Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identification and characterization of small rnas and proteins expressed by the human parvovirus B19 St. Amand, Janet Lynn 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_st_amand_janet_lynn.pdf [ 4.29MB ]
JSON: 831-1.0086585.json
JSON-LD: 831-1.0086585-ld.json
RDF/XML (Pretty): 831-1.0086585-rdf.xml
RDF/JSON: 831-1.0086585-rdf.json
Turtle: 831-1.0086585-turtle.txt
N-Triples: 831-1.0086585-rdf-ntriples.txt
Original Record: 831-1.0086585-source.json
Full Text

Full Text

IDENTIFICATION AND CHARACTERIZATION OF SMALL RNAS ANDPROTEINS EXPRESSED BY THE HUMAN PARVOVIRUS B19byJANET LYNN ST. AMANDB.Sc. (Hon), Simon Fraser University, 1986A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDepartment of Biochemistry(Genetics Program)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril, 1992© Janet L. St. Amand, 1992In 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.Department of________________The University of British ColumbiaVancouver, CanadaDate f* ‘, 1DE-6 (2/88)11ABSTRACTThe human pathogenic parvovirus B19 has a strict tissue tropism andwill only replicate in a subset of erythroid progenitor cells. However, it isshown that COS-7 cells transfected with SV4O-B19 hybrid vectors express themajor B19 RNAs and proteins. In addition, capsid proteins synthesized inthese cells self-assemble into virus particles that by EM are morphologicallyvery similar to native B19 virions. Cytoplasmic RNA from transfected COS-7cells was used to prepare a cDNA library using a method which enriched thelibrary for B19 cDNAs. A second cDNA library was prepared from B19 infectedhuman bone marrow cells that had been isolated from a patient with achronic myelogenous leukemia using PCR to amplify B19-specific cDNAs.The libraries were probed with a series of RNA probes derived fromdifferent regions of the B19 genome in pYT1O3 and selected cDNAs weresequenced and compared. Two size classes of small, abundant, polyadenylatedRNAs were identified; the 700 and 800 nt size class of RNA is the product oftranscriptional processing in the middle of the B19 genome downstream froman unusual polyadenylation signal, ATTAAA or AATAAC (map unit 49).These transcripts contain an ORF within their second exon which is the samereading frame used for the translation of the nonstructural proteins;however, there are no AUG translation initiation codons within this regionof the ORF. The second most abundant size class of RNA in B19 infected cellsare the 500 and 600 nt transcripts which are made from three exons andterminated at a normal polyadenylation signal (AATAAA) near the righthand end of the genome (map unit 97). A 94 aa ORF within the third exon isinvariant in the two RNA species.111Antisera were raised in rabbits immunized with synthetic peptideswhose amino acid sequences were derived from hydrophilic regions withinthese ORFs. The antisera were used in B19 expression studies using a rabbitreticulocyte lysate to translate SP6 RNA polymerase-generated transcripts invitro, and in COS-7 cells transfected with SV4O on plasmids utilizing ahuman cytomegalovirus promoter (pCMV5) or the B19 P6 promoter(pSVOd/BW). The antisera were also used to detect small proteins in B19infected human leukemic cells. It was shown that the 500 and 600 nt mRNAsdirect the synthesis of at least two Ii. kDa proteins. Site-specific mutagenesiswas used to show that more than one protein is translated from the sametranscript by a leaky ribosomal scanning mechanism. Using indirectimmunofluorescence the H kDa proteins were localized to the cytoplasm oftransfected COS-7 cells; however their localization in B19 infected humanerythroid cells was at least partially nuclear. A stop codon introduced into the94 aa ORF prevented synthesis of the 11 kDa proteins but did not affect theexpression of the major structural and nonstructural proteins in transfectedCOS-7 cells.Using the same analysis it was shown that the most abundant size classof RNA, the 700 and 800 nt transcripts, do not appear to direct the synthesis ofa protein from the NS reading frame. Two different anti-peptide serarecognized the nonstructural proteins expressed in transfected COS-7 cells andin B19 infected leukemic cells but failed to detect a potential 15 kDapolypeptide.ivTABLE OF CONTENTSPageAbstract iiTable of Contents ivList of Figures xiAcknowledgements xivAbbreviations XVI INTRODUCTION 11.1 Historical Perspective I1.2. Review of Parvoviruses 21.2.1. General Characteristics 21.2.2. Classification of Parvoviruses 31.2.3. Virion Structure 41.2.4. Genome Organization 61.2.5. Parvovirus Proteins 71.2.6. Parvovirus Life Cycle 91.2.7. Oncosuppression 111.3. B19 Pathology 121.4. B19 Tropism 131.5. B19 Gene Expression 131.6. Small Viral Regulatory Proteins 161.7. The COS Cell Expression System 191.8. The Present Study 20VII MATERIALS AND METHODS 222.1. Materials 222.2. Strains and Media 232.2.1. Plasmids 232.2.2. Bacteria 262.2.3. COS-7 Cells 272.2.4. Human Bone Marrow Cells 272.2.5. B19 Virus 282.3. Basic Molecular Cloning Techniques 282.3.1 Isolation of DNA Fragments from Agarose Gels 282.3.2. Cloning DNA Fragments into Plasmid Vectors 282.3.3. Preparation and Transformation of Competent Cells 292.4. Purification of Plasnild DNA 302.4.1. Small Scale Plasmid DNA Isolation 302.4.2. Large Scale Plasmid DNA Isolation 302.5. Isolation of Single-Stranded DNA from Phagemids 302.6. Double-Stranded DNA Sequencing 312.7. Preparation of Labeled Hybridization Probes 322.7.1. DNA Probes 322.7.2. RNA Probes 322.8. DEAE-Dextran Mediated Transfection of DNA into COS-7 Cells 332.9. B19 Infection of Human Leukemic Bone Marrow Cells 332.10. Isolation of Low Molecular Weight DNA from Cultured Cells 342.11. Southern Blotting 342.12. Isolation of RNA from Cultured Cells 35vi2.13. Northern Blotting 352.13.1. RNA Gels 352.13.2. RNA Transfer 362.14. Hybridization of Filters 362.15. cDNA Libraries 372.15.1. Construction of B19 Human Bone Marrow CellcDNA Library 372.15.2. Construction of B19 COS Cell cDNA Library 392.15.3. Screening the cDNA Libraries 402.16. Peptides and Antisera 402.17. In Vitro Translation, Immunoprecipitation, and WesternBlotting 412.17.1. In Vitro Translation of SP6 Generated RNAs 412.17.2. Lysis of Mammalian Cells in Sample Buffer 422.17.3. Western Blotting 432.17.4. Immunoprecipitation of Radiolabeled Proteins 432.18. Metabolic Radiolabeling of Proteins Expressed in COS-7 Cells 432.19. Indirect Immunofluorescence 442.19.1. Transfected COS Cells 442.19.2. B19 Infected Bone Marrow Cells 462.20. Site-Specific Mutagenesis 462.20.1. Mutation of the Second ATG in the 94 aa ORF 462.20.2. Mutation of the Third ATG in the 94 aa ORF 482.20.3. Mutation of the First ATG in the 94 aa ORF 502.20.4. Creating a Stop Codon in the 94 aa ORF 502.20.5. Mutation of an ACG to an ATG Codon in the687 nt cDNA 51vii2.21. Isolation of B19 Particles from Transfected COS Cells 512.22. Electron Microscopy of B19 Capsids 522.22.1. Direct EM 522.22.2. Immune EM 522.23. Expression of 11 kDa Proteins Fused to the Yeast GAL-4 DNABinding Domain 52III RESULTS 543.1. Expression of the Major B19 Proteins in COS-7 Cells 543.2. Replication of B19 in Human Leukemic Cells 543.3. Screening the cDNA Libraries 583.3.1. Identification and Sequence of B19 Splice Junctions incDNAs from Transfected COS Cells 583.3.2. Identification and Sequence of B19 Splice Junctions incDNAs from Infected Human Leukemic Bone MarrowCells 623.4. Northern Hybridization ArLalysis of RNA from TransfectedCOS Cells 633.5. Expression of the 518 and 638 nt RNAs 653.5.1. In Vitro Expression of 518 and 638 nt RNAs in a RabbitReticulocyte Lysate 653.5.2. COS Cell Expression of 11 kDa Proteins in pSVOd/M70and pCMV/518 Transfected Cells 693.5.3. Expression of 11 kDa Proteins in B1.9 Infected HumanBone Marrow Cells 723.6. Characterization of Multiple Forms of 11 kDa Protein 723.6.1. Phosphatase Treatment of the 11. kDa Proteins 72viii3.6.2. Mutagenesis of Translational Initiation ATG Codons atthe 5’ End of the 94 aa ORF 753.6.2.1. Expression of the Second ATG Codon Mutant 753.6.2.2. Expression of the Second and Third ATG CodonMutant 753.6.2.3. Expression of the First ATG Codon Mutant 783.7. Expression of B19 Structural and Nonstructural Proteins inthe Absence of the 11 kDa Proteins 783.8. Expression of the 687 and 807 nt RNAs 813.8.1. In Vitro Translation of 687 and 807 nt RNA in a RabbitReticulocyte Lysate 813.8.2. COS Cell Expression of a Putative 15 kDa Protein inpSVOd/A170 and pCMV/687 Transfected Cells 813.8.3. Expression of Putative 15 kDa Protein in B19 InfectedHuman Bone Marrow Cells 853.9. Expression of 687 nt cDNA Containing an ACG to ATGMutation 853.10. Immunofluorescence 873.10.1. Localization of 11 kDa Proteins in Transfected COSCells 873.10.2. Localization of NS Proteins in Transfected COS Cells 873.10.3. Localization of 11 kDa Proteins in B19 InfectedHuman Leukemic Cells 903.10.4. Localization of NS Proteins in B19 Infected HumanLeukemic Cells 923.10.5. Immunofluorescence of Putative 15 kDa Protein 92ix3.11. Identification of B19 Particles in Transfected COS Cells 943.11.1. Identification of Capsid Protein by Western Blotting 943.11.2. Visualization of B1.9 Particles after Negative Stainingby Transmission EM 953.12. Searching for Sequence Similarities 953.13. Expression of 11 kDa Protein Fused to the Yeast GAL-4 DNABinding Domain 98IV DISCUSSION 1014.1. B19 Gene Expression in Transfected COS-7 Cells and B19Infected Human Chronic Myelogenous Leukemia Cells 1014.2. Presence of the 700-800 nt and 500-600 nt cDNAs in theTwo cDNA Libraries 1034.3. Comparison of Splice Junctions in the cDNAs fromTransfected COS Cells and B19 Infected CML Cells 1044.4. Expression of 11 kDa Proteins from the 500 and 600 nt cDNAs 1064.5. Expression of the Potential Protein in the 700 and 800 ntcDNAs 1104.6. If the 700-800 nt Class of RNAs are not Translated, What OtherFunction(s) Might They Have? 1124.7. Localization of B19 Proteins in Transfected COS-7 Cells andInfected Human Erythroid Precursor Cells 1174.8. What Might the Function of the 11 kDa Proteins Be? 1184.9. Summary of Mechanisms of Gene Expression in B19 1224.10. Future Directions 123V LITERATURE CITED 125xVI APPENDIX 143A. Restriction Fragments of pYT1O3 Cloned into pGEM4Z forSynthesis of RNA Probes 143B. Sequences of Oligonucleotides 144C. Nucleotide Sequences of Small B19 cDNAs 145D. Sequences of Peptides Used to Generate Antisera in Rabbits 149xiLIST OF FIGURESPage1. Comparison of the transcriptional maps of threerepresentative mammalian parvovirus genomes 52. B19 transcriptional map 153. Western blot analysis of B19 proteins synthesized inCOS-7 cells transfected with SV4O-B19 hybrid vectors 554. In vitro translation of T7 RNA polymerase-generatedtranscripts from the B19 major left-hand ORF 565. Replication of B19 in human leukemic cells 576. Expression of B19 structural and nonstructural proteinsin human leukemic bone marrow cells 597. Splice sites identified in the 500 and 600 nt class of B19transcript in transfected COS-7 cells 618. Sequence of the splice junctions of the small RNAs inB19 infected human leukemic cells 649. Northern hybridization analysis of RNA from B19transfected COS-7 cells 6610. Western blot analysis of in vitro transcribed and translated500 and 600 nt cDNAs 6711. Immunoprecipitation of 11 kDa proteins translated invitro from 518 and 638 nt RNAs 6812. Expression of 11. kDa proteins in transfected COS-7 cells 7013. Immunoprecipitation of 11 kDa proteins expressed invivo in transfected COS-7 cells and in vitro in a rabbitreticulocyte lysate 71xii14. Expression of 11 kDa proteins in B19 infected humanleukemic bone marrow cells 7315. Digestion of the 11 kDa proteins with potato acid phosphatasedoes not affect their mobility by SDS-PAGE 7416. The sequence context of three ATG codons at the 5’ end ofthe 94 aa ORF in the third exon of the 518 and 638 nt cDNAscompared with the Kozak consensus sequence for translationalinitiation and the 5’ end sequences of three ATG to CTGmutants 7617. COS-7 cell expression of ATG to CTG mutants 7718. Immunoprecipitation of proteins translated in vitro fromwild type and mutant 518 nt RNAs 7919. Summary of the 11 kDa proteins expressed in COS-7 cells aftertransfection with wild type and mutant pCMV/518 DNA 8020. The 11 kDa proteins are not made in COS-7 cells transfectedwith pSVOd/A170/11 kDa DNA 8221. COS-7 cell expression of B19 structural and nonstructuralproteins in the absence of the ii kDa proteins 8322. Expression of wild type and ACG to AUG mutant 687 ntRNAs in a cell free system 8423. Immunoprecipitation failed to detect the putative 15 kDaprotein 8624. Localization of the 11 kDa proteins to the cytoplasm oftransfected COS-7 cells by indirect immunofluorescence 8825. Localization of the NS proteins to the cytoplasm oftransfected COS-7 cells by indirect immunofluorescence 89xlii26. Localization of 11 kDa proteins in B19 infected humanleukemic cells 9127. Localization of NS proteins in B19 infected humanleukemic cells 9328. Direct electron microscopy of B19 particles made intransfected COS-7 cells 9629. Immune electron microscopy of B19 parvovirus particlesaggregated with a B19 convalescent serum 9730. Amino acid sequence of the 11 kDa proteins 9931. The 11 kDa protein lacks an activation domain 10032. Predicted secondary structures of 687 and 807 nt RNAs 114xivACKNOWLEDGEMENTSI would like to thank the many people who have contributed to thisthesis. First of all I thank my supervisor, Dr. Caroline Astell, for giving methe opportunity to work in her lab and for introducing me to virology. Iwould also like to thank the members of my supervisory committee; Dr.Shirley Gillam, Dr. Rob McMaster, and Dr. Peter Candido for their helpfuladvice and suggestions in these studies. I thank the Medical Research Councilof Canada whose funding made this work possible. Thanks to all members ofthe Astell lab past and present for the many discussions and criticisms of thiswork. I thank Caroline Beard for initiating the B19 expression studies in COScells and for helping me start this project. In addition, I thank Dr. BernardCohen for providing B19 virus and convalescent serum, Dr. KeithHumphries for supplying the CML bone marrow cells, Dr. Ian Clark-Lewis forsynthesizing peptides used to generate B19 antisera, and Drs. Sue Cotmoreand Peter Tattersall for providing B19 antibodies used in these studies. I thankMichael Weiss and Susan Wielesko for their aid and instruction on the use ofthe fluorescence microscope and the electron microscope. Thanks to DonaldMinato for his collaboration on particle isolation and EM work and to CohnHarris for testing the 11 kDa protein for transactivation activity. Finally, Iwould like to thank Patrick Tam for all our often heated but alwaysinteresting discussions about science, politics, and life.xvABBREVIATIONSaa amino acid(s)AAV adeno-associated virusAb antibodyADV aleutian disease virusamp ampicillinAR autoradiographyATP adenosine 5’-triphosphateBCIP 5-bromo-4-chloro-3-indoylphosphateBFU-E burst-forming unit-erythroidBMV brome mosaic virusbp base pair(s)BPV bovine parvovirusBSA bovine serum albuminCAT chioramphenicol acetyl transferasecDNA complementary DNACFU-E colony-forming unit-erythroidClAP calf intestinal alkaline phosphataseCIE counter-immunoelectrophoresisCML chronic myelogenous leukemiaCMV cytomegaloviruscpm counts per minuteCPV canine parvovirusCTP cytidine 5’-triphosphateCsC1 cesium chloridecys cysteinexvidAT1’ deoxyadenosine 5’-triphosphatedCTP deoxycytidine 5’-triphosphateDAT dsRNA-activated inhibitor of protein synthesisDEAE diethylaminoethylDEP diethylpyrocarbonatedGTP deoxyguanosine 5’-triphosphateDMEM Dulbecco’s Modified Eagle MediumDMSO dimethyl sulfoxideDNA deoxyribonucleic acidDNase deoxyribonucleasedNTP deoxynucleotide triphosphate mixDOC sodium deoxycholateDTT dithiothreitoldTTP deoxythymidine5t-triphosphateE. coli Escherichia coliEDTA ethylenediamine tetraacetic acidEM electron microscopyepo erythropoietinEtOH ethanolFCS fetal calf serumFPV feline parvovirusGST glutathione S-transferaseGTP guanosine 5’-triphosphateHBsAg hepatitis B surface antigenHBS HEPES buffered salineHEPES N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acidhis histidinexviiHIV-1 human immunodeficiency virus 1h.p.i. hours post infectionIg immunoglobulinJMDM Iscove’s Modified Dulbecco’s MediumIPTG isopropylthiogalactosidekb kilobase(s)kDa kilodalton(s)KLH keyhole limpet hemocyaninLTR long terminal repeatMAb monoclonal antibodyMCS multiple cloning sitemet methionineM-MLV Moloney murine leukemia virusMOPS morpholinopropane sulfonic acidmRNA messenger RNAm.u. map unitMVM minute virus of miceM W molecular weightNBT nitroblue tetrazolium chlorideNC nitrocelluloseNH2 aminonm nanometerNP-40 Nonidet P-40NS nonstructuralnt nucleotide(s)NTP nucleotide 5’-triphosphateOD optical densityxviiioligo oligonucleotideORF open reading frameon origin of replicationPAGE polyacrylamide gel electrophoresisPAP potato acid phosphatasePBS phosphate buffered salinePCR polymerase chain reactionPEG polyethylene glycolPMSF phenylmethylsulfonyl fluoridepoly(A) polyadenylationProK Proteinase KPTA phosphotungstic acidPVDF polyvinyldifluorideRF replicative formRI refractive indexRIA radioimmunoassayRNA ribonucleic acidRNase ribonucleaseRPHA reverse passive hemagglutinationrpm revolutions per minuteRRE rev-responsive elementRT room temperatureR V rat virus-likeSDS sodium dodecyl sulphateSPLV serum parvovirus-like virusSV4O simian virus 40TAR trans-activation response elementxixTBE Tris-borate EDTATCA trichioroacetic acidTE 10 mM Tris, 1 mM EDTATEMED N,N,N’,N’-tetramethylethylene diaminetet tetracyclineTris tris (hydroxymethyl)aminomethanetRNA transfer ribonucleic acidU unit(s)UTP uridine5t-triphosphateUV ultravioletVA virus-associatedvol volumeVP viral proteinVRC vanadyl-ribonucleoside complexesX-gal 5-bromo 4-chloro-3-indolyl-i-D-galactosidex g times gravity1INTRODUCTIONSmall proteins and RNAs are important in the life cycle of manyanimal viruses where they play vital roles in viral gene expression such astranscriptional activation of promoters and control of nuclear export orsplicing of viral transcripts. Viral proteins and RNAs are also involved inhost cell subversion where host cell protein synthesis is shut down and viralmRNAs are preferentially translated. Regulation of gene expression in thehuman pathogenic parvovirus B19 is not yet fully understood. This virus ispeculiar among parvoviruses since it has a strict tropism for erythroidprogenitor cells and cannot be propagated in a continuous cell line. Inaddition to transcripts that encode the major nonstructural and structuralproteins similar to other parvoviruses, B19 directs the synthesis of two sizeclasses of small, abundant transcripts, one of which is translated into a familyof 11 kDa proteins. The role of these RNAs and proteins in the B19 life cyclemay be important in our understanding of viral cell tropism andpathogenesis. Also, development of parvoviruses as vectors for human genetherapy is currently being investigated. Since the B19 promoter has highactivity in many cell types, and at least one parvovirus has been shown tointegrate site-specifically, B19 may become the vector of choice for genetherapy and therefore it is important to understand fully, gene expression inthis virus.1.1. Historical PerspectiveB19 was discovered serendipitously during routine testing of blood serafor hepatitis B surface antigen (HBsAg) (Cossart et at., 1975). At that time thesecond generation tests of reverse passive hemagglutination (RPHA) and2radioimmunoassay (RIA) for HBsAg were replacing the first generation testsof gel diffusion, counter-immunoelectrophoresis (CIE), and complementfixation. In a comparison between electrophoresis and a commercial RPHAtest, RPHA increased the yield of positive results by one third. However, threesera were positive by CIE but negative by both RPHA and RIA. One of thesesera which reacted only in electrophoresis was number 19. in panel . Itbecame apparent that this serum contained antigen which was distinct fromHB5Ag and by electron microscopy (EM) was shown to contain viral particlesof approximately 23 nm in diameter. In addition to full virus particles theserum contained disrupted fragments and empty shells which are consideredto be characteristic of parvoviruses. The reactivity of the newly discoveredvirus by electrophoresis was due to the fact that human sera were used, someof which contained antibodies to the serum parvovirus-like virus (SPLV),whereas in the RPHA and RIA tests hyperimmune sera from sheep, horse,and guinea pig were used which were monovalent for HB5Ag (Pattison, 1988).Subsequent serological testing suggested that antibodies against SPLV werecommon in the population but at that time it was not associated with aspecific clinical disease. In the next few years this “new” human virus wasdefinitively classified as a parvovirus (Summers et a!., 1983; Cotmore andTattersall, 1984). The name B19 was chosen on the recommendation of theStudy Group on Parvoviridae of the International Committee on Taxonomyof Viruses (ICTV) (Siegl et a!., 1985).1.2. Review of Parvoviruses1.2.1. General CharacteristicsParvoviruses are small, nonenveloped, DNA viruses which infect avariety of animal species from insects to man (for reviews see Ward and3Tattersall, 1978; Berns, 1984; Pattison, 1988; Tijssen, 1990). The virions containa molecule of single-stranded DNA of 4.7 to 6.0 kb in length (Siegi, 1984)enclosed within an icosahedral coat composed of two or three proteins withoverlapping amino acid sequences. A characteristic feature of parvoviruses isthe presence of palindromic sequences at either end of the genome resultingin the formation of stable duplex structures (Bourguignon et a!., 1976). Thehairpins provide a primer for DNA synthesis by host cell DNA polymeraseand serve to maintain the integrity of the terminal sequences.1.2.2. Classification of ParvovirusesThe parvoviruses have been historically divided into three genera:Parvovirus, Dependovirus, and Densovirus. This classification system isbased on the requirement for a helper virus for viral DNA replication, thenature of the termini and the sense of the DNA strand that is packaged intovirions, and the host species which is infected. The autonomousparvoviruses are vertebrate viruses which are helper independent in the cellsof their normal host species. They usually have unique termini and packagepredominantly minus sense DNA strands into virions. The dependovirusesalso infect vertebrates but are dependent on adenovirus or herpes viruscoinfection for lytic infection. The dependovirus genome is terminallyredundant and equal numbers of both plus and minus strands are packagedinto separate virions. The densoviruses, parvoviruses which infect insects,have not been studied to the same extent as have the vertebrate parvoviruses.The densoviruses which have been characterized replicate autonomously,contain identical inverted terminal repeats, and package both plus and minusDNA strands (Tijssen et a!., 1990).4Evidence that this classification scheme is in need of revision isexemplified by B19 which is autonomously replicating yet contains identicalterminal repeat sequences and packages DNA strands of both senses. Inaddition, the requirement of the dependoviruses for a helper function hasbeen challenged since the adeno-associated virus, AAV5, replicatesindependently, albeit at a reduced level, in mutagen-treated cells (Yakinogluet a!., 1988) and in chemically synchronized cells (Yakobson et a!., 1987) It hasalso been shown in Luill, a mammalian autonomous parvovirus, thatidentical ends are not required for the encapsidation of equal amounts of plusand minus strand DNA (Diffoot et a!., 1989). A newer scheme may classifyparvoviruses on the basis of the number and genomic position of theirtranscriptional promoters, splice sites, and polyadenylation signals (Figure 1).1.2.3. Virion StructureThe parvovirus virion is an icosahedron of 20 to 25 nm in diameterwith a molecular weight of 5.5 to 6.2 X 106 Da. The three-dimentional atomicstructure of canine parvovirus (CPV) has recently been determined to aresolution of 3.25 A (Tsao et a!., 1991). In CPV, full capsids contain sixty copiesof a combination of the coat proteins VP-2 and VP-3 and some VP-i and areicosahedral with a T 1 symmetry according to the nomenclature of Casparand Kiug (Caspar and Klug, 1962). Parvovirus particles do not appear tocontain lipids, carbohydrates, cellular encoded enzymes, or low molecularweight histone-type proteins (Tattersall and Cotmore, 1988). Chargeneutralization of the packaged DNA may result from interaction of a highlybasic region in the amino terminal of the largest structural polypeptide(Tattersall, 1978a). Recent reports suggest that in minute virus of mice (MVM)the DNA is covalently associated at its 5$ end with a virally encoded protein5B19 0 100I I I I I I Im.u.TATAAATAAC AATAAAA1rAAAAAV2 100I I I I I I I I I I Im.u.TATAAATAAAMVMp 0I I I I I I I I I I IA m.u.-TATAAATAAAFigure 1. Comparison of the transcriptional maps of three representativemammalian parvovirus genomes.The thick lines represent RNA transcripts. Introns are shown byinterruptions in the solid lines. The map positions of the functional TATAsequences and polyadenylation signals are indicated by vertical lines at thebottom of the figure.6which is antigenically related to NS-1 (Cotmore and Tattersall, 1989; Faust eta!., 1989).1.2.4. Genome OrganizationIn the vertebrate parvoviruses studied thus far all of the protein codingsequences are confined to a single DNA strand which by definition is positivesense. However, at least some densonucleosis viruses encode proteins onboth the viral and complementary DNA strands (Giraud and Bergoin, 1991).The parvovirus genome contains two large blocks of open reading frame(ORF) which together span almost the entire genome and a number ofsmaller ORFs which vary in size and location between the different viruses(Shade et a!., 1986). In general, the large left-hand ORF encodes the majornonstructural (NS) proteins and the right-hand ORF encodes the structural(VP) proteins. All parvovirus genomes have a promoter which maps to m.u.4 to 6 (the genome consists of 100 map units) and, with the exception of B19,an additional pomoter at m.u. 38 to 40 which initiates the capsid transcripts(Green and Roeder, 1980; Pintel et a!., 1983; Lebovitz and Roeder, 1986). AAVcontains a third promoter at m.u. 19 (Lusby and Berns, 1982) (Figure 1). Inparvoviruses with more than one promoter there is a temporal regulation ofgene expression with an early nonstructural gene product regulating a latepromoter encoding the capsid proteins in MVM and H-i (Rhode, 1985; Rhodeand Richard, 1987; Clemens and Pintel, 1988; Doerig et a!., 1988; Tullis et a!.,1988; Doerig et a!., 1990) or both the capsid and its own promoter in AAV(Labow et a!., 1986; Tratschin et a!., 1986). All RNA transcripts arepolyadenylated and in MVM, H-i, CPV, and feline parvovirus (FPV) coterminate at the extreme right-hand end of the genome. Bi9 and aleutiandisease virus (ADV) produce transcripts which terminate in the middle of the7genome in addition to RNAs which terminate at the right side of the genome(Ozawa et al., 1987; Alexandersen et a!., 1988). In MVM a 142-147 nt P4transcript has been identified which may arise from prematuretranscriptional termination (Ben-Asher and Aloni, 1984; Resnekov andAloni, 1989). Readthrough or termination of transcription may depend onsecondary structure in a manner which is analogous to attenuation inprokaryotes (Yanofsky, 1981) and it is speculated that a viral protein may beinvolved in this process (Krauskopf et a!., 1991). The coding capacity of thelimited parvovirus genome is increased by the extensive use of alternativesplicing to produce more than one protein from the same transcription unit.1.2.5. Parvovirus ProteinsParvoviruses encode two to four nonstructural proteins. The largenonstructural protein, NS-1 in MVM and H-I, and rep in AAV, has beenshown to be required both for viral DNA replication and transactivation ofthe capsid promoter (Rhode, 1985; Labow et a!., 1986; Tratschin et a!., 1986;Rhode and Richard, 1987; Clemens and ,Pintel, 1988; Doerig et a!., 1988; Tulliset a!., 1988; Doerig et al., 1990). Biochemical studies have shown that rep fromAAV, which is a nonphosphorylated nuclear protein, has ATPase, helicase,and site specific endonuclease activities (Tm and Muzyczka, 1990). Thecorresponding NS-1 from MVM is a nuclear phosphoprotein with the sameactivities (Cotmore and Tattersall, 1986; Wilson et a!., 1991). Computersequence analysis has shown that there is significant homology between thenonstructural proteins of the vertebrate parvoviruses suggesting a commonfunction for these proteins (Astell et a!., 1987). However, since B19 contains asingle promoter (Blundell et a!., 1987; Doerig et a!., 1987; Ozawa et a!., 1987;Liu et a!., 1991) B19 NS proteins probably lack a transactivation activity. The8function of a second, smaller nonstructural protein, NS-2, is less clear. InMVM, NS-2 shares 84 amino terminal amino acids with NS-1 but differs at itscarboxyl end due to a splicing event. Additional alternative splicing producesthree different NS-2 proteins whose functional significance is unknown(Jongeneel et a!., 1986; Morgan and Ward, 1986; Cotmore and Tattersall, 1990).Attempts to construct stable cell lines which constitutively express the majornonstructural protein of MVM and H-1 have routinely failed, providingindirect evidence that the protein is cytotoxic in transfected cells (Rhode, 1987;Astell and St. Amand, unpublished results). A similar observation was maderegarding the NS proteins of B19 (Ozawa et a!., 1988a) where plasmidscontaining the B19 genome were only stably transfected into HeLa cells aftermutation of the left-hand ORF prevented synthesis of the NS proteins. Theseresults allow speculation that the major parvoviral nonstructural proteinmay be the causative agent in host cell lysis during infection.In vertebrate parvoviruses at least two viral capsid proteins, VP-i andVP-2, are encoded by the major right hand ORE with the sequence of VP-2entirely contained within that of VP-I. In the rat virus-like (RV)parvoviruses, such as MVM, after the DNA is packaged into virions a thirdcapsid protein, VP-3, is produced by proteolytic cleavage of 15 to 20 aminoacids from the amino terminus of VP-2 (Tattersall