Identification and characterization of porins in Pseudomonas aeruginosa.

Earlier studies have shown that the major porin species in Pseudomonas aeruginosa outer membrane is protein F (OprF), which produces channels wider than those produced by Escherichia coli porins. In contrast, Yoshihara and Nakae ((1989) J. Biol. Chem. 264, 6297-6301) reported that protein F has no pore-forming activity as measured by the flux of L-arabinose, and that the channels in P. aeruginosa outer membrane, being produced by proteins C, "D," and "E," are much narrower than E. coli porin channels. In this study, we followed the protein purification scheme of Yoshihara and Nakae as closely as possible, and found that protein F had a specific activity for pore formation similar to that of proteins D1, D2, and E2. Furthermore, proteoliposome reconstitution assays showed conclusively that the channels formed by protein F, as well as by unfractionated outer membranes, allowed the diffusion of a tetrasaccharide, stachyose, at a significant rate, indicating that these channels are much larger than E. coli porin channels. It appears likely that in the study of Yoshihara and Nakae protein F was inadvertently inactivated during purification. We further suggest a hypothesis that resolves the apparent conflict between the presence of large diameter channels and the low permeability of the outer membrane in P. aeruginosa.     

Identification of three porins in the outer membrane of Bacteroides fragilis

Abstract Four outer membrane proteins were purified to homogeneity from isolated outer membranes of Bacteroides fragilis ; three ( M _r 51000, 92000 and 125 000) had pore-forming activity in reconstituted liposomes as determined by swelling assay. Membrane vesicles containing the M _rmr 55 000 outer membrane protein showed no detectable pore-forming activity. The three B. fragilis porins formed pores that allowed the penetration of uncharged saccharides of M _r lower than 340–400, even though the efficiency of solute diffusion showed slight differences. The diffusion rates of glucose through the porins appeared to be lower than those through Escherichia coli porins.     

In vitro demonstration by the rate assay of the presence of small pore in the outer membrane of Pseudomonas aeruginosa

Determination of the rates of saccharide diffusions by the proteoliposomes showed that the outer membrane of Pseudomonas aeruginosa only possesses small diffusion pores and that protein F might have not been involved in the pore formation. Proteoliposomes containing stachyose or Dextan T-10 showed the same relative diffusion rates as measured by the liposome swelling method. Slopes of the lines, diffusion rate vs saccharide Mr, in the liposomes made of the P. aeruginosa and E. coli B outer membranes appeared to be -7.4 and -3.5, respectively. Intercepts of the lines with x-axis in the liposomes containing the P. aeruginosa and E. coli B outer membrane appeared to be about Mr, 220 and 320, respectively. Relative diffusion rates of saccharides through the liposome membranes reconstituted from the protein F-deficient outer membrane were superimposable with that of the protein F-sufficient outer membrane.     

In vitro assembly of the functional porin trimer from dissociated monomers in Pseudomonas aeruginosa

The molecular weights of monomeric and oligomeric forms of the newly identified porins, protein D2, of the outer membrane of Pseudomonas aeruginosa appeared to be 47,000 and 137,000, respectively, as determined by the light scattering technique. Presence of the trimeric aggregates of the homologous subunits in the intact outer membrane, the liposome membrane, and the non-ionic surfactant were confirmed through cross-linking experiments and immunoblotting techniques. The protein D2 monomers prepared in 0.1% of sodium dodecyl sulfate at 23 degrees C spontaneously reassembled into the trimeric aggregate when the surfactant dropped below critical concentration. The diffusion rates of saccharides and beta-lactam antibiotics through the liposome membranes reconstituted from the reassembled protein D2 trimers were indistinguishable from those of the native protein D2. This study shed some light on the porin trimer assembly as well as on the mechanism of carbapenem diffusion through the protein D2 pores.     

Transmembrane permeability channels in vesicles reconstituted from single species of porins from Salmonella typhimurium.

