Solvent determined conformation of gramicidin affects the ability of the peptide to induce hexagonal HII phase formation in dioleoylphosphatidylcholine model membranes

It is shown by 31P-NMR and small angle X-ray scattering that induction of an hexagonal HII phase in dioleoylphosphatidylcholine model membranes by external addition of gramicidin A' depends on the solvent which is used to solubilize the peptide. Addition of gramicidin from dimethylsulfoxide or trifluoroethanol solution leads to HII phase formation whereas addition of the peptide from ethanol does not. This solvent dependence is shown by circular dichroism to be correlated with the peptide conformation. The channel conformation appears to be responsible for HII phase formation by gramicidin.     

Phase separation and hexagonal HII phase formation by gramicidins A, B and C in dioleoylphosphatidylcholine model membranes. A study on the role of the tryptophan residues

The role of the tryptophan-residues in gramicidin-induced HII phase formation was investigated in dioleoylphosphatidylcholine (DOPC) model membranes. 31P-NMR and small angle X-ray diffraction measurements showed, that gramicidin A and C (in which tryptophan-11 is replaced by tyrosine) induce a similar extent of HII phase formation, whereas for gramicidin B and synthetic analogs in which one tryptophan, either at position 9 or 11 is replaced by phenylalanine, a dramatic decrease of the HII phase inducing activity can be observed. Modification of all four tryptophans by means of formylation of the indole NH group leads to a complete block of HII phase formation. Sucrose density centrifugation experiments on the various peptide/lipid samples showed a quantitative incorporation of the peptide into the lipid. For all samples in a 1/10 molar ratio of peptide to lipid distinct bands were found, indicative of a phase separation. For the gramicidin A'/DOPC mixture these bands were analyzed and the macroscopic organization was determined by 31P-NMR and small-angle X-ray diffraction. The results demonstrate that a quantitative phase separation had occurred between a lamellar phase with a gramicidin/lipid ratio of 1/15 and a hexagonal HII phase, which is highly enriched in gramicidin. A study on the hydration properties of tryptophan-N-formylated gramicidin in mixtures with DOPC showed that this analog has a similar dehydrating effect on the lipid headgroup as the unmodified gramicidin. In addition both the hydration study and sucrose density centrifugation experiments showed that, like gramicidin also its analogs have a tendency to aggregate, but with differences in aggregation behaviour which seemed related to their HII phase inducing activity. It is proposed that the main driving force for HII phase formation is the tendency of gramicidin molecules to self-associate and organize into tubular structures such as found in the HII phase and that whether gramicidin (analogs) form these or other types of aggregates depends on their tertiary structure, which is determined by intra- as well as intermolecular aromatic-aromatic stacking interactions.     

Gramicidin-induced hexagonal HII phase formation in negatively charged phospholipids and the effect of N- and C-terminal modification of gramicidin on its interaction with zwitterionic phospholipids

The effect of gramicidin on macroscopic structure of the negatively charged membrane phospholipids cardiolipin, dioleoylphosphatidylglycerol and dioleoylphosphatidylserine in aqueous dispersions was investigated and compared with the effect of gramicidin on dioleoylphosphatidylcholine. It was shown by small-angle X-ray diffraction, 31P nuclear magnetic resonance and freeze-fracture electron microscopy that in all these lipid systems gramicidin is able to induce the formation of a hexagonal HII phase. 31P-NMR measurements indicated that the extent of HII phase formation in the various lipids ranged from about 40% to 60% upon gramicidin incorporation in a molar ratio of peptide to lipid of 1 : 10. Next, the following charged analogues of gramicidin were prepared: desformylgramicidin, N-succinylgramicidin and O-succinylgramicidin. The synthesis was verified with 13C-NMR and the effect of these analogues on lipid structure was investigated. It was shown that, as with gramicidin itself, the analogues induce HII phase formation in dioleoylphosphatidylcholine, lower and broaden the bilayer-to-HII phase transition in dielaidoylphosphatidylethanolamine and form lamellar structures upon codispersion with palmitoyllysophosphatidylcholine. Differential scanning calorimetry measurements indicated that, again like gramicidin, in phosphatidylethanolamine the energy content of the gel-to-liquid-crystalline phase transition is not affected by incorporation of the analogues, whereas in phosphatidylcholine a reduction of the transition enthalpy is found. These observations were explained in terms of a similar tendency to self-associate for gramicidin and its charged analogues. The results are discussed in the light of the various factors which have been suggested to be of importance for the modulation of lipid structure by gramicidin.     

