1082 ( 1991) 265-274 •'~' 1991 ElsevierScience Publishers B.V. (~)05-2760/Q1/$O3.50 A D O N I S 000527609100129Z

265

Biochimi(', et Biophystc'a A (.ta.

BBAL1P53604

Properties of phosphatidylethanolamine-containing phospholipid-apolipoprotein complexes modified by lecithin-cholesterol acyltransferase Elizabeth A. Bonomo*, Janet E. Matsuura and John B. Swaney Department of Biologwal Cher~ustrv Hahnemann C'nirer.~it.vPhiladelpl~ia, P.4 (USA.)

IReceived 6 August 1990) Key words: Lecithin-cholesterolac)ltransferase: Apolipoprotein: Cross-linking:Chole.~terol:Choleste%,lester: Gradient gel electrophoresis

The effect of the inclusion of phosphatidylethanolamine (PE), a phospholipid with unusual packing properties, on the substrate properties of protein-lipid complexes toward lecithin-cholesterol acyllransferase (LCAT) has been studied. Recombinant particles of apolipoprotein A-I with dimyristoylpho.~phatidylcholine (DMPC), dilauroylphosphatidylethanolamine (DLPE) and cholesterol were prepared at a molar ratio of 1:140:14 ( A - l / D M P C / c h o l e s t e r o l ) or 1: 70: 70:14 (A-! / DMPC / DLPE / cholesterol); the efficiency of cholesterol incorporation into complexes containing phosphafidylethanolamine was found tc be very pH-dependent, with enhanced cholesterol incorporation at elevated pH values. By incubating the complexes with either purified human LCAT or the d > !.21 g / n d fraction of ra~ serum as a source of LCAT activity, it was found that a high degree of cholesterol esterlficatlon could be achieved with either complex; however, the DLPE-containing complex possessed a much smaller Stokes' diameter than the DMPC-only particle despite compositional similarities between these complexes. With respect to particle diameter the DLPE-conraining particles behaved more like complexes prepared with egg yolk lecithin than did complexes Wepared with DMPC alone. When human LDL was added to the incubations to provide a source of additional cholesterol, the products were markedly different. Concomitant with an increased cholesteryl ester core was an increase in the protein stoichiometry in both types of particles, from 2 to 3 or 4 apo A-! per particle. The proportion of DLPE to D1V.'PC in the products was rechtced from 1.1 to 0.3: I, reflecting a preferential hydrolysis of PE by LCAT, and the Stokes' diameters of the DMPC-only and the DLPE-containing complexes were closely similar. We conclude that the presence of elevated proportions of certain phospl~ipid species may significantly alter both the phy~,~cal properties of the particles and their substrate properties with regard to reactions with enzymes of lipid metabolicnt. Introduction Plasma lipoproteins are pseudomicellar lipid-protein assemblies that consist of a protein and polar lipid coat

Abbreviations: apo A-I. apolipoprotein A-I; Chol, cholesterol; DLPE, dilauroylphosphatidylethanolamine; DMPE, dimyristoylphosphatidylethanolamine: DMPC, dimyristoylphosphatidylcholine: DMS, dimethylsuberimidate; GGE, pore limit gradient gel electrophoresis: HDL, high-density lipoproteins; LCAT, lecithin-cholesterol acyltransferase; PC, phosphatidylcholine; PE, phosphatidylethanolamine; SDS-PAGE, polyacrylamide gel etectrophoresis in sodium dodecylsulfate;. * Present address: Chemistry Department, Philadelphia College of Pharmacy and Science, Philadelphia,PA 19104, U.S.A. Correspondence: J.B. Swaney, Dept. of Biochemistry IMS 411), Hahnemann University.Philadelphia,PA 19102. U.S.A.

