Fig 2: Model for the role of PARG in the repair of TOP1-DPCs.a PARP1 and TOP1 form cellular protein complexes (present study and refs. 22,34). b TOP1-DPC trapped by CPT is rapidly modified with PAR by PARP1 and with ubiquitin (by RNF4 and potentially other E3 ligases, not shown). The PARylation recruits TDP1, PARG, and USP7 to the TOP1-DPC. c TOP1-DPC PARylation is readily and rapidly reversed by PARG, enabling the 26S proteasome to target the ubiquitylated TOP1-DPC for degradation. d TDP1 hydrolyzes the TOP1 peptide to expose the DNA ends for repair. e In the presence of PARGi, TOP1-DPC dePARylation is blocked and the persistent PAR polymers on TOP1-DPC obstruct the proteasome hence stabilize TOP1-DPC. f The stabilization of TOP1-DPC triggers USP7 to deubiquitylate the DPC to recycle the ubiquitin molecules.

Fig 3: Correcting TDP1 splice-site mutation restores TDP1 protein expression and function in HAP1 cells. (A) Sanger sequencing chromatograms of two clones each from RPE-1, HAP1 and HAP1 STAR cells, where the HAP1 STAR clones had had their endogenous TDP1 splice acceptor mutant ‘AA’ successfully edited into functional ‘AG’ via DNA-templated HDR. Two clones (HAP1 STAR Clone #3 & #30) were identified. A silent mutation of TDP1 c.672C > T was also observed in the edited HAP1 STAR clones. This mutation was intentionally included in the editing template to minimize re-annealing of gRNA to the target DNA sequence upon successful genome editing. (B) Separation of DNA products amplified using reverse transcribed TDP1 cDNA in edited (HAP1 STAR) and unedited HAP1 clones as well as in RPE-1 clones on a 2% agarose gel to test for TDP1 mRNA length and purity. The estimated size (bp) of DNA bands without exon skipping (full length of mRNA) for each PCR (i, ii, iii and iv) is annotated. See top part of Figure 2D. The presence of any DNA bands of smaller sizes demonstrate the phenomenon of exon skipping in TDP1 mRNA. The DNA ladder (left) is made up of DNA bands at 100 base-pair intervals. (C) Total cell lysates of HAP1 STAR and other unedited clones (HAP1 and RPE-1) were immunoblotted to detect the recovery of TDP1 protein expression upon editing. PCNA serves as a loading control. (D) Cell viability dose-response curves to camptothecin of HAP1 STAR clones relative to unedited HAP1 clones. Around 10 000 HAP1 cells were seeded in technical duplicates 24 h before camptothecin treatment. Treated cells were incubated for 3 days before measuring the cell viability with Alamar Blue fluorescent indicator. Experimental data were the average of three biological repeats ± S.E.M. For the comparison against ‘WT’ HAP1 at 10 nM camptothecin, the P-value (**** P< 0.0001) was calculated using two-way ANOVA. (E) Quantified AUC based on the cell viability dose-response curves in D. The P-value (**** P< 0.0001) was calculated through ordinary one-way analysis of variance (ANOVA). Despite showing only the comparison of HAP1 STAR clones #3 and #30 against the unedited HAP1 polyclonal cells, their comparison against unedited HAP1 clones (#3 and #5) were also significant with P-value≤ 0.0001. (F) Representative images of alkaline ‘comets’ for ‘WT’ HAP1, HAP1 TDP1KO, HAP1 STAR #3, and HAP1 STAR #30 upon different treatment conditions (UNT: untreated, CPT: camptothecin, H2O2: hydrogen peroxide). H2O2 (100 μM) treatment was included as a positive control for DNA single-stranded break induction. (G) Quantified Olive tail moment of the ‘comets’, which reflects the amount of DNA single-stranded breaks. Experimental data were the mean (bar height) of the population medians in 3 biological repeats (dots) ± S.E.M. 82–150 comets were analyzed for each condition in each repeat. Different cell lines were compared against ‘WT’ HAP1 in each treatment condition using ordinary two-way ANOVA with the following reported P-values: (ns) P> 0.05 and (****) P< 0.0001.

