MOLECULAR CARCINOGENESIS4: 196-202 (1991)
Ras Gene Mutation and Amplification in Human Nonmelanoma Skin CancersWilliam E. Pierceall, Leonard H. Goldberg, Michael A. Tainsky, Wpas Mukhopadhyay, and Honnwara N. Ananthaswamy' Departments of Immunology (WEe HNA), Tumor Biology (MAT), and Thoracic Surgery (TM), The University of Texas M. D. Anderson Cancer Center, Houston, Texas, and Department of Dermatology (LHG), Baylor College of Medicine, Houston, Texas
Our previous studies have shown that human skin cancers occurring on sun-exposed body sites frequently contain activated Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells. In this study, we analyzed human primary squamous cell carcinomas (SCCs) and basal cell carcinomas (BCCs) occurring on sun-exposed body sites for mutations in codons 12, 13, and 61 of Ha-ras, Ki-ras, and N-ras oncogenes by amplification of genomic tumor DNAs by the polymerase chain reaction, followed by dot-blot hybridization to synthetic oligonucleotide probes designed t o detect single base-pair mutations. In addition t o the primary human skin cancers, we also analyzed Ha-ras-positive NIH 3T3 transformants for mutations in the Ha-ras oncogene. The results indicated that all three NIH 3T3 transformants, 11 of 24 (46%) SCCs, and 5 of 16 (31%) BCCs contained mutations a t the second position of Ha-rascodon 12 (GGC-GTC), predicting a glycine-tovaline amino acid substitution, whereas only 1 of 40 skin cancers (an SCC) displayed a mutation in the first position of Ki-ras codon 12 (GGT-AU), predicting a glycine-to-serine amino acid change. In addition, three of the SCCs contained highly amplified copies of the N-ras oncogene in their genomic DNA. Interestingly, two of the SCCs containing amplified N-ras sequences also had G-T mutations in codon 12 of the Ha-ras oncogene. These studies demonstrate that mutations in codon 12 of the Ha-ras oncogene occurred a t a high frequency in human skin cancers originating on sun-exposed body sites, whereas mutation in codon 12 of Ki-rasor amplification of N-ras occurred a t a low frequency. Since the mutations in the Ha-ras and Ki-ras oncogenes were located opposite potential pyrimidine dimer sites (C-C), it is likely that these mutations were induced by ultraviolet radiation present in sunlight. Key words: UV carcinogenesis, human skin cancer, ras oncogene, point mutation, gene amplification, PCR INTRODUCTION Specific mutations in the ras gene family have been implicated in the development of a variety of human cancers 11-41, The ras family of oncogenes is made up of three members, Ha-ras, Ki-ras, and N-ras, coding for protein products of 21,000 daltons and sharing approximately 70% sequence homology [21. Ras proteins are localized at the cell membrane and display intrinsic GTPase activities that function in growth control through yet-to-bedefined signal transduction pathways I1-41. RJS gene activation has been found in a wide variety of tumors and ranges from 10-20% Ha-fa5 activation in bladder carcinomas [5-71 to 85-1 00% Ki-ras activation in colorectal[8,9] and pancreatic tumors [ 10-1 21. Most transforming ras mutations have been found in codons 12, 13, or 61 in all three ras family members and appear to precede the onset of neoplasia I131. Cellular proto-oncogenescan be activated both by point mutations and by chromosomal translocations, suggesting that there may be a direct link between exposure to agents that damage DNA and genetic change leading to malignancy[1-3,14-19]. Zarbletal. 11 51 haveshownthat in rat tumors induced by nitrosomethylurea, the point mutation (a G-+Atransition) responsible for the activation of the c-Ha-ras-1 proto-oncogene was always located at codon 0 1991 WILEY-LISS, INC.
12, whereas in rat tumors induced by dimethyl[a]benzanthracene, the point mutation was at codon 61 of the c-Hafa5 locus. Similarly, Quintanilla et at. [I61 have reported that over 90% of tumors, including premalignant papillomas, induced by dimethyl[a]benzanthracene and 12-0tetradecanoylphorbol-13-acetatecontaineda specific A-T transversion at the second nucleotide of codon 61 of the c-Ha-rasoncogene. One hypothesis suggested by these observations is that a particular class of carcinogens may activate specific oncogenes, leading to the development of tumors. Since the DNA damage induced by ultraviolet (UV) radiation is unique and differs from the lesions induced by any other carcinogen, it is quite possiblethat UV radiation can cause specific alterations in oncogene structure that can lead to the induction of tumors. To study this, several investiga-
'Corresponding author: Department of immunology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd.. Box 178, Houston, TX 77030. Abbreviations: BCC, basal cell carcinoma; KA, keratoacanthoma; PCR, polymerasechain reaction; SCC, squamouscell carcinoma; SDS, scdlum dodecyl sulfate; SSC, standard saline citrate; SSPE, standard saline phosphate-ethylenediarninetetraaceticacid; UV, ultraviolet; XP, xeroderma pigmentosum.
197 RAS GENE MUTATIONS IN HUMAN SKIN CANCERS Cell Lines tors have analyzed human skin cancers for the presence of activated oncogenes. Ananthaswamy et al. [20,2 11found Three NIH 3T3 cell lines (designatedAS-2T, DG4T and WS4T) that were transformed with genomic DNA from that genomic DNAs from nonmelanoma human skin canthree human SCCs and known to contain the transfected cers occurring on sun-exposed body sites contained actihuman Ha-ras sequences (201were also analyzed for mutavated c-Ha-ras oncogenes capable of inducing morphologic tions in Ha-rascodon 12 and 61. and tumorigenic transformationwhen introduced into NIH 3T3 cells by DNA-mediated gene transfer. Suarez et al. [22] Preparation of Cellular DNA have reportedthat two of eight nonmelanoma skin tumors High molecular weight DNA was extracted by the from patients with xeroderma pigmentosum (XP) contained method of Wigler et al. [26], with minor modifications. A-T point mutations at codon 61 of the N-ras oncogene Tumor and normal tissues were minced thoroughly on ice and that this mutation occurred opposite a T-T sequence with dissection scissors and washed twice with ice-cold in the antisense strand. Recently, van der Schroeff et al. phosphate-bufferedsaline. The minced tissueswere resus(231 analyzed DNA from 30 basal cell carcinomas (BCCs) pended in lysis buffer (1YO sodium dodecyl sulfate (SDS), and 12 squamous cell carcinomas (SCCs) and found ras 150 mM NaCI, 10 mM Tris-HCI, pH 8.0, 10 mM ethyleneoncogene mutations in four BCCs and one SCC. Three diaminetetraacetic acid (EDTA)) containing 200 pg/mL proBCCs and one SCC contained mutations in codon 12 of teinase K. The viscous lysates were heated at 65°C for 15 Ki-ras (GGT+TGT or GAT, changing glycine to cysteine or min and then incubated overnight at 37°C. Equal volumes aspartic acid) and one BCC contained a mutation in codon of 650 mM NaCI, 10 mM Tris-HCI, pH 8.0, were added to 61 of the Ha-rasoncogene(CAG-CAT, changing glutamine the lysates and extracted twice with an equal volume of to histidine). Also, White and Balmain [24] have detected buffer-saturated phenol. The aqueous phase was extracted a G-;T point mutation at the second nucleotide of codon once with an equal volume of phenol/chloroform/isoamyl 12, opposite a potential pyrimidine dimer site in the alcohol (25:24: 1) and once more with an equal volume antisense strand of the Hams oncogene. Corominas et of chloroform/isoamylalcohol (24: 1). DNA was precipitated al. [25] reported that 30% of human keratoacanthomas from the aqueous phase with two volumes of cold abso(KAs), a benign skin cancer that undergoes periodic regreslute ethanol. The precipitated DNA was recovered with a sion, and 13% of SCCs contained specific mutations (A+T Pasteur pipet and washed successively with two changes transversions) in the second position of codon 61 of the of 70% ethanol and two changes of 100% ethanol. The c-Ha-ras oncogene. Thus, these results suggest that alDNA precipitatewas air dried and dissolved in 10 mM Tristhough activation of ras oncogenes may occur in human skin cancers, it is not a frequent event. HCI, pH 8.0, 1 mM EDTA and dialyzed twice against 10 Because we previously observed an activated Ha-rasbut mM Tris-HCI, pH 8.0, 1 mM EDTA. not Ki-ras or N-ras oncogene by NIH 3T3 transfection assays Southern Blot Hybridization in three of eight human skin cancers [20], we sought to Human tumor DNAs were cleaved with EcoRl or BamHl investigate the nature of the activating ras mutations in restriction enzymes (Bethesda Research Laboratories, the Ha-ras-positive NIH 3T3 transformants and in primary Gaithersburg, MD) under the conditions recommendedby human nonmelanoma skin cancers occurring on sunthe manufacturer.Digested DNAs were subjected to 0.8% exposed body sites using differentialoligodeoxynucleotide agarose gel electrophoresis, and DNA fragments were transhybridization in combination with the polymerase chain ferred to nitrocellulose filters by the Southern method [27]. reaction (PCR). Our results indicate that all three NIH 3T3 The resulting blots were baked at 80°C under vacuum and transformants as well as 16 of 40 (40%) human skin canhybridized with 32P-labeledHa-ras, Ki-ras, or N-ras probes cers studied contained mutations at the second position (Oncor Inc., Gaithersburg, MD) at 34°C according to the of Ha-ras codon 12 and only 1 of 40 (2.5%) human skin manufacturer's instructions. After hybridization for 24 h, cancers studied contained a mutation in the first position the filters were washed four times for 5 min each at room of Ki-ras codon 12. In addition, three of the SCCs contemperaturewith 2 x standard salinecitrate(SSC; 300mM tained highlyamplifiedcopies of the unmutated N-rasgene in their genomic DNA. sodium chloride, 30 mM sodium citrate), 0.5% SDS, and then three times for 20 min each at 56°C with 0.1 YO SSC, MATERIALS AND METHODS 0.1YO SDS. The blots were dried and exposed to Kodak Patients and Tumor Tissues XAR-5filmwithintensifyingscreensfor24-48 hat - 70°C. Skin tumors from sun-exposed body sites (faces, ears, Synthetic Oligonucleotide Primers and Probes and necks) of 40 Caucasian patients of both sexes, 40-89 Synthetic oligonucleotideprimers and probes were synyears old, were resectedsurgically, and either the DNA was thesized using an Applied Biosystems oligodeoxynucleotide extracted immediately orthetumorswerefrozen at - 70°C synthesizer (Foster City, CAI and the sequences described for later DNA extraction.The tumors were classified as SCC by Verlaan-de Vries et al. (281, except for Ki-ras oligomer or BCC by histologicalexamination. In some cases, matchprobes, which were obtained from New England Nuclear ing unexposed (lower back) skin from the same patient (Boston, MA). Sequences of synthetic 5' and 3' primers was used as a control. None of the patients had underused to amplify specific ras gene codons were as follows: gone chemotherapy or radiation treatment for skin canHa-ras codon 12113, GACGGAATATAAGCTGGTGG and cers prior to surgical removal of the tumors.
