Two Homologous Apolipoprotein AI Mimetic Peptides

Two related 18-amino acid, class A, amphipathic helical peptides termed 3F-2 and 3F14 were chosen for this study. Although they have identical amino acid compositions and many similar biophysical properties, 3F-2 is more potent than 3F14 as an apolipoprotein AI mimetic peptide. The two peptides exhibit similar gross conformational properties, forming structures of high helical content on a membrane surface. However, the thermal denaturation transition of 3F-2 is more cooperative, suggesting a higher degree of oligomerization on the membrane. Both 3F-2 and 3F14 promote the segregation of cholesterol in membranes containing phosphatidylcholine and cholesterol, but 3F-2 exhibits a greater selectivity for partitioning into cholesterol-depleted regions of the membrane. Magic angle spinning/NMR studies indicate that the aromatic residues of 3F-2 are stacked in the presence of lipid. The aromatic side chains of this peptide also penetrate more deeply into membranes of phosphatidylcholine with cholesterol compared with 3F14. Using the fluorescent probe, 1,3-dipyrenylpropane, we monitored the properties of the lipid hydrocarbon environment. 3F-2 had a greater effect in altering the properties of the hydrocarbon region of the membrane. The results are consistent with our proposed model of the effect of peptide shape on the nature of the difference in peptide insertion into the bilayer.

There is growing evidence that certain apo 1 A-I mimetic, class A amphipathic helical peptides can be used to inhibit atherosclerosis (1). The oral administration of peptide 4F synthesized from all-D amino acids (D-4F) protects mice from diet-induced atherosclerosis without altering plasma cholesterol levels (2,3). Preliminary studies also suggest that oral administration of D-4F to LDL receptor null and apo E null mice causes the rapid formation and clearance of small high density lipoprotein-like particles containing peptide, cholesterol, apo A-I and paraoxonase, an enzyme capable of converting pro-inflammatory high density lipoprotein into anti-inflammatory high density lipoprotein (4). The peptide 4F has also been tested in vitro and shown to effectively inhibit lytic peptide-induced hemolysis, inhibit oxidized phospholipid-induced monocyte chemotaxis, scavenge lipid hydroperoxides from LDL (5), and maintain endothelial nitric-oxide synthetase activity in the presence of atherogenic concentrations of LDL (6). To study the relationship of peptide structure to anti-atherosclerogenic potency, we studied the properties of four related 18-amino acid, class A amphipathic helical peptides (5). All of these peptides had identical amino acid compositions and very similar physical properties, yet two of these peptides, 3F-1 and 3F-2, were more potent in inhibiting lytic peptide-induced hemolysis, inhibiting oxidized phospholipid-induced monocyte chemotaxis, and scavenging lipid hydroperoxides from LDL compared with the analogs 3F 3 and 3F 14 (5). In the present work we compare the interaction with phospholipid bilayers with and without cholesterol, of the most potent peptide of this series, 3F-2 (Ac-DKWKAVYDKFAEAFKEFL-NH 2 ), and the least potent among these peptides, 3F 14 (Ac-DWLKAFYDKVAEKFKEAF-NH 2 ) (Fig. 1). We have previously shown that the potent analog 4F that has four, rather than three Phe, is capable of forming cholesterol-rich domains by preferentially interacting with regions of the membrane that are depleted of cholesterol (7). 3F-2 exhibits similar biological potency to 4F but has somewhat less activity in inhibiting oxidized phospholipid-induced monocyte chemotaxis, about the same activity in scavenging lipid hydroperoxides from LDL and greater activity in inhibiting lytic peptide-induced hemolysis (5). In contrast, although 3F 14 is a class A amphipathic helical peptide with some anti-atherosclerogenic activity, it has a much weaker potency than either 3F-2 or 4F in the activities mentioned. This difference in potency is reflected in differences in the red edge effect in Trp emission from the peptide and in the quenching of 2-(3 (diphenylhexatrienyl)propanoyl)-1-hexadecanoyl-sn-glycero-3-phosphocholine (5). In this study we further evaluate the interaction of 3F-2 and 3F 14 with model membranes using NMR, DSC, CD, and fluorescence.

