Glycation of apolipoprotein E impairs its binding to heparin: identification of the major glycation site

Biochimica et Biophysica Acta 1454 (1999) 296^308 www.elsevier.com/locate/bba

Glycation of apolipoprotein E impairs its binding to heparin: identi¢cation of the major glycation site Vladimir V. Shuvaev a;b , Junichi Fujii a , Yoshimi Kawasaki a , Hidehiko Itoh c , Rieko Hamaoka a , Anne Barbier b , Olivier Ziegler d , Ge¨rard Siest b , Naoyuki Taniguchi a; * a

Department of Biochemistry, Osaka University Medical School, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan b Centre du Me¨dicament, Universite¨ de Nancy I, 30 rue Lionnois, 54000 Nancy, France c Department of Internal Medicine, Ikeda Municipal Hospital, 3-5-1, Jounan, Ikeda, Osaka 563-0025, Japan d Service de Me¨decine G du CHU de Nancy, Universite¨ de Nancy I, Hoªpital Jeanne d'Arc, 54201 Toul, France Received 14 December 1998; received in revised form 3 June 1999; accepted 8 June 1999

Abstract The increased glycation of plasma apolipoproteins represents a possible major factor for lipid disturbances and accelerated atherogenesis in diabetic patients. The glycation of apolipoprotein E (apoE), a key lipid-transport protein in plasma, was studied both in vivo and in vitro. ApoE was shown to be glycated in plasma very low density lipoproteins of both normal subjects and hyperglycemic, diabetic patients. However, diabetic patients with hyperglycemia showed a 2^3-fold increased level of apoE glycation. ApoE from diabetic plasma showed decreased binding to heparin compared to normal plasma apoE. The rate of Amadori product formation in apoE in vitro was similar to that for albumin and apolipoproteins A-I and A-II. The glycation of apoE in vitro significantly decreased its ability to bind to heparin, a critical process in the sequestration and uptake of apoE-containing lipoproteins by cells. Diethylenetriaminepentaacetic acid, a transition metal chelator, had no effect on the loss of apoE heparin-binding activity, suggesting that glycation rather than glycoxidation is responsible for this effect. In contrast, glycation had no effect on the interaction of apoE with amyloid L-peptide. ApoE glycation was demonstrated to be isoform-specific. ApoE2 showed a higher glycation rate and the following order was observed : apoE2 s apoE4 s apoE3 . The major glycated site of apoE was found to be Lys-75. These findings suggest that apoE is glycated in an isoform-specific manner and that the glycation, in turn, significantly decreases apoE heparin-binding activity. We propose that apoE glycation impairs lipoprotein-cell interactions, which are mediated via heparan sulfate proteoglycans and may result in the enhancement of lipid abnormalities in hyperglycemic, diabetic patients. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Apolipoprotein E; Glycation; Heparin binding activity; Diabetes

Abbreviations: AL, amyloid L-peptide; apo, apolipoprotein ; DETAPAC, diethylenetriaminepentaacetic acid; ELISA, enzyme-linked immunosorbent assay; HDL, high density lipoprotein(s) ; HSPG, heparan sulfate proteoglycan(s); Kd , dissociation constant of equilibrium; LDL, low density lipoprotein(s); LRP, LDL receptor-related protein; RP-HPLC, reverse phase high performance liquid chromatography; MALDI-TOF-MS, matrix-assisted laser desorption ionization time-of-£ight mass spectrometry; SPR, surface plasmon resonance; VLDL, very low density lipoprotein(s) * Corresponding author. Fax: +81 (6) 68793429; E-mail: [email protected] 0925-4439 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 4 4 3 9 ( 9 9 ) 0 0 0 4 7 - 2

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

1. Introduction Atherosclerosis represents a major complication in diabetes and is a major cause of morbidity among diabetic patients [1^3]. Several lipid abnormalities accompanying diabetes, including increased levels of VLDL and LDL and decreased levels of HDL have been proposed as a major contributing factor in atherogenesis in diabetic patients. These abnormalities are severer in patients under poor, glycemic control and tend to decline at conditions of well-controlled glucose homeostasis [4^7]. Glycation, a non-enzymatic reaction between the free amino groups of protein and glucose, is believed to be an important atherogenic factor in diabetic patients [8], and it has been shown that the major plasma lipoproteins, VLDL, LDL, and HDL, are susceptible to glycation. Moreover, the glycation of speci¢c apolipoproteins A-I, A-II, B, C-I and E has also been described in hyperglycemic, diabetic patients [9^12]. Extensive glycation of apoB reduces its a¤nity to heparin and can totally block the recognition of LDL by the LDL receptor [13]. The glycation of apoA-I alters the self-association and lipid-binding properties of the protein [14] and also leads to the abnormal association of apoA-I with HDL and decreased lecithin-cholesterol acyltransferase (LCAT) activation by apoA-I in vivo [15,16]. Glycated HDL shows a decrease in its functional ability to bind to the HDL receptor-binding site on cells and to promote intracellular cholesterol e¥ux [17,18]. In contrast, glycation of VLDL, which contains several apolipoproteins as components, has been less well studied. It has, however, been shown that glycation decreases both the hydrolysis of VLDL by lipoprotein lipase and its degradation in skin ¢broblasts [19,20]. Although the glycation of plasma apoE has been detected in hyperglycemic, diabetic patients, the functional consequences of the process have not yet been assessed. ApoE is a key lipid-transport apolipoprotein in plasma and an important cholesteroltransport protein in the brain [21,22], serving as a ligand for several types of receptors, such as the LDL receptor and the LDL receptor-related protein (LRP). The binding of apoE to cellular heparan sulfate proteoglycans (HSPG) is a critical sequestration event in LRP-mediated lipoprotein uptake [23]. The