198
PIERCEALL ETAL.
TGGATGGTCAGCGCACTCTT; Ha-ras codon 61, AGACGTGCCTGllGGACATCand CGCATGTACTGGTCCCGCAT; Ki-ras codon 12/13, GACTGAATATTAAACTTGTGG and CTATTGTTGGATCATATTCC; Ki-ras codon 61, TTCCTACAGGAAGCAAGTAG and CACAAAGAAAGCCCTCCCCA; N-rascodon 12113, GACTGAGTACAAACTGGTGG and CTCTATGGTGGGATCATATT; N-ras codon 61, GGTGAAACCTGTITGTTGGA and ATACACAGAGGAAGCCTTCG . The synthetic oligodeoxynucleotideprobes used to detect point mutations in Ha-ras and Ki-ras genes were as follows: Ha-ras codon 12, GTGGGCGCC-GGTGTGGG (Gly), GTGGGCGCCAAGGTGTGGG (Ser), GTGGGCGCCLCGGTGTGGG (Cys), GTGGGCGCCCGCGGTGTGGG(Arg), GTGG GC GC CG& G GTGTGG G (Asp), GTGGG C G C C G S GGTGTGGG (Val), and GTGGGCGCCEGGTGTGGG (Ala); Ki-ras codon 12: GTGGAGCTGSGGCGTAG (Gly, wildtype). GTTGGAGCTWGGCGTAG (Arg), GTTGGAGCTAGTG G C GTAG (Ser), GTTGGAGCTTTTG G C GTAG (Cys), mGGAGCTGLTGGCGTAG (Ala), GTGGAGCTGLTGGCGTAG (Asp), GllGGAGCTGJGGCGTAG (Val). Polymerase Chain Reaction Amplification of genomic DNA was performed as described by Saiki et al. [29] for Ha-rasand Ki-ras codons 12 and 61 and for N-ras codons 12, 13, and 61. The reaction mixture containing 0.5 pg of genomic DNA and 0.6 pg of each primer was denatured by incubating at 95°C for 5 min, then 200 p M of each deoxynucleotidetriphosphate in 1 x GeneAmp (Perkin-Elmer Cetus, Norwalk, CT) reaction buffer containing 1.5 mM MgCI2,50 mM KCI, 10 mM Tris-HCI, pH 8.3,2.5 U Taq polymerase (Perkin-Elmer Cetus), and 1 U of Perfect Match polymeraseenhancer (Stratagene Inc., La Jolla, CA)were added to bring the final reaction volume to 100 pL. The addition of Perfect Match polymerase enhancer increases the specificity of primer extension reactions by destabilizing many mismatched primer-template complexes that would otherwise result in a heterogeneous population of amplified molecules 130). This reaction mixture was overlayed with 100 p L of mineral oil. PCR was carried out in a DNA thermal cycler (PerkinElmer Cetus) for 30 cycles. Each cycle consisted of denaturation at 94°C for 1 min, primer annealing at 42°C for 2 min, and extension at 72°C for 1 min. The last extension cycle was continued an additional 9 min to insure that all final product DNAs were complete. Dot-Blot Hybridization One-tenth of the reaction mixture was denatured at 37°C for 1 h in a final concentration of 0.4 N NaOH and then neutralized with the addition of sodium acetate, pH 5.2, to a final concentration of 0.3 M. SSC was next added, to a final concentration of 2 x ; this mixture was then applied to nylon filters (Genescreen; Du Pont Co., Boston, MA) using a Bio-Dot apparatus (Bio-Rad Labs, Richmond, CAI. Each well was washed once with 50 p12 x SSC. The filters were exposed to 254 nm UV radiation and then baked at 80°C under vacuum to insure maximum DNA cross-linking and binding. Specific synthetic DNA probes (50 ng each) were end labeled using [Y-~~PIATP (Amersham Corp.,
Arlington Heights, IL) and T4 polynucleotidekinase (Boehringer Mannheim Corp., Indianapolis, IN) as described by Maniatiset al. [31]. Nylon filters were prehybridizedin 5 x SSPE (0.75 M sodium chloride, 50 mM sodium phosphate monobasic, 5 mM EDTA, pH 8.0), 0.1% SDS, 5 x Denhardt's solution, and 100 pg/mL sonicated salmon sperm DNA for 2 h a t 50°C. The prehybridization mix was removed and fresh mix containing the labeled probe was added and allowed to hybridize for 2 h at 50°C. Filters were washed once at room temperature for 10 min, once a t 56°C for 30 min, and then exhaustively at the melting temperature T (), less 1°C as calculated by the Wallace rule for melting temperature of DNA duplexes (T, = 4"CIG-C bond + 2"C/A-T bond).All washes were in 5 x SSPE, pH 8.0. Filters were allowed to air dry briefly and then were exposed to Kodak XAR-5 film for 1-4 h at - 70°C. RESULTS PCR Amplification Specificity Upon completion of PCR amplification of Ha-ras, Ki-ras, and N-ras codons 12, 13, and 61 using the specific 5' and 3' primers, one tenth of the reaction mixture was analyzed routinely on a 1.8% agarose gel to confirm that the amplification step was successful and specific. In all cases, the PCRamplified DNA of the different ras gene codons gave rise to the expected band size and specificity (data not shown). Dot-Blot Hybridizations to Synthetic Oligodeoxynucleotide Probes PCR-amplified DNA from 40 human skin cancers was analyzed for the presence of point mutations in codons 12, 13, and 61 of N-ras, Ki-ras, and Ha-ras by dot-blot hybridizationto syntheticoligonucleotideprobes designed to detect single base-pair mutations. The results indicated that none of the human skin cancers analyzed contained mutations in codons 12, 13, or 61 of N-ras, or codon 61 of Ha-ras or Ki-ras, oncogenes. However, 11 of 24 SCCs and 5 of 16 BCCs had a G-+T base change a t the second position of Ha-ras codon 12 (Figure 1). This base mutation at the second position of Ha-ras codon 12 predicts an amino acid change from glycine to valine. The PCRamplified DNA from these 16 human skin cancers hybridized to both the Ha-raswild-type(glycine)codon 12-specific (GGC) and the valine-specific (GTC) oligonucleotide probes a t equal intensities (Figures 1A and 151, but not to the aspartic acid-specific (GAC) or the alanine-specific (GCC) probes. In addition, no PCR-amplified DNA from any of 40 human skin cancers hybridized to the Ha-/as codon 12 first position-specific oligonucleotide probes, indicatingthat there were no mutations at this position. As controls, we analyzed matching unexposed skin tissue from 12 skin cancer patients (1 1 SCC and 1 BCC) whose tumors contained G-T mutations in codon 12 of Ha-ras for mutations in Ha-ras codonl2. The results shown in Figure 2 indicate that the PCR-amplifiedmaterial from the 12 unexposed skin tissues hybridized to the GGC probe but not to the GTC probe. This suggests that G-;T mutations in codon 12 of the Ha-ras gene were not present in matching unexposed skin tissues.
199
RAS GENE MUTATIONS IN HUMAN SKIN CANCERS
A. (GGC) 1
2
3
4
5
B. (GTC) 6
1
2
3
4
5
6
Figure 1. Dot-blot h bridization of PCR-amplified human tumor DNAs with 32P-la!eled Ha-ras codon 12 wild-type and mutant oligodeoxynucleotide probes. PCR-amplifiedDNA from T24 bladder carcinoma cell line was used as a positive control for Ha-ras codon 12 mutation [42] and is spotted in well A l . Amplified DNAs from 24 SCCs are spotted in wells A2-El and from 16 B C G in wells E2-G5. (A) Dot-blot hybridization with the wild-type (GGC) probe, which indicates that the samples have at least one normal Ha-ras codon 12 allele. The DNAs in wells C2, C5, and G2 did not amplify for reasons unknown. The genomic DNA from these three samples was redialyzed against 10 mM Tris, pH 8.0. and 1 mM EDTA, amplified by PCR, and then reanalyzed for Ha-rascodon 12 mutations. The results indicated that all three samples hybridized to the wild-ty e probe but not to any of the mutant probes (data not shown). (BrDot-blot hybridization of an identically spotted filter using a GTC probe, which indicates that a G-T base substitution has occurred in the second position of codon 12. Eleven of 24 SCCs and 5 of 16 BCCs DNAs were positive for this mutation. None of the samples hybridized with other mutant probesspecificfor positions 1 or 2, indicating that no additional mutations occurred at this codon.
