UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The bacterial production of antimicrobial, cationic peptides and their effects on the outer membranes… Piers, Kevin Lee 1993

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

Item Metadata

Download

Media
831-ubc_1994-89430x.pdf [ 3.52MB ]
Metadata
JSON: 831-1.0099224.json
JSON-LD: 831-1.0099224-ld.json
RDF/XML (Pretty): 831-1.0099224-rdf.xml
RDF/JSON: 831-1.0099224-rdf.json
Turtle: 831-1.0099224-turtle.txt
N-Triples: 831-1.0099224-rdf-ntriples.txt
Original Record: 831-1.0099224-source.json
Full Text
831-1.0099224-fulltext.txt
Citation
831-1.0099224.ris

Full Text

THE BACTERIAL PRODUCTION OF ANTIMICROBIAL, CATIONIC PEPTIDESAND THEIR EFFECTS ON THE OUTER MEMBRANES OF’ GRAM-NEGATWEBACTERIAbyKEVIN LEE PIERSB.Sc. (Hons., Genetics), The University of Alberta, 1987A THESIS SUBMITTED IN PARI’IAL FULFILLMENT OF THEREQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(GENETICS PROGRAM)We accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIADecember, 1993© Kevin Lee Piers, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)Department of Microbiology and ImmunologyThe University of British ColumbiaVancouver, CanadaDate December 20, 1993DE.6 (2/88)ABSTRACTNatural polycationic antibiotic peptides have been found in manydifferent species of animals and insects and shown to have broad antimicrobialactivity. To permit further studies on these peptides, bacterial expressionsystems were developed. Attempts to express these peptides with an N-terminalsignal sequence were unsuccessful due to the lability of the basic peptides.Therefore, different fusion protein systems were tested, including fusions toglutathione-S-transferase (GST) and Staphylococcus aureus protein A. For GST,fusions to the defensin, human neutrophil peptide 1 (HNP- 1), or a syntheticcecropin/melittin hybrid (CEME) were generally unstable if found in the solublefraction of lysed cells, but were stable if found as insoluble inclusion bodies. Inthe course of these studies, we developed a novel method of purilring inclusionbodies, using the detergent octyl-polyoxyethylene, as well as establishingmethods for preventing fusion protein proteolytic breakdown. Cationic peptidescould be successfully released from the GST carrier protein with high efficiencyby chemical means (cyanogen bromide digestion) and with low efficiency byenzymatic cleavage (using factor Xa). Fusions of protein A to cationic peptideswere expressed in the culture supernatant of S. aureus clones and after affinitypurification, CNBr digestion and column chromatography, pure cationic peptidewas obtained. CEME produced by this procedure had the same amino acidcontent, amino acid sequence, gel electrophoretic mobility and antibacterialactivity as CEME produced by protein chemical procedures.Three catiorilc peptides, CEME, CEMA and melittin, were all found tohave a broad range of antibacterial activity at concentrations that werecomparable to conventional antibiotics. All three were found to permeabiizethe outer membrane of Pseudomorias aeruginosa and Enterobacter cloacae tolysozynie and the hydrophobic probe l-N-phenylnaphthylamine. CEMAIIpermeabiized membranes at concentrations 2- to 5-fold lower than CEME and20-fold lower than melittin. In some cases, it disrupted membranes better thanpolymyxin B, a known potent permeabilizer. CEMA also had the highestbinding affinity for purified P. aeruginosa LPS and whole cells, although CEMEand melittin also bound strongly.These data are discussed with special reference to the mechanism bywhich these peptides cross the outer membrane of Gram-negative bacteria. It isproposed that they utilize the self-promoted uptake pathway which has beensuggested previously for other cationic antibiotics. As well, the potential for theuse of cationic peptides as therapeutic antibiotics is discussed.IIITABLE OF CONTENTSABSTRACT iiTABLE OF CONTENTS ivLIST OF FIGURES vifiLIST OF TABLES xLIST OF ABBREVIATIONS xiACKNOWLEDGMENTS xiflDEDICATION xivINTRODUCTION 1A. Pseudomonas aeruginosa 11. Clinical Considerations 12. Outer Membrane Structure 23. Outer Membrane Permeabffity and Antibiotic Uptake 64. Self-Promoted Uptake of Antibiotics 9B. Cationic Peptides 111. Introduction 112. Defensins 113. Cecropins 184. Melittin 225. Synthetic and Hybrid Peptides 26C. Fusion Protein Technology 301. Introduction 302. Glutathione-S-transferase 343. Protein A 354. Inclusion Body Formation and Protein Degradation 365. Expression of Cationic Peptides 38D. Aims of This Study 40MATERIALS AND METHODS 42A. Strains, Plasmids and Growth Conditions 42B. Genetic Manipulations 421. General DNA Techniques 422. DNA Fragment Isolation 483. DNA Sequencing 484. Polymerase Chain Reaction 495. Transformation and Electroporation 49iv6. Oligonucleotide Pufication . 50C. Vector Construction 511. Direct Expression Vectors 512. GST Expression Vectors 573. Protein A Expression Vectors 58D. Immunological Techniques 591. Production and Purification of Antibodies 592. Western Blotting 59E. Electrophoresis 601. SDS-PAGE 602. AU-PAGE 60F. Protein Expression 601. Direct Expression 602. GST Fusions 613. Protein A Fusions 61G. Fusion Protein Purification 621. GST Fusion Proteins 622. Protein A Fusion Proteins 63H. Peptide Release 641. Factor Xa 642. Cyanogen Bromide 64I. Peptide Purification 65J. Gel Overlay Assay 65K. Peptide Analysis 661. Preparation of Samples 662. Amino Acid Sequencing and Analysis 66L. Assays 671. Protein Concentration Estimation 67a. Lowry 67b. Dinitrophenylation 672. Killing Assay 683. Lysozyme Lysis 684. 1-N-phenylnaphthylamine Uptake 695. Dansyl Polymyxin B Displacement 69a. Lipopolysaccharide Isolation 69b. Dansyl Polymyxin B Synthesis 70Vc. Assay.706. Minimum Inhibitory Concentration 71RESULTS 73CHAPTER ONE. The Production of Cationic Peptides as GST FusionProteins 73A. Introduction: Why Fusion Proteins Are Necessary 73B. Construction of the pGEX-KP Vector 74C. The Production and Purification of GST/HNP-1 74D. Factor Xa Cleavage of GST/HNP-1 85E. The Production of GST/proHNP- 1 91F. The Production of GST/CEME 94G. The Production and Purification of GST/pr0CEME 96H. The Release of CEME From GST/proCEME Using CNBr 98I. Summary 99CHAPTER TWO: The Production of Cationic Peptides as Protein AFusion Proteins 102A. Introduction 102B. Construction of the Vectors 102C. Expression of pPA-CEME in .E. coiL 104D. Expression and Purification of Protein A/Cationic PeptideFusion Proteins in S. aureus 104E. Release and Purification of CEME 106F. Purification of Other Cationic Peptides 110G. Summary 113CHAPTER THREE: Antimicrobial Activity of Cationic Peptides 114A. Introduction 114B. Killing of P. aerugthosaHl87byCEME 115C. Minimum Inhibitory Concentrations 115D. The Effect of Cations on the MIC of Cationic Peptides 120E. Synergy Studies with Cationic Peptides and Other Antibiotics 121F. Summary 125CHAPTER FOUR: Membrane Permeabilizing Activities of CationicPeptides 126A. Introduction 126B. Lysozyme Lysis Assays 126C. 1-N-phenylnaphthylamine Uptake Assay 130viD. Dansyl Polymyxin B Displacement Assays.134E. Summary 136DISCUSSION 140A. General 140B. Direct Expression 141C. Comparison of Different Fusion Protein Systems 142D. A New Strategy for Solubiizing Inclusion Bodies 148E. Comparison of Fusion Protein Cleavage Methods 150F. Do Cationic Peptides Cross the Outer Membrane via the SelfPromoted Uptake Pathway9 152G. Catiornc Peptides As Therapeutic Agents 156REFERENCES 160VIILIST OF FIGURESFigure 1: Schematic representation of the outer membrane andpeptidoglycan of P. aeruginosa 4Figure 2: Schematic diagram of the self-promoted uptake model 10Figure 3: Amino acid sequences of selected cationic peptides 14Figure 4: Schematic diagram of mammalian defensin structure 16Figure 5: Schematic diagram of cecropin A structure 20Figure 6: Schematic diagram of melittin structure 24Figure 7: Schematic representation of the strategy used to anneal thesix oligonucleotides that encoded the HNP-1 gene 55Figure 8: Construction of pKP19O and pKP 160 56Figure 9: Construction of pGEX-KP 75Figure 10: Construction of GST/cationic peptide fusion proteins 76Figure 11: Production of GST/HNP- 1 77Figure 12: Schematic diagram of GST fusion protein affinitypurification 79Figure 13: Affinity purification of GST/HNP-1 80Figure 14: Purification of GST/HNP- 1 inclusion bodies 84Figure 15: Factor Xa cleavage of GST/HNP1 86Figure 16: Optimal factor Xa cleavage of GST/HNP- 1 87Figure 17: Amino acid sequence of GST 89Figure 18: Cleavage of denatured/renatured GST by factor Xa 90Figure 19: Production and affinity purification of GST/proHNP-1 93Figure 20: Production and affinity purification of GST/CEME 95Figure 21: Production and purification of soluble and insolubleGST/proCEME 97Figure 22: CNBr release of active CEME from GST/pr0CEME 100Figure 23: Construction of protein A/cationic peptide fusion proteins 103Figure 24: Production of protein A/cationic peptide fusion proteins 105VIIIFigure 25: Bio-Gel P100 chromatography of CNBr-digested PA/CEME 108Figure 26: Reverse phase chromatography of CEME 109Figure 27: Antibacterial activity of CEME 112Figure 28: Killing of P. aerugirtosa H187 by recombinant CEME 116Figure 29: Peptide-mediated lysozyme lysis of P. aerugiriosa H309 128Figure 30: Peptide-mediated lysozyme lysis of E. cloacae 218R1 129Figure 31: Peptide-mediated NPN uptake in P. aeruginosa H309 132Figure 32: Peptide-mediated NPN uptake in E. cloacae 218R1 133Figure 33: Inhibition of dansyl polymyxin B binding to P. aeruginosaH103 LPS by various compounds 135Figure 34: Inhibition of dansyl polymyxin B binding to P. aeruginosaH309 whole cells by various compounds 138ixLIST OF TABLESTable I: Cationic Peptides 12Table II: Commonly Used Affinity Tags 31Table III: Methods for the Site-Specific Cleavage of Fusion Proteins 33Table 1V: Strains 43Table V: Plasmids 45Table VI: Oligonucleotides 52Table VII: Summary of Strains Used to Prevent ProteolyticDegradation of GST/HNP- 1 82Table VIII: Percent Yield of CEME from a 601 Fermentor Run 111Table IX: MIC Values of Various Antimicrobial Agents 117Table X: Effects of Mg2 and Na Cations on the MICs of CationicPeptides Against P. aeruginosa H309 122Table XI: Effects of Sub-MIC Levels of Cationic Peptides on the MICsof Common Antibiotics 123Table XII: Values for Various Compounds Against P. aeruginosaLPS and Whole Cells 137xLIST OF ABBREVIATIONSAp ampicfflinAU acid ureabp base pairCEMA CEME analogueCEME cecropin/melittin hybrid peptide (CA1-8M1-18)CFU colony forming unitCm chioramphenicolCNBr cyanogen bromideDEAE diethylaminoethylEDTA ethylenediamine tetraacetic acidFPLC fast protein liquid chromatographyGST glutathione-S-transferaseHEPES (N-2-hydroxyethylpiperazine-N-2-ethanesulfo ic acid) sodium saltHNP human neutrophil peptideIGF-I human insulin-like growth factor IIgG immunoglobulin GIPTG isopropyl-f3-D-thiogalactosideKan kanamycinkb kilobase pairkDA kilodaltonKCN potassium cyanideLB-S Luria-Bertani, no salt brothLBNS Luria-Bertani, normal salt brothLPS lipopolysaccharideMCP macrophage cationic peptideMIC minimum inhibitory concentrationxiMTPBS mouse tonicity phosphate-buffered salineNMR nuclear magnetic resonanceNPN 1 -N-phenylnaphthylamineOD optical density0-POE octyl-polyoxyethylenePA truncated S. aureus protein APAGE polyacrylamide gel electrophoresisPCR polymerase chain reactionPEG polyethylene glycolPMBN polymyxin B nonapeptidePMSF phenylmethylsulfonyl fluorideRP reverse phaserpm revolutions per minuteSDS sodium dodecyl sulfateTet tetracyclineTFA trifluoroacetic acidTLCK Na-p-tosyl-L-lysine chloromethyl ketoneTPCK N-tosyl-L-phenylalanine chloromethyl ketoneTSS transformation and storage solutionTST 50 mM Tris-HC1 pH 7.6, 150 mM NaC1, 0.05% Tween 20UV ultra violetWCLB whole cell lysing bufferZ synthetic IgG binding domainXIIACKNOWLEDGMENTSThe financial support of the Medical Research Council of Canada isgratefully acknowledged.I would like to thank Dr. Bob Hancock for his enthusiastic guidancethroughout this work. I would also like to thank all the members of his lab overthe duration of my stay for providing a more than pleasant workingenvironment.Specifically, I would like to say a special thank you to the followingpeople:- the 233 gang (Eileen, Rebecca and Anand) for all the laughs and thestimulating discussions (some of them even about science!)- Nancy, Susan and Manjeet for all their help- my fellow hockey poolers, especially the “core” members Terry, Loverneand Mo: thanks for the money!- Doug Kilburn for the “timely suggestion” and George Spiegelman forcritically reading this thesisFinally, I would like to thank my parents, Ed and Shirley, who alwaysencouraged me in my studies and provided me with the opportunities to pursuethem.XIIIDEDICATIONThroughout the duration of these studies, my wife, Gloria, has been theone person to share every up and down. Whether I needed a shoulder forsympathy, an ear for my frustrations or a smile of encouragement, she alwayshad them just at the right time. This thesis, therefore, is not only a reflection ofmy work, it is a testimony to the love, patience and dedication that she hasshown to me over these last years. Gloria, it is to you that I dedicate thisthesis.xivINTRODUCTIONA. Pseudomonas aeruginosa..1. Clinical Considerations.Pseudomonas aeruginosa is a Gram-negative, rod-shaped organism thatcan grow with limited minerals on a large number of carbon and energysources. These minimal growth requirements enable P. aerugirtosa to exist invery diverse environments, including soil, standing water and sediments. Italso has other properties such as slime layers and adhesins that make itdifficult to remove from colonized sites by mechanical means. P. aerugirtosa isnot usually a pathogen for healthy human beings, since it is unable topenetrate the host’s first defense barrier, namely the skin and mucosa.However if that physical barrier is compromised and host defenses are weak, P.aeruginosa becomes prevalent in human infections of bums, eye injuries, andsites of major surgery and medical interventions involving instruments such ascatheters or shunts. Another occurrence of P. aeruginosa infections is inpatients whose non specific defense or immune systems are under duress, suchas premature infants, patients with cystic fibrosis or neoplasia, and patientswho are undergoing immunosuppressive drug treatment or whole bodyirradiation treatment. P. aeruginosa can therefore be categorized as anosocomial opportunistic human pathogen.Many different toxins which are produced by P. aeruginosa have beendescribed (Doring et al., 1987) and it has been suggested that these may allcontribute variably in different P. aeruginosa infections (Nicas and Iglewski,1985). Given this multifactorial virulence, it is not suprising that P. aerugiriosais able to cause a wide variety of different medical problems such as1bacteremia, urinary tract infections, endocarditis and gastrointestinal infections(Pollack, 1990).The clinical significance of this pathogen has been increasing steadily.During the period between 1975 and 1984, the relative frequency of P.cterugiriosa in nosocomial infections (expressed as a percentage of all isolates)increased from 6.3% to 11.4%, making it second only to E. coil as the mostfrequently acquired hospital pathogen. Not only is the increase in P. aeruglnosainfections a cause for concern, but also the fact that these infections areassociated with a high mortality rate (Young, 1984; Hancock and Bell, 1989).The increase in the incidence of P. aeruginosa infections is largely atestimony of its resistance to many forms of antimicrobial therapy (Bryan,1979). Antibiotics that have been used to treat P. aeruginosa in the past, suchas streptomycin, carbenicihin, tetracycline, chloramphenicol and trimethoprim,have become ineffective due to the emergence of resistant strains (Bryan, 1979).One of the characteristics of P. aeruginosa that confers such intrinsic resistanceis the low permeability of its outer membrane (Angus et al., 1982). Althoughthe outer membrane of P. aeruginosa is not remarkably different to those ofother Gram-negative organisms, its permeability has been estimated to betwelve-fold (Nicas and Hancock, 1983b) to 100-fold (Yoshimura and Nikaido,1982) lower than that of the E. coil outer membrane.2. Outer Membrane Structure.The envelope of P. aeruginosa is made up of three distinct components:the cytoplasmic or inner membrane, the periplasmic space and the outermembrane. The cytoplasmic membrane is a phospholipid bilayer membranethat is studded with as many as one hundred different proteins (Cronan et al.,21987). The periplasm can be dissected into inner and outer periplasmic spaces.The former is relatively un-crosslinked and therefore appears as a gelatinusstructure (Hobot et al., 1984). The latter, conversely, is highly crosslinked andplays a large role in the structural integrity of the cell (Oliver, 1987). It containspeptidoglycan (murein) which is a macromolecular combination ofpolysaccharides and peptide (Park, 1987). In addition to these components, theperiplasm is filled with an array of proteins that carry out binding, scavengingand detoxifying functions (Oliver, 1987).The outer membrane of P. aeruginosa, unlike the uniform cytoplasmicmembrane, is an asymmetric lipid bilayer (Figure 1). The inner leaflet consistsof the chloroform-methanol extractable phospholipids that includephosphatidylethanolamine, phosphatidylglycerol, diphosphatidyiglycerol(cardiolipin), and some other unknown phospholipids (Nikaido and Hancock,1986). In all cases, these phospholipids possess two fatty acid chains thatextend from the polar head group. Interspersed among the phospholipids arelipoproteins such as OprL and OprI. The latter, which is the Pseudornonasequivalent of the E. coli Braun Lipoprotein (Mizuno, 1979), is covalentlyattached to the peptidoglycan (Mizuno and Kageyama, 1979) thus providing an“anchor” for the outer membrane. In addition to these lipoproteins, there aremany other different outer membrane proteins in P. aeruginosa (for reviews, seeNikaido and Hancock, 1986; Hancock et al., 1990).The component of the Gram-negative outer membrane that distinguishesit from the cytoplasmic membrane is the lipopolysaccharide (LPS), which isfound only in the outer leaflet (Smit et al., 1975). It is an amphipathic moleculeconsisting of three domains: a) the hydrophific 0-antigen side chain thatextends outward from the cell, b) a core polysaccharide region consisting ofvarious conserved components, and c) the hydrophobic lipid A tail which3Figure 1: Schematic representation of the outer membrane and peptidoglycanof P. aerugiriosa.4anchors the molecule into the membrane. Lipid A consists of a 4-phosphoglucosaminyl- (1-6) -glucosamine- 1-phosphate backbone (Wilkinson,1983) to which four to six saturated fatty acids are either ester or amide linked(Karunaratne et al., 1992). These linkages can be either directly on the sugarbackbone, or through the hydroxyl groups of other fatty acid chains (Kropinskiet at, 1985). The fatty acid composition of the lipid A has been found toinclude 3-hydroxydodecanoic acid, 2-hydroxydodecanoic acid, 3-hydroxydecanoic acid, and decanoic acid (Kropinski et al., 1985) as well asminor amounts of others (Pier et at, 1981). The core region of LPS consists ofD-glucose, D-galactosamine, L-rhaninose, L-glycero-D- mannoheptose and 2-keto-3-deoxyoctulosonic acid (KDO) (Rowe and Meadow, 1983; Wilkinson, 1983;Kropinski et at, 1985). Besides these saccharide components, the core alsocontains L-alanine, ethanolamine and phosphate (Rowe and Meadow, 1983;Wilkinson, 1983). Although the core constituents remain the same in differentP. aerugirtosa strains, variations in sugar ratios, sugar arrangement andphosphate composition have been found (Kropinski et at, 1985; Rivera et al.,1988a). The 0-antigen side chain, conversely, varies greatly between strains.This highly immunogenic component of LPS is composed of unbranched tn- ortetra-saccharide repeat units (Kropinski et al., 1985) which are rich in Nacetylated sugars and whose repeat number varies from strain to strain. The0-antigen side chain structures of all P. aeruginosa serotypes have now beenelucidated (Yu et al., 1988).Another key feature of the outer membrane is the interaction betweenLPS molecules. The presence of negatively charged groups on or near thedisaccharide backbone of LPS, combined with the juxtaposition of the LPSmolecules in the outer membrane, suggest that a strong electrostatic repulsionwould exist thus causing a destabilization of the membrane (Nikaido and Vaara,51985). It has been found that neighboring LPS molecules are stabilized bydivalent cation bridges (Schindler and Osbom, 1979). The importance of thesecations, such as Mg2 and Ca2, is shown by treating intact cells with thedivalent cation chelator ethylenediaminetetraacetate (EDTA) which disrupts thecells and results in the loss of LPS molecules (Leive, 1965). The combination ofLPS and divalent cations results in a stable hydrophilic barrier that is virtuallyimpermeable to many hydrophobic molecules (see below). In addition todivalent cations there are protein-LPS interactions that are also crucial tomaintaining outer membrane integrity. An example is P. aerugthosa protein Hiwhich is induced in low Mg2 conditions and results in increased resistance topolymyxin B, gentamicin, and EDTA (Nicas and Hancock, 1980). It has beenhypothesized that since OprH has a putative p1 of 8.6 and therefore is positivelycharged at physiological pH, it will interact with the negatively charged LPS inlimiting divalent cation conditions (Bell et al., 1991), resulting in a stable outermembrane.3. Outer Membrane Permeability and Antibiotic Uptake.The outer membrane of P. aeruginosa has been shown to contributesignificantly to the intrinsic resistance of this organism to antibiotics by formingan effective permeability barrier. In addition to antibiotics, this barrier restrictsthe access of many hydrophobic and large hydrophiic compounds (for reviews,see Nikaido and Vaara, 1985; Nikaido and Hancock, 1986). There is a group ofcompounds able to disrupt the organization of the outer membrane and allowthe uptake of normally impenetrable molecules. This group is collectivelyknown. as outer membrane permeabifizers.6The antibiotic polymyxin B is one of the most well studiedpermeabilizers. Its structure consists of a polycationic decapeptide ‘head’region (in which seven of the amino acids form a cyclic structure) and a fattyacyl tail. The fatty acyl tail has been implicated in the association with, andsubsequent lysis of, the cytoplasmic membrane (Teuber and Bader, 1976), sincepolymyxin B nonapeptide (PMBN), a polymyxin B derivative that lacks this fattyacyl chain, shows very weak killing activity (Vaara and Vaara, 1983). Bothpolymyxin B and PMBN exhibit potent membrane permeabiizing activity on theouter membranes of enteric bacteria as evidenced by their enhancement ofhydrophobic antibiotic uptake (Vaara and Vaara, 1983). Polymyxin B was alsoshown to permeabilize the outer membrane of P. aeruginosa to the proteinlysozyme, the hydrophobic fluorescent probe 1-N-phenylnaphthylamine (NPN)and the chromogenic -1actam nitrocefin (Hancock and Wong, 1984). Thispermeabiization could be antagonized by the addition of Mg2, indicating thatpolymyxin B must initially associate with the negatively charged sites on theLPS molecules that are normally occupied by Mg2. Evidence supporting thisassociation came from the observation that dansylated polymyxin B had astrong (0.4 riM) affinity for purified LPS and that Mg2 could competitivelydisplace the dansyl polymyxin from the LPS (Moore et al., 1986).Studies with bacteria that are resistant or sensitive to polymyxin B orother polycations seem to corroborate the theory that polycationic compoundsinitially interact with the negatively charged sites on LPS. A P. aeruglnosa tolAmutant that was susceptible to aminoglycosides was shown to have altered LPSthat resulted in an increased affinity for polymyxin B (Rivera et al., 1988b).Conversely, poiymyxin resistant (pmr) mutants of S. typhimurium (Makelä et al.,1978) and E. coil (Meyers et al., 1974) have been isolated and shown to haveincreased esterification of their LPS molecules (Vaara et al., 1981; Peterson et7at, 1987). The esterification results in a less acidic LPS and consequently in adecreased binding affinity for polymyxin B and other cationic compounds.Antibiotics are taken up through a number of different pathways. Many13-lactams are believed to cross the outer membrane by diffusing through thesmall water filled channels of porin proteins (Zimmermann and Rosselet, 1977).Much of the evidence supporting this hypothesis comes from the fact that some3-1actam MIC values for porin-deficient mutants are much higher than theirisogenic wild type strains (Hancock and Bell, 1989). Some of these mutants,however, still demonstrate residual uptake of antibiotics such aschioramphenicol and tetracycline, suggesting that there may be otheralternative pathways (Hancock and Bell, 1989). In some cases, porins involvedin substrate-specific uptake may become an uptake pathway for an antibioticthat mimics the structure of the substrate. For example, the 13-lactamimipenem is taken up through the basic amino acid-specific channel OprD in P.aerugirtosa (Trias and Nikaido, 1990a; Trias and Nikaido, 1990b).A second pathway, the partitioning of hydrophobic antibiotics into theouter membrane, is virtually non existent in wild type P. aeruginosa, asindicated by high MIC values for these hydrophobic antibiotics (Nikaido andHancock, 1986) and the inability of NPN to penetrate the outer membrane (Lohet al., 1984). The uptake of these compounds can occur in deep rough mutants(Hancock and Bell, 1989) or be enhanced by membrane-disrupting compoundssuch as EDTA (Loh et at, 1984), suggesting that the stabifized LPS interactionson the surface of the outer membrane, serve as a barrier to this group ofantibiotics.84. Self-Promoted Uptake of Antibiotics.Mutant P. aeruginosa cells that overproduce the Hi protein (OprH) andcells grown in limited Mg2 conditions (which induce OprH production) werefound to be resistant to polymyxin B and gentamicin (Nicas and Hancock, 1980)but had no change in the susceptibility of to other antibiotics such as f3-lactamsand tetracyclines (Nicas and Hancock, i983a). From these data it washypothesized that OprH blocked an uptake pathway that was utilized bycationic compounds. Given the membrane permeabilizing activities (Hancockand Wong, 1984) and LPS binding capabilities (Moore et al., 1986) of thesecationic compounds, a third antibiotic uptake pathway, called self-promoteduptake, was proposed (Figure 2; Hancock et al., 1981). This pathway involvesthree steps: (1) the initial displacement of the Mg2 ions from the LPS cross-bridges by the cationic antibiotic (Figure 2B); (2) the subsequent disruption ofthe outer membrane structure which can permeabiize it to other, normallyexcluded compounds (Figure 2C); and (3) the eventual uptake of the compounditself into the periplasmic space (Figure 2D). The nature of the final uptakeprocess is not well understood, but its efficacy is undoubtedly related to boththe structure of the disrupted membrane and the conformation of thecompound once it has bound to the LPS. Recently, Sawyer et al. (1988) wereable to show that the interactions of antimicrobial rabbit macrophage defensins(see below) with the outer membrane of P. aeruginosa were consistent with theself-promoted uptake hypothesis (discussed below).9A BFigure 2: Schematic diagram of the self-promoted uptake model.A, Typical Gram-negative bacterial outer membrane with a Mg2crossbridge (outer membrane proteins are omitted for simplicity); B,Displacement of Mg2 by a cationic antibiotic (polymyxin B, Px, is used as anexample) and the initial perturbations of the lipid bilayer; C, Further disruptionof the outer membrane resulting in the uptake of normally excludedcompounds; D, Uptake of the antibiotic across the outer membrane.•Mg.C Nitrocefin- NPND10B. Cationic Peptides.1. Introduction.Antibiotics produced by microorganisms have long been used to fightmany infectious diseases. Within the past several decades, however, a newbreed of antimicrobial agents, termed “peptide antibiotics”, has been discovered.Zeya and Spitznagel (1966) were the first to describe such peptides inmammalian polymorphonuclear leukocytes and noted that they tended to becationic in nature. Since then, further investigation has led to the discovery ofmany different peptides from a wide range of organisms (examples in Table I).Considering the location of some of these peptides in the organisms, it is notcertain that their primary role is host defense. Nonetheless, they all show abroad range of antimicrobial activity and therefore are potentially useful inantibiotic therapy.2. Defensins.The term “defensins” was originally used to describe a family ofantimicrobial and cytotoxic peptides from mammalian neutrophils that rangedfrom 29-35 amino acids in length (for reviews, see Ganz et al., 1990; Lehrer etaL, 1990; Lehrer et at, 1993). They are invariably cationic, containing betweenfour and ten arginine residues, and include six conserved cysteine residues thatform three disulfide bonds (Figure 3). The first, N-terminal cysteine forms adisulfide bond with the last, C-terminal cysteine, resulting in an effectivelycyclic peptide (Seisted and Harwig, 1989). Defensins constitute 5-7% of thetotal cellular protein in neutrophils and 30-50% of the total protein in theneutrophil’s primary granules (Rice et at, 1987). They are actually synthesized11TableI:CationicPeptides.PeptideOriginSizeaStructurebReferenceIndolicidinBovineneutrophils13UndeterminedSeisted,etat,1992ApidaecinsApisme11feragut18UndeterminedCasteels,etat,1989MagaininsXenopuslaevisskin23he1Zasloff,1987MelittinApisme11feravenom26he1HabermannandJentsch,1967DefensinsHuman29-3013-sheetGaflZ,etal.,1985Rabbit3334f3-sheetSeisted,etat,1984GuineaPig3113-sheetSeistedandHarwig,1987Rat29-3213-sheetEisenhauer,etal.,1989Equine4613-sheetCouto,etat,1992CecropinP1Pigsmallintestine31heLee,etat,1989CryptdinMousePanethcells3513-sheetEisenhauer,etat,1992CecropinsHyalophoracecropia3537a-helixSteiner,etal.,1981con’t...r\)TableI:CationicPeptides(cont).PeptideOriginSizeaStructurebReferenceCharybdotoxinScorpionvenom37p-sheetMiller,etaL,198513-defensinsBovineneutrophils38-42UndeterminedSelsted,etal.,1993SarcotoxinsSarcophagaperegrina39a-helixOkadaandNatori,1983PhormicinsPhonniaterrartovae40MixedLambert,etal.,1989SapecinSarcophagaperegrirta40MixedMatsuyamaandNatori,1988SeminalpiasminBovinesemen48a-helixTheilandScheit,1983BactenecinsBovineneutrophils42UndeterminedFrank,etal.,1990DiptericinPhorniiaterranovae82UndeterminedDimarcq,etal.,1988HymenoptaecinApismellferagut93UndeterminedCasteels,etal.,1993AttacinsHyalophoracecropia187-188UndeterminedHuitmark,etal.,1983a,numberofaminoacidresiduesb,predominantsecondarystructurefoundinthepeptideC,)HNP- 1Ii I IIACYCRIPACIAGRRYGTCIYQGRLWAFCCCecropin AKWKLFKKIEKVGQNIRjGIIKAGPAVAVVGQATQ1AKMelittinGIGAVLKVLTTGLPALISWIKRKRQQCEMEKWKLFKKIGIGAVLKVLTPGLPALISCEMAKWKLFKKIGIGAVLKVL’ITGLPALKLTKdefensin pre pro regionpre: MRTLAILAAILLVALQAQApro:Figure 3: Amino acid sequences of selected cationic peptides.Designations include: HNP-1, human neutrophil peptide 1; CEME,cecropin/melittin hybrid (CA(1-8)M(1-18)); CEMA, a frameshift analogue ofCEME. Positively charged amino, acids are in bold and negatively chargedamino acids are underlined. The disuffide bond array for mammalian defensinsis demonstrated for HNP- 1.14as 94-95 amino acid prepropeptides. The negatively charged residues in thepro region (Figure 3) virtually balance the positively charged residues in thedefensin sequence, which led to speculation that the pro region was necessaryfor host protection prior to processing and packaging of the mature defensininto the granules (Michailson et at, 1992). From the granules, the peptides arereleased into the microbe-containing phagosome through degranulation wherethey exert their microbiocidal activity. The importance of defensins to the hostdefense system is evidenced by the high infection rate that occurs in patientswhose granular components are defective or missing (Ganz et at, 1988).Defensins show a broad range of antimicrobial activity. Atconcentrations of 10-100 jig/mL, they are able to kill Gram-negative and Gram-positive bacteria, although the are generally more active against the latter(Lehrer et al., 1993). For Gram-negative bacteria, rough strains are moresensitive to defensin activity than their isogenic smooth strains (Ganz et al.,1990). In addition to bacteria, various defensins have been shown to killCandida albicans (Selsted et al., 1985), Treponema paUidum (Borenstein et al.,1991), and some enveloped viruses such as herpes simplex virus (Lehrer et al.,1985).The first information concerning the 3-dimensional structure ofdefensins came with the elucidation of the solution structure of rabbitneutrophil defensin NP-S (Pardi et at, 1988). The basic structural backbone ofthe defensin molecule was a triple-stranded antiparallel 13-sheet (Figure 4A),with the charged amino acids at one end of the molecule and the hydrophobicones at the other end. This basic structure was confirmed by the crystalstructure of HNP-3 (Hill et al., 1991) which showed the human defensin as anelongated ellipsoid. The HNP-3 crystallized as a basket-shaped dimer (Figure4B) which, due to the monomer’s amphipathicity, had an apolar base and a15ABFigure 4: Model of the mammalian defensin structure.A, The triple-stranded anti-parallel 13-sheet structure of an HNP-3monomer. The disuffide bonds are represented by lightening bolts. B, An HNP3 dimer with a poiar top and an apolar bottom. Both figures were reproducedby copyright permission from Hifi, et al. (1991) © American Association for theAdvancement of Science.016polar top. The invariant Gly’8 in the defensin amino acid sequence was foundat the dimer interface, suggesting that dimerization is required for activity. Thisis further supported by the fact that upon dimerization, a water-filled channel isformed. These observations led to the speculation that HNP-3 monomers wedgetheir hydrophobic base into lipid membranes, and eventually form dimer-poresor multimer-pores (Hill et al., 