The Cockayne syndrome protein B is involved in the repair of 5-AZA-2′-deoxycytidine-induced DNA lesions

Oncotarget35069 www.oncotarget.com The Cockayne syndrome protein B is involved in the repair of 5-AZA-2′-deoxycytidine-induced DNA lesions Estefanía Burgos-Morón2,*, José Manuel Calderón-Montaño2,*, Nuria Pastor1, Andreas Höglund3,5, Ángel Ruiz-Castizo1, Inmaculada Domínguez1, Miguel López- Lázaro2, Nabil Hajji4, Thomas Helleday3, Santiago Mateos1 and Manuel Luis Orta1 1Department of Cell Biology, Faculty of Biology, University of Seville, 41012 Seville, Spain 2Department of Pharmacology, Faculty of Pharmacy, University of Seville, 41012 Seville, Spain 3Science for Life Laboratory, Division of Translational Medicine and Chemical Biology, Department of Medical Biochemistry and Biophysics, Karolinska Institute, S-171 21 Stockholm, Sweden 4Department of Medicine, Division of Experimental Medicine, Centre for Pharmacology & Therapeutics, Toxicology Unit, Imperial College London, Hammersmith Campus, London, W12 0NN UK 5Present address: Sprint Bioscience AB, 141 57 Huddinge, Sweden *The authors contributed equally to this work and should be regarded as joint First Authors Correspondence to: Manuel Luis Orta, email: [email protected] Keywords: CSB; 5-azadC; DNMT1; DNA damage; transcription Received: March 02, 2018 Accepted: September 10, 2018 Published: October 12, 2018 Copyright: Burgos-Morón et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT The Cockayne Syndrome Protein B (CSB) plays an essential role in Transcription- Coupled Nucleotide Excision Repair (TC-NER) by recruiting repair proteins once transcription is blocked with a DNA lesion. In fact, CSB-deficient cells are unable to recover from transcription-blocking DNA lesions. 5-Aza-2′-deoxycytidine (5-azadC) is a nucleoside analogue that covalently traps DNA methyltransferases (DNMTs) onto DNA. This anticancer drug has a double mechanism of action: it reverts aberrant hypermethylation in tumour-suppressor genes, and it induces DNA damage. We have recently reported that Homologous Recombination and XRCC1/PARP play an important role in the repair of 5-azadC-induced DNA damage. However, the mechanisms involved in the repair of the DNMT adducts induced by azadC remain poorly understood. In this paper, we show for the first time the importance of CSB in the repair of azadC- induced DNA lesions. We propose a model in which CSB initiates a signalling pathway to repair transcription blocks induced by incorporated 5-azadC. Indeed, CSB-deficient cells treated with 5-azadC show a delay in the repair of trapped DNMT1, increased levels of DNA damage and reduced survival. INTRODUCTION After millions of years of evolution, cells have evolved complex mechanisms to repair DNA breaks and prevent mutations. Although it has been known for many years that transcriptional stress plays an important role in genomic instability [1–5], it was in the mid-eighties when an additional DNA repair mechanism associated with transcriptionally active genes [6–7] was discovered. This specialized DNA repair processes, called Transcription-Coupled Repair (TCR), couples RNA polymerase blocks with the efficient removal of DNA lesions in the transcribed strand. This pathway is considered as a branch of the nucleotide excision repair pathway (NER). In humans, mutations in NER lead to a variety of DNA repair disorders, including Cockayne syndrome (CS), in which there is a deficiency in TCR. Two complementation groups of CS, designed CSA and CSB have been identified. Cells with mutations in any of these proteins cannot resume transcription after the UV-induced blockage of RNA polymerase [8, 9]. www.oncotarget.com Oncotarget, 2018, Vol. 9, (No. 80), pp: 35069-35084 Research Paper Oncotarget35070 www.oncotarget.com CSB is a 168 kDa protein related to the SWI/SNF family of ATP-dependent chromatin remodelers; this protein has nucleosome remodeling activity and binds to core histone proteins in vitro. When transcription fork is blocked, CSB protein is recruited and strongly interacts