XAB2 Dynamics during DNA Damage-Dependent Transcription Inhibition

Lise-Marie DONNIO

1,#

, Elena CERUTTI

1,#

, Charlène MAGNANI

, Damien NEUILLET

, Pierre-Olivier MARI

and Giuseppina GIGLIA-MARI

1,*

1 : Institut NeuroMyogène (INMG), CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1,

8 avenue de Rockefeller, 69008 LYON

* To whom correspondence and request for materials should be addressed.

G G-M : PHONE : 04 26 68 82 63

EMAIL : ambra.mari@univ-lyon1.fr

# L.-M.D. and E.C. contributed equally to this work

AUTHOR CONTRIBUTIONS: P.-O.M. and G.G.-M. designed research; L.-M.D.; E.C.; D.N.; C.M. performed

research; P.-O.M., E.C., L.-M.D. and G.G-M. analyzed data; and L.-M.D., E.C. and G.G.-M. wrote the paper.

COMPETING INTEREST STATEMENT: The authors declare no conflict of interest.

CLASSIFICATION: biological sciences; Cell Biology

KEYWORDS: XAB2; CSA; CSB; pre-mRNA splicing; TC-NER; RNA polymerase 2; R-loops

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

ABSTRACT

Xeroderma Pigmentosum group A (XPA)-binding protein 2 (XAB2) is a multi-functional protein

that plays a critical role in distinct cellular processes including transcription, splicing, DNA repair and

mRNA export. In this study, we detailed XAB2 involvement during Nucleotide Excision Repair (NER), a

repair pathway that guarantees genome integrity against UV light-induced DNA damage and that

specifically removes transcription-blocking damage in a dedicated process known as Transcription-

Coupled repair (TC-NER). Here, we demonstrated that XAB2 is involved specifically and exclusively in TC-

NER reaction and solely for RNA Polymerase 2 transcribed genes. Surprisingly, contrary to all the other

NER proteins studied so far, XAB2 does not accumulate on the local UV-C damage but on the contrary is

remobilized after damage induction. This fast change in mobility is restored when DNA repair reactions

are completed. By scrutinizing from which cellular complex/partner/structure XAB2 is released, we have

identified that XAB2 is detached after DNA damage induction from the DNA:RNA hybrids, commonly

known as R-loops, and from the CSA and XPG protein and this release is thought to contribute to the

DNA damage recognition step during TC-NER. Importantly, we have disclosed a role for XAB2 in retaining

RNAP2 on its substrate.

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

INTRODUCTION

The DNA molecule, contained in the nucleus of our cells, forms the instruction manual for proper

cellular functioning. Unfortunately, the integrity of our DNA is continuously challenged by a variety of

endogenous and exogenous agents (e.g. ultraviolet light, cigarette smoke, environmental pollution,

oxidative damage, etc …). These DNA lesions interfere with DNA replication, transcription and cell cycle

progression and lead to mutation and cell death, which may cause cancer, inborn disease or aging

(Chatterjee & Walker, 2017).

To prevent the deleterious consequences of persisting DNA lesions, all organisms are equipped

with a network of efficient DNA repair systems. One of these systems is the Nucleotide Excision Repair

(NER) which removes helix-distorting DNA adducts caused by ultraviolet light (UV) such as Cyclo-

Pyrimidine Dimers (CPD) and 6-4 Photoproducts (6-4PP) (Giglia-Mari et al, 2011).

In mammals, the different steps of NER require about thirty different proteins that are recruited

one by one into the DNA damage, as demonstrated by the different studies of NER proteins kinetics

(Moné et al, 2004; van den Boom et al, 2004; Politi et al, 2005; Zotter et al, 2006; Rademakers et al,

2003). The first step of NER consists of damage recognition, followed by the opening of the DNA duplex,

dual incisions on both sides of the damage, excision of 24–32 oligonucleotides containing the damage

and, finally, gap filling by repair DNA synthesis.

NER is divided into two sub-pathways depending on where DNA lesions are located within the

genome. The global genome repair (GG-NER) will detect and repair lesion throughout the genome,

whereas transcription-coupled repair (TC-NER) is associated with RNA polymerase II (RNAP2) to repair

lesions on the transcribed strand of active gene (Marteijn et al, 2014).

The NER system has been linked to rare human diseases classically grouped into three distinct

NER-related syndromes. These include the highly cancer prone disorder Xeroderma Pigmentosum (XP)

and the two progeroids diseases: Cockayne Syndrome (CS) and Trichothiodystrophy (TTD). Importantly,

CS and TTD patients are not cancer-prone but present severe neurological and developmental features

(Hanawalt, 1994).

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

Xeroderma Pigmentosum group A (XPA)-binding protein 2 (XAB2) is an evolutionarily highly

conserved protein of 100 kDa and consists of fifteen tetratricopeptide repeat (TPR) motifs that play a

role in protein-protein interactions. XAB2 protein was identified as a protein interacting with XPA, a NER

factor, using a yeast two-hybrid system (Nakatsu et al, 2000). Next, it has been shown that this protein

interacts also with the TC-NER-specific factors, CSA and CSB, as well as elongating RNAP2 (Nakatsu et al,

2000). XAB2 is also essential for early mouse embryogenesis as demonstrated by the preimplantation

lethality observed in XAB2 knockout mice (Yonemasu et al, 2005).

Downregulation of XAB2 using either anti-XAB2 or siRNA specifically inhibited normal RNA

synthesis and the recovery of RNA synthesis after UV irradiation (Nakatsu et al, 2000; Kuraoka et al,

2008). Furthermore, injection of anti-XAB2 in GG-NER deficient cells results in a significant reduction of

UV-induced Unscheduled DNA Synthesis during repair (Nakatsu et al, 2000). These results suggest the

involvement of XAB2 in transcription and in TC-NER.

Further studies have shown that XAB2 is a component of Prp19/XAB2 complex (Aquarius [AQR],

XAB2, Prp19, CCDC16, hISY1 and PPIE) or Prp19/CDC5L-related complex required for pre-mRNA splicing

(Kuraoka et al, 2008). XAB2, as well as PRP19 and AQR, has been involved in the DNA damage response

(Onyango et al, 2016; Maréchal et al, 2014; Sakasai et al, 2017). Indeed, XAB2 is important for

homologous recombination (HR) by promoting the end resection step (Onyango et al, 2016); PRP19 is a

sensor of RPA-ssDNA after DNA damage (Maréchal et al, 2014) and AQR contributes to the maintenance

of genomic stability via regulation of HR (Sakasai et al, 2017).

Moreover, AQR has also a role in the removal of R-loops, a three-stranded nucleic acid structure

composed of a DNA:RNA hybrid and the associated non-template single-stranded DNA (Sollier et al,

2014). These structures can form during transcription when an RNA molecule emerging from the

transcription machinery hybridizes with the DNA template. They are found abundantly in human gene

promoters and terminators where RNA processing takes place (Wang et al, 2018).

Despite the knowledge acquired in the last decades on XAB2 and its different cellular roles, little

is known on the exact crosstalk and dynamics between its diverse cellular functions, specifically between

the DNA repair function, the involvement in transcription and its splicing activity. In this work, we

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

described the molecular dynamics of XAB2 within the cell after UV-damage induction and during the TC-

NER repair process. We determine in vivo that, in absence of XAB2, the transcription-coupled repair

reaction is impaired and hence the transcription restart after UV damage is abolished. Surprisingly,

unlike all the other NER proteins studied so far, the mobility of XAB2 is increased after DNA damage and

XAB2 shows no accumulation on the local UV damage. This faster dynamic is not restored until DNA

repair is completed and in damaged TC-NER-deficient cells, XAB2 remains more mobile. Interestingly, we

demonstrate that, after DNA damage induction, XAB2 is not released from the splicing complex but is

detached from R-loops, a recently identified substrate of XAB2 (Goulielmaki et al, 2021).

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

RESULTS

XAB2 is involved in TC-NER process

Two decades ago, Tanaka’s research group demonstrated the involvement of XAB2 in the NER

pathway (Kuraoka et al, 2008; Nakatsu et al, 2000). However, the dynamics of XAB2 during the DNA

Repair process remained to be elucidated. We aimed to study the molecular dynamics of XAB2 and the

shuttling between its different functions when DNA damage is induced. Firstly, we wanted to verify that

XAB2 is exclusively involved in the TC-NER reaction and not in other steps or pathways repairing UV-

lesions.