et a!., 1976). In CPV theinfectious particle contains sixty protein subunits which are predominantlyVP-2 although both VP-i and VP-3 are required for virus infectivity (Tsao eta!., 1991). It has been shown for B19 that VP-2 alone can self-assemble intoparticles which are morphologically identical to native B19 particles byelectron microscopy (EM) (Brown et al., 199ia; Kajigaya et a!., 1991). However,VP-i is required to produce neutralizing antibodies from these particlessuggesting that at least part of the unique VP-i sequence is on the outside of9the virion. There is no evidence for a third capsid protein in B19. In AAV,VP-2 is initiated at an ACG codon on the same transcript which produces VP-3 from an AUG triplet (Becerra et al., 1985). VP-3, which represents 86% of thetotal virion protein mass in AAV, is analogous to VP-2 in the RV and B19viruses.In addition to the large ORFs which encode the major NS andstructural proteins parvoviruses contain small ORFs which vary in size andlocation among the members of the family. The coding potential of thesesmall OREs has not yet been explored.1.2.6. Parvovirus Life CycleProductive infection is initiated by adsorption of the virion to cell-surface receptors. Internalization is thought to take place through coated pits(Linser et a!., 1977) and it is not clear if the particle is uncoated in thecytoplasm or in the nucleus. These initial steps of viral attachment, entry, andaccumulation in the nucleus can proceed in all cells which have a functionalparvovirus receptor regardless of their position in the cell cycle (Rhode, 1973;Siegi and Gautschi, 1973). However, viral DNA replication and geneexpression can only occur during the S-phase of the cell cycle and is entirelydependent on host cellular factors expressed at that stage (Tennant et a!., 1969;Hampton, 1970; Tattersall, 1972; Rhode, 1973; Siegi and Gautschi, 1973).Unlike SV4O and adenovirus, parvoviruses cannot induce resting cells in G0to enter the mitotic cycle. In addition to the requirement for a dividing cellthe differentiation state of the cell is also important in viral replication(Mohanty and Bachmann, .1974; Miller et a!., 1977; Tattersall, 1978b). In theprototype (p) strain of MVM, proteins encoded by both halves of the genomeappear to be synthesized almost simultaneously within the first two hours of10S phase suggesting that both promoters are operational at this time. However,late in infection capsid transcripts predominate due to an increase ininitiation from the P38 promoter mediated by NS-1 (Schoborg and Pintel,1991).Viral DNA synthesis occurs predominantly through a self-primingmechanism employing the palindromic termini. During replication in thecell nucleus, the incoming single-stranded DNA is initially converted to adouble-stranded monomer replicative form (RF) and then to a dimer (andhigher order) RF employing host replication enzymes. These forms areresolved presumably by viral NS-1 through a series of site specific nicks andligations to produce progeny single-stranded DNA genomes for packaging.Replication is thought to occur by a rolling hairpin method involving ahairpin transfer to regenerate the 5’ ends of the replicated DNA strands(Tattersall and Ward, 1976). A consequence of this hairpin transfer is thegeneration of two different DNA sequences at the left-hand and right-handends of the genome, as has been demonstrated in AAV (Lusby et a!., 1982;Srivastava et a!., 1983). These sequences, which are the inverted complementof each other, result from imperfections in the terminal palindromes. Thetwo sequence orientations have been designated “flip” and “flop” and arefound in the same abundances in packaged virions. In the case of MVM andH-I, the right-hand end of the genome contains two sequence orientations(“flip” and “flop”) which are present in equal amounts (Rhode and Claussen,1982; Astell et a!., 1983; Astell et a!., 1985). However, the left-hand end ofMVM, H-i, H-3, and KRV contain a single unique sequence (“flip”) (Astell eta!., 1979; Astell et al., 1985). A modified rolling hairpin model has beenproposed to explain the absence of the “flop” sequence orientation at the lefthand end of MVM (Astell et al., 1985). Progeny virions are packaged within11the nucleus (Richards et al., 1977) and released from infected cells followingnuclear degeneration and rupture of the plasma membrane by mechanismsas yet undefined (Cotmore and Tattersall, 1987).The life cycle of the dependovirus, AAV, can be either latent or lytic. Inthe absence of helper virus (adeno- or herpesvirus) AAV integrates sitespecifically into the distal portion of the q arm of human chromomosome 19(q14.3-ter) (Kotin et al., 1990; Berns et a!., 1991; Nahreini and Srivastava, 1991;Samuiski et a!., 1991). The physical structure of integrated AAV genomessuggests that the integrated viral DNA exists as a tandem repeat of multiplecopies in a head-to-tail arrangement (Kotin and Berns, 1989). In tissue culturecells, the integrated DNA is stable for more than 50 rounds of replication butcan be rescued by superinfection with helper virus. The virus then undergoesa lytic cycle culminating in host cell lysis and release of progeny virions.Because the site of integration is localized to within 100 bp of chromosome 19,interest in AAV as a vector for human gene therapy has increased (Samulskiet a!., 1.991).1.2.7. OncosuppressionUnlike other nuclear DNA viruses parvoviruses have never beenimplicated in oncogenesis (Berns, 1990). On the contrary, parvoviruses appearto play a role in oncosuppression (Toolan, 1967; Toolan and Ledinko, 1968;Parker et a!., 1970; Campbell et a!., 1977; Mousset and Rommelaere, 1982;Toolan et aL, 1982; Cornelis et a!., 1984; Heilbronn et al., 1984; Cornelis et a!.,1986). Originally, parvoviruses were isolated from neoplastic tumor cellssuggesting that the parvovirus was the agent of tumorigenesis (Killam andOlivier, 1959; Toolan, 1961). However, it is now clear that this was anopportunistic association since parvoviruses replicate exclusively in rapidly12dividing cells. The intercellular milieu of the dedifferentiated, transformedcell may allow the parvovirus to replicate and subsequently lyse and kill theaffected cell; however, the precise mechanism of oncosuppression is notunderstood.1.3. B19 PathologyB19 has been shown to be the causative agent of erythema infectiosum(also known as fifth disease) (Pattison et a!., 1981; Serjeant et a!., 1981;Anderson et a!., 1983), a common childhood exanthem. In adults, B19infection often results in a polyarthralgia/arthritis syndrome (Reid et a!., 1985;White et a!., 1985). In the majority of cases the disease is mild and short-livedwith an excellent prognosis. However, in individuals suffering from chronichemolytic anemias such as homozygous sickle cell disease, hereditaryspherocytosis, pyruvate kinase deficiency, and 1-thalassemia, B19 infection isassociated with transient reticulocytopenic aplastic crisis (Pattison et a!., 1981;Duncan et at., 1983; Kelleher et at., 1983; Mortimer, 1983; Rao et a!., 1983). B19infection can also be persistent (Frickhofen and Young, 1989). In individualswith an immunodeficiency, either congenital or acquired, persistent B19parvovirus results in chronic pure red cell aplasia (Kurtzman et a!., 1987;Kurtzman et a!., 1988). In utero infection has been associated with a fewspontaneous abortions and hydrops fetalis with ensuing fetal death due tosevere anaemia and congestive heart failure (Anand et a!., 1987; Andersonand Hurwitz, 1988; Rodis et a!., 1988). B19 is a relatively new virus and thefull spectrum of its pathology has probably not been uncovered. As thesensitivity of diagnostic testing is increased a more complete picture of B19disease will likely emerge.131.4. B19 TropismThe B19 parvovirus is cytotoxic to human progenitor cells of theerythroid lineage (Mortimer et al., 1983; Young et a!., 1984). The target cell isthe burst-forming unit-erythroid (BFU-E) and the more mature colony-forming unit-erythroid (CFU-E) (Srivastava and Lu, 1988). Genetic studies ofB19 have been hampered by an inability to replicate the virus in a continuouscell line. B19 has been propagated in explanted human bone marrow (Ozawaet a!., 1986; Ozawa et a!., 1987; Takahashi et a!., 1990), fetal liver (Yaegashi eta!., 1989; Brown et a!., 1991b), erythroleukemia cells (Takahashi et a!., 1989),and cord blood (Sosa et a?., 1991; Zhou et a?., 1991). In studies presented herewe infected bone marrow tissue, isolated from a patient suffering from achronic myelogenous leukemia (CML), with B19 parvovirus. This marrowwas enriched in erythroid as well as myeloid progenitor cells (Eaves et a?.,1980).1.5. B19 Gene ExpressionThe sequence of an almost full length (5112 nt) clone of the B19genome was reported by Shade et a?., (1986). This clone contained the entirecoding region but had deletions in both the 5? and 3? hairpin ends.Subsequent sequencing of clones of intact ends has shown they containidentical inverted terminal repeats of 383 nt (Deiss et a!., 1990). This wouldresult in a full length B19 genome of 5596 nt. A single, functional promoter atm.u. 6 is active both in vitro and in vivo in heterologous cell types (Blundellet a?., 1987; Doerig et a?., 1987; Ozawa et a?., 1987; Liu et a!., 1991). Thetranscription start site was mapped to nucleotide 350 using reversetranscription of RNAs synthesized in vitro using a HeLa nuclear cell extract(Blundell et a?., 1987). At least nine overlapping B19 transcripts have been14identified in infected bone marrow cells by SI nuclease analysis using probesderived from different regions of the cloned B19 genome in the plasmidpYT1O3 (Ozawa et a!., 1987). From this data a transcription map has beenproduced which however does not precisely define the DNA sequence at thespliced junctions (Figure 2). All nine transcripts are polyadenylated and threeRNAs terminate in the middle of the genome and are polyadenylated using avariant polyadenylation signal which is either ATTAAA (nt 2639) orAATAAC (nt 2645). The other six transcripts co-terminate at the right-handend of the DNA which contains a consensus polyadenylation sequenceAATAAA (Proudfoot and Brownlee, 1976).The only unspliced transcript likely encodes the nonstructural proteinswith reported molecular weights of 77 kDa (major species), 52 kDa and 34 kDa(minor species) (Ozawa and Young, 1987) or 71 kDa, 63 kDa, and 52 kDa(Cotmore et a!., 1986). All other transcripts are spliced at least once andcontain a short untranslated leader sequence at their 5’ end. The two longestright-sided transcripts likely encode the minor capsid protein, VP-i, which is83 kDa and the other two long transcripts likely encode the 58 kDa majorcapsid protein, VP-2. The sequence of VP-2 is entirely contained within VP-i.In infected cells the amount of VP-I. produced is only 4 to 10% that of VP-2(Ozawa and Young, 1987). The difference in the abundances of the two capsidproteins can be explained by the differences in the levels of VP-i and VP-2transcripts. In addition, a leader sequence in the VP-i transcripts which isspliced out of the VP-2 transcripts contains multiple AUG triplets and hasbeen shown to down-regulate translation from these mRNAs in vitro (Ozawaet al., 1988b).Two additional size classes of polyadenylated RNAs are producedwhich by northern analysis are the most abundant transcripts in B19 infected15map units0 P6 20 40 60 80 100I I I I I I I IProtein2.3 NS-10.8 IIIIIIIIIIIIIIIIII..I— 90.73.15 VP-i3.0 VP-i2.3 VP -22.2 V P.2ii kDa11 kDa0 49 92 100I Ipolyadenylation A1TAAA AATAAAsignals AATAACFigure 2. B19 transcriptional map.The thick lines represent RNA transcripts and introns are shown byinterruptions in these lines. The known proteins encoded by the transcriptsare noted on the right. The positions of the promoter and polyadenylationsignals are indicated (modified from Ozawa et at., 1987).16human bone marrow cells (Ozawa et al., 1987). Two of these transcripts of 700and 800 nt which terminate in the middle of the genome share an ORF withthe NS gene however there are no AUG translation initiation codons withinthis ORF. The two remaining transcripts of 500 and 600 nt are made fromthree exons and contain a 94 amino acid ORF in their third exon derivedfrom the extreme right-hand end of the genome. In this study it is shown thatthese transcripts are bicistronic directing the synthesis of two 11 kDapolypeptides initiated from two different AUG triplets on the same mRNA.Since B19 contains a single promoter, gene expression cannot beregulated by differential promoter strength or transactivation of a secondpromoter by an early gene product such as occurs in MVM, H-i, and AAV.Instead, transcript abundances may be controlled by alternative 3’ endprocessing or alternative splicing events. It is not known whether a B19 viralfunction is involved in gene regulation but, if so, the abundant RNAs andsmall proteins would be prime candidates for this activity.1.6. Small Viral Regulatory ProteinsRecent studies have demonstrated many short ORFs within animalvirus genomes. Expression of these small OREs and in some cases elucidationof the function of the small proteins has been demonstrated. Small proteinsfrom HIV-1 have been shown to regulate gene expression temporally from asingle promoter. This complex retrovirus synthesizes at least six smallproteins: Tat, Rev, Nef, Vir, Vpr, and Vpu (Cullen and Greene, 1990). (Relatedproteins are expressed by other lentiviruses). Tat and Rev are essential forvirus replication while Vpu, Vif, Vpr, and Nef proteins serve accessoryfunctions that enhance replication and/or infectivity.17The 86 aa Tat protein transactivates HIV-1 LTR dependent geneexpression by approximately two orders of magnitude and is essential forreplication of HIV-1 in culture (Arya et al., 1985; Sodroski et al., 1985). Thiseffect is mediated by Tat binding to a stem-loop structure at the 5’ end of allHIV-1 mRNAs in the 5’ untranslated leader sequence termed the TAR (transactivating responsive) element. Tat binding was recently shown to involve asingle arginine side chain which contacts a bulged region of TAR and isessential for transactivation (Calnan et a!., 1991).The 19 kDa (116 aa) Rev protein is the second essential activatorprotein of HIV-1 and is conserved among members of the lentivirus family ofretroviruses. Late in infection, Rev, a phosphoprotein localized in thenucleolus of the cell, mediates the nucleocytoplasmic export of unsplicedstructural mRNAs that were otherwise sequestered in the nucleus (Felber eta!., 1989; Malim et a!., 1989). This activity is mediated by binding of themultiple Rev monomers to an RNA target sequence present in the envmRNA termed the RRE (Rev responsive element) (Malim and Cullen, 1991).The RRE forms a complex stem-loop structure (Heaphy et a!., 1990; Malim eta!., 1990) which binds Rev resulting in the expression of the viral structuralproteins including Gag and Env late in infection. The Rev protein may actindirectly by inhibiting the interaction of cellular splicing factors with theviral pre-mRNA (Chang and Sharp, 1989), or directly by facilitating theinteraction of the incompletely spliced transcripts with a component of thenuclear export pathway (Malim et a!., 1989).Nef is a 27 kDa myristylated plwsphoprotein associated withcytoplasmic membrane structures. The function of Nef is unclear althoughone report suggested that Nef had the properties of a G-protein (a family ofGTP-binding proteins involved in signal transduction) (Guy et a!., 1987; Guy18et al., 1991). The Nef ORF is conserved among primate lentiviruses suggestingan important role for this small protein in the viral life cycle (Cullen andGreene, 1990).The function of Vpr, an 11 kDa, 96 aa protein which like Nef is notrequired for replication of HIV-1 in culture, is unknown. The 23 kDa Vifprotein has been reported to be important in virion infectivity by anunknown mechanism (Fisher et al., 1987; Strebel et al., 1987). Vpu, an 81 aaprotein, is phosphorylated in vivo and is associated with the cytoplasmicmembranes of infected cells (Strebel et at., 1988). This small protein which isunique to HIV-1 is thought to be involved in virion release (Cullen andGreene, 1990).The sequence of the vaccinia virus genome revealed that it containstwenty ORFs that could potentially encode polypeptides of 9 to 13 kDa (Goebelet at., 1990). One of these small proteins, a basic 11 kDa phosphorylatedpolypeptide, has recently been shown to be involved in virion maturationand assembly (Zhang and Moss, 1991).The U511 gene of herpes simplex virus 1 (HSV-1) encodes an 11 kDabasic, site-specific, RNA-binding protein. This small polypeptide was shownto regulate the expression of a truncated form of the mRNA encoding anessential protein in the UL34 ORF. The U511 gene product binds to the 3’ endof the A34 transcript in a sequence and conformation dependent mannerpreventing the accumulation of this transcript in infected cells (Roller andRoizman, 1991) in a manner which is reminiscent of the binding of the Revprotein of HIV-1 to the RRE on the incompletely spliced nuclear RNAs.In the bovine coronavirus (BCV) the region between the spike (S) andmembrane (M) genes was shown to encode four small nonstructuralpolypeptides of 4.9, 4.8, 12.7, and 9.5 kDa (Abraham et at., 1990). Three mRNAs19that potentially express these proteins were identified in BCV infected cells.The functions of these small proteins are unknown but analogous smallproteins have been identified in other coronaviruses where some have beenproposed to have an anchoring function during virus assembly or maintain amembrane association of the viral polymerase during replication.A number of ORFs which potentially encode small proteins haverecently been identified in influenza viruses (Lamb and Horvath, 1991). TheseORFs were previously “hidden” employing unusual mechanisms for theirtranslation. The 14 kDa NS2 protein is derived from splicing of the NS1transcript and the 97 amino acid M2 polypeptide is produced after splicing ofthe M1 RNA. As is the case in retroviruses both unspliced and splicedmRNAs are found in the cytoplasm of influenza virus infected cells. Anothersmall protein, the 100 amino acid NB glycoprotein, is translated from abicistronic mRNA containing two AUG initiating codons separated by fournucleotides utilizing a different reading frame than that encoding the 466amino acid neuraminidase glycoprotein. In the RNA segment 7 the initiationcodon of the 12 kDa BM2 protein overlaps with the termination codon of theM protein such that translation of the two tandem cistrons is coupled.Information concerning the identification and function of polypeptidesencoded by small ORFs found in viral genomes is likely to increasedramatically in the next few years as a result of the interest generated byrecently identified proteins (eg. Rev) and the development of sensitivetechniques to detect them.1.7. COS Cell Expression SystemThe COS cell system provides a means of studying cloned genes in amammalian cell line. Bacterial plasmids containing a minimal SV4O origin20sequence are amplified in the presence of T antigen produced by the host cell(Mellon et al., 1981). This host-vector system can be used to study promotersbecause the origin fragment lacks sequences essential for efficient earlytranscription. Since the DNA replicates episomally, gene expression is notcomplicated by position effects that might occur if the DNA was integratedinto a chromosome. Also, because this is a mammalian cell line the RNAshould be spliced and foreign proteins correctly modified after translation.1.8. The Present StudyThe objective of these studies was to investigate gene expression in thehuman pathogenic parvovirus B19 at the RNA and protein levels.Specifically, my goal was to precisely map the splice junctions of the two sizeclasses of small, abundant transcripts and to determine if either class of RNAwas translated into protein. At the start of this project, the only cells known tosupport a B19 infection were human bone marrow cells. Because we did nothave a ready supply of bone marrow donors, the COS cell expression systemwas used to study B19 gene expression. Previous work from this laboratoryhad defined the genomic sequence of two B19 isolates (Shade et a!., 1986;Blundell et a!., 1987) and identified a single functional promoter at m.u. 6.(Blundell et a!., 1987). The COS cell expression studies were initiated byCaroline Beard who cloned the B19 genome from pYT1O3 (Cotmore et a!.,1986) into the SV4O on vector, pSVOd (Mellon et al., 1981). The SV4O-B19hybrid vector was shown to express the known B19 nonstructural andstructural proteins (Beard et a!., 1989). This system was the source of viralmRNA used to construct a B19 cDNA library from COS cells transfected withthis SV4O-B19 hybrid vector.21In the present studies I have shown that B19 can replicate in humanbone marrow cells that were isolated from an individual with a chronicmyelogenous leukemia. Cytoplasmic RNA isolated from these infected CMLcells was used to construct a second cDNA library utilizing a technique whichenriched the library for B19 cDNAs. The cDNA libraries (COS cell and CMLcell) were screened and selected B19 cDNAs from both libraries weresequenced and compared. Two size classes of small, abundant cDNAs wereinvestigated to determine if either class could direct the synthesis of apolypeptide. Antibodies were raised in rabbits by immunization withsynthetic peptides that were derived from antigenic regions within thepotential proteins encoded by the largest ORFs of these cDNAs. The rabbitantisera were used to detect proteins made from the corresponding RNAs invitro in a cell free system, in transfected COS-7 cells, and in B19 infectederythroid progenitor cells. At least two 11 kDa proteins encoded by thesmallest RNAs were identified. Site-specific mutagenesis was used to showthat multiple forms of the 11 kDa polypeptide arise from translationalinitiation at more than one AUG codon on the same transcript. The cellularlocalization of the 11 kDa proteins was identified using indirectimmunofluorescence and a preliminary functional analysis of these smallproteins was begun. Using the same analysis I have shown that the other sizeclass of small, abundant RNA is probably not translated into protein.22MATERIALS AND METHODS2.1. MaterialsAll chemicals used were reagent or analytical grade and werepurchased from BDH Inc., Fisher Scientific Co., or Sigma Chemical Co. unlessotherwise specified. Polyacrylamide gel electrophoresis reagents and proteinmolecular weight standards were supplied by either Bio-Rad Laboratories orBethesda Research Laboratories (BRL). Acetonitrile was from AldrichChemical Company. Ultrapure phenol, SDS, Tris, BSA, DTT, and agarosewere obtained from I3RL. GeneClean kits were purchased from Biocan andSep-Pak C18 cartridges were from Millipore. Electron microscopy (EM) gradeparaformaldehyde was obtained from J.B.S. Chemical.Bactotryptone, yeast extract, and bactoagar were from DifcoLaboratories. Ampicillin was supplied by Ayerst Laboratories. Cell culturemedia and fetal bovine serum were purchased from Gibco Canada Ltd. anderythropoietin was from Terry Fox Laboratories (Vancouver, B.C.). Salts,buffers, and amino acids used in tissue culture were supplied by SigmaChemical Company or Gibco Canada Ltd. DMSO was from Eastman KodakCo. or Fisher Scientific Co.Radioisotopes were purchased from New England Nuclear orAmersham. NTPs, dNTPs, and ddNTPs were supplied by Pharmacia P-LBiochemicals. Yeast tRNA was from BRL. Oligonucleotide primers weresynthesized on an Applied Biosystems DNA synthesizer in the BiochemistryDepartment at U.B.C.Restriction endonucleases were obtained from BRL, United StatesBiochemical Company (USB), New England Biolabs, or Promega Biotec andused according to the suppliers’ specifications. Calf intestinal alkaline23phosphatase (ClAP) was from Promega Biotec. Taq DNA polymerase wassupplied by Perkin Elmer-Cetus Corp. M-MLV reverse transcriptase, Kienowfragment, T4 DNA polymerase, T4 DNA ligase, SP6 and T7 RNApolymerases, were from BRL. DNA sequencing kits and modified T7 DNApolymerase were purchased from USB. DNaseI and RNaseA were fromSigma Chemical Company.Rabbit reticulocyte lysate was purchased from Promega Bidtec orBiocan. Conjugated anti-rabbit and anti-human secondary antibodies wereprovided by Jackson Laboratories or BRL. NBT, BCIP, and streptavidin TexasRed were from BRL. Heat-killed Staph A cells were obtained from Zymed andProtein A Sepharose CL-4B was from Pharmacia P-L Biochemicals. Normalgoat serum was provided by the animal care facility at U.B.C. Freund’sadjuvant was obtained from BRL.Transfer membranes for blotting of nucleic acids (GeneScreenPlus,Colony/PlaqueScreen) were obtained from New England Nuclear ResearchProducts and membranes for protein transfer were from Schleicher & Schuell(Optibind NC) or Millipore Corp. (Immobilon PVDF).X-OMAT AR film was supplied by Eastman Kodak Co. and Curix RP-1film was from Agfa-Gevaert.2.2. Strains and Media2.2.1. Plasmids819 DNA, from the serum of a child with homozygous sickle-celldisease during the early stages of a reticulocytopenic aplastic crisis, wasmolecularly cloned in two parts into pAT153 (Cotmore et a!., 1986). Theplasmid pYT1O3 contains the entire coding sequence of the B19-Au genome(5112 bp) but is missing an A residue at nt 3940 generated during the24molecular cloning and creating a frameshift mutation in the capsid genes(Shade et al., 1986). The B19 sequence from plasmid pYT1O3 was cloned intopSVOd (Beard et a!., 1989). The resultant plasmid, pSVOd/Bl9wt, wasmodified by exchanging the Eco RI to Sma I fragment (nt 1-2070) with thesame fragment from a B19 Wi clone (a separate B19 isolate) to obtainpSVOd/B19z170. The Sma I to Kpn I fragment (nt 2070-4080) was replacedwith the same restriction fragment from another B19-Au clone to correct theframeshift mutation in pYT1O3. The resulting plasmid, pSVOd/B19M70(A4),was used in these studies and for simplicity is referred to here aspSVOd/B19A170. [The original clone with the frameshift mutation in thecapsid genes is now labeled pSVOd/B19A170(A3)].Due to the instability of thepalindromic termini in E. coli, these B19 sequences contain hairpin deletionsat the left-hand terminal and right-hand end and therefore the plasmidwould not be expected to be infectious in permissive cells. A full-length cloneof B19 has recently been constructed (Deiss et a!., 1990) but was not used inthese studies. (N.B. All nucleotide positions within the B19 genomementioned in this thesis correspond to those published by Shade et a!., 1986.Because of the left and right hand hairpin deletions this numbering system isdifferent from that of the true full-length genome.)pGEM-4Z (Promega Biotec) was used to clone the ff19 human and COScell cDNA libraries in addition to restriction fragments of the B19 sequencefrom pYT1O3. The vector contains SP6 and T7 RNA polymerase promotersites flanking the multiple cloning site used to generate RNA transcripts foruse as hybridization probes and for translation in vitro. In pGEM-3Z theorientation of the SP6 and T7 polymerase sites are reversed.pGEM-3Zf’ and pGEM-3Zf’ (Promega Biotec) were used to generatesingle-stranded DNA templates for sequencing and mutagenesis. These25plasmids are essentially pGEM-3Z containing the Fl origin sequence of asingle-stranded bacteriophage (in either orientation) for the generation ofsingle-stranded DNA in the presence of the helper phage M13K07 (Vieira andMessing, 1987) or R408 (Russel et a!., 1986).pSELECT-1 (Promega Biotec) is a phagemid vector containing twogenes for antibiotic resistance. The tetracycline (tet) resistance gene is alwaysfunctional but the ampicillin (amp) resistance gene contains a filled-in PstIsite causing a frameshift mutation such that bacterial cells containing theplasmid must be propagated in the presence of tet. The amp resistance gene isrepaired with an oligonucleotide (oligo) which is annealed to the single-stranded DNA along with a mutagenic oligo, providing a selection in site-directed mutagenesis. The vector also contains SP6 and T7 RNA polymerasebinding sites and an Fl origin of replication.pSVOd, a gift of Dr. Tom Maniatis (Harvard University, Boston, MA),is a derivative of pBR322 containing 243 nt of SV4O origin sequences butlacking a functional promoter (Mellon et a!., 1981). The pBR322 sequenceswhich are inhibitory to replication in mammalian cells have been removed.This plasmid was used to construct SV4O-B19 hybrid vectors for theexpression of B19 genes in COS-7 cells (Gluzman, 1981).pCMV5 (Andersson et a!., 1989) contains the SV4O origin of replicationand the major immediate early promoter and enhancer from the humancytomegalovirus genome. This vector was used to express individual clonedcDNAs in COS-7 cells in the absence of the other B19 proteins.pM2, constructed by Dr. Ivan Sadowski (Dept. of Biochemistry, UBC), isalso a COS cell expression plasmid. The vector contains an SV4O early regionwhich consists of the origin of replication and early promoter sequences. Thecloned DNA is expressed as a fusion protein with the 147 aa DNA binding26domain of the yeast GAL-4 protein. This plasmid is co-transfected into COScells with a reporter plasmid, pG5BCAT, containing five GAL-4 DNA bindingsites upstream from a core promoter (Elb TATA box) driving the expressionof the bacterial chioramphenicol acetyl transferase (CAT) gene (Martin et a!.,1990). CAT assays on cell extracts are used to determine if the fusion partnerof GAL-4 is a transcriptional activator (Sadowski and Ptashne, 1989).2.2.2. BacteriaDH5cx [F- Ø8Od1acZAM15 A(ZYA-argF)U169 endAl recAl hsdRl7(rkmk) deoR thi-1 supE44 ? gyrA96 relAl] is a recombination-deficientsupressing strain that was used for all routine cloning of plasmids. Theø8OlacZAMl5 permits a-complementation with the amino terminus of 13-galactosidase encoded in pUC vectors (Hanahan, 1983; BRL, 1986).RZ1032 [hfrKLl6 P0/45 lysA6l dutl ungl thu relAl Zbd-279:TnlOsupE44Ol is a dut- ung strain of E. coli that was used for the preparation ofuracil-containing single stranded DNA templates for in vitro mutagenesis(Kunkel et a!., 1987).JM109 [recAl supE44 endAl hsdRl7 gyrA96 relAl. thi A(lac-proAB)F’(traD36 proAB !aqI LacZAMI5)] cells were used to propagate pSELECT andpGEM-Zf’ vectors for the production of single-stranded DNA. JM109 is recAand lacks the E. coli K restriction system (Yanisch-Perron