297

high-a¤nity binding of apoE to Alzheimer AL has also been established [24^26]. Levels of apoE, as well as apoE:apoB ratios are signi¢cantly increased in the plasma triglyceride-rich lipoprotein fractions from donors with non-insulin-dependent diabetes mellitus [27]. The present paper describes a study of the glycation of apoE in vivo and in vitro. Glycation was monitored as the formation of the Amadori product, the earliest stable product of the Maillard reaction, which represents the most important product for a protein such as apoE with a short half-life [28]. In vivo, glycated apoE was found in the VLDL fraction of both normal subjects and diabetic patients. However, the extent of apoE glycation was 2^3-fold higher in hyperglycemic, diabetic patients, as compared to non-diabetic subjects. The rate of apoE glycation in vitro was similar to that of albumin, and the glycation of apoE signi¢cantly impaired its binding to heparin as estimated by surface plasmon resonance (SPR), a biosensor technology used for biospeci¢c interaction studies [29,30]. Moreover, we identi¢ed Lys-75 as the most likely site of apoE glycation. 2. Materials and methods 2.1. Materials and lipoprotein isolation Recombinant human apoE isoforms were puri¢ed from Escherichia coli as described previously [31]. Synthetic AL [1^40] peptide was purchased from Bachem (Switzerland). Sensor chips CM5 and SA were obtained from Biacore AB (Sweden). Native heparin was obtained from Sigma (USA), and polyclonal anti-apoE antibodies were purchased from Calbiochem (UK). Plasma was obtained from fasting donors and diabetic patients. VLDL (d 6 1.006 g/ml) were isolated from plasma by ultracentrifugation in an RP 65-859 rotor (Hitachi, Japan) at 40 000 rpm for 23 h at 18³C. ApoE3 was puri¢ed from normal and diabetic plasma VLDL as described elsewhere [32]. 2.2. Preparation of glycated proteins Proteins at a ¢nal concentration of 50 Wg/ml were

BBADIS 61852 30-7-99

298

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

incubated in a 50 mM phosphate bu¡er, 0.15 M NaCl, 5 mM sodium azide, pH 7.4, supplemented with 100 mM D-glucose at 37³C. Control samples were incubated in the same bu¡er without glucose. In some experiments 1 mM DETAPAC was added to prevent oxidation. Afterwards the incubation samples were frozen until used in assays. 2.3. Immunoblotting and ELISA of glycated proteins Samples of proteins or VLDL preparations were reduced with sodium borohydride to convert Amadori products to hexitol-lysine derivatives. One Wl of freshly prepared 0.1 M sodium borohydride in 0.01 M NaOH was added to 20 Wl of the sample, which was then incubated for 2 h at room temperature. Reduced samples were subjected to 12% polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate, followed by transferring to nitrocellulose membranes under semi-dry conditions using a Semi-phor TE-70 (Hoefer). The membrane was blocked with 3% BSA and reacted with a monoclonal anti-hexitol-lysine antibody [33] or polyclonal anti-apoE antibodies (Calbiochem). The chemiluminescence method was used for the detection of peroxidase activity with an ECL kit (Amersham, UK). The membranes were exposed to the X-ray ¢lm (Fuji, Japan) and the bands corresponding to the speci¢c proteins were semiquanti¢ed by scanning the ¢lm. Glycation index of apoE was estimated as the Amadori product level determined by ¢lm scanning of the apoE-containing band after normalization to the apoE level of the band. Comparison of glycation indexes of samples in the same membrane was performed. ELISA was performed essentially as previously described [33]. 2.4. Heparin biotinylation Heparin was biotinylated using Biotin-LC-NHS (Pierce, USA) [34]. Five mg of heparin from porcine intestine mucosa (Sigma, USA) was dissolved in 2 ml of 50 mM bicarbonate bu¡er, pH 8.5. Biotinylation was performed in the presence of Biotin-LC-NHS, pre-dissolved in dimethylformamide at a molar ratio of heparin:biotin of 6:1 to exclude overlabeling. The mixture was then incubated for 1 h at room temperature, unreacted biotin was removed and the bu¡er

was changed to 25 mM Tris, pH 8.3, by centrifugation on a Centricon-3 (Amicon, USA). 2.5. Immobilization of ligands on sensor chips SPR experiments were performed on a BIAcore 2000 apparatus (Pharmacia Biosensor, Sweden). AL(1^40) was immobilized on a CM5 sensor chip by amino coupling as described previously [25]. Biotinylated heparin was immobilized on a streptavidinprecoated SA sensor chip by injection of 35 Wl at a concentration of 0.25 mg/ml at a £ow rate of 5 Wl/ min to reach saturation of the surface. 2.6. SPR experiments Binding experiments were performed at 25³C at a £ow rate of 5 Wl/min. The running bu¡er was that used for the glycation reaction. To study the binding of proteins to heparin, 20 Wl of analyte in the running bu¡er was injected over the heparin-containing sensor chip for the association phase (contact time: 4 min). After 4 min of dissociation, the surface was regenerated by injection of 10 Wl of 1.0 M NaCl in phosphate bu¡er, pH 7.4. Interaction of proteins with immobilized AL(1^40) was assessed by injection of 20 Wl of analyte. Regeneration was performed by the injection of 10 Wl of 4.0 M guanidine-HCl in phosphate bu¡er. Dissociation and association rate constants kdiss and kass were calculated by analysis of sensorgrams using BIAevaluation software (Pharmacia Biosensor). The equilibrium dissociation constant Kd was calculated as the ratio of mean kdiss /kass values. 2.7. Preparation of lysyl-endopeptidase-treated peptides and mass spectrometry Control and in vitro glycated apoE were S-carboxymethylated and treated with 1% (w/w) lysyl-endopeptidase (Wako Pure Chemicals) for 24 h at 37³C. The peptides obtained were separated by RP-HPLC on a Cosmosil C18 column (4.6U250 mm, Nacalai Tesque). Chromatography was performed at a £ow rate of 1 ml/min using a 15^75% gradient of acetonitrile. Detection was performed at 214 nm. The resolved peaks were collected, dried and dissolved in 50% acetonitrile. The molecular weight of the iso-

