Two-step Mechanism of Binding of Apolipoprotein E to Heparin IMPLICATIONS FOR THE KINETICS OF APOLIPOPROTEIN E-HEPARAN SULFATE PROTEOGLYCAN COMPLEX FORMATION ON CELL SURFACES*

The interaction of apolipoprotein E (apoE) with cell-surface heparan sulfate proteoglycans is an important step in the uptake of lipoprotein remnants by the liver. ApoE interacts predominantly with heparin through the N-terminal binding site spanning the residues around 136–150. In this work, surface plasmon resonance analysis was employed to investigate how amphipathic (cid:1) -helix properties and basic residue organization in this region modulate binding of apoE to heparin. The apoE/heparin interaction involves a two-step process; apoE initially binds to heparin with fast association and dissociation rates, followed by a step exhibiting much slower kinetics. Circular dichroism and surface plasmon resonance experiments using a disulfide-linked mutant, in which opening of the N-terminal helix bundle was prevented, demonstrated that there is no major secondary or tertiary structural change in apoE upon heparin binding. Mutations of Lys-146, a key residue for the heparin interaction, greatly reduced the favorable free energy of binding of the first step without affecting the second step, suggesting that electrostatic interaction is involved in the first binding step. Although lipid-free apoE2 tended to bind less than apoE3 and apoE4, there were no significant differences in rate and The apoE-enriched remnant lipoproteins are first captured through the fast association of apoE with cell-surface HSPG. Because of the fast dissociation rate of this interaction, remnant lipoproteins are rapidly transferred to the LRP to form a stable complex (characterized by slow dissociation) for endocytosis. Some initial complexes of HSPG-remnant particle would undergo the second binding step, forming a more stable complex for internalization.

Apolipoprotein E (apoE) 1 is a key protein regulating lipid transport in human plasma and brain (1)(2)(3)(4). It mediates hepatic clearance of remnant lipoproteins as a high affinity ligand for the low density lipoprotein receptor (LDLR) family, including LDLR, LDLR-related protein (LRP), and cell-surface heparan sulfate proteoglycans (HSPGs) (5). In the liver, HSPGs act in concert with LRP to complete the interaction of remnant particles with LRP in a process known as the HSPG-LRP pathway, in which apoE initially interacts with HSPG on the cell surface and is then transferred to the LRP for internalization (6). The ability of apoE to interact with members of the LDLR family and with HSPG can also be significant for cell signaling events (7). Binding of apoE to LRP activates cAMPdependent protein kinase A and inhibits platelet-derived growth factor-stimulated migration of smooth muscle cells (8). Inhibition of smooth muscle cell proliferation by apoE is, on the other hand, mediated by its binding to HSPG (9). In addition, the interaction of apoE with HSPG has been implicated in neuronal growth and repair and, consequently, is involved in the progression of late onset familial Alzheimer's disease (10 -12).
ApoE comprises a single polypeptide chain of 299 amino acid residues and contains two independently folded functional domains, a 22-kDa N-terminal domain (residues 1-191) and a 10-kDa C-terminal domain (residues 222-299) (2,13). The Nterminal domain is folded into a four-helix bundle of amphipathic ␣-helices and contains the LDLR binding region (around residues 136 -150 in helix 4) (14,15). The C-terminal domain also contains amphipathic ␣-helices that are involved in binding to lipoprotein particles (2,16,17). Both the N-and C-terminal domains contain a heparin binding site (18,19). The N-terminal domain site is located between residues 136 and 147, overlapping with the LDLR binding region (20,21), whereas the C-terminal site involves basic residues around lysine 233 (22). Although both sites are functional in the separated fragments, only the N-terminal site is available for interaction in both the lipid-free and lipidated states of the intact apoE molecule (22).