et a!., 1985).BMH 71-18 mut S [thi supE iX(lac-proAB) (mutS::TnlO) (F’ proABlaqIZAM15)] is a mismatch repair minus strain of E. coli which preventsrepair of the newly synthesized unmethylated DNA strand leading to highmutation efficiencies after site-directed in vitro mutagenesis.Bacteria were routinely grown in YT medium (8 g tryptone, 5 g yeastextract, 5 g NaC1 per liter) supplemented with either 100 tg/ml ampicillin or2750 iig/ml tetracycline to selectively maintain plasmids. TYP broth (16 gbactotryptone, 16 g yeast extract, 5 g NaCI, 2.5 gK2HPO4per liter) was used toculture JMI.09 cells containing pGEM-Zf’ phagemid DNA when generatingphagemids for single-stranded DNA isolation. RZ1032, JM109, and BMH 71-18mut S cells were maintained on minimal plates supplemented with 1. mMthiamine-HC1 to select for the Ft factor.2.2.3. COS-7 CellsCOS-7 cells are derived from the simian kidney cell line CV-1 (Mellonet at., 1981). This cell line has been transformed with an origin-defective SV4Ovirus such that the integrated viral sequences produce SV4O T antigen and arepermissive for SV4O viral replication. Transfected SV4O-ori plasmids replicateto high copy number (200,000-400,000 copies/cell). COS-7 cells were cultured asmonolayers in DMEM containing 10 mM Hepes pH 7.4 and 10% fetal bovineserum (FCS) at 370 in 5% CO2. The cells were routinely split by trypsinization1:10 every 3 days. After approximately 50 passages the cells were discarded andfresh cells (low passage number cells were resuspended in DMEM containing10% DMSO and stored in liquid nitrogen) were cultured.2.2.4. Human Bone Marrow CellsHuman bone marrow aspirate cells from a patient with a chronicmyelogenous leukemia (CML) were provided by Dr. Keith Humphries (TerryFox Laboratories, U.B.C.). The cells were passed over a Percoll density gradient(density=1.066) and light density cells were collected and stored at _1350 as asource rich in hematopoietic progenitors (Eaves et al., 1980).282.2.5. B19 VirusSerum containing B19 virus was supplied by Dr. R. Gascoyne(ChildrentsHospital, Vancouver, B. C.). The serum was from a 13 year old girlwith hereditary spherocytosis who was undergoing treatment for a suddenonset aplastic crisis. Viral DNA was quantitated by dot blot hybridizationusing cloned plasmid pYT1O3 as the standard. B19 positive human serum wasalso provided by Dr. Bernard Cohen (PHLS Virus Reference Laboratory,London, England).2.3. Basic Molecular Cloning Techniques2.3.1. Isolation of DNA Fragments from Agarose GelsDNA was isolated from regular agarose gels using GeneClean kits bythe method specified by the manufacturer. DNA was also purified from lowmelting-point (LMP) agarose gels by melting the gel slice in a small volume ofTE [10 mM Tris (pH 8), 1 mM EDTA] for 3 mm at 680, followed by one phenoland two or three phenol/chloroform extractions. DNA was EtOH precipitatedfrom the final aqueous layer using 10 ig of yeast tRNA as a carrier.2.3.2. Cloning DNA Fragments into Plasmid VectorsIn a typical ligation experiment, fragment and vector DNA were ligatedin a 1:1 molar ratio in a 15 tl reaction volume containing 1 U of T4 DNAligase in a buffer of 50 mM Tris (pH 7.6), 10 mM MgCl2, 1 mM ATP, 1 mMdithiothreitol (DTT), and 5% (w/v) polyethylene glycol-8000 (PEG-8000). TheDNA was ligated for 2 h at RT if the ends were cohesive or 4 h at RT if theends were blunt. Vector DNA that was digested with only one enzyme wasdephosphorylated with calf intestinal alkaline phosphatase (ClAP) prior toligation to suppress self-ligation. Briefly, 1 ig of linearized plasmid DNA was29reacted with ClAP (0.1 U/pmol ends) in a buffer containing 1 mM ZnC12, 1mM MgC12, and 10 mM Tris (pH 8.3) in a 50 tl volume for 60 mm at 370• Ifthe 5’ termini were blunt or recessed the concentration of ClAP was increasedto 1 U/pmol ends and the reaction temperature was increased to 550• Thereaction was stopped by the addition of 1 tl of 0.5 M EDTA and heated at 650for 20 mm followed by extraction with phenol/chloroform and EtOHprecipitation.2.3.3. Preparation and Transformation of Competent CellsCompetent cells were prepared by innoculating 100 p.1 of an overnightculture into 35 ml of fresh YT broth. When the OD at 590 nm was 0.200 thecells were centrifuged for 10 mm at 5000 rpm at 40• The pellet wasresuspended by vortexing in 1/2 vol of 50 mM CaC12 and left for 30 mm onice. The cells were recentrifuged as above and the final pellet was gentlyresuspended in 1/10 vol of 50 mM CaC12 containing 15% glycerol. Aliquots ofcompetent cells (200 p.1) were rapidly frozen in a dry ice/ethanol bath andstored at _700. The efficiency of DH5cL cells prepared this way was 5 x 106 to 2 xio transformants/p.g plasmid DNA.For transformation, 1/3 of the ligation reaction was diluted 5X in waterto a volume of 25 p.1. This was added to 100 p.1 or 200 p.1 of thawed, competentcells and left on ice for 30 mm. The cells were briefly heat-shocked at 420 for 90sec, then YT was added to 1 ml and the cells were recovered at 370 for anadditional 20 mm in a water bath with mild agitation. Fifty microliters of thetransformation mixture was spread onto YT agar plates containing theappropriate antibiotic. If the plasmid and bacteria allowed xcomplementation 50 p.1 X-gal (2% in dimethyformamide) and 50 p.1 IPTG (100mM) were spread onto the plate prior to the bacteria.302.4. Purification of Plasmid DNA2.4.1. Small Scale Plasmid DNA IsolationPlasmid DNA was routinely isolated from 1.5 ml of bacterial cells bythe alkaline lysis method (Sambrook et at., 1989). If the plasmid DNA was tobe sequenced a modification of the boiling lysis method was preferred(Holmes and Quiqley, 1981). The pellet from 1.5 ml of an overnight culture ofbacterial cells was resuspended in 100 il STET [10 mM Tris (pH 8.0), 50 mMEDTA, 5% Triton X-100, 8% sucrose, 0.5 mg/mi lysozymel and boiled for 2mm. Cell debris was pelleted by centrifugation for 15 mm at 4Q Isopropanol(100 pi) was added to the supernatant and the lysate was recentrifuged asabove. The final pellet was washed with 70% EtOH, dried, and resuspended in25 jil of dH2O or TE.2.4.2. Large Scale Plasmid DNA IsolationPlasmid DNA was purified from 100 ml and 500 ml bacterial culturesby the alkaline lysis method followed by either precipitation withpolyethylene glycol or equilibrium centrifugation in CsC1-ethidium bromidegradients (Sambrook et at., 1989).2.5. Isolation of Single-Stranded DNA from PhagemidsSingle-stranded DNA was generated from pGEM-Zf’ phagemids byinnoculating 100 tl of an overnight culture of the cells containing thephagemid DNA into 5 ml of TYP broth containing amp (100 pg/m1) in a 50ml flask. After shaking vigorously at 370 for 30 mm the cells were infectedwith either helper phage M13K07 or R408 at an m.o.i. of 10-20. The cells weregrown as above for an additional 6-8 h and the supernatant was harvested bypelleting the cells twice at 12,000 x g for 15 mm at 40 Phage was precipitated by31adding 0.25 vol of phage precipitation buffer [3.75 M ammonium acetate (pH7.5), 20% PEG (MW 8000)1 to the supernatant. The mixture was kept on ice for30 mm, then centrifuged again at 12,000 x g for 15 mm. The pellet wasresuspended in 400 .tl of TE and extracted once with chioroform/isoamylalcohol (24:1) and then with phenol/chloroform (50:50) until there was novisible material at the interface. Single-stranded DNA was recovered by EtOHprecipitation and the final DNA pellet was suspended in 20 il dH2O. Theamount of single-stranded DNA was estimated by agarose gel electrophoresisof a 2 jil sample. Moderate amounts of helper phage single-stranded DNAwas also recovered by this method but this was not found to interfere withany subsequent reactions.2.6. Double-Stranded DNA SequencingDouble-stranded DNA was prepared for sequencing by boiling 2-4 tg oftemplate and 1-2 pmol primer for 2 mm in 0.2 N NaOH followed by EtOHprecipitation. The DNA was resuspended in 8 il of dH2O and 2 tl of 5Xsequencing buffer [200 mM Tris (pH 7.5), 100 mM MgC12, 250 mM NaC1I andannealed for 15 mm at 370• Sequencing was performed in micro-titer platesusing modified T7 DNA polymerase and the sequencing mixes supplied byUSB incorporating x[32P]dATP (3000 Ci/mmol) into the newly synthesizedDNA strands. The reaction products were separated on a 38 cm x 18 cm x 0.4cm 8% polyacrylamide gel containing 8 M urea run at 32 W constant power.The dried gel was subjected to autoradiography (AR) using Curix RP-1 film.322.7. Preparation of Labeled Hybridization Probes2.7.1. DNA ProbesDNA probes were made by random hexamer labelling (Feinberg andVogeistein, 1983). Linearized plasmid DNA or gel-purified fragment DNA(50-100 ng) was boiled for 5 mm with 1.4 jtl random 6-mers (90 OD units/mi)and chilled on ice to anneal. This mixture was added to 5 p,l 1.0 M HEPES (pH6.6), 5 p,l dNTPs (100 jiM each dCTP, dGTP. dTTP), 10 p,g BSA, 3-5 p,ia[32P]dATP (3000 Ci/mmol) and 5 U Kienow in 25 p,l total volume. After 4 hincubation at RT the reaction was stopped with 100 p,l stop mix [20 mM NaCl,20 mM Tris (pH 7.5), 2 mM EDTA, 0.25% SDS, 1 IIM ATPI. The probe waspurified from unincorporated nucleotides on a spun column of Sephadex G25 or G-50 fine or by EtOH precipitation using 10 p,g tRNA as a carrier.2.7.2. RNA ProbesRNA probes were made utilizing vectors containing SF6 or T7 RNApolymerase promoter sites using a standard transcription protocol modifiedfrom the procedure of Melton et al. (Melton et a!., 1984) and described in thePromega Protocols and Applications Guide (1990). First, the vector DNA waslinearized with a suitable restriction enzyme to produce‘1run-off” transcriptsderived from insert sequences only. Following digestion, the DNA wasextracted with phenol/chloroform and precipitated with EtOH. All solutionsused for transcription were made with DEP-treated (diethyl pyrocarbonatetreated) dH2O to inactivate RNases. A typical reaction contained 4 p,l 5Xtranscription buffer [200 mM Tris (pH 7.5), 30 mM MgCl2, 10 mM spermidine,50 mM NaC1], 2 iii 100 mM DTT, 2 p,l 1 mg/ml BSA, 20 U RNasinribonuclease inhibitor, 4 p,l NTPs (2.5 mM each ATP, GTP, CTP; 0.25 mMUTP), 1 p,l linearized template DNA (0.5 mg/mi), 1 p,l x[32P]UTP (80033Ci/mmol), and 5-10 U SP6 or T7 RNA polymerase in a total reaction volumeof 20 jil. The reaction was incubated at 37400 for 60 mm. The DNA templatewas removed by the addition of RQ1 RNase-free DNase (1 U/iig DNA)followed by incubation at 370 for an additional 15 mm. The labeled RNA wasextracted with phenol/chloroform and precipitated with EtOH using 10 igyeast tRNA as a carrier. The RNA sample was resuspended in DEP-treateddH2O and was used within 24 h.Incorporation of label was quantitated by precipitation of 1 il of thetranscription reaction or the random primer labeling reaction in 5% TCAfollowed by Cerenkov counting of the dried filters.2.8. DEAE-Dextran Mediated Transfection of DNA into COS-7 CellsCOS-7 cells were transfected using DEAE-dextran followed by a DMSOshock (Lopata et a!., 1984). The day prior to transfection a confluent dish ofcells was split into four new plates. For transfection the cells, at approximately40% confluency, were washed twice with DMEM which was notsupplemented with fetal calf serum (FCS) and then incubated in the samemedium containing 200 jig/mi DEAE-dextran and 2-5 ig plasmid DNA per100 mm dish in a total volume of 3 ml. After 8 h of incubation at 370 themedium was removed and the cells were shocked for 3 mm in 10% DMSO inHBS (21 mM HEPES, 135 mM NaC1, 5 mM KC1, 0.8 mM Na2HPO4,5 mMdextrose). The treated cells were then incubated in DMEM plus 10% FCS foranother 48-72 h before harvesting.2.9. B19 Infection of Human Leukemic Bone Marrow CellsMarrow cells ( approximately 4x107 cells) were prepared for infectionby quickly thawing at 370 and diluting lOX in IMDM containing 10% FCS and3425 ig DNaseI to reduce cell clumping. After pelleting, cells were incubated inIMDM with 10% serum and 50 ig/ml DNaseI for 15 mm. One half of the cellswere infected with 20 p1 819-containing serum (approximately 60 ng 819DNA) in 400 i1 IMDM at 40 for 2 h followed by incubation in IMDM, 20%FCS, and erythropoietin (1.5 U/mi) at 370 in 5% CO2. Two hours postinfection (h.p.i.), 10% of both the infected and control uninfected cells wereisolated and low molecular weight DNA was isolated. The remaining cellswere harvested 48 h.p.i.2.10. Isolation of Low Molecular Weight DNA from Cultured CellsLow molecular weight DNA was isolated by a modification of the Hirtprocedure (Hirt, 1967). Cells were lysed in 10 mM Tris (pH 7.5), 10 mM EDTA,0.6% SDS for 5 mm, then NaC1 was added to a final concentration of 1.0 M.After an overnight incubation at 40 the lysates were centrifuged at 12,000 x gand the supernatants were extracted with phenol/chloroform andprecipitated with EtOH. The DNA was resuspended in TE and digestedsequentially with 40 jig/mi RNaseA and 200 jig/mi ProteinaseK (ProK)followed by another phenol/chloroform extraction and EtOH precipitation.2.11. Southern BlottingDNA was electrophoresed on a 0.8% agarose gel run in TBE andtransferred to a nylon membrane (GeneScreen Plus) using a vacuum blottingsystem (LKB 2016 VacuGene). The gel was placed in contact with themembrane supported on a porous screen and a vacuum of 50 cm dH2O wasapplied. The DNA was depurinated for 20 mm in 0.25 M HC1, denatured forthe same time in 1.5 M NaC1, 0.5 N NaOH; and neutralized in 1 M Tris, 2 MNaCl, (pH 5). Finally, the DNA was transferred in 20X SSC (3 M NaCI, 0.3 M35sodium citrate) for 1 h. The filter was removed from the vacuum blottingapparatus, rinsed briefly in 2X SSC, then dried for 10 mm in a 600 oven.2.12. Isolation of RNA from Cultured CellsTotal cytoplasmic RNA was isolated by the method described byKaufman and Sharp (Kaufman and Sharp, 1982). COS-7 cells, growing inmonolayers, were washed twice and collected by scraping in ice cold PBS. Forbone marrow suspension cultures the cells were immediately pelleted, thenresuspended in ice cold PBS, and washed by repelleting and resuspending incold PBS. The cell pellet from a 100 mm dish (or flask) was resuspended in 250il hypotonic lysis buffer [10 mM Tris (pH 7.4), 10 mM NaC1, 3 mM MgC12, and25 jil vanadyl-ribonucleoside complexes (VRCYI and after a 5 mm incubationNonidet P-40 (NP-40) was added to a final concentration of 0.5%. The lysatewas vortexed and centrifuged for 5 mm. The supernatant was transferred to afresh tube containing 250 p1 2X ProK buffer [300 mM NaC1, 200 mM Tris (pH7.5), 40 mM EDTAJ, 50 p1 10% SDS, and 50 jig ProK. Incubation was at 370 for30 mm. The lysate was extracted once with phenol and two times withphenol/chloroform followed by EtOH precipitation. The final RNA pellet wasresuspended in 50 p1 DEP-treated dH2O and 2 p1 of each RNA sample was runon a 1% agarose gel containing 0.5% SDS to check the integrity of the RNApreparation.2.13. Northern Blotting2.13.1. RNA GelsRNA was electrophoresed through gels containing formaldehyde by amodification of the method described by Sambrook et a!. (Sambrook et a!.,1989). The gel was prepared by melting agarose in DEP-treated dH2O, then36adding lox MOPS buffer [0.2 M MOPS (pH 7.0), 80 mM NaOAc, 10 mM EDTAin DEP-treated dH2O] and deionized formaldehyde (pH >4) to a finalconcentration of 1X and 2 M respectively. RNA samples were prepared in 50%formamide, 2 M formaldehyde, and 1X MOPS buffer, heated at 560 for 15 mm,then cooled on ice prior to loading. Gels were prerun at 70 V for 5 mm andthen run at 50 V in 1X MOPS buffer for 4 h. The size of the RNA transcriptswas determined either by running[32P]-labeled DNA markers or unlabelledRNA standards in one lane of the gel. If RNA standards were used themarker lane was cut out of the gel after electrophoresis and stained for 1 h in0.1 M NH4OAc containing 0.5 ig ethidium bromide and the RNA bands werevisualized under UV light.2.13.2. RNA TransferRNA was transferred to a nylon membrane (GeneScreen Plus) bycapillary elution. After electrophoresis the gel was rinsed briefly in DEPtreated dH2O. The membrane was prepared by first wetting in dH2O thensoaking in 20X SSPE (3 M NaC1, 0.2 M NaHPO4HO,20 mM EDTA) for 30mm. The RNA was transferred in 20X SSPE by capillary action for 12-16 husing the method described in Sambrook et al. (Sambrook et al., 1989). Aftertransfer, the filter was rinsed in 2X SSPE, air dried for 15 mm, then baked at8O0for2h.2.14. Hybridization of FiltersFilters were initially soaked in 5X SSPE, then prehybridized in a sealedplastic bag containing 50% formamide, 5X SSPE, 1.0% SDS, 0.1% Tween 20,and 100 iig/ml yeast tRNA for 1-2 h at 450• The filters were hybridized in thesame solution plus probe DNA or RNA (specific activity 107108 cpm/JIg37nucleic acid) for at least 16 h at 42480. After hybridization, the membraneswere washed in 1X SSPE and 0.1% SDS for 1 h at RT followed by washing in0.1X SSPE and 0.1% SDS for 1 h at 60630. The filters were covered with SaranWrap and autoradiographed without drying.2.15. cDNA Libraries2.15.1. Construction of B19 Human Bone Marrow Cell cDNA LibraryTotal cytoplasmic RNA was isolated from human bone marrow cells 48h after B19 infection. To synthesize the first cDNA strand 10 pg of total RNAwas heated at 650 for 5 mm then cooled immediately on ice. The RNA wasadded to a 35 p1 cocktail consisting of 8 p1 5X M-MLV (Moloney murineleukemia virus) reverse transcriptase buffer [1X= 50 mM Tris (pH 8.3), 75 mMKC1, 3 mM MgC12, 10 mM DTT], 1 U/il RNasin, 4 jig BSA, 1 iil each dNTP (10mM), 42 pmol oligo(dT) primer (see below) and 200 U M-MLV reversetranscriptase in a total volume of 40 jil. The reaction was incubated at 370 for 1h, another 200 U of reverse transcriptase were added, and the reaction wascontinued at the same temperature for another hour. The RNA was degradedwith the addition of 140 mM NaOH, 1.4 mM EDTA and incubation at 1000 for5 mm. The cDNA was extracted with phenol/chloroform and precipitatedwith EtOH using 10 jig tRNA as a carrier.Second strand cDNA synthesis and rapid cDNA amplification wascarried out using the polymerase chain reaction (PCR) (Saiki et a!., 1988) by amodification of the procedure of Frohman et a!. (Frohman et a!., 1988). Analiquot of the first strand cDNA (5-10%) was added to 2 p1 of each dNTP (10mM), 5 p1 lOX PCR buffer [1X= 10 mM Tris (pH 8.4), 0.05% Tween 20, 0.05%NP-40}, 6 p115 mM MgC12, 1 pmol oligo(dT), 20 pmol B19 specific oligo, 20pmol MCS oligo and 1 U Taq DNA polymerase in a total volume of 50 p1. The38sequence of the B19 specific oligonucleotide was derived from a region of the5’ leader sequence which is common to all 819 transcripts and the MCS oligois a truncated oligo(dT) containing the multiple cloning sequence without the17 T residues so that it has a similar annealing temperature to the B1.9 oligo(see below). The primers used to generate the library contained restrictionendonuclease sites at their 5’ ends to facilitate subsequent cloning of thecDNA products.B19 specific primer (27mer)5’ TCTAGAATTCTCTTTCTGGGCTGCTTT 3’X1a iI IEcoRl1st strand primer (37mer)5’ CGAGCATGCGTCGACAGGCAT 3’ii 17SphI TaqISal IHincilAccIThe samples were overlayed with light mineral oil and the DNA wasamplified in a thermocycler block. The amplification program consisted of 5cycles of: 940 denaturation for 15 sec, 450 annealing for 30 sec, and 720extension for 2 mm; followed by 30 cycles of: 940 denaturation for 15 sec, 550annealing for 30 sec, and 720 extension for 2 min.The final cycle was followedby a 10 mm incubation at 720. The PCR reaction products were extracted withchloroform to remove the mineral oil and precipitated with EtOH. 10% of39each reaction was analyzed on an agarose gel and the products from threePCR reactions were pooled to construct the library.The amplified cDNA was digested with Eco RI and Sph I, cloned intopGEM-4Z, and transformed into library competent DH-5a cells.2.15.2. Construction of B19 COS Cell cDNA LibraryThe COS cell cDNA library was constructed by Caroline Beard(presently at Whitehead Institute, Boston, MA). Plasmid pSVOd/B19 wasprepared by cloning the Eco RI fragment from pYTIO3 into the unique Leo RIsite of pSVOd. This fragment contains the promoter and the entire proteincoding region of the B19-Au genome but lacks —187 nt of the left-handhairpin and —297 nt of the right-hand hairpin (Deiss et a!., 1990). The B19-Augenome in pYTIO3 is missing an adenine residue at nt 3940 creating aframeshift mutation resulting in the expression of truncated VP proteins.This was corrected by replacing the 2 kb Sma I to Kpn I fragment (nt 2070-4080)in pSVOd/B19 with the same fragment obtained from another B19-Au clone.Plasmid pSVOd/B19 Afi 11 was made after digestion of pSVOd/B19 with AflII(unique site at 576 within the NS coding region), filling in with Kienowfragment, and recircularizing the plasmid. It was previously shown thatreplication is enhanced in SV4O-B19 hybrid plasmids in the absence of NSgene expression (Beard et a?., 1989).The cDNA library was made by transfecting pSVOd/B19 Afi 11 intoCOS-7 cells. Total cytoplasmic RNA was isolated and polyadenylated RNAwas selected by oligo d(T) cellulose chromatography. First-strand cDNA wassynthesized as above but second strand was made using Kienow fragment andthe B19 specific primer instead of PCR. The cDNA products were digested40with Sph I and Eco RI and cloned into pGEM-4Z as described for the humanbone marrow cell B19 library.2.15.3. Screening the cDNA LibrariesThe cDNA libraries were screened by transferring bacterial coloniesgrown on agar plates to nylon discs (Colony/PlaqueScreen). The discs werefirst treated with 0.5 N NaOH to lyse the cells and denature the DNA followedby neutralization in 1.0 M Tris pH 7.5. The filters were initially hybridizedwith an RNA probe corresponding to the Eco RI to Barn HI fragment (nt 1-3900) of the B19 genome in pYT1O3 to detect B19 cDNAs. To identify specificcDNAs the libraries were reprobed with labeled RNA derived from the HindIII fragment (nt 600-1540) for the NS transcript, the right-hand Pst I fragment(nt 3145 to 3860) for the VP transcripts, the Ava I to Pst I fragment (nt 2080-3145) for the 700 and 800 nt size class of cDNA, or the Barn HI to Eco RIfragment (nt 3900-5112) to detect the 500 and 600 nt class of cDNA (refer toAppendix A for the pYTIO3 restriction map).2.16. Peptides and AntiseraPeptide 1, containing amino acids corresponding to aa 58-86(PNTKDIDNVEFKYLTRYEQHVIRMLRLC) of the 94 amino acid (aa) potentialpolypeptide encoded by the extreme right-hand ORF, and peptide 2, derivedfrom aa 34-51 (ASWEEAFYTPLADQFRELGGC) of the proteins encoded by themajor left-hand ORF were synthesized by Dr. Ian Clarke-Lewis (BiomedicalResearch Center and Dept. of Biochemistry, U.B.C.). Antigenic (hydrophilic)regions of the encoded proteins were predicted by computer analysis using theP.C. Gene ANTIGEN program (IntelliGenetics) following the method of Hoppand Woods (1981).41For each synthetic peptide two female rabbits were immunized bysubcutaneous injection of 300 jig to 1 mg of the peptide conjugated to keyholelimpet hemacyanin (KLH) in complete Freund’s adjuvant. IncompleteFreund’s adjuvant was used for subsequent injections at 4 to 6 week intervals.Sera were collected 10 to 14 days after the last injection. The test bleeds wereallowed to clot at 370 for 30 mm. The clot was ringed with a Pasteur pipet andleft to contract overnight at 40 The supernatant was removed and centrifugedfor 10 mm at 10,000 x g at 40 to pellet the remaining unclotted red blood cells.The serum was aliquotted into small volumes and stored at 200.Anti-capsid sera were also raised in rabbits by immunization with afusion polypeptide consisting of a fragment of VP-2 fused to the GST proteinfrom the parasitic helminth Schistosoma japonicum expressed in E. coli(Smith and Johnson, 1988). This work was done by Don Minato in thislaboratory.Rabbit antisera specific for B19 NS proteins and VP proteins wereobtained from Drs. Susan Cotmore and Peter Tattersall (Yale University; NewHaven, CT) (Cotmore et al., 1986). Monoclonal antibodies VRL/B19-3 andVRL/B19-11, raised against B19 virus, and human B19 convalescent serumwere provided by Dr. Bernard Cohen (PHLS Virus Reference Laboratory,London, UK).2.17. In vitro Translation, Immunoprecipitation, and Western Blotting2.17.1. In vitro Translation of SP6 Generated RNAsThe cDNA library cloned into pGEM-4Z provided a system to producebiologically active RNAs using SP6 RNA polymerase. RNA for translationwas transcribed in vitro incorporating the 5’ cap analogue, GpppG, by themethod described in the Promega Protocols and Applications Guide (1990).42The RNA was not radiolabeled and the standard transcription reaction wasmodified so that it included 0.5 mM 5’GpppG (the same concentration as CTP,ATP, and UTP) and reduced the concentration of GTP to 1/10th that of theother NTPs which permitted the incorporation of the 5t cap analogue at theinitial nucleotide position (always G) in the RNA transcript. The templateRNA was heated at 670 for 10 mm and immediately cooled on ice. In a typicaltranslation reaction: to 35 tl nuclease treated rabbit reticulocyte lysate wasadded 40 U RNasin ribonuclease inhibitor, 1 p.1 of a 1.0 mM mixture of aminoacids minus methionine (met) or cysteine (cys), 2 p.1 of RNA substrate (5-20p.g/ml final), and 4-5 p.1 translation grade [35S1 L-met or [35S1 L-cys (800Ci/mmol) in a total volume of 50 p.1. Brome mosaic virus (BMV) RNA wastranslated as a positive control and a reaction with no added RNA was usedas the negative control. The translation reactions were incubated for 60-90mm at 300. To analyze the translation products 5 p.1 of each reaction wasboiled in 20 p.1 1X sample buffer [10% glycerol, 2% SDS, 0.02% bromophenolblue, 62.5 mM Tris (pH 6.8), 1 p.1 14.4 M 13-mercaptoethanoll and the proteinswere separated by SDS-PAGE (Laemmli, 1970). The dried gels were subjectedto AR using X-OMAT AR film.2.17.2. Lysis of Mammalian Cells in Sample BufferProteins were harvested from transfected COS-7 cells or infected bonemarrow cells after washing the cells twice with PBS and lysing in 100-200 p.1 of1X sample buffer heated to 850. The viscous lysate was scraped into aneppendorf tube and boiled for 10 mm. The chromosomal DNA was shearedby repeated passage through a 26 /2 gauge needle and the lysate wascentrifuged for 10 mm in a microfuge. Aliquots (5-10 p.1) of the supernatantwere analyzed by SDS-PAGE and western blotting.432.17.3. Western BlottingWestern blotting was used to detect specific proteins. Afterelectrophoresis the proteins were transferred to nitrocellulose or PVDFmembranes by electroblotting at 100 V for 1 h in 25 mM Tris, 192 mM glycine.and 20% (v/v) methanol (Towbin et al., 1979). Non-specific protein bindingwas blocked with 5% FCS for 1 h to overnight. The filters were incubated atRT for 1. h with primary antibody (Ab) [1:1000 dilution for Abs raised againstpeptides] followed by incubation with a secondary Ab which was conjugatedto alkaline phosphatase (at a concentration specified by the supplier).Immunoreactive proteins were visualized using NBT/BCIP (nitrobluetetrazolium chloride /5-bromo-4-cloro-3-indolylphosphate) colour reactionsand molecular weight was determined using pre-stained standards.2.17.4. Immunoprecipitation of Radiolabeled ProteinsFor immunoprecipitations 5 il of the translation reaction wasincubated with 5 tl of antiserum and 1.90 il RIPA buffer (150 mM NaC1, 1%NP-40, 0.5% DOC, 0.1% SDS, 50 mM Tris pH 8.0) at 40 for at least 1 h. Protein ASepharose (25 p1) was added to the samples and incubation was continuedunder the same conditions for I h to overnight. After washing two times inRIPA buffer and once in 10 mM Tris (pH 7.5), 0.1% NP-40; the bound proteinswere released from the beads by boiling the samples for 2 mm in samplebuffer. Proteins were separated on Laemmli gels and the dried gels wereexposed to X-OMAT AR film.2.18. Metabolic Radiolabeling of Proteins Expressed in COS-7 CellsForty-five hours after transfection COS cells were washed twice withprewarmed DMEM lacking met (or cys) and containing no serum. Each 60mm44dish of cells was then incubated at 370 and 5% CO2 for 20 mm in 3 ml of thesame medium to deplete the intracellular pools of sulfur-containing aminoacids. The medium was replaced with 500 p1 DMEM containing 500 jiCiprotein labeling mix (an E. coli hydrolysate containing —70% [S] L-met,—20% [5S1 L-cys, and a number of oxidized by-products) and incubated for anadditional 3 h as above. The dishes were gently rocked every half hour toensure that all cells were covered with medium. After labeling, the dishes ofcells were placed on a glass plate resting on a bed of crushed ice and washed 2times with ice cold PBS. The cell monolayers were lysed with the addition of500 p1 ice cold RIPA buffer containing :100 jig/mi PMSF(phenylmethylsulfonyl fluoride) and 1 p.g/ml aprotinin for 20 mm at 00.Lysates were removed to eppendorf tubes and centrifuged at 12,000 x g for 2mm at 40 The supernants were transferred to fresh eppendorf tubes andeither quick frozen in a dry ice/EtOH bath or stored on ice. Lysates wereusually precleared with preimmune sera prior to immunoprecipitation withantisera. Preimmune or irrelevent sera (10 p1) were added to 200 p1 of the celllysate and incubated at 40 for 1 h. Non-specific proteins were precipitated afteran additional 30 mm incubation at 40 with 10 p1 of a 10% suspension of heatkilled, formalin-fixed, Staph A cells. The precleared lysate was then added to afresh, chilled eppendorf tube containing 300 p1 RIPA buffer and 5 p1 of thespecific antisera. Immunoprecipitation was carried out as describedpreviously for in vitro translated proteins.2.19. Indirect Immunofluorescence2.19.1. Transfected COS CellsAcid-etched 22 mm coverslips were prepared by boiling in 0.1 N HC1for 10 mm, rinsing thoroughly with dH2O and storing in 95% EtOH. The45coverslips were sterilized by flaming, inserted into 35 mm tissue culturedishes, and seeded with approximately io COS-7 cells. After the cells hadattached to the coverslips they were transfected with 1 ig DNA/dish in a 1. mlvolume of growth medium by the DEAE-dextran method (describedpreviously). Forty-eight hours after transfection the cells were rinsed withPBS and fixed in 3.7% EM grade paraformaldehyde in PBS for 10 mm at RT.After rinsing with PBS, the cells were permeabilized using 0.25% saponin(w/v) in PBS and 10% goat serum for 45 mm at RT. All subsequent wash andantibody incubation steps contained 0.1% saponin and 5% normal goat serumand were performed at RT. Goat serum was used