Biochimica Elsevier

BBALIP

et Biophysics

13

Acta, 1046 (1990) 73-80

53469

Kinetic evidence for phosphatidylethanolamine and triacylglycerol as preferential substrates for hepatic lipase in HDL subfractions: modulation by changes in the particle surface, or in the lipid core Christine AzCma, Pedro Marques-Vidal, Anne Lespine, Gilles Simard, Hugues Chap and Bertrand Perret INSERM

Unit6 326, Phospholipides

membranaires

(Revised

Key words:

Signalisation

celhdaire et Lipoproteines,

Hapita

Purpan, Toulouse (France)

(Received 19 February 1990) manuscript received 16 May 1990)

Hepatic

lipase; HDL;

Phospholipid;

Triacylglycerol

Human HDL subfractions, HDL, (d: 1.085-1.125) and HDL, (d: 1.125-1.19) labelled with 2-[‘4C]linoleoylphosphatidylethanolamine and tri-[ 3H]oleoylglycerol, were incubated with partially purified hepatic triacylglycerol lipase, isolated from human post-heparin plasma. Kinetics of hydrolysis of these two HDGlipid substrates were followed and were compared to those previously obtained on phosphatidylcholine (G. Simard et al (1989) Biochim. Biophys. Acta 1001, 225-233). (1) The apparent K, obtained for HDL-triacylglycerol was half that for HDL-phosphatidylethanolwas higher for the latter. Hence, despite a lower affinity, more molecules of amine, but the estimated V,, phosphatidylethanolamine than of triacylglycerol were found hydrolysed. A strong correlation was observed between the hepatic lipase activity added and the maximal degradation rates for phosphatidylethanohunine measured in HDL, and HDL,. (2) A linear relationship was observed in both HDL, and HDL, between the respective degradations of the two substrates. The number of phosphatidylethanolamine molecules hydrolysed exceeded that of triacylglycerol by 30% in HDL, and by 70% in HDL,. HDL, were 2- and 4-times more reactive than HDL, for the hydrolysis of phosphatidylethanolamine and triacylglycerol, respectively, talcing the V,,/K, ratio as an indicator of catalytic efficiency. In both HDL subfractions, the calculated V,,/K, value was 30-50-fold higher for PE and TG than for PC. (3) HDL particles were modified either on their surface by selective enrichment in free cholesterol or in their inner-core by replacement of esterified cholesterol by triacylglycerol in presence of a source of neutral lipid transfer activity. A mild cholesterol enrichment stimulated the phosphatidylethanohunine and triacylglycerol reactivities by 30&O% towards hepatic lipase, whereas increasing the triacylglycerol concentration in HDL was followed by a proportional increase in the amounts of triacylglycerol hydrolysed with no effect on phospholipid degradation.

Introduction

Hepatic triacylglycerol lipase is involved in the clearance of plasma HDL phospholipids and cholesterol [l-5]. The enzyme exerts both triacylglycerol lipase and phospholipase activities or circulating lipoproteins [6,7], or on artificial models [8-111. We recently observed that triacylglycerol is very rapidly consumed during kinetic analysis of HDL-phos-

Abbreviations: H-TGL, hepatic triacylglycerol lipase; PE, phosphatidylethanolamine; TG, triacylglycerol; PC, phosphatidylcholine; PL, phospholipids; FC, free-cholesterol. Correspondence: B. Perret, INSERM Unit6 326, Phospholipides membranaires Signalisation cellulaire et Lipoproteines, Hi3pital Purpan, 31059 Toulouse Cedex, France OOOS-2760/90/$03.50

0 1990 Elsevier Science Publishers

B.V. (Biomedical

phatidylcholine hydrolysis by hepatic lipase, in vitro [12]. Regarding phospholipids, the clearance of injected HDL-radiolabelled phospholipids in rats is much faster for phosphatidylethanolamine than for phosphatidylcholine, a process inhibited by an antiserum against hepatic lipase [13]. However, no kinetic analysis was so far reported regarding the behaviour of phosphatidylethanolamine (PE) and triacylglycerol (TG), compared to phosphatidylcholine (PC), as substrates for hepatic triacylglycerol lipase (H-TGL), when present together in human HDL particles. In the present study, mild conditions of lipolysis were used so as to follow the kinetics of TG and PE hydrolysis, in HDL-subfractions, by action of a partially purified H-TGL, isolated from human postheparin plasma. Both molecules appeared as preferential substrates, compared to phosphatidylcholine. Division)

