UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Interaction of complement with human cytomegalovirus Spiler, Owen Bradley 1994

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

Item Metadata


831-ubc_1995-983493.pdf [ 4.54MB ]
JSON: 831-1.0088915.json
JSON-LD: 831-1.0088915-ld.json
RDF/XML (Pretty): 831-1.0088915-rdf.xml
RDF/JSON: 831-1.0088915-rdf.json
Turtle: 831-1.0088915-turtle.txt
N-Triples: 831-1.0088915-rdf-ntriples.txt
Original Record: 831-1.0088915-source.json
Full Text

Full Text

INTERACTION OF COMPLEMENT WITHHUMAN CYTOMEGALOVIRUSbyOWEN BRADLEY SPILLERB. M. L. Sc., The University of British Columbia, 1989A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHThOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF PATHOLOGYWe accept this thesis as conformingto the required standardTHE UNWERSITY OF BRITISH COLUMBIADecember 1994© Brad Spiller, 1994In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of_____________The University of British ColumbiaVancouver, CanadaDate____DE-6 (2188)ABSTRACT.In the absence of specific anti-human cytomegalovirus (HCMV)antibodies, complement has a negligible neutralizing effect on the virions.Under these conditions, the presence of activated C3 fragments, but not C9, werefound on the HCMV virions. There are no known viral complement inhibitorsencoded in the HCMV genome, but the presence of host-encoded complementinhibitors, CD55, CD46, and CD59, on the HCMV virions may explain thevirion’s ability to regulate complement. In the presence of anti-HCMVantibodies, complement enhanced the neutralizing ability of theimmunoglobulins by 2-3 fold. Complement activation in the presence ofantibodies occurred primarily by the classical complement activation pathwayand was complete, as assessed by the presence of C9.CD55 (decay-accelerating factor) and CD46 (membrane co-factor protein)regulate complement at the level of C3 and belong the regulators ofcomplement activation (RCA) gene cluster. CD59 regulates complement at thelevel of the terminal lytic pathway and does not belong to the RCA gene cluster.HCMV infection of fibroblasts and glioblastoma cells resulted in a 3-8 foldincrease in the expression of CD55 and CD46 by 72 h p.i., but infection withherpes simplex virus or adenovirus had no effect. An increase in C3 convertaseregulation was also found on the HCMV-infected cells using a purifiedcomplement component assay. By contrast, CD59 expression was decreased onHCMV-infected cells by 50% by 72 h p.i., similar to the decrease observed forHLA class I.CD55 expression was increased in HCMV-infected cells at the level ofprotein, mRNA, and transcription of the gene as assessed by using a variety oftechniques. Similarly, the levels of CD59 mRNA decreased in the HCMVinfected cells, paralleling the observations for protein expression, but radiolabel11pulse-chase analysis identified a decreased CD59 protein survival. Indirectevidence suggested that the immediate early or early genes from HCMV wereresponsible for the altered expression of host complement inhibitors, but CD55promoter activity and protein expression were unaffected by the presence ofisolated HCMV immediate early genes.111Table of ContentsAbstract iiTable of Contents ivList of Tables viiiList of Figures ixList of Abbreviations xiAcknowledgment xviiDedication xviii1. BACKGROUND 11.1. HCMV biology and lifecycle 21.2. Cell types infected by HCMV 51.2.1 Post-transplantation/transfusion HCMV infection 61.2.2 In vitro HCMV infection of hematologic cells 71.2.3 Infection of non-hematologic cell types 91.3. In vivo control of HCMV infection 101.3.1. Cellular immune response to HCMV infection 101.3.2. Humoral immune response to HCMV infection 111.3.3 Complement and antibody-mediated HCMV neutralization 121.3.4 The complement cascade 121.4. Host-encoded complement inhibitors 151.4.1. The regulators of complement activation (RCA) gene cluster. .. CD55 or decay-accelerating factor (DAF) CD46 or membrane cofactor protein (MCP) CD35 or complement receptor 1 (CR1) CD21 or complement receptor 2 (CR2) Remaining members of the RCA gene cluster 221.4.2 CD59 or membrane inhibitor of reactive lysis (MIRL) 23iv1.4.3 C8 binding protein (C8bp) or homologous restriction factor 241.5. Complement-mediated cytolysis and virus-infected cells 241.5.1 Soluble vaccinia complement inhibitor 251.5.2 Herpesviridae complement inhibitors 251.5.3 Complement regulation on HCMV-infected cells 261.6. Alterations in host protein expression induced by HCMVinfection 272. MATERIALS AND METHODS 312.1. Materials 312.2. Cells and viruses 312.3. Antibodies 322.4. Serum samples 332.5. IgG depletion from serum samples 342.6. Virus Infection 352.7. Standard Plaque Assay 362.8. Functional neutralization assay 372.9. Western blot analysis 382.10. Complement component co-purification with HCMV virions 392.11. Immunocytochemistry 412.12. Flow cytometry studies 422.13. Northern blot analysis 432.14. Radio-immunoprecipitation studies 442.15. Alternative pathway C3 convertase activity assay 452.16. CD55 promoter constructs 462.17. Other plasmids 472.18. Plasmid transfection into cells 472.19. Measurement of CAT activity 48v2.20. p3-gal activity assay.493. RESULTS 503.1 CHAPTER 1. Interaction of complement with HCMV virions 503.1.1 The role of complement in neutralization of HCMV virions. ...503.1.2 Neutralizing titer correlation with amount of specific antiHCMV antibody 573.1.3 Interaction of complement with purified virions 573.1.4 Host Complement Inhibitors associated with HCMV Virions. ..663.2 CHAPTER 2. Changes in complement inhibitor expression onadherent cells infected with HCMV 693.2.1 Complement inhibitor expression on uninfected cells 693.2.2 Complement inhibitor expression alteration with HCMVinfection 723.2.3 Viral specificity of CD55/CD46 increased expression 783.2.4 Increased CD55 is of host origin 803.2.5 Functional properties of increased CD55 on HCMV infectedcells 823.3. CHAPTER 3. Effects of HCMV infection on THP-1 cells 843.3.1. HCMV infection of undifferentiated THP-1 cells 853.3.2. HCMV infection of phorbol ester-differentiated THP-1 cells 913.4. CHAPTER 4. Mechanism of CD55 expression increase 953.4.1 Northern blot analysis of HCMV-infected cells 953.4.2. Further investigation of decreased CD59 expression onHCMV-infected cells 983.4.3. Requirements for CD55 promoter activity 1013.4.4. Effect of HCMV infection on CD55 promoter constructs 1033.4.5. Effect of HCMV late gene repressor on CD55 upregulation 109vi3.4.6. No Upregulation of CD55 promoter by isolated HCMV earlygenes 1124. DISCUSSION 1165. SUMMARY 1296. REFERENCES 130viiList of TablesTable 1. Monoclonal antibody description and sources 33Table 2. Mean cellular fluorescence of non-specific antibody bindingto uninfected and HCMV-infected cells 76Table 3. Flow cytometry analysis of fibroblasts under different viral andchemical exposure 79Table 4. HCMV associated with THP-1 cells 84Table 5. Mean cellular fluorescence of non-specific antibody binding touninfected and HCMV-infected cells 89Table 6. Transient transfection assessment for co-transfection of the CD55promoter/reporter and isolated HCMV IE genes 114Table 7. Flow cytometry analysis of CD55 expression following transienttransfection with isolated HCMV TE genes 115viiiList of FiguresFig. 1: Human cytomegalovirus and an infected cell 4Fig. 2: Human cytomegalovirus lytic cycle 5Fig. 3: Complement cascade 14Fig. 4: Neutralization assay with seronegative serum 52Fig. 5: Effect of complement on seronegative serum neutralization 53Fig. 6: Neutralization assay with seropositive serum 54Fig. 7: Effect of complement on seropositive serum neutralization 55Fig. 8: Plaque assay following pre—incubation with serum samples 56Fig. 9: Comparison of ELISA absorbance and extinction dilution 58Fig. 10: Comparison of serum neutralization and antibody titer 59Fig.11: Depletion and elution of IgG using protein G (Western blot) 60Fig. 12: IgG bound to HCMV virions under varying conditions(Western blot) 62Fig. 13: C3 bound to HCMV virions under varying conditions(Western blot) 63Fig. 14: C9 bound to HCMV virions under varying conditions(Western blot) 65Fig. 15: Host complement inhibitors on HCMV virions (Western blot) 68Fig. 16: Indirect immunofluorescence analysis of fibroblasts 70Fig. 17: Indirect immunofluorescence analysis of glioblastoma cells 71Fig. 18: Comparison of complement inhibitor Mr (Western blot) 72Fig. 19: Composite flow cytometry histograms for CD55 74Fig. 20: Effect of HCMV infection on protein expression for fibroblasts.(flow cytometry) 75Fig. 21: Effect of HCMV infection on protein expression for glioblastomacells. (flow cytometry) 78ixFig. 22: Confirmation of host source of CD55 (flow cytometry) 81Fig. 23: Confirmation of host source of CD55 (Western blot) 82Fig. 24: Decreased C3 convertase activity on HCMV-infected cells 83Fig. 25: Effect of HCMV infection on protein expression forundifferentiated THP-1 cells (flow cytometry) 87Fig. 26: Effect of HCMV infection on CR1 expression forundifferentiated TI-IF-i cells (flow cytometry) 90Fig. 27: Effect of phorbol esters on protein expression duringdifferentiation of TFIP-1 cells (flow cytometry) 92Fig. 28: Effect of HCMV infection on protein expression fordifferentiated THP-i cells (flow cytometry) 93Fig. 29: Effect of HCMV infection on CR1 expression for differentiatedTHP-i cells (flow cytometry) 94Fig. 30: Effect of HCMV infection on mRNA levels (Northern blot) 96Fig. 31: Effect of HCMV infection on mRNA levels over time(Northern blot) 97Fig. 32: Effect of HCMV infection on CD59 synthesis (autoradiograph) 99Fig. 33: Effect of HCMV infection on CD59 expression (flow cytometry) 101Fig. 34: Plasmid containing CD55 promoter and reporter gene 102Fig. 35: Basal CD55 promoter element requirements in HeLa cells 103Fig. 36: Effect of HCMV infection on full-length CD55 promoter 105Fig. 37: Basal CD55 promoter element requirements in HCMV-infectedcells 106Fig. 38: No effect of HCMV incubation with CD55 promoter inHeLa cells . . . .108Fig. 39: Effect of PAA on CD55 promoter activity in HCMV-infected cells 110Fig. 40: Effect of PAA on protein expression in HCMV-infected cells.(flow cytometry) 111xList of AbbreviationsAbs absorbanceAD169 common lab strain of HCMVAd5 human adenovirus type 5Ag antigenANOVA analysis of varianceAst astrocyte(s)31H factor H (soluble complement regulator)bp basepair(s) of nucleic acid(s)BSA bovine serum albuminC’ complementC3, etc third complement component, etcC3b large activation fragment of C3C3dg inactive, covalently-bound C3 fragment, CR2 ligandC4b large activation fragment of C4C4d inactive, covalently-bound C4 fragmentC4bp C4-binding proteinC8bp C8-binding proteinCAT chloramphenicol acetyl-transferaseCD cluster designation (international protein classification)CD4 cell-surface marker for T-cell (helper subtype) and monocytesCD8 cell-surface marker for T-cell (cytotoxic/suppressor subtype)CD34 hematologic stem cell markerCD35 complement receptor 1 (CR1)CD46 Membrane co-factor protein (MCP)CD55 decay-accelerating factor (DAF)CD59 Membrane inhibitor of reactive lysis (MIRL)xicDNA complementary DNA from mRNACML chronic myelogenous leukemiaCMV cytomegalovirusQ’E cytopathic effectCPM counts per minuteCR1 complement receptor 1 (CD35)CR2 complement receptor 2 (CD22)CR3 complement receptor 3 (CD11b/CD18)CSF cerebrospinal fluidCTP cytosine triphosphateD day(s)DH5o competent E. coli for transformationDHFR dihydrofolate reductaseDMEM Dulbecco’s minimum essential mediumDNA deoxyribonucleic acidE early HCMV gene(s)E2F eukaryotic transcription factor 2EDTA ethylenediaminetetraacetic acidEGTA ethyleneglycol-bis-(beta-aminoethylester) tetraacetic acidELAM endothelial-leukocyte adhesion moleculeELISA enzyme-linked immunosorbent assayFab variable, antigen-binding region of an immunoglobulinFc constant region of an immunoglobulinFC flow cytometry solutionFACS fluorescence activated cell sortingFBS fetal bovine serumFITC fluorescein isothiocyanatexiifos a human oncogenegB glycoprotein B (major envelope protein for CMV or HSV)gC glycoprotein C (major envelope protein for HSV)gH glycoprotein H (major envelope protein for CMV or HSV)gp4l part of the major HIV virion envelope proteingpll5 part of the CMV gB virion envelope proteingp350 EBV complement regulating proteinGPI glycophosphoinositolgpffl synonym for CMV g13 virion envelope proteinGVB gelatin containing veronal-buffered saline solutionh hour(s)HIV human immunodeficiency virusHLA human leukocyte antigenh p.i. hours post infectionHSV-1 herpes simplex virus type 1ICAM member of the integrin family of proteins]E immediate early HCMV gene(s)JE1 immediate early HCMV gene #1Ig immunoglobulinIgG gamma class of immunoglobulin1gM mu class of immunoglobulinIL interleukinInab specific CD55-deficient phenotypemt internationalINF interferonIRS internal repeat sequencekb kilobase pairsxli’kDa kilodaltonsL late HCMV gene(s)Lac Z 3-galactosidase geneLPS bacterial cell wall lipopolysaccarideM molarMAC membrane-attack complexMan mannosemCi millicurries (radioactivity)MEM minimal essential mediumMHC major histocompatibility antigen (mouse HLA)mm minute(s)MIRL synonym for CD59ml milliliter(s)m m millimeter(s)mM millimolarMOl multiplicity of infectionMono monocyteMr relative molecular mobilitymRNA messenger ribonucleic acidmtr group of immortalizing genes found in the HCMV genomemyc a human oncogeneN a sodium saltng nanogram(s)NK natural killer cellsn m nanometer(s)NHS normal human serumpl5E complement activating protein from C-type retrovirusesxivp55 part of the CMV gB virion envelope proteinp86 the CMV gH virion envelope proteinp130 part of the CMV gB virion envelope proteinPAA phosphoacetic acidPAGE polyacrylamide gel electrophoresisPBM peripheral blood mononuclear cellPBS phosphate buffered salinePHA phytohemaglutaninp.i. post-infectionP]PLC phosphoinositol-specific phospholipase CPFU plaque forming unit(s)Plt platelet(s)PMN polymorphonuclear cellsPNH paroxysmal nocturnal hemoglobinuriaRab C’ rabbit serum containing active rabbit complementRc/CMV eukaryotic expression vectorRCA regulators of complement activation (gene cluster)RNA ribonucleic acidrpm revolutions per minuteRSB reticulocyte standard bufferSCR short consensus repeatSDS sodium dodecylsulfateSIV simian immunodeficiency virusSW3OTi/SW5O swinging bucket rotors rated for 30 K and 50 K respectivelyTBS tris-buffered salineTBST tris-buffered saline containing 0.05% Tween 20 detergentTHP-1 monocytic leukemia cell linexvTRS terminal repeat sequencesU unitsU373-MG glioblastoma cell lineUL unique long region of the cytomegalovirus genomeU unique short region of the cytomegalovirus genomeug microgram(s)ul microliter(s)U V ultra violet lightVB S veronal-buffered salinevol volumew/v weight/volumew/w weight/weightX-gal chromogenic substrate for f3-galactosidasexviAcknowledgmentsI wish to acknowledge; My wife Linda for her patience and support,Frank Tufaro, Xiaoning Wu, Melanie Hanna, Doug Lublin, Roger Lippe andWilf Jefferies for their contributions to data presented in this thesis, themembers of Dana Devine’s and Frank Tufaro’s laboratory past and present(especially Jihan Marjan and Bruce Banfield) for helpful discussions, themembers of my advisory committee: Dr’s Walker, Thomas and Stiver for soundadvice through the years, Dr O’Kusky for providing me with motivation andtesting my personal resolve to it’s limits, and Dr. Robert Fulghum for teachingme that the unorthodox approach is often the most productive. Most of all Iwish to thank Dr Dana Devine for her willingness to take in stray cats andgraduate students and the opportunity to work under her tutelage.xviiDeIication;This thesis is dedicated to Dana Devine, who always believed; to my wife Lindawho always faced my difficulties with me; and to my parents, Bob and Lorna,without whose constant support and encouragement this would not have beenpossible.xviii1. BACKGROUND:This dissertation focuses on a proposed mechanism by which humancytomegalovirus (HCMV)-infected cells evade clearance by specific antibodies andcomplement, and briefly addresses the potential role of complement in the in vivoclearance of extracellular HCMV virions. HCMV, an enveloped, double-strandedDNA virus, is a member of the herpesvirus family, and as such replicates in thenucleus and is capable of latent infection. The means by which the virus evades thehost immune response during latent infection is not well understood. HCMV is aubiquitous viral agent. Depending on the population studied, up to 100% ofindividuals may have been infected with the HCMV by middle-age, most without anyobvious signs or symptoms (Borysiewicz et al 1983, 1986, Schrier et al 1986, Riddell etal 1991). HCMV antibody prevalence rates in North America range from 30% to 80%,while some areas in Africa and the Far East reach 100% seropositivity (Tegtmeier1989, Ho 1991). Asymptomatic infections in normal individuals are oftenaccompanied by virus shedding in urine, saliva, semen, breast milk, and virus canoccasionally be isolated from blood. HCMV also accounts for about 15% of all cases ofmononucleosis, a non-life threatening and self-resolving illness (Rinaldo et al 1980,Kaariainen et al 1966, Klemola et al 1965).The most severe pathologies are associated with congenital HCMV infection orinfection during immunosuppression. Approximately 1-2% of infants are infected inutero and another 6-60% (depending on geographic location) become infected duringthe first 6 months of life as the result of birth canal or breast milk transmissions (Ho1991, Reynolds et al 1973). HCMV transmission to the fetus during gestation canoccur during primary maternal infection or during re-infection/reactivation of latentHCMV infection. However, the pathological consequences seem to be moreassociated with a maternal primary infection (Ho 1991). Although HCMV is aubiquitous viral agent and evidence of infection can approach 100% of certainpopulations, there are some rare diseases in which HCMV has been speculated to play1a pathological role. These include various cancers, atherosclerosis, Guillain-Barrésyndrome, Charcot-Marie-Tooth disease, post-perfusion syndrome, diabetes mellitus,hepatitis, hemolytic anemia! thrombocytopenia, and gastrointestinal disease(reviewed in Huang and Kowalik 1994).Other than congenital HCMV infection, the other large patient group at risk forcomplications arising from HCMV infection of immunosuppressed individuals.Patients can become immunosuppressed by HIV infection or iatrogenically, throughorgan or bone-marrow transplantation procedures. The transplantation group is ofmajor concern, since surgical procedures and post-transplant care can require thetransfusion of large numbers of blood products. The ability of HCMV to be carriedalong with these blood products appears to be quite efficient (Bowden 1991). Besidesthe concerns associated with patient morbidity and mortality, HCMV infection hasbeen reported to be associated with a higher rate of graft rejection or the developmentof graft-versus-host disease (Lopez et al 1974, Lonnqvist et al 1984, Rinaldo et al 1980).Furthermore, many strategies are being developed to decrease the risk of HCMVinfection from blood transfusion including transfusing only blood products that donot contain anti-HCMV antibodies or filtering out the leukocytes, which seem toharbor infectious HCMV (Bowden 1991). All of these modalities are costly: massscreening of blood products, filters, infusion of intravenous immunoglobulins, post-rejection complications and re-transplantation. Further understanding of the HCMVbiology is necessary before one can develop more effective strategies to reducemorbidity and mortality associated with HCMV infection.1.1 HCMV biology and life cycle.HCMV is a member of the herpesvirus family. Inclusion into the familyHerpesviridae is based on the architecture of the virion. A typical herpesvirionconsists of (1) a core containing a linear, double-stranded DNA, (2) anicosadeltahedral capsid, containing 162 capsomeres with a hole running down the2long axis, (3) an amorphous, sometimes asymmetric material that surrounds thecapsid, designated the tegument, and (4) an envelope containing viral glycoproteinspikes on its surface (Roizman 1990; Figure 1A).HCMV is the largest of the herpesvirus family; an HCMV virion is between150-200 nm and has a genome of approximately 230 kilobase pairs. Analysis of thegenome of the AD169 laboratory strain of HCMV has identified 208 open-readingframes and potential protein products (if open reading frames are limited to <300 bp;Chee et al 1990). The HCMV genome has a complex arrangement; sequences fromboth termini are repeated in an inverted orientation and juxtaposed internally,dividing the genome into two components, each of which consists of uniquesequences flanked by inverted repeats. Therefore, since both components can invertrelative to each other, extracted DNA from virions exists as four equimolar isomers.Since not all herpesviruses have the ability to isomerize, the HCMV genome isreferred to as a type E genome which indicates this ability (Roizman 1990). HCMV isfurther classified into the Betaherpesvirinae subfamily, which indicates a restrictedhost range (non-exclusive to this subfamily), a long reproductive cycle, ability toestablish a latent infection, hematogenous spread and possible site of latency, and thatinfected cell frequently become enlarged (cytomegalia).While the cellular receptor for HCMV has not yet been elucidated, the viralenvelope glycoproteins gH (or p86; see section 1.3.2) and gB (or p130/55; see section1.3.2) have been reported to be responsible for HCMV ligand-cellular receptorinteractions (Rasmussen 1991). Upon absorption, uncoating, and entry of viral DNAinto the nucleus, expression of the HCMV genome is sequentially regulated (Figure1B). According to the kinetics of gene expression, the genes encoded by HCMV canbe categorized into three kinetic classes: immediate-early (IE), early (E), and late (L)genes (Stinski 1990; Figure 2). IE genes of HCMV encode a group of regulatoryproteins with strong transactivating activities. IE expression is not only needed toactivate subsequent E gene expression, but the IE gene products also transactivate the3A. Envelope (lipid bilayer)______Glycoproteins‘-7/ DNA (inside/ capsid)TegumentCapsid (composed of162 capsomeres)Cross-sectionthrough capsidFigure 1. (A) Schematic representation of the herpesvirion, seen through a crosssection of the envelope with spikes (glycoproteins) protruding from its surface. (B)Diagrammatic representation of the HCMV replication cycle. The abbreviationsdenote the immediate early (IE), early (E), and late (L) gene expression.4B.‘ AdherencePenetration Gene ExpressionUncoatingCapsid Assembly EgressQEp\Circularization DNA replicationBuddin0.—4k$I/272 Hrs - Extracellular VirionsFigure 2. Diagrammatic representation of the HCMV replication cycle in humanfibroblasts.promoters of certain cellular genes (see section 1.6). HCMV E genes encode enzymesand factors involved in viral DNA replication and aid in the regulation of late geneexpression. HCMV L genes are expressed following the initiation of HCMV DNAreplication and these genes encode the structural components of the virion. Capsidassembly occurs in the nucleus and infectious intracellular HCMV can be detected at55 h post-infection. The release of extracellular HCMV can be detected after 72 h post-infection in vitro and continues until cell death.1.2 Cell types infected by HCMV.The CMV family of viruses are species specific. With few exceptions CMVfrom one species will not infect cells from another (Plummer 1973). Furthermore,unlike HSV, another member of the herpesvirus family, HCMV will not infect all celltypes isolated from a permissive host. Some of the most compelling evidence for inI I I0 12 245vivo cell specificity comes indirectly from investigation of HCMV-associatedpathology of immunosuppressed patients.1.2.1 Post-transplantation/transfusion HCMV infection.The largest body of evidence for the site of HCMV persistence or latency wasidentified serendipitously via transmission of HCMV through blood products totransplant recipients. HCMV-seronegative transplant recipients can acquire primaryHCMV infection from either contaminated blood products or from the transplantedorgan (or bone marrow) itself, if these are obtained from HCMV seropositiveindividuals. Individuals who have serological evidence of previous exposure toHCMV, on the other hand, can reactivate latent HCMV or be re-infected. Eighteen toeighty-three percent of HCMV-seronegative transplant recipients who receive bloodproducts or allografts from HCMV-seropositive donors will become infected withHCMV depending on the transplantation setting (data summarized in Bowden 1991).Furthermore, 70% of seropositive patients, for whom no attempt at HCMV screeningof products is made, have evidence of active HCMV infection post-transplantation,but the origin of the HCMV was not elucidated (Bowden et al 1986). This contrastssharply to the 0-6% incidence of HCMV infection of HCMV-seronegative recipientsreceiving transplants and blood products from HCMV-seronegative donors (datasummarized in Bowden 1991). However, even these HCMV infections in seronegativerecipients receiving products from seronegative donors may reflect the limitations inthe screening procedure; HCMV DNA has been reported occasionally in monocytes inthe absence of detectable specific antibody (Taylor-Wiedeman et al 1991). It is therelatively high rate of transmission of HCMV via blood products which lends supportto the hypothesis that a major latency reservoir for HCMV is contained within theblood. The decrease in HCMV transmission following leukofiltration of bloodproducts as well as reports that HCMV can be found in buffy coat preparations fromblood donors, congenitally infected children, and transplant recipients strongly6suggests the mononuclear peripheral blood leukocyte fraction is the site of latency forHCMV (Diosi et al 1969, Kaariainen et al 1966, Lang and Noren 1968, Fiala et al 1975,Smith et al 1993). One multi-center controlled study reported that 21% of newborninfants transfused with unfiltered blood products developed HCMV infection,whereas a group of infants transfused with leukodepleted blood products did notdevelop HCMV infection (Gilbert et al 1989). The rare transmission of HCMV whenmeasures are taken to reduce the transmission via blood products may suggest thatthere are additional cell types involved in HCMV latency/persistence.1.2.2 In vitro HCMV infection of hematologic cells.Several in vitro studies have attempted to elucidate the particular cell which isinfected in the leukocyte fraction. At the least differentiated level, there is someevidence that clonogenic bone marrow progenitors (CD34+ cells) are permissive forHCMV infection (Maciejewski et al 1992). However, even though HCMV virionsappear to enter a large number of CD34+ cells and some viral progeny is produced,the infected cells quickly lose their expression of the CD34 antigen, indicatinginduction of cellular differentiation, prior to expression of HCMV late genes. Ofinfected CD34+ cells which express HCMV late genes following loss of the CD34antigen, all of the productively infected cells were found to belong to themyelomonocytic lineage (Maciejewski et al 1992). These findings provide additionalevidence that the cells related to the leukocyte fraction are responsible for harboringHCMV. Additionally, the use of polymerase-chain reaction techniques recentlyidentified the presence of HCMV DNA, which does not necessarily indicate aproductive infection, in the monocyte fraction of healthy donors (Taylor-Wiedeman etal 1991).Infectious HCMV has occasionally been found in buffy-coat or mononuclearcell preparations (Rinaldo et al 1977, Fiala et al 1975, Jordan 1983, Stanier et al 1989)obtained from patients with clinical HCMV infection; but only rarely has the virus7been isolated from healthy donors (Diosi et al 1969). A low number of purified T-cells(3 per million) isolated from transplant patients produced plaques when co-culturedwith fibroblasts; purified B-cells did not (Gamett 1982). However, a minimalcontamination of T-cells with monocytes could not be ruled out. Early studies usingin situ hybridization found that 0.03-2% of the peripheral blood mononuclear (PBM)cells from 8 asymptomatic individuals hybridized strongly with a probe for theHCMV immediate early gene-i (IE-1; Schrier et al 1985). Further analysis of oneindividual, who had 2% HCMV IE-1 (+) PBM, found that 2.4 % of the CD4(+) cells and0.8% of the CD8 (+) cells hybridized with an IE1 probe. However, it is important tonote that the authors alluded to similar findings with monocytes, and that allmonocytes also express CD4 which was used to identify the T-helper cells. Otherstudies relying on indirect immunofluorescence with monoclonal antibodiesidentified only LE HCMV genes, but not late genes nor the presence of anyintracellular virions by electron microscopy, in peripheral blood cells (Rice et al 1984).Determination of cell types found IE expression in monocytes>NK cells>>Blymphocytes >CD8(+) T cells>>CD4(+) T cells. Interestingly, two separate groupsfound that low passage HCMV isolated from patients was far more capable ofinfecting PBMs than fibroblast-conditioned laboratory strains of HCMV (Rice et al1984, Einhorn and Ost 1984). However, lack of detection of late HCMV geneexpression maybe indicative of an abortive infection since IE1 genes can be expressedin non-permissive cell lines (De Marchi 1983, La Femina and Hayward 1983).Rice et al (1984) indicated that the most receptive cell type for HCMV IE geneexpression was the monocyte and examinations of biopsies of transplanted organsfrom patients with HCMV disease indicated that mononuclear inflammatory cells arethe predominantly infected cell type (Gnann et al 1988, Wiley et al 1986). Directed bythese findings other investigators stimulated primary monocytes or monocytic celllines to become fully permissive to HCMV infection (Weinshenker et al 1988, Latheyand Spector 1991, Ibanez et al 1991). A monocytic leukemia cell line, THP-1, was8found to allow the expression of late HCMV proteins and release infectious virus onlyafter differentiation by phorbol esters (Weinshenker et al 1988). However, a similartreatment of promyelocytic or T cell lines (HL6O, HUT 102, and Molt-4) did not resultin late gene expression or virus production (Weinshenker et al 1988). Primarymonocytes could be induced to express late HCMV genes and produce significantamounts of HCMV if they were first co-cultured with phytohemagglutinin Pstimulated T-cells then treated with hydrocortisone (Lathey and Spector 1991) ordifferentiated with Con A (Ibanez et al 1991). In all cases however, permissive HCMVinfection was only induced after the monocytic cells were stimulated to formmultinucleated giant cells.1.2.3 Infection of non-hematologic cell types.The most common in vitro model of HCMV infection uses human fibroblastswith the laboratory strain of HCMV, AD169. Classically, investigation of in vivodistribution of HCMV utilized histological evidence of specific nuclear andcytoplasmic inclusions. Examination of tissue from individuals acutely infected withHCMV, for the presence of cytomegalic cells, has identified the virus in a wide varietyof organs (Myerson et al 1984, Wiley et al 1986, Ho 1982). Generally, infection hasbeen restricted to cells of epithelial and endothelial origin. However, evidence wasput forward at the 1994 International Herpesvirus Workshop to suggest the fibroblastmay play a central role as an in vivo host cell. Sinzer et al (1994) studied samples fromlung and gastrointestinal tract of patients with symptomatic HCMV infections andfound that the most common HCMV-infected cell type identified was the fibroblast.Furthermore, these investigators noted expression of late genes in endothelial cells,monocytes/macrophages, fibroblasts, smooth muscle cells and epithelial cells. Thesedata suggest that while endothelial cells, monocytes/macrophages andpolymorphonuclear cells may play a crucial role in the hematogenous spread of thevirus, fibroblasts, smooth muscle cells and epithelial cells are probably as important to9the multiplication and spread of HCMV via cell-cell contact in infected tissues. Invitro studies also confirm that cells of epithelial (Smith 1986, Heieren et al 1988,Numazaki et al 1989a,b), endothelial (Ho et al 1984, Smiley et al 1988, Lathey et al1990), smooth muscle (Tumilowicz et al 1985), and fibroblast (Compton 1993, Smith1986) origins are permissive for HCMV infection.1.3. In vivo control of HCMV infection.Regardless of the cell types infected in vivo , pathological consequence ofHCMV infection is associated with immunosuppression, for example post-transplant,infectious mononucleosis, HCMV meningitis/encephalitis. HCMV infection may alsobe acquired congenitally. However, in the immunocompetent host a significantimmune response to HCMV can be seen.1.3.1. Cellular immune response to HCMV infectionCellular immunity plays a central role in controlling HCMV. Post-transplantation studies have demonstrated that recovery from HCMV infectioncorrelates closely with the levels of HCMV-specific cytotoxic responses (Quinnan et al1982, Rook et al 1984, Reusser et al 1991). More recently, provocative studies whereCD8+ and CD4+ cells from transplant recipients were collected before transplantation,pre-stimulated against HCMV, and re-introduced after transplantation, observed aprotection from subsequent HCMV disease (Cheng-Rong et al 1994). Even though thecytotoxic effects of T-cells are more commonly associated with CD8+ cells, therecovery of CD4+ HCMV-specific lymphoproliferative responses was found to beobligatory for the endogenous reconstitution of CD8+ cytotoxic effects (Quinnan et al1982, Reusser et al 1991). However, upon immunosuppression many of these patientsexperience a reactivation of latent HCMV suggesting that cellular immunity maycontrol HCMV infections, but it does not clear the virus from the host.101.3.2. Humoral immune response to HCMV infection.Antibody-mediated complement cytolysis is one of the primary mechanisms bywhich the host immune system eliminates virus-infected cells (Sissons and Oldstone1980). HCMV infection of immunocompetent individuals results in a normal humoralresponse and the subsequent generation of stable levels of anti-HCMV IgG antibodies.Acute HCMV infection leads to the rapid generation of anti-HCMV antibodies of the1gM class (Langenhuysen 1972). A positive anti-HCMV 1gM response correlated withactive infection in mononucleosis patients and the titer was highest during viremia(Rasmussen et al 1982), and subsequently decreased over a period of a few months,while the IgG class of anti-HCMV antibodies increased and remained stable(Langenhuysen 1972, Rasmussen et al 1982). The levels of 1gM anti-HCMV antibodieswere also found to increase after reactivation (or re-infection) of HCMV in patientsand the amount of anti-HCMV 1gM measured was speculated to be related to theseverity of both primary and recurrent HCMV infections (Rasmussen et al 1982).Anti-HCMV antibodies, which arose following HCMV mononucleosis, werefound to react with HCMV proteins of 66, 50, 135 and 42 kDa within 2 weeks aftersymptoms (Hayes et al 1987). The longest response to develop was antibody against a92 kDa protein, which rose slowly for a month or more. The kinetics of anti-HCMVantibodies against the other HCMV proteins were intermediate between these twopatterns. The antibody response to HCMV antigens located in the nuclear andextracellular virions continued to increase for more than three months aftersymptoms, whereas response to the cytoplasmic HCMV antigens peaked within oneto two months (Hayes et al 1987).Several groups have also investigated which epitopes on the HCMV virion aremost important for neutralizing anti-HCMV antibodies. The most common epitopesappear to be envelope glycoproteins of approximately 130 kDa, 86 kDa, and 55 kDa(Britt 1984, Nowak et al 1984, Pereira et al 1984, Rasmussen et a! 1985a). The 55 kDaprotein (p55) was reported to be a post-transcriptionally processed form of the 13011kDa protein (p130) (Pereira et al 1984, Rasmussen et al 1985a). The gene of thep130/55 glycoprotein has been identified (Mach et a! 1986) and shown to behomologous to the HSV glycoprotein B (gB) protein (Cranage et al 1986). The secondprotein (p86) was identified as glycoprotein H (gH) which also shares homology withHSV gH (Cranage et al 1988, Gompeis and Minson 1986). A major difference incomplement requirement was observed when these proteins were inoculated intomice or guinea pigs: the anti-gB antibodies generated require the additional presenceof complement to neutralize HCMV virions, while anti-gH antibodies wereneutralizing indepeident of complement (Pereira et al 1982, 1984, Nowak et al 1984,Britt 1984, Rasmussen et al 1984,1985a,b). Interestingly, if the antibodies were raisedagainst gB expressed in E. coli, and therefore unglycosylated, the resultantneutralizing anti-gB antibodies raised in mice were complement independent (Britt1988).1.3.3 Complement and antibody-mediated HCMV neutralization.The role of complement-mediated neutralization enhancement has been welldocumented. Addition of guinea pig complement to commercially available humanIgG preparations has been found to increase the ability of those preparations toneutralize HCMV, as determined by plaque assay, from between 2- to 16 fold (Lewiset al 1986, Eizuru et al 1988). Other groups have found that the enhancedneutralization observed with complement addition is roughly the same as thatobserved when 0.24 mg/mi of rabbit anti-human IgG was added (Rundell and Betts1982). However, the mechanism by which complement enhances antibody-mediatedneutralization, for example by steric hindrance or virolysis, has not been elucidated.1.3.4 The complement cascade.The complement system consists of at least 20 immunologically non-crossreactive and distinct plasma proteins. Together they represent a significant12proportion of the plasma proteins as their cumulative concentration exceeds 3 mg/mi.Upon interaction with activators, these proteins interact with membranes and withone another in an orderly sequential manner (Figure 1.1). The complementcomponents can be grouped into two well characterized activation pathways, theclassical and the alternative pathways. Activation of the classical pathway is mostoften initiated by the binding of the first complement component (Cl) to antigen-antibody complex; however, other molecules, including polyanions, certain viralproteins, and lipid A of lipopolysaccharides, may also mediate classical pathwayactivation in the absence of antibody. Cl exists in the serum as a calcium-dependentcomplex of three subunits Clq, Cir, and Cis. Clq binds to the Fc region ofimmunoglobulin complexes, triggering the autoactivation of Cir, which then cleavesand activates Cis. Activated Cis then cleaves C4 into C4a and C4b. C4b has areactive internal thioester bond, which is exposed upon cleavage and mediates thecovalent attachment of C4b to the cell surface through an ester or amide bond toprotein, carbohydrate, or lipid moieties. C4b then binds C2 in the presence ofmagnesium, The nearby activated Cls cleaves C2 into C2a, which remains bound toC4b and forms the classical C3 convertase, and C2b, which diffuses away. C4b2a iscapable of cleaving C3 into C3a and C3b, which also has an internal thioester bondand covalently attaches to the surface of the cell. The cleavage of C3 signals thebeginning of the common terminal lytic pathway (discussed below), and the end ofthe classical activation pathway.The alternative complement activation pathway does not require the presenceof immunoglobulins. Activation of the alternative pathway is more commonlyassociated with bacterial cells, soluble immune complexes, or tumor cells. Anessential feature of the alternative pathway is the amplification of C3 convertase. C3b,which is generated by low-grade fluid-phase C3 spontaneous activation (or theclassical pathway), binds to the cell surface and factor B binds to C3b in the presenceof magnesium to form the precursor of the alternative C3 convertase, C3bB. Factor D13Classical Activation PathwayAlternative Activation PathwayC3bC3bFactor B+ Factor D+ ProperdinC3bBbC6,C7C3bC5b67(C8C5b678 cIC5b678999999...