to block non-specificprotein-protein interactions especially by the secondary antisera which wasraised in goat. After washing three times in PBS the cells were incubated for 1h with the primary antisera (dilution of 1:200 for rabbit polyclonal sera raisedagainst synthetic peptide). The wash step was repeated and secondary antibodywas added at the dilution recommended by the supplier and incubated asabove. The secondary antibody used was either goat cr-rabbit IgG Fc conjugatedto rhodamine or biotinylated goat c-rabbit IgG Fc. The secondary antibodyincubations and all subsequent steps in the procedure were carried out in thedark to reduce photobleaching of the fluorophore. When the biotinylatedsecondary antibody was used an additional 1/2 h incubation withstreptavidin-Texas Red was carried out before the final wash step. Thecoverslips were inverted and mounted on regular glass slides in 90% glycerolin PBS containing 2% Dabco (w/v) to reduce photobleaching. Phase contrastlight microscopy and indirect fluorescence microscopy were performed on aZeiss Universal photomicroscope equipped with an epi-fluorescence head.Photographic slides were taken using Kodak Ektachrome 400 ASA film.462.19.2. B19 Infected Bone Marrow CellsFor immunofluorescence studies of B19 infected human bone marrowcells the coverslips were treated with poiy L-lysine (10 mg/mi) for 30 mm atRT. Using a Pasteur pipette, three to four drops of cells were seeded onto thecoated coverslips and left for 30 mm at RT to adhere. The cells were thenwashed with PBS and incubated with the 10 and 20 Ab as described above forCOS-7 cells. Alternatively, infected marrow cells were attached directly toglass slides using a cytospin. The cells were diluted to 106 cells/mi, 3 dropsfrom a Pasteur pipette were dropped onto a glass slide and the cells were spunat 1600 rpm for 5 mm. The slides were subsequently treated in the same wayas coverslips following the procedure described above.2.20. Site-Specific MutagenesisThe 5’ end of the 94 aa ORF in the 518 and 638 nt cDNAs contains threeclosely spaced ATG codons. Since more than one 11 kDa protein is made fromthis ORF a mutational analysis was performed to determine if translation wasinitiated from more than one AUG codon.2.20.1. Mutation of the Second ATG in the 94 aa ORFThe second in frame ATG in the ORF encoding the 11 kDa proteins isin the best Kozak consensus sequence for translation initiation (Kozak, 1986).This codon was mutated to a CTG codon by changing the first A to a C using amodification of the method of Kunkel et al. (Kunkel et al., 1987). In thisprocedure a small number of uracil residues were substituted for thymineresidues in the template single-stranded DNA strand after replication in adut, ung E. coil strain. A mutagenic oligonucleotide was annealed to thetemplate priming the synthesis of a second DNA strand. When the47heteroduplex DNA was transformed into a dut+, ung+ strain the uracilcontaining template strand was preferrentially degraded with consequentsuppression of the production of wild type plasmid.The Barn HI to Eco RI fragment (nt 3900-5112) of the B19 genome inpYT1O3 was cloned into the phagemid pGEM3Zf’ (pGEM-3Zf’/BE) andtransformed into the RZ1032 strain of E. coli. Helper phage R408 was used toprepare uracil-containing single-stranded DNA templates (describedpreviously).The mutagenic oligonucleotides were 17-22 nt long, chemicallyphosphorylated at their 5’ termini, and designed such that there would be asingle base pair (bp) mismatch in the middle of the sequence. A DNAsequence analysis was routinely performed to ensure that the last 6 nt of theoligo hybridized uniquely to the target sequence. (refer to Appendix B for theDNA sequences of the mutagenic oligonucleotides.)Synthetic oligonucleotides were purified using a Sep-Pak C18 cartridge.The cartridge was prepared by washing with 10 ml acetonitrile followed by 10ml of dH2O. The oligo was resuspended in I ml 0.5 M NH4OAc, 10 mMMg(OAc)2 and applied to the column. The column was washed with 10 mldH2O and the oligo was eluted in 3 fractions of 1 ml each with 20%acetonitrile in dH2O. The yield of oligonucleotide was determined bymeasuring the 0D260 of a diluted sample.The mutagenesis protocol is described in Sambrook et a!. (Sambrook eta!., 1989). In a typical reaction 0.5 pmol of uracil-containing single-strandedDNA were added to 10 pmol of mutagenic oligo and 1 il lox PE1 buffer [200mM Tris (pH 7.5), 100 mM MgC12, 500 mM NaC1, 10 mM DTTI in a totalvolume of 10 pi. The mixture was heated for 5 mm at a temperature 200greater than the melting temperature (Tm) of the mutagenic oligo [Tm=484(G÷C) +2(A+T)J, then allowed to cool slowly to room temperature in thesame beaker in which it was heated. The components of the PE3 mixture; 1 E’1lox PE2 buffer [200 mM Tris (pH 7.5), 100 mM MgCl2, 100 mM DTT], 4 p1 2.5mM dNTPs, I p110 mM ATP, 5 U T4 DNA polymerase, 1 U T4 DNA ligaseand dH2O to 10 p1, were assembled on ice. The ice-cold PE3 mixture was addedto the annealed oligo/template and the reaction was incubated for 5 mm at 00,5 mm at RT, then 2 h at 370• Both RZ1032 and DH5cz cells were transformedwith 5 uI of the synthesis reaction. Fewer transformants from DH5c cellsindicated that the uracil system was working. A second transformationensured proper segregation of mutant and wild type plasmids. Initially,plasmid DNA prepared from 12 isolated colonies from the transformed DH5cplate was used to transform 12 aliquots of DH5a cells. Later, plasmid DNAwas prepared directly from the pool of transformed cells and this was used ina second transformation. This method routinely produced >60% mutants sothe DNA was readily screened by sequencing.The same mutagenic oligonucleotide and the same procedure was usedto mutate the 518 bp cDNA cloned into pGEM-3Zf. The entire cDNA wassequenced to ensure that there were no additional nucleotide changes andwas then cloned into pCMV5 for COS cell expression studies.To construct a mutant plasmid in pSVOd the Kpn I to Aft III DNAfragment (nt 4080 to 4945 of the B19 sequence) in pSVOd/A170 was replacedwith the same restriction fragment from pGEM-3Zf’/BE containing the A to Cmutation.2.20.2. Mutation of the Third ATG in the 94 aa ORFThe third ATG codon in the 94 aa ORF was also changed to a CTG byoligonucleotide-mediated mutagenesis. The DNA sequence containing the49first mutation was used as the template for the second mutant oligo to createa new mutant with a single ATG codon at the 5? end of the ORF.The target sequence in pGEM-3Zf/BE was mutated in the same way asdescribed above. The Kpn I to Afi III fragment containing the two A to C pointmutations was used to replace the same fragment in pSVOd/M70.The pSELECT system was used to mutate the third ATG in the 94 aaORF encoded in the 518 nt cDNA. This system uses a second mutagenic oligoto confer amp resistance to the mutant DNA strand and was used according tothe method outlined by the supplier (Promega). The cDNA containing thefirst mutation was cloned into pSELECT-1 and single-stranded DNA wasisolated. Mutagenic oligo (1.25 pmol) and amp repair oligo (0.25 pmol) wereannealed to single-stranded template DNA (0.05 pmol) in a buffer containing20 mM Tris (pH 7.5), 10 mM MgC12,and 50 mM NaC1, by heating the mixtureat 820 for 5 mm then cooling slowly to RT. The DNA synthesis components;10 mM Tris (pH 7.5), 0.5 mM dNTPs, 1 mM ATP, 2 mM DTT, 2.4 U T4 DNApolymerase, and 2 U T4 DNA ligase, were assembled on ice and added to theannealing reaction. The reaction was incubated at 370 for 90 mm to performmutant strand synthesis and ligation.A 5 jil fraction of the synthesis reaction was used to transform themismatch repair minus E. coil strain BMH 71-18 mut S. The usualtransformation procedure was modified such that the cells were recovered for1 h in 4 ml YT at 370 after heat shocking and then amp (125 iig/ml) was addedand the incubation was continued overnight. The next day, plasmid DNAwas isolated from the mixture of amp resistant cells and this DNA was usedto transform DH5cx cells. Single colonies were isolated from the secondtransformation and plasmid DNA was prepared. Mutants were again50screened by sequencing and this method resulted in >60% of the clonescontaining the desired mutation.The 518 nt cDNA containing the two ATG to CTG mutations wascloned into pGEM-3Z1 for in vitro translation and cloned into pCMV5 for invivo expression in COS cells.2.20.3. Mutation of the First ATG in the 94 aa ORFThe first ATG, in a poor sequence context for translation initiation, waschanged to a CTG leaving the other ATG codons in the 94 aa ORF intact. ThepSELECT mutagenesis system was used to mutate the 518 nt cDNA asdescribed for the mutation of the third ATG. The new mutant cDNAsequence was cloned into pGEM-3Zf’ for in vitro translation and into pCMV5for COS cell expression.The same mutagenic oligonucleotide was used to mutate the B19sequence in pGEM-3Zf/BE after producing uracil-containing single-strandedtemplate DNA. The Kpn I to Af1 III fragment containing the mutant sequencewas then cloned into pSVOd/A170 as before.2.20.4. Creating a Stop Codon in the 94 aa ORFThe sequence AAA which occurs immediately Y to the third ATGcodon in the 94 aa ORF was changed to a TAA stop codon by changing an A toa T residue to create a mutant B19 genome which does not produce the 11 kDaproteins. The method used for mutagenesis was essentially that described formutation of the ATG codons in pGEM-3Zf’/BE. A Kpn I to Af1 III restrictionfragment containing the mutant sequence was used to replace the samefragment in pSVOd/A170.512.20.5. Mutation of an ACG to a ATG Codon in the 687 nt cDNAA translation initiation ATG codon was created at the 5’ end of the 142aa ORF encoded in the 687 nt cDNA by changing the C in the first ACG codonto a T residue. The 687 nt cDNA was cloned into pGEM-3ZF’ and uracilcontaining single-stranded DNA was prepared and used as a template formutagenesis by the method described previously. The mutant 687 nt cDNAwas cloned into pCMV5 for COS cell expression studies.2.21. Isolation of B19 Particles From Transfected COS CellsEight 100 mm plates of COS-7 cells were each transfected with 3 pgpSVOd/M70 using the DEAE-dextran method and the cells were harvested 72h later. After washing with ice cold PBS, the cells were scraped in 5 ml PBS,pelieted, and then resuspended in 8 ml PBS. The cells were lysed bysonication and cell debris was pelleted by centrifugation at 600 x g for 15 mmat 40 Two ml of the supernatant from the low speed spin were layered on topof 3 ml of 40% sucrose (w/v in PBS) and spun in a SW 50.1 rotor at 40,000 rpmfor 5 h at 40• The supernatant was discarded and the pellet was resuspended in1/10th vol (200 p1) PBS using a bath sonicator to achieve an even suspension.The samples were combined and layered onto 12 ml of a CsC1 solution of: 50mM Tris (pH 8.8), 5 mM EDTA, 0.1% sarcosyl, and 419.5 mg/mi CsCl. Thegradients were centrifuged in a Beckman SW 41 rotor at 28,000 rpm for 35 h at250. Fractions of 500 p1 were collected and the refractive indices (RI) weremeasured. Fractions with a RI corresponding to the density of empty particles(1.31 g/ml) and full particles (1.4 g/ml) were diluted in PBS and concentratedusing micro-concentrator tubes (Centricon 30,000). To demonstrate viralcapsids 10% of the concentrated sample was boiled in SDS sample buffer,52subjected to SDS-PAGE, transferred to NC, and immunoblotted with VP-2antisera.2.22. Electron Microscopy2.22.1. Direct EMA small sample of the CsC1 purified particle preparation was negativelystained with 2% (w/v) phosphotungstic acid (PTA), pH 7.4, on Formvarcarbon-coated copper grids and examined with a transmission electronmicroscope (Zeiss EM 1OC) at 60 kV.2.22.2. Immune EMImmune EM was used to agglutinate the B19 particles to facilitate theirvisualization in the electron microscope. B19 positive serum (50 il) wasmixed with 50 tl of B19 convalescent serum (clumping serum) and stored for1 h at RT. The sera were diluted to 3 ml with PBS and centrifuged at 9,500 rpmfor 1 h at 40 to pellet the immune complexes. The pellet was resuspended in20 p1 of PBS and a small sample was negatively stained with 2% PTA, pH 7.4.Grids were examined at 63,000X magnification using a Zeiss EM 1OCtransmission electron microscope at 60 kV.2.23. Expression of 11 kDa Proteins Fused to the Yeast GAL-4 DNABinding DomainPCR was used to amplify the coding region of the 518 nt cDNA and toadd restriction endonuclease sites at either end of the fragment for cloninginto the mammalian expression vector pM2. The 5’ oligo was derived fromnt 4710 to 4727 in the B19 sequence of pYT1O3 and contained a Barn HIrestriction site (refer to Appendix B for oligo sequence). The 3’ oligo annealed53to the sequence from nt 5013 to 4996 on the viral strand and had an Eco RIrestriction site at its 5’ end. The Nde I to Eco RI fragment from pYT1O3 (nt4680 - 5112) was gel-purified and used as the template for PCR. The reactionwas essentially that described for generation of the cDNA library in B19infected leukemic cells with the following differences: 100 pmol of eachprimer were added to 10 ng template DNA and 10% DM50 was included inthe reaction. The PCR program consisted of 29 cycles of: 940 for 1 mm, 500 for 1mm, and 720 for 1 mm; followed by 1 cycle of 940 for 1 mm, 500 for 1 mm, and720 for 10 mm. The amplified product was isolated on LMP agarose, digestedwith Barn HI and Eco RI, and cloned into the same site in pGEM-7Z. The Xba Ito BarnHI fragment from this clone containing the 11 kDa protein codingsequence was cloned into pM2 in frame with the yeast GAL-4 sequence (aa 1-147) such that a fusion protein consisting of the GAL-4 DNA-binding domainand the 11 kDa polypeptide was expressed in COS cells. This construct was cotransfected into COS cells with a reporter plasmid, pG5BCAT, whichcontained five GAL-4 binding sites upstream from a core promoter (ElbTATA box) driving the expression of a bacterial CAT gene (Martin et a!., 1990).CAT assays (Gorman et a!., 1982) were performed on cell lysates to determineif the 11 kDa protein had an activation domain.54RESULTS3.1. Expression of the Major B19 Proteins in COS-7 CellsCOS-7 cells transfected with SV4O-B19 hybrid vectors synthesized themajor B19 structural and nonstructural proteins (Figure 3). It was also shownthat COS cell expression of the B19 proteins was elevated in transfections withthe M70 clone compared with the near full length clone. The A170 sequencecontains an additional left-hand hairpin deletion from nt 43 to 172 inclusivein the B19 genome in pYT1O3. The capsid proteins VP-2 of 58 kDa (majorspecies) and VP-i of 83 kDa (minor species) were produced in COS cells inabout the same relative abundance as in B19 infected erythroid progenitorcells. Two B19 nonstructural proteins of about 71 kDa and 63 kDa weresynthesized in transfected COS cells in approximately equimolar amounts(Figure 3). Nonstructural polypeptides with the same mobility on an SDS gelas those made in COS cells were made from synthetic RNAs derived from theleft-hand side of the genome in in vitro translations in a cell free systemsuggesting that more than one protein is made from the same 2.3 kbtranscript (Figure 4).3.2. Replication of B19 in Human Leukemic CellsA primary culture of human bone marrow cells, previously isolatedfrom an individual with a CML, was shown to support a B19 infection.Monomer RF and single-stranded DNA were identified by Southern blottingof low molecular weight DNA using an RNA probe corresponding to nt 1 to3900 of the B19 genome in pYTIO3 (Figure 5). DNA isolated from a postadsorption, pre-replication sample was not detected on another Southern blotsuggesting that the virus is indeed replicated in these cells. Western blots of554 -Figure 3. Western blot analysis of B19 proteins expressed in COS-7 cellstransfected with SV4O-B19 hybrid vectors.B19-specific polypeptides were detected using rabbit anti-capsid (lanes 1-5) andrabbit anti-nonstructural sera (lanes 6-10). The detection system included asecondary antibody, goat anti-rabbit IgG conjugated to alkaline phosphatase,which was detected by the NBT/BCIP colour reaction. The markers were prestained protein standards. Lanes 1 and 10 contained lysates of COS cellstransfected with pSVOd vector DNA; lanes 2 and 9, pSVOd/B19A170; lanes 3and 8, pSVOd/Bl9wt; lanes 4 and 7, pSVOd/B19(A3)Xbw which is missing anA residue at nt 3940 and thus produces a truncated VP-2 protein of 27 kDa andalso contains a filled-in Xba I site in the left hand ORF encoding the NSproteins; and lanes 5 and 6, pSVOd/A207 which contains a Exolli deletionfrom nt 1-207 inclusive. The major 58 kDa VP-2 and minor 83 kDa VP-iproteins are indicated in lanes 2 and 3. Two nonstructural proteins of 71 and63 kDa were detected in lane 9 (Beard et a!., 1989).12 34583 kd5akda27 kd $106.789(kd)---. -66-’—./ -36-—‘j—29—.t200-97 -68-43 -29 -18- I56CzFigure 4. In vitro translation of T7 RNA polymerase-generated transcriptsfrom the B19 major left-hand ORERNA was translated in a rabbit reticulocyte lysate incorporating135S]met intonewly synthesized polypeptides. Proteins were separated by SDS-PAGE (4-12%) and the gel was dried and autoradiographed at RT. Lanes marked t327contain proteins synthesized in the lysate primed with RNA derived fromthe B19 left-hand ORF. The control lanes BMV and no RNA contain proteinstranslated from BMV RNA or translations without exogenous RNArespectively. The lane marked BMV&A327 contains proteins synthesizedfrom a mixture of BMV and A327 RNA. The marker lane (M) contains [14C]labeled high molecular weight protein standards. Two proteins ofapproximately 63 and 71 kDa that are absent in the control lanes are producedin the translations of RNAs derived from the B19 left-hand ORF (A327 RNA).These proteins were specifically immunoprecipitated with anti-NS proteinsera suggesting that they are B19 nonstructural proteins (data not shown).- - —1570000‘•0I—‘I23.1—9.4monomerRF—*I=6.6ssDNA—I-2.32.0—0.6—0.1Figure5.ReplicationofB19inhumanleukemiccells.AprimarycultureofhumanbonemarrowcellsisolatedfromapatientwithaCMLwasinfectedwith20jilofB19-containingserum(60iig/ml)!2Xi07cells.LowmolecularweightDNAwasisolatedbyHirtextraction48h.p.i.,electrophoresedona1%agarosegelandtransferredtoanylonmembranebyvacuumblotting.TheSouthernblotwashybridizedwithanRNAprobederivedfromnt1.-3900oftheB19genomeinpYT1O3.Autoradiographywasperformedat700withanintensifyingscreen.Thecontrollane(pYT1O3/EcoRI)containsthe5.1kbnearfull-lengthB19genomeinpYT1O3whichisreleasedfromthevectorsequencesafterdigestionwithEcoRI.MonomerRFandsingle-strandedDNAareindicated.58infected cell lysates detected immunoreactive proteins identified as VP-i (83kDa) and VP-2 (58 kDa) using an anti-capsid serum and as NS proteins usingan anti-nonstructural protein serum (Figure 6). In Bi9 infected CML bonemarrow lysates the major non-structural protein was estimated to beapproximately 71 kDa and at least one other lower molecular weight NSprotein was detected by western blotting of cell lysates (Figure 6). The reportedmolecular weights of the B19 nonstructural proteins are 77 kDa (majorspecies), 52 kDa and 34 kDa (minor species) in human bone marrow cellcultures (Ozawa et a!., 1987) and 71 kDa, 63 kDa, and 52 kDa in transpiacentallyinfected fetal liver tissue (Cotmore et a!., 1986). Cytoplasmic RNA from theseB19 infected CML marrow cells was used to construct a cDNA library.3.3. Screening the cDNA LibrariesSelected cDNAs from the B19 human leukemic cell cDNA library wereisolated, sequenced, and compared with cDNAs generated from a Bi9 COS cellcDNA library.3.3.1. Identification and Sequence of B19 Splice Junctions in cDNAs fromTransfected COS-7 CellsThe B19 COS cell cDNA library constructed by Caroline Beard(presently at Whitehead Institute, Boston, MA) was initially screened usingan RNA probe derived from nt 1 to 3900 of the B19 genome in pYT1O3 (referto Appendix A for a map of the restriction fragments used to make RNAprobes). Twenty-four B19 positive colonies were selected and the plasmidDNA from these cells was sequenced. Eleven of these clones contained intactB19 sequences and all were from the 500 and 600 nt size class of RNAsuggesting that this class of RNA is the most abundant species in transfected5914.3 -6.2 -M3 4Figure 6. Expression of B19 structuralleukemic bone marrow cells.and nonstructural proteins in humanProtein lysates from B19 infected CML bone marrow cells were fractionatedon 4-12% (lanes 1 and 2) or 4-15% polyacrylamide-SDS gels (lanes 3 and 4) andelectroblotted to nitrocellulose membranes. Specific B19 proteins weredetected with anti-capsid serum (lanes I and 2) or a rabbit anti-serumprepared by immunizing rabbits with a synthetic peptide (peptide 2)corresponding to an antigenic region within the left-hand ORF (lanes 3 and4). Lanes 1 and 3 contain proteins from mock infected cells; lanes 2 and 4contain proteins from B19 infected cells. The major 58 kDa VP-2 protein andthe minor 83 kDa VP-i protein are indicated in lane 2. Lane 4 shows themajor 71 kDa NS protein and at least two other stained bands of lowermolecular weight. The markers are pre-stained protein standards.Ml 297 -68-29 -14.3 -43 -29 -NS— VP-260COS cells. By sequence analysis this size class of RNA was made from threeexons. A 57 nt leader exon corresponding to nt 350 to 406 of the B19 genomein pYT1O3 was spliced to a second exon which exhibited variability in its spliceacceptor site; nt 1925, 1952, or 2030, and variability in its 5’ splice donor site; nt2172 or 2183. This exon was spliced to a third exon of 306 nt derived from nt4704 to 5010 (Figure 7).The library was further screened with an RNA probe corresponding tothe Ava I to Pst I fragment (nt 2080 to 3145) to detect the 700 to 800 nt size classof cDNA which is made from the small RNAs which terminate in the middleof the genome. Fifteen B19 positive clones were identified out ofapproximately 400 colonies. Sequencing of these clones revealed that allterminated at nt 5010 and were of the 500 to 600 nt size class. This probe wasnot specific for the 700 and 800 nt transcripts since it hybridizes to 153 nt of the518 nt cDNA. Because the 700 and 800 nt class of cDNA was not readilydetected using this probe this suggests that these transcripts are not abundantin transfected COS cells indicating that the variant polyadenylation signal atm.u. 49 may not be efficiently recognized in these cells.An RNA probe corresponding to the left-hand Hind III fragment (nt600 to 1540) was used to screen the library for NS cDNAs. After screeningapproximately 8000 colonies no NS cDNA clones were identified supportingthe suggestion that the variant middle polyadenylation signal may affectpolyadenylation or transcriptional termination in COS cells resulting in anunderrepresentation of the NS transcripts in the library.Using an RNA probe, made from the right-hand Pst I fragment (nt 3145to 3860) and specific for the capsid proteins, a cDNA clone corresponding to aVP-2 transcript was identified. This cDNA contained the same 57 nt leaderexon found in the 500 and 600 nt cDNAs spliced to a second exon derived61readingframe 23map units 0 20 40 60 80 100I I I I I I I I Inucieotides 0 1I00 21100 31100 41100 5000II’350406 2030 2183 oo 5184704350406 1952 217255350408 1952 2183 p010 596350 405 1925 2183 palo 623Figure 7. Splice sites identified in the 500 and 600 nt class of B19 transcriptsin transfected COS-7 cells.The B19-COS cell cDNA library was screened with an RNA probe derivedfrom nt 1-3900 of the B19 genome in pYT1O3 and cDNAs were sequenced asdescribed. cDNA (RNA) sequences are represented by thick lines and intronsby interruptions in these lines. The major open reading frames and their mappositions in the B19 genome are shown at the top of the figure. The cDNAscontain an ORF within their third exon in reading frame 2 indicated by theblack box. The second exon showed variability in splicing at both donor andacceptor sites. Since the 94 aa ORF is entirely within exon 3 the alternativesplicing pattern does not affect the size of the protein product(s).62from nt 2030 to 2183 and a third exon from nt 3045 to 5010. The longer cDNAscorresponding to the 3.0 kb VP-i transcripts (minor species) (see Figure 2)were not detected using this probe.In addition to variability in splice donor and splice acceptor sites thesecond exon in both the 500 to 600 nt cDNAs and the VP-2 cDNA contained anumber of single bp changes. There was a single A to G change at nt 2121, a Gto A substitution at nt 2164, and a G to A transversion at nt 2075 in three ofthe five small cDNAs that were completely sequenced. The VP-2 clonecontained a G to A change at nt 2125. In order to determine if these changeswere occurring at the DNA or RNA level replicated plasmid DNA wasisolated from transfected COS cells, digested with Dpn I to degrade inputDNA, and sequenced in the region of the second exon. The results suggestedthat, although B19 DNA was often deleted and rearranged when replicated inCOS cells, single base pair substitutions were not prevalent. The nucleotidechanges in the cDNAs may have resulted from first strand synthesis by MMLV reverse transcriptase which is more error-prone than E. coli DNApolymerase I since it lacks the 3’ to 5’ exonuclease activity which acts as anediting function (Gerard, 1983).3.3.2. Identification and Sequence of B19 Splice Junctions in cDNAs fromInfected Leukemic Bone Marrow CellsThe initial screening of the human leukemic cell B19 cDNA librarywith an RNA probe corresponding to the Eco RI to Barn HI fragment (nt 1 to3900) detected the 700 to 800 nt size class of cDNA exclusively, suggesting thattranscripts of this class are the most prevalent in B19 infected cells. In the firstscreening 21 B19-positive colonies were picked and the plasmid DNA fromthese clones was sequenced. Fifteen of the inserts were 807 nt and 6 were 68763nt in length. Both cDNAs contained the common 57 nt 5’ exon. This wasspliced to a second exon derived from nt 1910 to 2659 or nt 2030 to 2659(Figure 8). Since these transcripts terminate in the middle of the B19 genomeand are polyadenylated this confirms that a variant polyadenylation signal,ATTAAA or AATAAC, is utilized in human cells.The human B19 cDNA library was also screened with a probe derivedfrom the right-hand end of the genome (nt 3900 to 5110) to identify cDNAscorresponding to the 500 and 600 nt size class of RNA. These cDNAs werefound by sequencing to contain three exons: the 5’ leader sequence from nt350 to 406, the middle exon from nt 1910 or 2030 to nt 2183, and a third exonfrom nt 4704 to 5010. Two splice acceptor sites for the second exon result intwo sizes of transcript; 638 and 518 bases. The additional variability in splicejunctions in exon 2 of these cDNAs as demonstrated in the COS library wasnot observed in the human cell library. In addition, there were no single bpchanges detected in the second exon of the 500 and 600 nt cDNAs.3.4. Northern Hybridization Analysis of RNA from Transfected COS CellsSince there were no B19 cDNAs which corresponded to transcriptsterminating in the middle of the genome isolated from the COS cell librarythis suggested that these transcripts were either not made or that they werenot polyadenylated. To distinguish between these two possibilities RNA fromB19 transfected COS cells was analyzed by northern blotting. CytoplasmicRNA was isolated from COS cells transfected with pSVOd/B19M70 DNA,blotted to nylon membranes, and probed with RNA probes derived fromdifferent regions of the B19 genome to identify specific transcripts. Ahybridization probe corresponding to nt 1 to 3900 identified all nine B19transcripts. A second probe derived from nt 2435 to 2880 hybridized only to641reading 2frame3 V/4W/#AtW///4V/4W//4V/#/%%V/#/Amap 0 20 40 60 60 100units I I I I I I I I II I I Inucieotides 0 2000 3000 4000 5000nucieotides350 406 1910 2659807350 406 2030 2659687350406 iio 2183 1°638350406 20302183 4104518Figure 8. Sequence of the splice junctions of the small RNAs in ff19 infectedhuman leukemic cells.The B19-human cell cDNA library was screened with RNA probes derivedfrom either nt 1-3900 or nt 3900-5110 of the B19 sequence in pYT1O3. cDNAsderived from the 700 and 800 nt size class of RNA which terminate in themiddle of the genome were the most abundant species in the library. ThesecDNAs share an ORF with the NS protein(s) in their second exon as indicatedby the gray boN. Two splice acceptor sites, nt 1910 or nt 2030, and a single splicedonor site, nt 2183, were identified in exon 2 in both size classes of RNA.In the figure RNA sequences are represented by thick lines and intronsequences by interruptions in these lines. The major open reading frames andtheir map positions in the B19 genome are shown at the top.65VP-I (3.15 kb, 3.0 kb), NS (2.3 kb), and the 700 and 800 nt transcripts. The twoclasses of small RNAs ran together on the RNA gel; however, this secondprobe did clearly identify a band in the appropriate position suggesting thatthe 700 and 800 base transcripts are made in COS cells (Figure 9).3.5. Expression of the 518 and 638 nt RNAs3.5.1. In Vitro Expression of 518 and 638 nt RNAs in a RabbitReticulocyte LysateSP6 RNA polymerase was used to generate biologically active 518 and638 nt RNAs in vitro. The RNAs were translated in a rabbit reticulocyte lysateand newly synthesized proteins were labeled with [35S] met. The polypeptideswere separated by SDS-PAGE, electroblotted to NC, and immunoreactiveproteins were detected by colour reaction. Antisera, raised in two differentrabbits against a synthetic peptide (peptide 1) derived from an antigenicsequence encoded in the 94 aa ORE of the 500 and 600 nt cDNAs, were used todevelop the western blot. The results show that a protein of the appropriatesize of 11 kDa is synthesized from the 518 and 638 nt RNAs but is not made inthe absence of exogenous RNA and that both antisera bind the 11 kDa protein(Figure 10). The two RNAs direct the synthesis of the same protein since the94 aa ORE is entirely within the third exon which is invariant between thetwo transcript species.Proteins translated in vitro were immunoprecipitated with the twoantisera and separated on Laemmli gels. The results show that the rabbitantisera recognized more than one 11 kDa protein in the translations of 518and 638 nt RNAs (Figure 11).66A BpSVOd i\170-9.5-7.5- 4.4- 2.4- 1.4-0.24Figure 9. Northern hybridization analysis of RNA from B19 transfectedCOS-7 cells.Cytoplasmic RNA from COS-7 cells transfected in duplicate with eitherpSVOd vector DNA or pSVOd/B19A170 DNA was isolated 48 h posttransfection, fractionated through a formaldehyde agarose gel and blotted to anylon membrane. The same amount of RNA was loaded in each lane of thegel. The filter was cut in half and blot A was hybridized with a probe derivedfrom nt 1-3900 of the B19 sequence in pYT1O3 which identifies all B19transcripts. Blot B was hybridized with a specific probe derived from nt 2435-2880 which detects the VP-i transcripts (3.15 and 3.03 kb), NS transcript (2.3kb), and the 700 and 800 RNAs. Blot B shows that the 700 and 800 nt RNAs aremade in B19 transfected COS-7 cells.