foreign dna integration—perturbations of the genome—oncogenesis

276

Foreign DNA Integration—Perturbations ofthe Genome—Oncogenesis

WALTER DOERFLER,a URTE HOHLWEG,a KNUT MÜLLER,a RALPH REMUS,a,b HILDE HELLER,a,b AND JENNIFER HERTZa,c

aInstitute of Genetics, University of Cologne, D-50931 Köln, GermanybInstitut für Humangenetik und Anthropologie, Universität Düsseldorf,D-40225 Düsseldorf, GermanycCaredata, San Francisco, California 94105, USA

ABSTRACT: We have been interested in the consequences of foreign DNAinsertion into established mammalian genomes and have initially studied thisproblem in adenovirus type 12 (Ad12)–transformed cells or in Ad12-inducedhamster tumors. Since integrates are frequently methylated de novo, it appearsthat they might be modified by an ancient defense mechanism against foreignDNA. In cells transgenic for the DNA of Ad12 or for the DNA of bacteriophage�, changes in cellular methylation and transcription patterns have beenobserved. Thus, the insertion of foreign DNA can have important functionalconsequences that are not limited to the site of foreign DNA insertion. Thesefindings appear to be relevant also for tumor biology and for the interpretationof data derived from experiments with transgenic organisms. For most ani-mals, the main portal of entry for foreign DNA is the gastrointestinal tract.Large amounts of foreign DNA are regularly ingested with the supply of nutri-ents. Starting in 1987/1988, we have been investigating the fate of orally admin-istered foreign DNA in mice. Naked DNA of bacteriophage M13 and the clonedgene for the green fluorescent protein (GFP) of Aequorea victoria have beenused as test molecules. Moreover, the plant-specific gene for the ribulose-1,5-bisphosphate carboxylase (rubisco) has been followed in mice after feedingsoybean leaves. At least transiently, food-ingested DNA can be traced todifferent organs and, after transplacental transfer, to fetuses and newborns.There is no evidence for germ line transmission or for the expression of orallyadministered GFP DNA.

KEYWORDS: Genome; Foreign DNA insertion; DNA methylation; Ad12; Tumor

LONG-TERM GOAL AND CONCEPT OF PROJECT

The long-term genome-wide consequences of the insertion of foreign DNA intoan established mammalian genome have been the focus of our research and will besummarized in this communication. We set out to study the mechanism of tumorinduction by the human adenovirus type 12 (Ad12) in newborn Syrian hamsters(Mesocricetus auratus).1,2 In each tumor cell, multiple copies of Ad12 DNA with a

Address for correspondence: Walter Doerfler, Institut für Genetik, Universität zu Köln,Weyertal 121, D-50931 Köln, Germany. Voice: +49-221-470-2386; fax: +49-221-470-5163.

[email protected]

277DOERFLER et al.: FOREIGN DNA INTEGRATION

genome size of 34,125 base pairs (bp) are covalently linked to the DNA of the cell.The chromosomal insertion site is identical in all cells of a given tumor, but variesfrom tumor to tumor. The characteristics of the adenoviral integration sites andpossible mechanisms involved in the insertional event are summarized in the thirdsection of this report. The Ad12-induced tumors are most likely of clonal origins.Revertants of the tumor cells, which have lost all or a part of the multiple copies ofintegrated Ad12 DNA, arise occasionally and retain the oncogenic phenotype evenafter the viral genomes have been lost.3,4 This finding supports a hit-and-runmechanism of viral oncogenesis.

The details of Ad12 DNA insertion appear to be very similar, if not identical, formany types of foreign DNA that are integrated into mammalian genomes. Apparently,it does not matter whether the foreign DNA reaches the nucleus of a cell via virusinfection or by one of the somewhat artificial transfection procedures that have beendeveloped to transport foreign DNA into the nuclei of cells. Mammalian cells andtheir genomes apparently harbor a highly flexible set of mechanisms to accept andrecombine with foreign DNA. It is a challenging question to consider the evolu-tionary significance of the integration reaction. The presently available data on thehuman genome have revealed that some human genes appear likely to have resultedfrom horizontal transfer from bacteria at some point in the vertebrate lineage.5

Foreign DNA integrates, particularly those consisting of viral DNA, frequentlycomprise between 5 to 30 copies of the viral genome. This number can add up to amegabase (Mb) of foreign DNA, for example, for Ad12 or bacteriophage λ DNA.This size of an insert cannot be accommodated without disturbances for the stabilityof the genome. This perturbation might be transmitted to neighboring DNAsequences, to their chromatin structure, and possibly to adjacent chromosomes thatare in contact with the insertion site of the chromosome targeted by foreign DNAinsertion. There is only scanty information about the consequences of foreign DNAintegration for the recipient genome. We have studied the following:

(a) the de novo methylation of the integrated foreign DNA;(b) the alterations in DNA methylation and transcription patterns of cellular

DNA sequences also at sites remote from the site of integration.

Over an evolutionary time span, viral infections and the integration of retroviralgenomes, or of retrotransposons, have contributed to a considerable extent to theaccumulation of repetitive DNA sequences in mammalian genomes, as has been veryprecisely documented, for example, for the human genome.5 Viruses have evolvedhighly specialized mechanisms to transport their genomes into the nuclei of hostcells and to fix their genomes permanently by integration into the host genome.