(Membrane Attack Complex)Figure 3. Diagrammatic representation of the complement cascade. Lines drawnover complement components indicate an activated state.C4c2C3CommonTerminalLyticPathway14cleaves factor B to form the activated convertase, C3bBb, and this process isaccelerated by the binding of properdin which stabilizes the interaction of C3b andfactor B. The alternative pathway C3 convertase also generates C3b which will formthe C5 convertase and the classical and alternative pathways join into the commonterminal pathway at this point.The common terminal lytic pathway starts with the cleavage and activation ofC5. C4b2a3b and C3bBb both have the ability to cleave C5 into C5b and C5a. C6 thenbinds to and stabilizes C5b (otherwise, dissociation of C5b from the C5 convertasecomplex occurs). The C5b6 complex then reacts with C7 to form a trimolecularcomplex, C5b67. This complex then weakly associates with the membrane and theassociation of C8 makes the complex more hydrophobic which in turn causes astronger binding of C5b678 to the membrane. The association of C9 initiates the finalinsertion of this complex across the membrane, and additional binding of multiple C9subunits causes the formation of the lytic pore which can be visualized by electronmicroscopy (Berry and Almeida 1968). This terminal complex of complement is calledC5b-9 or the membrane attack complex (MAC). The mechanism by which the MACcauses cell death has been reviewed elsewhere (Morgan 1989), and some investigatorssuggest that complete formation of the lytic pore is not required to induce cell death(Pramoonjago et al 1992, Dankert and Esser 1985, Morgan et al 1987).1.4. Host-encoded complement inhibitors.The complement system provides a potent means of recognizing andeliminating foreign elements. It is critical, however, to focus these actions on foreignparticles and to prevent inadvertent attack against host tissue. Integral to thisfunction are regulatory proteins of the complement system. Such componentsprovide a means of separating “self’ from “non-self” during complement attack. Priorto 1980, several plasma proteins were known to fulfill this role. However, in the early1980’s two additional regulatory proteins of the complement system, decay15accelerating factor and membrane cofactor protein, were identified; closely followedby the discovery of CD59 in the mid 80’s (discussed below). This section will discussseveral aspects of host complement inhibition and some of the viral immune evasionmechanisms which have evolved and mimic the actions of these host proteins.1.4.1. The regulators of complement activation (RCA) gene cluster.The human RCA gene cluster is a 900 kb region located at the q32 region on thelong arm of chromosome 1 (Weis 1987; Lublin 1987,1988) which encodes a family ofproteins that are structurally and functionally related. Pulsed-field gel electrophoresisand Southern blot analysis has revealed that the following genes are located there (inorder): CD46 (membrane co-factor protein) - CD35 (complement receptor 1) - CD21(complement receptor 2) - CD55 (decay-accelerating factor)- C4 binding protein(C4bp) x subunit - C4bp I subunit. Another structurally related complementregulator is located around 500 kb from this group and can also be considered part ofthe RCA gene cluster (Rey-Campos et al 1988; Carroll et al 1988; Bora et al 1989;Pardo-Manuel et al 1990). Additionally, there appears to be some partial duplicationof the CD35, C4bpx, and CD46 genes which are also located in this gene cluster, buttheir roles in complement regulation, if any, remain to elucidated (Hourcade et al1992, 1990b,1988; Wong et al 1989; Pardo-Manuel et al 1990). All of the characterizedproteins share the ability to bind C3 or C4 fragments, act as a co-factor for factor I-mediated inactivation of C3 or C4 and/or dissociate C3 convertases. All of theseproteins also show a particular structural organization based on the presence of aninternal repeat of -60 amino acids that share a framework of highly conservedresidues called a short consensus repeat (SCR; Reid et al 1986; Aheam and Fearon1989, Coyne 1992). These similarities and the juxtaposition of the genes may supportthe hypothesis of a common ancestral gene which diversified by partial geneduplication. Discussed below are the relevant complement regulating proteins:decay-accelerating factor (CD55), membrane cofactor protein (CD46), complement16receptor 1 (CD35), complement receptor 2 (CD21), C4-binding protein, factor H,membrane inhibitor of reactive lysis (CD59), and C8-binding protein. CD55 or decay-accelerating factor (DAF).CD55 is constitutively expressed on most cells with the exception of naturalkiller cells (Nicholson-Weller et al 1986). CD55 is not a typical membrane boundprotein since it has no transmembrane region, but is linked to the cell surface by aglycophospholipid (GPI) anchor (Davitz et al 1986, Mahoney et al 1992). This anchoris added in the endoplasmic reticulum and on most cells is susceptible to cleavagewith phosphoinositol-specific phospholipase C (PIPLC; Davitz et al 1986, Thomas et al1990, Caras et al 1987). CD55 also exists as two isoforms; erythrocyte CD55 is a 70kDa protein which has a slightly altered GPI composition (Davitz et al 1986, Roberts etal 1988, Kinoshita et al 1985) while and the rest of the cell types which express CD55have a 78 kDa protein (Kinoshita et al 1985). This relative mass difference has beenspeculated to be due to differences in the glycosylation of the core protein (Lublin et al1986). Deficiencies of CD55 have been reported to be in part responsible for theincreased hemolysis associated with paroxysmal nocturnal hemoglobinuria (PNH); anacquired myelodysplastic disease in which all GPI-anchored proteins are deficient ormissing entirely (Selvaraj et al 1988, Fujioka and Yamada 1992, Norris et al 1994,Ohashi et al 1994). However, some individuals who are selectively deficient in CD55mRNA levels and protein expression with normal levels of the other GPI-linkedproteins have been described (Inab phenotype; Tate et al 1989, Telen et al 1988, Merryet al 1989).CD55 regulates complement by inactivating the classical and alternative C3convertases through acceleration of the dissociation of the bimolecular enzymes(C4b2a and C3bBb) into subunits; this occurs only on the surface of cells which bearCD55 (Holguin et al 1992, Fujita et al 1987, Medof et al 1984). CD55 binds to the intactconvertases with a higher affinity than to the separate components (Pangburn 1986),17and CD55 rapidly dissociates C2 and factor B only after being cleaved to their activefragments (Fujita et al 1987). CD55 also binds to C3b alone and may competitivelyinhibit convertase formation (Pangbum 1986). The overall ability of CD55 to regulateC3 deposition is greater than CD46, however, together they are more effective thansingly (Iwata et al 1994). Although CD55 is composed of four short consensus repeats(SCR) of -P60 amino acids each, the entire C3 convertase dissociating activity wasmapped (by molecular manipulation of cDNA constructs and blocking by monoclonalantibodies) to the third SCR (Coyne et al 1992). However, the spacing of SCR3 seemsto play an important role. It is interesting to note that the GPI anchor played no rolein the efficiency of CD55’s regulatory activity since transmembrane versions, createdby manipulating cDNA constructs, resulted in molecules with identical complementregulatory activity (Lublin and Coyne 1991). CD46 or membrane cofactor protein (MCP).CD46 is expressed constitutively on most cells with the exception of red bloodcells (Cho et al 1991, Cole 1985, Yu 1986, Seya 1988, McNearey 1989). Unlike CD55,CD46 is bound to the cell surface through a transmembrane region. A distinguishingstructural characteristic of CD46 on SDS-PAGE is the presence of two broad proteinspecies with relative masses of 59-68 kDa and 50-58 kDa (Johnstone et al 1993). Thesetwo forms differ by the presence or absence of a heavily 0-glycosylated regionencoded by exon 8, but still consist of 4 SCR’s. However, the tissue distribution of thedifferent isoforms is less well defined than the distributions for CD55; some tissuesexpress both forms of protein and this further varies between individuals (Johnstoneet al 1993). A detailed study reports that of the 14 introns found in the CD46 gene, alarge inter-organ and inter-individual variation exists in the combination of intronsincluded in mRNA species (Johnstone et al 1993). This is in contrast to CD55 whichhas a major and minor mRNA species in all cell types but produces only one protein(Thomas and Lublin 1993).18CD46 regulates complement activation by interacting with cell-bound C3b, andto a lesser extent C4b, and acting as a co-factor for factor I, which cleaves C3b and C4bto inactive fragments (Seya and Atkinson 1989, Adams et al 1991). Interestingly, eventhough factor I can cleave C3b to the smaller C3c and C3dg fragments in the presenceof factor H (see below), there seems to be some controversy as to whether factor I cancleave C3b to C3dg or only to iC3b in the presence of CD46 (Seya and Atkinson 1989,Adams et al 1991). CD46 has minimal activity when it is soluble, and therefore, CD46is considered an intrinsic cofactor for the cleavage of C3b bound to the same cellsurface (Seya and Atkinson 1989). SCR deletion mutants were constructed todetermine which of the four SCR of CD46 contribute to ligand binding and cofactoractivity. The third and fourth SCR were important for both ligand binding andcofactor activity of C3b and C4b, but deletion of SCR1 only decreased the binding ofC4b (Adams et al 1991). Further, deletion of SCR2 did not decrease binding of C3b,but deletion of SCR2 abolished the cofactor activity. This suggests that binding ofCD46 to complement fragments is not always sufficient for cofactor activity. It hasalso been reported that CD46 preferentially associates with and more efficientlyinactivates the C3b dimers in the alternative C5 convertase than other membranebound forms of C3b (Seya et al 1991), and may have minimal regulatory activitytowards classical pathway activation (Kojima et al 1993). CD35 or complement receptor 1 (CR1).The tissue distribution of CD35 is more restricted than CD55 or CD46.Erythrocytes, B lymphocytes, a subset of T lymphocytes, monocytes, neutrophils,eosinophils, glomerular podocytes, and follicular dendritic cells are the only cell typesdocumented to express CD35 (Fearon 1980, Wilson et al 1983, Reynes et al 1985,Gelfand et al 1975, Kazatchkine et al 1982). A soluble form of CD35 also has beenfound in plasma in picomolar concentrations that sedimented as a broad peak inultracentrifugation (Yoon and Fearon 1985). Like CD46, CD35 is attached to the cell19surface by a transmembrane region. However, unlike CD46 and CD55, CD35 is madeup of 30 SCR’s and has a relative molecular mass of 190 kDa. Since every eighth SCRis a highly homologous repeat, this very large protein can be further organized into 3long homologous repeats, a structural organization which is unique to CD35 (Ahearn1989). However, this organization only holds true for the most common allotype. Atotal of four allotypic forms of CD35 have been identified. The size and frequency ofoccurrence (within brackets) is as follows: 190 kDa (0.82) > 220 kDa (0.18) >.160 kDa(<0.01) = 250 kDa (<0.01). These forms seem to differ by increments of 30 kDa whichis comparable to the Mr of one long homologous repeat (LHR) and other evidencewhich suggests unequal translocation as being the mechanism by which they arose isreviewed in Fearon and Ahearn 1989.Molecular cloning and deletion experiments have identified separate C3b andC4b recognition sites within CD35 (Klickstein et al 1988). The C3 recognition siteswere found to exist in both the LFIR-B and -c regions, while LHR-A was found tocontain the primary C4 recognition site. CD35 can regulate the alternative pathwayactivation by three mechanisms: impairing uptake of factor B by C3b, displacing Bbfrom the C3bBb convertase, and acting as a cofactor for the cleavage of C3b to iC3b,C3c and C3dg by factor I (Fearon 1979, Seya and Atkinson 1989, Seya et al 1991,Pangbum 1986). Similarly, CD35 inhibits the classical pathway by impairing uptakeof C2 by C4b, displacing C2a from the C4b2a convertase, and promoting the cleavageof C4b to C4c and C4d by factor I (lida and Nussenzweig 1981, Pangburn 1986). ThusCD35 has similar complement activation regulating mechanisms as both CD55 andCD46, with one important difference: the tissue distribution of CD35 is limited and itmainly acts in an extrinsic fashion, i.e. on C3b/C4b deposited on bystander cellsurfaces (Ross and Medof 1985, Medof et al 1982, 1983). This implies that CD35 playsa minimal role in intrinsic cellular protection from complement mediated lysis.Finally, CD35 has recently found a prominent place in protection of in vivo models20from complement mediated tissue damage, following intravenously administeredsoluble, recombinant CD35 (Piddlesden et al 1994, Homeister et al 1993). CD21 or complement receptor 2 (CR2).The tissue distribution of CD21 is even more restricted than CD35; it isexpressed in mature B lymphocytes, human thymocytes, rare T lymphoblastoid celllines, pharyngeal and cervical epithelium, and in fofficular dendritic cells (Ross et al1973, lida et al 1983, Tsoukas and Lambris 1988, Menezes et al 1977, Tatsumi et al1985, Fingeroth et al 1988, Young et al 1986, Sixbey et al 1987, Reynes et al 1985).CD21 is attached to the cell surface by a transmembrane region, and the extracellulardomain is composed of 15 SCR’s (Weis and Fearon 1985). There are no LHR’s withinCD21 as there are in CD35; however, there is a less conserved repeating pattern ofhomology involving every fifth SCR (Ahearn and Fearon 1989). The mature, cell-surface expressed CD21 has an Mr of 145 kDa and the N-linked oligosaccharides havebeen linked to extended biological half-life (Weis and Fearon 1985).CD21 binds to the C3d fragment of C3, which is available for binding in iC3b,C3dg, and C3d (Weis et al 1984). Of the 15 SCR’s contained within CD21, both the N-terminus SCR1 and SCR2 were required for the binding of C3dg and the Epstein-Barrvirus (Lowell 1989). However, the complement regulatory role of CD21 is not clear.Some investigators document that CD21 is responsible for the activation of thealternative pathway (Theofilopoulos et al 1974, Baker et al 1977, Schreiber 1980,Ramos et al 1985), while others reported CD21 factor 1-cofactor activity for thecleavage of iC3b to C3dg and C3c (Mitomo et al 1987). The role of CD21 is moreprominent in B-cell activation and as the Epstein-Barr virus receptor (reviewed inAhearn and Fearon 1989). Remaining members of the RCA gene cluster.Factor H is a soluble serum protein which is present at a high concentration inserum (350 .tg/m1). It is an elongated molecule of approximately 160 kDa, and iscomposed of 20 SCRs (Reid et al 1986, DiScipio 1992, Ripoche et al 1988). Electronmicroscopic analysis found that factor H was an asymmetric, elongated molecule,with one end slightly larger and more rounded than the other (Smith et al 1983,DiScipio 1992). Five allelic variants of factor H have been identified using isoelectricfocusing and neuraminidase treatment (Rodriguez de Cordoba and Rubenstein 1984,1987). Factor H not only acts a cofactor for factor I cleavage of C3b to C3c and C3d,but also accelerates the decay and inhibits the reformation of alternative C3 and C5convertases (Whaley and Ruddy 1976a, Fisher and Kazatchkine 1983). The nature ofthe cell surface to which C3b is attached effects the activity of factor H; the presence ofsialic acid greatly enhances the factor I cofactor activity (Fearon 1978, Pangburn andMuller-Eberhard 1978). Factor H binds to the C3d portion of C3 and can compete forbinding with CD35 (Pangburn 1986). Both the C3b-binding and cofactor activitieshave been reported to reside in the 38 kDa N-terminus (Alsenz et al 1984, 1985).Factor H also binds properdin, but at a site distinct from the C3b-binding site(DiScipio 1981), and very weak C4b-binding ability for factor H has also been reported(Pangbum 1986, Whaley and Ruddy 1976b). Finally, there is one report of a mutantRaji human B-cell line which did not secrete factor H into the medium, but expressedfunctional factor H on it’s cell surface (Demares 1989). This could implicate factor Has a cell surface complement regulator; however, it remains to be determined whetherthis phenomenon is replicated in vivo , or whether it is related to the recently reportedfactor H receptor (Avery and Gordon 1993).C4bp is the only other functionally characterized complement regulator in theRCA. Like factor H, C4bp is soluble and is found in serum at a concentration of 150j.tg/ml; unlike factor H, C4bp is composed of a combination of 2 separate subunitswhich are both encoded in the RCA gene cluster. C4bp is about 570 kDa and has a22unique spider-like molecular structure with seven extended tentacles as observed byhigh resolution electron microscopy (Dahlback et al 1983). This ‘spider’ is composedof 7 cL-chains, which each has a C4b binding site and one -chath (Dahlback et al 1983,Nagasawa et al 1983). Even though each C4bp molecule has the potential to bind 7molecules of C4b, four C4b were shown to be bound per C4bp at physiologicalstrength (Ziccardi et al 1984). The 3-chain does not bind C4b, but instead binds andregulates the coagulation factor protein S (reviewed in Dahlback 1991). Despite thesedifferences, each ce-chain of C4bp still contains 8 SCRs, which are distinctive to theRCA family. The binding of C4bp to C4b is important to the regulation of the classicalpathway C3 convertase (C4b2a). The binding of C4bp to C4b accelerates the naturaldecay of the C2a subunit from the C4b2a complex and functions as a cofactor for thecleavage of C4b into iC4b and C4d (Adams et al 1991).1.4.2 CD59 or membrane inhibitor of reactive lysis (MIRL).CD59 is not a member of the RCA gene cluster; it is located at the p13 region ofthe short arm of chromosome 11 (Heckl-Ostreicher et al 1993). Furthermore, CD59does not share the SCR structural organization or functional similarities with the RCAgene cluster. It does have a GPI anchor like CD55, CD59’s absence has also beenimplicated as a contributor to the pathology of PNH (Ohashi et al 1994, Fujioka andYamada 1992, Yamashina et al 1990). CD59 has a wide tissue distribution (Men et al1991). It is a small 18 kDa, highly glycosylated protein (Holguin et al 1989, Davies etal 1989), and acts to inhibit complement activation. This occurs by inhibition of theincorporation of C8 and multiple C9 molecules into the terminal complement lyticcomplex, which results in abolition of the formation of the lytic pore (Rollins et al1991). However, there is some controversy surrounding the initial findings of speciesrestriction for the effectiveness of CD59’s activity (Rollins et al 1991, Morgan et al1994).231.4.3 C8 binding protein (C8bp) or homologous restriction factor.This is the least well characterized complement regulator, and its associationwith the other complement inhibitors by genomic location is unknown. C8bp wasfirst isolated from red blood cells and found to have an Mr = 65 kDa (Schonermark etal 1986, Zalman et al 1986). C8bp was also found on monocytes and lymphocytes, andC8bp’s sensitivity to PIPLC suggests that, like CD55 and CD59, it is attached to the cellsurface by a GPI anchor (Hansch et al 1988). Furthermore, C8bp has also beenreported to be absent on PNH red blood cells (Zalman et al 1987, Hansch et al 1988).More recently another group has identified a related protein they call membraneattack complex inhibitory protein (MW), but whether this is the same as C8bp remainsto be elucidated (Watts et al 1990).Functionally, C8bp resembles CD59. C8bp blocks the formation of themembrane attack complex (MAC) by binding to C8, C9, or both (Schonermark et al1986, Zalman et al 1986), and incorporation of C8bp into PNH cells reduced theirsusceptibility to reactive lysis (Zalman et al 1987, Hansch et al 1988). A urinary formof C8bp was also reported. The soluble C8bp appeared to lack the GPI anchor, butalso appeared to inhibit the assembly of the MAC at a different site that theerythrocyte-bound form (Zalman et al 1989).1.5. Complement-mediated cytolysis and virus-infected cellsAntibody-mediated destruction of virus-infected cells is an effective methodfor virus clearance by the immune system (reviewed in Sissons and Oldstone 1980).However, certain viruses have evolved evasion strategies which allow virus-infectedcells to inhibit complement attack, or at least to decrease complement activation longenough to produce viral progeny. So far these viruses include vaccinia and a fewmembers of the herpesviridae as discussed below.241.5.1 Soluble vaccinia complement inhibitor.Vaccinia produces a 35 kDa major secretory protein, which is structurallyrelated to C4 binding protein and is so abundant it can be visualized in cellsupematants by staining of polyacrylamide gels with Coomassie blue. This proteinwas not found in attenuated strains of vaccinia, and was only found to be effectiveinhibiting activation of the classical pathway (Kotwal and Moss 1988, Kotwal et al1990).1.5.2 Herpesviridae complement inhibitorsThe strategy evolved by members of the herpesvirus family is somewhatdifferent; the complement inhibiting activity is found within virion envelope proteins.These are expressed on the surface of virus-infected cells and have additionalfunctions important to infectivity. The Epstein-Barr virus (EBV) encodes a 350 kDaglycoprotein (gp350) which mediates the binding to the EBV receptor, complementreceptor 2 (CD21, see below), accelerates the decay of the alternative pathway C3convertase (but not the classical C3 convertase), and has cofactor activity for factor I (aserum complement regulating protein) which in turn mediates the cleavage of C3b tothe inactive fragment C3dg and C4b to inactive C4c (Nemerow et al 1987, Mold et al1988). It is interesting to note that EBV only infects B-cells (which express CD21), andthat gp350 enhances the production of C3dg on the surface of EBV virions and EBVinfected cells. Thus the surface of EBV infected cells and EBV virions have twoseparate ligands, gp350 and C3dg, for the receptor, CD21. Similarly, herpes simplextype 1 and 2 (HSV-1 and HSV-2) have complement inhibiting functions which colocalize with the glycoprotein C (gC) a heparin-binding protein important to HSVvirion attachment to permissive cells (Herold et al 1991). The complement inhibitingfunction of gC was identified as an ability to bind C3b and accelerate the decay of thealternative pathway C3 convertase (Harris et al 1990, McNearney et al 1987, Hatano etal 1988, Smiley et al 1985, Fries et al 1986). However, gC was found to lack both factor25I co-factor activity or classical C3 convertase decay accelerating activity (Fries et al1986). Additionally, two other members of the herpesviridae which have proteins ofhigh homology to HSV gC, porcine pseudorabies virus and bovine herpesvirus-1,were both found to have C3b binding ability and may have complement regulatoryfunctions (Huemer et al 1993, Huemer et al 1992). However, two separateinvestigators have reported that HCMV does not have a C3b-binding protein (Smileyet al 1985) or activity like EBV’s gp350 (Nemerow et al 1987).1.5.3 Complement regulation on HCMV-infected cellsEven though an HCMV-encoded complement inhibiting protein has not beenidentified, HCMV infected cells still seem to have an enhanced ability to regulatecomplement. The serum levels of complement-fixing anti-HCMV antibodies increasefollowing infection and are detectable for a prolonged period following HCMVinfection (Spencer and Anderson 1972). Other studies have identified antigenicHCMV proteins which are expressed on the surface of infected cells early in infection(Amadei et al 1983). However, two separate groups have reported that serum fromlatently infected patients, which contain anti-HCMV antibodies and complement, areincapable of lysing HCMV-infected cells in vitro , except under special circumstances(Betts and Schmidt 1981, Middeldorp et al 1986). The special circumstances are metby the use of serum isolated from patients within the acute phase of HCMV infection,and this activity disappeared within 6-8 weeks after onset of infection. This activitywas suggested by Betts and Schmidt (1981) to be caused by the presence of 1gM classanti-HCMV antibodies. In contrast, Middeldorp et al (1986) found that cytolyticantibodies were distributed through both lgG and 1gM class antibodies. This authorattributed the cytolytic activity to the presence of antibodies binding to late membraneexpressed HCIVIV proteins. However, under all conditions the cells could only belysed 72 h post-infection (p.i.), by which time fibroblasts are producing infectiousextracellular virus (McAllister et al 1963; see Figure 1.2). Furthermore, 20% of latently26infected patients were reported to have antibodies directed against membraneassociated HCMV antigens (Middeldorp et al 1985), and in the samples which hadcytolytic activity, binding of antibodies was observed transiently at 6 h p.i. andincreased from 18 h p.i., long before the ability of the HCMV-infected cells to be lysedat 72 h p.i. (The and Langerthuysen 1972, Middeldorp et al 1985, 1986). This earlybinding of potentially cytolytic anti-HCMV without complement-mediated lysis until72 h p.i., suggests the presence of a complement inhibitor.1.6 Alterations in host protein expression induced by HCMV infection.The most intensely studied alteration of protein expression induced by HCMVinfection is the decrease in HLA class I; due to the reduced efficiency of cytotoxic Tcell-mediated cytolysis of HCMV-infected cells, presumably caused by reduced cellrecognition (Laubscher et al 1988, Warren et al 1994). Initially, HCMV infection offibroblasts was reported to increase the expression of HLA class I (Grundy et al 1988),and 3-2-microglobulin was reported to bind to HCMV virions (Grundy et al 1987a,McKeating et al 1987) and enhance HCMV infection of fibroblasts (Grundy et al198Th). These initial findings resulted in the speculation that HLA class I was thecellular receptor for HCMV. However, shortly thereafter, numerous groups reporteda decrease of HLA class I on the surface of HCMV infected fibroblasts (Yamashita et al1993, Barnes and Grundy 1992, Gilbert et al 1992, Beersma et al 1993). The level ofHCMV interference with 1-ILA class I expression is thought to be after thetranscriptional level since no alteration in the mRNA levels have been observed(Browne et al 1990). Early reports supported the hypothesis that a viral protein (H301,UL18) bound and sequestered beta-2-microglobulin, thus preventing the normalassociation and expression of HLA class I chains (Stannard 1989, Tysoe-Calnon et al1991, Browne et al 1990). However, deletion of the H301 (UL18) gene from the HCMVgenome did not abolish the FILA class I down-regulation (Browne et al 1992).27More recent reports identified an accelerated HLA class I heavy chain half-life,coinciding with infection by HCMV (Beersma et al 1993, Warren et al 1994). Thesteady state levels of HLA class I heavy chains (both free chains and those complexedwith beta-2-microglobulin) were reduced, even though pulse-chase analysis indicatedthat initial synthesis was not affected (Beersma et al 1993). HCMV-infection was alsoaccompanied by the loss of post-translational modifications associated with thecis/medial Golgi compartment, and the possible presence of a ‘virus-assisted’ HLAheavy chain-specific protease was therefore suggested. However, no alterations in thecellular levels of -2-microglobulin were detected nor were novel 3-2-microg1obulincomplexes with viral proteins seen when immunoprecipitating with antibodiesspecific for 13-2-microglobulin.Conversely, HCMV infection of cells also causes increase of total cellularmRNA as well as an increase of host gene products involved with DNA replication,including human thymidine kinase and hamster dthydrofolate reductase (Estes andHuang 1977, Benson and Huang 1990, Wade et al 1992). Other intracellular proteinsfound to increase with human HCMV infection include heat shock protein-70 andcertain proto-oncogenes (Santomenna and Colberg-Poley 1990, Boldogh et al 1991,Boldogh et al 1990). HCMV infection of endothelial cells has also been found torapidly increase the secretion of interleukin-6 (IL-6; Almeida et al 1994). HCMVinfection of the monocytic leukemia cell line, THP-1, resulted in the increase in IL-i 3mRNA levels (Iwamoto 1990), but not in secreted protein (Kline et al 1994); however,it did result in an increased secretion of an IL-i receptor agonist. Lastly, HCMVinfection of endothelial cells also resulted in an increased polymorphonuclear cellsand indirect evidence suggested that the increased adherence was mediated by anincrease in ELAM-1 (Span et al 1991). However, the mechanism of adherence appearsto be indirect, since the presence of anti-IL-i antibodies partly inhibited thisphenomenon.28This thesis investigates the influence of HCMV infection on the expression ofthe host-encoded complement inhibitors. The influence of membrane-boundinhibitors is emphasized, because under the conditions of the in vitro experimentsmentioned above (Middeldorp et al 1986, Betts and Schmidt 1981), soluble factorswould be removed and could not mediate the inhibition of complement. This doesnot exclude the possibility that HCMV infection influences the expression of solublecomplement inhibitors and/or that the increased expression of soluble complementregulators would not be as effective at protecting HCMV-infected cells.The overall goal of this work is to understand how HCMV-infected cells evadeantibody-mediated C’-cytolysis. The elucidation of the interaction of HCMV with it’shost cell may lead to the development of techniques to manipulate the patient’simmune response to clear endogenous and exogenous HCMV and decrease themorbidity, mortality of HCMV-associated disease, and graft rejection. The specificaims of this thesis