Xeroderma Pigmentosum p48 Gene Enhances Global Genomic Repair and Suppresses UV-Induced Mutagenesis

Molecular Cell, Vol. 5, 737–744, April, 2000, Copyright 2000 by Cell Press

Xeroderma Pigmentosum p48 Gene Enhances Global Genomic Repair and Suppresses UV-Induced Mutagenesis Jean Y. Tang,1,2 Byung Joon Hwang,1,2,6,7 James M. Ford,1,3,7 Philip C. Hanawalt,4 and Gilbert Chu1,2,5 1 Department of Medicine 2 Department of Biochemistry 3 Department of Genetics 4 Department of Biological Sciences Stanford University Stanford, California 94305

Summary UV-damaged DNA-binding activity (UV-DDB) is deficient in some xeroderma pigmentosum group E individuals due to mutation of the p48 gene, but its role in DNA repair has been obscure. We found that UVDDB is also deficient in cell lines and primary tissues from rodents. Transfection of p48 conferred UV-DDB to hamster cells, and enhanced removal of cyclobutane pyrimidine dimers (CPDs) from genomic DNA and from the nontranscribed strand of an expressed gene. Expression of p48 suppressed UV-induced mutations arising from the nontranscribed strand, but had no effect on cellular UV sensitivity. These results define the role of p48 in DNA repair, demonstrate the importance of CPDs in mutagenesis, and suggest how rodent models can be improved to better reflect cancer susceptibility in humans.

Introduction Nucleotide excision repair (NER) removes a broad spectrum of lesions. Many of the lesions are medically important, including cyclobutane pyrimidine dimers (CPDs) and 6–4 photoproducts induced by ultraviolet radiation (UV) from the sun, intrastrand cross-links generated by the anticancer drug cisplatin, and benzo(a)pyrene adducts derived from tobacco smoke (Friedberg et al., 1995). NER recognizes these lesions, excises the damaged oligonucleotide, and restores the DNA by replication of the complementary undamaged DNA strand. Recognition of lesions for NER involves two mechanisms: transcription-coupled repair (TCR) and global genomic repair (GGR). TCR repairs lesions on the transcribed strand of expressed genes (Hanawalt, 1994), while GGR repairs lesions from both nontranscribed genomic DNA and the nontranscribed strand of expressed genes. TCR appears to be initiated by the arrest of RNA polymerase II at lesions (Hanawalt, 1994), but recognition of lesions for GGR has been poorly understood (Wood, 1999). 5

To whom correspondence should be addressed (e-mail: chu@ cmgm.stanford.edu). 6 Present address: Division of Biology, California Institute of Technology, Pasadena, California 91125. 7 These authors contributed equally to this work.

Inherited defects in NER occur in xeroderma pigmentosum (XP), an autosomal recessive disease associated with UV sensitivity and skin cancer susceptibility (Bootsma et al., 1998). This association has contributed to the clinical impression that sun sensitivity is a marker for skin cancer risk in the general population (Naylor, 1997). XP includes seven genetic complementation groups, corresponding to seven gene products involved in NER. Groups A, B, D, F, and G are defective in both TCR and GGR. Group C is defective only in GGR (Venema et al., 1991). XP variant defines individuals with the clinical symptoms of XP despite normal NER. XP group E is biochemically heterogeneous, with the absence of a highly specific UV-damaged DNA-binding activity (UV-DDB) in some individuals (Chu and Chang, 1988) but not others (Keeney et al., 1992). UV-DDB requires the expression of two subunits, p125 (or DDB1) and p48 (or DDB2) (Takao et al., 1993; Dualan et al., 1995; Hwang et al., 1996). The p48 gene is inactivated by missense mutations in those XP group E cells lacking UV-DDB (Nichols et al., 1996; Hwang et al., 1998). In normal cells, p48 expression is rate limiting for UV-DDB (Hwang et al., 1998), and p48 transcription is induced by the p53-dependent response to DNA damage (Hwang et al., 1999). Several observations have challenged the role of UVDDB in NER. Among XP cells, group E cells have the mildest defect in NER as measured by UV-induced unscheduled DNA synthesis and are the least sensitive to UV (Bootsma et al., 1998). Furthermore, UV-DDB is not required for NER in cell-free extracts (Aboussekhra et al., 1995; Kazantsev et al., 1996). Balanced against these observations is the recent discovery that XP group E cells either lacking or expressing UV-DDB are deficient in the GGR of CPDs (Hwang et al., 1999). Despite the imperfect correlation between the loss of UV-DDB and loss of GGR, we proposed that UV-DDB is involved in GGR. A direct test of this proposal was not possible because available XP group E cells were refractory to transfection. Wild-type hamster cell lines also fail to express UVDDB (Hwang et al., 1998) and are deficient in GGR of CPDs (Bohr et al., 1985). Here, we establish that several primary tissues from hamsters and mice likewise express very low levels of UV-DDB, defining a specific biochemical difference between rodents and humans in DNA repair. Transfection of hamster cells with human p48 enhanced GGR of CPDs, suppressing UV-induced mutagenesis without affecting survival. These results establish the role of p48 in DNA repair, illustrate the possible dissociation of sun sensitivity from cancer risk, and question the validity of rodent models for assessing cancer risk in humans. Results Hamster Cells Are Deficient in UV-DDB UV-DDB is absent or expressed at very low levels in 10 Chinese hamster cell lines (Hwang et al., 1998). To

