81
Eiochimica et Biophysics Acta, 1046 (1990) 81-88
Elsevier BBALIP 53465
Effect of apoprotein cross-linking on the metabolism of human HDL, in rat Claude Senault ‘, Florence H. Mahlberg 2, Guy Renaud I, Anik Girard-Globa and George K. Chacko 2
’
’ Unit&INSERM 286, Fact& de Mkdecine Xavier-&hat,
Paris (France) and 2 Department of Physiology and Biochemistry, Medical College of Pen~s~~lvania,P~ilade~hia, PA (U.S.A.)
(Received 19 March 1990)
Key words: HDL; HDL receptor; Apoprotein catabolism; Cholesteryl ester selective uptake
Apo E-free human high-densi~ Ii~protein (HDL,) was i&&d with ‘=I in apoprotein and with 3H in choIestery1 linoleyl ether (a non-hydrolyzable analogue of eholesteryl ester). The labeled HDL, was modified by cross-liuking of apoproteins with dimethylsuberimidate (DMS) to inhibit binding to HDL specific receptors. The control and the DMS HDL, were characterized with respect to their rate of clearance from rat blood, in vivo binding to major rat organs and in vitro binding to purified rat liver lasma membranes. Both 12’1 and 3H labels from control HDL, were cleared from Q rat blood mon~x~nenti~iy, but H at a faster rate than “‘1 (3H fl,2 = 3.0-4.1 h; ‘“I t,,:, = 7.0-7.7 h). This difference is consistent with reports of the nonendocytotic selective uptake of HDL-associated cholesteryl ester. DMS modification did not affect the rate of 3H clearance whereas it increased the rate of “‘1 clearance (HDL, z,,, = 7.7 h, DMS HDL, t,,, = 4.1 h). Both in vivo binding to rat organs and in vitro binding to rat liver membranes confirmed that DMS modification inhibited the specific binding of HDL, but also suggested that the modification produced saturable binding of HDL to a separate class of sites. Thus, the present data do not rule out the involvement of direct HDGcell intention in the selective uptake of HDL eholesteryl ester. However, results suggest that the binding of HDL to its specific cell surface sites is not necessary for this uptake.
Introduction
Apo E-free ~~-density lipoprotein (HDL) binds to a variety of cell surfaces and isolated membranes, with properties typical for a l&and-receptor type interaction [1,2]. A membrane-associated HDL binding protein has been identified by ligand blotting [3]. However, very little is known about the role of this putative HDL receptor. Some of the postulated functions of HDL binding sites include: (1) efflux of unesterified cholesterol from peripheral cells [4-61; (2) intemalization and catabolism of HDL in the liver [7,8]; and (3) preferential delivery of free and esterified cholesterol to liver and steroidogenic tissues without apoprotein de-
Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; DMS, dimethylsuberimidate; CEt, cholesteryl ether; TNM, te~~trome~ane; SDS-PAGE, sodium dodecyl sulfate-polyacryiamide gel electrophoresis. Correspondence: G.K. Chacko, Department of Physiology and Biochemistry, Medical College of Pennsylvania, Philadelphia, PA 19129, U.S.A. US-27~/~/$03.50
gradation [9-121. It has also been suggested that the mitogenic effect of HDL on certain cells is mediated through the NDL binding sites [331. To investigate the possible role of HDL binding sites in various processes, we sought to inhibit HDL binding by chemical modification. Thus, we have previously shown that treatment with tetranitromethane (TNM) or cross-linking of apoproteins by ~methylsube~~date (DMS) inhibited the binding of HDL to its specific binding sites on isolated rat liver membranes and on cells in culture [14,15]. Using these chemically modified HDL preparations we have shown recently that, at least in human skin fibroblasts and in rat Fu5AH hepatoma cells, HDL binding is not a prerequisite for either the efflux or influx of unesterified cholesterol between cells and lipoproteins [16,17]. In the present report, we have used DMS modification of HDL to probe the functional role of the HDL binding sites in the in viva catabolism of HDL in rats. The apoprotein in human HDL, was labeled with i2? and the cholesteryl ester moiety was traced with [3Hfcholesteryl linoleyl ether (a non-hydrolyzable analogue of cholesteryl ester). DMS modification of human HDL, did not decrease but
@ 1990 Eisevier Science Publishers B.V. (Biomedical Division)
82 rather increased the rate of apoprotein clearance from the blood; clearance of cholesteryl linoleyl ether was not affected. Material and Methods Rat serum albumin, bovine serum albumin (BSA) fatty acid free and dimethylsuberimidate, (DMS) were from Sigma (St. Louis, MO). [1,2-3H]Cholesteryl linoleyl ether (spec. act. 31 Ci/mmol) was purchased from Amersham, France. All other chemicals were of reagent grade. Lipoprotein preparations Human HDL, (1.125 1.215) that had been prelabeled with [3H]cholesteryl linoleyl ether, followed by reisolation of the labeled lipoprotein by ultracentrifugation. The labeled lipoprotein was analyzed in triplicate by agarose gel electrophoresis (barbital buffer, pH 8.0). One gel was stained with Oil-Red-O and scanned by densitometry. The others were cut into 5 mm sections and analyzed for radioactivity: directly for 125I label and after extraction by the procedure of Dole [22] for 3H label. The initial preparation showed, in addition to HDL-associated radioactivity, another radioactive component that barely moved from the origin of the agarose gels. This contaminant was removed by gel filtration on a column of 0.5 M Bio-gel Agarose (Bio-Rad), eluting with 0.15 M NaCl containing 0.02% EDTA (pH 7.4). The purified sample exhibited a single radioactive band which co-migrated with control HDL,. Its specific activity was 16 dpm per ng cholesteryl ester. Similarly, doubl? labeled HDL,, with “‘1 in the apoprotein and H in the cholesteryl ether (‘251(apo)[ 3H](CEt)-HDL,) was prepared from 12?-HDL3 and [3H]cholesteryl linoleyl ether using the procedure of Roberts et al. [21], followed by gel filtration chromatog-
raphy on a 0.5 M Bio-gel Agarose column. DMS modified HDL, was prepared as previously described [15]. Briefly, HDL, (2 mg protein/ml) was reacted with DMS (2 mg/ml) in 0.09 M triethanolamine buffer (pH 9.6) for 2 h at room temperature. After incubation, the reaction mixture was applied to a column of Sephadex G-50 and the lipoprotein was eluted with 0.15 M NaCl. containing 0.01% EDTA (pH 7.4). Fractions containing protein were combined to obtain DMS HDL,. SDSPAGE analysis followed by Coomassie blue staining revealed extensive cross-linking of the apoproteins. The cross-linked apoproteins, seen as a high molecular weight band with a mean molecular weight of 95000 on a 3-27% gradient slab gel represented more than 80% of the Coomassie blue staining components of DMS HDL,
WI. Membrane preparations Rat liver plasma membranes were isolated using the Percoll gradient centrifugation procedure of PastorAnglada et al. [23]. The 5’-nucleotidase activity of the purified membranes, assayed according to Grunnet and Bogesen [24] was more than lo-times that of the liver homogenate.
Binding of labeled liEproteins to isolated membranes The binding of I-HDL, and iz51-DMS HDL, to isolated rat liver plasma membranes was determined according to the procedure of Chacko [25]. Briefly, aliquots of membranes (200 pg protein) were incubated with varying concentrations of labeled lipoprotein at room temperature for 1 h in a total volume of 0.2 ml, containing 0.150 M NaCl, 0.5 mM CaCl,, 10 mM Tris-HCl (pH 7.4) and 1% bovine serum albumin. After incubation, 0.175-ml aliquots of incubation mixture were centrifuged in a Beckman 42.2 Ti rotor at 30000 rpm for 15 min to recover the membranes. The membrane pellets were washed once with 0.175 ml of incubation medium and the tubes containing the membranes were assayed for radioactivity in a gamma scintillation spectrophotometer. Non-specific binding of labeled lipoprotein was determined in samples run in parallel that also contained a lOO-fold excess of the corresponding unlabeled lipoprotein. The difference between the ‘251lipoprotein bound to the membrane in the absence and presence of excess unlabeled lipoprotein was taken as the amount of specific binding. In order to determine the selective uptake of cholesteryl ester, 200~pg protein aliquots of liver plasma membranes were incubated with identical concentration of ‘251-HDL3 and [ 3H](CEt)HDL, (10 pg protein/ml) as above and the amount of i2’I and 3H radioactivity associated with the washed membranes were determined, “‘1 directly by gamma counting and 3H, after extraction [22], by liquid scintillation counting.