were directed to test the hypothesis that HCMV infection of hostcells leads to an altered complement regulation on the cell surface. We set abouttesting this hypothesis by determining:1) The degree of complement binding and neutralization of free HCMVvirions.2) The effect of viral infection on the surface expression of host-encodedcomplement inhibitors.3) The effect of viral infection on the transcription of host-encodedcomplement inhibitors.For these studies I utilized fibroblasts since they are the most commonly used celltype for in vitro experiments and relate directly to HCMV complement-resistance(Middeldorp et al 1986, Betts and Schmidt 1981). A second adherent permissiveglioblastoma cell line, U373-MG, was also used to address cell specificity issues; and29finally, I used the THP-1 monocytic leukemia cell line because they are reported to bepermissive for HCMV infection and represent the cell type most probably involved asthe HCMV persistence reservoir (Weinshenker 1981).302. MATERIALS AND METHODS:2.1. Materials.Bovine serum albumin (BSA, fraction five), EDTA (ethylene-diaminetetraacetate) and EGTA(ethyleneglycol-bis-(beta-aminoethylester) tetraacetic acid)were obtained from Sigma (St. Louis, MO). Scintillation cocktails (Ready Safe andReady Prot.), and all ultracentrifuge rotors were from Beckman (Palo Alto, California).The scintillation cocktail used for the kinetic partitioning CAT assay, Econofluor, andall radioactive reagents were purchased from Dupont-NEN (Mississauga, Ontario)except[32PJ-labeled dCTP which was purchased from ICN (St. Laurent, Quebec). Alltissue culture products were purchased from Gibco/BRL (Burlington, Ontario).Purified complement components: C3, factor B, and factor D were provided by Dr.Devine, and cobra venom factor was purchased from Jackson Immunoresearch(Avondale, PA).2.2. Cells and viruses.Primary explant fibroblasts were derived from human neonatal foreskin(HuFF), using the method described by Korn et al (1983), and cultured in DMEM(Gibco/BRL; Burlington, Ontario) supplemented with 100 U penicillin/ml, 100 mgstreptomycin/ml, and 5% FCS. Primary fibroblasts were used between the secondand ninth passage. The glioblastoma cell line, U373-GB, was obtained from AmericanType Culture Collection (ATCC: catalogue # HTB-17). These cells were cultured inMEM with non-essential amino acids supplemented with 1 mM sodium pyruvate,Earle’s basic salt solution, 100 U penicillin/ml, 100 mg streptomycin/ml, and 5%FCS. The human epithelial cell line, HeLa, and the monocytic leukemia cell line, THP1, were also purchased from the ATCC, and propagated in Eagle’s DMEMsupplemented with 100 U penicillin/ml, 100 mg streptomycin/ml, and 5% FCS.31The laboratory strain of human cytomegalovirus used for these studies wasAD169, obtained from the ATCC. Patient isolates of HCMV were a gift of Dr. EvaThomas, Children’s Hospital, Vancouver, B.C. The identification of patient isolates asHCMV was made using both plaque assays and immunocytochemical assays atChildren’s Hospital, Vancouver, B.C. The KOS strain of Herpes Simplex-i (HSV-1)was a gift from Dr. Frank Tufaro (University of British Columbia, Vancouver, B.C.).The Ad5 strain of Human Adenovirus was a gift of Dr. Wilf Jefferies (University ofBritish Columbia, Vancouver, B.C.).2.3. Antibodies.The description and sources for murine monoclonal antibodies used in thisstudy are listed in Table 1. Peroxidase-conjugated goat anti-rabbit IgG (used forWestern blot analysis) or fluorochrome-conjugated (either phycoerythrin or FITC)goat anti-mouse IgG or anti-rabbit antisera (used for flow cytometry and indirectimmunofluorescence analysis, respectively) were purchased from JacksonImmunoresearch (Avondale, PA). Western blot analysis of complement inhibitors oncells or virions utilized: the polyclonal rabbit anti-CD55 was raised using DAFisolated from RBC by the method of Nicholson-Weller et al (1982); the polyclonalrabbit anti-CD59, a gift of Dr. C.J. Parker (University of Utah); the polyclonal anti-CR1, a gift of Dr. B. Paul Morgan (University of Wales); and the monoclonal antiCD46 (M75- see Table 1).Western blot analysis of complement components C3 and C9 utilizedpolyclonal antibodies raised in goats and purchased from Quidel (San Diego, CA) anddetected with a peroxidase-conjugated anti-goat IgG antibody purchased fromJackson Immunoresearch (Avondale, PA). Western blot analysis of human IgGutilized a peroxidase conjugated primary antibody raised against human lgG heavyand light chains (Jackson Immunoresearch). Normal rabbit sera was used as a controlin Western blot analysis and was obtained from Cedarlane Labs (Hornby, Ontario).32The control antibodies used in flow cytometry consisted of an isotype-matchedmonoclonal anti-MHC class II (Isotype IgG2 gift of Dr. W. Jefferies) or a mousemonoclonal antibody raised against Aspergillus niger glucose oxidase (isotype IgGi,DAKO laboratories, Santa Barbara, CA), which is neither expressed nor inducible inmammalian tissues.Table 1. Monoclonal antibody description and sources.Antibodies Ag recognized Source11H4, 12G6, 11D7 CD55 Dr. W.F. RosseBR1C229 CD59 Dr. B.P. MorganMIRL1 CD59 Dr. C.J. Parker3D9 CR1 Dr. C.J. ParkerM75 CD46 Dr. D.M. LublinJ4-48 CD46 Dako LaboratoriesW6/32 HLA class I Dr. W. Jefferies9.3F10 HLA class II Dr. W. JefferiesCat #9466SA C3d BRL (Burlington, ON)1. Affiliations of sources: Dr. W.F. Rosse - Duke University; Dr. B.P. Morgan- University of Wales; Dr.C.J. Parker - University of Utah; Dr. D.M. Lublin - Washington University; Dr. W. Jefferies- Universityof British Columbia. DAKO laboratories products are distributed from Santa Barbara, CA.2.4. Serum samples.Serum samples were obtained from healthy volunteers in our and neighboringlaboratories. Samples were drawn by venipuncture using serum-separator typevacutainer tubes. The blood was allowed to clot at room temperature for 10 min.prior to centrifugation at 3000 xg. The serum was immediately removed and divided33into 250 il aliquots into sterile eppendorf tubes. The serum samples were frozenimmediately at 700C and used within 4 months of collection to ensure that thecomplement activity in the samples remained optimal. Some aliquots were heat-inactivated at 56°C for 30 mm. to remove the complement activity when required.Aliquots from all serum samples were also sent the Vancouver branch of theCanadian Red Cross transfusion and blood services to be tested for the presence ofHCMV antibodies using the Abbott EIA total HCMV antibody test (Hamilton, ON).All serum samples were coded and all experiments performed blind as to the status ofthe presence of HCMV antibodies.2.5. IgG depletion from serum samples.The anti-HCMV antibody status of many of the individuals was measured atintervals of 4-6 months. Serum samples which contained anti-HCMV IgG antibodies,but were negative for anti-HCMV 1gM were depleted of IgG, resulting in the removalof all anti-HCMV antibodies. These depleted sera were confirmed as seronegative byAbbott ETA. For IgG depletion, 700 tl of fresh serum was loaded on a 1 ml protein Gsepharose column which had been equilibrated with 20 mM sodium phosphate(pH=7.0). The serum was allowed to enter the column, the flow was then stoppedand the serum allowed to incubated in the sepharose for 30 mm. at 4°C. Flowthrough the column was resumed and 0.5 ml fractions collected; the depleted serumwas easily identified by it’s yellow color and the serum containing fractions werepooled and concentrated to the original 700 iii volume. The depleted serum wasconcentrated using a Amicon centrifugal concentrator with a molecular weight cut-offof 10 kDa to avoid losing the smaller complement components (i.e. factor D). Thecomplement activity of the depleted serum was checked by the clinical hematologylaboratory at University Hospital and found to be at the low end of the normal range(classical pathway=114-202, alternative pathway=92-152). The TgG retained by theprotein G-sepharose column was eluted using a 0.1 M glycine buffer (pH=3.0), and 0.534ml fractions collected. The absence of IgG in the depleted serum and presence of IgGin the eluted fractions was determined by Western blotting analysis (see below) usinga peroxidase-conjugated polyclonal antibody directed against human IgG (see Figure11 in results section).2.6. Virus Infection.The laboratory strain of human cytomegalovirus, AD169, was used to infectcell lines at a multiplicity of infection (MOI)=10 for investigations of complementinhibitor expression changes. Mock-infected controls consisted of cells incubated withUV-inactivated HCIvIV using conditions previously reported; briefly, half of the viralstock to be used for infection of cells was put in the UV-stratalinker (254 nm) for 15mm. on maximum irradiation just prior to incubation with cells (Nishiyama and Rapp1980). UV-inactivated HCMV produced no plaques when subsequently tested bystandard plaque assay techniques, and data collected for cells incubated with UVinactivated HCMV was statistically indistinguishable from data collected foruninfected cells. Infection of fibroblasts with human adenovirus type 2 was carriedout at an MOI=50, whereas infection with the KOS strain of herpes simplex virus-i(HSV-1) was performed at an MOI=10. Culture conditions for adenovirus and HSV-iwere the same as those described above for HCMV.Stocks for HCMV were propagated by infecting 150 cm2 T-flasks containingfibroblasts at 85% confluency at an MOI<0.01, to minimize the production of non-infectious enveloped-particles and dense bodies (virions containing no HCMV DNA;Irmiere and Gibson 1983) and defective interfering particles (incomplete and incorrectHCMV genomes; Ramierez et al 1979, Stinski et al 1979), which leads to a decrease ininfectivity. These flasks were harvested when 100% cytopathic effects (cell roundingand vacuolization) was observed, usually between 10-14 days p.i., with a cell scraper.Intracellular virions were released by passing the harvested cell debris through a 25-gauge needle three times. The cell debris was then removed by centrifugation of the35samples at 600 xg, leaving the virions in the supematant. This supernatant usuallycontained roughly 5x106plaque-forming units (pfu) per ml, and was sufficient forviral infections of adherent cells. However, HCMV stocks could be furtherconcentrated by centrifugation of the precleared supernatant at 20,000 xg for 1 hr at4°C, and the magnitude of the concentration could then be controlled by the amountof media the pellet was resuspended in. Concentration of the HCMV stocks wasnecessary to attain the necessary MOl for the infection of T1-IP-1 cells and as apreparatory step for HCMV virion purification, and the concentration procedure alsoserved to remove any soluble factors (i.e. cytokines or interferons, etc.) in the media ofthe original stock and reduce extraneous variables. Propagation of the HSV-1 andhuman adenovirus stocks was performed by the labs from which they were obtained.2.7. Standard Plaque Assay.One confluent 100 mm dish of human fibroblasts was disaggregated usingtrypsin (0.25% bovine trypsin with EDTA, Gibco/BRL, Burlington, Ontario),resuspended in 12 mis of media and split between two 6-well dishes (Gibco/BRL,Burlington, Ontario). Next morning, the cells were usually at about 85% confluency,which is optimal for viral infection. The virus sample to be assayed was then seriallydiluted; 50 p1 added to 450 p1 media, the dilution was mixed, and 50 p1 removed foraddition to the next tube containing 450 p1 media, etc. One hundred microliters ofundiluted virus and each serial dilution were added to a corresponding labeled wellin the 6-well dish (in duplicate). Unconcentrated virus stocks were tested fromdilutions of 100 to io (effective dilution of 10-1 to 10-6, taking into account the factthat only 100 j.tl was added to each well), while concentrated stocks were tested todilutions as high as 10-8. Mock-infected controls were also run in parallel to compareuninfected cell morphology. The 6-well dishes were incubated for 1 h in the cellincubator and rocked every 10 minutes to stop the cell layer from drying out.Following the incubation, the virus was aspirated and the unbound virus removed by36washing with 1 ml of PBS. The dishes were incubated at 37°C, 5% CO2 for 10-14 daysand plaque development was checked daily by inverted microscopy. To ensure theplaques formed were the product of the initial infection and not subsequent release ofvirus into the supernatant by initially infected cells, the cells were overlaid withDMEM with 10% FBS containing 1% agarose or overlaid with DMEM with 5% FBScontaining 0.05 mg/mi of human IgG (containing high titers of anti-HCMV antibodiesas determined by functional neutralization assay). Both of these methods wereequally effective at stopping the formation of extraneous plaques. Visualization of theplaques was enhanced by fixing the cells for 10 min. at room temp with 5%formaldehyde in saline solution and then staining the cell layer with 0.05% methyleneblue in distilled water for 30 min. After rinsing the cell layer with distilled water andallowing them to dry, the plaques were counted under an inverted microscope andappeared as dark blue cells exhibiting CPE against a pale blue background.2.8. HCMV neutralization assay.Six-well dishes of fibroblasts were prepared as stated above, but the amount ofvirus was kept constant for each well (roughly 40-80 pfu per well, calculated anddiluted from known frozen HCMV stocks which had been previously plaqueassayed). For each well, 125 p.1 of HCMV stock was placed in a sterile eppendorf tube.The control well, which yielded between 40 and 80 plaques was mixed 1:1 with media,while the other viral aliquots were mixed 1:1 with human serum samples (eithersamples of undiluted serum or samples of the serum serially diluted in cell media).Each serum sample was tested as undiluted (1/2 dilution since it was mixed 1:1 withvirus) and also diluted by increments of 1/2 to a dilution of 1/1024. The sampleswere incubated with the virus aliquot for 30 min. at 37°C then duplicate 100 p.1samples were taken from each tube and incubated with the cell layers in the 6-welldishes for 1 h. Controls were incubated with serum only to identify a potentialpresence of HCMV in the serum sample. Following the infection of the cell layers for371 h at 37oC, the samples were aspirated and the unbound virus removed by rinsingthe cell layers with PBS. The cell layers were then allowed to develop as listed for theplaque assay. Variations of the described neutralization assay were necessary forsome experiments and these are identified in the results section.2.9. Western blot analysis.Whenever Western blot analysis was being used to compare protein expressionamongst cell types, or between HCMV-infected and uninfected cells of the same type,the numbers of cells per lane were equalized to a common cell number. Ifcomparisons were being made between cells and HCMV virions, the amount ofprotein per lane was equalized. Cells were disaggregated using 15 mM EDTA inphosphate-buffered saline (PBS) containing 30 mM sodium azide, centrifuged at 4000xg, and resuspended in 100 p.1 of EDTA/PBS and counted using a hemocytometer.The cell concentration of all samples was equalized by increasing the volume of themost concentrated samples. An equal amount of 2X non-reducing sample buffer (0.12M Tris (pH 6.8), 4% SDS, 20% glycerol, and 0.1% Triton X-100) was added to eachsample and the cells were lysed by repeated freeze-thaw cycles, followed by heatingin a boiling water bath for 5 mm. prior to loading on the SDS-polyacrylamide gel(SDS-PAGE; Laemmli, 1970). Whenever polyacrylamide gels were being used toinvestigate the presence of complement components on the surface of HCMV virions,each lane was loaded with an equivalent amount of protein or an equal volume if theprotocol had initially equalized the amount of virions and serum. Furthermore, whencomplement components were being investigated, the gels were run under reducingconditions (the 2X sample buffer contained 10% 1-mercaptoethano1).Four different concentrations of polyacrylamide were used for the lowerresolving portion of the gel depending on the size of the protein being investigated:5%, 7.5%, 10%, and 15% (in order for larger proteins to smaller proteins). The lowerresolving portion of the SDS-PAGE always contained the same concentrations of SDS38(0.01%), Tris (pH=8.8) (0.375 M), ammonium persulfate (0.01%), and TEMED (0.005%).Regardless of the polyacrylamide concentration of the resolving portion of the gel, theupper stacking portion of the gel was constant: 3.3% polyacrylamide with 0.26 M Tris(pH=6.8), 0.1% SDS, 0.1% ammonium persulfate, and 0.005% TEMED.Amersham rainbow-prestained protein mass standards (Oakville, ON) andBRL prestained protein mass standards (Burlington, ON) were concomitantly run oneach gel and used to estimate the relative molecular mass of the proteins beinginvestigated. The gels were run at a constant 30 mA per side until the bromophenolblue stain, from the sample buffer, was just above the end of the gel. The gels werethen removed and equalized for 30 mm. in Western blot transfer buffer (25 mM Tris,192 mM glycine, and 20% (v/v) methanol [pH=8.41) and then electrophoreticallytransferred (1.0 Amp Hour) to nitrocellulose paper. After blocking overnight with 3%powdered milk and 3% BSA in Tris-buffered saline containing 0.05% Tween 20(TBST), the membranes were incubated with polyclonal or monoclonal antibodies (inTBST) directed against the protein of interest. The primary antibody was detected bya subsequent incubation with a horse radish peroxidase-conjugated secondaryantibody (directed against the heavy and light chains of the immunoglobulin subclassfor the species the primary antibody was raised in) after the excess primary antibodywas removed by washing 3X in TBST. Excess secondary antibody was removed bythree washes in TBST, and immunoreactive bands were detected with the AmershamECL chemiluminescence Western blot detection kit and Amersham Hyperfilm X-rayfilm (Oakville, ON). Background staining patterns were always determined by preprobing the Western blot with the appropriate secondary antibody before the primaryantibody and any bands observed (usually none) were noted.2.10. Complement component co-purification with HCMV virions.The specific association of complement components with purified HCMVvirions was determined by incubating serum samples with virions under varying39conditions. HCMV virions were purified from cell fragments by discontinuoussucrose gradient centrifugation. The HCMV stocks used for the purification weregenerated by inoculating 5 large T-flasks at an MOI<0.01, allowing the monolayer toreach 100% CPE, and concentrating the virus down to 1 ml using the protocol listed inthe virus propagation section above. The discontinuous sucrose gradient was madeby layering 2 mis of each of the following in a 12 ml polyallomer tube, in order: 60%,50%, 40%, 30%, and 20% (w/w) sucrose in Tris-buffered saline (TBS) (12.5 mM Tris(pH7.4), 0.88% sodium chloride). The concentrated HCMV stock was resuspendedin 1 ml of reticulocyte standard buffer (10 mM Tris (pH=7.4), 10 mM KC1, 1.5 mMMgC12, 0.25 M sucrose) and 0.5 ml was loaded per discontinuous gradient. Thegradients were centrifuged in a Beckman SW3OTi rotor at 18,000 rpm for 2 h at 4°C,and the HCMV virions were found as an opalescent band at the interface between the50% and 60% sucrose layers. The band was extracted using a 3 cc syringe and a 20-gauge needle, and resuspended in 3.5 mls TBS and the virions were pelleted at 32,000rpm for 30 mm. at 4°C using a Beckman SW5O.1 rotor. As a control, an equal numberof fibroblasts used for initial virus propagation were disrupted by passing the cells 3Xthrough a 25-gauge needle, then treating the resultant suspension as though it werethe supernatant containing HCMV from a viral propagation; even though no bandswere seen on the discontinuous sucrose gradient nor any pellet in subsequentcentrifugation steps, the samples were collected at the same level and treated identicalto those samples containing HCMV virions (cell lysate control).For optimal complement activation conditions, the virion pellet wasresuspended in veronal-buffered saline (GVB; 5.0 mM sodium barbitol, 0.15 M NaC1,and 0.2% gelatin [pH 7.5]) containing 2.5 mM MgC12 and 0.15 mM CaC12(GVB2j.One hundred microliters containing 5x106pfu of HCMV virions in GVB wasincubated with each of the following: 100 iil of GVB2,100 pl HCMV-seropositiveserum, 100 jil seropositive serum with 15 mM EDTA or 15 mM EGTA, 100 i1seropositive serum depleted of IgG, and HCMV-seronegative serum. A control40containing only HCMV-seropositive serum and 100 jil GVB was also included toensure the post-incubation washes removed the unbound complement components.All mixtures were incubated for 30 mm. at 37°C then 3.5 ml of GVB without cationscontaining 15 mlvi EDTA (GVB-EDTA) was added and the virions pelleted bycentrifugation at 32,000 rpm in a Beckman SW 50.lTi rotor for 30 mm. The pellet wasresuspended in GVB-EDTA and the centrifugation procedure repeated. The pelletwas then resuspended in RSB and the non-specifically associated complementcomponents were removed from the HCMV virions by layering the mixture on top ofa 30% sucrose pad (3.5 ml) and centrifuging at 32,000 rpm in a Beckman SW 50.lTirotor for 1 h. This last wash step was repeated three times; the percentage of sucroserequired and the number of washes required to remove the unbound complementcomponents were empirically determined in the preliminary studies by using 100 iilof seropositive serum, without HCMV virions. Following the wash steps the resultingpellet was resuspended in 100 j.il of 2X reducing sample buffer and 25 .tl for eachsample was used for SDS-PAGE and Western blot analysis using polyclonalantibodies raised in goats against human IgG or human complement components C3orC9.2.11. Immunocytochemistry.For immunofluorescence studies on the glioblastoma cell line, U373-MG, andprimary fibroblasts, one hundred to 200 cells were grown overnight at 37°C on acid-etched coverslips. Polyclonal antibodies were used at a concentration of 10 ig/ml,the monoclonal anti-HLA class I was used at 5 jig/ml, and the fluorescein-conjugatedanti-mouse or anti-rabbit IgG secondary antibodies were used at a concentration of 1.tg/ml. All procedures were carried out at 4°C and all solutions were filtered prior tothe addition of antibodies and contained 30 mM sodium azide to prevent antibodypatching and capping. No aldehyde fixative can be used for the investigation of CD55expression since this completely abrogates antibody binding to this antigen.41Background staining was determined using normal rabbit serum or the monoclonalantibody directed against HLA class II as the primary antibodies. Cells to beexamined by epifluorescence were washed twice with phosphate-buffered saline(PBS; pH=7.4) then with PBS containing 1% bovine serum albumin (BSA). Primaryand secondary antibody incubations were for 30 mm. and separated by three rinseswith PBS/1% BSA. Following the secondary antibody incubation, the cells wererinsed once with PBS/1% BSA, once with PBS, mounted, and photographed with aZeiss Axiophot microscope equipped with epifluorescence optics.2.12. Flow cytometry studies.Adherent cell lines were disaggregated by incubation with 15 mM EDTA at4°C and suspended in cold flow cytometry (FC) solution (1% BSA, 15 mM EDTA, 15mM sodium azide in PBS). Exposure to EDTA did not interfere with antibodyrecognition of CD55, CD46, CD59, CR1 or HLA class I. All procedures were carriedout at 4°C and all solutions contained 30 mlvi sodium azide to prevent endocytosis orantibody patching and capping. Briefly, the cells were rinsed in the FC solution,centrifuged for 4 min. at 4,000 rpm in a Beckman TLA-100 centrifuge, andresuspended in FC solution at 5x106 cells/ml. Fifty microliters of cells were incubatedfor 30 min. with an equal volume of primary monoclonal antibody. The saturatingamounts of each monoclonal antibody were predetermined for each cell line.Following three washes in FC solution, cells were incubated in fluorescent-labeledsecondary antibodies. After the unbound secondary antibodies were removed, cellswere analyzed on an EPICS Profile flow cytometer and using Cytologic software(Coulter, Hialeah, FL). Non-specific antibody binding was determined by usingmonoclonal anti-MHC class II as the primary antibody (5 j.tg/ml) or the monoclonalcontrol antibody directed against a non-mammalian protein (see antibody section; 1j.tg/ml). Repeated measures analysis of variance (ANOVA) was used to initiallydetermine statistically significance. One-way ANOVA and post-hoc analysis using42Fisher’s least signfficant test and Bonferroni’s correction for multiple comparisons wasused to identify differences between groups.2.13. Northern blot analysis.Total RNA was extracted from HCMV-infected or uninfected cells at varyingtimes post-infection using the guanidium isothiocyanate method (Turpin andGriffiths, 1986). Briefly, monolayers of fibroblasts or glioblastoma cells, which wereuninfected or infected with HCMV (MOI=10), were grown in three 150 mm for eachcondition investigated (i.e. mock- vs. HCMV-infected) and harvested at various times.At the time of harvest, the monolayers were washed once with PBS and 7 mis ofguanidium isothiocyanate solution (4.5 M guanidium isothiocyanate, 5 mM Na-citratepH 7.0, 0.1 M 2-mercaptoethanol, 0.5% sarkosyl) was added to each of the threemonolayers sequentially so that total volume did not exceed 8.5 mls; all residualswere collected by scraping. The cells were disrupted by three passages through a 25-gauge needle. The cell lysate is layered on top of 4 mis of 5.7 M cesium chloridecontaining 50 mM EDTA (pH=7.4), which is in an 11 ml polyallomer quickseal tube.The tubes were balanced, sealed, and centrifuged at 60,000 rpm for 6 hours at 20°C ina Beckman near vertical rotor. The RNA pellet was resuspended in 10 mM Tris-HC1(pH=7.4) containing 2 mM EDTA and 0.5% SDS, then extracted twice with 12 mls of1:4 butanol:chloroform. The aqueous phase was then ethanol precipitated with 12 mlsof 100% ethanol following the addition of 1/10 vol. of 3.0 M Na acetate. The totalcellular RNA pellet was then resuspended in 100 jil of distilled water and the RNAconcentration measured at 260 nm. RNA was stored at 700C after the sample hadbeen adjusted to a constant RNA concentration of 10 jig! jil.Thirty micrograms (3 jii) of total RNA were separated on a formaldehyde (6%)-agarose (1%) gel after being mixed with 5 j.tl RNA sample buffer (55 mM MOPS, 6%formaldehyde, 5% glycerol, 0.1 mM EDTA, 0.04% bromophenol blue, 0.04% xylenecyanol) and being denatured at 65°C for 15 mm. Total RNA from HeLa cells (10 jig)43was included as the positive control for CD55. The RNA was osmotically transferredto non-charged nylon membranes using 3 M NaC1/0.3 M Na-citrate, and hybridizedwith radiolabeled RNA anti-sense probes specific for CD46, CD55 or CD59 mRNA.Radio-labeled anti-sense probes were made using a random primer kit (BethesdaResearch Laboratories, Burlington, ON) and[32P]-labeled dCTP (ICN; St. Laurent,Quebec). Plasmids containing CD55 cDNA were provided by D.M. Lublin,Washington University, plasmids containing CD59 cDNA were provided by W.F.Rosse, Duke University, and plasmids containing CD46 cDNA were provided by J.P.Atkinson, Washington University.2.14. Radio-immunoprecipitation studies.Monolayers of fibroblasts growing in 60 mm dishes were infected with HCMVat a MOI=10, while control cells were uninfected. After 1 h the virus was removedand replaced with complete medium. At 24 h p.i. cells were washed three times withcysteine-free medium and labeled for 50 mm. with 100 j.tCi/ml [35S]-cysteine (1117Ci/mmol) in cysteine-free medium containing 5% dialyzed FCS. At the completion oflabeling, cells were either harvested immediately or washed three times andincubated in DMEM with 10% FBS (which contained excess cold cysteine) for 2 or 4 h.Cells were harvested by washing the monolayer with cold PBS and incubated for 15min. with 1 ml cold lysis buffer (10 mM Tris-HC1 pH 7.4, 150 mM NaC1, 1% NonidetP-40 (NP-40), 1% Na-deoxycholate).Immunoprecipitations were performed by first adding 1/10 volume of amixture of 10% NP-40, 10% Na-deoxycholate, and 1% SDS to aliquots of cell lysates tobe precipitated. The samples were precleared by incubation with protein G-sepharose(pre-washed in 20 mM phosphate (pH=7.4), containing 2% BSA) while rocking at 4°Cfor 2 h. The protein G-sepharose was pelleted by using a benchtop Beckmancentrifuge, the pellet was saved as the preclear control, and the supernatanttransferred to a new tube. Ten microliters of rabbit polyclonal anti-CD59 antibody44was added to each sample and incubated, while rocking at 4°C, for 2 h; at which timeprotein G-sepharose was added to capture the anti-CD59 antibody as described abovefor preclearing. All protein G-sepharose pellets (including the preclear controls) werewashed with 0.5 ml of the following solutions; wash buffer #1 (20 mM Tris-HC1 pH7.5, 150 mM NaC1, 1% NP-40), wash buffer #2 (20 mM Tris-HC1 pH 8.8, 150 mM NaC1,1% NP-40, 0.2% SDS), wash buffer #3 (20 mM Tris-HC1 pH 6.8, 150 mM NaC1, 1% NP -40,0.2% SDS). The final pellet was resuspended in 2X reducing sample buffer, heatedat 1000C for 5 mm. and separated by SDS-PAGE. Gels were fixed for 45 mm. in 7%acetic acid containing 25% methanol, dried, and exposed to Kodak XAR-5 film forautoradiography (usually between 3 days and 2 weeks).2.15. Alternative pathway C3 convertase activity assay.The protocol used to measure C3 convertase activity was a modified version ofone previously published for platelets (Devine et al 1987). HCMV-infected oruninfected cells were harvested and incubated with 30 mM NaN3, 30 mM 2-deoxy-D-glucose, 10 mM gluconic acid-delta-lactone in PBS to block cell metabolism. Equalnumbers of infected or uninfected cells were suspended in magnesium-free GVB(0.2% gelatin in veronal-buffered saline), 1 mg of C3, and 200 j.tl of preformed fluidphase C3 convertases (cobra venom factor (CoVF) factor B, and factor D). After 30min. at 37°C, fluid phase C3 convertases and any unbound C3 were washed away,and half of the cells were removed for quantitation of baseline C3 deposition. Thecells were resuspended in GVB with 5 mM MgCl 50 jil factor B and 50 jil factor Dwere added and incubated for 10 min. at 37°C. One milligram of C3 was added andthe incubation was stopped after 1 h at 37°C by the addition of 20 mM EDTA. Afterwashing, the amount of cell-bound C3 on all cell preparations was determined byflow cytometry using monoclonal anti-C3d (Gibco/BRL, Burlington, ON). In theabsence of factor B, anti-C3d binding was equivalent to that seen with non-immuneantibody controls. Data were pooled for time points between 36 and 72 h p.i. (N=4),45as no significant trend could be established between these time points. Statisticalsignificance was determined by one-tailed paired t-tests.2.16. CD55 promoter constructs.The CD55 bacteriophage genomic clones were previously identified, mapped,and partially sequenced (Post et al 1990). A 4.6-kb Hindill fragment encoding the 5?flanking region, 5’-untranslated region/signal peptide, and part of short consensusrepeat 1 was subcloned into pBluescript KS (Statagene, San Diego, CA). Subsequentrestriction digests and manipulations were performed as described in Thomas andLublin (1993), and I received the following 5’-deletion constructs of the CD55promoter from Dr. DM Lublin who is located at Washington University (all of whichcontained the transcriptional start site and 84 additional base pairs of the transcribedregion; see figure 34): -2800, -796, -206, -77, -54, and -36. The basal promoter activity ofall of these, except the -2800 to +84 construct, were previously reported whentransfected into K562, EBV-transformed B-cells, Molt 4 and HeLa cells (Thomas andLublin, 1993). I created two additional5t-deletions from the -796 to +84 construct toinvestigate the large region left between the -796 and -206 from the receivedconstructs. The CD55 promoter had been inserted at the BssH II restriction site in theSP65 polylinker region, which meant the Pst I and Xba I (Gibco/BRL, Burlington, ON)restriction sites from the SP65 polylinker region were located between the +84 end ofthe CD55 promoter and the start of the 5’-end of the chioramphenicol acetyltransferase (CAT) reporter gene. By removing the Pst I and Xba I restrictionfragments from the -796 to +84 CD55 promoter in SP65-CAT plasmid, then insertingthe isolated fragment into a SP65-CAT plasmid (devoid of previous CD55 promoterinsertions); the -425 to +84 and -275 to +84 CD55 promoter constructs were created.The restricted fragments were isolated by separating the cut plasmid on a 1% agarosegel, staining the gel with ethidium bromide, excising the band of appropriate sizeunder UV-light visualization, purifying the DNA using the Geneclean purification46system (Gibco/BRL, Burlington, ON), and ligating the fragment into SP65-CAT (precut with the appropriate restriction enzyme) with T4 ligase (Gibco/BRL, Burlington,ON). Since two orientations were possible, the correct orientation of single colonyplasmid DNA was identified by Sac I restriction enzyme (Gibco/BRL, Burlington,ON) digest after separation on a 1% agarose gel and staining with ethidium bromide.The SP65-CAT plasmid contained the ampidilin resistance gene so all plasmids werepropagated in transformed DH5 E. coli (subcloning efficiency) grown in 2X YTmedia (1.6% tryptone, 1.0% yeast extract, 0.5% NaCl) containing 100 jig/mi ampicillin,or on plates made from 2X YT media containing 1.5% agar and 100 .tg/ml ampicillin.2.17. Other plasmids.The (3-galactosidase ((3-gal) or Lac-Z gene in the Rc/CMV plasmid (Invitrogen,San Diego, CA) was a gift from Dr. X. Wu (University of British Columbia) and wasutilized to measure the efficiency of transfection. The isolated HCMV immediateearly (IE) genes were obtained from various sources: Plasmids containing the Ut 36,UL37, UL38 and UL36-38 IE genes were a generous gift of Dr. AM Colberg-Poley(George Washington University, Washington D.C.) and are described in ColbergPoley et al (1992). Plasmids containing the TEl, 1E2, and US3 TE genes were a generousgift of Dr. R Ruger (Boehringer Manrtheim, Penzberg, Germany) and are described inColberg-Poley et al (1992).2.18. Plasmid transfection into cells.Plasmid DNA to be used for transfection was purified from RNA, proteins, andbacterial DNA by banding the plasmid DNA in 1.2 g/ml cesium chloride bycentrifugation at 100,000 rpm at 25°C in a Beckman near-vertical rotor in a Beckmanbenchtop ultracentrifuge. The bands were visualized by pre-addition of ethidiumbromide; after extracting the band with a 20-gauge needle and 3 cc syringe, theethidium bromide was removed by 3 consecutive n-butanol extractions followed by47ethanol precipitation. The DNA concentration was measured at 260 nm and the DNAconcentration standardized to 1 jig! jil prior to storage at -70°C. The glioblastoma,fibroblast, and HeLa cells grown in 6-well dishes at 70% confluency were transfectedusing the calcium chloride method: 5 jig plasmid DNA was co-precipitated with 20 jigof salmon sperm DNA in HEPES-buffered saline (pH=7.0) with 2 M CaC12 (per well).The cells were shocked with 20% glycerol in HEPES-buffered saline (pH=7.0) 12 hlater, and some of the glioblastoma cells and fibroblasts were infected at 24 h posttransfection for 1 h. For transient co-transfection assays investigating, 5 jig of eachplasmid was added to the transfection mixture (including a plasmid containing the 3-gal gene), no super-infection by HCMV was performed, and cells were harvested at 48h post-transfection for measurement of CAT activity or flow cytometry. The cellswere then harvested 48 h after being glycerol shocked and the intracellular CATactivity and mock controls were cells treated identically, but the 5 jig of plasmid DNAwas replaced with sterile water.2.19. Measurement of CAT activity.The intracellular CAT activity, which could only have originated from theplasmid, since mammalian cells do not naturally contain this gene, was assayed usinga kinetic partitioning assay (Neumann et al, 1987). The cells were harvested bytrypsinization and each well was resuspended in 1 ml PBS, put in a 1.7 ml eppendorftube, and pelleted by centrifugation in a benchtop Beckman centrifuge. Each cellpellet was resuspended in 100 jil of PBS and the intracellular CAT was released bythree cycles of freeze-thaw using a 100% ethanol!dry ice bath and a 37°C water bath.The cell debris was cleared by centrifugation (15,000 rpm in a Biofuge 17 centrifuge)and 30 jil of each sample was put in a scintillation vial. After mixing each samplewith 210 jil of the CAT reaction mixture (100 mM Tris-HC1 (pH=7.8), 1.0 mM[14C]-acetyl CoA (0.1 jiCi), and 1.0 mM chloramphenicol) and 5 ml of the hydrophobicscintillation fluid, Econofluor (Dupont NEN research products, Boston, MA), was48carefully layered on top of the reaction mixture. Since the covalent attachment of thechioramphenicol to the[‘4C]-acetyl-CoA makes the latter more hydrophobic, the basicprinciple of this assay utilizes the fact that only the covalently linked[14C] will enterthe hydrophobic scintillation fluid and be detected by the scintillation counter.Therefore, by counting the vials at regular intervals and graphing the cpm as afunction of the accumulated time for that count, the amount of CAT present in thesample is directly proportional to the slope of the graph. This technique also allowedthe investigators to measure the activity in the linearly increasing portion of thereaction and avoid artifactually low readings from points at which the concentrationof the substrates became limiting. CAT activity measured for each sample wasperformed in triplicate, and each finding was replicated at least three times.2.20. 