On the other hand, the main portal of entry of foreign DNA in mammalianorganisms is the gastrointestinal tract, which is constantly exposed to the oral uptakeof foreign DNA by food ingestion. For over a decade, our laboratory has studied thefate of foreign DNA in the gastrointestinal tract of mice. This DNA is not completelydegraded to the mononucleotide level. A small proportion of food-ingested DNAsurvives in the form of fragments that can be detected in the nuclei of the intestinalepithelia, of cells in the Peyer’s patches of the intestinal wall, of white blood cells,of spleen, and of liver cells. Food-ingested foreign DNA can also transgress theplacental barrier. There is, however, no evidence for germ line transmission. Germ

278 ANNALS NEW YORK ACADEMY OF SCIENCES

cells seem to be protected. In splenic cells, food-ingested foreign DNA fragmentshave been found in covalent linkage to mouse pseudogenes. Food-derived foreigngenes do not seem to be detectably transcribed in any of the organs in which thisDNA has been detected.6–9

Since integrated foreign DNA in mammalian and plant cell systems becomes fre-quently methylated de novo, possibly as a consequence of the activity of an ancienthost defense mechanism,10 it has been reasoned that all organisms must have beenexposed to foreign DNA since the earliest times of their evolution. In our laboratory,this argument has led to an extensive series of studies on the fate of food-ingestedDNA as the main source of foreign DNA in nature.

GENERAL SIGNIFICANCE OF STUDY

Our data gleaned from basic research can be related to problems of generalinterest in biomedical research. The insertion of foreign DNA into established mam-malian genomes plays an important role in several areas of experimental biology.The following topics will be considered here:

(1) Oncogenesis: For many years, we have pursued the working hypothesisthat the integration of foreign DNA is a conditio sine qua non at least inviral oncogenesis.11 The perturbations in chromatin structure and altera-tions in cellular and viral transcription and methylation patterns in thewake of foreign DNA insertion are thought to render a decisive contribu-tion to the transition from a normal mammalian cell to a tumor cell.12,13

Our results on the fate of food-ingested DNA in the mammalian organismraise questions about the role of food-ingested DNA in the generation ofspontaneously occurring tumors.

(2) In transgenic organisms, foreign DNA has been inserted by homologous orheterologous recombination into animal or plant genomes. So far, theconsequences for the stability of the recipient genomes have not beenthoroughly investigated. Therefore, the interpretations of the resultsderived from such experiments have to be viewed with caution.

(3) Similar reservations hold for gene fixation regimens in somatic genetherapy.

A SYNOPSIS OF THE CHARACTERISTICS OF FOREIGN (AD12, AD2, �) DNA INTEGRATION INTO MAMMALIAN GENOMES

Most of the analyses performed in our laboratory have used Ad12 DNA; in someinstances, Ad2 or bacteriophage λ DNA was used. It is likely that many of theobservations described in this chapter hold true for any type of foreign DNA that isintegrated into a recipient mammalian genome. However, since the mechanism(s) ofinsertional recombination in mammalian cells seems to be very flexible and com-plex, we do not claim that the following catalogue of parameters will be complete orall-inclusive. Overviews on this topic have been published previously.12–14


Adenovirus DNA is integrated into the host genome in transformed cells and inAd12-induced hamster tumor cells.2,11,15–18 Free Ad12 DNA has never been foundin these cells.

In abortively and productively infected human cells, Ad12 DNA associates withthe chromosomes early after infection.19

A symmetric recombinant (SYREC) carries the 2081 left-terminal Ad12 DNAnucleotides that flank a long palindrome of human cellular DNA on both termini.Hence, Ad12 DNA can recombine with human cellular DNA also in productivelyinfected cells. This recombinant DNA molecule with the length of viral DNA ispackaged into Ad12 virions due to the presence of the viral packaging signal.20,21

The nucleotide sequence analyses of several integration sites of Ad12 or Ad2DNA into the host genome have not revealed a unique or specific cellular targetsequence for the insertion.14

Most sites of viral DNA integration in transformed and tumor cells or of viral-cellular DNA recombination products generated in a cell-free system arecharacterized by patchy homologies from 2 to 20 nucleotides in length between therecombination partners.18,22–25

The mechanism of this insertional recombination resembles mostly that ofheterologous recombination, albeit short sequence homologies between therecombination partners frequently support the integration reaction.

In the process of integration, terminal viral nucleotides (2–174 bp, in differentintegration events) can be deleted.26 At the site of insertion, cellular DNA can alsobe deleted.27 At one site, however, not a single cellular nucleotide has been missingin comparison to the preinsertion cellular DNA sequence.28

In about 60 different Ad12-induced tumors, each tumor cell in a given tumorcarries the Ad12 integrates at one and rarely at two chromosomal sites that areidentical in all cells of one tumor, but differ from tumor to tumor.2

In Ad12-induced tumor cells or Ad12-transformed cells, up to 30 copies of viralDNA are integrated, at least some of them intact. There is a pearl-like array of Ad12integrates that are, however, separated from each other by cellular or rearranged viralDNA.19,29

The cellular preinsertion sites are frequently transcriptionally active in cells priorto Ad12 infection and viral DNA integration.30,31 Transcriptionally active chromatinmay be predisposed to recombination with foreign DNA.

