PSIP1/LEDGF reduces R-loops at transcription sites to maintain genome integrity

Abstract

R-loops that accumulate at transcription sites pose a persistent threat to genome integrity. PSIP1 is a chromatin protein associated with transcriptional elongation complex, possesses histone chaperone activity, and is implicated in recruiting RNA processing and DNA repair factors to transcription sites. Here, we show that PSIP1 interacts with R-loops and other proteins involved in R-loop homeostasis, including PARP1. Genome-wide mapping of PSIP1, R-loops and γ-H2AX in PSIP1-depleted human and mouse cell lines revealed an accumulation of R-loops and DNA damage at gene promoters in the absence of PSIP1. R-loop accumulation causes local transcriptional arrest and transcription-replication conflict, leading to DNA damage. PSIP1 depletion increases 53BP1 foci and reduces RAD51 foci, suggesting altered DNA repair choice. Furthermore, PSIP1 depletion increases the sensitivity of cancer cells to PARP1 inhibitors and DNA-damaging agents that induce R-loop-induced DNA damage. These findings provide insights into the mechanism through which PSIP1 maintains genome integrity at the site of transcription.

Fig. 1. PSIP1 depletion leads to the accumulation of R-loops.

a Schematic showing PSIP1 domains. b Immunofluorescence images and γ-H2AX foci in wild-type, Psip1^–/– and Psip1^–/– p75R MEFs (n > 1500 nuclei, three independent experiments; median values in red bar; p-values obtained using two-tailed Mann–Whitney test). c Slot-blot images showing the binding of p75/p52 isoforms of PSIP1 with RNA-DNA hybrids. RNASEH-treated genomic DNA and IgG served as a control. Normalised band intensity (right) with mean ± SD (n = 3 independent experiments; p-values by one-way ANOVA followed by Tukey’s test). d Immunoblots of S9.6, H3 and IgG immunoprecipitated lysates from RWPE-1 nuclear extract with indicated antibodies. e Representative images and dot-plot showing PLA foci between S9.6 and PSIP1 in control and RNASEH1 overexpressed RWPE-1 cells (n > 1400 nuclei observed over three independent experiments; median values indicated with a red line; p-values obtained using two-tailed Mann–Whitney test). f Slot-blot for R-loop using S9.6 antibody from cells as in b. The normalised band intensity (right) was plotted as mean ± SD (n = 3 independent experiments; p-values by one-way ANOVA followed by Tukey’s test). g Immunoblot with PSIP1/p75 and β-actin for lysate from PSIP1-shRNA (PSIP1-sh) and non-targeting control shRNA (control-sh). h S9.6 immunofluorescence in RWPE-1 cells (left) and quantification (representative images; n > 4000 nuclei from three independent experiments; median values are indicated with the red line; p-values by two-tailed Mann–Whitney test). i Like f, but for PSIP1 knockdown (PSIP1-sh) and control knockdown (control-sh) RWPE-1 cells. R-loops isolated from cells overexpressing RNASEH1 also was used for slot-blot. The normalised band intensity has been plotted as mean ± SD (n = 3 independent experiments; p-values by one-way ANOVA followed by Tukey’s test). j Heatmaps showing E. coli spike-in normalised CUT&Tag reads for PSIP1/p75 in control and PSIP1 knockdown RWPE-1 cells. k Heatmap showing R-loop and PSIP1/p75 levels in PSIP1-KD and control RWPE-1 cells. l Genome-browser track (hg38) showing the CUT&Tag signal (read counts) for PSIP1 and R-loop in control and PSIP1-KD RWPE-1 cells. Source data are provided as a Source Data file.

Fig. 2. PSIP1 depletion leads to R-loop-mediated DNA damage.

a Western blot showing the levels of γ-H2AX in control and PSIP1 KD RWPE-1 cells. b Representative IF images and dot-plot showing the γ-H2AX foci in control and PSIP1 KD RWPE-1 cells. The cells overexpressing RNASEH1 were also used as controls. The number of foci per cell was quantified and plotted as a dot-plot (n > 8000 nuclei from three independent experiments; Median values are indicated with a red line; p-values by two-tailed Mann–Whitney test). c Like b but for 53BP1 foci (representative images; n > 3000 cells for each group from three independent experiments; median values are indicated with red bars; p-values by two-tailed Mann–Whitney test). d Immunoblot images showing the bulk levels of 53BP1 protein in control, PSIP1 KD and upon RNASEH1 overexpression in HEK293T cells. e Heatmap showing CUT&Tag reads (normalised to E. coli reads) for γ-H2AX and PSIP1/p75 across S9.6 peaks gained in PSIP1-KD HEK293T cells. f Genome-browser tracks showing CUT&Tag data for PSIP/p75, R-loops (S9.6 ab) and γ-H2AX in control and PSIP-KD HEK293T cells. g Representative images and dot-plot of PLA between R-loops (S9.6) and γ-H2AX antibodies in control and PSIP-KD RWPE1 cells (representative images; n > 1000 nuclei observed from three independent experiments; median values indicated with red bar; p-values obtained using two-tailed Mann–Whitney test). Source data are provided as a Source Data file.