83 In vivo metabolic studies
Male Sprague-Dawley rats (IFFA CREDO France) weighing 225-250 g were used. They were fed standard rat chow and were fasted overnight prior to metabolic studies. Radiolabeled HDL, was injected into the femoral vein under light ether anesthesia (1251, 4. lo6 cpm/animal; 3H, 0.3. lo6 cpm/animal). 3 min after injection, a small sample of blood was obtained from the orbital plexus for determination of the initial plasma radioactivity. Blood samples of about 150 ~1 were then collected at the specified times from the tail vein, allowed to clot on ice for 15 min and centrifuged to obtain the serum. Serum was analyzed for 1251radioactivity by counting aliquots in a gamma counter. TCAsoluble counts were subtracted from the total 1251counts. For 3H radioactivity determination, lipids were extracted according to Dole [22] and the extract analyzed by liquid scintillation counting. Serum decay data were fitted to a monoexponential function for calculation of half time (t,,*) using regression line analysis. The tissue-space distribution of labeled lipoproteins and rat serum albumin in rat was evaluated according to the procedure of Koelz et al. [26]. The labeled proteins were injected into the femoral vein under light ether anesthesia. After 10 min, the animal was anesthetized and blood was collected from the abdominal aorta. Tissues were removed and weighed. They were cut into small pieces, carefully rinsed in ice-cold 0.15 M NaCl and analyzed either directly for “‘1 or after Folch extraction [27] for ‘H radioactivity. Tissue-space, which represents the volume (~1) of plasma cleared per gram of tissue, was calculated using the equation: tissue space = tissue radioactivity (cpm/g)/plasma radioactivity (cpm/pl) (261. Protein was determined by the method of Peterson [28] using bovine serum albumin as the standard. Total lipids from lipoprotein preparations were extracted according to the procedure of Folch et al. [27]. Lipids were assayed enzymatically using Wako kits for phospholipids (Biolyon, France) and Boehringer kits for total and unesterified cholesterol (Boehringer Mannheim, France). Statistical analysis The values presented
are means f S.E. The significance of differences was tested by Student’s t-test. Results fi$2ct of DMS modification on serum decay of human I-HDL,
When trace amounts of native ‘*?-HDL,, were injected intravenously into rats and the 1251counts in the serum were measured with time, the TCA-precipitable counts disappeared from the serum monoexponentially with a half-time (t1,2) of 7.7 f 0.5 h (Fig. 1). This value
100
1
2
3
4
5
6 Time
7
a
hours)
1. Effect of DMS modification on the rate of disappearance of label from rat blood following injection of radiolabeled human HDL, (4.106 cpm/animal): n -w, ‘251-HDL,; A-A, ‘*‘I-DMS 12?-acetyl HDL,. At the indicated times after HDL,; o-o, injection, blood was collected and serum was separated and analyzed for TCA-precipitable counts. Values are expressed as ‘k; of the initial radioactivity (to) obtained from the blood collected at 3 min after injection from the orbital plexus. Values are means from 20 animals preeach for ‘*‘I-HDL, and ‘*‘I-DMS HDL, and five lipoprotein parations. The data for ‘*‘I-acetyl HDL, were mean values from three animals with one lipoprotein preparation. t,,* values for 1251-HDLs, 4.1 kO.6 and ‘*‘I-DMS HDL, and 1251-acetyl HDL, were 7.7+0.5, 0.1 h, respectively.
is slightly higher than the value reported by Van To1 et al. ~291 (t,,, = 6.2 h) but lower than that reported by Roheim et al. [30] (tl,* = 10.5 h), both studies using rat HDL in rat. DMS modified ‘*?-HDL,, was also cleared monoexponentially from the blood (Fig. l), but at a significantly faster rate than control HDL, (t,,, = 4.14 f 0.6 h). This increased turnover of ‘*‘I-DMS HDL, was found in five different lipoprotein preparations and in a total of 20 rats. In order to investigate whether this increased rate of clearance might be related to the recently reported uptake mechanism for acetylated HDL, [31], a sample of ‘251-HDL, was acetylated and its clearance from the blood was studied. As shown in Fig. 1, acetylated ‘*‘I-HDLs was removed from the rat blood with a t,,, of about 0.1 h, much faster than that of I*?-DMS HDL,. In vivo binding of ‘2sI-HDL,
and 12’I-DMS HDL,
In order to identify the organs responsible for the increased uptake, and thus the faster disappearance from the blood of 125I-DMS HDL,, the in vivo binding of native and DMS modified HDL, to various organs of rat was determined 10 min after injection, using the tissue space technique [26] (Table I). The specificity of binding was assessed by using “‘1 rat serum albumin as a non-specific marker of tissue space .[26]. All tissues examined were found to bind ‘251-HDL3; however when these binding values were corrected for non-specific binding by subtracting the 1251rat serum albumin space, only liver and adrenal tissues showed a specific binding. The tissue space for native was highest in the adrenal. For “‘1-DMS HDL, the
tissue spaces were similar in all organs except the liver and, to a smaller extent, the spleen. The liver tissue space was significantly larger for ‘*‘I-DMS HDL, than for native HDL, (P < O.OS), while adrenal space was unchanged. A small specific compartment also appeared in the spleen. Binding of ‘751-HDL, and “‘I-DMS HDL, to rat liver plasma membranes The binding of native and DMS-modified HDL, to purified rat liver plasma membranes was studied. In Fig. 2 is shown the binding of ‘251-HDL, (A) and of 12’I-DMS HDL, (B) to membranes in the absence and the presence of an excess of the respective unlabeled HDL,. The specific binding was saturable for both the native and DMS-modified HDL,. Saturation of binding sites occurred at an HDL concentration less than 100 pg HDL protein/ml. Scatchard analysis of the binding data gave similar Kd (26 pg/ml for native HDL, and 24 pg/ml for DMS HDL,); there was also no significant difference in the maximum binding capacity, B,,, (Table II). In order to determine whether the native and DMS-modified HDL, are binding to the same or different binding sites, competition experiments were performed. In Fig. 3 (A) is compared the ability of unlabeled native and DMS HDL3 to compete with 12’1HDL, for its binding sites. Whereas the native HDL, was very effective in competing with ‘251-HDL , DMSmodified HDL, was not. Similarly, when %DMS HDL, was used as the ligand (Fig. 3 (B)) native HDL, was less effective than the DMS-modified HDL,, especially at low concentration, to inhibit the binding of the labeled lipoprotein; thus, at 10 pg protein, the inhibitions by native and DMS-modified HDL, were 15 and 55%, respectively. These results would suggest that, although both native and DMS-modified HDL, bind to isolated rat liver membranes with apparent similar binding properties, they are binding to different sites.
1251-mX3, pgglml
1251-DMS HDL-~~,pglrnl Fig. 2. Binding of 1251-HDL, (A) and tz51-DMS HDL, (B) to rat liver plasma membranes as a function of lipoprotein concentration. Aliquots of membranes (200 ng protein) were incubated with the indicated concentration of labeled lipoprotein in 0.2 ml of incubation medium containing 0.15 M NaCl, 0.5 mM CaCl,, 10 mM Tris-HCl (pH 7.4) and 1% bovine serum albumin for 1 h at room temperature in the absence (+) and presence (0) of lOO-times excess of the corresponding unlabeled lipoprotein. After incubation, the amount of labeled lipoprotein bound to the membranes was determined as described in Materials and Methods. The amount of labeled lipoprotein bound specifically to the membranes (A) was obtained by subtracting the labeled lipoprotein bound to the membrane in the presence of excess unlabeled lipoprotein from that bound in its absence. The data in the figure are from one representative of four separate binding experiments.
TABLE Effect
Effect of DAYS modification on the preferential clearance of HDL [-‘H]cholesteryl linoleyl ether Using doubly labeled lipoprotein preparations it has been shown in rats [11,12] that the core cholesteryl ester
I
of DMS modification on the in vivo binding of “‘I HDL, to major organs of ran
In vivo binding, expressed as tissue space (~1 plasma/g of rats in each group is shown in parentheses.
tissue) was calculated
as described
in Materials
and Methods.
Values are means f S.E.; NO.
Tissue space (PI/g) liver
kidney
spleen
adrenal 84k
testes
heart
6
21*6
56kll
iz51-rat serum albumin
(3)
70f
2
91*15
7X*9
12’1 HDL,
(7)
74k
8
64klO
61+6
238 + 38
9*4
4Oi
4
‘251-DMS HDL,
(6)
105*13
68klO
84+8
275+75
9*3
41k
6
85 TABLE
III
Comparison of the in vivo binding of “?-apoprotein
and [‘H]CEi
In viva binding expressed as tissue space (~1 plasma/g of rats in each group is shown in parentheses.
from “.‘I-HDL,
tissue) was calculated
and [‘H](CEt)-HDL,
as described
in Materials
respectively, to major organs of rat and Methods.
Values are means f SE.; No.
Tissue space (pi/g) liver “‘I-HDL3
(7)
74*
[ 3H](CEt)-HDL,
(6)
350*30
TABLE
8
kidney
spleen
64+10
611
64k
8
adrenal
testes
heart
38
9k4
40*
1273 f 123
15*4
6
122 + 49
238+
4
51*19
II
Comparison of the binding of “‘I-HDL, liver plasma membranes
and ‘251-DMS HDL,
to rat
Values are mean f S.E., from four separate binding experiments performed as described in the legend to Fig. 2. K, and B,,, values are derived from Scatchard analysis of the binding data.