The integrated Ad12 genomes, in general, are very stable, even after the transferand continuous propagation of tumor cells in culture for many generations.2,17

Occasionally, the Ad12 integrates can be lost in part or completely.32 The cellsdevoid of Ad12 DNA, however, retain their oncogenic potential. This observationhas supported the idea of a hit-and-run mechanism of Ad12 oncogenesis.3,4

Ad12 DNA integrated into the hamster cell genome becomes de novo methylatedin specific patterns.17,33 Virion DNA extracted from purified Ad12 particles is notmethylated.34 An inverse correlation between DNA methylation and Ad12 geneactivity has been described for the first time in Ad12-transformed cells.33,35

De novo methylation appears to be regionally initiated in paracentrally locatedsegments of the collinearly integrated Ad12 genomes36 and spreads from thesedistinct regions to other parts of the genome.37,38

As a further consequence of Ad12 or other foreign (bacteriophage λ) DNAintegration, cellular DNA methylation and transcription patterns can be altered even


at sites remote from the integration locus.39–42 The loci affected by these genomicperturbations may be related, in an as yet unknown way, to the site of foreign DNAintegration.

Most of the parameters described here hold true also for bacteriophage λ DNAthat has become integrated into hamster cell DNA after transfection of cells selectedfor the expression of the cotransfected neomycin phosphotransferase gene.39,40

Apparently, the type of foreign DNA and the mode of transfer into the recipient cellnucleus do not affect the cellular mechanism(s) responsible for the insertionalrecombination with foreign DNA.

DE NOVO METHYLATION OF INTEGRATED FOREIGN DNA

We have investigated this problem in two different experimental systems:(i) integrated Ad12 genomes and (ii) the mouse B lymphocyte tyrosine kinase (BLK)gene, which has been reintegrated by homologous or heterologous recombinationinto the mouse genome.43 Our current, still-limited understanding of problems relatedto the de novo methylation of integrated DNA in these systems is summarized below.

Integrated Ad12 Genomes in Hamster Tumor Cells

In our first experiments in this project, integrated Ad12 genomes in hamstertumor and transformed cells could not be cleaved with methylation-sensitive restric-tion enzymes like HpaII or HhaI,17 whereas the HpaII isoschizomer MspI cleavedthe inserted Ad12 DNA to the same pattern as Ad12 virion DNA that we had shownto be nonmethylated.34 After this discovery on the de novo methylation of integratedforeign DNA in mammalian genomes, we started a series of experiments on the bio-logical function of DNA methylation in mammalian cells. A further discovery33,44

established, for the first time, inverse correlations between the levels of promotermethylation and promoter activity. Soon afterwards, we documented that thesequence-specific methylation of viral promoters led to the inactivation of thesepromoters.45–48

In the years that followed, we investigated DNA methylation patterns in thehuman genome because we reasoned that DNA methylation must have a biologicalfunction that reaches beyond the long-term shutdown of promoters.35 This moregeneral role was thought to be in the establishment of chromatin domains for whichpatterns of DNA methylation might present the initial scaffold on which to buildmore complex structures by specific DNA-protein interactions. A number of humangenomic segments were therefore investigated for specific DNA methylationpatterns.49–58 Patterns of methylation were highly specific for different parts of thehuman genome and were different from cell type to cell type. The patterns of DNAmethylation in the human genome were frequently found to be highly conservedbetween individuals.49–57

We have started a program to elucidate the establishment and mechanisms of denovo methylation by using integrated Ad12 genomes as models.36 Currently, ourresults can be summarized as follows: Upon the integration of Ad12 DNA into thegenome of Ad12-induced tumor cells, the de novo methylation appears to be initi-ated at specific sites in the genome. The results gleaned from analyses using bisulfite


genomic sequencing indicate that the initiation is not at a specific nucleotide or anarrowly restricted set of nucleotides, but rather in a region of the integratedgenome. Moreover, there is considerable variability from tumor to tumor withrespect to where exactly within this specified region de novo methylation commences(U. Hohlweg, A. Schramme, R. Remus, and W. Doerfler; unpublished experiments).From the site(s) of initiation of de novo methylation, this modification spreadsgradually and progressively, but not uniformly, across major parts of the Ad12genome. Moreover, certain parts of the Ad12 genome, particularly those that areactively transcribed in the tumor cell, remain unmethylated or become only hypo-methylated. Further work will be required to understand this mechanism in depth.We surmise that a chromatin-like structure across integrated genomes of foreignderivation, like the Ad12 genome, might have a decisive influence on the initiationand spreading37,38 of de novo methylation.