In humans, apoE is a polymorphic protein with three major isoforms, apoE2, apoE3, and apoE4, each differing by cysteine and arginine at positions 112 and 158 (11). ApoE3, the most common form, contains cysteine and arginine at these positions, respectively, whereas apoE2 contains cysteine and apoE4 contains arginine at both sites. Both apoE3 and apoE4 bind with equally high affinity to the LDLR, but apoE2 has defective binding to the LDLR and is associated with type III hyperlipoproteinemia (23). ApoE4 is associated with an increased risk for coronary artery disease and is a major risk factor for Alzheimer's disease (3,24,25).
A cluster of arginine and lysine residues located on the polar face of the fourth amphipathic ␣-helix in the N-terminal bundle represents the binding site for cell-surface receptors (2). Despite the apparent accessibility of these basic residues, interaction of apoE with lipid is necessary for its high affinity binding to the LDLR (26). Recent studies indicate that lipid binding induces opening of the helix bundle in the N-terminal domain (27,28); this increases exposure of lysines 143 and 146 to the aqueous phase and thereby enhances interaction with acidic elements of the LDLR (29). Disruption of the amphipathic ␣-helix spanning residues 140 -150 abolishes LDLR binding, indicating that this structural motif in apoE is critical for function (30). In contrast to binding to the LDLR, the stringency for binding of apoE to the LRP or HSPG appears to be less severe. Lipid association of apoE is not required for binding to the LRP (31) or HSPG (22,32), although the same apoE domain spanning residues 136 -150 is involved in the binding (21,33). ApoE2 that is highly defective in LDLR binding activity (Ͻ2% of normal apoE3 activity) has significant binding activity to LRP (40 -50% of apoE3) (23) and HSPG (50 -90% of apoE3) (6). The detailed molecular features that control high affinity binding of apoE to the LRP and HSPG are not yet defined fully.
Previously, we have characterized the effect of point mutations of the basic residues present in the heparin binding sites on the equilibrium parameters defining interaction of apoE with heparin (22). To predict the distribution of apoE to different receptors, such as LDLR, LRP, and HSPG, however, it is also important to establish the kinetics of apoE-receptor association and dissociation because receptor/ligand interac-tions at the cell surface do not occur at equilibrium. In this study, surface plasmon resonance (SPR) measurements were employed to determine the affinity and kinetics of the interaction of engineered apoE molecules in the lipid-free and lipidated states with heparin. Using this approach, we provide novel insights into the mechanism of apoE/heparin interaction.

EXPERIMENTAL PROCEDURES
Materials-1,2-Dimyristoyl phosphatidylcholine (DMPC) was purchased from Avanti Polar Lipids (Pelham, AL), and stock solutions were stored in chloroform:methanol (2:1) under nitrogen at Ϫ20°C. Brain natriuretic peptide and ␤-mercaptoethanol were from Sigma. Ultrapure guanidine HCl was from ICN Pharmaceuticals (Costa Mesa, CA). Porcine intestinal mucosa heparin (average molecular weight of 13,500 -15,000) and its biotin conjugate were purchased from Calbiochem. Biotinylated heparin was dissolved in water and extensively dialyzed to remove any contaminating free biotin before use.
Expression and Purification of Proteins-The full-length human apoE2, apoE3, apoE4 and their 22-kDa, 12-kDa, and 10-kDa fragments were expressed and purified as described previously (34). All mutants of apoE were made using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The cDNA was ligated into a thioredoxin fusion expression vector pET32a(ϩ) (Novagen, Madison, WI) and transformed into the Escherichia coli strain BL21 star (DE3) (Invitrogen). The resulting thioredoxin-apoE fusion proteins were expressed and purified as described previously (28). For the apoE3 22-kDa (L141K/K143L/K146L/L148K) mutant that was expressed as insoluble protein, the proteins were solubilized from the cell pellet using 6 M urea (30). The apoE preparations were at least 95% pure as assessed by SDS-PAGE. In all experiments, the apoE sample was freshly dialyzed from 1% ␤-mercaptoethanol and 6 M guanidine HCl solution into Tris buffer (10 mM Tris-HCl, 150 mM NaCl, 0.02% NaN 3 , 1 mM EDTA, pH 7.4) before use. The disulfidelinked apoE4 22-kDa mutant was dialyzed from 6 M guanidine HCl solution to maintain interhelical disulfide bonding (27). DMPC complexes with the apoE3 22-kDa variants were prepared as described previously (29).