with RNA pol II. This protein acts as a chromatin remodeling factor displacing nucleosomes and recruiting some protein complexes, including the CSA complex, core NER factors (XPA, TFIIH, XPG, XPF-ERCC1, and RPA) and histone acetyltransferase p300 (that also works as a chromatin remodeling factor) [8]. The CSA complex acts by ubiquitination and subsequent degradation of CSB, Figure 1: 5-azadC induces transcription-coupled DSBs. (A) Experimental schedule followed. AA8 cells were treated for 6h with 15 μM of 5-azaC. DRB (30μM) was present during the last 3h. (B) In order to evaluate DNMT1 foci, DNMT1-GFP transfected cells were treated as described above, fixed in paraformaldehyde and scored for DNMT1 foci. Cells containing >5 DNMT1 foci were scored as positive. (C) Alternatively the number of DNMT1 foci was scored using Image J software. (D) In order to analyze chromosomal abnormalities after 5-azadC treatment (50μM), cells were treated with colcemid and metaphase spreads were made as described in the material and methods section. The means and SEM of three independent experiments are shown. Data were statistically analyzed using Student’s t-test. Data were considered statistically different when P <0.05 (*) or P <0.01 (**). Oncotarget35071 www.oncotarget.com RNA pol II, CSA itself and histones [10]. This clearance of proteins is needed for DNA repair and subsequent resumption of transcription. Apart from its roles in transcription coupled nucleotide excision repair (TCNER) and chromatin remodeling, CSB is thought to be involved in oxidative damage [11], crosslink repair [12], telomere maintenance [13], transcription associated DNA recombination [14], double strand break repair choice and checkpoint activation [15]. 5-aza-2′-deoxycytidine (5-azadC), also called decitabine, is a cytidine analogue that is incorporated randomly in the genome during replication. This drug is effective in the treatment of Myelodysplastic Syndromes and Acute Myeloid Leukemia (AML), this latter especially in elderly patients [16, 17]. Its mechanism of action involves the covalent trapping of DNA methyltransferases (DNMTs) onto DNA, generating a whole hypomethylation state [18]. Therefore, this drug can reactivate the expression of Tumour Suppressor Genes whose promoters are highly hypermethylated [19]. Trapped DNMTs onto DNA generate DNA damage, which also contributes to the anticancer properties of this nucleoside [20–22]. The mechanisms involved in the repair of the DNMT adducts induced by azadC remain poorly understood. We recently reported that these bulky lesions can interfere with replication forks and induce double strand breaks (DSBs) that are repaired by Homologous Recombination (HR) involving Fanconi Anemia (FA) proteins (21). Also, we have proposed that XRCC1 and PARP could play a role in the repair of DNMT adducts [22]. In the present paper, we investigate the role of CSB in the repair of the lesions induced by 5-azadC. We show that CSB is important in the repair of the lesions induced by 5-azadC in a process that is independent of classic TC-NER. We found that a transcription coupled DNA damage response (TC-DDR) is activated shortly after 5-azadC incorporation in a CSB dependent manner. Furthermore, our results revealed that CSB-deficient cells displayed a delay in the repair of DNMT1 adducts, resulting in hypersensitivity to 5-azadC. Finally, we demonstrate that CSB and transcription act in the same pathway to repair 5-azadC-induced DNA lesions and promote survival. RESULTS 5-azadC induces transcription-dependent double strand breaks We have recently reported that 5-azadC induces DNA damage that depends on active replication, which suggests that trapped DNMT collapses with oncoming replication forks into DSBs (21). Here we wanted to investigate the possible role of transcription elongation in the generation of DNA breaks after 5-azadC treatment. In order to distinguish between replication and transcriptionassociated DNA DSBs, we exposed the cells to 5-azadC for 6 h. This short exposure time impossibilities