Integration and transcription of group C human adenovirus sequences in the DNA of five lines of transformed rat cells

Identification of a region on the adenovirus E1A gene responsible for induction by phorbol ester tumor promoter

12-O-Tetradecanoylphorbol-13-acetate (TPA) treatment induces human adenovirus (Ad) early region 1A (E1A) messenger ribonucleic acid expression in infected or Ad-transformed cells. Here, we report that deletion analysis has identified a TPA-responsive element (TRE) in the E1A enhancer region. Deletion analysis indicates that the TRE is located upstream of the E1A cap site between nucleotides -237 and -47. Incubation of extracts from TPA-treated cells with radioactively labeled deoxyribonucleic acid (DNA) fragments containing the TRE (-237 to -47) form specific DNA-protein complexes as demonstrated by gel shift analysis and Southwestern blotting. These experiments provide evidence that novel protein-DNA complexes are formed on a region of the E1A promoter required for TPA-enhanced expression. We speculate that these DNA-binding proteins may interact with the TRE and play a critical role in the mechanism through which TPA upregulates transcription from the Ad E1A gene.

Isolation and characterization of novel adenovirus type 12 E1A mRNAs by cDNA PCR technique

The technique of cDNA polymerase chain reaction was used to study the heterogeneity of adenovirus type 12 early region 1A mRNAs in infected and transformed cells. In addition to 13 S and 12 S mRNAs, three transcripts of 10 S, 9.5 S, and 9 S could be isolated from infected HeLa cells. A fourth transcript of 11 S was found in transformed cells. The 9 S mRNA is generated by removing one intron (nucleotide (nt) 589 to nt 1143) from the RNA precursor. For the 11 S and 10 S mRNAs the primary transcript is spliced twice. The first splicing event removes a common intron from nt 589 to nt 715; the second splicing event removes introns deleted also during the processing of either the 13 S mRNA (leading to the 11 S mRNA) or the 12 S mRNA (leading to the 10 S mRNA). The 9.5 S mRNA is also generated by removing two introns (first splice junction: nt 588/852; the second splice junction is identical to that of the 12 S mRNA). In comparison to the 13 S and 12 S transcripts, the four smaller mRNAs alter the translational reading frames with the beginning of their second exons. In vitro translation of these mRNAs resulted in protein products with molecular weights of 10,100 (11 S and 10 S mRNAs), 10,500 (9.5 S mRNA), and 6200 (9 S mRNA) kDa.

Structure of viral DNA in a rat cell line, GY1, transformed by Ad12 HindIII fragment-G

The cell line GY1, established by transformation of a rat cell line 3Y1 with the Ad12 HindIII fragment-G (leftmost 6.8%, nucleotide 1 to 2322), contains more than 100 viral copies per haploid genome. The viral DNAs in this cell line were cloned into a phage vector, lambda gtWES lambda B, and recloned into pBR322 with their flanking cellular DNAs. Independently isolated 39 clones were analyzed by restriction enzyme cleavage and Southern blot hybridization experiment and divided into 11 classes. Some of classes contained multiple identical clones, at maximum 16 clones. It may be interpreted that amplification of some of the recombined sequences had occurred after the multiple integrations of transfected DNAs within cells. Using five clones from different classes the sequences of recombination sites were determined. Viral DNAs deleted with varying degrees at both ends were flanked by quite different cellular sequences in different clones and no common sequences were revealed around viral-cellular junctions. Tandemly repeated viral DNAs were found in one of the clones to be integrated in a head to tail manner into cellular DNA. The linkage of these two viral DNAs had occurred at the site where parental viral DNAs shared 2 bp. Palindrome structures could be constructed around viral-cellular and viral-viral junction sites and around the regions of parental viral DNAs corresponding to the junction sites in all of the cases investigated.

Detection of cellular proteins associated with human adenovirus type 5 early region 1A polypeptides

