Recovery of protein synthesis to assay DNA repair activity in transcribed genes in living cells and tissues

Abstract Transcription-coupled nucleotide excision repair (TC-NER) is an important DNA repair mechanism that protects against the negative effects of transcription-blocking DNA lesions. Hereditary TC-NER deficiencies cause pleiotropic and often severe neurodegenerative and progeroid symptoms. While multiple assays have been developed to determine TC-NER activity for clinical and research purposes, monitoring TC-NER is hampered by the low frequency of repair events occurring in transcribed DNA. ’Recovery of RNA Synthesis’ is widely used as indirect TC-NER assay based on the notion that lesion-blocked transcription only resumes after successful TC-NER. Here, we show that measuring novel synthesis of a protein after its compound-induced degradation prior to DNA damage induction is an equally effective but more versatile manner to indirectly monitor DNA repair activity in transcribed genes. This ‘Recovery of Protein Synthesis’ (RPS) assay can be adapted to various degradable proteins and readouts, including imaging and immunoblotting. Moreover, RPS allows real-time monitoring of TC-NER activity in various living cells types and even in differentiated tissues of living organisms. To illustrate its utility, we show that DNA repair in transcribed genes declines in aging muscle tissue of C. elegans. Therefore, the RPS assay constitutes an important novel clinical and research tool to investigate transcription-coupled DNA repair.


INTRODUCTION
DNA damage induced by various environmental and metabolism-deri v ed agents continuously threatens the integrity and functionality of DN A. DN A damage interferes with essential DNA-transacting processes like transcription and replication and contributes to aging and causes mutagenesis, leading to cancer ( 1 ).Various dedicated DNA repair pathways maintain genomic integrity by removing DNA damage depending on the type of lesion, its genomic location and the cell cycle stage.Nucleotide excision repair (NER) is an important DNA repair mechanism that protects organisms against cancer and aging by removing different DNA helix-distorting lesions, such as those induced by UV light, cancer therapeutics like cisplatin, metabolism-deri v ed aldehydes and various genotoxic environmental chemicals including poly cy clic aromatic hydrocarbon and aromatic amines found in smoke and cooked food ( 2 , 3 ).
NER is initiated by lesion detection via global genome NER (GG-NER) anywhere in the entire genome, or via transcription-coupled NER (TC-NER) e xclusi v ely in the e93 Nucleic Acids Research, 2023, Vol. 51, No. 18 PAGE 2 OF 12 tr anscribed str ands of acti v e genes.Lesion detection via GG-NER is mediated by the XPC protein, which continuously probes DNA and recognizes lesions by interacting with the DNA strand opposite of the lesion and inserting a ␤-hairpin domain into the distorted DNA duplex (4)(5)(6).XPC e xists in comple x with CETN2 and RAD23B and is aided by the CRL4 DDB2 ubiquitin ligase complex when lesions are difficult to detect and / or when chromatin needs to be re-organized ( 7 , 8 ).Lesion detection via TC-NER occurs when lesions b lock forwar d progression of elongating RN A pol ymerase II (Pol II), which triggers the stable association of the CSB protein with Pol II ( 9 ).CSB then recruits the CRL4 CSA ubiquitin ligase complex and additional TC-NER factors ( 10 ).Lesion detection via either XPC or CSB leads to sequential recruitment of core NER factors, including TFIIH and XPA, which unwind DNA and check for the presence of damage, and the endonucleases ERCC1-XPF and XPG that excise a 22-30 bp DNA stretch containing the lesion ( 11 , 12 ).The resulting gap is then filled in by DNA synthesis mediated by replication factors, DN A pol ymerases and ligases.
The biological significance of NER is illustrated by the se v ere cancer-prone, de v elopmental and pr oger oid symptoms associated with hereditary mutations in NER genes ( 13 ).GG-NER deficiency causes xeroderma pigmentosum (XP), which manifests as photosensiti v e skin and strong cancer predisposition ( 14 ).Mutations in TC-NER factors cause either the mild UV-sensiti v e syndrome (UV S S) or the much more se v ere Cockayne syndrome (CS) ( 15 , 16 ).Mutations in genes involved in both GG-NER and TC-NER can cause se v ere XP, often also combined with CS features, or a photosensiti v e form of trichothiodystrophy (TTD) ( 17 , 18 ).UV S S is mainly characterized by telangiectasia and sun sensitivity of the skin, whereas CS is characterized by a pleiotropic range of se v ere symptoms including growth failur e, progr essi v e organ decline and neurodegeneration and segmental progeria.It is currently not understood why different mutations in TC-NER factors lead to this wide array of symptoms, but it is hypothesized that this may be related to differences in clearance of lesion-stalled Pol II ( 19 ).TTD is characterized by brittle hair and nails, ichthyosis and progressi v e mental and physical retardation and is thought to be mainly caused by problems with gene expression ( 20 ).Inter estingly, r esear ch on pr oger oid features of TC-NER deficiency disorders, in humans and mouse models, has re v ealed tha t accumula ting DNA damage interfering with transcription is one of the major underlying causes of aging ( 21 , 22 ).To study the etiology of aging, and how this affects organs differ ently , it is ther efor e important to be ab le to inv estigate DNA repair activity in transcribed genes in different types of cells in vivo .
Both the clinical diagnosis of NER disorders as well as the discovery of novel genes involved in NER depends heavily on the availability of reliable and straightforward assays that can discriminate GG-NER and TC-NER deficiency.GG-NER activity can be determined by measuring unscheduled DNA synthesis (UDS) that r estor es the singlestranded DN A ga p resulting from DN A damage excision.Originally, this assay was used to link DNA repair defects to XP and was based on GG-NER-dependent incorporation of radioacti v ely labeled thymine analogs ( 23 ).Nowa-days, these have been replaced by 5-ethynyl-2 -deoxyuridine, which can be visualized by fluorescent labeling using a click chemistry r eaction ( 24 , 25 ).T C-NER activity is mor e difficult to detect largely due to the fact that this repair pathway only removes a minority of lesions, i.e.only those that block transcription.Originally, TC-NER was demonstrated using a Southern blot-based assay that showed prefer ential r epair in the tr anscribed str ands of acti v e genes ( 26 ), w hich was subsequentl y used to show TC-NER deficiency in cells from CS patients ( 27 ).Since then, multiple assays have been developed to monitor TC-NER activity either directly, such as the gene-specific qPCR ( 28 ), comet-FISH ( 29 ), the amplified UDS in GG-NER deficient cells ( 30 ) and strand-specific ChIP-Seq ( 31 ) assays, or indirectly, such as the host cell reactivation ( 32 ) and the Recovery of RNA synthesis (RRS) assays.Because of its ease of use, the RRS assay is commonly used in NER r esear ch and in the clinic.It is based on the notion that DNA damage inhibits transcription and that this can only resume if TC-NER is acti v e and remov es the damage ( 33 ).Transcription recovery can be measured by labeling RNA with radioacti v e or bromo-uridine or, as is nowadays more common, 5ethynyluridine that can be fluorescently labeled using click chemistry ( 24 , 25 ).A drawback of RRS and most other techniques is that these assays only work in cells in culture and cannot easily be used to determine repair activity in realtime or in vivo .
Here, we show the utility of a nov el v ersatile assay to indirectl y monitor DN A repair in transcribed genes based on the idea that not only transcription but also translation will only resume after DNA damage induction if DNA repair is functional.We show that this 'recovery of protein synthesis' (RPS) assay is robust and reliable and can be performed by monitoring the novel synthesis of different proteins, using both fluorescence imaging as well as immunoblotting as r eadout.Mor eover, the RPS assay can be used to assay DNA repair activity in real-time both in living cells in culture as well as in vivo in dif ferentia ted cell types of a living organism.

