Study of the lipid binding characteristics of the apolipoproteins from human high density lipoprotein

57

Biochimica et Biophysics Acta, 441 (1976) 57-67 @IElsevier Scientific Publishing Company, Amsterdam

- Printed in The Netherlands

BBA 56816

STUDY OF THE LIPID BINDING CHARACTERISTICS APOLIPOPROTEINS FROM HUMAN HIGH DENSITY I. ELECTRON MICROSCOPIC AND GEL FILTRATION SYNTHETIC PHOSPHATIDYLCHOLINES

GERT MIDDELHOFF *, MARYVONNE and W. VIRGIL BROWN

ROSSENEU

Department of Medicine, School of Medicine, Calif. 92093 (U.S.A.) (Received

OF THE LIPOPROTEIN STUDIES WITH

**, HUB. PEETERS

**

University of California, San Diego, La Jolla,

February 4th, 1976)

Summary The characteristics of the lipid - protein complex produced by the addition of the major apolipoproteins (apo AI and apo AII) of human high-density lipoprotein to synthetic phospholipids has been studied. Under the in vitro conditions utilized, apo AI binds to 1,2dimyristoyl-sn-glycerophosphocholine and 1,2dipalmitoyl-sn-glycerophosphocholine liposomes, but does not alter their morphologic characteristics. This binding occurs at temperatures above or below that of the transition (T,) of the lipid bilayer. In contrast, apo AI1 spontaneously generates small, homogeneous disc-shaped lipid-protein complexes (50 X 100 A) from large phospholipid globules or from liposomes prepared with these lipids. This type of complex was only formed when the lipid/ape AI1 mixtures were warmed above the transition temperatures. The incorporation of apo AI into this small complex with apo AI1 may be greatly facilitated or inhibited depending on the sequence of addition of the various components. Under optimal circumstances, a maximum of 1 molecule of apo AI is incorpo-

Abbreviations: ifDL: High density lipoprotein (human. 1.063 < d < 1.21). APO AI: The major a~olipoprotein constituent of human HDL. This protein has a single peptide chain of 28 000 d&one. with carboxyl terminal dutamine and untno terminal uputtc acid. APO AII: an lpolipo~mtein of HDL which is composed of two identical polypeptide chains each of 8700 daltons connectud by a disulflde bond at cystetne in the sixth positton. The cubox~l kminal amtno acid is glutamine and the amino terminal is pnrolidone carboxylic acid. DMPC: 13dimyristoyl~n-alycerophosphocholine. DPPC: 1,2dipalmitoyl-mOycerophosphocholine. Tt: tram&ion temperature of the lipid bilayer. l Medizinische Univerdtiib Kliaik. 6900 Heidelberg. Bergbeimer Strasr+e68. West Germany. ** Simon Stevin Instituut. Jeruzalemstraat 34. Brugge. Belgium.

rated with each molecule phospholipids.

of apo AI1 into complexes

with these two synthetic

Introduction The lipid components of the high-density lipoprotein (HDL) can be entirely removed from the protein by organic solvent extraction. If these lipid extracts are again added to the apolipoproteins in aqueous solutions under conditions of sonic irradiation, particles can be produced which are very similar to the native lipoprotein [ 11. It therefore appears that the basic information required for the determination of the particle size and shape is contained within the molecules of the lipoprotein. The present studies have been undertaken to gain more specific knowledge about the nature of this interaction of lipid with the protein components. The apolipoproteins of human HDL are composed of two major proteins (apo AI and apo AII) and a series of minor protein constituents [2,3]. These two major apolipoproteins have been studied extensively [4,5], and, although they differ in their primary structure [4,5], several common characteristics have been noted. The carboxyl terminal portion of the peptide chain appears to have a higher affinity for lipid [6,7], both apoproteins contain amphipathic helixes which are suggested to be the binding regions for phospholipid [4,8], and each inhibits the phospholipid-requiring enzyme /3-hydroxybutyrate dehydrogenase [ 6,7]. In previous recombination studies, the amount of apo AI which could be incorporated into phospholipid liposomes was much less than that of apo AI1 [g-l1 1. A large increase in the amount of apo AI incorporated was noted when equimolar mixtures of this apoprotein were preincubated with apo AI1 prior to introduction of the lipid [lo]. In the present study, the morphologic characteristics of the lipid * protein complex produced in aqueous media by incubation of apo AI and apo AI1 individually or in combination has been studied using pure synthetic phosphatidylcholine liposomes as ligand. It is clear from .these studies that both apolipoproteins interact with phospholipid bilayers, but that their interactions have quantitative and qualitative differences. Methods I. Preparation