Antisera prepared against synthetic peptides corresponding to the amino and carboxy termini of human adenovirus type 5 (Ad5) early region 1A (E1A) proteins were used to identify polypeptides that are associated with these viral species in lytically infected KB cells. Proteins were sought which coprecipitated with E1A polypeptides using both sera and which were not recognized in extracts from mock-infected cells by either serum. Four such species were identified with apparent molecular weights of 68K, 65K, and a doublet at about 105K. A fifth species migrating with a molecular weight in excess of 250K was also identified consistently with E1A-C1 but not E1A-N1 serum. Addition of an excess of the appropriate synthetic peptide to the immunoprecipitation mixtures prevented the precipitation of all of these species. Mixing experiments demonstrated that all species were cellular proteins expressed in normal uninfected KB cells and in addition showed that an association with E1A proteins could take place in vitro. Studies carried out with the mutants pm975 and hr1 indicated that while the 105K doublet and the greater than 250K species were found with the products of both the 1.1- and 0.9-kb E1A mRNAs, 65K and 68K appeared to be primarily associated with those of the 1.1-kb mRNA. Finally, the 105K doublet and greater than 250K were shown to be phosphoproteins. These data indicated that Ad5 E1A proteins may function in a complex with cellular polypeptides which includes species of 105K, 68K, 65K, and possibly a large protein of greater than 250K.

Analysis of multiple forms of human adenovirus type 5 E1A polypeptides using an anti-peptide antiserum specific for the amino terminus

Studies were carried out to further characterize the proteins coded for by the early 1A region (E1A) of human adenovirus type 5 (Ad5) in lytically infected cells using an antiserum prepared against a synthetic octapeptide corresponding to the amino termini of E1A products as well as another anti-peptide serum specific for the carboxy termini. Both sera precipitated the same collection of four major and two minor acidic phosphoproteins and it was confirmed that each of the 1.1- and 0.9-kb E1A mRNAs code for two major and one minor species. These data also indicated that none of the E1A species was produced by proteolytic cleavage. The deletion mutant dl313 which lacks DNA coding for the last 70 C-terminal amino acids of E1A products also produced multiple species which suggested that post-translational modifications involved in their generation do not take place in this region of the proteins. The N-terminal serum was effective in detecting neither the truncated 1.1-kb mRNA product predicted for the mutant hr1 nor the product of the small 0.6-kb E1A mRNA, suggesting that these species are either very short-lived in infected cells or exist in a conformation in which the amino terminus is inaccessable to the antibody.

Immortalization of rat embryo fibroblasts by an adenovirus 2 mutant expressing a single functional E1a protein

Expression of the adenovirus E1a and E1b genes is required for transformation of nonpermissive rodent cells. Differential splicing of the E1a precursor RNA molecules results in the production of two early mRNAs, 13S and 12S, which encode a 289-amino-acid-residue (289R) and 243R protein, respectively. Previously we constructed a mutant virus, dl231, which can only produce normal 289R protein from the E1a gene. In this report we demonstrate that dl231 induced focal transformation of primary rat embryo fibroblasts at 20% of the level of wild-type virus. dl231 transformants were immortalized and produced normal levels of E1a 13S and E1b mRNAs but only minute levels of defective E1a 12S mRNA. These transformants only minimally expressed the transformation phenotype and were similar to untransformed cells. Unlike wild-type transformants, they had a more fibroblastic morphology, were contact inhibited, grew to only low saturation density, and were limited in their ability to grow in an anchorage-independent manner in soft agar. We conclude that the 289R E1a protein can mediate immortalization of primary cells and that the 243R E1a protein is required to elicit the full transformation phenotype.

Mapping the 5' ends, 3' ends, and splice sites of mRNAs from the early E3 transcription unit of adenovirus 5

Using nuclease gel analyses, the sites in the approximately 4000 nucleotide (nt) E3 transcription unit of adenovirus 5 (Ad5) that encode the 5' ends, 3' ends, and 5' and 3' splice sites of the approximately 10 E3 mRNAs were determined. Transcription initiation of all mRNAs occurs at two major (nt 1 and 8) and approximately two minor sites, situated 20-30 nt 3' to a TATA box. There are two major 3' end sites (nt 2227 and 3308), located approximately 20 nt downstream from ATTAAA and an AATAAA sequences, respectively. Thus, ATTAAA, as well as the usual AATAAA, apparently can function as a 3' end signal. There are two 5' splice sites (nt 372 and 923), both with GT at the intron boundary. There are four 3' splice sites (nt 766, 1817, 2201, and 2880), all with AG at the intron boundary. The nt 1817, 2201, and 2880 3' splice sites are located immediately upstream from open reading frames, such that splicing at the different sites allows synthesis of completely different proteins.

