Nov 25, 2015 - A, using a Lauda Film balance. Thirty to sixty micrograms of peptide dissolved at a concentration of 150 pg/ml in the same buffer as the.
THEJOURNAL
OF BIOLOGICAL CHEMISTRY Vol. 255, No. 22, Issue of November 25, pp. 10651-10657,1980 Printed in U.S.A.
Chain Length-Function Correlationof Amphiphilic Peptides SYNTHESIS AND SURFACE PROPERTIES OF A TETRATETRACONTAPEPTIDE SEGMENT OF APOLIPOPROTEIN A-I* (Received for publication, May 1, 1980, and in revised form, June 18, 1980)
Daikichi Fukushima, Shinji Yokoyama, Daniel J. Kroon, FerencJ. Kezdy, and Emil Thomas Kaiser$ From the Departments of Chemistry and Biochemistry, The University of Chicago, Chicago, Illinois 60637
The segment correspondingto residues 121 to 164 of human plasma apolipoprotein A-I (apo A-I) has been synthesized by the Merrifield solid phase method. The peptide binds to unilamellar phospholipid vesicles and to phospholipid-cholesterol mixed vesicles. The surface affinity of the peptide measured in this way indicated that the mechanism of binding is the same as that of apo A-I (144-165)and apo A-I itself. The peptide appears to be a globular monomer in aqueous solution, with 17% a helix content. The peptide bound to vesicles activates 1ecithin:cholesterol acyltransferase: compared to apo A-I, t h e peptide is about 30% as efficient in the activation of cholesterol esterification and of phospholipid hydrolysis when the surface is saturated b y the activator. For a variety of amphiphilicpeptides and for apo A-I, the 1ecithin:cholesterol acyltransferase-activating ability correlateswell with their a helix contents in 50%trifluoroethanol.
The major physiological role of apolipoproteins appears to be to stabilize the surface of small, phospholipid-emulsified lipid particles and to provide a signal for the biochemical processing of these particles (1). Apoliprotein A-I (apo A-I),’ the majorproteincomponent of circulating high density plasma lipoproteins (HDL), is particularly interesting from the point ofviewof functional adaptation, since it adsorbs selectively to the surface of the smallest of the plasma lipoproteins and it activates 1ecithin:cholesterol acyltransferase (2), the key enzyme in cholesterol transport in the circulation. The apo A-I molecule is composed of segments of high internal homology (3) and these segments of 22 amino acids appear to have a high amphiphilic helix-forming potential (4, 5). It has been suggested that the length and cross-section of these amphiphilic helices render them particularly apt to be intercalated between the phospholipid polar head groups on the surface of particles with a small radius of curvature (6). Our previous studies with segments of apo A-I (7) and with their synthetic analogue (8) showed indeed that the affinity of apo A-I for curved phospholipid surfaces is closely approximated * This research was supported by United States Public Health Service Program Project HL-18577, by United States Public Health Service HL-15062 (SCOR),and by United States Public Health Service Cardiovascular Pathophysiology and Biochemistry Traineeship 57’32 HL 07237 (D. J. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article musttherefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. To whom correspondence should be addressed. The abbreviations used are: apo A-I, human plasma apolipoprotein A-I; HDL, human plasma high density lipoproteins; Peptide I, amphiphilic docosapeptide defined in Fig. 1; MOPS, 3-(4-morpholino)propanesulfonic acid.