(a) GGC
Three (Figure 1, wells A3, B2, and C 1 ) of 16 SCCs that exhibited G-T mutations in codon 12 of Ha-rasoncogene were shown previously to contain Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells [ZOI. To determine whether these three NIH 3T3 transformants (designated AS-2T, DG-4T, and WS4T) contained the same G+T mutations, genomic DNAs from these cell lines were amplified by PCR and analyzed by dot-blot hybridization to Ha-ras codon 12 wildtype and mutant-specific synthetic oligodeoxynucleotide probes. All three NIH transformantshybridizedto the wildtype (GGC) and valine-specific (GTC) probes (Figure 2 a and b), but not to other Ha-rds first or second codonspecific probes (data not shown).The hybridizationof PCRamplified NIH 3T3 transformant DNAs to the GCC probe could be due to the presence of the normal human Ha-ras allele in addition to the mutant allele. When the genomic DNAs from the same 40 human skin cancers were amplified by PCR using Ki-ras codon 12specific primers and analyzed for mutations at this codon, only 1 of 40 skin tumors hybridized to both the wild-type (GGT) and the serine-specific (AGT) 32P-labeledsynthetic oligodeoxynucleotide probes but not to other first- or second-position probes. The data for the mutation-positive SCC and two representative mutation-negative SCCs are shown in Figure 3. Amplification of ras Oncogenes The genomic DNAs from all 40 human skin cancers were analyzed for possible amplification of ras genes by Southern blot hybridizationto 32P-labeledHa-ras, Ki-ras, or N-ras probes. None of 40 skin tumors contained amplified copies of the Ha-ras or Ki-ras oncogenes (data not shown). However, 3 of 40 human skin cancers contained highly amplified copies of the N-rasoncogene.The Southern blot of four SCCs shown in Figure 46 indicated that three of
(b) GTC
1 1
2
3
4
1
2
3
4
A
2
3
A m (GGT)
B C
D Figure 2. Dot-blot hybridization of PCR-amplified NIH 3T3 transformant DNAs and unexposed skin DNAs with 32P-labeled Ha-ras codon 12 wild-type and mutant oligodeoxynucleotide probes. Wells Al-A3 contain amplified DNA from the three NIH 3T3 transformants (AS-2T, DG4T, and M 4 T , respectively).Amplified DNA from 12 matching unexposed skin tissues of 12 skin cancer patients are spotted in wells 81-84, Cl-C4, and D1-D4, respectively. (a) Dot-blot hybridization with the wild-type (GGC) probe. (b) Dot-blot analysisof an identically spotted filter using a GFC probe.
Figure 3. Dot-blot hybridization of PCR-amplified human tumor DNAs with 32P-labeledwild-type and mutant oligonucleotide Ki-ras codon 12 probes. PCR-amplified DNAs from all 40 human skin cancers hybridized with the wild-type (GCT) Ki-ras codon 12 probe, but only 1 of 40 skin cancer DNAs hybridized with the mutant (AN) probe. Data for three SCCs are shown here, of which one is positive (well 1) and two are negative (wells 2 and 3). for this mutation. (A) Hybridization to the wild-type probe. (B) Hybridization to the mutant probe. The data indicate that a G-A base substitution occurred at the first position of Ki-ras codon 12 in the mutation-positiveSCC.
200
PIERCEALL ETAL.
A s 1 2 3 4
B 1 2 3 4
e 7 . 2 kb
Figure 4. Amplification of N-rassequences in human skin cancers. Genomic DNAs (10 pg each) from human skin cancers were digested with EcoRl and analyzed b Southern blot hybridization to "P-labeled N-ras probe, as iescribed in Materials and Methods. Data from four representativeSCC samples are shown here. Three of the SCC DNAs (lanes 2-4) show N-rasgene amplification (panel B). A picture of the ethidium bromide-stained gel before Southern blotting is shown in panel A. Lane S contains the molecular size standard of Hindll-cut A DNA. The molecular sizes of the six bands from top to bottom are 23.1, 9.4,6.6,4.4,2.3, and 2.0 kb, respectively.
them (lanes 2, 3, and 4) contained highly amplified copies of N-ras sequences. This amplification of the N-ras gene in three SCCs was not due to differences in the amount of DNA loaded onto each lane because the intensity of the ethidium bromide-stained gel was roughly the same in each lane (Figure 4A). Surprisingly, two of these SCCs (lanes 2 and 3) also contained G+T mutations at codon 12 of the Ha-ras oncogene (Figure 16, wells 65 and D4). DISCUSSION In our previous study, we had shown that genomic DNA from three of eight human skin cancers occurring on sunexposed body sites contained activated Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells [20,211. To identify the mutations responsible for the activation of ras oncogenes in human skin cancers, we analyzed the three Ha-ras-positive NIH 3T3 transformants for mutations in codons 12 and 61 of Ha-ras oncogene; in addition, we analyzed 40 primary human skin cancers (24 SCCs and 16 BCCs) occurring on sun-exposed body sites for the presence of activating fa5 mutations by PCR amplification of tumor DNAs, followed by dot-blot hybridization to synthetic oligodeoxynucleotide probes designed to detect mutations at codons 12, 13, and 61 of Ha-ras, Ki-ras, and N-ras oncogenes. We found that 16 of 40 skin tumors analyzed (1 1 of 24 SCCs and 5 of 16 BCCs) contained identical G-T base changes at the second position of the Ha-ras codon 12. In addition, the three NIH 3T3 transformants that were shown previously
to contain transfected Ha-rassequences[20] also contained G+T mutations identical to those found in corresponding primary SCCs. This G+T mutation predicts a glycine-tovaline amino acid substitution. In contrast, only 1 of 40 human skin cancers had a G-A mutation in the first position of Ki-ras oncogene, predicting a glycine-to-serine amino acid substitution.The remaining 23 skin tumors did not contain any mutations at codons 12, 13, or 61 of Ha-rds, Ki-ras, or N-ras oncogenes. Similarly, none of 12 matching unexposed skin tissues from skin cancer patients analyzed contained G+T or other Ha-rds codon 12 mutations. Even though activation of ras oncogenes in human skin cancers has been reported previously, two aspects of our finding are striking: the specificity and the frequency of the mutated Ha-ras oncogene. Of all possible ras gene mutations for which we tested, only the second position of codon 12 of Ha-ras was mutated at a high frequency, and this mutation was exclusively a G-T base change. In contrast, only one of 40 human skin cancers tested contained a G+A mutation at the first position of Ki-rascodon 12, resulting in an amino acid substitution of glycine to serine. A second feature of this study is the frequency with which the Ha-ras mutation occurred in human skin cancers. We found that 40% of all nonmelanoma skin cancers examined (46% of SCCs and 3 t % of BCCs) contained this G-T base change at the second position of the Ha-ras codon 12. In contrast, previous reports have indicatedthat ras activation occurs at a relatively low frequency in human skin cancers. Corominas et al. [251 reported that although 30% of human KAs, a benign tumor that undergoes periodic regression, contained specific mutations (A+T transversions) in the second position of codon 61 of the Ha-ras oncogene, only 13% of human SCCs contained mutations in the Ha-ras oncogene. This observation has led to the hypothesis that activated Ha-ras oncogenes may play a role in initiation but not in tumor progression [25]. Suarez et al. 1221have reportedthat two of eight nonmelanoma skin tumors from XP patients contained A-T point mutations at codon 61 of the N-ras oncogene. More recently, van der Schroeff et al. (231found that of 12 SCC samples examined, only one contained a mutation (GGT+TGT) in codon 12 of the Ki-rasoncogene, and of 30 BCCs examined, three had Ki-ras mutations at codon 12 (GGT-TGT or GAT) and one had a mutation at Ha-ras codon 61 (CAG-CAT). The frequency of ras mutations that we observed in human skin cancers, although higher than that reported by other investigators, is not unexpected because ras gene mutations are known to occur at 80-90% frequencies in other types of human cancers such as colorectal[8,91 and pancreatic tumors [lo-1 21. Epidemiologic studies have indicated that UV radiation present in sunlight is responsible for the induction of most human skin cancers [32,33]. Our studies also support an active role for UV radiation in the induction of mutation in human skin cancers. The mutationsfound in our human skin cancers were primarily G+T base changes in the sense strand of Ha-ras codon 12 and, in one instance, a G+A base change in the sense strand of Ki-ras codon 12. The
20 1
RAS GENE MUTATIONS IN HUMAN SKIN CANCERS
specific base change in codon 12 of both Ha-rasand Ki-ras occurred directly opposite cytosine doublets in antisense DNA strands. Therefore, it is distinctly possible that the mutation in codon 12 of Ha-rasand Ki-ras may have arisen from induction of pyrimidine dimers in the antisense strand by UV radiation. By examining the following sequence of Ha-rasand Ki-ras codon 12, one can predict that a pyrimidine dimer (C-C) could be formed as a result of UV exposure. Ha-ras: Ki-ras:
sense strand antisense strand sense strand antisense strand
5'-GGC-3'
3 '-CCG - 5' 5'-GGT-3' 3'-CCA-5'
Although UV radiation is known to induce both pyrimidine dimers and (6-4) photoproducts 134-361, it appears from our studies and from studies by others that pyrimidine dimers, and not (6-4) photoproducts, are the principal premutagenic lesions responsible for the activation of ras oncogenes in human skin cancers occurring on sunexposed body sites. First, we observed that all the mutations (G-T or G-A) found in our human skin cancers were located opposite C-C doublets, potentialtargets for pyrimidine dimer formation. White and Balmain [24] have also found this identical G-T mutation in a human BCC. Second, Suarez et al. [22] have shown that the mutations in codon 61 of the N-ras oncogene in nonmelanoma skin cancers from XP patients were located directly opposite a T-T doublet in the antisense strand. Similarly, van der Schroeff et al. [23] have shown that the mutations in codon 12 of Ki-ras in human BCCs and SCCs occurred directly opposite a C-C doublet. Taken together these results suggest that UV-induced pyrimidine dimers are probably the premutagenic lesions involved in induction of mutation and subsequent transformation of cells. In addition to the mutations in the Ha-ras and Ki-ras oncogenes, amplification of N-ras was detected in some human skin cancers. Three (all SCCs) of 40 human skin cancers analyzed contained highly amplified copies of the N-rasoncogene. Interestingly, two of these SCCs also possessed a G-T mutation in the second position of Ha-ras codon 12. However, the third SCC contained only amplified copies of the N-rasgene and no other known rasgene mutations. Suarez et al. [371 reported that skin tumors from XP patients contained amplified Ha-ras, N-ras, and c-rnyc mRNA transcripts, in addition to a mutated N-ras and a rearranged Ha-ras oncogene. The presence of several oncogene alterations in the same tumor may have resulted from the accumulation of unrepaired UV-induced DNA lesions. These results suggest that gene amplification, in addition to mutation in rds genes, may also play a role in carcinogenesis. Mutational activation of ras oncogenes has also been detected in human melanomas. However, most of the mutations in human melanomas have been shown to occur at codons 12, 13, and 61 of the N-ras oncogene [38-401 and occasionally at codon 61 of the Ha-rasoncogene[41]. Because there is a growing concern that the incidence of
both melanoma and nonmelanoma human skin cancers may increase because of the recently reported decrease in stratospheric ozone, it is important to understand the molecular mechanisms by which UV radiation present in sunlight induces human skin cancers. ACKNOWLEDGMENTS
This research was supported by Grant R01-CA-46523 from the National Institutes of Health (awarded to HNA). William Pierceall is the recipient of a Predoctoral Fellowship from the National Institutes of Health under Training Grant T32-CA-09598. Received November 9, 1990; revised February 19, 1991, accepted February 20,1991.