Live-cell confocal laser-scanning microscopy
For li v e cell imaging, cells were grown on 24-mm coverslips and imaged using a Leica SP5 confocal microscope equipped with an environmental chamber set to 37 • C and 5% CO2.Confocal images were recor ded e v ery hour after UV irradia tion, as indica ted.Da ta collection and analysis was performed using LAS X software (Leica).

Fluorescence imaging
For imaging of fluorescence le v els, cells were grown on 24-mm coverslips and fixed using 3.6% formaldehyde (Sigma) diluted in PBS. 4 ,6-diamidino-2-phenylindole (DAPI; Sigma) staining was performed in PBS for 15 min at room temperature.Subsequently, coverslips were washed twice with PBS supplemented with 0.1% Triton-X-100 and once with PBS and mounted with aqua-poly / mount (Polysciences Inc).Cells were imaged using a Zeiss LSM700 microscope equipped with a Plan-Apochromat 40x / 1.3 Oli DIC M27 immersion lens (Carl Zeiss Microimaging Inc.).EGFP-FKBP F36V or EGFP-Androgen receptor expression in the nucleus was quantified in Fiji ImageJ.

C . eleg ans strains and experimental handling
Caenorhabditis elegans ( C. elegans ) wer e cultur ed accor ding to standar d methods ( 40 ) S1. HAL526, HAL534 and HAL535 str ains were gener ated by crossing CA1202 with xpc-1(tm3886) and / or csb-1(ok2335) mutants ( 42 ) and genotyped by PCR and sequencing.For the recovery of protein synthesis assay, animals of similar age were depleted of AID::GFP expression by culturing for 2 h on NGM plates containing 100 M auxin (3-indoleacetic acid; Sigma).Directly after depletion, animals were mock-trea ted or irradia ted with 120 J / m 2 UV-B (Philips TL-12 tubes, 40W) and allowed to recover for 48 h on NMG plates containing OP50 E. coli food.AID::GFP le v els wer e measur ed by imaging fluor escence in body wall muscle cells in the head of living animals on a Leica TCS SP8 microscope (LAS AF software, Leica).AID::GFP e xpression le v els were quantified in Fiji ImageJ.