3-gal activity assay.For all transient co-transfection assays and transfected cells used for flowcytometry, parallel cell cultures were also created to investigate the transformationefficiency. The monolayers were grown in 6-well dishes and the cells were washedtwice with PBS prior to fixation with 4% formaldehyde and I3-gal activity investigated.Cells containing the 13-gal gene turn blue when incubated overnight with PBScontaining 2 mM MgC1, 10 mM FeCN2,10 mM FeCN3,and the chromogenic substratefor 13-gal, 1 .ig/ j.tl (0.1%) X-gal (Gibco/BRL, Burlington, ON). The transfected cellmonolayer was observed under the microscope: the total number of cells and numberof blue cells were counted for 10 fields, and the transfection efficiency estimated fromthe ratio of these two numbers.493. RESULTS:3.1 CHAPTER 1. Complement effects on HCMV virions.The presence of complement has been reported to enhance the ability ofspecific anti-HCMV antibodies to neutralize HCMV virions, as detected by plaqueassay (Lewis et al 1986, Eizuru et al 1988). However, the role of complement alone inHCMV neutralization has not been addressed, nor has the pathway responsible forthe enhanced antibody neutralization. These questions will be addressed in thischapter.3.1.1 The role of complement in neutralization of HCMV virions.The first set of experiments were designed to determine the role of complementin the neutralization of HCMV. Most of the experiments in this section utilize avariation of the standard plaque assay, called the neutralization assay. In this assaythe amount of virus was kept constant (roughly 60-80 plaques per 9.6 cm2)but thehuman serum mixed and incubated with the virus was serially diluted. Therefore, themore neutralization activity present in the serum the greater the serum dilutionrequired before plaques were observed. The dilution of serum which yielded half theplaques compared to virus mixed with media only is referred to as the NT50,and wasreported as the inverse of the dilution (i.e. a dilution of 1/64 is reported as 64). Thepresence of anti-HCMV antibodies was also measured on the same samples by theAbbott EJA method at the Vancouver branch of the Canadian Red Cross BloodTransfusion Services. Interestingly, incubation of seronegative serum with virusresulted in NT50 ranging from <2 to 8 (3 representative samples of more than 10experiments shown in Figure 4). Heat-inactivation of the complement in seronegativeserum samples from donors #1,2, 7, and 15 did not effect the results, nor did theaddition of rabbit complement to sample #7 (Figure 4 and 5), which suggests theseobservations cannot be attributed to the effect of complement. The NT50 values forthe seropositive individuals #9, 20, 26, and 27 ranged from 64 to >256 (Figure 6 and 7),50but unlike the seronegative samples, heat inactivation of the serum sample #9decreased the neutralization activity and adding rabbit complement to sample #9increased the activity greatly (Figure 7). The difference between the effects of theintrinsic complement and the added rabbit complement likely represents the relativedilution of the intrinsic complement, coincident with the dilution of the antibodies.By adding an equal amount of undiluted serum to an aliquot of a viral stock,the difference in neutralizing activity between seronegative and seropositive serumsamples is better illustrated by standard plaque assay (Figure 8). A twenty percentdecrease was observed when seronegative serum was mixed with the HCMV stockcompared to a 3800-fold decrease when mixed with seropositive serum. Again heat-inactivation of complement did not vary these effects in the seronegative sample;however, heat-inactivation of the seropositive serum decreased the neutralization byhalf. In a separate experiment using a different set of donors, depleting the IgG fromthe seropositive sample using protein G-sepharose reduced that serum sample’sabffity to neutralize HCMV to the level of a seronegative serum sample.It is important to note that the HCMV serological status for all individuals wasgauged by their Abbott ETA test results. There may be some limitations for this testsince it is a first generation ELISA. Repeat testing of all individuals before and afterassured that the individuals were not in the process of seroconverting, since theAbbott ETA is designed to detect the IgG class of immunoglobulins. More sensitive,specific peptide ELISA tests are available now, but these tests have limitations as well,and none of the assays available are very good at determining the presence of weaklycross-reacting antibodies directed against other herpesviruses. However, thesimilarity between seronegative serum and the IgG-depleted serum in the pages thatfollow provide the best argument for the accuracy of the Abbott ETA determinations.51L Figure 4. Serum samples obtained from three individuals (coded as #1, #2, and#15),which were determined to be negative for the presence of anti-HCMVantibodies by the Abbott EIA method, were subsequently used in a functionalHCMV neutralization assay. The number of plaque-forming units present, after a30 mm. incubation of a virus aliquot with serial dilution for each serum, wasexpressed as a percentage of the number of plaques formed when only tissueculture media was added to the HCMV aliquot. The percentage of plaques formedis compared to the final dilution of serum used in the graph below. Open symbolsrepresent the same samples tested after the intrinsic complement components wereheat inactivated (56°C, 30 mm.). Each point represents the average determined ontwo aliquots of the same sample run simultaneously.100— 80Ce 60 —.—— — — — — —— —— —0-- #1 no C’ — — NT50• 40 —A-- #2-i-C’—h— #2noC’20 —*-- #15+C’—e-- #l5noC’2 4 8 16 32 64Dilution (inverse)52SI- Figure 5. Using the same neutralization assay in figure 4, HCMV aliquots wereincubated with serial dilutions of another serum sample (found to be seronegativeby the Abbott EIA method). The variables tested include neutralization activitywith the intrinsic complement activity left intact (closed square), neutralizationactivity with the intrinsic complement activity removed via heat inactivation (opencircle), or neutralization activity in the presence of additional rabbit complement(to a final concentration which is two-fold higher than the intrinsic complementconcentration; closed triangle). Each point represents the average determined ontwo aliquots of the same sample run simultaneously. Identical results wereobtained when seronegative serum samples #42 and #15 were used.10040200•• #7+ C’-o #7noC’i #7RabC’——NT502 4 8 16 32 64Dilution (inverse)53..L Figure 6. Serum samples obtained from three individuals (coded as #20, #26,and #27) were determined to have anti-HCMV antibodies by the Abbott EIAmethod and were subsequently used in the HCMV neutralization assay outlined inFigure 4. The graph below demonstrates the percentage of plaques formed as afunction of the dilution of serum. Heat inactivation of intrinsic complementcomponents was not used as a variable in this experiment because of the highdilution factors involved. Each point represents the average of two aliquots of thesame sample run simultaneously.100_______________* Patient #26— 80 4 Patient #20* Patient #278 60N’1’50 —— — — — —04 8 16 32 64 128 256 512Dilution (inverse)54Figure 7. Using the same neutralization assay in figure 4, HCMV aliquots wereincubated with serial dilutions of another serum sample (found to be seropositiveby the Abbott EIA method). The variables tested include neutralization activitywith the intrinsic complement activity left intact (closed square), neutralizationactivity with the intrinsic complement activity removed via heat inactivation (opencircle), or neutralization activity in the presence of additional rabbit complement(to a final concentration which is two-fold higher than the intrinsic complementconcentration; closed triangle). Each point was determined in duplicate andidentical results were obtained when seropositive serum sample #26 was used.100... #9+C’80 #9 Ct: zzz8 16 32 64 128 256Dilution (inverse)55Figure 8. A standard plaque assay was performed on an HCMV stock afteraliquots of HCMV had been incubated with an equal volume of medium (viralstock), seronegative serum (#42), or seropositive serum(#20). The effect of theintrinsic complement was determined by heat-inactivating (56° C, 30 mm.) bothserum samples in a parallel incubation (labeled as 3heated). The presence ofHCMV in the serum samples was also determined and found to be the same as themock infection. Each value was determined in duplicate and the actual values aregiven at the top of each bar.107 3.25x106i06io5io41021010 00• i I I&•04C,,2.6x10 2.6x10$563.1.2 Neutralizing titer correlation with amount of specific anti-HCMV antibody.An Abbott EIA was used to assess the level of anti-HCMV antibody. Althoughthe supporting documentation provided with this test kit suggested that the amountof colorimetric product produced (measured by absorbance) was directlyproportional to the amount of anti-HCMV antibody present in the sample, we failedto find a close correlation in all samples. In general, dilution of the serum samples diddecrease the absorbance, but many of the samples did not decrease linearly, somevalues even increased slightly for serial points of sample dilution (Figure 9).Comparisons were then made between the anti-HCMV antibody titer and thefunctional neutralizing ability of the serum sample as determined by the NT50 assaydiscussed in the previous section. If the NT50 values are graphed as a function of theantibody titer, there is stifi no correlation observed (Figure 10). This may reflect theepitopes identified by the various antibodies produced during affinity maturation.3.1.3 Identification of complement activation pathway with purified virions.These experiments were designed to identify the complement activationpathway responsible for the enhancement of antibody-mediated neutralization. Therole of antibodies was addressed by two different methods: First, sucrose purifiedvirions were incubated with both seropositive and seronegative serum samples, andsecond, with seropositive serum samples in which the IgG was removed by a proteinG-Sepharose. The Western blots shown in Figure 11 demonstrate the fractionscollected after IgG-depletion using a protein G-Sepharose column (A), and thesubsequent elution proffle of the bound IgG with a 0.1 M glycine buffer (B). Theactivity of both the alternative and classical activation complement pathways of thedepleted serum were within the normal range as assessed by functional hemolyticassays performed in the routine clinical laboratory at University Hospital (Vancouver,B.C.). The classical pathway function was unaffected by IgG depletion because theantibody used in the assay is an exogenous anti-rabbit erythrocyte antibody.57.1’ Figure 9. Serum samples from 6 individuals tested positive for anti-HCMVantibodies by the Abbott BIA method. These samples were then serially diluted inPBS, and the absorbance of the dilutions determined by this method. The cut-offabsorbance (dotted line) was the value=0.123; absorbance readings below thisabsorbance were considered seronegative. The absorbance readings do not alwaysdecline linearly, and higher initial values do not always correspond to a higherextinction dilution (dilution which falls below the cut-off value).1.5—--Positive #5-+—Positive #10—A--Positive #20—B—Positive #21——Positive #230______________Cut-off value: Abs = 0.123.01 2 4 8 16 32 64 128 256Dilution (inverse)581 Figure 10. Seven seropositive samples, for which the antibody titer wasdetermined (by extinction dilution as demonstrated in Figure 9) were subsequentlytested in the functional HCMV neutralization assay. This graph depicts the NT5Ovalues graphed as a function of the extinction dilution: again the correlation wasfairly low(R2=O.315)300250200150I I I I II4 8 16 32 64 128 256 512NT50 (inverse)59Figure 11. (A). Western blotAanalysis for IgG present infractions collected after123456 78203- using a protein-G sepharose974 column to deplete IgG from70.8-4— Han HCMV-seropositive43.6-serum sample. Samples28.3- -4- L from the first 8 fractionswere separated underB.reducing conditions by SDSPAGE. After transfer to1 2 3 4 5 6 7 8 nitrocellulose the Western203- blot was probed with a97.4 polyclonal anti-human IgG-4’.- H antibody. The first fourfractions contained a28.3-4— L majority of the total proteinand were pooled andconcentrated to their original volume (see materials and methods). As seen in (A)these fractions contained negligible amounts of the heavy chain of IgG (53 kDa),while this band is readily apparent in the 1/20 dilution of a sample of HCMVseropositive serum (NHS) run as a control. (B). The same column wassubsequently rinsed with a 0.1 M glycine buffer and samples of the eluted fractionstreated the same as for (A). The 53 kDa heavy-chain of IgG is readily apparent infractions 5-8, and are much stronger than that observed for the 1/20 NHS control.Pre-probing both (A) and (B) with the secondary antibody yielded no signal ineither case.60Sucrose-density gradient purified HCMV virions were incubated with serumsamples under various conditions for 30 minutes at 37 OC, and the unbound solubleproteins removed by a series of sucrose-density gradient purifications. lii addition tounaltered seropositive serum, HCMV virions were also incubated with seropositiveserum in the presence of EGTA or EDTA. EGTA is a specific calcium chelator whichinhibits the classical complement activation pathway, but does not affect thealternative pathway since it only requires magnesium. However, EDTA chelates bothmagnesium and calcium and serves to identify non-specific association ofcomplement components with HCMV virions since activation cannot occur withoutthese divalent cations. Under these various conditions, a large amount of IgG wasbound to the virions incubated with seropositive serum, regardless of the presence ofEDTA or EGTA (Figure 12). No IgG co-purified with virions incubated with IgGdepleted seropositive serum, and a very weak signal was observed for virionsincubated with seronegative serum. The antibodies associated with virions afterincubation with seronegative serum may represent cross-reacting antibodies directedagainst other herpesviruses or false-negative results generated by the ELISA methodused to test the serum. However, these possibilities are unlikely: equivalent amountsof C3 deposition are seen between virions incubated with seronegative serum andserum depleted of all IgG (see below), and these results were consistent amongst fourseparate seronegative serum samples in my preliminary investigations. Only theheavy (H)-chain was observed by PAGE under reducing conditions which mayindicate a higher avidity of the polyclonal anti-human IgG anti-serum for 53 kDa Hchains than 25 kDa light (L)-chains.We next studied the complement activation of the classical and alternativepathways. The co-purification of C3 with virions incubated with seropositive serumin the presence of EDTA (where no complement activation should occur) indicated alow level of non-specific association of C3 with HCMV virions under theexperimental conditions used (Figure 13). The C3 associated with the virions61Figure 12. Western blot analysis for IgG bound to the surface of virions. Virionswere first incubated with serum samples then unbound proteins removed bysucrose density gradient purification. Samples of virions were separated underreducing conditions by SDS-PAGE, and a large amount of the 53 kDa heavy chainof IgG co-purified with virions incubated with seropositive serum, regardless ofthe presence of EDTA or EGTA. Negligible amounts of IgG were observed whenseropositive serum depleted of IgG was incubated with purified virions, and theweak band observed for seronegative serum and virions indicates the non-specificassociation under these conditions. A 1150 dilution of HCMV-seropositive serum(NHS) was used as a control and pre-probing of the blot with secondary antibodyonly yielded no signal. IgG-depleted = seropositive serum depleted of IgG with asepharose column as in Fig. 11, pos. = seropositive serum.cJ‘I_C.’.2VJ 0 0— Cl)70.8-a43.6-p28.3-62-L Figure 13. Western blot analysis for C3 bound to the surface of virions. Virionswere first incubated with serum samples then unbound proteins removed bysucrose density gradient purification. Samples of virions were separated underreducing conditions by SDS-PAGE, and the largest amount of the 75 kDa 13-chainof C3 co-purified virions incubated with seropositive serum, in the absence ofEDTA or EGTA. The amount of C3 associated with the virions incubated withseropositive serum in the presence of EDTA is the lowest and represents the non-specifically associated C3. The amount of C3 associated with virions incubatedwith IgG depleted seropositive serum, seropositive serum with EGTA, andseronegative serum was intermediate and represents the background plus a lowactivation of the alternative pathway. The higher Mr smearing indicatescovalently-bound C3 a-chain, and the 43 kDa band represents the proteolyticallyprocessed C3dg fragment of the a-chain. These findings were consistent regardlessof whether the incubations were run at low or high ionic strength. The samecontrols and abbreviations as listed for Fig. 12 apply.C Gb .nil L203-97.4-70.8-43.6-2R-4-. c3dg-.-/ ---—i63incubated with seronegative serum, seropositive serum + EGTA (where only thealternative pathway is active), and IgG-depleted seropositive serum were verysimilar, indicating some alternative pathway activation does occur. However, muchmore C3 is associated with the virions incubated with seropositive serum in theabsence of EGTA and EDTA. This indicates that the activation of the classicalpathway is the largest contributor to complement activation products on the virionsin the presence of virion-bound antibody. Furthermore, greater evidence of covalentlinkage of the x-chain of C3, as visualized by higher bands and darker smearing at ahigh molecular weight by SDS—PAGE, was prevalent for the seropositive serumsample. Interestingly, a band was seen at 43 kDa which is the expected molecularweight for the C3dg fragment of C3. The molecular weight is slightly higher than theband seen in the NHS control lane, and it is possible that the band in the control lanerepresents the smallest fragment, C3d. In vitro , the C3g fragment, which has an Mr of5 kDa, can be cleaved from C3dg by using trypsin (Ross and Medof 1985). Thus, C3dmay be the fragment produced during the processing and repeated freezing andthawing of the control serum (the same aliquot was used as a control for multiplegels), rather than the fragment produced by C3b inactivation on the surface of virions.The levels of C3dg corresponded to the general levels of C3 co-purifying with thevirions, and the possibility of a cross-reaction with the H-chain of IgG was ruled outby subsequent staining of the same blot with anti-human IgG antibody; whichidentified a band about 8 kDa larger. This implies that C3 bound to HCMV virions isprocessed by factor Ito inactive fragments. These results were identical regardless ofwhether the experiment was performed under low ionic strength conditions (allcomponents suspended in GVB; Figure 13) or under physiological ionic strengthconditions (all components suspended in TBS).Finally, the extent of complement activation was investigated by staining forthe presence of C9 associated with the virions. No C9 was found to associate with thevirions in the presence of EDTA, suggesting minimal non-specific association. The64Figure 14. Western blot analysis for C9 bound to the surface of virions. Virionswere first incubated with serum samples then unbound proteins removed bysucrose density gradient purification. Samples of virions were separated underreducing conditions by SDS-PAGE, and the largest amount of the 71 kDa C9protein co-purified with virions incubated with seropositive serum, in the absenceof EDTA or EGTA. Low amounts of C9 were associated with virions incubatedwith IgG depleted seropositive serum and seronegative serum, while the amount ofC9 associated with virions incubated with seropositive serum and EGTA wasintermediate. No C9 was associated with the virions in the presence of EDTA,indicating negligible background. These findings were consistent regardless ofwhether the incubations were run at low or high ionic strength. The same controlsand abbreviations as listed for Fig. 12 apply.‘a. .97.4-...70.8- -‘- 4—C943.6-.28.3-65amount of C9 associated with virions incubated with seronegative serum and IgGdepleted seropositive serum was low, a little more was observed with seropositiveserum + EGTA, and the most C9 was associated with virions when incubated withseropositive serum in the absence of EDTA or EGTA. Again, these results wereconsistent whether the experiment was conducted under low or physiological ionicstrength conditions (Figure 14). This suggests that minimal activation of complementoccurs in the absence of anti-HCMV antibodies; but when anti-HCMV antibodies arepresent, activation occurs mainly by the classical pathway and proceeds tocompletion with the deposition of C9 into the virions.3.1.4 Host Complement Inhibitors associated with HCMV Virions.The presence of what appears to be C3dg with HCMV virions suggests thatsome complement regulation may also occur on the virions. To address the possiblepresence of host encoded complement inhibitors on the virions, Western blottingtechniques were utilized on sucrose-density gradient purified HCMV virions todetermine whether host-encoded complement inhibitors were present. Further,Western blot results were compared for virions, HCMV-infected cells, and uninfectedcells. A cell lysate control was prepared by lysing an equal number of fibroblasts(equal to the amount used to propagate the virus) using a fine needle and the celldebris was then treated identically to the virus preparation. This control addressedthe possible non-specific co-migration of host cell complement inhibitors to the samedensity at which HCMV virions are found, thus causing artifacts. Figure 15demonstrates the presence of CD46, CD55 and CD59 in the lanes with purifiedHCMV virions. The lysate control also identifies the presence of a small amount ofCD59, indicating a small non-specific co-migration of CD59 at the same density as theHCMV virions. However, a much larger amount of CD59 is present in the purifiedvirion sample; this enrichment suggests CD59 is associated with the virions.Interestingly, no CD46 or CD55 was detected in the lysate control, while vast amounts66of CD55 and a lesser amount of CD46 were associated with the HCMV virions. Sinceequal amounts of purified virions were used in Figure 15 A, B, and C, the relativeenrichment of CD55 and CD59 associated with the HCMV virions is probably areflection of their common glycolipid anchor, as compared to the transmembranecellular attachment of CD46. The ease of incorporation of glycolipid anchoredproteins into exogenous membrane surfaces was a phenomenon which aided in theinitial determination of the complement regulating ability of CD55 and CD59 (Davitzet al 1986, Holguin et al 1989). Further, the apparent molecular mass for the CD55associated with virions was more variable than the CD55 found on uninfectedfibroblasts or fibroblasts infected with HCMV for 72 h (Fig 15C). Cross-reactivity ofthe anti-CD55 monoclonal antibody with a viral protein was ruled out by obtainingsimilar results with monoclonal anti-CD55 antibodies directed against differentepitopes. There is also no homology between the region bound by the anti-CD55monoclonal antibody and the prospective proteins determined using the publishedsequence of AD169 and the computer programs Geneworks and DNA StriderThe Mr of the CD46 on the purified virions is not as variable as the virion CD55,nevertheless, the virion CD46 had a higher Mr than the uninfected fibroblast orfibroblasts infected with HCMV for 72 h. These differences in CD55 and CD46 Mrbetween cells and virions may represent altered glycosylation occurring in cells whichhave been infected with HCMV for 10-12 days during HCMV propagation. Otherauthors have noted that the SDS-PAGE separation of HCMV-encoded virion glycoproteins results in disperse bands (Benko and Gibson 1986). There even appears to bemore glycoprotein size heterogeneity in HCMV-encoded glycoproteins whencompared to simian CMV. In preliminary experiments in which the HCMV-infectedfibroblasts were harvested at times later than 72 Hp.i., the altered forms of CD55 wereseen and became predominant compared to uninfected cell CD55 as the cellsapproached death.67‘1- Figure 15. Western blot analysis for the presence of CD46 (A), CD55 (B), or CD59(C) in sucrose density gradient purified virions as compared to HCMV-infected anduninfected human foreskin fibroblasts (HuFF). The cell lysate control representsuninfected HuFFs mechanically disrupted and then exposed to the samepurification procedure as the virus stocks.‘ All proteins were separated under nonreducing conditions in the presence ofc)A. 0.05% NP-40 detergent. The samples usedThc 8 for CD55 and CD46 detection wereseparated on 7.5% polyacrylamide gels,97.4-while samples used for CD59 detection70.8- 4 were separated on 15% polyacrylamide gels.A polyclonal rabbit anti-CD59 and mouse43.6-monoclonal anti-CD55 antibodies were28.3- 1 used, but no signal was seen with preprobing with the secondary antibody.el C4 Gjt.. - t,_• gC)B. C.z.—97.4- 43.6-28.3-708-179-436- 151-28.3-683.2 CHAPTER 2. Changes in complement inhibitor expression on adherent cellsinfected with HCMV.Concurrent with my studies of purified HCMV virions, I investigated thepossible alterations in host complement inhibitor expression on HCMV-infected cells.3.2.1 Complement inhibitor expression on uninfected cells.Only certain human cell types are fully permissive for HCMV infection. Iselected two different fully-permissive cell types: primary explant human foreskinfibroblasts, which are the most common model for in vitro HCMV infection, and thehuman glioblastoma cell line, U373-MG. Using indirect immunofluorescence (IIF)techniques, the expression of HLA class I, CD46, CD55, and CD59 were confirmed onuninfected fibroblasts (Figure 16) and glioblastoma cells (Figure 17). CD55 expressionon glioblastoma cells was negligible by Ill techniques, but Western blot (Figure 18),flow cytometry (see section 3.2.2), and Northern blot (see section 3.4.1) techniquesconfirmed low but consistent constitutive expression of CD55 on this cell type. Thelevels of CD46 on both cell types (Fig 16E and 17E) seemed relatively weak comparedto CD59 expression; however, this probably reflects the smaller amounts of amonoclonal antibody (anti-CD46) bound to the investigated antigen (single epitoperecognition) as compared to a polyclonal antibody (anti-CD59; multiple epitoperecognition). Western blot techniques demonstrated that the Mr of CD55 found onfibroblasts, glioblastoma cells, and primary fetal astrocytes was the same as that seenfor CD55 expressed on monocytes (78 kDa), and not the lower Mr observed forerythrocyte CD55 (72 kDa) (Figure 18 A and B). However, the Mr for CD59 wasidentical (18-20 kDa) across all of the cell types investigated (Figure 18 C), as was theMr for CD46 (65 kDa; Figure 18 D).69Figure 16. Indirect Immunofluorescence analysis investigating the presence ofcell surface proteins on fibroblasts. Primary explant foreskin fibroblasts weregrown on coverslips as seen by light microscopy in A. These cells were incubatedwith mouse monoclonal antibodies directed against HLA class I (B) or CD46 (E)and visualized with a FITC-conjugated anti-mouse IgG antibody, or the cells wereincubated with rabbit polyclonal antibodies raised against CD55 (D) or CD59 (F)and visualized with a FITC-conjugated anti-rabbit IgG antibody. Backgroundstaining was determined by using non-immune rabbit serum and the FITCconjugated anti-rabbit IgG secondary antibody (C). Magnifications: A, B, and C=40X, D, E, and F= 63X.70Figure 17. Indirect Immunofluorescence analysis investigating the presence ofcell surface proteins on the glioblastoma, U373-MG, cell line. Glioblastoma cellswere grown on coverslips as seen by light microscopy in A. These cells wereincubated with mouse monoclonal antibodies directed against HLA class I (B) orCD46 (E) and visualized with a FITC-conjugated anti-mouse IgG antibody, or thecells were incubated with rabbit polyclonal antibodies raised against CD55 (D) orCD59 (F) and visualized with a FITC-conjugated anti-rabbit IgG antibody.Background staining was determined by using non-immune rabbit serum and theFITC-conjugated anti-rabbit IgG secondary antibody (C). Magnifications: A, B, andC= 63X, D, E, and F= 40X.71B.—— 0D.97.4-monocytic leukemia cell line, THP-1 (Mono).secondary antibody yielded no signal.Figure 18. Western blot analysison whole cell lysates toinvestigate relative molecularmass for CD55 (A and B), CD59(C) and CD46 (D) on a number ofdifferent cell lines. Proteins in A,B, and D were separated on 7.5%polyacrylamide gels, whileproteins in C were separated on a15% polyacrylamide gel. Relativeprotein mobilities were comparedbetween fibroblasts (HuFF),glioblastoma cells (U373),astrocytes (Ast), erythrocytes(RBC), platelets (Pit), and aPre-probing with conjugated3.2.2 Complement inhibitor expression alteration with HCMV infection.The cell surface expression of CD55, CD46, CD59, and HLA class I proteinsduring HCMV infection of human fibroblasts was measured using monoclonalantibodies and flow cytometry. Flow cytometry profiles for CD55 (Figure 19)demonstrated increases in cell fluorescence following infection with HCMV(MOI=1O). Figure 19 also shows the binding of the isotype control antibody, whichwas identical for uninfected and HCMV-infected fibroblasts. Moreover, increases incell fluorescence appeared to be uniform, and the mean fluorescence increased withA.C.97.4-70.8-97.4-70.8-43.6-28.3- —.i72the time of infection, which suggests that the majority of cells in the population wereinfected. The time-course and magnitude of these results also suggest that CD55expression was not enhanced as a result of the immediate release of internal pools ofCD55, as has been described for neutrophil responses to a variety of stimuli (Bergerand Medof 1987). Figure 20 displays the cumulative data for CD55, CD46, CD59, andHLA class I. CD55 expression on HCMV-infected fibroblasts increased two-fold(p<O.Ol) at 24 h post-infection (p.i.), while CD46 expression did not increasesignificantly until 48 h p.i. (p<O.Ol). By 72 h p.i., there was a 10-fold increase in CD55(p<O.OO1) and a 3.4-fold increase in CD46 (p<O.Ol) as compared to mock-infected cells.Interestingly, CD59 expression decreased in response to HCMV infection, whichsuggests that the enhanced cell surface complement inhibitor expression mediated byHCMV is not universal (Figure 20). By 24 h p.i., the CD59 expression had decreased30% (p<O.O5),a trend which continued until by 72 h p.i. the expression of CD59 onHCMV-infected fibroblasts was 50% that measured on mock-infected controls(p<O.Ol). I also observed down-regulation of HLA class I expression, which has beenreported previously (Gilbert et al 1993, Barnes and Grundy 1992, Yamashita et al1993), but the HLA class I down-regulation was not significant until 48 h p.i.