The alpha -chains of C4b-binding protein mediate complex formation with low density lipoprotein receptor-related protein.

C4b-binding protein (C4BP) is a heparin-binding protein that participates in both the complement and hemostatic system. We investigated the interaction between C4BP and low density lipoprotein receptor-related protein (LRP), an endocytic receptor involved in the catabolism of various heparin-binding proteins. Both plasma-derived C4BP and recombinant C4BP consisting of only its alpha-chains (rC4BPalpha) bound efficiently to immobilized LRP, as determined by surface plasmon resonance analysis. Complementary, two distinct fragments of LRP, i.e. clusters II and IV, both associated to immobilized rC4BPalpha, and binding could be inhibited by the LRP antagonist receptor-associated protein. Further analysis showed that association of rC4BPalpha to LRP was inhibited by heparin or by anti-C4BP antibody RU-3B9, which recognizes the heparin-binding region of the C4BP alpha-chains. In cellular degradation experiments, LRP-expressing fibroblasts effectively degraded (125)I-labeled rC4BPalpha, whereas their LRP-deficient counterparts displayed a 4-fold diminished capacity of degrading (125)I-rC4BPalpha. Finally, initial clearance of C4BP in mice was significantly delayed upon co-injection with receptor-associated protein. In conclusion, our data demonstrate that the alpha-chains of C4BP comprise a binding site for LRP. We propose that LRP mediates at least in part the catabolism of C4BP and, as such, may regulate C4BP participation in complement and hemostatic processes.

C4b-binding protein (C4BP) 1 is a plasma protein, which serves as a regulator of the complement system (1). C4BP binds the complement protein C4b, which results in enhancement of factor I-mediated degradation of C4b and inhibition of the classical pathway C3 convertase (C4b2a) complex (1). In plasma, C4BP may serve as a carrier protein for at least two other plasma proteins: the vitamin K-dependent anticoagulant Protein S and serum amyloid P component (1). In addition, C4BP may interact with coagulation factor VIII as well (2). The majority of the C4BP molecules (ϳ80%) consist of seven identical ␣-chains and a unique ␤-chain, whereas other isoforms lack either one of the ␣-chains or the ␤-chain (3). The ␤-chain is involved in the interaction with Protein S (4), and complex formation with serum amyloid P component (SAP) and complement proteins is mediated by the ␣-chains (5,6). In addition, the ␣-chains have been found to contain binding sites for bacterial surface proteins from Streptococcus pyogenes (7) and heparin (8). The heparin interactive site, which overlaps with the C4b interactive site, encompasses a cluster of positively charged amino acids involving Arg residues at positions 39, 64, and 66 (9).
The average plasma concentration of C4BP is ϳ260 nM (150 g/ml) (10), although its levels may increase up to 4-fold during inflammation, infection, or tissue damage (11,12). Plasma levels represent a balance between C4BP production and removal. At present, little is known concerning the molecular mechanisms that control the removal of C4BP from the circulation. The notion that C4BP is able to interact with heparin opens the possibility that C4BP may interact with heparan sulfate proteoglycans (HSPG) exposed at the cellular surface. Alternatively, C4BP may interact with cellular receptors like the low density lipoprotein receptor-related protein (LRP), which is known to recognize heparin-binding proteins.
LRP, also known as the ␣ 2 -macroglobulin receptor, is a member of the low density lipoprotein receptor family of endocytic receptors (13,14). It consists of a noncovalently linked heavy and light chain. The 85-kDa light chain comprises the transmembrane and cytoplasmic domains, whereas the ligand binding regions are located within the 515-kDa heavy chain (15). The heavy chain contains four domains enriched in low density lipoprotein receptor class A domains, generally referred to as clusters I-IV. It has been reported that clusters II and IV play a prominent role in ligand binding to the receptor (16).
