Poly(ADP-ribosyl)ation of TIMELESS limits DNA replication stress and promotes stalled fork protection

SUMMARY

Poly(ADP-ribosyl)ation (PARylation), catalyzed mainly by poly(ADP-ribose) polymerase (PARP)1, is a key posttranslational modification involved in DNA replication and repair. Here, we report that TIMELESS (TIM), an essential scaffold of the replisome, is PARylated, which is linked to its proteolysis. TIM PARylation requires recognition of auto-modified PARP1 via two poly(ADP-ribose)-binding motifs, which primes TIM for proteasome-dependent degradation. Cells expressing the PARylation-refractory TIM mutant or under PARP inhibition accumulate TIM at DNA replication forks, causing replication stress and hyper-resection of stalled forks. Mechanistically, aberrant engagement of TIM with the replicative helicase impedes RAD51 loading and protection of reversed forks. Accordingly, defective TIM degradation hypersensitizes BRCA2-deficient cells to replication damage. Our study defines TIM as a substrate of PARP1 and elucidates how the control of replisome remodeling by PARylation is linked to stalled fork protection. Therefore, we propose a mechanism of PARP inhibition that impinges on the DNA replication fork instability caused by defective TIM turnover.

Graphical Abstract

In brief

Rageul et al. report that TIMELESS harbors poly(ADP-ribose) (PAR)-binding motifs, which are essential for the PARylation of TIMELESS by PARP1. Defects in this modification prevent proteasomal degradation of TIMELESS and cause DNA replication fork instability. The proteotoxic stress caused by dysregulated TIMELESS turnover underlies a mechanism of PARP inhibitor action.

INTRODUCTION

Poly(ADP-ribose) polymerases (PARPs) comprise a family of ADP-ribosyl transferases that catalyze the transfer of ADP-ribose moieties from NAD⁺ to trigger poly(ADP-ribosyl)ation, or PARylation, of proteins involved in a variety of cellular pathways, including genome maintenance.¹ PARP1 is a founding member that is known for its essential role in the DNA damage response. Upon recognition of DNA breaks, PARP1 undergoes an allosteric change that leads to the opening of a catalytic domain responsible for conjugating linear or branched PAR polymers onto target proteins as well as onto PARP1 itself.^2,3 The PAR modification occurs rapidly and is counterbalanced by its degrading enzyme PAR glycohydrolase (PARG) and ADP-ribosylserine hydrolase.⁴ Aspartate, glutamate, and lysine are known as major target residues for the modification, and recent studies have revealed that serine ADP-ribosylation is predominant under DNA damage, which is augmented by histone PARylation factor 1 (HPF1), a cofactor of PARP1 that jointly creates a catalytic core (CAT) that is necessary for serine modification.^5–8 Similar to other posttranslational modifications, covalently attached PAR polymers are read by distinct classes of PAR-binding motifs, including PAR-binding zinc finger, macrodomain, WWE, and PAR-binding motif (PBM), that are present in noncovalent binding partners.⁹ This in turn recruits downstream reader proteins, allowing regulation through protein-protein interactions, localization, and stability.¹⁰

Emerging evidence supports the essential role of PARP1 at DNA replication forks. Specifically, unligated Okazaki fragment intermediates during lagging strand synthesis induce PARP activity, activating DNA single-strand break (SSB) repair to fill single-stranded DNA (ssDNA) gaps.¹¹ PARP1 activity restricts the speed of fork elongation in part via the regulation of CDKN1A/p21.¹² In response to camptothecin (CPT)-induced replication damage, PARP1 is required for replication fork reversal, a process involving the regression of a stalled fork to form a four-way junction to stabilize the damaged fork and act as an intermediate for repair and restart.^13,14 PARP1 recruits MRE11 to regulate limited processing of DNA and stabilizes reversed forks by restricting the RECQ1 helicase, which would otherwise promote uncontrolled fork restart.^15,16 In addition, coactivator-associated arginine methyltransferase 1 (CARM1) directly interacts with PARP1 and promotes its activity, thereby favoring fork reversal, while limiting ssDNA gap accumulation and error-prone damage tolerance.¹⁷ Accordingly, PARP inhibitors (PARPis) exploit the replication vulnerability of cancer cells, in which the inhibition of PARP activity specifically kills breast cancer gene 1/2 (BRCA1/2)-deficient tumors that exhibit compromised DNA repair and stalled fork protection.¹⁸ Multiple mechanisms of PARPi action have been proposed, including the inhibition of SSB repair, formation of a PARP1-DNA complex (i.e., PARP trapping), and accumulation of replication-associated ssDNA gaps.^19–21 Nevertheless, it is still not fully understood how the multifaceted function of PARP1 counteracts the instability of stalled forks. This knowledge would reveal the additional determinants of PARPi action and acquired drug resistance.

In this study, we identify a role of PARP1 in preserving fork integrity via its PARylation of TIMELESS (TIM) within the replisome. TIM and its heterodimeric partner TIPIN (TIM-interacting protein) are the key constituents of the fork protection complex (FPC), which tethers the cell division cycle 45 (CDC45)/minichromosome maintenance 2–7 (MCM2–7)/Go-Ichi-Nii-San (GINS) (CMG) helicase and replicative polymerase activities to prevent replisome uncoupling and guide efficient fork progression.²² The TIM-TIPIN complex is localized at the leading edge of CMG, where it grips double-stranded DNA (dsDNA) and directly interacts with the MCM subunits, thereby facilitating parental strand separation while restraining replisome activity under replication stress.^23–25 We previously showed that TIM and its regulatory partner silencing defective 2 (SDE2) prevent the nucleolytic degradation of reversed forks.²⁶ Nevertheless, a detailed mode of TIM action that modulates the replisome activity is lacking. Intriguingly, previous studies revealed that TIM and PARP1, but not PARP2 or PARP3, directly interact.^27,28 Although the participation of the PARP1-TIM complex in the early step of DNA double-strand break signaling was revealed,²⁹ the physiological role of this interaction at DNA replication forks remains uncharacterized. Here, we demonstrate that the TIM-PARP1 interaction primes TIM for PAR-dependent proteolytic degradation to control the replisome dynamics under replication stress. Not only is proper turnover of TIM necessary for preventing its aberrant accumulation at replication forks to suppress replication stress but it also allows for remodeling of the replisome necessary for stalled fork protection. Our study elucidates an important physiological relevance of the PARP1 activity that protects replication fork integrity in coordination with the replication machinery. We propose that PARPis compromise fork stability in part by impeding TIM proteolysis and thereby causing a proteotoxic stress analogous to PARP trapping, which exacerbates the replication stress of BRCAness.

RESULTS

The TIM-PARP1 interaction at replication forks is essential for preserving fork integrity

Since our prior study revealed that TIM works on the fork stabilization pathways that are shared by the PARP1 function,²⁶ we hypothesized that the TIM-PARP1 interaction may constitute an important mechanism to maintain DNA replication fork integrity. We confirmed the interaction between endogenous TIM and PARP1 from cells in S phase (Figure 1A). In line with the TIM localization at ongoing replication forks,²⁶ we visualized the association of PARP1 at biotin-5-ethynyl-2ʹ-deoxyuridine (EdU)-labeled replication forks using a proximity-ligation assay (PLA) (Figure 1B). We also observed the TIM-PARP1 interaction by PLA, which was enriched in EdU⁺ cells, indicating that TIM and PARP1 integrate into the replication machinery during DNA replication (Figure 1C). Previous X-ray crystallography and biochemical studies showed that the PARP1-binding (PAB) region of TIM within its C terminus directly interacts with the CAT within the C terminus of PARP1 via a series of electrostatic interactions grouped at two independent sites (Figure 1D).^27,28 We verified that mutations of key residues in these sites from TIM or PARP1 abrogate the TIM-PARP1 interaction, leading to the generation of a TIM E1049Q/E1056Q/T1078D (EQ/EQ/TD) mutant and a reciprocal PARP1 mutant harboring K940G/K943Q/D993G (KG/KQ/DG) mutations (Figures 1E, 1F, and S1A). As previously described, the T1078D mutation by itself abolished the TIM-PARP1 interaction, indicating that site I constitutes a major interface for the PARP1 binding.²⁸ The TIM EQ/EQ/TD mutant maintained its interaction with its obligate partner TIPIN, indicating that it is a separation-of-function mutant that is able to preserve the integrity of the FPC (Figure S1B). To determine the role of the TIM-PARP1 interaction in replication fork stability, we reconstituted TIM wild-type (WT) or EQ/EQ/TD mutant into U2OS cells by knocking down endogenous TIM and expressing small interfering RNA (siRNA)-resistant FLAG-tagged TIM variants from a single FRT (Flp-recombination target) locus in response to doxycycline (Figure 1G). Using this Flp-In system, we were able to reconstitute TIM-depleted cells with WT or EQ/EQ/TD mutant to near-endogenous levels (Figure 1G). The expression of FLAG-TIM resistant to either cDNA- or 3ʹ UTR-targeting siRNA oligonucleotides was sufficient to suppress RPA32 phosphorylation at S4/S8, a marker for replication stress and strand breakage, elevated by TIM knockdown, confirming that TIM is functionally complemented in this system (Figure S1C). We also verified that the reconstituted FLAG-TIM WT interacts with endogenous PARP1, whereas the mutant failed to do so (Figure 1H). The proximity between PARP1 and EdU-labeled replication forks was significantly decreased in cells expressing the TIM EQ/EQ/TD mutant, further supporting that the TIM-PARP1 interaction occurs at ongoing replication forks (Figure S1D). DNA combing showed that TIM Flp-In cells expressing the TIM mutant exhibit impaired fork progression and fail to counteract the hyper-resection of hydroxyurea (HU)-induced stalled forks in TIM-knockdown cells (Figures 1I and 1J). Together, these results underscore the importance of the TIM-PARP1 interaction in preserving the integrity of DNA replication forks.

Figure 1. The TIM-PARP1 interaction at DNA replication forks is essential for preserving fork integrity.

(A) coIP of endogenous TIM and PARP1 in 293T cells released from G1/S block.

(B) Left: PLA of PARP1 with EdU-labeled replication forks in U2OS WT versus PARP1 KO cells. Right: quantification of cells positive for PARP1:EdU PLA foci. Scale bar: 10 μm; **p < 0.01, Student’s t test.

(C) Left: The PARP1:TIM PLA in U2OS cells. S phase cells were marked with EdU. Right: percentages of cells positive for PLA foci in EdU positive or negative cells. Scale bar: 10 μm; **p < 0.01, Student’s t test.

(D) Schematic of TIM-PAB and the catalytic domain of PARP1.

(E) Anti-FLAG coIP of TIM variants with endogenous PARP1 in 293T cells. EV, empty vector.

(F) Anti-myc coIP of TIM-myc with FLAG-PARP1 variants.

