[CANCER RESEARCH 56. 2375-2381, May 15. 1996]

Mechanisms for the Involvement of DNA Methylation in Colon Carcinogenesis1

Christoph Schmutte, Allen S. Yang, TuDung T. Nguyen, Robert W. Beart, and Peter A. Jones2

Departments of Biochemistry and Molecular Biology ¡C.S.. A. S. Y.. T. T. N.. P. A. J.I and Colorectal Surgery ¡K.W. B.¡,University of Southern California, School of Medicine,Los Angeles. California 90033-0800

ABSTRACT

C —¿�>T transitions at CpG sites are the most prevalent mutations found

in the p53 tumor suppressor gene in human colon tumors and in thegermline (Li-Fraumeni syndrome). All of the mutational hot spots aremethylated to 5-methylcytosine, and it has been hypothesized that themajority of these mutations are caused by spontaneous hydrolytic deami-

nation of this base to thymine. We have previously reported that bacterialmethyltransferases induce transition mutations at CpG sites by increasingthe deamination rate of C —¿�»U when the concentration of the methylgroup donor S-adenosylmethionine (AdoMet) drops below its A1M,suggest

ing an alternative mechanism to create these mutations. Unrepaired uracilpairs with adenine during replication, completing the ( ' —¿�•T transition

mutation. To determine whether this mechanism could contribute to thedevelopment of human colon cancer, we examined the level of DNA(cytosine-S)-methyltransferase (MTase) expression, the concentration ofAdoMet, and the activity of uracil-DNA glycosylase in human colon

tissues, and searched for the presence of mutations in the MTase gene.Using reverse transcription-PCR methods, we found that average MTaseniKNA expression levels were only 3.7-fold elevated in tumor tissues

compared with surrounding normal mucosa from the same patient. Also,no mutations were found in conserved regions of the gene in 10 tumorssequenced. High-performance liquid Chromatographie analysis of extracts

from the same tissues showed that AdoMet concentrations were notreduced below the A,,, value for the mammalian enzyme, and the concentration ratio of AdoMet:5-adenosylhomocysteine, the breakdown product

of AdoMet and the competitive MTase inhibitor, did not differ significantly. Finally, extracts from the tumor tissue efficiently removed uracilfrom DNA. Therefore, biochemical conditions favoring a mutagenic pathway of C —¿�>U —¿�>T were not found in a target tissue known to undergo ahigh rate of C —¿�»T transitions at CpG sites.

INTRODUCTION

The dinucleotide CpG has been found to have a special role in thegenome and in the development of human cancers. Most of the CpGsin human DNA are methylated at the 5-position of cytosine, which is

the only known covalent modification in DNA. Areas of the genomewithout methylation are clustered as "CpG islands," which are often

associated with promoter regions of genes (1,2) and are generallymethylated on the inactive X chromosome (3) and on imprinted genes(4). Changes in global genomic methylation and methylation patternsare among the most consistent findings in the development of humancancers (5-7). This has led to the hypothesis of an epigenetic mech

anism for cancer development, in which changes in methylationinterfere with expression of proto-oncogenes and tumor suppressor

genes (8). In addition, involvement of methylation is also the mostplausible explanation for the high rate of C —¿�»T transition mutations

at CpG sites, which are frequently found in several human cancers inkey genes such as the p53 tumor suppressor gene, especially in colontumors and in the germline (Li-Fraumeni syndrome; Ref. 9). Five of

six mutational hot spots in the p53 gene are CpG sites; 47% of all

Received 11/7/95; accepted 3/14/96.The costs of publication of this article were defrayed in pan by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by National Cancer Institute Grant R35 CA49758." To whom requests for reprints should be addressed, at University of Southern

California/Morris Cancer Center, 1441 Easttake Avenue. Los Angeles. CA 90033. Phone:(213) 764-0816; Fax: (213) 764-0102.

mutations at these sites in colon tumors are C —¿�»T or G —¿�>A

transitions (10); and all of these hot spots in p53 have been found tobe methylated (9, 11, 12), which is consistent with a methylation-

induced mutational mechanism. Other genes, such as the factor IXgene, show a similar accumulation of transition mutations at CpG sites(13). In addition, CpG sites are underrepresented in human DNA bya factor of 5 (14), and it is believed that transition mutations from C—¿�>T or G —¿�»A on the opposite strand, respectively, caused the loss

of CpGs during evolution (15, 16). We have argued that these mutations are caused by an endogenous mechanism, because epidemiolog-

ical studies have not revealed any regional differences in the patternsof germline mutations (17), and no strand bias has been found in therepair process (10).

Two mechanisms have been proposed to explain why these CpGsites are mutational hot spots in human tumors (9, 18). Amino groupsof cytosine and 5mC3 are less stable in DNA than the amino groups

of guanine and adenine and hydrolyze spontaneously, forming uraciland thymine, respectively (19, 20; for review, see Ref. 21). Differences in the repair efficiencies of the U:G versus T:G mismatchescreated by this process are also likely to contribute substantially to theformation of mutational hot spots at CpG sites (22, 23). On the otherhand, the methylation process itself could increase the deaminationrate at the target cytosine. During the normal transfer reaction of themethyl group, the MTase flips its target cytosine out of the DNA helixinto the catalytic pocket of the enzyme, where a cysteine residue of theenzyme forms a covalent intermediate with the C-6 of the cytosine

ring, destroying the aromatic character of the base and activating theC-5 position to accept the methyl group from the cofactor AdoMet(24-26). ßelimination of the cysteine moieties of the MTase restores

the aromatic character of the base and completes the methylationprocess (Fig. 1, lower pathway). When AdoMet is present in concentrations below the Km value for the enzyme, the half-life of the

activated intermediate is increased, destabilizing the amino group inthe 4-position of the cytosine ring. We have shown previously (27)

that the bacterial MTase M.Hpall increases the hydrolytic deamination rate of C —¿�>U in vitro by a factor of IO4 over background under

these conditions. Because uracil codes for thymine during replication,unrepaired mismatches would be expected to result in C —¿�»T transi

tion mutations (Fig. 1, upper pathway). Subsequently, we (16) andothers (28, 29) have demonstrated the same mutational mechanism forthe MTases M.Hhal, M.Sssl, and M.EcoRII in vitro and in vivo,suggesting that the ability to deaminate the target cytosine is acommon feature of all MTases, thus offering an alternative pathwayfor creating the high rate of C —¿�»T transition mutations at methylation

sites found in human tumors.Low AdoMet concentrations might not only lead to an increase in

the numbers of mutations via this pathway but also reduce the overallDNA methylation level, because fewer methyl groups would beavailable for transmethylation reactions. Interestingly, as mentionedbefore, the content of 5mC in DNA isolated from various colontumors is significantly reduced compared with that in normal tissue (5,

' The abbreviations used are: 5mC, 5-methylcytosine; AdoMet. S-adenosylmethionine;

AdoHcy. S-adenosylhomocysteine; AdoEt. 5-adenosylethionine; MTase. cytosine (DNA-5)-methyltransferase; UDG. uracil-DNA glycosylase; GAPDH. glyceraldehyde-3-phos-phate dehydrogenase; SSCP. single-stranded conformational polymorphism; RT. reversetranscription; HPLC, high-performance liquid chromatography.