“1I—— .r -W ——18—14____ ____63Figure 10. Western blot analysis of in vitro transcribed and translated 500and 600 nt cDNAs.SP6 transcripts synthesized from the cloned cDNAs were translated in a rabbitreticulocyte lysate. Proteins were fractionated by SDS-PAGE (4-12%) andelectroblotted to nitrocellulose. A B19 specific 11 kDa protein was identified bybinding of rabbit antiserum generated by immunization with a syntheticpeptide (peptide 1) derived from antigenic sequences within the 94 a ORFcontained within the third exon of these small cDNAs. The lanes marked 518and 638 contained the translation products primed by the 518 nt and 638 ntRNAs respectively. The control lane contained the products of a translationreaction without added exogenous RNA. Protein standards were pre-stainedlow molecular weight markers. A component of the lysate (probablyhemaglobin) appears to affect the mobility of the 11 kDa protein, marked withan arrow, as evidenced by the curve of the 11 kDa bands.C-670Cj CL1 %D U Lr%DEkDa—43Figure 11. Immunoprecipitation of 1.1 kDa proteins translated in vitro from518 and 638 nt RNAs.Proteins labeled with[35S]met were selected by incubation with antisera raisedin two different rabbits against peptide 1 followed by adsorption of theimmune complexes with Protein A Sepharose. The immunoprecipitatedproducts were separated by SDS-PAGE (4-12%) and the dried gel wasautoradiographed at RT. The lanes marked 518 and 638 containimmunoreactive proteins translated from 518 nt and 638 nt RNAsrespectively. The arrow indicates the 11 kDa proteins. The control lanescontain proteins from translations with either no exogenous RNA or BMVRNA added to the lysate.68—o 0ci, cr 0 tti eL(J D çj In ‘.,D c-- —=kDa—43—29— 18— 14—6—369..:... ...3.5.2. COS Cell Expression of 11 kDa Proteins in pSVOdM170 and pCMV/518Transfected CellsCOS cells were transfected with pSVOd/A170 and the cells were lysed 48to 72 h later in SDS sample buffer. The A170 clone was chosen for this and allsubsequent COS cell expression studies since it had been previouslydemonstrated that this clone expressed B19 proteins to a higher level than didthe near full-length B19 clone from pYT1O3 (Beard et al., 1989). Proteins in thelysate were separated by SDS-PAGE, transferred to NC, and the western blotwas developed with the antiserum directed against peptide 1. The resultsclearly demonstrated that at least two 11 kDa immunoreactive proteins aresynthesized in COS cells transfected with pSVOd/A170 that are not made incells transfected with vector sequences alone (Figure 12).Proteins from transfected COS cells metabolically labeled for 3 h with[35S1 met were immunoprecipitated with peptide 1 antiserum and separatedon an SDS gel. The mobility of the 11 kDa proteins made in COS cells wascompared with that of the same polypeptides synthesized in vitro. The 11 kDaproteins appeared to migrate at the same molecular weight whethersynthesized in COS cells or in a reticulocyte lysate (Figure 13). These resultssuggested that the two forms did not result from post-translationalmodifications since these changes would not be expected in vitro.Western blots of cell lysates and immunoprecipitations of radiolabeledproteins from COS cells transfected with pCMV/518 or pCMV/638 (data notshown) produced the same result as that with pSVOd/A170 DNA. Thisconfirmed that the 11 kDa proteins were made from the sequence encoded inthe 518 nt and 638 nt cDNAs and that the other B19 proteins did not affect theexpression of these small proteins in COS cells.700 0r-—<a-4‘0 ‘0 r0 000> >cn cnkDa— 43— 29— 18—14—6—3Figure 12. Expression of 11 kDa proteins in transfected COS-7 cells.COS-7 cells were transfected with either pSVOd or pSVOd/B19M70 DNA.After 48 h cells were lysed in sample buffer and proteins were fractionated bySDS-PAGE (4-15%). Proteins were electroblotted to nitrocellulose and probedwith a rabbit antiserum raised against peptide 1. The arrow shows twoproteins of approximately 11 kDa expressed in A170 transfected cells.71< <a a—S •SC> > > >ci) ci) ci) ci) 0Figure 13. Immunoprecipitation of 11 kDa proteins expressed in vivo intransfected COS-7 cells and in vitro in a rabbit reticulocyte lysate.Lysates from transfected COS-7 cells, metabolically labeled with[35Sjmet, wereimmunoprecipitated with rabbit anti-il kDa serum (raised against peptide 1)and Protein A Sepharose as described in Materials and Methods. SP6transcripts of the 518 nt cDNA translated in a rabbit reticulocyte lysatecontaining[35Slmet were immunoprecipitated with the same antiserum. Theimmunoprecipitated products were separated by SDS-PAGE (4-15%) and thedried gel was autoradiographed at RT. The 11 kDa proteins synthesized inCOS-7 cells appear to have the same mobility as those produced in thereticulocyte lysate suggesting that the two forms do not arise from posttranslational modifications. The markers were pre-stained protein standards.The lane marked control was an in vitro translation with no exogenous RNAadded to the lysate and the BMV lane was a translation of BMV RNA.723.5.3. Expression of 11 kDa Proteins in B19 Infected Human Bone MarrowCellsThe 11 kDa proteins were also detected on western blots of cell lysatesfrom B19 infected human CML bone marrow cells (Figure 14) demonstratingthat these proteins are expressed in the natural host cell of the virus. At leasttwo immunoreactive proteins are made in these cells with the fastermigrating species appearing to be more abundant. Another western blot of thesame proteins was probed with a human B19 convalescent serum. Thisserum identified the major structural polypeptides but failed to detect the 11kDa proteins suggesting that the convalescent serum does not containantibodies to the 11 kDa proteins (data not shown).3.6. Characterization of Multiple Forms of the 11 kDa Protein3.6.1. Phosphatase Treatment of the 11 kDa ProteinsTo determine if either of the two 11 kDa proteins was phosphorylatedCOS cells transfected with pSVOd/A170 DNA were labeled with [35S] met andcell lysates were immunoprecipitated with antiserum to the 11 kDa proteins.Each sample was divided into two tubes and half of the tubes were treatedwith potato acid phosphatase (PAP) prior to SDS-PAGE. If the difference inmobility between the two 11 kDa proteins was due to phosphorylation onewould expect a shift in the mobility of one or both bands on the gel in thePAP-treated versus the control immunoprecipita tions. The results indicatedthat both the phosphatase treated and the untreated 11 kDa proteins migratedwith the same mobilities suggesting that the difference between the twoproteins was not due to phosphorylation (Figure 15).73M CB19kDa43-29-14-6-Figure 14. Expression of 11 kDa proteins in B19 infected human leukemicbone marrow cells.Western blot showing the expression of two 11 kDa proteins in B19 infectedcells. Infected (B19) and control (C) cells were harvested 48 h.p.i., lysed insample buffer, and fractionated by SDS-PAGE (4-15%). The proteins wereelectroblotted to NC and the 11 kDa proteins were detected using an anti-ilkDa serum (anti-peptide 1 serum) followed by binding with AP-conjugatedsecondary antibodies. Reactive proteins were stained by the NBT/BCIP colourreaction. The marker lane (M) contained low molecular weight pre-stainedprotein standards.74432918.414.36.2Figure 15. Digestion of the 11 kDa proteins with potato acid phosphatase doesnot affect their mobility by SDS-PAGE.COS-7 cells were transfected with pSVOd or pSVOd/B19A170 DNA and newlysynthesized proteins were labeled with[35S]met. The 11 kDa proteins wereimmunoprecipitated with anti-li kDa serum and digested with PAP. Theproducts were separated by SDS-PAGE (4-15%) and the dried gel was subjectedto autoradiography. Markers were pre-stained low molecular weight proteinstandards. The mobility of the ii kDa proteins was not altered in thephosphatase-treated (A170/PAP) versus the undigested (A170) protein samplessuggesting that neither of the small polypeptides is phosphorylated.753.6.2. Mutagenesis of Translational Initiation ATG Codons at the 5’ End ofthe 94 aa ORFThe 5’ end of the 94 aa ORF in the 518 and 638 nt cDNAs contains threeclosely spaced ATG codons (refer to Appendix C for the complete sequence ofboth cDNAs: the three ATGs are underlined). Since the data showed thatmore than one 11 kDa protein was translated from this ORF a mutationalanalysis was performed to determine if translation was initiated at more thanone of these ATG codons. This method was chosen instead of amino terminalsequencing since the NH2 terminal is usually blocked in proteins expressed inmammalian cells thereby necessitating the purification of ig quantities ofprotein. Expression of the Second ATG Codon MutantThe second ATG codon from the 5’ end of the 94 aa ORF is in a strongsequence context for translation initiation (Kozak, 1986) (Figure 16). Mutantplasmids which contained a CTG in place of the second ATG codon, whenexpressed in vitro, still directed the synthesis of two 11 kDa proteins althoughthe level of expression of the higher mobility species was reduced. The sameresult was found in vivo in COS cells transfected with mutant pSVOd/M70or pCMV/518 plasmids (Figure 17). Expression of the Second and Third ATG Codon MutantThe third ATG codon in the 94 aa ORF is 3 nt downstream from thesecond ATG and this ATG is also in a good consensus sequence for initiationof translation. To determine if translation was initiating from this codon itwas also mutated to a CTG codon. The results of expression studies in vitroand in COS cells suggested that although synthesis of the higher mobility 1176-3 +41) ACTCTACAGATGC2) ACCACAGACATGG3) GACATGGATATGAKozak GCCGCCCCATGGl8nt 3ntwt:-- ATG ATG --- ATG --muti: --ATG CTG---ATG-mut 2:-- ATG CTG--- CTG --mut3:-- CTG ATG---ATG-Figure 16. The sequence context of the three ATG codons at the 51 end ofthe 94 aa ORF in the third exon of the 518 and 638 nt cDNAs compared withthe Kozak consensus sequence for translational initiation and the 5’ endsequences of three ATG to CTG mutants.The +4 and -3 nucleotide positions relative to the ATG triplet are critical fortranslation initiation at a given initiation codon.77432918146.23Figure 17. COS-7 cell expression of ATG to CTG mutants.Immunoprecipitation of [35S]met labeled proteins from COS-7 cellstransfected in duplicate with pSVOd (lanes 1 and 2), pSVOd/B19A170 (lanes 3and 4), pSVOd/B19M70 muti (lanes 5 and 6), and pSVOd/Bi9zi7O mut2DNA (lanes 7 and 8) and immunoprecipitated with anti-li kDa serum. Theexpression of the higher mobility ii kDa protein was reduced when thesecond ATG in the 94 aa ORF was mutated to an CTG and further decreasedin the double mutant where both the second and third ATGs have beenmutated to CTGs. A low level of expression from a non-ATG codon isapparent in the double mutant (lanes 7 and 8). The markers were pre-stainedlow molecular weight protein standards.12 34567878kDa protein was diminished in the double mutant there was still someexpression presumably due to translation initiation at a non-ATG codon(Figure 17). Expression of the First ATG Codon MutantThe first ATG codon in the 94 aa ORF is found in a poor Kozakconsensus sequence for translation initiation. This codon was changed to aCTG codon leaving the other two 5’ ATG codons unaltered. Expression of SP6generated transcripts of this mutant DNA in vitro in a reticulocyte lysateshowed that by SDS-PAGE the lower mobility species of 11 kDa proteindisappeared suggesting that the first AUG, although in a poor Kozak sequencecontext, is used for initiation of translation. Also, these results confirm thattranslation does initiate downstream from this first AUG triplet (Figure 18).The identical result was observed in vivo in transfected COS cells (Figure 19)confirming that the two bands, separable by SDS-PAGE, arise fromtranslational initiation of the ii kDa proteins at more than one codon.3.7. Expression of B19 Structural and Nonstructural Proteins in theAbsence of the 11 kDa ProteinsTo determine if the 11 kDa proteins affect the expression of the majorstructural and nonstructural proteins a null mutant was constructed byinserting a stop codon immediately 3’ to the third ATG in the 94 aa ORF inthe plasmid pSVOd/A170. Transfected COS cells were metabolically labeledwith [355] met and the harvested proteins were immunoprecipitated witheither anti-li kDa protein serum, anti-capsid serum or anti-NS proteinserum. The results indicated that the ii kDa proteins are not synthesized in791 2 3 4 5 643 -29 -18.4 -3-Figure 18. Immunoprecipitation of proteins translated in vitro from wildtype and mutant 518 nt RNAs.SF6 RNA polymerase-generated transcripts were translated in vitro in a rabbitreticulocyte lysate incorporating[35Slmet into newly synthesized polypeptides.The 11 kDa B19 proteins were immunoprecipitated with anti-li kDa serumand separated by SDS-PAGE (4-15%). Lane 2 shows immunoreactive proteinsfrom the translation of the wild type 518 nt RNA, lane 3 is translated 518muti RNA (2nd ATG mutated), lane 4 is translated 518 mut2 RNA (2nd and3rd ATGs mutated), and lane 5 represents the proteins translated from 518mut3 RNA (1st ATG mutated). Lanes 1 and 6 contain immunoprecipitatedproducts from the translations of BMV RNA and the lysate withoutexogenous RNA, respectively. The markers were pre-stained low molecularweight protein standards.80Figure 19. Summary of the 11 kDa proteins expressed in COS-7 cells aftertransfection with wild type and mutant pCMV/518 DNA.Transfected COS-7 cells were metabolically labeled with [35S]met andimmunoreactive proteins were precipitated with anti-li kDa serum. Thetransfections were performed in duplicate and the transfecting DNA isindicated above the lanes on the gel. WT is pCMV/518 wild type DNA. Mut 1DNA contains a mutation in the 2nd ATG in the 94 aa ORF. Mut 2 DNAcontains mutations in both the 2nd and 3rd ATGs in the 94 aa ORE and mut 3DNA contains a mutation in the 1st ATG codon in the 94 aa ORF. Themarkers were pre-stained low molecular weight protein standards.pCMV wt muti mut2 mut31 I I I I I I I81cells transfected with the stop mutant DNA (Figure 20) and that theexpression of the major structural and nonstructural proteins in COS cells isnot affected in the absence of the 11 kDa proteins (Figure 21). This experimentwas repeated using synchronized COS cells with the same result (data notshown).3.8. Expression of the 687 and 807 nt RNAs3.8.1. In Vitro Translation of 687 and 807 nt RNA in a Rabbit ReticulocyteLysateSP6-generated transcripts from the 687 and 807 nt cDNAs weretranslated in a cell free system incorporating [35S] cys to label newlysynthesized proteins. (The sequences of the 687 and 807 nt cDNAs are givenin Appendix C). The large ORF contained within these cDNAs does notcontain any ATG translation initiation methionine codons so met cannot beused to label the putative protein. Immunoprecipitation of in vitro translatedpolypep tides synthesized from the 687 nt RNA using rabbit antiserum raisedagainst a synthetic peptide (peptide 2) whose sequence is within this ORFsuggested that a 15 kDa protein could be detected at extremely low levels(Figure 22, lane 687).3.8.2. COS Cell Expression of a Putative 15 kDa Protein in pSVOd/A170 andpCMV/687 Transfected CellsRadiolabeled proteins from COS cells transfected with pSVOd/A170were immunoprecipitated with antiserum raised against peptide 2 whosesequence was derived from antigenic regions contained within the largestORF of the 687 and 807 nt cDNAs. Since the same ORF is used to make the NSproteins the immune sera was predicted to also detect the NS proteins. The82pSVOd z17O 11 kDa—------‘ ——-—4_—-----, . —43 -29 -18 -14 -6- —3-Figure 20. The 11 kDa proteins are not made in COS-7 cells transfected withpSVOd/A170/11 kDa DNA.COS-7 cells were transfected in duplicate with either pSVOd, pSVOd/M70, orpSVOd/z17O/11 kDa DNA. Forty-five hours after transfection the cells weremetabolically labeled with [35S]met and immunoreactive proteins wereimmunoprecipitated from cell lysates using anti-fl kDa serum (raised againstpeptide 1). The proteins were resolved by SDS-PAGE (4-15%) and the dried gelwas autoradiographed at RT. The markers were pre-stained low molecularweight protein standards. Two 11 kDa proteins only appear in the lanesderived from pSVOd/A170 transfections (A170 lanes) showing that theintroduction of a stop codon (TAA) in the 94 aa ORF prevents the synthesis ofthe 11 kDa polypeplides in the pSVOd/A170/11 kDa- mutant (11 kDa- lanes).830 Io e 0> F.. > F..U) I- Q) IMI II II II II—II I, -‘-‘ t’ a . i. -20068.42918.4 *I II INS capsidantiseraFigure 21. COS-7 cell expression of B19 structural and nonstructural proteinsin the absence of the 11 kDa proteins.COS-7 cells were transfected in duplicate with pSVOd vector DNA,pSVOd/B19A170 DNA, or pSVOd/B19A170 11 kDa- DNA and metabolicallylabeled with[35S]met after 48 h. B19 nonstructural proteins were detected byimmunoprecipitation of cell lysates with antiserum raised against peptide 2and B19 capsid proteins were detected in the same way by binding to an anticapsid serum. The proteins were fractionated by SDS-PAGE (4-12%) and thedried gel was autoradiographed at RT. The markers were pre-stained highmolecular weight protein standards. The gel indicates that the levels of the 71and 63 kDa NS proteins and the 58 kDa capsid protein are the same with orwithout the 11 kDa proteins. The minor capsid protein, VP-i, was notdetected in this experiment.84zEz .E >1’ 0kDa43-29-18.4-14.3-6.2-3-_________________Figure 22. Expression of wild type and ACG to AUG mutant 687 nt RNAs ina cell free system.SP6 RNA polymerase-generated transcripts were translated in a rabbitreticulocyte lysate containing [35S] cys to label newly synthesized polypeptides.The putative 15 kDa protein was immunoprecipitated with antiserum raisedagainst peptide 2 and the products were fractionated by SDS-PAGE (4-15%). Afaint immunoreactive band can be detected in the wild type lane (687) while astrong band is present at —15 kDa in the AUG mutant lanes (687mut andBMV&687mut). The control lanes (BMV or no RNA) are indicated. Themarkers were pre-stained low molecular weight protein standards.85results indicate that the antisera precipitated two NS proteins of 71 and 63 kDabut failed to detect a 15 kDa protein (Figure 23). If translation initiates from anon-ATG codon the level of protein expression may be too low to bedetectable using these methods. Western blotting and immunoprecipitationof proteins expressed from pCMV/687 plasmids transfected into COS-7 cellsalso failed to identify an immunoreactive 15 kDa protein.3.8.3. Expression of Putative 15 kDa Protein in B19 Infected Human BoneMarrow CellsWestern blots of proteins from B19 infected leukemic cells weredeveloped with a rabbit serum raised against peptide 2 or with a humanconvalescent serum against a B19 viral infection. The rabbit antiserumrecognized a major NS protein of approximately 71 kDa in addition to one ortwo other lower molecular weight polypeptides but did not detect a smaller 15kDa protein (Figure 6). The convalescent serum did not detect either the NSor the putative small 15 kDa polypeptide (data not shown).3.9. Expression of 687 nt cDNA Containing an ACG to ATG MutationIn an attempt to determine if the 687 nt RNA was translatable an ACGcodon (underlined in Appendix C) at the 5’ end of the large ORF in the 687 ntcDNA was changed to an ATG codon. SF6 polymerase RNAs transcribed fromthe mutant cDNA and translated in a reticulocyte lysate directed theproduction of an immunoreactive protein of the appropriate molecularweight of 15 kDa as well as a minor band (‘—16 kDa) of slightly slower mobility.Immunoprecipitation of radiolabeled proteins synthesized in the cell-freesystem with peptide 2 antiserum also detected a protein of the samemolecular weight (‘--15 kDa) synthesized from the wild type 687 nt RNA. This86/.1iix-NSFigure 23. Immunoprecipitation failed to detect the putative 15 kDa protein.COS-7 cells were transfected in duplicate with pCMV vector (lanes pCMV),pCMV/687 (lanes 687), pCMV/687mut (lanes 687mut), pSVOd vector (lanespSVOd), or pSVOd/B19A170 DNA (lanes A170), and metabolically labeledwith [35S] cys 45 hours after transfection. Protein lysates wereimmunoprecipitated with either anti-15 kDa serum (raised against peptide 2)or anti-nonstructural protein serum (as indicated below the gel), and theproducts were separated by SDS-PAGE (4-15%). Both antisera detected the 71and 63 kDa NS proteins in the A170 lanes (marked with an arrow) which aretranslated in the same reading frame as the putative 15 kDa protein but theanti-15 kDa protein serum failed to detect a B19 specific 15 kDa polypeptide(lanes 687, 687mut, and A170). The markers were pre-stained low molecularweight protein standards.kJ43.29-18.4—14.3 -6.2 -3-. F’ ø —-ii:—- NSa-iS kDa87protein was of much lower abundance than was the mutant form and couldonly be detected when the autoradiogram was overexposed (Figure 22).COS cells were transfected with a pCMV/687 clone which contained theACG to ATG mutation. Western blots of cell lysates andimmunoprecipitations of radiolabeled proteins with antisera against peptide 2failed to convincingly identify an immunoreactive 15 kDa proteinsynthesized in these cells (data not shown). The 15 kDa polypeptidé madefrom this ORF may be of low abundance or high instability and therefore notreadily detectable by these methods.3.10. Immunofluorescence3.10.1. Localization of 11 kDa Proteins in Transfected COS CellsIndirect immunofluorescence was used to localize the 11 kDa proteinsto the cytoplasm of transfected COS cells (Figure 24). The localization was thesame whether the cells were transfected with pSVOd/i17O or pCMV/518DNA. Control cells, transfected with vector sequences alone, did not fluorescedemonstrating that the signal was specific for the 11 kDa proteins.3.10.2. Localization of NS Proteins in Transfected COS CellsAntiserum raised against peptide 2 was used to localize the NS proteinsin pSVOd/A170 transfected COS cells. The unexpected result was that theseproteins also localized to the cytoplasmic compartment of transfected COScells (Figure 25). Previous studies have reported that the major B19 NSprotein is found in the nucleus of infected cells (Cotmore et. al., 1986; Ozawaand Young, 1987). In addition, the conserved function of the parvovirus NSproteins in viral DNA replication requires a nuclear localization for at leastone of the NS proteins. These results suggested that the localization of the NS88A BFigure 24. Localization of the 11 kDa proteins to the cytoplasm of transfectedCOS-7 cells by indirect immunofluorescence.COS-7 cells transfected with pSVOd/Bl9M7O DNA on 25 mm coverslips werefixed 48 h after transfection in 3.7% paraformaldehyde and the ii kDaproteins were detected using a rabbit anti-li kDa serum directed againstpeptide 1 followed by binding with a biotinylated secondary antibody andstreptavidin-Texas Red as described in Materials and Methods. Photographywas performed on a Zeiss Universal photomicroscope using KodakEktachrome slide film (400 ASA) with a 40X neofluar objective. Figure A isthe bright field image and Figure B is immunofluorescence. Control plates ofcells transfected with pSVOd vector DNA did not produce a fluorescentsignal. COS-7 cells transfected with pCMV/518 DNA produced the same resultas with pSVOd/B19zi7O DNA. In both cases the ii kDa proteins appear to becytoplasmic.89A BFigure 25. Localization of the NS proteins to the cytoplasm of transfectedCOS-7 cells by indirect immunofluorescence.COS-7 cells growing on 25 mm coverslips were transfected with pSVOd/B19A170 DNA and fixed 48 h after transfection in 3.7% paraformaldehyde asdescribed. The NS proteins were detected using an antiserum directed againstpeptide 2 followed by binding with a secondary antibody conjugated torhodamine as described in the Materials and Methods section. Photographywas performed on a Zeiss Universal photomicroscope using a 63X neofluarobjective. Figure A is the bright field image and Figure B is the fluorescentimage. Another antiserum which is directed against the NS proteinsproduced the same result, i.e., localization of the nonstructural proteins to theCOS cell cytoplasm.90proteins in COS cells was aberrant and therefore the localization of the 11 kDaproteins in these cells may also be incorrect.3.10.3. Localization of 11 kDa Proteins in B19 Infected Human LeukemicCellsB19 infected human leukemic bone marrow cells growing insuspension cultures were attached to microscopic slides by using a cytospin at48 h.p.i. Indirect immunofluorescence using an antiserum raised againstpeptide 1 detected an immunoreactive protein that was present in both thenucleus and the cytoplasm (Figure 26). The number of cells which exhibitedthe signal was very low; approximately 1 in 100 cells was clearly positive. Thecultures contained cells with the characteristic features of B19 infectionincluding cells with marginated chromatin and giant pronormalblast cells.A second experiment was performed to determine the time afterinfection when the 11 kDa protein could be detected by this method. Atintervals of 12 hours, starting 24 h.p.i., infected CML bone marrow cells wereattached to polylysine-coated coverslips and fixed with paraformaldehyde.The final time point was taken at 72 h.p.i. The cells were incubated withantiserum and the 11 kDa proteins were detected by indirectimmunofluorescence. The results showed that at 36 h.p.i. a small number ofB19 infected cells weakly expressed the 11 kDa proteins. The signal wasstronger and more cells were positive at 48 h.p.i. At later time points theculture contained many cells whose nuclei appeared to have broken downand it was difficult to assign the fluorescent signal to a cellular compartment.91A BFigure 26. Localization of 11 kDa proteins in B19 infected human leukemiccells.Infected and control cells were attached to microscopic slides using a cytospin.The cells were fixed and permeabilized and the 11 kDa proteins were stainedwith Texas Red as described. Photography was performed on a ZeissUniversal photomicroscope using Kodak Ektachrome slide film (400 ASA)with a 100X neofluar objective. Figure A is the bright field image and Figure Bis immunofluorescence. The 11 kDa proteins are partially nuclear in thisassay. Control mock-infected cells did not exhibit a fluorescent signal.923.10.4. Localization of NS Proteins in B19 Infected Human Leukemic CellsForty-eight h.p.i. B19 infected CML bone marrow cells were attached tomicroscopic slides using a cytospin and the cells were stained forimmunofluorescence. The fluorescent NS protein signal was detected both inthe nucleus and the cytoplasm of the infected cells (Figure 27). The nuclearsignal appeared to be stronger than the cytoplasmic signal and also strongerthan the nuclear signal of the 11 kDa proteins in the same cells (compareFigure 27 with Figure 26). The results indicate that at least one of the two NSproteins is entirely or partially localized in the nucleus of B19 infected cells.This is consistent with the predicted function of the NS proteins in viralDNA replication.3.10.5. Immunofluorescence of the Putative 15 kDa Protein in COS CellsBecause I had not been able to detect convincingly the putative 15 kDaprotein expressed in transfected COS cells or B19 infected CML cells usingeither immunoprecipitation or western blots, it was decided to use in situimmunofluorescence which may be a more sensitive technique. Hence, COScells were transfected with the plasmids pCMV/687 and pCMV/687 mutantDNA containing the ATG methionine codon replacing an ACG threoninecodon. The cells were fixed 48 h after transfection and the putative 15 kDaprotein was stained using antiserum raised against peptide 2 and a secondaryantibody conjugated to Texas Red. This technique also failed to detect eitherthe wild type or mutant 15 kDa protein. It was not possible to use the peptide2 antiserum to look for immunofluorescence of the 15 kDa protein in B19infected CML human bone marrow cells since this serum also reacts with theNS proteins.93A BFigure 27. Localization of NS proteins in B19 infected human leukemic cells.Infected and control cells were attached to microscopic slides using a cytospinand fixed and stained as described. The primary antiserum was a rabbit serumraised against peptide 2 and the fluorophore was Texas-Red. Photography wasperformed as described in Figure 26. Figure A is the bright field and Figure Bis the immunofluorescent image. The NS protein signal is mostly nuclear inthese cells. There were no individual cells which fluoresced brightly in themock-infected culture.943.11. Identification of B19 Particles in Transfected COS Cells3.11.1. Identification of B19 Proteins by Western BlottingSince B19 virus cannot be propagated in vitro the availability ofantigen for diagnostics and vaccine development is limited. Viral antigen isnormally obtained from viremic individuals. However, in most cases,patients who exhibit symptoms of B19 disease no longer have high titers ofvirus in their serum (Cohen, 1988). The COS cell system was investigated as aconvenient source of 819 antigen. Western blot analysis had shown that COScells transfected with pSVOd/B19 vectors produced both VP-i and VP-2 of thecorrect sizes and in the appropriate abundances. Studies were undertaken todetermine if these viral proteins could self-assemble into particles.Capsids were purified from COS cells transfected with pSVOd/A170DNA. Fractions of the appropriate density on a CsC1 gradient were diluted inPBS, the proteins were concentrated into a smaller volume, and thenresolved by SDS-PAGE. The results showed that a 58 kDa protein was presentin a CsC1 fraction with the density expected for 819 empty particles. Theprotein was identified as VP-2 by western blotting using an anti-capsid serum(data not shown). In addition, the minor VP-i protein was not detected in thispreparation. Since the amount of VP-i may be only 4% that of VP-2 theminor capsid protein may not be detected by this method. To determine if theii kDa protein was also present in this particle preparation a western blot ofthis capsid fraction was probed with anti-il kDa serum. These studies failed todetect this small protein suggesting that these proteins are unlikely tocontribute to the Bi9 capsid structure or, like VP-i, they are such a minorcomponent that they cannot be detected by this method.953.11.2. Visualization of B19 Particles after Negative Staining byTransmission EMThree tl (3%) of the concentrated particle fraction was stained with anequal volume of 2% PTA, pH 7.4, and the dried grid was examined at 50,000Xmagnification by transmission EM. Both isolated particles and small clumpsof particles were readily detected in the microscope. The particles appeared tohave the appropriate shape and were the expected size of parvovirus particles(Figure 28). This result confirms that B19 viral proteins produced in COS cellscan form particles that are morphologically very similar to native particles(Figure 29). The native particles in Figure 29 were agglutinated with immunesera so they form large clusters. The spike-like features which are visiblebetween the B19 particles are antibody molecules from the convalescentserum.Capsids from COS cells transfected with a pSVOd/M70 11 kDa clonewhich did not express the 11 kDa proteins, still produced particles providingfurther evidence that the 11 kDa protein is unlikely to be a structural protein(Cohen et a!., 1991).3.12. Searching for Sequence SimilaritiesThe sequence in the 11 kDa proteins was compared with sequences inGenBank, EMBL, Pirform, and Swiss-Prot using the FASTA program (Pearsonand Lipman, 1988) at a ktup of 2. There were no sequence similarities ofsignificance identified by this method, and hence, no hints as to the functionof the small proteins.96Figure 28. Direct electron microscopy of B19 particles made in transfectedCOS-7 cells.B19 particles purified by CsC1 density gradient centrifugation were negativelystained with 2% PTA (pH 7.4) on Formvar carbon-coated grids and visualizedwith a Zeiss EM 1OC transmission electron microscope at 60 kV at amagnification of 50,000X.97Figure 29. Immune electron microscopy of B19 parvovirus particlesaggregated with a B19 convalescent serum.The virions were negatively stained with 2% PTA (pH 7.4) on Formvarcarbon-coated grids and visualized using a Zeiss EM 1OC transmissionelectron microscope at 60 kV at a magnification of 63,000X. The spike-likeprojections surrounding the particles are antibody molecules from theclumping serum.983.13. Expression of 11 kDa Protein Fused to the Yeast GAL-4 DNA-BindingDomainOne interesting feature of the 11 kDa protein is the presence of 14 prolineresidues clustered within a 40 aa linear sequence in the protein (Figure 30).Proline rich regions of certain proteins have been implicated intranscriptional regulatory activities (Mermod et a!., 1989; Mitchell and Tjian,1989; Madden et al., 1991). To test if the 11 kDa protein was an activator thecoding region was fused in frame to the DNA binding domain of the yeastGAL-4 protein in the plasmid pM2. This DNA was transfected into COS cellsalong with a reporter plasmid, pG5BCAT, containing five GAL-4 DNAbinding sites upstream from a minimal promoter (Elb TATA box) driving theexpression of a bacterial chioramphenicol acetyl transferase (CAT) gene(Martin et a!., 1990). CAT assays performed on cellular extracts from thesetransfections demonstrated that the 11 kDa protein does not appear to havean activation activity in COS cells (Figure 31).99I 10 20 I 30 I 40 I 50MQNNTTDMDM KSLKNCGQPK AVCTHCKHSP PCPQPGCVTK RPPVPPRLYLPPPVPIRQPN TKDIDNVEFK YLTRYEQHVI RNLRLCNMYQ NLEKZI 60 I 70 I 80 I 90 I 94Figure 30. Amino acid sequence of the 11 kDa proteins.The proline-rich region is underlined. The smaller protein species ismissing seven amino acids from the NH2 terminal.100pSG pM2-1 1 pSG-NS1 pSG-VP• •.. •4. .