REFERENCES 1. Bishop JM. Cellular oncogenes and their retroviruses. Annu Rev Biochem 52:301-354, 1983. 2. Barbacid M. Rasgenes. Annu Rev Biochem 56:779-827, 1987. 3. Bos 11. The ras gene family and human carcinogenesis. Mutat Res 195:255-271, 1988. 4. Bos JL. Ras oncogenes in human cancer: A review. Cancer Res 4914682-4689,1989. 5. Fujita I,Yoshida 0, Yuasa Y, Rhim IS,Hatanaka M, Aaronson SA. Ha-ras oncogenes are activated by somatic alterations in human urinary tract tumors. Nature 309:464-466, 1984. 6. Malone PR, Visvanathan KV, Ponder BA, Shearer RJ, Summerhayes IC. Oncogenesand bladder cancer. Br J Urol57:664-667, 1985. 7. Visvanathan KV, Pocock RD, Summerhayes IC. Preferential and novel activation of H-ras in human bladder carcinomas.Oncogene Res3:77-86, 1988. 8. Bos JL, Fearon ER, Hamilton SR, et al. Prevalenceof ras gene mutations in human colorectal cancers. Nature 327:293-297, 1987. 9. Forrester KC, Almquera C, Han K, Grizzle WE, Perucho M. Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature 327:298-303, 1987. 10. Almoguera C, Shibata D, Forrester K, Martin J, Arnheim N, krucho M. Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53:549-554, 1988. 11. Smit VTHMB, Boot AIM, Smit AMM, Fleuren GJ, Cornelisse CJ, Bos 11. K-ras codon mutations occur very frequently in pancreaticadenocarcinomas. Nucleic Acids Res 16:7773-7782, 1988. 12. Grunewald K, Lyons J, Frohlich A, et al. High frequency of Ki-ras codon 12 mutations in pancreatic adenocarcinomas. Int J Cancer43:1037-1041, 1989. 13. Kumar R, Sukumar 5, Barbacid M. Activation of ras oncogenes precedingthe onset of neoplasia.Science 248: 1 101-1 104, 1990. 14. Land H, Parada LF, Weinberg RA. Cellular oncogenesand multistep carcinogenesis. Science 222:771-778, 1983. 15. Zarbl H, Sukumar 5,Arthur AV, Martin-Zanca D, Barbacid M. Direct mutagenesis of Ha-ras-1 oncogenes by Knitroso-N-methylurea during initiation of mammary carcinogenesis in rats Nature 315 382-385.1985 16. Quintanilla M, Brown K, Ramsden M, Balmain A. Carcinogenspecific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature 322:78-80, 1986. 17. Brown K, Buchmann A, Balmain A. Carcinogen-induced mutations in the mouse c-Hams gene provide evidence of multiple pathways for tumor progression. Proc Natl Acad Sci USA 87:538542, 1990. 18. Balmain A, Brown K. Oncogene activation in chemical carcinogenesis.AdvCancerRes51:147-182, 1988. 19. Suarez HG. Activated oncogenes in human tumors. Anticancer Res 9:1331-1334,1989, 20. Ananthaswamy HN. Price JE, Goldberg LH, Bales ES. Detection and identification of activated oncogenes in human skin cances occurringonsun-exposedbodysites. CancerRes48:3341-3346,1988. 21. Ananthaswamy HN, Price JE, Goldberg LH, Bales ES. Simultaneous transfer of tumorigenic and metastatic phenotypes by transfection with genomic DNA from a human cutaneous squamous cell carcinoma. J Cell Biochem 36:137-146, 1988. 22. Suarez HG, Daya-Grosjean L, Schlaifer D, et al. Activated oncogenes in human skin tumors from a repair-deficient syndrome, xeroderma pigmentosum. Cancer Res 49: 1223-1 228, 1989.
202
PIERCEALL ETAL.
23. van der Schroeff JG, Evers LM, Boot AJM, Bos JL. Ras oncogene mutations in basal cell carcinomas and squamous cell carcinomasof human skin. J Invest Dermatol94:423-425, 1990. 24. White SI, Balmain A. G to T mutation in codon 12 of the human Harvey ras oncogene derived from a basal cell carcinoma. J Invest Dermatol91:407, 1988. 25. Corominas M, Kamino H, Leon J, Pellicer A. Oncogene activation in human benign tumors of the skin (keratoacanthomas):Is HRAS involved in differentiation as well as proliferation?Proc Natl Acad Sci USA86:6372-6376, 1989. 26. Wigler M, Pellicer A, Silverstein 5, Axel R. Biochemical transfer of single-copy eucaryotic genes using total cellular DNA as donor. Cell 14:725-731, 1978. 27. Southern EM. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503-517, 1975. 28. Verlaan-de Vries M, Bogaard ME, van den Elst H, van Boom JH, van der Eb AJ,60s JL. A dot-blot screening procedure for mutated ras oncogenes using synthetic oligodeoxynucleotides. Gene 50:313-320,1986, 29. Saiki RK, Gelfaud DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491,1988. 30. Nelson K, Mathur El. Perfect MatchTMenhancer: Limits false priming events during amplification reactions. Strategies 3: 17-19, 1990. 31. Maniatis T, kitsch EF, Sambrook 1. Molecular Cloning, A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1982. 32. Urbach F. Evidence and epidemiology of ultraviolet-inducedcancersin man. NCI Monogr 50:s-10.1978.
33. Scotto J, Fears TR. Skin cancer epidemiology: Research needs. NCI Monogr 50:169-177,1978. 34. Setlow JK. The molecularbasis of biolwical effectsof ultravioletradiationand photoreactivation. CurrTop Radiat Res2: 195-248,1966. 35. Lippke JA. Gordon K, Brash DE, HaseltineWA. Distribution of UV light-induced damage in a defined sequence of human DNA: Detection of alkaline-sensitive lesions a t pyrimidine nucleosidecytidinesequences. Proc Natl Acad Sci USA78:3388-3392,1981. 36. Rosenstein BS, Mitchell DL. Action spectra for the induction of pyimidine (64) Wrimidone photoproductsand cyclobutanedimers in normal human skin fibroblasts. Photochem Photobiol45:775781, 1987. 37. Suarez HG, Nardeux PC, Andeol Y, Sarasin A. Multiple activated oncogenes in human tumors. Oncogene Res 1 :201-207, 1987. 38. van't Veer U, 8urgering BMT, hrsteeg R, et al. N-ras mutations in human cutaneous melanoma from sun-exposedbody sites. Mol Cell Biol9:3114-3116, 1989. 39. Padua RA, Barrass NC, Currie GC. Activation of N-ras in a human melanoma cell line. Mol Cell Biol5:582-585, 1985. 40. Keijzer W, Mulder Me Langeveld JCM, et al. Establishment and characterization of a melanoma cell line from a xeroderma pigmentosum patient: Activation of N-ras at a potential pyrimidine dimer site. Cancer Res49: 1229-1 235, 1989. 41. Shukla VK, Hughes DC, Hughes LE, McCormick F, Padua RA. Ras mutations in human melanotic lesions: K-ra5 activation is a frequent and early event in melanoma development. Oncogene Res 5: 121-127, 1989. 42. Reddy EFI Reynolds RK, Santos E, Barbacid M. A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300: 149152,1982.
