Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cloning, sequencing and physicochemical characterization of a hemagglutinin from Escherichia coli 09:H10:K99 Lutwyche, Peter 1993

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

Item Metadata


831-ubc_1993_fall_phd_lutwyche_peter.pdf [ 7.8MB ]
JSON: 831-1.0061792.json
JSON-LD: 831-1.0061792-ld.json
RDF/XML (Pretty): 831-1.0061792-rdf.xml
RDF/JSON: 831-1.0061792-rdf.json
Turtle: 831-1.0061792-turtle.txt
N-Triples: 831-1.0061792-rdf-ntriples.txt
Original Record: 831-1.0061792-source.json
Full Text

Full Text

CLONING, SEQUENCING AND PHYSICOCHEMICALCHARACTERIZATION OF A HEMAGGLUTININFROM Escherichia coil 09:H10:K99ByPETER LUTWYCHEBSc., University of Warwick, 1986A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDEPARTMENT OF CHEMISTRYWe accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAAugust 1993© Peter LutwycheIn 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 C H A/1 T The University of British ColumbiaVancouver, CanadaDate I 1+^cDE-6 (2/88)AbstractA mannose-resistant hemagglutinating (erythrocyte aggregating)protein was cloned from Escherichia coli 09:H10:K99. This hemagglutininwas determined to be different from the two mannose-resistanthemagglutinins that this strain was known to possess, F41 and K99, andwas named Heat Resistant Agglutinin 1 (HRA1) on the basis of one of itsphysical properties. The HRA1 gene present on the recombinant plasmidpETE1 was localized and identified unambiguously by subcloning. Thenucleotide sequence of the gene was determined and found to consist of a792 base pair open reading frame coding for a protein of 29 kDa. Thisprotein has a predicted prokaryotic N-terminal secretory signalsequence. The protein sequence derived from the genetic informationshowed no significant similarity to database protein sequences.Physical and chemical attempts to isolate HRA1 from the cell membranemet with limited success. N-terminal sequence analysis of a pronounced25 kDa band present on polyacrylamide gels of crude membranepreparations of bacteria harbouring pETE1 correlated with the predictedN-terminal amino acid sequence of HRA1 after cleavage of the signalpeptide. Four other open reading frames were found on the cloned DNAfragment. Three of the deduced proteins from these regions containedpredicted signal sequences and membrane associated areas and one showed50 % identity with E.coli lipoprotein, suggesting that a region of DNAassociated with outer membrane structure and function was cloned.Partition in various aqueous two-phase polymer systems showed thatexpression of HRA1 caused pronounced cell-surface changes in hostbacteria. Viscometric analysis showed that the agglutination eventiimediated by HRA1 was less pronounced than that of F41. This was likelydue to the fact that F41 is known to be an exposed high molecular weightmultivalent adhesin, whereas evidence suggests that HRA1 is a monovalentmolecule that is closely associated with the bacterial membrane.iiiTable of ContentsPageAbstract^ iiTable of Contents^ ivList of Figures viiList of Tables^ xiAcknowledgements xii1. Introduction^ 11.1. General Background^ 11.2. The Gram-Negative Bacterial Membrane^ 11.3. Bacterial Adhesins 51.4. E.coli Adhesins^ 71.4.1. Type 1 (Mannose sensitive) Adhesins^ Biochemistry^ Genetics Receptors 101.4.2. Mannose Resistant Fimbrial Adhesins^ Colonization Factors S Fimbriae^ P Fimbriae The K88 Antigen The K99 Antigen^ The F41 Adhesin Other E.coli Fimbrial Adhesins^ 221.4.3. Afimbrial Adhesins of E.coli^ The Z Antigens The NFA Group of Afimbrial Adhesins^ AIDA-1 (Adhesin Involved in DiffuseAdherence)-1^ AFA-1 (Afimbrial Adhesin I)^ 251.5. Viscometric studies of Bacterial Adhesins 261.5.1. Viscometric analysis of Erythrocyte Suspensions ^ 271.5.2. Viscometric analysis of Aeromonas salmonicida 438^ 281.5.3. Viscometric analysis of E.coli Adhesins^ 301.6. Partition of Bacterial Cells in Aqueous Two-phase Systems ^ 331.6.1. Principles of Aqueous Two-phasePolymer Partitioning^ 331.6.2. Bacterial Partition 342. Cloning of Hemagglutinin Gene 372.1. Background and Aims^ 372.2. Materials and Methods 382.2.1. Bacterial Strains and Growth Conditions^ 382.2.2. Enzymes and Reagents^ 382.2.3. Buffers^ 392.2.4. Plasmid 392.2.5. Antibodies 392.2.6. Gel Electrophoresis of DNA^ 40iv2.2.7. Gel Electrophoresis of Proteins^ 402.2.8. Western Blotting^ 422.2.9. DNA Extractions 422.2.10. Construction of E.coli 09:H10:K99 chromosomal DNALibrary^ 452.2.11. Library Screening^ 472.2.11.1. Hemagglutination Assay^ 472.2.11.2. Oligonucleotide Probe 482.2.11.3. Antibody Colony Blots 492.2.11.4. Immunoblotting of SDS-PAGE Samples^ 502.2.11.5. Enzyme-Linked Immunosorbent Assay 502.3. Results^ 532.3.1. Random Fragment Cloning^ 532.3.2. Subsequent Library Screening 542.3.2.1. Oligonucleotide Probe Hybridization^ 542.3.2.2. Antibody Colony Blots^ 572.3.2.3. Western Blotting 582.3.2.4. Enzyme-Linked Immunosorbent Assay^ 602.4. Discussion^ 623. Subcloning and Sequencing Hemagglutinin Gene^ 644.3.1.^Background^ 643.2.^Materials and methods^ 64^3.2.1.^Bacterial Strains, Growth conditions and Plasmids^ 643.2.2.^Electroporation 653.2.3.^Sequencing Polyacrylamide Gel Electrophoresis^ 663.2.4.^Restriction mapping^ 663.2.5.^Subcloning^ 673.2.6.^Nucleotide Sequencing 713.2.6.1.^Strategy^ 713.2.6.2.^Growth and Purification of SingleStranded DNA Sequencing Templates^ 733.2.6.3.^Sequencing Reactions^ 743.2.6.4.^Sequencing Data Handling 763.3.^Results^ 773.3.1.^Restriction Map of pETE1^ 773.3.2.^Subcloning and Localization of Agglutinin Gene^ 773.3.3.^Identification of Open Reading Frames^ 793.3.4.^Genebank Homology Searches^ 933.4.^Discussion 95Biochemical Studies on Proteins Encoded by pETE1^ 984.1.^Background^ 984.2.^Materials and Methods^ 984.2.1.^Bacterial Strains and Growth Conditions^ 984.2.2.^Physical Methods for the Isolation of theAgglutinin^ 994.2.2.1.^Temperature / Vortex Experiments^ 994.2.2.2.^Detergent Extraction Experiments 1014.2.2.3.^Osmotic Shock^ 1014.2.2.4.^Sonication 1024.2.3.^Erythrocyte Receptor Studies 1024.2.3.1.^Species Hemagglutination Profile^ 1024.2.3.2.^Binding Studies with Erythrocyte Ghosts ^ 1034.2.3.3. Glycophorin Experiments^ 1064.2.4. Maxicell Analysis of proteins encoded by pETE1^ 1064.2.5. N-terminal Sequencing of Hemagglutinin 1084.3. Results^ 1094.3.1. Membrane Association of Hemagglutinin^ 1094.3.2. Comparison with F41^ 1174.3.2.1. Agglutination Strengthand Specificity^ 1174.3.2.2. Erythrocyte Receptor Results^ 1194.3.2.3. Maxicell Results 1234.3.3. N-terminal Sequence^ 1234.4. Discussion^ 1275. Physicochemical Characterization of the Agglutinin ^ 1295.1. Background^ 1295.2. Materials and Methods^ 1295.2.1. Aqueous Two-Phase Partitioning^ 1295.2.1.1. Preparation of the Phase Systems^ 1305.2.1.2. Radiolabelling of Bacteria^ 1315.2.1.3. The Partitioning Experiment 1325.2.1.4. Bacteria-Erythrocyte BindingPartition Experiments^ 1335.2.2. Viscometry of Bacterial Strains 1335.2.2.1. Bacterial Strains 1335.2.2.2. Growth of Human ColonAdenocarcinoma Cells^ 1345.2.2.3. Viscometric Hemagglutination Assay^ 1345.2.3. Cell Electrophoresis^ 1355.2.4. Microscopy^ 1365.2.4.1. Electron Microscopy^ 1365.2.4.2. Optical Photomicroscopy 1365.3. Results^ 1385.3.1. Influence of pETE1 on Bacterial Partition^ 1385.3.2. Viscometric Analysis of HemagglutinationInduced by E.coli Strains^ 1485.3.3. Bacterial Electrophoretic Mobility^ 1545.3.4. Photomicrographs of the Agglutination Event^ 1555.3.5. Electron Microscope Results^ 1605.4. Discussion^ 1606. Concluding Discussion 1657. Abbreviations^ 1758. Glossary of Terms 1779. References^ 17910. Appendix 193viList of FiguresPage1. The envelope of gram-negative bacteria^ 22. E.coli protein secretory signal sequences 43. The type 1 fimbriae operon^ 94. The pap operon^ 125. Putative assembly mechanism of pap pili^ 146. Sialic acid^ 177. Comparison of N-terminal sequences^ 198. Glycophorin A^ 219. N-terminal homology of CS31A and K88^ 2210. Diagrammatic view of a couette type viscometer^ 2911. Adhesion event between E.coli F41 adhesin and erythrocytesas determined by viscometry^ 3112. Differential agglutination of MM and NN erythrocytes by F41adhesin as determined by viscometric assay^ 3213. Dextran^ 3514. Poly(ethylene glycol)^ 3515. Diagrammatic representation of random fragmentcloning procedure^ 4616. Diagrammatic representation of a generalized enzyme-linkedimmunoassay^ 5117. Hemagglutination microtitre assay^ 5418. N-terminal amino acid sequence of F41 and oligonucleotideprobe pool designed to it^ 5519. Southern blot of probe hybridized to partially digested F41chromosomal DNA^ 5620. Autoradiogram of blotted library colonies after hybridizationwith oligonucleotide probe^ 5721. Photograph of antibody colony blot^ 5822. Western blotting of bacteria against a-F41 and a-K99 antibodies..59viipage23. Graph of absorbance against antibody dilution for whole-cellELISA experiment^ 6124. Construction of pPL7 7025. Restriction map of pETE1^ 7826. Subclones of pETE1 showing open reading frame areas^ 8027. DNA sequence in the region of the plprotl putative start codon^ 8128. Nucleotide sequence and deduced amino acid structureof plprotl^ 8229. plprotl predicted secretory signal sequence^ 8330. Semi-graphical display of secondary structure prediction forplprotl by the methods of Garnier and Gascuel and Golmard^ 8531. Nucleotide sequence and deduced amino-acid structureof plprot2^ 8732. Construction of plasmid pPL7N^ 8833. Nucleotideof plprot3sequence and deduced amino acid structure90 34. Nucleotide sequence and deduced amino acid structureof plprot4 ^ 9135. Nucleotide sequence and deduced amino acid structureof plprot5 ^ 9236. Identity between plprotl and N.gonorrhoeae opacity protein^ 9337. Identity between plprot5 and E.coli major outer membranelipoprotein precursor^ 9438. Predicted antigenic determinant on plprotl^ 9639. Comparison of implicated and identified binding siteamino acid sequences^ 9740. SDS-PAGE and hemagglutination profiles of samples fromtemperature/vortex experiments^ 11041. SDS-PAGE of detergent-solublized proteins^ 11242. SDS-PAGE of proteins released by osmotic shock 11343. SDS-PAGE and hemagglutination profiles of crude membranefragments prepared by sonication^ 115viiipage44. SDS-PAGE and hemagglutination profiles of bacteriaharbouring various control plasmids^ 11645. Nitrocellulose immobilized erythrocyte ghost proteinsblotted against biotinylated bacterial proteins^ 12146. Nitrocellulose immobilized bacterial proteins blottedagainst biotinylated erythrocyte ghosts^ 12247. Coomassie stain and autoradiograph of SDS-PAGE ofmaxicells^ 12448. N-terminal sequence homology of plprotl^ 12649. Diagrammatic representation of cell electrophoresisapparatus^ 13750. Bacterial partition in [(10,7.5)5] and [(6,10)5]phase systems^ 13951. Partition differences of JM101(pETE1) in variousPEG/dextran systems^ 14052. Partition of various pETE1 subclones in [(7.4,4.7)55] ^ 14153. Partition of JM101(pPL7) in [(5,4)5] and [(7.4,4.7)55] ^ 14354. Bacterial partition in charge sensitive and dextran/ficollsystems^ 14455. Partition of B/R and B/R(pETE1) in PEG/dextran systems^ 14656. Bacteria-erythrocyte partition binding experiment^ 14757. Viscometric analysis of hemagglutination mediated byE.coli strains in the absence and presence of mannose^ 14958. Hemagglutinating behavior of various E.coli strains 15159. Increased hemagglutination of B/R(pETE1) at lowershear rates^ 15260. Aggregation of cob 201 cells mediated by B/R(pETE1)^ 15361. Photomicrograph of erythrocyte ghosts agglutinatedand distorted by E.coli F41^ 15662. Photomicrographs of agglutination of human erythrocyteghosts mediated by JM101(pETE1)^ 15763. Photomicrographs of agglutination of human erythrocyteghosts mediated by JM101^ 158ixpage64. Photomicrograph of JM101(pETE1) agglutinating colo 201cells in the presence of 10 mM mannose^ 15965. Electron micrographs of uranyl acetate stained bacteria^ 161xList of TablesPage1. Restriction enzymes utilized during restriction mappingand subcloning^ 672. Constructed pETE1 subclones^ 773. Predicted amino acid composition of plprotl^ 834. Comparison of plprotl secondary structure prediction^ 845. Disruption experiments^ 1006. Detergents utilized for solubilization of membrane protein^ 1017. Erythrocyte ghost-bacteria binding experiment^ 1058. Agglutination scale^ 1119. Agglutination profile for washed erythrocytes ofdifferent species^ 118xiAcknowledgementsI would like to thank Don Brooks for suggesting this project, forhis advice, encouragement and optimism throughout and also for manningthe microscope when things got rough! Thanks to everyone in the Brookslab, especially Ray for introducing me to partitioning, Rosemarie forhelp with the viscometry and John Cavanagh for sharing his accumulatedknowledge of protein chemistry and 60's pop songs. I'll fondly rememberthe Brooks lab parties and drinking sessions. Thanks to Tony Warren andeveryone in the cellulase group for always patiently answering my oftenbasic molecular biology questions. Thanks to Bonnie for help with theelectron microscope and most of all thanks to Jane for her constant truefriendship and much more. This thesis is dedicated to my parents, Joanand Ron.xii1. Introduction1.1. General BackgroundBacterial infections are ubiquitous in humans and other speciesand have been the focus of intense scientific scrutiny for many years.At present, although much has been learned about bacterial physiology,pathogenic bacteria still present an enormous medical and veterinaryproblem. For example, it has been estimated that one third of thedeaths of children under 5 in developing countries are due toenterotoxigenic bacterial infection (1).An important initial step in bacterial pathogenesis is theadherence of the bacterial cell to the host tissue. This adherenceinvolves recognition of specific receptor molecules on the host cellsurface by specialized bacterial membrane structures called adhesins.A large number of these bacterial systems have been described and someof their receptors elucidated. There follows an overview of thebacteria cell surface environment with special reference to adhesiveagents present on the surface of the Gram-negative bacterium,Escherichia coli.1.2. The Gram-Negative Bacterial MembraneBacterial membranes are very complex systems. The importantfeatures of a Gram-negative (e.g. E.coli) membrane are shown in Figure1. There are two lipid/protein membrane bilayers, an inner1Per:plasmic bondingproteinPeptidoglycanMinor outermembraneproteinDivalent ions(*)stabilising theouter membrane15 -20nm.Lipopolysaccharvit chains(extending outwards up to30 nm)OUTERMEMBRANEPERIPLASMIC SPACEPoemFret lipoproteinBound lipoproteinCYTOPLASMICMEMBRANE11, Phospholip idFigure 1. The envelope of gram-negative bacteria.(from reference 68)2(cytoplasmic) membrane containing the enzymes of the electron transportchain and various transport proteins, and an outer membrane containingintegral porins and other proteins with lipopolysaccharide (LPS) chainsanchored on the external side extending outwards. It is the LPS thatgives rise to the 0 serotype designation (serotyping is a method ofidentification and classification of bacteria based upon the antigenicidentity of the outer membrane). The periplasmic space between themembranes contains peptidoglycan, lipoprotein and various periplasmicproteins.Almost all periplasmic and outer membrane proteins are synthesizedin the cytosol with a secretory signal peptide sequence of 20-30 aminoacids at the N-terminus. These signal peptides have very little aminoacid sequence identity but always consist of a positively chargedN-terminal followed by a long sequence of hydrophobic residues before ashort processing site usually containing alanine residues (68)(Fig. 2).This signal sequence is required for passage of the protein through theinner membrane and is cleaved at the outer face of the membrane by theinner-membrane associated enzyme signal peptidase.The outer membrane also anchors flagella, which are locomotoryorganelles consisting of a basal body extending through the membraneconnected to a hook region which in turn anchors the filament. Thefilament contains 95 % of the mass of the flagellum and consists of apolymer of a single protein subunit, flagellin ( 60 kDa in E.coli) thatis some 20 nm in diameter and 10-20 pm long. The flagellum comprisesthe H antigen in serological analysis.The K serological determinants are generally polysaccharidemoieties, sometimes forming a glycocalyx (capsule) around the bacteria.Another class of surface antigens expressed by some E.coli strains are3Ara-binding proteincMKXTK I LVLGAVILTAGLSXGAXAI EN+ + LipoproteincMKATK I LVLGAVILGSTLLAGI CSS+ + LamBcMMITLRKI LPLAVAVAAGVMSAQAMAI VDF+ +OmpAcMKKITAIAIVALAGFATVAQAIAPK+ +OmpFcMMKRINILAVIVPALLVAGTANAIAEI+ +Figure 2. E.coli protein secretory signal sequences (adapted from 68)Amino acids are shown in single letter code (see Appendix); Xdenotes undetermined residues. All sequences contain a positivelycharged N-terminal region followed by a long region of hydrophobicresidues. The signal peptidase cleavage point is designated "c".4known as curli (82). Curli are coiled wiry fibres with a diameter ofabout 2 nm, made of a 17 kDa structural subunit, curlin. The curlinsubunit is very unusual as it appears to have no signal peptide; itsN-terminal region is highly homologous to a subunit protein of thin,aggregative fimbriae in a Salmonella enteriditis strain (83). Curlimediate binding to fibronectin, and it is suggested that they play arole in colonization.Fimbriae, often known as pili or adhesins, constitute anotherclass of bacterial surface structure. Many variations in theseorganelles have been described. Generally they consist of smallerfilaments than the flagellae, made up of repeating protein subunits;these structures mediate adhesion to host cells and other surfaces.Non-fimbrial surface adhesins have however also been described. Often, arelated side-effect of these adhesins is agglutination of differentspecies' erythrocytes (hemagglutination) due to the presence ofappropriate receptor molecules on the red cell surface (12). In fact,this effect was first noticed by Guyot in 1908 (86), long beforemechanisms of bacterial pathogenesis were elucidated. Hemagglutinationby pathogenic bacteria has simplified their study and has led to adhesinmolecules often being described as hemagglutinins.1.3. Bacterial AdhesinsVirulence factors in many Gram-negative bacteria have beendescribed. Those found in E.coli will be reviewed in Section 1.4.Pseudomonas Aeroginosa, an opportunistic agent of local infections and amajor cause of lung infection in cystic fibrosis patients, produces pilithat have been shown to bind to specific receptors on host epithelial5cells (2). These pill consist of a single protein subunit of 15 kDacalled pilin, which is chromosomally encoded. The pilin gene has beencloned and sequenced (3) and is translated as prepilin, with a leaderpeptide that is removed enzymatically leaving an N-terminalphenylalanine which is methylated. A whole class ofN-methyl-phenylalanine pill has been described (5). In the relatedbacteria, Pseudomonas cepacia, it has been shown that therespiratory mucin-binding adhesin is a 22 kDa protein associatedwith surface pill that is dissimilar to the epithelial-binding pilinmolecule (4).Helicobacter pylori, a Gram-negative human gastric pathogen,produces an adhesin with a subunit of 19,600 Da that appears to beafimbrial, mediates agglutination of human and rabbit erythrocytes andshows a limited sequence homology with an adhesin from Vibrio cholerae(6). Type 1 mannose sensitive adhesins (i.e. adhesion inhibited bymannose in solution) have been studied in Salmonella enterica (7) andShigella flexneri (8) and Moraxella lacunta, a causative agent ofconjunctivitis, is known to express pili involved in pathogenesis (9).Other Gram-negative geni where fimbrial systems have been exploredinclude Neisseria, Klebsiella, Serratia and Proteus (10). Clearly,adhesive surface structures are a near-universal feature of theGram-negative bacteria.Few Gram-positive strains possessing adhesive structures havebeen described. One genus, that of the Corynebacterium, show productionof bundles of fimbriae that have been correlated with increasedvirulence (11). These fimbriae show very limited mannose-resistanthemagglutination against trypsinized sheep erythrocytes (13).61.4. Escherichia coli AdhesinsEscherichia coli, a Gram-negative bacterium, is the most commonorganism found in the large bowel of humans and many other species. Itgenerally exists in a symbiotic relationship with the host, although itis also an opportunistic pathogen commonly associated withgastrointestinal and urinary tract infections (14).Escherichia coli produces some fourteen or more families ofadhesin distinguishable by morphological, serological, chemical andreceptor binding characteristics (15). These can be grouped into twoclasses - mannose-sensitive (type 1) and mannose-resistant adhesins.After attachment of bacteria to the host cell, disease can be causedin several ways: invasion of the host cells by the bacteria, productionof toxins or other unelucidated pathways. In E.coli infections of thegut, this leads to a classification of bacteria as enteroinvasive,enterotoxigenic (ETEC) or enteropathogenic (EPEC).1.4.1. Type 1 (Mannose -sensitive) Adhesins1.4.1.1. BiochemistryThe majority of Escherichia coli stains express type 1 fimbriaeand these structures have been extensively studied. Morphologically,type 1 pili average 7 nm in width and 2 pm in length (17) The assemblyof the pili has been studied by immunoelectron microscopy to determinethat subunits are added to the base of the growing pilus (18) . Duguidet al (12) determined that the fimbriae mediate strong mannose-sensitive7agglutination of erythrocytes from most species, including pig, dog,horse, guinea-pig and rabbit. Human erythrocytes are agglutinatedmoderately, sheep weakly, and only ox red blood cells show noagglutination. Brinton et al (16) purified type 1 pili and found that aswell as the major 17 kDa structural pilin monomer there are three minorproteins present as integral components with one of these proteins (28kDa) being the actual mannose-binding adhesin. Recent research hasshown cryptic mannose binding sites present at intervals along the pilusthat are inactive under normal conditions but that are exposed byfreeze-thaw cycles (19). Immunological variation of type 1 adhesins hasbeen described by Klemm et al (33). They found three variants with verysimilar N-terminal amino acid sequences which they named 1A, 1B and 1C.The expression of type 1 fimbriae is known to be phase-variable(expression controlled at the genetic level), with cells being eitherpiliated or non-piliated. Isolates appear to phase-vary randomly inliquid culture but many are non-piliated when grown on plates (20).Variation with temperature is also known, with 37 °C favoured over 30 °Cfor fimbriae production (22). GeneticsThe operon (that is, the region of DNA) responsible for thecorrect expression of type 1 fimbriae was mapped on the bacterialchromosome by Schmidt et al (26) and cloned by Orndorff and Falkow (21).They showed that the operon has eight open reading frames (sequences ofuninterrupted DNA prefaced by a start codon and terminated with a stopcodon, usually encoding a single protein) determining the production ofproteins involved in the biosynthesis and expression of pili (Fig. 3).8pPKL63 (type 1)nag^ u..^!LW^r71171r71;1 w 22.0 23^13.1^24.3 44 14.7^14.3^211.41Figure 3. The type 1 fimbriae operon (from 101).Numbers indicate size of protein products (x 1000 Da). Arrowindicates direction of transcription/translation.The protein product of fimA (italics indicate genes, standard casewith capitalized initial letter denotes protein) was found to be thestructural pilus subunit and has since been determined to be underthermal expression control (22). The products of fimB and fimE areinvolved in regulation of expression; fimE has been shown to stimulatethe fim invertible control element from on to off when bacteria aregrown on solid media (23). The proteins FimC and FimD are involved inlocation of the fimbria, with FimC acting as a periplasmic chaperone (aprotein that aids the folding and location of other proteins) molecule(24) and FimD being anchored in the outer membrane (28). The tipadhesin protein was determined to be the product of the fimff gene (25).The genes fimF and fimG appear to regulate the length of the pilus (29). ReceptorsThe mannose inhibition of hemagglutination that type 1adhesins exhibit shows that the fimbriae probably recognize mannose-likesugar residues on erythrocytes and other cells. Nakazawa and coworkers(30) have studied attachment of E.coli expressing the type 1 adhesin tothe small intestine of piglets, Durno et al (31) have shown that type 1fimbriae mediate attachment of bacteria to a-mannosyl residues onsurface glycoproteins of rabbit epithelial cells and Dal Nogare hasfound that type 1 pill bind to a mannose-containing glycoprotein ontracheal epithelial cells (32).1.4.2. Mannose-resistant Fimbrial AdhesinsMany mannose-resistant (MR) fimbrial agglutinins have beendescribed. Generally, the occurrence of MR adhesins of certain types iscorrelated with somatic 0 antigens, the lipopolysaccharide complexes ofthe bacterial outer membrane. Production of most MR agglutinins istemperature sensitive. Synthesis is usually repressed when grown at 18°C and heating bacteria to 65 °C most often inactivates the adhesin(60). Important MR systems are reviewed below. Colonization FactorsColonization Factor Antigen I (CFAI) was found on an E.coli straincausing diarrhea in humans (34). Related, but immunologically distinct,mannose-resistant hemagglutinins CFAII, CS (coli surface)I, CSII andCSIII were later described (35), and CSIV and CSV have recently been10identified (36,37). The CFAII and CSIII adhesins are afimbrial (seesection 1.4.3.). McConnell et al (39) have shown that polyclonal rabbitantisera prepared to these antigens showed significant cross-reactivitybetween types.Colonization factor antigen 1 has been isolated and purified toascertain a subunit molecular weight of 15 kDa (38). The organizationof the operon was found to be somewhat different to other E.coli adhesinoperons, with two coding regions separated by 40 kbp (kilo base pairs);the first region contains the structural determinants and the secondregion the regulator (40). The first part of the CFAI operon has beencloned and the four genes it contains sequenced (40).The receptors for these adhesins are largely undetermined,although Evans et al (44) have shown that the erythrocyte receptor forCFAII is more common among black donors than non-black donors. S FimbriaeThe S fimbria, a subset of a larger heterologous group ofoligosaccharide-binding (X) adhesins, are primary determinants in E.colimediated meningitis in newborn infants (41). The pili consist ofa fimbrial subunit protein polymer and a tip adhesin, the genes of whichhave been cloned and sequenced (42). Other involved genes have alsobeen identified (42). The adhesin mediates MR hemagglutination of humanerythrocytes and the 0-linked sialyloligosaccharide on glycophorin A, amajor erythrocyte membrane glycoprotein, has been identified as thereceptor on the red blood cell (43). The binding behavior ofS-fimbria to cultured human umbilical vein endothelial cells has beenstudied (44). Interestingly, adhesion of S-fimbriated E.coli to11buccal epithelial cells is inhibited by human milk fat globule membranecomponents, leading to the suggestion that these components inhibitbacterial growth in the entire intestine, providing protection frominfection (45). P FimbriaeFimbriae mediating adhesion to the uroepithelium in bacterialinfection of the urinary tract are known as P or pap (pili associatedwith pylonethritis) fimbriae. The genetic structure of this system isvery well known, as are its receptors: digalactoside-containingglycolipids found on the surface of erythrocytes and uroepithelialcells. The chromosomal operon mediating expression of these fimbriaewas cloned by Hull et al (46) and is shown in Figure 4.