SMARCAL1 ubiquitylation controls its association with RPA-coated ssDNA and promotes replication fork stability

Abstract

Impediments in replication fork progression cause genomic instability, mutagenesis, and severe pathologies. At stalled forks, RPA-coated single-stranded DNA (ssDNA) activates the ATR kinase and directs fork remodeling, 2 key early events of the replication stress response. RFWD3, a recently described Fanconi anemia (FA) ubiquitin ligase, associates with RPA and promotes its ubiquitylation, facilitating late steps of homologous recombination (HR). Intriguingly, RFWD3 also regulates fork progression, restart and stability via poorly understood mechanisms. Here, we used proteomics to identify putative RFWD3 substrates during replication stress in human cells. We show that RFWD3 interacts with and ubiquitylates the SMARCAL1 DNA translocase directly in vitro and following DNA damage in vivo. SMARCAL1 ubiquitylation does not trigger its subsequent proteasomal degradation but instead disengages it from RPA thereby regulating its function at replication forks. Proper regulation of SMARCAL1 by RFWD3 at stalled forks protects them from excessive MUS81-mediated cleavage in response to UV irradiation, thereby limiting DNA replication stress. Collectively, our results identify RFWD3-mediated SMARCAL1 ubiquitylation as a novel mechanism that modulates fork remodeling to avoid genome instability triggered by aberrant fork processing.

Fig 1. Proteomics identification of RFWD3 interactors and putative substrates.

(A) Domain organization of RFWD3 mutants. (B) Relative enrichment of interacting partners of RFWD3 C315A compared to C315A/I639K in HU-treated HEK293T cells, as determined by SAINT analysis of proteomics data. (C) STRING representation of the RPA-ssDNA-centered RFWD3 interactome. (D) Co-localization of SMARCAL1 and RFWD3 (WT, I639K, C315A, and C315A/I639K mutants) at micro-irradiation stripes. (E) Quantification of RFWD3 recruitment to SMARCAL1 stripes in U2-OS cells. The bars correspond to the mean +/− standard error of the mean. Experiments were performed 3 times and correspond to n = 854 SMARCAL1+ stripes. (F) HEK293T cells were transfected either with SFB-RFWD3 C315A or C315A/I639K and treated or not with 2 mM HU for 3 h. Native streptavidin pulldown was performed to isolate RFWD3 along with interacting partners followed by SDS-PAGE and blotting with the indicated antibodies. Summary data displayed in Fig 1E can be found in S1 Data. HU, hydroxyurea; RPA, replication protein A; ssDNA, single-stranded DNA.

Fig 2. DNA damage induces SMARCAL1 ubiquitylation on chromatin.

(A) HEK293T cells were transfected with control or SMARCAL1-targeting siRNA and 24 h later with a Strep-HA ubiquitin expression plasmid for 24 h and treated with 1 μm CPT for 3 h before harvest. Ubiquitylated proteins were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibodies. (B, C) HEK293T cells expressing Strep-HA ubiquitin were (B) continuously treated with 1 μm CPT or 2 mM HU for 3 h or exposed to 10 Gy IR or 50 J/m² UV-C and collected 3 h later or (C) treated with 100 ng/ml MMC for 24 h. Ubiquitylated proteins were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibody. (D) HEK293T cells expressing Strep-HA ubiquitin were treated with 2 mM HU for 3 h and fractionated to obtain either soluble or chromatin-associated protein fractions. Ubiquitylated proteins from these fractions were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibodies. CPT, camptothecin; HU, hydroxyurea; IR, irradiation; MMC, mitomycin C; UV, ultraviolet.

Fig 3. RFWD3 promotes SMARCAL1 ubiquitylation in vivo and in vitro.

(A, B) HEK293T cells were transfected with control or RFWD3-targeting siRNAs and 24 h later with a Strep-HA ubiquitin plasmid for 24 h and were treated or not with 2 mM HU for 3 h. Ubiquitylated proteins were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibodies. (C, D) Cells complemented by stable integration of siRNA-resistant HA-tagged (C, D) WT RFWD3 cDNA and (D) C315A and I639K mutant were transfected with control or RFWD3 3′UTR-targeting siRNAs and 24 h later with a Strep-HA ubiquitin construct for 24 h and cells were treated or not with 2 mM HU for 3 h. Ubiquitylated proteins were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibodies. (E–G) RFWD3 directly ubiquitylates SMARCAL1. In vitro ubiquitylation assays were performed using recombinant human SMARCAL1 and RFWD3 proteins (E, F) purified from insect cells or (G) WT or C315A RFWD3 purified from human cells (E–G) in the presence of UBE2D1 (E2), UBA1 (E1), ubiquitin, and ATP. Ubiquitylation was stopped in Laemmli buffer and SDS-PAGE and immunoblotting were performed using the indicated antibodies. HU, hydroxyurea.

Fig 4. SMARCAL1 ubiquitylation does not promote its degradation.

(A) HEK293T cells were transfected with control or p97-targeting siRNA and 24 h later with a Strep-HA ubiquitin expression plasmid for 24 h and treated with 2 mM HU for 3 h before harvest. Ubiquitylated proteins were collected by denaturing Strep-Tactin pulldown and blotted with the indicated antibodies. (B, C) HEK293T cells were transfected with Strep-HA ubiquitin and treated with (B) either 2 mM HU for 3 h or (C) 2 mM HU or 100 ng/ml MMC for 24 h and (B, C) treated or not with 5 μm MG132 for 2 h before harvest. Ubiquitylated proteins were collected by Strep-Tactin pulldown and blotted with the indicated antibodies. (D) Total extracts from HEK293T cells treated or not with 2 mM HU 4 h before the addition of 50 μg/ml CHX for the indicated times were blotted with the indicated antibodies. CHX, cycloheximide; HU, hydroxyurea; MMC, mitomycin C.

