Decitabine cytotoxicity is promoted by dCMP deaminase DCTD and mitigated by SUMO-dependent E3 ligase TOPORS

The nucleoside analogue decitabine (or 5-aza-dC) is used to treat several haematological cancers. Upon its triphosphorylation and incorporation into DNA, 5-aza-dC induces covalent DNA methyltransferase 1 DNA–protein crosslinks (DNMT1-DPCs), leading to DNA hypomethylation. However, 5-aza-dC’s clinical outcomes vary, and relapse is common. Using genome-scale CRISPR/Cas9 screens, we map factors determining 5-aza-dC sensitivity. Unexpectedly, we find that loss of the dCMP deaminase DCTD causes 5-aza-dC resistance, suggesting that 5-aza-dUMP generation is cytotoxic. Combining results from a subsequent genetic screen in DCTD-deficient cells with the identification of the DNMT1-DPC-proximal proteome, we uncover the ubiquitin and SUMO1 E3 ligase, TOPORS, as a new DPC repair factor. TOPORS is recruited to SUMOylated DNMT1-DPCs and promotes their degradation. Our study suggests that 5-aza-dC-induced DPCs cause cytotoxicity when DPC repair is compromised, while cytotoxicity in wild-type cells arises from perturbed nucleotide metabolism, potentially laying the foundations for future identification of predictive biomarkers for decitabine treatment.

1) Every manuscript requires a Data Availability section (even if only stating that no deposited datasets are included).Primary datasets or computer code produced in the current study have to be deposited in appropriate public repositories prior to resubmission, and reviewer access details provided in case that public access is not yet allowed.Further information: embopress.org/page/journal/14602075/authorguide#dataavailability2) Each figure legend must specify -size of the scale bars that are mandatory for all micrograph panels -the statistical test used to generate error bars and P-values -the type error bars (e.g., S.E.M., S.D.) -the number (n) and nature (biological or technical replicate) of independent experiments underlying each data point -Figures may not include error bars for experiments with n<3; scatter plots showing individual data points should be used instead.
3) Revised manuscript text (including main tables, and figure legends for main and EV figures) has to be submitted as editable text file (e.g., .docxformat).We encourage highlighting of changes (e.g., via text color) for the referees' reference.5) Point-by-point response letters should include the original referee comments in full together with your detailed responses to them (and to specific editor requests if applicable), and also be uploaded as editable (e.g., .docx)text files.6) Please complete our Author Checklist, and make sure that information entered into the checklist is also reflected in the manuscript; the checklist will be available to readers as part of the Review Process File.A download link is found at the top of our Guide to Authors: embopress.org/page/journal/14602075/authorguide7) All authors listed as (co-)corresponding need to deposit, in their respective author profiles in our submission system, a unique ORCiD identifier linked to their name.Please see our Guide to Authors for detailed instructions.8) Please note that supplementary information at EMBO Press has been superseded by the 'Expanded View' for inclusion of additional figures, tables, movies or datasets; with up to five EV Figures being typeset and directly accessible in the HTML version of the article.For details and guidance, please refer to: embopress.org/page/journal/14602075/authorguide#expandedview9) Digital image enhancement is acceptable practice, as long as it accurately represents the original data and conforms to community standards.If a figure has been subjected to significant electronic manipulation, this must be clearly noted in the figure legend and/or the 'Materials and Methods' section.The editors reserve the right to request original versions of figures and the original images that were used to assemble the figure.Finally, we generally encourage uploading of numerical as well as gel/blot image source data; for details see: embopress.org/page/journal/14602075/authorguide#sourcedataAt EMBO Press, we ask authors to provide source data for the main manuscript figures.Our source data coordinator will contact you to discuss which figure panels we would need source data for and will also provide you with helpful tips on how to upload and organize the files.
Further information is available in our Guide For Authors: https://www.embopress.org/page/journal/14602075/authorguide In the interest of ensuring the conceptual advance provided by the work, we recommend submitting a revision within 3 months (15th Jan 2024).Please discuss the revision progress ahead of this time with the editor if you require more time to complete the revisions.Use the link below to submit your revision: Link Not Available ------------------------------------------------Referee #1: In this manuscript, Carnie et al. investigate the determinants of cytotoxicity associated with 5-aza-dC, also known as decitabine, through CRISPR/Cas9 screens.Decitabine is used in the treatment of certain blood disorders, such as MDS and AML.The therapeutic mechanism of 5-aza-dC is believed to be related to its ability to trap DNA methyltransferases (DNMTs) onto DNA, resulting in DNA hypomethylation as well as the formation of DNA-protein crosslinks (DPCs) containing DNMT1.However, the exact mechanisms underlying its action and patient responses remain unclear.
This study used CRISPR/Cas9 screens to identify genes affecting sensitivity or resistance to 5-aza-dC.It revealed that, in DPCrepair-proficient cells, 5-aza-dC's cytotoxicity results from the conversion of 5-aza-dC to 5-aza-dUMP by DCTD, suggesting that the origin of cytotoxicity is independent of DNA hypomethylation and DNMT1-DPCs.
To gain more insights, the authors performed another CRISPR/Cas9 screen, this time in DCTD KO cells, and found that DPC repair pathways become more prominent determinants of 5-aza-dC sensitivity under this condition.Using a modified iPOND technique (called iPOND-DPC) to identify the proximal proteome of DNMT1-DPCs, the authors identified, among other known DPC repair factors, TOPORS (a SUMO1-and ubiquitin E3 ligase) as a factor that confers protection against DNMT1-DPCs.They propose that TOPORS is recruited to decitabine-induced DNMT1-DPCs in a SUMO-dependent but ubiquitin-independent manner, facilitating the proteolysis of DNMT1-DPCs in parallel to RNF4 and UBE2K.This is an important study that challenges current assumptions regarding decitabine's mechanisms of action.This study also reveals a novel DPC repair mechanism mediated by TOPORS, highlighting the critical role of SUMO modifications in DPC repair.
The presented data are of high quality.The proposed model is supported both by genetic evidence on 5-aza-dC sensitivities and by cellular data on TOPORS recruitment to SUMOylated DNMT1.The study could be enhanced by further investigations into the mechanism underlying the SUMO-dependency of TOPORS recruitment.Additionally, whether TOPORS' ubiquitin E3 ligase activity is required for the repair of DNMT1-DPCs is an important question to address to substantiate their model that TOPORS polyubiquitinates DNMT1-DPCs for proteasomal degradation.Comments are listed below for the authors to consider during the revision.
Major points: 1.Does DCTD KO result in increased DPC burden during 5-aza-dC treatment?Fig. 2b appears to show that the DNMT1-DPC level is higher in DCTD KO cells compared to WT, suggesting that DCTD might funnel the 5-aza-dC metabolite away from DNMT-DPC formation in WT cells.However, the authors do not seem to acknowledge this, given that they state "DNMT1-DPC induction in DCTD KO cells was similar to that in WT cells, (Fig 2b)" (Line 152).2. It would be helpful to highlight the ranking of DNMT1 in Fig. 1b and Fig. 2c.In Fig. 1b, DNMT1 KO would not confer resistance as cytotoxicity in WT cells is independent of DNMT1.In Fig. 2c, on the other hand, DNMT1 KO would be expected to provide some resistance to 5-aza-dC if DCTD KO results in increased DNMT1-DPCs (as mentioned in Point #1). 3. How TOPORS promotes DPC tolerance is not addressed sufficiently.Are the RING domain and SIMs of TOPORS important for DNMT1-DPC repair in Fig. 5g?Rescue experiments with WT or mutant TOPORS would help address this question.Additionally, are the TOPORS SIMs important for the DNMT1 interaction in Fig. 4i,j? 4. Does TOPORS contribute to SUMOylation of DNMT1-DPCs?Mailand's group has reported that there are SUMO E3 ligases for DNMT1-DPCs redundant to PIAS4, but TOPORS was not included in their study (Ref #37). 5. Line 251: "While we detected an interaction between WT GFP-TOPORS and a high molecular weight ubiquitylated interactor, this was not overtly 5-aza-dC-inducible but undetectable in co-immunoprecipitates of GFP-TOPORS CCAA (Fig 4i)."Could this be TOPORS auto-ubiquitination? TOPORS pull-down under denaturing condition could clarify this possibility.The same experiment could also be used to test if the SUMO conjugates in the pull-down is reduced under a denaturing condition.That would support authors' interpretation that the SUMO chains are conjugated to interacting proteins (presumably DNMT1) but not to TOPORS itself.
Minor points: 1. Supplemental Fig. 3a,b: Is there any TOPORS antibody that can validate the lack of TOPORS proteins in TOPORS KO and siTOPORS cells? 2. Line 64: "In addition to DNMT1 DNA-protein crosslink (DPC) induction," requires a prior introduction of DNMT1-DPC formation.

