@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Medicine, Faculty of"@en, "Biochemistry and Molecular Biology, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Tam, Patrick"@en ; dcterms:issued "2009-04-15T21:14:58Z"@en, "1994"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Minute virus of mice (MVM) belongs to the Parvovirus genera of the Parvoviridae family of eukaryotic viruses. Its genome consists of approximately 5 kb of primarily single stranded negative sense DNA. Since the primary sequences at the genomic termini are palindromic, duplex hairpin structures are found at the termini. It is clear that parvoviral terminal hairpins are important for viral DNA replication. In order to determine the sequence requirements for MVM DNA replication, an in vivo DNA replication assay was developed. In this assay, two plasmids were required. First, the plasmid vector, pPTLR, was constructed such that it encoded a minigenome of MVM containing 411 nt and 807 nt of the left and right termini, respectively. This minigenome was designed to contain only the terminal sequences since the analysis of DNAs found within defective interfering (DI) particles suggested that the cis-acting sequences required for viral DNA synthesis are found at or near the termini (Faust and Ward, 1979). When pPTLR was co transfected into mouse LA9 cells or COS-7 cells with a second plasmid which expresses MVM NS-1, the major viral non-structural protein, the viral minigenome was rescued from the plasmid sequences and replicated in the host cell. The replicated DNA exhibited heterogeneous termini suggesting that both terminal hairpins were functional during viral DNA replication. Deletion analysis of the pPTLR minigenome suggested that in addition to previous studies which partially defined the requirements of the right end palindrome, two additional regions of the MVM genome are important for viral DNA replication. First, analysis of the left terminal hairpin suggested that an element(s) between MVM nucleotide position 11 and 25 (or 31) is important.. Although the nature of the left terminal hairpin allows minor deletion mutants to regenerate wild type terminal hairpins, it is hypothesized that the bubble sequence which, results from nucleotide positions 25-26 being mispaired with nucleotide positions 91-89 in the duplex hairpin, is altered during DNA replication. The potential effect of this change in the DNA template on viral DNA replication is discussed. Second, deletion analysis of the region internal of the right hairpin suggested that two adjacent elements, A (nucleotide positions 4489-4636) and B (nucleotide positions 4636-4695) are important for minigenome DNA replication. Two Rsa I restriction fragments which partially span elements A and B, Rsa A (nucleotide positions 4431-4579) and Rsa B (nucleotide positions 4579-4662), were used to probe nuclear extracts for sequence specific DNA binding proteins in electrophoretic mobility shift assays. A number of DNA-protein interactions were discovered. The binding site of one relatively abundant cellular factor, MVM DNA replication factor B5 or MRF B5, was determined. MRF B5 protected two regions of the Rsa B probe, site I (—nucleotide positions 4589-4610) and site II (—nucleotide positions 4616-4646), in DNase I footprinting assays. Although the function of MRF B5 in viral DNA replication is unknown, it is speculated that it might activate DNA replication at the right terminus. Analysis of the replication of identical end minigenome mutants showed that viral DNA can also be replicated using either two left termini (LL) or two right termini (RR). Replication of the RR minigenome was observed to be substantially greater that of the LL minigenome. However, it was demonstrated that the Xba I (nucleotide position 4342)/Sau3a (nucleotide position 4741) fragment, which contained elements A and B, could stimulate DNA replication of LL type minigenomes. These data confirm the finding that sequences at both termini encode origins of DNA replication."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/7128?expand=metadata"@en ; dcterms:extent "3681518 bytes"@en ; dc:format "application/pdf"@en ; skos:note "CIS-ACTING SEQUENCES FOUND WITHIN THE MVM GENOMEREQUIRED FOR DNA REPLICATIONbyPatrick TamB.Sc. (Hon), University of British Columbia, 1988A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDoctor of PhilosophyinTHE FACULTY OF GRADUATE STUDIESDepartment of Biochemistry and Molecular BiologyWe accept this thesis conforming to the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember, 1994© Patrick Tam, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)____Department of Biochemistry and Molecular BiologyThe University of British ColumbiaVancouver, CanadaDate Oct.5, 1994DE-6 (2/88)11AbstractMinute virus of mice (MVM) belongs to the Parvovirus genera of theParvoviridae family of eukaryotic viruses. Its genome consists of approximately 5kb of primarily single stranded negative sense DNA. Since the primary sequencesat the genomic termini are palindromic, duplex hairpin structures are found at thetermini. It is clear that parvoviral terminal hairpins are important for viral DNAreplication.In order to determine the sequence requirements for MVM DNA replication,an in vivo DNA replication assay was developed. In this assay, two plasmids wererequired. First, the plasmid vector, pPTLR, was constructed such that it encoded aminigenome of MVM containing 411 nt and 807 nt of the left and right termini,respectively. This minigenome was designed to contain only the terminalsequences since the analysis of DNAs found within defective interfering (DI)particles suggested that the cis-acting sequences required for viral DNA synthesisare found at or near the termini (Faust and Ward, 1979). When pPTLR was cotransfected into mouse LA9 cells or COS-7 cells with a second plasmid whichexpresses MVM NS-1, the major viral non-structural protein, the viral minigenomewas rescued from the plasmid sequences and replicated in the host cell. Thereplicated DNA exhibited heterogeneous termini suggesting that both terminalhairpins were functional during viral DNA replication.Deletion analysis of the pPTLR minigenome suggested that in addition toprevious studies which partially defined the requirements of the right endpalindrome, two additional regions of the MVM genome are important for viral111DNA replication. First, analysis of the left terminal hairpin suggested that anelement(s) between MVM nucleotide position 11 and 25 (or 31) is important..Although the nature of the left terminal hairpin allows minor deletion mutants toregenerate wild type terminal hairpins, it is hypothesized that the bubble sequencewhich, results from nucleotide positions 25-26 being mispaired with nucleotidepositions 91-89 in the duplex hairpin, is altered during DNA replication. Thepotential effect of this change in the DNA template on viral DNA replication isdiscussed. Second, deletion analysis of the region internal of the right hairpinsuggested that two adjacent elements, A (nucleotide positions 4489-4636) and B(nucleotide positions 4636-4695) are important for minigenome DNA replication.Two Rsa I restriction fragments which partially span elements A and B, Rsa A(nucleotide positions 4431-4579) and Rsa B (nucleotide positions 4579-4662), wereused to probe nuclear extracts for sequence specific DNA binding proteins inelectrophoretic mobility shift assays. A number of DNA-protein interactions werediscovered. The binding site of one relatively abundant cellular factor, MVM DNAreplication factor B5 or MRF B5, was determined. MRF B5 protected two regions ofthe Rsa B probe, site I (—nucleotide positions 4589-4610) and site II (—nucleotidepositions 4616-4646), in DNase I footprinting assays. Although the function of MRFB5 in viral DNA replication is unknown, it is speculated that it might activate DNAreplication at the right terminus.Analysis of the replication of identical end minigenome mutants showed thatviral DNA can also be replicated using either two left termini (LL) or two righttermini (RR). Replication of the RR minigenome was observed to be substantiallygreater that of the LL minigenome. However, it was demonstrated that the Xba I(nucleotide position 4342)/Sau3a (nucleotide position 4741) fragment, whichcontained elements A and B, could stimulate DNA replication of LL typeivminigenomes. These data confirm the finding that sequences at both terminiencode origins of DNA replication.VTable of ContentsPageAbstract iiTable of Content vList of Figures ixAbbreviations xiAcknowledgements xvI. Introduction 11.1. Review of Parvoviruses 21.1.1. General Characteristics 21.1.2. Classification of Parvoviruses 21.2. Molecular Biology of MVM 41.2.1. Historical Aspects 41.2.2. Genomic Organization of MVM(p) 51.2.3. Viral Transcripts 51.2.4. MVM(p) Polypeptides 71.2.5. Structure of MVM Terminal Hairpins 111.2.6. MVM DNA Replication 151.2.7. AAV DNA Replication 191.3. Origins of DNA Replication 201.3.1. SV4O Origin of DNA Replication 211.3.2. Bovine Papillomavirus Type 1 Origin of DNA Replication 271.3.3. Herpes Simplex 1 (HSV-1) and Epstein-Barr Virus (EBV) 28Origin of DNA Replicationvi1.3.4. The Yeast Chromosomal ARS1 Origin of DNA Replication 311.4. The Present Study 32II. Materials and Methods 332.1. Materials 332.2. Bacteria 332.3. Mammalian Cell Lines 342.3.1. COS-7 Cells 342.3.2. LA9 Cells 342.4. Basic Cloning Techniques 352.4.1. Isolation of Plasmid DNA 352.4.2. Isolation of DNA Fragments From Agarose Gels 352.4.3. Cloning of Restriction Fragments Into Plasmid Vectors 362.4.4. Preparation and Transformation of Competent Cells 362.5. DNA Sequencing 372.6. Plasmid Constructions 382.6.1. Construction of pCMVNS-1 382.6.2. Construction of pPTLR 392.6.3. Construction of Left (3’) Palindrome Deletions 392.6.4. Construction of Internal Left End (ILE) and Internal Right 40End (IRE) Deletions2.6.5. Construction of Minigenomes Containing Two Left Termini 41or Two Right Termini2.7. Transfection of DNA Into Mammalian Cells 422.8. Isolation of Low Molecular Weight DNA From Mammalian Cells 422.9. Use of Restriction Endonuclease, Exonuclease III and 51 43Nuclease to Characterize Replicated DNAvii2.10. Southern Blotting and Hybridization 442.11. Preparation of 32P labeled DNA Probes 452.12. Western Blot Analysis 452.13. Preparation of Nuclear Extracts 462.14. Fractionation of Nuclear Extracts 482.15. Electrophoretic Mobility Shift Assay 482.16. DNase I Footprinting Assay 492.17. Preparation of 32P labeled DNA Probes for DNA-Protein 49Interaction Studies2.17.1. Rsa A and Rsa B Restriction Fragments 492.18.2 Synthetic Oligonucleotides 50Ill. Results 513.1. Replication of a Defective LR Minigenome in COS-7 51and LA9 Cells3.2. Analysis of the Termini of LR Replicative Intermediates 593.3. Replication of Left Hand (3’) Terminal Deletions 643.4. Replication of Internal Right End (IRE) Deletions 723.5. Replication of Internal Left End (ILE) Deletions 863.6. Replication of Identical End Minigenomes 883.7. Activation of LL Minigenomes 983.8. Identification of DNA Binding Activites in Nuclear Extracts 105of LA9 and MVM(p) Infected LA9 Cells3.9. Fractionation of Nuclear Extracts 1093.10. Determination of MRF B5 DNA Binding Site by DNase I 110Footprintingvii’IV. Discussion 1194.1 Replication of MVM minigenomes 1194.2 Analysis of the Left Hand (3’) Terminal Deletion Mutants Suggest 120That Bubble Sequences Are Important For MVM DNA Replication4.3 Replication of Identical End Genomes 1244.4 Origins of DNA Replication Contain Multiple Cis-acting Elements 1274.5 Possible Roles For Cis-Acting Elements Found Inboard of the 129Right PalindromeV. Bibliography 134ixList of FiguresPage1. Parvoviridae Family 32. MVM(p) Transcription Map 63. Structure of the MVM(p) Termini 124. Modified Rolling Hairpin Model of MVM DNA Replication 155. Model of AAV DNA Replication 196. Organization of Several Eukaryotic Origins of DNA Replication 227. Schematic Diagram of pPTLR MVM Minigenome and pCMVNS-1 538. Replication of the pPTLR Minigenome 549. Expression of NS-1 in LA9 and COS-7 Cells 5810. Exonuclease Ill/SI Nuclease Digestion Patterns of the 60Minigenomic Replicative Intermediates11. Heat Denaturation/Quick Chill Properties of Minigenomic 61Replicative Intermediates12. Replication of the Left Hand (3’) Terminal Deletion Mutants 6513. Schematic Diagram of Internal Right End (IRE) and Internal 73Left End (ILE) Deletion Mutants14. Replication of Internal Right End (ILE) Deletion Mutants 7415. Schematic Diagram of DNA Fragments Used to Rescue the 77Replication of IRE Deletion Mutants16. Replication of IRE SXR and SXL Rescue Mutants 7817. Replication of IRE SVR Rescue Mutants 8018. Rsa A and Rsa B Activates DNA replication of the d1411-4695 83IRE Deletion Mutant19. Replication of Internal Left End (ILE) Deletion Mutants 8720. Replication of Identical End Genomes 90x21. Exo ITT/Si Nuclease Digestion Patterns and Heat 95Denaturation/Quick Chill Properties of the MinigenomicRR Replicative Intermediate22. Activation of DNA Replication of LL Minigenomes 9923. Exo III Digestion Patterns and Heat Denaturation/Quick 103Chill Properties of the Minigenomic LL Replicative Intermediate24. Cellular Proteins in Nuclear Extracts Bind to Rsa A and Rsa B 10725. Determination of MRF B5 Binding Site 11126. MRF B5 Can Saturate Rsa B in EMSA 11727. Competition EMSA Suggest MRF B3 and MRF B4 are Components 118of MRF B528. Possible Repair Mechanism of Left Hand (3’) Terminal Hairpin 122Deletions Mutants29. Organization of the Putative Replication Elements Found at 130the Right Terminus.xiAbbreviationsAAV adeno-associated virusACS ARS consensus sequenceAd2 adenovirus type 2ARS autonomously replicating sequenceATP adenosine 5’-triphosphateBPV-1 bovine papillomavirus type 1CAT chioramphenicol acetyl transferasecDNA complementary DNACMV cytomegaloviruscpm counts per minutedATP deoxyadenosine 5’-triphosphatedCTP deoxycytidine 5’-triphosphateDEAE diethylaminaoethyldGTP deoxyguanosine 5’-triphosphateDI defective interfering particlesdLL dimer LLdLR dimer LRdRF dimer replicative formDMEM Dulbecco’s modified Eagle mediumDMSO dimethyl sulphoxideDNA deoxyribonucleic acidDNase I deoxyribonuclease IdRR dimer RRDS dyad of symmetry elementxliDTT dithiothreitoldTTP deoxythymidine 5’-triphosphateDUE DNA unwinding elementsE.coli Escherichia coliEBNA-1 Epstein-Barr nuclear antigen IEBV Epstein-Barr virusEDTA ethylenediamine tetra-acetic acidEGTA ethylene glycol-bis(f-aminoethyl ether) N,N,N’,N’,tetraacetic acidEP early palindromeEtBr ethidium bromideexo III exonuclease IIIFR family of repeatsHEPES N-2-hydroxyethylpiperazine-N’-2ethanesulfonic acidHSV-l herpes simplex virus type 1ILE internal left endIP input plasmidIRE internal right endkb kilo base(s)kbp kilo base pair(s)kDa kilo dalton(s)mm minute(s)mLL monomer LLmLR monomer LRMRF MVM(p) replication factormRF monomer replicative formmRR monomer RRxliiMVM(i) minute virus of mice, immunosupressive strainMVM(p) minute virus of mice, prototypical strainNS non-structural proteinnt nucleotide(s)OD optical densityPBS phosphate buffered salinePCV packed cell volume(s)PMS plasmid maintenance sequencepol polymeraseRF replication factorRNA ribonucleic acidRNase ribonucleaserpm revolutions per minuteRT room temperatureRV rat virus-likeSDS sodium dodecylsulphateSSC standard saline citrateSSPE standard saline phosphate EDTASV4O simian virus 40T-ag 5V40 large tumour antigenTAE tris-acetate-EDTATBE tris-borate-EDTATris tris(hydroxymethylamino)methanetRNA transfer RNAU unit(s)UV ultravioletV voltsVP viral proteinxivxvAcknowledgementsFirst, I wish to thank Dr. Caroline Astell for the opportunity to work in herlaboratory. I also thank Dr. Michael Smith and Dr. Frank Tufaro for being on mysupervisory commitee. The research presented in this thesis would not have beenpossible without the constant encouragement of Amy, Jan, and the other currentmembers and ex-members of this lab. Special thanks go to Wesley and Brendan ofthe Sadowski lab for technical advice and late night talks. Last but not least, I wishto thank my parents for putting up with all the nonsense throughout the years.1I IntroductionStudies on the initiation and regulation of DNA replication in mammaliancells have been generally unyielding because of a lack of useful genetic tools andapproaches. However, the study of small defined genomes of some mammalianDNA viruses has allow detailed genetic and biochemical dissection of essentialcellular processes such as DNA replication and gene expression, albeit in thecontext of viral function. Well studied viruses such as simian virus 40 (SV4O) oradenovirus have served as relatively simple model systems for understanding suchcellular processes. Initially, the parvovirus minute virus of mice (MVM) wasthought to be another simple model system to study mechanisms of DNAreplication. Although the mechanism of MVM DNA replication is still not fullyunderstood, what is currently known shows that MVM requires complex reactionsat or near the sequences encoding the hairpin termini. Understanding themolecular mechanisms of MVM DNA replication may contribute to a fullerunderstanding of parvovirus replication as well as mammalian DNA replication. Aclear account of DNA replication and its regulation is essential to understanding thegrowth and development of organisms. In addition, this understanding may alsocontribute to the comprehension of disease states which arise from aberrant growthand replication of cells.21.1. Review of parvoviruses1.1.1. General CharacteristicsParvoviruses form a large family of physically similar viruses which infectboth invertebrate (insects) and vertebrate (mammals) animal species (Berns, 1990;Tijssen, 1990). A typical parvovirus virion is composed of 4.6 to 6.0 kb of singlestranded DNA encapsidated in an icosahedral coat 20-25 nm in diameter. Thegenomic DNAs of all parvoviruses analyzed to date have terminal palindromeswhich are able to form hairpin structures (Astell, 1990). The hairpin sequences ofseveral parvoviruses (MVM, AAV) are essential for viral replication.1.1.2. Classification of parvovirusesHistorically, parvoviruses have been divided into three genera: Parvovirus,Densovirus and Dependovirus (Tijssen, 1990) (Fig. 1). This classification system isbased on two criteria: requirement of a helper virus for viral replication and thehost species which is infected. Autonomous parvoviruses are vertebrate viruseswhich are able to replicate in the absence of a helper virus (Cotmore and Tattersall,1987). Originally, it was thought that autonomous parvoviruses have non-identicalterminal hairpins and package only minus sense DNA. It has since been shown thatsome autonomous parvovirus contain identical terminal palindrome and packageboth plus and minus sense DNA. Dependoviruses, (also called the adenoassociated viruses or AAV) (Berns and Bohenzky, 1986), are vertebrate viruseswhich were originally thought to3Parvoviridae FamilyA. Genus ParvovirusKilham rat virus (KRV)Minute virus of mice (MVM)LuIJI RV-likeH-iFeline parvovirus (FPV)Canine parvovirus (CPV)B-19Simian parvovirus (SPV)Bovine parvovirus (BPV)Aleution disease virus (ADV)B. Genus DependovirusAdeno-associated virus (AAV)C. Genus DensovirusesGalleria DNVBombyx DNVFig. 