Statistical analysis and softw ar e
Software used is listed in Supplementary Table S5.Statistical analyses were performed using Graph Pad Prism version 9 for Windows (GraphPad Software, La Jolla California USA).In each graph, mean values and S.E.M. error bars are shown and the number of experiments and cells or C. elegans tested is indicated in the legend.All underlying data is reported in Supplementary Table S6.For the li v e-cell imaging assays an unpaired t-test was used.For all other experiments, we applied a one-way ANOVA with correction for multiple comparison.S6.

Recovery of protein synthesis after DNA damage depends on transcription-coupled NER
As the RRS assay is based on measuring the ability of cells to recover transcription after DNA damage induction, we reasoned that the ability of cells to re-synthesize a degraded protein could be used as indirect TC-NER readout as well (Figure 1 A).After protein degradation, ongoing transcription will produce new proteins, which will still happen if cells incur transcription-blocking DNA damage that is re-paired.Howe v er, in the absence of TC-NER, transcription blockage will persist and less or no new protein will be produced.We first tested this idea using GFP, as imaging the resynthesis of a fluorescent reporter will allow the monitoring of TC-NER activity in real-time in living cells.We therefore generated a DNA construct expressing EGFP fused to the FKBP F36V degradation tag (dTAG) ( 38 ) and to a nuclear localization signal, which can be efficiently knocked-in at the AAVS1 locus using CRISPR-Cas9 in any cell type ( 43 ) and selected for by blasticidin (Figure 1 B).Proteins fused to the mutant FKBP F36V dTAG are ra pidl y degraded by the proteasome via polyubiquitination by the CRL4 Cereblon E3 ubiquitin ligase complex, after addition of a heterobifunctional dTAG ligand such as dTAG-13 to cells ( 38 ).Rapid protein depletion using this system before DNA damage induction will ther efor e enable us to determine if, after DNA damage induction, novel protein is synthesized and therefor e DNA r epair has taken place.After establishing human osteosarcoma U2OS cells stab ly e xpressing EGFP-FKBP F36V in the nucleus, we found by fluorescence imaging of fixed cells that addition of dTAG-13 led to rapid loss of the nuclear fluorescent signal, w hich re-a ppeared upon w ashing aw ay dTAG-13 (Figure 1 C).Subsequently, we tested if EGFP protein synthesis is dependent on TC-NER after DNA damage induction.We treated the cells with control siRN A or siRN A against the GG-and TC-NER factor XPF, the GG-NER factor XPC and / or the TC-NER factor CSB, after which we depleted EGFP.We first tested re-appearance of EGFP-FKBP F36V expression in the absence of UV irradiation and observed that this was similarly robust in all siRNA-treated cells (Supplementary Figure S1B).Hereafter, we UV irradiated the cells and determined EGFP-FKBP F36V protein synthesis r ecovery (Figur e 1 D).Untr eated cells, which wer e not exposed to dTAG-13 and UV-irradiation, were used for comparison.In control siRNA-treated cells, the EGFP fluorescence le v els returned to similar le v els as in untreated cells within 16 h (Figure 1 E and F).In sharp contrast, EGFP protein synthesis was completely abolished in XPFdepleted cells and almost completely abolished in CSBdepleted cells.Depletion of XPC led to a minor reduction in EGFP le v els and further reduced EGFP le v els in CSBdepleted cells to the same le v el as in XPF-depleted cells.These results demonstrate that protein resynthesis after its depletion can be used as readout of TC-NER capacity after DNA damage induction.
Our results suggest that in UV-irradiated cells GG-NER via XPC is responsible for repairing a minor fraction of lesions in transcribed genes, indica ting tha t this method not only monitors TC-NER activity but any type of DNA repair activity that removes transcription-blocking DNA lesions.UV-C irradiation mainly induces (6-4) photoproducts (6-4PPs) and cyclobutane pyrimidine dimers (CPDs).These lesions are ra pidl y removed with equal efficiency by T C-NER ( 44 ), but CPDs ar e m uch less efficientl y reco gnized and slower r epair ed b y GG-NER than b y TC-NER ( 5 , 45 ).The small fraction of XPC-dependent repair therefor e likely r eflects GG-NER of 6-4PPs rather than CPDs.Strikingly, using a moderate dose of 6 J / m 2 UV-C irradiation, we observed a persisting complete depletion of the EGFP-FKBP F36V protein in NER deficient cells, indicating that transcription of the encoding transgene was fully inhibited.This UV dose is estimated to produce on average less than one lesion per 10 000 bp ( 44 ), suggesting that our EGFP-FKBP F36V transgene of 1122 bp is not damaged in each cell.It is ther efor e not likely that Pol II is physically blocked by DNA damage in the EGFP-FKBP F36V transgene in e v ery cell measured.Indeed, transcription shutdown in response to DNA damage does not only occur in cis due to direct physical blockage of elongating Pol II in a particular gene, but also takes place in trans , i.e. through genome-wide inhibition of Pol II-mediated transcription initiation and elongation (46)(47)(48)(49)(50). Ther efor e, the lack of EGFP protein synthesis after this moder ate UV irr adiation of cells with depletion of XPF and CSB most likely reflects the genome-wide shutdown of transcription due to DNA damage induction in a subset of genes.Because of this, it is possible to use the translation of even a relatively short gene like the EGFP-FKBP F36V transgene to monitor DNA damageinduced transcriptional responses and the activity of TC-NER or other repair pa thways tha t remove transcriptionblocking lesions.In analogy to the RRS assay, we term our assa y theref ore 'Recovery of Protein Synthesis' (RPS).