74 Most studies dealing with the phospholipase activity of hepatic lipase on HDL are based on PC hydrolysis [6,9,12,14]. We recently observed that the latter was enhanced following free cholesterol enrichment of the particle outershell, whereas replacement of cholesteryl ester by triacylglycerol in the inner core had no influence. The effects of such modifications were thus investigated here, regarding the rates of hydrolysis of two preferential substrates for hepatic lipase. It was observed that changes affecting either the HDL surface or its lipid core differentially orientate the particles reactivity towards the enzyme. Materials and Methods Materials Bovine serum albumin, essentially fatty acid-free (BSA); L-cY-lysophosphatidylcholine from egg yolk; lpalmitoyl,2-oleoyl-sn-glycero-3-phosphocholine; l-palmitoyl,2-oleoyl-sn-glycero-3-phosphoethanolamine and triolein were from Sigma (St Louis, MO, U.S.A.). Tri[ 3H]oleoylglycerol and l-palmitoyl,2-[l-‘4C]linoleoyl-sn-glycero-3-phosphethanolamine (59 mCi/mmol) were from Amersham (Les Ulis, France). Standard heparin was from Choay laboratories (Paris, France) and heparin-Sepharose CLdB was from Pharmacia (Uppsala, Sweden). Chromatography gel AcA 34 was from LKB (Bromma, Sweden). Celite (545 grade) was from Merck (Darmstadt, F.R.G.). Isolation and labeling of HDL Plasma pool was obtained from healthy volunteer donors. HDL and HDL subfractions (HDL,, HDL,) were isolated by sequential ultracentrifugation in the density intervals of 1.085-1.125 for HDL, and 1.1251.190 for HDL, and were washed at their lower density limit. Lipoproteins were kept at 4” C under nitrogen in a dark place. Prior to use, HDL were extensively dialysed against 10 mM Tris-HCl buffer (pH 7.40) containing 150 mM NaCl, 0.01% (w/v) sodium azide and 0.25 mM EDTA. Purity of the fractions was assessed by electrophoresis on a 2-3% polyacrylamide gel (Sebia, Issy-les-Moulineaux, France). For the labelling of HDL with [‘4C]-phosphatidyll-palmitoyl,2-[l-‘4C]linoleoyl-snethanolamine, glycero-3-phosphoethanolamine (5 PCi) and a mixture of unlabelled phospholipids made of 1-palmitoyl,Z oleoyl-sn-glycero-3-phosphoethanolamine (250 nmol) and l-palmitoyl,2-oleoyl-sn-glycero-3-phosphocholine (750 nmol) were put dry and resuspended in 2 ml Tris/NaCl buffer. Phospholipids were sonicated for 25 min at 4 o C and the solution was centrifuged at 150 000 x g. The supematant containing phospholipid vesicles was incubated with HDL, or HDL, (10 pmol total phospholipid) for 240 min at 37 o C under mild shaking. All preparations were concentrated 3-4-fold by filtra-

tion (Amicon cones, Amicon, Danver, MA, U.S.A.) and HDL were recovered by gel filtration chromatography on an AcA 34 column (0.5 X 50 cm) equilibrated with Tris/NaCl buffer. HDL-triacylglycerol labelling was performed as follows: trace amounts of [3H]trioleolylglycerol (20 PCi) were added to phosphatidylcholine (1 pmol) containing liposomes. prepared as above. Incubations with HDL subfractions (10 pmol phospholipids) were carried out at 37 o C for 180 min. 3H-labelled HDL were reisolated by gel filtration. The specific radioactivities ranged from 2500-40 000 dpm/nmol [‘4C]phosphatidylethanolamine and 540013 500 dpm/nmol [ 3H]triacylglycerol. The chemical composition of each lipoprotein preparation was determined before use, and is expressed as percent weight. In HDL,, protein accounted for 43.8 ( k 1.7%); phospholipid, 30.8 (+ 2.4%); esterified cholesterol, 20.6 ( + 0.7%); free cholesterol, 3.4 ( + 1.2%) and triacylglycerol, 1.5 (f 0.1%); (mean f S.E., n = 4). HDL, displayed the following average weight composition: protein, 50.3 ( f 1.7%); phospholipid, 28.0 (f 3.8%); esterified cholesterol, 18.7 (& 1.8%); free cholesterol, 1.6 (+0.5%) and triacylglycerol, 1.45 (*O.lW), from six preparations. Labelling HDL with [‘4C]phosphatidylethanolamine or with [ 3H]triolein resulted in a parallel recovery of the different components, yielding very similar chemical compositions. The average contributions of PE to total phospholipids were 4.9 + 0.3% for HDL, and 4.6 _t 0.2% for HDL,. Isolation of the post-heparin plasma hepatic triacylglycerol lipase (H- TGL) Partial purification of H-TGL from human postheparin plasma was achieved by heparin affinity chromatography as described previously [15]. H-TGL was eluted at 0.75 M NaCl, in 5 mM barbital buffer (pH, 7.40) containing 10% glycerol and fractions of maximal activity were pooled. Heparin 15 IU/ml and BSA 0.25% (w/v) were added to stabilise H-TGL activity. The enzyme preparation was dialysed for 120 min at 4” C with several changes against Tris/NaCl buffer, so as to remove glycerol. The preparation was then concentrated lo-15-fold in a centrifuge concentrator (Amicon, Danver, MA, U.S.A.) and stored at - 80 o C. H-TGL activity was assayed according to NilssonEhle and Schotz [16] using [3H]triolein as a substrate in the presence of 1 M NaCl. Enzymatic activity is expressed as mIU (nmol of free fatty acid released per min). Enrichment of HDL, with cholesterol Incorporation of cholesterol into HDL, was performed as follows: HDL, (1 mg protein/ml) were mixed with Celite 545 (50 mg/ml) loaded with 4% (w/w) cholesterol according to Avigan [17]. Incubations were

