Transformation of immortalized woodchuck hepatic cell lines with the c-Ha-ras proto-oncogene

Woodchuck hepatocytes were immortalized with the simian virus40 T antigen (SV40 T-ag) oncogene and utilized in an oncogenictransformation assay. Transfection of these cell lines withan activated c-Ha-ras oncogene (EJ6.6) resulted in the transformationof cells to a phenotype characterized by anchorage-independentgrowth in soft agar. Colonies of transformed cells were subclonedand up to 80% were positive for oncoprotein expression detectedby immunoblotand Northern blot procedures. When compared withthe parental cell lines, ras-transformed derivatives were alteredboth morphologically and in growth rate. The tumorigenic potentialof c-Ha-ras transformed cells was demonstrated in severe combinedimmunodeficient (SCID) mice. There was a latency period of 1to 4 weeks before tumors were detectable and a period of over7 weeks was required for tumors to reach a diameter of 1 cm.Histologically, tumorsderived from cell lines fully transformedby the SV40 T-ag had the appearance of well differentiated hepatocellularcarcinoma (HCC) while tumors derived from c-Ha-ras transformedcell lines had the appearance of poorly differentiated HCC.The capacity to induce oncogenic transformation events in immortalizedwoodchuck hepatic cell lines should provide the opportunityto study the cooperative effects of hepadnaviral genes in hepatocarcinogenesisin vitro.     

Quantitation of early clonal expansion of two mutant 61st codon c-Ha-ras alleles in DMBA/TPA treated mouse skin by nested PCR/RFLP

It has been hypothesized that tumor promotion in mouse skininvolves clonal expansion of initiated cells with activatedc-Harvey (Ha)-ras oncogene to give rise to benign tumors. Wehave used the two stage mouse skin carcino-genesis model using7, 12-dimethylbenz[a]anthracene (DMBA) as the initiator and12-0-tetradecanoyl-phorbol-13-acetate (TPA) as the tumor promoterto quantitate the number of mutated c-Ha-ras alleles in mouseepidermal DNA. Epidermal samples were harvested over a 12-weekperiod before the appearance of papillomas. Three 61st codon(i.e. CAA) c-Ha-ras mutations, CTA (T2), CGA (G2) and CAT (T3)were quantitated by newly developed nested PCR/RFLP assays.During TPA promotion the number of T2 mutant copies showed aprogressive increase starting at 4 weeks after initiation andthe number of T3 mutant alleles showed an increase startingat 6 weeks. By 12 weeks after initiation, TPA-promoted mouseepidermis averaged     

Association of deficient DNA repair during G2 phase with progression from benign to malignant state in a line of human skin keratinocytes transfected with ras oncogene

Human skin keratinocytes after malignant neoplastic transformationby infection with Kirsten murine sarcoma virus (KiMSV) or transfectionwith pSV² ras (containing an activated c-Ha-ras oncogene) showeda DNA repair deficiency(ies). The repair deficiency was manifestas an abnormally high frequency of chromatid breaks and gapspersisting after X-ray-induced DNA damage inflicted during theG₂ phase of the cell cycle. Non-tumorigenic control cells atthat time were clearly repair-efficient. By analyzing benignand malignant tumorigenic HaCaT-ras clones, we could excluderas p21 oncoprotein expression as the causal mechanism for repairdeficiency, since both clone types expressed similar levelsof the mutated protein and only the malignant tumorigenlc cellsshowed repair deficiency. The results suggest that mutated p21ras provided the human keratinocytes with a growth advantagein vivo (benign tumor growth), but acquisition of repair deficiencyis required for progression from benign to malignant state.     

Overexpression of c-K-ras, c-N-ras and transforming growth factor {beta} co-segregate with tumorigenicity in morphologically transformed C3H 10T1/2 cell lines

