OTUD5 limits replication fork instability by organizing chromatin remodelers

Abstract

Proper regulation of replication fork progression is important for genomic maintenance. Subverting the transcription-induced conflicts is crucial in preserving the integrity of replication forks. Various chromatin remodelers, such as histone chaperone and histone deacetylases are known to modulate replication stress, but how these factors are organized or collaborate are not well understood. Here we found a new role of the OTUD5 deubiquitinase in limiting replication stress. We found that OTUD5 is recruited to replication forks, and its depletion causes replication fork stress. Through its C-terminal disordered tail, OTUD5 assembles a complex containing FACT, HDAC1 and HDAC2 at replication forks. A cell line engineered to specifically uncouple FACT interaction with OTUD5 exhibits increases in FACT loading onto chromatin, R-loop formation, and replication fork stress. OTUD5 mediates these processes by recruiting and stabilizing HDAC1 and HDAC2, which decreases H4K16 acetylation and FACT recruitment. Finally, proteomic analysis revealed that the cells with deficient OTUD5-FACT interaction activates the Fanconi Anemia pathway for survival. Altogether, this study identified a new interaction network among OTUD5-FACT-HDAC1/2 that limits transcription-induced replication stress.

Graphical Abstract

Figure 1.

OTUD5 mitigates replication fork stress. (A) OTUD5 knockdown leads to DNA replication stress, as indicated by increased frequency of 53BP1 foci. HeLa cells were transfected with the indicated siRNAs (20 nM), and 72 h later cells were fixed and co-stained with 53BP1 (green) and Cyclin A (red) antibodies. On the right is quantification for 53BP1 foci in Cyclin A negative cells. The experiments were conducted in triplicate. The number of foci is counted using ImageJ (n = 100 in each). (B) 53BP1 is enriched at selected CFSs in OTUD5 knockdown cells. (Top) Schematic for Common fragile site and primer binding location on FRA3B, FRA7H and FRA16D. (Bottom) qPCR quantification of 53BP1 ChIP in HeLa cells transfected with either control or OTUD5 siRNA, the experiments were conducted in triplicate. Statistical analysis was performed using ‘one way ANOVA’ (***P < 0.0005). (C) The single strand binding protein RPA is increased at nascent forks upon OTUD5 depletion. (Left) Representative images of PLA signals between EdU and RPA32. The signals are increased in HeLa cells transfected with OTUD5 siRNA. (Right) Quantification of percentage of cells with more than 5 PLA signals, with normalization to biotin–biotin signals. The experiment was conducted in triplicate (n = 50 in each experiment). (D) OTUD5 and UBR5 depletion causes increased formation of micronuclei. (Left) Representative images of increased micronuclei in HeLa cells depleted of OTUD5 or UBR5 by siRNAs. (Right) Quantification of the percentage of cells with micronuclei (n = 100). (E) OTUD5 and UBR5 depletion cause increased sensitivity to hydroxyurea. HeLa cells were transfected with indicated siRNAs, then 48 h later treated with indicated concentration of hydroxyurea. Cells were incubated for 10 additional days, fixed then stained by crystal violet. The staining intensities measured by using Gen5 software on a Synergy 2. Assays were performed in triplicates. (F) qPCR quantification of OTUD5 ChIP in HeLa cells either untreated or treated with hydroxyurea (2 mM) (***P < 0.0005). (G) OTUD5 and UBR5 proteins are present at nascent forks. (Left) Representative images of PLA signals between EdU and OTUD5 or UBR5. HeLa cells were labeled with 50μM EdU for 15mins then treated with or without hydroxyurea for 2 h. (Right) Quantification of the PLA signals per nucleus (n = 100).

Figure 2.