that a second round of DNA replication takes place over a 5-azadC substituted DNA (Figure 1A). DNA damage was monitored by the scoring of chromosome aberrations as indicators of DSBs [23]. Cells were co-treated with the transcription inhibitor DRB during the last 3 h to see if transcription elongation is required for the formation of chromosomal abnormalities. Under these conditions, when DNMT1-GFP transfected AA8 cells were treated with 5-azadC, they presented increased levels of DNMT1 foci as compared to control cells (Figure 1B–1C). Of interest, we could not find any significant difference in the levels of trapped DNMT1 between 5-azaC treated and co-treated cells. This data indicates that similar levels of the initial lesions are present in both cases. It is worth mentioning that a 6 h treatment with 5-azadC induced chromosome aberrations (Figure 1D); this suggests that the breaks arising as a consequence of collapsed replications forks is not the only mechanism by which 5-azadC induces DNA damage. Indeed, transcription inhibition abrogates the levels of 5-azadC induced chromosome aberrations (Figure 1D). Also, we observed that this abrogation was not caused by a reduction in the mitotic index (Supplementary Figure 1A) or by variations in cell cycle profiles (Supplementary Figure 1B and 1C). Furthermore, our experimental schedule is similar to other previously published in relation with Topoisomerase I adducts [23]. A transcription dependent DNA damage response (TC-DDR) is activated after 5-azadC incorporation It has been shown that Topoisomerase I-DNA complexes can block transcription forks resulting in the activation of a DDR involving ATM-CHK2, γ-H2AX, 53BP1 and DNA-PK [24, 25]. H2AX is phosphorylated at Serine 139 in response to several types of DNA damage such as DNA DSBs [26] and replication stress [27]. 53BP1 foci are a more restrictive marker for DSBs and are thought to be indicative of a NHEJ repair [28, 29]. Using the same experimental design described above (Figure 1A), we wanted to analyze if a TC-DDR is activated. We used human cells defective in CSB from a Cockayne syndrome patient complemented with human CSB (CS1AN CSB-GFP). According to the results presented before, 5-azadC induced an increase in the phosphorylation of H2AX which was reduced when cells were co-treated with DRB (Figure 2A). Similar results were found when examining 53BP1 foci (Figure 2B). CSB has been recently proposed to have a role in the repair of Topoisomerase I cleavable complexes stabilized by Camptothecin (CPT) [30]. Under our experimental design we found an accumulation of CSB in foci after a treatment with 5-azadC that seems to be transcription dependent (Figure 2C). Oncotarget35072 www.oncotarget.com We found that γ-H2AX co-localized with CSB foci, and that the frequency of 5-azadC induced co-localized foci was reduced when transcription was inhibited (Figure 2D). When looking at co-localization in DNMT1-GFP transfected cells, we found that DNMT1 foci co-localized with γ-H2AX. This data indicates the activation of a DDR when 5-azadC was administered at short times without the interference of a second replication round (Figure 2E). Under these conditions, γ-H2AX also co-localized with 53BP1 foci (Figure 2F). All these data suggest that transcription has a role in the sensing of the lesions induced by 5-azadC. Figure 2: A transcription dependent DNA damage response is activated after 5-azadC incorporation. CS1AN CSB-GFP complemented cells were treated following the schedule presented in Figure 1A and γ-H2AX (A), 53BP1 (B) and CSB (C) foci were analyzed. The average of γ-H2AX foci colocalizing with CSB per cell is representated in (D). Representative micrographs are depicted in (E). (F) Co-localization between DNMT1 and γ-H2AX foci. AA8 cells were transfected with DNMT1-GFP plasmid, treated for 6 h with 5-azadC and immunomarked for γ-H2AX. (G) Co-localization between 53BP1 and γ-H2AX. AA8 cells were treated for 6 h with 5-azadC and cells were fixed and marked for immunofluorescence. Confocal micrographs