2375

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpG TRANSITION MUTATIONS

NormalB-.P"«*

ReactiveIntermediate

PremntagenlcLesion

M«**»

Uracll-DNA Glycosylaae

Fig. I. Proposed mechanisms for the formation of C —¿�»T transition

mutations at CpG sites induced by methyltransferase and by spontaneousdeamination of 5mC and its repair.

Methyltransferase 5mC:G

r.-G mismatch repair

30-32),4 although regions of hypermethylation of CpG islands have

been found (33, 34). In contrast, the overall decrease in 5mC contentcannot be explained by reduced MTase activity; in fact, levels ofMTase expression are increased in this tissue type (35, 36; see below),and high levels of MTase activity have also been found in various celllines (37). Further evidence supporting the hypothesis that alterationsin DNA methylation are causally related to colon cancer developmenthave been provided by several studies on the MTase and AdoMet forthe development of cancer in rodents. Laird et al. (38) found areduction in tumor incidence in min~ mice with reduced activity of

MTase, due to heterozygosity of the MTase gene. In addition, treatment of rats with the drug 5-azacytidine, which reduces MTaseactivity, further reduced the number of polyps detected in these mice.Moreover, overexpression of MTase leads to transformation of NIH/3T3 cells (39). In another approach, feeding rats folate- or methi-onine-deficient diets, precursors in the biosynthesis of AdoMet, led tolower concentrations of AdoMet in the liver, and the DNA washypomethylated within 1 week of treatment (40-42). In addition, therate of liver cancer has been found to be elevated under these conditions. Smith et al. (43) sequenced relevant exons in the p53 gene ofhepatomas induced by feeding Fischer 344 rats with a choline-defi-cient diet and found two C —¿�»T transitions at CpG sites of nine

mutations detected. It has been shown that diets low in methionine andfolate also increase the risk of colon cancer in humans (44), andinterestingly, many transformed cells are methionine dependent (45).These findings are consistent with a model that MTase increases therate of transition mutations at CpG sites in genes such as p53, whichare critical for cancer development (Fig. 1, upper pathway), and lowAdoMet concentrations could play a role in inducing a mutatorphenotype. Interestingly, these findings, except for the accumulationof CpG transition mutations, can also be explained by the epigeneticmodel of carcinogenesis, and both models are not exclusive.

The above studies all point to alterations in DNA methylationpatterns and control of the DNA methylation process as havingimportant potential roles in colon cancer development. Alterations inMTase expression levels, mutations in the MTase gene, and variationsof AdoMet concentrations may contribute not only to changes in DNAmethylation levels and, therefore, to alterations in gene expression,but also to enzymatic deamination of cytosine residues at mutationalhot spots (Fig. 1). Therefore, we examined these aspects of DNA

4 P. Cheng. C. Schmulte. K. F. Cofer, F. Felix. M. L. Yu. and L. Dubeau. Differential

role of DNA methylation in the development of benign and malignant ovarian tumors,submitted for publication.

methylation in colon cancer specimens and the adjacent normal co-Ionic epithelium.

MATERIALS AND METHODS

Colon Specimen. Human colon cancer tissue and surrounding normalmucosa were obtained from colon cancer patients undergoing colectomy.Normal mucosa was stripped from fat and muscle tissues. Samples were frozenin liquid nitrogen or ethanol and dry ice within 30 min after colectomy andkept at -70°C until use. Colon carcinomas were at least 1 cm in diameter and

of high grade and stage.Analysis of the Methylation Status ofp53 Codons 248,282, and 283. A

previously described PCR-based methylation assay was modified to determine

the methylation status at mutational hot spots of the p5J gene (11). GenomicDNA ( 100 ng) was digested for l h at 37°Cwith 10 units of C/oI, Hpall, Rsal,

or Mspl. Hpa\\ was used as a methylation-sensitive enzyme for the analysis of

the methylation status of codons 248 (CGC) and 282 (CGC), and C/01 wasused to determine whether codon 283 (CGC) was methylated. The digestedDNA (10 ng) was PCR amplified using primers that flanked the codons to bestudied. The primers for codon 248 were 5'-GTTGCCTCTGACTGTACCAC-CATC-3' and 5'-CAAGTGGCTCCTGAC-3'. The primers for codon 282 and283 were 5'-CCTATCCTGAGTAGTGGT-3' and 5'-TCGCTTAGTGCTC-CCTGG-3'. The PCR was performed at 94°Cfor 75 s, 62°Cfor 75 s, and 72°C

for 90 s for 26 cycles, which was in the linear range of the assay. The PCRproducts were run on 1.2% agarose gels, transferred to a Zeta probe membrane,and probed with the p53 cDNA (46).

Quantitative RT-PCR Analysis of MTase Expression. A quantitativeRT-PCR assay was used to measure the level of MTase mRNA. as described

previously (47). Total mRNA was extracted from normal and cancerous colonmucosa. mRNA was reverse transcribed using random hexamers and Molonymurine leukemia virus reverse transcriptase (Boehringer Mannheim). ThecDNA was then PCR amplified using primers for DNA MTase (5'-GTGC-GAGACACGATGTC-3' and 5'-TTGTCCAGGATGTTGCCG-.V) andGAPDH (S'-TGAGGCTGTTGTCATACTTCTC-S' and 5'-CAGCCGAGC-CACATCG-3'). The PCR conditions were 94°Cfor 45 s, 55°Cfor 30 s, and

72°for 90 s for 22 and 27 cycles for MTase and GAPDH, respectively. The

PCR conditions were optimized to assure that the reaction was in the linearphase of amplification. To assure consistency between experiments, a stock ofHeLa cell cDNA was serially diluted and run for each experiment.