+ ’
by these short peptides andthat thesurface affinity of apo AI is predominantly the result of summing of the activity of the segments with high amphiphilic helix-forming potential. More recently, we observed that the activation of the lecithin:cholesterol acyltransferase-catalyzed reaction by apo AI occurs through the activation of the phospholipid monolayer by the surface-bound apoprotein (9). Preliminaryresults showed, however, that in this activation the intact apolipoprotein activates the enzyme to ahigher extent than one of its short docosapeptide segments. Activation of 1ecithin:cholesterol acyltransferase occurred, on the other hand, by a synthetic docosapeptide, Peptide I, (Fig. 1) designed solely for optimal amphiphilic helix-forming ability (9). The latterpeptide also showed a higher affinity for phospholipid surfaces than did the apo A-I segment, apo A-I (144-165)’ (7,8). Thus, optimal 1ecithin:cholesterol acyltransferase activation still could be due either to high surface activity, as expressed by the dissociation constant of the surface-bound peptide, or to the larger size and, perhaps, intramolecular cooperative interactions in apo A-I. In orderto decide between these two possibilities we synthesized a tetratetracontapeptide segment of apo A-I (apo A-I (121-164)), containing the docosapeptide segment linked to anothersegment of high amphiphilic helixforming potential (6). For historical reasons, the docosapeptide, apo A-I (144-165), has been synthesized with Glu at positions 146 and 147 (7). These residues of apo A-I have been shown to be Gln (10) rather than Glu, and the tetratetracontapeptide has been prepared according to the latest amino acid sequence. We feel, however, that comparison of the surface properties of the Glu-containing docosapeptide and the Gln-containing tetratetracontapeptide is valid since the substitutions occur only in the hydrophilic face, whereas the hydrophobic face which interacts with the vesicle remains unchanged. Changing the amino acid side chain from Gln to Glu does not alter much the overall helix-forming potential of the docosapeptide. On the other hand, the introduction of two additional negative charges might contribute significantly to the surface behavior of the peptide, and for this reason, the surface properties of apo A-I (144-165) were also measured at lower pH values where the carboxylates are at least partially protonated. The measurement of the surface activity and 1ecithin:cholesterol acyltransferase activation of these peptides could then shed light on the role of possible intersegment interactions in the mechanism of surface binding and activation of the enzyme. In the present paper we wish to describe the synthesis and Peptide fragments of apo A-I shall he designated by the abbreviation apo A-I (x - y) where x and y are the residue numbers of the NH2-terminal and COOH-terminal residues of the peptide, respectively, corresponding to their position in the amino acid sequence as described by Brewer et al. (10).
10651
10652 ApoA-I(144-165)
Segments
Apolipoprotein
Leu-Gly-Glu-Glu-Met-Arg-Asp-Arg-Ala-Arg-AlaHis-Val-Asp-Ala-Leu-Arg-Thr-His-Leu-Ala-Pro
ApoA-IflZ1-164)
Pro-Leu-Arg-Ala-Glu-Leu-Gln-Glu-Gly-Ala-Arg-
Gln-Lys-Leu-His-Glu-Leu-Gln-Gln-Lys-Leu-SerPro-Leu-Gly-Gln-Gln-Met-Arg-Asp-Arg-Ala-ArgAla-His-Val-Asp-Ala-Leu-Arg-Thr-His-Leu-Ala Peptide I
Pro-Lys-Leu-Glu-Glu-Leu-Lys-Glu-Lys-Leu-LysGlu-Leu-Leu-Glu-Lys-Leu-Lys-Glu-Lys-Leu-Ala