Molecular Cell 738

determine whether the low level of UV-DDB was an artifact of cell culture or a property shared by primary hamster tissues, we measured UV-DDB in extracts of peripheral blood lymphocytes isolated directly from Syrian golden hamsters. Hamster lymphocytes contained barely detectable levels of UV-DDB (Figure 1A, lanes 5 and 6), while extracts from human peripheral blood lymphocytes contained at least 30-fold higher levels of UV-DDB (lanes 2 and 3). Similarly, UV-DDB was expressed at barely detectable levels in hamster V79 cells (lanes 9 and 10), but expressed at high levels in human HeLa cells (lanes 7 and 8). Low levels of UV-DDB were also seen in primary mouse tissues, including peripheral blood lymphocytes, spleen, heart, lung, kidney, and skin (data not shown). The lack of UV-DDB in rodent cells is not due to absence of the p48 or p125 gene, since treatment of hamster cells with the demethylating agent azacytidine induces expression of UV-DDB (Hwang et al., 1998). Thus, UVDDB may be transcriptionally suppressed in many rodent tissues. Hamster Cell Lines Transfected with p48 Express UV-DDB The failure of hamster cells to express UV-DDB presented an opportunity to define its role in DNA repair, since hamster cell lines are highly receptive to DNA transfection. An expression vector lacking a cDNA insert or containing either p48 or FLAG-p48 was transfected into V79 hamster cells, and clones were screened for UV-DDB by EMSA (Figure 1B). Little UV-DDB was detected in the parental V79 cell line or in V79 cells stably transfected with the control vector. However, clones with p48 message or FLAG-p48 protein have significant levels of UV-DDB (Figure 1B). Thus, expression of p48 conferred UV-DDB to the hamster cells. Three clones were selected for further analysis: 1A, 3B4, and 5E. Expression of p48 had no significant effect on the division times of these clones: 16 hr for 1A (vec), 15 hr for 3B4 (p48), and 16 hr for 5E (FLAG-p48). p48 Is Required for Global Genomic Repair of CPDs but Not 6–4 Photoproducts To test the effect of p48 on GGR of 6–4 photoproducts, unreplicated DNA from UV-irradiated cells was probed with a monoclonal antibody against 6–4 photoproducts. Expression of p48 in hamster cells had no effect on the already rapid repair of 6–4 photoproducts (Figure 2A). To test the effect of p48 on GGR of CPDs, unreplicated DNA from UV-irradiated cells was probed with a monoclonal antibody against CPDs. GGR of CPDs was undetectable in the hamster clone transfected with control vector, consistent with previous reports that hamster cell lines are defective in GGR of CPDs (Bohr et al., 1985), but occurred at significant levels in hamster clones expressing FLAG-p48 or p48 (Figure 2B). p48 Is Not Required for Transcription Coupled Repair of CPDs To determine whether p48 has a role in TCR, we measured CPD repair in the transcribed and nontranscribed strands of the DHFR gene. Wild-type hamster cells (wt) and hamster cells transfected with control vector showed

Figure 1. p48 Confers UV-DDB to Hamster Cells (A) Deficiency of UV-DDB in hamster cells. UV-DDB was measured in human and hamster peripheral blood lymphocytes and human (HeLa) and hamster (V79) cell lines. Whole-cell extracts were incubated with a 148 bp 32P-labeled DNA probe, which was either left intact (⫺) or irradiated with 5000 J/m2 of UV (⫹), and resolved by electrophoretic mobility shift assay (EMSA) (Hwang et al., 1999). B1 and B2 mark the mobility of UV-DDB/DNA complexes. F marks the mobility of free DNA probe. (B) Expression of UV-DDB in cells transfected with p48. Hamster V79 cells were stably transfected with an expression vector lacking a cDNA insert (vec) or containing cDNA encoding p48 or FLAG-p48. Extracts (2 ␮g) were assessed for UV-DDB by EMSA. Levels of p48, p125, and actin mRNA were measured by Northern blot of total RNA (10 ␮g per lane), and levels of FLAG-p48 protein were measured by immunoblot with anti-FLAG antibodies (2 ␮g/ml) (Hwang et al., 1998).

negligible levels of CPD repair in the nontranscribed strand (Figure 2C) as previously reported (Mellon et al., 1987). Hamster cells expressing p48 showed a significantly higher level of CPD repair on the nontranscribed strand, consistent with the role of p48 in GGR. Expression of p48 had no effect on the proficient repair of CPDs from the transcribed strand. p48 Expression Does Not Affect UV Survival To determine the physiological significance of GGR of CPDs, we measured the effect of p48 on UV survival. Colony formation after UV was indistinguishable among

p48 Enhances DNA Repair and Suppresses Mutagenesis 739

Figure 3. p48 Expression Does Not Affect UV Survival but Suppresses Mutagenesis Figure 2. p48 Enhances Global Genomic Repair of CPDs (A) GGR of 6–4 photoproducts. (B) GGR of CPDs. Repair was measured after exposure to UV in hamster cells expressing p48, FLAGp48, or vector. Each point represents the mean of three independent experiments. (C) TCR of CPDs. Removal of CPDs from the nontranscribed and transcribed strands of the DHFR gene was measured after exposure to UV.