LRP is widely distributed among tissues, such as liver, brain, and placenta, and is expressed in an array of cell types including parenchymal cells, Kupffer cells, neurons, astrocytes, smooth muscle cells, monocytes, and fibroblasts (17). LRP has traditionally been reported as a receptor that is involved in hepatic clearance of numerous proteins (18), although recent studies demonstrate that LRP contributes to cellular signaling processes as well (19). Ligands bound by LRP belong to a spectrum of structurally and functionally unrelated proteins (13,14,18). These include apolipoproteins, lipases, proteinases, proteinase/inhibitor complexes, Kunitz-type inhibitors, matrix proteins, and several others.
The mechanism by which binding to LRP is mediated varies between ligands. First, ligands (e.g. ␣ 2 -macroglobulin/protease complexes) may bind directly from the circulation to LRP (20,21). Alternatively, binding may be promoted by so called accessory proteins. This possibility is exemplified by the urokinasetype plasminogen activator receptor, which facilitates internal-ization of urokinase complexed with its inhibitor plasminogen activator inhibitor-1 by LRP (22,23). Furthermore, LRP-mediated degradation may be preceded by sequestration of the ligands by HSPG. Examples hereof include ␤-amyloid precursor protein (24), tissue factor pathway inhibitor (25), activated factor IX (26), and thrombospondin (27).
In the present study, we assessed binding of C4BP to LRP by surface plasmon resonance (SPR) employing purified components. Our data show that C4BP is able to interact with LRP with moderate affinity, and that binding involves the C4BP ␣-chain and the cluster II and IV regions of LRP. Furthermore, we found that LRP mediates the delivery of rC4BP␣ to the intracellular degradation pathway in mouse fibroblast cells, and that the initial clearance of C4BP in mice is delayed upon co-injection with the LRP-antagonist receptor-associated protein. Our data suggest that LRP contributes to the catabolism of the complement protein C4BP.

EXPERIMENTAL PROCEDURES
Materials-The Biacore2000 biosensor system and reagents, including an amine-coupling kit and CM5 biosensor chips (research grade), were from Biacore AB (Uppsala, Sweden). Cell culture plates, Dulbecco's modified Eagle's medium (DMEM), DMEM/F-12 medium, fetal calf serum, penicillin, and streptomycin were from Invitrogen (Breda, The Netherlands). Unfractionated heparin was purchased from Sigma (Zwijndrecht, The Netherlands). Protein G-Sepharose was from Amersham Biosciences, Inc.
Proteins-Plasma-derived C4BP (pd-C4BP) and Protein S were purified as described (28,29). Recombinant C4BP, consisting of the ␣-chains but lacking the ␤-chain (rC4BP␣), was produced using stably transfected baby hamster kidney cell lines, purified until homogeneity by immunoaffinity chromatography as reported previously (30), and stored in 125 mM NaCl, 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4) at Ϫ20°C. Purified rC4BP␣ was labeled with Na 125 I (Amersham Biosciences, Inc.) using the IODO-GEN method (Pierce) as described (8). Free Na 125 I was removed by chromatography on a PD-10 column (Amersham Biosciences, Inc.) equilibrated in 125 mM NaCl, 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4), and 125 I-labeled rC4BP␣ was stored in small aliquots at Ϫ20°C. Specific radioactivity was 4.0 (Ϯ 1.3) ϫ 10 5 cpm/pmol rC4BP␣ (mean Ϯ S.E.; n ϭ 6). Each radiolabeled rC4BP␣ preparation was compared with unlabeled rC4BP␣ for binding to immobilized LRP employing SPR. In all cases labeled and unlabeled preparations displayed similar sensorgrams, demonstrating that both association and dissociation characteristics were unchanged upon labeling. On one occasion, this was investigated in more detail by determining affinity constants, which proved to be similar for radiolabeled and unlabeled rC4BP␣ (3.5 and 55.2 nM versus 5.6 and 66.2 nM, respectively). Purified full-length LRP (31) was kindly provided by Dr. S. K. Moestrup (University of Aarhus, Aarhus, Denmark). Receptor-associated protein fused to glutathione S-transferase (GST-RAP) (32) was prepared as described previously (33). Recombinant LRP fragments encompassing LRP cluster II and cluster IV were produced using stably transfected baby hamster kidney cell lines, which were kindly provided by Dr. H. Pannekoek (University of Amsterdam, Amsterdam, The Netherlands). Clusters were purified employing GST-RAP affinity chromatography as described (16) and stored at 4°C in 125 mM NaCl, 1 mM CaCl 2 , 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4). Monoclonal antibody RU-3B9 (34) was purified from ascites using protein G-Sepharose as recommended by the manufacturer. RU-3B9 Fab fragments were prepared using the ImmunoPure Fab preparation kit (Pierce) as instructed. Bovine serum albumin (fraction V) was obtained from Sigma. Proteins were quantified by a BCA protein assay (Pierce) using albumin as a standard.