are depicted. The means and SEM of two-three independent experiments are shown. Statistical significance was determined using Student’s t-test, n.s. = non-significant, **P < 0.01, *P < 0.05. Oncotarget35073 www.oncotarget.com CSB mutant cells have defects in the signaling of DNA damage induced by 5-azadC Once described that the initial DNA lesions induced by 5-azadC are sensed in a transcription dependent fashion, we wanted to address the importance of CSB in the setup of this process. We use UV61 and CS1AN cells (CSB deficient) and data were compared with AA8 (wild type) or CS1AN CSB-GFP (CSB complemented) cells. Cells were treated for 6 hour with 5-azadC and the frequency of DNMT1 foci was monitored in wild type and CSB deficient cells. Our results showed no difference in the number of DNMT1 foci in both cell lines after 5-azadC treatment (Figure 3A–3B). Therefore, the frequency of initial lesions seems to be similar, which makes possible to compare the DDR activated in both cell lines. We found that UV61 cells were unable to phosphorylate H2AX in response to 5-azadC (Figure 3C). Accordingly, we could not find an increase in the co-localization between γ-H2AX and DNMT1 in response to 5-azadC in CSB mutant cells (Figure 3D). When we wanted to analyse γ-H2AX and 53BP1 foci in human cells, a similar pattern was found, which demonstrates a similar behaviour of hamster and human cells (Figure 3E–3F). A classical feature of CSB-deficient cells is related with its transcriptional incapability to recover from a transcription block [8]. In order to monitor the effect of 5-azadC in the transcription rates, we incubated AA8 and UV61 cells with increasing doses of 5-azadC for 6 h. Then, cells were pulse-labelled for 15 minutes with 3H-Uridine and transcription rates were measured in a scintillation counter. Our data shows a consistent decrease in transcription rates in UV61 cells from the first dose assayed (5μM). However, the block started to be evident only when high doses of the drug were assayed for wild type cells (Figure 3G). Altogether, our results show that CSB is involved in the signaling of 5-azadC lesions required to resume transcription. Delayed repair of trapped DNMT1 in CSB mutant cells Trapped DNMTs are the primary lesion induced by 5-azadC responsible for tumour cell death [31]. It has been recently shown that trapped Topoisomerases by poisons are obstacles to transcription, causing blocks that act as sensors for the repair of the adducts and the resumption of transcription [24]. Moreover CSB has been recently implicated in this mechanism [30]. Based on the evidence reported above, our data suggests an implication of CSB in the signaling of trapped DNMT1. If this is the case, it is conceivable to think that CSB deficient cells would be defective in the repair of trapped DNMT1 induced by 5-azadC. To address this question, we performed cellular fractionation to isolate chromatin in which we detected trapped DNMT1. We show that DNMT1-GFP transfected AA8 and UV61 cells presented similar levels of native and ectopic DNMT1 in the whole lysates (Figure 4A). In order to monitor the repair of trapped DNMT1, DNMT1-GFP transfected cells were treated with 15μM of 5-azadC for 24 h, washed and allowed to recover for different times in drug free media. When looking at the levels of remaining DNMT1-GFP trapped in chromatin, we found a clear increase in UV61 cells compared to wild type cells (Figure 4B). Taking into account that there is some repair during the treatment time, a plausible explanation would be a defect in the repair in CSB deficient cells, responsible for the increased level of trapped DNMT1 observed after treatment. Also, we think that this repair defect in CSB mutant cells would be the cause for the delayed repair observed as well. This result was confirmed by fluorescence microscope in DNMT1-GFP transfected cells (Figure 4D). In this case, DNMT1 foci were scored over time after a 24 hour treatment with 5-azadC. Our data show that CSB cells present more DNMT1 foci all along the repair