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

299

lated peptides was estimated using matrix-assisted laser desorption ionization time-of-£ight mass spectrometry (MALDI-TOF-MS) on a Voyager DL RP mass spectrometer (PerSeptive Biosystems). Samples for mass spectrometry were prepared in 50% acetonitrile, 0.1% trichloroacetic acid using 0.5% K-cyano4-hydroxycinnamic acid as a matrix [35]. 2.8. Amino acid sequencing Protein sequencing of the separated glycated peptide was performed on a gas-phase protein sequencer G1000A and PTH-amino acids were quanti¢ed by RP-HPLC on a liquid chromatograph (1090, Hewlett-Packard). 3. Results 3.1. ApoE glycation in vivo and its binding to heparin ApoE glycation in human plasma was studied in normal subjects and diabetic patients. Hexitol-lysine contents in the apoE band and the apoE protein contents were estimated by Western blotting using the same membrane with the anti-hexitol-lysine antibody and the anti-apoE antibody, respectively. Western blotting analysis of plasma VLDL demonstrated that apoE was glycated in both control and diabetic samples (Fig. 1A). The extent of glycation was calculated as a ratio of apoE Amadori product to apoE protein. Glycation levels were increased 2^3-fold in hyperglycemic patients, compared to non-diabetic subjects (12.8 þ 2.3 vs. 5.2 þ 2.3, respectively) (Fig. 2, Table 1). Moreover, in vitro glycation of apoE

Table 1 Comparison of normal individuals and diabetic patients n Age (years) Glucose (mg%) Cholesterol (mg%) Triglycerides (mg%) ApoE glycation index (arbitrary units)

Control

Diabetes

3 72 þ 15 89 þ 16 161 þ 44 76 þ 12 5.2 þ 2.3

3 65 þ 6 158 þ 19 158 þ 66 94 þ 18 12.8 þ 2.3

ApoE glycation index was calculated after the analysis of plasma VLDL. Mean and S.D. are shown.

Fig. 1. Comparison of glycation of apoE in vitro and in vivo. (A) ApoE glycation in vitro in the presence of 100 mM glucose after 0, 1 and 7 days (lanes 1^3, respectively; 0.5 Wg apoE), plasma VLDL from normal subjects (lanes 4^6) and hyperglycemic diabetic patients (lanes 7^9; 0.25 Wg apoE). Western blotting analysis was performed using anti-hexitol-lysine antibodies. Molecular masses of recombinant and plasma proteins are 40 and 34 kDa, respectively. (B) Estimation of the level of apoE glycation in vitro as compared to in vivo. The mean of apoE glycation in normal plasma VLDL was taken as 100%. Relative glycation was calculated as the ratio (Amadori product)/(protein) by scanning densitometry of apoE bands on Western blot membranes which were stained using antibodies for Amadori product and for apoE, respectively.

was comparable with the level of apoE glycation that was observed in in vivo glycation (Fig. 1B). Recombinant apoE used in in vitro studies showed lower mobility on the gel because the protein had a His tag consisting 34 amino acids useful for fast puri¢cation [31]. Binding of normal and diabetic apoE to heparin was estimated using SPR to see whether there is any e¡ect of glycation. Samples of apoE in 10 mM phosphate bu¡er, 0.15 M NaCl, pH 7.4, were injected over a sensor chip containing immobilized heparin. Binding of apoE from diabetic patients was signi¢-

BBADIS 61852 30-7-99

300

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

3.2. In vitro glycation of apoE

Fig. 2. ApoE glycation in plasma VLDL of normal individuals and hyperglycemic, diabetic patients. Relative glycation in the groups was calculated as the ratio (Amadori product)/(protein) by scanning densitometry of apoE bands on Western blot membranes after reaction with antibodies for Amadori product and apoE, respectively. Means and S.D. of three individuals in each group are shown.

cantly decreased as compared to normal plasma apoE (Fig. 3), suggesting involvement of the glycation reaction in the decreased a¤nity. Since available apoE proteins from patients were limited, we used the recombinant apoE proteins for detailed analysis in the following experiments.

For in vitro glycation, 50 Wg/ml of the protein in the glycation bu¡er (50 mM phosphate bu¡er, 0.15 M NaCl, 5 mM NaN3 , pH 7.4) was incubated without glucose or in the presence of 25 or 100 mM Dglucose. Western blotting analysis clearly showed that the incubation of apoE at 37³C for 1 week had no e¡ect on its molecular weight (Fig. 4A), indicating that fragmentation of apoE had not occurred. In contrast, Amadori products were accumulated in the apoE sample after incubation with 100 mM glucose as shown by Western blotting of the reduced apoE using an anti-hexitol-lysine antibody (Fig. 4B). 3.3. Quanti¢cation of formation of Amadori products by ELISA Glycation of apoE in 25 and 100 mM glucose showed that the rate of Amadori product formation was a nearly linear function of glucose concentration and incubation time in the ranges employed (Fig. 5A). The glycation rate of apoE was detected to be comparable to that for albumin (Fig. 5B). The results indicate that apoE reacts with and is readily glycated by glucose. A comparison of glycation of three com-

Fig. 3. SPR analysis of the binding of puri¢ed apoE from normal (1) and diabetic (2) plasma VLDL to immobilized heparin. Twenty Wl of 1 WM (34 Wg/ml) apoE in 10 mM phosphate bu¡er, 0.15 M NaCl, pH 7.4, was injected at a £ow rate of 5 Wl/min for association phase, followed by the dissociation phase.