MATERIALS AND METHODS
Lipids-The lipids used in this study were obtained from Avanti Polar Lipids (Alabaster, AL). The purity of the phospholipids was verified by measuring the cooperativity and temperature of the phase transition using DSC.
Peptide Synthesis-The peptides were synthesized by the solid phase method with a Protein Technologies PS-3 automatic peptide synthesizer using the procedures described previously (2,8). The peptides were purified using a preparative HPLC system (Beckman Gold), and the purity of the peptides was ascertained by mass spectral analysis and analytical HPLC. * This work was supported by Grant MT-7654 from the Canadian Institutes of Health Research and Grants HL 34343 and RO1 HL 65663 from the National Institutes of Health. 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.
Concentrations of Peptide and Lipid-The concentrations of peptide solutions in buffer were determined spectrophotometrically using the absorbance at 280 nm and an extinction coefficient of 6970 cm Ϫ1 M Ϫ1 , calculated from the Tyr and Trp content. Phospholipid concentration was determined by phosphate analysis (9).
Preparation of Samples for DSC and NMR Experiments-Phospholipid and cholesterol were dissolved in chloroform/methanol (2/1, v/v). For samples containing peptide, an aliquot of a solution of the peptide in methanol was added to the lipid solution in chloroform/methanol. The solvent was then evaporated under a stream of nitrogen with constant rotation of a test tube so as to deposit a uniform film of lipid over the bottom third of the tube. The last traces of solvent were removed by placing the tube under high vacuum for at least 2 h. The lipid film was then hydrated with 20 mM PIPES, 1 mM EDTA, 150 mM NaCl with 0.002% NaN 3 , pH 7.40, and suspended by intermittent vortexing and heating to 50°C for 2 min under argon. The samples used for NMR analysis were hydrated with the same buffer made in 2 H 2 O adjusted to a pH meter reading of 7.0 (pD ϭ 7.4). The samples used for NMR were incubated 24 h at 4°C to allow conversion of anhydrous cholesterol crystals to the monohydrate form. For the NMR measurements, the samples were spun in an Eppendorf centrifuge at room temperature. The resulting hydrated pellet was transferred to a 12-l capacity Kel-F spherical insert of an 18 ϫ 4-mm ZrO 2 rotor, attempting to pack the maximal amount of lipid into the rotor while maintaining it wet.
Differential Scanning Calorimetry-Measurements were made immediately after sample preparation using a Nano Differential Scanning Calorimeter (Calorimetry Sciences Corporation, American Fork, UT). The scan rate was 2°C/min, and there was a delay of 5 min between sequential scans in a series to allow for thermal equilibration. DSC curves were analyzed by using the fitting program, DA-2, provided by Microcal Inc. (Northampton, MA) and plotted with Origin, version 5.0.
Centrifugation Assay for Membrane Binding of 3F Peptides-The fraction of peptide bound to the lipid after three heating and cooling cycles between 0 and 100°C was determined by separating the free peptide by centrifugation. The vesicles with bound protein were pelleted at 200,000 ϫ g for 90 min at 25°C. A clear supernatant was separated from the solid pellet and assayed for protein by absorption at 278 nm. The absorption of a blank was subtracted, and a base line, set at 350 nm for each sample, was used to correct for residual scattering. In addition, the amount of lipid in both pellet and supernatant fractions was determined by phosphate analysis. Cholesterol concentration was determined by a modified version of the procedure of Zlatkis et al. (10). Briefly, after extraction with chloroform, the organic phase was dried under nitrogen and then placed under vacuum to completely remove the solvent. To the dried samples or to the dry cholesterol standards, glacial acetic acid was added, followed by a reagent composed of FeCl 3 and 85% H 3 PO 4 in sulfuric acid. The samples were placed in a boiling water bath for 3 min to develop a purple color. After cooling the absorbance was read at 550 nm.
