Structure and function of the adenovirus origin of replication

Template requirements for in vivo replication of adenovirus DNA

The adenovirus (Ad) DNA origin of replication was defined through an analysis of the DNA sequences necessary for the replication of plasmid DNAs with purified viral and cellular proteins. Results from several laboratories have shown that the origin consists of two functionally distinct domains: a 10-base-pair sequence present in the inverted terminal repetition (ITR) of all human serotypes and an adjacent sequence constituting the binding site for a cellular protein, nuclear factor I. To determine whether the same nucleotide sequences are necessary for origin function in vivo, we developed an assay for the replication of plasmid DNAs transfected into Ad5-infected cells. The assay is similar to that described by Hay et al. (J. Mol. Biol. 175:493-510, 1984). With this assay, plasmid DNA replication is dependent upon prior infection of cells with virus and only occurs with linear DNA molecules containing viral terminal sequences at each end. Replicated DNA is resistant to digestion with lambda-exonuclease, suggesting that a protein is covalently bound at both termini. A plasmid containing only the first 67 base pairs of the Ad2 ITR replicates as well as plasmids containing the entire ITR. Deletions or point mutations which reduce the binding of nuclear factor I to DNA in vitro reduce the efficiency of plasmid replication in vivo. A point mutation within the 10-base-pair conserved sequence has a similar effect upon replication. These results suggest that the two sequence domains of the Ad origin identified by in vitro studies are in fact important for viral DNA replication in infected cells. In addition, we found that two separate point mutations which lie outside these two sequence domains, and which have little or no effect upon DNA replication in vitro, also reduce the apparent efficiency of plasmid replication in vivo. Thus, there may be elements of the Ad DNA origin of replication which have not yet been identified by in vitro studies.

Constitutive and retinoic acid-inducible expression of cytomegalovirus immediate-early genes in human teratocarcinoma cells

Human teratocarcinoma stem cells are nonpermissive for human cytomegalovirus (HCMV) but become permissive after being induced to differentiate by treatment with retinoic acid. We show that in uninduced teratocarcinoma stem cells, and also in transformed human 293 cells expressing adenovirus E1a gene products, the HCMV immediate-early (IE) 68,000-molecular-weight polypeptide (68K polypeptide) was not expressed, and consequently input viral genomes were not replicated. However, after differentiation of the teratocarcinoma cells, synthesis of the HCMV IE 68K polypeptide was induced, and viral DNA replication occurred. In contrast to our observations for HCMV, simian cytomegalovirus (SCMV) displayed constitutive expression of its analogous IE 94K polypeptide, and the input SCMV genomes were replicated in both uninduced stem cells and 293 cells. Since little, if any, HCMV IE RNA was detectable in human teratocarcinoma or 293 cells after infection under IE conditions, we suggest that a direct transcriptional block to permissivity occurs in these cells. The presence of tandemly repeated sequences which bind nuclear factor I protein in the promoter for the SCMV IE 94K polypeptide gene but not in the promoter for the HCMV IE 68K polypeptide gene may allow the expression of the simian but not of the human IE gene product in transformed cells.

Purification of a cellular, double-stranded DNA-binding protein required for initiation of adenovirus DNA replication by using a rapid filter-binding assay

A rapid and quantitative nitrocellulose filter-binding assay is described for the detection of nuclear factor I, a HeLa cell sequence-specific DNA-binding protein required for the initiation of adenovirus DNA replication. In this assay, the abundant nonspecific DNA-binding activity present in unfractionated HeLa nuclear extracts was greatly reduced by preincubation of these extracts with a homopolymeric competitor DNA. Subsequently, specific DNA-binding activity was detected as the preferential retention of a labeled 48-base-pair DNA fragment containing a functional nuclear factor I binding site compared with a control DNA fragment to which nuclear factor I did not bind specifically. This specific DNA-binding activity was shown to be both quantitative and time dependent. Furthermore, the conditions of this assay allowed footprinting of nuclear factor I in unfractionated HeLa nuclear extracts and quantitative detection of the protein during purification. Using unfrozen HeLa cells and reagents known to limit endogenous proteolysis, nuclear factor I was purified to near homogeneity from HeLa nuclear extracts by a combination of standard chromatography and specific DNA affinity chromatography. Over a 400-fold purification of nuclear factor I, on the basis of the specific activity of both sequence-specific DNA binding and complementation of adenovirus DNA replication in vitro, was affected by this purification. The most highly purified fraction was greatly enriched for a polypeptide of 160 kilodaltons on silver-stained sodium dodecyl sulfate-polyacrylamide gels. Furthermore, this protein cosedimented with specific DNA-binding activity on glycerol gradients. That this fraction indeed contained nuclear factor I was demonstrated by both DNase I footprinting and its function in the initiation of adenovirus DNA replication. Finally, the stoichiometry of specific DNA binding by nuclear factor I is shown to be most consistent with 2 mol of the 160-kilodalton polypeptide binding per mol of nuclear factor I-binding site.