Fig 4: HAP1 cells are hypersensitive to camptothecin and deficient of TDP1 protein. (A) Cell viability dose-response curves of HAP1, human retinal pigment epithelial (RPE-1) and human osteosarcoma (U2OS) cell lines to camptothecin treatment. (B) Chemogenetic CRISPR-Cas9 knockout screens in RPE-1 (49), HEK293A (48) and HAP1 (47) cells subjected to camptothecin treatment, highlighting the relative impacts on gRNAs targeting TDP1 (•) and TDP2 (•) in the cell populations interpreted as Zscore (54) or log2 fold-change. (C, left) Cell viability dose-response curves to camptothecin of HAP1 and (D, left) RPE-1 cells in the presence or absence of TDP2. For the comparison in (C, left) at 2 nM camptothecin, the P-value (**** P< 0.0001) was calculated using two-way ANOVA. (C & D, right) Quantified area under curve (AUC) data from the corresponding dose-response curves. In (A), (C) & (D), 5000 HAP1 cells and 1000 RPE-1 and U2OS cells were seeded in technical duplicates 24 h before camptothecin treatment. Treated cells were incubated for 4 days before measuring the cell viability with Alamar Blue fluorescent indicator. Experimental data were the average of 3 biological repeats ± s.e.m. (E) Comparison of the differential AUCs (AUCTDP2-KO/AUCWT) between (C) HAP1 and (D) RPE-1 cells. The P-value (*** P< 0.001) was calculated through an unpaired two-tailed t-test (i.e. one-way ANOVA). (F) Total cell lysates of individual cell lines were immunoblotted for TDP1 protein. PCNA serves as a loading control. The sizes (kDa) of reference proteins in PageRuler prestained protein ladder are indicated on the right. (G) Normalized read counts of TDP1 mRNA in HAP1, RPE-1 and U2OS cell lines. RNA read counts for each cell line were sourced from RNA-Seq data available online and analyzed using DESeq2. Each column represents an RNA-Seq experiment with multiple biological replicates (black dots, n = 2–4). Adjusted p-values of TDP1 RNA read count comparison against HAP1 cells are shown. (H) Comparison of TDP1 mRNA level (normalized against PUM1 as a stable housekeeping gene) between different cell lines via RT-qPCR. Normalized TDP1 mRNA levels were compared against HAP1 cells. Experimental data were the average of three technical replicates ± S.E.M.

Fig 5: TDP1 splice variant impairs TDP1 RNA splicing and causes exon skipping in HAP1 cells. (A) Pairwise comparison of the sequenced non-coding DNA (intronic regions adjacent to the exon ends) of TDP1 in HAP1, U2OS and RPE-1 cells against the reference genome. Introns have been compressed into fixed lengths (50 bp width; black connecting lines) to showcase the entire TDP1 gene, where the blue horizontal rectangles are individual exons except for exon 6 (magenta rectangle). Genomic variants in the TDP1 non-coding sequences for individual cell lines are highlighted as a ‘lollipop’ and coloured according to the variant type: (•) variant in the intron, (•) variant that affects splice donor/acceptor site, (•) variant in the 5′ or 3′ untranslated region according to the whole-exome sequencing. (B) Sanger sequencing chromatograms of RPE-1, U2OS, HAP1 and HAP1-BE3 cells showing the genomic sequence of TDP1 around exon 6. The TDP1 c.660–1G > A splice-site mutation (‘AA’ splice acceptor site) is unique to HAP1 and HAP1-BE3 cells. (C) Sashimi plots of TDP1 RNA expression in HAP1 (top), U2OS (middle) and RPE-1 (bottom) cells where the peaks indicate the RNA coverage at the coding sequence of different exons and the peak height reflects the depth of mapped reads from RNA-Seq results. The peaks at exons 3, 6, 7 and 8 were annotated and the dotted line boxes highlight the peak heights of exons 6 and 7 in HAP1, U2OS and RPE-1 cell lines. (D) Top: Schematics of the expected PCR amplicon sizes (in bp) of multiple targeted PCRs – i, ii, iii and iv (with reverse primers annealing to exon 6, 7, 8 or 9 respectively) of the TDP1 cDNA, taking into account when exon 6 or exon 6 + 7 skipping occur. Note: forward primers annealing to exon 3 are not identical in different PCRs. Bottom: Separation of DNA products amplified using reverse transcribed TDP1 cDNA in different cell lines (R: RPE-1, U: U2OS, H: HAP1, HB: HAP1-BE3) on a 2% agarose gel to test for TDP1 mRNA length and purity. The estimated size (bp) of DNA bands for each PCR (i, ii, iii and iv) is annotated (smaller PCR products due to exon skipping are annotated in magenta). The DNA ladder (bottom left) is made up of DNA bands at 100 base-pair intervals.