• .•• •• •• ••. a S • S SFigure 31. The 11 kDa protein lacks an activation domain.COS-7 cells were cotransfected in duplicate with 2 tg of activator DNA (eitherpSG, pM2-11, pSG-NS1, or pSG-VP) and 2 ig of reporter DNA (pG5BCAT).The cells were lysed 48 hours after transfection and CAT activity wasdetermined using standard methods (Gorman et at., 1982). CM ischloramphenicol and A and B are its mono-acetylated derivatives, 1-acetateand 3-acetate chloramphenicol, respectively. C indicates the diacetylatedspecies, 1,3-diacetate chloramphenicol. pM2-11 DNA encodes a 29 kDa fusionprotein containing the yeast GAL-4 DNA-binding domain (aa 1-147) fused tothe full-length 11 kDa protein (94 aa). pSG-VP DNA encodes a fusion proteinconsisting of the yeast GAL-4 DNA binding domain and the herpes simplexvirus (HSV) VP-16 activation domain and was the positive control in thisassay. pSG-NS1 encodes a fusion protein containing the same yeast GAL-4DNA-binding domain fused to aa 541-672 of the NS-1 protein of MVM. Thisregion of MVM NS-1 also encodes an activator by this assay. The negativecontrol, pSG containing the GAL-4 DNA binding domain alone, does exhibita low level of activation activity. This is a consistent result with theseplasmids (Ivan Sadowski, personal communication). The chromatographsuggests that the 11 kDa protein does not have an activation domain (lanespM2-11).The pSG and pM2 vectors share essentially the same DNA sequencesexcept that pM2 has a unique Eco RI site and there are other minor differencesin the multiple cloning sites.101DISCUSSION4.1. B19 Gene Expression in Transfected COS-7 Cells and B19 InfectedHuman Chronic Myelogenous Leukemia CellsIn the studies described in this thesis I have used two experimentalsystems to analyze gene expression of B19 parvovirus; transfected COS-7 cellsand B19 infected human CML bone marrow cells. When SV4O-B19 hybridvectors are transfected into COS cells, the DNA is amplified to high copynumber, is transcribed into RNAs corresponding to the same size classesobserved for B19 infected human bone marrow cells, and the RNAs aretranslated into the major structural and nonstructural proteins (Beard et at.,1989; Figure 3). Similarly, B19 virus infected into CML cells undergoesreplication [detected by the identification of monomer RF and progeny viralstrands (Figure 5)1, the DNA is transcribed into the known RNA size classes,and the structural and nonstructural proteins are synthesized.The two capsid proteins of 83 kDa (minor species) and 58 kDa (majorspecies) are made in transfected COS cells (Figure 3) and B19 infected CMLcells (Figure 6). In transfected COS cells the viral structural proteins wereshown to assemble into particles which by EM are morphologically verysimilar to those produced in B19 infected erythroid progenitor cells (compareFigures 28 and 29). We do not know if infectious mature viral particles areassembled in B19 infected CML cells, but presumably they are.The nonstructural proteins expressed by transfected COS cells migrateas 71 kDa and 63 kDa species (Figure 3). These proteins are made inapproximately equimolar amounts, presumably using a single 2.3 kb mRNAtranscript (Figure 2). This assumption is based on the observation that invitro an unspliced RNA transcript corresponding to the left half of the102genome is translated into two polypeptides of 71 and 63 kDa (Figure 4).Although most eukaryotic mRNAs are monocistronic, a number oftranscripts (predominantly in viral systems) use more than one initiation sitefor protein synthesis (Kozak, 1986a). In B19 infected CML cells, proteins whichreact with anti-NS sera migrate with somewhat different molecular weights;—71 kDa (major species), —50 kDa, —38 kDa (doublet), and —35 kDa (Figure 6).The sizes of the B19 nonstructural proteins have been reported by others to be71 and 63 kDa (major species) and 52 kDa (minor species) in liver tissue froma transpiacentally infected fetus (Cotmore et a!., 1986), and 77 kDa (majorspecies), 52 kDa and 34 kDa (minor species) in erythroid bone marrow cellcultures (Ozawa and Young, 1987). The discrepancy between the number andsizes of these NS proteins may be due to posttranslational modification(cleavage ?) in different permissive cells (erythroid precursors in fetal liver,bone marrow, and CML cells) or normal degradation of these proteins in thedifferent cell preparations.Of particular interest in B19 gene expression is the syrtthesis of twoclasses of small, abundant RNAs which are unique in that no other vertebrateparvovirus has been shown to make RNAs of these size classes (700-800 ntand 500-600 nt). The 700 and 800 nt RNAs have been mapped using Sinuclease to the left half of the B19 genome whereas the 500 and 600 nt RNAscorrespond to three exons which are derived from the extreme left end,middle, and right end of the genome (Ozawa et a!., 1987; Figure 2). The focusof this study was to prepare cDNA libraries from transfected COS-7 cells andB19 infected CML cells and determine the translation potential of these smallRNAs.1034.2. Presence of the 700-800 nt and 500-600 nt cDNAs in the Two cDNALibrariesThe most abundant species of cDNA isolated from the B19 humanCML cell library were those of the 700 and 800 nt size class (Figure 2). Sincethis cDNA library was constructed by PCR one would expect the smallertranscripts to be preferentially amplified, hence the library may not representtrue RNA abundances. However, the relatively high abundance of the 700and 800 nt cDNAs in the library is consistent with northern hybridizationanalyses which suggested that the 700 and 800 base transcripts are the mostprevalent species in B19 infected cells followed by the 500 and 600 nt RNAs(Ozawa et al., 1987).On the other hand, the 700 and 800 nt cDNAs were not detected in thetransfected COS cell cDNA library suggesting that the corresponding RNAsare either not expressed at high levels or not efficiently polyadenylated inCOS cells. In these studies it was shown by northern analysis using a specificprobe which distinguished between this size dass of RNA and the 500 and 600nt RNAs that these transcripts are indeed made in B19 transfected COS cells(Figure 9). Therefore, the 700 and 800 nt RNAs may not be efficientlypolyadenylated and since the first cDNA strand was made by reversetranscription using an oligo(dT) primer, this would result in theunderrepresentation of these sequences in the library. Perhaps the unusualpolyadenylation signal, ATTAAA or AATAAC, in the middle of the B19genome is not effectively recognized in COS cells. These variant sequenceshave been reported to be functional in 12% (ATTAAA) and 1% (AATAAC) ofeukaryotic genes (Wickens and Stephenson, 1984; Birnstiel et a!., 1985). Inaddition, a GU-rich downstream element identified to be important inpoly(A) site selection and 3t end-processing in SV4O (McDevitt et a!., 1986;104Weiss et a!., 1991) is present downstream from the canonical polyadenylationsequence at m.u. 97 but absent from the region downstream of the poly(A)signal at m.u. 49. Since COS cells are a simian cell line the lack of thissequence in the middle of the B19 genome may favour polyadenylation andtranscriptional termination at the right-hand end of the genome. In order toclarify this observation poiy A RNAs from transfected COS cells should beselected on a poly U Sepharose column and compared with unselected RNAsby northern blot analysis. This would establish whether the 700 and 800 ntRNAs are polyadenylated. The fact that this class of RNA can be detected intransfected COS cells but is not present in the library suggests that 3’ endprocessing may be aberrant (i.e. cleavage may occur without polyadenylation.).In the B19 COS cell cDNA library, the most abundant cDNAs were ofthe 500 and 600 nt size class. This library was made using reverse transcriptaseand DNA polymerase (Kienow fragment) hence, in contrast to PCR, smallertranscripts were not preferentially amplified. However, due to the RNase Hactivity and low processivity of reverse transcriptase, full-length cDNAs aremore difficult to obtain with this enzyme than are the shorter species (Matsonet al., 1980) so the abundances of cDNAs may be skewed towards the smallerspecies. Nevertheless, northern hybridization analysis has suggested that the500 and 600 nt RNAs are in fact the most prevalent transcripts in B19transfected COS cells (C. Beard, unpublished results).4.3. Comparison of Splice Junctions in the cDNAs from Transfected COSCells and B19 Infected CML CellsIt was shown that there was variability in two splice junctions in thesecond exon of the 500 and 600 nt RNAs made in COS cells. The splice donorsite of the first exon and the splice acceptor site of the third exon were105invariant in all clones sequenced and identical to those of the same cDNAsproduced in human cells. The splice acceptor sites for the second exon were nt1925, 1952, or 2030 in the COS cell system and nt 1910 or 2030 in the humancells. The splice donor site for this exon was nt 21.72 or 2183 in COS cellcDNAs and always nt 2183 in the human cell cDNA library (compare Figures7 and 8). The GT-AG consensus sequence in the 5’ and 3’ termini of the firstintron (Mount, 1982) was conserved in all cases suggesting that another signalor factor is important for correct splicing of B19 transcripts in human cells. Itis interesting that the splice acceptor site at nt 1910 was not seen in the COScell generated B19 cDNAs because this site is preferred over nt 2030 in thenatural host cell of the virus. In any case the variability in splicing involvingthe second exon of the 500 and 600 nt RNAs does not affect the proteinsynthesized from these transcripts since it is encoded entirely within the thirdexon.There are many examples of alternative splicing of RNA in differentcell types, including splicing to generate calcitonin in thyroid cells and CGRP(calcitonin gene-related peptide) in cells derived from neural tissues (Amaraet at., 1982; Emeson et al., 1989). In a recent study of the splicing of SV4O earlypre-RNA, the ratio of small t antigen (t) mRNA to large T antigen (T) mRNAproduced in human embryonic kidney 293 cells was 10- to 20-fold higher thanthat made in HeLa cells (Ge and Manley, 1990). The small t and large TmRNAs are processed from a single pre-RNA which can be spliced at twoalternative 5’ splice sites and a single 3’ splice site. A protein factor, ASF,purified from 293 cell extracts was shown in vitro to promote the use of thethe 5’ proximal splice site (Ge and Manley, 1990). It was also reported that anessential splicing factor, SF2, influences 5’ splice site selection in HeLa cells(Kramer et at., 1990). Comparison of ASF and SF2 suggests that they are the106same protein(s) indicating that cell-specific differences in the concentrationsor activities of general splicing factors could regulate alternative splicing(Maniatis, 1991). Regulation of alternative splicing has also be shown to becontrolled by specialized proteins such as occurs in the sex-determinationpathway of Drosophila. Sex lethal (Sxl) is an RNA-binding protein thatregulates alternative splicing of its own pre-RNA and that of transformer (tra)pre-mRNA by binding to specific splice sites thus preventing their use (Bell etaL, 1988).In this study the spliced junctions of five different B19 transcripts havebeen precisely mapped by direct sequencing of their cDNAs. In addition sinceSi nuclease and RNase mapping studies predicted that the spliced junctionsof the other three spliced transcripts were shared with these five sequencedRNAs we are confident that we have an accurate transcription map forparvovirus B19. Although NS cDNAs were not isolated in these studies, likethe 700 and 800 nt mRNAs, the NS transcripts are also processed andpolyadenylated using the unusual poly(A) signal ATTAAA or AATAAC atm.u. 49. Hence, we might expect the 2.3 kb NS mRNA to be of very lowabundance. However, since the NS proteins were readily detected intransfected COS cells (Figure 3) this low abundance RNA must be translatedvery efficiently, or the NS proteins must be very stable once they aresynthesized. Alternatively, the NS proteins may be synthesized from a full-length unspliced transcript.4.4. Expression of 11 kDa Proteins from the 500 and 600 nt cDNAsSmall, abundantly expressed RNAs are a unique feature of B19 geneexpression and in the studies reported here it has been shown that at least oneclass, the 500 and 600 nt transcripts, encode a family of 11 kDa proteins. The107518 and 638 nt RNAs are made from 3 exons, the length difference resultingfrom alternative splicing involving the second exon. Since the proximal 3’splice acceptor site at nt 1910 is preferred over the site at nt 2030 the longer 638nt species is the more abundant transcript in B19 infected erythroidprogenitor cells. The 94 aa ORF encoding the 11 kDa proteins is entirelywithin the third exon; hence, both transcripts encode the same 11 kDaproteins. This ORF is not used for either the structural or NS proteins andtherefore the translated polypeptides do not share any amino acid sequencehomology with other B19 proteins.In vitro translation products of SP6 RNA polymerase generatedtranscripts from these small cDNAs showed that at least two 11 kDa proteinscould be resolved by SDS-PAGE. These proteins were detected using an antiserum directed against an antigenic amino acid sequence contained withinthe 94 aa ORF. The two 11 kDa proteins did not merely reflect infidelity oftranslation initiation by the reticulocyte ribosomes since 11 kDa proteins ofthe same apparent MW by SDS-PAGE were synthesized in vivo in COS cellstransfected with SV4O-B19 hybrid plasmids (Figure 12). The two smallproteins were also detected on western blots of proteins from B19 infectedhuman leukemic marrow cells (Figure 14) suggesting that these polypeptidesmay have a role in the B19 life cycle.There was no evidence for a post-translational modification of either11 kDa protein. Amino acid sequence analysis predicted that the proteins donot contain a signal sequence or hydrophobic, membrane-spanning region.Although the amino acid sequence predicted that the proteins could bephosphorylated at serine or threonine residues by protein kinase C or caseinkinase II it was shown in these studies that these small polypeptides areapparently not phosphoproteins (Figure 15).108Expression studies of the small cDNAs and the sequence at the 5’ endof their respective mRNAs suggested that protein synthesis initates at morethan one AUG codon and this results in the synthesis of more than one 11kDa protein. Bifunctional and polyfunctional mRNAs have been widelyreported in viral systems (Kozak, 1986a). Proteins can be synthesized on thesame transcript from two different initiation codons in the same readingframe or in different reading frames. In most cases the 5’ proximal AUG is ina weak context for initiation (Kozak, 1986b). The scanning model predicts thatthe 40S ribosomal subunit carrying Met-tRNAmet and associated factorsmigrates along the mRNA in a 5’ to 3’ direction until it reaches the first AUGcodon (Kozak, 1983; Kozak, 1989). If the first AUG is in an optimal context[CCACCAUGG for eukaryotes (Kozak, 1986b)] the 40S subunit couples withthe 60S ribosomal subunit and protein synthesis initiates uniquely at this site.If the sequence around this first AUG is sub-optimal some of the smallribosomal subunits and factors may bypass this site and migrate furtherdownstream until reaching another AUG codon which is in a morefavourable context and initiate protein synthesis at this more favourable site.Translation initiation has also been reported to initiate from aninternal AUG when the 5’ proximal AUG is in a favourable context by amechanism of reinitiation by the ribosome after translation of a leaderpeptide. Termination of protein synthesis at a stop codon has been reported tolead to reinitiation at a nearby AUG in another frame (Liu et a!., 1984). In asituation that is not consistent with the scanning model, translation ofpicornavirus mRNAs which lack a 5’ terminal cap structure has been shownto be mediated by internal ribosome binding (Pelletier and Sonenberg, 1988).In addition, at least one cellular protein, BiP (immunoglobulin heavy-chainbinding protein), has recently been shown to initiate internally under109conditions where the cap recognition mechanism is inoperative (Macejak andSarnow, 1991).Using site-directed mutagenesis it was shown that three 11 kDaproteins could be translated from the same transcript by initiation at threedifferent AUG codons. Since the Kozak consensus sequence around the firstAUG triplet in the 94 aa ORF is a weak translation initiation signal (lackingthe important purine at position -3 as well as the preferred G in position +4)(Figure 16) a leaky scanning mechanism of initiation was predicted. Thismechanism could also explain translation initiation from non-AUG codons.After mutating both the second and third AUG codons in the 94 aa ORF thereappeared to be a low level of expression downstream from the first AUGsuggesting that protein synthesis was initiating at a CUG or other non-AUGtriplet. Since this first AUG is in a weak context some of the scanningribosomal subunits could still bypass the 5’ proximal AUG until encounteringand stalling at the first CUG codon. Both downstream CUG triplets were in agood sequence context for translation initiation so a low level of proteinsynthesis from this site was not unexpected. An upstream CUG codon in afavourable sequence context has been shown to initiate translation in thepim-1 oncogene of murine and human cells (Saris et a!., 1991). In that reportit was shown that the synthesis of a 44 kDa protein from an upstream CUGcodon was as efficient as that of a 34 kDa protein initiated from a downstreamAUG codon on the pim-1 mRNA.Mutation of the first weak AUG eliminated the higher molecularweight translation product suggesting that protein synthesis was initatinguniquely at the second AUG which is consistent with the scanning model.From the data and the model it is predicted that two 11 kDa proteins are madefrom the 518 and 638 nt mRNAs one of which is seven aa longer than the110other at its amino terminal. NH2-terminal amino acid sequences oftendetermine the cellular distribution and perhaps the activity of proteins(Kozak, 1988), therefore, the long and short forms of the 11 kDa protein mayhave functional significance.4.5. Expression of the Potential Protein in the 700 and 800 nt cDNAsThe most abundant RNAs in B19 infected cells, the 700 and 800 nttranscripts, contain a large ORF within their second exon. This ORF is sharedwith the NS proteins but in this region does not contain any AUG translationinitiation codons. Although AUG seems to be the exclusive initiator codon invertebrate (Kozak, 1987) and yeast mRNAs (Cigan and Donahue, 1987) thereare reports of inefficient but detectable translation initiation from non-AUGcodons mostly in viral systems. The possibility that B19 utililizes an alternatecodon for translation initiation was explored. A threonine, ACG, codon hasbeen reported to initiate translation in a Sendai virus protein (Curran andKolakofsky, 1988; Gupta and Patwardhan, 1988) and in a related humanparvovirus, AAV (Becerra et al., 1985). In AAV, both VP-2 and VP-3 aretranslated from the same transcript. By this mechanism there is a low level oftranslation initiated from an upstream ACG codon to produce VP-2 and ahigh level of translation initiation from a downstream AUG codon to makeVP-3. In this way the abundances of these two structural proteins areregulated independently of mRNA concentration such that VP-3 represents86% of the viral protein in the AAV virion (Carter et al., 1990). A candidateACG codon at the 5’ end of the 129 aa ORF in the 687 nt RNA (nt 2062 in thecloned B19 genome) could potentially initiate synthesis of a polypeptide witha molecular weight of approximately 15 kDa.111The data reported in this thesis suggest that the 687 and 807 nt RNAsare translatable in vitro. A protein of the expected MW was detected at verylow levels in translations of SP6 RNA polymerase generated transcripts of the687 nt cDNA in a cell-free system (Figure 22). This protein was specificallyimmunoprecipitated with antisera raised against a synthetic peptidecontaining antigenic amino acid sequences encoded within this potentialpolypeptide. Mutant transcripts, containing an AUG in place of the ACGcodon (aa 14 in the 129 aa ORF) expressed high levels of this 15 kDa protein invitro. This protein migrated at the same MW by SDS-PAGE as the productfrom the wild type 687 nt RNA suggesting that protein synthesis may initiateat or near this ACG codon in the rabbit reticulocyte lysate. Expectedly, thelevel of expression of the wild type product was much lower than that of theAUG mutant. In lysates of COS cells transfected with SV4O-B19 hybrid vectorsa protein in the region of 15 kDa could not be detected reproducibly byimmunoprecipitation or western blotting. It has been reported that initiationat non-AUG codons occurs far more efficiently in vitro in a rabbit reticulocytelysate than in vivo in COS cells and may be due to artificially highconcentrations of magnesium or spermidine in cell-free extracts (Andersonand Buzash-Pollert, 1985; Peabody, 1987; Gupta and Patwardhan, 1988). Themutant 687 nt cDNA cloned into the pCMV vector also failed to expressconvincingly a 15 kDa polypeptide (Figure 23) suggesting that if the 15 kDaprotein is synthesized from the ACG codon at nt 2062 or from another nonAUG codon it would probably not be detected by these methods. The 15 kDaprotein was not seen on a western blot of proteins from B19 infected humanleukemic cells suggesting again that if a protein is synthesized from thesetranscripts it is expressed at a very low level or it is unstable and therefore notreadily identified.112The larger, more abundant 807 nt RNA contains another ORFencoding a 71 aa polypeptide. This ORF does contain an AUG codon at its 5’end which does not appear in the smaller 687 nt RNA (refer to Appendix Cfor the nucleotide sequence; the ATG is underlined). It is possible that a smallprotein is synthesized from the 807 base transcript using this reading frameand work is in progress to determine if this is the case.4.6. If the 700-800 nt Class of RNAs are not Translated, What OtherFunction(s) Might They Have?If the 700 and 800 nt transcripts are not translated into proteins theymay be functional as RNAs. Short, non-coding, VA (virus-associated) RNAsare common to all adenoviruses and have been shown to be involved intranslational control and in counteracting host antiviral defenses.Adenoviruses encode two distinct VA RNAs, each about 160 nt long: a majorspecies VAT, and a minor species VAIl. These RNAs are transcribed by hostRNA polymerase III, are GC-rich, and can adopt stable secondary structureswhich are important for their function (Mathews and Shenk, 1991). The VARNAs are present in exceedingly high amounts late in adenovirus infection:..108 VAT and 106 VAIl molecules per infected cell (Soderlund et a!., 1976).VAT RNA blocks activation of the interferon induced eTF-2c protein kinase,DAI, which inactivates the protein synthesis initiation factor eTF-2 thusshutting down host cell protein synthesis (Kitajewski et a!., 1986; Katze et a!.,1987). It has been suggested that host and viral mRNAs segregate into twodifferent translational pools late in infection. Translation of host cellularmRNAs is inhibited in one pool as a result of eIF-2cz phosphorylation whileviral mRNA is protected from inhibition in another pool by the specificinteraction with VAT RNA (O’Malley et a!., 1989).113Epstein-Barr virus also encodes two RNA polymerase III transcriptswhich are similar in size and secondary structure to the VA RNAs. EBERRNAs also bind to and block activation of DAT but at a reduced efficiencywhen compared with VAT RNA (Clarke et a!., 1991).The TAR element (Tat-responsive element) of the human immunodeficiency virus type 1 (HIV-1) may also regulate DAT activity. The TARsequence is located in the 5’ untranslated region of all HIV-1 mRNAs and in aclass of nonpolyadenylated, promoter-proximal, —60 nt RNAs thataccumulate in the cytoplasm of infected cells (Kao et al., 1987; Laspia et a!.,1989). This sequence has been shown to assume a stem-loop structure, to bindDAT, and to block activation of DAT in vitro (Gunnery et a!., 1990; Roy et a!.,1991). Therefore, TAR RNA may also serve to neutralize this cellular defensemechanism during HTV-1 infection.A key difference between these DAT-interfering RNAs and the 700 and800 base transcripts is that the B19 RNAs are presumably polyadenylated andtranscribed by RNA polymerase IT, not polymerase ITT as are the VA and EBERRNAs. However, analysis of the B19 RNAs has predicted that that thesetranscripts assume extensive regions of secondary structure and it remains tobe determined if these structures are related to the function of the molecules(Figure 32).One explanation for the presence of abundant, left-sided, splicedmRNAs is that they may represent a mechanism to control the expression ofNS-1. In B19 it has been shown in these studies that there are at least twononstructural proteins translated from a single 2.3 kb transcript. There isindirect evidence that the NS-1 protein in MVM, H1, and B19 may becytotoxic (Rhode, 1987; Ozawa et a?., 1988a) hence, clearly the expression ofthese proteins should be controlled. In parvovirus genomes with a capsid114AT1T TC6C Tc.Ta c6C C6TCATC 6r CT C.16CA C ATfi C ATTTATCCTTCCTA CC AR GA ACT TC TGR 6 CC.6c66 cTCcA Ac A A6T CCCRCT aGTTGAR TAT T TTCT 6G Aaw C- ‘6 6A AC A ARFigure 32. Predicted secondary structures of the 687 and 807 nt RNAs.The secondary structures were predicted using the M-FOLD RNA foldingprogram of Michael Zuker (Division of Biological Sciences, NationalResearch Council of Canada) on a UNIX system and displayed using theMOLECULE program (Lapalme et al., 1982) modified by John Thompson(Dept. of Biological Sciences, Carnegie-Mellon University, Pittsburgh, PA).Structure A represents the folded structure of the first 200 nt of the 807 ntRNA and structure B represents that of the first 200 nt of the 687 nt RNA.Structure C is derived from nt 2101-2300 of the DNA sequence in the secondexon of the 700 and 800 nt RNAs. The predicted AG values for the threestructures are -29.7, -39.3 and -46.2, respectively.