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

301

sulted in the signi¢cant loss of its heparin-binding activity as compared with apoE incubated without glucose. The a¤nity of heparin-apoE interaction was also decreased for glycated apoE. The apparent

Fig. 4. In vitro glycation of apoE. Recombinant apoE3 at a concentration of 50 Wg/ml was incubated in 50 mM phosphate bu¡er, 0.15 M NaCl, 5 mM NaN3 , pH 7.4 in the presence of 100 mM D-glucose at 37³C. Western blot analysis was performed using anti-apoE (A) and anti-hexitol-lysine antibodies (B).

mon apoE isoforms revealed that apoE2 had the highest glycation rate (Fig. 5C). The order of apoE isoforms, in terms of sensitivity for glycation, was as follows: apoE2 s apoE4 s apoE3 . Preliminary experiments also demonstrated that apolipoproteins A-I and A-II exhibited similar rates of Amadori product formation to that of albumin (data not shown). To examine the possible inhibition of apoE glycation by heparin, the protein was incubated with 100 mM glucose in the presence of 1% heparin. ELISA analysis showed that heparin had no detectable e¡ect on total Amadori product accumulation in apoE. However, the inhibition of speci¢c glycation site(s) by heparin cannot, at this time, be excluded. 3.4. E¡ects of apoE glycation on the binding properties to heparin and AL Functional characterization of glycated apoE was performed using SPR. Heparin-binding and ALbinding activities were analyzed. The degree of binding of apoE to heparin was reduced, compared with control apoE incubated without glucose (Fig. 6A and B, respectively). Spikes at the beginning and at the end of injection re£ect a changing of running and sample bu¡ers, i.e. bulk e¡ects. Incubation of apoE in the presence of 100 mM glucose for 1 week re-

Fig. 5. Characterization of glycation of apoE. (A) E¡ect of glucose concentration on the glycation rate of recombinant apoE3 incubated in the absence (1) and in the presence of 25 mM (2) or 100 mM (3) D-glucose. (B) Comparison of glycation rates of BSA (1), human plasma apoE3 (2), and recombinant apoE3 (3) incubated in the presence of 100 mM D-glucose. (C) Comparison of the glycation of apoE isoforms incubated in the presence of 100 mM D-glucose. ELISA was carried out after reducing the samples with sodium borohydride, using anti-hexitol-lysine antibodies.

BBADIS 61852 30-7-99

302

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

Fig. 6. SPR analysis of apoE binding to immobilized heparin. ApoE was incubated in the absence (A) or in the presence of 100 mM glucose (B). Binding of glycated recombinant apoE3 to amyloid L-peptide (C). ApoE was incubated in 100 mM glucose. The incubation time (days) is indicated at the right of each sensorgram. Twenty Wl of the protein sample at a concentration of 50 Wg/ml was injected at a £ow rate of 5 Wl/min for the association phase, followed by the dissociation phase.

Kd for the interaction was calculated to be 15 þ 6 nM for apoE3 without incubation and 43 þ 2 nM for glycated apoE (8 days of incubation) vs. 24 þ 2 nM for matched control apoE. To distinguish possible e¡ects

of glycoxidation, 1 mM DETAPAC was added to apoE samples before incubation. DETAPAC had no e¡ect on the loss of its binding to heparin during glycation. The data suggest that glycation of apoE

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

303

Fig. 7. RP-HPLC of glycated apoE, after treatment with lysyl-endopeptidase, on a Cosmosil C18 column. The eluent was monitored by absorbance at 214 nm (A). Fractions were collected and glycation of the peptides was analyzed by ELISA (B). The most extensively glycated fractions were analyzed by mass spectrometry and by amino acid sequencing.

rather than glycoxidation is involved in the observed diminished binding to heparin. The analysis of apoE binding to AL demonstrated that glycation had no e¡ect on this interaction (Fig. 6C), suggesting that glycation of apoE does not play an important role in its binding to AL. The ¢nding was consistent with our previous observation [25], indicating that hydrophobic interactions are the driving forces for the binding. 3.5. Glucose binding sites on apoE The recombinant apoE had His tag [31] in which no additional lysine residue is present and thus could not be glycated. Since the N-terminal K-amino group of apoE is blocked in the recombinant protein, likely candidates for glycation sites are the O-amino groups of lysine residues. To assess the preferential site(s) of apoE glycation, the glycated protein was cleaved with lysyl-endopeptidase and the apoE peptides sep-

arated by RP-HPLC (Fig. 7A). ELISA analysis of the separated apoE peptides indicated that the peptides were glycated to di¡erent extents and that one peak (fraction 16) was highly glycated (Fig. 7B). The glycated peptides were further characterized by MALDI-TOF-MS and by amino acid sequencing. Fractions 15 and 16 were found to contain the same two peptides. In total these two peaks represent 20% of the total Amadori products detected in the apoE peptides. A comparison of amino acid sequences and molecular mass of these peptides to those of apoE showed that the peptides correspond to apoE peptides 76^95 and 73^95 (Table 2). The fact that a glycated Lys-containing peptide cannot be cleaved by lysyl-endopeptidase and Lys cannot be detected at cycle by an automated amino acid sequencer, indicates that Lys-75 represents the glycated residue. MALDI-TOF-MS data showed an increase of 165 Da in mass for the latter peptide, consistent with formation of the glucose adducts.