of the apolipoproteins

Apo AI and apo AI1 were prepared from high density lipoprotein (d, 1.0631.21) isolated from the plasma of normal fasting donors by ultracentrifugation [12]. The delipidated proteins were isolated by a slight modification of the method previously reported using DEAE-cellulose chromatography [ 131. The purity of the preparations was monitored by polyacrylamide gel electrophoresis in 6 M urea at pH 8.9 [14], and by double immunodiffusion in 1% agarose using appropriate antisera [ 151. Liposomes were prepared from 1,2dimyristoyl-sn-glycerophosphocholine (DMPC) and 1,2dipalmitoyl-sn-glycerophosphocholine (DPPC) by sonic irradiation, using a Branson cell disrupter equipped with a microtip probe follow-

59

ing the method of Huang 1161. An energy setting of 60-70 W was used for a period of 1 min with intermittent cooling for a total irradiation time of up to 30 min. The sample was flushed with a nitrogen stream during sonication. Sufficient 12-I “C]dip~mitoyl-s~-glycerophospho~hol~e was added to provide a specific activity between 0.5 X 10m2and 2.0 X 10m2 Ci/mol of lipid. Radioactivity was determined by addition of up to 0.5 ml of the aqueous solutions (0.1 M NaCl, 0.01 M sodium barbital, pH 7.4) con~ning the liposomes to 10 ml of scintillation counting fluid. The latter was prepared by addition of Triton X-100 containing 5.5 g of PPO, 0.125 g of dimethyl POPOP per 1 to an equal volume of toluene. The counting efficiency for 14Cwas 92%.

II. Preparation of lipid ’ protein complexes The apolipoproteins were incubated with the liposomes at two temperatures, chosen to be respectively above and below the transition temperature (Tt ) of the lipid. These were 15 and 35°C for DMPC and 25 and 50°C for DPPC. After incubation for 5-20 mm the mixtures were routinely stored at room temperature for 12-16 h before separation of the lipid - protein complexes from the unbound components by gel filtration or ul~a~en~fu~tion. Gel filtration experiments utilized 6% agarose gel beads (BioGel A-15 M) packed in columns of 60 X 2 cm or 55 X 1.5 cm equilibrated with 0.1 M NaCl, 0.01 M sodium barbital and saturated in advance with the appropriate phospholipid. The columns were characterized by determination of the elution volume of Blue Dextran (Pharmacia Fine Chemicals), human LDL, HDL, aldolase (Abbott Labs) and bovine serum albumin (Per&x). Ultracentrifugal flotation was performed either in a Beckman preparative ultracentrifuge using an SW 50.1 rotor, operated at 210 000 X g for 48 h or in an Inflations B-60 ultracent~fuge using an A321 rotor at 220 000 X g for 24 h. The density of the solution was adjusted by the method of Have1 et al. [ 121, or by addition of crystals of NaCl/NaBr. Selected preparations were negatively stained with a 2% phosphotun~tic acid solution (pH 6.8) and examined on Formvar carbon-coated copper grids using a Phillips (Model 311) electron microscope. Protein determinations were performed following the procedure of Lowry et al. [ 181 using bovine serum albumin as standard. The lipid-containing samples were extracted with diethyl ether after color development [19]. The phospholipid determinations were done by the method of Bartlett [20] and, in certain experiments, by the method of Eibl and Lands [ 211. The transition temperatures of the liposomal preparations were determined by monitoring the go-degree light scattering, using a spectrofluorometer equipped with a thermopile (Science Products, Dover, New Jersey) which had an output of 0.1 mV per degree. The temperature of the sample and the light emission were directly recorded with an XY recorder. Results Liposomes prepared from DPPC were studied initially. When chromatography on 6% agarose was begun immediately (within 10 min) after their preparation by sonic irradiation, approximately half of the total lipid eluted in the