Evidence that a second tumor antigen coded by adenovirus early gene region E1a is required for efficient cell transformation

The expression of the adenovirus (Ad) early coding region 1a (E1a) is required for virus-induced cell transformation and for the activation of other viral early genes and some cellular genes. Two overlapping early mRNAs of 13S and 12S that are transcribed from this region code for a 289-amino acid protein and a 243-amino acid protein, respectively. Earlier studies have shown that the 289-amino acid protein is essential for cell transformation. We have constructed an Ad type 2 (Ad2) deletion mutant (dl231) in which the intervening sequence for the 13S mRNA is precisely removed. Mutant dl231 is completely viable in human KB cells and produces normal amounts of 13S mRNA but much reduced amounts of a defective 12S mRNA. Mutant dl231 induces focal transformation of established rat embryo fibroblasts at a frequency one-fifth to one-half that of wild-type virus. However, the transformed cells are defective in their ability to form anchorage-independent colonies on semisolid medium. Therefore, our results demonstrate that the 243-amino acid protein is required for full transformation of rat embryo cells.

Mapping a new gene that encodes an 11,600-molecular-weight protein in the E3 transcription unit of adenovirus 2

The DNA sequence of the early E3 transcription unit of adenovirus 2 (Ad2) (J. Hérissé et al., Nucleic Acids Res. 8:2173-2192, 1980), indicates that an open reading frame exists between nucleotides 1860 and 2163 that could encode a protein of Mr 11,600 (11.6K). We have determined the DNA sequence of the corresponding region in Ad5 (closely related to Ad2) and have established that this putative gene is conserved in Ad5 (a 10.5K protein). To determine whether this protein is expressed, we prepared an antiserum in rabbits against a synthetic peptide corresponding to amino acids 66 to 74 in the 11.6K protein of Ad2. The peptide antiserum immunoprecipitated a ca. 13K-14K protein doublet, as estimated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, from [35S]methionine-labeled Ad2- or Ad5-early-infected KB cells. The antiserum also immunoprecipitated a 13K-14K protein doublet translated in vitro from Ad2 or Ad5 early E3-specific mRNA purified by hybridization to Ad2 EcoRI-D (nucleotides -236 to 2437). The synthetic peptide successfully competed with the 13K-14K protein doublet in immunoprecipitation experiments, thereby confirming the specificity of the antiserum. As deduced from the DNA sequence, the 11.6K protein (and the corresponding 10.5K Ad5 protein) has a conserved 22-amino-acid hydrophobic domain, suggesting that the protein may be associated with membranes. We conclude that a gene located at nucleotides 1860 to 2143 in the Ad2 E3 transcription unit (nucleotides 1924 to 2203) in the Ad5 E3 transcription unit) encodes an 11.6K protein (10.5K in Ad5).

Mutations in the gene encoding the adenovirus early region 1B 19,000-molecular-weight tumor antigen cause the degradation of chromosomal DNA

The adenovirus mutant Ad2ts111 has been previously shown to contain a mutation in the early region 2A gene encoding the single-stranded-DNA-binding protein that results in thermolabile replication of virus DNA and a mutation in early region 1 that causes degradation of intracellular DNA. A recombinant virus, Ad2cyt106, has been constructed which contains the Ad2ts111 early region 1 mutation and the wild-type early region 2A gene from adenovirus 5. This virus, like its parent Ad2ts111, has two temperature-independent phenotypes; first, it has the ability to cause an enhanced and unusual cytopathic effect on the host cell (cytocidal [cyt] phenotype) and second, it induces degradation of cell DNA (DNA degradation [deg] phenotype). The mutation responsible for these phenotypes is a single point mutation in the gene encoding the adenovirus early region 1B (E1B) 19,000-molecular-weight (19K) tumor antigen. This mutation causes a change from a serine to an asparagine in the 20th amino acid from the amino terminus of the protein. Three other mutants that affect the E1B 19K protein function have been examined. The mutants Ad2lp5 and Ad5dl337 have both the cytocidal and DNA degradation phenotypes (cyt deg), whereas Ad2lp3 has only the cytocidal phenotype and does not induce degradation of cell DNA (cyt deg+). Thus, the DNA degradation is not caused by the altered cell morphology. Furthermore, the mutant Ad5dl337 does not make any detectable E1B 19K protein product, suggesting that the absence of E1B 19K protein function is responsible for the mutant phenotypes. A fully functional E1B 19K protein is not absolutely required for lytic growth of adenovirus 2 in HeLa cells, and its involvement in transformation of nonpermissive cells to morphological variants is discussed.