Referee #2:
General summary & opinion of study -significance, questions, and findings The study by Carnie et al seeks to identify factors that mediate cytotoxicity of decitabine (5-aza-dC), a deoxycytidine analogue used in the treatment of haematological malignancies.Decitabine is a nucleoside analogue, a class of therapy which are all prodrugs and thus require bioactivation inside cells.As a result, these therapies can have multiple active metabolites and typically poly-pharmacologic (and poorly resolved) modes of action.Understanding the molecular underpinnings of how these drugs are metabolised and subsequently exert their cytotoxic effects is critical to improving how these therapies are used in the clinic.Thus, I find the question posed in the study, and the findings presented, to be highly relevant and of great interest.
To identify factors controlling decitabine cytotoxicity, the authors employ a multi-omic approach, encompassing a series of whole genome CRISPR screens in a haploid cell line (HAP1), together with a new proteomic approach (based upon iPONDimmunoprecipitation of nascent DNA) which can define the proximal proteome of DNA protein crosslink (DPC) lesions, which are understood to be a component of how decitabine exerts cytotoxicity.By this way, the authors uncover an alternative metabolic route for decitabine and characterise factors involved in DPC-dependent and independent pathways, including the identification of dCMP deaminase (DCTD) and the E3 ligase TOPORS.Specific major concerns essential to be addressed to support the conclusions 1. Validation of findings in additional relevant cell models As written by the authors in the introduction, decitabine is clinically used in myelodysplastic syndromes (MDS), acute myeloid leukaemia (AML) and chronic myelogenous leukaemia (CML).The principal cell model used in the study is HAP1, which is derived from KBM-7, a CML line, however they are quite different in morphology and characteristics (for e.g., they are adherent).Other cell lines include the osteosarcoma line U2-OS and RPE-1 cells.I understand these are great models for the screen and downstream mechanistic experiments, but given the intent is to identify predictive biomarkers for decitabine treatment, additional more disease relevant models (e.g., AML or more CML lines) to suggest wider applicability of the findings (specifically controlling decitabine cytotoxicity) would be an important addition to the study -to validate these key findings more broadly.

Relevant study omitted from the manuscript
The first major finding in the study is the identification of DCTD and DCTPP1 (dCTPase) as decitabine resistance factors in the HAP1 cell line, this is shown in Figure 1 and discussed between lines 121 and 133.There is a study -Requena et al., PMID: 27325794, entitled "The nucleotidohydrolases DCTPP1 and dUTPase are involved in the cellular response to decitabine"which is very relevant to this section and the discussion of these results.This paper shows recombinant DCTPP1 can hydrolyse 5-aza-dCTP (thus, line 124 can be updated with this reference as this has been shown and so no need to speculate that DCTPP1 can catalyse this reaction).The study also suggests TYMS inhibition and uracil misincorporation can be important components of decitabine mode of action.So the findings in the present study -arrived at in an unbiased whole genome screen -further support this previously published model (see model in the above reference).This also calls into question Figure 1c, comparing existing model of decitabine action and additional model based on screen, as TYMS inhibition and DNA incorporation was already proposed in that study.Would revise.Here you are supporting a pre-existing model with genetic evidence in HAP1 cells.
3. Consideration of dNTP pool perturbations upon decitabine cytotoxicity.Loss of DCTD or DCTPP1 would be expected to lead to an increase in the dCTP pool.This is supported from pombe studies with DCTD (PMID: 22927644) and knockdown studies in human cells with DCTPP1 (PMID: 27325794) (perhaps there are also other examples out there).It's typically thought that active metabolites of nucleoside analogues compete with their endogenous counterparts (e.g., 5-aza-dCTP vs dCTP), and thus an elevation in the dCTP pool could also be expected to give rise to decitabine resistance.Can the authors integrate this into their current model and test the relevant contribution?dCTP is also a feedback inhibitor of DCK -the enzyme which activates decitabine -can the authors verify that in DCTD KO cells decitabine is being phosphorylated to similar levels?To exclude this possible explanation of resistance.

Verification of commercial KO lines absent
The absence of the target protein in the HAP1 DCK, DCTD, and TOPORS cell lines are not shown.
Minor concerns that should be addressed

Figure legends for western blots
In the figure legends it is not indicated how many experiments the Western blots are representative of.n=? 6.Comparison to DCK KO?When stating that DCTD KO causes "profound" decitabine resistance (line 127), it would be informative to compare degree of resistance against DCK KO (which is included in the study but not shown) for reference.As here very little to no decitabine should be activated.

Toning down language to reflect models used
The title says "central mediators" and numerous points in manuscript discuss results in "wild-type cells" (e.g., line 31, line 102) which reads with very broad implications, but those results are largely obtained in HAP1 cells.So I think either being more specific, or toning down, these sentences would be good.
Additional non-essential suggestions for improving the study (which will be at the author's/editor's discretion) In addition to the suggested experiments (comment #1), have the authors considered looking at publicly available pharmacogenomic datasets?Decitabine is included in the Broad (CTRP) and maybe other large-scale drug screening datasets (accessible via DepMap) -does decitabine cell killing in the panels of haematological cell lines included (100+) correlate with basal expression of factors identified in the screens?DCTD, TYMS, TOPORS etc.There are many reasons why a correlation wouldn't exist, but if it does, that could support a broader relevance of the findings beyond the cell models used.