Morphologic transformation and tumorigenicity are separate cellularphenotypes in transformed 1OT1/2 cells. We have investigatedthe levels of expression of genes for c-myc, c-H-ras, c-K-ras,c-N-ras, TGFß and Rb in 42 morphologically transformed1OT1/2 cell lines, in an attempt to define the molecular mechanisntsgoverning morphologic transformation and tumorigenicity in the1OT1/2 cell system. The 101 1/2 cell lines investigated generallyoverexpressed mRNAs for c-myc, c-H-ras, and TGFß relativeto the levels expressed by wild- type 10T1 cells (levels ofexpression >1.5-fold that of wild- type 1011/2 cells). Incontrast, only half of these cell lines overexpressed mRNAsfor c-N-ras and/or Rb relative to wild- type 1OT1/2 cells, andonly 25%; overexpressed c-K-ras mRNA. The mean levels of mRNAexpression for each of c-K-ras, c-N-ras and TGFß genesin tumorigenic cell lines were significantly greater than themean levels of expression in non-tumorigenic cell lines, suggestingan association between twnorigenicity and the levels of expressionof these specific genes. In contrast, levels of expression forc-myc, c-H-ras and Rb genes were not correlated with tumorigenicity.Cell lines that coexpressed high levels of c-K-ras, c-N-rasand TGFß genes were likely to be tumorigenic (11/12cell lines were tumorigenic), whereas cell lines that coexpressedlow levels of these genes were unlikely to be tumorigenlc (1/10cell lines were tumorigenic). High expression of TGFßwas sufficient for tumorigenicity in the absence of high levelsof expression of c-K-ras and c-N-ras (5/5 cell lines were tumorigenic).Elevated expression of either c-K-ras or c-N-ras alone was Insufficientfor tumorlgenlclty, however, coordinate overexpresslon of bothc-K-ras and c-ras was associated with tumorigenicity irrespectiveof the expression status for TGFß (13/15 cell lineswere tuinorigenic). These results suggest that overexpresslonof c-ras, c-H-ras and TGFß are conunonly associatedwith, and possibly mechanistically related to, the process ofmorphologic transformation in 1OTI/2 cells. In addition, theseresults suggest that progression from morphologic transformationto tumorigenicity in 1011/2 cell lines is frequently accompaniedby overexpression of c-K-ras and c-N-ras, and by enhancementof the level of overexpresslon of TGFß     

Malignant transformation of mouse BALB/c3T3 cells induced by NaNO2

The addition of sodium nitrite (NaNO₂; 5–20 mM) for 72h to mouse BALB/c3T3 cells resulted in the induction of transformedfoci (type III foci) in a dose-dependent manner. The cells isolatedfrom the NaNO₂-induced transformed foci produced progressivelygrowing tumors when inoculated into nude mice subcutaneouslyat an inoculum size of 1 x 10⁶ cells per site. In contrast,the original untreated cells did not take even at an inoculumsize of 1 x 10⁶ cells per site. The possibility that NaNO₂ mightreact with cellular or medium components to make carcinogenicN-nitrosamines and that these might induce cell transformationwas examined and almost excluded. Thus, nitrite itself seemsto have a cell transforming activity. Recent evidence suggeststhat NO₂^– is produced by the activated macrophage of mammals.We also detected NO₂^– production in culture media in themouse macrophage-like cell Line J774–A1 after lipopolysaccharide(LPS) treatment, and also in the human promyeloleukemia cellline HL60 after differentiation into macrophage-like cells by12-O-tetradecanoyl phorbol-13-acetate and further activationby LPS.     

Cathepsins D,B, and L in transformed human breast epithelial cells

To investigate the regulation of lysosomal enzymes during carcinogenesis, we measured cathepsins (Cats) D, B, and L in MCF-10F, which is a human breast epithelial cell line, and cells evolved after treatment with carcinogen and transfected with c-Ha-ras oncogene. The clones used in this study, MCF-10FTras, D3, D3-1, and D3-1Tras, expressed no estrogen receptors and gradually increased invasive potential, while oncogenetransfected lines were also tumorigenic in SCID mice [16,19]. Cats D, B, and L were determined in the cells and in cell media using enzyme-linked immunosorbent assay (ELISA), specific enzyme activity measurements, and immunocytochemistry. The major intra- and extracellular lysosomal proteinase in these cells was Cat D (30–180 pm/mg), followed by Cat B (2–10 pm/mg) and Cat L (1–5 pm/mg). An inverse relationship between intracellular Cat D levels and invasive potential of carcinogen-treated and c-Ha-ras oncogene-transfected cell lines was observed. No significant changes in extracellular concentration of Cat D precursor in this series of cell lines was observed. Intracellular levels of Cats B and L were unchanged or slightly lower in carcinogentreated D3 and D3-1 cells, as well as in MCF-10FTras. On the other hand, in D3-1Tras cell line, evolving from c-Ha-ras transfected D3-1 line, 3.5 fold and 4.4 fold increases in Cat B and Cat L, respectively, but a 2 fold decrease in Cat D, were observed compared to the parental cell line. Immunocytochemical staining showed a granular, polarized perinuclear and cytoplasmic staining of cathepsins in all cell lines. Cysteine proteinases stained more frequently and more intensely in D3-1Tras compared to other lines, confirming the immunochemical assays. We hypothesize that several molecular events, caused by a carcinogen and an oncogene such as c-Ha-ras, are needed to increase Cat B and Cat L, but not Cat D, expression. Therefore, the cysteine and aspartic lysosomal proteinases are differentially expressed in the breast cell lines with more invasive phenotype.     