^Pilus^ Pilus TipMajor^ Assembly FibrillumRegulation Pilus Machinery SubunitsSubunitE-EB:1--1 A L-{ H^C^7 D J  --•KtElF,-1 G 1,-Pilus Shaft Pilus^Outer Membrane^Periplasmic^Tip^Maior^Links^Gal a (1-4)Protein Anchor Usher^Chaperone Length^Tip^Adhesin Gal-binding^Regulation Compo-^to^Adhesinnent^FibrillumFigure 4. The pap operon (from 49).12The organization of this operon is very similar to that of type 1fimbriae. The product of papA is a 16,500 Da protein that forms thepilus subunit; about 1,000 of these subunits are arranged helically toform hair-like appendages on the outside of the cell (47). It has beenshown that mutants lacking PapA do not produce fimbriae, but do stillbind the receptor (53). The digalactoside-binding adhesin is PapG (48),which is located on the tip of the fimbriae. The tip itself is mainlycomposed of repeating PapE protein subunits, with small amounts of PapFand K also present (49). The PapF protein is required for the correctpresentation of the adhesin at the distal end of the tip structure; PapKjoins the tip to the major part of the fimbria and regulates its length(55). All of these proteins are translated with secretory signalpeptides. Figure 5 is a diagram detailing the biosynthesis of the papfimbriae.The molecular structures of some of these proteins are known insome detail. The three-dimensional structure of the periplasmicchaperone PapD has been solved to a resolution of 2.5 A and a PapAbinding area similar to that of an immunoglobulin has been proposed(50). The 88 kDa protein PapC has been given the name of "molecularusher", since it regulates an ordered assembly of the correct componentson the outer membrane. This protein seems to be conserved inpilus-producing bacteria and interacts with the chaperone-subunitcomplexes on the periplasmic side of the outer membrane (51). Normarket al have studied the structure of PapG, the digalactoside-bindingprotein and have found that it consists of two domains, one for receptorbinding (the amino-terminal end) and one that is involved inincorporation of the protein into the pilus (52). The amino-acid13Adhesin: GTip FibrAlum Major component: EMinor components: F, KPilus ShaftMajor Component: ATip FibrillumAnchor: H1. lbrgeting2. Chaperone Uncapping3 . PolymerizationPreassemblyComplexesOuterMembranePeriplasmCytoplasmicMembraneFigure 5. Putative assembly mechanism of pap pill (from 49).14primary structure has been deduced from the nucleotide sequence andshows some homology with the CFAI protein group (54).The receptor for the pap pili is the a-D-galactopyranosyl-(1-4)-g-D-galactopyranose moiety present on the glycolipids on the surface ofcells lining the urinary tract (56). The production of pap piliis phase variable; there is evidence to suggest that, as well as the papoperon regulatory proteins, DNA methylation upstream of the regulatoryregion of the operon occurs, controlled by the mbf (methylation blockingfactor) locus (57) and involving Lrp (leucine-responsive regulatoryprotein) (58). The K88 AntigenThe K88 adhesive antigen was discovered by Orskov et al (59) whoalso distinguished two variants K88ab and K88ac. The K designation wasassigned although it was found that the antigen was in fact a proteinrather than a polysaccharide (88). A third variant K88ad has since beendescribed (61). The major fimbrial subunit has a molecular weight ofbetween 23.5 kDa and 27.5 kDa depending on the variant. It has a 21amino acid N-terminal signal sequence and contains no cysteine residues,which implies hydrophobic or electrostatic interactions hold thesubunits together in the pili (60). The plasmid-encoded geneticdeterminant of these fimbriae was cloned (62) and is found to be similarto other fimbriae determinants, with components for structure, adhesion,construction and expression grouped in the operon. However, despitethis similarity it appears that, in contrast to other systems, the majorcomponent of the fimbriae also carries the adhesive property (63).Jacobs et al modified the K88 fibrillar subunit by site-directed15mutagenesis and found that the replacement of phenylalanine 150 withserine resulted in production of fimbriae that were essentiallyindistinguishable from wild-type, but which had lost the adhesivecapacity of the pili (64). They ascertained that the region containingthis phenylalanine residue shared some sequence identity with aconserved region in the gonococcal pilin and in the B-fragment ofdiphtheria toxin, supporting a common receptor binding function. Thereceptor for K88 is unknown, though there is some evidence indicatinggalactosyl residues are important (60). The K99 AntigenThe K99 fimbriae have a helical structure with a diameter of 4.8nm, made of structural protein subunits of 18,500 Da (65). Productionof fimbriae is dependent upon utilizing minimal growth media The operoncodes for seven proteins involved in biosynthesis and deletion mutantsconstructed by de Graaf and coworkers (65) indicated that allpolypeptides were required for MR agglutination of horse erythrocytes.The adhesive function was determined to reside on the structural subunit(105) and was found to be specific for carbohydrates on erythrocytes(66), with sialic acid residues (sialic acid is, strictly, N-acetylneuraminic acid (Fig. 6) although the term "sialic acids" generallycomprises all N- and 0-acyl derivatives of neuraminic acid isolated fromnatural materials) being implicated. The equine erythrocyte receptorwas determined by Smit et al (72) to be the glycolipidNeu5Gc-a(2,3)-Gal-g(1,4)-Glc-g(1,1) -ceramide (where Neu = neuraminicacid, Gal = Galactose and Glc = glucose). This specificity was confirmedby Lindahl and Carlstedt (67) by studying binding of purified K9916fimbriae to sections of pig small intestine. The fimbriae were found tobind to high molecular weight mucin glycopeptides and the binding wasabolished when these glycopeptides were enzymatically desialated. TheN-terminal amino acid sequence of the structural subunit shows someidentity with the K88 fimbrial protein and limited identity with the F41adhesin described below (69) (Figure 7).HO^,HNr''HOH2C /C„HH3C - C -N^OHHOHCo - OH0Figure 6. Sialic acid (N-Acetyl-g-D-neuraminic acid)(from 102) The F41 adhesinThis MR adhesin was discovered on a K99 negative mutant of the K99reference strain 841M (serotype 0101:K(A):NM) (69). As with K99, theproduction of this antigen was found to be dependent on the compositionof the growth media with biosynthesis being repressed by rich media andoptimum production occurring when a minimal salt medium was used. Theagglutinin molecule was very easy to detach from the cell by simplyvortexing a bacterial suspension for five minutes; denaturingpolyacrylamide gel electrophoresis showed that the fimbriae were17composed of subunits of molecular weight 29,500 Da. The production ofthis antigen appeared to be restricted to the serotypes 09 and 0101.The amino acid composition Was determined, as was the N-terminal aminoacid sequence, which was compared to those of other adhesins (Fig. 7).Limited sequence identity was found with K99, but not with K88, CFAI ortype 1.The F41 chromosomal genetic determinant was cloned from adifferent porcine enterotoxigenic E.coli strain (VAC1676,0101:K30:F41:H-) by Moseley et al (70). Their results showed that therewas significant homology between the genes encoding F41 and thoseencoding K88, as DNA probes specific for the K88 operon hybridized tothe cloned F41 determinant. In later work, Moseley and Andersondescribed the genetic organization of the operon, detailing severalproteins that together mediated full expression of the F41 fimbriae.The accessory proteins showed high sequence identity with the equivalentproteins in the K88 operon but there was little similarity in thefimbrial subunits themselves apart from the signal peptide sequence(71). Moseley and Anderson conjecture that the homology between adhesinsynthesis and assembly systems indicates the ability of the bacterium torapidly evolve new fimbria differing in antigenic structure. The genecoding for the fimbrial subunit was sequenced and its deduced amino acidstructure was found to differ significantly from the amino acidcomposition determined by analysis of the purified F41 structural unitfrom E.coli B41M reported by de Graaf and Roorda (69), although theN-terminal sequences are almost identical.181 20F41^A DW.1■11/T EGQPG D I L II G GI E X .11■11■P S 11■11.VK99 T G T I N F N G L I ETATSXIE P A V111■K88^WM T G D F N G SVDIGGIS A D G Y G.11■1.CFAI^VEKNITVTA SVDPVIDLLQ A D GTypel^AATTVNGGTVHFKGEVVNAFigure 7. Comparison of adhesin N-terminal sequences(adapted from 69). Regions of identity are boxed. Forclarity, not all identity relationships are shown.19The receptor for the F41 adhesin on human erythrocytes isglycophorin A (73). The agglutinating activity of the purified adhesinis greatest when the erythrocytes used are of blood group MM and leastwhen the blood group is NN (76). The M/N antigen lies in the first andfifth amino acid residues of the N-terminal region of glycophorin A asshown in Figure 8 (74). Glycophorin MM has an N-terminal serine and aglycine at position five; glycophorin NN has leucine and glutamic acidin these respective locations. Neuraminidase treatment of theerythrocytes (which removes the terminal sialic acid residue) alsodecreases agglutination showing that both the amino acid sequence andthe carbohydrate at the N-terminus are important in the recognition ofthe receptor by the adhesin.A thermodynamic study of the binding of radiolabelled purified F41adhesin to radiolabelled glycophorin A' utilizing an aqueous two-phasepolymer system has indicated a stoichiometry of 1 adhesin monomer to 1glycophorin molecule. This indicates that the F41 system does notrequire a specialized tip adhesin, but that binding sites are anintegral part of the structural subunit (76). This suggestion wasrecently strengthened by epitope mapping of the adhesin with monoclonalantibodies, where 23 antibodies recognizing five highly conservedepitopes were produced (77). The antibody binding was studied byimmunoelectron microscopy and showed that the epitopes were equallydistributed along the fimbrial structure. The binding of F41 toglycoproteins in the bovine and porcine colostrum was studied by MatsLindahl (75). He found that the adhesin was carbohydrate specific,binding to periodate-sensitive oligosaccharides present on theglycoproteins.20(Lau(Ser1occococooecumeaeocoe30^4000000400:000000000050^600000000000000000000070^8000000000000 se0000000 Lis90^100000000000000110^12000000000000( a )^ 1300 0-glycansgal 131+3 gal NAc -. Ser^gal pl-■3gal NAc -0.Ser^gal /31-.3gal NAcar1-.Ser(Thr) 3^(Thr) 3^6^(Thr)t ^neuNAca2 neuNAca2 neuNAca2O N-g/ycans(neuNAca2•6) gal A..I 4gIc NAc^man at e3neuNAca 2..6 gal /31.4 gIcNAc f31-•2 man al^4gIcNAc(316 4^ 6man (314.4 gk NAc p1-.4gIcNAci31-.Asntuc al0 or 1( b)Figure 8. Glycophorin A (from 103)(a). Primary structure showing glycosylated amino acid residues(b). Generalized structures of the carbohydrate units211.4.2.7. Other E.coli Fimbrial AdhesinsOther fimbrial adhesins include CS31A, present on E.coli isolatesfrom diarrheic or septicemic calves. The CS31A determinant is encodedon a 105 MDa plasmid. The fimbriae produced are thin (2 nm) and consistof a structural subunit of 29,500 Da (78). Amino acid sequencing of theN-terminus of this protein showed that it shares a high degree ofsequence identity with K88 (Fig. 9) however relocation of the plasmidinto a non-hemagglutinating strain does not confer hemagglutination eventhough CS31A fimbriae are produced.CS31A^G T TGDFNGS F D M N G T ITAD A T KK88^WM TGDFNGS V D I G G S ITAD G Y G.11■• .11■11.Figure 9. N-terminal homology of CS31A and K88(Adapted from 78)The CS31A determinant was cloned from genomic DNA using F41determinant DNA probes (79), indicating that the entire determinant wasrelated to that of F41. The clone harbouring the recombinant CS31Aantigen was able to adhere to cultured epithelial cells.The 987P fimbriae are composed of subunits of 20,000 Da and aremorphologically identical to type 1 pill, although they show a MRhemagglutination pattern (60). The plasmid encoded 987P operon has beencloned and shown to contain open reading frames for eight proteins (80).22Analysis of these genes has shown that, although all gene products areinvolved in pili biosynthesis, the lack of expression of a singleprotein never resulted in the separation of fimbriation from adhesion,strongly suggesting that the fimbrial subunit is also the receptorbinding adhesin (81).The bundle-forming pilus present on some enteropathogenicE.coli strains is so-called because filaments of the repeating 19.5 Ksubunit aggregate into rope-like bundles that bind together individualbacterial cells (85). This effect occurs when the bacteria are grown onblood agar. The structural gene was cloned (167) and it was found thatwhen this gene was used as a DNA hybridization probe, allenteropathogenic E.coli strains showed homologous regions, as did mostof the Salmonella strains tested. The authors suggest that this pilusis involved in initial colonization early in the infective process.1.4.3. Afimbrial Adhesins of Escherichia coliMuch less is known about these structures, several of which havebeen identified in the last ten years. Afimbrial adhesins were firstdescribed by Duguid et al in 1979 (84). In a survey of 387 strains ofE.coli, they found that approximately ten percent of strains that causehemagglutination do not appear to produce fimbriae when viewed under theelectron microscope. These bacteria are thought to express non-fimbrial(or afimbrial) adhesins. Since then several afimbrial systems have beeninvestigated, some of which are detailed below. The 2 Antigens.Three uropathogenic MR-hemagglutinating Escherichia coli strainswere studied by Orskov et al (87). They found that the strains possesseda common surface antigen (Z1) when grown at 37 °C, but not at 18 °C.One of the strains produced an additional antigen 22. Unlike other MRstrains, heating to 100 °C did not destroy the agglutination. A mutantthat produced much less of the Z1 antigen was non-adherent, so the 21,and possibly the 22 moieties were held responsible for thehemagglutination. The Z1 antigen was identified as a 14,400 Da.protein. Immunoelectron microscopy of the embedded, sectioned bacteriashowed that the protein formed a capsule around the bacterium. Inelectron microscopy studies after staining the bacteria with uranylacetate, this capsule appeared as a mesh of fine filaments. The authorstied their results in with results of a group who had isolated anonfimbrial MR hemagglutinin from two enteropathogenic E.coli strains,of monomer molecular weight of about 14,500 Da (89). This group hadnoticed a capsule on the cells of both strains despite the fact that nopolysaccharide K antigens were included in the serotype designations; itwas proposed that this may indicate a similar adhesive protein capsule. The NFA Group of Afimbrial AdhesinsNon-fimbrial adhesin (NFA)-1 and NFA-2 were first reported byGoldhar and coworkers (92) as consisting of soluble proteins ofmolecular weights 21,000 and 19,000 Da respectively. Both mediate amannose resistant hemagglutination of human erythrocytes and were shown24to bind to cultured human kidney cells. Biosynthesis of both is haltedwhen the bacteria are grown at 18 °C. Adhesive bacteria of the E.colistrain expressing NFA-1 showed an extracellular capsule-like layer,whereas non-adhesive cells did not. The hemagglutinating strainexpressing NFA-2 did not have this layer. The proteins tended to formaggregates of high molecular weight, and Goldhar et al. state that theproteins recognize some glycoprotein on the erythrocyte. The adhesinsNFA-3 and NFA-4 have been described (95,96). The NFA-3 protein has amolecular weight of 17,500 Da and is specific for glycophorin NN. Nonfimbrial adhesin 4 is 28,000 Da and binds glycophorinm, but shows nosensitivity to the presence of sialic acid residues. AIDA-I (Adhesin Involved in Diffuse Adherence I)This adhesin was cloned from a large plasmid present in anenteropathogenic E.coli strain (90). A 100 kDa protein was identifiedas the product of the smallest subclone that still mediated adherence toHeLa cells. This molecular weight is similar to that of anotherafimbrial adhesin, EAF (enteropathogenic adherence factor), also codedon a large plasmid (91). These adherence factors do not showhemagglutination and seem to belong to a separate family of adhesinsbeing discovered on enteropathogenic E.coli strains. AFA-1 (Afimbrial Adhesin-1)The adhesin AFA-1 exists on the bacterial surface and as anaggregate in culture supernatants. It has a monomer molecular weight of16,000 Da and has been found to agglutinate only human and gorilla25erythrocytes. Glycophorin A is not bound by AFA-1 (93). The operonencoding AFA-1 biosynthesis was cloned from chromosomal DNA (94) andfound to contain genes coding for five polypeptides. The nucleotidesequence of the structural subunit gene, afaE, was determined and wasfound to have no similarity to known fimbrial or afimbrial adhesins.Nowicki et al (97) have studied the receptor binding specificities ofAFA-1 and the related afimbrial adhesins AFA-3, the Dr hemagglutinin andF1845. They have identified the common erythrocyte receptor as the Drantigen, a cell membrane protein called DAF (decay accelerating factor).The protein takes part in the regulation of the complement cascade. Theresearchers, using enzyme treated erythrocytes and monoclonalantibodies, found that different regions of the protein were recognizedby the different adhesins, showing that the agglutinins were distinct,but related.1.5. Viscometric Studies on Bacterial AdhesinsAlthough bacterial adhesins are the subject of much research,little attention has been paid to the effect of the in vivophysiological environment in which these adhesins act. Bacteria areable to colonize mucosal surfaces despite fluid and mechanical shearcaused by blood, mucus, urine and gastrointestinal content flow. Theviscometer, a device that measures the viscosity of a liquid undershear, provides a convenient in vitro method for studyingbacteria-erythrocyte interactions under these conditions (98).261.5.1. Viscometric Analysis of Erythrocyte SuspensionsEinstein showed that the viscosity of a very dilute suspension ofnon-interacting particles under shear was a linear function of thevolume fraction occupied by these particles (149). This equation onlyholds true for suspensions where the volume fraction of the particles isless than 2 %. For higher particle concentrations, the relationship isbased upon higher order terms in the volume fraction (150), and theoryexists to model the viscosity of non-interacting suspensions withparticle volume fractions of the order of 10 % (151). In theviscometric experiments described from the literature below, and forthose in this work, the viscosity of a concentrated (a50 %) erythrocytesuspension is measured. These suspensions behave as a complexnon-Newtonian fluid, with the cells being non-spherical, interactingelectrostatically and deforming under shear. There is as yet no theorythat allows modeling of such a complex system. It is knownexperimentally, however, that all other factors being equal, red cellaggregation can greatly enhance the viscosity of the erythrocytesuspension (152).The viscometer consists of a rotating cup and a central bob (Fig.10). There is a small gap between the cup and bob into which thesolution under test is placed. The cup is then rotated at a constantangular velocity, producing a roughly constant shear rate (velocitygradient) throughout the sample, assuming Newtonian flow. The centralbob is held still by the application of an electrical torque that isproportional to the shear stress exerted upon it by the liquid, andhence proportional to the viscosity of the liquid (viscosity = shearstress/shear rate). A signal proportional to the shear stress is27provided continuously by the instrument. Normally a liquid will show aviscosity that will not vary over time. Samples of material thataggregates will show an increase in apparent viscosity relative to thenon-aggregated condition. If the sample is a cell suspension aggregatedby multivalent ligands, the viscosity may increase as shear continues.The hemagglutination event that occurs between erythrocytes and bacteriais an example of this. The apparent viscosity of thebacteria-erythrocyte mixture at constant shear rate is increased overthat of a non-aggregating suspension because bonds between the cells arebeing formed and energy is required to break them if the suspension isto flow in the velocity gradient between the cup and the bob. Thisenergy dissipation is the source of the extra shear stress and thereforethe increase in viscosity.The technique of viscometry is useful for the study of bacteria-erythrocyte interactions because a quantitative index of the degree ofagglutination (and hence the strength of the agglutination event) canbe expressed by the value R, which is the ratio of the shear stress inthe presence of aggregation to that in its absence.1.5.2. Viscometric Analysis of Aeromonas salmonicida 438.The hemagglutination event mediated by Aeromonas salmonicida 438was studied by measuring the viscosity of a mixture of erythrocytes andbacteria under shear, and it was found that adhesion was enhanced byshear forces, as indicated by a viscosity increase (99). The adhesionreaction was found to consist of two phases, a fast (10 s) initialviscosity rise followed by a slower increase which reached a plateau atapproximately 30 min. The slower, later phase of the reaction was284Torque-measuring bobGuard ring-.1- Rotating cup maintainedat constant temperatureCell suspension in 1 mm gapStationary wall of bobCross-section of concentric cylinder viscometerApproximate flow profile in gap^Rotating wall of cupTransverse section showing flow profileFigure 10. Diagrammatic view of a couette type viscometer29determined to be induced by the shear, as bacteria-erythrocyte mixturesthat were preincubated without shear for certain lengths of time alwaysshowed lower viscosity when shear was applied than mixtures that hadbeen subjected to shear for the entire time.1.5.3. Viscometric Analysis of E.coli AdhesinsWhen erythrocytes were sheared with purified F41 adhesin the sameeffect was noticed as for the Aeromonas salmonicida bacteria (Figure11). A model for this shear enhancement involving a shear inducedrearrangement of the membrane proteins involved was put forward (100).This model states that after initial contact and adhesion takes place,there is a shear induced rearrangement of the receptor proteins in thefluid membrane increasing the probability of more adhesive interactionsoccurring. As these interactions build, the adhesion becomesprogressively stronger.Viscometric analysis was also used to determine the M/N bloodgroup specificity of the F41 adhesin (76). Although the agglutinationdifference between MM and NN erythrocytes was unnoticed in a microtitreassay at room temperature, at 37 °C in the viscometer the F41 mediatedagglutination of NN erythrocytes was noticeably less than that of thecells carrying the MM antigen (Figure 12).30I1010^Time (min)^20Figure 11. Adhesion event between E.coll F41 adhesin anderythrocytes as determined by viscometry (from 100).Adhesin added to erythrocyte suspension at the arrow.31•■••■■••I0^ 20Time (min.)Figure 12. Differential agglutination of MM and NN erythrocytes byF41 adhesin as determined by viscometric assay (from 76).(shear rate = 49 s -1 )321.6. Partition of Bacterial Cells in Aqueous Two-phase Systems1.6.1. Principles of Aqueous Two-phase Polymer PartitionAqueous solutions of polymers are often incompatible above certainconcentrations such that if two are mixed they will rapidly separateinto two phases, each of which is enriched in one of the polymers, withthe denser solution becoming the lower layer. An example of thisphenomenon is the phase system composed of dextran (Dx, Fig. 13) andpoly(ethylene glycol) (PEG, Fig. 14), which consists of a dextran-richlower phase, and a PEG-rich upper. Another system used in this thesisis dextran\Ficoll (Ficoll is a synthetic branched polymer made by thecopolymerization of sucrose and epichlorohydrin, (104)). A generalphase diagram for the PEG/Dx system is shown in Appendix 2. These phasesystems can be made up in buffers to suit a wide range ofphysiologically sensitive applications. When cells such as erythrocytesare added to these systems, the cells partition unequally among theupper and lower phases and the interface, which acts as a thirdcompartment (107). Some ions, such as phosphate, show preferentialpartition in one phase over the other, leading to an electrostaticpotential difference between phases. Such systems are termedcharge-sensitive and can be useful for separations based on theelectrostatic properties of cell surfaces (108).The partition of cells in phase systems can be influenced by awide range of properties of the cell surfaces. In charge-sensitivesystems, the charge on the cells is an important factor in theirpartition (108), and other determinants such as cell surface-polymer33interactions (including polymer adsorption to cells) also appear to besignificant (108). From the hundreds of papers that have been publishedon cell partition it is clear that the cell distribution in a givensystem is very sensitive to both the properties of the system and of thecell surface. There is as yet no theory that allows the cell surfacestructure to be predicted from the partition behavior, however (109).1.6.2. Bacterial PartitionThe first partitioning studies with bacteria were performed byAlbertsson (who also developed the partitioning technique). He used themultiple partition method of countercurrent distribution in PEG\dextransystems to separate what was thought to be a single strain of E.coliinto two populations based upon differences in their surfacecharacteristics (110). In E.coli, the 0 and K antigenic determinantshave been investigated as sources of partition differences. Thecapsular polysaccharide K antigens on the surface of Gram-negativebacteria contain a large amount of acidic carbohydrate material andeffectively render the bacteria negatively charged, which to a greatextent dictates partition in charge-sensitive systems. In these systemsthe bacteria tend to show affinity for the phase which is relativelypositive in potential (the PEG-rich phase in dextran/PEG charge-sensitive systems). Lipopolysaccharide (LPS) structure variesconsiderably between strains of bacteria and has been shown to influencepartition in non charge-sensitive systems. In one set of experiments,isolated LPS from two different Salmonella typhimurium strains waspartitioned in a dextran\PEG system. The LPS partitioned verydifferently, with one type strongly favouring the PEG rich upper phase34Figure 13. DextranH- (-0-012-CH2- ] n—OHFigure 14 Poly(Ethylene Glycol)35and the other strongly favouring the dextran rich lower phase. Thisdifference was also seen when the whole bacteria were partitioned,indicating that the LPS was the major determinant of the partitiondifference (111)The role of fimbriae in determining partition has also beenstudied. Using a PEG\dextran system with differing amounts of thehydrophobic PEG-palmitate derivative in the upper phase, it was foundthat the partition of piliated strains of Neisseria gonorrhoeae was lessaffected by the PEG-palmitate (and hence said to be less hydrophobic)than non-piliated strains (112). Kihlstrom and Magnussen (113) foundthat hemagglutinating properties (probably mediated by fimbriae) ofYersiniae species correlated with increased tendency to hydrophobicinteraction. As mentioned earlier, the valency of the F41 adhesin wasdetermined using a PEG/dextran phase system to generate an equilibriumbinding isotherm based on the differing partitions of the F41 adhesinand its receptor, glycophorin.362. Cloning of Hemagglutinin Gene2.1. Background and AimsThe Escherichia coli strain 09:1110:K99 is known to express at leasttwo mannose-resistant agglutinins, K99 and F41 (114). As describedearlier, the K99 antigen is an 18,500 Da protein that is plasmid encodedand agglutinates horse erythrocytes, but not human (65), whereas F41 hasa molecular weight of 29,500 Da, is chromosomally encoded andagglutinates human erythrocytes as well as several other species (69).The original aim of this study was to clone and investigate theF41 adhesin, which has since been cloned by Moseley (70). To this endthe cloning procedure was designed to take advantage of the fact thatonly the F41 adhesin agglutinates human red blood cells. A simplemicrotitre assay was used to screen quickly insert-containing colonies.This initial assay was then backed up with colony blots usingantibodies raised against the purified F41 and K99 adhesins and by anoligonucleotide probe designed from the N-terminal amino acid sequenceof the F41 protein. On the basis of the results of these assays, aclone was chosen for further study which showed the interesting propertyof exhibiting agglutinating activity with absence of both the K99 or F41antigens, and that did not hybridize to the oligonucleotide probe.372.2. Materials and Methods2.2.1. Bacterial Strains and Growth ConditionsE.coli 09:H10:K99 (K99 + ,F41 + ) was originally isolated by R.E.Isaacson (115) and was kindly supplied by T. WadstrOm. Hereafter, thisstrain is referred to as " F41 ". E.coli JM101 (116) was used as thehost strain for cloning and maintenance of recombinant plasmids. Bothstrains were grown in Luria-Bertani (LB) broth (117), which wassupplemented with Difco agar (1.5 %) if solid media was required. Cellswere grown aerobically at 37 °C, and the media was supplemented withampicillin (100 µg/ml) for maintenance of plasmid-containing clones.Tetracycline (15 µg/ml) was added for the original screening of thelibrary for clones that contained a foreign insert in the pBR322plasmid.2.2.2. Enzymes and ReagentsAll restriction enzymes were purchased from either BoehringerMannheim, Laval, Quebec, New England Biolabs, Beverly, MA, USA, orBethesda Research Laboratories (BRL), Gaithersburg, MD, USA. Modifyingenzymes came from the same sources. Ethidium bromide(3,8-diamino-5-ethyl-6-phenylphenanthridium bromide) was obtained fromSigma, St. Louis, MO, USA. For probe labelling, 7-32P-ATP was purchasedfrom Amersham, Arlington Heights, IL, USA. X-Phosphate(5-bromo-4-chloro-3-indolyl-phosphate) and nitroblue tetrazolium, thecolour development reagents used in alkaline phosphatase conjugated38antibody assays, were purchased from Bio-Rad (Richmond, CA, USA). Allother chemicals used were of analytical grade and obtained from chemicalsuppliers.2.2.3. BuffersRestriction enzyme buffers used were those supplied by themanufacturers with the enzymes. Buffers for DNA suspension, ligations,and modifying reactions were as described by Maniatis et a/ (117). Forgel electrophoresis of DNA, tris-acetate-EDTA (TAE, 0.04 M tris-acetate,0.002 M EDTA, pH 8.0) was used (117). Phosphate-buffered saline (PBS,130 mM NaCl, 16.7 mM Na 2 HPO and 3.0 mM NaH 2 PO 4, pH 7.4) was used to4wash and suspend erythrocytes in the hemagglutination assay, and forwashing filters in the antibody assay. Where desired this wassupplemented with 10 mM mannose.2.2.4. PlasmidThe plasmid pBR322 (118) was used for the cloning procedure. Thisplasmid utilizes the ColE1 origin of replication and contains genes thatcode for proteins that make the host cell resistant to ampicillin andtetracycline.2.2.5. AntibodiesThe monoclonal antibodies against the F41 adhesin were made byNancy Hamilton in the Brooks laboratory and the K99 antisera was kindlydonated by Dr. Steve Acres from the Veterinary Infectious Diseases39Organization (VIDO), Saskatoon, SK.2.2.6. Gel Electrophoresis of DNAAgarose gel electrophoresis for visualization and preparation ofrestriction digested DNA was performed in submerged horizontal gelsystems (117). Gel composition was routinely 0.8 % agarose / 0.5 µg/mlethidium bromide in TAE buffer. Ethidium bromide complexes with DNA byintercalating between stacked base-pairs and allows visualization of theDNA as it fluoresces orange under ultraviolet illumination when bound.The applied voltage and electrophoresis time varied with experiment.The gels were photographed with a Polaroid 57 4x5 Land Film camera withorange filter and were illuminated for photographs with ashort-wavelength UV transilluminator.2.2.7. Gel Electrophoresis of ProteinsProtein samples were fractionated on denaturing polyacrylamidegels by a modification of the procedure of Davis (119) and Ornstein(120). The gels were cast and run on the Bio-Rad Mini Protean II slabgel system. Acrylamide:bis solution (30:08) was made up as follows:Acrylamide (60 g) and N,N'-Methylene-bis-acrylamide (1.6 g) weredissolved in distilled water to a final volume of 200 ml. The solutionwas filtered through Whatman #1 filter paper and stored at 4 °C in thedark. For the resolving gel, the acrylamide:bis stock solution (12.5ml) was mixed with 1.875 M tris-base pH 8.8 (6 ml), 0.2M EDTA (300 gl),water (10.9 ml) and SDS (300 gl) and degassed under vacuum. To thissolution was added 300 gl of freshly-made ammonium persulphate40((NH 4 ) 2 S 2 0 8, 10 %) and 15 pl TEMED, and the solution was carefullypipetted between the glass plates of the casting apparatus (1 mmspacers). Water was layered onto the solution surface and thepolymerization was complete in 30 min. The water was removed, the geltop rinsed, a sample comb inserted and a stacking gel composed ofacrylamide:bis (2.5 ml), tris.HC1 pH 6.8 (1.88 ml), 0.2 M EDTA (150 pl),10 % SDS (150 pl), 10.3 ml H20, 10 % ammonium persulphate (150 pl) and7.5 Al TEMED was pipetted on top of the resolving gel. The quantitiesof solutions described were sufficient to pour two gels each. Whenpolymerization was complete (30 min), the comb was removed and the gelloaded into the running chamber. The chamber was filled with runningbuffer, a 4:1 dilution of the stock solution, 14.4 g/1 glycine and 1 g/1SDS in 0.05 M tris-base.Protein samples were prepared for electrophoresis by mixing withan equal volume of sample buffer ( 4.0 ml water, 1.0 ml 0.5 M tris-HC1,pH 6.8, 0.8 ml glycerol, 1.6 ml 10 % SDS, 0.4 ml 2-mercaptoethanol and0.2 ml 0.05 % (w/v) bromphenol blue.) and boiling for 5 min. Thebromphenol blue was omitted when preparing samples to be sent forN-terminal sequencing. Samples (5-20 Al) were applied to the gel and acurrent of 150 mA was applied until the tracking dye reached the bottomof the gel. Gels were then removed from the apparatus and were eitherstained or western blotted as described below. Staining the gels wasperformed for two hours or overnight with Coomassie brilliant blue R-250(2 g/1) in 35 % EtOH, 10 % acetic acid. The gels were destained withtwo changes of 35 % EtOH, 10 % acetic acid over several hours. The gelswere generally then allowed to destain further overnight in water beforebeing photographed with a Polaroid 57 4x5 Land Film camera and dried forthree hours on a Bio-Rad 543 gel drier.412.2.8. Western BlottingProtein samples size-fractionated on SDS-PAGE gels were blottedonto either nitrocellulose (NC, Bio-Rad) or PVDF (polyvinylidenefluoride, 148) (ImmobilonTm, Millipore) using a Bio-Rad western blottingapparatus, by an adaption of the method of Towbin et al (121). Fornitrocellulose blotting, the transfer buffer used was tris-base (3.025g/1, pH 8.3) and glycine (14.41 g/1) in 20 % methanol. For Immobilonblotting the buffer used was CAPS (2.213 g/1 3'-[cyclohexylaminol-1-propanesulphonic acid in 10 % methanol). The membrane was wetted (NCin blotting buffer, PVDF in Me0H) and sandwiched against the gel in theblotting clamp, using filter paper pads to keep the two in tightcontact. The apparatus was set up so that the membrane was nearest tothe positive electrode, and the gel was electroblotted for two hours at1 A, or overnight at 0.5 A. Usually, prestained protein standards wererun on the SDS-PAGE gel so that the success of the blotting could beascertained. After blotting, the protein samples were probed withantibodies or utilized in other manners described later.2.2.9.