75 carried out at 37” C for 8 h and 16 h under mild shaking. The samples were centrifuged at 30000 x g for 20 min and the supernatant was filtered through a millipore-filter (0.45 pm). HDL, were then reisolated by gel-filtration chromatography, as above. Enrichment of HDL,

with triacylglycerol

HDL, were modified in the presence of VLDL and with/without the d > 1.21 plasma fraction according to Simard et al [12]. HDL, (0.20 mM cholesterol total, 0.025 mM free cholesterol) and VLDL (0.145 mM free cholesterol, 1.39 mM triacylglycerol) were incubated in 50 mM Tris-HCl (pH, 8.0), 0.135 M NaCl and 1% bovine serum albumin (w/v) and in presence or absence of the d > 1.21 plasma fraction (25%, v/v), the lecithin-cholesterol acyltransferase activity being here inhibited by addition of 0.4 mM dithionitrobenzoic acid. Incubations were carried out at 37 o C for 360 min in a shaking water-bath. At the end of the incubation, VLDL and modified HDL were separated by ultracentrifugation at d: 1.070 and HDL were further reisolated at d: 1.21. HDL were extensively dialysed at 4’ C against Tris/NaCl buffer and their chemical composition (given in the Results section) was determined before use for in vitro incubations with the partially purified H-TGL. In vitro incubations of HDL subfractions with partially purified H-TGL

H-TGL was incubated with native or modified HDL and HDL,), labelled with subfractions (HDL, [‘4C]phosphatidylethanolamine (PE) and [ 3H]triacylglycerol (TG) in a concentration range of 0.12 to 2.4 mM HDL-phospholipids. The enzyme activity added ranged between 15 and 65 mIU/ml. In preliminary experiments, it was verified that the rates of HDL-PE and HDL-TG hydrolysis were proportional to the enzyme concentrations in the range tested. Incubations were carried out in 50 mM Tris-HCl (pH, 8.20), containing BSA 2% (w/v), NaCll50 mM, CaCl, 1 mM, in a final volume of 0.4 ml. Incubations were for 60 min at 37 o C, under mild shaking. At the end of the reaction, lipids were immediately extracted according to Bligh and Dyer [18] after acidification with 12 ~1 formic acid/ml aqueous solution. This procedure enables a full extraction of lysophospholipids. Analytical procedures

Lipids were analysed by mono-dimensional thin-layer chromatography on Silica-gel G,, plates (Merck, Darmstadt, F.R.G.), using the solvent system of Skipski et al [19] to separate phospholipid classes. Neutral lipids were separated using petroleum ether/diethyl oxyde/ acetic acid (165 : 35 : 2, v/v) as a solvent. A radioactivity scanning was performed for each plate using an automatic TLC-linear analyser (Berthold LB 2842, LKB,

Bromma, Sweden). Spots were scraped off and assayed for their radioactivity, using a Packard tri-carb 4 530 liquid scintillation counter with automatic quenching correction (Packard Instrumentation International, Zurich, Switzerland). Proteins were measured according to Lowry et al [20] using BSA as a standard. Total and unesterified cholesterol were determined enzymatically [21], using commercial kits (Boehringer, Mannheim, F.R.G.). Phospholipids were estimated as the lipid phosphorus content according to Bbttcher et al [22]. HDL triacylglycerol content was assayed enzymatically [23], in a centrifugal autoanalyser (Cobas Bio, Roche, France). Calculations and statistics