The well-known standard assay used to quantify NER activity is the Unscheduled DNA Synthesis

(UDS), which measure replication activity outside of the S-phase after UV treatment. This technique

quantifies the refilling of the single-strand DNA gap by the DNA replicative machinery. When we

performed UDS assay in XAB2 silenced cells, no decreased level of UDS was observed (as well as in mock-

treated cells) (Figure 1A, blue and black columns and Figure S1). As a positive control, when we silenced

the excision repair factor XPF, we observed a strong reduction in UDS level (Figure 1A, red column and

Figure S1). This result shows that XAB2 is not involved in GG-NER sub-pathway, but does not exclude an

involvement of XAB2 in TC-NER sub-pathway.

The commonly used assay measuring TC-NER activity is the RNA Recovery Synthesis (RRS). In this

assay, the newly transcribed RNA is measured via the incorporation of a nucleoside analog coupled to a

fluorophore. The experiment is conducted at different time point after UV irradiation (0, 3, 16 and 24h),

in order to quantify, 3h after UV damage, the decline in transcriptional activity and 16-24h after UV

irradiation, the restart of transcriptional activity after DNA repair. When we performed this assay in

XAB2 silenced cells no restart of transcription after UV damage was observed (Figure 1B, blue column

and Figure S2A), as well as in siXPF-treated cells due to the inability to repair DNA lesions (Figure 1B, red

column and Figure S2A) and in contrast with siMock-treated cells (Figure 1B, black columns and Figure

100

S2A). Surprisingly, contrary to Tanaka’s group which observed a decreased of nascent RNA synthesis in

101

absence of XAB2, we observed in our experiment an increase of transcription in siXAB2-treated cells

102

without irradiation (Figure S2B) (Kuraoka et al, 2008; Nakatsu et al, 2000). This result demonstrated an

103

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

involvement of XAB2 in the TC-NER sub pathway, but did not discriminate between a role in the repair

104

reaction per se or in the restart of transcription after repair (RTR) (Mourgues et al, 2013).

105

In order to discriminate this point, we performed an assay designed previously in our group that

106

specifically measures repair replication during TC-NER: the TCR-UDS assay (Mourgues et al, 2013). For

107

this assay, we performed UDS assay in GG-NER deficient cells using XPC (Xeroderma Pigmentosum

108

complementation group C) mutant cells (XP4PA-SV). The cells were transfected with siRNA and then

109

locally irradiated with UV-C through a filter. In order to precisely localize DNA-damaged areas, a yH2AX

110

co-immunofluorescence labeling was performed, and repair-replication was quantified. In siXPF-treated

111

XPC-negative cells, both the GG-NER and the TC-NER pathways are compromised and as expected low

112

TCR-UDS levels were observed compared to siMock-treated cells (Figure 1C, red and black column and

113

Figure S3). Knockdown of XAB2 results in a decrease of TCR-UDS levels (Figure 1C, blue column and

114

Figure S3). This result demonstrates a role of XAB2 in the repair reaction itself, its silencing preventing

115

the DNA-synthesis associated with the excision of UV lesions on actively transcribed genes.

116

Recently we demonstrate that a fully functional NER mechanism is necessary for repair of

117

ribosomal DNA (rDNA), genes transcribed by the RNA polymerase I (RNAP1) (Daniel et al, 2018). To

118

investigate the involvement of XAB2 in the repair of ribosomal genes, the level of RNAP1 transcription

119

was measured at different time points after UV irradiation by using a specific ribosomal RNA probe

120

coupled to a fluorophore, as described previously (Daniel et al, 2018). This probe recognizes the 5’ end of

121

the rDNA transcript, the 47S pre-rRNA (upstream from the first site cleaved rapidly during rRNA

122

processing) (a sketch of the 47S is depicted in figure S4A). In siMock treated cells, we observed a

123

decrease of 47S levels 3h after UV-C exposure and the restart of RNAP1 transcription 40h after

124

irradiation (Figure 1D, black column and Figure S4C). The TC-NER-deficient siCSB-treated cells presented

125

a low level of rRNA synthesis even 40h after UV-C exposure (Figure 1D, violet column and Figure S4B and

126

S4C). A restart of RNAP1 transcription was observed in the absence of XAB2, as for the control (Figure

127

1D, blue column and Figure S4C).

128

All these results demonstrated a function of XAB2 in the TC-NER repair reaction specifically and

129

exclusively for RNAP2 transcribed genes.

130

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

131

Figure 1. XAB2 involvement in DNA repair

132

A) Quantification of Unscheduled DNA Synthesis assay (UDS) determined by EdU incorporation after

133

local damage (LD) induction with UV-C (100J/m²) in WT cells (MRC5 cells) treated with siRNAs against

134

indicated factors. The graph is the average of two independent experiments and error bars represent the

135

SEM obtained from at least 30 LDs. B) Quantification of RNA Recovery Synthesis (RRS) assay determined

136

by EU incorporation after UV-C (10J/m²) exposure in WT cells treated with siRNAs against indicated

137

factors. Error bars represent the SEM obtained from at least 50 cells and data are representative of three

138

independent experiments. C) Quantification of TCR-UDS assay determined by EdU incorporation after LD

139

induction with UV-C (100J/m²) in GG-NER deficient cells (XPC-/- cells) treated with siRNAs against

140

indicated factors. Error bars represent the SEM obtained from at least 15 LDs and data are

141

representative of two independent experiments. D) Quantification of RNA-FISH assay showing the 47S

142

pre-rRNA level after UV-C (16J/m²) exposure in WT cells treated with siRNAs against indicated factors.

143

Error bars represent the SEM obtained from at least 30 cells and data are representative of two

144

independent experiments. For all graph, p-value of student’s test compared to No UV or siMock

145

condition : ***<0.001.

146

147

148

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

XAB2-splicing complex mobility is released from the DNA damage area.

149

XAB2 is included in a splicing complex composed of five other proteins, namely Aquarius (AQR),

150

PRP19, CCDC16, PPIE and ISY1 (Kuraoka et al, 2008). In order to explore how the XAB2 splicing complex is

151

behaving after local damage induction, the localization of XAB2, AQR, PRP19 and CCDC16 was revealed

152

by immunofluorescence assays at different time points after local UV-irradiation of the cells. In this

153

assay, the fluorescence signal from each protein in the damaged area (visualized by a co-staining of

154

yH2AX) was compared to the rest of the nucleus. Unexpectedly, in contrast with all other NER proteins

155

studied so far, we observed a relatively rapid (1 hour after UV-irradiation) release of XAB2, AQR, PRP19

156

and CCDC16 from the damaged area (Figure 2A and 2B). The localization of the XAB2-splicing complex is

157

re-established after the completion of DNA repair reactions, when the transcription is fully restarted, 16h

158

after irradiation (Figure 2B).

159

We thus verified if the entire XAB2-splicing complex was involved in TC-NER or whether only

160

XAB2 played a role in this process. In order to measure the repair capacity of cells silenced for XAB2-

161

related proteins, we performed UDS, TCR-UDS and RRS experiments in AQR/PRP19/CCDC16/PPIE/ISY1-

162

siRNAs treated cells (Figure S5, S6, S7 and S8) and compared the results with XPF-siRNAs treated cell

163

lines. Our results clearly show that none of the cells silenced for XAB2-related proteins are deficient in

164

DNA Repair and both GG-NER (Figure S6) and TC-NER (Figure S7 and S8) are proficient in absence of

165

AQR/PRP19/CCDC16/PPIE/ISY1.

166

In order to investigate whether XAB2 release from damaged areas was dependent of the TC-NER

167

reaction, the localization of XAB2 was detected and quantified within locally damaged areas in TC-NER-

168

deficient cells, CSA (CS3BE) and CSB (CS1AN) mutant cells. Interestingly, the absence of CSA and CSB did

169

not hinder the release of XAB2 from locally damaged areas but this release continued 16h after UV-C

170

exposure (Figure 2C blue and red curve compared to black curve and figure S9), suggesting that the re-

171

establishment of the proper localization of XAB2 within the nucleus after the DNA repair process

172

depends either on the repair process per se or on the restart of transcription after completion of the

173

DNA repair reactions.

