Newly identified CHO ERCC3/XPB mutations and phenotype characterization

Abstract

Nucleotide excision repair (NER) is a complex multistage process involving many interacting gene products to repair a wide range of DNA lesions. Genetic defects in NER cause human hereditary diseases including xeroderma pigmentosum (XP), Cockayne syndrome (CS), trichothiodystrophy and a combined XP/CS overlapping symptom. One key gene product associated with all these disorders is the excision repair cross-complementing 3/xeroderma pigmentosum B (ERCC3/XPB) DNA helicase, a subunit of the transcription factor IIH complex. ERCC3 is involved in initiation of basal transcription and global genome repair as well as in transcription-coupled repair (TCR). The hamster ERCC3 gene shows high degree of homology with the human ERCC3/XPB gene. We identified new mutations in the Chinese hamster ovary cell ERCC3 gene and characterized the role of hamster ERCC3 protein in DNA repair of ultraviolet (UV)-induced and oxidative DNA damage. All but one newly described mutations are located in the protein C-terminal region around the last intron–exon boundary. Due to protein truncations or frameshifts, they lack amino acid Ser751, phosphorylation of which prevents the 5′ incision of the UV-induced lesion during NER. Thus, despite the various locations of the mutations, their phenotypes are similar. All ercc3 mutants are extremely sensitive to UV-C light and lack recovery of RNA synthesis (RRS), confirming a defect in TCR of UV-induced damage. Their limited global genome NER capacity averages ∼8%. We detected modest sensitivity of ercc3 mutants to the photosensitizer Ro19-8022, which primarily introduces 8-oxoguanine lesions into DNA. Ro19-8022-induced damage interfered with RRS, and some of the ercc3 mutants had delayed kinetics. All ercc3 mutants showed efficient base excision repair (BER). Thus, the positions of the mutations have no effect on the sensitivity to, and repair of, Ro19-8022-induced DNA damage, suggesting that the ERCC3 protein is not involved in BER.

Introduction

A large variety of agents including ultraviolet (UV) radiation and reactive oxygen species can alter normal DNA bases and subsequently produce mutations or block cellular processes such as replication or transcription. In order to survive this impact, cells have evolved several mechanisms to repair the damage and restore the DNA structure. Depending on the type of damage, two major excision repair pathways are involved: base excision repair (BER) and nucleotide excision repair (NER) (1). BER primarily corrects non-helix-distorting lesions, such as oxidative damage to DNA caused as a by-product of normal aerobic metabolism or environmental influences (2). NER processes structures distorting the DNA helix, such as the main UV-C photoproducts, cyclobutane pyrimidine dimers (CPDs) and 6-4 photoproducts. Based on the transcriptional state of DNA, there are two subclasses of NER: global genome repair (GGR or GG-NER) occurring in transcriptionally inactive DNA of the entire genome and transcription-coupled repair (TCR or TC-NER) acting on the transcribed strand of active genes. The damage is repaired in a similar manner although the sub-pathways differ in the damage recognition step. In GG-NER, the damage recognition is provided by the XPC–hHR23B complex and can be greatly enhanced by the damaged DNA-binding complex (UV-DDB, also known as XPE) (3,4). TCR is initiated by DNA damage in transcribed strands that can arrest the transcription machinery [for the most recent reviews, see (5,6)]. Unlike TCR of UV-induced DNA damage, which was demonstrated two decades ago (7), the existence of TCR of non-bulky oxidative lesions remains controversial. While studies using shuttle vectors containing a defined oxidative lesion strongly suggest the presence of TCR of oxidative damage in rodent cells (8,9), analysis of gene-specific repair of 8-oxoguanine (8-oxoG) in endogenous genes does not confirm its preferential repair (10).