60

was i 1.25 Fig. 1. Electronmicrographs of liposomes prepared from DPPC. The phospholipid concentration mM. Preparation of samples was performed as described under Meth .ods. The phospho~pid dispersion R was (25°C) for 12 h (B) . The applied to tbe grid immediately after sonic&on (A) and aiso after incubation length of the white bar indicates 1000 A in all figures.

void volume and the remainder appeared as a sect bnd peak in the included volume. This second peak eluted slightly before hum an LDL and the mean pbarticle diameter was estimated to be approximately 250 A. When the lipost >mes were incubated at 25°C for 12-16 h, the absorba nce of the solution incre !ased

61

up to 50% and all phospholipid eluted in the void volume. These changes in the elution profiles of the liposomes were investigated by electron microscopy. The preparations examined i~ediately after sonication (Fig. 1A) were found to contain particles with a wide range of sizes, with the majority between 200 and 300 A in their greatest dimensions. During a 12-16 h incubation at 25”C,
I e E

=

.05

2.5

.04

2.0

.03

I.5

B

; I

z

g

Q :

.02

1.0

E

.OI

20

‘-%----

60

80

100

EFFLUENT

120

VOLUME

140

(ml1

I60

Ia

Fig. 2. Gel filtration experiment with mixture of DPPC and ape AI. The incubation of DPPC and apo AI (at a lipid/protein molar ratio of 50 : 1) was performed at 5O’C for 10 min. followed by 12 h at room temperature (24’0. The chromatography on the 6% agarose column (50 .X 2 cm) was performed at room temperature. Protein (e-) and phospholipid (o---- --0) concentrations in e&ted fractions are shown.

62

containing no added protein, appearing as discs and ellipsoidal figures of 250500 A in their greatest and 100-150 Ir( in their smallest dimensions. The ultracentrifuge experiments gave results compatible with those above. DMPC liposomes floating at density 1.08 contained no measurable protein after preincubation with apo AI. After adjustment of the infranatant to density 1.23, a complex was isolated by flotation with a lipid/protein ratio (m/m) of approximately 120. When very low ratios of lipid/protein were used in the original incubation mixture (i.e. less than 30 m/m), final ratios of 70-80 mol of DPPC per mol of protein were found. Binding experiments with apo AU. Incubation of apo AI1 with freshly prepared DPPC liposomes at 25°C for 12 h appeared to retard their aggregation and coalescence, as measured both by gel filtration (Fig. 3A) and by electron microscopy. However, inserting a brief (10 min) incubation period above the T, markedly altered the elution profile (Fig. 3B). A small lipid * protein complex was observed with an elution volume identical to that of human HDL. Essentially all added apo AI1 was found in this complex when the original incubation

.“i” : -02

-05

I -

4

1 ::

-001 -

z. E

d

I

z E

I

iba,

0 ‘+

:Ah A

Q.

, z 125 E

B

: -04

IO

--.03 -

- 0.75

-02

-

- 0.5

-.Ol

-

120

I40

I60

EFFLUENT VOLUME ( ml ) 3. Gel filtration experiments with mixtures of DPPC and ape AIL Conditions for incubation and column chromato$traphy WE identical to those described in Fig. 2. Protein and phospholipid concentrations are indicated by the aame symbols. The lipid/protein molar ratio was 20 : 1. In the experiment depicted in the upper panel (A) the sample wps incubated for 12 h at room temperature (24%. In the experbnent indicated in the lower panel (B). a lO-min incubation Period at 50°C WBS added. Fig.