wild-type hamster cells (V79, AA8) and hamster cells transfected with p48 or control vector (Figure 3A), demonstrating that GGR of CPDs does not affect UV survival. By contrast, decreased colony survival was observed in mutant cell lines with defects in the Cockayne syndrome B gene (UV61) and the XP group D gene (UV5) as previously reported (Friedberg et al., 1995). p48 Suppresses UV-Induced Mutagenesis The V79 cell line was derived from a male hamster and contains one copy of the X-linked HPRT gene. In V79 cells transfected with vector, the spontaneous mutation rate in the HPRT gene was very low, 0.7 mutations per 105 cells. When asynchronously growing cells were exposed to UV, the mutation rate increased to 15 per 105 cells at a dose of 2 J/m2 and 55 per 105 cells at a dose of 10 J/m2 (Figure 3B). Similar mutation rates (6.2 per 105 cells per J/m2) were reported previously for V79 cells (Zdzienicka et al., 1988). Hamster cells expressing p48 also had a low spontaneous mutation rate, 1 per 105 cells. When the cells expressing p48 or FLAG-p48 were exposed to UV, the mutation rate was 7.5 per 105 cells at a dose of 2 J/m2 and 20 per 105 cells at a dose of 10 J/m2. Thus, expression of p48 in hamster cells suppressed UV-induced mutagenesis 2- to 2.7-fold. Suppression of mutagenesis was similar in hamster cells growing asynchronously or synchronized in G1 at the time of UV exposure.

(A) UV survival. Wild-type hamster cells (V79 or AA8), V79 hamster cells transfected with p48 or vector, and hamster cells with mutations in the Cockayne syndrome group B gene (CSB⫺) or the XP group D gene (XPD⫺) were exposed to UV from a Westinghouse IL782–30 germicidal lamp at an incident dose of 0.33 J/m2/s. Surviving colonies were counted after 2 weeks. (B) UV-induced mutagenesis. Mutations in the HPRT gene were measured in hamster cells stably transfected with p48, FLAG-p48, or vector. Cells were grown asynchronously (asyn) or synchronized in G1 (syn) by serum starvation, exposed to different doses of UV, and selected for HPRT mutations with 6-thioguanine.

p48 Suppresses Mutations Only from Nontranscribed DNA Hamster cells were exposed to UV (2 J/m2) and grown in divided populations to select for independent HPRT mutant clones. Table 1 lists independent mutations arising in the hamster cells transfected with vector or p48. Most mutations were single base pair substitutions, 29% of which were C-to-T transitions. Significantly, 3 tandem base pair mutations at dipyrimidines were observed: a CT-to-TC transition, a CC-to-TT transition, and a TT-toAA transversion. Such mutations have been reported almost exclusively in cells from UV-induced skin tumors, and very rarely for internal tumors (Giglia et al., 1998). The damaged DNA strand that led to each HPRT mutation could be inferred in most cases from sequence context. For hamster cells transfected with vector, the vast majority of UV-induced lesions (94%) occurred at sites containing adjacent pyrimidines, where CPDs and 6–4 photoproducts could have formed. Only 22% (7 of 32) of the mutations were attributable to dipyrimidine lesions on the transcribed strand, while 72% (23 of 32) were from the nontranscribed strand (Table 1), consistent with the poor GGR of CPDs in hamster cells. By contrast, for hamster cells expressing p48, 36% (13 of

Molecular Cell 740

Table 1. UV-Induced HPRT Mutations from Hamster Cells Transfected with Vector or p48 Mutations in Cells Transfected with Vector (1A) Strand of Mutagenic Lesiona ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ NP UD

Mutations in Cells Transfected with p48 (3B4)

Nucleotide Position in ORF

Base Change with Adjacent Sequence in Nontranscribed Strand

Amino Acid Change

Strand of Mutagenic Lesiona

62 71 74 110 145 181 194 194 221 223 254 296 297 305 325 325 374 463 473 487 490/491 508 596 124 229 247 539 544 589 606 611 Del 486–537

ATT (T⬎A) ATT GTA (T⬎A) TCC TTC (C⬎T) TAA TTA (T⬎A) TCC AGA (C⬎T) TTG CAT (C⬎T) ACA CCC (T⬎G) CTG CCC (T⬎A) CTG AAT (T⬎G) CTT TTC (T⬎C) TTG CAC (T⬎C) GAA ATT (T⬎G) TAT TTT (T⬎A) ATC GAC (T⬎G) GAA GAT (C⬎T) AGT GAT (C⬎T) AGT CTT (T⬎G) AAC AAC (C⬎T) CCA TGG (T⬎G) TAA AGC (T⬎G) TGC TTG (CT⬎TC) GGT TCT (C⬎T) GAA ACT (T⬎C) CAG GTG (A⬎T) TTA GCT (G⬎T) ACC ATT (A⬎T) AAG TTG (G⬎T) ATT TTT (G⬎T) AAA AAT (G⬎T) AGT TTT (G⬎T) AAT ATC (A⬎C) TAT Not defined

L⬎Stop I⬎N P⬎L I⬎N L⬎F H⬎Y L⬎R L⬎H F⬎C P⬎L L⬎P F⬎C F⬎L L⬎R Q⬎Stop Q⬎Stop L⬎Stop P⬎S V⬎G L⬎V L⬎S R⬎Stop F⬎S I⬎F D⬎Y K⬎Stop G⬎V E⬎Stop E⬎Stop L⬎F H⬎P Trunc

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ NP NP NP UD UD UD UD UD

Nucleotide Position in ORF

Base Change with Adjacent Sequence in Nontranscribed Strand

Amino Acid Change

114 221 223 245 245 274 295 297 347 410 464/465 488 532 542 542/543 40 119 119 139 247 538 538 538 543 556 580 606 628 349 522 584 246 Del 323–390 Del 385–405 Del 385–402 Del 533–614