^DNA ExtractionsFor total cellular DNA, E.coli 09:H10:K99 cells were grownovernight in 500 ml LB, harvested by centrifugation at 4 °C, washed in20 mM tris pH 8 / 100 mM NaC1, and resuspended in 40 ml 20 mM tris / 100mM NaC1 / 5 mM EDTA. Lysozyme was added to 2 mg/ml, and the cells wereincubated for 30 min at 37 °C with gentle mixing. Sodium dodecylsulphate (SDS) was added to a final concentration of 0.5 % and the cells42were incubated for 30 min at 65 °C, after which time 10 mg Proteinase K(BRL) was added. The lysed cells were incubated at 30 °C for 3 h andwere extracted three times with an equal volume of phenol (saturatedwith THE buffer - 10 mM tris.Cl pH 8.0, 100 mM NaCl, 1 mM EDTA pH 8.0).To remove RNA, 10 mg/ml RNase was added to a final concentration of 50pg/ml, incubated for 30 min at room temperature, then extracted twicewith phenol:chloroform (1:1). The purified DNA was then precipitatedfrom the aqueous phase with 95 % ethanol (2 volumes), washed with 70 %ethanol by resuspension followed by centrifugation, air-dried andresuspended in sterile distilled H20. The concentration and purity ofthe DNA was determined by measuring the absorbance of a dilute solutionat 260 and 280 nm. The 260 nm reading gives the concentration of DNA(1 a.u. g 50 pg ml -1 ), and the ratio A260 /A280 gives a measure of itspurity. Pure preparations of DNA have an A260/A280 ratio of 2.0,contamination with protein or phenol lowers this ratio (117).Large scale plasmid preparations were made by ethidiumbromide-CsC1 density gradient ultracentrifugation. Overnight cultures(1000 ml) were harvested by centrifugation at 6000 g for 15 min andresuspended in 20 ml GET (50 mM glucose, 25 mM tris.Cl pH 8.0 and 10 mMEDTA). The cells were lysed gently with NaOH (0.2 N) and SDS (1 %) toallow the cells to release plasmid DNA yet retain most high molecularweight chromosomal DNA associated with the cell debris. Potassiumacetate (7.5 ml, solution is made up with potassium acetate and aceticacid so that the final concentration of potassium is 3 M, acetate is 5M.) was added to precipitate the cell debris, which was removed bycentrifugation.43The covalently closed circular plasmid DNA was isolated from thecleared lysate supernatant by centrifugation to equilibrium in a CsC1density gradient (2 g/ml) containing ethidium bromide ( 25 pg/m1). Theethidium bromide serves two purposes - it allows visualization of theDNA bands in the gradient and it separates the covalently closed plasmidDNA from the remaining chromosomal DNA and linearized plasmid DNA bands.It does the latter by decreasing the density of the DNA as itintercalates. Intercalation of ethidium bromide forces a partialunwinding of the DNA helix and since closed circular plasmid DNA hasmore topological constraints than open-ended DNA it binds less andretains a higher density. After 40 h at 42,000 rpm in a Beckmanultracentrifuge (Titanium 70.1 rotor), two ethidium bromide stainedbands were seen, the lower corresponding to the covalently closedplasmid DNA. The plasmid band was extracted by puncture of thepolyallomer centrifuge tube with a 25 gauge needle and syringe andpurified by precipitation with two volumes 95 % ethanol, centrifugationand washing of the pellet in 70 % ethanol. The DNA was redissolved intris.Cl pH 8.0-EDTA (TE) buffer and its concentration and purityassessed by reading the absorbance of a dilute sample at 260 and 280 nm.Plasmid minipreps were performed by gentle alkaline lysis of thecells as above, followed by ammonium acetate protein precipitation,phenol / chloroform extraction and ethanol precipitation of the DNA(117). Precipitated DNA was dissolved in TE containing RNase (50µg/ml)and purity and concentration were estimated by agarose gelelectrophoresis.442.2.10. Construction of E.coli 09:H10:K99 Chromosomal DNA LibraryAn E.coli 09:H10:K99 genomic DNA library was constructed using afragmentation cloning procedure (Figure 15), where random chromosomalDNA fragments are packaged into plasmid vectors and introduced into hostbacteria. The bacteria are then grown on plates so that each clonecontaining a plasmid incorporating foreign DNA gives rise to a singlecolony, which can then be assayed for production of the desired geneproduct (117). Genomic DNA from the above strain was partially digestedwith the restriction enzyme Sau3AI to create a spectrum of DNA sizefragments. These fragments were then ligated into the complementaryBamHI site of pBR322 with T4 DNA ligase (2 Weiss units) at 16 °Covernight. The plasmid pBR322 codes for proteins that confer resistanceto the antibiotics ampicillin and tetracycline. Escherichia coli JM101cells were made competent (ready to take up foreign DNA) by CaC1 2treatment (135). The cells (5 ml) were grown overnight, then 500 gl ofthe culture was diluted into 25 ml LB and incubated at 37 °C for 2-3 huntil the optical density at 650 nm of the bacterial suspension wasbetween 0.2 and 0.4 (exponential phase). The cells were centrifuged at6,000 rpm in a Sorvall centrifuge at 4 °C for fifteen minutes. Thebacteria were resuspended in 10mM sterile NaCl (10 ml) to wash andcentrifuged to pellet. The cells were then suspended in 12 ml 30 mMsterile CaC12 and incubated on ice for 20 min, centrifuged andresuspended in 1.5 ml 30 mM CaC1 2. Ligation mix (12 gl) was added to 200gl of the competent cells and the mix was incubated on ice for 30 min.The cells were then subjected to heat-shock at 42 °C and incubated inLB-0.2 % glucose at 37 °C for 1.5 h before plating on LB ampicillin45Origin ofreplicationCleovogesiteTc resistance,'"RestrictionenclowdeoseForeign DNACleavagesiteRestrictionendonuc I easeI tTTAAAnnealingAAT TUriAITDNA ligase sealsnicksAATTIrri 1 111TTAAChromosomeRecombinantplamtdTransformation, TcselectionRecombinantplasmtdFigure 15. Diagrammatic representation of random fragment cloningprocedure (from 122)46plates and overnight growth at 37 °C.Plasmid-containing colonies were selected for foreign inserts bytransferring colonies with a sterile toothpick to a grid position onboth an LB ampicillin and an LB tetracycline plate, since cloning intothe BamHI site of pBR322 generally inactivates the plasmid'stetracycline resistance gene. After overnight growth, those coloniesthat grew on ampicillin but not tetracycline were selected for furtherscreening.2.2.11. Library Screening2.2.11.1. Hemagglutination AssayFor the initial clone screening, bacterial colonies were scrapedfrom a plate with a sterile toothpick and mixed in a round-bottomedmicrotitre well with 50 pl of a 1 % v/v fresh human erythrocytesuspension (3x washed in PBS,). The plates then stood for 2 h at 23 °Cbefore being read. Unagglutinated erythrocytes formed a tight button inthe bottom of the well, whereas agglutinated cells were spread over thewhole well surface by adhesive interactions with neighbouring cells.Clones of interest were then titred against the erythrocytes by addingdoubling dilutions of the bacteria in PBS (initial concentration 3.9x10 9cells/mil to 50 pl of red cells in PBS-mannose in microtitre platesuntil no agglutination was seen.472.2.11.2. Oligonucleotide ProbeA 20-nucleotide oligonucleotide pool was designed to a seven aminoacid region of the published N-terminal sequence of the F41 protein(76). This entailed working backwards from the amino acid sequence tothe sequence of three nucleotide codons that determine each amino acid.Since most amino acids are coded by more than one codon, points wherethere is more than one base possibility will occur. At these points,proportions of each possible base are synthesized into the nucleotidechain, and hence a pool of oligonucleotides is built up. The sevenamino acid region with the fewest possible codons was chosen as thetemplate for the probe. The nucleotide probe sequence is shown insection The oligonucleotide pool was synthesized by the UBCNucleotide Synthesis Facility on an Applied Biosystems DNA synthesizer,and purified by chromatography using a C 18SEP-PAKTm (Millipore).The oligonucleotide probe was checked for specificity byhybridization to F41 chromosomal DNA that had been partially digestedwith Sall, electrophoresed on a 0.7 % agarose gel and blotted tonitrocellulose by the method of Southern (123). The hybridization wascarried out in an analogous manner to the colony hybridization describedbelow.The insert-containing clones were grown on plates in grids of 100,colonies were transferred to nitrocellulose filters and alkaline lysedwith 0.5 M NaOH, 1.5 M NaCl. The nitrocellulose filters were thenneutralized in 1.5 M NaC1, 0.5 M tris.Cl pH 8.0 and the DNA permanentlyimmobilized on the filter by baking at 80 °C for 2h (117). An 80 pmolsample of the probe was labelled with 32P using 7-32P-ATP and T448polynucleotide kinase, and hybridized to the filters as previouslydescribed (117). The filters were soaked in 6x SSC (20x SSC = 175.3 g/1NaC1, 88.2 g/1 sodium citrate, pH 7.0) for 5 min then incubated for 2 hin prewashing solution (50 mM tris.Cl pH 8.0, 1 M NaC1, 1 mM EDTA, 0.1 %SDS). Sites on the nitrocellulose that may bind DNA non-specificallywere then blocked by 5 h incubation in prehybridization solution (5xDenhardts solution (50x = 5 g ficoll, 5 g polyvinylpyrrolidone, 5 g BSAin 500 ml H20), 5x SSPE (174 g/1 NaC1, 27.6 g/1 NaH 2 PO 4 .H2 0, 7.4 g/1EDTA, pH 7.4), 0.1 % SDS and 100 Ag/m1 denatured salmon sperm DNA). TheDNA probe was denatured (100 °C, 5 min), was added to the filters andincubated overnight. The probe binds to any regions of the immobilizedbacterial DNA to which it is complementary. Filters were then washedwith high stringency in 1 and 2x SSC, dried and autoradiographedovernight. Any colonies that showed a dark spot on the x-ray film werechosen from replica plates for further study. Antibody Colony BlotsThis assay was adapted from Helfman and Hughes (124). A generaldiagram for enzyme-conjugated antibody assays is shown in Fig. 16.Briefly, bacterial colonies grown in grids on plates were lysed withchloroform onto a nitrocellulose filter and the DNA enzymaticallydegraded with DNase 1. The filters were then blocked with bovine serumalbumin (BSA) to prevent non-specific antibody adsorbtion, washed in PBSand introduced to the antiserum (1/1000 dilution) in PBS-1% BSA for 3 hat room temperature. The antibodies in the serum bind very specificallyto any proteins present against which they were raised. The filters49were washed 3x in PBS-0.5% tween 20 (poly(oxyethylene) 20-sorbitan-monolaurate, a detergent) and transferred to a Petrie dishcontaining either alkaline phosphatase-conjugated goat anti-rabbit IgG(for polyclonal primary antibodies), or alkaline phosphatase-conjugatedgoat anti-mouse IgG (for monoclonal primary antibodies), for 3 h at roomtemperature. These secondary antibodies are specific for regions of theprimary antibody, and bind to them. The washes were repeated, then thefilters were incubated in 0.1 M tris.Cl (pH 9.6) to which was addedX-phosphate (50 pg/m1) and nitroblue tetrazolium (10 pg/m1),pseudosubstrates of alkaline phosphatase that are digested to yield anintense purple colour. After developing for 2-20 min, the filters werewashed once in PBS and air-dried. Immunoblotting of SDS-PAGE SamplesProtein samples were electrophoresed as in Section 2.2.6. andtransferred to nitrocellulose as in Section 2.2.7. The proteins onthe nitrocellulose sheet were then incubated with antibody and developedin the same manner as described in Section above. Enzyme-Linked Immunosorbent AssayThe enzyme-linked immunosorbent assay or ELISA is very similar inprinciple to the above assay, with the exception of the detection enzymeand the fact that the bacteria are bound intact to flat-bottomed wellsin a 96 well plastic microtitre plate. This means that only proteinsexpressed on the surface of the bacteria will react with the primaryantibody. Low-binding micro-ELISA plates (Falcon/Becton Dickinson,50Antigen immobilized on plate or membraneadd specific primary antibody /1\antibody bound to antigen litadd enzyme-coupled secondary antibody specificfor primary antibodyadd pseudosubstrates that are digested tocoloured productMeasure colour, visually orspectrophotometricallyFigure 16 Diagrammatic representation of a generalized enzyme-linkedimmunoassay51Lincoln Park, NJ, USA) were coated overnight at 4 °C with whole bacteria(100 gl of 3.9x109 cells/ml in 15 mM Na2CO3, 35 mM NaHCO3 , pH 9.6, perwell). The plates were washed 3x in PBS-tween and blocked with 3 % BSAin PBS at 23 °C for 30 min. Diluted antibody (usually 1/2000 dilution,100 gl per well) in PBS-tween was added. After incubation for 1 h at 370c, the plates were washed and blocked with 0.5 % BSA in PBS-tween for30 min at 37 °C. The plates were again washed and 100 gl of horseradishperoxidase-conjugated second antibody was added at 1/2000 dilution toeach well. The plates were incubated for 2 h at 37 °C, washed and 200 plfreshly made tetramethylbenzidine solution (Sigma), (0.04 % in 25 mMcitric acid, 51 mM Na2HPO4 , pH 5.0) was added per well. The reactionwas stopped after 2-10 min by addition of 50 gl 4 M H 2SO4 , and theplates were read at 450 nm on an SLT EAR 400 AT ELISA plate reader.522.3. Results2.3.1. Random Fragment CloningFrom the ampicillin-resistant colonies that resulted from thetransformation of E.coli JM101 with the ligation products of the cloningexperiment, 5000 were plated onto grid positions on LB amp and LB tetplates. Of these clones, approximately 1500 did not grow ontetracycline, indicating a foreign DNA insert. The insert-containingclones were screened for hemagglutination. Two were found that showedpartial agglutination and one that showed strong agglutinating activity(clone 8-82) was chosen for further study. The plasmid that thestrongly agglutinating clone harboured was named pETE1 and serialdilution hemagglutination experiments showed that this bacterium wascapable of full agglutination of 1 % Hct human erythrocytes in thepresence of mannose, at concentrations of 6.0 x10 7 cells/ml. Thisfigure is equivalent to that for F41 grown on minca plates, andindicates agglutination at erythrocyte:bacteria ratios of about 2:1. Aphotograph of the microtitre serial dilution assay is shown in Figure17. The plasmid pETE1 conferred the hemagglutination phenotype ontoE.coli JM101, B23 and C600 when purified and retransformed into thesestrains.53JM101JM101(pETE1)4109^III 11111^ c)_+,1:400 00*(3• 111041000.411041, 41111"11 ^+ 0-c c^CoCt•i#41+4,40^i4hivieFigure 17. Hemagglutination microtitre assayAssay was carried out as described in section withinitial bacteria concentrations of 3.9 x10 9 cells/ml. Well 6 from theleft was judged to be the last well where full agglutination occurred.2.3.2. Subsequent Library Screening2.3.2.1. Oligonucleotide Probe HybridizationThe oligonucleotide probe pool synthesized is shown in Figure 18.The last nucleotide of the final codon was omitted as it was anambiguity point which the nucleotide synthesizer could not synthesize atthe 3' end. The probe was designed to hybridize to the non-coding strandof the F41 gene.54H 2N - Ala - Asp - Trp - Thr - Glu - Gly - Gln -5' -GCAGATTGGACTGAAGGTCA -3'C^C^C^G^CT A AG G^GFigure 18. N terminal amino acid sequence of F41 and oligonucleotideprobe pool designed to it.The 32P-labelled probe was first hybridized to a Southern blot ofF41 genomic DNA, partially digested with Sall, to check its specificity.The results (Fig. 19) showed that the probe did recognize F41 DNA andnot JM101 DNA, although hybridization to the bacteriophage A DNAmolecular weight standards was seen. The library was then screened andtwo colonies showed dark spots on X-ray film (Fig. 20), the colonies28-61 and 12-07. Colony 8-82 (JM101(pETE1)) did not hybridize to theprobe.55Figure 19. Southern blot of probe hybridized to partially digested F41chromosomal DNA.(a). Original 0.7 % agarose gel (ethidium bromide stained)(b). Autoradiograph of membrane-blotted DNA from (a) afterhybridization with probe as described in section are not the same size. Gel wells are arrowed.lane 1 : A-HindIII molecular weight standards,lane 2 : A-HindIII/EcoRI molecular weight standards,lane 3 : F41 total DNA Sall digestlane 4 : JM101 total DNA Sall digest.56Figure 20. Autoradiogram of blotted library colonies afterhybridization with oligonucleotide probe.Experiment carried out as described in Section Arrowpoints to positive colony 12- Antibody Colony BlotsThe library colonies lysed on nitrocellulose filters wereincubated with anti-F41 mouse monoclonal antibodies. After incubationwith the alkaline phosphatase-conjugated secondary antibody anddevelopment, one colony, 3-49, showed faint purple colour (Fig. 21).This weak colour was seen when the experiment was repeated. Thecolonies that were positive in the oligonucleotide probe assay werenegative in this assay, as was JM101(p1IE1).57Figure 21. Photograph of antibody colony blotExperiment was performed as in Section This filtershows a weak positive reaction from colony 3-49 (arrowed). Western BlottingWhen electrophoresed and blotted against a-F41 monoclonalantibodies as described earlier, no library clones showed any positivebands. Blotting against a-K99 antibody produced several non-specificbands common to both JM101 and JM101(pETE1). Purified F41 adhesinshowed a minor 18 kDa band against a-K99 antibody This data is shown inFigure 22.58Figure 22. Western blotting of bacteria against a) anti F41 andb) anti K99 antibodies.Experiment performed as described in Section 1 = F41 purified adhesin2 = JM1013 = JM101(pETE1)a) 1 = purified F41 adhesin2 = F41 bacteria (minca grown)3 = F41 bacteria (LB grown)4 = JM1015 = JM101(pETE1)6 = clone 3-49Anti-K99 band in F41 preperation is arrowed.592.3.2.4. Enzyme-linked Immunosorbent AssayTest ELISAs were carried out using dilutions of F41 bacteria ineither carbonate-bicarbonate or glyoxal buffer as positive controls andJM101 as a negative control on both high binding and low binding plates.The results are shown in Fig 23. As the assay appeared to besatisfactory for identifying bacteria expressing the F41 antigen,library clones were tested at 1:500 dilution. No clones were found thatshowed a significant level of accessible F41 antigen expression. Theclones found to be positive in the assays described above were negativein this assay including clone 3-49 and JM101(pETE1).600.002^0.001^0.0005antibody dilutionI -1,- F41 -13- JM -A- HRP --)g- -HRPFigure 23. Graph of absorbance against antibody dilution for whole-cellELISA experiment as described in Section of antibody used for subsequent screening is arrowed. Notest clone showed an absorbance significantly greater than that ofJM101.F41 = E.coli 09:H10:K99, JM = E.coli JM101HRP = horseradish peroxidase alone blank, -HRP = blank without HRP butwith bacteria612.4. DiscussionThe cloning experiments were not successful in the generation of arecombinant plasmid that coded for an intact, functional F41hemagglutinin protein that was expressed on the cell surface. Theclones that hybridized with the oligonucleotide probe probably didcontain some portion of the genetic determinant but not the fullsequence of information necessary for construction and surfaceexpression of the protein. Hybridization seen between the probe and thebacteriophage A DNA molecular weight standards is unusual given thestatistical specificity of a 20-base DNA sequence. Ambiguity points inthe probe, where more than one base could complete the correct codon,raise the probability of non-specific hybridizations. This effect wasnot seen in the library screening as JM101 chromosomal DNA showed nosequence homology with the radioactive oligonucleotide.The clone 3-49, which showed a repeatable weak positive signal incolony blots, was negative in the ELISA assay possibly indicating thatthe F41 antigen was being expressed inside the cell but not exported tothe surface and assembled. This conjecture cannot however explain theWestern blot results where whole cell lysates of 3-49 electrophoresed,blotted onto nitrocellulose and incubated with a-F41 antibodies showedno antibody-recognized band. The denaturing conditions used in thecolony blots (chloroform) were different from those used for theelectrophoresis (SDS, g-mercaptoethanol, heat) and could possibly havecreated a cross-reacting false a-F41 epitope on a recombinant proteinpresent in a library colony.62Since these experiments were performed, Moseley (70) has clonedthe F41 determinant and has shown that a number of genes are necessaryfor the full expression of the adhesin, as is the case with many otheradhesin systems. It may be that ligation of a simple partial digest ofF41 DNA into pBR322 generates a large proportion of recombinant plasmidscontaining small inserts that do not contain the necessary geneticinformation for correct expression. An approach based upon selection ofF41 chromosomal DNA fragments of greater than 6 kb for ligation into avector would seem to have a greater chance of success.The experiments did however allow discovery of a mannose-resistant hemagglutinin that was different from either the F41 antigenor the K99 antigen that the parent strain is known to express. Thishemagglutinin was capable of agglutinating human erythrocytes in amicrotitre assay to the same extent as the original F41 bacteria. Sincethe E.coli strain used for this work, 09:H10:K99, is different from theF41 strains used by most other researchers (E.coli B41 (0101:K99)(60,69) and E.coli VAC1676 (0101:K30:F41:H- )(70), it was postulatedthat the agglutinin coded by pETE1 was an as yet uncharacterized surfaceprotein on this strain and as such was worthy of further investigation.633. Subcloning and Sequencing of pETE13.1. BackgroundOnce the recombinant plasmid pETE1 had been identified as thesource of a gene coding for a protein that caused mannose-resistantagglutination of human erythrocytes, the process of restriction mappingthe plasmid was undertaken. This led to the construction of a number ofsubclones based on the pTZ19R vector, some that retained thehemagglutination determinant and some that lost the genotype. Thenucleotide sequences of some of the pTZ-based recombinant subclones werethen determined by the dideoxy-chain termination method of Sanger et al(125). Single-stranded DNA templates were generated by utilizing theM13 origin of replication present in pTZ to grow the plasmid asbacteriophage. The sequence was then determined by utilizing a seriesof custom oligonucleotide primers to "walk along" the DNA. Sequencingresults (processed using PCGene) were used to generate more subcloneswith specific properties.3.2. Materials and Methods3.2.1. Bacterial Strains, Growth conditions and PlasmidsThe bacterial host for plasmids was JM101, grown as described inSection 2.2.1. Plasmid pTZ19R (117) was used for subcloning and togenerate single-stranded sequencing templates. This plasmid is alsoknown as a phagemid, a name that reflects its ability to propagate as adouble-stranded plasmid or a single-stranded bacteriophage.64For preparation of single-stranded DNA for sequencing from pTZplasmids, cells were grown in TPY broth (16 g tryptone, 16 g yeastextract, 5 g NaCl and 2.5g K2HPO4 in 1 1 distilled water, autoclaved 15min) in the presence of kanamycin (70 Ag / ml).When pUC-based plasmids were used (pTZ), isopropylg-D-thiogalactoside (IPTG, Sigma) (5 mM) and 5-bromo-4-chloro-3-indolyl-g-D-galactopyranoside (X-gal, Sigma) (20 µg/ml) were added to LB agarplates for blue/white screening of transformants. The IPTG induces thelacZ gene present on pUC based plasmids, which codes for the enzymeg-galactosidase. The X-gal is a pseudosubstrate for the enzyme thatproduces a blue coloured product which stains the bacterial colony.When the lacZ gene is disrupted by the cloning of foreign DNA into pUC'smultiple cloning site, g-galactosidase is not produced and the colonyremains off-white.3.2.2. ElectroporationAs an alternative to CaC1 2 -transformation of bacteria, plasmidswere often introduced to host cells by electroporation. Electroporationis a technique that transiently permeabilizes cell surfaces with anelectric pulse permitting the highly efficient uptake of DNA (153).Electrocompetent cells (made by P. Miller using the method of Dower etal (154)) were electroporated typically with 1 Al of ligation mix andpulsed at 25 kV, 0.3 s in a Bio-Rad Gene Pulser electroporationapparatus. The cells were rescued by incubation in LB broth (1 ml) for1 hour at 37 °C. The cells (2-200 Al) were then plated on LB amp andgrown overnight.653.2.3. Sequencing Polyacrylamide Gel ElectrophoresisRadiolabelled products of DNA sequencing reactions werefractionated according to size on urea-polyacrylamide gels. The gelswere formulated by adding 7.5 ml acrylamide-bisacrylamide (40 % : 0.8(Bio-Rad)) to 42.5 ml TBE, giving a final acrylamide concentration of6 %. Urea (25 g) was dissolved in the solution by heating to 55 °C, andthe solution was degassed under vacuum. Polymerization was initiated byaddition of 25 gl TEMED (N,N,N',N', tetraethylmethylene diamine,Bio-Rad) and 300 pl ammonium persulphate (100 mg/ml), the acrylamidesolution was poured between glass gel plates (40.0 cm x 85 cm, with agap (gel thickness) of 4 mm) and left to polymerize fully. The gelswere run vertically in a Bethesda Research Laboratories SI modelelectrophoresis system for up to 8.5 hours at 1500 V. The running bufferused was TBE.3.2.4. Restriction MappingRestriction analysis of plasmids of interest was undertaken forsize estimation and for subcloning. This involved digests of theplasmid DNA with a range of restriction enzymes that cut infrequently(i.e. those with a six-base recognition sequence), followed by agarosegel electrophoresis to count the bands produced and estimate theirmolecular weights. These primary digests (after phenol:chloroformextraction and resuspension in the correct buffer if necessary) werethen digested further with a range of restriction enzymes and thecleavage pattern and molecular weights of products noted. In this way66cleavage sites were arranged in an internally consistent manner inunambiguous locations. The restriction enzymes used are shown inTable 1.Table 1. Restriction enzymes utilized during restriction mapping andsubcloning.Enzyme 5'- site -3' Cuts in pBR322 Cuts in pETE1BamHI G GATCC 1 0EcoRI G AATTC 1 2HindIII A AGCTT 1 1PstI CTGCAIG 1 3ApaLI GITGCAC 3 4ApaI