Quantitative Analysis of AdoMet and AdoHcy Concentrations in ColonTissue. Colon tissue was homogenized in 2 ml 0.2 N perchloric acid, centri-

fuged for 30 min at full speed in an Eppendorf centrifuge, and subsequentlyfiltered through a 0.2-^.m syringe filter. All operations were performed on iceor at 4°C.AdoEt was added before tissue homogenization as an internal

standard. The crude extract was analyzed by ion-pair HPLC using a slightly

modified method of Wagner et al. (48). A Waters HPLC apparatus equippedwith a 440 absorbance detector and a 10-cm Brownlee RP-18 column guarded

by a Brownlee Aquaphore ODS precolumn was used. The columns were

2376

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpG TRANSITION MUTATIONS

adjusted to 40°C.AdoMet, AdoHcy, and AdoEt were separated by a linear

gradient, starting with 100% of eluent A [0.1 M NaH2PO4, 2% acetonitrile, and8 mM octadecanoylsulfonic acid (pH 2.65)] and leading in 40 min to eluent B[0.2 M NaH2PO4, 26% acetonitrile, and 8 mM octadecanoylsulfonic acid (pH3.25)]. The flow rate was set at 1 ml/min. Absorbance was monitored at 254nm. Under these conditions, the retention times of the standard substanceswere 18.1 min (AdoMet), 19.6 min (AdoHcy), and 21.5 min (AdoEt). AdoMetand AdoHcy were identified in tissue extracts on the basis of their retentiontimes and cochromatography with reference substances. The purity of thepeaks was verified by variation of the mobile phase pH and the concentrationof acetonitrile. Standard substances were obtained from Sigma Chemical Co.

Mutational Analysis of the Human Methyltransferase cDNA from Human Colon Tissues. cDNA from total RNA was prepared as described above.PCR was performed using primers for regions I-III (ml, 5'-CAAGCCTGT-GAGCCGAGCGAGC-3'; m2, 5'-CCAGCCATGACCAGCTTCAGCAG-

GATGTT-3'), regions IV-VI (m3, 5'-AAGGGAGACGTGGAGATGCTGT-GCG-3'; m4, 5'-GGAGCGCTTGAAGGAGACAAAGTTCCTGA-3'), and

regions IX and X (m5, 5'-CAGCACAACCGTCACCAACCC-3'; m6, 5'-CAACATACAAAGCTTGATCTCCAAGCCAATG-3'; see Fig. 5). PCR

products were analyzed for mutations using a SSCP technique and by directsequencing.

Determination of UDG Activity. We used a reporter plasmid (pSV2neoJ)

containing an ampicillin- and a neomycin-resistance gene. An inactivatingpoint mutation in the neomycin-resistance gene was introduced by site-directed

mutagenesis, changing GTGC to GCGC, as described previously (16). A U:Gmismatch at this site (GUGC) was created by incubating the plasmid withM.Hhal MTase (a generous gift of Dr. Sha Mi, Cold Spring Harbor Laboratory) for 16 h without AdoMet, which leads to deamination of cytosine touracil at the recognition site (16). The plasmid was purified by phenol-

chloroform extraction and ethanol precipitated, and 100 ng were incubatedwith various concentrations of extract from human normal and cancerous colontissue (23) in 50 mM Tris-HCl (pH 7.5), 10 mM EDTA, 2 /xg BSA, and 5 ¿IM

DTT in a total volume of 10 /il. Purified UDG (Boehringer Mannheim) wasalso serially diluted and added to the experiment as a control. Subsequently,the plasmid was isolated by phenol-chloroform extraction and ethanol precipitation and electrotransformed into Escherichia coli NR8052 (^[pro-lac], thi,

ara, trpE9777, ungi), which are devoid of UDG activity. After incubation on

SOB plates containing either ampicillin to score the transformation efficiencyor kanamycin, the reversion frequency was determined by dividing the numberof kanamycin-resistant colonies by the number of colonies on the ampicillin

plates. The efficiency of uracil repair was scored by comparing the reversionfrequencies of bacteria transformed with the plasmid, which was incubatedwith colon tissue extracts, with the reversion frequencies obtained from bacteria transformed with plasmids incubated without extracts.

RESULTS

Analysis of the Methylation Status of Mutational Hot Spots inthe p53 Gene. Previous studies in this (9, 11) and other laboratories(12) have shown that all CpG sites in the p53 gene are methylated inhuman WBC, kidney, muscle, bladder, and sperm DNA. We firstconfirmed that the CpG dinucleotides at codons 248, 282, and 283,which are mutational hot spots in colorectal cancer (49), were methylated in normal colonie mucosa, adenomas, and carcinomas. Fig. 2shows a typical analysis with primers flanking codons 248 (Fig. 2A)and 283 (Fig. 25). Methylation analysis of codon 248 showed strongsignals in template DNA cut with Cfol, which cuts outside of thesequence analyzed. No signals were generated when the templateDNA was precut with Mspl, which is insensitive to methylation of theinternal cytosine of the CCGG recognition sequence. On the otherhand, the CCGG sequences were methylated in all tissues, becausestrong signals were generated after digestion with Hpall. Similarresults were obtained for codons 282 and 283, which are substrates forthe methylation-sensitive enzymes Hpall and Cfol, respectively (Fig.

25). In this case, Rsal, which cuts outside of the sequence analyzed,was used as a control. Therefore, all of the mutational hot spots

A. Normal Adenoma Carcinoma

uo re CL. n ro£ O. (o £ o.O I 2 O I

B.Normal Adenoma Carcinoma

Fig. 2. Methylation analysis of codons 248. 282, and 283 of the p53 tumor suppressorgene in human colon tissues. DNA was isolated and incubaled with restriction enzymesfollowing PCR. using primers that flank codons 248 or 282/283. PCR products wereseparated by agarose gel electrophoresis. A, codon 248 (CGG): Cfol cuts outside of thesequence analyzed: Hpall cuts only unmethylated CCGG; and Mspl is insensitive formethylation at the CpG site. B, codons 282 (CGG) and 283 (CGC): Rsal cuts outside ofthe sequence analyzed: and Hpall and Cfol cut only if their recognition sites areunmethylated.

examined were methylated in the target colon tissue, showing thatthese sites are targeted by MTase.

Analysis of Methyltransferase Expression Levels. We showedearlier that the levels of active MTase enzyme are often markedlyelevated in transformed cell lines relative to their nononcogeniccounterparts (37). Also, El-Deiry et al. (35) found considerable over-

expression of the MTase mRNA in colorectal tumors, a situation thatmight lead to an enhanced rate of MTase-induced cytosine deamina

tion. According to our findings in the in vitro studies with the bacterialMTases M.Hpall and M.Hhal (16, 27), deamination of cytosine at thetarget site is dependent on the concentration of MTase and thecofactor AdoMet. Deamination was only measurable when AdoMetconcentrations dropped below its Km. To test whether this mechanismmight contribute to mutations at methylation sites in the human p53gene, we first measured MTase mRNA expression and AdoMet concentrations in colon carcinoma tissues and normal mucosa. RT-PCR

analysis using primers specific for region VIII of MTase (see Fig. 5for diagram) and primers specific for GAPDH as a control revealedthat MTase expression was increased in six of seven tumors comparedwith the normal epithelium from the same patient (Fig. 3). However,increases were generally modest and far less than the 60-200-folddifferences reported by El-Deiry et al. (35) but were similar to

increases of MTase activity levels reported by Issa et al. (36). Becausetumor tissue presumably contains a greater percentage of dividingcells than normal tissue, it remains to be seen whether increases inMTase expression are due to an increase in the proportion of dividingcells or due to an acute increase in MTase expression per individualcell.