To test for the presence of lipoprotein particles in the supernatant, it was analyzed by nondenaturing PAGE using the method of Laemmli (11). An aliquot of the supernatant containing 3 g of peptide was loaded on a 4 -20% gel and run under native conditions for 20 h. The gel was then stained with colloidal blue and destained with water.
Circular Dichroism-CD spectra were recorded on an AVIV model 215 spectropolarimeter using a quartz cell with a 0.1-cm path length. The cuvette was placed in a jacketed cell holder maintained at the desired temperature with circulating thermostatic fluid. The lipid was solubilized by incubation with the peptide followed by sonication to ensure minimal effects of light scattering.
Temperature heating or cooling scans were performed at a rate of 2°C/min, with a 30-s equilibration pause at each point for thermal equilibration. Ellipticity at 222 nm was measured as a function of temperature between 25 and 95°C and then cooling back to 25°C to assess the reversibility of denaturation.
Large Unilamellar Vesicles (LUVs)-Lipid films were made by dissolving appropriate amounts of lipid in a mixture of chloroform/methanol (2/1, v/v) and dried in a test tube under nitrogen to deposit the lipid as a thin film on the wall of the tube. Final traces of solvent were removed in a vacuum chamber attached to a liquid nitrogen trap for 2-3 h. Dried films were kept under argon gas at Ϫ30°C if not used immediately. The films were hydrated with buffer, vortexed extensively at room temperature, and then subjected to five cycles of freezing and thawing. The homogeneous lipid suspensions were then further processed by 10 passes through two stacked 0.1-m polycarbonate filters (Nucleopore Filtration Products, Pleasanton, CA) in a barrel extruder (Lipex Biomembranes, Vancouver, Canada), at room temperature. LUVs were kept on ice and used within a few hours of preparation.
Tryptophan Fluorescence-Fluorescence emission spectra of the peptides in the presence and absence of LUVs were measured at 25°C using an SLM Aminco Series II luminescence spectrometer. The excitation wavelength was varied between 280 and 310 nm. The emission scans were recorded and then corrected for inner filter and instrumental effects.
Excimer Formation in the Fluorescent Probe PC 3 P-The physical properties of the hydrophobic region of the membrane and the effects of the 3F peptides were monitored by measuring the fluorescent emission of PC 3 P (Molecular Probes, Eugene, OR). It has been shown that the ratio of the intensities of excimer and monomer emission is sensitive to the "fluidity" of the membrane interior (12). LUVs were prepared containing lipids and PC 3 P at a 100:1 molar ratio in Hepes buffer (10 mM Hepes, 0.14 M NaCl, 1 mM EDTA, pH 7.4). Florescence measurements were made in 1 ϫ 1-cm-square quartz cuvettes containing 2 ml of Hepes buffer at room temperature. The excitation wavelength was 344 nm using slits with 4-nm band pass in both emission and excitation. Emission scans were made between 360 and 550 nm before and after the addition of peptide. The results are summarized as the ratio of the intensity of excimer emission at 476 nm to that of monomer emission at 389 nm (I e /I m ). The experiments were repeated twice with two independent preparations. The error in I e /I m for duplicate measurements with the same preparations was between 1 and 2%, but between independently prepared samples the absolute value of the ratios varied between 5 and 10%; however, they always exhibited similar relative values with and without either of the two peptides.
1 H NOESY MAS NMR-High resolution MAS spectra were acquired using a spinning rate of 4 kHz in a Bruker DRX 500 NMR spectrometer. Probe temperature was 24 Ϯ 1°C. The two-dimensional NOESY spectra were obtained using mixing times of 50 and 300 ms. The resonances were assigned based on their close similarity to literature values for phosphatidylcholine (13), cholesterol (14), and amino acid residues (15).