The complexity and importance of GGR and TCR are linked to three distinct, genetically heterogeneous human diseases: xeroderma pigmentosum (XP), Cockayne syndrome (CS) and trichothiodystrophy (TTD). These rare autosomal recessive hereditary disorders are each characterized by high photosensitivity but very different clinical features (11,12). The genetic basis of these diseases can be mutations in different NER repair genes or different mutations in the same NER gene. Mutations in the CSA or CSB genes result in CS, whereas certain mutations in the ERCC3/XPB, XPD or XPG genes result in overlap of disease, the combined XP/CS syndrome. XP–CS individuals suffer from the clinical features of XP and are also afflicted by severe neurological and developmental abnormalities typical for CS caused by demyelination (13).

The ERCC3/XPB gene (hereafter referred to as ERCC3) is notable for its association with all three disorders. It was located on chromosome 2q21 and encodes a 782 amino acid long protein (14). The gene is ∼45 kb long and consists of 14 exons. The following functional domains were identified: nuclear localization signal, DNA-binding domain, DNA damage recognition domain, seven helicase domains, a unique RED motif (arginine, glutamine and asparagine) responsible for DNA unwinding and a flexible thumb motif [(14,15), Supplementary Figure S1, available at Mutagenesis Online]. The ERCC3 protein is an essential ATP-dependent DNA helicase that functions as an integral component of the transcription factor IIH (TFIIH) protein complex (16), which is required for transcription initiation as well as for NER and TCR. Its high evolutionary conservation has allowed identification of homologous proteins in different organisms, including hamster. Previously, it was shown that human and hamster ERCC3 proteins are 96% identical (17). Up to now, only six mutations in six families are reported, consistent with essentiality of the protein for the above-mentioned cellular processes and cell viability (18). Thus, ercc3 mutant hamster cell lines are a valuable tool for studying the role of ERCC3 protein in distinct metabolic processes, thereby contributing to a better understanding of the mechanisms of mutagenesis.

In the present study, we performed molecular analysis of Chinese hamster ovary (CHO) mutant cell lines derived from the AA8 parental line, and we discovered new mutations in the ERCC3 gene. Further, we investigated the impact of two different DNA-damaging agents, photosensitizer Ro19-8022 (RO) and UV-C light on wild-type versus ercc3 mutant cells to study the processing of different DNA lesions (primarily 8-oxoG and UV-photoproducts). In addition, we analysed recovery of RNA synthesis (RRS) after these two agents as an indirect marker of functional TCR. We show that all ercc3 mutants are able to remove formamidopyrimidine DNA-glycosilase (Fpg)-sensitive sites from the whole genome, suggesting effective BER as well as non-participation of ERCC3 protein in this repair pathway. In contrast, the mutant cell lines have severely reduced GGR efficiency of UV-induced DNA damage and display no RRS, confirming the involvement of ERCC3 protein in both NER sub-pathways.

Materials and methods

Cell culture conditions and treatments

The CHO-AA8 wild-type and the UV-sensitive ercc3 mutant cell lines (UV23, UV24, UV68, UV78, UV113, UV179 and UV218) were derived from the mutant hunt of Busch et al. (19). The ercc6/csb mutant UV61 was provided by Andrew Collins. All cell strains used in this study were cultured in Dulbecco's Modified Eagle’s Medium and F-10 (Ham) Nutrient Mixture (1:1) supplemented with 10% foetal bovine serum, 100 U/ml penicillin and 100 μg/ml streptomycin (Life Technologies Corporation, Carlsbad, CA, USA) and maintained at 37°C in a humidified atmosphere of 5% CO₂. The exposure of cells to RO was carried out in phosphate-buffered saline (140 mM NaCl, 3 mM KCl, 8 mM Na₂HPO₄, 1 mM KH₂PO₄, 0.1% glucose, pH 7.4) on ice using 1000 W halogen lamp (Philips PF811) at a distance of 33 cm for 2 min. Irradiation of monolayer cells was performed using a germicidal UV lamp (254 nm) at a dose rate of 1 J/m²/sec. [R]-1-[(10-chloro-4-oxo-3-phenyl-4H-benzo[a]quinolizin-1-yl)-carbonyl]-2-pyrrolidinemethanol (Ro19-8022; RO) was a generous gift from Hoffmann–LaRoche AG, Basel, Switzerland. T4 endonuclease V (endoV) and Fpg DNA glycosylase were obtained from K. J. Angelis, Institute of Experimental Botany, Academy of Sciences of Czech Republic, Prague, Czech Republic.