SPR Analysis-Binding studies were performed employing a Bia-core2000 biosensor system, and SPR analysis was done essentially as described (26,33). LRP or rC4BP␣ was immobilized on a CM5 sensor chip at the indicated densities using the amine-coupling kit as instructed by the supplier. Routinely, a control channel was activated and blocked using the amine-coupling reagents in the absence of protein. Binding to coated channels was corrected for binding to noncoated channels (Ͻ5% of binding to coated channels). SPR analysis was performed in 125 mM NaCl, 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4) at 25°C at indicated flow rates. Regeneration of the surface of the LRP sensor chip was performed by subsequent application of 100 mM H 3 PO 4 and 25 mM CaCl 2 . The rC4BP␣ sensor chip was regenerated using 100 mM H 3 PO 4 .
Analysis of Quantitative SPR Data-For analysis of association and dissociation curves of the sensorgrams, BiaEvaluation software was used (Biacore AB). Interaction constants were determined by performing nonlinear global fitting of data corrected for bulk refractive index changes. Data were fitted to various models available within the software. For binding of rC4BP␣ to immobilized LRP, a model describing the interaction between rC4BP␣ and two independent binding sites (heterologous ligand, parallel reactions) was found to provide the best fit of the experimental data. Accuracy of the fits was judged from residual plots and statistical parameters employing previously described equations (35).
Statistical Analysis-Statistical significance of clearance data (see Fig. 4) and of association and dissociation rate constants (see Table I) were calculated using Student's unpaired t test employing the InStat program (GraphPad Software, Inc.).
Cellular Degradation Experiments-Cellular degradation of rC4BP␣ was examined using mouse fibroblast cell lines MEF-1 (American Tissue Culture Collection, CRL-2214), or their counterparts, which are genetically deficient for LRP, PEA-13 (American Tissue Culture Collection, CRL-2216). MEF-1 and PEA-13 cells have been isolated from embryos resulting from the mating of mice heterogenous for LRP gene disruption (36). The MEF-1 cells express LRP endogenously, whereas PEA-13 cells have been demonstrated to contain two (via homologous recombination) disrupted alleles for the LRP gene. Cells were seeded at least 48 h before the experiment and grown to 90 -95% confluence in 24-well plates in DMEM supplemented with 10% (v/v) fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Before incubation, cells were extensively washed with DMEM/F-12 medium. 125 I-labeled rC4BP␣ was mixed with nonlabeled rC4BP␣ to a 1:1 molar ratio, and the mixture was then added to the cells in a final volume of 250 l in incubation medium (DMEM/F-12 medium supplemented with 1% (w/v) bovine serum albumin and 2 mM CaCl 2 ). Final concentration of rC4BP␣ was 100 nM. After a 1-h incubation at 4°C, cells were washed three times with 500 l of incubation medium to remove nonbound material. Subsequently, incubation was allowed to proceed at 37°C in a volume of 250 l. At indicated time points, samples were taken to determine the amount of degraded material. Degraded material is defined as the radioactivity that is soluble in 10% trichloroacetic acid. In all experiments a control was included in which the amount of degradation was examined in the absence of cells.