A sensitive and rapid gel retention assay for nuclear factor I and other DNA-binding proteins in crude nuclear extracts

The paper describes a rapid and sensitive assay for DNA binding proteins which interact with specific and defined binding sites. It exploits the observation that complexes of proteins and small synthetic DNA fragments (40 bp) containing the protein/DNA binding site can enter native polyacrylamide gels and remain stably associated during electrophoresis under non-denaturing conditions. The assay was applied to nuclear factor I, to its identification and purification from porcine liver, to an analysis of its binding site on adenovirus type 5 DNA and to an exploration of other potential binding sites for DNA binding proteins within the inverted terminal repetition of adenovirus DNA. The extreme sensitivity of the assay which surpasses that of conventional footprint assays by at least two orders of magnitude permitted the identification of nuclear factor I-like activities in Saccharomyces cerevisiae.

Transcriptional regulation of the human cytomegalovirus major immediate-early gene is associated with induction of DNase I-hypersensitive sites

Human teratocarcinoma cells were used to examine structural features associated with expression of the major immediate-early (IE) gene of human cytomegalovirus. By immunofluorescence, comparison of RNA levels, and in vitro transcription of nuclei, we showed that the major IE gene is inactive in undifferentiated but active in differentiated cells. Therefore, the block in human cytomegalovirus replication in teratocarcinoma cells appears to be at the transcriptional level, in one of the initial genes transcribed. In addition, the in vitro transcription experiments demonstrated that in permissive infections the gene was transcriptionally inactive late in infection. A comparison of the structural features of the promoter region with the active and inactive IE genes showed the presence of constitutive and inducible DNase I-hypersensitive sites. The majority of the constitutive sites existed at -175, -275, -375, -425, and -525 relative to the cap site in an area which has been shown to be capable of simian virus 40 enhancer function. In contrast, the inducible DNase I sites were located outside this region at -650, -775, -875, and -975.

Structure of transcriptionally active chromatin

Transcriptionally active or potentially active genes can be distinguished by several criteria from inactive sequences. Active genes show both an increased general sensitivity to endonucleases like DNase I or micrococcal nuclease and the presence of nuclease hypersensitive sites. Frequently, the nuclease hypersensitive sites are present just upstream of the transcription initiation site covering sequences that are crucial for the promoter function. Viral or cellular transcription enhancer elements are also associated with DNase I hypersensitive sites. At least for the SV40 enhancer, it was shown by electronmicroscopic studies that the DNase I hypersensitive DNA segment is excluded from nucleosomes. It is highly plausible that the binding of regulatory proteins to enhancer or promoter sequences is responsible for the exclusion of these DNA segments from nucleosomes and for the formation of nuclease hypersensitive sites. We speculate that the binding of such proteins may switch on a change in the conformation and/or the protein composition of a chromatin segment or domain containing one to several genes. Biochemical analysis of fractionated nucleosome particles or of active and inactive chromatin fractions have revealed differences in the composition as well as in the degree of modification of histones in these two subfractions of the chromosome. However, until present it is impossible to define unambiguously what are the crucial structural elements that distinguish between particles present on active and inactive chromatin.

Sequence analysis of a KpnI family member near the 3' end of human beta-globin gene

We determined the complete nucleotide sequence (6125 bp) of a full-length member of human KpnI family, designated T beta G41, which is located about 3 kb downstream from the beta-globin gene. Comparison of the sequence with the KpnI family sequence compiled by Singer revealed that a new 131 bp sequence is present in the T beta G41. Hybridization analyses showed that a few thousand of human KpnI family members are carrying this additional sequence. Computer search of DNA databases for T beta G41-homologous sequence showed that some T beta G41-homologous sequences were closely associated with pseudogenes. The T beta G41 sequence also showed significant sequence homology with ChBlym-1, a transferrin-like transforming gene of chicken. Furthermore, an amino acid sequence deduced from the T beta G41 nucleotide sequence revealed a relatively-high homology to those of human transferrin and lactotransferrin.