Analysis of AdoMet Concentrations. Ion-pair HPLC was used to

measure concentrations of AdoMet and the breakdown productAdoHcy in extracts of mucosal and carcinoma tissue samples. Toprevent breakdown of AdoMet, an acidic total cell extract was prepared, filtered, and injected directly into the HPLC. AdoEt, which haschemical stability similar to AdoMet under these conditions, was used

2377

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpG TRANSITION MUTATIONS

HeLa cDNA

Patient

B C

o°g88NTNTNTNT

TUMOR TISSUE

NORMAL MUCOSA

A B C D E F G

FACTORT/N 1.9 0.7

Fig. 3. Expression of methyltransferase mRNA in human colon tissue. Total RNA wasisolated from human colon tissue (N. normal: T, tumor) and HeLa cells. RT-PCR wasperformed with primers specific for methyltransferase (MTase) and GAPDH. respectively,and products were separated by agarose gel electrophoresis (A). Results were summarizedas the ratio of MTase:GAPDH (ß).

from these data that AdoMet is not limiting for methylating DNA byMTase in this late stage of tumor development.

Mutational Analysis of Human Methyltransferase cDNA. Because we found adequate concentrations of AdoMet for DNA meth-

ylation in the tumor tissue, we hypothesized that a mutation in theAdoMet-binding pocket of MTase might create an enzyme that po

tentially inhibits efficient AdoMet binding. This could prevent efficient cytosine methylation without reducing the DNA-binding abilityand thus increasing the frequency of C —¿�»T transitions at the target

site. Indeed, we have recently shown that point mutations in the highlyconserved domain I of the bacterial MTase M.H/xdl reduce the abilityof the enzymes to methylate DNA and also increase the deaminationrate at the target C dramatically in vitro and in vivo (50).

Therefore, we analyzed these highly conserved regions of theMTase cDNA to test whether a similar mutation in domains codingfor the AdoMet-binding pocket of the human MTase, which might

lead to a mutator phenotype. could be involved in the mutationalprocess in the colon tumors we analyzed (51, 52). RT-PCR analysis

was performed using primers flanking the highly conserved regionsI-III, IV-VI, and IX and X of the MTase gene (Fig. 5), which are

involved in AdoMet binding and catalysis. Portions of the MTase genewere screened for mutations by SSCP. Analysis of 10 tumor samplesshowed no mutations by SSCP. To confirm these results, PCR products amplified from these conserved regions using cDNA derivedfrom seven colon carcinoma samples as a template were directlysequenced. No mutations or polymorphisms were detected (data notshown). Although the number of samples analyzed was small, the datashowed no evidence for a mutated MTase, which might confer amutator phenotype in human colon cancer.

U:G Mismatch Repair. MTase-induced deamination of cytosine

leads to a premutagenic U:G mismatch. Repair of this mismatch isinitiated by UDG, which excises uracil, leaving an a basic site. Base

as an internal control. The concentrations of both substances werehigher in the tumors than in normal tissue, and the concentrations ofAdoMet were 8.8 ±4.4 nmol/g wet tissue in samples from normalmucosa and 30.6 ±12.3 nmol/g wet tissue in tumor tissue (Fig. 4).Average AdoHcy concentrations were 4.2 ±2.0 nmol/g wet tissue innormal mucosa and 10.5 ±7.3 nmol/g wet tissue in tumor tissue. Theratio of AdoMet:AdoHcy was not significantly different comparingnormal (2.1 ± 2.2) and tumor (2.9 ± 1.7) tissue. Extracts frommucosal tissue from two patients without cancer showed AdoMet andAdoHcy concentrations in the same range compared with values innormal tissue from cancer patients (Fig. 4, / and 2). We conclude

LDVESGCG

EMWD

ENVRNFVS

LCGGPPCQGFS

QVGNAVPPPL\

RWSVRECARSQGFPDTY

Fig. 5. Conserved regions in the catalytic domain of the human MTase cDNA. Thecatalytic domain starting at an amino acid position of approximately 900 contains 10conserved regions (rectangles). Amino acids thai are involved in AdoMet binding in thebacterial MTase M.Hhal (26) by forming hydrogen bonds to this compound are underlined. The catalytic cysteine, which forms a covalent bond to the 5-position of the targetC, is bold. Primers used for RT-PCR were rl and r2; primers for mutational analysis ofthe MTase cDNA were ml-m6.

Fig. 4. Concentration of AdoMet and AdoHcy in human colontissue samples. Extracts from human colon carcinomas and surrounding mucosa were analyzed for AdoMet and AdoHcy usingHPLC. (A-G. tissue from cancer patients: / and 2, tissue frompatients without cancer).

TUMOR TISSUE

NORMAL MUCOSA

2378

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpO TRANSITION MUTATIONS

100 -

80 -

£ 60 -

40 -J

20 -

0.0001 0.001 0.01

HgProteinFig. 6. G:U mismatch repair activity in human colon tissue. pSV2«ecí'plasmili with

one U:G mismatch in a kanamycin resistance gene was incubated with human colon tissueextract (A-C, tissue extract from three different patients). Unrepaired, the plasmid is ableto confer kanamycin resistance to E. coli deficient in UDG. Repair activity was scored bycomparing the reversion frequency of bacteria transformed with the plasmid. which wasincubated with the colon extract, to bacteria transformed with the untreated plasmid (see"Materials and Methods").

excision is the rate-limiting step in this repair pathway. Subsequently,

several enzymes, including DNA polymerase ßand DNA ligases, areinvolved in restoring the C:G bp (53). We used a genetic assay basedon the reversion to neomycin resistance to determine the activity ofUDG in colon tissue samples. Site-directed mutagenesis was used to

create a pSV2««/plasmid, in which the sequence GTGC was converted to GCGC, inactivating the neomycin resistance conferred bythis plasmid and creating a target for M.Hhal. A U:G mismatch wasformed at this site by incubation of the plasmid with Hluil MTase inthe absence of AdoMet (16, 27). This plasmid was then incubatedwith different concentrations of extracts from normal and canceroustissue and subsequently transfected into bacteria deficient in UDG tomeasure U:G mismatch repair efficiencies in human colon tissues.The results for three different tumor samples and adjacent normalmucosa from the same patient showed that repair of U:G mismatcheswere very efficient, and no significant difference between normal andcancerous tissue was found (Fig. 6). Interestingly, we found largeinterindividual variations in the efficiency of uracil excision. Thisobservation has been described previously for human colon and smallintestine tissue samples (54). These data, together with our recentresults using an alternative assay (23), indicate that UDG activity isnot decreased in colon tumors.