K, (pg protein/ml)
%HDL,
‘251-DMS HDL,
26
24
+5
ol
kl
1
%,, (Bg protein/mg membrane protein)
2
3
4
5
6
Time (hours ) 1.7fO.3
2.4kO.7
of HDL is cleared faster than the surface apoprotein component. In order to assess the possible role of HDL binding in this preferential transfer process, we studied the effect of DMS modification on the clearance from rat plasma of doubly labeled HDL, with 1251 in apoprotein and 3H in cholesteryl linoleyl ether, a nonhydrolyzable tracer of cholesteryl ester. As is shown in Fig. 4, when the unmodified lipoprotein was used, [ 3H]cholesteryl ether was removed much faster than ‘*‘I apoprotein. The clearance of both components was
Fig. 4. Differential rate of disappearance of ‘25I and 3H labeled from (4.106 rat blood following injection of ‘25I (apo)-[3H](CEt)-HDLa At the indicated times after cpm ‘*‘I and 0.3.106 cpm ‘H/animal). injection, blood was collected and serum was separated and analyzed and ‘H (0 -0) counts as described in for i25I (M-w) legend to Fig. 1 and in Materials and Methods. Average of three experiments, each performed in three rats.
monoexponential and the 7.0 + 1 h, respectively. The major organs that ential uptake and thus the of [3H]cholesteryl ether,
half-times
were 3.0 f 0.8 and
were involved in the preferfaster clearance from plasma were determined by in vivo
DMS-HDL3
o\”
ol____ 510
25
OJ 50
75 w
UKl “Df-3
5K)25
I
1
1
,
50
75
loo
PQ ‘4”L3
Fig. 3. Effect of HDL, and DMS HDL, on the binding of ‘*‘I-HDL, (A) and lz5 I-DMS HDL, (B) to rat liver plasma membranes. Each tube contained 200 gg membrane protein, 10 pg protein/ml of ‘*‘I-HDL, (A) or 125I-DMS HDL, (B) and indicated concentrations of either HDL, 0) or DMS HDL, (O0). After incubation at room temperature for 1 h, the amount of ‘*‘I-HDL, (A) and ‘*‘I-DMS HDL, (B) (Obound to the membranes was determined as described in Materials and Methods. The data are average from a single experiment done in triplicate and are typical of results in three experiments.
86 TABLE
V
Effect of DMS [3H](CEi)-HDL,
f
‘“M
modification on the binding to rat liver plasma membranes
6
I of HDL, bound (‘25I or 3H) i2?-HDL,
“‘1-DMS
, 1
,
,
,
,
(
2
3
4
5
6
and
3H/‘251
2.15 +0.17
[3H](CEt)-HDL,
10-l
“51-HDL_q
Aliquots of membranes (200 pg protein) were incubated with 10 pg protein/ml each of labeled lipoprotein in 0.2 ml of incubation medium containing 0.15 M NaCl. 0.5 mM CaCI,, 10 mM Tris-HCI (pH 7.4) and 1% bovine serum albumin for 1 h at room temperature. After incubation, membranes were washed, recovered and analyzed for “‘1 and 3H counts (as described in Materials and Methods) to determine the amount of lipoprotein bound. Values represent percentage of total lipoprotein in the incubation medium bound to the membranes and are mean f SE. of three experiments.
Time (hours)
i
of
8.42 f 3.29
HDL,
[ 3H](CEt)-DMS
3.87+ 1.30
1.96f0.19 HDL,
8.53 f 2.51
4.45 f 0.70
Time (hours)
Fig. 5. Effect of DMS modification on the rate of disappearance of 12’1 (A) and ‘H (B) labels from rat blood following injection of ‘25I (apo)-[3H](CEt)-HDL3 (4. lo6 cpm, “‘1 and 0.3. lo6 cpm, 3H/animal). At the indicated times after injection, blood was collected and serum was separated and analyzed for i*‘I and 3H counts as described in legend to Fig. 1 and in Materials and Methods. (A) D-m, 1251, control lipoprotein; A-A, 1251, DMS modified 0, ‘H, control lipoprotein; l ----lipoprotein; (B) o0, 3H, DMS modified
lipoprotein. Average of three experiments, formed in three rats.
clearance, half-times were 6.8 + 1.1 and 3.5 k 0.9 h, respectively. It also appeared that in the case of DMSmodified HDL,, both 1251 and 3H labels were cleared at similar rates. This lack of involvement of specific HDL binding in the selective uptake of cholesteryl ether was further studied by examining the effect of DMS modification of labeled lipoproteins in the in vivo binding to major organs of rat and in vitro binding to isolated rat liver plasma membranes. As is shown in Table IV, DMS modification has no significant effect on the in vivo binding of [3H]CEt)-HDL3 to major organs of rat; the high tissue space of [3H]cholesteryl linoleyl ether to liver and adrenals remained unaffected after DMS modification of the lipoprotein. We also examined the preferential binding/uptake of 3H-labeled cholesteryl linoleyl ether by isolated rat liver plasma membranes. When isolated rat liver plasma membranes were incubated with equal concentration of [3H](CEt)-HDL3 or ‘251HDL, (10 pg protein/ml), relatively higher percentage of 3H label was taken up by the membranes than of the lz51 label (Table V); the 3H: 12’1 ratio for the unmodified lipoproteins was 3.9. This preferential binding/ uptake by the membranes was not affected by DMS modification of the lipoproteins. Selective uptake of HDL cholesteryl ester has been reported previously by
each per-
binding experiments. In Table III is shown a comparison of the in vivo binding (uptake) of 3H and iz51 labels from the labeled HDL, to major organs of the rat 10 min after intravenous injection of the lipoprotein. Liver and adrenal, and to a small extent spleen, showed 1 linoleyl ether over referential uptake of [ 3H]choleste P 35, I labels, expressed 251 apoprotein. The ratio of 3H to as space volume, were 4.6 for liver and 5.3 for adrenal. In Fig. 5 is compared the clearance of 125I (A) and 3H (B) labels from the doubly labeled HDL, before and after DMS modification of the lipoprotein. DMS modification of the lipoprotein had no significant effect on the rate of clearance of [3H]cholesteryl linoleyl ether, half-times were 3.7 f 0.6 and 3.0 h 0.5 h, respectively. However, in agreement with the data presented in Fig. 1. DMS modification increased the rate of apoprotein TABLE
IV
Effect of DMS modification
on the in viuo binding of [_‘HI(CEt)-HDL,
In vivo binding, expressed as Tissue space (cl plasma/g of rats in each group is shown in parentheses.
to major organs of rat
tissue) was calculated
as described
in Materials
and Methods.
Values are means f S.E.; No.
Tissue spaces (PI/g)
[ 3H](CEt)-HDL, [ ‘H](CEt)-DMS
HDL,
liver
kidney
(6)
350 f 30
64k
(6)