Fig. 3. R-loop accumulation at promoters leads to transcription-replication conflict.

a Distribution of PSIP1, R-loop (S9.6 ab), and γ-H2AX CUT&Tag peaks from RWPE-1 cells around the gene transcription start sites (promoters), gene bodies and intergenic regions. b Heatmaps showing the log2 fold change in read counts between control and PSIP1-KD CUT&Tag reads for PSIP1/p75, R-loops (S9.6 ab) and γ-H2AX in RWPE-1 cells across the NCBI reference genes. c upSet plot and Venn diagram (right) showing the unique and overlapping peaks obtained from CUT&Tag reads for PSIP1, S9.6 and γ-H2AX in HEK293T cells. The x-axis shows the number of peaks, and the Y-axis shows the number of intersections. d Average profile with SD (shaded regions of blue and red) of TT-seq (RPKM) across protein-coding genes (left), dotted box shows promoter region and around the centre of the S9.6 peaks gained in PSIP KD RWPE-1 cells (right). e Slot blot using S9.6 antibody in control and PSIP-KD RWPE-1 cells after treatment with transcriptional inhibitor (triptolide 50 nM; 36 h), methylene blue staining of the same DNA served as a loading control. Normalised intensity of S9.6 intensities was plotted as mean ± SD (n = 3 independent experiments; p-values by one-way ANOVA followed by Tukey’s multiple comparison test). f Representative PLA image between α-PCNA and α-RNAPII antibodies obtained from control and PSIP1 KD RWPE-1 cells (left). The number of PLA foci per cell observed between PCNA and RNAPII was quantified and plotted as dot blot (n > 1000 cells over three independent experiments; red line shows the median value that has been indicated; p-values obtained from two-tailed Mann–Whitney test). Source data are provided as a Source Data file.

Fig. 4. PSIP depletion leads to reduced HR repair at the R-loop-induced DNA damage.

a Mean % survival ± SD of PSIP-KD and control-KD RWPE-1 cells treated with different concentrations of illudin-S, etoposide and aphidicolin drugs (n = 3 independent experiments; p-values by multiple unpaired t-tests). b Western blot for γ-H2AX and DNA-PK (right) in control-KD and PSIP1-KD cells in cells treated with phleomycin (1 μg/mL) followed by repair for the indicated time. β-actin served as a loading control. c Western blots for PSIP1/p75, γ-H2AX, for γ-H2AX IPed HEK293T extracts in control, PSIP1 KD, in the presence of camptothecin (CPT@10 μM; 2 h) (+) and DMSO (–). γ-H2AX IP was also performed after overexpressing the cells with RNASEH1. IgG is a negative control; input extracts are blotted (below). d Representative immunofluorescence microscopic images showing 53BP1 foci in DMSO-treated or CPT (10 μM; 2 h) treated RWPE-1 cells. The median and number of foci per cell are plotted as dot plots (n > 6000 cells in each group over three independent experiments; median values are indicated with a red line and p-values are obtained using a two-tailed Mann–Whitney test). e Like d, but for RAD51 foci. Dot plots showing the number of foci per cell with median values indicated (n > 6000 cells in each group over three independent experiments; p-values obtained by two-tailed Mann–Whitney test). Source data are provided as a Source Data file.

Fig. 5. PSIP1 deficiency increases the sensitivity of cancer cells to drugs that induce transcription-coupled DNA damage.

a Western blotting of S9.6, H3 and IgG immunoprecipitated (IP) from RWPE-1 cell nuclear extract with PSIP1/p75, PARP1, γ-H2AX and PCNA (negative control) antibodies. Lysate treated with RNASEH before the pulldown was used as a negative control. IgG and H3 IP were negative controls, and 5% of the nuclear extract was used as input. b Western blotting for PSIP1-IP with PSIP1, PARP1, γ-H2AX and PCNA antibodies, IgG served as a negative control. 10% of the nuclear extract was used as input. c Western blotting for eGFP-PARP1 and HA-PSIP1 co-IPs was performed using GFP-trap beads. αGFP ab was to detect eGFP-PARP1 and αHA to detect HA-PSIP1/p75 and HA-p52, along with γ-H2AX and PCNA antibodies. d TGCA expression data showing levels of PSIP1 transcripts in cancers of the bladder (BLCA; n = 408), breast (BRCA; n = 1097), cervical (CESC; n = 305), kidney chromophobe (KICH; n = 67), kidney renal clear cell (KIRC; n = 533), Kidney renal papillary cell (KIRP; n = 290), lung (LUAD; n = 515), lung squamous cell (LUSC; n = 503), pancreatic (PAAD; n = 178) and prostate (PRAD; n = 497). The data was obtained from the UALCAN online tool. The centre line indicates a median value; the boxes and whiskers indicate 25th to 75th and 10th to 90th percentiles, respectively. e Western blot showing the levels of PSIP1 in prostate normal and cancer cells and their sensitivity to olaparib as determined by CCK-8 assay (f). Mean survival ± SD is plotted (n = 3 independent experiments). g Western blot confirming the depletion of PSIP1 by shRNA in LNCaP cells. h Cell survival assay showing % survival of control and PSIP1 knockdown LNCaP cells with indicated doses of olaparib determined using the CCK-8 assay (mean ± SD; n = 3 independent experiments). i Similar to h but for indicated doses of illudin-S. j Working model showing the role of PSIP1 in reducing R-loop level at transcription sites to minimise transcription-replication conflict leading to DNA damage. Source data are provided as a Source Data file.