Initiation of SV40 DNA replication in vivo: location and structure of 5' ends of DNA synthesized in the ori region

Initiation of simian virus 40 DNA replication in vitro: aphidicolin causes accumulation of early-replicating intermediates and allows determination of the initial direction of DNA synthesis

Aphidicolin, a specific inhibitor of DNA polymerase alpha, provided a novel method for distinguishing between initiation of DNA synthesis at the simian virus 40 (SV40) origin of replication (ori) and continuation of replication beyond ori. In the presence of sufficient aphidicolin to inhibit total DNA synthesis by 50%, initiation of DNA replication in SV40 chromosomes or ori-containing plasmids continued in vitro, whereas DNA synthesis in the bulk of SV40 replicative intermediate DNA (RI) that had initiated replication in vivo was rapidly inhibited. This resulted in accumulation of early RI in which most nascent DNA was localized within a 600- to 700-base-pair region centered at ori. Accumulation of early RI was observed only under conditions that permitted initiation of SV40 ori-dependent, T-antigen-dependent DNA replication and only when aphidicolin was added to the in vitro system. Increasing aphidicolin concentrations revealed that DNA synthesis in the ori region was not completely resistant to aphidicolin but simply less sensitive than DNA synthesis at forks that were farther away. Since DNA synthesized in the presence of aphidicolin was concentrated in the 300 base pairs on the early gene side of ori, we conclude that the initial direction of DNA synthesis was the same as that of early mRNA synthesis, consistent with the model proposed by Hay and DePamphilis (Cell 28:767-779, 1982). The data were also consistent with initiation of the first DNA chains in ori by CV-1 cell DNA primase-DNA polymerase alpha. Synthesis of pppA/G(pN)6-8(pdN)21-23 chains on a single-stranded DNA template by a purified preparation of this enzyme was completely resistant to aphidicolin, and further incorporation of deoxynucleotide monophosphates was inhibited. Therefore, in the presence of aphidicolin, this enzyme could initiate RNA-primed DNA synthesis at ori first in the early gene direction and then in the late gene direction, but could not continue DNA synthesis for an extended distance.

DNA replication in vitro erases a Xenopus 5S RNA gene transcription complex

Replication of the Xenopus laevis 5S RNA gene in vitro is unimpeded by the presence of a complete transcription complex assembled on the internal control region of the gene. The transcription complex is disrupted by passage of the replication fork, specific transcription factors are displaced, and the daughter 5S RNA genes are inactivated. We have been unable to demonstrate that any "memory" of the preexisting transcription complex is transmitted to the daughter DNA duplexes following replication.

Replication and supercoiling of simian virus 40 DNA in cell extracts from human cells

Soluble extracts prepared from the nucleus and cytoplasm of human 293 cells are capable of efficient replication and supercoiling of added DNA templates that contain the origin of simian virus 40 replication. Extracts prepared from human HeLa cells are less active than similarly prepared extracts from 293 cells for initiation and elongation of nascent DNA strands. DNA synthesis is dependent on addition of purified simian virus 40 tumor (T) antigen, which is isolated by immunoaffinity chromatography of extracts from cells infected with an adenovirus modified to produce large quantities of this protein. In the presence of T antigen and the cytoplasmic extract, replication initiates at the origin and continues bidirectionally. Initiation is completely dependent on functional origin sequences; a plasmid DNA containing an origin mutation known to affect DNA replication in vivo fails to replicate in vitro. Multiple rounds of DNA synthesis occur, as shown by the appearance of heavy-heavy, bromodeoxyuridine-labeled DNA products. The products of this reaction are resolved, but are relaxed, covalently closed DNA circles. Addition of a nuclear extract during DNA synthesis promotes the negative supercoiling of the replicated DNA molecules.

Isolation of the origin of replication associated with the amplified Chinese hamster dihydrofolate reductase domain

Autoradiography of restriction digests of DNA labeled in early S phase indicates that replication of the amplified dihydrofolate reductase (DHFR) domain of methotrexate-resistant CHOC 400 cells initiates within a 6.1-kilobase pair (kb) EcoRI-doublet located on the 3' side of the DHFR gene. To localize the DHFR origin fragment, synchronized CHOC 400 cells were either pulse labeled with [3H]thymidine in vivo or permeabilized and incubated with [32P]dATP under conditions that support limited chromosomal DNA replication. The temporal order of replication of amplified fragments was determined by hybridization of the in vivo or in vitro replication products to cloned fragments spanning the earliest-replicating portion of the DHFR domain. At the G1/S boundary, the labeled products derived from the replication of amplified sequences, either in whole or permeabilized cells, are distributed about an amplified 4.3-kb Xba I fragment that maps 14 kb downstream from the DHFR gene. As cells progress through the S phase, bidirectional replication away from this site is observed. These studies indicate that the 4.3-kb Xba I fragment contains the origin of replication associated with the amplified DHFR domain.

Bidirectional promoter elements of simian virus 40 are required for efficient replication of the viral DNA

Mutants of simian virus 40 (SV40) lacking parts of the 72- and 21-base-pair repeat regions were made deficient in large T antigen by recombination with dlA 4000, a mutant containing a frameshift deletion near the amino terminus of the T antigen genes. These double mutants were transfected into COS cells, and the amounts of replicated viral DNA were measured at various times thereafter. It was found that deletion of either the 72- or 21-base-pair repeat region did not significantly reduce the accumulation of viral DNA. However, cells transfected with mutants lacking both of these promoter elements accumulated 100-fold less viral DNA than cells transfected with wild-type SV40. This indicates that the 72- and 21-base-pair repeat regions are each sufficient for supplying a function required for efficient replication of SV40 DNA. In addition, the ability of either of these regions to support efficient replication was gradually reduced as the number of promoter elements within each was decreased. Since the 72- and 21-base-pair repeat regions bidirectionally induce transcription, our results indicate that bidirectional promoter elements play a role in the replication of viral DNA. However, fewer of these elements are required for efficient replication than for efficient transcription.