GGGCCIC 0 0KpnI GGTAC C 0 0NcoI CICATGG 0 2Stul AGGICTT 0 0XhoI CITCGAG 0 0AccI GTIMNAC 2 3PvuI CGATICG 1 2XmaI CICCGGG 0 2Smal CCCIGGG 0 2XbaI T CTAGA 0 0Sall G TCGAC 1 2HincII GTPylPuAC 2 4SphI GCATGIC 1 2M = A or C; N = T or G Py = T or C; Pu = A or G3.2.5. SubcloningSubclones of the plasmid pETE1 were generated using the enzymeHincII, which recognizes the hexanucleotide 5'G T Py Pu A C3' (Py =Pyrimidine, either Thymine (T) or Cytosine (C) and Pu = Purine, eitherAdenine (A) or Guanine (G)). This enzyme cuts after the thirdnucleotide and creates blunt ends. In pBR322 the enzyme makes two cuts,whereas in pETE1 it makes four, meaning there are two HincII sites in67the foreign DNA region. Plasmid DNA (approx 2.3 gg) was digested withHincIl (8 U) in a volume of 40 gl for three hours at 37 °C. Stop dye(10 gl) was added and the DNA was fractionated by electrophoresis on a0.8 % agarose gel. Of the four bands that are obtained byelectrophoresis of HincII-digested pETE1, one (3.0 kb) is entirelypBR322 DNA, two (2.3 kb and 1.3 kb) contain both vector and foreign DNA,and one (1.0 kb) is entirely foreign DNA.The three insert DNA-containing bands were excised from the gelwith a scalpel and purified using the Geneclean kit (Fisher). Thisprotocol involves melting the gel in 750 gl 6 M NaI and addition of 5 AlGlassmilkTM , a suspension of silica matrix in water. In high saltconcentrations, DNA adsorbs to the glass microspheres that compose thematrix. The glassmilk was then pelleted by centrifugation for 30 s athigh speed in a benchtop microfuge. The pellet was washed 3x byresuspension in 750 Ml NEW wash (NaCl/Ethanol/Water) followed bycentrifugation at high speed for 5 s. After the final wash, the pelletwas resuspended in 10 gl sterile distilled H 20, heated to 65 °C for twominutes to elute the DNA from the silica, then centrifuged for 30 s.The purified DNA was then transferred in the supernatant to a cleaneppendorf tube and checked for purity and yield by agarose gelelectrophoresis.The vector chosen for religation and recircularization of thefragments was pTZ19R, a pUC-based plasmid with ampicillin resistance andthe lacZ gene. The vector also has a single HincII site in themultiple cloning site. Cesium-pure pTZ19R (5 pg) was digested with 1 UHincII overnight at 37 °C in a volume of 20 gl. The overnight digestwas diluted to 200 gl with TE and extracted once with phenol:chloroformthen once with chloroform, precipitated with ethanol, and redissolved in68100 Al TE. The concentration was estimated by electrophoresis andligations were set up with each of the Genecleaned pETE1 fragments asfollows. In an eppendorf tube were mixed 10A1 pTZ18R-HincII, 8 Al pETE1fragment, 6 Al 5x ligation buffer, 1 Al 10 mM ATP and 20 Weiss units T4DNA ligase (high conc. 15 U/µ1), with sterile distilled H 2O to 30 Al.The ligations were incubated overnight at room temperature thentransformed into CaC1 2-competent JM101 (see detailed description Section2.2.10.). Transformation mix (50 Al) was plated onto LB-amp-IPTG-X-Galplates (4 for each fragment ligation) and incubated at 37 °C overnight.White colonies were selected and stocks were made by as in Section2.2.1. Agarose gel electrophoresis of HincII-digested subclonesconfirmed the ligations were successful.Hemagglutination studies on the subclones (see 2.3.1.) led to thedecision to generate another subclone containing the 1.3 kb and 1.0 kbHincII fragments together in pTZ18R. To do this, pETE1 DNA (2.3 Ag) wasdigested with HincII for one hour in a total volume of 40 pl. Aliquots(5 Al) were taken at 0, 30 s, 1 min, 2 min, 5 min, 10 min, 30 min and1 h, added to stop dye (4 Al) and run on an agarose gel. A digest time(5 min) was chosen such that the plasmid DNA was partially digested andthe transient band running at about 2.3 kb (just below the 2.3 kb bandthat is present in a complete digest) was at a maximum. A digest wasthen performed under exactly the same conditions for 5 min and theentire digest was subjected to agarose gel electrophoresis. The 2.3 kbband was excised, Genecleaned and ligated into the HincII site ofpTZ18R, and the plasmid was then transformed into competent JM101.Clones that gave rise to off-white colonies on LB-amp-IPTG-X-Gal plateswere cultured. Restriction digests with HincII of isolated plasmidsfrom these clones followed by electrophoresis showed that some bacteria69Figure 24. Construction of pPL7.Lane 1 : high Mwt standards, lane 2 : HincII digest of pETE1Lane 3 : HincII digest of pPL770harboured a recombinant plasmid that released a 2.3 kb fragment, and oneclone harboured the desired product, a plasmid that released both the1.3 kb and 1.0 kb regions. The plasmid size was confirmed to beapproximately 6.2 kb (2.86 (pTZ) + 2.3 + 1.0) by digestion with Sall,which cuts the plasmid once only. This plasmid was named pPL7(Figure 26).Once nucleotide sequencing had determined important open readingframe areas in the subclones' DNA, further subclones were produced insimilar manner to investigate these regions. The subclones and theirproperties are shown in Figure 29, Section Nucleotide Sequencing3.2.6.1. StrategyIt was decided to sequence the F41 DNA stretches in the subclonespPL2, pPL3 and pPL4 as these plasmids are based on pTZ19R, whichcontains the intergenic (origin) region of bacteriophage M13 (126).This means that when the bacteria are grown in the presence of kanamycinwith the helper phage M13K07, virions containing single-strandedphagemid DNA are produced. Bacteriophage M13K07 is a derivative of M13with a mutated gene II, a plasmid origin of replication (p15A) and akanamycin resistance gene. When M13K07 infects bacteria that harbourpTZ plasmids, the incoming single-stranded DNA is converted into doublestranded by bacterial enzymes. This form then uses its plasmid originof replication to replicate without the need for viral gene products.However, M13K07 does code for all the proteins necessary for propagationas a virus, and as these gene products build up, they begin to generate71progeny single stranded DNA. The mutated gene II product of M13K07interacts less efficiently with its own origin of replication than itdoes with the origin present on the pTZ plasmids so there ispreferential production of single-stranded phagemid DNA, which can bepurified for use as a sequencing template.The dideoxy chain-termination method (125) was used forsequencing. This method entails firstly producing the DNA to besequenced as a single strand. A single stranded oligonucleotide primer(usually 18-21 bases) complementary to a known region of the template isthen annealed to the DNA strand. The Klenow fragment of the enzyme DNApolymerase (contains the full polymerase activity but lacks the 5'-3'exonuclease activity of the intact enzyme (117) then synthesizes acomplementary strand to the template in a 5' to 3' direction, startingat the 3' end of the primer. The enzyme needs the four deoxy bases assubstrates, and one of them (usually dATP) is radioactively labelled sothat the synthesized strand can be visualized by autoradiography. Theenzyme can also use dideoxy bases as substrates, but when they areincorporated into the growing chain they halt synthesis, as they have nofree 3' hydroxyl group. The polymerization is performed in four tubes,each tube having a low concentration of one of the four dideoxynucleotides included. As the reaction continues, each tube develops apopulation of radioactive DNA strands of varying lengths,having acommon 5' end and varying 3' ends dependent upon incorporation of thedideoxy analogue of a specific base. These DNA populations aredenatured and electrophoresed side by side, with visualization of thebands by autoradiography. The sequence is then read directly from thebottom of the gel (shortest fragments) upwards.Reverse primer, a primer that is complementary to a sequence in72pTZ19R just to the 5' side of the HincII site, was used to sequencearound 300 - 350 by into the fragments. After that the DNA wassequentially sequenced by synthesizing custom oligonucleotidescomplimentary to the sequenced DNA about 50 nucleotides from the end ofreadable sequence from each primer. In this way the recombinant DNAfragments were "walked along" until vector DNA or information from aprimer synthesized in the opposite direction was encountered. Growth and purification of single-stranded DNA sequencingtemplatesEscherichia coli JM101 strains harbouring the pTZ-based sequencingphagemids were inoculated into 2 ml TPY medium containing ampicillinand the helper phage M13K07 (1x10 9 pfu (plaque-forming units)) andincubated at 37 °C for one hour. Kanamycin (70 pg/m1) was added andincubation continued overnight. Cells (1.5 ml) were centrifuged for 5min in a microfuge and the supernatant (which contained the virions) wastransferred to a clean tube. Poly(ethylene glycol) (PEG) 6000 ( 200 glof 25 % in 3.75 M ammonium acetate) was added to the supernatant, whichwas vortexed and stored on ice for 10 min. Centrifugation at 17,000 rpmfor 10 min, 4 °C and removal of all liquid left a tight pellet ofbacteriophage, which was resuspended in TE (20 pl). To this was added200 pl of 4M NaC10 4 , which allows the DNA to bind to a glass fibrefilter in the next step. Whatman GF/C filters were cut with a number 3cork borer to make 7 mm diameter discs. These discs were placed intowells in an ELISA microtitre plate which had each had a small hole boredin the bottom by a hot 21 gauge needle. The microtitre plate wasattached to a base that was connected to a vacuum pump, samples were73loaded into the wells and the vacuum was slowly applied. The vacuumdrew the liquid through the filter paper, leaving the DNA bound. TheDNA was washed 4x by passing 200 Al of 70 % ethanol through the filtersand gentle suction was continued until the filters were dry. Thefilters were placed into capless 0.5 ml eppendorf tubes with needlepunctures in the bottom then each 0.5 ml tube was placed inside acapless 1.7 ml eppendorf. Sterile distilled water (20A1) was carefullyadded to the filter and left at room temperature for 5 min, then bothtubes were centrifuged for 30 s and the eluted single-stranded DNA wasrecovered from the large tube. The quality and quantity of the DNA waschecked by agarose gel electrophoresis. Sequencing ReactionsFor the annealing of the primer to the template, 5 Al of thesingle-stranded DNA prepared by the glass fibre filter method wasincubated with 1 pmol of primer (concentration adjusted so that 1 pmol =1 Al) at 68 °C for 5 min in annealing buffer (56 mM tris.Cl, 20 mM MgC1 2and 70 mM NaC1, pH 7.5, final volume = 10 Al). The initial reverseprimer was purchased from Boeringer Mannheim, subsequent primers werecustom synthesized on an Applied Biosystems DNA synthesizer, andpurified by chromatography on a C18 Sep-PakTM. The annealing mixturewas allowed to cool slowly to room temperature then 1 Al of 0.1Mdithiothreitol (DTT) was added, along with 2M1 of nucleotide labellingmix ( 3.0 AM deaza-GTP, 1.5 pM dCTP and 1.5 AM dTTP). The radiolabel,35S-a-thio-dATP (0.5 Al, 10 MCi / Al, Amersham, Arlington Heights, IL)and finally 3 U of T7 DNA polymerase (Sequenase version 1.0 (UnitedStates Biochemical, Cleveland, Ohio), diluted to 1.5 U / Al) were added74and the reaction allowed to continue for five minutes at roomtemperature. After this time 3.5 gl of the reaction mix was added toeach of the four termination mixes. The termination mixes were made upas 8 MM concentrations of dideoxy nucleotides in 20 mM NaCl and 2.5 glof each mix was pipetted into an individual well of a Nunc microtitreplate (i.e. one well for ddATP, one for ddCTP etc.). The terminationreaction continued for five minutes until stopped by addition of 4 plstop buffer (20 mM EDTA (pH 7.5), 0.05 % (w/v) xylene cyanol FF, 0.05 %(w/v) bromphenol blue in 95 % deionized formamide) to each well. Thetwo dyes in this mix allow tracking of the electrophoresis. The finalmixes were heated at 95 °C for two minutes and 2 gl of each sample wasloaded onto a denaturing urea (50 % w/v) polyacrylamide (6 %) gel inadjacent wells in the order A,C,G,T. The electrophoresis took place at1500 V for up to 8.5 h.Usually, to gain the most information from one primer, three setsof samples were loaded and run for 3, 5, and 8.5 h. The reactions thatwere run for the shortest time (i.e. those giving information onsequence closest to the primer) were often produced using a 10x dilutionof the labelling mix to retard the T7 polymerase activity and producesmaller labelled DNA fragments. The incubation times for both labellingand termination reactions were reduced to 3 min in these cases also.After electrophoresis, the gel was backed with Whatman filterpaper and dried for two hours under vacuum in a Bio-Rad gel drier. Thedry gel was then exposed to X-ray film (Amersham) overnight anddeveloped in a Kodak automatic developer. The autoradiograph was readmanually on a light box.753.2.6.4. Sequence Data HandlingSequence data from autoradiographs was entered into the program PCGeneTM version 4.0 (Intelligenetics) for open reading framedeterminations, nucleotide to protein translations, consensus sequences,prokaryotic protein secretory signal sequences and secondary structureprediction. Searches for homology with known sequences were made usingthe non-redundant SwissProt + PIR + SPUpdate + GenPept + GPUpdatedatabase and the non-redundant GenBank + EMBL + EMBLupdate + GBupdatedatabase by sending sequences by electronic mail to the BLAST server atthe National Institute of Health, Bethesda, Maryland, USA.763.3. Results3.3.1. Restriction map of pETE1From agarose gel electrophoresis, the plasmid pETE1 was found to beapproximately 8.0 kb in size, indicating an F41 DNA insert of around3700 bases. The restriction map generated by single and double enzymedigests is shown in Figure Subcloning and the localization of the agglutinin geneThe subclones generated are shown in Table 2 and figure 26.Table 2. Constructed pETE1 subclonesName Description HApPL2(A)pPL2(B)pPL3(A)pPL3(B)pPL4(A)pPL4(B)pPL7pPL7NpBApETE1 1.3kb HincIl fragment in pTZ19R vectoras above, but opposite orientationpETE1 1.0kb HincIl fragment in pTZ19R vectoras above, but opposite orientationpETE1 2.3kb HincIl fragment in pTZ19R vectoras above, but opposite orientationpETE1 1.3 + 1.0 HincIl Fragments in pTZ19RpPL7 with a 395bp NcoI deletion in ORF1(see Section 3.3.2) (construction Figure 32)A control plasmid for partitioning studies.pBR with a 17bp BanII disabling deletion inthe tet gene------+++--HA = hemagglutination activity in microtitre assay77XmalSmalAcclAcclSallHincllSphlPstlAcclApaLlApaLlSphl SallHincllAcclHIXmalPstlNcolHincll NcolPvulHindlIlEcoRIApaLlPstlPvulHincllFigure 25. Restriction map of pETE1F41 DNA insert is shown as thin line.^= Baal sites thatwere not regenerated upon incorporation of foreign fragment.78The first subclones were made from HincII fragments of 2.3, 1.3 and1.0 kb ligated into the HincII site of pTZ19R. The subclones were testedfor hemagglutination with negative results, therefore a subclonecontaining both the 1.3 and 1.0 fragments was made (Figure 24, Section3.2.5.). This clone, named pPL7, showed a mannose-resistant agglutinationof human erythrocytes in a microtitre assay to a level equivalent to thebacteria containing the pETE1 plasmid. It was decided to sequence theforeign DNA present in pPL2 and pPL3. Location of an incomplete openreading frame in pPL7, interrupted by the 1.0\2.3 fragment HincIIjunction, led to the sequencing of the pPL4 plasmid also.3.3.3. Identification of Open Reading FramesOpen reading frames were discovered in several locations on theplasmids (Fig. 26). The interrupted tetracycline resistance gene frompBR322 runs into the insert DNA for 47 bases until a stop codon isreached. After just one base there is an initiation codon (ATG) followedby a 792 base-pair open reading frame that would code for a protein of29,024 Da. The initial region of DNA sequence is shown in Figure 27.This putative protein was named plprotl. The open reading frame spansthe HincII site between the 1.3 and 1.0 fragments. The start codon doesnot appear to have a consensus promoter/initiation site or ribosomebinding site (RBS) (131). This suggests that an ATG codon later in thereading frame could be the true starting codon. There are two suchcodons close downstream, the first of which has a putative ribosomebinding site (GGAG). The nucleotide sequence and its translation areshown in Figure 28, amino acid composition shown in Table 3.79pETE1 insert = 3.4 kb1.0 kb2.3 kbpPL72.3 kbFigure 26. Subclones of pETE1 showing open reading frame areasHincII fragments were ligated into HincII site of pTZ19R= open reading frameEl = pBR322 DNA; tet = tetracycline resistance gene (truncated),-+ = direction of putative transcription/translation80start --stop —Figure 27. DNA sequence in the region of the piprotl putative startcodonfrom bottom - ATC1111GCATGG1111AGAATGAGCAACCCCATTTATCGTTCGGGG^SFAYF* MSNPIYRSG....piprot1 ,481Insert'^ -35CATGGCGACCACACCCGTCCTGTGGATCGATTCGATCGAAAACATTCCCTTTACCCTATCMATTPVLYIDSIENIPFTLS-10^1TTTTGCATGGTTTTAGaATGAGCAACCCCATTTATCGTTCGGGGAAAGTGATAGCGGAGAFAYF* MSNPIYRSGKVIAETGATTGAAATGAATAAGGTTATTGCGGTTTCAGCGCTTGCCATGGCAGGCATGTTTTCGAMIEMNKVIAVSALAMAGMFSCCCAGGCTCTGGCTGATGAGAGCAAAACAGGCTTTTATGTGACCGGTAAAGCAGGTGCTTTQALAIDESKTGFYVTGKAGACTGTTATGTCACTTGCAGACCAGCGTTTCCTGTCGGGGAATGGAGAGGAAACATCAAAATS VMSLADQRFLSGNGEETSKATAAAGGCGGTGATGGCCATGATACGGTATTCAGTGGCGGTATCGCGGCCGGTTATGATTYKGGDGHDTVFSGGIAAGYDTTTACCCGCAGTTCAGTATTCCGGTTCGTACGGAACTGGAGTTTTACGCTCGTGGAAAAGFYPQFSIPVRTELEFYARGKCTGATTCGAAGTATAACGTAGATAAAGACAGCTGGTCAGGCGGTTACTGGCGTGATGACCADSKYNVDKDSWSGGYWRDDTGAAGAATGAGGTGTCAGTCAACACACTGATGCTGAATGCGTACTATGACTTCCGGAATGL KNEVSVNTLMLNAYYDFRNACAGTGCATTCACACCATGGGTATCTGCAGGGATTGGCTACGCCAAGGAAATTCACCAGAD SAFTPWVSAGIGYAKEIHQAAACAACCGGTATCAGTACCTGGGATTATGGGTACGGAAGCAGTGGTCGCGAATCGTTGTK TTGISTWDYGYGSSGRESLCACGTTCAGGCTCTGCTGACAACTTCGCATGGAGCCTTGGCGCGGGTGTCCGCTATGACGS RSGSADNFAWSLGAGVRYDTAACCCCGGATATCGCTCTGGACCTCAGCTATCGCTATCTTGATGCAGGTGACAGCAGTG^ TPDIALDLSYRYLDAGDSSTGAGTTACAAGGACGAGTGGGGCGATAAATATAAATCAGAAGTTGATGTTAAAAGTCATG^ SYKDEWGDKYKSEVDVKSHACATCATGCTTGGTGTGACTTATAACTTCTGAcaaaactgctcctgaaagataataattacD IMLGVTYNF*----A^ L____ccatcttctgtaattaaacgaagactccctggcagtaattatactgcagggttgtttgttFigure 28. Nucleotide sequence and deduced amino acid structure ofplprotl.Predicted secretory signal sequence is underlined. Arrows show possibleinverted repeats; • = termination codon; a = single nucleotide betweenstop codon of possible tet-fusion protein and start codon of plprotl ORFOther potential plprotl start codons are double underlined.82Table 3. Predicted amino acid composition of plprotlResidue No. of aa's %ageAla 23 8.7Arg 10 3.7Asn 10 3.7Asp 23 8.7Cys 0 0Gln 4 1.5Glu 12 4.5Gly 28 10.6His 3 1.1Ile 11 4.1Leu 14 5.3Lys 16 6.0Met 8 3.0Phe 11 4.1Pro 5 1.8Ser 31 11.7Thr 13 4.9Trp 6 2.2Tyr 19 7.1Val 17 6.4(%age = number of residues/total residues in protein)The predicted protein contains three putative secretory signalsequences as predicted by the method of Von Heijne (127). The potentialsignal sequences are given a score based on a survey of known signalpeptides. The highest-scoring plprotl predicted signal sequence is shownbelow.6^10^ 20^ 30^39yrsgkviaemiemnkviaysalamagmfstqala40D E S Figure 29. plprotl predicted secretory signal sequenceAmino acids shown in single letter code (see abbreviations).Vertical line indicates predicted cleavage point, "+" denotes positivelycharged residues.83The lower scoring predictions also identified this region butpredicted cleavage between residues 30-31 and 37-38. This signalsequence is fairly long, but there are other potential start codons(coding for methionine residues) present in the signal region, one ofwhich appears to have a consensus ribosome binding site (Met at position15). This codon could be the true initiation codon of plprotl. Thiswould not affect the size or amino acid composition of the matureprotein. The mature plprotl protein would contain 225 amino acids andhave a molecular weight of 24,896 Da. All of the predicted signalsequences conform to the -3,-1 rule which states that the amino acidresidues in the -3 and -1 positions relative to the cleavage point shouldbe small and uncharged.The predicted protein plprotl was scanned for potential membraneassociated regions by the methods of Klein et al (128) and Eisenberg etal (155). No section of this protein had a sufficiently largeconcentration of hydrophobic residues to be positively identified as aputative membrane spanning sequence. This point is discussed further inChapter 6. An analysis of the possible secondary structure by themethod of Garnier (129) predicted a predominantly extended form withappreciable helical content, whereas a prediction according tothe method of Gascuel and Golmard (130) indicated a predominantly coiledform. (Table 4 and Figure 30).Table 4. Comparison of plprotl secondary structure predictionsconformation Garnier Gascuel and GolmardHelical 31.4 % (83 aa) 36.7 % (97 aa)Extended 45.8 % (121 aa) 14.3 % (38 aa)Turn 11.7 % (31 aa)Coil 10.9 % (29 aa) 48.8 % (129 aa)84predicted signal sequencecleavage point10^20 30 40 50 60MSNPIYRSGKVIAEMIEMNKVIAVSALAMAGMFSTQALADESKTGFYVTGKAGASVMSLA...............................*******"X-XXXXXXX)0000000G<XXXXXXX)COOCX*** *----****XX--XXX^70^80^90^100^110^ 120DQRFLSGNGEETSKYKGGDGHDTDSVAVIAAGYDFYPQFSIPVRTFIFFYARGKADSKYNGARN XX--*****XXX-\\\\\\***^\\ -\ -* XXXXXXXXXXX----G+G^X***********************X----XX*********-***XXXXXXXXXXXXXX*-^130^140^150^160^170^180VDKDSWSGGYWRDDLKNEVSVNTLMLNAYYDFRNDSAFTPWVSAGIGYAKEIHQKYYGISGARN _-_\\\*\\\_-\\***XXXXX^\\*\^***XXXXXX-_*\\_-G+G^***********XXXXXXX__-_XXXXXXXXXX********^***XXXXX*******-^190^200^210^220^230^240TWDYGYGSSGRESLSRSGSADNFAWSLGAGVRYDVTPDIALDLSYRYLDAGDSSVSYKDEGARN ---\\\****^\*\*** _x*\**^ \\^XXG+G^--*******************XXXXX*****--*****XXXXXXXXXXX****---****250^260WGDKYKSEVDVKSHDIMLGVTYNFGARN XXXXXXXXXXXXXXXGARN ***XXXXXXXXXXXXFigure 30. Semi-graphical display of secondary structure prediction forplprotl by the methods of Garnier and Gascuel and GolmardGARN = Garnier prediction; G+G = Gascuel and Golmard predictionX = Helical Conformation; - = Extended Conformation; * = CoilConformation: \ = Turn ConformationGARNG+G^*XXX - - -XXX85The second open reading frame begins 169 by later, is 522 by andcodes for a putative protein of 18,394 Da (named plprot2). This openreading frame has consensus transcription/translation initiation andtermination sites and spans the HincII site between the 1.0 and 2.3fragments. The nucleotide sequence and translation of this open readingframe are shown in Figure 31.Hemagglutination observations for the plasmid pPL7 can berationalized with the preceding information as follows. Bacteriaharbouring the plasmid pPL2, containing just the 1.3 kb HincII fragmentfrom pETE1, do not agglutinate red cells. Included on the 1.3 fragmentis the tet promoter from pBR322, which promotes transcription/translation of a truncated tetracycline resistance protein (156). Thistruncated protein was therefore not responsible for the agglutinationactivity present in pPL7, even though the protein signal sequence isintact so presumably the truncated form is being exported through thebacterial inner membrane but perhaps is not inserted into the outermembrane. The gene coding for plprot2 is not complete in pPL2. In pPL7,the full plprot1 open reading frame is present, but only part of the geneencoding plprot2. The pPL7 plasmid carries the hemagglutinindeterminant, therefore the likelihood was that the hemagglutinin was theprotein product of the complete open reading frame (ORF), plprotl. Itwas also possible however that the truncated product of the incompleteORF could be mediating the adhesion.To resolve this uncertainty, the plasmid pPL7N was constructed(see Figure 32). This plasmid has a deletion in the plprotl open readingframe and cells containing this plasmid are completely inactive in amicrotitre hemagglutination assay, even with very high initial86-35^ -10^1tctggtttttattctcttcacaataatgttgtcgatatactccgATGAATATCAGAAAGCMNIRKTGTTTTATACGGACAATATATCCCGGATTTTATTGTTTTTATTCTTCCCGGGCATGTGTGL FYTDNISRILLFLFFPGMCTAATAACCACTGCCGCATGCGGATACACTGAGAAGAATGCCACCGGGAATGTGTTGCTGC^ ITTAA1 C'GYTEKNATGNVLLTGTTTCTCCTTCTGATTCTTGCACACAGAAATACCCTTACATCCATTACAGCTCTGTTATL FLLLILAHRNTLTSITALL0110.11■•••TTCTGTTCTGTTGTGCACTGTATGCACCTGTCGGTATGACGTACGGAAAAATCAACAACAFLFCCALYAPVGMTYGKINN1•0•■••■••^■•■••■■•■••GTTTTATTGTCGCGTTGTTGCAGACCACGGCGGATGAGGCTGCGGAGTTTACCGGGATGAS FIVALLQTTADEAAEFTGMTTCCTGTTTATCATTTCTGGTCAGTGCCGCGATTCTGGTGTTCATGGTTATTTTCTGGCGIPVYHFWSVPRFWCSWLFSGGACACACCACCGTGGTCGACGAAACATACGCCGATGGCGACCCAGCGTCAGACGCCTGCTG HTTVVDETYADGDPASDACACGGCAAAAAAAGCGTCGTCTACGCAATCTCGTCCGGTGAATGCTGGTGCGAAGAAACGGYGKKSVVYAISSGECWCEET—...s.CCTAAagcggcggttggacgttaaaaaatattcccggcaactgacacgctacgattA -tttttatcattctgcgtacatacgcaagFigure 31. Nucleotide sequence and deduced amino acid structure ofplprot2Predicted signal sequence : solid underliningPredicted membrane spanning segments : dashed underlining87HincII HincIIinsert 2300 by792 byORF1^ ORF2ItetNcoI^NcoIf 395 by iDigest pPL7 withNcoI, recircularizeand transformJM101Figure 32. Construction of plasmid pPL7NEli = pBR322 DNA; tet = truncated tetracycline resistance geneMI = Open Reading Frames (not to scale)pPL7 = pETE1 HincII fragment (from partial digest) inserted intopTZ19R HincII site.JM101(pPL7) shows strong mannose-resistant hemagglutinationJM101(pPL7N) shows no hemagglutinating activity88concentrations of bacteria. This data localizes the agglutinin to theplprotl open reading frame.The putative plprot2 protein was predicted to be an integralmembrane protein with a prokaryotic secretory signal sequence, which whencleaved would leave a mature protein of 135 amino acids, 14,802 Da.Interestingly, plprot2 was also predicted to have a membrane lipoproteinlipid attachment site, a site which generally occurs just after thesignal sequence. The enzyme signal peptidase II recognizes a conservedsequence and cuts in front of a cysteine residue to which aglyceride-fatty acid lipid is attached (132). This consensus sequenceappears in plprot2 as shown below.