Lipid-peptide association and activation of lecithin:cholesterol acyltransferase. Effect of alpha-helicity.

A series of apolipopeptides, in which single proline substitutions were made at various sites in the 20-residue sequence, have been synthesized and tested. These peptides have nearly the same hydrophobic content, but very different helical contents, in a structure-making solvent. The affinity of these peptides for phospholipids was evaluated on the basis of their intrinsic tryptophan fluorescence and equilibrium dialysis against model high density lipoproteins. Proline substitutions at one end of the peptide had little or no effect on the fluorescence, circular dichroism, affinity for model high density lipoproteins, or activation of human plasma lecithin:cholesterol acyltransferase. By contrast, there was a dramatic change in all of these variables as the site of substitution was moved progressively closer to the middle of the peptide. All of these data suggested that a helix breaker that is substituted at the midpoint of a helical surface-associating peptide will greatly reduce its affinity for phospholipid surfaces. These results demonstrate that helicity and hydrophobicity are independent determinants of the affinity of an apolipopeptide for a phospholipid surface.

The plasma lipoproteins are water-soluble macromolecular particles that transport lipids in blood. They are composed of a monolayer of phospholipids, cholesterol, and specific apoproteins, surrounding a central core of nonpolar lipids, cholesteryl esters and triglycerides (1,2). Apoproteins exchange between lipoproteins by a mechanism that probably involves a transfer through the aqueous phase (3-5). A general model, the amphipathic helical theory, has been proposed for lipidapoprotein interactions (6). Hypothetically, an apoprotein has helical regions in which polar and nonpolar residues lie on opposite faces of the helix. The nonpolar face of the helix penetrates the lipid matrix and the polar face interacts with the aqueous phase. The theory is supported by numerous studies of native apoproteins and model apolipopeptides (7)(8)(9)(10)(11)(12)(13)(14). The affinity of an apopeptide for a phospholipid depends upon 1) the potential to form a-helical structures; 2) the length of the amphipathic peptide; 3) its hydrophobic content.
* This study was supported by Grants HL27341 and HL30914 from the National Institutes of Health and Grant Q-906 from The Welch Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. These criteria are not independent and, therefore, are not easy to study independently with native apolipoproteins. We have recently reported the lipid-binding characteristics of a family of acyl peptides whose hydrophobic content increased with the acyl chain length (15). The results clearly demonstrated that, for a given putative a-helicity and a constant peptide length, the free energy of association of an apolipopeptide to a lipid matrix is directly related to its hydrophobicity. The substitution of proline changes the calculated hydrophobicity very little but, as expected, has a dramatic effect on the predicted structure.

MaterMls
The LAP-20,' which has the sequence, Val-Ser-Ser-Leu-Leu-Ser-Ser-Leu-Lys-Glu-Tyr-Trp-Ser-Ser-Leu-Lys-Glu-Ser-Phe-Ser, was synthesized as previously described (16). By the same methods, a family of peptides in which a single proline was substituted for leucine at residues 5, 8, and 15, for valine at residue 1, for phenylalanine at residue 19, and for tyrosine at residue 11 was also obtained. DMPC and POPC were obtained from Avanti Biochemicals, Birmingham, AL.
[3H]Cholesterol and [3H]DMPC were purchased from New England Nuclear. Human apo-A-I was purified as previously described (17). Lecithin:cholesterol acyltransferase was purified to homogeneity by the method of Albers et al. (18). A buffer composed of 150 mM NaC1, 0.01% EDTA, 0.01% azide, and 10 mM Tris, pH = 7.4, was used except where noted.