Pyrimidine dimers block simian virus 40 replication forks

UV light produces lesions, predominantly pyrimidine dimers, which inhibit DNA replication in mammalian cells. The mechanism of inhibition is controversial: is synthesis of a daughter strand halted at a lesion while the replication fork moves on and reinitiates downstream, or is fork progression itself blocked for some time at the site of a lesion? We directly addressed this question by using electron microscopy to examine the distances of replication forks from the origin in unirradiated and UV-irradiated simian virus 40 chromosomes. If UV lesions block replication fork progression, the forks should be asymmetrically located in a large fraction of the irradiated molecules; if replication forks move rapidly past lesions, the forks should be symmetrically located. A large fraction of the simian virus 40 replication forks in irradiated molecules were asymmetrically located, demonstrating that UV lesions present at the frequency of pyrimidine dimers block replication forks. As a mechanism for this fork blockage, we propose that polymerization of the leading strand makes a significant contribution to the energetics of fork movement, so any lesion in the template for the leading strand which blocks polymerization should also block fork movement.

Selection of template initiation sites and the lengths of RNA primers synthesized by DNA primase are strongly affected by its organization in a multiprotein DNA polymerase alpha complex

Synthesis of (p)ppRNA-DNA chains by purified HeLa cell DNA primase-DNA polymerase alpha (pol alpha-primase) was compared with those synthesized by a multiprotein form of DNA polymerase alpha (pol alpha 2) using unique single-stranded DNA templates containing the origin of replication for simian virus 40 (SV40) DNA. The nucleotide locations of 33 initiation sites were identified by mapping G*pppN-RNA-DNA chains and identifying their 5'-terminal ribonucleotide. Pol alpha 2 strongly preferred initiation sites that began with ATP rather than GTP, thus frequently preferring different initiation sites than pol alpha-primase, depending on the template examined. The initiation sites selected in vitro, however, did not correspond to the sites used during SV40 DNA replication in vivo. Pol alpha 2 had the greatest effect on RNA primer size, typically synthesizing primers 1-5 nucleotides long, while pol alpha-primase synthesized primers 6-8 nucleotides long. These differences were observed even at individual initiation sites. Thus, the multiprotein form of DNA primase-DNA polymerase alpha affects selection of initiation sites, the frequency at which the sites are chosen, and length of RNA primers.

Evolutionarily selected replication origins: functional aspects and structural organization

A selective replicative pressure occurs during the evolution of simian virus 40 variants. When the replication origin is duplicated as an inverted repeat, there is a dramatic enhancement of replication. Having regulatory sequences located between the inverted repeat of ori magnifies their enhancing effect on replication. A passage 20 variant and a passage 45 variant containing three pairs of an inverted repeat of ori replicated more efficiently than a passage 13 variant containing nine copies of ori arranged in tandem. A 69-base-pair cellular sequence inserted between inverted repeats of ori of both passage 40 and 45 variants enhanced simian virus 40 DNA replication. Differences in replication efficiencies became greater as the total number of replicating species was increased in the transfection mixture, under conditions where T antigen is limiting. In a competitive environment, sequences flanking the replication origin may be inhibitory to replication.

Simian virus 40 DNA replication: functional organization of regulatory elements

The efficiency of simian virus 40 (SV40) DNA replication is dependent on the structural organization of the regulatory region. The enhancing effect of the G + C-rich 21-base-pair (bp) repeats on SV40 DNA replication is position and dose dependent and to some extent orientation dependent. The inverted orientation is about 50% as effective as the normal orientation of the 21-bp repeat region. Movement of the 21-bp repeat region 180 or 370 bp upstream of the ori sequence abolishes its enhancing effect, whereas no replication is detected if the 21-bp repeat region is placed downstream of the ori sequence. The dose-dependent enhancement of the 21-bp repeat of SV40 DNA replication as first described in single transfection by Bergsma et al. (D. J. Bergsma, D. M. Olive, S. W. Hartzell, and K. N. Subramanian, Proc. Natl. Acad. Sci. USA 79:381-385, 1982) is dramatically amplified in mixed transfection. In the presence of the 21-bp repeat region, the 72-bp repeat region can enhance SV40 DNA replication. In the presence of the 21-bp repeats and a competitive environment, the 72-bp repeat region exhibits a cis-acting inhibitory effect on SV40 DNA replication.

Domain structure of the simian virus 40 core origin of replication

The simian virus 40 core origin of replication consists of nucleotides 5211 through 31. These 64 base pairs contain three functional domains with strict sequence requirements and two spacer regions with relaxed sequence specificity but precise positional constraints. The early domain extends for 10 contiguous base pairs between nucleotides 5211 and 5220. A 9-base pair spacer from sequences 5221 through 5229 separates the early domain from the 23-base pair central palindrome that directs the binding of T antigen. The late end of the core between nucleotides 12 and 31 also contains spacer and sequence-specific functions that are not yet completely mapped. We propose that the sequence-specific domains are interaction sites for viral and cellular proteins, determinants of DNA conformation, or both. The spacers would position these signals at required distances and rotations relative to one another.

Functional organization of the simian virus 40 origin of DNA replication

To define the sequence elements involved in initiation of DNA synthesis at the simian virus 40 origin of replication, we determined the relative replication efficiencies in vitro and in vivo of templates containing a variety of mutations within the origin region. Replication of the mutants in vitro was assayed by the cell-free DNA replication system that we recently described (J.J. Li and T.J. Kelly, Proc. Natl. Acad. Sci. USA 81:6973-6977, 1984; J.J. Li and T.J. Kelly, Mol. Cell. Biol. 5:1238-1246, 1985), and replication in vivo was assayed after transfection of the mutant templates into COS-1 cells. The minimal origin of replication defined by both assays included a 15-base-pair (bp) imperfect inverted repeat, a 27-bp perfect inverted repeat, and a 17-bp A/T-rich region. T-antigen binding site I was not required for DNA replication, but its presence increased replication efficiency severalfold both in vitro and in vivo. Although SP1 binding sites and enhancers had little or no effect on replication in vitro, the presence of either element markedly increased replication in vivo. Thus, the biological role of these elements is not restricted to stimulating transcription but may be more general.