(G) Top: schematic of the Flp-In T-REx system to express siRNA-resistant FLAG-TIM WT or EQ/EQ/TD by doxycycline (dox) mediated derepression of the promoter. Bottom: verification of U2OS Flp-In cell lines reconstituting TIM WT or EQ/EQ/TD in response to increasing doses of dox, following transfection of TIM siRNA.

(H) Anti-FLAG coIP of endogenous PARP1 by TIM WT or EQ/EQ/TD from the Flp-In cells.

(I) Dot plot of the DNA fiber iodo-deoxyuridine (IdU) track length from Flp-In cells. Red bars indicate the median from at least 150 tracks. ****p < 0.0001, n.s., not significant; two-way ANOVA with post hoc test and false discovery rate (FDR) adjustment.

(J) Dot plot of the IdU/chloro-deoxyuridine (CldU) ratios from Flp-In cells treated with 4 mM HU for 4 h. ****p < 0.0001; two-way ANOVA.

(B and C) Mean ± SD from 3 biological replicates. (I and J) Representative plot from 3 biological replicates.

See also Figure S1.

TIM is PARylated by PARP1 in a PBM-dependent manner

The direct interaction between TIM and PARP1 raises the possibility that TIM is modified by PARP1. Intriguingly, we identified two previously uncharacterized PBMs, which are defined by two positively charged amino acid residues, lysine or arginine, at position 5 and 6, preceded by a positively charged residue at position 1 and followed by a residue with an alkyl side chain at position 7 (Figure 2A). Both PBMs from TIM exhibit a high degree of similarities with known PBMs, and the second PBM (PBM2) is located adjacent to the PAB region at its C terminus. Point mutations in two lysine residues from the recombinant PBM2 polypeptide (K949A/K950A) substantially decreased its interaction with biotinylated PAR chains in vitro (Figure 2B). Similarly, the PBM1 K537A/K538A mutant peptide exhibited a less efficient interaction with PAR chains (Figure 2C). To evaluate the role of TIM PBMs in the physical and functional interactions with PARP1, we established an in vitro system in which auto-PARylation of recombinant PARP1 was triggered by NAD⁺ in the presence of dsDNA and suppressed by olaparib, a PARPi (Figure 2D). Maximum PARP1 auto-PARylation was achieved by ~50 mM of NAD⁺ in our condition (Figure S2A). The KG/KQ/DG TIM binding-deficient PARP1 mutant underwent auto-PARylation comparable to WT, indicating that disruption of the TIM binding interface does not impair the catalytic activity of PARP1 (Figure S2B). This is consistent with a previous structure revealing that the TIM PAB faces opposite to the CAT of PARP1.²⁷ The addition of FLAG-TIM produced by in vitro transcription and translation (IVTT) did not affect the auto-PARylation of PARP1 (Figure S2C). Notably, FLAG-tagged TIM underwent PARylation by PARP1 as revealed by anti-FLAG immunoprecipitation (IP) in a denaturing condition and detection of conjugated biotin-NAD⁺ after in vitro reaction with a limited amount of PARP1 (Figure 2E, lanes 7 and 8). PARylated PARP1 was not captured in this IP, indicating that PAR signals are from immunoprecipitated FLAG-TIM (Figure 2E, lanes 1 and 5). Strikingly, point mutations in either TIM PBM1 or PBM2 significantly decreased TIM PARylation; PBM2 had a major effect, and mutations in both PBM1 and PBM2 abolished TIM PARylation (Figure 2F, lanes 9 and 12, and 2G). As a control, the TIM EQ/EQ/TD mutation abrogated the PARylation of TIM WT or PBM2 mutant, indicating that TIM recognizes auto-PARylated PARP1 (Figure S2D). To further confirm the specific recognition of auto-PARylated PARP1 by TIM, we set up the PARylation of histone H3.1 in vitro (Figure S2E). In this condition, IVTT FLAG-TIM efficiently coimmunoprecipitated PAR conjugates of PARP1, but not those of histone H3.1, suggesting that TIM specifically interacts with auto-PARylated PARP1 (Figure S2F). These results define TIM as a substrate of PARP1 and indicate that the recognition of auto-PARylated PARP1 by TIM PBMs is necessary for PARP1 to PARylate TIM (Figure 2H). Intriguingly, PARylation of bromodomain 7 (BRD7) by PARP1 was previously shown to be dependent on its intact PBM,³⁰ suggesting that recognition of auto-PARylated PARP1 via the PBM of substrates may be a general mechanism to engage PARP1 for efficient PARylation.¹⁹

Figure 2. TIM is PARylated by PARP1 in a PBM-dependent manner.

(A) Schematic of two conserved PBMs within TIM with the relative positions of binding partners TIPIN and PARP1.

(B) Top: schematic of the glutathione S-transferase-tagged recombinant TIM fragment containing amino acids with PBM2 WT or K949A/K950A (KA). Bottom: slot blot loaded with corresponding recombinant PBM2 and incubated with biotinylated PAR chains (pADPr).

(C) Top: schematic of synthesized PBM1 26 amino acid-peptides, either WT or with 4 mutations in basic residues. Bottom: slot blot of corresponding peptides incubated with pADPr.

(D) In vitro auto-PARylation assay with recombinant PARP1, activating dsDNA, and 50 μM NAD⁺, as well as 10 μM olaparib, where indicated.

(E) Anti-FLAG IP of IVTT FLAG-TIM WT reacted with recombinant PARP1 in the presence of 25 μM of biotinylated NAD⁺.

(F) Anti-FLAG IP of IVTT FLAG-TIM reacted with recombinant PARP1 in the presence of 25 μM of biotinylated NAD⁺.

(G) Quantification of (F). Mean ± SD from three biological replicates. *p < 0.05; ***p < 0.001; ****p < 0.0001; one-way ANOVA.

(H) Model for the recognition of auto-PARylated PARP1 by TIM (1) and subsequent PARP1-dependent TIM PARylation (2).

(I and J) Anti-FLAG IP of FLAG-TIM in denaturing conditions followed by anti-PAR western blot (WB) from 293T cells treated with 1 μM CPT and 10 μM PARGi (PDD00017273) with or without 10 mM olaparib pretreatment for 1 h.

(K) coIP of endogenous PARP1 and PARylated PARP1 by anti-FLAG TIM IP in chromatin-enriched fractions of U2OS cells treated with 1 μM CPT and 10 μM PARGi for 1 h.

(L) In vivo PARylation of FLAG-TIM in cells transfected with 2 independent siRNAs against HPF1 and treated with CPT/PARGi.

See also Figure S2.

We next determined whether the roles of TIM PBMs in TIM PARylation are physiologically relevant. CPT triggered cellular PARylation, which was detected as early as 30 min and reached a peak at 1 h (Figure S2G). In this condition, we also observed the PARylation of immunoprecipitated FLAG-TIM in cells challenged by CPT, which was abrogated by olaparib (Figure 2I, lanes 8 and 9). Importantly, CPT-induced TIM PARylation was substantially decreased by the PBM1/2 mutation in comparison to WT, indicating that recognition of PARP1 PAR chains by TIM PBMs is necessary for efficient TIM PARylation in cells (Figure 2J, lanes 10 and 12). The PAB TIM mutant showed a defect in PARylation as well (Figure S2H). Notably, FLAG-TIM was able to coimmunoprecipitate the auto-PARylated form of PARP1 that was increased by CPT, whereas the TIM PBM1/2 mutant failed to do so (Figure 2K, lanes 8 and 10). To further corroborate our finding that revealed a damage-inducible TIM PARylation, we knocked down HPF1, a cofactor of PARP1/2 that preferentially induces serine ADP-ribosylation under DNA damage (Figure S2I). Depletion of HPF1 with two independent siRNAs reduced cellular PARylation induced by CPT and also resulted in a substantial decrease in TIM PARylation (Figures 2L and S2I). Conversely, mutations in a series of serine residues identified from published proteomics studies abrogated CPT-induced TIM PARylation,^31,32 suggesting that TIM is a physiological substrate of the HPF1:PARP1 complex (Figure S2J). Together, these data suggest that TIM undergoes PARP1-dependent PARylation, which is pronounced under replication damage. This defines TIM PBMs as a regulatory element to potentiate TIM PARylation by PARP1.

PAR-dependent TIM degradation counteracts DNA replication stress

Recognition of PAR by PBMs engages multiple downstream processes, including regulation of protein stability. Notably, exogenous TIM is expressed at higher levels in PARP1 knockout (KO) U2OS cells, which we generated by CRISPR-Cas9 gene editing, and was suppressed by ectopic PARP1 (Figures 3A and 3B). The ability of PARP1 to decrease TIM levels was counteracted by proteasome inhibition via MG132, indicating that PARP1 down-regulates TIM through the ubiquitin (Ub)-proteasome system (Figure 3A, lanes 5 and 6). Importantly, the TIM PBM1/2 mutant was expressed at much higher levels in cells compared to TIM WT, and its elevated expression was not further increased in PARP1 KO cells (Figure 3C). Furthermore, mutations in PBM1, PBM2, or both elevated the levels of FLAG-TIM when transiently expressed in cells, but not when FLAG-TIM variants were produced in vitro, suggesting that the cellular machinery of proteolysis is responsible for the degradation of PARylated TIM (Figure 3D).

Figure 3. PAR-dependent TIM degradation counteracts DNA replication stress.

(A) Left: TIM-myc and FLAG-PARP1 were transiently expressed in U2OS or PARP1 KO cells and analyzed by WB. Where indicated, 10 mM MG132 was treated for 4 h.

(B) Quantification of TIM with each MG132 condition normalized to its corresponding untreated condition. *p < 0.05; **p < 0.01; Student’s t test.

(C) WB analysis of FLAG-TIM WT or PBM1/2 levels in U2OS or PARP1 KO cells.

(D) Expression of TIM variants in U2OS cells versus IVTT expression.

(E) Anti-FLAG IP of FLAG-TIM in denaturing conditions followed by anti-PAR WB in 293T cells treated with 10 μM PARGi with or without 1 μM CPT for 1 h.

(F) Ubiquitination assay of TIM-myc WT and PBM1/2 overexpressed in 293T cells with His-Ub and treated with 1 μM CPT and 10 μM MG132.

(G) Analysis of pRPA32 S4/S8 in U2OS cells transfected with increasing amounts of FLAG-TIM WT or PBM1/2.

(H) Structure depicting the MCM PI of TIM that includes anchor and DBM, and their conservation throughout multiple species. Adapted from PDB: 7PFO and created using Chimera.

(I) coIP of FLAG-TIM PBM1/2 or PBM1/2 ∆PI (amino acid D273–281 and amino acid ∆316–322) by endogenous MCM6 in 293T cells.

(J) Analysis of pRPA32 S4/S8 in U2OS cells induced by FLAG-TIM WT, PBM1/2, or PBM1/2 ∆PI.