Recovery of protein synthesis monitors TC-NER capacity in real-time in living cells
After observing that the RPS assay can be performed by imaging cells that wer e fix ed at defined time points, we tested if RPS can similarly be monitored in real-time in living cells.We ther efor e imaged cells tr eated with siRNA and depleted of EGFP-FKBP F36V for 9 h by li v e cell confocal microscopy.EGFP fluorescence signal recovered similarly in cells that were not UV irradiated and treated with either control or CSB siRNA (Supplementary Figure S2).In UV irradiated cells treated with control siRNA, the fluorescent EGFP signal r ecover ed also alr eady visibly within 1-2 h and continued to increase for 9 h without reaching plateau le v els yet.In contrast, in CSB-depleted cells, hardly any fluorescence signal was observed after 1-2 h and only a low recovery of fluorescence signal was visible a t la ter time points (Figure 2 A and B; Supplementary movies S1 and S2).These results indica te tha t measuring RPS in living cells r eflects r ealtime recovery of gene expression and thus TC-NER activity.In future studies, it might therefore be possible to use the RPS assay to determine and compare TC-NER activity between individual cells, which could for instance be combined with single cell sequencing or proteomics techniques to determine if any observed differences have a molecular (epi)genetic and / or proteomic basis ( 51 ).

Recovery of protein synthesis can be performed with chemicals that affect TC-NER
To determine if RPS accurately reflects TC-NER activity in response to DNA dama ging a gents other than UV irradiation, we exposed siRNA-treated and EGFP-depleted cells to cisplatin for 2 h (Figure 3 A).Cisplatin is a commonly used chemotherapeutic drug that mainly creates 1,2d(GpG) and 1,2-d(ApG) intr astr and crosslinks tha t ef fecti v ely inhibit transcription and are repaired via TC-NER (52)(53)(54)(55).Similar as after UV irradiation, we observed that the EGFP fluorescence signal in control siRNA-treated cells r ecover ed after cisplatin exposur e (Figur e 3 B and C).In contrast, no fluorescence recovery was observed at all in CSB-depleted cells after cisplatin exposure.These results illustra te the versa tile utility of the RPS assay to determine T C-NER and transcription r estart activity in r esponse to different types of genotoxic insults.
We subsequently tested if the RPS assay is compatible with pharmacological inhibition of NER (Figure 3 D).We  S6.
exposed EGFP-depleted cells for 4 h to spironolactone, which induces rapid and re v ersib le degradation of TFIIH subunit XPB (Supplementary Figure S3A) and ther efor e interferes with NER and transcription ( 56 ).Spironolactone did not strongly delay RPS in cells that were not exposed to UV irradiation, despite the importance of XPB for transcription initiation (Supplementary Figure S3B).Howe v er, increasing concentrations of spironolactone led to a correspondingly stronger RPS defect in UV irradiated cells (Figure 3 E and Supplementary Figure S3B, which depict two independent sets of experiments).This shows that the RPS assay can be effecti v ely used to monitor the impact of chemical inhibitors on DNA repair of transcription-blocking lesions.Together, these results indicate that the RPS assay is likely very suited for screening purposes.RPS could be used to screen for genotoxic compounds that block transcription and r equir e T C-NER to over come the transcription blockage, lik e cisplatin.Moreo ver, this type of screening could be combined with high-content screening approaches using genomic siRNA, shRNA or sgRNA libraries to identify the factors involved in repair of these lesions.Additionally, RPS could be used in high-content screening approaches, using fixed or living cells, to screen for compounds that inhibit transcription-coupled DNA repair.