Different metastatic potentials of ras- and src-transformed BALB/c 3T3 A31 variant cells

The metastatic phenotype of tumor cells is thought to be induced by an aberrant signaling cascade or cascades that are different from those required for tumorigenicity. Oncogene-transfected cells with different tumorigenicities and metastatic potentials have been used to identify such pathways and responsible molecules. However, oncogenes that can induce tumorigenicity in recipient cells also frequently induce the metastatic phenotype at the same time. The difficulty in obtaining cell lines that are tumorigenic but not metastatic has hampered such studies. In this report, we transfected the activated c-Ha-ras oncogene into BALB/c 3T3 A31 variant cells and found that the transfectants were tumorigenic but they did not form metastatic lung nodules in the experimental metastasis assay. The phenotype was very stable and was maintained during cultivation. On the other hand, the metastatic potentials of either the transfected cells or the original variant cells could be induced by transfection of the v-src oncogene. The src transfectants formed extensive nodules in lung when injected into the tail veins of congeneric mice. The cell motility of the metastatic src transfectants on Matrigel-coated dishes was greater than that of the ras transfectants. The src transfectants were also invasive in Matrigel when analyzed on a filter. These variant cells transformed by the ras and src oncogenes will be a useful system for identifying the signaling cascades responsible for the metastatic potential of tumors. © 1996 Wiley-Liss, Inc.     

Characterization of activated proto-oncogenes in chemically transformed syrian hamster embryo cells

The Syrian hamster embryo (SHE) cell transformation model has been used by many investigators to study the multistep process of neoplastic transformation induced by chemical carcinogens. In this study we have attempted to determine if activated proto-oncogenes are present in the transformed cells induced by a variety of chemical carcinogens. Twelve carcinogen-induced hamster cell lines, established by treatment of normal SHE cells with benzo[a]pyrene, diethylstilbestrol, or asbestos, were examined. One spontaneously transformed cell line (BHK-A) was also studied. Some of the cell lines were also tested for oncogene activation at the preneoplastic stage, before they acquired tumorigenic potential. DNAs from normal, preneoplastic, and neoplastic cells were tested by transfection into mouse NIH 3T3 cells, and morphologically transformed foci were scored on the contact-inhibited monolayer of 3T3 cells. The frequency of focus formation for normal SHE cell DNA was <0.0008 foci/μg DNA, while approximately 40% (5 of 12) of the DNAs from carcinogen-induced, tumorigenic hamster cell lines induced foci at a frequency of ? 0.012 foci/μg DNA. The other seven carcinogen-induced cell lines and the BHK-A cells were negative (<0.002 foci/μg DNA). When the DNAs from transformed foci induced by the five positive cell lines were retransfected into NIH 3T3 cells, the frequency of secondary foci of 3T3 cells was as much as 50-fold higher (1.34 foci/μg DNA) than with the primary transfectants. DNAs from transformed foci or tumors derived from transformed foci were screened by Southern blot analyses with known oncogenes and with a hamster repetitive DNA probe for the presence of transfected hamster oncogenes. Newly acquired hamster Ha-ras sequences were detected in transformed 3T3 cells induced by four of the five hamster tumor DNAs. Immunoprecipitation of lysates of several secondary transformants with a ras monoclonal antibody (Y13–259) showed altered gel mobility of the p21^ras protein consistent with a mutation at codon 12. These activated ras genes were detected by the NIH 3T3 assay in the tumorigenic hamster cells but not in the preneoplastic, immortal cell from which they were derived. The activated Ha-ras proto-oncogene was detected in cell lines induced by each of the three different carcinogens studied. Cells from transformed foci inauced by DNA from one of the hamster tumor cell lines (BP6T) contained hamster sequences but did not show newly acquired Haras, Ki-ras, or N-ras genes on Southern analysis or altered p21^ras protein. The transforming gene in this cell line appears to be a non-^ras oncogene. These observations indicate that ～40% of the chemically transformed Syrian hamster tumor cell lines have activated Ha-ras oncogenes. The activation of Ha-ras proto-oncogene is a late, postimmortalization step in the neoplastic progression of SHE cells. Only one cell line with a non-^ras oncogene was detected in the NIH 3T3 focus assay, and ～60% of the cell lines were inactive in this assay, indicating the need to develop alternative assay systems for oncogene activation. Some of the preneoplastic Syrian hamster cell lines may be useful for this purpose.     

