UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The development of the Pseudomonas aeruginosa outer membrane protein OprF as a presentation vector for… Wong, Rebecca Suk Yi 1995

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

Item Metadata


831-ubc_1995-060845.pdf [ 3.8MB ]
JSON: 831-1.0088315.json
JSON-LD: 831-1.0088315-ld.json
RDF/XML (Pretty): 831-1.0088315-rdf.xml
RDF/JSON: 831-1.0088315-rdf.json
Turtle: 831-1.0088315-turtle.txt
N-Triples: 831-1.0088315-rdf-ntriples.txt
Original Record: 831-1.0088315-source.json
Full Text

Full Text

THE DEVELOPMENT OF THE FSEUDOMONASAERUGINOSA OUTERMEMBRANE PROTEIN OPRF AS A PRESENTATION VECTOR FORFOREIGN ANTIGENIC DETERMINANTSbyREBECCA SUK Yl WONGB.Sc. (Co-op), Simon Fraser University, 1990A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Microbiology and Immunology)We accept this thesis as conformingto tJ required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember 1995© Rebecca Suk Yi Wong, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)__________________________Department of c1The University of British ColumbiaVancouver, CanadaDate ;72t ;ft/ (7’7DE-6 (2/88)11ABSTRACTA variety of systems have been developed to improve the presentation offoreign antigenic determinants (‘epitopes’) by inserting them in the context ofcarrier proteins. The goals of this study were to develop the Pseudomonasaeruginosa outer membrane protein OprF as a carrier for foreign epitopes and tostudy the effect of the mode of presentation on the antigenicity of the presentedepitope. The model epitope used in this study was the 4-amino acid repeatingepitope (NANP) of the circumsporozoite protein of the malaria parasite,Plasmodium falciparum. Linker-insertion mutagenesis was carried out to create11 “permissive” sites which allowed the insertion of 4 extra amino acids. Two seriesof OprF::malarial epitope hybrid proteins, the positional hybrids and the multiple-repeat hybrids, were constructed by inserting oligonucleotides encoding the epitopeinto the linker-insertion sites of oprF. The effects of the insertion position and thelength of the epitope on its antigenicity were studied by ELISA using outermembranes and by whole cell dot blot analysis. It was shown that the antigenicityof the epitope varied when inserted at different positions of OprF, while it increasedwith the length of the epitope at two of the three insertion positions studied. Thesedata were employed to revise the membrane topology model of OprF and haveimproved our understanding of the epitopes recognized by the OprF-specificmonoclonal antibodies. Generalizations about the influence of surrounding aminoacids on the antigenicity of the inserted epitope are proposed. A targeted study of111immunogenicity showed that a 19-amino acid malarial epitope was significantlymore immunogenic than a 7-amino acid epitope when inserted at an N-terminalinsertion site of OprF. A parallel immunogenicity study of two versions ofglutathione S-transferase (GST) : :malarial epitope fusion proteins demonstratedthat neither an 11- nor a 19- amino acid epitope fused to the C-terminus of GSTwas immunogenic. This study demonstrated for the first time that OprF can beused as a carrier to generate and detect anti-epitope antibodies in immunizedanimals and in immunoassays respectively.ivTABLE OF CONTENTSABSTRACT iiTABLE OF CONTENTS ivLIST OF FIGURES xLIST OF TABLES xiiiLIST OF ABBREVIATIONS xvACKNOWLEDGEMENTS xviiINTRODUCTION 11. Epitope presentation systems 11.1 Introduction 11.2 E. coli outer membrane protein presentation systems 11.3 Other presentation systems 52. Applications of epitope presentation systems 72.1 Vaccines 72.2 Immunopurification 82.3 Detection and production of anti-peptide antibothes 92.4 Construction of random libraries 93. The carrier protein: OprF 103.1 Bacterial outer membrane 103.2 General structures of porins 113.3 OprF 12V4. The model epitope: the malarial epitope 174.1 Life cycle of the malaria causative agent 17Plasmoclium falciparum 174.2 The tetrapeptide repeating epitope NANP 175. Factors affecting antigenicity and immunogenicity 205.1 Factors affecting antigenicity 205.2 Factors affecting immunogenicity 226. Aims of this study 24METHODS AND MATERIALS 251. Bacterial strains, plasmids and media 252. General recombinant DNA techniques 253. General protein and immunological techniques 283.1 SDS-PAGE and immunoblottings 283.2 Antibodies 293.3 Indirect immunofluorescence labeffing 293.4 Trypsin sensitivity assays 313.5 Protein assays 324. Construction of pRW3 325. Linker-insertion mutagenesis of oprF 325.1 Mutagenesis with kanamycin resistance cassette 325.2 Mutagenesis at the Sail site 365.3 Determination of linker-insertion sites 36vi6. Construction of OprF::malarial epitope hybrid proteins. 366.1 Positional hybrids 366.2 Multiple-repeat hybrids 397. DNA sequencing 408. Construction of glutathione S-transferase (GST)::malarial epitope fusionproteins 429. Isolation of outer membranes 429.1 Triton X-100 extraction 429.2 Sucrose gradient centrifugation 429.3 Removal of inclusion bodies 4410. Expression of oprF and oprF derivatives in E. coli 4510.1 Expression of oprFin different E. coli strains 4510.2 Expression of an oprF derivative in different inductionconditions 4511. Protein purification 4611.1 OprF::malanal epitope hybrid proteins 4611.2 GST::malarial epitope fusion proteins 4611.3 Extraction from SDS-polyacrylamide gel 4712. Antigenicity studies 4712.1 Outer membrane ELISA 4712.2 Whole cell dot blot analysis 4912.3 Statistical analyses 49vii13. Immunization studies 5013.1 Immunization with OprF::MElOaa2l5 andOpF 5013.2 Immunization with OprF::MEaa26 multiple-repeat hybridsand GST::malarial epitope fusion proteins 5013.3 Determination of antibody titers 5013.4 Characterization of antisera by Western immunoblot analysis 51RESULTSChapter one: Construction and characterization of OprF linker mutants 531.1 Introduction 531.2 Expression of oprF in E. coli 541.2.1 Construction of pRW3 541.2.2 Expression of oprFin different E. coli host strains 551.3 Semi-random linker mutagenesis with a kanamycin cassette 581.4 Site-directed mutagenesis at the Sail site 591.5 Determination of insertion sites 591.6 Expression and cellular localization of linker mutants 611.7 Monoclonal antibody reactivities of linker mutants 671.8 Membrane configuration of linker mutants in E. coli 691.8.1 Trypsin sensitivity assays 691.8.2 Immunofluorescence labeffing 731.9 Summary 75viiiChapter two: Construction, characterization and purification of OprF::malarial epitope and GST::malarial epitope hybrid proteins 772.1 Introduction 772.2 Construction of OprF::malarial epitope hybrid proteins 782.2.1 Positional hybrids 782.2.2 Multiple-repeat hybrids 802.3 Characterization of OprF::malarial epitope hybrid proteins 822.3.1 Expression of hybrid proteins 822.3.2 Cellular localization of hybrid proteins 842.3.3 Surface exposure of the epitope 882.3.4 Monoclonal antibody reactivity of hybrid proteins 912.4 Purification of OprF::malarial epitope hybrid proteins 932.4.1 Induction experiments 932.4.2 Detergent extractions 952.4.3 FPLC purification 982.4.4 Purification of inclusion body-contaminated outermembrane preparations 982.5 GST::malarial epitope fusion proteins 1012.5.1 Construction and purification of fusion proteins 1012.5.2 Binding of fusion proteins with epitope-specificmonoclonal antibodies 1032.6 Summary 103ixChapter three: Study of the effect of mode of presentation on antigenicityand immunogenicity 1063.1 Introduction 1063.2 Antigenicity study 1073.2.1 Approaches 1073.2.2 Position effect 1083.2.3 Length effect 1133.3 Immunogenicity study 1193.3.1 Immunogenicity of OprF::ME lOaa2 15 1213.3.2 Immunogenicity of OprF::ME7aa26 and OprF::MEl9aa26.. 1243.3.3 Immunogenicity of GST::ME11 and GST::ME19 1283.4 Summary 131DISCUSSION 134General 134Linker-insertion mutagenesis 135Effects of amino acid insertions in OprF 136Membrane topology of OprF 140Binding epitopes of OprF-specific monoclonal antibodies 144Antigenicity and mode of presentation 147Immunogenicity 151REFERENCES 159xLIST OF FIGURESFigure 1. Schematic diagram of the p-barrel structure of a porin 13Figure 2. Proposed membrane topolopy model of OprF 15Figure 3. The life cycle of Plasmodium falciparum 18Figure 4. Stereo drawings of two of the predicted structures of the(NANP)6peptide 21Figure 5. Construction of pRW3 33Figure 6. Schematic representation of semi-random linker-mutagenesiswith a kanamycin resistance cassette 34Figure 7. Nucleotide and encoded amino acid sequences of theoligonucleotides used for the construction of OprF::malarialepitope hybrid protein 37Figure 8. Construction of GST::malarial epitope fusion proteins 43Figure 9. Expression of oprF in different E. coli host strains 56Figure 10. Restriction mapping of linker-insertion sites 60Figure 11. Cellular localization of OprF linker mutants 65Figure 12. Expression of OprF linker mutants 66Figure 13. Trypsin sensitivity of linker mutants in outer membranes 71Figure 14. Expression of OprF::malarial epitope positional hybrids 83Figure 15. Cellular localization of OprF::malarial epitopepositional hybrids 85xiFigure 16. Expression of OprF::malarial epitope multiple-repeat hybrids.. 87Figure 17. Presence of inclusion bodies in outer membrane samples 89Figure 18. Surface exposure of the malarial epitope 90Figure 19. Western immunoblots of OprF::malarial epitope multiple-repeathybrids 94Figure 20. Expression of an oprF derivative in different inductionconditions 96Figure 21. Purification of OprF::malarial epitope hybrid proteins 97Figure 22. FPLC profile of a MonoQ column separation of the octylPOE/EDTA soluble OprF hybrid expressed by pRW3O7. 1M 99Figure 23. Removal of inclusion bodies from outer membranepreparations by octyl-POE extraction 100Figure 24. Purification of GST::malarial epitope fusion proteins 102Figure 25. Binding of GST::malarial epitope fusion proteins withepitope-specffic monoclonal antibodies 104Figure 26. Binding of an OprF-specffic polyclonal serum and the malarialepitope-specific mAb pf2A.10 with OprF and OprF::malarialepitope hybrid 109Figure 27. Effect of insertion position on the antigenicity of the malarialepitope 111Figure 28. Effect of insertion of multiple copies of the malarial epitope onantigenicity at insertion sites aa26 and aa213 of OprF 114xiiFigure 29. Effect of the length of the epitope on its antigenicity at insertionsite aa26 of OprF 116Figure 30. Effect of the length of the epitope on its antigenicity at insertionsite aa’96 of OprF 117Figure 31. Effect of the length of the epitope on its antigenicity at insertionsite aa213 of OprF 118Figure 32. ELISA titrations of anti-OprF and anti-malarial epitoperesponses induced in BALBIc mice immunized with OprFand OprF::MElOaa2l5 by ELISA 122Figure 33. ELISA titrations of anti-OprF and anti-malarial epitoperesponses induced in C57BL/6J mice immunized withOprF::ME7aa26 and OprF::MEl9aa26 by ELISA 125Figure 34. Western immunoblot analysis of the sera from miceimmunized with OprF::ME7aa26 and OprF::MEl9aa26 127Figure 35. ELISA titrations of anti-GST and anti-malarial epitoperesponses induced in C57BL/6J mice immunized withGST::ME11 and GST::ME19 129Figure 36. Western immunoblot analysis of the sera from miceimmunized with GST::ME11 and GST::ME19 130Figure 37. Proposed membrane topology model of OprF 143xliiLIST OF TABLESTable I. Examples of epitope presentation systems 2Table II. Bacterial strains and plasmids 26Table III. OprF epitopes recognized by monoclonal antibodies 30Table IV. Summary of insertion sites of 11 linker-insertion mutantsand one site-directed insertion mutant 62Table V. Summary of six of the deletion mutants isolated during linker-insertion mutagenesis 63Table VI. Summary of monoclonal antibody reactivity of linker mutants 68Table VII. Summary of trypsin sensitivity assays of linker mutantsin E. coli outer membranes, DH5ci and C386 whole cells 72Table VIII. Results from indirect immunofluorescence labelling of F. coliC386 cells expressing OprF linker mutants 74Table IX. Summary of OprF::malarial epitope positional hybrids 79Table X. Summary of OprF::malarial epitope multiple-repeat hybrids.. 81Table XI. Summary of monoclonal antibody reactivity of OprF::malarialepitope positional hybrids 92Table XII. Summary of antigenicity of the malarial epitope inOprF::mai.arial epitope positional hybrids 112Table XIII. Summary of antigenicity of the malarial epitope inOprF::malarial epitope multiple-repeat hybrids 120xivTable XIV. Summary of antibody responses induced in mice immunizedwith wild type OprF or OprF::MElOaa2l5 123Table XV. Summary of antibody responses induced in mice immunizedwith OprF::ME7aa26 and OprF::MEl9aa26 126Table XVI. Summary of antibody responses induced in mice immunizedwith GST::ME11 and GST::ME19 132Table XVII. Predicted primary and secondary structures atthe insertion sites 149xvLIST OF ABBREVIATIONSA optical densityaa amino acid position ‘n’amp ampicillinbp base pairBSA bovine serum albuminCSP circumsporozoite proteinDEAE diethylaminoethylEDTA ethylenediamine tetraacetic acidELISA enzyme-linked immunosorbent assayFCS fetal calf serumFMDV foot-and-mouth disease virusFPLC fast protein liquid chromatographyGST glutathione S-transferaseh hour(s)HIV- 1 human immunodeficiency virus-iIgG immunoglobulin GIPTG isopropyl thio- p-D-galactopanosidekb kilobase pairkDa kilodaltonKm kanamycinxviLPS lipopolysaccharidesMSP merozoite surface proteinmAb monoclonal antibodymm minute(s)Octyl-POE octyl-polyoxyetheleneOprF P. aeruginosa major outer membrane protein FoprF gene encoding OprFPBS phosphate-buffered saline (0.14 M NaC1/2.7mM KC1/1.47mMKH2PO4/2OmM NaHPO4pH7.4)PVDF polyvinylidene difluorides second(s)SDS sodium dodecyl sulfateSDS-PAGE SDS polyacrylamide gel electrophoresisTGEV transmissible gastroenteritis coronaviruswt wild typexviiACKNOWLEDGEMENTSI am grateful to my supervisor Dr. Robert E.W. Hancock for his guidanceand support in this project. I would like to thank the members of my supervisorycommittee, Drs. Finlay, Kilburn and Spiegelinan, for their time and advice. Specialthanks to members of the Department, especially members of the Hancock lab, fortheir technical assistance, moral support and friendship during my study. Last butnot least, I would like to acknowledge my friends and family for their warmth,encouragement and understanding.The financial support of the Natural Sciences and Engineering ResearchCouncil and the Medical Research Council of Canada is also gratefullyacknowledged.1INTRODUCTION1. Epitope presentation systems1.1 IntroductionProgress in molecular biology has allowed the engineering of heterologousproteins that carry components from two or more different host proteins. A varietyof systems have been developed to improve the presentation of foreign antigenicdeterminants (‘epitopes’) by inserting them in the context of carrier proteins.Carrier proteins that have been utilized in these systems include bacterial outermembrane proteins, subunits of bacterial cellular appendages such as pili andflagella, bacterial secreted proteins, ifiamentous phage surface structural proteins,and viral surface coat proteins. Examples of these epitope presentation systems arelisted in Table I. In general, the incorporation of passenger epitopes intoappropriate sites in these carrier proteins does not seriously interrupt the structureand function of the carriers. As a result, the passenger epitopes are usuallytargeted to the same cellular compartment as the carrier proteins.1.2 E. coli outer membrane nrotein yresentation systemsIt has been found that certain regions, identified as “permissive sites”, ofouter membrane proteins are flexible enough to accommodate extra amino acidsequences without affecting the biogenesis, folding and localization of these2Table I. Examples of epitope presentation systemsCarriers Epitopes ReferencesE. coli outer membrane proteinsOmpAFhuATraTHepatitis B virus pre S2epitopePoliovirus C3 epitopeHIV-1 GP11O epitopeHIV- 1 V3 loopChiamydia MOMP epitopeVR 1 and VR2 of class 1 OMPof N. meningitidisRandom peptidesFoot-and-mouth diseasevirus (FMDV) VP1 epitopeMycobacterial T-cell epitopeMalarial antigen fragmentsAntibody fragmentPoliovirus C3 epitopePoliovirus C3 epitopeCharbit et al., 1987van der Werfet al., 1990Charbit et al., 1990Charbit et al., 1993Hayes et al., 1991McCarvil et al., 1993Brown, 1992Agterberg et al.,1990aJanssen et al., 1994aSchorr et al., 1991Francisco et al., 1993Moeck et al., 1994Taylor et al., 1990Non-outer membrane proteinsE. coli pgalactosidaseE. coli MalEFMDV VP 1 epitopeRandom peptidesPoliovirus C3 epitopeHIV- 1 V3 loonBroekhuijsen et al., 1986Lenstra et al., 1992Leclerc et al., 1990Charbit et al., 1993To be continuedLamBPhoE3Table I. Examples of epitope presentation systems (continued)Non-outer membrane uroteins (continued)E. coli C1pG TGEV spike protein Sprepilin epitopeE. coli Pap pili IgG binding domain ofprotein AE. coli P- FMDV VP1 epitopeffinbriaeE. coli Type 1 Hepatitis B surface antigenfimbriae epitopeFMDV VP1 epitopePoliovirus C3 epitopeCholera toxin epitopeHepatitis B surface antigenepitopeInfluenza haemagglutininepitopeP. aeruginosa OprF andOprI fragmentsMalaria MSP fragmentsFilamentous nhae coat nroteinspill Malaria CSP repeating Cruz et at., 1988epitopeAntibody variable domains McCafferty et al., 1990Antibody Fab fragments Barbas et at., 1991HIV-l Gag p24 Tsunetsugu et at., 1991pVIII HIV-l p17 epitope Minenkova et at., 1993To be continuedSalmonellaflagellinDer Vartanian et at.,1994Steidler et at., 1993Van Die et at., 1990Hedegaard et at., 1989Hedegaard et at., 1989Hedegaard et at., 1989Newton et al., 1989Wu et at., 1989McEwen et at., 1992von Specht et at., 1995Ling et at., 1994Glutathione Stransferase4Table I. Examples of epitope presentation systems (continued)Gram-nositive systemM6 protein E7 protein of human Pozzi et al., 1992(S. pyogenes) papifiomavirusProtein A Malaria blood-stage antigen Hansson et al., 1992(S. aureus) Streptococcal albumin Hansson et al., 1992binding receptorViral uroteinsInfluenza A Chloramphenicol Percy et al., 1994virus acetyltransferaseneuraminidaseHepatitis B HIV-1 antigenic determinant Michel et al., 1993surface antigenMalaria CSP epitope Rutgers et al., 1988Adenovirus Poliovirus VP1 capsid Crompton et al., 1994hexon protein epitopeCowpea mosaic FMDV VP1 epitope Usha et al., 1993virusHuman HIV-1 V3 loop Smith et al., 1994rhinovirus 14Poliovirus VP1 Human papifiomavirus Jenkins et al., 1990protein epitopeHIV-1 gp4l epitope Evans et al., 19895proteins. Foreign epitopes that are inserted in the “permissive” surface-exposedloop regions of these outer membrane proteins have been shown to be detectable onthe cell surface. The system using the Escherichia coli outer membrane proteinLamB as the carrier is one of the most developed. LamB is the porin responsiblefor maltose uptake and the receptor for A phage (Szmelcman and Hofnung, 1975).The poliovirus C3 epitope has been used as a passenger epitope to identify 11permissive sites in LamB (Charbit et al., 1991). Other foreign antigenicdeterminants such as the hepatitis B virus preS2, HIV-1 gp 110 or V3 loop epitopes,or a Chiamydia Major Outer Membrane Protein (MOMP) epitope have also beeninserted into selected permissive sites of this carrier protein (Charbit et al., 1987;Charbit et al., 1990; Hayes et al., 1991). The phosphate-starvation-inducible porinPhoE of E. coli, has been used as a carrier to present the foot-and-mouth diseasevirus (FMDV) VP1 epitope and a Mycobacterial T cell epitope (Agterberg etal.,1991a; Janssen et al., 1994a). Likewise, the outer membrane protein OmpA hasbeen employed as a carrier to present malarial antigen fragments on the surface ofa Salmonella vaccine strain (Schorr et al., 1991). In most of these studies, theforeign epitopes in the context of the carrier proteins have been shown to beimmunogenic in test animals.1.3 Other oresentation systemsSubunits of bacterial cellular appendages such as flagella and pugenerally contain variable regions that allow the insertion of foreign amino acid6sequences. A number of viral epitopes have been inserted into the subunits of theseappendages and are found to be incorporated into the corresponding structures.Epitopes have also been fused to the periplasmic proteins MalE and j3-galactosidase, where the immunogenicity of the inserted epitopes has been reported(Leclerc et al., 1990; Charbit et al., 1993; Broekhuijsen et al., 1986). In addition tothese E. coli proteins, the Salmonella flagellin has been used to express epitopesfrom Cholera toxin, hepatitis B surface antigen and influenza haemagglutinin(Newton et al., 1989; Wu et al., 1989; McEwen et al., 1992). The vaccine potentialof an attenuated vaccine strain of Salmonella expressing a recombinant flageffinhas been demonstrated.A phage display system has also been developed to express foreign geneticinformation in the context of the bacteriophage structural surface proteins. Boththe major coat protein pVIII and the minor coat protein p111 of the filamentousphage have been used to display various foreign epitopes (see Scott and Craig, 1994for review). Furthermore, the potential of gram-positive bacterial surface proteinsas carriers to present foreign epitopes on the surface of gram-positive bacteria hasbeen investigated (Pozzi et al., 1992; Hansson et al., 1992). Another major categoryof these epitope presentation systems involves the use of viral coat proteins ascarriers (Pable I). A similar repertoire of foreign epitopes has been inserted into theantigenic regions of these coat proteins and the vaccine potentials of some of thesesystems have been studied.The system using glutathione S-transferase (GST) as the carrier protein7deserves a special mention because it was also used in this study. The proteincarrier was originally identified in Schistosomajaponicum (Smith et al., 1986) andcan be expressed as an active, soluble protein in E. coli (Smith et al., 1988). Thisprotein is a commonly used affinity tag for the purification of fusion proteins. Theaffinity of GST for reduced glutathione allows the purification of soluble GST fusionproteins by adsorption to glutathione beads and subsequent desorption using freereduced glutathione (Smith and Johnson, 1988). Due to the ease of purification,GST has also been used as a carrier to induce immune response against smallpeptides or antigenic fragments (e.g., Ling et al., 1994; von Specht et al., 1995).2. Applications of epitope presentation systems2.1 VaccinesRecombinant live bacterial vaccines consist of attenuated strains ofenteric bacteria expressing heterologous peptides derived from pathogenic agents.It has been reported that when intact cells are used as immunogens, the insertedepitope must protrude sufficiently from the outer membrane to stimulate anantibody response (Leclerc et al., 1991). Therefore, the insertion of the peptidewithin the surface-exposed loop of an outer membrane protein carrier is likely tofacilitate the immunogenicity of the peptide. Moreover, the surface exposure of theinserted epitope may be advantageous because most of the strongly antigenicregions of outer membrane proteins reside in the surface-exposed loops; hence, the8location of the inserted epitope in these regions may enhance its interaction withB cells. Furthermore, the association of the peptide with surface moieties such aslipopolysaccharides (LPS) may provide an adjuvant effect to promote theimmunogenicity of the peptide. The attenuated strains of Salmonella or E. coli cancolonize the intestinal tract without causing infection to the host, and hence canprovide a refuge for the recombinant protein so that it can persist to elicit a morelasting immune response. In these situations, the surface exposure of the insertedepitope may be beneficial for the targeting of the epitope to the gut-associatedlymphoid tissues.2.2 ImmunopurificationThe affinity purification of antigens or antibodies usually requires one ofthe ligands to be in an immobilized form. Traditionally, this involves the largescale purification of these molecules, followed by covalent linkage of the proteinsto a solid-phase matrix. If the binding epitopes in these protein antigens have beenidentified, these epitopes can be inserted into an outer membrane protein carrierand expressed on the cell surface of bacteria. The resultant recombinant bacteriathus represent a source of readfly available whole cell affinity adsorbent. The useof such a system can not only circumvent the necessity for large scale purificationof the protein and the subsequent chemical linkage of the protein to a solid matrix,and hence provide a more time and cost efficient alternative for the preparation ofreagents for affinity purification.92.3 Detection and production of anti-peptide antibodiesSome diseases can be diagnosed by the presence of specific antibodies inthe patient’s serum. These antibodies are often directed against peptide antigensassociated with the pathogens. If the peptide(s) that are reactive with theseantisera are identified, then oligonucleotides encoding the peptide(s) can begenetically inserted into the DNA sequences of the carrier protein so as to expressthe peptide in the context of the carrier protein. Since chemically-synthesizedpeptides often do not bind to antibodies efficiently by themselves, this method islikely to improve the presentation of the peptide(s) for interaction with therespective antibodies. On the other hand, the peptide/carrier hybrid can also beused as an immunogen to raise anti-peptide antibodies. In this case the carrierprotein is likely to provide a T cell epitope which is required for an effectiveantibody response, thus circumventing the need to synthesize the peptidechemically and then link it to a carrier protein. Combining these two applications,it has been demonstrated that the LamB and MalE epitope presentation systemscan be used in a complementary fashion to induce and detect anti-peptideantibodies without the use of chemically synthesized peptides (Martineau et al.,1991).2.4 Construction of random librariesAnother application of epitope presentation systems is for theconstruction of random peptide libraries. These libraries can be used for the10identification of epitopes or ‘mimotope& (antigenic sequences that mimic epitopes)that bind to specific antibodies. Random libraries using phage pill protein, pgalactosidase and rhinovirus as carriers have been constructed. These libraries,expressing 6- to 15- residue random peptides encoded by the inserted degenerateoligonucleotides, have been used to successfully identify linear and conformationalepitopes recognized by specific antibodies (Stoute et al., 1995; Lenstra et al., 1992;Smith et al., 1994). Likewise, random peptide libraries can also be used to identifysubstrate binding or adhesion motifs. For example, the screening of a LamBrandom peptide library has led to the successful identification of an iron-oxideadhesion domain (Brown, 1992). On the other hand, these vectors can also be usedto express combinatorial antibody libraries which have shown potential for theidentification of useful antigen binding domains (McCafferty et al., 1990; Barbaset al., 1991; Francisco et al., 1993). The number of applications of the randomlibrary approach is likely to increase upon further study. Future advances in thisarea may lead to powerful applications in the area of drug design and to thedevelopment of diagnostic markers and vaccines.3. The carrier protein: OprF3.1 Bacterial outer membraneThe cell envelope of gram-negative bacteria is composed of an inner orcytoplasmic membrane and an outer membrane, separated by the peptidoglycan11layer and periplasm. Unlike the cytoplasmic membrane, the outer membrane is anasymmetric bilayer containing lipopolysaccharides on the outer leaflet andphospholipids or lipids in the inner monolayer. The outer membrane of gram-negative bacteria represents the primary barrier between the cell and itsenvironment. Proteins associated with or embedded in the outer membraneperform a variety of cellular functions, including nutrient uptake, receptor activityand the maintenance of structural integrity. In addition, the outer membrane alsofunctions as a permeability barrier to exclude the entry of harmful substances suchas destructive enzymes and detergents, and it limits the passage of antibiotics. Thehigh level of antibiotic resistance in P. aeruginosa, for instance, is attributed in partto the low permeability of the outer membrane (Nakae, 1995).3.2 General structures of normsPorins are outer membrane proteins that form transmembrane water-filled channels which allow the uptake of small hydrophilic molecules. Accordingto their substrate specificity, porins can be divided into two functional classes.General or non-specific porins such as OmpF and OmpC in E. coli allow the generaldiffusion of water-soluble molecules smaller than the exclusion limit of thechannels (Nikaido and Vaara, 1987). Specific porins such as the P. aeruginosaOprP and E. coli LamB proteins have specffic binding sites for the uptake ofphosphate ions and maltose respectively (Hancock and Benz, 1986; Szmelcman andHofnung, 1975). Unlike other membrane proteins which have membrane-spanning12cc-helical segments, porins have short amphipathic stretches of residues thattraverse the outer membrane in p-sheet structure. To date, crystal structures offour porins have been resolved by X-ray diffraction. These porins include theRhodobacter capsulatus porin, the E. coli general porin OmpF, the phosphate-starvation-inducible porin PhoE and the maltoporin LamB (Weiss et al., 1991;Cowan et al., 1992; Schirmer et al., 1995). All of these established structuresindicate that these proteins have p-barrel structures comprised of transmembraneanti-parallel p-sheet segments of 7-14 amino acids in length. The p-strands foldback and form a barrel structure which constitutes the framework of the channel.The structures of the B. capsulatus porin, OmpF and PhoE contain 16 p-strandswhile that of LamB contains 18 p-strands. The neighbouring strands are joined bylong cell surface loops and small periplasmic turns (Cowan et al. 1992; Fig. 1). Acommon feature of these porins is that at least one of the surface loops folds backinto the centre of the channel to form an eyelet region that constricts the size of thechannel.3.3 OvrFOprF is the major outer membrane protein ofF. aeruginosa and is presentin about 2x105 copies per cell (Angus et al., 1982). The protein is 325 amino acidsin length and has an apparent molecular mass of 35 kDa. Circular clichroism datarevealed that OprF has 62% a-sheet structure, which is consistent with thepredominance of n-sheet structure in the other outer membrane proteins (Siehnel13L6L5 L8v” \‘ \\\\\J/ L? j\L3/5rwAc IT3L)QT2rTiFigure 1, Schematic diagram of the -barre1 structure of a porin. Thickarrows represent transmembrane n-strands; L, surface loops; T,periplasmic turns; N, N-terminus; C, 0-terminus. Reproducedwith permission from Cowan et aL, 1992.14et al., 1990). Heat treatment of OprF increases the apparent molecular mass of theprotein as monitored by SDS-PAGE, indicating the presence of a compact p-sheetstructure (Siehnel et al., 1990). The amino acid sequence of OprF contains fourcysteine residues, and the mobility of the protein on SDS-PAGE is modifiable bytreatment with 2-mercaptoethanol (Hancock and Carey, 1979). These findingssuggested that the cysteine residues are involved in disulphide bond formation.A number of approaches have been used to study and predict themembrane topology of OprF. These approaches include the prediction of secondarystructures by computer programs, the comparison of amino acid sequences withother outer membrane proteins, TnFhOA mutagenesis, oprF gene deletion analysis,and linker- and epitope-insertion mutagenesis (described in this study). The datagenerated from these studies have proven to be of use in confirming and refiningthe working model of OprF membrane topology. One of the most recent topologymodels of OprF is shown in Figure 2.The primary amino acid sequence of OprF is distinct from the classicaltrimeric porins such as the E. coli OmpF, OmpC and PhoE porins (Duchene et al.,1988). However, OprF shares strong C-terminal homology with the OmpA-relatedouter membrane proteins (Duchene et al., 1988; Woodruff and Hancock 1989) aswell as several proteins from unrelated species including the B. sub tilis MotA (DeMot and Vanderleyden, 1994). The highly conserved C-terminal motif in theseproteins has been proposed to share a common role in peptidoglycan association (DeMot and Vanderleyden, 1994).15‘V..NT.A V vP-T. T ET CETC S 0 Q-KP N s HS 0 ,N-DV.’