(K) The PLA between TIM and PARP1. A total of 10 μM olaparib was treated for 24 h, and 10 μM EdU was added for the last 30 min before fixation. Scale bar: 10 μm ****p < 0.0001; two-way ANOVA.

(B and K) Mean ± SD from 3 biological replicates.

See also Figure S3.

The requirement of intact PBM1/2 motifs for the steady levels of cellular TIM suggests that PARylation may regulate TIM degradation in unperturbed conditions. We were able to detect transient TIM PARylation when a PARG inhibitor was applied to suppress the disassembly of PAR chains, which was further increased after CPT treatment (Figure 3E). The TIM PARylation without damage was substantially decreased when the PBM1/2s were mutated (Figure S3A). Accordingly, the polyubiquitination of TIM in response to CPT was compromised in the PBM1/2 mutant, suggesting that TIM PARylation is necessary for Ub-dependent degradation of TIM (Figures 3F and S3B). The stability of TIM is supported by its direct interaction with TIPIN that docks into a hydrophobic interface of TIM.²³ The interaction between TIM and TIPIN remained intact following CPT treatment, excluding the possibility that TIM PARylation disrupts the interaction of TIM with TIPIN, causing degradation (Figure S3C). Rather, given that selected E3 ligases are known to function as PAR-targeted Ub ligases (PTUbLs) that use PAR as a recognition signal, these results point to the existence of a signaling cascade for TIM degradation, in which recognition of auto-PARylated PARP1 by TIM PBMs engages PARP1 to catalyze TIM PARylation, leading to the recruitment of PTUbL(s) to TIM and subsequent polyubiquitination.

Because cellular TIM levels are regulated by its PAR modification, and constant PARP1 activity is coupled to DNA replication,¹¹ we next sought to determine the effect of dysregulated TIM proteolysis in replication fork stability. Notably, increasing the expression of ectopic TIM PBM1/2 mutant intensified the RPA32 phosphorylation at S4/S8 compared to WT, indicating that the integrity of DNA replication forks is compromised (Figure 3G). We observed that a large portion of TIM is fractionated into the nonchromatin fractions, whereas PARP1 is mainly localized in the chromatin, indicating that a surplus of TIM exists outside of the replication fork, which may be distinct from the population that integrates into the replisome to promote fork elongation (Figure S3D). We reasoned that a defect in TIM turnover may lead to the aberrant engagement of excessive TIM to the replisome, thus hindering the progression of CMG and causing replication stress. The previous cryoelectron microscopy structure of the human replisome revealed that TIM harbors a highly conserved MCM plugin (PI) domain that is required for tethering the TIM-TIPIN complex to the replisome, which includes the ‘‘anchor’’ that attaches to MCMs and the ‘‘DNA-binding motif (DBM)’’ that interacts with parental dsDNA (Figure 3H).²³ Therefore, we introduced deletion of the PI domain (∆PI) on top of PBM1/2 mutations, in which both the anchor and the DBM regions are absent, and this mutant subsequently failed to be pulled down by MCM6 (Figure 3I). Notably, the DPI mutant suppressed the pRPA32 S4/S8 elevated by the PBM1/2 mutations, indicating that the illegitimate engagement of TIM PBM1/2 to CMG is detrimental (Figure 3J). The interaction between TIM and PARP1 was dramatically increased upon treatment of multiple PARP inhibitors, suggesting that the downstream TIM degradation step is impaired (Figure S3E). The proximity between TIM and PARP1 also increased upon olaparib treatment (Figure 3K). Furthermore, the association of TIM to EdU-labeled replication forks increased when proteasome was inhibited (Figure S3F). Together, these results suggest that PARP1-dependent TIM PARylation is coupled to TIM degradation; its dysregulation leads to DNA replication stress due to the excessive accumulation of TIM that becomes inadvertently engaged with the replisome.

Damage-dependent TIM depletion at replication forks requires TIM PARylation

Given our data showing that the TIM-PARP1 interaction is necessary for preventing the destabilization of stressed forks, we further hypothesized that PAR-dependent TIM degradation under replication damage may constitute an essential step for the protection of stalled forks. To test this idea, we synchronized cells at the G1/S boundary and released them into S phase in the presence of CPT to cause fork stalling at a stabilized topoisomerase I-DNA complex. Cell-cycle analysis showed that when released from synchronization, U2OS cells complete the transition to G2/M within 8 h, which is essentially blocked by CPT (Figure S4A). Although endogenous TIM without damage maintained its levels at the chromatin-enriched (P) fraction until 8 h, at which point cells completed DNA synthesis and entered G2/M, we observed the downregulation of TIM levels soon after release into S phase under CPT damage (Figure 4A, lanes 3 and 7). The decrease in endogenous TIM from the P fraction was prevented by MG132, a proteasome inhibitor, or bortezomib (BTZ), a selective inhibitor of the 26S proteasome, indicating that Ub-dependent degradation of TIM at stalled forks is responsible for TIM downregulation (Figure 4B). TIM levels in soluble fractions were largely unchanged, indicating that degradation occurs specifically in the context of DNA (Figures S4B and S4C). Inhibition of Ub-activating enzyme (E1) activity similarly prevented chromatin-associated TIM degradation after CPT treatment (Figure S4D, lanes 10 and 11). The TIM degradation in the P fraction was impaired by olaparib (Figures 4C and S4E) and in PARP1 KO cells (Figure S4F), suggesting that PARP1 activity is required for TIM degradation at stalled forks.

Figure 4. Damage-dependent TIM depletion at replication forks requires TIM PARylation.

(A) U2OS cells were synchronized at the G1/S boundary by double thymidine block and released in the presence or absence of 1 μM CPT for 0, 2, 4, and 8 h to monitor TIM levels at the chromatin (P) fraction.

(B) As in (A), but cells were pretreated in the last hour of the second thymidine block with 10 μM MG132 or 1 μM BTZ.

(C) U2OS cells were synchronized and released for 1 h with or without 10 μM olaparib, followed by 1 μM CPT treatment for 4 h.

(D) Reconstitution of TIM WT and PBM1/2 in Flp-In cells.

(E) Flp-In cells were synchronized and treated with 1 μM CPT during the release.

(F) Quantification of TIM levels -normalized to the WT untreated condition (0 h release).

Mean ± SD from 3 biological replicates. *p < 0.05; **p < 0.01; two-way ANOVA.

See also Figure S4.

To determine the role of regulated TIM degradation at stalled forks, we again implemented the Flp-In system to reconstitute TIM-depleted cells with FLAG-TIM WT or PBM1/2 mutant. The PBM1/2 mutant expressed at a higher level than WT, indicating that the PBM1/2 mutation stabilizes cellular TIM in our system (Figure 4D). Importantly, in contrast to WT, the PBM1/2 mutant failed to be downregulated when cells were released into S phase in the presence of CPT, suggesting that TIM PARylation is necessary for TIM degradation at stalled forks (Figure 4E, lanes 5 and 11, 4F, and S4G). Defective TIM degradation was accompanied by the elevated levels of pRPA32 at S4/S8, which indicates increased replication damage, specifically in the chromatin-enriched fraction (Figures 4E and S4H). Collectively, these results suggest that PARP1-mediated degradation of TIM occurs in response to fork-stalling DNA lesions, which is necessary for counteracting replication stress.

Conditional degradation of TIM is required for stalled fork protection and recovery

We next sought to identify the mechanism through which regulated TIM turnover is linked to replication fork integrity under replication damage. In unchallenged conditions, Flp-In cells reconstituted with the PBM1/2 mutant exhibited a fork progression defect, as revealed by DNA combing (Figure 5A), which may reflect the notion that the aberrant accumulation of TIM at replication forks induces replication stress during fork progression. We then applied HU to uniformly induce fork stalling and facilitate the analysis of fork recovery and protection by DNA combing. Similar to CPT, HU triggered a time-dependent increase in PARylation in both asynchronous and G1/S-synchronized cells (Figure S5A). Moreover, the PBM1/2 mutant exhibited impaired PARylation in response to HU (Figure S5B). In this condition, we observed a significant defect of PBM1/2 mutant-expressing cells in recovering from HU-induced stalled forks in comparison to WT (Figure 5B). Furthermore, cells expressing the PBM1/2 mutant failed to rescue the hyper-resection of stalled forks caused by TIM depletion, indicating that a defect in TIM degradation causes compromised stalled fork protection (Figure 5C). Stalled fork protection is often mediated by fork reversal via the actions of RAD51 recombinase and several SWI/SNF family translocases, including SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin, subfamily a like 1 (SMARCAL1).^33,34 Co-depletion of SMARCAL1 rescued the hyper-resection of stalled forks in PBM1/2-expressing cells, suggesting that TIM degradation is necessary for the proper protection of reversed forks (Figure 5D). To further show that the compromised fork protection in the PBM1/2 mutant-expressing cells occurs at reversed forks, we examined the ssDNA exposure of the nascent DNA strands that are marked by BrdU foci upon a short BrdU incubation before fork stalling as a readout of fork reversal.^35–37 The extent of BrdU foci was similar for both WT and PBM1/2 mutants, arguing for the formation of ssDNA resulting from 3ʹ end resection at reversed forks in both cases (Figure 5E). On the contrary, the PBM1/2 mutation impaired the induction of the PLA between RAD51 and EdU-labeled replication forks under HU damage, suggesting that defective RAD51 loading onto reversed forks underlies the defect in stalled fork protection observed in the PBM1/2 mutant-expressing cells (Figure 5F).

Figure 5. Conditional degradation of TIM is required for stalled fork protection and recovery.

(A) Dot plot of DNA fiber IdU track length from the Flp-In cells expressing FLAG-TIM WT or PBM1/2.

(B) Fork recovery measured as the portion of the number of red tracks divided by the number of red-only and red-green fibers.

(C) Dot plot of the DNA fiber IdU/CldU track length ratio from the indicated Flp-In cells.

(D) Fork protection assay as in (C) from Flp-In PBM1/2 transfected with siRNA control or SMARCAL1.

(E) Left: schematic of fork reversal with the exposure of ssDNA resulting from 3ʹ end resection visualized by native BrdU labeling. Center: representative images of BrdU foci in indicated Flp-In cells that were pulsed with 10 μM BrdU for 25 min, followed by 4 mM HU for 2 h. Right: quantification of BrdU⁺ cells, positivity defined by >10 foci. Scale bar: 10 μm.

(F) Left: representative images of RAD51-EdU PLA foci in Flp-In cells treated with 3 mM HU for 5 h. Scale bar: 10 μm. Right: quantification of cells positive for RAD51-EdU PLA foci.

(G) Dot plot of the DNA fiber IdU/CldU track length ratio from the indicated Flp-In cells.

(A–G): ****p < 0.0001; two-way ANOVA with post hoc test and FDR adjustment, except for (D): Kruskal-Wallis with post hoc test and FDR adjustment. (A, C, D, and G) Representative plot from at least 2 biological replicates. (B, E, and F) Mean ± SD from 3 biological replicates.

