R.E.W. Hancock Laboratory: PORINS OF GRAM NEGATIVE BACTERIA

Porins of Gram negative bacteria

Model membrane studies using liposomes and black lipid bilayer membranes were utilized to characterize in detail a broad range of porins from Gram negative bacteria. These include the general porins OprF of P. aeruginosa, phosphate-regulated porin PhoE, virulence plasmid associated porin K, and general porins NmpC and CE1198 porin of E. coli; and the general porins of Yersinia pestis, Campylobacter fetus and C. jejuni, Aeromonas salmonicida and A. hydrophila, Pseudomonas cepacia, Pseudomonas syringae, Spirochaetia aurantia,Wolinella recta, Treponema denticola, Helicobacter pylori, Treponema pallidum, Xenorhabdis and Leptospira. Characterization of the copper binding NosA porin from Pseudomonas stutzeri and the anion-specific major outer membrane protein from Bordatella pertussis. Characterization and gene cloning of the phosphate-specific OprP porin, pyrophosphate-specific OprO porin of P. aeruginosa, the imipenem-specific porin OprD and the glucose-specific OprB porin of P. putida. Demonstration of a specific binding site within the channel for each of the substrate-specific porins and measurement of the substrate binding affinity. First description of large families of porins in Helicobacter pylori (the Hop/Hor family), and P. aeruginosa (the OprM, OprD and TonB families).

Highlights.

1. OprF. Detailed characterization of OprF of P. aeruginosa. A controversy in the literature had centred around whether or not OprF is the major porin for P. aeruginosa and whether the exclusion limit of the P. aeruginosa outer membrane was >3000 daltons (as proposed by Hancock and Nikaido) or <200 daltons (as proposed by Nakae). Since this controversy was based on model membrane studies, an in vivo method was devised for measuring the exclusion limit for sugars. An OprF deficient mutant of P. aeruginosa was created by in vitro insertion mutagenesis and gene replacement techniques. A raffinose operon from an E. coli plasmid was cloned into P. aeruginosa and shown to permit growth on the trisaccharide raffinose and the disaccharide melibiose in an outer membrane-limited fashion, thus indicating an exclusion limit greater than 500 daltons. The role of OprF and the outer membrane was demonstrated by the slower growth of the OprF deficient mutant on both substrates and the concentration and sugar size dependence of the growth rate. These data were confirmed by measuring rates of plasmolysis using light scattering by cells resuspended in high concentrations of saccharides (as monitored by electron microscopy). A limiting pore diameter of 1.54 nm was determined for OprF (cf 1.18 nm for OmpF of E. coli). A structural model has been devised for OprF utilizing TnphoA mutagenesis and subcloning methods, chemical and proteolytic peptide isolation, linker, and epitope, insertion mutagenesis, and overlapping peptide synthesis in conjunction with mapping of monoclonal antibodies as well as circular dichroism spectra and structure predictive methods. A role for OprF in cell structure and shape determination was revealed by examining the properties of an interposon mutant, and the homology of the carboxy terminus of OprF to OmpA was demonstrated by sequence comparison, immunological means and complementation. Use of OprF as a target for immunotherapeutic intervention was demonstrated. Epitope insertion mutagenesis was utilized to create variants of the oprF gene which accepted extra amino acids (epitopes) inserted into the surface loops of OprF (added by ligation of oligonucleotides into the gene after "marking" these sites with Pst1 sites). A variation has been designed to permit the identification of epitopes useful as vaccines or diagnostics. Characterization of OprF-deficient, multiple antibiotic-resistant clinical and laboratory mutants was undertaken. Evidence for the role of an ECF-Sigma factor SigX, that regulates OprF expression, has been obtained. Recently we demonstrated that the protein could be truncated at amino acid 162 to yield a beta-barrel N-terminal domain that was modeled in 3 dimensions (appearing on the cover of the September 2000 issue of Journal of Bacteriology) by threading (using Insight II to thread the sequence to the structure of the crystallized protein OmpA (despite only 9% identity in these domains). This has indicated that this portion of OprF forms small channels. We also demonstrated through a deletion mutagenesis study that the C-terminal domain is involved in peptidoglycan binding, and cell structure and shape determination. We presume, based on these studies, that it somehow is involved in formation of larger channels by intact OprF.

2. OprP. This phosphate-specific porin has become one of the best-characterized anion channels known. A combination of sophisticated model membrane studies with chemical modifications, mathematical modelling, protein chemistry, and classical and molecular genetics have been utilized to characterize OprP. OprP is part of the phosphate starvation-inducible regulon and has a defined role in high affinity phosphate uptake, together with a periplasmic phosphate-binding protein which we have also characterized biochemically and genetically. OprP has been purified and reconstituted into planar lipid bilayers, and shown to form channels which are specific for anions over cations and which preferentially bind phosphate compared to other anions. The protein has been crystallized, and a 2-dimensional model as a 16-stranded ß-barrel constructed on the basis of epitope insertion mutagenesis. A critical lysine (K-121) was demonstrated by site-directed mutagenesis and model membrane studies to be located in the third surface loop between ß-strands 5 and 6, and involved in the phosphate binding site. Two other lysines involved funnelling phosphates to this binding site were similarly characterized. We have crystallized this protein and hope to be able to determine a 3D structure model soon.

3. OprD. Pseudomonas aeruginosa clinical isolates that have developed resistance to the broad spectrum beta-lactam antibiotic, imipenem, lack an outer membrane protein OprD. Trias and Nikaido provided data that showed that OprD is a specific porin for imipenem and basic amino acids (of which imipenem is an analogue). We have confirmed this by planar lipid bilayer studies. The gene for OprD and its regulatory region, opdE were cloned and sequenced. Interposon mutants and hyperexpressing derivatives were isolated and used to genetically prove that OprD could only be utilized by imipenem and related carbapenem b-lactam antibiotics but not other antibiotics. Sophisticated alignment and membrane topology predictions permitted alignment of OprD with the general porin superfamily, a first for a specific porin, and consequent prediction of a 2D membrane topology model. PCR-mediated site specific deletions were used to delete external loop regions (deletions in membrane spanning segments are non-permissive). Seven of 8 loop regions were proven to be accurately predicted. Deletion of 8 amino acids of loop 2 (3 separate deletions spanning the entire loop) or loop 3 caused loss of the imipenem binding site. Deletion of 8 amino acids of loops 5, 7 or 8 resulted in larger channels that resulted in enhanced susceptibility to antibiotics other than imipenem. These data support the general model for OprD and assign specific roles to selected regions of OprD. We also demonstrated that the oprD gene is highly regulated both positively and negatively with factors such as ArgR, MexT, salicylate, catabolite repression and certain C-and N-sources influencing expression. This protein has also been crystallized and we hope to solve the structure. Analysis of the Pseudomonas aeruginosa genome sequence revealed 18 homologs. We have started to characterize these and have evidence for the role of one in vanillate uptake.