Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Structure-function studies of the Pseudomonas aeruginosa porin Oprp Sukhan, Anand 1996

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

Item Metadata

Download

Media
831-ubc_1997-196585.pdf [ 6.96MB ]
Metadata
JSON: 831-1.0088292.json
JSON-LD: 831-1.0088292-ld.json
RDF/XML (Pretty): 831-1.0088292-rdf.xml
RDF/JSON: 831-1.0088292-rdf.json
Turtle: 831-1.0088292-turtle.txt
N-Triples: 831-1.0088292-rdf-ntriples.txt
Original Record: 831-1.0088292-source.json
Full Text
831-1.0088292-fulltext.txt
Citation
831-1.0088292.ris

Full Text

STRUCTURE-FUNCTION STUDIES OF THE PSEUDOMONAS AERUGINOSA PORTN O P R P by A N A N D S U K H A N B.Sc, University of Manitoba, 1985 M.Sc., University of Manitoba, 1989 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Microbiology and Immunology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A December 1996 © Anand Sukhan f 1 9 9 6 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department Date r DE-6 (2/88) i i ABSTRACT The gene encoding the Pseudomonas aeruginosa phosphate-specific porin OprP was subjected to both linker and epitope insertion mutagenesis. Nine of the 13 linker mutants expressed protein at levels comparable to the wild type gene and were shown to be properly exposed at the cell surface. Four of the linker mutants expressed protein at reduced levels which were not detectable at the cell surface. A foreign epitope from the circumsporozoite form of the malarial parasite Plasmodium falciparum was cloned into the linker sites of 12 of the 13 mutants. Seven of the resultant epitope insertion mutants expressed surface-exposed protein. Two of these mutants presented the foreign epitope at surface-accessible regions as assessed using a malarial epitope-specific monoclonal antibody. The data from these experiments were used to create a topological model of the OprP monomer which was tested and supported by further mutagenic procedures. In order to determine the function that individual lysine residues serve in the transport of anions through the channels of this protein, the first nine amino-terminal lysine residues of OprP were substituted with glutamates. The mutant proteins were purified and the channels they formed were characterized by reconstituting the purified porins in planar lipid bilayer 74 121 126 membranes. In comparison to the wild-type protein, the Lys , Lys , and Lys mutants all 74 displayed reduced levels of conductance at KC1 concentrations below 1 M and the Lys and 121 Lys mutants no longer exhibited a saturation of conductance at high anion concentrations. In addition, the ability of phosphate ions to inhibit the conductance of CI - ions through the 121 channels formed by the Lys mutant was greatly reduced, while their ability to inhibit the CI iii 74 conductance of the Lys mutant was reduced by approximately 2-fold. To clarify the roles that 74 121 126 Lys , Lys , and Lys play in regulating the channel characteristics of OprP, these amino acids were replaced with either glycine or glutamine residues. Analysis of these mutants revealed that 74 126 both Lys and Lys may serve to funnel anions toward the binding site, but only the presence 121 of Lys is required tor the formation of the inorganic phosphate-specific binding site of OprP. iv TABLE OF CONTENTS ABSTRACT i i T A B L E OF CONTENTS iv LIST OF FIGURES ix LIST OF TABLES xi ACKNOWLEDGEMENTS xii INTRODUCTION 1 Pseudomonas aeruginosa 1 The cell wall of gram negative bacteria 4 The inner membrane 4 The periplasmic space 5 The outer membrane 6 Porins of gram negative bacteria: Functional aspects 7 Porins of gram negative bacteria: Structural aspects 11 Phosphate transport across the outer membrane of gram negative bacteria 16 Aims of this thesis 18 MATERIALS AND METHODS 19 Bacterial strains and plasmids 19 Media and growth conditions 19 Chemicals 22 General recombinant D N A techniques 23 V 1. Enzymes and plasmid manipulation 23 2. Agarose gel electrophoresis 23 3. Synthetic oligonucleotides 24 4. D N A sequencing 24 General protein and immunological techniques 25 1. SDS-polyacrylamide gel electrophoresis 25 2. Western immunoblotting 25 3. Antibodies 26 4. Indirect immunofluorescence 27 Construction of plasmid pAS27 28 Linker insertion mutagenesis 31 Malarial epitope insertion 31 Site-directed lysine mutagenesis 32 Isolation of outer membrane fractions 34 Detergent solubilization of OprP 34 FPLC purification of OprP 35 Purification of OprP from SDS-polyacrylamide gels 36 Planar lipid bilayer experiments 37 Single-channel conductance 37 Anion selectivity 37 Phosphate ion specificity 38 RESULTS 39 vi A. Structural analysis of OprP 1. Linker insertion mutagenesis 39 1.1 Introduction '. '. 39 1.2 Expression of OprP in an E. coli background 39 1.3 Linker insertion mutagenesis of OprP at a unique Eco RV restriction enzyme site 40 1.4 Semi-random linker mutagenesis of OprP 40 1.5 Expression of the linker insertion mutant forms of OprP in E. coli 42 1.6 Surface localization of the OprP linker insertion mutants 45 1.7 Functional analysis of selected linker insertion mutants. 46 1.8 Summary 49 2. Epitope insertion mutagenesis of OprP 52 2.1 Introduction 52 2.2 Insertion of the malarial epitope into the linker sites of OprP mutants 52 2.3 Expression of the malarial epitope insertion mutant proteins 53 2.4 Surface localization of the epitope insertion mutant proteins 57 2.5 Surface localization of the inserted malarial epitopes 59 2.6 Summary 61 3. Modeling of the OprP monomer 63 3.1 Introduction 63 3.2 Structural implications of linker insertion mutagenesis 63 3.3 Structural implications of epitope insertion mutagenesis 65 vii 3.4 Testing of the proposed OprP topological model i) Deletion mutagenesis of the proposed fifth surface-exposed loop 66 3.5 Testing of the proposed OprP topological model ii) Epitope insertion mutagenesis of the proposed seventh surface-exposed loop 69 3.6 Summary 69 B. Functional analysis of OprP 4. Site-directed mutagenesis of specific lysine residues of OprP 72 4 ! Introduction 72 4.2 Site-directed mutagenesis of lysine residues 73 4.3 Expression of amino-terminal lysine substitution mutant proteins 74 4.4 Single-channel conductance of amino-terminal lysine substitution mutant proteins 74 4.5 Anion selectivity of amino-terminal lysine substitution mutant proteins... 79 4.6 Phosphate specificity of amino-terminal lysine substitution mutants 81 4.7 Channel characteristics of L y s ^ G l y , Lys->Gln and Lys 7 4 ' 1 2 1 ' 1 2 6 ->Glu mutant proteins 85 4.8 Expression and membrane localization of carboxy-terminal lysine substitution mutant proteins 85 4.9 Summary 87 DISCUSSION 90 Topological model of OprP 91 The role of specific lysine residues in defining the channel characteristics of OprP 99 Summary REFERENCES ix LIST OF FIGURES Figure 1. Procedure used in the creation of plasmid pAS27 29 Figure 2. Restriction map of plasmid pAS27 30 Figure 3. Western immunoblot of outer membranes of E. coli CE1248 containing OprP 41 Figure 4. Western immunoblot of outer membranes of E. coli CE1248 containing OprP linker insertion plasmids pPLl-pPL13 44 Figure 5. Indirect immunofluorescence of E. coli CE1248 containing plasmid pAS27 or plasmid pTZ19U 47 Figure 6. Western immunoblot of E. coli CE1248 containing OprP-encoding plasmids with the malarial epitope inserted at sites PL1, PL2 and PL4 to PL13 54 Figure 7. Western immunoblot of E. coli CE1248 containing OprP-encoding plasmids with the malarial epitope inserted at sites PL1, PL2 and PL4 to PL13 55 Figure 8. Indirect immunofluorescence oiE. coli strain CE1248 containing the OprP-encoding plasmid with the malarial epitope inserted at site PL10 60 Figure 9. Topological model of the monomelic structure of OprP based on linker and epitope insertion 64 Figure 10. Whole cell Western immunoblot of E. coli CE1248 expressing the OprP mutants deleted for sections of the proposed fifth surface-exposed loop 68 Figure 11. Indirect immunofluorescence of E. coli CE1248 containing the OprP-encoding plasmid with the malarial epitope inserted in the proposed seventh surface-exposed loop 70 Figure 12. Topological model of the monomeric structure of OprP 71 Figure 13. Whole cell lysate Western immunoblot of outer membranes of E. coli CE1248 expressing the OprP substitution mutant proteins 75 121 Figure 14. Single-channel events of wild-type OprP and Lys —>Glu mutant porin 77 Figure 15. KC1 concentration-conductance relationship of selected OprP Lys—>Glu mutants 78 Figure 16. Phosphate-induced inhibition of chloride conductance 83 Figure 17. Phosphate-induced inhibition of chloride conductance 84 Figure 18. Whole ceil Western immunoblot of E. coli CE1248 expressing the OprP substitution mutant proteins Lys361—>Glu and Lys385—> Glu 88 Figure 19. Topological model of the monomeric structure of OprP 89 Figure 20. Topological model of the monomeric structure of OprP 106 Figure 21. Schematic diagram of the p-barrel structure of the E. coli porin OmpF 117 Figure 22. Conceptual model for the structure of OprP 118 Figure 23. Hydropathy plots for OprP 119 LIST OF TABLES xi Table I. Strains and plasmids used in this study 20-21 Table IL Codon changes of OprP substitution mutants 33 Table III. Insertion sites of OprP linker mutants 43 Table IV. Expression and surface exposure of OprP linker mutant proteins 48 Table V. Channel activity of selected OprP linker mutant proteins 50 Table VI. Expression and surface exposure of OprP malarial epitope insertion mutant proteins 58 Table VII. Single-channel conductance of OprP Lys—>Glu mutant proteins 76 Table VIII. Anion selectivity of OprP Lys—>Glu substitution mutant proteins 80 Table LX. Phosphate inhibition of chloride conductance through the OprP Lys—>Glu substitution mutant proteins 82 Table X. Channel characteristics of Lys->Gly and Lys->Gln mutant proteins 86 xii ACKNOWLEDGEMENTS I would like thank my research supervisor Robert E. W. Hancock for taking me into his lab and giving me the benefit of his experience. I would also like to thank all the members of the Hancock lab, both past and present, who have help me along the way. The assistance of the following members of the Hancock lab was indispensible to the completion of this thesis; Manjeet Bains, Susan W. Farmer, Eileen G. Rawling, Richard J. Siehnel and Rebecca S.-Y. Wong. In addition; I would also like to thank the members of my advisory committee as well as various members of the Department of Microbiology and Immunology for their helpful advice. The financial support of the Medical Research Council of Canada is gratefully acknowledged. Finally, I would like to thank my parents whose unfailing support throughout the past five and a half years made this experience far less painful than it otherwise might have been. 1 INTRODUCTION Pseudomonas aeruginosa The gram negative bacterium Pseudomonas aeruginosa is a member of a large group of rod-shaped aerobic microorganisms (Palleroni, 1986). The cell wal l of this organism is typical of that of most gram negative bacteria, being composed of two l ipid bilayers separated by a thin periplasmic space. The external surface of the cell wal l is covered with a number of pi l i and also contains a single polarly located flagellum. This latter structure has been shown to provide the cells with a means of locomotion. Members of the Pseudomonas genus are often identified by the characteristic pigments which they produce. In the case of P. aeruginosa, cultures of the organism release the fluorescent compound pyoverdine into the growth media. This molecule, along wi th the nonfluorescent compound pyochelin, acts as an iron chelating scavenger (Cox, 1980; Cox and Adams, 1985). The growth requirements of P. aeruginosa are quite minimal. This allows the organism to occupy a wide variety of different environments; some of which would be rather inhospitable towards other more fastidious bacteria. Various species of Pseudomonas have been isolated from samples of soil and lake water (Palleroni, 1986), and there is evidence that they may inhabit seawater as wel l (Tanoue et a l , 1995). Members of this genus of bacteria have also been found to inhabit the areas surrounding the roots of certain plants (Scher et a l , 1988). While the prevalence of this bacterium in the environment should make it the cause of frequent human disease, a lack of efficient defenses against the immune system means that the organism is rarely able to establish a long term infection. However, in the case of immunocompromised hosts (eg. AIDS patients or individuals undergoing chemotherapy), the organism has been shown to be capable of causing severe problems (Schimpff et a l , 1970). 2 The pathogenicity of P. aeruginosa has been demonstrated in infections of patients suffering from burns of the skin (Pruitt and McManus, 1984)". The injured tissues of these patients lack many of the common mechanisms of defense. This deficiency allows bacteria, which would normally be quickly eliminated from these surfaces, the opportunity to establish a foothold on the body of the prospective host. Once an infection has been established, P. aeruginosa is able to elaborate a number of intriguing mechanisms which aid in the spread of the bacteria throughout the body. Several different proteases are secreted by P. aeruginosa which serve to destroy the surrounding host tissues and allow the bacteria to invade further into the body (Morihara and Tsuzuki, 1977). Other enzymes act to release various molecules from the host environment which are essential for the long term survival of the microorganism (Vasil, 1986). In addition to these proteins, the previously mentioned iron chelators pyoverdine and pyochelin are also released into the surrounding environment. These molecules bind to available iron(III) ions which are normally extremely insoluble. The iron-chelator complex is bound by the bacterium via a surface-exposed receptor and is then brought into the cell cytoplasm by the actions of various membrane associated proteins (Meyer et al., 1990). The internalized molecule is stripped of its i ron load and shunted back out into the external media where the process is repeated. This mechanism is an efficient means by which P. aeruginosa is able to obtain sufficient quantities of iron(III) which would normally be bound up by various host chelating proteins (Sriyosachati and Cox, 1986). Another category of patients who must endure the effects of P. aeruginosa infections are those who suffer from the genetically acquired disease cystic fibrosis (CF). Symptoms of this disease include an inability to adequately clear fluids from the lungs. Individuals suffering this genetic affliction have been found to produce a mutated form of a chloride channel which is found in the plasma membrane of epithelial cells which line the lungs (Welsh and Liedke, 1986). The lack of a normal chloride channel results in 3 the unusually high levels of fluid accumulation which are characteristic of the disease. The large volumes of fluid present in the lungs of these patients not only make breathing extremely laborious but they also produce a detrimental effect on the immunological competency of the surrounding tissues. This deficiency i n local immunity allows bacteria the chance to settle in for a prolonged occupation. The ability of P. aeruginosa to produce a viscous surface layer is thought to enhance its ability to infect the lungs of C F patients. This mucoid layer, which is secreted by the bacteria into the extracellular environment, is composed of a polymer of mannuronic and glucuronic acid residues (Pedersen et al., 1989). The viscosity of this substance, along wi th its high concentration of negative electrical charge, serves to inhibit the deposition of antibodies on the surface of the invading bacteria. This surface layer not only obstructs the phagocytosis of this microorganism by indigenous leukocytes (Bayer et al., 1991), but it may also serve to increase the ability of the bacteria to bind to the epithelial cell surface of the lungs (Marcus and Baker, 1985). This mucoid layer therefore qualifies as a virulence factor which is elaborated by P. aeruginosa when the organism inhabits the lungs of CF patients. While P. aeruginosa is not a common source of human infections, clinical studies have shown that once an individual has become colonized by this bacterium, the eradication of such an infection is often very difficult. This is due to a high degree of intrinsic resistance to antibiotics which is common among the various pseudomonads (Bryan,. 1979). These high levels of resistance are thought to be due to the unique features of this bacterium's outer membrane. The membrane, which is extremely impermeable in P. aeruginosa, functions as an initial screen by which the organism is able to discriminate between various molecules present in the extracellular environment. The low permeability of the P. aeruginosa outer membrane limits the extent to which various molecules (such as antibiotics) are capable of entering the cell. 4 The cell wall of gram negative bacteria The cell walls of all gram negative bacteria are composed of three separate domains; the inner membrane, the periplasmic space and the outer membrane. Each of these structures serve a unique role i n maintaining an equilibrium between the extracellular environment and the internal milieu of the cell. They each also play a role in the uptake of nutrients by the cell. In addition to these three components, many bacteria produce an extracellular capsule layer. This surface structure has been implicated i n providing the cells with a means of inhibiting the phagocytic capacity of neutrophils and macrophages (Horwitz and Silverstein, 1980) in a manner similar to that of the P. aeruginosa mucoid layer. The inner membrane The inner membranes of most gram negative bacteria are composed of a small variety of different phospholipids (Nikaido and Vaara, 1987). These molecules are arranged in a symmetrical bilayer membrane wi th the polar head groups oriented towards either the cytoplasm of the cell or the aqueous matrix of the periplasmic space. The hydrophobic tails of the phospholipids are faced towards each other and associate by means of strong, nonpolar interactions. The arrangement of the lipids in such a manner provides a formidable barrier to the non-specific transit of hydrophilic molecules across the membrane. Passage of such molecules is limited by the hydrophobic core of the membrane. However, this structure provides little impediment to the transport of even moderately hydrophobic compounds. The l ip id molecules are interspersed with a number of different proteins which serve various functions. Some of these proteins act to generate energy which is needed for the active transport of nutrients into the cell (Dassa and Hofnung, 1985) while others function as receptors that bind certain substrates which find their way into the periplasmic space (Kristjansson 5 and Hollcher, 1979). The bound molecules are then brought through the membrane by the action of these receptors in concert with-those of other inner membrane associated proteins. Other inner membrane proteins act as pumps which serve to expel noxious compounds from the cell interior (Li et al., 1995); still other proteins carry out the synthesis of various components of the cell wall (Tuomanen, 1986). The periplasmic space The area separating the inner and outer membranes of gram negative bacteria is referred to as the periplasmic space. The name of this region is deceptive, since it is composed not of empty space, but of a viscous, gelatinous substance. This material, which has been termed peptidoglycan, was shown to be comprised of linear strands of two different amino sugars (N-acetylglucosamine and N-acetylmuramic acid). Short peptides (normally four amino acid residues long) which serve to cross-link the individual strands via peptide bonds are bound to the N-acetylmuramic acid residues. The peptidoglycan layer acts as a stabilizing force in the cell wall . The covalent or strong non-covalent attachment of this layer to proteins located i n the outer membrane has been observed in a number of different bacteria (Nikaido and Vaara, 1987). In addition to the structural role played by the constituents of the periplasmic space, this area of the cell also serves to determine the types of molecules which are able to enter the cell cytoplasm as well as the rates at which these compounds cross the membranes. Several substrate-binding proteins have been shown to reside i n this region of the cell wall . The specificity of these proteins is usually towards molecules which are essential to the growth and survival of the cell. By maintaining i n the periplasm a low concentration of free substrate, these proteins ensure the continuous flow of those molecules through the outer membrane. Both the maltose-binding protein of E. coli and the phosphate-binding protein of P. aeruginosa have been shown to be responsible for keeping the free periplasmic concentrations of their respective substrates far below that 6 of the external environment (Kellerman and Ferenci, 1982; Poole and Hancock, 1984). Transport of appropriately charged molecules into the periplasmic space is also thought to be enhanced by the presence of an electrical potential (the Donnan potential) which is maintained across the outer membrane by a number of periplasmically-localized fixed charges (Stock et a l , 1977). The outer membrane Like the inner membrane, the outer membrane of gram negative bacteria is composed of a number of phospholipid molecules. In addition, this outer layer also contains lipopolysaccharide (LPS) which is only found in the cell walls of bacteria (Nikaido and Vaara, 1987). This molecule is comprised of the l ip id A core and a hydrophilic polysaccharide tail. The l ipid A component is anchored i n the membrane while the hydrophilic tail extends away from the surface. The arrangement of the outer membrane is such that the LPS molecules form the external face of the membrane while the phospholipids line the inner face. The individual LPS molecules have been shown to be both covalently and non-covalently attached to certain outer membrane proteins (Hancock and Karunaratne, 1990). They have also been found to form inter-molecular interactions via the mutual attraction of certain sites on these molecules to various divalent cations (eg. M g ^ ) . The presence of these ions has been shown to be important to the maintenance of the stability of the outer membrane (Hancock et al., 1994). Like most biological membranes, the outer membranes of gram negative bacteria contain a number of proteins which serve various functions. One of the major protein constituents of the outer membrane of E. coli is OmpA. While this protein has been shown to play a role i n maintaining the structural stability of the outer membranes of this bacteria, more recent studies have led to the proposal that O m p A may also function as a channel (Sugawara and Nikaido, 1992). Another protein which serves to stabilize the structure of the E. coli outer membrane is the Braun lipoprotein. This protein is 7 localized on the inner face of the outer membrane and has been shown to be covalently bound to the peptidoglycan layer of the periplasmic space (Braun and Sieglin, 1970). Since the presence of the hydrophilic LPS side chains on the external surface provides an effective means by which to limit the uptake of hydrophobic compounds and the nonpolar nature of the inner l ipid core does not allow the passage of hydrophilic molecules, the outer membrane provides an excellent barrier to nonspecific transport through the cell surface. The impermeability of this bilayer necessitates the inclusion of certain proteins in the outer membrane which are capable of facilitating the uptake of various nutrients from the environment. The permeability differences displayed by various gram negative bacteria have been shown to be at least partially dependent on differences in the number and the types of these proteins which are expressed in the outer membrane of these organisms. Porins of gram negative bacteria: Functional aspects The presence of porins in the outer membranes of gram negative bacteria serves to provide these organisms wi th a means by which to obtain various growth factors from the extracellular environment. These proteins form channels which span the membrane and allow the passage of molecules from the external medium into the periplasmic space (Hancock, 1987a). Functional analyses of these channels have been accomplished using a number of different techniques. These include uptake experiments using whole cells expressing the porin of interest (Nikaido, 1986) as wel l as model membrane studies of purified proteins (Hancock, 1987b). The latter category of methods provide a finer level of biophysical examination of the channels as they allow better control of the environment to which the porins are exposed. However, both types of experimental systems have been used to gain valuable information on the function of this class of bacterial outer membrane proteins. 