TTC (C⬎T) TCA AAT (T⬎C) CTT TTC (T⬎G) TTG ACA (T⬎A) TAA ACA (T⬎A) TAA AGA (T⬎G) CCA GAT (T⬎C) TTA TTT (T⬎G) ATC AAG (T⬎G) TAT TAA (T⬎G) TGA ACC (CC⬎TT) AAA GCT (T⬎G) GCT GAC (T⬎G) TTG GAT (T⬎A) TGA GAT (TT⬎AA) GAA GAT (G⬎T) AAC ATG (G⬎A) AGT ATG (G⬎A) AGT ACT (G⬎A) AAA ATT (A⬎T) AAG GTT (G⬎A) GAT GTT (G⬎A) GAT GTT (G⬎A) GAT TTT (G⬎T) AAA GAC (A⬎T) AGT CTT (G⬎A) ACT TTT (G⬎T) AAT AGT (G⬎T) AAA GTT (A⬎T) TTG ATA (T⬎A) AGG ACT (A⬎C) TAA CAT (⫺T) AAA Not defined Not defined Not defined Not defined

P⬎L F⬎S F⬎V I⬎N I⬎N S⬎A F⬎L F⬎L V⬎G I⬎S P⬎L L⬎W F⬎V F⬎Y F⬎Stop E⬎Stop G⬎E G⬎E E⬎K K⬎Stop G⬎R G⬎R G⬎R F⬎Stop K⬎Stop D⬎N L⬎F E⬎Stop I⬎F Y⬎Stop Y⬎S Trunc Trunc Trunc Trunc Trunc

a The strand of the mutagenic lesion was either the transcribed (⫹) or nontranscribed (⫺) strand depending on which strand contained a dipyrimidine sequence at the site of the mutation. NP, no pyrimidine dimer was associated with the mutation. UD, strand with the mutagenic lesion was undetermined. Deletion (Del) in the cDNA led to a truncated (Trunc) open reading frame (ORF), but the mutation in genomic DNA was not defined.

36) of the mutations were attributable to the transcribed strand, while 42% (15 of 36) were from the nontranscribed strand. The effect of p48 on the strand specificity of mutations was statistically significant (p ⫽ 0.041 by Fisher’s exact test). To calculate mutation rates for transcribed and nontranscribed DNA, the percentage of mutations found on each strand was multiplied by the overall mutation rate for the cell line (Figure 4A). Expression of p48 did not strongly affect the mutation rate on the transcribed strand, but decreased the mutation rate 3.7-fold on the nontranscribed strand, consistent with the role of p48 in repairing nontranscribed DNA. UV induces CPDs at both TT and non-TT (CC, CT, or TC) dipyrimidine sites. In hamster cells transfected with control vector, more mutations arose from non-TT than from TT dipyrimidines (Table 1). This bias was seen on the transcribed strand (1 mutation from TT, 6 from nonTT) and the nontranscribed strand (7 from TT, 12 from non-TT). When p48 was expressed in hamster cells,

there was no significant effect on mutations from the transcribed strand (3 from TT, 10 from non-TT). However, on the nontranscribed strand, the mutations from non-TT dipyrimidines declined significantly (9 from TT, 3 from non-TT). The difference in the effect of p48 on mutations arising from TT and non-TT dipyrimidines was statistically significant (p ⫽ 0.037 by Fisher’s exact test), suggesting that UV-DDB targets non-TT CPDs for repair more efficiently than TT CPDs (Figure 4B). Discussion The Role of UV-DDB in Intact Cells Expression of p48 conferred UV-DDB to hamster cells and enhanced removal of CPDs from genomic DNA and from the nontranscribed strand of a transcribed gene. Thus, UV-DDB is required for the GGR of CPDs, presumably by binding to CPDs on nontranscribed DNA and recruiting the core NER complex. GGR of 6–4 photoproducts was not affected by p48

p48 Enhances DNA Repair and Suppresses Mutagenesis 741

Figure 4. p48 Suppresses Mutations Arising from Pyrimidine Dimers in the Nontranscribed Strand of DNA (A) Strand bias. UV-induced mutations in the HPRT gene were measured in hamster cells stably transfected with vector or p48. The percentage of mutations on the transcribed strand (TS) and nontranscribed strand (NTS) was multiplied by the overall mutation rates to obtain strand-specific mutation rates. (B) TT versus non-TT bias. Mutation rates are shown for TT and non-TT dipyrimidine sequences. Cases in which the dipyrimidine could have been either TT or non-TT (Table 1) were all on the NTS, and these were not included in the analysis.

