Modular antibodies reveal DNA damage-induced mono-ADP-ribosylation as a second wave of PARP1 signaling

Summary PARP1, an established anti-cancer target that regulates many cellular pathways, including DNA repair signaling, has been intensely studied for decades as a poly(ADP-ribosyl)transferase. Although recent studies have revealed the prevalence of mono-ADP-ribosylation upon DNA damage, it was unknown whether this signal plays an active role in the cell or is just a byproduct of poly-ADP-ribosylation. By engineering SpyTag-based modular antibodies for sensitive and flexible detection of mono-ADP-ribosylation, including fluorescence-based sensors for live-cell imaging, we demonstrate that serine mono-ADP-ribosylation constitutes a second wave of PARP1 signaling shaped by the cellular HPF1/PARP1 ratio. Multilevel chromatin proteomics reveals histone mono-ADP-ribosylation readers, including RNF114, a ubiquitin ligase recruited to DNA lesions through a zinc-finger domain, modulating the DNA damage response and telomere maintenance. Our work provides a technological framework for illuminating ADP-ribosylation in a wide range of applications and biological contexts and establishes mono-ADP-ribosylation by HPF1/PARP1 as an important information carrier for cell signaling.


In brief
Longarini et al. generate high-affinity and sensitivity antibodies for mono-ADPribosylation detection in various applications and show that serine mono-ADP-ribosylation constitutes a second wave of PARP1 signaling. They reveal a role of mono-ADPr as a recruitment signal for RNF114, modulating the DNA damage response and telomere maintenance.
INTRODUCTION PARP1, a much-studied target for cancer therapy, plays key roles in the DNA damage response (DDR) by covalently transferring ADP-ribose from NAD + to a target substrate, generating ADP-ribosylation (ADPr). 1 For more than 50 years, this enzyme has been studied exclusively in the context of poly-ADPr on aspartate and glutamate. [2][3][4] The unexpected identification of histone serine ADPr (Ser-ADPr) marks 5 has quickly led to the establishment of serine as the primary target residue for PARP1 upon DNA damage [6][7][8][9] and to the discovery of HPF1 10 (in complex with PARP1) and ARH3 as the writer and the eraser, respectively, of Ser-ADPr. 6,9,11 When bound to PARP1, HPF1 catalyzes the addition of single units of ADP-ribose to serine residues while blocking their addition to poly-ADPr chains. 12 HPF1 rapidly dissociates, permitting PARP1 to extend the initial modification to poly-ADP-ribose. 13,14 Recent reports have shown the prevalence of cellular mono-ADPr upon DNA damage 15,16 and the impact of specific histone mono-ADPr marks on chromatin structure in biochemical assays. 17  (legend continued on next page) ll OPEN ACCESS Technology abundance was interpreted as an intermediate in the formation and/or degradation of poly-ADPr, and direct cellular roles of mono-ADPr by HFP1/PARP1 remained unknown. Here, we consider the possibility that monomeric ADPr is a fully fledged histone mark, written and erased by dedicated enzymes, and recognized by specific effectors. To test this hypothesis and enable sensitive, versatile mono-ADPr detection both within and beyond PARP1 signaling, we have developed modular antibodies based on SpyTag technology.

DESIGN
Despite the clinical development of PARP inhibitors and their broad biological significance, the chemical nature of ADPr has long hampered our ability to study this PTM. Accordingly, considerable efforts have recently been invested in developing new tools for ADPr research. The Kraus laboratory has pioneered the conversion of protein domains recognizing ADPr into antibody-like reagents. 19 Moreover, sophisticated approaches have improved the chemical synthesis of ADP-ribosylated substrates, 20 and our phospho-guided enzymatic strategy has provided antigens for generating broad-and site-specific antibodies. 15 However, detection of ADPr in cellular contexts is still in its infancy compared with other PTMs. Although it is becoming clear that most PARPs and other transferases conjugate monomeric ADPr, 21 researchers still lack a widely applicable toolbox for sensitive detection of this modification in cellular contexts.
Given the difficulties in generating conventional anti-ADPr antibodies, most currently available tools are recombinant, either domain-based reagents or phage-display antibodies. 15,19,22 We aimed to exploit the recombinant nature of these tools to expand their functionality by applying SpyTag technology. 23 In this protein ligation system, the SpyTag peptide and the SpyCatcher domain covalently bond when brought together; hence, antibodies generated as SpyTagged antigen-binding fragments (Fabs) can be ligated to various domains and chemical labels via the SpyCatcher, yielding an expandable library of antibody formats. 24 The availability of the SpyCatcher reagents allows the implementation of this strategy in any biological laboratory. Here, we have developed an entire toolkit for flexible and sensitive detection of mono-ADPr ( Figure 1A) by applying affinity maturation and the SpyTag/SpyCatcher system to our previously generated antibodies 15 and show the utility of these modular antibodies in immunoblotting, immunofluorescence, immunoprecipitation, enzymelinked immunosorbent assay (ELISA), and live-cell imaging.