Referee #3:
The authors conducted genetic screens to investigate the cytotoxic effects of 5-aza-dC.Their research revealed that the conversion of 5-aza-dC into 5-aza-dUMP, facilitated by the enzyme DCTD, is responsible for its cytotoxicity.Interestingly, the absence of DCTD can lead to resistance to the chemotherapy drug decitabine, despite these cells showing elevated levels of DPC lesions (DNMT1-DNA crosslinks) induced by decitabine, which is commonly used in the treatment of Myelodysplastic syndromes.
To further explore how elevated levels of DNMT1-DNA crosslinks are repaired, the authors conducted another genetic screening in DCTD-KO cells.Their findings indicate that DNMT1-DNA crosslinks are repaired by the SUMO1/ubiquitin E3 ligase TOPORS, which is responsible for degrading these DNMT1-DNA lesions and thereby promoting cell survival.
The findings of this study hold significant importance.Their investigation into the use of decitabine (5-aza-dC) for the treatment of haematological cancers has unveiled its capacity to induce changes in DNA.Notably, the unexpected discovery of DCTD's role in resistance adds a crucial layer to our understanding of how this drug operates.
These insights provide a deeper understanding of the mechanisms behind the cytotoxic effects of decitabine, highlighting the critical role of nucleotide metabolism in its effectiveness.I found this paper interesting and believe it's worth publishing in EMBO J, but it requires major revisions.
There are two major drawbacks in this manuscript.First, the most important and clinically relevant discovery of this manuscript, which is that 5-aza-dC toxicity is based on a deamination pathway regulated by DCDT, and not by DNMT1-DNA crosslinks and consequent hypomethylation, is superficially investigated.To support this statement, further experiments are needed.If their hypothesis is correct, then DCTD-KO cells should strongly depend on the BER pathway for sensitivity to 5-aza-dC.Therefore, the authors should re-sensitize HAP1-DCDT-KO cells to 5-aza-dC by inactivating the BER pathway or the glycosylase responsible for recognizing and repairing this lesion (e.g., UNG, XRCC1).
Second, the role of TOPORS in DNMT1-DNA proteolysis repair should be further supported.Based on the experiments performed, it's not clear whether the TOPORS repair model of DNMT1-protein lesion is post-replicative or replicative.A more detailed mechanistic insight into TOPORS is necessary.It's also unclear how this specific DPC (DNMT1) is finally degraded after being ubiquitinated/sumoylated by TOPORS.Is it through the proteasome or an SPRTN-dependent process?In summary, these findings have significant implications for both clinical applications and the understanding of DPC repair.However, the study needs significant improvement.
Major Comments: 1.The major discovery of the manuscript regarding 5-aza-dC cytotoxicity acting through its deamination pathway should be further supported.Consider experiments to demonstrate that DCTD-KO cells depend on the BER pathway for sensitivity to 5aza-dC.
2. In Figure 2b, the authors should analyze the cell cycle profiles of DCTD-KO and WT cells to determine whether the resistance of DCTD-KO cells is due to differences in cell cycle progression.E.g., the resistance of DCTD-KO cells could be due to suppressed cell cycle progression, and consequent arrest of these cells in the G1-phase.
3. In Figure 3, the authors discuss DNMT1-DPC repair mechanism in the context of the post-replicative pathway.However, they are using iPOND, a classical biochemical method to isolate factors associated with DNA replication forks.Moreover, cell synchronization and subsequent treatment clearly indicate the isolation of replication-related proteins.To clarify this, their iPOND experiments must be analyzed with well-recognized DNA replication factors such as PCNA and DNA polymerases delta or epsilon, and done in the presence of a thymidine chase to determine whether DNA replication fork moves in response to 5aza-dC treatment.In addition, they have identified several major components for replisome disassembly (p97-UBXDN7) and ICL-protein FANCA, all of which are also replication-related protein complexes.Altogether, this indicates that their model of DNMT1 repair is very likely replication-dependent.4. To clarify whether TOPORS is a replication-dependent or independent DPC-repair protein, the authors should conduct iPOND chase experiments and monitor TOPORS recruitment and progression at nascent and mature chromatin, with appropriate controls (PCNA and histone H3). 5.In Figure 4, use a Ub inhibitor in Fig. 4g to confirm their mass-spec data (from Fig 3f) and directly demonstrate that the TOPORS recruitment to DNMT1 PLA foci is only SUMO-dependent.To support their statement, they should also use the TOPORS-SIM-mutated variant to demonstrate whether TOPORS works as a STUbL in this pathway or not.
6.In addition, the survival assays in TOPORS-ko cells should be restored with TOPORS-wt and also investigated with TOPORS-variants defective for E3-ub ligase activity and SUMO binding.
7. In lines 158-169, the Rank plot demonstrating the loss of DPC repair genes in DCTD KO cells being sensitive to 5-aza-dC treatment and the increase in DNMT1-DPCs in response to 5-aza-dC cytotoxicity appears to be contradictory to the earlier statement (lines 151-156) where the authors discussed 5-aza-dC cytotoxicity in a DPC-independent manner.This inconsistency should be addressed or clarified.
8. In lines 473-485, in the relevant section, the authors describe the iPOND assay as a modified version, but upon closer examination, it seems that they have essentially followed the intact protocol from the referenced paper.Their method is simply iPOND and not POND-DPC.9. Additionally, aside from presenting the data regarding SUMOi with and without 5-aza-dC as demonstrated in Figure 4j, it is advisable to incorporate experiments involving mutated SUMO binding motif of TOPORS.This would help determine if TOPORS is indeed recruited to the DNMT1-DPCs sites as a STUbL.10.Considering the possibility of DNMT1-DPC degradation by the proteasome system, it would be intriguing to observe the accumulation of these DPCs in the presence of a proteasome inhibitor.
Minor comments 1.In lines 55-60, the authors effectively address the observed treatment variability in patients, highlighting the significance of low-dose drug exposure in resistance and relapse.Their insightful analysis aligns with a recurring theme in disease management and treatment strategies seen in the broader literature.As seen with SN-38 treatment, a drug causing DNA damage and forming TOP1cc, there is a significant convergence in anticancer drug therapy outcomes (PMID: 35869071, PMID: 37240063).Notably, these treatments also result in unsuccessful repair processes and genetic instability, contributing to the discussion.I recommend that the authors incorporate references to these papers (PMID: 35869071, PMID: 37240063), as they provide possible explanations for patients' responses to these drugs.This inclusion will enhance the comprehensiveness and relevance of the study.
2. In Figure 4 a-f, Supplementary figure 4 e-g: It would be beneficial for the manuscript if the authors could consider including representative images of clonogenic survival assays in the supplementary section.

