Introduction

The human genome is under perpetual attack from a range of endogenous and exogenous genotoxic agents (Lindahl 1993). Genome stability is maintained by a set of integrated DNA repair pathways, which comprise at least 130 gene products (Wood et al. 2001). Of these pathways the most versatile is nucleotide excision repair (NER), a highly conserved mechanism that requires at least 30 polypeptides (Aboussekhra et al. 1995). Lesions removed by NER include UV-induced photoproducts. Loss of function of some NER genes is incompatible with life whilst defects in a subset of proteins can result in developmental abnormalities or cancer.

Xeroderma pigmentosum (XP) is characterised by hyperphotosensitivity to the UV constituent of sunlight. The hallmark of XP is a predisposition to skin neoplasia, with a >1,000-fold incidence of UV-induced squamous cell carcinoma and malignant melanoma. Classical XP is caused by mutations in one of seven genes XPA-XPG. Cockayne syndrome (CS) patients do not have a predisposition to UV-induced skin neoplasia but exhibit a battery of clinical manifestations, such as segmental progeria, cachetic dwarfism, deafness, retinal degeneration, neurological dysfunction and skeletal-associated abnormalities including osteoporosis and a bird-like face. Death occurs at a mean age of 12 years. Mutations in the CKN1 (CSA) and CSB genes are responsible for CS.

Whilst CS and XP cells have defective NER, the striking differences between the syndromes are explained by the fact that the NER mechanism is divided into two distinct sub-pathways. Global genome repair (GGR) facilitates the removal of lesions from the non-transcribed strand and transcriptionally silent regions of the genome. Transcription-coupled repair (TCR) preferentially removes lesions from the transcribed strand of active genes. In GGR, the recruitment of the 106 kDa XPC protein complexed with the 43 kDa hHR23B protein allows interaction with damaged DNA, facilitating the recognition of UV-induced lesions (Sugasawa et al. 1998). Conversely, the arrest of an elongating RNA polymerase II at a lesion on the transcribed strand of an active gene elicits the recruitment of TCR-specific proteins, including CSA. The 44 kDa CSA protein is encoded by the CKN1 gene and contains multiple WD-40 repeats (Henning et al. 1995). The two NER sub-pathways converge at the point where the presence of a lesion is confirmed. Excision, re-synthesis and ligation follow using a common mechanism.

Both XPC and CKN1 genes were cloned as a result of screening cDNA libraries for genes that corrected the UV sensitivity of XP-C and CS-A cell lines. The mutations present in the original CS-A line used for library screening (CS3BE) have not previously been identified (Henning et al. 1995) although it was reported that the cell line appeared to produce no CSA protein (van Gool et al. 1997). Here we report a novel mutation in the CKN1 gene of this patient and the defect in a previously uncharacterised XP-C cell line, XP8CA.

Materials and methods

Cell lines and culture conditions

MRC-5 foetal lung fibroblasts were purchased from ECACC and were grown in monolayers in Eagle’s minimum essential medium with Earle’s salts (EMEM, Invitrogen), supplemented with 10% foetal calf serum (FCS; Imperial Laboratories, London), 1× MEM non-essential amino acids (NEAA) without L-glutamine. XP-C (GM02996, XP8CA), CS-A (GM01856, CS3BE) and XP-A (GM05509, XP12BE) diploid primary human fibroblasts were purchased from Coriell and grown as monolayers in EMEM with Earle’s salts with 10% FCS, 2× MEM NEAA without L-glutamine, 2× essential amino acids MEM, 2 mM L-glutamine, 2× MEM vitamins. All media were supplemented with penicillin (10,000 U/ml) and streptomycin (10 μg/ml) and cells grown at 37°C in a humidified incubator with 5% CO2. G418 (400 μg/ml) was used for cells undergoing drug selection.

Telomerase immortalisation of cell lines

The catalytic sub-unit of telomerase, human telomerase reverse transcriptase (hTERT), was transduced into MRC-5, XP8CA, CS3BE and XP12BE primary cells using an amphotropic retroviral vector, pBABE neo hTERT, using protocols previously described (Wyllie et al. 2000).

PCR amplification of XPC and CKN1 gene fragments from isolated genomic DNA

Genomic DNA was isolated and purified using RNase A treatment and proteinase K digestion, followed by a standard phenol/chloroform extraction. PCR primer sets (Table 1), used to amplify XPC and CKN1 coding regions and intron–exon boundaries from XP8CA and CS3BE genomic DNA were devised using NCBI sequences NT_022517.17 and NT_006713.14, respectively. Thermocycling was performed on an MJ Tetrad. Amplification conditions were 95°C for 12 min, followed by 33–35 cycles of 30 s each at 94°C, the primer specific annealing temperature, and 72°C, ending with a final 10-min extension step at 72°C.

Table 1 Forward and reverse primer sets for CKN1 and XPC amplification from genomic DNA
Full size table

Screening XPC and CKN1 for genomic DNA variants

PCR products were purified using an exonuclease I, shrimp alkaline phosphatase PCR purification protocol (Amersham). PCR amplification products were prepared for sequencing using the BigDye Terminator version 3.1 Cycle Sequencing Reaction Ready Kit (Applied Biosystems) and excess dye terminators removed using Montage Seq96 Sequence Reaction Cleanup Kits (Millipore). Sequence analysis of XPC and CKN1 was performed on an ABI PRISM 3100 Genetic Analyser.