8 The majority of porins examined to date have been shown to form non-specific channels. This means that they allow the passive diffusion of molecules through the outer membrane provided that the size of the molecule is below the exclusion limit of the channel. Porins found in E. coli which have been shown to display these traits include the wel l characterized proteins OmpF and OmpC. Some of the earliest studies of the mechanisms of porin function have been performed on these proteins (Nikaido and Rosenberg, 1983). Functional analyses have shown that the OmpF and OmpC channels share many common features, including an estimated diameter of slightly more than 1 nm and an exclusion limit of approximately 600 daltons. The channels of these porins have also been shown to exhibit a slight (OmpF) to moderate (OmpC) selectivity towards the transport of cations over anions. However, these two porins differ in their single-channel conductances as measured using artificial membrane systems. When reconstituted into bilayers, purified OmpF exhibits an average single-channel conductance of 2.1 nS in I M KC1. Using these same methods the single-channel conductance of the OmpC porin was shown to be 1.5 nS (Nikaido and Rosenberg, 1983). These apparent differences in function have been determined to be due to differences i n the electrochemical nature of the amino acid residues which are located within the channels of these two proteins. Proteins analogous to OmpF and O m p C have been identified i n the outer membranes of other gram negative bacteria (Lee and Schnaitman, 1980). Unlike the constitutively expressed OmpF and OmpC porins, PhoE is only expressed in the outer membrane when cells experience phosphate deprivation (Overbeeke and Lugtenberg, 1980, Poole and Hancock, 1986). The production of this porin under these conditions is related to its selectivity for the transport of anions. The negatively-charged phosphate ions are able to more easily enter the cell via these anion-selective channels. However, the ability of PhoE to transport many anions as efficiently as it does phosphate ions has placed this protein in the category of the general diffusion 9 porins (Benz et al., 1984). Despite these differences in ion selectivity, PhoE channels have been shown to possess conductance levels and exclusion limits similar to those of OmpF and OmpC (Benz and Hancock, 1982). This is consistent wi th the fact that the amino acid sequence of PhoE is quite similar to the sequences of these other two porins (Jeanteur et a l , 1991). Another E. coli porin which is expressed in response to particular growth conditions is the maltose-specific porin LamB. Expression of this protein in the outer membrane is induced upon growth of the organism i n low concentrations of carbon sources other than maltose (Schwartz, 1987). Studies wi th both whole cells and isolated protein have shown that this porin forms channels that exhibit a distinct preference for the transport of maltose and maltodextran sugars (Wandersman et al., 1979; Benz et al., 1986). This specificity has been shown to be due to the presence of a substrate-binding site. Mutants which have been isolated by selecting for growth on large maltodextran molecules have been found to express LamB proteins which contain deletions in specific regions of the protein sequence (Misra and Benson, 1988). These regions are thought to be involved i n maintaining the substrate-specific nature of this particular porin. The related enteric bacterium Salmonella enterocolitica has been shown to express a LamB protein which is very homologous to the E. coli protein (Schulein and Benz, 1990). Several proteins which have been shown to exhibit porin-like behaviour have been identified in the outer membrane of P. aeruginosa. These include the nonspecific channel forming protein OprF (Benz and Hancock, 1981) as well as the specific porins OprD (Trias and Nikaido, 1990a), OprB (Hancock and Carey, 1980) and OprP (Hancock et al, 1982). While OprF has been shown to be the most common constitutively expressed porin found in P. aeruginosa, both the low level of channel activity of this protein combined wi th the low diffusion rate of solutes through the few active channels appear to be responsible for the relative impermeability displayed by the outer membrane of this bacteria. The exclusion limit of the channels formed by OprF has been 10 experimentally determined to be well above that of the E. coli general diffusion porins (in vitro estimates place it at over 3000 daltons, whereas in vivo data suggests a channel diameter of at least 1.56 nm). However, despite this apparently larger channel size, the rate of diffusion of solutes through the channels of OprF was about two orders of magnitude lower than the rates afforded by the E. coli porins OmpF and OmpC. This apparent contradiction has been theorized to be due to the fact that only a small proportion of the OprF porins form the large channels while the majority of the proteins form channels wi th considerably smaller exclusion limits. Planar l ip id bilayer experiments wi th OprF have shown that addition of this protein to bilayers results i n the incorporation of channels which can be divided into two categories according to measured conductance. The vast majority of channels exhibited conductances of 0.34 nS while a small minority of larger channels with conductances of approximately 5 nS were also observed (Benz and Hancock, 1981; Nikaido et al., 1991). The homologous E. coli outer membrane protein O m p A has recently been shown to form channels which share some of the characteristics displayed by the OprF channels (Sugawara and Nikaido, 1992; 1994). Since the differences in function displayed by these two proteins from that of the classical general diffusion porins would dictate that their three-dimensional structure must differ somewhat from that of the other channel forming proteins, the topological arrangements of O m p A and OprF w i l l not be discussed. Both OprD and OprP represent specific porins which are expressed in the outer membrane of P. aeruginosa under specific environmental conditions. The wild-type expression of OprD is low and the conditions that result in enhanced expression have not been well defined. The substrate specificity of the channels formed by this porin mean that the transport of essential molecules across the outer membrane is not solely dependent on passive diffusion. The proposed binding site of OprD is thought to be specific for the structure of basic amino acids (Trias and Nikaido, 1990a). However, it appears that the structural similarity of the antibiotic imipenem to that of these amino 11 acids allows this molecule to also gain entry into the cells via the channels of this porin (Trias and Nikaido, 1990b). Strains which express OprD have been shown to be significantly more sensitive to the presence of imipenem than OprD deficient strains. The growth of P. aeruginosa under phosphate-starvation conditions (< 0.2 mM) induces the expression of the phosphate-specific porin OprP (Hancock et al., 1982). The channels formed by this protein demonstrate a strong selectivity for the transport of anions over cations (Benz et al., 1983). This feature is also exhibited by the analogous phosphate-starvation inducible E. coli porin PhoE (Benz et al., 1984). However, unlike the E. coli porin, OprP has been shown to possess a binding site which makes the channels specific for phosphate ions. The presence of this binding site means that the channels of OprP transport phosphate ions at approximately a 100-fold higher rate than they do other similarly sized anions (Hancock and Benz, 1986). Given the low general outer membrane permeability of P. aeruginosa, this specificity is thought to be essential to this porin's role in the acquisition of phosphate ions when the extracellular concentrations are extremely low. The sequence of the recently cloned P. aeruginosa poly-phosphate specific porin OprO was shown to be quite similar to that of OprP (76% identity, 16% conserved substitutions) (Siehnel et a l , 1992). The differences in substrate-specificity displayed by these two proteins appear to be due to minor variations in the amino acid sequences. Porins of gram negative bacteria; Structural aspects A number of different techniques have been employed to determine the structure of bacterial porins. These include both genetic and biophysical approaches. The data gathered from these various studies have shown that these proteins share a number of common structural features, including a high percentage of (3-sheet content. The differences in function exhibited by the various porin channels appears to be due to 12 specific regions of their amino acid sequences, and i n fact, only a small number of individual residues have been implicated in defining the unique channel characteristics of these different proteins. Insertion and deletion mutagenesis of the genes encoding various different porins has shown that only certain sections of the expressed proteins were permissive for these types of alterations (Boulain et al., 1986; Bosch and Tommassen, 1987; Roque and McGroarty, 1990). Data previously obtained from circular dichroism analysis indicated that these proteins contain a large number of (3-strands. Thus, it seemed likely that the sites which were permissive for the insertion or deletion of small peptides were located in loop regions which connected the individual p-strands. The insertion of a foreign epitope into these permissive sites allowed for the identification of those loop regions which were exposed at the surface of whole cells (Charbit et al., 1986; Agterberg et al. 1987). Although a number of topological models have been based on data generated by the previously described methods, the most accurate procedure available for the assessment of the three-dimensional configuration of a given protein is that of X-ray crystallography. While this technique has been used to determine the structures of several ct-helical proteins, it has only been used sparingly for the analysis of porin structures. This has been due to the difficulties encountered i n crystallizing these proteins. Recent successes in growing suitable crystals has allowed this technique to be used to create structural models for a small number of porins. These models have been compared to those generated by genetic analyses to strengthen the structural predictions for these proteins. The first porin structure to be solved by X-ray diffraction analysis was a porin of Rhodobacter capsulatus. The model which was based on this analysis showed this protein as being arranged in a complex of three identical monomers (Weiss et al., 1990). Chemical cross-linking studies had previously provided evidence that these proteins 13 formed oligomeric structures in the outer membrane (Angus and Hancock, 1988). The individual subunits were arranged around a central axis and were tilted at a slight angle (approximately 16°) toward that axis. Each monomer assumed the conformation of a p-barrel which was composed of 16 strands. The barrel enclosed a central water-filled channel. The amino acid sequences of the p-strands tended to alternate between hydrophobic and hydrophilic residues. This would seem appropriate if one side of the P-sheet were to face the interior of the channel while the other side faced the l ip id core of the membrane. This is in contrast to the sequences of the membrane spanning segments of ct-helical membrane proteins which contain predominately hydrophobic residues (Taylor, et al., 1994). The P-strands of the porin monomer structures were shown to be connected by loop regions. The loops exposed to the periplasmic space were found to be relatively shorter than those exposed to the cell surface. These larger surface-exposed loops appear to play a much more important role in defining the channel characteristics of the protein than the periplasmically-exposed loops. Six of the eight surface-exposed loops (loops 1, 4, 5, 6, 7 and 8 ) were shown to form an umbrella-like structure over part of the opening to the central channel. This umbrella appears to limit the size of the molecules that are capable of passing through the channels of this protein. The third surface-exposed loop was shown to be positioned in a way which allowed it to exert a substantial influence over the passage of molecules through the channel. This region of the protein was seen to fold back into the channel and create an eyelet structure approximately half-way down the channel. This eyelet is believed to constrict the internal diameter of the pore and further limit the diffusion of solutes. The crystal structures of the highly homologous E. coli general diffusion porins OmpF and PhoE were found to be remarkably similar to that of the R. capsulatus porin (Appendix-Figure 21) (Cowan et a l , 1992). This is somewhat surprising, given that the amino acid sequence of this protein does not show a great deal of homology wi th that 14 of the other three porins. Each of the E. coli proteins were also found to exist in the outer membrane as a trimer of three identical subunits. Like the structure of the R.. capsulatus porin, the models of the E. coli porins showed the third surface-exposed loop as being folded back into the channel and the second loop as interlocking with the neighbouring monomer. The locations of the loop regions in the crystal models of these porins coincided wi th the placement of these regions in topological models based on mutagenic procedures (van der Ley et al., 1987; Benson et al., 1988). The sections of these proteins which were proposed as composing the loop regions were shown to be the same sections that were permissive for insertion and deletion mutagenesis. In addition to verifying the location of the loops, specific alterations in the amino acid sequences of these proteins have shown that certain residues located i n and around these loops serve to define their individual channel properties (Misra and Benson 1988). Combinatorial mutagenesis of OmpF, OmpC, and PhoE have shown that regions located in the amino-terminal half of these proteins are involved in determining the specific channel characteristics of these porins (Benz et al., 1989). Genetic manipulations which resulted i n the combination of different portions of these proteins helped to identify several discrete regions implicated as serving to define both the exclusion limits and the ion selectivities of the channels. These experiments have shown that the third surface-exposed loop contains a number of residues which play a role in the functioning of the channels. Mutagenic studies have shown that the region between residues 108 and 133 in OmpF are responsible for the cation selective nature of the channels (Benson et al., 1988). Other experiments have shown that Lys 1 2 5 in PhoE serves to govern the ion selectivity of the channels of this protein (Bauer et a l , 1989). A l l of these residues are contained in the third surface-exposed loops of the proposed topological models. Thus, the results of these experiments helped to strengthen the proposal of an interior location for this particular surface-exposed loop region. 15 Since the structure of the maltose-specific LamB porin was found to contain a large amount of (3-sheet structure, it was assumed that the topology of this protein would be similar to that of the other better characterized members of the family. However, as there exists little similarity between the amino acid sequence of LamB and the other porins (Jeunteur et a l , 1991), and as this protein contains approximately 25% more residues than the other porins, it was difficult to construct a model of LamB based on the previously solved porin structures. A recently published crystal model of this protein showed the structure to be quite similar to that of the other porins, wi th one noted exception; the p-barrels were constructed out of 18 strands instead of 16 (Schirmer et al., 1995). Mutagenic analyses revealed that residues located i n both the second and third surface-exposed loops were important for the substrate-specificity of this porin (Heine et al., 1987). In addition to these sites, certain residues arranged throughout the sequence of the protein were established as also directing the functions of the channels formed by this protein. These amino acids have been proposed to be involved in forming a diffuse maltose binding site which serves to specifically guide the substrate through the channels and into the periplasmic space. The genes encoding the P. aeruginosa specific porins OprD and OprP have been cloned and expressed in an E. coli background (Huang and Hancock, 1993; Siehnel et al., 1988). The evident homology of the OprD sequence with that of the classical porins allowed for the construction of a topological model of this protein which was based on that homology (Huang et al., 1995). Subsequent experimental testing of this model showed that OprD does indeed appear to assume a conformation similar to that of the majority of previously characterized porins (Huang et a l , 1996). These experiments also indicated that residues contained i n the second and third surface-exposed loops of this protein might be involved in forming the substrate binding site of the channels. Unlike the sequence of OprD, the amino acid sequence of OprP displays little homology to that of the other porins (Siehnel et al., 1990). In spite of this fact, OprP 16 shares many biophysical properties with these proteins, including a high proportion of (3-sheet structure (Worobec et al., 1988) and a native oligomeric configuration (Angus and Hancock, 1983). Lysine-specific chemical modifications of this protein were shown to adversely affect the single-channel conductance as well as the anion selectivity and the phosphate specificity of the OprP channels (Hancock and Benz, 1986). Similar modifications of PhoE were also shown to affect the channel properties of this porin (Darveau et al., 1984). The results of the OprP modification experiments indicated that lysine residues contained in the sequence of this protein were involved i n determining the various channel characteristics. Phosphate transport across the outer membranes of gram negative bacteria The acquisition of inorganic phosphate (phosphate ions) and phosphorylated compounds is an essential function of growing microorganisms. Many bacteria have been shown to possess two separate systems for the transport of phosphate ions; one which operates in a phosphate rich environment and a second which functions under conditions of phosphate deprivation. This second group of genes is part of what is often referred to as the Pho regulon (Wanner and Latterll, 1980). The products of the E. coli Pho regulon are among the most extensively studied representatives of this category of bacterial proteins. Many other gram negative bacteria contain genes which appear to be homologous to components of this regulon (Filloux et al., 1988; Lee et al., 1989; Scholten et al., 1995). Therefore, information regarding the E. coli Pho regulon has been invaluable to gaining an understanding of the means by which other gram negative bacteria respond to conditions of phosphate deprivation. The Pho regulon of E. coli has been found to include at least 31 genes which are arranged i n eight separate operons (Wanner, 1993). The expression of these genes is induced by low extracellular concentrations of phosphate. Under conditions of 17 phosphate deprivation, the inner membrane localized PhoR sensor protein along wi th the response regulator protein PhoB act in concert to activate transcription of the components of the regulon. The phosphorylation of PhoB by the sensor protein occurs under conditions of phosphate deprivation. The activated PhoB protein i n turn binds to regions on the chromosomal D N A which are upstream of the individual Pho operons. These regions contain common consensus sequences and are usually referred to as Pho boxes. The Pho regulons of many enteric bacteria have also been shown to be under the control of similar upstream genetic elements. The gene products of the E. coli Pho regulon include an inner membrane permease (PstB), a periplasmically localized phosphate binding protein (PstS) and an outer membrane porin (PhoE). Along with the actions of a few other proteins, these components of the Pho regulon serve to increase the rate at which the organism is able to assimilate phosphate from the surrounding environment (Wanner, 1993). The common purpose provided by these different proteins engendered their inclusion i n a particular subgroup of the Pho regulon; the Pst transporter system. Proteins homologous to members of the E. coli Pst system have been identified in such bacteria as Shigella flexneri (Scholten et a l , 1995) and Salmonella typhimurium (Bauer et al., 1985). The Pst system of P. aeruginosa has been shown to be quite similar to that of E. coli. The genes of this system are also under the control of upstream Pho boxes and the products of these genes include a periplasmic phosphate binding protein (Poole and Hancock, 1984) and an outer membrane localized channel forming protein (OprP) (Hancock et al., 1982). Mutants which are deficient in one or more components of this system are unable to transport inorganic phosphate when the extacellular concentration of this nutrient is low (Poole and Hancock, 1984). This lack of an intact high affinity phosphate transport system results in significantly slower rates of growth for these mutants i n comparison to the w i ld type strains. The genes encoding both the periplasmic phosphate-binding protein and OprP have been cloned and expressed i n E. 18 coli (Worobec et al., 1988; Siehnel et a l , 1988). The P. aeruginosa periplasmic phosphate binding protein was found to be extremely homologous to the E. coli PstC protein. However, the sequence of OprP displayed little homology wi th the E. coli phosphate starvation inducible porin PhoE (Siehnel et al., 1990). Functional studies have revealed that the P. aeruginosa porin is far more selective for the transport of anions than the E. coli protein. In addition, the channels of OprP were shown to possess a phosphate-specific binding site which was proposed to be formed by the presence of one or more lysine residues (Benz and Hancock, 1986). Aims of this thesis The purpose of this thesis is to indirectly ascertain the structure of the P. aeruginosa phosphate-starvation inducible porin OprP and to examine how this structure relates to the functions of this protein. A topological model of the structure of OprP was created using data generated by insertion and deletion mutagenesis of the gene encoding this protein in combination wi th previously generated data regarding the hydrophobicity and propensity to form (3-turns of individual segments of protein sequence. Subsequently, the roles of specific residues i n the transport of phosphate ions through the channels of this porin were assessed by means of site-directed amino acid substitution mutagenesis. The results of this second set of experiments were combined with those which were utilized in the creation of the topological model in order to gain a greater understanding of the mechanism of phosphate uptake across the P. aeruginosa outer membrane. 