RESULTS
DSC-Lipid mixtures of SOPC containing 0, 0.3, 0.4, and 0.5 mol fractions cholesterol were analyzed by DSC in the presence of 0, 5, 10, and 15 mol % 3F-2 or 3F 14 . As examples, we present the results from mixtures containing 0 or 15 mol % of each of the peptides (Fig. 2). The DSC curves are presented as the excess heat capacity/mol cholesterol as in our earlier paper on 4F (7). Both peptides clearly promote the separation of choles-FIG. 1. Helical wheel representation of 3F-2 and 3F 14 and molecular models of these peptides with surrounding phosphatidylcholine. The wheel is projected along the axis of the helix from the N to the C terminus with the hydrophobic side facing downward. The primary structure is given above each wheel diagram. The amino acid composition of both peptides is the same. The sequence is different. The plus and minus signs denote the charges on the amino acids at neutral pH. The red color denotes an acidic residue, the blue color denotes a basic residue, and the bold black denotes aromatic residues.
terol into crystalline domains at higher mol fractions of cholesterol and of peptide (Table I). The cholesterol crystals formed are in a metastable state and disappear after sequential heating and cooling scans between 0 and 100°C (Fig. 2). This has been found previously, albeit to a lesser extent, with the peptide 4F (7). Scanning only up to 50°C can eliminate much of this loss (16), but the higher temperatures are required to observe the unfolding transition of the peptide. 3F 14 causes  14 . Scan rate 2 K/min. The rows correspond to lipid compositions of 7:3 (top row), 6:4 (middle row), and 1:1 (bottom row) SOPC:cholesterol, respectively. The columns correspond to lipid alone (first column) or lipid with the addition of 15% of either 3F-2 (second column) or 3F 14 (third column). Lipid concentration is 2.5 mg/ml in 20 mM PIPES, 1 mM EDTA, 150 mM NaCl with 0.002% NaN 3 , pH 7.40. Sequential heating and cooling scans between 0 and 100°C. The numbers are the order in which the scans were carried out, with scans 1 and 3 being heating scans, each of which was followed by one of the cooling scans 2 or 4. Scans were displaced along the y axis for clarity of presentation. separation of cholesterol crystals at lower mol fractions of cholesterol and peptide than is required with 3F-2. A more marked difference is observed in the peak at the lowest transition temperature that is ascribed to the gel to liquid crystalline transition of SOPC. The enthalpy of this transition can be reasonably estimated from cooling scans. For pure SOPC (data not shown), the transition occurs at 5.5°C on cooling at 2 K/ min with a transition enthalpy of 4 kcal/mol (16). With pure SOPC without cholesterol, the addition of 15 mol % 3F-2 lowers the enthalpy to 2.2 kcal/mol and to 2.4 kcal/mol with 3F 14 (data not shown). With mixtures of cholesterol and SOPC the addition of increasing concentrations of 3F-2 eliminates this transition, whereas with 3F 14 the enthalpy of this transition is slightly increased (Table II).
The highest temperature transition observed in the DSC scans corresponds to that of the unfolding of the peptide. It occurs at about 65°C for both peptides; with 3F 14 the transition is readily observed only with SOPC in the absence of cholesterol (not shown), but the transition is quite prominent in scans with 3F-2 (Fig. 2). The transition temperature is slightly lower on cooling than on heating.
CD-The CD spectra in buffer of both 3F-2 and 3F 14 exhibit some dependence on peptide concentration (5). The addition of SOPC or SOPC:cholesterol (1:1) to 3F-2 at 100 M peptide results in little change in the secondary structure of 3F-2 ( Fig.  3) but slightly increases the magnitude of the CD for 3F 14 (Fig.  3). The temperature dependence of the spectrum shows a substantial loss of secondary structure on heating to 95°C with both 3F-2 and with 3F 14 (Fig. 4). The thermal transition is broad with some hysteresis on heating and cooling.