SPR Measurements-Experiments were performed on Biacore-X and -2000 SPR instruments (Biacore, Inc., Uppsala, Sweden). A streptavidin SA sensor chip was pretreated with three consecutive 5-l injections of 50 mM NaOH in 1 M NaCl to remove nonspecifically bound contaminants. For immobilization of heparin on a SA chip, an injection of biotinylated heparin (10 g/ml) in Tris buffer was made at a flow rate of 5 l/min followed by a 10-l injection of 2 M NaCl. Typically, 30 -200 resonance units of heparin were immobilized, and the effects of mass transport were not significant because of the low surface density of ligand (35). An untreated flow cell was used as a control. For kinetic measurement of apoE interaction with heparin, a 30-l injection of the apoE sample was passed over the sensor surface at a flow rate of 10 l/min. At the end of the sample plug, the same buffer was passed over the sensor surface to facilitate dissociation. After 3 min of dissociation time, the sensor surface was regenerated for the next sample using a 10-l pulse of 2 M NaCl. The resultant sensorgrams were analyzed using BIAevaluation software (version 4.1). The response curves of various analyte concentrations were globally fitted to either the 1:1 Langmuir model or the two-state binding model described by the following equation (36,37), where the equilibrium constants of each binding step are K 1 ϭ k a1 /k d1 and K 2 ϭ k a2 /k d2 , and the overall equilibrium binding constant is calculated as K A ϭ K 1 (1 ϩ K 2 ) and K d ϭ 1/K A . In this model, the analyte (A) binds to the ligand (B) to form an initial complex (AB) and then undergoes subsequent binding or conformational change to form a more stable complex (AB x ). For the apoE 10-and 12-kDa fragments, binding responses in the steady-state region of the sensorgrams (R eq ) were also plotted against analyte concentration (C) to determine the overall equilibrium binding affinity. The data were subjected to nonlinear regression fitting according to the following equation, where R max is the maximum binding response, and K d is the dissociation constant. Circular Dichroism (CD) Spectroscopy-Far UV CD spectra were recorded from 195 to 250 nm at 25°C using a Jasco J-820 spectropolarimeter. After dialyzing from 1% ␤-mercaptoethanol and 6 M guanidine HCl solution, the apoE sample was diluted to 25 g/ml in Tris buffer (pH 7.4) to obtain the CD spectrum. For the apoE/heparin mixture sample, apoE was mixed with heparin (apoE:heparin ratios of 0.5-2 w/w) prior to the measurement. The results were corrected by subtracting the buffer base line or a blank sample containing an identical concentration of heparin. The ␣-helix content was calculated from the molar ellipticity at 222 nm, [] 222 , according to the following equation (38,39).
Analytical Procedures-Protein concentrations were determined by either the Lowry procedure (40) or the absorbance coefficient at 280 nm. Phospholipid concentrations were determined with an enzymatic assay kit (Wako Chemicals, Richmond, VA).

Comparison of ApoE Binding to Immobilized Heparin and
Heparin-Sepharose Gel-We have previously characterized the effects of point mutations of lysines in the N-and C-terminal heparin binding sites on the equilibrium binding to heparin using heparin-Sepharose gel (22). In this study, we made SPR measurements for the apoE/heparin interaction to examine the real-time binding kinetics. Fig. 1A shows sensorgrams of binding of apoE C-terminal fragments to the immobilized heparin on a sensor chip. Because binding equilibrium was achieved during the injection of the 10-and 12-kDa fragments of apoE, we determined binding isotherms from the relationship between the equilibrium binding response (R eq ) and protein concentration (Fig. 1B). These isotherms were similar to those obtained using heparin-Sepharose gel (Fig. 1C), and calculated K d values from both methods were comparable (for example, K d for the apoE 10-kDa fragment/heparin interaction are 29 and 76 g/ml from Fig. 1, B and C, respectively); these results validate the SPR method for studies of the apoE/heparin interaction (41).