overlaying a core of triacylglycerol and cholesteryl ester. The surface properties of the different lipoprotein classes is presumed to result from the differences in apoprotein content among these lipoproteins, but could potentially also be influenced by the polar lipid constituents. It has been known for some ti:ne that among the circulating lipoproteins, the major polar lipid is phosphatidylcholine (PC). which amounts to 60-80% of the total phospholipids present [1]. Although the proportion of phosphatidylethanolamine (PE) has been generally found to be low (less than 6% of total phospholipids), lipoproteins have been more recently described that have higher proportions of PE. For example, a species of human apo E-free very low density lipoprotein has been described in human plasma in which 9-11% of total phospholipid is PE [21. Furthermore, it has been reported that nascent lipoproteins from rat hepatocytic Goigi fractions are enriched in PE [3,41. Haase and

266 Stoffei [5] have reported that secreted A-l, the major protein of high-density lipoprotein, is mainly associated with newly synthesized PE. It is possible that the assembly of lipoproteins involves the abstraction of membrane lipids [6]; the enrichment of nascent lipoprotein with PE ,' ould then be consistent with the relative enrichment of the luminal side of the endoplasmic reticulum membrane wi,.h PE [7]. It has been speculated that the enrichment of PE in Golgi lipoproteins might play a role in facilitating the secretion of these particles [3]. After secretion into the circulation, the proportion of PE in the lipoproteins seems to diminish; this [,a~ be attributable to phospholipid exchange or to conversion of PE to other products, such as lyso PE. For example, hepatic lipase has been reported to preferentially hydrolyse PE to the lyso form [8,9]: similarly, Pownall and coworkers [10] found that when various phospholipids were incorporated at a low percentage in an inert matrix, PE was the most active substrate for lecithin-cholesterol acyltransferase (LCAT). There has been considerable recent interest in the ab:lity of LCAT to dramatically transform synthetic, model HDL and these investigations have been recently rcviewed by Nichols [11]. However, essentially all of the studies of this sort carried out to date have involved recombinant particles of apolipeprotein A-I and various phosphatidylcholines. We have recently reported a method for incorporating large proportions of PE into synthetic HDL and have characterized these products [12]. It was shown that the presence of PE diminished the protein-lipid interactions and altered the size and density of the product. In this study we have utilized this approach to investigate the effect of inclusion of cholesterol on the properties of PE-rich model HDL and on the properties of the products of the reaction of these particles with the LCAT enzyme. Materials and Methods

Ma;erials Plasma was obtained from male and female human subjects. HDL. LDL and the d > 1.21 g/ml fraction were isolated from human plasma by sequential flotation ultracentrifugation [13]. Apo A-I was isolated from H DL apolipoproteins by Sephadex G-150 (Pharmacia) column chromatography after reduction of the A-II protein with mercaptoethanol to improve the purification [6]. Serum was obtained from male Sprague-DawIcy rats (270-350 g); this was centrifuged to obtained the d > 1.21 fraction as a source of LCAT activity. 5,5'-Dithio-bis(2-nitrobenzoic acid) was purchased from Pierce Chemical Co. (Rockford, 1L). Bio-Beads SM-2 were acquired from Bio-Rad Laboratories (Richmond. CA). Phospholipids were purchased from Avanti Polar Lipids (Birmingham, AL) and sodium

cholate was obtained from Sigma (St. Louis. MO). [3HlCholesterol and [14C]formaldehyde were obtained from Amersham (Arlington Heights, IL). Thin-layer chromatography using Whatman Linear-K 5D Preadsorbant TLC plates was employed to verify the purity of the phospholipids. Phospholipids were separated using a mobile phase mixture of chloroform/ methanol/water (65 : 25 : 4, v/v). All reagents checked for purity yielded a single spot and were used without further purification. Similarly, the purity of the tritiated cholesterol was verified by thin-layer chromatography. Apolipoprotein was radiolabeled by reductive methylation with [~4C]formaldehyde using the method of Jentoft and Dearborn [14]. Under the conditions employed, only about 4-5% of the lysines were modified and no changes in lipid associative properties were noted.

Preparation of lipid-,po A-I complexes PE-containing lipid-protein complexes were routinely prepared by using a modification of a method for preparing PC-containing complexes that we recently described [15} except that the time of centrifugation for the removal of the beads from the incubation mixture was increased from 2 min to 30 rain to pellet any unreacted lipid.

