Partial purification of a potassium channel with low permeability for sodium from tonoplast membranes of Hordeum vulgare cv. Gerbel

A potassium-specific tonoplast channel was identified by reconstitution of tonoplast polypeptides into planar lipid bilayer membranes. Highly purified tonoplast membranes were solubilized in Triton X-100-containing buffer and fractionated by size-exclusion chromatography. The protein fractions were assayed for ion channel activity in a planar bilayer system, and the potassium channel was routinely recovered in specific fractions corresponding to an apparent molecular mass of 80 kDa. In symmetrical electrolyte solutions of 100 mM potassium chloride, the potassium channel had a single-channel conductance of 72 pS. Substates of the channel with conductances of 17, 33 and 52 pS were frequently observed. After identification of the channel in low or high KCl, addition of sodium acetate or sodium chloride caused only insignificant conductance changes. This result suggested that the channel was not or little permeable for sodium or chloride, whereas it had similar single-channel conductance for rubidium and caesium ions as compared with potassium ions. The channel is presumably responsible for the equilibration of potassium between the vacuole and the cytosol. The role of the channel in the physiology of the barley cell under salt stress is discussed.

Reconstitution of vacuolar ion channels into planar lipid bilayers

Vacuolar ion channels were characterized after reconstitution into planar lipid bilayers. (1) Channel activity was observed after incorporation of tonoplast-enriched microsomal membranes, purified tonoplast membranes or of solubilized tonoplast proteins. (2) Channels of varying single-channel conductances were detected after reconstitution. In symmetrical 100 mmol l-1 KCl, conductances between 1 and 110 pS were frequently measured; the largest number of independent reconstitution events was seen for single-channel conductances of 16-25 pS (28 experiments), 30-42 pS (26), 49-56 pS (15) and 64-81 pS (15). Channel current usually increased linearly with voltage. (3) In asymmetrical solutions, cation-, non-selective and, for the first time for the tonoplast, anion-selective channels were detected. Ca(2+)-dependent regulation of channel opening was not observed in our reconstitution system. (4) Permeability was also observed for Cl-, NO3-, SO4(2-) and phosphate. (5) After fractionation of tonoplast proteins by size exclusion chromatography, ion channel activity was recovered in specific fractions. (6) Some of these fractions catalyzed sulfate transport after reconstitution into liposomes. The results suggest that different channels are active at the tonoplast membrane at a larger number than has been concluded from previous work.

Regulation of adenylyl cyclase from Paramecium by an intrinsic potassium conductance

Hyperpolarization of the cell membrane of Paramecium stimulates adenosine 3',5'-monophosphate (cAMP) formation. Manipulations of the K+ resting conductance of the ciliate by adaptation in different buffers affected excitability of the cAMP generating system. Blockade of K+ channels inhibited hyperpolarization-stimulated cAMP formation. A mutant of Paramecium that is unable to control its K+ resting conductance had a defect in cAMP formation. Purified adenylyl cyclase, when incorporated into an artificial lipid bilayer membrane, revealed properties of a voltage-independent K+ channel. This indicates that the adenylyl cyclase of Paramecium has a secondary function as carrier of the K+ resting conductance. A hyperpolarization-activated K+ efflux appears to directly regulate adenylyl cyclase activity in vivo.

Investigation of the selectivity of maltoporin channels using mutant LamB proteins: mutations changing the maltodextrin binding site

Wild-type and seven mutant maltoporins were purified and their channel-forming activities studied after reconstitution into black lipid membranes. The proteins were assayed for alterations at the maltodextrin binding site by measuring the sugar-dependent blockage of ion flux through these channels. Some substitutions (R8H, W74R) caused reduced channel affinity for all maltodextrins without changing single channel conductivities. The channel with a GlySer insertion after residue 9 was also poorly blocked by sugars but unique to this protein, the channel showed a striking, almost exponential increase of affinity with increasing maltodextrin chain length. In mutants with AspPro insertions after residues 79 and 183, there was an increase in affinity for glucose and maltose but not longer maltodextrins. The additional negative charge in the AspPro insertion mutants increased the cation selectivity of maltoporin channels, as did the decrease in positive charge resulting from the R8H substitution. A mutant with a W120C substitution also showed an increased affinity for glucose and maltose but reduced affinity for longer maltosaccharides. In contrast, a Y118F substitution resulted in an 8-fold increase in maltotriose affinity, but lesser improvements for other sugars. These results are interpreted to reflect changes in subsites contributing to an extended binding site within the channel, which in turn determines the overall sugar affinity of maltoporin.