Solubilization of Lipid and Peptide-Mixtures of SOPC, cholesterol, and the peptide 3F-2 or 3F 14 were centrifuged after the DSC experiments. In the presence of peptide, lipid is partially solubilized (Fig. 5). A larger fraction of cholesterol is solubi-  ND  2200  ND  ND  2400  30%  500  800  345  278  515  515  530  40%  350  0  0  0  470  570  265  50%  210  0  0  0  300  lized, compared with SOPC in samples containing 0.3 mol fraction cholesterol. It was confirmed by PAGE that lipoprotein particles were formed after either a mixture of SOPC and cholesterol at a molar ratio of 1:1 or 7:3 and also containing 15 mol % of 3F-2 or 3F 14 was incubated at room temperature for 3 h and then centrifuged. Particles of ϳ100 Å size were seen (Fig. 6), confirming that soluble lipoprotein particles were formed with both peptides. The fact that the major fraction of both lipid and peptide is in the insoluble fraction at the low concentrations used for DSC indicates that in the case of the much higher concentrations used for NMR, the major fraction of peptide and lipid are found in the pellet and not in solubilized micellar form. This is in agreement with our finding of a bilayer shaped static 31 P NMR powder pattern for this lipid mixture (see below).
Tryptophan Fluorescence-The fluorescence emission spectra of 3F-2 and 3F 14 were measured in buffer and in the presence of lipid (Fig. 7). The emission maximum of 3F-2 in buffer is 336 nm compared with 338 nm for 3F 14 in buffer. In the presence of lipid, either with or without cholesterol, the emission is close to 333 nm. The results indicate that the Trp residue inserts into the bilayer. These values are somewhat blue-shifted compared with our previous study, but in a different lipid system and at a somewhat higher peptide concentration. In addition, there may be some time-dependent changes accounting for part of the difference. There is an ϳ2-fold increase in the emission intensity of 3F-2 in the presence of SOPC without cholesterol, which is smaller in the presence of 1:1 SOPC:cholesterol. This could be a result of 3F-2 inserting more deeply into bilayers not containing cholesterol, but there may also be contributions from scattering or small differences in peptide concentration (Fig. 7). Trp emission maximum is not sensitive to the presence of cholesterol, indicating that the polarity of the Trp environment is not greatly altered. This behavior is quite similar to that found with 4F (7). 3F 14 has almost identical emission intensity in the presence of SOPC either with or without cholesterol (Fig. 7), suggesting that this peptide does not preferentially interact with cholesterol-depleted domains. The blue shift caused by the addition of the lipid is somewhat greater with 3F 14 than with 3F-2, suggesting that the Trp of the former peptide is more deeply buried in the membrane.
We have also determined the effect of cholesterol on the red edge excitation shift. In accord with our previous findings (5), there is no red edge excitation shift with 3F-2, but there is with 3F 14 , indicating that Trp residues are more rigid with 3F-2 in presence of cholesterol containing membranes compared with 3F 14 . It should be noted that 3F-2 has Trp at the center of the nonpolar face, whereas Leu appears at the center of the nonpolar face of in 3F 14 . This could be interpreted as a consequence of Trp tending to push the peptide up toward the lipid water interface, whereas Leu at the center of the nonpolar face increases interaction with the lipid acyl chain and hence results in a deeper penetration into phospholipid bilayer. Cholesterol has no effect on this phenomenon (Fig. 8).
Excimer Formation in the Fluorescent Probe PC 3 P-The I e /I m ratio of PC 3 P (1% of lipid) was determined in LUVs of POPC and POPC:cholesterol (1:1) with and without the addition of 3F-2 or 3F 14 at a 10:1 lipid to peptide ratio (Fig. 9). In agreement with previous results (12), in the present work we also find that cholesterol markedly lowers the I e /I m ratio. This is likely a consequence of the lower rate and extent of molecular motion in the presence of cholesterol decreasing the rate of conformational change in the fluorescent probe. The effect of the peptide is smaller and tends to increase with higher concentrations of probe, suggesting that there are both inter-and intra-molecular formation of excimers. This ratio is insensitive to the presence of 3F 14 but is affected by 3F-2, indicating that 3F-2 has a greater effect on hydrocarbon packing and/or dynamics.