1991). This is consistent with the data fromKagan et al. (1990) who showed that rabbit defensins can form voltagedependent ion-permeable channels in planar lipid bilayers.These structural data, however, do not provide any information on howthese peptides cross the asymmetric outer membrane in Gram-negativebacteria. Sawyer et al. (1988) showed that rabbit MCP-i and MCP-2 (equivalentto NP-i and NP-2) were able to bind to and permeabilize the outer membrane ofP. aemginosa and that these defensins could bind purified LPS. These dataresulted in the proposal that defensins are taken up across the outer membraneby the self-promoted uptake pathway. Human defensins were also shown,albeit only at bactericidal or bacteriostatic concentrations, to perrneabiize theouter membrane of various Gram-negative bacteria to rifampicin and Triton X100 (Viljanen et al., 1988). A detailed examination of HNP-i-mediatedbactericidal activity against E. coU revealed that the defensin caused sequentialpermeabiization of the outer and inner membranes, and that it was likely thatthe latter event that resulted in cell death (Lehrer et al., 1989). This study alsoshowed that HNP- 1 was a relatively poor permeabiizer of the outer membranesince it required up to 50 jig/mL for 20 mm before outer membranepermeabiization was detected.Another group of defensins was isolated from a number of differentinsects such as Phormia terranovae (Lambert et al., 1989), Sarcophagaperegrina (Matsuyama and Natori, 1988), and others (for a review, see17Hoffmann and Hetru, 1990)(Table I). These peptides were initially called insectdefensins since they were small, cationic and contained six cysteine residues.Upon further investigation, however, the positions of these cysteine residues inthe amino acid sequence, and even the disulfide array were different from thoseof mammalian defensins (Seisted and Harwig, 1989; Lepage et al., 1991). Theinsect defensins are also slightly longer (38-43 residues) and less cationic (p18.0-8.5)(Hoffmann and Hetru, 1990) than their mammalian counterparts. Thebiggest differences between these two groups of peptides are in their 3-dimensional structures. While the mammalian defensins are dominated by [3-sheet structures (Hifi et al., 1991), the insect defensins have a loop, x-helix, and13-sheet structures (Hoffmann and Hetru, 1990). Therefore, while these twogroups share similarities in name, size, charge and antibacterial activity, theycannot be considered structural homologues.3. Cecropins.Along with defensins, insects produce another family of cationicantimicrobial peptides in response to a bacterial challenge. This group iscollectively called cecropins, after the moth Hyalophora cecropia in which theywere first discovered (Hultmark et al., 1980). Cecropin-like peptides have nowbeen found in many different insects (for reviews, see Boman and Hultmark,1987; Boman et at, 1991; Huitmark, 1993), as well as in mammalian cells (Leeet al., 1989)(Table I). The peptides are 35-39 amino acids in length and havenet positive charges that range from +3 to +8. Unlike the defensins, they do notcontain cysteine residues (Figure 3) and therefore adopt a completely differenttertiary structure (see below). A number of residues are conserved throughoutthe cecropin family, including Trp2, Phe5,Lys6,Glu9,Gly’2,Arg’6,Ala22,Gly23,18Pro24,Ala25, and Ala32. The role of these residues remains unclear; however,they may be involved in inter-protein contacts in oligomeric structures (Durellet at, 1992). The necessity of Trp2 was demonstrated since upon its deletion,the lytic activity of the peptide decreased dramatically (Steiner et at, 1988).In response to an infection, H. cecropia produces three major cecropins,A and B (Huitmark et al., 1980) and D (Huitmark et aL, 1982). The genes forthese peptides have been cloned and sequenced (Xanthopoulos et al., 1988;Gudmundsson et at, 1991) revealing that, like defensins, they were producedas prepropeptides. The amino acid sequence was analyzed for potentialsecondary structures which led Steiner (1982) to propose that the peptideconsisted of two a—helices joined by a hinge region (Figure 5A). This was laterconfirmed by the two-dimensional NMR structure determination of cecropin A(Holak et al., 1988). The strongly basic N-terminal a-helix is almost perfectlyamphipathic, while the more hydrophobic C-terminal a-helix is less so. Thesetwo regions are joined by the flexible hinge region encoded by Ala-Gly-Pro-Ala,which is conserved in most cecropins. The design of cecropin-like modelpeptides and cecropin analogs demonstrated that all three structural featureswere required for full antibacterial activity (Steiner et al., 1988; Fink et al.,1989).Cecropins have a broad spectrum of antibacterial activity against Gram-positive and Gram-negative bacteria as determined by zone inhibition assays(Boman et al., 1991). Cecropins A and B have a much broader range of activitythan cecropin D which may reflect the high positive charge densities of their N-terminal a-helices as compared to cecropin D. None of the cecropins are able tolyse eukaryotic cells (Steiner et al., 1981; Wade et al., 1990) in contrast todefensins, which possess potent cytotoxic activity (Lehrer et at, 1993). Like thedefensins, however, cecropins were able to form voltage dependent ion channels19ABKWKLFKKIEKVGQNIRDGjIKAGpAyyvGQATQIAK cecA 1-37Figure 5: Model of the structure of cecropin A.A, Cecropin A primary and secondary structures showing its helix-hinge-helix nature. The charged residues are shown to indicate the arnphipathicnature of the N-terminal a-helix. Reproduced by copyright permission fromSteiner, et al. (1988). B, Antiparallel dimer of cecropin A in a membranebilayer. The N-terminal a-helices are embedded into the lipid head-group layerand the C-terminal a-hellces traverse the hydrophobic phase of the membrane.Selected residues are highlighted. Reproduced by copyright permission fromDurell, et al. (1992) © Biophysical Society.N-terminal helices20ABFigure 6: Model of the structure of melittin.A, Monomeric, membrane-bound melittin showing the helix-hinge-helixstructure and the C-terminal random coil motif. Circles denote positivelycharged amino acids. B, Tetrameric melittin associated with a lipid membrane.The shaded spheres represent the non-helical C-terminal residues, and thecylinders represent the x-helices with the hydrophobic sides (white) facing themembrane and the hydrophilic sides (shaded) facing each other, thus forming achannel. Both figures were reproduced by copyright permission from Vogel andJahnig (1986) © Biophysical Society.CQNH2‘U24in planar lipid membranes (Christensen et aL, 1988). By studying a number ofsynthetic analogs, these researchers were able to determine that the flexiblehinge region separating the two amphipathic x-helices was required for pore-forming activity. The observed correlation between the peptides’ bactericidaland pore-forming activities led to the conclusion that the lethal effect ofcecropins on bacterial cells was the formation of large pores in the cytoplasmicmembrane which led to cell lysis (Christensen et al., 1988). Although ionchannel formation has not been demonstrated for the porcine cecropin P1, itwas shown to kill E. coil bacteria by cell lysis (Boman et al., 1993) whichsuggests that all cecropins have a similar mechanism of activity. Lipid bilayerswhich had either a positive surface charge or included cholesterol were highlyresistant to the pore-forming action of cecropins. The latter result wasconsistent with the insensitivity of eukaryotic cells to lysis by cecropins, andthe former suggested that the positively charged cecropins required a negativelycharged membrane to initiate bactericidal activity. Although required, thisbinding was not sufficient for killing, since bacteria that were resistant tocecropin activity, were still found to bind large amounts of the peptide (Steineretal., 1988).The observations on cecropin binding and lysis led to the formulation ofa three step mechanism of cecropin activity on lipid membranes (Christensen eta!., 1988). The first step involved the electrostatic adsorption of cecropinoligomers to the negatively charged surface of the membrane. In the secondstep, the hydrophobic C-terminal os-helix would insert into the membrane whilethe amphipathic N-terminal x-helix would remain at its interfacial position.This step could only occur with peptides that possessed the flexible hingebetween the two helices. Finally, an applied voltage (positive to negative goinginto the membrane) would drive the N-terminal ce-helix into the membrane, with21its hydrophilic residues forming a water-filled channel. This model wasexpanded by Durell et al. (1992) who, using atomic-scale computer models,proposed that cecropins form antiparallel dimers whose C-terminal x-helicescan span the membrane (Figure 5B). These dimers would then aggregate into ahexagonal geometry (based on pore dimensions from Christensen et aT., 1988),resulting in a small ion channel formed by six adjacent C-terminal a-helices.This “type I” channel would further undergo a conformational change in whichthe hydrophilic residues of the N-terminal cx-helices would move from themembrane surface to the pore lining, creating a larger (“type II”) channel.Although these proposals may accurately describe the mechanism of cecropinaction on the cytoplasmic membrane, they cannot be applied to the morecomplex, asymmetric outer membrane of Gram-negative bacteria. It isinteresting to note that despite their high potency against Gram-negativebacteria, the effects of cecropins on outer membranes have never been studieddirectly.4. Melittin.The venom of the honey bee Apis me11fera contains, as its major proteincomponent, a 26 amino acid peptide that is highly cationic (Figure 3;Habermann and Jentsch, 1967) and possesses strong antibacterial and potenthemolyuc activities (for a review, see Dempsey, 1990). Many investigators havestudied the structure/function relationship of melittin, particularly with respectto its interaction with membranes. This has resulted in some conflicting ideason how this peptide antibiotic exerts its mechanism of activity.Melittin adopts different structures depending on its environment. Inwater, or in other low ionic strength environments, it exists as a monomer that22has no detectable secondary structure. In higher ionic strength buffers, orother environments that promote self association, the peptide forms a tetramerconsisting of monomeric subunits that are predominantly in a helicalconformation (Dempsey, 1990). Residues 1-21 of melittin form a hydrophobica-helix that is bent in the middle at a glycine residue (Figure 6A). The C-terminal portion of the peptide is very hydrophilic and is not believed to form ana-helix. This structure differs from the a-helical peptide cecropin A in a fewways. First the polarity of the peptides is reversed, with the N-terminus ofcecropin A and the C-terminus of melittin being hydrophilic, while the C-terminus of cecropin A and N-terminus of melittin are hydrophobic. Secondly,the bend in the peptides is found in a different position. In melittin, the bend isin the middle of the N-terminal hydrophobic a-helix, while in the cecropins, itseparates the hydrophilic and hydrophobic a-helical segments. Thesedifferences have ramifications for hybrid peptides created from melittin andcecropin A (see below).The amphipathicity of the N-terminal a-helix of melitthi suggests that themonomers must aggregate when interacting with lipid membranes. Figure 6Bshows a schematic model of tetrameric melittin in a membrane. Thehydrophilic faces of the four a-helices face the inside of the tetramer andprobably form a pore (see below). Although the antibacterial properties ofmelittin have been documented (Boman et al., 1989a), it is the hemolyticactivity that continues to be the focus of intense research. Melittin has beenshown to cause cell lysis at concentrations of greater than 1 ig/mL (DeGrado etal., 1982; Hider et al., 1983). It is also able to form voltage-dependent ionpermeable channels in lipid bilayers (Tosteson and Tosteson, 1981). Theconductance of these channels was shown to change with the fourth power ofmelittin concentration indicating that tetrameric association of the monomers23forms the channel (Tosteson and Tosteson, 1981). This conclusion seemed tobe inconsistent with the observation that melittin could form multi-state pores,indicating that there was heterogeneity in the pore structures (Hanke et aL,1983). This is in contrast to the cecropins which showed discreet channel sizes(Christensen et al., 1988).The structure of melittin has been examined to determine what featuresare responsible for its channel-forming and hemolytic activities. Blondelle andHoughten (1991) created a series of 24 individual omission analogues ofmelittin to determine which residues were necessary for hemolytic activity.Their results showed that the deletion of any residue in the hydrophobic a-helixresulted in a marked decrease in the hemolytic activity, while the omission ofany C-terminal residues had no effect. This was in contrast to Werkmeister etal. (1993) who found that either the incubation of melittin with a C-terminalspecific monoclonal antibody or the deletion of one of the two Lys-Arg motifs inthe C-terminus significantly reduced the hemolytic activity of the peptide. Thissuggested that the C-terminus was necessary for hemolysis, and that thedeletion of one residue from that sequence (Blondelle and Houghten, 1991) wasnot sufficient to reduce that activity. This was supported by the fact that amelittin analogue that was missing the last four amino acids (and therefore oneLys-Arg motif) had a 30% weaker affinity for phospholipid membranes and amuch lower lytic activity compared to native melittin (Otoda et al., 1992).Interestingly, this shortened melittin analogue could still form voltage-dependent ion channels in lipid bilayers (Otoda et al., 1992). This indicatedthat membrane binding and cell lysis were not directly related to ion channelformation. Therefore, one can only conclude that the actions of meittin on lipidmembranes are varied and complex. At low concentrations, melittin will formheterogeneously sized ion channels, probably consisting of melittin aggregates25that do not significantly perturb the membrane, while at higher concentrations,it will cause cell lysis.Other activities of melittin have been studied and recently reviewed(Dempsey, 1990; Fletcher and Jiang, 1993). Again, since the focus of melittinresearch has been its interaction with phospholipid membranes, virtually noresearch on the peptide’s interaction with the outer membrane has been done.One study, however, showed that melittin bound strongly to purified Salmonellatyphimurium LPS (David et at, 1992), suggesting that melittin may be taken upby the self-promoted uptake pathway since the first step in this pathway is thebinding of the compound to LPS.5. Synthetic and Hybrid Peptides.The discovery of cationic peptides and their potent antimicrobialactivities has led to attempts to improve this activity. Most of this research hasfocused on the a-helical peptides since their structural features are relativelysimple (compared to those of defensins) and their basic functional domainshave been identified (see above).Early studies in the field of synthetic and hybrid peptides focused on thecreation of peptides that were predicted to have amphipathic a-helix structure.Lee et al. (1989) synthesized a number of peptides that contained varyingnumbers of repeated tn-, tetra-, or pentapeptide fragments that would give riseto an a-helical structure. They tested these peptides for activity against sixorganisms and found a direct correlation between the peptide’s ability to form astable, amphipathic a-helix and its antibacterial activity against Gram-positivebacteria. (It is interesting to note that these synthetic peptides were inactiveagainst Gram-negative bacteria, suggesting that cationicity and cc-helicity are26not sufficient to cross the outer membrane). The peptide that possessed thehighest bactericidal activity (Ac-(Leu-Ala-Arg-Leu)3-NHCH)also had thehighest a-helix content in a lipid environment (Lee et al., 1989). This peptidewas also the best facilitator at forming ion channels in lipid membranes atbactericidal concentrations (Anzai et al., 1991). The conductance of thesechannels ranged from 2-750 pS which indicated a heterogeneity in theconformation or assembly of the channels, as was demonstrated for melittin(Hanke et al., 1983). The requirement for amphipathic and a-helical structurewas supported by Blondelle and Houghten (1992) who showed that both thesefeatures were necessary for bacterial cell lysis.In addition to creating new peptides, analogues of some naturallyoccurring peptides have been synthesized in an attempt to augment theiractivity. One example is the magainins, a small group of peptides from the skinof the African frog Xenopus laevts (Table I), which are 23 amino acids in lengthand possess broad spectrum antimicrobial activity (Zasloff, 1987). Syntheticmagainin analogues with enhanced a-helical structure were shown to have a50-fold increase in their antimicrobial activity (Chen et al., 1988). The additionof groups of ten or more arginine or lysine residues to the N- or C-terminal endsof magainin-2 resulted in MIC reductions of 5- to 10-fold (Bessalle et al., 1992).The peptide with positive chain extension at the C-terminal end had no increasein hemolytic activity. This was in contrast to melittin whose positively chargedC-terminus was shown to be necessary for hemolytic activity (Otoda et al.,1992; Werkmeister et al., 1993).As described above, two other naturally occurring peptides, cecropin Aand melittin, have been altered to enhance their activity. In addition to changesin the parent peptides, much work has gone into hybrid peptides that containsequences from both cecropin A and melittin. Boman et al. (1989a) described27the first hybrid peptides which were the same size as melittin, and containeddifferent combinations of the hydrophobic and hydrophiic regions of cecropin Aand melittin. In this process, they synthesized a peptide consisting of the first13 amino acids of cecropin A followed by the first 13 amino acids of melittin(CA(1-13)M(1-13)). This peptide was 100-fold more active against S. aureusthan cecropin A and had antimalarial activity that was 10-fold better thancecropin B (Boman et al., 1989a). Unlike melittin, however, this cecropin-likepeptide possessed no hemolytic activity. The two-dimensional structure ofCA(1-13)M(1-13) revealed that it still consisted of a helix-hinge-helix structure,but that the hydrophobic C-terminal a-helix came from the N-terminalsequence of melittin (Sipos et al., 1991). The conclusion from this work wasthat hemolytic activity was dependent on a specific design and location of thebasic “head” region, while antibacterial activity was not sequence specific(Boman et al., 1989a).Wade et al. (1990) investigated whether or not these peptides interactedwith chiral membrane components to initiate antibacterial activity. Theysynthesized D-enantiomers of melittin, cecropin A, magainin-2 (all normallyfound in the L-form) and a few cecropin-melittin hybrids, and found that thesepeptides showed the same antibacterial and ion channel forming activities astheir corresponding L-forms. They concluded that these peptides do not requirespecific interaction with chiral receptors or enzymes for antibacterial activitybut rather an appropriate hydrophobic environment (Wade et al., 1990).During these studies, a new hybrid peptide was constructed [CA(1-8)M(1-18),hereafter referred to as CEME; Figure 3], which had even better antimicrobialactivity and lower hemolytic activity than CA(1-13)M(1-13). Although the twodimensional structure for CEME has not been elucidated, its amino acidsequence indicates that it may have two hinge regions, one from the G-I-G-A28domain of melittin (which is the domain responsible for the hinge in CA( 1-13)M(1- 13)), and another from the G-L-P domain of the melittin sequence whichprovides the only bend in native melittin (Figure 6A). The presence of twoflexible hinge regions would predict an overall reduction of the a-helical contentof CEME, which was demonstrated by circular dichroism analysis (Wade et al.,1990). Confirmation of this double-hinge structure and its ramifications forantibacterial activity requires further investigation.Using CEME as a parent peptide, a number of truncated peptides weresynthesized to determine how short the peptide could be made withoutcompromising its bactericidal activity (Andreu et al., 1992). C-terminaldeletions brought the peptide size down to 18 amino acids with no apparentloss of activity as compared to CEME. A series of 15-mers was created bycombining the first 7 amino acids of cecropin A with different 8 amino acidsegments from the N-terminal sequence of melittin. Although most of thesepeptides showed good activity, they all had slightly higher lethal concentrationsthan CEME, and in some cases were only weakly active against specific bacteria(Andreu et aL, 1992). The most active of these 15-mers were peptides that hada disrupted G-I-G-A domain and therefore no bend in the a-helix. Thissuggested that shorter peptides required a full length a-helix to disruptmembranes, but the mechanism of action of these shorter hybrid peptides wasnot explored further.A recent, ambitious study used 30 chimeric peptides to define further thestructural requirements for antibacterial activity (Wade et al., 1992). Theresults confirmed that the general structural requirements were anamphipathic N-terminal a-helix, a flexible hinge region, and a hydrophobic Cterminal a-helix, but no new requirements were discovered. None of the29peptides tested in this study improved on the potent, broad host range activityof CEME.C. Fusion Protein Technology.1. Introduction.Recent advances in the area of gene cloning and expression have placednew demands on the field of protein purification. It is now commonplace totake a cloned gene, whether it is prokaryotic or eukaryotic, and express it in aprokaryotic host such as E. coiL This process has resulted in the purification ofmany different proteins in quantities that would have been unattainable ifpurifying them from original sources. However, while providing larger amountsof starting material, this procedure did not simplify the tedious process ofpurifying the desired protein from a complex array of cellular components.The development of affinity tag fusion protein technology (reviewed inSassenfeld, 1990; Uhlén and Moks, 1990) was a way to simplify the purificationprocedure. In one example of this technology, the gene encoding the protein ofinterest is fused to a gene encoding a protein that has a strong binding affinityto a specific ligand. The heterologous protein produced from such a gene fusioncan be purified by affinity chromatography on a matrix to which the ligand isbound. In other examples, the affinity tag is a short, poly-amino acid “tail” thatis added on to the protein, increasing its charge density or hydrophobicity andenabling its purification by ion exchange or hydrophobic chromatography.Examples of these two types of affinity tags are found in Table II.This technology provides several advantages for producing foreignproteins in a host such as E. coiL Many heterologous proteins produced bydirect expression of a cloned gene are proteolytically degraded or insoluble.30TableII:CommonlyUsedAffinityTags.AffinityTag3-ga1actosidaseCellulose-bindingdomainF1agTMGlutathione-S-transferaseMaltosebindingproteinPolyargininePolyhistidineProteinASyntheticIgG-bindingdomainBindingMatrixTPEG-SepharoseCelluloseAnti-FlagTMantibodyGlutathioneagaroseCrosslinkedamyloseCationicexchangeNitrolotriaceticacidSepharose(Ni2)IgG-SepharoseIgG-SepharoseElutionConditions0.1Mboratewater0.1MglycinepH3.0or2-5mMEDTA5mMreducedglutathione10mMmaltoseNaC1gradientlowpHor250mMimidazole0.5MaceticacidpH3.40.5MaceticacidpH3.4ReferenceU]lman,1984Ong,etal.,1989Hopp,etal.,1988SmithandJohnson,1988Mama,etal.,1988Sassenfeld,1984Hochuli,etal.,1988Uhlén,etal.,1983Moks,etal.,1987ac)There aie several examples of proteins that, when produced as fusion proteins,had decreased susceptibilities to proteolysis and increased solubiity, comparedto when they were produced directly (Marston, 1986). Another consideration offusion protein production is the localization of the gene product. Use of asecretory system can have many advantages, such as the proper formation ofdisulfide bonds in the oxidative environment outside the cell, increased stabifityby the avoidance of intracellular proteases, increased solubility, theenhancement of purification procedures and, in some cases, the avoidance ofhost toxicity from the heterologous protein. Unfortunately, not all proteins arecompatible with the secretory requirement of membrane translocation.The major obstacle of fusion protein technology is how to obtain thetarget protein free of the fusion partner with no extra amino acids on the N- orC-terminal ends. One advantage of the technology is the ability to cleavespecifically the protein in vitro, as opposed to relying on the in vivo removal ofthe formylmethionine residue or the signal sequence which can result inheterogeneous N-terminal ends. In order to utilize this advantage, site-specificcleavage sites must be engineered in the fusion protein to allow the removal ofall extraneous amino acids from the target protein. There are many differentcleavage methods which are divided into chemical and enzymatic methods(Table III). Chemical methods tend to be quite efficient but are not feasible withlarger target proteins which usually contain the low specificity recognitionsequences internally. The cleavage conditions for these methods are also quiteharsh and could be detrimental to the protein. In contrast, most enzymaticmethods, especially the enclopeptidases, are very specific but can be expensiveto scale up and are sometimes quite inefficient due to the inaccessibility of therecognition site to the enzyme. Therefore, the method best used to cleave the32TableIII:MethodsfortheSit&SpecfficCleavageofFusionProteins.MethodRecognitionSequence/CleavageSiteReferenceChemicalAcidxD4p.x.Szoka,etal.,1986CNBr-X-M-J-X-Itakura,etal.,1977BNPS-skatole-x-w-.I.-x-Knott,etal.,1988Hydroxylamine-x-N-L-G-x-Moks,etal.,1987bEnzymaticCarboxypeptidaseA-x4-x1x1RorKHochuli,etal.,1988CarboxypeptidaseB-X--Ror-X-L-KSassenfeld,1984ChymotrypsinDahiman,etal.,1989Collegenase-x-p-x-.L-G-p-x-GerminoandBastia,1984Enterokinase-X-D-D-D-K-.L-X-Maroux,etal.,1971FactorXa-x-I-E-G-R-i--x-NagaiandThøgersen,1987Thrombin-X-L-v-P-R-L-G-S-X-SmithandJohnson,1988Trypsin-x-R-L-x-,or-X-K-L-X-Shine,etaL,1980CA)CA)fusion protein will vary on an individual basis and depend on a number offactors including protein size and lability.2. Glutathione-S-transferase.One commonly used affinity tag is glutathione-S-transferase (GST),which was originally identified in Schistosomajaponicum (Smith et al., 1986)and can be expressed as an active, soluble protein in E. coil (Smith et al., 1988).This protein was the basis for a number of fusion protein vectors which wereinitially used to purify over 30 eukaryotic polypeptides (Smith and Johnson,1988). Since then, this system has been used in the purification of manydifferent proteins such as the murine leukemia inhibitory factor (Gearing et al.,1989), the Plasmodium faiciparum antigen Pf155/RESA (Stahl et al., 1990),epidermal growth factor receptor sequences (Koland et al., 1990), ratinterleukin-6 (Frorath et al., 1992), Poa plX grass pollen allergens (Olsen andMohapatra, 1992), the transactivation domain of Vmw65 from herpes simplexvirus type 1 (Donaldson and Capone, 1992), mouse mammary tumor virusprotease (Menéndez-Ariau et al., 1992), and human cellular retinoic acidbinding protein II (Redfern and Wilson, 1993). The affinity of GST for reducedglutathione allows the purification of soluble GST fusion proteins by adsorptionto glutathione agarose beads and subsequent desorption using free reducedglutathione (Smith and Johnson, 1988). The GST moiety of the purified fusionproteins can be released by specific proteolytic cleavage using thrombin orfactor Xa (Smith and Johnson, 1988). In several instances, the GST fusionproteins were found to be partially degraded (Koland et al., 1990; Olsen andMohapatra, 1992), although there does not appear to be a correlation betweenthe amino acid sequence of the target protein and the lability of the fusion34protein. Another potential problem with this system occurs when the fusionproteins form insoluble inclusion bodies (Smith and Johnson, 1988; discussedin detail below) which require harsh purification techniques that areincompatible with that of affinity chromatography. This problem was recentlyaddressed by Hartman et al. (1992) who described conditions which enabled theconversion of insoluble fusion protein aggregates to a soluble form amenable toglutathione agarose affinity purification.3. Protein A.Another affinity tag that has been used extensively is the IgG-bindingdomain of S. aureus protein A (Nilsson and Abrahmsén, 1990). The gene wasoriginally isolated by LOfdahl et at (1983) and its complete sequence confirmedearlier suggestions that it consisted of two functional domains (Uhlén et al.,1984). The N-terminal region has a 58 amino acid unit which is repeated 5times and is believed to be responsible for IgG binding. The C-terminal regionhas an 8 amino acid unit which is repeated 12 times and is thought to beinvolved in the binding of the protein to the cell wall. The former region wasused as the basis for a number of fusion protein vectors that had a truncatedprotein A gene (Uhlén et al., 1983). These vectors were first used to createprotein A/[3-galactosidase fusion proteins that could be purified by adsorptionto IgG-Sepharose and desorption using either a glycine buffer pH 3.0 (Lofdahl etal., 1983) or excess pure protein A (Uhlén et al., 1983) depending on the pHstabffity of the target protein. The original vectors were improved to give onesthat were temperature inducible (pRIT2) or allowed for efficient secretion of thefusion protein to the E. coil periplasm or the S. aureus extracellular medium(pRITS)(Nilsson et al., 1985a). These vectors have since been used to produce35fusion proteins to alkaline phosphatase (Nilsson et al., 1985a), human insulin-like growth factor I (Moks et al., 1987a), human apoA-1 protein (Monaco et at,1987), a transactivation domain of the herpes simplex virion protein VP16(Stringer et at, 1990), yeast calmodulin (Stirling et al., 1992), and a tyrosinekinase domain of the c-src protein (Saya et al., 1993). A new series of fusionprotein vectors was constructed using different multiplicities of a synthetic DNAfragment that encoded a small IgG-binding domain (designated “Z”). This Zregion lacked asparagine-glycine dipeptide sequences as well as methionineresidues which rendered it resistant to hydroxylamine and CNBr treatmentsrespectively. This allowed the specific release of a target protein that alsolacked these residues, by engineering one of these chemical cleavage sites at thejunction of the fusion protein. This was already demonstrated with theproduction of IGF-I (Moks et al., 1987a). The ZZ-IGF-I protein was purified byIgG-Sepharose affinity chromatography, cleaved with hydroxylamine to releaseIGF-I, and passed over the affinity matrix a second time to obtain pure IGF-I inthe flow through. This process was scaled up to a 1000 L which led to theproduction of large quantities of IGF-I (Moks et al., 198Th).4. Inclusion Body Formation and Protein Degradation.The use of affinity tags for the production and purification of foreignproteins in E. coil is not without problems. Many foreign proteins producedeither directly or as fusion proteins in E. coli are found to form insoluble,electron dense aggregates (reviewed in Marston, 1986; Kane and Hartley, 1988;Fischer et al., 1993). It has been, and continues to be, a mystery as to whysome heterologous proteins form inclusion bodies while others do not. Earlyhypotheses suggested that solubifity limitations, the size of the protein, the type36of promoter, the extent of expression and the formation of incorrect intra- andinter-chain disuffide formation played key roles in inclusion body formation.However, upon comparing different insoluble and soluble fusion proteins withrespect to these parameters, no correlation could be determined (Kane andHartley, 1988; Schein, 1989). It should be noted that with a complex processsuch as inclusion body formation, the lack of any such correlations does notnecessarily mean that these factors are not important for any specific protein.For example although not all cysteine-containing proteins form inclusionbodies, improper disulfide bond formation may contribute to the stabifity of theinclusion bodies in ones that do (Mitraki and King, 1989).One parameter that is consistently observed to affect inclusion bodyformation is growth temperature, with lower temperatures (30°C or lower)favoring soluble protein production (Schein and Noteborn, 1988). Mitraki andKing (1989) suggested that inclusion bodies were formed by the aggregation ofspecific, partially folded intermediates and that this was dependent ontemperature and pH. This would explain why some heterologous proteins areinsoluble and others are not since each has its own specific folding pathways.This hypothesis was supported by a study in which the soluble production of aprotein A/3-ga1actosidase fusion protein was dependent on specific temperatureand pH conditions (Strandberg and Enfors, 1991). In addition, they showedthat changing the amino acid sequence around the junction between the twoproteins abolished inclusion body formation, perhaps as a result of alterationsin the abffity of the fusion proteiii.to fold properly.The formation of inclusion bodies results in advantages anddisadvantages with respect to fusion protein purification. Substantialpurification of the protein by the removal of cytoplasmic contaminants can beobtained by differential solubilization of the inclusion bodies and has resulted37in protein that was used for crystallography (Nagai et al., 1988). As well, someproteins