Overexpression in Escherichia coli and functional analysis of a novel PPi-selective porin, oprO, from Pseudomonas aeruginosa

Immediately upstream from and adjacent to the oprP gene, which codes for the phosphate-specific porin OprP of Pseudomonas aeruginosa, lies the PR region (oprO), which cross-hybridizes with oprP DNA. To determine the function of this region, the oprO gene was expressed behind the lactose promoter in Escherichia coli, and the resultant OprO protein was purified and reconstituted into planar lipid bilayers. OprO formed sodium dodecyl sulfate-stable trimers, cross-reacted immunologically with OprP, and, like OprP, formed an anion-specific, phosphate-selective porin. However, it demonstrated lower affinity for and higher maximal conductance of both chloride and phosphate than did the OprP channel. Examination by macroscopic conductance inhibition experiments of the affinity of OprO for phosphates of different lengths revealed a preference for PPi and tripolyphosphate over Pi, suggesting that OprO functioned as a PPi-selective polyphosphate channel, in contrast to OprP, which has a marked preference for Pi.

The cation-selective substate of the mitochondrial outer membrane pore: single-channel conductance and influence on intermembrane and peripheral kinases

The polyanion-induced substate of the outer mitochondrial membrane was studied in vivo and in vitro. Study of the substate in artificial bilayers showed that it is highly cation selective. The induction of the substate in intact mitochondria leads to a complete inhibition of the intermembrane kinases, such as creatine kinase and adenylate kinase, which were excluded from the external ATP pool. Peripheral kinases, such as hexokinase, were blocked when they utilized internal ATP. The results with intact mitochondria suggested the existence of two regions of the outer membrane containing channels of different states, which may be involved in the regulation of intermembrane and peripheral kinases.

Role of frequency and amplitude of repetitive HCG stimulations for sustained progesterone secretion from the bovine corpus luteum in vitro

This investigation aimed at evaluating a role for frequencies and amplitudes of repeated HCG stimulations for the optimal maintenance of progesterone (P4) secretion from the bovine corpus luteum in vitro. Slices (100-120 mg) of midluteal bovine corpora lutea were perifused with medium M199 (0.05% BSA, pH 7.2, 38.5 degrees C) and the perifusion effluent collected at 15 minute intervals for 20-29 hours. Unstimulated P4 release (n = 5) was distinctly pulsatile (by Pulsar pulse algorithm), with pulses occurring every 90 +/- 6 minutes (mean +/- SEM) and pulse amplitudes of 14.4 +/- 1.1 ng. Conversely, no pulses were detected in two control perifusions. Unstimulated P4 release increased during the first 5 perifusion hours (from 39.3 +/- 4.6 to 50.3 +/- 5.6 ng/15 min, p less than 0.01), but then appeared to decline (to 29.3 +/- 1.3 ng/15 min, p less than 0.05) towards the end of the perifusion periods. Hourly pulses of HCG (6.7 mM) did not change the P4 pulse amplitudes (16.6 +/- 2.0 ng), the pulse periodicities (105 +/- 15 min) and overall release rates (34.7 +/- 5.7 ng/15 min), nor did they prevent the decline in P4 secretion towards the end of perifusions (n = 5). In contrast, 2-hourly HCG stimulations maintained stable P4 release rates throughout the perifusion periods (34.7 +/- 6.8 ng/15 min), with P4 pulses of similar amplitudes (14.7 +/- 1.7 ng), but of lower periodicities (135 +/- 2 min, p less than 0.05) than during unstimulated conditions (n = 5).(ABSTRACT TRUNCATED AT 250 WORDS)

Haemolysin-derived synthetic peptides with pore-forming and haemolytic activity

Escherichia coli haemolysin (Hlya) is a pore-forming protein which belongs to the family of 'Repeat-toxins' (RTX) (Lo et al., 1987; Lally et al., 1989; Kraig et al., 1990). A model for the pore-forming structure of HlyA has been proposed (Ludwig et al., 1991) which consists of eight transmembrane segments all present in this hydrophobic region of HlyA. We report here that two synthetic peptides of 10 and 8 amino acids in length (Pep1 and Pep2, respectively), which are derived from transmembrane segment V, are able to form pores in an artificial lipid bilayer. In addition, Pep1 exhibits strong haemolytic activity when tested on human red blood cells (HRBCs). The haemolytic activity of Pep1 and of E. coli haemolysin is completely inhibited by antibodies raised against Pep1.