174

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

175

Figure 2. Splicing complex released from DNA damage

176

A) Representative confocal images of immunofluorescence against XAB2, AQR, PRP19 or CCDC16 (green)

177

and γH2AX (red) one hour after local damage (LD) induction with UV-C (60J/m²). LD are indicated by

178

dashed lines. Scale bar 3µm. B) Quantification of the immunofluorescence signal of the different splicing

179

proteins on the LD after different times of recovery. C) Quantification of XAB2 signal on LD in different

180

cell lines after different times of recovery. For both graph, the signal from the local damage has been

181

subtracted from the background of each cell. Error bars represent the SEM obtained from at least 20

182

cells and data are representative of two independent experiments.

183

184

185

XAB2 dynamic during TC-NER

186

To further analyze the mobility of XAB2 within the nuclei, we performed SPOT-FRAP (Fluorescent

187

Recovery After Photo-Bleaching) experiments. In this technique, fluorescence molecules are photo-

188

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

bleached in a small spot by a high intensity laser pulse; then the recovery of fluorescence within the

189

bleached area is monitored over time (Figure S10A). With no treatment, this measure of fluorescence

190

recovery corresponds to the protein mobility within the living cells (Figure S10A black curve). After

191

perturbation of the nuclear environment (e.g. DNA damage), a protein can become less mobile if it

192

physically interacts with a new substrate or a bigger complex (Figure S10A, green curve); more mobile if

193

the protein is released from its substrate (Figure S10A, blue curve); or eventually will not change its

194

mobility (Figure S10A red curve).

195

We stably transfected a vector expressing a fluorescently version of XAB2 (XAB2-GFP, Figure

196

S10B) in different SV40-immortalized human fibroblast: wild-type cells (MRC5, Figure S10C), CSA

197

deficient cells (CS3BE) and CSB deficient cells (CS1AN). In order to determine the minimum dose of UV-C

198

needed to detect a significant difference in XAB2 mobility, MRC5 XAB2-GFP cell were irradiated with

199

doses of UV-C ranging from 2 to 16 J/m² and SPOT-FRAP experiments were performed at different time

200

point after irradiation (Figure 3A). Interestingly, we observed a dose-dependent increase in mobility of

201

XAB2 (Figure 3A). Doses of UV-C as weak as 2 and 4 J/m² induced a moderate increase in mobility 3

202

hours post-irradiation and a re-establishment of the basal XAB2 mobility within 16 hours post-irradiation

203

(Figure 3A, blue and yellow bar). High UV-C doses (16 J/m²) induced an increased mobility of XAB2 more

204

rapidly (1-hour post-irradiation) and a re-establishment of the control mobility 24 hours post-irradiation

205

(Figure 3A, red bar). At intermediate doses of 8 J/m² of UV-C, we observed a significant increase in XAB2

206

mobility during repair (3h after UV-C exposure) and the following return to the normal condition once

207

repair is completed and transcription restarted (16h after irradiation) (Figure 3A, green bar).

208

Interestingly, in CSA and CSB mutant cells, the increase in XAB2 mobility is also observed after 8 J/m²

209

irradiation but lasted until 24h after UV-C exposure (Figure 3B) witnessing the fact that in these cells

210

DNA damage is not repaired and hence initial XAB2 mobility is not restored. Interestingly, without

211

damage, XAB2 mobility is reduced in TC-NER deficient cells compared to wild-type cells for still unknown

212

reasons (Figure 3B, black histogram).

213

The results of these experiments directed us to explore the possibility that the change in XAB2

214

mobility was due to transcription inhibition and not really to the repair process itself. In order to verify

215

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

this hypothesis XAB2 mobility was measured after DRB (transcription inhibitor) treatment. The results

216

show that XAB2 increased mobility in transcription inhibition conditions is very similar to the one

217

measured upon UV treatment (Figure 3C).

218

As XAB2, the mobility of the late-stage spliceosomes change after UV irradiation and this

219

mobilization depends on DNA damage response (DDR) signaling pathways (Tresini et al, 2015). Key

220

mediators of DDR are the ATM and ATR kinases, which induced cell cycle arrest and facilitate DNA repair.

221

To demonstrate that the change of XAB2 mobility is not due to the UV-Damage Response, we realized

222

the same measures in presence of ATR and ATM inhibitors. Both drugs did not modify the increase of

223

XAB2 mobility after UV irradiation (Figure 3C and Figure S10D).

224

Taken together, these results demonstrated that XAB2 mobility changes after DNA damage is

225

triggered and sustained by the transcriptional inhibition.

226

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

227

Figure 3. XAB2 dynamic during TC-NER

228

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

A) FRAP analysis of XAB2-GFP mobility in WT cells. Cells were treated or not with different doses of UV-C

229

(2 to 16J/m²) and XAB2 mobility was measured at different time points after UV-C exposure. The No UV

230

condition was used to calculate the change in bound fraction. B) FRAP analysis of XAB2-GFP expressed in

231

WT cells (MRC5-SV) and TC-NER deficient cells (CSA -/- and CSB -/-). Cells were treated or not with 8J/m²

232

of UV-C. The No UV condition of the WT cell lines was used to calculate the change in bound fraction. C)

233

FRAP analysis of XAB2-GFP mobility in WT cells. Cells were either untreated (dark line) or treated with

234

100µg/ml of DRB for 2h (red line) or with 10J/m² of UV-C for 3h (blue line). Inhibitor of ATR pathway was

235

added at 10µM in the medium 1h before irradiation (purple line). For all graph, error bars represent the

236

SEM obtained from at least 10 cells and data are representative of two independent experiments.

237

238

239

XAB2 is not released from the splicing complex during DNA repair reactions.

240

The increase of XAB2 mobility after UV-induced transcription inhibition could be explained by

241

either the release of XAB2 from a bigger complex and/or the release from an immobile (or nearly

242

immobile) substrate such as the chromatin or a DNA-related substrate.

243

In order to distinguish between these two possibilities, we firstly investigated whether XAB2

244

dissociates after DNA damage induction from the splicing complex described earlier. XAB2 was

245

immunoprecipitated together with AQR (Figure S11A). Interestingly, XAB2 was strongly and consistently

246

immunoprecipitated 1-hour post-irradiation which corresponded to the time in which XAB2 mobility is

247

increased and at the same time point more AQR is also immunoprecipitated. No clear release of XAB2

248

from AQR was observed at different time points.

249

In parallel, we also verified by PLA (Proximity Ligation Assay) whether the binding of XAB2 to

250

AQR was modified after UV-C irradiation (Figure S11B). 2 hours after UV-C irradiation, instead of a

251

release of XAB2 from AQR, we measured a stronger interaction (Figure S11B). However, this stronger

252

interaction could very well result from the increase of AQR concentration 1h after UV irradiation (Figure

253

S11C).

254

In conclusion, immunoprecipitation or PLA experiment could not explain the increased mobility

255

of XAB2 after DNA damage, indeed XAB2 is not released from splicing complex during DNA repair

256

reactions.

257

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

258

259

XAB2 is released from R-loops during DNA repair reactions.

260

Interestingly, while trying to immunoprecipitate XAB2 interacting partners during DNA Repair

261

reactions, we could observe that systematically and consistently more XAB2 could be immuno-

262

precipitated from nuclear extracts 1 to 3 hours after UV-C irradiation in WT cells (Figure S11A and 4A). At

263

16 hours post irradiation, we observed that the amount of XAB2 immunoprecipitated was comparable to

264

the level observed in non-irradiated cells. Moreover, in CSA-/- and CSB-/- cell lines, in which UV-lesions

265

on transcribed strands of genes are not repaired, the amount of XAB2 immunoprecipitated was

266

remained high all along the time course of the experiment (3 and 16 hours) (Figure 4A). These results

267

suggest that in TCR-deficient cells the binding between XAB2 and its substrate is not restored.