-Ile-Thr-Thr-Ala-Ala-CYS-Gly-Tyr-Thr-Glu-Lys-32^ 37^ 42The sequencing of the 2.3 kb HincII fragment revealed three openreading frames in the opposite direction (Fig. 26) that code for proteinsof 24,623 Da (plprot3), 9,555 Da (plprot5) and 7,801 Da (plprot4). Thenucleotide sequences and deduced amino acid sequences for these putativeproteins are shown in Figures 33,34 and 35. The plprot3 and plprot5proteins each contain predicted signal peptides which, when cleaved,would create mature proteins of 22,681 Da and 8,079 Da respectively.Although these proteins were not responsible for any of the agglutinationeffect (because pPL4, which contains only this fragment, shows noagglutinating activity), they were scanned for homology with otherproteins in the genebank.89-35^ -10^1aaaagaacaactcatctatatccgggataaacgcaacggagaggtgaaaaacagATGAAAM KATAATACTTCTGTTTTTAGCAGCCCTGGCAAGTTTTACCGTACACGCACAGCCCCCCTCAIILLFLAALASFTVHAQiPPSCAGACCGTAGAACAAACAGTCCGGCATATTTATCAGAACTATAAATCAGATGCCACTGCCQ TVEQTVRHIYQNYKSDATACCTTATTTTGGTGAAACCGGAGAGCGGGCGATAACTTCTGCGCGTATTCAACAGGCGCTTP YFGETGERAITSARIQQALACCCTGAACGACAATCTTACGCTGCCGGGCAATATTGGCTGGCTGGATTATGATCCGGTTTLNDNLTLPGNIGWLDYDPVTGTGATTGTCAGGATTTTGGCGATCTGGTGCTAGAAAGCGTTGCGATAACCCAAACTGACCDCQDFGDLVLESVAITQTDGCCGATCATGCCGATGCCGTTGTGCGCTTTCGTATCTTTAAAGATGATAAAGAAAAGACCADHADAVVRFRIFKDDKEKTACGCAGACACTGAAAATGGTGGCGGAAAATGGTCGTTGGGTCATTGACGATATTGTCAGCTQTLKMVAENGRWVIDDIVSAATCATGGCAGCGTCTTACAAGCAGTTAATAGCGAGAATGAAAAAACGCTGGCCGCTTTAN HGSVLQAVNSENEKTLAALGCTTCGTTGCAAAAAGAACAGCCGGAAGCCTTTGTTGCCGAACTCTTTGAACATATTGCTASLQKEQPEAFVAELFEHIAGATTATAGCTGGCCGTGGACGTGGGTGGTTTCCGACTCTTACGCCAGGCGGTTAATGCCTD YSWPWTWVVSDSYARRLMPTCTATAAAACCACCTTCAAGACGGCCAATAATCCCGATGAAGATATGCAAATAGaacggcS IKPPSRRPI IPMKICK-aatttatttacgacaatccgatctgttttggcgaagagtcgctattttcFigure 33. Nucleotide sequence and deduced amino acid structure ofplprot3Putative signal sequence is underlined.90-35^ -10^1ggcctggtggtgatcccgaaatcggtcacaccttcacgtattgccgaaactttgATGTCTM SGGGATTTCCGTCTCGACAAAGACGAACTCGGCGAAATTGCAAAACTCGATCAGGGCAAGCG ISVSTKTNSAKLQNSIRASGTCTCGGTCCCGATCCTGACCAGTTCGGCGGCTAACATGCAAATTCTCCCGGTGGCGGTA^ SVPILTSSAANMQILPVAVATGTTCCGCTACCGGACTTTTCAGAAATCATTTATTCCCCTCGCGTCCCGCCCGTTGTTAMFRYRTFQKSFIPLASRPLLCTCTTCCTTGTTCAGGAATGCCAAATATAAggacatcatcatgcagagccggaagctcttL FLVQECQI-aaaagaacaactcatctatatccgggataaacgcaacggagaggtgaaaaacagatgaaaataatacttctgtttttFigure 34. Nucleotide sequence and deduced amino acid structure ofplprot49135caaaaactgccgccagattaaagcaatcagctcaattaaataacaattagccggaacaat-10^1aaataaaacccaacactatATGAAAACGAGGTTCACCGTGGGAGCTGTTGTTCTGGCAACMKTRFTVGAVVLAT^4CTGCTTGCTCAGTGGCTGCGTCAATCAGCAAAAGGTCAATCAGCTGGCGAGCAATGTGCACLLSGCVNQQKVNQLASNVQAACATTAAATGCCAAAATCGCCCGGCTTGAGCAGGATATGAAAGCACTACGCCCACAAATTLNAKIARLEQDMKALRPQICTATGCTGCCAAATCCGAAGCTAACAGAGCCAATACGCGTCTTGATGCTCAGGACTATTTYAAKSEANRANTRLDAQDYFTGATTGCCTGCGCTGCTTGCGTATGTACGCAGAATGATAAaaaatccccggtaggcgtgtDCLRCLRMYAE- -cagttgccgggaatattttttaacgtccaaccgccgcttFigure 35. Nucleotide structure and deduced amino acid sequence ofplprot5Putative signal sequence is underlined.923.3.4.2. Genebank Homology SearchesEach putative protein amino acid sequence was scanned for homologywith other known proteins. No significant similarity with known proteinsequences was found for plprot2, plprot3 and plprot4. The proteinplprotl showed limited similarity to a region of the Neisseriagonorrhoeae opacity protein (133) which has been implicated in thevirulence of these organisms, and other Neisseria gonorrhoeae outermembrane proteins. The region of strongest identity is shown inFigure 36.The putative protein plprot5 showed 50 % identity and 74 %similarity with the precursor form of the major E.coli outer membranelipoprotein, a 77 amino acid protein that is covalently attached to themurein component of the inner side of the bacterial cell wall (134). Thealignment of the two sequences is shown in figure 37.140^ 1681plprotl SVNTLMLNAYYDFRNDSAFTPWVSAGIGY+V+ +L L+A YDF+ +^F P+ + A + Yopacity AVSSLGLSAVYDFKLNDKFKPYIGARVAY661^ 747Identity = 37 % Similarity = 68+ indicates functional similarityFigure 36. Identity between plprotl and N.gonorrhoeae opacity proteinIdentical and functionally similar residues are indicated inmiddle row.932^ 311 1plprot5 KTRFTVGAVVLATCLLSGCVNQQKVNQLAS+ T +^+GAV+L + LL+GC+^K + + Q L + +lipo^RTKLVLGAVILGSTLLAGCSSNAKIDQLST1 13^ 3232^ 611 1plprot5 NVQTLNAKIARLEQDMKALRPQIYAAKSEA+VQTLNAK+ + L^D+ A + R^+ AAK + Alipo^DVQTLNAKVDQLSNDVTAIRSDVQAAKDDA1 133^ 6263^711 1plprot5 NRANTRLDAQRAN RLDQlipo^ARANQRLDNQ64^72Figure 37. Identity between plprot5 and E.coli Major outer membranelipoprotein precursorIdentical and functionally similar (+) residues are indicated inthe middle row.50 % identity, 72 % similarity943.4. DiscussionThe region of DNA that was cloned contained open reading framescoding for four putative periplasmic, membrane or secreted proteins.These proteins mostly showed no significant similarity to amino acidsequences present in protein databases, although plprot5 was found to be50% identical to a known E.coli outer membrane lipoprotein. Thislipoprotein is covalently attached to the murein layer in the periplasmand the gene for this protein is found in single copy on the chromosome(134). These findings suggest that an area of the F41 genome determiningmembrane structure and function was cloned into pBR322.The nucleotide sequence of the plprotl ORF was determined for bothDNA strands as shown in Appendix 4. The nucleotide sequences of theother open reading frames were generated predominantly from one strandonly, although in the case of plprot2 there is significant dual strandinformation. The nucleotide sequences of plprot2,3,4 and 5 may thereforecontain errors although the presence of putative termination signals inthe cases of plprot2 and plprot5 suggest that these possible errors maybe relatively minor base substitutions rather than frameshift errors.The consistency of identity that plprot5 shows with lipoproteinthroughout its sequence also indicates that the sequence information forthis ORF is correct.Subcloning results identify the 29,024 Da protein plprotl as theagglutinating agent. Based on physical hemagglutination propertiesdiscussed in Chapter 4, plprotl was named Heat Resistant Agglutinin 1(HRA1). This protein has a putative secretory signal sequence which,95when cleaved, would produce a mature protein of 24,896 Da.Antigenic determinants of the adhesin (HRA1) were predicted bycomputer using the method of Hopp and Woods (136), which searches forregions of highest hydrophilicity. This method performed very well on aset of control proteins, with the highest predicted antigenic determinantcorrelating to a known determinant in 100 % of the cases. Lower scoringpredictions were correct 66 % of the time. Antigenic determinants arelikely to be exposed at the surface of the protein and can, in someinstances, be correlated to receptor binding areas. The highest scoringpredicted antigenic determinant on HRA1 is shown below.132^ 138- Arg - Asp - Asp - Leu - Lys - Asn - Glu -+Figure 38. Predicted antigenic determinant on HRA1The receptor binding site of the adhesin K99 has been identifiedusing site-directed mutagenesis (105) to be the region shown in Figure 39below. This region also corresponds to the highest predicted antigenicdeterminant in the protein. This motif appears to be implicated in thebinding of three other sialic acid binding lectins, CFA1 (Colonizationfactor antigen 1), CT-B (Cholera toxin B) and LT-B (E.coli heat-labiletoxin B) (106, Figure 39).96CFA1^Lys - Lys - Val - Ile - Val - LysCT-B^Lys - Lys - Ala - Ile - Glu - ArgLT-B^Lys - Lys - Ala - Ile - Glu - ArgK99 Lys - Lys - Asp -^Asp - Arg(HRA1^Arg - Asp - Asp - Leu - Lys - Asn - Arg)Figure 39. Comparison of implicated and identified binding site aminoacid sequences (adapted from 106).Although the HRA1 antigenic determinant is quite different insequence to the determined binding site regions, it has a similar residuecomposition and it is interesting to speculate on the possible receptorbinding importance of this area. This region would be an excellentcandidate for site directed mutagenesis experiments to assess anyadhesive function.Another interesting point about the deduced amino acid structure ofHRA1 is that there are no cysteine residues present. The K88 and CFA1adhesins (60,38) and a number of other outer membrane proteins contain nocysteine residues and it has been suggested (68) that this is necessaryto protect the structure from conformational changes induced bydifferences in the redox potential of the external environment.974. Biochemical Studies on Proteins Encoded by pETE14.1. BackgroundOnce it was established that the recombinant plasmid pETE1 andsome of its subclones mediated a mannose-resistant agglutination ofhuman erythrocytes, attempts were made to isolate the adhesive proteinfrom the surface of the cell. When physical, chemical and geneticmethods of doing this met with limited success, experiments weredesigned to identify the receptor on the red cell membrane. Maxicellanalysis on the clone was undertaken in an attempt to elucidateplasmid-encoded proteins. When a partial purification of the adhesiveprotein was achieved, its N-terminal amino acid sequence was determinedand compared to information derived from the nucleotide sequence.4.2. Materials and Methods4.2.1. Bacterial Strains and Growth ConditionsThe bacteria used were as in Section 2.2.1., with the addition ofEscherichia coli B23 and C600 (117) which were used as plasmidhosts in the dissociation studies. Escherichia coli CSR603 (138) wasobtained from the American Type Tissue Collection (ATCC), and used inthe maxicell experiments.For adhesin production studies, bacteria were grown on Mincamedium (139, 1.36 g/1 KH 2 PO 4 , 7.56 g/1 Na 2 HPO 4, 1 g/1 glucose, 1 g/1casamino acids (Difco) and 1m1/1 trace salts solution (40 mM Mg 2+ , 5 mM98Mn2+ , 0.6 mM Fe2+ and 2.7 mM Ca 2+ ), autoclaved 20 min) or LB (Sect.2.2.1.). In maxicell studies, bacteria were grown in M9 minimal media(117) ( 6 g/1 Na2HPO4 , 3 g/1 KH2PO4 , 1.0 g/1 NH 4C1, pH7.4. Afterautoclaving 20 min, MgSO 4 was added to 20 mM, glucose to 0.2 % and CaC1 2to 0.1 mM. These solutions were previously filter sterilized.). For the35S-labelling of bacteria, the MgSO 4 was omitted from the media. Mediawas solidified by the addition of 1.5 % agar and supplemented withampicillin, IPTG and cycloserine (200 µg/ml) (Sigma) where necessary.4.2.2. Physical Methods for the Isolation of the Agglutinin4.2.2.1. Temperature/vortex ExperimentsThe initial series of isolation attempts was based on observationsthat large amounts of F41 adhesin were removed from the cell surface bysimple vortexing (69). Escherichia coli F41 and E.coli JM101, B23, B/Rand C600 harbouring recombinant plasmids were grown on solid mediaovernight and harvested by scraping the cells into PBS. The cells werewashed once with PBS and resuspended in PBS to an approximateconcentration of 5x10 9 cells/ml as measured by filling a microhematocrittube with bacterial suspension, centrifugation for 10 min and measuringthe percentage of the tube filled with packed cells (1 % a 2.5x109cells/ml). The cells were then subjected to cycles of heating,vortexing and blending in a Waring blendor as detailed in Table 5.99Table 5. Disruption experimentsExpt. Experimental Details1. vortex 5 min2. heat 60 °C 30 min, vortex 5 min3. heat 60 t 30 min, vortex w/glass beads 5 min4. Waring blend 5 min5. heat 60 t 1 h, Waring blend 5 min6. heat 60 t 3 h, Waring blend 10 min7. heat 60 °C 1 h, Waring blend 20 min8. add 10 mM EGTA, heat 60 °C 1 h, Waring blend 20 min9. heat 70t 1 h, vortex 5 min10. heat 100 t 5 min, vortex 5 min11. heat 100 t 5 min, Waring blend 20 minsAfter these treatments, the bacteria were centrifuged for 5 min athigh speed in a microfuge, the supernatant was transferred to a new tubeand the pellet was resuspended in the same volume PBS. The supernatantwas usually concentrated 40x by centrifugation in Amiconmicroconcentrators with a 10,000 Da cutoff. Both the supernatant andthe pellet were then tested for hemagglutinating activity in amicrotitre serial-dilution assay (see Section, electrophoresedon a 12.5 % denaturing polyacrylamide gel, and sometimes blotted ontonitrocellulose for receptor studies with biotinylated red cell ghosts(see Section 2.2.8.). Often several of the above experiments would beperformed in sequence to the same bacteria resuspended in PBS.1004.2.2.2. Detergent Extraction ExperimentsThe bacteria were grown and washed as above and incubated withvarious detergents for 1 h, 60 °C, at concentrations below the criticalmicelle concentration (CMC). The detergents used are shown in table 6.Table 6. Detergents utilized for solubilization of membrane proteinsDetergent CMC 1 PropertiesOctyl GlucosideCHAPS2DeoxycholateSDS25 mM0.49 %4-6 mM84 mMnon-ioniczwitterionicanionicanionic1CMC = Critical Micelle Concentration2CHAPS = 3-[(3-chloramidopropy1)-dimethylammonio]-1-propanesulfonateThe bacteria were then centrifuged and the supernatant wasseparated from the pellet, which was resuspended in the same volume. Forhemagglutinations, both the resuspended pellet and the supernatant wereextensively dialysed (2 days) against 50 mM phosphate buffer pH 7.4.For SDS-PAGE, the supernatants were concentrated for 90 min in Centriconfilters with a 10,000 Da cutoff. Osmotic ShockEscherichia coli F41, JM101 and JM101(pETE1) were grown overnightand suspended to 12.5x109 cells/ml in 10 ml 20 % sucrose-PBS. The cells101were incubated overnight at 4 °C then poured rapidly into 75 ml 5 mMphosphate pH 8.0 and mixed well. The bacteria were centrifuged and thesupernatants concentrated in Centricon filters with a 10,000 MWt cutoff.The supernatants and resuspended pellets were tested forhemagglutination and electrophoresed on SDS-PAGE. SonicationCrude membrane preparations were made by sonic disruption ofE.coli F41, JM101, JM101(pETE1) and a number of subclones in a Sonifier350 cell disrupter (Branson). The bacteria were grown overnight in 5 mlLB, harvested by centrifugation and resuspended in 0.5 ml 50 mMphosphate buffer pH 7.4. The cells were then sonicated for 30 s, 50pulse, setting 3, on ice. The resulting suspension was centrifuged andthe supernatant (containing the cytosolic proteins) was separated fromthe pellet (crude outer and inner membranes), which was resuspended inthe same volume. Hemagglutinations and SDS-PAGE were performed on bothpreparations.4.2.3. Erythrocyte Receptor Studies4.2.3.1. Species Hemagglutination ProfileThe bacterium containing the recombinant plasmid, E.coliJM101(pETE1), was tested for hemagglutination against a variety oferythrocytes from different species. The host cell, JM101, and F41 wereused as controls. The bacteria were suspended to 1.75x10 10 cells/ml,and the erythrocytes were washed 3x in PBS and used at a 1 % hematocrit.102The erythrocytes came from the following species - goat, rat, mouse,guinea pig, cow, rabbit, chicken, cat, pig, dog and human. Binding Studies with Erythrocyte Ghosts(a) Erythrocyte Ghost PreparationThe results from the species' hemagglutination profile showed thatthere were some interesting differences between E.coli JM101,JM101(pETE1) and F41. Erythrocytes from two species, pig and dog, werechosen for further study as they interacted very differently inhemagglutination tests against these bacteria (see Section cell ghosts were prepared by washing the cells (5 ml) twice in PBS,then making the cells back up to 5 ml with PBS. The cells were lysed bypouring this suspension into 20 volumes ice-cold 10 mM phosphate pH 8.0and incubating on ice for 30 min. The solution was centrifuged for 30min at 6,000 g and the supernatant carefully pipetted off to leave aloose pellet of erythrocyte membranes (ghosts). The ghosts were washedby suspension in 10 mM phosphate pH 8.0 followed by centrifugation,until they lost most of their red colour (4-6 washes). The last washwas in water, and the pellet after centrifugation was frozen at -20 °Cuntil needed.(b). Biotinylation of Ghosts and BacteriaThe erythrocyte ghosts were biotinylated by suspension of 200 plghosts in 400 pl carbonate-bicarbonate buffer pH 8.5 and addition of2 pl NHS-LC biotin (Pierce, 200 mg/ml in H20). The reaction was left on103ice for 1 h, centrifuged and resuspended in tris-buffered saline (TBS,20 mM tris,500 mM NaC1 pH 7.5). The NHS-LC biotin is a water solubleanalogue of biotin with an extended spacer arm that reacts with aminegroups. For this reason the reaction was performed in carbonate-bicarbonate buffer, and any excess biotin quenched with tris. Bacteriawere biotin-labelled in an identical manner, with cells suspended in 400Al carbonate-bicarbonate to a concentration of 12.5x10 10 cells/ml.After biotinylation the cells were centrifuged, resuspended in 50 Al TBSand heated for 1 h at 60 °C. The cells were then vortexed for 5 min andcentrifuged. The supernatant (containing any biotinylated proteins thatwere removed from the cell surface) was transferred to a new tube andthe pellet (biotinylated bacteria) resuspended in 200 Al TBS.Biotinylated protein molecular weight standards for SDS-PAGE wereprepared in a similar fashion.(c). Erythrocyte-bacteria Binding ExperimentThis experiment entailed electrophoresing non-labelled erythrocyteghosts or bacteria whole cell lysates on a 12.5 % SDS-polyacrylamide gelthen western blotting the proteins onto nitrocellulose. Thenitrocellulose was rinsed in TBS, blocked for 1 h with TBS / 0.05 %tween / 3 % BSA, washed twice in TBS-tween and incubated for 3 h withsamples of either biotinylated erythrocyte ghosts (50 Al in 5 ml TBS),biotinylated bacteria (50 Al / 5 ml) or biotinylated bacterial vortexsupernatants (50 Al / 5 ml). Table 7 details this procedure.104Table 7. Erythrocyte ghost-bacteria binding experimentgel 1 gel 2lane sample blotted w/ lane sample blotted a123standardpigdogbiotinylatedF41 vortex sn123standardpigdogbiotinylatedF41 bacteria456standardpigdogbiotinylatedJM101 vortexsn456standardpigdogbiotinylatedJM101bacteria78910standardpigdogbiotinylatedJM101(pETE1)vortex sn78910standardpigdogbiotinylatedJM101(pETE1)bacteriagel 3 gel 4lane sample blotted a lane sample blotted a1234standardF41snJM101snJM(pl)snbiotinylateddog ghosts1234standardF41snJM101snJM(pl)snbiotinylatedpig ghosts6789standardF41ptJM1OlptJM(p1)ptbiotinylateddog ghosts6789standardF41ptJM1OlptJM(pl)ptbiotinylatedpig ghostssn = supernatant; pt = resuspended bacterial cell pellet after vortexingThe nitrocellulose filters were then washed twice in TBS-tween andincubated for 2 h with streptavidin-conjugated alkaline phosphatase(Pierce, 5 gl in 5 ml TBS). The filters were washed again and incubatedin 0.1 M tris pH 9.6 containing X-phosphate (50 µg/ml) and nitrobluetetrazolium (10 µg/ml) for 5-10 min for the colour to develop. Blotswere rinsed in water and air-dried.1054.2.3.3. Glycophorin ExperimentsThe human erythrocyte receptor for the F41 agglutinin is known tobe glycophorin, a major red blood cell membrane glycoprotein. Todetermine whether the agglutinin encoded by pETE1 also recognizedglycophorin, two experiments were performed. The first experiment was aglycophorin inhibition assay involving incubating 50 ml bacteria(12.5x109 cells/ml) with 50 pl glycophorin (5.0 mg/ml) (prepared by J.Cavanagh using the method of Marchesi and Andrews, 140) for 10 min at23 °C, then performing a standard hemagglutination microtitre assay withserial dilutions of the bacteria against washed 1 % human or pigerythrocytes in PBS. The second experiment involved incubatingerythrocytes (1 ml 10 % hematocrit) with Vibrio cholerae neuraminidase(Sigma, 1 U), for 2 h at 37 °C, followed by washing of the cells andmicrotitre hemagglutination with serial dilutions of bacteria.Neuraminidase is an enzyme that removes the terminal sialic acid residuefrom oligosaccharides present on erythrocyte surface glycoproteins(141).4.2.4. Maxicell Analysis of Proteins Encoded by pETE1Escherichia coli CSR603 was obtained from the American TypeCulture Collection. This bacterium has mutations in its recAI and uvrA6genes which halt the functions of the gene products, proteins that allowrecombination and repair DNA that has been damaged by UV light (138).When irradiated with UV therefore, the bacterial DNA is extensivelydegraded and replication ceases. If the bacterium contains a plasmid,106then that too will be degraded, but as there are multiple copies in thecell any that escaped a UV hit will continue to replicate with plasmidlevels increasing tenfold by 6 h after illumination. When DNA damage isat a maximum, the non-dividing cell (maxicell) will contain mostlyplasmid encoded proteins, as the cells transcription/translation systemsare still functional. If 35S-methionine is introduced as the onlysulphur source, then the plasmid proteins will be radioactivelylabelled during synthesis (Sancar and Rupp (142)).Plasmids pETE1 and pBR322 were transformed into CaC1 2 competentCSR603 (see Section 2.2.6.). The plasmid-containing bacteria were thengrown overnight in 5 ml LB amp. A 500 gl aliquot of the culture wasadded to 25 ml M9 minimal media containing ampicillin and the cells weregrown at 37 °C to exponential phase. This took approximately 12 h, asthe recAl mutation causes the bacteria to grow very slowly (117). Thecells were then exposed to UV light (a germicidal lamp held 30 cm fromthe cultures) for 30 s and incubation continued for 1 h at 37 °C.Cycloserine (200 pg/ml) was added to the cultures and incubationcontinued for 16 h. After this time the cells were centrifuged (6,000rpm, 15 min) and washed twice with M9 salts. The bacteria wereresuspended in 50 gl M9 media minus sulphate and 35S-methionine (10mCi/ml) was added to 5 gCl/ml. The cells were incubated for 2 h at 37 °Cthen washed twice with M9 salts and resuspended in SDS-PAGE loadingbuffer (40 gl). Samples (15 gl) were run on two 12.5 % gels, one ofwhich was stained with Coomassie blue, the other of which was blottedonto nitrocellulose for 2 h at 1 A. The nitrocellulose was rinsed inwater, air dried and exposed to X-ray film for as long as 4 days.1074.2.5. N-terminal Sequencing of HemagglutininGel electrophoresis of crude membrane fractions of JM101(pETE1)and JM101(pPL7) prepared by sonication showed the presence of a distinctband at m 25 kDa compared to JM101 (Section 4.3.1.). These samples wereblotted to PVDF (Immobilon) membrane and lightly stained with Coomassieblue. When the membrane was dry, the band was cut from the membrane andsequenced by automated Edman degradation at the University of Victoriaprotein sequencing facility.1084.3. Results4.3.1. Membrane Association of HemagglutininTemperature/vortex experiments on E.coli strains harbouring pETE1showed that the agglutination effect was strongly associated with thecell surface and could not easily be removed. Figure 40 shows thehemagglutination patterns of some supernatants from the disruptionexperiments, along with SDS-PAGE results from these samples. A widevariety of disruption experiments were attempted and the samples shownin Figure 40 represent the best results achieved. The supernatantsamples even when concentrated showed very little hemagglutination, butin the samples that produced a small amount of activity, SDS-PAGE showedthe existence of a unique band of approximately 25,000 Da. The pETE1plasmid was introduced into the different E.coli strains described inSection 4.2.1. and always conferred on the bacteria the agglutinatingability, but was always strongly associated with the cell surface. Inthe centrifuged pellets of cells containing pETE1 after these disruptiontreatments there remained a strong agglutinating presence which was veryresistant to inactivation by the heat and mechanical forces used. Infact, it was found that the bacteria containing pETE1 were moreresistant to these forces than the original bacteria, either JM101 orother strains. After identical treatment, the control cells wouldexperience some degree of lysis as judged by the turbidity ofcentrifugation supernatants, whereas plasmid-containing cells wouldremain intact. This stabilization effect was also seen in sonicationexperiments.109hemagglutination profileFigure 40. SDS-PAGE and hemagglutination profiles of samples fromtemperature/vortex experiments(see Table 8, next page, for hemagglutination strength index)lane 3,6 = molecular weight standardslane 1 = B/R concentrated supernatant after experiment 6 (Section4.2.2.1)lane 2 = B/R(pETE1) conc. sn. after expt. 6lane 4 = B/R conc. sn. after expt. 4lane 5 = B/R(pETE1) conc. sn. after expt. 4lane 7 = B/R conc. sn. after expt. 10lane 8 = B/R(phIE1) conc. sn. after expt. 10110Table 8. Agglutination scale (used throughout this document)+ + + + = agglutination to approximately 6.0 x107 bacteria/ml (or anerythrocyte:bacteria ratio of 2:1).+ + + = agglutination to approximately 2.1 x108 bacteria/ml+ + = agglutination to approximately 1.0 x10 9 bacteria/ml+ = agglutination to approximately 4.0 x10 9 Bacteria/ml(+) = trace partial agglutination- = no visible agglutinationMicrotitre Hemagglutination assays carried out as described in section2.2.10.1Experiments with detergents met with little success. Thedetergents removed considerable amounts of material from the cell wall(SDS-PAGE Fig. 41), but there were no discernible differences betweenthe control bacteria and the plasmid-containing cells. After dialysisand concentration, no supernatant sample contained agglutinatingproperties. Osmotic shock of the bacterial cells gave a largenumber of protein bands on SDS-PAGE, but again no differences were seenbetween the control and the test bacteria (Figure 42).111Figure 41. SDS-PAGE of detergent-solublized proteinslane 1 = molecular weight standardslane 2 = B/R supernatant after octyl glucoside treatmentlane 3 = B/R(pETE1) supernatant after octyl glucoside treatmentlane 4 = B/R supernatant after SDS treatmentlane 5 = B/R(pETE1) supernatant after SDS treatmentlane 6 = B/R supernatant after CHAPS treatmentlane 7 = B/R(pETE1) supernatant after CHAPS treatmentlane 8 = B/R after deoxycholate treatmentlane 9 = B/R(pETE1) after deoxycholate treatment112Figure 42. SDS-PAGE of proteins released by osmotic shockBacteria were subjected to osmotic shock as described in Section4.2.2.3.lane 1 = B/R supernatantlane 2 = B/R(pETE1) supernatantlane 3 = B/R concentrated supernatantlane 4 = B/R(pETE1) concentrated supernatant113As the evidence suggested that the agglutinating factor was verytightly bound in the bacterial membrane, sonic disruption of the cellswas tried to obtain crude membrane preparations. The crude membranepreparations from JM101(pETE1) still showed very strong red blood cellaggregation and SDS-PAGE of these membranes contained a distinct bandover the cloning vehicle that ran at around 25,000 Da. The gel is shownin Figure 43. Subclones and controls were sonicated and the membranepellets and supernatants tested for HA and electrophoresed. The resultsshowed that hemagglutinating activity and the 25 kDa protein band alwayscoexisted. Membranes prepared from subclones pPL2, 3 and 4, containingthe 1.3, 1.0 and 2.3 kb HincII fragments respectively, were inactive inHA microtitre and did not contain an obvious band (Fig. 43). Bacteriacontaining pBR322 were also negative. The plasmid pPL7 however behavedsimilarly to pETE1 with both the agglutination and the band present. Asubclone of pPL7 with a 395 by NcoI deletion in the open reading frameof plprotl, called pPL7N, showed no band and no agglutination (Fig. 44).This result locates the hemagglutination effect to that open readingframe.114Figure 43. SDS-PAGE and hemagglutination profile of crude membranefragments prepared by sonication.lanes 2-7, bacterial supernatants after sonication and centrifugationlanes 9-14, crude membrane pellets after sonication and centrifugationlanes 1,8 and 15, protein molecular weight standards2 and 9 = JM1013 and 10 = JM101(pETE1)4 and 11 = JM101(pPL7)5 and 12 = JM101(pPL4)6 and 13 = JM101(pPL3)7 and 14 = JM101(pPL2)115Figure 44. SDS-PAGE and hemagglutination profiles of crude membranepreperations harbouring pPL7N control plasmid.lane 1 = molecular weight standardslane 2 = JM101 crude membrane pelletlane 3 = JM101(pPL7) crude membrane pelletlane 4 = JM101(pPL7N) crude membrane pellet1164.3.2. Comparison with F414.3.2.1. Agglutination Strength and SpecificityThe bacteria containing pETE1 agglutinated human red blood cellsto the same degree in a microtitre assay as the parent, F41, but withoutthe requirement of specialized media. The agglutination was of the samemagnitude whether grown on plates or liquid media, and was unchanged inthe presence of 10 mM mannose. Unlike F41 or other mannose resistanthemagglutinins characterized from enterotoxigenic E.coli strains,agglutinating activity was not destroyed by heating at 60 °C for thirtyminutes, or even by transiently boiling the bacteria. For this reasonthe hemagglutinin was named Heat Resistant Agglutinin 1 (HRA1).Bacteria harbouring pETE1 did not appear to be fimbriated (see Section5.3.6.). The hemagglutination was not inhibited by glycophorin, the F41receptor, and neuraminidase treatment of erythrocytes prior tohemagglutination by JM101(pETE1) had no effect. For furthercharacterization, the agglutinating activity of the clone against arange of species erythrocytes was investigated. The speciesagglutination profile is shown in Table 9.These results are interesting for several reasons -(1). The cloning vehicle, JM101, agglutinates certain species'erythrocytes very strongly, indicating the presence of some type ofsurface adhesin. This observation may be explained by later resultsindicating that this strain expresses the mannose sensitive type 1adhesin, though the type 1 fimbriae are known to agglutinateerythrocytes from a broader range of species (12).117Table 9. Agglutination profile for washed erythrocytes of thespecies indicated; T = 21 °C; suspending medium was PBS, no mannose.Experiment performed as Section of Hemagglutination lF41 Adhesin F41 cell JM101-JM101(pETE1)Goat +++ +++ - +Rat + + + + + + + + + + + + + +Mouse - + + + + +Guinea Pig + + + + + + + + + +Cow - + - +Rabbit - + + + + +Chicken + + + + + + - + +Cat ++++ ++++ - +Pig ++++ ++++ - +++ Dog - + - + + +Human +++ ++++ - +++Agglutination scale as described in Table 8.(2). F41 cells express more than just the F41 adhesin. Thisbacterium is known to express the K99 adhesin also. Unfortunately asample of K99 +F41 E.coli was not available at the time of the test, soit cannot be determined which species' agglutination is caused by thisagent. Studies in the literature (65) have shown that K99 can causeagglutination of horse, sheep and cow erythrocytes, but not guinea pigor human. It can be concluded therefore that K99 is not responsible forthe agglutinating behavior of JM101(pETE1) because of the stronghemagglutination of human red cells mediated by this strain.118(3). Dog erythrocytes are agglutinated weakly by F41 bacteria, butnot at all by the purified adhesin or by JM101. This species howevershows strong aggregative behavior with JM101(pETE1), which couldindicate the amplification of a minor agglutinin on F41 by cloning intoa plasmid vector.(4). The cloned agglutinin does not appear to be a non-specific"sticky" protein due to the large differences in strength of adherencebetween species. Erythrocyte receptor resultsExperiments with erythrocyte ghosts failed to unequivocallyelucidate either the receptor on the red cell or the bacterial proteinresponsible for the adherence. Of the set of experiments where the redcell membranes were electrophoresed and blotted against the biotinylatedbacteria supernatants, only F41 showed binding to the immobilizedmembrane proteins (Fig 45). The binding was not specific to a singleband, but several bands were obviously more favoured than others; forexample, the band at approximately 55 kDa corresponding to glycophorin,the known receptor for F41. Some differences were seen between thebinding patterns of F41 to pig and dog ghosts. No binding was observedfor JM101 as expected, and no binding was seen when JM101(pETE1)supernatant was incubated with the membrane. This is not surprising asvery little, if any, of the agglutinating agent is removed from the cellsurface by vortexing. Unfortunately, when the same experiment wasattempted with the whole bacteria, there was no binding from any of thestrains indicating that the bacteria were too large to be tightly boundto the membrane proteins in the nitrocellulose.119Experiments in which the bacterial vortex supernatant proteinswere blotted onto the membrane and then incubated with biotinylatederythrocyte ghosts showed that both pig and dog ghosts recognized a29,000 Da protein from the F41 bacteria supernatant (Figure 46). Thiswould correspond to the F41 adhesin and is interesting since thisadhesin when purified shows strong affinity for pig erythrocytes but notdog. The experiments with whole cells however showed many non-specificdiffuse bands for all samples which made interpretation impossible.This is unfortunate as the adhesin present on JM101 is stronglyassociated with the cell surface and could not be purified for cell-freebinding experiments. Both JM101 and JM101(pETE1) supernatants showcommon ghost-recognized bands in Figure 46, which may be due tonon-specific interactions, and in the bacterial supernatant blot againstdog ghosts, a faint 25 kDa band is visible in the JM101(pETE1) lane.These results are very inconclusive. The receptor for theJM101(pETE1) agglutinin on the erythrocyte was not determined, and theagglutinin itself was not identified with certainty, the onlypossibility being the weak 25 kDa band seen against a non-specificbackground. The weak band seen may have occurred because theinteraction was not of sufficient strength to withstand thenitrocellulose membrane washing procedure or because denaturing theproteins on SDS-PAGE destroyed their specificity. Evidence for theformer speculation was found during the viscometric analysis (Section5.2.2.), and evidence for the latter is provided by the weak interactionseen between blotted F41 whole cell digests and biotinylated ghosts.Also, the biotinylation of amine groups on the bacterial proteins and on120Figure 45. Nitrocellulose immobilized erythrocyte ghost proteinsblotted against biotinylated bacterial proteins.a) Erythrocyte ghosts blotted against biotinylated F41 supernatants.b) Erythrocyte ghosts blotted against biotinylated 24101(pETE1)supernatants. All other samples were identical to this blot.lane 1 = biotinylated molecular weight standardslane 2 = pig erythrocyte ghostslane 3 = dog erythrocyte ghosts121Figure 46. Nitrocellulose immobilized bacterial proteins blottedagainst biotinylated erythrocyte ghostsExperiment performed as described in Section pig erythrocyte ghosts, b) dog erythrocyte ghostslane 1 = molecular weight standardslane 2 = F41 supernatantlane 3 = JM101(pETE1) supernatantlane 4 = JM101 supernatant122the erythrocytes may have interfered with the binding reaction (the K99adhesin erythrocyte binding site incorporates lysine residues and HRA1contains a similar sequence; see Section 3.4.). Maxicell ResultsMaxicell results are shown in Figure 47. The autoradiograph showsproduction of a radiolabelled protein of 25 kDa by JM101(pETE1). Theg-lactamase encoded by the ampicillin resistance gene of pBR322is probably the band visible at approximately 30 kDa. There is a highbackground suggesting that the UV exposure time was not optimal, butlengthening the exposure time led to a drastic loss of labelling. Nobands are seen in the JM101(pBR322) control lane indicating thatlabelling did not occur efficiently in this sample for unknown reasons.The Coomassie blue stain of these bacteria showed enhancement ofproduction of these proteins as expected.4.3.3. N -terminal SequenceThe N-terminal sequence of the 25k band visible in the SDS-PAGE ofthe crude membrane preparations of JM101(pETE1) and JM101(pPL7) wasdetermined. The band sent for analysis was impure, as shown on SDS-PAGE(Fig 43). There is a faint contaminating band in the JM101 membranefraction lane running at approximately the same molecular weight.The N-terminal sequence was determined to be -H2N-Asp-Glu-Ala-Gly-Thr-Phe-Phe-Tyr-Arg(?)