RESULTS
SpyTag-based modular antibodies enable sensitive and versatile detection of mono-ADPr Although detection of ADPr has been significantly advanced by recent introduction of specific tools, 15,19,22,25 important ADPr events have remained undetectable. A striking illustration is that histone mono-ADPr was undetectable in the absence of DNA damage, although the high levels detected in undamaged cells lacking the serine mono-ADP-ribose hydrolase ARH3 14,15,26 indicate that it is constantly formed. For broader investigations of mono-ADPr, we have re-engineered the recombinant antibodies generated with our HPF1/PARP1-based chemical biology approach. 15 To increase their sensitivity and versatility, we applied a modular platform 24 based on the SpyTag system 23 to couple multiple features onto Fabs (Figure 1A). Among the formats we obtained for all our antibodies, 15 site-directed labeling with three copies of horseradish peroxidase (HRP) through SpyTag coupling 24 increased the immunoblotting sensitivity dramatically compared with the conventional IgG format ( Figure S1A; Methods S1). We further improved the detection of mono-ADPr by using our serine mono-ADP-ribosylated peptides 15 for phage-display-based affinity maturation of AbD33204, whose binding affinity was $4.4 mM, 15 to obtain AbD43647, a mono-ADPr-specific SpyTag antibody with a binding affinity of 248 nM and very minor unspecific binding (Figures 1B and S1B-S1K). Even when the cellular levels of poly-ADPr were massively boosted by inhibiting the poly-ADPr eraser PARG, AbD43647 did not show any cross-reactivity toward poly-ADPr ( Figures 1C and S2A). An ELISA competition assay indicates that binding of AbD43647 to ADP-ribose requires adenine and a free 2 0 -hydroxyl group on the adenineproximal ribose ( Figure S2B). The high affinity of AbD43647 together with the HRP-conjugated format for immunoblotting rendered HPF1-dependent histone mono-ADPr detectable in WT cells in the absence of exogenous DNA damage ( Figure 1D). The signal was significantly improved when immunoblotting was preceded by immunoprecipitation with AbD43647 IgG, illustrating the synergy of multiple conjugate formats of an affinitymatured antibody to maximize mono-ADPr detection ( Figures 1D, S2C, and S2D). Upon H 2 O 2 treatment, mono-ADPr was dramatically reduced in HPF1-KO cells, compared with WT cells (Figures 1E, S2E, and S2F), and restored by HPF1-WT expression in HPF1-KO cells, but not HPF1-E284A ( Figure S2G), a catalytically inactive mutant that still interacts with PARP1. 12 Recently, ARH3 has been implicated in ALT-mediated telomere maintenance, 14 suggesting a role for mono-ADPr at telomeres. To test this, we used the TRF1-FokI system, which specifically cleaves telomeric DNA to induce telomere DDR. 27 Our antibodies reveal the formation of mono-ADPr foci at telomeres with WT-TRF1-FokI, but not with the catalytically dead (D450A) mutant. Compared with WT cells, this mono-ADPr was increased in ARH3-KO cells and abrogated in HPF1-KO cells ( Figures 1F and S2H).
To demonstrate the broad value of AbD43647, we have gone beyond PARP1 signaling and protein ADPr. In untreated cells, AbD43647 gives a clear mitochondrial signal, 28  Technology HPF1 or ARH3 and undetectable by AbD33205 (Figures 1G and  S2I). Moreover, we tested our antibodies on DNA ADPr. 29 Although AbD33205 detects ADP-ribosylated DNA weakly, AbD43647 gives a strong signal, especially in the HRP-coupled format ( Figures 1H and S2J). Thus, the range of applications for our new high-affinity antibody AbD43647 is much wider than for the ''anti-protein mono-ADPr'' AbD33205. 15 Fluorescence-based sensors reveal DNA damageinduced serine mono-ADPr as a second wave of PARP1 signaling With these tools for sensitive detection of endogenous mono-ADPr in various applications (Figure 1), we set out to investigate mono-ADPr dynamics in the DDR. To visualize mono-ADPr in living cells, we took advantage of the multiple formats of our SpyTagged recombinant antibodies ( Figure 1A) and combined laser irradiation with bead loading, a simple method for introducing proteins into the cells. 30 Reasoning that the recruitment kinetics of mono-ADPr-specific, fluorescence-based probes should reflect the dynamics of DNA damage-induced mono-ADPr in living cells, we conjugated the Fab version of AbD33205 to a fluorescent dye ( Figure S3A). A key feature of the Fab format is its small size, allowing the sensor to enter the nucleus after bead loading into the cytoplasm (Figures 2A,  S3B, and S3C). To test the specificity of our Fab-based probe, we treated the cells with olaparib and observed that PARP1 inhibition abolished the accumulation of the mono-ADPr sensor at DNA damage sites ( Figure S3D). Recruitment of the probe was also abolished in HPF1-KO cells ( Figure 2B), corroborating the dependence of serine mono-ADPr on HPF1 in cells ( Figure 1E). PARP1 signaling is one of the earliest pathways activated during the DDR, as illustrated by the rapid formation of poly-ADPr upon DNA damage. 31,32 Surprisingly, we observed a more gradual mono-ADPr response ( Figures 2B and S3D). To directly compare the dynamics of these two forms of ADPr in living cells, we bead loaded the anti-mono-ADPr fluorescent Fab and expressed GFP-WWE, an established probe for poly-ADPr. 25,31,33 Strikingly, live-cell microscopy revealed that, in contrast to the ultrafast, transient formation of poly-ADP-ribose, mono-ADPr levels increase more gradually within the first few minutes after damage before reaching a plateau and then slowly declining (Figure 2C). To confirm these observations, we used the macrodomain of MacroD2, which is specific for mono-ADPr and unable to hydrolyze Ser-ADPr, 11,34 as a fluorescence-based sensor for mono-ADPr in live-cell imaging. In ARH3-KO cells, which have high and long-lasting levels of mono-ADPr upon DNA damage, 14,15,26 the recruitment of this macrodomain persists longer compared with WT cells, reaching a plateau 6 min after damage ( Figure 2D). By contrast, its recruitment is either abolished or greatly diminished in the olaparib-treated and HPF1-KO cells, respectively ( Figures S3E and S3F), 10,34 establishing this macrodomain construct as a second sensor of mono-ADPr. The accumulation of this probe confirmed that mono-ADPr is a delayed, persistent signal upon DNA damage induction by microirradiation ( Figure 2E), in line with the results obtained with our Fabbased fluorescent sensor ( Figure 2C). Although there are differences in the precise recruitment kinetics between these two probes ( Figures 2C and 2D), this is not surprising considering the varied binding preferences of each mono-ADPr-binding tool, the distinct dynamics of different mono-ADPr targets 15 and the possibility that MacroD2 hydrolyzes mono-ADPr at non-serine residues. 34 We further corroborated our observation of transient poly-ADPr and persistent mono-ADPr by immunoblotting and immunofluorescence using orthogonal detection reagents. In H 2 O 2 -treated cells, poly-ADPr peaks at 20 min followed by a rapid decline. Conversely, mono-ADPr peaks at 30 min and remains elevated much longer ( Figures 2F and 2G).
Together, these results demonstrate that, upon DNA damage, serine mono-ADPr is a delayed, persistent signal in the DDR distinct from poly-ADPr, implicating this modification as the second wave of PARP1 signaling and suggesting a dual role for PARP1 as both an early and late responder to DNA damage.
Cellular HPF1/PARP1 ratios regulate mono-ADPr levels Considering the distinctive poly-and mono-ADPr dynamics observed (Figure 2), we were intrigued by the recent report that HPF1 dissipates from DNA lesions more slowly than PARP1, 32 a behavior reminiscent of prolonged mono-ADPr. Given the dependence of mono-ADPr on HPF1 ( Figures 1E and 2B), we hypothesized that the HPF1/PARP1 ratio dynamically regulates the mono-ADPr response to DNA damage, supported by the increased generation of mono-ADPr by higher concentrations of HPF1 observed in biochemical reactions ( Figure 3A). 13,35 Strikingly, we observe a dramatic increase in cellular mono-ADPr upon HPF1 overexpression ( Figures 3B and 3C). By livecell imaging, we detect much higher mono-ADPr levels at laser tracks when we overexpress HPF1-WT compared with HPF1-E284A, demonstrating that the increase in mono-ADPr relies on the catalytic activity of HPF1 ( Figure 3D). In ARH3-KO cells, where mono-ADPr is already elevated under basal conditions, 14,15,26 overexpression of HPF1 further increased the mono-ADPr levels ( Figure 3E). Collectively, these findings demonstrate that mono-ADPr depends on cellular HPF1/ PARP1 molarity ratios, implying the transient association of HPF1 to PARP1 12,13 as a mechanism regulating the cellular   (Figures 3B-3E). Although it was recently reported that HPF1 modulates the number and length of poly-ADP-ribose chains present at DNA lesions, 32 our modular antibodies reveal the specific effect of HPF1 on cellular mono-ADPr.
Having explored regulation of mono-ADPr at the initiation phase, we investigated whether the mono-ADPr wave results from poly-ADPr degradation. Previously, we have shown that mono-ADPr partly depends on PARG, 15 although the precise dynamic and substrate specificity were not defined. In a timecourse western blotting experiment, we discovered that the degree of dependence on PARG differs for mono-ADPr substrates ( Figure 3F). Consistent with our previous study, 15 mono-ADPr on PARP1 and core histones largely depends on PARG. Intriguingly, mono-ADPr of other primary substrates is virtually independent of PARG ( Figure 3F) and, therefore, must result directly from initial serine mono-ADPr. When specifically looking at mono-ADPr dynamics at DNA lesions, we observed that prolonged mono-ADPr is not decreased by PARG inhibition, indicating that mono-ADPr is not a remnant of poly-ADPr (Figures 3G and  3H). Thus, although mono-ADPr on PARP1, most of which is not on chromatin, 14 originates from poly-ADPr removal, the HPF1/PARP1-dependent mono-ADPr wave at DNA lesions appears to be independent of PARG action.

Identification of mono-ADPr readers by chromatin proteomics
Our evidence of serine mono-ADPr as a second wave of PARP1 signaling controlled in cells by the HPF1/PARP1 ratio (Figures 2 and 3) prompted us to determine the consequences of this PTM as a regulatory signal, complementary to but distinct from poly-ADPr. Considering its dynamics at DNA lesions, we reasoned that mono-ADPr might function as a recruitment signal at later stages of DNA repair signaling. To identify putative readers of histone mono-ADPr, we applied three complementary proteomics approaches. For two strategies, we broadened our chemical biology strategy, 15 generating three biotinylated ADP-ribosylated peptides (H3S10ADPr, H3S28ADPr, and H4S1ADPr) corresponding to primary histone mono-ADPr marks 5 for a SILAC-based peptide pull-down approach ( Figure 4A). Among the interactors, we identified DTX3L/PARP9, a complex that binds mono-and oligo-ADPr and mediates the attachment of ADP-ribose to ubiquitin, 36,37 and ALC1, an established poly-ADPr binder. [38][39][40] In addition, we identified proteins that were unknown as ADPr interactors, including ALYREF/THOC4 and SIRT6 ( Figures 4B, S4A, and S4B; Table S1), which is intriguing, given the SIRT6 participation in BER and DSB repair 41,42 and its promotion of XRCC1 and POLB recruitment to DNA damage. 33 Next, to probe mono-ADPr interactors in a more physiological context, we generated nucleosomes carrying mono-ADPr on H3S10. Although specialized approaches for generating ADP-ribosylated nucleosomes were recently proposed, 17,35 we designed an approach combining our phospho-guided strategy 15 to ADP-ribosylate long peptides site-specifically and a widely accessible approach for ligating peptides to a fully assembled tailless nucleosome ( Figures 4C and S4C-S4F). Nucleosome pull-down screens unexpectedly revealed several members of the Septin family as interactors of H3S10 mono-ADP-ribosylated nucleosomes ( Figures 4D and S4G; Table S1). Immunoblotting after nucleosome pull-down confirmed mono-ADPr binding of SEPT2 and SIRT6, an interactor identified only with peptide pull-down-based proteomic analyses ( Figures S4G and S4H). This suggests that in comparison with peptide pull-downs, the abundant recombinant histones hamper proteomic identification of low-abundance mono-ADPr interactors and underscores the complementarity of these two approaches.
Finally, to identify proteins whose recruitment to chromatin is regulated specifically by mono-ADPr directly in cells, we combined quantitative proteomics of chromatin-associated proteins with targeted modulations of cellular mono-ADPr described above ( Figures 4E, S4I, and S4J). As expected, upon H 2 O 2 treatment, we observed a broad rearrangement of the chromatinassociated proteome ( Figure 4F; Table S1) and enrichment of proteins involved in the DDR ( Figure S4K), validating our approach. We reasoned that we could specifically determine the mono-ADPr-dependent chromatin-associated proteome by analysis of H 2 O 2 -treated HPF1-KO cells complemented with HPF1-WT or HPF1-E284A. Two ubiquitin E3 ligases, RNF114 and the DTX3L/PARP9 complex, were exclusively enriched on chromatin under HPF1-WT overexpression compared with HPF1-E284A overexpression ( Figure 4G). To explore the chromatin-associated proteome beyond exogenously induced DNA damage, we abolished mono-ADPr in ARH3-KO cells with olaparib ( Figure S4L) and observed reduced chromatin associations of RNF114 and DTX3L/PARP9, as well as several known PARP1modulated proteins, including XRCC1, LIG3, and APLF ( Figure 4H). Direct comparisons between WT and ARH3-KO cells revealed that chromatin binding of endogenous RNF114 and DTX3L/PARP9 is controlled by mono-ADPr (Figures 4I, 4J, and  S4L; Table S1). To corroborate the interactions of DTX3L and RNF114 with mono-ADPr, we immunoprecipitated GFP-RNF114 and GFP-DTX3L and observed co-precipitation of  Technology mono-ADPr-ribosylated H3 and PARP1, interactions that were abolished by olaparib ( Figures 4K and S5A-S5C). Using AbD33644, our site-specific H3S10/S28mono-ADPr antibody, 15 we observed that these sites are recognized by RNF114 and that serine-to-alanine substitutions almost completely disrupted the interaction ( Figures S5D and S5E). Continuous H 2 O 2 treatment in ARH3-KO cells produces persistently high levels of mono-ADPr up to 90 min after induction of treatment, 14,15 and accordingly, proteomics revealed greater levels of chromatin-bound RNF114 and DTX3L/PARP9 after 90 min of H 2 O 2 treatment, which were abolished by olaparib ( Figure S5F). Under such high levels of mono-ADPr, we also observed recruitment of other known ADPr binders, such as PARP12, 40 to chromatin. These observations are not limited to DNA damage by H 2 O 2 , since methylmethanesulfonate (MMS) treatment also robustly enriched RNF114 on chromatin ( Figure S5G). Collectively, these proteomics datasets provide a systematic overview of protein binding to chromatin mono-ADPr on physiologically relevant substrates and represent a resource for future studies.
Chromatin mono-ADPr functions as a recruitment signal for RNF114 Among the identified putative readers of mono-ADPr (Figure 4), we were particularly intrigued by RNF114. 43 Figure 2D), we observed enhanced late accumula-tion of RNF114 at the sites of damage ( Figure 5D). To confirm that mono-ADPr drives RNF114 accrual at DNA lesions, we reasoned that, due to the high activity of PARG, late inhibition of PARP1 would abolish poly-ADPr but have only a minor effect on mono-ADPr. As expected, late olaparib treatment resulted in fast degradation of poly-ADPr demonstrated by the rapid decrease in the levels of WWE and the poly-ADPr reader APLF ( Figures 5E and 5F). In contrast, mono-ADPr remained largely unaffected, and importantly, RNF114 recruitment was not impaired by late olaparib treatment ( Figures 5G and 5H). Conversely, abolishing mono-ADPr by ARH3 overexpression while preserving poly-ADPr prevented the recruitment of RNF114 ( Figures 5I-5K). In addition, although the poly-ADPr binders ALC1 and APLF 38,44 are immediately recruited to DNA lesions, accumulation of RNF114 clearly follows the dynamics of the mono-ADPr wave ( Figures 5L and S6D). Altogether, these data provide conclusive evidence that mono-ADPr recruits RNF114 to DNA lesions. Next, considering our chromatin proteomics analysis of ARH3-KO cells ( Figures 4H-4J), we expected RNF114 recruitment to chromatin under physiological conditions when histone mono-ADPr is elevated in the absence of exogenous DNA damage. Using fluorescence correlation spectroscopy (FCS), 45 we observed a reduced diffusion of RNF114 and the mono-ADPr binding domain in the ARH3-KO compared with WT cells, suggesting increased binding to mono-ADP-ribosylated chromatins. This effect was abolished by long-term olaparib treatment and boosted by HPF1 overexpression, whereas the recruitment of the poly-ADPr binder APLF was unaffected ( Figures 5M and S6E), indicating that mono-ADPr-mediated chromatin association of proteins might represent a response to HPF1/PARP1 signaling beyond the repair of exogenous DNA damage. Overall, these results demonstrate that recruitment and chromatin retention of RNF114 depends on mono-ADPr.