Origin of adenovirus DNA replication. Role of the nuclear factor I binding site in vivo

An assay is described that detects in vivo a single round of initiation and DNA synthesis directed by a linear molecule containing an exposed single copy of an adenovirus (Ad) origin of replication. This and a previously described assay, which measures multiple rounds of DNA replication, were used to identify DNA sequences within the Ad2 and Ad4 origins of replication that are important for ori function. Linear DNA molecules containing sequences from the Ad2 or Ad4 genome termini were cotransfected with homologous and heterologous helper virus, and net amounts of DNA synthesis were compared. Linear molecules containing the Ad4 inverted terminal repeats were replicated 20-fold better in the presence of the homologous helper, whereas both Ad2 and Ad4 inverted terminal repeats were utilized efficiently by Ad4. DNA sequence analysis of the Ad2 ori and the corresponding region in Ad4 indicated that, although there are only ten variant base-pairs, eight are located within the Ad2 DNA sequence recognized by the cellular protein nuclear factor I. This protein is required to achieve the maximal rate of Ad2 DNA replication in vitro, and these differences therefore identify DNA sequences that are crucial to Ad2 ori function. The Ad4 ITR does not contain a functional nuclear factor I binding site, and deletion analysis has demonstrated that this region of the Ad4 genome is not required for ori function. In contrast to Ad2, the DNA sequences required for the initiation of Ad4 DNA replication were shown to reside entirely within the terminal 18 base-pairs of the Ad4 inverted terminal repeat.

Sequence-specific DNA binding of the Epstein-Barr virus nuclear antigen (EBNA-1) to clustered sites in the plasmid maintenance region

Latently infected B lymphocytes continuously express an Epstein-Barr Virus nuclear antigen (EBNA-1) required in trans for maintenance of the plasmid state of the EBV genome. Filter binding assays and DNAase I footprinting analyses revealed that the carboxy-terminal domain of EBNA-1 protects binding sites at three different loci in the 172,000 bp EBV genome. Two of these loci correspond to essential elements within an 1800 bp segment defined as the minimal region required for plasmid maintenance (ori-P). Binding to each of 20 X 30 bp tandem repeats in the "sink" locus protects 25 bp centered over a 12 bp palindromic consensus sequence TAGCATATGCTA. The nearby dyad symmetry "origin" locus contains two 46 bp protected regions each encompassing two paired core binding sites. The demonstration of sequence-specific binding at multiple loci suggests that EBNA-1 has pleiotropic functions, which may include control of copy number and segregation of the EBV plasmids as well as initiation of replication.

Activation of the major immediate early gene of human cytomegalovirus by cis-acting elements in the promoter-regulatory sequence and by virus-specific trans-acting components

Upstream of the major immediate early gene of human cytomegalovirus (Towne) is a strong promoter-regulatory region that promotes the synthesis of 1.95-kilobase mRNA (D. R. Thomsen, R. M. Stenberg, W. F. Goins, and M. F. Stinski, Proc. Natl. Acad. Sci. U.S.A. 81:659-663, 1984; M. F. Stinski, D. R. Thomsen, R. M. Stenberg, and L. C. Goldstein, J. Virol. 46:1-14, 1983). The wild-type promoter-regulatory region as well as deletions within this region were ligated upstream of the thymidine kinase, chloramphenicol acetyltransferase, or ovalbumin genes. These gene chimeras were constructed to investigate the role of the regulatory sequences in enhancing downstream expression. The regulatory region extends to approximately 465 nucleotides upstream of the cap site for the initiation of transcription. The extent and type of regulatory sequences upstream of the promoter influences the level of in vitro transcription as well as the amount of in vivo expression of the downstream gene. The regulatory elements for cis-activation appear to be repeated several times within the regulatory region. A direct correlation was established between the distribution of the 19 (5' CCCCAGTTGACGTCAATGGG 3')- and 18 (5' CACTAACGGGACTTTCCAA 3')-nucleotide repeats and the level of downstream expression. In contrast, the 16 (5' CTTGGCAGTACATCAA 3')-nucleotide repeat is not necessary for the enhancement of downstream expression. In a domain associated with the 19- or 18-nucleotide repeats are elements that can be activated in trans by a human cytomegalovirus-specified component but not a herpes simplex virus-specified component. Therefore, the regulatory sequences of the major immediate early gene of human cytomegalovirus have an important role in interacting with cellular and virus-specific factors of the transcription complex to enhance downstream expression of this critical viral gene.