RPS can be used as diagnostic tool for TC-NER activity with any degradable protein
After establishing the RPS assay by monitoring novel synthesis of dTAG-mediated degraded EGFP in U2OS cells, we tested if instead also other proteins and / or other cell types, including patient-deri v ed fibrob lasts, can be used for RPS.We ther efor e first attempted RPS by monitoring expression of the androgen receptor (AR) after its depletion with Bavdegalutamide (ARV-110) (Figure 4 A).ARV-110 is a PROTAC degrader that induces AR ubiquitination and subsequent degradation ( 57 ).To easily quantify AR levels, we used human hepatoma Hep3B cells stably overexpressing AR fused to EGFP and incubated with the synthetic AR ligand R1881 that induces nuclear localization of AR ( 34 , 35 ).Imaging showed that EGFP-AR protein le v els wer e alr eady slightly r educed 16 h after UV irradiation in CSB-depleted cells without the addition of ARV-110 to cells.Even more so, we observed that after ARV-110  S6.
addition, the EGFP-AR protein le v els onl y clearl y recovered in cells treated with control siRNA, but not in cells treated with siCSB (Figure 4 B).
We subsequently tested if we could similarly use this PROTAC-media ted degrada tion of endo genousl yexpressed non-tagged proteins to perform RPS in human fibroblasts of CS patients, using immunoblotting instead of fluorescence imaging.As the AR protein is not clearly e xpressed in fibrob lasts (Supplementary Figure S3C), we instead monitored RPS of the PTK2 / FAK protein (Figure 4 C), which can be efficiently degraded by exposure to the PROTAC FAK degrader 1 ( 58 ).We depleted PTK2 in TC-NER proficient control C5RO fibroblasts and in fibrob lasts deri v ed from CSB-deficient (CS1AN) and CSAdeficient (CS3BE) CS patients ( 59 ).Immunoblotting of cell lysates from these fibroblasts after UV-irradiation showed that PTK2 protein synthesis only clearly r ecover ed in the TC-NER pr oficient contr ol cells (Figure 4 D and E).These results indica te tha t monitoring RPS of any protein that can be inducibly depleted or degraded, in any cell type of choice, can be used to monitor TC-NER activity and / or transcription recovery after DNA damage induction.Moreover, the successful application of RPS in patient fibroblasts indicates that this assay could be useful for clinical diagnosis of CS, as versa tile alterna tive to RRS.As we noticed that resynthesis of PTK2 requires substantially longer time than resynthesis of ectopically expressed AR, it will be useful to test and compare additional PROTACs and their protein targets to determine which are most efficient and cost-effecti v e for potential future (clinical) applications of the RPS assay.

RPS monitors DNA r epair of tr anscription-blocking lesions in young and old tissues of living organisms
To test if the RPS assay can be applied in vivo , in differentiated cells of a living organism, we tested if novel protein synthesis in UV irradiated muscle cells of C. elegans depends on TC-NER activity (Figure 5 A).NER is well-conserved in C. elegans and particularly TC-NER is acti v e in differentia ted soma tic tissues ( 60 , 61 ).XPC-1 and CSB-1 are C. elegans orthologs of human XPC and CSB, respecti v ely.We  S6.
gener ated tr ansgenic wild type and XPC-1-and CSB-1-deficient animals expressing GFP fused to an auxininducible degradation tag (AID::GFP) and Arabidopsis TIR1 (fused to mRuby) ( 41 ) under control of the eft-3 promotor driving ubiquitous expression including in muscle cells.TIR1 forms an E3 ubiquitin ligase complex that can be activated by culturing animals on the auxin plant hormone indole-3-acetic acid, which leads to ubiquitylation of the AID tag and subsequent proteasomal degradation of AID::GFP in body wall muscle cells (Figure 5 B).48 h after UV irradiation of C. elegans cultured on auxin for 2 h (Figure 5 A), we imaged living animals by confocal microscopy and observed that in wild type and XPC-deficient animals the AID::GFP fluorescence had returned to the same le v el as that of animals not treated with auxin (Figure 5 B and C).Contrarily, AID::GFP fluorescence le v els in UV-irradiated CSB-1-deficient animals remained strongly r educed.Pr e-vious survi val e xperiments had suggested that XPC-1 can partially compensate for the lack of repair in acti v e genes in somatic cells of TC-NER-deficient C. elegans ( 42 , 62 ).In line with this, we observed that additional loss of XPC-1 in CSB-1-deficient animals further reduced AID::GFP fluorescence le v els after UV.These results confirm that XPC-1 acts partially redundant to CSB-1, as was also observed in human cells (Figure 1 E and F).
As the RPS assay accurately reflects the repair of transcription-blocking DNA damage in active genes in muscle cells, we wondered whether it can be used to re v eal changes in DNA repair activity when animals grow older, as DNA repair is proposed to decline with age ( 63 ).We ther efor e compar ed AID::GFP protein le v els after UV irradiation in one-day-old wild type adult animals to those in six-day old animals.W hereas UV-irradia ted one-dayold animals and unirradiated one-day-old and six-day-old  S6.
animals all fully r ecover ed GFP fluor escence le v els after culturing on auxin, strikingly, UV-irradiated six-day-old animals showed no fluorescence recovery at all (Figure 5 D and E).These results suggest that aging C. elegans lose their ability to repair DNA damage in acti v ely transcribed genes, at least in muscle cells.These results are in line with previously observed reduced repair in six-day-old animals using genespecific qPCR assays ( 64 ).Together, our results show that the RPS assay can be applied in living, multicellular organisms to compare the capacity to remove transcriptionblocking lesions, such as by TC-NER, between tissues, different genetic backgrounds and / or between de v elopmental stages or during aging.It will be interesting to investigate if a similar reduction of transcription-associated DNA repair capacity can be determined in aging tissues of e.g.mouse models of aging, as accum ulating DN A damage leading to transcription stress is considered one of the driving forces of the aging process ( 22 , 65 ).For such a purpose, variations to the RPS method may be used or may be further de v el-oped.A similar principle as applied here, i.e. measuring the ability of protein expression, was previously used in C. elegans to show that UV-induced de v elopmental arrest in XPA deficient animals is due to transcription blockage ( 66 ).Instead of using degradable GFP, the ability of UV-irradiated animals to express heat-shock promoter-dri v en GFP was tested.Although it may not be straightforward to use heatshock inducible promoters in mammalian cells or model systems, besides degradable proteins alternatively other inducible transcriptional reporter systems could be used to measure the ability of protein synthesis as read-out for TC-NER capacity.