OTUD5 mitigates transcription-replication conflicts. (A) OTUD5 or UBR5 depletion causes increased proximity of PCNA and an elongating form of RNA polymerase II. Representative images of PCNA-RNA Pol II (S2-P) PLA in HeLa cells (left) transfected with indicated siRNAs, and quantification are shown (right). The experiment was conducted in triplicate (n = 50 in each experiment). (B) RNA Pol II (S2-P) is more present in nascent forks upon OTUD5 or UBR5 depletion. Representative images of PLA between EdU and RNA Pol II (S2-P) in HeLa cells transfected with indicated siRNAs (left). The percentage (normalized to biotin-biotin signals) of PLA signal positive cell is counted and plotted (right). The experiment was conducted in triplicate (n = 50 in each experiment). (C) DNA–RNA hybrids are increased in OTUD5 or UBR5-depleted cells. (Top) Representative images of S9.6 (red) and V5 immunostaining (green; for the V5-RNH1 transfected cells, S9.6 signals were quantified from V5-positive cells only). HeLa cells were transfected with indicated siRNAs. After 48 h, cells were transfected with V5-RNase H1 and then incubated for an additional 24 h. Cells were fixed and subjected to immunostaining. (Bottom) quantification of nuclear intensity of S9.6 staining (n = 100 for non-transfected control cells and n = 60 for V5 positive cells, ****P < 0.0001). (D) R-loops are significantly enriched in selected CFSs in OTUD5 or UBR5 knockout cells. ChIP using the S9.6 antibody followed by qPCR amplification with the indicated primers shows that R-loop level are increased at CFS in HeLa cells depleted of OTUD5 or UBR5, the experiment was conducted in triplicate (***P < 0.0005). (E) Mutant RNH1 proteins are enriched in CFSs in OTUD5-depleted cells. qPCR quantification of RNaseH1 (D210N) ChIP in HeLa cells transfected with indicated siRNAs. Depleting OTUD5 increases catalytically inactive RNaseH1 fold enrichment at CFS, the experiment was conducted in triplicate (***P < 0.0005). (F) RNAPII is increased at CFSs in OTUD5 knockdown cells. ChIP using the RNA Pol II (S2-P) antibody. HeLa cells were transfected with indicated siRNAs. Following qPCR amplification shows that elongating RNA pol II enriched at common fragile sites. The experiment was conducted in triplicate (***P < 0.0005).

Figure 3.

Uncoupling the OTUD5-FACT complex leads replication stress. (A) FACT interaction with UBR5-OTUD5 proteins is lost in the OTUD5^D537A KI cells, as shown using α-SPT16 co-IP. Parental HeLa cell and OTUD5^D537A KI cells were harvested and lysed, and anti-SPT16 IP assay was performed. (B) The reverse co-IP using α-OTUD5 consistently finds similar results. Lysate from Parental HeLa cell and OTUD5^D537A KI cells were subject to anti-OTUD5 IP. (C) Disruption of the OTUD5-SPT16 interaction is confirmed using PLA in OTUD5^D537A KI cells. PLA (anti-OTUD5 + anti-SPT16 antibodies) was performed in HeLa cells. The number of PLA was counted and plotted using ImageJ (n = 50, ****P > 0.0001). (D) ChIP using the SPT16 antibody followed by qPCR amplification with indicated primers was performed. Enrichment of SPT16 at CFS increase in OTUD5^D537A KI cell, the experiment was conducted in triplicate (***P > 0.0005). (E) The OTUD5^D537A KI cells display an increase in micronuclei (MN) and cytoplasmic bridges. Representative images of micronuclei and ultrafine bridges in OTUD5^D537A KI cell. Knock-in cell harboring D537A point mutation shows increased micronuclei and ultrafine bridges (left). Percentage of micronuclei is plotted (right). Cells were counted across three different experiments. (F) KI cells exhibit increased formation of 53BP1-NBs. (Top) Representative images of parental HeLa cell or OTUD5^D537A KI cells stained with 53BP1 and Cyclin A. (Buttom) Quantification of 53BP1 foci in Cyclin A negative cells (n = 100). (G) An increase of 53BP1-NBs is observed in HCT116 KI clones. (Top) Representative images of parental HCT116 cell or OTUD5^D537A KI cells stained with 53BP1 (green) and Cyclin A (red). 53BP1 nuclear body is indicated by arrows. (Bottom) Quantification of 53BP1 foci in Cyclin A negative cells (n > 92 in each case). (H–K) Parental HeLa cell and OTUD5^D537A KI cells were treated with indicated concentration of aphidicolin (APH), AZ20, hydroxyurea (HU) and bleomycin. Cells were incubated for 10 additional days, fixed, then stained using crystal violet.

Figure 4.