See also Figure S5.

To further understand how the defective TIM degradation under replication damage affects the processing of stalled forks, we turned our attention to the remodeling of the replisome, since TIM is a direct binding partner and regulator of CMG. Recent studies suggest that the replicative helicase does not dissociate from a stalled fork but is repositioned in a parental ssDNA bubble that is formed ahead of a reversed fork.^38,39 We determined whether regulated TIM turnover promotes this process. We used the PLA between MCM6 and EdU-labeled forks as a readout for the uncoupling of the replisome and repositioning of CMG. Flp-In cells expressing TIM WT exhibited a significant decrease in PLA signals upon HU treatment, whereas the PBM1/2-mutant cells failed to do so, indicating that the replisome remodeling is impaired (Figure S5C). The PLA signals from the PBM1/2 mutant cells were lower to start with, which may reflect the fork progression defect observed in these cells during unchallenged conditions.

We then tested whether disruption of the TIM-CMG interaction bypasses the requirement of TIM degradation and thus alleviates the phenotypes of defective fork protection. To this end, we generated another Flp-In cell line that reexpresses the PBM1/2 mutant with an additional mutation in the MCM PI in a doxycycline-dependent manner (PBM1/2 ∆PI) (Figure S5D). The TIM PBM1/2 ∆PI-expressing cells exhibited higher levels of nascent DNA strand exposure compared to WT or PBM1/2, indicating that fork remodeling and processing is more active (Figure 5E). Importantly, the additional mutations in the TIM MCM PI restored the RAD51 loading onto stalled forks and counteracted the hyper-resection of stalled forks in PBM1/2 mutant-expressing cells (Figures 5F and 5G). This suggests that TIM needs to be disengaged in a timely manner to facilitate CMG repositioning and that the aberrant engagement of TIM to CMG impedes the processing of stalled forks. Interestingly, RAD51 loading was further stimulated by the PBM1/2 ∆PI reconstitution when compared to WT, indicating that abnormal circumvention of TIM-dependent replisome remodeling may dysregulate fork reversal and protection processes (Figure 5F). Together, these data provide evidence that PAR-dependent regulated TIM turnover is essential for remodeling of the replisome under replication damage, which allows for subsequent DNA-protein transactions necessary for stalled fork protection and recovery.

Defective TIM degradation sensitizes BRCA2-deficient cells to replication stress

Our results thus far suggest that increased TIM engagement at replication forks may cause the cytotoxicity associated with noncovalent DNA-protein crosslinks, similar to PARP trapping induced by PARPis. Defective PARP1 turnover and accumulation at DNA lesions is known to be an underlying cause of the replication stress induced by PARPis, which results in synthetic lethality, with cells exhibiting BRCAness. We therefore determined whether defective TIM degradation creates the replication stress that further sensitizes cells to chemotherapy or PARPis in conjunction with BRCA2 deficiency (Figure 6A). We observed that cells reconstituted with TIM PBM1/2 exhibit modest but statistically significant hypersensitivity to CPT in comparison to the WT counterpart upon BRCA2 knockdown, indicating that TIM-trapped replication forks are vulnerable to additional replication-associated DNA lesions (Figure 6B). Moreover, depleting BRCA2 in the PBM1/2-mutant cells rendered them more sensitive to olaparib when compared to the WT cells without BRCA2, indicating that defective TIM degradation potentiates the cytotoxicity of PARPis in combination with BRCA2 deficiency (Figure 6C). Together, these results reveal an additional mechanism of PARPi action besides its own trapping to DNA, whereby accumulation of TIM at stalled forks induces DNA replication stress that further compromises DNA replication fork integrity.

Figure 6. Defective TIM degradation sensitizes BRCA2-deficient cells to replication stress.

(A) WB analysis of siRNA BRCA2 in Flp-In cells.

(B and C) Cytotoxicity of Flp-In WT or PBM1/2 cells transfected with siRNA control or BRCA2 and treated with increasing doses of CPT or olaparib, measured by ATP-dependent luminescence. At least n = 4; *p < 0.05; **p < 0.01; two-way ANOVA.

(D) Model for the role of PAR-dependent TIM degradation in DNA replication and stalled fork protection. See Discussion for details.

DISCUSSION

In this study, we propose a mechanism through which PAR-mediated degradation of TIM promotes replication fork integrity (Figure 6D). Both PARP1 and TIM associate with the replication machinery, and recognition of auto-PARylated PARP1 via the conserved PBM motifs in TIM engages PARP1 to catalyze TIM PARylation, thereby priming TIM for proteolysis.

On the one hand, PAR-dependent TIM degradation prevents inadvertent accumulation of TIM at active forks, which is required for restricting the engagement of surplus TIM with the replisome that is present outside of replication forks. We demonstrated that a defect in TIM PARylation elevates the replication stress that is dependent on MCM proteins and DNA interactions of TIM. Cellular levels of TIM near DNA replication forks may be constantly monitored by the replication-associated PARP1 activity. The transient but continuous PARylation at ongoing replication forks has been noted, which represents the PARP1-dependent back-up Okazaki fragment processing.¹¹ Our study suggests that basal levels of PARylation during DNA synthesis may be necessary for controlling replisome integrity and thus suppressing replication stress via regulation of TIM proteolysis. Similarly, controlled PARylation has been linked to promote the optimal recruitment of DNA repair factors to replication forks.⁴⁰ The reason why cells express TIM in excess is not clear. Unlike the replicative helicases that cannot be reloaded during S phase, TIM may undergo dynamic turnover during replication to control the rate of fork progression and readily respond to replication-blocking lesions. Intriguingly, it was shown that a surplus of DNA-bound MCMs exists to restrain fork speed, indicating that the pace of replication forks is tightly controlled by the multiple constituents of the replisome.⁴¹

On the other hand, conditional PARP1-dependent TIM degradation at stalled forks contributes to the remodeling of the replisome, which constitutes an essential step for protecting reversed forks. Accordingly, defective TIM proteolysis, either by mutations in PBMs or PARP1 inhibition, leads to impaired RAD51 loading at reversed forks, exacerbated fork degradation, and defective stalled fork recovery. Disrupting the interaction of the nondegradable TIM mutant with MCMs and dsDNA alleviates these adverse phenotypes, suggesting that TIM needs to be temporarily removed from stalled forks via degradation, opening the way for CMG to reposition ahead of a stalled fork. In line with this notion, recent studies have reported a dynamic remodeling of the replisome necessary for the processing of stalled forks. The strand exchange activity of RAD51 circumvents the replicative helicase, which remains bound to the stalled fork, and generates a paranemic joint to allow DNA translocases to catalyze fork reversal.³⁸ Multiple nascent strand degradation steps occur before and after fork reversal upon replisome uncoupling, whereas the replicative helicase is relocated to a ssDNA bubble ahead of the fork and is ready to resume replication processes.³⁹ Repositioning of CMG occurs in a SMARCAL1- and RAD51-dependent manner,³⁸ suggesting that CMG displacement does not necessarily precede fork reversal, which is consistent with our data showing that fork reversal occurs in the PBM mutant-expressing cells (Figure 5E). Given that TIM resides at the front of CMG to restrict its uncontrolled progression, we speculate that PAR-dependent TIM degradation increases the mobility of CMG along the stalled fork and helps position it in a way that facilitates the processing of the reversed fork, including generation of optimal 3ʹ-DNA ends necessary for RAD51 loading and protection. Thus, our study elucidates a mechanism underlying TIM-dependent fork protection regulated by PARP1 catalysis; the dynamic remodeling of stalled forks governed by PAR modifications is one of the key functions of PARP1 in preserving replication fork integrity.

The fate of the replisome at stalled forks remains largely elusive. The ataxia telangiectasia and Rad3-related (ATR)-mediated replication checkpoint is known to stabilize a stalled fork and facilitate its restart to prevent fork collapse by directly regulating regression and enzymatic processing of stalled forks.⁴² Fork collapse often refers to the status of inactivation of a replication fork unable to complete DNA synthesis rather than a physical dissociation of the replication components, and the replisome stability is largely independent of the replication checkpoint in both yeast and human cells.^43–45 Nevertheless, similar to other replisome subunits, TIM and CLASPIN in the FPC undergo a progressive dissociation from the nascent DNA at HU-stalled forks, indicating that disengagement of the replisome may participate in the processes of fork remodeling and subsequent DNA-protein transactions for fork protection and restart.⁴⁵ The ATR-checkpoint kinase 1 (CHK1) pathway is also known to restrict replisome activity at stalled forks.^46,47 In a pathological condition of ATR inactivation, replisome dysfunction and fork collapse ensue, which is dependent on RNF4, a small Ub-like modifier (SUMO)-targeted Ub E3 ligase, and polo-like kinase 1 (PLK1), a known mediator of CLASPIN degradation, suggesting that protein degradation constitutes a key mechanism for regulating dynamic replisome remodeling, and excessive fork regression and cleavage may occur if this process is not properly controlled.^48,49

Defining TIM as a substrate of PARP1 signifies the emerging concept of PAR-dependent substrate PARylation and subsequent PAR-dependent posttranslational modification (e.g., ubiquitination), by promoting the enzyme-substrate interaction and recruitment of modifying enzymes.¹⁹ It is currently not known why the recognition of PAR chains by TIM PBMs is essential for TIM PARylation. Analogous to the interaction between Ub and Ub-binding domains, as exemplified by the ubiquitin-binding zinc finger motif recognizing monoubiquitinated Fanconi anemia D2 (FANCD2) and proliferating cell nuclear antigen (PCNA) proteins, additional interacting interfaces may provide both increased affinity and specificity.⁵⁰ Alternatively, since the CAT of PARP1 is on the opposite side of the TIM-binding interface and PBM2, engagement of auto-PARylated PARP1 by TIM PBM2 may trigger a local conformational change of TIM favorable for its PARylation. Several TIM PARylation sites are enriched adjacent to PBM2 and the PAB region, indicating that PAR recognition may direct specific residues for PARylation.³²

Lastly, our study sheds light on a mechanism of PARPi related to dysregulated TIM at replication forks, which may result in the proteotoxic stress that causes additional replication stress and compromises fork stability. Trapping of PARP1 to DNA upon PARP inhibition underlies the hypersensitivity of BRCA1/2-deficient tumors to PARPis. Therefore, ‘‘TIM trapping’’ may have a similar adverse effect that elevates replication stress and exacerbates the replication fork instability.^20,51 Such replication-coupled proteolytic activity has been noted in the stable topoisomerase I-DNA complex at CPT-induced lesions and non-covalent Ub-specific protease 1 (USP1) accumulation onto DNA when the USP1 self-cleavage is inhibited.⁵² A dramatic increase in the association of TIM and PARP1 upon PARPis likely results from defective TIM degradation due to the lack of TIM PARylation. It would be interesting to further determine the roles of positive PARP1 regulators, such as CARM1 and HPF1,^17,53 in controlling TIM proteolysis at both active and stalled forks. Overall, understanding the extent to which the TIM-PARP1 interaction at replication forks modulates the PARPi cytotoxicity may lay the foundation for expanding the usage of PARPis beyond BRCAness—for instance, stratifying tumors according to TIM expression, which is often upregulated during tumorigenesis.⁵⁴

Limitations of the study

Our data do not pinpoint the exact mechanism through which TIM turnover regulates replisome remodeling under replication stress. It is not clear whether the removal of TIM passively leads to the displacement of CMG to ssDNA ahead of the stalled fork or additional factors or signaling may be required for driving the remodeling. The exact position of CMG was not visualized, and how this process is coordinated with fork reversal was not characterized. Another limitation is lack of information on PTUbLs involved in TIM degradation. A few members of PTUbLs have been identified, such as RNF146/Iduna, HUWE1, TRIP12, CHFR, and DELTEX family members, which contain distinct PAR-binding motifs along with RING or HECT E3 ligase domains.^55–59 The identity of PTUbLs was not pursued in this study. How replication stress signaling controls the activity of PTUbLs will provide additional insight into the replisome dynamics regulated by proteolysis at stalled forks.