time course. These findings agree very well with data presented above and highlight the importance that CSB has in the repair of trapped DNMT1. Increased levels of DNA damage and reduced survival in CSB mutant cells after 5-azadC treatment Recently, we have shown that trapped DNMTs interfere with replication and generate DNA damage [21]. If CSB mutant cells have defects in the repair of trapped DNMT1, the levels of DNA damage should increase over time. To test this hypothesis, we treated cells with 5-azadC for 24 h and allowed to repair for another 24 h. DNA damage was evaluated by the scoring of γ-H2AX, 53BP1 and RAD51 foci (Figure 5A–5D). We found that, in contrast to the data obtained when we used short treatments (Figure 2), CSB mutant cells display increased levels of DNA damage foci. In our opinion this apparent contradiction can be explained considering that foci arising after short treatments indicate initial signaling of the repair of 5-azadC lesions. On the contrary, many of the foci visualized after long treatments could be related with the signaling of secondary lesions i.e. collisions with the replication fork with trapped DNMT1. On the other hand, CSB deficient cells showed a high increase in 5-azadC induced RAD51 foci as compared to wild type cells (Figure 5D). We hypothesize that CSB cells could present increased levels of collapsed forks as a result of unrepaired DNMT1 triggering RAD51- dependent HR for repair. Of interest, CSB deficient cells showed increased micronuclei frequency, which is indicative of residual DNA breaks (Figure 5E). Oncotarget35074 www.oncotarget.com Next we wanted to examine if DNA repair defects observed in CSB mutant cells have a negative consequence in survival. We performed clonogenic assays both in hamster and human cells (Figure 6A–6B). As can be seen, CSB mutant cells (UV61 and CS1AN) presented reduced survival when exposed to 5-azadC. In good agreement with this, CSB mutant cells also showed a reduced growth when they were chronically exposed to 5-azadC (Figure 6C–6D). Cell cycle profiles agree very well with the above data and showed increased levels of cell death (sub-G1 cells) in CSB mutant cells when treated with 5-azadC (Figure 6E–6H). Figure 3: CSB mutant cells have defects in the signaling of the initial DNA damage induced by 5-azadC. (A-B) Similar levels of trapped DNMT1 in wild type AA8 and CSB defective UV61 cells. DNMT1 CSB-GFP cells were treated for 6 h with 15μM of 5-azadC, fixed and prepared for scoring of DNMT1 foci. Cells > 5 foci were scored as positive (A). Alternatively the number of DNMT1 foci/cell is also represented (B). (C) CSB cells have defects forming γ-H2AX foci in response to a short 5-azadC treatment (6 h). (D) 5-azadC induces DNMT1-γH2AX co-localization in a CSB dependent fashion. DNMT1-GFP transfected cells were treated for 6 h with 5-azadC, fixed and immunolabelled for γH2AX. Mander’s co-localization index was employed (Image J). (E and F) Importance of CSB in the formation of γ-H2AX and 53BP1 foci in human cells. CS1AN and CS1AN CSB-GFP cells were treated for 6 h with 5-azadC (15μM), fixed and immunolabelled for the indicated proteins. A clear increase in foci was observed for complemented cells. However CSB cells couldn’t form repair foci. (G) CSB cells present transcription blocks after a short treatment with 5-azadC. AA8 and UV61 cells were treated with increasing doses of 5-azadC for 6 h and labelled with tritiated uridine for 15 minutes in order to monitor transcription rates. CSB cells presented blocks in transcription elongation from the first dose assayed. However for wild type cells blocks are evident only for higher doses assayed. Oncotarget35075 www.oncotarget.com Transcription and CSB work in the same pathway to promote survival after treatment with 5-azadC Apart from TC-NER, CSB has been implicated in several processes such as Base Excision Repair [11, 32, 33], nucleosome positioning [34] and mitochondrial DNA repair [35]. As far as we know, many of these processes are not closely related to