DNA binding site for a factor(s) required to initiate simian virus 40 DNA replication

Efficient initiation of DNA replication in the absence of nonspecific DNA repair synthesis was obtained by using a modification of the system developed by J.J. Li and T.J. Kelly [(1984) Proc. Natl. Acad. Sci. USA 81, 6973-6977]. Circular double-stranded DNA plasmids replicated in extracts of CV-1 cells only when the plasmids contained the cis-acting origin sequence for simian virus 40 DNA replication (ori) and the extract contained simian virus 40 large tumor antigen. Competition between plasmids containing ori and plasmids carrying deletions in and about ori served to identify a sequence that binds the rate-limiting factor(s) required to initiate DNA replication. The minimum binding site (nucleotides 72-5243) encompassed one-half of the simian virus 40 ori sequence that is required for initiation of replication (ori-core) plus the contiguous sequence on the late gene side of ori-core containing G + C-rich repeats that facilitates initiation (ori-auxiliary). This initiation factor binding site was specific for the simian virus 40 ori region, even though it excluded the high-affinity large tumor antigen DNA binding sites.

Expression of the late gene of simian virus 40 under the control of the simian virus 40 early-region promoter in monkey and mouse cells

We constructed a recombinant plasmid (pVNR4) with the simian virus 40 (SV40) early promoter positioned 30 nucleotides upstream from the major SV40 late transcription initiation site at residue 325. After transfection of the recombinant plasmid DNA into COS and mouse L cells, the transcripts of the SV40 late region were analyzed by S1 nuclease and primer extension analysis. The following are the principal findings. (i) The 16S and 19S late RNAs used the characteristic wild-type splice; no detectable levels of 19S unspliced RNA were observed. (ii) The majority of the late RNAs were heterogeneous and initiated in the early region (upstream and downstream from the Hogness-Goldberg sequence), and a minor population initiated at residue 325, the principal 5' terminus of the wild-type late RNA. (iii) During SV40 lytic infection there was a shift in initiation sites used to transcribe the early region from sites that are downstream to sites which are upstream (up RNA) of the origin of DNA replication. We observed that unlike lytic infection, T antigen and viral DNA replication were not needed for the appearance of up RNA in mouse L cells. (iv) In mouse L cells late RNAs were made, and the residue 325 5' end was utilized in the absence of T antigen or DNA replication. (v) In COS cells we found down RNA and up RNA transcribed from the extrachromosomally replicating plasmid but only down RNA produced by the integrated SV40 genome.

Functional interactions of the simian virus 40 core origin of replication with flanking regulatory sequences

We constructed a matched set of plasmids to investigate the interactions of essential core sequences of the simian virus 40 replication origin with flanking regulatory sequences. Deletions of either T-antigen-binding region I or the 21-base-pair repeated promoter elements reduced replication to 50 to 70% of wild-type levels. The simultaneous deletion of both regions decreased replication to less than 5% of wild-type levels. Thus, the double deletion greatly amplified the defects of the single deletions. We conclude that region I and the 21-base-pair repeats have related rather than independent functions in DNA synthesis. Insertion of a synthetic region I or the adenovirus 2 major late promoter at the late side of isolated core sequences in place of the 21-base-pair repeats failed to restore replication. In contrast, insertion of a single 72-base-pair enhancer element stimulated replication of the core origin more than fivefold. Thus, three distinct regulatory elements appear to facilitate core DNA replication by related mechanisms. Flanking sequences have only a small direct effect on T-antigen binding to naked core DNA. Possible mechanisms of action include the regulation of transcription or of chromatin structure.

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.

T antigen and template requirements for SV40 DNA replication in vitro

A cell-free system for replication of SV40 DNA was used to assess the effect of mutations altering either the SV40 origin of DNA replication or the virus-encoded large tumor (T) antigen. Plasmid DNAs containing various portions of the SV40 genome that surround the origin of DNA replication support efficient DNA synthesis in vitro and in vivo. Deletion of DNA sequences adjacent to the binding sites for T antigen either reduce or prevent DNA synthesis. This analysis shows that sequences that had been previously defined by studies in vivo to constitute the minimal core origin sequences are also necessary for DNA synthesis in vitro. Five mutant T antigens containing amino acid substitutions that affect SV40 replication have been purified and their in vitro properties compared with the purified wild-type protein. One protein is completely defective in the ATPase activity of T antigen, but still binds to the origin sequences. Three altered proteins are defective in their ability to bind to origin DNA, but retain ATPase activity. Finally, one of the altered T antigens binds to origin sequences and contains ATPase activity and thus appears like wild-type for these functions. All five proteins fail to support SV40 DNA replication in vitro. Interestingly, in mixing experiments, all five proteins efficiently compete with the wild-type protein and reduce the amount of DNA replication. These data suggest that an additional function of T antigen other than origin binding or ATPase activity, is required for initiation of DNA replication.

The distribution and properties of RNA primed initiation sites of DNA synthesis at the replication origin of Escherichia coli chromosome

RNA-linked DNA molecules were obtained from E. coli dnaCts cells synchronously initiating a new round of chromosome replication. The deoxynucleotides at the transition from primer RNA to DNA were 32P-labeled, and their positions were located on the nucleotide sequence of 1.4 kb genomic region (position -906 to +493) including the oriC and its leftside flanking region. In the r-strand (the counterclockwise strand), many strong transition sites were mapped in the left half portion of the oriC and a few weak sites in the left outside region. In the 1-strand (the clockwise strand), no transition sites were found inside the oriC but many weak sites were found in the left outside region. The results support the initiation mechanism in which the first leading strand synthesis starts with the r-strand counterclockwise from the oriC that is followed by the 1-strand synthesis on the displaced template strand on the left of oriC. Primer RNA molecules attached to the strong r-strand transition sites were only a few residues in length. Properties of the transition sites were discussed.