Prospects and limitations of the RPS assay
Our results indicate that the RPS assay is an equally effecti v e method as the widely used RRS to monitor the capacity of cells to remove transcription-blocking DNA damage from genes and / or to restart transcription after DNA  ( 52 , 67 ).The RPS assay, howe v er, possesses a gr eater degr ee of versatility than RRS, because various read-outs such as imaging and western blotting can be used, and especially because protein recovery can be imaged in real-time in living cells and in vivo as well.Moreover, because the fluorescence signal of degron-tagged GFP is monitor ed dir ectly in fix ed or living cells, without the need of a click-chemistry reaction to couple fluorescent azide to 5ethynyluridine as in RRS ( 24 , 25 ), the RPS assay is essentially easier to perform and possibly better compatible with immunofluorescence of other proteins.Another advantage and versatile aspect of the RPS assay is that it will likely work with any protein that can be degraded, either using a degron tag or a PROTAC degrader.Ther efor e, differ ent adaptations of the assay will be best suitable for different purposes.For instance, recovery of degron-tagged GFP in combination with siRNA-mediated knockdown of proteins is suited to ra pidl y test if a protein of interest may be involved in repair of transcription-blocking DNA damage.This version of the assay will ther efor e also be suited for genetic screening purposes.Alternati v ely, it may be possib le to b leach the GFP signal instead of degrading the fluorescent protein, and image the recovery of fluorescence to monitor DNA repair activity, which may be easier for use in some in vivo a pplications.Additionall y, the compoundinduced degradation of a relati v ely long gene may be more suited for use in an RPS assay that needs to be sensiti v e to low doses of DNA damage, as a longer gene will incur more DNA dama ge.For dia gnostic purposes, howe v er, the use of PROTAC degraders and immunoblotting will likely be an easy and suitable approach, as this can be applied by laboratories that do not have adequate imaging facilities as well.Se v eral proteins and histone modifications have been implicated in transcription restart following DNA damage induction and DNA repair, rather than DNA repair itself (68)(69)(70)(71)(72).It is important to emphasize that the RPS assay, although versatile in its applicability with different cell types both in vitro and in vivo , cannot be used to distinguish between T C-NER and r estart defects.To make such a distinction, results should be compared to those obtained in assays dir ectly measuring T C-NER such as the amplified UDS in GG-NER deficient cells ( 30 ) or strand-specific ChIP-Seq ( 31 ) assays.Also, it is unclear to what extent translation itself is affected by the DNA damage doses that we used.We did not clearly observ e ne w protein synthesis after depletion of GFP for 4 h in completely NER-deficient cells, suggesting that also no translation took place from mRNA transcripts tha t alread y existed before DNA damage induction.This is striking as GFP mRNA, and thus possibly also the mRNA of GFP tagged with a degron, has a half time of more than 4 h ( 73 ).Ther efor e, this suggests that besides transcription also protein synthesis is inhibited, as also previously reported to occur after UV irradiation, e v en though this was observ ed after relati v ely high UV doses (74)(75)(76)(77)(78). Addition-ally, translation may be inhibited as a result of reduced expression of genes encoding protein synthesis factors, as observed in C. elegans after UV irradiation ( 72 ).Thus, possibly, our assay could be useful to identify factors involved in regulating the translational response to DNA damage as well.As the RPS assay is easily adjustable for use with different cell types, degradable target proteins and / or experimental readouts, it is expected that the assay will be very useful for future screening purposes utilizing living cells or e v en tissue models such as organoids or various model organisms.