Referee #1:
In this manuscript, Carnie et al. investigate the determinants of cytotoxicity associated with 5-aza-dC, also known as decitabine, through CRISPR/Cas9 screens.Decitabine is used in the treatment of certain blood disorders, such as MDS and AML.The therapeutic mechanism of 5-aza-dC is believed to be related to its ability to trap DNA methyltransferases (DNMTs) onto DNA, resulting in DNA hypomethylation as well as the formation of DNAprotein crosslinks (DPCs) containing DNMT1.However, the exact mechanisms underlying its action and patient responses remain unclear.
This study used CRISPR/Cas9 screens to identify genes affecting sensitivity or resistance to 5-aza-dC.It revealed that, in DPC-repair-proficient cells, 5-aza-dC's cytotoxicity results from the conversion of 5-aza-dC to 5-aza-dUMP by DCTD, suggesting that the origin of cytotoxicity is independent of DNA hypomethylation and DNMT1-DPCs.
To gain more insights, the authors performed another CRISPR/Cas9 screen, this time in DCTD KO cells, and found that DPC repair pathways become more prominent determinants of 5-aza-dC sensitivity under this condition.Using a modified iPOND technique (called iPOND-DPC) to identify the proximal proteome of DNMT1-DPCs, the authors identified, among other known DPC repair factors, TOPORS (a SUMO1-and ubiquitin E3 ligase) as a factor that confers protection against DNMT1-DPCs.They propose that TOPORS is recruited to decitabine-induced DNMT1-DPCs in a SUMO-dependent but ubiquitinindependent manner, facilitating the proteolysis of DNMT1-DPCs in parallel to RNF4 and UBE2K.
This is an important study that challenges current assumptions regarding decitabine's mechanisms of action.This study also reveals a novel DPC repair mechanism mediated by TOPORS, highlighting the critical role of SUMO modifications in DPC repair.
The presented data are of high quality.The proposed model is supported both by genetic evidence on 5-aza-dC sensitivities and by cellular data on TOPORS recruitment to SUMOylated DNMT1.The study could be enhanced by further investigations into the mechanism underlying the SUMO-dependency of TOPORS recruitment.Additionally, whether TOPORS' ubiquitin E3 ligase activity is required for the repair of DNMT1-DPCs is an important question to address to substantiate their model that TOPORS polyubiquitinates DNMT1-DPCs for proteasomal degradation.Comments are listed below for the authors to consider during the revision.
We thank the reviewer for their supportive comments and helpful feedback.
Major points: 1.Does DCTD KO result in increased DPC burden during 5-aza-dC treatment?Fig. 2b appears to show that the DNMT1-DPC level is higher in DCTD KO cells compared to WT, suggesting that DCTD might funnel the 5-aza-dC metabolite away from DNMT-DPC formation in WT cells.However, the authors do not seem to acknowledge this, given that they state "DNMT1-DPC induction in DCTD KO cells was similar to that in WT cells, ( 2. It would be helpful to highlight the ranking of DNMT1 in Fig. 1b and Fig. 2c.In Fig. 1b, DNMT1 KO would not confer resistance as cytotoxicity in WT cells is independent of DNMT1.In Fig. 2c, on the other hand, DNMT1 KO would be expected to provide some resistance to 5-aza-dC if DCTD KO results in increased DNMT1-DPCs (as mentioned in Point #1).
We have added DNMT1 to these plots as requested (new Fig 1B and new Fig 2C).DNMT1 does not score as a hit in any of these screens, consistent with our observation described above that siRNA-mediated DNMT1 depletion does not confer 5-aza-dC resistance to either WT or DCTD KO cells.3. How TOPORS promotes DPC tolerance is not addressed sufficiently.Are the RING domain and SIMs of TOPORS important for DNMT1-DPC repair in Fig. 5g?Rescue experiments with WT or mutant TOPORS would help address this question.Additionally, are the TOPORS SIMs important for the DNMT1 interaction in Fig. 4i,j?As requested, we have performed structure-function analysis of TOPORS in clonogenic survival assays, immunoprecipitation and PLA experiments.Indeed, inducible re-expression of TOPORS WT rescued the 5-aza-dC sensitivity of TOPORS KO cells, but this was not the case for expression of TOPORS' RING domain mutants (TOPORS C103A C1036A; TOPORS CCAA ) or TOPORS mutated in the 6 candidate SUMO-interaction motifs (SIMs; TOPORS ΔSIM ; new Fig 5B-C, new Fig EV4C -E).Furthermore, we established that ubiquitin ligase-defective TOPORS CCAA retains its ability to interact with high molecular weight SUMOylated proteins and DNMT1 in a 5-aza-dC-inducible manner (new Fig 5D ), while TOPORS ΔSIM does not (new Fig 5F), suggesting that TOPORS is recruited through its SIMs to SUMOylated DNMT1-DPCs where its ubiquitin ligase activity is important for polyubiquitylation and subsequent degradation of the DNMT1-DPCs.Unfortunately it was extremely challenging to establish TOPORS KO cells re-expressing TOPORS, so we have not been able to extend these studies to PxP assays due to time constraints.However, we feel that the clonogenic survival assays with RING-mutant TOPORS alongside the existing understanding of DNMT1-DPC repair and effect of TOPORS depletion on DNMT1-DPC degradation strongly support our model that TOPORS polyubiquitylates DNMT1-DPCs.  5. Line 251: "While we detected an interaction between WT GFP-TOPORS and a high molecular weight ubiquitylated interactor, this was not overtly 5-aza-dC-inducible but undetectable in co-immunoprecipitates of GFP-TOPORS CCAA (Fig 4i)."Could this be TOPORS auto-ubiquitination? TOPORS pull-down under denaturing condition could clarify this possibility.The same experiment could also be used to test if the SUMO conjugates in the pull-down is reduced under a denaturing condition.That would support authors' interpretation that the SUMO chains are conjugated to interacting proteins (presumably DNMT1) but not to TOPORS itself.
As suggested by the reviewer, we have performed stringent GFP-TOPORS WT and GFP-TOPORS CCAA pulldowns in 1M NaCl and it does indeed appear that the ubiquitylated factor is TOPORS itself, with the ubiquitylation signal not appearing in stringent IPs of GFP-TOPORS CCAA (new Fig 5E).A single band of similar size to unmodified GFP-TOPORS does remain, possibly suggesting some residual TOPORS-independent ubiquitylation.Interestingly, we were also able to detect conjugates of both SUMO2/3 and SUMO1 in these experiments, but these were not appreciably 5-aza-dC-inducible and might therefore account for some of the background SUMO signal seen in dC-treated IP samples in other experiments (such as in new Fig 5D, shown above).These SUMO signals were increased in stringent immunoprecipitates from GFP-TOPORS CCAA , but the reason for this remains unclear.Overall, this experiment suggests that TOPORS is endogenously SUMOylated and autoubiquitylated, but that the 5-aza-dC-inducible portion of heavily SUMOylated interactors likely represents DNMT1.Unfortunately, despite testing a wide range of commercially available TOPORS antibodies, we have been unable to identify one that we can independently validate.
We have rectified this omission and added this explanation.

Referee #2:
General summary & opinion of study -significance, questions, and findings The study by Carnie et al seeks to identify factors that mediate cytotoxicity of decitabine (5aza-dC), a deoxycytidine analogue used in the treatment of haematological malignancies.Decitabine is a nucleoside analogue, a class of therapy which are all prodrugs and thus require bioactivation inside cells.As a result, these therapies can have multiple active metabolites and typically poly-pharmacologic (and poorly resolved) modes of action.Understanding the molecular underpinnings of how these drugs are metabolised and subsequently exert their cytotoxic effects is critical to improving how these therapies are used in the clinic.Thus, I find the question posed in the study, and the findings presented, to be highly relevant and of great interest.
To identify factors controlling decitabine cytotoxicity, the authors employ a multi-omic approach, encompassing a series of whole genome CRISPR screens in a haploid cell line (HAP1), together with a new proteomic approach (based upon iPOND -immunoprecipitation of nascent DNA) which can define the proximal proteome of DNA protein crosslink (DPC) lesions, which are understood to be a component of how decitabine exerts cytotoxicity.By this way, the authors uncover an alternative metabolic route for decitabine and characterise factors involved in DPC-dependent and independent pathways, including the identification of dCMP deaminase (DCTD) and the E3 ligase TOPORS.
We thank the reviewer for their supportive comments and feedback.Specific major concerns essential to be addressed to support the conclusions 1. Validation of findings in additional relevant cell models As written by the authors in the introduction, decitabine is clinically used in myelodysplastic syndromes (MDS), acute myeloid leukaemia (AML) and chronic myelogenous leukaemia (CML).The principal cell model used in the study is HAP1, which is derived from KBM-7, a CML line, however they are quite different in morphology and characteristics (for e.g., they are adherent).Other cell lines include the osteosarcoma line U2-OS and RPE-1 cells.I understand these are great models for the screen and downstream mechanistic experiments, but given the intent is to identify predictive biomarkers for decitabine treatment, additional more disease relevant models (e.g., AML or more CML lines) to suggest wider applicability of the findings (specifically controlling decitabine cytotoxicity) would be an important addition to the study -to validate these key findings more broadly.
We agree with the reviewer that expanding these observations into more clinically-relevant cell lines will be needed to allow extension of our findings to the search for meaningful clinical biomarkers.However, despite efforts to address this point we have been unable to establish efficient gene editing of relevant cell lines as the use of these cell lines is not yet established in our labs.
We have become aware that TOPORS loss has been reported by other groups in conference abstracts to sensitise AML cell lines to 5-aza-dC and we have now cited these abstracts in the manuscript (Truong et al, Blood 2022; Kaito et al, Blood 2023).Having contacted these groups, we have learned that their work is much more AML-orientated and also at an advanced stage with another journal.We do not want to simply repeat other people's work, so have instead focussed on strengthening the mechanistic understanding of our phenotypic observations.To reflect our focus on mechanistic studies, we have toned down our claims relating to AML/MDS treatment.