Aggregates of the "major" outer membrane proteins, "porins," of Salmonella typhimurium form diffusion channels in reconstituted vesicle membranes. The aggregate consists of three species of porins with apparent molecular weights of 34,000, 35,000, and 36,000 when active aggregates are subjected to sodium dodecyl sulfate-acrylamide gel electrophoresis after heating in the presence of sodium dodecyl sulfate (Nakae, J. Biol. Chem. 251:2176-2178, 1976). Single species of porins were isolated by solubilization of membranes and subsequent gel filtration in the presence of sodium dodecyl sulfate from the mutant strains of Salmonella typhimurium that produced only single species of porin. The single species of porins of either 34,000, 35,000, or 36,000 daltons formed diffusion channels when assayed for sucrose permeability in the vesicle membranes reconstituted from porins, phospholipids, and lipopolysaccharides. The exclusion limits of the pores made of single species of porins were not distinguishable from each other and from the exclusion limits of the pores made of the porin aggregates from the wild-type strain, when the permeability of vesicle membranes to radioactive di-, tri-, and tetrasaccharides and to various sizes of radioactive polyethylene glycol was determined. Porin-deficient mutants produced residual amounts of porin amounting to 1 to 5% that produced by the parent strain. This residual porin made diffusion channels when the isolated porins were incorporated into the vesicle membrane and assayed for permeability of saccharides.     

Purification and characterization of the pore forming protein of yeast mitochondrial outer membrane

One of the major outer membrane proteins of yeast mitochondria was isolated and purified. It migrated as a single band with an apparent molecular weight of 30 kDa on a SDS-electrophoretogram. When reconstituted in lipid bilayer membranes the protein formed pores with a single channel conductance of 0.45 nS in 0.1 M KCl. The pores had the characteristics of general diffusion pores with an estimated diameter of 1.7 nm. The pore of mitochondrial outer membranes of yeast shared some similarities with the pores formed by mitochondrial and bacterial porins. The pores switched to substates at voltages higher than 20 mV. The possible role of this voltagedependence in the metabolism of mitochondria is discussed.     

Isolation and characterization of transposon Tn5-induced mutants of Pseudomonas perfectomarina defective in nitrous oxide respiration.

Outer membranes of Haemophilus influenzae type b were fractionated to yield Triton X-100-insoluble material and lipopolysaccharide and phospholipids. Liposomes reconstituted from lipopolysaccharide and phospholipids were impermeable to sucrose (Mr, 342) and to a high-molecular-weight dextran (average Mr, 6,600). When the Triton X-100-insoluble material was introduced into the reconstituted liposomes, the vesicles became permeable to sucrose, raffinose (Mr, 504), and stachyose (Mr, 666) and fully retained dextrans of Mr greater than 1,500. Inulin (average Mr, 1,400) was tested for its efflux from the reconstituted outer membrane vesicles; 62% of the added inulin was trapped. The molecular weight exclusion limit for the outer membrane of H. influenzae type b was therefore estimated at approximately 1,400. A protein responsible for the transmembrane diffusion of solutes was purified from H. influenzae type b by extraction of whole cells with cetyl trimethyl ammonium bromide. When this extract was passed over DEAE-Sepharose, three protein-containing peaks (I, II, and III) were eluted. Peaks I and II contained mixtures of proteins as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis; when tested for their pore-forming properties, these proteins were unable to render liposomes of lipopolysaccharide and phospholipid permeable to sucrose. Peak III contained only one molecular species of protein of molecular weight 40,000; this protein acted as a porin in reconstituted vesicles. The molecular weight exclusion limit for 40,000-molecular-weight protein matched the estimate of approximately 1,400 which was determined for outer membranes. A series of homologous saccharides of increasing degree of polymerization was prepared from agarose by hydrolysis with beta-agarase and fractionation on gel filtration chromatography. These oligosaccharides of Mr, 936, 1,242, 1,548, and 1,854 were assayed for retention by the complete vesicles containing 40-kilodalton protein and lipopolysaccharide and phospholipids. All of these oligosaccharides were lost by efflux through the porin. Since the molecular conformation of the largest oligosaccharide is an elongated semirigid helix, it is suggested that the pore formed by the 40-kilodalton protein does not act as a barrier to the diffusion of this compound.     