Methods
Fluorescence Measurements-Fluorescence spectra were recorded on an SLM photon counting spectrofluorimeter operated at ambient temperature using an excitation wavelength of 280 nm. Emission spectra were recorded between 300 and 450 nm using 2 nm slits throughout. Each peptide was incubated overnight at 23 "C with various concentrations of single bilayer vesicles of DMPC or POPC produced by sonication.
Density Gradient Ultracentrifugation-Approximately 0.5 mg of peptide was lyophilized in a 13-X 100-mm screw-cap culture tube. The peptide was solubilized in 1 ml of 0.1 M NaCl and 0.01 M Na2HP0,, pH 7.4, and the concentration was determined by recording the UV spectrum between 320 and 240 nm on a Cary 15 spectrophotometer. A dispersion of [3H]DMPC (22 pCi/mmol) was prepared by sonication of 30 mg in 2.8 ml of 0.1 M NaCl with a probe sonicator. The dispersion was maintained at 4 "C under nitrogen during the 15 min required to clarify the solution. The dispersion was centrifuged for 30 min at 14,000 rpm in a Beckman 5-21 centrifuge to remove any titanium. The DMPC and peptide were mixed in appropriate quantitites to obtain molar ratios of lipid to peptide of 501. The screw-cap tubes were placed in a 24 "C bath for 24 h. The volume of the peptide-lipid mixture was adjusted to 1. Na2HP04, 0.10 M NaCl, and 1.31 M CsC1, pH 7.4, was placed in the bottom of a 5-ml polyallomer tube and the peptide-lipid solution was carefully added to the top. The tube containing the peptide-lipid mixture was rinsed with 1 ml of the CsCl solution and the rinse solution was added to the top of the centrifuge tube. The tubes were placed in the buckets of a Beckman SW 50.1 rotor and centrifuged at 45,000 rpm at 20°C for 72 h. The gradients were fractionated from the top of the tube into 250-pl aliquots with a peristaltic pump, Densiflow, and a drop-counting fraction collector. The density of each fraction was determined from the refractive index. Equilibrium Dialysis-M-HDL were prepared by mixing POPC and apo-A-I (1001 molar ratio) in the presence of cholate at room temperature; the detergent was removed as previously described (19). The peptides were radioiodinated by the chloramine-?' method (20). One ml of M-HDL, at various concentrations, was placed inside the dialysis tubing. A trace of each labeled peptide was added to the outside compartment in 0.02 ml of 50% 2-propanol. At equilibrium, the concentration of POPC was determined (21), and the amount of peptide bound to the M-HDL was calculated from the difference between the radioactivities of the inside and outside compartments. L,ecithin:Cholesterol Acyltransferase Assay-Enzyme activity was assayed by using a modification of the method of Glomset (22) in which minicolumns of silica gel (1 ml) constructed from Pasteur pipettes were substituted for thin layer plates (23). Further details are provided in the legend to Fig. 5.
Circular Dichroic Spectra-Circular dichroic spectra of the peptides in the helix-making solvent, hexafluoropropanol, were recorded on a Jasco J-500A spectropolarimeter operated under ambient conditions. The helical content of each peptide was calculated from the molar ellipticity at 222 nm (24).