Replication and supercoiling of simian virus 40 DNA in cell extracts from human cells

Soluble extracts prepared from the nucleus and cytoplasm of human 293 cells are capable of efficient replication and supercoiling of added DNA templates that contain the origin of simian virus 40 replication. Extracts prepared from human HeLa cells are less active than similarly prepared extracts from 293 cells for initiation and elongation of nascent DNA strands. DNA synthesis is dependent on addition of purified simian virus 40 tumor (T) antigen, which is isolated by immunoaffinity chromatography of extracts from cells infected with an adenovirus modified to produce large quantities of this protein. In the presence of T antigen and the cytoplasmic extract, replication initiates at the origin and continues bidirectionally. Initiation is completely dependent on functional origin sequences; a plasmid DNA containing an origin mutation known to affect DNA replication in vivo fails to replicate in vitro. Multiple rounds of DNA synthesis occur, as shown by the appearance of heavy-heavy, bromodeoxyuridine-labeled DNA products. The products of this reaction are resolved, but are relaxed, covalently closed DNA circles. Addition of a nuclear extract during DNA synthesis promotes the negative supercoiling of the replicated DNA molecules.

DNA primase-DNA polymerase alpha from simian cells: sequence specificity of initiation sites on simian virus 40 DNA

Unique single-stranded regions of simian virus 40 DNA, phage M13 virion DNA, and several homopolymers were used as templates for the synthesis of (p)ppRNA-DNA chains by CV-1 cell DNA primase-DNA polymerase alpha. Intact RNA primers, specifically labeled with an RNA capping enzyme, were typically 6 to 8 ribonucleotides long, although their lengths ranged from 1 to 9 bases. The fraction of intact RNA primers 1 to 4 ribonucleotides long was 14 to 73%, depending on the template used. RNA primer length varied among primers initiated at the same nucleotide, as well as with primers initiated at different sites. Thus, the size of an RNA primer depended on template sequence. Initiation sites were identified by mapping 5' ends of nascent RNA-DNA chains on the template sequence, identifying the 5'-terminal ribonucleotide, and partially sequencing one RNA primer. A total of 56 initiation events were identified on simian virus 40 DNA, an average of 1 every 16 bases. Some sites were preferred over others. A consensus sequence for initiation sites consisted of either 3'-dCTTT or 3'-dCCC centered within 7 to 25 pyrimidine-rich residues; the 5' ends of RNA primers were complementary to the dT or dC. High ATP/GTP ratios promoted initiation of RNA primer synthesis at 3'-dCTTT sites, whereas low ATP/GTP ratios promoted initiation at 3'-dCCC sites. Similarly, polydeoxythymidylic acid and polydeoxycytidylic acid were the only effective homopolymer templates. Thus, both template sequence and ribonucleoside triphosphate concentrations determine which initiation sites are used by DNA primase-DNA polymerase alpha. Remarkably, initiation sites selected in vitro were strikingly different from initiation sites selected during simian virus 40 DNA replication in vivo.

Mouse DNA polymerase alpha-primase terminates and reinitiates DNA synthesis 2-14 nucleotides upstream of C2A1-2(C2-3/T2) sequences on a minute virus of mice DNA template

The distribution of termination and initiation sites in a 5081-nucleotide minute virus of mice DNA template being copied by a highly purified mouse DNA polymerase alpha-DNA primase complex in the presence of GTP has been examined. The 3'-hydroxyl termini (17 in all) were clustered at six sites that were located 2-14 nucleotides upstream of C2A2C2, C2AC3, or C2A2T2 sequences. When either [alpha-32P]- or [gamma-32P]GTP was included in the DNA polymerase reaction mixtures, nascent DNA became radiolabeled. Analysis of the 32P-labeled material following treatment of the DNA with tobacco acid pyrophosphatase, bacterial alkaline phosphatase, or ribonuclease T1 revealed the presence of oligoribonucleotide chains averaging 5-7 nucleotides long and beginning with 5' GTP residues. Eight presumptive DNA primase initiation sites were located opposite C4 or C5 sequences 3-9 nucleotides upstream of one of the three closely related hexanucleotides C2A2C2, C2AC3, and C2A2T2. RNA-DNA junctions were found 3-10 nucleotides downstream of DNA primase initiation sites. The results indicate that hexanucleotides having the general formula C2A1-2(C2-3/T2), herein referred to as psi, are involved in promoting termination of DNA synthesis and/or de novo initiation of RNA-primed DNA chains by DNA polymerase alpha-primase.

Method of mapping DNA replication origins

We have developed a method which allows determination of the direction in which replication forks move through segments of chromosomal DNA for which cloned probes are available. The method is based on the facts that DNA restriction fragments containing replication forks migrate more slowly through agarose gels than do non-fork-containing fragments and that the extent of retardation of the fork-containing fragments is a function of the extent of replication. The procedure allows the identification of DNA replication origins as sites from which replication forks diverge. In this paper we demonstrate the feasibility of this procedure, with simian virus 40 DNA as a model, and we discuss its applicability to other systems.

DNA primase-DNA polymerase alpha from simian cells: sequence specificity of initiation sites on simian virus 40 DNA

Unique single-stranded regions of simian virus 40 DNA, phage M13 virion DNA, and several homopolymers were used as templates for the synthesis of (p)ppRNA-DNA chains by CV-1 cell DNA primase-DNA polymerase alpha. Intact RNA primers, specifically labeled with an RNA capping enzyme, were typically 6 to 8 ribonucleotides long, although their lengths ranged from 1 to 9 bases. The fraction of intact RNA primers 1 to 4 ribonucleotides long was 14 to 73%, depending on the template used. RNA primer length varied among primers initiated at the same nucleotide, as well as with primers initiated at different sites. Thus, the size of an RNA primer depended on template sequence. Initiation sites were identified by mapping 5' ends of nascent RNA-DNA chains on the template sequence, identifying the 5'-terminal ribonucleotide, and partially sequencing one RNA primer. A total of 56 initiation events were identified on simian virus 40 DNA, an average of 1 every 16 bases. Some sites were preferred over others. A consensus sequence for initiation sites consisted of either 3'-dCTTT or 3'-dCCC centered within 7 to 25 pyrimidine-rich residues; the 5' ends of RNA primers were complementary to the dT or dC. High ATP/GTP ratios promoted initiation of RNA primer synthesis at 3'-dCTTT sites, whereas low ATP/GTP ratios promoted initiation at 3'-dCCC sites. Similarly, polydeoxythymidylic acid and polydeoxycytidylic acid were the only effective homopolymer templates. Thus, both template sequence and ribonucleoside triphosphate concentrations determine which initiation sites are used by DNA primase-DNA polymerase alpha. Remarkably, initiation sites selected in vitro were strikingly different from initiation sites selected during simian virus 40 DNA replication in vivo.

