Intercellular connexin channels (gap junction channels) have long been thought to mediate molecular signaling between cells, but the nature of the signaling has been unclear. This study shows that connexin channels from native tissue have selective permeabilities, partially based on pore diameter, that discriminate among cytoplasmic second messenger molecules. Permeability was assessed by measurement of selective loss/retention of tracers from liposomes containing reconstituted connexin channels. The tracers employed were tritiated cyclic nucleotides and a series of oligomaltosaccharides derivatized with a small uncharged fluorescent moiety. The data define different size cut-off limits for permeability through homomeric connexin-32 channels and through heteromeric connexin-32/connexin-26 channels. Connexin-26 contributes to a narrowed pore. Both cAMP and cGMP were permeable through the homomeric connexin-32 channels. cAMP was permeable through only a fraction of the heteromeric channels. Surprisingly, cGMP was permeable through a substantially greater fraction of the heteromeric channels than was cAMP. The data suggest that isoform stoichiometry and/or arrangement within a connexin channel determines whether cyclic nucleotides can permeate, and which ones. This is the first evidence for connexin-specific selectivity among biological signaling molecules.

The peptide antibiotic nisin is shown to disrupt valinomycin-induced potassium diffusion potentials imposed on intact cells of Staphylococcus cohnii 22. Membrane depolarization occurred rapidly at high diffusion potentials while at low potentials nisin-induced depolarization was slower suggesting that nisin requires a membrane potential for activity. This assumption was proven in experiments with planar lipid bilayers (black lipid membranes). Macroscopic conductivity measurements indicated a voltage-dependent action of nisin. The potential must have a trans-negative orientation with respect to the addition of nisin (added to the cis-side) and a sufficient magnitude (ca.-100 mV). With intact cells the threshold potential was lower (-50 to -80 mV at pH 7.5 and below -50 mV at pH 5.5). Single channel recordings resolved transient multi-state pores, strongly resembling those introduced by melittin into artificial bilayers. The pores had diameters in the range of 0.2-1 nm, and lifetimes of few to several hundred milliseconds. The results indicate that nisin has to be regarded as a membrane-depolarizing agent which acts in a voltage-dependent fashion.

A procedure for the purification of a very hydrophobic lipase from Pseudomonas sp. strain ATCC 21808 was elaborated by avoiding the use of long-chain detergents in view of subsequent crystallization of the enzyme. The purification procedure included chromatography on Q-Sepharose in the presence of n-octyl-o-D-glucopyranoside, Ca2" precipitation of fatty acids, and Octyl-Sepharose chromatography. The enzyme was purified 260-fold to a yield of 35% and a specific activity of 3,300 U/mg. The molecular weight was determined as 35,000; a polyacrylamide gel under nondenaturing conditions revealed a band at 110,000, and the isoelectric point proved to be at 4.5 to 4.6. The lipase crystallized with different salts and ethylene glycol polymers in the presence of n-octyl-o-D-glucopyranoside and one alkyloligooxyethylene compound (C.EY) in the range from C5E2 to C8E4. The crystals diffract to a resolution of about 0.25 nm. Precession photographs revealed that they belong to space group C2 with lattice constants of a = 9.27 nm, b = 4.74 nm, c = 8.65 nm, and f = 122.30, indicating a cell content of one molecule per asymmetric unit of the crystal. In hydrolysis of triglycerides, the lipase showed substrate specificity for saturated fatty acids from C6 to C12 and unsaturated long-chain fatty acids. Monoglycerides were hydrolyzed very slowly. The N-terminal sequence is identical to that of the lipase from Pseudomonas cepacia. Treatment with diethyl-p-nitrophenylphosphate affected the activities toward triolein and p-nitrophenylacetate to the same extent and with the same velocity.Lipases are surface-active enzymes because binding to emulsified triglyceride substrates markedly increases their hydrolytic activity compared with their activity toward dissolved substrates (24). By biochemical (24) and threedimensional structural studies (5, 26), they were shown to act as serine hydrolases with a serine-histidine-aspartate triad at the active site, like serine proteases. The sequence around the reactive serine residues in the lipases from human pancreas (26) and Mucor miehei (5) is conserved. Homologous sequences have been found in all of the lipases sequenced (1, 6). However, the mechanism of activation at interfaces and the way the bulky substrate gains access to the buried active site (5, 26) is not understood. The threedimensional structures of other lipases may provide further insight into these questions. Several Pseudomonas species have been shown to produce lipases (16,22,23). Some of them have been sequenced, i.e., Pseudomonas fragi (2, 15), P. cepacia (19), and P. aeruginosa (28), and until 1990, crystallization of only one Pseudomonas lipase has been described (23). The enzyme of Pseudomonas sp. strain ATCC 21808 is an attractive lipase. It is stable at high temperature (50 to 60°C) and over a broad pH range (pH 5 to 10), according to Kobayashi, who has also described its activity on natural oils and fats, its intestinal absorbability, and its effect on hyperlipemia (13). By using the purification procedure he described, he...

The cationic peptide antibiotics Pep 5, nisin and subtilin depolarize bacterial and artificial membranes by formation of voltage-dependent multi-state pores. Studies with non-energized liposomes indicated that the peptides do not span the membrane in the absence of a membrane potential. The effects of Pep 5 and nisin on neutral membranes, as studied by membrane fluidity, phase transition points and carboxyfluorescein e&x, were small compared to melittin. Acidic liposomes were affected more strongly, indicative of primarily electrostatic interactions with phospholipid head groups. Subtilin may slightly enter the hydrophobic core as suggested by tryptophan fluorescence quenching and liposome fusion experiments.

The cationic bactericidal peptides Pep 5 and nisin render membranes permeable to low‐Mr compounds. All Gram‐positive bacteria treated with these peptides showed an immediate efflux of entrapped radioactive markers. The uptake of α‐[14C]methylglucoside by the phosphoenolpyruvate‐dependent phosphotransferase system was stimulated by Pep 5, supporting previous results that pep 5 abolishes the membrane potential. Oxygen consumption was inhibited, presumably due to lack of ADP. Escherichia coli became sensitive to Pep 5 and nisin when the outer membrane was bypassed by osmotic shock or by formation of cytoplasmic membrane vesicles. In contrast, Mycoplasma cells and erythrocytes were unaffected by Pep 5 and nisin in concentrations up to 1 mM. Human lung fibroblasts released only small amounts of ATP when treated with Pep 5 and nisin in μM concentrations. Eukaryotic and Mycoplasma cells were disrupted more effectively by the bee venom peptide melittin, which displays overall structural similarities to Pep 5 and nisin. Various artificial membranes were not affected by Pep 5, nisin, or melittin.