268

AQR has a role in the removal of DNA:RNA hybrids, commonly known as R-loops structure

269

(Sollier et al, 2014) and recently XAB2 was found to be involved in the R-loops resolution (Goulielmaki et

270

al, 2021). We thus decided to investigate if XAB2 has also a link with R-loops, a DNA-related substrate.

271

Indeed, we could find by immunoprecipitation that XAB2 directly interacts with R-loops (Figure 4B) and

272

that in cells silenced for XAB2 or AQR compared to control cells, we observed an accumulation of R-loops

273

(Figure 4C and Figure S12A). We thus decided to examine whether XAB2 is released from R-loops after

274

DNA damage induction by performing a PLA assay using a specific R-loops antibody (S9.6). We measured

275

a strong and consistent reduction, more than 40% reduction, of the interactions between R-loops and

276

XAB2 and between R-loops and AQR (Figure 4D and Figure S12B). This reduced interaction is not caused

277

by a reduction in either R-loops, XAB2 or AQR concentration during DNA repair (Figure 4E and figure

278

S12B). To verify that this result is specific for RNA-loops and not for a non-specific signal coming from the

279

interaction of XAB2 with messenger RNAs, we performed the same assays (PLA and IF) in presence of

280

RNAse H which specifically degrades R-loops structures and not the RNA (Figure S13A and S13B). Our

281

results show that RNAse H decreases the quantity of RNA-loops (Figure S13C-F) and in doing so it

282

decreases also the interaction with both XAB2 (Figure S13C and S13D) and AQR (Figure S13E and S13F).

283

To verify whether XAB2 increase in mobility is due to a decreased interaction with the mRNA we

284

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

performed PLA assays in UV-irradiated cells (Figure S14A and S14C). A decrease in the interaction

285

between XAB2 and mRNA was observed (Figure S14A and S14C) but this decrease corresponded to the

286

decrease of mRNA caused by UV-dependent transcription inhibition (Figure S14B and S14C) invalidating

287

the results observed in the PLA XAB2-mRNAs (Figure S14A).

288

These results clearly demonstrate that XAB2 is released from R-loops during DNA repair

289

reactions.

290

291

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

292

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

Figure 4. XAB2 and AQR are released from R-loop during DNA repair

293

A) Immunoprecipitation of XAB2 in nuclear extract from different cell lines treated with 10J/m² of UV-C

294

at different times. Bound proteins were revealed by Western blotting with antibodies against XAB2.

295

INPUT, 10% of the lysate used for immunoprecipitation (IP) reaction. B) Immunoprecipitation of R-loops

296

in non-crosslinked chromatin extract from WT cells. XAB2 bound to R-loops was revealed by Western

297

blotting. INPUT, 10% of the lysate used for immunoprecipitation (IP) reaction. C) Quantification of

298

immunofluorescence (from at least 50 cells) against R-loops in WT cells treated with siRNAs against

299

indicated factors. D-E) Quantification of fluorescent signal in the nucleus against the couple XAB2_R-

300

loops or AQR_R-loops from PLA experiment (D) or from the immunofluorescence done in parallel of PLA

301

assay (E). Error bars represent the SEM obtained from at least 50 cells.

302

For all graph, data are representative of two independent experiments. For R-loops quantification, the

303

nucleolar signal was subtracted from the nuclear signal. P-value of student’s test compared to No UV or

304

siMock condition: **<0.01;***<0.001.

305

306

307

XAB2 is released from CSA during DNA repair.

308

Because XAB2 has been found to participate specifically in TCR-NER repair reactions (Figure 1C),

309

we wanted to investigate whether part of the increased mobility of XAB2 observed after UV-induction

310

was due to a release from repair complexes. We measured, by PLA, the amount of interactions between

311

XAB2 and CSA, CSB, XPB and XPG proteins, observed during TC-NER. Amongst all the proteins tested, we

312

could observe a clear and consistent release from the CSA protein 2h after UV irradiation (Figure 5A) and

313

from the XPG protein 1h after UV irradiation (Figure 5C). The corresponding IF did not show a decreased

314

quantity of CSA or XPG (Figure 5B and 5D) which validated the specificity of the XAB2-CSA and XAB2-XPG

315

decreased interaction at those time points. On the contrary, no clear decrease in interaction was

316

observed between XAB2 and CSB (Figure S15A and S15B) or XPB (figure S15C and S15D).

317

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

318

Figure 5: XAB2 are released from CSA and XPG during DNA repair

319

A to D) Quantification of fluorescent signal in the nucleus against the couple XAB2-CSA (A and B) and

320

XAB2-XPG (C and D) from PLA experiment (A and C) or from the immunofluorescence done in parallel of

321

PLA assay (B and D). Data are the average of at least two independent experiments and error bars

322

represent the SEM obtained from at least 80 cells. P-value of student’s test compared to No UV

323

condition: *<0.05; **<0.01; ***<0.001.

324

325

326

XAB2 depletion modify the mobility of the RNAP2.

327

Because XAB2 was found to interact with RNAP2 (Nakatsu et al, 2000; Kuraoka et al, 2008) and

328

because the increased mobility observed is dependent on the UV-dependent transcription inhibition

329

step, we wanted to explore whether the depletion of XAB2 might play a role in the overall mobility of

330

RNAP2. In order to investigate this point, we performed FRAP experiments on RNAP2-GFP expressing

331

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

cells in presence and absence (depletion) of XAB2 (Figure 6) (Donnio et al, 2019). Interestingly, depletion

332

of XAB2 greatly increase the mobility of RNAP2 (Figure 6, dark green curve versus dark blue curve).

333

Moreover, UV-irradiation intensify this remobilization (Figure 6, light green curve versus light blue

334

curve), demonstrating that XAB2 might play a role in maintaining the RNAP2 bound to the substrate in

335

absence or presence of UV-lesions on the DNA.

336

337

Figure 6: Pol2 behavior with XAB2

338

Strip-FRAP analysis of RNAP2-GFP expressing WT cells treated or not with UV-C (10J/m²) after siRNA

339

mediated knockdown of the indicated factors. Error bars represent the SEM obtained from at least 10

340

cells and data are representative of two independent experiments.

341

342

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

DISCUSSION

343

Helix distorting lesions continuously challenge the cell survival by interfering with and blocking

344

fundamental cellular functions, such as transcription and replication. In order to prevent the deleterious

345

effects of these events, cells have developed different mechanisms to restore an undamaged DNA

346

molecule and allow the restart of cellular processes. The importance of rapidly reestablishing perturbed

347

cellular functions is underlined by the presence of a repair mechanism directly coupled with

348

transcription, like the Transcription-Coupled Nucleotide Excision Repair (TC-NER).

349

Two phases can be distinguished during TC-NER events: (i) the actual repair reaction of the

350

damaged strand via the TC-NER sub-pathway and (ii) the resumption of transcription after repair. The

351

experiment known as ‘RNA Recovery Synthesis’ (RRS) allow measuring the restart of transcription after

352

DNA repair completion and thus the involvement of a protein in the TC-NER process. However, this assay

353

does not discriminate between the two phases of the TC-NER. By consequence, proteins that are fully

354

proficient in the repair reaction but fail in the transcription restart after DNA repair completion might

355

have the same defect in RRS as proteins that are fully deficient in the DNA repair reaction.

356

Up to now, all studies have demonstrated the involvement of XAB2 in TC-NER process using only

357

RRS experiments (Nakatsu et al, 2000; Kuraoka et al, 2008). In this study, we used a specific test called

358

TCR-UDS, developed in house (Mourgues et al, 2013), and demonstrated that indeed XAB2 is solely

359

needed for the repair reaction per se (Figure 1B and 1C).

360

XAB2 has been found as part of the pre-mRNA splicing complex composed of AQR, PRP19,

361

CCDC16, PPIE and ISY1 (Kuraoka et al, 2008). However, none of these proteins seems to take part in the

362

repair process, underlying the peculiar function of the splicing factor XAB2 in TC-NER repair (Figure S5).