Characterization of SH groups in porin of bovine heart mitochondria. Porin cysteines are localized in the channel walls

Porin from bovine heart mitochondria contains probably two cysteines (Cys126 and Cys230 in human porin, Kayser, H., Kratzin, H. D., Thinnes, F. P., Götz, H., Schmidt, W. E., Eckart, K. & Hilschmann, N. (1989) Biol. Chem. Hoppe-Seyler 370, 1265-1278). Reduced and oxidized forms of these cysteines were investigated in purified protein and in intact mitochondria using the agents dithioerythritol, cuprous(II) phenantroline, diamide and performic acid. Furthermore, intact mitochondria were labelled with the sulfhydryl-alkylating agents N-[14C]ethylmaleimide, eosin-5-maleimide and N-(1-pyrenyl)-maleimide. Affinity chromatography of bovine heart porin was performed with cysteine-specific material. The results can be summarized as follows: (1) Porin has one reduced and two oxidized forms of apparent molecular masses between 30 and 35 kDa. The native form of porin is the reduced 33 kDa form. The oxidized forms only appear after denaturation with SDS. (2) The 35-kDa reduced and the 33.5-kDa oxidized forms of porin show the same pore-forming properties after reconstitution of the protein into lipid bilayer membranes. (3) Labelling of cysteines by eosin-5-maleimide and N-(1-pyrenyl)-maleimide suggested their location at a boundary between the water-phase and the lipid-phase. Incubation of intact mitochondria with N-ethylmaleimide prior to eosin-5-maleimide and N-(1-pyrenyl)maleimide treatment resulted in the inhibition of the fluorescent labelling. Among the cysteines present in the primary structure, Cys126 is the most sensitive to N-ethylmaleimide binding. (4) Bovine heart mitochondrial porin covalently bound to Affi-Gel 501 (with a 1.75 nm long spacer), but not to Thiopropyl-Sepharose 6B (with a 0.51 nm spacer). This suggests that at least one of the cysteines is localized between 0.51 nm and 1.75 nm deep in the protein micelle.

How do androgens affect episodic gonadotrophin secretion in postmenopausal women?

In the absence of any significant ovarian oestrogen secretion, as in post-menopausal women, the hypothalamic-pituitary axis may still be influenced by the androgens which continue to be produced. The episodic secretion of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) by postmenopausal women was accordingly assessed following short-term androgen antagonism induced by flutamide, a specific androgen receptor blocker. Blood samples were collected at 10-min intervals for 10 h in nine women before and during flutamide administration (750 mg/day for 6 days) for the determination of gonadotrophin and sex hormone concentrations by radioimmunoassay. On both occasions, 25 micrograms of gonadotrophin-releasing-hormone (GnRH) was injected intravenously 8 h after initiation of the blood collections. Flutamide administration decreased (P less than 0.01 or less) androgen concentrations (testosterone, androstenedione and dehydroepiandrosterone sulphate) in relation to baseline values, but did not alter oestrogen (oestrone and oestradiol) or sex-hormone-binding globulin levels. The LH and FSH pulse characteristics (frequency, amplitude, interpulse interval and transverse mean levels) determined by a cluster algorithm in the gonadotrophin secretory profiles did not differ before and during androgen blockade. By contrast, androgen antagonism increased LH (P less than 0.01) and tended to enhance FSH (P = 0.10) FSH release in response to GnRH stimulation. Hence, short-term androgen receptor blockade with flutamide did not greatly affect episodic gonadotrophin secretion. However, the combined evidence of the enhanced gonadotrophin release observed in response to GnRH stimulation and the unchanged gonadotrophin secretion during androgen antagonism suggests that alterations in the magnitude, but not the frequency, of hypothalamic GnRH release had occurred. Even in the presence of substantial serum androgen concentrations, the gonadotrophin pulse rhythm in hypogonadal women constitutes the maximal-rate GnRH-LH release pattern.