1. Parvoviridae Family.The family Parvoviridae comprises of three genera. Within the Parvovirus genus,MVM is classified in the RV-like subgroup whose members have similar capsidstructure characteristics. The Dependovirus genus consists of adeno-associatedviruses (AAVs) which is a group of serotypically distinct viruses. Most molecularstudies of AAVs have been carried out on the human AAV-2 serotype. TheDensovirus genus includes a number of viruses which infect insects.4require a helper virus for productive viral replication. It has since been shown thatAAV is able to replicate in cell lines in the absence of a helper virus. AAV genomestypically contain either plus and minus sense DNA with identical terminalpalindromes. The Densoviruses are parvoviruses which infect insect species(Tijssen et al., 1990). Although the molecular characterization of Densovirusgenomes has lagged behind that of the Parvovirus and Dependovirus genera, theDNA sequences of several Densoviruses have now been published. Interestingly,this genus contain viruses which use both strands to code protein sequences. Incontrast, all vertebrate parvoviruses contain one coding strand. The Densovirusgenomes have been shown to contain plus and minus sense DNAwith identicalterminal palindromes and (Bando et al., 1990).1.2. Molecular Biology of MVM1.2.1. Historical AspectsMVM was first isolated as a contaminant virus in mouse adenovirus stocks(Crawford, 1966). The contaminant virus was separable from adenovirus andpolyomavirus by CsC1 density equilibrium centrifugation. Material from arelatively dense band on CsC1 gradients (1.43g/ml) was found to containhemagglutinating activity. Furthermore, mouse embryo cultures infected with thenewly discovered virus also produced hemagglutinating activity. MVM waseventually plaque purified and distributed to other laboratories interested in themolecular biology of MVM (Tattersall, 1972). Thus, this original plaque purifiedstrain was designated the prototypical strain of MVM, MVM(p). Animmunosupressive strain of MVM designated MVM(i) has also been isolated andsequenced (Sahli et a!., 1985; Astell et al., 1986).51.2.2. Genomic Organization of MVM(p)MVM(p) has been cloned into plasmid vectors and the complete nucleotidesequence determined in part from cloned fragments and viral DNA (Astell et a!.,1979; Astell et a!., 1983; Merchlinsky et a!., 1983). By convention, the nucleotidesequence of MVM(p) is numbered beginning at the first nucleotide of the 5’ terminalof the positive sense strand or complementary strand DNA. The 3’ terminal hairpinand 5’ terminus hairpin of the viral DNA are also designated as the left and righthairpin, respectively, and are used interchangeably in this thesis.1.2.3. Viral TranscriptsAll vertebrate parvoviruses studied to date appear to have their entireprotein coding sequences found on one DNA strand, the plus strand. MVM(p)contains two overlapping transcription units driven by promoters at map unit 4(P4) and 38 (P38) (Fig.2)(Cotmore and Tattersall, 1987, Pintel et a!., 1983). Threemajor size classes of spliced and polyadenylated cytoplasmic transcripts designatedRi (4.8 kb), R2 (3.3 kb) and R3 (3.0 kb) have been identified (Pintel et a!., 1983). Allthree classes of transcripts are polyadenylated at a single site found near the rightterminal palindrome (Clemens and Pintel, 1987). In addition, sequence analysis ofcDNAs demonstrated that one of three possible small introns, involving twopotential splice donor and splice acceptor sites between map units 46 and 48, isremoved in all mature transcripts (Jongeneel et a!., 1986;6P4 P38 pol,AI—Ri NS-iF_ _R2___ __NS-2L---Jr VP-iR3 VP-2L____________I—VP-2Fig. 2. MVM(p) Transcription Map.The top schematic shows the MVM(p) genome containing the left (3’) and right (5’)hairpins sequences (black boxes), the transcriptional promoters located at map unit4 and 38 (P4 and P38) and the polyadenylation signal (poly A) located atnucleotide position 4885. The transcript map below shows the three classes oftranscripts Ri, R2 and R3. Open boxes and thin horizontal lines represent theprotein coding portion and the non-coding portion of the transcript. All transcriptshave one of three possible introns removed due to the utilization of two splicedonor sites (nt 2280 and 2316) and two acceptor sites (nt 2399 and 2377). Inaddition, mature R2 transcripts also have a large intron removed between nt 514and 1990. The protein synthesized from each transcript is indicated on the right.7Morgan and Ward, 1986). Synthesis of transcripts coding for NS-1 (Ri) and NS-2(R2) is directed by P4 and starts approximately at nucleotide 201 (Ben-Asher andAlom, 1984). In addition to the removal of the small intron, R2 transcripts also havea large intron removed between map units 10 (nt 514) and 40 (nt 1990) (Jongeneel etat., 1986). Thus the N-terminal 84 amino acid residues of NS-1 and NS-2 areidentical (Fig. 2). Three possible C-terminal tails of NS-2 are generated by themultiple splicing events removing the small intron. These splicing events do notaffect the NS-i protein since the termination ochre codon (TAA) at nucleotide 2277occurs before the two potential splice donor sites at nt 2280 and 2316. Synthesis ofthe R3 transcripts is directed by P38 and starts at approximately nucleotide 2005(Ben-Asher and Aloni, 1984). Again the multiple splicing events between map units46 and 48 result in R3 transcripts which code for two possible translation products,VP-i and VP-2. The two R3 transcripts utilizing the splice donor at nt 2280 code forthe smaller VP-2 protein. R3 transcripts utilize the downstream splice donor at nt2316 code for the larger VP-i protein. This splicing pattern allows the translationalstart codon at nt 2286 to be used for the initiation of the VP-i protein. Utilization ofalternative splicing is thought to promote the correct molar amounts of VP-i andVP-2 (approximately one to five) observed in infected cells and virions (Tattersall eta!., 1976; Cotmore, 1990).1.2.4. MVM(p) PolypeptidesIn vitro translation of transcripts isolated from MVM(p) infected cells showsfour major viral encoded primary translation products are synthesized (Cotmore eta!., 1983; Cotmore and Tattersall, 1986b). These products were detected usingantisera raised against bacterial fusion proteins containing portions of the NS-1 andNS-2 ORFs or sera from animals infected with different vertebrate parvoviruses. A8third viral capsid protein, VP-3, thought to be a proteolytic cleavage product of VP-2, probably arises during virus particle maturation (Tattersall et at., 1976).Sequence analysis of the large ORF in the Ri transcript predicts that the largenon-structural protein of MVM(p), NS-1, is 672 amino acids in length with amolecular weight of approximately 77 kDa (Astell et at., 1983, Cotmore et at., 1983).It has been shown that NS-1 is a nuclear phosphoprotein which migrates as an 83kDa polypeptide in SDS-polyacrylamide gels (Cotmore et at., 1983; Cotmore andTattersall, 1986a). The NS-i proteins from parvoviruses from the RV-like groupshow a high degree of antigenic conservation since sera from animals infected withany of the RV viruses recognize MVM(p) NS-1 (Cotmore and Tattersall, 1987).Further comparison shows that a 405 nucleotide sequence (nt 1428-1833) foundwithin the MVM(p) NS-1 ORF, encodes a 135 amino acid sequence which shares asignificant degree of sequence homology with the other parvoviral large non-structural proteins as well as SV4O and polyomavirus large T-antigens and the Elprotein of papillomaviruses (Astell et at., 1987). Within this conserved 135 aminoacid domain, a conserved purine nucleotide binding motif was identifiedsuggesting that NS-i requires the binding and subsequent hydrolysis of ATP for itsfunction. It has subsequently been shown that purified recombinant NS-l from thebaculovirus expression system contains ATP binding, ATPase and ATP-dependentDNA helicase activities (Wilson et at., 1991). Point mutations within the purinenucleotide binding domain abrogated the ability of mutant NS-l to replicateMVM(p) minigenomes in vivo, and severely reduced the ATP-dependent DNAhelicase activity in purified mutant protein preparations (Jindal et al., 1994). Inanother study, mutation of the lys 405 residue within the purine nucleotide bindingdomain also eliminated the ability of mutant NS-l (in nuclear extracts) to replicatecloned MVM(p) palindromic junction fragments in vitro (Nuesch et al., 1992).9Interestingly, all mutant proteins were localized in the nucleus and were able totransactivate the p38 promoter (except for mutations at lys 405) (Nuesch et a!., 1992;Jindal et a!., 1994).In addition to the replication function(s) of NS-1, it has also beendemonstrated that NS-i from MVM(p) (Doerig et a!., 1988) as well as H-i (Rhode,i985; Rhode, i987; Rhode and Richard, i987) are able to transactivate the P38promoter. Also, the NS-i of B-19 (Doerig et a!., 1990) and Rep gene products fromAAV (Labow et a!., i986; Trempe and Carter, 1988) are able to regulate their ownpromoters. Although genetic analysis of MVM NS-1 has shown that it cantransactivate P38, there is no evidence that the protein is able to bind to DNA invitro. At this point it is not known if NS-1 requires a cognate DNA bindingprotein(s) to localize itself near the promoter for transcriptional activation.However, the activation domain of NS-1 has been mapped to the C-terminal 129residues by assaying the ability of various Ga14-NS-1 fusion proteins totransactivate a Gal4-CAT construct (Harris and Astell, 1994).Relatively little is known about the activities and functions of the smaller NS2 proteins (25 kDa). It has been shown that NS-2 is required for efficient DNAreplication and virus production in certain cell types (Naeger et a!., 1990; Naeger eta!., 1993). In addition, NS-2 is required for viral replication and pathogenesis ininfected mice since mice infected with a mutant MVM(i) deficient in NS-2 synthesiscaused an asymptomatic infection whereas wild type MVM(i) causes a lethalinfection (Brownstein et a!., 1992). NS-2 has been shown to be phosphorylated andlocalized to both the cytoplasmic and nuclear compartment of infected cells(Cotmore and Tattersall, 1990). It appears that heavily phosphorylated forms ofNS-2 are found primarily in the cytoplasm whereas non-phosphorylated forms of10NS-2 were shown to be found in the nucleus as well as the cytoplasm (Cotmore andTattersall, 1990). Indirect immunofluorescence experiments and pulse chaseexperiments show that NS-2 molecules appear to be rapidly degraded compared toNS-1. Since all three forms of the NS-2 protein appear to share the samedegradation, phosphorylation and localization patterns (Cotmore and Tattersall,1990), the significance of the three alternative C-terminal tails of NS-2 remains to bedetermined.Differential splicing of the R3 class of transcripts results in two transcripts(Fig. 2) encoding VP-2 (64 kDa) and one transcript encoding VP-i (83 kDa)(Labieniec-Pintel and Pintel, 1986). Although the entire protein sequence of VP-2 isfound within the sequence of VP-i (Tattersall et al., 1977), the larger VP-i proteincontains an additional N-terminal region which contains a significant number ofbasic residues. It is thought that this basic N-terminal domain may interact with thess genomic DNA inside the capsid (Tsao et al., 1991). A third viral coat protein, VP-3 (62 kDa), is derived by a proteolytic cleavage of a short N-terminal peptide of VP-2 (Tattersall et al., 1976). Although the nature of the proteolytic cleavage has notbeen characterized, it has been shown that purified empty capsids contain only VP-1 and VP-2, but preparations of full infectious virus contains at least a small portionof VP-3 (Tattersall et al., 1976; Santaren et al., 1993). Pulse chase experiments alsoshow that the proportion of VP-3 in infected cells increases during the infectionprogression. These data suggest the VP-2 to VP-3 conversion takes place duringvirion maturation or internalization (Tattersall et a!., 1976; Santaren et a!., 1993). Thesignificance of this conversion is not known since full virions appear to beinfectious irrespective of the VP-2/VP-3 ratios. In addition, all viral structuralproteins have been shown to be phosphorylated (Santaren et a!., 1993). It is thought11that phosphorylation of the capsid proteins may contribute to the morphogenesis ofthe virion.1.2.5. Structure of MVM Terminal Hairpins.The genomic DNA of MVM contains primarily minus sense single strandedDNA (Crawford et al., 1969; Bourguignon et a!., 1976) with imperfect palindromicsequences which may form stable hairpin duplexes at the left and right termini(Bourguignon et a!., 1976). Imperfections in the palindromic sequence result inmispaired nucleotides in the hairpin structure (Fig. 3). Unlike AAV, which containinverted terminal repeats (Lusby et a!., 1980), MVM and other RV-like virusescontain a unique primary sequence at each terminus (Astell, 1990). The terminalhairpins provide parvoviruses with a novel method of circumventing the problemthat all linear DNA molecules have in replicating their 5’ ends. This problem resultsfrom the fact that all DNA polymerases require a primer (3’ OH) for DNAsynthesis. Cavalier-Smith proposed the hairpin transfer mechanism as a method bywhich linear DNA molecules containing terminal palindromes can replicate their 5’ends (Cavalier-Smith, 1974). Imperfections in the palindromic sequences whichundergo hairpin transfer result in two sequence orientations termed the flip andflop forms (for example, compare Fig. 3C and 3D).Sequence determination of the left hand hairpin of MVM(p) shows that it is115 nucleotides in length and assumes a stem plus arms or Y shaped configuration(Fig. 3A) (Astell et a!., 1979). Mismatches between nucleotide12AAGCCG GCC 26 25C c ‘ /G T GA 3C GTCACACGTCACTTACGT ACATGGTTGGTCAGTTCTAAAAATGCGG CAGTGTGCAGTGAATGCAG5’G CT ACC 89 91AGFig. 3. Structure of the MVM(p) termini.The nucleotide sequence of the viral 3’ terminus is shown in the hairpinconfiguration (A). Only 104 out of 115 nt of this left hand (3’) terminus are basepaired in the hairpin configuration. The bubble structure found within the stemresults through a mismatch between nucleotide position 25-26 to 89-91. Thenucleotide sequence of the right hand (5’) viral termini is shown in variousconfigurations. B shows the perfect base pairing between nucleotide positions 4944-4993 to 5149-5100. This portion of the palindrome is in common to the otherstructures shown in C,D and E. The nucleotide sequence of the remaining DNA(position 4994-5099) in the flip (C) and flop (D) forms is shown in straight hairpinsconfiguration. The flip and flop forms are inverted complement of each other. Eshows the nucleo tide sequence of the flip form of the hairpin in the stem plus armsstructure. Both flip and flop forms may assume a stem plus arms structure. Thenucleotide positions of the imperfections in the hairpin structures are indicated.135149 5100SalATTAGTATTACTATGTTTTTAGGGTGGGAGGGTGGGAGATACATGTGTTCB 4......4944 49935099TGCTATGAGCGAACTGGTACTGGTTGGTTGC— GCTCAACCAACCAGACCGGCT FLIPC CGATACTCGCTTGACCATGACCAACCAACG A A CGAGTTGGTTGGTCTGGCCGT4994 5024 5026 50475099 5069 5067 5046I ITT AGCTATGAGCGAACTGGTACTGGTTGGTTGC GCTCAACCAACCAGACCGGC A FLOPI) CGATACTCGCTTGACCATGACCAACCAACG— CGAGTTGGTTGGTCTGGCCGAI I4994 5044GCCTGCTATAGCGCTATA5Q99 GC 5049i GCI TAGCTATGAGCGAACTGGTAC GACCGGCE CGATACTCGCTTGACCATG CTGGCCGT T (FLIP)4994 CG 5046ATATCGCGATATGC GAGCA AGFig. 3B-E14positions 25-26 and 89-91 in the stem predicts that a bubble structure is formed inthe stem of the hairpin. Although the bubble structure and sequence is a conservedfeature among some members of the RV-like viruses (H-i, KRV, ADV, Lulli) (Astellet a!., 1979), its significance if any is has not been determined. It has been predictedthat the bubble may play a role in the asymmetrical resolution of dimer RFmolecules (Astell et at., 1983). Only one sequence orientation (flip) was found at thisterminus in both the genomic DNA and intracellular replicative DNA forms (Astellet a!., 1983; Astell et a!., 1985). The right hand (5’) hairpin was determined to be 206nucleotides in length and exists in both the flip and flop sequence orientation(Fig.3C and D) (Astell et at., 1983; Astell et a!., 1983). In addition to the extendedhairpin conformation, the flip and flop forms of the right hand (5’) palindrome canpotentially form a stem plus arms structure (Fig.3E). Deletion mutants whicheliminate the potential stem plus arms structure but not the formation of anextended hairpin were unable to replicate (Salvino et at., 1991). Genetic analysis ofAAV hairpins also suggest that the stem plus arms conformation of the terminalpalindrome is important for DNA replication (Lefebrve et a!., 1984). In fact, allparvovirus sequences determined to date have terminal palindromes which canpotentially form a stem plus arms configuration (Astell, 1990). Although thesignificance of this conserved structural feature is not fully understood, these datasuggest that the stem plus arms structure is important for parvovirus DNAreplication. Since the primary sequence of the terminal palindromes of unrelatedparvovirus genomes are different, this suggests that the topology rather than thesequence of a stem plus arms palindrome may be important in the hairpin function.15nickVparC.