363

We also demonstrated that, unlike all the other NER protein studied so far, XAB2 protein is

364

released from the damage induced by UV-C exposure (Figure 3). Concomitantly, we also observed an

365

increased XAB2 cellular mobile fraction after UV irradiation (Figure 4D). This release from transcription-

366

blocking DNA lesions and increased mobility is surprising and atypical for a repair protein. However, this

367

behavior has also been observed for late-stage spliceosomes (Tresini et al, 2015) and could be explained

368

by the importance for the cell to rapidly provide access to the repair machinery.

369

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

Surprisingly, the increased mobility of XAB2 after UV irradiation occurs also in absence of CSA

370

and CSB protein, in presence of transcription inhibitors and in presence of ATR and ATM inhibitors

371

(Figure 2, 3 and Figure S6). These results strongly suggest that XAB2 remobilization is independent from

372

the repair process but it is more a result of the damage-dependent transcription inhibition. Accordingly

373

with this result, the return of XAB2 normal mobility is CSA and CSB dependent, suggesting that a proper

374

repair and re-establishment of the transcription process is needed to restore the mobility of XAB2.

375

Previously, we reported that a complete TC-NER mechanism is required to repair the UV-lesions

376

present on active ribosomal DNA, genes transcribed by RNA Polymerase 1 (RNAP1) (Daniel et al. 2018).

377

Notably, both Cockayne syndrome proteins (CSA and CSB) are implicated in this specific repair reaction,

378

as well as the UV-stimulated scaffold protein A (UVSSA), a protein required for the stabilization of CSB

379

specifically after UV irradiation (Higa et al, 2016). Interestingly, XAB2 is involved only in repair of RNAP2-

380

transcribed genes and not in repair of RNAP1-transcribed genes (Figure 1D), confirming a probable

381

specific interaction with RNAP2 and reinforcing the idea that RNAP1 and RNAP2 repair processes are

382

distinct although they share common proteins.

383

FRAP experiments in CSA and CSB mutant cells show a more immobile XAB2 fraction than in WT

384

cells (Figure 3B) already without any damage induction. Tanaka’s group found that XAB2 interacts in vitro

385

with CSA and CSB protein in the absence of DNA damage (Nakatsu et al, 2000). In addition, several

386

studies demonstrated an involvement of CSB in transcription regulation (Boetefuer et al, 2018). As both

387

CSB and XAB2 are necessary during the transcription process, it is therefore possible that the absence of

388

CSB will modify the mobility of XAB2. However, it is not excluded that in CSA and CSB mutant cells, a low

389

level of unrepaired oxidative damage (de Waard et al, 2004) might interfere with the proper XAB2

390

mobility, eventually modifying the amount of R-loops within these cells. Nevertheless, it is difficult to

391

precisely estimate the exact number of R-loops between different cell types and this hypothesis is

392

difficult to clearly assess.

393

The UV-induced remobilization of XAB2 is not explained by its release from the splicing complex,

394

as demonstrated by co-immunoprecipitation and PLA experiment, but from both a chromatin-specific

395

structures: R-loops and from repair proteins, CSA and XPG (Figure 4 and 5). An R-loop is a three-stranded

396

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

nucleic acid structure, composed of a DNA:RNA hybrid and the associated non-template single-stranded

397

DNA. This structure arises naturally in organisms from bacteria to humans, and have a multitude of roles

398

in the cell (Belotserkovskii et al, 2018). In this work, we show that R-loops are a substrate for XAB2 and

399

after DNA damage induction the interaction between XAB2 and R-loops are strongly reduced, this might

400

explain the increase mobility of XAB2 during the TC-NER reaction. However, we also show that XAB2

401

interactions with CSA and with XPG are reduced after UV irradiation and during repair reactions (Figure

402

5). So probably, a combination of reduced interactions of XAB2 results in an increased cellular mobility.

403

A study demonstrates that absence of AQR induces R-loops formation which are actively

404

processed into DNA double-strand breaks by XPF and XPG, the nucleotide excision repair endonucleases

405

(Sollier et al, 2014). Without any damage, we also observe an increased level of cellular R-loops in

406

absence of AQR, additionally we observed the same increase in cells silenced for XAB2 (Figure 4B),

407

suggesting an involvement of these two proteins in R-loops resolution, as recently also demonstrated by

408

Goulielmaki et al, 2021.

409

Transcription process and R-loops formation are finely interconnected. Indeed, R-loops

410

formation can cause transcription blockage but transcription blockage due to DNA damage appears to

411

result in R-loop formation (Steurer & Marteijn, 2017; Mullenders, 2015). Moreover, after UV irradiation,

412

transcription activity declines and mRNAs levels are drastically reduced (Figure S9). PLA assays show that,

413

after UV irradiation, XAB2 and total RNAs interaction is reduced, however, because the total amount of

414

RNAs is diminished after transcription block, we assume that XAB2_RNAs interaction is lessened because

415

of the intrinsic reduced amount of RNAs. Differently from total RNAs, in our study, we observe that R-

416

loops do not decrease in number after UV-damage and transcription inhibition (Figure 4). Therefore, the

417

interaction XAB2_R-loops and AQR_R-loops is specifically hindered after UV-damage (Figure 4).

418

Because XAB2 interacts in vitro with CSA and CSB protein (Nakatsu et al, 2000) and recently

419

studies have shown an interaction with XPG (Goulielmaki et al, 2021), we decided to verify whether

420

these interactions were modified during the DNA repair process and the concomitant transcription

421

inhibition. Our results show a clear and consistent reduction of interaction between XAB2 and CSA

422

(Figure 5A and 5B), and between XAB2 and XPG (Figure 5C and 5D) but not between XAB2 and CSB or

423

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

XAB2 and XPB (Figure S10). A working hypothesis would then be that XAB2 needs to be released from

424

CSA and XPG to allow a proper repair reaction.

425

The physical interaction between XAB2 and RNAP2 has already been established (Nakatsu et al,

426

2000; Kuraoka et al, 2008), however the exact relation between XAB2 and RNAP2 has not been

427

disclosed. In this study, we have clearly demonstrated by FRAP experiments that the mobility of RNAP2 is

428

severely affected in the absence of XAB2, namely, RNAP2 is released from its substrate and its mobility is

429

strongly increased (Figure 6). When cells are subjected to UV-damage, a small RNAP2 immobile fraction

430

can be observed, probably reflecting paused RNAP2 molecules on DNA damaged sites. However, in the

431

absence of XAB2 and in presence of a DNA damage, RNAP2 mobility is even more increased than without

432

damage, confirming that XAB2 might stabilize RNAP2 onto the transcribed strands of active genes in

433

presence of damage. This result leads to the hypothesis that XAB2 might collaborate with RNAP2 in the

434

detection of damage during TC-NER and hence be implicated in the first step of this process.

435

In conclusion, we described here an increased mobility of the protein XAB2 during the DNA

436

damage dependent transcription inhibition. This increased mobility might partly be explained by the

437

release of XAB2 from its substrate R-loops and its partner CSA and XPG. Importantly, we demonstrated

438

that XAB2 plays a role of anchoring RNAP2 to its substrate with and without DNA damage. The absence

439

of XAB2 might hinder the overall transcription activity of the cells as well as it severely affects the TC-NER

440

capacity.

441

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

MATERIAL AND METHODS

442

Cell culture and treatments

443

The cells used in this study were: (i) wild-type SV40-immortalized human fibroblasts (MRC5); (ii) XPC–

444

deficient SV40-immortalized human fibroblast (XP4PA, GG-NER deficient); (iii) CSA–deficient SV40-

445

immortalized human fibroblast (CS3BE, TC-NER-deficient); (iv) CSB-deficient SV40-immortalized human

446

fibroblast (CS1AN, TC-NER-deficient); (v) MRC5 stably expressing XAB2-GFP; (vi) CS3BE stably expressing

447

XAB2-GFP (vii) CS1AN stably expressing XAB2-GFP and (vii) MRC5 stably expressing RNAP2-GFP.

448

Immortalized human fibroblasts were cultured in DMEM (sigma) supplemented with 1% of penicillin and

449

streptomycin (Gibco) and 10% fetal bovine serum (Corning) and incubated at 37°C with 5% CO

450

XAB2-GFP and RNAP2-GFP stably expressing cell lines were produced by transfecting XAB2-GFP or

451

RNAP2-GFP using FuGENE® 6 Transfection Reagent (Promega) according to the manufacturer’ protocol.