Pituitary gonadotropin responsiveness to repeated gonadotropin releasing hormone (GnRH) stimulations in patients with chronic anovulation

To evaluate whether repeated gonadotropin releasing hormone (GnRH) stimulations were superior to single GnRH administrations for the accurate assessment of pituitary gonadotropin responsiveness, the GnRH-stimulated luteinizing hormone (LH) and follicle stimulating hormone (FSH) responses of 49 hyperandrogenic patients (HA) were compared with those of 20 hypogonadotropic patients (HH) and of 24 normally cycling women (N). Blood samples were obtained at frequent intervals during GnRH administrations (25 micrograms twice within 2 h). Unstimulated LH concentrations were higher (p less than 0.001) in HA than in N and HH women. However, basal FSH levels differed only in HA from HH women (p less than 0.001). Following either GnRH stimulation, increased (p less than 0.01) LH and FSH releases were noted in all N, HA and HH women. The GnRH-stimulated LH and FSH responses to either GnRH injections were highest (p less than 0.01) in HA and lowest (p less than 0.01 vs. N) in HH women. The net LH and FSH increases over unstimulated concentrations (delta LH or FSH) in response to either GnRH stimulation were highest (p less than 0.01 or less) in HA women. By contrast, no differences were determined in the delta LH and FSH levels between the first and second GnRH stimulations within each group. These observations document different unstimulated and stimulated gonadotropin concentrations in normal cycling and anovulatory women. Gonadotropin responses to single GnRH administrations differ for anovulatory patients. Since the gonadotropin responses to the second GnRH stimulation are comparable to those during the first GnRH injections, repeated GnRH stimulations may not help to distinguish the degree of pituitary responsiveness in ovulatory from anovulatory women.

Identification of two porins in Pelobacter venetianus fermenting high-molecular-mass polyethylene glycols

Porins were purified from cells of the anaerobic gram-negative bacterium Pelobacter venetianus grown with 20-kDa polyethylene glycol. After treatment of the cell envelope fraction with sodium dodecyl sulfate-containing solutions, the murein contained only two major peptidoglycan-associated proteins of 14 and 23 kDa. Both proteins were released from the peptidoglycan by the detergent Triton X-100. Genapol X-80 released only the 23-kDa protein. This protein was purified by chromatography on a hydroxyapatite column. It did not form sodium dodecyl sulfate-resistant oligomers. Reconstituted in lipid bilayer membranes, the 23-kDa protein formed cation-selective channels with a single-channel conductance of 230 pS in 1 M KCl. The channel is not a general-diffusion pore, since its conductance depends only moderately on the salt concentration. The channel conducted ammonium much better than potassium or rubidium ions, suggesting that it is probably involved in ammonium uptake. The outer membrane of P. venetianus contains a further, non-murein-associated pore with an unknown molecular mass. It is also cationically selective and has a single-channel conductance of 1.6 nS in 1 M KCl, which suggests that its effective diameter is similar to that of porins from enteric bacteria.

Purification of glucose-inducible outer membrane protein OprB of Pseudomonas putida and reconstitution of glucose-specific pores

A 43,000 molecular-weight, glucose-inducible, organic acid-repressible protein (OprB) was identified in the outer membrane of Pseudomonas putida. OprB was surface expressed in whole cells, had a high beta-sheet content, and was heat modifiable, as demonstrated by 125I-labeling, circular dichroism spectroscopy, and mobility on sodium dodecyl sulfate-polyacrylamide gel electrophoresis. OprB was extracted from outer membrane preparations by using 2% Lubrol PX with 10 mM EDTA and purified by DEAE-Sephacel ion exchange chromatography following ammonium sulfate precipitation. Reconstitution experiments with black lipid membranes showed that OprB formed small, cation-selective pores which bound glucose (KS = 110 mM) and other carbohydrates. However, the binding site of OprB appeared to be distinct from that of the maltodextrin-specific porin LamB from Escherichia coli.