(p<0.OOl). However, in this thesis I am using the down-regulation of HLA class I asan indirect measure of the uniformity of HCMV infection.73Figure 19. Composite ofCONTROl. UNINFECTED representative flow cytometricanalysis profiles using monoclonalj anti-CD55 antibody to detect surfaceexpression on uninfected fibroblasts(top) and on HCMV-infected (strainAD169) at 24,48, and 72 hours postCMV-INFECTED: 24 Hinfection. Binding of isotype/\ matched control antibody is alsoshown in the top histogram (left) andwas identical for HCMV-infected.. and uninfected cells.CUV-INFECTED: 48 HcUV4NFECTED: 72 H,/1LOG FLUOREScENcE74I. Figure 20. Flow cytometric analysis of the cell surface expression of CD55, CD46,CD59, and HLA class I on mock-infected (gray bars) and HCMV-infected (blackbars) human foreskin fibroblasts (HuFF) at 24 hour intervals. Predeterminedsaturating amounts of specific monoclonal antibodies were used for each antigen,and no change was observed when an isotype-matched control antibody was used.Error bars indicate one standard deviation, N=3.50 HLA140____302010024 48 72200 CD59 p<O.OO1 p<O.Ol150 - I10050I investigated the possible role of an Fc-receptor in the enhanced antibodybinding by using isotype-matched control antibodies as well as binding of thephycoerythrin-labeled secondary antibody only. Since, the Fc portion of the antibodyremoved as part of the phycoerythrin labeling process, it is uninfluenced by Fcreceptors and these results are summarized in Table 2. No difference in the nonspecific antibody binding was measured between uninfected cells and cells infectedwith HCMV (MOI=1O, 72 h p.i..) indicating the increased antibody binding and meancell fluorescence resulted from increased CD55 and CD46 antigen expression. Inpreliminary experiments, I blocked the Fc receptor on HCMV infected cells by pre75CD55 p<0.01_—ip<0.1I.p<0.1Ip<0.1400300100I:402024CD4648p<0.1I.p<0.1 -24 48072 24Time Post-Infection (Hours)48 72incubating cells with 5% HCMV-seronegative human serum, and since the secondaryantibody is pre-absorbed against human serum components, this did not contribute tothe signal. No difference was measured in the mean cellular fluorescçnce for cellswhich had the Fc receptor blocked and those which had no pretreatment, indicatingthe Fc receptor had minimal influence on the binding of the mouse monoclonalantibodies used.Table 2. Mean cellular fluorescence of non-specific antibodybinding to uninfected and HCMV-infected cells.1HCMV (72 h p.i.) No HCMV Statistics2Isotype controland secondary Abs3 0.260 ± 0.02 0.216 ± 0.02 NSSecondary antibodyonly 0.207± 0.02 0.163 ± 0.02 NSStatistics NS NS1. Mean cellular fluorescence is given as the mean +1- standard deviation, each condition performed intriplicate. Cells were infected with HCMV at an MOI=1O and harvested at 72 h p.i.2. Comparisons were performed amongst all groups by T-tests with the additional use of Bonferroni’scorrection for multiple comparisons. NS=not significant.3. The secondary antibody was a phycoerythrin-conjugated Fab fragment of a donkey anti-mouse Hand L chain antibody. The isotype matched control was a mouse monoclonal anti-HLA class IIantibody. Abs=antibodies76To investigate whether or not the HCMV-induced increase in CD55 and CD46was restricted to fibroblasts, I replicated the experiments using another permissivecell line, the U373-MG glioblastoma cell line. Figure 21 represents a comparisonamongst glioblastoma cells which were mock-infected, infected with the AD169 strainof HCMV, or infected with a low passage patient isolate of HCMV at a MOI=1O for 72h. It is readily apparent that the relative increases in CD55 and CD46 are identical tothose measured on HCMV-infected fibroblasts. Figure 21 also demonstrates theexpression of CD55 on glioblastoma cells is significantly above background. Similarto the findings for fibroblasts, a 19% and 23% decrease in cell-surface CD59 expressionwas observed for glioblastoma cells infected with AD169 strain and wild type,respectively (p>O.Ol). Moreover, the relative CD55 and CD46 increases are identicalfor the lab and wild type strain of FICMV, indicating that this phenomenon is not aproduct of conditioning due to long-term tissue culture propagation of the virus. Thestatistics listed in Fig. 21 are comparisons of the mean cellular fluorescence of mockinfected cells to HCMV-infected cells (either AD169 or wild type). No statisticaldifference was found between the mean cellular fluorescence for cells infected withAD169 or wild type HCMV, and these results were identical for two separate isolatesof wild type HCMV.77cJU0——1,)L Figure 21. Flow cytometric analysis of the cell surface expression of CD55, CD46,and HLA class I on mock-infected (gray bars) glioblastoma cells and cells infectedwith the AD169 (black bars) HCMV or a low passage patient isolate of HCMV(striped bars). Two separate patient isolates were investigated and gave identicalresults, only one of these is shown. The error bars indicate one standard deviationand the statistics given are comparisons between HCMV-infected and mock-infected cells; = this is the only sample in which the difference between HCMVstrains used approached significance (p=O.O7).200D Uninfected• AD 169-72HL Wild type-72Hp<o.05p<O.o150*p<0.1p<0.01I IHLA1 CD55 CD46 CD590Control3.2.3 Viral specificity of CD551CD46 increased expression.I considered the possibility that the CD55 and CD46 increased expression wasa non-specific phenomenon, and that any viral stress may replicate these findings. Toaddress this possibility I infected fibroblasts and glioblastoma cells with HSV-1 andadenovirus, both of which decrease the expression of HLA class I. The results of flowcytometry analysis using monoclonal antibodies (N=3 for each experimental group)are summarized in Table 3. Optimal viral infection of fibroblasts was gauged by thedecrease of HLA class I on 95% of the cells for each virus.78At2Ohp.i.:HLA12CD55CD46At 72 h p.i.:HLA 1CD55CD46At 72 h:HLA1CD55CD46Table 3. Flow cytometry analysis of fibroblasts underdifferent viral and chemical exposure.lMock HSV HCMV60.2 ± 5.2 13.9 ± 0.6 ***3 38.3 ± 3.1*26.3 ± 2.9 26.1 ± 2.8 37.9 ± 3•9*22.5 ± 1.2 19.5 ± 0.9 29.2 ± 1.5Mock65.5 ± 4.925.6 ± 2.523.7 ± 1.5Mock62.0 ± 3.525.9 ± 1.321.4 ± 0.5Adenovirus24.1 ± 2.0 **24.7 ± 0.922.1± 0.91 j.tg/ml LPS ‘94.5 ± 7.0 **25.1 ± 0.9ND51. Mean cellular fluorescence is given as the mean ± standard deviation, each condition performed intriplicate. Cells were infected with MCMV and HSV at an MOI=1O and Adenovirus at an MOI=50.Cells were harvested at times indicated.2. Predetermined saturating concentrations of monoclonal antibodies were used for analysis. HLA1=HLA class I.3. Statistics represent comparisons performed amongst all groups by ANOVA then Fishers leastsignificant difference tests with the additional use of Bonferroni’s correction for multiple comparisons.Differences found to be statistically different from the values for mock-infected cells are given as*p<Q•Q5, **p<0•01, and ***p<0•(J14. Fibroblasts were incubated for 72 h in DMEM with 5% FBS containing 1 or 15 .tg/ml bacteriallipopolysaccharide (LPS).5. ND=not done.HCMV21.4 ± 1.9 **150.1± 13.6***55.2 ± 1.6 **15 jig/mi LPS80.3 ± 6.9 *26.5± 2.122.2 ± 1.979Since HSV has a shorter lytic cycle these conditions were met by a MOI=10 at 20 h p.i.;for adenovirus these conditions were met by a MOI=50 at 72 h p.i. Table 3demonstrates that the increase in CD55 and CD46 expression measured on fibroblastswere unique for HCMV-infection since, under these conditions, no changes in CD55or CD46 were measured on HSV- or adenovirus-infected fibroblasts. Furthermore,incubating 1 to 15 pg/ml of bacterial lipopolysaccharide for 72 h with fibroblasts toinduce activation were also unsuccessful in altering the expression of CD55 andCD46, while expression of HLA class I was significantly increased.3.2.4 Increased CD55 is of host origin.The largest increase in expression on HCMV-infected cells was measured forCD55. This raised the concern that the antibody against CD55 was cross reacting witha viral epitope since all experiments used the same monoclonal antibody, 1H4. Twoseparate methods were employed to exclude this possibility. First, flow cytometryanalysis was repeated with the addition of two other CD55-specific monoclonalantibodies. As shown in Figure 22(A), each antibody is directed against a uniqueepitope on CD55. The results for HCMV-infected and mock-infected fibroblasts at 72h p.i., (Figure 22B), showed a similar relative increase in binding of anti-CD55antibody for HCMV-infected cells. Second, Western blot analysis using an anti-CD55polyclonal antiserum detected a 78 kDa protein in HCMV-infected fibroblasts (Figure23), which is consistent with the reported molecular weight for CD55. Moreover, the78 kDa band was more intense in HCMV-infected cells compared with uninfectedcontrols, and no additional bands were detected in the HCMV-infected cells.Furthermore, a search of Genebank did not reveal any homology between CD55 andthe HCMV strain AD169 genome. Taken together, these results are consistent withthe hypothesis that HCMV caused the enhanced expression of host cell CD55.Further confirmatory evidence is provided by Northern blot analysis (see section3.4.1).80..L Figure 22. Flow cytometric analysis of the cell surface expression of CD55 usinga panel of monoclonal antibodies against different epitopes as shown in (A).Relative increases in mean cellular fluorescence were seen on HCMV-infected(black bars) compared to mock-infected (gray bars) human foreskin fibroblasts(HuFF) with all anti-CD55 monoclonal antibodies. Error bars indicate one standarddeviation, and statistical analysis indicated all HCMV-infected cells had asignificant increase in CD55 expression as compared to matched uninfectedcontrols.A 1h1D7 701 BNH,—“1SCR1-____1H4 5O-SCR2- 40-SCR312G6 3O-SCR420-R-0 0-R4 10-S/T region-I III FMonoclonal Antibody81.—c)c.) 0200 kDa-116 kDa97kDa-69 kDa46 kDaFigure 23. Western blot analysis of wholecell lysates from HCMV-infected ormock-infected fibroblasts harvested at 72h p.i. was performed using rabbitpolyclonal anti-CD55 antibodies. Theintensity of the 78 kDa band detected forthe infected cells was greater than thatseen for the mock-infected cells and noadditional bands were observed. Preprobing of the blot with only thesecondary antibody yielded no signal.Molecular weights indicated on the figurewere determined using Amershamprotein standards.3.2.5 Functional properties of increased CD55 on HCMV infected cells.Lastly, I endeavored to show that an increase in CD55 translates into afunctional enhancement of complement regulation for the HCMV-infected cells. Totest this hypothesis I used a purified complement protein assay to assess the cellsurface activity of the alternative pathway C3 convertase, C3bBb (Devine and Rosse1987). Figure 24 shows that similar amounts of C3b were bound to the surface ofHCMV-infected or control cells from the fluid phase in the first stage of the assay.This process is known to be independent of the presence of CD55 or CD46. However,when C3bBb was formed at these sites in the second CD55-sensitive stage of theassay, the complexes on the surface of HCMV-infected cells were significantly lesseffective at activating C3 than C3bBb on uninfected controls (Fig 24; p<O.O5). Theseresults demonstrate that the increased levels of CD55 on the surface of HCMV82infected cells mediated an additional functional ability in regulating the C3-converting enzyme complexes..1- Figure 24. Effect of HCMV infection on C3 convertase function. HCMVinfected (black bars) or mock-infected (gray bars) fibroblasts were exposed to C3deposition from a fluid phase C3 convertase (A) followed by cell-surface C3convertase formation and subsequent C3 conversion (B) to assess CD55 function.Cell-bound C3 was assessed using a monoclonal anti-C3d antibody and flowcytometry. Error bars represent standard error of the mean.350—.B___Z300• Uninfected•HCMV250__________________hio 200150c-)100c, 50 ——0Fluid phase CoVF C3DepositionIC3 addition br Cell-surfaceC3 convertase833.3. CHAPTER 3. Effects of HCMV infection on THP-1 cells.The monocyte or macrophage has been speculated to be the HCMV reservoir invivo , and the THP-1 monocytic leukemia cell line is permissive to HCMV infection,but only after the cells are differentiated with phorbol esters (Weinshenker et al 1988).To confirm these reports, the amounts of virus present after HCMV infection of THP1 cells were assayed by plaque assay on fibroblasts at regular intervals (Table 4).Table 4. HCMV associated with THP-1 cells.Plague-forming units per ml in cell supernatant1Differentiated2 Virus3 Day 0 Day 2 Day 3 Day 6No AD169 3.7x106 4.3x105 1.8x105 ND4Yes AD169 9.0x106 NT) 8.5x103 1.9x103Yes Patient 2.0x106 1.0x105 2.7x103 NDYes & washed5 Patient 0 80 100 250No & washed5 Patient 0 0 0 01. Amount of HCMV in the supematant of infected cells was determined by removing a sample bysterile technique and performing a plaque assay using fibroblasts.2. HCMV infection of THP-1 cells was performed while the cells were undifferentiated (non-permissivestate), or after the THP-1 cells had been incubated with 300 ng/ml PMA for 72 h (permissive state).3. The HCMV stock used was either the laboratory strain, AD169, or a low passage patient isolate ofHCMV.4. ND=not done.5. To remove the high background of initial HCMV, the cells were extensively washed with PBSfollowing a 1 h infection.84If the inoculating dose of HCMV was not removed from the cultures a constantdecline in HCMV was seen, regardless of differentiation state, suggesting no (or anundetectable) production of HCMV above the inoculating dose. However, when theexperimental conditions were altered to include extensive washing of the cellsfollowing initial infection of cells, low amounts of HCMV were produced only by thedifferentiated THP-1 cells. Detectable levels of HCMV were first detected at 2 daysp.i. and the extracellular levels continued to increase to 6 days p.i., while no HCMVwas produced by undifferentiated THP-1 cells. The amounts of HCMV produced bydifferentiated TI-IP-1 cells was comparable to the amounts produced by othermonocytic cell or peripheral lymphocyte cultures previously reported (St. Jeor andWeisser 1977, Lathey and Spector 1991). The low amounts of virus produced fromdifferentiated THP-1 cells were not due to a minority of the cells being infectedbecause greater than 80% of the cells had evidence of infection as assessed by HLAclass I decrease. Although the measuring the decrease in HLA class I is not asaccepted as an infectious focus assay or in situ hybridization for viral products, it is aphenomenon very specific to fully permissive infection of certain viruses. Further, areview of recent publications indicates that monocyte HLA class I expressionincreases or remains unchanged when exposed to cytokines, interferons o, 3, o, and y,or bacterial cell products, indicating that this is a valid measure of HCMV infection(Chen et al 1994) However, higher viral titers may have been obtained if cell lysateshad been used rather than culture supernatant, as a high degree of cell association ofthe virus has been speculated for monocytes (Ibanez et al 1991). In this section, I wifireport the effects of HCMV infection of differentiated and undifferentiated THP-1cells on host complement inhibitor and HLA class I expression.3.3.1. HCMV infection of undifferentiated THP-1 cells.The undifferentiated THP-1 cells are non-adherent and most resemblemonocytes. Routinely, 2x106 cells were pelleted and resuspended in 150 pi of85concentrated virus stock (roughly 2x108pfu per ml) in a sealed tube, at 37°C for 1 h,with mixing every ten mm. The cells were then suspended in 10 mls of fresh DMEMwith 5% FBS and allowed to culture for a further 72H before being harvested. Theinfection was carried out this way to optimize infection and because the method usedfor infection of adherent cells was inappropriate. The virus stock was concentratedfor two reasons. First, it was the only way to incubate the cells with an adequate MOlto ensure all cells would have at least one infectious virion per cell enter it. Second,concentrating the virus resulted in removing soluble chemical mediators, such ascytokines which could have been produced by the fibroblasts during HCMVpropagation, and thus, reduced the confounding variables. Additionally, sincemonocytes are capable of engulfing proteins and this could trigger intracellular effectswhich alter the expression of surface proteins, a parallel incubation of a freshly UVinactivated (254 nm) portion of the virus stock was analyzed as well. The UVinactivated virus was always tested for the presence of plaque forming units bystandard plaque assay, and none were ever found.However, in the previous studies of adherent cell infection with HCMV, theuniform decrease in HLA class I was used as a gauge of HCMV infection. Even whennon-differentiated THP-1 cells were incubated with HCMV at an MOI=100, the HLAclass I expression did not decrease; in fact, as shown in Figure 25, the expression ofHLA class I actually significantly increased. The increase in HLA class I was notentirely due to the effects of live HCMV, since the effects of the antigen load, asmeasured with UV-inactivated (254 nm) HCMV, also caused a smaller, significantincrease fri HLA class I expression. However, the increase in ITLA class I expressionmeasured when THP-1 cells were incubated with live HCMV was uniformly muchhigher than that for cells incubated with UV-inactivated HCMV in all experimentsperformed. Interestingly, preliminary studies in which undifferentiated THP-1 cellswere infected with HSV-1 (MOI=5 for 20 h) resulted in a significant decrease in HLAclass I86I- Figure 25. Flow cytometry analysis of cell surface expression of CD55, CD46,CD59, and HLA class I (HLA 1) on THP-1, monocytic leukemia, cells followingincubation with live HCMV (MOI=50) or UV-inactivated HCMV for 72 h p.i.Vertical statistics indicate comparisons to uninfected cells, while statistics abovethe umbrella bracket indicate comparisons between cells incubated with live andUV-inactivated HCMV. HLA class II (HLA 2) is not expressed on this cell line andserves as a second isotype-matched antibody control.70________________Uninfected n<o.oi60 UV-HCMV___• HCMV-72H50°V40Control HLA 2 HLA 1 CD59expression. This indicates the HLA class I increased expression is specific for HCMVunder these circumstances.The increase in CD55 and CD46 expression, while observed was not any wherenear the magnitude of increased expression observed for HCMV infection of adherentcells (Figure 25). The incubation of undifferentiated THP-1 cells with UV-inactivatedHCMV did not significantly increase the expression of CD55, but the expression ofCD55 increased 62% (p<O.O1) at 72 h p.i. These results were the same when either themonoclonal antibody 1H4 or 11D7, directed against CD55, were used for flowp<0.1CD5587cytometry analysis. Incubation of undifferentiated THP-1 cells with UV-inactivatedHCMV did increase the expression of CD46 (p <0.05); however, the increase in CD46expression following incubation with live HCMV was always greater (p<O.Ol), andlike the effects of HCMV infection on adherent cells, the increase in CD46 expressionwas always a lower magnitude than that observed for CD55. These results wereconsistent when either the monoclonal antibody M75 or J4-48, directed against CD46,were used for flow cytometry analysis. The lower magnitude of CD55 and CD46expression increase, and the opposite effect on the expression of HLA class I, mayreflect the limited expression of HCMV genes reported in undifferentiated THP-1 cells(Weinshenker et al 1988). The increased expression of CD46, CD55, and HLA class Iwas observed regardless of whether the HCMV viral stock used for infection was theAD169 strain or the patient isolates of HCMV used in the previous section.No significant change in CD59 was observed under any of the conditions,regardless of whether the monoclonal antibody BR1C229 or MIRL1, directed againstCD59, were used (Figure 25). The expression of HLA class II was previously reportedto be negative for TFIP-1 cells (Tomoda et al 1992, Vey et al 1992), even followingdifferentiation with phorbol esters; thus monoclonal antibodies directed against [-ILAclass II were used as an isotype-matched negative control. Following HCMVinfection, HLA class II expression did not change (Table 5). Binding of a secondisotype-matched control antibody raised against Aspergillus niger glucose oxidase, aprotein which is neither expressed nor inducible in mammalian cells, was also foundto be unaltered by HCMV infection (Table 5). This suggests that HCMV infectiondoes not induce HLA class II expression on undifferentiated THP-1 cells, and that theincrease in expression of CD55, CD46, and HLA class I is not due to increases in nonspecific binding.88Table 5. Mean cellular fluorescence of non-specific antibodybinding to uninfected and HCMV-infected cells.1HCMV (72 h p.i.) No HCMV Statistics2HLA class jj3 2.23 ± 0.15 2.63 ± 0.36 NSControl3 1.23 ± 0.10 1.25 ± 0.05 NS1. Mean cellular fluorescence is given as the mean ± standard deviation, each condition performed intriplicate. Undifferentiated THP-1 cells were infected with HCMV at an MOI=50 and harvested at 72 hp.i.2. Comparisons were performed between HCMV-infected and uninfected cells tested with the sameantibody by T-tests with the additional use of Bonferroni’s correction for multiple comparisons.NS=not significant.3. Monoclonal antibodies directed against HLA class II (isotype IgGi) or against a yeast protein(control; isotype IgG2a), which is neither expressed nor inducible in eukaryotic cells, were used asisotype-matched controls to investigate non-specific antibody binding by the cells.The THP-1 cells also express low levels of another member of the RCA genecluster, complement receptor 1 (CR1 or CD35; Thieblemont et al 1993), as domonocytes and macrophages (Aheam and Fearon 1989, Thieblemont et al 1993). Theeffects of HCMV infection on CR1 expression were more subtle than the effect onCD55 and CD46 expression. The basal expression of CR1 is very low, but as shown inFigure 26 the percent of cells (above background) expressing CR1 increased withHCMV infection; while the expression of the control antibody did not change. Nosignificant increase in CR1 expression was seen with undifferentiated TI-IF-i cellsincubated with UV-inactivated HCMV, although it routinely approached significance.However, the patient isolate of HCMV appeared to have a greater ability to increase89CR1 expression than the AD169 laboratory strain of HCMV (Figure 26). These valuesare given as changes in percent of the population after subtraction of the background.Percent positive cells were used owing to the variability in the population andposibility that all cells may not be infected with HCMV..1. Figure 26. Flow cytometry analysis of CR1 expression on undifferentiated THP-1cells after 72 h incubation with live HCMV or UV-inactivated HCMV. The onlystatistical significance identified was between HCMV infected and uninfectedcells. Results from two representative experiments are shown using a laboratorystrain, AD 169 (A), or a patient isolate of HCMV (B). Control isotype-matchedantibodies directed against HLA class II or a yeast protein (see Materials andMethods) gave identical results for background.<0.0530 30 I25 UiecteTI 25UV-AD169 I20. 2015. 15io. toojp<0.5El Uninfected‘J UV-Wild typeU Wild type-72HControl CR1 Control CR1903.3.2. HCMV infection of phorbol ester-differentiated THP-1 cells.The morphology of the TFIP-1 cells following differentiation into macrophagelike cells is quite different from undifferentiated THP-1 cells. In addition to increasedphagocytic abilities and other functional abilities reported by others (Tsuchiya et al1982, Mehta and Lopez-Berestein 1986), the expression of integrins and otheradhesion molecules increased on THP-1 cells concomitant with their conversion to apredominantly adherent cell morphology (Prieto et al 1994). Once the THP-1 cells areadherent, they are much easier to infect with HCMV and the infection procedure isthe same as for fibroblasts and glioblastoma cells. The THP-1 cells are fullypermissive for HCMV infection following differentiation; extracellular HCMV isproduced and can be detected by plaque assay (see Table 4 above). Unfortunately,the basal expression levels of CD46, CD55, CR1, and HLA class I also increased(Figure 27) after differentiation, which made interpreting the effects of HCMVinfection post-differentiation a little difficult.HCMV infection of differentiated THP-1 cells resulted in the significantdecrease of HLA class I (Figure 28), which is opposite from the effects of HCMVinfection of undifferentiated THP-1 cells. This indicated the induction of a permissivestate for HCMV infection. However, incubation of differentiated THP-1 cells withUV-inactivated HCMV resulted in a small increase in the expression of HLA class I,similar to that seen when UV-inactivated HCMV was incubated with undifferentiatedTHP-1 cells. A 33% increase in CD55 expression (p<O.O5) was measured on HCMVinfected TFIP-1 cells after differentiation, while incubation of cells with UVinactivated HCMV did not alter CD55 expression. HCMV infection did not effectCD46 expression on differentiated THP-1 cells, but CD59 expression was increased by30% on HCMV-infected THP-1 cells (Figure 28). These last two results are markedlydifferent than those observed for the infection of fibroblasts and glioblastoma cells,and may represent a difference in the HCMV lytic cycle in monocytic cells.91HCMV infection of the differentiated THP-1 cells also decreased the expressionof CR1, which is contrary to the effect of infecting undifferentiated THP-1 cells withHCMV. Since the relative expression of CR1 on differentiated cells is still lowcompared to the other cell surface proteins, the percent of cells (above background)expressing CR1 is the best measure for alterations. Figure 29 shows a 3-fold decreasein the percent of cells positive for CR1 (p<O.O5). No explanation for the contrastingeffects of HCMV on differentiated versus undifferentiated cells is readily apparent..1. Figure 27. Flow cytometry analysis of alterations in expression of CD46, CD55,CD59, HLA class I, and CR1 following differentiation with 300 nglml for 72 h.Isotype-matched control antibody is directed against HLA class II, anddifferentiation did not induce significant binding with this antibody.p<o.oO1300.I0p<0.0l-IJ No PMA•PMAp<0.01I__IControl HLA1 CD5992I. Figure 28. Flow cytometry analysis of cell surface expression of CD55, CD46,CD59, and HLA class I (HLA 1) on differentiated THP-1, monocytic leukemia, cellsfollowing incubation with live HCMV (MOI=50) or UV-inactivated HCMV for 72 hp.i. The THP-1 cells were differentiated by exposure to 300 ng/ml phorbol estersfor 72 h prior to infection with the laboratory strain, AD169, of HCMV. Statisticswhich are listed vertically indicate comparisons to uninfected cells, while statisticsabove the umbrella bracket indicate comparisons between cells incubated with liveand UV-inactivated HCMV.100? 90C70____50.ç) 30C10.0.EJ UninfectedUV-AD169• AD169-72Hp<0.1LrV-,-Control HLA 1 CD46 CD5593L Figure 29. Flow cytometry analysis of CR1 expression on differentiated THP-1cells after 72 h incubation with live HCMV or UV-inactivated HCMV. The THP-1cells were differentiated by exposure to 300 nglml phorbol esters for 72 h prior toinfection with the laboratory strain, AD169, of HCMV. The only statisticalsignificance identified was between HCMV infected and uninfected cells. Controlisotype-matched antibodies were directed against HLA class II.U Uninfected2 UV-AD169• AD169-72HIControlIC,).—.—C,)0C0)0)15-10-5-0-p<0.5CR1943.4. CHAPTER 4. Mechanism of CD55 expression increase.There are many mechanisms available to the cell with the common result ofincreased surface expression of CD55. This chapter will provide evidence suggestingthat permissive HCMV infection of cells causes a increase in the transcription of theCD55 gene leading to an accumulation of CD55 mRNA species.3.4.1 Northern blot analysis of HCMV-infected cells.One possibility to account for the observed increase in CD55 expression is thatHCMV infection upregulates the synthesis of CD55 mRNA. To investigate thispossibility, total RNA was isolated from human fibroblasts and glioblastoma cellswhich were infected with the AD169 strain of HCMV for 72 h and uninfected controls.Thirty micrograms of total RNA from each was separated on a 1% formaldehyde—denaturing agarose gel and transferred to a nylon membrane. The amount of CD55mRNA was determined by probing the nylon membrane with a radiolabeled anti-sense CD55 probe, the comparison for infected and uninfected fibroblasts andglioblastoma cells are shown in figure 30 A and B, respectively. Ten micrograms oftotal RNA from HeLa cells was included as the positive control for CD55 mENAspecies (Thomas and Lublin 1992). Figure 30 C demonstrates the amounts of CD59mRNA present in HCMV-infected and uninfected fibroblasts in the same samplesused in figure 30 A and B, using a radiolabeled anti-sense CD59 probe. The levels ofCD55 mRNA were greatly increased in HCMV infected cells, either fibroblasts orglioblastoma cells, at 72 h p.i., which parallels the flow cytometry and Western blotanalysis findings. However, the amount of CD59 appeared to be diminished inHCMV infected fibroblasts compared to the uninfected cells.Further investigation of mRNA levels was performed by purifying total RNAfrom glioblastoma cells which were infected with HCMV, collecting samples frominfected and uninfected cells at 24 h intervals corresponding to the time pointsinvestigated by flow cytometry (section 3.2.2). Thirty micrograms of each sample was95I Figure 30. Northern blot analysis of HCMV-infected or mock-infected cellsharvested at 72 h p.i. was performed using a radiolabeled probe made from CD55cDNA on fibroblasts (A), or glioblastoma cells (B) or using a probe made fromCD59 cDNA on glioblastoma cells (C). Thirty micrograms of total RNA was loadedfor HCMV-infected cells and matched uninfected cells. Ten micrograms of totalRNA from HeLa cells were included as a control for CD55 mRNA. Sizes of mRNAare listed next to each band and match those previously reported for CD55 andCD59 (Thomas and Lublin 1993, Holguin et al 1993).A. B. C.C C C c2.5kb- f:2.2kb118kbL J 19kb1.2kb-separated under the same conditions described above and the nylon membranes wereprobed with radiolabeled anti-sense CD55, CD59, or CD46 probes (Figure 31). Theamount of CD55 mRNA was increased by 24 h p.1. and continued to accumulate to the72 h p.i., which parallels the flow cytometry findings (Figure 20). No alterations inthe splicing pattern or ratio of minor and major transcripts were observed for CD55which could account for the difference in the Mr of CD55 found on purified virions,suggesting the difference in the size of virion CD55 (Figure 15, p 67) may be due toalterations in post-transcriptional modifications.96The amount of CD46 mRNA did not demonstrate a consistent increase over thecourse of HCMV infection (Figure 31 B). Although the there appears to be a largeincrease in the CD46 mRNA at 24 H p.i., the differences between mock-infected CD46mRNA levels through the 24 to 72 H time points makes it difficult to comment on thesignificance of the differences between mock-infected and HCMV-infected cells. Thismethod may not be sensitive enough to detect alterations in CD46 mRNA, since cellsurface expression only changes by 2-3 fold, as opposed to the 8-10 fold increase ofCD55. Alternately, increased transcription of the CD46 gene may not be themechanism responsible for the increased cell surface expression. The amount ofCD59 mRNA appears to be decreased by 24 h p.i. and this decrease continuesthroughout all time points collected relative to samples collected for mock-infectedACD552.5kb-1.8kb-B:CD46 43.3 kb4__.1- Figure 31. Northern blot analysisof HCMV-irtfected or mock-iniectedglioblastoma cells harvested at 24 hintervals after infection. Thirtymicrograms of total RNA fromHCMV-infected or mock-infectedcells for each time point wereseparated on a denaturingformaldehyde-agarose gel andtransferred to a nylon membrane.Matching membranes were probed• with radiolabeled anti-sense probesmade from CD55 cDNA (A), CD46cDNA (B), or CD59 cDNA (C).CCD59 :2.2 kb- g’1.2 kb97cells (Figure 31 C); however, the magnitude of the decreaseis inconsistent with theflow cytometry findings which show a very small decrease in cell surface expressionover 72 h p.i. for HCMV-infected glioblastoma cells.3.4.2. Further investigation of decreased CD59 expression on HCMV-infected cells.Northern blot analysis consistently reported a decrease in CD59 mRNA levelswhen cell were infected with HCMV. However, flow cytometry studies reportedeither a decrease or little change depending on the antibody used as described below.In order to understand the variability in these results, CD59 expression was furtherinvestigated using radiolabeled immunoprecipitation analysis. Equal numbers ofHCMV-infected and uninfected fibroblasts were labeled with[35S]-cysteine for 50minutes at 24 h p.i. and harvested immediately or following a 2 or 4 h chase withexcess cold cysteine (Figure 32). Non-specific bands were identified by performing aninitial mock-immunoprecipitation by using all the reagents normally used, butomitting the addition of antibody. The polyclonal anti-CD59 antibody was thenadded and the samples were then subjected to a second round ofimmunoprecipitation. The mock immunoprecipitation was labeled as the“prec1eared lane, and run adjacent to the other lanes. Figure 32 (A) shows that theamount of CD59 labeled by the initial radioactive pulse was slightly decreased at 24 hp.i.; however, processing of the precursor of 16.5 kDa to the mature 17-19 kDa form,seen after a 2 h chase with the uninfected fibroblasts was negligible followinginfection with HCMV. Furthermore, a second band of 34 kDa was present only in theHCMV-infected sample at the initial harvest and not in any of the other samples.Figure 32 (B), which differs only by the longer post-labeling chase, confirms thesefindings; again at 24 h p.i. the initial pulse of the infect cells indicates the amount ofCD59 precursor was slightly decreased and no mature form can be seen. The proteinlabeling in (B) was not as strong as in (A), and the 34 kDa band in the HCMV-infected98L Figure 32. Autoradiograph from proteins labeled with[35S]-cysteine,immunoprecipitated using a polyclonal anti-CD59 antibody and protein Gsepharose, and separated on a 15% polyacrylamide gel under reducing conditions.[1-4C]-labeled bovine serum albumin was run in addition to unlabeled standards inboth A and B. The left half of both gels represents samples from fibroblastsinfected with HCMV for 24 h, while the right half represents samples harvestedfrom uninfected cells. The lanes labeled as pre-clear pulse represent non-specificassociation of proteins with protein G-sepharose and indicate background bandingpatterns, since no immunoprecipitating antibody was added. The cells were‘pulsed’ with 100 tCi[35S1-cysteine for 50 mm. at 37°C and lanes labeled pulse aresamples collected immediately following labeling. Some samples were collectedafter the radiolabeled cysteine was removed and the cells incubated (“chased”) with‘cold’ cysteine for 2 h (A) or 4 h (B). The large arrows indicate the CD59 precursorat around 16 kDa, while the arrowhead indicates the presence of a 34 kDa bandfound only with the HCMV-infected cells in (A).HCMV -24 Hp.i. Uninfected•1I) -- ?I u c,B. HCMV -24 Hp.i.C’•C Lfl Lflc)• cc,28.3-17.9--15.1-UninfectedC CUjI99is too faint to be distinguished; however, this band has consistently appeared in allpulse-chase experiments run.The possible explanations for the different rates of CD59 decrease measured bydifferent method include: 1) The steady-state half-life of cell surface CD59 isextremely long and the reduced production of mature CD59 does not represent asignificant proportion of total, and therefore, was not observed by methods such asflow cytometry. 2) The relative dilution of the CD59 mRNA by the HCMV encodedmRNA species exaggerates the actual decrease in CD59 mRNA, and could account forthe inability to observed increased CD46 mRNA levels in HCMV-infected cells. 