Sequencing the ERCC3 genes

To isolate the cDNA of ERCC3 gene, total RNA was obtained by the TRIzol method as described by the manufacturer (Life Technologies Corporation) followed by synthesis of the first cDNA strand by the ThermoScript™ reverse transcriptase (RT)-polymerase chain reaction (PCR) System (Life Technologies Corporation). The synthesized cDNA and gene-specific primers (17) were used for subsequent PCR with ThermalAce™ DNA polymerase. Sequencing was performed using the ABI PRISM BigDye™ Terminator v3.1 Cycle Sequencing Kit (Life Technologies Corporation) at the ABI PRISM 310 Genetic Analyzer.

Survival analysis

The survival curves based on the clonogenic assay were performed using a defined number of cells UV-C irradiated (0–15 J/m²) or exposed to RO (0–1.5 μM). After 7 days cultivation, the colonies were stained with methylene blue (1%) and counted. Survival was expressed as ratio of the plating efficiency of treated cells to untreated control. In the case of RO treatment, the control cells were exposed to visible light only. RO itself (without photosensitization) is not cytotoxic (10).

Recovery of RNA synthesis

The 4 × 10⁵/60 mm cells were treated with 0.1 μM RO or irradiated with 10 J/m² of UV. Nascent RNA was labelled in vivo for 30 min with [³H]uridine (2 μCi/ml) at 0, 2, 4, 6, 8 and 24 h after treatment as described previously (17). The amount of radioactivity was measured using a Beckman 1801 liquid scintillation counter. The values of relative total RNA synthesis were obtained by the ratio of incorporated [³H]uridine in treated to untreated cells.

Single-cell gel electrophoresis (comet assay)

We measured 8-oxoG removal after 0.05 μM of RO and CPD removal after 0.25 J/m² of 254 nm UV irradiation using the comet assay (20,21) as described previously (17). Briefly, exponentially growing cells were treated (or not) and processed immediately or allowed to repair for 4 or 24 h in fresh medium. After lysis of the cells, slides were overlaid either with Fpg DNA glycosylase (diluted 1:300) or T4 endoV (diluted 1:1000) and incubated at 37°C for 30 min (Fpg DNA glycosylase) or for 15 min (T4 endoV) in a humidified atmosphere. The percentage of removed CPDs was calculated as described previously (17). The percentage of repair of oxidized bases was calculated as the amount of damage removed after repair period (4 h) compared with the level of initial 8-oxoG DNA damage.

Statistical analysis

All data for statistical analysis were repeated using a minimum of three independent experiments. Averages and standard deviations are reported. Student's t-test was used to compare averages. P-values of 0.05 or less were considered statistically significant (*P < 0.05, **P < 0.01 and ***P < 0.001). The relationship between dose and effect of DNA-damaged agent was analysed using inhibitory concentration (IC)₅₀ (22).