.G T TY N N RN DGD S DT K Q K GS ‘,DP p 5RT G K N G E Hb-S.PF1 V RK 0 R 1•fl T R r1 c ri C E Cri F1 I HI I Q ru Q F; I A I I I TC v 011G1 N Q I GIl H l NI Q A Y I E I oh E’ iK I D I l r—i I RjF [l1, H T IQIIWIIAIIVIIMIA Y IVIIGIIAEIIi N I] E’ I TI A M 161 I I V1 I VI IF I N’ I ElI Al I Al VEI1Li’ Gil SI L A I DI’ Al I EI I RvI AD I çi I vii TAI II L I A’ S G I A I L I A1 I L’ 1N I Li Vi I N Ii A II A I ISI:; : ‘y’l LI C N ILl CL I A v AG I AK vi I °I I K K DR I E A I L.!IC W R K W N K I Di R N R PI Fl IA IE VA s L_XiE1 H P L I F’ LfJ K I D IC u C L Y F Lis RLiS D T C Y NI T pV F E GG 0 NT K ECL‘(F S KKVFigure 2. Proposed membrane topolopy model of OprF. The top of this modelis proposed to face the exterior of the cell. The transmembrane pstrands are indicated by rectangular boxes. Reproduced withpermission from Siehnel et at. (1990).16OprF serves a structural role in maintaining the cell shape and cell wallintegrity (Gotoh et al., 1989; Woodruff and Hancock, 1989). It has beendemonstrated that OprF can complement the cell shape defect in an E. coli OmpAdeficient mutant (Woodruff and Hancock, 1989). It has also been reported thatOprF is required for the growth of P. aeruginosa in low osmolarity medium (Nicasand Hancock, 1983). In addition to its structural role, OprF also functions as aporm which was proposed to be responsible for the molecular mass exclusion limitof 3000 to 9000 daltons through the outer membrane ofF. aeruginosa (Hancock andNikaido, 1978; Hancock et al., 1979). However, other model membrane studiesreported that OprF only forms small channels and that the exclusion limit of theP. aeruginosa outer membrane is too low for the penetration of disaccharides (342daltons) (Caulcott et al., 1984, Yoneyama and Nakae, 1986). More recent studieshave shown quite conclusively that OprF forms channels that allow the diffusionof substrates of at least the size of tetrasaccharides in both model membrane(Nikaido et al., 1991) and intact cell (Beffido et al., 1992) experiments. One possibleexplanation to this paradox is that OprF forms channels of two different pore sizes(Woodruff et al., 1986). In vivo, only a small portion of OprF form the large pores,and hence contributes to the low outer membrane permeability of P. aeruginosa.Purified OprF has been shown to be a B cell mitogen (Chen et al., 1980).Immunizing animals with purified or partially purified OprF preparations protectsthe animals from subsequent challenge with P. aeruginosa in various models.These findings suggested that OprF is immunogenic and a potential candidate for17a P. aeruginosa vaccine (Gilleland et al., 1984; Matthews-Greer and Gilleland,1987; Gifieland, et aL, 1988).4. The model epitope: the malarial epitope4.1 The life cycle of the malaria causative aaent: Plasmodium falcivarumMalaria is a parasitic disease that afflicts hundreds of millions of peoplein a broad tropical band around the world. The disease is spread by Anophelesmosquitos infected with the protozoan parasite Plasmodium falciparum. Themalaria parasite has a complex life cycle involving intracellular and extracellularstages in both the human host and the mosquito vector. The stage that infectsman, the sporozoite, is present in the salivary gland of the mosquito and injectedinto the victim’s bloodstream when the mosquito takes a blood meal. The sporozoitethen finds its way to a liver cell, where it undergoes a series of transformations andis released into the blood stream of the victim as merozoite, which is the blood stageof the parasite. Each merozoite then invades a red blood cell and multiplies. Somemerozoites become male and female gametocytes, which are then taken up by amosquito. After further transformations, sporozoites appear in the mosquito’ssalivary gland and are ready for another infective cycle (Fig. 3).4.2 The tetraieptide reneatin etitone NANPThe circumsporozoite protein (CSP) is the major surface antigen of the18zoitesOocystQokineteZygote) o( BloodGameteIn mosquitogutFigure 3. The life cycle of Plasinodium falciparum. Reproduced withpermission from Fig. 83-3, p. 1037. Medical microbiology. S. Baron(ed.), Churchifi Livingstone Publishing Co..19sporozoite stage of the parasite. The central portion of the protein contains 37tandem repeats of the tetrapeptide Asn-Ala-Asn-Pro (NANP), with 4 interspersedAsn-Val-Asp-Pro (NVDP) variants (Dame et aL, 1984). It has been reported thatthe repetitive sequence encompasses the immunodominant region of the proteinand antibodies raised against this region are potent inhibitors of invasion and ofthe development of sporozoites in cultured hepatocytes (Young et al., 1985; Zavalaet aL, 1985). Thus, the central tandem repeat region of CSP has engenderedconsiderable interest as a potential candidate for a malaria vaccine. Both the wholeCSP and the tandem repeat portion of the protein have been presented by carrierproteins such as the surface antigen of hepatitis B virus (Rutgers et al., 1988) andthe pITT protein of ifiamentous phage (Cruz et al., 1988). It has been documentedthat in mice the repeat (NANP)>2can only be recognized as a T cell epitope byanimals with a H2b background (Good et al., 1986). In view of this, identified T cellepitopes have been used in conjunction with the NANP repeating epitope in orderto stimulate an anti-NANP response (Good et al., 1987). The incorporation of thesynthetic NANP peptides as multiple antigen peptide (MAP) has also been shownto be able to overcome the genetic restriction in non-responsive animals (Pessi etal., 1991; Carvo-Calle et al., 1993).Due to the intense medical interest in the use of CSP as a component ina malaria vaccine, knowledge of the conformation of the repeating epitopecomprising the immunodominant region of the protein will be useful in the designof an effective molecule. Various methods have been used to predict the three-20dimensional structure of the repeating tetrapeptide and different conclusions havebeen drawn. Theoretical investigation using energy minimization and moleculardynamics methods indicated that a right-handed helical conformation is likely tobe adopted in aqueous solutions while a left-handed helical conformation should befavoured in non-polar environment (Gibson and Scheraga, 1986; Fig. 4). A similarstudy suggested that the most stable structure of the repeating tetrapeptide is aright-handed helix with 12 residues per turn (Brooks et al., 1987). However, theChou-Fasman predictive algorithm indicates a high p-turn content in the syntheticpeptide (NANP)8 (i.e. NANP repeated eight times). Circular clichroismmeasurements showed that the presence of prolines in these repeats induces anincrease in the p-turn content (Fasman et al., 1990). Proton nuclear magneticresonance revealed that a repeating structural motif is formed by the NPNA(instead of NANP) cadence (Dyson et al., 1990). To date, the X-ray crystallographystructure of the CSP or the repeating region has not yet been resolved. The choiceof the NANP repeat as the model epitope in this study was based on the simplicityof its repeating pattern, the well-documented immunodoniinance of the epitope, andthe availability of epitope-specific antibodies from a collaborator.5. Factors affecting antigenicity and immunogenicity5.1 Factors affecting antieenicitvThe antigenicity of a molecule refers to its ability to interact with21ABFigure 4. Stereo drawings of two of the predicted structures of the (NANP)6peptide. A) The left-handed helical conformation. B) The right-handed helical conformation. In both views, the carboxyl terminusis at the top. Reproduced with permission from Gibson andScheraga, 1986.22antibodies. In principle, peptide antigenic determinants can be divided into twostructural categories: the continuous epitope, which consists of a contiguous stretchof the amino acid sequence; and the conformational epitope, which is an assembledtopographic site consisting of amino acid residues separated in the primarysequence but brought together during the folding of the protein. It is generallybelieved that the surface accessible regions of a protein usually contain antigenicdeterminants that can be recognized by the immune system, and thus stimulate theproduction of specific antibodies (Hopp and Woods, 1981). Since proteins insolution tend to fold in a way that exposes the hydrophilic amino acid residues tothe surface, amino acid sequences with high local hydrophilicity are quitefrequently predicted to have high antibody affinity (Berzofsky, 1985). The overallstrength of an antibody-antigen interaction is governed by 3 major factors: theintrinsic affinity of the antibody for the epitope, the valency of the antibody andantigen, and the spatial configuration of the interacting compounds. With amonoclonal antibody of defined specificity, it seems logical that the presence ofrepeating epitopes may increase avidity and hence overall stability. However, theeffects of multivalency may also involve spatial configurations that may imposesteric constraints on the interactions.5.2 Factors affecting immunoenicitvImmunogenicity, the ability to elicit an immune response, is determinedby the intrinsic chemical structures of a molecule and by the ability of the host23animal to recognize the molecule. The mechanisms of an immune response can bedivided into 2 categories: humoral, which mainly involves the production ofcirculatory antibodies, and cellular, which functions to target specific immune Tcells against fungi, intracellular pathogens and cancer cells, etc.. In general, for anantigen to elicit a good antibody response, both a B cell epitope and a T cell epitopeare required. The B cell epitope is recognized by B cell surface receptors tostimulate the production of antibodies of its own specificity. The T cell epitope,which results from antigen processing, is presented on the surface of an antigenpresenting cell or B cell in conjunction with a MHC class II molecule for binding tospecific T cell receptor. Clonal expansion of both B cells and T cells can then occurin parallel, leading to the production of specific antibodies by plasma cells. For asmall antigen (hapten) which is not likely to encompass a T cell epitope, thisepitope can be obtained by conjugating the hapten to a carrier protein. The wholehapten/carrier conjugate can then be used as an immunogen to elicit an anti-haptenantibody response. In addition to the intrinsic properties of a compound, theimmunogenicity of the compound also depends on the extrinsic factors such as theprocessing pathways in antigen presenting cells, the set of MHC moleculesavailable for antigen presentation, the presence of specific T suppressor cells, thedelivery system and the route of adminstration, etc. (Gammon et at., 1987;Gregoriadis, 1990; Monaco, 1992).246. Aims of this studyA number of outer membrane proteins have been developed as carriersfor foreign antigenic determinants in different epitope presentation systems.Limited studies have shown that the flanking amino acid residues and the lengthof the inserted epitope can affect the antigenicity and immunogenicity of theinserted epitope (Agterberg et al., 1990a; Van der Werfet al., 1990; Janssen et al.,l994b). In light of these findings, it is conceivable that determining the optimalparameters for epitope presentation would increase the potential utility of thesepresentation systems.The goals of this study were to develop the F. aeruginosa outer membraneprotein OprF as a carrier for foreign antigenic determinants, to employ this proteinto study the effect of the mode of presentation on the antigenicity of the presentedepitope, and to perform a limited study correlating antigenicity withimmunogenicity. The “permissiveness” of different regions of OprF to accommodateextra amino acid sequence was first examined by linker-insertion mutagenesis.The feasibility of OprF as a carrier for foreign antigenic determinants was theninvestigated by using the malarial tetrapeptide repeating epitope NANP as themodel epitope. Also, the influence of the insertion position and the length of theinserted epitope on the antigenicity of the epitope was examined by using a seriesof OprF::malarial epitope hybrid proteins. Finally, this was correlated with theimmunogenicity of the epitope presented in different ways in the context of OprF.25METHODS AND MATERIALS1. Bacterial strains, plasmids and mediaThe bacterial strains and plasmids used in this study are listed in TableII. Bacterial strains were grown in Luria broth (LB) {1%(w/v) tryptone, 0.5%(wlv)yeast extract, 0.5%(wlv) NaC1} unless otherwise stated. Strains of C158background were grown in Luria broth supplemented to a final concentration of1.7% (w/v) NaC1 and 0.1% (wlv) glucose to suppress the expression of OmpF andLamB respectively. When plasmids were present, media were supplemented with75 .tg/ml of ampiciilin or with 50 .tg/ml each of ampicillin and kanamycin. All mediacomponents were purchased from Difco Laboratories, Detroit, Michigan.2. General recombinant DNA techniquesGeneral DNA techniques were performed as described in Ausubel et al.(1987) and in Sambrook et al. (1989). Competent cells for transformation wereprepared using the CaCl2method (Hanahan, 1983). DNA fragments were isolatedeither by elution from preparative agarose gel onto DEAE paper (Schleicher &Schuell) or by using the GENECLEAN kit (BlO 101 Inc. La Jolla, CA). Restrictionenzymes and DNA modifying enzymes were purchased from Bethesda ResearchLaboratories (BRL, Burlington Canada) or Boehringer Mannheim (Mannheim,TableII.BacterialstrainsandplasmidsStrainorplasmidGenotype,phenotype,orrelevantpropertiesSourceand/orreferenceE.coliDH5F’F’8OdlacZM15(lacZYA-argF)U169deoRHanahan,1983recAlendAlhsdR17(r, mK)supE44AgyrA9Gthi-1relAlC158aroAilvmethispurE4lprocyc-lxylllacY29Foulds&Chai,1979rpsL97tsx63ompAompCC386lppompAtsxSonntagetal.,1983C466ara-14leuB6azi-6lacYl proCl4tsx-67A(ompT-Elishetal.,1988fepC)266entA4O3,AtrpE38rfbDlrpsL109xyl-5mtl-1thi-1C443(DH5F’IQ)F’8OdlacZAM15A(1acZYA-argF)U169deoRBRLrecAlendAl hsdR17(r,mj)supE44AgyrA96thi-1relAl[F’ proABlacIqZAMl5zzf::Tn5[Km9P.aeruginosaH103PAO1Cm’ prototrophHancock&Carey,197911692LPSA-,B-,sameasrd7513inreferenceLightfoot&Lam,1991TobecontinuedTableII.Bacterialstrainsandplasmids(continued)StrainorplasmidGenotype,phenotype,orrelevantpropertiesPlasmidspTZ18R/19RpHJ13pRW1cloningvector,AmprpTZ18Rwitha4.5kbinsertcarrying2separatefragmentsof oprF.pTZ19Rwitha4.8kbHindillIEcoRI insertcarryingtheentireoprFinonecontinuousfragment.pTZ19Rwitha1.47kbHindIII/KpnI insertcarryingtheentireoprFpUCtypeplasmidwithadrug-resistancemarkerfromTh903,jrandflprglutathioneS-transferasefusionproteinexpressionvectorSourceand]orreferencepRW3pUC4KAPApGEX- iNPharmaciaHancocklabThisstudyThisstudyPharmaciaPharmacia28Germany). Oligonucleotides were synthesized on an Applied BiosystemsIncorporated (ABI, Foster City, CA) 392 DNAIRNA synthesizer according tomanufacturer’s instructions. Oligonucleotides were purified by passing through aSep-Pak C18 cartridge (Waters, Division of IVliffipore) and eluting with 20%acetonitrile followed by ethanol precipitation. For oligonucleotides used in ligations,the sense and antisense strands of each set were annealed by heating an equalamount of each strand (100 jiM) in 2 mM MgC12/50 mM NaC1/20 mM Tris-HC1pH7.5 at 90°C for 15 mm, followed by gradual cooling to 23°C.3. General protein and immunological techniques3.1 SDS-PAGE and immunob1ottinsSDS-PAGE procedures were performed as described in Hancock andCarey (1979). Colony and Western immunoblotting procedures were as describedin Mutharia and Hancock (1985). The percentage of acrylamide used was 11%unless otherwise stated. The amount of proteins loaded per lane was normalizedby protein assays (described in section 3.5). For some Western immunoblots,polyvinylidene difluoride (PVDF) membrane (Millipore Corporation, Bedford, MA)was used and immunodetection was done without blocking (Mansfield, 1994).Briefly, the PVDF membrane was air-dried completely after protein transfer,followed by incubation with primary antibody for 1 h at 37°C. After 2 x 10 swashing in PBS, secondary antibody was added and incubated for 30 mm at 37°C,29followed by 2 x 10 s washing in PBS and colour development.3.2 AntibodiesThe OprF-specific mAbs and polyclonal serum used in this study were asdescribed by Finnen et al. (1992). The epitopes recognized by the OprF-specificmAbs have been delineated by Rawling et al. (1995) and are summarized in TableIII. The isolation of the malarial epitope-specific mAbs pf2A. 10 and pf5A4. 1 weredescribed by Wirtz et al. (1987).3.3 Indirect immunofluorescence labe11inImmunofluorescence labeffing was performed as follows. Overnightcultures of strains containing the specified plasmids were harvested and washedtwice in PBS. Slides were coated with poly-L-lysine (Sigma Chemical Co., St.Louis, Mo; average MW’—’25,000) by flooding with poly-L-lysine solution (1 mg/ml)in a moist chamber for 15-20 mm and then rinsing thoroughly with distified water.Samples of washed cells were smeared onto the poly-L-lysine coated slides andallowed to air dry briefly. Slides were then incubated with an OprF-specific mAbor a malarial epitope-specific mAb (both at 1/100 dilution) in PBS containing 1%fetal calf serum (FCS) for 30 mm at 23°C. After washing with excess PBS, slideswere incubated with fluorescein isothiocyanate-conjugated goat anti-mouse IgG(BRL) at 1/20 dilution in PBS/1% FCS for 30 mm at 23°C. Following PBS washing,one drop of mounting medium (Sigma) was added to the slides and the cells were30Table III. OprF epitopes recognized by monoclonal antibodiesa.Monoclonal Amino acid Type of epitope Surfaceantibody positions of epitope exposure ofepitope<MA7-1 55-62 linear +MA7-2 237-244 linear +1-MA7-3 188-230 conformationaib +MA7-4 188-2 78 conformational +MA7-5 188-2 78 conformational +MA7-6 198-240 conformational +MA7-7 188-278 conformational +MA7-8 152-2 10 conformational +MA4-4 152-2 10 conformational +1VIA5-8 307-3 14 linear +a Data are summarized from Rawling et al., 1995b Delined by non-reactivity with overlapping 8-amino acid peptides synthesizedon pins.+, surface-exposed; +1-, surface exposure only in an LPS altered rough strain.31examined under a Zeiss microscope fitted with a halogen lamp, a condenser andfilters for fluorescence microscopy at 525 nm for emission of fluoresceinisothiocyanate.3.4 Trvusin sensitivity assaysE. coli C386-derived strains containing different pRW3-derived plasmidswere grown in Luria broth supplemented with Amp (75 i.gIml) to an A of 0.8.Samples of 1.5 ml of the cultures were harvested and washed twice with 20 mMTris-HC1 pH 7.4 containing 5 mM MgC12, and then resuspended in 500 jil of thesame buffer. Trypsin (TPCK treated, Sigma) was added to a final concentration of0.1 mg/ml of cell resuspension, followed by incubation at 37°C for 60 mm. Untreated samples were incubated in the same conditions, except that trypsin wasomitted. Proteolysis was stopped by heating at 88°C for 10 mm in solubilizationbuffer (2% SDS, 10% glycerol, 62.5mM Tris-HC1 pH6.8). OprF in outer membranesamples was digested at a trypsin concentration of 0.1 mg/ml. The reactions werecarried out as described above. In both cases, the trypsinized samples wereanalyzed by SDS-PAGE and Western immunoblotting with the specifiedmonoclonal antibodies. As controls, the cleavage of OprF in P. aeruginosa intactcells to a 28 kDa core fragment and the complete cleavage of bovine serum albuminby trypsin to low molecular weight peptides were demonstrated.323.5 Protein assaysThe modified Lowry assay was performed as described by Sandermannand Strominger (1972). The bicinchoniriic acid (BCA) protein assay was performedin 96-well microtiter plates with a sample volume of 10 jt1 and the addition of 200iil of BCA reagent (Sigma). The plates were then incubated at 37°C for 30 mm andthe A550 was determined with a BioRad model 3550 ELISA microplate reader.4. Construction of pRW3Figure 5 outlines the subcloning of oprF. The 4.5 kb Smal fragmentcontaining the 3’ end of oprF was excised from pHJ13 and re-ligated in the oppositeorientation into the vector to construct pRW1, so that the coding region of oprF wasin one continuous fragment. The 1.47 kb HindIII/KpnI partial fragment containingthe entire oprF gene from pRW1 was then subcloned into the vector pTZ19R toobtain pRW3.5. Linker-insertion mutagenesis of oprF5.1 Mutapenesis with kanamycin resistance cassetteFigure 6 shows a schematic diagram of the procedures. The plasmidpRW3 was linearized separately by partial digestion with restriction enzymes:RsaI, HaeIII, ThaI or AluI (Boehringer Mannheim), all of which leave blunt ends33pHJI3AH KIM pp K s lIE___I I ISmal digeshonand religationpRWIKIMH K p p MEAiArAi’subclonmg of 1.47 kb HIKfragment into pTZ19RpRW3H K\II—lA7kb—IEcoRIB SadKpnIBsmI NaeIfl—onMiul f\SailAsullpRW3oprF l00KmpBaiI/ — 3000 4274 bpsNcoIKpnIAatII 2000Eco47—3P—iac p001322-onRindIIJ 4.Figure 5. Construction of pRW3.A. Schematic diagram of the subcloning of oprF. The thick solid line representsthe cloning vector pTZ19R; the hatched box represents the OprF coding region.The arrows under each box indicate the direction of transcription of oprF. H,Hindill; K, KpnI; S, Sail; P, PstI; M, Smal; E, EcoRI. B. Restriction map ofpRW3. The position and direction of transcription of oprF and the ampiciliinresistance marker (Amp) are indicated. Abbreviations: fl-on, fi phage originof replication; pBR322-ori, ColE 1 origin of replication; P-lac, lac promoter; bps,base pairs.34ABH4,GTCGACCTGCAGGCAGCTGGGTCPH)nCCCTGCAGGTC4&AC)nGGGCGTCCAGCTGP(OprF .ve)PstILinearize with blunt-end enzymesInsert Hind fragment of Km cassetteW Select for Km resistanceOR —(OprF +ve)Screen with OprF-specific mAbfor loss ofOprF reactivityCut with PstI and remove Km cassetteReligateOprF linker mutant with l2bpinsert containing PstI siteFigure 6. Schematic representation of semi-random linker-mutagenesis witha kanamycin resistance cassette.A. The kanamycin resistance cassette used for the linker-insertion mutagenesis.The solid box represents the region that encodes an aminoglycoside 3’-phosphotransferase gene conferring kanamycin, neomycin, and G418 resistance.The nucleotide sequences in the flanking regions are specified to indicate thesymmetric restriction enzyme sites. H, Hind; P, PstI. B. Procedures for linker-insertion mutagenesis. The solid and hatched boxes represent the kanamycinresistance cassette and the oprF gene respectively; Km, kanamycin, OprF -ye,non-reactive with OprF-speciflc (C-terminal reactive) mAb MA5-8; OprF +ve,reactive with MA5-8.1.3kb35after digestion. In the cases of AluI and ThaI, partial digestions were performedin the presence of ethidium bromide (20 and 50 jig/mi respectively) to increase therecovery of DNA molecules cleaved at a single site. After partial digestions, thereaction mixtures were resolved by preparative agarose gels and the full sizedlinear form of the piasmid was isolated by elution onto DEAE paper (Schleicher &Schuell). The four poois of linearized pRW3, each corresponding to a separaterestriction enzyme used, were ligated separately with a 1.3 kb HincIl fragmentcontaining the kanamycin resistance cassette from pUC4KAPA. Following ligationand transformation, cells were plated on Luria agar plates containing 50 jig/mi eachof kanamyciri and ampicillin. The doubly resistant colonies were further screenedby colony immunoblotting for loss of expression of OprF using the OprF C-terminal-specific mAb MA5-8 (Rawling et al, 1995 and Table III). Piasmid DNA from theclones that did not express OprF were extracted and then digested with PstI, whichonly recognized sites in the flanking sequences of the kanamycin resistancecassette, and hence cleaved the cassette from the plasmid. Following re-ligation ofthe PstI digestion mixtures and transformation, recombinants were screened forkanamycin sensitivity and the recovery of immunoreactive OprF by colonyimmunoblotting using the same OprF-specffic mAb. The OprF expressing (OprF’),kanamycin sensitive (KmS) clones presumably contained mutated forms of pRW3with a 12 bp insertion at sites originally interrupted by the kanamycin resistancecassette.365.2 Mutaenesis at the Sail siteThe plasmid pRW3O7 was constructed by inserting a self-hybridizing Sailadaptor oligonucleotide (5’TCGACCTGCAGG3’) which contained a PstI site into theSail site corresponding to position 188 (aa’88)in the mature OprF sequence. As aresult, the 4 amino acids DLQV were added after the valine residue at aa’88.5.3 Determination of linker-insertion sitesThe 12 bp insertion in the linker mutants carried a unique PstI site.Plasmid DNA was prepared from 100 OprF, Kms clones and the linker-insertionsites were mapped by restriction enzyme digest using double digestions withPstI/HindIII and PstI/SalI, where HindIII and Sail recogiiized unique sites atnucleotide positions -63 and +726 of the oprF gene coding sequence respectively(Duchene et ai., 1988). Clones with the same restriction pattern were grouped andthe exact position of insertion was confirmed by DNA sequencing of at least onerepresentative from each group.6. Construction of OprF::malarial epitope hybrid proteins6.1 Positional hybridsSince the linker insertions occurred in different reading frames, three setsof synthetic oligonucleotides were required to accommodate the three possible readingframes at the PstI sites (Fig. 7A). In the course of cloning, it was realized that37A. Phase 1P N A N P N A N P N A G H ACCG AAC GCC AAC CCG AAC CCC