‘ 5Fig. 4. Modified Rolling Hairpin Model of MVM DNA Replication.The schematic diagram shows the Modified Rolling Hairpin Model ofMVM DNA replication (Astell et al., 1985). Vpar and Vprog denotethe parental and progeny viral strands. The complementary strand(C) represents the plus strand. The arrows indicate the Y hydroxylgroups which are used to prime DNA synthesis. The open circlesrepresent NS-1 which is known to be covalently attached to the 5’ends of viral DNAs. Details of this model are discussed in the text.VparCVparI234VpmgVpar/ligateCVprognickVparppCVparpC67nick161.2.6. Model of MVM DNA ReplicationA modification of the Rolling Hairpin Model for MVM DNA replication(Tattersall and Ward, 1976), which took into account the observed DNA sequencesfound at the right and left end, has been proposed (Fig. 4) (Astell et al., 1983; Astellet a!., 1985). In this model, the left hand (3’) hairpin is used to prime DNA synthesisusing the incoming single stranded DNA as the template (Fig. 4, Step 2). Theresulting monomer replicative form (mRP) is then further amplified using the righthand (5’) hairpin to prime DNA synthesis to produce the dimer replicative form(Fig. 4, step 2-4). This replicative intermediate contains essentially two genomeequivalents arranged in a head to head arrangement with the left (3’) terminalsequence joining the two genomes. The DNA joining the two monomer genomes istermed the 3’ junction or 3’ dimer bridge fragments. Since there is only onesequence orientation, flip, found at the genomic 3’ hairpin and in intracellularreplicative intermediates, the 3’ junction fragment must be resolved in anasymmetric manner to generate only the flip sequence orientation at the 3’ terminus(Fig.4, steps 4-6). The Modified Rolling Hairpin Model predicts that a site specificnickase recognize a specific site on the Vparental strand (Fig. 4, step 4). The 3’hydroxyl group provided by the nicking event primes DNA synthesis across the 3’junction fragment to a second nick introduced at the corresponding Vprogenystrand (Fig. 4, step 5). A ligation event which crosslinks the Vparental 3’ hairpinwith the Vprogeny strand producing a molecule which is covalently closed at bothtermini. The second mRF molecule generated contains extended hairpin forms atboth termini. The two resulting mRF molecules are used as templates for furtherDNA amplification or for packaging genomic DNA. Synthesis of genomic singlestranded DNA is primed at the 5’ hairpin sequence displacing the genomic single17stranded DNA which is then packaged (Fig. 4, step 7). Multiple rounds of hairpintransfer at the 5’ palindrome sequences followed by strand displacement generatesss genomic DNA (Fig. 4, step 7) containing equal proportions of flip and flopsequence configurations at the 5’ terminus. In contrast, the specific manner inwhich the 3’ junction fragments are resolved retains the unique flip sequenceorientation at the 3’ terminal. Site specific nicks introduced at the right handpalindrome sequence must be at least 18 nt inboard of the viral DNA sequence sinceintracellular DNA replicative forms are an additional 18 nt longer than the genomicviral DNA (Astell et a!., 1985). The finding that NS-1 is found covalently linked tothe ss genomic DNA, outside MVM particles, with approximately 24 nt foundoutside of the particles (Cotmore and Tattersall, 1989b) confirms this hypothesis. Itis thought the NS-1 and the external 5’ nucleotides are not required for infectionsince removal of the genome linked NS-1 and the external ss DNA does not affectinfectivity. Presumably, removal of the exposed ss DNA (and NS-1) in vivoaccounts for the “extra” 18 nt found in intracellular replicative intermediates that isnot found in ss genomic DNA.Evidence which supports the model has been recently reported. Since NS-1is found covalently bound to the 5’ ends of intracellular replicative forms as well asthe genomic single stranded DNA, it implies that NS-1 acts as the site specificnickase (Cotmore and Tattersall, 1988; Cotmore and Tattersall, 1989). Furtherevidence to support an asymmetrical resolution mechanism of the 3’ junctionfragment was obtained when recombinant forms of NS-1 were shown to specificallyresolve the junction fragments in vivo and in vitro. First, it was shown that cloned 3’and 5’ junction fragments of MVM(p) NS-1 were resolved only when plasmidsencoding the junction fragments were transfected into mouse cells and co-infectedwith MVM(p) (Cotmore and Tattersall, 1992). Further experiments showed that18extracts containing recombinant NS-1 from the vaccinia or baculovirus expressionsystems were able to resolve cloned 3’ junction fragments into the predictedcovalently closed and extended hairpin structures in an in vitro replication system(Cotmore et al., 1993; Liu et a!., 1993). In addition, the 5’ junction fragment were alsoresolved into extended ends with NS-1 covalently attached to the 5’ ends (Cotmoreet a!., 1992; Cotmore and Tattersall, 1992). Although NS-1 has been shown topossess the activities (ATPase, helicase and nickase) thought to be required to directthe resolution of dimer replicative form (dRF) molecules, the precise details ofsequence specific nicking and role of cellular proteins in the resolution reactionhave not been determined fully.1.2.7. AAV DNA ReplicationReplication of genomes containing inverted terminal repeats, such as AAV, isthought to follow a similar, but not identical scheme to that of MVM (Fig.5). SinceAAV genomic DNA contains single strands from both polarity (plus and minus)(Mayor et a!., 1969) and both flip and flop sequence orientation are found at the lefthand (3’) and right hand (5’) terminus (Lusby et a!., 1980), it suggests that bothtermini utilize the hairpin transfer mechanism to replicate the genomic termini.Sequence inversion was further demonstrated when a plasmid containing the AAVgenome with flip sequences at both ends generated AAV genomes containingapproximately equal proportions of flip and flop at both ends (Samulski et al., 1982).In order to demonstrate that AAV terminal hairpins undergo hairpin transfer, an invitro replication system using extracts from19A BTRS NE DNATRS+TRS_________23a__5_6UFig. 5 Model of AAV DNA Replication(A) AAV DNA replication uses the terminal hairpin primers (arrows) toinitiate synthesis of the mRF (2). The covalently closed hairpin is thenresolved in a series of reactions initiated by a site specific nick atthe terminal resolution site (TRS). The newly created 3’ OH then servesas a primer for synthesis of the palindrome. The fully extended terminusof the resulting intermediate can then be denatured to form terminalhairpin structures to facilitate DNA synthesis and strand displacementto generate genomic ss DNA and a duplex mRF. (B) The terminalresolution reaction shows synthetic no end (NE) DNA is resolved20AAV/adenovirus infected HeLa cell extracts were shown to resolve covalentlyclosed AAV hairpin “no end” (NE) DNA (Fig.5A step 2-4 and 5B) (Snyder et at.,1990). In addition, it was also shown that purified AAV Rep68, one of the AAVreplication proteins, contain ATP-dependent site specific nickase, DNA helicase andterminal hairpin binding activities in vitro. (Tm and Muzyczka, 1990; Snyder et at.,1990). Like MVM(p) NS-1, Rep68 is found covalently attached to the 5’ end ofnicked hairpin DNA (Tm and Muzyczka, 1990). Although it has not beendemonstrated, it’s likely that Rep proteins can also resolve AAV junction fragments.1.3. Origins of DNA replicationThe initiation of DNA replication has been hypothesized to depend on twodeterminants: a cis-acting element termed the origin or the replicator and a transacting factor termed the origin activator or initiator (Stillman, 1989). Identificationof either determinant in mammalian cells has been hampered by a lack of usefulgenetic techniques. Yet origins of DNA replication and their initiator proteins havebeen identified in mammalian DNA tumor viruses such as SV4O, bovinepapillomavirus (BPV-1), herpes simplex virus type 1 (HSV-1), Epstein-Barr virus(EBV) and in the budding yeast, Saccharomyces cerevisiae (Kelly et al., 1988;Challberg and Kelly, 1989; Marahrens and Stillman, 1992). For DNA tumor virusessuch as SV4O, adenovirus and BPV-1, the development of in vitro replicationsystems using cellular extracts or fractionated cellular extracts has allowed detailedbiochemical studies of initiation and regulation of viral DNA replication (Kelly etat., 1988; Stillman, 1989). Although in many cases, the origin and initiator proteinhave been identified, the mechanism of initiation have not been fully described.The most complete description of the initiation of DNA replication comes from thestudies of 5V40.211.3.1. SV4O Origin of DNA replicationThe interactions between SV4O and its host cell have been amenable todissection by scientists partly because of its apparent simplicity. The viralchromosome is 5243 bp of duplex circular DNA containing one origin of DNAreplication and two divergent transcriptional units coding for six proteins (Fanningand Knippers, 1992), The SV4O origin of DNA replication is embedded in thecomplex non-coding region of the SV4O genome located between the early and latetranscriptional units (Fig. 6A)(Deb et a!., 1986; Dean et at., 1987; Deb et a!., 1987). Thearrangement of the transcriptional control and DNA replication elements in acentral control region allows the virus to efficiently coordinate viral gene expressionand replication of the viral genome in the host cell. The minimal sequence which issufficient and necessary for the initiation of DNA replication in vitro is containedwithin a 64 bp region termed the on core (Fig. 6A) (Deb et a!., 1986). Although thetranscriptional control elements such as the enhancers and Spi binding sites are notrequired for DNA replication in vitro, maximal DNA replication in vivo requires thepresence of these auxiliary replication elements (DeLucia et a?., 1986; Hertz andMertz, 1986; DePamphillis, 1988). Only one viral encoded protein is required forthe initiation of DNA replication at the on core, the SV4O large T-antigen (T-Ag).This multifunctional protein plays several key roles in the viral life cycle: thecoordinate control between transcription and DNA replication, the initiation ofDNA replication at the on core and the transformation of infected cells (Fanningand Knippers, 1992). This discussion will only encompass functions of T-Ag withrespect to its role as the initiator protein at the on core.22Fig. 6A Organization of Several Eukaryotic Origins of DNA Replication.Eukaryotic origins of replication from DNA tumor viruses (SV4O, BPV-1, EBV, HSV1) and yeast (ARS1) contain multiple cis-acting elements. Each box representfunctional elements found in the respective origin of replication and are discussedin detail in the text. In the case of SV4O, both the control region and the core on(exploded view) is shown. The core on contains three elements, T-ag site II, whichis composed of four binding sites (arrows) for T-ag arranged in a perfect 27 bppalindrome, an A/T rich element and an early palindrome (EP). Transcriptionalcontrol elements like the 72 bp enhancers, GC boxes and TATA box and the T-agbinding sites I, II and III (thick lines) are shown. The BPV-1 origin of replication iscomposed of an 18 bp inverted repeat and an A/T rich element. Flanking the coreelements are two E2 binding sites (E2 BS1 1 and E2 BS12). The Epstein-Barr virusoriP contains two elements, the family of repeats (FR) element and dyad symmetry(DS) element. Each element is made up of degenerate copies of a 30 bp sequence(small box), the EBNA-1 binding site, arranged in either 30 tandem copies (FR) or intwo pairs of inverted repeats (DS). The DS and FR elements are separated byapproximately 1 kbp of DNA. The HSV-1 oriS contains three binding sites for UL9.Site I and II are inverted with respect to each other and are separated by an A/Trich element. A third UL9 binding site is found flanking UL9 site I. Yeast ARS1contains four elements. The A element contains a perfect 11/11 match with the arsconsensus sequence (ACS) The B1 and B2 sites have a 9/11 match to the ACS. TheB3 site corresponds to the binding site of the yeast transcription factor ABF1.23AT-Ag sitesIII II IEnhancers GC — Early transcriptsSV4O__1fl*ri fLJLate transcripts I II A/T H-*-*4-4-H EP IT-Ag site IIon core______JRBPV-1 I E2 BSII I I A/T 1. —‘ [ I E2 BS12 II Ion coreFR1kbDSEBV oriP iiiiiiiiiiiiiiiiiiiii N IJ-tJ--O—IJUL9 site III UL9 site I UL9 site IIHSVoriS I I-I I A/T I II ABF1 IARS1 I B3 I I B2 I I Bi I I A I9/11 9/11 9/11 11/11 ACS matches24helicaseFig. 6B Model of DNA replication at a replication fork by a multiproteincomplex.In vitro replication of SV4O DNA using highly purified proteins suggest that leadingand lagging strand synthesis requires a polymerase switching mechanism (Wagaand Stiiman, 1994). Thick and thin black lines represent the parental and newlysynthesized DNA strands, respectively. Pol oc/primase complex is required tosynthesize DNA primers for both leading and lagging (as shown) strand synthesis.Dual pol /RF-C/PCNA complexes elongate DNA as the helicase (SV4O T-ag)unwinds the DNA at the head of the fork. RF-A binds and stabilizes the ss DNA.After RNA primers (black box) are removed by MPF exonuclease and RNase H,lagging strand synthesis is completed when the 3’ end of the current Okazakifragment reaches the 5’ end of the previous fragment and is ligated with DNAligase I. This figure is adapted from that of Waga and Stillman (1994).DNA ligase IRNase HMF1 exonucleaseRF-AprimasePCNA—pol 625Although three T-ag binding sites have been identified, only site II, found inthe 64 bp on core, is required for initiation of DNA replication (Deb et al., 1986). Theon core also includes a 17 bp A/T rich element and 15 bp palindrome termed theearly palindrome (EP). Both of these elements are essential for the initiation ofDNA replication. T-Ag binding site II contains four copies of the pentanucleotidesequence, GAGGC, arranged in two pairs in opposite orientation forming a 27 bppalindromic sequence. Each pentanucleotide is capable of binding one molecule ofT-Ag (Tjian, 1978). Although a tetrameric T-Ag complex is able to form on site II(Mastrangelo et at., 1985), it can not initiate DNA replication. However, when T-Agis incubated with DNA containing the on core in the presence of ATP at 37° C, T-Agcomplexes with the on core give footprints covering the entire on core (Deb andTegtmeyer, 1987; Borowiec and Hurwitz, 1988a). Further studies usingsedimentation centrifuga tion and scanning transmission electron microscopetechniques revealed that T-Ag assembles into two hexamers forming a two lobedstructure which covers the entire on core on both strands of DNA (Mastrangelo eta!., 1989). The assembly of the two lobed structure requires ATP but not ATPhydrolysis since the complex structure is able to form in the presence of non-hydrolyzable analogs of ATP (Borowiec and Hurwitz, 1988a). This suggests thatATP acts as an allosteric molecule which induces a shift in T-Ag conformation toallow specific protein-protein interaction at the on core.The formation of the T-Ag two lobed structure on the on core inducesstructural changes in both the early palindrome and A/T rich element (Borowiecand Hurwitz, 1988b; Borowiec et at., 1990). Although the structural alteration in theSV4O on core do not require the hydrolysis of ATP, complete strand separationrequires the helicase activity of T-Ag and ATP hydrolysis (Wiekowski et at., 1988).26The activated on then recruits a three subunit protein, replication factor A (RF-A),which acts as a single stranded DNA binding protein (Melendy and Stillman, 1993)(see Fig. 6B). This protein presumably prevents the reannealing of the denaturedDNA and allows further unwinding of the duplex DNA at the origin. The TAg/RF-A/ori complex then associates with a DNA polymerase a (pol a) /primasecomplex. A RNA primer is synthesized by primase and extended by pol a(Murakami et al., 1992). The cellular replication factor C (RF-C) protein thenrecognizes and binds to the 3 end of the newly synthesized DNA primer(Tsurimoto and Stiiman, 1991b; Tsurimoto and Stillman, 1991a). RF-C then allowsDNA polymerase 6 (pol 8) and proliferating cell nuclear antigen (PCNA), aprocessivity factor for poi 6 (Prelich et a!., 1987; Prelich and Stillman, 1988), to beloaded into the replication fork. The processive RF-C/PCNA/pol 6 complex thenelongates the nascent DNA primer to synthesize the continuous leading strand ofthe replication fork (Tsurimoto et a!., 1990; Tsurimoto and Stillman, 1991b).The lagging strand synthesis is also primed by Okazaki fragmentssynthesized from the p01 a/primase complex. The lagging strand is thought toloop around in such a manner that the RNA primed Okazaki fragment is able to beelongated by a second RF-C/PCNA/pol 6 complex formed at the replication fork(Waga and Stillman, 1994). The RNA primer of the Okazaki fragments is removedby a 5’ to 3’ MPF exonuclease and RNase H (Ishimi et a!., 1988; Turchi and Bambara,1993; Waga and Stillman, 1994). The Okazaki fragments are then ligated tocomplete synthesis of the lagging strand. The dual RF-C/PCNA!pol 6 complex atthe