Journal: bioRxiv

Article Title: SFPQ Promotes Homologous Recombination via mRNA Stabilization of RAD51 and Its Paralogs

doi: 10.1101/2025.09.08.674956

Figure Lengend Snippet: (A) Representative images show DAPI (blue), EdU (green), pATM (magenta), SFPQ (cyan), and merged (right) staining in siNTC-treated DIvA U2OS cells under break and no break conditions. Breaks were induced with 4-hydroxytamoxifen for 24 hours, and images were captured using a 10X magnification. Cells (∼20,000 per well) were imaged across three wells per condition (16 fields per well; 48 images total) and quantified in Cell Profiler for nuclear intensity, foci count, and cell-cycle stage based on EdU/DAPI. Data are representative of n=48 images. (B) Quantification of SFPQ–pATM and pATM–γH2AX co-localization in G2-phase cells. Violin plots show correlation coefficients of SFPQ and pATM (left) and pATM and γH2AX (right) in G2 cells with or without DNA breaks. (C) ChIP-seq data representing SFPQ-bound chromatin at 122 defined AsiSI sites under uncut (noDSB) and cut (+4OHT, 4 hours) conditions (left). ChIP-seq data representing SFPQ bound to RNU sites (right). Immunoprecipitation was performed using SFPQ polyclonal antibody. Normalized ChIP-seq signal was plotted for ±1.5 kb around AsiSI sites. SFPQ occupancy profiles are shown for two independent replicates with DSB induction (dark blue and light blue) and for the noDSB control (yellow).

Article Snippet: U2OS DR-GFP cells (female; provided by Jeremy Stark’s laboratory, City of Hope, Duarte, California, USA) and wild-type U2OS cells (female; ATCC) were both grown in DMEM supplemented with 10% FBS and 1% P/S.

Techniques: Staining, ChIP-sequencing, Immunoprecipitation, Control

Journal: bioRxiv

Article Title: SFPQ Promotes Homologous Recombination via mRNA Stabilization of RAD51 and Its Paralogs

doi: 10.1101/2025.09.08.674956

Figure Lengend Snippet: (A) (Top) SFPQ mean intensity: Violin plots (with embedded boxplots) show the single-cell distribution of nuclear SFPQ mean fluorescence intensity in DIvA U2OS cells under no break (untreated) and break (4-hydroxytamoxifen, 4-OHT) conditions. Each dot is one nucleus; boxplots denote median and interquartile range. Cell-cycle phase (G1, S, G2) was assigned per cell using EdU incorporation (green) and DAPI DNA content (blue). (Bottom) SFPQ foci per cell: Violin plots (with embedded boxplots) show the number of SFPQ nuclear foci per cell under the same conditions and cell-cycle stratification. Quantification: Cells were left untreated or treated with 4-OHT to induce AsiSI-mediated DSBs, then stained for SFPQ (cyan), EdU, and DAPI. Images were analyzed in Cell Profiler to segment nuclei, call SFPQ foci, compute per-nucleus mean intensity and foci counts, and assigned cell-cycle stage from EdU/DAPI features. (B) Non–pre-extracted immunofluorescence staining of pATM and SFPQ in DIvA U2OS cells with or without DSB induction. Cells were left untreated or treated with 4-hydroxytamoxifen (4-OHT) to induce AsiSI-mediated DSBs and stained for DNA (DAPI, blue), EdU incorporation (green), phosphorylated ATM (pATM, magenta), and SFPQ (cyan). Images were acquired without cytoskeletal (CSK) pre-extraction to visualize total nuclear staining patterns. Merged images show nuclear co-localization of pATM and SFPQ signals in the presence and absence of DNA damage.

Article Snippet: U2OS DR-GFP cells (female; provided by Jeremy Stark’s laboratory, City of Hope, Duarte, California, USA) and wild-type U2OS cells (female; ATCC) were both grown in DMEM supplemented with 10% FBS and 1% P/S.