AAC CCC AAC GCC GGG CAT GCAACGTGGC TTG CGG TTG GGC TTG CCC TTG GGC TTG CGG CCC GTPhase 2N P N A N P N A N P N A CAC CCG AAC GCC AAC CCG AAC CCC AAC CCG AAC CCA TGC AACGG CCC TTG CGG TTC CCC TTG CCC TTG GGC TTG CGTPhase 3N A N P N A N P N A L D V QG AAC CCC AAC CCA AAC GCG AAT CCG AAT GCT CTA GAC TTG CAACGTC TTG CCG TTG GGT TTG CGC TTA CCC TTA CGA CAT CTG AB. 7-amino acid insertN P N A N P NTCGAAAC CCG AAC GCT AAT CCA AATTTG GGC TTG CGA TTA GGT TTAGATC11-amino acid insertN P N A N P N A N P NTCGAAAC CCC AAC GCT AAT CCA AAC CCC AAC CCT AATTTC GGC TTG CGA TTA GCT TTG CGG TTG CCA TTAGATC15-amino acid insertN P N A N P N A N PTCGAAAC CCC AAC CCT AAT CCA AAC GCC AAC CCCTTC GGC TTC CGA TTA GGT TTG CGG TTC GGGN A N P NAAT GCA AAT CCC AATTTA CGT TTA GGC TTAGATC19-amino acid insertN P N A N P N A N P NTCGAAAT CCA AAC GCC AAC CCG AAC GCA AAC CCC AATTTA CGT TTC CCC TTG CCC TTG CGT TTG GGG TTAA N P N A N P NCCA AAT CCT AAC CCC AAC CCA AATCGT TTA GCA TTG CGC TTG CGT TTAGATC38Figure 7. Nucleotide and encoded amino acid sequences of theoligonucleotides used for the construction of OprF::malarialepitope hybrid proteins.A. The three sets of oligonucleotides used for the construction of the positionalhybrids. Each set encodes the malarial epitope sequence in one of the threepossible reading frames at the PstI cleavage sites. The PstI compatible ends forthe ligation into the PstI sites generated by the linker-insertion mutagenesisprocedures are in italics. The unique restriction enzyme sites engineered in theoligonucleotides are underlined (SphI in frames 1 and 2, XbaI in frame 3). B.The four sets of oligonucleotides encoding different lengths of the repeatingepitope. All four sets carried XhoI and XbaI sites (in italics) on each endrespectively for the ligation into the corresponding sites generated by the PstIadaptors for directional cloning. The amino acid sequences are indicated in oneletter code.39maintaining the same codon usage in each NANP repeat caused the problem of hairpinloop formation when more than one copy of the insert was ligated in oppositeorientation. Therefore, a different codon usage for the NANP repeats was chosen forthe Phase 3 oligonucleotides. Each set of the annealed synthetic oligonucleotidesencoding the malarial epitope was ligated into the PstI sites of the various oprF linkermutant plasmids. After transformation, the recombinants were screened by colonyimmunoblotting with the OprF-specific, N-terminus reactive mAb MA7-1 (Rawling etaL, 1995, and Table III) and the malarial epitope-specific mAb pf2A.lO (Wirtz et al.,1987) separately. Plasnnd DNA from transformants that reacted positively with bothmonoclonal antibothes were extracted and analyzed by restriction analysis and DNAsequencing.6.2 Multinle-reneat hybridsThree sites of OprF (aa26, aa196 and aa213)were chosen for further study of thelength effect of the epitope on its antigenicity and immunogenicity. HybridOprF::malarial epitope plasmid constructs carrying different lengths of the epitopewere generated as follows:i) Insertion of PstI adaptors: The three selected sites were of two different reathngframes. To simplify the cloning procedures and the number of oligonucleotidesrequired, two sets of adaptors were synthesized and inserted into the PstI sites of thecorresponding linker mutants. The insertion of the adaptors created two unique sitesfor directional cloning (XbaI and XhoI) and also adjusted the reading frames at the40three chosen sites so that only one set of malarial epitope oligonucleotides would berequired for all three sites.ii) Insertion of malarial epitope encoding oligonucleotides: Four sets ofoligonucleotides, representing 7 amino acids (NPNANPN), 11 amino acids{(NPNA)2NP }, 15 amino acids {(NPNA)3NP } and 19 amino acids {(NPNA)4NP } ofthe malarial epitope were synthesized (Fig. 7B). The oligonucleotides contained XhoIand XbaI compatible ends and were designed so that the XhoI site would be destroyedafter the ligation. Transformants were screened by colony immunoblotting with themalarial epitope-specffic mAb pf2A. 10. Plasmid DNA from positive clones wasextracted and the incorporation of the oligonucleotides was confirmed by restrictiondigest analysis.The resultant hybrid proteins were designated as OprF::ME(X)aa(Y), whereME refers to jpa1arial pitope, X refers to the number of amino acids inserted, and Yrefers to the amino acid position of the insertion.7. DNA sequencingAutomated DNA sequencing was carried out with the Applied BiosystemsIncorporated (ART, Foster City, CA.) model 373A DNA sequencing system using thepolymerase chain reaction and dye-terminator chemistry as described by themanufacturer’s protocols. Sequence analyses were performed using the ABI 675DNA sequence editor program. Template DNA was prepared using Qiagen columns41(Qiagen Inc., Chatsworth, CA 91311) according to the manufacturer’s protocols.To determine the exact position of the insertion sites of the representative oprFlinker-insertion mutants, plasmid DNA from the corresponding kanamycinresistant clones was used as template. The sequencing primers used were 21 meroligonucleotides (5’ATGTAACATCAGAGATTTTGA3’ and5’TATGAGTCAGCAACACCTTCT3’) that hybridized to opposite strands of thekanamycin resistance cassette, approximately 50 bp from the ends of the cassette(Oka et al., 1981). The directions of extension from these primers were outwardfrom the cassette so that the oprF sequences flanking the insertion sites could beidentified. DNA sequencing to determine the number and orientation of themalarial epitope insert in the oprF::malarial epitope hybrid plasmids was carriedout by using primers that hybridized to oprF gene sequences at appropriatedistances upstream of the insertion sites. The sequences of the primers(5’- 3’) were: FP1,‘9TTAGGCGTTGTCATCGGCTCG39; FP2,402AACATGGCCAACATCGGCGCT2;FP3,577CCGGAACCGGTTGCCGACGTT9FP4, 879 GAGCGTCGTGCCAACGCCGTT99; FP5,702GTCGTACGCGTACAGCTGGACGTG25(numbers in superscripts indicate thepositions of the first and last nucleotides in the oprF gene sequence as described byDuchene et al., 1988).428. Construction of glutathione S-transferase (GST)::malarial epitopefusion proteinsTwo sets of annealed synthetic oligonucleotides, each encoding 11 and 19amino acids corresponding to the malarial epitope, were ligated into the BainHIand EcoRI sites of the pGEX cloning vector (Fig. 8). Transformants were screenedby colony immunoblotting with the malarial epitope-specific mAb pf2A. 10. Theresultant fusion proteins, GST: :ME 11 and GST: :ME 19, expressed 11 amino acids{P(NANP)2N } and 19 amino acids {P(NANP)4N } respectively at the C-terminusof glutathione S-transferase (GST).9. Isolation of outer membranes9.1 Triton X- 100 extractionThe outer membranes of OprF linker mutants were isolated by selectiveTriton X- 100 solubilization of cell envelopes as described by Schnaitman (1971).9.2 Sucrose gradient centrifuationOvernight cultures (1 L) of C 158-derived strains were harvested andresuspended in 10 ml of 20% sucrose, 10mM Tris-HC1 pH 8.0. DeoxyribonucleaseI (50 jtg/ml) was then added to each cell resuspension, followed by incubation at23°C for 20 mm. Cell lysis was achieved by two passages through a French43A EcoRISTOP i’SmaICOdQflSNIfr BamHISj26Ampy- ‘4/7 tacPsil j/ff pGEX-1‘49kbEcoRVORI—i1qB 11 amino acidsD P N A N P N A N P N A Q LGAT CCG AAC GCC AAT CCG AAT GCG AAC CCA AAC GCA CAG CTGGC TTG CGG TTA GGC TTA CGC TTG GGT TTG CGT GTC GAC TTAA19 amino acidsD P N A N P N A N P N A N P NGAT CCG AAC GCC AAT CCG AAT GCG AAC CCA AAT GCT AAC CCC AACGC TTG CGG TTA GGC TTA CGC TTG GGT TTA CGA TTG GGG TTGA N P N A Q LGCA AAT CCT AAC GCA CAG CTGCGT TTA GGA TTG CGT GTC GAC TTAAFigure 8. Construction of GST::malarial epitope fusion proteins.A. Schematic representation of the pGEX-l cloning vector. Abbreviations: Sj26,the gene encoding a 26 kDa glutathione S-transferase (GST); tac, tac promotor;P,irip’ p-lactamase gene; ORI, origin of replication. B. The nucleotide andamino acid sequences of the oligonucleotides encoding the malarial epitope. Theoligonucleotides carried BamHI and EcoRI compatible ends (in italics) forthrectional cloning into the multiple cloning sites of pGEX- 1. Fig. 8a isreproduced from Smith and Johnson, 1986, with permission.44pressure cell at 15,000 psi. The lysed cells were centrifuged at 1,700 x g for 10 mmto remove cell debris. The supernatants were then applied onto a 2-step sucrosegradient {50%/70%(w/v)} and centrifuged at 100,000 x g in a SW28 rotor (Beckman)for at least 6 h at 4°C. The lower band that formed at the interface of the 50% and70% sucrose layers was collected and the sucrose was diluted with at least twovolumes of distilled water, followed by centrifugation at 200,000 x g in a 6OTi rotor(Beckman) for 1 h. The final pellets were resuspended in 1 ml of distilled water.The protein concentration in each sample was determined by a modified Lowryprotein assay.9.3 Removal of inclusion bodiesOuter membrane samples containing the series of OprF::malarial epitopemultiple-repeat hybrid proteins at aa26 were contaminated with inclusion bodies.The membrane bound form of these hybrid proteins was obtained by octylpolyoxyethelene (octyl-POE) extraction. Briefly, the loosely-bound proteins in thepreparations were removed by resuspending the outer membrane pellets in 0.5%octyl-POE, followed by incubation at 37°C for 30 mm and centrifugation at 4°C for15 mm at 13,000 x g. The OprF::malarial epitope hybrids were released byresuspencling the insoluble fractions in 3% octyl-POE, 10mM EDTA, followed byincubation at 37°C for 30 mm and centrifugation in a microfuge at 4°C for 15 mmat 13,000 x g. Extraction with 3% octyl-POE, 10mM EDTA was repeated. Themembrane bound form of the proteins was found to be contained in the45supernatant.10. Expression of oprF and oprF derivatives in E. coil10.1 Expression of ovrFin different B. coli strainsThe plasmid pRW3 was transformed into different B. coli host strains bythe CaC12method (Hanahan, 1983). IPTG was added at 1 mM final concentrationto mid-log phase cultures and growth was continued at 37°C for another 4 h. Un-induced cells were harvested at the same time as the induced cultures. Cellenvelopes were prepared by centrifugation of whole cell lysates at 200,000 x g for1 h. The expression level of oprF was examined by Western immunoblotting of cellenvelope samples with the OprF C-terminal-specific mAb MA5-8.10.2 Exyression of an 0DrF derivative in different induction conditionsFresh LB broth (50 ml in 250 ml flask) was inoculated at 1/100 dilutionwith an overnight stationary phase culture of C158 containing pRW3O7. 1M. IPTG,at 0.2 miVi or 1 m]Vl final concentration, was added to cultures either at the time ofinoculation or when the cells were at mid-log phase (A “0.5-0.6). IPTG inductionwas carried out at 30°C or 37°C for 3 h or 16 h as indicated. Outer membranes wereprepared using differential Triton X-100 extraction as described in Section 9.1.Expression of OprF: :hybrid protein was examined by Western immunoblotting withan OprF-speciflc mAb.4611. Protein purification11.1 OyrF::malarial epitone hybrid }JroteinsOprF and OprF::malarial epitope hybrid proteins were purified fromplasmid-containing derivatives of E. coli strain C 158. Outer membrane samplescontaining OprF or OprF::malarial epitope hybrid proteins were prepared by a 2-step sucrose gradient centrifugation as described in Section 9.2. The samples werethen extracted sequentially with 0.5% octyl-POE, 3% octyl-POE /50 mM NaC1, and3% octyl-POE/10 mM EDTA. Extractions were performed by resuspending theinsoluble fractions in the detergent solutions, followed by incubation at 37°C for 1h and centrifugation at 200,000 x g for 1 h. Supernatants from the 3% octylPOE/10 mM EDTA extractions contained predominantly OprF or OprF::malarialepitope hybrid proteins. The detergent-extracted proteins were further purified byFPLC using aji anion exchange column, MonoQ (Pharmacia), and elution with anNaC1 gradient. Column buffer contained 0.5% octyl-POE, 10 mM EDTA, 10 mMTris-HC1 p118.0. Purified OprF or OprF::malarial epitope hybrid proteins wereeluted in the flow through fractions while the contaminants bound to the column.11.2 GST::malarial enitone fusion uroteinsThe GST::malarial epitope fusion proteins were purified by affinitychromatography using glutathione agarose beads as described in Smith andJohnson, 1988. Briefly, the procedures involved breaking of cells using a French-47pressure cell, centrifugation to remove cell debris, incubation of the cellsupernatant with glutathione agarose beads to allow binding of the GST::malarialepitope fusion protein to the matrix, washing away of non-binding proteins, andelution of the fusion protein with 5-10 mM reduced glutathione. The anti-GSTpolyclonal serum was kindly provided by Dr. Michael Gold (Department ofMicrobiology and Immunology, U. of British Columbia).11.3 Extraction from SDS-uolvacrvlamide gelProtein samples were separated by preparative SDS-PAGE (11%). Bandsof interest were excised and the proteins were eluted into 0.1% SDS, 10mM EDTA,10mM Tris-HC1 pH8.0 by incubation at 4°C for 16 h. The eluted proteins werequantitated by SDS-PAGE, followed by Coomassie blue staining and measurementof the intensity of the bands by scanning densitometry using a protein + dnaJmageware apparatus (protein + dna Imageware Systems, PDI, NY, U.S.A.). Theconcentration of the samples was extrapolated from a standard curve obtained fromprotein samples with known concentrations.12. Antigenicity studies12.1 Outer membrane ELISAOuter membrane samples containing OprF::malarial epitope hybridproteins were diluted to various concentrations (from 0.5 to 20 tg/ml) in carbonate48buffer (l5iulVlNa2CO3/35mM NaHCO3/3mM NaN3pH 9.6). Dilutions of the outermembrane samples (100 jil) were used to coat the bottom of 96 well plates byincubation at 4°C for 16 h. The wells were then washed twice with PBS containing5 mM MgC12 and blocked by incubation with 3% BSAJPBS at 37°C for 1 h. Afterwashing, 100 .tl of primary antibody (1/2000 of rabbit-anti-OprF antiserum or1/2000 of pfA. 10) was added. After incubation (3 7°C, 1 h) and washing, 100 jil ofhorseradish peroxidase-conjugated secondary antibody was added to each well(37°C, 1 h). 3,3’5,5’ Tetramethylbenzidine (TMB) (Pierce Chemical Co., USA) wasused as a chromogenic substrate and the reactions were stopped after 5-10 mm bythe addition of 1 MH3P04. The A450 readings of the wells were obtained using aBioRad ELISA microplate reader (model 3550) with a 450 nm ifiter. To normalizethe expression levels of the hybrid proteins, each index was the ratio of the A450readings when pf2A. 10 was used as the primary antibody to the A450 readingswhen the OprF-specific polyclonal antibody was used as the primary antibody. Foreach experiment, a plot of A450 readings versus the concentrations of coatingantigen was drawn for each antibody, only values that corresponded to the linearportion of the binding curve were used for the calculation of antigenicity indices.Due to the presence of inclusion bodies in the outer membrane samplescontaining the multiple-repeat hybrids carrying an insertion at aa26, the membranebound protein solubiiized in 3% octyl-POE was used in ELISA. The samples werediluted at least 40 fold in carbonate buffer and the concentration of the detergentin the other samples was adjusted so as to standardize the effect of the detergent49on the antigen-antibody interactions in all of the samples.12.2 Whole cell dot blot analysislVlid-logarithmic growth phase cells of strain C 158 expressing the hybridplasmids were harvested, washed twice with PBS and diluted in PBS to lx 108,2x107,4x106,and 8x105 cells/gl. One t1 of each cell resuspension was spotted ontonitrocellulose filters, and the blotting procedures were performed as described inMutharia and Hancock (1983). The intensities of the dots were quantitated bydensitometry with the protein + dna Imageware (PDI) systems using the QuantityOne software. Each antigenicity index was the mean of the ratios of anti-malarialepitope reactivity to anti-OprF reactivity obtained from four sets of dotsrepresenting different numbers of cells (8x105 to lx 108 cells).12.3 Statistical analysesThe antigenicity indices of the inserted epitope in the positional hybridswere compared by using F-tests. The differences discussed in the text as significanthad p values <0.05. The relationship between the antigenicity and the length of theepitope was analyzed by linear regression. The value of correlation coefficient (r)lies between -1 to +1, where r=0 indicates no linear relationship, r>0 indicates apositive linear relationship (the closer to 1, the stronger the correlation), and r<0indicates a negative linear relationship.5013. Immunization studies13.1 Immunization with OnrF::MElOaa2l5 and OprFTwo groups of 6-8 week old female BALB/c mice (H2d background) wereimmunized subcutaneously with 10 .tg of FPLC-purified OprF or OprF::MElOaa2 15with TitermaxTM (CytRx Corp., Norcross, Georgia) on days 0 and 14. On day 28, theanimals were injected with 2x108 cells of heat-killed E. coli expressing thecorresponding OprF or OprF::malarial epitope hybrid. Serum samples wereobtained by tail-bleeding on days 7, 21 and 35. The control group was injected with100 jil of PBS for all three injections.13.2 Immunization with OirF::MEaa26 multivle-reveat hybrids andGST::malarial eyitoye fusion yroteinsGroups of 6-8 week old female C57BL/6J mice(11..2b background) wereimmunized subcutaneously with 20 jig of immunogens on days 0 and 21 and with10 jig of immunogens on day 35 each suspended with 200 jig of Adjuvax (AlphaBeta Technology, Worcester, MA) as an adjuvant in 200 jil total volume. Serumsamples were obtained by tail-bleeding on days 0 and 28 and by whole body bleedon day 45.13.3 Determination of antibody titersThe anti-OprF titer in serum samples was determined by ELISA using51FPLC-purifled OprF from P. aeruginosa as the coating antigen (500 ng/mi). Theanti-malarial epitope titer was determined by using affinity-purified GST::ME19(2 jig/mi) or gel-purified OprF::MEl9aa26 (1 jig/mi) as the coating antigens.GST::ME19 and OprF::MEl9aa26 were chosen because these proteinsdemonstrated highest binding to the malarial epitope-specific mAbs in ELISA ascompared to the corresponding proteins carrying the shorter versions of the epitope.In addition, the anti-malarial epitope peptide titer was determined by ELISA usingthe chemically synthesized peptide NANPNANPNANP (NANP)3 (API, Edmonton,Alberta) as the coating antigen. The peptide (lOjtg/mi) in PBS was covalentlylinked to the wells of ReactiBindTM maleic anhydride activated polystyrene plates(Pierce) by incubation at 4°C for 16 h. The secondary antibody used in the assayswas horse radish peroxidase-conjugated goat anti-mouse IgG (heavy and lightchains) (BioRad).13.4 Characterization of antisera by Western immunoblot analysisTo detect the presence of anti-OprF antibodies in the antisera, FPLCpurified OprF (20 jig per gel) was resolved by SDS-PAGE and was transferred ontoPVDF membrane, the filter was then cut into slices and incubated with serumsamples from the immunized animals at 1/1000 dilution or with a 1/3000 dilutionof MA7-2 as a positive control. The presence of anti-GST antibodies in the antiserawas detected in a similar manner using affinity-purified GST (20 jig per gel). Thepresence of anti-malarial epitope antibodies in the groups immunized with52GST::malarial epitope fusion proteins or OprF::malarial epitope hybrid proteinswas detected by using outer membrane preparation containing OprF::MElOaal96(50 g per gel) or affinity-purified GST::ME19 fusion protein (20 ig per gel)respectively. Antisera from immunized animals were diluted 1/100. The malarialepitope-specific mAb pf2A. 10 (1/3000 dilution) was used as a positive control.Subsequent incubations with secondary antibody and enzymatic staining werecarried out as described by Mutharia and Hancock (1983).53RESULTSChapter one: Construction and characterization of OprF linker mutants1. IntroductionLinker-insertion mutagenesis, either random or site-directed, has beenemployed to study the topology of several E. coli outer membrane proteins,including the maltoporin LamB (Boulain et al., 1986), the phosphate-starvation-inducible porin PhoE (Bosch and Tommassen, 1987) and the major peptidoglycanassociated protein OmpA (Freudi et al., 1986). The mutagenesis introduces extraamino acid residues at specific or non-specific sites of these membrane proteins,while leaving the rest of the proteins intact. As a result, it represents a more subtlemodification of the protein in comparison to the other genetic approaches employedto study membrane topology such as alkaline phosphatase and -galactosidasegene fusions Manoil, 1991). In general, the extended surface loop regions are morelikely to accommodate extra amino acids without gross perturbation of the proteinstructure. Indeed, the 3-dimensional structures of PhoE and LamB confirmed thatall of the known insertion sites occurred within these ioops (Cowan et at., 1992;Schirmer et at., 1995).This chapter describes the expression of oprF in E. coli, a semi-randomlinker-insertion mutagenesis of oprF and the results of the characterization of the54linker mutants. The data obtained have raised the possibility that certain regionsof OprF can be used to express longer foreign amino acid sequences.1.2 Expression of oprF in E. coliWhen attempts to produce recombinant OprF were initiated, plasmidvectors that allowed stable expression of cloned genes in P. aeruginosa were notavailable; hence, E. coli was chosen as the background strain for the expression ofoprF. Earlier attempts to subclone oprF into high copy number plasmids wereunsuccessful, probably due to the efficient expression of oprF from its own promoterin E. coli leading to over-expression lethality (Woodruff et al., 1986). Plasmids thatcontained P. aeruginosa DNA containing oprF sequence were already available inthis laboratory. However, the linker mutagenesis procedure required an OprFencoding plasmid that did not contain a PstI site. Therefore, the subcloning of oprFwas necessary to generate a plasmid that would allow the expression and the linkermutagenesis of oprF in E. coli.1.2.1 Construction of uRW3The plasmid pHJ13, made by Helen Jost in this laboratory, was used forthe initial subcloning of oprF. The plasmid contained two fragments of P.aeruginosa chromosomal DNA containing respectively the 5’ and 3’ portions of theOprF coding sequence inserted in the cloning vector pRK4O4 (Methods and55materials section 4, Fig. 5A). The putative -10 site of oprF in pHJ13 was mutatedby adding a G:C nucleotide pair between nucleotides -9 and -10 to create a HindIIIsite. This procedure weakened the oprF promoter, avoiding the over-expressionlethality in E. coli. To put the oprF coding region in one continuous reading frame,the Smal fragment in pHJ13 had to be inserted in a reverse orientation (Fig. 5A).Therefore, the plasmid pHJ13 was digested with Smal and the fragments were religated. The final plasmid, pRW3, contained a 1.47 kb HindIII/KpnI fragmentcarrying the entire oprF gene with a mutated promoter in the cloning vectorpTZ 19R. The transcription of the gene was in the same direction as the lacpromoter (Fig.5B). In addition to the elimination of the PstI sites, the subcloningalso removed most of the chromosomal DNA flanking oprF, thus reduced the sizeof the plasnud by approximately 3 kb and could potentially improve the efficiencyof further genetic manipulation.1.2.2 Expression of ovrFin different E. coli host strainsSince a certain genetic background in the host strain might beadvantageous for the optimal expression of oprF, the plasmid pRW3 wastransformed into different E. coli host strains to examine expression levels. The cellenvelope samples of strains carrying pRW3 were analyzed by SDS-PAGE (Fig. 9).In most of the strains examined, the band corresponding to OprF was not readilyobserved in Coomassie blue stained gel (Fig. 9), indicating that the expression levelof oprFin these E. coli strains was modest at best. Western immunoblotting withA56B50 —DH5 C466 C443-- + --+-- +OprFDH5LI C466 C443-- +-- +-- +35.1— — — .- .4Figure 9. Expression of pRW3 in different E. coli host strains.A. SDS-PAGE of cell envelope proteins from strains expressing pRW3. Proteinsamples (20 jig/lane) were heated at 100°C for 10 mm in solubilization buffer (2%SDS, 10% glycerol, 62.5mM Tris-HC1 pH6.8) before loading. B. Westernimmunoblot of the same samples with an OprF-specific mAb MA7-1. Symbols:--, un-induced; +, induced with 1mM IPTG at 37°C for 4 h. Each lane contained20 jig of proteins. Numbers on the left indicated the positions of the relevantmolecular mass standards (kfla). The position of OprF is indicated by an arrowhead.57an OprF-speciflc monoclonal antibody showed that the expression level of oprFvaried in different host strains. For instance, the un-induced levels of OprF inDH5a and C443(DH5cIQ) were comparatively higher than that in the membraneprotease OmpT-deflcient strain (C466). Similar levels of expression in the DH5and DH5IQ strains were unanticipated since the presence of the lac repressor wasexpected to suppress the expression from the lac promoter under un-inducedcondition. Although the basal expression level of oprF was lower in C466 (OmpT),it increased more significantly than that in the two DH5cc strains upon IPTOinduction. Since C466 is OmpT-deflcient, the lack of this membrane protease mightpermit tolerance of a higher amount of OprF in the outer membrane of K. coli.In the course of the subcloning, the 1.47 kb HindIII/KpnI fragmentcontaining oprF was also incorporated into the cloning vector pTZ 18R, resulting inthe transcription of the gene in the opposite orientation to the lac promoter.Expression of oprF from this plasnud was not observed, implying that the mutatedoprF promoter was not functional. Furthermore, when oprF was in the sameorientation as the lac promoter, the level of OprF production was increased uponIPTG induction, suggesting that the transcription of the gene was under the controlof the lac promoter. Since DH5 is a widely used strain for genetic manipulationand since the basal level of oprF expression appeared to be sufficient for ourpurposes, it was used as the background strain for most of the characterization inthe later stages of this study.581.3 Semi-random linker mutagenesis with a kanamycin resistancecassetteThe kanamycin resistance cassette used for the mutagenesis containedthe gene encoding an aminoglycoside 3’-phosphotransferase, which conferskanamycin resistance. The gene was flanked by symmetric restriction enzymesites. The restriction enzyme sites included PstI, which were flanked by Hincil, ablunt-end cutting enzyme (Methods and materials section 5, Fig. 6A). The plasmidpRW3 was linearized separately by partial digestion with 1 of 4 blunt-end cuttingrestriction enzymes as described in Methods and materials section 5.1. There werea total of 74 cleavage sites in pRW3 that were recognized by the four enzymesutilized, and 37 of these were within oprF. Low enzyme concentrations andlorethidium bromide were used to favour the production of singly cut plasmids. Afterligation of the restriction enzyme-linearized plasmid pRW3 with the 1.3 kb Hinclifragment of the kanamycin resistance cassette, plasmid DNA from 100 clones thatappeared to have the insertions within oprF were digested with PstI. The PstIdigestion removed the kanamycin resistance cassette but left behind a residue of12 nucleotide pairs in length, which was between the Hincli sites and PstI sitesflanking both sides of the cassette (Fig. GA). After re-ligation, 44 of the 100kanamycin sensitive clones had regained the ability to produce OprF, asdetermined by colony immunoblotting with the OprF C-terminal-specffic mAb MA5-8. Presumably, these clones represented insertion sites in OprF that could59accommodate the insertion of 4 extra amino acids without affecting the productionof the protein. The rest of the kanamycin sensitive clones were unable to produceimmunoreactive OprF or produced OprF that reacted only weakly with MA5-8 oncolony immunoblots. Restriction enzyme analysis was used to map the insertionsites in each of the 100 plasmids. Sites in the 44 OprF-expressing plasmids couldbe placed into 10 unique groups (e.g. Fig. 10). The remaining 56 plasmids includedthose which demonstrated gene rearrangements or deletions, probably due tomultiple cleavages of pRW3 by the restriction enzymes prior to ligation with thekanamycin resistance cassette.1.4 Site-directed mutagenesis at the Sail siteThe unique SalT site corresponding to amino acid position 188 (aa’88) ofthe mature OprF was a potentially interesting site to study because of its locationin the cysteine-containing region. To obtain the same 4-amino acid insertion at theSalT site, a 12 bp adaptor containing a PstI site was inserted. The characterizationof this mutant was performed simultaneously with the rest of the linker-insertionmutants.1.5 Determination of insertion sitesThe