RNF114 recruitment to DNA lesions is mediated by its zinc-finger domains
To determine whether RNF114 directly interacts with serine mono-ADPr, we tested the ability of recombinant RNF114 to bind a histone mono-ADP-ribosylated peptide. 15 Although RNF114 has not been reported to bind serine mono-ADPr, we were encouraged by evidence of mono-ADPr-dependent interaction between RNF114 and PARP10. 46 We observed binding  Table S1. Technology to H3(1-21)S10ADPr, but not its unmodified counterpart, and only weak binding to poly-ADPr, especially when compared with known poly-ADPr binders ( Figures 6A and 6B). The inability of RNF114 to bind mono-ADPr upon zinc depletion by EDTA ( Figure 6A) and the existence of a PBZ motif 44 made us hypothesize that zinc fingers (Zns) of RNF114 are responsible for its mono-ADPr-dependent recruitment to DNA damage sites. To test this, we deleted the three Zns of RNF114 individually ( Figures 6C and S6F) and monitored their recruitment to DNA lesions. The recruitment was abolished when either Zn2 or Zn3 was deleted, whereas the Zn1 deletion mutant mostly retained its ability to accumulate at DNA damage sites ( Figures 6D and  S6G). Zn2 and Zn3 are atypical C2H2 Zns that together constitute the drought-induced 19-protein type, zinc-binding domain (Di19_Zn-bd). 47 Mutation of the highly conserved cysteine C176 of Di19_Zn-bd prevented the recruitment of RNF114 to DNA damage sites ( Figure 6E) and its binding to a mono-ADP-ribosylated peptide ( Figure 6F). Consistently, immunoprecipitated GFP-RNF114-WT, but not GFP-RNF114-C176A, co-precipitated mono-ADP-ribosylated H3 and PARP1 ( Figure 6G). From these results, we concluded that the Di19_Zn-bd domain is essential for mono-ADPr-dependent recruitment of RNF114 to DNA lesions.
RNF114 modulates the alternative lengthening of the telomeres pathway and the DDR The loss of ARH3-KO was previously associated with impaired telomere extension by the ALT pathway, 14 suggesting a role of serine mono-ADPr in telomere maintenance. To investigate serine mono-ADPr in this process, we evaluated the formation of ALT-associated PML bodies (APBs) foci in HPF1-KO cells.
Complementation with HPF1-WT rescued APB foci formation to the level of unperturbed WT cells, an effect that was not observed with HPF1-E284A ( Figure 7A). These results were verified through the siRNA depletion of HPF1 in U2OS cells and LM216J, an additional ALT cancer cell line ( Figure S7A). Given the occurrence of mono-ADPr at the sites of telomeric damage ( Figure 1F) and its functional significance 14 ( Figure 7A), we hypothesized that RNF114 could be recruited to telomeres. We observed the localization of RNF114-WT at the sites of telomeric damage in WT cells, but not in HPF1-KO cells ( Figure 7B). Telomeric accumulation of RNF114 dramatically increased in ARH3-KO cells as expected, given the higher levels of mono-ADPr (Figure 1F). By contrast, GFP-RNF114-C176A failed to accumulate at damaged telomeres in WT and ARH3-KO cells ( Figure 7B).
Interestingly, we also observed the HPF1-dependent recruitment of other readers of mono-ADPr (Figure 4), namely SEPT6 and PARP9, to damaged telomeres ( Figures S7B and S7C). Our observation of RNF114 recruitment to damaged telomeres ( Figure 7B) prompted us to investigate whether RNF114 also plays a role in telomere maintenance. Similar to siHPF1, siRNF114 alone significantly reduced the APB formation ( Figures 7C, S7D, and S7E). Importantly, simultaneous siRNAs of HPF1 and RNF114 failed to further decrease the APB foci ( Figures 7C, S7D, and S7E), indicating participation in the same pathway. To further investigate the role of RNF114 at telomeres, we generated the RNF114-KO cells ( Figure S7F). Consistent with the results obtained with siRNAs of RNF114, APB formation is reduced in RNF114-KO cells and is rescued by the expression of RNF114-WT, but not RNF114-C176A ( Figure 7D), indicating the importance of RNF114 as a mono-ADPr effector in this process. We obtained similar results with an assay to visualize the ALT telomere DNA synthesis, for which APBs are functionally important. 48 ALT DNA synthesis, measured by the number of EdU+ telomeres, is decreased in RNF114-KO cells, and RNF114-WT, but not RNF114-C176A, expression rescued this phenotype ( Figure 7E). Taken together, these results suggest a role of mono-ADPr in ALT telomere maintenance and implicate RNF114 as an effector of mono-ADPr.
Next, we investigated the role of RNF114 in DDR and assessed the sensitivity of RNF114-deficient cells to DNAdamaging agents. Consistent with the role of RNF114 in the DDR, RNF114-KO cells exhibited significantly higher sensitivity to MMS and IR treatment than WT cells ( Figures 7F and S7G). 53BP1 is a key regulator of non-homologous end-joining (NHEJ)-mediated double-strand break repair, 49 and PARP1 loss is associated with a reduction of 53BP1 foci accumulation. 50 Therefore, we investigated whether RNF114 plays a role in this pathway. RNF114-KO-and siRNF114-treated cells exhibited significantly reduced 53BP1 foci formation upon DNA damage ( Figures 7G and S7H), without markedly altering gH2AX foci ( Figures 7H and S7I). RNF114-WT, but not RNF114-C176A, rescues the reduced 53BP1 foci formation upon DNA damage in the RNF114-KO cells ( Figure 7I). RNF114 depletion also significantly impaired the foci formation of RIF1, a key effector of 53BP1 49 ( Figures 7K and S7J). Using a live-cell probe specific for H2AK13/15Ub, 51 the histone mark responsible for 53BP1 recruitment to DNA lesions, in RNF114-KO cells, we observed lower levels of this mark at DNA lesions ( Figure 7J) and, accordingly, a decrease in foci  Figure S6 and Table S1.