For both substrates, hydrolysis was measured as the relative decrease of [14C]PE or [3H]TG compared to the same particles incubated without enzyme. Data on substrate hydrolysis were matched against the appearance of lipolysis products, i.e., 3H-labelled monoacyl and diacylglycerols and 3H-labelled free fatty acids for [ 3H]TG hydrolysis and [‘4C]lyso PE and 14C-labelled free fatty acids for [14C]PE. Degradation rates were converted to mass substrates degraded by taking into account the specific radioactivities of [14C]PE, and [3H]TG, in each HDL. Michaelis constants, K, and VInax, were estimated in each case using the double-reciprocal plot (l/u = f l/S) and the u = fu/S expressions in which S represents the concentration of unhydrolysed, remaining, substrate. Linear regression factors varied from 0.84 to 0.99. Results are expressed as means f S.E. Results I. Kinetics of PE and TG hydrolysis in HDL subfractions, as mediated by H-TGL

Increasing concentrations of doubly labelled HDL were incubated with comparable H-TGL activities (Figs. 1A and B). For a given amount of either HDL, or HDL, phospholipids, the respective concentrations of PE and TG were almost similar. Hydrolysis of the two substrates was hardly saturable, particularly in the case of phosphatidylethanolamine. This was reflected in the estimated kinetic parameters. The apparent K, values were twice as less for triacylglycerol than for phosphatidylethanolamine (0.025 mM TG and 0.056 mM PE, respectively, in HDL,; 0.037 mM TG and 0.061 mM PE, respectively, in HDL,), the corresponding total phospholipid concentrations in HDL ranged between 0.55 and 1.3 mM, values close to normal plasma levels. Taking the V,,/K, as an indicator of catalytic efficiency, HDL, were 2- and 4-times more reactive than HDL, for the hydrolysis of PE and TG, respectively. However the catalytic efficiencies of phosphatidylethanolamine and triacylglycerol appeared very com-

76

0022

0.043

0 075

0.086

TG (mM)

Fig. 1. Kinetics of PE and TG hydrolysis in HDL, (1A) and HDL, (1B) subfractions as mediated by H-TGL. Saturation kinetics for HDL, (1A). doubly Iabelled with [3H]triacylgIycerol (TG, n -W) and [‘4C]phosphatidylethanolamine (PE. 13-0) were followed using 65 mIU/ml enzymatic activity during 60 min incubations, in a range of 0.48 to 1.8 mM HDL-phospholipids corresponding to 0.024-0.089 mM HDL-PE and 0.021-0.077 mM HDL-TG. HDL, (lB), doubly labelled likewise were incubated in a concentration range of 0.48 to 2.4 mM phospholipids, corresponding to 0.022-0.110 mM HDL-PE and 0.021-0.107 mM HDL-TG with H-TGL (45 mIU/ml) during 60 min. Inserts display the corresponding double reciprocal-plots.

TABLE

I

Kinetic analysis of the hydrolysis ofHDL-PE K, and V, corresponding

and HDL-TG

by H-TGL

values are means (i-SE.) from n experiments. to 0.005-0.107 mM HDL-TG and to 0.006-0.118

Substrates

The concentration mM HDL-PE. V,,,

ranged

between

is calculated

0.12

and

for 200 mIU/ml

2.4 mM hepatic

HDL

n

K, (S mM_‘)

Vrnm (nmol S hydrolysed &‘.h-‘)

Vk, /Km

HDL, [3H]TG [14C]PE [‘QPC *

2 4 6

0.025 (0.020-0.030) 0.056 ( f 0.014) 0.318 (*0.052)

179.50 (136-223) 293.50 (+ 17.1) 41.10 ( * 6.6)

7180 5241 129

HDL, [3H]TG [14C]PE [‘V]PC *

4 4 6

0.037 (+ 0.008) 0.061 ( f 0.012) 0.430 ( f 0.037)

68.78 ( f 24.0) 178.90 (k27.3) 28.60 (f 1.2)

1859 2933 67

(S)

* Taken from Simard

phospholipids,

lipase activity.

et al [12].

m-m,, TG hydrolysed

ml-’

h -’

“mo,

TG hydrolysed

ml-’

-h-l

Fig. 2. Comparison between the amounts of PE and TG hydrolysed by the action of H-TGL in HDL, (2A) and HDL, (2B), respectively. Doubly labelled HDL, [3H]TG and [14C]PE, in a concentration range of 0.12-2.4 mM HDL phospholipids, corresponding to 0.006-0.118 mM HDL,-PE, 0.0055-0.110 mM HDLs-PE and 0.0052-0.103 mM HDL,-TG, 0.0053-0.107 mM HDL,-TG, were incubated with 15-65 mIU/mI hepatic lipase activity for 60 min. Comparative hydrolyses of both substrates are from individual points, taken from two (in the case of HDL,) and four (HDL,) different experiments. Linear regression factors were 0.992 (A) and 0.995 (B).