Kinetic Analysis of Heparin Binding to Full-length ApoE3 and Its 22-and 10-kDa Fragments- Fig. 2 shows a typical sensorgram for the binding of full-length apoE to heparin. The kinetic data were not fitted well by a 1:1 Langmuir binding  (Fig. 2, inset). However, significantly improved fit ( 2 ϭ 7.4) was obtained using a two-state binding model, indicating that binding of apoE to heparin involves either a sequential two-step process or some conformational change (36,37). The response curve of heparin binding of the apoE3 22-kDa fragment also showed two-state binding kinetics (Fig. 3A). Changing the injection time revealed that the dissociation rate was progressively decreased after longer injection (contact) time (Fig. 3A, inset), indicating that the stability of the initial apoE-heparin complex increases over time (36,37). Such an effect of contact time on dissociation rate was not seen in the apoE 10-kDa binding to heparin, because the contribution of the second binding step to the overall process was very small (Fig. 3B). Table I summarizes the kinetic rate constants and the derived affinity constants for full-length apoE3 and its 22and 10-kDa fragments obtained using the two-state binding model. There was good agreement in the K d values for fulllength apoE3 between our data and a previous report (32) but some discrepancy for the apoE3 22-kDa fragment (21), because all previous studies applied the 1:1 Langmuir binding model to the kinetic data. In addition, the K d value for the 10-kDa fragment in Table I (15 g/ml) is similar to that obtained by the steady-state analysis (29 g/ml from Fig. 1B), further validating the two-state binding model.
Effects of Heparin Binding on the Secondary and Tertiary Structure of ApoE-To examine the possibility that the conformational change in apoE occurs during binding to heparin, far UV CD measurements were employed to evaluate the secondary structure of apoE. As shown in Fig. 4, no change in the spectra of the apoE 22-kDa fragment was observed in the absence or presence of heparin. The ␣-helix contents derived from the molar ellipticity at 222 nm were in the range of 47-49% up to a heparin: apoE weight ratio of 2 (Fig. 4, inset), indicating that there is no change in the secondary structure in apoE upon heparin binding. The four-helix bundle of the apoE 22-kDa fragment is known to open upon lipid binding (27,28). To further confirm that any such conformational change in apoE is not involved in the two-state

Two-step Binding of ApoE to Heparin
binding to heparin, we tested the heparin binding of the triple interhelical disulfide-linked apoE4 22-kDa mutant in which the opening of the four-helix bundle is completely restricted (27). SPR sensorgrams of heparin binding of this mutant still exhibited two-state binding kinetics (data not shown), indicating that a conformational change, such as the opening of the helix bundle, was not involved in the two-state binding for the apoE/heparin interaction.
Effects of Lysine Mutations in the Heparin Binding Sites of ApoE on the Binding Kinetics-To explore the molecular mechanism of the two-state heparin binding of apoE, we used mutants with substitutions at lysines 146 and 233; these residues are located in the N-and C-terminal heparin binding sites, respectively, and contribute to an ionic interaction with heparin (21,22). As shown in Fig. 5, large decreases in binding responses were observed with the apoE3 K146E mutant, whereas the apoE3 K233E mutant exhibited responses similar to the wild type (WT), consistent with the previous observations using heparin-Sepharose gel in which only Lys-146 plays a dominant role in heparin binding of full-length apoE (22). Fig.  6A summarizes relative changes in rate constants for each mutant as well as for mutants K146Q and K233Q. Replacement of the lysines significantly affected the first association rate but had no effect on the rate constants for the second step. The relative contributions of each step to the overall free energy of binding were estimated from the affinity constants for the two individual steps (Fig. 6B) (42). Although the first binding step contributes the largest part of the free energy of binding in all proteins, significant decreases in the favorable free energy of binding for the first step were observed in the K146Q and K146E mutants, without any changes in the second step.