Adenovirus sequences required for replication in vivo

We have studied the in vivo replication properties of plasmids carrying deletion mutations within cloned adenovirus terminal sequences. Deletion mapping located the adenovirus DNA replication origin entirely within the first 67 bp of the adenovirus inverted terminal repeat. This region could be further subdivided into two functional domains: a minimal replication origin and an adjacent auxillary region which boosted the efficiency of replication by more than 100-fold. The minimal origin occupies the first 18 to 21 bp and includes sequences conserved between all adenovirus serotypes. The adjacent auxillary region extends past nucleotide 36 but not past nucleotide 67 and contains the binding site for nuclear factor I.

Site-specific DNA binding of nuclear factor I: analyses of cellular binding sites

Nuclear factor I is a cellular site-specific DNA-binding protein required for the efficient in vitro replication of adenovirus DNA. We have characterized human DNA sequences to which nuclear factor I binds. Three nuclear factor I binding sites (FIB sites), isolated from HeLa cell DNA, each contain the sequence TGG(N)6-7GCCAA. Comparison with other known and putative FIB sites suggests that this sequence is important for the binding of nuclear factor I. Nuclear factor I protects a 25- to 30-base-pair region surrounding this sequence from digestion by DNase I. Methylation protection studies suggest that nuclear factor I interacts with guanine residues within the TGG(N)6-7GCCAA consensus sequence. One binding site (FIB-2) contained a restriction endonuclease HaeIII cleavage site (GGCC) at the 5' end of the GCCAA motif. Digestion of FIB-2 with HaeIII abolished the binding of nuclear factor I. Southern blot analyses indicate that the cellular FIB sites described here are present within single-copy DNA in the HeLa cell genome.

Adenovirus DNA replication in vitro: site-directed mutagenesis of the nuclear factor I binding site of the Ad2 origin

The template requirements for efficient adenovirus DNA replication were studied in vitro in a reconstituted system with cloned DNA fragments, containing the Ad2 origin region, as templates. Replication is enhanced by nuclear factor I, a cellular protein that binds specifically to the Ad2 origin. This stimulation is shown to be strongly dependent on the concentration of the adenovirus DNA binding protein. Using synthetic oligonucleotides we have constructed plasmids with base substitutions in the nuclear factor I binding region. Footprint analysis and competition filter binding studies show that two of the three small blocks of conserved nucleotides in this region are involved in the binding of nuclear factor I. The binding affinity can be influenced by the base composition of the degenerate region just outside these two blocks. In vitro initiation and DNA chain elongation experiments with the mutants demonstrate that binding of nuclear factor I to the Ad2 origin is necessary for stimulation. However, binding alone is not always sufficient since a mutation which only slightly disturbs binding is strongly impaired in stimulation of DNA replication by nuclear factor I.

DNA sequence of a 3.8 kilobase pair region controlling Drosophila chorion gene amplification

During Drosophila oogenesis, two clusters of chorion genes and their flanking DNA sequences undergo amplification in the ovarian follicle cells. Amplification results from repeated rounds of initiation and bidirectional replication within the chorion gene regions, possibly from a single origin, producing nested replication forks. Previously we have shown that following reintroduction into the Drosophila genome, a specific 3.8 kilobase pair DNA segment from the amplified third chromosome domain could induce developmentally regulated amplification at its site of insertion. Here we present the complete nucleotide sequence of this "amplification control element" and of genes encoding the chorion structural proteins s18-1 and s15-1, which are contained within it. Sequences that may be involved in the regulation of chorion gene amplification and expression are identified.

Interaction of specific nuclear factors with the nuclease-hypersensitive region of the chicken adult beta-globin gene: nature of the binding domain

We have studied the properties of a factor or factors that bind selectively to the 5' flanking region of the chicken adult beta-globin (beta A-globin) gene. We previously showed that these components, when bound with histones on plasmids containing the region, confer on the complex a pattern of hypersensitivity to nuclease digestion similar to that in the nucleus. We have now measured the abundance of the factor(s) in partially purified preparations, and the affinity constants for binding to specific and nonspecific DNA sequences. Footprinting studies of the specific complex with DNAase I and II reveal two discrete protected regions within the hypersensitive domain. When these regions are physically separated, they interact with the factor(s) independently, suggesting that each region binds one or more distinct components. The footprint patterns of our complexes in vitro agree with the patterns observed in intact chicken erythrocyte nuclei. These complexes thus are on the transcriptionally active beta A-globin gene in vivo.