Results

Determination of the sequence of mutant ERCC3 cDNAs

The new CHO ercc3 mutations were identified using sequencing of cDNAs, which were obtained by RT-PCR of total extracted RNA using gene-specific primers. Mutations are summarized in Table I and Supplementary Figure S1 (available at Mutagenesis Online). In UV23 cell line, the adenine to guanine substitution in the splice acceptor site (AG 3′) of the last intron (c.2218-2A > G) shifted splicing to the nearest pseudo AG site inside the last exon. At the RNA level, it caused the deletion of the alternative transcript by 26 nt (r.[= 2218_2243del]). At the protein level, it caused the internal deletion of eight amino acids of the last exon, the ninth one being changed from serine to leucine due to the frameshift. A new stop codon was created seven amino acids downstream (p.A740_S748>LfsX7), truncating the protein by 37 amino acids. In mutant UV68, we discovered a point transversion mutation (c.1411G > T) that caused an amino acid change (p.V471F). This substitution of the neutral non-polar hydrophobic amino acid phenylalanine for the amino acid valine of similar properties is located directly inside the helicase domain III of the ERCC3 protein. In mutants UV78 and UV179, we found a transversion mutation (c.2191G > T) altering the codon for neutral non-polar amino acid glycine to stop codon TGA (p.G731X). The insertion of a premature stop codon causes a truncation of the C-terminus of the ERCC3 protein by 52 amino acids. The ERCC3 protein of UV113 cell line is truncated by 28 amino acids due to the frameshift at amino acid position 740 (p.A740GfsX16) next to the insertion mutation (c.2218_2219insG). UV218 has tandem double mutations, the first being a transversion (c.2225G > C) giving rise to the substitution of the neutral non-polar hydrophobic amino acid proline for the basic amino acid arginine (p.R742P). The second mutation is the insertion of cytosine (c.2227_2228insC), which causes a frameshift at amino acid position 743 (p.R743PfsX13). The mutation creates a premature stop codon in position 755 and truncation of the protein by 28 amino acids.

Table I.

CHO ERCC3 mutations

Cell line	Type of change	cDNA	Protein (1-letter) (change deduced only)	Type of mutation	Protein truncated by
27-1a	Substitution	c.1075A > G	K359E	Missense	0 AA
MMC-2a	Substitution	c.2215C > T	Q739X	Nonsense	44 AA
UV23	Substitution	c.2218-2A > G	(A740_S748>LfsX7)	Splicing	37 AA
				Frameshift
UV24a	Substitution	c.(1144T > C; 2215C > T)	(S382P; Q739X)	Missense	44 AA
				Nonsense
UV68	Substitution	c.1411G > T	V471F	Missense	0 AA
UV78 UV179	Substitution	c.2191G > T	G731X	Nonsense	52 AA
UV113	Insertion	c.2218-2219insG	A740GfsX16	Frameshift	28 AA
UV218	Substitution	c.(2225G > C;2227-2228insC)	(R742P; R743PfsX13)	Missense	28 AA
	Insertion			Frameshift

^aMutations were identified in (17). AA, amino acids.

Sensitivity of mutants to RO and UV-C

To test the sensitivity of mutants to photosensitizer, the colony-forming ability assay was employed. Consequently, each ercc3 mutant and parental cells were treated with increasing concentrations of RO and the survival fraction was determined (Figure 1A). We found a modest but significant difference between the mutants and parental line, which was more evident at higher concentrations. By comparison of the range of IC₅₀ values of ercc3 mutants, lines UV23, UV78 and UV113 were slightly more sensitive than the rest of the mutants (Table II).

Fig. 1.

Cytotoxic effects of RO (Panel A) and UV-C (Panel B) on the wild-type parental cell line AA8 and the ercc3 mutants. The survival curves show averages of at least three independent experiments, and error bars represent standard deviation.

Table II.

Comparison of ercc3 mutant cell lines according to IC₅₀

Cell line	AA8	UV23	UV24	UV68	UV78	UV113	UV179	UV218
IC50 UV-C (J/m2)	10.44	1.47***	0.82***	2.85***	1.36***	1.27***	1.39***	1.19***
IC50 Ro19-8022 (μM)	0.70	0.56**	0.59*	0.62*	0.44***	0.52**	0.61*	0.58**

The values show the average of at least three independent experiments.

Human cells from ERCC3 patients display high sensitivity to UV-C radiation, corresponding to a deficiency in DNA repair (18,23). Therefore, we studied the effect of UV-C radiation on the survival of ercc3 mutant lines using the clonogenic assay (Figure 1B). In comparison with the AA8 parental line, IC₅₀ values of mutant lines ranged from 0.82 to 2.85 J/m² with the most sensitive being UV24 and UV68 being the most resistant (Table II). These results confirm the role of hamster ERCC3 protein in the repair of UV-induced DNA damage.