exact linker-insertion sites in at least one representative from each60Figure 10, Restriction mapping of linker-insertion sites.10 11 12Restriction digest patterns of plasmid DNA from 12 of the kanamycinresistant clones. Each three lanes correspond to plasmid DNA from one clonetreated in the following ways (from left to right): i. uncut, ii. PstI digest, iii.PstJiHindIll double digest. Each plasmid had 2 PstI sites, one on each side ofthe cassette, and 2 HindIII sites, one within the cassette and one at thepromotor region of oprF. In the PstI digest lanes, the 1.3 kb and the 4.3 kbfragments represent the kanarnycin cassette and the pRW3 plasmidrespectively. In the PstIIHindIII double digest lanes, the 600 bp and 700 bpfragments (marked by open triangles) correspond to the kanamycin cassettewhile the 4 kb and the fourth fragments (marked by solid circles) correspondto pRW3. The sizes of the fourth fragments indicated the distance of theinsertion sites from the Hind.III site, Numbers on the left indicate the positionsof molecular size markers (bp).234 678961of the 10 groups of mutant plasmids encoding immunoreactive OprF weredetermined by DNA sequencing. The nucleotide positions of the linker insertionsand the identities of the 4 inserted amino acids for 11 linker-insertion mutants and1 site-directed insertion mutant (pRW3O7) are summarized in Table IV. Seven ofthe mutant plasmids that did not express immunoreactive OprF were also analyzedby sequencing CPable V). The results revealed that two of the mutant plasmids hadincorporated the 12 bp insert at nucleotide positions +433 and +795 of oprFrespectively, but the reading frames at these insertion sites both led to thetranslation of stop codons from the 12 bp insert. The other four mutant plasmidsanalyzed represented deletions of part of the oprF sequence, and the study of thesemutants was not further pursued in this work. Only one (pRW3O3) of the sevenmutant plasmids analyzed showed the incorporation of a 12 bp insert in the OprFcoding region without any other genetic alteration or change of reading frame. Theinability of this clone to demonstrate an OprF positive phenotype on colony orWestern immunoblots suggested the “non-permissiveness” of this insertion site.1.6 Expression and cellular localization of linker mutantsIn addition to the signal peptide, the primary sequence of maturemembrane proteins is believed to carry the targeting signal for the export of theseproteins across the bacterial membrane (Maclntyre and Henning, 1990). Therefore,it is possible that insertion of extra amino acids in the primary sequence of OprF62Table IV. Summary of insertion sites of 11 linker-insertion mutants and onesite-directed insertion mutant and the identities of the insertedamino acids.Insertion Insertion .. ApparentAmino acidsPlasmids sites sites. mol. massea a inserted(nucleotides) (amino acid) (kDa)pRW3O1 77 Gly-2 TCRS 41pRW3O2 148 A1a26c PAGP 36pRW3O3 198 Glu-42 DLQV NEpRW3O5 463 Ala-131° PAGP 40pRW3O6 476 Gly-135 TCRS 35pRW3O7b 636 Val- 188 PAGP 35pRW3O8 658 A].a-196° PAGP 35pRW3O9 710 Arg-213 TCRS 35pRW31O 717 Gln-215 DLQV 35pRW311 764 Ser231d TCRS 36pRW312 939 Arg-290 TCRS 35pRW314 1001 G1y310 TCRS 28a Position 1 is the translational start site (Duchene et al., 1988). The amino acidnumbers correspond to the mature native OprF.b pRW3O7 was generated by inserting a Sail adaptor that contained a PstI siteinto the Sail site corresponding to aa’88.The alanine residue at these insertion sites was replaced by a glycine.d The serine residue at the insertion site was replaced by an arginine.The apparent molecular mass of wild type OprF expressed by pRW3 is 35kDa.NE, no expression.63Table V. Summary of six of the deletion mutants isolated during linker-insertion mutagenesis.Insertion sites Insertion sites Amino acidsPlasnrids (nucleotides)a (amino acid) inserted or mutationspRW3O4 433 Tyr-121 stop CodoflbpRW313 1001 Gly-310 deletioncpRW315 795 Tyr-245 stop codonbpRW316 898 Val-279 TCRS + 24 aa dpRW317 399 AIa-114 pAGp+83aadpRW318 296 Gly-78 DLQV+ 97 aada Position 1 is the translational start site (Duchene et al., 1988).b The first codon encoded by the linker is a stop codon.The linearized pRW3 was cleaved at multiple sites so that the rest of the codingregion was deleted.d The linearized pRW3 was cleaved at multiple sites. The extra amino acidsencoded were due to a frame shift and represented the translated sequencebefore the first stop codon was encountered.64might affect its transport to the outer membrane. To examine the cellularlocalization of the OprF linker mutants, the outer membrane of F. coli containingthe linker mutant plasmids was isolated by using Triton X-lOO extractionprocedure (Schnaitman, 1971). SDS-PAGE analysis of the outer membranefractions demonstrated the presence of OprF and OprF linker mutants, suggestingthe association of these proteins in the outer membrane of F. coli (Fig. 11). Theelectrophoretic mobility of all of the linker mutants was modified by pre-treatmentwith 2-mercaptoethanol, indicating that the inserted amino acid residues did notperturb the formation of the OprF cysteine disulphide bonds (Fig. 11, lanes 6 & 10).The apparent molecular mass of the OprF linker mutants carrying aninsertion at aa2 and aa’3’ (encoded by plasmids pRW3O 1 and pRW3O5 respectively)was noticeably greater than that of the wild type (Table IV). The protein expressedby pRW3O5 migrated with a mobility similar to that of the heat-modified (unfolded)form of OprF (Fig. 12, lane 3), suggesting that the incorporation of the 4-amino acidlinker may have increased the susceptibility of the protein to denaturation byheating in SDS. Plasmids pRW3O9 (aa213) (Fig. 11, lane 9) and pRW311 (aa231)(data not shown) each directed the expression of an intense band with an apparentmolecular mass of 70 kDa, likely corresponding to the trimenc form of OprFMutharia and Hancock, 1985). After 2-mercaptoethanol treatment, a much moreintense monomer band with an apparent molecular mass of 35 kDa was observedin the same samples (e.g. Fig. 11, lane 5), suggesting that insertion of the linkerat aa213 and aa23’ may enhance the association of SDS-stable oligomers.651 2 3.4.5 678495—• . -__---—4. 4-4(4--__I..— 1.• ?___—__ __- -Figure 11, Cellular localization of OprF linker mutants.SDS-PAGE of outer membrane samples of E. coli DH5F’ strains carryingthe pRW3-derived. plasmids. Samples were prepared by using Triton X-l00extraction procedures (Schnaitman, 1971) and were incubated at 37°C for 10mm in solubilization buffer with (lanes 2-8) or without (lanes 9 and 10) 4% 2-mercap-toethanol before loading. The gel was stained with Coomassie blue afterelectrophoresis. Each lane contained —‘16 jig protein from each sample.Plasmids present in the lanes were: 2, pRW3O2 (aa26); 3, pRW3O5 (aa131); 4,pRW3O6 (aa’35); 5, pRW3O9 (aa213); 6, pRW31O (aa215); 7, pRW3; 8, pTZ19R; 9,pRW3O9; 10, pRW3 10. The amino acid positions of the insertions are indicatedin the brackets. OprF monomer bands are indicated by triangles; the OprF SDSstable trimer is indicated by an arrow head; the position of OmpA is indicatedby a solid circle. Positions of relevant molecular mass markers (kDa) areindicated on the left. Due to the low level of expression, protein expressed bypRW3O5 (lane 3) was localized by Western immunoblotting.325I661234567849.5--32.5——‘IIb1Figure 12, Expression of OprF linker mutants.Western immunoblot analysis of outer membrane samples of E. coliDH5F’ strains expressing the pRW3 derived plasmids. The OprF-specffic mAbMA7-5, which recognizes an epitope that is not interrupted by the insertions,was used. Samples were heated at 100°C for 10 mm in solubilization bufferbefore loading. Each lane contained about 8 ig protein. Plasmids present were,lanes: 2, pRW3O2; 3, pRW3O5; 4, pRW3O6; 5, pRW3O9; 6, pRW31O; 7, pRW3;8, pTZ19R. Numbers on the left indicate the positions of the relevant molecularmass standards (kDa). OprF monomer bands are indicated by arrow heads.The position of the heatmodifled form of OprF is indicated by a circle. Bandscorresponding to oligomeri.c and LPSassociated forms of OprF are visible insome lanes.67The level of production of two of the linker mutants, encoded by pRW3O2and pRW3O6, was noticeably lower than that of the others, as determined by theirabundance relative to the other E. coli proteins (Fig. 11, lanes 2 & 4), indicatingthat insertions at these sites may lead to reduced protein production or unstableproducts. Mutants with insertion sites at the C-terminal end of the proteins (e.g.those encoded by pRW312 and pRW314) produced OprF variants that weresubstantially but not completely degraded to smaller fragments, including apredominant 28 kDa fragment (Table IV). This confirmed the results of Finnen etal. (1992), that the C-terminal regions of OprF were required for the resistance ofthe protein to cellular proteases.1.7 Monoclonal antibody reactivity of linker mutantsMost of the OprF-specffic mAbs available in the laboratory recognizedconformational epitopes and thus can be used as probes to examine the generalconformation of the OprF linker mutants. Outer membrane samples containing thelinker mutants were analyzed by Western and colony immunoblotting using theseries of OprF-specific monoclonal antibodies (Table VI). The results demonstratedthat the OprF derivatives expressed by 5 of the plasmids (pRW301, 302, 306, 309and 310) were reactive with all 10 monoclonal antibodies, indicating the retentionof native OprF structure. In 6 other mutants, specific OprF epitopes weredisrupted by the insertion of the 4-amino acid linker. However, the reactivity ofTableVI.Summaryofmonoclonalantibodyreactivityof OprFlinkermutants.MonoclonalantibodyreactivityaPlasmidInsertionsite(aaposition)7.47-27-37-47-57-67-77-84-45-8pRW3O1Gly-2w+++++++++pRW3O2Ala-26++++++++++pRW3O3Glu-42---------.-pRW3O5Ala-131-+++++++++pRW3O6Gly-135++++++++++pRW3O7Val-188+++++++-++pRW3O8Gly-196+++++++--+pRW3O9Arg-213++++++++++pRW31OGln-215++++++++++pRW311Ser-231++---+-+++pRW312Arg-290++---+-+++pRW314Gly-310+w---w-++-aMeasuredbycolonyimmunoblotandWesternimmunoblotanalysesof outermembranesamples.Symbols:+,reactivityequivalenttowildtypeOprFexpressedbypRW3;-,noreactivity; w,weakreactivity.0069these mutated proteins with the majority of the monoclonal antibodies suggestedthe retention of substantial native OprF structure in these mutants. OprFexpressed by pRW3O3 was an exception. Despite the fact that DNA sequencingdemonstrated that only 12 bp were inserted and no premature stop codon or changein reading frame occurred, it did not produce any OprF product that could bedetected by the OprF-specific monoclonal antibodies on immunoblots or visualizedin Coomassie blue stained SDS-PAGE gel of outer membrane samples and wholecell lysates. Thus it was assumed that this site was “non-permissive” for theinsertion of 4 amino acids.1.8 Membrane configuration of linker mutants in E. coilTo permit conclusions regarding the structure of OprF to be drawn basedon linker-insertion mutagenesis in E. coli, it was necessary to examine whether thestructure of OprF and its linker mutants in E. coli reflected that of OprF in P.aeruginosa. To further examine the configurations of the OprF derivatives in theouter membrane of E. coli, trypsin accessibility assays and indirectimmunofluorescence labeffing experiments were conducted.1.8.1 Trvnsin sensitivity assaysOuter membrane porins tend to be protease resistant (Paul andRosenbusch, 1985) by virtue of their extensive n-sheet structure with linking70surface loops that are tightly packed or folded in towards the porin channel (Weisset al., 1991; Cowan et al., 1992; Schirmer et al., 1995). It was previouslydemonstrated that purified OprF or OprF in outer membrane preparations arepartly cleaved by trypsin to a core 28 kDa fragment, and that increasingconcentrations of trypsin or increasing length of treatment time fails to causefurther proteolysis (Mutharia and Hancock, 1985).Trypsin treatment of outer membranes from E. coli DH5ciF’ expressingthe parental plasmid pRW3 resulted in substantial retention of full-sized OprF andpartial proteolysis to a 28 kl)a fragment that could be detected by the OprF-speciflcmAb 7-1 (Fig. 13, lanes 8, 9, and 10). Similar results were obtained after trypsintreatment of outer membranes from cells containing plasmids pRW3O2 andpRW3O6 (data not shown), which encoded OprF linker mutants with N-terminalinsertions in OprF (Fig. 13, lane 1, Table VII). Mutant proteins with C-terminalinsertions in OprF (i.e. those encoded by pRW3O9, pRW31O, pRW311 and pRW312)were completely cleaved to the 28 kDa fragment after trypsin treatment (Fig. 13,lanes 4-7; Table Vil). Proteins encoded by plasmids pRW3O7 and pRW3O8, whichcarried insertions in the central cysteine clisuiphide region (aa’88 and aa’96respectively), were cleaved by trypsin to a 24 kDa fragment, instead of (RW307;Fig. 13, lane 2) or in addition to (pRW3O8; Fig. 13, lane 3) the 28 kDa fragment.Based on previous studies (Finnen et al., 1992), such a 24 kDa fragment might beexpected if cleavage occurred near aa9° within the cysteine-containing region,suggesting localized modification of OprF by these insertions rendered this region71MWstd(kDc495-3a527.5...Figure 13. Trypsin sensitivity of linker mutants in outer membranes.Western immunoblot analysis of trypsinized outer membrane samplescontaimng OprF linker mutants. Samples were treated with trypsin (0.1 mg/mi)at 37°C forGO mm, and then heated at 88°C for 10 mm in solubilization buffer.The plasmids corresponding to the OprF linker mutants contained in each laneare: 1, pRW3O2; 2, pRW3O7; 3, pRW3O8; 4, pRW3O9; 5, pRW31O; 6, pRW311;7, pRW312; 8, pRW3; 9, pRW3; 10, pRW3 (untreated). OprF monomer band isindicated by an arrow head. The 28 kDa and 24 kDa trypsin-resistant corefragments are indicated by open and solid triangles, respectively. The OprF N-terminal specific inAb MA7- 1 was used for immunodetection. The positions ofrelevant molecular mass standards (kfla) are indicated on the left.472Table VII. Summary of trypsin sensitivity assays of OprF linker mutantsin E. coli outer membrane, DH5ii and 0386 whole cells.Plasmid Apparent Apparent mol. mass after trypsin treatmentamol. massa (kDa)(kDa) Outer DH5ft whole C386 wholemembrane cell cellpRW3 35 35, 28 35 28pRW3O1 41 ND ND NDpRW3O2 36 36, 28 36 36, 28pRW3O5 40b 24, 20, 18b NDpRW3O6 35 35, 28 35 NDpRW3O7 35 24, 35 35 24, 28pRW3O8 35 35, 24, 28 35 24, 28pRW3O9 35 28 35 28pRW31O 35 28 35,28 NDpRW311 36 28 ND NDpRW312 35 28 28 28pRW314 28C ND 28 NDa As estimated on Western immunoblot with MA7-1. Where more than oneband appeared, they are listed in order of abundance. ND, not determined.b Tested with MA4-4 since this mutant OprF derivative was non-reactive withMA7- 1.35 kDa was observed as a minor band.73susceptible to trypsin. Plasmid pRW3O5 (aa’31) expressed an OprF linker mutantthat showed unexpected trypsin-cleavage pattern with a predominant 24 kDaproduct and minor 20 kDa and 18 kDa products (Table VII). This implied asubstantial localized disruption of OprF structure, consistent with the observedsusceptibility to heat denaturation.Trypsin treatment of whole cells of E. coli DH5F’ expressing the wildtype or mutant oprF plasmids did not result in proteolysis of OprF or its linkermutant derivatives (Table VII), with the exceptions of those C-terminal insertionmutants encoded by plasmids pRW3 10, pRW3 12 and pRW3 14. On the other hand,the proteolysis patterns of OprF and the linker mutants in whole cells of C386(ompA, lpp) resembled that of these proteins in isolated outer membranes. Sincethe C-terminal of OmpA is highly homologous to that of OprF, the differences in theresults from these two host strains might have been due to the interaction of thisregion of OmpA and OprF, affecting the accessibility of the trypsin cleavage site inOprF and some of its mutants.1.8.2 Immunofluorescence labeffinTo examine whether the OprF linker mutants had surface-exposedregions, immunofluorescence labeffing was carried out using E. coli strainsexpressing selected linker mutant plasmids and monoclonal antibodies which bindto surface epitopes in the N-terminus (MA7-1), central region (MA7-8), and Cterminus (MA5-8) of OprF (Table VIII). To avoid the presence of OmpA, which74Table VIII. Results from indirect immunofluorescence labelling of intactE. coli C386 cells containing different plasmids.Plasmid Immunofluorescence with monoclonai antibothesaMA7-1 MA7-8 1VIA5-8no plasmid---pRW3 ++ ++ +pRW3O2 ++ ++ +pRW3O7 ++ + +pRW3O8 +- +pRW312 ++ ++ +a ++,positive labeffing; +, weak labelling; -,negative labelling.75might affect the accessibility of OprF epitopes, an OmpA-deficient strain (C386)was used as the background strain for the expression of the plasmids. In each case,regardless of the trypsin susceptibility of the respective OprF mutants in intactcells, immunoreactivity followed precisely the pattern observed in both colonyimmunoblots and Western immunoblots (Table VI). Taken together, the OprFlinker mutants were probably inserted in the outer membrane in the nativeconformation, as reflected by their trypsin resistance and surface exposure.1.9 SummaryThe results presented here demonstrated the identification of 11 unique“permissive” sites in OprF, which were sites that allowed the insertion of 4 extraamino acids without grossly affecting the production, folding and stability of theprotein. The characterization of OprF linker mutants provided compeffing evidencethat the general conformation and membrane configuration of these proteins in E.coli were very similar to that of OprF in P. aeruginosa. The evidence included: 1).The presence of the mutant proteins in purified E. coli outer membranes, 2). Theapparently correct formation of disulphide bonds as judged by the 2-mercaptoethanol modifiability of the OprF mutants, 3). The reactivity of theproteins with at least three of the mAbs MA7-3 through MA7-8 and MA4-4, whichapparently recognize conformational epitopes (Finnen et al., 1992; Rawling et al.,1995), 4). The demonstration of a trypsin-resistant core structure in most mutants76contained in outer membranes and the general resistance of OprF to trypsincleavage in intact cells, and 5). The correct surface localization of some of the OprFepitopes as examined by immunofluorescence labeffing of intact cells with OprFspecific monoclonal antibodies.The information obtained from this study established the“permissiveness” of the characterized linker-insertion sites in OprF, thus openingup the possibility that OprF could be used as a carrier for the presentation offoreign amino acid sequences. The 12 bp insertion in oprF resulted from the linkermutagenesis procedure provided a unique PstI site, which was useful for thecloning of foreign DNA sequences.77Chapter two: Construction, characterization and purification of OprF::malarialepitope and GST::malarial epitope hybrid proteins2.1 IntroductionEpitope-insertion studies have been used to demonstrate the potential ofouter membrane proteins as carriers for the expression of foreign antigenicdeterminants (Charbit et aL, 1991; Agterberg et aL, 1990b). The examination of thesurface exposure of the inserted epitope can also provide information about themembrane topology of the carrier protein. To further investigate the flexibility andlimitations of OprF as a carrier for epitope presentation, an epitope-insertion studywas carried out using the four-amino acid repeating epitope (NANP) of thecircumsporozoite protein of the human malarial parasite, Plasmodium falciparum,as a model epitope. In this chapter, the construction and characterization of twoseries of OprF::malarial epitope hybrid proteins, the positional and multiple-repeathybrids, will be described. In addition, the construction and characterization of twoversions of GST::malarial epitope fusion proteins, which were used to monitor theanti-malarial epitope response in serum samples from immunized animals, will alsobe described.782.2 Construction of OprF::malarial epitope hybrid proteinsPrevious linker-insertion mutagenesis had identified “permissive” sitesin OprF that can accommodate 4 extra amino acid residues. To further explore the“permissiveness” of the sites, a 10-amino acid malarial epitope was geneticallyinserted into these sites to generate a series of OprF::malarial epitope positionalhybrids. In addition, 4 different lengths of the malarial epitope were inserted into3 of the “permissive” sites to generate a series of multiple-repeat hybrids. Thesetwo series of OprF::malarial epitope hybrids not only helped to explore the“permissiveness” of the OprF insertion sites, they also provided a tool for the studyof the effects of insertion position and length of the epitope on epitope presentationin the OprF system.2.2.1 Positional hybridsThe previous linker-insertion mutagenesis study generated a series ofoprF linker mutants that carried a unique PstI site at different positions of thegene. To construct the series of OprF::malarial epitope hybrid proteins expressingthe epitope at different positions of OprF, oligonucleotides encoding the malarialepitope sequence (NANPNANPNA) were inserted into the PstI sites of the oprFlinker mutants. Three sets of oligonucleotides were required to accommodate thethree possible reading frames at the PstI sites. The positions and reading framesof the insertions were confirmed by DNA sequencing. Table IX summarizes the79Table IX. Summary of OprF::malarial epitope positional hybrids.Plasmid Insertion site Amino acids Surface exposureinserte& of the epitope’pRW3O2.1M Ala-26 PAP(ME)GHAGP +pRW3O2.2M Ala-26 PA{P(ME)GHA}2GPpRW3O6.2M Alal35b TC{NP(ME)CRS +pRW3O7.1M Val-188 DLQ(ME)LDVQV +pRW3O8.1M Ala-196 PAP(ME)GHAGP +pRW3O9.1M Arg-213 TCNP(ME)CRS +pRW3O9.3M Arg-213 TC{NP(ME)C}3RSpRW31O.1M Gln-215 DLQ(ME)LDVQV +pRW311.1M Ser-231 TCNP(ME)CRS +pRW311.5M Ser-231 TC{NP(ME)C}5RSepRW312.1M Arg-290 TCNP(ME)CRS +pRW312.4M Arg-290 TC{NP(ME)C}4RSpRW314.1M Gly-310 TCNP(ME)CRS +a ME=NANPNANPNA.b The site at aa’35 was also found to be permissive for the expression of twocopies of the epitope insert, but a hybrid that carried a single copy of theepitope was not obtained and therefore this site was not included in thisstudy.The numbers of insert in these cases were estimated by 2% agarose gel.d, the malarial epitope was detectable on the cell surface by indirectimmunofluorescence studies using a malarial epitope-specific monoclonalantibody.80series of OprF::malarial epitope positional hybrids and the identities of the insertedamino acid residues. The amino acid residues immediately flanking the malarialepitope varied according to the reading frame at the linker-insertion sites. In thecourse of cloning, hybrids that had incorporated multiple copies of the insert werealso isolated. However, due to the volume of work involved and the clarity ofpresentation, only two of these hybrids were chosen for further characterization.2.2.2 Multinle-reneat hybridsTo study the length effect on epitope presentation in the OprF system,OprF::malarial epitope hybrids with different lengths of the epitope insert wererequired. Three “permissive” sites in OprF (aa26, aa’96 and aa213) were selected forthe construction of such multiple-repeat hybrids. The choice of these sites wasbased on their positions in the protein (the N-terminus, the middle region and theC-terminus) and the stability of the corresponding positional hybrids. Table Xsummarizes the multiple-repeat hybrids constructed in this study. These hybridscarried 7, 11, 15 and 19 amino acids corresponding to the malarial epitope, eachwith an increment of one tetramer repeat. In addition, nine flanking amino acidresidues, which were the result of the previous linker-insertion mutagenesisprocedures and genetic cloning, were also added.81Table X. Summary of OprF::malarial epitope multiple-repeat hybrids.Insertion site Plasmid Amino acids insertedaAla-26 pRW3O2.7 PAARNPNANPNLDAGPpRW3O2. 11 PAAR(NPNA)2NPNLDAGPpRW3O2. 15 PAAR(NPNA)3NPNLDAGPpRW3O2. 19 PAAR(NPNA’)1NPNLDAGPMa- 196 pRW3O8.7 PAARNPNANPNLDAGPpRW3O8. 11 PAAR(NPNA)2NPNLDAGPpRW3O8. 15 PAAR(NPNA)3NPNLDAGPpRW3O8. 19 PAAR(NPNA)1NPNLDAGPArg-213 pRW3O9.7 TCTRNPNANPNLDCRSpRW3O9. 11 TCTR(NPNA)2NPNLDCR5pRW3O9. 15 TCTR(NPNA)3NPNLDCRSpRW3O9. 19 TCTR(NPNA)1NPNLDCRSa The amino acid residues corresponding to the malarial epitope areunderlined. The flanking amino acids PA_GP and TC_RS were the resultsof the previous linker-insertion mutagenesis procedures.822.3 Characterization of OprF::malarial epitope hybrid proteinsTo examine if the “permissive” sites previously identified by linker-insertion mutagenesis were “permissive” for the expression of the longer epitopesequence, the OprF::malarial epitope hybrid proteins were characterized in termsof their expression, cellular localization and reactivity with the series of OprFspecific monoclonal antibodies.2.3.1 Expression of hybrid vroteinsThe expression of the hybrid proteins was examined by Westernimmunoblotting of whole cell lysates of strains carrying the hybrid plasmids. Thehybrid plasniids containing the epitope-encoding oligonucleotides at eight differentsites expressed proteins that were reactive with both OprF-specific and malarialepitope-specific niAbs on Western immunoblots (Fig. 14). The apparent molecularmass of these proteins was slightly higher than that of native OprF, which wasconsistent with the presence of additional malarial epitope sequences in the hybridproteins. Plasmids pRW3O2.2M (aa26) and pRW3O9.3M (aa’96) encoded proteins ofhigher apparent molecular mass than plasmids carrying a single copy of the insertat the same sites (Fig.14, compare lanes 1 & 2, and lanes 5 & 6). The lanescorresponding to plasmids pRW3O7. 1M (aa’) and pRW3O8. 1M (aa’96) (lanes 3 and4) showed a more prominent upper band which apparently corresponded to theheat-modified form of the protein. This implied that the presence of extra aminoA83a1 2 3 4 5 6 7 8 91011129 101112BkDa1 2 3 4 567 8WI I—.j - —- —Figure 14. Expression of OprF::ma].arial epitope positional hybrids.Western immunoblots of whole cell lysates of E. coli DH5F’ strainsexpressing various OprF::malarial epitope hybrid proteins after reaction withA) an OprF specific monoclonal antibody MA7- 1 and B) a malarial epitopespecific mAb pf2A. 10. Samples were resuspended in solubilization buffer andheated at 100°C for 10 mm before loading. Plasmids expressed in the samplesin lanes were: 1, pRW3O2.lM; 2,pRW3O2.2M; 3, pRW3O7.1M; 4, pRW3O8.1M;5, pRW3O9.1M; 6, pRW3O9.3M; 7, pRW31O.lM; 8, pRW311.1M; 9, pRW312.1M;10, pRW3l4.1M; 11, pTZ19R; 12, pRW3. The bands corresponding to wild typeOprF and the N-terminal degradation product are indicated by a solid and anopen triangle, respectively. In some lanes, the bands corresponding to OprFdimers, oligomers and protease degradation products are visible. The positionsof relevant molecular mass standards (kDa) are indicated on the left.84acids in the cysteine-containing region of OprF might have affected the localconformation, and thus rendered the protein more susceptible to heat denaturation(Hancock and Carey, 1979). Lanes 7 to 10 demonstrated an increase in abundanceof the 28 kiJa degradation product, which failed to react with the malarial epitopespecific mAb, suggesting that it represented the N-terminal part of these proteinslacking the malarial epitope sequences. This result was consistent with previousfindings that C-terminal perturbations rendered these OprF derivatives moresusceptible to cellular proteases (Finnen et at., 1992).Oligonucleotides encoding the malarial epitope were also inserted intosites corresponding to aa2 and aa’3’ of OprF, but no hybrid proteins were detectedon Western immunoblots. These two sites appeared to be either “non-permissive”for the insertion of more than 4 amino acids (encoded by the linker insertion) or“non-permissive” for the expression of the malarial epitope sequence.2.3.2 Cellular localization of hybrid vroteinsThe insertion of extra amino acid residues into the mature sequence of amembrane protein affects its primary and possibly secondary structure andtherefore might interfere with its transport to the native subcellular compartment.SDS-PAGE analysis of outer membrane preparations from strains carrying thehybrid plasmicis was performed to examine the outer membrane localization of theOprF::malarial epitope hybrid proteins. It was shown that the hybrid proteins wereassociated with the outer membranes (Fig. 15). Due to the genetic background of856 7 8 91011128 9101112.4——..z.-. AFigure 15. Cellular localization of OprF::malarial epitope positional hybrids.A. Sucrose gradient outer membrane preparations of E. coli C158 (ompA, ompC,phoE) strains expressing OprF::malarial epitope positional hybrid proteins.Samples were heated at 100°C for 10 mm in solubilization buffer before loading.The gel was stained with Coomassie blue after electrophoresis. B. Westernimmunoblot analysis of the same samples with MA7- 1. Plasmids expressed inthe strains were: 1, pTZ19R; 2, pRW3; 3, pRW3O2.1M(aa26) ; 4,pRW3O2.2M(aa26); 5, pRW3O7. 