Techniques: Fluorescence, Staining, Immunofluorescence, Extraction

Journal: bioRxiv

Article Title: SFPQ Promotes Homologous Recombination via mRNA Stabilization of RAD51 and Its Paralogs

doi: 10.1101/2025.09.08.674956

Figure Lengend Snippet: (A) mRNA-seq log₂ fold changes of RAD51 paralogs and pooled transcripts in the indicated Gene Ontology (GO) categories in DIvA U2OS cells treated with siSFPQ compared to siNTC control for 72 hours in the absence of DSBs. Data represent the mean of three biological replicates. Individual p-values were adjusted for multiple comparisons. Aggregate p-values were combined by Fisher’s method. (B) Differential transcript utilization analysis for RAD51 paralogs. mRNA-seq data from siNTC versus siSFPQ DIvA U2OS cells were analyzed for transcript isoform usage. Bars represent the likelihood ratio statistic for each gene, with blue bars indicating genes showing significant shifts in transcript utilization (RAD51B, RAD51C) upon SFPQ depletion. Grey bars indicate genes without significant changes. (C) Western blot analysis of SFPQ and RAD51 protein levels of the three biological replicates used for mRNA-seq following siNTC or siSFPQ treatment. Total protein staining is shown as a loading control. (D) Representative images show DAPI (blue), EdU (green), RAD51 (magenta), SFPQ (cyan), and merged (right) staining in siNTC-treated DIvA U2OS cells under break and no break conditions, pre-extracted with CSK. Breaks were induced with 4-OHT for 4 hours. Cells (∼20,000 per well) were imaged across four wells per condition (16 fields per well; 64 images total) and quantified in Cell Profiler for nuclear intensity, foci count, and cell-cycle stage based on EdU/DAPI. (E) Quantification of SFPQ and RAD51 foci per cell in DIvA U2OS cells following siRNA treatment and DNA damage induction. Violin plots show the distribution of foci counts across conditions with or without 4-hydroxytamoxifen (4-OHT) treatment and following transfection with non-targeting control (NTC), RAD51-targeting, or SFPQ-targeting siRNAs. Data are representative of n=64 images. (F) Violin plots showing correlation coefficients of SFPQ and RAD51 in G2 cells with or without DNA breaks. Quantification of SFPQ-RAD51 foci co-localization in G2-phase cells was performed using Cell Profiler analysis of single-cell fluorescence signals.

Article Snippet: U2OS DR-GFP cells (female; provided by Jeremy Stark’s laboratory, City of Hope, Duarte, California, USA) and wild-type U2OS cells (female; ATCC) were both grown in DMEM supplemented with 10% FBS and 1% P/S.

Techniques: Control, Western Blot, Staining, Transfection, Fluorescence

Journal: bioRxiv

Article Title: SFPQ Promotes Homologous Recombination via mRNA Stabilization of RAD51 and Its Paralogs

doi: 10.1101/2025.09.08.674956

Figure Lengend Snippet: (A) Differential expression analysis of mRNA-seq data comparing DSB versus no-DSB conditions in siNTC-treated DIvA U2OS cells (n=3 biological replicates). Mean log₂ fold change for the same targets is shown as . No significant expression differences were detected for these targets upon DSB induction in control cells. (B) ChIP-seq data showing SFPQ abundance at sites upstream and downstream of RAD51-paralog genes both without (noDSB) or with (+DSB) 4 hours of DSB induction. Data displayed is the average signal across all 6 RAD51 paralogs. (C) mRNA-seq log₂ fold changes of transcript expression of the indicated gene or GO category in DIvA U2OS cells treated with siSFPQ compared to siNTC control for 72 hours in the absence of DSBs. Data represent the mean of three biological replicates. Individual p-values were adjusted for multiple comparisons. Aggregate p-values were combined by Fisher’s method. (D) Western blot of DIvA U2OS cells treated with siSFPQ with or without p53 inhibition by PFT-α (30 µM) for 24 hours. Lysates were blotted for SFPQ, HSP70, MDM2 and RAD51. (E) (Left) Western blot of p53-null K562 cells treated with siSFPQ. Total protein staining is shown as a loading control. (Right) Quantification of SFPQ and RAD51 normalized band intensities relative to total protein is graphed.

Article Snippet: U2OS DR-GFP cells (female; provided by Jeremy Stark’s laboratory, City of Hope, Duarte, California, USA) and wild-type U2OS cells (female; ATCC) were both grown in DMEM supplemented with 10% FBS and 1% P/S.