The major outer membrane protein of Acidovorax delafieldii is an anion-selective porin

The major outer membrane protein (Omp34) of Acidovorax delafieldii (formerly Pseudomonas delafieldii) was purified to homogeneity and was characterized biochemically and functionally. The polypeptide has an apparent molecular weight (Mr) of 34,000, and it forms stable oligomers at pH 9.0 in the presence of 10% octylpolyoxyethylene or 2% lithium dodecyl sulfate below 70 degrees C. The intact protein has a characteristic secondary structure composition, as revealed by Fourier transforming infrared spectroscopy (about 60% beta sheet). These features and the amino acid composition are typical for porins. The purified Omp34 is associated with 1 to 2 mol of lipopolysaccharide per mol of the monomer. Pore-forming activity was demonstrated with lipid bilayer experiments. Single-channel and selectivity measurements showed that the protein forms highly anion-selective channels. The unusual dependence of the single-channel conductance on salt concentration suggests that the porin complexes bear positive surface charges, accumulating negatively charged counterions at the pore mouth.

Mutations affecting pore formation by haemolysin from Escherichia coli

By introduction of site-specific deletions, three regions in HlyA were identified, which appear to be involved in pore formation by Escherichia coli haemolysin. Deletion of amino acids 9-37 at the N-terminus led to a haemolysin which had an almost threefold higher specific activity than wild-type and formed pores in an artificial asolectin lipid bilayer with a much longer lifetime than those produced by wild-type haemolysin. The three hydrophobic regions (DI-DIII) located between amino acids 238-410 contributed to pore formation to different extents. Deletion of DI led to a mutant haemolysin which was only slightly active on erythrocyte membranes and increased conductivity of asolectin bilayers without forming defined pores. Deletions in the two other hydrophobic regions (DII and DIII) completely abolished the pore-forming activity of the mutant haemolysin. The only polar amino acid in DI, Asp, was shown to be essential for pore formation. Removal of this residue led to a haemolysin with a considerably reduced capacity to form pores, while replacement of Asp by Glu or Asn had little effect on pore formation. A deletion mutant which retained all three hydrophobic domains but had lost amino acids 498-830 was entirely inactive in pore formation, whereas a shorter deletion from amino acids 670-830 led to a mutant haemolysin which formed abnormal minipores. The conductivity of these pores was drastically reduced compared to pores introduced into an asolectin bilayer by wild-type haemolysin. Based on these data and structural predictions, a model for the pore-forming structure of E. coli haemolysin is proposed.

Autonomous progesterone secretion from the bovine corpus luteum in vitro

While the regulation of progesterone secretion from the corpus luteum by LH has been convincingly demonstrated, the secretory patterns in the absence of any pituitary LH inputs are yet unclear. Consequently, we investigated the progesterone secretion by an in vitro perifusion system to characterize spontaneous progesterone release from the isolated bovine corpus luteum. Slices (120 mg) of midluteal corpora lutea were placed in perifusion chambers and continuously perifused by Medium-199 for 160-320 min. Progesterone was determined by radioimmunoassay in the effluent fractions collected at 2-min intervals. The spontaneous progesterone release from all bovine corpora lutea was pulsatile. Pulses were observed at mean (+/- SEM) intervals of 17.7 +/- 1.5 min with amplitudes of 6.7 +/- 0.5 ng and release rates of 29.5 +/- 2.4 ng.ml-1.(2 min)-1 (N = 5). Addition of 6.7 nmol/l hCG to the perifusion medium appeared to increase the pulse amplitudes and release rates (195 +/- 25% over unstimulated conditions), but did not change the pulse frequencies (N = 3). Perifusions with calcium-free medium containing 50 mumol/l verapamil and 20 mmol/l EGTA tended to suppress the pulse frequencies and amplitudes of this spontaneous progesterone release, whereas addition of hCG reversed this decrease again (N = 3). When prostaglandin activity was inhibited by 100 mumol/l indomethacin added to the perifusion medium, both pulse frequencies and amplitudes of this progesterone release were enhanced (N = 4). During perifusions with 100 mumol/l of the anti-progesterone ZK 96.734, both the pulse frequencies and amplitudes increased (N = 4). These studies demonstrate an episodic progesterone release from the bovine corpus luteum perifused in vitro.(ABSTRACT TRUNCATED AT 250 WORDS)

Characterization of pore-forming activity in liver mitochondria from Anguilla anguilla. Two porins in mitochondria?