RRS following DNA damage

It is assumed that RRS following DNA damage reflects the removal of lesions from the transcribed strand of active genes. To determine whether photosensitizer-induced DNA damage can inhibit transcription, we measured total RRS by in vivo labelling nascent transcripts with [³H]uridine following RO treatment. Since after 8 h recovery time AA8 parental cells had dramatically decreased RRS (up to 80% at a 1 μM; data not shown), cells were treated with a sub-lethal concentration of 0.1 μM. RNA synthesis was markedly reduced soon after treatment (Figure 2A) to below 40% in some mutants. Mutant lines UV24, UV68, UV113 and UV218 showed the most inhibition (Figure 2A). Interestingly, the ercc6/csb line UV61 is as proficient in RRS as parental cells. However, within 24 h, all mutant cell lines restored transcription activity similar to the level of parental cells.

Fig. 2.

RRS. Cells were treated and pulse labelled with [³H]uridine for 30 min and the data normalized to the values of untreated cells. Each data point represents the average of at least three independent experiments; error bars represent the SD. (A) Treatment with RO (0.1 μM) plus visible light; (B) 10 J/m² UV-C.

An important role of TCR is to efficiently remove CPD lesions blocking transcription after UV-C exposure. To determine whether RNA synthesis is impaired by UV-C in the ercc3 mutant lines, we measured the kinetics following 10 J/m² UV-C light (Figure 2B). While the AA8 parental line recovered after 6–8 h, none of the mutants did. Their RNA synthesis was inhibited to 30–50% compared with parental cells after 8 h post-treatment, implying a defect in TCR. As expected, the ercc6 cell line UV61 is severely deficient in RRS.

Removal of DNA damage from the whole genome

It is assumed that oxidative damage is removed by an intricate network of DNA repair mechanisms that are integrated with other cellular processes such as transcription and replication and even use some common proteins. To test for differences in GGR of oxidative damage between parental and mutant lines, the modified comet assay using Fpg DNA glycosylase was employed. Fpg is a multifunctional BER enzyme that predominantly recognizes 8-oxoG base damage and creates alkali-labile sites (that are converted to single-stranded breaks). DNA experiencing breaks migrates further in the gel than intact DNA, creating an image resembling a comet. To distinguish among BER capacities of the mutant lines, we used a very low concentration (0.05 μM) of RO to avoid saturating the response. As 50% of the Fpg-sensitive sites in hamster cells are repaired within 2 h (24), the treated cells were post-incubated for 4 h to test the level of global BER of 8-oxoG. The remaining lesions were measured and compared with the BER capacity of the parental cell line (Figure 3A). However, there were no significant differences in the repair of RO-induced DNA damage among any of the cell lines, suggesting that all ercc3 mutant lines are proficient in global BER of oxidative damage.

Fig. 3.

The level of DNA damage in parental and ercc3 mutant lines. (A) Cells were treated with 0.05 μM RO and remaining DNA damage was estimated by Fpg endonuclease modified comet assay 4 h after treatment. (B) Cells were irradiated with 0.25 J/m² and remaining DNA damage was estimated by T4 endoV modified comet assay 24 h after UV irradiation. The numbers on the top of columns represent the percentage of repair. (***) Denotes P < 0.001 significance of difference between mutant cell line and corresponding wild-type cells.

Rodent cells are known to repair CPDs primarily in transcriptionally active DNA, but their survival is similar to that of human cells, which remove a much higher proportion of CPDs from the whole genome (the so-called ‘rodent repair paradox’). CPD removal was monitored by the modified alkaline comet assay using T4 endoV after 0.25 J/m² of UV irradiation and 24 h repair. Again, we used a low UV fluence to distinguish among the NER capacities of the mutant lines. Wild-type cells showed normal repair of CPDs (i.e. ∼20%) (Figure 3B), which is the same level of repair as seen after higher UV doses. However, the six mutant cell lines had only 6–12% residual NER capacity (Figure 3B).