19 MATERIALS A N D METHODS Bacterial strains and plasmids Strains and plasmids used in this study are listed in Table I. The P. aeruginosa strain H I 03 was utilized in the purification of w i ld type OprP which was used for the production of polyclonal anti-OprP antiserum. The E. coli strain DH5-ocF' was used in all procedures involved in creating the oprP mutant plasmids. E. coli strain CE1248 was used as the host background in all OprP expression experiments. This strain does not produce O m p C and OmpF due to a regulatory mutation and is also deficient in the expression of PhoE due to a gene deletion. In addition, CE1248 contains a mutation in the phoR gene which allows for the constitutive expression of the Pho regulon. The gene encoding OprP was originally cloned into the expression vector p M M B 6 6 H E to create plasmid pRSP-3. This latter plasmid was used as the source of the oprP gene which was subcloned into the commercial phagemid vector pTZ19U to create plasmid pAS27 (see below). Plasmid pAS27 was the parent plasmid used in the creation of all the oprP insertion and substitution mutants. Plasmid p U C 4 K A P A used as the source of the Hinc II fragment which contained the kanamycin resistance cassette. This fragment was used in the creation of the linker insertion mutants. Plasmids pUC18 and pUC19 were used to create the vector pUC18/19 which was utilized i n the creation of plasmid pAS27. Media and growth conditions Cells were grown overnight at 37 °C on agar plates or in broth cultures. The P. aeruginosa strain was initially streaked on to Mueller-Hinton agar plates and grown 15 O d <u V H ,<L5 15 O 15 O VH d o CO 15 VH 15 P< o VH O H 4-» > 15 O O a 15 O T3 s CO or (73 • 1—1 g V H o o3 C/3 PQ oo O N o3 d N o o3 < O H d a N o o3 O o T3 O N N O oo PH CO i— l -=! 2 1 — 1 T-H 15 -e-V Q < a 15 V H O N IS O N C50 O N 13 15 VH 15 oo U ON O N i>1 u d O o o d o3 nd O H O VH o 4-1 o V H C5H u i— i o < AH o Q •S s o3 •rH o o3 o3 rd fin O 15 r > W) d 'd o to: 'I LO: 03 O N i— i u oo u r D 03 •rH o 03 o3 rd §3 bO CO S ° r d ^ 03 H * a d CO VH 03 3 15 D ft S U .2 &1 I — I CO d U SH <5 u 03 •rH o 03 03 -d a, V H O 4-1 o 15 r> d d o O N 1—I N H P H 21 o T3 o 1-1 pi o CO oo oo 03 o 03 03 ,—< CCi 03 13 a 4 3 03 CO c/1 O Pi as 13 13 H 03 -4-> 03 </l 4 3 H 03 H 3 CN l-i J3 & -a U >> 3 I/) 4 3 H H -i-> i/i </) 4 3 •5H tn 03 ft O ft -*-» 03 13 03 ft >> O 03 o c3 O OO 1—I U ft T3 ' i ' f t gp • 1-1 id o i— i N H ft T3 CN ft a o xi u 03 i / l a « »—l VJ_/ c/p k> 43 43 'a ft 03 03 03 rj "a ^ 4 3 j -"ft T 3 a v I a o 5 o t o 03 O o PI 03 c/1 T3 "3 £ ft c/i c/3 T3 c/5 (73 cn i cu CO ft CN CO < ft cn CU I 1—I •J ft cn i— i i-l OH I CU CN ft S PI o 1/1 4 2 i/i c/5 •"J 42 ^ ^3 CO c/i <! 43 ft ft J2 03 H 22 overnight at 37 °C. The cells were asceptically scraped from the plates and used to inoculate 1 litre flasks containing a defined phosphate minimal medium (phosphate-sufficient conditions) (Hancock et al., 1982). These cultures were grown overnight and were then used to inoculate a 60 litre fermentor containing phosphate-deficient minimal medium (phosphate-deficient conditions). The E. coli strains were grown overnight i n Luria-Bertani (LB) broth (lOg tryptone, 5g yeast extract, 5g N a C l / litre) or on LB agar plates (LB broth + 2% agar). When cells were harbouring plasmids, the media was supplemented wi th 50 u g / m l ampicillin. Broth cultures were agitated on a shaker at 37 °C. When strain CE1248 was used, 0.4% glucose was included to inhibit expression of LamB. In addition, 5 u M isopropyl thio-D-galactoside (IPTG) was included in cultures used in the indirect immunofluorescence experiments i n order to optimize expression of the plasmid-encoded proteins. Media components were purchased from Difco Laboratories (Detroit, MI). Chemicals KC1, K 2 H P O 4 and KH2PO4 used in the planar bilayer experiments were purchased from Fisher Scientific. KC1 was used unbuffered (pH 6.0) while equal molar concentrations of K 2 H P O 4 and K H 2 P O 4 were mixed to achieve a p H of 8.0. A l l other chemicals were purchased from either Fisher Scientific, Sigma Chemicals, B D H or I C N Biomedicals Inc. 23 General recombinant D N A techniques 1. Enzymes and plasmid manipulation A l l recombinant D N A techniques were performed as described by Sambrook et al. (1989). Restriction endonucleases and D N A modifying enzymes were purchased from G i b c o / B R L , Boehringer-Mannheim, New England Biolabs or Pharmacia and used wi th the accompanying buffers at the recommended concentrations. A l l restriction digests were performed at 37 °C. Ligations were performed for 2 h at room temperature or overnight at 16 °C using T4 D N A ligase. Plasmid transformation of competent cells was accomplished using the CaCf2 method. Plasmids were isolated from overnight cultures of E. coli by the alkaline lysis method followed by phenol/chloroform extraction and ethanol precipitation. D N A fragments were isolated from agarose gels by either centrifuging agarose slices containing the fragments through a plug of siliconized glass wool followed by phenol extraction and ethanol precipitation or by using a commercial D N A isolation kit (Gene Clean II, Bio 101 Inc. La Jolla, C A ) which utilized a DNA-b ind ing silica matrix. 2. Agarose gel electrophoresis Analysis of restriction endonuclease digested plasmids was accomplished by electrophoresis through 1% agarose gels containing either I X TBE buffer or I X T A E buffer and 0.5 u g / m l ethidium bromide. The 10X loading buffer contained 0.25 % bromophenol blue, 0.25 % xylene cyanol FF and 30 % glycerol in water. Gels were run at a constant current of 85 volts for small gels (10 cm) and 120-150 volts for large gels (25 cm). After electrophoresis the D N A bands were observed using a short-wave U V trans-illuminator and photographed using polaroid high-speed black and white film and a polaroid MP-4 Land camera fitted with a lens covered by a #15 deep yellow filter. The agarose was purchased from Bio-Rad Industries. 24 3. Synthetic oligonucleotides Oligonucleotides used as primers for site-directed mutagenesis and D N A sequencing were synthesized on an Appl ied Biosystems Model 392 automated D N A synthesizer (Applied Biosystems Incorporated, Foster City, C A ) . After an overnight incubation in N H 4 O H at 55 °C to cleave the base-protecting groups from the 5'-end of the oligonucleotides, the D N A was dried by centrifugation under vacuum. The dried D N A was resuspended in 30% N H 4 O H and precipitated by adding two volumes of isopropanol followed by centrifugation at 12,000 g for 15 min. The purified D N A was resuspended in sterile d H 2 0 and quantitated by monitoring U V absorbance at 260 nm. Double-stranded oligonucleotides used for the mutation of an oprP PstI site and the insertion of the malarial epitope into the oprP gene were annealed as follows. Equal quantities of each strand of D N A were combined in sterile d H 2 0 , heated to 90 °C for 15 min and allowed to cool slowly to room temperature. 4. D N A sequencing Plasmid D N A was sequenced using the ABI automated fluorescent sequencing system wi th a Model 373 sequencer and P C R protocols provided by ABI . Template D N A was prepared by either using Quiawell-8 plasmid kits (Quiagen) or the polyethylene glycol precipitation method (Sambrook et a l , 1989). Primers were synthesized on an A B I D N A synthesizer. The primers and templates were combined at the recommended concentrations, and were mixed with the sequencing premix which contained both fluorescent and nonfluorescent nucleotides as wel l as the D N A polymerase ampliTaq. The samples were then subjected to 26 rounds of a P C R cycle which consisted of a 30 second incubation at 96 °C followed by a 30 second incubation at 50 °C and a 3 minute incubation at 60 °C. 25 After amplification the D N A was purified by centrifugation through 1 m l spin columns composed of G-50 Sephadex (Pharmacia) followed by ethanol precipitation. The samples were then air dried, resuspended in four u l of a 5/1 mixture of fo rmamide /EDTA (pH 8.0) and heated at 90 °C for two minutes. The samples were immediately placed on ice and were loaded on to the gels within an hour. The gels, which contained 50% w / v urea and 4.75% acrylamide, were run at a constant power of 30 Watts for 12 hours. Automatic analysis of the sequencing data was accomplished wi th the aid of the ABI 373 analysis software program. General protein and immunological techniques 1. SDS-polyacrylamide gel electrophoresis Protein electrophoresis on SDS-polyacrylamide gels was performed as previously described. Separating gels contained 12% acrylamide/bis (Biorad) and stacking gels contained 2.5%. The running buffer contained 3.1g Tris, l g SDS and 14.4g glycine per litre. Whole cell lysates or isolated outer membranes were heated at 100 °C for 10 min prior to loading on to the gels except when nondenatured trimers were required. Undenatured proteins were not heated prior to loading. The 2X reduction mix contained 4% SDS, 20% glycerol, 125mM Tris-HCl and 40mM E D T A . Electrophoresis was carried out with a constant current of 120 V until the dye front reached the bottom of the gel. Both unstained and pre-stained M W t markers were purchased from Bio-Rad. Proteins were visualized by staining the gels in a solution of 45% methanol/45% d H 2 0 / 10% acetic acid containing 0.28% Coomassie Brilliant Blue. 2. Western immunoblotting For Western immunoblotting, unstained gels were transferred to nitrocellulose membranes (Schleicher and Schuell) as previously described (Towbin et al., 1979). After 26 electrophoresis, the gels were soaked for 10 min in I X blotting buffer (25mM Tris, 192mM glycine, 20% methanol). The gels were then placed on top of a piece of pre-soaked nitrocellulose cut to fit the gel and sandwiched between two pieces of filter paper i n a trans-blotting apparatus (Bio-Rad). The gel was then submerged in an electrophoresis chamber filled wi th blotting buffer. The electrotransfer of the proteins to the nitrocellulose membrane was carried out with a constant current of 100 V for 1 hr. Ice packs were placed in the chambers during the electrotransfer i n order to prevent over heating of the buffer. After blocking with 3% bovine serum albumin (BSA) in phosphate buffer saline (PBS) (pH 7.4), the nitrocellulose membranes were incubated with either a 1/1000 dilution of anti-OprP rabbit serum or a 1/5000 dilution of anti-malarial epitope monoclonal antibody in 1% B S A / P B S for 2 hrs at room temperature on a shaker. After washing 2X with PBS, the membranes were incubated wi th a 1/2000 dilution of alkaline phosphatase conjugated goat anti-rabbit or goat anti-mouse antibody (Bio-Rad). The membranes were washed 2X with PBS and the bound antibodies were detected by incubating the membranes in 0.1 M Tr is -HCl (pH 9.6) containing 40 m M M g C l 2 , 5 m M BCIP and 10 m M NBT. Blots were development until the positive control exhibited a dark band the membranes were washed with PBS to stop further development. 3. Antibodies Monomer specific anti-OprP rabbit serum was made during the course of a previous study (Poole and Hancock, 1986). Trimer-specific antiserum was made using F P L C purified OprP. A 3 kg female N e w Zealand White rabbit was injected subcutaneously wi th 100 ug of purified OprP suspended in 1.5 m l Freunds complete adjuvant. Two more injections at 5 week intervals with 100 ug of purified OprP suspended in 1.5 ml Freunds incomplete adjuvant were carried out. Three days after the 27 final injection, 20 m l of blood was drawn and the serum was assessed for reactivity wi th OprP by incubating dilutions wi th nitrocellulose bound purified protein. In order to remove nonspecific antibodies, aliquots of serum were preabsorbed wi th washed intact E. coli cells (DH-5aF' or CE1248). The pellets obtained after centrifuging 10 m l of an overnight culture of E. coli were resuspended i n 1 m l of the anti-OprP rabbit serum. After being incubated for 45 min at room temp on an end-over end-shaker, the cells were pelleted and the serum was reabsorbed against a fresh batch of cells. The preabsorbed serum was divided into 50 ul aliquots and stored at -20 °C. The anti-malarial epitope monoclonal antibody p.f. 2A.10 was obtained from R. Wir tz (Dept. of Entomology, Walter Reed Institute of Research, Washington, DC) and was prepared as previously described (Wong et al., 1995). Alkaline phosphatase conjugated goat anti-rabbit IgG and goat anti-mouse IgG were purchased from Bio-Rad Industries and fluorescein isothiocyanate conjugated secondary antibodies were purchased from Gibco BRL. 4. Indirect immunofluorescence The detection of surface exposed proteins and epitopes was accomplished using the method of Hofstra et al. (1979). Aliquots of cells (100 uls) grown overnight were pelleted, washed wi th PBS and incubated with a 1/100 dilution of primary antibody (polyclonal anti-OprP or monoclonal anti-malarial epitope antibodies) in P B S / 1 % BSA for 1.5 hr at room temperature. The cells were then washed 2x wi th 1 m l of PBS and resuspended in 100 (il of 1% B S A / P B S containing a 1/20 dilution of fluorescein isothiocyanate (FITC) conjugated secondary antibody (goat anti-rabbit or goat anti-mouse). After a 1.5 h incubation at room temperature cells were washed 2x, resuspended i n 1% B S A / P B S and dried onto poly-lysine coated slides (Sigma). Fluorescence was monitored using a Zeiss microscope fitted wi th a halogen lamp and 28 filters set for emission at 525 nm. Cells were photographed using Kodak T-Max 400 black and white print film. Construction of plasmid pAS27 The oprP gene was subcloned into pTZ19U in two separate steps. First a vector lacking EcoRI and Psfl recognition sites was created by ligating a 900 base pair (bp) Scal/Hindlll fragment from pUC18 to a 1.7 kilobase pair (kb) Scal/Hindlll fragment from pUC19 to form the vector pUC18/19. A 700 bp Hindlll/Hindlll fragment from pRSP-3 encoding the first 95 amino acids of OprP was gel purified and cloned into the Hindlll site of pUC18/19 (Figure 1). The resultant plasmid was then digested wi th EcoRI and Psfl, releasing a 27 bp fragment. The vector was then ligated to a 27 bp synthetic oligonucleotide which replaced the excised fragment but contained a single basepair change which destroyed the Pstl recognition site while maintaining the encoded amino acid sequence. After verifying the proper orientation of the oligonucleotide by D N A sequencing, the HindUl/HindUl fragment was excised and gel purified. A 1.2 kb Pstl/Pstl fragment from pRSP-3 encoding amino acids 35-406 of OprP was cloned into the Pstl site of pTZ19U. This construct was then digested wi th HindUl and ligated to the gel purified 700 bp Hindlll/Hindlll fragment. Finally, the plasmid was digested wi th Pstl and ligated to an 18 bp synthetic oligonucleotide which encoded the last four amino acid residues of OprP and contained a single basepair change which destroyed the Psfl recognition site. Proper incorporation of the oligonucleotide was again verified by D N A sequencing. This final construct was named pAS27 (Figure 2). pRSP-3 ( own construct) EcoRl HiBtnH,.PstlHindin — L - H I R U > — Pstl Hiiiilll digest Hind in^ l s t lHMH Hindlll 4 Huidlll digest Hiudlll ligation EcoRl , ,PstljlmdJII EcoRl* * EcoRl pjfi 1=1 ligation EcoRl Hindlll, * Hindlll Hindlll digest + fragment isolation Hi«tdin Hindlll ligation Hindni —i r-Hindlll _1 Pstl Xbal Pstl digest EcoRL. I * Hll Hindlll  indlTI * Xbal Figure 1. Procedure used in the creation of plasmid pAS27 E c o R I K p n l B a m H l l Figure 2. Restriction map of plasmid pAS27. 31 Linker insertion mutagenesis Insertion of 12 bp linkers into the oprP gene was accomplished as described previously (Wong et al., 1993). Plasmid pAS27 was partially digested by the three frequently cutting blunt end restriction endonucleases Alul, Rsal and Thai i n the presence of varying concentrations of ethidium bromide (10-50 ng/ul). Prior to the large-scale plasmid digestion and isolation, small trial digestions were performed with various concentrations of ethidium bromide in order to establish conditions optimal for the isolation of singly cut plasmid DNA. The linker insertion mutant PL1 was created using Eco RV digested plasmid pAS27. After digestion, cut plasmid was gel purified and ligated in equimolar ratios to a 1.3 kb Hindi fragment which was isolated from the kanamycin resistance plasmid pUC4KAPA. The ligation products were transformed into E. coli strain DH5-aF' and clones were selected for both ampicillin and kanamycin resistance. The resultant colonies were picked onto LB agar plates and screened for OprP expression by immunoblotting using preabsorbed anti-OprP (trimer specific) polyclonal rabbit serum. Plasmids were isolated from OprP negative clones and digested with Pstl, releasing the kanamycin resistance gene and leaving a 12 bp residual linker sequence at the initial insertion site. The religated linker mutant plasmids were transformed back into DH5-aF' and characterized by restriction enzyme digestion. Clones displaying unique digestion patterns were selected for D N A sequencing . Malarial epitope insertion The linker-mutant plasmids were linearized by digestion with Pstl and ligated to 47 bp synthetic oligonucleotides encoding the malarial epitope at molar ratios of 1:100. The oligonucleotides were engineered so as to introduce a unique restriction enzyme recognition site (Sphl or Xbal as underlined below) when inserted into the plasmid. 32 Separate oligonucleotides were utilized for each reading frame to ensure correct translation. This resulted in the following inserted sequences: CCG A A C GCC A A C CCG A A C GCC A A C CCG A A C GCC GGG CAT G C A or AC CCG A A C GCC A A C CCG A A C GCC A A C CCG A A C G C A TGC A or G A A C GCC A A C C C A A A C GCG AAT CCG AAT GCT C T A G A C TTG C A . Clones were initially screened by colony immunoblot analysis using the anti-malarial epitope specific monoclonal antibody p.f. 2A.10. The plasmid D N A of selected immunopositive clones was isolated and sequenced in order to verify the proper incorporation of the inserted epitope. Site-directed lysine mutagenesis The oprP substitution mutants were created using a two-step recombinant PCR method (Higuchi et al., 1988; Ho et al., 1989) with the oprP containing plasmid pAS27 used as the template. Mutagenic oligonucleotides (27-mers) contained mismatches which corresponded to a substitution mutation in the encoded amino acid sequence (Table II). Reactions (100 ul total volume) contained 100 ng plasmid template (pAS27), 100 pg of primer, 200 mM each of the four dNTPs, 1-2 units Vent D N A polymerase (New England Biolabs) and IX Vent D NA polymerase buffer. The D N A was then subjected to 26 rounds of an amplification reaction which consisted of a 1 min denaturation step at 96 °C followed by a 1 min combined annealing/synthesis step at 72 °C. The amplified fragments were purified from agarose gels and the purified fragments were subcloned into appropriately digested and purified plasmid pAS27. Resultant colonies were screened for the presence of the desired construct by restriction 33 Table II. Codon changes of oprP substitution mutants. Plasmid Codon Amino acid wild-type mutant wild-type mutant pOPE13 AAG GAG Lys 1 3 Glu 1 3 pOPE15 AAG GAG Lys 1 5 Glu 1 5 pOPE25 AAG GAG Lys 2 5 Glu 2 5 pOPE30 AAG GAG Lys 3 0 Glu 3 0 pOPE74 AAG GAG Lys 7 4 Glu 7 4 pOPE109 AAG GAG Lys 1 0 9 Glu 1 0 9 pOPE121 AAG GAG Lys 1 2 1 Glu 1 2 1 pOPE126 A A A GAA Lys 1 2 6 Glu 1 2 6 pOPE181 AAG GAG Lys 1 8 1 Glu 1 8 1 pOPG74 AAG GGG Lys 7 4 Gly 7 4 pOPG121 AAG GGG Lys 1 2 1 Gly 1 2 1 pOPG126 A A A GGA Lys 1 2 6 Gly 1 2 6 pOPQ121 AAG CAG Lys 1 2 1 Gin 1 2 1 pOPQ126 AAA CAA Lys 1 2 6 Gin 1 2 6 34 enzyme mapping. Plasmids displaying the correct restriction digest maps were sequenced in order to verify the presence of the desired mutation as well as to ensure that no errors in the D N A were incorporated during the amplification. Isolation of outer membrane fractions Outer membranes were purified from both P. aeruginosa and E. coli by a method which utilized a two-step sucrose density gradient. After overnight growth, cells were pelleted by centrifigation at 3000 g for 20 min. The bacteria were resuspended i n 20 % sucrose containing 10 mM Tris-HCl (pH 8.0) and 50 ug/ml deoxyribonuclease and broken by French pressing. Unbroken cells were pelleted by centrifugation and the supernatant was collected and stored on ice. Two-layered sucrose density gradients (50%/70%) were set up i n transparent, disposable centrifuge tubes (Beckman Instruments Inc., Palo Alto CA, USA) and the supernatant was layered on top of the gradient. The tubes were then placed into a Beckman SW-41 swinging bucket rotor and spun at 100,000 g for 6-12 hours at a constant temp of 4 °C. After centrifugation, the outer membrane fraction, which migrated to the interface of the two different sucrose densities, was collected by poking a heated needle into the bottom of the tubes and draining the contents into collection vessels. The membrane fractions were diluted with distilled H 2 O to reduce the sucrose concentration to below 20 % and were then pelleted by centrifugation at 200,00 g for 1 hour. Detergent solubilization of OprP The nonionic detergent n-octyl-polyoxyethylene (octyl-POE) was used as previously described (Garavito and Rosenbusch, 1986) to selectively solubilize OprP contained in isolated outer membrane fractions. Pelleted outer membranes were 35 resuspended in 0.5 % octyl-POE containing 10 m M Tr is -HCl (pH 8.0) and incubated at 37 °C for 1 hour. The membranes were then pelleted by a 1 hour centrifugation and the supernatant was collected and stored on ice. The pellet was resuspended i n 3% octyl-POE containing 10 m M Tris -HCl (pH 8.0) and incubated at 37 °C for 1 hour. The membranes were again pelleted by centrifugation, the supernatant was saved and the pellet was resuspended in 3% octyl-POE/50 m M E D T A containing 10 m M Tr i s -HCl (pH 8.0) and incubated at 37 °C for 1 hour. The membranes were then centrifuged a final time, the supernatant collected and the insoluble pellet was resuspended i n distilled H2O. The three detergent solubilized fractions, along with the resuspended pellet were subjected to SDS-PAGE on 12% polyacrylamide gels. After electrophoresis, the protein bands were visualized by either staining the gels with a 0.28% solution of Coommassie Brilliant Blue or by electrotransferring the proteins onto nitrocellullose membranes and incubating the membranes with an anti-OprP polyclonal antiserum. OprP was found to be predominantly solubilized in the 3% octyl-POE/50 m M E D T A supernatant fractions. FPLC purification of OprP OprP used in the production of anti-OprP rabbit antiserum was purified from isolated P. aeruginosa outer membranes by F P L C . The octyl-POE solubilized protein solutions were placed into dialysis bags with a M W cutoff of 30 k D and were then dialyzed overnight against 500 times volume 0.08% N,N-dimethyl-dodecylamide-N-oxide (LDAO) containing 10 m M Tris-HCl (pH) and 10 m M E D T A at 4° C. This was done in order to replace the octyl-POE with L D A O which possesses a lower critical micelle concentration. The dialyzed protein was then concentrated using Amicon centrifuge columns. The concentrated protein sample was filtered through a syringe filter wi th a 0.2 um cut off and the collected filtrate was stored on ice. 36 The protein sample was loaded