transcription elongation. Therefore, we analyzed if the function of CSB is epistatic with transcription elongation in response to 5-azadC. We treated AA8 and UV61 cells with 5-azadC for 24 h. During the last 12 h cells were cultured in the presence/absence of the transcription inhibitors DRB or Figure 4: Delayed repair of trapped DNMT1 in CSB mutant cells. (A) Validation of the fractionation protocol followed. AA8 and UV61 cells were transfected with DNMT1-GFP plasmid before lysis. Whole (W) or fractionated lysis were carried out. Cytoplasmic (Cyt), nuclear soluble (Nu.sol.) and chromatin (Chr.) fractions were obtained. Our data show that endogenous and ectopic DNMT1 are equally expressed in both cell types. The expected localization of GADPH, Fibrillarin and γ-H2AX in the cytosolic, nuclear, and chromatin enriched fractions, respectively is shown. (B) Repair of chromatin bound DNMT1-GFP in wild type AA8 and CSB deficient UV61 cells. Cells were treated for 24h with 15μM of 5-azadC and allowed to repair in drug free media for the indicated times. Cells were fractionated and chromatin fractions were loaded analyzed by SDS-PAGE. (C) A densitometric analysis of the chromatin bound DNMT1-GFP in relation to the chromatin marker H3 is shown. (D) Alternatively DNMT1-GFP transfected AA8 and UV61 cells were treated for 24h with 15μM of 5-azadC and allowed to repair in drug free media for different times, fixed and scored for DNMT1 foci in a fluorescence microscope. The means and SEM of two independent experiments are shown. Data were considered statistically different when P < 0.05 (*) or P < 0.01(**). Oncotarget35076 www.oncotarget.com α-Amanitin. Next, we monitored DNA repair foci (Figure 7A–7C) or micronuclei (Figure 7D) induction. An increase of DNA repair foci induced by 5-azadC was observed when transcription was blocked in wild type cells but not in UV61 cells (Figure 7A–7C). The same pattern was observed when looking at micronuclei induction (Figure 7D). The lack of potentiation in micronuclei induction in CSB deficient cells cannot be explained by a delay in cell cycle, as no difference in the progression of the cell cycle was observed (Supplementary Figure 2). Overall Figure 5: Increased levels of DNA damage in CSB mutant cells after a prolongued 5-azadC treatment. (A–D) AA8 and UV61 cells were treated for 24 h with 5-azadC (2μM) washed and allowed to repair for 24h. Then were fixed and prepared for immunological staining of γ-H2AX (A-B), 53BP1 (C) and RAD51 (D). (E) In order to monitor residual DNA damage after a 24h treatment with 5-azadC, we performed micronucleus assay. Cells were treated with increasing doses of 5-azadC, washed and allowed to repair in cytochalasin B containing media for 18h. Cells were then fixed and scored for the number of micronuclei in binucleated cells. The means and SEM of three independent experiments are shown. Data were statistically analysed using Student’s t-test. Data were considered statistically different when P < 0.05 (*) or P < 0.01(**). Oncotarget35077 www.oncotarget.com these results suggest that transcription works in the same pathway than CSB to repair lesions induced by 5-azadC. Finally, we wanted to see if the epistatic interaction could be also observed when looking at cell survival (Figure 7E). In good agreement with the data presented above, a clear sensitization by DRB was obtained just in wild type cells. As a whole, data suggest that transcription works upstream to CSB in response to lesions induced by 5-azadC. DISCUSSION Several DNA damaging agents are known to form DNA-protein cross-links (DPC), including topoisomerase I and II poisons, 5-azadC, PARP inhibitors and cisplatin. In recent years, some mechanisms involved in the repair of these DNA insults have been elucidated [36], although many questions remains unanswered for the DPC induced by 5-azadC. Figure 6: CSB promotes survival after 5-azadC incorporation. (A and B) Survival after 5-azadC treatments. AA8 and UV61 cells (A) or CS1AN and CS1AN CSB-GFP cells (B) were seeded at low density, treated