Relevant study omitted from the manuscript
The first major finding in the study is the identification of DCTD and DCTPP1 (dCTPase) as decitabine resistance factors in the HAP1 cell line, this is shown in Figure 1 and discussed between lines 121 and 133.There is a study -Requena et al., PMID: 27325794, entitled "The nucleotidohydrolases DCTPP1 and dUTPase are involved in the cellular response to decitabine" -which is very relevant to this section and the discussion of these results.This paper shows recombinant DCTPP1 can hydrolyse 5-aza-dCTP (thus, line 124 can be updated with this reference as this has been shown and so no need to speculate that DCTPP1 can catalyse this reaction).The study also suggests TYMS inhibition and uracil misincorporation can be important components of decitabine mode of action.So the findings in the present study -arrived at in an unbiased whole genome screen -further support this previously published model (see model in the above reference).This also calls into question Figure 1c, comparing existing model of decitabine action and additional model based on screen, as TYMS inhibition and DNA incorporation was already proposed in that study.Would revise.Here you are supporting a pre-existing model with genetic evidence in HAP1 cells.
We thank the reviewer for bringing this relevant and complementary study to our attention.As suggested, we have updated our text, figures and discussion accordingly.

Consideration of dNTP pool perturbations upon decitabine cytotoxicity.
Loss of DCTD or DCTPP1 would be expected to lead to an increase in the dCTP pool.This is supported from pombe studies with DCTD (PMID: 22927644) and knockdown studies in human cells with DCTPP1 (PMID: 27325794) (perhaps there are also other examples out there).It's typically thought that active metabolites of nucleoside analogues compete with their endogenous counterparts (e.g., 5-aza-dCTP vs dCTP), and thus an elevation in the dCTP pool could also be expected to give rise to decitabine resistance.Can the authors integrate this into their current model and test the relevant contribution?dCTP is also a feedback inhibitor of DCK -the enzyme which activates decitabine -can the authors verify that in DCTD KO cells decitabine is being phosphorylated to similar levels?To exclude this possible explanation of resistance.
We agree that assessing the impact of DCTD loss on nucleotide metabolism as a whole will likely be important for understanding exactly how DCTD loss confers 5-aza-dC resistance.However, we have been unable to establish this within the timeframe allocated for revisions.Given the competitive nature of this work, we feel that a full mechanistic explanation of DCTD-mediated 5-aza-dC toxicity is beyond the scope of this manuscript.
Importantly, however, elevated dCTP pools or defects in 5-aza-dC phosphorylation (that might be expected in DCTD KOs) would be anticipated to cause 5-aza-dC resistance by causing a dramatic reduction in DNMT1-DPC induction.However, our PxP assay in new Fig

Verification of commercial KO lines absent
The absence of the target protein in the HAP1 DCK, DCTD, and TOPORS cell lines are not shown.
Unfortunately despite testing multiple commercially-available TOPORS antibodies, we are yet to find one that specifically targets TOPORS.With respect to DCTD and DCK, we have now added western blots confirming these knockouts (new Fig EV1A-B).Minor concerns that should be addressed

Figure legends for western blots
In the figure legends it is not indicated how many experiments the Western blots are representative of.n=?These details have been added throughout.

Comparison to DCK KO?
When stating that DCTD KO causes "profound" decitabine resistance (line 127), it would be informative to compare degree of resistance against DCK KO (which is included in the study but not shown) for reference.As here very little to no decitabine should be activated.
We agree that this is an informative comparison and have conducted the recommended clonogenic survival assays, which show that, as expected, DCK KO cells are more resistant to 5-aza-dC than DCTD KOs (new Fig 1D-E).

Toning down language to reflect models used
The title says "central mediators" and numerous points in manuscript discuss results in "wildtype cells" (e.g., line 31, line 102) which reads with very broad implications, but those results are largely obtained in HAP1 cells.So I think either being more specific, or toning down, these sentences would be good.
We have adjusted the wording in these places as suggested, including in the title which now reads "The dCMP deaminase DCTD and the SUMO-dependent E3 ligase TOPORS mediate decitabine cytotoxicity".
Additional non-essential suggestions for improving the study (which will be at the author's/editor's discretion) In addition to the suggested experiments (comment #1), have the authors considered looking at publicly available pharmacogenomic datasets?Decitabine is included in the Broad (CTRP) and maybe other large-scale drug screening datasets (accessible via DepMap)does decitabine cell killing in the panels of haematological cell lines included (100+) correlate with basal expression of factors identified in the screens?DCTD, TYMS, TOPORS etc.There are many reasons why a correlation wouldn't exist, but if it does, that could support a broader relevance of the findings beyond the cell models used.
We have checked through public data accessible via DepMap and cBioPortal but have not identified overt correlations for DCTD, TOPORS or TYMS.