BBADIS 61852 30-7-99

304

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

Table 2 Comparison of peptides from the most highly glycated fraction with apoE fragments Cycle 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Estimated mass Predicted mass

Number of fraction 15 (component 1)

15 (component 2)

16

Ser (40.0) Glu (158.3) Leu (125.9) Glu (136.6) Glu (149.0) Gln (109.0) Leu (101.6)

Ala (3.86) Tyr (2.91) Lys (...) Ser (0.94) Glu (149.0) Leu (3.32) Glu (12.6)

2283 2285

2814 2647

Ala (8.59) Tyr (7.53) Lys (...) Ser (1.64) Glu (10.64) Leu (5.73) Glu (5.75) Glu (5.60) Gln (2.92) Leu (4.75) Thr (2.72) Pro (2.54) Val (4.59) Ala (4.55) Glu (3.54) Glu (4.49) Thr (1.84) Arg (3.44) Ala (3.34) Arg (3.95) Leu (2.57) Ser (...) Lys (...) 2813 2647

Yields of each PTH-amino acid are shown in parentheses.

Thus, we conclude that Lys-75 of apoE represents a preferential site for glycation. 4. Discussion Atherosclerosis represents a complication in most diabetic patients [1,2]. Moreover, macrovascular disease is the major cause of death in about 80% of diabetic patients [1,36]. Glycation of apolipoproteins may be a factor in impaired lipid metabolism and lipid abnormalities, such as increased levels of triglyceride-rich lipoproteins and LDL, and decreased HDL levels, which accompany diabetes [3,8]. ApoE glycation in plasma VLDL of control and hyperglycemic diabetic subjects was analyzed. A close correlation between the levels of glycemic control in diabetic patients and the glycation of lipoproteins has been observed [37]. The major portion of glycated apoE has been observed in the triglyceride-

rich lipoproteins of hyperglycemic diabetics [10]. Using monoclonal anti-hexitol-lysine antibodies we were able to detect the presence of glycated apoE in control, non-diabetic subjects, which was not detected in earlier studies [10]. ApoE glycation in VLDL was 2^3-fold higher in hyperglycemic diabetic patients, compared to control subjects. Analysis of apoE heparin-binding activity demonstrated that the protein from diabetic patients reveals signi¢cantly lower binding than the normal protein. The results suggest a possible involvement of apoE glycation in the impaired degradation of VLDL in plasma of poorly controlled diabetics. In vitro apoE glycation was characterized using anti-hexitol-lysine antibodies. These antibodies recognize the reduced Amadori product of the Maillard reaction, which is then converted to advanced glycation end products (AGEs) and oxidation products. Amadori product formation may be the most important criterion of glycation for proteins with a short

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

half-life, such as apoE [28]. A comparison of the rate of apoE glycation with that of proteins such as albumin, apoA-I, and apoA-II suggests that the amino groups of apoE are readily accessible for glycation. Glycated apoE showed a diminished total binding and a¤nity to heparin, indicating that at least some apoE molecules lost their ability to bind to heparin or bound with lower a¤nity. Thus, the lysine residue(s) of the heparin-binding site(s) in apoE may be involved in the Amadori rearrangement. After 1 week of incubation in 100 mM glucose, the heparin-binding activity of apoE was signi¢cantly decreased. It should be noted that the incubation of apoE without glucose also results in some loss of heparin-binding activity perhaps due to partial denaturation. Treatment with DETAPAC demonstrated that inhibition of apoE oxidation did not a¡ect the decrease in its binding to heparin. We concluded that the glycation of apoE, rather than glycoxidation, is responsible for the loss of its binding to heparin. Moreover, earlier we observed that the mild oxidation of apoE does not a¡ect its binding to heparin (unpublished data). Lysine residues in apoE are known to be involved in the binding of apoE to heparin. Mutant forms of ApoE with additional lysines show increased a¤nity to heparin and proteoglycans [38,39]. In contrast, lack of lysine may result in loss of heparin-binding activity [40]. Similar results were obtained in vitro [41]. Thus, blocking of O-amino groups in lysine residues due to glycation is likely involved in the loss of the heparin binding activity of the protein. However, possible conformational changes due to the amino acid modi¢cation cannot be excluded either. The heparin-binding activity of apoE is strongly linked to its HSPG-binding activity. The latter is, in turn, a critical interaction for the sequestration of apoEcontaining lipoproteins on the cell surface, a limiting step in the uptake and degradation of triglyceriderich lipoproteins via the LDL receptor-related protein (LRP)-mediated pathway [42] as well as in HSPG independent pathway [43,44]. The binding of apoE to HSPG has also been proposed to be an important factor in the e¡ects of apoE on cell proliferation [45,46]. Moreover, the heparin-binding site of apoE is part of a larger receptor-binding site, which suggests that apoE glycation may signi¢cantly a¡ect the interaction of apoE-containing lipoproteins with apoE receptors. Further studies should elucidate