on to a Mono-Q F P L C column (Pharmacia) which was attached to a Pharmacia F P L C system. After washing the column with the buffer A (0.08% N,N-dimethyl-dodecylamide-N-oxide (LDAO); 10 m M Tr is -HCl (pH); 10 m M EDTA) , OprP was eluted from the column by passing a linear N a C l gradient over the column. Fractions which were shown by U V detection at 280 nm to contain large concentrations of protein were analyzed by SDS-polyacrylamide gel electrophoresis. Those fractions that were found to contain predominately OprP were pooled and concentrated by filtration. The concentrated protein sample was reapplied to the Mono-Q column and subjected to a second round of purification as described above. Pooled fractions of protein obtained from this second passage over the column were found to be virtually free of contaminating proteins as determined by SDS-PAGE followed by Coomassie Blue staining of the gel. Purification of OprP from SDS-polyacrylamide gels Small amounts of mutant protein required for planar l ip id bilayer analysis were purified from preparative SDS-polyacylamide gels. Unheated detergent solubilized OprP was loaded on to 12% SDS-polyacrylamide gels and electrophoresed as described above. After electrophoresis, the vertical edge of the gel was removed and stained wi th Coommassie brilliant blue in order to visualize OprP. The stained slice of gel was then used as a guide to localize OprP on the unstained gel, and the band was excised wi th a razor blade. The protein was eluted from the gel slices by incubating the slices overnight in approximately 300 ul of 10 m M Tris -HCl (pH 8.0) containing 0.1% SDS. 37 Planar l i p i d bilayer experiments Analysis of the single-channel conductance, anion selectivity and phosphate specificity of mutant forms of OprP was accomplished using the planar lipid bilayer technique as previously described (Hancock and Benz, 1986). The main apparatus consisted of a Teflon chamber which contained two individual compartments. The wall separating the two compartments contained a small hole (0.1 mm-single channel conductance; 0.9 mm-anion selectivity) which permitted the free passage of bathing solution. Prior to use the, the area surrounding the hole was coated with a small amount of 2% oxidized cholesterol and dried with a hair dryer. The chambers were filled with equal volumes of bathing solution and calomel electrodes which were connected to a voltage source and a current amplifier were placed in either chamber. A membrane was painted over the hole by wiping a small amount of oxidized cholesterol over the hole with a plastic wand and the current was monitored using either a chart recorder (which was connected to a current amplifier) or a voltage meter. Once a stable base line had been established the experiment was commenced. Single-channel conductance The bathing solutions used contained 100 mM, 1 M or 3 M KC1. Nanogram quantities of purified protein were added to the bathing solution and increasing steps i n conductance were monitored with a chart recorder. At least 100 individual channel events were observed for each mutant protein examined. Anion selectivity When the ionic selectivity of the mutant proteins was measured the electrodes were connected directly to a current/voltage meter and the bathing solution was 50 mM KC1. Upon addition of the mutant proteins a voltage of 20 mV was applied and the 38 membrane current was allowed to increase approximately 100 fold before the applied voltage was turned off and the meter was switch to read voltage. A t this point 50 ul aliquots of 50 m M and 3 M KC1 were added to either compartment of the Teflon chambers. Changes in the zero-current membrane potential were recorded after each salt solution addition. The zero-current membrane potential of the proteins was calculated by using the Goldman-Hodgkin Katz equation. Phosphate ion specificity The effect that the substitution of individual amino-terminal lysine residues had on the phosphate-binding site of OprP was assessed by measuring the inhibition of conductance induced by the addition of small amounts of 1 M potassium phosphate to the bathing solution (0.1 M KC1). Equimolar concentrations of K H 2 P 0 4 and K 2 H P 0 4 were mixed to achieve a buffered p H of 8.0. The single-channel conductance of the individual mutant proteins in the presence of increasing phosphate ions was compared to the channel conductance of the same proteins i n phosphate-free bathing solution. One hundred individual channel events were measured for each protein examined. 39 RESULTS A. Structural analysis of OprP Chapter 1. Linker insertion mutagenesis of OprP 1.1 Introduction Since there is not a great deal of similarity between the amino acid sequence of OprP and the other well characterized members of the porin family, a method was needed which would differentiate between loop regions and those regions which formed the trans-membrane (3-barrel. The semi-random insertion of 12 base pair linkers into the genes of certain bacterial porins has previously been successfully used to generate topological models of these proteins. Subsequent crystallization studies have shown that this technique is able to predict the location of loop regions wi th a reasonably high degree of reliability. While this technique does not allow for the differentiation between internally-exposed and externally-exposed loop regions, the inclusion of a unique restriction enzyme site within the inserted linkers provides a convenient means by which to later insert a foreign epitope which could serve as a tag. This chapter describes the creation and characterization of 13 linker insertion mutants of OprP which were generated by both site-directed and semi-random mutagenesis. 1.2 Expression of OprP i n an E. coli background The OprP gene was subcloned into the phagemid vector pTZ19U as described i n the Materials and Methods section to create the OprP expression plasmid pAS27. This construct was created in and purified from the E. coli strain DH-5ocF'. After sequencing of the regions of the vector which were mutated so as to eliminate two Pst I sites, the purified plasmid was transformed into the porin deficient E. coli strain CE1248. Outer 40 membranes from E. coli DH-5aF'(pAS27) and CE1248(pAS27) were purified by centrifugation of cell envelopes through sucrose density gradients. The purified membranes were analyzed by SDS-polyacylamide gel electrophoresis and Western immunoblot analysis with anti-OprP specific polyclonal rabbit serum. The presence of OprP was detected in both backgrounds, but as expected, the porin deficient strain CE1248 expressed a slightly higher amount of OprP than did DH-5ocF'. For this reason CE1248 was chosen as the background strain to be used in all experiments involved i n the expression of the mutant forms of OprP (Figure 3) while the heartier DH5-otF' strain was used for the initial construction of all the mutants. Inclusion of IPTG in the growth media was not shown to increase the expression of OprP as determined by SDS-polyacrylamide gel electrophoresis. 1.3 Linker insertion mutagenesis of OprP at a unique Eco RV restriction enzyme site As an initial test of the usefulness of this technique, the kanamycin resistance cassette was ligated to Eco RV digested and gel purified pAS27. After digesting wi th Hinc II to release the kanamycin resistance gene the religated plasmid was purified and sequenced. The 12 base pair linker was found to be properly incorporated into the Eco RV site. The reading frame of the inserted linker was such that the resulting four amino acid sequence D L Q V was inserted into OprP. 1.4 Semi-random linker insertion mutagenesis of OprP The insertion of the 12 base pair linker into multiple sites of the oprP gene was accomplished by first purifying partially digested plasmid as described in Materials and Methods. The enzymes used in this procedure cut at four base pair recognition sites which were scattered throughout the sequence of pAS27. Since all clones which had the kanamycin resistance cassette inserted into the coding sequence,of the oprP gene would no longer be capable of protein expression, OprP negative clones were selected for by 41 A B Figure 3 . Western immunoblot of outer membranes of E. coli CE1248 containing OprP reacted with anti-OprP trimer specific (A) or monomer specific (B) rabbit antiserum.The sample reacted with the monomer specific antiserum was heated at 100° C for 10 min. Prestained molecular mass markers on the left were (from the top) 106, 80, 49.5, 32.5, 27.5 and 18.5 kDa. 42 immunobloting recombinants with anti-OprP (trimer-specific) rabbit serum. After a Hinc II digestion which served to release the kanamycin resistance cartridge, the religated plasmids were purified and subjected to restriction enzyme digestion analysis. The clones were grouped according to their restriction digestion patterns and representative plasmids were sequenced. Twelve unique linker insertion mutants were isolated by this method (Table III). 1.5 Expression of the linker insertion mutant forms of OprP in E: coli Competent E. coli CE1248 cells were transformed with each of the 13 OprP linker insertion mutant plasmids and the resulting transformants were used to inoculate 5 ml LB broth cultures which contained 50 ul/ml ampicillin to maintain the plasmid and 0.4% glucose to inhibit the expression of LamB. The cultures were grown overnight at 37 °C and cells from 100 ul aliquots were pelleted, lysed and electrophoresed on SDS-polyacrylamide gels. After staining the gel with Coomassie Blue, all 13 mutants were shown to be expressed to varying degrees (data not shown). In order to determine if these mutant proteins were properly localized to the outer membranes of the host cells, 100 ml cultures of each clone grown under the same conditions as described above were used to prepare isolated outer membrane fractions. The purified outer membranes were assessed for the presence of OprP by both SDS-polyacryamide gel electrophoresis and Western immunoblotting. The amounts of the mutant forms of OprP in outer membranes was found to mirror the amounts of these proteins in whole cells. Figure 4 depicts a Western immunoblot analysis of outer membranes containing the thirteen linker insertion mutants along with the wild-type OprP expressing plasmid pAS27. The negative control lane contained outer membranes purified from E. coli CE1248 containing the phagemid pTZ19U. Although all mutants were expressed to some degree, certain insertion sites appeared to be less capable of tolerating the presence of the inserted linker than other 43 Table in. Insertion sites of OprP linker mutants Plasmid Insertion site Amino acids inserted (amino acid) pAS27 p P L l 9 D L Q V pPL2 60 TCRS pPL3 75 PAGP pPL4 82 D L Q V pPL5 124 TCRS pPL6 148 D L Q V pPL7 154 TCRS pPL8 190 PAGP pPL9 208 TCRS pPLlO 226 TCRS p P L l l 287 TCRS pPL12 309 D L Q V pPL13 333 D L Q V 44 1 2 3 4 5 6 7 8 9 10 11 12 13TZ AS — OprP Figure 4. Western immunoblot of outer membranes of E. coli CE1248 containing OprP linker insertion plasmids pPLl-pPL13 (lanes 1-13 respectively) and positive and negative controls plasmid pAS27 (AS) and pTZ19U (TZ). The sample were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80,49.5, 32.5, 27.5 and 18.5 kDa. 45 sites. The OprP insertion mutants PL3, PL8, PL12 and PL13 were all expressed at significantly lower levels than that of the wild-type protein. A l l of the insertion mutants were shown to possess electrophoretic mobilities similar to that of wild-type OprP. Insertion of the linker into the PL2 site appeared to induce the production of a degradation product with an apparent molecular mass of approximately 30 kDa and two smaller products of approximately 23 and 25 kDa; insertion at the PL13 site appeared to induce the production of a single product with a molecular mass slightly less than that of the native protein. It is possible that the presence of the linker at these sites resulted i n a slight distortion in the folded structure of the protein which exposed a proteolytic site. However the insertion of these linkers could not have caused a profound distortion of the folding patterns since these proteins reacted with the trimer-specific polyclonal serum as well as the wild-type protein, as assessed by colony immunoblotting (data not shown). A number of the mutant proteins appeared to possess increased trimer stability. Insertion at sites PL5, PL7, PL9, PL10 and PL11 resulted i n mutant proteins which were partially resistant to heat denaturation in SDS. Normally, nondenatured trimers do not react with monomer-specific antiserum. The fact that the high molecular weight bands of these mutants were able to bind the antibody indicated that the heat treatment affected the secondary structure of the protein without causing the dissociation of the trimers into individual monomers. 1.6 Surface localization of the OprP linker insertion mutants To determine if the mutant forms of OprP were efficiently folded and incorporated into the outer membrane of the host cells, clones were analyzed by indirect immunofluorescence for surface exposure of the mutant proteins. Cells containing the pPL plasmids, pAS27 or pTZ19U were incubated with anti-OprP (trimer-specific) rabbit serum followed by a goat anti-rabbit fluorescein isothiocyanate-46 conjugated secondary antibody. The cells were applied to poly-L-lysine coated glass slides and viewed under a fluorescence microscope. Nine of the 13 linker insertion mutants fluoresced at levels comparable to those exhibited by the wild-type protein (Figure 5A). These same proteins were previously shown by Western immunoblotting to be efficiently expressed and localized to the outer membrane. Cells harbouring the other four linker insertion mutant plasmids did not fluoresce at perceptively higher levels than that of cells carrying the negative control plasmid pTZ19U (Figure 5B; Table IV). These four clones were also previously shown to be expressed at greatly reduced levels in the outer membrane in comparison to the wild-type protein. Therefore the absence of detectable surface exposure displayed by these mutant forms of OprP could be due to either an inability of the proteins to be properly incorporated into the outer membrane or the fact that the binding of antibodies to cells expressing such a small amount of surface-exposed protein did not lead to perceptively greater fluorescence than the non-specific binding of these same antibodies to OprP-negative cells. 1.7 Functional analysis of selected linker insertion mutants The majority of the mutant proteins generated by the linker insertion method appeared to suffer few detrimental effects on either expression or surface exposure due to the presence of the inserted sequence. These findings however did not exclude the possibility that the channels formed by these mutants could have been affected in some manner. To determine if the insertion of the linkers had any effect on the functional aspects of the mutant proteins, several of the insertion mutants whose expression appeared to be unaffected by the mutagenic procedure were purified from SDS-polyacrylamide gels as described in the Materials and Methods section and analyzed by the planar lipid bilayer method. 47 Figure 5. Indirect immunofluorescence of E. coli strain CE1248 containing plasmid pAS27 (A) or plasmid pTZ19U (B) reacted with anti-OprP (trimer-specific) antiserum. 48 Table IV. Expression and surface exposure of OprP linker mutant proteins Plasmid Protein expression Surface reactivity (by SDS-PAGE) pAS27 ++ ++ p P L l ++ ++ pPL2 ++ ++ pPL3 + -pPL4 ++ ++ pPL5 ++ ++ pPL6 ++ ++ pPL7 ++ ++ pPL8 + -pPL9 ++ ++ pPLlO ++ ++ p P L l l ++ ++ pPL12 + -pPL13 + -* As assessed by indirect irnmunoftuoresence using an anti-OprP specific antiserum 49 A l l of the mutant proteins tested were shown to be capable of forming stable channels when added to the salt solution bathing a lipid bilayer composed of oxidized cholesterol (Table V). The single-channel conductances of these proteins were not found to be significantly different from that of wild-type OprP; the variation was less than 10%. This would suggest that the sites at which these linkers had been inserted into OprP were not critical to maintaining the three-dimensional folding pattern of the protein. 1.8 Summary The insertion of 12 base pair linkers into the gene encoding OprP was used to differentiate between sections of this protein which were structurally flexible and sections which were relatively nonpermissive for the insertion of the linker. Four of the thirteen mutants created were shown to have been significantly affected by the mutagenesis. Their levels of expression were drastically reduced in the outer membranes of the host cells and there was no evidence to show that they were capable of being properly exported to and expressed at the cell surface. The reasons for the reduction i n expression of these proteins is unknown but it is possible that the presence of the linkers at these particular sites inhibited the progressive folding kinetics of these proteins which created the opportunity for various cellular proteases to degrade the proteins before they were able to be exported to the outer membrane. Since the most structurally demanding sections of OprP should consist of those involved i n forming the (3-barrel, it would not be unreasonable to assume that these four sites may well be located within sections of (3-strands. Nine of the insertion mutants were expressed at levels comparable to the wild-type protein in both the outer membrane and at the cell surface. The sites at which these mutations occurred must therefore not be involved with maintaining the folded structure of the mature surface localized protein. The sections of porins which have 50 Table V. Channel activity of selected OprP linker mutant proteins Plasmid Average single-channel conductance (nS) ± SD a pAS27 0.23 ± 0.04 p P L l 0.23 ± 0.04 pPL2 0.22 ± 0.04 pPL3 N D b pPL4 0.25 ± 0.04 pPL5 0.24 ± 0.04 pPL6 ND pPL7 ND pPL8 ND pPL9 0.23 ± 0.04 pPLlO 0.23 ± 0.04 p P L l l 0.24 + 0.05 pPL12 ND pPL13 ND a. Average of 100 single-channel events in picoSiemens ± S.D. b. N.D. = not determined 51 previously been shown to be most amenable to the insertion of foreign sequences are those that form the internal and external loop regions. It is possible that the reason the expression of these mutants was not noticeably affected was because the sites of these mutations were located in loop regions of OprP. 52 Chapter 2. Epitope insertion mutagenesis of OprP 2.1 Introduction The insertion of foreign epitopes has previously been used to determine the topological structure of a number of porins. In these studies, the surface exposure of the inserted epitopes was assessed by reacting whole cells expressing the epitope insertion mutant with an epitope-specific antibody. Those clones which reacted with the antibody were assumed to carry the epitope in a surface localized loop region. A comparison of the 3-dimensional crystal structure of proteins subjected to epitope insertion mutagenesis revealed that this method was able to accurately predict which sections of these proteins were exposed to the external surface. In order to gain further insight into the topological arrangement of wild-type OprP, an epitope from the circumsporozoite form of the malarial parasite Plasmodium falciparum was inserted into the unique Psf I restriction enzyme site contained within the linker of 12 of the 13 OprP linker insertion mutants. It was hoped that this procedure would both help to determine which sites were flexible enough to accommodate the insertion of a larger foreign peptide as well as aid i n the identification of specific sites which are located in the extracellular domain of OprP. 2.2 Insertion of the malarial epitope into the linker sites of OprP mutants Oligonucleotides which encoded the repeating amino acid sequence (PNAN) 2. 3 were synthesized and purified as described in the Materials and Methods section. The oligonucleotides were annealed and ligated to twelve of the OprP linker insertion mutant plasmids which had previously been linearized by digesting with Psf I. Clones were screened for expression of the malarial epitope by colony immunoblotting with the malarial epitope-specific monoclonal antibody p i . 2A.10. Plasmids purified from clones which were found to react with the epitope-specific antibody were assessed for 53 the presence of the inserted oligonucleotide by restriction endonuclease mapping. Those plasmids which presented the correct restriction pattern were sequenced by the di-deoxy terminator method. The results of the sequencing experiments demonstrated that the malarial epitope was correctly inserted into the Pst I site of all twelve of the selected linker insertion mutants. 2.3 Expression of the malarial epitope insertion mutant protein The effect that the insertion of the malarial epitope had on the expression of the twelve mutant proteins was assessed by Western immunoblotting of outer membrane fractions which were purified from cells harbouring the epitope insertion mutant plasmids. The nitrocellulose-bound outer membrane proteins were reacted with anti-OprP polyclonal antiserum. In order to optimize protein expression, all of the plasmids were mobilized into the porin deficient E. coli strain CE1248 which served as the background strain for all of the protein expression experiments. As seen in Figures 6 and 7 all of the epitope insertion mutant proteins were expressed in the outer membrane of the host cells. However the levels of expression of these mutants were, in most cases, reduced in comparison to the wild-type protein. The one noticeable exception was the mutant which contains the epitope inserted at the PL10 site. The insertion of the foreign amino acid sequence at this site did not appear to have any effect on the expression of the mutant protein. The three epitope insertion mutants whose linker mutant parents displayed the greatest reductions in expression (insertions at sites PL8, PL12 and PL13) also exhibited low levels of expression. In addition, two other epitope insertion mutant proteins (sites PL6 and PL7) were expressed at noticeably lower levels than their linker insertion parents. It may be that these sites are less flexible than those of the remaining seven mutants. A l l of the epitope insertion mutant proteins demonstrated slightly reduced electrophoretic mobilities compared to wild-type OprP. 54 1 2 4 5 6 7 8 9 10 11 12 13TZ AS OprP Figure 6. Western immunoblot of outer membranes of E. coli CE1248 containing OprP-encoding plasmids with malarial epitope inserted at sites PL1, PL2 and PL4 to PL13. Positive and negative controls were plasmids pAS27 (AS) and pTZ19U (TZ). The blots were reacted with anti-OprP (monomer-specific) antiserum. The samples were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80,49.5, 32.5, 27.5 and 18.5 kDa. 55 1 2 4 5 6 7 8 9 10 11 12 13 TZ AS Figure 7. Western immunoblot of outer membranes of E. coli CE1248 containing OprP-encoding plasmids with malarial insertions at sites PL1, PL2 and PL4 to PL13. Positive and negative controls were plasmids pAS27 (AS) and pTZ19U (TZ). The blots were reacted with the anti-malarial epitope-specific antibody pf 2A.10. The samples were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80, 49.5, 32.5, 27.5 and 18.5 kDa. 56 Insertion of the malarial epitope into linker site PL5 appeared to induce the production of two degradation products with approximate molecular masses of 25 and 30 kDa. Surprisingly, inclusion of the epitope at insertion site PL2 did not induce the production of a degradation product even though the insertion of the linker alone did (Figure 4). The increased trimer stability displayed by some of the linker mutant proteins was reversed after insertion of the malarial epitope into the linker sites. Western immunoblotting of the outer membranes from the epitope insertion mutants with the anti-malarial epitope-specific antibody was done to verify that the epitope was indeed expressed in the mutants prior to the analysis of their surface exposure in whole cells. If the mutant proteins did indeed contain the malarial epitope then the bands corresponding to OprP would bind to the anti-malarial epitope-specific antibody regardless of if the epitope was inserted in a periplasmically-exposed or a surface-exposed region of OprP. While the degree to which the mutant proteins reacted with the epitope-specific antibody was reduced i n comparison to their reactivity with the OprP-specific antibodies, a careful inspection of the alkaline-phosphatase stained membrane revealed that a band which bound to the anti-malarial epitope antibody was present in the outer membrane of every one of the epitope insertion mutants examined. The reason the anti-epitope-specific antibody failed to react with the mutant proteins to the same extent as the anti-OprP-specific antibodies was likely because the anti-epitope antibody consisted of a single monoclonal antibody which bound to a single epitope contained within the mutant proteins. The anti-OprP-specific antiserum contained a number of different antibodies which bound to a number of different epitopes distributed throughout the sequence of OprP. Therefore it would be expected that when this polyclonal antiserum was used as the primary antibody i n the Western immunoblot analysis the intensity of the reaction would be greater than when the anti-malarial epitope-specific antibody was used. The outer membranes of both clones carrying the 57 phagemid vector pTZ19U as well as those carrying plasmid pAS27 lacked a band which reacted with the anti-epitope-specific antibody. 