with 5-azadC for 24h and allowed to repair in fresh media for 7-9 days to form colonies. Colonies were fixed and stained with methylene blue/methanol 4 g/L. Colonies with more than 50 cells were scored. Data are plotted in relation to the survival of control cells (100 %). (C and D) Cell growth after a continuous exposure to 5-azadC (15μM). Cells were collected at the indicated times and viable cells were scored (tripan blue negative). (E–H) Cell cycle profiles after a continuous exposure to 5-azadC (15μM) in AA8-UV61 (E), CS1AN-CS1AN CSB-GFP (F) and quantifications of subG1 cells (G-H). The means and SEM of three independent experiments are shown. Data were statistically analized using Student’s t-test. Data were considered statistically different when P < 0.05 (*) or P < 0.01(**). Oncotarget35084 www.oncotarget.com RH, Hoeijmakers JH, Mullenders LH, Vermeulen W, Marteijn JA. The core spliceosome as target and effector of non-canonical ATM signalling. Nature. 2015; 523:53–58. 40. Petermann E, Helleday T. DNA replication-associated lesions: importance in early tumorigenesis and cancer therapy. Biochem Soc Trans. 2007; 35:1352–54. 41. Sakai A, Sakasai R, Kakeji Y, Kitao H, Maehara Y. PARP and CSB modulate the processing of transcription-mediated DNA strand breaks. Genes Genet Syst. 2012; 87:265–72. 42. Guirouilh-Barbat J, Redon C, Pommier Y. Transcriptioncoupled DNA double-strand breaks are mediated via the nucleotide excision repair and the Mre11-Rad50-Nbs1 complex. Mol Biol Cell. 2008; 19:3969–81. 43. Mullenders L. DNA damage mediated transcription arrest: step back to go forward. DNA Repair (Amst). 2015; 36:28–35. 44. Nakano T, Katafuchi A, Matsubara M, Terato H, Tsuboi T, Masuda T, Tatsumoto T, Pack SP, Makino K, Croteau DL, Van Houten B, Iijima K, Tauchi H, Ide H. Homologous recombination but not nucleotide excision repair plays a pivotal role in tolerance of DNA-protein cross-links in mammalian cells. J Biol Chem. 2009; 284:27065–76. 45. Flohr C, Bürkle A, Radicella JP, Epe B. Poly(ADP-ribosyl) ation accelerates DNA repair in a pathway dependent on Cockayne syndrome B protein. Nucleic Acids Res. 2003; 31:5332–37. 46. Zhou W, Doetsch PW. Transcription bypass or blockage at single-strand breaks on the DNA template strand: effect of different 3′ and 5′ flanking groups on the T7 RNA polymerase elongation complex. Biochemistry. 1994; 33:14926–34. 47. Tirado-Magallanes R, Rebbani K, Lim R, Pradhan S, Benoukraf T. Whole genome DNA methylation: beyond genes silencing. Oncotarget. 2017; 8:5629–37. https://doi.org/10.18632/oncotarget.13562 48. Shukla S, Kavak E, Gregory M, Imashimizu M, Shutinoski B, Kashlev M, Oberdoerffer P, Sandberg R, Oberdoerffer S. CTCF-promoted RNA polymerase II pausing links DNA methylation to splicing. Nature. 2011; 479:74–79. 49. Deaton AM, Bird A. CpG islands and the regulation of transcription. Genes Dev. 2011; 25:1010–22. 50. Dan J, Chen T. Genetic Studies on Mammalian DNA Methyltransferases. Adv Exp Med Biol. 2016; 945:123–50. 51. Orren DK, Dianov GL, Bohr VA. The human CSB (ERCC6) gene corrects the transcription-coupled repair defect in the CHO cell mutant UV61. Nucleic Acids Res. 1996; 24:3317–22. 52. Dinant C, Ampatziadis-Michailidis G, Lans H, Tresini M, Lagarou A, Grosbart M, Theil AF, van Cappellen WA, Kimura H, Bartek J, Fousteri M, Houtsmuller AB, Vermeulen W, Marteijn JA. Enhanced chromatin dynamics by FACT promotes transcriptional restart after UV-induced DNA damage. Mol Cell. 2013; 51:469–79. 53. Yamaguchi Y, Takagi T, Wada T, Yano K, Furuya A, Sugimoto S, Hasegawa J, Handa H. NELF, a multisubunit complex containing RD, cooperates with DSIF to repress RNA polymerase II elongation. Cell. 1999; 97:41–51. 54. Peterlin BM, Price DH. Controlling the elongation phase of transcription with P-TEFb. Mol Cell. 2006; 23:297–305. 55. Méndez J, Stillman B. Chromatin association of human origin recognition complex, cdc6, and minichromosome maintenance proteins during the cell cycle: assembly of prereplication complexes in late mitosis. Mol Cell Biol. 2000; 20:8602–12.