Uncoupling the OTUD5-FACT complex increases FACT loading and leads to transcription-induced genomic instability. (A) Uncoupling of FACT from OTUD5 increases elongating RNAPII in HCT116 cells. EdU-RNA Pol II (S2-P) PLA was performed in parental HCT116 cell and OTUD5^D537A KI cells. On the right is the quantification of percent of PLA signal numbering more than 5 with normalization to biotin-biotin. The experiment was performed in triplicate (n = 75). (B) Uncoupling of FACT from OTUD5 increases elongating RNAPII in HeLa cells. EdU-RNA Pol II (S2-P) PLA was performed in parental HeLa cells and OTUD5^D537A KI cells (n = 100, ****P > 0.0001). (C) Increased DNA–RNA hybrids occur at selected CFSs in a OTUD5^D537A KI cells as compared to wild type. qPCR quantification of S9.6 ChIP in parental HeLa cell or OTUD5^D537A KI HeLa cells (***P > 0.0005). (D) DNA–RNA hybrids are increased in CFSs upon SPT16 knockdown. qPCR quantification of S9.6 ChIP in HeLa cell transfected with control or SPT16 siRNA (***P > 0.0005). (E) Increased RNAPII presence occurs at CFSs in OTUD5^D537A KI clones. qPCR quantification of RNA Pol II (S2-P) ChIP in parental HeLa cell or OTUD5^D537A KI HeLa cell (***P > 0.0005). (F) S9.6 signal intensity is increased in the nucleoplasm of KI clones, which is partially reduced by expressing wild type RNH1. Nuclear S9.6 intensity with V5 positive in parental HCT116 cell or OTUD5^D537A KI cell. Cells were transfected with V5-RNaseH1, incubated for 24 h. Cells were fixed and treated with S9.6 antibody and α-V5 antibody (n > 70, ****P > 0.0001). (G) γH2AX foci are increased in the KI clones. (Left) Representative image of γH2AX immunostaining in parental HeLa cell or OTUD5^D537A KI HeLa cells. (Right) Quantification of percentage of cells with γH2AX foci numbering more than 5 is plotted (n = 100). (H) Foci are reduced by expressing RNH1. Quantification of percentage of cells with γH2AX foci numbering more than 5 is plotted (n = 50). Cells were transfected with or without V5-RNase H1 (for the V5-RNH1 transfected cells, S9.6 signals were quantified from V5-positive cells only). (I) Depleting XPG significantly reduces γH2AX foci in KI clones. Quantification of percentage of cells with γH2AX foci numbering more than 5 is plotted (n = 50). Cells were transfected with control or XPG siRNA. (J) Replication elongation is reduced in OTUD5^D537A KI HeLa cells. Parental HeLa and indicated OTUD5 KI clones were pulse labeled with CldU and IdU for 30 min each prior to DNA fiber analyses. IdU lengths were measured from dual labeled replication tracts and plotted. (K) IdU lengths were measured from bidirectional replication forks and ratios are plotted. P values were derived using Kruskal–Wallis one-way ANOVA corrected for multiple comparisons using Dunn's method with P< 0.05 as cutoff. 100 fibers were measured for figure J and at least 60 fibers were measured for figure K. (L–M) Treatment of transcription inhibitor cordyceptin rescues the elongation defect of OTUD5^D537A KI cells (L is for CldU and M is for IdU lengh).

Figure 5.