2.4 Surface localization of the epitope insertion mutant proteins To determine if the insertion of the malarial epitope into the linker site of the twelve mutant proteins had any detrimental effects on the export of the proteins to the outer membrane, whole cells expressing the mutant proteins were examined using the method of indirect immunofluorescence. The specificity of the primary antibody was towards wild-type OprP. The degree to which whole cells expressing the epitope insertion mutant proteins bound the antibody would reflect the surface exposure of that particular mutant. In addition, the fact that the anti-OprP antibody used i n this assay only bound to the trimer form of the protein meant that a positive reaction reflected the proper trimerization of a given mutant. As shown in Table VI the reactivity of the individual clones with the anti-OprP antibody closely reflected the expression of these proteins i n the outer membranes as determined by Western immunoblot analysis. Cells expressing the five proteins which demonstrated significantly decreased expression levels in isolated outer membrane preparations did not appear to bind the antibody at the cell surface to a greater extent than did those clones carrying the negative control plasmid pTZ19U. As stated previously in Section 1.6, the lack of detectable surface exposure could be due either to the actual lack of surface exposed protein or to the insensitivity of this assay in discerning a difference between the specific binding of the antibodies to such a small quantity of protein and the non-specific binding of antibodies to the outer membrane proteins of the negative control. A l l the other seven proteins were shown by this method to be exported to the outer membrane at levels comparable to that of the wild-type protein. It would therefore seem likely that if the malarial epitope was contained in 58 Table VI. Expression and surface exposure of OprP malarial epitope insertion mutant proteins Surface reactivity* Insertion site Protein expression Anti-OprP Anti-malarial (by SDS-PAGE) polyclonal Ab epitope M A b none ++ ++ — PL1 ++ ++ -PL2 ++ :++ -PL4 ++ ++ -PL5 ++ ++ -PL6 + - -PL7 + - -PL8 + - -PL9 ++ ++ -PL10 ++ ++ ++ PL11 ++ ++ ++ PL12 + - -PL13 + - -* As assessed by indirect immunofluoresence using either an anti-OprP specific antiserum or the malarial epitope specific MAb p.f. 2A.10. 59 a surface exposed region of any of these seven mutant proteins, the cells expressing those proteins would bind the anti-malarial epitope-specific antibody at the cell surface. 2.5 Surface localization of the inserted malarial epitopes To determine if any of the sites at which the malarial epitopes were inserted into OprP faced the external surface of the cell, whole cells expressing the mutant proteins were subjected to indirect immunofluorescence with the anti-malarial epitope-specific monoclonal antibody p i . 2A.10 used as the primary antibody. Clones expressing wild-type OprP were used as the negative control. Of the twelve clones examined by this method, only two were found to bind the anti-malarial epitope-specific antibody at levels significantly higher than the negative control (Table VI). Cells expressing the mutant proteins with the epitope inserted at sites PL10 and PL11 exhibited high levels of fluorescence after incubation with the antibodies (Figure 8). These same mutants were found to be capable of reacting in whole cells with antibodies directed against the native trimer form of OprP (Table VI). It can therefore be assumed that the accessibility of these inserted epitopes to the antibody was due to the fact that they were actually located in surface-localized sites rather than to a distortion of the native conformation of the protein which may have caused the exposure of a section of the protein which would normally be inaccessible at the cell surface. None of the other ten mutant expressing clones bound the anti-malarial epitope-specific antibody to a greater extent than the clone expressing wild-type OprP. These results suggest that either the sites at which the epitope was inserted into these mutants was not accessible to the cell surface or that the structure of the epitope was somehow distorted by the surrounding amino acid sequence into which it was inserted. In the case of the poorly expressed mutant proteins, the lack of binding of the anti-epitope-specific antibody may have been due to the fact that the expression levels of these mutant proteins was below the detection sensitivity of this assay. Figure 8. Indirect immunofluorescence of E. coli strain CE1248 containing the OprP-encoding plasmid with the malarial epitope inserted at site PL 10. Cells expressing the mutant protein with the epitope inserted at site PL 11 displayed similar levels of fluorescence. The primary antibody used was the malarial epitope specific monoclonal antibody p.f. 2A.10. A l l of the other malarial epitope insertion mutant expressing cells well as cells expressing wild type OprP fluoresced at the same levels as the negative control in Figure 5B. 61 2.6 Summary A 15 amino acid peptide corresponding to an epitope of the malarial pathogen P. falciparum was inserted into the Psf I site of twelve of the thirteen linker insertion mutant forms of OprP which had previously been created (see Chapter 1). The purpose of this procedure was to both assess the permissiveness of the insertion sites for the insertion of a longer length of foreign sequence and to determine if any of these sites were located in surface-exposed regions. The presence of the epitope insertion mutants i n the outer membranes of host cells was determined by Western immunoblotting using either anti-OprP or anti-malarial epitope-specific antibodies. The expression of a number of these proteins was shown to be adversely affected after insertion of the epitope. This would suggest that the insertion sites of these specific mutants possess a limited degree of flexibility. The majority of the mutant proteins were expressed at levels comparable to that of wild-type OprP, suggesting that the insertion sites of these proteins were flexible enough to accommodate the presence of the foreign peptide. The fact that all of the epitope insertion mutant proteins bound the anti-malarial epitope-specific antibody indicated that each of these constructs did indeed contain the epitope and that it was inserted in both the correct orientation and in the proper reading frame. The results of the immunofluorescence experiments using the anti-OprP antiserum as the primary antibody mirrored those of the Western immunoblot analysis of isolated outer membranes; those mutant proteins which were present in the isolated outer membranes at greatly reduced levels compared to the wild-type protein were not shown to be present at detectable levels at the cell surface of whole cells. Indirect immunofluorescence of whole cells expressing the epitope insertion mutant proteins was used to determine if any of the insertion sites were located i n regions of the protein which are accessible to the antibodies at the cell surface. The results of these experiments indicated that two insertion sites of the seven efficiently 62 expressed mutants were surface exposed. While these results could be interpreted as meaning that only a low proportion of the sites into which the epitope was inserted were surface localized, a negative immunofluorescence result does not necessarily mean that a particular insertion site was not exposed to the cell surface. Both steric hindrance imposed by the presence of other sections of the protein as well as a lack of local flexibility could have resulted in a negative result for a particular mutant even though the epitope was indeed inserted into a site which faces the cell surface. 63 Chapter 3. Modeling of the OprP monomer 3.1 Introduction The results of both the linker insertion and epitope insertion mutagenesis of OprP were used to create a topological model of the structure of OprP. Certain data regarding the structure of previously examined porins were used to help create a framework into which these results could be placed. The general amphipathic nature of the fJ-strand regions as well as the actual number of trans-membrane strands exhibited in the structural models of the other porin proteins was maintained i n the proposed model of OprP. The orientation of the various loop regions of the OprP model were partially determined by the results of the immunofluorescence experiments with the anti-malarial epitope-specific antibody. After establishing an initial model for the folding pattern of the OprP monomers, this model was tested by the site-directed insertion of the malarial epitope into the proposed site of a surface-exposed loop as well as by the deletion of sections of another proposed surface-exposed loop. 3.2 Structural implications of linker insertion mutagenesis The analysis of a number of bacterial porins by the method of linker insertion mutagenesis has revealed the fact that the regions of these proteins which are the most amenable to the insertion of foreign peptides are those which are located i n the loop regions. These same studies showed that insertions occurring in the trans-membrane (3-strand regions of these proteins served to inhibit expression and/or export to the outer membrane. With these findings in mind, the amino acid sequence of OprP has been overlaid with the results of the linker insertion mutagenesis experiments (Figure 9). The insertion sites of mutants whose expression levels were significantly reduced by the presence of the four amino acid insert were assumed to be located in a trans-membrane 64 10 (ff) G E K D T T I 1 G A D T T V T Q 8 F K L Q Q B IL Q IA D IY Q IR _fj D G Y E L Y A R 2 ® , Y T G G T A Y R D ~W~ K Q I N Y D nsr R N Y F D T Y F R G F K L N V P N F G T Y T V s A IE D F J_J G A S D L E K A T 8 <D K W V T A L E R N W D N V • (3 N N N E S F V 8 G N V 8 7 ® A Q I G T D G D S V K R Y N L R G V F ® P L H Q M R P R I R T D V A S D E L D 9 ® Y A Y Q L G L H Y V N P G V 8 T N G G N D A G 8 N G N R G L W D A E K 11 ( § ) V T R R L Y E A Q A 8 D R E D L K A S G Y F A G L A W A G E L G G j l 2 | <g) R W V L S Y V V I N D D E V K I 8 D N 3| w A G 1 8 A T R E V G D A K G K T R E K N E N A 8 1 K D T K A K A V L Y G N V A N S W V Y K A V N E A P K I N G D D 8 G D G L IV Ml R L IQ Y IV F L D G A K F D Figure 9. Topological model of the monomeric structure of OprP based on linker and epitope insertion. Boxes enclose the proposed transmembrane p-strands. Circles indicate insertion sites PL 1 to PL 13 (number 1 to 13 respectively). 65 p-strand. A l l of the other insertion sites were placed i n either periplasmically-exposed or surface-exposed loops. A topological model of OprP which was based on the p-turn prediction method of Paul and Rosenbusch has previously been proposed (Martin, N., unpublished results). While the model proposed i n the current study differs considerably from this earlier version, the location of certain p-turn regions has been retained. In particular, the regions corresponding to the first surface-exposed loop, the second periplasmically-exposed loop and the fourth periplasmically-exposed loop in Figure 9 are also located in loop regions of the previously proposed model (Appendix-Figure 22). In addition, certain sections i n the current model which form the P-strands are also found to compose the P-strands of the previous model. The placement of these strands was partially based on the amphipathic nature of the amino acid sequences contained i n those sections of the protein (Appendix-Figure 23). 3.3 Structural implications of epitope insertion mutagenesis The insertion of foreign epitopes into the amino acid sequences of several outer membrane proteins has successfully been used i n the past to construct three-dimensional structures of the folding patterns of these proteins. Provided that the mutagenesis does not induce substantial changes i n the folding, this procedure can be used to differentiate between those loops which are exposed to the cell surface and those which may face the periplasmic space. The insertion of the malarial epitope into OprP resulted in the identification of two regions of the protein which are likely to be exposed to the cell surface (sites PL10 and PL11 in Figure 9). These sections of OprP were flexible enough to tolerate the insertion of the foreign peptide without manifesting profound reductions in expression. The mutant proteins were also shown to have retained their native conformation at the cell surface as determined by indirect 66 immunofluorescence experiments with the anti-OprP (trimer-specific) polyclonal antiserum. Although only two of the linker insertion sites were identified as being exposed to the cell surface, there was reason to believe that epitopes inserted into certain other sites of OprP were also contained in surface exposed loops, despite the fact that they were unable to bind to the anti-malarial epitope antibody. The accessibility of these epitopes to the outer surface may have been limited if they were located at the interface between a trans-membrane (3-strand and a surface-exposed loop. The junction between these two structural features would likely be flexible enough to accommodate the insert without a substantial effect on expression. However, the epitope contained at this site might not be extended far enough from the cell surface to allow for the efficient binding of the anti-malarial epitope-specific antibody. Insertion sites PL2, PL5 and PL9 have all been assigned to surface-exposed loop regions. The lack of accessibility of these epitopes to the antibody could be due to the reason stated above. The lack of binding of the antibody to the epitope inserted into site PL5 can not be explain by the proposed location of the epitope within the loop. This site has been postulated to occupy a region located in the center of the third surface-exposed loop. Theoretically, an epitope located at this site should therefore be capable of binding the anti-epitope-specific antibody when expressed i n the outer membrane of whole cells. However the crystal structures of a number of bacterial porins have shown that the third surface-exposed loops of these proteins are folded such that they are at least partially buried within the central channel of the native monomers. Epitopes inserted into these loops would therefore not necessarily be accessible to the outer surface despite the fact that the loops are actually facing the external surface of the cell. 3.4 Testing of the OprP topological model: i) Deletion mutagenesis of the proposed fifth surface-exposed loop 67 In order to test the validity of the proposed topological model of OprP, two sections of the proposed fifth surface-exposed loop were both individually and concomitantly deleted and the mutant proteins were assessed for expression in the outer membrane and surface exposure in whole cells. This region was originally placed in a surface-exposed loop because of the apparent surface-exposure of a malarial epitope which was cloned into this site. The section between insertion sites PL9 and PL10 (17 amino acids) and the section between insertion site PL10 and the site in the protein which corresponded to a downstream Ssf II restriction enzyme site in the D N A sequence (16 amino acids) were deleted from the protein. In addition, the entire section of the protein encompassing the region between insertion site PL9 and the Ssf 71 site was also deleted. These three deletion mutants were expressed in E. coli CE1248 and the outer membranes were purified and analyzed by Western immunoblotting with anti-OprP-specific antiserum. As shown in Figure 10, the mutants which resulted from the individual deletion of the sections of protein sequence which border the PL10 insertion site were expressed at levels comparable to the wild-type protein. Both of these mutants also appeared to possess slightly increased electrophoretic mobilities compared to wild-type OprP. A faint band which bound the anti-OprP antibodies was shown to be present i n the lane containing the mutant protein which was generated from the deletion of both sections of OprP. The intensity of this band was significantly reduced in comparison to either the wild-type protein or the other two deletion mutant proteins. The electrophoretic mobility of this protein was also considerably greater than that of wild-type OprP. The surface-exposure of these three mutant proteins was assessed by indirect immunofluorescence with the anti-OprP antiserum. The two mutants (OprP APL9-PL10; OprP APLIO-Ssf) which did not display appreciable decreases i n expression as determined by Western immunoblotting were shown to be properly exported and trimerized at the cell surface. However the clone expressing the mutant with the largest 68 A B C D Figure 10. Whole cell Western immunoblot of E. coli CE1248 expressing the OprP mutants deleted for sections of the proposed fifth surface-exposed loop. Lane A corresponds to the mutant OprP APL9-PL10, lane B corresponds to the mutant OprP APLIO-Ssr and lane C corresponds to the mutant OprP HsPlB-Sstl. Lane D corresponds to wild-type OprP. The samples were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80, 49.5, 32.5, 27.5 and 18.5 kDa. 69 deletion (OprP APL9-Ssf) did not bind the anti-OprP specific antibodies to a significantly higher degree than did the cells which served as the negative control. 3.5 Testing of the OprP topological model: ii) Epitope insertion mutagenesis of the proposed seventh surface-exposed loop As a further test of the topological model, the malarial epitope was inserted into a site which has been proposed (based on the linker-insertion mutagenesis) as being located within the seventh surface-exposed loop of OprP. The outer membranes from cells expressing this mutant protein were purified and examined by Western immunoblot analysis with both anti-OprP and anti-malarial specific antibodies. The protein was expressed at levels comparable to the wild-type and the malarial epitope inserted into this site was expressed in the outer membranes (data not shown). Whole cells expressing this mutant protein were subjected to indirect immunofluorescence with the anti-malarial epitope-specific antibody i n order to determine if this site was accessible to the external surface of the cell. In comparison to cells expressing the wild-type protein, cells expressing the epitope insertion mutant fluoresced strongly when examined by immunofluorescence microscopy (Figure 11). This result indicated that the site at which this epitope had been inserted was contained in a surface-exposed loop of the OprP monomer. 3.6 Summary The results of the linker and epitope insertion experiments performed on OprP were used to generate a topological model of the structure of the native monomer. This model was partially based on the findings of previous studies performed on the structure of other bacterial porins. The validity of the proposed model was then tested by performing additional mutagenesis of the protein. The results of these experiments were found to be consistent with the proposed structure of OprP (Figure 12). Figure 11. Indirect immunofluorescence of E. coli strain CE1248 containing the OprP-encoding plasmid with the malarial epitope inserted in the proposed seventh surface-exposed loop. The primary antibody used was the anti-malarial epitope specific antibody p.f. 2A.10. The negative control fluoresced at levels similar to Figure 5B. 71 G D T E K D T T S F K L Q Q R G E L Y A R R Y T G G T A Y R D F D T Y F R G F K L N V P N F G T T V T G w Y K T Y V Q s A N E Y D D F L Y 3 G R A N S V D L E K A T S s K W V T A L E R N L T Y D I A D W V K D N V G N N N E S F V s G S L F A M G G _¥_ V S S A Q I M ^ ° N M : R N R * T s D : v s S s |q D V T D G D K R G N L V N P G N S ^ T s N s N N G : N - G :» » G F G G S s A V E G L W K D A E K V T R R L Y E A Q A S F A G L A W A G E L G W D S ® D R E D L K A S G Y Y A Q L A Y T I T G E P R ® S V A V I N D D E V K I S D Y R Y F L E W A G I T R E V G D A K G K T R N A S I K D T K A K A V L Y G N V A N S W V Y K A V N E A G N T G E K N E P K L D G A K F D N G D D S G D G Figure 12. Topological model of the monomeric structure of OprP. The sections of the fifth surface-exposed loop which were deleted are marked with hatched lines. Circles indicate the sites at which the inserted malarial epitopes were shown to be surface-exposed. 72 B. Functional analysis of OprP Chapter 4. Site-directed mutagenesis of specific lysine residues of OprP 4.1 Introduction It has been previously determined by chemical modification experiments that lysine residues present in both OprP as well as the anion-selective E. coli porin PhoE play a role i n determining the specific channel characteristics exhibited by these porins. Modifications which destroyed the positive charges of the lysine residues resulted in porins which displayed significantly lower single-channel conductances i n comparison to the wild-type counterparts. In addition, the lysine-specific modification of OprP was shown to cause a loss i n the phosphate specificity of the modified proteins. Since the amino acids which have been identified in other porins as being important to the establishment of specific channel characteristics tend to be located in the amino-terminal half of these proteins, the lysine residues contained in the amino-terminal half of OprP were subjected to mutagenesis using a recombinant PCR method. The mutant proteins were purified and analyzed by the planar lipid bilayer method i n order to determine if there had been any alteration in the channel function. After the identification of three specific lysine residues which appeared to play a role i n defining the channel characteristics of OprP, a triple mutant in which all three of these lysine residues were substituted with glutamate residues was created and the mutant protein was purified and analyzed in the same manner as the single-substitution mutant proteins. In addition to the role that the amino-terminal lysine residues of OprP were proposed to play in determining the channel characteristics of this porin, it was proposed that certain lysine residues located in the carboxy-terminal half of this protein might be important for the export of the newly synthesized protein to the outer 73 membrane of the host cell. The individual or concomitant mutagenesis of two lysine residues which are contained in a four amino acid motif of a mitochondrial porin was shown to adversely affect the export of this protein (Smith et al., 1995). This particular motif occurs in the carboxy-terminal half of a number of different porins. Two such four amino acid sequences are located in the carboxy-terminal half of OprP and thus, it seemed possible that such residues might similarly be involved in the export of this protein to the outer membrane of host cells. In order to test this hypothesis, the first lysine residue of each of these two sequences was substituted with a glutamate residue by the same procedure used to mutate the amino-terminal residues. The mutant proteins were expressed i n E. coli and their presence in the outer membrane was assessed by Western immunoblotting of outer membranes which had been subjected to a selective solubilization with octyl-POE. Whole cells expressing the substitution mutants were also assessed for the presence of surface-exposed protein by indirect immunofluorescence. 