63

mixture contained a molar ratio of lipid/protein above 30. Below a ratio of 30, the excess protein eluted as a separate peak. The isolated lipid * protein complex had a molar ratio of 20 + 7 in a series of six such experiments. Electron micrographs revealed ellipsoidal particles of 50-60 A in thickness and 80-100 a in diameter. These results were interpreted to be consistent with a single type of particle, a round disc of approximately 100 A in diameter. In some fields, they were stacked in rouleaux. This stacking phenomena was. felt to be related in part to the concent~tion of material applied to the grid. The origin of the small particles might have been due either to the stabilization by apo AI1 of very small liposomes generated during the sonic irradiation, or alternatively by direct interaction of the apolipoprotein with much larger regions of bilayer lipid. The second possibility seems more probable for the following reasons: electron microscopic studies of liposomes prepared within seconds of their exposure to ultra-sound revealed the diameter of the smallest particles to be 200 A or more. A turbid suspension, prepared by simply stirring the DPPC into the buffer before addition of the apolipoproteins, cleared immediately on warming above Tt in the presence of apo AIL. On gel filtration and by electron microscopy, a major portion of the lipid had been converted to 50 X 100 w discs, indistinguishable from those produced with the sonicated lipid. Thus, with no sonic i~adiation utilized at any phase of the experiment, this apolipoprotein produced a small homogeneous lipid * protein complex at temperatures above the transition of the lipid bilayer. Recombination with mixtures of apo AI and apo AII. The possibility of cooperativity between the two proteins in the binding process was considered. Initial experiments involved sequential additions to the liposomes of apo AI1 followed by apo AI. No increased incorporation of apo AI was observed after fractionation of the incubation mixture by gel filtration or by ultracentrifugation. If the two proteins were together in solution prior to the addition’of the lipid or if the apo AI was added to the liposomes before apo AII, a marked increase in apo AI binding was observed. In a typical experiment, Table I, a series of ultracent~fugation tubes confining increasing quantities of apo AI were incubated with a constant amount of apo AI1 at room temperature. Addition of DPPC to these mixtures at a constant molar ratio to apo AI1 of 50 : 1 was followed by an incubation period (10 min) at 50” C. Protein and phospholipid

TABLE I BINDING OF APO AI TO DPPC Ultracentrifugation (for details, see text) APO AI/AI1 incubation

APO AI (alone) 0.5 1.0 2.0

APO AI In infrenatatlt

APO AI Bound

m)

(5%) --

90-95 O-3 15-20 40-50

5 97 80 50

Ratio bound APO AI/AH

0.5 0.0 1.0

64

were measured in the infranatant after ultracentrifugation at density 1.21. Polyacrylamide gel electrophoresis was used to analyze the infranatant protein (after dialysis) and the protein in the supernatant after delipidation by diethyl ether (Fig. 4). As noted earlier, most (approx. 90%) of the apo AI was recovered in the infranatant with no phospholipid, when added alone to the liposomes at a lipid/protein molar ratio of 50 : 1. However, when added in the presence of a molar excess of apo AII, virtually complete binding of apo AI occurred. When the molar ratio of apo AI was equal to or greater than apo AII, the ratios of the two proteins in the bound form approached but did not exceed unity (Table I). As the amount of apo AI increased further (up to a ratio of 4 in some experiments) only apo AI was found in the infranatant (Fig. 4). Thus, apo AI, although added simultaneously, was not able to compete effectively with apo AI1 for regions of lipid. The lipid - protein complex isolated after incubation (see above) of DPPC liposomes with a solution containing equimolar apo AI and apo AI1 was indistinguishable both in its elution volume and by electron microscopy from the particles produced by addition of apo AI1 alone.

Fig. 4. Varying quantities of apo AI were added to a constant amount of apo AI1 to provide mok~~atios (ape AI/ape AII) from 0.25 to 2.0. Liposomes were then added to each tube to provide a phosphoIipid/ apo AI1 molar ratio of 50 : 1 as described in text. The phospholipid . protein complexes were separated from unbound protein by flotation at d = 1.21. The lipid was then extracted from the SuPematant fraction using dietbyl ether. Equal volumes of the aqueous phase were applied to a series of gels containing 8 M urea buffer. 20 c(g of apo AI (a and g) and of ape AI1 (b and h) were applied as standards for the analysis of both the supematant and infmnatant fractions. For the gels shown, the molar ratios of ape AI to ape AI1 were 0.5 Cc), 1.0 (d), 1.0 (e). 2.0 (f)