Dynamic properties of gramicidin A in phospholipid membranes

The flexibility of the tryptophan side chains of gramicidin A and the rotational diffusion of the peptide in methanolic solution and in three membrane systems were studied with deuterium nuclear magnetic resonance (NMR). Gramicidin A was selectively deuterated at the aromatic ring systems of its four tryptophan side chains. In methanolic solution, the tryptophan residues remained immobile and served as a probe for the overall rotation of the peptide. The experimentally determined rotational correlation time of tau c = 0.6 X 10(-9) s was consistent with the formation of gramicidin A dimers. For gramicidin A incorporated into bilayer membranes, quite different results were obtained depending on the chemical and physical nature of the lipids employed. When mixed with 1-palmitoyl-sn-glycero-3-phosphocholine (LPPC) at a stoichiometric lipid:peptide ratio of 4:1, gramicidin A induced the formation of stable bilayer membranes in which the lipids were highly fluid. In contrast, the gramicidin A molecules of this membrane remained completely static over a large temperature interval, suggesting strong protein-protein interactions. The peptide molecules appeared to form a rigid two-dimensional lattice in which the interstitial spaces were filled with fluidlike lipids. When gramicidin A was incorporated into bilayers of 1,2-dioleoyl-sn-glycero-3-phosphocholine or 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) above the lipid phase transition, the deuterium NMR spectra were motionally narrowed, indicating large-amplitude rotational fluctuations. From the measurement of the quadrupole echo relaxation time, a rotational correlation time of 2 X 10(-7) s was estimated, leading to a membrane viscosity of 1-2 P if the rotational unit was assumed to be a gramicidin A dimer. (ABSTRACT TRUNCATED AT 250 WORDS)     

Folding and membrane insertion of the pore-forming peptide gramicidin occur as a concerted process

Many antibiotic peptides function by binding and inserting into membranes. Understanding this process provides an insight into the fundamentals of both membrane protein folding and antibiotic peptide function. For the first time, in this work, flow-aligned linear dichroism (LD) is used to study the folding of the antibiotic peptide gramicidin. LD provides insight into the combined processes of peptide folding and insertion and has the advantage over other similar techniques of being insensitive to off-membrane aggregation events. By combining LD data with conventional measurements of protein fluorescence and circular dichroism, the mechanism of gramicidin insertion is elucidated. The mechanism consists of five separately assignable steps that include formation of a water-insoluble gramicidin aggregate, dissociation from the aggregate, partitioning of peptide to the membrane surface, oligomerisation on the surface and concerted insertion and folding of the peptide to the double-helical form of gramicidin. Measurement of the rates of each step shows that although changes in the fluorescence signal cease 10 s after the initiation of the process, the insertion of the peptide into the membrane is actually not complete for a further 60 min. This last membrane insertion phase is only apparent by measurement of LD and circular dichroism signal changes. In summary, this study demonstrates the importance of multi-technique approaches, including LD, in studies of membrane protein folding.     