DISCUSSION

The mechanism by which DNA methylation contributes to thedevelopment of human tumors is currently a subject of intense interest(55). In particular, the recent observations that widespread methylation changes occur during tumor development (30-32, 34, 56).4 and

that mice deficient in MTase activity have a decreased rate of polypformation (38), have stimulated interest in this field. Hypomethylationof proto-oncogenes (57-59) and hypermethylation of tumor suppres

sor genes (47, 60, 61) can alter their expression levels, providing anepigenetic model for tumor development (8, 62). On the other hand,methylation of cytosine to 5mC leads to a higher rate of C —¿�»T

mutations compared with that in unmethylated sites (9, 63), thuscontributing to a genetic pathway of carcinogenesis.

In this study, we focused on mechanisms for the generation of C —¿�»

T mutations and measured parameters that might favor the inductionof these mutations during the DNA methylation process by humanMTase (Fig. 1). This pathway, which has been validated for severalbacterial MTases (16, 27-29), is particularly attractive, because it is

specific for methylation sites and may be triggered by a diet low inmethyl group-providing substances. The mechanism is also consistent

with the observation that inactivation of MTase and reduction ofmethylation reduces the incidence of polyp formation in a mousemodel system (38), and that levels of MTase expression and MTaseactivity have been reported to be elevated in colorectal tumors (35,36). Analysis of normal and transformed colonie tissues confirmedthat sites that are hot spots for C —¿�»T mutations in the p53 tumor

suppressor gene are methylated and were, therefore, targeted byMTase. RT-PCR analysis revealed that the average increase of MTaseexpression was 3.7-fold in colon carcinomas compared with normal

adjacent mucosa (Fig. 3). These findings are much lower than whathas been reported earlier by El-Diery et al. (35). This group found that

values for MTase mRNA expression ranged up to several hundredfold higher in human colon cancer tissue than in tissue samples frompatients with benign disorders. Our data are similar to the 5.4-fold

higher MTase enzyme activity detected by Issa el al. (36) in coloncancers compared with normal mucosa. In hepatomas of LEG rats,MTase mRNA expression was also only 2-fold higher compared withnormal liver, and MTase activity was increased approximately 3-fold

(64).The decisive parameter that determines whether the MTase methy-

lates its target cytosine or induces deamination (Fig. 1, upper andlower pathwa\s) is the concentration of AdoMet available to theenzyme during the methylation process. In our analysis, AdoMetconcentrations were increased in some tumors or not significantlychanged in carcinoma tissue compared with normal mucosa from thesame patient or from patients without cancer. Assuming that theintracellular water content is in the range of 70-85% (65, 66), the

average intracellular concentration of AdoMet in normal mucosa canbe calculated to be approximately 11 ^.M, which is similar to thatwhich has been found in rat colon (17 nmol/mg protein; Ref. 67), andaverage concentrations in the carcinoma tissue were approximately 38/J.M(Fig. 4). The Km of human MTase from HeLa cells for AdoMethas been determined to be 3.25 /UM,and Km values of other eukaryoticMTases are in the same range (rat, 0.5-6.6 /J.M;Refs. 68 and 69; rice,2.6-6 /UM;Ref. 70; and wheat, 5-6 /XM;Ref. 71). Although this isapproximately 10-100-fold higher than the Km of bacterial MTases

for AdoMet (24, 72), it is unlikely that extensive induction of cytosinedeamination by MTase is possible under these high concentrations ofAdoMet. In addition, we found that the concentration of AdoHcy thathas been shown to inhibit the deamination reaction (27) was higher inthe tumor tissue as well. It should be noted that a direct comparisonof concentration data obtained from tumor and mucosal tissue isdifficult, because mucosal tissue is heterogeneous, and rapidly proliferating cells are located only in the crypt base (73). Therefore, cautionshould be used in using MTase expression levels measured for theentire mucosa as indicators for MTase levels in colon cells at the basesof the crypts. Thus, we cannot exclude that cytosine deaminationoccurs as a side reaction at very low efficiency and might contributeto the burden of DNA repair. Moreover, it has been shown that a dietlow in methionine, which is a direct precursor in the biosynthesis ofAdoMet, and low in other methyl group-providing factors can reduceintracellular AdoMet levels and also reduce DNA methylation (40-

42). Therefore, it is possible that special conditions in human tissuemight result in intracellular concentrations of AdoMet below the Kmof MTase, creating conditions favorable for the induction of cytosinedeamination (Fig. 1, upper pathway). Because it is not possible to

2379

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpG TRANSITION MUTATIONS

follow fluctuations of AdoMet or AdoHcy concentrations over time intissue during the initial or later stages of tumor development, atransient shortage of AdoMet might enable the MTase to inducedeaminations at its target cytosine in colon tissue. However, theconcentrations of AdoMet we measured in the normal colon or tumortissue exclude the possibility that a continuous shortage of AdoMetcaused by a defect in AdoMet synthesis exists in these cells. Inaddition, a drop of AdoMet concentration by a factor of approximately4 to concentrations below the Km for MTase would be expected toaffect all enzymes using this cofactor for transmethylation reactionsand, thus, would be more cytotoxic than mutagenic for a cell.

Recently, we have shown (50) that bacterial MTases carryingmutations in the AdoMet-binding pocket, which presumably preventsefficient binding of the cofactor, can induce C —¿�>U deaminations in

vitro and in vivo (Fig. 1, upper pathway) at normal concentrations ofAdoMet, even in the presence of a functional UDG repair system.Therefore, we screened tumors for the presence of similar mutationsin the AdoMet-binding pocket of the human MTase gene. The fact that

such mutations were not observed in the colon tumors argues againstthis mechanism being of importance in colon carcinogenesis.

An additional factor determining whether a MTase-induced deami-nation reaction contributes to the generation of C —¿�»Tmutations is the

ability of cells to repair U:G mismatches generated by this process.Therefore, we measured the ability of colon extracts to repair thesemismatches. Excision of uracil was efficient in all samples tested (Fig.6; Ref. 23). Moreover, UDG expression has been shown to be maximal in S-phase (74, 75), when MTase activity peaks. Although it is

possible that MTase forms a tight complex with the U:G mismatchformed by the deamination reaction, or that a small fraction of uracilgets methylated by MTase, thus escaping repair by UDG (16, 76), thehigh efficiency of UDG repair found in the human colon extractsmakes it unlikely that a U:G mismatch persists in DNA until replication completes the C —¿�»T mutation.