Ras Gene Mutation and Amplification in Human Nonmelanoma Skin CancersWilliam E. Pierceall, Leonard H. Goldberg, Michael A. Tainsky, Wpas Mukhopadhyay, and Honnwara N. Ananthaswamy' Departments of Immunology (WEe HNA), Tumor Biology (MAT), and Thoracic Surgery (TM), The University of Texas M. D. Anderson Cancer Center, Houston, Texas, and Department of Dermatology (LHG), Baylor College of Medicine, Houston, Texas
Our previous studies have shown that human skin cancers occurring on sun-exposed body sites frequently contain activated Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells. In this study, we analyzed human primary squamous cell carcinomas (SCCs) and basal cell carcinomas (BCCs) occurring on sun-exposed body sites for mutations in codons 12, 13, and 61 of Ha-ras, Ki-ras, and N-ras oncogenes by amplification of genomic tumor DNAs by the polymerase chain reaction, followed by dot-blot hybridization to synthetic oligonucleotide probes designed t o detect single base-pair mutations. In addition t o the primary human skin cancers, we also analyzed Ha-ras-positive NIH 3T3 transformants for mutations in the Ha-ras oncogene. The results indicated that all three NIH 3T3 transformants, 11 of 24 (46%) SCCs, and 5 of 16 (31%) BCCs contained mutations a t the second position of Ha-rascodon 12 (GGC-GTC), predicting a glycine-tovaline amino acid substitution, whereas only 1 of 40 skin cancers (an SCC) displayed a mutation in the first position of Ki-ras codon 12 (GGT-AU), predicting a glycine-to-serine amino acid change. In addition, three of the SCCs contained highly amplified copies of the N-ras oncogene in their genomic DNA. Interestingly, two of the SCCs containing amplified N-ras sequences also had G-T mutations in codon 12 of the Ha-ras oncogene. These studies demonstrate that mutations in codon 12 of the Ha-ras oncogene occurred a t a high frequency in human skin cancers originating on sun-exposed body sites, whereas mutation in codon 12 of Ki-rasor amplification of N-ras occurred a t a low frequency. Since the mutations in the Ha-ras and Ki-ras oncogenes were located opposite potential pyrimidine dimer sites (C-C), it is likely that these mutations were induced by ultraviolet radiation present in sunlight. Key words: UV carcinogenesis, human skin cancer, ras oncogene, point mutation, gene amplification, PCR INTRODUCTION Specific mutations in the ras gene family have been implicated in the development of a variety of human cancers 11-41, The ras family of oncogenes is made up of three members, Ha-ras, Ki-ras, and N-ras, coding for protein products of 21,000 daltons and sharing approximately 70% sequence homology [21. Ras proteins are localized at the cell membrane and display intrinsic GTPase activities that function in growth control through yet-to-bedefined signal transduction pathways I1-41. RJS gene activation has been found in a wide variety of tumors and ranges from 10-20% Ha-fa5 activation in bladder carcinomas [5-71 to 85-1 00% Ki-ras activation in colorectal[8,9] and pancreatic tumors [ 10-1 21. Most transforming ras mutations have been found in codons 12, 13, or 61 in all three ras family members and appear to precede the onset of neoplasia I131. Cellular proto-oncogenescan be activated both by point mutations and by chromosomal translocations, suggesting that there may be a direct link between exposure to agents that damage DNA and genetic change leading to malignancy[1-3,14-19]. Zarbletal. 11 51 haveshownthat in rat tumors induced by nitrosomethylurea, the point mutation (a G-+Atransition) responsible for the activation of the c-Ha-ras-1 proto-oncogene was always located at codon 0 1991 WILEY-LISS, INC.
12, whereas in rat tumors induced by dimethyl[a]benzanthracene, the point mutation was at codon 61 of the c-Hafa5 locus. Similarly, Quintanilla et at. [I61 have reported that over 90% of tumors, including premalignant papillomas, induced by dimethyl[a]benzanthracene and 12-0tetradecanoylphorbol-13-acetatecontaineda specific A-T transversion at the second nucleotide of codon 61 of the c-Ha-rasoncogene. One hypothesis suggested by these observations is that a particular class of carcinogens may activate specific oncogenes, leading to the development of tumors. Since the DNA damage induced by ultraviolet (UV) radiation is unique and differs from the lesions induced by any other carcinogen, it is quite possiblethat UV radiation can cause specific alterations in oncogene structure that can lead to the induction of tumors. To study this, several investiga-
'Corresponding author: Department of immunology, The University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Blvd.. Box 178, Houston, TX 77030. Abbreviations: BCC, basal cell carcinoma; KA, keratoacanthoma; PCR, polymerasechain reaction; SCC, squamouscell carcinoma; SDS, scdlum dodecyl sulfate; SSC, standard saline citrate; SSPE, standard saline phosphate-ethylenediarninetetraaceticacid; UV, ultraviolet; XP, xeroderma pigmentosum.
197 RAS GENE MUTATIONS IN HUMAN SKIN CANCERS Cell Lines tors have analyzed human skin cancers for the presence of activated oncogenes. Ananthaswamy et al. [20,2 11found Three NIH 3T3 cell lines (designatedAS-2T, DG4T and WS4T) that were transformed with genomic DNA from that genomic DNAs from nonmelanoma human skin canthree human SCCs and known to contain the transfected cers occurring on sun-exposed body sites contained actihuman Ha-ras sequences (201were also analyzed for mutavated c-Ha-ras oncogenes capable of inducing morphologic tions in Ha-rascodon 12 and 61. and tumorigenic transformationwhen introduced into NIH 3T3 cells by DNA-mediated gene transfer. Suarez et al. [22] Preparation of Cellular DNA have reportedthat two of eight nonmelanoma skin tumors High molecular weight DNA was extracted by the from patients with xeroderma pigmentosum (XP) contained method of Wigler et al. [26], with minor modifications. A-T point mutations at codon 61 of the N-ras oncogene Tumor and normal tissues were minced thoroughly on ice and that this mutation occurred opposite a T-T sequence with dissection scissors and washed twice with ice-cold in the antisense strand. Recently, van der Schroeff et al. phosphate-bufferedsaline. The minced tissueswere resus(231 analyzed DNA from 30 basal cell carcinomas (BCCs) pended in lysis buffer (1YO sodium dodecyl sulfate (SDS), and 12 squamous cell carcinomas (SCCs) and found ras 150 mM NaCI, 10 mM Tris-HCI, pH 8.0, 10 mM ethyleneoncogene mutations in four BCCs and one SCC. Three diaminetetraacetic acid (EDTA)) containing 200 pg/mL proBCCs and one SCC contained mutations in codon 12 of teinase K. The viscous lysates were heated at 65°C for 15 Ki-ras (GGT+TGT or GAT, changing glycine to cysteine or min and then incubated overnight at 37°C. Equal volumes aspartic acid) and one BCC contained a mutation in codon of 650 mM NaCI, 10 mM Tris-HCI, pH 8.0, were added to 61 of the Ha-rasoncogene(CAG-CAT, changing glutamine the lysates and extracted twice with an equal volume of to histidine). Also, White and Balmain [24] have detected buffer-saturated phenol. The aqueous phase was extracted a G-;T point mutation at the second nucleotide of codon once with an equal volume of phenol/chloroform/isoamyl 12, opposite a potential pyrimidine dimer site in the alcohol (25:24: 1) and once more with an equal volume antisense strand of the Hams oncogene. Corominas et of chloroform/isoamylalcohol (24: 1). DNA was precipitated al. [25] reported that 30% of human keratoacanthomas from the aqueous phase with two volumes of cold abso(KAs), a benign skin cancer that undergoes periodic regreslute ethanol. The precipitated DNA was recovered with a sion, and 13% of SCCs contained specific mutations (A+T Pasteur pipet and washed successively with two changes transversions) in the second position of codon 61 of the of 70% ethanol and two changes of 100% ethanol. The c-Ha-ras oncogene. Thus, these results suggest that alDNA precipitatewas air dried and dissolved in 10 mM Tristhough activation of ras oncogenes may occur in human skin cancers, it is not a frequent event. HCI, pH 8.0, 1 mM EDTA and dialyzed twice against 10 Because we previously observed an activated Ha-rasbut mM Tris-HCI, pH 8.0, 1 mM EDTA. not Ki-ras or N-ras oncogene by NIH 3T3 transfection assays Southern Blot Hybridization in three of eight human skin cancers [20], we sought to Human tumor DNAs were cleaved with EcoRl or BamHl investigate the nature of the activating ras mutations in restriction enzymes (Bethesda Research Laboratories, the Ha-ras-positive NIH 3T3 transformants and in primary Gaithersburg, MD) under the conditions recommendedby human nonmelanoma skin cancers occurring on sunthe manufacturer.Digested DNAs were subjected to 0.8% exposed body sites using differentialoligodeoxynucleotide agarose gel electrophoresis, and DNA fragments were transhybridization in combination with the polymerase chain ferred to nitrocellulose filters by the Southern method [27]. reaction (PCR). Our results indicate that all three NIH 3T3 The resulting blots were baked at 80°C under vacuum and transformants as well as 16 of 40 (40%) human skin canhybridized with 32P-labeledHa-ras, Ki-ras, or N-ras probes cers studied contained mutations at the second position (Oncor Inc., Gaithersburg, MD) at 34°C according to the of Ha-ras codon 12 and only 1 of 40 (2.5%) human skin manufacturer's instructions. After hybridization for 24 h, cancers studied contained a mutation in the first position the filters were washed four times for 5 min each at room of Ki-ras codon 12. In addition, three of the SCCs contemperaturewith 2 x standard salinecitrate(SSC; 300mM tained highlyamplifiedcopies of the unmutated N-rasgene in their genomic DNA. sodium chloride, 30 mM sodium citrate), 0.5% SDS, and then three times for 20 min each at 56°C with 0.1 YO SSC, MATERIALS AND METHODS 0.1YO SDS. The blots were dried and exposed to Kodak Patients and Tumor Tissues XAR-5filmwithintensifyingscreensfor24-48 hat - 70°C. Skin tumors from sun-exposed body sites (faces, ears, Synthetic Oligonucleotide Primers and Probes and necks) of 40 Caucasian patients of both sexes, 40-89 Synthetic oligonucleotideprimers and probes were synyears old, were resectedsurgically, and either the DNA was thesized using an Applied Biosystems oligodeoxynucleotide extracted immediately orthetumorswerefrozen at - 70°C synthesizer (Foster City, CAI and the sequences described for later DNA extraction.The tumors were classified as SCC by Verlaan-de Vries et al. (281, except for Ki-ras oligomer or BCC by histologicalexamination. In some cases, matchprobes, which were obtained from New England Nuclear ing unexposed (lower back) skin from the same patient (Boston, MA). Sequences of synthetic 5' and 3' primers was used as a control. None of the patients had underused to amplify specific ras gene codons were as follows: gone chemotherapy or radiation treatment for skin canHa-ras codon 12113, GACGGAATATAAGCTGGTGG and cers prior to surgical removal of the tumors.