expression in hamster cells. However, XP group E cells show a delay in the repair of 6–4 photoproducts (Hwang et al., 1999; Itoh et al., 1999). UV-DDB binds to 6–4 photoproducts (Treiber et al., 1992) in addition to CPDs (Hwang and Chu, 1993; Reardon et al., 1993), cisplatin adducts, and many other lesions (Payne and Chu, 1994). These observations suggest that 6–4 photoproducts are targeted for repair by both UV-DDB and a second protein (perhaps XPC/HR23B), which substitutes fully for the absence of UV-DDB in hamster cells but only partially in human XP group E cells. Expression of p48 suppressed mutagenesis from nontranscribed DNA, with a larger effect on non-TT CPDs than on TT CPDs (Figure 4). UV-DDB has a weak affinity for TT CPDs in the predominant cis-syn conformation (Reardon et al., 1993). Consistent with this, when DNA was damaged with a UV dose of 100 J/m2 to produce predominantly cis-syn TT CPDs, binding by UV-DDB was largely unaffected when the CPDs were removed by pretreatment of the DNA with photolyase (Hwang and Chu, 1993). On the other hand, when the DNA was damaged with higher UV doses to produce a significant number of non-TT CPDs, binding by UV-DDB decreased after pretreatment of the DNA with photolyase. Thus, both binding and mutagenesis data suggest that UVDDB recognizes non-TT CPDs more efficiently than TT CPDs. Three damage-specific DNA-binding proteins have now been implicated in NER: the XPA protein (Jones and Wood, 1993), the XPC/HR23B heterodimer (Sugasawa et al., 1998), and UV-DDB. During NER in cell-free extracts, binding of XPC/HR23B precedes XPA protein in one report (Sugasawa et al., 1998), but not in another (Wakasugi and Sancar, 1999). Additional experiments are required to determine the order in which UV-DDB, XPC/HR23B, and XPA protein direct the GGR of CPDs.

Failure of Cell-Free Systems to Reveal the Role of UV-DDB in NER Although UV-DDB appears to be dispensable for the reconstitution of NER in cell-free systems (Aboussekhra et al., 1995; Kazantsev et al., 1996), our experiments with intact cells reveal a critical role for UV-DDB in GGR. Indeed, the results of the cell-free experiments can now be better understood. The failure of UV-DDB to stimulate repair in the system of Kazantsev et al., who utilized an oligonucleotide substrate containing a 6–4 photoproduct, is consistent with our conclusion that 6–4 photoproducts are targeted for repair by either UV-DDB or a second protein in intact cells. Aboussekhra et al. utilized naked DNA substrates containing both CPDs and 6–4 photoproducts after exposure to a high dose of UV. Addition of UV-DDB resulted in only a 2-fold stimulation of NER. Furthermore, extracts from XP group E cells failed to reveal a repair defect (Otrin et al., 1998). Perhaps the high dose of UV (450 J/m2) required to generate the substrate for this assay does not fully test the very high specificity of UV-DDB for UV-damaged DNA (Hwang and Chu, 1993). Alternatively, the naked DNA substrate does not test possible interactions between chromatin and p48, which contains a WD domain conserved among several chromatin remodeling proteins (Hwang et al., 1998). CPDs on Nontranscribed DNA Do Not Affect Survival When hamster cells acquired GGR of CPDs, there was no effect on UV survival. Cells deficient in GGR of CPDs may survive by replicating unrepaired DNA with translesion DNA polymerases, such as polymerase ␩, which is mutated in XP variant (Johnson et al., 1999; Masutani et al., 1999). By contrast, UV61 hamster cells, which have a specific defect in TCR of CPDs due to mutation of the Cockayne syndrome B gene, are extremely UVsensitive. Failure to repair CPDs on the transcribed DNA strand is associated with cell death, perhaps because bypass polymerases do not exist for RNA transcription as they do for DNA replication. XP group E cells are deficient in the GGR of CPDs, but unlike hamster cells, have a mild UV sensitivity. Our results suggest that the UV sensitivity in XP group E cells can be attributed to their mild deficiency in repairing 6–4 photoproducts. Consistent with this, XP group C cells have a severe defect in the GGR of 6–4 photoproducts and are much more sensitive to UV (van Hoffen et al., 1995). CPDs Are a Major Source of UV-Induced Mutations Previous reports have disagreed about the role of CPDs in mammalian mutagenesis (Zdzienicka et al., 1992; Vreeswijk et al., 1998). In these studies, unrecognized genetic differences may have affected the mutagenesis assay in comparisons between nonisogenic cell lines. This problem was circumvented by transfection of mouse fibroblasts with bacteriophage T4 endonuclease V (Kusewitt et al., 1998) or transfection of human XP group A cells with photolyase (Asahina et al., 1999), which specifically target CPDs for repair. UV-induced mutagenesis was suppressed, but the experiments involved ectopic repair systems not endogenous to mouse

Molecular Cell 742

or human cells. We utilized transfected hamster cells isogenic except for UV-DDB expression, which can be induced by demethylation with azacytidine and is therefore indigenous to hamster cells (Hwang et al., 1998). In our experiments, GGR of CPDs led to a 3.7-fold suppression of mutagenesis from the nontranscribed DNA strand (Figure 4A). Thus, CPDs are indeed a major source of UV-induced mutations.

Implications for Rodent and Human Carcinogenesis The discovery that UV-DDB was suppressed in many rodent tissues suggests that models of human carcinogenesis in rodents may have serious shortcomings. In the hairless mouse, which has been used as a model for UV-induced skin cancer, nearly all of the p53 mutations in the tumors arise from the nontranscribed strand (Dumaz et al., 1997), while in humans, only 54% of p53 mutations arise from the nontranscribed strand (Giglia et al., 1998). Inactivation of the Cockayne syndrome group B gene increases the risk for UV-induced skin cancer in mice (van der Horst et al., 1997) but not in humans (Nance and Berry, 1992). Our findings indicate that mice will reflect the skin cancer susceptibility of humans more accurately if they are engineered to express similar levels of p48. A dissociation between UV sensitivity and mutagenesis occurs in human diseases of DNA repair. Defective GGR of CPDs in XP group E leads to relatively mild UV sensitivity that is nevertheless associated with skin cancer. Conversely, defective TCR of CPDs in Cockayne syndrome causes severe UV sensitivity without affecting skin cancer risk, presumably because GGR of CPDs is sufficient to suppress mutagenesis. Our observation of a dissociation between UV sensitivity and cancer susceptibility has clinical implications. Sensitivity to sunburn has been accepted as a clinical indicator of skin cancer risk (Naylor, 1997). Our data suggest that this is not always true and that variations in GGR should be considered in assessing skin cancer risk.