Referee #3:
The authors conducted genetic screens to investigate the cytotoxic effects of 5-aza-dC.Their research revealed that the conversion of 5-aza-dC into 5-aza-dUMP, facilitated by the enzyme DCTD, is responsible for its cytotoxicity.Interestingly, the absence of DCTD can lead to resistance to the chemotherapy drug decitabine, despite these cells showing elevated levels of DPC lesions (DNMT1-DNA crosslinks) induced by decitabine, which is commonly used in the treatment of Myelodysplastic syndromes.
To further explore how elevated levels of DNMT1-DNA crosslinks are repaired, the authors conducted another genetic screening in DCTD-KO cells.Their findings indicate that DNMT1-DNA crosslinks are repaired by the SUMO1/ubiquitin E3 ligase TOPORS, which is responsible for degrading these DNMT1-DNA lesions and thereby promoting cell survival.
The findings of this study hold significant importance.Their investigation into the use of decitabine (5-aza-dC) for the treatment of haematological cancers has unveiled its capacity to induce changes in DNA.Notably, the unexpected discovery of DCTD's role in resistance adds a crucial layer to our understanding of how this drug operates.
These insights provide a deeper understanding of the mechanisms behind the cytotoxic effects of decitabine, highlighting the critical role of nucleotide metabolism in its effectiveness.I found this paper interesting and believe it's worth publishing in EMBO J, but it requires major revisions.
There are two major drawbacks in this manuscript.First, the most important and clinically relevant discovery of this manuscript, which is that 5-aza-dC toxicity is based on a deamination pathway regulated by DCDT, and not by DNMT1-DNA crosslinks and consequent hypomethylation, is superficially investigated.To support this statement, further experiments are needed.If their hypothesis is correct, then DCTD-KO cells should strongly depend on the BER pathway for sensitivity to 5-aza-dC.Therefore, the authors should resensitize HAP1-DCDT-KO cells to 5-aza-dC by inactivating the BER pathway or the glycosylase responsible for recognizing and repairing this lesion (e.g., UNG, XRCC1).
Second, the role of TOPORS in DNMT1-DNA proteolysis repair should be further supported.Based on the experiments performed, it's not clear whether the TOPORS repair model of DNMT1-protein lesion is post-replicative or replicative.A more detailed mechanistic insight into TOPORS is necessary.It's also unclear how this specific DPC (DNMT1) is finally degraded after being ubiquitinated/sumoylated by TOPORS.Is it through the proteasome or an SPRTN-dependent process?In summary, these findings have significant implications for both clinical applications and the understanding of DPC repair.However, the study needs significant improvement.
We thank the reviewer for their supportive comments and feedback, which have helped to improve our study.
Major Comments: 1.The major discovery of the manuscript regarding 5-aza-dC cytotoxicity acting through its deamination pathway should be further supported.Consider experiments to demonstrate that DCTD-KO cells depend on the BER pathway for sensitivity to 5-aza-dC.
We agree with the reviewer that understanding the contribution of BER to 5-aza-dC cytotoxicity is of importance.However, it is a complex question to pick apart, because the BER machinery could feasibly act on genome-embedded 5-aza-dC, 5-aza-dU and dU (dU incorporation upon 5-aza-dC treatment has been noted previously: PMID: 27325794).BER could therefore in principle contribute to 5-aza-dC cytotoxicity in both WT and in DCTD KO cells.Indeed, in our 5-aza-dC screens, XRCC1, a central mediator of SSB repair, appears as a sensitivity hit in both WT and DCTD KO cells, supporting the notion that BER and SSB repair are relevant for more than one aspect of decitabine cytotoxicity -we have updated our figures to highlight this hit in our screens (new Fig 1B,new Fig 2D; also see above in our response to reviewer 1).Furthermore, in polyclonal pools of WT and DCTD KO cells edited via CRISPR/Cas9 to inactivate PARP1, we have found that both WT and DCTD KO cells are sensitised to 5-aza-dC by PARP1 depletion (new Fig EV1H -J).No glycosylases scored as hits in any of our CRISPR screens, and by using genetic and cell biology approaches we have been unable to identify the glycosylase(s) driving BER upon 5-aza-dC treatment, possibly because of redundant effects of different glycosylases on different base lesions (5aza-dC, 5-aza-dU, dU).We hope to address this question in follow-up work using reconstituted biochemical systems but feel that this lies beyond the scope of the present study.
We have updated our text to incorporate the data mentioned above and to discuss the role of BER in the cellular response to 5-aza-dC more thoroughly.2. In Figure 2b, the authors should analyze the cell cycle profiles of DCTD-KO and WT cells to determine whether the resistance of DCTD-KO cells is due to differences in cell cycle progression.E.g., the resistance of DCTD-KO cells could be due to suppressed cell cycle progression, and consequent arrest of these cells in the G1-phase.
As requested, we have performed cell cycle analysis of WT and DCTD KO cells, also combining EdU incorporation with 5-aza-dC treatment (new Fig EV1G).The percentage of cells incorporating EdU within the 30-minute pulse was not substantially affected by DCTD loss.Therefore, the resistance of DCTD KO cells to 5-aza-dC is not caused by reduced DNA synthesis and 5-aza-dC incorporation.3.In Figure 3, the authors discuss DNMT1-DPC repair mechanism in the context of the postreplicative pathway.However, they are using iPOND, a classical biochemical method to isolate factors associated with DNA replication forks.Moreover, cell synchronization and subsequent treatment clearly indicate the isolation of replication-related proteins.To clarify this, their iPOND experiments must be analyzed with well-recognized DNA replication factors such as PCNA and DNA polymerases delta or epsilon, and done in the presence of a thymidine chase to determine whether DNA replication fork moves in response to 5-aza-dC treatment.In addition, they have identified several major components for replisome disassembly (p97-UBXDN7) and ICL-protein FANCA, all of which are also replication-related protein complexes.Altogether, this indicates that their model of DNMT1 repair is very likely replication-dependent.
We were probably not clear enough in our introduction as to how 5-aza-dC induces DNMT1-DPCs.These DPCs occur behind the replication fork but in a replication-coupled manner, with DNMT1 and its cofactor UHRF1 interacting with PCNA to methylate newly-synthesised DNA.As such, DNMT1-DPC repair is replication-independent (see Borgermann et al 2019, EMBO J and Weickert et al 2023, Nat Comms), despite occurring in the vicinity of replisomes.Using iPOND, in combination with 5-azadC labelling, we can isolate postreplicative DNA containing DNMT1-DPCs.However, it is notable that several replicationassociated proteins appear as sensitivity hits in our CRISPR screens, which we speculate could arise from remnants of degraded DNMT1-DPCs perturbing replisomes in the next Sphase following initial DPC induction.We have revised our text to make the basis of DNMT1-DPC induction and repair and its relationship with replication clearer.4. To clarify whether TOPORS is a replication-dependent or independent DPC-repair protein, the authors should conduct iPOND chase experiments and monitor TOPORS recruitment and progression at nascent and mature chromatin, with appropriate controls (PCNA and histone H3).
As stated above, DNMT1-DPC repair has already been shown to be replication-independent.Therefore, we feel that attempting the proposed experiment would not improve the study.Furthermore, we are unfortunately not able to detect TOPORS via iPOND-western blots due to the lack of suitable reagents.5.In Figure 4, use a Ub inhibitor in Fig. 4g to confirm their mass-spec data (from Fig 3f) and directly demonstrate that the TOPORS recruitment to DNMT1 PLA foci is only SUMOdependent.To support their statement, they should also use the TOPORS-SIM-mutated variant to demonstrate whether TOPORS works as a STUbL in this pathway or not.
We have performed the PLA experiments suggested by the reviewer.In agreement with our iPOND data, in contrast to SUMO E1i, Ub E1i did not compromise TOPORS recruitment to DNMT1-DPCs as assessed by PLA (new Fig 4G -H).HA-TOPORS ΔSIM was expressed less at the RNA level than HA-TOPORS WT (new Fig EV4E) but had higher PLA signal in dC-treated cells than HA-TOPORS WT (new Fig EV4F-G).Despite this, the PLA signal induction of HA-TOPORS ΔSIM with GFP-DNMT1 upon 5-aza-dC treatment was much less dramatic than that between HA-TOPORS WT and GFP-DNMT1 (new Fig EV4F-G).To support these observations, we used co-immunoprecipitation of GFP-TOPORS WT and GFP-TOPORS ΔSIM and found that GFP-TOPORS ΔSIM is deficient in its ability to interact with high molecular weight SUMO conjugates and heavily modified (presumably SUMOylated) DNMT1 after 5aza-dC treatment (new Fig 5F).Together, these experiments strengthen the interpretation that TOPORS acts in a SUMO-dependent manner.6.In addition, the survival assays in TOPORS-ko cells should be restored with TOPORS-wt and also investigated with TOPORS-variants defective for E3-ub ligase activity and SUMO binding.
We are pleased to now include data demonstrating that doxycycline-inducible expression of WT TOPORS, but not E3 ligase-defective (TOPORS CCAA ) or SUMO-binding-defective mutants (TOPORS ΔSIM ) restores 5-aza-dC tolerance of TOPORS KO cells (new Fig 5B-C; new Fig EV4C-D; also shown in this letter in response to Reviewer 1's Major Point 3 above).
7. In lines 158-169, the Rank plot demonstrating the loss of DPC repair genes in DCTD KO cells being sensitive to 5-aza-dC treatment and the increase in DNMT1-DPCs in response to 5-aza-dC cytotoxicity appears to be contradictory to the earlier statement (lines 151-156) where the authors discussed 5-aza-dC cytotoxicity in a DPC-independent manner.This inconsistency should be addressed or clarified.
We recognise that we did not explain our reasoning with sufficient clarity.Having found that DCTD KO cells do not experience substantially fewer DNMT1-DPCs than WT cells upon 5aza-dC, we concluded that DCTD was driving a mechanism of toxicity independent of DNMT1-DPCs, likely through its deamination of 5-aza-dCMP to 5-aza-dUMP.Consistently, depletion of DNMT1 did not affect 5-aza-dC sensitivity of WT or DCTD KO cells (new Fig EV1D-F, and see above in response to Reviewer 1's Major Point 1).However, if DPC repair factors are lost (e.g. by TOPORS KO), 5-aza-dC-induced DNMT1-DPCs cannot be repaired and thus contribute to toxicity.This is even more apparent in DCTD KO cells, in which the balance between DPC-dependent and DPC-independent cytotoxicity is shifted, resulting in stronger phenotypes upon loss of DPC repair factors upon 5-aza-dC-treatment than in WT cells.
We have updated the text to make this line of reasoning clearer.
8. In lines 473-485, in the relevant section, the authors describe the iPOND assay as a modified version, but upon closer examination, it seems that they have essentially followed the intact protocol from the referenced paper.Their method is simply iPOND and not POND-DPC.
We combined iPOND with 5-azadC co-labelling to identify the proximal proteome of DNMT1-DPCs.Following the reviewer's suggestion, we have rephrased our discussion and now simply refer to these assays as iPOND.9. Additionally, aside from presenting the data regarding SUMOi with and without 5-aza-dC as demonstrated in Figure 4j, it is advisable to incorporate experiments involving mutated SUMO binding motif of TOPORS.This would help determine if TOPORS is indeed recruited to the DNMT1-DPCs sites as a STUbL.
As suggested, we have mutated the six TOPORS SIMs and performed GFP-TOPORS ΔSIM immunoprecipitations (new Fig 5E; also shown in this letter in response to Reviewer 1, Major Point 5 above).Consistent with TOPORS acting in a SUMO-dependent manner, GFP-TOPORS ΔSIM retained some interaction with unmodified DNMT1 but failed to interact with high molecular weight SUMO conjugates or heavily-modified (presumably SUMOylated) DNMT1.
10. Considering the possibility of DNMT1-DPC degradation by the proteasome system, it would be intriguing to observe the accumulation of these DPCs in the presence of a proteasome inhibitor.Indeed, DNMT1-DPC repair has already been shown to be entirely dependent on both SUMOylation and proteolysis by the proteasome and SPRTN [see Borgermann et al 2019, Liu et al 2021and Weickert et al 2023].We have revised our text to make this point clearer.
Please note that it is The EMBO Journal policy for the transcript of the editorial process (containing referee reports and your response letter) to be published as an online supplement to each paper.If you do NOT want this, you will need to inform the Editorial Office via email immediately.More information is available here: https://www.embopress.org/transparentprocess#Review_Process