De Novo Methylation of Foreign DNA Integrated into the Mouse Genomeby Homologous or Heterologous Recombination

In one set of experiments, we have reinserted the mouse gene for the BLK intothe mouse genome by homologous recombination into its authentic genomic site onone of the chromosome 14 alleles.43 In this case, the previously unmethylated BLKgene, which had been cloned and propagated in a methylation-deficient bacterialhost, became remethylated in exactly the same pattern that had preexisted on theunmanipulated mouse alleles. When the BLK gene landed, however, at heterologoussites somewhere in the mouse genome, different patterns of de novo methylationwere observed. We submit that different sites in the mouse genome might have“memory signals” for the establishment of their specific methylation patterns. Thesesignals could be related to specific chromatin properties at a given genomic site.Furthermore, foreign genes, which had been attached to the BLK gene, like theluciferase gene under the control of the weak (Ad2E2AL) or the strong (early SV40)promoter, seemed to be differently methylated depending on promoter strength.Tethering of the gene to a weak promoter frequently resulted in hypermethylation.On the other hand, linkage of the gene to a strong promoter would result in hypo-methylation or absence of de novo methylation. This dependence on promoterstrength did not hold for all sites in the genome, but was most clearly observed whenthe construct had been reinserted by homologous recombination.

GENOME-WIDE PERTURBATIONS IN THE MAMMALIAN GENOMEUPON FOREIGN DNA INSERTION

We have started to investigate the structural and functional consequences of theinsertion of foreign DNA into established mammalian genomes.41,42 The de novomethylation of the integrated DNA and alterations in the patterns of DNA methyla-tion in the recipient genomes at the site of insertion and remote from it have been ofparticular interest. By using different methods, including the bisulfite protocol of thegenomic sequencing technique,59,60 we have documented extensive changes in thepatterns of DNA methylation at several cellular sites remote from the loci ofinsertion of the DNA of Ad12 and lesser changes in cells transgenic for the DNA of


bacteriophage λ.39,40 Since λ DNA is not transcribed in transgenic mammalian cells,alterations of methylation patterns subsequent to foreign DNA insertion are possiblynot dependent on foreign gene transcription. It has been shown earlier that cellularDNA sequences immediately abutting the foreign DNA integrates also exhibitchanges in DNA methylation.61 It is presently unknown by what mechanisms theinsertion of foreign DNA affects the organization and function of the recipientgenome. Does the site of foreign gene integration determine where the remoteeffects occur and is there a requirement for a critical size of integrate? We surmisethat the acquisition of many kilobases or even a megabase of inserted DNA alters thechromatin topology and thus influences the function of specific parts of the genomeon chromosomes that are in contact with the site of foreign DNA integration in theinterphase nucleus.

A collection of cellular DNA segments and genes was analyzed and searched forchanges in DNA methylation and transcription. The technique of methylation-sensitive representational difference analysis (MS-RDA) was based on a subtractivehybridization protocol after selecting against DNA segments that were heavilymethylated and hence rarely cleaved by the methylation-sensitive endonucleaseHpaII. The MS-RDA protocol led to the isolation of several cellular DNA segmentsthat were indeed more heavily methylated in λ DNA–transgenic hamster cell lines.42

By applying the suppressive subtractive hybridization technique to cDNA prepara-tions from nontransgenic and Ad12-transformed or λ DNA–transgenic hamstercells, several cellular genes with altered transcription patterns were cloned fromAd12-transformed or λ DNA–transgenic hamster cells. Many of the DNA segmentswith altered methylation, which were isolated by a newly developed amplicon sub-traction (MS-AS) protocol,41 and cDNA fragments derived from genes with alteredtranscription patterns were identified by their nucleotide sequences.42 In controlexperiments, no differences in gene expression or DNA methylation patterns weredetectable among individual nontransgenic BHK21 cell clones.

In one mouse line transgenic for the DNA of bacteriophage λ, hypermethylationwas observed in the imprinted Igf2r gene in DNA from heart muscle. Two mouselines transgenic for an adenovirus promoter-indicator gene construct showed hypo-methylation in the interleukin 10 (IL10) and Igf2r loci. We conclude that the inser-tion of foreign DNA into an established mammalian genome can lead to alterationsin cellular DNA methylation and transcription patterns. It is conceivable that thegenes and DNA segments affected by these alterations depend on the site(s) offoreign DNA insertion.

PERSISTENCE OF FOOD-INGESTED FOREIGN DNAIN THE MAMMALIAN ORGANISM

When foreign DNA, like that of oncogenic viruses or of constructs transfectedinto cells, is permanently fixed in established mammalian or plant genomes bycovalent integration, the transgenic DNA is often de novo methylated in specificpatterns.17,33,37,43,62–64 This mechanism of long-term gene silencing has been docu-mented in many eukaryotic systems and has been interpreted as an ancient cellulardefense against the activity of foreign genes.10,65 The existence of this ubiquitousdefense machinery raised the question of its evolutionary origin and prompted a


search for the main portal of entry of foreign DNA into organisms. In mammals, thegastrointestinal tract and the large amounts of food-associated DNA undoubtedlyqualify as the most prominent candidates. We have therefore started to study the fateof food-ingested foreign DNA in the gastrointestinal tract of mice.6–9

Information on the stability and persistence of macromolecules in the intestinalsystem has previously not been available. On the other hand, DNA has been shownto survive for centuries, at least in fragmented form, in the remains of even extinctorganisms.66,67 In previous investigations on the persistence of foreign DNA duringthe gastrointestinal passage of food, we have orally administered test DNA asunprotected purified molecules to mice in model experiments. The DNA of bacterio-phage M13 or a vector plasmid carrying the cloned gene for the green fluorescentprotein (GFP) from Aequorea victoria tethered to different viral promoters (HCMV,SV40, or RSV) has been shown to persist up to 24 hours in fragmented form inminute amounts in different parts of the gastrointestinal canal. This DNA has alsobeen found to gain access to cells of the intestinal wall, those of the Peyer’s patches,peripheral white blood cells, and cells in the spleen and liver. Furthermore, food-ingested DNA can transgress the placental barrier in pregnant mice. However, onlya few cells in the embryo or fetus take up the foreign DNA, all of them into theirnuclei.8