1 H NOESY MAS NMR-Slices from the two-dimensional NOESY spectrum of 3F-2 in the presence of POPC, at a 1:10 peptide to lipid ratio, are presented for the spectral region of the aromatic side chain resonances using 50-or 300-ms mixing times (Fig. 10). The 31 P NMR powder pattern demonstrated that the major fraction of the lipid was in a bilayer arrangement, although there was a significant isotropic component in the spectra of both peptides in the presence of POPC. There are cross-peaks between the aromatic resonance and the protons from the lipid, particularly those of the CH 2 protons, indicating insertion of the peptide into pure POPC membranes. The sign of the nuclear Overhauser enhancement with 50-ms mixing time is negative, corresponding to rapid molecular motion, on this time scale, between the peptide and lipid. In addition, for several of the slices, peaks in the aromatic region, in addition to that on the diagonal (chemical shift of the slice), are clearly observed with 3F-2. This indicates that the aromatic resonances are in close proximity, likely as a result of stacking of the aromatic groups, both in bilayers of POPC (Fig. 10) as well as with a 1:1 mixture of POPC and cholesterol (Fig. 11). The 31 P NMR powder pattern of 3F-2 with a 1:1 mixture of POPC and cholesterol showed only a minor isotropic component for this sample. The NOESY slices with 3F-2, for the samples with cholesterol, generally have somewhat larger cross-peaks. This is particularly evident for the cross-peaks to the acyl CH 3 group in the spectrum with 300 ms mixing time (Fig. 11).
We performed a similar analysis with the 3F 14 peptide. The static 31 P NMR powder patterns for this peptide were similar to those for 3F-2 showing a bilayer shape pattern with a significant isotropic component in the absence of cholesterol but only a minor isotropic peak in the equimolar mixture of POPC and cholesterol. The slices for the sample of 3F 14 with POPC alone (Fig. 12) generally had cross-peaks of positive sign (i.e. negative NOE), opposite to that with 3F-2. In addition, only the peak on the diagonal was observed in the aromatic region, indicating that the aromatic groups are less stacked in 3F 14 than in 3F-2, as might be anticipated on the basis of the larger cluster of aromatic residues seen in the helical wheel projection of 3F-2 (Fig. 1). The cross-peaks between the lipid and 3F 14 are particularly weak in the presence of cholesterol (Fig. 13), suggesting that this peptide is largely excluded from cholesterol-containing membranes. In addition, the aromatic region of the one-dimensional spectrum is particularly well resolved compared with other cases.
Interaction of the peptide with lipid can also be assessed by monitoring the changes in the chemical shifts of the lipid resonances on introduction of the peptide (Table III). These changes in chemical shifts may arise from ring current effects as well as from changes in the polarity of the environment. In the case of 1 H MAS/NMR, only changes in the spectrum of the phospholipid can be assessed, because resonances are not observed from protons of cholesterol in these lipid mixtures (13), and the lower concentration of peptide makes it difficult to discern its resonances. The addition of 3F-2 or 3F 14  tial location of the peptide in the bilayer. Because cholesterol resonances are not observed in 1 H MAS/NMR, we also measured 13 C MAS/NMR. The changes in chemical shift are generally small, and there is no large difference between 3F-2 and 3F 14 . In mixtures containing cholesterol the changes in chemical shifts are comparable for POPC and for cholesterol (not shown). DISCUSSION Using in vitro assays expected to correlate with protection against atherosclerosis, 3F-2 and 3F 14 exhibit quite different potency (5); however, the differences in their biophysical properties have been found to be relatively small (this work and Ref. 5). Both peptides are class A amphipathic helices and therefore can fold into a helical structure resembling amphipathic helices contained in exchangeable plasma apolipoproteins, i.e. class A amphipathic helices (17). Although the position of the Trp residue in the two peptides is different, the nature and location of the other residues on the hydrophilic face of the helical conformation of these two peptides are identical, as is their total amino acid composition. The HPLC elution profile and monolayer collapse pressure indicate that the two peptides have similar hydrophobicities, but the 3F-2 is somewhat less hydrophobic (5). Both peptides can rapidly solubilize 1-palmitoyl-2-oleoyl phosphatidylcholine at an equimolar ratio of peptide and lipid (5). The CD spectra of the two peptides are similar, and we show in this work that the secondary structure is independent of the presence of cholesterol (Fig. 3). We also show that both peptides undergo a loss of secondary structure on heating (Fig. 4). However, the thermal denaturation of 3F-2 is more prominent in DSC scans of 3F-2 than of 3F 14 (Fig. 2). Because a similar amount of helicity is lost upon heating of the two peptides, we suggest that the transition of 3F-2 is more cooperative because this peptide has a higher degree of oligomerization on the membrane surface compared with 3F 14 .