OPEN ACCESS
Technology accumulation of RNF168, the writer of H2AK13/15Ub (Figure 7L). Altogether, these results are consistent with the role of RNF114 in the DDR and particularly in the NHEJ pathway. Direct assessments of RNF114 contribution to NHEJ with U2OS-EJ5 reporter cells 52 revealed that siRNF114 decreases the frequency of NHEJ events, similar to XRCC4, a core factor required for NHEJ ( Figure 7M).
These results shed new light on HPF1/PARP1 signaling by linking the mono-ADPr effector RNF114 to telomere maintenance and DNA repair.

DISCUSSION
Although recently introduced tools have significantly advanced ADPr research, 15,19,22,25 our ability to study many forms of this elusive PTM still lags far behind other important PTMs, such as phosphorylation and ubiquitination, for which many mature tools have been made available over the course of several decades. To bridge this technical gap, in this work, we have applied the SpyTag technology 23,24 and affinity maturation to the recombinant antibodies generated with Ser-ADPr-ribosylated peptides according to our phospho-guided enzymatic approach. 15 This has advanced detection of ADPr at multiple levels, creating a blueprint for the future development of antibodies for the sensitive and versatile detection of various elusive targets. The expandable family of formats thus available for every generated recombinant antibody ( Figure 1A) may even help ADPr ''leapfrog'' other PTMs, which rely primarily on non-recombinant antibodies, such as rabbit IgG. As illustrated here for our mono-ADPr antibodies, this system allows straightforward creation of fluorescent probes for livecell imaging ( Figure 2). Importantly, the ongoing development of site-specific ADPr antibodies 15 will soon enable us to track the dynamics of histone and other serine mono-ADPr marks in living cells. This prospect is particularly appealing, considering the distinct dynamics of different serine mono-ADPr targets on DNA damage, as shown for PARP1 and H3. 15 Furthermore, the high sensitivity that we achieve by combining a bivalent HRP-coupled Fab format with affinity maturation extends the reach of these antibodies to the investigation of low-level mono-ADPr events in various signaling pathways regulated by many known mono-ADP-ribosyltransferases. 21 Limitations of the study The SpyTag technology 23 has the fundamental limitation that it cannot be applied to non-recombinant antibodies, such as the poly-ADPr antibody clone 10H, and the SpyTag needs be added to existing recombinant reagents by cloning. Although an increasing number of SpyTagged anti-ADPr antibodies and SpyCatcher formats are becoming available, in most cases, the users must couple their own SpyTagged antibody to SpyCatcher reagent following a simple 1 h protocol. 24 With the SpyTag-based antibodies described here, we show that serine mono-ADPr is the second wave of PARP1 signaling in the context of DNA damage. Its immediate formation and large size make poly-ADPr one of the earliest signals produced during the DDR, and it constitutes a potent ''emergency'' trigger of DNA repair initiation, 53 but its toxicity 14 makes it unsuitable as an enduring signal. If poly-ADPr were its only signal, the reach of PARP1 signaling would be severely limited by the transient nature of poly-ADPr. Instead, we propose that serine mono-ADPr extends the reach of PARP1 signaling in the form of a second, enduring PTM with which biological processes are regulated over an extended period of time. The persistence of mono-ADPr and transience of poly-ADPr explain the recent puzzling observations that mono-ADPr is more abundant than poly-ADPr in cells upon DNA damage. 15,16 Dynamic modulation of the PARP1/HPF1 ratio in the chromatin milieu constitutes the molecular basis of the dual activity of PARP1 as a poly-ADPr and mono-ADPr transferase and acts as a cellular mechanism to regulate the levels of chromatin mono-ADPr ( Figure 3). By identifying the readers of chromatin mono-ADPr, we illustrate how such a two-speed signaling pathway operates in the recruitment of proteins to the sites of DNA damage. Although the recruitment of poly-ADPr readers is immediate and largely temporary ( Figures 2 and 5L), 31 the assembly of a mono-ADPr reader, exemplified here by RNF114, is progressive and enduring ( Figure 5). Chromatin retention of mono-ADPr readers also occurs in the absence of exogenous DNA damage, as we have shown for cells lacking ARH3. Although the toxicity of high poly-ADPr levels has been suggested as the cause of neurodegeneration in patients with ARH3 deficiency 14 in agreement with the current model of mono-ADPr as a primer for poly-ADPr, 12,13 toxic accumulation of mono-ADPr readers on chromatin might constitute a key factor underlying this inherited neurodegenerative disorder. 26,54-56 Our work provides the conceptual and technological framework for further dissecting mono-ADPr signaling pathways. Further progress may include the discovery that additional cellular processes are mediated by serine mono-ADPr. Beyond basic research, we anticipate that these new tools and principles will aid the ongoing development of PARP inhibitors that specifically target the composite active sites of the serine mono-ADPr writer HPF1/PARP1. 57 We hope that our findings of serine mono-ADPr as a distinct signal shaping a second wave of PARP1 signaling will trigger a broad reinterpretation, enabled by our modular antibodies, of the scope of the HPF1/PARP1 signaling complex by suggesting a further level of control to the DDR. We propose serine mono-ADPr as a general mechanism by which PARP1 regulate various biological processes with implications for development of clinical inhibitors.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

INCLUSION AND DIVERSITY
We support inclusive, diverse, and equitable conduct of research. One or more of the authors of this paper self-identifies as a member of the LGBTQIA+ community.   Anti-pan-ADPr antibody Bio-Rad AbD33641 Anti-protein mono-ADPr antibody Bio-Rad AbD33205 Anti-mono-ADPr antibody Bio-Rad AbD33204 Anti-PARP1-S499ADPr Bio-Rad AbD34251 Anti-H3-S10/28ADPr Bio-Rad AbD33644

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Ivan Matic (imatic@age.mpg.de).

Materials availability
The antibody generated in this study (AbD43647) as well as the previously published antibodies 15  d Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS
Cell culture, SILAC labeling, and drug treatments U2OS cell lines were obtained, authenticated by STR profiling and confirmed mycoplasma free by ATCC cell line authentication services. Cells were routinely tested for mycoplasma contamination. ARH3-KO, PARP1-KO and HPF1-KO U2OS cell lines were generously provided by Ivan Ahel (University of Oxford). For the generation of RNF114-KO cell lines, the sgRNAs 5'-GCCGCTTA-CACGTGTCCGCA-3' aimed towards exon 1 (Clone 1) and 5'-AGCCGAAGAAGCCTGTCTGT-3' aimed towards exon 3 (Clone 2) of the RNF114 gene were designed using ChopChop https://chopchop.cbu.uib.no. 63 The sgRNA encoding plasmids were cloned using the Guide-it CRISPR/Cas9 System (TAKARA Bio Inc., Japan) with the pGuide-it-ZsGreen1 vector following the manufacturer's protocol. U2OS cells were separately transfected with the resulting plasmids using XfectÔ Transfection Reagent (TAKAR Bio Inc., Japan) by following the manufacturer's instructions. Two days after transfection, the cells were washed twice in PBS, trypsinized and resuspended in PBS with 10 % bovine serum and single zsGreen expressing cells were sorted into 96 well plates using FACS (FACS & Imaging Core Facility, MPI Age). Single cells were regrown and screened by western blotting using anti-RNF114 antibody. Clones lacking RNF114 expression were further validated by sequencing. Genomic DNA of potential clones and wildtype U2OS cells was isolated using NucleoSpin Tissue Kit (Macherey-Nagel, Germany).The sgRNA's target region of Clone 1 was then amplified using 5'-GGAGGAACGGGATACTTAGGAG-3' and 5'-GCAGAGCGGCAGCAAGATGG-3' as forward and revers primer, respectively. Additionally, 5'-CAAACTGGCCAGATCCCCATA-3' and 5'-GAGCGGCAGCAAGATGGC-3' were used as forward and revers primer for clone 1. The sgRNA's target region of Clone 2 was amplified using 5'-ATTTGTCGTCTTGTCTCCATGA-3' and 5'-GGAAAGACATGGACTCTTCCAC-3' as forward and revers primer, respectively. After gel extraction (GenElute Gel Extraction Kit, Sigma-Aldrich, USA) the amplicons were sequenced using the respective PCR primer. Sequencings were analyzed for deletions in RNF114 Exon 1 (Clone 1) or Exon 3 (Clone 2) by applying the sequencing to ICE CRISPR Analysis Tool https://ice.synthego. com 64 and subsequent checking for frame shifts leading to premature stop codons of the gene. Each cell line was cultured in Glutamax-DMEM supplemented with 10% bovine serum and 100 U/ml penicillin/streptomycin at 37 C and 5% CO 2.
For SILAC labeling, 65 U2OS cells were grown in DMEM for SILAC (Thermo Scientific) supplemented with unlabeled L-lysine (Silantes, Germany) and L-arginine (Silantes, Germany) for the light condition or isotopically labeled L-lysine ( 13 C 6 15 N 2 , Silantes, Germany) and L-arginine ( 13 C 6 15 N 4 , Silantes, Germany) for the heavy condition. Both light and heavy DMEM were supplemented with 10% dialyzed FBS (Thermo Scientific). Cells were cultured for more than seven generations to achieve complete labeling. Incorporation efficiency (>99%) was determined by MS.
To induce DNA damage, the cell medium was aspirated and replaced with 37 C complete DMEM containing 1 or 2 mM  For Hoechst presensitization (live-cell microscopy experiments), growth medium was aspirated and replaced with fresh medium containing 0.3 mg/mL Hoechst 33342 for 1 h at 37 C. Immediately prior to imaging, growth medium was replaced with CO 2 -independent imaging medium (Phenol Red-free Leibovitz's L-15 medium (Life Technologies) or Molecular Probes Live Cell Imaging Solution (Invitrogen) supplemented with 20% fetal bovine serum, 2 mM glutamine, 100 mg/mL penicillin and 100 U/mL streptomycin).
To generate stable cell lines complemented with GFP empty vector (EV), GFP-RNF114-WT, or GFP-RNF114-C176A; U2OS RNF114-KO cells were plated in 6 cm dishes and transiently transfected with the corresponding plasmid (EV, GFP-RNF114-WT, GFP-RNF114-C176A) using lipofectamine 2000 according to manufacturer's protocol at 1:2 ratio with 2 mg of DNA. After 24 h, the cells were transferred to 15 cm dishes and grown for 24 h. Afterwards, the media was replaced with complete DMEM supplemented with 500 mg/mL G-418 solution for 14 days to select for resistant cells with stably integrated plasmids. Then, cells were collected and sorted by cytometry to select alive cells expressing GFP. For each condition, cells with similar expression levels were selected and bulk sorted before plating into 24-well plates in complete DMEM without G-418.