77 TABLE

II

HDL,-PE

and HDL,-TG

hydrolysis by H-TGL as a function of the free cholesterol/phospholipid

molar ratio

Free cholesterol was incorpored into HDL, by incubating the particles with cholesterol-laden celite for 8 or 16 h. Control particles were incubated with unmodified celite for 8 h. Data are from four experiments and represent the rate of PE and TG hydrolysis (PM for 60 min incubation) at two concentrations of HDL phospholipids: 0.72 mM and 1.5 mM, using 45 mIU/mf hepatic triacylglycerol lipase activity. Statistical significance referred to control-HDL,: a, P < 0.05; b, P < 0.01; c, P -C0.001. FC/PL

HDL, (n=4)

control-HDL,

molar ratio

1.5 mM HDL phospholipids

0.72 mM HDL phosphohpids

PE hydrolysis

TG hydrolysis

(pM/60

(pM/60

mm)

min)

(nM)/60

4.22 f 0.46

0.103 f 0.021

3.85 f 0.18

3.16 f 0.41

6.65 i 0.25

Chol-celite

8h

0.174 + 0.014

5.02 + 0.05 ’

4.13 f0.16

9.04kO.40

chol-celite

16 h

0.221 f 0.011

4.28 f 0.21

2.93 + 0.17

7.5OkO.41

parable within each subfraction (Table I). The radioactivity lost from [3H]TG was recovered for 81.5 f 1.9% as H-labelled free fatty acids, for 15.9 + 2.1% as [3H]monoacylglycerols and for 3.9 k 2.6% as [3H]diacylglycerols. Regarding the lipolysis products of [14C]PE, 2-[‘4C]acyl-sn-glycerophosphoethanolamine accounted for 83.1 & 1.4% and free fatty acids for 16.8 f 1.4%. II. Comparative hydrolysis of PE and TG in HDL particles The respective degradations of phosphatidylethanolamine and triacylglycerol were matched together, scanning various degrees of lipolysis at different substrate concentrations (Figs. 2A and B). In both HDL, and HDL,, a linear relationship was observed between the two processes. With equivalent amounts of the two substrates in HDL particles, the hydrolysis of PE exceeded that of TG by 30% in HDL, and by 75% in HDL,. Hence, despite a lower affinity, more molecules of phosphatidylethanolamine than of triacylglycerol undergo hydrolysis and this is illustrated by the differences in the estimated V,, (Table I). III. Modulations of PE and TG hydrolysis in HDL by changes in the surface or core lipid composition As the phospholipase and triacylglycerol lipase activities of H-TGL are directed against two different domains of the HDL structure, attempts were made to modify either the particles surface or the lipid core, and to check thereafter the HDL behaviour towards the two activities. Modification of HDL, in the presence of cholesterol-laden Celite led to a selective enrichment in free cholesterol: the free cholesterol/protein ratio was raised from 0.055 pmol/mg + 0.011 in control particles up to 0.109 pmol/mg * 0.010 (P < 0.02). Conversely, the phospholipid content was about 10% lowered during the procedure (from 0.536 pmol/mg + 0.005 to 0.493 pmol/mg + 0.026, NS). Hence, the FC/PL molar ratio was doubled during the particle surface modification

TG hydrolysis

PE hydrolysis (pM/60 min)

b

mm)

6.72 + 0.90 a 5.66 f 0.46

(Table II). On the other hand, the relative proportions of esterified cholesterol and triacylglycerol were almost kept constant. Reisolation of the lipoprotein particles by gel-filtration showed similar elution patterns for control and cholesterol-enriched HDL. Notably, the peak elution volume was unchanged. Cholesterol enrichment had a biphasic effect on HDL-TG and HDL-PE reactivities towards hepatic lipase. With a 70% increase in the free cholesterol/ phospholipid ratio, there was a 30-608 stimulation in the hydrolysis of both substrates. This was significant compared to control particles, at two different HDL concentrations (Table II). This effects was partly reverted at higher levels of cholesterol enrichment (115%) so that the stimulation for the most sterol-enriched particles was minimal (O-358, NS). For comparison, the degradation of phosphatidylcholine induced by HTGL (200 mIU/mL) was chemically measured in control particles and in the 8 h-Celite/cholesterol-incubated HDL. The latter displayed a 37% higher rate of PC hydrolysis, in agreement with previous observations [12]. Further cholesterol enrichment of the HDL, rising the free sterol/phospholipid ratio up to 0.71 mol/mol resulted in no change of the PE or TG hydrolysis values, compared to control particles (not shown). HDL, were also modified in their core lipid composition by incubation with an excess of VLDL and in presence or absence of a source of cholesteryl ester transfer activity. The latter treatment little affected the HDL surface composition but there were reciprocal changes in the respective ratios of esterified cholesterol/ protein and triacylglycerol,’ protein: from, 0.587 to 0.433 and from 0.114 to 0.232, respectively. Hence, particles could be arranged as a function of their relative proportions of the two core components (Table III). As described in earlier studies, triacylglycerol enrichment in HDL, leads to a slight shift in the density distribution of the particles, which thus sediment at a position intermediate between HDL, and HDL, [12]. The values of TG hydrolysis increased proportionally

