USP1-trapping lesions as a source of DNA replication stress and genomic instability

The deubiquitinase USP1 is a critical regulator of genome integrity through the deubiquitylation of Fanconi Anemia proteins and the DNA replication processivity factor, proliferating cell nuclear antigen (PCNA). Uniquely, following UV irradiation, USP1 self-inactivates through autocleavage, which enables its own degradation and in turn, upregulates PCNA monoubiquitylation. However, the functional role for this autocleavage event during physiological conditions remains elusive. Herein, we discover that cells harboring an autocleavage-defective USP1 mutant, while still able to robustly deubiquitylate PCNA, experience more replication fork-stalling and premature fork termination events. Using super-resolution microscopy and live-cell single-molecule tracking, we show that these defects are related to the inability of this USP1 mutant to be properly recycled from sites of active DNA synthesis, resulting in replication-associated lesions. Furthermore, we find that the removal of USP1 molecules from DNA is facilitated by the DNA-dependent metalloprotease Spartan to counteract the cytotoxicity caused by “USP1-trapping”. We propose a utility of USP1 inhibitors in cancer therapy based on their ability to induce USP1-trapping lesions and consequent replication stress and genomic instability in cancer cells, similar to how non-covalent DNA-protein crosslinks cause cytotoxicity by imposing steric hindrances upon proteins involved in DNA transactions.

. Cell cycle effects of USP1 autocleavage mutants. a, HCT116 cell lines were treated with or without 2 mM HU for 4 hours and subjected to immunoblotting analysis with the indicated antibodies. b, Schematics of different classes of replication structures from DNA fiber analysis used to determine the frequency of origins for each of the indicated HCT116 cell lines. The percentage of new origin firing was calculated as the sum of both bi-directional redgreen-red tracts and red only tracks over the total number of replication structures. Data for % origin firing are represented by mean and -/+ SD for three independent experiments (n=200 events) and p-values were calculated using t-test with Welch's correction (ns=no significance, *p<0.05, **p<0.01, two-tailed). c, USP1 WT and mutant cells were pulse-labeled with 10 µM EdU for 30 min and analyzed by flow cytometry for DNA synthesis (EdU) and DAPI staining for DNA content (10,000 events per sample). Percentages of cells in G1, S, and G2/M phases are graphed for three biological replicates. Error bars indicate SEM. d, Asynchronously growing HCT116 cell lines were fixed and co-stained for Cyclin A and 53BP1 (counterstained with DAPI). Representative images for each cell line are shown. Cells that were Cyclin A negative with 53BP1 nuclear bodies were quantified and graphed with mean and -/+ SD for three independent experiments (n=300 Cyclin A negative cells per experiment). P-values were calculated using ttest with Welch's correction (ns=no significance, **p<0.01, ***p<0.001, two-tailed). Scale bar = 20 µM e, Cells were treated with either control or Polk siRNA for 72 hours and subjected to DNA fiber analysis. Replication origins firing during the second labeling period (red only tracts) were analyzed for replication fork speed. CldU tract lengths of elongating forks are plotted with mean and -/+ SD for 3 independent experiments (n=100 red only origins), and p values were calculated using the Mann-Whitney rank-sum t-test (ns=no significance, *p<0.05, ***p<0.001, ***p<0.0001), two-tailed). Figure 3. A representative sample of genomic DNA fractionation of HCT116 cells for Ok-seq analysis. a, Alkaline gel image showing sucrose gradient of HCT116 USP1C90S#1 genomic DNA. Fractions 1-9 (below >300bp) as indicated within the red box are concentrated for further processing to generate the adaptor-ligated Okazaki fragments library. b, Representative PCR gel image of amplified libraries of Okazaki fragments for HCT116 USP1C90S#1 and HCT116 USP1C90S#2 cell lines. c, Tapestation image of amplified libraries in (b) after removal of any adaptor-dimers and primer-dimers as part of quality control. (upper panel) Schematic showing anticipated Okazaki fragment distributions around replication origins with strong (blue) versus weak (red) localized firing efficiencies. (lower panel) Schematic representation of Okazaki fragment distributions arising from replication termination at TTS under normal conditions (black) versus delocalized termination under replication stress (red). b, Percentage of replication forks moving left to right around TSS sorted by total RNA-seq read depth quartile (FPKM), (from ref 64 ), for the indicated cell lines. P values were calculated using the Kruskal-Wallis test, using the regions 50-30kb upstream and 1-10kb downstream of the TSS. Effect sizes, shown as eff, were calculated using the same regions. The values were calculated comparing the USP1 mutant cell lines against the USP1 wild-type cell line. All statistics are presented in the Methods. c, Percentage of replication forks moving left to right around TSS of actively transcribed genes (FPKM>median), binned by gene length according to quartiles for transcribed genes. d, Same analysis as in (c), binned for long genes of 50-100 kb (upper panel) or >100 kb (lower panel). e, Percentage of replication forks moving left to right around TTS binned by RNA-seq read depth quartile for the indicated cell lines. f, Percentage of replication forks moving left to right around TTS of actively transcribed genes (FPKM>median), binned by gene length according to quartiles for transcribed genes. g, Same analysis as in (f), binned for long genes of 50-100 kb (upper panel) or >100 kb (lower panel).