Effect of pH on cholesterol incorporation into complexes In studying the effect of pH on cholesterol incorporation, [14C]apo A-l, phospholipid, and [3H]cholesterol were incubated with cholate at a molar ratio of 1 : 140:14:190 in the presence of various buffers. The following buffers were used: 50 mM sodium phosphate (pH 5.5) or 50 mM sodium acetate (pH 5.5), 50 mM sodium phosphate (pH 7.5) or 50 mM Hepes (pH 7.5), 50 mM sodium borate (pH 9.5), and 50 mM sodium phosphate (pH 11.0) or 50 mM CAPS (pH 11.0); all buffers contained 150 mM sodium chloride, 1 mg/ml EDTA, and 0.1 mg/ml sodium azide. After the removal of cholate by extraction on Bio-Beads columns, the eluate was chromatographed on 1.0 × 30 cm column of Superose 6B (Pharmacia/LKB) and fractions were collected. By dual label counting of radioactivity, the cholesterol/protein molar ratio was averaged over the fractions containing the lipid-protein complex.

Purification of human LCA T LCAT was purified from hyperlipemic human plasma by using a procedure based on the method of Matz and Jonas [161. In brief, the density fraction of 1.71.-1.25 g/ml from human plasma was isolated by centrifugation, dialyzed, and passed over an Affi-Gel Blue column (1.6 cm × 28 cm; Bio-Rad Laboratories) in series with a DEAE Sepharose CL-6B column (1.6 c m × 8 cm; Pharmacia-LKB). The first two peaks following the void volume peak were pooled, dialyzed, and passed over a

267 hydroxyapatite column (1.0 × 18 cm: Bio-Rad Laboratories) and eluted with a gradient of 10-60 mM sodium phosphate. "the extent of purification achieved ranged between 1200-15000-fold in different preparations. LCAT activity was measured by using the assay of Chen and Albers [17] which was modified by substituting complexes of apo A-I, DMPC, and [-~Hlcholesterol as the LCAT substrate and by increasing the reaction time to 1 h.

Reaction of lipid / A-I complexes with LCA T For incubations utilizing purified human LCAT, the preparations of lipid-protein complexes were adjusted to a pH of 7.0 with Hepes. LCAT was added to the complexes to yield a final concentration of 50 units/ml, which is comparable with activity levels we measured in human plasma; mercaptoethanol was added to a final concentration of 5 mM to maintain enzyme activity. After incubation at 37 °C for various times the samples were used for pore limit gradient gel electrophoresis. In later experiments, the incubated samples were subjected to FPLC separation on a column of Superose 12B to remove ,qgge'egated material prior to electrophoresis and chemical analysis; the fractions were pooled for analysis based upon the absorbance tracing at 280 nm and upon the profile of radioactive cholesterol. in cases where the d > 1.21 g/ml fraction of plasma was used as a source of LCAT activity, the samples were centrifuged at d= 1.21 g/ml after incubation, followed by gel filtration on a Superose 12B column; this procedure effec*,!v*ly removed any high molecular weight aggregates that were observed in studies with purified LCAT. In time-course studies, aliquots were removed and the LCAT reaction stopped by the addition of DTNB to a concentration of 1.5 mM. In some studies, LDL was added at a ratio of 0.5 : 1 (LDL free cholesterol to recombinant phospholipid) and the ,,1'> 1.21 g/ml fraction added to yield a final concentration equal to two-thirds that in native plasma Where LDL was added after an initial incubation, additional d > 1.21 g/ml material was added to restore the two-thirds plasma concentration. For these samples centrifugation was required at both d = 1.063 g/ml and at d = 1.21 g/ml and gel filtration was found not to be required.