Site-specific DNA binding of nuclear factor I: analyses of cellular binding sites

Nuclear factor I is a cellular site-specific DNA-binding protein required for the efficient in vitro replication of adenovirus DNA. We have characterized human DNA sequences to which nuclear factor I binds. Three nuclear factor I binding sites (FIB sites), isolated from HeLa cell DNA, each contain the sequence TGG(N)6-7GCCAA. Comparison with other known and putative FIB sites suggests that this sequence is important for the binding of nuclear factor I. Nuclear factor I protects a 25- to 30-base-pair region surrounding this sequence from digestion by DNase I. Methylation protection studies suggest that nuclear factor I interacts with guanine residues within the TGG(N)6-7GCCAA consensus sequence. One binding site (FIB-2) contained a restriction endonuclease HaeIII cleavage site (GGCC) at the 5' end of the GCCAA motif. Digestion of FIB-2 with HaeIII abolished the binding of nuclear factor I. Southern blot analyses indicate that the cellular FIB sites described here are present within single-copy DNA in the HeLa cell genome.

A method for mapping intranuclear protein-DNA interactions and its application to a nuclease hypersensitive site

We have devised a method for mapping sites on DNA within the nucleus that are protected against nuclease attack by interaction with bound protein or other factors. This "footprinting" method uses an end-labeled sequence-specific DNA probe, which is annealed to the DNA from nuclear digests under carefully controlled conditions. The annealed complexes are treated with single-strand-specific nuclease, the resulting duplex molecules are electrophoresed on gels, and the gels are autoradiographed. The high sensitivity and resolution of the method have made it possible to obtain a detailed map of DNase I cutting patterns in the 5' flanking sequence of the chicken adult beta (beta A)-globin gene within nuclei from various tissues. In nuclei from adult erythrocytes, this domain is hypersensitive to nucleases. However, we detect within the domain two well-defined regions that are protected against attack, indicating the presence of one or more bound factors. Nuclei from oviduct or 5-day-old embryonic erythrocytes, in which the domain is not hypersensitive, show limited and different patterns of protection.

Expression in Escherichia coli of a fusion protein product containing a region of the adenovirus DNA polymerase

The bulk of an open reading frame extending from map coordinates 23.3 to 14.2 in region E2b of the adenoviral genome has been cloned and expressed from a chimeric plasmid in Escherichia coli. The cloning strategy used created a fusion protein of 124,000 daltons, which contained greater than 98% adenovirus-encoded sequences. Antiserum raised against this protein reacted with the authentic 140,000-dalton adenovirus DNA polymerase. Another serum raised against a synthetic hexapeptide whose sequence corresponded to the predicted carboxyl terminus of adenovirus-encoded DNA polymerase also reacted with the fusion protein and authentic adenovirus DNA polymerase. These results demonstrate that the cloned region of DNA encodes the adenovirus DNA polymerase.

The TGGCA protein binds to the MMTV-LTR, the adenovirus origin of replication, and the BK virus enhancer

TGGCA-binding proteins are nuclear proteins with high affinity for double-stranded DNA homologous to the prototype recognition sequence 5'YTGGCANNNTGCCAR 3'. Their ubiquitous tissue distribution in higher vertebrates characterizes them as a class of highly conserved proteins which may exert a basic function. To obtain clues to this function, specific binding sites were mapped on three viral genomes. Recognition sites were identified in the enhancer region of the BK virus, in the LTR of the mouse mammary tumor virus, and in the origin of replication of adenovirus 12. The TGGCA-binding protein from HeLa cells appears to be identical to nuclear factor I described by others, which stimulates initiation of adenovirus DNA replication in vitro. However, data from MMTV, BKV, and from cellular genes suggest that this specific protein-DNA interaction may also be involved in the control of gene activity.

High-affinity binding site for a specific nuclear protein in the human IgM gene

Proteins binding to specific regions of DNA with high affinity frequently govern or regulate reactions at the gene level. We have identified a high-affinity binding site in the immunoglobulin mu gene that binds a specific nuclear protein, and have now characterized it fully using nuclear factor 1 (NF-1), a protein purified from the nuclei of HeLa cells and required for the in vitro replication of adenovirus (Ad) DNA. NF-1 protects a 25-base pair (bp) double-stranded segment of DNA which shares a consensus sequence, 5' TGGA/CNNNNNGCCAA 3', with similar binding sites in the Ad-5 terminal repeat and the human c-myc gene. Although this site differs from the enhancer region, a biological function is suggested by the fact that it is DNase I hypersensitive in immunoglobulin-producing lymphoblastoid cells. The binding site for the NF-1 protein in the mu gene, by analogy with the site in the Ad-5 terminal repeat, may represent one component of a cellular origin of replication; alternatively, it may be responsible for the activation of the chromatin in this region.