Clearance of Human C4BP in C57BL/6J Mice-We used 14 -16week-old C57BL/6J mice. Three to six mice were housed in each cage and fed a standard chow diet and water ad libitum. Mice were injected intravenously with pd-C4BP (10.9 mg/kg) alone or in combination with GST-RAP (30 mg/kg) into the tail vein. Blood samples were collected in polypropylene Eppendorf tubes, containing approximately 0.1 volume of 129 mM trisodium citrate, at indicated time points from anesthetized mice by retro-orbital venous plexus puncture. Plasma was prepared by centrifugation of the blood at 2500 ϫ g for 20 min at room temperature. Residual C4BP was determined using an immunosorbent assay specific for human C4BP. For each time point, 3 mice were used.

C4BP
Binding to Immobilized LRP-Binding of pd-C4BP to LRP was investigated by SPR analysis using purified components. An increase in response was observed when pd-C4BP (30 nM) was passed over immobilized LRP (7 and 11 fmol/mm 2 ), demonstrating that pd-C4BP is able to associate with LRP (Fig.  1A). As the highest response is observed at the highest density of LRP, binding appears to be dose-dependent. Replacement of pd-C4BP solution by buffer resulted in a gradual decline of the response, indicating that pd-C4BP dissociates from LRP and that binding is reversible (Fig. 1A). In plasma, C4BP is able to associate with other plasma proteins, like Protein S and SAP. Our data therefore do not fully exclude that the observed response originates from traces of Protein S or SAP present within the pd-C4BP preparation. However, no additional response was observed upon subsequent injection of anti-Protein S or anti-SAP antibodies after injection of C4BP preparations (data not shown). In addition, no association of purified Protein S to immobilized LRP was detected (Fig. 1A). Apparently, the observed increase in response originates from binding of C4BP to LRP.
LRP Binding Involves ␣-Chains of C4BP-To further char-acterize the interaction between C4BP and LRP, SPR analysis was performed employing purified recombinant C4BP that consists only of the ␣-chains (30). As shown in Fig. 1B, recombinant rC4BP␣ (2-30 nM) efficiently associated to immobilized LRP (11 fmol/mm 2 ) in a dose-dependent manner. Binding appeared to be similar for plasma-derived and recombinant C4BP. The interaction between rC4BP␣ and immobilized LRP was studied in more detail by assessment of the apparent association and dissociation rate constants, which are summarized in Table I. Experimental data were fitted most appropriately using a model describing the interaction of rC4BP␣ with two classes of binding sites (heterologous ligand, parallel interactions). The resulting apparent affinity constants (K d(app) ) values were 2.4 Ϯ 1.9 nM and 71.4 Ϯ 42.5 nM, respectively. These data demonstrate that LRP is able to bind the C4BP ␣-chain with moderate affinity in a reversible and dose-dependent manner.
Binding of Recombinant LRP Fragments to Immobilized rC4BP␣-The observation that C4BP interacts with two different binding sites may suggest heterogeneity of LRP because of its immobilization. Alternatively, LRP may comprise distinct regions that are able to interact with C4BP ␣-chains. To identify LRP regions involved in binding C4BP ␣-chains, purified recombinant receptor fragments were used. These fragments, designated cluster II and IV, respectively, have been established to encompass the ligand binding domains of LRP (16). When either cluster (200 nM) was incubated with immobilized recombinant rC4BP␣ (10 fmol/mm 2 ), reversible binding of cluster II and IV to rC4BP␣ was observed ( Fig. 2A). The specificity of the interaction was subsequently assessed by investigating the binding of cluster II or IV to immobilized rC4BP␣ in the presence of various concentrations of the LRP-antagonist GST-RAP. Indeed, GST-RAP (0 -750 nM) efficiently interfered with binding of either cluster (150 nM) to immobilized rC4BP␣ (Fig.  2B). Thus, both recombinant LRP fragments encompassing the ligand binding domains, i.e. clusters II and IV, comprise a binding site for rC4BP␣, and binding is inhibited in the presence of GST-RAP.