354*45
74k22
8
spleen
adrenal
testes
heart
122 f 49
1273k123
15*4
51*19
21Ok93
1017 f 217
15*1
49*11
87 Israeli et al. [32] in a bovine tion.
adrenal
membrane
prepara-
Discussion Unlike low-density lipoprotein (LDL), which is taken up by the apo B/E receptor pathway as a whole particle and degraded to its components in the lysosomes [33], the catabolism of HDL appears to involve more than one route [2]. At least four distinct pathways have been described for the catabolism of HDL: (1) internalization of the whole HDL particle followed by lysosomal degradation [7,8]; (2) preferential uptake of core cholesteryl ester with incomplete degradation of apoprotein [11,12]; (3) preferential uptake of apoprotein [34]; and (4) uptake of unesterified cholesterol [35-371. The purpose of the present study was to investigate the possible role of HDL specific binding sites in the first two of these pathways, by comparing the metabolism of appropriately labeled HDL, before and after DMS modification. DMS modification of HDL, has been shown to inhibit its specific binding to isolated membranes and to cells in culture [15]. TNM modification is another way to inhibit the binding of HDL to its specific binding sites [14,38]. However, this modification results in extensive chemical alteration of HDL molecule, including cross-linking of lipids to apoproteins, cross-linking of apoproteins to each other, nitration of tyrosine residues and change of surface charge [14,38]. Furthermore, TNM-modified HDL is rapidly cleared from rat blood and is taken up preferentially in liver, kidney and spleen [39]. Recently, it has been shown that rat liver endothelial cells contain scavenger receptors that recognize and catabolize TNM-modified HDL [40]. Contrary to our expectations, the modification of HDL with DMS increased, rather than reduced the rate of apoprotein clearance of ‘251-HDL, from rat blood. To understand the mechanism of this increased clearance, in vitro binding to purified liver plasma membranes and in vivo binding to major organs of the native and the DMS-modified HDL, were studied. The results of these studies suggest that although DMS-modified HDL, does not bind to the HDL specific binding sites, it does bind to a different class of binding sites, with apparently similar binding affinity and capacity. In addition, the binding capacity of 12’I-HDL, to liver and spleen but not to other organs was higher, suggesting the possible participation of reticuloendothelial cells in the catabolism of DMS-modified HDL,. Scavenger cell receptors for chemically modified HDL, with properties different from those for acetyl LDL [41] have recently been reported [31,40]. Thus, Murakami et al. [31] have identified receptors that recognize acetyl HDL in rat liver sinusoidal cells. Rat liver endothelial cells appear to have scavenger receptors that internalize and degrade
TNM-modified HDL, [40]. Whether these or similar receptors in reticuloendothelial cells are responsible for the increased clearance of DMS HDL, apoprotein from rat blood remains to be studied. Although the phenomenon of selective uptake of HDL cholesteryl ester has been well documented [11,12], the mechanism of the process has not been identified. The two extreme possibilities are: (1) that HDL binds to the cell surface and cholesteryl ester is transferred into the cell; or (2) that the entire HDL particle is internalized with subsequent release of the cholesteryl ester-depleted HDL from the cell by retroendocytosis [5]. Binding of HDL to the cell surface is necessary in both mechanisms. In our present study we find that DMS modification has no effect on the rate of clearance of 3H label from rat blood of ‘*?(apo)-[ 3H](CEt)-HDL3 (Fig. 5). In vivo binding experiments are also in agreement with this observation; similar amounts of control and DMS-modified [3H](CEt)-HDL, bind to liver and adrenal tissues, the major organs that selectively take up cholesteryl esters (Table IV). Furthermore, selective uptake of [3H]cholesteryl ether from [3H](CEt)-HDL3 by isolated rat liver plasma membranes was not affected by DMS modification of the lipoprotein (Table V). One explanation for these observations is that the control and the DMS-modified HDL, bind to different binding sites. In the case of control HDL,, binding results in the delivery of cholesteryl ester without internalization and degradation of HDL apoprotein. However, in the case of DMS HDL,, the lipoprotein is internalized as a whole particle and lysosomally degraded in a manner similar to acetyl HDL, [31] or TNM HDL, [40]. An alternate explanation is that both control and DMSHDL, are binding to the same binding sites. Subsequent to binding, part or all of the apoprotein component of control HDL, dissociates from the bound lipoprotein and thus escapes internalization. However, with DMS HDL,, since apoproteins are cross-linked to one another, this exchange is reduced, resulting in the uptake of both ‘lipid and apoprotein components. Clearly, further experiments are needed to decide among the different possibilities. Because of its ability to bind to cell surfaces, although not to HDL specific sites, and undergo internalization, DMS-modified HDL does not seem to be suitable for in vivo functional studies of HDL receptor. The present results do not give any indication that binding of HDL to its ‘specific’ binding sites is essential for the apoprotein catabolism or for the cholesteryl ester selective uptake. Acknowledgments We are grateful to Dominique Vacher for technical assistance, to Dr. William J. Johnson for reading the manuscript and for offering helpful suggestions, and to Gerri Lacrosse for typing. This research was supported
88 by NIH grant HL 37550 from National Heart, Lung and Blood Institute. F.H.M. was supported in part by the ComitC des Aides & la Recherche Fournier-Dijon, France. G.K.C. was a recipient of INSERM Senior Fellowship for visiting scientists. References 1 Chait, A. (1983) in Special Topics in Endocrinology and Metabolism (Cohen, M.P. and Foa. P.P., eds.). Vol. 5, pp. l-53, AR. Liss, New York. 2 Eisenberg, S. (1984) J. Lipid Res. 25, 1017-1058. 3 Graham, D.L. and Oram, J.F. (1987) J. Biol. Chem. 262. 74397442. 4 Oram, J.F., Brinton. E.A. and Bierman, E.L. (1983) J. Clin. Invest. 72, 1611-1621. 5 Schmitz, G., Robenek, H., Lohmann, V. and Assmann, G. (1985) EMBO J. 4, 613-622. 6 Barbaras, R., Grimaldi, P., Negrel, R. and Ailhaud, G. (1986) Biochim. Biophys. Acta 888, 143-156. 7 Nakai, T., Otto, P.S., Kennedy, D.L. and Whayne, T.F., Jr. (1976) J. Biol. Chem. 251, 4914-4921. 8 Bachorik, P.S., Franklin, F.A.. Virgil, D.G. and Kwiterovich. P.O., Jr. (1982) Biochem. 21, 5675-5684. Rev. 3. 9 Gwynne, J.T. and Strauss, J.F.. III (1982) Endocrinol. 299-329. 10 Nestler, J.E., Bamberger, M., Rothblat, G.H. and Strauss, J.F., III (1985) Endocrinology 117, 502-510. 11 Glass, C., Pittman, R.C., Civen, M. and Steinberg, D. (1985) J. Biol. Chem. 260, 744-750. 12 Leitersdorf, E., Stein, 0.. Eisenberg, S. and Stein, Y. (1984) Biochim. Biophys. Acta 796, 72-82. D., Cohen, D.C. and Massogha, S.L. (1983) J. 13 Gospodarowicz, Cell. Physiol. 117, 76-90. 14 Chacko, G.K. (1985) J. Lipid Res. 26, 745-754. 15 Chacko, G.K., Mahlberg, F.H. and Johnson. W.J. (1988) J. Lipid Res. 29, 319-324. 16 KarIin, J.B., Johnson, W.J., Benedict, C.R., Chacko, G.K., Phillips, M.C. and Rothblat, G.H. (1987) J. Biol. Chem. 262, 12557-12564. 17 Johnson, W.J., Mahlberg, F.H., Chacko, G.K., Phillips, M.C. and Rothblat, G.H. (1988) J. Biol. Chem. 263, 14000-14106. 18 Weber, K. and Osbom, M. (1969) J. Biol. Chem. 244, 440664412. 19 Bilheimer, D.W., Eisenberg, S. and Levy, R.I. (1972) B&him. Biophys. Acta 260, 212-221.
20 Basu, S.K., Goldstein, J.L., Anderson, R.G.W. and Brown, MS. (1976) Proc. Nat]. Acad. Sci. USA 73, 3178-3182. 21 Roberts, D.C.K., Miller, N.E., Price, S.G.L., Crook, D.. Cortese, C., La Ville, A., Masana, L. and Lewis, B. (1985) Biochem. J. 226. 319-322. 22 Dole, V.P. (1956) J. Clin. Invest. 35, 150-154. M., Remesar, R. and Bourdel, G. (1987) Am. J. 23 Pastor-Anglada, Physiol. 252, E408-E413. 24 Grunnet. L. and Bogesen, E. (1976) Biochim. Biophys. Acta 419. 365-378. 25 Chacko, G.K. (1982) Biochim. Biophys. Acta 712. 129-141. 26 Koelz, H.R., Sherrill. B.C.. Turley, S.D. and Dietschy, J.M. (1982) J. Biol. Chem. 257, 8061-8072. 27 Folch, J., Lees, M. and Sloane-Stanley, G.H. (1957) J. Biol. Chem. 2263497-509. 28 Peterson, G.L. (1977) Anal. Biochem. 83. 346-356. 29 Van Tol, A., Van Grent. T., Van’t Hooft, F.M. and Vlaspolder, F. (1978) Atherosclerosis 29, 439-448. D., Stein. 0. and Stein, Y. (1971) 30 Roheim, P.S., Rachmilewitz, B&him. Biophys. Acta 248, 315-329. 31 Murakami, M.. Horiuchi, S., Takata. K. and Morino, Y. (1987) J. Biochem. 101, 729-741. 32 Israeli, A., Leitersdorf, E., Stein, 0. and Stein Y. (1987) Biochim. Biophys. Acta 902, 128-132. P.T. and Goldstein, J.L. (1981) Science 33 Brown, M.S.. Kovanen, 212, 628-635. 34 Glass, C.K.. Pittman, R.C., Keller, G.A. and Steinberg, D. (1983) J. Biol. Chem. 258, 7161-7167. C.C., Vlahcevic, Z.R., Berman, M., Meadows, J.G.. 35 Schwartz. Nisman, R.M. and Swell. L. (1982) J. Clin. Invest. 70, 105-116. W.C. (1980) Trend. Biocbem. Sci. 5. 36 Jansen, H. and Hulsmann, 265-268. 37 Portman, O.W., Alexander, M. and O’MaIley. J.P. (1980) Biochim. Biophys. Acta 619, 545-558. 38 Brinton, E.A., Oram, J.F., Chen, C.-H., Albers, J.J. and Bierman, E.L. (1986) J. Biol. Chem. 261, 495-503. 39 Nestler, J.E., Chacko, G.K. and Strauss, J.F., III (1985) J. Biol. Chem. 260, 7316-7321. 40 Kleinherenbrink-Stins, M.F., Schouten, D., Van der Boom, J., Brouwer, A., Knook, D.L. and Van Berkel, T.J.C. (1989) J. Lipid Res. 30, 511-520. 41 Brown, M.S. and Goldstein, J.L. (1983) Annu. Rev. B&hem. 52. 223-261.