DA T A A V AILABILITY
The EGFP-FKBP F36V plasmid was deposited at Addgene (#199765).Other plasmids, cell lines and C. elegans strains generated in this study are available upon reasonable request.

SUPPLEMENT ARY DA T A
Supplementary Data are available at NAR Online.

e93 18 PAGEFigure 1 .
Figure 1.Recovery of protein synthesis after DNA damage depends on transcription-coupled NER .(A) Schematic depiction of the rationale of the Recovery of Protein Synthesis (RPS) assay.First a protein (indicated by green shading in the nucleus) is degraded in cells, after which cells are exposed to a DNA dama ging a gent and the novel synthesis of the protein is monitored in time.The protein will only be produced in TC-NER proficient cells after DNA damage induction.(B) Scheme of the transgene knocked-in at the AAVS1 locus to dri v e EGFP-FKBP F36V from the PGK promoter and blasticidin (for selection) from the hEF1a promoter.(C) Representati v e images of fixed U2OS cells stably expressing EGFP-FKBP F36V , before treatment (untreated) and directly (depleted) or se v eral hours (as indicated) after treatment with dTAG13.(D) Timing scheme of the RPS assay using UV-C irradiation.(E) Representati v e images of fixed EGFP-FKBP F36V -expressing U2OS cells transfected with control siRNA (siCtrl) or siRNA against CSB (siCSB), XPF (siXPF) or XPC (siXPC) that wer e either untr ea ted, incuba ted with dTAG13 for 4 h ('depleted') or incubated with dTAG13 and irradiated with 6 J / m 2 UV-C and left to recover for 16 h ('UV + recovery').Scale bars 5 m (F) Quantification of GFP e xpression le v els from siRNA-treated cells imaged and treated as explained in (E) .Bars depict the mean with S.E.M. of individual cells measured in three independent experiments.The number of cells measured in respecti v e or der shown in the graph is 292, 274, 217, 257, 275, 181, 398, 250, 216, 326, 238, 260, 292, 291 and 206.Statistical differences were determined by one-way ANOVA with correction for multiple comparison.Numbers in the graph indica te p-values.Source da ta for the graph can be found in Supplementary TableS6.

e93 18 PAGEFigure 2 .
Figure 2. Recovery of protein synthesis monitors TC-NER activity in living cells.(A) Representative live cell imaging pictures of EGFP-FKBP F36Vexpressing U2OS cells transfected with control (siCtrl) or CSB (siCSB) siRNA.Cells were incubated with dTAG13 for 8 h, irradiated with 6 J / m 2 UV-C and imaged e v ery hour as indicated.Scale bars, 10 m (B) Quantification of GFP signal of EGFP-FKBP F36V -expressing U2OS cells treated and imaged as explained in (A) .Depicted is the mean and S.E.M. of individual cells measured in two independent experiments.The number of measured cells in respecti v e or der shown in the graph is 108, 125, 133, 138, 142, 136, 146, 129, 118 and 119 for siCTRL, and 98, 95, 99, 98, 100, 98, 96, 110, 100 and 89 for siCSB.Sta tistical dif fer ences wer e determined with unpair ed t-test.Numbers in the graph indicate p-values.Source data for the graph can be found in Supplementary TableS6.

PAGE 7 OF 12 NucleicFigure 3 .
Figure 3. Recovery of protein synthesis after cisplatin or chemical NER inhibition.(A) Timing scheme of the Recovery of Protein Synthesis assay using cispla tin.(B) Representa ti v e images of fixed EGFP-FKBP F36V -expressing U2OS cells transfected with control (siCtrl) or CSB (siCSB) siRNA that were either untrea ted, incuba ted with dT AG13 for 4 h ('depleted') or with dT AG13 f or 4 h and 100 M cisplatin f or 2 h and left to recover f or 16 h ('Cisplatin + recovery').Scale bars, 5 m (C) Quantification of GFP expression levels from siRNA-treated cells imaged and treated as explained in (B) .Mean and S.E.M. of three independent experiments.(D) Timing scheme of the Recovery of Protein Synthesis assay using spironolactone.(E) Quantification of GFP e xpression le v els fr om mock or spir onolactone-tr eated EGFP-FKBP F36V -expr essing U2OS cells that wer e either untr ea ted, incuba ted with dTAG13 for 4 h ('depleted') or with dTAG13 for 4 h and then irradiated with 6 J / m 2 UV-C and left to recover for 16 h ('UV + recovery').Bars depict the mean with S.E.M. of individual cells measured in three independent experiments.The number of cells measured in respecti v e or der shown in the graph is (C) 288, 310, 250, 330, 251 and 199.(E) 362, 344, 263, 391, 320, 191, 334, 257 and 192.Sta tistical dif fer ences wer e determined by one-way ANOVA with correction for multiple comparison.Numbers in the graphs indicate p-values.Source data for the graphs can be found in Supplementary TableS6.