METHOD DETAILS
Transfections and siRNA treatments For transient expression, cells were transfected for $ 24-48 h using polyethylenimide (PEI) for mass spectrometry and immunoprecipitation experiments, lipofectamine 2000 (Sigma) for FLAG-TRF1-FokI and live-cell microscopy, and XtremeGENE HP (Sigma) for live-cell microscopy according to manufacturer's instructions. Details of plasmids generated, used, and origin are in the supplementary materials.

Chromatin extraction
For cellular fractionation experiments, U2OS cells were chemically fractionated using a subcellular protein fractionation kit (Thermo Fisher Scientific, USA), according to manufacturer's instructions with minor modifications. All buffers were kept on ice and freshly supplemented with 1x protease inhibitor (Thermo Fisher Scientific, USA), 1 mM Olaparib, and 1 mM ADP-HPD. Cells were gently scraped from the surface of a 10 cm tissue culture dish into PBS and recovered by centrifugation at 500 xg for 5 min at 4 C. 100 ml of CEB buffer was added to the cell pellet and the tube incubated for 10 min on ice. Following centrifugation at 500 xg for 5 min at 4 C, the supernatant (cytoplasmic extract) was transferred to a clean tube on ice and stored at -20 C. 100 ml of MEB buffer was added to the cell pellet, vortexed for 5 sec, and incubated for 10 min at 4 C with gentle agitation. Following centrifugation at 3000 xg for 5 min at 4 C, the supernatant (membrane bound extract) was transferred to a clean tube on ice and stored at -20 C. 50 ml of NEB buffer was added to the cell pellet, vortexed for 15 sec, and incubated for 30 min at 4 C with gentle agitation. Following centrifugation at 5000 xg for 5 min at 4 C, the supernatant (nuclear soluble extract) was transferred to a clean tube on ice and stored at -20 C. 50 ml of NEB buffer supplemented with Micrococcal Nuclease and CaCl 2 was added to the pellet, vortex for 15 sec, and incubated at 37 C for 6 min. After incubation, the tube was vortexed for 15 sec and centrifuged at 16,000 xg and the supernatant (chromatin bound nuclear extract) was transferred to a clean pre-chilled tube on ice.
Of the final 50 ml of the chromatin bound extract, 10 ml were used for immunoblotting and the remaining 40 ml were processed with acetone precipitation (as described in STAR Methods) for LC-MS/MS preparation.

Acetone precipitation of proteins
For mass-spectrometry analysis of chromatin samples, proteins were purified and desalted using a standard acetone precipitation protocol. Briefly, 4x volumes of ice-cold acetone was added to the protein sample and incubated at -20 C overnight. Following incubation, the samples were centrifuged at 16,000 xg at 4 C for 10 min and the supernatant was aspirated and discarded. The protein pellet was washed in ice-cold acetone, centrifuged at 16,000 xg for 5 min at 4 C and the supernatant discarded. This step was repeated once more. The final protein pellet was air-dried to remove all traces of acetone.

Technology
Recombinant protein production E. coli codon-optimized coding sequences for RNF114 and corresponding mutants (see Figure 6C) were transferred into pGEX-6P-3 vectors. Transformed BL21 strains were grown until OD 600 = 0.6 and protein expression induced by 0.2 mM IPTG for 16h at 18 C. Bacteria cells were collected, resuspended in lysis buffer (50 mM HEPES pH 7.9, 200 mM NaCl, 2.5 mM MgCl2, 5% glycerol, 1 mM DTT, supplemented with 0.25 mM PDMS, EDTA-free protease inhibitor cocktail, 1 mg/ml lysozyme, 75 U/ml benzonase) at 1:4 bacteria weight to buffer volume ratio and incubated on ice for 30 min. The suspension was probe-sonicated for 30 sec on ice, then centrifuged at 18,000 xg for 45 min at 4 C. The supernatant was collected and recombinant proteins purified by incubation with Glutathione Sepharose 4B beads overnight at C, washed trice with ice-cold PBS and eluted with elution buffer (50 mM HEPES pH 7.9, 200 mM NaCl, 1 mM DTT, 20 mM reduced glutathione). The eluate was aliquoted and stored at -80 C.
Preparation of nuclear extract for peptide and nucleosome immunoprecipitation U2OS WT cells were washed in ice-cold PBS, cells were scraped from the surface of a 10 cm tissue culture dish into PBS and recovered by centrifugation at 500 xg for 5 min at 4 C. The pellet was resuspended in hypotonic buffer (10 mM HEPES pH 7.9; 10 mM KC1; 0.1 mM EDTA; 0.1 mM EGTA; 1 mM DTT; 0.5 mM PMSF, supplemented with 1 mM Olaparib, 1 mM ADP-HPD, and 1x EDTA-free protease inhibitors), incubated for 15 min on ice. Following incubation, NP-40 to a final concentration of 0.6% was added and the sample vortexed for 10 sec. The sample was centrifuged at 14,000 xg for 30 sec at 4 C. The supernatant (cytosolic fraction) was discarded. The pellet was washed once in hypotonic buffer and centrifuged at 14,000 xg for 30 sec at 4 C. The pellet was resuspended in icecold salt extraction buffer (20 mM HEPES pH 7.9, 420 mM NaCl, 0.2 mM EDTA, 0.1% NP-40, 20% glycerol, supplemented fresh with 1 mM Olaparib, 1 mM ADP-HPD, 1x protease inhibitors (EDTA free), 1 mM DTT) and incubate with end-over-end rotation for 30 min at 4 C. After incubation, the sample was centrifuged at 16,000 xg for 10 min at 4 C. Following centrifugation the supernatant (nuclear extract) was collected, diluted to a final concentration of 150 mM NaCl and $ 0.6 mg/mL proteins, flash frozen in liquid nitrogen, and stored at -80 C until further processing for nucleosome and peptide pulldowns.

Peptide and nucleosome immunoprecipitation
Peptide and nucleosome immunoprecipitation were performed as previously described. 66 For peptide pulldowns, 1 mg of biotinylated H3S10ADPr or H3WT (control) peptide was bound to 75 mL Dynabeads MyOne C1 streptavidin (Pierce) in binding buffer (20 mM HEPES pH 7.9, 150 mM NaCl, 0.1 % NP-40, 20% glycerol, supplemented fresh with 1 mM DTT) for 20 min at RT. Following extensive washing of unbound peptides, the beads-peptide mix was incubated with 500 mg of ice-cold nuclear extract (see STAR Methods for nuclear extract preparation) with end-over-end rotation for 2 h at 4 C. Following incubation, the beads were washed twice in binding buffer, and twice in binding buffer without NP-40 and glycerol (20 mM hepes pH 7.9, 150 mM NaCl, supplemented fresh with 1 mM DTT). Elution and digestion for mass spectrometry were performed with incubation with 50 mL of digestion buffer (50 mM TEAB pH 8.5, 2.5 mM TCEP, 10 mM CAA, 0.002 mg/mL Lys-C, 0.02 mg/mL Trypsin) for $ 14-16 h at 37 C. The eluate was collected and processed with stagetips as described in STAR Methods.
Nucleosome pulldowns were performed as described above, with minor modifications. Briefly, 5 mg of WT nucleosome or H3S10ADPr nucleosome (with biotinylated 601 DNA sequence) was bound to 75 mL Dynabeads MyOne T1 streptavidin (Pierce).