Activation of H-ras oncogene in 3-methylcholanthrene-transformed human cell line

DNA prepared from the 3-methylcholanthrene (3MC)-transformedhuman 312H cell line induced foci on NIH/3T3 cells, whereasDNAs prepared from 7,12-dimethylbenz[a]-anthracene-transformedand the dimethylsulfoxide control 312H cell lines failed toinduce foci. The transformed gene from the 3MC-transformed 312Hcells was identified as an activated form of the human cellulartransforming H-ras oncogene. Analysis of the ras oncogene p21product in this transformant by immunoprecipitation and gelelectrophoresis suggested that this gene was activated by mutationin the 61st codon. These findings demonstrate that activationof a member of the ras gene family can occur in a chemicallytransformed human cell line.     

Detection of distinct transforming genes in X-ray induced tumors

DNAs from mouse skin tumors (papillomas, squamous cell carcinomas,basal cell carcinomas and pilomatrixomas) initiated with X-irradiationand promoted with 12-O-tetradecanoyl-phorbol-13-acetate (TPA)demonstrated dominant transforming activity by the productionof transformed foci in the mouse recipient tine, NIH3T3. Dominanttransforming activity was not found in DNA isolated from normalmouse epidermis or from the corresponding liver. The NIH3T3transformants induced with squamous cell carcinoma DNA grewin soft agar and formed tumors in nude mice. Southern blot analysisof primary NIH3T3 transformant DNAs carrying oncogenes fromradiation-initiated squamous cell carcinomas indicated thatthe oncogenes responsible for the transformation of the recipientcells were not Ha-ras, Ki-ras or N-ras genes, nor were theyerbB, B-lym, met, neu or raf. The data presented indicate thatDNAs from radiation-initiated mouse skin tumors contain dominanttransforming genes that are detectable by DNA-mediated genetransfer. The oncogene sequences activated in these radiation-initiatedtumors are distinct non-ras transforming genes.     

Inactivation of a tumor suppressor function in immortal Syrian hamster cells by N-methyl-N'-nitro-N-nitrosoguanidine and by 5-aza-2'-deoxycytidine

Clonal lines of immortal Syrian hamster cells were previouslyisolated that either suppressed (supB⁺) tumorigenicity in hybridswith a malignant hamster cell line (BP6T) or had lost this suppressionability (supB^–). Neither line was tumorigenic or showedanchorage-independent growth in normal growth medium. SupB^–cells, but not supB⁺ cells, grew in agar supplemented with thegrowth factors EGF, PDGF and insulin (EPI), providing a selectiveassay for the supB^–phenotype. After treatment of supB⁺cells with either N-methyl-N'-nitro-N-nitrosoguanidine (10–300ng/ml) or 5-aza-2'-deoxycytidine (25–250 ng/ml), and anexpression period of 4–8 weeks, a dose-dependent increasein altered cells that grew in agar supplemented with EPI wasobserved. Cell lines derived from colonies in agar showed persistentEPI-stimulated growth in agar, and decreased suppression ofgrowth in agar for hybrids with BP6T cells. Thus, carcinogen-inducedloss of the tumor suppressor phenotype has been demonstrated.     