have been shown to be protected from proteolysis when produced asinclusion bodies (Cheng et al., 1981). One major disadvantage of inclusionbodies is that one must denature them with high concentrations of a chaotropicagent, such as urea or guanidine-HC1, to solubilLze them (Marston, 1986). Inthe case of disulfide-containing proteins, a sulfhydryl reducing agent is alsorequired (for a review, see Fischer et al., 1993). This denaturation necessitatesthe refolding of the heterologous protein to recover its biological activity. Thisprocess comes with its own set of problems and complications which have to besolved for each individual protein, although recent advances in protein refoldinghave improved the efficacy of this procedure (reviewed in Schein, 1990).Another problem associated with fusion protein production is that ofproteolytic degradation. There are believed to be 8 proteases in E. coil, 5 ofwhich are found in the cytoplasm (Swamy and Goldberg, 1981), where mostheterologous proteins accumulate. There are a few different ways to preventthis potential problem. One is to use a cocktail of protease inhibitors such asPMSF, EDTA, pepstatin A and leupeptin which inhibit serine, metallo, acid andthiol proteases respectively (Deutscher, 1990). Another alternative is to use E.coil hosts that are defective in one or more protease genes such as ion and htpR(Goff et al., 1984) or ompT (Elish et al., 1988). One conclusion that can bedrawn from all this work is that each heterologous protein is different, thusmaking it difficult to predict their solubility and stability characteristics in aspecific recombinant production system.5. Expression of Cationic Peptides.38The study of cationic peptides such as cecropins and defensins has beensomewhat limited by the paucity of the material and the complexity of theirpurification schemes. Although chemical synthesis has been used to makesmall peptides like NP-i (Rao et al., i992) and cecropin A (Boman et al.,i989b), for larger peptides and those with complex disuffide arrays this is noteconomically feasible. Therefore, there have been several attempts to establishexpression systems in which to produce these peptides.The first successful attempts at producing cecropin A from H. cecropiawere accomplished in a baculovirus system. Hellers et al. (i991) cloned a eDNAfragment encoding the preprocecropin A gene into a baculovirus which wasused to infect insect hosts. The cecropin A peptide was properly processed to amature, active form and secreted into the hemolvmph of the insects, althoughonly 70% of the peptide was properly amidated at the C-terminus (Hellers et al.,i99 1). This recombinant construct, however, was not able to produce anypeptide in cultured insect cells. When the cecropin A gene was fused to asynthetic IgG-binding domain of S. aureus protein A (ZZ; Nilsson et al., i987)and placed in a baculovirus system, pure fusion protein was obtained from bothcultured insect cells and larvae hosts (Andersons et al., i99i). Active cecropinA was obtained by CNBr cleavage of the fusion protein. Although thebaculovirus system could be used to produce biologically active cecropin A, it isnot a system that lends itself well to scale up.Saccharomyces cerevisiae has also been used as a host for theproduction of cationic peptides. The insect defensin A gene from P. terranovaewas fused to the gene encoding a leader sequence from the yeast pheromonemating factor a (MFcci) and expressed in yeast (Reichhart et al., i992). Theprotein expressed from this gene was processed and secreted, resulting in activedefensin A in the culture supernatant.39A scorpion insectotoxin, 15A, was expressed in yeast, F. coil, and tobaccoplants (Pang et al., 1992). The peptide, which is 35 amino acids in length andcontains 4 disulfide bonds, was secreted from both bacterial and yeast cells,and produced in the cytoplasm of the tobacco plant cells. Although the peptidecould be purified from each host, no biological activity was detected. Thepeptides produced in yeast and bacteria were both found to contain extra aminoacid residues at their N-termini, which may have had a detrimental effect ontheir activity. This work confirmed that producing cationic peptides by directexpression in bacteria was not very feasible. In fact, producing these peptidesby fusion protein technology has been equally unsuccessful. Fusions ofmammalian defensins to f3-galactosidase (T. Ganz, personal communication)and cecropin A to the ZZ domain (Andersons et al., 1991) were found to behighly susceptible to proteolysis when produced in bacterial expressionsystems. There has been only one known report of successful production of acationic peptide in bacteria, that being the disulfide-containing peptidecharybdotoxin (Park et al., 1991). This peptide was produced in F. coil as afusion protein to the gene 9 protein of the T7 phage. The peptide wassuccessfully released from the gene 9 protein by factor Xa and refolded to give apeptide that was shown to have biological activity. It is unknown if this systemhas been attempted for the production of other cationic peptides.D. Aims of This Study.The goals of this thesis were (1) to establish a bacterial expressionsystem for cationic peptides in order to purify large quantities for further study,(2) to study the outer membrane permeabiizing activity and the bactericidalactivity of selected peptides, and (3) to search for evidence that would support40the hypothesis that these peptides cross the outer membrane via the selfpromoted uptake pathway. HNP- 1 and CEME were the peptides chosen for thisstudy due to their documented, strong activity against P. aeruginosa. Inaddition, CEMA, a C-terminal variant of CEME that contains two extra positivecharges, and melittin were also studied.41MATERIALS AND METHODSA. Strains, Plasmids and Growth Conditions.All strains used in this study are listed in Table 1V, and all plasmids usedare found in Table V. Most strains were grown on Luria Broth (1% (w/v)tryptone, 0.5% (w/v) yeast extract) supplemented with 0.5% (w/v) NaC1 (LBNS)and 1.5% (w/v) agar. Plasmids were maintained using antibiotics such asampicillin (75 jig/mL for E. coil), kanamycin (25 jig/mL for E. coil) andchioramphenicol (10 ig/mL for S. aureus). For membrane permeabilizationstudies such as lysozyrne lysis assays and NPN uptake assays the bacteria weregrown in LB without any salt supplement (LB-S). Bacteria used in dansylpolymyxin displacement assays and minimum inhibitory concentration (MIC)assays were also grown in LB-S. All medium components were obtained fromDifco Laboratories, Detroit, Michigan.B. Genetic Manipulations.1. General DNA Techniques.All general DNA techniques such as DNA isolation, agarose gelelectrophoresis, radioactive labeling of oligonucleotides, colony blotting andSouthern and Northern blotting were performed as described in Ausubel et al.(1987) and Sambrook et al. (1989). Other methods used included RNA isolation(von Gabain et al., 1983), and slot lysis gel electrophoresis (Sekar, 1987). DNArestriction and modifying enzymes (Bethesda Research Laboratories (BRL),Burlington, Canada; Boehringer Mannheim, Mannheim, Germany; Pharmacia,42TableIV:Strains.StrainE.coilDH5cxBL21HMS174K38UT5600C474(SC122)C475C476UB1005DC2SC9251SC9252DescriptionsupE44hsdRl7recAlendAlgyrA96thi-1rel.A1F-hsdsgalnonsuppressinghost(rBmBjF-hsdsrecArtj’nonsuppressinghost(rjç mKjHfrCR)ompTproteasemutantlac(am)trp(arn)pho(am),nal(am)supdsrpsLSC122ionSC122htpRl65-TnlOParentforDC2AntibioticsupersusceptiblemutantParentfor5C9252PolymyxinBresistantReference/SourceHanahan,1983StudierandMoffatt,1986StudierandMoffatt,1986RusselandModel,1984Elish,etaL,1988Goff,etaL,1984Goff,etal.,1984Goff,etal.,1984Richmond,etal.,1976Richmond,etal.,1976Meyers,etaL,1974Meyers,etal.,1974cont...0)TableIV:Strains(con’t).StrainS.aureusRN4220SAPOO17P.aeruginosaH103H309K799(orH187)Z61(orHl88)S.typhimuriumC587(14028s)C590(MS7953s)E.cloacae218S218R1DescriptionTransformationrecipientandexpressionhost,methicihinsensitiveMethicfflinresistant,clinicalisolatePAO1prototrophH103containingplasmidRP1Wildtypeisolate;parentforZ61AntibioticsupersusceptiblemutantParentofC590phoP/phoQmutant;defensinsensitiveClinicalisolateandparentof218R113-lactamaseoverproducingstrainReference/SourceKreiswirth,etal.,1983T.Chow,UBCHancockandCarey,1979HancockandWong,1984Angus,etal.,1982Angus,etal.,1982Fields,etal.,1989Fields,etal.,1989Marchou,etal.,1987Marchou,etal.,1987.TableV:Plasmids.PlasmidDescriptionReference/SourcepTZ19Rgeneralcloningvector,ApRPharmaciapKP196pTZ19Rcontaininga196bpSmaJfragmentencodingaribosomeThisstudybindingsite,theE.colialkalinephosphatasesignalsequenceandtheHNP-1genepT7-517RNApolymeraseexpressionvector,Ap1TaborandRichardson,1985pKP190p17-5containinga190bpSstI/HindIIIfragmentfrompKP196Thisstudyp17-717RNApolymeraseexpressionvectorthatcontainstheribosomeTaborandbindingsitefrom17gene10,ApRRichardson,1985pKP160p17-7containinga160bpNdeI/HirtdIIIfragmentfrompKP196whichThisstudylackstheribosomebindingsitepGP1-2AP15Abasedplasmidthatencodesthe17RNApolymeraseunderTaborandcontrolofthePLpromotor.ItalsocontainstherepressorcI-857,KanRRichardson,1985pGEX-3Xglutathione-S-transferasefusionproteinexpressionvector,ApRPharmaciapPCRpTZ19Rcontainingthe500bpBctI/EcoRIfragment thathadbeenThisstudychangedbyPCRpGEX-KPpGEX-3XderivativewithanSphI/HirtdIII/EcoRImultiplecloningsiteThisstudycon’t...01TableV:Plasmids(con’t).PlasniidDescriptionReference/SourcepGEX-HNP-1pGEX-KPwitha97bpSphI/HindIIIfragmentencodingtheHNP-1geneThisstudypGEX-proHNP-1pGEX-HNP-1witha207bpSphIfragment encodingtheHNP-1preproThisstudyregionpGEX-CEMApGEX-KPwitha110bpSphI/HindIIIfragmentencodingtheCEMAgeneThisstudypGEX-CEMEpGEX-CEMAwitha25bpNaeI/EcoRIfragment replacedwithThisstudyoligonucleotidesKandLcorrectingthe2bpdeletionpGEX—pr0CEMEpGEX-CEMEwitha207bpSphIfragmentencodingtheHNP-1preproThisstudyregionpGEX-CEME-SpGEX-CEMEwithadaptorsatthe5’and3’endsthatcreateaCEMEThisstudySailfragmentpGEX-CEMA-SpGEX-CEMAwithadaptorsatthe5’and3’endsthatcreateaCEMAThisstudySailfragmentpGEX-HNP1-BpGEX-HNP1withadaptorsatthe5’and3’endsthatcreateanHNP-1ThisstudyBamHIfragmentpGEX-proHNPl-BpGEX-proHNPlwithadaptorsatthe5’and3’endsthatcreateaThisstudyproHNP-1BamHIfragmentcon’t...0)TableV:Plasmids(con’t).PlasmidDescriptionReference/SourcepRIT5proteinAbasedfusionproteinexpressionvectorforuseinE.coilorS.Pharrnaciaaureus,j\mpR, CmRpPA-CEMEpRIT5witha136bpSailfragmentencodingtheCEMEgeneThisstudypPA-CEMApRIT5witha142bpSailfragmentencodingtheCEMAgeneThisstudypPA-HNP-1pRIT5witha112bpBamHIfragmentencodingtheHNP-1geneThisstudypPA-proHNP-1pRIT5witha295bpBamHIfragmentfrompGEX-proHNP-1encodingThisstudytheproHNP-1geneRP1broadhostrangeplasmidencodingtheTEM-213-lactamase;NeoRKanRNicasandTet’Amp’Hancock,1983bUppsala, Sweden) and DNA ligase (BRL) were used according to themanufacturers method.2. DNA Fragment Isolation.Specific DNA fragments were isolated by the band intercept technique(Winberg and HammarskjOrd, 1980) using DEAE paper and manufacturer’smethod (Schleicher and Schuell Inc., Keene, N.H.). Alternatively, a method ofisolating DNA fragments directly from agarose plugs was used. Briefly, a pieceof agarose containing the fragment of interest was cut from a gel and placed ina 1.5 mL eppendorf tube containing siiconized glass wool. A small hole waspoked in the bottom of this tube which was then placed in another 1.5 mLeppendorf tube. Upon spinning these tubes at 5000 g for 10 mm in amicrofuge, the liquid and DNA contained in the agarose plug was collected inthe bottom tube. The DNA was extracted once with water-saturated N-butanoland ethanol precipitated.3. DNA Sequencing.Plasmid DNA for sequencing was isolated using Qiagen columns (QiagenInc., Chatsworth, California) following the manufacturer’s protocol. Sequencingreactions were set up according to the manufacturer’s method, containing 1 igof template DNA, 3.2 pmol of primer and components from an AppliedBiosystems Inc. (ABI; Foster City, California) Taq DyeDeoxy Terminator CycleSequencing Kit. Sequencing reactions were carried out using an Ericompthermocycler (98°C for 1 see, 50°C for 15 see, 60°C for 4 mm; 25 cycles), run on48an ABI 370A automated DNA sequencer, and analyzed using ABI 373A DataCollection and Analysis programs for the Macintosh computer.4. Polymerase Chain Reaction.Reactions were set up to include 1 ng of template DNA, 20 pmoles ofeach oligonucleotide primer, 0.8 mM dNTPs, and 1 unit of DNA Taq polymerase(BRL) in 10 mM Tris-HC1 pH 8.2, 50 mM KC1, 0.8 mM MgC12 and 0.01% gelatin.Reactions were covered with mineral oil and placed in an Ericomp thermocyclerfor 35 cycles of 94°C for 1 mm, 50°C for 1 mm and 72°C for 1 mm, with a finalstep of 72°C for 10 miii to complete all chain extensions. The entire contents ofthe reaction tube were placed on a piece of paraflim and rolled around toremove the mineral oil. Samples were then diluted and analyzed by agarose gelelectrophoresis.5. Transformation and Electroporation.Although some transformations of E. colt cells with plasmid DNA weredone using 0.1 M CaCl2 (Sambrook et al., 1989), most were done using themethod of Chung et al. (1989). Briefly, cells to be transformed were grown to an0D600 of 0.3-0.4 in LBNS. For each transformation, 1 mL of these cells werepelleted in a 1.5 mL microcentrifuge tube for 1 mm and resuspended in 100 jiLof TSS solution (LBNS broth containing 10% (w/v) PEG 8000, 5% (v/v) dimethylsulfoxide, 25 mM MgCl2,pH 6.5). DNA ( 1 ng) was added to the cells and themixture was placed on ice. After 30 mm, 0.9 mL of LBNS was added to the cellswhich were then incubated at 37°C for 1 hr to allow the expression of the49plasmid’s antibiotic resistance gene. Aliquots of the cells were plated onselective medium and grown overnight at 37°C.The method for electroporation of S. aureus cells was similar to that ofCompagnone-Post et al. (1991). Cells were grown in 200 mL of LBNS to an0D600 of 0.4-0.6 before being chilled and harvested. Cells were washed twicewith 80 mLs of cold electroporation buffer (7 mlvi HEPES pH 7.2, 272 mMsucrose) and finally resuspended in 2.4 mL of the same buffer. Between 10 and100 ng of plasmid DNA was added to 160 iL of cells and stored on ice for a fewminutes. Electroporation was carried out in 0.1 cm electroporation cuvettesusing a Bio-Rad (Mississauga, Canada) Gene Pulser set to 200 ohms, 25 jiF and1.8 kV. Immediately following electroporation, 1 mL of SMMP medium (2.8%(w/v) Bacto Pennassay Broth with 275 mM sucrose, 11 mM maleic acid, 11 mMMgC12 and 0.25% (w/v) BSA) was added. The cells were incubated for 90 mmat 37°C before plating for transformants on LBNS plates containing 10 jig/mL ofchioramphenicol.6. Oligonucleotide Purification.Oligonucleotides were obtained from T. Atkinson (Department ofBiochemistry, University of British Columbia). Oligonucleotides longer than 50bases were gel purified according to the method of Atkinson and Smith (1984)using 8% polyacrylamide gels. The oligonucleotides were visualized by UVshadowing and cut out of the gel with a sterile scalpel. To elute theoligonucleotides, the polyacrylarnide gel slices were placed in 1.5 mL Eppendorftubes containing 1 mL of 0.5 M ammonium acetate and incubated overnight at37°C. The eluted oligonucleotides were further purified on C18 SEP-PAKcartridges (Waters, Milford, Massachusetts) as described by Atkinson and Smith50(1984). The 0.5 M ammonium acetate solution containing the oligonucleotideswas loaded onto a prepared C18 SEP-PAK column, washed with water andeluted with 40% acetonitrile. The oligonucleotides were either lyophilized orethanol precipitated before quantification by A260 absorbance. In the case ofoligonucleotides of < 50 bases, these were dissolved in 0.5 M ammoniumacetate and purified directly on C18 SEP-PAK columns using 20% acetonitrile asthe eluting solution.C. Vector Construction.1. Direct Expression Vectors.Six overlapping, complementary oligonucleotides (Table VI, A-F) encodingthe gene for HNP- 1 (with E. coil biased nucleotides at the wobble position ofeach codon), preceded by the alkaline phosphatase signal sequence and aribosome binding site, were annealed together (heated for 3 mm at 90°C andcooled slowly to room temperature) and ligated into pTZ19R that was previouslycleaved with SmaI to form pKP196 (Figure 7). From this plasmid, a 190 bpSstI/HindIII fragment containing the HNP- 1 gene was cloned into pT7-5 tocreate pKP 190 (Figure 8). This plasmid was transformed into K38 (pGP1-2)which possessed a plasmid-encoded 17 RNA polymerase gene under the controlof the temperature inducible ? Pj promoter and would allow HNP- 1 expressionfrom the 17 RNA polymerase promoter. A 160 bp NdeI/HirtdIII fragment frompKP196 that no longer contained the ribosome binding site was inserted intop17-7 to make pKP 160 (Figure 8). This was also transformed into K38 (pGP1-2)for HNP- 1 expression.51TableVI:Oligonucleotides.NameSequence5’-3’DescriptionGGGAGCTCCTAACTAACTAAGGAGGAGACATATGAAACAAAGCACTATrGCACTGGCACTCTI’ACCGTTACTGmACCCCCCAGTGCAATAGTGCTITGYITCATATGTCTCCTCCTTAGYI’AGTPAGGAGCTCCTGTGACAAAAGCCGCATGCTACTGCCGTATACCGGCCTGCATCGCGGGCGAACGTCGTTACGGTACAGGCCGGTATACGGCAGTAGCATGCGGCITITGTCACAGGGGTAAACAGTAACGGTAAGAGTGCCTGCATCTACCAGGGCCGTCTGTGGGCGTrCTGCTGCTAAAAGCYCGCGCGAAGC]TITAGCAGCAGAACGCCCACAGACGGCCCTGGTAGATGCAGGTACCGTAACGACGTTCGCCCGCGATGCCGATGGCCATCATACG’TTATATAGCTGAC81merusedinconstructionofHNP-1gene56merusedinconstructionofHNP-1gene65merusedinconstructionofHNP-1gene64merusedinconstructionofHNP-1gene50merusedinconstructionofHNP-1gene76merusedinconstructionofHNP-1gene30merusedinconvertingpGEX-3XtopGEX-KPGCGGGAKITCAAGCITGCATGCACGACCTI’CGATCAGA42merusedinconvertingpGEX-3XtoTCCGpGEX-KPCGGGGATCCGCATATGAAATGGAAACTGITCAAGAAGATCGGCATCGGCGCCGTGCTGAAAGTGCTGACCACCGGTCTGCCGGCGCTGAAGCTAACTAAGTA102merencodingCEMAcon’t...A B C D E F G H I01 F’)TableVI:Oligonucleotides(con’t).NameSequence5’-3’DescriptionAGCYACTAGTPAGCTCAGCGCCGGCAGACCGGTGGTCAGCACTTTCAGCACGGCGCCGATGCCGATCTCYGAACAGTITCCATITCATATGCGGATCCCCGCATGGGCGCTGAAGCTAACTAAGTAAGCTJ’GAATPCAAGCTTACYPAG’rFAGCTrCAGCGCCMCCATATGAGGACCCTCGCCATCC’ITGCTGCCAYrCTCCTGGTGGCCCTGCAGGCCCAGGCTGAGCCACTCCAGGCAAGAGCTGATGAGGYGCAGCAGCCCCGGAGCAGANTTGCAGCTGACATCCCAGAAGTGGYrGYITCCCTI’GCATGGGACGAAACGTGGCTCCAAAGCATCCAGGTCAAGGAAAAACATGGCATG0CCATGYf1TI’CCTTGAGCCTGGATGC1TFGGAGCCAAGC1TCGTCCCATGCAAGGGAAACAACCACTCTGGGATGTCAGCTGCAATCTGCTCCGGGGCTGCTGCAACCTCATCAGCTCTTGCCTGGAGTGGCTCAGCCTGGGCCTGCAGGGCCAGCAGGAGAATGGCAGCAAGGATGGCGAGGGTCCTCATATGGCATGAGCYGTCGACACGTCGACATCGAAGGTCGTGCATGCACGACC’ITCGATGTCGACGCATG110merencodingCEMA27merusedinconvertingCEMAtoCEME31merusedinconvertingCEMAtoCEME110merusedinconstructionofpreprocartridge91merusedinconstructionofpreprocartridge109merusedinconstructionofpreprocartridge92merusedinconstructionofpreprocartridge12merencodingaHiridIIItoSaLTadaptor24merencodingfactorXarecognitionsiteandanSphItoSalTadaptor24merencodingfactorXarecognitionsiteandanSphItoSaLTadaptorcon’t...J K L P Q R S01 Ci)TableVI:Oligonucleotides(con’t).NameSequence5’-3’DescriptionTCGGATCCATGGCATG15merencodingamethionineresidueandanSphItoBamHladaptorUCCATGGATCCGCATG15merencodingamethionineresidueandanSphItoBamHladaptorVAA’ITCGGATCCG12merencodinganEcoRltoBamH1adaptorWTATGGGATCCCA12merencodinganNdeItoBamHladaptorXCCAAAATCGGATCTGATCGAAGG23merusedassequencingprimerYCAGATCGTCAGTCAGTCACG20merusedassequencingprimer01A CD FELigate into Smal site of pTZ19RFigure 7: Schematic representation of the strategy used to anneal the sixoligonucleotides that encoded the HNP- 1 gene.B55Figure 8: Construction of pKP19O and pKP16O.The HNP-1 gene was inserted into p17-5 and p17-7 to form pKP19O andpKP 160 respectively as described in Materials and Methods, section C. 1. Onlyrelevant restriction sites are shown. The large arrow indicates the position anddirection of the 17 RNA polymerase promotor. AP, alkaline phosphatase; rbs,ribosome binding site.SstISstt NdeI HindIJISphI—T7 RNA Polymerase Promoter562. GST Expression Vectors.The multiple cloning site of the fusion protein expression vector pGEX3X had to be changed before the HNP- 1 gene could be cloned into it. Therefore,a 500 bp BamHI/BalI fragment was isolated from pGEX-3X and subjected toPCR using oligonucleotides G and H, the latter which had a 21 base overhangencoding SphI, Hindlil and EcoRI restriction enzyme sites. The PCR reaction(Materials and Methods, section B.3.) was performed in an Ericompthermocycler. The reaction products were blunt-ended using Kienow fragmentand ligated into pTZ19R cleaved with Smal to produce pPCR. A 200 bpBclI/EcoR[ fragment from pPCR was used to replace the BclI/EcoRI fragmentfrom pGEX-3X to produce pGEX-KP, which now had a multiple cloning siteconsisting of SphI, HindIII and EcoRI instead of BamHI, SmaI and EcoRI. A 90bp SphI/HindIII fragment from pKP196 containing just the HNP-1 gene wascloned into pGEX-KP to create pGEX-HNP- 1, which was sequenced usingoligonucleotides X and Y as primers to ensure it was the correct construct.The plasmid pGEX-CEME was created by a two step process. Twocomplementary oligonucleotides encoding SphI and HirtdIII sticky ends and theCEME gene preceded by a methionine codon were cloned into pGEX-KP. Uponsequencing, however, a 2 bp deletion was discovered, which resulted in a geneencoding a peptide (CEMA) with an additional 2 amino acids and a net chargechange of ÷2 as compared with CEME (oligonucleotides I and J). This plasmidwas termed pGEX-CEMA. In order to fix the frameshift mutation, a NaeI/EcoRIfragment containing the deletion was removed from pGEX-CEMA and replacedwith oligonucleotides K and L to form pGEX-CEME. This correction wasconfirmed by DNA sequencing.57The plasmids pGEX-proHNP- 1 and pGEX-proCEME were created usingoligonucleotides M-P which encode the pre pro region of the HNP- 1 gene (Daheret al., 1988). These oligonucleotides, which have SphI sticky ends whenannealed, were inserted into SphI-cleaved pGEX-HNP-1 and pGEX-CEME. Theinsertion orientation of the prepro fragment was confirmed by restrictionenzyme analysis.3. Protein A Expression Vectors.The limited restriction sites available for cloning into pRIT5 (Nilsson eta!., 1985a) and the requirement for obtaining the correct reading frame,necessitated changing the 5’ and 3’ ends of the CEME, CEMA and HNP- 1 genes.For CEME and CEMA, the 3’ ends were altered by inserting the self-annealingoligonucleotide Q encoding a Hindill to Sail adapter into the Hindill sites ofpGEX-CEME and pGEX-CEMA. These were further altered at the 5’ end bycloning, into the SphI sites, the annealed oligonucleotides R and S whichencode an SphI to Sail adapter including a factor Xa recognition site. Theorientation of this asymmetric insertion was confirmed by DNA sequencing.The Sail fragments from the resulting two vectors (pGEX-CEME-S and pGEXCEMA-S) were cloned into pRIT5 to produce pPA-CEME and pPA-CEMA. TheHNP- 1 gene was modified in a similar manner. An SphI to BamHI adaptercontaining a methionine codon (oligonucleotides T and U) was inserted into theSphI site on the 5’ end of the HNP- 1 gene (the orientation was confirmed byDNA sequencing), while an EcoPJ to BamHI adapter (oligonucleotide V) wasused at the 3’ EcoRI site. The resulting BamHI fragment from pGEX-HNP-1-Bwas cloned into pRIT5 to form pPA-HNP- 1. The plasmid pGEX-proHNP- 1 waschanged to pGEX-proHNP-1-B using oligonucleotide V at the 3’ end and58oligonucleotide W, an NdeI to BamHI adapter, at the 5’ end. The cloning of thisBamHI fragment into pRIT5 resulted in the production of pPA-proHNP- 1. AllpPA plasmids were isolated and electroporated into S. aureus RN4220D. Immunological Techniques.1. Production and Purification of Antibodies.Antibodies were prepared as described by Harlow and Lane (1988). Apreparation of GST/CEMA was solubffized in SDS-PAGE loading buffer, heatedat 100°C for 10 mm, and loaded onto a preparative SDS polyacrylamide gel.GST/CEMA contained in gel slices was passively eluted into water andlyophiized. The sample was resuspended and mixed with an equal volume ofFreund’s complete adjuvant before injecting 1 mg of protein subcutaneouslyinto a rabbit. Booster shots of 1 mg of protein mixed with Freund’s incompleteadjuvant were given at 3 and 7 weeks. After 10 weeks, the rabbits were bledand the serum was isolated.2. Western Blotting.Western immunoblotting (Mutharia and Hancock, 1983) andimmunodetection (Mutharia and Hancock, 1983; Harlow and Lane, 1988) weredone as previously described.59E. Electrophoresis.1. SDS-PAGE.SDS-PAGE was performed as previously described (Hancock and Carey,1979).2. AU-PAGE.AU-PAGE was done according to the method of Panyim and Chalkey(1969). Briefly, 15% polyacrylamide gels containing 5 M urea (w/v) and 5%acetic acid were pre-electrophoresed at 100 V for two hours (for mini gels) or150 V overnight (for large gels) in 5% acetic acid. The buffer was replaced withnew 5% acetic acid before the samples were loaded. Samples were solubiizedin 5% acetic acid and diluted 2:1 with sample buffer (9 M urea in 5% acetic acidcontaining methyl green as a tracking dye). Gels were run in the reversepolarity (towards the cathode) at either 100 V or 250 V until the methyl greenmigrated off the bottom of the gel.F. Protein Expression.1. Direct Expression.Attempts to express HNP-1 from pKP 190 and pKP 160 using the 17 RNApolymerase system were performed as previously described (Ausubel et al.,1987). Briefly, K38 (pGP1-2) cells containing pKP19O or pKP16O were grown inLBNS with kanamycin and ampidillin at 37°C to an 0D600 of 0.5. A sample ofcells was removed (as an uninduced control) before the culture was shifted to42°C for 30 mm to induce the T7 RNA polymerase. Again, a sample of cells was60removed before the culture was shifted back to 37°C and 200 ig/mL ofrifampicin simultaneously added to prevent E. coiL RNA polymerase activity.The cells were harvested and the basic proteins extracted as previouslydescribed (Lindahi and Zengel, 1979). Proteins extracted from all strains wereanalyzed by SDS-PAGE and AU-PAGE.2. GST Fusions.Cells of E. coil strain DH5a containing pGEX-KP, pGEX-CEME, pGEXpr0CEME, pGEX-CEMA, pGEX-HNP-l or pGEX-proHNP-1 were grown at 37°Cto mid log phase and IPTG added to a final concentration of 0.2 mM to inducefusion protein expression. Growth was allowed to continue for 3 h before cellswere harvested. To test for fusion protein production, 500 jiL of cells werepelleted in a microfuge for 1 mm and resuspended in 30-50 iiL of water. Anequal volume of 2X SDS-PAGE whole cell lysing buffer (2X WCLB; 0.125 M TrisHC1 pH 6.8, 20% (v/v) glycerol, 4% (w/v) SDS and 10% (v/v) 2-mercaptoethanol) was added, the sample heated at 100°C for 10 mm andloaded onto an SDS-PAGE gel.3. Protein A Fusions.Initially, expression of PA/CEME fusion protein was attempted in E. coiLDH5o cells harboring pRIT5 or pPA-CEME were grown at 37°C until mid-logphase. Samples of whole cells and culture supernatants were analyzed by SDSPAGE and Western immunoblotting for the presence of fusion protein. Fusionproteins were detected on Western immunoblots using an IgG antiserum whichbound to the carrier portion of the molecule. Plasmids pRIT5, pPA-CEME, pPA61HNP- 1, and pPA-proHNP- 1 were also expressed in S. aureus strain RN4220. Inthis case, S. aureus cells harboring the various plasmids were grown at 37°C toan 0D600 of between 1.5-1.8 before the cells were removed by centrifugation.Culture supernatants (20 tL) were analyzed as above by Westernimmunoblotting. Large scale (20 L or 60 L) expression of pPA-CEME was donein a LH fermentor (100L working volume, 5000 series) and the cells removed bya Sharples centrifuge T-1P. After the addition of 1 mM sodium azide, thesupernatant was clarified using a 0.45 iM membrane in a Pellicon CassetteSystem with tangential flow filtration (Millipore). The supernatant was thenconcentrated about 10-fold by ultrafiltration using a 1 kDa Omega FiltronCentrasett Screen Channel. Large scale growth of RN4220(pPA-CEME) and theclarification and concentration of the culture supernatant were carried out by L.Robillo and G. Lesnicki (Biotechnology Laboratory, University of BritishColumbia)G. Fusion Protein Purification.1. GST Fusion Proteins.Cell pellets were resuspended in 1/100 volume of 50 mM Tris-HC1 pH8.0, 50 mM NaC1 and 1 mM EDTA and passed twice through a French pressurecell at 15,000 psi. The lysate was fractionated by centrifugation at 2500 g for 5mm and the fusion protein recovered either in the supernatant fraction or inthe pellet as insoluble inclusion bodies. In the former situation, thesupernatant was incubated for 1 h at 4°C with glutathione agarose beads(sulfur linkage, Sigma, St. Louis, Missouri) previously swelled and equilibratedin mouse tonicity phosphate buffered saline (MTPBS; 150 mM NaCl, 16 mMNa2HPO4and 4 mM NaH2PO4,pH 7.2). The beads were washed three times62with one volume of MTPBS by resuspension and centrifugation. The fusionprotein was eluted with 5 mM reduced glutathione in 50 mM Tris-HC1 pH 8.0.The cellular production of inclusion bodies prevented the use of affinitychromatography to purify the fusion protein. Therefore, two methods ofdifferential solubilization were employed to purify the inclusion bodies. ForGST/HNP-1, the insoluble pellet was washed sequentially with: 1) 10 mM TrisHC1 pH 8.0, 10 mM EDTA and 1% Triton X-100; 2) H20; 3) 3 M urea (APISTARquality, BDH Chemicals, Vancouver, Canada; previously deionized with AG501-X8 mixed bed resin, Bio-Rad) in 5 mM Tris-HC1 pH 8.0; and 4) H20. Eachwash consisted of resuspending the pellet thoroughly with a pipet andimmediately pelleting it at 20 K for 15 mm in an ultracentrifuge using a Ti50 or60 rotor. Finally, the inclusion bodies were solubiized with 8 M urea in 5 mMTris-HC1 pH 8.0. For GST/pr0CEME, the insoluble pellet was extracted using adifferent procedure. After washing the pellet with 10 mM Tris-HC1 pH 8.0, 10mM EDTA and 1% Triton X-100, it was sequentially extracted twice with 10 mMTris-HC1 pH 8.0, 10 mM EDTA and 3% (v/v) octyl-polyoxyethylene (0-POE;Bachem Bioscience, Philadelphia), and twice with 10 mM Tris-HC1 pH 8.0. Thisprocedure removed all the contaminating membrane proteins and resulted in areasonably pure inclusion body preparation.2. Protein A Fusion Proteins.PA/CEME and PA/CEMA were purified as previously described (Moks etal., 1987a). Briefly, clarified culture supernatants were adjusted to pH 7.6 withNaOH and passed over an IgG-Sepharose column (bed volume of 30 mL,Pharmacia), previously equffibrated with TST buffer (50 mM Tris-HC1 pH 7.6,150 mM NaCl and 0.05% Tween 20). The column was washed sequentially with6310 volumes of TST and 5 volumes of 5 mM ammontum acetate pH 5.0. Sampleloading and column washing was done at a flow rate of 100-150 mL/hr whichwas controlled by a Pharmacia P-i peristaltic pump. The fusion protein waseluted at a flow rate of 50 mL/hr with 0.5 M acetic acid pH 3.4 and lyophilized.H. Peptide Release.1. Factor Xa.Inclusion body preparations of GST/HNP- 1 previously solubilized in 8 Murea and 5 mM Tris-HC1 pH 8.0 were placed in Spectra/Por 3 dialysis tubing(Spectrum Medical Industries, Los Angeles, California) and sequentially dialyzedovernight against 1 M urea in 5 mM Tris-HC1 pH 8.0, and factor Xa cleavagebuffer (50 mM Tris-HC1 pH 8.0, 100 mM NaC1 and 1 mM CaCl2). If necessary,the sample was concentrated using a Centricon 10 (Amicon, Oakvfile, Canada).Factor Xa (Boehringer Mannheim) was added at an enzyme to substrate ratio ofapproximately 1:25, and the reaction was allowed to proceed at 37°C for 60 h.Samples were taken at 12 h intervals and analyzed on a 12-25% gradient SDSPAGE gel.2. Cyanogen Bromide.Whether the sample was an inclusion body preparation of GST/proCEMEor purified, lyophilized PA/CEME or PA/CEMA, the fusion proteins weresolubilized in 70% formic acid to which 1 M CNBr was added. Proteinconcentrations in these reaction were such that the molar excess of CNBr tomethionine residues in the fusion protein was at least 2000. The sample wasflushed with nitrogen, tightly capped, and incubated in the dark at 25°C for 18-6424 h. The reaction was quenched by diluting to 5% formic acid with H20followed by lyophiization.I. Peptide Purification.The CNBr cleaved PA/CEME or PA/CEMA fusion protein wasresuspended in 5% formic acid, and approximately 100 mg loaded onto a BioGel P100 column (bed volume of 300 mL, flow rate of 10 mL/h, Bio-Rad), andeluted with 1% acetic acid. Two mL fractions were collected, lyophiized andanalyzed by AU-PAGE. Fractions containing the peptide were pooled,lyophiized and resuspended in 0.1% TFA. In the case of a large scalepreparation, a 15 mg sample was loaded onto an FPLC Bio-Sil C18 reversephase column (bed volume of 50 mL, flow rate of 2 mL/min, Bio-Rad),previously equilibrated with 0.1% TFA. The sample was eluted with a 60 mLgradient of 50% acetonitrile in 0.1% TFA. Samples containing the peptide ofinterest were pooled, lyophiized and again resuspended in 0.1% TFA. Finalpurification of the peptide was performed on a Pharmacia PepRPC HR 5/5column (bed volume of 1 mL, flow rate of 0.7 mL/rnin). One mg of protein wasloaded onto the column and the peptide was eluted with the following gradientof acetonitrile in 0.1% TF’A: 3 mL, 0-30%; 5 mL, 30%; 20 mL, 30-50%.Fractions (0.5 mL) were lyophilized and analyzed by AU-PAGE to confirmhomogeneity.J. Gel Overlay Assay.Detection of antibacterial activity of proteins separated by AU-PAGE wasdescribed previously (Hultmark et al., 1980). Briefly, purified or partially65purified samples of antibacterial peptides were electrophoresed on a 15% acidurea gel which was then incubated in Mueller-Hinton broth containing 0.2 Msodium phosphate buffer pH 7.4 for 1 h. The gel was overlaid with 5 mL of thesame media containing 0.6% agar and about colony forming units of E. coilDC2, and then again with another 5 mL of agar. The gel was incubatedovernight at 37°C which resulted in zones of lysis that corresponded to themigration site of the antibacterial peptide.K. Peptide Analysis.1. Preparation of Samples.Peptides were subjected to AU-PAGE on a 15% gel and electroblottedonto Immobilon membrane (Millipore, Bedford, Massachusetts) using themanufacturer’s method. Briefly, the transfer membrane was wetted with 100%methanol for a few seconds, rinsed in water, and equilibrated in transfer buffer(0.7% acetic acid, 10% methanol). Electroblotting was carried out in a Bio-RadMini Trans-Blot electrophoretic apparatus in the direction