4.2 Site-directed mutagenesis of lysine residues The substitution of lysine residues was accomplished using a two-step PCR method as described in the Materials and Methods section. Mutagenic oligonucleotides which were homologous to the wild-type oprP sequence save for either one or two base pair substitutions were synthesized and used as primers for the primary amplification step. Each pair of primary PCR products was then purified, combined and amplified by a second round of PCR using primers homologous to sections of the gene which surround the mutation site. The mutagenized fragments were subcloned back into oprP, the religated plasmids were purified and the entire amplified segment was sequenced. A l l of the substitution mutants were found to contain the correct nucleotide substitutions and to be free of any random errors. 74 4.3 Expression of amino-terminal lysine substitution mutant proteins The plasmids encoding the lysine substitution mutant forms of OprP were used to transform E. coli CE1248 and the presence of the mutant proteins was determined by Western immunobloting of isolated outer membranes with the anti-OprP-specific antiserum. A l l of the mutants appeared to be expressed at levels equal to that of wi ld -type OprP (Figure 13). The proteins were purified from the outer membrane fractions by selective detergent solubilization followed by elution from SDS-polyacrylamide gels. 4.4 Single-channel conductance of amino-terminal lysine substitution mutant proteins In order to determine whether the substitution of individual lysine residues wi th glutamates had an effect on the conductance of the channels formed by these mutant forms of OprP, the single-channel conductance in different concentrations of KC1 was determined for each of the mutant proteins using the planar l ip id bilayer method (Table VII). The average conductance of six of the mutant proteins (Lys 1 3, Lys 1 5 , Lys 2 5 , Lys 3 0 , Lys 1 0 9 and Lys 1 8 1) was similar to that of wild-type OprP at all tested salt concentrations. In contrast, three of the mutant proteins displayed distinctly altered channel characteristics. In 1 M KC1, the Lys 7 4 and Lys 1 2 6 mutants exhibited levels of conductance that were approximately one-half of that of the wild-type protein, while the Lys 1 2 1 mutant possessed a conductance of approximately one-third that of wild-type OprP (Figure 14). The conductance of the channels formed by these three mutants as wel l as w i ld -type OprP was plotted as a function of increasing salt concentration (Figure 15). As shown previously, the single-channel conductance of wild-type OprP plateaued as the salt concentration approached 1 M indicating the presence of an anion-binding site. However, both the Lys 7 4 and the Lys 1 2 1 mutants formed channels that displayed linear concentration-conductance relationships at up to 3 M KC1, suggesting the free diffusion of the CI - ions through the channel. The Lys 1 2 6 mutant channel conductance, although 75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Figure 13. Whole cell lysate Western immunoblot of outer membranes of E. coli CE1248 expressing the OprP substitution mutant proteins. The lanes are labelled as follows: I) Lysl3->Glu, 2) Lys15-»GIu, 3) Lys 2 5->Glu, 4) LyS30->Glu, 5) L y s ^ G l u , 6) Lys 7 4->Gly, 7)Lys109-»Glu, 8) Lys 1 2 1-K31u, 9) Lys121-»Gly, 10) Lys121-»Gln, II) Lys 1 2 6->Glu, 12) L y s 1 2 6 ^ G l y , 13) Lys 1 2 6->Gln, 14) pAS27-wild-type OprP. The samples were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80, 49.5, 32.5, 27.5 and 18.5 kDa. 76 Table VII. Single-channel conductance of OprP Lys->Glu mutant proteins OprP mutation Average single-channel conductance (pS)a' 0 . 1 M K C 1 1 M K C 1 3 M K C 1 Wild type 103 ± 45 230 ± 42 261 ± 5 1 L y s 1 3 ^ G l u 78 ± 2 5 194 ± 3 3 220 ± 42 Lys 1 5 -+Glu 74 ± 2 4 207 ± 46 N . D b -L y s 2 5 ^ G l u 88 ± 3 0 210± 51 236 ± 64 L y s 3 0 ^ G l u 93 ± 4 0 206 ± 60 308 ± 60 L y s 7 4 ^ G l u 41 ± 14 116 + 34 363 ± 8 3 L y s 1 0 9 ^ G l u 85 ± 14 221 ± 3 4 N.D. L y s 1 2 1 ^ G l u 10 ± 2 74+11 248 ± 36 L y s 1 2 6 ^ G l u 25 ± 6 99 ± 17 103 ± 23 Lys 1 8 1 ->Glu 79 ± 14 220 ± 22 N.D. a. Average of 100 single-channel events expressed in picoSiemens ± S.D. b. N.D.= not determined Figure 14. Single-channel events of wild-type OprP and Lys 1 2 1 —»Glu mutant porin. Chart recorder tracings of the increasing single-channel steps resulting from the incorporation of A) wild-type OprP and B) the Lys 1 2 1 -»Glu substitution mutant channels into lipid membranes. The bathing solution was 1 M KC1 (pH 6.0). 78 Figure 15. KC1 concentration-conductance relationship of selected OprP Lys-+Glu mutants. The single-channel conductance of three of the Lys—•Glu OprP mutants along with wild-type OprP was determined in a number of KC1 concentrations and the values were plotted as a function of the salt concentration. (•-OprP, 0-Lys 7 4 -*Glu, • -Lys 1 2 1->Glu, •-Lys 1 2 6->Glu) 79 significantly lower than wild-type OprP, tended to follow the same pattern as the wild-type protein, with the conductance reaching a plateau as the salt concentration approached 1 M. 4.5 Anion selectivity of amino-terminal lysine substitution mutant proteins It has been shown previously that the lysine-specific chemical modification of OprP resulted in a substantial decrease i n the anion selectivity of the channels (Hancock and Benz, 1986). Unmodified OprP demonstrates a more than 100-fold preference for anions over cations as measured by the zero-current membrane potential method. The chemically modified forms of OprP were shown to possess levels of anion selectivity that were far below that of the wild-type protein. It was therefore assumed that the modification of the lysine residues was responsible for the decreased anion selectivity of the chemically-treated proteins. In order to determine if any of the nine amino-terminal lysine residues of OprP play a role in determining the ion selectivity of this protein, the lysine substitution mutants were subjected to analysis using the zero-current membrane potential method. As shown in Table VIII, the anion selectivity of the nine substitution mutants varied somewhat from that of the wild-type protein. However, these variations could not be correlated with the previously observed differences i n single-channel conductance (Table VII). Even the three mutants which demonstrated distinctly different average single-channel conductances compared to the wild-type protein did not demonstrate any substantial decreases in anion selectivity. In addition, not only were all of these mutants shown to possess a distinct preference for transporting anions over cations, but the selectivity displayed by these mutants was significantly greater than the anion selectivity of OprP which had been subjected to lysine-specific chemical modification (Hancock and Benz, 1986). Table VIII. Anion selectivity of OprP Lys->Glu substitution mutant proteins Zero-current membrane potential OprP mutation (Vmaxf Wild type -36.8 ±0.08 L y s 1 3 ^ G l u - 39.2 ± 0.03 Lys 1 5->Glu - 43.5 ± 0.05 L y s 2 5 ^ G l u - 4 1 . 0 ± 0 . 1 7 L y s 3 0 ^ G l u - 30.8 ± 0.40 L y s 7 4 ^ G l u -42.0 ±0.10 L y s 1 0 9 ^ G l u -39.7 ±0.58 Lys 1 2 1 -+Glu -37.3 ±0.25 Lys 1 2 6 ->Glu -36.7 ±0.08 Lvs 1 8 1 ->Glu N.D. b ' a. Maximum potential difference obtained with a 10-fold ion gradient across the membrane ± S.D. (average of three experiments) b. N.D.= not determined 81 4.6 Phosphate specificity of amino-terminal lysine substitution mutant proteins To determine whether any of the Lys—>Glu substitutions had an effect on the Fi binding site of OprP, the ability of small concentrations of phosphate ions to inhibit the single-channel conductance of the mutant proteins was measured. The single-channel conductance of each mutant protein in 0.1 M KC1 was determined prior to the addition of phosphate ions (Table VII). Increasing amounts of potassium phosphate were added to the bathing solutions, and the resultant channel conductances were measured. These data were then used to calculate the percent inhibition and the I s o concentration of the added phosphate ions (Table IX). The majority of the mutant proteins exhibited degrees of conductance inhibition similar to or greater than the wild-type protein, which displayed a 74% decrease in conductance after the addition of 3.3 mM potassium phosphate. The Lys 7 4 mutant demonstrated a slightly lowered affinity for phosphate ions, with a maximum inhibition of 58% and an I s o concentration of 1.95 mM compared to 0.96 mM for wild-type OprP. The Lys 1 2 1 substitution had a profound effect on the ability of the protein to bind phosphate ions. This mutant exhibited a maximum inhibition of 30%, and while the I 5 0 for this mutant could not be measured under the conditions used to examine the other mutant proteins, additional experiments revealed that it was above 10 mM. The Lys 1 2 6 mutant channel conductance, although greatly reduced compared to the wild-type protein, appeared to be inhibited by the presence of phosphate ions to a similar degree. Figure 16 shows the channel conductances of the Lys 7 4, Lys 1 2 1 and Lys 1 2 6 mutant proteins along with wild-type OprP plotted as a function of increasing phosphate ion concentration and Figure 17 shows the percentage inhibition of these same mutants as a function of phosphate ion concentration. Table IX. Phosphate inhibition of chloride conductance through the OprP Lys—>GIu substitution mutant proteins Phosphate inhibition of chloride conductance OprP mutation Maximum inhibition (%) I50 (mM) Wild type 74 0.96 L y s 1 3 ^ G l u 89 0.62 L y s 1 5 ^ G l u 87 1.20 L y s 2 5 ^ G l u 79 1.00 Lys 3 0 -»Glu 81 0.73 L y s 7 4 ^ G l u 58 1.95 L y s 1 0 9 ^ G l u 79 0.84 L y s 1 2 1 ^ G l u 30 >10 L y s 1 2 6 ^ G l u 84 0.81 L y s 1 8 1 ^ G l u 80 0.90 83 Figure 16. Phosphate-induced inhibition of chloride conductance. The ability of selected OprP Lys-*Glu mutants to bind phosphate ions was assessed by measuring the inhibition of single-channel conductance (in 0.1 M KC1) induced by the addition of increasing concentrations of potasium phosphate ions. Single-channel conductance was plotted as a function of phosphate ion concentration. (•-OprP, o- Lys 7 4—•Glu, •-Lys121->Glu, •-Lys126->Glu) 84 100 0 0.00 0.50 1.00 1.50 2.00 2.50 3.00 Concentration phosphate (mM) 3.50 Figure 17. Phosphate-induced inhibition of chloride conductance. The ability of selected OprP Lys Glu mutants to bind phosphate ions was assessed by measuring the inhibition of single-channel conductance (in 0.1 M KC1) induced by the addition of increasing concentrations of phosphate ions.Percent inhibition was plotted as a function of phosphate ion concentration..(»-OprP, o- Lys 7 4->Glu, •-Lysm-*Glu ,-n-Lys 1 M->Glu) 85 4.7 Channel characteristics of Lys->Gly, Lys->Gln and Lys 7 4 / 1 2 1 , 1 2 6->Glu mutant proteins To further examine the roles Lys 7 4, Lys 1 2 1 and Lys 1 2 6 play in determining the electrochemical nature of the channels formed by OprP, these amino acids were again substituted with the neutrally-charged residues Gly or Gin, and the single-channel conductance was determined for each of these mutant proteins (Table X). In addition, a triple mutant with Lys 7 4, Lys 1 2 1 and Lys 1 2 6 all substituted with glutamates was also created and analyzed. Substituting Lys 7 4 with Gly resulted in a channel with a conductance comparable to that of the wild-type protein in 1 M KC1. However, the channel conductance of this mutant i n 0.1 M KC1 was similar to that of the Lys 7 4—>Glu mutant protein. The phosphate-induced conductance inhibition of the Lys 7 4—>Gly mutant was comparable to that of wild-type OprP. Substituting the Lys 1 2 6 residue with either Gly or Gin resulted in channels that displayed reduced levels of conductance in comparison to the wild-type protein. In the case of the Gly substitution, the channel conductance at both 0.1 M and 1 M KC1 was lower than the Lys 1 2 6—>Glu mutant. Substituting Lys 1 2 1 with either Gly or Gin resulted i n channels with reduced conductance at both 0.1 M and 1 M KC1. These mutant proteins also formed channels that were as severely impaired in their ability to bind phosphate ions as the initial Lys 1 2 1->Glu mutant protein. The single-channel conductance of the Lys 7 4 , 1 2 1 , 1 2 6—>Glu triple mutant was somewhat lower than any one of the single substitution mutants in both 0.1 M and 1 M KC1; however, the phosphate-induced inhibition of conductance of this mutant channel was i n the range of the Lys 1 2 1 single mutant. 4.8 Expression and membrane localization of carboxy-terminal lysine substitution mutant proteins The expression and presence in outer membranes of the Lys 3 6 1—>Glu and the Lys 3 8 5—>Glu mutant proteins was assessed by SDS-polyacrylamide gel electrophoresis 86 Table X. Channel characteristics of OprP Lys->Gly and Lys->Gln mutant proteins OprP mutation Average single-channel conductance (pS)a' 0.1 M K C 1 1 M KC1 Maximum inhibition1'' Wild type 103 ± 4 5 230 ± 4 2 74 % L y s 7 4 ^ G l u 41 ± 14 116 ± 3 4 58 % L y s 7 4 ^ G l y 39 ± 9 205 ± 3 5 72 % L y s 1 2 1 ^ G l u 10 ± 2 74 ± 11 30% L y s 1 2 1 ^ G l y 8 ± 2 84 ± 2 6 16 % L y s 1 2 1 ^ G l n 10 ± 3 67 ± 1 8 24% L y s 1 2 6 ^ G l u 25 ± 6 99 ± 17 79% L y s 1 2 6 ^ G l y 13 ± 3 52 ± 8 N.D.C-L y s 1 2 6 ^ G l n 13 ± 2 87 ± 2 8 N.D. L y s 7 4 , 1 2 U 2 6 ^ 4 ± 1 47 ± 9 20% a. Average of 100 single--channel events expressed in picoSiemens ± S.D. b. Maximum conductance inhibition after addition od 3.3 m M potassium phosphate c. N.D.= not determined 87 and Western immunoblot analysis of outer membranes purified from clones expressing these mutant proteins (Figure 18). Both of the substitution mutants were found to be expressed at levels comparable to the wild-type protein. In addition, not only were these proteins absent from the supernatant of outer membranes solubilized i n 0.5% octyl-POE (a procedure that should solubilize proteins which were produced as inclusion bodies) but, based on the results of the immunofluorescence experiments, cells expressing these proteins bound anti-OprP antibodies to the same extent as cells expressing wild-type OprP. These findings indicate that neither of these two lysine residues are important for the biogenesis or export of OprP to the outer membrane of the host cell. 4.9 Summary A recombinant PCR site-directed mutagenesis method was used to individually mutate nine lysine residues contained i n the amino-terminal half of OprP and two lysine residues located in the carboxy-terminal half of the protein. The amino-terminal substitution mutants were purified and analyzed for alterations in channel functions using the planar bilayer method. Three of these amino-terminal lysine residues (Lys 7 4, Lys 1 2 1 and Lys 1 2 6) were shown to be involved in determining certain channel characteristics of OprP. One particular residue (Lys 1 2 1) which was postulated in Chapter 3 to be located in the third surface-exposed loop (Figure 19) was shown to be involved in maintaining the phosphate-specific nature of the channels formed by OprP. Two carboxy-terminal lysine residues in OprP which were suspected of playing a role in the export of the protein to the outer membrane were substituted with glutamates. These mutant proteins were expressed in an E. coli background and their association with isolated outer membrane fractions was assessed. The results of these experiments revealed that the substitution of these two lysine residues had no appreciable effect on the export of OprP to the outer membrane of host cells. 88 Figure 18. Whole cell Western immunoblot of E. coli CE1248 expressing the OprP substitution mutant proteins Lys 3 6 1 -»Giu (A) and Lys 3 8 5 -> Glu (B). Lane C corresponds to wild-type OprP. The blots were reacted with the OprP-specific rabbit polyclonal serum. The samples were heated at 100° C for 10 min prior to loading. Prestained molecular mass markers on the left were (from the top) 106, 80, 49.5, 32.5, 27.5 and 18.5 kDa. 89 R O G D T E K D T T S F K L G G R L Q A D Y G R F D G Y E L Y A R R W G G G T A Y R D Y T T V T G S R N V G F L D E T 121 (g) Y A F T R S G S F 126 (K) K W L N V P N F G T Y T V s . A E D F Y G A S D V T A L E R N W N D N V G N N N E S F V s G S L F A M G G _y_ v s s A Q I T D G D S V K R M G M R P R I R T D V A S D E L D V S T N G G N D A G S N G N R G L Y R F N Y G L A G R Y S G Q s V L A F G V A L E P H G L V L H V W E N K P G D A E K V T R R L Y E A Q A S F A G L A W A G E L G W D S R D R E D L K A S G Y Y A Q L A Y T I T G E P R V V N D D E V K I S D Y R Y F L E W A G I A T R E V G D A K G K T R N A S I K D T K A K G N T Q E K N E P K I L D G A K F D N G D D S G D G A V L L Y V G N M V A R N S L W V Q Y K Y A V V N E A F Figure 19. Topological model of the monomeric structure of OprP. Circles indicate lysine residues shown to play a role in defining the channel characteristics of OprP. 90 DISCUSSION Previous studies of the outer membranes of gram negative bacteria have revealed the existence of a class of proteins which have been termed porins. These proteins form trans-membrane channels which allow the exchange of small hydrophilic molecules between the external environment and the periplasmic space (Nikaido, 1992). Analysis of the structure and function of various porins has shown that they appear to possess a number of common features which are apparently necessary for their particular function. A l l of the porins examined to date traverse the outer membrane in the form of a p-barrel. While the majority of porins whose structures have been determined are composed of 16 p-strands (Cowan et al., 1992), the monomers of the maltoporin LamB are i n the form of an 18-stranded p-barrel (Schirmer et al., 1995). The loop regions of all of these proteins were found to be the most variable with regard to both the individual residues and the lengths of the loops. However, the external loops were always found to be longer and more variable than the internally-exposed loops. Both the ion-selective and substrate-specific characteristics of several different porins have been determined to be due to the presence of a small number of amino acid residues located primarily in the amino-terminal halves of these proteins. The sites which these residues occupy appear to be conserved within the proposed structures of the various porins. Combinatorial mutagenesis studies of OmpC and PhoE have shown that the residues involved in determining the individual ion selective natures of these porins are concentrated in the amino-terminal halves of the proteins (Benz et a l , 1989). Studies of PhoE have also shown that the substitution of one particular lysine residue located in the amino-terminal end with a negatively-charged glutamate residue also effected the ion selectivity of the channels formed by this protein (Bauer et al., 1989). The results of this study have been used to create a topological model of the OprP monomer. Those sections of the protein which form loop regions have been 91 differentiated from those that compose the p-strands by means of linker insertion mutagenesis. Three surface-exposed loops have been positively identified by means of epitope insertion mutagenesis followed by indirect immunofluorescence using an anti-epitope-specific antibody. In addition, the location of three lysine residues which have been shown to play a role in maintaining the channel characteristics of OprP have been identified. Two of these residues appear to accelerate the rate at which the anions pass through the channel. The third lysine residue, which has been proposed as being located i n the third surface-exposed loop, appears to take part in the formation of the anion binding site. Topological model of OprP The proposed topological model of OprP is similar to those of many other bacterial porins. The individual monomers are composed of a P-barrel which encloses a central channel. The most recent models of the E. coli general diffusion porins depict these proteins as possessing p-strands with amino acid sequences which alternate between hydrophobic and hydrophilic residues (Cowan et a l , 1992). This predicted amphipathicity is thought to be necessary for the simultaneous exposure of these strands to both the nonpolar outer membrane core and the internal lumen of the channel. The side chains of the amino acids which project into the aqueous channels would likely display a hydrophilic character. However, those residues exposed to the fatty acyl chains of the external l ip id membrane would need to be hydrophobic. Interactions of the external surface of these proteins wi th the surrounding l ip id bilayer have been proposed to be facilitated by a band of hydrophobic residues which encircles the trimer. The boundaries of these bands are largely composed of phenylalanine and tyrosine residues (Cowan et a l , 1992) 92 The general amphipathic nature of the P-strands displayed by other porin models has been maintained in the model of OprP. The number of individual strands which form the p-barrel has been set at 16. This is in keeping with the number of p-strands found i n most porins, but it is less than the number of strands proposed to exist in the structure of LamB (Schirmer et a l , 1995). The length of the amino acid sequences of both LamB and OprP exceed those of most other porins by at least 20%. It might therefore seem reasonable to assume that since the p-barrels of LamB contain 18 strands, so too must the p-barrels of OprP. The results of this study show that this is not likely to be the case. The insertion sites of the OprP linker mutant proteins PL3, PL8, PL12 and PL13 were all placed within p-strands. Incorporation of the four amino acid residue insert at these sites resulted i n a substantial decrease in the amount of protein expressed in the outer membranes and to a loss of detectable surface-exposed protein in whole cells. The placement of the p-strands i n certain topological models of LamB and PhoE was partially based on the effect that the insertion of foreign sequences into these sites had on the expression of the mutant proteins (Charbit et a l , 1986; Agterberg et al., 1987). Those regions of these proteins where the insertion of the foreign sequence resulted in a significant decrease in protein expression were placed in the p-strands. Subsequent models based on X-ray crystallography revealed that the assignment of these sites to the domain of the p-strands accurately reflected their true location in the structure of these proteins (Schirmer et al., 1995; Cowan et al., 1992). Although the expression levels of the mutant forms of OprP with the foreign insert located in proposed p-strand regions were reduced in comparison to the wild-type