3)The monoclonal anti-CD59 antibody used in the flow cytometry studies has a lowaffinity and did not accurately assess the cell surface levels of CD59. The actualreason is probably a combination of all three, but the only possibility easily assessed isthe last. To address this possibility the levels of CD59 expression were investigatedon HCMV-infected and uninfected glioblastoma cells at 72 h p.i., using two anti-CD59antibodies which recognize separate epitopes (Figure 33). The basal levels of CD59-associated fluorescence on the surface of uninfected cells appear to be 8-fold higherwhen the monoclonal antibody BR1C229 is used. Furthermore, a significant decreasein the expression of CD59 was only detected on the HCMV infected cells, ascompared to the uninfected controls, when BR1C229 was used. Therefore, thediscrepancies between the decreases in CD59 expression (as measured by flowcytometric, Northern blot, and radio-immunoprecipitation methods) must be due to along catabolic half-life for CD59 with some dilutional effects caused by theaccumulation of viral mRNA.100L Figure 33. Flow cytometry analysis of cell surface expression of CD59 onglioblastoma cells using two different monoclonal antibodies directed againstCD59. Gray bars represent uninfected cells and black bars indicated cells infectedfor 72 h with HCMV. Even though the monoclonal antibodies are of the samemouse IgG class and both were used at predetermined saturating conditions, theBR1C229 monoclonal antibody is much better at detecting surface expression ofCD59. The only statistical significance found for CD59 decreased expression waswith the BR1C229 monoclonal antibody.200_______________UninfectedI HCMV-72HI •1•Control MIRL13.4.3. Requirements for CD55 promoter activity.There are a number of mechanisms which could conceivably result in theaccumulation of CD55 mRNA reported in section 3.4.1. To investigate the possiblerole of HCMV-infection on upregulation of CD55 transcription, I obtained the CD55promoter inserted into a plasmid with a chloramphenicol-acetyl-transferase (CAT)reporter gene and several CD55 promoter 5’-deletion constructs (Figure 34). The basicrequirements for the basal expression have been previously published (Thomas andLublin 1992) for the HeLa cell line and a few other cell lines. However, I chose to usea different method of detecting CAT than the method used by Thomas and Lublin,and a few additional constructs were added including: -275, -425, -592, and -2800 toBRIC 229101+84. Figure 35 shows the relative CAT expression in HeLa cells under the influence ofthe different 5’-deletion constructs of the CD55 promoter. These findings are verysimilar to those reported by Thomas and Lublin: no CAT activity was seen until 77 bpof the promoter upstream of the transcriptional start site were present. However,nearly all of the full promoter activity was present with the construct containing -206bp, and no statistically significant addition of activity was observed when up to -2800bp were present. Figure 35 does suggest a weak pattern in which the activity of the-206, -275, and -425 constructs are slightly less than the -592, -796, and -2800 promoterconstructs, but it does not stand up to statistical analysis.1- Figure 34. The CD55 promoter was isolated from a bacteriophage genomic clonescontaining human CD55. A fragment containing 2800 bp upstream from thetranscriptional start site to 84 bp downstream from the transcriptional start site(+84) was inserted into the plasmid SP65 containing the chloramphenicol-acetyltransferase (CAT) reporter gene. Deletion mutants were made by removing 5segments of DNA with either restriction enzymes (-796 to -206 constructs) or bydesigning appropriate PCR primers (-77 to -36 constructs) (Thomas and Lublin1993).-2800 bpCD55Promoter-77, -54, -36Chioramphenicolacetyltransferase102.1’ Figure 35. Basal CD55 promoter activity was measured 48 h after transfectioninto HeLa cells. Intracellular CAT levels were measured using[14C]-Acetyl CoAfor each deletion construct measured. Almost all basal promoter activity waspresent when only 206 bp upstream from the start site were present. Mock activity(reporter plasmid only) CAT activity arbitrarily set=1, N=6 per point, bars=SEM.3.4.4. Effect of HCMV infection on CD55 promoter constructs.Unfortunately, the HeLa cell line is not permissive for HCMV infection;therefore, the CD55 promoter construct containing -796 to +84 bp was transfected intoglioblastoma cells and fibroblasts. Twenty-four hours after plasmid transfection ofboth cell types, half of the cultures were super-infected with the AD169 strain ofHCMV and the other half were kept as uninfected controls. Forty-eight hours post-viral infection (i.e. 72 h post-transfection) HCMV-infected and uninfected cells wereharvested and the intracellular CAT activity assessed (Figure 36). The levels ofrelative CAT activity of the -796 bp CD55 promoter construct were increased 10-12fold in both cell types following HCMV infection compared to the levels in uninfected-2800Relative CAT activity2 3103cells. This indicates that the transcription of the CD55 promoter is upregulated byHCMV infection and the accumulation of CD55 mRNA is, at least in part, caused bythis increased transcription.To address the basic requirements of the HCMV upregulation of transcription,compared to the basal transcription requirements, the 5’-deletion constructs of theCD55 promoter were used. The various constructs were transfected into theglioblastoma cell line because it survived the transfection procedure more readilythan the fibroblasts. The super-infection of transfected cells was carried out asdetailed above, and 48 h post-viral infection, the intra-cellular CAT activity wasassessed. Figure 37 demonstrates that, unlike the basal transcription requirements,the upregulation of the CD55 promoter activity was first seen in the -275 to +85promoter construct. No difference was seen amongst the CD55 promoter constructslarger than -275, up to and including the -2800 to +84 construct. Interestingly, nodifference was seen in the basal levels of activity between the -206 and -275 constructin the HeLa cell line (Figure 35), but all of the HCMV upregulation activity seems tobe between these constructs.104SI- Figure 36. Intracellular CAT activity from fibroblasts or glioblastoma cellswhich were transfected with the -796 to +84 CD55 promoter construct thenharvested 72 h later (gray bars). The black bars represent cells treated exactly thesame as those shown in gray except that they were infected with HCMV 24 h aftertransfection. A large increase in CD55 promoter activity was seen for both celltypes when super-infected with HCMV. CAT activity is arbitrarily set atuninfected cell activity=1, N=6.• Uninfected• HCMV-48 H15c-)105-0•IFibroblast Glioblastoma105.1- Figure 37. Intracellular CAT activity was measured for glioblastoma cells whichwere first transfected with the promoter constructs listed above, then super-infected with HCMV for 48 h prior to harvesting. Enhanced promoter activityrequired the presence of 275 bp upstream from the start site. This is different fromthe requirements of only 206 bp for the basal transcription. Mock activity (reporterplasmid only) CAT activity arbitrarily set=1, N=6 per point, bars=SEM.Relative CAT activityCAT-2800 +84106The basal activity of the CD55 promoter constructs was too low in thefibroblasts and glioblastoma cells to be assessed by the CAT measurement protocol Ihad selected. The reason it works for the HeLa cell line is readily apparent from thedifferent intensities of the mRNA present between 10 pg of total HeLa RNA and 30 tgof total uninfected glioblastoma cell RNA seen by Northern blot analysis in Figure 30.Even though the HeLa cells are not permissive for HCMV infection, the possibilitythat the presence of HCMV virions was altering CD55 promoter activity withoutactually infecting the cells was addressed by incubating HeLa cells, transfected withthe -796 CD55 promoter construct, with an amount of HCMV equivalent to aMOI=100 as determined by infection of fibroblasts (Figure 38).No morphological changes indicative of cytopathic effect were observed byobservation under the inverted microscope, while such changes were quite apparentin permissive cells infected with HCMV at the same time point. Furthermore, noalteration in the CD55 promoter activity was observed in HeLa cells incubated withHCMV. This indicates that the upregulation of the CD55 promoter activity requirespermissive HCMV infection. The HCMV responsive element appears to be located inthe region between -275 and -206 or may overlap the -206 restriction site, but analysisof this region did not identify any known transcription factor binding sites or knownenhancer sequences.107Figure 38. Intracellular CAT activity from HeLa cells which were transfectedwith the -796 to +84 or -54 to +84 CD55 promoter construct. At 24 h posttransfection half of the transfected cells were incubated for 1 h with a highconcentration of HCMV (black bars), then the cells were harvested 48 h later. TheHeLa cells are not permissive for HCMV infection and incubation with a highamount of HCMV did not increase the CDS5 promoter activity when compared tothe transfected cells not incubated with HCMV (gray bars), indicating the effects ofHCMV on the CD55 promoter in fibroblasts and glioblastoma cells (Figure 36) wasnot an artifact. CAT activity is arbitrarily set at background activity=1, N=6, errorbars=SEM.0 1HCMV+—+5CD5 5Dromoter—54—54—796—796Relative CAT activity2 3 4HNS6NS1083.4.5. Effect of HCMV late gene repressor on CD55 upregulation.A 2-fold increase in CD55 expression is measured by 24 h p.i. on HCMVinfected fibroblasts, and 24 h p.i. is also the point at which the HCMV genome beginsto replicate and the late HCMV genes are first expressed in infected fibroblasts(DeMarchi et al 1980). Therefore, it is probable that the HCMV immediate early ordelayed early genes are responsible for the upregulation of CD55. It has been shownthat infecting fibroblasts with HCMV in the presence of phosphonoacetic acid (PAA)completely inhibits the expression of the HCMV late gene expression and genomereplication (Stinski 1977). Duplicate cultures infected with HCMV for 84 h p.i. in thepresence of PAA did not produce infectious extracellular virus as assessed by plaqueassay; while the culture infected with HCMV in the absence of PAA routinelyproduced between iO-iO pfu/ml of HCMV. Figure 39 demonstrates the effect ofPAA on HCMV super-infection of glioblastoma cells transfected with the -796 CD55promoter construct, as described in the previous section. The addition of PAA didnot effect the CD55 promoter activity when the cells were super-infected with HCMV,since the promoter activity increased between 8-9 fold, compared to uninfected cells(p<O.OO1), regardless of whether PAA was present or not.The effect of FAA was also tested using flow cytometric methods. Theinvestigation utilized four groups of glioblastoma cells: one was mock-infected, onewas mock-infected in the presence of 0.1 mg/ml PAA, one was infected with AD169with no PAA, and the last group was infected with AD169 in the presence of 0.1mg/ml FAA. Forty-eight hours p.i. all groups (each in triplicate) were harvested andcell surface expression of CD55 and HLA class I were assessed using monoclonalantibodies and flow cytometry (Figure 40). The decrease in HLA class I, used toassess the completeness of HCMV infection, was found to be significantly decreasedby HCMV infection in the presence or absence of PAA (p<O.OO1). However, PAAalone was also observed to decrease the expression of HLA class I. Although thedecrease was less than that observed for HCMV infected cells, it was observed to be109separate from the effects of HCMV (p<O.O5). The expression of CD55 was found to besignificantly increased on HCMV-infected cells in the presence and absence of FAA(p<O.O5 and p<O.OOi, respectively). However, the increase in CD55 expression on cellsinfected with HCMV in the presence of PAA was found to be significantly less than inthe absence of PAA (p<O.Ol).Figure 39. Intracellular CAT activity from glioblastoma cells which weretransfected with the -796 to +84 CD55 promoter construct and super-infected withHCMV, in the presence or absence of 0.1 mg/ml phosphonoacetic acid (PAA), thenharvested 48 h later. The large increase in CD55 promoter activity still occurred inthe presence of PAA, suggesting the involvement of HCMV early or immediateearly genes. CAT activity is arbitrarily set at uninfected cell activity=1, N=6, errorbars=SEM.Relative CAT activity0 2 4 6 8PAA HCMV— —— ++ +10 12110.1- Figure 40. Flow cytometric analysis of HCMV-infected or uninfectedgliobiastoma cells was performed using predetermined saturating amounts ofmonocional anti-CD55 and HLA class I antibodies. The effect of 0.1 mg/mi PAA (48hour exposure) on the cell surface protein expression for HCMV-infected(harvested at 48 hours post-infection) and uninfected cells was determined (seelegend in figure for designations). Error bars indicate one standard deviation.Statistical significance between adjacent samples is indicated by an umbrellabracket, while statistical comparison between PAA or no PAA matched pairs islisted at a 45° angle. No difference in the isotype matched monoclonal antibodywas observed.70.__________________60____50 —40.30. Z1= II5.2014.C10ControlUninfectedE’ Uninfected PAA• HCMV 48 Hp.i.HCMV PAAHLA1 CD55111There are two possible explanations for these results: 1) The late genes of HCMVplay a partial role in the increased expression of CD55, or 2) the presence of PAAadversely affects the proper synthesis of some host proteins, as demonstrated forHLA class I. The basal expression CD55 on the glioblastoma cells did not seem to beaffected by PAA, but the basal expression of CD55 on glioblastoma cells is alsoextremely low. Additionally, the CD55 promoter data shown in Figi.ire 39demonstrated no effect of PAA on HCMV upregulation of CD55, and tend to supportthe latter possibility.3.4.6. No Upregulation of CD55 promoter by isolated HCMV early genes.All of the results reported in this thesis suggest that the HCMV immediateearly (IE) or early (E) genes play a role in the upregulation of CD55 expression. Toaddress this possibility, I obtained all of the isolated HCMV immediate early genes(that are known to date) inserted in eukaryotic expression vectors from a number ofsources (see materials and methods). These IE genes were transiently co-transfectedwith the -796 CD55 promoter and the CAT activity assessed (Table 6), and the IEgenes were also transfected into cells and their effects on the endogenous CD55expression were assessed by flow cytometry (Table 7). As shown in Tables 6 and 7,none of these genes had any effect on CD55 promoter activity or CD55 cell surfaceexpression, either alone or in combination. To investigate the possibility that theseresults were negative due to low transfection efficiency, a plasmid containing the LacZ driven by an HCMV IE promoter was added to all transfections, and thetransfection efficiency was assessed on duplicate cultures by staining for Lac Zactivity. In the presence of X-gal, a substrate for 3-galactosidase (Lac Z) successfullytransfected cells turn blue. All of the cultures of the experiments reported in Table 6and 7 had transfection efficiencies from 8-12 % as assessed by microscopy. At theselevels, the flow cytometric analysis should have been able to identify a subpopulation upregulated within the untransfected cells, and the transient co112transfection assay should be unaffected since super-infection of cells transfected bythe same procedure yielded good results (see section 3.4.4). Therefore, these resultssuggest that these HCMV IE genes are not directly responsible for the upregulation ofCD55 or CD46 on glioblastoma cells.113TABLE 6. Transient transfection assessment for co-transfection of the CD55promoter/reporter and isolated HCMV IE genes.’IE genes2 mg/mi3 Promoter construct4 Relative CAT activityNone 0 -796 to +84 1.00 ± 0.06TEl, 1E2 5 -796 to +84 1.06 ± 0.06TEl, 1E2, US3,UL36-38 5 -796 to +84 1.20 ± 0.20None 0 -275 to +84 1.00 ± 0.10IE1,1E2,US3,UL36-38 5 -36 to +84 0.95 ± 0.13TEl, 1E2, US3,UL36-38 5 -275 to +84 1.00 ± 0.13AD1695 -796 to +84 30.1 ± 5.11. Promoter activity was assessed using transient co-transfection of the isolated HCMV immediateearly genes and CD55 promoter construct listed for each value. N=3 for each value reported as mean ±standard deviation, relative CAT activity in cells transfected with the CD55 promoter only werearbitrarily set as equal to 1.00. Isolated HCMV immediate early genes in expression vectors weretransfected into glioblastoma cells using the CaC1 method listed in materials and methods.2. The IE genes were transfected in the combinations listed, and IE genes transfected alone yieldedidentical results.3. This column lists the amount of each IE construct used in that transfection.4. This column lists the CD55 promoter construct utilized for the IE genes listed.5. A positive control using the longest CD55 promoter and super-infected with the AD169 strain ofHCMV was included to ensure the CD55 promoter used for co-transfection with the IE genes wasresponsive to HCMV infection.114IE genes2NoneIE11E2mg/mi3055% cells in gate 341.53 ± 0.570.96 ± 0.250.63 ± 0.21MCF of the top 5% of cells434.79 ± 0.5633.79 ± 0.0836.76 ± 4.031. Isolated HCMV immediate early genes in expression vectors were transfected into glioblastomacells using the CaCl method listed in materials and methods. The effect of transfection on theendogenously expressed CD55 was assessed by flow cytometry using monoclonal antibodies directedagainst CD55. N=3 for each value reported as mean ± standard deviation.2. The IE genes were transfected in the combinations listed, and IE genes transfected alone yieldedidentical results.3. This column lists the amount of each IE construct used in that transfection.4. Since the transfection efficiency for this method is roughly 10%, the transfected cells with increasedCD55 on the cell surface would be observed as a sub-population. Therefore, gate 3 was arbitrarily setat 1.5% for the cells which were not transfected with IE genes, so that any sub-population present withincreased CD55 would be identifiable. Gate 2 was set at 5% for the cells which were not transfectedwith IE genes and the mean cellular fluorescence (MCF) is reported so that the increased CD55expression would be minimally diluted by the fluorescence of the larger, untransfected population.5. A positive control using glioblastoma cells infected with the AD169 strain of HCMV was included toensure CD55 expression was responsive to HCMV infection in these experiments.TABLE 7. Flow cytometry analysis of CD55 expression following transienttransfection with isolated HCMV IE genes.1None 0 1.63 ± 0.12 32.96 ± 1.17IE1,1E2,US3,UL36-38 1 1.57 ± 0.76 32.45 ± 1.42rEl, 1E2, US3,UL36-38 2 1.87 ± 0.26 35.68 ± 2.72TEl, TE2, US3,TJL36-38 5 1.60 ± 0.14 34.03 ± 0.78AD1695 75.4 ±2.101154. DISCUSSIONOther investigators (Lewis et al 1986, Eizuru et al 1988, Rundell and Betts 1982)have reported that complement enhances the neutralizing ability of HCMVseropositive serum. My studies confirm this finding, and elaborate the extent of thecomplement binding to virions.The accessory role for complement in assisting viral neutralization has beenpreviously described for other viruses as well. The neutralization of equine arteritisvirus requires both complement and specific antibody (Radwan and Burger 1973).Complement was also found to be essential for the neutralization for equine herpesvirus-i in vivo , but only during the first 2 days after infection; following this periodthe enhancing effect of complement decreased (Snyder et al 1981). Interestingly, adifference in the mechanism of complement enhanced neutralization was foundbetween homotypic and heterotypic complement and antibody sources for avianinfectious bronchitis virus (AIBV; Berry and Almeida 1968). Incubation of fowl seracontaining indigenous antibodies with AIBV virions resulted in an increased proteinhalo around the virus particles when viewed by electron microscopy while incubationof AII3V virions with serum from rabbits previously inoculated with the virus resultedin the formation of 100 angstrom holes in the virion envelope. The lack of evidencefor the formation of the MAC, as assessed by electron microscopy, between rabbit andfowl complement suggests that the AIBV virions contain an inhibitor for fowlcomplement which interferes with the complete activation of complement. Thepresence of a species specific inhibitor would not be surprising, since an avian viruswould have evolved against the selective pressure of an avian immune system. Thesefindings are comparable to the results presented in this thesis, except that the sourceof the complement inhibitors on HCMV virions appears to be host cell from which thevirus was derived (an option which is also possible for the AIBV). The presence of thehost complement inhibitors on the HCMV virions may also explain the lack ofnoticeable neutralizing ability of the seronegative sera, even though I found that116incubation of seronegative serum with HCMV virions resulted in the deposition of C3on the virions. Reports of viral proteins which directly activate the classicalcomplement pathway exist (Bartholomew et al 1978), but the C3 deposited on theHCMV virions probably arose from minimal activation of the alternative pathwaysince similar amounts of C3 were deposited on HCMV virions incubated with HCMVseronegative serum and serum in the presence of EGTA, a classical complementpathway inhibitor.I found the presence of antibodies greatly increased the deposition of C3 on theHCMV virions and was required to achieve a level of complement activation whichresulted in significant amounts of C9 being associated with the HCMV virions.Studies reported for herpes simplex type-i (HSV-1), using C4-, C5-, C6-deficientguinea pig serum, suggested that the presence of a functional terminal lytic pathwaywas not necessary for neutralization enhancement by complement (Daniels et al 1970).Furthermore, Daniels et al (1970) found that the addition of purified C4 to optimumamounts of activated Clq on HSV virions enhanced virus neutralization, but that thefurther addition of purified C2 and C3 only further enhanced the virus neutralizationif the amount of C4 present was sub-optimal. These minimal requirements ofcomplement for HSV neutralization are probably not reflective of the requirements forHCMV neutralization; C3 was present on HCMV virions incubated with HCMVseronegative serum, but with a negligible effect on HCMV neutralization. However,these conclusions are tenuous since the presence of complement-activating antibodieson the virion would undoubtedly result in much greater quantities of the earlycomplement components being deposited on the virion.Complement assisted neutralization of some type C oncoviruses results invirolysis, as assessed by release of radiolabeled RNA and tegument enzymes (Cooperet al 1976, Spear et al 1990); indicating complement activation by these viruses goes tocompletion. Complement activation by HCMV virions appears to go to completiononly in the presence of specific anti-HCMV antibodies as assessed by the presence of117C9 with the HCMV virions. However, whether true virolysis occurs is questionablesince the virions maintain enough integrity to survive through 3x 30% sucrosepurification steps post-incubation (see materials and methods).The absolute requirement of complement for efficient neutralization wasreported for the initial infection of horses with equine herpes virus-i (Snyder et al1981). However, the enhancing abilities of complement for antibody neutralizationdecreased with time, and the same effect has been reported for anti-HSV antibodies(Yoshino et al 1977). The neutralizing enhancement of complement for equine herpesvirus-i and HSV could be a combination of the switch from 1gM to IgG production,since 1gM is far more efficient at activating complement, or the rapid affinitymaturation of the antibodies being produced, which would result in the increase inthe intrinsic neutralizing abilities of the antibodies and subsequently reduce theapparency of complement’s effect.The pattern of antibody response to HCMV infection has been delineated in theliterature. Patients with primary HCMV infections were found to seroconvert withinthe first 2-10 weeks and the mean antibody response titer reached a plateau around 10weeks after the first symptoms were noticed, but the antibody response in patients reinfected or with a reactivated infection attained higher levels much earlier in infection(Pass et al 1983). Spencer and Andersen (1972) found that the neutralizing antibodytiters rose later than antibodies detectable by indirect fluorescent assay or complementfixing assay. The enhancing effect of complement in antibody-mediatedneutralization in my studies was investigated using serum samples from volunteerswho had been seropositive for at least one year and, therefore, represent convalescentserum samples. It should be noted that my experiments were not carried out withpooled serum, but were performed using individual characterized sera in eachexperiment. To my knowledge no reports exist which indicate that the neutralizingabilities of early antibodies generated against HCMV infection are enhanced to agreater or lesser degree than antibodies generated at later stages of infection, and118samples from patients recently infected with HCMV were not readily available to me.However, one murine monoclonal antibody directed against the HCMV-envelopeglycoprotein p130!55 required complement to neutralize HCMV virions while asecond antibody directed against another envelope glycoprotein, p86, did not; eventhough they were both the same IgG subclass (IgG2a). This suggests that themechanism of complement enhancement of antibody neutralization is much morecomplex than simple subclass differences between early and late antibodies. Britt et al(1988) found that neutralizing monoclonal antibodies generated after inoculation of anon-glycosylated recombinant from of HCMV p130/55 did not require complementfor neutralization while antibodies generated from glycosylated p130/55 did;suggesting the specificity of the antibody is more important to complementrequirements.Finally, host encoded CD55, CD46, and CD59 were found on purified HCMVvirions. The greater presence of CD55 and CD59 on the virions, as compared to CD46,probably represents the difference in transfer of glycolipid-anchored proteins tobudding virions compared to transmembrane proteins. However, the apparentmolecular mass for the host-encoded complement inhibitors associated with virionswas more variable than that found on uninfected fibroblasts or fibroblasts infectedwith HCMV for 72 h (Fig 15C). A similar finding has been found when comparing themolecular weights of virion glycoproteins purified from simian CMV and HCMVvirions (Benko and Gibson 1986). It was indicated by these authors that the disperseappearance of the electrophoretic proffle of HCMV virion glycoproteins makes itdifficult to assign accurate molecular weights. Some of the variability was accountedfor by heterogeneity of the glycosylation; the proteins appear to be heavilyglycosylated and incompletely processed since some virion glycoproteins appear tohave both high marinose and complex type carbohydrates on the same protein (Benkoand Gibson 1986). It is feasible, therefore, that the glycosylation of host proteins maybe affected with progressive HCMV infection. In fact, the slight alteration in host119protein processing may account for the unique ability of sera from patients acutelyinfected with HCMV to induce the lysis of HCMV-infected fibroblasts (Betts andSchmidt 1981, Middeldorp et al 1986). The early immune response to HCMV mayinclude antibodies which are directed against modified epitopes on CD55, CD46, andCD59. This could decrease the ability of HCMV-infected cells, even with theupregulated expression of CD46 and CD55, to regulate complement, and would alsoresult in the addition of more cell-bound complement activating antibodies.However, since these represent autoantibodies the affinity maturation process wouldlikely eliminate the production of these antibodies and may explain the absence ofcytolytic abilities in HCMV-seropositive serum from convalescent patients.Complement represents one of the first lines of defense against foreignantigens, and host-encoded complement inhibitors associated with the virion mayprovide a reprieve from complement-mediated clearance. A second report of virionscontaining CD55, CD46, and CD59 was published after this body of work wascompleted: these host complement inhibitors being associated with HIV and SWvirions in that report (Montefiori et al 1994). Similar to Dr. Cooper’s findings forHCMV (1993), Montefiori et al found that monoclonal antibodies directed againstCD46 and CD59 blocked the complement regulating ability of the virions and resultedin a complement-dependent reduction in HIV and SW infectivity. It is possible thatmany of the enveloped viruses passively carry along host complement regulators andthat this represents a viral evolution strategy for evading the host immune system.The presence of complement regulators on HW and SW virions is interestingsince both viruses also activate complement, even in the absence of anti-HIV antibody(Solder et al 1989). However, the C3 fragments left on the virion surface are capableof mediating immune adherence to complement receptor-bearing cells, andcomplement activation has been shown to enhance the binding of 1-IIV to CR2 by 10-fold (Montefiori et al 1992). Thus complement can alter the distribution of host cellsfor the virus. Similarly, the EBV glycoprotein gp350 was found to enhance the120cleavage of bound C3 to C3dg which binds to the EBV receptor, CR2, which is also thereceptor for gp35O (Mold et al 1988). It is possible, therefore, that the regulatedpresence of C3 fragments on the surface of HCMV virions represent a mechanismwhich directs HCMV to bind to cells expressing C3 receptors, such as themonocyte/macrophage which may be the HCMV reservoir in vivo.Concurrent with my studies of purified HCMV virions, I investigated the effectof HCMV-infection on host protein expression. The upregulation of CD55 or decay-accelerating factor on the cell surface induced by non-viral effects has been seen.Shibata et al (1991) reported that incubating mesangial cells with human complementcomponents which were activated by unrelated immune complexes resulted in a 2-fold increase in the expression of CD55. This is relevant to HCMV-infected cells, sincein vivo anti-HCMV antibodies would activate complement and may increase theCD55 expression further than the direct effects which I have presented in thisdissertation. The increased CD55 expression that I measured for phorbol esterstimulated THP-1 cells was previously reported for phorbol ester stimulatedendothelial cells (Bryant et al 1990), and was found to be associated with proteinkinase C activation. The fact that HCMV infection of THP-1 cells which were predifferentiated by phorbol esters resulted in a further increase suggests that theHCMV-induced increase maximized this stimulation pathway , or more possibly, actsvia a second unrelated pathway. Interestingly, incubation of endothelial cells withtumor necrosis factor, IL-i, interferon gamma (IFNy) and some lectins did not alterCD55 expression, but incubation of endothelial cells with the lectin wheat germagglutinin increased CD55 expression 5-fold and incubation with lectins ConA andP1-IA increased CD55 expression by two-fold (Bryant et al 1991). Not surprisingly,there appears to be some difference in response amongst cell types; incubation withIL-i and IFNy had a negligible effect on CD55 and CD46 expression on endothelialcells (Moutabarrik et al 1993, Bryant et al 1991), but were found to increase both CD55and CD46 expression on cultured thyroid cells (Tandon et al 1994). Exposure of121endothelial cells to histamine also resulted in a two-fold increase in CD55 expressionas measured by flow cytometry, without affecting the expression of CD46 or CD59,but the amount of CD55 shed into the supematant also greatly increased (Tsuji et al1994), indicating that flow cytometry measurements may underestimate themagnitude of CD55 increase in my studies.Increased expression of the other complement inhibitors have also beenreported. As mentioned above, CD46 expression was increased on cultured thyroidcells exposed to IL-i and IFNy (Tandon et al 1994), and interestingly, elevated CD46levels are commonly associated with leukemic tumor cell lines, except B-cell lines, andother malignancies (Seya et al 1994, Seya et al 1990, Hara et al 1992, Cho et al 1991).While I found that the expression of CD55 and CD46, which are closely related andbelong to the RCA gene cluster, were both increased with HCMV infection ofadherent cells, alterations in CD46 and CD55 appear to be mostly independent of oneanother in most other reports (Tsuji et al 1994, Moutabarrik et al 1993). IncreasedCD59 expression had been reported for endothelial cells incubated with IL-i 13(Moutabarrik et al 1993), phorbol esters (Men et al 1993, Holguin et al 1993), and theexpression of CD59 has been reported to be elevated on colonic adenocarcinoma cells(Bjorge et al 1994). However, I only found an increased CD59 expression on the THP1 cells when infected with HCMV or treated with phorbol esters; CD59 expressionwas decreased on adherent cells infected with HCMV. The only other reports ofdecreased CD59 expression include: endothelial cells incubated with IL-i 13(Moutabarrik et al 1993), phorbol ester stimulated HL6O, promyelocytic leukemia cellline (Sedlak et al 1993), and CD8(+) lymphocytes infected with the retrovirus, humanimmunodeficiency virus-i (HIV-1; Weiss et al 1992).I only observed an increase in CR1 expression on undifferentiated THP-1 cells,since the fibroblasts and glioblastoma cells did not express this complement inhibitor.Other investigators have reported similar increases in CR1 on monocytes isolatedfrom patients with rheumatoid arthritis (McCarthy et al 1992). Tn my experiments, the122small increase in CR1 probably represents activation of the undifferentiated THP-1cells because these cells are only permissive for HCMV infection after differentiation(Weinshenker et al 1988), CR1 pools have been reported to be in the secretory vesiclesof neutrophils (Sengelov et al 1994), and other reports show CR1 expression increasedon monocytes under conditions of activation (Leino and Liius 1992). Conversely, thedecreased expression of CR1 on differentiated THP-1 cells following HCMV infectionis much more interesting. CR1 expression has been reported to be decreased onerythrocytes in patients infected with FIIV (Pascual et al 1994), on erythrocytes frompatients connective tissue diseases, on erythrocytes from patients with a high amountof circulating immune complexes (Corvetta et al 1991, Tausk and Gigli 1990), and onleukocytes isolated from patients with chronic myelogenous leukemia (CML; Lanza etal 1991, Lanza and Castoldi 1992). However, the relevance of decreased erythrocyteCR1 are minimal, since