65

Discussion Liposomes prepared from DMPC and DPPC tended to aggregate and fuse rapidly as indicated by light absorption and electron microscopic studies. When either apo AI or apo AI1 was added to mixtures of such liposomes at temperatures below the transition of the lipid bilayer, the particle size was stabilized. This indicated that both apolipoproteins interact with the bilayer lipid in the ordered state, perhaps preventing fusion through addition of net surface charge. When incubated with apo AI alone, the structure of the liposomes and quantity of protein bound were not further changed by increasing the temperature above T,.In contrast, when apo AI1 was present, warming above the transition temperature of the lipid bilayer caused a rapid disruption of the lipid particles with the generation of a rather homogeneous complex which appeared as a disc 100 A in diameter and 50 A in thickness. These particles had lo-fold more molecules of protein per unit lipid than the liposomes isolated with apo AI. It has been recently suggested that the apolipoproteins bind to phospholipid surfaces through amphipathic helixes. These proposed helical regions of the protein have both a non-polar and polar surface with the latter containing charged groups arranged to allow ionic interactions with the zwitterionic phospholipid head groups. Apo AI1 contains two such helical regions in each of the monomeric units [ 81 and the much larger apo AI contains 13 such regions [ 41. In light of these marked differences in molecular size and structure, differences in the characteristics of their interaction with lipid are not surprising. Differences in binding might also relate to protein-protein interactions. Certainly, self-association of apo AI could reduce its capacity to bind lipid, and the enthalpy changes associated with lipid binding are compatible with the existence of aggregates [ 171. As noted by others previously [lo], the presence of apo AI1 in the incubation medium clearly facilitates the binding of apo AI to phospholipid. The present experiments indicate that apo AI must be added before or simultaneously with the apo AII. Once formed, the apo AI1 lipid complex does not show increased binding of apo AI over that observed with phospholipid alone. This facilitation of binding could first require the interaction of apo AI with the lipid bilayer surface, perhaps through the amphipathic helixes. The formation of more hydrophobic bonds with apo AI may then occur in a reaction directly induced by the binding of apo AII. The optimum incorporation of apo AI with an equal molar ratio of apo AI1 suggests a specific protein-protein interaction or the alteration of a discrete region of the lipid bilayer by apo AI1 allowing the simultaneous incorporation of one molecule of apo AI. The heat involved in these interactions of protein with lipid has been measured in a series of microcalorimetric experiments reported in the accompanying article. These data also indicate that the binding of apo AI to phospholipid is optimally affected by the addition of apo AI1 with apo AI in an equimolar concentration. It is tempting to believe that the small disc-shaped particle generated from phospholipid liposomes in the presence of apo AI1 has meaning which may be pertinent to physiologic processes. The Stokes radius of the particle is virtually identical to that of human HDL. Particles of similar shape and appearance have been found to occur in liver perfusates [22] and in human plasma [23] under