We have recently chosen a natural scenario and fed soybean leaves to mice.9 Thedistribution of the plant-specific gene for the nucleus-encoded ribulose-1,5-bis-phosphate carboxylase (rubisco) gene has been studied in the murine organism. Therubisco gene or fragments of it can be recovered from 2 hours up to 49 hours afterfeeding in the intestine and up to 121 hours in the cecum. Thus, plant-associated,naturally fed DNA is more stable in the intestinal tract than naked DNA. Rubiscogene-specific PCR products have also been amplified from DNA of the spleen andliver. There is no evidence for the expression of orally administered genes asassessed by the RT-PCR method. Moreover, mice have been continuously fed dailywith GFP DNA for eight generations and have been examined for the transgenic stateby assaying DNA from tail tips, occasionally from internal organs of the animals, byPCR. The results have been uniformly negative and argue against the germ linetransfer of orally ingested DNA. Upon the intramuscular injection of GFP DNA,authentic GFP DNA fragments have been amplified by PCR up to 17 months post-injection in DNA from muscle, but only up to 24 hours postinjection in DNA fromorgans remote from the site of injection. Upon GFP DNA injection, GFP fragmentscan also be retrieved from the intestinal contents up to 6 hours postinjection. Appar-ently, the organism eliminates injected foreign DNA via the liver-bile-intestinalroute.9

In contrast to earlier work, when we orally administered naked DNA to mice,6–8

we have now followed the fate of the plant-specific, nucleus-encoded rubisco genein mice after feeding soybean leaves. This approach reconstitutes conditions moreakin to a natural scenario and has the advantage of following a nucleus-encodedplant-specific gene with no homologies to mouse or Escherichia coli DNA. Never-theless, the results presented are very similar to those reported on the fate of orallyadministered naked M13 or pEGFP-C1 DNA.6–8 A major difference lies in the longerpersistence of the rubisco gene and in the maintenance of the high molecular massof this gene in the intestinal contents, possibly because some of this DNA might stillbe included in subcellular leaf structures in the gut or might be tightly associated


with proteins. Gut contents or feces examined microscopically still contained intactremnants of leaves.

Food-ingested DNA fragments that survive the gastrointestinal passage, perhapsbecause they are protected in DNA-protein complexes, are distributed in the organismvia the blood and/or lymph circulation. It is likely that peripheral white blood cellsserve as vectors for this transport.7,8 It is unknown what role the immune systemplays in the defense against foreign DNA. Food-ingested and, surprisingly, intra-muscularly injected foreign DNA molecules are excreted via the gastrointestinalroute. Injected DNA can also be traced to the kidney.

After the oral application of the GFP gene construct under SV40-, RSV-, or HCMV-promoter control, 50 µg daily for 21 days, transcription of this foreign gene in variousorgan systems has not been detected by the sensitive RT-PCR method. In a previousreport,8 we have demonstrated that, after oral application, the persistence of the entirepromoter-GFP gene construct in organ DNA can be documented by PCR. Expressionof the same construct in the injected mouse muscle, although not at remote sites, hasbeen shown by UV microscopy and by RT-PCR after intramuscular injection. Thus,the construct can be transcribed and translated at least in muscle tissue of mice. Afterfeeding, one can expect only a tiny, if any, portion of the construct molecules to persistintact, and these molecules would have to reach cells offering a milieu with an optimalcombination of transcription and translation factors conducive to the efficient expres-sion of the GFP gene. So far, we have not been able to adduce any evidence for theexpression of orally administered foreign (pEGFP-C1) DNA.

As might be expected, there is no indication for the germ line transmission oforally ingested DNA. The data available argue for transplacental transfer to the fetusto a limited extent when pregnant animals are fed. The nuclei of single cells in smallclusters have been found positive by the fluorescent in situ hybridization (FISH)technique, and the test DNA-specific signals have been located exclusively in thenuclei and in rare instances in association with both chromatids in fetal cells.8

Apparently, cells of the germ line are spared from this transmission. At this time, itis of little use to speculate about the long-term effects of the presence of foreignDNA in individual nuclei in the fetus or in the animals that ingested the DNA.

The results on the fate of food-ingested foreign DNA in the mammalian organismhave been discussed among specialists concerned about food in general and aboutgenetically modified organisms in the food chain in particular. Although it will bemandatory to consider this problem case by case, the food-consuming public can bereassured by the realization that all kinds of foreign genes in almost limitless com-binations have been part of the food chain throughout the evolution of the speciesHomo sapiens and other species as well. For millennia, these genes and their break-down products with high recombinatorial capacities have been constant partners inour gastrointestinal inner milieu and that of other species.