FIIV cannot directly infect erythrocytes and immunecomplexes do not exist in my in vitro system; however, neutrophils from AIDSpatients were reported to have a decreased ability to upregulate surface expression ofCR1 in response to activation (Tausk and Gigli 1990). Another possible explanationmay relate to the trypsin sensitivity of CR1 (Pascual et al 1994); thus, patients infectedwith HIV could have more circulating proteases, which could also be released frommy differentiated TI-IF-i cells when infected with HCMV. Alternately, the expressionof CR1 may indicate an alteration in cell differentiation, such as with CML, whichmay occur with HCMV infection.The HCMV genes involved in upregulation of certain host proteins have beenstudied by other investigators. I was interested to know whether these genes couldalso be responsible for the changes that I observed in host-encoded complementproteins. The HCMV immediate early genes 1 and 2 (IE1 and 1E2) were found totransactivate the hamster dihydrofolate reductase promoter through the transcriptionfactor E2F (Wade et al 1992), using transient co-transfection assays. The rat braincreatine kinase promoter, which was linked to a chloramphenicol acetyl-transferase123(CAT) reporter gene, was found to respond to transient co-transfection of the isolatedHCMV IE2 gene, with minimal increase with the addition of IE1 gene and asuppression of enhancement with the addition of the HCMV immediate early genesUS3 or UL36-38 (Colberg-Poley et al 1992). These results are in contrast to the effectsof the same genes on the human heat-shock protein-70 (hsp-70) promoter in the HeLacell line, which is non-permissive for HCMV infection (in decreasing order ofenhancement): UL36-38/US3 > UL36-38/IE1 > IE1 > TEl /US3 (Colberg-Poley et al1992). However, some differences existed between hsp-7O promoter responsiveness inHeLa cells and in fibroblasts (which are HCMV-permissive); in fibroblasts only the TEcombinations of TEl /1E2 and IE1/UL36-38 were effective at increasing hsp-70promoter activity (Colberg-Poley et al 1992). However, I used these same constructscontaining the isolated HCMV TE genes in fibroblasts and no effect was observedusing the CD55 promoter. These results may indicate that increased CD55 and CD46expression on HCMV-infected adherent cells is not mediated through directtransactivation via HCMV TE genes, rather other possibilities must be consideredincluding alterations in cytokines, signal transduction or autocrine pathways.More relevant to the results presented in this dissertation, isolated IE geneshave been used to transactivate cellular promoters in THP-1 cells. The effects ofHCMV IE1 and 1E2 on IL-113 mRNA levels and IL-i 3 promoter activities wereinvestigated using Northern blotting techniques and transient co-transfection ofpromoter-CAT constructs (Iwamoto et al 1990). Maximal effect of TEl and 1E2 genesrequired additional stimulation of THP-1 cells with LPS, and peaked at 3h post-addition of LPS, decreasing rapidly there after. The IE1 /1E2 genes were also found toenhance the IL-i (3 promoter activity in LPS-stimulated THP-1 cells, in a dosedependent fashion, but the HCMV 1E2 gene alone had no effect even though itsynergistically increased the transactivation activity of TEl on the IL-1(3 promoter(Iwamoto et al 1990). Unfortunately, subsequent investigations found that eventhough IL-i (3 mRNA levels were increased, the levels of released IL-i 13 protein were124found to be unaltered by transfection of LPS-stimulated THP-1 cells with TEl and 1E2(Kline et al 1994). The transfection of these cells with IE1 and 1E2 did, however, resultin increased mRNA and secreted protein levels of ]L-1 receptor agonist (Kline et al1994). The increased protein IL-i receptor agonist protein secretion required thetransfection of both ]E1 and 1E2, and was actually suppressed when IE1 was usedalone. Transient co-transfection assays investigating the effect of TEl and 1E2 on theHLA class I (A2 allele) promoter in the Jurkat (immortalized T-lymphocyte) cell linefound a 4-fold increase in promoter activity when the cells were unstimulated and an11.5-fold increase in stimulated Jurkat cells (Burns et al 1993). The effect of TE geneson the HLA class I promoter may explain the increased lILA class I expression Imeasured on the unstimulated THP-1 cells. However, Weinshenker et al (1988)reported that infection of THP-1 cells prior to differentiation with phorbol esters didnot lead to expression of HCMV IE genes, which disputes this hypothesis.The increases in CD55 and CD46 on unstimulated, HCMV-infected THP-1 cellswas a fraction of the increases observed for HCMV-infected adherent cell lines.However, the differences between the effects in HCMV infection on HLA class Iexpression on differentiated and undifferentiated THP-1 cells may help to explainthis. After the differentiation of THP-1 cells with phorbol esters, HCMV infection ofTHP-i cells results in the significant decrease in HLA class I expression observed onHCMV-infected adherent cells, indicating a release of the block of the HCMV geneexpression which mediates this effect. It is reasonable to assume that differentiationof the THP-1 cells also results in the release of the block of the HCMV-mediatedincrease in CD55 and CD46 expression; however, the phorbol esters used todifferentiate the THP-1 cells vastly increased the expression of the complementinhibitors, confirming the findings of other investigators (Bryant et al 1990, Holguin etal 1993). Therefore, the increase in CD55 and CD46 may mask the increases inducedby fully-permissive HCMV infection. The fact that an additional increase in CD55 ismeasured on differentiated, HCMV-infected TI-IF-i cells suggests that the125upregulation of CD55 is not maximized by phorbol ester stimulation, or that HCMVand phorbol esters act through separate stimulation pathways. However, there is noexplanation apparent for the lack of CD59 decrease on differentiated TI-IF-i cellsinfected with HCMV. Perhaps, like the differences reported by Colberg-Poley et al(1992) for hsp-70 promoter responsiveness between HeLa cells and fibroblasts, thegene(s) responsible for CD59 down-regulation are differentially expressed amongstcell types.I also investigated the promoter requirements for the HCMV-induced increasein CD55 promoter activity. A majority of the basal CD55 promoter activity was foundin the construct containing 206 bp upstream from the transcriptional start site, but allof the enhanced CD55 promoter activity in HCMV-infected cells required 275 bpupstream from the transcriptional start site. This is in contrast to the requirements ofthe HLA class I (A2) promoter identified by Bums et al (1993): the basal and HCMVinduced increase in HLA-A2 promoter both required only the minimal 116 bpupstream from the transcriptional start site. However the HLA class I promotercontains 2 CCAAT boxes and a presumed TATA box in this region (Burns et al 1993),while the CD55 promoter lacks both of these elements (Thomas and Lublin 1993).Furthermore, the HLA-A2 promoter responded to transfection with the TEl and 1E2genes, while the CD55 promoter did not; suggesting different HCMV genes areresponsible. The hamster DHFR promoter was reported to require the sequenceCCCGACTGCAATTTCGCGCCAAACTTGG to respond to the HCMV TEl and 1E2genes (Wade et al 1992), which is also the sequence required for this gene to respondto adenovirus and required the transcription factor E2F. However, the region of theCD55 promoter in question (-275 to -206) does not contain any sequences withhomology to the DHFR sequence, CD55 expression is not affected by adenovirusinfection, and the CD55 promoter also does not respond to co-transfection with theTEl and 1E2 genes. The HCMV DNA polymerase gene is upregulated by the HCMVIE genes and Kerry et al (1994) reported the sequence of an inverted repeat found in126the DNA polymerase promoter required for a majority of the transactivating activityof HCMV infection. However, the CD55 promoter does not contain this sequenceeither. Given the responsiveness of CD55 expression to IL-i, IFNy, and histamine insome cell types (Tandon et al 1994, Tsuji et al 1994), the alterations in cytokineproduction induced by HCMV infection (Almeida et al 1994, Iwamoto et al 1990), andthe lack of responsiveness of the CD55 promoter and expression to transfection withisolated HCMV IE genes suggests that the effect of HCMV infection on CD55 andCD46 expression may be indirect and possibly linked to alterations in autocrineproduction. The lack of increase in CD55 and CD46 on fibroblasts incubated withbacterial lipopolysaccharide (LPS) ,adenovirus, or HSV in Table 3 address some of thepossibilities. LPS was found to induce granulocyte/macrophage colony stimulatingfactor tumor growth factor 1 3, Jnterleukin-loL (IL-ia), IL-i, JL-6, and IL-8, but nottumor necrosis factor-a (Xing et al 1993, Huleihel et al 1990, Schwachula et al 1994).Therefore, since LPS treatment of fibroblasts did not increase CD55 or CD46expression, one may assume that the CD55 and CD46 expression increase is notrelated to these cytokines. Further, since adenovirus and HSV infection results in theexpression of interferon a and 13 (Daly and Reich 1993, Neilsch et al 1992), but did notresult in increased expression of CD55 or CD46, one may also assume that the CD55and CD46 expression increase is not related to interferon.Beyond the increased complement regulating activity induced by HCMVinfection, there may be other, further reaching implications to HCMV infection. CD59and CD55 have both been associated with tyrosine kinase pathways (Shenoy-Scaria etal 1992, Morgan et al 1993, Stefanova and Horesji 1991), indirectly to signaltransduction (Shibuya et al 1992) and CD55 and CD59 are found to be closelyassociated on the cell surface, indicating intentional organization (Stefanova andHoresji 1991). Alterations in CD55 to CD59 ratios as seen in HCMV infection ofadherent cells may have major implications for the signal transduction pathwaysassociated with them. Signal transduction through CD55 was found to be associated127with altering cytokine production and glucose consumption on monocytes (Shibuya etal 1992) and lymphocyte proliferation (Shenoy-Scaria et al 1992); while signaltransduction through CD59 was found to be associated with calcium transients andcell activation in T lymphocytes. The implications of increased CD46 decreased CR1with various carcinomas (Lanza et al 1991, Lanza and Castoldi 1992, Seya et al 1994,Seya et al 1990, Hara et al 1992, Cho et al 1991) also implies a potential role for HCMVin immortalization. Finally, increases in CD55 expression have been reported todecrease the function of natural killer cells (Finberg et al 1992), which potentiallycould act as a further immunological evasion strategy for the one cytotoxic cell whoseeffectiveness would not be reduced by the decreased HLA class I expression onHCMV-infected cells.1285. SUMMARY.The overall objective of this project was to describe the interaction betweencomplement and human cytomegalovirus (HCMV) virions and to determine the effectof HCMV infection on host complement inhibitor expression. There were twoobjectives in this thesis. The first objective was to assess the ability of complementto neutralize purified HCMV virions. Complement alone, in the absence of specificanti-HCMV antibody, was found to have a negligible neutralizing effect, even thoughsmall amounts of C3 were found to be deposited on the HCMV virion. Addition ofspecific anti-HCMV antibody resulted in the activation of both the alternative andclassical pathways, deposition of large amounts of C3 on the HCMV virion, andformation of the membrane attack complex as evidenced by the deposition of C9.Although complement activation was clearly seen in seronegative specimens mixedwith HCMV virions, the apparent arrest of the cascade at the C3 step suggested thepresence of complement inhibitors on the HCMV virion. Further investigationidentified the presence of host complement inhibitors, CD46, CD55, and CD59 onvirions. These findings potentially explain the complement regulatory activity ofHCMV virions in the absence of virus-encoded complement inhibitors in the HCMVgenome. This is a novel observation as no reports exist in the literature of anyidentifiable complement regulators present on the HCMV virion, irrespective oforigin.The second objective was to determine whether HCMV infection altered theexpression of host-encoded complement inhibitors. HCMV infection resulted in anincrease of the C3 regulating proteins, CD55 and CD46, on adherent cells, but adecrease in the terminal complement pathway inhibitor, CD59. These effects weredetermined not to be cell type specific as the effect was seen on both glioblastoma cellsand fibroblasts. Furthermore, virus-induced increases in CD46 and CD55 werespecific to HCMV infection; they were not seen when cells were infected with eitherHSV-1 or human adenovirus. Preliminary studies found an enhanced complement129regulating function associated with the enhanced CD55 expression on HCMV-infectedcells. This finding provides the first evidence for a new mechanism whereby virus-infected cells enhance resistance to complement-mediated cytolysis.To further characterize the mechanism by which HCMV infection enhancesexpression of CD55 and decreases CD59 expression, comparisons were made in theRNA and protein synthesis pathways between HCMV-infected and uninfected cells.While the results presented here do not absolutely identify the contributions of allpossible pathways, infection of permissive adherent cells clearly resulted in theaccumulation of CD55 mRNA. Furthermore, evidence was provided that this mRNAincrease was accompanied by a large increase in CD55 promoter activity whichappeared to be mediated through a novel HCMV-responsive element locatedupstream from the elements required for basal transcription. Conversely, thedecreased CD59 expression appeared to result from a combination of decreased CD59mRNA levels and dysfunctional protein processing which could result in enhancedcatabolism. These studies have identified CD59 as only the second host protein todecrease with HCMV infection.Interestingly, addition of isolated, identified immediate early HCMV genes,singly or in combination, were not able to reproduce the increased CD55 promoteractivity or CD55 expression, suggesting that the mechanism is much more complexthan that reported for other upregulated host proteins. Further investigation of CD55promoter activity and CD55 protein expression using larger gene segments in cosmidexpression vectors as well as measurement alterations in cytokine production inHCMV infected cells may clarify the mechanism by which HCMV mediates theincreased CD55 expression. Finally, since the specific deficiency of CD55 (Inabphenotype) is reported not to be associated with clinically significant episodes ofhemolysis in vivo , future studies for HCMV infection therapy might considerincluding a specific CD55 inactivator to enhance the effectiveness of the normalimmune response in HCMV clearance.130REFERENCES.Adams EM, Brown MC, Nunge M, Krych M, and AtkinsonJP. 1991. Contribution ofthe repeating domains of membrane cofactor protein (CD46) of the complementsystem to ligand binding and cofactor activity. J Immunol 147:3005-11.Ahearn JM and Fearon DT. 1989. Structure and function of the complementreceptors, CR1 (CD35) and CR2 (CD21). Adv Immunol 46:183-219.Almeida GD, Porada CD, St. Jeor S, and Ascensao JL. 1994. Human cytomegalovirusalters interleukin-6 production by endothelial cells. Blood. 83:370-6.Alsenz J, Lambris JD, Schulz TF, and Dietrich MP. 1984. Localization of thecomplement-component-C3b-binding site and the cofactor activity for factor i in the38 kDa tryptic fragment of factor H. Biochem J 224:389-98.Alsenz J, Schultz TP, Lambris JD, Sim RB, and Dietrich MP. 1985. Structural andfunctional analysis of the complement component factor H with the use of differentenzymes and monoclonal antibodies to factor H. Biochem J 232:841-50.Amadei C, Tardy-Panit M, Couillin P. Loulon M, Cabau N, Boue A, and Michelson S.1983. Kinetic study of the development and localization of human cytomegalovirusinduced antigens using monoclonal antibodies. Ann Virol (Inst Pasteur) 134E:165.Avery VM and Gordon DL. 1993. Characterization of factor H binding to humanpolymorphonuclear leukocytes. J Immunol 151:5545-53.131Baker PJ, Lint TF, Mortensen RF, and Gewurz H. C567-initiated cytolysis of lymphoidcells: description of the phenomenon and studies on its control by C567 inhibitors. JImmunol 118:198-202.Barnes PD and Grundy JE. 1992. Down-regulation of class I HLA heterodimer and2-microg1obulin on the surface of cells infected with cytomegalovirus. J Gen Virol73:2395.Bartholomew RL, Esser AF, and Muller-Eberhard HI. 1978. Lysis of oncornavirusesby human serum: isolation of the viral complement (Cl) receptor and identification aspl5E. J Exp Med 147:844-853.Beersma MFC, Bijimakers MJE, and Ploegh HL. 1993. Human cytomegalovirusdown-regulates HLA class I expression by reducing the stability of class I H chains. JImmunol 151:4455.Benko DM, and Gibson W. 1986. Primate cytomegalovirus glycoproteins: lectinbinding properties and sensitivities to glycosidases. J Virol 59:703-13.Benson JD and Huang E-S. 1990. Human cytomegalovirus induces expression ofcellular topoisomerase II. J Virol. 64:9.Berger M and Medof ME. 1987. Increased expression of complement decay-accelerating factor during activation of human neutrophils. J Clin Invest 79:214-20.Berry DM and Ahnieda JD. 1968. The morphological and biological effects of variousantisera on avian infectious bronchitis virus. J Gen Virol 3:97-102.132Betts RF and Schmidt SG. 1981. Cytolytic 1gM antibody to cytomegalovirus inprimary cytomegalovirus infection in humans. J Infect Dis 143:821-6.Bjorge L, Vedeler Ca, Ulvstad E, and Matre R. 1994. Expression and function of CD59on colonic adenocarcinoma cells. Eur J Immunol 24:1597-1603.Boldogh I, Abubakar 5, and Albrecht T. 1990. Activation of proto-oncogenes: animmediate early event in human cytomegalovirus infection. Science. 247:561.Boldogh I, Abubakar 5, Deng CZ, and Albrecht T. 1991. Transcriptional activation ofcellular oncogenes fos, jun, and myc by human cytomegalovirus. J Virol. 65:1568.Bora NS, Lublin DM, Kumar By, Hockett RD, Holers VM, and Atkinson JP. 1989.Structural gene for human membrane cofactor protein (MCP) of complement maps towithin 100 Kb of the 3’ end of the C3b/C4b receptor gene. J Exp Med 169:597.Borysiewicz LK, Morris S. Page J, and Sissons JGP. 1983. HCMV-specific cytotoxic Tlymphocytes: Requirements for in vitro generation and specificity. Eur J Immunol13:804.Borysiewicz LK, and Sissons JGP. 1986. Immunobiology of CMV infection. ImmunolToday 7:57.Bowden RA, Sayers M, Floumey N, Newton F, Banaji M, Thomas ED, and Meyers JD.1986. Cytomegalovirus immune globulin and seronegative blood products to preventprimary cytomegalovirus infection after marrow transplantation. N Engi J Med314:1006.133Bowden RA. 1991. Cytomegalovirus infection in transplant patients: methods ofprevention of primary cytomegalovirus. Transplant Proceedings 23 (suppi 3):136-138.Britt WJ. 1984. Neutralizing antibodies detect a disulfide-linked glycoproteincomplex within the envelope of human cytomegalovirus. Virol 135:369-78.Britt WJ, Vugler L, and Stephens EB. 1988. Induction of complement-dependent and-independent neutralizing antibodies by recombinant-derived humancytomegalovirus gp55-116 (gB). J Virol 62:3309-18.Browne H, Smith G, Beck 5, and Minson T. 1990. A complex between the MHC classI homologue encoded by human cytomegalovirus and 132-microglobulin. Nature347:770.Browne H, Churcher M and Minson T. 1992. Construction and characterization of aHCMV mutant with the UL18 (Class I homologue) gene deleted. J Virol 66:6784-7.Bryant RW, Granzow CA, Siegel MI, Egan RW, and Billah MM. 1990. Phorbol estersincrease synthesis of decay-accelerating factor, a phosphatidylinositol-anchoredsurface protein, in human endothelial cells. J Immunol 144:593-8.Bryant RW, Granzow CA, Siegel MI, Egan RW, and Billah MM. 1991. Wheat germagglutinin and other selected lectins increase synthesis of decay-accelerating factor inhuman endothelial cells. J Inimunol 147:1856-62.Burns U, Waring JF, Reuter JJ, Stinski MF, and Ginder GD. 1993. Only the HLA classI gene minimal promoter elements are required for transactivation by humancytomegalovirus immediate early genes. Blood 81:1558.134Caras 1W, Weddell GN, Davitz MA, Nussenzweig V, Martin DW. 1987. Signal forattachment of a phospholipid membrane anchor in decay accelerating factor. Science238:1280-3.Caras 1W, Weddell GN, Davitz MA, Nussenzweig V, and Martin DW. 1990. Signalfor attachment of a phospholipid membrane anchor in decay-accelerating factor.Science 238:1280-3.Carroll MC, Aliquot EM, Katzman PJ, Klickstein KB, Smith JA, and Fearon DT. 1988.Organization of the genes encoding complement receptors type 1 and 2, decay-accelerating factor, and C4-binding protein in the RCA locus on human chromosome1. JExp Med 167:1271Chee MS, Bankier AT, Beck S. Bohni R, Brown CM, Cerny R, Horsnell T, HutchinsonDA, Kouzarides T, Martignetti JA, Freddie E, Satchwell SC, Tomlinson P, Weston KM,and Barrell BG. 1990. Analysis of the protein-coding of the sequence of humancytomegalovirus strain AD169. Current Topics in Microbiology and Immunology154:125-69.Chen Y-H, Bock G, Vornhagen R, Steindi F, Katinger H, and Dierich MP. 1994. HIV-1gp4l enhances MHC class I and ICAM-1 expression on H9 and U937 cells. Tnt ArchAllergy Immunol 104:227-31.Cheng-Rong L, Greenberg PD, Gilbert MJ, Goodrich JM, and Riddell SR. 1994.Recovery of HLA-restricted CMV-specific T-cell responses after allogeneic bonemarrow transplant: correlation with CMV disease and effect of ganciclovirprophylaxis. Blood 83:1971-9.135Cho S-W, Oglesby TJ, Hsi B-L, Adams EM, and Atkinson JP. 1991. Characterizationof three monoclonal antibodies to membrane co-factor protein (MCP) of thecomplement system and quantification of MCP by radioassay. Clin Exp Immunol83:257-61.Colberg-Poley AM, Santomenna LD, Harlow PP. Benfield PA, and Tenney DJ. 1992.Human cytomegalovirus US3 and UL 36-38 immediate-early proteins regulate geneexpression. J Virol 66:95-105.Cole JL, Housley GA, Dykman, Macdermott RP, and Atkinson JP. 1985.Identification of an additional class of C3-binding membrane proteins of humanperipheral blood leukocytes and cell lines. Proc Natl Acad Sci USA 82:859.Corvetta A, Pomponio G, Bencivenga R, Luchetti MM, Spycher M, Spaeth PJ, DanieliG. 1991. Low number of complement C3b/C4b receptors (CR1) on erythrocytes frompatients with essential mixed cryoglobulinemia, systemic lupus erythematosus andrheumatoid arthritis: relationship with disease activity, anticardiolipin antibodies,complement activation and therapy. J Rheumatol 18:1021-5.Coyne KE, Hall SE, Thompson ES, Arce MA, Kinoshita T, Fujita T, Anstee DJ, RosseW, and Lublin DM. 1992. Mapping of epitopes, glycosylation sites and complementregulatory domains in human decay-accelerating factor. J Immunol 149:2906-13.Compton T. 1993. An immortalized human fibroblast cell line is permissive forhuman cytomegalovirus infection. J Virol 67:3644-8.136Cooper NR, Jensen FC, Welsh RM, and Oldstone MBA. 1976. Lysis of RNA tumorviruses by human serum: direct antibody-independent triggering of the classicalcomplement pathway. J Exp Med 144:970-983.Corvetta A, Pomponio G, I3encienga R, Luchetti Mlvi, Spycher M, Spaeth PJ, andDanieli G. 1991. Low number of complement C3b/C4b receptors (CR1) onerythrocytes from patients with essential mixed cryoglobulinemia, SLE, and RA:relationship with disease activity, anticardiolipin antibodies, complement activationand therapy. J Rheumatol 18:1021-5.Cranage MP, Kouzarides T, Bankier AT, Satchwell S, Weston K, Tomlinson P, BarrelB, Hart H, Bell SE, Minson AC, and Smith GL. 1986. Identification of the humancytomegalovirus glycoprotein B gene and induction of neutralizing antibodies via itsexpression in recombinant vaccinia virus. EMBO J 5:3057-63.Cranage MP, Smith GL, Bell SE, Hart H, Brown C, Bankier AT, Tomlinson P, Barrel B,and Minson AC. 1988. Identification and expression of a human cytomegalovirusglycoprotein with homology to the Epstein-Barr virus BXLFZ product, varicella zostervirus gpIII and herpes simples virus type I gH. J Virol 62:1416-22.Dahiback B, Smith CA, Muller-Eberhard HJ. 1983. Visualization of human C4b-binding protein and its complexes with vitamin-K dependent protein S andcomplement protein C4b. Proc Natl Acad USA 80:3461-65.Dahlback B. 1991. Protein S and C4b-binding: components involved in the regulationof the protein C anticoagulant system. Thrombosis and Hemostasis 66:49-61.137Daly C, and Reich NC. 1993. Double-stranded RNA activates novel factors that bindto interferon-stimulated response element. Molec Cell Biol 13:3756-64.Daniels CA, Borsos T, Rapp HJ, Snyderman R, and Notkins AL. 1970. Neutralizationof sensitized virus by purified components of complement. Proc Natl Acad Sci USA65:528-35.Dankert JR and Esser AF. 1985. Proteolytic modification of human complementprotein C9: loss of poly (C9) and circular lesion formation without impairment offunction. Proc Natl Acad USA. 82:2128-32.Davies A, Simmons DL, Hale G, Harrison RA, Tighe H, Lachmann PJ, and WaldmannH. 1989. CD59, an LY-6-like protein expressed in human lymphoid cells, regulatesthe action of the complement membrane attack complex on homologous cells. J ExpMed 170:637-54.Davitz MA, Low MG, and Nussenzweig V. 1986. Release of decay-accelerating factor(DAF) from the cell membrane by PIPLC. J Exp Med 163:1150-61.DeMarchi JM, Schmidt CA, and Kaplan AS. 1980. Patterns of transcription of humancytomegalovirus in permissively infected cells. J Virol 35:277.DeMarchi JM. 1983. Nature of the block in the expression of some early virus genesin cells abortively infected with human cytomegalovirus. Virology 129:287-97.Demares MJ. 1989. Membrane-associated complement factor H on lymphoblastoidcell lines Raji express a co-factor activity for the factor I-mediated cleavage of C3b.Immunol 67:553-6.138Devine DV, RS Siegel, and WF Rosse. 1987. Interactions of the platelets inparoxysmal nocturnal hemoglobinuria with complement. J Clin Invest 79:131.Diosi P, Moldovan E, and Tonescu N. 1969. Latent cytomegalovirus infection inblood donors. Br J Med 4:660-2DiScipio RG. 1981. The binding of human complement proteins C5, factor B, 31H andproperdin to complement fragment C3b on zymosan. Biochem J 199:485-96.DiScipio RG. 1992. Ultrastructures and interactions of complement factors H and I. JImmunol 149:2592-9.Einhorn L and Ost A. 1984. Cytomegalovirus infection of human blood cells. J InfectDis 149:207-14.Eizuru Y, Ueno I, and Minamishima Y. 1988. Evaluation of immunoglobulin Gpreparations for anti-cytomegalovirus antibodies with reference to neutralizingantibody in the presence of complement. J Clin Microbiol 26:1881-3.Estes JE and Huang E-S. 1977. Stimulation of cellular thymidine kinases by humancytomegalovirus. J Virol. 24:13.Fearon DT. 1978. Regulation by membrane sialic acid of 1H-dependent decaydissociation of amplification C3 convertase of the alternative complement pathway.Proc Nati Acad Sci USA 75:1971-5.139Fearon DT. 1979. Regulation of the amplification C3 convertase of humancomplement by an inhibitory protein isolated from the human erythrocyte membrane.Proc Natl Acad Sci USA 76:5867-71.Fearon DT. 1980. Identification of the membrane glycoprotein that is the C3breceptor of the human erythrocyte, polymorphonuclear leukocyte, B lymphocyte, andmonocyte. J Exp Med 152:20-30.Fearon DT and Aheam JM. 1989. Complement receptor type 1 (C3b/C4b receptor;CD35) and complement receptor type 2 (C3d/Epstein-Barr virus receptor; CD21).Current Topics in Microbiol Immunol 153:83-98Fiala M, Payne JE, Berne TC, Moore W, Henle W, Montgomerie JS, Chetterjee SN, andGuze LB. 1975. Epidemiology of cytomegalovirus after transplantation andimmunosuppression. J Infect Dis 132:421-33Finberg RW, White W, and Nicholson-Weller A. 1992. Decay-accelerating factorexpression on either effector or target cells inhibits cytotoxicity by human naturalkiller cells. J Immunol 149:2055-60.Fingeroth JD, Clabby ML, and Strominger JD. 1988. Characterization of a Tlymphocyte Epstein-Barr virus/C3d receptor (CD21). J Virol 62:1442-7.Fischer E and Kazatchkine MD. 1983. Surface-dependent modulation by H of C5cleavage by the cell-bound alternative pathway C5 convertase of human complement.J Inimunol 130:2821-24.140Fries LF, Friedman HM, Cohen GH, Eisenberg RJ, Hammer CH, and Frank MM. 1986.Glycoprotein C of herpes simplex virus 1 is an inhibitor of the complement cascade. JImmunol 137:1636-41.Fujioka S and Yamada T. 1992. Longer in vivo survival of CD59- and DAF-almostnormal positive and partly positive erythrocytes in paroxysmal nocturnalhemoglobinuria as compared with negative erythrocytes: a demonstration bydifferential centrifugation and flow cytometry. Blood 79:1842-5.Fujita T, Inoue T, Ogawa K, lida K, and Tamura N. 1987. The mechanism of action ofdecay-accelerating factor. J Exp Med 166:1221-8.Garnett HM. 1982. Isolation of human cytomegalovirus from peripheral blood T cellsof renal transplant patients. J Lab Clin Med 99:92-7.Gefland MC, Fran MM, and Green I. 1975. A receptor for the third component ofcomplement in the human renal glomerulus. J Exp Med 142:1029-34.Gilbert GL, Hayes K, Hudson IL, and James J. 1989. Prevention of transfusion-acquired cytomegalovirus infection in infants by blood filtration to removeleukocytes. Lancet 1:1228.Gilbert MJ, Riddel SR, Li C-R, and Green berg PD. 1993. Selective interference withclass I MI-IC presentation of the major immediate-early protein following infectionwith human cytomegalovirus. J Virol 67:3461.141Gnann JW, Ahimen J, Svalander C, Olking L, Oldstone MB, and Nelson JA. 1988.Inflammatory cells in transplanted kidneys are infected by human cytomegalovirus.Am J Pathol 132:239-48.Gompels U and Minson A. 1986. The properties and sequence of glycoprotein H ofherpes simples virus type 1. Virol 153:230-47.Grundy JE, McKeating JA, and Griffiths PD. 1987a. Cytomegalovirus strain AD169binds 32-microglobulin in vitro after release from cells. J Gen Virol 68:777.Grundy JE, McKeating JA, Ward PJ, Sanderson AR, and Griffiths PD. 198Th. 32-microglobulin enhances the infectivity of CMV and when bound to the virus enablesclass I HLA molecules to be used as a virus receptor. J Gen Virol 68:793-803.Grundy JE, Ayles HM, McKeating JA, Butcher RG, Griffiths PD, and Poulter LW.1988. Enhancement of Class I HLA antigen expression by cytomegalovirus: role inamplification of virus infection. J Med Virol 25:483.Hansch GM, Weller PF, and Nicholson-Weller A. 1988. Release of C8 binding protein(C8bp) from the cell membrane by PIPLC. Blood 72:1089-92.Hara T, Kojima A, Fukuda H, Masaoka T, Fukumori Y, Matsumoto M, and Seya T.1992. Levels of complement regulatory proteins, CD35 (CR1), CD46 (MCP) and CD55(DAF) in human hematological malignancies. Br J Haematol 82:368-73.Harris SL, Frank I, Yee A, Cohen GH, Eisenberg RJ, and Friedman HM. 1990.Glycoprotein C of herpes simplex virus type 1 prevents complement-mediated celllysis and virus neutralization. J Infect Dis 162:331-7.142Hatano H, Oh JO, Teug Ou KH, and Minasi P. 1988. Induction of Fc and C3breceptors on rabbit corneal cells by herpes simplex virus. Invest Ophthal Vis Sci29:1352-6.Hayes K, Afford C, and Britt W. 1987. Antibody response to virus-encoded proteinsafter cytomegalovirus mononucleosis. J Infect Dis 156:615-21.Heckl-Ostreicher B, Ragg S, Drechlser M, Scherthan H, and Royer-Pokora B. 1993.Localization of the human CD59 gene by fluorescence in situ hybridization andpulsed-field gel electrophoresis. Cytogenet Cell Genet 63:144-6.Heieren MH, Kim Y, and Balfour HH. 1988. Human cytomegalovirus infection ofkidney glomerular visceral epithelial and tubular epithelial cells in culture.Transplantation 46:426-32.Herold BC, WuDunn D, Soltys N and Spear PG. 1991. Glycoprotein C of herpessimplex virus type 1 plays a principal role in the adsorption of virus to cells and ininfectivity. J Virol 65:1090-8.Ho M. 1991 Cvtomegalovirus biology and infection. 2nd edn Plenum. New York.Ho DD, Rota TR, Andrews CA, and Hirsch MS. 1984. Replication of humancytomegalovirus in endothelial cells. J Infect Dis. 150:956.Holguin MH, Fredrick LR, Bernshaw NJ, Wilcox LA, and Parker CJ. 1989. Isolationand characterization of a membrane protein from normal human erythrocytes that143inhibits reactive lysis of the erythrocytes of paroxysmal nocturnal hemoglobinuria. JClin Invest. 84:7-17.Holguin MH, Martin CB, Bernshaw NJ, and Parker CJ. 1992. Analysis of the effects ofactivation of the alternative pathway of complement on erythrocytes with an isolateddeficiency of decay accelerating factor. J Immunol 148:498-502.Holguin MH, Martin CB, Weis JH, and Parker CJ. 1993. Enhanced expression of thecomplement regulatory protein, membrane inhibitor of reactive lysis (CD59), isregulated at the level of transcription. Blood 82:968-77,Homeister JW, Satoh PS, Kilgore KS, and Lucchesi BR. 1993. Soluble complementreceptor type 1 prevents human complement-mediated damage of the rabbit isolatedheart. J Immunol 150:1055-64.Hosenpud JD, Chou S. and Wagner CR. 1991. Cytomegal.ovirus-induced regulationof MHC class I antigen expression in human aortic smooth muscle cells.Transplantation. 