Characterization of phospholipid-apo A-1 complexes To quantitate protein in the presence of lipid a modification of the Lowry procedure was used [18]. Total phospholipid phosphorus was determined by acid digestion and colorimetry [19]. To determine the phosphatidylcholine concentration, an enzymatic assay kit (Biochemical Diagnostics, Edgewood, NY: phospholipids B kit) was used. The concentration of PE was derived by subtracting the PC concentration from the concentration of total phospholipid phosphorus. Total

and free cholesterol were measured enzymaticaily, using reagents prepared by Fermco (Elk Gro,,e Village. IL); a sample of control human serum (Control H, Sigma) was analyzed on each occasion to verify the day-t~-day reproducibility of the results. The size of the lipid-protein complexes was determined by pore limit gradient gel electrophoresis (GGE). Electrophoresis was conducted for a total of 3000 V. h using precast PAA 4/30 gels purchased from Pharmacia (Piscataway, N J) [20]. Protein standards (Pharmacia) for estimating Stokes diameters were purchased from Pharmacia and the following values for Stokes" diameter were used: thyroglobulin, 17.0 nm; ferritin, 12.2 nm: catalase, 10.2 nm; lactate dehydrogenase. 8.1 nm; and albumin, 7.1 mn [211. Autorad~ography of gels containing tritiated cholesterol was performed after soaking the gels in Amofluor (National Diagnostics, Somervi~,le, N J) and dt3'ing on a slab gel dryer. To determine the extent of lipid-protein complex formation and the protein stoichiometry of the complexes, a cross-linking reagent, dimethylsuberimidate (DMS, 20 rag/rot in 1 M triethanolamine-HCl, pH 9.7), was added to aliquots of the samples and allowed to react at room temperature for 2 h [22]. The cross-linked self-associated forms of lipid-free apo A-I range from monomers to pentamers and these bands were used as markers to determine the number of protein cheins per particle in the lipid-protein complexes.

Fluorescence spectroscopy To evaluate any changes in the environment of the tryptophans in the lipid-protein complexes the wavelength maximum for intrinsic tryptophan fluorescence emission spectrum was measured using a Hitachi U-2000 fluorescence spectrophotometer at an excitation wavelength of 280 nm. Samples for analysis were isolated before or after modification with LCAT in the density interval of 1,063-1.21 g/ml, dialyzed against saline, EDTA (pH 7.4), and diluted to 0,020 mg protein/ml with buffer. Results

pH dependence of cholesterol incorporation into cornflexes In order to establish the effect of the incorporation of cholesterol into PE-containing recombinant complexes, dried lipid mixtures were dispersed into buffers at a variety of pH values and the extent of cholesterol uptake into the lipid-protein complexes after removal of detergent was determined (Table I). It was found that cholesterol incorporation into the product was enhanced at elev~,ted pH values to some degree with A-I:DMPC complexes, but to an even greater extent with A-I:DMPC:DLPE complexes, with a sharp break

268 FABLE I

Effect uf plt on clmlevterol incorporation into complexes Complexesof [laCIA-I. phospholipid,and ['tHIcholesterolwere prepared and the molar ratio of cholesterolto protein was determinedby dual-isotope scintillation counting; data are averages from 3-4 scparilte e x p e r i m e n t s .

Chol/A-I molarratio+ S.D. plq: 5.5 7.5 9.5 ll.0 A-I/I')MP(." 9.0+1.4 10.6+1.2 11,9+0.9 13.9+2.1 A-I/DMPC/DLPE 5.6_+0.8" 5.0_+0.2" 14.0+1.3 t' 12.8+3.5 differencessignificantat P < 0.02 by t-test v.~A-i/DMPC. h difference.significantat P < 0.04 by t-test vs A-I/DMPC.

100

.//

80

A-hDMPC:DMPE:Cho~

60

L~EC : ho!

40

20

//////I

0

between pH 7.5 and 9.5 (Fig. 1). These results are consistent with studies of cholesterol affinity for multilamellar vesicles, which similarly show a strong pH dependence for DLPE [23]. For subsequent preparation of complexes as substrates for LCAT, a pH 9.5 buffer was used to optimize the cholesterol content of the substrate. The effect of including various amounts of cholesterol on the size distribution of the resultant p h o s p h o l i p i d - c h o l e s t e r o l - p r o t e i n complexes was studied by gradient gel electrophoresis. It was found that inclusion of cholesterol caused the particles to migrate slightly farther, reflecting a small decrease in the Stoke's diameter of these particles (data not shown). Time-course Jor LCA T modification To establish the time-course for the cholesterol esterificatio;~ reaction upon lipid-protein complexes, a purified LCAT was added to incubation mixtures containing the following complexes: A-I:DMPC:Chol (1 : 1 4 0 : 1 4 m o l / m o l ) , A - I : D M P C : D L P E : C h o l ( 1 : 7 0 : 7 0 : 14), and A - I : D M P C : D M P E : C h o l (1 : 70 : 70 : 14). As shown in Fig. 2, cholesterol esterifi-

~,00 . . . . . . . . . . .