Co-immunoprecipitation
48 h post-transfection U2OS cells were either left untreated or treated with 2 mM H2O2 for 30 min. Following one wash in ice-cold PBS, cells were scraped from the surface of a 10 cm tissue culture dish into PBS and recovered by centrifugation at 500 xg for 5 min at 4 C. The cell pellet was then resuspended in 100 ml of Lysis Buffer (20 mM HEPES pH 7.9, 300 mM NaCl, 2.5 mM MgCl 2 , 0.5% NP-40, 20% glycerol, 750 U/ml Benzonase, 10 mM Olaparib, 1X ADP-HPD, 1X EDTA-free protease inhibitor) and incubated on end-over-end rotation for 1 h at 4 C. After quenching the reaction with 100 ml of Quenching Buffer (20 mM HEPES pH 7.9, 30 mM EDTA, 10 mM Olaparib, 1X ADP-HPD, 1X EDTA-free protease inhibitor) the lysate was clarified at 20,000 xg for 10 min at 4 C and the supernatant from this step was collected in fresh tubes, supplemented with 300 ml of Dilution Buffer (20 mM HEPES pH 7.9, 300 mM NaCl, 0.5 mM EDTA, 0.5% NP-40, 10 mM Olaparib, 1X ADP-HPD, 1X EDTA-free protease inhibitor) to a final volume of 500 ml. 50 ml aliquots were kept as Input samples and the remaining 450 ml of cell lysate was mixed with 10 ml bed volume of GFP-Trap magnetic particles M-270 (Chromotek, Munich, DE) and incubated for 1 h with end-over-end rotation at 4 C. Beads were recovered using a magnetic separation rack and washed 5 times with Washing Buffer (10 mM HEPES pH 7.9, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40). 50 ml of 2X Laemmli sample buffer (50 mM DTT) was then added and bound proteins were eluted by heating the beads at 95 C for 5 min. Supernatants were collected in fresh tubes as Bound samples and resolved by SDS-PAGE, followed by detection of proteins by immunoblotting.

Histone purification
Histones were purified as previously described, 5 with minor modification. Briefly, cells were treated as indicated in the figure legends, washed twice with ice-cold PBS, collected by scraping in PBS and recovered by centrifugation at 500 xg for 5 min at 4 C. Cells were lysed by rotation in 0.1 M H 2 SO 4 on end-over-end rotation for 2 h at 4 C. The lysate was centrifuged at 2,200 xg at 4 C for 20 min. The resulting pellet containing non-soluble proteins and cell debris was discarded. The supernatant was transferred to a clean pre-chilled tube and supplemented with NaCl, EDTA, and DTT to a final concentration of 0.5 M NaCl, 2 mM EDTA, 1 mM DTT. 1M Tris-HCl pH 8.0 was added until the pH was neutralized. For ion exchange chromatography, sulfopropyl (SP)-Sepharose resin (S1799, Sigma-Aldrich) was packed into a column and pre-equilibrated with 10 bed volumes of binding buffer. The neutralized supernatant was passed through the column. The resin was washed with 10 bed volumes of binding buffer and 30 bed volumes of washing buffer (50 mM Tris-HCl, pH 8.0; 0.6 M NaCl; 2 mM EDTA; 1 mM DTT). Proteins were eluted with elution buffer (50 mM Tris-HCl, pH 8.0; 2 M NaCl; 2 mM EDTA; 1 mM DTT) in six fractions. Eluted proteins were precipitated overnight in 4% (vol/vol) PCA at 4 C. The fractions were then centrifuged at 21,000 xg at 4 C for 45 min, and the resulting pellets were washed with 4% PCA in water (2 3 1 ml), 0.2% HCl in acetone (2 3 1 ml) and acetone (2 3 1 ml). The pellet was air-dried to remove all traces of acetone.
Partial FASP Partial FASP was performed as previously described. 5 Briefly, protein pellets obtained from histone purification were resuspended in 200 ml of 50 mM TEAB, pH 8.5 and loaded onto a pre-washed Vivacon 10 kDa cut-off filter. The filter was centrifuged at 10,000 xg for 20 min at RT to concentrate the proteins and then washed twice with 50 mM TEAB, pH 8.5. Protein digestion was performed with 50 ml of Trypsin Gold (Promega) at a ratio 1:2000 (trypsin:protein) for 20 min at RT. The digestion was stopped by adding 400 ml of 0.15% TFA and the resulting peptides were collected by centrifugation at 10,000 xg for 20 min at RT. 200 ml of 50 mM TEAB, pH 8.5 was added to the filter, and filters were centrifuged once more, collecting both flowthroughs in the same tube. The flowthrough (partial FASP fraction) was stage-tipped for downstream mass spectrometry processing.
Mono-ADPr immunoprecipitation of purified untreated histones Cells were grown on 245 mm square dishes and harvested in ice-cold PBS on ice. Afterwards histones were purified as described above with special care using ice-cold buffers and keeping samples on ice. Concentration of purified histones was determined using Pierce Rapid Gold BCA Protein Kit. 40 mg histones were incubated with 20 mg mono-ADPr antibody (AbD43647, IgG) overnight in a thermoshaker at 4 C. To ensure efficient binding a final protein concentration of 1 mg/ml was targeted using a 10x binding buffer (final concentration 1x: 50 mM MOPS, 10 mM Na2HPO4, 150 mM NaCl). Protein A agarose beads (ThermoFisher, #20333) were prewashed in 1x binding buffer using a centrifugation column (Pierce, #89868) and mixed with the antibody/histone complex. The bead/sample mixture was incubated for 2 h at RT using a thermoshaker. Columns were spin at 0.6 xg for 1 min at RT and washed three times in 2 bed volumes 1x binding buffer. Histones were eluted twice in 1 bed volume 0.15% TFA and elutions were pooled and dried using a speedvac. Dried samples were recovered in an appropriate volume of 1X LDS-sample buffer containing DTT and immunoblotted as described.
Protein digestion and cleanup for mass spectrometry Dried protein pellets from acetone precipitation were resuspended in digestion buffer (50 mM TEAB, pH 8.5; 2.5 mM TCEP; 10 mM CAA; 0.002 mg/mL Lys-C; 0.02 mg/mL Trypsin) and incubated for $14-16 h at 37 C. The digestion was stopped with the addition of 0.15% TFA. Digested peptides were cleaned-up on C18 stagetips according to a standard protocol 5 with 100% MeOH conditioning buffer; 30% ACN, 0.15% TFA equilibration and elution buffer; 0.15% TFA washing buffer. After the elution step, the eluate was dried to completion in a speedvac (Eppendorf). The dried peptides were resuspended in 0.1% FA for injection.

LC-MS/MS data acquisition
Three different quantitative proteomics methodologies were used in this study. For quantitation of histone modifications, a SILAC approach was used in order to make quantitation independent of possible variations in the partial digestion procedure.
Data Independent Acquisition (DIA) data were collected using Boxcar-DIA methodology on an Orbitrap Fusion coupled to an Easy-nLC 1000 (Thermo Scientific). The column was a 50 cm in-house packed emitter (medium Poroshell 120, C18, 1.9um. Emitter CoAnn (MS Wil) -75 micron -15 micron tip). Running buffers were 0.1% FA in Water (A) and 0.1%FA in 80% Acetonitrile (B). A 60-minute gradient from 4%-30% B was run at 300 nl/min followed by a steeper washing phase and a fast wash gradient at 400 nl/min.
Boxcar-DIA acquisition -MS1 spectra were collected for m/z 400-800 Th at a resolution of 120k and an AGC of 300%. Targeted SIM scans were collected in 2 sets of 8 overlapping user-defined windows spanning the precursor range, also at 120k. DIA MS2 spectra were collected in 14 Th windows from 400-800 Th, with HCD fragmentation (NCE = 27) and a resolution of 15k over the m/z range 145-1450 Th. Exact method details can be extracted from the raw files provided.
Tandem-Mass-Tag (TmT) MS3 acquisitions were performed on an Orbitrap Fusion LUMOS equipped with a FAIMs-Pro interface and coupled to an Easy-nLC 1200 with a 50 cm Acclaim Pep-map column (Thermo Scientific) and a 20 micron CoAnn emitter (MS Wil). LC buffers were identical to those described above. A 90-minute gradient of 6-31% B was run at 250 nl/min and MS data were collected with FAIMs compensation voltage (CV) alternating between -50 and -70 volts. MS1 was collected at 60k resolution with m/z 350-1500. Multiply-charged precursors were selected by a Topspeed method (1 sec cycle time per FAIMs voltage), isolated (width 0.7 Th) in the ion trap and fragmented by CID (energy 35%). The daughter ions were analyzed in the ion trap at ''Turbo'' resolution and 10 MS2 masses were selected for re-isolation for MS3. MS3 for TmT reporter quantification was performed with HCD (NCE=65) and analyzed in the orbitrap at 50k resolution. Again, method details may be extracted from the raw files provided.
Data-dependent Acquisition (DDA) SILAC data were collected by a TopSpeed method on the Orbitrap Fusion LUMOS system described above. A 110-minute gradient ran from 1-31% Buffer B at 250 nl/min followed by a steeper washing phase. MS1 spectra were acquired for m/z 350-1500 at 60k resolution and 250% AGC alternating between two FAIMs CVs (-50 and -70). Topspeed cycle time was 1 sec for each FAIMs channel. Precursors with charge states 3-9 (highest first) were isolated (quadrupole width 1.6 Th) and fragmented by electron transfer dissociation (ETD). MS2 spectra were collected at 60k resolution with an AGC of 100%. Again, method details may be extracted from the raw files provided.
In vitro HPF1/PARP1 ADP-ribosylation assay Recombinant PARP1 (100 nM), H3 (1 mM), HPF1 (concentration ranging from 100 nM to 5000 nM, as indicated), and 1 mg/mL of sonicated DNA were incubated with or without 200 mM NAD + in a PARP reaction buffer (50 mM Tris-HCl, pH 7.5; 50 mM NaCl; 1 mM MgCl 2 ) for 20 min at RT. The reaction was stopped with the addition of 4x Laemmli loading buffer supplemented with 5 mM DTT. The samples were them boiled for 5 min at 95 C and stored at -20 C until further processing for immunoblotting.
Detection of ADP-ribosylated genomic DNA by dot blot assay Control and TaqDarT-modified gDNA was prepared as previously described. 29 A 1:1 dilution series of the TaqDarT-modified gDNA was prepared and dotted along with the control gDNA both at a maximum of 1000 ng onto nitrocellulose membranes (Amersham Protran 0.45 NC nitrocellulose). The membranes were crosslinked with 1200 J using a Stratalinker UV crosslinker and immunoblotted for gDNA (autoanti-dsDNA, DSHB, 1:200) or for ADPr-gDNA using the indicated antibodies at a concentration of 1 mg/mL in 5 % (w/v) powdered milk in PBST for 1 hr at RT. Where applicable, secondary peroxidase-couple antibodies (Dako) were incubated at RT for 45 min and samples visualised on hyperfilms (GE) using an ECL western blotting detection kit (Pierce).