None of the parameters tested suggest that biochemical conditionsfavor a MTase-induced pathway for the formation of C —¿�»U —¿�>T

transitions in human tissue (Fig. 1, upper pathway). Spontaneousdeamination of 5mC, therefore, is the most likely mechanism bywhich the majority of these mutations are formed (Fig. 1, lowerpathway). Nevertheless, it is difficult to exclude the possibility thatlow-efficiency side reactions of enzymatic methylation do not con

tribute to the generation of these mutations. The mechanisms bywhich methyl-deficient diets contribute to carcinogenesis and the

reason for the decreased number of polyps formed in transgenic micewith lower MTase levels (38) may, therefore, be best explained by anepigenetic pathway.

REFERENCES

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

29.

30.

31.

32.

33.1. Bird. A. P. CpG-rich islands and the function of DNA methylation. Nature (Lond.),

321: 209-213, 1986.2. Gardiner-Garden. M.. and Frommer. M. CpG islands in vertebrate genomes. J. Mol. 34

Biol.. 196: 261-282. 1987.3. Riggs. A. D.. and Pfeifer. G. P. X-chromosome inactivation and cell memory. Trends

Genet.. 8: 169-174, 1992.

4. Bartolomei. M. S.. Zenel. S., and Tilghman, S. M. Parental imprinting of the mouseH19 gene. Nature (Lond.), 365: 764-767, 1991.

5. Gama-Sosa, M. A., Slagel, V. A., Trewyn, R. W., Oxenhandler, R., Kuo, K. C.Gehrke. C. W., and Ehrlich, M. The 5-methylcytosine content of DNA from humantumors. Nucleic Acids Res.. //: 6883-6894. 1983.

6. Baylin. S. B.. Hoppener, J. W., de Bustros, A., Steenbergh. P. H.. Lips, C. J.. andNelkin. B. D. DNA methylation patterns of the calcitonin gene in human lung cancersand lymphomas. Cancer Res.. 46: 2917-2922. 1986.

7. Baylin, S. B., Makos, M., Wu. J. J., Yen. R. W., de Bustros, A., Venino, P.. andNclkin. B. D. Abnormal patterns of DNA melhylation in human neoplasia: potentialconsequences for tumor progression. Cancer Cells (Cold Spring Harbor), 3: 383-390,1991.

8. Jones. P. A., and Buckley. J. D. The role of DNA methylation in cancer. Adv. CancerRes., 54: 1-23, 1990.

35.

36.

37.

38.

39.

Rideout. W. M. I.. Coetzee. G. A.. Olumi. A. F.. and Jones. P. A. 5-Methylcytosineas an endogenous mutagen in the human LDL receptor and p53 genes. Science(Washington DC). 249: 1288-1290. 1990.

Greenblatt. M. S.. Bennett, W. P.. Hollstein. M., and Harris. C. C. Mutations in thep53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis.Cancer Res.. 54: 4855-4878. 1994.Magewu. A. N.. and Jones. P. A. Ubiquitous and tenacious methylation of the CpGsite in codon-248 of the pSJ gene may explain its frequent appearance as a mutationalhot-spot in human cancer. Mol. Cell. Biol.. 14: 4225-4232. 1994.Tornaletti. S.. and Pfeifer. G. P. Complete and tissue-independent methylation of CpGsites in the p53 gene: implications for mutations in human cancers. Oncogene. IO:1493-1499. 1995.

Bottema. C. D.. Ketterling. R. P.. Vielhaber. E.. Yoon. H. S.. Gostout. B., Jacobson.D. P.. Shapiro. A., and Sommer. S.S. The pattern of spontaneous germ-line mutation:relative rales of mutation at or near CpG dinucleotides in the factor IX gene. Hum.Genet.. 91: 496-503, 1993.

Schorderet. D. F.. and Gartier. S. M. Analysis of CpG suppression in methylated andnonmethylated species. Proc. Nati. Acad. Sci. USA. 89: 957-961, 1992.Sved, J.. and Bird, A. The expected equilibrium of the CpG dinucleotide in vertebrategenome under a mutation model. Proc. Nati. Acad. Sci. USA. 87: 4692-4696. 1990.Yang. A. S., Shen, J. C.. Zingg. J. M., Mi. S.. and Jones. P. A. Hhal and Hpall DNAmethyltransferases bind DNA mismatches, methylate uracil and block DNA repair.Nucleic Acids Res.. 23: 1380-1387. 1995.

Spruck. C. H.. Rideout. W. M.. and Jones. P. A. DNA methylation and cancer. In: 1.P. Jost and H. P. Saluz (eds.), DNA Methylation: Molecular Biology and BiologicalSignificance, pp. 487-509. Basel: BirkhäuserVerlag. 1993.

Coulondre. C., Miller, J. H., Farabaugh. P. J., and Gilbert. W. Molecular basis of basesubstitution hotspots in Escherichia coli. Nature (Lond.). 274: 775-780, 1978.Wang, R. Y.. Kuo. K. C.. Gehrke. C. W.. Huang. L. H.. and Ehrlich. M. Heat- andalkali-induced deamination of 5-methylcytosine and cytosine residues in DNA. Bio-chim. Biophys. Acta. 697: 371-377. 1982.

Shen. J. C.. Rideout. W. M., and Jones. P. A. The rate of hydrolytic deamination of5-methylcytosine in double-stranded DNA. Nucleic Acids Res.. 22: 972-976. 1994.

Lindahl. T. Instability and decay of the primary structure of DNA. Nature (Lond.).362: 709-715. 1993.Shenoy, S.. Ehrlich. K. C.. and Ehrlich. M. Repair of thymine.guanine and uracil.-guanine mismatched base-pairs in bacteriophage M13mpl8 DNA heteroduplexes, J.Mol. Biol.. ¡97:617-626, 1987.

Schmutte. C.. Yang. A. S.. Bean. R. W.. and Jones. P. A. Base excision repair of U:Gmismatches at a mutational hotspot in the ¡i53gene is more efficient than baseexcision repair of T:G mismatches in extracts of human colon tumors. Cancer Res..55: 3742-3746, 1995.Wu, J. C., and Santi, D. V. Kinetic and catalytic mechanism of Hhal methyltrans-ferase. J. Biol. Chem.. 262: 4778-4786. 1987.

Chen. L.. MacMillan. A. M.. Chang. W.. Ezaz, N. K.. Lane, W. S.. and Verdine, G. L.Direct identification of the active-site nucleophile in a DNA (cytosine-5)-methyl-transferase. Biochemistry, 30: 11018-11025, 1991.

Klimasauskas. S.. Kumar. S., Roberts, R. J.. and Cheng. X. Hhal methyltransferaseflips its target base out of the DNA helix. Cell. 76: 357-369, 1994.Shen. J. C.. Rideout. W. M.. and Jones. P. A. High-frequency mutagenesis by a DNAmethyltransferase. Cell. 71: 1073-1080. 1992.