Analysis of CS3BE CKN1 mRNA

Total RNA was isolated using TRIzol and reverse transcribed with SuperScript II (both Invitrogen). Amplification of CKN1 cDNA was performed using Taq polymerase (Promega) using the primers 5′-GGT TTT TGT CCG CAC GCC AA-3′ (forward) and 5′-GTG GAG ACC AGG AAA CTG CT-3′ (reverse). PCR products were ligated into the TA cloning vector pCR2.1 and transformed into INVαF′ Escherichia coli One Shot cells (Invitrogen). Individual colonies were expanded and sequenced (as above) in both directions using M13 forward and reverse primers.

Immunoblotting

Whole-cell extracts were prepared, as previously described (Evans et al. 2003). Proteins were separated on 8% (XPC) and 12% (CSA) SDS polyacrylamide gels, transferred to Immobilon polyvinylidene difluoride (PVDF) membranes (Millipore) and exposed to either an anti-XPC mouse monoclonal antibody (ab6264; Abcam) or anti-CSA goat polyclonal antibody (W-16: sc-10997; Santa Cruz). Blots were visualised using an appropriate HRP-coupled secondary antibody and an ECL system using Hyperfilm (Amersham). Equal protein loading was confirmed by staining the PVDF membrane with India ink.

Results and discussion

Primary human diploid fibroblasts have a limited lifespan in culture due to telomere-dependent replicative senescence. In order to provide sufficient material for biochemical analysis, we infected MRC-5, CS3BE, XP8CA and XP12BE with a retrovirus expressing the catalytic subunit of telomerase (hTERT) and isolated subclones. The extended lifespan of these lines confirmed that hTERT expression had immortalised the cultures (not shown). The repair-defective cell lines also retained their UV sensitivity (not shown). Genomic DNA from these cell lines was isolated and the 12 exons of CKN1 and 16 exons of XPC were amplified by PCR (Table 1). PCR products were sequenced and base changes identified by comparing to consensus cDNA sequences (NM_000082 [CKN1]; NM_004628 [XPC]).

The original assignment of patient XP8CA to XP complementation group C was made of the basis of cell fusion studies. In XP8CA, a homozygous 2 bp TG deletion, was identified in codon 547 at position 1639. This would result in a frame shift and a predicted protein product of 63.4 kDa. Two further homozygous changes were identified: 2061A>G (R687R) and 2815A>C (K939Q). Immunoblotting using an XPC-specific antibody confirmed the absence of full-length protein in XP8CA. XPC protein was present in CS-A and XP-A and normal MRC-5 cell extracts (Fig. 1). XPC protein levels were slightly reduced in the CS-A cell extract. XPC is an inducible protein (Adimoolam and Ford 2002), and the lower expression in CS3BE cells reflects a different steady state level compared to other cell lines. The presence of a further mutation at K939 is likely to have no effect because firstly, translation terminates prior to the mutation, and secondly, K939 is adjacent to the final amino acid of the normal protein. The majority of previously reported XPC mutations result in truncated proteins (Chavanne et al. 2000), and an identical mutation to that in XP8CA has been reported in the patient XP4PA (Li et al. 1993). The relationship between patients XP8CA and XP4PA is unknown although they were originally identified in different laboratories.

Fig. 1
figure 1

Loss of expression of XPC and CSA proteins in mutant cell lines. a Protein extracts (20 μg) from a normal (MRC-5) and DNA repair-defective XP-A, XP-C and CS-A cell lines were separated on SDS-PAGE gels, transferred and membranes probed with an anti-XPC antibody. A band of approximately 125 kDa corresponding to XPC protein was absent in XP8CA consistent with the predicted truncated product. b Protein extracts (20 μg) from MRC-5 and CS3BE (CS-A) were separated as above and transferred protein probed with an anti-CSA antibody. CS3BE cell extracts showed no expression of a protein of approximately 44 kDa (upper band) that was present in MRC-5 extracts

Full size image

CS3BE, assigned to CS complementation type A, was a compound heterozygote for CKN1 mutations 37G>T (E13X) and 479C>T (A160V). Immunoblot analysis of CS3BE and MRC-5 cell extracts using an anti-CSA antibody detected a non-specific cross-reacting band (Fig. 1). However, a band corresponding to CSA was present in MRC-5 and absent in CS3BE, confirming the earlier report that the protein was not expressed in the cell line (van Gool et al. 1997). It was confirmed that the mutations were on separate alleles by reverse transcription of CS3BE total RNA followed by PCR. Products were sub-cloned and sequenced, and in clones where 37G>T was present, the 479C>T mutation was absent. No clones were recovered containing the 479C>T mutation although cDNA containing the wild-type 37G appeared to have spliced out the exon containing 479C>T. The instability of mRNA from the second allele is likely to explain the lack of CSA protein in the cells.

To date, the availability of CKN1 mutation data is limited because although mutations in the CKN1 gene account for approximately 20% of CS (Stefanini et al. 1996), only eight have been characterised (Cao et al. 2004; Ren et al. 2003). Here we report a third missense mutation in CKN1. The Utah Genome Center SNP database (http://www.genome.utah.edu/genesnps) and a previous study (Cao et al. 2004) indicate that this is not an SNP. The availability of genomes from other species has shown that the CSA protein is highly conserved (data not shown; http://www.ensembl.org). The two previously reported missense mutations (Q106P and A205P) lie in the WD-40 regions. The A160V mutation reported here lies between the second and third WD-40 repeats. If undetectably low levels of full-length protein are expressed in CS3BE, this is likely to be inactive as the sites of all three missense mutations (Q106, A205 and A160) are highly conserved in mammalian, chicken and puffer fish CSA protein homologues. Unlike Q106 and A205, A160 is retained in Schizosaccharomyces pombe and Arabidopsis thaliana. The role of CSA in DNA repair is still poorly understood, and the existence of this highly conserved region suggests a possible avenue for elucidating the function of this protein.