78 TABLE

III

Modulations of the hydrolysis of HDL, -PE and HDL, -TG mediated by H-TGL by changes in the lipid core Particles were modified by incubation in presence of VLDL and with/without a cholesteryl ester transfer activity containing plasma fraction, (Post-VLDL-HDL, and CET-HDL,, respectively). Data are from triplicate assays in one experiment (means f S.E.) and represent the PE and TG hydrolysis measured after 60 min incubations with 45 mIU/ml H-TGL activity, using 0.72 mM HDL phospholipids. Statistical significance referred to control HDL,: a, P i 0.01; b, P i 0.001. TG/CE molar ratio

PE hydrolysis yM/60 mm

0.072

4.02 f 0.20

(n=3)

0.194

4.44kO.13

12.67 f 1.03 b

CET-HDL, (n = 3)

0.535

4.43 + 0.17

22.44k

Control

TG hydrolysis pM/60 min

HDL,

(n=5)

3.31 kO.25

Post-VLDL-HDL,

5.13 =

with the particles content in this substrate. By contrast, the rate of PE hydrolysis remained stable and was unaffected by the core lipid modification. Discussion In agreement with other studies using artificial models, phosphatidylethanolamine and triacylglycerol present in the composition of human HDL appear as preferential substrates for hepatic lipase, and both molecules display comparable catalytic efficiencies. Modulations of the free cholesterol content of the particle outershell can enhance the HDL reactivity towards both PE-phospholipase and triacylglycerol hydrolase activities and this corroborates earlier findings on the hydrolysis of phosphatidylcholine by hepatic lipase [12]. On the other hand, changes in the HDL lipid core, affected only the level of TG hydrolysis, the latter being roughly proportional to the amount of triacylglycerol present in the core composition. The relative degradation rates for both substrates appear very comparable. As the two molecules contribute in almost similar proportions to the particles composition (TG/total phospholipids: 4.3-4.45 . lop2 mol/mol and PE/total phospholipid 4.6-4.9. lop2 mol/mol in HDL subfractions), the amounts of HDLphosphatidylethanolamine and triacylglycerol degraded in vitro by action of hepatic lipase is of the same order of magnitude. However, in terms of free fatty acid production, 2 to 3-times more would be generated from TG than from PE. We had previously observed that the relative rates of TG hydrolysis were lo-times more elevated than those of phosphatidylcholine [12]. Consistent with that observation, the calculated K, for phosphatidylethanolamine and triacylglycerol are lo-fold lower than the K,

for HDL-phosphatidylcholine, and the catalytic efficiencies, thus appear multiplied by 20-30-fold for the two former substrates compared to phosphatidylcholine (5241 and 2933 for HDL, and HDL,-PE, 7180 and 1859 for HDL, and HDL,-TG, versus 129 and 67 for HDL, and HDL,-PC, respectively) [12]. Kinetics of HDL-TG hydrolysis have been already studied by Shinomiya et al [7], who reported higher K, values (0.18 mM), yet the calculations of V,,,/K, are very close to our own estimates. Although hepatic lipase is supposed to act at the amphipathic outer surface of the lipoprotein particle, K, values were calculated from the total TG concentration since, in a previous study [12], all TG molecules appeared available for hydrolysis. Thus it is likely that, as the low amounts of surface TG get progressively degraded, they are replaced by TG molecules coming from the inner core, most of the lipolysis products being trapped by albumin. With all the reservations about applying Michaelis constants to lipolytic enzymes acting at a lipid/water interface, it is, however, possible to estimate a particle’s K, towards hepatic lipase. This was calculated from the hydrolysis data of the three substrates considered, the values for the hydrolysis of phosphatidylcholine being taken from two different studies [6,12], and from their relative contribution to the HDL chemical composition. For instance, considering an average molecular weight of 3.5-4.0. lo2 kDa for HDL, [24], the estimated particle’s K, would be 2.5-3.5 . 10e6 M based on phosphatidylcholine degradation [6,12], 3.8-4.5 . lop6 M based on TG and 7.0-8.2. lop6 M from PE hydrolysis. Hence, all values are in a narrow range, and the better K, obtained from phosphatidylcholine degradation could be explained by the overwhelming predominence of lecithin, relative to the other substrates, in HDL composition. The present comparison of three HDL substrates towards hepatic lipase is in agreement with other studies on artificial systems [8,9,11,25]. In monolayer models, the rates of hydrolysis for phosphatidylethanolamine and triacylglycerol were comparable and exceeded by 30-fold those measured on phosphatidylcholine [ll]. As well in emulsions, the catalytic efficiencies appeared better for phosphatidylethanolamine than for phosphatidylcholine species, yet increasing concentrations of the latter could inhibit the degradation of phosphatidylethanolamine [25]. As already described for phosphatidylcholine [6,12], HDL, appears much more reactive than HDL, towards the hepatic lipase-mediated hydrolysis of triacylglycerol and phosphatidylethanolamine. This cannot be explained by changes in the relative proportions of the three potentially competing substrates. which are almost identical in the two subfractions considered. Rather, differences in reactivity might be related to the size of the particles, to the proportion of lipid molecules per