replication fork allows leading and lagging strand synthesis to take place inconcert. This polymerase switching mechanism on both the leading and laggingstrand synthesis is supported by experiments using the in vitro replication of eitherSV4O DNA or synthetic template DNA with highly purified proteins (Tsurimoto et27a!., 1990; Tsurimoto and Stillman, 1991b; Waga and Stillman, 1994). Since thecomplete replication of naked SV4O DNA can be reconstituted in vitro with purifiedT-Ag and highly purified cellular proteins mentioned above plus purifiedtopoisomerase I and II, activation of the core on by other trans-acting factors such astranscription factors can now be explored biochemically in the SV4O system.1.3.2. Bovine Papillomavirus Origin of ReplicationPapillomaviruses are small double stranded DNA tumor viruses which areable to be maintained extrachromasomally in terminally differentiated cells (Kelly eta!., 1988; Challberg and Kelly, 1989). BPV is presented as an attractive model formolecular studies on DNA replication for many of the same reasons as SV4O. TheBPV-1 genome contains 7945 bp of covalently closed circular DNA. The BPV-1origins of DNA replication or plasmid maintenance sequences (PMS) were definedgenetically by testing plasmids containing cloned restriction fragments of the BPV-1genome for their ability to be maintained extrachromasomally in BPV-1transformed mouse cells (Lusky and Botchan, 1984; Lusky and Botchan, 1986).Although two segments of the BPV-1 genome were able to maintain plasmids(PMS-1 and PMS-2), two-dimensional gel mapping techniques (Yang and Botchan,1990) localized the origin of replication to a small region of the upstream regulatoryregion (URR) near PMS-1. The on core of BPV-1 has been shown to contains twoelements: an 18 bp inverted repeat and an 8 bp A/T rich element (Fig. 6A). An invitro replication system for BPV-1 showed that two proteins, El (68 kDa) and E2 (48kDa), were required to efficiently initiate DNA replication when supplementedwith soluble cell extracts from uninfected cells (Yang et a!., 1991). El has beenidentified as the initiator protein for BPV-l DNA replication (Lusky and Botchan,1986) since El, alone, is capable of supporting DNA replication in vitro (Yang et a!.,281991). E2 was first characterized as a transcriptional activator, but it is now knownthat E2 is also required for DNA replication in vivo. Although El alone is capable ofsupporting low levels of DNA replication in vitro, it is only able to do so atrelatively high El concentrations. Upon addition of E2, the in vitro replication ofBPV-1 DNA is greatly stimulated. Thus DNA replication of BPV-l is absolutelydependent on the presence of E2 when El is present at low concentrations. Since itwas shown that El is capable of forming a complex with E2, it was thought that theactivation of DNA replication by E2 is a result of the formation of the E1-E2complex (Mohr et a?., 1990). The E1-E2 complex was shown to protect a regioncovering the 18 bp inverted repeat found in the on core (Yang et a?., 1991).Curiously, the two E2 binding sties flanking the on core are not required for DNAreplication in vitro, It is thought that E2 functions by targeting the El initiatorprotein to the on core through the El-E2 complex. Once the El protein is localizedand stabilized to the on core via the El-E2 complex, it is believed that the El proteinacts as an ATP-dependent helicase to unwind DNA at the on (Yang et a?., 1991) toinitiate DNA replication.1.3.3. Herpes Simplex Virus 1 (HSV-l) and Epstein-Barr Virus (EBV) Origins ofDNA ReplicationHSV-1 (153 kbp) and EBV (172 kbp) are large linear double stranded DNAviruses which belong to the herpesvirus family of viruses. Although the EBVgenome is linear, EBV is maintained as supercoiled plasmid in the nucleus of EBVimmortalized cells (Middleton et a?., 1991). The EBV origin of DNA replication,oriP, was first defined by testing recombinant plasmids containing segments of theEBV genome to replicate in EBV transformed or EBV nuclear antigen 1 (EBNA-l)expressing cell lines (Yates et a?., 1984; Yates et a?., 1985). Deletion analysis of oriP29shows that it is composed of two functional elements, both of which containmultiple copies of a degenerate 30 bp sequence (Fig. 6A) (Reisman et al., 1985). Oneelement contains 20 copies of the 30 bp sequence arranged as direct repeats and istermed the FR element (for family of repeats). The second element contains aregion of dyad symmetry, termed the DS element, where four copies of the repeatsequence are arranged as a pair of inverted repeats. The DS and FR elements areseparated by approximately 1 kbp of DNA. The exact spacing the these elementsappears not to be critical since the intervening DNA can be lengthened or shortenedwithout impairing oriP activity (Reisman et al., 1985). EBNA-1 (80 kDa) is the viraltrans-acting initiator protein which binds to both the FR and DS elements (Rawlinset a!., 1985), Although both elements are required for oriP activity, the initialunwinding and DNA synthesis is thought to occur at the DS element (Gahn andSchildkraut, 1989). Initiation is thought to include a DNA looping mechanismwhich links the DS and FR elements. Biochemical (Middleton and Sugden, 1992)and electron microscopy studies (Frappier and ODonnell, 1991) show that EBNA-1complexes binding to each of the DS and FR elements can interact with each otherthrough protein-protein interaction, looping out the intervening DNA. This mayexplain why spacing between the two elements is not critical for oriP function. HowEBNA-1 promotes the initiation of DNA replication from the looping structure isnot known. EBNA-1 does not appear to have a helicase or an ATP binding activitythat is seen in the other replicator proteins discussed. Thus, in addition to thestructural effects induced by EBNA-1, it may be possible that EBNA-1 is able torecruit cellular replication factors to the activated oriP promote formation ofinitiation complexes.HSV-1 is known to contain three functional origins of replication. OriL isfound in the middle of the long unique region of the HSV-1 genome whereas oriS is30found in the inverted repeat sequences found flanking the unique short region ofthe HSV-1 genome (Challberg and Kelly, 1989). Most molecular studies have usedoriS since plasmids containing oriL appear to be highly unstable in bacteria. Unlikethe large EBV oriP, the HSV-1 origin of replication, oriS, is relatively compact. OriSconsists of a 45 bp palindrome (Deb and Doelberg, 1988; Lockshon and Galloway,1988) containing two high affinity binding sites (site I and site II) for the HSV-1origin binding protein, UL9 (82 kDa) (Elias and Lehman, 1988; Koff and Tegtmeyer,1988; Olivo et a!., 1988). The two high affinity UL9 binding sites are inverted withrespect to each other and are separated with an AlT rich element. In addition, athird, low affinity UL9 binding site (site III) flanking site I appears to be importantfor DNA replication. The UL9 protein is thought to bind to each site as a dimer(Koff and Tegtmeyer, 1988) since UL9 ia a dimer in solution (Fierer and Challberg,1992). After both high affinity sites are filled, protein-protein interactions betweenthe two UL9 dimers occurs, looping out the intervening A/T rich DNA (Koff et a!.,1991; Rabkin and Hanlon, 1991). Chemical probing experiments suggest thatlooping out of DNA distorts the A/T rich element (Koff et a!., 1991), which thenmay promote the formation of replication complexes. It has been recently shownthat purified UL9 protein contains an ATP-dependent helicase activity (Fierer andChallberg, 1992; Boehmer et a!., 1993; Dobson and Lehman, 1993) which isstimulated by the viral encoded ss DNA binding protein, ICP8. This findingsuggests that UL9 (complexed with ICP8 (Boehmer and Lehman, 1993)) is thehelicase which unwinds the DNA at oriS. The mechanism of assembly ofreplication complexes on the activated oriS has not been reported yet.311.3.4. The yeast chromosomal ARS1 origin of replicationThe origins of replication described in this review have been limited to a setof mammalian DNA viruses thus far. Identifying yeast chromosomal origins ofreplication was relatively simple since short chromosomal sequences can be clonedinto plasmid vectors and tested for their ability to replicate in yeast along with thehost chromosome in plasmid stability assays (Fangman and Brewer, 1991).Sequences which are able to maintain plasmids in yeast were termed autonomouslyreplicating sequences or ARS. An ARS consensus sequence (ACS),(T/A)THA(T/C)(A/G)TTT(T/A) (Van Houten and Newlon, 1990), was defined bycomparing several ARS sequences from yeast chromosome III. Recently, thechromosomal ARS1 origin was shown to contain four functional elements(Marahrens and Stillman, 1992) (Fig 6A). The A element contains an 11/11 match tothe ACS. The B1 and B2 elements are adjacent to the A element and contain a 9/11match to the ACS. The B3 element is located most distal to the A element andcorresponds to the binding site for the yeast transcription factor ABFI. It wasshown that the B3 element could be replaced with the binding sites for eithertranscription factors RAP1 or Ga14 without loss of ARS activity. Only the A elementis essential for ARS activity since plasmids containing linker insertions into of eachof the Bi, B2 or B3 elements, individually, are maintained in yeast strains(Marahrens and Stiliman, 1992). A multiprotein complex, origin recognitioncomplex (ORC), was purified from yeast and shown to bind to the A element ofARS1 and other A type elements in other chromosomal ARS sequences, in DNase Ifootprinting assays (Bell and Stiliman, 1992). Although the binding of ORC to the Aelement is dependent on the presence of ATP, it is not clear if the hydrolysis of ATPis required since purified ORC preparations do not appear to have ATPase activity.SDS-PAGE analysis of purified ORC shows that ORC contains possibly six32polypeptides (120, 72, 62, 56, and 50 kDa). Although the ORC is considered theinitiator in yeast, its role in the initiation of chromosomal DNA replication has notbeen elucidated.1.4. The present studyThe objective of this study was to identify elements within the MVM(p)genome which are necessary for viral DNA replication. Although it was clear fromprevious studies that the viral terminal palindromes were essential for DNAreplication, sequences inboard of the terminal palindromes had not beeninvestigated. Since the MVM(p) genome has been cloned into a plasmid vector,pMM984 and this has been shown to be infectious, the first objective was to developan in vivo DNA replication system based on the original infectious clone. After theDNA replication assay was developed, mutant plasmid constructs of the viralgenome (“minigenomes”) were generated to determine sequences which affect viralDNA replication. Deletion mutants made in the left hand (3’) palindrome, internalright end (IRE) and internal left end (ILE) regions were tested. Mutant genomescontaining identical termini were also constructed and tested. It was determinedthat sequences found internal of the right palindrome were important for DNAreplication of MVM minigenomes. This sequence was used as a probe to detectDNA binding factors in nuclear extracts. The factors which bind to elements whichaffect the DNA replication of the mutant virus may be potential replication factors.33II. Materials and Methods2.1. MaterialsChemicals were purchased from BDH, Fisher Scientific, or Sigma ChemicalCo. unless otherwise specified. Polyacrylamide and agarose gel electrophoresissupplies were obtained from GIBCO/Bethesda Research Laboratories (BRL) or BioRad Laboratories.Tissue culture media and supplies were purchased from eitherGIBCO/Bethesda Research Laboratories or Sigma Chemical Co. Bactotryptone,yeast extract and Bactoagar were supplied by Difco Laboratories. Ampicillin(Penbritin) was purchased from Ayerst Laboratories.All restriction and DNA modification enzymes were purchased from eitherGIBCO/BRL, Promega or New England Biolabs unless otherwise specified. DNAsequencing kits using the Sequenase enzyme were supplied by United StatesBiochemical Company (USB).2.2. BacteriaThe recombination deficient strains of E.coli, JC8111 (Boissy and Astell, 1985)and Sure (Stratagene), were used to maintain and propagate plasmids whichcontain the right hand (5’) terminal palindrome sequence of MVM(p) since it hasbeen shown that specific deletions in the right hand terminal sequences tend to begenerated and propagated in standard strains. The DH5x strain was used for allother routine cloning. Bacteria were routinely grown in YT medium (8 g tryptone,345 g yeast extract and 5g NaC1 per liter) supplemented with either 100 p.g/mlampicillin or 10 ig/ml tetracycline (when necessary) for the selection of plasmids.2.3. Mammalian Cell Lines2.3,1. COS-7 CellsCOS-7 cells (obtained from T. Maniatis, Harvard University) were derived bytransforming the simian kidney cell line, CV-1, with an origin-defective SV4Ogenome which is integrated into the host cell genome (Gluzman, 1981). In this cellline, SV4O large T-Antigen is expressed and is thus able to maintain plasmids whichcontain the SV4O origin of replication. COS-7 cells were cultured in Dulbecco’sModified Eagle Medium (DMEM) supplemented with 10 mM HEPES-NaOH,pH=7.4 and 10% fetal bovine serum at 37° C in 5% C02. Cells were passagedapproximately once every three days by trypsinization and dilution (1:10).2.3.2. LA9 CellsMurine LA9 cells (Littlefield, 1964) used in this study are a ouabain resistantisolate of the HGPRT- mouse fibroblast cell line A9 and were obtained from P.Tattersall (Yale University). These cells have been shown to be permissive toinfection by MVM(p) and are routinely used for the molecular analysis of MVM(p)(Tattersall, 1972). LA9 cells were grown in DMEM supplemented with 10 mMHEPES-NaOH, pH=7.4 and 5% fetal bovine serum and passaged as describedabove.352.4. Basic Cloning TechniquesMany of the basic cloning procedures and the growth of bacteria weredescribed previously (Sambrook et a!., 1989).2.4.1. Isolation of Plasmid DNALarge scale isolation of plasmid DNA was performed using the alkaline lysismethod followed by either density equilibrium centrifugation in CsCl-ethidiumbromide (EtBr) gradients or the polyethylene glycol precipitation method. FinalDNA concentrations were determined by A260 readings. Small scale isolation ofDNA was performed by the alkaline lysis method followed by a phenol/chloroformextraction and ethanol precipitation.2.4.2. Isolation of DNA Fragments From Agarose GelsRestriction fragments were routinely isolated from agarose gels afterelectrophoresis in Tris-Acetate-EDTA (0.04 M Tris-Acetate, 1 mM EDTA pH=8.0)(TAE) buffer. DNA was stained with 0.2 jig EtBr/ml added to the gel and runningbuffer and visualized by exposure to UV light. DNA from gel slices was isolated byelectroelution followed by ethanol precipitation or by adsorption to glass beadsusing the GeneClean DNA purification reagents (BlO 101 Inc.). Purified DNA wasredissolved in 5-10 ii TE (10mM Tris-HC1, pH=8.0, 1mM EDTA).362.4.3. Cloning of Restriction Fragments into Plasmid VectorsSmall restriction fragments were routinely cloned into plasmid vectors.Typically, vector and insert fragments were ligated in a 10 p.1 reaction containing 50mM Tris-HC1, pH=7.6, 10 mM MgC12, 1 mM ATP, 1mM dithiotheritol (DTT) and 1U T4 DNA ligase. The ligation reaction was allowed to proceed for 2 hours at roomtemperature before transformation of competent E.coli with the ligation products.When blunt-ended restriction fragments were cloned, the vector DNA fragmentwas dephosphorylated using calf intestinal alkaline phosphatase (ClAP). Thereaction was terminated by adding SDS to a final concentration of 1% andincubating at 75°C for 15 mm. Vector DNA was then extracted once withphenol/chloroform and precipitated with 3 volumes of ethanol.Dephosphorylation of 5’ ends of DNA inhibits self-ligation and increases theefficiency of cloning blunt-ended molecules Blunt ended ligation reactions wereincubated at room temperature for 4 hours or at 16° C overnight beforetransformation.2.4.4. Preparation and Transformation of Competent CellsCompetent Sure cells and DH5cc were prepared using the CaC12 method.Briefly, 100 p1 of a fresh overnight culture was inoculated into 35 ml of YT mediumand allowed to grow in a 37° C waterbath, shaking vigorously. Cells wereharvestedwhen cultures reached 0D590= 0.2 by centrifugation (10 mm X 4000 rpm).The cells were resuspended in 15 ml ice cold 50 mM CaCl2by gentle aggitation witha vortex mixer. After incubation on ice for 30 mm with occasional agitation, thecells were’ recentrifuged and resuspended in 3.0 ml ice cold 50 mM CaCl2containing 15% glycerol. Two hundred microliter aliquots were flash-frozen in an37ethanol/dry ice bath before being stored at 800 C. JC8111 cells were madecompetent by the Hanahan method (Hanahan, 1983) since this strain can not bemade competent using the standard CaCl2method. In this method, 100 ml cultureswere grown in SOB medium to a 0D590=0.7 before being harvested bycentrifugation. The cells were resuspended in 30 ml ice cold frozen storage buffer(FSB) and incubated on ice for 30 mi After recentrifugation, the cells wereresuspended in 8.0 ml of ice cold FSB. After the addition of 240 iii of DMSO, cellswere flash-frozen in 200 p1 aliquots as described above.In standard transformations, 200 p1 of frozen competent cells were allowed tothaw on ice before being added to ice cold plastic culture tubes containing 2-5 p.1 ofa ligation reaction. The transformation mixture was incubated on ice for a further30 mm. Cells were heat shocked at 42° C for 90 seconds and 0.8 ml YT was added.Transformed cells were incubated in a 37° C water bath, gently shaking, for 45-60mm. Usually 20 and 200 p.1 of each transformation mixture were spread on YT-agarplates containing the appropriate antibiotics.2.5. DNA SequencingPlasmid