452

Selection was performed with G418 at 2 mg/ml.

453

DNA damage was inflicted by UV-C light (254 nm, 6-Watt lamp). Cells were globally irradiated with a UV-

454

C dose of 2, 4, 8, 10 or 16 J/m

or locally irradiated with a UV-C dose of 60 or 100 J/m

through a filter

455

with holes of 5 µm of diameter (Millipore). After irradiation, cells were incubated at 37°C with 5% CO

for

456

different periods of time. Inhibitor of ATR pathway (VE821) was added at 10µM in the medium 1h before

457

irradiation.

458

459

Transfection of small interfering RNAs (siRNAs)

460

The small interfering RNA (siRNAs) used in this study are : siMock, Horizon, D-001206-14 (10nM); siXAB2,

461

Horizon, L-004914-01 (20nM); siXPF, Horizon, M-019946-00 (10nM); siCSB, Horizon, L-004888-00

462

(10nM). The final concentration used for each siRNA is indicated in parentheses. All siRNA from Horizon

463

are a pool of four different siRNA.

464

Cells were seeded in six-well plates and allowed to attach for at least 24h. Coverslip were added inside

465

the well if needed for experiment. 24h and 48h after seeding, cells were transfected with siRNA using

466

Lipofectamine® RNAiMAX reagent (Invitrogen; 13778150) or GenJet (Tebu-Bio), according to the

467

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

manufacturer’ protocol. Experiments were performed between 24h or 72h after the second transfection.

468

Protein knockdown was confirmed for each experiment by western blot.

469

470

Recovery of RNA synthesis (RRS) assay

471

MRC5 cells were grown on 18 mm coverslips. siRNA transfections were performed 24h and 48h before

472

RRS assay. RNA detection was done using a Click-iT RNA Alexa Fluor Imaging kit (Invitrogen), according to

473

the manufacturer’s instructions. Briefly, cells were UV-C irradiated (10 J/m²) and incubated for 0, 3, 16

474

and 24 h at 37°C. Then, cells were incubated for 2 hours with 5-ethynyl uridine. After fixation and

475

permeabilization, cells were incubated for 30 min with the Click-iT reaction cocktail containing Alexa

476

Fluor Azide 594. After washing, the coverslips were mounted with Vectashield (Vector). The average

477

fluorescence intensity per nucleus was estimated after background subtraction using ImageJ and

478

normalized to not treated cells.

479

480

RNA Fluorescence In Situ Hybridization (RNA FISH)

481

Cells were grown on 18 mm coverslips, washed with warm PBS and fixed with 4% paraformaldehyde for

482

15 min at 37° C. After two washes with PBS, cells were permeabilized with PBS + 0.4 % Triton X-100 for 7

483

min at 4° C. Cells were washed rapidly with PBS before incubation (at least 30 min) with pre-

484

hybridization buffer : 15% formamide in 2X SSPE pH8.0 (0.3M NaCl, 15.7mM NaH

O and 2.5mM

485

EDTA). 35 ng of probe was diluted in 70 µl of hybridization mix (2X SSPE, 15% formamide, 10% dextran

486

sulphate and 0.5 mg/ml tRNA). Hybridization of the probe was conducted overnight at 37° C in a

487

humidified environment. Subsequently, cells were washed twice for 20 min with pre-hybridization buffer

488

and once for 20 min with 1X SSPE. After several washing with PBS, the coverslips were mounted with

489

Vectashield containing DAPI (Vector). The probe sequence (5’ to 3’) is: Cy5-

490

AGACGAGAACGCCTGACACGCACGGCAC. Images of the cells were obtained using a Zeiss LSM 780 NLO

491

confocal laser scanning microscope and the following objective: Plan-Apochromat 63X/1.4 oil DIC M27.

492

493

Unscheduled DNA synthesis (UDS or TCR-UDS)

494

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

MRC5 or XP4PA cells, WT or GG-NER-deficient cells respectively, were grown on 18 mm coverslips. siRNA

495

transfections were performed 24h and 48h before UDS assays. After local irradiation at 100 J/m

with

496

UV-C through a 5 µm pore polycarbonate membrane filter, cells were incubated for 3 or 8 hours (UDS

497

and TCR-UDS respectively) with 5-ethynyl-2’-deoxyuridine (EdU), fixed and permeabilized with PBS and

498

0.5% triton X-100. Then, cells were blocked with PBS+ solution (PBS containing 0.15% glycine and 0.5%

499

bovine serum albumin) for 30 min and subsequently incubated for 1h at room temperature with mouse

500

monoclonal anti-yH2AX antibody (Ser139 [Upstate, clone JBW301]) 1:500 diluted in PBS+. After extensive

501

washes with PBS containing 0.5% Triton X100, cells were incubated for 45min at room temperature with

502

secondary antibodies conjugated with Alexa Fluor 594 fluorescent dyes (Molecular Probes, 1:400 dilution

503

in PBS+). Next, cells were washed several times and then incubated for 30 min with the Click-iT reaction

504

cocktail containing Alexa Fluor Azide 488. After washing, the coverslips were mounted with Vectashield

505

containing DAPI (Vector). Images of the cells were obtained with the same microscopy system and

506

constant acquisition parameters. Images were analyzed as follows using ImageJ and a circle of constant

507

size for all images: (i) the background signal was estimated in the nucleus (avoiding the damage, nucleoli

508

and other non-specific signal) and subtracted, (ii) the locally damaged area was defined by using the

509

yH2AX staining, (iii) the average fluorescence correlated to the EdU incorporation was then measured

510

and thus an estimate of DNA synthesis after repair was obtained.

511

512

Immunofluorescence

513

Cells were plated on 12 or 18 mm coverslips in order to reach 70% confluence on the day of the staining.

514

After two washes with PBS, cells were fixed with 2% paraformaldehyde for 15 min at 37° C. Cells were

515

permeabilized by three short washes followed by two washes of 10min with PBS + 0.1 % Triton X-100.

516

Blocking of non-specific signal was performed with PBS+ (PBS, 0.5 % BSA, 0.15 % glycine) for at least 30

517

min. Then, coverslips were incubated with primary antibody diluted in PBS+ for 2 h at RT or overnight at

518

4°C in a moist chamber. After several washes with PBS + 0.1 % Triton X-100 (three short washes and two

519

of 10min) and a short wash with PBS+, cells were incubated for 1h at RT in a moist chamber with

520

secondary antibody coupled to fluorochrome (Goat anti-mouse Alexa Fluor® 488 [A11001, Invitrogen] or

521

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

594 [A11005] and Goat anti-rabbit Alexa Fluor® 488 [A11008] or 594 [A11012], 1/400 dilution in PBS+).

522

After the same washing procedure with PBS instead of PBS + 0.1 % Triton X-100, coverslips were finally

523

mounted using Vectashield with DAPI (Vector Laboratories).

524

The variation of fluorescence in the locally irradiated zone has been calculated as for the UDS

525

experiment.

526

527

Proximity Ligation Assay (PLA)

528

PLA experiments were done using Duolink™ II secondary antibodies and detection kits (Sigma-Aldrich,

529

#DUO92002, #DUO92004, and #DUO92008) according to the manufacturer’s instruction. The cells were

530

fixed and permeabilized with the same procedure as immunofluorescence. After blocking 1h at 37°C with

531

the Blocking Solution from the kit, incubation of the primary antibodies diluted in Antibody Diluent was

532

performed at 4°C overnight. Duolink™ secondary antibodies were added and incubated for 1h at 37°C. If

533

secondary antibodies were in close proximity (<40 nm), they were ligated together to make a closed

534

circle by the Duolink™ ligation solution. The DNA is then amplified (rolling circle amplification) and

535

detected by fluorescence 594 (red dot). Coverslips were mounted using Vectashield with DAPI (Vector

536

Laboratories).