Wyszynski. M.. Gabbara, S., and Bhagwat, A. S. Cytosine deaminations catalyzed byDNA cytosine methyltransferases are unlikely to be the major cause of mulalional hotspots at sites of cytosine methylation in Escherichia cali. Proc. Nati. Acad. Sci. USA,91: 1574-1578. 1994.

Bandaru. B.. Wyszynski. M.. and Bhagwat. A. S. Hpall methyltransferase is mutagenic in Escherichia coli. J. Bacteriol.. 177: 2950-2952. 1995.Goelz. S. E.. Vogelstein, B.. Hamilton. S. R., and Feinberg, A. P. Hypomethylationof DNA from benign and malignant human colon neoplasms. Science (WashingtonDC), 22«:187-190, 1985.

Feinberg. A. P.. Gehrke, C. W.. Kuo. K. C.. and Ehrlich. M. Reduced genomic5-methylcytosine content in human colonie neoplasia. Cancer Res.. 48: 1159-1161.

1988.Diala. E. S.. Cheah. M. S., Rowith, D., and Hoffman. R. M. Extent of DNAmethylation in human tumor cells. J. Nati. Cancer Inst., 71: 755-764, 1983.

de Bustros. A.. Nelkin, B. D.. Silverman. A.. Ehrlich. G.. Poiesz. B.. and Baylin. S. B.The short arm of chromosome 11 is a "hot spot" for hypermethylation in human

neoplasia. Proc. Nati. Acad. Sci. USA, «5:5693-5697. 1988.

Makos. M.. Nelkin, B. D.. Lerman. M. I.. Latif, F., Zbar. B., and Baylin, S. B. Distincthypermethylation patterns occur at altered chromosome loci in human lung and coloncancer. Proc. Nati. Acad. Sci. USA, 89: 1929-1933. 1992.El-Deiry, W. S., Nelkin. B. D.. Celano. P.. Yen, R. W.. Falco. J. P.. Hamilton, S. R.,

and Baylin, S. B. High expression of the DNA melhyltransferase gene characterizeshuman neoplastic cells and progression stages of colon cancer. Proc. Nati. Acad. Sci.USA. 88: 3470-3474. 1991.Issa, J. P., Venino. P. M.. Wu, J.. Sazawal, S., Celano, P.. Nelkin. B. D.. Hamilton.S. R., and Baylin, S. B. Increased cytosine DNA-methyltransferase activity duringcolon cancer progression. J. Nat!. Cancer Inst., 85: 1235-1240, 1993.

Kautiainen. T. L., and Jones. P. A. DNA methyltransferase levels in tumorigenic andnontumorigenic cells in culture. J. Biol. Chem.. 26/: 1594-1598. 1986.

Laird. P. W., Jacksongrusby. L., Fazeli, A.. Dickinson. S. L.. Jung, W. E.. Li, E..Weinberg. R. A., and Jaenisch. R. Suppression of intestinal neoplasia by DNAhypomethylation. Cell, 81: 197-205. 1995.Wu. J.. Issa. J. P.. Herman, J.. Bassett. D. J.. Nelkin. B. D.. and Baylin, S. B.Expression of an exogenous eukaryotic DNA methyltransferase gene induces transformation of NIH 3T3 cells. Proc. Nati. Acad. Sci. USA. 90: 8891-8895, 1993.

2380

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

DNA METHYLATION AND CpG TRANSITION MUTATIONS

40. Wainfan, E., and Poirier. L. A. Methyl groups in carcinogenesis: effects on DNAmethylation and gene expression. Cancer Res., 52 (Suppl.).- 2071s-2077s. 1992.

41. Christman. J. K., Sheikhnejad. G.. Dizik, M.. Abileah, S.. and Wainfan. E. Reversibility of changes in nucleic acid methylation and gene expression induced in rat liverby severe dietary methyl deficiency. Carcinogenesis (Lond.), 14: 551-557, 1993.

42. Pogribny. I. P., Basnakian, A. G., Miller, B. J.. Lopalina. N. G., Poirier. L. A., andJames, S. J. Breaks in genomic DNA and within the p53 gene are associated withhypomethylation in livers of folate/methyl-deficient rats. Cancer Res.. 55: 1894-

1901. 1995.43. Smith, M. L.. Yeleswarapu, L.. Scalamogna, P.. Locker, J., and Lombardi, B. p53

mutations in hepatocellular carcinomas induced by a choline-devoid diet in maleFischer 344 rats. Carcinogenesis (Lond.), 14: 503-510, 1993.

44. Giovannucci, E., Rimm, E. B., Ascherio, A., Stampfer. M. J., Colditz, G. A., andWillett, W. C. Alcohol, low-methionine-low-folate diets, and risk of colon cancer inmen. J. Nati. Cancer Inst.. 87: 265-273, 1995.

45. Coalson. D. W.. Mechara. J. O., Stern, P. H.. and Hoffman. R. M. Reduced availability of endogenously synthesized methionine for S-adenosylmethionine formationin melhionine-dependent cancer cells. Proc. Nati. Acad. Sci. USA, 79: 4248-4251,

1982.46. Feinberg. A. P.. and Vogelstein. B. A technique for radiolabelling DNA restriction

endonuclease fragments to high specific activity. Anal. Biochem., 132: 6-13, 1983.47. Gonzalez-Zulueta, M., Bender. C. M.. Yang, A. S., Nguyen, T. T.. Bear!, R. W., Van

Tornout. J. M.. and Jones, P. A. Methylation of the 5' CpG island of the pl6/CDKN2

tumor suppressor gene in normal and transformed human tissues correlates with genesilencing. Cancer Res.. 55.- 4531-4535, 1995.

48. Wagner, J., Claverie, N., and Danzin, C. A rapid high-performance liquid Chromato

graphie procedure for the simultaneous determination of methionine, ethionine,S-adenosylmethionine, S-adenosylethionine, and the natural polyamines in rat tissues.Anal. Biochem.. 140: 108-116, 1984.

49. Harris, C. C., and Hollstein, M. Clinical implications of the p53 tumor-suppressorgene. N. Engl. J. Med., 329: 1318-1327, 1993.

50. Shen. J. C., Zingg, J. M., Yang, A. S., Schmutte. C.. and Jones. P. A. A mutant DNA(cytosine-5)-methyltransferase functions as a mutator enzyme. Nucleic Acids Res.,32: 4275-4282, 1995.

51. Kumar, S., Cheng, X., Klimasauskas, S., Mi, S., Posfai, J., Roberts, R. J., and Wilson,G. G. The DNA (cytosine-5) methyltransferases. Nucleic Acids Res., 22: 1-10, 1994.