Techniques: Quantitative Proteomics, Expressing, Control, ChIP-sequencing, Western Blot, Inhibition, Staining

Journal: bioRxiv

Article Title: SFPQ Promotes Homologous Recombination via mRNA Stabilization of RAD51 and Its Paralogs

doi: 10.1101/2025.09.08.674956

Figure Lengend Snippet: (A) Cycloheximide (CHX) ± carfilzomib (Carf) protein stability assay in DIvA U2OS cells. Cells were transfected with either non-targeting control (siNTC) or SFPQ-targeting (siSFPQ) siRNAs for 72 h, then treated with CHX alone or CHX + Carf to inhibit protein synthesis and proteasomal degradation, respectively. Lysates were collected at 0-, 2-, and 4-hours post-drug treatment from three independent biological replicates. (B) RAD51 abundance from normalized to total protein and then to 0 hr. condition. Data points represent individual replicates; lines indicate the mean. (C) RIP-seq analysis of SFPQ binding across RAD51 family paralogs in melanoma cells. Read coverage tracks show SFPQ-associated RNA fragments aligned to the genomic loci of RAD51B, RAD51C, RAD51D, XRCC2, and XRCC3. Peaks indicate regions of enriched SFPQ binding, with annotations of exon–intron structure shown below each track. Model for SFPQ-mediated stabilization of RAD51 mRNA and its impact on homologous recombination (HR). In the presence of SFPQ, the protein binds to RAD51 mRNA, promoting transcript stabilization. Stable RAD51 mRNA ensures sufficient RAD51 protein production, enabling efficient RAD51 filament formation on DNA and supporting robust HR (left). Upon SFPQ loss, RAD51 family mRNAs are destabilized, leading to reduced RAD51 protein abundance. This reduction impairs HR efficiency (right).

Article Snippet: U2OS DR-GFP cells (female; provided by Jeremy Stark’s laboratory, City of Hope, Duarte, California, USA) and wild-type U2OS cells (female; ATCC) were both grown in DMEM supplemented with 10% FBS and 1% P/S.

Techniques: Stability Assay, Transfection, Control, Binding Assay, Homologous Recombination, Quantitative Proteomics

Journal: Cell Reports Medicine

Article Title: UBA1 inhibition sensitizes cancer cells to PARP inhibitors

doi: 10.1016/j.xcrm.2024.101834

Figure Lengend Snippet: TAK243 treatment reduces homologous recombination (A) Schematic of the direct repeat (DR)-GFP assay. Stable cells are generated expressing the DR-GFP construct. The first GFP repeat contains an I-SceI endonuclease site, which initially prevents expression, and the second GFP is a truncated fragment overlapping the region with the I-SceI site. Transient transfection with I-SceI induces a double-strand break in the first GFP, which can be repaired by HR using the second truncated GFP fragment, resulting in GFP expression in cells that successfully perform HR. (B) DR-GFP U2OS cells were treated with DMSO, TAK243, or Mirin, were transfected with I-SceI, and GFP-positive cells were analyzed by flow cytometry. Mirin treatment was used as a positive control. At least 5,000 single and viable cells were analyzed for each condition. Graph represents the mean with standard deviation. ANOVA with Holm-Sidak post hoc test. ∗∗∗∗ p < 0.0001. n = 3. See gating strategy in Figure S5 . (C) Representative image of DNA damage-induced Rad51 foci in OVCAR8 and HCC1806 cells. Cells were treated with either DMSO or TAK243 for 48 h prior to irradiation (10 Gy) or mock irradiation, and then processed 4 h after irradiation to stain for cycling cells (EdU, red) and Rad51 foci (green). Scale bar. 10 μm. (D) Quantification of RAD51 foci (at least 50 cells per group) in EdU+ DMSO and TAK243-treated cells. Line represents mean, dots represent individual cells. ANOVA with Holm-Sidak post hoc test. ∗∗∗∗ p < 0.0001. n = 3. (E) OVCAR8 cells were transfected with His-ubiquitin. 48 h post-transfection cells were treated with 2 mM HU and then harvested after 2 h. His-tagged ubiquitinated proteins purified from whole-cell lysates were blotted for endogenous RPA1 and RPA2. Inset values indicate average band intensity from two independent experiments.