of the cathode for 2 hat 150 V. The membrane was stained with 0.5% (w/v) Ponceau S in 1% aceticacid for 2 mm and destained with water until protein bands were visible(Ausubel et aL, 1987).2. Amino Acid Sequencing and Analysis.The band containing the peptide of interest was excised and analyzed foramino acid content using an Applied Biosystems amino acid analyzer (model470) and the N-terminal sequence determined using an Applied Biosystemssequencer (model 420). The amino acid analysis and peptide sequencing was66carried out by S. Keiland (Department of Biochemistry and Microbiology,University of Victoria, B.C.).L. Assays.1. Protein Concentration Estimation.a. Lowry.Most protein concentrations were estimated with a modified Lowry assay(Sandermann and Strominger, 1972). A 1 mg/mL solution of bovine serumalbumin was used as a standard.b. Dinitrophenylation.Since CEME and CEMA only have one aromatic residue in their primarystructures, a protein assay that measures the presence of free amino groups(Bader and Teuber, 1973) was used instead of the modified Lowry assay. A 1mg/mL solution of poiymyxin B was used as the standard. Briefly, a 50 IlLsample of peptide was mixed with Na2B4O710H2 (final concentration 0.8%)and 1-fluoro 2,4 dinitrobenzene (final concentration 10 mM), and incubated at37°C for 1 h. One mL of 2 N HC1 and 1 mL of n-butanol were added, and themixture vortexed. After centrifugation at 100 rpm for five minutes, theabsorbance of the butanol phase was read in a spectrophotometer at 420 nm.The values obtained needed to be adjusted to reflect the number of free aminogroups in the respective peptides compared to the standard (CEME had 6,CEMA had 8, and polymyxin B had 5).672. Killing Assay.This assay was performed as described previously (Lehrer et al., 1983).Briefly, reactions were carried out in 100 iL volumes and contained 106 CFU ofP. aeruginosa H187 in a low ionic strength buffer (10 mM potassium phosphate,pH 7.4), and either 2.5 Jig/mL or 5.0 Jig/mL of CEME. After either 20 or 60 mmat 37°C, samples of the bacteria were removed, diluted and plated to obtain aviable count.3. Lysozyme Lysis.The uptake of lysozyme into whole cells due to membranepermeabilization by various compounds was previously described by Hancocket al. (1981). Overnight cultures of P. aeruginosa H309 or .E. cloacae 218R1grown in LB-S were diluted 1 in ü in fresh medium and grown to an 0D600 of0.5-0.6. The cells were harvested in a Silencer H-103N clinical centrifuge at1800 g for 10 mm, washed once with one volume of assay buffer (5 mM HEPESpH 7.2, 5 mM KCN), and resuspended in the same buffer to an 0D600 of 0.5.Assays consisted of 600 jiL of cells with 50 jig/mL of chicken egg whitelysozyme and varying concentrations of cationic compounds. Cell lysis wasmeasured as a decrease in the 0D600 in a Perkin-Elmer dual beamspectrophotometer. Parallel experiments performed without lysozyme enabledthe measurement of the lytic activity of the compounds themselves. To testwhether or not permeabffity to lysozyme could be inhibited by divalent cations,various concentrations of MgC12 were added to the assay after the addition oflysozyme and before the addition of the test compound.684. i-N-phenylnaphthylamine Uptake.This assay was previously described by Loh et al. (1984). Cells wereprepared exactly the same as for the lysozyme lysis assay. i-Nphenylnaphthylamine (NPN) was dissolved in acetone at a concentration of 500jiM. NPN fluorescence was measured in a Perkin-Elmer 650-i OS fluorescencespectrophotometer using excitation and emission wavelengths set to 350 nmand 420 nm respectively, with slit widths of 5 nm. The assay was standardizedby adding 20 jiL of NPN (final concentration of 10 jiM) and 10 jiL of a 0.64mg/mL solution of polymyxin B (final concentration of 6.4 jig/mL) into 1 mL ofcells, and adjusting the resulting fluorescence to read 90% deflection on thechart recorder (90 arbitrary units). Various compounds were tested by adding10 jiL of different concentrations to a cuvette containing 1 mL of cells and 10jiM NPN. Permeabilizing activity was designated as the total fluorescence minusthe fluorescence due to NPN alone. Following the fluorescence measurement,the 0D600 of the cells was measured to ensure no significant cell lysis hadoccurred. Control experiments showed that neither acetone nor testcompounds alone resulted in an increase in fluorescence in the absence of NPN.5. Dansyl Polymyxin B Displacement.a. Lipopolysaccharide Isolation.P. aerugiriosa LPS was isolated as previously described (Darveau andHancock, 1983). The procedure was performed by Susan Farmer.69b. Dansyl Polymyxin B Synthesis.Dansyl polymyxin B was synthesized as described by Schindler andTeuber (1975). Briefly, 40 mg of polymyxin B and 10 mg of dansyl chloridewere mixed in 2 mL of 60 mM NaHCO3 and 40% acetone and incubated in thedark for 90 mm. The unreacted dansyl chloride was separated from the dansylpolymyxin B by gel filtration on a Sephadex G-50 column. The fractionscontaining dansyl polymyxin were extracted with 1/2 volume of n-butanol andevaporated to dryness in a desiccator at 37°C. The dansyl polymyxin B wasresuspended in 5 mM HEPES pH 7.0, quantified by dinitrophenylation andstored in aliquots at -20°C. This procedure was carried out by Susan Farmer.c. Assay.The method of Moore et al. (1986) was used to test how much dansylpolymyxin B was needed to saturate the binding sites on LPS. Briefly, 5 iLsamples of 100 iIM dansyl polymyxin B were titrated into 1 mL of 3 ig/mL ofLPS until a maximum fluorescence was reached. The fluorescence wasmeasured in a Perkin-Elmer 650-1 OS fluorescence spectrophotometer with anexcitation wavelength of 340 nm and an emission wavelength of 485 nm usingslit widths of 5 nm. A final concentration of dansyl polymyxin B giving 90-100% maximum fluorescence (2.5 iM) was chosen and used in all subsequentexperiments. For the binding displacement assays (Moore et al., 1986), 2.5 IIMdansyl polymyxin B was added to 3 jig/mL of H103 LPS in 5 mM HEPES pH7.2. Samples (5 jiL) of the test compounds were added and the decrease influorescence due to displacement of the dansyl polymyxin B from the LPS wasrecorded. The addition of the compound was continued until it resulted in onlya small (<5%) decrease in fluorescence. The data were plotted as the fraction of70dansyl polymyxin B bound as a function of the concentration of compound (I).The relative affinities of the compounds for the binding sites on LPS weredetermined by calculating the values directly from the graph. 150represented the concentration of compound that resulted in 50% maximaldisplacement of dansyl polymyxin B from the LPS. All experiments wereperformed a minimum of three times.For dansyl polymyxin B binding inhibition assays using whole cellsinstead of purified LPS, H309 cells were prepared in the same way as for thelysozyme lysis assay. The assay consisted of 10 jiL of cells at an 0D600 of 0.5,990 iL of 5 mM HEPES pH 7.2 and 5 mM KCN, and a concentration of dansylpolymyxin B that had been determined to result in 90-100% binding saturation.This concentration varied from day to day but usually was between 2.5 jiM and3.5 jiM. Compounds were titrated in, and 150 values determined as describedabove.6. Minimum Inhibitory Concentration.These assays were done according to the broth dilution method(Amsterdam, 1991). Briefly, cells were grown overnight at 37°C in LB-S anddiluted 1 in 10 000 in the same medium to give concentrations of about 10 toiü CFU/mL. Serial dilutions of the antimicrobial substances in LB-S wereperformed in a 96 well microtitre plate. Subsequently, 10 jiL of bacteria werepipetted into 100 jiL volume of the diluted antibiotic, and the plates incubatedovernight at 37°C. Samples of the bacterial inoculum were plated to ensurethey were within the proper inoculum range. The next day the microtitre plateswere scored for growth in the wells, and the MIC determined as the lowestantibiotic concentration that inhibited growth. To determine the effect of71cations on the MIC values of the various compounds, either 5 mM MgC12 or 80mlvi NaC1 were included in the LB-S medium. In the synergy MIC studies, thegiven sub-MIC concentrations of the peptides were included in the LB-Smedium.72RESULTSCHAPTER ONE. The Production of Cationic Peptides as GST FusionProteins.A. Introduction: Why Fusion Proteins Are Necessary.The HNP- 1 gene was initially cloned into the T7 RNA polymeraseexpression vectors pT7-5 and pT7-7 (Materials and Methods, section C. 1.).Attempts at producing HNP-1 from either of these two constructs ((Materialsand Methods, section C. 1.) were unsuccessful as determined by SDS-PAGE andAU-PAGE (data not shown). To ensure that the lack of peptide production wasnot a transcriptional problem, Northern blots were performed with uninducedand induced cultures of K38 (pGP1-2)(pKP 160) using oligonucleotide D (TableVI) as a probe. The results showed a small transcript being produced in theinduced cultures that was not present in the uninduced culture (data notshown). Based on these experiments, it was concluded that although thetranscription of the gene was occurring, the protein was either not beingproduced at all, or was being produced and subsequently degraded by the hostorganism.The inabifity to produce HNP- 1 by direct expression of the gene in E. coliled to attempts to produce it as a fusion protein. Certain requirements of afusion protein expression system had to be met if it were to be useful forproducing cationic peptides. First, the carrier protein had to have a highaffinity for a specific ligand to provide a simple means of purification. Second,the affinity tag had to contain a site specific cleavage sequence (eitherenzymatic or chemical) to allow the release of the cationic peptide. Third, thevector had to possess restriction enzyme sites that would enable the insertion ofthe gene immediately adjacent to the specific cleavage site. This would allow73the release of the peptide from the carrier protein without any extra amino acidsat the N-terminus of the peptide. This is critical in the case of cationic peptidessince it has been shown that the addition of one or more amino acids to thepeptide can alter its biological activity (Bessalle et al., 1992). Since the GSTfusion protein system, specifically the vector pGEX-3X, fulfilled these criteria, itwas used to express the genes encoding HNP- 1 and CEME.B. Construction of the pGEX-KP Vector.The vector pGEX-3X did not have the restriction enzyme sites that werecompatible with the ends of the HNP- 1 gene. Therefore, the multiple cloningsite (BamHI/SmaI/EcoRl) was changed to include the appropriate sites(SphI/HindIII/EcoPJ) necessary for the insertion of the HNP- 1 gene. This wasaccomplished using a series of genetic manipulations (Materials and Methods,sections B .3. and C .2.; Figure 9). When the HNP- 1 gene (Figure 1 OB) wascloned into the resulting vector, pGEX-KP (Materials and Methods, section C.2.;Figure 1 OA), the first amino acid codon of the HNP- 1 gene (Ala) was immediately3’ to the last amino acid codon of the factor Xa recognition site (Arg)(Figure 9).Theoretically, therefore, when the fusion protein (GST/HNP- 1) produced by thisconstruct (pGEX-HNP- 1) was purified and cleaved with factor Xa, there wouldbe no extraneous N-terminal amino acids on the HNP- 1 peptide.C. The Production and Purification of GST/HNP-l.The production of the GST/HNP- 1 fusion protein was investigated byinducing the expression of the gene with IPTG for varying amounts of time(Figure 11). The vector pGEX-3X was used as a control, and pGEX-KP was74Cut out BamHI-BaII fragment from pGEX-3Xprimer500bp IBallBallBclIIsolate PCR fragmentBcIIFactor XaI E G RA1CSphlLigateBallHindlil EcoRlHindIllFigure 9: Construction of pGEX-KP.All genetic manipulations are described in Materials and Methods,sections B.3. and C.2. The I-E-G-R sequence is the factor Xa recognition siteand the A-C residues are encoded by the SphI site and represent the first twoamino acids of HNP- 1. The arrow between them shows the site of factor Xacleavage.4Ball Sphl Hindill EcoRlPCRIBlunt ligateinto pTZ19UCut with Bclland EcoRlIsolate smallfragmentIsolate largeEcoRIIBcllfragmentSphlBcIlBcllEcoRlpG EX-KPBcII75A GSTI,’’.,’’’HNP-1 ConstructS Met CEME HPre Pro CassetteFigure 10: Construction of GST/cationic peptide fusion proteins.A, The pGEX-KP cloning vector. B, DNA fragment containing the HNP-1gene which was inserted into pGEX-KP using the SphI and HiridIII sites to formpGEX-HNP-1. C, DNA fragment containing the CEME gene, preceded by amethionine codon, which was cloned into pGEX-KP using SphI and Hindill toform pGEX-CEME. D, DNA fragment encoding the mammalian defensin preproregion which was inserted into the SphI sites of both pGEX-HNP- 1 and pGEXCEME to create pGEX-proHNP- 1 and pGEX-pr0CEME respectively. GST,glutathione-S-transferase; ptac, tac promotor; Amp, 3-lactamase gene; on,origin of replication; fXa, factor Xa recognition site; S, Sphi; H, Hindu!; E, EcoRI;Met, methionine codon.pGEX-KP AmpE. coil onlac IqS HNP-1BCHCEME ConstructSDPreMet SPro II761234 M5 6789101112III— -_—144Figure 11: Production of GST/HNP-1.Whole cell lysates of DH5a(pGEX-HNP-1)(lanes 1-4), DH5cL(pGEX-KP)(lanes 5-8), and DH5cc(pGEX-3X)(lanes 9-12) were electrophoresed throughan 11% SDS polyacrylamide gel and stained for protein with Coomassie blue.Approximately 15 jig of protein were loaded in each lane except lane 5 and 9 (5jig). Cells were grown to an 0D600 of approximately 0.5 and induced with 0.2mM IPTG for: 0 h (lanes 1,5 and 9); 1 h (lanes 2,6 and 10); 2 h (lanes 3,7 and11); or 3 h (lanes 4,8 and 12). Lane M, molecular weight standards (in kDa).Arrow, GST/HNP- 1; open circle, induced 30 kDa 13-lactamase; open square,native GST.—9467—43—0—3O77included to ensure that the genetic changes made to the vector had no negativeeffects on gene expression. The results showed that pGEX-3X and pGEX-KPboth produced a protein with an apparent molecular weight of 26 kDa (Figure11, lanes 6-8 and 10-12). The vector pGEX-HNP-l produced a heterologousprotein of 29 kDa which migrated just below the induced f3-lactamase protein(lanes 2-4). Maximal expression seemed to occur between 2 and 3 h. Otherstudies showed that if expression were allowed to continue, the fusion proteinbegan to be degraded after 6 h and was no longer detectable after overnightgrowth (data not shown).Once stable expression had been demonstrated, the next step was topurify the fusion protein. The affinity of GST for glutathione attached to anagarose matrix was utffized as a purification technique. The scheme for such apurification (Figure 12) contained all the basic elements of affinity tagpurification: the specific binding of the desired fusion protein to the matrix(usually from a heterogeneous population of proteins such as a cell extract), therelease of this protein from the matrix by competitive displacement, the specificrelease of the peptide from the carrier protein, and the final purification of thepeptide using a second passage over the affinity column to remove thecontaminating affinity tag. In this case, a soluble cell extract from a cultureinduced for GST/HNP- 1 production was passed over a glutathione agarosecolunm, and the fusion protein subsequently eluted using reduced glutathione(Materials and Methods, section G.1.).The results of this first stage in HNP- 1 purification are found in Figure13. Various samples were retained throughout the purification procedure tomonitor the state and location of the fusion protein. Lane 2 showed that thefusion protein was indeed being produced. The fact that lanes 3 and 4, thesoluble cellular fractions before and after being incubated with glutathione78AFigure 12: Schematic diagram of GST fusion protein affinity purification.A, Purification of the fusion protein from the soluble fraction of whole celllysates. B, Release of the target protein by factor Xa cleavage and itspurification by a second passage over the affinity matrix. Closed squares, GST;open circles, target protein.wash with buffer elute with 5 mM101531I ISD01careduced glutathioneSupernatant oflysed whole cells*10Factor XaSDBSDI.— II. —II •I. II• II —.1I — Itm• II II •—II• .11\_ /Voo%o 0O°Oc?°0000o%O oOc9o 0•R —0 •00 RO• • 0— 0 O0 —•purefusionproteinpurepeptide7920.1—14.4—789Figure 13: Affinity purification of GST/HNP-1.A culture of DH5o(pGEX-HNP-l) was induced for fusion proteinproduction (Materials and Methods, section F.2.) and the fusion protein purifiedby affinity chromatography (Materials and Methods, section G. 1.). Sampleswere retained after various steps in the procedure and subjected to 12% SDSPAGE prior to being stained for protein with Coomassie blue. Lanes: 1,uninduced whole cell lysate; 2, induced whole cell lysate; 3, soluble fractionbefore incubation with glutathione agarose beads; 4, soluble fraction afterincubation with glutathione agarose beads; 5, insoluble pellet fraction; 6,glutathione agarose bead-bound fraction before elution with 5 mM reducedglutathione; 7, glutathione agarose bead-bound fraction after elution with 5 mMreduced glutathione; 8, eluted fraction; 9, whole cell lysate of inducedDH5x(pGEX-KP); M, molecular weight standards (in kDa). Open circle,GST/HNP-1; open square, GST. Approximate protein content in each lane: 1-4, 15 jig; 5, 50 jig; 6-9, 10 jig.12345M694— --67—__—43—..b_=_ —30-p____80agarose beads, showed virtually no difference suggested that there was verylittle soluble GST/HNP- 1 in the cell. This was confirmed by the presence of alarge protein band with the apparent molecular weight of GST/HNP- 1 in theinsoluble pellet fraction (lane 5), which suggested that most of the fusionprotein was produced as inclusion bodies (see below). Nonetheless, someGST/HNP- 1 was produced as soluble protein that bound to glutathione agarosebeads (lane 6) and was released by competitive displacement with freeglutathione (lane 8). When samples of the protein that bound to and wereeluted from the affinity column were analyzed by SDS-PAGE, two distinctprotein bands were detected (Figure 13, lanes 6-8). The migration distance ofthese two bands corresponded to those of GST/HNP- 1 (compare with lane 2)and GST alone (compare to lane 9), which suggested that the fusion protein wassusceptible to proteolytic degradation, particularly at the fusion junctionbetween the carrier molecule and the peptide. Different conditions throughoutthe purification scheme were employed in an attempt to reduce the breakdownof this fusion protein. Neither lower temperatures nor the use of proteaseinhibitors such as PMSF, TLCK, TPCK or EDTA were able to enhance thestability of the protein (data not shown). The pGEX-HNP- 1 vector wastransformed into a number of different strains that were either mutants thatprevented the expression of certain proteases or the parents of these mutantstrains to investigate whether such hosts could produce stable GST/HNP- 1fusion protein. Table VII summarizes the data from these strains with respectto protein production and stability. Although several strains showed betterexpression of the fusion protein than DH5cs, none of them, including theprotease-deficient strains, were able to prevent the subsequent degradation ofthe fusion protein.81Table VII: Summary of Strains Used to Prevent Proteolytic Degradation ofGST/HNP 1.Strain Relevant Phenotype Expressiona DegradationbIn Whole Cell PurifiedLysates ProteinDH5a Original host used + — ++BL21 Host for T7 RNA +++ +polymerase expressionHMS174 Host for T7 RNA ++÷ +polymerase expressionUT5600 ornpr — ND NDC474 Parent of C475 and ++ — ++C476C475 lon ++ + ++C476 htpR — ND NDa very strong; ++, strong; +, weak; —, not detectedb >70% degraded; i-i-, 30-70% degraded; +, <30% degraded; —, nodegradation detected; ND, not determined82As indicated above, the induction of GST/HNP- 1 expression led to theformation of insoluble inclusion bodies (Figure 13, lane 5). This phenomenonhas been documented for many recombinant proteins (Kane and Hartley, 1988)and, in fact, can be useful in a purification scheme (Marston, 1986). Sinceinclusion bodies are quite resistant to mild conditions of solubilization, manyproteins in a heterogeneous pellet can be removed by such treatment before theinclusion bodies are solubilized using harsher conditions. Many differentselective solubiization conditions were tested on an inclusion body preparationof GST/HNP- 1 before a final purification scheme was established (Materials andMethods, section G. 1.). This procedure used a lower induction temperature(30°C), which allowed the eventual solubilization of the inclusion bodies withmilder conditions (Stein and Noteborn, 1988), and differential ureasolubiizations to obtain a pure preparation of GST/HNP- 1. Figure 14 showsthat many of the contaminating proteins were removed through the procedure(lanes 2-4) and indicates the purity of the solubiized inclusion bodies (lane 5).The loss of some of the fusion protein which was either solubiized by 3 M urea(lane 4) or left in the pellet after the 8 M urea treatment (lane 6) was expecteddue to the heterogeneous nature of the inclusion bodies. The protein that wassolubilized by the 8 M urea migrated slightly faster in SDS-PAGE (lane 5)possibly due to carbamylation which neutralizes positive charges (Stark, 1965;Marston and Hartley, 1990). This carbamylation occurred despite the use ofdeionized urea and the inclusion of 5 mM Tris in the solubilization buffer tominimize the amount of cyanate ions present.83123M45694—67—43—30—__20.1—14.4Figure 14: Purification of GST/HNP-1 inclusion bodies.An induced culture of DH5ci(pGEX-HNP-1) was lysed and the insolublepellet extracted with a series of different solutions (Materials and Methods,section G. 1.). Samples were electrophoresed on a 12% SDS polyacrylamide geland stained for protein with Coomassie blue. Lanes: 1, induced whole celllysate; 2, soluble fraction from lysed cells; 3, soluble extract after 10 mM TrisHC1 pH 8.0, 10 mM EDTA, 1% Triton X-100 wash; 4, soluble extract after 3 Murea wash; 5, soluble extract after 8 M urea solubiization; 6, insoluble pelletafter all washes and solubilizations; M, molecular weight standards (in kDa).Arrow indicates GST/HNP-1. Approximate protein content in each lane: 1-3,15 jig; 4-5, 3 jig; 6, 50 jig.84D. Factor Xa Cleavage of GST/HNP- 1.Inclusion bodies solubilized by differential urea extraction were dialyzedagainst factor Xa cleavage buffer using a two step procedure (Materials andMethods, section G. 1.) to avoid the precipitation of the protein that can occur ifthe urea was removed too quickly. The protein was subjected to factor Xacleavage using a number of different incubation temperatures and times (Figure15). Although incubation at 4°C can be used to prevent non-specific proteolysisof the fusion protein, this condition did not result in the cleavage of anyGST/HNP-1 (data not shown). Incubation at 23°C resulted in some fusionprotein cleavage as indicated by the reduction of the GST/HNP- 1 band in lanes3-5; however it led to very little, if any, protein corresponding to the size of HNP1, even after long incubation times (lane 5). Increasing the temperature to 37°Cimproved the cleavage of the fusion protein and resulted in the appearance of aprotein band of approximately 3500 daltons, the predicted molecular weight ofHNP-l. This protein band was electroblotted onto an Immobilon transfermembrane and analyzed for its N-terminal amino acid sequence. The analysisshowed that the HNP- 1 peptide was present in this band. Therefore, the HNP- 1peptide was released from the GST/HNP- 1 fusion protein by factor Xa but onlyunder conditions (37°C for 60 h with an enzyme to substrate ratio of 1:25;Figure 16) that are much harsher than normally required (Nagai andThøgersen, 1984; Smith and Johnson, 1988). The reason for this may bebecause the factor Xa recognition site was inaccessible to the large (55 kDa)factor Xa molecule. This inaccessibility may arise because of the highlyconstrained and compact structure of the defensin molecule, which mightappear as a bulky addition to the GST protein. In addition, thedenaturation/renaturation step of inclusion body purification may have led to8516.9—14.4 —8.1—6.2”2.512345M67 8Figure 15: Factor Xa cleavage of GST/HNP 1.Solubiized GST/HNP- 1 inclusion bodies were dialyzed against factor Xacleavage buffer (Materials and Methods, section H. 1.) and subjected to factor Xacleavage at 23°C (lanes 3-5) or 37°C (lanes 6-8). Samples were removed at 2 h(lanes 3 and 6), 6 h (lanes 4 and 7), and 24 h (lanes 5 and 8) andelectrophoresed on a 12-25% gradient SDS polyaciylainide gel which was thenstained for protein with Coomassie blue. Other lanes: 1, induced whole celllysate of DH5o(pGEX-KP); 2, uncleaved, solubilized GST/HNP- 1 inclusionbodies; M, molecular weight standards (in kDa). The arrow indicates theprotein band in which HNP- 1 was found. Equal volumes of the reactionmixture were loaded for lanes 2-8.861216.9 —14.4—8.1—__6.2”2.5—Figure 16: Optimal factor Xa cleavage of GST/HNP-l.Samples of solubilized GST/HNP- 1 inclusion bodies were incubated at37°C for 60 h with (lane 2) or without (lane 1) factor Xa at an enzyme tosubstrate ratio of 1:25. Samples were then subjected to 12-25% gradient SDSPAGE and stained for protein with Coomassie blue. The arrows indicate themajor GST cleavage products (see text). 5 jig of protein was loaded in lane 1and 10 jig of protein was loaded in lane 2.87incorrect disulfide bond formation and inaccurate protein folding and thuscontributed to the inaccessibility of the target cleavage site.The sequencing analysis of the 3500 dalton molecular weight bandrevealed the presence of another N-terminal sequence that started with AsnLys-Lys-Phe-Glu-Leu-Gly. Upon examination of the GST amino acid sequence(Smith et al., 1986), it was discovered that this contaminating peptidecorresponded to an internal peptide fragment (Figure 17), which suggested itarose from the non-specific factor Xa cleavage of GST. This suggestion wassupported by the fact that this peptide sequence in GST is preceded directly byan arginine residue, which is the amino acid found at the end of the factor Xarecognition site. Further N-terminal amino acid sequence investigation of theother two major bands produced in the factor Xa digest of GST/HNP- 1 (Figure16, lane 2, indicated by arrows), revealed that they also were proteolyticfragments of GST which corresponded to its first 42 ( 5.0 kDa) and last 145 (16 kDa) amino acids (Figure 17). The latter peptide, as with the 31 amino acidproteolytic fragment, was produced by cleavage of the protein after an arginineresidue, which suggested that there were two sites in the GST/HNP- 1 fusionprotein, other than the authentic recognition site, that were susceptible tocleavage by factor Xa. Since GST is not usually cleaved internally by factor Xa(Smith and Johnson, 1988), the possibility existed that these sites becamesusceptible due to improper refolding of the GST moiety during the renaturationof the GST/HNP- 1 fusion protein after inclusion body solubilization with 8 Murea. To test this hypothesis, purified GST was denatured using 8 M urea, andrenatured using the same conditions as for GST/HNP- 1 (sequential dialysisagainst 1 M urea in 5 mM Tris-HC1 pH 8.0, and factor Xa cleavage buffer). ThisdenaturedIrenatured GST was subjected to factor Xa cleavage as described forGST/HNP- 1 and analyzed by gradient SDS-PAGE (Figure 18). The results88TKLILGYWKIKGLVQPTRLLLEYLEEKYEEHLYERDEGDKWRNKKFELGLEFPNLPYYIDGDVKLTQSMAIIRYIADKHNMLGGCRKERAEISMLEGAVLDIRYGVSRIAYSKDFETLKVDFLSKLPEMLKMFEDRLCHKTYLNGDHVTHPD FMLYDALDVVLYMDPMCLDAFPKLVCFKKRIEAIPQIDKYLKSSKYIAWPLQGWQATFGGGDHPPKFigure 17: Amino acid sequence of GST.The arrows indicate non-specific factor Xa cleavage sites indenatured/renatured GST/HNP- 1 and GST.8912 34M567891011M121314—16.9—14.4i_8.16.2—2.5Figure 18: Cleavage of denatured/renatured GST by factor Xa.GST that was denatured by 8 M urea then renatured (lanes 1-7), andsolubiized GST/HNP- 1 inclusion bodies (lanes 8-14) were cleaved with factorXa at 37°C using different enzyme to substrate ratios and incubation times.Samples were and electrophoresed on a 12-25% gradient SDS polyacrylamidegel which was then stained for protein with Coomassie blue. Conditions:enzyme to substrate ratio, 1:25 (lanes 2-4 and 9-11) and 1:12 (lanes 5-7 and12-14); time, 8 h (lanes 2,5,9 and 12), 24 h (lanes 3,6,10 and 13), and 60 h(lanes 4,7,11 and 14). Other lanes: 1, uncleaved GST; 8, uncleaved GST/HNP1; M, molecular weight standards (in kDa). Equal volumes of the GST cleavagereaction were loaded in lanes 1-7. Equal volumes of the GST/HNP-1 cleavagereaction were loaded in lanes 8-14.90showed that GST treated in this manner was indeed susceptible to non-specificfactor Xa cleavage, yielding the three proteolytic fragments found in factor Xadigests of GST/HNP-1 (Figure 18, lanes 6 and 7, 13 and 14). The fact that thecleavage of GST was not as extensive as GST/HNP- 1 may indicate that thepresence of the HNP- 1 moiety contributed to the improper refolding of the GSTmolecule in GST/HNP-l.E. The Production of GST/proHNP- 1.Although HNP- 1 was successfully produced and released in the GSTfusion protein system, further purification would have been difficult consideringthe small amount of peptide released and the contaminating GST fragment ofidentical molecular weight. The formation of GST/HNP- 1 inclusion bodies alsoprevented the use of affinity chromatography to purify the fusion protein,although recent studies have been able to accomplish the affinity purification ofother GST fusion proteins that form inclusion bodies (Hartman et al., 1992). Itis not entirely understood why some recombinant proteins form inclusionbodies while others do not (Kane and Hartley, 1988), but one hypothesis is thatinclusion bodies are formed due to improper folding of protein intermediates(Mitraki and King, 1989). Therefore, in the case of fusion proteins, anyproperties of the target sequence that contribute to the misfolding of the carrierprotein can be the cause of inclusion body formation. Some of the propertieswhich could apply to HNP- 1 include a high charge density or any number ofcysteine residues that could result in improper intra- and interchain disulfidebond formation. Although little could be done about incorrect disulfide bondformation, an attempt was made to minimize the effect of the high positivecharge density of the HNP- 1 molecule on proper fusion protein folding. In vivo,91HNP1 is produced in the neutrophils as a prepro molecule (Harwig et al.,1992). The pro region of this molecule is highly negatively charged (Figure 3)and is believed to neutralize and stabilize the positively charged defensinmolecule until it is properly compartmentalized into primary granules(Michailson et al., 1992) during which the pro peptide is removed and thedefensin forms its native configuration. It was hypothesized that the inclusionof a prepro defensin cartridge between the GST and HNP- 1 genes might resultin the stabilization of the fusion protein as well as enhance its solubility byneutralizing the positive charges on the HNP- 1 moiety. Therefore, fouroligonucleotides (M-P, Table VI) encoding the prepro defensin sequence (Figure3) on an SphI fragment (Figure 1OD) were synthesized, annealed and insertedinto the SphI site of pGEX-HNP- 1 to form pGEX-proHNP- 1. Proper insertionorientation was confirmed by restriction endonuclease digestion analysis.Production of GST/proHNP- 1 was performed (Materials and Methods, sectionsF.2. and G. 1.) to determine if the presence of the prepro sequence preventedproteolysis of the fusion protein. The results indicated that GST/proHNP- 1 wasproduced in F. coli (Figure 19, lane 2). Although some of the protein wassoluble as indicated by the difference between the soluble fraction before (lane3) and after being incubated with glutathione agarose beads (lane 4), much ofthe protein was still found in the insoluble pellet as inclusion bodies (lane 5).When the soluble fusion protein was purified on glutathione agarose beads, itwas again found to be proteolytically degraded (lanes 6-8) although apparentlynot to the same degree as GST/HNP-1 (Figure 13, lanes 6-8). It was determinedfrom this experiment that the presence of the prepro defensin cartridge inpGEX-proHNP- 1 had a slight effect on the stability but not on the solubiuty ofthe fusion protein.921 234M56 7894—67— e.43—J_._ 030—20.1—14.4—Figure 19: Production and affinity purification of GST/proHNP-l.A culture of DH5cc(pGEX-proHNP- 1) was induced for fusion proteinproduction (Materials and Methods, section F.2.) and the fusion protein purifiedby affinity chromatography (Materials and Methods, section G. 1.). Sampleswere retained after various steps in the procedure and subjected to 12% SDSPAGE prior to being stained for protein with Coomassie blue. Lanes: 1,uninduced whole cell lysate; 2, induced whole cell lysate; 3, soluble fractionbefore incubation with glutathione agarose beads; 4, soluble fraction afterincubation with glutathione agarose beads; 5, insoluble pellet fraction; 6,glutathione agarose bead-bound fraction before elution with 5 mM reducedglutathione; 7, glutathione agarose bead-bound fraction after elution with 5 mMreduced glutathione; 8, eluted fraction; M, molecular weight standards (in kDa).Open circle, GST/proHNP- 1; open square, GST. Approximate protein content ineach lane: 1-5, 20 pg; 6-8, 5 g.93F. The Production of GST/CEME.The cecropin A/melittin hybrid peptide CEME was chosen for this studybecause it represented the x-helical (non-defensin-like) family of cationicpeptides and it possessed potent antibacterial activity against P. aeruginosa(Wade et al., 1990). The hypothesis that this peptide would be less likely toform inclusion bodies in the GST fusion protein system than HNP- 1 came fromthe fact that it contained no cysteine residues and consequently no disulfidebonds. The gene encoding CEME was created by a two step process (Materialsand Methods, section C.2.) because of a two base pair deletion in the originalgene. It should be noted that although a factor Xa cleavage site was present inthis construct, the mature CEME peptide was preceded directly by a methionineresidue (Figure bC). Thus CNBr cleavage would ensure that no extra aminoacids were present at its N-terminal