branch migration. B and D, the fork substrates were prebound with RPA and incubated with SMARCAL1 (SM1) for 60 min (B) or 10, 30, or 60 min (D). Products of the reactions were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown.

FIGURE 4.

A SMARCAL1 mutant that binds to the RPA70N domain is inhibited by RPA on the normal fork substrate. A, schematic of the lag gap fork regression assay. The ³²P-labeled strands are indicated by asterisks. The physiological reaction that is mimicked is indicated in parentheses. BM, branch migration. B, the fork substrates were prebound with RPA and incubated with increasing concentrations of SMARCAL1-RPA70BD for 20 min at 30 °C. Products of the reactions were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown.

FIGURE 5.

RPA mutants that are defective in binding DNA do not stimulate SMARCAL1 as well as wild-type RPA. A, representative RPA binding assay. The fork regression substrate was incubated with the indicated concentrations of wild-type RPA, DBD-A, or DBD-B for 20 min, and products were analyzed by native gel electrophoresis. The concentrations used in the remodeling assays were 6 nm wild-type RPA and 30 nm DBD-A and DBD-B. B, representative autoradiograph of the fork restoration assay with RPA mutants. The substrates and products are indicated. C–E, the restoration and regression substrates were prebound with WT RPA, DBD-A, or DBD-B as indicated for 20 min and then incubated with SMARCAL1 for increasing times. The products were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown.

FIGURE 6.

RPA bound to an 8-nt gap is sufficient to stimulate SMARCAL1 regression of a stalled fork. A, schematic of the stalled fork substrate used in the assay. The ssDNA region is limited to 8 nucleotides, allowing the binding of only RPA DBD-A and DBD-B. B, the fork substrates were prebound with RPA and incubated with SMARCAL1 for increasing times. Products of the reactions were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown.

FIGURE 7.

An RPA protein containing only DBD-A, DBD-B, and a SMARCAL1-interacting surface is sufficient to stimulate SMARCAL1. A, schematic of the RPA DBD-C mutant (mutation in DBD-C that abrogates binding to DNA) bound to the substrate. B, the fork substrates were prebound with the RPA DBD-C mutant and incubated with 1 nm SMARCAL1 for increasing times. Products of the reactions were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown. C, schematic of the FAB-RPA protein (RPA that has only DBD-A and DBD-B along with the RPA70N domain) bound to the DNA substrate. D, the fork substrates were prebound with FAB-RPA and incubated with the indicated SMARCAL1 (SM1) proteins at 1 nm for increasing times. Products of the reactions were analyzed by native gel electrophoresis. The means ± S.D. from three experiments are shown.