FIG. 1. Primary sequence of synthetic peptide fragments of apo A-I and of Peptide I. At the time when apo A-I (144-165) was prepared, only the original sequence reported by Baker et al. (21) was available to us. The residues now known to be Gln a t positions 146 and 147 (10) were thought to be Glu, and for that reason apo A-I (144-165) was prepared with Glu at thecorresponding positions.
purification of the 121-164 fragment of apo A-I. With the help of the physiochemical characterization of this peptide we hope to show that cooperativity is not a major factor in the surface affinity of apo A-I fragments, and that their 1ecithin:cholesterol acyltransferase-activating ability is governed by the intrinsic surface a f f i t y of the peptides. EXPERIMENTAL PROCEDURES3
Peptide Synthesis and Purification-The synthesis and puriiication of Peptide I and apo A-I(144-165) were described in earlier publications (9, 11). The solid phase synthesis, purification, amino acid analysis, and sequencing of apo A-I (121-164) are described in the Miniprint section following this paper. Circular Dichroism-The circular dichroic spectra of solutions of apo A-I (121-164) and apo A-I itself were measured using a Cary 60 spectropolarimeter. The buffers were 0.01 M acetate at pH 4.0, 0.01 M phosphate a t pH 7.0, and 0.01 M Tris-HC1 buffer a t pH 9.0. A mixture of trifluoroethanol with 0.01 M phosphate buffer, pH 7.0, was also used to induce the peptides to form a-helical structures. The fractions of helical structure of the peptides described in the present report were estimated from their (8)222 ,,, values (deg cm2/dmol) using the equation: fraction a helix = ([~Izzz, + 3000)/(36,000 + 3,000) (22). Gel Permeation Chromatography-A Sephadex G-50 column (1.4 X 100 cm) calibrated with globular proteins was employed for determination of the apparent molecular weight of apo A-I (121-164) in 0.16 M KC1 solution. Binding of the Peptide to Unilamellar Vesicles and Assay of Lecithin:Cholesterol Acyltransferase Activity-Egg yolk lecithin was purchased from Avanti Chemical Co. and cholesterol was from Supelco Co., Ltd. 2-[9,10-3H]Oleoyl lecithin was synthesized from [9,103H]oleicacid (New England Nuclear) and egg lysolecithin, employing rat iiver microsomes (13). [4-'4C]Cholesterol was purchased from Amersham, England. The chemical or radiochemical purity of the materials were shown to be at least 99% pure by thin layer chromatography on Silica GelG, using either the solvent system chloroform: methano1:water (70305, v/v/v) or hexane:diethylether:acetic acid (65:35:1, v/v/v). Radioactivity was measured using a O d u o r - t o l uene-Triton X-100 mixture andan Isocap 300, Nuclear Chicago scintillation counter (9). Unilamellar vesicles of pure lecithin and a mixture of lecithin and cholesterol (4:l molar ratio) were prepared by the method of Korn and Batzri (9, 14). The peptides were incubated with the vesicles under the following conditions: 7.3 to 9.1 X M lecithin, 0.6 to 6 The synthesis of peptide apo A-I (121-164) and its purification (including additional tables, figures, and references) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass.Full size photocopies are available from the Journal of Biological Chemistry, 9550 Rockville Pike, Bethesda, Md. 20014. Request Document No.8OM-873, cite authors, and include a check or money order for $1.20 per set of photocopies. Full sized photocopies are also included in the micro!ilm edition of the Journal that is available from Waverly Press.