-

,

½" ~0

< rl

60

© LJ --

___---

go 20

,-;E

6.0

8.0

10.0

pH Fig. 1. Effect o f p H o n the i n c o r p o r a t i o n o f c h o l t ~ t e r o l i n t o c o m plexes of a p o A-I with D M P C ((9) o r 1 : 1 D M P C : D L P E (O).

I

1

I

2 3 TIME (hours)

,r/

I

24

Fig. 2. Time-courseof cholesterolesterificationby purified LCAT in recombinant complexesof A-l/DMPC/cholesterol (1:140:14) (o). A-l/DMPC/DLPE/cholesterol (1 : 70 : 70 : 14) (e), and A-I/ r~MPC/DMPE/cholesterol(I : 70: 70 : 14) (O). cation proceeded at a much greater rate in PE-containing complexes, but that after 24 h of reaction most of the cholesterol in each of the substrate particles was esterified. Thus, to achieve a similar degree of cholesterol esterification in the complexes, subsequent reactions with LCAT were carried out for 24 h. Effect of LCA T modification upon particle size The progress of LCAT-mediated cholesterol esterification as a function of time was also assessed by monitoring the size of the lipid-protein complexes by gradient gel electrophoresis. Some studies were performed utilizing purified LCAT and the incubated samples were applied directly to the gels (Fig. 3) to eliminate the risk of damage to the particles stemming from ultracentrifugation; however, similar results were observed when the d > 1.21 g / m l fraction of serum was used as a source of LCAT (which necessitated ultracentrifugal re-isolation). As shown in Fig. 3A, Al:DMPC:[3H]cholesterol complexes possess two major bands, with diameters of approx. 10.4 and 11 nm. The only significant change resulting from modification by LCAT is a shift at early time points to enhance the 10.4 nm band. On the other hand, reaction of A'.:DMPC:DLPE:[3H]cholesIeTol complexes with purified LCAT produces a dramatic change in electrophoretic properties (Fig. 3B). The complex at 0 time shows essentially a single band with a Stokes' diameter of 10.1 nm, but after 2-4 h of LCAT reaction there is a nearly-quantitative conversion of this band to a particle of about 7.6 nm. Some material of very large diameters is also produced, but these bands are also produced in a sample which is incubated in the absence of LCAT, suggesting that aggregation may take place. Autoradiograms were also obtained from the gradient gels which show the distribution of [3H]cholesterol

269

Diam.

- "•L ~

17.012.2-

12.2-!!i:!~::

10.2..

10.26.1-

6°1,.. .

7 . 1 - •.!~i'~ t: 1

7.12

3456

7

1

2

3456

7

D C

/I

~

• •

1

~

?.

.;' :i';

• : . , .-'.~

2

ii:!ia Li'4"•. •i5 !:'•i~Qi;~

Fig. 3. Gradient gel electrophoresis of A-I/DMPC/[3H]cholesterol complexes (Panels A and C) and A-I/DMPC/DLPE/|aH]cholesteroi complexes (Panels B and D) after various periods of modification by purified I,CAT: panel,,, A and B are photographs of Coomassie stained gels while panels C and D are autoradiograms of the same gels. Lane 1 contains protein staodards and lanes 2-7 contain the protein lipid complexes treated as follows: hours of incubation with LCAT-Lane 2-0 h; Lane 3-1 h: Lane 4-2 h: Lane 5-4 h; Lane 6-24 h; Lane 7-incubated 24 h in the absence of LCAT.