end. The plasmid pGEX-CEME, wasexpressed in F. coli as described for both pGEX-HNP- 1 and pGEX-proHNP- 1(Materials and Methods, section F.2.). The samples prepared from theproduction and purification of GST/CEME were run in duplicate on SDS-PAGE(Figure 20A) and one set of samples was electroblotted onto nitrocellulose forWestern immunoblotting (Figure 20B). The primary antibody used to detect thefusion protein was a rabbit polyclonal serum raised against GST/CEMA(Materials and Methods, section D. 1.) As predicted, virtually all of the proteinproduced by this vector was found in the soluble cellular fraction (Figure 20B,lane 3) while little, if any, was detected, even with the ccGST/CEMA antibody, inthe insoluble pellet fraction (Figure 20B, lane 5). As with GST/HNP-1, however,GST/CEME fusion protein was highly susceptible to proteolytic degradation.SDS-PAGE of a sample of GST/CEME that was purified using glutathioneagarose beads revealed that almost all of the protein was degraded to the94A123456794—67—43-30— — —20.1—14.4—23456794—67—43—30—20.1—14.4—Figure 20: Production and affinity purification of OST/CEME.A culture of DH5x(pGEX-CEME) was induced for fusion proteinproduction (Materials and Methods, section F.2.) and the fusion protein purifiedby affinity chromatography (Materials and Methods, section G. 1.). Samplesretained throughout the procedure were run in duplicate on an 11% SDSpolyacrylamide gel and either A, stained for protein with Coomassie blue, or B,electroblotted onto nitrocellulose for Western immunoblotting. Lanes: 1,uninduced whole cell lysate; 2, induced whole cell lysate; 3, soluble fractionbefore incubation with glutathione agarose beads; 4, soluble fraction afterincubation with glutathione agarose beads; 5, insoluble pellet fraction; 6,glutathione agarose bead-bound fraction before elution with 5 mM reducedglutathione; 7, purified GST. The primary antibody used in B was a rabbitpolyclonal xGST/CEMA serum. Open circle, GST/CEME; open square, GST.Molecular weight standards are indicated in kDa. Approximate protein contentin each lane: 1-4, 25 jig; 5-7, 5 jig.B195molecular weight of GST (Figure 20A and B, lanes 6 and 7).This is in contrast tothe sample of GST/HNP- 1 which showed that almost half of the purified fusionprotein remained intact throughout the purification procedure (Figure 13, lanes6-8).G. The Production and Purification of GST/pr0CEME.In an attempt to stabilize the GST/CEME fusion protein, anoligonucleotide cartridge encoding the anionic prepro defensin sequence (Figure3) was cloned into the SphI site of pGEX-CEME to form pGEX-proCEME. Thisplasmid was transformed into DH5o and the gene expressed as previouslydescribed (Materials and Methods, section F.2.). Samples subjected to SDSPAGE were either stained for protein content with Coomassie blue or Westernimmunoblotted to detect the fusion protein (Figure 21). Although theGST/pr0CEME fusion protein was only weakly detected in a sample from aninduced culture stained with Coomassie blue (Figure 21A, lane 2), itsproduction was confirmed with the Western blot (Figure 21B, lane 2). It wasinteresting to note that the fusion protein in this sample was already partiallydegraded. The addition of the prepro sequence to this fusion protein resulted ina shift from soluble protein (very little GST/pr0CEME in lanes 3 and 4), as wasthe case for GST/CEME, to insoluble inclusion bodies (lane 5). The smallamount of protein that was produced in soluble form (lane 8) was almostcompletely degraded to the size of GST (lane 9). Therefore, although theinclusion of the prepro region did not stabilize the soluble form of the fusionprotein, it did result in the formation of insoluble inclusion bodies whichprotected the heterologous protein from extensive proteolysis.96A1234B56789012345678994—67 —43—30—0—201144—Figure 21: Production and purification of soluble and insoluble GST/pr0CEME.A culture of DH5cL(pGEX-proCEME) was induced for fusion proteinproduction (Materials and Methods, section F.2.) and the fusion protein purifiedeither by affinity chromatography or differential solubiization of the insolublepellet fraction (Materials and Methods, section G.1.). Samples retainedthroughout the procedure were run in duplicate on an 11% SDS polyacrylamidegel and either A, stained for protein with Coomassie blue, or B, electroblottedonto nitrocellulose for Western irnmunoblotting. Lanes: 1, uninduced wholecell lysate; 2, induced whole cell lysate; 3, soluble fraction before incubationwith glutathione agarose beads; 4, soluble fraction after incubation withglutathione agarose beads; 5, insoluble pellet fraction; 6, heated and 7,unheated samples of insoluble pellet fraction after 3% octyl -POE extraction; 8,glutathione agarose bead-bound fraction; 9, purified GST. The primaryantibody used in B was a rabbit polyclonal xGST/CEMA serum. Open circle,GST/pr0CEME; open square, GST. Molecular weight standards are indicated inkDa. Approximate protein content in each lane: 1-4, 25 jig; 5-7, 15 jig; 8-9, 5jig.94—67—43 —30—20.197Since the CEME peptide had to be released from GST/proCEME usingCNBr, the inclusion bodies eventually needed to be solubilized in 70% formicacid. Such conditions would solubiize most protein contained in the insolublepellet fraction. Therefore, a novel method of purifying the GST/proCEMEinclusion bodies was developed. The strategy in this protocol was to solubiizeall the contaminating proteins, thus leaving only the inclusion bodies to besolubilized by 70% formic acid. Since outer membrane proteins are known tobe the major contaminants in inclusion body preparations (Veeraragavan, 1989)the protocol used 3% 0-POE, a detergent that has been shown to solubilizesuch proteins (Siehnel et al., 1992). The pellet was sequentially washed asdescribed in Materials and Methods, section G. 1. The selective extraction ofouter membrane proteins was easily monitored by SDS-PAGE analysis of heated(100°C for 10 mm) and non-heated samples, since outer membrane proteins areknown to be heat-modifiable (Hancock and Carey, 1979). Both the heated andnon-heated samples of post-3% 0-POE extraction pellets were analyzed by SDSPAGE and Western immunoblotting (Figure 2 1A and B, lanes 6 and 7). Theresults showed little contamination and little difference between the twosamples, which indicated there were no detectable membrane proteincontaminants in the pellet. Western blots of the 3% 0-POE supernatantsamples showed that no GST/pr0CEME inclusion bodies were solubilized underthese conditions (data not shown).H. The Release of CEME From GST/pr0CEME Using CNBr.The GST/pr0CEME inclusion bodies were solubilized in 70% formic acidand subjected to CNBr cleavage (Materials and Methods, section H.2.). Analysisof the cleavage products by AU-PAGE revealed a complex array of peptides98(Figure 22A, lane 1). This was expected given the fact that the inclusion bodypreparation was not entirely homogeneous and that the GST protein contains 8methionine residues (Figure 17). Nonetheless, a peptide band whichcorresponded to the migration of chemically synthesized CEME was detected(Figure 22A, lane 1, arrow) Identical samples were tested for antibacterialactivity using a gel overlay assay (Materials and Methods, section J.; Figure22B). The results showed a zone of bacterial lysis in the CNBr-cleavedGST/pr0CEME sample that corresponded to the zone produced by syntheticCEME. This confirmed that biologically active CEME could be produced in thissystem. However further purification of this peptide from the complicatedmixture of CNBr cleavage fragments was presumed to be too difficult andpotentially inefficient given the fact that the CEME band was a minor species.I. Summary.Both HNP- 1 and CEME were produced as GST fusion proteins. Whenmade as soluble proteins, they were highly susceptible to proteolyticdegradation. GST/HNP- 1 formed insoluble inclusion bodies which werepurified using differential urea solubilization. Cleavage of this fusion proteinwith factor Xa released a peptide fragment that was confirmed by N-terminalamino acid sequencing to be HNP- 1. The inclusion of the defensin preproregion in the fusion protein (GST/proHNP- 1) rendered it only slightly morestable than GST/HNP-1. Compared to GST/HNP-1, GST/CEME was moresoluble but also more susceptible to proteolysis. The presence of the preproregion in GST/pr0CEME dramatically increased the stability of this fusionprotein, presumably in part because it resulted in the formation of inclusionbodies. These inclusion bodies were purified by removing contaminating99A B12 2-Figure 22: CNBr release of active CEME from GST/pr0CEME.GST/pr0CEME inclusion bodies were solubiized and cleaved with CNBr(Materials and Methods, section H.2.). Samples of the CNBr digest (lane 1) andpurified CEME obtained by chemical synthesis methods (lane 2) wereelectrophoresed in duplicate on a 15% acid urea polyacrylamide gel and eitherA, stained for protein with Coomassie blue, or B, tested for antibacterial activityusing a gel overlay assay (Materials and Methods, section J.). Arrow indicatesthe CEME peptide. Approximate protein content in each lane: 1, 90 jig; 2, 10jig.100proteins with 3% 0-POE, and cleaved with CNBr to release active CEMEpeptide, as confirmed by a gel overlay assay.101CHAPTER TWO: The Production of Cationic Peptides as Protein A FusionProteins.A. Introduction.Protein A from S. aureus has a high binding affinity for the Fe portion ofIgG molecules (Forsgren and Sjoquist, 1966). This property was the basis of anumber of gene fusion vectors (Uhlén et al., 1983; Nilsson et al., 1985a) whosefusion proteins could be purified by IgG affinity chromatography. Severalproteins have been produced in this system including 13-galactosidase andalkaline phosphatase (Nilsson et aL, 1985a) and human insulin growth factor(Moks et al., 1987a). This chapter describes the use of the protein A fusionprotein system for producing cationic peptides. The vector pRIT5 (Figure 23A)was chosen for these studies because of its host range (it possesses both E. coltand S. aureus origins of replication) and the presence of the protein A signalsequence which directs export of the fusion protein to the periplasm when in E.colt or to the external medium when in S. aureus, and thus may serve to protectit from intracellular proteases.B. Construction of the Vectors.The restriction enzyme sites available on pRIT5 were not compatible withthose of the gene cartridges that encoded HNP- 1, CEME or CEMA. Rather thanchanging the multiple cloning site of the vector several times to create thenecessary reading frames for the cationic peptide genes, the genes themselveswere modified using a series of oligonucleotides (Materials and Methods, sectionC.3.). Figure 23A-C are schematic representations of the changes made to thegenes (compare with Figure 8A-C). The genes encoding CEME and CEMA were102ASignalProtein APre Pro CassetteHLHtrE BEFigure 23: Construction of protein A/cationic peptide fusion proteins.A, The pRIT5 vector; B-D, DNA fragments from Figure 1OB-D encodingCEME, HNP-1 and the mammalian defensin prepro region. Their restrictionendonuclease sites were altered using oligonucleotide adaptors. The genes werecloned into pRIT5 using a Sail fragment for CEME and a BamHI fragment forHNP- 1 and preproHNP- 1 (Materials and Methods, section C.3.). Protein A,truncated protein A gene of S. aureus; SS, protein A signal sequence; pPA,protein A promotor; Amp, [3-lactamase gene; CAT, chloramphenicol acetyltransferase; or, origin of replication; S, SphI; L, Sail; B, BamHI; H, Htn.dIII; N,NdeI; E, EcoRI; fXa, factor Xa recognition site; Met, methionine codon.CATS. aureus onE. coli onCEMEfXaS L S B MetB II_Li IB MetS\/sN BNCEME ConstructHNP-1HNP-1 ConstructPre ProMet S103excised from pGEX-CEME-S and pGEX-CEMA--S using Sail, and inserted intothe SalT site of pRIT5 to form pPA-CEME and pPA-CEMA respectively.Similarly, BamHI fragments from pGEX-HNP- 1-B and pGEX-proHNP- 1-Bencoding the HNP- 1 and preproHNP- 1 genes respectively were cloned into theBamHI site of pRIT5 to form pPA-HNP- 1 and pPA-proHNP- 1.C. Expression of pPA-CEME in E. coiLFusion proteins produced by pRIT5 in E. coil are transported to theperiplasm. Both pRIT5 and pPA-CEME were transformed into DH5a andexpressed in cells grown to mid-log phase. Samples of whole cells wereanalyzed by SDS-PAGE and Western immunoblot using IgG antibody to detectthe protein A moiety (Figure 24). There was no detectable difference betweenthe pRIT5 and pPA-CEME samples (Figure 24B, lanes 1 and 2), indicatingextensive degradation of the PA/CEME fusion protein. In fact, it appears as ifthe protein A carrier molecule itself is subject to proteolysis in E. coil (comparelanes 1 and 3). This result is not unexpected since other researchers using thesame vector system to produce human IGF-I found that using E. coil as a hostorganism led to the production of unstable fusion protein (Nilsson et aL, 1985b;Moksetal., 1987a).D. Expression and Purification of Protein A/Cationic Peptide Fusion Proteinsin S. aureus.Since no stable fusion protein was produced in the E. coil periplasm, allfour plasmids, pRIT5, pPA-CEME. pPA-HNP-1, and pPA-proHNP-1, wereelectroporated into S. aureus (Materials and Methods, section B.4.) where the104A123456 7Figure 24: Production of protein A/cationic peptide fusion proteins.Fusion proteins were produced in E. coil DH5o (lanes 1 and 2) or S.aureus RN4220 (lanes 3-6)(Materials and Methods, section F.3.). Whole celllysates (lanes 1 and 2) and extracellular supernatants (lanes 3-6) were run induplicate on an 11% SDS polyacrylamide gel and either A, stained for proteinwith Coomassie blue or B, transferred to nitrocellulose for Westernimmunoblotting. Lanes: 1, DH5cz(pRIT5); 2, DH5a(pPA-CEME); 3,RN4220(pRIT5); 4, RN4220(pPA-CEME); 5, RN4220(pPA-HNP- 1); 6,RN4220(pPA-proHNP-1). Lane B7 is a 5 jig sample of PA/CEME purified on anIgG Sepharose column (Materials and Methods, section G.2.). Open square,PA/CEME; open circle, native S. aureus protein A. Molecular weight standardsare shown in kDa. Equal volumes of culture supematants were loaded in lanes3-6.12B0‘a105fusion protein would be exported to the external medium. The proteins wereexpressed (Materials and Methods, section F.3.) and samples of culturesupernatant analyzed by SDS-PAGE and Western immunoblot (Figure 24). Theprotein A carrier molecule and the PA/CEME fusion protein were barely visibleon a Coomassie-stained polyacrylamide gel (Figure 24A, lanes 3 and 4) butPA/HNP- 1 and PA/proHNP- 1 were not detected (lanes 5 and 6). The Westernimmunoblot, however, clearly showed that the proteins were produced andexported with little (Figure 24B, lanes 3 and 4) or no (lanes 5 and 6) proteolyticdegradation. Once expression of the genes had been established, PA/CEMEwas chosen as the fusion protein to be purified. PA/CEME was affinity purifiedfrom culture supernatants of S. aureus RN4220(pPA-CEME) on an IgGSepharose column (Materials and Methods, section G.2.). A sample of theprotein eluted off the column showed that only PA/CEME and native S. aureusprotein A were present in the sample (Figure 24B, lane 7). These were the onlytwo bands present in a Coomassie-stained SDS polyacrylamide gel (data notshown).E. Release and Purification of CEME.Purified PA/CEME fusion protein was pooled, lyophilized and treatedwith CNBr to release the CEME peptide (Materials and Methods, section H.2.).The protein A moiety of the fusion protein contains three methionine residuesother than the one that precedes the CEME peptide (Uhlén et aL, 1984; Nilssonet al., 1985a), which resulted in a total of five CNBr fragments including CEME.Native protein A present in the sample contained a total of six methionineresidues (Uhlén et al., 1984) which gave rise to seven fragments, four of whichwere different from those produced by the protein A carrier molecule (which is a106truncated version of native protein A). Fortunately, all the CNBr fragmentsfrom the protein A carrier molecule had p1 values that ranged from 3.9 to 4.7compared to 11.3 for CEME. This difference in p1 values was evident when asample of CNBr-cleaved PA/CEME was analyzed by AU-PAGE (Figure 25, laneD) which separates proteins on the basis of charge and size. The CEME peptidewas well separated from the major contaminating protein, indicating that thesedifferences in p1 values could be used in subsequent purification techniques.The PA/CEME CNBr digest was resuspended in 5% formic acid andfractionated on a Bio-Gel P100 column by isocratic elution with 1% acetic acid(Materials and Methods, section I.). Protein content of the fractions wasmonitored by absorbance at 280 nm (Figure 25A). A sample from every thirdfraction, beginning with number 27, was electrophoresed on an acid ureapolyacrylamide gel (Figure 25B) which revealed that fragments from the CNBrdigest eluted from the column well ahead of CEME. Fractions which containedCEME (Figure 25A, indicated by bar) were pooled and lyophilized.The final purification step was by reverse phase chromatography using aPepRPC FPLC column (Materials and Methods, section I.). If large samples wereto be purified, an initial purification step was performed on a larger Bio-Sil C18reverse phase column which possessed a higher loading capacity but resultedin poorer resolution (data not shown). Samples from Bio-Gel P100 or Bio-SilC18 chromatography that contained CEME were dissolved in 0.1% TFA andapplied in 1 mg quantities to the PepRPC column. Peptides were eluted fromthe column using a tri-phasic acetonitrile gradient (Figure 26; Materials andMethods, section I.). Under these conditions, CEME was eluted byapproximately 44% acetonitrile as a homogeneously pure peptide thatcomigrated with chemically synthesized CEME on an acid urea polyacrylamidegel (Figure 26, inset). A sample of this purified CEME was electroblotted onto107ACC—— —— —— — —Figure 25: Bio-Gel P100 chromatography of CNBr-digested PA/CEME.Purified PA/CEME was cleaved with CNBr, lyophilized, resuspended andsubjected to gel exclusion chromatography (Materials and Methods, section I.).A, Proffle of protein content in colunm fractions. The bar indicates the fractionsthat were pooled for RP-FPLC. B, 15% AU-PAGE in which every third fractionbetween 27 and 75 (inclusive) was analyzed and stained for protein withCoomassie blue. D, 50 ig of PA/CEME CNBr digest; C, 15 pg of syntheticCEME control; arrow, CEME. Approximately 5-10 jig from each column fractionwas loaded.1080.4 100C0.30.290807060C‘-“-IC.)40“p30201000 10 20 30 400.10.0Fraction #Figure 26: Reverse phase chromatography of CEME.Pooled fractions from Bio-Gel P100 chromatography were applied to aPepRPC HR 5/5 column and eluted with an acetonitrile gradient (Materials andMethods, section I.). Fractions were monitored for protein content and analyzedby AU-PAGE. Inset: 15% AU-PAGE showing a 5 jig sample from fraction 33(lane 1), and 5 jig of chemically synthesized CEME (lane 2).109an Immobjion transfer membrane (Materials and Methods, section K. 1.) andanalyzed for both amino acid content and N-terminal amino acid sequence. Allthe data was consistent with the amino acid sequence and content of authenticCEME. The yield of the CEME peptide from a 60 L fermentor run ofRN4220(pPA-CEME) was not overly impressive (Table VIII). It should be noted,however, that this was the only attempt at large scale production of CEME.Although the bacterial production of cationic peptides was a goal of this study,optimization of production was not the focus, and therefore no attempt wasmade to improve the yield of the peptide in this system.To determine whether antibacterial activity was retained during thepurification process, samples of isolated PA/CEME fusion protein, CNBrdigested PA/CEME, purified CEME from the PepRPC column, chemicallysynthesized CEME, and melittin were subjected to AU-PAGE (Figure 27A) andtested for activity using a bacterial overlay assay using E. coli DC2 as the testorganism. Figure 27B shows that CEME produced by recombinant DNAtechniques had antibacterial activity comparable to melittin and CEMEproduced by chemical synthesis, while the uncleaved PA-CEME showed noactivity.F. Purification of Other Cationic Peptides.The CEMA peptide, which is the CEME analogue with two extra Cterminal amino acids and two additional positive charges (Figure 3), waspurified to homogeneity from culture supernatants of RN4220(pPA-CEMA)using the same procedure as for CEME by Dr. M. Brown. Dr. M. Brown alsoused a similar scheme to purify PA/HNP- 1. The fusion protein was isolated,cleaved with CNBr, and passed over a Bio-Gel P100 column. This resulted in a110Table VIII: Percent Yield of CEME from a 60 1 Fermentor Run.Protein Sample Concentration Volume Total Protein Estimated Yieldamount of(mg/mL) (mL) (mg) CEME (%)(mg)Supernatent after 0.la 50000 5000 N/A N/A0.4 jim filtrationSupernatent after 0.26a 5000 1309 N/A N/A1 kDa cut-offconcentrationAfter IgG 106 6 636 22.3c 100SepharosepurificationAfter CNBr 32 14 448 15.7C 70treatmentAfter P100 gel ioa 5 50 8d 36ifitrationpurificationAfter RP-FPLC 33 1.5 5 4.5using Bio-SilAfter RP-FPLC 054b 1.1 0.6 O.6 2.7using PepRPCa Determined by the modified Lowry assay.b Determined by the dinitrophenylation assay.C Estimation assuming that 50% of the protein bound to the IgG Sepharosecolumn is native S. aureus protein A and that CEME is 7% of the PACEME fusion protein.d Estimation based on AU-PAGE of the sample.111A12345Ii—BFigure 27: Antibacterial activity of CEME.Various samples were electrophoresed in duplicate on a 15% acid ureapolyacrylarnide gel and either A, stained for protein with Coomassie blue, or B,tested for antibacterial activity using a gel overlay assay (Materials andMethods, section J.). Lanes: 1, 5 jig of purified, uncleaved PA/CEME; 2, 50 jigof CNBr-cleaved PA/CEME; 3, 5 jig of purified, recombinant CEME; 4, 5 jig ofchemically synthesized CEME; 5, 5 jig of melittin.112partially purified preparation of HNP- 1, as determined by comigration withpurified HNP- 1 (a gift from M. Selsted) on AU-PAGE (data not shown). Thispartially purified HNP- 1 was also tested for antimicrobial activity using the geloverlay assay which showed that the HNP- 1 preparation had no detectableantibacterial activity. There are a few possibilities as to why this peptide didnot show any antibacterial activity in this assay. First, the assay may not besensitive enough to detect HNP- 1 activity, in part due to the high ionic strengthof the buffer in which it is performed. Second, during the production of thefusion protein, proper disulfide bond formation in HNP- 1 may not haveoccurred, which would result in an inactive peptide (Selsted and Harwig, 1989).It should be noted here that no attempt was made in this study to refold thedisulfide bonds which may have been improperly formed throughout thepurification procedure.G. Summary.The protein A fusion protein system was used to produce various cationicpeptides. When E. colt was used as a host organism, both the PA/CEME fusionprotein and the protein A carrier molecule were extensively degraded. However,all fusion proteins produced in S. aureus showed little or no proteolysis, whichindicated that export to the culture supernatant was a critical step in theformation of stable fusion proteins. CEME was purified to homogeneity using acombination of affinity, gel exclusion and reverse phase chromatographicmethods. The amino acid sequence of this peptide was confirmed and itsbiological activity demonstrated using a gel overlay assay. HNP- 1 produced inthe system was partially purified, but possessed no antibacterial activity.113CHAPTER THREE: Antimicrobial Activity of Cationic Peptides.A. Introduction.The antimicrobial properties of CEME and melittin against several Gram-positive and Gram-negative bacteria have been previously documented (Wade etal., 1990). These data showed that both peptides had a broad range of activity,with CEME being more active against Gram-negative bacteria and melittinbeing more active against Gram-positive bacteria. The key difference betweenthese two is that melittin was lytic for erythrocytes which prohibited it frombeing considered as a therapeutic agent. Many attempts have been made tocreate cecropin/melittin hybrids peptides that have improved antibacterialactivity but no hemolytic activity (Boman et al., 1989a; Andreu et al., 1992;Wade et al., 1992). These studies helped to define the structural requirementsfor antibacterial activity but did not address the actual mechanism of thatactivity.This chapter focuses on the use of antibacterial activity assays as ameans for investigating the mechanism of action of cationic peptides,particularly with regards to the self-promoted uptake model. This was doneprimarily through MIC assays using various mutant strains as test organisms,and different cations as antagonists. As well, synergy studies using otherconventional antibiotics were undertaken to provide information on the mode ofthe uptake of these peptides, as well as their therapeutic potential. Thepeptides used in these studies were CEME, CEMA and melittin. CEMA was ofparticular interest because of its two extra positive charges, since it has beenshown with magainins that extending the polypeptide chain with positivelycharged amino acids could augment antibacterial activity (Bessalle et al., 1992).114B. Killing of P. aeruginosa H187 by CEME.The antimicrobial activity of CEME was initially quantified by a killingassay (Lehrer et al., 1983) in which 106 CFU of P. aeruginosa H187 wereincubated with either 2.5 jig/mL or 5.0 jig/mL of peptide in a low ionic strengthbuffer (10 mM potassium phosphate, pH 7.4). After either 20 or 60 mm,samples of the bacteria were diluted and plated to obtain a viable count. Thedata was plotted as the log of CFU/mL as a function of time (Figure 28). Theresults showed that P. aeruginosa H187 was highly susceptible to the action ofCEME, with greater than 99.9% of the bacteria killed by 2.5 ig/mL of peptide in20 mm. Under these conditions, CEME was able to kill P. aeruginosa moreeffectively (a decrease in viability of 3.2 log orders) than the reported value forrabbit lung macrophage cationic protein 1 (2.1 log order decrease; Lehrer et al.,1983). Further incubation caused little or no decrease in viability compared tothe control. This could be due to interference by bacterial components, such asLPS, released from initially lyséd cells, or it may represent a less susceptiblesubpopulatmon of cells.C. Minimum Inhibitory Concentrations.The minimum inhibitory concentrations of CEME, CEMA and melittinwere determined for a number of different organisms (Materials and Methods,section L.5.; Table IX). Generally, CEME and CEMA had similar MIC valueswhich were consistently lower than those of melittin. The exception to this wasS. aureus, which had an MIC value for CEME and CEMA that was equivalent tothat observed for melittin. This is in agreement with data from Wade et al.(1990) who demonstrated that melittin had a lethal concentration which was115Figure 28: Killing of P. aerugthosa H187 by recombinant CEME.P. aerugirtosa H187 (10 CFU/mL) was incubated with 0, 2.5, or 5.0ig/mL of CEME for 20 or 60 mm before plating the bacteria out for viabilitycounts (Materials and Methods, section L.2.). The data was plotted as the log ofCFU/mL as a function of time.116Time (mm)TableIX:MICValuesofVariousAntimicrobialAgents.BacteriumStrainRelevantPhenotypeMIC(JIg/mL)aPXGMCEFCEMECEMAMELP.aeruginosaH309Wildtype0.5112.42.88H187ParentofH1880.5124.82.88H188Antibioticsensitive0.060.250.031.21.48F.coliUB1005ParentofDC20.510.52.42.88DC2Polymyxinsensitive0.060.50.060.60.74SC9251ParentofSC92520.0620.121.21.48SC9252Polymyxinresistant440.51.21.48S.typhmuriumC587ParentofC590140.252.45.616C590Defensinsensitive0.2520.250.61.48F.cloacae218SParentof218R10.50.50.52.42.88218R1j3-lactamresistant0.50.5>162.41.48S.aureusRN4220Methicihinsensitive>8289.6>5.68SAPOO17Methicihinresistant>8>8>169.6>5.68aPX,polymyxinB;GM,gentamicin;CEF,ceftazidime;MEL,melittin.lower than CEME for Gram-positive bacteria, but higher for Gram-negativebacteria. The MICs of the peptides were usually higher than those of polymyxinB, gentamicin (an aminoglycoside) and ceftazidime (a 3-lactam), but thisdifference was deceiving due to the respective molecular weights of thesecompounds. For example, when the MICs for P. aeruginosa H309 wereconverted to jiM, the MIC values (polymyxin B, 0.3 jiM; gentamicin, 1.0 jiM;ceftazidime, 1.8 jiM; CEME, 0.9 jiM; CEMA, 1.0 jiM; melittin, 2.8 jiM) showedthat CEME and CEMA were equally, or more effective at killing this organism ona molar basis than the often utilized anti-pseudomonal antibiotics gentamicinand ceftazidime.The MIC values of these compounds against certain mutant strainsprovided initial evidence regarding the uptake pathway of cationic peptides.Both antibiotic supersusceptible mutants (P. aerugtnosa Z6 1 and E. colt DC2)were two- to four-fold more sensitive to the cationic peptides (the exceptionbeing melittin against Z61) as compared to the parental strains. This increasedsensitivity for E. colt DC2 can be explained by the nature of its mutation.Studies on the outer leaflet of DC2 showed that there was a marked decrease inthe esterification of the LPS molecules, which rendered them more negativelycharged than the parent strain (Rocque et al., 1988). This could provide thecationic peptides with an increased number of accessible sites on the LPS withwhich to interact, thus increasing the susceptibility of the organism to thesecompounds. Strain Z61 has an outer membrane alteration resulting inincreased susceptibility to most antibiotics (Angus et al., 1982). However,although analysis of the Z6 1 LPS is underway, no specific conclusions as towhy this organism is susceptible to cationic peptides could be made.The LPS of the polymyxin B resistance mutant SC9252, in contrast toDC2, was found to have increased esterification and consequently, fewer118negative charges (Peterson et al., 1987). SC9252 had increased resistance topolymyxin B which is known to interact at these sites (Schindler and Osborn,1979). In contrast, SC9252 showed no increased resistance to any of thecationic peptides, despite the hypothesis that these molecules interact withnegatively charged sites on the LPS. The discrepancy could be explained if thespecific sites of esterification in SC9252 were not the primary sites of cationicpeptide interaction. An indication of this came from the MIC values for theparental strain SC925 1. Against this strain, polymyxin B had an MIC that wasat least 20-fold lower than the cationic peptides. This suggested that thenegatively charged sites in SC925 1 that were esterified in SC9252 were targetedto a greater extent by polymyxin B than the cationic peptides. This site-specificinteraction may reflect the accessibility of the sites to the different structuresthat polymyxin B and the various cationic peptides adopt in an aqueousenvironment. The importance of the relationship between the tertiary structureof compounds and their interaction with the outer membrane has beendemonstrated by Vaara (1991) who showed that linear analogues of polymyxinB nonapeptide could not permeabiize the outer membrane.The S. typhimurtum defensin sensitive strain was an interesting organismto test, since it has a mutation in the phoP/phoQ two-component virulenceregulon (Fields et al., 1989). The PhoQ protein is a transmembrane sensorkinase that contains an anionic domain predicted to be in the periplasmicregion, and therefore it was hypothesized that in wild-type cells this domainwould interact with, and neutralize defensins (Miller et al., 1989).Characterization of phoP and phoQ mutants revealed that this hypothesis wasincorrect and that defensin sensitivity was conferred by the lack of synthesis ofPhoP-activated genes (Miller et al., 1990). The defensin sensitive mutant wastwo- to four-fold more susceptible to the cationic peptides than the wild-type119parent. Only the MIC of ceftazidime, which is taken up across the outermembrane by a porin-mediated pathway different from that of the peptides,remained unchanged. These results suggested that the mutation was onlyaffecting the self-promoted uptake pathway, possibly at the sites of initialantibiotic contact on the surface of the cell. This conclusion, however, remainsspeculative since little is known about the genes that are activated by PhoP.The peptides were tested against a clinical isolate of E. cloacae (2 18S)and its f3-lactam resistant mutant (218R1, a j3-lactamase overproducer;Marchou et al., 1987). Not surprisingly, the compounds proposed to be takenup by the self-promoted uptake pathway were equally active against both ofthese, with ceftazidime being the only antibiotic that showed a higher MIC forthe mutant. Similarly methicillin resistance in S. aureus had no apparent effecton cationic peptide susceptibifity.D. The Effect of Cations on the MIC of Cationic Peptides.One of the key proposals of the self-promoted uptake hypothesis is theinitial interaction between the cationic antibiotic and the negatively chargedsites on the surface of the outer membrane (Hancock et al., 1981). In cellsgrown under physiological conditions, these sites are occupied by divalentcations. Therefore, one can envision that, in the presence of Mg2 ions, the MICvalues of compounds proposed to be taken up by the self-promoted uptakepathway would increase due to the Mg2 ions competing for the negativelycharged binding sites. Indeed it has been demonstrated that the presence of 5mlvi MgC12 can increase the MIC of polymyxin B four-fold (Nicas and Hancock,1980).120The MIC determination was repeated for strain H309 in the presence of 5mM Mg2 or 80 mlvi Na to determine whether or not the antibacterial activity ofthe cationic peptides would be inhibited (Materials and Methods, section L.5.;Table X). The results showed that the MIC of all three peptides weredramatically increased in the presence of Mg2 but only minimally in thepresence of a higher concentration of Na. Other antibiotics were also affectedby the presence of Mg2, albeit to a much lesser extent. The differencesbetween the effects of Mg2 on the MIC of cationic peptides, and polymyxin Band gentamicin (which are all proposed to be taken up via the self-promoteduptake pathway), could reflect a