-Thr-Gly-Gly(?)-AlaQuestion marks indicate inconclusive sequence data.123Figure 47. Coomassie stain, (a), and autoradiograph, (b), of SDS-PAGEof maxicells.lane 1 = molecular weight standardlane 2 = CSR603(pETE1)lane 3 = CSR603(peR322)124The predicted N-terminal sequence of the mature plprot1 protein(that is, after cleavage of the highest-scoring predicted signalpeptide) from the nucleotide sequence is show below, aligned with thedetermined sequence.PREDICT: Asp-Glu-Ser-Lys-Thr-Gly-Phe-Tyr-Val-Thr-Gly-Lys-AlaDETERM: Asp-Glu-Ala-Gly-Thr-Phe-Phe-Tyr-Arg-Thr-Gly-Gly-AlaThe predicted sequence shows 62 % sequence identity with thedetermined sequence which indicates strongly that plprot1 is present asa 25,000 Da protein in the crude membrane preparation and is thecausative agent of agglutination. The sequence information derived fromthe N-terminal analysis is not identical to the sequence predicted fromthe determined nucleotide sequence however, and there are severalpossible reasons for this. Firstly, there could be errors in thenucleotide sequence data. The nucleotide sequence information in thisregion is very clear and was determined independently for both strands(see Appendix 4), which eliminates the possibility of codon readingerrors. Secondly, the inconsistencies in the sequences could arise froma contaminating protein present in the sample sent for analysis. Thisappears to be a likely explanation as SDS-PAGE of JM101 control crudemembrane preparations (p.115) shows the presence of a faint band withapproximately the same molecular weight. The analysts report statedthat the sequence data contained contaminating signals. Lastly, thepronounced 25 kDa band present in SDS-PAGE of crude membranepreparations of JM101(pETE1) and JM101(pPL7) could be a differentprotein, although evidence that this band is not pronounced in membranepreparations of JM101(pPL7N) (which contains a deletion in the plprotl125ORF) makes this an unlikely possibility.Using the PCgene program ALIGN (which utilizes the method of Myersand Miller (143)), possible homology with other adhesin N-terminalsequences was investigated. The results are shown in figure 48.•plprotl DESKTGI^IFYVTGKAGALVMSLADNFA4^WTTGDFNGSFNMNGAIAADVYKGCS31A^GTTGDFNGSFDMNGTITADATKK88^WMTGDFNGSVDIGGSITADGYGF41^ADWTEGQPGDIL IGGEITFTS V• = Functional similarityBy inserting one space into the plprotl sequence, six identicalresidues can be aligned with the NFA4 N-terminal sequence (27.3 %), withone residue showing a functional similarity.•plprotl DESKTGLIFIYVTGK AGA S VMSL A DNFA4 WTTGDFNGSFNMNGA I A A D VYKG■■■■I •••■••ICS31A G TTGDFNGSFDMNG T I T A D A T KK88 W MTGDFNGSVDpismonm■••I G G S I T A D G Y GF41 A D W TEGQPGDIL I G G E I T XPSV.111■1•■•■•• = Functional similarityFigure 48. N-terminal sequence homology of plprotlInserting two gaps into plprotl and one into NFA4, seven residuescan be aligned with NFA4 (32 %) with one showing functional similarity.This data suggests that plprot1 could be distantly related to thenon-fimbrial adhesin 4 family of adhesins, although the degree ofsimilarity is not high.1264.4. DiscussionThis section determines the pETE1-coded agglutinin to be verystrongly associated with the bacterial membrane. N-terminal amino acidsequencing of a distinct 25 kDa band present in SDS-PAGE of crudemembrane preparations of JM101(pETE1) showed that this band probablycontained the protein product plprotl, coded on ORF1 of the pETE1plasmid. This protein was shown by deletion subcloning (Section 3.2.4.)to be responsible for the hemagglutination effect. The agglutinin isvery different from the F41 adhesin and most other adhesins in thatheating to 60 °C does not inactivate the protein. For this reason theagglutinin was named Heat Resistant Agglutinin 1 (HRA1). Heat ResistantAgglutinin 1 showed a small degree of N-terminal sequence homology withthe NFA4 family of adhesins.The species hemagglutination profile showed that HRA1-mediatedagglutination is somewhat species specific, but an attempt to identifythe erythrocyte receptor gave equivocal results.Subclone pPL4 was very much more resistant to cellular disruptionby sonication than the other bacteria, requiring sonication timesthree-fold higher than other clones to achieve the same degree ofdisruption, measured visually as a clearing of the suspension. Thisstabilization effect is likely due to the gene products of the 2.3 kbHincII fragment that this subclone contains. These products (detailedin Section 3.3.3.) include the plprot5 protein which is highlyhomologous to the major E.coli lipoprotein, a protein that is known tostabilize the outer membrane (137). The strain containing this plasmid127may express more of the plprot5 protein than pETE1, since the fragmentis in a high copy number pTZ vector. The expression of amembrane-stabilizing protein could also explain the greater resistanceto lysis under severe physical conditions (temperature/vortex) thatJM101(pETE1) exhibited compared to JM101.1285. Physicochemical Characterization of the Agglutinin5.1. BackgroundTwo non-traditional physical chemical techniques, namely aqueoustwo-phase partitioning and low-shear viscometry, were employed for studyof the recombinant strains expressing the agglutinin. Both systems arevery sensitive to cell surface differences between closely relatedbacteria, and can give qualitative and quantitative information on thedegree of difference. Cell electrophoresis was used to calculate theelectrophoretic mobility of the clone, and phase-contrast microscopy wasemployed to visualize the adherence event. Electron microscopy was usedto search for surface appendages on both F41 and the recombinantorganism.5.2. Materials and Methods5.2.1. Aqueous Two-Phase PartitioningThe principles of partitioning cells in aqueous two-phase polymersystems are explained in Section 1.6. In this work, a variety ofpolymer systems with varying properties were utilized to assess cellsurface differences between the bacterial strains.1295.2.1.1. Preparation of the Phase SystemsPhase systems were made up by mixing appropriate weights of thefollowing stock solutions.(a). dextran T500 ( M = 461,700, Pharmacia, Uppsala, Sweden, lotno. FD 160-27), 20 g in 100 g PBS. Dextran stock solutionconcentrations were determined polarimetrically (108).(b). dextran T40 ( M = 40,000 Pharmacia, lot no. 40601), 30 g in100 g PBS.(c). poly(ethylene glycol) (PEG) 8000 (MMwa 8,000, Union Carbide,New York, NY, lot no. B-688-0232-2), 30 g in 100 g PBS. PEG stocksolution concentrations were measured by refractive index measurements(108).(d). ficoll 400 (M = 400,000 (by light scattering), Pharmacia,lot no. IL-33505), 40 % w/w.The phase systems used in this work were then prepared.(a). [(5,4)5] : 5 % w/w dextran T500; 4 % w/w PEG 8000; 150 mMNaCl, 6.84 mM Na 2 HP0 4 , 3.16 mM NaH 2 PO 4, pH 7.20(b). [(5,4)I] : 5% w/w dextran T500; 4 % w/w PEG 8000; 75.2 mMNa 2 HP0 4 , 34.8 mM NaH2 PO 4, pH 7.20(c). [(7.4,4.7)55] : 7.4 % w/w dextran T40; 4.7 % w/w PEG 8000;37.6 mM Na 2 HP0 4 , 17.4 mM NaH2 PO 4, pH 7.20(d). [(10,7.5)5] : 10 % w/w dextran T500; 7.5 % w/w PEG 8000; 150mM NaC1, 6.84 mM Na 2 HP0 4 , 3.16 mM NaH2 PO 4, pH 7.20(e). [(6,10)5] : 6 % w/w dextran T500; 10 % Ficoll 400; 150 mMNaCl, 6.84 mM Na 2 HP0 4 , 3.16 mM NaH2 PO 4, pH 7.20130(f). [(7.4,4.7)5] : 7.4 % w/w dextran T40; 4.7 % w/w PEG 8000;150 mM NaC1, 6.84 mM Na 2 HP0 4 , 3.16 mM NaH2 PO , pH 7.20(g). [(5,3.5)1] : 5 % w/w dextran T500; 3.5 % w/w PEG 8000; 75.2mM Na 2 HP0 4 , 34.8 mM NaH 2PO , pH 7.20(h). [(6,10)I] : 6% w/w dextran T500; 10 % w/w Ficoll 400; 75.2mM Na 2 HPO , 34.8 mM NaH 2PO , pH 7.204 ^4(i). [(7.5,10)55] : 7.5 % w/w dextran T500; 10 % w/w Ficoll 400;37.5 mM Na 2 HP0 4 , 17.4 mM NaH2 PO , pH 7.20[(7.5,10)5] : 7.5 % w/w dextran T500; 10 % w/w Ficoll 400;150 mM NaC1, 6.84 mM Na 2 HPO , 3.16 mM NaH2 PO , pH 7.204 ^ Radiolabelling the BacteriaThe bacterial strains used were E.coli F41, JM101, B/R,JM101(pBR322),JM101(pETE1), JM101(pPL2), JM101(pPL3), JM101(pPL4),JM101(pPL7), B/R(pETE1) and B/R(pPL7). Another control subclone wasconstructed by excising a 17 base pair DNA fragment from pBR322,recircularization of the plasmid and electroporation (see Table 2Section 3.2.5) back into JM101. This was designed to inactivate thetetracycline resistance gene to create a plasmid that differed frompETE1 only in the foreign DNA content. The clone was selected by itsability to grow on LB amp and inability to grow on LB tetracyclineplates.The bacteria were all radiolabelled in the same way. Carbon-14labelled amino acid mix (50 gl, 10 pCi/ml, Amersham) was spread on thecentral portion of minca or LB plates containing the appropriateantibiotics and inducers. The radiolabel was allowed to soak into theplates for 5 min, then 25 gl of a -70 °C DMSO stock bacteria suspension131was spread in the same area. The bacteria were then grown at 37 °Covernight and were harvested by swabbing into PBS and washed thrice.Bacteria concentration was estimated by centrifugation inmicrohematocrit tubes (see Section The Partitioning ExperimentThe phase systems were made up in 50 ml polyallomer tubes andallowed to equilibrate for 1 h at room temperature. The top phase wascarefully drawn off into a separate tube leaving approximately 1 ml nearthe interface. The lower phase was drawn off by puncturing the tubewith a needle and collecting. Lower phase (1 ml) was aliquotted into a5 ml polycarbonate test tube and 1.5 ml of upper phase (the load mix)placed in a separate 5 ml tube. Bacteria (200 pl, 2x109 cells/ml) orerythrocytes (2 % hematocrit, 3x washed, in PBS) were added to the loadmix, mixed and 1 ml of the suspension was then added to the lower phase.The experiment was always done in triplicate. The tubes were capped andinverted rapidly 20 times, the caps quickly removed to avoid drippingand the systems were left to equilibrate for 20 min at room temperature.After this time, 200 Al of each phase was sampled from the upper, lowerand load mixes. For 14C-labelled bacteria, the sample was transferredinto a scintillation vial, to which was added 7 ml scintillation fluid(Atomlight, Du Pont). The vials were vortexed and counted in a PhillipsFW3400 scintillation counter. For erythrocytes, the samples were addedto 12 ml diluent (Hematall azide-free isotonic, Fisher) and the cellswere counted by impedance (coulter) counting with an Electrozone cellcounter. The partition ratio was expressed as the quantity of cellspresent in the upper or lower phase as a percentage of the total cells132added to the system. Bacteria-Erythrocyte Binding Partition ExperimentsWhen optimum phase systems had been found (i.e. those that showedthe greatest difference between the bacteria and the red cells) abinding experiment was performed. Red blood cells (350 gl, 2hematocrit) were preincubated with 350 ill radiolabelled bacteria (2x10 9cells/ml) for 30 min at 37 °C. The mixture (200 gl) was then added tothe phase system load mix as before and the partition carried out. Whenthe phase systems had reached equilibrium, two 500 Al aliquots weretaken from each phase, one of which was counted for red cells in thecell counter, and the other of which was counted for bacteria in thescintillation counter. Controls were performed with erythrocytes aloneand bacteria alone under the same conditions.5.2.2. Viscometry of bacterial strainsThe principles and rational of the viscometric assay for studyingbacterial adhesion to erythrocytes are explained in Section 1.5. Theadhesion of bacteria to human colon adenocarcinoma 201 (colo 201) cellswas also examined by this technique. Bacterial strainsThe bacteria used were Escherichia coli of the following strains.(a). JM101 and recombinant clones JM101(pETE1) and JM101(pPL7)(b). F41133(c). K12W, R1316, B/R and K802, strains obtained from R. Graham thatwere scanned in the viscometer for agglutination in the absence ofmannose to find a plasmid host with no detectable agglutinationbackground.(d). recombinant clones B/R(pETE1) and B/R(pPL7)The bacteria were grown overnight on solid media, either minca orLB with the appropriate antibiotics and harvested into HBSS-hepes (Hanksbalanced salt solution, 5 mM KC1, 0.3 mM KH 2PO4 , 138 mM NaCl, 4 mMNaHCO3 and 0.3 mM Na 2 HP0 4, with 5 mM hepes (4-(2-hydroxethyl)-1-piperazineethanesulphonic acid); pH 7.2, 278 mOs). Growth of Human Colon Adenocarcinoma Cells.Human colon adenocarcinoma cells, line 201 (145), were obtainedfrom the ATCC. The cells were grown in Dulbecco's Modified Eagle medium(DME, Gibco, for composition see Appendix) with 20 % fetal bovine serum(Flow Laboratories at 37 °C in a 10 % CO2 atmosphere. The cells wereharvested by centrifugation and suspended in HBSS-hepes to a density of4.4x106 cells/ml. Viscometric Hemagglutination AssayThe viscometry was performed in a Contraves LS-2 couetteviscometer (Contraves, airich, Switzerland). Erythrocyte suspensionswere prepared by centrifugation of whole blood to remove plasma,followed by washing of the red cells 3x in HBSS-hepes. The erythrocyteswere used at a hematocrit of 47 %.The bacteria studied were washed 3x in HBSS-hepes and suspended in134HBSS-hepes to approximately 10x10 10 cells/ml (45% "bugcrit"). Red cells(800 pl) were put into the well of the viscometer and allowed toequilibrate at 37 °C for 20 min then sheared for one minute (for shearrates, see Results section) before addition of 100 pl bacteria. Acontrol was carried out by adding instead 100 pl of buffer. The reactionwas allowed to proceed for 5-20 min and the shear stress increase wasplotted as a function of time with a Fisher Recordall 500 chartrecorder.5.2.3. Cell ElectrophoresisCell electrophoresis was carried out to estimate the charge onE.coli F41, JM101 and JM101(pETE1). The electrophoretic mobility ofhuman erythrocytes was also determined as a control. Cells weresuspended in 0.15 M NaC1 buffered to pH 7.3 with NaHCO 3 in thecylindrical chamber of a Rank Mark I (Cambridge, U.K.) cellelectrophoresis apparatus. The cells were then viewed at 200xmagnification with a horizontal microscope lens equipped with a waterimmersion objective focused at the stationary layer (146). A diagramof this apparatus is shown in Figure 49. A voltage (40V, electric field4V/cm) was applied across the sample and movement of the cells against afixed grid background of known dimensions was observed and timed with ahand timer to ± 0.5 s. Mobility measurements were made (at least 20 foreach sample) and electrophoretic mobilities (velocity/electric field)determined for each cell using the formula-135p = 1.D.Let Vwhere t = averaged time for one transit (across calibration grid)D = size of grid division (in pm)Le = electrical distance between electrodes (cm), determined asdescribed (146).V = voltagem = electrophoretic mobility (in pm.sec 1 V icm)5.2.4. Microscopy5.2.4.1. Electron MicroscopyEscherichia coli F41, JM101 and JM101(pETE1) were stained withuranyl acetate and viewed with a Phillips electron microscope atmagnifications of up to 40,000x for evidence of piliation. Optical PhotomicroscopyPhase contrast photomicroscopy was used to study the aggregationof human erythrocyte ghosts and cultured colo 201 cells with E.coli F41,JM101 and JM101(pETE1) in the presence and absence of mannose. Thecells were spotted onto glass slides and viewed without coverslips usinga Carl Zeiss 67547 photomicroscope with a 400x objective.136Figure 49. Diagrammatic representation of cell electrophoresisapparatus.a) Front view - movement of cells is measured at the stationary layerin the region of the optical flat.b) Side view showing microscope lens.1375.3. Results5.3.1. Influence of pETE1 on Bacterial PartitionThe results from the bacterial partition in two-phase polymersystems are shown graphically in Figures 50-56. As the uncertainty inthe averaged partition of cells in these experiments (expressed as apercentage) was usually very low, of the order of 1-2 %, error bars havebeen omitted from the figures unless determined to be unusually high.There are many notable features of these experiments -(1). F41 grown on minca shows a very different partition in the[(10,7.5)15 system than F41 grown on LB. This shows that this system isa very sensitive indicator of the levels of F41 antigen expressioninduced by the different media (Figure 50). The sensitivity of thissystem can be appreciated by reference to Western blot results for F41antigen production on the different media in Section JM101(pETE1) partitions very differently from both JM101 and F41in several PEG\dextran systems indicating that the cell surface ofJM101(pETE1) is very different from that of JM101 (Figure 51). Thisdifference could only have arisen from the introduction of the pETE1plasmid.(3). JM101(pBR322) and JM101 harbouring the constructed controlplasmid pBR322-BanII deletion partition in a very similar manner toJM101 in the phase systems (Figure 52). This provides good evidence1380.° 3o.1008060c*:t4020CL 0200 406080A 13JFL FM JM P1^FL FM JM P1Bacterial StrainFigure 50. Bacterial partition in A) [(10,7.5)5] and B) [(6,10)5]phase systems.FL = F41 grown on LB; FM = F41 grown on mincaJM = JM101; P1 = JM101(pETE1)Cells are partitioned between the upper and lower phases and theinterface.139806045CLg- 400 20tcoa 0a)w 02CLa) 406080F J P1 F J P1^F J P1bacterial strainF J P1A 13i 2_1Figure 51. Partition of JM101(pETE1) in variousPEG/dextran systems.A = [(5,4)5]^ F = F41 (minca grown)B = [(5,4)I] J = JM101C = [(7.4,4.7)5]^ P1 = JM101(pETE1)D = [(7.4,4.7)55]1408070t 5 60a.o.3 50400H2302010-2 01020J P7 P1 PN P2 P3 P4 BR BAbacterial strainFigure 52. Partition of various pETE1 subclones in [(7.4,4.7)55]J = JM101^ P3 = JM101(pPL3)P7 = JM101(pPL7)^ P4 = JM101(pPL4)P1 = JM101(pETE1) BR = JM101(pBR322)PN = JM101(pPL7N)^BA = JM101(pBR322-BanII deletion)P2 = JM101(pPL2)141that the partition difference seen for JM101(pETE1) is due to theproducts of the F41 DNA insert in that plasmid.(4). E.coli JM101 harbouring pPL7 behaves in a very similar manner toJM101(pETE1) in all systems studied (Figure 53). Often the effect ofpPL7 is greater than that of pETE1. This indicates that the partitiondifference seen with pETE1 originates in the 1.3 + 1.0 kb HincIIfragments, and may also show a higher level of expression caused by thevector pTZ19R, which has a higher copy number than pBR322.(5). The subclones pPL2, pPL3 and pPL4 containing the single HincIIfragments partition very similarly to JM101 in [(7.4,4.7)55] (Figure52). This, together with previous points, is good evidence for thelocalization of the partition difference to the open reading framespanning the HincII site between the 1.3 and 1.0 fragments. This datais consistent with hemagglutination data and implicates theagglutinating agent as the cause of the partition difference seen. Inthe subclone pPL7N, which contains an NcoII deletion in the open readingframe of plprotl, the partition difference is abolished (as is thehemagglutination). This further supports identification of the proteinHRA1 as the agent of the partition difference.(6). In dextran\Ficoll systems, the F41 bacteria partition is highlysensitive to phosphate concentration, partitioning strongly in the upper(dextran rich) phase at low phosphate concentrations (system[(7.5,10)5], Figure 54). When introduced to a system containing higherconcentrations of phosphate, [(7.5,10)55], the bacteria partitionstrongly into the lower layer. The JM101 and JM101(pETE1) bacteria showpreference for the lower layer in both systems.14280AI^I^I I^r^I^IB60Zi50. 40m202015 4006080P1 P7 J F P1 P7 J Fbacterial strainFigure 53. Partition of JM101(pPL7) in A) [(5,4)5] and B) [(7.4,4.7)55]J = JM101F = F41 (minca grown)P1 = JM101(pETE1)P7 = JM101(pPL7)1431008060a4020acN 0A )2J20040 10601111 1^1^[^1F JP1 F JP1 F JP1^F JP1bacterial strainF JP1 F JP180Figure 54. Bacterial partition in charge-sensitive systems anddextran/ficoll systems.A = [(5,4)51 (non charge-sensitive)B = [(5.4)I1C = [(5,3.5)11D = [(6,10)I)E = [(7.5,10)5)F = ((7.5,10)55)F = F41J = 314101P1 = JM101(pETE1)144(7). Partition in the charge sensitive systems used, [(5,4)I] and[(5,3.5)1] (Figure 54), indicates that although all bacterial strainsare drawn up into the relatively positive upper PEG-rich phase (verystrongly in the case of the [(5,3.5)1] system), the inherent negativecharge on JM101(pETE1) is not disproportionately large or small comparedto the JM101 control strain. The F41 strain shows the greatest changein partition when its partition in the charge sensitive system iscompared to that in the non charge-sensitive analogue, [(5,4)5]. TheF41 bacteria move from the lower phase through the interface topartition strongly in the upper. This movement suggests the presence ofa large accessible negative charge on this strain. There is nosuggestion from this data (nor from cell electrophoresis measurements,Sect. 5.3.3.) that the hemagglutination and partition effects seen inJM101(pETE1)are due to a non-specific highly positively charged proteinon the cell surface. Rather, they likely arise from the overexpressionof an active bacterial membrane adhesive protein.(8). E.coli B/R, which was shown to have no agglutinating activityin the viscometer and hence is postulated not to express type 1fimbriae, has a different partition from JM101 in several systems. Thedifference between B/R and B/R(pETE1) is even more pronounced thanthat seen for JM101 in [(5,4)5] and [(5,4)1] and is in the samedirection (Figure 55).(9). Bacteria-erythrocyte binding experiments in two-phase systemsproduced equivocal results, generally because the three strains undertest (F41, JM101 and JM101(pETE1)) partitioned so differently that itwas impossible to choose a test system in which all strains partitioneddifferently from the erythrocytes. This difference was necessary in145100CB P1bacterial strain80I-a)a 6040c20a015 2004060B P1 B P1Figure 55. Partition of B/R (B) and B/R(pETE1) (P1)In PEG/dextran systemsA = [(5,4)5]B = [(5,4)1]C = [(7.4,4.7)55]146504030?Daa 20=tr) 2030R Rif Fft^1F^R RA 1 Jft 1 JSample1^1^1^1R R/p P/r P304050 1^1^1^1Figure 56. Bacteria-Erythrocyte Partition Binding ExperimentR = Erythrocyte; F or f = F41; J or J = JM101; P or p = JM101(pETE1)R/x = Partition of erythrocyte/bacteria mix measured by countingerythrocytes (where x = f,j or p)X/r = Partition of erythrocyte/bacteria mix measured by countingbacteria (where X = F,J or P)System used = [(7.4,4.7)55]147order to observe a change in the bacteria or erythrocyte partitioncoefficient, which would imply that an interaction between cell typeswas occurring. In the systems chosen, [(5,4)5], [(5,3.5)1],[(7.4,4.7)5] and [(7.4,4.7)55], some effects can be seen, with thebacterial partition coefficient moving when partitioned witherythrocytes. All strains seemed to show effects, F41 being thestrongest, often agglutinating the red cells too much to allow countingin the coulter counter. The results are difficult to compare betweenstrains because of the differences in partition. The JM101 aggregationseen was inhibited when the partition was performed in 10mM mannose.5.3.2. Viscometric Analysis of Hemagglutination Induced by E.coliStrainsViscometric data from experiments studying the effect of shear onthe red cell aggregation induced by JM101(pETE1) showed a surprisinglylarge rise in shear stress when the control JM101 was examined (Figure57). This agglutination was inhibited by mannose. Although type 1agglutinins are known to agglutinate human erythrocytes, the observationwas surprising because no such aggregation was ever seen in microtitreassays, even with very large initial bacteria concentrations and at 37°C. Viscometry of JM101 and JM101(pETE1) in the presence of 10 mMmannose at a shear rate of 43 s -1 , 37 0C (Fig. 57) shows that thebacteria harbouring the pETE1 plasmid produce a modest but readilymeasurable increase in the viscosity of the erythrocyte suspension overtime compared to the cloning vehicle, JM101. This observation can be14885 70 1 62^3^4time (min)32.50.51JM101 (-mannose)^JM(p1) (-mannose) --)K— buffer control--IF— JM101 (+mannose)^JM(p1) (+mannose) —A— F41 (+mannose)Figure 57. Viscometric analysis of hemagglutination mediated byE.coli strains in the absence and presence of 10 mM mannoseShear rate = 43 s -1 , 47 % Hct RBC in HBSS, 37 °C.149rationalized, based on previous observations, as indicating that thebacteria-erythrocyte interaction caused by HRA1 is weaker than theaggregation caused by E.coli F41 or Aeromonas salmonicida 438, with thebonds between the cells being either lower in number, slower to form, ormore easily broken. This shows that viscometry measures differentproperties of the aggregative event from the microtitre hemagglutinationassay. This point is discussed further in the discussion at the end ofthis chapter (Section 5.3.7.).The observation that the mannose-sensitive aggregation associatedwith type 1 pili induced by JM101 seen in the viscometer is entirelyabsent in microtitre assays could indicate a shear-enhancement (100) ofthe agglutination. To investigate this point further, five E.colistrains were examined for mannose-sensitive hemagglutination in theviscometer (Fig. 58) and E.coli B/R was chosen as a strain with nobackground agglutination. The plasmid pETE1 was introduced into thisstrain by electroporation and the same mannose-resistanthemagglutination effect was seen (Fig. 59). The adhesive effect ofB/R(pETE1) was more obvious at low shear rates (2.6s -1 , Figure 59).The B/R(pETE1) bacteria was also shown to agglutinate colo 201 cellsquite strongly at low shear rates (Figure 60). This experiment was notcarried out at higher shear rates because the colo cells were notmechanically stable.1500^1^2^3^4^5^6^8time (min)-U- C600 -9- K12W --)4(-- B/R---4-- R1316 -Ar- K802Figure 58. Hemagglutinating behavior of various E.coli strainsShear rate = 43 s -1 , 47 % Hct RBC in HBSS, 37 °C.1513.00-2.50-1.001. ^ 0.500^1^2^3^:4^5^6^7time (min)I -II-- B/R (2.6s-1)^-B- B/R(p1)(2.6s-1) —AC— B/R(43s-1)^-A.- B/R(p1)(43s-1) 1Figure 59. Increased hemagglutination of B/R(pETE1) at lower shearrates.Shear rates as legend, 47 % RBC in HBSS, 37 °C.1521o I 2^3^4^5^6^7time (min)8—1*-- B/R^B/R(p1)+mannoseFigure 60. Aggregation of cold 201 cells mediated by B/R(pETE1) at ashear rate of 2.6 s-1 , 47 % Hct RBC in HBSS, 37 °C.1535.3.3. Bacterial Electrophoretic MobilityThe electrophoretic mobilities determined by cell electrophoresisin 150mM NaCl were are shown below.Erythrocytes.^A = -1.08 +/- 0.01 pan. sec 1 . V 1 . cmE.coli F41.^A = -0.78 +/- 0.05 Am. sec-1 . V-1 . cmE.coli JM101^A = -1.28 +/- O. 05Am. sec 1 . V 1 . cmE.coli JM101 (pETE1) A = -1.38 +/- 0.05 tun. sec 1 . V 1 . cmThe erythrocyte measurement produced the accepted value (146).The data shows that there is very little difference in electrophoreticmobility between JM101 and JM101(pETE1). This is further evidence thatthe agglutination and partition effects seen in JM101(pETE1) are not dueto a non-specific charge effect since it would be expected that theadhesin would exhibit a net positive charge to cause adhesion to anegatively charged red cell surface. Exposure of such a protein wouldreduce the negative mobility, which was not observed. The mobilityvalue derived for F41 is appreciably lower than those derived for JM101and JM101(pETE1), perhaps indicating a lower surface charge, a findingthat appears at odds with data from the partitioning experiments(Section 5.3.1). It should be noted that the above measurement issimply a measure of the rate of movement of the cells in an electricfield however. To calculate the surface charge distribution otherparameters related to cell surface structure have to be taken intoaccount (146), parameters which were not available in this limitedstudy.1545.3.4. Photomicrographs of the agglutination eventPhotomicrographs show that JM101(pETE1) does show a bindingaffinity for erythrocytes but that the effect appears very weak comparedto F41. Figure 61 shows red cell ghosts agglutinated by F41. The cellsare heavily clumped and distorted. In contrast Figure 62 shows theagglutination mediated by JM101(pETE1) in the presence of mannose. Theerythrocyte ghosts show loose clumping and no distortion. Figure 62also shows a close-up of JM101(pETE1) cells binding to a singleerythrocyte ghost. The cell JM101 is shown with ghosts in the absenceand presence (Fig. 63) of mannose. In the absence of mannose the ghostsare strongly agglutinated, whereas with mannose present there appears tobe little or no interaction between the bacteria and the cell membranes.Figure 64 details the aggregation of colo 201 cells by JM101(pETE1) inthe presence of mannose.155Figure 61. Photomicrograph of erythrocyte ghosts agglutinated anddistorted by F41 bacteria.