(Figs. 3C and 3D). By comparison with the Coomassiestained patterns, it is evident that, except after prolonged modification by LCAT, bands that stain poorly for protein contain a disproportionate amount of

appear to aggregate to very high molecular weight forms. On the other hand, when active LCAT is present, this aggregation is counteracted by the conversion to products so that the LCAT-modified product is the major

labelled cholesterol and that samples containing PE

band both by protein staining and by localization of the

270 cholesterol label. Thus, both with A - I - D M P C cholesterol complexes and with A - I - D M P C - D L P E cholesterol complexes the product seen after 24 h of LCAT modification is more h o m o g e n o u s in size than before incubation or after incubation in the absence of LCAT. Because of the presence of high molecular weight material in these preparations, in later experiments we routinely purified the complexes after incubation by gel filtration on a Superose 12B column, which readily separated the aggregated material from the p r o t e i n - l i p i d products or, alternatively, we isolated the d = 1.063-1.21 g / m l fraction. Most preparations showed less aggregation than was present in Fig. 3, as evidenced by the a m o u n t of a high molecular b a n d by gel filtration.

Isolation and characterization of products after 24 h of modification by LCA T To further characterize the p r o d u c t s of L C A T modification of such c o m p l e x e s , D M P C - o n l y and D M P C / D L P E containing complexes were incubated either with purified h u m a n L C A T or with rat or h u m a n d > 1.21 g / m l fraction of serum as sources of L C A T activity. Following a 20 h incubation, the complexes were re-isolated by ultracentrifugation at 1.063-1.21 g/ml. The chemical composition of the products is presented in Table II. U n d e r the conditions of our experiments the degree of cholesterol esterifieation was somewhat higher with the d > 1.21 g / m l fraction of rat or

h u m a n serum ( > 90%) than that achieved by purified L C A T (60-85%): we attribute the lower degree of esterificalion by purified L C A T to reduced stability of the enzyme activity during the incubation, and to the absence of albumin to remove products. The composition of either D M P C - o n l y or D M P C / D L P E complexes tended to be rather similar regardless of the source of L C A T activity. T h e p h o s p h o l i p i d - p r o t e i n ratios in the isolated products were lower than in both the starting materials and in the products not subjected to the centrifugal re-isolation procedures. These products also had little or no aggregated material evident in the G G E profiles. Protein stoichiometry was assessed by chemical cross-linking with dimethyl-suberimidate. In all eases, the major b a n d seen with both modified and u n m o d ified complexes c o r r e s p o n d e d to the A-I dimer (data not shown); this indicates that there are two A-I molecules per particle a n d that this stoichiometry was unaffected by L C A T modification.

Effect of LDL on the LCA T modification of recombinants It was of interest to learn whether the presence of a source of additional free cholesterol would influence the nature of the L C A T modification of the complexes, Inclusion of h u m a n L D L at a p r o p o r t i o n of 0.5 tool L D L free cholesterol per tool of r e c o m b i n a n t phospholipid was f o u n d to dramatically increase the a m o u n t of total cholesterol in the p r o d u c t a n d to substantially reduce the p r o p o r t i o n of p h o s p h o l i p i d (Table I1). In the

TABLE I[

Compositwno/lipid-proteincomplexesafter LCAT modification Complexes containing protein, pl',ospholipid, and cholesterol were prepared and incubated for 24 h with purified human LCAT or the d > 1.21 g/ml fraction of rat or human serum as a source of LCAT activity. In some experiments human LDL was added to provide a source of additional free cholesterol. Composition (molar ratio) a apolipoprotein A-I A-i/DMPC/CHOL (1:140:14 mol/mol) b + purified LCAT ~ 1 + human d > 1.21a 1 + rat d > 1.21 1 + rat d > 1.21 + hLDL ~ l

DMPC

DLPE

75± 7 66±13 65 ± 10 37 + 9

unesterified cholesterol

esterified cholesterol

2.6±0.7 0.6±0.3 0.7±0.4 2.5±1.3

5.7±1.3 13 ±2.0 12 ±0.9 38 ±0.2

1.0±0.5 1.9±0.3 0.6±0.1 6.1±1,8

4.9±0.8 15 ±0.9 17 ±3.5 37.4±1.3

A-I/DMPC/DLPE/C_'HOL(1:70:70:14 tool/tool) h + purified LCAT 4. human d > 1.21d + rat d>l.21 + rat d>I.21+hLDL ~