198
PIERCEALL ETAL.
TGGATGGTCAGCGCACTCTT; Ha-ras codon 61, AGACGTGCCTGllGGACATCand CGCATGTACTGGTCCCGCAT; Ki-ras codon 12/13, GACTGAATATTAAACTTGTGG and CTATTGTTGGATCATATTCC; Ki-ras codon 61, TTCCTACAGGAAGCAAGTAG and CACAAAGAAAGCCCTCCCCA; N-rascodon 12113, GACTGAGTACAAACTGGTGG and CTCTATGGTGGGATCATATT; N-ras codon 61, GGTGAAACCTGTITGTTGGA and ATACACAGAGGAAGCCTTCG . The synthetic oligodeoxynucleotideprobes used to detect point mutations in Ha-ras and Ki-ras genes were as follows: Ha-ras codon 12, GTGGGCGCC-GGTGTGGG (Gly), GTGGGCGCCAAGGTGTGGG (Ser), GTGGGCGCCLCGGTGTGGG (Cys), GTGGGCGCCCGCGGTGTGGG(Arg), GTGG GC GC CG& G GTGTGG G (Asp), GTGGG C G C C G S GGTGTGGG (Val), and GTGGGCGCCEGGTGTGGG (Ala); Ki-ras codon 12: GTGGAGCTGSGGCGTAG (Gly, wildtype). GTTGGAGCTWGGCGTAG (Arg), GTTGGAGCTAGTG G C GTAG (Ser), GTTGGAGCTTTTG G C GTAG (Cys), mGGAGCTGLTGGCGTAG (Ala), GTGGAGCTGLTGGCGTAG (Asp), GllGGAGCTGJGGCGTAG (Val). Polymerase Chain Reaction Amplification of genomic DNA was performed as described by Saiki et al. [29] for Ha-rasand Ki-ras codons 12 and 61 and for N-ras codons 12, 13, and 61. The reaction mixture containing 0.5 pg of genomic DNA and 0.6 pg of each primer was denatured by incubating at 95°C for 5 min, then 200 p M of each deoxynucleotidetriphosphate in 1 x GeneAmp (Perkin-Elmer Cetus, Norwalk, CT) reaction buffer containing 1.5 mM MgCI2,50 mM KCI, 10 mM Tris-HCI, pH 8.3,2.5 U Taq polymerase (Perkin-Elmer Cetus), and 1 U of Perfect Match polymeraseenhancer (Stratagene Inc., La Jolla, CA)were added to bring the final reaction volume to 100 pL. The addition of Perfect Match polymerase enhancer increases the specificity of primer extension reactions by destabilizing many mismatched primer-template complexes that would otherwise result in a heterogeneous population of amplified molecules 130). This reaction mixture was overlayed with 100 p L of mineral oil. PCR was carried out in a DNA thermal cycler (PerkinElmer Cetus) for 30 cycles. Each cycle consisted of denaturation at 94°C for 1 min, primer annealing at 42°C for 2 min, and extension at 72°C for 1 min. The last extension cycle was continued an additional 9 min to insure that all final product DNAs were complete. Dot-Blot Hybridization One-tenth of the reaction mixture was denatured at 37°C for 1 h in a final concentration of 0.4 N NaOH and then neutralized with the addition of sodium acetate, pH 5.2, to a final concentration of 0.3 M. SSC was next added, to a final concentration of 2 x ; this mixture was then applied to nylon filters (Genescreen; Du Pont Co., Boston, MA) using a Bio-Dot apparatus (Bio-Rad Labs, Richmond, CAI. Each well was washed once with 50 p12 x SSC. The filters were exposed to 254 nm UV radiation and then baked at 80°C under vacuum to insure maximum DNA cross-linking and binding. Specific synthetic DNA probes (50 ng each) were end labeled using [Y-~~PIATP (Amersham Corp.,
Arlington Heights, IL) and T4 polynucleotidekinase (Boehringer Mannheim Corp., Indianapolis, IN) as described by Maniatiset al. [31]. Nylon filters were prehybridizedin 5 x SSPE (0.75 M sodium chloride, 50 mM sodium phosphate monobasic, 5 mM EDTA, pH 8.0), 0.1% SDS, 5 x Denhardt's solution, and 100 pg/mL sonicated salmon sperm DNA for 2 h a t 50°C. The prehybridization mix was removed and fresh mix containing the labeled probe was added and allowed to hybridize for 2 h at 50°C. Filters were washed once at room temperature for 10 min, once a t 56°C for 30 min, and then exhaustively at the melting temperature T (), less 1°C as calculated by the Wallace rule for melting temperature of DNA duplexes (T, = 4"CIG-C bond + 2"C/A-T bond).All washes were in 5 x SSPE, pH 8.0. Filters were allowed to air dry briefly and then were exposed to Kodak XAR-5 film for 1-4 h at - 70°C. RESULTS PCR Amplification Specificity Upon completion of PCR amplification of Ha-ras, Ki-ras, and N-ras codons 12, 13, and 61 using the specific 5' and 3' primers, one tenth of the reaction mixture was analyzed routinely on a 1.8% agarose gel to confirm that the amplification step was successful and specific. In all cases, the PCRamplified DNA of the different ras gene codons gave rise to the expected band size and specificity (data not shown). Dot-Blot Hybridizations to Synthetic Oligodeoxynucleotide Probes PCR-amplified DNA from 40 human skin cancers was analyzed for the presence of point mutations in codons 12, 13, and 61 of N-ras, Ki-ras, and Ha-ras by dot-blot hybridizationto syntheticoligonucleotideprobes designed to detect single base-pair mutations. The results indicated that none of the human skin cancers analyzed contained mutations in codons 12, 13, or 61 of N-ras, or codon 61 of Ha-ras or Ki-ras, oncogenes. However, 11 of 24 SCCs and 5 of 16 BCCs had a G-+T base change a t the second position of Ha-ras codon 12 (Figure 1). This base mutation at the second position of Ha-ras codon 12 predicts an amino acid change from glycine to valine. The PCRamplified DNA from these 16 human skin cancers hybridized to both the Ha-raswild-type(glycine)codon 12-specific (GGC) and the valine-specific (GTC) oligonucleotide probes a t equal intensities (Figures 1A and 151, but not to the aspartic acid-specific (GAC) or the alanine-specific (GCC) probes. In addition, no PCR-amplified DNA from any of 40 human skin cancers hybridized to the Ha-/as codon 12 first position-specific oligonucleotide probes, indicatingthat there were no mutations at this position. As controls, we analyzed matching unexposed skin tissue from 12 skin cancer patients (1 1 SCC and 1 BCC) whose tumors contained G-T mutations in codon 12 of Ha-ras for mutations in Ha-ras codonl2. The results shown in Figure 2 indicate that the PCR-amplifiedmaterial from the 12 unexposed skin tissues hybridized to the GGC probe but not to the GTC probe. This suggests that G-;T mutations in codon 12 of the Ha-ras gene were not present in matching unexposed skin tissues.
199
RAS GENE MUTATIONS IN HUMAN SKIN CANCERS
A. (GGC) 1
2
3
4
5
B. (GTC) 6
1
2
3
4
5
6
Figure 1. Dot-blot h bridization of PCR-amplified human tumor DNAs with 32P-la!eled Ha-ras codon 12 wild-type and mutant oligodeoxynucleotide probes. PCR-amplifiedDNA from T24 bladder carcinoma cell line was used as a positive control for Ha-ras codon 12 mutation [42] and is spotted in well A l . Amplified DNAs from 24 SCCs are spotted in wells A2-El and from 16 B C G in wells E2-G5. (A) Dot-blot hybridization with the wild-type (GGC) probe, which indicates that the samples have at least one normal Ha-ras codon 12 allele. The DNAs in wells C2, C5, and G2 did not amplify for reasons unknown. The genomic DNA from these three samples was redialyzed against 10 mM Tris, pH 8.0. and 1 mM EDTA, amplified by PCR, and then reanalyzed for Ha-rascodon 12 mutations. The results indicated that all three samples hybridized to the wild-ty e probe but not to any of the mutant probes (data not shown). (BrDot-blot hybridization of an identically spotted filter using a GTC probe, which indicates that a G-T base substitution has occurred in the second position of codon 12. Eleven of 24 SCCs and 5 of 16 BCCs DNAs were positive for this mutation. None of the samples hybridized with other mutant probesspecificfor positions 1 or 2, indicating that no additional mutations occurred at this codon.