DNA to be sequenced by the dideoxy chain termination DNAmethod (Sanger et al., 1977) was isolated using the standard mini-prep method.Usually 2 p.g of each plasmid was boiled with 2-5 pmol of the appropriate primerfor 2 mm in 0.2 M NaOH in a final volume of 20 p.1 and neutralized with 3 p.1 3 Msodium acetate (pH=5.2). The DNA was then precipitated with ethanol,centrifuged, washed and dessicated. The resulting DNA was then redissolved in 10p.1 of lx sequencing buffer (40 mM Tris-HC1, pH=7.5, 20 mM MgCl2 and 50 mMNaCl) and incubated at 37° C for 15 mm before proceeding with the labelling and38termination reactions using the DNA sequencing mixes and Sequenase enzymesupplied by USB. Labelled DNA fragments in the termination reactions were thenseparated by electrophoresis through a 6 or 8% polyacryamide gel containing 8 Murea and Tris-Borate-EDTA (TBE) buffer (89 mM Tris-Borate, pH=8.3, 2 mM EDTA).After electrophoresis, gels were dried and subjected to autoradiography.2.6. Plasmid ConstructionThe original infectious clone of MVM(p), pMM984 (obtained from D.C.Ward, Yale University), contains the double stranded sequence of MVM(p) clonedinto the BamHI site of pBR322 (Merchlinsky et a!., 1983). This clone was shown toproduce infectious virions upon transfection into cells permissive for MVM(p).Both pMM984 and the variant pCA4.O, which contains a Smal site adjacent to theleft end BamHI site of MVM(p), were used as starting points for most plasmidconstructions.2.6.1. Construction of pCMVNS-1The NS-1 expression vector, pCMVNS-1, was constructed in two steps. First,the 2.3 kbp Hga I fragment of pMM984 was blunt ended using T4 DNA polymeraseand ligated to phosphorylated BamHI linkers. After digestion with BamHI, thisinsert was then cloned into the BamHI site of pUC 19 (Wilson et a!., 1991). TheBamHI fragment was transferred to the BamHI site of the pCMV-5 SV4O originbased mammalian expression vector (Andersson et a!., 1989). The expression of NSIis directed by the strong immediate early CMV promoter.392.6.2. Construction of pPTLRSince most of the internal coding portion of the MVM(p) genome is notrequired in cis for the replication of the viral DNA, I sought to make a minigenomicconstruction which contained only the terminal sequences. The right terminal 0.8kbp Xba I (4342)/Barn HI fragment of pMM984 was cloned into the Xba I/Barn HIsites of pUC 19 and designated pUCRH. The 0.7kb Hind III/Pst I fragmentcontaining the 411 bp of the left terminus was subsequently cloned into pUCRH.This plasmid, designated pPTLR, contains both the right and left terminalpalindrome sequences as well as the flanking internal sequences. The P4 promoterand transcriptional start site remains intact at the left end. The singlepolyadenylation site and 65 bp tandem repeat were also left intact in the rightterminal fragment.2.6.3. Construction of Left Palindrome DeletionsIn order to analyze the sequence requirements of the left palindrome, the leftend containing the 0.4 kbp Sma 1/Alu I (408) fragment of pCA4.0 was cloned intothe Sma I site of pUC 19 and designated pCA4O8. Plasmid pCA4O8F, containing theSma 1/Alu I fragment oriented such that the 3’ palindrome sequence is proximal tothe EcoRI end of the multiple cloning site, was digested with Sst I and Sma I andused to generate exonuclease III deletions (Henikoff, 1984). After sequence analysis,clones containing deletions from MVM(p) nucleotide one to nucleotide 12, 26, 31,57, and 78 bp were selected. The 0.9 kbp Xba I (4342)/Sph I fragment of pMM984which contains the 5’ terminal palindrome was then cloned into the Xba I/Sph I siteof the deletion clones resulting in a set of minigenomes containing deletions in theleft terminal palindrome.402.6.4. Construction of Internal Left End (ILE) and Right End (IRE) DeletionsDeletion of the internal portions of the minigenome encoded within pPTLRwere made in both the internal left end (ILE) and internal right end (IRE) regions.Unidirectional IRE deletions were generated using XbaI/PstI digested pPTLR as atemplate for exonuclease III digestion (Henikoff, 1984). ILE deletions weregenerated by digesting pPTLR with HincII/NcoI and HinclI/Styl. The DNAfragments were blunt ended using T4 DNA polymerase and ligated. Plasmids fromrecombinant clones were first sized on agarose gels. Clones containingapproximately the desired size deletions were then sequenced to determine theexact nucleotide junctions.In order to use the EcoRV (381) site within the MVM(p) minigenomes forsubsequent analysis of the IRE deletions, the EcoRV site in the vector sequences wasremoved by replacing the EcoRV(381)/HindIII left end containing fragment withthe EcoRV(381)/HindIII fragment from pCA4O8F. The subsequent series of clonesis essentially identical to the parental clones except it contains 408 bp (Alu 1(408)) ofthe left terminus instead of 411 bp. with respect to the minigenomic sequences. TheSV4O origin of DNA replication was cloned into the resulting minigenomes byligating a blunt ended 0.3 kbp SmaI/PstI fragment of pSV(-), which contains the BglI/PvuII fragment of the SV4O genome (SV4O nucleotides 5235-270), into the uniqueEcoRV site, resulting in the SVR series of minigenomes. In order to use the XbaI(4342)/Sau3a(4741) fragment (SX fragment), the SX fragment from pUCRH wassubcloned into the BamHI/XbaI site of pUC 19 and designated pUCXbaSau. Thenthe blunt ended SmaI/PstI fragment pUCXbaSau which contains the SX fragment41was ligated into the unique EcoRV (381) site of the modified minigenomes resultingin the SXR series of minigenomes..2.6.5. Construction of Minigenomes Containing Two Left Termini or Two RightTerminiThe 4.5kb NheI (pBR322 sequence)/XbaI(4342) fragment of pMM984 wascloned into the XbaI site of pCA4O8F to create a plasmid, pPTLL, encoding a 4.7 kbpMVM(p) minigenome containing two left hairpins. A smaller version of the LLminigenome was constructed by deleting the internal XhoI(2070)/XbaI(4342)restriction fragment from pPTLL, resulting in pPTLLX. In order to introducerestriction fragments into the LLX minigenome, the pPTLLX was digested withEcoRV and Bgl II linkers were ligated onto the EcoRV blunt ends. After digestionwith Bgl II, the vector containing the identical left terminal fragments (pPTLLX-Bgl)was purified and used in ligation reactions. In order to insert the SX fragment intopPTLLX-Bgl, the Smal/Hincli fragment of pUCXbaSau, which contains the SXfragment, was purified before the ligation of Bgl II linkers. After digestion with BglII, the fragment was purified and ligated into the pPTLLX-Bgl vector giving rise tothe LLSX series of minigenomes. In addition to the LLXS clones, the 350 bpBamHI/BglI fragment of the FUS3 gene of yeast was also inserted into pPTLLX-Bglvector giving rise to the LLFUS3 series of minigenomes.The right end containing 0.9 kbp EcoRl/Hindill fragment of pUCRH wasligated to the large EcoRl/Hind 111(3996) 5.1 kbp fragment of pMM984. Theresulting plasmid, designated pPTRR, encodes an MVM(p) minigenome containingtwo right termini oriented as inverted terminal repeats with the MVM(p)XbaI(4342)/HindIII(3996) fragment separating the two right termini. Construction42of minigenomes containing inverted terminal right or left end repeats withoutspacer DNA was unsuccessful.2.7. Transfection of DNA into Mammalian CellsDNA was introduced into COS-7 and LA9 cells by the DEAE-dextrantransfection method (Lopata et al., 1984). Cells were usually passaged at dilution of1:5 the night before the transfection. The next morning, the cells were washed twicewith DMEM before the transfection mixture was added to the cells. Thetransfection mixture was made by adding 5-10 jig of the appropriate plasmid DNAsto 0.3 ml DEAE-Dextran solution(Mr=20,000) (2 mg/mi). The mixture was made upto a final volume of 3.0 ml with DMEM. After the addition of the transfectionmixture, the cells were incubated at 37° C in 5% CO2 for 8 hours. The cells wereshocked with PBS (0.20 g KC1, 0.24 g KH2PO4, 8.0 g NaCl, 1.4 g NA2PO4)containing 10% DMSO for 2 mm (LA9 cells) or 4-5 mm (COS-7 cells). Thisprocedure has been shown to increase transfection efficiency in a variety of celllines. The transfected cells were allowed to grow in complete medium for 3 daysbefore harvesting for further analysis.2.8. Isolation of Low Molecular Weight DNA From Mammalian CellsLow molecular weight DNA was isolated by a modified Hirt extractionprocedure (Hirt, 1967). Transfected cells were washed twice with PBS then lysedwith 1.0 ml of Hirt lysis buffer ( 100 mM NaC1, 1% SDS, 10 mM EDTA, pH=8.0)added to the culture dish. After incubation at room temperature for 15-20 mm, thelysate was transferred to a fresh 1.5 ml microfuge tube. NaC1 was added to a finalconcentration of 1.1 M and the lysate was mixed by inversion 10-15 times. After the43extract was incubated on ice overnight, the fluffy white precipitate containingmainly genomic DNA was pelieted by centrifugation for 30 mm at 4°C in amicrofuge. The supernatant was transferred to a tube and digested with proteinaseK (BRL) at a final concentration of 500 jig/mi at 37° C for 2 hours. After onephenol/chloroform extraction, the nucleic acids were precipitated with one volumeof isopropanol.2.9. Use of Restriction Endonuclease, Exonuclease III and Si nuclease toCharacterize Replicated DNAIn order to characterize the DNA in Hirt extracts, samples were digestedwith various restriction and other nucleases. DpnI and MboI digestions wereperformed in a buffer supplied by the manufacturer. Duplicate exonuclease III(exolli) digestions of Hirt DNA samples were performed in a 100 p.1 volumecontaining 66 mM Tris-HC1 pH=8.0, 77 mM NaCl, 5 mM MgCl2,10 mM DTT and660 U exolil for 1 hour at 37°C. The reactions were terminated by the addition of100 p.1 of ice cold 2X Si nuclease buffer (60 mM potassium acetate pH=4.6, 500 mMNaCl, 2 mM ZnC12,10% glycerol). Ten units of Si nuclease were added to one ofthe duplicate sets of exolil reactions and incubated at 37°C for 30 mm. All sampleswere extracted once with phenol/chloroform and the DNA was precipitated withethanol before agarose gel electrophoresis and Southern blot analysis (Southern,1975).442.10. Southern Blotting and HybridizationDNA was electrophoresed through a 1.0% agarose gel containing TAE bufferand 0.2 pg/ml EtBr at 10-20 volts for 16-20 hours. The DNA was then transferred toa nylon membrane (GeneScreenPius, Dupont) using the LKB 2016 VacuGenevacuum blotting system. The gel was first soaked in 0.25 M HC1 for 10-15 mm(until the xylene cyanol dye turned yellow). The gel was then laid on top of thenylon membrane supported by a porous screen. A vacuum of 40-50 cm ofH20wasapplied. DNA in the gel was then denatured by adding a denaturation solution (1.5M NaC1, 0.5 M NaOH) on top of the gel for approximately 15-20 mm. After thedenaturation solution was removed, the gel was neutralized by adding aneutralization solution (1.5 M Tris-HC1, pH=5.0, 1.5 M NaC1) on top of the gel foranother 15-20 mm. The transfer was then allowed to proceed in 20X SSC (3 MNaCl, 0.3 M sodium citrate, pH=7.0) for one hour. The nylon membrane was thenwashed in 5X SSC for 5 mm to remove any agarose left on the membrane andallowed to dry.Nylon membranes were prehybridized in 10 ml of the hybridization solution(5X SSPE, 1.0% SDS, 0.1 % Tween 20, 50% formamide and 100 jig/ml denaturedsheared salmon sperm DNA) before the addition of the denatured 32P labeled DNAprobe. Hybridization was carried out in a water bath at 43-44°C for at least 16hours. Membranes were then washed twice in 1XSSPE, 0.1% SDS at roomtemperature for 30 mm each, and twice in 0.1X SSPE, 0.1% SDS at 65°C for 30 mmeach. The membranes were wrapped in Saran Wrap and subjected toautoradiography.452.11. Preparation of 32P labeled DNA Probes32P labeled DNA probes were prepared by the random hexamer primingmethod (Feinberg and Vogeistein, 1983). Usually, the template DNA wasdenatured in boiling water for 5 mm in a total volume of 8.0 p.1 containing 50-100 ngof a purified restriction fragment, 1.6 p.1 of 100 A260 units/mi of random hexamers.The reaction was cooled on ice for 5 mm before the addition of 5 p.1 1 M HEPESNaOH, 5 p.1 5X dNTP (100 mM of each dCTP, dGTP, dTTP, 50 mM 13-mercaptoethanol and 25 mM MgC12), 5 p.1 3000Ci/mmol [a32P]dATP, 1 p.110mg/mi acetylated BSA and 6U DNA plymerase I (Klenow fragment). The labelingreaction was allowed to proceed at room temperature for 4 hours. The labeledDNA was then precipitated with ethanol and washed 5-6 times with 1.0 ml 70%ethanol to remove the majority of the unincorporated label. The labeled probe wasdissolved in 100 p.1 of dH2Oand denatured in boiling water for 5 mm before beingused in Southern hybridization experiments. Probes contained approximately 108cpm/ug DNA as determined by TCA preciptiation.2.12. Western Blot AnalysisIn order to determine the levels of NS-1 in transfected LA9 cells and COS-7cells, lysates of transfected cells were probed with the NS-1 specific monoclonalantibody, CE1O (Yeung et al., 1990), using the Enhanced Chemiluminescencedetection (ECL) kit supplied by Amersham. Transfected cells were washed twicewith 10 ml PBS before lysis in 200 p.1 sample buffer (62.8 mM Tris-HC1, pH=6.8, 2%SDS, 5% 13—mercaptoethanol, 10% glycerol and 0.05% bromophenol blue). Theviscous lysate was transferred to a microfuge tube and boiled for 10 mm. Aliquots(2-10 p.1) were loaded onto a 4%/12% discontinuous polyacrylamide mini-gel (Mini46Protean II, Bio-Rad) and electrophoresed for one hour at 100 V (Laemmli, 1970).After electrophoresis, proteins were transferred to nitrocellulose membranes byelectroblotting in transfer buffer (25mM Tris, 192 mM glycine pH=8.3 and 20 %methanol). The membranes were incubated with undiluted hybridoma culturesupernatant containing the CE1O primary antibody for one hour. The membraneswere washed and incubated with the secondary goat anti-mouse antibodyconjugated to horseradish peroxidase (1:3000 dilution) for one hour.Immunoreactive proteins were detected using a chemiluminescence reaction duringexposure to Kodak X-AR film as described by the protocol from the supplier.2.13. Preparation of Nuclear ExtractsNuclear extracts were prepared from isolated nuclei from MVM(p) infectedor uninfected LA9 cells (Dignam et a!., 1983). Large scale growth of LA9 cells wasdone in Beilco spinner culture flasks. Ten to 20 subconfluent plates of LA9 cellswere trypsinized, pelleted by centrifugation and resuspended in Joklik’s ModifiedEagles Medium (S-MEM) supplemented with 5% fetal bovine serum, 10 mMHEPES-NaOH, pH=7.4. The culture was then introduced into a 250 ml spinnerculture flask and gassed with 5% C02 balanced air. Cells were allowed to grow insuspension at 37° C with stirring at a 3.5 setting on a Belico 4-Spin magnetic stirrer.Cultures were maintained continuously at a cell density of 1.0 x 106 to 6x10 cellsper ml. Approximately 2x108 LA9 cells were infected at a multiplicity of infection(moi) of 10 in a final volume of 50 ml containing DMEM supplemented with 1%fetal bovine serum and 10 mM HEPES-NaOH, pH=7.4 for one hour. The cells werethen diluted to a final volume of 1000 ml of complete medium in a 2 liter spinnerflask and gassed with 5% C02 balanced air. Infected cells were allowed to grow for16 hours before harvesting.47Nuclear extracts were prepared using the protocol from Dignam et a!. (1983).All steps were performed at 4° C. Infected or uninfected LA9 cells were harvestedby centrifugation. The cells were washed once in 5 PCV (packed cell volume) of icecold PBS and once in 5 PCV of buffer A (10 mM HEPES-KOH, pH=7.9, 10 mM KC1,1.5 mM MgCl2,0.5mM DTT, 0.5 mM phenylmethylsulfonyl fluoride (PMSF) and 1.tg/ml each antipain, leupeptin and pepstatin). The resulting cells wereresuspended in 2 PCV of buffer A and homogenized in a 25 ml DounceHomogenizer using 20 strokes of the loose fitting “B” pestle (Kontes). The ruptureof the cells was monitored visually by phase contrast microscopy. The nuclei werepelleted in a 15 ml COREX tube (2000 x g, 10 mm). The supernatant was removedand the pellet was recentrifuged (25,000 x g, 20 mm) to remove the residualcytoplasmic material. The nuclei were resuspended in 2 ml buffer C (20 mMHEPES-KOH, pH=7.9, 0.6 M KC1, 1.5 mM MgCl2, 0.5mM DTT, 0.2 mM EDTA, 0.5mM PMSF, 1 ig/ml each of antipain, leupeptin and pepstatin, and 20% glycerol)and homogenized again using 10 strokes of the B pestle in a Dounce homogenizer.The ruptured nuclei were stirred for 30 mm and centrifuged (25,000 x g, 30 mm).The supernatant was then dialyzed against one liter of buffer D (20 mM HEPESKOH, pH=7.9, 0.1 M KC1, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5mM DTT, 0.5 mMPMSF and 25% glycerol) for 8 hours. Particulate matter in the dialysate wasremoved by centrifugation (25,000 x g, 30 mm). The supernatant, designatednuclear extract, was stored at -70° C in 100-200 p1 aliquots.482.14. Fractionation of Nuclear ExtractsTypically, 5-10 ml of uninfected LA9 nuclear extract was loaded on a 10 mlDEAE-Sephacel (Pharmacia) column equilibrated with buffer D containing 1 jig ofeach/mi antipain, leupeptin and pepstatin. The column was washed with 50 ml ofbuffer D followed by 50 ml buffer D containing 0.6 M KC1. Fractions wereimmediately assayed for binding activities and relative protein concentration(A280). Fractions containing DNA binding activity were pooled and loaded on a 10ml heparin-agarose (Sigma) column equilibrated with buffer D. The heparinagarose column was then washed with 60 ml of buffer D containing 0.1 M KC1, 0.3M KC1 and 0.6 M KC1. Peak protein fractions were assayed for DNA bindingactivity, dialyzed against one liter of buffer D for 6-8 hours and stored at -70° C.2.15. Electrophoretic Mobility Shift Assay (EMSA)Electrophoretic mobility shift assays (EMSA) were performed in a finalvolume of 20 p1 containing 10 mM Tris-HC1 (pH=7.5), 50 mM NaCl, 1.0 mM DTT,250 ng/pI double stranded poly dI-dC (Pharmacia) and 2.5-5.0 jig of nuclearextracts. The proteins were