UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of the cell surface of caulobacter crescentus Walker, Stephen George 1994

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

Item Metadata


831-ubc_1994-954035.pdf [ 3.44MB ]
JSON: 831-1.0088236.json
JSON-LD: 831-1.0088236-ld.json
RDF/XML (Pretty): 831-1.0088236-rdf.xml
RDF/JSON: 831-1.0088236-rdf.json
Turtle: 831-1.0088236-turtle.txt
N-Triples: 831-1.0088236-rdf-ntriples.txt
Original Record: 831-1.0088236-source.json
Full Text

Full Text

STUDIES OF THE CELL SURFACE OF CAULOBACTER CRESCENTUSBySTEPHEN GEORGE WALKERB. Sc. Hons., The University of Western Ontario, 1984M. Sc., The University of Guelph, 1987A THESIS SUBMITI’ED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Microbiology and Immunology)We accept this thesis as conformingto the required standard.....September 1994THE UNIVERSITY OF BRITISH COLUMBIA© Stephen George Walker, 1994In presenting this thesis in partial fulfillment of therequirements for an advanced degree at the University of BritishColumbia, I agree that the Library shall make it freely availablefor reference and study. I further agree that permission forextensive copying of this thesis for scholarly purposes may begranted by the head of my department or by his or herrepresentatives. It is understood that copying or publication ofthis thesis for financial gain shall not be allowed without mywritten permission.(Signature)____________________Department of IV\J-(RO iSIoCOyThe University of British ColumbiaVancouver, CanadaDateAbstractCaulobacter crescentus is a Gram-negative eubacteria whichproduces a surface layer (S-layer). S-layers are paracrystallineassemblies of protein that cover the outer surface of some eubacteriaand archaebacteria cells. The method by which the protein subunitscomposing the S-layer of C. crescentus, RsaA, interact to form thearray and attach to the cell was examined in this thesis.The S-layer was extracted from the cell surface of Ccrescentus NA1000 by treating cells with a pH 2 solution or asolution containing 10 mM ethylene glycol-bis(f-aminoethylether)N,N,N,N’-tetraacetic acid (EGTA). The extracted extract wasexamined by sodium dodecyl sulfate (SDS) - polyacrylamide gelelectrophoresis (PAGE) and found to consist of nearly pure RsaA. Theisolated S-layer was amorphous in structure but could reassemblein vitro into a crystalline array in the presence of calcium ions.Two mutants of C. crescentus NA1000, JS1001 and JS1002,selected for the ability to grow in the absence of calcium had theadditional phenotype of being unable to attach the S-layer to the cellsurface although they produced a wild-type RsaA. These mutantsshed the S-layer into the surrounding medium during growth.Methods were developed to identify, isolate and purify the cellsurface molecules of the wild-type and S-layer attachment-defectivestrains. It was determined that the mutant strains did not produce asmooth lipopolysaccharide (LPS) although they produced the wild-type rough LPS and extracellular polysaccharide. The smooth LPS(termed the S-layer associated oligosaccharide or SAO) was veryhomogeneous in length as determined by SDS-PAGE and silverstaining. RsaA negative strains that could form and attach an S-layeron the cell surface if RsaA was expressed on a plasmid vector werealso shown to produce SAO. All S-layer attachment-defectivemutants examined did not produced SAO. Two cosmids wereidentified that partially restored the production of SÃO in the mutantstrain JS1001 however restoration of S-layer attachment did notoccur.IIITABLE OF CONTENTSPageAbstract iiTable of Contents ivList of Figures ixList of Tables xivList of Abbreviations xvAcknowledgments xviiDedication xviii1. Introduction 11 .1 Bacterial surface arrays 21.2 The S-layer of Caulobacter crescentus 112. Materials and Methods 162.1 Chemicals 162.2 Bacterial strains 162.3 Growth media 162.4 Isolation of calcium-independent mutants 192.5 Growth studies 192.6 Colourimetric assays 202.7 Isolation and purification of cell surface molecules 202.7.1 LPS isolation 20iv2.7.2 EPS isolation 222.7.3 RsaA isolation 232.7.4 SAO isolation and purification 242.8 Antisera production 252.8.1 Production of antiserum to low pHextracted RsaA 252.8.2 Antisera to SAO 262.9 In vitro crystallization of S-layer 262. 10 Electrophoretic methods 272.10.1 SDS-PAGE 272.10.2 Western blotting 282.10.3 Sample preparation 282. 11 Silver staining 292. 12 Cell preparation for thin-section electronmicroscopy 302. 13 Negative stain electron microscopy 312.14 Transposon mutagenesis of NA1000 322. 15 Complementation of mutant strains usingcosmids 322. 16 Carbohydrate and lipid chemical analysis 333. Results 343. 1 Analysis of S-layer attachment by Westernblotting 343.2 Isolation and purification of EPS 38V3.2.1 Assessment of EPS cell association 383.2.2 Chemical characterization of EPS 393.3 Isolation and purification of LPS 413.3.1 Electrophoretic analysis of LPS 413.3.2 Isolation LPS, monitored by lipid analysis 433.3.3 Colourimetric analysis of LPS 453.3.4 Detailed chemical analysis of LPS 453.4 Identification of SÃO 473.5 Isolation and purification of SÃO 523.5.1 Extraction of cell surface molecules 523.5.2 Examination of cell extracts 523.5.3 Examination of extracted cells 553.5.4 Purification of SÃO 593.6 Purification of S-layer protein 613.6.1 Extraction of the S-layer of C. crescentusNA1000 613.6.2 In vitro recrystallization of NA1000S-layer 653.6.3 Anti-RsaA sera 653.7 Comparison of S-layers among Caulobacterisolates 683.7.1 S-layer extraction 683.7.2 Western blot analysis of extracted proteins 733.7.3 Polysaccharide analysis 793.8 Metal ion requirements for C. crescentus growthviand S-layer assembly 853.8.1 Influence of calcium on cell growth 863.8.2 Influence of metal ions on growth rate 863.8.3 Influence of metal ions on S-layercrystallization 903.8.4 Influence of Ca2+ or Sr2+ concentration onS-layer crystallization 953.8.5 The localization of non-crystallized S-layerprotein 953.9 Genetic studies of the calcium-independentphenotype 1003.9.1 Production and screening of a Tn5 library 1003.9.2 Complementation of mutants with a cosmidlibrary 1044. Discussion 1074.1 The C. crescentus surface polysaccharides 1074.1.1 C. crescentus “rough” LPS 1 074.1 .2 C. crescentus SAO 1114.1.3 C. crescentus EPS 1184.2 RsaA extraction and in vitro recrystallization 1 2 14.3 Distribution of RsaA- and SAO-like molecules inenvironmental Caulobacter isolates 1234.4 Ionic requirements for C. crescentus NA1000viigrowth and expression/crystallization of RsaA 1 284.5 Genetic studies of the calcium-independent /S-layer attachment-defective phenotype 1 354.6 Conclusions 1374.6.1 The relationship between calcium-independence and loss of SAO 1 374.6.2 The role of Ca2 or Sr2 in thecrystallization of S-layer 1404.6.3 The role of SAO in S-layer attachment 1404.7 Summary 1415. References 1426. Appendix I 1696.1 Isolation of Calcium-independent mutants 1 696.2 Production of anti-SAO sera 1696.3 Colony immunoblot for identification of S-layermutants 1707. Appendix II 171Figure 1 172Figure 2 174Figure 3 176Table I 177Table II 178Table III 1 79VIIILIST OF FIGURESFigure Page1. Detection of RsaA in Caulobacter crescentusstrains by Western Blot analysis. 3 52. Fractionation of the EPS of CB2A and NA1000 onSephacryl S-400. 403. Analysis of the LPS of CB2A and NA1000 bySDS-PAGE and silver stained. 424. Analysis of the LPS of various Caulobactercrescentus strains by SDS-PAGE and silverstained. 445. Gas chromatographic analysis of the fatty acidmethyl esters from the LPS of NA 1000. 466. Analysis of the LPS of NA1000 containingcontaminating SAO by SDS-PAGE and silverstained or examined by Western blotting usinga-SAO sera. 487. Analysis of various Caulobacter crescentusstrains by SDS-PAGE. (A) Silver stained to detectLPS and SAO. (B) Examined by Western blottingusing a-SAO sera. 508. Analysis of proteinase K treated NA 1000NaC1IEDTA extract by SDS-PAGE and silverixstained or examined by Western blotting usingx-SAO sera. 5 39. Analysis of NA1000 NaC1JEDTA extract bySDS-PAGE and coomassie-blue stained orexamined by Western blotting using x-RsaA sera. 5410. Gas chromatographic analysis of the aiditolacetates from NA1000 EPS isolated by themethod of Darveau and Hancock (1983) (A) orNaC1/EDTA extraction (B). 5 611. Transmission electron micrograph of thinsectioned NA1000 cell (A) and NA1000 cellextracted with NaC1/EDTA (B). 5 712. Analysis of the carbohydrates purified fromJS1003 NaC1/EDTA extracts by SDS-PAGEand silver staining. 6 013. Analysis of low pH extract from NA 1000 bySDS-PAGE and stained by Coomassie blue orexamined by Western blotting using x-RsaA sera. 6 214. Analysis of urea-solubilized protein fromJS1001 cultures by SDS-PAGE and stained byCoomassie blue. 6415. Negative-stain transmission electron micrographof the in vitro recrystallization product of RsaA. 6 616. Analysis of low pH extract of NA 1000 and wholecell lysate of NA1000 by Western blottingxusing unabsorbed cc-RsaA sera. 6717A. Analysis of low pH extracts from Caulobacterstrains by SDS-PAGE and Coomassie blue staining. 6917B. Analysis of EGTA extracts from Caulobacterstrains by SDS-PAGE and Coomassie blue staining. 7 118. Analysis of low pH (A) or EGTA (B) extractsfrom Caulobacter strains by Western blotting usingx-RsaA sera. 7 419. Analysis of proteinase K-treated whole celllysates of Caulobacter strains by SDS-PAGE andsilver staining. 8020. Analysis of proteinase K-treated whole celllysates of Caulobacter strains by Western blottingusing c-SAO sera. 8 321. The influence of calcium concentration on thegrowth of Caulobacter crescentus NA1000 andJS1001 cultured in M10Higg medium. 8722. The influence of calcium concentration on thegeneration time of Caulobacter crescentusNA1000 cultured in M10Higg medium. 8 823. The influence of the metal ion supplement on thegeneration time of Caulobacter crescentusNA 1000 cultured in M10Higg medium. 8 924. The influence of the metal ion supplement on thegeneration time of Caulobacter crescentusxiJS1001 cultured in M10Higg medium. 9 125. The influence of the metal ion supplement on thelag phase of Caulobacter crescentus NA1000cultured in M10Higg medium. 9226. The influence of the metal ion supplement on thelag phase of Caulobacter crescentus JS1001cultured in M10Higg medium. 9327. Negative-stain transmission electron micrographof strontium mediated crystallization of RsaAin a Caulobacter crescentus JS1001 colony. 9428. Analysis of whole cell lysates of Caulobactercrescentus JS1001 cells, cultured in M10Higgliquid medium supplemented with a chloride metalsalt, by Western blotting using x-RsaA sera. 9 829. Analysis of whole cell lysates of Caulobactercrescentus JS1001 cells, cultured on M10Higgmedium plates supplemented with a chloride metalsalt, by Western blotting using x-RsaA sera. 9 930. Analysis of proteinase K-treated NaC1JEDTAextracts of Caulobacter crescentus NA1000 Tn5mutants by SDS-PAGE and silver staining. 1023 1. Analysis of proteinase K-treated cell lysatesof Caulobacter crescentus strains by Westernblotting using x-SAO sera. 1 0532. Analysis of proteinase K-treated cell lysatesxiiof Caulobacter crescentus strains by Westernblotting using o-SAO sera. 1 0633. A representation of the cell surface ofCaulobacter crescentus NA1000. 1 08XIIILIST OF TABLESTable PageI. Bacterial strains. 1 7II. Relevant characteristics of Caulobacter strains. 7 7III The influence of the template and cationconcentration on the crystallization of RsaA. 96xivList of AbbreviationsC terminal carboxy terminalcm centimetercFU colony forming unitDDW distilled deionized waterDNA deoxyribonucleic acidDNase deoxyribonucleaseEDTA ethylendiaminetetra-acetic acidEGTA 1 ,2-Di(2-aminoethoxy)ethane-NNN’N’ -tetra-acetic acidEPS extracellular polysaccharideFWC freshwater CaulobacterF1EPES N-2-hydroxyethylpiperazine-N’ -2-ethane sulfonic acidct gas chromatographyg gravityh hourkDa kilodaltonkV kilovoltsKDO 2-keto-3-deoxyoctonateLPS lipopolysaccharideL litreMCS marine Caulobacter strainM molarmg milligrammm minutexvml milliliterMS mass spectrometrymicrogrammicrolitre$.Lm micrometerN normalnm nanometerNMR nuclear magnetic resonanceN terminal amino terminalODyj optical density at 600 nmohmsPAGE polyacrylamide gel electrophoresisPCII phenol-chloroform-hexanePYE peptone yeast extractPBS phosphate-buffered salineRNA ribonucleic acidRNase ribonucleaseSÃO S-layer associated oligosaccharideS-layer surface layerSDS sodium dodecyl sulfateSEC steric exclusion chromatographyTEM transmission electron microscopy‘]lC thin-layer chromatographyTris Tris(hydroxymethyl)methylaminexviAcknowledgmentsI would like to thank my supervisor Dr. John Smit for hisguidance and encouragement during my studies. I would also like tothank Dr. R. E. W. Hancock, Dr. R. S. Molday and Dr. G. B. Spiegelmanfor serving on my advisory committee. A special thanks goes to Dr.Spiegelman for reading and early draft of this thesis.I would like to thank Dr. Neil Ravenscroft and Dr NedraKarunaratne for analyzing the carbohydrates purified during mywork.xviiDedicationThis thesis is dedicated to my parents Dr. and Mrs. G. R. Walker.I would like to thank them for their love, support and guidancethroughout my life. Thanks Mom and Dad.xviii1. IntroductionWithin cells enzymes synthesize biomolecules by the formationof covalent bonds between substrates. The biomolecules that areformed, be they proteins, nucleic acids, lipids or polysaccharides,must then be organized into functional structures. The componentsof the functional, or supramolecular, structure interact mainlythrough the formation of weak (hydrogen, ionic, hydrophobic) bonds.Supramolecular structures can be composed of identical biomoleculessuch as the protein actin in the case of an actin filament or can becomposed of a variety of different biomolecules as in a bacterialribosome. The supramolecular structure can be constructed eitherby the process of “instructed morphogenesis” or “self-assembly”(Cohen 1977; Sitte 1981).Supramolecular structures fashioned by instructedmorphogenesis require the action of components that are notretained in the final structure. Thus the final components of thestructure do not contain all the information required for assembly.These additional components may take the form of atemplate/scaffold or a proteolytic cleavage event (Kellenberger1990). Supramolecular structures built through self-assembly do notrequire any additional components to achieve the final functionalform. All the information required for assembly is contained in thebiomolecules that make up the structure. The only requirements area sufficient concentration of the subunits and suitable environmental1conditions (Sitte 1981). Thus the isolated components of asupramolecular structure that forms through self-assembly can,under appropriate conditions, spontaneously reassemble into thefinal structure in vitro.Bacterial surface layers (S-layers) are examples ofsupramolecular structures built through the self-assembly process.This thesis reports on studies of the mechanisms by which theprotein subunits that form the S-layer of Caulobacter crescentusinteract to produce a crystalline array and the method by which theS-layer associates with the bacterial cell surface. The introductionserves as a brief review on bacterial S-layers and the S-layer of C.crescentus. For more in depth information on S-layers the reader isreferred to published reviews by Baumeister et a!. (1988), Koval(1988), Messner and Sleytr (1992), Sleytr and Messner (1983, 1988),and Smit (1987).1. 1 Bacterial surface arraysS-layers are two dimensional crystalline arrays ofproteinaceous subunits forming surface layers on prokaryotic cells(Sleytr et al. 1988). The subunits are usually composed of a singleprotein or glycoprotein species which self-assembles to form acharacteristic lattice (Sleytr and Messner 1983). The subunits varyin molecular weight, between species and strains, from 30 - 220 kDa.S-layers are common components of prokaryotic cell design, beingfound on over 200 species of eubacteria and archaebacteria, however2in comparison to other wall components little is known about thesestructures (Messner and Sleytr 1992). This lack of knowledge is due,for the most part, to the absence of S-layers on the enteric bacteriawhich are the most thoroughly studied of all prokaryotic groups.Historically, research on S-layers has focused on structuralstudies to determine how bacteria maintain these layers. S-layersare identified in transmission electron microscopy studies by virtueof their characteristic periodic morphology (Smit 1987). The detailedstructure of the S-layer subunits are not readily discernible howeverdue to the limited resolution in these images (Hovmöller et a!.1988a). To overcome this, computerized image processingtechniques are used to obtain an unbiased averaged image in theform of a two dimensional density map (Amos et al. 1982). Bycombining two dimensional information from a tilt series of the samespecimen, three dimensional reconstructions with 1.3 nm resolutionhave been produced (Baumeister and Engelhardt 1987; Chalcroft etal. 1986). Such analytical techniques have determined that S-layerproteins, within the crystal lattice, consist of a large core domain anda smaller connecting domain. This asymmetrical conformation allowsthese subunits to arrange themselves into a variety of patterns ormorphological units the most common containing 2, 4 or 6 monomers.Crystallization of these building blocks results in hexagonal (p6),tetragonal (p4.) or linear (p2) lattice types (Saxton and Baumeister1986). The resulting patterns are very heterogeneous betweenspecies and strains of the same species with respect to lattice3symmetry and the centre-to-centre spacing of the unit cell. Threeand two dimensional reconstructions indicate that the S-layers arealso asymmetric with respect to the two surfaces. The surface facingthe external environment is smooth in character while the surfaceproximal to the cell is generally rough (Sleytr and Messner 1988). Ithas been assumed that these layers maintain a standard pore sizeunder all growth conditions, however the S-layer of Aeromonassalmonicida (Garduflo and Kay 1992; Stewart et al. 1986) and somethermophilic Bacillus Sp. (Sleytr and Sara 1986) have been shown toundergo structural transformations which apparently alter theporosity.Figure 1 in appendix II illustrates some of the features of 5-layers using the S-layer of C. crescentus as an example. Two S-layermonomers are shown in Fig. 1A (appendix II) illustrating the twodomains. Six S-layer monomers assemble to produce a hexagonal“unit cell” or “morphological unit” (Fig. 1B; appendix II). Unit cellsthen interact to form the final array structure (Fig. 1C; appendix II).Kinetic studies have determined that crystallization of S-layermonomers into the final array proceeds by a two step mechanism asillustrated in Fig. 1 (appendix II) and the first step occurs at a fasterrate than the second (Sleytr and Messner 1983).The majority of S-layer producing bacteria have a single layer,although double S-layers have been found on Aquaspirillummetamorphum (Beveridge and Murray 1975), A. “Ordal” (Beveridgeand Murray 1976c), A. serpens MW5 (Kist and Murray 1984), A.4sinuosum (Smith and Murray 1990), Lampropedia hyalina (Austinand Murray 1990), Nitrocystis oceanus (Remsen et al. 1970) andBacillus brevis 47 (Tsuboi et al. 1982), and three layers have beenobserved on Waisby’s “square bacterium” (Stoeckenius 1981). Thedouble layer of A. “Ordal” (Beveridge and Murray 1976c) containsan outer layer of hexagonal symmetry and an inner layer oftetragonal symmetry. The two S-layers of A. serpens MW5 are bothhexagonal in symmetry but are antigenically unrelated (Koval et a!.1988) while the proteins of the two S-layers of Bacillus brevis 47differ in molecular weight they are produced as a cotranscriptionalunit. An S-layer is usually composed of a single protein. However,the very complex S-layers of L. hyalina (Austin and Murray 1990),Flexibacter polymorphus (Ridgeway and Lewin 1983) andChiamydia trachomatis (Chang et al. 1982) are composed of two ormore polypeptides. The pathogen Campylobacter fetus produces anS-layer that undergoes an antigenic shift. A single strain canproduce S-layer proteins that vary by molecular weight andantigenic character (Dubreuil et al. 1990; Wang et a!. 1990).Biochemical studies of S-layers have demonstrated that generalsimilarities exist between the protein subunits produced by diversespecies. Most subunits are held together, and to the underlying cellsurface, by non-covalent (hydrophobic, ionic, hydrogen or polar)bonds and are similar with respect to amino acid composition (Kovaland Murray 1984a; Messner and Sleytr 1992; Sleytr and Messner1983). They generally contain a large proportion of acidic andhydrophobic amino acids and little or no sulphur-containing amino5acids. These proteins contain a high proportion of random coil, 20 -35 percent beta sheet and very little alpha helix (< 2 - 14 percent)(Koval 1988; Sleytr and Messner 1988). However, not all S-layerproteins conform to these generalizations. The very thermophilicand sheathed archaebacterial (Konig and Stetter 1986), Chiamydiace.Sp. (Newhall and Jones 1983), and the inner tetragonal layer of A.sinuosum (Smith and Murray 1990) have subunits that arecovalently bonded and are highly resistant to denaturation byphysical or chemical methods. Subunits can also be modified withcovalently attached carbohydrates. Glycoprotein containing subunitswere first identified in Halobacterium salinarium (Mescher andStrominger 1976) and have since been located in a number ofarchaebacterial and eubacterial species. These bacterialglycoproteins contain substantial differences compared to those ineukaryotes with respect to both the glycan chains and linkages(Messner and Sleytr 1988; Messner and Sleytr 1991).Although similarities exist between S-layer proteins producedby unrelated species, analysis of the S-layer genes has identifiedlittle to no sequence homology (Gilchrist et al. 1992; Messner andSleytr 1992). Genetic (Gilchrist et al. 1992; Messner and Sleytr1992), ultrastructural (Hovmöller et al. 1988b) and biochemical(Sleytr and Messner 1983) comparisons have led to a generalconsensus among researchers that S-layers are of a non-conservednature and have arisen independently in species by convergentevolution. An S-layer may evolve in a given species to fulfill a6specific function(s); However, the protein subunits of all species mustbe capable of three major tasks: secretion, self-assembly or bondingwith adjacent subunits, and attachment to the underlying cell surface(Smit 1987). It is assumed that the similarities between S-layersubunits with respect to general amino acid composition, method ofsubunit bonding and gross morphology are due to the proteins allhaving to fulfill the above tasks.The mechanisms by which S-layers are secreted have not beenstudied in great detail in comparison to the mechanisms involved inassembly and attachment to the cell surface. Of the 18 sequenced Slayer genes all but 4; Campylobacter fetus (Blaser and Gotschlich1990), Caulobacter crescentus (Gilchrist et al. 1992), Rickettsiaprowazekii (Carl et al. 1990), and R. rickettsii (Gilmore et al. 1989),contain a cleaved N-terminal signal sequence. Many outermembrane proteins are believed to be transported from theperiplasm to the outer membrane through adhesion zones (Bayer1979) However, Belland and Trust (1985) have indicated that the Slayer monomer of Aeromonas salmonicida is transported from thecytoplasm to the distal side of the outer membrane by a mechanismwhich includes a step where the protein is free in the periplasm. Alinkage between S-layer and lipopolysaccharide (LPS) translocationfrom the cytoplasm to the outer membrane has been suggested forAcinetobacter 199A and Aeromonas salmonicida (Thorne et al.1976; Belland and Trust 1985).The non-covalent forces responsible for subunit-subunit and7subunit-cell surface stability in S-layers are determined byidentifying conditions under which the crystalline arrays willdisintegrate and then reassemble into a regular array (Beveridge1981; Koval and Murray 1984a; Smit 1987; Sleytr and Messner1983). Within the same S-layer, the subunit-subunit and subunit-cell surface bonds may be of a different nature although it is oftendifficult to differentiate between the two interactions (Smit 1987). Ithas been noted that when metal ions are required for reassembly anS-layer Ca2 is the ion of choice. Only Bacillus brevis (Tsuboi et al.1982) and Sporosarcina ureae (Beveridge 1979) have a absoluterequirement for Mg2+ while Aeromonas salmonicida appears torequire both Ca2 and Mg2 (Garduno et al. 1992b).The cell surface molecules with which S-layers interact tomaintain cell association have been identified in only a few bacterialspecies. The S-layer of Deinococcus radiodurans is anchored to thecell via proteins (Thompson et al. 1982) whereas Clostridiumdifficile utilizes neutral cell surface polysaccharides (Masuda andKawata 1981). Other Gram-positive species may use anionic sites onthe peptidoglycan (Beveridge 1981; Hastie and Brinton 1979). TheS-layers of Gram-negative eubacteria have been shown to interactwith the outer membrane via protein(s) in Acinetobacter 199A(Thorne et al. 1975), and perhaps Spirillum putridiconchylium(Beveridge and Murray 1976b), and LPS in Aeromonas salmonicida(Belland and Trust 1985), A. hydrophila (Dooley and Trust 1988)and Carnpylobacter fetus (Yang et a!. 1992). Spirillum serpens8requires both LPS and lipid (Chester and Murray 1978). For speciespossessing a double S-layer, the upper layer will often onlyreassemble in the presence of the lower layer as illustrated withAquaspirillum serpens MW5 (Kist and Murray 1983).Although much is known about the structure and biochemicalcomposition of S-layers no one definitive function has been proposedfor these layers. Because S-layer producing bacteria are founded inalmost every environmental niche it is unlikely that all S-layers havean identical function. It is assumed that S-layers have evolved toserve different functional roles due to particular environmentalstresses that a species encounters in its habitat. The function 5-layers serve must be important to the survival of the cell when oneconsiders the great energetic cost of the layer. S-layer monomersaccount for up to 10 percent of the total cellular protein and theenergy expenditure is even greater when glycosylization occurs(Sleytr and Messner 1988). Considering this cost, it is not surprisingthat many S-layer producing strains lose them upon laboratorycultivation in the absence of environmental stress (Blaser et al. 1985;Buckmire 1971; Luckevich and Beveridge 1989; Stewart andBeveridge, 1980).The observation of pore-like structures formed within S-layersindicates that these layers may act as a molecular sieve and restrictthe diffusion of molecules larger than the exclusion limit of the pore.S-layers with exclusion limits less than that of a harmful moleculewould thus protect the cell. S-layers protect some bacterial strainsfrom lysozyme (Nermut and Murray 1967), various proteases (Sleytr91976), predation (Buckmire 1971; Koval and Hynes 1991) and hostvirulence factors (Ishiguro et al. 1981; Blaser et a!. 1988)presumably