EMBO Press Author Checklist USEFUL LINKS FOR COMPLETING THIS FORM
The EMBO Journal -Author Guidelines EMBO Reports -Author Guidelines Molecular Systems Biology -Author Guidelines EMBO Molecular Medicine -Author Guidelines Please note that a copy of this checklist will be published alongside your article.

Abridged guidelines for figures 1. Data
The data shown in figures should satisfy the following conditions:

Reporting
Adherence to community standards Information included in the manuscript?
In which section is the information available?
(Reagents and Tools Have primary datasets been deposited according to the journal's guidelines (see 'Data Deposition' section) and the respective accession numbers provided in the Data Availability Section?

Data availability statement
Were human clinical and genomic datasets deposited in a public access-controlled repository in accordance to ethical obligations to the patients and to the applicable consent agreement?

Not Applicable
Are computational models that are central and integral to a study available without restrictions in a machine-readable form?Were the relevant accession numbers or links provided?

Not Applicable
If publicly available data were reused, provide the respective data citations in the reference list.

Not Applicable
The MDAR framework recommends adoption of discipline-specific guidelines, established and endorsed through community initiatives.Journals have their own policy about requiring specific guidelines and recommendations to complement MDAR.
Fig 2b)" (Line 152).The observed increase in DNMT1-DPCs in DCTD KO cells was not apparent in all of our replicates.Overall, DCTD KO cells seem to experience slightly less DNMT1-DPC induction than their WT counterparts -we have replaced the original figure with one more representative (new Fig 2B) and also show a further replicate below.However, this modest reduction in DPC formation is unlikely to account for the dramatic resistance phenotype 15th Mar 2024 1st Authors' Response to Reviewers given that siRNA-mediated depletion of DNMT1 did not confer 5-aza-dC resistance in either WT or DCTD KO cells (new Fig EV1D-F).This is in contrast to TOPORS KO cells, in which siDNMT1 fully restores 5-aza-dC sensitivity to WT levels (new Fig 4E-F), indicating that 5aza-dC-induced DNMT1-DPCs contribute to toxicity in cells with defective DPC repair, but not in WT or DCTD KO cells.New Fig 2. (B) DNMT1-DPC formation assessed by PxP in WT, DCTD KO and DCK KO HAP1 cells treated with 5-aza-dC (1μM) for the indicated times; representative of 3 independent experiments.New Fig EV1.(E) Clonogenic survival assays with siRNA-transfected WT and DCTD KO cells from treated with 5-aza-dC; n = 3 biological replicates, error bars ± SEM.New Fig 4. (E) Clonogenic survival assays in WT and TOPORS KO HAP1 cells transfected with indicated siRNAs and treated with 5-aza-dC; n = 3 biological replicates, error bars ± SEM.
New Fig 5. (B-C) Clonogenic survival assays in TOPORS KO cell lines with (B) or without (C) doxycyclineinduced expression of the indicated forms of TOPORS, treated with 5-aza-dC; n = 4 biological replicates, error bars ± SEM. (D) Co-immunoprecipitation of GFP from extracts of HeLa cells expressing GFP (EV), GFP-TOPORS WT or GFP-TOPORS CCAA , released from thymidine block into S-phase and treated with dC or 5-aza-dC for 1 hour, followed by western blotting for indicated proteins; representative of 3 independent experiments.(F) Co-immunoprecipitation and western blotting as in (D) but with cells expressing GFP (EV), GFP-TOPORS WT or GFP-TOPORS ΔSIM ; representative of 3 independent experiments.
4. Does TOPORS contribute to SUMOylation of DNMT1-DPCs?Mailand's group has reported that there are SUMO E3 ligases for DNMT1-DPCs redundant to PIAS4, but TOPORS was not included in their study(Ref #37).This is a pertinent question, given the reported SUMO1 ligase activity of TOPORS.In PxP assays, we observe increased SUMOylation (most dramatically SUMO1 modification) of DNMT1-DPCs in cells depleted of bothRNF4 and TOPORS (new Fig 6G), a phenomenon also observable in stringent GFP-DNMT1 IPs in cells depleted of just TOPORS(new Fig  6H).These findings suggest that rather than promoting DNMT1-DPC SUMOylation, TOPORS acts downstream of DNMT-DPC SUMOylation, with TOPORS depletion impairing proteasomal degradation of DNMT1-DPCs and thus stabilising SUMOylated DNMT1-DPC intermediates.Combined with our structure-function analyses outlined above, this strengthens our argument that TOPORS is recruited to SUMOylated DNMT1-DPCs and promotes their polyubiquitylation and degradation.New Fig 6. (G) HeLa WT or RNF4 KO were transfected with the indicated siRNAs, synchronised by double thymidine block and treated with 5-aza-dC (10μM).DNMT1-DPCs were isolated using PxP and analysed by western blotting using the indicated antibodies; representative of 3 independent experiments.Asterisks (*) indicate SPRTN dependent DNMT1 cleavage fragment.(H) Stringent immunoprecipitation of GFP-DNMT1 and subsequent western blotting from U2OS cells constitutively expressing GFP-DNMT1, transfected with siCtrl or siTOPORS and treated with dC or 5-aza-dC as indicated after release from a thymidine block into S-phase; representative of 3 independent experiments.
New Fig. 5. (E) Western blots following co-immunoprecipitation of the indicated GFP expression constructs from HeLa cells under stringent, denaturing conditions after 5-aza-dC treatment in S-phase-synchronized cells; representative of 3 independent experiments.Minor points: 1. Fig EV. 3a,b: Is there any TOPORS antibody that can validate the lack of TOPORS proteins in TOPORS KO and siTOPORS cells?
2B (see our response to reviewer 1 above) demonstrates at most only a mild reduction in DNMT1-DPCs that is unlikely to account for the dramatic cellular resistance phenotypes.Indeed, DNMT1 depletion by siRNA did not protect WT or DCTD KO cells from 5-aza-dC, and only conferred resistance to DPC-repair-deficient TOPORS KO cells (see new Fig EV1D-F and Fig 4E-F; see our response to reviewer 1 above).
New Fig EV1.(A-B) Western blot in WT and DCTD KO (A) or DCK KO (B) HAP1 cells with the indicated antibodies; representative of 3 (A) and 2 (B) independent experiments.The DCTD-specific band in (A) is marked with a red asterisk (*).