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.

Method of mapping DNA replication origins

We have developed a method which allows determination of the direction in which replication forks move through segments of chromosomal DNA for which cloned probes are available. The method is based on the facts that DNA restriction fragments containing replication forks migrate more slowly through agarose gels than do non-fork-containing fragments and that the extent of retardation of the fork-containing fragments is a function of the extent of replication. The procedure allows the identification of DNA replication origins as sites from which replication forks diverge. In this paper we demonstrate the feasibility of this procedure, with simian virus 40 DNA as a model, and we discuss its applicability to other systems.

The role of palindromic and non-palindromic sequences in arresting DNA synthesis in vitro and in vivo

The nature of specific DNA sequences that arrest synthesis by mammalian DNA polymerase alpha in vitro was analyzed using circular, single-stranded M13 or phi X174 virion DNA templates annealed to a unique, terminally labeled, DNA primer. This method rigorously defined both the starting nucleotide position and the direction of synthesis, as well as making the amount of radioactivity proportional to the number rather than the length of nascent DNA chains. The precise nucleotide locations of arrest sites were determined over templates with complementary sequences by cloning unique DNA restriction fragments into M13 DNA and isolating virions containing either the Watson or Crick strand. Results were correlated with the locations of palindromic (self-complementary) sequences, repeated sequences, and repeated sequences with mirror-image orientation. Two classes of DNA synthesis arrest sites were identified, distinct in structure but equivalent in activity. Class I sites consisted of palindromic sequences that formed a stable hairpin structure in solution and arrested DNA polymerase on both complementary templates. The polymerase stopped precisely at the base of the duplex DNA stem, regardless of the direction from which the enzyme approached. Class II sites consisted of non-palindromic sequences that could not be explained by either secondary structure or sequence symmetry elements, and whose complementary sequence was not an arrest site. Size limits, orientation and some sequence specificity for arrest sites were suggested by the data. Arrest sites were also observed in vivo by mapping the locations of 3'-end-labeled nascent simian virus 40 DNA strands throughout the genome. Arrest sites closest to the region where termination of replication occurs were most pronounced, and the locations of 80% of the most prominent sites appeared to be recognized by alpha-polymerase on the same template in vitro. However, class I sites were not identified in vivo, suggesting that palindromic sequences do not form hairpin structures at replication forks.

Unusual regulation of simian virus 40 early-region transcription in genomes containing two origins of DNA replication

As part of our efforts to create multifunctional vectors for the transduction of animal cells, a set of simian virus 40 recombinants were constructed which contain an inverted duplication of the region including the origin of viral DNA replication (ori) and the early-region promoter. The unusual aspects of the structure of these recombinant genomes revealed several unexpected features of their function. In particular, transcription from the early-region promoters on these recombinants occurred primarily after the start of DNA replication, and, in that sense, these promoters behaved as if they were late-region promoters. This behavior results from the fact that these genomes contain multiple ori segments, and, therefore, they replicate earlier and faster than wild-type virus DNA, thereby causing a precocious shift in the initiation of early-region transcription from sites downstream of ori to sites located upstream of ori. The abnormal expression from multiple ori genomes is consistent with our present notions regarding the replication-dependent shift in early-region transcriptional start sites (Buchman et al., Mol. Cell. Biol. 4:1900-1914). Since our experiments demonstrate that RNAs initiated upstream of ori contribute to T-antigen formation late in infection, we suggest that the shift in early-region transcription starts modulates large T-antigen production in concert with viral DNA replication.

Complex regulation of simian virus 40 early-region transcription from different overlapping promoters

During simian virus 40 lytic infection there is a shift in initiation sites used to transcribe the early region, which encodes large T and small t antigens. Early in infection, transcription is initiated almost exclusively from sites that are downstream of the origin of DNA replication, whereas transcripts produced later are initiated mainly from sites on the upstream side. We have used mutant virus and specially constructed plasmid DNAs to investigate the factors regulating this transcriptional shift. In our studies simian virus 40 large T antigen appears to mediate the shift in transcription in two ways: first, T antigen represses transcription at the downstream sites late in infection by binding to the region where these RNAs are initiated; second, T antigen promotes transcription from sites on the upstream side by its ability to initiate replication or amplification, or both, of the template DNA. In addition, transcription from the downstream sites is heavily dependent on enhancer sequences located in the 72-base-pair repeat region, whereas transcription from the upstream sites late in infection does not require enhancer sequences. Thus, different overlapping promoters regulate simian virus 40 early-region expression in a manner that apparently coordinates the production of large T antigen with the increase in viral DNA.

Multiple initiation sites of DNA replication flanking the origin region of lambda dv genome

Early replicative intermediates of lambda dv plasmid were prepared by an in vitro replication system in the presence of 2',3'-dideoxycytidine 5'-triphosphate, an inhibitor of DNA chain elongation. Short-chain DNAs produced from regions near the replication origin were purified from the intermediates. A fraction of the DNAs was covalently linked to primer RNA. The transition sites from primer RNA to DNA synthesis were mapped along the nucleotide sequence of the genome, by eliminating the RNA by alkaline hydrolysis and labeling the freshly exposed 5' ends of DNA with 32P. The transition sites were found to be located on both sides of the ori region, which includes four 19-base-pair repeats where one of the lambda specific initiator proteins, O, binds. No transition arose within the ori region. The transition sites are multiple on both sides of the ori region and are clustered in one of the two strands in such a way that DNA syntheses from the two sides converge. The frequency of the "leftward" DNA synthesis is several times higher than that of "rightward" synthesis, reflecting the asymmetric bidirectional replication of lambda dv DNA.