by exclusion.The most obvious function for an S-layer is in the type threearchaebacterial wall, consisting of a plasma membrane surroundedby an S-layer (Kandler and Konig 1985), where it determines the cellshape (Sleytr et al. 1986a). Shape determination and structuralintegrity has also been attributed to the S-layer sheaths of somemethanogenic archaebacteria (Patel et al. 1986). A number of otherfunctions have been suggested for S-layers. Beveridge and Murray(1976a) proposed that the electronegative character of most S-layersmay serve to concentrate essential cations from a dilute environmentor protect cells by immobilizing toxic ions (Beveridge 1979).Alternatively, S-layer interaction with soluble ions may also act tobuffer the environment immediately surrounding the cell and inhibitlarge changes in pH (Stewart and Beveridge 1980). S-layers havealso been implicated as a means to promote bacterial adhesion tomacrophages (Garduflo et al. 1992a; Trust et al. 1983), epidermalcells (Baumeister and Hegerl 1986), porphyrin and immunoglobulin(Kay et al. 1988; Phipps and Kay 1988), fibronectin and laminin (Doiget al. 1992; Kay and Trust 1991), bacteriophage (Edwards and Smit1991; Howard and Tipper 1973) and between bacteria viaautoagglutination (Evenberg and Lugtenberg 1982).This brief review of bacterial S-layers illustrates that althoughthese structures appear “similar” at a superficial level (two10dimensional arrays composed of acidic proteins lacking cysteine andof similar secondary structure) detailed ultrastructural, biochemicaland genetic studies have revealed that these layers are veryheterogeneous, sometimes even between strains of the same species,and are evolutionaraly unrelated. Therefore, when anuncharacterized S-layer is studied it is difficult to predict how theprotein subunits are secreted to the outer membrane, assembled intoan array and attached to the cell surface.1.2 The S-layer of C. crescentusC. crescentus is a Gram-negative eubacterium whichundergoes a sequence of morphological changes at specific polarmembrane sites during its life cycle (for review, see Poindexter 1964and 1981; Shapiro 1976). Swarmer cells express a single flagellum,bacteriophage receptors, pili and an adhesive substance termedholdfast all at one cell pole. All polar features but the holdfast arelost and a stalk develops, at the same pole, as the swarmerdifferentiates into a stalked cell. The stalk, which is an outgrowth ofthe cell envelope and contains no cytoplasmic material, remainsthrough all subsequent generations. Swarmers are produced bygrowth and division of the stalked cell with the swarmer cell polarsurface appendages being produced at the pole distal to the stalk cell.With the exception of the pilus (Smit and Agabian 1982a) control ofthe production of the polar structures is linked to DNA replication. Asingle round of replication occurs during the life cycle and these11polar events are initiated at the midpoint by an unknown signal. Themajority of research conducted with C. crescentus has focused ondissecting the developmental process at the genetic level (Dingwall etal. 1990; Shapiro 1993). Throughout the entire life cycle of C.crescentus the cell surface is completely covered with an S-layer(Smit et al. 1981).The C. crescentus S-layer is of hexagonal symmetry (Smit et al.1981). Smit et al. (1992) have produced a three-dimensionalreconstruction to a resolution of 2.0 nm (see Fig. 1; appendix II). Thereconstruction shows that the morphological unit is formed by sixprotein subunits that are arranged on a p6 lattice. The subunitsforming the array contain a heavy domain, that interacts to form acentral hexagonal core, and a lighter domain that connects adjacentmorphological units. The centre to centre distance between themorphological units is 22 nm. The interaction of the heavy domainregions produces a central pore that has a diameter of 2.5 - 3.5 nm.A similar size of gap is found in the space between the unit cells.These breeches in the array would allow the passage of globularproteins no larger than approximately 17 kDa (Smit et al. 1992).The gene encoding the protein that forms the S-layer (rsaA)has been cloned (Smit and Agabian 1984) and sequenced (Fisher etal. 1988; Gilchrist et al. 1992). The predicted molecular weight ofRsaA is 98,132 and the predicted p1 is 3.46. Apart from the removalof the initial methionine no N-terminal or C-terminal processing ofthe protein occurs during secretion (Gilchrist et aT. 1992).12RsaA is a major cellular protein accounting for 5-7% of totalprotein synthesis (Smit et al. 1981). Indirect immunocytochemicalmethods have been used to examine where new S-layer isincorporated into the crystalline lattice during cell growth. Twosystems of S-layer incorporation were distinguished by Smit andAgabian (1982b). During growth newly synthesized S-layer isincorporated at random locations on the cell body but only new 5-layer was incorporated onto the growing stalk and the newly formedpole. Because array components are synthesized uniformly duringthe cell cycle (Agabian et al. 1979) and a cell possesses only one copyof rsaA (Smit and Agabian 1984), control of S-layer incorporationmust work at the level of assembly rather than transcription. Toaccount for the two mechanisms of S-layer assembly, it waspostulated that on the cell body S-layer arrived via transientadhesion zones between the inner and outer membrane, whereas atthe newly formed pole and stalk more stable sites of membraneadhesion exist. At present no information is available on how the 5-layer protein is secreted or targeted to the cell surface.Calcium has been implicated in the assembly of the Ccrescentus S-layer and/or in the attachment of the layer to the cellsurface. Gilchrist et a!. (1992) identified 4 or 5 putative calciumbinding motifs in the C-terminal region of the predicted amino acidsequence of rsaA. Similar glycine-rich repeats have been identifiedin hemolysins and proteases of other species which require calciumfor biological activity. Apart from the homology analysis there is nodirect demonstration that the S-layer binds calcium. Poindexter13(1982) showed that wild-type C. crescentus has a growthrequirement for calcium and that mutants could be selected which nolonger require calcium. Analysis of the mutants revealed that theycould no longer attach RsaA to the cell surface and that in highdensity liquid cultures a macroscopic debris, presumed to consist ofRsaA, was formed.Smit et a!. (1992; unpublished studies; see appendix I [methodA]) isolated a number of calcium-independent mutants from C.crescentus NA1000. It was demonstrated that these mutantsproduced large sheets of non-cell-associated, but assembled S-layer,when cultured on calcium-containing agar. Smit et a!. (1992) showedthat these sheets were composed of two S-layers that associate viathe side of the array which is proximal to the cell surface in thewild-type situation. The two layers were in such precise alignmentthat three-dimensional image reconstruction was required to resolvethat a double layer existed. When the calcium-independent mutantswere grown on calcium-free plates no assembled S-layer wasdetected. This data indicated that the mutants were not defective inS-layer assembly and that calcium was involved in the assemblyprocess. The S-layer gene was cloned from three calciumindependent mutants and each was individually introduced into C.crescentus JS1003. The chromosomal copy of rsaA of the parentstrain, NA1000, was deleted to produce JS1003 (Smit, unpublished).Negative-stain electron microscopy and indirect immunofluorescencemicroscopy (Smit et al. unpublished) showed that a wild-type S-layer14was produced on the cell surface of JS1003 when any of the S-layergenes from the calcium-independent mutants were expressed on aplasmid vector. This implied that the mutation resulting in the Slayer attachment-defective phenotype was not located in rsaA butin some other gene whose product is involved in the attachment thearray to the cell surface.The central goal of this thesis was to identify, isolate andcharacterize the cell surface molecule(s) of C. crescentus whichinteract with the RsaA in order to attach the S-layer to the cell. Thecell surface of calcium-independent / S-layer attachment-defectiveC. crescentus strains and the wild-type strain, NA1000, wereexamined in order to identify any differences between the wild-typeand mutant strains. An examination of the LPS of the Caulobacterstrains determined that the mutants did not produce a smooth LPS.The smooth LPS, termed the “S-layer associated oligosaccharide”(SAO), was produced by all wild-type / S-layer attachmentcompetent strains but not by any of the calcium-independent /attachment-defective mutants.Some of the data presented in this thesis has been previouslypublished as Ravenscroft et al. 1991, Ravenscroft et al. 1992, Walkeret a!. 1992 and Walker et al. 1994.152. Materials and Methods2.1 ChemicalsUnless otherwise stated all chemicals were purchased fromSigma (Sigma Chemical Company, St. Louis, MO) and were ofanalytical grade.2.2 Bacterial strainsThe Caulobacter strains used in this thesis are described inTable I. All Caulobacter strains were grown at 30°C andEscherichia coli B and DH5x were grown at 37°C.2.3 Growth mediaPeptone-yeast extract (PYE) medium (Poindexter 1964) wasused for the growth of of all Caulobacter strains unless otherwisespecified and contains per litre; 2 g peptone (Difco Laboratories,Detroit, MI), 1 g yeast extract (BDH Inc., Darrnstadt, Ger.), 0.01% CaC12(BDH), and 0.02% MgSO4 (BDH). Cells grown in PYE medium wereroutinely harvested for experiments at mid-logarithmic growthphase (OD600 = 0.6 - 0.7). Solid media, of all types, contained 1.2%agar (Difco). For growth of freshwater Caulobacter isolates PYEliquid was supplemented with riboflavin at 2 ig/m1.M3Higg medium (Smit et al. 1981) contains 5 mM imidazole -HC1 (pH 7.0), 2 mM KH2PO4 (pH 6.8), 0.3% glucose, 0.3% L-glutamicacid (monosodium salt; pH 7.0), 0.05% NH4C1, and 1% modified Hutner16TABLE I. Bacterial strainsBacterial strain Description or genotype Reference ofsourceC. crescentusCB2A S.1ayer minus variant of wild-type Smit and AgabianCB2, Tpr, Rfr, Apr. 1984CB2NY66R Spontaneous S-layer plus mutant J. Poindexter*of CB2, Tpr, Apr.CB2NY66Rmg1 Calcium-independent mutant of J. Poindexter*CB2NY66R, Tpr, Apr.NA1000 Variant of wild-type strain CB15, Smit and AgabianATCC19O89, synchronous cultures 1984readily prepared from this strain,Tpr, Apr.JS1001 Calcium-independent of NA000. 3. Smit’JS1002 Calcium-independent of NA1000. I. Smit351003 NA 1000 with rsaA interrupted with 3. Smit*KSAC Kmr cassette.JS1004 JS1001 with rsaA interrupted with J. Smit*KSAC Kmr cassette.JS1005 JS1002 with rsaA interrupted with J. Smit*KSAC Kmr cassette.Fresh Water Caulobacter isolates42 FWC strains Isolated from aquatic and wastewater MacRae andsources. Smit, 1991Legend: * unpublished strain, Km’ = kanamycin resistant, Tpr = trimethoprimresistant, Apr = Ampicillin resistant, Rf = Rifampicin resistant17mineral base (Cohen-Bazire et a!. 1957). Cells grown in M3Higg orM 10Higg were routinely harvested for experiments at mid-logarithmic growth phase (0D600 = 2.0 - 3.0). M3Higg and M10Higgmedia were used for physiological studies to determine the cationrequirements for Caulobacter growth and their influence on S-layerstructure. The final metal ion concentration in M3Higg medium is:2.2 mM MgSO4;454 p.M CaC12; 38 p.M ZnSO4; 25.1 p.M FeSO4; 9.1 p.MMnSO4;1.6 p.M CuSO4;0.9 p.M Co(N03)2;0.5 p.M Na2B4O7;and 0.15 p.M(NH)6Mo7O24.M10Higg medium is identical to M3Higg medium except that theHutner mineral base was prepared without CaCI2 and 18 M2 distilled- deionized water, produced by a Barnsted “NANOpure” ultrapurewater system, was used. All containers used to prepare or storeM 10Higg medium were washed with 10 mM Na-EDTA (FisherScientific Co., Nepean, Ont.) (pH 8.0), to remove any trace calcium,then washed with 18 M2 distilled - deionized water. Only new 16 x150 mm S/P® diSPo® culture tubes (Baxter Healthcare Corporation,McGaw Park, IL) were used for growth of cells in M10Higg medium.The agar used for M10Higg medium plates was suspended in 50 mMEDTA (pH 8.0) and allowed to settle. The supernatant was decantedand the procedure was repeated four times. The agar was thenwashed in, the same manner, twice with 18 M2 distilled - deionizedwater. The slurry was then suction-filtered through a hardened -ashless Whatman (Whatman International Ltd., Maidstone, England)541 filter. The agar was then dried under negative pressure in a18dessicator oven at 85°C. Plastic petri dishes (Fisher) were used forsolid M10Higg medium.L medium was used for the growth of E. coil (Miller, 1972) andcontains per L; 10 g tryptone (Difco), 5 g yeast extract and 5 g NaC1.When required antibiotics were added to media at thefollowing concentrations (in p.g/ml): Ap(100) [sodium salt], Km(50)[sulfate salt] (ICN Biomedicals, Inc., Cleveland, OH), Sm(50 or 10 whenpSUP2O21 [Simon et al. 1983] was used in NA1000) [sulfate salt],Tc(10 or 4 when pLAF5 [Keen et al. 1988] was used in JS1004)[chloride salt] (P-L Biochemicals, Inc., Milwaukee, Wis). Allantibiotics were prepared and stored as recommended by Sambrooket al (1989).2.4 Isolation of calcium independent mutantsThe calcium independent mutant CB2NY66Rmg1 was isolated byDr. J. Poindexter (unpublished). The calcium independent mutantmutants JS1001 and JS1002 were isolated by Dr. John Smit(unpublished). See appendix I (method A) for the method by whichthese mutants were selected.2.5 Growth studiesFor studies on the ion requirements of NA1000 and JS1001 thefollowing procedure was used. Cells were grown in M3Higg mediumto mid-log phase (0D600 = 2.0-3.0), harvested by centrifugation, andwashed four times in M10Higg by centrifugation and resuspension.19Five ml of M10Higg (with or without a metal ion [chloride salt]supplement) was inoculated with 5 x 106 washed cells and incubatedat 30°C on a tube roller (VWR Scientific Canada, Ltd., London, Ont) at60 rpm. Plate counts determined that an 0D600 = 1.0 of C.crescentus NA1000 grown in M3Higg medium containsapproximately 1 x iO CFU’s per ml. Growth rates were estimatedbetween 0D600 = 0.100 and Colourimetric assaysAn LKB Biochrom Ultraspec II UV/VIS spectrophotometer wasused for all colourimetric assays. Protein levels were determined bythe method of Markwell et al. (1978) using egg white lysozyme as astandard or using the Bio-Rad protein assay (Bio-Rad laboratories,Missassauga, Ontario), which is based on the method of Bradford(1976), using bovine gamma globulin as a standard. 3-deoxy-2-octulosonic acid (KDO) was estimated by the method of Karkhanis etal. (1978) using authentic KDO (ammonium salt) as a standard.Inorganic phosphate was determined by the method of Ames andDubin (1960) using K2HPO4 as a standard. Sugars were estimated bythe method of Dubois et al. (1956) using D-glucose as a standard.Uronic acids were estimated by the method of Dische (1947) using Dglucuronic acid as a standard.2.7 Isolation and purification of cell surface molecules2.7.1 LPS isolation. C. crescentus strains CB2A and NA1000were grown in PYE as 500 ml cultures, in 2 L Erlenmeyer flasks, on a20rotary shaker (200 rpm) and harvested during late log-phase (0D600= 0.6-0.7). Cells were harvested with a Sorvall RC-5B centrifuge (DuPont Instruments, Wilmington, Delaware) using a GSA rotor (Du Pont)(10,000 xg for 10 mm) and washed once with 0.1 M HEPES(Research Organics, Inc., Cleveland, Ohio) buffer (pH 7.2). LPS wasisolated using a modification of the method of Darveau and Hancock(1983). After nuclease digestion of the disrupted cells, the cell lysatewas made to contain 0.1 M EDTA, 2% SDS, and 10 mM Tris-HC1 (pH8.0) and was then incubated at 37°C for 2 h. The extendedincubation was required to completely dissociate the Caulobactercell membranes. The published procedure was followed untilcompletion of the final ultracentrifugation step. The supernatantfrom this ultracentrifugation step contained “crude” EPS. The LPSpellet was resuspended in 10 mM Tris-HCI (pH 8.0) and washed fivetimes by ultracentrifugation (200,000 xg for 2 h at 15°C), using aBeckman (Beckman Instruments, Inc., Palo Alto, CA) L8-55ultracentrifuge and a Type 6OTi rotor (Beckman), and resuspension.The final pellet was considered to be the “crude1’ LPS fraction.The crude LPS was extracted following the Sonesson et al.(1989) modification of the Galanos et al. (1969) procedure. Thefreeze-dried crude LPS preparation was extracted three times withphenol:chloroform:hexane, 2:5:8, (PCH) (at 8 ml per g of original dryweight of the cells) and centrifuged at 2000 xg. The pooledsupernatants were evaporated using a rotary evaporator and thephenol was removed by dialysis against distilled deionized water (421x 2L). The LPS was recovered by lyophilization and washed threetimes with 10 ml of chloroform:methanol (2:1) and insoluble materialwas pelleted by centrifugation (200,000 xg for 2 h). The precipitatewas dried under a stream of nitrogen gas, dissolved in distilleddeionized water then freeze-dried to yield the “puret’ LPS fraction.2.7.2 EPS isolation. The “crude” EPS obtained during the LPSisolation procedure (see above) was freeze dried, resuspended inDDW to 1/10 the original volume, treated with bovine pancreaticRNase (25 jig/mi at 37°C for 24h), dialyzed against 4L DDW at 4°C for24h and then ultracentrifuged at 200,000 xg (30 hr at 4°C) toremove any remaining LPS. The EPS was then freeze dried,resuspended in 0.1 M pyridinium acetate buffer (pH 7.0) andfractionated by steric-exciusion chromatography (SEC) on a SephacrylS-400 (Pharmacia LKB Biotechnology, Uppsala, Sweden) column (60 x2 cm) using 0.1 M pyridinium acetate buffer (pH 7.0) as eluent.Fractions (2.5 ml) were collected using a Pharmacia FRAC-100fraction collector. Each fraction was analysed for carbohydrate todetermine peak locations.The nature of the cell surface EPS was investigated byassessing the degree to which the carbohydrate remained associatedwith the cells. The yield of EPS from unwashed cells was comparedto yield from cells that were washed 5 times by centrifugation andresuspension in 0.1 M HEPES buffer (pH 7.2).222.7.3 RsaA isolation. C. crescentus NA1000 was used in aseries of experiments to determine an effective method forextracting the S-layer protein from the cell surface. In a typicalexperiment a 100 ml culture was grown in PYE to an 0D600 = 0.6-0.7and harvested by centrifugation (10,000 xg for 10 mm). The cellpellet was washed twice in 1 volume of 10 mM HEPES buffer (pH 7.2)by suspension and centrifugation and then resuspended in 15 ml ofthe same buffer. One ml of the cell suspension was placed into 1.7ml microfuge tubes, pelleted and the supernatant removed. Twohundred tl of one of the following agents was used to resuspend thecells: 10 mM EDTA in 10 mM HEPES buffer (pH 7.5); 10 mM EGTA in10 mM HEPES buffer (pH 7.5); 100 mM HEPES (pH 2, 4, 6, 7.5, 8 or10); 200 mM glycine-HC1 buffer (ICN) (pH 2, 3 or 4); 100 mM TRISbuffer (ICN) (pH 7.2); 0.5% 8-mercaptoethanol (Bio-Rad) in 10 mMHEPES (pH 7.5); 1 M urea (BDH); 1 M guanidine-HC1 (BRL); 10 mMNaCI (Fisher); 10 mM CaC12; and 100 mM HEPES (pH 7.5) withincubation at 65°C. The samples were incubated for 15 mm at roomtemperature (unless otherwise stated) and then the cells werepelleted by centrifugation. Ten il of supernatant was analyzed bySDS-PAGE and the proteins in the gel were visualized by Coomassieblue staining (See below).Subsequently, as a standard method for isolation of S-layerprotein, 5 ml cultures of Caulobacter strains were grown to 0D600 =0.6 and the cells harvested by centrifugation. The cells were washedtwice by centrifugation and resuspension with 5 ml of 10 mM HEPES23(pH 7.2) and then the washed pellet was suspended in 200 j.tl of 100mM HEPES (pH 2) or 200 il of 10 mM EGTA in 10 mM HEPES (pH7.5). The cell suspension was incubated for 10 mm at 20°C, pelletedby centrifugation and the supernatant retained for examination bySDS-PAGE. The acid samples were immediately adjusted to pH 7with 5 N NaOH (BDH).JS1001 and JS1002 produce a macroscopic particulate “debris”in high density cultures. A debris sample (approximately 100 mgwet weight) was collected with a pasteur pipet and suspended in 10mM HEPES buffer (pH 7.2). The material was pelleted in amicrocentrifuge (5 sec, 14,000 rpm) and the supernatant discarded.The pellet was suspended in the same buffer and washed a total ofthree times. The washed pellet was suspended 400 ml of 8 M urea(pH 8.5) at room temperature for 8 h. Insoluble material wasremoved by centrifugation for 10 mill in an Eppendorf centrifugeand the supernatant was dialyzed against 10 mM HEPES (pH 7.2) at4°C.2.7.4 SÃO isolation and purification. SAO was isolated andpurified from JS1003 cultured in PYE. All procedures were carriedout at room temperature (20 - 22°C). Cells were washed bysuspension and centrifugation (10,000 xg; 10 mm) with 20 mMHEPES (pH 7.2). The washed cell pellets were suspended in 0.77 MNaC1 / 0.12 M EDTA (pH 7.2) [at 25 mi/lO g (wet weight) of cells(Kabir 1986)], stirred for 5 mill and then pelleted by centrifugation.The supernatant was saved and the pellet was extracted a second24time. The combined supernatants were ultracentrifuged at 225,000xg. The pellet was resuspended in PBS (consisting of 1.23 g Na2HPO4,0.18 g NaH2PO4 and 8.5 g NaC1 per L) by sonication and dialyzedagainst PBS at 4°C. The extracts, containing approximately 5 mgprotein and 450 pg KDO per ml, were mixed with an equal volume ofSDS-PAGE sample buffer and heated at 100°C for 10 mm then cooledto room temperature. Proteinase K was added to a finalconcentration of 0.5 mg/mi and the sample was placed at 60°Covernight. The sample was then heated at 100°C for 10 mm andfractionated by preparative SDS-PAGE (12 x 14 cm separating gelcast using 1.5 mm spacers). The region of the gel containing SAO wasexcised, using prestained molecular weight markers (Gibco BRL LifeTechnologies, Inc.) as a guide to estimate its location, placed in adialysis membrane [12 - 14 Kda cutoff. (Spectrum)] containing SDSPAGE running buffer and electroeluted at 100 mA. The SAO wasconcentrated and washed extensively with water using an Centricon30 microconcentrator (Amicon Canada Ltd., Oakville, Ont.).2.8 Antisera production2.8.1 Production of antiserum to low pH extracted RsaA.S-layer protein was extracted from NA1000 using 100 mM HEPES(pH 2) as described above. A New Zealand white female rabbit wasimmunized with this preparation, after combining with an equalquantity of Freund’s incomplete adjuvant, by an initial injectioncontaining 1 mg protein and subsequent booster injections at days2521, 28 and 35 with 0.3 mg protein. Sera with the highest titer wascollected on days 55 and 62 and was processed by standard methodsaccording to Heide and Schwick (1978). Serum activity wasdetermined by the Ouchterlony double diffusion assay (Ouchterlony1949) and Western immunoblot analysis. The antisera (c.c-RsaA) wasadsorbed against whole cells of JS1003, and against Westernimmunoblots of JS1003 cell lysate. The pre-immune sera gave noactivity against RsaA by Ouchterlony or Western immunoblot assays.Unless otherwise noted the antibody was used at a concentration of1:50,000 for use in Western blot experiments.2.8.2 Antisera to SAO. Smit and Merker (unpublished)produced a sera that completely labeled the cell surface of S-layernegative C. crescentus strains while not labeling S-layer producingstrains. The epitopes on the cell surface that the antibody recognizedwere unknown. See appendix I (method B) for details on antigenpreparation.2.9 In vitro crystallization of S-layerLow pH and EGTA extracted NA1000 S-layer samples and 5-layer isolated by urea extraction of the macroscopic debris from fromJS1001 cultures were dialyzed [Spectra/Por cellulose dialysis tubingwith a molecular weight cutoff of 12 - 14 KDa (Spectrum MedicalIndustries, Inc., Los Angeles, CA)] overnight against 10 mM HEPES(pH 7.5 at 4°C). Portions of the dialyzed samples were examined by26SDS-PAGE and negative-stain transmission electron microscopy(TEM). The remainder was dialyzed overnight at 4°C against 10 mMHEPES (pH 7.5) containing one of: 1 mM MgC12, 1 mM SrC12, 1 mM, 5mM or 10 mM CaC12. Each sample was then examined by negative-stain TEM.2.10 Electrophoretic methods2.10.1 SD S-PAGE. Samples were analysed by SDS-PAGE usingthe buffer system of Laemmli (1970). Polyacrylamide (Bio-Rad) -bisacrylamide (Bio-Rad) stock solutions contained 29.2 g and 0.8 g,respectively, per 100 ml. Sample buffer consisted of 40% 0.5M TrisHC1 (pH 6.8), 40% glycerol, 4% SDS, 4% 3 mercaptoethanol and 0.005%bromophenol blue (Bio-Rad). Samples loads were normalized byassaying for protein or KDO. When C. crescentus S-layer was to beanalysed by SDS-PAGE, samples were not heated prior toelectrophoresis because little or no S-layer protein will enter the gelif heated to 100°C in sample buffer (Smit and Agabian 1984).Protein molecular weights were estimated using Bio-Rad lowmolecular weight protein standards. If gels were to be used forWestern blotting, prestained protein molecular weight standards(BRL Life Technologies, Burlington, Ontario) were employed. Gelswere stained with: A) 0.1% Coomassie brilliant blue R-250 (Bio-Rad)in fixative (40% methanol [Fisher], 10% glacial acetic acid [Fisher])then destained in fixative or B) one of the silver stains outlinedbelow.272.10.2 Western Blotting. Following SDS-PAGE proteins orcarbohydrates were transferred to nitrocellulose membranes(Schleicher and Scheull, Inc., Keene, N.H.) by the method of Burnette(1981). After blotting, membranes were processed as described bySmit and Agabian (1984). All primary antibody was used at adilution of 1:50,000 for cc-RsaA and 1:20,000 for x-SAO. Goat xrabbit antibody coupled to horseradish peroxidase (Antibodies Inc.,Davis, CA) secondary antibody was used at a dilution of 1:2000. Theblots were developed using 4-chloro-l-naphthol as described bySmit and Agabian (1982b).2.10.3 Sample preparation. Five ml of PYE grown cells (0D600= 0.6-0.7) or 1 ml ofM3Higg/M10i g grown cells (0D600 = 2.0-3.0)were pelleted and washed with 10 mM HEPES (pH 7.2) by suspensionand centrifugation. The pellet was suspended in 250 jil of 10 mMTris-HC1 / 1 mM EDTA (pH 7.2), frozen at -20°C and then thawed atroom temperature. A sample was removed to estimate proteinconcentration. To the remainder of the sample 1 pA of proteinasefree bovine pancreatic DNase (0.5 mg/ml), 20 j.tl of lysozyme (10mg/ml) (ICN), and 3 p.1 of 1 M MgC12 were added and the samplelysate was incubated at room temperature for 15 mm. If lysate wasto be used for detection of LPS, cell lysate containing 1 p.g of KDO wassuspended in 20 p.1 of SDS-PAGE sample buffer, heated at 100°C for10 mm, cooled to room temperature, made to 0.5 mg/mi with28Proteinase K and incubated at 60°C for 1 h. This method is referredto as the modification of the sample preparation method of Hitchcockand Brown (1983) for the qualitative analysis of Caulobacter LPS.Cells were prepared for analysis of LPS by the method ofHitchcock and Brown (1983) as follows. The cells from 5 ml ofculture (0D600 = 0.6-0.7) were pelleted and washed with 10 mMHEPES (pH 7.2) by suspension and centrifugation then resuspendedto a concentration of 200 Klett units (blue filter, Klett-Summersoncolourimeter) in the same buffer. One and one half ml was thenpelleted and the pellet was resuspended in 50 pl of a sampledissociation solution containing 2% SDS, 1 M Tris-HC1 (pH 6.8), 4% 13mercaptoethanol, 10% glycerol, and 0.005% bromophenol blue. Thesample was heated at 100°C for 10 mm, cooled to room temperatureand 10 p.1 of the above sample dissociation solution containing 2.5mg/ml of proteinase K was added. The sample was then incubated at60°C for 1 h.2. 