537

538

Cytostripping

539

To remove the background generated by some antibodies or EdU incorporation, the cytoplasm of the

540

cells was removed before fixation. After two washes with cold PBS, cells were incubated on ice 5 min

541

with cold cytoskeleton buffer (10mM PIPES pH6,8; 100mM NaCl; 300mM Sucrose; 3mM MgCl2; 1mM

542

EGTA; 0,5% Triton X100) followed by 5mn with cold cytostripping buffer (10mM Tris HCL pH7,4; 10mM

543

NaCl; 3mM MgCl2; 1% Tween 40; 0,5% sodium deoxycholate). After 3 gently washes with cold PBS, cells

544

were fixed.

545

546

Images acquisition

547

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

For RRS, images of the cells were obtained using an Andor spinning disk : Olympus IX 83 inverted

548

microscope, equipped with a Yokaga CSU-X1 Spinning disk Unit and BOREALIS technology for

549

homogeneous illumination. The acquisition software is IQ3.

550

For RNAFish, UDS, TCR-UDS and IF of splicing complex after local damage, images of the cells were

551

obtained using a Zeiss LSM 780 NLO confocal laser scanning microscope and the following objective:

552

Plan-Apochromat 63X/1.4 oil DIC M27 or 40X/1.3 oil DIC. The acquisition software is ZEN.

553

PLA and some IF has been performed on a Zeiss Z1 imager right using a 40x/0.75 dry objective. The

554

acquisition software is Metavue.

555

All images were analyzed with ImageJ software.

556

557

Primary antibodies used for IF and PLA

558

Primary antibodies used for Immunofluorescence and PLA experiments were anti-XAB2 (mouse, sc-

559

271037 [Santa Cruz Biotechnology], 1/1000 dilution and rabbit, A303-638A [Béthyl], 1/500 dilution), anti-

560

AQR (IPB160 rabbit, A302-547A [Béthyl], 1/500 dilution), anti-CCDC16 (rabbit, HPA027211 [atlas

561

antibodies], 1/250 dilution), anti-PRP19 (rabbit, ab27699 [abcam], 1/500 dilution), anti-DNA:RNA hybrid

562

clone S9.6 (mouse, MABE1095 [Merck Millipore], 1/100 dilution and rabbit, Ab01137-23.0 [Absolute

563

antibody], 1/100 dilution), anti-CSA (rabbit, GTX100145 [genetex], 1/400 dilution), anti-XPG (rabbit,

564

sc84663 [santa-cruz], 1/1000 dilution), anti-Pol2tot (rabbit, A300-653A [Béthyl] 1/400 dilution, anti-

565

Pol2Ser2P (rabbit, ab5095 [abcam] 1/100 dilution) and anti-Pol2Ser5P (rabbit, 13523S [cell signaling]

566

1/1000 dilution).

567

568

Fluorescence Recovery after Photo-Bleaching (FRAP)

569

FRAP experiments were performed on a Zeiss LSM 780 NLO confocal laser scanning microscope (Zeiss),

570

using a 40x/1.3 oil objective, under a controlled environment (37°C, 5% CO2). A narrow region of interest

571

(ROI) centered across the nucleus of a living cell was monitored every 20 ms (1% laser intensity of the

572

488 nm line of a 25 mW Argon laser) until the fluorescence signal reached a steady state level (after ≈ 2

573

s). The same region was then photo-bleached for 20 ms at 100% laser intensity. Recovery of fluorescence

574

in the bleached ROI was then monitored (1% laser intensity) every 20 ms for about 20 seconds. Analysis

575

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

of raw data was performed with the ImageJ software. All FRAP data were normalized to the average pre-

576

bleached fluorescence after background removal.

577

Analysis of SPOT FRAP data was performed as follows (Figure S6). The average fluorescence (over all

578

cells) of the No UV condition was subtracted from the average fluorescence of the UV treated conditions.

579

The obtained difference between the two FRAP curves was then summed point by point, starting from

580

the bleach up to the following 100 measurements i.e. the area between the curve of interest and the

581

untreated condition curve.

582

583

Protein extraction

584

Cells cultured in 10-cm dishes were harvested by trypsinization and the pellet was washed once with PBS

585

supplemented with the Protease Inhibitor Cocktail (PIC, Roche). The extraction of nuclear proteins has

586

been performed using the CelLytic™ NuCLEAR™ Extraction kit (Sigma-Aldrich) complemented with PIC.

587

Protein concentration was determined using the Bradford method. The samples were diluted with

588

Laemmli buffer (10% glycerol, 5% β-mercaptoethanol, 3% sodium dodecyl sulfate, 100mM Tris-HCL [pH -

589

.8], bromophenol blue), and heated 95°C before loading on an SDS-PAGE gel.

590

591

Co-immunoprecipitation

592

For co-immunoprecipitation, 10 µl of protein G magnetic beads (Bio-adembead, Ademtech) were used

593

per IP. 1µg of anti-XAB2 antibody (rabbit, A303-638A, Bethyl) were bound to the beads in PBS with BSA

594

(3%) during 2h at 4°C with rotation. 100 µg of nuclear extracts were then incubated with beads-

595

antibodies complex for 2h at 4°C with rotation. After two washes at 100 mM salt, two washes at 150mM

596

and one wash at 100mM, beads were boiled in 2x Laemmli buffer and eluted samples loaded on a SDS

597

PAGE gel.

598

599

RNA/DNA Hybrid IP

600

Non-crosslinked MRC5 cells were lysed in 85 mM KCl, 5 mM PIPES (pH 8.0), and 0.5% NP-40 for 10 min

601

on ice. Pelleted nuclei were resuspended in RSB buffer (10 mM Tris-HCl pH 7.5, 200 mM NaCl, 2.5 mM

602

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

MgCl2) with 0.2% sodium deoxycholate, 0.1% SDS, 0.05% sodium lauroyl sarcosinate and 0.5% Triton X-

603

100, and extracts were sonicated for 10 min (Diagenode Bioruptor, 60 cycles high power, 10sec ON and

604

20 sec OFF). Extracts were then diluted 1:4 in RSB with 0.5% Triton X-100 (RSB + T) and subjected to IP

605

with the S9.6 antibody overnight at 4°C. RNaseA was added during IP at 0.1 ng RNaseA per microgram

606

genomic DNA. Then protein G dynabeads (Invitrogen) washed with RSB + T was added and incubated for

607

3h. Beads were washed 4x with RSB + T; 2x with RSB; and eluted in 2x Laemmli buffer for 10 min at 95°C

608

before loading on SDS-PAGE.

609

610

Western-blot

611

Proteins were separated on a SDS-PAGE gel composed of bisacrylamide (37:5:1), blotted onto a

612

polyvinylidene difluoride membrane (PVDF, 0.45μm Millipore) The membrane was blocked in PBS-T (PBS

613

and 0.1 % Tween 20) with 5 % milk and incubated for 2 h at RT or overnight at 4° C with the following

614

primary antibodies diluted in milk PBS-T : Rabbit anti-XAB2, A303-638A (Bethyl) 1/1000. Subsequently,

615

membrane was washed repeatedly with PBS-T and incubated 1 h at RT with the following secondary

616

antibody diluted 1/5000 in milk PBS-T: Goat anti-rabbit IgG HRP conjugate (170-6515; BioRad). After the

617

same washing procedure, protein bands were visualized via chemiluminescence (ECL Enhanced

618

Chemiluminescence; Pierce ECL Western Blotting Substrate) using the ChemiDoc MP system (BioRad).

619

620

Statistical analysis

621

Error bars represent the Standard Error of the Mean (SEM) of the biological replicates. Microsoft Excel

622

was used for statistical analysis and plotting of all the numerical data. Statistics were performed using a

623

test of student to compare two different conditions (siMock vs siRNA X or No UV irradiation vs after

624

irradiation) with the following parameters: two-tailed distribution and two-sample unequal variance

625

(heteroscedastic).

626

627

628

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

ACKNOWLEDGMENTS AND FUNDING SOURCES

629

This work has been funded by Agence Nationale de la Recherche (ANR-14-CE10-0009), Institut National

630

du Cancer (PLBIO17-043 et PLBIO19-126), La Ligue contre le Cancer (218398) et Électricité de France

631

(contrat 218398).