STAR☆METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Hyungjin Kim (hyungjin.kim@stonybrook.edu).

Materials availability

Reagents generated for this study will be made available with some restrictions including MTA completion. Reagents obtained from other sources should be requested from those investigators.

Data and code availability

• All data are available in the main text and supplemental information.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Cell culture and cell lines

U2OS and HEK293T cell lines were acquired from the American Tissue Culture Collection (ATCC). Parental U2OS Flp-In T-REx cells were a kind gift from Dr. Daniel Durocher (The Lunenfeld-Tanenbaum Research Institute, Canada). Cells were cultured in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, following standard culture conditions and procedures.

Generation of U2OS Flp-In T-REx TIM WT and mutant cell lines

The host U2OS cell line stably expresses the Tet-repressor (T-REx) and carries a single FRT locus (Flp-In). All TIM constructs used to generate the Flp-In TIM cell lines (WT, EQ/EQ/TD, PBM1/2, or PBM ∆PI) are encoded in pcDNA5/FRT/TO plasmids and contain 7 silent mutations to prevent siRNA targeting. Each U2OS Flp-In T-REx TIM cell line was generated by co-transfecting one pcDNA5/FRT/TO TIM construct with the pOG44 plasmid encoding the Flp recombinase (Invitrogen). Forty-eight hours after transfection using the X-tremeGENE 9 (Roche), hygromycin selection was started and stably transfected cells recovered after 4 weeks. Doxycycline was then used at 100 ng/mL to drive the expression of the transgene, usually after knocking-down the endogenous TIM by reverse transfecting TIM siRNA oligo-1 using Lipofectamine RNAiMAX, as indicated.

Generation of U2OS PARP1 KO single-cell clones

Parental U2OS cells were plated for 60–70% confluency one day before transfection which was done using GeneJuice (Millipore Sigma) according to the manufacturer’s instructions. After 30 h of transfection, cells were trypsinized and seeded in 6-cm dishes in media containing 2 μg/mL puromycin. A transient selection was allowed for 51 h before replenishing cells with regular media. Surviving cells recovered in 10 days, when they were seeded at a single cell density in 96-well plates. Single-cell clones were expanded in 24-well then 6-well plates and collected for Western blotting to detect PARP1 expression levels. Genomic DNA from PARP1 KO single-cell clones was extracted and the sgRNA-targeted region was amplified by Polymerase Chain Reaction (PCR) before sub-cloning into pUC19 vector to sequence single alleles and confirm that PARP1 was indeed knocked-out.

METHOD DETAILS

Plasmid construction

pcDNA4 FLAG-hTIMELESS (TIM)-Myc-^6xHis and pcDNA3 FLAG-TIPIN were a gift from Dr. Aziz Sanç ar (Chapel Hill, NC) (Addgene plasmids #22887 and #22889, respectively). Using Q5 site-directed mutagenesis (NEB), we generated a pcDNA4 FLAG-hTIM-Myc vector by deleting the ^6xHis tag from the TIM-encoding plasmid. That template was also used for PCR amplification of a D906 to V1000 fragment and sub-cloning into pGEX-6P-1 using BamHI and XhoI restriction sites. Flp-In competent constructs were cloned using the pcDNA5/FRT/TO backbone which was a gift from Dr. Michael Frohman (Stony Brook University, NY). Using site-directed mutagenesis (SDM) of TIM WT in the pcDNA4 construct, we first introduced 7 silent mutations to prevent siRNA targeting. The siRNA-resistant construct was then sub-cloned into pcDNA5/FRT/TO using KpnI and NotI restriction sites. TIM expressed from this vector has one FLAG tag at its N terminus. Further rounds of SDM with the pcDNA5/FRT/TO-TIM plasmid was performed to generate the EQ/EQ/TD or PBM1/2 point mutations. Deletion of the DBM and Anchor domains in the MCM plug-in from the PBM1/2 construct to generate PBM ∆PI was achieved using Q5 site-directed mutagenesis (NEB). pCMV 3xFLAG-PARP1 was a gift from Dr. Thomas Muir (Addgene plasmid #111575) and was also PCR-amplified to be sub-cloned into a pcDNA3 HA-PARP1 construct with one HA tag in its N terminus. pCl-His-hUbiquitin was a gift from Dr. Astar Winoto (Addgene plasmid #31815). pcDNA3.1 HA-HPF1 was custom-cloned from NM_017867.3 (GenScript) with one HA tag in its N terminus. PCR primers containing restriction sites were used to amplify cDNAs for sub-cloning and PCR primers with mutations or deletions were used for SDM. After amplification, the PCR products were cleaned-up using PCR purification (Qiagen), followed by restriction enzyme digestion of both insert and vector, purification by gel extraction (Qiagen), and ligation. The ligated product was transformed into DH5α competent cells, and individual colonies were inoculated in Luria-Bertani (LB) media for DNA extraction using DNA miniprep or midiprep kits (Promega). All mutations were confirmed by Sanger DNA sequencing (Stony Brook Genomic Facility). PCR primers information can be found in the key resources table.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Biotin (mouse)	Jackson ImmunoResearch	Cat#200-002-211; RRID:AB_2339006
Biotin (rabbit)	Bethyl Laboratories	Cat#A150-109A; RRID:AB_67327
BRCA2 (Ab-1)	Millipore Sigma	Cat#OP-95; RRID:AB_2067762
BrdU (CldU) (BUI/75 ICR1)	Abcam	Cat#ab6326; RRID:AB_305426
BrdU (IdU) (B44)	BD	Cat#347580; RRID:AB_400326
pCHK1 S345	Cell Signaling Technology	Cat#2341; RRID:AB_330023
Cyclin A (B-8)	Santa Cruz	Cat#sc-271682; RRID:AB_10709300
FLAG M2	Millipore Sigma	Cat#F1804; RRID:AB_262044
HA (6E2)	Cell Signaling Technology	Cat#2367; RRID:AB_10691311
ɣH2AX (Phospho-Histone H2A.X Ser139)	Cell Signaling Technology	Cat#2577; RRID:AB_2118010
His-tag	Cell Signaling Technology	Cat#2365; RRID:AB_2115720
Histone H3	Abcam	Cat#ab1791; RRID:AB_302613
HPF1	Millipore Sigma	Cat#HPA043467; RRID:AB_10793949
HSC70 (B6)	Santa Cruz	Cat#sc-7298; RRID:AB_627761
pKAP1 S824	Bethyl Laboratories	Cat#A300-767A; RRID:AB_669740
MCM6 (H-8)	Santa Cruz	Cat#sc-393618; RRID:AB_2885187
Myc (9E10)	Santa Cruz	Cat#sc-40; RRID:AB_627268
ORC-2	BD Biosciences	Cat#551178; RRID:AB_394085
Poly ADP-ribose (10HA)	R&D Systems/Trevigen	Cat#4336-BPC-100; RRID:AB_2721257
PARP1 (46D11)	Cell Signaling Technology	Cat#9532; RRID:AB_659884
PARP1	Bethyl Laboratories	Cat#A301-376A; RRID:AB_937941
PARP1 (F-2)	Santa Cruz	Cat#sc-8007; RRID:AB_628105
RAD51 (D4B10)	Cell Signaling Technology	Cat#8875; RRID:AB_2721109
RPA32	Millipore Sigma	Cat#MABE285; RRID:AB_11205561
pRPA32 S4/S8	Bethyl Laboratories	Cat#A300-245A; RRID:AB_2779097
TIMELESS	Bethyl Laboratories	Cat#A300-960A; RRID:AB_805856
TIMELESS	Bethyl Laboratories	Cat#A300-961A; RRID:AB_2779580
TIPIN	Bethyl Laboratories	Cat#A301-474A; RRID:AB_999573
α-Tubulin	Santa Cruz	Cat#sc-32293; RRID:AB_628412
mouse IgG2b isotypic control	Thermo Fisher Scientific	Cat#02-6300; RRID:AB_2532949
rabbit IgG isotypic control	Millipore Sigma	Cat#12-370; RRID:AB_145841
goat anti-mouse Alexa Fluor 488	Thermo Fisher Scientific	Cat#A-11001; RRID:AB_2534069
goat anti-rat Alexa Fluor 594	Thermo Fisher Scientific	Cat#A-11007; RRID:AB_10561522
anti-Mouse IgG HRP	Cell Signaling Technology	Cat#7076; RRID:AB_330924
anti-Rabbit IgG HRP	Cell Signaling Technology	Cat#7074; RRID:AB_2099233
Bacterial and virus strains
NEB 5-alpha Competent E. coli (High Efficiency)	NEB	Cat#C2987
NEB Stable Competent E. coli (High Efficiency)	NEB	Cat#C3040
BL21 (DE3) Competent E. coli	NEB	Cat#C2527H
Chemicals, peptides, and recombinant proteins
Olaparib (AZD2281)	Selleckchem	Cat#S1060
Talazoparib	MedChemExpress	Cat#HY-16106
Veliparib	MedChemExpress	Cat#HY-10129
Hydroxyurea (HU)	Millipore Sigma	Cat#H8627
5-iodo-2ʹ-deoxyuridine (IdU)	Millipore Sigma	Cat#I7125
5-chloro-2ʹ-deoxyuridine (CldU)	Millipore Sigma	Cat#C6891
5-ethynyl-2ʹ-deoxyuridine (EdU)	Thermo Fisher Scientific	Cat#A10044
5-bromo-2ʹ-deoxyuridine (BrdU)	Millipore Sigma	Cat#B5002
Thymidine	Millipore Sigma	Cat#T1895
β-nicotinamide adenine dinucleotide hydrate (NAD+)	Millipore Sigma	Cat#N7004
Biotin-NAD+	R&D Systems	Cat#6573/131U
Biotinylated recombinant poly(ADP-Ribose) chains	R&D Systems/Trevigen	Cat#4336-100-02
RNAiMAX transfection reagent	Thermo Fisher Scientific	Cat#13778150
GeneJuice transfection reagent	Millipore Sigma	Cat#70967
X-tremeGENE 9 DNA transfection reagent	Millipore Sigma	Cat#XTG9-RO
cOmplete, EDTA-free Protease Inhibitor Cocktail	Millipore Sigma	Cat#11873580001
Halt Phosphatase Inhibitor Cocktail	Thermo Fisher Scientific	Cat#78420
MG132	Millipore Sigma	Cat#C2211
Camptothecin (CPT)	Millipore Sigma	Cat#C9911
Doxycycline hyclate (Dox)	Millipore Sigma	Cat#D9891
PDD 00017273 (PARGi)	Tocris Bioscience	Cat#PDD 00017273
Puromycin	Millipore Sigma	Cat#P8833
Bortezomib	Millipore Sigma	Cat#5043140001
Biotin Azide	Thermo Fisher Scientific	Cat#B10184
Alexa Fluor 488 Azide	Thermo Fisher Scientific	Cat#A10266
Streptavidin, HRP conjugate	Thermo Fisher Scientific	Cat#S911
Paraformaldehyde (PFA) 16%	EMS	Cat#A15710
Bovine Serum Albumin (BSA)	Millipore Sigma	Cat#A9647
Triton X-100	Millipore Sigma	Cat#T8787
Vectashield DAPI medium	VectorLabs	Cat#H-1200
Hexadimethrine bromide (polybrene)	Millipore Sigma	Cat#H9268
Hen egg white lysozyme	Millipore Sigma	Cat#L4919
Isopropyl β-d-1-thiogalactopyranoside (IPTG)	Millipore Sigma	Cat#I6758
L-glutathione, reduced (GSH)	Millipore Sigma	Cat#G4251
7-amino-actinomycin D (7-AAD)	Thermo Fisher Scientific	Cat#00-6993-50
Red Ponceau S	Boston BioProducts	Cat#ST-180
FLAG M2 affinity gel	Millipore Sigma	Cat#A2220
Anti-c-Myc agarose conjugate	Millipore Sigma	Cat#A7470
Ni-NTA agarose	Qiagen	Cat#30210
Glutathione agarose	Thermo Fisher Scientific	Cat16100
Protein G Sepharose	GE HealthCare	Cat#17-0618-01
TAK-243	MedChemExpress	Cat#HY-100487
Recombinant hPARP1 enzyme	R&D Systems/Trevigen	Cat#4668-100-01
TIM PBM1 wt peptide: GNLVVQNKQKKRRKKKKKVLDQAIVS	Genscript	N/A
TIM PBM1 mut peptide: GNLVVQNKQKKAAKKKAAVLDQAIVS	Genscript	N/A
Recombinant GST-PBM2 wt	This study	N/A
Recombinant GST-PBM2 KAKA	This study	N/A
Recombinant human Histone H3.1	NEB	Cat#M2503S
Critical commercial assays
Click-iT Cell Reaction Buffer Assay kit	Thermo Fisher Scientific	Cat#C10269
Duolink In Situ Detection Reagents Red	Millipore Sigma	Cat#DUO92008
Duolink In Situ PLA Probe Anti-Rabbit PLUS	Millipore Sigma	Cat#DUO92002
Duolink In Situ PLA Probe Anti-Mouse MINUS	Millipore Sigma	Cat#DUO92004
Duolink In Situ Wash Buffers	Millipore Sigma	Cat#DUO82049
TnT Quick Coupled Transcription/Translation System	Promega	Cat#L1170
FiberPrep DNA Extraction kit	Genomic Vision	Cat#EXT-001
Q5 Site-Directed Mutagenesis Kit	NEB	Cat#E0554
CellTiter-Glo luminescent cell viability assay	Promega	Cat#G7571
4-15% Mini-PROTEAN TGX Stain-free precast gels	Bio-Rad	Cat#4568083
Experimental models: Cell lines
U2OS	ATCC	CVCL_0042
HEK293T	ATCC	CVCL_0063
U2OS PARP1 KO	this study	N/A
U2OS Flp-In T-Rex	Dr. D. Durocher	N/A
U2OS Flp-In T-Rex TIM WT	this study	N/A
U2OS Flp-In T-Rex TIM EQ/EQ/TD	this study	N/A
U2OS Flp-In T-Rex TIM PBM	this study	N/A
U2OS Flp-In T-Rex TIM PBM ∆PI	this study	N/A
Oligonucleotides
PARP1 activating 16-mer: GGAATTCCGGAATTCC	Integrated DNA Technologies	N/A
siRNA BRCA2: ttGAAGAATGCAGGTTTAATA	Qiagen	N/A
siRNA HPF1-1: ctCCAGTGACCTTCGAAAAGA	Qiagen	N/A
siRNA HPF1-2: ttGGTTGTTCCAGTAGATAAA	Qiagen	N/A
siRNA SMARCAL1: caGCTTTGACCTTCTTAGCAA	Qiagen	N/A
siRNA TIM-1: ggGTAGCTTAGTCCTTTCAAA	Qiagen	N/A
siRNA TIM-2: aaGAGCTAAGAAGCCTAGGGG	Qiagen	N/A
Allstars negative control: caGGGTATCGACGATTACAAA\	Qiagen	N/A
Mutagenesis primers for TIM and PARP1, see Table S1.	Integrated DNA Technologies	N/A
Recombinant DNA
pcDNA4 FLAG-hTIM-Myc-6xHis	Unsal-Kacmaz K et al.60	Addgene_22887
pcDNA4 FLAG-hTIM-Myc	This study	N/A
pcDNA5/FRT/TO	Dr. M. Frohman	N/A
pcDNA5/FRT/TO FLAG-hTIM	This study	N/A
pOG44	Invitrogen	Cat#V600520
pcDNA3 FLAG-TIPIN	Unsal-Kacmaz K et al.61	Addgene_22889
pcDNA3.1 HA-HPF1	GenScript	N/A
px459 sgRNA PARP1 GenCRISPR	GenScript	N/A
pCMV 3xFLAG-PARP1	Liszczak et al.62	Addgene_111575
pcDNA3 HA-PARP1	This study	N/A
pGEX-6P-1 TIM D906-V1000	This study	N/A
Software and algorithms
Fiji	NIH	https://imagej.net/software/fiji/; RRID: SCR_0022285
NIS-Elements, BR	Nikon	SCR_002776
GraphPad Prism	GraphPad Software	https://www.graphpad.com/; RRID: SCR_002798
Chimera	UCSF	https://www.cgl.ucsf.edu/chimera/; RRID: SCR_004097
Attune NxT Software	Thermo Fisher Scientific	RRID:SCR_019590