difference in the nature of the initial contact ofthese compounds with the outer membrane. Nonetheless, this evidence wasconsistent with the hypothesis that the initial step in the antibacterialmechanism of cationic peptides is an association with the negatively chargedMg2 binding sites on LPS molecules.E. Synergy Studies with Cationic Peptides and Other Antibiotics.The ability of the cationic peptides to augment the activity of differentantibiotics was tested. Previous studies have shown that some cationicmembrane permeabilizing agents such as PMBN and lysine2o, at sub-MIClevels, were able to increase the sensitivity of bacteria to a number of differentantibiotics (Vaara and Vaara, 1983). Therefore, MIC assays using variousantibiotics were performed in the presence of 1/2 or 1/4 MIC levels of cationicpeptides to determine whether any synergy existed between them (Materials andMethods, section L.5.; Table XI). Generally the peptides had very little effect onthe MIC of antibiotics that are proposed to be taken up through porins(ceftazidime, imipenem, and tetracycline), although CEMA at 1/2 MIC levels did121Table X: Effects of Mg2 and Na Cations on the MICs of Cationic PeptidesAgainst P. aeruginosa H309.Compound MIC (jig/mL)No addition + 5 mM Mg2 + 80 mM NaPolymyxin 0.5 1 0.5Gentamicin 1 4 1Ceftazidime 1 2 2CEME 2.4 38.4 4.8CEMA 2.8 22.4 5.6Melittin 8 >64 16122Table XI: Effects of Sub-MIC Levels of Cationic Peptides on the MICs of CommonAntibiotics.Compound MIC (jig/mL) in the presence ofNo CEME (ig/mL) CEMA (jig/mL) Meittin (jig/mL)peptide0.6 1.2 0.7 1.4 2 4Polymyxin 1 0.5 0.25 0.5 0.12 0.5 0.06Ceftazidime 4 4 4 2 2 4 2Imipenem 4 4 4 4 2 4 4Tetracycline 8 8 8 8 4 8 8Novobiocin 256 256 256 256 128 256 256Fusidic Acid 1024 1024 1024 1024 512 1024 1024123reduce their MICs two-fold. This phenomenon was also observed for novobiocinand fusidic acid, two antibiotics that are believed to cross the membrane via thehydrophobic uptake pathway. Any compound possessing membranepermeabilizing activity at sub-lethal concentrations (which included CEME andmelittin) would be expected to enhance the uptake of these hydrophobicantibiotics by disrupting the outer membrane. This, however, was notobserved. It is possible that due to the high antimicrobial activity of CEME,concentrations of the peptide that are required to permeabilize the outermembrane to such antibiotics result in cell death and thus mask itspermeabilizing activity. This has been shown for polymyxin B, which could notenhance the uptake of novobiocin or fusidic acid (Vaara and Vaara, 1983),despite evidence that it could permeabilize outer membranes at sub-lethalconcentrations (this study; Hancock and Wong, 1984). In contrast, PMBN wasable to augment hydrophobic antibiotic activity at sub-lethal concentrationsprobably because it had such a high MIC (>100 jig/mL; Vaara and Vaara,1983). Therefore, one could test its abffity to permeabiize the outer membraneto hydrophobic antibiotics at relatively high concentrations without killing thecells.The sub-MIC levels of cationic peptides rendered strain H103 4- to 16-fold more susceptible to polymyxin B (Table Xl). Given the fact that the cationicpeptides and polymyxin B are both proposed to be taken up by the self-promoted uptake pathway, these results are consistent with additive effects ofthe membrane permeabilizing and killing activities of these two compounds.This suggestion, however, is incomplete since melittin showed the greatestinfluence on the polymyxin B MIC despite having the lowest permeabilizingactivity of the three peptides tested (see below). This raised the question ofwhich compound of the two was responsible for outer membrane124permeabiltzation and which one was responsible for killing by the disruption ofthe cytoplasmic membrane. In the cases of CEME and CEMA with polymyxinB, it was considered likely that both compounds contributed equally topermeabilization and killing, since their membrane permeabilization activities(see below) and MIC values against H309 were quite similar. However, sincemelittin showed greater synergy with polymyxin B, it was reasonable toconclude that polymyxin B, with its stronger outer membrane disrupting abffity,was permeabilizing the outer membrane for melittin which would subsequentlytarget the cytoplasmic membrane. This interpretation would imply that melittinwas more effective once it reached its target site and that the rate limiting stepin its antibacterial activity was its passage across the outer membrane.F. Summary.The antimicrobial activities of CEME, CEMA, and melittin were examinedto provide initial information regarding their mechanism of action. All thecationic peptides, especially CEME and CEMA, were shown to kill a number ofdifferent organisms at molar concentrations that were comparable to someconventional antibiotics. MIC assays performed against some mutant strainsgave initial, yet unconfirmed evidence that the cationic peptides wereinteracting with the divalent cation binding sites on LPS molecules. This wassupported by experiments that demonstrated the inhibitory effects of Mg2 onthe MIC of cationic peptides. Finally, synergy studies with the cationic peptidesand various conventional antibiotics showed that with the exception ofpolymyxin B, the peptides were unable to substantially augment antibioticactivity.125CHAPTER FOUR: Membrane Permeabilizing Activities of CationicPeptides.A. Introduction.One of the proposed requirements for compounds taken up by the self-promoted uptake pathway is they must be able to permeabilize the outermembrane (Hancock et al., 1981). The outer membrane permeabilizingactivities of many different compounds have been studied (Hancock and Wong,1984; Vaara, 1992). With respect to cationic peptides, and defensins inparticular, there has been some disagreement as to whether or not they possessthe ability to disrupt membranes. Rabbit defensins (macrophage cationicproteins; Sawyer et at, 1988) and human defensins (Lehrer et al., 1989) wereshown to permeabiize the outer membrane as measured by the uptake of NPNor the crypticity of periplasmic f3-lactamase respectively. Viljanen et at (1988)argued that sub-lethal concentrations of human defensins could notpermeabilize the outer membrane to rifampicin as measured by fractionalinhibitory concentration assays. Therefore, this chapter describes experimentsthat were performed to determine whether or not these cationic peptides havemembrane permeabilizing activity and therefore fulfill one of the requirementsof the self-promoted uptake model.B. Lysozyme Lysis Assays.Lysozyme is a 14 kDa basic protein that is unable to penetrate intactouter membranes, but can diffuse across ruptured membranes to exert its lyticactivity (Hancock and Wong, 1984). Because of its large size, one would expect126that significant destabilization of the membrane would be required in order forit to penetrate through to the peptidoglycan.The catiomc peptides were tested for their ability to facilitate the uptakeof lysozyme by permeabilizing the outer membrane (Materials and Methods,section L.2.). Briefly, the assay mixture contained prepared cells, lysozyme(which demonstrated no lytic activity by itself) and cationic peptides at variousconcentrations. Cell lysis due to the cationic peptide-enhanced uptake oflysozyme was measured as a decrease in 0D600 (Figures 29 and 30, solid lines).The data showed that CEMA was a stronger permeabiizer for lysozyme thanpolymyxin B in either P. aeruginosa H309 or E. cloacae 218R1. At lowconcentrations, CEME demonstrated better activity than polymyxin B against P.aerugtnosa H309 while with E. cloacae 2 18R1, these two compounds showedsimilar activities. The assays were also performed in the absence of lysozyme(Figures 29 and 30, dashed lines). These studies revealed that CEME andparticularly melittin (P. aerugthosa data only) possessed significant lytic activity,compared to permeabiizing activity, at the concentrations tested. In light ofthis activity, melittin did not appear to be a good permeabiizer compared to theother peptides, yet it was still 5- to 10-fold better than gentamicin (Hancock etal., 1981; data not shown). These data supported earlier suggestions in thisstudy that although melittin appeared to have difficulty penetrating themembrane barrier, it had strong lytic activity at its target site. In contrast,CEMA had little or no lytic activity at concentrations which resulted inextensive membrane permeabiization. Since lysozyme has been shown to bindLPS (Ohno and Morrison, 1989), it could be suggested that this binding mayinterfere with the interaction of the cationic peptides with LPS. If this was asignificant factor in this assay, one might expect to see antagonism rather thana synergistic effect between lysozyme and the cationic peptides. At low1270.504)4)I.’4)Figure 29: Peptide-mediated lysozyme lysis of P. aeruginosa H309.Various peptide antibiotics were tested for their ability to permeabilizethe outer membrane of P. aeruginosa H309 to lysozyme (Materials and Methods,section L.3.). Lysis was measured by a decrease in 0D600 as a function ofpeptide concentration.0.40.30.20.10.1Compound Concentration (ug/mi)1280.50.C)4)C)C)0.1Figure 30: Peptide-mediated lysozyme lysis of E. cloacae 218R1.Various peptide antibiotics were tested for their ability to permeabilizethe outer membrane of E. cloacae 218R1 to lysozyine (Materials and Methods,section L.3.). Lysis was measured by a decrease in 0D600 as a function ofpeptide concentration.0.40.30.2Compound Concentration (ug/mi)129concentrations, neither the cationic peptides nor lysozyme were able to lysecells, but when used simultaneously, lysis was observed. Therefore, it wasconcluded that lysozyme either did not bind appreciably to cellular LPS or itcould be readily displaced by the cationic peptides, which possess high LPSbinding affinities (see below). Additional experiments demonstrated that Mg2÷concentrations as low as 1 mM were able to prevent lysozyme lysis induced bythe cationic peptides. This indicated that the cationic peptides interacted withdivalent cation binding sites on the LPS to initiate outer membranepermeabiization.C. 1-N-phenylnaphthylamine Uptake Assay.1 —N-phenylnaphthylamine (NPN) is an uncharged, hydrophobicfluorescent probe that has been used to study outer membranepermeabiization (Loh et al., 1984; Sawyer et at, 1988). When NPN is mixedwith cells, it fluoresces weakly since it is unable to breach the outer membranepermeability barrier. Upon membrane destabilization, however, it can partitioninto the hydrophobic environment of the membrane where it emits a brightfluorescence. One advantage of NPN is that it is uncharged and therefore is notexpected to bind to anionic sites on LPS and interfere with the activity of thecationic peptides. In addition, its small size and hydrophobicity enable it toinsert into membranes more easily and therefore may be used to detect moresubtle disruptions of the outer membrane.NPN uptake assays were performed on P. aerugthosa H309 and .E. cloacae218R1 (Materials and Methods, section L.3.). Various concentrations of cationicpeptides were added to cuvettes containing cells and NPN. The resultingincreases in fluorescence (measured in arbitrary units) were plotted as a130function of the compound concentration (Figures 31 and 32). For P. aeruginosaH309, CEMA and polymyxin B were found to have virtually identical membranepermeabiizing activities. CEME and melittin also had similar activities, butwere weaker permeabiizers than CEMA and polymyxin B. These results (CEMApolymyxin B > CEME melittin) were different to those obtained in thelysozyme lysis assays (Figure 29; CEMA> CEME> polymyxin B > melittin).Analysis of these data revealed that similar concentrations (0.25 - 1.0 ig/mL) ofCEMA permeabilized the P. aeruginosa H309 outer membrane to lysozyme andNPN. In contrast, the concentrations of polymyxin B which were able topermeabiize cells to NPN, were not sufficient to permeabiize them to lysozyme.These data could be explained if polymyxin B caused more subtle membraneperturbations at low (< 1 pg/mL) concentrations and larger disruptions athigher (> 1 Ig/mL) concentrations. The similarities of CEMA activity in bothassays could be explained if the peptide caused substantial perturbations atboth low and high concentrations. Similar arguments could be made for CEME(cf CEMA), which showed no difference in the two assays and melittin (cfpolymyxin B), which showed much lower permeabiization concentrations in theNPN assay. Under the conditions of the NPN assay, little or no cell lysisoccurred as measured by 0D600.The results from NPN uptake experiments performed on E. cloacae2 18R1 showed that the cationic peptides were able to enhance the uptake of thehydrophobic probe at concentrations similar to those in the P. aeruginosa H309experiments. Although conclusions regarding the differences between the outerleaflets of F. cloacae 2 18R1 and P. aeruginosa H309 cannot be made, theresults do indicate that cationic peptide-induced outer membranepermeabiization is a phenomenon that is not restricted to the Pseudomonads.131C.)C.)Figure 31: Peptide-mediated NPN uptake in P. aerugthosa H309.P. aeruginosa H309 cells were incubated with NPN in the presence ofvarious concentrations of cationic peptide antibiotics (Materials and Methods,section L.4.). Enhanced uptake of NPN was measured by an increase influorescence due to the partitioning of NPN in the hydrophobic membrane. Thedata was plotted as arbitrary fluorescence units as a function of peptideconcentration.Compound Concentration (ug/mi)1320C.)4)C.)0Compound Concentration (ug/mi)Figure 32: Peptide-mediated NPN uptake in E. cloacae 218R1.E. cloacae 2 18R1 cells were incubated with NPN in the presence ofvarious concentrations of cationic peptide antibiotics (Materials and Methods,section L.4j. Enhanced uptake of NPN was measured by an increase influorescence due to the partitioning of NPN in the hydrophobic membrane. Thedata was plotted as arbitrary fluorescence units as a function of peptideconcentration.0.1 1133D. Dansyl Polymyxin B Displacement Assays.Another proposal of the self-promoted uptake model is the ability of thecompounds that access this pathway to bind to the divalent cation binding sitesof LPS. This is proposed to promote membrane destabffization, leading to theuptake of the molecule. The above experiments in this study providedsuggestive evidence that cationic peptides might bind to LPS to initiate theirantimicrobial effects. MIC assays on mutants with altered LPS anionicity(Chapter 3.B.), and Mg2 inhibition of the peptide’s antibacterial (Chapter 3.C.)and permeabiizing (Chapter 4.B.) activities were consistent with the proposalthat the sites at which the cationic peptides bind are also the ones at whichMg2 binds to form cross bridges between adjacent LPS molecules.To further investigate this question, dansyl polymyxtn B displacementassays were performed using purified LPS or whole cells. Dansyl polymyxin Bhas been shown to bind LPS, resulting in enhanced fluorescence of the dansylgroup (Moore et al., 1984). This property led to the development of an assay fordetermining the LPS-binding affinities of certain antibiotics (Materials andMethods, section L.4.c.; Moore et aL, 1986). Dansyl polymyxin B was added toa sample of LPS until approximately 90% of the binding sites were occupied asindicated by 90% of maximal fluorescence enhancement. Cationic peptides orother polycations were then titrated in and the displacement of dansylpolymyxin B monitored by the decrease in fluorescence. The results, whichwere plotted as the fraction of dansyl polymyxin B bound versus the compoundconcentration (Figure 33), indicated that all the compounds, with the exceptionof gentamicin, had high binding affinities for purified LPS. To quantify theseaffinities, the 150 value, which was the concentration of compound that resultedin 50% maximal displacement of the dansyl polymyxin B, was calculated for1340IIFigure 33: Inhibition of dansyl-polymyxin B binding to P. aerugthosa H103 LPSby various compounds.Cationic antibiotics were titrated into cuvettes containing LPS that haddansyl polymyxin B bound to it (Materials and Methods, section L.5.c.).Displacement of the dansyl polymyxin B from the LPS was measured as adecrease in fluorescence. The fraction of dansyl polymyxin B bound at a givencompound concentration was calculated as l-{(maximum fluorescence —fluorescence at compound concentration)/maximum fluorescence}. The datawas plotted as the fraction of dansyl polymyxin B bound as a function ofcompound concentration.Compound Concentration (uM)135each compound (Table XII). These values were read directly from graphs suchas the one in Figure 33. The values showed that CEMA had the highest affinityfor purified LPS, followed closely by polymyxin B, and then CEME and melittin.To test whether these compounds bind to LPS in native outermembranes, the assay was repeated using P. aerugthosa H309 whole cellsinstead of purified LPS (Figure 34). It was clear from the graph that all threecationic peptides showed a slightly higher affinity for cellular LPS thanpolymyxin B, which was confirmed by the calculated I5 values (Table XII). Incontrast only CEMA (and CEME at low concentrations) was shown to be abetter permeabiizer than polymyxin B. This indicated that the initial bindingaffinity of the compound to the LPS and the subsequent ability to permeabiizethe outer membrane are not strictly related. Nevertheless, one must becautious of any conclusions drawn from these whole cell experiments becauseof the complexity of the system. The levels of competing divalent cations, thepossible release of LPS molecules, the influence of dansyl polymyxin B whichitself perturbs the permeabffity barrier, potential conformational changes in thepeptides themselves, and the effects of the cationic peptides on membranestructure are all aspects that may affect the results. Nevertheless, the peptideswere apparently able to bind LPS in the context of whole cells.E. Summary.The membrane permeabilizing activities and LPS binding affinities ofCEME, CEMA and melittin were investigated. CEMA was found to be a potentperrneabiizer of outer membranes to both lysozyme and NPN in P. aeruginosaand F. cloacae. CEME also demonstrated strong permeabilizing activity,although not as good as CEMA. Different results obtained in the two assays136Table XII: 150 Values (in jiM) for Various Compounds Against P. aerugtrtosaLPS and Whole Cells.aCompoundPolymyxinGentanilcinMgC12 •6H20CEMECEMAMelittinP. aerugiriosaHl03 LPS0.93 ± 0.0312.17 ± 0.58850 ± 1321.30 ± 0.370.70± 0.101.41 ± 0.05P. aerugthosa H309 Whole Cells0.85 ± 0.1319.0 ± 5.2127 ± 310.41 ± 0.080.33 ± 0.030.43 ± 0.06a Each value is the average of at least three trials ± the standard deviation.137‘ 0.75CII0.25Figure 34: Inhibition of dansyl-polymyxin B binding to P. aemginosa H309whole cells by various compounds.Cationic antibiotics were titrated into cuvettes containing whole cells thathad dansyl polymyxin B bound to them (Materials and Methods, section L.5.c.).Displacement of the dansyl polymyxin B from the whole cells was measured asa decrease in fluorescence. The fraction of dansyl polymyxin B bound at agiven compound concentration was calculated as l-{(maximum fluorescence —fluorescence at compound concentration)/maximum fluorescence). The datawas plotted as the fraction of dansyl polymyxin B bound as a function ofcompound concentration.0.5Compound Concentration (uM)138suggested there may be degrees of permeabilizing activity, with polymyxin Band melittin showing minor perturbations at low concentrations and largerdisruptions at high concentrations, while CEME and CEMA showed majordisruptions even at low concentrations. Dansyl polymyxin B displacementassays with pure LPS demonstrated that the LPS binding affinities of thecompounds (CEMA> polymyxin B > CEME> melittin) correlated well with theirmembrane permeabilizing activities as determined by the NPN uptake assay.This correlation was not maintained in dansyl polymyxin B displacement assayswhich used whole cells instead of purified LPS. Nonetheless, all the cationicpeptides showed high LPS binding affinities and strong membranepermeabilizing activities, which are both required of compounds proposed to betaken up by the self-promoted uptake pathway.139DISCUSSIONA. General.This study describes the establishment of a bacterial expression systemthat allows the production of antimicrobial cationic peptides in bacteria. Anumber of different systems were investigated, including direct expression andthe expression of fusion proteins. In the latter systems, different proteolyticcleavage methods were used to remove the carrier protein. Various fusionproteins required different purification techniques, and during these studies, anovel method for the purification of inclusion bodies was developed. Two a-helical peptides, CEME and CEMA, were produced as fusion proteins to S.aureus protein A, and, after being cleaved from the affinity tag, were purified tohomogeneity. These peptides were shown to be biologically active, possessing abroad host range of antibacterial activities. The peptides, together with otherantibiotic controls, were also tested for their ability to permeabilize the outermembranes of Gram-negative bacteria. Overall, the data from the killing andmembrane permeabilization studies indicated that cationic peptides cross theouter membrane by the self-promoted uptake pathway.This work is the first report of the successful production of an a-helical,antimicrobial peptide in bacteria. During these studies, a unique method ofpurifying inclusion bodies was developed. Although many previous studieshave focused on the interaction of a-helical antimicrobial peptides with thecytoplasmic membrane, none have investigated the effects of these moleculeswith the outer membrane. The experiments in this thesis, therefore, representnew information regarding the mechanism of action of antibacterial cationicpeptides on Gram-negative bacteria.140B. Direct Expression.My first attempt at producing catiomc peptides in bacteria, was to clonethe HNP- 1 gene into the direct expression vector pT7-5. Intuitively, it did notseem feasible to produce an antibacterial peptide directly in bacteria because ofthe potential for host toxicity. It was hypothesized that by exporting the peptideto the periplasm, and thus removing the peptide from any potential cytoplasmictargets, this threat could be minimized. Therefore, the signal sequence of E. coilalkaline phosphatase was included in the genetic construct to drive this export.This hypothesis was highly speculative since the mechanism of action ofdefensins was, at the time of these experiments, largely unknown. In fact,exporting the peptide to the periplasm may enhance host toxicity sincedefensins were subsequently shown to form ion channels in planar lipidmembranes (Kagan et al., 1990). Nonetheless, the alkaline phosphatase signalsequence was included, since the export of the peptide to the periplasm hadother potential advantages such as proper formation of disulfide bonds (whichis necessary for defensin activity; Selsted and Harwig, 1989), an increasedchance of stability and solubility, and simplification of the purificationprocedure. All of these advantages proved to be irrelevant since no HNP- 1 wasdetected in any of the expression attempts using the p17-5 system.Since translation in the p97-5 system was driven by a synthetic ribosomebinding site included in the cloned fragment that encoded the alkalinephosphatase signal sequence and the HNP- 1 peptide, it was possible that it wasnot recognized by the E. coil translation machinery. Therefore the alkalinephosphatase signal sequence and the HNP- 1 gene were transferred to the p17-7vector which contained a strong ribosome binding site that was known tofunction in E. coli (Tabor and Richardson, 1985). This, however, did not result141in the production of any HNP- 1 peptide, despite the fact that a transcript thathybridized to a portion of the HNP- 1 gene was being produced. These resultscould be explained by one of two scenarios. First, the peptide was not producedat all, perhaps due to a non functional translational start site. Second, andmore plausible, the peptide was produced, but rapidly degraded by proteases inthe cytoplasm or periplasm. Indeed, only one report of successful directexpression of a disulfide-containing cationic peptide has been documented, butthe isolated peptide possessed no biological activity (Pang et al., 1992).C. Comparison of Different Fusion Protein Systems.The initial affinity tag fusion protein system that was used in this studywas that of GST. It was chosen since it provided a quick method of fusionprotein purification by affinity chromatography and since it had a specificproteolytic cleavage site that allowed for the release of the target protein (Smithand Johnson, 1988). As well, many fusion proteins produced by this systemare soluble although some do form inclusion bodies (Smith and Johnson, 1988;Hartman et al., 1992). When HNP- 1 was produced as a GST fusion protein(Figure 13), two interesting observations were made. First, most of the proteinwas found in the insoluble pellet as inclusion bodies. Since inclusion bodiesare believed to be formed by the aggregation of improper folding pathwayintermediates (Mitraki and King, 1989), it is possible that the addition of the 30amino acids of HNP- 1 to the GST carrier protein interfered with the properfolding of the protein, which resulted in aggregation. On the basis of thisargument, however, one would expect that most GST fusion proteins wouldform inclusion bodies, due to interference by the target sequence in properfolding mechanisms. This, however, is not the case. Another explanation for142the formation of these aggregates is that the cysteine residues of HNP- 1 andGST formed incorrect disulfide bonds and thus obstructed the proper foldingpathway of the fusion protein that normally leads to a soluble product. Thislatter explanation is more likely since many cysteine-containing eukaryoticpeptides form inclusion bodies in E. coli (Fischer et al., 1993). The secondobservation on the expression of GST/HNP- 1 was that the small amount ofsoluble protein detected was degraded. The presence of a protein band in thesample of soluble, purified fusion protein that comigrated with GST, indicatedthat the GST moiety retained its proteolysis-resistant nature even whenproduced as a fusion protein. This further implied that the soluble fusionprotein consisted of two distinct protein domains that were differentiallysusceptible to proteolysis. The fact that some intact fusion protein was stillpresent, suggested that the conformation adopted by the HNP- 1 moiety waspartially resistant to proteolytic degradation, or that more than oneconformation was adopted by the fusion protein.An attempt was made to increase the stability and solubility of theGST/HNP- 1 protein by inserting the prepro defensin gene cartridge at thefusion joint. The anionic prepro region is believed to stabilize the cationicdefensin peptide during its synthesis and compartmentalization in theeukaryotic host (Michailson et al., 1992). Therefore it was hypothesized thatthe presence of the prepro region in GST/proHNP- 1 might interact with andprotect the HNP- 1 peptide from proteolytic degradation. There is evidence thatthe alteration of amino acid sequences in the region linking the carrier andtarget proteins can abolish the formation of inclusion bodies (Strandberg andEnfors, 1991). It was possible, therefore, that the inclusion of the prepro regionwould also enhance the solubility of the fusion protein. The production ofGST/proHNP- 1 (Figure 19) revealed that the prepro region had no effect on the143solubiity of the fusion protein since it was still primarily found in the insolublepellet. This result showed that the structural features of HNP- 1 thatcontributed to inclusion body formation were unaffected by the preprosequence. This indicated that the positive charge density of HNP- 1, whichmight have been neutralized in GST/proHNP- 1, may not have been acontributing factor in the aggregation process. Alternatively, the prepro regionmay not have neutralized the positive charges on the HNP- 1 peptide. Incontrast to the solubility of the fusion protein, the stability of the solubleprotein was enhanced by the prepro region. This enhancement could reflect thedifferent conformation adopted by the HNP- 1 peptide in response to thepresence of the prepro region. It must be emphasized that the effects onstability applied only to the soluble protein since it appeared as though theprotein contained in inclusion bodies was completely stable.The GST system was also used to produce the x—helical peptide CEME(Figure 20). GST/CEME had different properties than GST/HNP- 1. First, itwas found almost exclusively in the soluble fraction of lysed cells. Since CEMEand HNP- 1 are similar in size and have similar charge densities, this resultprovided even more evidence that the cysteine residues played a key role in theformation of GST/HNP- 1 inclusion bodies. Second, while purified GST/HNP- 1was found to be partially degraded, GST/CEME was found to be completelydegraded. This extensive degradation may have been due to the structure ofthe CEME moiety. Since CEME has a random coil configuration in an aqueousenvironment (Wade et aL, 1990), it may be present as a peptide appendageprotruding out from the proteolysis-resistant GST carrier protein, thus makingit highly susceptible to proteolytic degradation by E. coU proteases. This is incontrast to HNP- 1, whose tight, j3-sheet structure may contribute to its partialresistance of cellular proteases.144Using the same logic as for GST/proHNP- 1, the prepro defensin genecartridge was inserted into pGEX-CEME in an attempt to stabilize the solublefusion protein. Unlike GST/proHNP- 1, the production of GST/proCEME(Figure 21) revealed that the prepro region had no effect on the stability of thefusion protein. The inability of the prepro region to protect CEME fromproteolysis could be a reflection of the ultrasensitivity of the CEME peptide tocellular proteases. Part of the problem may lie in the prepro sequence itself,since it was derived from mammalian defensins. In contrast to the pro region ofdefensins, the cecropin A pro region is quite short and is charge neutral(Gudmundsson et al., 1991). It is possible, therefore that a cecropin-likepeptide such as CEME is unable to interact with a defensin pro region in a waythat would protect it from proteolytic degradation. The other interestingobservation in the production of GST/pr0CEME was that it formed inclusionbodies. In an indirect way, this was an advantage since the inclusion bodiesprovided a source of stable GST/pr0CEME that could be purified. Nonetheless,since GST/CEME was able to fold properly into a soluble product, the preproregion must have interfered with that process in GST/pr0CEME. In contrast,the formation of GST/proHNP- 1 inclusion bodies could not be attributeddirectly to the presence of the pro region since the GST/HNP- 1 fusion proteinalready formed inclusion bodies.An interesting result was obtained when the GST fusion protein systemwas used to produce the CEMA peptide (data obtained by Dr. M. Brown in ourlaboratory). The GST/CEMA fusion protein was found to be both soluble andstable when purified on glutathione agarose beads. This is interesting whencompared to CEME which differs from CEMA by three amino acids and twopositive charges at the C-terminus. Taken by itself, this result suggests that145there are only fine lines that separate insolubility from solubility, which candetermine whether a fusion protein is stable or unstable.The other affinity tag fusion protein system that was used in this studywas that of protein A from S. aureus (Nilsson et at, 1985a). The vector chosenfor this study was pRIT5 since it could be propagated in both E. colt and S.aureus. When E. colt was used as a host to produce PA/CEME, no stableprotein was detected (Figure 24). This instability was also observed when thesame vector was used to produce human IGF-I in E. colt (Nilsson et al., 1985b;Moks et al., 1987a). Recently, shorter protein A fragments and syntheticfragments that maintained IgG binding capabilities were used as affinity tags.Fusion protein produced in these systems have been shown to leak into theexternal medium when grown in E. coU (Abrahmsén et at, 1986; Moks et al.,1987b). These vectors were not used in the present study, but were no moresuccessful when used by Dr. M. Brown in our laboratory.A number of different cationic peptides, including CEME, CEMA, andHNP- 1 were produced in S. aureus using the pRIT5 system. The peptides wereexported to the culture supernatant in a stable, soluble form (Figure 24). Somedegradation products were seen in samples of PA/CEME, but this has beenshown to occur with other fusion protein produced by this system in S. aureus(Nilsson et al., 1985a; Moks et at, 1987a).The production of cationic peptides as fusion proteins has a number ofadvantages. First, purification can be performed in a single step when thecarrier protein is an affinity tag (Sassenfeld, 1990). Second, the presence of afusion molecule can prevent the antibacterial peptide from being active againstthe host organism (as was shown for PA/CEME, Figure 27). Third, theheterologous protein can be used to elicit an antibody response without priorconjugation to a hapten (LOwenadler et at, 1987). Fourth, the fusion partner146can be manipulated to improve peptide stability, as seen above in certaininstances with the insertion of the defensin prepro region. Fifth, the peptidecan be fused to the carrier molecule in such a way that it can be released bychemical or enzymatic proteolytic cleavage without leaving any extra aminoacids on the N-terminus (see below). This is important with respect tofunctional studies using the peptide since it has been shown that the additionof one or more amino acids can alter its biological activity (Bessalle et al., 1992;Pangetal., 1992).The use of specific fusion protein systems provide additional advantages.The GST system had a high expression level of fusion protein, especially whenused in