(400x magnification)156b)Figure 62. Photomicrographs of agglutination of human erythrocyteghosts mediated by JM101(pEIL1) (400x magnification, phase contrast,10mM mannose).a) + mannose, b) + mannose enlargement157b)Figure 63. Photomicrographs of agglutination of human erythrocyteghosts mediated by JM101. (400x magnification, phase contrast)a) no mannose, b) 10 mM mannose158Figure 64. Photomicrograph of JM101(pLIE1) agglutinating colo 201In the presence of 10 mM mannose. (400x magnification, phase contrast)1595.3.5. Electron microscope resultsElectron micrographs of uranyl acetate stained F41, JM101 andJM101(pEiLl) failed to show evidence of piliation (Figure 65). Theresults for F41 are consistent with previous findings in our laboratory(114), even though this strain is known to express the F41 and K99agglutinins, both of which have been described as fimbrial. The JM101strain also shows a smooth surface despite the type 1-like agglutinationit shows in the viscometer.5.3.7. DiscussionAqueous two-phase partitioning studies determined that therecombinant plasmid pETE1 induced pronounced cell surface changes in thehost bacteria. Partitioning of subclones of this plasmid showed thatthese effects could be localized to the product of an open reading framepresent on the foreign DNA insert in the plasmid, plprotl (see Section3.3.2.). This proteinaceous agent of the partition difference has alsobeen identified as the hemagglutinating agent (Section 3.3.2.). Thepartitioning results show that this protein is expressed on the cellsurface and is exposed to the surrounding environment, and that thedifference it induces in the cell surface of JM101 is not solely due toa charge effect. This suggestion is strengthened by the electrophoreticmobility studies, and combined with earlier results regarding thespecific agglutination by JM101(pETE1) of erythrocytes from differentspecies (Section, strongly indicates that the protein160Figure 65. Electron micrographs of uranyl acetate stained bacterialcell surfaces.a). F41 (30,000x) b). JM101 (30,000x) c). JM101(pETE1) (30,000x)161plprotl (HRA1) is an E.coli surface lectin, with unknown receptorspecificity.Viscometry showed that the aggregation of erythrocytes mediated byJM101(pETE1) in the presence of mannose was weaker than that associatedwith F41 expression, producing small viscosity increments at moderateshear rates. At low shear rates B\R(pETE1) produced an appreciableagglutination of human colon adenocarcinoma cells, suggesting that theprotein could have physiological relevance with regard to bacterialattachment to the intestinal epithelial cells. Photomicrograph resultscorrelate with this hypothesis.The bacterium JM101(pETE1) showed a very pronouncedhemagglutination profile in the microtitre assay (Section,agglutinating erythrocytes down to bacterial concentrations ofapproximately one bacteria to two red cells, a very similar degree tothe F41 strain. The fact that JM101(pETE1) shows a weaker aggregatingeffect in the viscometric assay reflects the different parametersmeasured by the two assays. The microtitre assay indicates the presenceand accessibility of agglutinating agents on the bacterial cell surface.Since there is very little shear involved, even a weak interactionbetween the bacteria and the erythrocyte will be sufficient to causeagglutination providing that there is more than one adhesive molecule oneach bacteria (and more than one receptor on each erythrocyte). Thisagglutination will be macroscopically indistinguishable from that of abacterial strain with a much physically stronger adhesive system if theaccessibility of the agglutinins are equivalent. By providing avariable destructive shear force, the viscometric assay can give a162semi-quantitative index of the physicochemical strength of the adhesiveinteraction.Physical strength of adhesive interaction is not the only reasonthat the agglutination effect may appear weak in a viscometric assaycompared to a microtitre assay. The kinetics of the bindinginteractions must be taken into account. In the microtitre assay theagglutinins have a relatively long time in which to recognize and adhereto their receptors on the erythrocyte as the only movement is a gentledownward settling of the red cells. By contrast, in the viscometricassay the shear rate determines that the adhesive interactions mustoccur quickly before the cells are separated by the fluid flow (thecharacteristic time constant is the inverse of the shear rate). Thefast interaction necessary in the viscometric assay favours highmolecular weight multivalent easily accessible adhesin systems such asthe F41 agglutinin. The HRA1 protein may well be present as a monomeranchored in the bacterial outer membrane (see Chapter 6) and as such mayhave a much less accessible binding site.The large mannose-sensitive agglutination of human erythrocytesunder shear exhibited by JM101 was unexpected given the inactivity ofthe strain in a microtitre assay. Previous reports (12) have indicatedthat E.coli strains expressing type 1 fimbria do show an agglutinationeffect in microtitre assays. Other reports (17) indicate that the type1 determinant can be activated by repeated culturing, with the effectbeing initially very weak. The difference between the extent ofagglutination observed in the microtitre assay and that observed in theviscometric assay could arise from shear enhancement of the event (100).The JM101 strain showed no evidence of type 1 fimbriae expression when163examined by electron microscopy. Hultgren et al (147) have alsodescribed mannose-sensitive hemagglutination in the absence of piliationin E.coli. The other strains studied also showed no piliation andalthough the findings for the F41 strain are in agreement with previousresults from our laboratory (114) it is possible that the experimentalconditions precluded the observation of these structures.164Chapter 6. Concluding DiscussionThis thesis documents the molecular cloning and nucleotidesequencing of a region of DNA isolated from Escherichia coli 09:H10:K99.The region contains five open reading frames, four of which code forputative proteins containing an N-terminal prokaryotic secretory signalpeptide, indicating that these proteins are either exported into theintermembrane periplasm, inserted into the outer membrane or secreted.Most of these proteins show little homology with protein sequencesavailable in the database. One, however, plprot5, shows 50 % identityalong virtually its entire length with the amino acid sequence deducedfrom RNA of the E.coli B major outer membrane lipoprotein (134). Thesedata suggest that the region of DNA cloned was involved in determinationof the structure and function of the outer membrane. The high homologyof plprot5 with lipoprotein is useful in mapping the cloned region onthe bacterial chromosome. The plprot5 protein probably serves a relatedfunction to E.coli B lipoprotein, though its primary structure issufficiently different to that protein to suggest that the molecule maygive rise to differences in outer membrane properties.Lipoprotein is an abundant protein in E.coli, and usuallyrepresents about 6 % of total cell protein (68). Approximately onethird of the protein is covalently bound to the peptidoglycan on theinner side of the cell membrane by its C-terminal lysine residue, andthe rest exists freely in region of the outer membrane. Maturelipoprotein bears a cysteine at the N-terminus which is is covalentlylinked to a diacyl glycerol molecule by a thioester bond.165Although a lipoprotein-like molecule would appear to be acandidate for modification of the surface of JM101 when the gene isintroduced in the pETE1 plasmid, possibly leading to the partitiondifference and hemagglutination effect, subcloning has shown that thisis not the case. Lipoprotein is not a receptor for any knownbacteriophages (68), although antibodies prepared against purifiedlipoprotein have been shown to bind to intact bacterial cells that havedefects in LPS structure (137). This data indicates that the majorityof lipoprotein molecules are usually inaccessible to the surroundingenvironment, either on the periplasmic side of the outer membrane or inthe membrane screened by the LPS. Aside from the major E.colilipoprotein, there are many other diverse lipid-modified proteins in thevicinity of the outer membrane (157) and it seems likely that plprot5 isrelated to this family. The protein plprot5 does not have an C-terminallysine for covalent attachment to peptidoglycan, but it does have anarginine residue close to the C-terminus that could fulfill the samefunction. The plprot5 protein does have a very similar signal sequenceand an N-terminal cysteine residue which could be modified in the samemanner as the previously described lipoprotein. The gene encodingplprot5 also has a double stop codon, as does lipoprotein, but thecodons are in reverse order.The location of a lipoprotein-like gene in this DNA fragment putsthe other determined protein sequences into the context of membranestructure and function. The plprot2 gene product lies next to thelipoprotein determinant and the 18.5 kDa peptide not only has a signalsequence but also contains a cysteine at the N-terminus of the matureprotein that has the correct consensus sequence for lipid modification.The protein is also predicted to have three transmembrane helices and is166classified as an integral membrane protein. It is therefore likely thatthis is a membrane-bound lipoprotein that plays a role in theconstruction or stability of the outer membrane. The 24.5 kDa plprot3protein coded on the 2.3 kb HincII fragment also contains a signalsequence, and therefore probably has some function related to bacterialmembrane maintenance.The protein identified as the hemagglutinin and the source of thepartition differences is the 25 kDa (mature) protein plprotl, named HRA1(heat resistant agglutinin 1). Like the proteins described above, thisprotein contains a high scoring predicted signal sequence. The proteinbears very little resemblance to other documented E.coli hemagglutinins,either in amino acid sequence or physical properties. The N-terminalregion of the mature protein shows weak similarity to the NFA4 nonfimbrial adhesin and some related agglutinins, which could indicate arelationship between the molecules, but other considerations make itseem likely that the HRA1 agglutinin is an entirely different system.These considerations include the fact that there is no requirement foran operon of genes for the correct expression of the hemagglutinationeffect (although it is conceivable that the expression of this proteinin JM101 is mediated by the proteins involved in the biosynthesis andexpression of the type 1 adhesin that this strain was found to possess).There is also the observation that HRA1 is extremely difficult toremove from the outer membrane of the cell, even with detergent.Despite this experimental result, HRA1 was not predicted to have anymembrane associated areas (See Section 3.3.3.). This phenomenon is wellknown in the field of E.coli outer membrane porins, where OmpF, OmpC andPhoE are known to be strongly associated with the outer membrane andextremely resistant to detergents and proteases, yet their sequences are167found to be predominantly polar, containing no long hydrophobic segments(158). This raises the possibility of HRA behaving in an analogousmanner to these proteins, existing as a monomer or small oligomer in themembrane and interacting in a similarly strong manner with the membranecomponents and LPS (149). Porins generally contain large amounts ofa-sheet secondary conformation (160,161,162). The secondary structureprediction of Garnier generated for HRA1 (Section 3.3.3.) showed apredominantly g-sheet form, but it should be remembered that thisprediction method is best applied to globular proteins and reports haveshown that even for these proteins the algorithms used are only accurateto approximately 50 % (163,164).The experimental result that heating the bacteria to 60 °C or evenbriefly to 100 °C does not destroy the agglutination together with thestrong membrane association and the lack of necessary expressionsystems also show that the molecule is different from known adhesins andthat it probably does not exist in a fimbrial structure. Fimbriae werenot apparent on JM101(pETE1) when the bacteria were viewed with anelectron microscope, nor were any seen on the source strain, F41.Although some researchers have described the F41 adhesin as fimbrial(65,70), in our laboratory no evidence has ever been collated toindicate that this is the case, although the adhesin does exist in vivoas a high molecular weight protein polymer.The HRA1 protein is not positively charged (predicted pI = 4.44)and so would appear to mediate the agglutination of erythrocytes byeither some specific interaction with a receptor on the cell membrane,or by in some way inducing a change in the lipopolysaccharide extendingoutwards from the membrane surface. With regard to the formerhypothesis, the highest predicted antigenic determinant of HRA1 showed168homology with the known receptor binding site of the K99 adhesin,showing that HRA1 could have the potential to interact with erythrocytesin the same manner as K99. The second-highest scoring segment (with ahydrophilicity index very similar to that of the first) appears veryclose to the C-terminus of the protein and is shown below.-Asp-Lys-Tyr-Lys-Ser-Glu-243^248This section could, if the molecule were exposed at its C-terminalend, also be a potential erythrocyte binding region. A recent report(165) has shown that the Hopp and Woods method for predicting antigenicdeterminants (the method used for HRA1) often correlates with importantprotein-protein and protein-DNA binding domains. With regard to thelatter hypothesis, free lipopolysaccharide has been shown to bind toeukaryotic cells, probably through the central hydrophobic lipid Acomponent (68). Hemagglutination of cells by LPS has not been reportedhowever, and if it occurred it is not likely to be specific, although itis conceivable that the HRA1 protein may disrupt the LPS in some mannerto increase lipid A exposure on the cell surface. There seems no reasonto believe that HRA1 is involved in LPS biosynthesis.Of the two hypotheses above, the more likely explanation seems tobe that the protein itself is the agglutinating agent. It was clonedfrom a strain that expresses two strong mannose-resistant hemagglutininsand would not necessarily be noticed, especially as it would probably bepresent in lower quantities when expressed from the chromosome ratherthan a plasmid. The species' hemagglutination profile of F41 compared169to the purified F41 adhesin indicates that there is anotheragglutinating agent present on the bacteria. It also appears that thepreparation of F41 adhesin from this strain is most often contaminatedwith K99, as SDS-PAGE of column purified F41 adhesin often exhibits aminor 18 kDa band that blots against a-K99 antibodies (114). Thispreparation shows no agglutination of dog erythrocytes, whereas the F41bacteria show a small agglutination and JM101(pETE1) shows pronouncedagglutination of this species. Therefore it is probable that theaggregation of dog erythrocytes is not due to the K99 adhesin, but thatit arises from an uncharacterized minor hemagglutinin on the bacterialcell surface, and that this effect is amplified in JM101(pETE1) bycloning the determinant into a plasmid vector.The aggregation of colo cells by JM101(pETE1) under shear and asviewed microscopically suggests that the adhesion mediated by thisprotein could be physiologically relevant in adherence of bacteria tothe intestinal epithelial cells. The interaction is weak compared tothe F41 and type 1 adhesion events, but may indicate that HRA1 playsa role in general colonization rather than pathogenesis.If HRA1 is to cause hemagglutination by specific interaction witha receptor on the erythrocyte, regions of the protein must be exposed tothe external environment. The outer membrane proteins OmpA (35,160 Da),OmpC (36,000 Da) and OmpF (37,200 Da) are known to be receptorstructures for bacteriophages (166) and so must be exposed at thesurface. Since these proteins are exposed, it is possible that HRA1could also achieve the exposure necessary for receptor recognition.In SDS-PAGE gels of crude membrane samples and cell lysates, thedistinct band that was seen to be unique to JM101(pETE1) compared toJM101 alone was the 25 kDa band corresponding to plprotl, even though170there were open reading frames coding for several other proteins presenton the plasmid. There are two possible reasons for this result.(1). The gene encoding the protein HRA1 is expressed in greateramounts than the genes coding for the other proteins because it is being"read-through" from the truncated tet gene which is under the control ofthe strong tet promoter. There is an inverted repeat between theplprotl open reading frame and the plprot2 start codon, so transcriptionwould cease and the plprot2 gene would be under the control of its ownpromoter. The other open reading frames are in the opposite orientationand would also be under the control of their own promoters.(2). The proteins are all expressed in differing amounts but aremostly unseen due to background bands present arising from JM101proteins. Certainly there would be little chance of seeing plprot5against a large background of JM101 lipoprotein. Since the clonedregion is probably involved in aspects of membrane structure, JM101probably has a chromosomal region specifying similar gene products.The use in this thesis of viscometry and aqueous two-phase polymerpartitioning to study the expression of a recombinant protein isunusual and has proved to be valuable. The effect of plasmid-producedproteins can be studied using the non-recombinant plasmid host (in thiscase JM101) as an excellent control. Since the background bacteria areidentical, any changes that are observed must arise from the products ofthe plasmid. Using molecular biological techniques, the regionresponsible for the effects can be manipulated and changes in behaviornoted.171Viscometric analysis gave valuable insight into the relativestrength of the hemagglutination mediated by HRA1, which correlated withother evidence to suggest that this agglutinin was different from othersstudied. The adhesive interaction was seen to be weaker than that ofother adhesins studied in this laboratory, however this could have beendue to a lower binding site accessibility as discussed in Section 5.4.Aqueous two phase partitioning showed that the hemagglutinincaused dramatic surface changes against the JM101 background and thatthere was no evidence to suggest that the hemagglutination effects weredue to the presence of a highly charged non-specific species on the cellsurface. Partitioning was especially valuable in interpreting theeffect of the Ncol deletion in the HRA1 open reading frame. Thedeletion not only destroyed the agglutinating activity of the host cellbut restored the cells partition to that of JM101, a useful confirmatoryresult.Future ExperimentsFrom the evidence presented in this thesis, it seems likely thatHRA1 is a strongly membrane-associated molecule rather than a"traditional" bacterial hemagglutinin. If this is the case, then thehemagglutination event that this molecule mediates is unusual and itwould be interesting to further study this system. Future experimentscould be conducted in several areas.(1). Screening of a variety of E.coli strains with an oligonucleotideprobe designed from a region of DNA in the open reading frame of plprotl172would determine whether HRA1 or similar proteins are a general featureof this genus or whether the protein is specific for this particularstrain.(2). Investigation of the possible binding sites identified asprobable antigenic determinants should be carried out firstly byattempting to inhibit the agglutination with synthetic peptides of thosesequences. If this approach is successful then site-directedmutagenesis could be undertaken to determine the contributions ofindividual amino acid residues in that area.(3). Monoclonal antibodies could be raised to HRA1 to allowvisualization on western blots and to aid purification.(4). Purification of the protein should be pursued further, either ina traditional manner, isolating outer membranes in a sucrose densitygradient and subsequent purification by membrane disruption and columnchromatography, or by subcloning the gene into a high expression vectorwithout the leader sequence determinant. This would lead to largeamounts of the protein being synthesized in the cytosol and not beingincorporated into the membrane. The protein could then be purified bylysis of the bacteria and chromatography. Once purified, the proteincould be radiolabelled or biotinylated and the erythrocyte ghostblotting experiment repeated to determine the receptor on theerythrocyte.(5). Aqueous two-phase partitioning could be applied to the aboveexperiments in several ways. Partitioning of site-directed mutants173would give interesting data on the role of individual residues indetermining the partition difference. Purified protein could beradiolabelled and partitioned with erythrocytes to further investigatethe adhesion process. If the erythrocyte receptor were elucidated thena binding isotherm could be constructed using phase systems in ananalogous manner to the method used for the valence determination of theF41 adhesin.1747. AbbreviationsA^adenosineAFA afimbrial adhesinAIDA^adhesin involved in diffuse adherenceAP alkaline phosphataseamp^ampicillinAMP adenosine monophosphateATCC^American type culture collectionATP adenosine triphosphateBSA^bovine serum albuminC cytidineCFA^colonization factor antigencolo human colon adenocarcinoma cellsCMC^critical micelle concentrationCS coli surfaceDa^daltonsDNA deoxyribonucleic acidDx^dextranEAF enteropathogenic adherence factorE.coli^Escherichia coliEDTA (ethylenedinitrilo)tetraacetic acidEGTA^(ethylenebis(oxyethylenenitrilo))tetraacetic acidELISA enzyme linked immunosorbent assayEPEC^enteropathogenic E.coliETEC enterotoxigenic E.coliG^guanosineGET glucose-EDTA-Tris bufferHA^hemagglutinationHBSS hanks balanced salt solutionHct^hematocrithepes 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acidHRA^heat resistant agglutininHRP horseradish peroxidaseIgG^immunoglobulin GIPTG isopropyl-g-D-thiogalactopyranosidekDa^kilo Daltonskbp kilo base pairsLB^Luria-Bertani mediumLPS lipopolysaccharideMR^mannose-resistantMS mannose-sensitiveMw^weight average molecular weightNC nitrocelluloseNFA^non-fimbrial AdhesinOmp outer membrane proteinORF^open reading framePAP pill associated with PylonethritisPBS^phosphate-buffered SalinePEG poly(Ethylene Glycol)PIR^protein information resourcePVDF polyvinylidene Fluoride175RBCRNASDSTTETAETBETBSTEMEDtetTHEtrisX-galX-phosphatered blood cellribonucleic acidsodium dodecyl sulphateThymidinetris-EDTA buffertris-acetate-EDTA buffertris-borate-EDTA buffertris-buffered salineN,N,N',N', tetraethylmethylene diaminetetracyclinetris-NaCl-EDTA buffer2-amino-2-(hydroxymethyl)-1,3-propanediol5-bromo-4-chloro-3-indolyl-B-D-galactopyranoside5-bromo-4-chloro-3-indolyl-phosphate8. Glossary of TermsBacteriophage - A virus that is specific for bacteria.Chaperone - A protein that aids the folding and location of newlysynthesized proteins.Clone - Arising from a single cell. In molecular biology cloning oftenrefers to the expression of foreign gene products in a culture startedfrom a single bacterial colony.Codon - A DNA base triplet that codes for one amino acid, or signals thestart or end of an open reading frame.Electrophoretic mobility - The rate of movement of a particle in anelectric field.Electroporation - The introduction of foreign DNA into a bacterial cellby transient permeabilization of the membrane with a high voltage pulse.Enteropathogenic - A bacteria that adheres to epithelial cells andcauses enteric disease in an unclear manner.Enterotoxigenic - A bacteria that causes enteric disease by adhesion toepithelial cells and release of a toxin.Epitope - Region of a protein that is recognized by an antibody.Hemagglutination - The ability of a substance to aggregateerythrocytes.Inverted repeats - Areas of self-complementary DNA that exist in anorientation such that when an RNA transcript is made it forms a hairpinloop, signalling the end of transcription.Klenow fragment - A subtilisin-cleaved fragment of DNA polymerase I thatcarries the 5'-3' polymerase activity and the 3'-5' exonuclease activityof the full enzyme, but not the 5'-3' exonuclease.Lipopolysaccharide - In bacteriology, this term refers topolysaccharide molecules possessing significant lipid moiety present onthe surface of the cell.Maxicells - Bacterial cells that, due to extensive damage of theirchromosomal DNA, do not divide but synthesize plasmid encoded proteinsalmost entirely.Monoclonal Antibody - An antibody derived from a single cell clone andproduced from a spleen cell-tumor cell hybrid culture.Open Reading Frame - A region of DNA prefaced by a start (ATG) codon andterminated by a stop codon, that usually codes for a single protein.177Operon - Clusters of related genes in close proximity, that may betranscribed in a single messenger RNA.Phage - see BacteriophagePhenotype - the expression of genetic information.Plasmid - An extracellular circular DNA molecule, often associated withantibiotic resistance in bacteria. Plasmids have the ability toreplicate independently of the chromosome.Restriction enzyme - An enzyme that recognizes and cleaves specific basesequences in DNA.Serotype - A classification system for bacteria based upon the antigenicproperties of their cell surfaces.Shotgun cloning - An experimental method of creating a library ofrecombinant plasmids containing random foreign chromosomal DNA fragmentseach in a seperate bacterial clone culture. The library can then bescreened for production of proteins of interest.Signal sequence - An amino acid segment present at the N-terminus ofnewly synthesized membrane and periplasmic proteins that enables theprotein to efficiently cross the cytoplasmic membrane. The sequence iscleaved at the outer face of the membrane to leave the mature protein.Transformation - In bacteriology, this term refers to the introductionof DNA into a bacteria by permeablization of the membrane with calciumchloride or other salts.1789. References1. World Health Statistics Annual. 1989. World HealthOrganization, Geneva, p.11.2. Doig, P., T. Todd, P.A. Sastry, K.K. Lee, R.S. Hodges, W.Paranchych, and R.T. Irvine. 1988. Role of Pili in Adhesionof Pseudomonas aeroginosa to Human Respiratory Epithelial Cells.Infect. Immun. 56 : 1641-1646.3. Pasloske, B.L., B.B. Finlay, and W. Paranchych. 1985. Cloning andSequencing of the Pseudomonas aeroginosa PAK pilin gene. FEBSLett. 183 : 408-412.4. Sajjan, S.U., and J.F. Forstner. 1992. Identification of the MucinBinding Adhesin of Pseudomonas cepacia Isolated from Patients withCystic Fibrosis. Infect. Immun. 60 : 4 : 1434-1440.5. Paranchych, W., and L.S. Frost. 1988. The Physiology andBiochemistry of pili. Adv. Microb. Physiol. 29 : 53-114.6. Doig, P., J.W. Austin, M. Kostrzynska, and T.J. Trust. 1992.Production of a Conserved Adhesin by the Human GastroduodenalPathogen Helicobacter pylori. J. Bacteriol. 174 : 8 : 2539-2547.7. Kukkonen, M., T. Raunio, R. Virkola, K. Lahteenmaki, P.H. Makela, P.Klemm, S. Clegg, and T.K. Korhoren. Basement Membrane Carbohydrateas a Target for Bacterial Adhesion : Binding of Type 1 Fimbriae ofSalmonella enterica and Escherichia coli to Laminin. 1993. Molec.Microbiol. 7 : 2 : 229-237.8. Gbarah, A., D. Mirelman, P.J. Sansonetti, R. Verdon, W. Bernhardt,and N. Sharon. 1993. Shigella flexneri Transformants ExpressingType 1 (Mannose-Specific) Fimbriae Bind to, Activate, and areKilled by Phagocytic Cells. Infect. Immun. 61 : 5 : 1687-1693.9. Marrs, C.F., F.W. Rozsa, M. Hackel, S.P. Stevens, and A.C. Glasgow,1990. Identification, Cloning, and Sequencing of piv, a New GeneInvolved in inverting the Pilin Genes of Moxarella lacunata. J.Bacteriol. 172 : 8 : 4370-4377.10. Pearce, W.A., and T.M Buchanan. 1980. Structure and Cell MembraneBinding Properties of Bacterial Fimbriae. In Bacterial AdherenceReceptors and Recognition, Series B, Vol. 6 (Ed. E.H. Beachey).291-340, Chapman and Hall (London).11. Yanagawa, R., and K. Otsuki. 1970. Some Properties of the Piliof Corynebacterium renale. J. Bacteriol. 101 : 3 : 1063-1069.12. Duguid, J.P., I.W. Smith, G. Dempster, and P.N. Edmunds. 1955.Non-Flagellar Filamentous Appendages ("Fimbriae") andHaemaggluinating Activity in Bacterium coli. J. Path. Bacter. 19 :335-348.17913. Honda, E., and R. Yanagawa. 1974. Agglutination of TrypsinizedSheep Erythrocytes by the Pili of Corynebacterium renale Infect.Immun. 10 : 1426-1432.14. Joklik, W.K., H.P. Willett, and D.B. Amos. 1984. ZinsserMicrobiology, 18th Edition, Appleton-Century-Crofts, Conn.15. Christiansen, G.D., W.A. Simpson and E.H. Beachey. 1985.Microbial Adherence in Infection, 6-23 in Mandell, G.L., R.G.Douglas Jr., and J.E. Bennett (ed.). Principles and Practise ofInfectious Diseases, 2nd Edition, Wiley and Sons, New York.16. Hanson, M.S., J. Hempel, and C.C. Brinton, Jr. 1988. Purificationof the Escherichia coli Type 1 Pilin and Minor Pilus Proteins andPartial Characterization of the Adhesin Protein. J. Bacteriol.170 : 8 : 3350-3358.17. Brinton, C.C. Jr., 1965. The Structure, Function, Synthesis, andGenetic Control of Bacterial Pill and a Molecular Model for DNA andRNA Transport in Gram Negative Bacteria. Trans. N.Y. Acad. Sci.27 : 1003-1054.18. Lowe, M.A., S.C. Holt, and B.I. Eisenstein. 1987. ImmunoelectronMicroscopic Analysis of Elongation of Type 1 Fimbriae inEscherichia coli. J. Bacteriol. 169 : 1 : 157-163.19. Ponniah, S., R.O. Endres, D.L. Hasty, and S.N. Abraham. 1991.Fragmentation of Escherichia coli Type 1 Fimbriae Exposes CrypticD-Mannose Binding Sites. J. Bacteriol. 