66

conditions in which there were low levels of the enzyme 1ecithin:cholesterol acyltransferase. These observations have led Hamilton [22] to suggest that the nascent HDL produced in the liver might have a disc-shaped bilayer structure. This type of particle could then be converted into a more globular form by the generation of cholesterol esters through the 1ecithin:cholesterol acyltransferase reaction as demonstrated by the in vitro experiments of Forte et al. [ 241. The end result would be the formation of the sphere of approximately 100 K diameter recognized as the native circulating HDL particle. If this schema were true, one could envisage the generation of nascent HDL particles at a rate determined by the addition of apo AI1 to some membrane-like structure containing apo AI in a more superficially bound configuration. Studies of circulating human plasma HDL do not fit well with the above speculations, since higher molar ratios of apo AI/AI1 have been found by most observers with measurements in some fractions ranging up to a ratio of 9 [25]. In other species, such as the pig 1271 and cow [28], the HDL contains virtually no apo AII. Apoprotein composition studies on nascent HDL freshly produced by the perfused liver should help resolve the question of an important role for apo AI1 in facilitating the incorporation of apo AI into this lipoprotein. Acknowledgments We would like to thank Mr. Pablito Tejada and Ms. Virginia Tejada for their excellent technical assistance. Dr. Juan Yguerrabide, Department of Biology, University of California, San Diego, provided the equipment and advice necessary for the measurement of lipid transition temperatures. This work was supported by the Special Center of Research on Atherosclerosis (HL-14197) and by a grant from the American He.& Association (70-1035). References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Hin. R. and SCMU. A.M. (1970) Biochim. Biophys. Acta 207.364-367 Edelstein, C.. Lim. C.T. and Scanu. A.M. (1972) J. Biot. Chem. 247. 5842-5849 Scanu. A.M., Lim, C.T. and Ed&kin, C.T. (1972) J. Biol. Chem. 247.5850-5855 Baker. H.N., Gotto. Jr. A.M. and Jackson, R.L. (1975) J. Biol. Chem. 250. 2725-2733 Brewer, H.B.. Lux, S.E.. Ronan, R. and John. K.M. (1972) Proc. Natl. Acad. Sci. U.S. 69.1304-1308 Lux, S.E.. John. KM., Flefscher. S.. Jackson, R.L. and Gotto, A.M. (1972) Biochem. Biophys. Res. Commun. 49,23-29 JackSOn. R.L.. Baker, H.N.. David, J.S.K. and Gotto. A.M. (1972) Biochem. Biophys. Res. Commun. 49,1444-1451 Segrest. J.D.. Jackson, R.L., Monisett, J.D. and Gotto. A.M. (1973) FEBS Lett. 38. 247-253 Middelhoff, G. and Brown, W.V. (1973) Circulation XLVIII. IV-79 Assmann, G. and Brewer. H.B. (1974) Proc. Natl. Acad. Sci. U.S. 71, 989-993 Stoffel, W.. Zierenberg, 0.. Tunggal. B.D. and Schreiber. E. (1974) Hoppe-Zeyler’s Z. Physiol. Chem. 355.1361-1390 Havel. R.J.. Eder. H.A. and Bragdon. J.H. (1955) J. Clin. Invest. 34. 1345-1353 Shore, B. and Shore, V. (1969) Biochemistry 8,451M516 Omstetn, L. (1964) Ann. N.Y. Aced. Sci. 121,321--349 Ouchterlony (1958) hog. Allergy 5. 1-78 Huang. C. (1969) BiochemistrV 8.344-352 Rosseneu. M.. Middelhoff, G., Peeters. H. and Brown. W.V.. Biockim. Biophys. Acta Lowry, O.H.. Rosebrough. N.J.. Farr. A.L. and Randall, P.J. (1951) J. Biol. Chem. 193.265-275 Gustafson. A., Alaupovic, P. and Furman, R.H. (1965) Biochemistry 4. 596-605 Bartlett. R. (1969) J. Biol. Chem. 234,466-468 Elbl, H. and Lands. W.E.M. (1969) Anal. Biochem. 30, 51-57

67 22 Hamilton, R.B. end Kayden. H.J. (1974) in Biochemistry of Diseese (Becker, F.F.. ed.). Vol. 5, P. 531, Dekker. New York 23 Forte, T.. Norum. K.R.. Glomset, J.A. end Nichols, A.V. (1971) J. CIin. Invest. 50,1141-1148 24 Forte, T.M.. Nichols, A.V., Gong, E.L.. Lux. S. end Levy. R.I. (1971) Biochim. Biophys. Acta 248, 381386 25 Kostner, G.M., Patsch. J.R.. Sailer, S., Braunsteiner, H. and Holesek. A. (1974) Europ. J. Biochem. 45.611-621 26 Jackson, R.L.. Baker, H.N.. Taunton, O.D.. Smith, L.E.. Garner, C.W. end Gotto. Jr., A.M. (1973) J. Biol. Chem. 248.2639-2644 27 Jonee, A. (1975) Biochim. Biophys. Acte 393.460-470