79 apo A-I, or, as recently suggested, by their free cholesterol to phospholipid molar ratio [12]. Indeed, a moderate cholesterol enrichment of the substrate HDL led to a 30-60% stimulation in the hydrolysis of both phosphatidylethanolamine and triacylglycerol. A similar behaviour of hepatic lipase was already reported for the TG degradation rate in a monolayer system, where the composition in polar lipids was varied [26] and for the PC hydrolysis in native and modified HDL [12]. As well, a strong stimulation of the HDL reactivity towards H-TGL-mediated phospholipid hydrolysis was noted following lipolysis of triacylglycerol-rich lipoproteins by lipoprotein lipase, a process which brings extra surface components to the HDL outershell [12,27,28]. Hence, modifications of the HDL enveloppe, seem to modulate the reactivities of the three substrates towards hepatic lipase, either through variations in the free cholesterol/phospholipid molar ratio, or more generally, through changes in the physical state and surface pressure of the lipoprotein monolayer. Since the hydrolyses of the three potential substrates present in HDL were stimulated, it is likely that the affinity of the whole HDL particle for hepatic lipase is increased upon cholesterol enrichment, as it was suggested before for HDL modified by the lipolysis of triacylglycerol-rich lipoproteins [12]. Alternatively, added cholesterol may affect the solubility of the different substrate in the HDL surface [29], and thus modify their accessibility to hepatic lipase. A different behaviour was observed in the HDL modified by the neutral-lipid-transfer reaction. As described in many previous studies [12,30-321, the modified particles display lower densities due to the replacement of cholesteryl ester molecules by triacylglycerol. The latter can thus contribute up to 35% of the apolar core instead of 6.7% in native particles. Phospholipid hydrolysis by hepatic lipase seems unaffected by the neutral-lipid-transfer-modifications, as is evident for phosphatidylethanolamine in the present study and for phosphatidylcholine in a previous report [12]. Conversely, the rate of TG hydrolysis appears proportional to the degree of enrichment in this substrate. Hence, it is likely that the hydrolysis of HDL triacylglycerol by hepatic lipase may depend on two factors: the affinity of the HDL particle for the enzyme, as estimated here for native HDL, and the relative wealth of the particle in triacylglycerol. However, one cannot exclude that the rate of TG hydrolysis could be influenced by slight changes in the physicochemical state of the modified HDL (decreased density, increased size.. .). For instance, large differences in the TG degradation rates are evident comparing the reactivities of HDL, and HDL,. The comparison of the different substrates for hepatic lipase in native and surface/core modified HDL may contribute to the still-unanswered debate: which one of the triacylglycerol or phospholipase activities of the

enzyme is predominant on HDL particles. Ultimately, this question will depend on the relative content of the particles in triacylglycerol, itself related to the competition between lipoprotein lipase and the cholesteryl ester transfer protein for the handling of VLDL and chylomicron triacylglycerol. Lipolysis of TG-rich lipoproteins will essentially bring surface components to the HDL system [28] and will enhance the particles reactivity towards the phospholipase action of H-TGL [12]. On the other hand, long persistence of TG-rich lipoproteins would favour the neutral-lipid-transfer to HDL, gradually enriched in triacylglycerol. The latter being a preferential substrate for H-TGL, the process would end up in the formation of small-sized HDL [24,30]. In addition, the very efficient clearance by H-TGL of HDL triacylglycerol, the latter deriving mostly from very-low-density lipoproteins and chylomicrons, indicates that the liver may be a major site of triacylglycerol lipolysis. The similar reactivity towards hepatic lipase of PE, a minor phospholipid class in HDL, and the existence in plasma of phospholipid exchange proteins [33], suggest that phosphatidylethanolamine could be also rapidly turned over. The behaviour of those different substrates towards a membrane-bound hepatic lipase, in a more physiological situation, needs further investigations. In conclusion, triacylglycerol and phosphatidylethanolamine in HDL are hydrolysed by hepatic lipase at comparable rates in vitro. Both appear as preferential substrates compared to phosphatidylcholine. Calculations on a particle basis suggest that the enzyme might interact with a whole HDL particle and then hydrolyse the various substrates according to their respective proportions in HDL. Moreover, the reactivities towards the phospholipase and triacylglycerol lipase activities of H-TGL appear differentially affected by changes in the particle’s outershell or in the inner core. Acknowledgment We thank tance.