Eiochimica et Biophysics Acta, 1046 (1990) 81-88
Elsevier BBALIP 53465
Effect of apoprotein cross-linking on the metabolism of human HDL, in rat Claude Senault ‘, Florence H. Mahlberg 2, Guy Renaud I, Anik Girard-Globa and George K. Chacko 2
’
’ Unit&INSERM 286, Fact& de Mkdecine Xavier-&hat,
Paris (France) and 2 Department of Physiology and Biochemistry, Medical College of Pen~s~~lvania,P~ilade~hia, PA (U.S.A.)
(Received 19 March 1990)
Key words: HDL; HDL receptor; Apoprotein catabolism; Cholesteryl ester selective uptake
Apo E-free human high-densi~ Ii~protein (HDL,) was i&&d with ‘=I in apoprotein and with 3H in choIestery1 linoleyl ether (a non-hydrolyzable analogue of eholesteryl ester). The labeled HDL, was modified by cross-liuking of apoproteins with dimethylsuberimidate (DMS) to inhibit binding to HDL specific receptors. The control and the DMS HDL, were characterized with respect to their rate of clearance from rat blood, in vivo binding to major rat organs and in vitro binding to purified rat liver lasma membranes. Both 12’1 and 3H labels from control HDL, were cleared from Q rat blood mon~x~nenti~iy, but H at a faster rate than “‘1 (3H fl,2 = 3.0-4.1 h; ‘“I t,,:, = 7.0-7.7 h). This difference is consistent with reports of the nonendocytotic selective uptake of HDL-associated cholesteryl ester. DMS modification did not affect the rate of 3H clearance whereas it increased the rate of “‘1 clearance (HDL, z,,, = 7.7 h, DMS HDL, t,,, = 4.1 h). Both in vivo binding to rat organs and in vitro binding to rat liver membranes confirmed that DMS modification inhibited the specific binding of HDL, but also suggested that the modification produced saturable binding of HDL to a separate class of sites. Thus, the present data do not rule out the involvement of direct HDGcell intention in the selective uptake of HDL eholesteryl ester. However, results suggest that the binding of HDL to its specific cell surface sites is not necessary for this uptake.
Introduction
Apo E-free ~~-density lipoprotein (HDL) binds to a variety of cell surfaces and isolated membranes, with properties typical for a l&and-receptor type interaction [1,2]. A membrane-associated HDL binding protein has been identified by ligand blotting [3]. However, very little is known about the role of this putative HDL receptor. Some of the postulated functions of HDL binding sites include: (1) efflux of unesterified cholesterol from peripheral cells [4-61; (2) intemalization and catabolism of HDL in the liver [7,8]; and (3) preferential delivery of free and esterified cholesterol to liver and steroidogenic tissues without apoprotein de-
Abbreviations: HDL, high-density lipoprotein; LDL, low-density lipoprotein; DMS, dimethylsuberimidate; CEt, cholesteryl ether; TNM, te~~trome~ane; SDS-PAGE, sodium dodecyl sulfate-polyacryiamide gel electrophoresis. Correspondence: G.K. Chacko, Department of Physiology and Biochemistry, Medical College of Pennsylvania, Philadelphia, PA 19129, U.S.A. US-27~/~/$03.50
gradation [9-121. It has also been suggested that the mitogenic effect of HDL on certain cells is mediated through the NDL binding sites [331. To investigate the possible role of HDL binding sites in various processes, we sought to inhibit HDL binding by chemical modification. Thus, we have previously shown that treatment with tetranitromethane (TNM) or cross-linking of apoproteins by ~methylsube~~date (DMS) inhibited the binding of HDL to its specific binding sites on isolated rat liver membranes and on cells in culture [14,15]. Using these chemically modified HDL preparations we have shown recently that, at least in human skin fibroblasts and in rat Fu5AH hepatoma cells, HDL binding is not a prerequisite for either the efflux or influx of unesterified cholesterol between cells and lipoproteins [16,17]. In the present report, we have used DMS modification of HDL to probe the functional role of the HDL binding sites in the in viva catabolism of HDL in rats. The apoprotein in human HDL, was labeled with i2? and the cholesteryl ester moiety was traced with [3Hfcholesteryl linoleyl ether (a non-hydrolyzable analogue of cholesteryl ester). DMS modification of human HDL, did not decrease but
@ 1990 Eisevier Science Publishers B.V. (Biomedical Division)
82 rather increased the rate of apoprotein clearance from the blood; clearance of cholesteryl linoleyl ether was not affected. Material and Methods Rat serum albumin, bovine serum albumin (BSA) fatty acid free and dimethylsuberimidate, (DMS) were from Sigma (St. Louis, MO). [1,2-3H]Cholesteryl linoleyl ether (spec. act. 31 Ci/mmol) was purchased from Amersham, France. All other chemicals were of reagent grade. Lipoprotein preparations Human HDL, (1.125 1.215) that had been prelabeled with [3H]cholesteryl linoleyl ether, followed by reisolation of the labeled lipoprotein by ultracentrifugation. The labeled lipoprotein was analyzed in triplicate by agarose gel electrophoresis (barbital buffer, pH 8.0). One gel was stained with Oil-Red-O and scanned by densitometry. The others were cut into 5 mm sections and analyzed for radioactivity: directly for 125I label and after extraction by the procedure of Dole [22] for 3H label. The initial preparation showed, in addition to HDL-associated radioactivity, another radioactive component that barely moved from the origin of the agarose gels. This contaminant was removed by gel filtration on a column of 0.5 M Bio-gel Agarose (Bio-Rad), eluting with 0.15 M NaCl containing 0.02% EDTA (pH 7.4). The purified sample exhibited a single radioactive band which co-migrated with control HDL,. Its specific activity was 16 dpm per ng cholesteryl ester. Similarly, doubl? labeled HDL,, with “‘1 in the apoprotein and H in the cholesteryl ether (‘251(apo)[ 3H](CEt)-HDL,) was prepared from 12?-HDL3 and [3H]cholesteryl linoleyl ether using the procedure of Roberts et al. [21], followed by gel filtration chromatog-
raphy on a 0.5 M Bio-gel Agarose column. DMS modified HDL, was prepared as previously described [15]. Briefly, HDL, (2 mg protein/ml) was reacted with DMS (2 mg/ml) in 0.09 M triethanolamine buffer (pH 9.6) for 2 h at room temperature. After incubation, the reaction mixture was applied to a column of Sephadex G-50 and the lipoprotein was eluted with 0.15 M NaCl. containing 0.01% EDTA (pH 7.4). Fractions containing protein were combined to obtain DMS HDL,. SDSPAGE analysis followed by Coomassie blue staining revealed extensive cross-linking of the apoproteins. The cross-linked apoproteins, seen as a high molecular weight band with a mean molecular weight of 95000 on a 3-27% gradient slab gel represented more than 80% of the Coomassie blue staining components of DMS HDL,
WI. Membrane preparations Rat liver plasma membranes were isolated using the Percoll gradient centrifugation procedure of PastorAnglada et al. [23]. The 5’-nucleotidase activity of the purified membranes, assayed according to Grunnet and Bogesen [24] was more than lo-times that of the liver homogenate.
Binding of labeled liEproteins to isolated membranes The binding of I-HDL, and iz51-DMS HDL, to isolated rat liver plasma membranes was determined according to the procedure of Chacko [25]. Briefly, aliquots of membranes (200 pg protein) were incubated with varying concentrations of labeled lipoprotein at room temperature for 1 h in a total volume of 0.2 ml, containing 0.150 M NaCl, 0.5 mM CaCl,, 10 mM Tris-HCl (pH 7.4) and 1% bovine serum albumin. After incubation, 0.175-ml aliquots of incubation mixture were centrifuged in a Beckman 42.2 Ti rotor at 30000 rpm for 15 min to recover the membranes. The membrane pellets were washed once with 0.175 ml of incubation medium and the tubes containing the membranes were assayed for radioactivity in a gamma scintillation spectrophotometer. Non-specific binding of labeled lipoprotein was determined in samples run in parallel that also contained a lOO-fold excess of the corresponding unlabeled lipoprotein. The difference between the ‘251lipoprotein bound to the membrane in the absence and presence of excess unlabeled lipoprotein was taken as the amount of specific binding. In order to determine the selective uptake of cholesteryl ester, 200~pg protein aliquots of liver plasma membranes were incubated with identical concentration of ‘251-HDL3 and [ 3H](CEt)HDL, (10 pg protein/ml) as above and the amount of i2’I and 3H radioactivity associated with the washed membranes were determined, “‘1 directly by gamma counting and 3H, after extraction [22], by liquid scintillation counting.