305

whether apoE glycation may change the recognition of apoE-containing lipoproteins by HSPG and apoE receptors. The data herein show that apoE is glycated in an isoform-speci¢c manner. ApoE2 was the isoform most sensitive to glycation. It is well known that individuals with apoE2 homozygosity may develop type III hyperlipoproteinemia if they develop additional risk factors such as obesity or diabetes [21,23]. Moreover, the impairment of apoE heparin-binding activity in apoE variants correlates with the severity of type III hyperlipoproteinemia [40]. Thus, it might be proposed that high susceptibility of apoE2 to glycation explains, in part, the hyperlipoproteinemia observed in diabetic patients with this isoform. SPR analysis of apoE-AL interactions demonstrated that glycation has no e¡ect on binding, indicating that the lysine residues of apoE have little, if any, involvement in this interaction. The data are in accordance with evidence which suggests that apoE binds to AL via hydrophobic forces [25,47]. In addition, several clinical studies failed to ¢nd a link between diabetes mellitus and Alzheimer's disease [48,49]. Preferential sites of glycation have previously been determined for apoA-I and apoB. Lys-239 has been proposed as the preferential site of apoA-I glycation [50] and the 67 amino acid region which is located at 1791 residues in the NH2 -terminal, putative LDL receptor-binding domain has been shown to be the major modi¢cation site in apoB, as evidenced by the presence of advanced glycosylation end products [51]. In the present work Lys-75 was found to be the preferred site of apoE glycation. The modi¢cation of Lys-75 was estimated to be responsible for 20% of the total apoE glycation. The reason for the high reactivity of the lysyl residue is obscure. The local charge density of the region surrounding Lys75 in the apoE molecule may suppress its protonation. The region is rich in glutamate residues (Glu70, Glu-77, Glu-79, Glu-80) [52], which may increase the reactivity of the lysine O-amino group and stimulate the Amadori rearrangement [53]. The high accessibility of Lys-75 to glucose due to its localization in the molecule might also be a reason for its reactivity. It is interesting to note that the distance between the N of the Lys-75 O-amino group and O of the Glu-79 carboxy group is quite small and corre-

BBADIS 61852 30-7-99

306

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

lates with the glycation rate of apoE isoforms. Using molecular simulation, the distance between Lys-75 and Glu-79 was calculated to be 2.78, 4.85 and î in apoE2 , apoE3 and apoE4 , respectively. 4.40 A Thus, apoE2 , the most sensible form to glycation, displays the closest localization to the putative catalyst. The role of this lysyl residue in the binding of apoE to heparin is not clear. Fragments apoE(129^ 169) and apoE(202^243) have been shown to be the sequences which are critical for heparin binding [41,54,55]. However, participation of lysyl residues other than those localized in these fragments cannot be ignored in apoE-heparin binding. The heparinbinding fragments of apoE contain several lysyl residues that may be modi¢ed by Amadori rearrangement. In conclusion, the glycation of apoE signi¢cantly decreases its heparin-binding activity suggesting an impairment of HSPG binding as well. ApoE was glycated in an isoform-speci¢c manner. Nevertheless, the precise mechanism for this remains to be elucidated. Lys-75 was identi¢ed as a preferred site of apoE glycation. Amadori products were detected in VLDL apoE of both non-diabetic and hyperglycemic diabetic subjects and diabetic apoE revealed lower binding to heparin. A higher apoE glycation level in hyperglycemic conditions indicates the possible involvement of apoE glycation in lipid abnormalities in diabetic patients under poor glycemic control.

[3]

[4]

[5]

[6]

[7]

[8] [9]

[10]

[11]

[12]

[13]

Acknowledgements We thank Drs. Ayumu Hoshi, Nobuko Miyazawa, Kristophor Baka and Norihiko Nakano for their technical assistance. This work was supported, in part, by a Grant-in-Aid for Scienti¢c Research (B) from the Ministry of Education, Science, Sports and Culture, Japan, and Research Grant (9A-2) for Nervous and Mental Disorders from the Ministry of Health and Welfare, Japan. References

[14]

[15]

[16]

[17]

[1] E.L. Bierman, Atherogenesis in diabetes, Arterioscler. Thromb. 12 (1992) 647^656. [2] R. O'Brien, K. Timmins, The role of oxidation and glycation

[18]

in the pathogenesis of diabetic atherosclerosis, Trends Endocrinol. Metab. 5 (1994) 329^334. B. Guerci, O. Ziegler, P. Drouin, Hyperlipidemia during diabetes mellitus. Recent developments, Presse Med. 23 (1994) 82^88. G. Schonfeld, C. Birge, J.P. Miller, G. Kessler, J. Santiago, Apolipoprotein B levels and altered lipoprotein composition in diabetes, Diabetes 23 (1974) 827^834. J. Gabor, M. Spain, N. Kalant, Composition of serum verylow-density and high-density lipoproteins in diabetes, Clin. Chem. 26 (1980) 1261^1269. A. Yegin, T. Ozben, H. Yegin, Glycation of lipoproteins and accelerated atherosclerosis in non-insulin-dependent diabetes mellitus, Int. J. Clin. Lab. Res. 25 (1995) 157^161. R. Bucala, Z. Makita, G. Vega, S. Grundy, T. Koschinsky, A. Cerami, H. Vlassara, Modi¢cation of low density lipoprotein by advanced glycation end products contributes to the dyslipidemia of diabetes and renal insu¤ciency, Proc. Natl. Acad. Sci. USA 91 (1994) 9441^9445. C.A. Colaco, B.J. Roser, Atherosclerosis and glycation, BioEssays 16 (1994) 145^147. L.K. Curtiss, J.L. Witztum, A novel method for generating region-speci¢c monoclonal antibodies to modi¢ed proteins. Application to the identi¢cation of human glucosylated low density lipoproteins, J. Clin. Invest. 72 (1983) 1427^ 1438. L.K. Curtiss, J.L. Witztum, Plasma apolipoproteins AI, AII, B, CI, and E are glucosylated in hyperglycemic diabetic subjects, Diabetes 34 (1985) 452^461. M. Panteghini, R. Bonora, F. Pagani, Determination of glycated apolipoprotein B in serum by a combination of a¤nity chromatography and immunonephelometry, Ann. Clin. Biochem. 31 (1994) 544^549. W. Kortlandt, H.J.M. van Rijn, D.W. Erkelens, Glycation and lipoproteins, Diabetes Nutr. Metab. 6 (1993) 231^ 239. A. Gugliucci-Creriche, A.J.C. Stahl, Glycation and oxidation of human low density lipoproteins reduces heparin binding and modi¢es charge, Scand. J. Clin. Lab. Invest. 53 (1993) 125^132. C. Calvo, C. Talussot, G. Ponsin, F. Berthezene, Non enzymatic glycation of apolipoprotein A-I. E¡ects on its selfassociation and lipid binding properties, Biochem. Biophys. Res. Commun. 153 (1988) 1060^1067. C. Calvo, C. Verdugo, Association in vivo of glycated apolipoprotein A-I with high density lipoproteins, Eur. J. Clin. Chem. Clin. Biochem. 30 (1992) 3^5. C. Calvo, N. Ulloa, R. Del Pozo, C. Verdugo, Decreased activation of lecithin :cholesterol acyltransferase by glycated apolipoprotein A-I, Eur. J. Clin. Chem. Clin. Biochem. 31 (1993) 217^220. P.B. Duell, J.F. Oram, E.L. Bierman, Nonenzymatic glycosylation of HDL resulting in inhibition of high-a¤nity binding to cultured human ¢broblasts, Diabetes 39 (1990) 1257^ 1263. P.B. Duell, J.F. Oram, E.L. Bierman, Nonenzymatic glyco-