M peptide; the mixture was incubated for 2 h at 22°C in 0.025 MOPS (pH 7.0 to 7.3) or 0.02 M citrate (pH 5.2) containing 0.16 M KCl. The binding assay used the ultrafiltration technique described previously (8,9). Thepeptide-vesicle solution was subjected to ultrafiltration (100 A pore Amicon XM 100 membrane), three 100-pl volumes being collected for analysis. The following reagents were added to each sample: 50 pl of 0.16 N KC1,250 pl ofisopropyl alcohol, 500 pl of 0.4 M borate buffer (pH 9.0) and 50 pl of an acetone solution of fluorescamine (2.5 mg/ml, Roche, NJ). The fluorescence intensity was measured using an Aminco-Bowman Spectrofluorheter a t &x = 390 nm, A=,, = 475 nm (12, 15). Lecithin:cholesterol acyltransferase was partially purified from frozen stored human citrated plasma, according to the procedure described elsewhere (9). The final, 1600-fold purified, enzyme preparation was free from apo A-I. Initial rates of reactions catalyzed by lecithhcholesterol acyltransferase were measured by monitoring either the esterifcation of ['4C]cholesterolor [3H]oleoyltransfer from 2-[9,10-3H]oleoyl lecithin to cholesterol. The substrate wasmixed lecithin-cholesterol unilamellar vesicles (molar ratio, 4:l). The activation of phospholipase Az activity of 1ecithin:cholesterolacyltransferase was determined as described elsewhere (9). Monolayers of the Peptides a t the Air- Water Interface-Water was distilled in an all glass apparatus. The aqueous phase for the monolayer studies was either 0.02 M MOPS, 0.16 M KC1 or 0.01 M Tris-HCI, 0.1 M NaCl, pH 7.4 buffer which had been foamed with air for 10 min in order to remove the possible surfactant contaminants. The surface pressure, II (dyn/cm), of the monomolecular layer of the peptides at theair-water interface was measured as a function of area, A, using a Lauda Film balance. Thirty tosixty micrograms of peptide dissolved at a concentration of 150 pg/ml in the same buffer as the subphase was spread on the surface of the buffer, using an all-glass syringe and a partially immersed glass plate. The monolayer formed from the peptide was compressed and expanded between 700 cm2and 200 cm2 at a rate of 2.2 cm2/s. X
M
RESULTS
The peptide apo A-I (121-164) is readily soluble in water, and it can be stored in neutral aqueous solution for several weeks without decomposition. Gel permeation chromatography using a Sephadex G-50 column (1.4 X 100 cm) indicated an apparent molecular weight, 6600,close to the expected molecular weight of the monomer, 5069. Gel filtration did not indicate any self-association even when the peptide was 3.3 x 10-~ M. The helicity of apo A-I (121-164) was estimated from CD measurements. The CD spectra of aqueous solutions of apo A-I (121-164) were pH-independent over the range pH 4 to 9 (0.01 M acetate, phosphate, or Tris-HC1 buffer). The a helix content of apo A-I (121-164) was calculated to be 17% from the mean residue ellipticity at 222 nm: 3700 f 400 deg cm2/ dmol, observed overthe pH range of our experiments (Fig. 2). The helix content, which indicated that the peptide has predominantly random structure in aqueous solution, was independent of the peptide concentration over the range of 2.65 x M to 1.08 x lop4 M, in contrast tothe case of the amphiphilic a-helical synthetic Peptide I. The latter showed an increase in the ellipticity measured at 222 nm with increasing peptide concentration, thereby indicating reversible tetramerization of the peptide (9).The mean residue ellipticity and the calculated per cent a helicity of apo A-I (144-165) were described in an earlier communication (7): the a helicities in aqueous solution were calculated to be -15% over the range pH 4.9 to 9.0 showing no concentration dependency. In the neutral phosphate buffer solution the a helicity increased to 35% and 40% in the presence of added trifluoroethanol, 25% and 50% (v/v), respectively. In the presence of 50%trifluoroethanol in neutral phosphate buffer the a helicity of apo A-I (121-164), calculated from the mean residue ellipticity of 222 nm, 19,900 deg cm2/dmol, was 60%. The CD spectrum of a solution of apo A-1 itself (1.2 X M ) in 50%trifluoroethanol, pH 7.3, buffer was also measured. The mean residue ellipticity at 222 nm was calculated
Apolkoprotein Segments
8 dyn/cm for apo A-I (144-165).In comparison, monolayers of apo A-I have a collapse pressure of 22 dyn/cm (21).Analysis of the II-A curve at low surface pressures for a weighed amount of peptide gave a molecular weightof 3100 for apo AI (144-165)and 5400 for apo A-I (121-164).Thus, both peptides form monomeric monolayers. The molecular areas expressed in terms of the area/amino acid (Aolx,where x is the number of amino acids per peptide) were 20.8 A2/arnino acid for apo A-I (144-165) and 18.8 A2/amino acidfor apo A-I (121-164).Analysis of the curve measured at higher surface pressures using the empirical equation also gave the monomeric molecular weight for apo A-I (121-164)(5300)and the same area per amino acid (18.8A2/amino acid). The instability of the monolayer of apo A-I (144-165)at the higher surface pressure made a comparable analysis for this peptide somewhat uncertain. The concentration dependency of the binding of apo A-I (144-165)and apo A-I (121-164)to lecithin and lecithin-cholesterol unilamellar vesicles is shown in Fig. 3. The results were analyzed in terms of the Langmuir isotherm according to the equation Pf = ( N -PC.P,IPb) - K d , where Pfand P b are the concentration of free and bound peptide, respectively, PC is the concentration of lecithin. N is the upper limit of Pf/PC, and Kd is the dissociation constant (16).Plots of Pf versus PC. Pf/Pt,are shown in Fig.4. Straight lines were obtained in each case, showingthat the binding of apo A-I (144-165)and
0
c
I-
O
E 0
cy
-1
5 cn 0
U
? -
I
-
0
\
cn
\
X
i
/
-2
2
10653
2 2 02 1 0 230 WAVE LENGTH, n m
3
FIG.2. Circular dichroism spectra of synthetic fragments of apo A-I, 22°C. Solid line, apo A-I (121-164) in 0.01 M phosphate, pH 7.0; dotted line, apo A-I (144-165) in the same buffer; dashed line, apo A-I (121-164) in 50% v/v trifluoroethanol, 0.01 M phosphate buffer,pH7.0; dotted-dashed line, apo A-I (144-165) in the same solution.