1M(aa’88); 6, pRW3O8. 1M(aa’96); 7,pRW3O9. 1M(aa213); 8, pRW3O9.3M(aa21); 9, pRW3 10. 1M(aa215); 10,pRW3 11. 1M(aa231); 11, pRW3 12. 1M(aa290); 12, pRW3 14. 1M(aa310). The aminoacid positions of the insertion sites are in brackets. The positions of relevantmolecular mass standards (kDa) are indicated on the left. The positions of thenative and heat-modified forms of the proteins are indicated by solid and opentriangles respectively. The band corresponding to an E. coli outer membraneprotein is indicated by a solid circle. Bands corresponding to OprF oligomersare visible in some lanes.12345AB 1234567— — — —53.2—34.9—86the host strain (ompA, ompC, phoE) and the growth conditions of the bacterialcultures used for the outer membrane preparations (high osmolarity in the presenceof 0.1% glucose to suppress the production of OmpF and LamB respectively), thehybrid proteins appeared to be one of the major species present in the preparations(Fig. 15). The slight differences in apparent molecular mass of the hybrids mighthave been due to the interactions between the inserted amino acid residues and thelocal OprF amino acid sequence which affected the electrophoretic mobilities ofthese proteins. Western immunoblot analysis of the same outer membranepreparations revealed minor and comparable levels of protease degradationproducts in all of the samples, implying that the 28 kDa product observed in thewhole cell lysates was not associated with the outer membrane (Fig. 15B).Similar to the positional hybrids, the OprF::malarial epitope multiple-repeat hybrids were also expressed in the outer membrane of E. coli (Fig. 16). Thelength increment of the inserted epitope in each set of the hybrids was reflected bythe stepwise increase in the apparent molecular mass of the hybrid proteins. Theseries of multiple-repeat hybrid proteins carrying insertions at aa26 formedinclusion bodies which fractionated with the outer membrane. The bandscorresponding to the inclusion body form of the hybrid proteins migrated at higherapparent molecular mass than the membrane bound form. In addition, while themembrane bound form of the proteins was 2-mercaptoethanol modifiable due to thepresence of disuiphide bonds, the gel mobility of the inclusion body form of theproteins was not affected by 2-mercaptoethanol, indicating the absence of87AB53.2-II1 2 3 4 534.9 %. —.—53.2 -34.9 -1 2 3 4 5Figure 16. Expression of OprF::malarial epitope multiple-repeat hybrids.SDS-7.5%PAGE outer membrane preparations carrying insertions at A).Ala-26 and B). Arg-213 of OprF. Samples were heated at 37°C for 10 mm insolubilization buffer before loading. The gels were stained with Coomassie blueafter electrophoresis. Lane 1, OprF with no insert; lanes 2 to 5 representhybrids carrying 7, 11, 15 and 19 amino acids of the epitope respectively.Samples corresponding to the aa26 hybrids were obtained by octyl-POEextraction of outer membrane samples. The positions of relevant molecular massstandards (kDa) are indicated on the left. Bands corresponding to OprF or OprFhybrid proteins are marked with solid circles.-4-.--88clisulphide bonds in these proteins (Fig. 17).2.3.3 Surface exuosure of the evitopeAccording to the membrane topology model of OprF, most of the malarialepitope insertion sites were proposed to be in the surface-exposed loop regions ofOprF. In other words, the inserted epitope should have been detectable on the E.coli cell surface. Previous studies have shown that the presence of OmpA, an E. coliouter membrane protein that shares C-terminal, homology with OprF (Duchene etal., 1988; Woodruff and Hancock, 1989), appeared to mask the binding of mAbs toOprF in intact E. coli cells (Martin et al., 1993). Therefore, an OmpA-deficientstrain C386 was chosen for the expression of the hybrid plasmids. The surfaceexposure of the malarial epitope was examined by indirect immunofluorescencelabelling of whole cells containing the hybrid proteins with the malarial epitopespecific monoclonal antibody, followed by a secondary antibody that was conjugatedto a fluorescent dye (Fig. 18). Due to the limitation of the instruments used influorescence microscopy, only the cells that were at the same depth of fieldappeared fluoresced in Figure 18. However, examination of the slides by varyingthe depths of field showed that the majority of cells (>90%) were labelled. Themalarial epitope expressed at all eight “permissive” sites was detectable on the cellsurface (Table IX), which suggested the placement of the insertion sites in thesurface-exposed ioop regions in the topology model of OprF. These results wereconsistent with the general assumption that surface loop regions are more likely to891 2 3 4 5 6 7 8 910111213Figure 17. Presence of inclusion bodies in outer membrane samples.SDS-PAGE of sucrose gradient outer membrane preparations containingboth the membrane bound and inclusion body forms of the OprF::malarialepitope multiple-repeat hybrid proteins carrying the inserted epitope at aa26.Samples were heated at 100°C for 10 mm in solubilization buffer without (lanes1 to 6) or with 4% 2-mercaptoethanol (2-ME) (lanes 8-13) before loading. Thegel was stained with Coomassie blue after electrophoresis. Samples containedin the lanes: 1 & 8, outer membrane samples from strain expressing pTZ 19R;2 & 9, OprF; 3 & 10, OprF::lVIE7aa26; 4 & 11, OprF::MEllaa26; 5 & 12,OprF::ME15aa26; 6 & 13, OprF::MEl9aa26; 7, molecular mass standards. Solidand open circles indicate the positions of the inclusion body and membranebound forms of the proteins respectively. The positions of relevant molecularmass standards (kDa) are marked on the left.L_ 2-ME -—.i L.... 2-ME —J90A.-L)B‘ 1,1 1 ••. j_ —s 4 ‘ ,U.. .14 rj. b ,It‘:• i•-‘‘-‘-:•:Ha4 , I ;(‘ • r‘..: -‘i.;;d’’’ —..,.4_*.4.1 . I I______________:Figure 18. Surface exposure of the malarial epitope.Indirect immunofluorescence labelling of C386 (pRW3O2.1M) with themalarial epitope-specific monoclonal antibody pf2A. 10. A) Labelled cellsobserved under fluorescence microscopy. B) the same field observed underphase contrast. The scale bar at the bottom left corner indicates 20im. Thedark box at the bottom right corner was an artefact from the scale bar slider.91be flexible enough to accommodate foreign peptide sequences (Hofnung, 1991).2.3.4 Monoclonal antibody reactivity of hybrid vroteinsThe insertion of foreign amino acid residues might affect the localconformation of the protein and hence disrupt the antibody binding sites of OprF.To examine the effect of epitope insertion in OprF on antibody binding, the outermembrane samples containing the hybrid proteins were analyzed by Westernimmunoblotting with ten OprF-specific mAbs and two malarial epitope-specificmAbs (Table XI). In general, the pattern of mAb reactivities of the OprF::malarialepitope hybrid proteins was similar to that of the OprF linker mutants. However,while the linker insertion at aa’88 disrupted only the MA7-8 epitope, the epitopeinsertion at the same site disrupted the MA4-4 epitope in addition to the MA7-8epitope. This difference suggested that the insertion of a longer amino acidsequence caused more extensive disruption of the local secondary structure, thusdestroying both the MA4-4 and MA7-8 epitopes. Moreover, while the linkerinsertion at aa31° (encoded by pRW3 14) disrupted the binding of MA7-3 and MA7-5,epitope insertion at the same sites restored weak reactivities with these two mAbs,suggesting that the insertion of the malarial epitope might have restored theantibody binding site(s) to a conformation that resembled to that of the native OprFenvironment. This speculation was consistent with previous conclusions that thesetwo monoclonal antibodies recognize conformational epitopes (Rawling et at., 1995).Malarial epitope inserted at all eight sites was recognized by the two malarialTableXI.SummaryofmonoclonalantibodyreactivityofOprF::malarial epitopehybridproteinscontainedinoutermembranesamplesInsertionMonoclonalantibodyreactivityaPlasmidssites(aaposition)IVIA7-1IVIA7-2MA7-3M.A7-4MA7-5MA7-6MA7-7MA7-8MA4-4MA5-8pf2A.1Opf5A4.1pRW3-++++++++++--pRW3O2.1MAla-26++++++++++++pRW3O2.2MA1a-26++++++++++++pRW3O7.1MVal-188+++++++--+++pRW3O8.1MGly-196+++++++--+++pRW3O9,1MArg-213++++++++++++pRW3O9.3MArg-213++++++++++++pRW31O.1MGln-215++++++++++++pRW311.1MSer-231++www+w+++++pRW312.1MArg-290++w--+-+++++pRW314.1MGly-310++w-w+-++-++aDeterminedbyWesternimmunoblotanalysesofoutermembranesamples.Symbols:+,reactivityequivalent towildtypeOprFexpressedbypRW3;-,noreactivity; w,weakreactivity.93epitope-specific mAbs.Western immunoblot analysis of the outer membrane samples containingthe multiple-repeat hybrids with an OprF-speciflc polyclonal antiserumdemonstrated that the increase in length of the insert did not increase the amountof degradation products in the outer membrane (Fig. 19A). The hybrid proteinsremained reactive with the OprF-specific mAbs MA7-1, MA7-2, MA7-6, 1VIA7-8(except for the series carrying insertions at aa’96) and 1VIA5-8, of which 1VIA7-6 andMA7-8 recognized conformational epitopes, indicating that the overall secondarystructure of OprF was stifi conserved in these hybrid proteins (Fig. 19B).2.4 Purification of OprF::malarial epitope hybrid proteinsPurified forms of the hybrid proteins could be useful as antigens for thestudy of antibody binding and as immunogens for immunogenicity studies.Therefore, efforts were made to establish a protocol for the purification of OprF orOprF: :malarial epitope hybrids from E. coli. To simplify the purificationprocedures, the E. coli strain C158 (ompA, ompC, phoF) was used as thebackground strain for the purification.2.4.1 Induction experimentsSince the expression of oprF was under the control of the lac promoterfrom the cloning vector, different IPTG induction conditions were investigated to94aa26 aalS6 aa213no. of aa ——%ins.rt.d 711 1619 7111519 7111519 wt—50-325.B .a26 aal9S aa213no. of aa—‘ —ns.rt.d 71115197It151g7111519“I,Ilk50—a.,. —.32.5Figure 19. Western immunoblots of OprF::malarial epitope multiple-repeathybrids with A) an OprF-specific polyclonal antibody and B) an OprF-specificmAb MA7-6. Each set of four lanes represented hybrids carrying an insertionat the site indicated; wt, OprF with no insert. Samples carrying the aa26 hybridswere supernatants from 3% octyl-POE extraction of outer membrane samples,while samples carrying the aa’ and aa213 hybrids were sucrose gradient outermembrane preparations. The percentages of acrylamide used in the SDS-PAGEprior to Western transfer in A and B are 11% and 9% respectively. Sampleswere heated at 10000 for 10 mm before loading. Bands corresponding tooligomeric and LPS-assocjated forms of OprF are visible in some lanes. Theposition of OprF is indicated by a•. The positions of relevant molecular massstandards (kDa) are indicated on the left.,95maximize the level of OprF production. Figure 20 shows the amount of OprFhybrid proteins in the outer membrane preparations of cultures grown underdifferent induction conditions. Induction at 30°C for 3 h with 0.2 mM or 1 mM ofIPTG did not increase the expression level significantly (lanes 3 and 4). On theother hand, induction with 1 miVi IPTG at 37°C or with 0.2 mM IPTG for 16 h at30°C increased the expression level considerably (lanes 5 to 7). Prolonged inductionappeared to lead to higher levels of protein production and degradation (lane 7).2.4.2 Detergent extractionsThe purification protocol for OprF was as described in Methods andmaterials section 11.1. Outer membrane samples prepared by sucrose gradientmethod were extracted with detergent in the presence of NaCl and EDTAsequentially. The three detergents tested, Triton X-100, Zwittergent 3-16 and octylPOE, gave similar results. Octyl-POE was chosen because of its mild nature whichwas hypothesized to preserve the structure of OprF after extraction. Figure 21shows a Coomassie blue-stained gel of samples from the sequential detergentextraction procedures. The addition of EDTA released the tightly bound outermembrane proteins including OprF or the OprF hybrid proteins (lanes 7&8). Theamount of the OprF hybrid protein in the residual insoluble fraction wassignificantly reduced (lane 9) as compared to that in the initial outer membranepreparations (lane 3).96A 1234567325:1111114B 3456749.5—l32,5—aFigure 20. Expression of an oprF derivative in different induction conditions.A. SDS-PAGE of outer membrane samples containing the hybrid proteinOprF::MElOaal88. Samples were heated at 37°C for 10 mm before loading.The gel was stained with Coomassie blue after electrophoresis. B. Westernimmunoblot of the same samples with the OprF-specific mAb MA5-8. Sampleswere heated at 100°C for 10 mm before loading. Lanes: 1, molecular massstandards; 2, un-induced; 3, 0.2 mM IPTG at 30°C for 3 h; 4, 1mM IPTG at 30°Cfor 3 h; 5, 1mM IPTG at 37°C for 3 h; 6, 1mM IPTG at 37°C for 6 h (i. e. addedat the time of inoculation); 7, 0.2mM IPTG at 30°C for 16 h. IPTG was added tothe cultures during the logarithmic growth phase unless otherwise stated. Theposition of the hybrid protein is indicated by an arrow head. The positions ofrelevant molecular mass standards (kDa) are indicated on the left.9749.5032.5_____Figure 21. Purification of OprF::malarial epitope hybrid proteins.SDS-PAGE analysis of samples from different steps of the purification ofthe OprF hybrid protein from C158 expressing pRW3O7.1M. The gel wasstained with Coomassie blue after electrophoresis. Lanes: 1, molecular massstandards; 2, whole cell lysates; 3, sucrose gradient outer membrane fraction;4, supernatant from 0.5% octyl-POE extraction; 5 & 6, supernatants from twosequential 3% octyl-POE/1M NaC1 extractions; 7 & 8, supernatants from twosequential 3% octyl-POE/lOmM EDTA extractions; 9, insoluble fraction of 3%octyl-POE/lOmM EDTA extraction. The positions of relevant molecular massstandards (kDa) are indicated on the left, The position of the OprF hybridprotein is indicated by an arrow.123 56789F -:-— ,4—982.4.3 FPLC nurificationBased on our laboratory’s experience of OprF purification from P.aeruginosa, a single chromatography step with an anion exchange column by FPLCwas introduced and found to be sufficient to yield purified OprF. Therefore, theOprF hybrids contained in the solubilized fraction after 3% octyl-POE/lOmM EDTAextraction were further purified by FPLC. In these experiments, OprF did not bindto the anion exchange column while the major contaminants in the samples boundand were eluted at a NaCl concentration of 0.3 M and 1 M respectively. Figure 22shows the FPLC proffle of the samples eluted from MonoQ column, indicating thatOprF was the predominant species in the flow through fractions.2.4.4 Purification of inclusion body-contaminated outer membrane nreparationsAttempts to prevent the formation of inclusion bodies in strainsexpressing the multiple-repeat hybrids at aa26 by growing the cultures at 30°C andharvesting at early-log growth phase were unsuccessful. Detergent extraction wastherefore used to isolate the membrane bound protein from the inclusion bodies(Piers et aL, 1993). It was found that extraction with 3% octyl-POE and 10mMEDTA selectively released the membrane bound protein into the supernatant whileleaving the inclusion body form in the insoluble fraction (Fig. 23).99A 1.0 P1.2. A2NaCI(M)L0.00 20 40 60 80 100F ra ct) 0 fl $B 12 34 56- 32.5*LFigure 22. FPLC profile of a MonoQ column separation of the octylPOE/EDTA soluble OprF hybrid expressed by pRW3O7. 1M.A. Elution proffle with a NaC1 gradient. B. SDS-PAGE of samples correspondingto the peak fractions. Lanes: 1 & 2; 3 & 4; 5 & 6; correspond to samples from thefirst, second and third peaks (from left to right) respectively. The position of theOprF hybrid protein is indicated by an arrow. The positions of relevantmolecular mass standards (kDa) are indicated on the right.100.1J•0325-12345Figure 23. Removal of inclusion bodies from outer membrane preparations byoctyl-POE extraction.SDS-PAGE of OprF::ME 19aa26..contairnng samples from the extraction.The gel was stained with ()oomassie blue after electrophoresis. Lanes: 1, outermembrane sample from a sucrose. gradient separation; 2, supematant from 0.5%octyl-POE extraction; 3 & 4, supernatants from the two sequential 3% octylPOE/lOmM EDTA extractions; 5, insoluble fraction after octyl-POE extractions.Solid and open circles indicate the positions of the inclusion body and membranebound forms of the protein respectively. The positions of relevant molecularmass standards (kDa) are marked on the left.I1012.5 GST::malarial epitope fusion proteinsNot only are synthetic peptides costly, they quite often do not interactefficiently with antibodies. Therefore, the genetic construction of 2 versions ofGST::malarial epitope fusion proteins was undertaken to provide a source of theepitope for the detection of anti-malarial epitope antibodies in the sera ofimmunized animals in the later stage of this study.2.5.1 Construction and purification of fusion proteinsThe fusion proteins were constructed by cloning hybridizedoligonucleotides encoding the malarial epitope into the 3’ end of the GST codingregion in the cloning vector pGEX-1N (Fig. 8). The resulting fusion proteins,GST::1VIE11 and GST::ME19, expressed 11 {P(NANP)2N and 19 {P(NANP)4N }amino acids respectively corresponding to the epitope at the C-terminus of thecarrier protein. The fusion proteins were expressed in E. coli DH5ci and purifiedby affinity chromatography using glutathione agarose beads as described by Smithand Johnson (1988). Figure 24 shows that both fusion proteins had higherapparent molecular mass than GST, in agreement with the presence of the extramalarial epitope amino acid residues. Both fusion proteins were recognized by ananti-GST polyclonal antiserum and the malarial epitope-specific mAb pf2A. 10 onWestern immunoblots (Fig. 24).BA123M 123M— so-3”— ,zS102Figure 24, Purification of QST::malarial epitope fusion proteinsA. SDS-PAGE of GST::malarial epitope fusion proteins purified by affinitychromatography. The gel was stained with Coomassie blue afterelectrophoresis. B. Western immunoblot of the affinity-purified protein with ananti-GST polycional serum. C. the same blot reacted with the malarial epitopespecific mAb pf2A.l0. Lanes: 1, GST::ME19; 2, GST::ME11; 3, GST; M,molecular mass standards. Samples were heated at 100°C for 10 mm insolubjjjzatjon buffer before loading. The positions of the relevant molecularmass standards (kDa) are indicated on the right.1032.5.2 Binding of fusion uroteins with euitone-snecific monoclonal antibodiesSince the GST::malaria]. epitope fusion proteins were constructed for useas coating antigens in ELISA to determine the anti-malarial epitope titers in thesera of immunized anima]s, the level of binding of these proteins with the epitopespecific mAbs in ELISA was also examined. As extrapolated from the bindingcurves, GST::ME11 had Km values of 1.25 jig/mi and 0.5 j.tg/ml for the epitopespecific mAbs pf2A. 10 and pf5A4. 1 respectively, whereas GST::1VIE 19 had Km valuesof 0.83 jig/ml and 0.35 jig/mi for pf2A.10 and pf5A4.1 respectively (Fig. 25). Thissuggested that GST::ME 19 had a slightly higher affinity for the malarial epitopespecific monoclonal antibodies, which might have been due to the presence of thelonger epitope sequence permitting more appropriate folding.2.6 SummaryIn this chapter, it was demonstrated that OprF could be used as a carrierfor the expression and surface exposure of a malarial epitope. Two series ofOprF::malarial epitope hybrids were constructed genetically. The first series, thepositional hybrids, consisted of OprF hybrid proteins that expressed the malarialepitope at different permissive sites within OprF. The second series, the multiple-rep eat hybrids, consisted of hybrids that expressed four different lengths of therepeating epitope at one of the three selected sites in OprF. Eight “permissive”sites were identified which could accommodate and express the model malarialA2A450 10C2A450 100123456789Conc. of coating antigen (uglmi)0123456789Conc. of coating antigen (uglmi)B2A4500D2A45000123456789Conc. of coating antigen (ug!mI)0123456789Conc. of coating antigen (uglmi)104Figure 25. Binding of GST::màlarial epitope fusion proteins with epitopespecific monoclonal. antibodies.ELISA of GST::ME11 with pf2A.l0 (A) and pf5A4.1 (B); and ELISA ofGST::ME19 with pf2A.10 (C) and pf5A4.1 (D). Curves represent differentdilutions of the antibodies: +, 1:1000; , 1:5000; 0, 1:25000; 4, 1:125000; A,1:625000.‘ZZZ-A - i—z—-4- - -.- =-r —-A -A- - —- - -- - rr ---- 4105epitope sequence. Insertion of the epitope sequence in the cysteine-containingregion of OprF increased the heat sensitivity of the hybrid proteins, whileinsertions in the C-terminus of the protein rendered the hybrid proteins moresusceptible to degradation by cellular proteases. All of the hybrids were expressedin the outer membrane and the inserted epitope at each of the “permissive” siteswas detectable on the cell surface. Western immunoblot analysis of the hybridproteins with the series of OprF-specific mAbs indicated that the proteins retainedsubstantial wild type conformation. Furthermore, a protocol for the purification ofOprF or OprF hybrid proteins from E. coli was also established. In addition to theOprF::malanal epitope hybrids, two versions of GST::malarial epitope fusionproteins were also constructed. Both fusion proteins were reactive with the twoepitope-specific mAbs tested.The availability of the two series of OprF::malarial epitope hybridproteins provided a set of tools for the study of the effects of insertion position andlength of the epitope on epitope presentation in the OprF carrier system, while theGST::malarial epitope fusion proteins represented a source of easily-purifiedepitope for the analysis of anti-epitope response in serum samples from immunizedanimals.106Chapter three: Effects of mode of presentation on antigenicity andimmunogenicity.3.1 IntroductionLimited studies of two of the outer membrane protein epitopepresentation systems have shown that the nature of the flanking amino acidsequences and the length of the inserted epitopes could influence the antigenicity(i.e., the ability to interact with antibody) and the immunogenicity (i.e. the abilityto stimulate an immune response) of the epitopes (Agterberg et al., 1990b; Van derWerfet al., 1990). These findings suggested that more extensive investigations ofthe position and length effects of epitope insertion in carrier proteins will help usto exploit the effectiveness of such presentation systems.The potential of OprF as a carrier protein for the presentation of a foreignmalarial epitope was clear from the results described in the previous chapter. Tofurther our understanding of the flexibility and limitations of the OprF system, thetwo series of OprF::malarial epitope hybrid proteins were used to examine theeffects of the insertion position and the length of the epitope on epitopepresentation in the OprF system. In this chapter, a broad survey of antigenicityof the inserted epitope is described and this led to a targeted study ofimmunogenicity of the epitope.1073.2 Antigenicity studyThe antigenicity of a molecule refers to its ability to interact withantibodies. In the OprF epitope presentation systems, the accessibility of theinsertion sites, the nature of flanking amino acid residues and the length of theinserted epitope might affect the interaction of the epitope with its specificantibodies. Therefore, a study was undertaken to compare the antigenicity of theepitope presented in different lengths and at different positions of OprF.3.2.1 AunroachesIn the process of establishing an assay to evaluate the antigenicity of theepitope, a whole cell ELISA and an antigen competition assay were employed. Inthe whole cell ELISA, E. coli cells expressing the hybrid proteins were used as thecoating antigens to capture a malarial epitope-specific monoclonal antibody. Thelevels of malarial epitope-specific antibody binding to the various strains of K colicells were quantified and used as measurements for the antigenicity of thepresented epitope. In the antigen competition assay, E. coli cells expressing thehybrids were used to adsorb the malarial epitope-specific antibody. The residualtiters of the epitope-specific antibody were then measured by ELISA using theGST::malanal epitope fusion protein, GST::ME19, as the coating antigen. Theamount of malarial epitope-specific antibody adsorbed by the various strains of E.coli cells was used as an indication of the antigenicity of the epitope in the108corresponding OprF::malarial epitope hybrid proteins. However, neither assaygave consistent results, presumably due to the high backgrounds in control bacterialacking the malarial epitope. This background could be caused by the non-specificbinding of the epitope-specific monoclonal antibodies, pf2A. 10 and pf5A4. 1, tosurface components on E. coli cells in these assays. It was found that pf5A4. 1 wasmore highly reactive with E. coli outer membrane preparations in ELISA thanpf2A. 10. Therefore, only pf’2A. 10 was used in the antigenicity assays. The relativeantigenicity of the inserted epitope was measured by whole cell dot blot analysisand outer membrane ELISA using the epitope-specific mAb pf2A. 10 and an OprFspecific polyclonal antiserum. These assays were quite reproducible and thebackground readings were low under the conditions utilized. Binding titrationcurves of the two antibodies to OprF and an OprF::malarial epitope hybrid proteinare shown in Figure Position effectThe antigenicity of single copies of a 10-amino acid malarial epitope(NANPNANPNA) expressed at different positions of OprF was compared by wholecell dot blot analysis and ELISA using outer membranes. To take into account thevarious expression levels of the hybrid proteins in the outer membrane, theantigenicity index of each inserted epitope was calculated as the ratio of the antiepitope reactivity to the anti-OprF reactivity of the corresponding hybrid protein.Both assays indicated that the epitope had different relative affinities for the109A_________________________— __.A —/ —A450 7’0.50//A1/1A0.000 5 10 15 20Conc. of protein (uglmi)B 1.007’A -4500.50 7VV0.00 I I0 5 10 15 20Conc. of protein (ugimi)Figure 26. Binding of an OprF-specffic polyclonal serum (A) and the malarialepitope-specific mAb pf2A. 10 (B) with outer membranes from E.coli expressing the following OprF variants. Symbols: •, vectorcontrol; A, OprF; 4, OprF::ME7aal96.110malarial epitope-specffic mAb pf2A. 10 when expressed at different positions inOprF (Fig. 27). However, generally siniilar antigenicity patterns were observed forgiven mutants in the context of both whole cells and isolated outer membranes. Forexample, the malarial epitope inserted at aa215 or aa31° consistently demonstratedlow relative antigenicity, whereas insertions at aa’ or aa’96 were significantlymore antigenic in both assays. In contrast, the epitope inserted at aa26 was moreantigenic in whole cell dot blot analysis, while the epitope inserted at aa213 andaa29°was significantly more antigenic when assayed by outer membrane ELISA.The dissimilarity could conceivably be due to the difference in the presentation ofthe epitope in whole cells as compared to an outer membrane environment.Although the hybrid proteins were likely to be in their native configuration in theouter membrane preparations, the isolation procedures might have removed partof the surface moieties such as lipopolysaccharides (LPS). The presence of cellsurface LPS could promote the presentation of the inserted epitope for antibodybinding at aa26 while reducing the accessibility for antibody binding at aa213 andaa290. As a result, significant dissimilarities in antigenicity indices were observedwhen assayed in whole cell and outer membrane environments. The antigenicityindices of the epitope in the positional hybrids are summarized in Table XII.Due to the cloning procedures, the flanking sequences of the epitopesinserted at aa213, aa231, aa29° and aa31° contained cysteine residues, indicating thatthey might participate in disulphide bond formation. However, results frompreliminary survey showed that the antigenicity of the epitope inserted at aa213 was1112015xC)>: 10CC)0,C50(0 C’ 10 0 0(0 O ‘ C) O i1 CJ C1 CI CM C)m c uPosition of insertionFigure 27. Effect of insertion position on the antigenicity of the malarialepitope.Solid and hatched bars represent results from outer membrane ELISA andwhole cell dot blot analyses respectively. To allow comparisons between the twomethods, the antigenicity index of the epitope inserted at aa’96 was used as aninternal standard and arbitrarily set to 10 and the rest of the values wereadjusted accordingly. Values were the means and standard deviations from sixindependent experiments for outer membrane ELISA and three independentexperiments for dot blot analyses. The indices that were discussed in theResults section as being significantly different were confirmed by F-tests(P<0 .05).112Table XII. Summary of antigenicity of malarial epitope in the OprF: :malarialepitope positional hybrids.Antigenicity index0. Insertion Amino acidsPlasmid a bsite inserted Outer Whole cellmembranepRW3- wt 0 0pRW3O2.1M Ala26 PAP(MGHAGP 7.06±0.58 16.49±1.24pRW3O2.2M Ala26 PA{P(MGHA}2GP 15.58±1.19 25.60±3.05pRW3O7.1M VaT’88 DLQ(MLDVQV 7.18±1.03 10.28±3.22pRW3O8.1M Ala’96 PAPQ1GHAGP 10 10pRW3O9.1M Arg213 TCNPQf.CRS 16. 