1 1 Silver stainingFollowing electrophoresis gels were stained using the Bio-Radsilver stain kit (Merril et al. 1981), a modification of the BioRadTMsilver stain kit (Cava et al. 1989), the method of Tsai and Frasch(1982) or a modification of the method of Tsai and Frasch.The Bio-Rad silver stain kit was used as directed by themanufacturer. In the modification of the Bio-Rad silver stain kitthe oxidizer solution was replaced with 0.7% sodium metaperiodate(BDH) dissolved in 0.65% isopropanol and 0.26% glacial acetic acid.29Briefly, the method of Tsai and Frasch (1982) involves an overnightfixation of the gel in 40% ethanol-5% acetic acid followed by a 5 mmoxidation with 0.7% periodic acid in 40% ethanol-5% acetic acid.After washing with water the gel is stained for 10 mm in anammonium-silver reagent then washed extensively with water. Thegel was developed at 25°C with a citric acid-formaldehyde solution.This development temperature inhibits the visualization of proteins(Hitchcock and Brown 1983). The Tsai and Frasch (1982) procedurewas modified in that the gels were fixed for only 1 h with twochanges of the fixation solution, the periodic acid oxidation step wasextended from 5 to 15 mm, and the staining step was extended to 20mm.2.12 Cell preparation for thin-section electron microscopyTreated and control cells were pelleted in a microcentrifugetube and resuspended in Burdett’s buffer (Burdett and Murray 1974)[5% acrolein, 0.25% glutaraldehyde (J. B. E. M. Services Inc., PointClaire - Durval, PQ) in 50 mM cacodylate (Electron MicroscopySciences, Fort Washington, PA) buffer (pH 7.4)] and incubated atroom temperature for 1 h then at 4°C overnight. Cells were thenpelleted, washed twice with 50 mM cacodylate buffer (pH 7.4) bycentrifugation and resuspension, and the washed pellet was thenresuspended in 0.8% tannic acid (Mallinckrodt, Inc., Paris, KY) (in 50mM cacodylate buffer [pH 7.4]) and incubated at room temperaturefor 30 mm. The cells were again washed twice in 50 mM cacodylate30buffer (pH 7.4) and enrobed in 2% nobel agar (in 50 mM cacodylatebuffer [pH 7.41). Blocks were post fixed (1% 0s04 [J. B. E. M.] and 0.5mg/ml ruthenium red in 50 mM cacodylate buffer [pH 7.4]) for lh at4°C, washed three times with 50 mM cacodylate buffer (pH 7.4),twice with water and en bloc stained in saturated aqueous uranylacetate (Fisher) for lh. Blocks were then washed twice with waterand dehydrated in an ethanol series, infiltrated with Spurr’s (3. B .E.M.) resin/ethanol [(1:1) for 30 mm], Spurr’s resin [100% for 30 mmtwo times] and finally embedded in Spurr’s resin and heatpolymerized at 65°C for 24h. Thin sections were cut then stainedwith uranyl acetate and Reynolds’ lead citrate (Reynolds 1963).Specimens were viewed in a Siemens lOlA electron microscopeoperating at 80 kV.2.13 Negative stain electron microscopyTo examine colonies for the presence of S-layer, a colony wassuspended in 10 p.1 of water containing 1 p.1 of bacitracin (1 mg/ml)and a carbon-stabilized, parlodion (3. B. E. M.)-coated 400 meshcopper grid was floated on top of the drop for a few mm. To examinesamples for in vitro reassembly of S-layer a grid was placed on adroplet containing RsaA. Grids were then lifted and excess liquidremoved by wicking with filter paper. After drying, the sample wasnegatively-stained using 2% aqueous ammonium molybdate(Mallinckrodt) (pH 7.5) or 2% methylamine tungstate. Specimenswere examined in a Siemens lOlA transmission electron microscope31operated at 60 kV.2.14 Transposon mutagenesis of NA1000A Tn5 library was constructed by the electroporation ofpSUP2O21 (Simon et al. 1983) into electrocompetent NA1000 cells.Transposition events were selected by plating on PYE supplementedwith Km and Sm. pSUP2O21 was isolated from E. coli S17-1 by amini-alkaline plasmid preparation procedure (Sambrook et al. 1989).Electrocompetent NA1000 were prepared as described by Gilchristand Smit (1991). A BioRadTM Gene Pulser, Pulse Controller, andcuvettes with 0.2-cm interelectrodal gaps was used as described bythe manufacturer. The Gene Pulser was set at 2.5 kV and 25 .tF. ThePulse Controller was set at 400 2. Twenty-thousand Km/Smresistant colonies were pooled to form the library.2.15 Complementation of JS1004 for SÃO production withan NA1000 cosmid libraryA NA1000 cosmid library, using pLAF5 (Keen et al. 1988) as avector, was supplied by Dr. L. Shapiro (Stanford University) and waselectroporated into electrocompetent JS1004. Nine hundred and fiftyTcr colonies were inoculated into 96 well microtiter plates containingPYE-Tc using sterile tooth picks and grown for 40 h at 30°C. Two j.tlof each culture, including JS1003 as a positive control and JS1004 asa negative control, was placed on a nitrocellulose sheet and allowedto dry for lh. Sterile DMSO was added to each microtiter well to a32final concentration of 5% and the plates were frozen at -70°C. Thedry nitrocellulose was processed in the same manner as a Westernblot using cz-SAO sera as the primary antibody. One of the 950electroporants reacted positively in the antibody screen. The cosmidDNA was obtained from this clone by an alkaline lysis method(Sambrook et a!. 1989) and the cosmid DNA was electroporated intoE. coli DH5x (BRL Laboratories, Gaithersburg, MD). The cosmid DNAwas isolated by an alkaline lysis method, digested with B amHI andelectrophoresed on a 0.7 % agarose (Bio Rad) gel using a Tris-acetateEDTA buffer system (Sambrook et al. 1989). The DNA fragmentsrunning lower than the top band (pLAF5 plus some CaulobacterDNA) were isolated using a GENECLEAN Il® Kit (Bio 101 Inc., La Jolla,CA). The isolated DNA was 32P labeled by nick translation (Rigby eta!. 1977) and used to screen a cosmid library by colony-blothybridization to identify overlapping cosmids (Maniatis et al. 1983).2.16 Carbohydrate and lipid chemical analysisDetailed chemical analysis on the isolated and purifiedcarbohydrates and lipids were carried out as described byRavenscroft et al. (1991; 1992). The rough LPS and EPS wasanalysed by Dr. N. Ravenscroft and the SAO was analysed by Dr. D. N.Karunaratne.333 Results3.1 Analysis of the S-layer attachment phenotype byWestern blottingSmit and coworkers (manuscript submitted) showed bynegative-stain electron microscopy that calcium-independentmutants of C. crescentus, when grown on calcium-containing PYEplates, produced sheets of assembled S-layer that were notassociated with the bacterial cells. The S-layer gene from thecalcium-independent mutant JS1001 was cloned and expressed inthe spontaneous S-layer minus C. crescentus CB2A and a wild-type,cell-associated, S-layer resulted. Thus although CB2A is S-layernegative it is S-layer “attachment competent”. When the cloned Slayer gene from the wild-type strain NA1000 was introduced intothe calcium-independent, S-layer negative, strain JS1004 non-cellassociated S-layer sheets were produced. These electron microscopystudies determined that the defect in calcium-independent mutants,resulting in the inability to attach the S-layer to the cell surface, wasnot due to a defect in the S-layer protein but in some other cellularlocus. Thus calcium-independent mutants exhibited an S-layer“attachment-defective” phenotype as well as the calciumindependent phenotype for which the mutants were selected.Western blotting experiments were conducted to corroboratethe electron microscopy observations of Smit and coworkers. S-layerattachment-defective mutants could be distinguished from wild-type34FIG. 1. Western blot analysis of whole cell lysates reacted withunadsorbed anti-RsaA antisera. “L” indicates cells were grown in aliquid culture and washed prior to analysis. “P” indicates cells weregrown on plates and were not washed prior to analysis. Lanes: 1,prestained molecular mass markers; 2, NA1000 (L); 3, JS1003 (P); 4,JS1003/pKT23O-A19 (L); 5, JS1003/pKT23O-Ca5 (L); 6, JS1002 (L);7, JSlOOl (P); 8, JS1004 (P); 9, JS1002 (P); 10, JS1005 (P); 11,NA1000 (L); 12, JS1003/pKT23O-CB2AD (L); 13, JS1005/pKT23O-A19 (L); 14, JS1005/pKT23O-A19 (P); 15, prestained molecular massmarkers; 16, CB2A (P); 17, CB2A/pKT23O-A19 (L); 18,CB2A/pKT23O-CalO (L); 19, CB2A/pKT23O-Ca5 (L); 20,CB2A/pKT23O-CB2AD (L); 21, JS1004/pKT23O-CalO (P); 22,JS1004/pKT23O-CalO (L); 23, prestained molecular mass markers;24, CB2NY66R (L); 25, CB2NY66Rmg1 (L); 26, CB2NY66Rmg1 (P). Onlythe region of the gel containing S-layer is shown. The prestainedmolecular mass markers are 97.4 and 200 kDa. Lanes were loadedwith samples containing 5 tg of protein as estimated by the methodof Markwell et al. (1978). Samples were fractionated on a resolvinggel containing 10% acrylamide prior to blotting onto nitrocellulose.Unless otherwise noted this and all subsequent polyacrylamide gelsutilized a stacking gel containing 4 % acrylamide. The discontinuousbuffer and sample dissociation method of Laemmli (1970) was usedin this and all subsequent polyacrylamide gels as described in theMaterials and Methods.3511 12 13 14 15 16 17 18 19 2021 22 23 24 25 2636strains on the basis of how the cells were prepared forelectrophoresis. S-layer protein could be detected in colonies ofcalcium-independent mutants only when the mutant cells werescraped directly from plate cultures (Fig. 1; lanes 7, 9, 14, 21 and 26)whereas washed liquid-grown calcium-independent cells werealmost negative for RsaA (Fig 1; lanes 6, 13, 22 and 25). Westernblots of plate-grown strains in which the chromosomal S-layer genehad been deleted or which were spontaneous S-layer negativemutants were RsaA negative (Fig. 1; lanes 3 and 16). Washed liquidmedium-grown cells of strains that were S-layer negative butattachment-competent produced a positive S-layer blot when anycloned rsaA gene, including those from attachment-defectivestrains, was expressed in the cell on a plasmid vector (Fig. 1; lanes 4,5, 12, 17, 18,19 and 20). When rsaA, cloned from an attachment-competent strain, was expressed in an attachment-defectivebackground the attachment-defective phenotype persisted in that Slayer protein was detected only when the cells were harvested fromplate cultures (Fig 1; lane 14). If the same strain is cultured in liquidmedium RsaA is not detected (Fig. 1; lane 13). These resultscorroborated the electron microscopy observations of Smit andcoworkers (manuscript submitted) which demonstrate that thedefect in “calcium-independent” mutants, resulting in the inability toattach the S-layer to the cell surface, was not due to a defect in theS-layer protein but in some other cellular locus.373.2 Isolation and purification of EPS from CB2A, NA1000and JS1001An EPS was identified on the surface of wild-type and calcium-independent Caulobacter strains. The EPS from mutant and wild-type strains was isolated and characterized to determine if analteration in the EPS was responsible for the S-layer attachment-defective phenotype.Fractionation, by steric-exciusion chromatography, of the crudeEPS isolated from NA 1000 and CB2A yielded similar profiles for bothstrains (Fig. 2). JS1001, a calcium-independent mutant of NA1000,produced the same fractionation profile as NA 1000 (not shown).Carbohydrate analysis showed that the void volume peak (A)contained a heteropolysaccharide whereas the second peak (C)contained only ribose. Additionally, the second peak (C) containedsignificant amounts of phosphate and had a maximum absorbance at260 nm so the peak was attributed to undegraded RNA. The CB2Acrude EPS also contained an additional minor peak (B) between theheteropolysaccharide (A) and the RNA (C) peaks. On basis ofexclusion limits reported for Sephacryl S-400 the appearance of theEPS peaks in the void volume indicates a minimum molecular massof 1-2 million daltons.3.2.1 Assessing the degree of EPS cell association.Bacterial cell surface carbohydrates are qualitatively categorized astrue capsules I BPS or as slime layers based on their ability to38maintain cell association. True capsules maintain cell associationduring growth or when the cells are subjected to mild sheer forceswhereas slime layers are easily detached from the cell (Boulnois andRoberts 1990). In one trial of the EPS isolation procedure NA1000,JS1001 and CB2A cells were washed five times by centrifugationwith 0.1 M HEPES buffer (pH 7.2) before extracting the EPS todetermine if the EPS could be readily washed from the cells. Theyield of EPS from this experiment was not significantly different thanthat of cells which had been washed only once or not at all (notshown). The culture supernatants from 500 ml batch cultures ofCB2A, JS1001 and NA1000 were freeze-dried, dialyzed against water,freeze-dried again and analyzed for carbohydrate by gaschromatography. The analysis did not produce a sugar profilesimilar to that produced by purified EPS (see below). It wastherefore concluded that the EPS produced by the 3 strains wassignificantly adherent to the cell.3.2.2 Chemical characterization of EPS from CB2A,NA1000 and JS1001. The isolated EPS was found to be in apurified state suitable for detailed structural analysis (see appendixII, Fig. 2.). Dr. Neil Ravenscroft concluded that the EPS isolated fromwild-type C. crescentus NA 1000 and the calcium-independentmutant JS1001 were identical. The EPS produced by C. crescentusCB2A differed from that of NA1000.390.3ACC02 a’o.a’o 0.1- .C0B z0.0 • •I 0.00 5 10 15 20 J’25Bluedextran traction No. GlucoseFIG. 2. Fractionation of EPS of CB2A () and NA 1000 (+) onSephacryl S-400. Carbohydrate monitored by the phenol-sulfuricacid assay as described in Materials and Methods (optical density at490 nm). Peak A contains CB2A and NA1000 EPS fraction. Peak Bcontains a minor polysaccharide found only in CB2A. Peak C containsRNA. Each fraction consisted of 2.5 ml.403.3 Isolation and purification of LPS from CB2A, NA1000and JS1001The LPS from mutant and wild-type Caulobacter strains wascharacterized and isolated to determine if an alteration in the LPSwas responsible for the S-layer attachment-defective phenotype.3.3.1 Electrophoretic analysis of LPS. SDS-PAGE of thepurified LPS and subsequent staining, by the method of Tsai andFrasch (1982) or using the Bio-Rad silver stain kit, revealed that theC. crescentus strains produced a rough LPS (Fig. 3). Bands of similarmobility were detected in purified LPS preparations (Fig 3, lanes 3 -6) No high molecular weight morphological heterogeneous forms,typical of smooth LPS, were noted. LPS profiles of these strainsprepared by the method of Hitchcock and Brown (1983) also showedno high molecular weight bands (Fig. 3, lanes 1 and 2) indicating thatthe isolation method did not select against recovery of smooth LPSspecies. The heterogeneity of the LPS bands is typical of themicroheterogeneity found in “rough” LPS species when examined bySDS-PAGE or TLC (Nowotny 1984). The samples prepared by theHitchcock and Brown method (Fig. 3; lanes 1 and 2) produced bandsthat were broader in the horizontal plane and more condensed in thevertical plane than the bands produced by the purified samples.This is most likely a result of the Hitchcock and Brown samplescontaining bulk cellular components, such as undigested protein,peptidoglycan and nucleic acids which slightly alter the mobility ofthe LPS through the polyacrylamide gel.The electrophoretic profile of a number of C. crescentus411 2 3 4 5 6FIG. 3. SDS-PAGE of LPS prepared by the methods of Hitchcock andBrown (1983) [lanes 1 and 2] and Darveau and Hancock (1983)[lanes 3 to 61. Gels were stained by the method of Tsai and Frasch(1982) [lanes 1 to 41 or by using the Bio-Rad silver stain kit [lanes5 and 6]. Odd-numbered lanes contained LPS from CB2A; evennumbered lanes contained LPS from NA1000. All lanes were loadedwith samples containing 1 p.g of KDO as estimated by the method ofKarknanis et al. (1978). The resolving gel contained 14% acrylamide.42strains, including the calcium-independent strains JS1001 andJS1002, were examined by the method of Hitchcock and Brown(1983) (Fig. 4). The LPS from all strains ran at the dye front of thegel. Purified LPS from JS1001 and JS1002 produced the samebanding pattern (gel not shown).3.3.2 Isolation of LPS, monitored by lipid analysis. Thecold ethanol extraction procedure of Darveau and Hancock (1983)yielded the crude LPS fraction. Analysis of the fatty acids of thecrude LPS showed the presence of saturated and mono-unsaturated16- and 18-carbon fatty acids together with 3-OH-dodecanoic acidindicating contamination by phospholipids (Fig. 5A; [Lelts et al.19821). Extraction of the “crude” LPS by PCH and chloroform-methanol yielded a “pure” LPS product that is largely free of theseC16 and C18 fatty acids (Fig. 5B). Lipid analysis of the solublecontaminates from the PCH and chloroform-methanol extractionsteps revealed negligible amounts of 3-OH-dodecanoic acid indicatingthat little LPS was lost during these purification steps (see below).The material extracted from the final “pure” LPS was also examinedby SDS-PAGE and stained with the Bio Rad’ silver stain kit andCoomassie blue. The gels showed that residual protein but notcarbohydrate was extracted. Coomassie blue stained gels showedthat a pronase-resistant protein of approximately 31 kDa was themajor protein contaminant in the “crude” LPS (data not shown). Thefinal freeze dried “purified” LPS was approximately 40% (by weight)43FIG. 4. SDS—PAGB of LPS prepared by the method of Hitchcock andBrown (1983) and stained by the method of Tsai and Frasch (1982).Lanes: 1, CB2A; 2, CB2NY66R; 3, NA1000; 4, JS1003; 5, JS1002; 6,JS1005; 7, JS1001; 8, JS1004; 9, CB2NY66Rmg1. All lanes wereloaded with samples containing 0.5 .tg of KDO as estimated by themethod of Karknanis et a!. (1978). The resolving gel contained 12.5%acrylamide.1 23444lighter that the “crude” LPS indicating that the two extractionprocedures removed substantial amounts of impurities. The PCHextraction resulted in a 10% decrease in the dry weight of the “crude”LPS whereas the chloroform-methanol wash decreased the weight by30%.3.3.3 Colourimetric analysis. The “pure” LPS was analyzedfor protein, phosphate and KDO. No protein was detected andphosphate and KDO were found to account for 0.5% and 12% of theLPS dry weight, respectively. The molar ratio of phosphate to KDO,determined colourimetrically, was approximately 1:3. When the LPSwas analysed for KDO under more severe hydrolysis conditions, thanthat recommended by Karkhanis et al. (1978), no increase in theamount of KDO per dry weight LPS was noted indicating that the KDOwas likely not phosphorylated (Caroff et al. 1987).The total amount of thiobarbiturate-positive material (KDO)contained in freeze dried cells of CB2A, NA1000 and E. coli B wasalso examined. For these studies identical weights of the three celltypes were simultaneously analyzed along with KDO standards. TheCaulobacter strains were found to contain less KDO per dry weightthan E. coli B; the ratio being E. coli B:CB2A:NA1000 = 1:0.83: Detailed chemical analysis. The purified LPS wasfound suitable for detailed chemical analysis. Dr. Neil Ravenscroftconcluded that LPS isolated and purified from wild-type C.crescentus strains CB2A and NA1000 (Ravenscroft et al. 1992) and45AHLLBFIG. 5. A GC trace of fatty acid methyl esters from the LPS ofNA1000 prepared and chromatographed as described in theMaterials and Methods. (A) Crude LPS preparation prior to organicextractions. (B) Pure LPS resulting from extraction of the crudesample with phenol/chloroform/hexane followed bychloroform/methanol as described in the Materials and Methods. S =methyl ester of octadocanoic acid as an internal standard. The crudeand purified LPS was prepared for GC and the major peak remainingin the purified LPS was identified as 3-OH-dodeconate (asdetermined by GC-MS) by Dr. Neil Ravenscroft.S46the calcium-independent mutant JS1001 (Ravenscroft, unpublished)were structurally and chemically identical (see appendix II, Table Iand II).3.4 Identification of an S-layer associated oligosaccharideDuring one experiment to isolate LPS from NA1000, by theDarveau and Hancock procedure (1993), an additionalpolysaccharide-containing molecule that migrated more slowly thanLPS on SDS-PAGE was detected (Fig. 6). This molecule (which wassubsequently referred to as the S-layer associated oligosaccharide orSÃO) did not stain with the Tsai and Frasch silver stain procedure forLPS but was reliably stained using the more general Bio-Rad silverstain kit (Fig. 6; lane 1). [Note: It will be determined that the SAO isa smooth LPS species.] When this LPS preparation was examined byWestern blot analysis, using anti-SÃO sera (see Materials andMethods), the SÃO band was specifically labeled (Fig. 6; lane 2). Noreactivity was seen with the rough LPS. The various steps in theDarveau and Hancock procedure were monitored by SDS-PAGE withsilver-staining and immunoblotting to determine where the SÃO waslost. It was found that the SÃO and significant amounts of rough LPSremained in the supernatant following the cold ethanol-Mg2+precipitation procedure (data not shown).The rapid LPS analysis method of Hitchcock and Brown did notreveal the SÃO along with the rough LPS (Fig. 3 and 4). Amodification of both the Hitchcock and Brown sample preparation47FIG. 6. Electrophoretic analysis of purified NA1000 LPS containingcontaminating SAO. Lane 1: SDS-PAGE of LPS stained using the BioRad silver stain kit. Lane 2: Western blot of LPS, fractionated bySDS-PAGE, reacted with cc-SAO sera and visualized as described inthe Materials and Methods. LPS was prepared by the method ofDarveau and Hancock (1983) and fractionated using a resolving gelcontaining 13% acrylamide. Both lanes were loaded with samplescontaining 0.5 jig of KDO as estimated by the method of Karknanis eta!. (1978). The arrow indicates the region of the gel containing SAO.48procedure and the Tsai and Frasch staining process was developed inorder to visualize SAO in polyacrylamide gels. This modifiedprocedure was used to examine a number of Caulobacter strains(Fig. 7A). Wild-type S-layer producing strains, NA1000 andCB2NY66R, and strains that produce a wild-type S-layer when rsaAis expressed on a plasmid (CB2A and JS1003), contained thisadditional polysaccharide (Fig. 7A; lanes 1, 4, 5, and 6). Acomparable band was not detected in the S-layer attachment-defective strains JS1001 and JS1002 (Fig. 7A; lanes 2 and 3). Eightother calcium-independent strains isolated by Smit (unpublished)were also examined by these methods and none contained the SAOband (data not shown). The modified procedure of Tsai and Fraschrevealed the SÃO in whole cell lysates of S-layer attachment-competent C. crescentus strains (Fig. 7A, lanes 1, 4- 6) althoughincreased sample loadings were required to that which is normallyused to detect LPS (note the overloading of the rough LPS in Fig. 7A).Tsai and Frasch (1982) recommends loading 1 to 5 .tg of LPS per laneto obtain a satisfactory LPS profile. Western blotting of thesepreparations, using anti-SÃO sera, showed that the SÃO of all strainsproducing the polysaccharide cross reacted with the antibody whichwas raised against CB2A cell membranes (Fig. 7B). JSlOOl andJS1002 showed only a faint immunoreactive band in the SÃO region.These experiments were repeated using cells directly removed fromplates and identical results were obtained (data not shown). Thisconfirmed that the SÃO was not sloughed off the cell surface of49FIG. 7. A) SDS-PAGE of whole cell lysates treated with proteinase Kas described in the Materials and Methods, using a modification ofthe method of Hitchcock and Brown (1983), and fractionated using aresolving gel containing 13% acrylamide. The gel was stained using amodification of the method of Tsai and Frasch (1982) as described inthe Materials and Methods. B) Western blot of samples shown in“A” reacted with c-SAO sera and visualized as described in Materialsand Methods. Lanes in “A” and “B “: 1, NA1000; 2, JS1001; 3,JS1002; 4, JS1003; 5, CB2A; 6, CB2NY66R. All lanes in “A” and “B”were loaded with samples containing 0.75 j.tg KDO as estimated bythe method of Karknanis et a!. (1978). The arrow in “A” and “B”indicates the region containing SAO.50It)calcium-independent mutants.3.5 Isolation and purification of SÃO3.5.1 Extraction of cell surface molecules. C. crescentus iskilled by the concentration of Na found in phosphate-bufferedsaline (PBS) (Poindexter 1964). When cells incubated in PBS wereexamined by negative-stain electron microscopy small vesicles wereobserved to be extracted from the cell surface. The isolated vesicleswere shown to contain SAO when examined by Western blottingusing anti-SÃO sera and the amount of material extracted from thecells increased if the PBS was supplemented with 10 mM EDTA (datanot shown). Multiple extractions were required to completelyremove all of this material from the cells and resulted in an extractthat contained a low concentration of SÃO. When 0.77 M NaC1 / 0.12M EDTA (pH 7.2) [25 ml/10 g (wet weight)] was used, in place of PBS/ 10 mM EDTA, 95% of the cellular KDO was solubilized with twoextractions. Therefore this method was used to isolate the SÃO.3.5.2 Examination of extract. The extract was subjected toultracentrifugation and the pelleted material from NÃ1000 wasshown to contain a protein/KDO at a ratio of 19.2 while pellets fromJS1003 extract had ratio of 10.7. Proteinase K treatment of thepelleted extract followed by SDS-PÃGE with silver-staining using theBio Rad kit and Western blotting using anti-SÃO sera revealed thatthe pelleted extract contained rough LPS and SAO (Fig. 8). Coomassie5297,40066,20045,00031,00021,500__________14,4001 2FIG. 8. Electrophoretic analysis of proteinase K-treated NaCl/EDTAextract of NA1000. Lanes: 1, Western blot, of the polyacrylamide gelshown in lane 2, reacted with x-SAO sera and visualized as describedin the Materials and Methods. 2, SDS-PAGE stained using the BioRad’ silver stain kit. 3, Molecular mass standards in daltons. Lanes1 and 2 were loaded with samples containing 0.5 pg KDO asestimated by the method of Karknanis et al. (1978). Samples werefractionated on a resolving gel containing 12% acrylamide. SAO is theband that runs adjacent to the 45,000 molecular mass marker. Therough LPS runs at the dye front.4-4-43534 97,400— 4 66,2004 45,0004 31,000_____4 21,50014,4002 3FIG. 9. Electrophoretic analysis of NaC1/EDTA extract of NA1000.Lanes: 1, Western blot, of the polyacrylamide gel shown in lane 2,reacted with cc -RsaA sera and visualized as described in theMaterials and Methods. 2, SDS-PAGE stained using Coomassiebrilliant blue R-250. 