(a) GGC
Three (Figure 1, wells A3, B2, and C 1 ) of 16 SCCs that exhibited G-T mutations in codon 12 of Ha-rasoncogene were shown previously to contain Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells [ZOI. To determine whether these three NIH 3T3 transformants (designated AS-2T, DG-4T, and WS4T) contained the same G+T mutations, genomic DNAs from these cell lines were amplified by PCR and analyzed by dot-blot hybridization to Ha-ras codon 12 wildtype and mutant-specific synthetic oligodeoxynucleotide probes. All three NIH transformantshybridizedto the wildtype (GGC) and valine-specific (GTC) probes (Figure 2 a and b), but not to other Ha-rds first or second codonspecific probes (data not shown).The hybridizationof PCRamplified NIH 3T3 transformant DNAs to the GCC probe could be due to the presence of the normal human Ha-ras allele in addition to the mutant allele. When the genomic DNAs from the same 40 human skin cancers were amplified by PCR using Ki-ras codon 12specific primers and analyzed for mutations at this codon, only 1 of 40 skin tumors hybridized to both the wild-type (GGT) and the serine-specific (AGT) 32P-labeledsynthetic oligodeoxynucleotide probes but not to other first- or second-position probes. The data for the mutation-positive SCC and two representative mutation-negative SCCs are shown in Figure 3. Amplification of ras Oncogenes The genomic DNAs from all 40 human skin cancers were analyzed for possible amplification of ras genes by Southern blot hybridizationto 32P-labeledHa-ras, Ki-ras, or N-ras probes. None of 40 skin tumors contained amplified copies of the Ha-ras or Ki-ras oncogenes (data not shown). However, 3 of 40 human skin cancers contained highly amplified copies of the N-rasoncogene.The Southern blot of four SCCs shown in Figure 46 indicated that three of
(b) GTC
1 1
2
3
4
1
2
3
4
A
2
3
A m (GGT)
B C
D Figure 2. Dot-blot hybridization of PCR-amplified NIH 3T3 transformant DNAs and unexposed skin DNAs with 32P-labeled Ha-ras codon 12 wild-type and mutant oligodeoxynucleotide probes. Wells Al-A3 contain amplified DNA from the three NIH 3T3 transformants (AS-2T, DG4T, and M 4 T , respectively).Amplified DNA from 12 matching unexposed skin tissues of 12 skin cancer patients are spotted in wells 81-84, Cl-C4, and D1-D4, respectively. (a) Dot-blot hybridization with the wild-type (GGC) probe. (b) Dot-blot analysisof an identically spotted filter using a GFC probe.
Figure 3. Dot-blot hybridization of PCR-amplified human tumor DNAs with 32P-labeledwild-type and mutant oligonucleotide Ki-ras codon 12 probes. PCR-amplified DNAs from all 40 human skin cancers hybridized with the wild-type (GCT) Ki-ras codon 12 probe, but only 1 of 40 skin cancer DNAs hybridized with the mutant (AN) probe. Data for three SCCs are shown here, of which one is positive (well 1) and two are negative (wells 2 and 3). for this mutation. (A) Hybridization to the wild-type probe. (B) Hybridization to the mutant probe. The data indicate that a G-A base substitution occurred at the first position of Ki-ras codon 12 in the mutation-positiveSCC.
200
PIERCEALL ETAL.
A s 1 2 3 4
B 1 2 3 4
e 7 . 2 kb
Figure 4. Amplification of N-rassequences in human skin cancers. Genomic DNAs (10 pg each) from human skin cancers were digested with EcoRl and analyzed b Southern blot hybridization to "P-labeled N-ras probe, as iescribed in Materials and Methods. Data from four representativeSCC samples are shown here. Three of the SCC DNAs (lanes 2-4) show N-rasgene amplification (panel B). A picture of the ethidium bromide-stained gel before Southern blotting is shown in panel A. Lane S contains the molecular size standard of Hindll-cut A DNA. The molecular sizes of the six bands from top to bottom are 23.1, 9.4,6.6,4.4,2.3, and 2.0 kb, respectively.
them (lanes 2, 3, and 4) contained highly amplified copies of N-ras sequences. This amplification of the N-ras gene in three SCCs was not due to differences in the amount of DNA loaded onto each lane because the intensity of the ethidium bromide-stained gel was roughly the same in each lane (Figure 4A). Surprisingly, two of these SCCs (lanes 2 and 3) also contained G+T mutations at codon 12 of the Ha-ras oncogene (Figure 16, wells 65 and D4). DISCUSSION In our previous study, we had shown that genomic DNA from three of eight human skin cancers occurring on sunexposed body sites contained activated Ha-ras oncogenes capable of inducing morphologic and tumorigenic transformation of NIH 3T3 cells [20,211. To identify the mutations responsible for the activation of ras oncogenes in human skin cancers, we analyzed the three Ha-ras-positive NIH 3T3 transformants for mutations in codons 12 and 61 of Ha-ras oncogene; in addition, we analyzed 40 primary human skin cancers (24 SCCs and 16 BCCs) occurring on sun-exposed body sites for the presence of activating fa5 mutations by PCR amplification of tumor DNAs, followed by dot-blot hybridization to synthetic oligodeoxynucleotide probes designed to detect mutations at codons 12, 13, and 61 of Ha-ras, Ki-ras, and N-ras oncogenes. We found that 16 of 40 skin tumors analyzed (1 1 of 24 SCCs and 5 of 16 BCCs) contained identical G-T base changes at the second position of the Ha-ras codon 12. In addition, the three NIH 3T3 transformants that were shown previously
to contain transfected Ha-rassequences[20] also contained G+T mutations identical to those found in corresponding primary SCCs. This G+T mutation predicts a glycine-tovaline amino acid substitution. In contrast, only 1 of 40 human skin cancers had a G-A mutation in the first position of Ki-ras oncogene, predicting a glycine-to-serine amino acid substitution.The remaining 23 skin tumors did not contain any mutations at codons 12, 13, or 61 of Ha-rds, Ki-ras, or N-ras oncogenes. Similarly, none of 12 matching unexposed skin tissues from skin cancer patients analyzed contained G+T or other Ha-rds codon 12 mutations. Even though activation of ras oncogenes in human skin cancers has been reported previously, two aspects of our finding are striking: the specificity and the frequency of the mutated Ha-ras oncogene. Of all possible ras gene mutations for which we tested, only the second position of codon 12 of Ha-ras was mutated at a high frequency, and this mutation was exclusively a G-T base change. In contrast, only one of 40 human skin cancers tested contained a G+A mutation at the first position of Ki-rascodon 12, resulting in an amino acid substitution of glycine to serine. A second feature of this study is the frequency with which the Ha-ras mutation occurred in human skin cancers. We found that 40% of all nonmelanoma skin cancers examined (46% of SCCs and 3 t % of BCCs) contained this G-T base change at the second position of the Ha-ras codon 12. In contrast, previous reports have indicatedthat ras activation occurs at a relatively low frequency in human skin cancers. Corominas et al. [251 reported that although 30% of human KAs, a benign tumor that undergoes periodic regression, contained specific mutations (A+T transversions) in the second position of codon 61 of the Ha-ras oncogene, only 13% of human SCCs contained mutations in the Ha-ras oncogene. This observation has led to the hypothesis that activated Ha-ras oncogenes may play a role in initiation but not in tumor progression [25]. Suarez et al. 1221have reportedthat two of eight nonmelanoma skin tumors from XP patients contained A-T point mutations at codon 61 of the N-ras oncogene. More recently, van der Schroeff et al. (231found that of 12 SCC samples examined, only one contained a mutation (GGT+TGT) in codon 12 of the Ki-rasoncogene, and of 30 BCCs examined, three had Ki-ras mutations at codon 12 (GGT-TGT or GAT) and one had a mutation at Ha-ras codon 61 (CAG-CAT). The frequency of ras mutations that we observed in human skin cancers, although higher than that reported by other investigators, is not unexpected because ras gene mutations are known to occur at 80-90% frequencies in other types of human cancers such as colorectal[8,91 and pancreatic tumors [lo-1 21. Epidemiologic studies have indicated that UV radiation present in sunlight is responsible for the induction of most human skin cancers [32,33]. Our studies also support an active role for UV radiation in the induction of mutation in human skin cancers. The mutationsfound in our human skin cancers were primarily G+T base changes in the sense strand of Ha-ras codon 12 and, in one instance, a G+A base change in the sense strand of Ki-ras codon 12. The
20 1
RAS GENE MUTATIONS IN HUMAN SKIN CANCERS
specific base change in codon 12 of both Ha-rasand Ki-ras occurred directly opposite cytosine doublets in antisense DNA strands. Therefore, it is distinctly possible that the mutation in codon 12 of Ha-rasand Ki-ras may have arisen from induction of pyrimidine dimers in the antisense strand by UV radiation. By examining the following sequence of Ha-rasand Ki-ras codon 12, one can predict that a pyrimidine dimer (C-C) could be formed as a result of UV exposure. Ha-ras: Ki-ras:
sense strand antisense strand sense strand antisense strand
5'-GGC-3'
3 '-CCG - 5' 5'-GGT-3' 3'-CCA-5'
Although UV radiation is known to induce both pyrimidine dimers and (6-4) photoproducts 134-361, it appears from our studies and from studies by others that pyrimidine dimers, and not (6-4) photoproducts, are the principal premutagenic lesions responsible for the activation of ras oncogenes in human skin cancers occurring on sunexposed body sites. First, we observed that all the mutations (G-T or G-A) found in our human skin cancers were located opposite C-C doublets, potentialtargets for pyrimidine dimer formation. White and Balmain [24] have also found this identical G-T mutation in a human BCC. Second, Suarez et al. [22] have shown that the mutations in codon 61 of the N-ras oncogene in nonmelanoma skin cancers from XP patients were located directly opposite a T-T doublet in the antisense strand. Similarly, van der Schroeff et al. [23] have shown that the mutations in codon 12 of Ki-ras in human BCCs and SCCs occurred directly opposite a C-C doublet. Taken together these results suggest that UV-induced pyrimidine dimers are probably the premutagenic lesions involved in induction of mutation and subsequent transformation of cells. In addition to the mutations in the Ha-ras and Ki-ras oncogenes, amplification of N-ras was detected in some human skin cancers. Three (all SCCs) of 40 human skin cancers analyzed contained highly amplified copies of the N-rasoncogene. Interestingly, two of these SCCs also possessed a G-T mutation in the second position of Ha-ras codon 12. However, the third SCC contained only amplified copies of the N-rasgene and no other known rasgene mutations. Suarez et al. [371 reported that skin tumors from XP patients contained amplified Ha-ras, N-ras, and c-rnyc mRNA transcripts, in addition to a mutated N-ras and a rearranged Ha-ras oncogene. The presence of several oncogene alterations in the same tumor may have resulted from the accumulation of unrepaired UV-induced DNA lesions. These results suggest that gene amplification, in addition to mutation in rds genes, may also play a role in carcinogenesis. Mutational activation of ras oncogenes has also been detected in human melanomas. However, most of the mutations in human melanomas have been shown to occur at codons 12, 13, and 61 of the N-ras oncogene [38-401 and occasionally at codon 61 of the Ha-rasoncogene[41]. Because there is a growing concern that the incidence of
both melanoma and nonmelanoma human skin cancers may increase because of the recently reported decrease in stratospheric ozone, it is important to understand the molecular mechanisms by which UV radiation present in sunlight induces human skin cancers. ACKNOWLEDGMENTS
This research was supported by Grant R01-CA-46523 from the National Institutes of Health (awarded to HNA). William Pierceall is the recipient of a Predoctoral Fellowship from the National Institutes of Health under Training Grant T32-CA-09598. Received November 9, 1990; revised February 19, 1991, accepted February 20,1991.