preincubated for ten mm at room temperature followedby addition of 10,000 cpm (0.2-1.0 ng) of the indicated 32P end labeled probe. Aftera 30 minute binding reaction, DNA-protein complexes were loaded onto a native4% polyacrylamide gel (Hoeffer SE600) containing 0.5 X TBE and 1% glycerol. Gelswere pre-electrophoresed for 2-3 hours at 200V in 0.5 X TBE/1% glycerol prior tosample loading. Samples were electrophoresed at 150 V for 2.0-2.5 hours before thegel was dried on Whatman 3MM paper and subjected to autoradiography.492.16. DNase I FootprintingIn DNase I footprinting experiments, proteins from column fractions wereused in binding reactions as described above except the final incubation volumewas 50 p1 and contained 10 mM MgC12 and 5 mM CaC12. The DNA in the bindingreaction was tpartially digested with DNase I by the addition of 2 ti of 1/25thdilution of the enzyme (Promega) in 150 mM NaC1, 1 mM CaCl2 and 50% glycerol.The DNase I digestion was stopped after 60 seconds with 50 p.1 DNase I stop buffer(8 M ammonium acetate, 40 mM EGTA and 100 p.g/ml yeast tRNA). Samples wereimmediately extracted once with phenol/chloroform and precipitated with ethanol.The precipitated DNA was dissolved in 2 p1 of formamide dye buffer and boiled for2 mm before being electrophoresed through denaturing 10% polyacrylamide/8 Murea sequencing gels. Samples were electrophoresed at 32 W constant power untilthe bromophenol blue dye reached approximately two-thirds the length of the gel.Gels were dried on Whatman 3MM paper and subjected to autoradiography.2.17. Preparation of Probes for DNA-Protein Interaction Studies2.17.1 RsaA and RsaB Restriction FragmentsThe RsaA and RsaB probes were prepared by cloning the respective Rsa Ifragments (nt 4431-4579 and 4579-4662) into the Sma I site of pUC19 and releasingthen with BamHI and EcoRI digestion. Digestion products were loaded onto a 5%polyacrylamide gel containing 1 X TBE and electrophoresed 16-20 hours at bOy.After the gel was stained by soaking in 1 X TBE containing 0.5 p.g/ml EtBr for 10-15mm, the desired EcoRI/BamHI fragments were isolated by electroelution.50The EcoRI/BamHI fragments containing Rsa A and Rsa B were end labeledusing DNA polymerase I (Kienow fragment). Approximately 300-500 ng of eachprobe was labeled in a 20 tl reaction volume containing 20 mM Tris HC1, pH=8.0, 7mM MgC12,40 .tCi [a32P]dATP (3000 Ci/mmol) and 5 U DNA polymerase I(Kienow fragment). The reaction was incubated for 15 mm at room temperaturebefore the probe was precipitated with ethanol. The radioactive probe was pelletedfor 30 mm in a microfuge and washed 5-6 times with 70 % ethanol to remove theunincorporated label. The pellet was dried briefly and dissolved in 20 p1 TE. Thespecific activity was determined by precipitating a liii sample with 5%trichioroacetic acid. The precipitable material was collected by suction filtrationthrough glass fibre filters (Whatman) and counted. The probe was then dilutedwith TB just before use to 10,000 cpm/iil.2.17.2 Synthetic OligonucleotidesThree pairs of synthetic oligonucleotides, B oligo site I(5’AGCTTTCATATATTATTAAGACTAATAAAGATACAA3’ and5’AGCTTTGTATCTTTATTAGTCTTAATAATATATGAA3’), B oligo site II(5’AATTCATAGAAATATAATATTACATATAGATTTAAGAAATAG3’ and5’AATTCTATTTCTTAAATCTATATGTAATATTATATTTCTATG3’) and FREBPoligo A (5?GATCCGGGAGCTGCATCCGGAGTAGG3 and5’GATCCTACTCCGGATGCAGCTCCCGG3’) were synthesized (ABI 391 DNAsynthesizer) and annealed before being used in competition EMSA experiments.51III. Results3.1 Replication of a Defective LR Minigenome in COS-7 and LA9 CellsIn order to study the sequence requirements for MVM DNA replication, atransient DNA replication assay was developed. In this assay system, a plasmidencoding a defective genome or minigenome of MVM(p) is co-transfected with avector expressing the viral encoded enzymes into cells permissive for MVM(p)DNA replication. Thus, the ability of the minigenome to replicate is dependent onhaving the essential sequences present on the plasmid containing the minigenome.Previous studies have shown that plasmids containing a single copy of either theleft or right palindrome were incapable of replicating in permissive cells(Merchlinsky et a!., 1983; Salvino et a!., 1991). The minigenome encoded by thepPTLR plasmid contains 411 nt of the left terminus fused to 807 nt of the rightterminus (Fig. 7). Thus the majority of the viral coding sequences have beendeleted. The hypothetical sizes of the monomer and dimer replicative forms,termed mLR and dLR, respectively, of the minigenome is expected to beapproximately 1.2 and 2.3 kbp respectively.In order to determine if the minigenome encoded by pPTLR replicates inCOS-7 or LA9 cells, pPTLR linearized at the unique EcoRI site was co-transfectedwith either pMM984 or pCMVNS-1 into the two cell lines (Fig. 8). Low molecularweight DNA was isolated from transfected cells and analyzed by Southern blotanalysis. When pPTLR was transfected in the absence of vectors expressing viralproteins, no replicative intermediates were seen at the expected molecular weight.When pPTLR was co-transfected with either pMM984 or pCMVNS-1,-I——0IliiiIRHIS-——IC-)(I)o Z00D_ D.(I)U) 0 Z>r >000D_ cL 0.—kDa20097684329*- NS-1123 45659clone, pCA4.O, was used rather that the SV4O based pCMVNS-1 (Fig. 9, lane 2 vs. 3)since the infectious clone is able to replicate in these cells. When the result of thewestern blot is compared to that of the replication assays, it appears that the level ofviral DNA replication is more or less proportional to the level of NS-1 expression.Since replication of the LR minigenome was more extensive in LA9 cells withpCA4.O and in COS-7 cells with pCMVNS-1 than the reciprocal experiments,subsequent co-transfection experiments were carried out using pCA4.O in LA9 cellsand pCMVNS-1 in COS-7 cells.3.2. Analysis of the Termini of the mLR Replicative IntermedatesIt was observed that at least 4 bands were present at the mLR position(designated mLR1, mLR2, mLR3 and mLR4) and at least 3 bands in the dLRposition (designated dLR1, dLR2, and dLR3) (Fig. 8A and B, lanes 5, 6 and 11).Since previous analysis of intracellular viral DNA showed that the termini existedin both the covalently closed hairpin form and open extended form (Astell et al.,1985; Cotmore et al., 1989), it was suspected that the multiple banding patterns atthe mLR and dLR position reflected this heterogeneity at viral termini ofminigenomic replicative intermediates found in co-transfected cells. In order toinvestigate this possibility, DNA samples from Hirt lysates were digested witheither exo III, an enzyme which digests double stranded DNA progressively fromeither a blunt end or 3’ recessed end to single stranded DNA, or with exo IIIfollowed by Si nuclease, an enzyme which digests single stranded DNA (Fig. 10).After exhaustive digestion with exo III, all mLRs were degraded to ssLR except formLR4 in each set of Hirt samples (Fig. 10, lanes 6, 8 and 11). This experiment showsthat mLR4 is devoid of 3’ OH ends suggesting that this intermediate60pCA4.0 pCA4.0 pCMVNS-l pCMVNS-iEcoRl pPTLR pPTLR pPTLRI II II IIExolli -++-++-++-++Si23.0 —1.1 —0.9 —123I ILPlasmid DNA A94—mMVM—I PFig. 10. Exonuclease III! Si Nuclease Digestion Patterns of the Minigenomic LRReplicative Intermediates.Hirt DNA samples of the indicated transfection experiments or EcoRI linearizedpCA4.0 plasmid were analyzed by nuclease digestions. Either undigested (lanes 1,4, 7 and 10) , exolil digested (lanes 2, 5, 8 and ii) or exolli and Si digested (lanes 3,6, 9 and 12) DNA samples were analyzed by Southern blotting as described in theMaterials and Methods.9.4 —6.7 —4.3 —2.3 —2.0 —‘4dLRmLRI 4—ssLR‘I.9 10 11 12IICOS-7Fig. 11. Heat Denaturation/Quick Chill Properties of Minigenomic LR ReplicativeIntermediatesHirt DNA samples from A9 transfected cells (A) or COS-7 transfected cells (B) werethermally denatured by incubation in boiling water for 5 minutes followed by arapidly chilling in ice water for 5 minutes (lanes 2 and 4). These DNAs along withuntreated samples (lanes 1 and 3) were analyzed by gel electrophoresis andSouthern blotting. (C) indicates the proposed structures for the various mLRspecies. The rabbit ear hairpin structure represents the left hairpin, and the straighthairpin structure represents the right hairpin. The arrow tip represents the 3’OHend of the DNA molecules, a substrate for exo III digestions.61A BF-FC. ++F- -ç FC.I II I I II23.0—a23.0—I t,2.3 — *—ssMVM 2.3 —2.0—4—dLR 2.04—ssLR1234.4— IPdLRmLR4— ssLR1 262CmLR1___________mLR2mLR3___________ ____________________mLR4Fig. I1C63contains covalently closed hairpins at both termini whereas the other intermediatesmust contain at least one open Y OH end. It was also observed that a portion ofmMVM in transfected LA9 cells (Fig.9, lanes 4) and other higher molecular weightminigenomic intermediates in transfected COS-7 cells (Fig. 9, lanes 10-12) wereresistant to exo III. The single stranded ssLR was confirmed to be mostly singlestranded DNA since it was removed by Si nuclease. In order to further characterizethese intermediates, Hirt DNA samples were boiled and rapidly cooled on ice waterto test for the presence of closed hairpin molecules (Fig. 11). Under theseconditions, mLR1 molecules were denatured to ssLR material whereas mLR2, mLR3and mLR4 were resistant to this treatment. Results from samples from transfectedLA9 (Fig. hA) and COS-7 cells (Fig. 11B) were the same. This analysis shows thatmLR2, mLR3 and mLR4 molecules can snap back to their double stranded form viathe hairpin sequences. These data also confirm that mLR4 contains covalentlyclosed hairpins at both termini. Furthermore, these data in conjunction with that inFig. 10 suggest that mLR2 and mLR3 contain one covalently closed hairpin at oneterminus (boil/chill resistant) and an open extended terminal at the oppositeterminus (exo III susceptible). Since the size of the right terminal hairpin (206 nt) islarger than the left terminal hairpin (1 l5nts), it is likely that mLR2 contains theclosed left hairpin and open extended right terminal and mLR3 contains the reverseconfiguration (Fig. 11C). Since it was also observed that a portion of replicativeintermediates were exo III resistant at the dLR level (in both A9 and COS-7 cells)and higher multimer levels (in COS-7 cells), it suggests that there may also existprocessed hairpin forms at each concatemer level, supporting the notion that baseconcatemers are processed to secondary concatemer forms.643.3. Replication of Left Hand (3’) Terminal Deletion Mutants.All parvovirus sequences determined thus far contain terminal palindromicsequences which can potentially fold into a stem plus arms structure (Astell, 1990).Deletion analysis of the right hand (5’) terminal palindrome of MVM(p) suggeststhat the potential to form the stem plus arms structure is required for viral DNAreplication (Salvino et a!., 1991). Furthermore, analysis of AAV terminal repeatsalso suggests that the structure of the stem plus arms is required for viral DNAreplication (Lefebrve et a!., 1984). In order to determine the structural requirementsof the 3’ terminal palindrome, a nested set of unidirectional deletions originating atthe extreme 3’ terminal palindrome was constructed (Fig. 12). Deletion clones withdeletions of 11, 25, 30, 56 and 77 nt, designated d112, d126 d131, d157 and dl 78,respectively, of the viral DNA sequences were chosen. Removal of the first 11 ntremoves only a small portion of one strand of the stem of the hairpin. Deletion of 25or 30 ntis predicted to remove a greater portion of the stem structure and alter orremove the bubble structure found in the middle of the stem. Further deletion of 56and 77 nt removes completely one strand of the stem and bubble feature, but theformer deletion only removes one arm while the latter deletion removes both arms.These deletions were then assembled into minigenome constructs similar topPTLR (except that the original left terminal fragment is shorter by 3 bp, seeMaterials and Methods) and transfected into LA9 and COS-7 cells . When thedeletion mutants were co-transfected into LA9 cells with pMM984, all minigenomeswere able to replicate to approximately comparable levels (Fig.12B). It wassurprising to see that deletion mutants in which either65T GGTCACACGTCACTTACGTG CAGTGTGCAGTGAATGCAGCG I AcGc TAG dl78d112A 3’ACATGGTTGGTCAGTTCTAAAAATTGTACCAACCAGTCAAGATTTTTACTATAFig. 12. Replication of Left Hand (Y) Terminal Deletion Mutants.The schematic diagram of the 3’ hairpin sequence is shown (A) with the deletionendpoints indicated by the arrows. Minigenomic constructs containing thedeletions in (A) were transfected into LA9 (B) and COS-7 cells (C, D and E). ThemRF and dRF of wild type MVM(p) (mMVM and dMVM) and the mutantminigenomes (mLR and dLR) are as indicated. IP represent input plasmid DNA.(A) and (B) show undigested Hirt DNA samples from transfected LA9 and COS-7cells. In addition, Dpn I (D) and Mbo I digested (E) Hirt DNA samples fromtransfected COS-7 cells are shown. In the case of the Mbo I digest, the Southern blotwas probed with a 32P radiolabeled 0.4 kbp BamHI/Ec0RV fragment containing the3’ terminal sequences. The bands indicated on this figure identify the 3’ bridgedimer fragment (3’ br) and the covalently closed and open extended 3’ hairpin.Although the pPTLR and pPTLRd112 minigenomes produce both types of 3’termini, pPTLRd126 and pPTLRd131 only generate covalently closed 3’ hairpins.Adl57Al31 d126(D-OCi)IIIIpMM984IpPTLR_________________________pMM984+pPTLRpMM984+pPTLRdI120pMM984+pPTLRdI26StiipMM984+pPTLRdI31p.•1pMM984÷pPTLRdI57•-pMM984+pPTLRd78LJU143ciIc “3 no-&a0C3IILiC30)-C;)IIII-LC CD CD CD a0)—1ItIC-)S11pCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1+pPTLRd[12pCMVNS1+pPTLRdI26•pCMVNS1+•pCMVNS1+pCMVNS1+pPTLRdI78U -uLJU3a.0-’I’Jo-LIII’.)_____SI— CgaiOCZIICA)0).CA)IIIIcipCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1+pPTLRdI12pCMVNS1+pPTLRdI26pCMVNS1+pPTLRdI31pCMVNS1+pPTLRdI57pCMVNS1+pPTLRdI78L_J30.IpCMVNS1$pPTLRSpCMVNS1+pPTLRIpCMVNS1+pPTLRdI12IpCMVNS1+pPTLRdI26SpCMVNS1+pPTLRdI31pCMVNS1+pPTLRdI57pCMVNS1+pPTLRdI78C r)irHOC.)II.)(D)O)-1.(IIIOP4C)1CD-.IIC,) ,I(7’0)7C 0 0.CO CD C,)CD 0.m—1S III liiiC.)C-U -a70one arm (pPTLRd157) or two arms (pPTLRd178) were removed from the stem plusarms structure were able to replicate as effectively as pPTLRd112 which onlyremoves a small portion at the base of the stem (Fig. 12B). In addition, it appearsthat the mLR intermediates of all the mutant 3’ palindromes are of approximatelythe same size. This result suggests that all mutants were able to repair their hairpinsequences to the wild type sizes. Although it is easy to envision that the pPTLRd112mutant is able to repair its hairpin sequences by using the existing hairpin to primeDNA synthesis to fill in the rest of the missing sequences, this mechanism can notbe applied to the pPTLRd178 mutant since the possibility of hairpin formation ishighly unlikely. It is more probable that the repair of pPTLRd178 mutant mayinvolve a recombination mechanism which uses wild type MVM(p) sequencesderived from pMM984 as a template for repair.In order to determine the replication levels of the 3’ palindrome deletionmutants in the absence of the wild type viral sequence, the minigenomic constructswere co-transfected into COS-7 cells with pCMVNS-1 (Fig. 12C, D, E). Replicationof pPTLRd112 replicates to levels similar to pPTLR (Fig.12C, lane 3 vs. 4). Incontrast, pPTLRd126 and pPTLRd131 replicated to a very low level (1-5%) (Fig.12Cand D, lanes 3 vs 5 and 6). Interestingly, there is a considerable level of the dLRgenerated by these mutants despite exhibiting relatively low levels of mLR. It alsoappears that only mLR2 and mLR4 (arrowheads, Fig. 12C, lanes 5 and 6) are theonly detectable monomers species generated by these two mutants. BothpPTLRd157 and pPTLRd178 also generated very low levels of mLR and nodetectable dLR (Fig.12C and D, lanes 7 and 8). Thus, in the absence of wild typeviral hairpin sequences, only pPTLRd112 was able to replicate to levels comparablewith pPTLR. It is probable that the pPTLRd112 mutant is able to repair its 3’ hairpinsequences probably by primer extension of the remaining hairpin structure. It was71interesting that both pPTLRd126 and pPTLRd131 could not replicate as efficiently aspPTLR or pPTLRd112 even though the formation of a useable hairpin at the 3’ endwould be predicted to form. Both pPTLRd126 and pPTLRd131 mutations alsodisrupt the bubble structure located at nt 25 and 26. If the hairpin is repaired byprimer extension, the mismatches which form the bubble in the wild type sequencewill be altered such that duplex DNA will be generated instead. Since thesemutants are able to generate dLR, but not all species of mLR, it may be possible thatthese mutants exert their effect during the resolution of dLR. This hypothesis isconsistent with data from a Southern blot of Mbo I digests of Hirt DNA probed theleft (3’) terminal sequences (Fig 12E). Clearly pPTLR gives rise to three prominentbands corresponding to the 3’ dimer bridge (3’ br, 1 .4kbp), open left hairpinfragment (0.81kb) and the covalently closed left hairpin (0.75 kbp) fragment (Fig12E, lane 3). Similarly, pPTLRd112 generated the same banding pattern but thefragments migrate at lower molecular weights since an extra Mbo I site wasintroduced during the cloning procedures (Fig 12E, lane 4). These data suggest thatboth pPTLR and pPTLRd12 clones generate both the open extended (0.41 kbp) andcovalently closed hairpin forms (0.35 kbp) at the left terminus. In contrast,pPTLRd126 and pPTLRd131 produced a 3’ dimer bridge fragment (3’ br, 0.7 kbp)and a band which co-migrates with the covalently closed hairpin. Interestingly, theband corresponding to the extended left hairpin is not present, suggesting thatmLRs with extended hairpin conformation are either not present or present atundetectable levels. It may be possible that replicative intermediates containing anopen extended left end might not be generated because the resolution of dLR typemolecules may be affected by the sequence changes during the repair of the 3’palindrome of these mutants. Further deletion of one (pPTLRd157) or both(pPTLRd178) arms of the left hairpin clearly inhibits viral DNA replication evenfurther. Although no dimer replicative intermediates were observed, a low level of72a single species of mLR exists. It may be possible that these intermediates areexcised from the plasmid sequences and are allowed to proceed with synthesis of amonomer species by an unknown mechanism possibly using the intact right handhairpin sequences to prime DNA synthesis. The structures of these intermediateswere not examined.3.4. Analysis of Internal Right End (IRE) DeletionsAlthough it is clear that the terminal palindrome sequences are essential forMVM(p) DNA replication, it is not clear whether sequences internal to the right orleft palindromes is required. Some evidence has been reported that the 65 bptandem repeat near the right terminus was important for DNA replication (Salvinoet a!., 1991). It is curious to note that the wild type immunosuppressive variant ofMVM(p), MVM(i), has only one copy of the 65 bp element (Sahli et a!., 1985; Astell eta!., 1986). It was shown that MVM(i), which normally infects T lymphocytes, couldacquire fibrotropism by changing codons 317 and 321 of the MVM(i) VP-2 gene tothe MVM(p) equivalent codons (Ball-Goodrich and Tattersall, 1992). Thisfibrotropic mutant virus and wild type MVM(p) were able to replicate their DNAequally well in LA9 fibroblasts (Ball-Goodrich and Tattersall, 1992). Although thisresult suggests that the 65 bp repeat may not be essential for MVM DNA replicationits role has not been elucidated. In order to perform a more comprehensive deletionanalysis in this region, a nested set of unidirectional exo III deletions was created inthe LR minigenome using the unique Xba I site as the origin of deletion (Fig.13B).In A9 cells it was shown that a LR minigenome containing a deletion of nt 411-4436—‘C———014uIIIIIIIIIIiuIpolyA(+)—Ildl4ll-4436II d1411-4489Id1411-4636BIREId1411-4695DeletionsI—d1411-477IId1411-4806Id1411-4853dl411-525(ILErIId1259-4342‘DeletionsLIId1140-4342Fig.13.SchematicDiagramofInternalRightEnd(IRE)andLeftend(ILE)DeletionMutants.