An adenovirus early region 1A protein is required for maximal viral DNA replication in growth-arrested human cells

Two closely related adenovirus early region 1A proteins are expressed in transformed cells. The smaller of these, which is 243 amino acids in length, is required for the transformation of primary rat cells and for the transformation of immortalized rat cells to anchorage-independent growth. This protein is not required for productive infection of exponentially growing HeLa cells but is required for maximal replication in growth (G0)-arrested human lung fibroblasts (WI-38 cells). To determine the function of this protein in viral replication in these G0-arrested cells, we compared viral early mRNA, early protein, and late protein synthesis after infection with wild type or a mutant which does not express the protein. No differences were found. However, viral DNA synthesis by the mutant was delayed and decreased to 20 to 30% that of wild type in these cells. Viral DNA synthesis was much less defective in growing WI-38 cells, and in the transformed human HeLa cell line it occurred at wild-type levels. Furthermore, the mutant which can express only the 243-amino-acid early region 1A protein induced cellular DNA synthesis in G0-arrested rat cells to the same level as wild-type virus. A mutant which can express only the 289-amino-acid early region 1A protein induced less cellular DNA synthesis in G0-arrested rat cells. We propose that the early region 1A 243-amino-acid protein alters the physiology of arrested permissive cells to allow maximal viral DNA replication. In nonpermissive rodent cells, the 243-amino-acid protein drives G0-arrested cells into S phase. This activity is probably important for the immortalization of primary cells.

Mutational mechanisms by which an inactive replication origin of bacteriophage M13 is turned on are similar to mechanisms of activation of ras proto-oncogenes

M13 viral strand synthesis is initiated by nicking of the viral strand of the duplex replicative form by the M13 gene II initiator protein at a specific site within a sequence of about 40 base pairs having dyad symmetry. Efficient replication of the M13 viral strand also requires the presence of an adjacent sequence of ca. 100 base pairs. Together these sequences constitute the minimal origin for M13 viral strand synthesis. A pBR322 derivative having a 182-base-pair insert of M13 DNA contains a functional M13 viral strand origin and, when provided with M13 gene functions in trans, replicates under conditions nonpermissive for the parent plasmid. Chimeric plasmids containing deletions within the sequence flanking the viral strand origin are unable to replicate under these conditions. We isolated spontaneous mutants of M13 based on their ability to activate replication of such plasmids. The mutations found in these strains, as well as several produced by oligonucleotide-directed mutagenesis, all result in the substitution of any of at least four different amino acids for a specific glycine residue near the amino-terminal end of the initiator protein. Other studies have shown that overproduction of the wild-type initiator protein also restores replication. These alternate mechanisms are discussed in terms of their striking similarity to the mechanisms of activation of the ras proto-oncogenes which can be activated either by increased expression of the wild-type protein or by substitution of any of several amino acids for a glycine residue near the amino terminus.

The origin of adenovirus DNA replication: minimal DNA sequence requirement in vivo

Adenovirus mini-chromosomes which contain two cloned, inverted adenovirus termini replicate in vivo when supplied with non-defective adenovirus as a helper. This system has been used to define the minimum cis acting DNA sequences required for adenovirus DNA replication in vivo. Deletions into each end of the adenovirus inverted terminal repeat (ITR) were generated with Bal31 exonuclease and the resulting molecules constructed into plasmids which contained two inverted copies of the deleted ITR separated by the bacterial neomycin phosphotransferase gene. To determine the effect of the deletion in vivo plasmids cleaved to expose the adenovirus termini were co-transfected with adenovirus type 2 DNA into tissue culture cells. The replicative ability of the molecules bearing adenovirus termini was assayed by Southern blotting of extracted DNA which had been treated with DpnI, a restriction enzyme which cleaves only methylated and therefore unreplicated, input DNA. Molecules containing the terminal 45 bp of the viral genome were fully active whereas molecules containing only 36 bp were in-active in this assay. Therefore sequences required for DNA replication are contained entirely within the terminal 45 bp of the viral genome. Thus, both the previously described highly conserved region (nucleotides 9-18) and the binding site for the cellular nuclear factor I (nucleotides 19-48) are essential for adenovirus DNA replication in vivo.