Cellular Degradation of 125 I-rC4BP␣ Is Mediated by LRP-The observation that LRP recognizes C4BP in a system employing purified components prompted us to investigate the contribution of LRP to the delivery of rC4BP␣ to the intracellular degradation pathway. This was addressed in experiments using mouse fibroblast cells genetically deficient for LRP (i.e. PEA-13 cells), or their counterparts expressing LRP endogenously (i.e. MEF-1 cells) (36). Employing LRP-expressing MEF-1 cells, an increase in the amount of degraded 125 I-labeled rC4BP␣ was observed in time (Fig. 3). However, when degradation was examined in the presence of 1 M GST-RAP, the amount of 125 I-labeled rC4BP␣ degraded was markedly reduced (Fig. 3), indicating that the degradation process involves a GST-RAP-sensitive receptor. In the presence of LRP-deficient PEA-13 cells, the amount of rC4BP␣ degraded was similar to that of MEF-1 cells in the presence of the LRP antagonist GST-RAP (Fig. 3). Together, these data strongly suggest that the cellular uptake and transport of 125 I-rC4BP␣ to the intracellular degradation pathway involves a LRP-dependent pathway.
In Vivo Clearance of Human pd-C4BP Is Modulated by GST-RAP-To investigate whether LRP also contributes to the clearance of C4BP in vivo, human pd-C4BP (10.9 mg/kg) was injected intravenously into the tail vein of C57BL/6J mice in the absence or presence of GST-RAP (30 mg/kg). At indicated time points, blood samples were collected, and plasma was FIG. 1. Binding of plasma-derived and recombinant C4BP to immobilized LRP. A, LRP immobilized at a CM5 sensor chip at a density of 7 or 11 fmol/mm 2 (lines II and I, respectively) was incubated with pd-C4BP (30 nM) in 125 mM NaCl, 0.005% (v/v) Tween 20, and 25 mM Hepes (pH 7.4) at a flow rate of 5 l/min for 2 min at 25°C. Ligand solution was replaced with buffer to initiate dissociation. Line III represents incubation of immobilized LRP (11 fmol/mm 2 ) with Protein S (400 nM) under similar conditions. B, seven different concentrations (2, 4, 7, 10, 15, 20, and 30 nM) of rC4BP␣ were passed over immobilized LRP (11 fmol/mm 2 ) at 25°C for 3 min at a flow rate of 20 l/min. The subsequent association and dissociation of rC4BP␣ is represented by the seven data curves shown. Lines represent the data curves and their fitted curves obtained using a model for heterologous ligand (parallel reaction) interactions. For both graphs, the signal is indicated in resonance units (RU) and is corrected for aspecific binding, which was less than 5% of binding to LRP-coated channels.

TABLE I
Kinetic parameters for the binding of rC4BP␣ to LRP Association and dissociation of various concentrations rC4BP␣ to LRP (11 fmol/mm 2 ) was assessed as described in the legend of Fig. 1. Buffer consisted of 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4) supplemented with either 50 or 125 mM NaCl. rC4BP␣ concentrations tested varied between 2 and 30 nM. The data obtained for all concentrations tested were analyzed to calculate apparent association rate constants (k on(app) ) and apparent dissociation rate constants (k off(app) ) as described using a two-site binding model (35). Each class of binding sites is referred to as 1 and 2, respectively. Apparent affinity constants (K d(app) ) were inferred from the ratio k off(app) :k on(app) . Data are based on three to six measurements using four or five different concentrations for each measurement. Data represent the average (Ϯ S.D.). p values were calculated employing Student's unpaired t test. subsequently analyzed for residual C4BP content. As shown in Fig. 4, human pd-C4BP was cleared in mice in a biphasic manner with an initial half-life of ϳ50 min, whereas, in the presence of GST-RAP, pd-C4BP was cleared at a slower rate (approximate half-life of 60 min). At the 15-and 30-min time points, the content of C4BP was significantly higher (p ϭ 0.0026 and p ϭ 0.0098, respectively) in the presence of GST-RAP than in the absence of this LRP-antagonist. Apparently, inhibition of GST-RAP-sensitive receptors, like LRP, is associated with a delay in initial clearance of C4BP.