BBADIS 61852 30-7-99

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308

[19]

[20]

[21]

[22] [23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32]

[33]

sylation of HDL and impaired HDL-receptor-mediated cholesterol e¥ux, Diabetes 40 (1991) 377^384. J.C.L. Mamo, L. Szeto, G. Steiner, Glycation of very low density lipoprotein from rat plasma impairs its catabolism, Diabetologia 33 (1990) 339^345. M. Yamamoto, S. Ranganathan, B.A. Kottke, Metabolism of glycosylated very-low-density lipoproteins in human skin ¢broblasts, Biochim. Biophys. Acta 875 (1986) 410^413. R.W. Mahley, Apolipoprotein E: cholesterol transport protein with expanding role in cell biology, Science 240 (1988) 622^630. K.H. Weisgraber, E. Apolipoprotein, Structure-function relationships, Adv. Protein Chem. 45 (1994) 249^302. R.W. Mahley, M.M. Hussain, Apolipoprotein E: impact of cytoskeletal stability in neurons and the relationship to Alzheimer's disease, Curr. Opin. Lipidol. 2 (1991) 170^ 176. W.J. Strittmatter, A.M. Saunders, D. Schmechel, M. Pericak-Vance, J. Enghild, G.S. Salvesen, A.D. Roses, Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease, Proc. Natl. Acad. Sci. USA 90 (1993) 1977^1981. V.V. Shuvaev, G. Siest, Interaction between human amphipathic apolipoproteins and amyloid beta-peptide: surface plasmon resonance studies, FEBS Lett. 383 (1996) 9^12. A.A. Golabek, C. Soto, T. Vogel, T. Wisniewski, The interaction between apolipoprotein E and Alzheimer's amyloid beta-peptide is dependent on L-peptide conformation, J. Biol. Chem. 271 (1996) 10602^10606. M. Eto, A. Chait, E.L. Bierman, Postprandial remnant-rich lipoprotein fractions from Type 2 diabetics (NIDDM) enhance incorporation of 14-C oleate into cholesteryl esters in human macrophages, Arteriosclerosis 10 (1990) 787a. R.E. Gregg, L.A. Zech, E.J. Schaefer, D. Stark, D. Wilson, H.B. Brewer Jr., Abnormal in vivo metabolism of apolipoprotein E4 in humans, J. Clin. Invest. 78 (1986) 815^821. U. Jo«nsson, L. Fa«gerstam, B. Ivarsson, B. Johnsson, R. Karlsson, K. Lundh, S. Lo«fas, B. Persson, H. Roos, I. Ronnberg, S. Sjolander, E. Stenberg, S. Stahlberg, C. Urbaniczky, H. Ostlin, M. Malmqvist, Real-time biospeci¢c interaction analysis using surface plasmon resonance and a sensor chip technology, Biotechniques 11 (1991) 620^627. M. Malmqvist, Biospeci¢c interaction analysis using biosensor technology, Nature 361 (1993) 186^187. A. Barbier, A. Visvikis, F. Mathieu, L. Diez, L.M. Havekes, G. Siest, Characterization of three human apolipoprotein E isoforms (E2, E3 and E4) expressed in Escherichia coli, Eur. J. Clin. Chem. Clin. Biochem. 35 (1997) 581^589. K.H. Weisgraber, S.C. Rall, R.W. Mahley, Human apoE heterogeneity. Cys-Arg interchanges in the amino acid sequence of apoE isoform, J. Biol. Chem. 256 (1981) 9077^ 9083. T. Myint, S. Hoshi, T. Ookawara, N. Miyazawa, K. Suzuki, N. Taniguchi, Immunological detection of glycated proteins in normal and streptozotocin-induced diabetic rats using anti