DNA replication and chromatin structure of simian virus 40 insertion mutants

Insertion of DNA segments into the nuclease-sensitive region of simian virus 40 alters both replication efficiency and chromatin structure. Mutants containing large insertions between the simian virus 40 origin of replication (ori site) and the 21-base-pair repeated sequences replicated poorly when assayed by transfection into COS-1 cells. Replication of mutants with shorter insertions was moderately reduced. This effect was cis-acting and independent of the nucleotide sequence of the insert. The nuclease-sensitive chromatin structure was retained in these mutants, but the pattern of cleavage sites was displaced in the late direction from the ori site. New cleavage sites appeared within the inserted sequences, suggesting that information specifying the nuclease-sensitive chromatin structure is located on the late side of the inserts. Accessibility to BglI (which cleaves within the ori site) was reduced in the larger insertion mutants. These results support the conclusion that efficient function of the viral origin of replication is correlated with its proximity to an altered chromatin structure.

Discontinuous DNA replication of Drosophila melanogaster is primed by octaribonucleotide primer

To investigate the precise structure of eucaryotic primer RNA made in vivo, short DNA chains isolated from nuclei of Drosophila melanogaster embryos were analyzed. Post-labeling of 5' ends of short DNA chains with polynucleotide kinase and [gamma-32P]ATP revealed that 7% of the DNA fragments were covalently linked with mono- to octaribonucleotide primers at their 5' ends. Octaribonucleotides, the major component (ca. 30%), formed the cap structure in the reaction with vaccinia guanylyltransferase and [alpha-32P]GTP, indicating that they were the intact primer RNA with tri- (or di-) phosphate termini, and the shorter ribooligomers were degradation intermediates. The intact primers started with purine (A/G ratio, 4:1), and the starting few ribonucleotide residues were rich in A.

S1-hypersensitive sites in eukaryotic promoter regions

We have examined by fine mapping the S1 nuclease-hypersensitivity of the 5' flanking regions of the human beta-globin and rat preproinsulin II genes and of the SV40 origin/enhancer region. In all cases S1-hypersensitive sites are located in known or presumed promoter/regulatory regions. Though a consensus DNA sequence is not evident, all of these sites reside in predominantly homopurine-homopyrimidine stretches. The alternate (non-B) DNA structure which is revealed by the enzymatic probe is a sequence-dependent feature of a short stretch of DNA, which is retained upon transplantation into a foreign environment. The alternate structure exhibits S1-nicking patterns uniquely different from those associated with the presence of Z-DNA.

Specific in vitro adenylylation of the simian virus 40 large tumor antigen

Incubation of the simian virus 40 (SV40) large tumor antigen (T) from either transformed or lytically infected cells with adenosine [8-3H]-, [alpha-32P]-, or [alpha-[35S]thio]-triphosphate in the presence of Mg2+ resulted in its labeling as defined by the appearance of an intact, appropriately immunoreactive band in NaDodSO4/polyacrylamide gels. Radioactivity remained associated with the protein after boiling in buffer containing 3% NaDodSO4, and 2-mercaptoethanol as well as after heating in 0.1 M HCl, 0.1 M NH4OH, or hydroxylamine, but it was dissociated after incubation in 0.1 M NaOH at 37 degrees C. After limited boiling of gel-purified [alpha-32P] ATP + T complex in 5.6 M HCl, o-[32P]phosphoserine was released, and snake venom phosphodiesterase or 0.5 M piperidine treatment of such a complex resulted in the liberation of [alpha-32P]AMP. The reaction proceeded when either purified, soluble T or insoluble, specifically immunoprecipitated antigen was used as substrate. ATP and dATP were the preferred nucleotide substrates by comparison with the other six standard ribonucleoside or deoxynucleoside triphosphates. Partial tryptic digests of T + [alpha-32P]ATP complexes revealed the presence of a single labeled peptide of Mr approximately equal to 12 - 14 X 10(3), and after exhaustive digestion, there was a single radioactive spot in the fingerprint. These data indicate that T can be adenylylated at a specific seryl residue(s) in a limited portion of the protein surface. Furthermore, adenylylation appears to be reversible and to proceed by a pyrophosphorylytic mechanism, since the nucleotide was released from the protein following incubation of adenylylated T with Mg2+, sodium pyrophosphate, and poly(dT).

Unusual regulation of simian virus 40 early-region transcription in genomes containing two origins of DNA replication

As part of our efforts to create multifunctional vectors for the transduction of animal cells, a set of simian virus 40 recombinants were constructed which contain an inverted duplication of the region including the origin of viral DNA replication (ori) and the early-region promoter. The unusual aspects of the structure of these recombinant genomes revealed several unexpected features of their function. In particular, transcription from the early-region promoters on these recombinants occurred primarily after the start of DNA replication, and, in that sense, these promoters behaved as if they were late-region promoters. This behavior results from the fact that these genomes contain multiple ori segments, and, therefore, they replicate earlier and faster than wild-type virus DNA, thereby causing a precocious shift in the initiation of early-region transcription from sites downstream of ori to sites located upstream of ori. The abnormal expression from multiple ori genomes is consistent with our present notions regarding the replication-dependent shift in early-region transcriptional start sites (Buchman et al., Mol. Cell. Biol. 4:1900-1914). Since our experiments demonstrate that RNAs initiated upstream of ori contribute to T-antigen formation late in infection, we suggest that the shift in early-region transcription starts modulates large T-antigen production in concert with viral DNA replication.

Complex regulation of simian virus 40 early-region transcription from different overlapping promoters

During simian virus 40 lytic infection there is a shift in initiation sites used to transcribe the early region, which encodes large T and small t antigens. Early in infection, transcription is initiated almost exclusively from sites that are downstream of the origin of DNA replication, whereas transcripts produced later are initiated mainly from sites on the upstream side. We have used mutant virus and specially constructed plasmid DNAs to investigate the factors regulating this transcriptional shift. In our studies simian virus 40 large T antigen appears to mediate the shift in transcription in two ways: first, T antigen represses transcription at the downstream sites late in infection by binding to the region where these RNAs are initiated; second, T antigen promotes transcription from sites on the upstream side by its ability to initiate replication or amplification, or both, of the template DNA. In addition, transcription from the downstream sites is heavily dependent on enhancer sequences located in the 72-base-pair repeat region, whereas transcription from the upstream sites late in infection does not require enhancer sequences. Thus, different overlapping promoters regulate simian virus 40 early-region expression in a manner that apparently coordinates the production of large T antigen with the increase in viral DNA.