A fast purification procedure for the isolation and purification of eukaryotic porin (De Pinto et al., (1987) Biochim. Biophys. Acta 905, 499-502) was applied to liver mitochondria of the fish Anguilla anguilla. A protein preparation was obtained which formed slightly anionically selective pores in reconstitution experiments with lipid bilayer membranes. The distribution of single-channel conductances had two maxima of 2.4 nS and 4.0 nS in 1 M KCl. Sodium dodecylsulfate electrophoretograms of the protein preparation showed the presence of two bands of very similar electrophoretic mobility (32 and 32.5 kDa). Both bands cross-reacted with antibodies raised against purified bovine heart porin and with antibodies raised against the 19 amino acids N-terminal end of human porin. No cross-reactivity was observed with antibodies against yeast porin. The peptide maps of the two bands showed slight differences. The possibility of the presence of two different porins in liver mitochondria of Anguilla anguilla is discussed. An extensive immunological comparison of different mitochondrial porins is presented.

Positive residues involved in the voltage-gating of the mitochondrial porin-channel are localized in the external moiety of the pore

The role of positive charges located on the hydrophilic surface of the mitochondrial outer membrane channel was investigated by studying the interaction between LDAO-solubilized porin and a cation-exchanger column. The binding of porin to the column material was inhibited when the elution buffer had a pH of 9 or when 2 mM dextran sulfate was added to the buffer at neutral pH. Interestingly, the addition of a synthetic copolymer of methacrylate, maleate and styrene known as a potent modulator of the voltage-dependence, did not influence the interaction between column material and porin. Incubation of porin with fluorescein isothiocyanate (FITC) resulted in the isolation of a porin fraction in which on average two lysines located on the surface of the pore-forming complex per 35 kDa polypeptide were modified. The voltage-dependence of the fluorescein isothiocyanate modified porin was strongly decreased as compared with the unmodified porin. The experiments presented here give the first biochemical evidence that positively charged lysine residues located on the surface of the channel-forming complex are responsible for the gating of the mitochondrial porin-channel.

Studies on human porin. III. Does the voltage-dependent anion channel "Porin 31HL" form part of the chloride channel complex, which is observed in different cells and thought to be affected in cystic fibrosis?

"Porin 31HL", of known primary structure, is an integral protein of the plasmalemma of human B cells (Thinnes, F.P. et al. (1989) This Journal 370, 1253-1264; Kayser, H. et al. (1989) This Journal 370, 1265-1278). Purified "Porin 31HL" from human B lymphocytes was reconstituted into lipid bilayer membranes, where it formed defined voltage-dependent channels. Five minutes preincubation with 100 microM 4,4'-diisothiocyanatostilbene-2,2'-disulfonate, potent inhibitor of chloride transport, altered the channel-forming properties of the protein, so that it now showed small irregular channels instead of distinct steps. In addition, the voltage-dependence of the channel was abolished by the action of 4,4'-diisothiocyanatostilbene-2,2'-disulfonate. Functional and structural similarities between "Porin 31HL" and porin preparations from other human tissues and from other species suggest that this protein may be part of the chloride channel complex, which is defective in cystic fibrosis.

Pore formation associated with the tail-tip protein pb2 of bacteriophage T5

Upon binding of bacteriophage T5 tails to purified FhuA receptor protein the tail-tip protein pb2 became extremely sensitive to trypsin and other proteases. However, when T5 tails were bound to FhuA integrated into liposomes, pb2 was found to retain some resistance to trypsin. Electron microscopic examination of tail-liposome complexes supported the idea that trypsin resistance of pb2 in such complexes was caused by insertion of the tail-tip into the liposomes. pb2 was isolated from tails by treatment with sodium dodecyl sulfate and was further purified by gel filtration using a fast protein liquid chromatography system. pb2 obtained with this procedure was most likely monomeric. It was extremely sensitive to trypsin. When reconstituted into black lipid bilayer membranes, it formed pores with an average single-channel conductance of 4.6 nanosiemens in 1 M KCl. Zero-current potential measurements showed only a very slight preference, if any, for cations over anions. The data are compatible with pb2 forming a large water-filled transmembrane channel. The functioning during infection of pb2 in cytoplasmic membrane depolarization and phage DNA uptake into the cell is discussed.