2
1
to be 25,295 deg cm2/dmol, yielding an a helicity of 73%. An attempt was made to measure the CD spectra of the peptides in neutral buffer in the presence of lecithin-cholesterol mixed vesicles. However, two experimental problems were encountered: low S/N in the measurement of the ellipticities due to light scattering from the vesicles and the relatively weak binding of the peptides to thevesicles (see later) which made the apparent change in the ellipticity small, even though the change in a helicity of the peptide on binding might be quite substantial. Consequently, it was not feasible for us to obtain significant values from the CD spectral measurements on the peptide vesicle complexes. Additionally, it should be noted that even if it had proven experimentally feasible to obtain reliable ellipticities for the peptide -vesiclecomplexes, a sound theoretical basis for interpreting such values measured in the presence of chiral phospholipid molecules in terms of the per cent a helicity of the peptides is not available. The monomolecular layer of apo A-I (121-164)was stable for at least 30 min, over a surface pressure (II)range of 0 10 dynes/cm. On the other hand, a monolayer of apo A-I (144164) disappeared rapidly from the surface especially when at II > 2 dynes/cm. The force-area curves can be analyzed at low surface pressure (1dyne/cm > II)using the equation II(A - do) = nkT where A. is the area excluded by the molecules and n is the number of molecules. The empirical equation II(A - do (1 - KII)) = nkT (K is a constant) proved to describe adequately the force-area curve for higher surface pressure range (n 2 1 dyne/cm). The II-A curves were the same for MOPS buffer as for Tris-HC1 bufferin the subphase. The stability of the peptide monolayers showed a sharp discontinuity at a certain high surface pressure and this “collapse pressure” was characteristic for each peptide. We found the collapse pressure to be 12 dyn/cm for apo A-I (121-164)and
0
le
t
4
-
-
0
FIG.3. Isotherms for the adsorption of synthetic apolipoprotein fragments to unilamellar vesicles, 22°C. A , binding of apo A-I (121-164) to egg lecithin vesicles (9.1 X M lecithin) in 0.025 M MOPS, 0.16 M KCl, pH 7.1(0); in 0.02 M citrate, 0.16 M KCl, pH 5.2
(0). B , binding to mixed lecithin/cholesterol (4:lmol/mol) unilamellar vesicles (7.9 X lo-‘ M lecithin). Apo A-I (121-164) in 0.025 M MOPS, 0.16 M KCl, pH 7.3 (e),in 0.02 M citrate, 0.16 M KCl, pH 5.2 (0). Apo A-I (144-165) in 0.02 M citrate, 0.16 M KCl,pH 5.2 (A). Binding was quantitated by the ultrafiltration method described under “Experimental Procedures.” Solid lines were calculated using the equation described in the text and the parameters from Table I.
10654
Apolipoprotein
0.2
0.1
0.c
-0.1
-,
0.2
a’ 0.1
0.c
- 0.1
0 FIG. 4. Plot Pf = (N~PC*PfIP6) of Fig. 3.
2 ( Pf/P,,)rPC,
4
6 g/l
of the data from Fig. 3 according to the equation - Ka. For symbols and conditions see the legend TABLE
Segments to apo A-I (121-164). If this were the case, then the Kd of the docosapeptide at low pH would be the one to be compared to that of the tetratetracontapeptide. In contrast, apo A-I (121164) showed little pH dependency for its binding affinity toward either lecithin or lecithin-cholesterol mixed vesicles. The effect of the incorporation of cholesterol into the vesicles on the affinity for the binding of apo A-I (121-164) was also rather small and only apparent at low pH. At pH 5.2 the dissociation constant increased by 44% upon incorporation of 20 mole % cholesterol into the vesicles. On the other hand, the capacity of the binding (N) increased approximately 2fold. The N values expressed in terms of weight of peptide/ weight of phosphatidylcholine are comparable for apo A-I (144-165) and for apo A-I (121-X4), and also for the model synthetic amphiphilic e-helical Peptide I (Table I) (9). The activities of human 1ecithin:cholesterol acyltransferase were measured in the presence of either apo A-I (144-165) or apo A-I (121-164), and the results were compared to those obtained previously with apo A-I and Peptide I (9). The rate of transesterification catalyzed by 1ecithin:cholesterol acyltransferase was measured by [‘4C]cholesteryl ester formation; [“Hloleate transfer from lecithin to cholesterol (Fig. 5) was measured by determining the rate of cholesteryl r3H]oleate formation in lecithin-cholesterol mixed vesicles (20 mol % of cholesterol). The [“Hloleic acid release in lecithin vesicles (Fig. 6) measured the phospholipase AZ activity of lecithin: cholesterol acyltransferase. The initial rates of all these reactions are displayed in the figures as functions of the concentration of the various peptides bound to the vesicles. Concentrations of surface-bound peptides were calculated from P, the total peptide concentration, by solving the second order equation P& = (P - 9) (N-PC - Pt,), using the parameters
I
Parameters of binding of synthetic apo A-I fragments to unilamellar vesicles of lecithin and of lecithin-cholesterol, according to the equation, Pf = (N. PC. PfIPb) - Kd Parameters were determined by linear regression analysis of the plots shown in Fig. 4. Mole B Peptide of chopH I&x 10’ &x lo* N x IO3 N x 10’ lesterol mol g Peptide/g