1 1 1 1

32+ 7 40+10 37 4-4 28± 5

41 ±16 35 ± 9 28 ± 1 9.5± 1.5

J Molar ratio+ range of values; based upon two experiments, within which each quantity was determined twice, in duolicate. " Molar ratio of starting material. Purified LCAT was added to bring the activity level to approx. 50 units/ml. 0 The d > |.21 g/ml of human or rat serum was dialyzed and added to bring the plasma concentration to a level approx, two-thirds that of native plasma. Human LDL (hHDL) was added to bring the ratio of LDL unesterified cholesterol equal to 50~ that of the total phospholipid in the substrate complexes.

271 17 nrn

12.2 nm

10.2 nm

8.1 nm

!

7 1 nm

G,

+LDL

,.-/7

/~ L..",'/

--nL

i

"\"~ aD

~

-LDL

.i//""x

,33

/':?'

..

J ",!)\

L___ I

:

1

t

I

0.200

0.400

0.600

0.800

1.000

., IT

I ~

~

543

i

/i/~\\

i

2

MIGRATION ----,

Rf

A.I/DMPC/cholesterol

Fig. 4. Gradient gel electrophoresis of re-isolated A-I/DM2C/ cholesterol complexes(solid line) and complexes (dashed line) after 20 h incubation with rat d > !.21 g/ml serum fraction with or without added human LDL. The RI.- value was based upon the migration of serum albumin and the migration of protein standards is indicated along the top of the diagram.

A.|/DMPC/DkPE/cholesterol

case of the DMPC/DLPE-containing complexes, there was a marked reduction in the ratio of PE to PC, presumably reflecting the preference of LCAT for this phospholipid species. The presence of LDL during the LCAT reaction also caused a dramatic change in the Stokes diameter of the products, as measured by gradient gel electrophoresis (Fig. 4). As might be expected, the size of the D M P C -

only complex increased somewhat (from 10.2 nm to 10.8 nm) as a consequence of the increased cholesteryl ester core. Of greater surprise was the result that the PE/PC complex yielded a product which was now

Fig. 5. SDS-PAGE of re-isolat~ complexes (solid lines) and A-I/'DMPC/DLPE/choh:'stcro| complexes (dashed lines) cross-linked with dimethylsubcrimidateafter 20 h of incubation with d > 1.21 g/ml fraction of rat serum with or without added human LDL The pr¢,,~mce of band~ corresponding to one A-! per particle for the PE-containing sample are belie~,ed to be mainly attributable to competition of the amino group on the .nhospholipid for the cross-linking reagent. The numbers along the x-axis denote the migration of A-I oligomers ranging from mono~r (28.3 kDa) to pentamer (142 kDa).

the generation of complexes with 3 A - I / c o m p l e x versus 2 A - l / c o m p l e x prior to L D L addition (Fig. 6, Table III). Within 4 h the observed changes were substantially complete, although at 24 h after addition complexes containing 4 A - l / p a r t i c l e were also observed, primarily in the PE-containing complex, The percentage of cholesterol in ester form, already in excess of 95%,

LDL

slightly larger in Stokes diameter to the D M P C - o n l y complex, rather than smaller, as is the case in the absence of a source of additional cholesterol (Fig. 4). Also striking was the protein stoichiometry of these

particles, as revealed by cross-linking (Fig. 5). Whereas products formed in the absence of added LDL possessed principally two A-I per particle, in the presence of LDL both types of complex were found to have predominantly three or four A-I per particle. Thus, it appears that particle fusion or protein transfer occurs to stabilize the particles during extended reaction with LCAT. In additional experiments, complexes were reacted with rat d > 1.21 g/ml fraction for time periods up to 24 h, at which time human LDL and additional rat d > 1.21 g/ml fraction were added and the incubation was extended for an extra 24 h. A limiting cholesterol esterification of 95-97% was achieved within the first 6 h for PE-containing complexes, with D M P C - o n l y cem-