-4r.—Cl,.-4__.4ClClñ_m.riI-I—ClClClCl.4Cl,r)r4.,.4-,(-I.—-4.4a’a’-4-4a’a’Cl-4Cl4—Cla’ClC.)w-4 ClClCl-4a’,ClCl C4ClI,CIa’-4CI a’-4a.CICI—a’a’a.I_ICIri a’-4(a.m—,,ClCl-1CIa’-4-4-4ClCll .4a’C4Cla..4CICICI-4_4ma.Cl’a’Cl41a’ClClCICICl Cl.,,rQi—CICIa.-4ClClCl-4ClCIClCICI r.a.m-4Cl1a’CICla.—.4a.116promoter (all of the parvovirus genomes studied to date with the exception ofB19 and perhaps BPV) there is a temporal expression of gene expression withan early gene product (NS-1 or rep) expressed from a left-hand promoteractivating a second promoter which regulates the expression of the viralproteins later in infection (Rhode, 1985; Labow et a!., 1986; Tratschin et a!.,1986; Clemens and Pintel, 1988; Doerig et a!., 1988; Tullis et a!., 1988; Doerig eta!., 1990). Since B19 has a single promoter, gene expression cannot beregulated in this way. The capsid transcripts are transcribed from the samepromoter as the nonstructural proteins hence there must be somemechanism to shift expression late in infection to the production of capsids.Splicing and/or poly(A) site selection may play an important role in thisprocess. It has been shown in these studies that 3’ end processing at m.u. 49 isvery efficient in the host cell of the virus. Perhaps early in infection mosttranscripts are processed here and the nonstructural proteins predominate.Late in infection there may be some mechanism which favours full-lengthtranscripts and/or increases splicing activity and these mid-length transcriptsmay be actively spliced to produce the 700 and 800 nt RNAs. In this scenariothe 700 and 800 nt RNAs would merely be degradation products of spliced NSmessages.Such a shift in transcriptional termination is found in adenovirusinfected cells. Early in infection transcripts which initiate from the major latepromoter terminate near the middle of the genome downstream from aconsensus poly(A) recognition signal. Later in infection, when this promoterbecomes more active, transcription initiates at the same site but continues tonear the end of the genome at m.u. 99 bypassing the termination signal atm.u. 48 (Nevins, 1983). Thus, transcriptional termination controls theexpression of the distal coding sequences encoding the structural proteins.117In the retrovirus, HIV-1, early in infection RNA expression is limitedto the fully-spliced (—2 kb) transcripts which encode Tat, Rev, and Nef. Laterin infection, singly spliced (—4 kb) and unspliced (—9 kb) transcripts encodingthe structural proteins predominate in the cytoplasm (Kim et a!., 1989). Thisswitch is mediated by the viral Rev protein which binds to its target sequenceon the env gene in the viral pre-RNA and promotes nucleocytoplasmicexport of the incompletely spliced RNAs. In this way the small regulatoryproteins are made first and then later in infection expression is shifted to theproduction of capsid proteins (Cullen and Greene, 1990).Another example of alternative splicing and/or termination is foundin the progression from the membrane-bound to secreted form ofimmi.mogloblin in B cells. The switch occurs by a change in RNA processingwhich activates a poly(A) site previously spliced from the message encodingthe membrane-bound t chain resulting in a shorter transcript encoding a tchain that does not contain the hydrophobic anchor sequences therebyproducing the secreted IgG (Early et a!., 1980; Rogers et al., 1980).4.7. Localization of B19 Proteins in Transfected COS-7 Cells and InfectedHuman Erythroid Precursor CellsIndirect immunofluorescence localized the 11 kDa proteins to thecytoplasm of B19 transfected COS cells. However, this result was challengedafter the NS proteins also localized to the cytoplasm using the same method.Immunofluorescence of B19 proteins in infected human cells showed that atleast one of the NS proteins is nuclear and that the 11 kDa proteins arepartially nuclear. One possibility which would explain why the COS celllocalization of the B19 proteins is aberrant is that the B19 nuclear signalsequence may not be recognized in COS cells. The nuclear localization signal118has not been identified in any parvovirus protein including those of B19 andthere are no apparent homologies between the nuclear targetting sequence oflarge T antigen from SV4O, which would be recognized in COS cells, and anyB19 nuclear protein. Alternatively, there may be a requirement for a factorsuch as a protein to import B19 proteins into the nucleus which is absent orinactive in COS cells. Several reports have indicated that proteins may enterthe nucleus due to interaction with other proteins (Moreland et at., 1987; Liand Rhode, 1990; Sommer et a!., 1991). Alternatively, correct intracellularlocalization may depend on having replicating intact viral DNA within thenucleus. Knowing the cellular compartment in which the 11 kDa proteins arelocalized may suggest a function for these proteins.4.8. What Might the Function of the 11 kDa Proteins Be?There is no precedent in other parvoviruses for abundantly transcribedsmall RNAs or for highly expressed small proteins. The limited codingcapacity of the B19 genome would predict that a polypeptide synthesized ininfected cells is probably functional. These proteins may be a unique feature ofB19 and may relate to B19 cell tropism or pathology. B19 is only distantlyrelated to the other mammalian parvoviruses with the only significanthomology evident in the nonstructural coding region (Astell et at., 1987).However, there are small ORFs existent in other parvoviruses which mayencode polypeptides which are analogous if not homologous to the 11 kDaproteins; in particular there are two potential 10 kDa proteins encoded in theintron of AAV-2 (Carter et a!., 1990). One of these putative polypeptides,designated intlO is in the same phase as the structural proteins but the ORF isfollowed by two UGA termination signals so read-through into VP-2 and VP3 would probably not occur. The other putative small protein, designated119replO, shares the same carboxyl terminal amino acid sequences as rep78 andrep52. (This is the same situation as that found between the potential proteinencoded in the 700 and 800 nt RNAs and the NS proteins in B19). Thesequence of the putative replO protein is rich in cysteine and histidineresidues which form a sequence motif that is repeated three times and isreminiscent of a retroviral RNA-binding domain (shown below andcompared with the replO sequence, where X is any amino acid other than Cysor His).-Cys- X2-Cys- X4-His- X4-Cys- retroviral zinc finger sequence-Cys- X2- His- X9-Cys- X2- Cys- X9-Cys- X2-His- X4-Cys- X2-Cys- X19--Cys- X2-His- X9-Cys- X2-Cys- replO sequenceDuring retroviral replication the gag polypeptide binds RNA by thecoordination of side chains from three Cys residues and one His residue witha zinc atom (Summers, 1990). The replO protein has never been detected andits genetic analysis is incomplete. Although the entire intron can be deletedwithout affecting known rep function (Tratschin et a!., 1984) a frameshiftmutation within the intron was reported to result in defective AAVreplication (Hermonat et al., 1984).Extensive searching of DNA and protein data bases did not reveal anysignificant similarity between the sequence of the 11 kDa proteins and that ofother known proteins. The primary amino acid sequence is rich in prolineresidues; an internal thirty amino acid stretch has a proline composition of>43%. Since proline-rich domains have been shown to be involved intranscriptional activation (Mermod et al., 1989; Mitchell and Tjian, 1989) the120larger 11 kDa protein fused to the yeast GAL-4 DNA binding domain wastested for transcriptional activation of a minimal promoter (Elb TATA box)driving the expression of a reporter CAT gene using transient expression intransfected COS cells. The results of this assay suggest that the 11 kDa proteindoes not have an activation domain (Figure 31). In a recent report theglutamine- and proline-rich amino terminal domain of the Wilms tumor(WT1) gene product was shown to have a repressor functiOn intranscriptional assays in COS cells (Madden et a!., 1991), however, the 11 kDaprotein was not investigated for repressor function.A functional classification of viral proteins separates the proteins intotwo broad groups: structural and nonstructural. Structural proteins such asVP-1 and VP-2 make up the capsid structure of the virion. The nonstructuralproteins such as NS-1 have enzymatic activities and may be involved inreplication and regulation of transcription. Evidence from these studiessuggests, but by no means proves, that the 11 kDa protein is nonstructural.The observation that two B19 convalescent antisera fail to recognize the 11kDa polypeptides but do detect the two capsid proteins on a western blotsuggest that the 11 kDa proteins are not part of the capsid structure (data notshown). However, we do not know if there are anti-il kDa antibodies in theseconvalescent sera which recognize native 11 kDa protein. Other studies haveshown that viral particles are assembled in COS cells and insect cellstransfected or infected with B19 genomes which cannot make the 11 kDaprotein suggesting that these small proteins are not involved in capsidassembly (Brown et a!., 1991; Kajigaya et a!., 1991). Also, rabbit polyclonal seraraised against VP-2, immunoprecipitate capsid proteins but fail to coprecipitate the 11 kDa proteins, and immune sera specific for the 11 kDaproteins precipitate the Ii kDa proteins but not the capsid proteins (Figure 20).121If the 11 kDa proteins were intimately associated with the virion one wouldexpect the proteins to be immunoprecipitated together. Lastly, in fractions ofpurified capsids, separated on a CsC1 gradient, the 11 kDa proteins were notidentified in those fractions containing VP-2 by western blot analysis (D.Minato and J. St. Amand, unpublished results). These results indicate that the11 kDa proteins are probably not structural proteins.Due to the paucity of a cell line in which B19 is infectious it is notpossible to perform a mutagenic analysis of the 11 kDa proteins in B19infected cells. In COS cells it was shown that the expression of the major B19structural and nonstructural proteins was not affected in the absence of the 11kDa proteins (Figure 21). However, these results became equivocal when thelocalization of the small proteins was found to be aberrant in these cells. The11 kDa proteins are at least partially nuclear in B19 infected cells and likelyperform some regulatory function.Small regulatory proteins are widespread in viral systems. In HIV-1,Tat and Rev are involved in transactivation and nucleocytoplasmic export ofincompletely spliced mRNAs respectively (Cullen and Greene, 1990). Othersmall proteins such as Nef and Vif have accessory functions that enhancereplication or infectivity and these activities may be difficult to determine in aheterologous tissue culture system. In these studies we have shown that the11 kDa proteins do not appear to regulate the expression of the major B19proteins in COS cells. However, there could be a requirement for a Rev-likeactivity in B19 infected cells which would control splicing or promote thenuclear export of the singularly spliced RNAs encoding the VP-i and thedoubly spliced VP-2. A time course of the expression of the 11 kDa proteins inB19 infected leukemic cells showed that the 11 kDa proteins could be detectedby indirect immunofluorescence early in infection (36 h.p.i.) at the same time122as the nonstructural proteins. A similiar time course study to monitor theabundance of cytoplasmic RNAs would determine if in fact there is a shiftfrom completely spliced right-sided mRNAs corresponding to the 500 and 600nt transcripts to the singularly spliced and doubly spliced VP-i and VP-2mRNAs as infection proceeds. This shift may involve masking the 3’ spliceacceptor site at nt 4704 so that the acceptor at nt 3045 is preferred and VP-2mRNAs are made. In vitro studies are currently in progress to determine ifthe 11 kDa proteins bind to B19 mRNA. The octapeptide, K/R-G-F/Y-G/A-FV-X-F/Y, a highly conserved RNA-binding motif present in the RNA-bindingdomains of a large group of RNA-binding proteins, is not present in the 11kDa proteins (Bandziulis et a!., 1989). If the 11 kDa proteins have an accessoryfunction(s) this could only be determined in the natural host cell of the virus.4.9. Summary of Mechanisms of Gene Expression in B19The human parvovirus, B19, has evolved novel ways to both controlthe expression of its genes and to expand the coding potential of its limitedgenome of 5.6 kb. Gene expression from a single promoter is regulated bydifferential splicing and 3’ end processing at more than one termination site(Ozawa et a!., 1987; St. Amand et a!., 1991; this thesis). In addition,translational supression of the minor capsid protein appears to occur due to astring of out-of-frame AUG triplets upstream from the initating AUG codonon the VP-I mRNAs (Ozawa et a!., 1988b). The number of proteinssynthesized is increased by using overlapping reading frames and thediversity of polypeptides is expanded using bicistronic mRNAs encoding twodifferent proteins from a single transcript as has been shown for the two 11kDa proteins and suggested for the two NS proteins (this thesis). It remains tobe seen what role the small proteins have in the B19 life cycle. Since small123RNAs and proteins are a novel feature of the B19 parvovirus, they may relateuniquely to gene expression in this virus. Knowledge of their function will beimportant in understanding the viral replication cycle and could be ofconsiderable significance if the B19 genome is developed as a vector for use ingene therapy.4.10. Future DirectionsThe COS cell system has been useful in studying the expression of B19RNAs and proteins from both the B19 and the CMV promoters. However, itis clear that there are differences in B19 expression between COS cells and thenatural host cell of the virus. In COS cells the transcripts terminating at theright-hand end of the genome appear to be favoured over those terminatingin the middle of the genome and the localization of B19 proteins has beenshown to be aberrant. In order to determine the function of the small proteinsand RNAs the expressing cells must reflect the host cell of the virus.The development of a continuous cell line in which B19 is infectiouswould greatly facilitate further study of B19 gene expression. In lieu of this,the fetal liver system appears to offer the most promise. One report suggestedthat hematopoietic cells from 1st trimester fetal liver cultures contain up to70% target cells for B19 infection (Morey et a!., 1991). The recently constructedfull-length B19 clone (Deiss et a!., 1990) might be used to transfect primarycultures of these cells, however, so far there are no reports indicating that thisclone is infectious. If the clone is infectious a mutagenic analysis of the 11 kDaproteins could be performed to assess their function in infected cells. In lieuof an infectious clone, it may be possible to use antisense oligonucleotides toblock the expression of the 11 kDa proteins in B19 infected cells. Extracts from124infected cells could also be used to determine if the 700 and 800 nt RNAs bindcellular or viral proteins.The COS cell expression system could be further developed fordiagnostics. The potentially serious consequences of B19 infection emphasizethe need for reliable clinical testing. Infections are usually diagnosed by thedetection of B19-specific 1gM or IgG. Since there is no convenient tissueculture system for generating B19 virus and the viremic period is very shortthere is a scarcity of viral antigen for such testing. It was shown that SV4O-B19transfected COS cells produce B19 capsid proteins which self-assemble intoparticles and that these B19 particles are antigenic in RIA andimmunofluorescence assays (Cohen et a!., 1991). Although the baculovirussystem produces a much higher level of capsid protein (—500,000particles/cell) than the COS cell system (—1000 particles/cell) there is someevidence that the COS cell generated antigen may perform better than thatmade in insect cells in a hemagglutination assay (Kevin Brown; personalcommunication). Since COS cells are a mammalian cell line one wouldexpect the B19 proteins to be modified in the same way as would occur inhuman cells.It will be interesting to discover the special features of B19 whichrestrict its replication to a small subset of erythroid progenitor cells and touncover the mechanisms behind its pathogenesis in these cells.125LITERATURE CITEDAbraham, S., Kienzle, T. E., Lapps, W. E., and Brian, D. A. (1990). Sequenceand expression analysis of potential nonstructural proteins of 4.9, 4.8, 12.7,and 9.5 kDa encoded between the spike and membrane protein genes of thebovine coronavirus. Virology 177: 488-495.Alexandersen, S., Bloom, M. E., and Perryman, S. (1988). Detailedtranscription map of Aleutian mink disease parvovirus. I. Virol. 62: 3684-3694.Amara, S. G., Jonas, V., Rosenfeld, M. G., Ong, E. S., and Evans, R. M. (1982).Alternative RNA processing in calcitonin gene expression generates mRNAsencoding different polypeptide products. Nature 298: 240-244.Anand, A., Gray, E. S., Brown, T., Clewley, J. P., and Cohen, B. J. (1987).Human parvovirus infection in pregnancy and hydrops fetalis. N. Engl. I.Med. 316(4): 183-186.Anderson, C. W., and Buzash-Pollert, E. (1985). Can ACG serve as an initiatorcodon for protein synthesis in eucaryotic cells. Mol. Cell. Biol. 5: 3621-3624.Anderson, L. J., and Hurwitz, E. S. (1988). Human parvovirus B19 andpregnancy. Clin. Perinatol. 15: 273-286.Anderson, M., Jones, S., Fisher-Hoch, S., Lewis, E., Hall, S., Bartlett, C., Cohen,B., Mortimer, P., and Pereira, M. (1983). Human parvovirus, the cause oferythema infectiosum (fifth disease). Lancet 1: 1378.Andersson, S., Davis, D. L., Dahiback, H., Jornvall, H., and Russell, D. W.(1989). Cloning, structure, and expression of the mitochondrial cytochrome P450 sterol 26-hydroxylase, a bile acid biosynthetic enzyme. J. Biol. Chem.264(14): 8222-8229.Arya, S. K., Guo, C., Josephs, S. F., and Wong-Staal, F. (1985). Trans-activatorgene of human T-lymphotrophic virus type III (HTLV-III). Science 229: 69-73.Astell, C. R., Chow, M. B., and Ward, D. C. (1985). Sequence analysis of thetermini of virion and replicative forms of minute virus of mice DNAsuggests a modified rolling hairpin model for autonomous parvovirus DNAreplication. I. Virol. 54(1): 171-177.Astell, C. R., Mol, C. D., and Anderson, W. F. (1987). Structural and functionalhomology of parvovirus and papovavirus polypeptides. I. Gen. Virol. 68: 885-893.126Astell, C. R., Smith, M., Chow, M. B., and Ward, D. C. (1979). Structure of the3’ hairpin termini of four rodent parvovirus genomes: Nucleotide sequencehomology at origins of DNA replication. Cell 17: 691-703.Astell, C. R., Thomson, M., Merchilinsky, M., and Ward, D. C. (1983). Thecomplete nucleotide sequence of the minute virus of mice, an autonomousparvovirus. Nucl. Acids Res. 11(4): 999-1018.Bandziulis, R. J., Swanson, M. S., and Dreyfuss, G. (1989). RNA-bindingproteins as developmental regulators. Genes Dev. 3: 431-437.Beard, C., St. Amand, J., and Astell, C. R. (1989). Transient expression of B19parvovirus gene products in COS-7 cells transfected with B19-SV4O hybridvectors. Virology 172: 659-664.Becerra, S. P., Rose, J. A., Hardy, M., Baroudy, B. M., and Anderson, C. W.(1985). Direct mapping of adeno-associated virus capsid proteins B and C: Apossible ACG initiation codon. Proc. Natl. Acad. Sci. USA 82: 7919-7923.Bell, L. R., Maine, E. M., Schedl, P., and Cline, T. W. (1988). Sex-lethal, aDrosophila sex determination switch gene, exhibits sex-specific RNA splicingand sequence similarity to RNA binding proteins. Cell 55: 1037-1046.Ben-Asher, E., and Aloni, Y. (1984). Transcription of minute virus of mice, anautonomous parvovirus may be regulated by attenuation. I. Virol. 52: 266-276.Berns, K., Kotin, R., Linden, M., and Beaton, A. (1991). Site specific integrationby AAV. 4th Parvovirus Workshop, Elsinore, Denmark, August 18-22.Berns, K. I., Ed. (1984). “The Parvoviruses”. New York, Plenum Pubi.Berns, K. I. (1990). Parvovirus replication. Microbiol. Rev. 54(3) 316-329.Birnstiel, M. L., Busslinger, M., and Strub, K. (1985). Transcriptiontermination and 3’ processing: the end is in site. Cell 41: 349-359.Blundell, M., Beard, C., and Astell, C. R. (1987). In vitro identification of a B19parvovirus promoter. Virology 157: 534-538.Bourguignon, G. J., Tattersall, P. J., and Ward, D. C. (1976). DNA of minutevirus of mice: self-priming, non-permutated, single stranded genome with 5’terminal hairpin duplex. J. Virol. 20(1): 290-306.BRL (1986). Focus 8(2): 9.Brown, C., Lent, J., Viak, J., and Spaan, W. (1991a). Assembly of empty capsidsby using baculovirus recombinants expressing human parvovirus B19structural proteins. I. Virol. 65(5): 2702-2706.127Brown, K., Mori, J., Cohen, B., and Field, A. (1991b). In vitro propagation ofparvovirus B19 in primary foetal liver culture. J. Gen. Virol. 72: 741-745.Calnan, B. J., Tidor, B., Biancalana, S., Hudson, D., and Frankel, A. D. (1991).Arginine-mediated RNA recognition: the arginine fork. Science 252: 1167-1171.Campbell, D. A., Staal, S. P., Manders, E. K., Bonnard, G. D., Oldham, R. K.,Salzman, L. A., and Herberman, R. B. (1977). Inhibition of in vitrolymphoproliferative responses by in vivo passaged rat 13762 mammaryadenocarcinoma cells. Ill. Evidence that Kilham rat virus is responsible forthe inhibitory effect. Cell. Immunol. 33: 378-391.Carter, B., Trempe, J. B., and Mendelson, E. (1990). Adeno-associated virusgene expression and regulation. in “Handbook of Parvoviruses”. Boca Raton,FL, CRC Press. 227-254.Caspar, D., and Klug, A. (1962). Physical principles in the construction ofregular viruses. Cold Spring Harbor Symp. Quant. Biol., Cold Spring Harbour,NY.Chang, D. D., and Sharp, P. A. (1989). Regulation by HIV Rev depends uponrecognition of splice sites. Cell 59: 789-795.Cigan, A. M., and Donahue, T. F. (1987). Sequence and structural featuresassociated with translational initiator regions in yeast- a review. Gene 59: 1-18.Clarke, P. A., Schwemmle, M., Schickinger, J., Hilse, K., and Clemens, M. J.(1991). Binding of Epstein-Barr virus small RNA EBER-1 to the double-stranded RNA-activated protein kinase DAT. Nuci. Acids Res. 19: 243-248.Clemens, K. E., and Pintel, D. (1988). The two transcription units of theautonomous parvovirus minute virus of mice are transcribed in a temporalorder. J. Virol. 62(4): 1448-1451.Cohen, B. J. (1988). Laboratory tests for the diagnosis of infection with B19virus, in “Parvoviruses and Human Disease”. Boca Raton, Florida, CRC Press.69-83.Cohen, B. J., Field, A. M., Clewley, J. P., Mori, J., Brown, K. E., St. Amand, J.,and Astell, C. R. (1991). Morphologic and antigenic properties of recombinantB19 parvovirus particles expressed in transfected COS-7 cells. 4th ParvovirusWorkshop, Elsinore, Denmark, August 18-22.Cornelis, J., Avalosse, B., Guetta, E., Tal, J., and Rommelaere, J. (1984).Transformation in vitro of rat cells correlates with an enhancement of theirsusceptibility to the production of and killing by parvovirus minute virus ofmice. Arch. mt. Physiol. Biochem. 92: 14-15.128Cornelis, J., Avalosse, B., Mousset, S., Mamba, M., and Rommelaere, J. (1986).Selective destruction by parvoviruses of human and mouse fibroblaststransformed with gamma-irradiation. mt. j. Radiat. Biol. 49: 529.Cossart, Y. E., Cant, B., Field, A. M., and Widdows, D. (1975). Parvovirus-likeparticles in human sera. Lancet i: 72-73.Cotmore, S., McKie, V., Anderson, L., Astell, C., and Tattersall, C. (1986).Identification of the major structural and nonstructural proteins encoded byhuman parvovirus B19 and mapping of their genes by procaryotic expressionof isolated genomic fragments. J. Virol. 60(2): 548-557.Cotmore, S., and Tattersall, P. (1984). Characterization and molecular cloningof a human parvovirus genome. Science 226: 1161-1165.Cotmore, S. F., and Tattersall, P. (1986). The NS-1 polypeptide of theautonomous parvovirus MVM is a nuclear phosphoprotein. Virus Res. 4:243-250.Cotmore, S. F., and Tattersall, P. (1987). The autonomously replicatingparvoviruses of vertebrates. Adv. Virus Res. 33: 91-174.Cotmore, S. F., and Tattersall, P. (1989). A genome-linked copy of the NS-1polypeptide is located on the outside of infectious parvovirus particles. I.Virol. 63(9): 3902-3911.Cotmore, S. F., and Tattersall, P. (1990). Alternate splicing in a parvoviralnonstructural gene links a common amino-terminal sequence todownstream domains which confer radically different localization andturnover characteristics. Virology 177: 477-487.Cullen, B. R., and Greene, W. C. (1990). Functions of the auxilliary geneproducts of the human immunodeficiency virus type 1. Virology 178: 1-5.Curran, J., and Kolakofsky, D. (1988). Ribosomal initiation from an ACGcodon in the Sendai virus PC/mRNA. EMBO J. 7: 245-251.Deiss, V., Tratschin, J.-D., Weitz, M., and Siegi, G. (1990). Cloning of thehuman parvovirus B19 genome and structural analysis of its palindromictermini. Virology 175: 247-254.Diffoot, N., Shull, B., Chen, K., Stout, E., Lederman, M., and Bates, R. (1989).Identical ends are not required for the equal encapsidation of plus- andminus-strand parvovirus Luill DNA. J. Virol. 63(7): 3180-3184.Doerig, C., Beard, P., and Hirt, B. (1987). A transcriptional promoter of thehuman parvovirus B19 active in vitro and in vivo. Virology 157: 539-542.129Doerig, C., Hirt, B., Antonietti, J.-P., and Beard, P. (1990). Nonstructuralprotein of parvovirus B19 and minute virus of mice controls transcription. J.Viral. 64(3): 387-396.Doerig, C., Hirt, B., Beard, P., and Antonietti, J.-P. (1988). Minute virus of micenon-structural protein NS-1 is necessary and sufficient for trans-activation ofthe viral P39 promoter. I. Gen. Virol. 69: 2563-2573.Duncan, J., Potter, C., Cappelini, M., Durtz, J., Anderson, M., and Weatherall,D. (1983). Aplastic crisis due to parvovirus infection in pyruvate kinasedeficiency. Lancet ii: 14-16.Early, P., Rogers, J., Davis, M., Calame, K., Bond, M., Wall, R., and Hood, L.(1980). Two mRNAs can be produced from a single immunoglobulin mugene by alternative RNA processing pathways. Cell 20: 313-319.Eaves, A. C., Henkelman, D. H., and Eaves, C. J. (1980). Abnormalerythropoiesis in the myeloproliferative disorders: An analysis of underlyingcellular and humoral mechanisms. Exp. Hematol. 8: 235-247.Emeson, R. B., Hedyran, F., Yeakley, J. M., Guise, J. W., and Rosenfeld, M. G.(1989). Alternative production of calcitonin and CGRP mRNA is regulated atthe calcitonin-specific splice acceptor. Nature 341: 76-80.Faust, E. A., Brudzynska, K., and Morgan, J. (1989). Characterization of novelpopulations of MVM virions containing covalent DNA-protein complexes.Virology 168: 128-137.Feinberg, A. P., and Vogeistein, B. (1983). A technique for radiolabeling DNArestriction endonuclease fragments to high specific actvity. Anal. Biochem.132: 6-13.Felber, B. K., Hadzopoulou-Cladaras, M., Cladaras, C., Copeland, T., andPaviakis, G. N. (1989). Rev protein of human immunodeficiency virus type 1affects the stability and transport of the viral mRNAs. Proc. Nati. Acad. Sci.USA 86: 1495-1499.Fisher, A. G., Ensoli, B., Ivanoff, L., Chamberlain, M., Petteway, S., Ratner, L.,Gallo, R. C., and Wong-Staal, F. (1987). The sor gene of HIV-I is required forefficient virus transmission in vitro. Science 237: 888-893.Frickhofen, N., and Young, N. S. (1989). Persistent parvovirus B19 infectionsin humans. Microb. Pathog. 7: 319-327.Frohman, M. A., Dush, M. K., and Martin, G. R. (1988). Rapid production offull-length cDNAs from rare transcripts: Amplification using a single gene-specific oligonucleotide primer. Proc. Nati. Acad. Sci. USA 85: 8998-9002.130Ge, H., and Manley, J. L. (1990). A protein factor, ASF, controls cell-specificalternative splicing of SV4O early pre-mRNA in vitro. Cell 62: 25-34.Gerard, G. F. (1983). Reverse transcriptase. in “Enzymes of Nucleic AcidSynthesis and Modification: DNA Enzymes”. Boca Raton, FL, CRC Press. 1.Giraud, C., and Bergoin, M. (1991). Expression of the Junonia coeniadensovirus (Jc DNV) genome: Development of a Jc DNV-derived insect cellexpression vector. 4th Parvovirus Workshop, Elsinore, Denmark, August 18-22.Gluzman, Y. (1981). SV4O-transformed simian cells support the replication ofearly SV4O mutants. Cell 23: 175-182Goebel, S. J., Johnson, G. P., Perkus, M. E., Davis, S. W., Winslow, J. P., andPaoletti, E. (1990). The complete DNA sequence of vaccinia virus. Virology179: 247-266.Gorman, C. M., Moffat, L. F., and Howard, B. H. (1982). Recombinant genomeswhich express chioramphenicol acetyltransferase in mammalian cells. Mol.Cell. Biol. 2: 1044-1051.Green, M. R., and Roeder, R. G. (1980). Definition of a novel promoter for themajor adeno-associated virus mRNA. Cell 22: 231-242.Gunnery, S., Rice, A. P., Robertson, H. D., and Mathews, M. B. (1990). Tat-responsive region RNA of human immunodeficiency virus 1 can preventactivation of the double-stranded RNA-activated protein kinase. Proc. Nati.Acad. Sd. LISA 87: 8687-8691.Gupta, K. C., and Patwardhan, S. (1988). ACG, the initiator codon for a Sendaivirus protein. I. Biol. Chem. 263: 8553-8556.Guy, B., Kieny, M. P., Riviere, Y., Le Peuch, C., Dott, K., Girard, M.,Montagnier, L., and Lecocq, J.-P. (1987). HIV F/3’ orf encodes a phosphorylatedGTP-binding protein resembling an oncogene product. Nature 330: 266-269.Guy, B., Riviere, Y., Dott, K., Regnault, A., and Kieny, M. P. (1991). Mutationalanalysis of the HIV nef protein. Virology 176: 413-425.Hampton, E. (1970). H-i growth in synchronized rat embryo cells. Can. I.Microbiol. 16: 266.Hanahan, D. (1983). Studies on transformation of Escherichia coli withplasmids. 1. Mo!. Biol. 166: 557.131Heaphy, S., Dingwall, C., Ernberg, I., Gait, M. J., Green, S. M., Karn, J., Lowe, A.D., Singh, M., and Skinner, M. A. (1990). l-IIV-I regulator of virion expression(Rev) protein binds to an RNA stem-loop structure located within the Revresponse element region. Cell 60: 685-693.Heilbronn, R., Schiehofer, J., and zur-Hausen, H. (1984). Selective killing ofcarcinogen-treated SV4O-transformed Chinese hamster cells by a defectiveparvovirus. Virology 136: 439-442.Hermonat, P., Labow, M. A., Wright, R., Berns, K. I., and Muzyczka, N. (1984).Genetics of adeno-associated virus: isolation and preliminary characterizationof mutants of adeno-associated virus type 2. 