DNA replication and chromatin structure of simian virus 40 insertion mutants

Insertion of DNA segments into the nuclease-sensitive region of simian virus 40 alters both replication efficiency and chromatin structure. Mutants containing large insertions between the simian virus 40 origin of replication (ori site) and the 21-base-pair repeated sequences replicated poorly when assayed by transfection into COS-1 cells. Replication of mutants with shorter insertions was moderately reduced. This effect was cis-acting and independent of the nucleotide sequence of the insert. The nuclease-sensitive chromatin structure was retained in these mutants, but the pattern of cleavage sites was displaced in the late direction from the ori site. New cleavage sites appeared within the inserted sequences, suggesting that information specifying the nuclease-sensitive chromatin structure is located on the late side of the inserts. Accessibility to BglI (which cleaves within the ori site) was reduced in the larger insertion mutants. These results support the conclusion that efficient function of the viral origin of replication is correlated with its proximity to an altered chromatin structure.

Discontinuous DNA replication of Drosophila melanogaster is primed by octaribonucleotide primer

To investigate the precise structure of eucaryotic primer RNA made in vivo, short DNA chains isolated from nuclei of Drosophila melanogaster embryos were analyzed. Post-labeling of 5' ends of short DNA chains with polynucleotide kinase and [gamma-32P]ATP revealed that 7% of the DNA fragments were covalently linked with mono- to octaribonucleotide primers at their 5' ends. Octaribonucleotides, the major component (ca. 30%), formed the cap structure in the reaction with vaccinia guanylyltransferase and [alpha-32P]GTP, indicating that they were the intact primer RNA with tri- (or di-) phosphate termini, and the shorter ribooligomers were degradation intermediates. The intact primers started with purine (A/G ratio, 4:1), and the starting few ribonucleotide residues were rich in A.

Essential nucleotides in the polyomavirus origin region

We have isolated a series of point mutants in the polyomavirus origin-intergenic control region by using a procedure which exploits the single-stranded nature of DNAs cloned into M13 phage both for the generation of mutants and for their analysis by DNA sequencing. In this report we describe the effects of cytosine X guanine----thymine X adenine base-pair substitutions in the polyomavirus origin region upon replication in mouse 3T6 cells of the M13-polyomavirus constructs. Our results indicate that sequences near the center of a 34-base-pair sequence with dyad symmetry are important for replication, whereas specific nucleotides near the ends of the dyad symmetry are not important. Furthermore, a putative large T antigen-binding site at nucleotides 60 to 80 can be mutated without affecting replication function as measured in this assay.

Critical spatial requirement within the origin of simian virus 40 DNA replication

We inserted a single base pair into the center of a 27-base-pair palindrome within the replication origin of simian virus 40. The mutation did not directly alter the symmetry of the palindrome or the protein-binding sequences within the palindrome. DNA binding studies showed that subunits of the simian virus 40 A protein (T antigen) bound to each of the four recognition pentanucleotides in the origin palindrome but did so with reduced affinity in comparison with wild-type origins. The mutant origin cloned in a plasmid DNA failed to replicate in COS cells. Thus, precise spatial interactions among subunits of A protein are necessary for stable origin binding and are crucial for subsequent steps in the initiation of DNA replication. Furthermore, any possible functional interactions of the simian virus 40 A protein with cellular DNA would require a great fidelity of protein binding arrangements to initiate cellular DNA replication.

Replication of adenovirus mini-chromosomes

We have isolated adenovirus origins of DNA replication from both the right and left ends of the genome, which are functional on linear autonomously replicating mini-chromosomes. The mini-chromosomes contain two cloned inverted adenovirus termini and require non-defective adenovirus as a helper. Replicated molecules are covalently attached to protein, and DNA synthesis is initiated at the correct nucleotide even when the origins are not located at molecular ends. The activity of embedded origins leads to the generation of linear mini-chromosomes from circular or linear molecules. These observations therefore suggest that sequences within the adenovirus origin of replication position the protein priming event at the adenovirus terminus. Experiments investigating the regeneration of deleted viral inverted terminal repeat sequences show a sequence-independent requirement for inverted sequences in this process. This result strongly suggests that repair results from the formation of a panhandle structure by a displaced single strand. On the basis of these observations we propose a model for the generation of adenovirus mini-chromosomes from larger molecules.

A cis-acting element from the Epstein-Barr viral genome that permits stable replication of recombinant plasmids in latently infected cells

The Epstein-Barr viral (EBV) genome of approximately equal to 170 kilobase pairs (kbp) is maintained as a plasmid in human B lymphoblasts transformed by the virus. We have identified a cis-acting element within 1.8 kbp of the viral genome that allows recombinant plasmids carrying it to be selected at high frequency and maintained as plasmids in cells latently infected by EBV. This functional element(s) requires a segment of DNA at least 800 bp and at most 1800 bp long, which contains a family of 30-bp tandem repeats at one end. Since this region confers efficient stable replication only to plasmids transfected into cells containing EBV genomes, its function probably requires trans-acting products encoded elsewhere in the viral genome.