REFERENCES 1. Bishop JM. Cellular oncogenes and their retroviruses. Annu Rev Biochem 52:301-354, 1983. 2. Barbacid M. Rasgenes. Annu Rev Biochem 56:779-827, 1987. 3. Bos 11. The ras gene family and human carcinogenesis. Mutat Res 195:255-271, 1988. 4. Bos JL. Ras oncogenes in human cancer: A review. Cancer Res 4914682-4689,1989. 5. Fujita I,Yoshida 0, Yuasa Y, Rhim IS,Hatanaka M, Aaronson SA. Ha-ras oncogenes are activated by somatic alterations in human urinary tract tumors. Nature 309:464-466, 1984. 6. Malone PR, Visvanathan KV, Ponder BA, Shearer RJ, Summerhayes IC. Oncogenesand bladder cancer. Br J Urol57:664-667, 1985. 7. Visvanathan KV, Pocock RD, Summerhayes IC. Preferential and novel activation of H-ras in human bladder carcinomas.Oncogene Res3:77-86, 1988. 8. Bos JL, Fearon ER, Hamilton SR, et al. Prevalenceof ras gene mutations in human colorectal cancers. Nature 327:293-297, 1987. 9. Forrester KC, Almquera C, Han K, Grizzle WE, Perucho M. Detection of high incidence of K-ras oncogenes during human colon tumorigenesis. Nature 327:298-303, 1987. 10. Almoguera C, Shibata D, Forrester K, Martin J, Arnheim N, krucho M. Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53:549-554, 1988. 11. Smit VTHMB, Boot AIM, Smit AMM, Fleuren GJ, Cornelisse CJ, Bos 11. K-ras codon mutations occur very frequently in pancreaticadenocarcinomas. Nucleic Acids Res 16:7773-7782, 1988. 12. Grunewald K, Lyons J, Frohlich A, et al. High frequency of Ki-ras codon 12 mutations in pancreatic adenocarcinomas. Int J Cancer43:1037-1041, 1989. 13. Kumar R, Sukumar 5, Barbacid M. Activation of ras oncogenes precedingthe onset of neoplasia.Science 248: 1 101-1 104, 1990. 14. Land H, Parada LF, Weinberg RA. Cellular oncogenesand multistep carcinogenesis. Science 222:771-778, 1983. 15. Zarbl H, Sukumar 5,Arthur AV, Martin-Zanca D, Barbacid M. Direct mutagenesis of Ha-ras-1 oncogenes by Knitroso-N-methylurea during initiation of mammary carcinogenesis in rats Nature 315 382-385.1985 16. Quintanilla M, Brown K, Ramsden M, Balmain A. Carcinogenspecific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature 322:78-80, 1986. 17. Brown K, Buchmann A, Balmain A. Carcinogen-induced mutations in the mouse c-Hams gene provide evidence of multiple pathways for tumor progression. Proc Natl Acad Sci USA 87:538542, 1990. 18. Balmain A, Brown K. Oncogene activation in chemical carcinogenesis.AdvCancerRes51:147-182, 1988. 19. Suarez HG. Activated oncogenes in human tumors. Anticancer Res 9:1331-1334,1989, 20. Ananthaswamy HN. Price JE, Goldberg LH, Bales ES. Detection and identification of activated oncogenes in human skin cances occurringonsun-exposedbodysites. CancerRes48:3341-3346,1988. 21. Ananthaswamy HN, Price JE, Goldberg LH, Bales ES. Simultaneous transfer of tumorigenic and metastatic phenotypes by transfection with genomic DNA from a human cutaneous squamous cell carcinoma. J Cell Biochem 36:137-146, 1988. 22. Suarez HG, Daya-Grosjean L, Schlaifer D, et al. Activated oncogenes in human skin tumors from a repair-deficient syndrome, xeroderma pigmentosum. Cancer Res 49: 1223-1 228, 1989.
202
PIERCEALL ETAL.
23. van der Schroeff JG, Evers LM, Boot AJM, Bos JL. Ras oncogene mutations in basal cell carcinomas and squamous cell carcinomasof human skin. J Invest Dermatol94:423-425, 1990. 24. White SI, Balmain A. G to T mutation in codon 12 of the human Harvey ras oncogene derived from a basal cell carcinoma. J Invest Dermatol91:407, 1988. 25. Corominas M, Kamino H, Leon J, Pellicer A. Oncogene activation in human benign tumors of the skin (keratoacanthomas):Is HRAS involved in differentiation as well as proliferation?Proc Natl Acad Sci USA86:6372-6376, 1989. 26. Wigler M, Pellicer A, Silverstein 5, Axel R. Biochemical transfer of single-copy eucaryotic genes using total cellular DNA as donor. Cell 14:725-731, 1978. 27. Southern EM. Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98:503-517, 1975. 28. Verlaan-de Vries M, Bogaard ME, van den Elst H, van Boom JH, van der Eb AJ,60s JL. A dot-blot screening procedure for mutated ras oncogenes using synthetic oligodeoxynucleotides. Gene 50:313-320,1986, 29. Saiki RK, Gelfaud DH, Stoffel S, et al. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487-491,1988. 30. Nelson K, Mathur El. Perfect MatchTMenhancer: Limits false priming events during amplification reactions. Strategies 3: 17-19, 1990. 31. Maniatis T, kitsch EF, Sambrook 1. Molecular Cloning, A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1982. 32. Urbach F. Evidence and epidemiology of ultraviolet-inducedcancersin man. NCI Monogr 50:s-10.1978.
33. Scotto J, Fears TR. Skin cancer epidemiology: Research needs. NCI Monogr 50:169-177,1978. 34. Setlow JK. The molecularbasis of biolwical effectsof ultravioletradiationand photoreactivation. CurrTop Radiat Res2: 195-248,1966. 35. Lippke JA. Gordon K, Brash DE, HaseltineWA. Distribution of UV light-induced damage in a defined sequence of human DNA: Detection of alkaline-sensitive lesions a t pyrimidine nucleosidecytidinesequences. Proc Natl Acad Sci USA78:3388-3392,1981. 36. Rosenstein BS, Mitchell DL. Action spectra for the induction of pyimidine (64) Wrimidone photoproductsand cyclobutanedimers in normal human skin fibroblasts. Photochem Photobiol45:775781, 1987. 37. Suarez HG, Nardeux PC, Andeol Y, Sarasin A. Multiple activated oncogenes in human tumors. Oncogene Res 1 :201-207, 1987. 38. van't Veer U, 8urgering BMT, hrsteeg R, et al. N-ras mutations in human cutaneous melanoma from sun-exposedbody sites. Mol Cell Biol9:3114-3116, 1989. 39. Padua RA, Barrass NC, Currie GC. Activation of N-ras in a human melanoma cell line. Mol Cell Biol5:582-585, 1985. 40. Keijzer W, Mulder Me Langeveld JCM, et al. Establishment and characterization of a melanoma cell line from a xeroderma pigmentosum patient: Activation of N-ras at a potential pyrimidine dimer site. Cancer Res49: 1229-1 235, 1989. 41. Shukla VK, Hughes DC, Hughes LE, McCormick F, Padua RA. Ras mutations in human melanotic lesions: K-ra5 activation is a frequent and early event in melanoma development. Oncogene Res 5: 121-127, 1989. 42. Reddy EFI Reynolds RK, Santos E, Barbacid M. A point mutation is responsible for the acquisition of transforming properties by the T24 human bladder carcinoma oncogene. Nature 300: 149152,1982.