632

633

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

REFERENCES :

634

Belotserkovskii BP, Tornaletti S, D’Souza AD & Hanawalt PC (2018) R-loop generation during

635

transcription: Formation, processing and cellular outcomes. DNA Repair 71: 69–81

636

Boetefuer EL, Lake RJ & Fan H-Y (2018) Mechanistic insights into the regulation of transcription and

637

transcription-coupled DNA repair by Cockayne syndrome protein B. Nucleic Acids Res 46: 7471–

638

7479

639

van den Boom V, Citterio E, Hoogstraten D, Zotter A, Egly J-M, van Cappellen WA, Hoeijmakers JHJ,

640

Houtsmuller AB & Vermeulen W (2004) DNA damage stabilizes interaction of CSB with the

641

transcription elongation machinery. J Cell Biol 166: 27–36

642

Chatterjee N & Walker GC (2017) Mechanisms of DNA damage, repair and mutagenesis. Environ Mol

643

Mutagen 58: 235–263

644

Daniel L, Cerutti E, Donnio L-M, Nonnekens J, Carrat C, Zahova S, Mari P-O & Giglia-Mari G (2018)

645

Mechanistic insights in transcription-coupled nucleotide excision repair of ribosomal DNA. Proc

646

Natl Acad Sci U S A 115: E6770–E6779

647

Donnio L-M, Lagarou A, Sueur G, Mari P-O & Giglia-Mari G (2019) CSB-Dependent Cyclin-Dependent

648

Kinase 9 Degradation and RNA Polymerase II Phosphorylation during Transcription-Coupled

649

Repair. Mol Cell Biol 39: e00225-18

650

Giglia-Mari G, Zotter A & Vermeulen W (2011) DNA Damage Response. Cold Spring Harb Perspect

651

Biol 3: a000745

652

Goulielmaki E, Tsekrekou M, Batsiotos N, Ascensão-Ferreira M, Ledaki E, Stratigi K, Chatzinikolaou G,

653

Topalis P, Kosteas T, Altmüller J, et al (2021) The splicing factor XAB2 interacts with ERCC1-

654

XPF and XPG for R-loop processing. Nat Commun 12: 3153

655

Hanawalt PC (1994) Transcription-coupled repair and human disease. Science 266: 1957–1958

656

Higa M, Zhang X, Tanaka K & Saijo M (2016) Stabilization of Ultraviolet (UV)-stimulated Scaffold

657

Protein A by Interaction with Ubiquitin-specific Peptidase 7 Is Essential for Transcription-coupled

658

Nucleotide Excision Repair. J Biol Chem 291: 13771–13779

659

Kuraoka I, Ito S, Wada T, Hayashida M, Lee L, Saijo M, Nakatsu Y, Matsumoto M, Matsunaga T, Handa

660

H, et al (2008) Isolation of XAB2 complex involved in pre-mRNA splicing, transcription, and

661

transcription-coupled repair. J Biol Chem 283: 940–950

662

Maréchal A, Li J-M, Ji XY, Wu C-S, Yazinski SA, Nguyen HD, Liu S, Jiménez AE, Jin J & Zou L (2014)

663

PRP19 transforms into a sensor of RPA-ssDNA after DNA damage and drives ATR activation via

664

a ubiquitin-mediated circuitry. Mol Cell 53: 235–246

665

Marteijn JA, Lans H, Vermeulen W & Hoeijmakers JHJ (2014) Understanding nucleotide excision repair

666

and its roles in cancer and ageing. Nat Rev Mol Cell Biol 15: 465–481

667

Moné MJ, Bernas T, Dinant C, Goedvree FA, Manders EMM, Volker M, Houtsmuller AB, Hoeijmakers

668

JHJ, Vermeulen W & van Driel R (2004) In vivo dynamics of chromatin-associated complex

669

formation in mammalian nucleotide excision repair. Proc Natl Acad Sci U S A 101: 15933–15937

670

Mourgues S, Gautier V, Lagarou A, Bordier C, Mourcet A, Slingerland J, Kaddoum L, Coin F, Vermeulen

671

W, Gonzales de Peredo A, et al (2013) ELL, a novel TFIIH partner, is involved in transcription

672

restart after DNA repair. Proc Natl Acad Sci 110: 17927–17932

673

Mullenders L (2015) DNA damage mediated transcription arrest: Step back to go forward. DNA Repair

674

36: 28–35

675

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint

Nakatsu Y, Asahina H, Citterio E, Rademakers S, Vermeulen W, Kamiuchi S, Yeo JP, Khaw MC, Saijo

676

M, Kodo N, et al (2000) XAB2, a novel tetratricopeptide repeat protein involved in transcription-

677

coupled DNA repair and transcription. J Biol Chem 275: 34931–34937

678

Onyango DO, Howard SM, Neherin K, Yanez DA & Stark JM (2016) Tetratricopeptide repeat factor

679

XAB2 mediates the end resection step of homologous recombination. Nucleic Acids Res 44:

680

5702–5716

681

Politi A, Moné MJ, Houtsmuller AB, Hoogstraten D, Vermeulen W, Heinrich R & Driel R van (2005)

682

Mathematical Modeling of Nucleotide Excision Repair Reveals Efficiency of Sequential

683

Assembly Strategies. Mol Cell 19: 679–690

684

Rademakers S, Volker M, Hoogstraten D, Nigg AL, Moné MJ, Zeeland AA van, Hoeijmakers JHJ,

685

Houtsmuller AB & Vermeulen W (2003) Xeroderma Pigmentosum Group A Protein Loads as a

686

Separate Factor onto DNA Lesions. Mol Cell Biol 23: 5755–5767

687

Sakasai R, Isono M, Wakasugi M, Hashimoto M, Sunatani Y, Matsui T, Shibata A, Matsunaga T &

688

Iwabuchi K (2017) Aquarius is required for proper CtIP expression and homologous

689

recombination repair. Sci Rep 7: 13808

690

Sollier J, Stork CT, García-Rubio ML, Paulsen RD, Aguilera A & Cimprich KA (2014) Transcription-

691

coupled nucleotide excision repair factors promote R-loop-induced genome instability. Mol Cell

692

56: 777–785

693

Steurer B & Marteijn JA (2017) Traveling Rocky Roads: The Consequences of Transcription-Blocking

694

DNA Lesions on RNA Polymerase II. J Mol Biol 429: 3146–3155

695

Tresini M, Warmerdam DO, Kolovos P, Snijder L, Vrouwe MG, Demmers JAA, van IJcken WFJ,

696

Grosveld FG, Medema RH, Hoeijmakers JHJ, et al (2015) The core spliceosome as target and

697

effector of non-canonical ATM signaling. Nature 523: 53–58

698

de Waard H, de Wit J, Andressoo J-O, van Oostrom CTM, Riis B, Weimann A, Poulsen HE, van Steeg H,

699

Hoeijmakers JHJ & van der Horst GTJ (2004) Different Effects of CSA and CSB Deficiency on

700

Sensitivity to Oxidative DNA Damage. Mol Cell Biol 24: 7941–7948

701

Wang IX, Grunseich C, Fox J, Burdick J, Zhu Z, Ravazian N, Hafner M & Cheung VG (2018) Human

702

proteins that interact with RNA/DNA hybrids. Genome Res 28: 1405–1414

703

Yonemasu R, Minami M, Nakatsu Y, Takeuchi M, Kuraoka I, Matsuda Y, Higashi Y, Kondoh H &

704

Tanaka K (2005) Disruption of mouse XAB2 gene involved in pre-mRNA splicing, transcription

705

and transcription-coupled DNA repair results in preimplantation lethality. DNA Repair 4: 479–

706

491

707

Zotter A, Luijsterburg MS, Warmerdam DO, Ibrahim S, Nigg A, van Cappellen WA, Hoeijmakers JHJ,

708

van Driel R, Vermeulen W & Houtsmuller AB (2006) Recruitment of the nucleotide excision

709

repair endonuclease XPG to sites of UV-induced dna damage depends on functional TFIIH. Mol

710

Cell Biol 26: 8868–8879

711

712

.CC-BY 4.0 International licenseperpetuity. It is made available under a

preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in

The copyright holder for thisthis version posted December 24, 2021. ; https://doi.org/10.1101/2021.12.23.473962doi: bioRxiv preprint