Sequence specificity for the initiation of RNA-primed simian virus 40 DNA synthesis in vivo

Analysis of the nucleotide sequences at the 5' ends of RNA-primed nascent DNA chains (Okazaki fragments) and of their locations in replicating simian virus 40 (SV40) DNA revealed the precise nature of Okazaki fragment initiation sites in vivo. The primary initiation site for mammalian DNA primase was 3'-purine-dT-5' in the DNA template and the secondary site was 3'-purine-dC-5', with the 5' end of the RNA primer complementary to either the dT or dC. The third position of the initiation site was variable with a preference for dT or dA. About 81% of the available 3'-purine-dT-5' sites and 20% of the 3'-purine-dC-5' sites were used. Purine-rich sites, such as PuPuPu and PyPuPu , were excluded. The 5'-terminal ribonucleotide composition of Okazaki fragments corroborated these conclusions. Furthermore, the length of individual RNA primers was not unique, but varied in size from six to ten bases with some appearing as short as three bases and some as long as 12 bases, depending on the initiation site used. This result was consistent with the average size (9 to 11 bases) of RNA primers isolated from specific regions of the genome. Excision of RNA primers did not appear to stop at the RNA-DNA junction, but removed a variable number of deoxyribonucleotides from the 5' end of the nascent DNA chain. Finally, only one-fourth of the replication forks contained an Okazaki fragment, and the distribution of their initiation sites between the two arms revealed that Okazaki fragments were initiated exclusively (99%) on retrograde DNA templates. The data obtained at two genomic sites about 350 and 1780 bases from ori were essentially the same as that reported for the ori region (Hay & DePamphilis , 1982), suggesting that the mechanism used to synthesize the first DNA chain at ori is the same as that used to synthesize Okazaki fragments throughout the genome.

Monkey B-lymphotropic papovavirus DNA: nucleotide sequence of the region around the origin of replication

We have determined the nucleotide sequence of the HindIII-B DNA segment of African green monkey B-lymphotropic papovavirus (LPV), which shows a highly restricted host range and whose genome is a 5.1-kilobase-long circular DNA. The segment, consisting of 1,123 base pairs, contained the origin of DNA replication, the putative control region for early transcription, and the region probably coding for the amino-terminal portion of T antigens. The symmetrical region at the center of replication origin, 5'-GAGGC CA GGGGCCCC TA GCCTC-3' (on the L strand), has diverged in the central portion from the corresponding regions of primate polyomaviruses simian virus 40, BK virus, and JC virus, but resembles that of mouse polyomavirus. The structure of the control region upstream of the replication origin was unique to LPV and contained several repeated sequences, the longest of which were two 60-base-pair tandem repeats. The amino-terminal region common to LPV small T and large T antigens showed some homology (41%) in the deduced amino acid sequence to that of both simian virus 40 and the mouse polyomavirus. Like other polyomaviruses, the probable carboxyl-terminal region unique to LPV small T antigen contained two sets of the Cys-x-Cys-x-x-Cys structure. These data show that, despite the unique structures in the control region, LPV is evolutionally related to the mouse polyomavirus and to simian virus 40.

Mouse primase initiation sites in the origin region of simian virus 40

The sites of initiation of DNA synthesis by purified mouse DNA primase in the origin-of-replication region of simian virus 40 (SV40) were examined. Using as template the separated strands of a cloned fragment of SV40 approximately equal to 300 base pairs (bp) long that includes the origin, we observed specific sites of initiation on the two strands. On the early strand that is the template for early mRNA synthesis, the primary starts are at four positions within 10 nucleotides of each other around nucleotide 5215 and an additional site around nucleotide 5147 that is used at one-sixth the frequency of the major sites. The major start sites on the early strand are within the 65-bp minimal origin of replication and lie between tumor antigen binding sites I and II. On the late strand that is the template for late mRNA synthesis, six major initiation sites were observed, each within the 3' C-C-C-G-C-C 5' sequence in the template that is repeated twice within each of the three 21-bp repeats that lie adjacent to the minimal origin, on its late side. A 6-bp deletion in the 65-bp minimal origin that eliminates its function as an origin reduced the major initiations around nucleotide 5215 on the early strand by 90% but did not affect initiations at the minor start site on the early strand or initiations on the late strand. Mouse DNA primase is able to recognize specific regions on the SV40 DNA. Those on the early strand are within the minimal origin of replication and those on the late strand are within the 21-bp repeat region necessary for maximum replication.

Naturally arising recombinants that are missing portions of the simian virus 40 regulatory region

When simian virus 40 (SV40) is serially passaged at high multiplicity, a heterogeneous collection of naturally arising variants is generated. Those which are the most abundant presumably have a selective replicative advantage over other defective and wild-type helper SV40s. Two such naturally arising host-substituted variants of SV40 have been characterized in terms of complete nucleotide sequence determination. Evolutionary variant ev-1101 (previously isolated by Lee et al., Virology 66:53-69, 1975) is from undiluted serial passage 13, whereas ev-2101 is newly isolated from undiluted serial passage 6 of an independently-derived evolutionary series. Both variants contain a five-times tandemly repeated segment of DNA consisting of viral Hin C and Hin A sequences that have recombined with a segment of host DNA that is not highly reiterated in the monkey genome. The monkey segment differs in the two variants as does the size of the viral segment retained. In two additional host-substituted variants, ev-1102 (previously isolated from serial passage 20 by Brockman et al., Virology 54:384-397, 1973) and ev-1108 (newly isolated from serial passage 40), the SV40 sequences derived from the replication origin are present as inverted repetitions. The inverted repeat regions of these two variants have been analyzed at the nucleotide sequence level and are compared with SV40 variant ev-1104 from passage 45 (previously characterized by Gutai and Nathans, J. Mol. Biol. 126:259-274, 1978). The viral segment containing the regulatory signals for replication and viral gene expression is considerably shortened in later serial passages as demonstrated by these five variants. It is of interest that the variants presumably arose due to their enhanced replication efficiency, yet are missing some of the sequence elements implicated in the regulation of replication. Furthermore, a comparison of the structure of the replication origin regions indicates that additional changes occur in the SV40 regulatory region with continued undiluted serial passage.

Binding of simian virus 40 a protein to DNA with deletions at the origin of replication

Previous studies with wild-type simian virus 40 DNA have shown that the sequence 5'-GAGGC-3' directs the binding of A protein (T antigen). The functional origin of replication contains four recognition pentanucleotides each of which is separated by a single base pair and arranged a two pairs of direct repetitions that are inverted relative to each other. Analysis of A protein binding to a series of nonviable mutants progressively deleting these contact sites leads to the following conclusions: (i) stable binding of subunits of A protein to three origin pentanucleotides is not sufficient for the initiation of DNA replication, (ii) the stability of DNA binding depends on interactions between bound protein subunits, and (iii) a single pentanucleotide is sufficient to bind and orient a subunit of A protein.