Uptake of Bacterial DNA via the Gastrointestinal Pathway

In 1997/1998, among the recloned M13 DNA sequences from splenic DNA,which was isolated from mice fed with M13 DNA, we reported that there were twoclones that carried authentic E. coli DNA sequences immediately abutting therecloned M13 DNA.7,8 At that time, we discussed the possibility that bacterial DNA


might have been taken up through the gastrointestinal wall by the same route as theM13 test DNA that had been administered in our experiments. This uptake andgenomic fixation might have been going on for evolutionary time spans. It has beenpostulated “that there is a steady stream of foreign DNA from the intestinal or otherexternal surfaces of the organism to the spleen and possibly other organs of the ani-mals”.8 These results and interpretations gain increased support and new relevancein light of the finding that bacterial genes have been found to be part of the humangenome.5

ACKNOWLEDGMENTS

We thank Petra Böhm and Susanne Scheffler for expert editorial work. At differenttimes, this research was supported by the Deutsche Forschungsgemeinschaft(SFB274-A1), the Bundesministerium für Bildung und Forschung (BEO-0311110),and the Bayerisches Staatsministerium für Landesentwicklung und Umweltfragen.Jennifer Hertz was the recipient of a stipend from the Humboldt Foundation.

REFERENCES

1. TRENTIN, J.J., Y. YABE & G. TAYLOR. 1962. The quest for human cancer viruses: a newapproach to an old problem reveals cancer induction in hamsters by human adenovirus.Science 137: 835–841.

2. HILGER-EVERSHEIM, K. & W. DOERFLER. 1997. Clonal origin of adenovirus type 12–induced hamster tumors: nonspecific chromosomal integration sites of viral DNA.Cancer Res. 57: 3001–3009.

3. KUHLMANN, I. et al. 1982. Tumor induction by human adenovirus type 12 in hamsters:loss of the viral genome from adenovirus type 12–induced tumor cells is compatiblewith tumor formation. EMBO J. 1: 79–86.

4. PFEFFER, A. et al. 1999. Integrated viral genomes can be lost from adenovirus type 12–induced hamster tumor cells in a clone-specific, multistep process with retention ofthe oncogenic phenotype. Virus Res. 59: 113–127.

5. INTERNATIONAL HUMAN GENOME SEQUENCING CONSORTIUM. 2001. Initial sequencingand analysis of the human genome. Nature 409: 860–921.

6. SCHUBBERT, R., C. LETTMANN & W. DOERFLER. 1994. Ingested foreign (phage M13)DNA survives transiently in the gastrointestinal tract and enters the bloodstream ofmice. Mol. Gen. Genet. 242: 495–504.

7. SCHUBBERT, R. et al. 1997. Foreign (M13) DNA ingested by mice reaches peripheralleukocytes, spleen, and liver via the intestinal wall mucosa and can be covalentlylinked to mouse DNA. Proc. Natl. Acad. Sci. U.S.A. 94: 961–966.

8. SCHUBBERT, R., U. HOHLWEG & W. DOERFLER. 1998. On the fate of food-ingestedforeign DNA in mice: chromosomal association and placental transmission to thefetus. Mol. Gen. Genet. 259: 569–576.

9. HOHLWEG, U. & W. DOERFLER. 2001. On the fate of plant or other foreign genes uponthe uptake in food or after intramuscular injection. Mol. Gen. Genomics 265: 225–233.

10. DOERFLER, W. 1991. Patterns of DNA methylation—evolutionary vestiges of foreignDNA inactivation as a host defense mechanism—a proposal. Biol. Chem. Hoppe-Seyler 372: 557–564.

11. DOERFLER, W. 1970. Integration of the deoxyribonucleic acid of adenovirus type 12into deoxyribonucleic acid of baby hamster kidney cells. J. Virol. 6: 652–666.

12. DOERFLER, W. 1996. A new concept in (adenoviral) oncogenesis: integration of foreignDNA and its consequences. BBA Rev. Cancer Res. 1288: 243–244.


13. DOERFLER, W. 2000. Foreign DNA in Mammalian Systems. Wiley/VCH. Weinheim/New York.

14. DOERFLER, W. et al. 1983. On the mechanism of recombination between adenoviral andcellular DNAs: the structure of junction sites. Curr. Top. Microbiol. Immunol. 109:193–228.

15. DOERFLER, W. 1968. The fate of the DNA of adenovirus type 12 in baby hamsterkidney cells. Proc. Natl. Acad. Sci. U.S.A. 60: 636–643.

16. GRONEBERG, J., Y. CHARDONNET & W. DOERFLER. 1977. Integrated viral sequences inadenovirus type 12–transformed hamster cells. Cell 10: 101–111.

17. SUTTER, D., M. WESTPHAL & W. DOERFLER. 1978. Patterns of integration of viral DNAsequences in the genomes of adenovirus type 12–transformed hamster cells. Cell 14:569–585.

18. KNOBLAUCH, M. et al. 1996. DNA methylation of adenovirus type 12 DNA integrationsites in the hamster cell genome. J. Virol. 70: 3788–3796.

19. SCHRÖER, J., I. HÖLKER & W. DOERFLER. 1997. Adenovirus type 12 DNA firmly associ-ates with mammalian chromosomes early after virus infection or after DNA transferby the addition of DNA to the cell culture medium. J. Virol. 71: 7923–7932.

20. DEURING, R., G. KLOTZ & W. DOERFLER. 1981. An unusual symmetric recombinantbetween adenovirus type 12 DNA and human cell DNA. Proc. Natl. Acad. Sci.U.S.A. 78: 3142–3146.