We have shown that another biologically potent class A helical peptide, 4F, promoted the formation of cholesterol-rich domains by preferentially interacting with the phospholipid component of cholesterol/SOPC mixtures (7). This finding provided an interesting contrast with the peptide LWYIK that FIG. 7. Fluorescence emission spectra peptides with and without cholesterol. The left panel is for 3F 14 , and the right panel is for 3F-2. Curve 1, 15 M peptide in buffer. Curve 2, 15 M peptide mixed with SOPC at a lipid to peptide ratio of 6; Curve 3, 15 M peptide mixed with SOPC:cholesterol (1:1) at a lipid to peptide ratio of 6. Fluorescence intensities normalized to make the emission maximum for peptide in buffer equal to 1.0. The spectra were acquired using an excitation wavelength of 280 nm at a temperature of 25°C. preferentially interacted with cholesterol (16). Rearrangement of cholesterol in membranes can result from preferential interaction of proteins with cholesterol-rich domains as well as with cholesterol-depleted domains (18). These observations are likely to be relevant to the mechanism of formation of "rafts" in biological membranes. The DSC shows that 3F-2 interacts preferentially with the phospholipid in mixtures of cholesterol and SOPC. This is indicated by the fact that this peptide is much more potent in lowering the phase transition of SOPC than is 3F 14 (Fig. 2 and Table II). 3F-2 also promotes the formation of cholesterol crystals at mol fractions of cholesterol much lower than are required for their formation in the absence of peptide ( Fig. 2 and Table I). The peptide 3F 14 also promotes the formation of cholesterol crystals. However, 3F 14 does not lower the enthalpy of the phase transition of SOPC in mixtures with cholesterol (Table II), indicating that it is not interacting preferentially with SOPC. 3F 14 also does not greatly increase the enthalpy of SOPC by removing cholesterol, as we had shown for LWYIK (16). We suggest that a contributing factor to the promotion of cholesterol segregation in the membrane is through an increase in the lateral pressure of the membrane "squeezing out" cholesterol. The greater preference of 3F-2 for cholesterol-  depleted domains is also common to another biologically potent peptide of this series, 4F (7). This correlation is not seen with the less biologically active peptide 3F 14 .
There is also a difference between the two peptides in how they interact with the membrane. The emission intensity from the Trp of 3F-2 is increased more by lipid, especially the lipid mixture containing cholesterol (Fig. 7), than is the case of 3F 14 . The behavior of the Trp fluorescence of 3F-2 is qualitatively similar to that of 4F (7), even though the Trp is in different positions for the two peptides. The similar behavior of the Trp fluorescence of two peptides, despite the different position of Trp in the helical wheel, suggests that 3F 14 penetrates more deeply into the bilayer. The greater number of cross-peaks between protons of the aromatic residues of 3F-2 ( Fig. 10) indicates that its side chains are stacked in the presence of lipid. In the presence of cholesterol particularly, 3F-2 exhibits several strong, negative cross-peaks with several lipid protons, including those from the terminal CH 3 group of the acyl chain. The peaks of negative sign of the NOESY suggests increased molecular motion that could allow transient access to the protons in the center of the bilayer. This conclusion appears opposite to that derived from red edge excitation shift, indicating a restricted motion of the Trp of 3F-2. We suggest that this difference is a consequence of the widely different time scale for the two measurements. Fluorescence decay occurs in nanoseconds, whereas proton relaxation occurs in milliseconds to seconds. Hence in the longer time scale of NMR, the relative position between the Trp and the lipid can change, but locally around the Trp the environment is rigid, and there is slow reorientation of the surrounding dipoles. According to the model we proposed (Fig. 14), 3F-2 would, because of its cylindrical shape, short length, depth of penetration, and aromatic side chains pushing the peptide axis up closer to the polarnonpolar interface, introduce a destabilization in the bilayer resulting in a decrease in the lipid order parameters. The decreased penetration of 3F-2 compared with 3F 14 is also indicated by the small shift in Trp emission wavelength caused by lipid (Fig. 7). Again there is a difference between peptide penetration into the membrane as assessed by NOE effects and by fluorescence. It is possible that the greater bilayer disruption caused by 3F-2 allows some penetration of water into the membrane, resulting in a smaller effect on Trp emission, even though the NOSEY spectra indicate a greater penetration of this peptide. In comparison, 3F 14 appears to be largely excluded from bilayers containing cholesterol. It shows only very weak cross-peaks between the aromatic resonances and the lipid protons (Fig. 11). This is also in accord with the smaller change in the intensity of emission from the Trp residue in the presence of cholesterol. However, when 3F 14 incorporates into the bilayer, it may be able to pack with the acyl chains with less disruption of the hydrocarbon portion of the bilayer, as suggested in Fig. 14.