(A)representstheparentalpPTLRconstructusedtogeneratetheIRE(B)andILEdeletion(C)mutants.Thesizeofeachdeletion(dl)isindicatedbythesizeofthehorizontalbarandthecorrespondingnucleotidepositions.IREdeletionsweregeneratedusingtheXbal(4342)siteastheoriginofunidirectionalexoifideletions.ILEdeletionswereobtainedbyfusingthe1-lincIIsite, adjacenttothePstI(411)site,witheithertheNcoI(259)orStyI(140)siteoftheMVMgenome.741 2 3 4 5 6 7 8 91011Fig. 14 Replication of Internal Right End (IRE) Deletion Mutants.The IRE deletion mutants (see Fig.13 for description) were linearized with EcoRIbefore transfection into either A9 cells (A) or COS-7 cells (B). Hirt DNA sampleswere analyzed by gel electrophoresis and Southern blotting. The immobilizedDNA was probed with a left terminal BamHI/Ec0RV (381) 32P radiolabeledrestriction fragment. Replication was monitored by the appearance of mLRs.dA.D If r in00 r c’ 0 inD 0000-: .• . .. — . -i— — — — — — —-- - - -+ + + + + +++00000qqcqq•123.0 —9.46.7 —4.3 —41nput pCA4.0.4—mMVMDip.4— ssMVMmLREl ssLR1.10.9pCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1+pPTLRd1411-4436pCMVNS1+pPTLRd1411-4489pCMVNS1÷pPTLRd1411-4636pCMVNS1+pPTLRdI411-4695pCMVNS1+pPTLRd1411-4778pCMVNS1+pPTLRd1411-4806pCMVNS1+pPTLRd1411-4853CIIN)M4aOjJCciJ(J-LDIIIIII(‘3(1100I. CC),IILJLJ76and nt 411-4489 (Fig. 14A, lane 4 and 5) did not affect replication compared to theLR control (lane 3), but when the deletion was extended to nt 4636 (Fig. 14A, lane 6),replication was abolished. Similarly, in COS-7 cells (Fig.14B), LR minigenomescontaining deletions of nt 411-4436 and nt 411-4489 (Fig. 14B, lane 4 and 5)replicated as well as the pPTLR control (lane 3). Unlike in LA9 cells, the LRminigenome deletion of nt 411-4636 was able to replicate, but at a much lower levelthan that of the pPTLR minigenome in COS-7 cells (Fig.14A and B, lane 6).Extension of the deletion to nt 4695 (Fig.14 B, lane 7) totally abolished replication ofthe LR minigenome. Although ability of d1411-4636 to replicate in COS-7 cells butnot A9 cells may reflect that the COS-7 cell system is a more sensitive replicationassay system than the A9 cell system, the basis for this difference is unknown. Inaddition, the data from this analysis suggested to us that there may be two specificDNA elements found between nt 4489-4436 (element A) and 4636-4695 (element B)which are required for efficient MVM DNA replication.In order to confirm this hypothesis, I attempted to rescue the replicationphenotype of d1411-4636 and d1411-4695 by inserting the MVM(p) Xba I(4342)/Sau3a (4741) fragment (SX fragment) into the unique EcoRV (381) site, a sitecommon to all the IRE deletion minigenomes (Fig. 15). When the MVM(p) specificfragment was inserted into the d1411-4636 and d1411-4695 mutants, replication wasrestored to control LR levels in both A9 (Fig. 16A, lanes 4-6) and COS-7 cells (Fig.16B, lanes 4-6). It was also observed that the ability of the MVM(p) specificfragment to rescue the dl4ll-4636 mutant was orientation independent since d1411-4636SXL and d1411-4636SXR replicated to similar levels in A9 and COS-7 cells (Fig.16A and B, lanes 4 vs 5). In addition, I also tried toBSXRISXLIIIBIIAIBIIAlSVRSV400riSNEnhancersI-’CILHlIInsertIIRHI--IL.polyA(-I-)Fig.15.SchematicDiagramofDNAFragmentsUsedtoRescueReplicationofIREDeletionMutants.(A)RepresentsthepPTLRconstruct.(B)TworestrictionfragmentswereusedtorescuetheIREdeletionmutants.TheMVMinsertistheXbaI(4342)ISau3a(4741)fragmentwhichcontainsbothelementsA(nt4489-4636)andB(nt4636-4695).SXRandSXLinsertionsindicatetherelativeorientationoftheXbaI(4342)/Sau3a(4741)restrictionfragment.TheSV4Oinsert(SVR)wastheBglI(5235)/PvuII(270)fragmentoftheSV4Ogenome.ThisfragmentcontainstheentireSV4Ocontrolregionincludingthetranscriptionalenhancersfoundinthe72bptandemrepeat(stippledboxes),6SP1bindingsites(crosshatchedbox)andminimaloriginofSV4ODNAreplication(blackbox).(C)AsingleEcoRV(381)siteintheIREdeletionmutantswasusedtocloneeithertheSV4O(SVRmutants)orMVM(SXRorSXLmutants)restrictionfragments.AILHl0zIIIIIBrMC’,liiiIAIBI__.___.L.polyA(+)78--+ +COCC,,I,,+C12 3 45 6Fig. 16. Replication of IRE SXL and SXR Rescue Mutants.The MVM(p) specific XbaI(4342)/Sau3a(4741) SX fragment was cloned into IREdeletion mutants d1411-4636 and d1411-4695 in either orientation (Fig.15). Theresulting SXL and SXR mutants were linearized with EcoRI before transfection intoA9 cells (A) or COS-7 cells (B). Hirt DNA samples were analyzed by Southernblotting and probed as in Fig.14.A-(ID‘-4+23.0—11L4.. mMVMIP14— ssMVMmLR—pCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1+pPTLRdI411-4636SXRpCMVNS1+pPTLRd1411-4636SxLpCMVNS1+pPTLRd1411-4695SXL1%.)O(iIIIIIJII80A‘.0 tr 00 ‘.Drro r— cu‘.D’.O F— 0000_4 — — —qoqq q qouciuc) U23.0—43- 4— mMVM23- UssMVMU El mLRSVR1.1— ImLR0.9—,123456789Fig. 17. DNA Replication of IRE SVR Rescue Mutants.The SV4O control region was cloned into replication defective IRE deletion mutantsas well as the parental pPTLR vector (Fig.15). The resulting SVR mutants werelinearized with EcoRI before transfection into either A9 cells (A) or COS-7 cells (B).Hirt DNA samples were analyzed by Southern blotting and probed as in Fig. 14.CNr’JOwIIIIII__pCMVNS14pPTLR_______pPTLRSVRpCMVNS1+pPTLRpCMVNS1-i-pPTLRSVRpCMVNS1-f-pPTLRd1411-4636SVR.pCMVNS1+pPTLRdI411-4695SVRpCMVNS1+pPTLRd1411-4778SVp.pCMVNS1÷pPTLRd1411-4806SVRIpCMVNS1+pPTLRd1411-4853SVRLJLJ44LJ82rescue replication of the defective minigenomes by introducing the wellcharacterized SV4O control region (Fig. 15). In Fig. 17, the results show that theSV4O control region was not able to rescue the replication defective IRE deletionmutants in either A9 cells (Fig. 17A, lanes 5-9) or COS-7 cells (Fig. 17B, lanes 7-10)since no replication intermediates were detected at the expected sizes. It was alsonoted that the d141 1-4636 SVR mutation replicates at a low level in COS-7 cells (Fig.17B, lane 6) but not A9 (Fig. 17A, lane 7) cells. The parental minigenome, d1411-4636, also exhibited the same phenotypes (Fig.14B, lane 6). It was concluded thatthe insertion of the SV4O fragment is a neutral mutation since its insertion into theLR control or dl4ll-4636 did not affect the replication in COS-7 cells. Insertion ofthe SVR fragment into pPTLR appeared to have a small negative influence on thereplication of the resulting minigenome, pPTLRSVR, when the results arenormalized with respect to the input plasmid (II’) DNA band (Fig.17A, lane 3 vs.lane 4). It was predicted that the SV4O origin of replication would have a positiveeffect if any on the replication of SVR mutants in COS-7 cells since they express theSV4O T-Ag which initiates SV4O DNA replication. However, this was not apparentin Fig. 17B and may be due to the fact that the input DNA was linearized prior totransfection. The last two experiments, taken together, suggested that thereplication defective phenotype of the two deletion mutants, d1411-4636 and d1411-4695, was not due to the fact I have reduced the size of the minigenome, but due tothe deletion of specific MVM(p) DNA elements, A and B, which lie between nt4489-4636 and 4636-4695, respectively.In order to further define the cis-acting sequence in this region, two small RsaI fragments, designated Rsa A (nt 4489-4579) and Rsa B (nt 4579-4665), found--AlRsaAIRsaBINcoIEcoRVRsaIRsaIRsalSau3aSau3a___________25938144314579466247414806pPTLRILHIIIIIIIIIRHIAIBILpolyA(+)NcoIEcoRVSau3aSau3a25938147414806___________________pPTLRdI4II-4695RVILHIIIIIIRHILpo1yA(--)NcoISau3aSau3a259.4.47414806__________________pPTLRd14H-4695ARILHIIIIIRHftsaALpolyA(+)NcoISau3aSau3a__________25947414806__________________pPTLRdI4I1-4695ALILHI1‘RAIIRHIsaL.polyA(+)NcoISau3aSau3a__________25947414806__________________pPTLRd14II-4695BRILHIi1IFRHI1RsaBLi..polyA(-t-)NcoISau3aSau3a__________25947414806__________________pPTLRd14II-4695BLILHIIIIIRHI[saBIL..polyA(+)Fig.18.RsaAandRsaBActivatesDNAReplicationofd1411-4695IRE.(A)Schematicdiagramofd1411-4695minigenomicconstructswhichcontaintheRsaA(nt 44314636)andRsaB(nt46364695)insertsclonedintotheuniqueEcoRV(381)siteisshown.ElementsA(nt44894636)andB(nt46364695)areboxedandareshownbeneaththepPTLRconstruct.ThecloningstrategyisasoutlinedinFig.15.TheRsaAandRsaBfragments(boxed)wereclonedineitherRorLorientationasindicatedbythearrowabovethebox.AllminigenomicconstructswerelinearizedwithEcoRlpriortotransfection.SouthernblotsofHirtDNAsamplesisolatedfromtransfectedCOS-7cells(B)undigestedand(C)digestedwithDpnIareshown.TheDpnIresistantmonomer(mLR)replicativeformsareasindicated.r’) C,)z 0.01C’) CD4 0.CD 0000-9)DD.OC,))o).I,IIIIII’.) Ci)pCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1+pPTLRdI411-4695RVpCMVNS1+pPTLRdI411-4695SVRpCMVNS1+pPTLRdI411 -4695SXRpCMVNS1+pPTLRdI411-4695ARpCMVNS1+pPTLRcII411-4695ALpCMVNS1+pPTLRdI411-4695BRpCMVNS1+pPTLRdI411-46958LV 0.01Coo,CD 0 CDCo-.0nP CD-.0Ci)CCi)0).CA)IIIIIIIIIpCMVNS1pPTLRpCMVNS1+pPTLRpCMVNS1÷pPTLRdI411-4695RVpCMVNS1÷pPTLRdI411-4695SVRipCMVNS1+pPTLRdI411 -4695SXRpCMVNS1+pPTLRdI411-4695ARpCMVNS1+pPTLRdI411-4695ALpCMVNS1+pPTLRdI411 -4695BRpCMVNS1+pPTLRdI411-4695BLdl3 I 086within the complementing SX fragment were cloned into the unique EcoRV (381)site of the pPTLRd141I-4695(RV) minigenome (Fig.18A). Fragments which are ableto activate DNA replication are predicted to either partially or fully rescue thereplication phenotype. As expected, when the XbaI(4342)/Sau3a(4741) (SX)fragment was inserted, the replication phenotype was fully restored (Fig. 18B andC, lane 6). When the SV4O origin of replication was inserted, the minigenomeremained replication deficient (Fig. 18B and C, lane 5). Again, this confirmed thatthe elements contained within the SX fragment were specific for MVM(p) DNAreplication. The presence of the Rsa A fragment (nt 4431-4579) appeared to activatereplication to approximately 60% of the control LR levels in the “R” orientation (Fig.18B and C, lane 7), but only 7% in the opposite “L” orientation (Fig. 18B and C, lane8). Furthermore, the presence of the adjacent Rsa B fragment (nt 4579-4662) alsostimulated replication to approximately 20-30% of LR levels in either orientation.These results suggested the elements within the two Rsa I fragments are specificallyrequired for efficient DNA replication of MVM(p) minigenomes. Since the largerSX fragment was able to fully rescue replication levels to LR levels and neither ofthe two Rsa I fragments were able to do so, it appears that the effects from the twoRsa I fragments may be additive.3.5. Replication of Internal Left End (ILE) DeletionsSince cis-acting sequences were found distantly inboard of the right hairpin,two left end internal deletions were constructed to determine if other DNAelements found distantly inboard of the left hairpin are required for efficient DNAreplication (Fig. 13). The d1259-4342 construct deleted the entire87A BH+ + ++ + ÷ 4 ,—4 ‘-4 —4qcqq23O4— mMVM 6.7 —I‘P -— ssMVMdLR : dLR1mR12345 12345Fig. 19. Replication of Internal Left End (ILE) Deletion Mutants.Two ILE deletion mutants d1259-4342 and dll4O-4342 were constructed byremoving the Nco I (259)/Hinc II or Sty I (140)/Hinc II restriction fragments frompPTLR (see Fig. 13C). The resulting plasmid constructs were linearized with EcoRIand transfected into LA9 (A) and COS-7 cells (B). Flirt DNA samples were analyzedby Southern blot analysis using a 32P labeled probe generated with the rightterminal Xba I(4342)/BamHI fragment.88NS-1 and NS-2 ORFs while the d1140-4342 construct further deleted the promoterelements including a functional Spi binding site found within P4 (Ahn et al., 1989;Pitluk and Ward, 1991). The results show that the d1259-4342 mutation wasreplication competent when compared with the LR control minigenome in both A9and COS-7 cells (Fig.19A and B, lane 3 vs 4). However, the d1140-4342 mutantreplicated to at most 60-70% of the control levels (Fig.18A and B, lane 3 vs 5). Thesedata suggested that deletion of the transcriptional control elements found in P4 mayreduce the efficiency of minigenome replication, but the signals for processing ofthe viral hairpins were still present since the multiple banding patterns at the mLRlevel remains unchanged. Interestingly, deletion of Spi binding sites within theSV4O control region also reduces but does not abolish DNA replication in vivo(Hertz and Mertz, 1986). It is possible that Spi may provide the same function inboth cases. This interpretation does not exclude the possibility that there may be asyet unidentified transcriptional control or replication elements found in the hairpinsequences.3.6. Replication of Identical End GenomesAlthough MVM(p) belongs to the autonomously replicating parvoviruseswhich contain two non-identical terminal hairpin sequences and package primarilyminus strand viral genomes, other parvoviruses such as AAV and B-19 containinverted terminal repeats (which result in identical hairpin termini) and packageboth plus and minus strands in approximately equal proportions. Can MVM(p)also engage in a mode of DNA replication that is similar to that of AAV? To thisend, two plasmids (pPTRR and pPTLL) were constructed to encode twominigenomes which contains either two right termini or two left termini (Fig.20).The RR minigenome is approximately 1.9kbp and consists of two copies of the right89terminal 807 bp fragment oriented as inverted terminal repeats separated by a 0.3kbp Hind III(3996)/Xba 1(4342) restriction fragment. The LL minigenome isapproximately 4.7 kbp and contains two copies of the left terminal 408 bp fragmentoriented as terminal inverted repeats separated by a 3.9 kbp Alu I (408)/Xba 1(4342)fragment. When these plasmids were transfected into COS-7 cells in the absence ofpCMVNS-1, no additional bands corresponding to the expected replicationintermediates were seen. However, when these constructs were co-transfected withpCMVNS-1, replication intermediates migrating at the predicted sizescorresponding to the monomer and dimer RF’s of the LL (mLL and dLL) and RR(mRR and dRR) minigenomes appeared (Fig. 20A, lanes 5, 6 and 7). Thesereplicative intermediates were shown to be both Dpn I resistant (Fig. 20B, lanes 5, 6and 7) and Mbo I sensitive (Fig. 20C, lanes 5, 6 and 7), confirming that theseintermediates were replicated in COS-7 cells. It was not surprising that replicationof the LL and RR minigenomes was dependent on NS-1 since it has been shownthat NS-1 is required for MVM DNA replication. This result also confirms that NS1 plays a role in the replication of both the right and left termini. These resultsagree with other experiments showing that NS-1 is able to resolve concatemerjunction fragments containing the Y and 5’ terminal DNA sequences (Cotmore et al.,1992; Cotmore and Tattersall, 1992; Cotmore et a!., 1993). These results also confirmthat each terminus contains a latent origin of DNA replication. Although both theLL and RR minigenomes replicated in COS-7 cells, it was observed that only the RRminigenome replicated at similar if not higher levels than the LR minigenome. Incontrast, the LL minigenome consistently replicated at much lower levels (<5%)than that of the LR and RR minigenomes. The reason for this difference isunknown, but it is possible that the replication origin at the right terminus may0000fl00pPl’LL.——.IIiIIIIii11111+14P4P4dEpFI’RRIRRIIIIIRBIpolyA(+)4i4-4-polyA(+)Fig.20.Replicationofidenticalendgenomes.Identicalendminigenomescontainingtwolefttermini(pPTLL)ortworighttermrni(pPTRR)areshownabove.TheplasmidswerelinearizedwitheitherwithSmaI(pPTLL)orEcoRI(pPTRR)beforeco-transfectionwithpCMVNS-1intoCOS-7cells.Undigested(A),Dpn1(B)orMboIdigested(C)FlirtsampleswereanalyzedbySouthernblotanalysis.Theinputplasmid(IP),mRF(m)anddRF(d)ofthevariantRRandRRminigenomesareindicated.Underthegelconditionsused,threemRRspecieswereobserved.OnlyonemLLintermediatewasobserved.AllputativereplicativeintermediateswereDpnIresistantandMboIsensitive.