Discussion

Mutations in the ERCC3 gene cause XP, XP/CS or TTD, which exhibit a broad spectrum of severe clinical features in humans (18,25). At the cellular level, a defect in ERCC3 results in high UV sensitivity, reduced post-UV RNA synthesis recovery and decreased DNA repair due to defective NER (18). Considering that mutations in ERCC3 can result in diverse symptoms and there are only nine XP-B patients with identified mutations, the ercc3 hamster cell mutants are an attractive system for exploring mutation versus phenotype.

Physical hemizygosity is known for several regions of the CHO genome (26–30). Our DNA sequencing data showed no double peaks in the site of mutations, i.e. all ERCC3 mRNAs were the same within wild-type cells and each mutant, confirming functional hemizygosity of the ERCC3 region. We do not know whether the second allele is silent or deleted as we did not perform genomic sequencing.

Remarkably, we identified one mutation within helicase domain III of ERCC3 protein (UV68) (Table I, Supplementary Figure S1, available at Mutagenesis Online). The ERCC3 protein, as a member of the helicase superfamily 2 with 3′ → 5′ polarity, possesses seven helicase domains which are highly conserved from yeast to humans (14). Specific mutations in helicase domains III and VI do not affect promoter opening in transcription or opening the DNA around the damage in NER, respectively (31,32). The ATPase activity of ERCC3 is essential for both events. However, helicase activity is necessary for promoter escape (31). For NER, ERCC3 functions as a wedge using ATP to keep both strands of the DNA around the lesion apart. As helicase domain III couples ATP hydrolysis to DNA unwinding (33), mutation V471F of UV68 cell line probably spoils this function of the ERCC3 protein. It is XPD that unwinds the DNA (32). The recent three-dimensional crystal structure of an archaea XPB homolog highlights the presence of additional functionally important motifs, the RED and the Thumb (15). However, there are no identified patients with mutations in helicase or Thumb domains (18), suggesting that such changes are not tolerated during development in human cells. Although the UV68 mutant was the least sensitive to UV light, it ranks among the lowest in CPD removal (Figures 1B and 3B). Since helicase domain III region is involved in a protein–protein interaction with protein 53 (p53) (34), the phenotype suggests an alteration in p53-mediated apoptosis (35). Moreover, UV68 is unable to restore RNA synthesis after UV irradiation, confirming a defect in TCR.

A putative mutation hot spot site in ERCC3 is located in the C-terminal region around the last intron–exon boundary (Supplementary Figure S1, available at Mutagenesis Online). The last exon has no known functional domain (except of Ser751—see below), and larger truncations may be lethal. Most base changes were substitutions although two insertions were also found. The base changes resulted in missense, nonsense, frameshift and splicing mutations causing truncation of the ERCC3 protein by 28 (UV113 and UV218), 37 (UV23) or 52 (UV78 and UV179) amino acids, respectively. With the exception of UV68, all the new mutants either miss Ser751 due to truncation of the protein or have instead the nonsense amino acid due to frameshift (Table I), which may lead to the complete inactivation of NER, as phosphorylation of Ser751 specifically prevents the 5′ incision of the lesion by ERCC1-xeroderma pigmentosum F endonuclease without modification of the damaged DNA opening or the 3′ incision (36). In spite of the different sites of mutations, there was little difference in efficiency of NER among hamster mutants after UV-C irradiation (Figure 3B) probably due to the same defect in NER, i.e. defect in 5′ incision. Although the wild-type AA8 cell line repaired ∼20% of UV-induced CPDs from the whole genome, it was able to efficiently recover RNA synthesis after UV irradiation. On the other hand, the GGR efficiency ranged from 6 to 12% in ercc3 mutant cells, and they were unable to repair actively transcribing genes because of impaired RRS after UV light (Figure 2B). It is important to note that GGR was measured by the comet assay after a very low UV dose of only 0.25 J/m² as higher doses saturated the assay sensitivity (data not shown) and did not allow us to distinguish between mutants. On the other hand, after such a low dose, there is no difference in survival among the ercc3 mutants (Figure 1B).