Experimental Procedures Cell Lines and Primary Lymphocytes Cell lines were grown in DMEM medium supplemented with 10% fetal bovine serum. Rodent blood was collected by cardiac exsanguination of Syrian golden hamsters (Charles River, Kingston, NY) or Balb/c mice (generously provided by I. Weissman). Lymphocytes were isolated from the rodent blood using Lympholyte-M density separation medium (Cedarlane, Hornby, Canada). Lymphocytes were isolated from human peripheral blood by Ficoll-Paque PLUS density centrifugation (Pharmacia Biotech, Uppsala, Sweden).

DNA Transfection To generate cell lines expressing FLAG-p48, cells were transfected by calcium phosphate with 2 ␮g of pRSVneo and either 23 ␮g of the pBJ5 expression vector lacking a cDNA insert (vec) or containing the human p48 open reading frame fused at its N terminus to the FLAG epitope (FLAG-p48) (Hwang et al., 1998). To generate cell lines expressing unmodified p48, cells were transfected by Lipofectamine-PLUS (Gibco–BRL, Gaithersburg, MD) with 1.3 ␮g of pRSVneo and 2.7 ␮g of the pBJ5 expression vector containing p48. Two days after transfection, cells were plated in medium containing 500 ␮g/ml of the antibiotic G418 (Gibco-BRL, Gaithersburg, MD) to select for stable transfectants expressing the neo gene.

DNA Repair Assays For GGR, the relative number of UV-induced photoproducts in unreplicated genomic DNA was determined by an immunoblot assay with mouse monoclonal antibodies specific for either CPDs or 6–4 photoproducts (Mori et al., 1991; Ford and Hanawalt, 1997). In brief, after exposure to UV, hamster cells were incubated in growth medium containing bromodeoxyuridine (BrdUrd). Incorporation of BrdUrd in newly replicated DNA permitted isolation of unreplicated DNA by cesium chloride density gradient sedimentation. Unreplicated DNA was adsorbed to a nylon membrane in triplicate using a slot-blot apparatus, and incubated with mouse monoclonal antibody directed against either CPDs or 6–4 photoproducts and horseradish peroxidase-conjugated goat anti-mouse antibody. For TCR, repair of CPDs was examined in the dihydrofolate reductase (DHFR) gene, as previously described (Mellon et al., 1987). In brief, unreplicated DNA from UV-irradiated cells was cleaved with KpnI, and treated or mock treated with T4 endonuclease V, which specifically nicks DNA at CPD sites. The DNA was resolved by electrophoresis under denaturing conditions, transferred to a nylon membrane, and hybridized with RNA probes specific for either the transcribed or nontranscribed strand of the DHFR gene. The ratio of intact restriction fragments in the endonuclease-treated and -untreated DNA was used to calculate the number of CPDs per fragment from Poisson statistics. HPRT Mutants UV-induced mutants in the hypoxanthine-guanine phosphoribosyl transferase (HPRT) gene were generated as previously reported (Zdzienicka et al., 1988), except that before UV irradiation, hamster cells were either grown asynchronously or arrested in G1 phase by starvation in medium containing low serum (0.2%) for 48 hr. G1arrested cells were exposed to UV and then switched to standard medium containing 10% fetal bovine serum. The cells were passaged for 7 days to allow loss of HPRT protein in the newly induced HPRT mutants, then selected in medium supplemented with 5 ␮g/ ml of 6-thioguanine (Sigma, St. Louis, MO). Surviving colonies were counted after 10 days. To isolate independent HPRT mutants from the 3B4 (p48) and 1A (vector) cell lines, cells were exposed to a UV dose of 2 J/m2, and immediately divided into 36 separate populations to ensure that each HPRT mutant was independently generated. Eight days after UV irradiation, each population was grown in medium containing 6-thioguanine to select for HPRT mutants. One colony was isolated from each dish. To define HPRT mutations, total cytoplasmic RNA was isolated from each HPRT mutant clone using the RNeasy Mini Kit (Qiagen, Chatsworth, CA). Synthesis of cDNA with reverse transcriptase was initiated with the primer (5⬘-TAATTTTACTGGGAACAT-3⬘). HPRT cDNA was amplified by PCR with high-fidelity Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA) using the forward primer (5⬘-TCT GCGGGCTTCCTCCTCAC-3⬘) and the reverse primer (5⬘-ATGGGAC TCCTCGTGTTTGC-3⬘). The DNA was purified and sequenced on both strands. Acknowledgments The authors thank A. van Zeeland, M. Zdzienicka, L. DeFazio, V. Goss, and T. Tan for helpful suggestions. The research was supported by the Medical Scientist Training Program and a Paul and Daisy Soros Fellowship (J. Y. T.), the Janet M. Shamberger Fellowship Fund (B. J. H.), Clinical Investigator Award K08-CA64330 from the National Cancer Institute (J. M. F.), Outstanding Investigator Grant CA44349 from the National Cancer Institute (P. C. H), and gifts from Graham and Jane Nissen and a Burroughs Wellcome Clinical Investigator Award (G. C.). Received December 6, 1999; revised March 7, 2000. References Aboussekhra, A., Biggerstaff, M., Shivji, M., Vilpo, J., Moncollin, V., Podust, V., Protic, M., Hubscher, U., Egly, J., and Wood, R. (1995).