Conformation of gramicidin in relation to its ability to form bilayers with lysophosphatidylcholine

The ability of gramicidin to induce bilayer formation in lysophosphatidylcholine (LPC) systems was investigated as a function of the conformation of the peptide. The conformation was varied by using different solvents to cosolubilize gramicidin and lipid. Using circular dichroism (CD), it was found that when codissolved in trifluoroethanol (TFE), after drying and subsequent hydration, gramicidin is mainly present in the single-stranded beta 6.3-helical configuration, whereas when using chloroform/methanol or ethanol as the solvent, it is proposed that the dominant conformation of gramicidin in the membrane is that of the double-stranded antiparallel dimer. Employing 31P NMR, the stoichiometry for bilayer formation was found to be 6 to 7 lipid molecules per gramicidin monomer, when samples were prepared from TFE, whereas a stoichiometry of 4 was found when chloroform/methanol or ethanol was the solvent. Upon heating the latter samples, a conversion was observed in the CD pattern toward that indicative of the beta 6.3-helical configuration. This change was accompanied by an increase in the extent of bilayer formation. Next, it was investigated whether the conformation of gramicidin and its ability to induce bilayer formation were dependent on the lipid acyl chain length. CD measurements of samples prepared from TFE indicated that gramicidin, independent of acyl chain length, was present in the beta 6.3-helical configuration but the intensity of the ellipticities at 218 nm increased with the length of the acyl chain. The extent of bilayer formation in these samples was found to be largely chain length independent.(ABSTRACT TRUNCATED AT 250 WORDS)     

The membrane as an environment of minimal interconversion. A circular dichroism study on the solvent dependence of the conformational behavior of gramicidin in diacylphosphatidylcholine model membranes

The conformation of gramicidin in diacylphosphatidylcholine model membranes was investigated as a function of the solvent in which peptide and lipid are initially codissolved. By use of circular dichroism it is demonstrated that, upon removal of the solvent and hydration of the mixed gramicidin/lipid film, it is the conformational behavior of the peptide in the organic solvent that determines its final conformation in dimyristoylphosphatidylcholine model membranes. As a consequence, parameters that influence the conformation of the peptide in the solvent also play an essential role, such as the gramicidin concentration and the rate of interconversion between different conformations. Of the various solvents investigated, only with trifluoroethanol is it possible directly to incorporate gramicidin entirely in the beta 6.3-helical (channel) configuration. It is also shown that the conformation of gramicidin in the membrane varies with the peptide/lipid ratio, most likely as a result of intermolecular gramicidin-gramicidin interactions at higher peptide/lipid ratios, and that heat incubation leads to a conformational change in the direction of the beta 6.3-helical conformation. Using lipids with an acyl chain length varying from 12 carbon atoms in dilauroylphosphatidylcholine to 22 carbon atoms in dierucoylphosphatidylcholine, it was possible to investigate the acyl chain length dependence of the gramicidin conformation in model membranes prepared from these lipids with the use of different solvent systems. It is demonstrated for each solvent system that the distribution between different conformations is relatively independent of the acyl chain length but that the rate at which the conformation converts toward the beta 6.3-helical configuration upon heating of the samples is affected by the length of the acyl chain.(ABSTRACT TRUNCATED AT 250 WORDS)     

Orientation and lipid-peptide interactions of gramicidin A in lipid membranes: polarized attenuated total reflection infrared spectroscopy and spin-label electron spin resonance

Gramicidin A was incorporated at a peptide/lipid ratio of 1:10 mol/mol in aligned bilayers of dimyristoyl phosphatidylcholine (DMPC), phosphatidylserine (DMPS), phosphatidylglycerol (DMPG), and phosphatidylethanolamine (DMPE), from trifluoroethanol. Orientations of the peptide and lipid chains were determined by polarized attenuated total reflection infrared spectroscopy. Lipid-peptide interactions with gramicidin A in DMPC bilayers were studied with different spin-labeled lipid species by using electron spin resonance spectroscopy. In DMPC membranes, the orientation of the lipid chains is comparable to that in the absence of peptide, in both gel and fluid phases. In gel-phase DMPC, the effective tilt of the peptide exceeds that of the lipid chains, but in the fluid phase both are similar. For gramicidin A in DMPS, DMPG, and DMPE, the degree of orientation of the peptide and lipid chains is less than in DMPC. In the fluid phase of DMPS, DMPG, and DMPE, gramicidin A is also less well oriented than are the lipid chains. In DMPE especially, gramicidin A is largely disordered. In DMPC membranes, three to four lipids per monomer experience direct motional restriction on interaction with gramicidin A. This is approximately half the number of lipids expected to contact the intramembranous perimeter of the gramicidin A monomer. A selectivity for certain negatively charged lipids is found in the interaction with gramicidin A in DMPC. These results are discussed in terms of the integration of gramicidin A channels in lipid bilayers, and of the interactions of lipids with integral membrane proteins.     