“3CC:C(JIIIIIIISpCMVNSISpPTLRIpPTLL5pPTRRpCMVNS1+PpTLRPCMVNS1+ppTLLPCMVNSI+ppTRRI’ll—.t——I.‘Thii-•wki—C•pPTLRpPTRRPPTLLpCMVNS1÷pPTLRpCMVNS1+PPTLLpCMVNS1÷pPTRR“3CN)N)I%JIIIIIIIIpPTLR$pPTRRI.IpPTLL-•.pCMVNS1+pPTLR•a4PCMVNS1i-PPTLL———94potentially be more active than that of the left terminus. Although the left terminalsequences provide resolution sites in Y dimer bridge fragments for dimerresolution, it might be predicted that the right terminal sequences contain a moreactive origin since this terminus serves as the origin of DNA replication for thesynthesis of higher concatemeric forms as well as single stranded genomic DNA.Since it was observed that mRR is composed of at least three bands, mRR1,mRR2 and mRR3, it was suspected that the terminal ends of the RR minigenomescontained various processed hairpin forms. When Hirt DNA samples from thepPTRR co-transfection were digested with exo III or with exo III followed by Sinudease, a similar digestion pattern was seen as with mLR (Fig. 21A). Again, onlythe smallest mRR, mRR3, was resistant to exolli (i.e. absence of 30H end) whereasmRRi and mRR2 were digested to smaller exoill digestion products, ei and ssRR(Fig. 21A, lane 6). These data suggest that mRR3 contains a closed hairpin structureat both termini. When these samples were subsequently digested with Si nuclease,two resistant bands migrating between el and ssRR were observed, esi and es2(Fig. 21A, lane 7). Since ssRR was degraded by Si nuclease, ssRR must be mostlysingle stranded DNA. In contrast, ei represents a product which contains bothsingle stranded and double stranded structures since el appears to be partiallyresistant to the Si nuclease treatment, generating the smaller double stranded esiand es2 products. I propose that el might represent an exolil resistant panhandlestructure that is present in the degradation pathway of mRR1 and/or mRR2. Toconfirm this, Hirt DNAs were boiled and rapidly cooled before gel electrophoresis(Fig. 2iA, lane 8). As suspected, both mRR2 and 3 were resistant to this treatment9523.0—9.4—6.7—12345678Fig. 21. Exo ITT/Si Nuclease Digestion Patterns and Heat Denaturation/Quick ChillProperties of the Minigenomic RR Replicative Intermediates.(A) Hirt DNA samples containing either minigenomic LR or RR replicativeintermediates were either digested with exolll (lanes 2 and 6), exoIll followed by Sinuclease (lanes 3 and 7), or boiled and rapidly chilled (lanes 4 and 8) before analysisby gel electrophoresis and Southern blotting. The exolil degradation products, eland ssRR (lane 6), and exoill/Si degradation products, esi and es2 (lane 7), areindicated. The band indicated by PH (lane 8) is predicted to be a panhandlestructure. (B) The proposed structure of mRRi, mRR2, mRR3 and PH is shownbased on data shown in (A). The arrowheads indicate (3’ OH) ends which aresusceptible to exolli digestion.ApCMVNS-1 pCMVNS-1+ +pPTLR pPTRRii 7exoill — + + — — + +Si - - +- —— ++4.3— IP LR—*2.3—2.01.1—0.9—dLR—*1mLRssLR—IPRRdRRmRRPe•4— esi4 es2.4—. ssRR96BmRRImRR2:R34aFig. 21B97showing that these intermediates contain at least one closed right hairpin end butthe heat denatured/quick chilled mRR1 material co-migrated with a similar band inthe undigested sample and migrated slightly slower with the exolil degradationproduct el. These data suggested that mRR1 contained two open extended righttermini. When mRR1 is heat denatured and rapidly cooled, a panhandle (PH)structure is predicted to form due to the presence of the inverted 5’ terminalsequence repeat in the duplex RR sequences. It may be possible that el is slightlysmaller than PH because of a limited exo III degradation of the terminal sequencesends during the exo III reaction. In Fig. 21B, the proposed structures of the mRRsbased on the exolil/Si degradation patterns, heat denaturation properties andelectrophoretic mobility are indicated. The presence of multiple mLLs were notobserved probably because the LL minigenome is relatively large and thus variousmLLs may not be resolved during electrophoresis.When the same transfection experiments were performed in A9 cells, eithervery low levels of replication (RR) or no replication (LL) were observed (data notshown). Although it is not know why this is the case, it may be possible that in theCOS-7 transfections, our helper plasmid is not an infectious clone, hence there is nocompeting replicating genome. When A9 cells are used, the infectious wild typehelper genome likely competes with the LL and RR minigenomes for replicationfactors. If this is the case, it suggests that the normal MVM(p) genomeconfiguration (LR) may be a more efficient replication template than the aberrantLL and RR minigenomes. There is other evidence for this line of reasoning. Studiesof MVM(p) defective interfering (DI) particles (Faust and Ward, 1979) showed thatthere are two types of DI particles, Type I particles contain the LR configurationwhereas Type II DI particles maintain the RR configuration. Upon further98undiluted serial passages, the population of the Type II particles is reduced relativeto the Type I particles suggesting that either the Type I genomes are preferentiallypackaged or Type I genomes have a competitive advantage during the replicationof the DNA. The data suggest that the latter explanation is more likely since the RRminigenome clearly replicates as efficiently as the LR minigenome only in theabsence of a competing LR genome.3.7. Activation of LL minigenome DNA replicationThe analysis of IRE deletion mutants suggested that there may be elementsfound within the SX fragment which are required for efficient replication of LRminigenomes. Since the SX fragment was able to rescue the replication of the dl4ll-4695 minigenome, it is possible that this fragment may also activate DNA replicationat the left terminus. To test this hypothesis, a series of MVM(p) minigenomes wasconstructed containing two left hand (3’) termini with either the activating SXfragment or an unrelated fragment from the yeast FUS3 gene (Fig. 22). First, thesmaller LL type minigenome, LLX (2.4 kbp), was constructed by deleting the internalXho I(2070)/Xba 1(4342) fragment from the large LL minigenome encoded bypPTLL. The LLX minigenome replicates poorly but consistently better than thelarger LL minigenome (Fig. 22C, lanes 5 vs. 6). A minigenome with one copy of theFUS3 restriction fragment replacing the internal 1.7 kbp EcoRV fragment of pPTLLX(LLFUS3.1) replicated at a very low level when compared to the LR control (Fig.22C, lane 11 vs 4). When four copies of the FUS3 fragment were inserted(pPTLLFUS3.4), no replication intermediates were detected (Fig. 22C, lane 12).99AEc0RV Xh T/V1- T\\”5tI4l381 0 J/ iwa \\ Ec0RV38IpPTLLX ‘fl i EpPTLLXS1F 1 I SX I—bpPTLLXS1R 1 I_ SX IpPTLLXS2R I SX I SX IpPTLLXS3F I SX I SX I SX IpPTLLFUS3.1 1 I FUS3 IpPTLLFUS3.4 E1 I FUS3 I FUS3 I FUS3 I FUS3 IFig. 22. Activation of Replication of MVM(p) LL Minigenomes.Plasmid clones encoding LL minigenomes are described in (A). The stippled boxesat the end of each line represents the left hairpin sequences. A deletion of the Xho I(2070) to Xba I (4342) restriction fragments from pPTLL (Fig. 20A) resulted in thepPTLLX clone. One (pPTLLXS1F or pPTLLXSIR), two (pPTLLXS2R) or three(pPTLLXS3F) copies of the SX replaced the 1.7 kbp internal EcoRV fragment of thepPTLLX vector. In addition, one (pPTLLFUS3.1) or four (pPTLLFUS3.4) copies ofthe FUS3 gene were also used to replace the internal EcoRV fragment of pPTLLX.The arrows below each SX or FUS3 fragment (box) indicate the relative direction ofeach insert. Southern blots of low MW DNAs isolated from transfected COS-7 cells(B) undigested or (C) Dpn I digested were performed as described in the Materialsand Methods. The monomer and dimer replicative forms of the LR (mLR and dLR)and RR genomes (mRR and dRR) are as indicated. The predicted monomer(triangle) and dimer (diamond) replicative forms of the various LL minigenomes isindicated to the left of each lane. Only two monomer forms are observed forminigenomes containing the LL type configuration.0.oCD rn-’Co rn 0.CD 0 -a -arF)c10’r)OO)i-Ci)0))IIIIIIt1,V‘V ‘V I VVIIYm•‘::I’,)-a I’.)(I)2-’J1pCMVNS1pPTLRpPTLLêpCMVNS1+pPTLRpCMVNS1+pPTLLpCMVNS1+pPTLLXpCMVNS1+pPTLLXS1FpCMVNS1+pPTLLXS1RpCMVNS1+pPTLLXS2RpCMVNS1+pPTLLXS3FpCMVNS1+pPTLLFUS3.1*‘S t41‘VC CpCMVNS1+pPTLLFUS3.4pCMVNS1+pPTRR30.0oi0)0.CDCDQCD-L0-L-Lr’)I’3r’JnnpCMVNS1pPTLRCD-’j-0C)Ci)0).C)IIIIIIIIrLrZpPTLLSpCMVNS1+pPTLRpCMVNS1+pPTLLpCMVNS1+pPTLLXpCMVNS1+pPTLLXS1FpCMVNS1+pPTLLXS1RIpCMVNS1+pPTLLXS2RJpCMVNS1+pPTLLXS3FpCMVNS1-i-pPTLLFUS3.1pCMVNS1+pPTLLFUS3.4pCMVNS1+pPTRR4£44(*L-LC)CL1J3102In contrast, when the internal 1.7 kbp EcoRV fragment of pPTLLX was replaced bythe SX fragment, the predicted Dpn I resistant monomers and dimers were easilydetectable (Fig. 22C, lanes 7 and 8). A single copy of the SX fragment activatedDNA replication approximately 30-60 times over that of the LLX minigenome asjudged by the Dpn I resistant material in the Hirt DNA samples (Fig. 22C, comparelane 6 with 7 or 8). It was noted that the LLXS1F and LLXSIR minigenomesreplicate equally well since they are identical minigenomes even though the SXfragment is oriented in opposite directions with respect to the plasmid sequences.Thus no comment can be made on the orientation dependence of the SX fragmenton the activation of LL type minigenomes. Although activation of the LLminigenomes by the SX fragment is unequivocal, some unexpected replicationintermediates were observed. When two (LLXS2R) or three (LLXS3F) copies of theSX fragment were used (Fig. 22C, lanes 9 and 10), monomer forms corresponding tothe LLXS1F or LLXSIR were present. These data suggested that LL minigenomescontaining multiple copies of the SX fragment may be able to excise copies of the SXfragment by either recombination or slipped mispairing during replication. Anintramolecular slipped mispairing mechanism has been implicated in thegeneration of deletion variants of parvovirus genomes (Faust and Hogan, 1990)I also observed only two monomer RF forms of the LLXS1F and LLXS1Rconstructs (Fig. 22B and C). In order to determine the nature of the termini ofmonomer replicative forms of the LLXS1F minigenome, Hirt DNA samples weredigested with exonuclease III (Fig. 23). Only mLLXS-2 was resistant to exonucleaseIll digest suggesting that both termini are covalently closed. To determine thestatus of mLLXS-1, Hirt DNA samples were denatured in boiling water andquickly cooled in ice water before gel electrophoresis. It was observed that both2.32.0V0CDFig 23 Exo III Digestion Patterns and Heat Denaturation/Quick Chill Properties ofMinigenomic LL Replicative Intermediates.Hirt DNA samples containing the indicated transfected DNA were eitherundigested (lane 1), exonuclease III digested (lane 2) or boiled and rapidly cooled(lane 3) before gel electrophoresis and Southern blot analysis. The mLLXS-2indicates a structure which contains covalently closed hairpin ends at both terminiwhereas mLLXS-1 contains one covalently closed hairpin end and one open duplex103pCMVNS1+pPTLLXS1 FC.)-o =x 0239.46.64.3—11.10.9 —mLLXS-1mLLXS-2C,terminus.104mLLXS-1 and 2 are both able to snap back suggesting that both molecules have atleast one covalently closed hairpin terminus. Taken together, mLLXS-2 mustcontain one open extended duplex palindrome terminus and one covalently closedhairpin terminus (Fig. 23 cartoon). The absence of detectable molecules containingtwo open duplex termini is consistent with what is known of MVM DNAreplication. First, the MVM left hairpin cannot undergo hairpin transfer like AAV(Fig. 5) to produce extended duplex palindrome sequence at either end since allends are required to remain in the flip sequence orientation. Second, the proposedasymmetric resolution mechanism of the Modified Rolling Hairpin Model of MVMDNA replication (Fig. 4) predicts that one covalently closed left (3’) hairpin isproduced in each dimer resolution reaction (Astell et a!., 1985).It is also interesting to note that minigenomes containing two right terminigenerate three monomer forms (Fig. 21) while minigenomes containing two lefttermini generates two monomer forms (Fig. 22). The difference appears to be thatthe RR minigenome is capable of generating a monomer replicative form withextended duplex palindromes at each end. The extended right palindrome may begenerated either by a hairpin transfer reaction or by a resolution of the 5’ bridgedimer fragment. The latter reaction has been demonstrated in vivo and in vitro.(Cotmore and Tattersall, 1992; Cotmore et a!., 1992). A closed hairpin at the rightterminus is thought to arise through hairpin formation followed by DNAreplication using the 3’OH as the primer. Clearly the presence of a mRR3 whichcontains two covalently closed hairpin termini (Fig. 21B) cannot be explained usingthe model of MVM DNA replication in Fig. 4. It may be possible that mRR3 arisesthrough recombination of two mRR2 type molecules. Alternatively, a ligation eventat the right terminus of mRR2 type molecules may also generate the mRR3. This105has been previously postulated to explain the existence of mLR4 type molecules,which also have covalently closed right and left termini (Cotmore et al., 1989). Inany case, if mRR3 is a legitimate replicative intermediate, the covalently closed endsmust be resolved in a manner similar to the resolution of AAV covalently closedhairpins. Curiously, the generation of mLLXS-2 type molecules (see Fig 23)presents a problem to the replication of LL type molecules. It is thought thatcovalently closed left hairpins cannot be resolved by a hairpin transfer typemechanism since the flip sequence orientation must be maintained at the leftterminus. If both ends are covalently closed, then this molecule may not participatein further amplification of the LL minigenome since both ends are “locked” into theclosed hairpin form. Thus amplification of LL type minigenomes may bedependent on the amplification of mLLXS-1 type molecules.3.8. Identification of Binding Activities in Nuclear Extracts of A9 and MVM(p)Infected A9 CellsSince it was determined that the SX fragment, as well as the RsaA and RsaBfragments are able to specifically rescue DNA replication of the dl4ll-4695 mutantMVM(p) minigenome, it is plausible that elements A (nt 4489-4636) and B (nt 4636-4695), as defined by deletion analysis, may constitute elements of a viral origin ofDNA replication. Previous studies have shown that sequence specific DNAbinding proteins such as transcription factors are able to activate DNA replicationin a variety of viral systems as well as yeast. In order to determine if proteins bindto elements A and B, nuclear extracts were prepared from uninfected and MVM(p)infected LA9 cells and probed with end labeled fragments which contain Rsa A orRsa B in electrophoretic mobility shift assays (EMSA) (Fried and Crothers, 1981). Inthis assay, 32P labeled restriction fragments were mixed with proteins from a106nuclear extract. DNA-protein complexes were separated from the unbound 32Plabeled probe by electrophoresis through a non-denaturing 4% polyacrylamide gel.At least six DNA-protein complexes were seen when Rsa A was used as a probe(Fig. 24A, lane 2). In order to determine specificity of these complexes, identicalbinding reactions were carried out in the presence of a 200 fold excess by weight ofeither unlabeled Rsa A, RsaB or 70 bp RsaI fragment of unrelated sequence frompUC19 designated Rsa7O. The slowest migrating complex was likely to be a nonspecific DNA-protein complex (NS) since it is completely removed upon addition ofany of the three unlabeled fragments. In contrast, five other DNA-proteincomplexes designated MVM(p) DNA Replication Factors (MRF) A2, A3, A4, A5 andA6 demonstrated specific competition. The majority of MRF A2 was competed offwith excess of unlabeled Rsa A and Rsa B probe whereas MRF A3, A4 and A5 werecompletely removed only in the presence of unlabeled Rsa A. Addition of Rsa B orRsa 70 did reduce the intensity of MRF A3, A4 and A5, but failed to completelyremove these complexes. These data suggest that these complexes are likely to bespecific for the Rsa A probe, but the protein involved may also have some nonspecific interactions with Rsa B and Rsa 70. MRF A2 appears to be specific for Rsa Band Rsa A but not Rsa 70. Since identical results were seen when nuclear extractsfrom MVM(p) infected LA9 cells were used, it suggests that factors binding toelements within the Rsa A probe are cellular in origin. When 32P labeled Rsa B wasused in EMSA experiments, four complexes were observed and were designatedMRF B2, B3, B4 and B5 (Fig. 24B). Again the majority of MRF B2 was competed offin the presence of an excess of unlabeled Rsa A or Rsa B but not Rsa 70. Thisobservation suggests the protein factor involved in MRFA2 and MRF B2 may be thesame as both bind specifically to107Fig. 24. Cellular Proteins in Nuclear Extracts Bind to Rsa A and Rsa B.Radiolabeled Rsa A (A) and Rsa B (B) were incubated in the presence of 5 jig ofeither uninfected A9 nuclear extracts (lanes 2-5) or MVM(p) infected LA9 nuclearextracts (lanes 7-10). Addition of a 200 fold excess of either unlabeled Rsa Afragment (lanes 3 and 8), Rsa B fragment (lanes 4 and 9) or the non-specific Rsa 70fragment (lanes 5 and 10) in competition binding reactions determined specificity ofbinding. Lanes 2 and 7 contained no unlabeled restriction DNA fragmentcompetitor. Lanes 1 and 6 contained no protein added to the binding reaction.Specific protein-DNA complexes were designated MVM(p) replication factors(MRFs). Non-specific (NS) complexes were as indicated.Aa,(3zz0xa,ci,0zMVM infected NEIdUninfected NE‘;?0Ifr’U ed‘I4— N S“ MRFA6MRFA5MRF A4MRFA3r1.I4— MRFA24— Free ProbeRsaA1 2 3 4 5 6 7 8 9 10MVM infected NE.