NER mainly repairs bulky DNA adducts and helix-distorting lesions but is considered to be a possible back-up system for BER to remove oxidative damage (37–40). The mild sensitivity of the CHO mutants to RO (Figure 1A) suggests the involvement of ERCC3 protein either in NER-dependent removal of a fraction of the oxidative lesions or in NER-independent repair (i.e. transcription-coupled BER) or some unknown repair. In fact, a group of bulky oxidative DNA lesions has been discovered. Biochemical studies demonstrated that these lesions could markedly block DNA replication and transcription and are repaired by NER (41). However, it is known that xpa knockout cells, which are devoid of global NER and TCR, are insensitive to acute oxidative stress induced by ionizing radiation (IR) (42,43), questioning the idea of an involvement of NER or at least XPA in the repair of oxidative lesions. On the other hand, ercc3^XP-CS primary mouse embryonic fibroblasts were slightly (∼1.3-fold) sensitive to IR (44) as were CHO ercc3 mutants to RO-induced oxidative damage (Figure 1A, Table II). The spectra of the oxidative damage induced by various oxidative agents obviously differ as does the sensitivity of ercc3 mutants. It seems that certain TFIIH mutations compromise the repair of certain oxidative DNA lesions occurring independently of XPA and classical NER (44).

It is less likely that the sensitivity of ercc3 mutants is caused by DNA damage-independent toxicity as RO treatment without photosensitization was shown to be non-toxic for hamster cells (10). Thorslund et al. (10) used the trypan blue exclusion method and 250 μM RO treatment. We have used different methods of cytotoxic analyses (i.e. colony-forming ability assay) and the doses of RO used were substantially lower (up to 1.5 μM).

Approximately 70% of the RO-induced DNA base modifications are 8-oxoG (24), which is repaired primarily by BER via OGG1 DNA glycosylase (45,46). Since global BER of oxidative DNA damage measured by comet assay is not reduced in ercc3 mutants, the hamster ERCC3 protein is not involved in this process (Figure 3A). However, photo-excited RO markedly reduced RRS in some ercc3 mutants, suggesting interference of oxidative damage with transcription (Figure 2A). It was recently shown that the 8-oxoG lesion does not block transcription in vitro (47–49). Moreover, the concentration of RO we used is too low (0.1 μM) to induce enough 8-oxoG (50) to directly hit the average transcription unit so some kind of damage signalling may be interfering with transcription. We cannot rule out the possibility that the transient delay of RRS creates a signal for cell cycle arrest or repair. Wild-type cells and some ercc3 mutants resumed RNA synthesis within 8 h after treatment. RNA synthesis in the other mutants recovered by 24 h. Notably, the RRS in the ercc6 mutant has the same kinetics as wild-type cells.

TC-BER of oxidative DNA damage remains a possibility although several relevant publications were retracted or partially retracted (51–56). Genetic analyses in Escherichia coli showed that 8-oxoG is subject to TCR (57), but in CHO cells, Thorslund (10) found no preferential repair of 8-oxoG in transcribed genes compared to non-transcribed regions and did not detect any strand bias in the repair of the DHFR. On the other hand, a host cell reactivation assay with a single 8-oxoG in a transcribed sequence showed faster removal than in a non-transcribed region, indicating TCR of 8-oxoG in the same cells (8).

In conclusion, each newly described ercc3 mutant has a similar, severe defect in GGR and in TCR measured as RRS, which confirms the known role of ERCC3 in these processes. Furthermore, these mutants are slightly sensitive to oxidative damage, and this sensitivity is not caused by a defect in global BER.

Supplementary data

Supplementary Figure S1 is available at Mutagenesis Online.

Funding

Science and Technology Assistance Agency (Slovakia) (APVT-51-003202, APVV-0208-07); Slovak Scientific Grant Agency (VEGA-2/4007/24, VEGA-2/7014/27); US–Slovak Science and Technology Joint Fund (031/2001); European Social Fund Project (13120200038) to E.K.P.; US Department of Energy, University of California; Lawrence Livermore National Laboratory (DE-AC52-07NA27344); National Cancer Institute/National Institutes of Health (CA89405, CA112566).