Production of a monospecific antiserum against the early region 1A proteins of adenovirus 12 and adenovirus 5 by an adenovirus 12 early region 1A-beta-galactosidase fusion protein antigen expressed in bacteria

Antisera were prepared against the amino acid sequences encoded within the N-terminal half of the adenovirus 12 (Ad12) early region 1A (E1A) gene. This was accomplished by construction of a plasmid vector which encoded the N-terminal 131 amino acids of Ad12 E1A joined in frame to the coding sequence of beta-galactosidase. After induced synthesis in Escherichia coli, the Ad12 E1A-beta-galactosidase fusion protein (12-1A-FP) was extracted with urea and used to raise antibodies in rabbits. The 12-1A-FP antisera immunoprecipitated major phosphoproteins of 39,000 and 37,000 apparent molecular weights from Ad12-transformed and infected cells. The 12-1A-FP antisera also immunoprecipitated E1A phosphoproteins from Ad5-transformed and infected cells. Immunospecificity of the 12-1A-FP antisera was demonstrated by the ability of 12-1A-FP antigen to block immunoprecipitation of E1A proteins. Furthermore, E1A proteins immunoprecipitated from in vivo-labeled cells comigrated with those translated in vitro by RNA that had been hybridization selected to E1A DNA.

Adenovirus type 5 early region 1b gene product is required for efficient shutoff of host protein synthesis

To determine the role adenovirus 5 early region 1b-encoded 21- and 55-kilodalton proteins play in adenovirus productive infection, mutants have been isolated which were engineered to contain small deletions or insertions at 5.8, 7.9, or 9.6 map units. By using an overlap recombination procedure involving H5dl314 (delta 3.7 to 4.6 map units) DNA cleaved at 2.6 map units with ClaI and the adenovirus 5 XhoI-C (0 to 15.5 map units) fragment containing the desired mutation, viral mutants were isolated by their ability to produce plaques on KB cell line 18, which constitutively expresses only viral early region 1b functions (Babiss et al., J. Virol. 46:454-465, 1983). DNA sequence analysis of the viral mutants isolated (H5dl118, H5dl110, H5in127, and H5dl163) indicates that all of the viruses contain mutations which affect the 55-kilodalton protein, whereas dl118 should also produce a truncated form of the 21-kilodalton protein. When analyzed for their replication characteristics in HeLa cells, all of the mutant viruses exhibited extended eclipse periods and effected yields that were reduced to 10% or less of that produced by H5sub309 (parent virus of the mutants which is phenotypically identical to wild-type adenovirus 5). When compared with characteristics of sub309, the early and late transcription and DNA replication of the mutants were similar, but synthesis of late polypeptides and late cytoplasmic mRNAs was greatly reduced. Quantitation of mutant virus-specific late mRNAs associated with polysomes revealed a threefold reduction when compared with that of sub309. Analysis of infected cell extracts further revealed that these mutants were incapable of efficiently shutting off host cell protein synthesis, suggesting that the 55-kilodalton protein plays a role in this process. These data suggest that early region 1b products may function by interacting with additional viral or host cell macromolecules to modulate host cell shutoff or that some late viral mRNA or polypeptide may potentiate this reaction.

Transformation by human adenoviruses

When, approximately 10 years ago, it was shown that the functions essential for cell transformation were localized in a small region of the adenovirus genome, a DNA segment which at that time was thought to be capable of encoding two or three average-sized proteins at most, it seemed reasonable to hope that an understanding of the mechanisms by which adenoviruses transform cells might be quickly achieved. While such optimism might be forgiven, it was quite clearly naive in the extreme. As a consequence of mRNA splicing and the use of overlapping reading frames the number of proteins encoded within E1 is 2-3-times greater than would have been predicted a decade ago, and post-translational modifications may add another dimension of complexity. In fact it has taken nearly all of the past decade just to identify the proteins encoded in E1 and to characterize them in the most rudimentary way. However, we have now entered a period in which new information is accumulating at an extremely rapid rate as a result of several major technical and fundamental advances. Chief among these are the use of recombinant DNA techniques, particularly site-directed mutagenesis, which combined with methods for introducing mutations made in cloned sequences back into infectious virus, clearly represents a powerful approach to studying the functions of transforming proteins. In addition, the ability to express transforming proteins in bacteria and to produce large amounts of highly purified proteins which previously were only just detectable in infected and transformed cells is a major breakthrough. Advances in immunological techniques, particularly the development of monoclonal antibodies and antisera against synthetic peptides, have enormously simplified the task of detecting and characterizing E1 proteins. Finally, recent results suggesting that adenovirus transforming proteins may be functionally and structurally similar to other oncogenes brings a new perspective to the study of oncogenic transformation. Have all the proteins involved in transformation by adenoviruses been identified? It seems probable that all those virally coded proteins which play a major role are now known but of course minor players in the cast could still be waiting in the wings. We have pointed out that viral functions encoded outside region E1 may have some importance at least in initiation of transformation by virions and have speculated on the possibility that one or more of these may be involved in the integration of viral DNA into the host cell chromosome.(ABSTRACT TRUNCATED AT 400 WORDS)