14±1.82 5.84±0.32pRW3O9.3M Arg213 TC{NP(MEC}3RS 20.23±1.91 12.82±2.44pRW31O.1M Gin215 DLQMLDVQV 3.52±0.63 3.93±0.14pRW311.1M Ser23’ TCNP(MCRS 7.91±0.74 6.57±0.51pRW312.1M Arg29° TCNP(MCRS 10.01±0.67 4.74±0.21pRW314.1M G1y3° TCNPQCRS 4.25±0.31 3.83±1.42a Position of the amino acid preceding the insertion. At insertion sites Ala’96 andSer231, the preceding amino acids were replaced by a glycine and argininerespectively.b= NANPNANPNA, the outer two amino acids on both sides of the flankingsequences (i.e. PA_GP, DL_QV or TC_RS) were the results of the previouslinker-insertion mutagenesis procedures.Results are presented as means±standard deviations as described in Figure 27legend. Antigenicity indices were calculated as described in Methods andmaterials section 12.113not affected by the presence of 2-mercaptoethanol, suggesting that the flankingcysteine residues were likely not be involved in disulphide bonding, andlor thatclisuiphide bonding did not affect the presentation of the epitope for antibodybinding.3.2.3 Length effectThe antigenicity of insertions containing 2 and 3 copies of the 10-aminoacid epitope at aa26 and aa213 was also measured. Insertion of multiple copies of theepitope at both sites resulted in a significant increase in antigenicity (Table XII,rows 2&3; 6&7). The hybrid carrying 2 copies of the epitope insert at aa26 (encodedby pRW3O2.2M) consistently demonstrated an antigenicity index which wasapproximately two-fold higher than that of the hybrid expressing one copy of theinsert at the same site (encoded by pRW3O2. 1M), indicating that the presence of anadditional copy of the epitope insert enhanced the ability of the inserted epitope tobind antibodies (Fig. 28). Similarly, the hybrid expressing three copies of the insertat aa213 (encoded by pRW3O9.3M) also demonstrated higher antigenicity than thehybrid expressing one copy of the insert at the same site (encoded by pRW3O9. 1M)(Fig. 28). The presence of an additional copy of the epitope might have improved theexposure of the epitope and thus its accessibility for antibody binding, or it mighthave increased the valency of antibody binding. The lesser influence of multipleinsertions at aa213 on the antigenicity might imply that this site is already relativelywell-exposed in the OprF protein, or that the effects of multivalency in this case11430 -25‘CV>1: 15Ca)0)1050Plasmid expressed in strainFigure 28. Effect of insertion of multiple copies of the malarial epitope onantigenicity at insertion sites aa26 and aa213 of OprF.Plasmids pRW3O2. 1M and pRW3O2.2M encoded hybrid proteins carryingone and two copies of the epitope at aa26 respectively; plasmids pRW3O9. 1M andpRW3O9.3M encoded hybrid proteins carrying one and three copies of theepitope at aa213 respectively. Solid and hatched bars represent results fromouter membrane ELISA and whole cell dot blot analyses respectively. Valuespresented are as described in Table XII foot notes (c).-IC%I115might have induced steric hindrance which limited the accessibility of the epitopeto antibody. In light of this finding, hybrids with increasing number of repeats ofthe tetramer unit at three different sites of OprF were constructed to investigatethe effect of the length of the epitope on its antigenicity.The three sets of multiple-repeat hybrids were constructed to express 7,11, 15 or 19 amino acids of the malarial epitope sequence at either aa26, aa’96 oraa213 of OprF. At insertion sites aa2° and aa’96, the antigenicity of the epitopeincreased as the length of the epitope increased Figs. 29 & 30). The assays for eachset were repeated three times and the r (correlation coefficient) values by linearregression in each independent experiment were between 0.93 14 and 0.9877, andbetween 0.9453 and 0.9875 for insertion sites aa26 and aa’96 respectively. In linearregression, r lies between -1 and +1, and when r is close to one this indicates apositive linear relationship (Ott, 1988). Therefore, these results demonstrated asignificant, positive relationship between the length of the epitope and itsantigenicity in these two cases. On the other hand, the antigenicity of the fourlengths of the epitope inserted at aa213 was not significantly different and did notseem to vary with the length of the insert (Fig. 31). According to the resultsobtained from the antigenicity study of the positional hybrids, the epitope insertedat aa213 was comparatively more antigenic than that inserted at aa26 and aa’96,probably due to the better exposure of the epitope at this site. Considering thisresult, the shortest version of the epitope inserted at aa213 might already beadequately accessible for antibody binding; hence the increase in length did not116xa,-oCC.,Ca,0)4-’C2I07 11 15 19 10No. of aa corresponding to the epitopeFigure 29. Effect of the length of the epitope on its antigenicity at insertionsite aa26 of OprF.The data from one representative experiment is shown. The r value(correlation coefficient) from linear regression analysis in this experiment was0.9314.—1172xa)C.)a) 107 11 15 19 10No. of aa corresponding to the epitopeFigure 30. Effect of the length of the epitope on its antigenicity at insertionsite aa’96 of OprF.The data from one representative experiment is shown. The r value(correlation coefficient) from linear regression analysis in this experiment was0.9875.xa,.>C.)a)0)43211187 11 15 19 10No. of aa corresponding to the epitopeFigure 31. Effect of the length of the epitope on its antigenicity at insertionsite aa213 of OprF.The data from one representative experiment is shown. The r value(correlation coefficient) from linear regression analysis in this experiment was0.0 17 1.0 —119lead to a significant improvement in antigenicity. On the other hand, in the caseof insertion sites aa26 and. aa’96, the longer inserted sequence might increase theantigenicity by improving the exposure of the epitope. This seemed to beparticularly obvious at aa26, where a more than two-fold increase in antigenicitywas observed as the length of the inserted epitope was increased from 7 to 11 aminoacids (Fig. 29). The antigenicity of the inserted malarial epitope in the multiple-repeat hybrids is summarized in Table XIII.3.3 Immunogenicity studyThe immunogenicity of a molecule refers to its ability to elicit an immuneresponse. This study concentrated on the antibody responses against theimmunogens. In general, to elicit a good antibody response, the immunogen isrequired to have a B cell epitope which binds to receptors on B cell surface, and aT cell epitope that can be recognized by MHC class II molecules and presented toT helper cells (Guillet et al., 1986; Brown et al., 1988). In an epitope presentationsystem, the flanking amino acid sequences and length of an epitope might affect itsinteraction with B cell receptors as well as its processing and presentation by MHCmolecules. This section describes the immunogenicity of a 10-amino acid epitopeinserted at aa215 and two different lengths of the epitope inserted at aa26 of OprF.In addition, the immunogenicity of the two lengths of the malarial epitope fused tothe C-terminus of GST was also investigated.TableXIII.SummaryoftheantigenicityofthemalarialepitopeinOprF::malarial epitopemultiple-repeathybrids.Antigenicityindicesfromindependent experimentsInsertionsitePlasmidAminoacidsinserteda(rvaluesY1st2nd3rdAla-26pRW3O2.7PAARNPNANPNLDAGP0.250.160.26pRW3O2.l1PAAR(NPNA)2NPNLDAGP0.80.60.48pRW3O2.15PAAR(NPNA’3NPNLDAGP0.940.841.01pRW3O2.19PAAR(NPNA)NPNLDAGP1.090.941.3(0.931)(0.959)(0.988)Ala-196pRW3O8.7PAARNPNANPNLDAGP0.531.270.81pRW3O8.llPAAR(NPNA)2NPNLDAGP0.731.460.91pRW3O8.15PAAR(NPNA)3NPNLDAGP0.81.721.52pRW3O8.19PAAR(NPNA)1NPNLDAGP1.261.821.71(0.945)(0.988)(0.961)Arg-213pRW3O9.7TCTRNPNANPNLDCRS1.171.553.48pRW3O9.llTCTR(NPNA)2NPNLDCRS1.4224.01pRW3O9.15TCTR(NPNA)3NPNLDCRS1.211.863.65pRW3O9.19TCTR(NPNA)1NPNLDCRS1.371.543.61(0.415)(-0.096)(0.017)aTheaminoacidresiduescorrespondingtothemalarial epitopeareunderlined.TheflankingaminoacidsPA_OPandTC_RSweretheresultsof thepreviouslinker-insertionmutagenesisprocedures.IDristhecorrelationcoefficientvaluebylinear regression, wherer>1indicatesapositivelinearrelationship.C1213.3.1 Immunogenicitv of OIrF: :ME lOaa2 15The immunogenicity of the epitope inserted at aa215 was examined byimmunizing BALB/c mice with OprF::MElOaa2l5. The two control groups includedanimals immunized with wild type OprF or PBS. The animals were immunized ondays 0 and 14 with FPLC-purified proteins and on day 28 with 2x108 heat-kifiedE. coli cells expressing the corresponding proteins. The anti-OprF and antimalarial epitope antibody responses were determined by ELISA using purifiedOprF and GST::ME19 and (NANP)3 peptide as coating antigens respectively(Methods and materials 13.3). No significant anti-OprF and anti-malarial epitopetiters were detected in antisera after two injections. Significant anti-OprF responsewas observed in all animals after the third injection, whereas an anti-epitoperesponse was only observed in one of the five animals immunized withOprF::MEaa2 15 (Fig. 32). However, characterization of the antiserum from theresponsive animal by Western immunoblotting failed to demonstrate the presenceof anti-malarial epitope antibodies in this serum. Therefore, the anti-epitoperesponse detected could have been due to non-specific binding of the antiserum tothe coating antigen. No significant anti-OprF or anti-epitope titers was detectedin the pre-immune sera or the sera from the PBS control group. These resultsindicated that despite the fact that the malarial epitope inserted at aa215 of OprFwas antigenic; it was not immunogenic when administered to BALB/c mice. Thetiters of the antisera are summarized in Table XIV.122—7’? -— — W • • • — • —3 4 5 6Log,0 dilutions of antisera-i,,OprF•LUJ,..—A and B, anti-OprF response; C and D, anti-malarial epitope response.The anti-OprF response and anti-malarial epitope responses were measuredusing purified OprF and GST::ME19 as coating antigens respectively. Theimmunogens used in each groups are indicated on the graphs. Symbols: •,pooled pre-immune sera; A, D, 4, A, •; serum samples from five differentOprFA Vi“\—OprF::MElOaa2l5‘E’, vc’4\isi, “\A2A6550C2A65503 4 5Log,0 dilutions of antiseraB2A6550D2455506OprF::ME1 Oaa2l 5LiAIiIE ± ,A —2 3 4 5 2 3 4 5Log,0 dilutions of antisera Log,0 dilutions of antiseraFigure 32. ELISA titrations of anti-OprF and anti-malarial epitope responsesinduced in BALBIc mice immunized with OprF andOprF::MElOaa2l5 by ELISA.animals.123Table XIV. Summary of antibody responses induced in mice immunizedwith wild type OprF or OprF::MElOaa2l5.ELISA titersaImmunogens Animals Anti-OprF Anti- AntiGST::ME19 (NANP)3OprF::ME1O- a 5.00±0.61 3.4±0.96 <2aa215 b 4.70±0.37 2.4±0.35 <2c 4.70 ± 0.37 2.8 ± 0.35 <2d 5.30±0.37 2.8±0.46 <2e 5.20±0.12 2.4±0.35 <2OprF a 5.37±0.57 <2 NDbb 4.37 ± 0.41 <2 NDc 5.27±0.64 <2 NDd 4.97±0.71 <2 NDe 5.57±0.41 <2 NDa Titers are reported as the log of dilutions of antisera that gave twice of theA655 readings of the pre-immune serum at 100-fold dilutions. Anti-OprF andanti-GST::ME19 responses were determined using purified OprF andGST::ME19 as coating antigens respectively. Anti-(NANP)3was determinedusing a synthetic peptide NANPNANPNANP as coating antigen. Antiserawere taken after three injections. The reported values are mean values ±standard deviations from three independent assays.1ND, not determined1243.3.2 Immunoenicitv of OurF: :ME7aa26 and OurF: :ME 19aa26To examine the effect of the length of the epitope on its immunogenicity,two OprF::malarial epitope hybrids, OprF::ME7aa26 and OprF::MEl9aa26,carrying a 7- and a 19- amino acid malarial epitope respectively at aa26, were usedas immunogens in an immunization study. C57BL/6J mice were immunized with20 .tg of the gel-purffied immunogens on days 0 and 21 and with 10 .tg of theimmunogens on day 35. Control groups included animals injected with wild typeOprF or PBS. The anti-OprF and anti-malarial epitope antibody responses weredetermined by ELISA using purified OprF, GST::ME19 and (NANP)3peptide ascoating antigens (Methods and materials 13.3). The antisera taken after the secondinjection showed significant anti-OprF titers (>10), but no anti-epitope titers weredetected in the same sera. After 3 injections, the anti-OprF titers increased to >i0in all three groups, while a significant anti-malarial epitope response was onlydetected in animals immunized with OprF::ME 19aa26 and a weak anti-malarialepitope response was detected in one of the five animals (animal a) immunized withOprF::ME7aa26 (Fig. 33, Table XV). As controls, the anti-OprF titers were similarin all three groups. Characterization of the antisera by Western immunoblottingdemonstrated the presence of anti-OprF and anti-malarial epitope antibodies in theantisera that showed the corresponding antibody response in ELISA (Fig. 34).Neither the pre-immune sera nor antisera taken from the PBS control groupshowed any significant anti-OprF and anti-malarial epitope response. The resultssuggested that the 19-amino acid epitope inserted at aa26 was significantly more125A 2—” B 2—”OprF::ME7aa26 OprF::MEl9aa26A655\A555\j03456 3456Log10 dilutions of antisera Log10 dilutions of antiseraC 2—’ D 2—”OprF::ME7aa26 OprF::MEl9aa26A655 A655&o 02 3 4 5 2 3 4 5Log10 dilutions of antisera Log10 dilutions of antiseraFigure 33. ELISA titrations of anti-OprF and anti-malarial epitope responsesinduced in C57BL/6J mice immunized with OprF::ME7aa26 andOprF::MEl9aa26 by ELISA.A and B, anti-OprF response; C and D, anti-malarial epitope response. Theanti-OprF and anti-malarial epitope responses were measured using purifiedOprF and GST::IvlEl9 as coating antigens respectively. The immunogens usedin each groups are indicated on the graphs. Symbols: •, pooled pre-immunesera; A, C],+, A, B; serum samples from five different animals.126Table XV. Summary of antibody responses induced in mice immunized withOprF::ME7aa26 and OprF::MEl9aa26.ELISA titersaImmunogens Animals Anti-OprF Anti- AntiGST::ME19 (NANP)3OprF a 5.91±0.17 <2 <2b 5.61±0.17 <2 <2c 5.71±0.30 <2 <2d 5.11±0.60 <2 <2e 5.11±0.52 <2 <2OprF::ME7- a 4.61±0.35 2.30±0.00 <2aa26 b 5.31± 0.62 <2 <2c 5.31±0.35 <2 <2d 5.31±0.62 <2 2.00±0.00e 5.71±0.30 <2 <2OprF::ME19- a 5.71±0.30 3.40±0.35 <2aa26 b 5.71±0.30 3.71±0.46 2.20±0.17c 5.51±0.46 3.20±0.31 <2d 5.31 ± 0.62 4.21 ± 0.35 2.50 ± 0.17e 5.71±0.30 3.05±0.26 2.15±0.21a Titers are reported as the log of dilutions of antisera that gave twice of the4readings of the pre-immune serum at 100-fold dilutions. Antisera were takenafter three injections. The antigens used were as described in Table XIV. Thereported values are mean values ± standard deviations from three independentassays.127Figure 34. Western immunoblot analysis of the sera from mice immunizedwith OprF: :ME7aa26 and OprF::ME 19aa26,A), Western immunoblot using purified E. coli OprF as the antigen withantisera from individual immunized mice at 1/1000 dilution. B). Westernimmunoblot using GST::ME19 as the antigen with antisera from individualimmunized mice at 1/100 dilution. Lanes: p1, pooled pre-inimune sera; a,b,c,d,e;sera from five mice immunized with the indicated proteins. Arrows representnative OprF (A) and GST::ME19 (B) respectively. Positive controls used wereMA7-2 and pf2A. 10 (both at 1/3000 dilution) for (A) and (B) respectively.OprF::ME7aa2Gp.1. a b c d eABI—iOprF::MEI 9aa26p.1. a b c d a 7-2OprF::ME19aa26p.1. a bOprF::ME7aa26p.i, a b c d aHCd • 2A,1O128immunogenic than the 7-amino acid epitope inserted at the same site. The titersof the antisera are summarized in Table XV.3.3.3 Immunoenicitv of GST::ME11 and GST::ME19Due to the ease of genetic cloning and the simplicity of the subsequentpurification of the fusion proteins, glutathione S-transferase is quite frequentlyused as a carrier to enhance the immunogenicity of small peptide immunogens(Ling et al., 1994; von Specht et al., 1995). Since two versions of GST::malarialepitope fusion proteins were already available, the immunogenicity of the epitopein these proteins was also studied so as to allow comparison between the GST andOprF carrier systems. The control groups in this experiment were immunized withGST or PBS. The anti-GST and anti-malarial epitope titers were determined byELISA using gel-purified GST and OprF::MEl9aa26 as coating antigensrespectively. Animals immunized with GST or the GST fusion proteins developeda significant anti-GST response after 2 injections. The anti-OST and anti-malarialepitope responses in antisera taken after 3 injections are shown in Figure 35. Allthree groups immunized with the protein immunogens generated significant antiGST titers (‘4O); however, no anti-malarial epitope response was observed in thegroups immunized with the fusion proteins. Figure 36 shows the presence of antiGST antibodies in the antisera by Western immunoblotting. The same antisera didnot react with OprF::MEl9aa26 on Western immunoblotting, indicating theabsence of malarial epitope-specific antibodies. The inability of the GST fusion129A 2—“ B 2GST::ME11 GST::ME19A655 A655 i \\ ‘,•‘\ \\o345603456Log10 dilutions of antisera Log10 dilutions of antiseraC 2—” D 2—”GST::ME11 GST::ME19A655 655 Io:1F.Log10 dilutions of antisera Log10 dilutions of antiseraFigure 35. ELISA titrations of anti-GST and anti-malarial epitope responsesinduced in C57BL/6J mice immunized with GST::ME11 andGST::ME 19.A and B, anti-GST response; C and D, anti-malarial epitope response. Theimmunogens used in each group are indicated on the graphs. The anti-GST andanti-malarial epitope responses were measured using purified GST andOprF::MEl9aa26 as coating antigens respectively. Symbols: •, pooled preimmune sera; A, E],•, A, , serum samples from five different animals.GST::MEI9piabcde_c+cIIFigure 36. Western immunoblot analysis of the sera from mice immunizedwith GST::ME11 and GST::ME19.Affinitypurffied GST was used as the antigen. Antisera examined werefrom individual immunized mice at 1/1000 dilution. Lanes: p.i., pooled preimmune sera; a,b,c,d,e, sera from five mice immunized with the indicatedproteins;—C, negative control using pf2A.lO; +0, positive control using an antiGST antiserum. The arrow indicates the position of GST.130GST::MEI Ip.i.abcde;i [131proteins to elicit an anti-malarial epitope response was unexpected. A possibleexplanation could be that the folding of the epitope in the GST fusion proteins didnot allow the epitope to be readily recognized by the components of the immunesystem. The titers of the antisera are summarized in Table XVI.3.4 SummaryIn this chapter, the antigenicity of the malarial epitope presented atdifferent positions and in different lengths in the OprF epitope presentation systemwas studied. The malarial epitope inserted at different positions of OprF displayeddifferent binding affinities for an epitope-specific mAb. For example, the malarialepitope inserted at aa’ or aa’96 was consistently more antigenic than that insertedat aa215 or aa310, while insertions at aa26 or aa213 were significantly more antigenicin the whole cell and outer membrane environments respectively. Insertion ofmultiple copies of the epitope at aa and aa213 resulted in higher levels of antibodybinding than with single copy; possibly due to the increase in valency andlor betterpresentation of the binding epitope. Among the three sets of multiple-repeathybrids, insertions at aa26 and aa’96, but not at aa213 demonstrated an increase inantigenicity with the increase in length of the epitope. This suggested that thecorrelation between length and antigenicity of the epitope was site-dependent inthe OprF system.The immunogenicity of the epitope genetically inserted in OprF or fused132Table XVI. Summary of antibody responses induced in mice immunized withGST::ME11 amd GST::ME19.ELISA titersaImmunogens AnimalsAnti- AntiAnti-GST OprF::ME19- (NANP)3aa26GST a 4.81± 0.53 <2 <2b 4.10±0.46 <2 <2c 4.81±0.61 2.15±0.21 <2d 4.41±0.46 <2 <2GST::ME11b a 4.90±0.76 2.10±0.17 <2b 4.00±0.76 <2 <2c 4.20±0.61 <2 <2d 4.40 ±0.63 <2 <2e 3.80 ±0.63 <2 <2GST::ME19b a 4.30±0.76 2.35±0.31 <2b 4.20±0.91 2.00±0.00 <2c 4.41±0.76 2.00±0.00 <2d 3.80±0.76 2.20±0.17 <2e 4.30±0.76 <2 <2a Titers are reported as described in Table XV footnotes. Antisera were takenafter three injections. The anti-GST and anti-OprF::MEl9aa26 responseswere determined by using purffied OST and OprF::MEl9aa26 as coatingantigens. The anti-(NANP)3response was determined by using the syntheticpeptide NANPNANPNANP as coating antigen. The reported values are meanvalues ± standard deviations from three independent assays.‘0 The ma]anal epitope sequence fused to the C-terminus of GST wereDP(NANP)2NAQL and DP(NANP)4NAQL for GST::ME11 and GST::ME19respectively.133to GST was also investigated. Despite its ability to interact with the malarialepitope-speciflc mAb, a 10-amino acid epitope inserted at the C-terminal insertionsite aa215 of OprF was not immunogenic. A 19-amino acid epitope inserted at theN-terminal insertion site aa26 was able to elicit a significant anti-malarial epitopeantibody response, whereas a 7-amino acid epitope inserted at the same site wasonly weakly immunogenic. For the GST-malarial epitope fusion proteins, it wasfound that neither an 11- nor a 19-amino acid epitope fused to the C-terminus ofGST could stimulate an anti-malarial epitope response in immunized animals.134DISCUSSIONGeneralThis study demonstrated the potential of OprF, the major outermembrane protein of P. aeruginosa, as a carrier for the presentation of foreignantigenic determinants. Semi-random linker-insertion mutagenesis was conductedto investigate the “permissiveness” of different regions of OprF to accommodateextra amino acid residues. The repeating epitope (NANP) of the circumsporozoiteprotein of the malarial parasite, P. falciparum, was used as a model epitope tofurther explore the usefulness of the insertion sites to express foreign antigenicdeterminants. The antigenicity of the malarial epitope inserted at differentpositions and in different lengths in OprF was compared. A targeted study wasalso undertaken to exanune the immunogenicity of the inserted epitope in selectedOprF::mal.arial epitope hybrid proteins.The results of the linker-insertion and epitope-insertion studies havegenerated useful information about the membrane topology of OprF. In addition,the analysis of the reactivities of the OprF linker mutants and the OprF::malarialepitope hybrid proteins with the series of OprF-specffic monoclonal antibodies haveimproved our understanding of the binding epitopes of these antibodies. This studyrepresents the first attempt to compare the antigenicity of the presented epitope ineight different insertion sites of the carrier protein in an epitope presentation135system. Moreover, it is also the ffrst study to systematically investigate the effectsof insertion position and length of the epitope on its antigenicity in an epitopepresentation system. Furthermore, this study revealed, for the first time, thatOprF can be used as a carrier for a foreign epitope to generate and detect antiepitope antibodies in immunized animals and in immunoassays respectively.Linker-insertion mutagenesisAccording to restriction enzyme site analysis, there were 37 sites withinoprF that were potential targets for the linker-insertion mutagenesis. However,only 13 unique sites (including 2 in which the inserted 12 nucleotide pairs weretranslated to a stop codon) were identified after the screening of 100 clones byrestriction enzyme digest analysis. Although attempts were made to identify siteswhere the insertion of the 4-amino acid linker resulted in no detectable OprFproduct (i.e. “non-permissive” sites), only one such site was identified. Six of theseven putative “non-permissive” clones analyzed by DNA sequencing showed thatthe lack of detectable OprF product was the result of deletions or the incorporationof stop codons translated from the 12 bp linker due to the reading frame at theinsertion sites. It was possible that insertion sites that were close to each otherhave been overlooked by the restriction digest analysis. In addition, it was alsolikely that more unique sites could have been identified by more exhaustivescreening.136Effects of amino acid insertions in OprFAlthough the results of the characterization of the OprF linker mutantsand the OprF::malarial epitope hybrid proteins suggested that these OprF variantsshared many similarities to the wild type, insertion of extra amino acids at specificsites did alter the wild type properties to a limited extent. For example, theinsertion of 4 amino acids at aa2 of the mature protein directed the expression of a41 kfla product which was of a higher apparent molecular mass than the wild typeOprF (35 kDa) (Table IV). It is noted that one of the inserted amino acids wasarginine. A number of studies have reported that the incorporation of positivecharges at the N-terminal end of the mature protein impedes the export ofmembrane proteins (IViaclntyre and Henning, 1990; Geller et at., 1993; Struyvé etat., 1993b). These studies suggested that the positively-charged residues mightaffect the export of proteins by disrupting membrane potential and interacting withpolar head groups of acid phospholipids, thus arresting the transport across thecytoplasmic membrane. On the other hand, it has also been documented that thepresence of positive charges at the extreme N-terminus of the mature protein caninterfere with signal peptide processing (Li et at., 1988). Therefore, one canspeculate that this 41 kDa band corresponding to the OprF variant carrying aninsertion at aa2 represented OprF with the signal peptide stifi attached. Thepresence of this OprF variant in the outer membrane preparation might have beencaused by contamination of the preparation with inclusion bodies, which were137formed as a consequence of defects in folding and the export pathway (Marston,1986). The insertion of a 10-amino acid malarial epitope at aa2 did not lead to anydetectable product. If the insertion of 4 amino acids already interfered with theexport pathway, it is conceivable that the insertion of an additional 10 amino acidsmight lead to a more severe effect, resulting in the degradation of the cytoplasmicintermediate.The series of multiple-repeat hybrids carrying an insertion at site aa26 alsoinduced incomplete formation of inclusion bodies as indicated by the migration ofsome of the protein at the expected position on SDS-PAGE (Fig. 17). The inclusionbody form of these proteins differed from the membrane associated form in twoaspects. First, the apparent molecular mass of the inclusion bodies was higherthan that of the membrane associated form of the proteins. As mentioned above,this could be due to the presence of signal peptide as a result of defective signalpeptide cleavage. Second, the inclusion body form was not modified by 2-mercaptoethanol, indicating the absence of disulphide bond. Since the proteinsrequired for disulphide bond formation in compartments external to the innermembrane, DsbA and DsbB, are located in the periplasm and cytoplasmicmembrane respectively (Bardwell, 1994), this observation is consistent with thecytoplasmic localization of the inclusion bodies. The partial defect in export of thisseries of OprF::malarial epitope multiple-repeat hybrids was apparently not due tothe length of the inserted sequence because the corresponding positional hybridwith a 10-amino acid insertion at the same site was completely inserted into the138outer membrane. Comparison of the flanking amino acid sequences of thepositional hybrid and the multiple-repeat hybrids revealed that an arginine residuewas introduced at the N-terminal side of the malarial epitope in the multiple-repeathybrids as a result of the cloning procedures. It has been documented that positivecharges introduced at positions up to +20 of the mature protein are able to inhibitexport (Kuhn et al., 1994). Therefore, it is possible that the introduction of thearginine residue at aa26 of the mature OprF might lead to the same effect. It shouldbe noted that the insertion of arginine residues at other sites (e.g. aa213, aa231, andmultiple-repeat hybrids at aa’96) did not affect the localization of the protein,indicating that this effect was site-specific.Linker insertion at aa’3’ directed the production of an OprF variant ofmolecular mass identical to that of the heat-modified, unfolded form of OprF andwith a distinct trypsin-cleavage pattern (Figs. 12 and 13). Thus it can be assumedthat this insertion influenced the SDS stability of the protein, probably by inducinga slight change in membrane configuration. This decrease in stability may haveexplained the result that the insertion of a 10-amino acid epitope at this site did