Purification and characterization of two kinds of porins from the Enterobacter cloacae outer membrane.

Two major outer membrane proteins of Enterobacter cloacae 206 were purified and identified as porins by using reconstituted vesicles. The 37-kilodalton porin forms a channel with a radius of 0.6 nm, which prefers positively charged substances to negatively charged ones, whereas the 39- to 40-kilodalton porin forms a larger channel with a radius of 0.8 nm, which has weaker selectivity for electric charges.     

Ion selectivity of gram-negative bacterial porins.

Twelve different porins from the gram-negative bacteria Escherichia coli, Salmonella typhimurium, Pseudomonas aeruginosa, and Yersinia pestis were reconstituted into lipid bilayer membranes. Most of the porins, except outer membrane protein P, formed large, water-filled, ion-permeable channels with a single-channel conductance between 1.5 and 6 nS in 1 M KCl. The ions used for probing the pore structure had the same relative mobilities while moving through the porin pore as they did while moving in free solution. Thus the single-channel conductances of the individual porins could be used to estimate the effective channel diameters of these porins, yielding values ranging from 1.0 to 2.0 nm. Zero-current potential measurements in the presence of salt gradients across lipid bilayer membranes containing individual porins gave results that were consistent with the conclusions drawn from the single-channel experiments. For all porins except protein P, the channels exhibited a greater cation selectivity for less mobile anions and a greater anion selectivity for less mobile cations, which again indicated that the ions were moving inside the pores in a fashion similar to their movement in the aqueous phase. Three porins, PhoE and NmpC of E. coli and protein P of P. aeruginosa, formed anion-selective pores. PhoE and NmpC were only weakly anion selective, and their selectivity was dependent on the mobility of the ions. In contrast, cations were unable to enter the selectivity filter of the protein P channel. This resulted in a high anion selectivity for all salts tested in this study. The other porins examined, including all of the known constitutive porins of the four gram-negative bacteria studied, were cation selective with a 3- to 40-fold preference for K+ ions over Cl- ions.     

Functional expression in Escherichia coli and membrane topology of porin HopE, a member of a large family of conserved proteins in Helicobacter pylori

HopE is one of the smallest members of a family of 31 outer membrane proteins in Helicobacter pylori and has been shown to function as a porin. In this study it was cloned into Escherichia coli where it was expressed in the outer membrane, as confirmed by indirect immunofluorescence using HopE-specific antibodies. HopE purified from E. coli reconstituted channels in planar bilayer membranes that were the same size as those formed by HopE purified from H. pylori. A model of the membrane topology of HopE was constructed and indicated that this protein formed a beta-barrel with 16 transmembrane amphipathic beta-strands. The accuracy of this model was tested by linker insertion mutagenesis, assuming that, like other porins, amino acid insertions were not tolerated in the transmembrane beta-strands but were tolerated in the adjoining loop regions. Generally, the results obtained with a series of 12 insertions of the sequence RSKDV and two substitutions were consistent with the topological model. The preponderance of amino acids that were conserved in the extended family of HopE paralogs were predicted to be within the membrane and comprised 45% of all residues in the membrane.     

The voltage-dependent activity of Escherichia coli porins in different planar bilayer reconstitutions

Because of conflicting results from differing techniques, the degree of voltage sensitivity of Escherichia coli porins in planar bilayers is still a matter of debate. In order to provide the first comparative study, OmpF porin was purified in three ways; firstly as native outer membrane vesicles, secondly as salt-extracted porin trimers in sodium dodecyl sulphate and thirdly as solubilised trimers extracted with octyl-polyoxyethylene (Octyl-POE). These methods represent the major approaches to porin isolation and purification. All three were reconstituted into Schindler-type bilayers. Detergent-solubilised OmpF was also reconstituted into Montal-Mueller- and Mueller-Rudin-type bilayers. In all cases voltage-dependent closing of OmpF was observed. Octyl-POE-extracted PhoE porin was similarly investigated in all three types of planar bilayer. Two membrane-formation techniques appeared genuinely to alter the voltage sensitivity of the porins they contained. Firstly, porins in membranes formed by the Montal-Mueller technique sometimes showed an increase in voltage sensitivity during the first 30 min after bilayer formation. Secondly, membranes formed by the Mueller-Rudin technique on thick polyethylene septa showed both poor solvent drainage and a significantly reduced porin voltage sensitivity.     