Establishment and characterization of hamster cell lines transformed by restriction endonuclease fragments of adenovirus 5

We have established a library of hamster cells transformed by adenovirus 5 DNA fragments comprising all (XhoI-C, 0 to 16 map units) or only a part (HindIII-G, 0 to 7.8 map units) of early region 1 (E1: 0 to 11.2 map units). These lines have been analyzed in terms of content of viral DNA, expression of E1 antigens, and capacity to induce tumors in hamsters. All cells tested were found to express up to eight proteins encoded within E1A (0 to 4.5 map units) with apparent molecular weights between 52,000 (52K) and 25K. Both G and C fragment-transformed lines expressed a 19K antigen encoded within E1B (4.5 to 11.2 map units), whereas an E1B 58K protein was detected in C fragment-transformed, but not G-fragment-transformed, lines. No clear distinction could be drawn between cells transformed by HindIII-G and by XhoI-C in terms of morphology or tumorigenicity, suggesting that the E1B 58K antigen plays no major role in the maintenance of oncogenic transformation, although possible involvement of truncated forms of 58K cannot be ruled out. Sera were collected from tumor-bearing animals and examined for ability to immunoprecipitate proteins from infected cells. The relative avidity of sera for different proteins was characteristic of the cell line used for tumor induction, and the specificity generally reflected the array of viral proteins expressed by the corresponding transformed cells. However, one notable observation was that even though all transformed lines examined expressed antigens encoded by both the 1.1- and 0.9-kilobase mRNAs transcribed from E1A, tumor sera made against these lines only precipitated products of the 1.1-kilobase message. Thus, two families of E1A proteins, highly related in terms of primary amino acid sequence, appear to be immunologically quite distinct.

Human adenovirus 2 E1B-19K and E1B-53K tumor antigens: antipeptide antibodies targeted to the NH2 and COOH termini

The human adenovirus 2 (Ad2) transforming region is located in the left 11.1% of the viral genome and encodes two early transcription units, E1A and E1B. Based on the amino acid sequence deduced from the Ad2 E1B DNA sequence (Gingeras et al., J. Biol. Chem. 257:13475-13491, 1982), we have prepared antibodies against synthetic peptides, 8 to 16 amino acids in length, encoded at the NH2 and COOH termini of the major E1B-19K and E1B-53K tumor antigens. The antipeptide antibodies immunoprecipitated the targeted E1B-19K or E1B-53K tumor antigens from extracts of Ad2-infected cells. The specificity of the peptide competition studies. Antipeptide antibodies directed to the NH2 and COOH termini immunoprecipitated the E1B-19K and E1B-53K tumor antigens from two Ad2-transformed rat cell lines, F17 and F4, providing evidence that identical tumor antigens are synthesized in Ad2-infected and Ad2-transformed cells. These results show that the E1B-19K and E1B-53K T antigens are not processed proteolytically at either the NH2 or COOH terminus. Our data provide strong evidence at the protein level that the E1B-19K and E1B-53K tumor antigens partially overlap in DNA sequence, with the E1B-19K initiating translation at the first ATG at nucleotide 1711 in translation reading frame 1 and the E1B-53K tumor antigen initiating translation at the second ATG at nucleotide 2016 in reading frame 3. This confirms the results of others on the N-terminal amino acid sequence of E1B-19K and theoretical deductions based on the DNA sequence. Our findings prove that the large E1B-53K T antigen initiates translation at the second ATG at nucleotide 2016 and not at equally plausible initiation codons located farther downstream at nucleotides 2202 and 2235. Thus, the E1B-53K T antigen is another example of a protein which initiates translation at an internal ATG rather than at the 5'-proximal ATG.