The NMR parameters that we have monitored are not very sensitive to the hydrocarbon packing or molecular motion in the interior of the membrane. We have therefore also studied the properties of the fluorescent probe PC 3 P that has been suggested to monitor changes in membrane motional properties (12,19). We do not wish to use the probe as the basis for a model of the interaction of the peptide with lipid but rather as a demonstration that the more active peptide, 3F-2, has a much larger effect on the packing properties of the hydrocarbon region of the membrane, compared with the less active 3F 14 . The effects are consistent with the proposed model for the interaction of these peptides with bilayers based on molecular shape (Fig. 14). In this model the 3F-2 peptide would introduce greater destabilization of the bilayer as a consequence of its cylindrical shape, short length, and aromatic side chains pushing the peptide axis up closer to the polar-nonpolar interface. Accumulation of the probe in the bilayer defect would dilute the PC 3 P and decrease intermolecular excimer formation and possibly also disfavor the conformation required for pyrene dimerization in the monomer. This is also consistent with the greater anti-atherogenic properties of 3F-2, because decreased membrane order has been suggested to be associated with increased risk for cardiovascular disease (20).
In summary, given the marked differences in biological activity between 3F-2 and 3F 14 , the nature of the interaction of these peptides with SOPC with or without cholesterol is remarkably similar. Nevertheless there are important differences between the two peptides that support our model of the difference in peptide "shape" (5). Although 3F-2 is less hydrophobic than 3F 14 by the criteria of HPLC volume and monolayer exclusion pressure (5), 3F-2 has stronger NOESY crosspeaks with protons more in the center of the bilayer (Fig. 10). We suggest that this is a consequence of a greater disordering of the bilayer caused by this peptide. 3F-2 also has two features that are more in common with those of 4F (7). These are a preferential broadening of the chain melting transition in mixtures of SOPC and cholesterol (Fig. 2) and a more cooperative unfolding transition of the peptide. The latter suggests oligomerization that may also contribute to a greater disruption of the bilayer order by insertion of a larger peptide aggregate. The disordering of the lipid could allow for the transfer of a Chemical shift differences are in parts per million between that of lipid alone and in the presence of 10 mol% peptide. A positive charge corresponds to a shift in the resonance to a lower frequency caused by the peptide.
FIG. 14. Biological activity of a Class A amphipathic helix depends on hydrophobic face-lipid acyl chain interaction. Top panel, a minimal effect on lipid acyl chain packing occurs in the wedgeshaped molecule 3F 14 . Bottom panel, the cylindrical shaped peptide, 3F-2, causes greater acyl chain perturbations, facilitating the entry of molecules such as water and lipid hydroperoxides into the hydrophobic milieu of the complex. oxidized lipids from the LDL surface to peptide-containing particles, thus rendering LDL less effective in inducing monocyte chemotaxis, an important step in the initiation of atherogenesis.