Effects of ApoE Isoform on the Binding Kinetics to Heparin-We compared the association and dissociation rate constants and equilibrium dissociation constants for the binding of apoE isoforms to heparin. A previous SPR study indicated that the three full-length apoE isoforms have similar affinities for heparin (32). However, as shown in Fig. 7A, full-length apoE2 exhibited a somewhat slower first association rate and higher K d value than full-length apoE3. Regarding the 22-kDa fragments, although apoE2 displayed a slower rate, and apoE4 a faster rate, in the first association phase compared with apoE3, the K d values were similar among the three isoforms (Fig. 7B). Overall, these results indicate that apoE2 tends to bind to heparin with somewhat less affinity than apoE3 and apoE4 because of the slower first binding step.

Effects of Substitution of Basic Residues and Disruption of ␣-Helical Structure in the N-terminal Heparin Binding Site on
Binding Kinetics of Lipidated ApoE-Previously, we have determined the structural requirements in the apoE receptor binding domain for effective binding to the LDLR by substituting basic residues or disrupting ␣-helical structure in this region (30). To examine the possibility that similar requirements are necessary for heparin binding, we used two apoE3 22-kDa mutants for SPR measurements. The first was a (K143R/K146R) mutant in which the conservative substitution of two lysines with arginines decreases the affinity for the LDLR to ϳ30% of the WT value. The second was a (L141K/K143L/K146L/L148K) mutant in which the exchange of lysines and leucines (which disrupts the ␣-helical structure spanning residues 140 -150 without altering the net charge) eliminates binding to the LDLR (30). Because lipid association is required for the high affinity binding of apoE to the LDLR (26), we used discoidal complexes of apoE 22-kDa

FIG. 4. Far UV CD spectra of apoE3 22-kDa alone (•) or incubated with heparin (broken line). Protein and heparin concentrations were both 25 g/ml.
The inset shows the change in ␣-helix content of apoE3 22-kDa as a function of the weight ratio of heparin to protein (the protein concentration was 25 g/ml).
fragments for SPR measurements to compare their LDLR binding abilities. The apoE3 22-kDa fragment-DMPC discoidal complex displayed a second order higher first association rate constant than the lipid-free protein, resulting in much higher affinity binding to heparin (cf. Tables I and II) (22). Among apoE isoforms complexed with DMPC, there was no significant difference in association and dissociation rate constants and equilibrium affinity (Table II). The mutation (K143R/K146R) had no effect on the kinetics and affinity for binding to heparin, indicating that charge specificity of basic residues is not required for effective binding to heparin. In contrast, the (L141K/K143L/K146L/L148K) mutant exhibited much lower responses in heparin binding than the WT apoE3 22-kDa fragment in the lipidated state (Fig. 8), suggesting that the amphipathic nature of the ␣-helix spanning residues 140 -150 is critical for the heparin binding of apoE. Interestingly, this mutation led to an increased favorable free energy of binding in the second step as well as a decrease in free energy of the first step (Fig. 9), but the overall K d value was not significantly different from WT (Table II). DISCUSSION The ability of apoE to bind to cell-surface glycosaminoglycans, including HSPG and heparin, is important for lipoprotein remnant metabolism. The binding of apoE to heparin has been studied extensively (18 -21), and we have recently defined the dominant role of the N-terminal site (residues 136 -150) in the heparin binding of both lipid-free and lipid-associated apoE (22). However, the structural requirements in this region, such as the basic residue organization and the amphipathic nature of ␣-helix for the high affinity interaction of apoE with heparin as opposed to the LDLR, have not been defined yet. In the present study, we characterized the kinetics of the interaction with heparin for apoE isoforms and various mutants lacking effective LDLR binding affinity. Knowledge of the association and dissociation rate constants obtained here is relevant to the understanding events at a cell surface, because many receptor/ ligand interactions do not occur at equilibrium. Two-step Mechanism of Binding of ApoE to Heparin-Our SPR data indicate that the binding of apoE to heparin is a two-state process involving sequential binding and/or a conformational change. It is known that, when a protein binds to heparin, it can induce a change in conformation within heparin and/or the protein (43). For the interaction between antithrombin and heparin, the initial interaction induces a conformational change in antithrombin that enables additional interactions between antithrombin and heparin, resulting in stronger binding (44). There is also evidence that the relatively flexible 2-sulfoiduronic acid residue in heparin can change its confor-  6. SPR rate constants and free energy of binding for the interaction of apoE3 mutants with heparin. A, relative change in SPR rate constants for each binding step of apoE mutants to heparin. B, free energy for each binding step of apoE mutants to heparin. Free energy was calculated according to ⌬G ϭ ϪRT ln K using binding constants for each step, K 1 and K 2 . *, p Ͻ 0.05; **, p Ͻ 0.01 compared with WT apoE3. mation upon protein binding, resulting in a better fit and enhanced binding (43). However, our CD measurements for the apoE/heparin mixture demonstrated that heparin binding does not induce a change in the secondary structure of apoE (Fig. 4). Taken together with the SPR data for the triple interhelical disulfide-linked apoE4 22-kDa mutant showing two-state binding to heparin, it appears that the two-state apoE/heparin interaction observed in SPR measurements is not due to any major conformational change in apoE upon heparin binding.