1. Viro!. 51: 329-339.Hirt, B. (1967). Selective extraction of polyoma DNA from infected mouse cellcultures. J. Mol. Biol. 26: 365-369.Holmes, D. S., and Quiqley, M. (1981). A rapid boiling method for thepreparation of bacterial plasmids. Anal. Biochem. 114: 193.Hopp, T. P., and Woods, K. R. (1981). Prediction of protein antigenicdeterminants from amino acid sequences. Proc. Natl. Acad. Sci. USA 78: 3824-3828.Tm, D. S., and Muzyczka, N. (1990). The AAV origin-binding protein rep-68 isan ATP dependent site specific endonuclease with DNA helicase activity. Cell61: 447-457.IntelliGenetics PC/Gene Software. Mountainview, California,Jongeneel, C. V., Sahli, R., McMaster, G. K., and Hirt, B. (1986). A precise mapof splice junctions in the mRNAs of minute virus of mice, an autonomousparvovirus. I. Virol. 59(3): 564-573.Kajigaya, S., Fujii, H., Field, A., Anderson, S., Rosenfeld, S., Anderson, L. J.,Shimada, T., and Young, N. S. (1991). Self-assembled B19 parvovirus capsids,produced in a baculovirus system, are antigenically and immunologicallysimiliar to native virions. Proc. Natl. Acad. Sci. USA 88: 4646-4650.Kao, S.-Y., Calman, A. F., Luciw, P. A., and Peterlin, B. M. (1987). Anti-termination of transcription within the long terminal repeat of HIV-1 by tatgene product. Nature 330: 489-493.Katze, M. G., DeCorato, D., Safer, B., Galabru, J., and Hovanessian, A. G. (1987).Adenovirus VAT RNA complexes with the 68,000 Mr protein kinase toregulate its autophosphorylation and activity. EMBO J. 6: 689-697.Kaufman, R. J., and Sharp, P. A. (1982). Construction of a modulardihydrofolate reductase cDNA gene: Analysis of signals utilized for efficientexpression. Mo!. Cell. Biol. 2: 1304-1319.132Kelleher, J., Naomi, M., Luban, L., Mortimer, P., and Kamimura, T. (1983).Human serum parvovirus: A specific cause of aplastic crisis in children withhereditary spherocytosis. J. Pediatr. 102: 720-722.Killam, L., and Olivier, L. J. (1959). A latent virus of rats isolated in tissueculture. Virology 7: 428-437.Kim, S., Byrn, R., Groopman, J., and Baltimore, D. (1989). Temporal aspects ofDNA and RNA synthesis during human immunodeficiency virus infection:evidence for differential gene expression. I. Virol. 63: 3708-3713.Kitajewski, J., Schneider, R. J., Safer, B., Munemitsu, S. M., Samuel, C. E.,Thimmappaya, B., and Sheñk, T. (1986). Adenovirus VAT RNA antagonizesthe antiviral action of the interferon-induced eIF2cx-kinase. Cell 45: 195-200.Kotin, R., Siniscalco, M., Samuiski, R. J., Zhu, X., Hunter, L., Laughlin, C. A.,McLaughlin, S., Muzyczka, N., Rocchi, M., and Berns, K. I. (1990). Site-specificintegration by adeno-associated virus. Proc. Natl. Acad. Sci. USA 87: 2211-2215.Kotin, R. M., and Berns, K. I. (1989). Organization of adeno-associated virusDNA in latently infected Detroit 6 cells. Virology : 460-467.Kozak, M. (1983). Comparison of initiation of protein synthesis in procaryotes,eucaryotes, and organelles. Microbiol. Rev. 47: 1-45.Kozak, M. (1986a). Bifunctional mR.NAs in eukaryotes. Cell 47: 481-483.Kozak, M. (1986b). Point mutations define a sequence flanking the AUGinitiator codon that modulates translation by eukaryotic ribosomes. Cell 44:283-292.Kozak, M. (1987). An analysis of 5’-noncoding sequences from 699 vertebratemessenger RNAs. Nuci. Acids Res. 15: 8125-8148.Kozak, M. (1988). A profusion of controls. I. Cell Biol. 107: 1-7.Kozak, M. (1989). The scanning model for translation: an update. I. Cell Biol..108: 229-241.Kramer, A. R., Conway, G. C., and Kozak, D. (1990). The essential pre-mRNAsplicing factor SF2 influences 5’ splice site selection by activating proximalsites. Cell 62: 35-42.Krauskopf, A., Bengal, E., and Aloni, Y. (1991). The block to transcriptionelongation at the minute virus of mice attenuator is regulated by cellularelongation factors. Mol. Cell. Biol. 11(7): 3515-3521.133Kunkel, T., Roberts, J., and Zakour, R. (1987). Rapid and efficient site-specificmutagenesis without phenotypic selection. Methods Enzymol. 154: 367-382.Kurtzman, C., Cohen, B., Meyers, P., Amunullah, A., and Young, N. S. (1988).Persistant B19 parvovirus infection as a cause of severe chronic anemia inchildren with acute lymphocytic leukemia. Lancet ii: 1159-1162.Kurtzman, G., Ozawa, K., Cohen, B., Hanson, G., Oseas, R., and Young, N. S.(1987). Chronic bone marrow failure due to B19 parvovirus infection. N. Engl.J. Med. 317: 287-294.Labow, M., Hermonet, P., and Berns, K. (1986). Positive and ne.gativeautoregulation of the adeno-associated virus type 2 genome. J. Virol. 60: 251-258.Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly ofthe head of bacteriophage T4. Nature 227: 680-685.Lamb, R. A., and Horvath, C. M. (1991). Diversity of coding strategies ininfluenza viruses. Trends in Genetics 7(8): 261-246.Lapalme, G., Cedergren, R. J., and Sankoff, D. (1982). An algorithm for thedisplay of nucleic add secondary structure. Nuci. Acids Res. 10: 8351-8356.Laspia, M. F., Rice, A. P., and Mathews, M. B. (1989). HIV-1 Tat proteinincreases transcriptional initiation and stabilizes elongation. Cell 59: 283-292.Lebovitz, R., and Roeder, R. (1986). Parvovirus H-i expression: Mapping ofthe abundant cytoplasmic transcripts and identification of promoter sites andoverlapping transcription units. I. Virol. 58: 271-280.Li, X., and Rhode, S. L., III. (1990). Study of the function of the smallparvovirus nonstructural protein NS-2 by mutagenesis. ASV Meeting, SaltLake City, Utah.Linser, P., Bruning, H., and Armentrout, R. (1977). Specific binding sites for aparvovirus, minute virus of mice, on cultured cells. J. Virol. 24: 211-221.Liu, C. C., Simonsen, C. C., and Levinson, A. D. (1984). Initiation oftranslation at internal AUG codons in mammalian cells. Nature 309: 82-85.Liu, J. M., Fujii, H., Green, S. W., Komatsu, N., Young, N. S., and Shimada, T.(1991). Indiscriminate activity from the B19 parvovirus P6 promoter innonpermissive cells. Virology 182: 361-364.Lopata, M. A., Cleveland, D. W., and Sollner-Webb, B. (1984). High leveltransient expression of a chloramphenicol acetyl transferase gene by DEAEdextran mediated DNA transfection coupled with a dimethyl or glycerolshock treatment. Mud. Acids Res. 12: 5707-5717.134Lusby, E. W., and Berns, K. I. (1982). Mapping of the 5’ termini of two adenoassociated virus 2 RNAs in the left half of the genome. I. Virol. 41: 518-526.Lusby, E. W., Fife, K. H., and Berns, K. I. (1982). Nucleotide sequence of theinverted terminal repetition in adeno-associated virus DNA. I. Virol. 34: 402-409.Macejak, D. G., and Sarnow, P. (1991). Internal initiation of translationmediated by the 5’ leader of a cellular mRNA. Nature 353: 90-94.Madden, S. L., Cook, D. M., Morris, J. F., Gashler, A., Sukhatme, V. P., andRaucher, F. J. I. (1991). Transcriptional repression mediated by the WT1Wilms tumor gene product. Science 253: 1550-1553.Malim, M. H., Bohnlein, S., Hauber, J., and Cullen, B. R. (1989). Functionaldissection of the HIV-1 trans-activator-derivation of a trans-dominantrepressor of Rev function. Cell 58: 205-214.Malim, M. H., and Cullen, B. R. (1991). HIV-1 structural gene expressionrequires the binding of multiple Rev monomers to the viral RRE:Implications for HIV-1 latency. Cell 65: 241-248.Malim, M. H., Hauber, J., Le, S.-Y., Maizel, J. V., and Cullen, B. R. (1989). TheHIV-1 rev trans-activator acts through a structured target sequence to activatenuclear export of unspliced viral mRNA. Nature 338: 254-257.Malim, M. H., Tiley, L. S., McCarn, D. F., Rusche, J. R., Hauber, J., and Cullen,B. R. (1990). HIV-I structural gene expression requires binding of the Revtrans-activator to its RNA target sequence. Cell 60: 675-683.Maniatis, T. (1991). Mechanisms of alternative pre-mRNA splicing. Science251: 33-34.Martin, K. J., Lillie, J. W., and Green, M. R. (1990). Evidence for interaction ofdifferent eukaryotic transcriptional activators with distinct cellular targets.Nature 346: 147-152.Mathews, M. B., and Shenk, T. (1991). Adenovirus virus-associated RNA andtranslational control. J. Virol. 65(11): 5657-5662.Matson, S. W., Fay, P. J., and Bambara, R. A. (1980). Mechanism of inhibitionof the avian myeloblastosis virus deoxyribonucleic acid polymerase byadriamycin. Biochem. 19: 2089-2096.McDevitt, M. A., Hart, R. P., Wong, W. W., and Nevins, J. R. (1986).Sequences capable of restoring poly(A) site function define two distinctdownstream elements. EMBO J. 5: 2907-2913.135Mellon, P., Parker, V., Glutzman, Y., and Maniatis, T. (1981). Identification ofDNA sequences required for transcription of the human al-globin gene in anew SV4O host-vector system. Cell 27: 279-288.Melton, D. A., Krieg, P. A., Rebagliati, M. R, Maniatis, T., Zinn, K., and Green,M. R. (1984). Efficient in vitro synthesis of biologically active RNA and RNAhybridization probes from plasmids containing a bacteriophage SP6 promoter.Nucl. Acids Res. 12: 7035.Mermod, N., O’Neill, E. A., Kelly, T. J., and Tjian, R. (1989). The proline-richtranscriptional activator of CTF/NF-1 is distinct from the replication andDNA binding domain. Cell 58: 741-753.Miller, R. A., Ward, D. C., and Ruddle, F. H. (1977). Embryonal carcinoma cells(and their somatic cell hybrids) are resistant to infection by the murineparvovirus MVM, which does infect other teratocarcinoma-derived cell lines.I. Cell. Physiol. 91: 393-402.Mitchell, P. J., and Tjian, R. (1989). Transcriptional regulation in mammaliancells by sequence-specific DNA binding proteins. Science 245: 371-378.Mohanty, S., and Bachmann, P. (1974). Susceptibility of fertilized mouse eggsto minute virus of mice. Infect. Immun. 9: 762.Moreland, R. B., Langevin, G. L., Singer, R. H., Garcea, R. L., and Hereford, L.M. (1987). Amino acid sequences that determine the nuclear localization ofyeast histone 2B. Mol. Cell. Biol. 7(11): 4048-4057.Morey, A. L., Ferguson, D., and Fleming, K. A. (1991). Human parvovirus B19infection of fetal haemopoietic cells in vitro. 4th Parvovirus Workshop,Elsinore, Denmark, August 18-22.Morgan, W. R., and Ward, D. C. (1986). Three splicing patterns are used toexcise the small intron common to all minute virus of mice RNAs. J. Virol.60(3): 1170-1174.Mortimer, P. (1983). Hypothesis: the aplastic crisis of hereditary spherocytosisis due to a single transmissable agent. J. Clin. Pathol. 36: 445-448.Mortimer, P., Humphries, R., Moore, J., Purcell, R., and Young, N. (1983). Ahuman parvovirus-like virus inhibits haematopoietic colony formation invitro. Nature 302: 426-429.Mount, S. M. (1982). A catalogue of splice junction sequences. Nucl. AcidsRes. 10: 459-472.Mousset, S., and Rommelaere, J. (1982). Minute virus of mice inhibits celltransformation by simian virus 40. Nature 300: 537-539.136Nahreini, P., and Srivastava, A. (1991). Site-specific integration of the adenoassociated virus 2 genome into the chromosomal DNA of normal humandiploid cells. 4th Parvovirus Workshop, Elsinore, Denmark, August 18-22.Nevins, J. R. (1983). The pathway of eukaryotic mRNA formation. Ann. Rev.Biochem. 52: 441-466.OtMalley, R. P., Duncan, R. F., Hershey, J. W. B., and Mathews, M. B. (1989).Modification of protein synthesis initiation factors and shut-off of hostprotein synthesis in adenovirus-infected cells. Virology 168: 112-118.Ozawa, K., Ayub, J., Kajigaya, S., Shimada, T., and Young, N. (1988a). The geneencoding the nonstructural protein of 1319 (human) parvovirus may be lethalin transfected cells. J. Viral. 62: 2884-2889.Ozawa, K., Ayub, J., and Young, N. (1988b). Translational regulation of B19parvovirus capsid protein production by multiple upstream AUG triplets. I.Biol. Chem. 263(22): 10922-10926.Ozawa, K., Ayub, J., Yu-Shu, H., Kurtzman, G., Shimada, T., and Young, N.(1987). Novel transcription map for the B19 (human) pathogenic parvovirus.J. Viral. 61(8): 2395-2406.Ozawa, K., Kurtzman, G., and Young, N. (1986). Replication of the B19parvovirus in human bone marrow cell cultures. Science 233: 883-886.Ozawa, K., and Young, N. (1987). Characterization of capsid and noncapsidproteins of B19 parvovirus propagated in human erythroid bone marrow cellcultures. J. Viral. 61(8): 2627-2630.Parker, J. C., Collins, M. J., Jr., Cross, S. S., arLd Rowe, W. P. (1970). Minutevirus of mice. II. Prevalence, epidemiology, and occurrance as a contaminantof transplanted tumors. I. Natl. Cancer Inst. 45: 305-310.Pattison, J., Jones, S., Hodgson, J., Davis, L., White, J., Stroud, C., and Murtaza,L. (1981). Parvovirus infections and hypoplastic crisis in sickle-cell anaemia.Lan cet ii: 664-665.Pattison, J. R., Ed. (1988).t’Parvoviruses and Human Disease”. Boca Raton, FL,CRC Press Inc.Peabody, D. S. (1987). Translation initiation at an ACG triplet in mammaliancells. J. Biol. Chem. 262: 11847-11851.Pearson, W. R., and Lipman, D. J. (1988). Improved tools for biologicalsequence comparison. Proc. Nati. Acad. Sci. USA 85: 2444-2448.137Pelletier, J., and Sonenberg, N. (1988). Internal initiation of translation ofeukaryo tic mRNA directed by a sequence derived from poliovirus RNA.Nature 334: 320-325.Pintel, D., Dadachanji, D., Astell, C. R., and Ward, D. C. (1983). The genome ofminute virus of mice, an autonomous parvovirus, encodes two overlappingtranscription units. Nuci. Acids Res. 11(4): 1019-1038.Proudfoot, N. J., and Brownlee, G. G. (1976). 3’ Non-coding region sequencesin eukaryotic messenger RNA. Nature 263: 211-214.Rao, K., Patel, A., Anderson, M., Hodgson, J., Jones, S., and Pattison, J. (1983).Infection with parvovirus-like virus and aplastic crisis in chronic haemolyticanemia. Ann. Intern. Med. 98: 930-932.Reid, D., Brown, T., Reid, T., Rennie, J., and Eastmond, C. (1985). Humanparvovirus-associated arthritis: A clinical and laboratory description. Lancet i:422-425.Resnekov, 0., and Aloni, Y. (1989). RNA polymerase II is capable of pausingand prematurely terminating transcription at a presise location in vivo andin vitro. Proc. Nat!. Acad. Sci. USA 86: 12-16.Rhode, S. L, III, and Claussen, B. (1982). DNA sequence of the 5’ terminuscontaining the replication origin of parvovirus replicative form DNA. J.Virol. 41: 990-999.Rhode, S. L., III (1973). Replication process of the parvovirus H-I. I. Kinetics ina parasynchronous system. I. Viro!. 4: 856-861.Rhode, S. L., III (1987). Construction of a genetic switch for inducible transactivation of gene expression in eucaryotic cells. J. Virol. 61(5): 1448-1456.Rhode, S. L., III, and Richard, S. M. (1987). Characterization of the transactivation-responsive element of the parvovirus H-i P38 promoter. I. Virol.61(9): 2807-2815.Rhode, S. L., III. (1985). trans-activation of parvovirus P38 promoter by the76K noncapsid protein. I. Virol. 55(3): 886-889.Richards, R., Linser, P., and Armentrout, R. W. (1977). Kinetics of assembly ofa parvovirus, minute virus of mice, in synchronized rat brain cells. I. Virol.22: 778-793.Rodis, J., Hovick, T. J., Quinn, D., Rosengren, S., and Tattersall, P. (1988).Human parvovirus infection in pregnancy. Obstet. Gynecol. 72: 733-738.138Rogers, J., Early, P., Carter, C., Calame, K., Bond, M., and Wall, R. (1980). TwomRNAs with different 3’ ends encode membrane-bound and secreted formsof immunoglobulin mu chain. Cell 20: 303-312.Roller, R. J., and Roizman, B. (1991). Herpes simplex virus 1 RNA-bindingprotein U11 negatively regulates the accumulation of a truncated viralmRNA. J. Virol. 65(11): 5873-5879.Roy, S., Agy, M., Hovanessian, A. G., Sonenberg, N., and Katze, M. G. (1991).The integrity of the stem structure of human immunodeficiency virus type 1Tat-responsive sequence of RNA is required for interaction with theinterferon-induced 68,000 MW protein kinase. I. Virol. 65: 632-640.Russel, M., Kidd, S., and Kelley, M. (1986). An improved filamentous helperphage for generating single-stranded plasmid DNA. Gene 45: 333-338.Sadowski, I., and Ptashne, M. (1989). A vector for expressing GAL4(1-147)fusions in mammalian cells. Nuci. Acids Res. 17: 7539.Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T.,Mullis, K. B., and Erlich, H. A. (1988). Primer-directed enzymaticamplification of DNA with a thermostable DNA polymerase. Science 239: 487-491.Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). “Molecular Cloning, ALaboratory Manual”. N. Y., Cold Spring Harbor Laboratory Press.Samuiski, R. J., Zhu, X., Xiao, X., Brook, J. D., Housman, D. E., Epstein, N., andHunter, L. A. (1991). Targeted integration of adeno-associated virus (AAV)into human chromosome 19. EMBO J. 10(12): 3941-3950.Saris, C., Domen, J., and Berns, A. (1991). The pim-1 oncogene encodes tworelated protein-serine/threonine kinases by alternative initation at AUG andCUG. EMBO J. 10(3): 655-664.Schoborg, R. V., and Pintel, D. J. (1991). Accumulation of MVM gene productsis differentially regulated by transcription initiation, RNA processing andprotein stability. Virology 181: 22-34.Serjeant, G., Mason, K., Topley, J., Serjeant, B., Pattison, J., Jones, S., andMohamed, R. (1981). Outbreak of aplastic crisis in sickle-cell anaemiaassociated with parvovirus-like agent. Lan cet ii: 595-597.Shade, R. 0., Blundell, M. C., Cotmore, S. F., Tattersall, P., and Astell, C. R.(1986). Nucleotide sequence and genome organization of human parvovirusB19 isolated from the serum of a child during aplastic crisis. J. Virol. 58(3):921-936.139Siegl, G. (1984). Biology and pathogenicity of autonomous parvoviruses. in“The Parvoviruses”. New York, Plenum Press. 297-388.Siegi, G., Bates, R. C., Berns, K. I., Carter, B. J., Kelly, D. C., Kurstak, P., andTattersall, P. (1985). Characteristics and taxonomy of Parvoviridae.Intervirology 23: 61.Siegi, G., and Gautschi, M. (1973). The multiplication of parvovirus Lulli in asynchronized culture system. I. Biochemical characteristics of virusreplication. Arch. Gesamte Virusforsch 40: 119-127.Smith, D. B., and Johnson, K. 5. (1988). Single-step purification ofpolypeptides expressed in Escherichia coli as fusions with glutathione 5-transferase. Gene 67: 31-40.Soderlund, H., Pettersson, U., Vennstrom, B., Philipson, L., and Mathews, M.B. (1976). A new species of virus-coded low molecular weight RNA from cellsinfected with Ad2. Cell 7: 585-593.Sodroski, J., Patarca, R., and Rosen, C. (1985). Location of the trans-activatingregion on the genome of human T-cell lymphotrophic virus type III. Science229: 74-77.Sommer, L., Hagenbuchle, 0., Wellauer, P. K., and Strubin, M. (1991). Nucleartargeting of the transcription factor PTF1 is mediated by a protein subunit thatdoes not bind to the PTF1 cognate sequence. Cell 67: 987-994.Sosa, C. P., Mahony, J. B., Luinstra, K. E., Sternbach, M., and Chernesky, M. A.(1991). Replication and cytopathology of human parvovirus B19 in humanumbilical cord blood in erythroid progenitor cells, submitted for publication.Srivastava, A., and Lu, L. (1988). Replication of B19 parvovirus in highlyenriched hematopoietic progenitor cells from normal human bone marrow.J. Virol. 62(8): 3059-3063.Srivastava, A., Lusby, E. W., and Berns, K. I. (1983). Nucleotide sequence andorganization of the adeno-associated virus 2 genome. I. Virol. 45: 555-564.St. Amand, J., Beard, C., Humphries, K., and Astell, C.R. (1991). Analysis ofsplice junctions and in vitro and in vivo translation potential of the small,abundant B19 parvovirus RNAs. Virology 183: 133-142.Strebel, K., Daugherty, D., Clouse, K., Cohen, D., Folks, T., and Martin, M. A.(1987). The HIV ‘A’ (sor) gene product is essential for virus infectivity. Nature328: 728-730.Strebel, K., Klimkait, T., and Martin, M. A. (1988). A novel gene of HIV-I, vpu,and its 16-kilodalton gene product. Nature 241: 1221-1223.140Summers, J., Jones, S. E., and Anderson, M. J. (1983). Characterization of thegenome of the agent of erythrocyte aplasia permits its classification as ahuman parvovirus. I. Gen. Virol. 64: 2527.Summers, M. F. (1990). High-resolution structure of an HIV zinc fingerlikedomain via a new NMR-based distance geometry approach. Biochem. 29: 329-340.Takahashi, T., Ozawa, K., Mitani, K., Miyazono, K., Asano, S., and Takaku, F.(1989). B19 parvovirus replicates in erythroid leukemic cells in vitro. I. Infect.Dis. 160: 548-549.Takahashi, T., Ozawa, K., Takahashi, K., Asano, S., and Takaku, F. (1990).Susceptibility of human erythropoietic cells to B19 parvovirus in vitroincreases with differentiation. Blood 75(3): 603-610.Tattersall, P. (1972). Replication of the parvovirus MVM. Dependence of virusmultiplication and plaque formation on cell growth. I. Virol. 10: 586-590.Tattersall, P. (1978a). Parvovirus protein structure and virion maturation. in“Replication of Mammalian Parvoviruses”. Cold Spring Harbor, N.Y., ColdSpring Harbor Press.Tattersall, P. (1978b). Susceptibility to minute virus of mice as a function ofhost cell differentiation. in “Replication of Mammalian Parvoviruses’. ColdSpring Harbor, N. Y., Cold Spring Harbor Press.Tattersall, P., Cawte, P. J., Shatkin, A. J., and Ward, D. C. (1976). Threestructural polypeptides coded for by minute virus of mice, a parvovirus. I.Virol 29(1): 273-289.Tattersall, P., and Cotmore, S. (1988). The nature of parvoviruses. in“Parvoviruses and Human Disease”. Boca Raton, FL, CRC Press. 5-41.Tattersall, P., and Ward, D. C. (1976). Rolling hairpin model for replication ofparvovirus and linear chromosomal DNA. Nature 263: 106-109.Tennant, R., Layman, K., and Hand, R. (1969). Effect of cell physiological stateon infection by rat virus. I. Virol. 4: 872-878.Tijssen, P., Ed. (1990). “Handbook of Parvoviruses11 Boca Raton, FL, CRCPress.Tijssen, P., Arella, M., and Kawase, S. (1990). Molecular biology ofdensonucleosis viruses, in “Handbook of ParvovirusesttBoca Raton, FL, CRCPress. 283-291.Toolan, H. (1961). Studies on a viral agent associated with human tumors.Proc. Am, Assoc. Cancer Res. 3: 273.141Toolan, H. W. (1967). Lack of oncogenic effect of the H-viruses. Nature 214:1036.Toolan, H. W., and Ledinko, N. (1968). Inhibition by H-i virus of theincidence of tumors produced by adenovirus 12 in hamsters. Virology 35: 475-478.Toolan, H. W., Rhode, S. L., III, and Gierthy, J. F. (1982). Inhibition of 7,12-dimethylbenz(a)anthracene-induced tumors in Syrian hamsters by priorinfection with H-i parvovirus. Cancer Res. 42: 2552-2555.Towbin, H., Staehelin, T., and Gordon, J. (1979). Electrophoretic transfer ofproteins from polyacrylamide gels to nitrocellulose sheets: Procedure andsome applications. Proc. Nat!. Acad. Sci. USA 76: 4350-4354.Tratschin, J. D., Miller, I. L., and Carter, B. J. (1984). Genetic analysis of adenoassociated virus. Properties of deletion mutants constructed in vitro andevidence for an AAV replication function. J. Virol. 51: 611-619.Tratschin, J. D., Tal, J., and Carter, B. J. (1986). Negative and positve regulationin trans of gene expression from adeno-associated virus vectors inmammalian cells by a viral rep gene product. Mol. Cell. Biol. 6(8): 2884-2894.Tsao, J., Chapman, M., Agbandje, M., Keller, W., Smith, K., Wu, H., Luo, M.,Smith, T., Rossmann, M., Compans, R., and Parrish, C. (1991). The three-dimensional structure of canine parvovirus and its functional implications.Science 251: 1456-1464.Tullis, G. E., Labieniec-Pintel, L., Clemens, K. E., and Pintel, D. (1988).Generation and characterization of a temperature-sensitive mutation in theNS-1 gene of the autonomous parvovirus minute virus of mice. I. Virol. 62:2736-2744.Vieira, J., and Messing, J. (1987). Production of single-stranded plasmid DNA.Methods Enzymol. 153: 3Ward, D., and Tattersall, P., Ed. (1978). in uReplication of MammalianParvoviruses”. Cold Spring Harbor, N. Y., Cold Spring Harbor Press.Weiss, E. A., Gilmartin, G. M., and Nevins, J. R. (1991). Poly(A) site efficiencyreflects the stability of complex formation involving the downstreamelement. EMBO J. 10: 215-219.White, D., Mortimer, P., Blake, D., Woolf, A., Cohen, B., and Bacon, P. (1985).Human parvovirus arthropathy. Lancet i: 419-421.142Wickens, M., and Stephenson, P. (1984). Role of the conserved AAUAAAsequence: four AAUAAA point mutants prevent mRNA Y end formation.Science 226: 1045-1051.Wilson, G. M., Jindal, H. K., Yeung, D. E., Chen, W., and Astell, C. R. (1991).Expression of minute virus of mice major nonstructural protein in insectcells: purification and identification of ATPase and helicase activities.Virology 185: 90-98.Yaegashi, N., Shiraishi, H., Takeshita, T., Nakamura, M., Yajima, A., andSugamura, K. (1989). Propagation of human parvovirus B19 in primaryculture of erythroid lineage cells derived from fetal liver. I. Virol. 63(6): 2422-2426.Yakinoglu, A., Heilbronn, R., Burkie, A., Schiehofer, J., and zur-Hausen, H.(1988). DNA amplification of adeno-associated virus as a response to cellulargenotoxic stress. Cancer Res. 48: 3123-3129.Yakobson, B., Koch, T., and Winocour, E. (1987). Replication of adenoassociated virus in synchronized cells without the aid of a helper virus. I.Virol. 61: 972-981.Yanisch-Perron, C., Viera, J., and Messing, J. (1985). Improved M13 phagecloning vectors and host strains: Nucleotide sequences of the M13mp18 andpUC19 vectors. Gene 33: 103-119.Yanofsky, C. (1981). Attenuation in the control of expression of bacterialoperons. Nature 289: 751-758.Young, N. S., Mortimer, P. P., Moore, J. G., and Humphries, R. K. (1984).Characterization of a virus that causes transient aplastic crisis. I. Clin.Inves fig. 73: 224-230.Zhang, Y., and Moss, B. (1991). Vaccinia virus morphogenesis is interruptedwhen expression of the gene encoding an li-kilodalton phosphorylatedprotein is prevented by the Escherichia coli lac repressor. I. Virol. 65(11): 6101-6110.Zhou, S. Z., Srivastava, C. H., Munshi, N. C., and Srivastava, A. (1991).Parvovirus B19 replication in human cord blood cells: A possible mechanismof virus-induced hydrops fetalis. 4th Parvovirus Workshop, Elsinore,Denmark, August 18-22.11EcoRlm ><XbaI480AfIllZ600HIndillm C) o><-‘C- 101—1540—HindlilCM[1080Aval1HindIllZ2880HIncil3145PstIII II I3860PstI13900BamHII0C1WI aIC0;C,’‘.SllOEcoRl144B. Sequences of OligonucleotidesPCR Primers5’ TCTAGAATTCTCTTTCTGGGCTGCTTT 3’5? CGAGCATGCGTCGACAGGCAT173’5’ CGAGCATGCGTCGACAGGCA 3’5? TCGGATCCATGCAAAACAACACCACA 3’5? CCGAATTCTTTTAACCACAACAAATG YMutagenic Oligos5’P-CTCTACAGCTGCAAAAC 3’5’P-CACCACAGACCTGGATATGAA 3’5’P-AGACCTGGATCTGAAAAGCCT 3’5P-CATGGATATGTAAAGCCTGAAG 3’5’P-CTCTAGTATGCCCATCC YB19 Sequencing Oligos5’ AACCACCCCAATTGTCA 3?5? GTGCACACGGCTTTTGGCT 3?5? CATCTGTAGAGTTCACGA 3?5’ CAGGGGCAGCTGCACCTT 3?5? TAGTGGCCCTGGCATGA 3’oligo B19-27oligo dT 36-1MCS oligoBarn HI & 5’ end 94 aa OREEco RI & 3’ end 94 aa ORFmutate 1st ATG in 94 aa OREmutate 2nd ATG in 94 aa OREmutate 3rd ATG in 94 aa OREcreate stop codon in 94 aa OREmutate ACG to ATG in 687 ntcDNAtop strand (nt 1932-1948)bottom strand (nt 4780-4762)bottom strand (nt 4704/2183 jct)bottom strand (nt 3059-3042)bottom strand (nt 4003-3987)145C Nucleotide Sequences of Small B19 cDNAsSequence of 518 nt cDNAI 10 I 20 I 30 I 40 I 50CTTTCTGGGC TGCTTTTTCC TGGACTTTCT TGCTGTTTTT TGTGAGCTRACTRACRGGCG CCTGGAACRC TGAAACCCCG CGCTCTRGTR CGCCCATCCCCGGGACCAGT TCAGGAGAAT CRTTTGTCGG AAGCCCAGTT TCCTCCGAAGTTGTRGCTGC RTCGTGGGAA GRRGCCTTCT ACRCRCCTTT GGCAGACCAGTTTCGTGAAC TCTACRGBTh. CAFIAACARCfl CCACAGACRT_GGRTfiI..AAAAGCCTGARGA RTTGTGGACA GCCRAARGCC GTGTGCRCCC ATTGTRAACACTCCCCACCG TGCCCTCRGC CRGGRTGCGT RACTRARCGC CCACCAGTRCCRCCCRGACT GTACCTGCCC CCTCCTGTAC CTATARGRCR GCCTAACACRRARGATATAG ACRATGTAGA ATTTAAGTRC TTAACCAGAT FiTGAACFiACATGTTATTRGA RTGTTAAGAT TGTGTAATRT GTATCRRRRT TTAGAAAAATAAACATTTGT TGTGGTTA 516I 10 I 20 I 30 I 40 I 50146Sequence of 638 nt cDNAI 10 I 20 I 30 I 40 I 50CTTTCTGGGC TGCTTTTTCC TGGACTTTCT TGCTGTTTTT TGTGAGCTRRCTAACAGATG CCCTCCACCC AGFICCTCCAA ACCRCCCCAA TTGTCRCAGRCACCAGTATC RGCAGCRGTG GTGGTGRAFIG CTCTGRAGAR CTCRGTGARAGCAGCTTTTT TRACCTCATC ACCCCAGGCG CCTGGAACAC TGRARCCCCGCGCTCTAGTR CGCCCATCCC C666RCCAGT TCAGGRGRAT CATTTGTCGGAAGCCCRGTT TCCTCCGARG TTGTAGCTGC RTCGTGGGAA GAAGCCTTCTACACACCTTT GGCRGRCCAG TTTCGTGAAC TCTACAGffI. CRARACRACACCACAGACAT_GGATff[.AAA AGCCTGRAGR ATTGTGGACA GCCRRRRGCCGTGTGCACCC RTTGTAAACA CTCCCCACCG TGCCCTCAGC CRGGATGCGTARCTAAACGC CCACCAGTAC CACCCAGACT GTRCCTGCCC CCTCCTGTACCTATAAGACA GCCTAACACA ARAGATATAG ACARTGTAGA ATTTRAGTACTTAACCRGAT RTGAACAACA TGTTATTAGA RTGTTAAGAT TGTGTARTATGTATCAAAAT TTAGRAAAAT ARRCRTTTGT TGTGGTTA 638I 10 I 20 I 30 40 I 50147Sequence of 687 nt cDNAI 10 I 20 I 30 I 40 I 50CTTTCTGGGC TGCTTTTTCC TGGACTTTCT TGCTGTTTTT TGTGAGCTAACTFIRCAGGCG CCTGGRRCAC TGFIRACCCCG CGCTCTRGTA_CGCCCATCCCCGGGRCCAGT TCRGGAGART CATTTGTCGG RRGCCCAGTT TCCTCCGAAGTTGTAGCTGC ATCGTGGGRA GAAGCCTTCT ACACACCTTT GOCAGACCAGTTTCGTGRRC TGTTAGTTGG GGTTGATTAT GTGTGGGACG GTGTAAGGGGTTTRCCTGTG TGTTGTGTGC AACRTRTTAA CAATAGTGGG GGAGGGTTGGGACTTTGTCC CCATTGCATT ARTGTAGGGG CTTGGTATRA TGGATGGAAATTTCGAGAAT TTACCCCAGA TTTGGTGCGA TGTAGCTGCC ATGT666AGCTTCTRATCCC TTTTCTGTGC TAACCTGCRA AARATGTGCT TACCTGTCTGGATTGCAAAG CTTTGTAGAT TATGAGTAAA RAAAGTGGCA AATGGTGGGAAAGTGATGRT AAATTTGCTA ARGCTGTGTR TCAGCAATTT GTGGAATTTTATGARAAGGT TACTGGAACA GRCTTAGFIGC TTATTCAAAT RTTAAAAGRTCATTATRATFI TTTCTTTAGA TRATCCCCTA GAAAACCCAT CCTCTCTGTTTGACTTAGTT GCTCGTATTR AAAATARCCT TAAAAAC 687I 10 I 20 I 30 I 40 I 50148Sequence of 807 nt cDNAI 10 I 20 I 30 ‘0 I 50CTTTCTGGGC TGCTTTTTCC TGGRCTTTCT TGCTGTTTTT TGTGRGCTRRCTRACAGflI. CCCTCCACCC AGACCTCCRA ACCACCCCAA TTGTCACFIGRCRCCAGTRTC AGCRGCRGTG GTGGTGAAAG CTCTGF1RGAA CTCAGTGAAAGCAGCTTTTT TAACCTCRTC ACCCCFIGGCG CCTGGAACAC TGRRACCCCGCGCTCTAGTR_CGCCCATCCC C606ACCAGT TCRGGAGFIAT CATTTGTCGGRRGCCCAGTT TCCTCCGRAG TTGTAGCTGC ATCGTGGGFIFI GRRGCCTTCTACRCACCTTT GGCAGACCRG TTTCGTGARC TGTTAGTTGG GGTTGRTTATGTGTGGGACG GTGTAAGGGG TTTACCTGTG TGTTGTGTGC ARCATATTARCAATAGTGGG GGA666TTGG GACTTTGTCC CCATTGCATT AATGTAGGGGCTTGGTATAA TGGRTGGARA TTTCGRGAAT TTRCCCCAGA TTTGGTGCGRTGTAGCTGCC ATGTGGGAGC TTCTRATCCC TTTTCTGTGC TAACCTGCAAAAAATGTGCT TACCTGTCTG GRTTGCAAAG CTTTGTRGRT TATOROTARAARRRGTGGCA ARTGGTGGGR AAGTGATGAT AAATTTGCTA AAGCTGTGTATCRGCAATTT GTGGRRTTTT ATGRAARGGT TRCTGGAACR GACTTAGAGCTTATTCRAAT RTTRAAAGAT CATTATRATA TTTCTTTAGA TRRTCCCCTAGARAACCCRT CCTCTCTGTT TGACTTRGTT GCTCGTRTTA ARARTAACCTTRARRAC 807I 10 I 20 I 30 I 40 I 50149D. Sequences of Peptides Used to Generate B19 Antisera in RabbitsPeptide 1N - Pro - Asn - Thr- Lys - Asp- Tie - Asp - Asn - Val - Glu -Phe-Lys-Tyr-Leu-Thr-Arg-Tyr-Glu-Gln-HisVal - Tie- Arg - Met - Leu- Arg - Leu- Cys- CPeptide 2 *N - Ala - Ser - Trp - Glu- Glu - Ala - Phe- Tyr - Thr - ProLeu-Ala-Asp-Gln-Phe-Arg-Giu-Leu-Gly-GlyCys-C* the last three amino acids do not correspond to B19 sequence


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items