DNA and siRNA transfection

Unless otherwise stated, plasmid transfection was performed using GeneJuice (Millipore) according to the manufacturer’s protocols. siRNA duplexes were transfected at 25 nM using Lipofectamine RNAiMAX (Invitrogen). siRNA sequence information can be found in the key resources table.

Drugs treatment

For all treatments and labeling that affected DNA replication, cell culture media was pre-equilibrated in the incubator the night before. Hydroxyurea (HU) was dissolved in water at 500 mM stock and stored at −20°C. HU treatment was at 2 mM, 4h for DNA fiber assay fork restart, 3 mM, 5h for RAD51-EdU or MCM6-EdU PLA, and 4 mM for BrdU staining or DNA fiber assay fork resection assays. Camptothecin (CPT) was dissolved in DMSO at 1 mM stock and stored at −20°C. CPT was diluted into cell culture media to 1 μM for the indicated times. For detecting PARylation, cells were treated with 10 μM PARGi (PDD 00017273) for the indicated times. PARGi stock was at 10 mM in DMSO and stored at −20°C MG132 and bortezomib (BTZ) were dissolved at 10 mM in DMSO and stored at −20°C. Cells were treated at 10 μM MG132 or 1 μM BTZ in cell culture media for the indicated times. PARP inhibitors (olaparib, veliparib, talazoparib) were dissolved at 10 mM in DMSO and stored at −80°C. Cells were treated at 10 μM PARP inhibitors in cell culture media for the indicated times. Drugs information can be found in the key resources table.

Cell lysis, fractionations and western blotting

U2OS cells or U2OS Flp-In cells were trypsinized and HEK293T cells were either trypsinized or scraped in PBS and the cell pellet washed in PBS. For whole cell lysates, cells were lysed in NETN300 buffer (50 mM Tris-HCl [pH 7.5], 300 mM NaCl, 0.2 mM EDTA, 1% NP40) complemented with protease inhibitor cocktail (Millipore Sigma) and phosphatase inhibitor cocktail (Thermo Fisher) for 40 min on ice. Lysates were then clarified by centrifugation at 14,000 rpm for 10 min at 4°C. For chromatin-enriched fractionation, cells were lysed in cytoskeleton (CSK) buffer (10 mM Tris-HCl [pH 6.8], 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 1 mM EDTA, 1mM EGTA, 0.1% Triton X-100) complemented with protease (Roche) and phosphatase inhibitor cocktails (Thermo Fisher) for 5 min on ice. Soluble proteins were separated by centrifugation at 1,500 g for 5 min at 4°C and saved as (S). Pellets were resuspended in PBS and an equal volume of 2X boiling buffer (100 mM Tris-HCl [pH 6.8], 4% SDS, 1.7 M β-mercaptoethanol (BME)) and incubated at 95°C for 10 to 15 min with occasional vortexing, and labeled as (P). For sub-cellular fractionations, cells were lysed using S lysis buffer (10 mM HEPES [pH 7.4], 10 mM KCl, 0.05% NP-40) for 10 min on ice and lysates clarified by centrifugation at 14,000 rpm for 10 min (supernatants containing cytosolic proteins were saved as the S fraction). The pellets (nuclei) were first washed with S lysis buffer and lysed in P1 low salt buffer (10 mM Tris-HCl [pH 7.5], 0.2 mM MgCl₂, 1% Triton X-100) and incubated on ice for 15 min and centrifuged (supernatants containing nuclear and nonchromatin-associated proteins were saved as the P1 fraction). The new pellets (chromatin) were first washed with PBS, resuspended in 0.2 N HCl and incubated on ice for 20 min. After centrifugation, supernatants were transferred into new tubes containing equal volumes of Tris-HCl pH 8.0 to neutralize the acid (these acid-soluble chromatin-associated proteins were saved as the P2 fraction). Protein concentration of supernatants was measured using the Bradford assay and 15–50 μg of protein was aliquoted and diluted 1:1 with 2X Laemmli sample buffer complemented with 5% BME. After boiling aliquots at 95°C for 5 min and letting them cool down to RT, samples were loaded onto SDS-PAGE gels and transferred to PVDF membranes (Millipore Sigma). Membranes were incubated with the indicated primary and HRP-conjugated secondary antibodies. HRP signal was detected by enhanced chemiluminescence (ECL) of western blotting substrates (Thermo Fisher) using the iBright imager (Thermo Fisher, #CL1000). Antibody information can be found in the key resources table.