p48 Enhances DNA Repair and Suppresses Mutagenesis 743

Mammalian DNA nucleotide excision repair reconstituted with purified protein components. Cell 80, 859–868. Asahina, H., Han, Z., Kawanishi, M., Kato, T., Ayaki, H., Todo, T., Yagi, T., Takebe, H., Ikenaga, M., and Kimura, S. (1999). Expression of a mammalian DNA photolyase confers light-dependent repair activity and reduces mutations of UV-irradiated shuttle vectors in xeroderma pigmentosum cells. Mutat. Res. 435, 255–262. Bohr, V., Smith, C., Okumoto, D., and Hanawalt, P. (1985). DNA repair in an active gene: removal of pyrimidine dimers from the DHFR gene of CHO cells is much more efficient than in the genome overall. Cell 40, 359–369. Bootsma, D., Kraemer, K., Cleaver, J., and Hoeijmakers, J. (1998). Nucleotide excision repair syndromes: xeroderma pigmentosum, Cockayne syndrome, and trichothiodystrophy. In The Genetic Basis of Human Cancer, B. Vogelstein and K. Kinzler, eds. (New York: McGraw-Hill), pp. 245–274. Chu, G., and Chang, E. (1988). Xeroderma pigmentosum group E cells lack a nuclear factor that binds to damaged DNA. Science 242, 564–567. Dualan, R., Brody, T., Keeney, S., Nichols, A., Admon, A., and Linn, S. (1995). Chromosomal localization and cDNA cloning of the genes (DDB1 and DDB2) for the p127 and p48 subunits of a human damage-specific DNA binding protein. Genomics 29, 62–69. Dumaz, N., van Kranen, H., de Vries, A., Berg, R., Wester, P., van Kreijl, C., Sarasin, A., Daya-Grosjean, L., and de Gruijl, F. (1997). The role of UV-B light in skin carcinogenesis through the analysis of p53 mutations in squamous cell carcinomas of hairless mice. Carcinogenesis 18, 897–904. Ford, J.M., and Hanawalt, P.C. (1997). Expression of wild-type p53 is required for efficient global genomic nucleotide excision repair in UV-irradiated human fibroblasts. J. Biol. Chem. 272, 28073–28080. Friedberg, E.C., Walker, G.C., and Siede, W. (1995). DNA Repair and Mutagenesis (Washington, DC: ASM Press). Giglia, G., Dumaz, N., Drougard, C., Avril, M., Daya-Grosjean, L., and Sarasin, A. (1998). p53 mutations in skin and internal tumors of xeroderma pigmentosum patients belonging to the complementation group C. Cancer Res. 58, 4402–4409. Hanawalt, P. (1994). Transcription-coupled repair and human disease. Science 266, 1957–1958. Hwang, B.J., and Chu, G. (1993). Purification and characterization of a protein that binds to damaged DNA. Biochemistry 32, 1657–1666. Hwang, B.J., Liao, J., and Chu, G. (1996). Isolation of a cDNA encoding a UV-damaged DNA binding factor defective in xeroderma pigmentosum group E cells. Mutat. Res. 362, 105–117. Hwang, B.J., Toering, S., Francke, U., and Chu, G. (1998). p48 activates a UV-damaged DNA binding factor and is defective in xeroderma pigmentosum group E cells that lack the binding activity. Mol. Cell. Biol. 18, 4391–4399. Hwang, B.J., Ford, J., Hanawalt, P.C., and Chu, G. (1999). Expression of the p48 xeroderma pigmentosum gene is p53-dependent and is involved in global genomic repair. Proc. Natl. Acad. Sci. USA 96, 424–428. Itoh, T., Mori, T., Ohkubo, H., and Yamaizumi, M. (1999). A newly identified patient with clinical xeroderma pigmentosum phenotype has a non-sense mutation in the DDB2 gene and incomplete repair in (6–4) photoproducts. J. Invest. Dermatol. 113, 251–257. Johnson, R., Kondratick, C., Prakash, S., and Prakash, L. (1999). hRAD30 mutations in the variant form of xeroderma pigmentosum. Science 285, 263–265. Jones, C., and Wood, R. (1993). Preferential binding of the xeroderma pigmentosum group A complementing protein to damaged DNA. Biochemistry 32, 12096–12104. Kazantsev, A., Mu, D., Nichols, A.F., Zhao, X., Linn, S., and Sancar, A. (1996). Functional complementation of xeroderma pigmentosum complementation group E by replication protein A in an in vitro system. Proc. Natl. Acad. Sci. USA 93, 5014–5018. Keeney, S., Wein, H., and Linn, S. (1992). Biochemical heterogeneity in xeroderma pigmentosum complementation group E. Mutat. Res. 273, 49–56. Kusewitt, D., Dyble, J., Sherburn, T., Ryan, S., and Ji, J. (1998).