New
Fig 1. (D) Clonogenic survival assays in WT and DCTD KO HAP1 cells treated with 5-aza-dC and stained 6 days later; n = 3 biological replicates, error bars ± SEM. (E) Representative images from (D) of cells at selected 5-aza-dC doses.
New Fig EV1.(H) Western blot with the indicated antibodies of polyclonal cell populations of WT and DCTD KO cells following CRISPR/Cas9-mediated depletion of PARP1 with the indicated sgRNAs; representative of 2 independent experiments.(I) Clonogenic survival assays on cells from (H) treated with 5-aza-dC; n = 3 biological replicates, error bars ± SEM. (J) Representative images from (I).
New Fig EV1.(G) Percentage of EdU-positive cells determined by flow cytometry following 5-aza-dC treatment in the indicated cell lines; n = 3 biological replicates, error bars ± SD.

New
Fig 4. (G) Proximity ligation assay in U2OS cells expressing GFP-DNMT1 and HA-TOPORS, released from a single thymidine block and treated with deoxycytidine (dC), 5-aza-dC and/or Ub E1i or SUMO E1i for 1 hour before pre-extraction of non-chromatin-bound proteins and fixation.(H) Quantification of per-nucleus mean PLA intensities from (G), normalised to the median from the 5-aza-dC-treated U2OS GFP-DNMT1/HA-TOPORS condition.Black dots display the median normalised PLA intensity of each biological replicate for each condition; n = 4 independent biological replicates error bars ± SEM.New Fig EV4.(E) Expression levels of TOPORS in U2OS GFP-DNMT1 cells measured by qPCR relative to GAPDH and normalised to U2OS GFP-DNMT1 cells expressing HA-EV; n = 2 biological replicates performed in technical triplicate.(F-G) Representative images (F) and quantification (G) of PLA in U2OS GFP-DNMT1 cells expressing HA-EV, HA-TOPORS WT or HA-TOPORS ΔSIM treated with dC or 5-aza-dC.New Fig 5. (F) Co-immunoprecipitation and western blotting with HeLa cells expressing GFP (EV), GFP-TOPORS WT or GFP-TOPORS ΔSIM treated with dC or 5-aza-dC after release from a thymidine block into S-phase; representative of 3 independent experiments.

In which section is the information available?
definitions of statistical methods and measures: (Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)

In which section is the information available?
(Reagents and ToolsTable, Materials and Methods, Figures, Data Availability Section) Short novel DNA or RNA including primers, probes: provide the sequences.Yes Reagents table, referred to in Methods

section Cell materials Information included in the manuscript? In which section is the information available?
(Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)Cell lines: Provide species information, strain.Provide accession number in repository OR supplier name, catalog number, clone number, and/OR RRID.

In which section is the information available?
Provide species, strain, sex, age, genetic modification status.Provide accession number in repository OR supplier name, catalog number, clone number, OR RRID.
(Reagents and ToolsTable, Materials and Methods, Figures, Data Availability Section) Laboratory animals or Model organisms:

In which section is the information available?
(Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)If collected and within the bounds of privacy constraints report on age, sex and gender or ethnicity for all study participants.

In which section is the information available?
, such as t-test (please specify whether paired vs. unpaired), simple χ2 tests, Wilcoxon and Mann-Whitney tests, can be unambiguously identified by name only, but more complex techniques should be described in the methods section; (Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)If your work benefited from core facilities, was their service mentioned in the acknowledgments section?Yes AcknowledgementsDesign-common tests

Please complete ALL of the questions below. Select "Not Applicable" only when the requested information is not relevant for your study. if
n<5, the individual data points from each experiment should be plotted.Any statistical test employed should be justified.Source Data should be included to report the data underlying figures according to the guidelines set out in the authorship guidelines on Data Each figure caption should contain the following information, for each panel where they are relevant: a specification of the experimental system investigated (eg cell line, species name).theassay(s)and method(s) used to carry out the reported observations and measurements.anexplicitmention of the biological and chemical entity(ies) that are being measured.anexplicitmention of the biological and chemical entity(ies) that are altered/varied/perturbed in a controlled manner.ideally,figurepanels should include only measurements that are directly comparable to each other and obtained with the same assay.plotsinclude clearly labeled error bars for independent experiments and sample sizes.Unless justified, error bars should not be shown for technical the exact sample size (n) for each experimental group/condition, given as a number, not a range; a description of the sample collection allowing the reader to understand whether the samples represent technical or biological replicates (including how many animals, litters, cultures, etc.).a statement of how many times the experiment shown was independently replicated in the laboratory.This checklist is adapted from Materials Design Analysis Reporting (MDAR) Checklist for Authors.MDAR establishes a minimum set of requirements in transparent reporting in the life sciences (see Statement of Task: 10.31222/osf.io/9sm4x).Please follow the journal's guidelines in preparing your the data were obtained and processed according to the field's best practice and are presented to reflect the results of the experiments in an accurate and unbiased manner.

Checklist for Life Science Articles (updated January Study protocol Information included in the manuscript? In which section is the information available?
(Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)If study protocol has been pre-registered, provide DOI in the manuscript.For clinical trials, provide the trial registration number OR cite DOI.

In which section is the information available?
(Reagents and ToolsTable, Materials and Methods, Figures, Data Availability Section) Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)Include a statement about sample size estimate even if no statistical methods were used.Were any steps taken to minimize the effects of subjective bias when allocating animals/samples to treatment (e.g.randomization procedure)?If yes, have they been described?Describe inclusion/exclusion criteria if samples or animals were excluded from the analysis.Were the criteria pre-established?If sample or data points were omitted from analysis, report if this was due to attrition or intentional exclusion and provide justification.NotFor every figure, are statistical tests justified as appropriate?Do the data meet the assumptions of the tests (e.g., normal distribution)?Describe any methods used to assess it.Is there an estimate of variation within each group of data?Is the variance similar between the groups that are being statistically compared? (

Sample definition and in-laboratory replication Information included in the manuscript? In which section is the information available?
(Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)In the figure legends: state number of times the experiment was replicated in laboratory.YesFigure legends, further details given in Methods section In the figure legends: define whether data describe technical or biological replicates.Yes Figure legends, further details given in Methods section

In which section is the information available?
(Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section)Include a statement confirming that informed consent was obtained from all subjects and that the experiments conformed to the principles set out in the WMA Declaration of Helsinki and the Department of Health and Human Services Belmont Report.State if relevant permits obtained, provide details of authority approving study; if none were required, explain why.
Studies involving human participants: State details of authority granting ethics approval (IRB or equivalent committee(s), provide reference number for approval.Not ApplicableStudies involving human participants:Not ApplicableStudies involving human participants: For publication of patient photos, include a statement confirming that consent to publish was obtained.Not Applicable Studies involving experimental animals: State details of authority granting ethics approval (IRB or equivalent committee(s), provide reference number for approval.Include a statement of compliance with ethical regulations.Not ApplicableStudies involving specimen and field samples:

Use Research of Concern (DURC) Information included in the manuscript? In which section is the information available?
(Reagents and ToolsTable, Materials and Methods, Figures, Data Availability Section) Could your study fall under dual use research restrictions?Please check biosecurity documents and list of select agents and toxins (CDC): https://www.selectagents.gov/sat/list.htmNot Applicable If you used a select agent, is the security level of the lab appropriate and reported in the manuscript?Not Applicable If a study is subject to dual use research of concern regulations, is the name of the authority

granting approval and reference number for
the regulatory approval provided in the manuscript?

and III randomized controlled trials
Table, Materials and Methods, Figures, Data Availability Section) State if relevant guidelines or checklists (e.g., ICMJE, MIBBI, ARRIVE, PRISMA) have been followed or provided.Not Applicable For tumor marker prognostic studies, we recommend that you follow the REMARK reporting guidelines (see link list at top right).See author guidelines, under 'Reporting Guidelines'.Please confirm you have followed these guidelines., please refer to the CONSORT flow diagram (see link list at top right) and submit the CONSORT checklist (see link list at top right) with your submission.See author guidelines, under 'Reporting Guidelines'.Please confirm you have submitted this list.Reagents and Tools Table, Materials and Methods, Figures, Data Availability Section) (