The most dominant type of interaction between heparin and proteins is ionic; clusters of basic residues on proteins form ion pairs with spatially defined acidic groups in heparin (45). In fact, molecular modeling for the interaction between the heparin binding site of apoE and heparin fragment suggested that eight polar residues including Lys-146 directly contact with sulfate or carboxyl groups on the heparin chain (21). Thus, the large reductions in the association rate and favorable free energy in the first binding step for Lys-146 mutants (Fig. 6) are consistent with the first step involving fast electrostatic interaction between basic residues in apoE and acidic groups in heparin. In addition, analysis of the dependence of K d values on sodium chloride concentration (data not shown) using the protein/polyelectrolyte interaction theory (46) indicated that there are 3.3 and 2.7 charged residues, respectively, involved in the electrostatic component of the overall and the first binding process in the apoE/heparin interaction. The similarity in these two values also supports the idea that electrostatic interaction is involved in the first binding step.
Besides the ionic interaction, it is known that there is a significant contribution to the binding to heparin by nonionic interactions such as hydrogen bonding and hydrophobic interactions in some cases (45). In fact, molecular modeling of the apoE/heparin interaction predicts formation of hydrogen bonding between some residues such as Lys-143 in apoE and heparin, and also hydrophobic interaction between the shallow groove of the ␣-helix of apoE and the saccharide chains of heparin (21). In this regard, the significant increase in the contribution of the second step to the free energy of binding observed for the (L141K/K143L/K146L/L148K) mutant (Fig. 9), in which disruption of the amphipathic ␣-helix in the heparin binding region destabilizes the N-terminal helix bundle and exposes the hydrophobic surface (30), implies that hydrophobic interaction between apoE and heparin contributes to the second binding step. In addition, we found that SPR data of the interaction of brain natriuretic peptide with heparin, in which the major contribution to the free energy of binding comes from hydrogen bonding interaction (47) displayed two-state binding, and the first binding step contributed most of the overall free energy of binding (data not shown). This suggests that hydrogen bonding may be involved mostly in the first binding step.
Structural Requirements in the Receptor Binding Region of ApoE for Heparin Interaction-Although lipid association is required for high affinity binding of apoE to the LDLR (26), lipid-free apoE can bind effectively to heparin (Table I). In addition, apoE2 that is highly defective in LDLR binding (23) exhibited similar heparin binding activity to apoE3 in both lipid-free and lipidated states ( Fig. 7 and Table II). These results are consistent with previous observations that the requirements for the high affinity binding of apoE to HSPG/ heparin are less stringent than for binding to the LDLR (6). It has been suggested that the multiple basic residues in the receptor binding region (residues 136 -150) of apoE are required to have a particular orientation for optimal binding to the LDLR (48). Indeed, comparison of the structure of residues 135-151 between lipid-free and lipid-associated apoE showed a significantly different curvature and that the basic residues are less scattered, forming a better aligned and more elongated hydrophilic surface in the lipidated form (49,50). In addition, our nuclear magnetic resonance studies (29,51) demonstrated that lipid interaction increases positive electrostatic potential around lysines 143 and 146 and that these effects are much less in apoE2. Taken together, it appears that the enhanced electrostatic potential around residues 136 -150 in apoE, required for the interaction with the LDLR, is not necessary for the high affinity binding to heparin.