83 In vivo metabolic studies
Male Sprague-Dawley rats (IFFA CREDO France) weighing 225-250 g were used. They were fed standard rat chow and were fasted overnight prior to metabolic studies. Radiolabeled HDL, was injected into the femoral vein under light ether anesthesia (1251, 4. lo6 cpm/animal; 3H, 0.3. lo6 cpm/animal). 3 min after injection, a small sample of blood was obtained from the orbital plexus for determination of the initial plasma radioactivity. Blood samples of about 150 ~1 were then collected at the specified times from the tail vein, allowed to clot on ice for 15 min and centrifuged to obtain the serum. Serum was analyzed for 1251radioactivity by counting aliquots in a gamma counter. TCAsoluble counts were subtracted from the total 1251counts. For 3H radioactivity determination, lipids were extracted according to Dole [22] and the extract analyzed by liquid scintillation counting. Serum decay data were fitted to a monoexponential function for calculation of half time (t,,*) using regression line analysis. The tissue-space distribution of labeled lipoproteins and rat serum albumin in rat was evaluated according to the procedure of Koelz et al. [26]. The labeled proteins were injected into the femoral vein under light ether anesthesia. After 10 min, the animal was anesthetized and blood was collected from the abdominal aorta. Tissues were removed and weighed. They were cut into small pieces, carefully rinsed in ice-cold 0.15 M NaCl and analyzed either directly for “‘1 or after Folch extraction [27] for ‘H radioactivity. Tissue-space, which represents the volume (~1) of plasma cleared per gram of tissue, was calculated using the equation: tissue space = tissue radioactivity (cpm/g)/plasma radioactivity (cpm/pl) (261. Protein was determined by the method of Peterson [28] using bovine serum albumin as the standard. Total lipids from lipoprotein preparations were extracted according to the procedure of Folch et al. [27]. Lipids were assayed enzymatically using Wako kits for phospholipids (Biolyon, France) and Boehringer kits for total and unesterified cholesterol (Boehringer Mannheim, France). Statistical analysis The values presented
are means f S.E. The significance of differences was tested by Student’s t-test. Results fi$2ct of DMS modification on serum decay of human I-HDL,
When trace amounts of native ‘*?-HDL,, were injected intravenously into rats and the 1251counts in the serum were measured with time, the TCA-precipitable counts disappeared from the serum monoexponentially with a half-time (t1,2) of 7.7 f 0.5 h (Fig. 1). This value
100
1
2
3
4
5
6 Time
7
a
hours)
1. Effect of DMS modification on the rate of disappearance of label from rat blood following injection of radiolabeled human HDL, (4.106 cpm/animal): n -w, ‘251-HDL,; A-A, ‘*‘I-DMS 12?-acetyl HDL,. At the indicated times after HDL,; o-o, injection, blood was collected and serum was separated and analyzed for TCA-precipitable counts. Values are expressed as ‘k; of the initial radioactivity (to) obtained from the blood collected at 3 min after injection from the orbital plexus. Values are means from 20 animals preeach for ‘*‘I-HDL, and ‘*‘I-DMS HDL, and five lipoprotein parations. The data for ‘*‘I-acetyl HDL, were mean values from three animals with one lipoprotein preparation. t,,* values for 1251-HDLs, 4.1 kO.6 and ‘*‘I-DMS HDL, and 1251-acetyl HDL, were 7.7+0.5, 0.1 h, respectively.
is slightly higher than the value reported by Van To1 et al. ~291 (t,,, = 6.2 h) but lower than that reported by Roheim et al. [30] (tl,* = 10.5 h), both studies using rat HDL in rat. DMS modified ‘*?-HDL,, was also cleared monoexponentially from the blood (Fig. l), but at a significantly faster rate than control HDL, (t,,, = 4.14 f 0.6 h). This increased turnover of ‘*‘I-DMS HDL, was found in five different lipoprotein preparations and in a total of 20 rats. In order to investigate whether this increased rate of clearance might be related to the recently reported uptake mechanism for acetylated HDL, [31], a sample of ‘251-HDL, was acetylated and its clearance from the blood was studied. As shown in Fig. 1, acetylated ‘*‘I-HDLs was removed from the rat blood with a t,,, of about 0.1 h, much faster than that of I*?-DMS HDL,. In vivo binding of ‘2sI-HDL,
and 12’I-DMS HDL,
In order to identify the organs responsible for the increased uptake, and thus the faster disappearance from the blood of 125I-DMS HDL,, the in vivo binding of native and DMS modified HDL, to various organs of rat was determined 10 min after injection, using the tissue space technique [26] (Table I). The specificity of binding was assessed by using “‘1 rat serum albumin as a non-specific marker of tissue space .[26]. All tissues examined were found to bind ‘251-HDL3; however when these binding values were corrected for non-specific binding by subtracting the 1251rat serum albumin space, only liver and adrenal tissues showed a specific binding. The tissue space for native was highest in the adrenal. For “‘1-DMS HDL, the
tissue spaces were similar in all organs except the liver and, to a smaller extent, the spleen. The liver tissue space was significantly larger for ‘*‘I-DMS HDL, than for native HDL, (P < O.OS), while adrenal space was unchanged. A small specific compartment also appeared in the spleen. Binding of ‘751-HDL, and “‘I-DMS HDL, to rat liver plasma membranes The binding of native and DMS-modified HDL, to purified rat liver plasma membranes was studied. In Fig. 2 is shown the binding of ‘251-HDL, (A) and of 12’I-DMS HDL, (B) to membranes in the absence and the presence of an excess of the respective unlabeled HDL,. The specific binding was saturable for both the native and DMS-modified HDL,. Saturation of binding sites occurred at an HDL concentration less than 100 pg HDL protein/ml. Scatchard analysis of the binding data gave similar Kd (26 pg/ml for native HDL, and 24 pg/ml for DMS HDL,); there was also no significant difference in the maximum binding capacity, B,,, (Table II). In order to determine whether the native and DMS-modified HDL, are binding to the same or different binding sites, competition experiments were performed. In Fig. 3 (A) is compared the ability of unlabeled native and DMS HDL3 to compete with 12’1HDL, for its binding sites. Whereas the native HDL, was very effective in competing with ‘251-HDL , DMSmodified HDL, was not. Similarly, when %DMS HDL, was used as the ligand (Fig. 3 (B)) native HDL, was less effective than the DMS-modified HDL,, especially at low concentration, to inhibit the binding of the labeled lipoprotein; thus, at 10 pg protein, the inhibitions by native and DMS-modified HDL, were 15 and 55%, respectively. These results would suggest that, although both native and DMS-modified HDL, bind to isolated rat liver membranes with apparent similar binding properties, they are binding to different sites.
1251-mX3, pgglml
1251-DMS HDL-~~,pglrnl Fig. 2. Binding of 1251-HDL, (A) and tz51-DMS HDL, (B) to rat liver plasma membranes as a function of lipoprotein concentration. Aliquots of membranes (200 ng protein) were incubated with the indicated concentration of labeled lipoprotein in 0.2 ml of incubation medium containing 0.15 M NaCl, 0.5 mM CaCl,, 10 mM Tris-HCl (pH 7.4) and 1% bovine serum albumin for 1 h at room temperature in the absence (+) and presence (0) of lOO-times excess of the corresponding unlabeled lipoprotein. After incubation, the amount of labeled lipoprotein bound to the membranes was determined as described in Materials and Methods. The amount of labeled lipoprotein bound specifically to the membranes (A) was obtained by subtracting the labeled lipoprotein bound to the membrane in the presence of excess unlabeled lipoprotein from that bound in its absence. The data in the figure are from one representative of four separate binding experiments.
TABLE Effect
Effect of DAYS modification on the preferential clearance of HDL [-‘H]cholesteryl linoleyl ether Using doubly labeled lipoprotein preparations it has been shown in rats [11,12] that the core cholesteryl ester
I
of DMS modification on the in vivo binding of “‘I HDL, to major organs of ran
In vivo binding, expressed as tissue space (~1 plasma/g of rats in each group is shown in parentheses.
tissue) was calculated
as described
in Materials
and Methods.
Values are means f S.E.; NO.
Tissue space (PI/g) liver
kidney
spleen
adrenal 84k
testes
heart
6
21*6
56kll
iz51-rat serum albumin
(3)
70f
2
91*15
7X*9
12’1 HDL,
(7)
74k
8
64klO
61+6
238 + 38
9*4
4Oi
4
‘251-DMS HDL,
(6)
105*13
68klO
84+8
275+75
9*3
41k
6
85 TABLE
III
Comparison of the in vivo binding of “?-apoprotein
and [‘H]CEi
In viva binding expressed as tissue space (~1 plasma/g of rats in each group is shown in parentheses.
from “.‘I-HDL,
tissue) was calculated
and [‘H](CEt)-HDL,
as described
in Materials
respectively, to major organs of rat and Methods.
Values are means f SE.; No.
Tissue space (pi/g) liver “‘I-HDL3
(7)
74*
[ 3H](CEt)-HDL,
(6)
350*30
TABLE
8
kidney
spleen
64+10
611
64k
8
adrenal
testes
heart
38
9k4
40*
1273 f 123
15*4
6
122 + 49
238+
4
51*19
II
Comparison of the binding of “‘I-HDL, liver plasma membranes
and ‘251-DMS HDL,
to rat
Values are mean f S.E., from four separate binding experiments performed as described in the legend to Fig. 2. K, and B,,, values are derived from Scatchard analysis of the binding data.
