Mark A. Roseman

Hydrophilicity of polar amino acid side-chains is markedly reduced by flanking peptide bonds.

Several amino acid side-chain hydropathy scales have been devised on the basis of solubility and water/organic solvent partitioning data obtained with free amino acids or side-chain analogs. In nearly all cases, these scales are based upon the structure-additivity assumption; it has been assumed that the transfer free energies of the amino acid side-chains are the same in these model compounds as they are in a polypeptide. This assumption is probably wrong. In the present study, deviations from additivity for amino acid side-chains are demonstrated by comparing a theoretically derived scale, which N-acetylamino acid amides. The results show that the flanking peptide bonds dramatically reduce the hydrophilicity of the polar side-chains, with deviations up to several kilocalories (1 kcal = 4.184 kJ) for the charged side-chains at pH 7.0. Further calculation shows that these deviations are due to reductions of 40 to 85% in the unfavorable transfer free energy of the polar functional groups. In addition, proximity of the neighboring amide bonds in the parent molecule (N-acetylglycine amide) decreases the hydrophilicity of the -CONH-backbone unit by 36%. This decrease is expected to be twice as large for -CONH- units in the interior of a polypeptide backbone. The significance of these observations is: (1) valid hydropathy scales can be obtained only with model peptides; (2) deviations from additivity are expected in all solvent systems, including non-polar solvents that are thought to mimic the interior of a membrane; (3) the spontaneous insertion of polypeptides into membranes is likely to occur much more readily than has been previously thought. In order to estimate the free energy of transferring the side-chains and the polypeptide backbone from water to the interior of a lipid bilayer, the results of this study are used to construct a hydropathy scale based upon the partitioning of solutes between water and non-polar solvents. The validity of hydropathy scales that are based on criteria other than solubility and water/organic solvent partitioning data is also discussed.

Hydrophobicity of the peptide CO···HN hydrogen-bonded group☆

The hydrophobicity of the peptide C=O ... H-N hydrogen-bonded group is an important parameter that determines the structure of proteins in water and in biological membranes, and therefore the free energy of transferring this group from water to non-polar solvents should be determined accurately. The essential work on this problem was carried out by Klotz and co-workers, and has been summarized elsewhere. Using N-methylacetamide as a model peptide, the free energies of the following processes were determined; (1) formation of the C=O ... H-N bond in water, (2) formation of the C=O ... N-N bond in CCl4, and (3) transfer of N-methylacetamide from water to CCl4. (4) From (3), the free energy of transferring the non-hydrogen bonded (C=O, H-N) group from water to CCl4 was calculated. When the free energies of (1), (2) and (4) are combined, one finds that the free energy of transferring the C=O ... H-N group from water to CCl4 is a surprising -1.4 kcal/mol (1 cal = 4.184 J). This number does not seem reasonable, since it implies that the C=O ... H-N group is about as hydrophobic as an isopropyl group, i.e. the side-chain of valine. In the present report, it is shown that this apparent hydrophobicity results from an underestimation of the free energy contribution that the methyl groups make to the transfer of N-methylacetamide from water to CCl4. When appropriate methyl group transfer free energies are used, one finds that the free energy of transferring the C=O ... H-N group from water to CCl4 is +0.62 kcal/mol. Therefore, this group is relatively insensitive to solvent polarity. A similar calculation shows that the free energy of transferring the C=O ... H-O hydrogen-bonded group from water to benzene is +0.55 kcal/mol.

Effect of cholesterol on the tight insertion of cytochrome b5 into large unilamellar vesicles

When cytochrome b5 is added to large unilamellar vesicles (LUVs) of 1-palmitoyl-2-oleoylphosphatidylcholine (POPC), it binds predominantly in a loose, or transferable form. Prolonged incubation of 30 degrees C leads to insertion in the physiological tight, nontransferable form, with a halftime for the loose --> tight conversion of approx. 9 days. In this study, the effect of cholesterol on the rate of tight insertion was determined. Tight binding was assayed by depleting the LUVs of loose cytochrome b5 with an excess of SUV acceptors and then separating the liposome populations by gel-filtration or velocity sedimentation. Incorporation of cholesterol into the LUVs was found to markedly increase the rate of tight insertion, even though cholesterol decreases the equilibrium binding constant and saturation level of protein binding. The effect is not a continuously increasing function of cholesterol content, but attains a maximum at 20-25% mol%, where the rate enhancement is approx. 10-fold over baseline. At higher cholesterol levels, the rate decreases, returning to baseline at 40 mol% cholesterol. These observations are highly unusual in that cholesterol generally decreases the membrane binding affinity and the permeability of solutes, and does so as a monotonic function of cholesterol concentration (above the liquid-crystalline phase transition of the phospholipids). It is suggested that tight insertion is enhanced by lipid-protein packing mismatches and by bilayer fluidity; the former increases monotonically with increasing cholesterol whereas the latter decreases monotonically. At 20-25 mol% cholesterol the optimum balance of these physical properties is obtained for tight insertion.

Tight insertion of cytochrome b5 into large unilamellar vesicles

Cytochrome b5 spontaneously binds to liposomes in a loose, or transferable form, whereas in vivo b5 binds post-translationally to the ER in the tight or nontransferable form. The mechanism of tight insertion is unknown, except that it does not require SRP or energy input. The present study shows that prolonged incubation of b5 with large unilamellar vesicles (LUVs) of phosphatidylcholine results in slow conversion of the loose to the tight form, with a halftime of days. However, the process is complex. When the b5-LUVs are depleted of loose b5, by transfer of b5 to sonicated vesicles, the tight b5 is found to be concentrated to near saturating levels in a small fraction of the LUVs. If the LUVs devoid of tight b5 are recovered and then reincubated with fresh b5, the same slow transformation recurs. Apparently, a new population of vesicles, containing tight b5, is generated during the prolonged incubation with the protein. The b5-enriched LUVs contain about the same level of trapped sucrose as does the original vesicle preparation, indicating that vesicle integrity is maintained throughout the process. When fresh b5 is added to these tight b5-containing LUVs, all the freshly bound protein rapidly inserts (< 2 h) into the tight configuration. Apparently, the newly formed tight-b5/LUV vesicle population is insertion-active. A model for these complex transformations is proposed.

Transbilayer migration (flip-flop) of 12-(4-azido-2-nitro-phenoxy)stearoyl glucosamine in large unilamellar phospholipid vesicles

The ability of the glycolipid photoprobe, 12-(4-azido-2-nitrophenoxy)-stearoyl[1-14C]glucosamine (12-APS-GlcN), to undergo transbilayer flip-flop and intermembrane transfer between liposomes was examined. It was found that probe which was incorporated into membranes during the preparation of large unilamellar vesicles (LUVs) could be rapidly and completely extracted by incubation of these donor vesicles (in the liquid-crystalline state) with probe-free acceptor vesicles.