21. DEURING, R. & W. DOERFLER. 1983. Proof of recombination between viral and cellulargenomes in human KB cells productively infected by adenovirus type 12: structureof the junction site in a symmetric recombinant (SYREC). Gene 26: 283–289.

22. GAHLMANN, R. et al. 1982. Patch homologies and the integration of adenovirus DNA inmammalian cells. EMBO J. 1: 1101–1104.

23. JESSBERGER, R., D. HEUSS & W. DOERFLER. 1989. Recombination in hamster cellnuclear extracts between adenovirus type 12 DNA and two hamster preinsertionsequences. EMBO J. 8: 869–878.

24. TATZELT, J. et al. 1993. Fractionated nuclear extracts from hamster cells catalyze cell-free recombination at selective sequences between adenovirus DNA and a hamsterpreinsertion site. Proc. Natl. Acad. Sci. U.S.A. 90: 7356–7360.

25. WRONKA, G. et al. 2001. Integrative recombination between adenovirus type 12 DNAand mammalian DNA in a cell-free system. Under revision.

26. STABEL, S. & W. DOERFLER. 1982. Nucleotide sequence at the site of junction betweenadenovirus type 12 DNA and repetitive hamster cell DNA in transformed cell lineCLAC1. Nucleic Acids Res. 10: 8007–8023.

27. SCHULZ, M. & W. DOERFLER. 1984. Deletion of cellular DNA at site of viral DNAinsertion in the adenovirus type 12–induced mouse tumor CBA-12-1-T. NucleicAcids Res. 12: 4959–4976.

28. GAHLMANN, R. & W. DOERFLER. 1983. Integration of viral DNA into the genome of theadenovirus type 2–transformed hamster cell line HE5 without loss or alteration ofcellular nucleotides. Nucleic Acids Res. 11: 7347–7361.

29. STABEL, S., W. DOERFLER & R.R. FRIIS. 1980. Integration sites of adenovirus type 12DNA in transformed hamster cells and hamster tumor cells. J. Virol. 36: 22–40.

30. GAHLMANN, R., M. SCHULZ & W. DOERFLER. 1983. Low molecular weight RNAs withhomologies to cellular DNA at sites of adenovirus DNA insertion in hamster ormouse cells. EMBO J. 3: 3263–3269.

31. SCHULZ, M. et al. 1987. Transcriptional activities of mammalian genomes at sites ofrecombination with foreign DNA. J. Virol. 61: 344–353.

32. GRONEBERG, J. et al. 1978. Morphological revertants of adenovirus type 12–transformed hamster cells. J. Gen. Virol. 40: 635–645.

33. SUTTER, D. & W. DOERFLER. 1980. Methylation of integrated adenovirus type 12 DNAsequences in transformed cells is inversely correlated with viral gene expression.Proc. Natl. Acad. Sci. U.S.A. 77: 253–256.

34. GÜNTHERT, U. et al. 1976. DNA methylation in adenovirus, adenovirus-transformedcells, and host cells. Proc. Natl. Acad. Sci. U.S.A. 73: 3923–3927.

35. DOERFLER, W. 1983. DNA methylation and gene activity. Annu. Rev. Biochem. 52: 93–124.


36. OREND, G. et al. 1995. The initiation of de novo methylation of foreign DNAintegrated into a mammalian genome is not exclusively targeted by nucleotidesequence. J. Virol. 69: 1226–1242.

37. TOTH, M., U. LICHTENBERG & W. DOERFLER. 1989. Genomic sequencing reveals a 5-methylcytosine-free domain in active promoters and the spreading of preimposedmethylation patterns. Proc. Natl. Acad. Sci. U.S.A. 86: 3728–3732.

38. TOTH, M., U. MÜLLER & W. DOERFLER. 1990. Establishment of de novo DNA methyla-tion patterns: transcription factor binding and deoxycytidine methylation at CpG andnon-CpG sequences in an integrated adenovirus promoter. J. Mol. Biol. 214: 673–683.

39. HELLER, H. et al. 1995. Chromosomal insertion of foreign (adenovirus type 12,plasmid, or bacteriophage λ) DNA is associated with enhanced methylation ofcellular DNA segments. Proc. Natl. Acad. Sci. U.S.A. 92: 5515–5519.

40. REMUS, R. et al. 1999. Insertion of foreign DNA into an established mammaliangenome can alter the methylation of cellular DNA sequences. J. Virol. 73: 1010–1022.

41. MÜLLER, K. & W. DOERFLER. 2000. Methylation-sensitive amplicon subtraction: a novelmethod to isolate differentially methylated DNA sequences in complex genomes.Gene Funct. Dis. 1: 154–160.

42. MÜLLER, K., H. HELLER & W. DOERFLER. 2001. Foreign DNA integration: genome-wide perturbations of methylation and transcription in the recipient genomes. J. Biol.Chem. 276: 14271–14278.

43. HERTZ, J., G. SCHELL & W. DOERFLER. 1999. Factors affecting de novo methylation offoreign DNA in mouse embryonic stem cells. J. Biol. Chem. 274: 24232–24240.