e93 18 PAGEFigure 4 .
Figure 4. Novel synthesis of PROTAC-degraded proteins to monitor TC-NER.(A) Timing scheme of the Recovery of Protein Synthesis assay using ARV-110.(B) Quantification of GFP fluorescence le v els in EGFP-AR-e xpr essing Hep3B cells that wer e either untr ea ted or incuba ted with 100 nM ARV-110 for 4 h ('depleted'), incubated with 100 nM ARV-110 for 4 h and then irradiated with 6 J / m 2 UV-C and left to recover for 16 h ('UV + recovery'), incubated with 100 nM M ARV-110 for 4 h and left to recover for 16 h ('no UV + recovery') or irradiated with 6 J / m 2 UV-C and left to recover for 16 h ('No depletion + UV + recovery').Bars depict the mean with S.E.M. of individual cells measured in three independent experiments.The number of cells measur ed in r especti v e or der shown in the graph is (B) 181, 157, 184, 159, 115, 163, 161, 211, 179 and 55. (C) Timing scheme of the Recovery of Protein Synthesis assay using PROTAC FAK degrader 1. (D) Immunoblot analysis of cell lysate from SV40-immortalized C5RO, CS1AN and CS3BE cells that were either untreated or incubated with 250 nM PROTAC FAK degrader 1 for 24 h ('depleted'), or incubated with 250 nM PROTAC FAK degrader 1 for 24 h and irradiated with the indicated UV dose and left to recover for 72 h ('indicated UV dose + recov ery').Immunob lots are stained with antibodies against PTK2 and Tubulin (as loading control).(E) Quantification of protein le v els based on immunoblots as shown in (D).Mean and S.E.M. of four (C5RO) or three (CS1AN and CS3BE) independent experiments.Statistical differences were determined by one-way ANOVA with correction for multiple comparison.Numbers in the graphs indicate p-values.Source data for the graphs can be found in Supplementary TableS6.

Figure 5 .
Figure 5. Recovery of protein synthesis assay in muscle cells of C. elegans.(A) Timing scheme of the Recovery of Protein Synthesis assay using UV-irradiated C. elegans.Depicted is a schematic drawing of the head of C. elegans with body wall muscle cells expr essing gr een color ed GFP.(B) Repr esentati v e confocal images of living wild type, csb-1 and xpc-1; csb-1 animals expressing AID::GFP (and TIR1::m-Ruby, not depicted) under control of the eft-3 promoter in body wall muscles, shown here in the head of C. elegans .Animals were either untreated, cultured on 100 M auxin for 2 h ('depleted') or cultured on 100 M auxin for 2 h and then irradiated with 120 J / m 2 UV-B and left to recover for 48 h ('UV + recovery').Scale bar, 25 m.(C) Quantification of GFP fluorescence le v els in muscle cells of wild type, xpc-1, csb-1 and xpc-1; csb-1 animals e xpr essing AID::GFP and tr eated and imaged as explained in (B).Mean and S.E.M. of three independent e xperiments.(D) Representati v e confocal images of living one-day old and six-day old wild type animals expressing AID::GFP in body wall muscles, shown here in the head of C. elegans .Animals were either untr eated, cultur ed on 100 M auxin for 2 h ('depleted'), cultured on 100 M auxin for 2 h and then irradiated with 120 J / m 2 UV-B and left to recover for 48 h (UV + recovery) or cultured on 100 M auxin for 2 h and left to recover for 48 h ('no UV + recovery').Scale bar, 25 m.(E) Quantification of GFP fluorescence le v els in muscle cells of one-day-old and six-day-old wild type animals expressing AID::GFP and treated and imaged as explained in (D).Bars depict the mean with S.E.M. of individual cells in C. elegans measured in three independent experiments.In each independent experiment at least three animals were tested.The number of cells measured in respecti v e or der shown in the graph is (C) 229, 138, 207, 178, 83, 167, 275, 151, 209, 263, 183 and 310 (E) 85, 75, 84, 70, 86, 68, 88 and 66.Statistical differ ences wer e determined by one-way ANOVA with correction for multiple comparison.Numbers in the graphs indicate p-values.Source data for the graphs can be found in Supplementary TableS6.

e93
Nucleic AcidsResearch, 2023, Vol.51, No. 18PAGE 10 OF 12 damage induction.Both the protein synthesis recovery in control cells as well as the lack of recovery in TC-NER deficient cells are equally robust and reproducible as transcription r ecovery measur ed by RRS, such as e.g.previously determined by RRS in our laboratory in control and TC-NER deficient cells using similar UV (6 J / m 2 ) or cisplatin (100 M) doses