notlead to any detectable product. Taking together the results of the linker-insertionand epitope-insertion studies, this might reflect the inflexibility of this region ofOprF to accommodate extra amino acid sequences.Linker insertions at aa’88 and aa’96 led to OprF variants with a differenttrypsin-cleavage pattern in outer membranes (Fig. 13), indicating exposure of atrypsin-accessible cleavage site in or adjacent to the cysteine disulphide loop. This139suggested that the insertion of 4 amino acids caused alterations in the localconformation, a notion supported by the loss of reactivities of these OprF variantswith the OprF disulphide bond-sensitive antibodies MA4-4 and MA5-8, and yet theretention of reactivities with the rest of the OprF-specific antibodies (Table VI).Likewise, the OprF::malarial epitope hybrid proteins with epitope insertions ataa’88 or aa’96 retained reactivity with most of the OprF-specific mAbs except forMA4-4 and MA7-8 I’able XI). Therefore, by the same token, one can assume thatthe decrease in heat stability in these OprF variants was due to localizedinstability, instead of an impact on the general folding of the n-sheet structure (Fig.14).It has been shown in this study that linker insertions at the C-terminalregion of OprF demonstrated enhanced susceptibility to trypsin cleavage in theouter membrane while epitope insertions at the same region increasedsusceptibility of the hybrid proteins to cellular proteases. The increase insusceptibility could be due to the introduction of a new protease cleavage site or theincrease in accessibility of existing cleavage site(s). Since the degradation productobserved in all the cases had the same apparent molecular mass of 28 kDa and acleavage product of the same apparent molecular mass was observed in wild typeOprF, it seems more likely that the insertion of extra amino acids increased theaccessibility of an existing cleavage site. This is reminiscent of the previous findingby Finnen et al. (1992) that C-terminal perturbations render the resulting OprFmutants more prone to protease action.140In cases where cysteine residues were present in the linkers translatedas TCRS, interactions of these exogenous cysteines with the cysteines in OprFmight disrupt the protein structure by forming alternate disuiphide bonds.However, the observations that the OprF variants with 4-amino acid insertion ataa’35 and aa213 migrated with similar mobility on SDS-PAGE, showed similar 2-mercaptoethanol modifiability and reacted with MA7-8 and MA4-4, which recognizeepitopes sensitive to reduction of the OprF disuiphide bonds, suggested that thecysteine residue present in the linkers of these variants did not participate indisulphide bonding with the endogenous cysteine residues. A recent study hasreported on cysteine substitution mutagenesis of LamB and the subsequentexamination of the reactivity of the mutated proteins with different monoclonalantibodies before and after thiol treatment (Notley et al., 1994). The data obtainedprovided new information to extend previous maps of the monoclonal. antibodybinding sites. Therefore, using the same approach, the cysteine residues introducedin the linkers might provide similar information to refine the mapping of the OprFepitopes recognized by specific antibodies.Membrane topology of OprFSiehnel et al. (1990) previously presented a model for OprF based on theapparent existence of two disulphide bridges between the four cysteines of OprF(Hancock and Carey, 1979), the p-turn prediction rules of Paul and Rosenbusch141(1985), and circular dichroism data suggesting that 62% of the secondary structureof OprF is in the form of p-sheet, a value typical for outer membrane proteins(Cowan et al., 1992; Schirmer et al., 1995). This model has been tested in part byThphoA mutagenesis and deletion analysis of the oprF gene (Finnen et al., 1992).Sequence comparison studies have revealed that the sequences in thesurface loop regions of outer membrane proteins are hypervariable. These regionsare possibly least spatially constrained because of their surface location. Both ofthese properties lead to the assumption that these regions are more likely totolerate extra amino acid insertions (Charbit et al., 1991). Based on thisassumption, linker- and epitope-insertion studies of other outer membrane proteinshave suggested that the “permissive” insertion sites are most likely to be within theloop regions of the proteins. Interestingly, the 3-dimensional structures of PhoEand LamB, which were resolved after the insertion studies, proved this assumptionto be correct (Cowan et al., 1992; Schirmer et al., 1995). Therefore, the insertionsites characterized in this study should have similar value in defining OprFtopology.With the data presented here, a new topological model of OprF wasconstructed partly based on the assumption that the “permissive” sites for linkerinsertion should be in the loop regions of the protein. It was considered that thesite at aa2 was probably periplasmic since all outer membrane proteins studied todate have N-termini that are in the periplasm. The insertion site at aa42 wasconsidered “non-permissive” since it resulted in no detectable product. Thus this142site was placed within the membrane. The remaining 9 sites were placed in surfaceloops. Linker insertion in five of these sites (aa’, aa’96, aa231, aa29° and aa310)interrupted the binding of specific monoclonal antibodies that have been shown tobind to surface-exposed epitopes (Martin et al., 1993), and were thus placed in thesurface-exposed loop regions. The placement of aa31° was supported by its locationwithin a flexible protein segment adjacent to a variable sequence (Jeanteur et al.,1991). The location of this site was further confirmed by a subsequent study whichshowed that the insertion point is within the linear, surface-exposed epitope forIVIA5-8 (Rawling et al., 1995). Insertions in the other 4 sites did not interrupt thebinding of specific monoclonal antibodies, but these sites were assigned on thesurface in keeping with the precedent in LamB and PhoE that “permissive”insertion sites are usually found in the surface ioop regions. Moreover, the resultsobtained from the malarial epitope-insertion study showed that the epitope insertedat these sites was detectable on the cell surface (Table IX), which supported thesurface localization of these sites. A revised topological model of OprF, which isconstructed in part based on the data from this study, is shown in Figure 37.The OprF topology model indicates that the insertion site at aa215 is at theend of a transmembrane segment. This segment has high homology to thetransmembrane segment 8 of OprF and to the consensus 16th transmembrane 3-strand of the porin superfamily (Jeanteur et al., 1991), and thus appears almostcertain to be in the membrane. However, this site was found to be “permissive” forthe incorporation of both the 4-amino acid linker and the 10-amino acid epitope.The linker-insertion sites are circled and the permissive malarial epitopeinsertion sites are indicated by solid triangles. The top of this model is proposedto face the exterior of the cell. The transmembrane p-strands are indicated byrectangular boxes.GL EKVV TvC0ANGAQ0 0LDKDVDuAALNKNRriJs101ITIIFII IIIKGFAEEVSN143NTS00RANGADN0H00ELiiS0SC:V0AVpEpApAAKS0T NE KU V KT v s0 HrirnI VI IN I101 IL II H I TII I Is II El ILl101 lo Ili I AlI I IiiI LI IvIA H[JJFigure 37. Proposed membrane topology of OprF.KQVpSTST0A23lGVAG ®:b0 AAV BN OK LS E144Examination of the inserted sequences suggested the possibility that sufficient-strand character was maintained in these sequences, which might have enabled theextension of this strand and hence the exposure of the epitope on the cell surface.Insertions at aa213 and aa23’ appeared to promote OprF trimer stability inthe absence of 2-mercaptoethanol (Fig. 11). In the 3-dimensional structures ofPhoE and LamB, it was revealed that the surface loop L2 is responsible for trimerassociation by extending from one monomer to an adjacent monomer. This leadsto the speculation that the two surface loops of OprF where the insertions occurredmight also be involved in OprF trimer/oligomer association. Extending these loopsby 4 amino acids might enable them to reach farther to the adjacent monomer, thusenhancing trimer stability. On the other hand, the reduction of the disulphide bondby 2-mercaptoethanol might have loosened the secondary structure in a mannerthat resulted in the dissociation of the trimers.Binding epitopes of OprF-specific monoclonal antibodiesThe examination of the reactivities of the linker-insertion and epitopeinsertion derivatives of OprF with the series of OprF-specffic monoclonal antibodiesalso provided information about the binding epitopes of these antibodies. Recently,Rawling et al. (1995) had delineated the OprF epitopes recognized by a series ofOprF-speciflc monoclonal antibodies available in our laboratory. The results of thisstudy were based on data generated by monoclonal. antibody reactivities with145overlapping synthetic peptides on pins and cyanogen bromide and papain cleavagefragments of OprF. Since the insertions of the 4-amino acid linker and the malarialepitope only disrupted a localized region of the protein while apparently leaving therest of the protein intact, these approaches represented a more subtle way to definethe boundary and nature of the binding epitopes.This study provided new information about the conformational epitopesrecognized by the OprF-specffic mAbs MA4-4 and MA7-8. Rawling et al. suggestedthat epitope(s) recognized by these antibodies are located at the same region fromaa’52-aa210. According to the linker-insertion study, insertion of 4 amino acids ataa’96 disrupted the epitopes for both MA4-4 and MA7-8, while insertion at aa’88disrupted only the MA7-8 epitope (Table VI), implying that these antibodies bindto overlapping but distinct epitope(s). Moreover, based on the observation that thepresence of 2-mercaptoethanol abolished the binding of these monoclonal antibodiesto OprF, the binding of MA4-4 and MA7-8 are believed to require the presence ofdisuiphide bond(s) (Mutharia and Hancock, 1983). In the present study, most ofthe OprF variants carrying linker and epitope insertions in the cysteine-containingregion (aa’88 and aa’96) resulted in the loss of MA4-4 and IVIA7-8 reactivities. Theinsertion of extra amino acid residues in this region could potentially haveinterfered with the formation of disuiphide bond and/or affected the conformationof the disulphide bond-containing region. The observation that these variants stillretained similar 2-mercaptoethanol modifiability as wild type OprF, and that the4-amino acid insertion at aa188 stifi retained MA4-4 reactivity suggested that the146disuiphide bonds were correctly formed. Hence, it is likely that the loss of MA4-4and MA7-8 reactivity was due to a change in the secondary structure(conformation) of the amino acids in the disuiphide bond-containing region ratherthan the disruption of disuiphide bond formation. Therefore, it appears likely thatthe binding of 1VIA4-4 and 1\/1A7-8 requires both the correct formation of thedisuiphide bonds and the correct secondary structure in this region.This study also shed new light on the conformational epitopes recognizedby the OprF-specific mAbs MA7-3, MA7-4, MA7-5 and MA7-7. The insertion of 4amino acids at the C-terminus of OprF (aa231, aa29° and aa310) abolished the bindingof all of these monoclonal antibodies (Table VI). Interestingly, based on thereactivity of protease digested peptides carrying 1-2 75 amino acids, Rawling et al.(1995) have delineated the MA7-3 epitope to aa’88- a23°and the IVIA7-4, MA7-5 andMA7-7 epitopes toaa1-aa278. The results presented here revealed that althoughsites aa° and aa31° do not comprise the epitope(s), changes at these sites could stifiaffect the conformation of the epitope(s). Alternatively, the presence of extra aminoacids at these sites might have reduced the accessibility of the epitope(s) forantibody binding. A number of studies have reported that changes at sites distantfrom the epitopes can interfere with the binding of antibodies to the epitopes(McCutcheon et al., 1993; Collawn et al., 1988). Therefore, it is possible that theinsertion of the 4-amino acid linker at aa29° and aa31° induced conformationalchanges in the epitopes recognized by MA7-3, MA7-4, MA7-5 and MA7-7, andconsequently interrupted the binding of these antibodies. The insertion of a 10-147amino acid epitope at aa29° and aa31° resulted in weak binding of MA7-3, whiledisrupted binding of MA7-4, MA7-5 and MA7-7 (Table XI). This is consistent withthe results from Rawling et al. (1995), which suggested that MA7-3 recognizes adifferent epitope than MA7-4, MA7-5 and MA7-7.Based on monoclonal antibody reactivities of overlapping syntheticpeptides on pins, Rawling et al. concluded that 1VIA7-1, 1VIA7-2 and MA5-8 recognizelinear epitopes of OprF. In general, the disruption of the linear binding epitopesby insertions in this study was consistent with the epitope boundaries as definedby this previous study (Rawling et al., 1995 and Table III). The only exception wasthat the OprF linker mutant carrying an insertion at aa’3’ was not reactive with1VIA7- 1 (Table VI), whose epitope has been mapped to aa55 to aa62 of OprF. Sinceinsertion at aa’3’ appeared to cause minimal changes in OprF membraneconfiguration in that the binding of the majority of mAbs was unaffected, the lossof MA7- 1 reactivity of this mutant might have been due to the masking of the MA7-1 epitope as a result of the extension of an adjacent ioop.Antigenicity and mode of presentationThe larger number of sites examined in this study permitted us toattempt to correlate the measured antigenicity of the epitope at the variousinsertion sites with the primary and secondary structures at these sites. Thepossible structures of each insertion site were analyzed using various structure148prediction methods (Table XVII). The structures at each insertion site wereanalyzed in the context of the entire protein, as well as in a segment of thesequence including the six OprF amino acid residues flanking either side of theinsertion. When only the flanking residues were taken into consideration, theGascuel and Golmard Basic Statistical Methods (GGBSM) analysis (Gascuel andGolmard, 1988) predicted that three or more amino acids were in extendedconformation on at least one of the flanking sequences of insertion sites aa196, aa29°and aa213, where epitope insertion showed medium to high antigenicity (Fig. 27).The antigenic determinant program of Hopp and Woods (1981) predicted that thesesites have comparatively low to medium hydrophiiicity on both sides of the flankingregions. When analyzed in the context of the entire amino acid sequence, theinsertion sites that exhibited high relative antigenicity were found in regions thatwere generally predicted to be more flexible in their local secondary structure andto have higher coil propensity (Karplus and Schulz, 1985). Although thecorrelations were not universal, the general trend of extended conformation andhigh flexibility of the local sequences at insertion sites which resulted in highrelative antigenicity of the inserted epitope seemed to suggest that these featuresmight improve the accessibility of the epitope.It is noteworthy that epitope inserted at aa213 and aa’5 demonstratedsignificantly different antigenicities despite the fact that these insertion sites areonly 2 amino acids apart (Fig. 27). The proposed location of aa215 is at the cellsurface end of a transmembrane segment of OprF. Based on the proposed model149Table XVII. Predicted primary and secondary structures at the insertion sites.No. residues in Average Probability of coilInsertion extended hydrophilicityc conformationa,e Flexibilitysites conformationa (%) B [normJRightb Leftb Right LeftAla-26 0 0 0.45 0.33 55 1.09Val-188 0 0 0.42 1.03 35 0.98Ala- 196 0 3 0.43 -0.4 72 1.08Arg-213 3 4 0.34 0.53 25 1.01Gln-215 0 1 0.28 0.53 27 0.95Ser-231 0 0 1.78 0.42 30 1.08Arg-290 0 6 0.77 0.75 47 1.07Gly-310 0 0 0.3 1.15 35 1.05a as predicted by the GGBSM program (Gascuel and Golmard, 1988).b right and left of the flanking sequences respectively.cas predicted by the Antigenic Determinant program (Hopp and Woods, 1981). Thehighest and lowest hydrophilicity values of the various regions of the entire proteinare 2 and -1.5, respectively.d as predicted by the Flexpro program (Karplus and Schulz, 1985), the numberscited are the B[norm] values. The B[norm] values of the whole protein range from0.820 to 1.129.analyzed in the context of the entire protein.150of OprF topology, it can therefore be hypothesized that part of the inserted epitopemight be at the membrane interface and/or that the exposure of the epitope wasshielded by the protruding surface loops, thus resulting in low antigenicity.The results of antigenicity studies suggested that LPS plays a role in thepresentation of epitope for antibody binding at aa26, aa213 and aa290. This indicatedthat these regions of OprF are involved in LPS association. Therefore, dependingon the type of antigen preparation chosen (i.e. whole cells or outer membranes), thechoice of insertion sites for optimal antigenicity may vary. The relatively highantigenicity of the inserted epitope at aa213 and aa290, and the fact that most OprFspecific monoclonal antibodies recognize epitope(s) that are located in this region,suggested the immunodominance of this region of OprF. This notion is supportedby a recent report which identified seven B cell epitopes, two of which are surface-exposed, in the C-terminal region (aa’90- a350)of OprF, confirming that this part ofthe protein is rich in B cell epitopes (von Specht et al., 1995).The antigenicity of epitope insertions at aa26 and aa’96 increased with thelength of the epitope, while that at aa213 did not (Table XIII). This may beexplained by the degree of exposure of the insertion sites. Since the 10-amino acidepitope inserted at aa213 was comparatively more antigenic, it appears that this siteis already well-exposed. On the other hand, the 10-amino acid epitope inserted ataa26 and aa19 only displayed relatively low to medium antigenicity, suggestingmediocre exposure at these sites. Therefore, it appears that increasing the lengthof the epitope at aa26 and aa’96 might have improved its exposure, thus facilitating151its presentation for antibody binding. This is reminiscent of the insertion of theFMIV epitope into the third loop of PhoE, which is a loop that is hidden inside thepore region of the protein. Researchers found that the insertion of one copy of theepitope does not result in the surface exposure of the epitope. However, insertionof multiple copies of the epitope forces this loop out of the pore, leading to exposureof the epitope (Struyvé et al., 1993a).Since the epitope-specific monoclonal antibody pf2A. 10 was used in theantigenicity assays, one should keep in mind that the antigenicity of the epitopediscussed in this study should only refer to its binding to this monoclonal antibody.Survey experiments using two other epitope-specific monoclonal antibodies toevaluate the correlation between antigenicity and the length of the epitope at aa’96and aa213 revealed the same findings as those observed using pf2A. 10 (i.e. positivecorrelation at aa’96 and no correlation at aa213). These results implied that thetrend of antigenicity observed using pf2A. 10 may apply to the binding of otherantibodies in general.ImmunogenicityIn general, the basic mechanisms of an immune response involve theproduction of antibodies and the development of cellular immunity. This studyconcentrated on the examination of antibody (more specifically, IgO) responseagainst the immunogens. Therefore, the scope of immunogenicity discussed in the152context of this study is limited to the induction of antibody response. Theimmunogenicity of the inserted malaria]. epitope in five selected proteins,representing two different carriers (OprF and GS’I), was investigated in this study.The inserted epitope in only one of the proteins demonstrated significantimmunogenicity. A number of factors might have caused the lack ofimmunogenicity of the inserted epitope. In this section I will attempt to addresssome of the possibilities.A 19-amino acid epitope inserted at aa26 of OprF (OprF::MEl9aa26) wassignificantly more immunogenic than a 7-amino acid epitope inserted at the samesite (OprF::ME7aa26) (Table XV). Since no anti-malarial epitope response wasdetected in the immunized animals after the first injection, it indicated that theresponse was T cell-dependent. To generate a T cell-dependent response, theimmunogen is required to have a B cell epitope which binds to the antigen receptoron B cells, as well as a T cell epitope that can be recognized by MHC class IImolecules and presented to T helper cells (Guillet et al., 1986; Brown et al., 1988).The difference in immunogenicity of the 2 versions of malarial epitope inserted atthe same position of OprF could be due to a number of factors. For instance, theincrease in length of the epitope might have increased the accessibility of theepitope for binding to the B cell antigen receptor, as was indicated by the resultsof the antigenicity study which demonstrated a positive correlation between lengthand antigenicity at aa26. Alternatively, since B cell activation requires thecrosslinking of B cell antigen receptors, the poor immunogenicity of the shorter153version of the epitope might reflect its poor ability to elicit B cell receptorsignalling. Nevertheless, it is quite possible that all of the factors mentioned aboveplayed a role in determiming the immunogenicity of the epitope in this experiment.The examination of the immunogenicity of the epitope in OprF::MEllaa26 andOprF::MEl5aa26 will further define the minimal length requirement for theimmunogenicity of the epitope andlor the relationship between antigenicity andimmunogenicity.Using the same logic, the low antigenicity of the epitope inserted at aa215could have been indicative of the lack of immunogenicity of this epitope in thecontext of OprF: :ME lOaa2 15. However, it is not likely that the intrinsic factorssuch as the antigenicity of the epitope were the sole factors affecting itsimmunogenicity in this case. Using the filamentous phage pill protein as carrier,Cruz et al., (1988) reported that although the inserted malarial epitope is antigenicin vitro, it is not necessarily immunogenic when administered to mice. An earlierstudy had also revealed that the malarial epitope is recognized as a T cell epitopeonly in mouse strains with 112b and H2k backgrounds (Good et al., 1986). Sincesignificant anti-OprF response was elicited in the BALB/c mice (H-2”) used in thisstudy, it indicates that OprF could recruit the required T cell response for theepitope. Therefore, the non-responsiveness of the animals immunized withOprF::MElOaa2l5 was likely due to inadequate B cell activation. Of course, theuse of a lower dosage (10 .tg instead of 25 jig as used in the other study) and the useof adjuvant might have also contributed to the non-responsiveness of the154immunized animals. However, the impact of these factors could not be easilyevaluated by the simple design of the experiment in this study.Successful antigen processing and presentation of the T cell epitope aretwo essential steps involved in an efficient antibody response. It has beendocumented that sequences outside a minimal epitope can affect the products ofprocessing and thus may determine whether a peptide is selected for presentation(Van der Werfet al., 1990; Del Valet al., 1991; Janssen et al., 1994b). Now that aneffective immunization protocol has been established, similar studies using theother 2 series of multiple-repeat hybrids carrying the epitope at aa’96 and aa213might provide information about the length requirement at these two sites. Acomparison of such results should elucidate the position effect (if one exists) on theimmunogenicity of the epitope in the OprF presentation system. Moreover, theseresults might also increase our general understanding of the effect of flankingamino acid residues on immunogenicity. However, such studies were beyond thescope of the current thesis.The inability of the GST::malarial epitope fusion proteins to generate ananti-malarial epitope response was unexpected (Table XVI). Since the protocolutilized for the immunogenicity study of the GST::malarial epitope fusion proteinswas the same as that for the OprF::MEaa26 hybrid proteins, it is not likely thatextrinsic factors such as the genetic background of the animals, the use of adjuvant,and the route of delivery were involved in the non-responsiveness of the immunizedanimals. The length of the epitope should not be a limiting factor either because155the length of the epitope m GST::ME19 is comparable to that in OprF::MEl9aa26.One possible explanation could be that the conformation of the epitope in theGST::malaria]. epitope fusion proteins prevents its interaction with the componentsof the immune system. Based on the previous experience in this laboratory,attempts made to release a fusion peptide from the GST protein by proteasecleavage at an engineered recognition site were unsuccessful, presumably due tothe folding of the epitope in the fusion protein which limits the accessibility of thecleavage site (Piers, 1993). Similarly, antigen processing also requires cleavage ofthe original proteins by proteases. Therefore, one is tempted to speculate that thefolding of the epitope in vivo might have rendered the protease cleavage siteinaccessible for antigen processing, resulting in its failure to elicit an efficientantibody response.In spite of its inability to stimulate an anti-malarial epitope antibodyresponse, the GST::malarial epitope fusion proteins were useful for the detectionof anti-epitope titers in antisera from animals immunized with the OprF::malarialepitope hybrid proteins. ELISA using the fusion protein as coating antigen for thedetermination of anti-malarial epitope titers was more sensitive than that using thesynthetic peptide (NANP)3as coating antigen (Table XV). The same difference insensitivity was also observed in ELISA using OprF::malarial epitope hybrid proteinas coating antigen as compared to that using the same synthetic peptide (TableXV1). On the other hand, the synthetic peptide could be recognized by both of themalarial epitope-specific mAbs pf2A. 10 and pf5A4. 1, suggesting that it is antigenic.156These findings are consistent with the general knowledge that synthetic peptidesalone do not interact effectively with antibodies, probably due to inadequatepresentation.Despite the apparent versatility of the OprF epitope presentation system,there were difficulties encountered in the course of this study that can be potentialpitfalls of the system. For example, when E. coli cells expressing the OprF hybridproteins were used as antigens in ELISA, the non-specificity binding of the epitopespecific monoclonal antibodies to whole cells was too high to permit any meaningfulinterpretation of the data. This problem was more prominent with the mAbMA5A4. 1, which appeared to have lower affinity for the inserted epitope. Titrationstudy of this monoclonal antibody with E. coli outer membrane containing OprF orone of the OprF::malarial epitope hybrid proteins showed that the dilution of theantibody that allowed reasonable sensitivity also resulted in significant backgroundreactivity. Therefore, if whole cells containing the OprF hybrid proteins are to beused as antigens for the detection of anti-epitope antibody, the non-specificity ofantibody binding should be carefully monitored.The instability of the plasmids encoding OprF::malarial epitope multiple-repeat hybrids in C 158 became a significant concern at the later stage of this study.Recombinant C158 strains had the tendency to lose the plasmids upon subculturingand subsequent growth in liquid medium. The ability of the host strain alone togrow in the presence of ampicillin remains puzzling. Since hydrophilic -lactamantibiotics such as ampiciilin are believed to be taken up by the bacterial cells via157the porin pathway, the lack of porins in the strain C158 may increase theampicillin resistance of the bacteria. An attempt to enhance plasmid stability byraising the ampicillin concentration in the medium to 200 jig/nil was unsuccessful.Therefore, efforts to improve the stability of the recombinant plasmids in the hoststrain would be necessary to ensure the efficiency of the OprF epitope presentationsystem.Using the malarial epitope as a model epitope, this study has shown thatthe OprF epitope presentation system can be used to raise and detect malarialepitope-specific antibodies. In addition, it has also been demonstrated that theOprF and GST systems can be used in a complementary fashion as a set of simpleand flexible tools to induce and monitor an anti-peptide response without the useof synthetic peptides. The ability of OprF to promote immunogenicity of a foreignepitope and its potential as a vaccine against P. aeruginosa infections suggestedthat it has the possibility for the development of a multivalent vaccine.In the course of this study, and using one of the OprF linker mutantplasmid vectors described here, a Pseudomonas pilin epitope was inserted in thecontext of OprF and shown to be recognized by its specffic antibody in ELISA (B.Finlay, personal communication). Currently, OprF is being used to express randomfragments of the ifiamentous haemagluttinin (FHA) gene of Bordetella pertussisand has shown some promising results for the mapping of antibody bindingepitope(s) and functional domains of the protein (A. Siebers and B. Finlay,unpublished results). Attempts are also being made to utilize OprF as a carrier for158a neutralizing epitope of TSST-1, the Toxic Shock Syndrome Toxin-i ofStaphylococcus aureus (E. Rubinchik and A. Chow, personal communication). Theincreased use of the OprF epitope presentation system should expand the repertoireof foreign epitopes that can be presented and further explore the potentialapplications of this system.159REFERENCESAgterberg, M., H. Adriaanse, H. Lankhof, R. Meloen, and J. Tommassen. 