OTUD5-FACT complex mitigates replication stress through recruiting and stabilizing HDAC1 and HDAC2. (A) SPT16 specifically binds to HDAC1 and HDAC2. Parental 293T cells and FLAG-SPT16 CRISPR KI 293T cells were lysed and anti-FLAG IP was performed. (B) OTUD5 interacts with HDAC1 and HDAC2. 293T cells and stably expressing FLAG-OTUD5 cells were lysed, and anti-FLAG IP was performed with 250U benzonase treatment. (C) HDAC1 and HDAC2 proteins do not interact with SPT16 in OTUD5^537A cells. Co-IP using anti-SPT16 antibody was performed in parental HeLa cells and OTUD5^D537A KI cells. (D) SPT16 and SSRP1 interactions with HDAC1 are lost in OTUD5^537A cells. Co-IP using anti-HDAC1 was performed in parental HeLa cells and OTUD5^D537A KI cells. (E) SPT16 and HDAC1 or HDAC2 interact in wild type cells but not in OTUD5^537A cells. PLA (anti-HDAC1 or anti-HDAC2 + anti-SPT16 antibodies) was performed in HeLa cells. The experiment was conducted in triplicate. The percentage of cells with PLA positive is counted and plotted using ImageJ (n = 100 in each experiment). (F) The C-terminus of OTUD5 is sufficient to bind FACT. (Top) Schematic of full length OTUD5 a C-terminal truncation. The GST tag is on the N-terminus. (Bottom) GST pulldown, bacterial expression of GST-OTUD5 were purified using glutathione beads. The beads were applied to whole cell lysate (WCL) of 293T cells for pulldown. (G) SPT16/H4K16-ac PLA signals are eliminated by knocking down SPT16. PLA (α-SPT16 + α-H4K16ac) was performed in HCT116 cell. Cells were transfected with indicated siRNAs and incubated for 72 h. Cells were fixed and used for PLA. Representative images (left) and its quantification (right) (n = 256, ***P > 0.0005, ****P > 0.0001). (H) The SPT16/H4K16-ac PLA signal is increased in KI clones. PLA (α-SPT16 + α-H4K16ac) was performed in parental HCT116 and OTUD5^D537A KI cells (n > 140, ****P > 0.0001). (I) SPT16 occupancy at CFSs is increased in HDAC1 and HDAC2 knockdown cells. qPCR quantification of SPT16 ChIP in HeLa cells transfected with control or indicated siRNAs (***P > 0.0005). Cells were incubated for 72 h followed by transfection, then fixed and subjected to ChIP, the experiment was conducted in triplicate. (J) HDAC1 and HDAC2 depletion increases R-loop formation. Representative images (left) and its quantification (right). siRNAs were transfected into HCT116 cells and stained with S9.6 antibody (For the V5-RNH1 transfected cells, S9.6 signals were quantified from V5-positive cells only) (n > 70, ****P > 0.0001). (K) The stability of HDACs can be rescued by the proteasome inhibitor MG132. Cycloheximide chase experiment was performed in cells transfected with siOTUD5, followed by cycloheximide treatment (25 μM, for the indicated h) and followed by MG132 treatment (10 μM for indicated time points). (L) The stability of HDACs can be rescued by re-expressing wild type OTUD5. Cycloheximide chase experiment was performed using HCT116 cells with transfection of indicated siRNA and plasmid.

Figure 6.

Fanconi Anemia proteins mitigate R-loop stress induced in OTUD5^D537A KI cells. (A) Schematic for phospho-site mass spectrometry. Parental HeLa cells and two clones of OTUD5^D537A CRISPR KI cells were harvested and extracted peptides were subject to ion chromatograms for relative quantitation with Skyline software. (B) FANCD2 foci are increased in the KI clones. Representative images of FANCD2 foci (left) in parental HCT116 and OTUD5^D537A KI HCT116 cell and its quantification (right) (n = 32). Cells were seeded on 12-well plates and immunofluorescence were performed after 24 h. (C) RNH1 eliminates FANCD2 foci in the KI cells. Representative images of FANCD2 foci (left) in parental HCT116 and OTUD5^D537A KI HCT116 cells are shown (quantification in right; n = 145 each; see methods). (D) FANCD2 foci are increased when OTUD5 or UBR5 is depleted. Representative images of FANCD2 foci (top). HCT116 Cells were transfected with indicated siRNAs then incubated for 48 h. Cells were additionally incubated for 24 h after transfection of V5-RNase H1. Quantification of FANCD2 foci in V5-positive cell (bottom) (n > 38). (E) FANCD2 prevents DSB formation in KI cells. Representative images show EdU and γH2AX foci in parental and OTUD5^D537A KI cells. Each cell line was transfected with siFANCD2. Graph shows quantification of γH2AX foci in EdU positive cells (n = 80, ***P< 0.0005, ****P< 0.00001). (F) Clonogenic assay shows that viability of OTUD5^D537A KI cells is reduced by depletion of FANCD2.

Figure 7.

Model. (Top) FACT-driven RNAPII elongation is regulated by OTUD5-HDAC1/2. HDAC1/2-mediated deacetylation of nucleosomes reduces FACT recruitment and local transcription. (Bottom) Inactivation (or deficiency) of OTUD5 or HDAC1/2 causes FACT to be deregulated, leading to transcription stress and replication fork collisions.