K, (pg protein/ml)
%HDL,
‘251-DMS HDL,
26
24
+5
ol
kl
1
%,, (Bg protein/mg membrane protein)
2
3
4
5
6
Time (hours ) 1.7fO.3
2.4kO.7
of HDL is cleared faster than the surface apoprotein component. In order to assess the possible role of HDL binding in this preferential transfer process, we studied the effect of DMS modification on the clearance from rat plasma of doubly labeled HDL, with 1251 in apoprotein and 3H in cholesteryl linoleyl ether, a nonhydrolyzable tracer of cholesteryl ester. As is shown in Fig. 4, when the unmodified lipoprotein was used, [ 3H]cholesteryl ether was removed much faster than ‘*‘I apoprotein. The clearance of both components was
Fig. 4. Differential rate of disappearance of ‘25I and 3H labeled from (4.106 rat blood following injection of ‘25I (apo)-[3H](CEt)-HDLa At the indicated times after cpm ‘*‘I and 0.3.106 cpm ‘H/animal). injection, blood was collected and serum was separated and analyzed and ‘H (0 -0) counts as described in for i25I (M-w) legend to Fig. 1 and in Materials and Methods. Average of three experiments, each performed in three rats.
monoexponential and the 7.0 + 1 h, respectively. The major organs that ential uptake and thus the of [3H]cholesteryl ether,
half-times
were 3.0 f 0.8 and
were involved in the preferfaster clearance from plasma were determined by in vivo
DMS-HDL3
o\”
ol____ 510
25
OJ 50
75 w
UKl “Df-3
5K)25
I
1
1
,
50
75
loo
PQ ‘4”L3
Fig. 3. Effect of HDL, and DMS HDL, on the binding of ‘*‘I-HDL, (A) and lz5 I-DMS HDL, (B) to rat liver plasma membranes. Each tube contained 200 gg membrane protein, 10 pg protein/ml of ‘*‘I-HDL, (A) or 125I-DMS HDL, (B) and indicated concentrations of either HDL, 0) or DMS HDL, (O0). After incubation at room temperature for 1 h, the amount of ‘*‘I-HDL, (A) and ‘*‘I-DMS HDL, (B) (Obound to the membranes was determined as described in Materials and Methods. The data are average from a single experiment done in triplicate and are typical of results in three experiments.
86 TABLE
V
Effect of DMS [3H](CEi)-HDL,
f
‘“M
modification on the binding to rat liver plasma membranes
6
I of HDL, bound (‘25I or 3H) i2?-HDL,
“‘1-DMS
, 1
,
,
,
,
(
2
3
4
5
6
and
3H/‘251
2.15 +0.17
[3H](CEt)-HDL,
10-l
“51-HDL_q
Aliquots of membranes (200 pg protein) were incubated with 10 pg protein/ml each of labeled lipoprotein in 0.2 ml of incubation medium containing 0.15 M NaCl. 0.5 mM CaCI,, 10 mM Tris-HCI (pH 7.4) and 1% bovine serum albumin for 1 h at room temperature. After incubation, membranes were washed, recovered and analyzed for “‘1 and 3H counts (as described in Materials and Methods) to determine the amount of lipoprotein bound. Values represent percentage of total lipoprotein in the incubation medium bound to the membranes and are mean f SE. of three experiments.
Time (hours)
i
of
8.42 f 3.29
HDL,
[ 3H](CEt)-DMS
3.87+ 1.30
1.96f0.19 HDL,
8.53 f 2.51
4.45 f 0.70
Time (hours)
Fig. 5. Effect of DMS modification on the rate of disappearance of 12’1 (A) and ‘H (B) labels from rat blood following injection of ‘25I (apo)-[3H](CEt)-HDL3 (4. lo6 cpm, “‘1 and 0.3. lo6 cpm, 3H/animal). At the indicated times after injection, blood was collected and serum was separated and analyzed for i*‘I and 3H counts as described in legend to Fig. 1 and in Materials and Methods. (A) D-m, 1251, control lipoprotein; A-A, 1251, DMS modified 0, ‘H, control lipoprotein; l ----lipoprotein; (B) o0, 3H, DMS modified
lipoprotein. Average of three experiments, formed in three rats.
clearance, half-times were 6.8 + 1.1 and 3.5 k 0.9 h, respectively. It also appeared that in the case of DMSmodified HDL,, both 1251 and 3H labels were cleared at similar rates. This lack of involvement of specific HDL binding in the selective uptake of cholesteryl ether was further studied by examining the effect of DMS modification of labeled lipoproteins in the in vivo binding to major organs of rat and in vitro binding to isolated rat liver plasma membranes. As is shown in Table IV, DMS modification has no significant effect on the in vivo binding of [3H]CEt)-HDL3 to major organs of rat; the high tissue space of [3H]cholesteryl linoleyl ether to liver and adrenals remained unaffected after DMS modification of the lipoprotein. We also examined the preferential binding/uptake of 3H-labeled cholesteryl linoleyl ether by isolated rat liver plasma membranes. When isolated rat liver plasma membranes were incubated with equal concentration of [3H](CEt)-HDL3 or ‘251HDL, (10 pg protein/ml), relatively higher percentage of 3H label was taken up by the membranes than of the lz51 label (Table V); the 3H: 12’1 ratio for the unmodified lipoproteins was 3.9. This preferential binding/ uptake by the membranes was not affected by DMS modification of the lipoproteins. Selective uptake of HDL cholesteryl ester has been reported previously by
each per-
binding experiments. In Table III is shown a comparison of the in vivo binding (uptake) of 3H and iz51 labels from the labeled HDL, to major organs of the rat 10 min after intravenous injection of the lipoprotein. Liver and adrenal, and to a small extent spleen, showed 1 linoleyl ether over referential uptake of [ 3H]choleste P 35, I labels, expressed 251 apoprotein. The ratio of 3H to as space volume, were 4.6 for liver and 5.3 for adrenal. In Fig. 5 is compared the clearance of 125I (A) and 3H (B) labels from the doubly labeled HDL, before and after DMS modification of the lipoprotein. DMS modification of the lipoprotein had no significant effect on the rate of clearance of [3H]cholesteryl linoleyl ether, half-times were 3.7 f 0.6 and 3.0 h 0.5 h, respectively. However, in agreement with the data presented in Fig. 1. DMS modification increased the rate of apoprotein TABLE
IV
Effect of DMS modification
on the in viuo binding of [_‘HI(CEt)-HDL,
In vivo binding, expressed as Tissue space (cl plasma/g of rats in each group is shown in parentheses.
to major organs of rat
tissue) was calculated
as described
in Materials
and Methods.
Values are means f S.E.; No.
Tissue spaces (PI/g)
[ 3H](CEt)-HDL, [ ‘H](CEt)-DMS
HDL,
liver
kidney
(6)
350 f 30
64k
(6)