different E. coiL hosts (Table VII). The problem with this system is itsunpredictability with respect to the stability and solubility of the producedfusion protein. This is clearly unacceptable when trying to establish a generalexpression system that can be used for the production of many differentcationic peptides. The protein A system overcomes many of these problems,primarily by exporting the fusion protein out of the cytoplasm. It has generallybeen shown that protein A fusion proteins directed to the external medium ineither F. coli or S. aureus are both stable and soluble (Nilsson et al., 1985b;Abrahmsén et al., 1986; Moks et al., 1987b). This was not true for the cationicpeptide fusion proteins in this study, since only when produced in S. aureuswere the fusion proteins stable. One drawback to the protein A system is itslow expression level. This is in part due to the systems promoter which is notvery strong and cannot be induced. Nonetheless, extracellular proteinaccumulations of 75 ig/mL (Moks et al., 1987b) and 100 .tg/mL (this study)can be obtained. These expression levels can undoubtedly be improved uponby manipulating the type of promoter and growth conditions.147D. A New Strategy for Solubiizing Inclusion Bodies.When insoluble inclusion bodies are formed, there are different methodsthat can be used to solubilize them (Marston, 1986; Fischer et al., 1993).Traditionally, inclusion body purification procedures consist of a series ofsolubiization steps that will selectively solubilize the inclusion bodies whileleaving many of the contaminating proteins in the insoluble pellet. Suchextractions often employ a chaotropic agent or detergent to denature theinclusion bodies. There are several disadvantages associated with thisprocedure. First, it can be difficult to obtain a pure sample of fusion proteinsince conditions required to solubilize the inclusion bodies will also bringcontaminants into solution (Schoner et al., 1985; Cheng et aL, 1990). This isnot a problem if the solubilized inclusion bodies are to be further purified byconventional techniques (Jayaram et al., 1989; Cheng et at, 1990). It is,however, a problem when the protein is to be immediately refolded and used forfunctional studies. In addition, if the solubffized fusion protein is to be cleavedby site-specific proteases, the denaturant must be removed carefully to avoidre-precipitation of the protein (Stein, 1990; Claassen et al., 1991). Second, thesolubilization procedure is protein specific and can even vary for a singleprotein depending on the growth conditions under which the inclusion bodiesare formed (Stein, 1989). Therefore, the solubilization conditions for eachprotein must be empirically determined. Third, due to the heterogeneity of theinclusion body composition, a single solubiization condition will not necessarilydissolve all of the inclusion bodies (Schoner et al., 1985). This was the case forthe solubiization procedure of GST/HNP- 1 inclusion bodies which resulted inthe loss of significant amounts of the protein (Figure 14). Fourth, due to the148harsh conditions necessary to solubilize the inclusion bodies, the fusion proteinmay be irreversibly modified (Marston and Hartley, 1990).Theoretically, another method of purifying inclusion bodies would be tosolubilize preferentially all the contaminating proteins in the insoluble pelletprior to solubilizing the inclusion bodies. To date, there has been no report ofsuch an approach, possibly because the washing conditions used in the pasteither solubilize the inclusion bodies, or do not selectively remove thecontaminating proteins. This problem was observed with urea in thesolubiization of GST/HNP- 1. In this study, a novel solubiization scheme wasdeveloped for GST/pr0CEME inclusion bodies based on the observation that themajor contaminating proteins in inclusion bodies are outer membrane proteins(Veeraragavan, 1989). The initial washes with Triton X- 100, Tris-HC1 pH 8.0,and EDTA remained the same as those described for the GST/HNP- 1 inclusionbodies. The novel extraction used 3% 0-POE, a detergent known to solubilizeouter membrane proteins (Siehnel et al., 1992). Analysis of the supernatantsand pellets after these extractions revealed that not only were thecontaminating proteins removed completely, but the 3% 0-POE did notsolubilize any of the fusion protein (determined by Western immunoblotting,Figure 21). This procedure resulted in a preparation of GST/pr0CEME proteinthat had only minor amounts of contaminating proteins. This method could beapplied to any inclusion body preparation, regardless of how they are eventuallyto be solubilized. For those preparations of fusion proteins that are to beproteolytically cleaved by chemical methods such as CNBr, the inclusion bodiescan be solubiized directly in 70% formic acid since the preparation is relativelypure. One of the key advantages of this solubiization protocol is that it avoidsthe loss of fusion protein through premature solubifization.149E. Comparison of Fusion Protein Cleavage Methods.In many instances, a purified fusion protein must be specifically cleavedto release the target protein of interest. Two different cleavage methods wereused in the study, one enzymatic (factor Xa) and one chemical (CNBr).Solubiized GST/HNP- 1 inclusion bodies were cleaved with factor Xa torelease HNP- 1 without any extra N-terminal amino acids (Figure 15). Therelease of HNP- 1 by factor Xa at 4°C did not occur, even though factor Xa hasbeen shown to be active at this temperature (Ellinger et al., 1989). Even attemperatures as high as 37°C, the reaction required 60 h and an enzyme tosubstrate ratio of 1:25 to cleave the protein. Conditions similar to these havebeen shown to be necessary for other fusion proteins (Nambiar et al., 1987;Baldwin and Schultz, 1989), but in other cases, the conditions were muchmilder (Nagai and Thøgersen, 1984; Ellinger et al., 1989). This suggests thatthe accessibility of the cleavage site varies between fusion proteins, and thatthis may have contributed to the inefficient cleavage of GST/HNP- 1.Furthermore, it may have been the aggregation of the fusion partner and thepeptide that was preventing the factor Xa enzyme from reaching its targetsequence. In addition to the inefficiency of the reaction, some non-specificcleavage of the fusion protein occurred (Figure 16). Although the factor Xacleavage site has a consensus sequence of Ile-Glu-Gly-Arg, there have beenreports of other sequences at which cleavage can occur (Carter, 1990;Greenwood, 1993). In all of these examples, the cleavage site was on the Cterminal side of an arginine residue. The proteolytic cleavage of GST/HNP- 1resulted in three distinct peptide products. Upon further analysis, it wasdiscovered that these peptide fragments corresponded to internal fragments ofGST. Based on N-terminal amino acid sequence analysis, it was determined150that the cleavages giving rise to these fragments occurred after specific arginineresidues. The four amino acid residue sequences preceding these cleavagesites, Asp-Lys-Trp-Arg and Ala-lie-lie-Mg, had not previously been reported asfactor Xa recognition sites (Carter, 1990; Greenwood, 1993). In GST that hadbeen previously denatured and renatured, these same two sites weresusceptible to factor Xa cleavage (Figure 18), suggesting that these internal siteswere made accessible to factor Xa as a consequence of the inability of GST torefold properly under these conditions. It is interesting to note that there aremany other arginine residues in the GST amino acid sequence that were notcleaved to any large extent by factor Xa, indicating that even the non-specific”cleavage was not random. Due to the inefficient and the expensive nature ofthe factor Xa cleavage reaction, this was not the method of choice for theproduction of large quantities of cationic proteins. Furthermore, the biggestadvantage of enzymatic cleavage, namely its high specificity, could not beutffized in the purification of HNP-1 from GST/HNP-1.The gene encoding CEME was preceded directly by a methionine codonwhich enabled the release of the peptide from a fusion protein by CNBrcleavage. Biologically active CEME was released from both GST/proCEME andPA/CEME fusion proteins (Figures 22 and 27). The major disadvantage ofusing CNBr on large fusion proteins is the generation of other peptidefragments due to the presence of methionine residues in the carrier molecule.Unlike the recognition site for factor Xa, the methionine residues are rarelyinaccessible since CNBr is a small chemical molecule and the reaction is carriedout in conditions under which most proteins are denatured. This problem wasevident in both fusion protein systems since GST (Smith et al., 1986) andprotein A (Uhién et al., 1984; Nilsson et al., 1985a) both contain severalmethionine residues. The difficulty of purifying the target protein from a CNBr151digest depends on the size and p1 of the various peptide fragments produced bythe reaction. In the GST/pr0CEME digest, such peptide fragments madefurther purification of the cationic peptide impractical. In contrast, all thecontaminating peptide fragments in the PA/CEME digest could be separatedfrom CEME by gel exclusion and reverse phase chromatographies. Since thiscleavage method is both inexpensive and efficient, it can be used in a large scaleprocess, as described for PA/CEME.F. Do Cationic Peptides Cross the Outer Membrane via the Self-PromotedUptake Pathway?The self-promoted uptake model was originally proposed as a mechanismof uptake for polycationic antibiotics such as polymyxin B and gentamicin(Hancock, 1984). It has been shown that peptides of the defensin family wereable to bind LPS (Sawyer et al., 1988) and permeabilize outer membranes ofvarious bacteria (Sawyer et al., 1988; Lehrer et al., 1989), two activities that areobserved for compounds proposed to cross the outer membrane via thispathway. The hypothesis that cationic peptides are taken up via the self-promoted uptake pathway was strengthened by the fact that several differentcationic peptides such as defensins (Sawyer et al., 1988), magainins (Rana etal., 1990) and melittin (David et al., 1992) are able to interact directly with LPS.Furthermore, the ability to permeabilize outer membranes has beendemonstrated for various defensins (Sawyer et al., 1988; Lehrer et al., 1989) aswell as bactenicins (Skerlavaj et al., 1990). There have been, however, nostudies to elucidate the interaction of cecropins and melittin with the outermembrane, despite their documented Gram-negative bactericidal activity (Wadeet al., 1990). Therefore, this study set out to obtain evidence to support the152hypothesis that these cationic peptide antibiotics cross the outer membranepermeability barrier by promoting their own uptake.The self-promoted uptake of a compound is proposed to be initiated byits interaction with the negatively charged sites on LPS molecules that arenormally occupied by Mg2÷ ions which form stabilizing cross bridges betweenadjacent LPS molecules. Using dansyl polymyxin B as a probe, the LPS bindingcapabilities of CEME, CEMA and melittin were examined and compared to thoseof other cations including polymyxin B, which has been shown to bind verytightly to LPS (Moore et at, 1986). Indeed the cationic peptides all showed highbinding affinities for pure LPS (Figure 33) or LPS in the context of the whole cellenvironment (Figure 34). The calculated 150 values even showed that withwhole cells, all the cationic peptides could displace dansyl polymyxin B morereadily than polymyxin B. This, however, may not reflect the affinities of thecompounds for LPS per se, but rather the accessibility of the sites to thedifferent compounds. For instance, since the a-helical peptides consist mainlyof random coil in an aqueous environment, they may more easily penetrate aslightly perturbed membrane (due to the bound dansyl polymyxin B) thanpolymyxin B, which has a constrained, cyclic structure. Regardless, theseexperiments clearly demonstrated the ability of these peptides to bind LPS.These studies also provided evidence that the peptides were binding tothe divalent cation binding sites on the LPS. First, dansyl polymyxin B thatwas displaced from LPS by the cationic peptides could also be displaced byMg2, which was consistent with a common binding site for all of thesecompounds. Second, the presence of Mg2 in MIC assays rendered the bacteriamore resistant to the action of cationic peptides (Table X, presumably due to itssaturation of divalent cation binding sites and the resulting stabilization of theouter membrane. Third, the ability of the cationic peptides to permeabiize the153outer membrane to compounds such as lysozyme was inhibited by Mg2÷.Taken together, these data suggest that the initial interaction of cationicpeptides with Gram-negative bacteria occurs at the divalent cation binding siteson the LPS.In the self-promoted uptake model, the binding of cationic compounds tothe LPS is proposed to accompany a localized destabilization of the outermembrane. It has been shown in this study that CEME, CEMA and melittJn allhave the ability to permeabiize the outer membrane of P. aerugthosa and E.cloacae to lysozyme and NPN (Figures 29-32). In fact, under some conditionsCEMA and CEME are better permeabilizers that polymyxin B, whose potentouter membrane disrupting ability has been well documented (Vaara, 1992).This permeabilization of the outer membrane, although necessary for self-promoted uptake, is not sufficient. The molecule must be able to dissociatefrom the LPS to which it is bound and move through the disrupted outermembrane into the periplasm. This may occur as a result of changes in thestructure of LPS or the peptide itself, once the outer membrane has beendestabilized. It is very difficult to examine the nature of such post-destabilization events, but some indirect evidence is available which supportsthe hypothesis that these peptides are, in fact, taken up. First, studies haveshown that defensins (Lehrer et al., 1989) and bactenicins (Skerlavaj et al.,1990) sequentially permeabiize outer and inner membranes. Second, manypeptides have been shown to form ion channels in planar lipid membranes(Hanke et al., 1983; Christensen et al., 1988; Kagan et al., 1990) which led tothe hypothesis that the disruption of cytoplasnitc membrane integrity was thecause of cell death (Christensen et al., 1988; Lehrer et al., 1989). Third, studieson the interaction of melittin with lipid A revealed that when bound to LPS,melittin had no hemolytic activity (David et al., 1992). This suggested that154melittin must be released from the LPS before it can exert its bactericidalactivity, and therefore it must be taken up into the periplasm. These data,together with the findings that CEME, CEMA and melittin caused outermembrane permeabiization at concentrations that are not bacteriolytic, suggestthat the disruption of the outer membrane is not sufficient to kill Gram-negative bacteria, and that killing requires both uptake and further action ofthe peptides. The evidence provided in this thesis strongly implies that thesecationic peptides are taken up by self-promoted uptake.Self-promoted uptake involves certain basic steps that all compoundstaken up by this pathway are thought to go through. This thesis providesevidence that different compounds proceed through these steps in variousways. For example, CEME and CEMA showed different permeabilizingproperties when compared to polymyxin B and melittin. It seems reasonablethat a more profound disruption of the outer membrane would be required topermeabilize it to lysozyme (a 14 kDa protein) than to NPN (a small hydrophobicchemical). While all four compounds, at low concentrations, enabled NPN toenter the membrane environment (Figure 31), only CEME and CEMA couldpromote the uptake of lysozyme at those same concentrations (Figure 29).Polymyxin B and melittin, on the other hand, required higher concentrations forthis, suggesting that they could only induce major outer membrane disruptionsat high concentrations. While this may not be surprising for polymyxin B, givenits different secondary structure, it is for melittin, since it has a similarsecondary structure to CEME and CEMA. The only difference in secondarystructure that the latter two peptides may have compared to melittin, is anadditional hinge region, but this has not been confirmed and therefore remainsspeculative. It is also interesting to note that the concentrations of melittinrequired to permeabiize the outer membrane to lysozyme, were only slightly155lower than those that lysed cells (Figure 29). This implies that the rate-limitingstep in the bactericidal mechanism of melittin was its ability to cross the outermembrane permeability barrier. This has also been suggested for polymyxin B(Vaara and Vaara, 1983) as well as human (Viljanen et al., 1988) and rabbit(Sawyer et al., 1988) defensins. In contrast, CEMA was able to permeabiize themembrane to lysozyme at concentrations >10-fold lower than its MIC,demonstrating that it could breach the permeability barrier relatively easily, butcould not kill cells unless a sufficient concentration accumulated in theperiplasm. Therefore, while all these cationic compounds are proposed to betaken up by the self-promoted uptake pathway, differences exist in theirinteractions with the outer membrane.G. Cationic Peptides As Therapeutic Agents.Cationic peptides have been referred to as “natural peptide antibiotics”,reflecting their broad range of antimicrobial activity. Work in the last fewdecades has focused on isolating and characterizing dozens of different peptidesin an effort to understand how they function. As this understanding increases,a logical extension of this work is to begin to develop them as therapeuticagents that can be used to combat various infectious diseases.One property common to many of the cationic peptides that preventsthem from being used as antibiotics, is their cytolytic activity againstmammalian cells. Melittin has been shown to have potent hemolytic activity(Habermann and Jentsch, 1967; Wade et al., 1990), and defensins demonstratea broad range of cytotoxicity against mammalian cells (Lehrer et at, 1993). Inan attempt to identify structural regions within proteins that are responsible forcytolytic activity, Kini and Evans (1989) compared the sequences of over 30156cytolysins. They concluded that a cationic site flanked by a hydrophobic regionwas a common feature to all these proteins. This conclusion, however, remainstoo general to aid in the design of antibiotic peptides without cytolytic activity.More helpful have been the studies in which analogues of various peptides havebeen synthesized that maintain high antibacterial activity but do not possesshemolytic activity (Boman et al., 1989a; Wade et al., 1990; Andreu et aL, 1992;Bessalle et al., 1992; Wade et at, 1992).Another problem that these peptides may encounter is proteolyticdegradation inside the host. One potential solution to this problem is to makeD-enantiomers of the peptides which would increase their resistance to hostproteases resulting in a longer in vivo half-life. It has already been shown thatD-forms of certain antibacterial peptides are as biologically active as theirnaturally occurring L-form counterparts (Wade et al., 1990). One drawback tothis solution is that the D-forms of the peptides could not be produced in abacterial expression system. Other solutions could include the alteration offlanking amino acids, which can influence proteolysis, or the encapsultion ofthe peptide in either liposomes or polymers to protect it during delivery. Analternative solution to the problem of peptide degradation would be toconcentrate on the 13-sheet, disulfide bond-containing peptides, since theirtightly packed tertiary structure renders them quite resistant to proteolyticdegradation (Fujll et at, 1993).Another aspect that must be addressed before the peptides can be usedas therapeutic agents is their size. There are a number of reasons why it is ofinterest to develop peptides that are smaller. First, the marketing of thepeptides would be enhanced if their molecular weight was in the range of otherconventional antibiotics. For example, CEMA, with a molecular weight of 2800daltons, is almost double that of polymyxin B. Second, a smaller peptide may157have improved solubility properties, which could have ramifications for deliveryand bioavailabiity in clinical trials. Third, when creating analogues of a certainpeptide in order to improve it, the task is significantly simplified by startingwith a shorter parent peptide. Therefore, one of the challenges of designingthese compounds would be to decrease the size of the peptides withoutcompromising antimicrobial activity. Andreu et al. (1992) investigated howmany C-terminal amino acids could be deleted from CEME without affecting thebactericidal activity. They managed to bring the length of the peptide down to18 amino acids (from 26) with no significant loss of antibacterial activity orincrease in hemolytic activity and suggested that this was possible because theflexible hinge region between the two x-helices was retained. When thedeletions brought the peptide down to 15 amino acids, there was a markedincrease in the lethal concentration against certain bacteria. It was suggestedthat this hinge containing pentadecapeptide no longer had the ability to spanthe apolar portion of the membrane and form ion channels (Andreu et al.,1992). The antibacterial activity of other 15 mers that did not have the flexiblehinge region was almost as good as the 18 and 20 mers. This suggested thateither, (1) the peptide formed a 3rn CL-helix that was able to span the membrane,(2) the peptide associated as multimers that were able to span the membrane,or (3) these short peptides have a mechanism of action that has not yet beenadequately described (Andreu et al., 1992; Wade et aL, 1992). It remainsabundantly clear that reducing these peptides to a therapeutically useful sizecannot be done by indiscriminately deleting residues from an active peptidewithout considering the structural ramifications.Another focus of future research may be to design peptides withimproved specific activity (LPS binding, membrane permeabilization, channelformation), which could be used in conjunction with conventional antibiotics.158An example of this is PMBN, which is not antibacterial (MIC >1 OOjig/mL), butpossesses potent membrane permeabilizing activity that can increase thesensitivity of various Gram-negative bacteria to different antibiotics (Vaara,1992). This thesis has described a CEME variant, CEMA, which has a 2-5-foldincrease in permeabiizing activity as compared to CEME. Synergy studies withCEMA and hydrophobic antibiotics showed that the peptide was unable toenhance bacterial susceptibility to these antibiotics. This could be because theMIC of CEMA is relatively low and since concentrations required to open themembrane to hydrophobic antibiotics would kill the cell, the permeabilizingeffect would be masked. This phenomenon has previously been documented forpolymyxin B (Vaara and Vaara, 1983). It would be interesting to createshortened CEMA analogues and examine whether they can retain themembrane permeabilization activity of CEMA regardless of the effects onantibacterial activity. By doing this, it may be possible to derive a peptide thatis short enough to use as a therapeutic, and yet still be useful in conjunctionwith other antibiotics. Indeed, there is a precedent for in vitro synergy betweencationic peptides and conventional antibiotics. Darveau et aL (1991) found thatsub-inhibitory concentrations of f3-lactam antibiotics such as cefepime wereable to potentiate the activity of magainin-2 (or vica versa) against E. coil in vitroand in an in vivo mouse model. Magainins have also been shown to have asynergistic effect with erythromycin on P. aeruginosa (MacDonald et al., 1991).In light of such studies it seems very possible that the next few decades will seethe emergence of cationic peptides as boni fide therapeutic agents.159REFERENCESAbrahmsén, L., T. Moks, B. Nilsson and M. Uhlén. 1986. Secretion ofheterologous gene products to the culture medium of Escherichia coiLNuci. Acids Res. 14: 7487-7500.Amsterdam, D. 1991. Susceptibility testing of antimicrobials in liquid media.In ‘Antibiotics in Laboratory Medicine”, V. Lorian (ed.). Williams andWilkins, Baltimore, 72-78.Andersons, D., A. Engstrom, S. Josephson, L. Hansson and H. Steiner. 1991.Biologically active and amidated cecropin produced in a baculovirusexpression system from a fusion construct containing the antibody-binding part of protein A. Biochem. J. 280: 219-224.Andreu, D., J. Ubach, A. Boman, B. Wählin, D. Wade, R.B. Merrifield and H.G.Boman. 1992. Shortened cecropin A-melittin hybrids. Significant sizereduction retains potent antibiotic activity. FEBS Lett. 296: 190-194.Angus, B.L., A.M. Carey, D.A. Caron, A.M.B. Kropinski and R.E.W. Hancock.1982. Outer membrane permeability in Pseudomonas aeruginosa:Comparison of a wild type with an antibiotic supersusceptible mutant.Antimicrob. Agents Chemother. 21: 299-309.Anzai, K., M. Hamasuna, H. Kadono, S. Lee, H. Aoyagi and Y. Kirino. 1991.Formation of ion channels in planar lipid bilayer membranes bysynthetic basic peptides. Biochim. Biophys. Acta 1064: 256-266.Atkinson, T. and M. Smith. 1984. Purification of oligonucleotides obtained bysmall scale (0.2 iimol) automated synthesis by gel electrophoresis. In“Oligonucleotide Synthesis: A Practical Approach”, M.J. Gait (ed.). IRLPress, Oxford, 35-81.Ausubel, F.M., R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smithand K. Struhi. 1987. Current Protocols in Molecular Biology. GreenePublishing Associates and Wiley-Interscience, New York, New York.Bader, J. and M. Teuber. 1973. Binding action of polymyxin B on bacterialmembranes to the 0-antigenic lipopolysaccharide of Salmonellatyphimurturn. Z. Naturforsch. Teil C. 28: 422-430.Baldwin, E. and P.G. Schultz. 1989. Generation of a catalytic antibody by site-directed mutagenesis. Science 245: 1104-1107.Bell, A., M. Baths and R.E.W. Hancock. 1991. Outer membrane protein OprHof Pseudomonas aeruginosa: Expression from the cloned gene andfunction in ethylenediaminetetraacetate resistance. J. Bacteriol. 173:4970-4976.Bessalle, R., H. Haas, A. Goria, I. Shalit and M. Fridkin. 1992. Augmentation ofthe antibacterial activity of magainin by positive-charge chain extension.Antiniicrob. Agents Chemother. 36: 313-317.160Blondelle, SE. and R.A. Houghten. 1991. Hemolytic and antimicrobialactivities of the twenty-four individual omission analogues of melittin.Biochem. 30: 4671-4678.Blondelle, SE. and R.A. Houghten. 1992. Design of model amphipathicpeptides having potent antimicrobial activities. Biochem. 31: 12688-12694.Boman, H.G. and D. Huitmark. 1987. Cell-free immunity in insects. Ann. Rev.Microbiol. 41: 103-126.Boman, H.G., D. Wade, l.A. Boman, B. Wãhlin and R.B. Merrifield. 1989a.Antibacterial and antimalarial properties of peptides that are cecropinmelittin hybrids. FEBS Lett. 259: 103-106.Boman, H.C., l.A. Boman, D. Andreu, Z.Q. Li, R.B. Merrifield, G. Schlenstedtand R. Zimmermann. 1989b. Chemical synthesis and enzymaticprocessing of precursor forms of cecropins A and B. J. Biol. Chem. 264:5852-5860.Boman, H.G., I. Faye, G.H. Gudmundsson, J.-Y. Lee and D.-A. Lidhoim. 1991.Cell-free immunity in Cecropia. A model system for antibacterialproteins. Eur. J. Biochem. 201: 23-31.Boman, H.G., B. Agerberth and A. Boman. 1993. Mechanisms of action onEscherichLa coiL of cecropin P1 and PR-39, two antibacterial peptides frompig intestine. Infect. Immun. 61: 2978-2984.Borenstein, L.A., M.E. Selsted, R.I. Lehrer and J.N. Miller. 1991. Antimicrobialactivity of rabbit leukocyte defensins against Treponema pallidum subsp.pallidum. Infect. Immun. 59: 1368-1377.Bryan, L.E. 1979. Resistance to antimicrobial agents: The general nature ofthe problem and the basis of resistance. In “Pseudomonas aeruginosa:Clinical Manifestations of Infection and Current Therapy”, R.G. Doggett(ed.). Academic Press, New York, 2 19-270.Carter, P. 1990. Site-specific proteolysis of fusion proteins. In “ProteinPurification: From Molecular Mechanisms to Large-Scale Processes” ACSSymposium Series No. 427, M.R. Ladisch et al. (eds.). AmericanChemical Society, Washington, D.C., 181-193.Casteels, P., C. Ampe, F. Jacobs, M. Vaeck and P. Tempst. 1989. Apidaecins:antibacterial peptides from honeybees. EMBO J. 8: 2387-2391.Casteels, P., C. Ampe, F. Jacobs and P. Tempst. 1993. Functional andchemical characterization of hymenoptaecirt, an antibacterial polypeptidethat is infection-inducible in the honeybee (Apis mellifera). J. Biol.Chem. 268: 7044-7054.Chen, H.C., J.H. Brown, J.L. Morell and C.M. Huang. 1988. Syntheticmagainin analogues with improved antimicrobial activity. FEBS Lett.236: 462-466.161Cheng, Y.E., D. Kwoh, T.J. Kwoh, B.C. Soltvedt and D. Zipser. 1981.Stabilization of a degradable protein by its overexpression in EscherichiacoiL Gene 14: 121-130.Cheng, Y.E., M.H. McGowan, C.A. Kettner, J.V. Schloss, S. Erickson-Viitanenand F.H. Yin. 1990. High-level synthesis of recombinant HIV- 1 proteaseand the recovery of active enzyme from inclusion bodies. Gene 87: 243-248.Christensen, B., J. Fink, R.B. Merrifield and D. Mauzerall. 1988. Channel-forming properties of cecropins and related model compoundsincorporated into planar lipid membranes. Proc. Natl. Acad. Sci. USA85: 5072-5076.Chung, C.T., S.L. Niemela and R.H. Miller. 1989. One-step preparation ofcompetent Escherlchia coiL transformation and storage of bacterial cellsin the same solution. Proc. Natl. Acad. Sci. USA 86: 2172-2175.Claassen, L.A., B. Ahn, H.-S. Koo and L. Grossman. 1991. Construction ofdeletion mutants of the Escherichia coil UvrA protein and theirpurification from inclusion bodies. J. Biol. Chem. 266: 11380-11387.Compagnone-Post, P., U. Malyankar and S.A. Khan. 1991. Role of host factorsin the regulation of the enterotoxin B gene. J. Bacteriol. 173: 1827-1830.Couto, M.A., S.S.L. Harwig, J.S. Cullor, J.P. Hughes and R.I. Lehrer. 1992.Identification of eNAP-1, an antimicrobial peptide from equineneutrophils. Infect. Immun. 60: 3065-307 1.Cronan, J.E., R.B. Gennes and S.R. Maloy. 1987. Cytoplasmic membrane. In“Escherichia coil and Salmonella typhlmurium: Cellular and MolecularBiology”, F.C. Neidhardt (ed.). ASM Publications, Washington, D.C., 31-55.Daher, K.A., R.I. Lehrer, T. Ganz and M. Kronenberg. 1988. Isolation andcharacterization of human defensin cDNA clones. Proc. Nail. Acad. Sd.USA 85: 7327-7331.Dahlman, K., P.-E. Strömstedt, C. Rae, H. Jörnvall, J.-I. Flock, J. CarlstedtDuke and J.-A. Gustafsson. 1989. High level expression in Escherichiacoil of the DNA-binding domain of the glucocorticoid receptor in afunctional form utilizing domain-specific cleavage of a fusion protein. J.Biol. Chem. 264: 804-809.Darveau, R.P. and R.E.W. Hancock. 1983. Procedure for isolation of bacteriallipopolysaccharides from both smooth and rough Pseudomortasaeruginosa and Salmonella typhimurlum strains. J. Bacteriol. 155: 831-838.Darveau, R.P., M.D. Cunningham, C.L. Seachord, L. Cassiano-Clough, W.L.Cosand, J. Blake and C.S. Watkins. 1991. 13-lactam antibiotics162potentiate magainin 2 antimicrobial activity in vitro and in vivo.Antimicrob. Agents Chemother. 35: 1153-1159.David, S.A., V.1. Mathan and P. Balaram. 1992. Interaction of melittin withendotoxic lipid A. Biochim. Biophys. Acta 1123: 269-274.DeGrado, W.F., G.F. Musso, M. Lieber, E.T. Kaiser and F.J. Kezdy. 1982.Kinetics and mechanism of hemolysis induced by melittin and by asynthetic melittin analogue. Biophys. J. 37: 329-338.Dempsey, C.E. 1990. The actions of melittin on membranes. Biochim.Biophys. Acta 1031: 143-161.Deutscher, M.P. 1990. Maintaining protein stability. In “Guide to ProteinPurification”, M.P. Deutscher (ed.). Academic Press, San Diego, 83-89.Dimarcq, J.-L., E. Keppi, B. Dunbar, J. Lambert, J.-M. Reichhart, D. Hoffmann,S.M. Rankine, J.E. Fothergill and J.A. Hoffmann. 1988. Insectimmunity. Purification and characterization of a family of novel inducibleantibacterial proteins from immunized larvae of the dipteran Phormiaterrartovae and complete amino-acid sequence of the predominantmember, diptericin A. Eur. J. Biochem. 171: 17-22.Doring, G., M. Maier, E. Muller, Z. Bibi, B. TümmLer and A. Kharazmi. 1987.Virulence factors of Pseuclomonas aeruginosa. In “Basic Research andClinical Aspects of Pseudomonas aeruginosa”, G. Döring, l.A. Holder andK. Botzenhart (eds.). Karger, Basel, Switzerland, 136-148.Donaldson, L. and J.P. Capone. 1992. Purification and characterization of thecarboxyl-terminal transactivation domain of Vmw65 from herpes simplexvirus type 1. J. Biol. Chem. 267: 1411-1414.Durell, S.R., G. Raghunathan and H.R. Guy. 1992. Modeling the ion channelstructure of cecropin. Biophys. J. 63: 1623-163 1.Eisenhauer, P.B., S.S. Harwig, D. Szklarek, T. Ganz, M.E. Selsted and R.I.Lehrer. 1989. Purification and antimicrobial properties of threedefensins from rat neutrophils. Infect. Immun. 57: 2021-2027.Eisenhauer, P.B., S.S.L. Harwig and R.I. Lehrer. 1992. Cryptdins:antimicrobial defensins of the small intestine. Infect. Immun. 60: 3556-3565.Elish, M.E., J.R. Pierce and C.F. Earhart. 1988. Biochemical analysis ofspontaeous fepA mutants of Escherichia coli. J. Gen. Microbiol. 134:1355-1364.Ellinger, S., R. Glockshuber, G. Jahn and A. Plückthun. 1989. Cleavage ofprocaryotically expressed human immunodeficiency virus fusion proteinsby factor Xa and application in western blot (immunoblot) assays. J.Clin. Microbiol. 27: 97 1-976.163Fields, P.1., E.A. Groisman and F. Heffron. 1989. A Salmonella locus thatcontrols resistance to microbicidal proteins from phagocytic cells.Science 243: 1059-1062.Fink, J., A. Boman, H.G. Boman and R.B. Merrifield. 1989. Design, synthesisand antibacterial activity of cecropin-like model peptides. Internat. J.Peptide Prot. Res. 33: 412-421.Fischer, B., I. Sumner and P. Goodenough. 1993. Isolation, renaturation, andformation of disulfide bonds of eukaryotic proteins expressed inEschertchia colt as inclusion bodies. Biotech. Bioeng. 41: 3-13.Fletcher, J.E. and M.-S. Jiang. 1993. Possible mechanisms of action of cobrasnake venom cardiotoxins and bee venom melittin. Toxicon 31: 669-695.Forsgren, A. and J. Sjoquist. 