Protein expression and purification

GST and GST-TIM PBM2 WT or K949A/K950A were expressed using BL21 (DE3) cells by incubating at 37°C to an absorbance of 0.6–0.8 at 600 nm, then inducing with 0.5 mM IPTG (Millipore Sigma) at 30°C for 6 h. After centrifugation, cells were resuspended in PBS with 1 mg/mL hen egg white lysozyme and 0.5 mM PMSF, rocked at 4°C for 40 min, and stored at −80°C. After thawing on ice, the mixture was sonicated on a QSonica Q500 Digital Sonicator at 50% amplitude with three cycles of 10 s pulses followed by 20 s recovery time. Triton X-100 was added to a final concentration of 0.5–1%, and the lysate was rocked for 30 min at 4°C. Lysates were centrifuged at 14,000 rpm for 15 min at 4°C, then filtered through a 0.45 μm polyethersulfone filter. Cleared lysates were frozen at −80°C in aliquots. For batch purification, 1 mL filtered lysates were thawed on ice and added to 50 to 100 μL of equilibrated Glutathione Agarose affinity beads (Thermo Fisher Scientific), then incubated at 4°C while rocking for 2 to 4 h. Beads were washed three times with wash buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 0.1 mM EDTA). The recombinant GST or GST-polypeptides were eluted with 50 mM Tris-HCl [pH 8.0] containing 10 mM reduced GSH while rocking at 4°C for 30 min. Elution was repeated once and both eluates were combined, aliquoted and stored at −80°C. The purified fractions were visualized by Coomassie staining against BSA to confirm purity and yield.

PAR overlay assay

Purified recombinant GST, GST-TIM PBM2 WT or K949A/K950A, and BSA were first diluted in TBS to achieve 6 ng/μL, corresponding to 1.2 μg in 200 μL. Peptides of 26 amino acids representing TIM PBM1 WT or mutant were custom-synthesized (GenScript) and resuspended at 600 μM stock in formic acid or water, respectively. Peptides were serial-diluted down to 5 and 2 nmol in 200 μL volume. Each 200 μL-sample was blotted onto a nitrocellulose membrane (BioRad, #1620115) by slot-blotting using the Bio-Dot SF microfiltration system (BioRad, #1706542), making sure all the wells contain the same volume of sample/buffer. The membrane was then stained in Red Ponceau S to control for equal loadings. After a few water and TBS washes to destain, membranes were incubated overnight with biotinylated recombinant poly(ADP-ribose) chains (R&D Systems/Trevigen) at 1:3,000 in TBS containing 0.1% Tween 20 (TBS-T). Non-specific binding was first removed by three washes in TBS-T containing 1 M NaCl. Membranes were then blocked with 5% milk in TBS-T for 1 h at RT, followed by incubation for 1 h with a primary rabbit anti-biotin antibody diluted at 1:1,000 in 1% milk in TBS-T. Membranes were washes three times in TBS-T and incubated with a secondary anti-rabbit HRP-conjugated antibody diluted at 1:4,000 in 1% milk in TBS-T for 45 min at RT. After three final washes with TBS-T, membranes were developed using enhanced chemiluminescence substrates and imaged using the iBright imager (Thermo Fisher).

In vitro PARP1 PARylation assay

PARP1 autoPARylation assay was carried out in 10 μL reaction volume, by first equilibrating 10 U of recombinant PARP1 (R&D Systems, #4668–100-01) with 2 μM 16-mer dsDNA in 1X reaction buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 4 mM MgCl₂) complemented with EDTA-free protease inhibitor cocktail (Thermo Fisher) for 5 min on ice. The reaction was started by swiftly adding NAD⁺ (or water for negative control) to the desired concentration as indicated, and immediately incubating the samples at 30°C for 20 min. Reactions were stopped by adding the same volume of 2X Laemmli sample buffer complemented with 5% BME, before western blotting.

For in vitro PARylation of a substrate (TIM) and its subsequent detection by immunoprecipitation, we first produced recombinant full-length FLAG-TIM using the In Vitro Transcription & Translation (IVTT) kit (Promega) following the manufacturer’s instructions. Briefly, TNT quick master mix, methionine and T7 TNT PCR enhancer were added to 250 ng expression plasmid DNA bearing a T7 promoter and incubated at 30°C for 90 min. The PARylation reaction was then scaled up to 60 μL reaction volume (from the auto-PARylation conditions), except for recombinant PARP1 which was used at a final concentration 10 times lower and for the addition of PARGi at 10 μM. FLAG-TIM PARylation reaction hence contained 12 μL of IVTT-produced recombinant FLAG-TIM added to the PARP1 reaction which was started by the addition of 25 μM biotinylated NAD⁺. Samples were incubated at 30°C for 20 min before being diluted into 5 volumes of NETN150 buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 0.2 mM EDTA, 1% NP40) complemented with protease and phosphatase inhibitor cocktails (Thermo Fisher) and PARGi at 10 μM. Samples were then incubated with pre-equilibrated anti-FLAG M2 agarose beads (Sigma-Aldrich) for immunoprecipitation as described below. After western-blotting, PARylated products were revealed by incubating the membranes with HRP-coupled Streptavidin (Thermo Fisher, #S911, 1:3,000) diluted in 1% milk.

For in vitro PARylation of Histone H3.1, the reactions were scaled up to 120 μL reaction volume, including recombinant PARP1 at a final concentration 10 times lower than for auto-PARylation conditions, 1 μM recombinant H3.1 (NEB) and PARGi at 10 μM. The auto-PARylation of PARP1 and trans-PARylation of H3.1 reaction was started by adding 50 μM of NAD⁺ for 20 min at 30°C. Then, 10 μL of IVTT-produced FLAG-TIM were added to 100 μL of PARylation products diluted in 7 volumes of NETN150 buffer complemented with protease inhibitor cocktail and PARGi at 10 μM. Samples were then incubated overnight with pre-equilibrated anti-FLAG M2 agarose beads (Sigma-Aldrich) for immunoprecipitation and western-blot analysis.

Immunoprecipitation (IP)

HEK293T transfected and treated as indicated were harvested 36 to 48 h post-transfection. Cells were lysed in NETN150 buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 0.2 mM EDTA, 1% NP40) complemented with protease and phosphatase inhibitor cocktails (Thermo Fisher) for 40 min on ice, and lysates clarified by centrifugation at 14,000 rpm for 10 min at 4°C. Clarified lysates were then incubated with pre-equilibrated anti-FLAG M2 agarose beads (Sigma-Aldrich) or anti-c-Myc agarose beads (Sigma-Aldrich) for 4 h at 4°C while rocking. Beads were then washed 3 to 4 times in NETN150, and proteins eluted by boiling for 5 min in 2X Laemmli sample buffer complemented with 5% BME, before western blotting. For U2OS cells transfected and treated as indicated, chromatin-enriched fractions were obtained by first gently lysing cells in cytoskeleton (CSK) buffer (10 mM Tris-HCl [pH 6.8], 100 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 1 mM EDTA, 1mM EGTA, 0.1% Triton X-100) complemented with protease and phosphatase inhibitor cocktails for 5 min on ice. Soluble proteins were separated by centrifugation at 1,500 g for 5 min at 4°C and discarded. Pellets were lysed in complemented NETN300 buffer for 40 min on ice, and clarified by centrifugation at 14,000 rpm for 10 min at 4°C. Supernatants were diluted vol:vol with NETN0 buffer (50 mM Tris-HCl [pH 7.5], 0.2 mM EDTA, 1% NP40) to bring sodium chloride concentration down to 150 mM. This mixture was incubated overnight with pre-equilibrated beads, as indicated. The following morning, beads were washed 4 times in NETN150 buffer and eluted as described before.

Endogenous IP

HEK293T cells were either synchronized in S phase by double thymidine block and released into fresh media for 2 h as described before, or reverse transfected with the indicated constructs for 42 h. Cells were scraped in PBS, washed and lysed in NETN150 buffer complemented with protease and phosphatase inhibitor cocktails (Thermo Fisher) for 40 min on ice, and lysates clarified by centrifugation at 14,000 rpm for 10 min at 4°C. Clarified lysates were then incubated with immunoprecipitating antibodies for 4 h at 4°C while rocking as follows: 5 μg of rabbit IgG isotypic control (Millipore, #12–370) or rabbit anti-TIM antibody (Bethyl Laboratories, #A300–960A); 2 μg mouse anti-MCM6 antibody (Santa Cruz, #sc-393618). Protein A-conjugated agarose or protein G Sepharose beads were equilibrated in NETN150 buffer for IP with rabbit or mouse antibodies, respectively. Immunocomplexes were added onto pre-equilibrated beads for overnight incubation at 4°C while rocking. The following morning, beads were washed 3 to 4 times with NETN150 buffer and eluted by boiling for 5 min in 2X Laemmli sample buffer complemented with 5% BME, before western blotting.

In vivo TIM PARylation assay

HEK293T cells transfected and treated as indicated were harvested 40 to 45 h post-transfection by scraping them in PBS on ice. After a PBS wash, cell pellets were loosened up in a small volume of TBS (50 mM Tris-HCl [pH 7.5], 150 mM NaCl), then lysed in 0.5% SDS RIPA buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.5% SDS, and 0.5% sodium deoxycholate) complemented with protease and phosphatase inhibitor cocktails (Thermo Fisher) for 20 min on ice, followed by incubation at 95°C for 15 min with occasional vortexing. Samples were then diluted 1:1 in TBS complemented with protease and phosphatase inhibitor cocktails before centrifugation at 14,000 rpm for 10 min at 4°C. Cleared lysates were then incubated with pre-equilibrated anti-FLAG M2 affinity beads for 4 h at 4°C with gentle rocking. Beads were subsequently washed four times with a 1:1 solution of 0.5% SDS RIPA and TBS, and proteins were eluted by boiling for 5 min in 2X Laemmli sample buffer complemented with 5% BME, before western blotting.

Ubiquitination assay

HEK293T cells were transfected with the indicated constructs 40 h before drug treatments, consisting of 1 μM CPT alone for 1 h before adding concentrated MG132 (10 μM final) for an additional 4 h period. In vivo ubiquitin assay was performed as previously described.⁶³ Cells were then harvested by scraping in PBS, washed and cell pellets were loosened by adding a small volume of PBS before lysis in SDS 1% in PBS and snap freeze in liquid nitrogen. Samples were then transferred at 95°C for 15 min with occasional vortexing, and diluted with 5 volumes of PBS before centrifugation at 14,000 rpm for 10 min at 4°C. Supernatants were carefully saved into new tubes and diluted 1:1 with PBS to bring SDS down to 0.1%. Imidazole was added to the samples at 10 mM final before incubation with pre-equilibrated Ni-NTA agarose beads (Qiagen) for 3 h at 4°C while rocking. Beads were subsequently washed five times with wash buffer (0.1% SDS, 10 mM Imidazole in PBS) and eluted by boiling for 5 min in 2X Laemmli sample buffer complemented with 5% BME, before western blotting.