Apo A-I (144-165)
Apo A-I (121-164)
0 0 20 20 0 0 20 20
6.0 7.4 5.2 7.3 5.2 7.1 5.2 7.3
6.9 97 49 >300 5.7 8.4 8.2 8.2
1.75 24.4 12.2 >76 2.87 4.24 4.13 4.16
ithin
lecithin
10.6 10.6 17.8
3.45 3.45 5.77
4.17 5.44 9.45 8.87
2.73 3.56 6.18 5.80
of apo A-I (121-164) to the vesicles obeys a single equilibrium. The value of N can be calculated from the slope and that of Kd from the ordinate intercept. The parameters thus measured are shown in Table I. As described earlier (17), the binding affinity of apo A-I (144-165) to lecithin vesicles showed a marked pH dependency, requiring the protonated form of a functional group with a pK value of 6.2. The binding of apo A-I (144-165) to mixed vesicles also shows the same marked pH dependency, and the affinity of the peptide toward the mixed vesicles decreased at all pH values by a factor of 7 relative to that toward lecithin vesicles, together with an increase of 67% in the limit of binding (N). The pH dependency of Kd might be due, in fact, partially to the presence of two additional caboxylates in the docosapeptide with respect
FIG. 5 (left). Activation of 1ecithin:cholesterol acyltransferase as a function of the percentage of maximal binding of synthetic peptides bound to the surface of mixed lecithincholesterol unilamellar vesicles. Open circles, rate of [‘4C]cholesterol ester formation activated by apo A-I (121-164). Dashed line shows the activation of the same reaction by apo A-I according to Ref. 9. Closed circles, activation of [3H]oleoyl transfer from lecithin to cholesterol by apo A-I (121-164). Dotted line shows the rate of the same reaction activated by apo A-l (9). Open triangles represent the rates of activation of [‘%]cholesterol ester&cation in the presence of apo A-I (144-165). Substrate concentrations were 3.6 X 10m4 M lecithin and 9 x 10e5 M cholesterol in 0.02 M MOPS, 0.16 M KCl, pH 7.4. Solid lines represent the tit of the data by linear regression analysis. FIG. 6 (right). Activation of 1ecithin:cholesterol acyltransferase catalyzed phospholipid hydrolysis as a function of the percentage of maximal amount of synthetic peptides bound to the surface of lecithin unilamellar vesicles. Open circles, activation by apo A-I (121-164); open triangles, activation by apo A-I (144-165). Dashed line shows activation by apo A-I as reported in Ref. 9. Substrate concentration was 3.7 X lo-’ M in 0.02 M MOPS, 0.16 M KC1 buffer, pH 7.4. Solid lines represent the fit of the data by linear regression analysis.
Segments
Apolipoprotein
10655
Under these conditions, a random coil peptide would have the rate on increasing peptide surface concentration was ob- displayed an apparentmolecular weight much larger than the true molecular weight of the monomer (18).The intramolecserved, whereas the plots of rate uersus concentrationin solution did not yield any linear correlation. Furthermore, the ular interactions would also explain why the tetratetracontarate of the 1ecithin:cholesterol acyltransferase-catalyzed re- peptide remains monomeric in solution, whereas the smaller action increased in all cases up to the maximal binding capac- docosapeptide Peptide I of somewhat similar composition, shows a high tendency to tetramerize. In the light of these ity, without any indication of a saturation effect. Using apo A-I (121-164) as theactivator, both themaximal considerations the experimental dissociation constant, K d , of activation of the esterification of ['4C]cholesterol in mixed the peptide vesicle complexis probably not atrue measure of vesicles and themaximal phospholipase Az activity inlecithin the surface affinity of the peptide. Similarly, the binding of vesicles were about 30% of those observed when apo A-I at a apo A-I to vesicles is also most likely governed to a large saturation level was used as theactivator. On the other hand, extent by the considerable energy required to disrupt the the maximal [3H]oleate transfer activated by apo A-I (121- tertiary and quarternary structure of the protein in aqueous 164) which measures side chain specificity was 8 times less solution (19). Because of the quite different tertiary structures of the than thatactivated by apo A-I itself. Near neutral pH,where 1ecithin:cholesterolacyltransferase various peptides in the aqueous phase, their experimentally has maximum activity, the affinity of apo A-I (144-165) to the measured helicity could not measure directly their intrinsic mixed lecithin-cholesterol vesicles is very low so that it was ability to form helices at the surface of the vesicle. On the 50% solution of trifluoroethanol is expected to experimentally impossible to decide whether apo A-I (144- other hand, the 165) could activate the enzyme-catalyzed ['4C]cholesterol es- optimize helix formation and also perhaps to disrupt tertiary terification and/or [3H]oleate transfer (Fig. 5). Since apo A-I and quaternary structures(20). Thus, the helicities measured (144-165) binds to lecithin vesicles to a measurable extent at in that solvent should reflect more closely the differences in neutral pH values, the activation of