1966. “Protein A” from S. aureus. Pseudo-immune reaction with human y-globulin. J. Immunol. 97: 822-827.Frank, R.W., R. Gennaro, K. Schneider, M. Przybylski and D. Romeo. 1990.Amino acid sequences of two proline-rich bactenecins. J. Biol. Chem.265: 18871-18874.Frorath, B., C.C. Abney, H. Berthold, M. Scanariril and W. Northemann. 1992.Production of recombinant rat interleukin-6 in Eschertchta colt using anovel highly efficient expression vector pGEX-3T. BioTechniques 12:558-563.Fujll, G., M.E. Selsted and D. Eisenberg. 1993. Defensins promote fusion andlysis of negatively charged membranes. Protein Science 2: 1301-1312.Ganz, T., M.E. Selsted, D. Szklarek, S.S. Harwig, K. Daher, D.F. Bainton andR.I. Lehrer. 1985. Defensins. Natural peptide antibiotics of humanneutrophils. J. Clin. Invest. 76: 1427-35.Ganz, T., J.A. Metcalf, J.I. Gallin, L.A. Boxer and R.I. Lehrer. 1988.Microbicidal/cytotoxic proteins of neutrophils are deficient in twodisorders: Chediak-Higashi syndrome and “specific” granule deficiency.J. Clin. Invest. 82: 552-556.Ganz, T., M.E. Selsted and R.I. Lehrer. 1990. Defensins. Eur. J. Haematol. 44:1-8.Gearing, D.P., N.A. Nicola, D. Metcalf, S. Foote, T.A. Willson, N.M. Gough andR.L. Williams. 1989. Production of leukemia inhibitory factor inEscherichta colt by a novel procedure and its use in maintainingembryonic stem cells in culture. Bio/Technol. 7: 1157-1161.Germino, J. and D. Bastia. 1984. Rapid purification of a cloned gene productby genetic fusion and site-specific proteolysis. Biotechnol. Lett. 10: 377-382.164Goff, S.A., L.P. Casson and A.L. Goldberg. 1984. Heat shock regulatory genehtpR influences rates of protein degradation and expression of the iongene in Escherichia coiL Proc. Nati. Acad. Sci. USA 81: 6647-6651.Greenwood, J.M. 1993. Use of the cellulose-binding domain of a cellulase fromCelluiomortas fimi for affinity purification of fusion proteins. Ph.D.,University of British Columbia.Gudmundsson, G.H., D.A. Lidhoim, B. Asling, R. Gan and H.G. Boman. 1991.The cecropin locus. Cloning and expression of a gene cluster encodingthree antibacterial peptides in Hyaiophora cecropia. J. Biol. Chem. 266:11510-11517.Habermann, E. and J. Jentsch. 1967. Sequenzanalyse des Melittins aus dentryptischen und peptischen Spaltstucken. Hoppe Seyler’s Z. Physiol.Chem. 348: 37-50.Hanahan, D. 1983. Studies on transformation of Escherlchia coil withplasmids. J. Mol. Biol. 166: 557-580.Hancock, R.E.W. 1984. Alterations in outer membrane permeability. Ann. Rev.Microbiol. 38: 237-264.Hancock, R.E.W. and A.M. Carey. 1979. Outer membrane of Pseudomonasaeruginosa. Heat- and 2-mercaptoethanol-modifiable proteins. J.Bacteriol. 140: 902-9 10.Hancock, R.E.W. and P.G.W. Wong. 1984. Compounds which increase thepermeability of the Pseudomonas aeruglnosa outer membrane.Antirnicrob. Agents Chemother. 26: 48-52.Hancock, R.E.W. and A. Bell. 1989. Antibiotic uptake into Gram-negativebacteria. In “Perspectives in Anti-Infective Therapy”, G.G. Jackson, H.D.Schiumberger and H.J. Zefler (eds.). Vieweg and Sohn, Braungshweig,21-28.Hancock, R.E.W., V.J. Raffle and T.I. Nicas. 1981. Involvement of the outermembrane in gentamicin and streptomycin uptake and killing inPseudomonas aeruglnosa. Antimicrob. Agents Chemother. 19: 777-785.Hancock, R.E.W., R. Siehnel and N. Martin. 1990. Outer membrane proteins ofPseudomonas. Mol. Microbiol. 4: 1069-1075.Hanke, W., C. Methfessel, H.-U. Wilmsen, E. Katz, G. Jung and G. Boheim.1983. Melittin and a chemically modified trichotoxin form alamethicintype multi-state pores. Biochim. Biophys. Acta 727: 108-114.Harlow, E. and D. Lane. 1988. Antibodies. A Laboratory Manual. Cold SpringHarbor Laboratory Press, Cold Spring Harbor, New York.Hartman, J., P. Daram, R.A. Frizzell, T. Rado, D.J. Benos and E.J. Sorscher.1992. Affinity purification of insoluble recombinant fusion proteinscontaining glutathione-S-transferase. Biotech. Bioeng. 39: 828-832.165Harwig, S.S., A.S. Park and R.I. Lehrer. 1992. Characterization of defensinprecursors in mature human neutrophils. Blood 79: 1532-1537.Hellers, M., H. Gunne and H. Steiner. 1991. Expression of post-translationalprocessing of preprocecropin A using a baculovirus vector. Eur. J.Biochem. 199: 435-439.Hider, R.C., F. Khader and A.S. Tatham. 1983. LyUc activity of monomeric andoligomeric melittin. Biochim. Biophys. Acta 728: 206-214.Hill, C.P., J. Yee, M.E. Seisted and D. Eisenberg. 1991. Crystal structure ofdefensin HNP-3, an amphiphilic dimer: mechanisms of membranepermeabiization. Science 251: 1481-1485.Hobot, J.A., E. Carlemaim, W. Villiger and E. Kellenberger. 1984. Periplasmicgel: new concept resulting from the reinvestigation of bacterial cellenvelope ultrastructure by new methods. J. Bacteriol. 160: 143-152.Hochuli, E., W. Bannwarth, H. Döbeli, R. Gentz and D. Stüber. 1988. Geneticapproach to facilitate purification of recombinant proteins with a novelmetal chelate adsorbent. Bio/Technol. 6: 1321-1325.Hoffmann, J.A. and C. Hetru. 1990. Insect defensins: inducible antibacterialpeptides. Immunol. Today 13: 411-415.Holak, T.A., A. Engstrom, P.J. Kraulis, G. Lindeberg, H. Bennich, T.A. Jones,A.M. Gronenborn and G.M. Clore. 1988. The solution conformation ofthe antibacterial peptide cecropin A: a nuclear magnetic resonance anddynamical simulated annealing study. Biochem. 27: 7620-7629.Hopp, T.P., K.S. Prickett, V.L. Price, R.T. Libby, C.J. March, D.P. Cerretti, D.L.Urdal and P.J. Conlon. 1988. A short polypeptide marker sequenceuseful for recombinant protein identification and purification.Bio/Technol. 6: 1204-1210.Hultmark, D. 1993. Immune reactions in Drosophila and other insects: amodel for innate immunity. Trends Genet. 9: 178-183.Huitmark, D., H. Steiner, T. Rasmuson and H.G. Boman. 1980. Insectimmunity. Purification and properties of three inducible bactericidalproteins from hemolymph of immunized pupae of Hyalophora cecropia..Eur. J. Biochem. 106: 7-16.Huitmark, D., A. Engstrom, H. Bennich, R. Kapur and H.G. Boman. 1982.Insect immunity: isolation and structure of cecropin D and four minorantibacterial components from cecropia pupae. Eur. J. Biochem. 127:207-217.Hultmark, D., A. Engstrom, K. Andersson, H. Steiner, H. Bennich and H.G.Boman. 1983. Insect immunity: Attacins, a family of antibacterialproteins from Hyalophora cecropia. EMBO J. 2: 57 1-576.166Itakura, K., T. Hirose, R. Crea and A.D. Riggs. 1977. Expression in Escherlchiacoil of a chemically synthesized gene for the hormone somatostatin.Science 198: 1056-1063.Jayaram, B., R. Devos, Y. Guisez and W. F’iers. 1989. Purification of humaninterleukin-4 produced in Escherichia coiL Gene 79: 345-354.Kagan, B.L., M.E. Seisted, T. Ganz and R.I. Lehrer. 1990. Antimicrobialdefensin peptides form voltage-dependent ion-permeable channels inplanar lipid bilayer membranes. Proc. Nati. Acad. Sci. USA 87: 210-214.Kane, J.F. and D.L. Hartley. 1988. Formation of recombinant protein inclusionbodies in Escherlchia coiL Trends Biotechnol. 6: 95-101.Karunaratne, D.N., J.C. Richards and R.E.W. Hancock. 1992. Characterizationof lipid A from Pseudomonas aeruginosa 0-antigenic B bandlipopolysaccharide by 1D and 2D NMR and mass spectral analysis.Arch. Biochem. Biophys. 299: 368-376.Kini, R.M. and H.J. Evans. 1989. A common cytolytic region in myotoxins,hemolysins, cardiotoxins and antibacterial peptides. Internat. J. PeptideProt. Res. 34: 277-286.Knott, J.A., C.A. Sullivan and A. Weston. 1988. The isolation andcharacterization of human atrial natriuretic factor produced as a fusionprotein in Escherlchia coiL Eur. J. Biochem. 174: 405-4 10.Koland, J.G., K.M. O’Brien and R.A. Cerione. 1990. Expression of epidermalgrowth factor receptor sequences as E. coil fusion proteins: applicationsin the study of tyrosine kinase function. Biochem. Biophys. Res.Commun. 166: 90-100.Kreiswirth, B.N., S. Lofdahl, M.J. Betley, M. O’Reilly, P.M. Schlievert, M.S.Bergdoll and R.P. Novick. 1983. The toxic shock syndrome exotoxinstructural gene is not detectably transmitted by a prophage. Nature305: 709-712.Kropinski, A.M., B. Jewell, J. Kuzio, F. Milazzo and D. Berry. 1985. Structureand functions of Pseudomorias aerugiriosa lipopolysaccharide. In“Pseudomonas aeruginosa: New Therapeutic Approaches from BasicResearch”, D.P. Speert and R.E.W. Hancock (eds.). Karger, Basel,Switzerland, 58-73.Lambert, J., E. Keppi, J.-L. Dimarcq, C. Wicker, J.-M. Reichhart, B. Dunbar, P.Lepage, A. Van Dorsselaer, J. Hoffmann, J. Fothergill and D. Hoffmann.1989. Insect immunity: isolation from immune blood of the dipteranPhormia terrartovae of two insect antibacterial peptides with sequencehomology to rabbit lung macrophage bactericidal peptides. Proc. Natl.Acad. Sci. USA 86: 262-266.Lee, J.—Y., A. Boman, C.X. Sun, M. Andersson, H. Jörnvall, V. Mutt and H.G.Boman. 1989. Antibacterial peptides from pig intestine: isolation of amammalian cecropin. Proc. Nati. Acad. Sci. USA 86: 9159-9162.167Lehrer, RI., M.E. Seisted, D. Szklarek and J. Fleishmann. 1983. Antibacterialactivity of microbicidal cationic proteins 1 and 2, natural peptideantibiotics of rabbit lung macrophages. Infect. Immun. 42: 10-14.Lehrer, R.I., K. Daher, T. Ganz and M.E. Selsted. 1985. Direct inactivation ofviruses by MCP-1 and MCP-2, natural peptide antibiotics from rabbitleukocytes. J. Virol. 54: 467-472.Lehrer, R.I., A. Barton, K.A. Daher, S.S. Harwig, T. Ganz and M.E. Seisted.1989. Interaction of human defensins with Escherichia coiL Mechanismof bactericidal activity. J. Clin. Invest. 84: 553-56 1.Lehrer, R.I., T. Ganz and M.E. Seisted. 1990. Defensins: natural peptideantibiotics from neutrophils. ASM News 56: 315-318.Lehrer, R.I., A.K. Lichtenstein and T. Ganz. 1993. Defensins: antimicrobialand cytotoxic peptides of mammalian cells. Ann. Rev. Immunol. 11:105-128.Leive, L. 1965. Release of lipopolysaccharide by EDTA treatment of EscherichiacoiL Biochem. Biophys. Res. Commun. 21: 290-296.Lepage, P., F. Bitsch, D. Roecklin, E. Keppi, J.-L. Dimarcq, J.-M. Reichhart, J.A.Hoffmann, C. Roitsch and D.A. Van. 1991. Determination of disulfidebridges in natural and recombinant insect defensin A. Eur. J. Biochem.196: 735-742.Lindahl, L. and J.M. Zengel. 1979. Operon-specific regulation of ribosomalprotein synthesis in Escherichia coiL Proc. Natl. Acad. Sci. USA 76:6542-6546.LOfdahl, S., B. Guss, M. Uhlén, L. Phiipson and M. Lindberg. 1983. Gene forstaphylococcal protein A. Proc. Nati. Acad. Sd. USA 80: 697-701.Loh, B., C. Grant and R.E.W. Hancock. 1984. Use of the fluorescent probe 1-N-phenylnaphthylamine to study the interactions of aminoglycosideantibiotics with the outer membrane of Pseudomortas aerugiriosa.Antimicrob. Agents Chemother. 26: 546-55 1.Löwenadler, B., B. Jansson, S. Paleus, E. Holmgren, B. Nilsson, T. Moks, G.Palm, S. Josephson, L. Phiipson and M. Uhién. 1987. A gene fusionsystem for generating antibodies against short peptides. Gene 58: 87-97.Mäkelä, P.H., M. Sarvas, S. Calcagno and K. Lounatmaa. 1978. Isolation andcharacterization of polymyxin resistant mutants of Salmonella. FEMSMicrobiol. Lett. 3: 323-326.MacDonald, D.L., C.J. Messier, W.L. Maloy and L.S. Jacob. 1991. Synergisticcombination of magainins and erythromycin on Pseudomonasaeruginosa. Program Abstr. 31st Intersci. Conf. Antimicrob. AgentsChemother., abstr. 422.168Mama, C.V., P.D. Riggs, A.G. Grandea III, B.E. Slatko, L.S. Moran, J.A.Tagliamonte, L.A. McReynolds and C. di Guan. 1988. An Escherichta coIlvector to express and purify foreign proteins by fusion to and separationfrom maltose-binding protein. Gene 74: 365-373.Marchou, B., F. Bellido, R. Charnas, C. Lucain and J.-C. Pechère. 1987.Contribution of 13-lactamase hydrolysis and outer membrane permeabilityto ceftriaxone resistance in Enterobacter cloacae. Antimicrob. AgentsChemother. 31: 1589-1595.Maroux, S., J. Baratti and P. Desnuelle. 1971: Purification and specificity ofporcine enterokinase. J. Biol. Chem. 246: 503 1-5039.Marston, F.A.O. 1986. The purification of eukaryotic polypeptides synthesizedin Escherichia colt Biochem. J. 240: 1-12.Marston, F.A.O. and D.L. Hartley. 1990. Solubiization of protein aggregates.In “Guide to Protein Purification”, M.P. Deutscher (ed.). Academic Press,San Diego, 264-276.Matsuyama, K. and S. Natori. 1988. Purification of three antibacterial proteinsfrom the culture medium of NIH-Sape-4, an embryonic cell line ofSarcophagaperegrina. J. Biol. Chem. 263: 17112-17116.Menéndez-Ariau, L., M. Young and S. Oroszlan. 1992. Purification andcharacterization of the mouse mammary tumor virus protease expressedin Escherlchia colt J. Biol. Chem. 267: 24134-24 139.Meyers, E., W.L. Parker and W.E. Brown. 1974. EM49: A new polypeptideantibiotic active against cell membranes. Ann. New York Acad. Sci. 235:493-501.Michailson, D., J. Rayner, M. Couto and T. Ganz. 1992. Cationic defensinsarise from charge-neutralized propeptides: a mechanism for avoidingleukocyte autocytotoxicity? J. Leuk. Biol. 51: 634-639.Miller, C., E. Moczydlowski, R. Latorre and M. Phffips. 1985. Charybdotoxin, aprotein thhibitor of singleCa2+activ ted K channels from mammalianskeletal muscle. Nature 313: 316-318.Miller, S.I., A.M. Kukral and J.J. Mekalanos. 1989. A two-componentregulatory system (phoP phoQ) controls Salmonella typhimuriumvirulence. Proc. Natl. Acad. Sci. USA 86: 5054-5058.Miller, S.I., W.S. Pulkkinen, M.E. Selsted and J.J. Mekalanos. 1990.Characterization of defensin resistance phenotypes associated withmutations in the phoP virulence regulon of Salmonella typhimurium.Infect. Immun. 58: 3706-37 10.Mitraki, A. and J. King. 1989. Protein folding intermediates and inclusion bodyformation. Bio/Technol. 7: 690-696.169Mizuno, T. 1979. A novel peptidoglycan-associated lipoprotein found in the cellenvelope of Pseudomonas aeruginosa and Escherlchia coil. J. Biochem.86: 991-1000.Mizuno, T. and M. Kageyama. 1979. Isolation and characterization of majorouter membrane proteins of Pseudomonas aeruginosa strain PAO withspecial reference to peptidoglycan-associated protein. J. Biochem. 86:979-989.Moks, T., L. Abrahmsén, E. Homgren, M. Bilich, A. Olsson, M. Uhlén, G. Pohi,C. Sterky, H. Hultberg, S. Josephson, A. Holmgren, H. Jornvall and B.Nilsson. 1 987a. Expression of human insulin-like growth factor I inbacteria: use of optimized gene fusion vectors to facilitate proteinpurification. Biochem. 26: 5239-5244.Moks, T., L. Abrahmsén, B. Osterlof, S. Josephson, M. Ostling, S. Enfors, I.Persson, B. Nilsson and M. Uhlén. 1987b. Large-scale affinitypurification of human insulin-like growth factor I from culture medium ofEscherlchia coiL Bio/Technol. 5: 379-382.Monaco, L., H.M. Bond, K.E. Howell and R. Cortese. 1987. A recombinantapoA- 1-protein A hybrid reproduces the binding parameters of HDL to isreceptor. EMBO J. 6: 3253-3260.Moore, R.A., L. Chan and R.E.W. Hancock. 1984. Evidence for two distinctmechanisms of resistance to poiymyxin B in Pseudomonas aeruginosa.Antimicrob. Agents Chemother. 26: 539-545.Moore, R.A., N.e. Bates and R.E.W. Hancock. 1986. Interaction of polycationicantibiotics with Pseudomortas aeruginosa lipopolysaccharide and lipid Astudied by using dansyl-polymyxin. Antimicrob. Agents Chemother. 29:496-500.Mutharia, L.M. and R.E.W. Hancock. 1983. Surface localization ofPseudomonas aeruginosa outer membrane porin protein F usingmonoclonal antibodies. Infect. Immun. 42: 1027-1033.Nagai, K. and H.C. Thøgersen. 1984. Generation of 3-globulin by sequence-specific proteolysis of a hybrid protein produced in Escherichia coli.Nature 309: 8 10-812.Nagai, K. and H.C. Thøgersen. 1987. Synthesis and sequence-specificproteolysis of hybrid proteins produced in Escherichta coil. Meth.Enzymol. 153: 461-481.Nagal, K., H.C. Thøgersen and B.F. Luisi. 1988. Refolding and crystallographicstudies of eukaryotic protein produced in Escherichia coil. Biochem. Soc.Trans. 16: 108-110.Nambiar, K.P., J. Stackhouse, S.R. Presnell and S.A. Benner. 1987. Expressionof bovine pancreatic ribonuclease A in Escherlchia coil. Eur. J. Biochem.163: 67-71.170Nicas, T.I. and R.E.W. Hancock. 1980. Outer membrane protein Hi ofPseudomorias aerugin.osa : Involvement in adaptive and mutationalresistance to ethylenediaminetetraacetate, polymyxin B and gentamicin.J. Bacteriol. 143: 872-878.Nicas, T.I. and R.E.W. Hancock. 1983a. Alteration of susceptibility to EDTA,polymyxin B and gentamicin in Pseudomonas aerugirtosa by divalentcation regulation of outer membrane protein Hi. J. Gen. Microbiol. 129:509-517.Nicas, T.I. and R.E.W. Hancock. i983b. Pseudornonas aerugthosa outermembrane permeability: isolation of a porin protein F-deficient mutant.J. Bacteriol. 153: 281-285.Nicas, T.I. and B.H. Iglewski. 1985. The contribution of exoproducts tovirulence of Pseudomonas aeruginosa.. Can. J. Microbiol. 31: 387-392.Nikaido, H. and M. Vaara. 1985. Molecular basis of bacterial outer membranepermeability. Microbiol. Rev. 49: 1-32.Nikaido, H. and R.E.W. Hancock. 1986. Outer membrane permeability ofPseudomonas aeruginosa. In “The Bacteria”, J.R. Sokatch (ed.).Academic Press, New York, 145-193.Nilsson, B. and L. Abrahmsén. 1990. Fusions to staphylococcal protein A.Meth. Enzymol. 185: 144-161.Nilsson, B., L. Abrahmsén and M. Uhlén. i985a. Immobilization andpurification of enzymes with staphylococcal protein A gene fusionvectors. EMBO J. 4: 1075-1080.Nilsson, B., E. Holmgren, S. Josephson, S. Gatenbeck, L. Philipson and M.Uhlén. 1985b. Efficient secretion and purification of human insulin-likegrowth factor I with a gene fusion vector in Staphylococci. Nucl. AcidsRes. 13: 1151-1162.Nilsson, B., T. Moks, B. Jansson, L. Abrahmsén, A. Elmblad, E. Holmgren, C.Henrichson, T.A. Jones and M. Uhlén. 1987. A synthetic IgG-bindingbased on staphylococcal protein A. Prot. Eng. 1: 107-i 13.Ohno, N. and D.C. Morrison. 1989. Lipopolysaccharide interaction withlysozyme. Binding of lipopolysaccharide to lysozyme and inhibition oflysozyme enzymatic activity. J. Biol. Chem. 264: 4434-4441.Okada, M. and S. Natori. 1983. Purification and characterization of anantibacterial protein from hemolymph of Sarcophaga peregrina (flesh-fly)larvae. Biochem. J. 211: 727-734.Oliver, D.B. 1987. Periplasm and protein secretion. In “Escherichia coil andSalmonella typhimuriunt Cellular and Molecular Biology”, F.C. Neidhardt(ed.). ASM Publications, Washington, D.C., 56-69.171Olsen, E. and S.S. Mohapatra. 1992. Expression and thrombin cleavage of PoapIX recombinant allergens fused to glutathione-S-transferase. Internat.Arch. Allergy Immunol. 98: 343-348.Ong, E., N.R. Gilkes, R.A.J. Warren, R.C. Miller and D.G. Kilburn. 1989.Enzyme immobilization using the cellulose-binding domain of aCeUulomonasflmi exoglucanase. Bio/Technol. 7: 604-607.Otoda, K., S. Kimura and Y. Imanishi. 1992. Interaction of melittin derivativeswith lipid bilayer membrane. Role of basic residues at the C-tenninaland their replacement with lactose. Biochim. Biophys. Acta 1112: 1-6.Pang, S.-Z., S.M. Oberhaus, J.L. Rasmussen, D.C. Knipple, J.R. Bloomquist,D.H. Dean, K.D. Bowman and J.C. Sanford. 1992. Expression of a geneencoding a scorpion insectotoxin peptide in yeast, bacteria and plants.Gene 116: 165-172.Panyim, S. and R. Chalkey. 1969. High resolution acrylamide gelelectrophoresis of histones. Arch. Biochem. Biophys. 130: 337-346.Pardi, A., D.R. Hare, M.E. Selsted, R.D. Morrison, D.A. Bassolino and A.2.Bach. 1988. Solution structures of the rabbit neutrophil defensin NP-S.J. Mol. Biol. 201: 625-636.Park, C.-S., S.F. Hausdorif and C. Miller. 1991. Design, synthesis, andfunctional expression of a gene for charybdotoxin, a peptide blocker of Kchannels. Proc. Nati. Acad. Sci. USA 88: 2046-2050.Park, J.T. 1987. The murein sacculus. In “Escherichia coiL and SalmonellatyphLmuriurit Cellular and Molecular Biology”, F.C. Neidhardt (ed.). ASMPublications, Washington, D.C., 23-30.Peterson, A.A., S.W. Fesik and E.J. McGroarty. 1987. Decreased binding ofantibiotics to lipopolysaccharides from polymyxin-resistant strains ofEscherichia coIl and Salmonella typhtmurium. Antimicrob. AgentsChemother. 31: 230-237.Pier, G.B., R.B. Markham and D. Eardley. 1981. Correlation of the biologicalresponses of C3H/HEJ mice to endotoxin with the chemical andstructural properties of the lipopolysaccharides from Pseudomonasaemginosa and Escherlchia colL J. Immunol. 127: 184-191.Pollack, M. 1990. Pseudomonas aeruginosa. In “Principles and Practice ofInfectious Diseases”, G.L. Mandell, R.G. Douglas and J.E. Bennett (eds.).Churchill Livingstone, New York, 1673-169 1.Rana, F.R., C.M. Sultany and J. Blazyk. 1990. Interactions between Salmonellatyphimurlum lipopolysaceharide and the antimicrobial peptide, magainin2 amide. FEBS Lett. 261: 464-467.Rao, A.G., T. Rood, J. Maddox and J. Duvick. 1992. Synthesis andcharacterization of defensin NP-i. Internat. J. Peptide Prot. Res. 40:507-514.172Redfern, C.P.F’. and K.E. Wilson. 1993. Ligand binding properties of humancellular retinoic acid binding protein II expressed in F. coil as aglutathione-S-transferase fusion protein. FEBS Lett. 321: 163-168.Reichhart, J.-M., I. Petit, M. Legrain, J.-L. Dimarcq, E. Keppi, J.-P. Lecocq, J.A.Hoffmann, and T. Achstetter. 1992. Expression and secretion in yeast ofactive insect defensin, an inducible antibacterial peptide from the fleshflyPhormia terranovae. Invert. Reproduct. Devel. 21: 15-24.Rice, W.G., T. Ganz, J.J. Kinkade, M.E. Selsted, R.I. Lehrer and R.T. ParmLey.1987. Defensin-rich dense granules of human neutrophils. Blood 70:757-765.Richmond, M.G., D.C. Clark and S. Wotton. 1976. Indirect method forassessing the penetration of beta-lactamase-nonsusceptible penicihinsand cephalosporins in Escherichia coil strains. Antimicrob. AgentsChemother. 10: 215-218.Rivera, M., L.E. Bryan, R.E.W. Hancock and E.J. McGroarty. 1988a.Heterogeneity of lipopolysaccharides from Pseudomonas aerugirtosa:analysis of lipopolysaccharide chain length. J. Bacteriol. 170: 512-521.Rivera, M., R.E.W. Hancock, J.G. Sawyer, A. Haug and E.J. McGroarty. 1988b.Enhanced binding of polycationic antibiotics to lipopolysaccharide froman aminoglycoside-supersusceptible, tolA mutant strain of Pseudomon.asaeruginosa. Antimicrob. Agents Chemother. 32: 649-655.Rocque, W.J., S.W. Fesik, A. Haug and E.J. McGroarty. 1988. Polycationbinding to isolated hipopolysaccharide from antibiotic-hypersusceptiblemutant strains of Escherlchia coiL Antimicrob. Agents Chemother. 32:308-313.Rowe, P., 5, N. and P.M. Meadow. 1983. Structure of the core from thelipopolysaccharide of Pseudomon.as aerugirlosa PACIR and its defectivemutants. Eur. J. Biochem. 132: 329-337.Russel, M. and P. Model. 1984. Replacement of the ftp gene of Escherlchia coilby an inactive gene cloned on a plasmid. J. Bacteriol. 159: 1034-1039.Sambrook, J., E.F. Fritisch and T. Maniatis. 1989. Molecular Cloning. ALaboratory Manual. Cold Spring Harbor Laboratory Press, Cold SpringHarbor, New York.Sandermann and Strominger. 1972. Purification and properties of C55-Isoprenoid alcohol phosphokinase from Staphylococcus aureus. J. Biol.Chem. 247: 5123-5131.Sassenfeld, H.M. 1984. A polypeptide fusion designed for the purification ofrecombinant proteins. Bio/Technol. 2: 76-81.Sassenfeld, H.M. 1990. Engineering proteins for purification. TrendsBiotechnol. 8: 88-93.173Sawyer, J.G., N.L. Martin and R.E.W. Hancock. 1988. The interaction ofmacrophage cationic proteins with the outer membrane of Pseudomonasaeruginosa. Infect. Immun. 56: 693-698.Saya, H., P.S.Y. Lee, T. Nishi, I. Izawa, M. Nakajima, G.E. Gallick and V.A.Levin. 1993. Bacterial expression of an active tyrosine kinase from aprotein A/truncated c-src fusion protein. FEBS Lett. 327: 224-230.Stein, C.H. 1989. Production of soluble recombinant proteins in bacteria.Bio/Technol. 7: 1141-1149.Stein, C.H. 1990. Solubility as a function of protein structure and solventcomponents. Bio/Technol. 8: 308-317.Stein, C.H. and M.H.M. Noteborn. 1988. Formation of soluble recombinantproteins in Escherichia colt is favored by lower growth temperature.Bio/Technol. 6: 29 1-294.Schindler, M. and M.J. Osborn. 1979. Interaction of divalent cations andpolymyxin B with lipopolysaccharide. Biochem. 18: 4425-4430.Schindler, P.R.G. and M. Teuber. 1975. Action of polymyxin B on bacterialmembranes: morphological changes in the cytoplasm and in the outermembrane of Salmonella typhtmurium and Eschertchta colt B.Antimicrob. Agents Chemother. 8: 94-104.Schoner, R.G., L.F. Ellis and B.E. Schoner. 1985. Isolation and purification ofprotein granules from Escherichta colt cells overproducing bovine growthhormone. Bio /Technol. 2: 151-154.Sekar, V. 1987. A rapid screening procedure for the identification ofrecombinant bacterial clones. 5: 11-13.Seisted, M.E. and S.S. Harwig. 1987. Purification, primary structure, andantimicrobial activities of a guinea pig neutrophil defensin. Infect.Immun. 55: 2281-2286.Seisted, M.E. and S.S. Harwig. 1989. Determination of the disulfide array inthe human defensin HNP-2. A covalently cydlized peptide. J. Biol. Chem.264: 4003-4007.Selsted, M.E., D. Szklarek and R.I. Lehrer. 1984. Purification and antibacterialactivity of antimicrobial peptides of rabbit granulocytes. Infect. Immun.45: 150-154.Selsted, M.E., D. Szklarek, T. Ganz and R.I. Lehrer. 1985. Activity of rabbitleukocyte peptides against Candtda albtcans. Infect. Immun. 49: 202-206.Selsted, M.E., M.J. Novotny, W.L. Morris, Y.-Q. Tang, W. Smith and J.S. Cullor.1992. Indolicidin, a novel bactericidal tridecapeptide amide fromneutrophils. J. Biol. Chem. 267: 4292-4295.174Seisted, M.E., Y.-Q. Tang, W.L. Morris, P.A. McGuire, M.J. Novotny, W. Smith,A.H. Henschen and J.S. Cullor. 1993. Purification, primary structures,and antibacterial activities of (3-defensins, a new family of antimicrobialpeptides from bovine neutrophils. J. Biol. Chem. 268: 6641-6648.Shine, J., I. Fettes, N.C.Y. Lan, J.L. Roberts and J.D. Baxter. 1980. Expressionof cloned f3-endorphin gene sequences by Escherlchia coiL Nature 285:456-461.Siehnel, R.J., C. Egli and R.E.W. Hancock. 1992. Polyphosphate-selective porinOprO of Pseudomonas aeruglnosa: expression, purification andsequence. Mol. Microbiol. 6: 2319-2326.Sipos, D., K. Chandrasekhar, K. Arvidsson, A. Engstrom and A. Ehrenberg.1991. Two-dimensional proton-NMR studies on a hybrid peptidebetween cecropin A and melittin. Resonance assignments and secondarystructure. Eur. J. Biochem. 199: 285-291.Skerlavaj, B., D. Romeo and R. Gennaro. 1990. Rapid membranepermeabilization and inhibition of vital functions of Gram-negativebacteria by bactenecins. Infect. Immun. 58: 3724-3730.Smit, J., Y. Kamio and H. Nikaido. 1975. Outer membrane of Salmonellatyphimurium: chemical analysis and freeze-fracture studies withlipopolysaccharide mutants. J. Bacteriol. 124: 942-958.Smith, D.B. and K.S. Johnson. 1988. Single-step purification of polypeptidesexpressed in Escherichia colt as fusions with glutathione-S-transferase.Gene 67: 3 1-40.Smith, D.B., K.M. Davern, P.G. Board, W.U. Tiu, E.G. Garcia and G.F. Mitchell.1986. Mr 26,000 antigen of Schistosoma japonlcum recognized byresistant WEHI 1 29/J mice is a parasite glutathione-S-transferase. Proc.Natl. Acad. Sd. USA 83: 8703-8707.Smith, D.B., M.R. Rubira, R.J. Simpson, K.M. Davem, W.U. Tiu, P.G. Boardand G.F. Mitchell. 1988. Expression of an enzymatically active parasitemolecule in Escherichia colt: Schistosoma japonicum glutathione-Stransferase. Mol. Biochem. Parasitol. 27: 249-256.Stahl, S., A. Sjölander, M. Hansson, P.-A. Nygren and M. Uhlén. 1990. Ageneral strategy for polymerization, assembly and expression of epitopecarrying peptides applied to the Plasmodium falciparum antigenPf155/RESA. Gene 89: 187-193.Stark, G.R. 1965. Reactions of cyanate with functional groups of proteins. III.Reactions with amino and carboxyl groups. Biochem. 4: 1030-1036.Steiner, H. 1982. Secondary structure of the cecropins: antibacterial peptidesfrom the moth Hyalophora cecropia. FEBS Lett. 137: 283-287.175Steiner, H., D. Hultmark, A. Engstrom, H. Bennich and H.G. Boman. 1981.Sequence and specificity of two antibacterial proteins involved in insectimmunity. Nature 292: 246-248.Steiner, H., D. Andreu and R.B. Merrifield. 1988. Binding and action ofcecropin and cecropin analogues: antibacterial peptides from insects.Biochim. Biophys. Acta 939: 260-266.Stirling, D.A., A. Petrie, D.J. Pulford, D.T.W. Paterson and M.J.R. Stark. 1992.Protein A-calmodulin fusions: a novel approach for investigatingcalmodulin function in yeast. Mol. Microbiol. 6: 703-7 13.Strandberg, L. and S.-L. Enfors. 1991. Factors influencing inclusion bodyformation in the production of a fused protein in Escherichia coil. Appi.Environ. Microbiol. 57: 1669-1674.Stringer, K.F., C.J. Ingles and J. Greenblatt. 1990. Direct and selective bindingof an acidic transcriptional activation domain to the TATA-box factorTFIID. Nature 345: 783-786.Studier, F.W. and B.A. Moffatt. 1986. Use of bacteriophage 17 RNA polymeraseto direct selective high-level expression of cloned genes. J. Mol. Biol.189: 113-130.Swamy, K.H.S. and A.L. Goldberg. 1981. E. coil contains eight solubleproteolytic activities, one being ATP dependent. Nature 292: 652-654.Szoka, P.R., A.B. Schrieber, H. Chan and J. Murthy. 1986. A general methodfor retrieving the components of a genetically engineered fusion protein.DNA5: 11-20.Tabor, S. and C.C. Richardson. 1985. A bacteriophage T7 RNApolymerase/promoter system for controlled exclusive expression ofspecific genes. Proc. Nail. Acad. Sci. USA 82: 1074-1078.Teuber, M. and J. Bader. 1976. Action of polymyxin on bacterial membranes.Binding capacities for polymyxin B of inner and outer membranesisolated from Salmonella typhimurlum G30. Arch. Microbiol. 109: 51-58.Theil, R. and K.H. Scheit. 1983. Amino acid sequence of seminaiplasmin, andantimicrobial protein from bull semen. EMBO J. 2: 1159-1163.Tosteson, M.T. and D.C. Tosteson. 1981. The sting. Melittin forms channels inlipid bilayers. Biophys. J. 36: 109-116.Trias, J. and H. Nikaido. 1990a. Outer membrane protein D2 catalyzesfacilitated diffusion of carbapenems and penems through the outermembrane of Pseudomonas aerugirtosa. Antimicrob. Agents Chemother.34: 52-57.Trias, J. and H. Nikaido. 1990b. Protein D2 channel of the Pseudomonasaeruginosa outer membrane has a binding site for basic amino acids andpeptides. J. Biol. Chem. 265: 15680-15684.176Uhlén, M. and T. Moks. 1990. Gene fusions for purpose of expression: anintroduction. Meth. Enzymol. 185: 129-143.Uhlén, M., B. Nilsson, B. Guss, M. Lindberg, S. Gatenbeck and L. Philipson.1983. Gene fusion vectors based on the gene for staphylococcal proteinA. Gene 23: 369-378.Uhlén, M., B. Guss, B. Nilsson, S. Gatenbeck, L. Philipson and M. Lindberg.1984. Complete sequence of the staphylococcal gene encoding protein A.A gene evolved through multiple duplications. J. Biol. Chem. 259:1695-1702.Ullman, A. 1984. One-step purification of hybrid proteins which have Igalactosidase activity. Gene 29: 27-31.Vaara, M. 1991. The outer membrane permeability-increasing action of linearanalogues of polymyxin B nonapeptide. Drugs Exp. Clin. Res. 17: 437-444.Vaara, M. 1992. Agents that increase the permeability of the outer membrane.Microbiol. Rev. 56: 395-411.Vaara, M. and T. Vaara. 1983. Polycations sensitize enteric bacteria toantibiotics. Antimicrob. Agents Chemother. 24: 107-113.Vaara, M., T. Vaara, M. Jensen, I. Helander, M. Nurminen, E.T. Rietschel andP.H. Mäkelä. 1981. Characterization of the lipopolysaccharide from thepolymyxin-resistant mutants of Salmonella typhimurtum. FEBS Lett.129: 145-149.Veeraragavan, K. 1989. Studies on two major contaminating proteins of thecytoplasmic inclusion bodies in Escherichia coiL FEMS Microbiol. Lett.61: 149-152.Viljanen, P., P. Koski and M. Vaara. 1988. Effect of small cationic leukocytepeptides (defensins) on the permeability barrier of the outer membrane.Infect. Immun. 56: 2324-2329.Vogel, H. and F. Jahnig. 1986. The structure of melittin in membranes.Biophys. J. 50: 573-582.von Gabain, A., J.G. Belasco, J.L. Schottel, A.C.Y. Chang and S.N. Cohen.1983. Decay of mRNA in Escherichia coiL: investigation of the fate ofspecific segments of transcripts. Proc. Nati. Acad. Sci. USA 80: 653-657.Wade, D., A. Boman, B. Wahlin, C.M. Drain, D. Andreu, H.G. Boman and R.B.Merrifield. 1990. AU-D amino acid-containing channel-forming antibioticpeptides. Proc. Nail. Acad. Sci. USA 87: 476 1-4765.Wade, D., D. Andreu, S.A. Mitchell, A.M.V. Silveira, A. Boman, H.G. Boman andR.B. Merrifield. 1992. Antibacterial peptides designed as analogs or177hybrids of cecropins and melittin. Internat. J. Peptide Prot. Res. 40:429-436.Werkmeister, J.A., A. Kirkpatrick, J.A. McKenzie and D.E. Rivett. 1993. Theeffect of sequence variations and structure on the cytolytic activity ofmelittin peptides. Biochim. Biophys. Acta 1157: 50-54.Wilkinson, S.G. 1983. Composition and structure of lipopolysaccharides fromPseudomona.s aerugthosa. Rev. Infect. Dis. 5: S941-S949.Winberg, G. and M.-L. Hammarskjord. 1980. Isolation of DNA from agarosegels using DEAE-paper. Application to restriction site mapping ofadenovirus type 16 DNA. Nuci. Acids Res. 8: 253-264.Xanthopoulos, K.G., J.Y. Lee, R. Gan, K. Kockum, I. Faye and H.G. Boman.1988. The structure of the gene for cecropin B, an antibacterial immuneprotein from Hyaiophora cecropia. Eur. J. Biochem. 172: 371-376.Yoshimura, F. and H. Nikaldo. 1982. Permeability of Pseudomonas aeruginosaouter membrane to hydrophiic solutes. J. Bacteriol. 152: 636-642.Young, L.S. 1984. The clinical challenge of infections due to Pseudomonasaemginosa. Rev. Infect. Dis. 6: S603-S607.Yu, A., E.V. Knirel, N.A. Kocharova, N.A. Paramonov, N.K. Kochetkov, B.A.Dmitriev, E.S. Stanislavsky and B. Lanyi. 1988. The structure of 0-specific polysaccharides and serological classification of Pseudomonasaerugirtosa. Acta Microbiol. Hung. 35: 3-24.Zasloff, M. 1987. Magainins, a class of antimicrobial peptides from Xenopusskin: isolation, characterization of two active forms and partial eDNAsequence of a precursor. Proc. NatL Acad. Sci. USA 84: 5449-5453.Zeya, H.I. and J.K. Spitznagel. 1966. Cationic proteins of polymorphonuclearleukocyte lysosomes. I. Resolution of antibacterial and enzymaticactivities. J. Bacteriol. 91: 750-754.Zimmermann, W. and A. Rosselet. 1977. Function of the outer membrane ofEscherlchia coil as a permeability barrier to 13-lactam antibiotics.Antimicrob. Agents Chemother. 12: 368-372.178

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0099224/manifest

Comment

Related Items