A mismatch between the length of gramicidin and the lipid acyl chains is a prerequisite for HII phase formation in phosphatidylcholine model membranes

Previously it was shown that gramicidin can induce HII phase formation in diacylphosphatidylcholine model membranes only when the lipid acyl chain length exceeds 16 carbon atoms (Van Echteld, C.J.A., De Kruijff, B., Verkleij, A.J., Leunissen-Bijvelt, J. and De Gier, J. (1982) Biochim. Biophys. Acta 692, 126-138). Using 31P-NMR and small angle X-ray diffraction we now demonstrate that upon increasing the length of gramicidin, the peptide loses its ability to induce HII phase formation in di-C18:1c-PC but not in the longer chained di-C22:1c-PC. It is concluded that a mismatch in length between gramicidin and the lipid acyl chains, when the latter would provide excess bilayer thickness, is a prerequisite for HII phase formation in phosphatidylcholine model membranes.     

Gramicidin A induced fusion of large unilamellar dioleoylphosphatidylcholine vesicles and its relation to the induction of type II nonbilayer structures.

The fusogenic properties of gramicidin were investigated by using large unilamellar dioleoylphosphatidylcholine vesicles. It is shown that gramicidin induces aggregation and fusion of these vesicles at peptide to lipid molar ratios exceeding 1/100. Both intervesicle lipid mixing and mixing of aqueous contents were demonstrated. Furthermore, increased static and dynamic light scattering and a broadening of 31P NMR signals occurred concomitant with lipid mixing. Freeze-fracture electron microscopy revealed a moderate vesicle size increase. Lipid mixing is paralleled by changes in membrane permeability: small solutes like carboxyfluorescein and smaller dextrans, FD-4(Mr approximately 4000), rapidly (1-2 min) leak out of the vesicles. However, larger molecules like FD-10 and FD-17 (Mr approximately 9400 and 17,200) are retained in the vesicles for greater than 10 min after addition of gramicidin, thereby making detection of contents mixing during lipid mixing possible. At low lipid concentrations (5 microM), lipid mixing and leakage are time resolved: leakage of CF shows a lag phase of 1-3 min, whereas lipid mixing is immediate and almost reaches completion during this lag phase. It is therefore concluded that leakage, just as contents mixing, occurs subsequent to aggregation and lipid mixing. Although addition of gramicidin at a peptide/lipid molar ratio exceeding 1/50 eventually leads to hexagonal HII phase formation and a loss of vesicle contents, it is concluded that leakage during fusion (1-2 min) is not the result of HII phase formation but is due to local changes in lipid structure caused by precursors of this phase. By making use of gramicidin derivatives and different solvent conformations, it is shown that there is a close parallel between the ability of the peptide to induce the HII phase and its ability to induce intervesicle lipid mixing and leakage. It is suggested that gramicidin-induced fusion and HII phase formation share common intermediates.     

Gramicidin channel kinetics under tension.

We have measured the effect of tension on dimerization kinetics of the channel-forming peptide gramicidin A. By aspirating large unilamellar vesicles into a micropipette electrode, we are able to simultaneously monitor membrane tension and electrical activity. We find that the dimer formation rate increases by a factor of 5 as tension ranges from 0 to 4 dyn/cm. The dimer lifetime also increases with tension. This behavior is well described by a phenomenological model of membrane elasticity in which tension modulates the mismatch in thickness between the gramicidin dimer and membrane.     

Influence of proteins on the reorganization of phospholipid bilayers into large domains.