the phospholipase Az the intrinsic helix-formingability of the peptides. We observed activity of 1ecithin:cholesterol acyltransferase by this peptide that in that solvent the 44-amino acid peptide was 59%helical could be observed. We found that it is about one-third of that whereas the 22-amino acid fragment was only 40% helical. by apo A-I (121-164) when extrapolated to maximum activa- Thus, mere extension of the peptide chain by a segment of identical helix-forming potential resulted in a signifkant intion. crease of the stability of the a helix. It was then not surprising DISCUSSION to find that in 50% trifluoroethanol apo A-I was 73% helical. The synthesis of a tetratetracontapeptide fragment of apo We feel that the helicities obtained in 50% trifluoroethanol A-I allowed us to study the role of chain length inthe surface are more amenable to structure-function analysis than if one properties and 1ecithin:cholesterol acyltransferase activation were to measure the helicity of the peptides in the peptide. of amphiphilic peptides. Our measurements of the affinity and vesicle complex. The helicities measured in 50% trifluoroethanol correlate the stoichiometry of the binding of peptides to phospholipid indeed very well with the ability of the peptides to activate vesicles and to cholesterol-phospholipid mixed vesicles showed the mode of binding to be independent of chain length. 1ecithin:cholesterolacyltransferase. It was most interesting to Extending the peptide chain from 22 to 44 amino acids in- observe that the model docosapeptide, Peptide I, which was creases somewhat the affinity of the peptide toward the phos- not asegment of apo A-I had a 61%helicity in trifluoroethanol, pholipid surface, but thefree energy of the binding is far from the same as the tetratetracontapeptide, and itactivated lecibeing additive, being increased from 5.5 kcal/mol to only 7.0 thin:cholesterol acyltransferase exactly to the same extent as kcal/mol. Taking the K d of the docosapeptide at low pH, the the twice as large tetratetracontapeptide. difference is even smaller. This lack of additivity is further Since the 1ecithin:cholesterol acyltransferase activation is shown by the free energy of binding of the intact apo A-I directly correlatable only with the extent of helix formation in which only amounts to 8.4 kcal/mol. If the reason for non- 50% trifluoroethanol, and not with the apparent stability of additivity were the binding of only part of the longer peptide, the vesicle-peptide complex, nor with the chain lengthper se then thestoichiometry expressed in termsof grams of peptide or sequence of the peptides, we would like to conclude that per g of lecithin, should show an increase in the amount of the activation of phospholipid surfaces can only occur by peptide bound at saturation corresponding to the increase in surface-bound helical segments. It appears, however, that molecular weight. This is clearly not the case, since the 22- activation of the 1ecithin:cholesterol acyltransferase reaction amino acid fragment and 44-amino acid fragment show stoi- also requires aprecise depth of penetration of the peptide into chiometries which are identical on a gram per g of lecithin the surface monolayer. This in turn is governed by the ratio basis and not on a mole per g of lecithin basis. This samefact of the areas of the hydrophilic and lipophilic portions. Thus, also rules out thepossibility that thetwo peptides would have a predominantlyhydrophilic helix will only penetrate marginradically different conformations when bound to the surface. ally, probably not more than a random coil.On the other The most likely explanation of the rather poor energy of hand, if more than half of the helix ishydrophobic, then deep binding of the larger peptide is that the conformations of the penetration will result in the disruption of organized monotwo peptides are different when dissolved in water. Since the layers. The optimal depth of penetration for enzyme activatetratetracontapeptide remains monomeric in solution as tion should occur when one-third of the surface of the helix is shown by gel fitration, we can rule out self-association as the hydrophobic, as is approximately the case for the peptides cause of the stabilization of the hydrated form. Thus, we studied. The marginal activation of lecithin-cholesterol acylwould like to propose that intramolecular, presumably hydro- transferase by apo A-I1 (23) could thus be explained by too phobic, interactions provide some tertiary structure for the deep a penetration into the surface layer, since the 6-31 and larger peptide, and that this structure has to be disrupted 52-73 segments would form a helices with at least half of the before binding to the surface of the vesicles can occur. Evi- surface being hydrophobic. Such helices would, of course, bind dence for such structure is provided by the gel permeation very strongly to the surface, thereby accounting for the inhichromatography, which indicated amolecular weight of about bition by apo A-I1 of the apo A-I-activatedenzymatic reaction 6,600 when the column was calibrated with globular proteins. (23).