3, Molecular mass standards in daltons. Lanes1 and 2 were loaded with samples containing 3 and 10 t g,respectively, as estimated by the method of Markwell et al. (1978).Samples were fractionated on a resolving gel containing 12%acrylamide.154blue staining of proteinase K treated extracts detected the minorband running above the SAO in lane 2 of Fig. 8 (data not shown).This band most likely represented the 66 kDa protein found in thenon-proteinase K treated extracts (see below). The proteinscontained in the pelleted extracts were examined by SDS-PAGE withCoomassie blue staining and Western blotting using anti-RsaA sera(Fig. 9). Many bands were detected with the major protein bandsbeing approximately 20 kDa, 66 kDa and 105 kDa (RsaA). The 105kDa protein reacted with -RsaA sera confirming that it was the Slayer protein.The supernatant following ultracentrifugation of the NA1000extract was found to contain a high concentration of carbohydrate.Chemical analysis of this purified carbohydrate indicated that it wasthe EPS (Ravenscroft et al. unpublished). Figure 10 shows that theBPS was not contaminated with RNA as was the case when the EPSwas isolated by the method of Darveau and Hancock (1983) beforethe SEC chromatographic step.3.5.3 Examination of extracted cells. Extracted and controlcells were prepared for and examined by thin-section electronmicroscopy (Fig. 1 1A and B). The extracted cells remained intact, nobreeches in the peptidoglycan layer were noted, and the bilayerappearance of the outer and cytoplasmic membrane was maintained.Two major ultrastructural changes were noted. The electron-densematerial between the peptidoglycan and the outer membrane wasextracted from treated cells (Fig. 1 1A and B; arrow in inset) and55A sRFBFIG. 10. GC trace of alditol acetates from EPS of NA1000 isolatedusing: A. The method of Darveau and Hancock (1983). B.NaC1/EDTA extraction. R = ribose; F = fucose; S = inositol standard.The same amount of inositol standard was included in both samples.The EPS was prepared for GC by Dr. Neil Ravenscroft.56FIG. 11. Thin-section TEM micrograph of: A. Control NA1000 cell. B.NaC1/EDTA extracted NA1000 cell. Note the even distribution ofchromosomal material and ribosomes in control cell (A) and therearrangement of cytoplasmic constituents into two distinct regionscontaining ribosomes or chromosomal material (B). Arrow in (B)shows ribosome-free region containing chromosomal material.Arrow in the inset figures illustrates electron-dense materialbetween peptidoglycan layer and inner leaflet of the outermembrane (A) which is absent in extracted cells (B). All of 100longitudinally sectioned control and extracted cells examined showedthe features demonstrated in this figure. Bar = 0.5 .im. Micrographtaken by Mr. S. H. Smith.57A.B.T58there was a redistribution of the nuclear material within thecytoplasm of these cells (Fig. 11B; arrow in main figure). Thisredistribution has been proposed by Whitfield and Murray (1956) tobe due to loss of the plasma membrane losing its selectivepermeability barrier to monovalent cations. A total of 100longitudinally sectioned control and extracted cells were examinedand all of the extracted cells contained the two ultrastructuralchanges.3.5.4 Purification of the SAO. Proteinase K treated extractsfrom JS1003 were subjected to SDS-PAGE and the region of the gelcontaining SAO was removed. The SAO was extracted from theacrylamide by electroelution then concentrated and washed byAmiconTM filtration. A portion of the SAO was subjected to SDS-PAGEto determine its purity. Figure 12 shows that it was free ofcontamination by protein and rough LPS. It was noted that althoughboth the SAO and rough LPS were stained by the Bio-Rad silver-stain kit this method was approximately five-fold less sensitive forthe detection of rough Caulobacter LPS in comparison to the methodof Tsai and Frasch (1982) (Gel not shown).Colourimetric analysis indicated that the purified SAOcontained, per mg dry wt; 25 j.ig KDO, 454 pg phenol-sulphuricpositive carbohydrate, 1 pg phosphate and 4.7 .tg of uronic acid. On amolar basis the KDO:Pi ratio was 3:1; the same ratio as that found inthe rough LPS. Detailed chemical analysis of the SAO, preformed byDr. D. N. Karunaratne, indicated the only major lipid present was the59-2 3FIG. 12. SDS-PAGE of carbohydrates, purified from NaC1/EDTAextracts from JS1003, using a resolving gel containing 12%acrylamide and stained using the Bio-Rad silver stain kit. Lanes: 1,purified rough LPS; 2, protein molecular mass standard; 3, purifiedSÃO. Lanes 1 and 3 were loaded with samples containing 1.0 pg and0.1 .tg of KDO, respectively, as estimated by the method of Karknaniset al. (1978). Molecular mass standards are, from top, 66.2, 45, 31,21.5, and 14.4 kDa. Note: The 31 kDa marker is poorly stained.60esterified fatty acid 3 -OH-dodecanoate. Carbohydrate analysisindicated three major sugars and a number of minor sugars (seeappendix II, Table III). The minor sugars were mannose, glucose,galactose and heptose which have been detected in the rough LPS.The major sugars were tentatively identified as 4,6-dideoxy-4-aminohexose, 3,6-dideoxy-3-amino hexose and glycerol.3. 6 Purification of C. crescentus S-layer proteinIt was of interest to determine a rapid and effective method toextract and isolate a relatively pure preparation of RsaA from wholecells. With purified RsaA it will enable both the in vitrocrystallization experiments and the production of polyclonal antiseraagainst RsaA.3.6.1 Extraction of the S-layer of C. crescentus NA1000.Of the agents tested, 100 mM HEPES at pH 2 was the most effectiveat extracting RsaA with the least contamination from other proteins.By Coomassie blue staining the preparations appeared to containnearly pure RsaA in this single step purification procedure (Fig. 13).The smearing of RsaA above the major band in the immunoblotshown in Figure 13 is commonly seen in acrylamide gels stained withCoomassie blue if sufficient RsaA protein is loaded on the lane. Thisis assumed to be a consequence of RsaA aggregating or polymerizingbefore or during electrophoresis (Smit and Agabian 1984). RsaA alsostained poorly in polyacrylamide gels by Coomassie blue stainingmethods presumably due to the low content of basic amino acids in61200,00097,400-4- 68,00043,000129,000FIG. 13. Electrophoretic analysis of proteins extracted from wholecells of NA1000 using 0.1 M HEPES (pH = 2.0). Lanes: 1, SDS-PAGEusing a resolving gel containing 12% acrylamide and stained withCoomassie brilliant blue R; 2, Western blot of gel shown in lane 1reacted with x -RsaA sera and visualized as described in theMaterials and Methods. 3, Molecular mass markers in kDa. Lanes 1and 2 were loaded with samples containing 3 jig of protein asestimated by the BioRadTM protein assay.62the protein (Wilson 1983). Lysine, histidine and arginine account foronly 2.3% of the total of amino acids in RsaA (Gilchrist et al. 1992).Occasionally, a minor amount of protein migrating with a fasterelectrophoretic mobility than RsaA was noted in low pH extracts.EGTA treatment also efficiently removed RsaA although thepreparations were more contaminated with other protein species.Other methods were less effective: HEPES at pH 4 extracted RsaA aswell as a number of other proteins while HEPES at pH 6, 7.5, 8, and10 did not extract RsaA. Glycine-HC1 at pH 2 yielded RsaA as aprominent protein but significant amounts of lower molecular weightproteins were also present. Glycine-HC1 treatment at pH 3 and 4showed further increases of other proteins. Similarly, 65°Ctreatment produced a prominent RsaA protein band but many lowermolecular weight proteins were also present. Guanidine-HC1, urea,Tris (pH 7.2), 8-mercaptoethanol and EDTA all extracted numerousproteins without RsaA predominating. NaCI and CaC12 treatments didnot yield significant amounts of protein.The macroscopic precipitate formed in high density cultures ofJS1001 was extracted with 8 M urea for 8 hrs and particulate matterwas then removed by centrifugation. The urea was removed bydialysis and the solubilized protein was examined by SDS-PAGE andCoomassie blue staining (Fig. 14). Western blotting using anti-RsaAsera confirmed that the major protein species was RsaA (not shown).6397,40066,200--45,000 -_____31,000a1FIG. 14. SDS-PAGE using a resolving gel containing 10% acrylamideand stained with Coomassie brilliant blue R. Lanes: 1, Molecularmass standards. 2, Protein solubilized by urea from macroscopicprecipitates produced in high density cultures of JS1001. Lane 2 wasloaded with a sample containing 5 jig of protein as estimated by theBio-Rad protein assay.2643.6.2 In vitro crystallization of the isolated NA1000 Slayer. After dialysis overnight against 100 mM HEPES buffer (pH7.5), protein samples extracted by the EGTA or low pH methods hadno visible turbidity and TEM negative-stain analysis showed onlyamorphous structures (not shown). Dialysis of the sample against 1mM MgC12, or 1 mM SrC12 did not promote crystallization of the Slayer protein. After overnight dialysis against 1 mM CaC12 thesample became turbid and TEM showed that the protein hadcrystallized into a regularly structured array of hexagonal symmetry(Fig. 15) with center-to-center spacing comparable to the native Slayer. Higher concentrations of CaCl2 also promoted turbid solutionsbut ordered S-layer regions were much more difficult to detect byTEM. RsaA solubilized by urea from the macroscopic precipitatesformed in cultures of JS1001 did not recrystallize.3.63 Anti-RsaA sera. The anti-RsaA sera was used at adilution of 1:100,000 for a Western blot of whole cell lysates ofNA1000 (Fig. 16). The blot indicated that the low pH extracted RsaAused as an antigen was not pure but contained a number of otherproteins. Preadsorption of the antisera with cell lysates of JS1003effectively removed antibody activity to proteins other than RsaA.The urea solubilized protein from JS1001 reacted with the anti-RsaAsera by Western blotting indicating that although it failed torecrystallize into a regular array it was in fact RsaA. The adsorbedRsaA antisera bound to NA1000 cells and did not bind to JS1003 or65FIG. 15. Negative-stain transmission electron micrograph of the Invitro recrystallization of RsaA. RsaA was purified by low-pHextraction from NA1000 cells. Recrystallization was mediated bydialysis of monomeric RsaA against 10 mM HEPES buffer (pH = 7.5)containing 1 mM CaC12 at 4°C for 18 hrs. Bar = 0.1 $.im. This electronmicrograph was taken by Mr. S. H. Smith.66FIG. 16. Western blot reacted with unadsorbed x-RsaA sera andvisualized as described in the Materials and Methods. Lanes: 1,Whole cell lysate of NA1000; 2, Low-pH extracted protein fromNA1000. Lanes 1 and 2 were loaded with samples containing 10 and1 .tg of protein, respectively, as estimated using the Bio-Rad proteinassay. The samples were fractionated by SDS-PAGE using a resolvinggel containing 12% acrylamide.1 267CB2A cells in indirect immunoflurescent microscopy experiments(Smith and Smit, unpublished).3.7 Comparison of S-layers among freshwaterCa ul oh a c te rsIt was of interest to determine if the cell surface ofenvironmental Caulobacter isolates had the same general characteras the laboratory strains. It will be shown that most environmentalCaulobacter isolates produce an S-layer, rough LPS and SÃO.3.7.1 S-layer extraction. The HEPES (pH 2.0) extractionmethod was applied to all of the freshwater Caulobacter (FWC)strains and, in general, proved to be a useful technique to specificallyextract the S-layer proteins. That is, only a single major highmolecular weight band, characteristic of S-layer proteins fromlaboratory strains of Caulobacter, was seen by SDS-PAGE. The SDSPAGE profiles of low pH extracts yielding an S-layer like band areshown in Figure 17A. The SDS-PAGE profiles of low pFT extracts fromFWC3O, -38, -40, and -43 did not show a prominent S-layer like bandand are shown in Figure 17A (lanes 2, 8, 30 and 31) for comparison.Strains from which an S-layer band could not be extracted using theHEPES (pH 2.0) method were extracted using HEPES buffer (pH 7.5)containing 10 mM EGTA. This treatment extracted a prominentprotein in almost every case (Fig. 17B). FWC5, -14 and -21 (Fig. 17B;lanes 17, 15 and 14 respectively) did not produce an extract showinga prominent S-layer like band using HEPES buffer (pH 7.5) containing68FIG. 17A. SDSPAGE of low pH extracted proteins from Caulobacterstrains. Lanes: 1, molecular mass markers; 2, FWC38; 3, FWC28; 4,FWC33; 5, FWC35; 6, FWC31; 7, FWC1; 8, FWC43; 9, NA1000; 10,FWC15; 11, molecular mass markers; 12, FWC2; 13, FWC44; 14,FWC19; 15, FWC17; 16, FWC2O; 17, NA1000; 18, FWC37; 19,NA1000; 20, molecular mass markers; 21, FWC16; 22, FWC26; 23,FWC18; 24, FWC22; 25, FWC25; 26, FWC46; 27, FWC11; 28, FWC23;29, NA1000; 30, FWC4O; 31, FWC3O; 32, FWC39; 33, FWC27; 34,FWC24. The gels were stained with Coomassie Brilliant blue and theresolving gel contained 10% acrylamide. Samples contained 3 jig ofprotein as estimated by the Bio-Rad protein assay with theexception of FWC26, -30, -38, -40, -43, and -46 (see text). Molecularmass markers are: 200, 97.4, 68, 43, and 29 kDa.69A-—————8 9 10 11 12 13 14 15 16 1718 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34F70FIG. 17B. SDS-PAGE of EGTA-extracted proteins from Caulobacterstrains. Lanes: 1, molecular mass markers; 2, NA1000; 3, FWC32; 4,FWC12; 5, FWC42; 6, FWC7; 7, FWC6; 8, FWC9; 9, FWC29; 10, FWC41;11, FWC45; 12, molecular mass markers; 13, NA1000; 14, FWC21;15, FWC14; 16, FWC4; 17, FWC5; 18, FWC8; 19, FWC34. The gelswere stained with Coomassie Brilliant blue and the resolving gelcontained 10% acrylamide. Samples contained 3 tg of protein asestimated by the Bio-Rad protein assay with the exception of FWC5,-14, -21, and -41 (see text). Molecular mass markers are: 200, 97.4,68, 43, and 29 kDa.71w:“‘*10 mM EGTA or the low pH method. As well, FWC -30, -38, -40 and-43 did not produce an extract containing a prominent S-layer likeband by using HEPES buffer (pH 7.5) containing 10 mM EGTA (notshown) or the low pH method (see above).Extracts containing 3 tg of protein were examined by SDS-PAGEin Figures 17A and 17B with the exception of FWC5, -14, -21, -26, -30, -38, -40, -41, -43 and -46. Extraction of these strains by bothmethods resulted in extracts containing very low concentrations ofprotein. Therefore 3 pg of protein could not be loaded on the SDSPAGE for these strains so 15 j.tl of extract from the method yieldingthe highest protein concentration was used.3.7.2 Western blot analysis of extracted proteins. Theprotein samples used for the Coomassie blue stained gels wereanalyzed by Western blotting using the anti-RsaA serum (Fig. 18).All of the broad S-layer-like bands in Figure 17 gave a positivereaction with the exception of FWC23 (see Fig 17A; lane 28 andFig.l8; lane 28). The antiserum was quite specific for the suspectedS-layer bands. Western blots of EGTA-extracted samples, whichsometime contained non-S-layer proteins, only detected oneprominent high molecular weight band.FWC5, -14, -21, -30, -38, -40, and -43 did not produce an 5-layer (MacRae and Smit 1991) and did not show a high molecularweight band by Coomassie blue staining (See Fig. 17A; lanes 2, 8, 30,and 31 for FWC38, -43, -40, and -30 respectively and Fig 17B; lanes73FIG. 18. Western blot analysis of proteins extracted fromCaulobacter strains using the low pH (A) or EGTA (B) method. Blotswere reacted with cL-RsaA serum. The SDS-PAGE resolving gelcontained 10% acrylamide. (A) Lanes: 1, molecular mass markers; 2,FWC38; 3, FWC28; 4, FWC33; 5, FWC35; 6, FWC31; 7, FWC1; 8,FWC43; 9, NA1000; 10, FWC15; 11, molecular mass markers; 12,FWC2; 13, FWC44; 14, FWC19; 15, FWC17; 16, FWC2O; 17, NA1000;18, FWC37; 19, NA1000; 20, molecular mass markers; 21, FWC16;22, FWC26; 23, FWC18; 24, FWC22; 25, FWC25; 26, FWC46; 27,FWC11; 28, FWC23; 29, NA1000; 30, FWC40; 31, FWC3O; 32, FWC39;33, FWC27; 34, FWC24. (B) Lanes: 1, molecular mass markers; 2,FWC32; 3, FWC12; 4, FWC42; 5, FWC7; 6, FWC6; 7, FWC9; 8, FWC29;9, FWC41; 10, FWC45; 11, FWC21; 12, FWC14; 13, FWC4; 14, FWC5;15, FWC8; 16, FWC34; 17, molecular mass markers. Samplescontained 3 .tg of protein as estimated by the Bio-Rad protein assaywith the exception of FWC5, -14, -21, -26, -30, -38, -40, -41, -43, and-46 (see text). Molecular mass markers are: 200, 97.4, 68, 43, and 29kDa. Note: The samples in this figure are not in the same order assamples in Figure 17A and 17B.74NC,-)(‘.J0toNU,C,)C”14, 15 and 17 for FWC21, -14, and -5) or positive reaction byWestern blotting (See Fig. 18A; lanes 2, 8, 30 and 31 for FWC28, -43,-40, -30 respectively and Fig 18B; lanes 11, 12 and 14 for FWC21, -14, and -5). The Coomassie blue stained gel of extracts from bothFWC14 and 5 showed a thin high molecular weight band (Fig 17B;lanes 15 and 17) but the bands did not react to anti-RsaA sera inimmunoblot experiments (Fig 18B; lanes 12 and 14). FWC23 was theonly strain without an S-layer, as determined by TEM, to show aprominent band by Coomassie blue staining (Fig. 17A; lane 28) but itdid not label by Western blot analysis (Fig. 18A; lane 28). FWC26, -41, and -46 have an S-layer and produced an S-layer band by thecriteria of Coomassie blue staining (Fig. 17A; lanes 22 and 26 forFWC26 and -46 respectively and Fig. 17B; lane 10 for FWC41) andWestern blot analysis (Fig. 18A; lanes 22 and 26 for FWC26 and -46respectively and Fig. 18B; lane 9 for FWC41) although the bandswere only weakly visible especially for the Western blots. Extractionof the S-layer protein from FWC26, -41 and -46 by both methodswas considered poor in that the concentration of solubilized proteinwas much less that obtained from other strains (see above). Theremaining 32 strains yielded significant amounts of protein with atleast one of the extraction procedures and gave positive results byboth Western immunoblots and Coomassie blue staining.The molecular weights of the S-layer proteins from the FWCstrains were estimated from their mobility in SDS-PAGE relative toprotein standards (Table II). The S-layer proteins were quite76TABLE II. Relevant characteristics of Caulobacter strainsS -layerStrain Anti-RsaA Polysaccharides AntiSAOproteinsize (kDa) a responseb detected C response dS-layer-producingstrainsFWC1 100 + + 1FWC2 130 + + 1FWC4 130e +- 0FWC6 180 + + 3FWC7 175 + + 3FWC8 120 + + 1FWC9 135 + + 3FWC11 110 + ÷ 1FWC12 130 + ÷ 3FWC15 110 + + 0FWC16 150 + + 3FWC17 105 ÷ + 2FWC18 130 + + 1FWC19 110 + + 2FWC2O 110 + + 0FWC22 105 + + 1FWC24 145 + + 3FWC25 105 + + 1FWC26 140 + + 3FWC27 145 + + 1FWC28 105 + + 1FWC29 125 + + 0FWC31 105 + + 1FWC32 135 + + 3FWC33 110 + + 1FWC34 110 + + 2FWC35 100 + + 1FWC37 150 + + 3FWC39 195 + + 4FWC41 135 + + 3FWC42 180 + + 2FWC44 105 + + 2FWC45 140 + + 3FWC46 110 + + 2Strains without Slayerswcs NF-- 0FWC14 NF- Multiple bands 0FWC21 NF- Multiple bands 0FWC23 155- + 0FWC3O NF-- 0FWC38 NF- Multiple bands 0FWC4O NP-- 0FWC43 NF- + 077Legend for Table II.a Molecular masses were estimated on the basis of SDS-PAGE.b Reaction with low pH or EGTA-extracted protein by Westernimmunoblot analysis using anti-RsaA serum.C Results of SDS-PAGE of proteinase K-treated samples stainedfor polysaccharides by a modification of the procedure of Tsai andFrasch (1982). +, single high-molecular-weight band. -, no silver-stained bands running slower than the rough lipopolysaccharideband.d Reaction by Western immunoblot analysis with anti-SAOserum to proteinase K-treated samples. The response is rated on ascale from 0 to 4, as follows: 0, no reaction; 1, slight reaction showinga doublet in the SAO region; 2, reaction showing the doublet andsome smearing above the top doublet band; 3, definite smearingreaction reminiscent of the silver-stained image of the SÃO; 4,reaction equal in intensity to that obtained with the C. crescentusNA1000 SÃO.e S-layer presence was not confirmed by negative-stain TEMalthough the protein was reactive with anti-Rsaà antibody.f NF, not found.78heterogeneous in molecular weight ranging from 100 to 190 kDa.Proteins greater than 100 kDa are difficult to size by their SDS-PAGEmobility and so the molecular weights reported are only estimates,but are useful for comparative purposes.3.7.3 Polysaccharide analysis. The FWC strains were alsoexamined for the electrophoretic mobility of their LPS using a rapidpurification and staining procedure that was modified to visualizeboth SAO and rough LPS. All strains produced a low molecularweight polysaccharide-containing molecule, presumably rough LPS,which migrated at the dye front in the gel system used for theseexperiments (Fig. 19). All strains producing an antibody reactive Slayer protein (Fig. 18, Table II) were found to also produce a singleslower-migrating carbohydrate band with the exception of FWC4(See Fig. 19; lane 20). This SAO-like species varied in apparentmolecular weight from 60 to 95 kDa (such values are only forcomparative purposes reflecting the approximate molecular weightof a protein migrating at that position). Twenty two strains alsoshowed a stained band running with the electrophoretic mobility of a43 kDa protein (see arrow in lane 1 of Fig. 19). Lane 2 showed amoderately stained band while lane 3 contained a strongly stainedband. A Coomassie blue stained gel of the proteinase K-treatedsamples also detected this band indicating the presence of aproteinase K-resistant protein (data not shown).The S-layer negative strains showed a variety of high79FIG. 19. SDS-PAGE of proteinase K-treated whole cell lysates ofCaulobacter strains stained using a modification of the method ofTsai and Frasch (1982). Lanes: 1, NA1000; 2, FWC9; 3, FWC18; 4,FWC41; 5, FWC45; 6, FWC24; 7, FWC27; 8, FWC16; 9, FWC25; 10,FWC2O; 11, FWC19; 12, FWC17; 13, FWC15; 14, FWC11; 15, FWC1;16, FWC23; 17, NA1000; 18, FWC39; 19, FWC2; 20, FWC4; 21, FWC5;22, FWC38; 23, FWC43; 24, FWC14; 25, FWC4O; 26, FWC29; 27,FWC8; 28, FWC46; 29, FWC44; 30, FWC35; 31, FWC33; 32, FWC28;33, FWC21; 34, FWC3O; 35, FWC7; 36, FWC37; 37, FWC12; 38,NA1000; 39, FWC22; 40, FWC34; 41, MCS6; 42, FWC32; 43, FWC26;44, NA1000; 45, FWC31; 46, FWC6; 47, FWC42. Each lane wasloaded with sample containing 0.75 j.tg of KDO as estimated by themethod of Karknanis et al. (1978). The resolving gel contained 13%acrylamide. The small arrow in lane 1 indicates the running positionof a proteinase-K resistant protein that was present in some samples.SAO runs as a poorly stained band above the small arrow in lane 1.801comolecular weight carbohydrate banding patterns, including noadditional bands (FWC5, -30 and -40; see Fig. 19, lanes 21, 34, and25 respectively), multiple bands reminiscent of the smooth LPS“ladder” seen with enteric bacteria (Peterson and McGroarty 1985)(FWC14, -21, -38; see Fig 19, lanes 24, 33, and 22 respectively), or asingle band (FWC23, -43 and MCS6; see Fig. 19, lanes 16, 23 and 41respectively).Western blot analysis, using anti-SÃO serum, of the proteinaseK-treated samples (Fig. 20) showed no reaction with the probablerough LPS species of any of the strains. When a reaction was seen inS-layer producing strains it coincided with the SÃO-like bands. Noimmunoreactive bands were observed in Western blots, using antiSAO sera, for S-layer minus strains (Table II, Fig. 20; lanes 9, 21, 11,12, 10, 22, 6).The majority of the “SÃO-like” carbohydrates, identified bysilver-staining (Fig. 19) reacted with the anti-SÃO serum in Westernblotting experiments (Fig. 20). However, there was a variation in thedegree of immuno-reactivity between strains. A qualitative analysisof the degree to which the SÃO-like molecules reacted with anti-SÃOsera is presented in Table II. The laboratory strain NÃ1000 staineddarkly and produced a major band that was smeared, reminiscent ofthe silver-stained image of the SÃO, and a minor band runningslightly faster than the major band (See Fig. 20; lanes 4, 17, 24, 28,37, and 47). The band produced by FWC39 was as intense as thatproduced by NA 1000 but no minor band was noted (Anti-SÃOresponse = 4 in Table II; Fig. 20; lane 25). Eight strains produced areaction that showed a definite smearing reaction, but was less82FIG. 20. Western blot analysis of proteinase-K treated whole celllysates of Caulobacter strains. Blots were reacted with c-SAO sera.Lanes: 1, molecular mass markers; 2, FWC32; 3, FWC26; 4, NA1000;5, FWC31; 6, FWC43; 7, FWC6; 8, FWC42; 9, FWC5; 10, FWC38; 11,FWC2I; 12, FWC3O; 13, molecular mass markers; 14, FWC7; 15,FWC37; 16, FWC12; 17, NA1000; 18, FWC22; 19, FWC34; 20, MCS6;21, FWC14; 22, FWC4O; 23, molecular mass markers; 24, NA1000;25, FWC39; 26, FWC2; 27, FWC4; 28, NA1000; 29, molecular massmarkers; 30, FWC9; 31, FWC18; 32, FWC41; 33, FWC45; 34, FWC24;35, FWC27; 36, FWC16; 37, NA1000; 38, molecular mass markers;39, FWC28; 40, FWC33; 41, FWC35; 42, FWC44; 43, FWC46; 44,FWC8; 45, FWC29; 46, molecular mass markers; 47, NA1000; 48,FWC23; 49, FWC1; 50, FWC11; 51, FWC15; 52, FWC17; 53, FWC19;54, FWC2O; 55, FWC25. Each lane was loaded with samplecontaining 0.75 j.tg of KDO as estimated by the method of Karknanis etal. (1978). Samples were fractionated by SDS-PAGE using resolvinggel that contained 13% acrylamide. Note: The samples in this figureare not in the same order as samples in Figure 19.83IU)(NU-)cDInU)C.-)(N00-)C.)C.)C.)intense than that seen for NA 1000, and produced a minor bandbelow the smearing band (Anti-SÃO response = 3 in Table II; see Fig.20; lane 3 [FWC26J as an example). Three strains, FWC6, 7 and 9produced the definite smearing band but did not contain the lowerminor band (Anti-SÃO response = 3 in Table II; see Fig. 20; lane 7[FWC6] as an example). Four strains showed some minor smearingabove a single band (Anti-SAO response = 2 in Table II; see Fig. 20;lane 42 [FWC44] as an example) and two strains showed minorsmearing above a doublet band (Anti-SÃO response = 2 in Table II;see Fig. 20; lane 53 [FWC19] as an example). Eight strains produced adoublet band with no smearing above them (Anti-SAO response = 1in Table II; see Fig. 20; lane 50 [FWC11J as an example) while fourstrains showed only a single immunoreactive band (Anti-SAOresponse = 1 in Table II; see Fig. 20; lane 26 [FWC2J as an example).Three strains producing a silver-stain positive band did not produceany immunoreactive bands (Anti-SÃO response = 0 in Table II;FWC15, -20, -29). Figure 20 illustrates that the “SÃO-like”carbohydrate produced by freshwater Caulobacters differed fromstrain to strain with respect to electrophoretic mobility andimmunological reactivity to the anti-SÃO sera.3.8 Metal ion requirements for C. crescentus growth andS-layer assemblyIt has been shown that wild-type C. crescentus NA1000requires calcium for growth. The experiments described below85examine this calcium requirement in more detail.3.8.1 Influence of calcium on growth of NA1000 andJS1001. Washed M3Higg grown NA1000 and JS1001 cells wereinoculated into M10Higg medium supplemented with variousconcentrations of calcium and the optical density was measured after48 h of incubation (Fig. 21). Calcium concentration had littleinfluence on the growth of the calcium-independent mutant JS1001whereas calcium concentrations of less than 75 p.M resulted indecreased growth yield of NA1000. Figure 22 illustrates that calciumbecame growth rate limiting for NA1000 below 250 p.M. WhenJS1003, which was NA1000 with rsaA interrupted with a Kmrcassette, was used in place of NA1000, the same growth patternswere observed. The washed NA1000 cells used to inoculate M10Higgmedium containing no metal ion supplement did not lyse or die.After 48 hrs of incubation, phase contrast microscopy showed thecells were elongated, tapered consisting of a swarmer and a stalkedcell frozen in mid-cell division. Addition of calcium resulted ingrowth following a brief lag period.3.8.2 Influence of metal ions on the growth rate ofNA1000 and JS1001. M10Higg medium, which contains 2.2 mMMg2+, was supplemented with various cations to a final concentrationof 500 p.M and the growth rate of NA1000 was determined by themethod outlined in the Materials and Methods section (Fig. 23).NA1000 did not grow in M10Higg medium or M10Higg medium865.3.I NA1000IiM CalciumFIG. 