Using large (5-10 microns) vesicles formed in the presence of phospholipids fluorescently labeled on the acyl chain and visualized using a fluorescence microscope, charge-coupled-device camera, and digital image processor, we examined the effects of membrane proteins on phospholipid domain formation. In vesicles composed of phosphatidic acid and phosphatidylcholine, incubation with cytochrome c induced the reorganization of phospholipids into large phosphatidic acid-enriched domains with the exclusion of phosphatidylcholine. Cytochrome c binding was demonstrated to be highest in the phosphatidic acid-enriched domain of the vesicle using the absorbance of the heme moiety for visualization. Both binding of cytochrome c and phospholipid reorganization were blocked by pretreatment of the vesicles with 0.1 M NaCl. The pore forming peptide gramicidin was examined for the effects of an integral protein on domain formation. Initially, gramicidin distributed randomly within the vesicle and showed no phospholipid specificity. Phosphatidic acid domain formation in the presence of 2.0 mM CaCl2 or 100 microM cytochrome c was not affected by the presence of 5 mol % gramicidin within the vesicles. In both cases, gramicidin was preferentially excluded from the phosphatidic acid-enriched domain and became associated with phosphatidylcholine-enriched areas of the vesicle. Thus, cytochrome c caused a major reorganization of both the phospholipids and the proteins in the bilayer.     

Enzyme-bound formylvaline and formylvalylglycine; an initiation complex for gramicidin A biosynthesis.

Gramicidin A is an antibiotic peptide produced by Bacillus brevis ATCC 8185, which also produces tyrocidines. An attempt was made to establish a cell-free enzyme system for gramicidin A synthesis. An enzyme fraction, Component I, was partially purified from crude extracts of the organism and proven to be involved in the synthesis of the formyl-Val-Gly- region of gramicidin A. The initiation of gramicidin A biosynthesis is a function of Component I, which activates valine and binds it as a thioester, and further formylates it in the presence of formyltetrahydrofolic acid. The formylvaline thus synthesized is transferred to the glycine moiety, which is also thioesterified to Component I. Elongation of the peptide chain takes place by a mechanism similar to those found for tyrocidines, gramicidin S, and bacitracin.     

Structure-activity relationship of gramicidin S analogues on membrane permeability

The previous study of the action of gramicidin S on bacteria (Katsu, T., Kobayashi, H. and Fujita, Y. (1986) Biochim. Biophys. Acta 860, 608-619) prompted us to investigate further the structure-activity relationship of the gramicidin S analogues on membrane permeability. Two types of the gramicidin S analogues were used in the present study: (1) cyclo(-X-D-Leu-D-Lys-D-Leu-L-Pro-)2, where X = Gly, D-Leu and D-cyclohexylalanine (D-cHxAla); (2) N,N'-diacetyl derivative of gramicidin S (diacetyl-gramicidin S) which lacks a cationic moiety of gramicidin S. All the analogues have a beta-sheet conformation as gramicidin S. The following cellular systems were used: Staphylococcus aureus as Gram-positive bacteria, Escherichia coli as Gram-negative bacteria, human erythrocytes, rat liver mitochondria and artificial liposomal membranes. It was found that gramicidin S and one of the type 1 analogues having X = D-cHxAla induced the efflux of K+ through the cytoplasmic membrane of all types of the cells. In addition, these two peptides had the ability to lower the phase transition temperature of dipalmitoylphosphatidylcholine. Accordingly, it was concluded that, if peptides can expand greatly the membrane structure of neutral lipids which constitute main parts of the biological membrane, they can stimulate the permeability of cells without any selectivity. The action of the type 2 peptide, diacetyl-gramicidin S, was strongly cell dependent. Although this peptide stimulated the efflux of K+ from mitochondria, it did not do so efficiently, if at all, from S. aureus, E. coli and erythrocytes. In experiments using liposomes, diacetyl-gramicidin S increased markedly the permeability of liposomes composed of egg phosphatidylcholine. The presence of egg phosphatidylethanolamine or cholesterol reduced its activity. These results on liposomes explained well the low sensitivity of diacetyl-gramicidin S against E. coli and erythrocytes in terms of lipid constituents of the membranes. The mechanism of action of diacetyl-gramicidin S was discussed from the formation of a boundary lipid induced by this peptide.     