Cell synchronization and flow cytometry

U2OS cells were synchronized at the G1/S boundary using a double thymidine block protocol. Briefly, cells were seeded at low density and treated the next day with 2 mM thymidine (Millipore Sigma) in cell culture media for 17 h (first block). Cells were washed three times with warm PBS and replenished with fresh thymidine-free media for 9 h. Finally, cells were then treated with 2 mM thymidine for an additional 17 h (second block) before being released into fresh media as indicated. For HEK293T cells, first block was 18 h, then release for 9 h, and second block for 15 h. Cell cycle analysis by flow cytometry was performed on cells harvested by trypsinization at different time points during the synchronization and release, as indicated. After a PBS wash, cells were washed in 1% BSA in PBS, vortexed briefly and centrifuged at 500g for 5 min at RT. Cells were resuspended and homogenized in a small volume of 1X PBS, before being fixed with 70% ethanol, which was added dropwise while vortexing at a low speed. Fixed cells were kept at 4°C overnight before being labeled. After fixation, ethanol was removed via centrifugation at 500g for 5 min at 4°C and cells were washed twice with PBS. Finally, cells were resuspended in PBS containing 200 μg/mL PureLink RNase and eBioscience 7-AAD viability staining solution (Thermo Fisher) and incubated for 30 min at 37°C. All samples were analyzed with the Attune NxT acoustic focusing cytometry (Thermo Fisher) and analyzed with the Attune NxT software V2.7 (Thermo Fisher).

BrdU staining for ssDNA detection

U2OS Flp-In cells were reverse transfected with siRNA oligos in 6-cm dishes, induced with doxycycline (100 ng/mL) and seeded in 12-well plates on glass coverslips 48 h after transfection. The following day, cells were pulse-labelled with 10 μM BrdU for 25 min in pre-equilibrated media, washed once with warm PBS and treated or not with 4 mM HU for 2 h. After a wash with cold PBS on ice, cells were permeabilized with PBS/0.5% Triton X-100 for 5 min on ice, washed again with PBS three times, and fixed with 4% paraformaldehyde for 10 min at RT. After three PBS washes, cells were blocked with 5% BSA for 1 h at RT, then incubated with mouse anti-BrdU antibody (BD, #347580, 1:10) in 1% BSA for 1 h at RT. After three PBS washes, cells were incubated with goat anti-mouse Alexa Fluor 488 IgG at 1:1,000 in 1% BSA for 45 min at RT. After three PBS washes, coverslips were mounted using DAPI-containing mounting medium (Vector Lab) and analyzed with the Eclipse Ts2R-FL inverted Nikon fluorescence microscope equipped with the Nikon DSQi2 digital camera. Cells with more than 10 foci were counted as positive. Percentages of positive cells were quantified and analyzed by Prism (GraphPad).

Proximity ligation assay between proteins, and on nascent DNA

The Proximity Ligation Assay (PLA) was performed to analyze the proximity between two proteins, using the Duolink In Situ red system (Sigma), following the manufacturer’s instructions. Briefly, U2OS cells or U2OS Flp-In cells transfected with the indicated siRNA oligos and induced or not with 100 ng/mL doxycycline were seeded onto glass coverslips one day before EdU labeling, when sufficient cell culture media was equilibrated overnight in a humidified incubator at 37^◦C under 5% CO₂. EdU, dissolved in DMSO, was diluted into the pre-equilibrated media at either 125 μM with 12 min labeling time for PLA on nascent DNA, or diluted at 10 μM with 30 min incubation for labeling cells in S phase. Following nascent DNA labeling, or in the case of PLA between two proteins, cells were washed with cold PBS on ice, and fixed with 4% paraformaldehyde for 10 min at RT. After three washes with PBS, cells were stored in fresh PBS, in a sealed plate protected from light at 4°C. To recover cells from storage, coverslips were first washed with cold PBS, then permeabilized with 0.3% Triton X-100 in PBS, for 3 min on ice or, in the case of RAD51-EdU PLA, with 0.25% Triton X-100 in PBS for 30 min at RT, and again washed three times with PBS. In the case of EdU labeling, cells were quickly blocked with BSA 1% in PBS for 10 min at RT while the Click-iT reaction cocktail was prepared, in order to conjugate the EdU alkyne with either Biotin azide (for nascent DNA) or Alexa Fluor 488 azide (for S phase cells), following the manufacturer’s instructions (Thermo Fisher Scientific, #C10269). Briefly, the reaction cocktail is freshly prepared before use, and is composed of 1X Click-iT reaction buffer, 2 mM CuSO₄, 10 μM Biotin-Azide/5 mM Alexa Fluor 488 azide and 10 mM sodium ascorbate. Coverslips were incubated with this cocktail for 30 min to 1 h at RT, protected from light and in a humidified chamber. Following azide conjugation, coverslips were washed once with PBS, before proceeding to the PLA assay. Coverslips were first blocked with a drop of Blocking Solution, for 1 h. From this step onward, all incubations were performed at 37°C, in a humidified chamber, and protected from light. Following blocking, coverslips were incubated with primary antibodies diluted in the Antibody Diluent, using the following dilution ratios: mouse anti-biotin (Jackson ImmunoResearch, #200–002-211, 1:2,000), rabbit anti-biotin (Bethyl Laboratories, #A150–109A, 1:3,000), rabbit anti-TIM (Bethyl Laboratories, #A300–961A, 1:500), mouse anti-PARP1 (Santa Cruz, #sc-8007, 1:250 to 1:500), mouse anti-MCM6 (Santa Cruz, #sc-393618, 1:500), rabbit anti-RAD51 (Cell Signaling Technology, #8875, 1:100). The coverslips were washed twice with Wash Buffer A at RT, before incubation with the PLUS and MINUS PLA probes for 1 h. After two washes with Wash Buffer A at RT, coverslips were incubated with the Ligation reaction for 30 min, before being washed again twice with Wash Buffer A. The Amplification reaction was then carried out on the coverslips, for 100 min, before washing the coverslips twice with Wash Buffer B. After a final wash with 1:100 Wash Buffer B, coverslips were mounted in the wet In Situ Mounting Medium with DAPI and fixed with nail polish. Coverslips were observed and imaged with the Eclipse Ts2R-FL inverted Nikon fluorescence microscope equipped with the Nikon DSQi2 digital camera. Fluorescence images were analyzed with NIS-Elements, Research BR software (Nikon), and quantification data was processed by Prism (GraphPad).

DNA fiber combing

Exponentially growing U2OS Flp-In cells were reverse transfected with the indicated siRNA and induced or not by doxycycline for at least 48 h before being pulse-labelled with 50 μM CldU for 30 min, washed three times with PBS, then pulse-labelled with 250 μM IdU for 30 min. In the case of resection studies, cells were further washed three times with PBS before replenishing media with 4 mM HU for 4 h. For fork restart assays, cells were treated with 2 mM HU for 4 h between the two thymidine analog labeling periods. All necessary media was equilibrated overnight at 5% CO₂ and 37°C. Cells were harvested by trypsinization, counted and pelleted so that 400,000 cells of each sample were kept to prepare the DNA fibers using the FiberPrep DNA extraction kit and the FiberComb Molecular Combing System (Genomic Vision, France), following the supplier’s instructions. In brief, cells were washed again with PBS before being embedded in low-melting point agarose, and cast in a plug mold. After full solidification, plugs were digested overnight with proteinase K. Over the next day, plugs were extensively washed before being shortly melted and digested with agarase overnight. The obtained DNA fibers were then combed onto silanized coverslips (Genomic Vision, France) that were subsequently baked for 2 h at 60°C. DNA was denatured for 8 min using 0.5 M NaOH in 1 M NaCl. All subsequent incubations during the staining process were done in humidified conditions at 37°C. In short, PBS-equilibrated coverslips were first dehydrated then blocked with 1% BSA for 30 min, before both primary antibodies were diluted together in 1% BSA (rat monoclonal anti-BrdU for CldU, 1:25, and mouse monoclonal anti-BrdU for IdU, 1:5) and incubated for 1 h. After washing the coverslips with PBS-Tween 0.05% (PBS-T), secondary antibodies were both diluted together in 1% BSA (Alexa Fluor 594 goat anti-rat and Alexa Fluor 488 goat anti-mouse, 1:100) and incubated for 45 min. After washing the coverslips with PBS-T and dehydrating them, they were mounted onto microscopic glass slides using ProLong Gold Antifade and cured overnight. DNA fibers were then imaged with the Eclipse Ts2R-FL inverted Nikon fluorescence microscope equipped with the Nikon DSQi2 digital camera, measured using the Fiji software, and analyzed with Prism (GraphPad).

ATP-based viability assay

U2OS Flp-In cells were reverse transfected with siRNA oligos in 6-well plates, induced with doxycycline and seeded in 96-well plates 48 h after transfection. Drug treatment started 24 h after seeding and cell viability was determined using the CellTiter-Glo luminescent cell viability assay (Promega) seven days post-transfection. Luminescence was measured using a GloMax Explorer microplate luminometer (Promega).

QUANTIFICATION AND STATISTICAL ANALYSIS

All statistical analyses were performed using Prism (GraphPad) and in all cases, n.s., p > 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001; ****, p < 0.0001. The type of statistical test used and their results for each experiment can be found in the figures and figure legends. When comparing only two groups of variables following a normal distribution, unpaired Student’s t-tests were performed with a 95% confidence interval, using two-tailed p values on a set of at least 3 replicates. When comparing more than two groups of variables, ANOVA-based statistical tests were used followed by multiple comparisons and post-hoc test using the two-stage step-up method of Benjamin, Krieger and Yekutieli to control for the False Discovery Rate (FDR) set at 5% (Q = 0.05). In these conditions, one-way ANOVA was used for normally distributed values, whereas Kruskal-Wallis was performed for non-normally distributed values. When the experimental set up involved two factors, a two-way ANOVA was performed, with the same type of post-hoc test and FDR adjustment. For DNA combing or PLA assays, dot or violin plots depict a representative result consisting of at least 150 fibers or 350 cells per condition, respectively, and representative of at least 2 independent biological replicates. For western blot quantifications and cell viability assays, graphs represent the average of a set of at least 3 independent biological replicates.

Highlights.

TIMELESS (TIM) and PARP1 interact at DNA replication forks

TIM recognizes auto-PARylated PARP1 by two conserved PAR-binding motifs (PBMs)

PBM-dependent TIM PARylation primes TIM for proteasomal degradation

Aberrant TIM accumulation causes replication stress and impairs stalled fork protection