N and Kd from Table I. In all cases, a linear dependency of
-
10656
Segments
Apolipoprotein
It appears then that although the physiological and physicochemical role of apolipoprotein A-I is fulfied by short segments of the apoprotein, the linking together of these segments enhances their ability to activate 1ecithin:cholesterol acyltransferase. Cooperative helix formation and cooperative binding to the surface seem to be the mechanisms conducive to increased efficiency, although the cooperative binding is almost totally counterbalanced by an increased stability of the tertiary structureof the peptide in water. Segmental helix formation is a feature of apo A-I necessitated by the high curvature of the HDL surface, and this segmentation necessarily limits the cooperativity between helices. It isthen reasonable to conclude that thephysiological action of apo AI is expressed by a multiplicity of weakly cooperative segments, rather than by a single active site. Acknowledgments-We thank Dr. R. H. Heinrikson and Ms. P. Keim for the sequence analysis of the peptide and Dr. A. M. Scanu for providing samples of apo A-I. REFERENCES 1. Smith, L. C., Pownall, H. J., and Gotto, A. M., Jr. (1978) Annu. Reu. Biochem. 47, 751-777 2. Glomset, J. A. (1979) in The Biochemistry of Atherosclerosis (Scanu, A. N., ed) pp. 247-274, Marcel Dekker Inc., New York 3. Fitch, W . M. (1977) Genetics 86, 623-644 4. Segrest, J . P., Jackson, R. L., Morrisett, J. D., and Gotto, A. M., dr. (1974) FEBS Lett 38, 247-253 5. Chou, P. Y., and Fasman, G. D. (1974) Biochemistry 13,211-222 6 . Edelstein, C., Kezdy, F. J., Scanu, A, M., and Shen, B. W . (1979) J. Lipid Res 20, 143-153 7. Kroon, D. J., Kupferberg, J. P., Kaiser, E. T., and Kezdy, F. J. (1978) J.Am. Chem. SOC.100,5975-5977 8. Fukushima, D., Kupferberg, J. P., Yokoyama, S., Kroon, D. J.,
FXDFRIMEVTN PROCEDLWES
Kaiser, E. T., and Kezdy, F. J. (1979) J. Am. Chem. SOC.101, 3703-3704 9. Yokoyama, S., Fukushima, D., Kupferberg, J . P., Kezdy, F. J., and Kaiser, E. T.(1980) J.Biol. Chem 255, 7333-7339 10. Brewer, H. B., Jr., Fairwell, T., LaRue, A., Roman, R., Houser, A,, and Bronzert, J. J. (1978) Biochem. Biophys. Res. Commun. 80,623-630 11. Kroon, D. J., and Kaiser, E. T. (1978) J. Org. Chem. 43, 21072113 12. Brand, L., and Gohlke, J. R. (1972) Annu. Rev. Biochem. 41,843868 13. Lards, W . E. M. (1960) J. Biof.Chem. 235,2233-2237 14. Batzri, S., and Korn, E. D. (1973) Biochim. Biophys. Acta 298, 1015-1019 15. Udenfriend, S., Stein, S., Bohlen,P., and Dairman, W . (1972) Proceedings of the Third American Peptide Symposium, pp. 655-663, Ann Arbor Science Publishers, Ann Arbor, Mich. 16. Adamson, A. W . (1967) Physical Chemistry of Surfaces, pp. 569 ff., Interscience Publishers, New York 17. Fukushima, D., Kaiser, E. T., Kezdy, F. J., Kroon, D. J., Kupferberg, J. P., and Yokoyama, S. (1980) Ann. N . Y.Acad. Sci. 348, 365-373 18. Nozaki, Y., Schechter, N. M., Reynolds, J. A., and Tanford, C. (1976) Biochemistry 15, 3884-3890 19. Vitello, L. B., and Scanu, A. M. (1976) J. Biol. Chem. 261. 11311136 20. Lotan, N., Berger, A., and Katchalski, E. (1972) Annu. Rev. Biochem. 41,869-902 21. Kupferberg, J . P., Kroon, D. J., Yokoyama, S., Fukushima, D., Shen, B. W., and Kaiser, E. T. (1980) Proceedings of the Sixth American Peptide Symposium, pp. 519-522, Pierce Chemical Co., Rockford, Ill. 22. Morrisett, J. D., David, J. S. K., Pownall, H. J., and Gotto, A. M., Jr. (1973) Biochemistry 12, 1290-1299 23. Albers, J . J . (1979) Scand. J. Clin. Lab. Invest. 38, Suppl. 150, 48-52 Additional references are found below.
Apolipoprotein Segments
10657 0.8 c
06-
g
36
04-
04 I
1
--
A
N