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

307

hexitol-lysine IgG, Biochim. Biophys. Acta 1272 (1995) 73^ 79. C. Foxall, K.R. Holme, W. Liang, Z. Wei, An enzymelinked immunosorbent assay using biotinylated heparan sulfate to evaluate the interactions of heparin-like molecules and basic ¢broblast growth factor, Anal. Biochem. 231 (1995) 366^373. M. Takahashi, Y. Lu, T. Myint, J. Fujii, Y. Wada, N. Taniguchi, In vivo glycation of aldehyde reductase, a major 3-deoxyglucosone reducing enzyme: identi¢cation of glycation sites, Biochemistry 34 (1995) 1433^1438. M.J. Garcia, P.M. McNamara, T. Gordon, W.B. Kannel, Morbidity and mortality in diabetics in the Framingham population. Sixteen year follow-up study, Diabetes 23 (1974) 105^111. J.D. Wagner, J.D. Bagdade, K.N. Litwak, L. Zhang, A.D. Bell-Farrow, Z.Q. Wang, W.T. Cefalu, Increased glycation of plasma lipoproteins in diabetic cynomolgus monkeys, Lab. Anim. Sci. 46 (1996) 31^35. G. Feussner, H. Scharnagl, C. Scherbaum, J. Acar, J. Dobmeyer, J. Lohrmann, H. Wieland, W. Marz, Apolipoprotein E5 (Glu212CLys): increased binding to cell surface proteoglycans but decreased uptake and lysosomal degradation in cultured ¢broblasts, J. Lipid Res. 37 (1996) 1632^1645. T. Yamamura, L.M. Dong, A. Yamamoto, Characterization of apolipoprotein E7 (Glu244CLys, Glu245CLys), a mutant apolipoprotein E associated with hyperlipidemia and atherosclerosis, J. Lipid Res. 40 (1999) 253^259. W.A. Mann, N. Meyer, W. Weber, S. Meyer, H. Greten, U. Beisiegel, Apolipoprotein E isoforms and rare mutations: parallel reduction in binding to cells and to heparin re£ects severity of associated type III hyperlipoproteinemia, J. Lipid Res. 36 (1995) 517^525. A.D. Cardin, D.A. Demeter, H.J. Weintraub, R.L. Jackson, Molecular design and modeling of protein-heparin interactions, Methods Enzymol. 203 (1991) 556^583. R.W. Mahley, Z.S. Ji, Remnant lipoprotein metabolism: key pathways involving cell-surface heparan sulfate proteoglycans and apolipoprotein E, J. Lipid Res. 40 (1999) 1^16. I.V. Fuki, K.M. Kuhn, I.R. Lomazov, V.L. Rothman, G.P. Tuszynski, R.V. Iozzo, T.L. Swenson, E.A. Fisher, K.J. Williams, The syndecan family of proteoglycans. Novel receptors mediating internalization of atherogenic lipoproteins in vitro, J. Clin. Invest. 100 (1997) 1611^1622. M. Al-Haideri, I.J. Goldberg, N.F. Galeano, A. Gleeson, T. Vogel, M. Gorecki, S.L. Sturley, R.J. Deckelbaum, Heparan sulfate proteoglycan-mediated uptake of apolipoprotein Etriglyceride-rich lipoprotein particles: a major pathway at physiological particle concentrations, Biochemistry 36 (1997) 12766^12772. T. Vogel, N.G. Guo, R. Guy, N. Drezlich, H.C. Krutzsch, D.A. Blake, A. Panet, D.D. Roberts, Apolipoprotein E: a potent inhibitor of endothelial and tumor cell proliferation, J. Cell. Biochem. 54 (1994) 299^308. M.A. Clay, G.M. Anantharamaiah, M.J. Mistry, A. Balasubramaniam, J.A.K. Harmony, Localization of a domain in

BBADIS 61852 30-7-99

308

[47]

[48]

[49]

[50]

V.V. Shuvaev et al. / Biochimica et Biophysica Acta 1454 (1999) 296^308 apolipoprotein E with both cytostatic and cytotoxic activity, Biochemistry 34 (1995) 11142^11151. W.J. Strittmatter, K.H. Weisgraber, D.Y. Huang, L.M. Dong, G.S. Salvesen, M. Pericak-Vance, D. Schmechel, A.M. Saunders, D. Goldgaber, A.D. Roses, Binding of human apolipoprotein E to synthetic amyloid beta peptide: isoform-speci¢c e¡ects and implications for late-onset Alzheimer disease, Proc. Natl. Acad. Sci. USA 90 (1993) 8098^ 8102. G.W. Small, M.J. Rosenthal, Coexistence of Alzheimer's disease and diabetes mellitus, J. Am. Geriatr. Soc. 40 (1992) 1075^1076. J.D. Curb, B.L. Rodriguez, R.D. Abbott, H. Petrovitch, G.W. Ross, K.H. Masaki, D. Foley, P.L. Blanchette, T. Harris, R. Chen, L.R. White, Longitudinal association of vascular and Alzheimer's dementias, diabetes, and glucose tolerance, Neurology 52 (1999) 971^975. C. Calvo, N. Ulloa, M. Campos, C. Verdugo, M. AyraultJarrier, The preferential site of non-enzymatic glycation of human apolipoprotein A-I in vivo, Clin. Chim. Acta 217 (1993) 193^198.

[51] R. Bucala, R. Mitchell, K. Arnold, T. Innerarity, H. Vlassara, A. Cerami, Identi¢cation of the major site of apolipoprotein B modi¢cation by advanced glycosylation end products blocking uptake by the low density lipoprotein receptor, J. Biol. Chem. 270 (1995) 10828^10832. [52] R.W. Mahley, T.L. Innerarity, S.C. Rall, K.H. Weisgraber, Plasma lipoproteins: apolipoprotein structure and function, J. Lipid Res. 25 (1984) 1277^1294. [53] N. Iberg, R. Fluckiger, Nonenzymatic glycosylation of albumin in vivo. Identi¢cation of multiple glycosylated sites, J. Biol. Chem. 261 (1986) 13542^13545. [54] A.D. Cardin, N. Hirose, D.T. Blankenship, R.L. Jackson, A.K. Harmony, Binding of a high reactive heparin to human apolipoprotein E: identi¢cation of two heparin-binding domains, Biochem. Biophys. Res. Commun. 134 (1986) 783^ 789. [55] K.H. Weisgraber, S.C. Rall, R.W. Mahley, R.W. Milne, L. Marcel, J.T. Sparrow, Human apolipoprotein E. Determination of the heparin binding sites of apolipoprotein E3, J. Biol. Chem. 261 (1986) 2068^2076.

BBADIS 61852 30-7-99