21. The influence of calcium on growth of NA1000 and JS1001.5 x 106 washed mid-logarithmic cells were inoculated into 5 mis ofM 10Higg medium containing 0 to 500 iiM calcium. Cultures wereincubated at 30°C and the optical density at 600 nm was measuredafter 48 hrs. Duplicate tubes were used for all concentrations ofcalcium and the experiment was repeated 3 times. The final opticaldensities for each calcium concentration varied by less than 5%between experiments.87—0-.. ‘-FIG. 22. Influence of calcium on the generation time of Caulobactercrescentus NA1000. 5 x 106 washed mid-logarithmic cells wereinoculated into 5 mis of M10Higg medium containing 65.2 to 1000 Mcalcium. Cultures were incubated at 30°C and the optical density at600 nm was followed during growth. The mean generation time wasdetermined for cultures between 0D600 = 0.100 to 1.000. Duplicatetubes were used for all concentrations of calcium and the experimentwas repeated 3 times. The bar indicates the standard deviation.p.M Calcium884. C C 0 — 0 0 + 0OøNZ<OLIC.)500 .tM cation added ZFIG. 23. Influence of the metal ion supplement on the generationtime of Caulobacter crescentus NA1000 cultured in M10Higgmedium. 5 x 106 washed mid-logarithmic cells were inoculated into5 mis of M10Higg medium supplemented to 500 iM with a chloridemetal salt. Cultures were incubated at 30°C and the optical densityat 600 nm was followed during growth. The mean generation timewas determined for cultures between 0D600 = 0.100 to 1.000.Duplicate tubes were used for all metal salts and the experiment wasrepeated 3 times. No growth was noted in unsupplemented mediumor medium supplemented with a monovalent cation. M+ = sodium,potassium or lithium. Unsupplemented M10Higg medium contained2.2 mM magnesium chloride. The bar indicates the standarddeviation.89supplemented with the mono-valent cations lithium, sodium orpotassium (chloride salts). Supplementation of M10Higg mediumwith one of 8 divalent or 2 trivalent cations allowed NA1000 to grow,although the resulting generation times were greater than thoseobserved in medium supplemented with calcium. Figure 24 indicatesthat the growth of JS1001 was not greatly decreased by growth inthe presence of any of the cations tested. The cation used tosupplement M10Higg medium also had a pronounced influence on thelag time for NA1000 (Fig. 25) that was not noted in JS1001 cultures(Fig. 26). The lag phase was defined as the number of hours for theculture to reach an 0D600 of 0.100. Increasing the concentration ofmagnesium in M10Higg medium to 3.0 mM did permit limited growthof NA1000, although culture lysis occurred as the 0D600 approached1.5 (data not shown).3.8.3 Influence of metal ions on S-layer crystallization.TEM was used to determine if crystallized S-layer was formed on thecell surface of NA1000 or in non-cell associated sheets in cultures ofJS1001. Assembled S-layer was observed in cultures of NA1000 orJS1001 only when M10Higg medium was supplemented with calciumor strontium. Figure 27 illustrates the non-cell associated sheets ofS-layer produced by JSlOOl cultured in the presence of strontium.Identical sheets were produced when JS1001 was cultured onM10Higg medium supplemented with calcium. TEM of NA1000 cellsdemonstrated the presence of crystallized S-layer on the cell surfacewhen grown on M10Higg medium supplemented with calcium or903. p.M cation addedFIG. 24. Influence of the metal ion supplement on the generationtime of Caulobacter crescentus JS1001 in M10Higg medium. 5 x 106washed mid-logarithmic cells were inoculated into 5 mls of M10Higgmedium supplemented to 500 p.M with a chloride metal salt.Cultures were incubated at 30°C arid the optical density at 600 nmwas followed during growth. The mean generation time wasdetermined for cultures between 0D600 = 0.100 to 1.000. Duplicatetubes were used for all metal salts and the experiment was repeated3 times. M+ = sodium, potassium or lithium. UnsupplementedM 10Higg medium contained 2.2 mM magnesium chloride. The barindicates the standard deviation.91. IFIG. 25. Influence of the metal ion supplement on the lag phase ofCaulobacter crescentus NA1000 in M10Higg medium. 5 x 106washed mid-logarithmic cells were inoculated into 5 mis of M10Higgmedium supplemented to 500 M with a chloride metal salt.Cultures were incubated at 30°C and the optical density at 600 nmwas followed during growth. Lag phase was defined as the numberof hours required for a culture to reach an 0D600 = 0.100. Duplicatetubes were used for all metal salts and the experiment was repeated3 times. No growth was noted in unsupplemented medium ormedium supplemented with a monovalent cation. M+ = sodium,potassium or lithium. Unsupplemented M10Higg medium contained2.2 mM magnesium chloride. The bar indicates the standarddeviation.— .—+z500 pM cation added9230201000 +500 pM cation addedFIG. 26. Influence of the metal ion supplement on the lag phase ofCaulobacter crescentus JS1001 in M10Higg medium. 5 x 106 washedmid-logarithmic cells were inoculated into 5 mis of M10Higg mediumsupplemented to 500 tM with a chloride metal salt. Cultures wereincubated at 30°C and the optical density at 600 nm was followedduring growth. Lag phase was defined as the number of hoursrequired for a culture to reach an 0D600 = 0.100. Duplicate tubeswere used for all metal salts and the experiment was repeated 3times. M+ = sodium, potassium or lithium. Unsupplemented M10Higgmedium contained 2.2 mM magnesium chloride. The bar indicatesthe standard deviation.93FIG. 27. Negative-stain TEM micrograph of strontium mediatedcrystallization of RsaA formed in a Caulobacter crescentus JS1001colony. A double S-layer sheet, as determined by optical diffraction,that is not associated with the bacterial cells is formed. This electronmicrograph was taken by Dr. J. Smit. Bar 0.1 jiM.94strontium (not shown). TEM of NA1000 cells grown on M10Higgmedium supplemented with ions other than strontium or calciumshowed that large amounts of material, resembling cell membranes,sloughed off the cells and that no crystallized S-layer could be found.3.8.4 Influence of calcium or strontium concentration onS-layer crystallization. NA1000 and JS1001 were grown onM 10Higg plates that were supplemented with 5, 3, 1 and 0.5 mMcalcium or strontium. Colonies were examined by TEM for thepresence of crystallized S-layer. Table III demonstrates that S-layercrystallization in the presence of calcium or strontium wasconcentration dependent. Higher concentrations of strontium orcalcium were required for non-cell associated S-layer sheets to formin cultures of JS1001 than were required to allow detection of Slayer assembled on the cell surface of NA1000. It was also notedthat higher concentrations of strontium were required forobservation of crystallized S-layer in both NA1000 and JS1001cultures than that required for calcium mediated crystallization of Slayer.3.8.5 Localization of non-crystallized S-layer protein.Washed NA1000 cells were used to inoculate M10Higg liquid mediumsupplemented to 500 j.tM with a cation. After growth to midlogarithmic phase the cells were pelleted by centrifugation andwashed twice with 20 mM FIEPES buffer (pH 7.2). Whole cell lysates95TABLE III. The influence of template and cation concentration on thecrystallization of RsaA.Metal iona Crystallized SlayerbNA1000 JS1001Calcium05C e1.0 + +3.0 + +5.0 + +Strontium0.51.0 +-3.0 + +5.0 + +a Chloride salt.b Forming an array on the cell surface of NA1000 and noncell associated sheets in JS1001.C Concentration in mM.d Crystallized S-layer observed by negative-stain TEM.e Crystallized S-layer not observed by negative-stain TEM.96of the washed cells were analysed by Western blotting using antiRsaA sera. Figure 28 shows that S-layer protein was detected only incultures grown in the presence of calcium or strontium.Washed NA1000 or JS1001 cells were used to inoculateM 10Higg plates supplemented to 1 mM with a metal cation Followinggrowth the cells were scraped from the plate and suspended in 10mM Tris - 1 mM EDTA buffer. The cell lysates were analysed byWestern blotting using anti-RsaA sera. This procedure was used toidentify any RsaA within the cells, attached to the cell surface orRsaA that was translocated to the cell surface but not attached to thecell. The method used to grow and prepare cells for analysis in Fig.28 would wash unattached S-layer from the cell surface and thus notdetect the protein. S-layer was detected only when unwashedJS1001 or NA1000 cells were cultured in the presence of calcium orstrontium (Fig. 29). Unadsorbed sera was used as a control todetermine if any of the non-RsaA proteins recognized by theunadsorbed sera (see Fig. 16) were repressed during growth inM 10Higg medium supplemented with various cations. Figure 28 and29 illustrate that only RsaA was inhibited by growth on cations otherthan calcium and strontium.97200,00097,40068,00043,000FIG. 28. Western blot reacted with unadsorbed x-RsaA sera of wholecell lysates of washed Caulobacter crescentus NA1000 cells grown inliquid M10Higg medium supplemented to 500 j.tM with a chloridemetal salt. Lanes: 1, Molecular mass markers in daltons; 2, calcium;3, strontium; 4, manganese; 5, nickel. Lanes 2 to 5 were loaded with10 jig of protein as estimated by the method of Markwell et al.(1978). Samples were fractionated by SDS-PAGE using a resolvinggel containing 10% acrylamide. Unsupplemented M10Higg mediumcontained 2.2 mM magnesium chloride.1 2 3 4 598FIG. 29. Western blot of whole cell lysates from unwashedCaulobacter crescentus cells. Cells were grown on M10Higg mediumplates supplemented to 1 mM with a chloride metal salt. The blotwas reacted with unadsorbed x-RsaA sera. Lanes 1, 3, 4, and 5NA1000. Lanes 6, 7, 8, and 9 = JS1001. Lanes: 1, calcium; 2,molecular mass markers; 3, nickel; 4, manganese; 5, zinc; 6, no ions;7, calcium; 8, manganese; 9, zinc. All lanes but 2 were loaded with10 .tg of protein as estimated by the method of Markwell et al.(1978). Samples were fractionated by SDS-PAGE using a resolvinggel containing 10% acrylamide. Unsupplemented M10Higg mediumcontained 2.2 mM magnesium chloride.1 2 3 4 5 6 7 8 9993.9 Genetic investigation of the Calcium-independencephenotypeAn attempt was made to characterize the calcium-independent,S-layer attachment-defective phenotype at the genetic level. Atransposon library was constructed in attempt to isolate the generesponsible for the phenotype. A cosmid library of wild-type C.crescentus NA1000 was used in an attempt to complement thecalcium-independent phenotype in the mutant strain JS1001.3.9.1 Production and screening of a transposon library.The suicide vector pSUP2O21 was used to produce a transposonlibrary of C. crescentus NA1000 containing 20,000 independenttransposon-insertion mutants. The library was screened on calcium-free M10Higg plates for the identification of calcium-independentmutants. Although the medium supported the growth of JS1001 andinhibited the growth of NA1000, no calcium-independent mutantscould be isolated from the transposon library.S-layer attachment-defective Tn5 mutants were isolated fromthe library by use of a colony immunoblot screen (Awram and Smit,unpublished; see appendix I [method C]). These attachment—defective mutants were analysed by SDS-PAGE and silver-staining todetermine the LPS banding pattern (Fig. 30). Figure 30 demonstratesthat these mutants all produce altered LPS banding patterns.However, as detailed below, none of these mutants were calciumindependent.When SÃO is stained using the modification of the Bio-Rad100silver-stain method it is visualized as a golden-yellow band while therough LPS stains black. The arrow points to the running position ofthe SÃO band in lane 1 of Figure 30A and 30B. The Tn5 mutantscould be grouped into 6 clusters on the basis of LPS bandingpatterns. Cluster 1 consisted of one mutant, Fl, that produced a bandwith the same electrophoretic mobility as SAO although the band wasstained black (Fig. 30A; lane 3). The diamond () beside lane 1 ofFigure 30A and 30B denotes the running position of LPS species thathave less electrophoretic mobility than the rough LPS in the Tn5mutant strains. Cluster 2 consisted of 5 mutants that produced aband that was golden-yellow in colour like the SÃO band although itran with a much greater electrophoretic mobility (Fig. 30A; lanes 4, 6and 7 for F2, 4 and 5 respectively and Fig. 30B; lanes 3 and 10 forF12 and 19 respectively). Cluster 3 consisted of 2 mutants thatproduced a doublet band that were golden-yellow in colour runningat the same electrophoretic mobility as the cluster 2 mutants (Fig.30Ã; lane 13 for Fil and Fig. 30B; lane 12 for F21). Cluster 4consisted of one mutant that produced a doublet that stained black(Fig. 30A; lane 9 for F7). Cluster 5 consisted of 7 Tn5 mutants whichproduced less rough LPS and a small amount of a golden-yellow highmolecular weight LPS species (Fig. 30A; lane 11 for F9 and Fig. 30B;lanes 4, 5, 6, 7, 8 and 9 for F13, 14, 15, 16, 17 and 18 respectively).Cluster 6 consisted of 5 mutants that produced a dark band at therunning position of the diamond (Fig. 30A; lane 5, 8, 10 and 12 forF3, F6, F8 and FlO respectively and Fig. 30B; lane 11 and 13 for F20101FIG. 30. SDS-PAGE of proteinase K treated NaC1/EDTA extracts ofCaulobacter crescentus Tn5 strains. The S-layer attachmentdefective Tn5 mutants are designated “Fl - F22”. See appendix I(method C) for details on the method used to isolate the mutants.The gel was stained using the modification of the Bio-Rad silver-stain kit. (A) Lanes: 1, NA1000; 2, JS1001; 3, Fl; 4, F2; 5, F3; 6, F4;7, F5; 8, F6; 9, F7; 10, F8; 11, F9; 12, FlO; 13, Fil. (B) Lanes: 1,NA1000; 2, JS1001; 3, F12; 4, F13; 5, F14; 6, Fl5; 7, Fl6; 8, F17; 9,F18; 10, F19; 11, F20; 12, F21; 13, F22. The arrows designate therunning position of wild-type SAO while the diamonds designate therunning positions of LPS species produced by the Tn5 mutants thatrun with a greater electrophoretic mobility than the rough LPS.Samples containing 0.5 pg KDO, as estimated by the method ofKarknanis et al. (1978), were loaded into each lane and fractionatedwith a resolving gel containing 13% acrylamide.102A0B1 2 345 6789 1011121301 2 3 4 5 6 78 910111213103and F22 respectively). An S-layer attachment-defective Tn5 mutantfrom each cluster was tested for the ability to grow in the absence ofcalcium. None of the four mutants examined were capable of growthin M10Higg liquid medium indicating that although they producedaltered SÃO they were not “calcium-independent” mutants.3.9.2 Complementation of JS1004 with an NA1000cosmid library. A cosmid library derived from NA1000 wasintroduced into the calcium-independent S-layer-negative strainJS1004 by electroporation. One of 680 cosmid containing clonesreacted with anti-SÃO sera in a dot blot immunological screen. Thiscosmid was designated “D”. A Western blot using anti-SAO sera ofJS1001 containing cosmid D indicated that although SAO wasproduced, it was at less than wild-type levels (Fig. 31). TheCaulobacter DNA contained in cosmid D was isolated and used as aprobe to identify 28 overlapping cosmids. Restriction digests ofthese cosmids indicated that 18 had unique banding patterns andthese 18 were introduced into JS1001 separately by electroporation.The electroporants were screened by Western blot analysis usinganti-SÃO sera and two were shown to produce SAO, although atlevels less than wild-type (Fig. 32). These cosmids were designatedD12 and D13. Washed cultures of JS1001 containing cosmid D12 orD13 were analysed by Western blotting using anti-RsaA sera. TheWestern blots indicated that these cells were unable to attach mostof the S-layer protein to the cell surface (not shown).10497,40068,00043,00029,000FIG. 31. Western blot reacted with cc-SÃO sera of proteinase Ktreated whole cell lysates. Cells were grown in PYE liquid media.Lanes: 1, molecular mass markers in daltons; 2, NA1000; 3, JS1001;4, JS1004 containing cosmid D; 5, JS10O1 containing cosmid D.Samples containing 1 pg KDO as estimated by the method ofKarknanis et al. (1978) were loaded into each lane and fractionatedby SDS-PAGE using a resolving gel containing 13% acrylamide.1 2 3 4 5105FIG. 32. Western blot reacted with cc-SAO sera of proteinase Ktreated whole cell lysates. Cells were grown in PYE liquid media.Lanes: 1, NA1000; 2, JS1001 3, JS1001 containing cosmid D12; 4,JS1001 containing cosmid D13. Samples containing 1 jig KDO asestimated by the method of Karknanis et al. (1978) were loaded intoeach lane and fractionated by SDS-PAGE using a resolving gelcontaining 13% acrylamide.1 2 3 41064 DiscussionFigure 33 is a model diagram of the C. crescentus cell surfacebased, in part, on the information contained in this thesis. The figureillustrates the presence of three major polysaccharide species one ofwhich, the SÃO or smooth LPS, plays a role in the attachment of theS-layer to the cell. Figure 33 also proposes that calcium is involvedin the formation of the S-layer and may also have a role inmembrane assembly. The following is a discussion of theexperimental evidence which forms the basis of the model.4. 1 C. crescentus cell surface polysaccharidesThe C. crescentus cell surface was examined and found toproduce three types of polysaccharides: an EPS, a “rough” LPS, and a“smooth” LPS termed the SÃO.4.1.1 C. crescentus “rough” LPS. The presence of LPS in C.crescentus strains was initially revealed by SDS-PAGE analysis ofwashed whole cells treated with proteinase K (Fig. 3 and 4). The highelectrophoretic mobility of the band sensitive to the Tsai and Fraschstain and the absence of bands of higher molecular weight indicatedthat the LPS species was “rough” in nature (Hitchcock et al. 1986).The LPS from all strains yielded similar electrophoretic profiles (Fig.4).The LPS of CB2A and NA1000 was isolated and purified using a107phospholipid 2+Ca S-layerproteinEPS[::1/l\ Ii\protein R - LPS SADporin proteinhydrophobicinteraction*negativechargeFIG. 33. Representation of the Caulobactersurface.crescentus NA1000 cell108modification of the method of Darveau and Hancock (1983). Thepurified NA1000 and CB2A LPS was analyzed by SDS-PAGE. Half ofthe gel was stained using the method of Tsai and Frasch (1982) whilethe other half was stained using the Bio-Rad silver-stain kit (Fig. 3).Both methods produced similar patterns of staining. The bandsresulting from samples prepared by the method of Hitchcock andBrown (1983) (Fig. 3, lanes 1 and 2) had a slightly slowerelectrophoretic mobility and were wider in the horizontal plane thanthat of the purified LPS. This may result from these samplescontaining bulk cellular components, such as undigested protein,peptidoglycan and nucleic acids, which alters the mobility of the LPSthough the gel. Figure 3 also demonstrates that the purified LPS didnot contain contaminants. The Bio-Rad silver-stain method detectsprotein, carbohydrate, nucleic acid and contaminating metal salts(Bio-Rad 1987) whereas the staining procedure of Tsai and Frasch(1982) is very sensitive for detection of LPS but is insensitive for thedetection of protein and nucleic acids (Hitchcock and Brown 1983).The Bio-Rad procedure allowed the visualization of contaminatingprotein and nucleic acid which were contained in the “crude” LPSpreparation.Detailed chemical analysis of the purified LPS from CB2A,NA1000 and JS1001 was conducted by Dr. N. Ravenscroft(Ravenscroft et al. 1992). The purified LPS was shown to consist oftwo definable regions: (i) an oligosaccharide region, consisting of aninner core of three residues of 2-keto-3-deoxyoctonate, two residuesof x-L-glycero-D -mannoheptose, and one x-D-glycero-D109mannoheptose and an outer core region containing one residue eachof x-D-mannose, x-D-galactose, and x-D-glucose, with the glucoselikely phosphorylated and (ii) a region equivalent to the lipid A ofarchetype LPS, consisting primarily of the esterified fatty acid 3-OH-dodecanoate (see Table I, appendix II).The lipid A-like region was resistant to conclusive analysis.The major or only fatty acid component, 3-OH-dodecanoic acid, is lesscommonly found in LPS than 3-OH-tetradecanoic acid, but has beenfound, for example, in the Lipid A from Pseudomonas aeruginosa(Bhat et al. 1990). Mild acid hydrolysis readily cleaved the LPS intoLipid A and core oligosaccharide fractions. Yet despite extensiveefforts, no amino or diamino sugars, typical of thet’backbone” regionof other Lipid A moieties (Mayer et al. 1988; Rietschel et al. 1990)were detected during multiple approaches for amino sugar analysis(Ravenscroft et al. 1992).There is no clear explanation for the high stability of the LipidA and thus its resistance to hydrolysis into assayable sugars. It isnotable, however, that Caulobacters are members of the x-2subdivision of the alpha proteobacteria, as defined by 16S rRNAsequence analysis (Stackebrandt et al. 1988; Stahl et al. 1992; Woese1987), a group that contains members producing “unusual” Lipid Astructures. The amino sugar 2,3-diamino-2,3-dideoxy-D- glucose(DAG) has been identified in the Lipid A backbone structures of somespecies in this phylogenetic group (Weckesser and Mayer 1988). Itmay be that a variation of the DAG-type Lipid A is present in C.110crescentus. High-voltage paper electrophoresis has proven toseparate and detect such unusual Lipid A sugars in other species(Mayer et al. 1988) and may be appropriate in future studies withthe Caulobacter Lipid A.It has been reported that C. crescentus whole membranescontain from two-thirds to ten-fold less KDO than that reported formembranes of rough mutants of Salmonella typhimurium and wild-type S. typhimurium, respectively (Agabian and Unger 1978). Thisstudy showed that the total amount of KDO in whole cells of NA 1000and CB2A is less than that found in E. coli B, but only to the extentof 20 to 30%. This indicates that KDO could be used as an outermembrane marker during procedures to separate membranefractions. The published methods for membrane separation andisolation in C. crescentus do not account for the missing KDO in the“outer membrane” fractions (Agabian and Unger 1978; Clancy andNewton 1982; Koyasu et al. 1980). Two of the protocols (Clancy andNewton 1982; Koyasu et al. 1980) used PBS in the procedure. It hasbeen since shown that PBS extracts LPS from the envelope of C.crescentus (Edwards and Smit 1991; Walker and Smit, unpublishedobservation) and the absence of KDO in these membranepreparations might be explainable on that basis.4.1.2 C. crescentus SAO. SAO was initially identified as acontaminant in a purified LPS sample (Fig. 6). This proteinaseresistant molecule was detected in whole cell lysates of wild-type S111layer producing strains or strains that are attachment-competentusing a modification of the method of Tsai and Frasch (Fig. 7). Asimilar band was not detected in calcium-independent strains whichare unable to attach the S-layer to the cell surface indicating that itmay play a role in S-layer attachment. Originally it was unclear ifthis band represented a species of LPS with a homogeneous length 0-antigen, as seen in Aeromonas species (Belland and Trust 1985;Dooley and Trust 1988), or if it was a unique carbohydrate species.The SAO was not detected by the SDS-PAGE and silver-stainingmethods of Hitchcock and Brown (1983) or Tsai and Frasch (1982)which are used widely to identify and qualitatively characterize LPSfrom many bacterial species. Also, the SAO did not precipitate withthe rough LPS during the cold ethanol-MgC12 step of the Darveau andHancock (1983) procedure but remained in the supernatant alongwith a significant amount of rough LPS. Thus an isolation andpurification procedure for SAO was determined in order tochemically characterize the molecule.The cell surface extraction procedure using 0.77 M NaC1 / 0.12M EDTA (pH 7.2) was effective at solublizing the cell surfacecomponents without releasing large amounts of cytoplasmicconstituents, although the redistribution of cytoplasmic material isindicative of the plasma membrane losing its selective permeabilitybarrier toward ions (Whitfield and Murray 1956). This extractionmethod provided a convenient and rapid method to obtain roughLPS, SAO and EPS from liquid cultures ranging in volumes of 1.5 ml112to 60 liters. For detailed chemical analysis, SAO was isolated fromJS1003 instead of NA1000 due to the deletion of rsaA in this strainwhich resulted in a lower protein concentration in the NaC1/EDTAextract. The SAO was then separated from the rough LPS by SDSPAGE and isolated from the polyacrylamide gel by electroelution.Detailed chemical analysis of the purified SAO was preformed by Dr.D. N. Karunaratne (See appendix II, Table III). The SAO was shownto be composed of lipid and polysaccharide. The major fatty acid wasidentified as 3-OH-dodecanoate which is the same fatty acid as thatfound in the rough LPS. Minor amounts of the same sugars detectedin the rough LPS were identified as well as large amounts of 4,6-dideoxy-4-amino hexose, 3 ,6-dideoxy-3 -amino hexose and glycerolall in equal proportions. Proton NMR studies on the purified SÃOhave determined that the amino group of both dideoxyamino hexosesare acetylated (W. R. Abraham, unpublished; see Table III, appendixII). Given this data and the fact that colorimetric assays indicatedthat SÃO and rough LPS have the same molar ratio of KDO:phosphateit is clear that SÃO is a species of smooth LPS with homogeneous-length 0-antigen.Dideoxyamino sugars are regarded as “rare” and “unusual”although they have been identified as component sugars in a numberof bacterial species LPS (for a list of species and references seeAshwell and Hickman 1971; Jann and Jann 1977; Kenne andLindberg 1983; LUderitz et al. 1968; and Wilkinson 1977). 