Molecular and functional characterisation of the Serratia marcescens outer membrane protein Omp1

Serratia marcescens outer membrane contains three different general diffusion porins: Omp1, Omp2 and Omp3. Omp1 was cloned and sequenced and it shows a great homology to the family of outer membrane porins that comprises the general porins of enteric bacteria. The gene for Omp1 was transferred into an expression plasmid and was expressed in Escherichia coli UH302 (E. coli UH302 pOM100), a porin deficient strain. Its expression confers a higher susceptibility towards different antibiotics to this strain. Omp1 was purified to homogeneity from outer membrane of E. coli UH302 pOM100. Reconstitution of the purified protein into black lipid bilayers demonstrated that it is a channel-forming component with a single-channel conductance of approximately 2 nS in 1 M KCl similar to that of other porins from enteric bacteria. Omp1 is slightly cation-selective. Its homology to already crystallised members of the family of enteric porins whose three-dimensional-structures are known and allowed the design of a topology model for Omp1. The charge distribution within a porin monomer is similar as in other general diffusion pores. The positively charged amino acids localised at the beta-strands opposite the external loop L3, which restrict the pore diameter in the porin monomer.     

Purification of a protein having pore forming activity from the rat liver mitochondrial outer membrane

A protein with pore-forming activity has been isolated from the outer membrane of rat liver mitochondria. The purification involves sucrose gradient centrifugation, differential centrifugation in the presence of Triton X-100, and DEAE-Sepharose and CM-Sepharose chromatography. The yield of the purified protein was approx. 2% of the total outer membrane proteins. The protein, when inserted into soya bean phospholipid vesicles, increases the [3H]sucrose permeability of the vesicles but had no effect on the permeability of high-molecular-weight [14C]dextran (Mr 70 000). The protein is very active, since as little as 3-4 micrograms of protein per mg of phospholipid is required for the complete release of [3H]sucrose from the vesicles. Sucrose diffusion channels could not be reconstituted with other membrane proteins such as rat liver cytochrome oxidase or cytochrome b5. Purified pore protein revealed a single band of apparent Mr 30000 when resolved by sodium dodecyl sulphate/polyacrylamide-gel electrophoresis. This polypeptide could be further resolved by isoelectric focusing into a major (pI7.9) and two relatively minor (pI7.6 and 7.2) components. Proteolytic mapping with V8 proteinase from Staphylococcus aureus suggests that these probably represent a single component showing charge heterogeneity. The reason for the charge heterogeneity is not known. The amino acid composition of the protein revealed 47.8% polar amino acids with a relatively high lysine content.     

Characterization of the Pseudomonas aeruginosa Lol system as a lipoprotein sorting mechanism

Escherichia coli lipoproteins are localized to either the inner or the outer membrane depending on the residue that is present next to the N-terminal acylated Cys. Asp at position 2 causes the retention of lipoproteins in the inner membrane. In contrast, the accompanying study (9) revealed that the residues at positions 3 and 4 determine the membrane specificity of lipoproteins in Pseudomonas aeruginosa. Since the five Lol proteins involved in the sorting of E. coli lipoproteins are conserved in P. aeruginosa, we examined whether or not the Lol proteins of P. aeruginosa are also involved in lipoprotein sorting but utilize different signals. The genes encoding LolCDE, LolA, and LolB homologues were cloned and expressed. The LolCDE homologue thus purified was reconstituted into proteoliposomes with lipoproteins. When incubated in the presence of ATP and a LolA homologue, the reconstituted LolCDE homologue released lipoproteins, leading to the formation of a LolA-lipoprotein complex. Lipoproteins were then incorporated into the outer membrane depending on a LolB homologue. As revealed in vivo, lipoproteins with Lys and Ser at positions 3 and 4, respectively, remained in proteoliposomes. On the other hand, E. coli LolCDE released lipoproteins with this signal and transferred them to LolA of not only E. coli but also P. aeruginosa. These results indicate that Lol proteins are responsible for the sorting of lipoproteins to the outer membrane of P. aeruginosa, as in the case of E. coli, but respond differently to inner membrane retention signals.     