RESULTS AND DISCUSSION
To investigate the importance of helical potential, we have synthesized a series of peptides with proline substituted at various positions of the previously reported LAP-20 peptide (12). Fig. 1 shows the Chou-Fasman predicted structures for these six peptides and their average hydrophobicities calculated using the hydrophobicity scale of Bull and Breese (25). Below, we present the results of our studies on the interaction of these peptides with phospholipids. These results demonstrate the importance of helical potential to protein-phospholipid interactions.
Lipid-Apolipopeptide Interaction-The spectroscopic changes that accompany the addition of POPC or DMPC to the model peptides were a function of the identity of the peptide. All peptides exhibited a blue shift that is characteristic of the transfer of the tryptophan to the hydrophobic environment of the lipid, although the magnitude of this effect varied markedly from one peptide to another (Fig. 2). A shift of 10 nm required the addition of 0.04, 1.2, 4.0, and 7.5 m M POPC to LAP-20, &Pro-, 15-Pro-, and 8-Pro-LAP-20, respectively. This indicated a decreasing affinity of the peptides for the lipid in the order given. Based upon the circular dichroic spectra, the helical contents of the peptides recorded in hexafluoropropanol varied from 32% for 11-Pro-LAP-20 to 59% for 1-Pro-LAP-20 (Table I). As the position of the proline was moved closer to the midpoint of the 20-residue peptide, the helical content decreased dramatically; this behavior was in fairly good agreement with the Chou-Fasman analysis in Fig. 1. The decreased helical content suggests that the helical potential is, as expected, a sensitive function of the location of the proline residues. Density gradient ultracentrifugation of the peptide:DMPC mixtures permitted isolation of complexes with these same peptides (Fig. 3). However, the 15-Pro-LAP-20 lipidpeptide complex was unstable to recentrifugation. In the complex with DMPC, the 1-and 5-Pro-LAP-20 formed the complexes with the highest helical contents, 63% and 54%, respectively; 19-Pro-LAP-20 had 47% helix. 15-Pro formed a quasi stable complex (24-38% a-helix) that was unstable to recentrifugation. The 8-and 11-Pro-LAP-20 did not form a stable complex or show changes in their helical contents in the presence of DMPC. Therefore, the restriction of helix propagation in a hydrophobic peptide imposed by the introduction of a proline residue prevents the formation of a stable phospholipidpeptide complex when the helical segment is less than 15 residues, irrespective of the hydrophobicity of the peptide.   The binding of the synthetic peptides to M-HDL, which was further quantified by equilibrium dialysis, was independent of the M-HDL concentration (Fig. 4); the value of K was obtained from the slopes of the indicated lines. These were 35,000 for LAP-20 and 1-Pro-LAP-20, and 5,700, 2,500, and 700, respectively, for 5-PrO-, ls-pro-, and 8-Pro-LAP-20. The free energies calculated from AG = -RT 1nK were -6.3 and -3.9 kcal/mol for LAP-20 (or 1-Pro-LAP-20) and8-Pro-LAP-20, respectively. Comparison of these binding constants with our previous data (15) on acylated peptides revealed that the transfer of a proline from the 1-position to the 8-position of LAP-20 had an effect that was equivalent to the removal of 8 methylene units from the acyl chain of the acylated peptide. These results demonstrated that the maximum expression of the hydrophobicity of an amphiphilic peptide is greatly reduced by the introduction of a helix breaker. Although helical potential and a high hydrophobicity were both hallmarks of the lipid-binding hypothesis of Segrest et al. (6), our results suggest that a-helicity and calculated hydrophobicity are independent determinants of the affinities of model peptides for lipid surfaces. Studies of these peptides with space-filling models showed that portions of the peptide on either side of the proline could be placed in a helix, but that the hydrophobic faces of both parts could not be in simultaneous contact with the same surface. This observation could explain why the lipophilicity of a peptide containing a helix breaker in the middle of the primary structure is so much lower than one with a similar hydrophobicity but without a helix breaking residue.
Lecithin:CholesteroZ Acyltransferase Activation-Apo-A-I and several synthetic peptides have been shown to activate 1ecithin:cholesterol acyltransferase in vitro (9, 12,14,15,26,27). Therefore, we compared the activation of each peptide with that of apo-A-I as follows. Initial velocities were meas- of peptide binding to the M-HDL (Fig. 6). This confirmed that there was a direct relationship between binding and activation as previously noted by Fukushima et al. (9). This relationship was also noted with a series of peptides with similar helical potentials but different hydrophobicities ( l a , so that helical potential alone does not lead to activation. Rather, our results support the concept that the model apopeptides stimulate 1ecithin:cholesterol acyltransferase through the activation of a phospholipid surface by their bound helical segments. The present results and the past results from this (15) and other laboratories (8-13) support the overall validity of the amphipathic helical theory of lipid-associating proteins; i.e. that both hydrophobicity and helical potential might function as independent determinants of this class of peptides with both being required for the interaction with phospholipid surfaces. Previously, we reported that the hydrophobicity of a peptide can be increased by the covalent attachment of a saturated acyl group to the amino terminus. The linear relationship between the length of the acyl group and the free energy of association with lipids suggested that the free energies of association of the peptide and acyl chains for lipid surfaces were additive (15). Subsequently, we found that the plasma turnover times of the synthetic peptides in the rat were a predictable function of their affinities for M-HDL (28); the peptides with high affinities for M-HDL had plasma lifetimes similar to that of HDL and were targeted to those tissues that have HDL receptors. The peptides with low affinities for M-HDL had much shorter turnover times and were degraded mainly in the kidneys. Application of a theoretical model (29) strongly suggested that the peptides were distributed between the M-HDL and the surrounding aqueous phase according to their respective hydrophobicities. The monomeric peptides distribute to many tissue compartments, whereas those bound to HDL are taken up by tissues with HDL receptors. A similar finding was reported by Glass et al. (30) who observed that apo-A-I was degraded in liver, adrenals, and ovaries, as well as kidneys. Collective consideration of these and the present data suggests that the magnitude of the equilibrium constant for the distribution of peptides and proteins between lipoproteins and water has important metabolic implications. Since there is a clear relationship between the affinity of peptides for M-HDL and their catabolism (28- 30), then any structural change that alters affinity should produce a change in the rate and site of catabolism. We suggest that, in addition to hydrophobicity, the a-helical potential of a lipid-associating peptide is an important regulator of its catabolism. For strong binding to native or model HDL, one requires both a high hydrophobicity and a-helical potential. High hydrophobicity with low helical potential, or vice versa, can lead to poor lipid-protein association.