4,6-dideoxy-4-amino hexose or 3 ,6-dideoxy-3 -amino hexose containing0-antigens often have altered solubility characteristics than that of113most LPS. When some, but not all, species possessing these sugarsare subjected to the hot phenol/water LPS isolation procedure ofWestphal et al. (1952) the 0-antigen is found in the phenol phasewhereas the rough LPS is found in the aqueous phase. The roughand smooth LPS of most bacteria is partitioned into the aqueousphase. It has been suggested that the phenol solubility of thedideoxyamino hexose containing LPS is due to the increased numberof non-polar groups (terminal methyl and N-acetyl residues) in thesesugars (Hickman and Ashwell 1966). Whatever the chemical basiscontributing to the phenol solubility of such an LPS it is clear thatthese 0-antigens possesses a hydrophobic character not found inmost LPS species. It is tempting to speculate that differences inhydrophobic character between the C. crescentus rough and smoothLPS accounted for their separation during the Darveau and Hancock(1983) procedure.The inability to stain all components of LPS after SDS-PAGEusing the standard methods of Hitchcock and Brown (1983) and Tsaiand Frasch (1982) is relatively uncommon but is not unique to C.crescentus. Cytophaga johnsonae (Godchaux et al. 1990),Campylobacter jejuni (Preston and Penner 1987), C. coli (Mandatoriand Penner 1989), E. coli 026 (Karch et al. 1984), Coxiella burneti(Hackstadt et al. 1985), and Neisseria gonorrhoeae (Mandrell et al.1986) also have LPS species which do not stain using the standardLPS silver stain and immunological or alternative staining proceduresmust be used to visualize these molecules. The Bio-Rad silver-stain114procedure detected both the rough LPS and the SÃO (Fig. 8) althoughits sensitivity towards the rough LPS was approximately five-foldless than that obtained using the method of Tsai and Frasch (1982).Therefore, different staining methods are required depending whatspecies of Caulobacter LPS are being examined.The mechanism by which macromolecules are stained duringvarious silver staining protocols is unknown (Deh et al. 1985;Goldman and Merril 1982; Kropinski et a!. 1986). Silver stains arebased on methods using either ammoniacal silver solutions (Oakleyet al. 1980) or silver nitrate (Merrill et a!. 1981). The method of Tsaiand Frasch (1982) uses ammoniacal silver whereas the Bio-Rad silverstain kit uses silver nitrate. A modification of the Bio-Rad silverstain (Cava et al. 1989) which substitutes sodium periodate fordichromate in the oxidation step was shown to stain SÃO inpolyacrylamide gels (Fig. 30). SAO was not stained when periodicacid was used to oxidize molecules prior to staining with ammoniacalsilver in the method of Tsai and Frasch (1982) (Fig. 3 and 4). Thisindicates that the oxidized SÃO does not react with ammoniacalsilver. However, more detailed comparisons between the two LPSstaining methods will have to be carried out in order to determinethe precise reason why SÃO does not stain by the method of Tsai andFrasch (1982).When SAO is detected by the Bio-Rad method or themodification of the Bio-Rad method it is stained a yellow-orangecolour which is a common staining characteristic of 0-glycosidically115linked carbohydrate containing molecules (Deh et a!. 1985). Whenphotographing such gels with black and white film the resultingimage of the SÃO is much less intense in comparison to the rough LPSwhich stains dark black. Therefore, the photographs in Figure 30 arenot accurate representations of the original polyacrylamide gel.Figures 8 and 12 show that the SAO was not resolved intodiscrete bands as has been shown for other 0-antigen ofhomogeneous length (Chart et a!. 1984; Dooley et al. 1985). Althougha variety of acrylamide concentrations and a number of gel protocolswere used, heterogeneity in this region was not identified. Thepurified SÃO was subjected to laser desorption time of flight massspectroscopy analysis to identify any microheterogeneity in thisregion (A. Rudiger, unpublished data; See Fig.3, appendix II). Theanalysis indicated that there is microheterogeneity present, but theaverage difference in mass between SAO molecules is approximately176 daltons. Therefore SDS-PAGE, under any conditions, would beunable to detect this microheterogeneity.The cell surface defect responsible for the S-layer attachment-defective phenotype of calcium-independent mutants of C.crescentus appears to be the inability to produce an 0-antigen ofuniform length (SÃO molecule). This is reminiscent of the defectthought to be responsible for the attachment-defective phenotype inAeromonas salmonicida and A. hydrophila (Belland and Trust1985; Dooley and Trust 1988). In mutants of both species inabilityto attach the S-layer to the cell surface has been correlated with116defects in the LPS. With A. salmonicida strains the inability toproduce a homogeneous-length smooth LPS results in an attachment-defective phenotype (Belland and Trust 1985). Wild-type A.hydrophila strains also produced a homogeneous-length 0-polysaccharide but mutants that generated only a core LPS could stillmaintain the S-layer. Mutants of A. hydrophila producing a deep-rough LPS were, however, found to be S-layer attachment-defective(Dooley and Trust 1988). Dooley and Trust (1988) have suggestedthat a homogeneous-length 0-antigen may be required by all Gram-negative S-layer producing species. However, the S-layer containingCampylobacter fetus has smooth LPS of heterogeneous-lengthreminiscent of the LPS of enteric bacteria (Perez-Perez et a!. 1986)although Yang et al. (1992) have implicated the LPS as the cellsurface molecule to which the S-layer attaches. The 0-antigen of A.salmonicida and A. hydrophila extends past the S-layer into theenvironment (Chart et al. 1984; Dooley et a!. 1988). In contrast, theCaulobacter crescentus (Smit, unpublished observation) andCampylobacter fetus (Fogg et a!. 1990; McCoy et al. 1975) 0-antigensdo not extend past the S-layer. At present only these limitednumber of Gram-negative S-layer producing species have beenexamined by SDS-PAGE to determine the LPS profile. Thereforebroad generalizations cannot be made, however, it appears that Slayers attach to the cell surface of Gram-negative bacteria via theLPS. Like S-layers themselves, the mechanism of attachment to thecell surface may prove to be a product of convergent evolution.1174.1.3 C. crescentus EPS. CB2A, NA1000 and JS1001 producedsufficient quantities of an EPS during growth in broth culture for it tobe isolated in an aqueous phase as a by-product of the generalpurpose LPS isolation procedure of Darveau and Hancock (1983).The inability to wash the EPS off the surface by repeatedcentrifugations and suspensions and the lack of significant amountsof polysaccharide located in the culture medium following growthindicated that the polymers were not a loosely associated “slime”layer but a true capsule or EPS layer (ørskov and ørskov 1990).Carbohydrate analysis of the purified EPS from NA1000, JS1001 andCB2A was conducted by Dr. N. Ravenscroft (Ravenscroft et al. 1991;see Fig. 2, appendix II). These studies showed that NA1000 andCB15A produce a unique neutral EPS. The EPS of CB2A contained Dglucose, D-gulose and D-fructose in a ratio of 3:1:1 whereas theNA1000 EPS contained D-galactose, D-glucose, D-mannose and Dfucose in approximately equal amounts. NMR and methylationanalysis confirmed that the polymers consist of repeating units,NA1000 consisting of a tetrasaccharide and CB2A a pentasaccharide,containing both cz- and 13-linked sugar residues. The repeating sugarunits indicated that the isolated polymers had the general features ofa bacterial EPS or capsule. The chemical differences in the EPS ofCB2A and NA1000 illustrate that they have evolved independentlyso as to present different chemical motifs to the externalenvironment. Like other bacterial species C. crescentus strains mayproduce many different EPS chemotypes. The EPS isolated from the118calcium-independent S -layer attachment-defective mutant IS 1001was chemically identical to that found in the parent strain NA1000.The classification of cell surface polysaccharides as capsules(EPS), slime layers or LPS are based both on chemical analysis of thepolysaccharide as well as the biophysical characteristics of thematerial (Costerton et al. 1981). The firm attachment of theCaulobacter EPS to the surface is a property shared with the LPSand it might be argued that the EPS is an LPS with a very long 0-antigen. LPS is often differentiated from EPS based on the criterionthat the LPS is pelleted by ultracentrifugation at 200,000 xg for 30 hin aqueous solution whereas the EPS remains in the supernatant(Whitfield and Valvano 1993). Based on this sedimentationdefinition the carbohydrate described above is considered to be anEPS. This term is also appropriate based on the chemical studies ofthe LPS and SÃO that are discussed above. The “rough” LPS and SÃOhad a completely different chemical composition from the EPS inboth CB2A and NA1000 and no KDO, a constituent of all LPS, wasdetected in the EPS fraction (compare Fig. 2 with Table II and TableIII, all in appendix II).It is still possible however, that the EPS fraction is technically alarge species of “smooth” LPS due to the method by which it isattached to the cell surface. The long 4 or 5 sugar repeat structureoligosaccharide might be anchored to the outer membrane byattachment to a single rough LPS molecule (consisting of lipid A andcore oligosaccharide moieties). Such an anchoring arrangement has119been suggested for the group I capsular polysaccharide antigens ofEscherichia coli. (Jann and Jann 1990). Anchoring of the EPS to thesurface might also be mediated by other lipids as has been shown forgroup II capsular polysaccharides of E. coil where the EPS is linkedto the cell surface by phosphatidic acid (Jann and Jann 1990). At thispoint, the means of apparent surface adherence for the CaulobacterEPS is unresolved because of the difficulty in purifying sufficientquantities of the “anchor” portion of an EPS from the large excess ofpolymerized repeat unit after the two regions of the molecule arecleaved.Since the EPS remained on the cell following washing of thecells by centrifugation and resuspension it might be expected thatthis layer would be visible by thin section TEM methods. However,no indication of an EPS layer on cells prepared for thin section TEMby standard methods have been reported even when dyes commonlyused to reveal polysaccharides (eg, ruthenium red) wereincorporated into the procedures (Poindexter 1964, Ravenscroft et a!1991; Smit et al. 1981). Graham et a!. (1991), as part of a largerstudy on the use of freeze-substitution methods, did not visualize anEPS layer in strain NA1000. Yet the same cryofixation/freezesubstitution technique has been used to successfully preserve andvisualize the EPS layer on Leptothrix discophora (Beveridge 1988)and E. coil K30 (Whitfield et al. 1989). However, the EPS of bothCauiobacter strains contain only neutral monosaccharides and thecationic dyes and heavy metals used to stain surface polysaccharides120may not react with the neutral polymers and thus they remain.Other methods of capsular stabilization, such as pre-treatment withantibody directed against the EPS or chemical dehydration andLowicryl embedding (Bayer 1990), may be required to visualize thelayer by transmission electron microscopy. Ravenscroft et al. (1991)visualized an EPS-like structure on C. crescentus CB2A using acryofixation/freeze substitution scanning electron microscopytechnique. In that procedure the sample is sputter coated withheavy metals following cryofixation/freeze substitution. Thus evenneutral molecules, such as the EPS, are rendered electron dense andtherefore be visible when examined by electron microscopy.The composition of the BPS polymers differed sufficientlybetween CB2A and NA1000. In this context it is of interest to notethat CB2A no longer produces an S-layer (Smit et al. 1981) but doesproduce and correctly assemble the S-layer protein from NA1000when rsaA is introduced into CB2A on a plasmid (Smit et al.unpublished; See Fig. 1; lane 17). Apparently S-layer assembly is notaffected by differing EPS molecules. Since the EPS produced by theS-layer attachment mutant JS1001 does not differ from its S-layerattachment competent parent strain, NA1000, it is consideredunlikely that EPS plays a role in S-layer attachment to the cellsurface.4.2 RsaA extraction and in vitro recrystallizationRsaA could be selectively extracted from whole cells of NA1000121using low pH. Coomassie blue stained gels indicated that the extractscontained almost exclusively RsaA (Fig. 13). However, Western blotsof whole cell lysates of NA1000 probed with the unadsorbed antiRsaA sera indicated that the sample used as an antigen containedcontaminating proteins (Fig. 16). The presence of contaminatingnon-RsaA in the low pH extracts indicated that further purificationby gel exclusion chromatography or HPLC would be required if theprotein was to be studied by high resolution methods. However, thepurification of RsaA by the one step low pH extraction was sufficientfor conducting the experiments discussed below.Low pH extraction has also been used for selective purificationof the S-layer of other bacteria including Sp irillum “Ordal”(Beveridge and Murray 1976a), S. putridiconchylium (Beveridge andMurray 1976b,c), Aeromonas hydrophila (Dooley and Trust 1988)and Campylobacter fetus (McCoy et al. 1975). For the Aeromonasand Campylobacter species a low pH extraction procedure using aglycine-HC1 buffer was effective. In contrast, better results inselectively removing RsaA were obtained using HEPES at low pH,recognizing that it is not a buffer in that range. Perhaps with C.crescentus, the protonated amino group of glycine at low pH alsodisrupts other membrane-associated proteins.The reassembly studies with the purified NA1000 proteinprovided definitive data that only RsaA is responsible for the visiblerepeated structure. Previously, Smit et al. (1981) had reported thatS-layer preparations from NA1000 (consisting of shed S-layerfragments isolated by differential centrifugation) contained two122other proteins, the “74K” and “20K”, and membrane material. Itcould not be resolved whether the additional proteins weremembrane-derived or were part of the visible S-layer structure.Since the reassembly experiments reported here involvedpreparations with very little contamination from the 74K and 20Kproteins it seems clear that these additional proteins are not part ofthe S-layer structure. The in vitro reassembly experiments alsoreinforce that calcium is specifically required for S-layer assembly;even the divalent strontium ion, which has a hydrated moleculardiameter most similar to calcium and which has substituted forcalcium in the in vitro reassembly of other S-layers (Beveridge1976c), was unable to replace calcium. However, as discussed below,strontium is capable of mediating in vivo RsaA assembly into an Slayer.4.3 Distribution of RsaA- and SAO-like molecules inenvironmental Ca ulobac ter isolatesNA1000, and its parent CB15 (Poindexter 1964), have beenmaintained in pure culture as laboratory strains for almost thirtyyears. Therefore it is of interest to determine if Caulobacters intheir natural environment possess equivalent cell surfaces as thatfound on strain NA1000 (see Fig. 33). It is generally accepted thatthe cell surface of many “domesticated” bacteria bare littleresemblance to that of strains growing in their natural environment(Beveridge and Graham 1991, Costerton et al. 1981; 1987). This123study was undertaken to determine if the cell surface ofCaulobacters in nature have a similar S-layer and LPS compositionas NA1000 and to determine the degree to which the S-layersproduced by various environmental strains are conserved.Table II is a summary of the results of this study. Theseresults indicate there is a similarity between the S-layers of the FWCisolates and that most of the isolates have a cell surface whichresembles that of NA1000. The similarity was demonstrated atseveral levels. The disruption methods used appeared in all cases tospecifically disrupt and extract the S-layer and the solubilizedprotein was also, in all cases but one (FWC23), immunologicallycross-reactive with anti-RsaA sera. It is conceivable that as in strainNA1000, calcium (or another divalent cation) is required for S-layerattachment or crystallization in the various freshwater isolates andthe two extraction methods used would disrupt calcium-mediatedionic bonding (i.e., EGTA is a calcium-selective chelator and theprotons of the low pH treatment would compete with calcium foranionic sites). In addition, oligosaccharide-containing moleculessimilar to SAO were present in all but one (FWC4) S-layer producingstrain and in most cases the oligosaccharide had at least a degree ofimmunological reactivity with the anti-SÃO sera. It can be arguedthen that there is not only a degree of conservation amongCaulobacter S-layer proteins but also a conservation of an SÃO-likemolecule which may participate in surface attachment. Conversely,in the case of the atypical Caulobacters, when there is no S-layer,124there seems to be a different surface architecture as well. However,more detailed examination of the S-layer-like proteins and SAO-likepolysaccharides would have to be conducted to strengthen theseinitial findings, which indicate that most Caulobacters in theenvironment have similar cell surface features as NA1000.FWC23 and FWC4 were exceptional strains in this study in thatthey could not be grouped with the typical or atypical Caulobacterstrains. A single prominent S-layer-like protein was extracted fromFWC23 and an SAO-like carbohydrate was detected by silver-staining but neither the extracted protein or the carbohydratereacted in Western blots (See Fig. 17A; lane 28; Fig. 18; lane 28; Fig.19; lane 16; Fig. 20; lane 48). A regularly structured array was notidentified on FWC23 by negative-stain TEM. EGTA extraction ofFWC4 yielded a high molecular weight S-layer like band, as well as alarge number of lower molecular weight bands, that reacted with theanti-RsaA sera by Western blotting (see Fig. 17B; lane 16 and Fig 18;lane 13). However, FWC4 lacked an SAO-like polysaccharide asdetermined by silver-staining and Western blotting (see Fig. 19; lane20 and Fig. 20; lane 27). Negative-stain TEM has also failed tovisualize a regular structure on the surface of this strain.In a study by Stahl et al. (1992) involving 16s rRNA analysis ofa number of these strains it was learned that the typical strains are arelatively closely-related subgroup of the freshwater Caulobacters,while examples of the atypical strains were different from thetypical cluster and from each other (Stahl et al. 1992). Nevertheless,125Caulobacters in the typical group were still measurably dissimilar.Since the group of S-layer producing Caulobacters arephylogenetically cohesive, yet clearly different from one another, itwas difficult to predict a priori whether the S-layer proteins wouldbe structurally similar. Indeed, it might be expected that the S-layerproteins of a collection of Caulobacter strains would show significantdifferences because they are not, for example, pathogenic strainswith an S-layer attuned to parasite-host interactions, as with someother S-layer producing species (Dooley and Trust 1988; Dubreuil etal. 1990; Kay et a!. 1984; Murray et a!. 1988). Thus, there mightseem to be little reason for genetic selection to favor a specific Slayer structure, particularly at the level of immunological similarity.The anti-RsaA sera cross-reaction was specific in the Westernblotting experiments of FWC’s, but the degree of labeling wasrelatively uniform between strains and significantly less than thatobtained with RsaA. It seems possible that there are conservedregions in the S-layer proteins that are required for formation andsurface attachment of the paracrystalline structure, while the rest ofthe protein is variable and may be dispensable. RsaA is a member ofthe group of smallest Caulobacter S-layer proteins (ca. 100 kDa) andtherefore may be one that contains the minimal amount of essentialassembly-attachment information. This may mean, in some cases(e.g., FWC39), that more than half of the protein serves some purposeother than essential structure information. Dubreuil et al. (1990)made a similar prediction of structurally nonessential regions in theS-layer of Campylobacter fetus strains.126The immunological findings are also reminiscent of genehybridization studies which analogously showed that the NA1000 Slayer gene (rsaA) could be used to identify most Caulobactersisolated from the environment, since most produced S-layers, butonly under reduced stringency conditions (MacRae and Smit 1991).It was hypothesized that conserved regions of the S-layer genes maybe responsible for the hybridization noted. Therefore, the datapresented in this thesis, that of MacRae and Smit (1991), and that ofStahl et al. (1992), indicates that the degree of S-layer structuralconservation noted between Caulobacter isolates may be aconsequence of common mechanisms of self-assembly, surfaceattachment, and possibly export mechanisms conserved during theevolution of the various Caulobacter strains.Comparative studies between the S-layer proteins of relatedbacterial strains have been conducted in other species. InAeromonas salmonicida there is a significant degree of structureconservation among the S-layer protein of strains isolated fromdiverse locals, as judged by N-terminal protein sequencing, Westernimmunoblot and ELISA analysis of a few strains andimmunofluorescence analysis of a larger group using antibodyprepared against one of the S-layer proteins (Kay et al 1984). On theother hand, with A. hydrophila, there were antigenic differencesamong strains and no N-terminal amino acid sequence homology ofthe S-layer protein between two A. hydrophila strains (Dooley et al.1988). In a similar analysis of Campylobacter fetus there were127significant differences in the S-layer proteins of closely-relatedstrains and the suggestion that some form of antigenic variation wasoccurring (Dubreuil et al. 1990). In Aquaspirillum serpens, twostrains were examined by peptide mapping and immunologicalmethods; there is apparently a degree of similarity between the Slayer proteins (Koval et al. 1988). A more general study of 39Bacillus stearothermophilus strains, focussing primarily onmolecular weight of the S-layer protein and appearance by electronmicroscopy, indicated remarkable variety not only in the presence orabsence of S-layer but also the basic geometry of the paracrystallinestructure and the size of the protein involved (Messner et al. 1984).A similar finding was made for several species of Desulfatomaculumnigrificans (Sleytr et al. 1986b). Studies with strains of Bacillussphaericus also noted variation in presence or absence, molecularweight and antigenicity of the S-layer proteins (Lewis et al. 1987;Word et al. 1983). A study of Bacillus brevis strains, which oftenproduce a double S-layer, showed that the middle wall protein to beimmunologically conserved between strains whereas the outer wallprotein was not (Gruber et al. 1988). Overall, these studies show thatthe degree of S-layer conservation within related strains varies fromspecies to species.4.4 Ionic requirements for C. crescentus NA1000 growthand expression I crystallization of RsaAWild-type C. crescentus NA1000 does not grow in M10Higg128medium and calcium titration experiments indicated that culturesbecame growth rate limited at concentrations less than 250 mMcalcium (Fig. 22). A number of other cations could substitute forcalcium and permit NA1000 to grow in the M10Higg minimalmedium. All divalent cations, with the exception of magnesium, andtrivalent cations tested permitted growth. Growth did not occur inminimal medium supplemented with the mono-valent cationssodium, potassium or lithium. Cell growth in the presence of ionsother than calcium occurred with greater mean generation times andlag periods. Furthermore, the growth rates with these other ionsvaried from experiment to experiment resulting in greater standarddeviations in comparison to calcium grown cells. Negative-stainelectron microscopy determined that cells from non-calciumsupplemented M10Higg medium blebbed large quantities ofmembranous material with the exception of cells grown in strontiumsupplemented M10Higg medium. This observation suggests thatcalcium and strontium act to maintain the integrity of the cellmembranes. Metal ions are known to play an important role inmaintaining the cell membranes of other bacterial species (Beveridge1981). These growth studies indicated that ions other than calciumor strontium could permit growth in M10Higg medium although suchcells were less healthy.The growth characteristics of NA1000 in M10Higg mediumsupplemented with 0.5 mM of one metal ion indicate that calciumand strontium are the preferred ions. The ability of other ions to129substitute for the preferred ions indicates that the membrane issomewhat flexible with respect to the cations used for stabilization ifmembrane stabilization is the role of these cations in the physiologyof NA1000. Magnesium is clearly the poorest substitute for calciumor strontium. The M10Higg medium, containing 2.2 mM magnesium,required an additional 800 mM magnesium to support any growth ofNA1000. Cells cultured in medium containing 3 mM magnesium hadthe longest lag periods and greatest mean generations times. This isin contrast to the metal ion preference shown by Escherichia coliand Pseudomonas aeruginosa where magnesium is the preferred ionfor stabilization of the outer membrane (Coughlin et al. 1983; Ferrisand Beveridge 1986; Nicas and Hancock 1983). Nicas and Hancock(1983) determined that growth medium containing 0.5 mM Mg2produced a wild-type outer membrane whereas medium containing0.02 mM Mg2+ produced altered outer membrane.The ability of various divalent and trivalent metal ions tosubstitute for calcium is somewhat surprising. However, once ametal ion is introduced into an aqueous environment such as aminimal medium, it is difficult to predict what chemical form themetal will adopt. For example, metals can be in the free ion form,adopt a form via the interaction with water molecules, or form acomplex through the interactions with hydroxyl or carbonate species(Collins and Stotzky 1989). Therefore, it is difficult to predict thesize and charge of the species of the metal ion that is active to permitcell growth. In order to unambiguously show that the metal ionspermitting growth are acting to stabilize the cell membranes,130quantitative studies of the metal content of the inner and outermembrane need to be conducted. Unfortunately, at present there isno method available to separate the inner and outer membrane of C.crescentus and all cell disruption methods used to isolate the cellmembrane fraction from the cytoplasm has resulted in thesolublization of large amounts of LPS (Walker and Smit, unpublisheddata). However, quantitation of the cations bound to purified LPSand to whole cells grown in the presence of various metal ions mayprove to be informative.Negative-stain electron microscopy indicated that crystallizedS-layer could be found on the cell surface of NA1000 only whengrown in M10Higg supplemented with calcium or strontium, buthigher concentrations of strontium were required. Higher strontiumconcentrations were also required for the crystallization of S-layersheets in cultures of JS1001. The native template for the S-layer, thecell surface of NA1000, could use a lower concentration of eitherdivalent cation to mediate S-layer crystallization than the templateprovided by apposed S-layer subunits to produce the double sheetsin JS1001 cultures. The in vitro crystallization experiments,discussed in section 4.2, indicated that under those conditionsstrontium would not substitute for calcium. The requirement for asuitable template and the effect that the template has on the cationconcentration required for crystallization is illustrated by Table III.The data indicates that if very high strontium concentrations hadbeen used in vitro, crystallization of RsaA may have occurred. The131effect of template quality on the concentration of cation required tomediate crystallization has been noted in other species. Koval andMurray (1984b) demonstrated that 10 mM calcium was required forS-layer to crystallize on naked envelopes of Aquaspirillum serpenswhereas 0.5 mM calcium would mediate S-layer crystallization onthe denuded cell surface.Unlike C. crescentus NA1000, other Gram-negative S-layerproducing species will grow under sever calcium limitation. Thegrowth of Aquaspirillum serpens VHA, Sp irillumputridiconchylium, and Azotobacter vinelandii in mediumcontaining no added calcium has been studied. A e r o m o n a ssalmonicida has been studied in growth medium containing 0.5 Mcalcium. Aquaspirillum serpens VHA will grow but continuoussubculturing in such medium results in eventual cell lysis (Koval andMurray 1984b). Spirillum putridiconchylium grows, however, thecells exuded membranous material into the medium (Beveridge andMurray 1976c). For both Azotobacter vinelandii (Doran et al. 1987)and Aeromonas salmonicida (Garduno et a!. l992b), the cells growbut produce an S-layer with an altered conformation.The substitution of calcium with a number of cations allowedthe growth of C. crescentus in M10Higg medium, but these cells didnot produce an S-layer. Thus, the cell is somewhat flexible withrespect to the ions that will allow growth but has a strict ionicrequirement for S-layer production. Only strontium can substitutefor calcium to mediate S-layer crystallization. The ability for132strontium to substitute for calcium for in vivo and in vitrocrystallization of S-layer, in other Gram-negative species, has beennoted for Spirillum putridiconchylium (Beveridge and Murray1976c), Azotobacter vinelandii (Doran et al. 1978), Aquaspirillumserpens MW5 (Kist and Murray, 1984), the outer S-layer ofLampropedia hyalina (Austin and Murray 1990), and Spirillumserpens VHA (Buckmire and Murray 1970). The ability forstrontium to substitute for calcium is not limited to the role theseions play in S-layer crystallization but has been noted in a number ofphysiological processes (Huh et al. 1991). For example, theattachment of Rhizobium leguminosarum to the roots of leguminousplants is mediated by a calcium salt bridge between a bacterialderived calcium binding protein, rhicadhesin, and the plant surface,and strontium can replace calcium in this interaction (Smit et al.1991). Calcium and strontium are very similar with respect to manyphysicochemical properties, such as charge character and ionic radii,and this similarity is is often considered the reason that the two ionscan substitute for one another in a number of biological processes(Fenton 1987; Huh et al. 1991; Martell 1961). However, there areonly limited examples of strontium participating in the physiology ofcells in their natural environments (Schultze-Lam and Beveridge1994).It was of interest to determine if the S-layer protein remainedcell associated during growth on ions other than calcium or if it wasreleased into the growth medium. During growth without calcium133Aquaspirillum serpens VHA secretes S-layer into the medium whilethe S-layer of Azotobacter vinelandii and Aeromonas salmonicidaremains attached to the cell surface (Doran et al. 1987; Garduflo et al.1992b; Koval and Murray 1984b). The experiments using liquidcultures indicated that unless cells were grown in the presence ofcalcium or strontium, the S-layer was not cell associated, thusindicating that it was secreted into the medium (Fig. 28). However,when NA1000 cells were scraped off plates and examined for S-layerprotein, it was detected only if the cells were cultured in thepresence of calcium or strontium. The same observation was madefor cultures of JS1001 even though the strain grows very well,although with a slightly greater generation time, in unsupplementedM 10Higg (Fig. 29). At present the location of the S-layer protein isunknown, but three possibilities exist. 