Ethylenediaminetetraacetate-extractable protein-lipopolysaccharide complex of Pseudomonas aeruginosa: characterization of protein components

Five major outer membrane proteins (D1, D2, E, G, and H1) of Pseudomonas aeruginosa, but not proteins F (porin), I (lipoprotein), and H2, were detected in high-molecular-weight protein-lipopolysaccharide complex(es) solubilized from sucrose-stabilized cells on exposure to ethylenediaminetetraacetate and tris(hydroxymethyl)aminomethane.     

Isolation and reconstitution of an N-ethylmaleimide-sensitive phosphate transport protein from rat liver mitochondria

An N-ethylmaleimide-sensitive phosphate transport protein has been isolated from rat liver mitochondria, substantially purified, and reconstituted into phospholipid vesicles. Purified inner mitochondrial membrane vesicles depleted of F1-ATPase by urea treatment proved to be the most satisfactory starting material. Treatment of these membrane vesicles with Triton X-100 resulted in solubilization of the phosphate transport protein. Further purification was achieved using hydroxylapatite powder. Polyacrylamide gel electrophoresis of the purified fraction in sodium dodecyl sulfate indicated the presence of two Coomassie blue-staining bands with apparent Mr's of 30,000 and 35,000. Labeling of the 35,000 Mr band by the Pi transport inhibitor diazobenzene sulfonate was reduced markedly by prior treatment of the mitochondria with the inhibitor N-ethylmaleimide. The purified fraction containing both proteins could be reconstituted into liposomes prepared from purified asolectin. Phosphate efflux from these vesicles was inhibited by N-ethylmaleimide, by the impermeant mercurial agent, p-chloromercuribenzoate, and by diazobenzene sulfonate. Treatment of the purified fraction with N-ethylmaleimide prior to incorporation into liposomes resulted in a reconstituted system incapable of catalyzing Pi efflux. These studies summarize the first detailed attempt to purify the Pi/H+ transport system from rat liver mitochondria and emphasize the need to commence the purification with purified inner membrane vesicles depleted of F1-ATPase. In addition, these studies show that the final fraction contains a reconstitutively active transport system which when incorporated into phospholipid vesicles has its essential sulfhydryl groups oriented outward. Finally, it is shown that the purified fraction also contains a 30,000 Mr component.     

Pore formation by the mitochondrial porin of rat brain in lipid bilayer membranes

The porin of the outer membrane of rat-brain mitochondria was isolated and purified. The protein showed a single band of apparent Mr 35,500 on dodecyl sulfate-containing polyacrylamide gels. The incorporation of rat-brain porin into artificial lipid bilayer membranes showed that it is able to form pores with an average single-channel conductance of 400 pS in 0.1 M KCI. The pores were found to be voltage-dependent and switched to substrates at higher transmembrane potentials. The voltage-dependence of the rat brain pore was considerably smaller than that of the other known eukaryotic porins. The possible role of the rat-brain porin in the regulation of transport process across the outer mitochondrial membrane is discussed.     

Diffusion of beta-lactam antibiotics into proteoliposomes reconstituted with outer membranes of isogenic imipenem-susceptible and -resistant strains of Enterobacter cloacae.