Epidermal papillomas and carcinomas induced in uninitiated mouse skin by tumor promoters alone contain a point mutation in the 61st codon of the Ha-ras oncogene

Previous results in a number of laboratories have demonstratedthat epidermal papillomas and carcinomas induced by the two-stageprotocol of initiation and promotion contain a point mutationin the 61st codon of the c-Ha-ras oncogene when the initiatingagent used is 7,12-dimethylbenz-[a]anthracene (DMBA). In thepresent report, we have analyzed DNA purified from ‘spontaneouslyinitiated’ papillomas and carcinomas induced in SENCARmouse epidermis after repetitive treatments with a tumor-promotingagent. Southern blot hybridization studies of tumor DNA digestedwith diagnostic restriction endonudeases demonstrated that sevenof nine papillomas and carcinomas contained a point mutationin the 61st codon of one allele of the c-Ha-ras oncogene. Theimplications of our findings with respect to the role whicha point-mutated Ha-ras protooncogene plays in initiation ofskin tumorigenesis are discussed.     

Dose-dependent ras mutation spectra in N-nitrosodiethylamine induced mouse liver tumors and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone induced mouse lung tumors

In a previous study, the spectrum of H-ras mutations detectedin B6C3F1 mouse liver tumors induced by 5, 50 or 150 µmol/kgbody wt of N-nitrosodiethylamine (NDEA) was similar to thatin spontaneous B6C3F1 mouse liver tumors, suggesting that activationof the H-ras gene in NDEA-Induced mouse liver tumors may notbe the direct result of the chemical interaction with the H-rasgene. In the present study, mutations in the H-ras oncogenefrom B6C3F1 mouse liver tumors induced by 5 or 50 µmol/kgbody wt of NDEA were characterized by DNA amplification withpolymerase chain reaction (PCR), single-strand conformationpolymorphism (SSCP) and direct sequence analysis. Twenty-oneof 66 NDEA-induced B6C3F1 mouse liver tumors contained activatedH-ras gene with 2 of 21 having a CG to AT transversion at thefirst base of codon 61, 17 of 21 having AT to GC transitionand 2 of 21 having an AT to TA transversion at the second baseof codon 61 in the H-ras gene. The predominant mutation, ATto GC transition (17/21, 81%) is consistent with the formationof O⁴-ethylthymine adduct, and is distinct from the predominantCG to AT transversion (50%) at the first base of codon 61 detectedin H-ras gene from NDEA-induced B6C3F1 mouse liver tumors ina previous study by Stowers et al. Mutations in the K-ras oncogenefrom 59 A/J mouse lung tumors induced by 0.53 mmol/kg body wtof 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) werealso characterized by using the above mentioned methods. Forty-sixof 59 NNK-induced A/J mouse lung tumors contained activatedK-ras genes. All 46 (100%) of the activated K-ras gene had GCto AT transitions at the second base of codon 12. The same mutationwas observed in 70% (7/10) of the K-ras oncogene from A/J lungtumors induced by 4.8 mmol/kg body wt (given in 21 doses) ofNNK. These data suggest that other factors in addition to genotoxiceffect might be involved in the induction of rodent tumors bysome carcinogens when given at higher doses. Therefore, furtherstudies to compare the dose-dependent differences in the profileof ras mutations induced by chemical carcinogens may help toassess human cancer risk. Mutation(s) in exons 5-8 of the p53gene was not found in these NDEA-induced mouse liver tumorsand NNK-induced mouse lung tumors.     

Separation of transforming amino acid-substituting mutations in codons 12, 13 and 61 of the N-ras gene by constant denaturant capillary electrophoresis (CDCE)

We used high fidelity PCR and constant denaturant capillaryelectrophoresis (CDCE) [Khrapko et al. (1994) Nucleic AcidsRes., 22, 364–369] to separate wild type and differentmutant N-ras exon 1 and 2 sequences. The set of plasmids containingN-ras cDNA, wild type or mutant sequences representing all transformingamino acid-substituting single base pair changes in codons 12,13 (exon 1) and 61 (exon 2), were amplified using Pfu polymerasein a limited cycle polymerase chain reaction. One of the primersused for the amplification of each exon included a 40 nucleotideGC rich sequence that created high and low melting domains.The amplified fragments 151 bp (exon 1) and 150 bp (exon 2)were run on the CDCE with the ‘denaturant zone’temperature of the capillary corresponding to the melting temperatureof 111 bp (exon 1) and 110 bp (exon 2) low melting domains.The separation was achieved between wild type and mutant sequencesas homoduplexes in 15 out of 19 cases, as a single base substitutionalters the electrophoretic mobility of a partially melted doublestranded fragment. The denaturation and reannealing of wildtype and mutant fragments together created wild type/ mutantheteroduplexes. All the heteroduplexes were well resolved fromwild type homoduplex. In the current form mutant sequences weredetected at a frequency of 10^–3 in the presence of wildtype. This study has resulted in obtaining electrophoretic spectrumof different N-ras mutants on CDCE as homoduplexes as well asheteroduplexes.     