1. The S-layer may bedegraded by a protease unless it is folded into a crystallized array.There are no reports in the literature to indicate that C. crescentusproduces a secreted protease, although it is known that unfolded orimproperly folded proteins are often more susceptible to proteases.2. The S-layer may also be folded or aggregated in a form that willnot enter a gel during SDS-PAGE. It is known that RsaA, if boiled inthe presence of SDS, will not enter a gel (Smit et al. 1981) and thatthe macroscopic precipitate, formed in calcium-containing liquidcultures of JS1001, composed of RsaA will not enter a gel unless firstextracted into 8 M urea. However, during growth incalcium/strontium minus liquid medium no macroscopic precipitateis formed in cultures of JS1001. 3. In the absence of calcium or134strontium the S-layer may be blocked at the level of transcription,translation or secretion. It is known that calcium and magnesium actto inhibit transcription of one of the major S-layer gene promoters ofBacillus brevis 47 (Adachi et a!. 1991). C. crescentus also contains agene, flbF, which is very similar to a gene that is conserved inYersinia species, lcrD, that has been implicated in calcium signaltransduction (Piano et a!. 1991; Ramakrishnan et al. 1991; Sanders etal. 1992). So there is some indication that C. crescentus may be ableto sense environmental calcium levels. Clearly, furtherexperimentation into the fate of the S-layer protein during growth onions other than strontium or calcium is required.4.5 Investigations of the genetic basis of the calcium-independent I S-layer attachment-defective phenotypeThe inability to isolate calcium-independent mutants from thetransposon library indicates that the locus defining the calcium-independent I S-layer attachment-defective phenotype is not atarget for Tn5 integration or consists of more than one chromosomalsite. If the latter explanation is true then the calcium-independentS-layer attachment-defective strains did not arise by a single stepmutational event. The method by which the calcium-independentmutants were selected involved an “enrichment” step in liquidmedium before plating and so it is possible that double mutants mayhave been selected (see appendix I; method A).The Tn5 library was screened by Mr. P. Awram using a colony135immunoblot procedure and S-layer attachment-defective mutantswere identified (see appendix I; method C). The cell surface of themutants were extracted with NaCJ[EDTA and the LPS was analyzedby SDS-PAGE and silver staining (Fig. 30). All of the Tn5 S-layerattachment-defective mutants were found to have altered LPSbanding patterns. However, none of the mutants tested were capableof growth in unsupplemented M10Higg medium. This indicates that amutational event resulting in the attachment-defective phenotypedoes not also result in a calcium-independent phenotype. Calculationof the reversion rates from the spontaneous calcium-independent /attachment-defective phenotype to the wild-type phenotype woulddetermine if the calcium-independent mutants resulted from a singleor double mutational event.A cosmid library of NA1000 was electroporated into the Slayer negative and calcium-independent strain JS1004 instead of theparent strain, JS1001, to allow better access of the anti-SAO sera tothe cell surface during an immunoblot screen designed to detectrenewed production of SAO. Smit et a!. (unpublished) has shown thatthe S-layer blocks access of antibody to the SAO. Two cosmids, D12and D13, were isolated that allowed production of SAO in JS1001although at less than wild-type levels. However, JS1001 containingcosmid D12 or D13 were still unable to anchor the S-layer to the cellsurface. Perhaps subcloning the Caulobacter DNA to anotherplasmid vector or deleting extraneous DNA will result in increasedSAO production and S-layer attachment. It has been indicated in136Aeromonas salmonicida and Acinetobacter 199A that transport ofS-layer and LPS are coupled (Belland and Trust 1985; Thorne et al.1976). If this is the case in C. crescentus complementation of the 0-antigen in trans may not result in a functional attachment betweenthe S-layer and the SÃO. However, more studies with cosmid D12and D13 must be undertaken before any conclusions can be made.4.6 Conclusions4.6.1 The relationship between calcium-independenceand loss of SÃO. Many of the experiments presented in this thesiswere designed to characterize the cell surface of the wild-type andthe calcium-independent mutants of C. crescentus with the intentionof discovering the defect in the mutants which rendered them Slayer attachment-defective. Figure 33 is a model of the wild-typecell surface based, in part, on the information obtained in this study.The wild-type cell produces three classes of polysaccharidecontaining molecules: the “rough” LPS, the SÃO and an EPS. The SÃOmolecule was found to be absent in all of the calcium-independentmutants examined whereas the rough LPS and the EPS wereunaltered in the mutants. Thus for C. crescentus growth withoutcalcium apparently required a mutational event that consistentlyresulted in the loss of the SÃO molecule. The attendant phenotypewas that production of RsaA continued and the protein couldcrystallize into an S-layer, if calcium or strontium was available, butthe S-layer did not attach to the cell surface. Since no other137alterations of the surface were detected and it was demonstrated inthese mutants that the attachment-defective phenotype could not beascribed to a change in RsaA it was concluded that SAO was anecessary surface component for the attachment of the S-layer.On noting the absence of SAO in all calcium-independentmutants, it was hypothesized that SÃO had a net negative charge. Ifthis was the case, then the simplest explanation for the necessity todelete SAO in order for the cell to be viable in the absence of calciumwas that calcium binds to and neutralizes the charge on the SÃOmolecule. Without calcium, it was envisioned that charge repulsionbetween adjacent SÃO molecules would destabilize the outermembrane and inhibit cell growth. However, chemical analysis ofpurified SÃO suggested that the 0-antigen is composed of threeneutral molecules. Therefore another hypothesis must be generatedto account for the relationship between a mutation that allows growin the absence of calcium and the SAO-negative phenotype.The growth studies of NÃ1000 and JS1001 in M10Higg mediumclearly demonstrates that wild-type C. crescentus requires thepresence of calcium or strontium ions to grow normally and producean S-layer, whereas calcium-independent mutants no longer havethis ionic growth requirement. Examination of the calciumindependent mutant cell surface revealed that SÃO was not presentand this was suggested as the structural basis for the S-layerattachment-defective phenotype. The Tn5 S-layer attachmentdefective mutants were all found to have an altered smooth LPS,thus strengthening the argument that the S-layer interacts with the138wild-type smooth LPS to remain attached to the cell surface.However, unlike JS1001 the Tn5 mutants did not share the additionalphenotype of being calcium-independent. The Tn5 attachment-defective mutants could not be isolated by plating the library oncalcium-free medium and once isolated, by the immunoblot screen,they were unable to grow in calcium-free liquid medium. Taken as awhole, these results can best be explained by entertaining thehypothesis that the spontaneous calcium-independent mutantsJS1001 and JS1002 did not arise from a single point mutationalevent. The inability to fully complement the calcium-independentmutant JS1004 with a cosmid also supports the notion of more thanone mutational event rather than a single point mutation withpleiotropic effects.If the spontaneous calcium-independent strains are doublemutants and the loss of SAO is not responsible for the calcium-independent phenotype some other alteration in the cell must haveoccurred. It is possible that an alteration in the phospholipidcomposition of these mutants has taken place. C. crescentus has anunusual phospholipid composition when compared to that of othereubacteria in that it produces no phosphatidylethanolamine (DeSiervo and Homola 1980; Contreras et al. 1978; Jones and Smith1979). Approximately 85% of the phospholipids of C. crescentusconsist of the acidic species phosphatidylglycerol and cardiolipin.Johnson and Ely (1977) noted that the addition of calcium to PYEmedium increased both the growth rate and yield of C. crescentus.139Contreras et a!. (1978) suggested that this was a consequence ofcalcium binding to and neutralizing the high negative charge in themembranes produced by the acidic phospholipids. It is known thatdivalent cations, and calcium in particular, interact with andinfluence the structure of acidic phospholipids (Cullis et al. 1983).4.6.2 The role of calcium or strontium in thecrystallization of the S-layer. Although a number of metal ionswere able to replace calcium to allow growth of NA1000 onlystrontium or calcium could mediate S-layer crystallization. Thelocation in the S-layer protein where calcium or strontium acting tomediate crystallization is unknown. Figure 1 in appendix IIillustrates the possible locations where metal ion I proteininteraction may occur. Three potential calcium binding sites on RsaAcan be suggested: 1. Within the S-layer monomer thus altering itsconformation to a form that will crystallize. 2. Between the largedomains of the S-layer monomer allowing crystallization of the unitcell. 3. Between the small domains of the S-layer monomer allowingcrystallization of the unit cells. Site directed mutagenesis of thepredicted calcium binding motifs, identified by the rsaA sequence(Gilchrist et al. 1992), of RsaA may help determine the actual site(s)where calcium acts to allow crystallization.4.6.3 The role of the SÃO in S-layer attachment. Loss oralteration of the SAO has been shown to result in the S-layer140attachment-defective phenotype. Chemical studies of the purifiedSAO indicate that the 0-antigen region if formed from sugars thatmay impart a hydrophobic character on the molecule. It is temptingto envision that the S-layer attaches to the cell surface by theinteraction of hydrophobic regions on the S-layer protein and theSAO. If this is the case, the hydrophobic regions of the protein whichinteract with SAO could also interact between S-layer subunits toform the double S-layer sheets observed in mutant C. crescentusstrains that do not produce SAO. Smit et al. (1992) demonstratedthat the two S-layers forming the non-cell associated sheets,produced by calcium-independent mutants, interact via the surfacesof the S-layer that in the wild-type situation was proximal to the cellsurface. Definitive proof that the SAO molecule is responsible for Slayer attachment may be obtained by studies of the S-layerattachment-defective mutants produced by Tn5 mutagenesis.4.7 SummaryThe information contained in this thesis has enhanced ourunderstanding of the physiology and cell surface architecture of C.crescentus. Methods were devised to identify, isolate and purifythree major cell surface molecules: the S-layer protein, LPS, and EPSof this organism. In vivo and In vitro studies have illustrated theimportance of calcium ions for both the growth of C. crescentus andfor the production and crystallization of RsaA.1415. ReferencesAdachi, T., H. Yamagata, N. Tsukagoshi, and S. Udaka. 1991.Repression of the cell wall protein gene operon in Bacillusbrevis 47 by magnesium and calcium ions. J. Bacteriol.173:4243-4245.Agabian, N., M. Evinger, and G. Parker. 1979. Generation ofasymmetry during development. J. Cell Biol. 81: 123-136.Agabian, N., and B. Unger. 1978. Caulobacter crescentus cellenvelope: Effect of growth conditions on murein and outermembrane protein composition. J. Bacteriol. 133:987-994.Amos, L. A., R. Henderson and P. N. T. Unwin. 1982. Threedimensional structural determinations by electron microscopyof two-dimensional crystals. Prog. Biophys. Molec. Biol. 39:183-231.Ashwell, G., and J. Hickman. 1971. The chemistry of the uniquecarbohydrates of bacterial lipopolysaccharides, p. 235-266. InG. Weinbaum, S. Kadis and S. J. Aji (ed.), Microbial toxins. Vol.IV. Bacterial endotoxins. Academic Press, New York.Austin, J. W., and R. G. E. Murray. 1990. Isolation and In vitroassembly of the components of the outer S Layer ofLampropedia hyalina. J. Bacteriol. 172:3681-3689.Baumeister, W. and H. Engelhardt. 1987. Three-dimensionalstructure of bacterial surface layers, p. 110-154. In J. R. Harrisand R. W. Home (ed.), Electron microscopy of proteins. Vol. 6.142Membranous structure. Academic Press. Inc., London.Baumeister, W., and R. Hegerl. 1986. Can S-layers make bacterialconnexons? FEMS Microbiol. Lett. 36:119-125.Bayer, M. E. 1990. Visualization of the bacterial polysaccharidecapsule, p. 129-175. In K. Jann and B. Jann (ed.), Bacterialcapsules. Current topics in Microbiology and Immunology Vol.150. Springer-Verlag, Berlin.Bayer, M. E. 1979. The fusion sites between outer membrane andcytoplasmic membrane of bacteria: Their role in membraneassembly and virus infection, p. 167-202. In M. Inouye (ed.),Bacterial outer membranes biogenesis and functions. JohnWiley and sons, New York.Belland, J. J., and T. 3. Trust. 1985. Synthesis, export, andassembly of Aeromonas salmonicida A-layer analyzed bytransposon mutagenesis. 3. Bacteriol. 163:877-881.Beveridge, T. J. 1979. Surface arrays on the cell wall ofSporosarcina ureae. J. Bacteriol. 139:1039-1048.Beveridge, T. 3. 1981. Ultrastructure, chemistry and function ofthe bacterial wall. mt. Rev. Cytol. 72:229-3 17.Beveridge, T. 3. 1988. The bacterial surface: generalconsiderations towards design and function. Can. J. Microbiol.34:363-372.Beveridge, T. 3. and L. L. Graham. 1991. Surface layers of bacteria.Ann. Rev. Microbiol. 55:684-705.Beveridge, T. J., and R. G. E. Murray. 1975. Surface arrays on the143cell wall of Spirillum metamorphum. J. Bacteriol. 124:1529-1544.Beveridge, T .J., and R. G. B. Murray. 1976a. Superficial cell-walllayers of Spirillum “Ordal” and their in vitro reassembly. Can.3. Microbiol. 22:567-582.Beveridge, T .J., and R. G. E. Murray. 1976b. Reassembly in vitroof the superficial cell wall components of Spirillumputridiconchylium. J. Ultrastruct. Res. 55:105-118.Beveridge, T .J., and R. G. E. Murray. 1976c. Dependence of thesuperficial layers of Spirillum putridiconchylium on Ca2 orSr2. Can. J. Microbiol. 22:1233-1244.Bhat, R., A. Marx, C. Galanos and R. S. Conrad. 1990. Structuralstudies of lipid A from Pseudomonas aeruginosa PAO1:Occurrence of 4-amino-4-deoxyarabinose. J. Bacteriol.172:6631-6636.Bio-Rad. 1987. The Bio-Rad silver stain. Bulletin 1089. Bio-RadLaboratories. Richmond, CA.Blaser, M. 3., E. C. Gotschlich. 1990. Surface array protein ofCampylobacter fetus: cloning and gene structure. J. Biol. Chem.265:14,529-14,535.Blaser, M. J., P. F. Smith, P. F. Kohier. 1985. Susceptibility ofCampylobacter isolates to the bactericidal activity of humanserum. J. Infect. Dis. 151:227-235Blaser, M. J., P. F. Smith, 3. E. Repine, and K. A. Joiner. 1988.Pathogenesis of Campylobacter fetus infection. Failure of144encapsulated Campylobacter fetus to bind C3b explains serumand phagocytosis resistance. J. Clin. Invest. 81:1434-1444.Boulnois, G. 3., and I. S. Roberts. 1990. Genetics of capsularpolysaccharide production in bacteria, p. 1-17. In K. Jann andB. Jann (ed.), Bacterial capsules. Current topics in Microbiologyand Immunology Vol. 150. Springer-Verlag, Berlin.Bradford, M. M. 1976. A rapid and sensitive method forquantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-254.Buckmire, F. L. A. 1971. A protective role for a cell wall proteinlayer of Spirillum serpens against infection by Bedellovibriobacteriovorus, p. 43. In Bacteriological Proceedings. Seventy-first annual meeting. American Society for Microbiology,Washington, D.C.Buckmire, F. L. A., and R. G. E. Murray. 1970. Studies on the cellwall of Spirillum serpens. I. Isolation and partial purificationof the outermost cell wall layer. Can. J. Microbiol. 16:1011 -1022.Burdett, I. D. J., and R. G. E. Murray. 1974. Electron microscopestudy of septum formation in Escherichia coli strains B and B/rduring synchronous growth. J. Bacteriol. 119:1039-1056.Burnette, W. N. 1981. “Western blotting”: electrophoretic transferof protein from sodium dodecyl sulfate-polyacrylamide gels tounmodified nitrocellulose and radiographic detection withantibody and radioiodinated protein. Anal. Biochem. 112:195-145203.Carl, M., M. E. Dobson, W. M. Ching, and G. A. Dasch. 1990.Characterization of the gene encoding the protectiveparacrystalline-surface-layer protein of Rickettsia prowazekii:presence of a truncated identical homolog in Rickettsia typhi.Proc. Nati. Acad. Sci. U.S.A. 87:8237-8241.Caroff, M., S. Lebbar, and L. Szabo. 1987. Do endotoxins devoid of3-deoxy-D-manno-2-octulosonic acid exist ? Biochem. Biophys.Res. Corn. 143:845-847.Cava, 3. R., P. M. Elias, D. A. Turowski, and K. D. Noel. 1989.Rhizobium leguminosarum CFN42 genetic regions encodinglipopolysaceharide structures essential for complete noduledevelopment on bean plants. I. B acteriol. 171:8-15.Chaicroft, I. P., H. Engelhardt, and W. Baumeister. 1986. Three-dimensional structure of a regular surface layer fromPseudomonas acidovorans. Arch. Microbiol. 144:196-200.Chart, H., D. H. Shaw, E. E. Ishiguro, and T. J. Trust. 1984.Structural and immunochemical homogeneity of Aeromonassalmonicida lipopolysaccharide. 3. Bacteriol. 158:16-22.Chang, J. -3., K. Leonard, T. Arad, T. Pitt, Y. -X. Zhan, and L, -H.Zhang. 1982. Structural studies of the outer envelope ofChiamydia trachomatis by electron microscopy. 3. Mol. Biol.161:579-590.Chester, I. R., and R. G. E. Murray. 1978. Protein-lipidlipopolysaccharide association in the superficial layer ofSpirillum serpens cell walls. J. Bacteriol. 133:932-941.146Clancy, M. J., and A. Newton. 1982. Localization of proteins in theinner and outer membranes of Caulobacter crescentus.Biochimica et Biophysica Acta 686:160-169.Cohen, C., 1977. Protein assemblies and cell form. TIBS 2:5 1-55.Cohen-Bazire, G., W. R. Sistrom, and R. Y. Stanier. 1957. Kineticstudies of pigment synthesis by non-sulfur purple bacteria. J.Cell. Comp. Physiol. 49: 25-68.Collins, Y. E., and G. Stotzky. 1989. Factors affection toxicity ofheavy metals to microbes. p. 31-90. In T. J. Beveridge and R. J.Doyle (ed.), Metal ions and bacteria. Wiley, New York.Contreras, I., L. Shapiro, and S. Henry. 1978. Membranephospholipid composition of Caulobacter crescentus. J.Bacterial. 135:1130-1136.Costerton, J. W., R. T. Irvin, and K. -J. Cheng. 1981. The bacterialglycocalyx in nature and disease. Ann. Rev. Microbiol. 35:299-324.Costerton, J. W., K. -J. Cheng, G. G. Geesey, T. I. Ladd, I. C. Nickel, M.Dasgupta and T. J. Marrie. 1987. Bacterial films in nature anddisease. Ann Rev. Microbiol. 41:435-64.Coughlin, R. T., S. Tonsager, and E. J. McGroarty. 1983. Quantitationof metal cations bound to membranes and extractedlipopolysaccharide from Escherichia coli. Biochemistry22:2002-2006.Cullis, P. R., B. de Kruijff, M. J. Hope, A. J. Verkleij, R. Nayar, S. B.Ferren, C. Tilcock, T. D. Madden, and M. B. Bally. 1983.147Structural properties of lipids and their functional roles inbiological membranes, p. 40-81. In R. C. Aloia (ed.), Membranefluidity in biology, Vol. 1. Academic press, New York.Darveau, R. P., and R. E. W. Hancock. 1983. Procedure for isolationof bacterial lipopolysaccharides from both smooth and roughPseudomonas aeruginosa and Salmonella typhimuriumstrains. J. Bacteriol. 155:831-838.Deh, M. E., J. K. Dzandu, and G. E. Wise. 1985. Sialoglycoproteinswith a high amount of 0-glycosidically linked carbohydratemoieties stain yellow with silver in sodium dodecyl sulfatepolyacrylamide gels. Anal. Biochem. 150:166-173.De Siervo, A. J., and A. D. Homola. 1980. Analysis of Caulobactercrescentus lipids. 3. Bacteriol. 143:1215-1222.Dingwall, A., L. Shapiro, and B. Ely. 1990. Flagella biogenesis inCaulobacter, p. 155-176. In 3. P. Armitage and 3. M. Lackie(ed.), Biology of the chemotactic response. Forty-sixthsymposium of the Society for General Microbiology. CambridgeUniversity Press, Cambridge.Dische, Z. 1947. A new specific color reaction of hexuronic acids. 3.Biol. Chem. 167:189-198.Doig, P., L. Emody, and T. J. Trust. 1992. Binding of laminin andfibronectin by the trypsin-resistant major structural domain ofthe crystalline virulence surface array protein of Aeromonassalmonicida. J. Biol. Chem. 267:43-49.Dooley, 3. S. G., R. Lallier, D. H. Shaw, and T. J. Trust. 1985.148Electrophoretic and immunochemical analysis of thelipopolysaccharides from various strains of Aeromonashydrophila. J. Bacteriol. 164:263-269.Dooley, J. S. G., and T.J. Trust. 1988. Surface protein composition ofAeromonas hydrophila strains virulent for fish: Identificationof a surface array protein. Bacteriol. 170: 499-506.Dooley, J. S. G., R. G. B. Murray, and T. J. Trust. 1988. The S-layerof pathogenic strains of Aeromonas hydrophila, p. 55-59. InU. B. Sleytr, P. Messner, D. Pum, and M. Sara (ed.), Crystallinebacterial cell surface layers. Springer-Verlag, Berlin.Doran, J. L., W. H. Bingle, and W. 3. Page. 1987. Role of calcium inassembly of the Azotobacter vine landii surface array. J. Gen.Microbiol. 133:399-413.Dubreuil, J. D., M. Kostrzynska, J. W. Austin, and T. J. Trust. 1990.Antigenic differences among Campylobacter fetus S-layerproteins. J. Bacteriol. 172:5035-5043.Edwards, P., and 3. Smit. 1991. A transducing bacteriophage forCaulobacter crescentus uses the paracrystalline surface layerprotein as a receptor. J. Bacteriol. 191:5568-5572.Evenberg, D., and B. Lugtenberg. 1982. Cell surface of the fishpathogenic bacterium Aeromonas salmonicida. II. Purificationand characterization of a major cell envelope protein related toautoagglutination, adhesion, and virulence. Biochim. Biophys.Acta 6684:249-254.Ferris, G. F., and T. J. Beveridge. 1986. Physicochemical roles of149soluble metal cations in the outer membrane of Escherichiacoli K-12. Can. 3. Microbiol. 32:594-601.Fenton, D. E. 1987. Alkali metals and group hA metals, p. 1-80.In G. Wilkinson, R. D. Gillard, and 3. A. McCleverty (ed.),Comprehensive coordination chemistry. Pergamon Press,Oxford.Fisher, J. A., J. Smit, and N. Agabian. 1988. Transcriptionalanalysis of the major surface array gene of Caulobactercrescentus. 3. Bacteriol. 170:4706-4713.Fogg, G. C., L. Yang, E. Wang, and M. 3. Blaser. 1990. Surface arrayproteins of Campylobacter fetus block lectin-mediated bindingto type A lipopolysaccharide. Infect. Immun. 58:2738-2744.Galanos, C., 0. Lüderitz and 0. Westphal. 1969. A new method forthe extraction of R-lipopolysaccharides. Eur. 3. Biochem. 9:245-249.Garduño, R. A., E. 3. Y. Lee, and W. W. Kay. 1992a. S-layermediated association of Aeromonas salmonicida with murinemacrophages. Infect. Immun. 60:4373-4382.Garduflo, R. A., B. M. Phipps, W. Baumeister, and W. W. Kay. 1992b.Novel structural patterns in divalent cation-depleted surfacelayers of Aeromonas salmonicida. 3. Struct. Biol. 109:184-195.Garduño, R. A., and W. W. Kay. 1992. A single structural type inthe regular surface layer of Aeromonas salmonicida. 3. Struct.Biol. 108:202-208.Gilchrist, A., J. A. Fisher, and 3. Smit. 1992. Nucleotide sequenceanalysis of the gene encoding the Caulobacter crescentus150paracrystalline surface layer protein. Can. J. Microbiol. 