M.P.

Branchu

for her

secretarial

assis-

References 1 Kuusi. T., Kirmunen, P.K.J. and Nikkila, E.A. (1979) FEBS Lett. 104, 384-388. 2 Kuusi, T., Saarinen, P. and Nikkila, E.A. (1980) Atherosclerosis 36. 589-593. 3 Jansen, H., Van Tol, A. and Hulsmann, W.C. (1980) B&hem. Biophys. Res. Commun. 92, 53-59. 4 Murase, T. and Itakura, H. (1981) Atherosclerosis 39, 293-300. 5 Daggy, B.P. and Bensadoun, A. (1986) B&him. Biophys. Acta 877, 252-261. 6 Shirai, K., Barnhart, R.L. and Jackson, R.L. (1981) B&hem. Biophys. Res. Commun. 100, 591-599. 7 Shinomiya, M., Sasaki, N., Barnhart, R.L., Shirai, K. and Jackson, R.L. (1982) Biochim. Biophys. Acta 713, 292-299.

80 8 Miller, C.H., Parce, J.W., Sisson, P. and Waite, M. (1981) Biochim. Biophys. Acta 665, 385-392. 9 Belcher, J.D., Sisson, P.J. and Waite. M. (1985) Biochem. J. 229, 343-351. 10 Jackson. R.L., Ponce, E., McLean, L.R. and Demel, R.A. (1986) Biochemistry 25, 1166-1170. 11 Laboda, H.M.. Ghck, J.M. and Phillips, M.C. (1986) Biochim. Biophys. Acta 876. 233-242. 12 Simard, G., Perret, B., Durand, S., Collet, X., Chap. H. and Douste-Blazy, L. (1989) Biochim. Biophys. Acta 1001, 225-233. 13 Landin, B., Nilsson. A.. Twu. J.S. and Schotz, M.C. (1984) J. Lipid Res. 25. 559-563. 14 Van Tol. A., Van Gent, T. and Jansen, H. (1980) Biochem. Biophys. Res. Commun. 94, 101-108. 15 Perret. B.P., Chollet, F., Durand, S., Simard, G., Chap, H. and Douste-Blazy, L. (1987) Eur. J. Biochem. 162, 279-286. 16 Nilson-Ehle. P. and Schotz, M.C. (1976) J. Lipid. Res. 17. 5366541. 17 Avigan, J. (1958) J. Biol. Chem. 234, 787-790. 18 Bligh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37, 911-917. 19 Skipski. V.P., Peterson, F.F. and Barclay, M. (1964) Biochem. J. 90, 374-378. 20 Lowry, O.H.. Rosebrough. N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193. 265-275.

21 22 23 24 25 26 27 28 29 30

31 32

Roschlau, P., Bernet. E. and Gruber, N. (1974) 2. Klin. Biochem. 12,404-407. Bottcher, C.J.F., Van Gent, C.M. and Pries. C. (1961) Anal. Chim. Acta 24, 203-204. Wahefeld, A.W. (1974) in Methoden der Enzymodischen Analyse (Bergmeyer, H.D., ed.), tome II, p. 178, Verlag Chemie, Weinhem. Eisenberg, S. (1984) J. Lipid. Res. 25, 1017-1054. Kucera, G.L.. Sisson, P.J.. Thomas, M.J. and Waite, M. (1988) B&him. Biophys. Acta 665, 1920-1928. Laboda, H.M.. Glick, J.M. and Phillips, M.C. (1988) Biochemistry 27, 2313-2319. Patsch, J.R.. Gotto, A.M.. Olivecrona, T. and Eisenberg. S. (1978) Proc. Nat]. Acad. Sci. USA 75, 4519-4523 Eisenberg, S. and Olivecrona. T. (1979) J. Lipid Res. 20, 614-623. Small. D.M. (1986) in Physical Chemistry of Lipids (D.J. Hanahan, ed.), pp. 345-394, Plenum Press, New York. Deckelbaum, R.J., Eisenberg, S., Oschry. Y., Granot, E., Sharon, I. and Bengtsson-Olivecrona, G. (1986) J. Biol. Chem. 261, 52015208. Patsch. J.R., Prasad, S.. Gotto, A.M. and Olivecrona. T. (1984) J. Clin. Invest. 74, 2017-2023. Simard, G., Loiseau. D., Girault, A. and Perret, B. (1989) Biochim. Biophys. Acta 1005, 245-252.