Induction of thymomas by N-methyl-N-nitrosourea in AKR mice: interaction between the chemical carcinogen and endogenous murine leukaemia viruses

AKR mice develop thymomas spontaneously when >6 months oldbut when young AKR mice are treated with N-methyl-N-nitrosourea(MNU) they develop thymomas at 3–6 months of age. In thisstudy the potential role of oncogene activation in the developmentof both the spontaneous and MNU-induced thymomas in AKR micehas been examined by DNA transfection into NTH3T3 mouse fibroblastsand by Southern analysis of tumour DNA. The results show thata high proportion of MNU-induced thymomas contain activatedcellular ras^K while no activated cellular ras genes were detectedin spontaneous thymomas. Southern analysis of tumour DNA revealedthat 2/30 spontaneous tumours and 2/52 MNU-induced tumours containedalterations in the c-myc gene while 5/29 spontaneous tumoursand 6/56 MNU-induced tumours contained alterations in the Pim-1gene. A more detailed analysis of the Pim-1 gene demonstratedthat the alterations observed in most MNU-induced and spontaneoustumours resulted from proviral integration at the 3' end ofthis gene. Our analyses also demonstrated that the majorityof MNU induced tumours, including those containing rearrangementsin the Pim-1 gene, lacked the somatically acquired recombinantMCF proviruses that are present in most spontaneous AKR lymphomas.These results provide evidence that (i) the mechanisms of developmentof MNU-induced and spontaneous tumours in AKR mice are distinctand (ii) the development of thymomas that contain proviral integrationsat the Pim-1 locus in the MNU-treated AKR mice involve cooperationbetween the chemical carcinogen and endogenous murine leukaemiaviruses.     

Cell cycle-dependent modulation of O6-methylguanine-DNA methyltransferse in C3H/10T1/2 cells

O⁶-methylguanine-DNA methyltransferase (MGMT) was measured inpartially synchronized cultures of C3H/10T1/2 mouse embryo cellsas a function of cell cycle. The degree of synchrony and progressionof the cell cycle were monitored by flow cytometry. The MGMTlevel was significantly reduced prior to the onset of S-phase.This reduction was concomitant with the inhibition of in vivorepair of O⁶-methylguanine in DNA of S-phase cells as observedearlier. The recovery of the MGMT level paralleled the progressionof synchronized cells into G₂. S-phase cells purified by cellsorting contained -15% of the MGMT present in G_o or early G₁cells. A comparison of the in vivo repair of O⁶-methylguanineand MGMT levels suggests that the lack of repair of O⁶-methylguaninein DNA of the mouse embryo cells is due only in part to a temporalloss of MGMT.     

GC-rich regions in genomes as targets for DNA alkylation

For many DNA-damaging agents, the extent of damage at any givenbase site is influenced by the DNA sequence surrounding thatsite. Most agents that alkylate the guanine N7 position, includingmechlorethamine (nitrogen mustard) and benzo(a)pyrene diol epoxide,alkylate oligo-guanine sequences preferentially. Since thesedata suggest that guaninecyosine(GC)-rich regions in genes couldbe preferred sites of damage by these agents, GenBank was searchedfor genes containing 30 bp sequences of > 90% GC (GC runs).While primate, rodent, other mammalian, vertebrate and animalvirus genes constituted 57% of the annotated entries, they included90% of the entries with the GC runs. In addition, the percentageof oncogenes in the group of the entries with GC runs was higherthan that in the overall database. One gene of interest containingGC runs was the human c-Ha-ras oncogene. All seven GC runs inthe c-Ha-ras gene are in the 5'flanking region, rather thanin the coding sequences. In fact, some of the GC runs are containedin Sp1-binding enhancer sequences. Gel analysis of the alkylationof cloned c-Ha-ras DNA by several carcinogenic alkylating agentsstrongly suggest that in this gene GC runs can be preferredsites of damage. These observations suggest mechanisms by whichDNA damage at sites other than oncogene coding sequences mayplay a role in carcinogenesis and/or chemotherapy.