Supplementary Figure 5. Incubation of Halotag ligand alone in Mock samples did not cause unspecific background localizations. a,
Average USP1 density per nucleus. (mean -/+ std, n=166,183,76, 109 nuclei for Mock, USP1-WT, USP1-C90S, and USP1-GG/AA, respectively, pvalues were calculated using the student's t-test). Boxes represent the 1 st (25%) and 3 rd (75%) percentile. b, Measurements of the average EdU density around USP1 (left), and the average USP1 density around EdU (right) are shown, based on at least two independent experiments (mean -/+ std, left: n= 183, 188, 71, 113 nuclei for Mock, WT, C90S, and GG/AA respectively; right n=207, 188, 71, 113 nuclei for Mock, WT, C90S and GG/AA respectively, p-values were calculated using the student's t-test). Boxes represent the 1 st (25%) and 3 rd (75%) percentile. c, Plots of the EdU density as a function of the USP1 density of the same nucleus (black dots, n = 183, 76, 109 nuclei for WT, C90S, and GG/AA, respectively). The linear regression (red line) is performed for all the samples, and no significant correlation is observed for all the samples. Figure 6. GG/AA mutant protein is enriched at replication forks by iPOND analysis. a, Immunoblotting analysis of nuclear soluble and chromatin-bound fractions prepared from USP1 WT and mutant cell lines. Histone H3 was used as a positive control marker for the chromatin fraction. Relative levels of USP1 (compared to WT=1 ) in the nuclear soluble and chromatin fractions were quantified by densitometry analysis. Mean with SD are plotted for 3 independent experiments and p-values were calculated using t-test with Welch's correction (*p<0.05, **p<0.01, two-tailed). b, HCT116 and USP1-GG/AA cells were pulse-labeled for 10 min with 10 µM EdU and then either collected directly (+EdU) or treated with 10 uM thymidine for 1 hr (+EdU + Chase) or 2 mM HU for 3 hours (+EdU +HU). iPOND and input samples were analyzed by immunoblotting with the indicated antibodies. Figure 7. SPRTN depletion causes chromatin-enrichment of the USP1-C90S protein. a, HCT116 cell lines were treated with or without an alternative siRNA for SPRTN (siSPRTN-7) and subjected to DNA fiber analysis of fork speed, as previously described. CldU tract lengths of elongating forks are plotted with mean and -/+ SD for 3 independent experiments (n=200 elongating forks), and p values were calculated using the Mann-Whitney rank-sum t-test (ns=no significance, *p<0.05, **p<0.01, ****p<0.0001, two-tailed). b, HCT116 USP1-C90S cells were treated with either siCtrl or siSPRTN-1 siRNAs and complemented with siRNA-resistant Flag-SPRTN constructs prior to immunoblotting analysis for the indicated proteins. c, HCT116 USP1-GG/AA#1 cells overexpressing Flag-SPRTN-WT were compared to EV for measurements of elongating CldU Tract length. CldU tract lengths are plotted for 3 independent experiments (n=200 elongating forks) with mean -/+ SD indicated and p values were calculated using the Mann-Whitney rank-sum t-test. d, U2OS Cerulean-PCNA cells expressing siRNA-resistant HaloTag-USP1 fusion proteins were treated with or without siUSP1-1 and ML323 as indicated and subjected to immunoblotting analysis. e, (left graph) Average USP1 localizations per nucleus as a function of time (mean -/+ SEM, n= 16, 33, 45, and 33 nuclei for WT, C90S, GG/AA, andQ672A, respectively). Lines are the fitted single-exponential decay model. (right graph) The fitted bulk tau_on of different USP1 molecules as indicated. Mean -/+ std is the fitting result of the time course of tau_on in left graph. f, Cells were treated with or without siSPRTN-1 siRNA for 72 h and subjected to immunoblotting analysis of prepared nuclear soluble and chromatin fractions. Relative levels of chromatin-bound USP1 were quantified using densitometry analysis (normalized to WT=1) and the mean with SD are plotted for 3 independent experiments. P-values were calculated using t-test with Welch's correction (ns=no significance, *p<0.05, two-tailed).

Supplementary Movie 1 (see Source Files
). Representative SR blinking movies for the same cell shown for AF647 for EdU signal as SR movie. Scale bar = 2000 nm. Only 500 frames are shown in order to reduce the file size.
Supplementary Movie 2 (see Source Files). Representative SR blinking movies for the same cell shown for JF549 for HaloTag USP1 signal as SR movie. Scale bar = 2000 nm. Only 500 frames are shown in order to reduce the file size.

Software and Algorithms
Software Source ImageJ v1.52a Where #/$ ( ) is the local density of species P/Q at a given location in the image; 〈⋯ 〉 denotes the average of all the locations across the image. As the P-Q pairs in the nucleus are randomly oriented, the 2D #$ ( ) = #$ ( , ) is integrated across from 0 to 2π, resulting in the 1D radial distribution function #$ ( ). 〈 # ( ) $ ( + )〉 is the pairing density 2 at distance , and thus 〈 # ( ) $ ( + )〉 〈 # ( )〉 ⁄ , or equivalently #$ ( )〈 $ ( )〉 , denotes the average local density of Q that pairs with each P at distance , and vice versa. The amplitude, , of the cross-correlation (equation 4) is the ( = ! ) where the cross-correlation reaches the maximum, and 〈 $ ( )〉 is used to estimate the average local density of Q pairing with each P. Note that if P and Q are randomly distributed (no crosscorrelation signal) around each other, the local pairing density is 0. This distinguishes from direct counting the local density around each other, which would result in high local surrounding density when P and Q are densely but randomly distributed around each other.