Article Snippet: U2OS DR-GFP cells were transfected with I-SceI (Addgene #26477) by using Lipofectamine 3000 as per the manufacturer’s instruction.

Techniques: Homologous Recombination, Generated, Expressing, Construct, Transfection, Flow Cytometry, Positive Control, Standard Deviation, Irradiation, Staining, Ubiquitin Proteomics, Purification

Journal: bioRxiv

Article Title: MND1 and PSMC3IP control PARP inhibitor sensitivity in mitotic cells

doi: 10.1101/2022.08.24.505108

Figure Lengend Snippet: A, B. Mnd1 co-localized with Rad51 in the presence or absence of exogenous DNA damage (A), but only co-localized with gH2ax upon exogenous DNA damage (B). Scatterplots are shown of percentage of cells with >5 PLA foci, normalized to the total number of cells imaged in multiple (n=10) fields of view. KB1P-G3B1 cells with a Mnd1 defect (sgMnd1), either expressing empty-vector or a vector containing Mnd1 cDNA (Mnd1+), were plated onto coverslips. Cells were exposed to 10 Gy IR or remained unexposed. Proximity ligation assays (PLAs) were performed following staining with anti-HA-tag (HA fused to Mnd1) and anti-RAD51 (A) or anti-gH2AX (B) antibodies. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Tukey’s post-test. Representative images are shown in Supplementary Figure 5A, B. C. Higher RAD51 foci levels were observed in MND1 mutant cells compared to wild-type cells upon olaparib or IR exposure. Scatter plot of RAD51 foci count per nucleus (n=min. 157) in each indicated cell line is shown. MCF10A TP53 -/- cells, either wild-type or with MND1 defect (clones A1 and B1) were plated onto coverslips. Cells were either exposed to 10 µM olaparib and then fixed after 16 hours or 10 Gy IR and then fixed after 4 hours. Vehicle cells remained untreated and were fixed simultaneously with the olaparib- or IR-exposed samples. Cells were co-stained with anti-RAD51 and anti-gH2AX antibodies. gH2AX foci quantification in these samples in shown Supplementary Figure 5C. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Bonferroni’s post-test. D. Increased Rad51 foci levels and altered kinetics of Rad51 resolution were observed upon IR in Mnd1 mutant cells compared to control cells. These phenotypes were partially reversed with ectopic Mnd1 expression. Scatter plot of RAD51 foci count per nucleus (n=min. 369) in each indicated cell line is shown. KB1P-G3B1 cells were plated onto coverslips, either expressing non-targeting control (sgNtc) or sgRNA targeting Mnd1 (sgMnd1, expressing either empty-vector or vector containing Mnd1 cDNA (Mnd1+). Cells were either exposed to 10 Gy IR and then fixed at the indicated timepoint or remained unexposed. Cells were stained with anti-RAD51 antibody. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Bonferroni’s post-test. Representative images shown in Supplementary figure 7A. E. Higher RAD51 foci levels were observed in PSMC3IP mutant cells compared to wild-type cells upon PARPi or IR exposure. Scatter plot of gH2AX foci count per nucleus (n= min. 181) in each indicated cell line is shown. MCF10A TP53 -/- cells, either wild-type or with PSMC3IP defect (clones C3 and C4) were plated onto coverslips. Cells were either exposed to 10 µM olaparib and then fixed after 16 hours or 10 Gy IR and then fixed after 4 hours. Vehicle cells remained untreated and were fixed simultaneously with the olaparib-or IR-exposed samples. Cells were co-stained with anti-RAD51 and anti-gH2AX antibodies. gH2AX foci quantification in these samples is shown in Supplementary Figure 7C. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Bonferroni’s post-test. Representative images shown in Supplementary Figure 7B. F. MND1 or PSMC3IP silencing reduced HR-mediated repair. Bar plot of % GFP+ cells relative to cells transfected with both non-targeting control siRNA (siNTC) and I- Sce I is shown. Schematic of assay shown in right panel. U2OS DR-GFP cells were transfected with siRNAs targeting MND1, PSMC3IP or non-targeting control, prior to expression of I- Sce I. A proportion of cells remained untransfected (mock) and another proportion of cells transfected with indicated siRNAs were not transfected with I- Sce I for controls of background GFP positivity. siRNA-mediated silencing of BRCA1 or BRCA2 were used as positive controls for HR deficiency. GFP+ cells were analyzed by flow cytometry. Representative FACS scatterplots shown in Supplementary Figure 8C. G. MND1 is required for cell cycle progression following DNA damage. Scatter plot of % phospho(p)HistoneH3+ cells is shown. pHistoneH3 was used as a mitotic marker . Cells were either left untreated, exposed to 250 ng/mL nocodazole for 16 hours, or exposed to 2.5 Gy IR 30 minutes prior to the 16-hour nocodazole exposure. Cells were fixed and stained with propidium iodide (PI) and anti-pHistoneH3. Representative FACS scatterplots shown in Supplementary Figure 9D. H. Mnd1 co-localizes with EdU-labelled nascent DNA, which is further increased by hydroxy urea (HU)-induced RF stalling. Scatterplots of SIRF assay are shown of cells with >2 PLA foci, normalized to the total number of cells imaged in multiple (n=15) fields of view. KB1P-G3B1 cells with Mnd1 defect, either expressing empty-vector or vector containing Mnd1 cDNA (Mnd1+), were plated on coverslips with EdU. Cells were either exposed to 2 mM hydroxy urea (HU) for two hours or remained unexposed. PLAs were performed following staining with anti-HA-tag (tagged to Mnd1) and anti-biotin antibodies. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Tukey’s post-test. Representative images shown in Supplementary Figure 9G. I, J. Mnd1 is important for Brca1-independent RF degradation upon replication-blocking DNA damage. Schematic of DNA fiber assay performed in Brca1 -deficient KB1P-G3 cells (I) and Brca1 -proficient KB1P-G3B1 cells (J), as described in with a few modifications (detailed in upper panels). Scatter plots showing quantification of IdU/CldU ratio of at least n=120 fibers per sample (lower panels). Pulse-labelling followed by RF stalling via hydroxy urea (HU) resulted in an increased track length ratio of Mnd1 -mutant cells (sgMnd1) compared to non-targeting control (sgNtc). Track length ratio was restored to wild-type levels with reconstitution of Mnd1 cDNA (Mnd1+) (J). Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Tukey’s post-test. K. Mnd1 is important for RF slowing, specifically fork reversal, upon replication-blocking DNA damage. Schematic of DNA fiber assay performed in Brca1-deficient KB1P-G3 cells, as described in with a few modifications (detailed in upper panel). Scatter plot showing quantification of total track lengths of at least n=120 fibers per sample (lower panel). Pulse-labelling followed by RF stalling via MMC resulted in increased RF progression, evidenced by increased track length, of Mnd1 -mutant (sgMnd1) cells compared to non-targeting control (sgNtc). Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Tukey’s post-test. L, M. Mnd1 loss increases micronuclei formation upon exposure to olaparib (L) or IR (M), which is reversed with ectopic Mnd1 expression. Scatterplots of % cells with micronuclei in each indicated sample are shown (n=20). KB1P-G3B1 expressing either sgRNA targeting Mnd1 (sgMnd1) or non-targeting control (sgNtc) were plated onto coverslips. Mnd1 -deficient cells either express an empty-vector or a vector containing Mnd1 cDNA (Mnd1+). Cells were either exposed to 8 Gy IR, 10 µM olaparib for 16 hours or remained untreated. Error bars represent the median and 95% CI. P -values were calculated via ANOVA with Tukey’s post-test. Representative images shown in Supplementary Figure 10O. Scatterplots showing Psmc3ip loss also increases micronuclei formation upon exposure to olaparib or IR shown in Supplementary Figure 10K, L) with representative image in Supplementary Figure 10R.

Article Snippet: DR-GFP U2OS (kindly gifted by Jeremy Stark (City of Hope, USA)), HEK293T (ATCC), CAL51 (DSMZ) and MDAMB-231 (ATCC) were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Gibco) supplemented with 10% FBS.

Techniques: Expressing, Plasmid Preparation, Ligation, Staining, Mutagenesis, Clone Assay, Control, Transfection, Flow Cytometry, Marker, Blocking Assay