The pattern of integration of viral DNA sequences in the adenovirus 5-transformed human cell line 293

The pattern of integrated adenovirus 5 DNA in the adenovirus 5-transformed human cell line 293 was analyzed by DNA blot hybridization. In contrast to a previous report, only sequences from the left end of the genome were detected. The viral DNA was contained in a unique DNA fragment generated by cleavage of 293 DNA with either EcoRI or BamHI, two enzymes which do not cut the integrated viral DNA. Further mapping studies indicated that the integrated sequence was probably colinear with viral DNA. The joining to host cell sequences occurred between viral nucleotide base pairs 1 and 270 (a site for BalI) on one end and 4123 (SmaI site) and 5372 (BalI site) at the other end. The viral DNA was not tandemly repeated. These results suggest that integration of viral sequences into the host genome probably occurred at a single site. If any subsequent duplications of viral DNA took place, then duplication of extensive cellular sequences also must have occurred.

Nucleotide sequence analysis of the linked left and right hand terminal regions of adenovirus type 5 DNA present in the transformed rat cell line 5RK20

A peculiar phenomenon is observed in several adenovirus type 2 or 5 (Ad2 or Ad5) transformed cell lines: the right hand and left hand terminal regions of the viral genome present in the viral DNA insertions of these cell lines are found to be linked together. A large part of the viral DNA insertion present in the Ad5 transformed rat cell line 5RK20 has been cloned in the lambda vector Charon21A, including the segment containing the linked terminal regions. Sequence analysis of the linkage region showed a perfect homology with the Ad5 DNA sequence and a direct linkage of basepair (bp) 63 of the left hand end of the viral genome to bp 108 of the right hand end. No cellular or rearranged viral sequences were present. Our findings suggest that the joining of viral sequences into the cellular genome.

Identification of adenovirus 2 early region 4 polypeptides by in vitro translation and tryptic peptide map analysis

The mRNA species encoded by early region 4 (E4) (map position [mp] 91.5 to 99.3) of adenovirus 2 were isolated from the polysomes of infected KB cells and were purified by hybridization to the cloned HindIII-F fragment (mp 89.5 to 97.3) or to EcoRI-C fragment (mp 89.7 to 100). The mRNA's were translated in vitro using [35S]methionine as a labeled precursor in rabbit reticulocyte lysates treated with micrococcal nuclease as well as in wheat germ lysates. Five major (35,000-molecular-weight [35K], 23K, 22K, 21K, 18K) polypeptides were observed when the reticulocyte lysate was used. The 23K, 22K, 21K, and 18K polypeptides were also observed with the wheat germ lysate, as well as a very prominent 11K polypeptide; the 35K polypeptide was not observed. Assignment of these polypeptides to E4 was further established by hybrid arrested translation. Two-dimensional gel electrophoresis of a wheat germ translate resolved five polypeptides ranging from 18K to 23K, the major 11K polypeptide, and polypeptides of 10K and 9K. The in vitro 23K to 18K and 11K polypeptides migrated to approximately the same positions on two-dimensional gels as did seven 26K to 21K polypeptides and an 11K polypeptide synthesized in vivo (Brackmann et al., J. Biol. Chem, 255:6772--6779, 1980). Two-dimensional tryptic peptide maps demonstrated that the 35K, 23K, 22K, 21K, and 18K polypeptides are related. The peptide map of 11K is different from those of the above polypeptides, although 11K may share one tryptic methionine polypeptide with them. These results indicate that E4 encodes a major 11K polypeptide, as well as major 35K, 23K, 22K, 21K, and 18K polypeptides.

Patch homologies and the integration of adenovirus DNA in mammalian cells

The hamster cell line HE5 has been derived from primary hamster embryo cells by transformation with human adenovirus type 2 (Ad2). Each cell contains 2-3 copies of Ad2 DNA inserted into host DNA at apparently identical sites. The site of the junction between the right terminus of Ad2 DNA and hamster cell DNA was cloned and sequenced. The eight [corrected] right terminal nucleotides of Ad2 DNA were deleted. The unoccupied cellular DNA sequence in cell line HE5 , corresponding to the site of the junction between Ad2 and hamster cell DNA, was also cloned; 120-130 nucleotides in the cellular DNA were found to be identical to the cellular DNA sequence in the cloned junction DNA fragment, up to the site of the junction. The unoccupied and the occupied cellular DNAs and the adjacent viral DNA exhibited a few short nucleotide homologies. Patch homologies ranging in length from dodeca - to octanucleotides were detected by computer analyses at locations more remote from the junction site. When the right terminal nucleotide sequence of Ad2 DNA was matched to randomly selected sequences of 401 nucleotides from vertebrate or prokaryotic DNA, similar homologies were observed. It is likely that foreign (viral) DNA can be inserted via short sequence homologies at many different sites of cellular DNA.