Recombinant ARH3/DarG treatment of ADPr-gDNA
A solution of 400 ng ADPr-gDNA was incubated with either buffer only, or buffer supplemented with 1 mM of recombinant ARH3 or DarG and incubated for 30 min at 37 C before performing a dot blot assay as described above.

Colony formation assay
For colony formation assay, cells were plated at 400 cells/well in 6-well plates and grown for 14 days in Glutamax-DMEM supplemented with 10 % bovine serum and 100 U/ml penicillin/streptomycin at 37 C and 5 % CO 2 . Approximately 12 h after plating, cells were treated by replacing medium with MMS-containing medium. Cells were incubated for 14 days and resulting colonies were stained with 0.5 % crystal violet in 25 % methanol for 30 min, washed with water, and air-dried. The plates were scanned with EVOS FL Auto 2, and quantification was performed with ImageJ/FIJI. NHEJ reporter assay U2OS-EJ5 cells (kind gift of Dr. Jeremy Stark (Duarte, CA), 52 were transfected with control siRNA, siXRCC4, siBRCA1 or siRNF114 using Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific) according to manufacturer's pro-tocol. 36 hours later, the cells were transfected with mCherry and I-SceI expression vectors, pmCherry-C1 and pCBASceI, respectively, using Xfect (Takara). pCBASceI was a gift from Maria Jasin (Addgene plasmid # 26477 67 ). 48 hours after I-SceI transfection the cells were subjected to analysis with CytoFLEX S flow cytometer (Beckman Coulter Life Sciences). The percentages of GFP+ cells were determined within the mCherry expressing cell populations, representing the frequency of NHEJ events. Quantification was done by Kaluza Analysis Software (Beckman Coulter Life Sciences).
Affinity maturation, determination of antibody affinity, and SpyTag coupling The antibody AbD33204 derived from the Fab phage display library HuCAL PLATINUM 68 was used for affinity maturation as previously described. 69 To maintain, and possibly extend, the broad mono specificity of the parental antibody, we used a selection strategy based on three different mono-ADPr peptides. During panning, a mixture of H3S10ADPr and PARP7S39ADPr (a mimic of the PARP7C39ADPr site) peptides were used as antigens, whereas the corresponding unmodified peptides were used for counter-selection. After initial panning for high-affinity clones, the PARP1S499ADPr peptide was added as an additional screening peptide. To identify successful affinity maturated clones, a subset of the clones was screened with ELISA (see ''indirect enzyme linked immunosorbent assay (ELISA)'' section in the STAR Methods for technical details) using H3S10ADPr, PARP7S39ADPr, PARP1S499ADPr, corresponding unmodified peptides, and free poly-ADPr. The corresponding antibody affinity was quantified (see below) to determine affinity and specificity compared to the parental antibody AbD33204. Affinity determination of monovalent HuCALÒ Fab antibodies generated by Bio-Rad AbD Serotec GmbH (Germany, Puchheim) was performed on an OctetÒ RH16 instrument (Sartorius) at 30 C using 384-well microplates (Greiner Bio-One, No. 781209) that were agitated at 1000 rpm. Each of the purified Fab antibodies (molecular mass 52 kDa) was measured at five concentrations between 10 and 0.65 mg/mL (200 nM -12.5 nM), and diluted in running buffer (PBS with 0.1% (w/v) bovine serum albumin (BSA), and 0.02% (v/v) TweenÒ 20). Biotinylated antigen H3-S10ad was immobilized on streptavidin (SA) biosensors (Sartorius, No. 18-5021) in phosphate-buffered saline (PBS) at 10 mg/mL for 10 min and blocked with 10 mg/ml biocytin. Typical immobilization levels were 2.6 ± 0.1 nm. After loading, sensors were kept in PBS until needed. Between measurements, the H3-S10ad-loaded biosensor surfaces were regenerated by exposing them to 10 mM glycine (pH 3.0), for 10 s followed by PBS for 30 s. Association phase was measured for 300 s, dissociation phase depending on the interaction for 200-400 s. All measurements were corrected for baseline drift by subtracting a control sensor exposed to running buffer only. Data were analyzed using a 2:1 interaction model for heterogeneous binding (fitting global, Rmax unlinked by sensor) on the Sartorius data Values are only shown for the binding event, which reflects the interaction as closely as possible. Spy-Tag coupling was performed as previously described. 24 Briefly, SpyTagged proteins and SpyCatchers were mixed in PBS and incubated for 1 h at room temperature. For bivalent SpyCatcher proteins (IgG-like and HRP-coupled formats), a 25% molar excess of SpyTagged proteins over SpyCatcher was chosen to ensure complete labeling of all SpyTag-SpyCatcher coupled products.
Indirect enzyme-linked immunosorbent assay (ELISA) Biotinylated peptides (2 mg/mL or 0.2 mg/mL) in PBS 0.1% Tween 20 (PBS-T) were immobilized on Pierce NeutrAvidin Coated Plates. After three washes with PBS-T, wells were blocked with PBS-T containing 5% BSA for 1 h at RT. Next, primary antibodies diluted to 2 mg/mL in PBS-T/5% BSA were added and incubated for 2 h at RT. After washing the wells five times with PBS-T, HRP-conjugated goat anti-human IgG F(ab') (1:5000 dilution in blocking buffer) was added and incubated for 1 h at RT. After washing the wells five times, QuantaBlu Fluorogenic Peroxidase Substrate Kit was used for detecting peroxidase activity following manufacturer instructions.

Competition ELISA
A solution of 61 nM H3S10ADPr biotinylated peptide in PBS-T was incubated on Pierce NeutrAvidin Coaded Plates for 30 min at RT. After three washes with PBS-T, wells were blocked with PBS-T containing 5% BSA for 1 h at RT. Meanwhile, adenosine, AMP, ADP, deoxy-AMP, ATP, GDP, CDP, ADP-ribose were dissolved in water to a final concentration of 17 mM and used to prepare three solutions with a final concentration of 688 mM, 43 mM, 5.5 mM. Equal volumes of a 0.1 mg/mL antibody solution (AbD33205, AbD33204, AbD43647 HRP-conjugated antibodies) was mixed with each small molecule solution to a final concentration of 0.05 mg/mL and 344 mM, 21.5 mM, 2.75 mM respectively. The antibody:small molecule solution was incubated for 1 h at RT. After blocking, the wells were washed three times with PBS-T and the antibody:small molecule solution was added and incubated for 2 h at RT. After washing the wells five times, QuantaBlu Fluorogenic Peroxidase Substrate Kit was used for detecting peroxidase activity following manufacturer instructions.
SDS cell lysate preparation U2OS cells with different genetic backgrounds (WT, HPF1-KO, ARH3-KO) were untreated or, where indicated, treated with H 2 O 2 and lysed in SDS buffer (4% SDS; 50 mM HEPES, pH 7.9), boiled for 5 min at 95 C, and sonicated for 10 cycles of 30 s on/off on a bioruptor (Diagenode) at 4 C.
Immunoblotting with on-membrane recombinant proteins treatment For the experiments in Figure S1J, immunoblotting was performed as described previously. After milk bloking, the membranes were incubated in 5 mL of 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM MgCl 2 with 1 mM recombinant ARH3, or 10 U of snake venom phosphodiesterase (svPDE) as indicated. The membranes were incubated for 3 h at room temperature before washing with TBS-T and proceeding with primary antibody incubation as detailed above.