Fig 5. SMARCAL1 ubiquitylation by RFWD3 impedes its interaction with the RPA complex.

(A) Schematic diagram of ubiquitylated lysines on SMARCAL1 identified by LC/MS-MS and database mining. (B, C) In vitro ubiquitylation reactions were performed and stopped by the addition of EDTA, prior to incubation with pre-formed RPA-coated biotinylated ssDNA. Native streptavidin pulldown of RPA-ssDNA was then carried-out and RPA-ssDNA bound and unbound protein fractions were separated by SDS-PAGE and blotted using the indicated antibodies. (D, E) Immunofluorescence was performed against SMARCAL1 and RPA70 in control or RFWD3 KO U2-OS cell treated or not with UV-C. Images were collected and SMARCAL1 and RPA70 colocalizing foci were automatically counted using CellProfiler. Data are presented as the mean ± SD (n = 3). A total of >300 cells were assessed per biological replicate. Significance was determined by one-way ANOVA followed by Bonferroni’s multiple comparisons test. (****) P < 0.0001. Summary data displayed in Fig 5E can be found in S1 Data. KO, knock-out; RPA, replication protein A; ssDNA, single-stranded DNA; UV, ultraviolet.

Fig 6. SMARCAL1 ubiquitylation regulates its activity at replication forks in vivo.

(A) Schematic representation of K27R and N33-ubiquitin fusion SMARCAL1 mutants. (B, C) SFB-tagged WT, K27R, N33-Ub, or ΔN30 SMARCAL1 were transiently expressed for 48 h in U2-OS cells. Native streptavidin pulldown was performed to isolate SMARCAL1 along with interacting partners followed by SDS-PAGE and blotting with the indicated antibodies. (C) Quantification of RPA32 interaction with WT, K27R, N33-Ub, and ΔN30 SMARCAL1 using 3 independent experiments. (D–G) SFB-tagged WT, K27R, N33-Ub, or ΔN30 SMARCAL1 were transiently overexpressed for 48 h in U2-OS cells. Cells were stained with DAPI and immunofluorescence against FLAG and γ-H2A.X was performed. Levels of Flag-SMARCAL1 and γ-H2A.X were automatically quantified in each nucleus using CellProfiler. (D) Histogram of WT and mutants SMARCAL1 recruitment on chromatin. (E) Representative images of transfected cells. (F) Histogram of the % of cells with pan-nuclear γ-H2A.X staining in FLAG-positive cells. Error bars represent SEM from 4 independent experiments; 2,603 cells were quantified and samples were compared with one-way ANOVA followed by Tukey’s multiple comparisons test. (***) P < 0.001, (****) P < 0.0001. (G) Box and whisker diagram of γ-H2A.X intensity normalized according to FLAG-intensity to control for SMARCAL1 expression levels. Error bars represent 5–95 percentile from 4 independent experiments; 2,603 cells were quantified and samples were compared with one-way ANOVA followed by Tukey’s multiple comparisons test. (****) P < 0.0001. (H) Down-regulation of MUS81 and SLX4 decreases pan-nuclear γ-H2A.X signal in SMARCAL1-expressing cells. U2-OS cells were transfected with control or MUS81 and SLX4-targeting siRNA and 24 h later were transfected with SFB-WT, K27R, or N33-Ub SMARCAL1. Box and whisker diagram of γ-H2A.X/FLAG-intensity in FLAG+ cells nuclei. Error bars represent 5–95 percentile from 3 independent experiments; 5,072 cells were quantified and samples were compared with Kruskal–Wallis’ test followed by Dunn’s multiple comparisons. (*) P < 0.05, (**) P < 0.01, (***) P < 0.001, (****) P < 0.0001. Summary data displayed in Fig 6C, 6D, 6F, 6G and 6H can be found in S1 Data. RPA, replication protein A.

Fig 7. SMARCAL1 and MUS81 mediate UV-induced replication stress in RFWD3 KO cells.

(A–C) Depletion of RFWD3 induces RPA accumulation on chromatin in UV-treated S-phase cells. (A) U2-OS transfected with control or RFWD3-targeting siRNAs were exposed to 2 J/m² UV-C and 4 h later processed for FACS to monitor RPA accumulation on chromatin. (B, C) RFWD3 KO U2-OS cells stably transfected with doxycycline-inducible WT, C315A or I639K RFWD3 cDNAs were induced or not for 24 h and exposed to 2 J/m² UV-C. DNA and chromatin-associated RPA were quantified by FACS 4 h post-irradiation. Three independent biological replicates were performed and representative FACS profiles are shown. Statistical significance was established by one-way ANOVA followed by Šidák’s test. (*) P < 0.05, (**) P < 0.01, (***) P < 0.001, (****) P < 0.0001. (D–F) U2-OS control or RFWD3 KO cells were transfected with SMARCAL1 or (G–I) MUS81-targeting siRNAs were exposed to the indicated UV-C doses and chromatin-associated RPA and γ-H2A.X was quantified as above. (J) Working model for the influence of RFWD3-mediated SMARCAL1 regulation on fork stability. Summary data displayed in Fig 7E, 7F, 7H and 7I can be found in S1 Data. KO, knock-out; RPA, replication protein A; UV, ultraviolet.