44. VARDIMON, L. et al. 1980. DNA methylation and viral gene expression in adenovirus-transformed and -infected cells. Nucleic Acids Res. 8: 2461–2473.

45. VARDIMON, L. et al. 1982. Expression of a cloned adenovirus gene is inhibited by invitro methylation. Proc. Natl. Acad. Sci. U.S.A. 79: 1073–1077.

46. KRUCZEK, I. & W. DOERFLER. 1983. Expression of the chloramphenicol acetyltrans-ferase gene in mammalian cells under the control of adenovirus type 12 promoters:effect of promoter methylation on gene expression. Proc. Natl. Acad. Sci. U.S.A. 80:7586–7590.

47. LANGNER, K-D. et al. 1984. DNA methylation of three 5′ C-C-G-G 3′ sites in thepromoter and 5′ region inactivates the E2a gene of adenovirus type 2. Proc. Natl.Acad. Sci. U.S.A. 81: 2950–2954.

48. LANGNER, K-D., U. WEYER & W. DOERFLER. 1986. Trans effect of the E1 region ofadenoviruses on the expression of a prokaryotic gene in mammalian cells: resistanceto 5′-CCGG-3′ methylation. Proc. Natl. Acad. Sci. U.S.A. 83: 1598–1602.

49. BEHN-KRAPPA, A. et al. 1991. Patterns of DNA methylation are indistinguishable indifferent individuals over a wide range of human DNA sequences. Genomics 11: 1–7.

50. KOCHANEK, S. et al. 1990. Interindividual concordance of methylation profiles in humangenes for tumor necrosis factors α and β. Proc. Natl. Acad. Sci. U.S.A. 87: 8830–8834.

51. KOCHANEK, S. et al. 1991. DNA methylation profiles in the human genes for tumornecrosis factors α and β in subpopulations of leukocytes and in leukemias. Proc.Natl. Acad. Sci. U.S.A. 88: 5759–5763.

52. ACHTEN, S. et al. 1991. Patterns of DNA methylation in selected human genes indifferent Hodgkin’s lymphoma and leukemia cell lines and in normal humanlymphocytes. Cancer Res. 51: 3702–3709.

53. MUNNES, M. et al. 1998. A 5′-CG-3′-rich region in the promoter of the transcriptionallyfrequently silenced RET protooncogene lacks methylated cytidine residues. Oncogene17: 2573–2584.

54. REMUS, R. et al. 2001. The state of DNA methylation in the promoter regions of thehuman red cell membrane protein (band 3, protein 4.2, and β-spectrin) genes. GeneFunct. Dis. 2: in press.

55. ZESCHNIGK, M. et al. 1997. Imprinted segments in the human genome: different DNAmethylation patterns in the Prader-Willi/Angelman syndrome region as determinedby the genomic sequencing method. Hum. Mol. Genet. 6: 387–395.

56. SCHUMACHER, A. et al. 1998. Methylation analysis of the PWS/AS region does notsupport an enhancer competition model of genomic imprinting on humanchromosome 15. Nat. Genet. 19: 324–325.


57. KOCHANEK, S., D. RENZ & W. DOERFLER. 1993. DNA methylation in the Alu sequencesof diploid and haploid primary human cells. EMBO J. 12: 1141–1151.

58. GENÇ, B. et al. 2000. Methylation mosaicism of 5′-(CGG)n-3′ repeats in fragile X,premutation and healthy individuals. Nucleic Acids Res. 28: 2141–2152.

59. FROMMER, M. et al. 1992. A genomic sequencing protocol that yields a positive displayof 5-methylcytosine residues in individual DNA strands. Proc. Natl. Acad. Sci.U.S.A. 89: 1827–1831.

60. CLARK, S.J. et al. 1994. High sensitivity mapping of methylated cytosines. NucleicAcids Res. 22: 2990–2997.

61. LICHTENBERG, U., C. ZOCK & W. DOERFLER. 1988. Integration of foreign DNA intomammalian genome can be associated with hypomethylation at site of insertion.Virus Res. 11: 335–342.

62. SELKER, E.U., D.Y. FRITZ & M.J. SINGER. 1993. Dense nonsymmetrical DNA methyla-tion resulting from repeat-induced point mutation in Neurospora. Science 262:1724–1728.

63. MEYER, P., I. NIEDENHOF & M. TEN LOHUIS. 1994. Evidence for cytosine methylation ofnon-symmetrical sequences in transgenic Petunia hybrida. EMBO J. 13: 390–399.

64. SCHUMACHER, A. et al. 2000. Epigenetic and genotype-specific effects on the stabilityof de novo imposed methylation patterns in transgenic mice. J. Biol. Chem. 275:37915–37921.

65. YODER, J.A., C.P. WALSH & T.H. BESTOR. 1997. Cytosine methylation and the ecologyof intragenomic parasites. Trends Genet. 13: 335–340.

66. PÄABO, S., J.A. GIFFORD & A.C. WILSON. 1988. Mitochondrial DNA sequences from a7000-year-old brain. Nucleic Acids Res. 16: 9775–9787.

67. PÄABO, S. 1989. Ancient DNA: extraction, characterization, molecular cloning, andenzymatic amplification. Proc. Natl. Acad. Sci. U.S.A. 86: 1939–1943.

foreign dna integration—perturbations of the genome—oncogenesis

Documents