protein, all of these proteins were expressed to some extent. The residual levels of expression exhibited by these proteins could be due to the ability of the inserted sequences to maintain the amphipathic nature of the P-strand, allowing the displaced part of the strand to be forced into the adjacent loop. This could explain why, although 93 all four mutant proteins exhibited reduced levels of expression, the inserts at sites PL12 and PL13, which consisted of the amphipathic sequence DLQV, inhibited the expression of the mutant proteins to a lesser extent than the inserts at sites PL3 and PL8 which consisted of the nonpolar sequence PAGP. The OprP insertion sites PL1, PL4, PL6 and PL7 were all placed in periplasmically-exposed loops. None of these mutant proteins displayed significant decreases in protein expression after incorporation of the 12 base pair linker into the gene encoding OprP. However, the mutant forms of OprP resulting from the insertion of the malarial epitope into the linker sites PL6 and PL7 manifested profound decreases in expression. These two insertion sites are located at the beginning of a stretch of 23 uncharged amino acid residues. No such large tract of uncharged residues has ever been identified in the structure of any other porin (Siehnel et al., 1990). The insertion of the 14 amino acid malarial epitope at these sites may have caused a disruption in this area which resulted in a destabilization of the folded structure and a reduction in the amount of the mutant protein that was be able to be transported to and inserted into the outer membrane. The mutant proteins which contained inserts at sites PL1 and PL4 appeared to tolerate the inclusion of the malarial epitope as evidenced by their ability to be exported and expressed in the outer membrane of whole cells. The use of the trimer-specific anti-OprP antiserum ensured that all the mutant proteins detected by this assay were indeed exposed at the cell surface in the form of the native trimer. None of the mutant proteins with the malarial epitope inserted into proposed periplasmically-exposed loops were shown to display the epitope at surface-accessible regions. The proposed periplasmic loops of OprP tend to be longer than those of the other previously examined porins. The sixth periplasmic loop in particular is more than twice the size of the largest such loops in the E.coli porins (Cowan et al., 1992). Placement of this section of OprP in the periplasmic space was dictated by two sets of results; i) sites surrounding the proposed loop were nonpermissive for the insertion of foreign 94 sequences (indicating the presence of p-strands), ii) two loops surrounding this region were positively identified as being exposed at the cell surface. This second set of results fixes the position of this region of OprP; it would be difficult for the protein to cross the membrane two more times within such a short stretch of sequence. Since OprP is thought to interact with the periplasmic phosphate-binding protein in P. aeruginosa (Poole and Hancock, 1984), these longer periplasmic loops may be necessary to facilitate that interaction. The long periplasmic loops in the proposed topological models of the siderophore-iron receptors FhuA (Koebnik and Braun, 1993) and FoxA (Baumler and Hantke, 1992) may be similarly involved in the proposed interactions between these proteins and cytoplasmic membrane associated proteins such as TonB in other gram-negative bacteria. The linker insertion sites PL2, PL5 and PL9 were all placed in surface-exposed loops despite the fact that none of the epitopes inserted into these sites were shown to be exposed at the cell surface. A l l three of these sites tolerated the insertion of the linker and the mutant proteins were able to incorporate the malarial epitope without suffering any major decreases in expression. These mutant proteins were also shown to be properly exported to the cell surface. The lack of surface exposure of the malarial epitopes inserted at the PL2 and PL9 sites can be explained if these sites are presumed to occupy sites which are at the junction between loop and P-strand regions. Epitopes inserted into such sites might not be efficiently exposed at the cell surface and would therefore not be accessible to the epitope-specific antibody. Struyve et al. (1993) similarly speculated that the reason that an epitope inserted into the proposed second surface-exposed loop of PhoE (Agterberg et al. 1990) was not shown to bind antibodies at the cell surface was because it was located at the interface between a loop and an adjacent P-strand. Insertion site PL5 was localized to the third surface-exposed loop of the proposed model of OprP. The analogous loops in the Rhodobacter capsulatus porin 95 (Weiss et al., 1990) and in PhoE, OmpF and OmpC (Cowan et a l , 1992) have all been shown to fold back into the channels and constrict their internal diameters. The surface-exposure of an epitope inserted into the third loop of PhoE was not achieved until three copies of the epitope had been inserted (Struyve et al., 1993). These experiments suggested that the third loops of bacterial porins are somewhat shielded from the exterior: surf ace of the cell. The lack of surface-exposure displayed by the epitope inserted into the proposed third loop of OprP can be similarly explained by the fact that this loop folds back down into the central channel. A n epitope inserted into this site would be partially shielded from the external surface and would thus be unable to bind to the antibody. If the third loop of OprP is folded back into the channel, it would likely exist i n an environment which is different from that of the other surface-exposed loops. The location of this loop would allow for the presence in its sequence of hydrophobic residues which could interact with other hydrophobic residues protruding from the sides of the barrel wall. The sequence of the proposed third loop of OprP contains a higher amount of aromatic residues than any of the other predicted loops in this model. These residues are located predominantly on one side of the loop, which presumably would face the interior wall. The other side of the loop contains a number of residues with polar or charged side chains. These residues, which would be expected to face the central opening of the pore, might play a role in defining the channel characteristics exhibited by this particular porin. Two lysine residues contained in this region of the third loop (Lys 1 2 1 and Lys 1 2 6) have been shown to be involved in the conductance of anions through the channels formed by OprP. While the sequences of the well characterized E. coli porins have been shown to lack any long stretches of uncharged residues (Mizuno et al., 1983), there exists in the OprP sequence a stretch of 24 consecutive uncharged amino acid residues, as well as three other tracts of 13 or more uncharged residues (Siehnel et al., 1990). The largest 96 region of uncharged amino acids has been proposed as lining the wall of one side of the channels formed by OprP. The section of the channel which this uncharged stretch of amino acids is thought to construct might interact with that section of the third loop which contains a high proportion of aromatic residues. The presence of uncharged residues at this site might provide a means by which to stabilize the conformation of the loop which takes part in forming the constriction zone. As noted above, the insertion of the malarial epitope into two sites which were located at the beginning of this sequence of residues resulted i n a loss of detectable protein at the surface of whole cells. Three sites i n OprP have been positively identified as being located in surface-exposed regions. Epitopes presented at insertion sites PL10 and PL11 were shown to be capable of binding the anti-epitope specific antibody at the cell surface. These same epitope-containing mutant proteins were also shown to bind the trimer-specific anti-OprP antibodies, suggesting that the proteins did not suffer any gross alterations in their conformations as a result of the insertion of the foreign epitope. The fact that these proteins were capable of binding to both the anti-OprP antibodies and the anti-epitope-specific antibodies on the surface of whole cells suggests that the insertion sites of these mutant proteins are located in regions that are exposed to the cell surface i n the wild-type protein. In addition to the PL10 and PL11 sites, another site i n OprP was identified as being located in a surface-exposed region by means of site-directed epitope insertion mutagenesis. Based on the initial linker and epitope insertion mutagenesis experiments a preliminary topological model of OprP was proposed. To test the validity of this model, a copy of the malarial epitope was inserted into a site in OprP which was proposed to be located in the seventh surface-exposed loop. Analysis of whole cells expressing this mutant protein by indirect immunofluorescence revealed that the epitope was indeed exposed at the cell surface. 97 The proposed topological model of OprP was further tested by means of site-directed deletion mutagenesis. Two sections of OprP which were proposed to be located i n the fifth surface-exposed loop were deleted and the mutant proteins were analyzed for changes in expression. The section between linker insertion sites PL9 and PL10 as well as the section between insertion site PL10 and a site which corresponded to a downstream Sstll site in the gene sequence were individually deleted and the deletion mutants were analyzed, by the electrophoresis of isolated outer membranes through SDS-polyacrylamide gels. Both of the mutant proteins which contained deletions of sections of the fifth loop were found to be expressed at levels comparable to the wild-type protein. In addition, these mutant proteins were also shown to be properly localized to the cell surface. The results of the deletion mutation experiments not only served to determine the extent of the fifth surface-exposed loop, they also provided further evidence for the placement of the insertion site PL9 at the interface between a loop and a p-strand. The lengths of the surface-exposed loops i n the proposed model of OprP tend to be longer than the external loops of the other well characterized porins. The proposed third surface-exposed loop is approximately the same size as those of the general diffusion porins. Deletion analysis of the fifth surface-exposed loop has revealed that this region is composed of at least 34 amino acid residues. This is as large as the third surface-exposed loop of the classical E. coli porins OmpF, OmpC and PhoE, which is the largest loop found in the structures of these proteins. In addition to the fifth loop, the topological model of OprP also possesses within its structure two other carboxy-terminal loop regions (loop 6 and loop 7) which contain 26 and 27 amino acid residues, respectively. The abundance of such large loop regions may help to explain the small exclusion limit which is characteristic of wild-type OprP. These large loops might form a barrier over the external mouth of the channel which would restrain the passage of those substrates which were incapable of negotiating the narrow entrance of the pore. 98 The existence of such large loop regions also helps to explain why OprP apparently contains only 16 p-strands despite the fact that the primary sequence of this protein is composed of approximately 20% more amino acid residues than the sequences of other porins which are predicted to contain similar numbers of P-strand structures. Although the D N A sequences which control the phosphate-starvation inducible regulon i n both E. coli and P. aeruginosa share a certain degree of homology, the porins which these organisms express in response to such conditions exhibit very little sequence similarity. A comparison of the nucleotide sequences of the genes encoding OprP and PhoE yielded a total alignment of 53-60% with consistently low alignment scores (Siehnel et al., 1990). A similar comparison of the sequences of PhoE and LamB yielded comparable results. These findings might lead to the assumption that since the sequence of OprP does not resemble that of the classical E. coli porins, it must therefore retain a different conformation in its native form from that of the other porins. However, the crystal structure of LamB shows that this protein assumes a similar conformation to that of the other previously crystallized porins, despite the fact that the sequence of LamB shares little homology with those proteins (Schirmer et a l , 1995). Accordingly, it is possible that the structure of OprP resembles that of the other characterized porins despite the apparent lack of sequence homology. The lack of homology between OprP and PhoE is particularly surprising in light of the fact that the trimeric forms of these two proteins cross-react with an anti-OprP trimer-specific antiserum (Poole and Hancock, 1986). These findings suggest that these two proteins share one or more common conformation epitopes. Thus, despite the lack of homology between the amino acid sequences of these two porins, their common function in the outer membranes of their respective hosts appears to dictate the assumption of similar folding patterns. In addition, a lysine residue which has been implicated as playing a role in the maintenance of the anion-selective nature of PhoE (Bauer et al., 1989) was shown to occupy a position similar to that of a lysine residue i n 99 OprP which has been shown to be involved in forming the anion binding site of this porin. A gene located upstream of the oprP gene on the P. aeruginosa chromosome was recently cloned and expressed i n E. coli (Hancock et al., 1992). This gene was found to encode a polyphosphate-selective porin which was designated OprO. This porin, which cross-reacted with OprP-specific antiserum, was shown to be highly homologous to OprP, with 76% identity and 16% conserved substitutions (Siehnel et al., 1992). A comparison of the amino acid sequence of OprO with the proposed topological model of OprP reveals that 29/39 of the non-conservative substitutions and two of the three one amino acid gaps are located in the loop regions. This would seem appropriate, since although the variability between closely related outer membrane proteins is usually concentrated in the loop regions, a certain amount of variation i n the amino acids residues which line the interior of the channels formed within these two proteins would be required to account for their differences in substrate-specificity. The role of specific lysine residues i n defining the channel characteristics of OprP It has previously been demonstrated that the channel characteristics of several general diffusion porins are dependent on the presence of one or more amino acids located in their amino-terminal domains (Heine et a l , 1988; Benson et a l , 1988, Benz et al., 1989). In this study we have identified three amino-terminal lysine residues in OprP which play a role in defining the channel characteristics of this porin. A model of the structure of OprP in which the three monomers each contribute one lysine residue to form a single central binding site has previously been proposed. While the results of the current study do show that three different lysine residues of OprP appear to be involved i n determining the channel characteristics, it is now believed that the monomers of this protein form three separate channels instead of a single central pore. 100 The substitution of Lys 1 2 1 with a glutamate residue yielded a protein which displayed channel conductances which were significantly smaller than those exhibited by wild-type OprP. These mutant forms of OprP were shown to conduct anions at a rate approximately 3-fold lower than that of the wild-type protein in 1 M KC1. Similar substitutions of Lys 7 4 and Lys 1 2 6 produced mutant proteins which displayed 2-fold and 2.5-fold decreases i n conductance, respectively. While the changes induced by these substitutions are profound, they do not approach the 10-fold reduction in conductance exhibited by chemically-modified forms of OprP (Hancock and Benz, 1986). Even a triple mutant i n which all three of these lysine residues had been substituted with glutamates displayed only a 5-fold decrease i n conductance. These findings suggest that there may be certain lysine residues contained in the carboxy-terminal end of OprP which also play a role in determining the channel conductance. Alternately, the severe reduction i n conductance displayed by the chemically-modified forms of OprP may have been due to factors unrelated to the actual loss of the positive charges of the modified lysine residues (eg. the presence of the modifying groups within and/or around the channel). Lysine-specific acetylation of OprP was shown to produce channels with conductances that were no longer saturated at high anion concentrations (Hancock and Benz, 1986). This result was proposed to be due to a modification of residues which are involved i n forming the anion-binding site. Of the 11 Lys—> Glu mutant forms of OprP created during the course of this study, only the Lys 7 4 and Lys 1 2 1 mutants exhibited losses in the ability to saturate at KC1 concentrations above 1 M. The Lys 1 2 6—>Glu mutant displayed saturation kinetics similar to that of wild-type OprP, despite the fact that the conductance of these channels was as severely affected at low salt concentrations as that of the Lys 7 4 and Lys 1 2 1 mutant porins. The conductance patterns of the other eight Lys—> Glu mutants did not differ significantly from that of the wild-type protein. Apparently only the Lys 7 4 and Lys 1 2 1 substitutions had a detrimental effect on the anion-binding site. 101 The phosphate-induced inhibition of channel conductance of the Lys 7 4—>Glu mutant was approximately 2-fold lower than that of the wild-type protein. Substituting this lysine residue with a glycine resulted in a protein with a phosphate-induced inhibition of conductance which was similar to that of the wild-type protein. This result can be explained, if Lys 7 4 is assumed to occupy a space proximate to the Pi-binding site. The positive charge of this residue would not be required for the formation of the binding site, however the placement of a negatively-charged residue at this location may have indirectly affected the interaction of phosphate (and chloride) ions with the binding site. According to the topological model of OprP, Lys 7 4 is located at the top of the fourth p-strand and would presumably face the interior of the channel. Substitution of Lys 1 2 1 with glutamate, glycine or glutamine residues resulted i n proteins with channel conductances which were severely impaired in their abilities to be inhibited by the presence of phosphate ions. This particular residue is located i n the third surface-exposed loop according to the topological model. The placement of this residue in the third loop is i n agreement with this loop's role in constricting the interior of the channels formed by several bacterial porins (Weiss et al., 1990; Cowan et al., 1992; Schirmer et al., 1995). The equivalent lysine residue i n PhoE (Lys 1 2 5) which has been established as being responsible for determining the anion selectivity of this porin (Bauer et a l , 1989) was also shown to be located in the third surface-exposed loop. Substituting Lys 1 2 6 with glutamate in OprP had no apparent effect on the Pi-binding site, despite the fact that this residue is also predicted to be located in the third surface-exposed loop. The mutagenesis of Lys 3 6 1 and Lys 3 8 5 did not yield any useful information on the role of these residues in OprP. The substitution of these amino acids with glutamate residues did not appear to. affect the ability of the mutant proteins to be properly transported to and inserted into the outer membrane of the host cells. This is i n contrast with the results obtained after the analogous mutagenesis of a mitchondrial porin. 102 Smith et al. (1995) showed that the independent substitution of two lysine residues located in the carboxy-terminal half of a mitochondrial porin with glutamates affected the ability of the mutant proteins to insert into membranes. These two residues were found to be present i n a sequence of the mitochondrial porin which appears to be conserved within a small number of similar porins. The presence of two such lysine containing motifs in the carboxy-terminal end of OprP suggested that they might play a similar role in the biogenesis of this protein. While the results of this study suggest that these lysine residues do not serve such a purpose, it should be noted that the method applied i n the analysis of the mutant mitochondrial proteins involved a cell-free transcription/translation system. A n analysis of the effects of individual amino acid substitutions on membrane insertion using this procedure would be significantly more sensitive than one performed using the method employed i n the current study of OprP. It may therefore be possible that the mutation of OprP at these sites had an effect on the ability of the mutant proteins to insert into membranes which was not detected. Analysis of the anion selectivity of the amino-terminal Lys—>Glu substitution mutants using the zero-current membrane potential assay proved to be inconclusive. While there were slight variations in the potential differences elicited by these mutants, these differences could not be correlated with the observed changes in anion conductance. These findings conflict with the facts that; i) the destruction by chemical modification of the positive charges of accessible lysine residues of OprP substantially decreased the anion selectivity of the channel (Hancock and Benz, 1986), ii) the substitution of a single lysine residue located in the third surface-exposed loop of PhoE with glutamate changed this porin from an anion-selective to a cation-selective channel (Bauer et al., 1989). It may be that the anion selectivity of OprP, which is much greater than that of PhoE, is due to the presence of several positively-charged residues, and thus would not be as adversely affected by the mutation of only one of these amino acids. 