52. Yen, R. W., Venino, P. M., Nelkin, B. D., Yu, J. J., El-Deiry. W., Cumaraswamy, A.,Lennon. G. G.. Trask, B. J., Celano, P., and Baylin. S. B. Isolation and characterization of the cDNA encoding human DNA methyltransferase. Nucleic Acids Res., 20:2287-2291, 1992.

53. Singhal, R. K., Prasad, R., and Wilson, S. H. DNA polymerase ßconducts thegap-filling step in uracil-initiated base excision repair in a bovine testis nuclearextract. J. Biol. Chem., 270: 949-957, 1995.

54. Myrnes, B., Giercksky, K. E., and Krokan, H. Intel-individual variation in the activityof O6-methylguanine-DNA methyltransferase and uracil-DNA glycosylase. Carcino

genesis (Lond.), 4: 1565-1568, 1983.55. Balmain, A. Exploring the bowels of DNA methylation. Curr. Biol.. 5: 1013-1016.

1995.56. Feinberg. A. P.. and Vogelstein, B. Hypomethylation distinguishes genes of some

human cancers from their normal counterparts. Nature (Lond.), 301: 89-92, 1983.57. Munzel, P. A., Pfohl, L. A., Rohrdanz, E., Keith, G., Dirheimer, G., and Bock, K. W.

Site-specific hypomethylation of c-mvc protooncogene in liver nodules and inhibitionof DNA methylation by ¿V-nitrosomorpholine. Biochem. Pharmacol., 42: 365-371,1991.

58. Sharrard, R. M., Royds. J. A., Rogers, S., and Shorthouse, A. J. Patterns of methy

lation of the c-mvc gene in human colorectal cancer progression. Br. J. Cancer., 65:667-672, 1992. '

59. Bhave, M. R., Wilson, M. J.. and Poirier, L. A. c-H-ras and c-K-ra.v gene hypomethylation in the livers and hepatomas of rats fed methyl-deficient, amino acid-defined diets. Carcinogenesis (Lond.), 9: 343-348. 1988.

60. Merlo. A.. Herman, J. G., Mao, L.. Lee, D. J., Gabrielson, E.. Burger. P., Baylin.S. B., and Sidransky, D. 5' CpG island methylation is associated with transcriptional

silencing of the tumour suppressor pl6/CDKN2/MTSl in human cancers. Nat. Med.,/: 686-692. 1995.

61. Herman, J. G., Merlo. A.. Lapidus. R. G.. Issa, J-P. J.. Davidson. N. E.. Sidransky, D.,

and Baylin, S. B. Inactivation of the CDKN2/pl6/MTSI gene is frequently associatedwith aberrant DNA methylation in all common human cancers. Cancer Res., 55:4525-4530, 1995.

62. Greger, V.. Passarge, E.. Hopping, W., Messmer. E., and Horsthemke. B. Epigeneticchanges may contribute to the formation and spontaneous regression of retinoblas-toma. Hum. Genet., 83: 155-158, 1989.

63. Duncan, B. K.. and Miller, J. H. Mutagenic deamination of cytosine residues in DNA.Nature (Lond.). 287: 560-561. 1980.

64. Miyoshi. E., Jain, S. K., Sugiyama, T., Fujii, J.. Hayashi, N., Fusamoto. H.. Kamada,T., and Tanaguchi, N. Expression of DNA methyltransferase in LEC rat duringhepatocarcinogenesis. Carcinogenesis (Lond.), 14: 603-605, 1993.

65. Wheatley, D. N., Inglis, M. S., Foster, M. A., and Rimington, J. E. Hydration, volumechanges and nuclear magnetic resonance proton relaxation times of HeLa S-3 cells inM-phase and the subsequent cell cycle. J. Cell Sci., 88: 13-23, 1987.

66. Hoffman, D., Kumar, A. M.. Spitzer. A., and Gupta. R. K. NMR measurement ofintracellular water volume in rat kidney proximal tubules. Biochim. Biophys. Acta,889: 355-360, 1986.

67. Halline, A. G., Dudeja, P. K.. and Brasitus, T. A. 1,2-Dimethylhydrazine-inducedpremalignant alterations in the S-adenosylmethionine/S-adenosylhomocysteine ratio

and membrane lipid lateral diffusion of the rat distal colon. Biochim. Biophys. Acta.944: 101-107, 1988.

68. Arnaud, M., Dante, R., and Niveleau, A. DNA methyltransferases in normal and aviansarcoma virus-transformed rat cells. Quantitation of 5-methyldeoxycytidine in DNAand enzyme kinetics study. Biochim. Biophys. Acta, 826: 108-112, 1985.

69. Ruchirawat, M., Becker, F. F., and Lapeyre, J. N. Indistinguishable physical andcatalytic properties of DNA methyltransferase from normal rat liver and a transplant-able rat hepatocellular carcinoma. Carcinogenesis (Lond.), 6: 877-882, 1985.

70. Giordano. M.. Mattachini. M. E.. Cella. R.. and Pedrali, N. G. Purification andproperties of a novel DNA methyltransferase from cultured rice cells. Biochem.Biophys. Res. Commun., 777: 711-719, 1991.

71. Theiss, G., Schleicher, R., Schimpff, W. G.. and Follmann. H. DNA methylation inwheat. Purification and properties of DNA methyltransferase. Eur. J. Biochem.. /67:89-96. 1987.

72. Rubin. R. A., and Modrich, P. EcoRI methylase. J. Biol. Chem.. 252: 7265-7272,

1977.73. O'Brien. M. J., O'Keane. J. C., Zauber. A., Gottlieb. L. S.. and Winawer. S. J.

Precursors of colorectal carcinoma. Biopsy and biologic markers. Cancer (Phila.). 70:1317-1327. 1992.

74. Slupphaug, G., Olsen, L. C., Heiland. D.. Aasland. R., and Krokan, H. E. Cell cycleregulation and in vitro hybrid arrest analysis of the major human uracil-DNAglycosylase. Nucleic Acids. Res.. 19: 5131-5137, 1991.

75. Weng, Y., and Sirover. M. A. Developmental regulation of the base excision repairenzyme uracil DNA glycosylase in the rat. Mutât.Res., 293: 133-141, 1993.

76. Klimasauskas, S., and Roberts, R. J. M. Hhal binds tightly to substrates containingmismatches at the target base. Nucleic Acids Res.. 23: 1388-1395. 1995.

2381

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

1996;56:2375-2381. Cancer Res   Christoph Schmutte, Allen S. Yang, TuDung T. Nguyen, et al.   CarcinogenesisMechanisms for the Involvement of DNA Methylation in Colon

  Updated version

  http://cancerres.aacrjournals.org/content/56/10/2375

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/56/10/2375To request permission to re-use all or part of this article, use this link

Research. on January 31, 2018. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from