354*45
74k22
8
spleen
adrenal
testes
heart
122 f 49
1273k123
15*4
51*19
21Ok93
1017 f 217
15*1
49*11
87 Israeli et al. [32] in a bovine tion.
adrenal
membrane
prepara-
Discussion Unlike low-density lipoprotein (LDL), which is taken up by the apo B/E receptor pathway as a whole particle and degraded to its components in the lysosomes [33], the catabolism of HDL appears to involve more than one route [2]. At least four distinct pathways have been described for the catabolism of HDL: (1) internalization of the whole HDL particle followed by lysosomal degradation [7,8]; (2) preferential uptake of core cholesteryl ester with incomplete degradation of apoprotein [11,12]; (3) preferential uptake of apoprotein [34]; and (4) uptake of unesterified cholesterol [35-371. The purpose of the present study was to investigate the possible role of HDL specific binding sites in the first two of these pathways, by comparing the metabolism of appropriately labeled HDL, before and after DMS modification. DMS modification of HDL, has been shown to inhibit its specific binding to isolated membranes and to cells in culture [15]. TNM modification is another way to inhibit the binding of HDL to its specific binding sites [14,38]. However, this modification results in extensive chemical alteration of HDL molecule, including cross-linking of lipids to apoproteins, cross-linking of apoproteins to each other, nitration of tyrosine residues and change of surface charge [14,38]. Furthermore, TNM-modified HDL is rapidly cleared from rat blood and is taken up preferentially in liver, kidney and spleen [39]. Recently, it has been shown that rat liver endothelial cells contain scavenger receptors that recognize and catabolize TNM-modified HDL [40]. Contrary to our expectations, the modification of HDL with DMS increased, rather than reduced the rate of apoprotein clearance of ‘251-HDL, from rat blood. To understand the mechanism of this increased clearance, in vitro binding to purified liver plasma membranes and in vivo binding to major organs of the native and the DMS-modified HDL, were studied. The results of these studies suggest that although DMS-modified HDL, does not bind to the HDL specific binding sites, it does bind to a different class of binding sites, with apparently similar binding affinity and capacity. In addition, the binding capacity of 12’I-HDL, to liver and spleen but not to other organs was higher, suggesting the possible participation of reticuloendothelial cells in the catabolism of DMS-modified HDL,. Scavenger cell receptors for chemically modified HDL, with properties different from those for acetyl LDL [41] have recently been reported [31,40]. Thus, Murakami et al. [31] have identified receptors that recognize acetyl HDL in rat liver sinusoidal cells. Rat liver endothelial cells appear to have scavenger receptors that internalize and degrade
TNM-modified HDL, [40]. Whether these or similar receptors in reticuloendothelial cells are responsible for the increased clearance of DMS HDL, apoprotein from rat blood remains to be studied. Although the phenomenon of selective uptake of HDL cholesteryl ester has been well documented [11,12], the mechanism of the process has not been identified. The two extreme possibilities are: (1) that HDL binds to the cell surface and cholesteryl ester is transferred into the cell; or (2) that the entire HDL particle is internalized with subsequent release of the cholesteryl ester-depleted HDL from the cell by retroendocytosis [5]. Binding of HDL to the cell surface is necessary in both mechanisms. In our present study we find that DMS modification has no effect on the rate of clearance of 3H label from rat blood of ‘*?(apo)-[ 3H](CEt)-HDL3 (Fig. 5). In vivo binding experiments are also in agreement with this observation; similar amounts of control and DMS-modified [3H](CEt)-HDL, bind to liver and adrenal tissues, the major organs that selectively take up cholesteryl esters (Table IV). Furthermore, selective uptake of [3H]cholesteryl ether from [3H](CEt)-HDL3 by isolated rat liver plasma membranes was not affected by DMS modification of the lipoprotein (Table V). One explanation for these observations is that the control and the DMS-modified HDL, bind to different binding sites. In the case of control HDL,, binding results in the delivery of cholesteryl ester without internalization and degradation of HDL apoprotein. However, in the case of DMS HDL,, the lipoprotein is internalized as a whole particle and lysosomally degraded in a manner similar to acetyl HDL, [31] or TNM HDL, [40]. An alternate explanation is that both control and DMSHDL, are binding to the same binding sites. Subsequent to binding, part or all of the apoprotein component of control HDL, dissociates from the bound lipoprotein and thus escapes internalization. However, with DMS HDL,, since apoproteins are cross-linked to one another, this exchange is reduced, resulting in the uptake of both ‘lipid and apoprotein components. Clearly, further experiments are needed to decide among the different possibilities. Because of its ability to bind to cell surfaces, although not to HDL specific sites, and undergo internalization, DMS-modified HDL does not seem to be suitable for in vivo functional studies of HDL receptor. The present results do not give any indication that binding of HDL to its ‘specific’ binding sites is essential for the apoprotein catabolism or for the cholesteryl ester selective uptake. Acknowledgments We are grateful to Dominique Vacher for technical assistance, to Dr. William J. Johnson for reading the manuscript and for offering helpful suggestions, and to Gerri Lacrosse for typing. This research was supported
88 by NIH grant HL 37550 from National Heart, Lung and Blood Institute. F.H.M. was supported in part by the ComitC des Aides & la Recherche Fournier-Dijon, France. G.K.C. was a recipient of INSERM Senior Fellowship for visiting scientists. References 1 Chait, A. (1983) in Special Topics in Endocrinology and Metabolism (Cohen, M.P. and Foa. P.P., eds.). Vol. 5, pp. l-53, AR. Liss, New York. 2 Eisenberg, S. (1984) J. Lipid Res. 25, 1017-1058. 3 Graham, D.L. and Oram, J.F. (1987) J. Biol. Chem. 262. 74397442. 4 Oram, J.F., Brinton. E.A. and Bierman, E.L. (1983) J. Clin. Invest. 72, 1611-1621. 5 Schmitz, G., Robenek, H., Lohmann, V. and Assmann, G. (1985) EMBO J. 4, 613-622. 6 Barbaras, R., Grimaldi, P., Negrel, R. and Ailhaud, G. (1986) Biochim. Biophys. Acta 888, 143-156. 7 Nakai, T., Otto, P.S., Kennedy, D.L. and Whayne, T.F., Jr. (1976) J. Biol. Chem. 251, 4914-4921. 8 Bachorik, P.S., Franklin, F.A.. Virgil, D.G. and Kwiterovich. P.O., Jr. (1982) Biochem. 21, 5675-5684. Rev. 3. 9 Gwynne, J.T. and Strauss, J.F.. III (1982) Endocrinol. 299-329. 10 Nestler, J.E., Bamberger, M., Rothblat, G.H. and Strauss, J.F., III (1985) Endocrinology 117, 502-510. 11 Glass, C., Pittman, R.C., Civen, M. and Steinberg, D. (1985) J. Biol. Chem. 260, 744-750. 12 Leitersdorf, E., Stein, 0.. Eisenberg, S. and Stein, Y. (1984) Biochim. Biophys. Acta 796, 72-82. D., Cohen, D.C. and Massogha, S.L. (1983) J. 13 Gospodarowicz, Cell. Physiol. 117, 76-90. 14 Chacko, G.K. (1985) J. Lipid Res. 26, 745-754. 15 Chacko, G.K., Mahlberg, F.H. and Johnson. W.J. (1988) J. Lipid Res. 29, 319-324. 16 KarIin, J.B., Johnson, W.J., Benedict, C.R., Chacko, G.K., Phillips, M.C. and Rothblat, G.H. (1987) J. Biol. Chem. 262, 12557-12564. 17 Johnson, W.J., Mahlberg, F.H., Chacko, G.K., Phillips, M.C. and Rothblat, G.H. (1988) J. Biol. Chem. 263, 14000-14106. 18 Weber, K. and Osbom, M. (1969) J. Biol. Chem. 244, 440664412. 19 Bilheimer, D.W., Eisenberg, S. and Levy, R.I. (1972) B&him. Biophys. Acta 260, 212-221.
20 Basu, S.K., Goldstein, J.L., Anderson, R.G.W. and Brown, MS. (1976) Proc. Nat]. Acad. Sci. USA 73, 3178-3182. 21 Roberts, D.C.K., Miller, N.E., Price, S.G.L., Crook, D.. Cortese, C., La Ville, A., Masana, L. and Lewis, B. (1985) Biochem. J. 226. 319-322. 22 Dole, V.P. (1956) J. Clin. Invest. 35, 150-154. M., Remesar, R. and Bourdel, G. (1987) Am. J. 23 Pastor-Anglada, Physiol. 252, E408-E413. 24 Grunnet. L. and Bogesen, E. (1976) Biochim. Biophys. Acta 419. 365-378. 25 Chacko, G.K. (1982) Biochim. Biophys. Acta 712. 129-141. 26 Koelz, H.R., Sherrill. B.C.. Turley, S.D. and Dietschy, J.M. (1982) J. Biol. Chem. 257, 8061-8072. 27 Folch, J., Lees, M. and Sloane-Stanley, G.H. (1957) J. Biol. Chem. 2263497-509. 28 Peterson, G.L. (1977) Anal. Biochem. 83. 346-356. 29 Van Tol, A., Van Grent. T., Van’t Hooft, F.M. and Vlaspolder, F. (1978) Atherosclerosis 29, 439-448. D., Stein. 0. and Stein, Y. (1971) 30 Roheim, P.S., Rachmilewitz, B&him. Biophys. Acta 248, 315-329. 31 Murakami, M.. Horiuchi, S., Takata. K. and Morino, Y. (1987) J. Biochem. 101, 729-741. 32 Israeli, A., Leitersdorf, E., Stein, 0. and Stein Y. (1987) Biochim. Biophys. Acta 902, 128-132. P.T. and Goldstein, J.L. (1981) Science 33 Brown, M.S.. Kovanen, 212, 628-635. 34 Glass, C.K.. Pittman, R.C., Keller, G.A. and Steinberg, D. (1983) J. Biol. Chem. 258, 7161-7167. C.C., Vlahcevic, Z.R., Berman, M., Meadows, J.G.. 35 Schwartz. Nisman, R.M. and Swell. L. (1982) J. Clin. Invest. 70, 105-116. W.C. (1980) Trend. Biocbem. Sci. 5. 36 Jansen, H. and Hulsmann, 265-268. 37 Portman, O.W., Alexander, M. and O’MaIley. J.P. (1980) Biochim. Biophys. Acta 619, 545-558. 38 Brinton, E.A., Oram, J.F., Chen, C.-H., Albers, J.J. and Bierman, E.L. (1986) J. Biol. Chem. 261, 495-503. 39 Nestler, J.E., Chacko, G.K. and Strauss, J.F., III (1985) J. Biol. Chem. 260, 7316-7321. 40 Kleinherenbrink-Stins, M.F., Schouten, D., Van der Boom, J., Brouwer, A., Knook, D.L. and Van Berkel, T.J.C. (1989) J. Lipid Res. 30, 511-520. 41 Brown, M.S. and Goldstein, J.L. (1983) Annu. Rev. B&hem. 52. 223-261.