103 The phosphate-starvation inducible E. coli porin PhoE has not been shown to possess a substrate-specific binding site (Bauer et al., 1988), although it does maintain a slight preference for the transport of anions over cations. Lysine-specific chemical modifications of PhoE were shown to destroy the anion selectivity of the channels (Bauer et al., 1989). In addition, the replacement of lysine residues contained in the amino-terminal half of this porin with glutamates was also found to affect the ion-selectivity of the channels formed. Four lysine residues were individually replaced and the mutant proteins were analyzed both in vivo and in vitro for changes in channel function. Three of the substitution mutant proteins exhibited minor alterations i n anion selectivity as measured using the zero-current membrane potential assay. The fourth substitution mutant form of PhoE displayed a reversal i n ion selectivity, changing from an anion selective channel to a cation selective channel. The residue substituted in this particular mutant protein (Lys 1 2 5) is located in the third surface-exposed loop of the crystal structure of PhoE. This would be the appropriate position for a residue involved in determining the ion preference of a porin, owing to the fact that this loop region folds back down and serves to constrict the internal diameter of the channel. A l l of the substitution mutant forms of PhoE were shown to retain single-channel conductances which were comparable to those of the wild-type protein. A recently published model of the maltose-specific porin LamB has provided an interesting rational for the substrate-specific behavior of this particular porin (Schirmer et al., 1995). In the crystal structure of this protein, a series of six aromatic residues (predominantly tyrosines and tryptophans) are arranged in a helical pathway which extends from the external mouth of the channel through the constriction zone and out to the periplasmic side. These residues are thought to participation i n stacking interactions with incoming sugar molecules. The aromatic residues at the cell surface would serve to both attract the maltose molecules to the entrance of the channel and align these molecules with the aromatic pathway. The sugars would then be guided 104 through the channel by means of a series of interactions with the hydrophobic faces of the aromatic side chains. The substrate specificity of this porin may therefore rely on the presence of a diffuse binding site which extends over almost the entire length of the channel. The configuration of this binding site would likely differ somewhat from that of the proposed binding site of OprP, since the structure of the phosphate ion is quite different from that of maltose. The E. coli porins OmpF and PhoE have both been determined to form nonspecific general diffusion channels (Nikaido and Rosenberg, 1983). However, a recent study of the electrostatic properties of these two porins has proposed a mechanism of transport which appears to resemble a form of facilitated diffusion (Karshikoff et al., 1994). A strong transverse electrostatic field which winds through the channels of these proteins was identified. The electrochemical nature of this field would be determined by the nature of the individual residues involved in creating it. Differences in the ion selectivities displayed by these two porins are thought to be due to the differences i n the amino acid side chains of those residues which create the electrostatic fields of each of these two proteins. While a complete assessment of all 23 lysine residues contained i n OprP has yet to be performed, the current analysis of amino-terminal residues has yielded some important information regarding the nature of the channels formed by this protein. The roles of Lys 7 4 and Lys 1 2 6 in OprP appear to be to form an electrostatic funnel which serves to focus the flow of anions toward the binding site. The Lys 1 2 1 residue seems to render a more critical function in this porin. Not only does the presence of this residue serve to increase the flow of anions through the channel, but it appears that this particular amino acid is required for the formation of the anion/Pi-binding site. Whether this is the only residue involved in maintaining the Pi specificity of OprP is still unknown. 105 Summary A topological model of the P. aeruginosa porin OprP has been proposed based on linker and epitope insertion mutagenesis (Figure 20). This model is consistent with the current crystal structures of several porins isolated from the outer membranes of other gram negative bacteria. The p-barrel is proposed as being comprised of 16 strands, and the longer more variable loops of the barrel were placed at the external surface. Three of these loops were identified by means of epitope insertion. The third surface-exposed loop was shown to contain two lysine residues (Lys 1 2 1, Lys 1 2 6) which were important for the transport of anions through the channels formed by this protein. One of these residues (Lys 1 2 1) was implicated as taking part in the formation of the anion binding site. A third lysine which also appears to affect the ability of the channels to conduct anions has been localized to the top of the fourth p-strand. 106 1D (g) G I s F K L G G R 1 G A D T T V T G • G V L V A R 3 Y T G G T A V R 0 ® , Q 1 N V • R N V F • T Y F R G F K L N V P N F G T A S D L E © A T S © ® W V T A L E R N I A | • W| V —ft D N V a 0 N N N E S F V M V s 7 ® A Q I G T D G • S V K R — T N L R G V © P L H ^ M \ " ^ P \ R 9 J G>_ A Y • L G L H V P G s \ \ \ s v G G N D A G S N G N R G L F G G 5 5 A V E G L W A E K 11 © D R E 0 L K A S G Y Y F A G L A W A G E L G w V D S 12 I T G E © R V V I N D 0 E V K 1 s D 1 a R Y F L © w A G I S A T R E V G 0 A K G K T R N A K D T K A K A V L Y G N V A N S W V Y K A V N E A E K N E P K I L • G A K F D N G D D S G D G T~ IV U| R L IQ Y IV F Figure 20. Topological model of the monomeric structure of OprP. Boxes enclose the proposed transmembrane P-strands. Red circles indicate the sites at which the inserted malarial epitopes were shown to be surface-exposed. Green circles indicate lysine residues shown to play a role in defining the channel characteristics of OprP. A l l other circles indicate the linker insertion sites PL 1 to PL 13 (number 1 to 13 respectively). The sections of the fifth surface-exposed loop which were deleted are marked with hatched lines. 107 REFERENCES Agterberg, M . H. , Adriaanse, and J. Tommassen. 1987. Use of outer membrane protein PhoE as a carrier for the transport of a foreign antigenic determinant to the cell surface of Escherichia coli K-12. Gene. 59: 145-150. Angus, B. L . and R. E. W. Hancock. 1988. Outer membrane proteins F, P and DI of Pseudomonas aeruginosa and PhoE of Escherichia coli: chemical cross-linking to reveal native oligomers. J. Bacteriol. 155: 1042-1051. Bauer, K., R. Benz, J. Brass and W. Boos. 1985. Salmonella typhimurium. contains an anion-selective outer membrane porin induced by phosphate starvation. 161: 813-816. Bauer, K., P. van der Ley, R. Benz and J. Tommassen. 1988. The pho-controlled outer membrane porin PhoE does not contain specific binding sites for phosphate or polyphosphates. J. Biol. Chem. 263: 13046-13053. Bauer, K., M . Struyv6, D. Bosch, R. Benz and J. Tommassen. 1989. One single lysine residue is responsible for the special intreaction between polyphosphate and the outer membrane porin PhoE of Escherichia coli. J. Biol. Chem. 264: 16393-16398. Baumler, A . J. and K. Hantke. 1992. Ferrioxamine uptake in Yersinia enterocolitica: characterization of the receptor protein Fox A. Mol . Microbiol. 6: 1309-1321. Bayer, A . S., D. P. Speert, S. Park, J. Tu, M . Witt, C. C. Nast and D. C. Norman. 1991 Functional role of mucoid exopolysaccharide (alginate) in antibiotic-induced and polymorphonuclear leukocyte-mediated killing of Pseudomonas aeruginosa. Infect. Immun. 59: 302-308. Benson, S. A. , J. L . Occi and B. A . Sampson. 1988. Mutations that alter the pore function of the OmpF porin of Escherichia coli. J. Mol . Biol. 203: 961-970. Benz, R. and R. E. W. Hancock. 1981. Properties of the large ion-permeable pores from protein F of Pseudomonas aeruginosa in lipid bilayer membranes. Biochim. Biophys.Acta. 649: 298-308. Benz, R. and R. E. W. Hancock. 1982. Porins from gram-negative bacteria in lipid bilayer membranes, pp. 123-134. In " Transport in biomembranes: model systems and reconstitution. Raven Press, New York. Benz, R., M . Gimple, K. Poole and R. E. W. Hancock. 1983. An anion-selective porin from the Pseudomonas aeruginosa outer membrane. Biochim. Biophys. Acta. 730: 387-390. 108 Benz, R., R. P. Darveau and R. E. W. Hancock. 1984. Outer membrane protein PhoE from Escherichia coli forms anion-selective pores in lipid bilayer membranes. Eur. J. Biochem. 140: 319-324. Benz, R., A . Schmidt, T. Nakae and G. H. Vos-Scheperkeuter. 1986. Pore formation of LamB of Escherichia coli in lipid bilayer membranes. J. Bacteriol. 165: 978-986. Benz, R., A . Schmid, P. Van der Ley and J. Tommassen. 1989. Molecular basis of porin selectivity: membrane experiments with OmpC-PhoE and OmpF-PhoE hybrid proteins of Escherichia coli K-12. Biochim. Biophys. Acta. 981: 8-14. Bosch, D. and J. Tommassen. 1987. Effects of linker insertions on the biogenesis and functioning of the Escherichia coli outer membrane pore protein PhoE. Mol. Gen. Genet. 208: 485-489. Boulain, J. C., A . Charbit and M . Hofnung. 1986. Mutagenesis by random linker insertion into the lamB gene of Escherichia coli K12. Mol. Gen Genet. 205: 339-348. Braun, V . and U . Sieglin. 1970. The covalent murein-lipoprotein structure of the Escherichia coli cell wall. The attachment site of the lipoprotein on the murein. Eur J. Biochem. 13: 336-346. Bryan, L . E.. 1979. Resistance to antimicrobial agents: the general nature of the problem and the basis of resistance. In "Pseudomonas aeruginosa: clinical manifestations of infection and current therapy." (R. G. Doggett, ed.), pp. 219-270. Academic press. New York. Charbit, A . , J. C. Boulain, A . Ryter and M . Hofnung. 1986. Probing the topology of a bacterial membrane protein by genetic insertion of a foreign epitope; expression at the cell surface. E M B O J. 5: 3029-3037. Cowan, S. W., T. Schirmer, G. Rummel, M . Steiert, R. Ghosh, R. A . Pautit, J. N . Jansonius, and J. P. Rosenbusch. 1992. Crystal structures explain functional properties of two E. coli porins. Nature. 358: 727-733. Cox, C. D.. 1980. Iron uptake with ferripyochelin and ferric citrate by Pseudomonas aeruginosa. J. Bacteriol. 142: 581-587. Cox, C. D. and P. Adams. Siderophore activity of pyoverdin for Pseudomonas aeruginosa. Infect. Immun. 48: 130-138. Darveau, R. P., R. E. W. Hancock and R. Benz. 1984. Chemical modification of the anion selectivity of the PhoE porin from the Escherichia coli outer membrane. Biochim. Biophys. Acta. 774: 67-74. 109 Dassa, E. and M . Hofnung. 1985. Homologies between integral proteins of the inner membrane of binding protein transport systems in enterobacteria. Ann. Inst. Pasteur Microbiol. 136A: 281-288. Filloux, A. , M . Bally, C. Soscia, M . Murgier and A. Lazdunski. 1988. Phosphate regulation in Pseudomonas aeruginosa: cloning of the alkaline phosphatase gene and identification of the phoB- and phoR-like genes. Mol . Gen. Genet. 212: 510-513. Garavito, R. M . and J. P. Rosenbusch. 1986. Isolation and crystallization of bacterial porin. Methods Enzymol. 125: 309-328. Hancock, R. E. W. and A. M . Carey. 1980. Protein DI - a glucose-inducible, pore-forming protein from the outer membrane of Pseudomonas aeruginosa. FEMS Microbiol. Lett. 8: 105-109. Hancock, R. E. W., K. Poole and R. Benz. 1982. Outer membrane protein P of Pseudomonas aeruginosa: regulation by phosphate deficiency and formation of small anion-specific channels in lipid bilayer membranes. J. Bacteriol. 150: 730-738. Hancock, R. E. W. and R. Benz. 1986. Demonstration and chemical modification of a specific phosphate binding site in the phosphate-starvation-inducible outer membrane porin protein P of Pseudomonas aeruginosa. Biochim. Biophys. Acta. 860: 699-707. Hancock, R. E. W.. 1987a. Role of porins in the outer membrane permeability. J. Bacteriol. 169: 929-933. Hancock, R. E. W. 1987b. Model membrane studies of porin function. In "Bacterial outer membranes as model systems. (Inouye, M . , ed.) pp. 187-225. John Wiley and Sons. New York. Hancock, R. E. W., C. Egli, R. Benz and R. J. Siehnel. 1992. Overexpression in Escherichia coli and functional analysis of a novel selective PPj porin, OprP, from Pseudomonas aeruginosa. J. Bacteriol. 174: 471-476. Hancock, R. E. W. and N . Karunaratne. 1990. LPS integration into outer membrane structures. In "Cellular and molecular aspects of endotoxin reactions."(Novotny, A. , Spitzer, J. J. and E. J. Ziegler, eds.) Vol. 1: pp. 328-342. A S M press. Washington. Hancock, R. E. W., C. Egli and N . Karunaratne. 1994. Molecular oranization and structural role of outer membrane molecules. In "Bacterial cell envelope" (Ghuysen, J. M . and R. Hakenbeck, eds.) pp. 263-279. Elsievier Science Publishers B V Amsterdam. Heine, H . G., J. Kyngdon and T. Ferenci. 1987. Sequence determinants in the lamB gene of Escherichia coli influencing the binding and pore selectivity of maltoporin. Gene. 53: 287-292. 110 Higuchi, R., B. Krummel and R. K. Saiki. 1988. A general method of in vitro preparation and specific mutagenesis of D N A fragments: study of protein and D N A interactions. Nucleic Acids Res. 16: 7351-7367. Horwitz, M . A . and S. C. Silverstein. 1980. Influence of the Escherichia coli capsule on complement fixation and on phagocytosis and killing by human phagocytes. J. Clin. Invest. 65: 82-94. Ho, S. N . , H . D. Hunt, R. M . Horton, J. K. Pullen and L. R. Pease, f 989. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene. 77: 51-59. Hofstra H., M . J. D. van Tol and J. Dankert. 1979. Immunofluorescence detection of the major outer membrane protein II in Escherichia coli O26K60. FEMS Microbiol. Lett. 6: 147-150. Huang, H . and R. E. W. Hancock 1993. Genetic definition of the substrate selectivity of outer membrane porin protein OprD of Pseudomonas aeruginosa. J. Bacteriol. 175: 7793-7800. Huang, H. , D. Jeunteur, F. Pattus and R. E. W. Hancock, f995. Membrane topology and site-specific mutagenesis of Pseudomonas aeruginosa porin OprD. Mol. Microbiol. 16: 931-941. Huang, H. and R. E. W. Hancock. 1996. The role of specific surface loops in determining the function of the imipenem-specific pore protein OprD of Pseudomonas aeruginosa. J. Bacteriol. 178: 3085-3090. Jeanteur, D., J. H . Lakey and F. Pattus. 1991. The bacterial superfamily: sequence alignment and structure prediction. Mol. Microbiol. 5: 2153-2164. Karshikoff, A. , V . Spassov, S. W. Cowan, R. Ladenstein and T. Schirmer. 1994. Electrostatic properties of two porin channels from Escherichia coli. J. Mol . Biol. 240: 372-384. Kellerman, O. K. and T. Ferenci. 1982. Maltose-binding protein of Escherichia coli. Methods Enzymol. 90 Pt E: 459-463. Kristjansson, J. K. and T. C. Hollocher. 1979. Substrate binding site for nitrate reductase of Escherichia coli is on the inner aspect of the membrane. J. Bacteriol. 137: 1227-f233. Koebnik, R. and V. Braun. 1995. Insertion derivatives containing segments of up to 16 amino acids identify surface- and periplasm-exposed regions of FhuA outer membrane receptor of Escherichia coli K-12. J. Bacteriol. 175: 826-839. Lee, D. R. and C. A . Schaitman. 1980. Comparison of outer membrane porin proteins produced by Escherichia coli and Salmonella typhimurium. 142: f Of 9-f 022. I l l Lee, T. Y . , K. Makino, H. Shinagawa, M . Amemura and A. Nakata. 1989. Phosphate regulon in members of the family Enterobacteriaceae; comparison of the phoB-phoR operons of Escherichia coli, Shigella dysenteria and Klebsiella pneumonia. J. Bacteriol. 171: 6593-6599. L i , X . Z., H . Nikaido and K. Poole. 1995. Role of mexA-mexB-oprM in antibiotic efflux in Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 39: 1948-1953. Marcus, H . and N . R. Baker. 1985. Quantitation of adherance of mucoid and nonmucoid Pseudomonas aeruginosa to hamster tracheal epithelium. Infect. Immun. 47: 723-729. Meyer, J. M . , D. Hohnadel, A. Khan and P. Cornells. 1990. Pyoverdin-facilitated iron uptake in Pseudomonas aeruginosa: immunological characterization of the ferripyoverdin receptor. Mol . Microbiol. 4: 1401-1405. Misra, R. and S. A . Benson. 1988. Genetic identification of the OmpC porin of Escherichia coli K-12. J. Bacteriol. 170: 3611-3617. Mizuno, T., M . Y . Chou and M . Inouye. 1983. A comparative study on the genes for three porins of the Escherichia coli outer membrane. D N A sequences of the osmoregulated ompC gene. J. Biol Chem. 258: 6932-6940. Morihara, K. and Ff. Tsuzuki. 1977. Production of protease and elastase by Pseudomonas aeruginosa strains isolated from patients. Infect. Immun. 15: 679-685. Nikaido, H . and E. Y . Rosenberg. 1983. Porin channels in Escherichia coli: studies with liposomes reconstituted from purified proteins. J. Bacteriol. 153: 241-252. Nikaido, H . 1986. Transport through the outer membrane of bacteria. Methods Enzymol. 125: 265-278. Nikaido, H. and M . Vaara. 1987. Outer membrane. In " Escherichia coli and Salmonella typhimurium: cellular and molecular biology ". (Neidhardt F. C , ed.) pp. 7-22. A S M press. DC. Nikaido, H. , K. Nikaido and S. Harayama. 1991. Identification and characterization of porins in Pseudomonas aeruginosa. J. Biol. Chem. 266: 770-779. Nikaido H . 1992. Porins and specific channels of bacterial outer membranes. Mol . Microbiol. 6: 435-442. Overbeek, N . and B. Lugtenberg. 1980. Expression of outer membrane protein e of Escherichia coli K12. Eur. J. Biochem. FEBS Lett.112: 229-232. 112 Palleroni, N . J. 1986. Taxonomy of the Pseudomonas In " The bacteria: a treatise on structure and function. Vol . X . The biology of Pseudomonas." (Sokatch, J. R., ed.) pp. 3-25. Academic press. Florida. Pedersen, S. S., F. Espersen, N . Hoiby and G. H. Shand. 1989. Purification, characterization, and immunological cross-reactivity of alginates produced by mucoid Pseudomonas aeruginosa from patients with cystic fibrosis. J. Clin. Microbiol. 27: 691-699. Poole, K. and R. E. W. Hancock. 1984. Phosphate transport in Pseudomonas aeruginosa: involvement of a periplasmic phosphate-binding protein. Eur. J. Biochem. 144: 607-612. Poole, K . and R. E. W. Hancock. 1986. Phosphate-starvation-induced outer membrane proteins of members of the families Enterobacteriaceae and Pseudomonadaceae: demonstration of immunological cross-reactivity with an antiserum specific for porin protein P of Pseudomonas aeruginosa. J. Bacteriol. 165: 987-993. Praitt, B. A . and A. T. McManus. 1984. Opportunistic infections in severely burned patients. Am. J. Med. 76 (suppl): 146-154. Rocque, W. J. and E. J. McGroarty. 1990. Structure and function of an OmpC deletion mutant porin from Escherichia coli K-12. Biochemistry. 29: 5344-5351. Sambrook, J., E. F. Fritsch and T. Maniatis. 1989. Molecular cloning: a laboratory manual. Second edition. Cold Spring Harbor Laboratory. Cold Spring Harbor New York. Scher, F. M . , J.W. Kloepper, C. Singleton, I. Zaleska and M . Laliberte. 1988. Colonization of roots by Pseudomonas and Serratia species: relationship to bacterial motility, chemotaxis, and generation time. Phytopathology. 78: 1055-1059. Schimpff, S. C , M . Moody and V. M . Young. 1970. Relationship of colonization with Pseudomonas aeruginosa to development of Pseudomonas bacteremia in cancer patients. Antimicrob. Agents Chemother. 10: 240-244. Schirmer, T., T. A . Keller, Y . F. Wang and J. P. Rosenbusch. 1995. Structural basis for sugar translocation through maltoporin channels at 3.1 A resolution. Science. 267: 512-514. Scholten, M . , R. Janssen, C. Bogaarts, J. van Strien and J. Tommassen. 1995. The pho regulon of Shigella flexneri. 15: 247-254. Schulein, K. and R. Benz. 1990. LamB (maltoporin) of Salmonella typhimurium: isolation, purification and comparison of sugar binding with LamB of Escherichia coli. Mol . Microbiol. 5: 2153-2164. 113 Schwartz, M . 1987. The maltose regulon. In " Escherichia coli and Salmonella typhimurium: cellular and molecular biology ". (F. C. Neidhardt, ed.) pp. 7-22. A S M press. Washington, DC. Siehnel, R. J., E. A . Worobec and R.E. W. Hancock. 1988. Cloning of the Pseudomonas aeruginosa porin protein P gene: evidence for a linked region of homology. J. Bacteriol. 170: 2312-2318. Siehnel, R. J., N . L . Martin and R. E. W. Hancock. 1990. Sequence and relatedness in other bacteria of the Pseudomonas aeruginosa oprP gene coding for the phosphate-specific porin P. Mol. Microbiol. 4: 831-838. Siehnel, R. J., C. Egli and R. E. W. Hancock. 1992. Polyphosphate-selective porin OprP of Pseudomonas aeruginosa; expression, purification and sequence. Mol . Microbiol. 4: 831-838. Silverstein, M . A . and S. C. Horowitz. 1980. Influence of the Escherichia coli capsule on complement fixation and on phagocytosis and killing by human phagocytes. J. Clin. Invest. 65: 82-94. Smith, M . D., M . Petrak, P. D. Boucher, K. N . Barton, L . Carter, G. Reddy, E. Blanchly-Dyson, M . Fortes, J. Price, K. Verner and R. B. McCauley. 1995. Lysine residues at positions 234 and 236 in yeast porin are involved in its assembly into mitochondrial outer membrane. J. Biol. Chem. 270: 28331-28336. Sriyosachati, S. and C. D. Cox. 1986. Siderophore-mediated iron acquisition from transferrin by Pseudomonas aeruginosa. Infect. Immun. 52: 885-891. Stock, J. B., B. Rauch, S. Roseman. 1977. Periplasmic space in Salmonella typhimurium. and Escherichia coli. J. Biol. Chem. 252: 7850-7861. Struyve, M . , J. Visser, H . Adriaanse, R. Benz and J. Tommassen 1993. Topology of PhoE porin: the 'eyelet' region. Mol. Microbiol. 7: 131-140. Sugawara, E. and H. Nikaido. 1992. Pore-forming activity of OmpA protein of Escherichia coli. J. Biol . Chem. 267: 2501-2511. Sugawara, E. and H . Nikaido. 1994. OmpA protein of Escherichia coli outer membrane occurs in open and closed forms. J. Biol. Chem. 269: 17981-17987. Tanoue, E., S. Nishiyama, M . Kamo, A. Tsugita. 1995. Bacterial membranes: possible source of a major dissolved protein in seawater. Geochim. Cosmochim. Acta. 59: 2643-2648. Taylor, W. R., D. T. Jones, and N . M . Green, 1994. A method for alpha-helical integral membrane protein fold prediction. Proteins. 18: 281-294. 114 Towbin, M . T., T. Stachlin and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. U S A 76: 4350-4354. Trias, J. and H . Nikaido. 1990a. Protein D2 channel of the Pseudomonas aeruginosa outer membrane has a binding site for basic amino acids and peptides. J. Biol. Chem. 265: 15680-15684. Trias, J. and H . Nikaido. 1990b. Outer membrane protein D2 catalyzes facilitated diffusion of carbapenems and penums through the outer membrane of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 34: 52-57. Tuomanen, E.. 1986. Newly made enzymes determine ongoing cell wall synthesis and the antibacterial effects of cell wall synthesis inhibitors. J. Bacteriol. 167: 535-543. van der Ley, P., P. Burm, M . Agterberg, J. van Meersbergen and J. Tommassen. 1987. Analysis of structure-function relationships in Escherichia coli K12 outer membrane porins with the aid of ompC-phoE and phoE-ompC hybrid genes.Mol. Gen. Genet. 209: 585-591. Vasil, M . L. . 1986. Pseudomonas aeruginosa: biology, mechanisms of virulence, epidemiology. J. Pediatr. 108: 800-805. Wandersman, C , M . Schwartz and T. Ferenci. 1979. Escherichia coli mutants impaired in maltodextrin transport. J. Bacteriol. 140: 1-13. Wanner, B. L . and P. Latterll. 1980. Mutants affected in alkaline phosphatase expression: evidence for multiple positive regulators of the phosphate regulon in Escherchia coli. Genetics. 96: 353-366. Wanner, B. L. . 1993. Gene regulation by phosphate in enteric bacteria. J. Cell. Biochem. 51: 47-54. Weiss, M . S., T. Wacker, J. Weckesser, W. Welte and G. E. Schulz. 1990. The three-dimensional structure of porin from Rhodobacter capsulatus at 3 A resolution. FEBS Lett. 267: 268-272. Welsh, M . J. and C. M . Liedke. 1986. Chloride and potassium channels in cystic fibrosis airway epithelia. Nature. 322: 467-470. Worbec, B., N . L . Martin, W. McCubbin, C. Kay, G. Brayer and R. E. W. Hancock. 1988. Large-scale purification and biochemical characterization of crystallization-grade porin protein P from Pseudomonas aeruginosa. Biochim. Biophys. Acta. 939: 366-374. 115 Wong, R. S.-Y., H. Jost and R. E. W. Hancock. 1993. Linker-insertion mutagenesis of Pseudomonas aeruginosa outer membrane protein OprF. Mol. Microbiol. 10: 283-292. Wong, R. S.-Y., R. A . Wirtz and R. E. W. Hancock. 1995. Pseudomonas aeruginosa outer membrane protein OprF as an expression vector for foreign epitopes: the effects of positioning and length on the antigenicity of the epitope. Gene. 158: 55-60. 116 APPENDIX 117 Figure 21. Schematic diagram of the p-barrel structure of the E. coli porin OmpF. Thick arrows represent transmembrane p-strands; L, surface loops; T, periplasmic turns; N, N-terminus; C, C-terminus. Reproduced with permission from Cowan et al. 1992. 118 Figure 22. Conceptual model for the structure of OprP based upon the p-turn prediction method of Paul and Rosenbusch (1985). (Nancy L. Martin, Ph.D. thesis) 119 Figure 23. Hydropathy plots (Kyte and Doolittle, 1982) for OprP. Areas above the dotted line correspond to hydrophobic regions of the protein while those areas below the dotted line indicate hydrophilic regions (Nancy L. Martin, Ph.D thesis). 

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0088292/manifest

Comment

Related Items