Previous studies of the relative strengths of heparin binding of arginine and lysine demonstrated that the former binds 2.5 times more tightly (52). Indeed, we found from SPR measurements that the lipid-free substitution mutant (K143R/K146R) displayed a 2.6 times higher affinity to heparin than WT (data not shown). However, as shown in Table II, this double substitution of lysines with arginines had no effect on the kinetics and affinity of the binding of lipidated apoE to heparin, although this mutant exhibits much reduced binding affinity to  the LDLR (30). It follows that there is no preference between lysine and arginine at these positions for the heparin binding of lipid-associated apoE. Disruption of the ␣-helical structure in region 140 -150 by the mutation (L141K/K143L/K146L/L148K) greatly affected the kinetics of binding to heparin ( Fig. 8 and Table II), but this mutant still displayed overall heparin binding affinity similar to WT (Table II). This suggests that an amphipathic ␣-helical structural motif in the receptor binding region of apoE is not critical for the effective binding to heparin; this is in contrast to the situation with LDLR binding. Such somewhat surprising results may arise because the appropriate arrangement of basic residues required for the interaction with heparin is still maintained in the discoidal complex of this mutant. Presumably, cooperative binding through multiple copies of the heparin binding sites of apoE molecules is important in determining the heparin binding ability of apoE bound to lipid particles (22). Implications for the Sequestration Role of HSPG in Remnant Lipoprotein Metabolism-Our SPR kinetic data for the apoE/ heparin interaction provide novel insights into the mechanism of the HSPG-LRP pathway for remnant lipoprotein uptake by cells (Fig. 10). In this pathway, apoE is postulated to interact initially with cell-surface HSPG and then transfer to the LRP for internalization (6). Because only limited kinetic parameters for the apoE/LRP interaction are available (33), we can only compare our kinetic parameters of binding of lipidated apoE to heparin with those to the related LDLR (53,54). Because of the fast association of lipidated apoE in the first binding step to heparin through long range and nondirectional ionic interactions (Table II), it follows that apoE-enriched remnant particles can be captured rapidly by the abundant HSPG on the cell surface. Fast dissociation in this step is likely to allow a rapid transfer of remnant particles to the LRP associated closely with HSPG (55). The LRP could then retain the remnant particles at the cell surface until endocytosis occurs because of the very slow dissociation rate of the apoE-receptor complex (ϳ10 Ϫ5 s Ϫ1 ) (53). Such a process would be more efficient than direct binding of remnant particles in solution to the LRP, because remnant particles concentrated on the cell surface could diffuse in two dimensions by rapidly exchanging between HSPG binding sites until they collide with the LRP to form the LRP-remnant particle complex. This lateral diffusion on the cell surface is reminiscent of the "scooting mode" proposed for the interfacial reaction of phospholipases (56). It is also possible that some initial complexes of HSPG-remnant particle would undergo the second binding step, forming a more stable complex for internalization without LRP involvement.
In summary, we have characterized the kinetics and affinity of the interaction of apoE with heparin using SPR. Our data show that the binding of apoE to heparin is a two-step process, the fast initial binding involving a polar interaction followed by relatively slow nonpolar interaction. We propose that the fast  Because of the fast dissociation rate of this interaction, remnant lipoproteins are rapidly transferred to the LRP to form a stable complex (characterized by slow dissociation) for endocytosis. Some initial complexes of HSPG-remnant particle would undergo the second binding step, forming a more stable complex for internalization. association and dissociation process in the apoE/HSPG interaction plays a role in the HSPG-LRP pathway, in which remnant lipoproteins are rapidly captured by HSPG and then transferred to the LRP to form a stable complex for internalization.