52:896.Hourcade D, Miesner DR. Atkinson JP, and Holers VM. 1988. Identification of analternative polyadenylation site in the human C3b/C4b receptor (complementreceptor type 1) transcriptional unit and prediction of a secreted form of complementreceptor type 1. J Exp Med 168:1255.Hourcade D, Miesner DR, Bee C, Zeldes W, and Atkinson JP. 1990. Duplication anddivergence of the amino-terminal coding region of the complement receptor 1 (CR1)gene. J Biol Chem 265:974144Hourcade D, Garcia AD, Post TW, Taillon-Miller F, Holers VM, Wagner LM, Bora NS,and Atkinson JP. 1992. Analysis of the human regulators of complement activation(RCA) gene cluster with yeast artificial chromosomes (YACs) Genomics 12:289Huang E-S, and Kowalik TF. 1994. In: Molecular aspects of human cytomegalovirusdiseases. Becker Y, and Daral G (eds). Springer-Verlag. New York. ppl-45.Huemer HP, Larcher C, and Coe NE. 1992. Pseudorabies virus glycoprotein ifiderived from virions and infected cells binds to the third component of complement.Virus Research 23:271-80.Huemer HP, Larcher C, van Drunen Littel-van den Hurk S, and Babiuk LA. 1993.Species selective interaction of Alphaherpesvirinae with the “unspecific” immunesystem of the host. Arch Virol 130:353-64.Huleihel M, Douvdevani A, Segal S. and Apte RN. 1990. Regulation of interleukin 1generation in immune-activated fibroblasts. Eur J Immunol 20:731-8.Ibanez CE, Schrier R, Ghazal F, Wiley C, and Nelson JA. 1991. Humancytomegalovirus productively infects primary differentiated macrophages. J Virol.65:6581-8.lida K and Nussenzweig V. 1981. Complement receptor is an inhibitor of thecomplement cascade. J Exp Med 153:1138-50.lida K, Nadler L, and Nussenzweig V. 1983. Identification of the membrane receptorfor the complement fragment C3d by means of a monoclonal antibody. J Exp Med158:1021-33.145Irmiere A and Gibson W. 1983. Use of a glycerol-tartrate gradient to separate HCMVvirions into virions, non-infectious, enveloped particles, and dense bodies. Virol130:118.Iwamoto GK, Monick MM, Clark BD, Auron FE, Stinski MF, and Hunninghake GW.1990. Modulation of interleukin 1 beta gene expression by the immediate early genesof human cytomegalovirus. J Clin Invest 85:1853-7.Iwata K, Seya T, Ariga H, and Nagasawa S. 1994. Expression of a hybrid complementcomponent regulatory protein, membrane cofactor protein decay accelerating factoron Chinese hamster ovary. J Immunol 152:3436-44.Johnstone RW, Russel SM, Loveland BE, and McKenzie IFC. 1993. Polymorphicexpression of CD46 protein isoforms due to tissue-specific RNA splicing. MolecImmunol 30:1231-41.Jordan MC. 1983. Latent infection and the elusive cytomegalovirus. Rev Infect Dis5:205-15.Kaariainen L, Klemola E, and Paloheimo J. 1966. Rise of HCMV antibodies in aninfectious-mononucleosis-like syndrome after transfusion. Br Med J 1:1270-2.Kazatchkine MD, Fearon DT, Appay MD, Mandet C, and Bariety J. 1982.Immunohistochemical study of the human glomerular C3b receptor in normal kidneyand in seventy-five cases of renal diseases. J Clin Invest 69:900-12.146Kerry JA, Priddy MA, and Stenberg RM. 1994. Identification of sequence elements inthe human cytomegalovirus DNA polymerase gene promoter required for activationby viral gene products. J Virol 68:4167-76.Kinoshita T, Medof ME, Silber R, and Nussenzweig V. 1985. Distribution of decay-accelerating factor in the peripheral blood of normal individuals and patients withparoxysmal nocturnal hemoglobinuria. J Exp Med 162:75-92.Kiemola E and Kaariainen L. 1965. HCMV as a possible cause of a disease resemblinginfectious mononucleosis. Br Med J 2:1099-1102.Klickstein LB, Bartow TJ, Miletic V, Rabson LD, Smith JA, and Fearon DT. 1988.Identification of distinct C3b and C4b recognition sites in the human C3b/C4breceptor (CR1, CD35) by deletion mutagenesis. J Exp Med 168:1699-1717.Kline JN, Geist U, Monick MM, Stinski MF, and Hunninghake GW. 1994. Regulationof expression of the IL-i receptor agonist (IL-ira) gene products of the humancytomegalovirus immediate early genes. J Immunol 152:2351-2357.Kojima A, Iwata K, Seya T, Matsumoto M, Ariga H, Atkinson JP, and Nagasawa S.1993. Membrane cofactor protein (CD46) protects cells predominantly fromalternative complement pathway-mediated C3-fragment deposition and cytolysis. Jlmmunol 151:1519-27.Korn J., Torres D., and Downe E. 1983. Fibroblast prostaglandin E2 synthesis. J ClinInvest 71:1240.147Kotwal GJ and Moss B. 1988. Vaccinia virus encodes a secretory polypeptidestructurally related to complement control proteins. Nature. 335:176-8.Kotwal GJ, Isaacs SN, McKenzie R, Frank MM, and Moss B. 1990. Inhibition of thecomplement cascade by the major secretory protein of vaccinia virus. Science250:827-9.Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the headof bacteriophage T4. Nature 227:680-685.La Femina and Hayward F. 1983. Replicative forms of human cytomegalovirus DNAwith joined termini are found in permissively infected human cells but not in non-permissive Balb/c-3T3 mouse cells. J Gen Virol 64:373-89.Lang DF and Noren B. 1968. Cytomegaloviremia following congenital infection. JPediatr 73:812.Langenhuysen MMAC. 1972. 1gM levels, specific 1gM antibodies and liverinvolvement in cytomegalovirus infection. Scand J Infect Dis 4:113-8.Lanza F, Fagioli F, Gavioli R, Spisani S, Malavasi F, Castoldi GL, and Traniello S.1991. Evaluation of CR1 expression in neutrophils from chronic myeloid leukemia:relationship between prognosis and cellular activity. Brit J Haematol 77: 66-72.Lanza F and Castoldi G. 1992. Complement receptor 1 (CR1) expression in chronicmyeloid leukemia. Leukemia & Lymphoma 8:35-41.148Lathey JL, Wiley CA, Verity MA and Nelson JA. 1990. Cultured human braincapillary endothelial cells are permissive for infection by human cytomegalovirus.Virol 176:266.Lathey JL and Spector SA. 1991. Unrestricted replication of human cytomegalovirusin hydrocortisone-treated macrophages. J Virol. 65:6371-5.Laubscher A, Bluestein HG, Spector SA, and Zvaifler NJ. 1988. Generation of humancytomegalovirus-specific cytotoxic T lymphocytes in a short-term culture. J ImmunolMethods 110:69.Leino L and Lilius EM. 1992. The up- and down-modulation of immunoglobulin GFc receptors and complement receptors on activated human neutrophils depends onthe nature of activator. J Leukoc Biol 51:157-63.Lewis RB, Matzke DS, Albrecht TB, and Pollard RB. 1986. Assessment of the presenceof cytomegalovirus-neutralizing antibody by a plaque-reduction assay. Rev Infect Dis8:S434-8.Lonnqvist B, Ringden 0, Wahren B, Gahrton G, and Lundgren G. 1984. HCMVinfection associated with and preceding chronic graft-versus-host disease.Transplantation 38:465.Lopez C, Simmons RL, Mauer SM, Najarian JS, and Good RA. 1974. Association ofrenal allograft rejection with virus infection. Am J Med 56:280.149Lowell CA, Klickstein LB, Carter RH, Mitchell JA, Fearon DT, and Ahearn JM. 1989.Mapping of the Epstein-Barr virus and C3dg binding sites to a common domain oncomplement receptor type 2. J Exp Med 170:1931-46.Lublin DM, Krsek-Staples J, Pangbum MK, and Atkinson JP. 1986. Biosynthesis andglycosylation of the human complement regulatory protein decay-accelerating factor.J linmunol 137:1629-35.Lublin DM and Coyne KE. 1991. Phospholipid-anchored and transmembraneversions of either decay-accelerating factor or membrane cofactor protein show equalefficiency in protection from complement-mediated cell damage. I Exp Med. 174:35-44.Lund-Johansen F, Olweus J, Symington FW, Arli A, Thompson JS, Vilella R, Skubits K,and Horejsi V. 1993. Activation of human monocytes and granulocytes bymonoclonal antibodies to glycosylphosphatidylinositol-anchored antigens. Eur JImmunol 23:2782-91.Mach M, Utz U, and Fleckenstein B. 1986. Mapping of the major glycoprotein gene ofhuman cytomegalovirus. J Gen Virol 67:1461-7.Maciejewski JP, Bruening EE, Donahue RE, Mocarski ES, Young NS, and St. Jeor SC.1992. Infection of hematopoetic progenitor cells by HCMV. Blood. 80:170-78.Mahoney JE, Urakaze M, HallS, DeGasperi R, Chang H, Sugiyama E, Waren CD,Borowitz M, Nicholson-Weller A, Rosse WE, and Yeh ETH. 1992. Defectiveglycosyiphosphatidylinositol anchor synthesis in paroxysmal nocturnalhemoglobinuria granulocytes. Blood 79:1400-3.150McAllister RIVI, Straw, RM, Filbert JE, and Goodheart CR. 1963. HCMV: cytochemicalobservations of intracellular lesion development correlated with viral synthesis andrelease. Virology 19:521-31.McCarthy D, Taylor MJ, Bernhagen J, Perry JO, and Hamblin AS. 1992. Leukocyteintegrin and CR1 expression on peripheral blood leukocytes of patients withrheumatoid arthritis. Annals of Rheumatic Diseases 51:307-12.McKeating JA, Griffiths PD, and Grundy JE. 1987. Cytomegalovirus in urinespecimens has host microglobulin bound to the viral envelope: a mechanism ofevading the host immune system? J Gen Virol 68:785.McNearney TA, Odell C, Holers VM, Spear PG, and Atkinson JP. 1987. Herpessimplex glycoproteins gC-1 and gC-2 bind to the third component of complement andprovide protection against complement-mediated neutralization of viral infectivity. JExp Med 166:1525-35.Medof ME, lida K, Mold C, and Nussenzweig V. 1982. Unique role of thecomplement receptor CR1 in the degradation of C3b associated with immunecomplexes. J Exp Med 156:1739-54.Medof ME, Lam T, Prince GM, and Mold C. 1983. Requirement for human red bloodcells in inactivation of C3b in immune complexes and enhancement of binding tospleen cells. J Immunol 130:1336-40.151Medof ME, Kinoshita T, and Nussenzweig V. 1984. Inhibition of complementactivation on the surface of cells after incorporation of decay-accelerating factor intotheir membranes. J Exp Med 160:1558-78.Mehta K and Lopez-Berestein G. 1986. Expression of tissue transglutaminase incultured monocytic leukemia (TFIP-1) cells during differentiation. Cancer Res46:1388-94.Menezes J, Seigneurin JM, Patel P, Bourkas A and Lenoir C. 1977. Presence ofEpstein-Barr virus receptors, but absence of virus penetration, in cells of an Epstein-Barr virus genome-negative human lymphoid T line (Molt4). J Virol 22:816-21.Men S, Waldmann H, and Lachmann PJ. 1991. Distribution of protectin (CD59), acomplement membrane attack inhibitor, in normal human tissues. 65:532-7.Men S, Maffila P. and Renkonen R. 1993. Regulation of CD59 expression on thehuman endothelial cell line EA.hy 926. Eur J Immunol 23:2511-6.Merry AH, Rawlinson V, Uchikawa M, Watts MJ, Hodson C, and Sims RB. 1989.Lack of abnormal sensitivity to complement-mediated lysis in erythrocyte deficiencyonly in decay-accelerating factor. Biochem Soc Trans 17:514.Middeldorp JM, Jongsma J, and The TH. 1985. Cytomegalovirus early and latemembrane antigens detected by antibodies in human convalescent sera. J Virol54:240-4.Middeldorp JM, Jongsma J, and The TH. 1986. Killing of human cytomegalovirusinfected fibroblasts by antiviral antibody and complement. J Infect Dis 153:48-55.152Mitomo K, Fujita T, and lida K. 1987. Functional and antigenic properties ofcomplement receptor type 2, CR2. J Exp Med 165:1424-9.Mold C, Bradt BM, Nemerow GR, and Cooper NR. 1987. Epstein-Barr virus regulatesactivation and processing of the third component of complement. J Exp Med 168:949-69.Montefiori DC, Zhou J, and Shaff DI. 1992. CD4-independent binding of HIV-1 to theB lymphocyte receptor CR2 (CD21) in the presence of complement and antibody. ClinExp Immunol 90:383-389.Montefiori DC, Cornell RJ, Zhou JY, Zhou JT, Hirsch VM, and Johnson PR. 1994.Complement control proteins CD46, CD55, and CD59, as common surfaceconstituents of human and simian immunodeficiency viruses and possible targets forvaccine protection. Virology 205:82-92.Morgan BP, Patel AK and Campbell AK. 1987. The ring-like classical complementlesion is not the functional pore of the membrane attack complex. Biochem Soc Trans15:659-660.Morgan BP. 1989. Complement membrane attack on nucleated cells: resistance,recovery and non-lethal effects. Biochem J 264: 1-14.Morgan BP, van den Berg CW, Davies EV, Hallet MB, and Horejsi V. 1993. Crosslinking of CD59 and of other glycosyl phosphatidylinositol-anchored molecules onneutrophils triggers cell activation via tyrosine kinase. Eur J Immunol 23:2841-50.153Moutabarrik A, Nakanishi I, Namiki M, Hara T, Matsumoto M, Ishibashi M,Okuyama A, Zaid D, and Seya T. 1993. Cytokine-mediated regulation of the surfaceexpression of complement regulatory proteins, CD46 (MCP), CD55 (DAF), and CD59on human vascular endothelial cells. Lymphokine Cytokine Res 12:167-72Myerson D, Hackman RC, Nelson JA, Ward DC, and McDougall JK. 1984.Widespread presence of histologically occult cytomegalovirus. Hum Pathol 15:430-9.Nagasawa S, Unno H, Ichihara C, Koyama J, and Koide T. 1983. Human C4b-bindingprotein. C4bp. Chymotryptic cleavage and location of the 48 kDa active fragmentwithin C4bp. FEBS letters 164:135-38.Nemerow GR, Mold C, Schwend VK, Tollefson V, and Cooper NR. 1987.Identffication of gp350 as the viral glycoprotein mediating aftachment of Epstein-Barrvirus (EBV) to the EBV/C3d receptor of B cells: sequence homology of gp350 and C3complement fragment C3d. J Virol 61:1416-20.Neumann JR. Morency CA, and Russian KO. 1987. A novel rapid assay forchloramphenicol acetyltransferase gene expression. BioTechniques 5:444-6.Nicholson-Weller, A., J. Burge, DT. Fearon, PF. Weller, and KF. Austen. 1982.Isolation of a human erythrocyte membrane glycoprotein with decay-acceleratingactivity for C3 convertases of the complement system. J Immunol 129:184.Nicholson-Weller A, Russian D, and Austen KF. 1986. Natural killer cells aredeficient in the surface expression of the complement regulatory protein, decayaccelerating factor. J Immunol 137:1275.154Nielsch U, Pine R, Zimmer SG, and Babiss LE. 1992. Induced expression of theendogenous beta interferon gene in adenovirus type 5-transformed fibroblasts. J Virol66:1884-90.Nishiyama Y. and Rapp F. 1980. Enhanced survival of ultraviolet-irradiated herpessimplex virus in human cytomegalovirus infected cells. Virology 100:189.Norris J, Hall S, Ware RE, Kamitani T, Chang H, Yeh E, and Rosse WF. 1994.Glycosyl-phosphatidylinositol anchor synthesis in paroxysmal nocturnalhemoglobinuria: partial or complete defect in an early step. Blood 83:816-21.Nowak B, Sullivan C, Sarnow P, Thomas T, Bricout F, Nicolas JC, Fleckenstein B andLevine AJ. 1984. Characterization of monoclonal antibodies and polyclonal immunesera directed against human cytomegalovirus proteins. Virol 132:325-38.Numazaki K, Goldman H, Bai XQ, Wong I, and Wainberg MA. 1989a. Effects ofinfection by HP/-i, cytomegalovirus, and human measles virus on cultured humanthymic epithelial cells. Microbiol Immunol 33:733-45.Numazaki K, DeStephano L, Wong, I, Goldman H, Spira, and Wainberg MA. 1989b.Replication of cytomegalovirus in human thymic epithelial cells. Med MicrobiolImmunol 178:89-98.Ohashi H, Hotta T, Ichikawa A, Kinoshita T, Taguchi, Kiguchi T, Ikezawa H, and SaitoH. 1994. Peripheral blood cells are predominantly chimeric of affected and normalcells in patients with paroxysmal nocturnal hemoglobinuria: simultaneousinvestigation on clonality and expression of glycophosphatidylinositol-anchoredproteins. Blood 83:853-9.155Pangburn MX and Muller-Eberhard HI. 1978. Complement C2 convertase: cellsurface restriction of 131H control and generation of restriction on neuraminidasetreated cells. Proc Nati Acad Sci USA. 75:2416-20.Pangburn MK. 1986. Differences between the binding sites of the complementregulatory proteins DAF, CR1, and factor H on C3 convertases. J Immunol 136:2216-2221.Pardo-Manuel F, Rey-Campos J, Hillarp A, Dahiback 13 and de Cordoba SR. 1990.Human genes for the alpha and beta chains of complement C4b-binding protein areclosely linked in a head-to-tail arrangement. Proc Natl Acad Sci USA 87:4529.Pascual M, Danielsson C, Steiger G, and Schifferli JA. 1994. Proteolytic cleavage ofCR1 on human erythrocytes in vivo: evidence for enhanced cleavage in AIDS. Eur JImmunol 24:702-8.Pass EF, Griffiths PD, and August AM. 1983. Antibody response to cytomegalovirusafter renal transplantation: comparison of patients with primary and recurrentinfections. J Infect Dis 147:40-6.Pereira L, Hoffman M, Gallo D, and Cramer N. 1982. Monoclonal antibodies tohuman cytomegalovirus: Three surface membrane proteins with uniqueimmunological and electrophoretic properties specify cross reactive determinants.Infect Immunol 36:924-32.156Pereira L, Hoffman M, Tatsuno M, and Dondero D. 1984. Polymorphism of humancytomegalovirus glycoproteins characterized by monoclonal antibodies. Virol 139:73-86.Piddlesden SJ, Storch MK, Hibbs M, Freeman AM, Lassmann, H, and Morgan BP.1994. Soluble recombinant complement receptor 1 inhibits inflammation anddemyelination in antibody-mediated demyelinating experimental allergicencephalomyelitis. J Immunol 152:5477-84Plummer G. 1973. Cytomegalovirus of man and animals. Prog Med Virol 15:92-125.Post TW, Arce MA, Lisewski MK, Thompson ES, Atkinson JP, and Lublin DM. 1990.Structure of the gene for human complement protein decay accelerating factor. JImmunol 144:740.Pramoonjago P, Kinoshita T, Hong K, Takata-Kozono Y, Kozono H, Inagi R, andInoue K. 1992. Bactericidal activity of C9-deficient human serum. J Immunol148:837-43.Prieto J, Eklund A, and Patarroyo M. 1994. Regulated expression of integrins andother adhesion molecules during differentiation of monocytes and macrophages.Cellular Immunology 156:191-211.Quinnan GV, Kirmani N, Rook AH, Manischewitz JF, Jackson L, Moreshi G, SantosGW, Saral R, and Burns WH. 1982. Cytotoxic T cells in cytomegalovirus infection:HLA-restricted T-lymphocyte and non-T-lymphocyte cytotoxic responses correlatewith recovery from CMV in bone marrow-transplant recipients. N Engl J Med 307:7.157Radwan AT and Burger D. 1973. Complement-requiring neutralization of equinearteritis virus by late antisera. Virol 51:71-7.Ramirez ML, Virmani M, Garon C, and Rosenthal U. 1979. Defective virions ofhuman cytomegalovirus. Virology 96:311-14.Ramos OF, Sarmay G, Klein E, Yefenof E, and Gergely J. 1985. Complement-dependent cellular cytotoxicity: lymphoblastoid lines that activate complementcomponent 3 (C3) and express C3 receptors have increased sensitivity to lymphocyte-mediated lysis in the presence of fresh human serum. Proc Natl Acad Sci USA82:5470-4.Rasmussen LE, Kelsall DC, Nelson R, Carney W, Hirsch M, Winston D, Preiksaitis J,and Merigan TC. 1982. Virus-specific IgG and 1gM antibodies in normal andimmunocompromised subjects infected with cytomegalovirus. J Infect Dis 145:191-9.Rasmussen LE, Nelson RM, Kelsall DC, and Merigan TC. 1984. Murine monoclonalantibody to a single protein neutralizes the infectivity of human cytomegalovirus.Proc Natl Acad Sci 81:876-80.Rasmussen L, Mullenax J, Nelson M, and Merigan TC. 1985a. Humancytomegalovirus polypeptides stimulate neutralizing antibody in vivo. Virol 145:186-90.Rasmussen L, Mullenax J, Nelson R, and Merigan TC. 1985b. Viral polypeptidesdetected by a complement-dependent neutralizing murine monoclonal antibody tohuman cytomegalovirus. J Virol 55:274-80.158Rasmussen L, Nelson M, Neff M, and Merigan TC. 1988. Characterization of twodifferent human cytomegalovirus glycoproteins which are targets for virusneutralizing antibody. Virol 163:308-18.Rasmussen LE. 1991. Gene products of cytomegalovirus and their immunologicsignificance. In: Human cvtomegalovirus: biology and infection. 2nd edn. Ho M(ed) Plenum, New York.Reid KBM, Bentley DR. Campbell RD, Chung LP, Sim RB, Kristensen T, and Tack BF.1986. Complement system proteins which interact with C3b or C4b. Immunol Today7:230-4.Reusser P. Riddell SR, Meyers JD and Greenberg PD. 1991. Cytotoxic T-lymphocyteresponse to CMV after human allogeneic bone marrow transplantation: Pattern ofrecovery and correlation with CMV infection and disease. Blood 78:1373.Rey-Campos J, Rubinstein P, and de Cordoba SR. 1988. A physical map of the humanregulator of complement activation gene cluster linking the complement genes CR1,CR2, DAF and C4BP. J Exp Med 167:664.Reynes M, Aubert JP, Cohen JHM, Audouin J, Tricotet V, Diebold J, and KazatchkineMV. 1985. Human follicular dendritic cells express CR1, CR2, and CR3 complementreceptor antigens. J Immunol 135:2687-94.Reynolds DW, Stagno 5, Hostly TS, Tiller M, and Alford CA. 1973. Maternalcytomegalovirus excretion and perinatal infection. N Engl J Med 289:1-5.159Rice GPA, Schrier RD, and Oldstone MBA. 1984. Cytomegalovirus infects humanlymphocytes and monocytes: virus expression is restricted to immediate-early geneproducts. Proc Natl Acad Sci 81:6134-8.Riddell SR, Rabin M, Geballe AP, Britt WJ, and Greenberg PD. 1991. Class I MHCrestricted cytotoxic T lymphocyte recognition of cells infected with human CMV doesnot require endogenous viral gene expression. J Immunol 146:2795.Rinaldo CR, Black PH, and Hirsch MS. 1977. Interaction of cytomegalovirus withleukocytes from patients with mononucleosis due to cytomegalovirus. J Infect Dis136:667-8.Rinaldo CR, Camey WP, Richter BS, Black PH, and Hirsch M. 1980. Mechanisms ofimmunosuppression in cytomegalovirus mononucleosis. J Infect Dis 141:488.Ripoche J, Day AJ, Harris TJR, and Sim RB. 1988. The complete amino acid sequenceof human complement factor H. Biochem J 249:593.Roberts WL, Santikarn S, Reinhold VN, and Rosenberry TL. 1988. Structuralcharacterization of the glycoinositol phospholipid membrane anchor of humanerythrocyte acetylcholinesterase by fast atom bombardment mass spectrometry. J BiolChem 263:18776-84.Rodriguez de Cordoba S and Rubenstein P. 1984. Genetic polymorphism of humanfactor H (131H). J Immunol 132:1906-8.160Rodriguez de Cordoba S and Rubenstein P. 1987. New alleles of C4-binding proteinand factor H and further linkage data in the regulator of complement activation(RCA) gene cluster in man. Immunogenetics 25:267-8.Roizman B. 1990. Herpesviridae: a brief introduction. In: Virology. 2nd edn. FieldsBN and Knipe DM (eds). Raven Press Ltd. New York.Rollins SA, Thao J, Ninomiya H, and Sims PJ. 1991. Inhibition of homologouscomplement by CD59 is mediated by a species-selective recognition conferredthrough binding to C8 within C5b-8 or C9 within C5b-9. J Immunol 146:2345-51.Rook AH, Quirinan GV, Fredrick WJR, Manischewitz JF, Kirimani N, Dantzier T, LeeBB, and Currier CB. 1984. Importance of cytotoxic lymphocytes during CMVinfection in renal transplant recipients. Am J Med 76:385.Ross GD, Polley MJ, Rabellino EM, and Grey HM. 1973. Two different complementreceptors on human lymphocytes. One specific for C3b and one specific for C3binactivator-cleaved C3b. J Exp Med 138:798-811.Ross GD and Medof ME. 1985. Membrane complement receptors specific for boundfragments of C3. Adv Immunol 37:217-67.Rundell BB and Betts RF. 1982. Neutralization and sensitization of cytomegalovirusby IgG antibody, Anti-IgG antibody, and complement. J Med Virol 10:109-118.Santomenna LD and Colberg-Poley AM. 1990. Induction of cellular hsp70 expressionby human cytomegalovirus. J Virol 64:2033.161Schonermark 5, Rauterberg EW, Shin ML, Loke S, Roelcke D, and Hansch GM. 1986.Homologous species restriction in lysis of human erythrocytes: a membrane-derivedprotein with C8-binding capacity functions as an inhibitor. J Immunol 136:1772-6.Schrier RD., Nelson J.A., and Oldstone M.B.A. 1985. Detection of humancytomegalovirus in peripheral blood lymphocytes in a natural infection. Science230:1048-51.Schrier RD and Oldstone MBA. 1986. Recent clinical isolates of CMV suppresshuman CMV-specific human leukocyte antigen-restricted cytotoxic T lymphocyteactivity. J Virol 59:127.Schwachula A, Riemann D, Kehlen A, and Langner J. 1994. Characterization of theimmunotype and functional properties of fibroblast-like synoviocytes in comparisonto skin fibroblasts and umbilical vein endothelial cells. Immunobiol 190(1-2):67-92.Sedlak J, Hunakova L, Duraj J, Grofova M, and Chorvath B. 1993. Modulation ofprotectin (CD59) cell surface expression on human neoplastic cell lines. Neoplasma40:337-40.Selvaraj P, Rosse WF, Silber R, and Springer TA. 1988. The major Fc receptor in bloodhas a phosphatidylinositol anchor and is deficient in paroxysmal nocturnalhemoglobinuria. Nature 333:565-7.Sengelov H, Kjeldsen L, Kroeze W, Berger M, and Borregaard N. 1994. Secretoryvesicles are the intracellular reservoir of complement receptor 1 in human neutrophils.J Immunol 153:804-10.162Seya T and Atkinson JP. 1989. Functional properties of membrane cofactor protein ofcomplement. Biochem J 264:581-8.Seya T, Hara T, Matsumoto M, and Akedo H. 1990. Quantitative analysis ofmembrane cofactor protein (MCP) of complement. High expression of MCP onhuman leukemia cell lines, which is down-regulated during cell differentiation. JImmunol 145:238-45.Seya T, Okada M, Matsumoto M, Hong K, Kinoshita T, and Atkinson JP. 1991.Preferential inactivation of the CS convertase of the alternative complement pathwayby factor I and membrane cofactor protein (MCP). Molec Immunol 28:1137-47.Seya T, Matsumoto M, Hara T, Hatanaka M, Masaoka T, and Akedo H. 1994.Distribution of C3-step regulatory proteins of the complement system, CD35 (CR1),CD46 (MCP), and CD55 (DAF), in hematological malignancies. Leukemia &Lymphoma 12:395-400.Shenoy-Scaria AM, Kwong J, Fujita T, Olszowy MW, Shaw AS, and Lublin DM. 1992.Signal transduction through decay-accelerating factor. J Immunol 149:3535-41.Shibata T, Cosio FG, and Birmingham DJ. 1991. Complement activation induces theexpression of decay-accelerating factor on human mesangial cells. J Immunol147:3901-8.Shibuya K, Abe T, and Fujita T. 1992. Decay-accelerating factor functions as a signaltransducing molecule for human monocytes. J Immunol 149:1758-62.163Sissons JGP and Oldstone MBA. 1980. Antibody-mediated destruction of virus-infected cells. Adv Immunol. 29:209-260.Sixbey JW, Davies DS, Young LS, Hutt-Fletcher L, Tedder TF, and Rickinson AB.1987. Human epithelial cell expression of an Epstein-Barr virus receptor. J Gen Virol68:805-11.Smiley ML, Hoxie JA, and Friedman. 1985. Herpes simplex virus type 1 infection ofendothelial, epithelial, and fibroblast cells induces a receptor for C3b. J Immunol134:2673-8.Smiley ML, Mar EC, and Huang ES. 1988. Cytomegalovirus infection and viral-induced transformation of human endothelial cells. J Med Virol 25:213.Smith CA, Pangburn MK, Vogel C-W, and Muller-Eberhard HJ. 1983. Structuralinvestigations of properdin and factor H of human complement. Immunobiology164:298.Smith JD. 1986. Human cytomegalovirus: demonstration of permissive epithelialcells and non-permissive fibroblastic cells in a survey of human cell lines. J Virol60:583-8.Smith KL, Cobain T, and Dunstan RA. 1993. Removal of cytomegalovirus DNA fromdonor blood by filtration. Br J Haematol 83:640.Snyder DB, Myrup AD, and Dutta SK. 1981. Complement requirement for virusneutralization by antibody and reduced serum complement levels associated withexperimental equine herpesvirus 1 infection. Infect Immun 31:636-40.164Solder BM, Schulz TF, Hengster P. Lower J, Larcher C, Bitterlich G, Kurth R, WachterH, and Dierich MP. 1989. HIV and 1-IIV-infected cells differentially activate thehuman complement system independent of antibody. Immunol Let 22:135-146.Span AHM, Mullers W, Miltenburg AMM, and Bruggeman CA. 1991.Cytomegalovirus induced PMN adherence in relation to an ELAM-1 antigen presenton infected endothelial cell monolayers. Immunol. 72:355-60.Spear GT, Sullivan BL, Landay AL, and Lint TF. 1990. Neutralization of humanimmunodeficiency virus type 1 by complement occurs by viral lysis. J Virol 64:5869-73.Spencer ES and Andersen HK. 1972. The development of immunofluorescentantibodies as compared with complement-fixing and virus-neutralizing antibodies inhuman cytomegalovirus infection. Scand J Infect Dis 4:109-12.Stanier F, Taylor DL, Kitchen AD, Wales N, Tryhorn Y, and Tyms AS. 1989.Persistence of cytomegalovirus in mononuclear cells in peripheral blood from blooddonors. Br Med J 299:897-8.Stannard LM. 1989. 132-microglobulin binds to the tegument of Cytomegalovirus: animmunogold study. J Gen Virol 70:2179.Stefanova I and Horejsi V. 1991. Association of the CD59 and CD55 cell surfaceglycoproteins with other membrane molecules. J Inimunol 147:1587-92.165Stinski M.F. 1977. Synthesis of proteins and glycoproteins in cells infected withhuman cytomegalovirus. J Virol. 23:751.Stinski MF, Mocarski ES, and Thomsen DR. 1979. DNA of human cytomegalovirus:size heterogeneity and defectiveness resulting from serial undiluted passage. J Virol31: 231-9.Stinski MF. 1990. Cytomegalovirus and its replication. In: Virology. 2nd edn. FieldsBN and Knipe DM (eds). Raven Press Ltd. New York.St. Jeor S and Weisser A. 1977. Persistence of cytomegalovirus in humanlymphoblasts and peripheral leukocyte cultures. Infect Immun. 15:402-9.Tandon N, Yan SL, Morgan BP, and Weetman AP. 1994. Expression and function ofmultiple regulators of complement activation in autoimmune thyroid disease.Immunol 81:643-7.Tate CG, Uchikawa M, Tanner MJ, Judson PA, Parsons SF, Mallinson G, and AnsteeDJ. 1989. Studies on the defect which causes absence of decay accelerating factor(DAF) from the peripheral blood cells of an individual with the Inab phenotype.Biochem J 261:489-93.Tatsumi E, Harada S Kuszynski C, Volsky D, Minowada J, Purtilo DT. 1985.Catalogue of Epstein-Barr virus (EBV) receptors on human malignant and nonmalignant hemopoetic cell lines. Leuk Res 9:231-8.Tausk F and Gigli I. 1990. The human C3b receptor: function and role in humandiseases. J Invest Dermatol 94:141S-5S.166Taylor-Wiedeman J, Sissons JG, Borysiewicz LK, and Sinclair JH. 1991. Monocytesare a major site of persistence of human cytomegalovirus in peripheral bloodmononuclear cells. J Gen Virol 72:2059-64.Tegtmeier GE. 1989. Posttransfusion cytomegalovirus infections. Arch Pathol LabMed 113:236.Telen MJ, Hall SE, Green AM, Moulds H. and Rosse WF. 1988. Identification ofhuman erythrocyte blood group antigens on decay-accelerating factor (DAF) anderythrocyte phenotype negative for DAF. J Exp Med 167:1993-8.The TH and Langenhuysen MMAC. 1972. Antibodies against membrane antigens ofcytomegalovirus infected cells in sera of patients with a cytomegalovirus infection.Chin Exp Immunol 11:475-82.Theofilopoulos AN, Bokish VA, and Dixon FJ. 1974. Receptor for soluble C3 and C3bon human lymphoblastoid (RAJI) cells. Properties and biological significance. J ExpMed 139:696-711.Thieblemont N, Haeffner-Cavaillon N, Ledur A, L’age-Stehr J, Ziegler-HeitbrockHWL, Kazatchkine MD. 1993. CR1 (CD35) and CR3 (CD11b/CD18) mediateinfection of human monocytes and monocytic cell lines with complement-opsonizedFIIV independently of CD4. Clin Exp Immunol 92:106-12.Thomas JR. Dwek RA, and Rademacher TW. 1990. Structure, biosynthesis, andfunction of glycosylphosphatidylinositols. Biochem 29:5413-22.167Thomas DJ and Lublin DM. 1993. Identification of 5’-flanking regions affecting theexpression of the human decay accelerating factor gene and their role in tissue-specificexpression. J Immunol 150:151-60.Tomoda T, Kurashige T, and Taniguchi T. 1992. Stimulatory effect of interleukin-1 13on the interferon-gamma-dependent HLA-DR production. Immunology 75:15-19.Tsoukas CD and Lambris JD. 1988. Expression of CR2/EBV receptor on humanthymocytes detected by monoclonal antibodies. Eur I Immunol 18:1299-1302.Tsuchiya S. Kobayashi Y, Goto Y, Okumura H, Nakae S, Konno T, and Tada K. 1982.Induction of maturation in cultured human monocytic leukemia cells by a phorboldiester. Cancer Res. 42:1530-6.Tsuji S, Kaji K, and Nagasawa S. 1994. Decay-accelerating factor on human umbilicalvein cells. Its histamine-induced expression and spontaneous rapid shedding fromthe cell surface. J Immunol 152:1404-10.Tumilowicz JJ, Gawlik ME, Powell BB, and Trentin JJ. 1985. Replication ofcytomegalovirus in human arterial smooth muscle cells. J Virol. 56:839.Turpen T.H. and Griffith O.M. 1986. Rapid isolation of RNA by a guanidiumthiocyanate/cesium chloride gradient method. BioTechniques 4:11.Tysoe-Calnon VA, Grundy JE, and Perkins SJ. 1991. Molecular comparisons of theB2-microglobulin-binding site in class I MHC cc-chains and proteins of relatedsequences. Biochem J 277:359-69.168Van den Berg CW and Morgan BP. 1994. Complement-inhibiting activities of humanCD59 and analogues from rat, sheep, and pig are not homologously restricted. JImmunol 152:4095-101.VanDorp WT, Jonges E, Bruggeman CA, Daha MR, Van Es LA, and Van der WoudeFJ. 1989. Direct induction of MI-IC class I, but not class II expression on endothelialcells by cytomegalovirus. Transplantation 48:469.Vey E, Zhang J-H, and Dayer JM. 1992. IFN-y and 1,25 (OH)2D3induce on THP-1cells distinct patterns or cell surface antigen expression, cytokine production, andresponsiveness to contact with activated T cells. J Immunol. 149: 2040-6.Wade M, Kowalik TF, Mudryj M, Huang E-S, and Azizkhan JC. 1992. E2F mediateddihydrofolate reductase promoter activation and multiprotein complex formation inhuman cytomegalovirus infection. Mol Cell Biol 12:4364.Warren AP, Ducroq DH, Lehner PJ, and Borysiewicz. 1994. Human cytomegalovirusinfected cells have unstable assembly of major histocompatibility complex class Icomplexes and are resistant to lysis by cytotoxic T lymphocytes. J Virol 68:2822-29.Watts MJ, Dankert JR. and Morgan BP. 1990. Isolation and characterization of amembrane-attack-complex-inhibiting protein present in human serum and otherbiological fluids. Biochem J 265:471-7.Weinshenker BG, Wilton S. and Rice GPA. 1988. Phorbol ester-induceddifferentiation permits productive human cytomegalovirus infection in a monocyticcell line. J Immunol 140:1625-31.169Weis JJ, Tedder TF, and Fearon DT. 1984. Identification of a 145,000, Mr membraneprotein as the C3d receptor (CR2) of human B lymphocytes. Proc Nati Acad Sci USA81:881-5.Weis H and Fearon DT. 1985. The identification of N-linked oligosaccharides on thehuman CR2/ Epstein-Barr virus receptor and their function in receptor metabolism,plasma membrane expression, and 1.igand binding. J Biol Chem 260:13824-30.Weiss L, Okada N, Haeffner-Cabaillon N, Hattori T, Faucher C, Kazatchkine MD, andOkada H. 1992. Decreased expression of the membrane inhibitor of complement-mediated cytolysis CD59 on T-lymphocytes of HP/-infected patients. AIDS 6:379-85.Whaley K, and Ruddy S. 1976a. Modulation of the alternative complement pathwayby 1H globulin. J Exp Med 144:1147-63Whaley K, and Ruddy S. 1976b. Modulation of C3b hemolytic activity by a plasmaprotein distinct from C3b inactivator. Science 193:1011-3.Wiley CA, Schrier RD, Denaro FJ, Nelson JA, Lampert PW, and Oldstone MBA. 1986.Localization of cytomegalovirus proteins and genome during fulminant centralnervous system infection in an AIDS patient. J Neuropath Exp Neur 45:127-39.Wilson JG, Tedder TF, and Fearon DT. 1983. Characterization of human Tlymphocytes that express the C3b receptor. J Immunol 131:684-9.Wong WW, Chill JM, Rosen MD, Kennedy E, Bonaccio ET, Morris MJ, Wilson JG,Klickenstein LB and Fearon DT. 1989. Structure of the human CR1 gene: molecular170basis of the structural and quantitative polymorphism and identification of new CR1-like allele. J Exp Med 169:847.Xing Z, Jordana M, Braciak T, Ohtoshi T, and Gauldie J. 1993. LPS induces expressionof GM-CSF, IL-8, and IL-6 in human fibroblasts: evidence for heterogeneity within therespiratory tract. Am J Resp Cell Mol Biol 9:255-63.Yamashina M, Ueda, E, Kinoshita T, Takami T, qima A, Ono H, Tanaka H, Kondo N,Orii T, Okada N, Okada H, Inoue K, and Kitani T. 1990. Inherited completedeficiency of 20-kilodalton homologous restriction factor (CD59) as a cause ofparoxysmal nocturnal hemoglobinuria. N Engl J Med 323:1184-9.Yamashita Y, Shimokata K, Mizuno S, Yamaguchi H, and Nishiyama Y. 1993. Down-regulation of the surface expression of class I IVIHC antigens by humancytomegalovirus. Virology 193:727.Yoon SH, and Fearon DT. 1985. Characterization of a soluble form of the C3b/C4breceptor (CR1) in human plasma. J Immunol 134:3332-3.Yoshino K, Hashimoto M, Shinkai K. 1977. Studies on the neutralization of herpessimplex virus. VIII. Significance of viral sensitization for inactivation by complement.Microbiol Immunol 21:231-41.Young LS, Clark D, Sixbey JW, and Rickinson AB. 1986. Epstein-Barr virus receptorson human pharyngeal epithelia. Lancet 1:240-2.171Yu GH, Holers VM, Seya T, Ballard L, and Atkinson JP. 1986. Identification of a thirdcomponent of complement-binding glycoprotein of human platelets. J Clin Invest78:494.Zalman LS, Wood LM, and Muller-Eberhard HJ. 1986. Isolation of a humanerythrocyte membrane protein capable of inhibiting expression of homologouscomplement transmembrane channels. Proc Nati Acad Sci USA. 83:6975-9.Zalman LS, Wood LM, Frank MM, and Muller-Eberhard HI. 1987. Deficiency of thehomologous restriction factor in paroxysmal nocturnal hemoglobinuria. J Exp Med165:572-7.Zalman LS, Brothers MA, and Muller-Eberhard HJ. 1989. Isolation of homologousrestriction factor from human urine. Immunological properties and biologic activity.J Immunol 143:1943-7.Ziccardi RJ, Dahlback B, and Muller-Eberhard HI. 1984. Characterization of theinteraction of human C4b-binding protein with physiological ligands. J Biol Chem259:13674-79.172


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