Absence of pantothenic acid in gramicidin S synthetase 2 obtained from some mutants of Bacillus brevis

The pantothenic acid content of gramicidin S synthetase 2(GS 2) was estimated microbiologically with enzymes obtained from the wild strain and gramicidin S-lacking mutant strains of Bacillus brevis. Four mutant enzymes from BI-4, C-3, E-1, and E-2 lacked pantothenic acid. Other mutant enzymes from BII-3, BI-3, BI-9, and BI-2 contained the same amount of pantothenic acid as the wild-type enzyme. Pantothenic acid-lacking GS 2 belonged to group V of mutant enzymes, which could activate all amino acids related to gramicidin S; their complementary enzyme, gramicidin S synthetase 1(GS 1), lacked racemizing activity. To ascertain whether 4'-phosphopantetheine is involved in the formation of D-phenylalanyl-L-prolyl diketopiperazine (DKP) and gramicidin S, combinations were tested of intact GS 1 from the wild strain with various mutant GS 2 either containing or lacking pantothenic acid. Only the combinations of wild-type GS 1 with mutant GS 2 containing pantothenic acid could synthesize DKP. Combinations with pantothenic acid-lacking GS 2 also failed to elongate peptide chains. Pantothenic acid-lacking GS 2 could bind the four amino acids which constitute gramicidin S as acyladenylates and thioesters, but the binding abilities were lower than those of the wild-type enzyme and other mutant enzymes containing the pantothenic group.     

Membranes of bacteria and mechanism of action of the antibiotic gramicidin S]

The cyclopeptide antibiotic gramicidin S taken at a concentration of 100--200 mkg/mg membrane protein rapidly increases the permeability of M. lysodeikticus protoplast membranes for substrates of respiratory chain and exogenous cytochromes c. Prolonged incubation of gramicidin S with protoplasts results in their lysis which is more fast at low temperatures. In contrast to natural gramicidin, a derivative of gramicidin S with acetylated amino groups does not inhibit either the micrococcus membrane dehydrogenase or the whole of respiratory chain and does not affect the osmotic barrier of protoplasts. Aliphatic diamines (at concentrations up to 0.1 M) and Ca2+ ions (10(-2) M) do not affect the functioning of the respiratory chain in isolated micrococcus membranes. Another derivative of the antibiotic with an increased distance of loaded amino groups from the cyclopeptide framework (diglycyl gramicidin S) affects the membrane in a way similar to that of natural gramicidin. Washing of gramicidin-treated membranes with NaCl enhances the inhibitory effect of the antibiotic on membrane enzymes. The data obtained suggest that in addition to ionic interactions some hydrophobic interactions also occur during gramicidin S binding to the bacterial membrane, probably at the expense of a hydrophobic peptide ring. It is assumed that gramicidin S, similar to Ca2+ and some other membranotropic agents provides for phase separation of negatively charged phospholipids from other groups of phospholipids, manifesting itself in an appearance of "frozen" sites on the membrane which destroys its barrier properties. This is due to the formation of ionic bonds of negatively charged phospholipids. Simultaneously, unlike Ca2+, gramicidin S, when interacting with membrane proteins, prevents their redistribution in more liquid parts of the membrane, which results in a situation when the respiratory enzymes become surrounded by alkyl chains with restricted motion.     

Tryptophan photolysis is responsible for gramicidin-channel inactivation by ultraviolet light

The decay of gramicidin fluorescence resulting from ultraviolet exposure was compared to the decay of conductance from gramicidin-containing planar bilayer membranes under the same conditions of illumination. The decay rate was the same for both processes. The fluorescence decay was identical whether gramicidin was dissolved in methanol or incorporated into lipid vesicles, indicating that the peptide conformation does not affect the sensitivity of gramicidin to photolysis. The correlation of fluorescence decay and conductance decay imply that conductance loss from gramicidin-doped membranes illuminated with ultraviolet light is due to photochemical modifications of the channel tryptophans rather than simply to disturbance of the conformation of gramicidin channels.