1990a.Outer membrane PhoE protein of Escherichia coli as a carrier for foreignantigenic determinants: immunogenicity of epitopes of foot-and-mouthdisease virus. Vaccine 8: 85-9 1.Agterberg, M., H. Adriaanse, A. van Bruggen, M. Karperien, and J. Tommassen.1990b. Outer-membrane PhoE protein of Escherichia coli K- 12 as anexposure vector: possibilities and limitations. Gene 88: 37-45.Angus, B.L., A.M. Carey, D.A. Caron, A.M.B. Kropinski, and R.E.W. Hancock.1982. Outer membrane permeability in Pseudomonas aeruginosa:comparison of a wild-type with an antibiotic-suspersusceptible mutant. Anti.Agents Chemother. 21: 299-309.Ausubel, F. M., R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith andK. Struhi. 1987. Current Protocols in Molecular Biology. Greene PublishingAssociates and Wiley-Interscience, New York, New York.Baflou, W. R., J. Rothbard, R.A. Wirtz, D.M. Gordan, J.S. Williams, R.W. Gore, I.Schneider, M.R. Hoffingdale, R.L. Beaudoin, W.L. Maloy, L.H. lVIiller, W.T.Hockmeyer. 1985. Immunogenicity of synthetic peptides fromcircumsporozoite protein of Plasmodium falciparum. Science 228: 996-999.Barbas, C.F., A.S. Kang, R.A. Lerner and S.J. Benkovic. 1991. Assembly ofcombinatorial antibody libraries on phage surfaces: the gene III site. Proc.Nati. Acad. Sci. USA 88: 7978-7982.Bardwell, J.C.A.. 1994. Building bridges: clisulphide bond formation in the cell.Molec. Microbio. 14: 199-205.Beffido, F., N.L. Martin, R.J. Siehnel, and R.E.W. Hancock. 1992. Reevaluation inintact cells of the exclusion limit and role of porin OprF in Pseudomonasaeruginosa outer membrane permeability. J. Bacteriol. 174: 5 196-5203.Berzofsky, J.A. 1985. Intrinsic and extrinsic factors in protein antigenic structure.Science 229: 932-939.Bosch, D. and J. Tommassen. 1987. Effects of linker insertions on the biogenesisand functioning of the Escherichia coli outer membrane pore protein PhoE.Mol. Gen. Genet. 208: 485-489.160Boulain, J.C., A. Charbit, and M. Hofnung. 1986. Mutagenesis by random linkerinsertion into the lamB gene of Escherichia coli K12. Mol. Gen. Genet. 205:339-348.Broekhuijsen, M.P., T. Blom, J. van Rijn, P.R. Pouwels, E.A. Kiasen, M.J.Fasbender and B.E. Enger-Valk. 1986. Synthesis of fusion proteins withmultiple copies of an antigenic determinant of foot-and-mouth disease virus.Gene 49: 189-197.Brooks, B.R., R.W. Pastor, and F.W. Carson. 1987. Theoretically determined three-dimensional structure for the repeating tetrapeptide unit of thecircumsporozoite coat protein of the malaria parasite Plasmodiumfalciparum. Proc. Nati. Acad. Sci. USA 84: 4470-4474.Brown, J.H., T. Jardetzky, M.A. Saper, B. Samraoui, P.J. Bjorkman, and D.C.Wiley. 1988. A hypothetical model of the foreign antigen binding site of classII histocompatibility molecules. [Correction in Nature 333: 786]. Nature 332:845-850.Brown, S. 1992. Engineered iron oxide-adhesion mutants of the Escherichia coliphage A receptor. Proc. Natl. Acad. Sci. USA 89: 865 1-8655.Calvo-Calle, J.M., G.A. de Oliveira, P. Clavijo, M. Maracic, J.P. Tam, Y.A. Lu, E.H.Nardin, R.S. Nussenzweig, and A.H. Cochrane. 1993. Immunogenicity ofmultiple antigen peptides containing B and non-repeat T cell epitopes of thecircumsporozoite protein of Plasmodium falciparum. J. Immunol. 150: 1403-1412.Caulcott, C.A., M.R.W. Brown, and I. Gonda. 1984. Evidence for small pores in theouter membrane of Pseudomonas aeruginosa. FEMS Microbiol. Lett. 21: 119-123.Charbit, A., J. Ronco, V. IVlichel, C. Werts, and M. Hofnung. 1991. Permissive sitesand topology of an outer membrane protein with a reporter epitope. J.Bacteriol. 173: 262-275.Charbit, A., E. Sobczak, M. lVlichel, A. Molla, P. Tiollais, and M. Hofnung. 1987.Presentation of two epitopes of the preS2 region of hepatitis B virus on liverecombinant bacteria. J. Immunol. 139: 1658-1664.Charbit, A., A. Molla, J. Ronco, J. Clement, V. Favier, E.M. Bahraoui, L.Montagnier, A. Leguern, and M. Hofnung. 1990. Immunogenicity andantigenicity of conserved peptides from the envelope of HIV- 1 expressed atthe surface of recombinant bacteria. AIDS 4: 545-551.161Charbit, A., P. Martineau, J. Ronco, C. Leclerc, R. Lo-Man, V. Michel, D.O’Callaghan, and M. Hofnung. 1993. Expression and immunogenicity of theV3 loop from the envelope of human immunodeficiency virus type 1 in anattenuated aroA strain of Salmonella typhimurium upon genetic coupling totwo Escherichia coli carrier proteins. Vaccine 11: 1221-1228.Chen, Y-H. U., R.E.W. Hancock, and R.I. Mishell. 1980. IVlitogenic effects ofpurified outer membrane proteins from Pseudomonas aeruginosa. Infect.Immun. 28: 178-184.Collawn, J.F., C.J. Wallace, A.E. Proudfoot, and Y. Paterson. 1988. Monoclonalantibodies as probes of conformational changes in protein-engineeredcytochrome c. J. Biol. Chem. 263: 8625-8634.Cowan, S.W., T. Schirmer, G. Rummel, M. Steiert, R. Ghosh, R.A. Pauptit, J.N.Jansonino, and J.P. Rosenbusch. 1992. Crystal structures explainfunctional properties of two .E. coli porins. Nature 358: 727-733.Crompton, J., C.I.A. Toogood, N. Waffis, and R. T. Hay. 1994. Expression of aforeign epitope on the surface of the adenovirus hexon. J. Gen. Virol. 75: 133-139.Cruz, de la. V.F., A.A. Lal, and T.F. McCutchan. 1988. Immunogenicity and epitopemapping of foreign sequences via genetically engineered ifiamentous phage.J. Biol. Chem. 263: 4318-4322.Dame, J. B., J.L. Wiffiams, T.F. McCutchan, J.L. Weber, R.A. Wirtz, W.T.Hockmeyer, W.L. Maloy, J.D. Haynes, I. Schneider, D. Roberts, G.S.Sanders, E. P. Reddy, C.L. Diggs, L.H. Miller. 1984. Structure of the geneencoding the immunodominant surface antigen on the sporozoite of thehuman malaria parasite Plasmodium falciparum. Science 225: 593-599.De Mot, R. and J. Vanderleyden. 1994. The C-terminal sequence conservationbetween OmpA-related outer membrane proteins and MotB suggests acommon function in both Gram-positive and Gram-negative bacteria,possibly in the interaction of these domains with peptidoglycan. Mol.Microbiol. 12: 333-334.Del Val, M., H.-J., Schlicht, T. Ruppert, M.J. Reddehase, and U.H. Koszinowski.1991. Efficient processing of an antigenic sequence for presentation to MHCclass I molecules depends on its neighbouring residues in the protein. Cell66: 1145-1153.162Der Vartanian, M., M.-C., Méchin, B. Jaffeux, Y. Bertin, I. Felix and B. GaillardMartinie. 1994. Permissible peptide insertions surrounding the signalpeptide-mature protein junction of the Cp1G prepilin: CS3 1A fimbriae ofEscherichia coli as carriers of foreign sequences. Gene 148: 23-32.Duchene, M., A. Schweizer, F. Lottspeich, G. Krauss, M. Marget, K. Vogel, B-U.von Specht, and H. Domdey. 1988. Sequence and transcriptional start siteof the Pseudomonas aeruginosa outer membrane porin protein F gene. J.Bacteriol. 170: 155-162.Dyson, H.J., A.C. Satterthwait, R.A. Lerner, and P.E. Wright. 1990.Conformational preferences of synthetic peptides derived from theimmunodominant site of the circumsporozoite protein of Plasmodiumfalciparum. Biochemistry. 29: 7828-37.Elish, M.E., J.R. Pierce, and C.F. Earhart. 1988. Biochemical analysis ofspontaneous fepA mutants of Escherichia coli. J. Gen. Microbiol. 134: 1355-1364.Evans, D.J., J. Mckeating, J.M. Meredith, K.L. Burke, K. Katrak, A. John, M.Ferguson, P.D. Minor, R.A. Weiss, J.W. Almond. 1989. An engineeredpoliovirus chimeara elicits broadly reactive HIV- 1 neutralizing antibodies.Nature 339: 385-388.Fasman, G.D., K. Park, and D.H. Schlesinger. 1990. Conformational analysis of theimmuno dominant epitopes of the circumsporozoite protein of Plasmodiumfalciparum and knowlesi. Biopolymers 29: 123-130.Finnen, R.L., N.L. Martin, R.J. Siehnel, W.A. Woodruff, M. Rosok, and R.E.W.Hancock. 1992. Analysis of the major outer membrane protein OprF fromPseudomonas aeruginosa using truncated OprF derivatives and monoclonalantibodies. J. Bacteriol. 174: 4977-4985.Foulds, J. and T-J. Chai. 1979. Isolation and characterization of isogenic E. colistrains with alterations in the level of one or more major outer membraneproteins. Can. J. Microbiol. 25: 423-427.Francisco, J.A., R. Campbell, B.L. Iverson, and G. Georgiou. 1993. Production andfluorescence-activated cell sorting of Escherichia coli expressing a functionalantibody fragment on the external surface. Proc. Natl. Acad. Sci. USA. 90:10444-10448.163Freud, R., S. Maclntyre, M. Degen, and U. Henning. 1986. Cell surface exposureof the outer membrane protein OmpA of Escherichia coli K12. J. Mol. Biol.188: 491-494.Gammon, G., N. Shastri, J. Cogswell, S. Wilbur, S. Sadegh-Nasseri, U. Krzych, A.Miller, and E. Sercarz. 1987. The choice of T-cell epitopes utilized on aprotein antigen depends on multiple factors distant from, as well as at thedeterminant site. Immunol. Rev. 98: 53-73.Gascuel, 0. and J. L. Golmard. 1988. A simple method for predicting thesecondary structure of globular proteins: implications and accuracy.Computer Applications in the Biosciences 4:357-365.Geller, B., H.-Y. Zhu, S. Cheng, A. Kuhn, and R.E. Dalbey. 1993. Charged residuesrender pro-OmpA potential dependent for initiation of membranetranslocation. J. Biol. Chem. 268: 9442-9447.Gibson, K.D. and H.A. Scheraga. 1986. Predicted conformations for theimmunodominant region of the circumsporozoite protein of the humanmalarial parasite Plasmodium falciparum. Proc. Natl. Acad. Sci. USA 83:5649-56 53.Gifieland, Jr. H.E., M.G. Parker, J.M. Matthews, and R.D. Berg. 1984. Use of apurified outer membrane protein F (porin) preparation of Pseudomonasaeruginosa as a protective vaccine in mice. Infect. Immun. 44: 49-54.Gilleland, Jr., H.E., L.B. Gifieland, and J.M. Matthews-Greer. 1988. Outermembrane protein F preparation of Pseudomonas aeruginosa as a vaccineagainst chronic pulmonary infection with heterologous immunotype strainsin a rat model. Infect. Immun. 56: 10 17-1022.Good, M.F., J.A. Berzofsky, W.L. Maloy, Y. Hayashi, N. Fujii, W.T. Hockmeyer, andL.H. Miller. 1986. Genetic control of the immune response in mice to aPlasmodium falciparum sporozoite vaccine. J. Exp. Med. 164: 655-660.Good, M.F., W.L. Maloy, M.N. Lunde, H. Margalit, J.L. Cornette, G.L. Smith, B.Moss, L.H. Miller, and J.A. Berzofsky. 1987. Construction of syntheticimmunogen: use of new T-helper epitope on malaria circumsporozoiteprotein. Science 235: 1059-1062.Gotoh, N., H. Wakebe, E. Yoshihara, T. Nakae, and T. Nishino. 1989. Role ofprotein F in maintaining structural integrity of the Pseudomonas aeruginosaouter membrane. J. Bacteriol. 171: 983-990.164Gregoriadis, G. 1990. Immunological adjuvants: a role for liposomes. Immunologytoday 3: 29-39.Guillet, J.G., M.-Z. Lai, T.J. Briner, J.A. Smith, and M.L., Gefter. 1986. Interactionof peptide antigens and classll major histocompatibioity complex antigens.Nature 324: 260-262.Hanahan, D. 1983. Studies on transformation of Escherichia coli with plasmids. J.Mol. Biol. 166: 557-580.Hancock, R.E.W. and H. Nikaido. 1978. Outer membranes of gram-negativebacteria: isolation from Pseudomonas aeruginosa PAO1 and use inreconstitution and definition of the permeability barrier. J. Bacteriol. 136:38 1-390.Hancock, R. E. W., and A. Carey. 1979. Outer membrane of Pseudomonasaeruginosa: Heat- and 2-mercaptoethanol-modifiable proteins. J. Bacteriol.140: 902-9 10.Hancock, R.E.W. and R. Benz. 1986. Demonstration and chemical modification ofa specific phosphate binding site in the phosphate-starvation-inducible outermembrane porin protein P. Biochimica et Biophysica Acta 860: 699-707.Hancock, R.E.W., G.M. Decad, and H. Nikaido. 1979. Identification of the proteinproducing transmembrane diffusion pores in the outer membrane ofPseudomonas aeruginosa PAO1. Biochim. Biophys. Acta 554: 323-329.Hansson, M., S. Stahl, T.N. Nguyen, T. Bächi, A. Robert, H. Binz, A. Sjölander, andM. Uhlén. 1992. Expression of recombinant proteins on the surface of thecoagulase-negative bacterium Staphylococcus xylosus. J. Bacteriol. 174:4239-4245.Hayes, L.J., J.W. Conlan, J.S. Everson, M.E. Ward, and I.N. Clarke. 1991.Chlamydia trachomatis major outer membrane protein epitopes expressedas fusions LamB in an attenuated aroA strain of Salmonella typhimurium;their application as potential immunogens. J. Gen. Microbiol. 137: 1557-1564.Hedegaard, L., and P. Klemm. 1989. Type 1 fimbriae of Escherichia coli as carriersof heterologous antigenic sequences. Gene 85: 115-24.Hopp, T.P., and K.R., Woods. 1981. Prediction of protein antigenic determinantsfrom amino acid sequences. Proc.Natl.Acad.Sci. USA 78: 3824-3828.165Janssen, R. and J. Tommassen. 1994. PhoE protein as a carrier for foreign epitopes.Intern. Rev. Immunol. 11: 113-12 1.Janssen, R., M. Wauben, R. van der Zee, and J. Tommassen. 1994a.Immunogenicity of a mycobacterial T-cell epitope expressed in outermembrane protein PhoE of Escherichia coli. Vaccine 12: 406-409.Janssen, R., M. Wauben, R. van der Zee, M. de Gast, and J. Tommassen. 1994b.Influence of amino acids of a carrier protein flanking an inserted T celldeterminant on T cell stimulation. Intern. Immunol. 6: 1187-1193.Jeanteur, D., J.H. Lakey, and F. Pattus. 1991. The bacterial porin superfamily:sequence alignment and structure prediction. Molec. lVlicrobiol. 5: 2153-2164.Jenkins, 0., J. Cason, K.L. Burke, D. Lunney, A. Gifien, D. Patel, D.J. McCanceand J.W. Almond. 1990. An antigen chimera of poliovirus induces antibodiesagainst human papifiomavirus type 16. J. Virol. 64: 120 1-1206.Karplus, P. A. and G. E. Schulz. 1985. Prediction of chain flexibility in proteins.A tool for the selection of peptide antigens. Naturwissenschaften. 72:2 12-213.Kuhn, A., D. Kiefer, C. Köhne, H.-Y. Zhu, W.R. Tschantz, and R. E. Dalbey. 1994.Evidence for a loop-like insertion mechanism of pro-OmpA into the innermembrane of Escherichia coli. Eur. J. Biochem. 226: 89 1-897.Lenstra, J.A., J.H. Erkens, J.G. Langeveld, W.P. Posthumus, R.H. Meloen, F.Gebauer, I. Correa, L. Enjuanes, K.K. Stanley, 1992. Isolation of sequencesfrom a random-sequence expression library that mimic viral epitopes. J.Immunol. Methods. 152: 149-157.Leclerc, C., P. Martineau, S. van der Werf, E. Deriaud, P. Duplay, and M. Hofnung.1990. Induction of virus-neutralizing antibodies by bacteria expressing theC3 poliovirus epitope in the periplasm. J. Immunol. 144: 3174-3 182.Leclerc, C., A. Charbit, P. Martineau, E. Deriaud, and M. Hofnung. 1991. Thecellular location of a foreign B cell epitope expressed by recombinant bacteriadetermines its T cell-independent or T cell-dependent characteristics. J.Immunol. 147: 3545-3552.Li, P., J. Beckwith, and H. Inouye. 1988. Alteration of the amino terminus of themature sequence of a periplasmic protein can severely affect export inEscherjchja coli. Proc. Natl. Acad. Sci. USA 85: 7685-7689.166Lightfoot, J., and J.S. Lam. 1991. Molecular cloning of genes involved withexpression of A-band lipopolysaccharide, an antigenically conserved form, inPseudomonas aeruginosa. J. Bacteriol. 173: 5624-5630.Ling, I.T., S.A. Ogun, and A.A. Holder. 1994. Immunization against malaria witha recombinant protein. Parasite Immunology 16: 63-67.Maclntyre, S. and U. Henning. 1990. The role of the mature part of secretoryproteins in translocation across the plasma membrane and in regulation oftheir synthesis in Escherichia coli. Biochimie 72: 157-167.McCafferty, J., A.D. Griffiths, G. Winter and D.J. Chiswell. 1990. Phage antibodies:ifiamentous phage displaying antibody variable domains. Nature 348: 552-554.McCarvil, J., A.J. McKenna, C. Grief, C.S. Hoy, D. Sesarclic, M.C.J. Maiden, andI. M. Feavers. 1993. Expression of meningococcal epitopes in LamB ofEscherichia coli and the stimulation of serosubtype-specific antibodyresponses. Mol. Micro. 10: 203-2 13.McCutcheon J.A., K.D. Smith, A. Valenzuela, K. Aalbers, and C.T. Lutz. 1993.HLAB*0702 antibody epitopes are affected indirectly by distant antigenresidues. Human Immunology. 36: 69-75.McEwen, J., R. Levi, R.J. Horwitz, and R. Arnon. 1992. Synthetic recombinantvaccine expressing influenza haemagglutinin epitope in Salmonella flageffinleads to partial protection in mice. Vaccine 10: 405-411.Manoil, C. 1991. Analysis of membrane protein topology using alkalinephosphatase and -galactosidase gene fusions. Methods in cell biology. 34:6 1-75.Mansfield, M.A. 1994. Rapid immunodetection on hydrophobic PVDF membraneswithout blocking. Poster presented at The American Society ForBiochemistry And Molecular Biology, Washington, DC, May 1994.Martin, N.L., E.G. Rawling, R.S.Y. Wong, M. Rosok, and R.E.W. Hancock. 1993.Conservation of surface epitopes in Pseudomonas aeruginosa outermembrane porin protein OprF. FEMS Microbiol. Lett. 113: 26 1-266.Martineau, P., A. Charbit, C. Leclerc, C. Werts, D. O’Callaghan, and M. Hofnung.1991. A genetic system to elicit and monitor anti-peptide antibodies withoutpeptide synthesis. Bio/Technology 9: 170-172.167Marston, F. A. 0. 1986. The purification of eukaryotic polypeptides synthesizedin Eseherichia coli. Biochem. J. 240: 1-12.Matthews-Greer, J.M. and H.E. Gilleland, Jr. 1987. Outer membrane protein F(porin) preparation of Pseudomonas aeruginosa as a protective vaccineagainst heterologous immunotype strains in a burned mouse model. J. infect.Dis. 155: 1282-1291.Michel, M.L., M. Mancini, K. Schlienger, and P. Tiollais. 1993. Recombinanthepatitis B surface antigen as a carrier of human immunodeficiency virusepitopes. Research in Virology. 144: 263-26 7.Minenkova 0.0., A.A. llyichev, G.P. Kishchenko, and V.A. Petrenko. 1993. Designof specffic immunogens using filamentous phage as the carrier. Gene 128:85-88.Moeck, G.S., B.S. Fazly Bazzaz, M.F. Gras, T.S. Ravi, M.J.H. Ratcliffe, and J.W.Coulton. 1994. Genetic insertion and exposure of a reporter epitope in theferrichrome-iron receptor of Escherichia coli K-12. J. Bacteriol. 176: 4250-4259.Monaco, J.J. 1992. A molecular model of 1VIIIC class-I-restricted antigen processing.Immunology today. 13: 173-179.Mutharia, L.M. and R.E.W. Hancock. 1983. Surface localization of Pseudomonasaeruginosa outer membrane porin protein F by using monoclonal antibodies.Infect. Immun. 42: 1027-1033.Mutharia, L. M. and R.E.W. Hancock. 1985. Characterization of two surface-localized antigenic sites on porin protein F of Pseudomonas aeruginosa. Can.J. Microbiol. 31: 38 1-386.Nakae, T.. 1995. Role of membrane permeability in determining antibioticresistance in Pseudomonas aeruginosa. Microbiol. Immunol. 39: 221-229.Newton, S. M. C., C. 0. Jacob, and B. A. D. Stocker. 1989. Immune response tocholera toxin epitope inserted in Salmonella flageffin. Science 244: 70-72.Nicas, T.I. and R.E.W. Hancock. 1983. Alteration of susceptibility to EDTA,polymyxin B and Gentan’ucin in Pseudomonas aeruginosa by divalent cationregulation of outer membrane protein Hi. J. Gen. lVlicrobiol. 129: 509-5 17.168Nikaido, H. and M. Vaara. 1987. Outer membrane, p. 7-22. In F.C. Neidhardt,J.L., Ingraham, K.B. Low, B. Magasanik, M. Schaechter, and H.E. Umbarger(ed.), Escherichia coli and Salmonella typhimurium: cellular and molecularbiology, vol. I. American Society for Microbiology, Washigton , D.C.Nikaido, H., K. Nikaido, and S. Harayama. 1991. Identification andcharacterization of porins in Pseudomonas aeruginosa. J. Biol. Chem. 266:770-779.Notley, L., C. Hillier, and T. Ferenci. 1994. Epitope mapping by cysteinemutagenesis: identification of residues involved in recognition by threemonoclonal antibodies directed against LamB glycoporin in the outermembrane of Escherichia coli. FEMS Microbio. Lett. 120: 341-348.Oka, A., H. Sugisaki, and M. Takanami. 1981. Nucleotide sequence of thekanamycin resistance transposon Tn903. J. Mol. Biol. 147: 217-226.Ott, L. In An introduction to statistical methods and data analysis, pp300-33l.PWS-KENT publishing Co., Boston.Paul, C., and J.P. Rosenbusch. 1985. Folding patterns of porin andbacteriorhodopsin. EMBO J. 4: 1593-1597.Percy, N., W.S. Barclay, A. Garcia-Sastre, P. Palese. 1994. Expression of a foreignprotein by influenza A virus. J. Virol. 68: 4486-92.Pessi, A., D. Valmori, P. Migliorini, C. Tougne, E. Bianchi, P.H. Lambert, G.Corradin, and G. Del Giudice. 1991. Lack of H-2 restriction of thePlasmodium falciparum (NANP) sequence as multiple antigen peptide. Eur.J. Immunol. 21: 2273-2276.Piers, K.L. 1993. Ph.D. thesis. University of British Columbia, Vancouver, Canada.Piers, K.L., M.B. Brown, and R.E.W. Hancock. 1993. Recombinant DNAprocedures for producing small antimicrobial cationic peptides in bacteria.Gene 134: 7-13.Pozzi, G., M. Contorni, M. R. Oggioni, R. Manganell, M. Tommasino, F. Cavalieri,and V. A. Fischetti. 1992. Delivery and expression of a heterologous antigenon the surface of Streptococci. Infect. Immun. 60: 1902-1907.Rawling, E.G., N.L. Martin, and R.E.W. Hancock. 1995. Epitope mapping of thePseudomonas aeruginosa major outer membrane porin protein OprF. Infect.Immun. 63: 1995 38-42.169Rickman, L.S. and S.L. Hoffman. 1991. Malaria, p.1037. In S. Baron (ed.), Medicalmicrobiology. Churchifi Livingstone Publishing Co., New York.Rutgers, T., D. Gordon, A.M. Gathoye, M. Hoffingdale, W. Hockmeyer, M.Rosenberg and M. De Wilde. 1988. Hepatitis B surface antigen as carriermatrix for the repetitive epitope of the circumsporozoite protein ofPlasmodium falciparum. Bio/technology 6: 1065-1070.Sambrook, J., E.F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratorymanual, 2nd ed. Cold Spring Harbour Laboratory, Cold Spring, Harbour,N.Y.Sandermann , H. Jr. and J. L. Strominger. 1972. Purification and properties of C55-isoprenoid alcohol phosphokinase from Staphylococcus aureus. J. Biochem.Chem. 247:5123-5131.Schirmer, T., T. A. Keller, Y.-F. Wang, and J. P. Rosenbusch. 1995. Structural basicfor sugar translocation through maltoporin channels at 3.1 A resolution.Science 267: 512-514.Schnaitman, C.A. 1971. Solubilization of the cytoplasmic membrane of Escherichiacoli by Triton X-100. J. Bacteriol. 108: 545-552.Schorr, J., B. Knapp, E. Hundt, H. A. Kupper, and E. Amann. 1991. Surfaceexpression of malarial antigens in Salmonella typhimurium: induction ofserum antibody response upon oral vaccination of mice. Vaccine 9: 675-681.Scott, J.K. and L. Craig. 1994. Random peptide libraries. Curr. Opin. in Biotechnol.5: 40-48.Siehnel, R. J., N. L. Martin, and R. E. W. Hancock. 1990. Function and Structureof the porin proteins OprF and OprP of Pseudomonas aeruginosa. pp. 328-342. In Pseudomonas biotransformations, pathogenesis, and evolvingbiotechnology. Silver, S. (ed.). Washington, D.C. American Society ofMicrobiology.Smith, D.B., K.M. Davern, P.G. Board, W.U. Tiu, E.G. Garcia and G.F. Mitchell.1986. Mr 26,000 antigen of Schistosomajaponicum recognized by resistantWEHI 129/J mice is a parasite glutathione S-transferase. Proc. Natl. Acad.Sci. USA 83: 8703-870 7.Smith, D. B. and K. S. Johnson. 1988. Single-step purification of polypeptidesexpressed in Escherichia coli as fusions with glutathione S-transferase. Gene67: 3 1-40.170Smith, D.B., M.R. Rubira, R.J. Simpson, K.M. Davern, W.U. Tiu, P.G. Board, andG.F. Mitchell. 1988. Expression of an enzymatically active parasite moleculein Escherichia coli: Schistosoma japonicum glutathione S-transferase. Mol.Biochem. Parasitol. 27: 249-256.Smith, A.D., D.A. Resnick, A. Zhang, S.C. Geisler, E. Arnold, and G.F. Arnold.1994. Use of random systemic mutagenesis to generate viable humanrhinovirus 14 chimeras displaying human immunodeficiency virus type 1 V3loop sequences. J. Virol. 68: 575-579.Sonntag, I., H. Schwarz, Y. Hirota, and U. Henning. 1978. Cell envelope and shapeof Escherichia coli: multiple mutants missing the outer membranelipoprotein and other major outer membrane proteins. J. Bacteriol. 136: 280-285.Steidler, L., E. Remaut, and W. Fiers. 1993. Pap pu as a vector system for surfaceexposition of an immunoglobulin G-bincling domain of protein A ofStaphylococcus aureus in Escherichia coli. J. Bacteriol. 175: 7639-7643.Stoute, J.A., W.R. Ballou, N. Kolodny, C.D. Deal, R.A. Wirtz, and L.E. Lindler.1995. Induction of humoral immune response against Plasmodiumfalciparum sporozoites by immunization with a synthetic peptide minotopewhose sequence was derived from screening a ifiamentous phage epitopelibrary. Infect. Immun. 63: 934-939.Struyvé, M., D. Bosch, J. Visser, and J. Tommassen. 1993a. Effect of differentpositively charged amino acids, C-terminally of the signal peptidase cleavagesite, on the translocation kinetics of a precursor protein in Escherichia coliK-12. FEMS Microbiol. Lett. 109: 173-178.Struyvé, M., J. Visser, H. Adriaanse, R. Benz, and J. Tommassen. 1993b. Topologyof PhoE porin: the ‘eyelet’ region. Molec. Microbiol. 7: 13 1-140.Szmelcman, S. and M. Hofnung. 1975. Maltose transport in Escherichia coli K-12:Involvement of the bacteriophage lambda receptor. J. Bacteriol. 124: 112-118.Taylor, I.M., J.L. Harrison, K.N. Timmis, and C.D. O’Connor. 1990. The TraTlipoprotein as a vehicle for the transport of foreign antigenic determinantsto the cell surface of Escheichia coli Kl2: structure-function relationships inthe TraT protein. Mol. Microbiol. 4: 1259-1268.Tsunetsugu-Yokota, Y., M. Tatsumi, V. Robert, C. Devaux, B. Spire, J-C.Chermann and I. Hirsch. 1991. Expression of an immunogenic region of HIVby a ifiamentous bacteriophage vector. Gene 99: 261-265.171Usha, R., J.B. Rohil, V.E. Spall, M. Shanks, A.J. Maule, J.E. Johnson, and G.P.Lomonossoff. 1993. Expression of an animal virus antigenic site on thesurface of a plant virus particle. Virology. 197: 366-374.Van der Werf, S., A. Charbit, C. Leclec, V. Mimic, J. Ronco, M. Girard, and M.Hofnung. 1990. Critical role of neighbouring sequences on theimmunogenicity of the C3 poliovirus neutralization epitope expressed at thesurface of recombinant bacteria. Vaccine 8: 269-2 77.Van Die, I., J. van Oosterhout, I. van Megen, H. Bergmans, W. Hoekstra, B. EngerValk, S. Barteling, and F. Mooi. 1990. Expression of foreign epitopes in Pfimbriae of Escherichia coli. Mol. Gen. Genet. 222: 297-303.von Specht, B.-H., B. Knapp, G. Muth, M. Broker., K.-D. Hungerer, K.-D. Diehl, K.Massarrat, A. Seemann, and H. Domdey. 1995. Protection ofimmunocompromised mice against lethal infection with Pseudomonasaeruginosa outer membrane protein F and outer membrane protein I fusionproteins. Infect. Immun. 63: 1855-1862.Weiss, M.S., U. Abele, J. Weckesser, W. Welte, E. Schiltz, and G.E. Schulz. 1991.Molecular architecture and electrostatic properties of a bacterial porin.Science 254: 1627- 1630.Wirtz, R.A., F. Zavala, Y. Charoenvit, G.H. Campbell, T.R. Burkot, I. Schneider,K.M. Esser, R.L. Beaudoin, and R.G. Andre. 1987. Comparative testing ofmonoclonal antibodies against Plasmodium falciparum sporozoites forELISA development. Bull. W.H.O. 65: 39-45.Woodruff, W. A., T. R. Parr,Jr., R. E. W. Hancock, L. F. Hanne, T. I. Nicas, and B.H. Iglewski. 1986. Expression in Escherichia coli and function ofPseudomona,s aeruginosa outer membrane porin protein F. J. Bacteriol. 167:473-479.Woodruff, W. A. and R. E. W. Hancock. 1989. Pseudomonas aeruginosa outermembrane protein F: structural role and relationship to the Escherichia coliOmpA protein. J. Bacteriol. 171: 3304-3309.Wu, J. Y., S. Newton, A. Judd, and W. S. Robinson. 1989. Expression ofimmunogenic epitopes of hepatits B surface antigen with hybrid flageffinproteins by a vaccine strain of Salmonella. Proc. Nati. Acad. Sci. USA 86:4726-4730.Yoneyama, H. and T. Nakae. 1986. A small diffusion pore in the outer membraneof Pseudomonas aeruginosa. Eur. J. Biochem. 157: 33-38.172Young, J. F., W.T. Hockmeyer, M. Gross, W.R. Ballou, R.A. Wirtz, J.H. Trosper,R.L. Beaudoin, M.R. Hollingdale, L.H. Miller, C.L. Diggs, and M. Rosenberg.1985. Expression of Plasmodium falciparum circumsporozoite proteins inEscherichia coli for potential use in a human malaria vaccine. Science 228:958-962.Zavala, F., J.P. Tam, M.R. Hoffingdale, A.H. Cochrane, I. Quakyi, R.S.Nussenzweig, and V. Nussenzweig. 1985. Rationale for development of asynthetic vaccine against Plasmodium falciparum malaria. Science 228:1436-1440.


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