plexes reaching 80% esterification at 6 h and 97% at the 24 h time point. With the addition of LDL, substantial changes were seen within the first hour which included the production of a product with smaller R F value and

i I

i I

¢ Iti

0.2

0.4

0.6

0.8

1.0

Rf Fig. 6. Pore limit gradient gel el~tmphoresis of DMl~.only and PE-containing complexes after incubation for up to 24 h at 37 C with rat d > 1.21 g/ml fraction, followed by the addition of human LDL with incubation for up to an additional 24 h. Gels were stained with Coomassie blue and ~anned with a densitometer: the x-axis corresponds to a migration distance relative to bovine serum "albumin. The time points presented are: 0 time (no incubation), solid line: 24 h incubation prior to LDL addition, dashed line: l h after addition of LDL open circles; 24 h after LDL addition, closed circles.

272 TABLE III

Trine-course of LCA T modification with and without added LDL Protein/lipid complexes were prepared and incubated with a source of I.CAT activity; at this time human LDL and additional LCAT actMty were added and the incubation was continued for an additional 24 h. Values represent the mean of two experiments. Composition (molar ratio) apolipoprotein A-! A-I/DMPC/CIIOL (1 :!40:14 mol/mol) prior to LD1. addition 0h 1 2h 1 6h 1 24 h 1 after addition of LDL

DMPC

1

1 1

unesterified cholesterol

es,erified cholesterol

A-l/ particle "

16 8 3

0 10 12

2 2 2

1

15

2

1 1 1

22 26 30

3 3 3

14 4 2

11 17

2 2 2

1

18

2

1

24

3

1 1

24 28

3 3,4

159 83 69 53

22 23 19 A-I/DMPC/DLPE/CHOL (I : 70 : 70 : 14 mol/mol) prior to I.DL addition 0h 1 70 2 I1 1 38 6h 1 41 24 h 1 28 after addition of LDL lh 1 16 4h 1 18 24 h 1 17 A-I/EYPC/CHOL (1:70:70:14 mol/mol) prior to LDL addition 0h l 160 b 24 h I 72 b after addition of LDL 24 h 1 41 h 1h

4h 24 h

DLPE

65 25 N.D. 16 3 3 3 _

16

0

2

1

12

2

-

1

26

3

" ptotei1~ stoichiometry was determined by chemical cross-linking.

b egg yolk phosphatidylcholine.

was not altered by a d d e d L D L , but the m o l a r ratio of total cholesterol to protein was m a r k e d l y increased within the first hour, from a b o u t 1 6 : 1 to 2 3 : 1 (Table III). This increase, perhaps along with a c o n c o m i t a n t drop in the p h o s p h o l i p i d - p r o t e i n ratio, a p p e a r s to be the trigger which induces the increase in size a n d protein content, a n d this process a p p e a r s similar to the changes in structure observed with physiological H D L [24-26]. These results s h o w t h a t the o u t c o m e is indep e n d e n t of whether L D L is a d d e d at the onset of L C A T action or after the initial L C A T - m e d i a t e d t r a n s f o r m a tion is already complete. Studies were also p e r f o r m e d in which complexes were prepared with A-l, egg yolk PC, a n d cholesterol, and these were i n c u b a t e d with rat d > 1.21 simultaneously with D M P C / A - I / c h o l e s t e r o l and DMPC/ D L P E / A - I / c h o l e s t e r o l complexes. It was f o u n d by gradient gel electrophoresis that complexes f o r m e d with egg yolk PC migrated m u c h faster after i n c u b a t i o n w i t h o u t a d d e d LDL, but s u b s e q u e n t l y m i g r a t e d slower after L D L was a d d e d to provide a d d i t i o n a l free cholesterol; these results were essentially tdentical to those reported by Nichols et al. [27]. In this regard, the

egg yolk PC complexes behaved more similarly to the PE-containing complexes than they did to the DMPConly complexes. The compositional changes upon in-

LIJ ¢,.) Z ,< 123 re"

0

tfl m