Altered UV resistance and UV mutational spectrum in repair-proficient murine fibroblasts expressing endonuclease V. Mutat. Res. 407, 157–168. Masutani, C., Kusumoto, R., Yamada, A., Dohmae, N., Yokoi, M., Yuasa, M., Araki, M., Iwai, S., Takio, K., and Hanaoka, F. (1999). The XPV (xeroderma pigmentosum variant) gene encodes human DNA polymerase ␩. Nature 399, 700–704. Mellon, I., Spivak, G., and Hanawalt, P. (1987). Selective removal of transcription-blocking DNA damage from the transcribed strand of the mammalian DHFR gene. Cell 51, 241–249. Mori, T., Nakane, M., Hattori, T., Matsunaga, T., Ihara, M., and Nikaido, O. (1991). Simultaneous establishment of monoclonal antibodies specific for either cyclobutane pyrimidine dimer or (6–4) photoproduct from the same mouse immunized with ultravioletirradiated DNA. Photochem. Photobiol. 54, 225–232. Nance, M., and Berry, S. (1992). Cockayne syndrome: review of 140 cases. Am. J. Hum. Genet. 42, 68–84. Naylor, M. (1997). Erythema, skin cancer risk, and sunscreens. Arch. Dermatol. 133, 373–375. Nichols, A., Ong, P., and Linn, S. (1996). Mutations specific to the xeroderma pigmentosum group E Ddb- phenotype. J. Biol. Chem. 271, 24317–24320. Otrin, V.R., Kuraoka, I., Nardo, T., McLenigan, M., Eker, A.P.M., Stefanini, M., Levine, A.S., and Wood, R.D. (1998). Relationship of the xeroderma pigmentosum group E DNA repair defect to the chromatin and DNA binding proteins UV-DDB and replication protein A. Mol. Cell. Biol. 18, 3182–3190. Payne, A., and Chu, G. (1994). Xeroderma pigmentosum group E binding factor recognizes a broad spectrum of DNA damage. Mutat. Res. 310, 89–102. Reardon, J.T., Nichols, A.F., Keeney, S., Smith, C.A., Taylor, J.S., Linn, S., and Sancar, A. (1993). Comparative analysis of binding of human damaged DNA-binding protein (XPE) and Escherichia coli damage recognition protein (uvrA) to the major ultraviolet photoproducts: T[c,s]T, T[t,s]T, T[6–4]T, and T[Dewar]T. J. Biol. Chem. 268, 21301–21308. Sugasawa, K., Ng, J., Masutani, C., Iwai, S., van der Spek, P., Eker, A., Hanaoka, F., Bootsma, D., and Hoeijmakers, J. (1998). Xeroderma pigmentosum group C protein complex is the initiator of global genome nucleotide excision repair. Mol. Cell 2, 223–232. Takao, M., Abramic, M., Moos, M., Otrin, V., Wootton, J., McLenigan, M., Levine, A., and Protic, M. (1993). A 127 kDa component of a UVdamaged DNA-binding complex, which is defective in some xeroderma pigmentosum group E patients, is homologous to a slime mold protein. Nucleic Acids Res. 21, 4111–4118. Treiber, D., Chen, Z., and Essigmann, J. (1992). An ultraviolet lightdamaged DNA recognition protein absent in xeroderma pigmentosum group E cells binds selectively to pyrimidine (6–4) pyrimidone photoproducts. Nucleic Acids Res. 20, 5805–5810. van der Horst, G.T., van Steeg, H., Berg, R.J., van Gool, A.J., de Wit, J., Weeda, G., Morreau, H., Beems, R.B., van Kreijl, C.F., de Gruijl, F.R., et al. (1997). Defective transcription-coupled repair in Cockayne syndrome B mice is associated with skin cancer predisposition. Cell 89, 425–435. van Hoffen, A., Venema, J., Meschini, R., van Zeeland, A., and Mullenders, L. (1995). Transcription-coupled repair removes both cyclobutane pyrimidine dimers and 6–4 photoproducts with equal efficiency and in a sequential way from transcribed DNA in xeroderma pigmentosum group C fibroblasts. EMBO J. 14, 360–367. Venema, J., van Hoffen, A., Karcagi, V., Natarajan, A., van Zeeland, A., and Mullenders, L. (1991). Xeroderma pigmentosum complementation group C cells remove pyrimidine dimers selectively from the transcribed strand of active genes. Mol. Cell. Biol. 11, 4128–4134. Vreeswijk, M., Overkamp, M., Westland, B., van Hees-Stuivenberg, S., Vrieling, H., Zdzienicka, M., van Zeeland, A., and Mullenders, L. (1998). Enhanced UV-induced mutagenesis in the UV61 cell line, the Chinese hamster homologue of Cockayne’s syndrome B, is associated with defective transcription coupled repair of cyclobutane pyrimidine dimers. Mutat. Res. 409, 49–56. Wakasugi, M., and Sancar, A. (1999). Order of assembly of human DNA repair excision nuclease. J. Biol. Chem. 274, 18759–18768.

Molecular Cell 744

Wood, R. (1999). DNA damage recognition during nucleotide excision repair in mammalian cells. Biochimie 81, 39–44. Zdzienicka, M., van der Schans, G., Westerveld, A., van Zeeland, A., and Simons, J. (1988). Phenotypic heterogeneity within the first complementation group of UV-sensitive mutants of Chinese hamster cell lines. Mutat. Res. 193, 31–41. Zdzienicka, M., Venema, J., Mitchell, D., van Hoffen, A., van Zeeland, A., Vrieling, H., Mullenders, L., Lohman, P., and Simons, J. (1992). (6–4) photoproducts and not cyclobutane pyrimidine dimers are the main UV-induced mutagenic lesions in Chinese hamster cells. Mutat. Res. 273, 73–83.