Technology
Direct immunofluorescence (IF) General IF protocol Cells were cultured on glass coverslips, treated as indicated, and fixed with ice-cold methanol for 20 min at -20 C or 4% formaldehyde at RT. Cells were washed three times, permeabilised with 0.5% Triton X-100 for 5 min, and blocked with 3% BSA for 5 min. Incubation with primary antibodies (AbD33205: 1:1000; AbD43647: 1:1000; PAR (10H): 1:100; a-53BP1 (612523, BD transduction laboratories) 1:5000, a-RIF1 (229656, abcam) 1:400.) was followed by two washes in PBS before adding the secondary antibodies and DAPI stain. Coverslips were incubated again in 0.5% Triton X-100, washed twice in PBS and mounted with Prolong Diamond Antifade (ThermoScientific). Cells were imaged using a Leica SP8-DLS inverted laser-scanning confocal microscope using 63X objective. Nuclei were identified based on DAPI signal and ADPr fluorescence intensities were quantified in the colocalizing DAPI stained regions using Fiji Software. For quantification 100 nuclei were counted per condition. Quantification of mean grey values microscopy Raw files from microscopy were imported to Fiji. After splitting channels, the DAPI based channel was processed to generate a mask using the following filters: Minimum, radius=4; Maximum, radius=4; Subtract Background, rolling=300; Enhance Contrast, satu-rated=0.01 normalize; Auto Threshold, method=Huang white. Afterwards, particles were analysed using size=10-Infinity and edges were excluded. The mask was used to define regions of interest (ROIs) which were transferred to the ADPr fluorescence channel and mean grey values were measured. Next, another ROI was drawn manually to measure background levels which were subtracted from previously measured values. All values were imported to GraphPad Prism and plotted in a grouped data table. Immunofluorescence with on-slide recombinant protein treatment For the experiments in Figure S1J, immunofluorescence was performed as described previously using methanol fixation. After milk bloking, the slides were incubated in 500 mL of 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM MgCl 2 with 1 mM recombinant ARH3, or 1 U of snake venom phosphodiesterase (svPDE) as indicated. The slides were incubated for 3 h at room temperature before washing with PBS and proceeding with primary antibody incubation as detailed above. TRF1-FokI telomere experiments Cells on glass coverslips were washed twice in PBS and fixed with 2% paraformaldehyde (PFA) for 10 min. Cells were permeabilized with 0.1% (w/v) sodium citrate and 0.1 % (v/v) Triton X-100 for 5 min and incubated with fresh blocking solution (1 mg/mL BSA, 10% normal goat serum, 0.1% Tween) for 30 min. Primary antibodies were diluted in blocking solution and added to cells for 1 h at RT or overnight in refrigerated conditions. Next, cells were washed 3 times with PBS for 5 min and incubated with Alexa-coupled secondary antibodies (594 nm) (Life Technologies) for 1 h at RT. Following 3 washes in PBS, cells were mounted on slides with Prolong Gold Antifade reagent with DAPI (Life Technologies). Once the Prolong Anti-fade polymerized and cured, cells were visualized by conventional fluorescence with 60X and/or 63X Plan l objective (1.4 oil) using a Nikon 90I. For this set of experiments sample size was not predetermined. For EdU methods U2OS cells were transfected with TRF1-FokI plasmid as described above. Cells were pulsed with EdU (10mM) 1 h before harvest. Cells on glass coverslips were washed twice in PBS and fixed with 2% paraformaldehyde (PFA) for 10mins. Cells were permeabilized with 0.1% (w/v) sodium citrate and 0.1 % (v/v) Triton X-100 for 5mins. The Click-IT Plus EdU Cell Proliferation Kit with Alexa Flour 647 (Invitrogen) was used to detect EdU.
Antibody labeling 100 mg anti-monoADPr antibody AbD33205 in monovalent format was labelled with 50 mg DyLight 550 NHS-Ester labelling dye (Thermo Scientific #632263) as follows: The antibody was diluted with PBS to 1 mg/ml and added to the dried fluorophore and the mixture was incubated for 1 h at room temperature protected from light. The excessive fluorophore was removed using a PD minitrap G-25 (sigma Aldrich #28-9180-07) according to the manifacture's instructions. The elution was concentrated on a 10 kDa cut-off filter to final 0.5 mg/ml. Afterwards, fluorophore/antibody ratio was measured by a nanodrop. The labelled antibody was stored at 4 C protected from light. See also Figure S2G.

Antibody loading
The antibody was loaded into the cell as described in Cialek et al. 71 with minor modifications. In brief, 10 5 U2OS cells were seeded onto 35 mM high glass bottom dishes (ibidi #81158) two days prior to bead loading. Glass beads (% 106 mm) were cleaned and acidic washed by a sodium hydroxide wash, followed by two sterile water washes and a final ethanol wash. Beads were dried under sterile conditions and irradiated by several rounds of UV light. Afterwards beads were transferred into a jar covered by a 105 mM polypropylene mesh (B€ uckmann #PP 105/106-25) to ensure an even monolayer of beads when added to the cells. 5 mg labelled antibody was diluted in 50 ml PBS. DMEM covering the cells was removed and cells were washed once with PBS. 50 ml antibody loading solution was added to the cells and immediately covered by a monolayer of glass beads. Afterwards, the dish was tapped 10 times on the bench using medium force and 2 ml DMEM medium was added to the cells to prevent drying. Cells were incubated for 5 minutes with antibody/bead mixture before carefully washing off the glass beads with DMEM. Fresh DMEM with 0.3 mg/ml Hoechst 33342 (Thermo Scientific #H3570) was added and incubated for 1 h at 37 C. DMEM was replaced by live cell imaging solution (Invitrogen #A14291DJ). At this stage, cells are ready for live cells imaging as described in the ''protein recruitment kinetics at sites of laser irradiation'' section of the STAR Methods. See also Figure S2G. Protein recruitment kinetics at sites of laser irradiation The monitoring of protein recruitment kinetics at sites of laser irradiation was performed as described previously. 31 In brief, images were acquired either on a Ti-E inverted microscope from Nikon equipped with a CSU-X1 spinning-disk head from Yokogawa, a Plan APO 60x/1.4 N.A. oil-immersion objective lens and a sCMOS ORCA Flash 4.0 camera; or on an Olympus Spin SR spinning disc system equipped with a CSU-W1 spinning-disk head from Yokogawa (50 micron pinhole size), a UPLSAPO 100XS/1.35 N.A. silicon-immersion objective lens and a Hammamatsu sCMOS ORCA Flash 4.0 camera. Laser irradiation of Hoechst-presensitized cells was performed along a 10 or 16 mm-line through the nucleus with a continuous 405 nm laser set at 125-130 mW at the sample level. Cells were maintained at 37 C with a heating chamber. Analysis of the image sequences was performed either automatically using a custom MATLAB routine as described previously, 31 or by manually segmenting the irradiated area as well as the whole nucleus on ImageJ/FIJI or Olympus CellSense. Mean fluorescence intensities with the irradiated area was background substracted, divided to the mean nuclear intensity to correct for imaging photobleaching, and then normalized to the signal prior to DNA damage.
Fluorescence correlation spectroscopy Fluorescence Correlation Spectroscopy (FCS) experiments were performed on a Zeiss LSM880 confocal setup equipped with a C-Apo 40x/1.2 N.A. water immersion objective. Fluorescence of GFP and mCherry was excited at 488 nm and 561 nm, respectively, and single emitted photons were detected and counted on the GaAsP spectral detector with spectral detection windows of 500-550 nm and 580-650 nm, respectively. Raw photon fluctuation traces were acquired for 20 seconds and detrended for slow fluctuations using the Fluctuation Analyzer 4G software 45 before autocorrelation. Cells were maintained at 37 C with a heating chamber. Autocorrelation curves were fitted with the following effective diffusion model: ( Equation 1) where N is the number of tagged molecules in the focal volume, D is the effective diffusion coefficient, u is the radial radius of the focal volume and s is the shape factor. N and D are fitted parameters while u and s are fixed and set to 160 nm and 6, respectively. For the experiment in Figure 5I, Effective diffusion coefficient measured by FCS for GFP-RNF114 (left) and mono-ADPr probe (right). Left: WT or ARH3-KO U2OS cells were transfected with GFP-RNF114 alone or together with mCherry-tagged HPF1-WT and treated or not with 1 uM Olaparib for 24h. Right: WT or ARH3-KO U2OS cells were transfected with mono-ADPr probe alone or together with GFP-tagged HPF1-WT and treated or not with 1 uM Olaparib for 24h.

Bioinformatic analysis of mass spectrometry samples
For DIA analysis, the protein_groups file was processed with an in-house R script using LIMMA 72 for significance testing of log 2 intensities. Multiple testing-corrected p-values <0.05 (-log 10 (adj. p val) > 1.3) were considered significant. The GO term enrichment analysis was performed with the DAVID online platform. 73 For TMT analysis, proteomics data was analyzed using MaxQuant, version 1.6.17.0. 61 The isotope purity correction factors, provided by the manufacturer, were included in the analysis. Differential expression analysis was performed using limma, version 3.34.9, 72 in R, version 3.4.3 (https://www.R-project.org/).

Statistics on microscopy data
To calculate significance between multiple groups and conditions an ordinary two-way analysis of variance (ANOVA) test was conducted and Tukey's multiple comparisons post-test was performed. Calculated p-values are represented by asterisks on the graph and listed in the figure legends.

Detailed protocol
Immunoblotting using the HRP-coupled SpyTag format (Methods S1, related to STAR Methods and Figure 1     Corresponding quantification (right) of AbD43647 signal overlapping with mitochondria staining. Data is mean ± SEM from a representative of n = 3 biological replicates. ns, not significant (unpaired Student's t test). Related to     The chromatin bound fraction was analyzed by immunoblotting using the indicated antibodies. independent experiment is shown. * P-value < 0.05 (two-tailed Student's t test). Scale bars: 5 µm. Related to Figure   7K.