Aerolysin of Aeromonas sobria: evidence for formation of ion-permeable channels and comparison with alpha-toxin of Staphylococcus aureus

Aerolysin from Aeromonas sobria AB3 was isolated and purified. The pure toxin formed sodium dodecyl sulfate-insoluble oligomers in a lipidic environment. The addition of aerolysin to the aqueous phase bathing lipid bilayer membranes resulted in the formation of ion-permeable channels which had a single-channel conductance of about 70 pS in 0.1 M KCl. This defines the toxin as a channel-forming component similar to other toxins but without any indication for an association-dissociation reaction, since the channels had a long lifetime at low voltages. At voltages higher than 50 mV, the aerolysin channel switched into a closed state with a low residual conductance. The single-channel conductance was a linear function of the total aqueous conductance, which suggested that the toxin oligomers formed aqueous channels with an estimated minimal diameter of about 0.7 nm. The aerolysin pores were found to be slightly anion selective. The pore-forming properties of aerolysin were compared with those of alpha-toxin of Staphylococcus aureus. Both aerolysin and alpha-toxin share secondary structure features, must oligomerize to form pores in lipid bilayer membranes, and form channels with similar properties.

Porin of Rhodobacter capsulatus: biochemical and functional characterization

The major outer membrane protein of Rhodobacter capsulatus 37b4 (capsule-free) was isolated. Strong porin-activity was observed after reconstitution into artificial lipid bilayer membranes with a single channel conductance of 3.15 nS in 1 M KCl. The porin migrated as a broad, single band (Mr above 90,000) on sodium dodecyl sulfate polyacrylamide gel electrophoresis and dissociated into a single species of polypeptides (Mr 36,000) on treatment with EDTA (10 mM at 30 degrees C, 20 min) or by heating (100 degrees C, 5 min). Analytical ultracentrifugation studies demonstrated the native porin to be a trimer. The monomers chromatofocused as a single, sharp peak on fast performance liquid chromatography and only one band, corresponding to an isoelectric point of about 4.0, was obtained on isoelectric focusing. Gas-phase sequencing of the 23 N-terminal residues revealed Glu-Val-Lys-Leu-Ser-Gly-Asp-Ala-Arg-Met-Gly-Val-Met-Tyr-Asn-Gly-Asp-Asp- X-Asn- Phe-Ser-Ser.

Isolation of Neisseria meningitidis mutants deficient in class 1 (porA) and class 3 (porB) outer membrane proteins

The class 1 major outer membrane protein of Neisseria meningitidis is a serious candidate for a meningococcal vaccine. To facilitate studies on the function of this protein, mutants were isolated that lacked this protein or the structurally related class 3 protein. These mutants were obtained by using the antibody-dependent bactericidal action of the complement system. The class 1 protein-deficient strain grew normally in vitro, whereas growth of the class 3 protein-deficient strain was slightly retarded. The class 3 protein-deficient strain displayed increased resistance to the antibiotics tetracycline and cefsulodin, which is consistent with the proposed role of the protein as a pore-forming protein. The class 1 protein was purified to homogeneity from the class 3 protein-deficient strain. Lipid bilayer experiments revealed that this protein also formed pores. The class 1 protein pores were cation selective.

LamB (maltoporin) of Salmonella typhimurium: isolation, purification and comparison of sugar binding with LamB of Escherichia coli

LamB (maltoporin) of Salmonella typhimurium was found to be more strongly associated with the murein than OmpF. It was purified in one step using a hydroxyapatite (HTP) column. Reconstitution of the pure protein with lipid bilayer membrane showed that LamB of S. typhimurium formed small ion-permeable channels with a single channel conductance of about 90 pS in 1 M KCl and some preference for cations over anions. The conductance concentration curve was linear, which suggested that LamB of S. typhimurium does not contain any binding site for ions. Pore conductance was completely inhibited by the addition of 20 mM maltotriose. Titration of the LamB-induced membrane conductance with different sugars, including all members of the maltooligosaccharide series up to seven glucose residues, suggested that the channel contains, like LamB (maltoporin) of Escherichia coli, a binding site for sugars. The binding constant of sugars of the maltooligosaccharide series increased with increasing number of glucose residues up to five (saturated). Small sugars had a higher stability constant for sugar binding relative to LamB of E. coli. The advantage of a binding site inside a specific porin for the permeation of solutes is discussed with respect to the properties of a general diffusion porin.