38:193-202.Gilchrist, A., and J. Smit. 1991. Transformation of freshwater andmarine Caulobacters by electroporation. I. Bacteriol. 173:921-925.Gilmore, R. D., Jr., N. Joste, and G. A. McDonald. 1989. Cloning,expression and sequence analysis of the gene encoding the 120kD surface-exposed protein of Rickettsia reckettsii. Mol.Microbiol. 3:1579-1586.Godchaux, W., III, L. Gorski, and E. R. Leadbetter. 1990. Outermembrane polysaccharide deficiency in two nonglidingmutants of Cytophaga johnsonae. J. Bacteriol. 172:1250-1255.Goldman, D., and C. R. Merril. 1982. Silver staining of DNA inpolyacrylamide gels: linearity and effect of fragment size.Electrophoresis 3:24-26.Graham, L. L., R. Harris, W. Villiger, and T. J. Beveridge. 1991.Freeze substitution of Gram-negative eubacteria: General cellmorphology and envelope profiles. J. Bacteriol. 173:1623-1633.Gruber, K., H. Tanahashi, A. Tsuboi, N. Tsukagoshi and S. Udaka.1988. Comparative study on the cell wall structure of protein-producing Bacillus brevis. FEMS Microbiol. Lett. 56:113-118.Hackstadt, T., M. G. Peacock, P. J. Hitchcock, and R. L. Cole. 1985.Lipopolysaccharide variation in Coxiella burnetii: intrastrainheterogeneity in structure and antigenicity. Infect. Immun.48:359-365.151Hastie, A. T., and C. C. Brinton. 1979. Specific interaction of thetetragonally arrayed protein layer of Bacillus sphaericus withits peptidoglycan sacculus. J. Bacteriol. 138:1010-1021.Heide, K., and H. G. Schwick. 1978. Salt fractionation ofimmunoglobulin, p. 7.1-7.11. In D.M. Weir (ed.), Handbook ofexperimental immunology. Vol. 1. Blackwell, Oxford.Hickman, J., and G. Ashwell. 1966. Isolation of a bacteriallipopolysaccharide from Xanthomonas camp estris containing3 -acetamido-3 ,6-dideoxy-D -galactose and D -rhamnose. 3. Biol.Chem. 241:1424-1428.Hitchcock, P. 3., and T. M. Brown. 1983. Morphologicalheterogeneity among Salmonella lipopolysaccharidechemotypes in silver-stained polyacrylamide gels. J. Bacteriol.154:269-277.Hitchcock, P. J., L. Leive, P. H. Makela, E. T. Rietschel, W.Strittmatter, and D. C. Morrison. 1986. Lipopolysaccharidenomenclature - past, present, and future. J. Bacteriol. 166:699-705.Hovmöller, S., A. Sjogren and D. N. Wang. 1988a. The threedimensional structure of bacterial surface layers, p. 60-64. InU. B. Sleytr, P. Messner, D. Pum, and M. Sara (ed.), Crystallinebacterial cell surface layers. Springer-Verlag, Berlin.Hovmöller, A. Sjogren, and D. N. Wang. 1988b. The structure ofcrystalline bacterial surface layers. Prog. Biophys. Molec. Biol.51:131-163.152Howard, L., and D. J. Tipper. 1973. A polypeptide bacteriophagereceptor: Modified cell wall protein subunits in bacteriophage-resistant mutants of Bacillus sphaericus strain P-i. J. Bacteriol.113:1491-1504.Huh, N. W., P. Berkowitz, R. G. Hiskey, and L. G. Pederson. 1991.Determination of strontium binding to macromolecules. Anal.Biochem. 198:391-393.Ishiguro, E. B., W. W. Kay, T. J. Trust. 1981. Loss of virulenceduring culture of Aeromonas salmonicida at high temperature.J. Bacteriol. 148:333-340.Jann, B., and K. Jann. 1990. Structure and biosynthesis of thecapsular antigens of Escherichia coli, p. 19-41. In K. Jann andB. Jann (ed.), Bacterial capsules. Current topics in microbiologyand immunology. Vol. 150. Springer-Verlag, Berlin.Jann, K., and B. Jann. 1977. Bacterial polysaccharides antigens, p.247-287. In I. Sutherland (ed.), Surface carbohydrates of theprokaryotic cell. Academic Press, New York.Johnson, R. C., and B. Ely. 1977. Isolation of spontaneously-derived mutants of Caulobacter crescentus. Genetics 86:25-32.Jones, D. E., and 3. D. Smith. 1979. Phospholipids of thedifferentiating bacterium Caulobacter crescentus. Can. J.Biochem. 57:424-428.Kabir, S. 1986. Composition and immunochemical properties ofthe cell surface proteins of Vibrio cholerae. J. Gen. Microbiol.132:2235-2242.153Kandler, 0., and H. Konig. 1985. Cell envelopes of archaebacteria,p. 413-457. In C.R. Woese and R.S. Wolfe (ed.), The Bacteria. Atreatise on structure and function. Vol. VIII. Archaebacteria.Academic Press, London.Karch, H., H. Leying, and W. Opferkuch. 1984. Analysis ofelectrophoretically heterogeneous lipopolysaccharides ofEscherichia coli by immunoblotting. FEMS Microbiol. Lett.22:193-196.Karkhanis, Y. D., J. Y. Zeitner, J. J. Jackson, and D. J. Carlo. 1978. Anew and improved microassay to determine 2-keto-3-deoxyoctonate in lipopolys accharide of Gram-negative bacteria.Anal. Biochem. 85:595-601.Kay, W. W., B. M. Phipps, R. A. Garduflo, and T. J. Trust. 1988. TheA-layer of Aeromonas salmonicida: Organization and function,p. 44-50. In U. B. Sleytr, P. Messner, D. Pum, and M. Sara (ed.),Crystalline bacterial cell surface layers. Springer-Verlag, Berlin.Kay, W. W., B. M. Phipps, B. E. Ishiguro, R. W. Olafson, and T. J.Trust. 1984. Surface layer virulence A-protein fromAeromonas salmonicida strains. Can. Biochem. Cell Biol.62:1064-1071.Kay, W. W., and T. J. Trust. 1991. Form and functions of theregular surface array (S-layer) of Aeromonas salmonicida.Experientia 47:412-414.Keen, N. T., S. Tamaki, D. Kobayashi, and D. Trollinger. 1988.Improved broad-host-range plasmids for DNA cloning in Gram-negative bacteria. Gene 70:191-197.154Kellenberger, E. 1990. Form determination of the heads ofbacteriophages. Eur. J. Biochem. 190:233-248.Kenne, L., and B. Lindberg. 1983. Bacterial polysaccharides, p.287-363. In G. 0. Aspinall (ed.), The polysaccharides. Vol. 2.Academic Press, New York.Kist, M. L., and R. G. E. Murray. 1984. Components of the regularsurface array of Aquaspirillum serp ens MW5 and theirassembly in vitro. J. Bacteriol. 157:599-606.Koval, S. F. 1988. Paracrystalline protein surface arrays onbacteria. Can. 3. Microbiol. 34:407-414.Koval, S. F., and S. H. Hynes. 1991. Effect of paracrystalline proteinsurface layers on predation by Bdellovibrio bacteriovorus. 3.Bacteriol. 173:2244-2249.Koval, S. F., M. L. Kist, and R. G. E. Murray. 1988. The S-layers ofAquaspirillum serpens, p. 36-39. In U. B. Sleytr, P. Messner,D. Pum, and M. Sara (ed.), Crystalline bacterial cell surfacelayers. Springer-Verlag, Berlin.Koval, S.F., and R. G. E. Murray. 1984a. The isolation of surfacearray proteins from bacteria. Can. 3. Biochem. Cell Biol.62:1181-1189.Koval, S. F., and R. G. E. Murray. 1984b. Effect of calcium on thein vivo assembly of the surface protein of Aquaspirillumserpens VHA. Can. J. Microbiol. 31:261-267.Konig, H., and K. 0. Stetter. 1986. Studies on archaebacterial Slayers. Syst. Appl. Microbiol. 7:300-309.155Koyasu, S., A. Fukuda, Y. Okada. 1980. The penicillin-bindingproteins of Caulobacter crescentus. 3. Biochem. 87:363-366.Kropinski, A. M., D. Berry, and E. P. Greenberg. 1986. The basis ofsilver staining of bacterial lipopolysaccharides inpolyacrylamide gels. 13:29-31.Laemmli, U.K. 1970. Cleavage of the structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.Lelts, V., P. Shaw, L. Shapiro and S. Henry. 1982. Synthesis andutilization of fatty acids by wild- type and fatty acidauxotrophs of Caulobacter crescentus. 3. Bacteriol. 151:1269-1278.Lewis, L., A. A. Yousten, and R. G. E. Murray. 1987.Characterization of the surface protein layers of the mosquitopathogenic strains of Bacillus sphaericus. J. Bacteriol. 169:72-79.Luckevich, M. D., and T. I. Beveridge. 1989. Characterization of adynamic S layer on Bacillus thuringiensis. 3. Bacteriol.171:6656-6667.Lüderitz, 0., K. Jann and R. Wheat. 1968. Somatic and capsularantigens of Gram-negative bacteria, p. 105-228. In M. Florkinand E. H. Stotz (ed.), Comprehensive biochemistry. Vol. 26. PartA, Extracellular and supporting structures. Elsevier publishingcompany, New York.MacRae, J. D., and I. Smit. 1991. Characterization of Caulobacters156isolated from wastewater treatment systems. Appi. Environ.Microbiol 57:751-758.Maniatis, T., E. F. Fritsch, and J. Sambrook. 1983. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.Mandatori, R., and 3. L. Penner. 1989. Structural and antigenicproperties of Campylobacter coli lipopolysaccharides. Infect.Immun. 57:3506-3511.Mandrell, R., H. Schneider, M. Apicella, W. Zollinger, P. A. Rice, and3. M. Griffiss. 1986. Antigenic and physical diversity ofNeisseria gonorrhoeae lipooligosaccharide. Infect. Immun.54:63-69.Martell, A, B., 1961. Some factors governing chelating tendenciesand selectivities in the interaction of ligands with metal ions.Fed. Proc. 20 (suppi. 10):35-38.Masuda, K., and T. Kawata. 1981. Characterization of a regulararray in the wall of Lactobacillus buchneri and itsreattachment to other wall components. J. Gen. Microbiol.124:81-90.Mayer, H., 3. H. Krauss, A. Yokota and 3. Weckesser. 1988. Naturalvariants of lipid A, p. 45-70. In H. Friedman, T.W. Klein, M.Nakano and A. Nowotny (ed.), Advances in experimentalmedicine and biology. Vol. 256. “Endotoxin”. Proceedings of theinternational symposium on endotoxins, a satellite symposiumof the fourth international conference on157immunopharmacology, May 11 - 13, 1988, at Jichi medicalschool, Tochigi-ken, Japan. ©1990 Plenum Press, New York.McCoy, E. C., D. Coyle, K. Burda, L. B. Corbeil, and A. J. Winter. 1975.Superficial antigens of Campylobacter (Vibrio) fetus:Characterization of an antiphagocytic component. Infect.Immun. 11:517-525.Merker, R. I., and J. Smit. 1988. Characterization of the adhesiveholdfast of marine and freshwater Caulobacters. App!. Environ.Microbiol. 54:2078-2085.Merril, C. R., D. Goldman, S. A. Sedman, and M. H. Ebert. 1981.Ultrasensitive stain for proteins in polyacrylamide gels showsregional variation in cerebrospinal fluid proteins. Science211:1437-1438.Mescher, M. F. and J. L. Strominger. 1976. Purification andcharacterization of a prokaryotic glycoprotein from the cellenvelope of Halobacterium salinarium. J. Biol. Chem.251:2005-2014.Messner, P., F. Hollaus, and U. B. Sleytr. 1984. Paracrystalline cellwall surface layers of different Bacillus stearothermophilusstrains. mt. J. Syst. Bacteriol. 34:202-210.Messner, P., and U. B. Sleytr. 1988. Evidence for the glycoproteinnature of eubacterial S-layers, p. 11-16. In U. B. Sleytr, P.Messner, D. Pum, and M. Sara (ed.), Crystalline bacterial cellsurface layers. Springer-Verlag, Berlin.Messner, P., and U. B. Sleytr. 1992. Crystalline bacterial cell-surface layers. Adv. Microb. Physiol. 33:213-275.158Messner, P., and U. B. Sleytr. 1991. Bacterial surface layerglycoproteins. Glycobiology 1:545-551.Miller, 3. H. 1972. Experiments in molecular genetics, Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.Murray, R .G. E., J. S. G. Dooley, P. W. Whippey, and T. 3. Trust.1988. Structure of an S layer on a pathogenic strain ofAeromonas hydrophila. J. Bacteriol. 170:2625-2630.Nermut, M. V., and R. G. E. Murray. 1967. Ultrastructure of the cellwall of Bacillus polymyxa. J. Bacteriol. 93:1949-1965.Newhall, W. I. and R. B. Jones. 1983. Disulfide-linked oligomers ofthe major outer membrane protein of Chlamydiae. 3. Bacteriol.154:998-1001.Nicas, T. I., and R. E. W. Hancock. 1983. Alteration of susceptibilityto EDTA, polymyxin B and gentamicin in Pseudomonasaeruginosa by divalent cation regulation of outer membraneprotein Hi. J. Gen. Microbiol. 129:509-517.Nowotny, A. 1984. Heterogeneity of endotoxins, p. 308-332. In E.T. Rietschel (ed.), Handbook of Endotoxin. Vol. 1. Chemistry ofEndotoxin. Elsevier Science Pub. Co., New York.Oakley, B. R., D. R. Kirsch, and N. R. Morris. 1980. A simplifiedultrasensitive silver stain for detecting proteins inpolyacrylamide gels. Anal. Biochem. 105:361 -363.Ørskov, F., and I. ørskov. 1990. The serology of capsular antigens,p. 43-63. In K. Jann and B. Jann (ed.), Bacterial capsules.Current topics in Microbiology and Immunology. Vol. 150.159Springer-Verlag, Berlin.Ouchterlony, 0. 1949. Antigen-antibody reactions in gels. ActaPathol. Microbiol. Scand. 26:507-515.Pate!, G. B., G. D. Sprott, R. W. Humphry, and T. J. Beveridge. 1986.Comparative analyses of the sheath structures ofMethanothrix concilii GP6 and Methanospirillum hungateistrains GP1 and JF1. Can. J. Microbiol. 32:623-631.Perez-Perez, G. I., M. J. Blaser, and J. H. Bryner. 1986.Lipopolysaccharide structures of Campylobacter fetus arerelated to heat-stable serogroups. Infect. Immun. 51:209-212.Peterson, A.A., and E.M. McGroarty. 1985. High molecular weightcomponents in iipopolysaccharides of S a 1 m o n e ii atyphimurium, Salmonella minnesota, and Escherichia coli. J.Bacteriol. 162:738-745.Phipps, B. M., and W. W. Kay. 1988. Immunoglobulin binding bythe regular surface array of Aeromonas salmonicida. J. Biol.Chem. 263:9298-9303.Piano, G. V., S. S. Barve, and S. C. Straley. 1991. LcrD, amembrane-bound regulator of the Yersinia pestis low-calciumresponse. J. Bacteriol. 173:7293-7303.Poindexter, J. S. 1964. Biological properties and classification ofthe Caulobacter group. Bacteriol. Rev. 28:231-295.Poindexter, 3. S. 1981. The Caulobacters: Ubiquitous unusualbacteria. Microbiol. Rev. 45:123-179.Poindexter, 3. S. 1982. S-layer assembly mutants of Caulobacter160crescentus, abstr. CC II 1, p. 171. Abstr. 82nd Annu. Meet. Am.Soc. Microbiol. 1982.Preston, M. A., and 3. L. Penner. 1987. Structural and antigenicproperties of lipopolysaccharides from serotype referencestrains of Campylobacter jejuni. Infect. Immun. 55:1806-1812.Ramakrishnan, G., 3- L. Zhao, and A. Newton. 1991. The cell cycle-regulated flagellar gene flbF of Caulobacter crescentus ishomologous to a virulence locus (lcrD) of Yersinia pestis. J.Bacteriol. 173:7283-7292.Ravenscroft, N., S. G. Walker, G .G. S. Dutton, and 3. Smit. 1991.Identification, isolation and structural studies of extracellularpolysaccharides produced by Caulobacter crescentus. 3.Bacteriol. 173:5677-5684.Ravenscroft, N., S. G. Walker, G. G. S. Dutton, and 3. Smit. 1992.Identification, isolation and structural studies of the outermembrane lipopolysaccharide of Caulobacter crescentus. I.Bacteriol. 174:7595-7605.Remsen, C. C., S. W. Watson, and H. G. Truper. 1970.Macromolecular subunits in the walls of marine photosyntheticbacteria. J. Bacteriol. 103:254-257.Reynolds, E. S. 1963. The use of lead citrate at high pH as anelectron opaque stain in electron microscopy. J. Cell Biol.17:208-213.Ridgeway, H. F., and R. W. Lewin. 1983. Subunit composition of161goblet-shaped particles from the cell wall of Flexibacterpolymorphus. Can. J. Microbiol. 29:1689-1693.Rietschel, E. T., L. Brade, 0. Hoist, V. A. Kuishin, B. Lindner, A. P.Moran, U. F. Schade, U. Zahringer and H. Brade. 1990. Molecularstructure of bacterial endotoxin in relation to bioactivity, p. 15-32. In A. Nowotny, 3. J. Spitzer, E. J. Ziegler (ed.), Cellular andmolecular aspects of endotoxin reactions. Proceedings of the 1stCongress of the International Endotoxin Society, San Diego.Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg. 1977.Labeling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase I. 3. Mo!. Biol.113:237-251.Sanders, L. A., S. Van Way, and D. A. Mullin. 1992.Characterization of the Caulobacter crescentus flbF promoterand identification of the inferred FlbF product as a homolog ofthe LcrD protein from a Yersinia enterocolitica virulenceplasmid. J. Bacteriol. 174:857-866.Saxton, W. 0., and W. Baumeister. 1986. Principals of organizationin S-layers. 3. Mol. Biol. 187:251-253.Schultze-Lam, S., and T. 3. Beveridge. 1994. Nucleation of celestiteand strontianate on a cyanobacterial S-layer. Appi. Environ.Microbiol. 60:447-453.Shapiro, L. 1976. Differentiation in the Caulobacter cell cycle.Ann. Rev. Microbiol. 30:377-407.Shapiro, L. 1993. Protein localization and asymmetry in thebacterial cell. Cell 73:841 -855.162Simon, R., U. Priefer, and A. Pühler. 1983. A broad host rangemobilization system for in vivo genetic engineering:Transposon mutagenesis in Gram-negative bacteria.Bio/Technology 1:784-790.Sitte, P. 1981. Role of lipid self-assembly in subcellularmorphogenesis, p. 401-421. In 0. Kiermayer (ed.),Cytomorphogenesis in plants. Springer-Verlag, New York.Sleytr, U. B. 1976. Self-assembly of the hexagonally andtetragonally arranged subunits of bacterial surface layers andtheir reattachment to cell walls. J. Ultrastruc. Research.55:360-377.Sleytr, U. B., and P. Messner. 1983. Crystalline surface layers onbacteria. Annu. Rev. Microbiol. 37:311-339.Sleytr, U .B., and P. Messner. 1988. Crystalline surface layers inProcaryotes. J. Bacteriol. 170:2891-2897.Sleytr, U. B., P. Messner, M. Sara, D. Pum. 1986a. Crystallineenvelope layers in Archaebacteria. Syst. Application. Microbiol7:310-313.Sleytr, U. B., P. Messner, D. Pum, and M. Sara. 1988. Prefix, p.VIIn U. B. Sleytr, P. Messner, D. Pum, and M. Sara (ed.),Crystalline bacterial cell surface layers. Springer-Verlag, Berlin.Sleytr, U. B., and M. Sara. 1986. Ultrafiltration membranes withuniform pores from crystalline bacterial cell envelope layers.Appi. Microbiol. Biotechnol. 25:83-90.Sleytr, U. B., M. Sara, Z. KUpcU, and P. Messner. 1986b. Structural163and chemical characterization of S-layers of selected strains ofBacillus stearothermophilus and Desulfatomaculumnigrificans. Arch. Microbiol. 146:19-24.Smit, G., D. M. J. Tubbing, J. W. Kijne, and B. J. J. Lugtenberg. 1991.Role of Ca2 in the activity of rhicadhesin from Rhizobiumleguminosarum biovar viciae, which mediates the first step inattachment of Rhizobiaceae cells to plant root hair tips. Arch.Microbiol. 155:278-283.Smit, J. 1987. Protein surface layers of bacteria, p. 343-376. InM. Inouye (ed.), Bacterial outer membranes as model systems.John Wiley and Sons Inc., New York.Smit, J., and N. Agabian. 1984. Cloning of the major protein of theCaulobacter crescentus periodic surface layer: Detection andcharacterization of the cloned peptide by protein expressionassays. J. Bacteriol. 160:1137-1145.Smit, J., and N. Agabian. 1982a. Cell surface patterning andmorphogenesis: biogenesis of a periodic surface array duringCaulobacter development. J. Cell Biol. 95:41-49.Smit, J., and N. Agabian. 1982b. Caulobacter crescentus pili:Analysis of production during development. Dev. Biol. 237-247.Smit, J., H. Engelhardt, S. Volker, S. H. Smith and W. Baumeister.1992. The S-layer of Caulobacter crescentus: Three-dimensional image reconstruction and structure analysis byelectron microscopy. J. Bacteriol. 174:6527-6538.Smit, J., D. A. Grano, R. M. Glaeser, and N. Agabian. 1981. Periodic164surface array in Caulobacter crescentus: Fine structure andchemical analysis. 3. Bacteriol. 146:1135-1150.Smith, S. H., and R. G. E. Murray. 1990. The structure andassociations of the double S layer on the cell wall ofAquaspirillum sinuosum. Can. 3. Microbiol. 36:327-335.Sonesson, A., E. Jantzen, K. Bryn, L. Larsson and J. Eng. 1989.Chemical composition of a lipopolysaccharide from Legionellapneumophila. Arch. Microbiol. 153:72-78.Stackebrandt, E., A. Fischer, T. Roggentin, U. Wehmeyer, D. Bomar,and 3. Smida. 1988. A phylogenetic survey of budding, and / orprosthecate, non-phototrophic eubacteria: membership ofHyphomicrobium, Hyphomonas, Pedomicrobium,Filomicrobium, Caulobacter and “Dichotomicrobium” to thealpha-subdivision of purple non-sulfur bacteria. Arch.Microbiol. 149:547-556.Stahl, D. A., R. Key, B. Flesher and J. Smit. 1992. The phylogeny ofmarine and freshwater Caulobacters reflects habitat. 3.Bacteriol. 174:2193-2198.Stewart, M., and T. 3. Beveridge. 1980. Structure of the regularsurface layer of Sporosarcina ureae. 3. Bacteriol. 142:302-309.Stewart, M., and T. 3. Beveridge, and T. 3. Trust. 1986. Twopatterns in the Aeromonas salmonicida A-layer may reflect astructural transformation that alters permeability. J. Bacteriol.166:120-127.Stoeckenius, W. 1981. Walsby’s square bacterium: fine structure165of an orthogonal procaryote. 3. Bacteriol. 148:352-360.Thorne, K. 3. I., R. C. Oliver, and A. M. Glauert. 1976. Synthesis andturnover of the regular arranged surface protein ofAcinetobacter Sp. relative to the other components of the cellenvelope. 3. Bacteriol. 127:440-450.Thorne, K. 3. I., M. 3. Thornley, P. Naisbitt, and A. M. Glauert. 1975.The nature of the attachment of a regularly arranged surfaceprotein to the outer membrane of an Acinetobacter sp.Biochim. Biophys. Acta 389:97-116.Trust, T. 3., W. W. Kay, and E. E. Ishiguro. 1983. Cell surfacehydrophobicity and macrophage association of Aeromonassalmonicida. Curr. Microbiol. 9:315-318.Tsai, C. M., and C. E. Frasch. 1982. A sensitive silver stain fordetecting lipopolysaccharides in polyacrylamide gels. Anal.Biochem. 119:115-119.Tsuboi, A., N. Tsukagoshi, and S. Udaka. 1982. Reassembly invitro of hexagonal surface arrays in the protein-producingbacterium Bacillus brevis 47. 3. Bacteriol. 151:1485-1497.Walker, S. G., D. N. Karunaratne, N. Ravenscroft, and I. Smit. 1994.Characterization of mutants of Caulobacter crescentusdefective in surface attachment of the paracrystalline surfacelayer. 3. Bacteriol. In Press.Walker, S. G., S. Fl. Smith, and 3. Smit. 1992. Isolation andcomparison of the paracrystalline surface layer proteins offreshwater Caulobacters. 3. Bacteriol. 147:1783-1792.Wang, E., Z. Pei, and M. I. Blaser. 1990. Antigenic shift in surface166array proteins of Campylobacter fetus, abstr. B-233, p. 65.Abstr. 90th Annu. Meet. Am. Soc. Microbiol. American Societyfor Microbiology, Washington, D.C.Weckesser, 3., and H. Mayer. 1988. Different lipid A types inlipopolysaccharides of phototrophic and related nonphototrophic bacteria. FEMS Microbiol. Rev. 54:143-154.Westphal, 0., 0. Lflderitz, and F. Bister. 1952. Uder die extraktionvon bakterien mit phenol/water. Z. Naturforsch. 7b: 148-155.Whitfield, C., G. Schoenhals and L. Graham. 1989. Mutants ofEscherichia coli 09:K30 with altered synthesis and expressionof the K30 antigen. J. Gen. Microbiol. 135:2589-2599.Whitfield, C., and M. A. Valvano. 1993. Biosynthesis andexpression of cell-surface polysaccharides in Gram-negativebacteria. Adv. Microb. Phys. 35:135-246.Whitfield, 3. F., and R. G. E. Murray. 1956. The effects of the ionicenvironment on the chromatin structures of bacteria. Can. 3.Microbiol. 2:245-260.Wilkinson, S. G. 1977. Composition and structure of bacteriallipopolysaccharides, p. 97 - 175. In. I. Sutherland (ed.),Surface carbohydrates of the prokaryotic cell. Academic Press,New York.Wilson, C. M. 1983. Staining of proteins on gels: Comparisons ofdyes and procedures. Meth. Enzym. 91:236-247.Woese, C. R. 1987. Bacterial evolution. Microbiol. Rev. 51:221-271.Word, N. S., A. A. Yousten, and L. Howard. 1983. Regularly167structured and non-regularly structures surface layers ofBacillus sphaericus. FEMS Microbiol. Lett. 17:277-282.Yang, L., Z. Pci, S. Fujimoto, and M. 3. Blaser. 1992. Reattachmentof surface array proteins to Campylobacter fetus cells. J.Bacteriol. 174:1258-1267.1686 Appendix IAppendix I outlines unpublished experimental methodsconducted by other researchers that are pertinent to this thesis.6.1 Isolation of Calcium-independent mutants.The following method was used for the isolation of calcium-independent mutants (Smit, unpublished). C. crescentus NA1000was grown in PYE medium and then subcultured to M10Higg liquidmedium. Four to six days of incubation were needed to developturbid growth (whereas with calcium-sufficient medium growthoccurred with overnight incubation). The cells were subcultured intothe same medium and incubated for two more days. Cells were thenplated at appropriate dilutions onto M10Higg plate medium. Coloniesthat grew were examined for their S-layer characteristics and theirability to grow in the absence or presence of normal concentrationsof calcium.6.2 Production of anti-SÃO sera.Antisera to the SAO was fortuitously raised during attempts toprepare antibody to the adhesive holdfast of strain CB2A (Merkerand Smit, unpublished). Colloidal gold particles (which bind to theholdfast [Merker and Smit, 19881) were added to cultures of CB2Acells. The cells were harvested by centrifugation, extensively treatedby sonic disruption to break the cells and treated with RNase and169DNase. The preparation was then subjected to CsC1 density gradientcentrifugation (50% CsC1, wlv). The colloidal gold particles (andassociated material) sedimented to the bottom of the gradient; thesewere collected and used for rabbit immunization in a similar fashionto the RsaA immunization. Analysis of cells incubated with the seraby indirect immunofluorescence microscopy showed that the serahad little activity to the holdfast material. However, the cell surfaceof S-layer minus but S-layer attachment competent strains CB2A andJS100I were completely labeled in immunofluorescence andimmunoelectron microscopy experiments using this sera. When Slayer producing strains were examined by the same procedure nocell surface labeling was noted (Smit, unpublished).63 Colony immunoblot for identifying S-layerattachment-defective and S-layer negative mutants.Mr. Peter Awram developed a colony immunoblot screen usinganti-RsaA sera that could differentiate between S-layer attachmentdefective, S-layer negative and wild-type NA1000 strains (Awramand Smit, unpublished). While screening the NA1000 transposonlibrary for mutants that no longer produced RsaA on the cell surface22 S-layer attachment-defective Tn5 mutants were also isolated.1707 Appendix IIAppendix II lists the results of experiments conducted by otherresearchers which are pertinent to this thesis. The source ofpreviously published data is listed in the tables or the figure legends.For unpublished data the researcher who provided the data isacknowledged.171FIG. 1. Three dimensional reconstruction of the S-layer ofCaulobacter crescentus NA1000. A. S-layer monomer. B. The unitcell formed by crystallization of 6 S-layer monomers. C. The S-layerformed by crystallization of unit cells. Thin arrows in A, B and Crepresent possible sites of calcium interaction with the protein. Thisfigure was adapted from Smit et al. (1992).172CA)DaVFIG. 2. Proposed structures for the Caulobacter crescentusexopolysaccharides. (A) The two possible structures for the CB2A EPSrepeating unit. (B) The structure of the NA1000 EPS repeating unit.Gic = D-glucose, Fuc = D-fucose, Gui = D-gulose, Gal = D-galactose andMan = D-mannose. Data from Ravenscroft et al. 1991.174A.—4’ 3)- Gic- (1— 3)- Fuc- (1—0’ 3)- Gic- (1 —0’4 41 1Gic Guior— 3)- Gic- (1— 3)- Gb- (1—3)- Fuc- (1—*4 41 1Gic GuiB.—*4)-Fuc-(1--3)GIc(1. 4)-Man-(1--31Gal1751.- -0.0065(.0L() 0)r—0.0060 c 0)_-0.0-)0o 00450.0040.x1eD.0003I. 200 - 260)0 iiz11000FIG. 3. Analysis of SAO by Laser desorption time of flight massspectroscopy. The peaks are an average of 176 daltons apart.Unpublished data from A. Rüdiger (GBF-Gesellschaft fürBiotechnologische Forschung).176TABLE I. Lipid analysis of Caulobacter crescentus rough LPStZAssignment %3-OH-C12:O 822-OH-C16:1 9C16:O 5C18:1 4a Data from Ravenscroft et al. 1992177Table II. Sugar composition of Caulobacter crescentus rough LPSSugar Residue per molecule2-keto-3-deoxyoctonate 3-L-glycero-D-mannoheptose 2cx-D-glycero-D-mannoheptose 1x-D-mannose 1x-D-galactose 1cDglucoseb 1a Data from Ravenscroft et al. 1992.b phosphorylated178Table III. Sugar composition of SAOSugar %Glycerol 13.64,6-dideoxy-4-amino hexoseb 12.03,6-dideoxy-3-amino hexoseb 15.0Mannose 1.7Glucose 0.2Galactose 0.2D-Glycero-D-manno-heptose 0.2L-Glycero-D-manno-heptose 0.5a D. N. Karunaratne, unpublished (University of British Columbia).b The amino is acetylated (W. -R. Abraham, unpublished [GBFGesellschaft für B iotechnologische Forschung]).179


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