The influence of outer membrane (OM) permeability on carbapenem susceptibility was studied in strains of Enterobacter cloacae, a species in which carbapenem resistance depends upon the conjunction of overproduction of the chromosomal cephalosporinase and reduction of OM permeability. Relative trans-OM diffusion rates were measured using the liposome swelling assay. Proteoliposomes were reconstituted with OM from the members of an isogenic set of E. cloacae strains, selected in vivo or in vitro, which produced either porins F and D (wild-type), or F or D only, or neither. For all but one mutant, and compared with the wild-type strain, the respective increases in MICs and decreases in trans-OM diffusion of carbapenems were: nil and 13 to 18%; 4- to 32-fold and 33 to 50%; > or = 64-fold and > or = 90%. Our results suggest (i) that carbapenems (and other beta-lactam antibiotics) diffuse through porins F and D, but more rapidly through porin F, and (ii) that OM permeability is the critical factor in determining the level of MICs of carbapenems for cephalosporinase-overproducing strains of E. cloacae. The OM of one particular low-level carbapenem-resistant and porin F- and D-deficient mutant was at least five times more permeable to carbapenems than the similarly porin-deficient high-level resistant mutants. We infer from this observation the possible existence of an alternative carbapenem penetration pathway which could be associated with two as yet uncharacterized overproduced OM proteins of about 22 and 47 kDa.     

Expression and porin activity of P28 and OMP-1F during intracellular Ehrlichia chaffeensis development

Ehrlichia chaffeensis, an obligatory intracellular gram-negative bacterium, must take up various nutrients and metabolic compounds because it lacks many genes involved in metabolism. Nutrient uptake by a gram-negative bacterium occurs primarily through pores or channels in the bacterial outer membrane. Here we demonstrate that isolated E. chaffeensis outer membranes have porin activities, as determined by a proteoliposome swelling assay. The activity was partially blocked by an antibody that recognizes the two most abundant outer membrane proteins, P28/OMP-19 and OMP-1F/OMP-18. Both proteins were predicted to have structural features characteristic of porins, including 12 transmembrane segments comprised of amphipathic and antiparallel beta-strands. The sodium dodecyl sulfate stability of the two proteins was consistent with a beta-barrel structure. Isolated native P28 and OMP-1F exhibited porin activities, with pore sizes similar to and larger than, respectively, that of OprF, which is the porin with the largest pore size known to date. E. chaffeensis experiences temperature changes during transmission by ticks. During the intracellular development of E. chaffeensis, both P28 and OMP-1F were expressed mostly in the mid-exponential growth phase at 37 degrees C and the late-exponential growth phase at 28 degrees C. The porin activity of proteoliposomes reconstituted with proteins from the outer membrane fractions derived from bacteria in the mid- and late-exponential growth phases at 28 degrees C and 37 degrees C correlated with the expression levels of P28 and OMP-1F. These results imply that P28 and OMP-1F function as porins with large pore sizes, suggesting that the differential expression of these two proteins might regulate nutrient uptake during intracellular E. chaffeensis development at both temperatures.     

Identification of a new pore in the mitochondrial outer membrane of a porin-deficient yeast mutant

Reconstitution experiments were performed on lipid bilayer membranes in the presence of detergent-solubilized mitochondrial outer membranes of a porin-free yeast mutant and of its parent strain. The addition of the detergent-solubilized material resulted in a strong increase in the membrane conductance which was not observed if only the detergent was added to the aqueous phase. Surprisingly, the membrane conductance induced by the detergent extracts of the mutant membrane was only a factor of 20 less than that caused by the outer membrane of the parent strain under otherwise identical conditions. Single-channel recordings of lipid bilayer membranes in the presence of mitochondrial outer membranes of the yeast mutant suggested the presence of a transient pore. The reconstituted pores had a single-channel conductance of 0.21 nS in 0.1 M KCl and the characteristics of general diffusion pores with an estimated effective diameter of 1.2 nm. The pores present in the mitochondrial outer membranes of the yeast mutant shared some similarities with the pores formed by mitochondrial and bacterial porins although their effective diameter is much smaller than those of the 'normal' mitochondrial porins which have a single-channel conductance of about 0.4 nS in 0.1 M KCl, corresponding to an effective diameter of 1.7 nm. Zero-current membrane-potential measurements suggested that the second mitochondrial porin is slightly cation-selective. Its possible role in the metabolism of mitochondria is discussed.