173 : 13 : 4195-4202.20. Schwan, W.R., H.S. Seifert, and J.L. Duncan. 1992. GrowthConditions Mediate Differential Transcription of fim Genes Involvedin Phase Variation of Type 1 Pili. J. Bacteriol. 174 : 7 :2367-2375.21. Orndorff, P.E., and S. Falkow. 1984. Organization and Expressionof Genes Responsible for Type 1 Piliation in Escherichia coli.J. Bacteriol. 160 : 61-66.22. Dorman, C.J., and N.N. Bhriain. 1992. Thermal Regulation of fimA,the Escherichia colt Gene Coding for the Type 1 Fimbrial SubunitProtein. FEMS Microbiol. Lett. 78 : 2-3 : 125-130.23. Blomfield I.C., M.S. McClain, J.A. Princ, P.J. Calie, and B.I.Eisenstein. 1991. Type 1 Fimbriation and fimE Mutants ofEscherichia coli K-12. J. Bacteriol. 173 : 17 : 5298-5307.24. Klemm, P. 1992. FimC, a Chaperone-like Periplasmic Protein ofEscherichia colt Involved in Biogenesis of Type 1 Fimbriae.Research in Microbiol. 143 : 9 : 831-838.25. Krogfelt, K.A., H. Bergmans, and P. Klemm. 1990. Direct Evidencethat the FimH Protein is the Mannose-Specific Adhesin ofEscherichia coli Type 1 Fimbriae. Infect. Immun. 58 : 6 :1995-1998.18026. Krallmann-Wenzel, U., M. Ott, J. Hacker, and G. Schmidt. 1989.Chromosomal Mapping of Genes Encoding Mannose Sensitive (Type 1)and Mannose Resistant F8 (P) Fimbriae of Escherichia coli018:K5:H5. FEMS Microbiol. Lett. 49 : (2-3) : 315-321.27. Stocks, S.J. 1989. Cell Seperations by Immunoaffinity Partition.PhD Thesis, University of British Columbia.28. Klemm, P., and G. Christiansen. 1990. The fimD Gene Required forCell Surface Localization of Escherichia coli Type 1 Fimbriae.Mol. Gen. Genet. 220 : 2 : 334-338.29. Klemm, P., and G. Christiansen. 1987. Three fim Genes Requiredfor the Regulation of Length and Mediation of Adhesion ofEscherichia coli Type 1 Fimbriae. Mol. Gen. Genet. 208 : 439-445.30. Nakazawa, M., M. Haritani, C. Sugimotom and M. Kashiwazaki. 1986.Colonization of Enterotoxigenic Escherichia coli Exhibiting MannoseSensitive Hemagglutination to the Small Intestine of Piglets.Microbiol. Immunol. 30 : 5 : 485-489.31. Durno, C., R. Soni, and P. Sherman. 1989. Adherence of VeroCytotoxin-Producing Escherichia coli Serotype 0157:H7 to IsolatedEpithelial Cells and Brush Border Membranes in Vitro : Role of Type1 Fimbriae (Pili) as a Bacterial Adhesin Expressed by Strain CL-49.Clinic. Investig. Med. 12 : 3 : 194-200.32. Dal Nogare, A.R. 1990. Type 1 Pill Mediate Gram NegativeBacterial Adherence to Intact Tracheal Epithelium. Am. J. Resp.Cell Mol. Biol. 2 : 5 : 433-440.33. Klemm, P., I. Orskov, and F. Orskov, 1982. F7 and Type 1-LikeFimbriae from Three Escherichia coli Strains Isolated from UrinaryTract Infections : Protein Chemical and Immunological Aspects.Infect. Immun. 36 : 2 : 462-468.34. Evans, D.G., R.P. Silver, D.J. Evans, Jr., D.G. Chase, and S.L.Gorbach. 1975. Plasmid-Controlled Colonization Factor Associatedwith Virulence in Escherichia coli Enterotoxigenic for Humans.Infect. Immun. 12 : 656-667.35. Smyth, C.J. 1982. Two Mannose-resistant Hemagglutinins onEnterotoxigenic Escherichia coli of Serotype 06:K15:H16 or H-Isolated from Travellers' and Infantile Diarrhoea. J. Gen.Microbiol. 128 : 2081-2096.36. Willshaw, G.A., M.M. McConnell, H.R. Smith, and B. Rowe. 1990.Structural and Regulatory Genes for Coll Surface Associated Antigen4 (CS4) are Encoded by Seperate Plasmids in EnterotoxigenicEscherichia coli Strains of Serotype 025:H42. FEMS Microbiol.Lett. 56 : 3 : 255-260.18137. Heuzenroeder M.W., B.L. Neal, C.J. Thomas, R. Halter, and P.A.Manning. 1989. Characterization and Molecular Cloning of thePCF8775 CS5 Antigen from an Enterotoxigenic Escherichia coli0115:H40 Isolated in Central Australia. Mol. Microbiol. 3 : 3 :303-31.38. Klemm, P. 1982. Primary Structure of the CFAI Fimbrial Proteinfrom Human Enterotoxigenic Escherichia coli Strains. Eur. J.Biochem. 124 : 2 : 339-348.39. McConnell M.M., H. Chart, and B. Rowe. 1989. Antigenic HomologyWithin Human Enterotoxigenic Escherichia coli Fimbrial ColonizationFactor Antigens (CS)1, CS2, CS4, and CS17. FEMS Microbiol. Lett.52 : 1-2 : 105-108.40. Jordi, B.J., G.A. Willshaw, B.A. van der Zeijst, and W. Gaastra.1992. The Complete Nucleotide Sequence of Region 1 of the CFAIFimbrial Operon. DNA Sequence 2 : 4 : 257-263.41. Orskov, I., and F. Orskov. 1985. Escherichia coli inextraintestinal infections. J. Hyg. Cam. 95 : 551-575.42. Schmoll, T., H. Hoschutzky, J. Morschhauser, F. Lottspeich, K.Jann, and J. Hacker. 1989. Analysis of Genes Coding for theSialic Acid Binding Adhesin and two other Minor Fimbrial Subunitsof the S-Fimbrial Adhesin Determinant of Escherichia coli. Mol.Microbiol. 3 : 12 : 1735-1744.43. Parkkinen, J., G.N. Rogers, T. Korhonen, W. Dahr, and J. Finne.1986. Identification of the 0-Linked Sialyloligosaccharides ofGlycophorin A as the Erythrocyte Receptors for S-FimbriatedEscherichia coli. Infect. Immun. 54 : 1 : 37-42.44. Donnenberg, M.S., and J.B. Kaper. 1992. EnteropathogenicEscherichia coll. Infect. Immun. 60 : 10 : 3953-3961.45. Schroten, H., F.G. Hanisch, R. Plogmann, J. Hacker, G. Uhlenbruck,R. Nobis-Bosch, and V. Wahn. 1992. Inhibition of Adhesion ofS-Fimbriated Escherichia coli to Buccal Epithelial Cells by HumanMilk Fat Globule Membrane Components : a Novel Aspect of theProtective Function of Mucins in the Nonimmunoglobulin Fraction.Infect. Immun. 60 : 7 : 2893-2899.46. Hull, R.A., R.E. Gill, P. Hsu, B.H. Minshew, and S. Falkov. 1981.Construction and Expression of Recombinant Plasmids Encoding Type 1or D-Mannose Resistant Pili from a Urinary Tract InfectionEscherichia coli Isolate. Infect. Immun. 33 : 933-938.47. Normark, S., D. Lark, R. Hull, M. Norgren, M. Baga, P O'Hanley, G.Schoolnik, and S. Falkow. 1983. Genetics of Digalactoside-BindingAdhesin from a Uropathogenic Escherichia coli Strain. Infect.Immun. 41 : 942-949.18248. Lindberg, F., B. Lund, L. Johansson, and S. Normark. 1987.Localization of the Receptor-Binding Protein Adhesin at the Tip ofthe Bacterial Pilus. Nature 328 : 84-87.49. Hal Jones, C., F. Jacob-Dubuisson, K. Dodson, M. Kuehn, L. Slonim,R. Striker, and S.J. Hultgren. 1992. Adhesin Presentation inBacteria Requires Molecular Chaperones and Ushers. Infect. Immun.60 : 11 : 4445-4451.50. Holmgren, A., and C.I. Brandon. 1989. Crystal Structure of PapDreveals an Immunoglobulin Fold. Nature 342 : 245-251.51. Norgren, M., M. Baga, J.M. Tennent, and S. Normark. 1987.Nucleotide Sequence, Regulation and Functional Analysis of the papCGene Required for Cell Surface Localization of Pap Pill ofUropathogenic Escherichia coli. Mol. Microbiol. 1 : 169-178.52. Hultgren, S.J., F. Lindberg, G. Magnusson, J. Kihlberg, and J.M.Tennent. 1989. The PapG Adhesin of Uropathogenic Escherichia coliContains Seperate Regions for Receptor Binding and for theIncorporation into the Pilus. Proc. Natl. Acad. Sci. USA 86 : 12 :4357-4361.53. Uhlin, B.E., M. Norgren, M. Baga, and S. Normark. 1985. Adhesionto Human Cells by Escherichia coli Lacking the Major Subunit of aDigalactoside-Specific Pilus-Adhesin. Proc. Natl. Acad. Sci. USA82 : 1800-1804.54. Lund, B., F. Lindberg, B. Marklund, and S. Normark. 1987. ThePapG Protein is the a-D-galactopyranose-Binding Adhesin ofUropathogenic Escherichia coli. Proc. Natl. Acad. Sci. USA 84 :5898-5902.55. Jacob-Dubuisson, F., J. Heuser, K. Dodson, S. Normark, and S.Hultgren. 1993. Initiation of Assembly and Association of theStructural Elements of a Bacterial Pilus Depend on Two SpecializedTip Proteins. EMBO Journal 12 : 3 : 837-847.56. Leffler, H., and C. Svanborg-Eden. 1980. Chemical Identificationof a Glycosphingolipid Receptor for Escherichia coli Attaching toHuman Urinary Tract Epithelial Cells and Agglutinating HumanErythrocytes. FEMS Microbiol. Lett. 8 : 127-134.57. Braaten, B.A., L.B. Blyn, B.S. Skinner, and D.A. Low. 1991.Evidence for a Methylation-Blocking Factor (mbf) Locus Involved inPap Pilus Expression and Phase Variation in Escherichia coli.J. Bacteriol. 173 : 5 : 1789-1800.58. van der Woude, M.W., B.A. Braaten, and D.A. Low. 1992. Evidencefor Global Regulatory Control of Pilus Expression in Escherichiacoli by Lrp and DNA Methylation : Model Building Based on Analysisof Pap. Mol. Microbiol. 6 : 17 : 2429-2435.18359. Orskov I., F. Orskov, W.J. Sojka, and J.M. Leach. 1961.Simultaneous Occurence of Escherichia coli B and L Antigens inStrains from Diseased Swine. Acta. Pathol. Microbiol. Scand.Section B, 53 404-422.60. Gaastra, W., and F. K. de Graaf. 1982. Host-Specific FimbrialAdhesins of Noninvasive Enterotoxigenic Escherichia coli Strains.Microbiol Reviews 46 : 2 : 129-161.61. Guinee P.A.M., and W.H. Jansen. 1979. Behavior of Escherichiacoli K Antigens K88ab, K88ac, and K88ad in Immunoelectrophoresis,Double Diffusion and Hemagglutination. Infect. Immun. 23 :700-705.62. Mooi, F.R., N. Harms, D. Bakker, and F.K. de Graaf. 1981.Organization and Expression of Genes Involved in the Production ofthe K88ab Antigen. Infect. Immun. 32 : 1156-1163.63. Jacobs, A.A.C., J. Venema, R. Leeven, H. van Pelt-Heerschap, andF.K. de Graaf. 1987. Inhibition of Adhesive Activity of K88Fibrillae by Peptides Derived from the K88 Adhesin. J. Bacteriol.169 : 2 : 735-741.64. Jacobs, A.A.C., B. Roosendaal, J.F.L. van Breemen, and F.K. deGraaf. 1987. Role of Phenylalanine 150 in the Receptor-BindingDomain of the K88 Fibrillar Subunit. J. Bacteriol. 169 : 11 :4907-4911.65. de Graaf, F.K., B.E. Krenn, and P. Klaasen. 1984. Organizationand Expression of Genes Involved in the Biosynthesis of K99Fimbriae. Infect. Immun. 43 : 1 : 508-514.66. Lindahl, M., R. Brossmer, and T. Wadstrom. 1987. CarbohydrateReceptor Specificity of K99 Fimbriae of Enterotoxigenic Escherichiacoli. Glycoconjugate J. 4 : 51-58.67. Lindahl, M., and I. Carlsedt. 1990. Binding of K99 Fimbriae ofEnterotoxigenic Escherichia coli to Pig Small Intestinal MucinGlycopeptides. J. Gen. Microbiol. 136 : 8 : 1609-1614.68. Hammond, S.M., P.A. Lambert, and A.N. Rycroft. 1984. TheBacterial Cell Surface. Kapitan Szabo Publishers, Washington D.C.69. de Graaf, F.K., and I. Roorda. 1982. Production, Purification,and Characterization of the Fimbrial Adhesive Antigen F41 Isolatedfrom Calf Enteropathogenic Escherichia coli Strain B41M. Infect.Immun. 36 : 2 : 751-758.70. Moseley, S.L., G. Dougan, R.A. Schneider, and H.W. Moon. 1986.Cloning of Chromosomal DNA Encoding the F41 Adhesin ofEnterotoxigenic Escherichia coli and Genetic Homology betweenAdhesins F41 and K88. J. Bacteriol. 167 : 799-804.18471. Anderson, D.G., and S.L. Moseley. 1988. Escherichia con F41Adhesin: Genetic Organization, Nucleotide Sequence, and Homologywith the K88 Determinant. J. Bacteriol. 170 : 4890-4896.72. Smit, M., W. Gaastra, J.P. Kamerling, J.F.G. Vliegenhart, and Graaf. 1984. Isolation and Structural Characterization of theEquine Erythrocyte Receptor for Enterotoxigenic Escherichia coliK99 Fimbrial Adhesin. Infect. Immun. 46 : 578-584.73. Lindahl, M., and T. Wadstrom. 1986. Binding to ErythroctyeMembrane Glycoproteins and Carbohydrate Specificity of F41 Fimbriaeof Enterotoxigenic Escherichia coli. FEMS Microbiol. Lett. 34 :297-300.74. Tomita, M., H. Furthmayr, and V.T. Marchesi. 1978. PrimaryStructure of Human Erythrocyte Glycophorin A. Isolation andCharacterization of Peptides and Complete Amino Acid Sequence.Biochemistry 17 : 4756-4770.75. Lindahl, M. 1989. Binding of F41 and K99 Fimbriae ofEnterotoxigenic Escherichia coli to Glycoproteins from Bovine andPorcine Colostrum. Microbiol. Immunol. 33 : 5 : 373-379.76. Brooks, D.E., J. Cavanagh, D. Jayroe, J. Janzen, R. Snoek, and T.J.Trust. 1989. Involvement of the MN Blood Group Antigen in ShearEnhanced Hemagglutination Induced by the Escherichia coli F41Adhesin. Infect. Immun. 57 : 2 : 377-383.77. van Zijderveld, F.G., F. Westenbrink, J. Anakotta, R.A. Brouwers,and A.M. van Zijderveld. 1989. Characterization of the F41Fimbrial Antigen of Enterotoxigenic Escherichia coli by usingMonoclonal Antibodies. Infect. Immun. 57 : 4 : 1192-1199.78. Girardeau, J.P., M. der Vartanian, J.L. 011ier, and M. Contrepois.1988. CS31A, a New K88-Related Fimbrial Antigen on BovineEnterotoxigenic and Septicemic Escherichia coli Strains. Infect.Immun. 56 : 8 : 2180-2188.79. Korth, M.J., R.A. Schneider, and S.L. Moseley. 1991. An F41-K88Related Determinant of Bovine septicemic Escherichia coli MediatesExpression of CS31A Fimbriae and Adherence to Epithelial Cells.Infect. Immun. 59 : 7 : 2333-2340.80. Isaacson, R.E., and P. Richter. 1981. Escherichia coli 987Ppilus: Purification and Partial Characterization. J. Bacteriol.146 : 784-789.81. Schifferli, D.M., E.H. Beachey, and R.K. Taylor. 1991. GeneticAnalysis of 987P Adhesion and Fimbriation of Escherichia coli: thefas Genes link both Phenotypes. J. Bacteriol. 173 : 3 : 1230-1240.82. Olsen, A., A. Jonsson and S. Normark. 1989. Fibronectin BindingMediated by a Novel Class of Surface Organelles on Escherichiacoli. Nature 338 652-655.18583. Arnqvist, A., A. Olsen, J. Pfeifer, D.G. Russell, and S. Normark.1992. The Crl Protein Activates cryptic Genes for Curli Formationand Fibronectin Binding in Escherichia coli HB101. Mol. Microbiol.6 : 17 : 2443-2452.84. Duguid, J.P., S. Clegg, and M.I. Wilson. 1979. The Fimbrial andNon Fimbrial Hemagglutinins of Escherichia coll. J. Med.Microbiol. 12 : 213-228.85. Hultgren, S.J., S. Normark, and S.N. Abraham. 1991. ChaperoneAssisted Assembly and Molecular Architecture of Adhesive Pili.Ann. Rev. Microbiol. 45 : 383-415.86. Guyot, G. 1908. Ueber die Bakterielle HUmagglutination (Bakterio-Haemoagglutination). Zentl. Bakt. I. Abt. Orig. 47 : 640.87. Orskov, I., A. Birch-Anderson, J.P. Duguid, J. Stenderup, and F.Orskov. 1985. An Adhesive Protein Capsule of Escherichia coli.Infect. Immun. 47 : 1 : 191-200.88. Stirm, S., I. Orskov, and F. Orskov. 1966. K88, an EpisomeDetermined Protein Antigen of Escherichia coli. Nature 209 :507-508.89. Williams, P.H., S. Knutton, M.G.M. Brown, D.C.A. Candy, and A.S.McNeish. 1984. Characterization of Nonfimbrial Mannose-ResistantProtein Hemagglutinins of Two Escherichia coli Strains Isolatedfrom Infants with Enteritis. Infect. Immun. 44 : 3 : 592-598.90. Benz, I., and M.A. Schmidt. 1989. Cloning and Expression of anAdhesin (AIDA-I) Involved in diffuse Adherence of EnteropathogenicEscherichia coli. Infect. Immun. 57 : 5 : 1506-1511.91. Nataro, J.P., M.M. Baldini, J.B. Kaper, R.E. Black, N. Bravo, andM.M. Levine. 1985. Detection of an Adherence Factor ofEnteropathogenic Escherichia coli with a DNA Probe. J. Infect.Dis. 152 560-565.92. Goldhar, J., R. Perry, J.R. Golecki, H. Hoschutzky, B. Jann, and K.Jann. 1987. Nonfimbrial, Mannose-Resistant Adhesins fromUropathogenic Escherichia coli 083:K1:H4 and 014:K?:H11. Infect.Immun. 55 : 8 : 1837-1842.93. Walz, W., M.A. Schmidt, A.F. Labaigne-Roussel, S. Falkow, and G.Schoolnik. 1985. AFA-1, a Cloned Afimbrial X-type Adhesin from aHuman Pyelonephritic Escherichia coli Strain. Eur. J. Biochem.152 : 315-321.94. Labaigne-Roussel, A., M.A. Schmidt, W. Walz, and S. Falkow. 1985.Genetic Organization of the Afimbrial Adhesin Operon and NucleotideSequence from a Uropathogenic Escherichia coli Gene Encoding anAfimbrial Adhesin. J. Bacteriol. 162 : 3 : 1285-1292.18695. Grunberg, J., R. Perry, H. Hoschutzky, B. Jann, K. Jann, and J.Goldhar. 1988. Nonfimbrial Blood Group N-Specific Adhesin (NFA-3)from Escherichia coli 020:KX104:H-, Causing Systemic Infection.FEMS Microbiol. Lett. 56 : 241-246.96. Hoschutzky, H., W. Nimmich, F. Lottspeich, and K. Jann. 1989.Isolation and Characterization of the Non-Fimbrial Adhesin NFA-4from Uropathogenic Escherichia coli 07:K98:H6. MicrobialPathogenesis 6 : 351-359.97. Nowicki, B., A. Labigne, S. Moseley, R. Hull, S. Hull, and J.Moulds. 1990. The Dr Hemagglutinin, Afimbrial Adhesins AFA-I andAFA-III, and F1845 Fimbriae of Uropathogenic and Diarrhea-Associated Escherichia coli Belong to a Family of Hemagglutininswith Dr Receptor Recognition. Infect. Immun. 58 : 1 : 279-281.98. Brooks, D.E., and T.J. Trust. 1980. An Improved Technique for theStudy of Bacterial Adhesion to Surfaces in Microbial Adhesion toSurfaces, 513-515, ed. R.C.W. Berkerly, J.M. Lynch, V. Molling,P.R. Butter, and B. Vincent. Ellis Horwood, Chichester.99. Brooks, D.E., and T.J. Trust. 1983. Enhancement of BacterialAdhesion by Shear Forces: Characterization of the HaemagglutinationInduced by Aeromonas salmonicida Strain 438. J. Gen. Microbiol.129 : 3661-3669.100. Brooks, D.E., and T.J. Trust. 1983. Interactions of Erythrocyteswith Bacteria under Shear. Ann. N.Y. Acad. Sci. 416 : 319-331.101. Oudega, B., and F.K. de Graaf. 1988. Genetic Organization andBiogenesis of Adhesive Fimbriae of Escherichia coli. Antonie vanLeeuwenhoek 54 : 285-299.102. Brossmer, R., G. Burk, V. Eschenfelder, L. Holmquist, R. Jackh, B.Neumann, and U. Rose. 1974. Recent Aspects of the Chemistry ofN-Acetyl-D-Neuraminic Acid. Behring Inst. Mitt. 55 : 119-123.103. Hughes, R.C. 1983. Glycoproteins. Chapman and Hall, New York,New York.104. BHP-activated Dextran Handbook. 1977. Pharmacia, Uppsala,Sweden.105. Jacobs, A.A.C., L.H. Simons, and F.K. de Graaf. 1987. The Roleof Lysine-132 and Arginine-136 in the Receptor Binding Domain ofthe K99 Fibrillar Subunit. EMBO J. 6 : 1805-1808.106. Jacobs, A.A.C., P.A. van den Berg, H.J. Bak, and F.K. de Graaf.1986. Localization of Lysine Residues in the Binding Domain of theK99 Fibrillar Subunit of Enterotoxigenic Escherichia coll.Biochim. Biophys. Acta 872 : 92-97.107. Albertsson, P.A. 1986. Partition of Cell Particles andMacromolecules (Third Edition). John Wiley and Sons, New York, NewYork.187108. Bamberger, S., D.E. Brooks, K.A. Sharp, J.M. van Alstine, and T.J.Webber. 1985. Preparation of Phase Systems and Measurement ofTheir Physicochemical properties. In Partitioning in AqueousTwo-Phase Systems, ed. H. Walter, D.E. Brooks, and D. Fisher,p85-130. Academic Press Inc. Toronto.109. Brooks, D.E., K.A. Sharp, and D. Fisher. 1985. TheoreticalAspects of Partitioning. In Partitioning in Aqueous Two-PhaseSystems, ed. H. Walter, D.E. Brooks, and D. Fisher, p11-85.Academic Press Inc. Toronto.110. Albertsson, P.A., and G.D. Baird. 1962. Counter-currentDistribution of Cells. Exptl. Cell Res. 28 : 296.111. Stendahl, 0., L. Edebo, K.E. Magnusson, C. Tagesson, and S.Hjerten. 1977. Surface-charge Characteristics of Smooth and RoughSalmonella Typhimurium Bacteria Determined by Aqueous Two-PhasePartitioning and Free Zone Electrophoresis. Acta Pathol.Microbiol. Scand. Sect. B: Microbiol. 85B : 334-340.112. Magnusson, K.E., E. Kihlstrom, A. Norqvist, J. Davies, and S.Normark. 1979. Effect of Iron on Surface Charge andHydrophobicity of Neisseria gonorrhoeae. Infect. Immun. 26 :402-407.113. Kihlstrom, E., and K.E. Magnusson. 1983. Haemagglutinating,Adhesive and Physicochemical Surface Properties of DifferentYersinia enterocolitica and Yersinia enterocolitica-like Bacteria.Acta Pathol. Microbiol. Immunol. Scand. Sect. B: Microbiol. 91B :113-119.114. Brooks, D.E. Personal communication.115. Isaacson, R.E. 1978. K99 Surface-antigen of Escherichia coliAnigenic Characterization. Infec. Immun. 22 : 2 : 555-559.116. Yannish-Perron, Celeste, Jeffrey Vieira, and Joachim Messing.1985. Improved M13 Phage Cloning Vectors and Host Strains:Nucleotide Sequences of M13mp18 and pUC19 Vectors. Gene 33 :103-199.117. Maniatis, T., E.F. Fritsch, and J. Sambrook. 1982. MolecularCloning, A Laboratory Manual. Cold Spring Harbour Laboratories,New York.118. Bolivar, F., R.L. Rodriguez, P.J. Greene, M.V. Betlach, H.L.Heyneker, H.W. Boyer, J.H. Crosa, and S. Falkow. 1977.Construction and Characterization of New Cloning Vehicles. II. AMultipurpose Cloning System. Gene 2 : 95-113.119. Davis, B.J. 1964. Disc Electrophoresis II. Method andApplication to Human Serum Proteins. Ann. N.Y. Acad. Sci. 121 :404-427.188120. Ornstein, L., 1964. Disc Electrophoresis I. Background andTheory. Ann. N.Y. Acad. Sci. 121 : 321-349.121. Towbin, H. , T. Staehelin, and J. Gordon. 1979. ElectrophoreticTransfer of Proteins from Polyacrylamide Gels to NitrocelluloseSheets : Procedure and some Applications. Proc. Natl. Acad. Sci.USA 76 : 4350-4354.122. Old, R.W., and S.B. Primrose. 1981. Principles of GeneManipulation (Second Edition). Blackwell Scientific Publications,London.123. Southern, E.M., 1975. Detection of Specific Sequences among DNAFragments Seperated by Gel Electrophoresis. J. Mol. Biol. 98 :503-515.124. Helfman, D.M., and S.H. Hughes. 1987. Use of Antibodies to ScreencDNA Expression Libraries Prepared in Plasmid Vectors. MethodsEnzymol. 152: 451-458.125. Sanger, F., S. Nicklen, and A.R. Coulson. 1977. DNA Sequencingwith Chain-Terminating Inhibitors. Proc. Natl. Acad. Sci. USA 74 :5463-5467.126. Zoller, M.J., and M. Smith, 1983. Oligonucleotide-directedMutagenesis of DNA Fragments Cloned into M13 Vectors. Methods inEnzymology 100 : 468.127. Von Heijne, G. 1985. A New Method for Predicting Signal SequenceCleavage Sites. Nucleic Acids Research 11 : 4683-4690.128. Klein, P., M. Kanehisa, and C. Delis'. 1985. The Detection andClassification of Membrane Spanning Segments. Biochim. Biophys.Acta 815 : 468-476.129. Garnier, J., D.J. Osguthorpe, and B. Robson. 1978. Analysis ofthe Accuracy and Implications of Simple Methods for Predicting theSecondary Structure of Globular Proteins. J. Mol. Biol. 120 :97-120.130. Gascuel, O. and J.L. Golmard. 1988. A Simple Method for Predictingthe Secondary Structure of Globular Proteins - Implications andAccuracy. CABIOS 4 : 357-365.131. Shine J., and L. Dalgano. 1974. The 3'-Terminal Sequence ofEscherichia coli 16S ribosomal DNA: Complementarity to Non-senseTriplets and Ribosome Binding Sites. Proc. Natl. Acad. Sci. USA71 : 1342-1346.132. Hussain, M., S. Ichihara, and S. Mizushima. Mechanism ofSignal Peptide Cleavage in the Biosynthesis of the MajorLipoprotein of the Escherichia coli Outer Membrane. 1982. J. Biol.Chem. 257 : 9 : 5177-5182.189133. Bhat, K.S., C.P. Gibbs, O. Borrera, S.G. Morrisn, F. Jahrnig,A. Stern, E.M. Kupsch, T.F. Meyer and J. Swanson. 1991. TheOpacity of Neisseria gonorrhoeae Strain MS11 are Encoded by aFamily of 11 Complete Genes. Mol. Microbiol. 5 : 1889-1901.134. Nakamura, K., R.M. Pirtle, I.L. Pirtle, K. Takeishi, and M.Inouye. 1980. Messenger Ribonucleic Acid of the Lipoprotein ofthe Escherichia coli Outer Membrane: II. The Complete NucleotideSequence. J. Biol. Chem. 255 : 1 : 210-216.135. Miller, H. 1987. Practical Aspects of Preparing Phage andPlasmid DNA: Growth, Maintenance, and Storage of Bacteria andBacteriophage. Meth. Enz. 152 : 145-170.136. Hopp, T., and K.R. Woods. 1981. Prediction of Protein AntegenicDeterminants from Amino Acid Sequences. Proc. Natl. Acad. Sci.USA 78 : 6: 3824-3828.137. Braun, V. 1975. Covalent Lipoprotein from Outer Membrane ofEscherichia coli. Biochim et Biophys Acta 415 : 335-377.138. Sancar, A., and Claud S. Rupert. 1978. Determination ofPlasmid Molecular Weights from Ultraviolet Sensitivities. Nature272 : 471-472.139. Guinee, P.A.M., J. Veldkamp, and W.H. Hansen. 1977. ImprovedMinca Medium for the Detection of K99 Antigen in CalfEnterotoxigenic Strains of Escherichia coli. Infection andImmunity 15 : 676-678.140. Marchesi, V.T., and E.P. Andrews. 1971. Glycoproteins: Isolationfrom Cell Membranes with Lithium Diiodosalicylate. Science 174 :1247-1248.141. Cassidy, J.T. et al. 1965. Purification and Properties ofSialidase from Clostridium Perfringens. J. Biol. Chem. 240 : 9 :3501-3506.142. Sancar, A., A. Hack, and W. Rupp. 1979. Simple Method forIdentification of Plasmid-Encoded Proteins. J. Bacteriol. 137 :692-693.143. Myers, E.W., and W. Miller. 1988. Optimal Alignments inLinear Space. CABIOS 4 : 1 : 11-17.144. Inouye, M., J. Shaw, and C. Shen. 1972. The Assembly of aStructural Lipoprotein in the Envelope of Escherichia coli. J.Biol. Chem. 247 : 24: 8154-8159.145. Semple, T.V., Laurie A. Quinn, Linda K. Woods, and George E.Moore. 1978. Tumor and Lymphoid Cell Lines from a Patient withCarcinoma of the Colon for a Cytotoxicity Model.^Cancer Research38 : 1345-1355.190146. Seaman, G.V.F. 1975. Elektrokinetic Behaviour of Red Blood Cells.In The Red Blood Cell. Ed. D.M. Sturgenor. Academic Press, NY :1135-1229.147. Hultgren, S.J., J.L. Duncan, A.J.Schaeffer, and S.K. Amundsen.1990. Mannose-sensitive Haemagglutination in the Absence ofPiliation in Escherichia coli. Mol. Microbiol. 4 : 8 : 1311-1318.148. Pluskal, M.G., M.B. Przekop, M.R. Kavonian, and D.A. Hicks. 1986.Immobilon PVDF Transfer Membrane: A New Membrane Substrate forWestern Blotting of Proteins. BioTechniques 4 : 272-283.149. Einstein, A. 1911. Berichtigung zu meiner Arbeit: Eine NeueBestimmung der Molekuledimensionen. Ann. Physik. 34 : 591-592.150. Ford, T.F. 1960. Viscosity - Concentration and Fluidity -Concentration Relationships for Suspensions of Spherical Particlesin Newtonian Liquids. J. Phys. Chem. 64 : 1168-1174.151. Heimenz, P.C. 1977. Principles of Celloid and Surface Chemistry.Marcel-Dekker Inc. New York, NY.152. Brooks, D.E. 1975. Red Cell Interactions in Low Flow States. InMicrocirculation. Ed. J.Grayson and W. Zing. Plenum Press, NY.33-52.153. Dower, W.J., J.F. Miller, and C.W. Ragsdale. 1988. High EfficiencyTransformation of Escherichia coli by High Voltage Electroporation.Nuc. Acids Res. 16 : 13 : 6127-6145.154. Dower, W.J., B.M. Chassy, J.T. Trevors and H.P. Blaschek. 1990.Protocols for Transformation of Bacteria by Electroporation. InGuide to Electroporation and Electrofusion. Ed. D.C. Chang, B.M.Chassy, J.A. Saunders, and A.E. Sowers. Academic Press. SanDiego, CA. 485-499.155. Eisenberg, D., E. Schwarz, M. Komaromy and R. Wall. 1984.Analysis of Membrane and Surface Protein Sequences with theHydrophobic Moment Plot. J. Mol. Biol. 179 : 125-142.156. Peden, K.W.C. 1983. Revised Sequence of the TetracyclineResistence Gene of pBR322. Gene 22 : 2-3 : 277-280.157. Hayashi, S. and H.C. Wu. 1990. Lipoproteins in Bacteria. J.Bioenergetics and Biomembranes. 22 : 3 : 451-471.158. Mizuno, T., M-Y. Chou and M. Inouye. 1983. A Comparative Study onthe Genes for the Three Porins of the Escherichia coli OuterMembrane. J. Biol. Chem. 258 : 11 : 6932-6940.159. Jap, B.K. and P.J. Walian. 1990. Biophysics of the Structure andFunction of Porins. Q. Rev. Biophys. 23 : 4 : 367-403.191160. Vogel, H., and F. Jahnig. 1986. Models for the Structure of OuterMembrane Porins of Escherichia coli Derived from Raman Spectroscopyand Prediction Methods. J. Molec. Biol. 190 : 191-199.161. Garavito, R.M., J. Jenkins, J.N. Jansoniu, R. Karlsson, and J.P.Rosenbusch. 1983. X-ray Diffraction Analysis of Matrix Porin,an Integral Membrane Protein from Escherichia coli Outer Membranes.J. Molec. Biol. 164 : 313-327.162. Cowan, S.W., T. Schirmer, G. Rummel, M. Steiert, R. Ghesh, R. A.Pauptit, J. N. Jasonius, and J. P. Rosenbusch. 1992. CrystalStructures Explain Functional properties of Two Escherichia coliPorins. Nature 358 : 727-733.163. Kabsch, W., and C. Sander. 1983. How Good are Predictions ofProtein Secondary Structure? FEBS Lett. 155 : 2 : 179-182.164. Busetta, B., and M. Hospital. 1982. An Analysis of the Predictionof Secondary Structures. Biochim. et Biophys. Acta. 701 :111-118.165. Hopp, T.P. 1993. Prediction of Protein Functional Sites fromAmino Acid Sequences : New Applications of the Hopp and WoodsMethod. Abstract in Protein Science a : supp.1 : 103.166. DiRienzo, J.M., K. Nakamura and M. Inouye. 1978. The OuterMembrane Proteins of Gram-Negative Bacteria : Biosynthesis,Assembly, and Functions. Ann. Rev. Biochem. 47 : 481-532.167. Donnenberg, M.S., J.A. Giron, J.P. Nataro and J.B. Kaper. 1992. APlasmid-Encoded Type IV Fimbrial Gene of EnteropathogenicEscherichia coli associated with Localized Adherence. Mol.Microbiol. 6 : 22 : 3427-3437.192Appendix 1Amino acid abbreviations and codonsamino acid abbreviations codonsalanine Ala A GCT,GCC.,GCA,GCGarginine Arg R AGA,AGGasparigine Asn N AAT,AACaspartic acid Asp D GAT,GACcysteine Cys C TGT,TGCglutamine Gln Q CAA,CAGglutamic acid Glu E GAA,GAGglycine Gly G GGT,GGC,GGA,GGGhistidine His H CAT,CACisoleucine Ile I ATT,ATC,ATAleucine Leu L CTT,CTC,CTA,CTG,TTA,TTGlysine Lys K AAA,AAGmethionine Met M ATGphenylalanine Phe F TTT,TTCproline Pro P CCT,CCC,CCA,CCGserine Ser S AGT,AGCthreonine Thr T ACT,ACC,ACA,ACGtryptophan Trp W TGGtyrosine Tyr Y TAT,TACvaline Val V GTT,GTC,GTA,GTG(termination codons = TAA, TAG, TGA)193Appendix 2General Phase Diagram for a PEG/dx/water Phase SystemDX (%w/w)The lines BC and B'C' connect the points representing thecomposition of the two phases at equilibrium. For example, B representsthe bottom phase composition and C the top phase composition. The totalcomposition is given by A and the ratio AC:AB will give the weight ratioof bottom to top phase. K represents the critical point. (From 27)194Appendix 3.Preperation of Dulbecco's Modified Eagle mediumDME was made up from powdered packets from Gibco (cat. no. 430-1600,low glucose medium with L -glutamine and sodium pyruvate). One litre ofmedium was made up at a time and contained components in the followingconcentrations -mg/1Inorganic SaltsCaC12.21120^264.9Fe(NO3)3.9H20 0.100KC1^ 400.0MgSO4 97.7NaC1 6400.0NaH2PO4^ 125.0Other Componentsglucose^1000.0phenol red, sodium salt 15.0sodium pyruvate^110.0Amino AcidsL-arginine.HC1^84.00L-cystine, disodium salt 56.78L-glutamine 584.0glycine^ 30.00L-histidine.HC1.H20^42.00L-isoleucine^104.8L-leucine 104.8L-lysine.HC1 146.2L-methionine^30.00L-phenylamine 66.00L-serine 42.00L-threonine^95.20L-tryptophan 16.00L-tyrosine 72.00L-valine^ 93.60VitaminsD-Ca pantothenate^4.00choline chloride 4.00folic acid^4.00i-inositol 7.00nicotinamide 4.00pyridoxal.HC1^4.00riboflavin 0.4000thiamin.HC1 4.00195Appendix 4.Sequencing Strategybeginning of F41 insertHincII^HincIIHincII21pBR322 DNApPL7...........■...0R ----->---...*--.4Rtr--..-..-•R.11(■■■■■■■■HincII^ HincII pPL4R5 3^4........................).0I0.5 1 12= plprotl ORF= plprot2 ORF3 = plprot3 ORFScale (in kb) 4 = plprot4 ORF5 = plprot5 ORFSequencing vector used was pTZ19R. Insert DNA shown only. Insertfragments were ligated into vector in both orientations. Althoughplasmid pPL7 is shown above, the plasmids pPL2 and pPL3 (containing thepPL7 HincII fragments seperately) were also used for sequencing. Theoligonucleotides used for these plasmids are shown on the pPL7 diagramR refers to M13 reverse primer, other primers used were designed toregions of sequenced DNA. Arrows indicate direction of sequencing.196


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