Interaction between rC4BP␣ and LRP Is Ionic Strength-dependent-It has been reported that binding of the C4BP ␣-chains to C4b is strongly ionic strength-dependent, whereas binding of streptococcal M proteins to these ␣-chains is of more hydrophobic nature (9). To investigate the nature of the ␣-chain/LRP interaction, association of rC4BP␣ (20 nM) to immobilized LRP (11 fmol/mm 2 ) was examined in 0.005% (v/v) Tween 20, 25 mM Hepes (pH 7.4), supplemented with various concentrations of NaCl (25-150 mM). As shown in Fig. 5, association of rC4BP␣ to LRP was sensitive to the concentration of NaCl. Optimal association was observed at 35 mM NaCl, whereas association was reduced over 6-fold at NaCl concentrations exceeding 100 mM. To further examine the effect of NaCl on the rC4BP␣/LRP interaction, the apparent association and dissociation rate constants were determined employing 50 mM NaCl. As for the assessment of the rate constants in the presence of 125 mM NaCl, binding of rC4BP␣ to LRP involved two classes of binding sites (Table I). With regard to the calculated rate constants, no significant differences were found, except for the class 2 association rate constant. Class 2 k on(app) determined at 50 mM NaCl proved to differ significantly from class 2 k on(app) determined at 125 mM NaCl (p ϭ 0.024). A 4-fold increase in association rate constant at 50 mM NaCl resulted in a subsequent 4-fold higher affinity of rC4BP␣ for immobilized LRP.
Inhibition of rC4BP␣/LRP Interaction by Monoclonal Antibody RU-3B9 -The observation that the rC4BP␣/LRP interaction displays a sensitivity to NaCl similar to that described for the C4b/C4BP interaction (Fig. 5; Ref. 9) may suggest that C4b and LRP bind to a similar region within the ␣-chains of C4BP. Previously, monoclonal antibody RU-3B9, which is directed against C4BP, has been shown to interfere with the binding of C4b to C4BP. Therefore, the effect of Fab fragments of this antibody (0 -320 nM) on the binding of rC4BP␣ (40 nM) to immobilized LRP (11 fmol/mm 2 ) was tested by SPR-analysis. Fab RU-3B9 indeed inhibited binding of rC4BP␣ to immobi-lized LRP, and binding was fully suppressed at 80 nM RU-3B9 (Fig. 6A). Inhibition appeared to be specific because Fab RU-3B9 was unable to affect binding of another LRP ligand (factor VIII light chain; Ref. 33) to LRP (data not shown). These data indicate that the LRP binding site may overlap with the RU-3B9 binding site within the C4BP ␣-chain.
Binding of rC4BP␣ to LRP Is Inhibited in the Presence of Heparin-It has been reported previously that the ␣-chaindirected antibody RU-3B9 prevents binding of C4BP to heparin (8). It was of interest, therefore, to examine the effect of heparin on binding of rC4BP␣ to LRP. As expected, normal binding was detected in the absence of heparin as assessed by SPR analysis (Fig. 6B). In the presence of increasing concentrations of heparin, however, a decrease of the resonance signal was observed. Half-maximal inhibition was obtained at 0.2 mg/ml heparin. We anticipated that, if binding of rC4BP␣ to LRP was inhibited by heparin or RU-3B9 in a system employing purified components, both components should also interfere with the intracellular degradation of rC4BP␣ by LRP-expressing cells. Indeed, the amount of 125 I-labeled rC4BP␣ by LRP-expressing MEF-1 cells was reduced 4-and 3-fold in the presence of RU-3B9 and heparin, respectively (Fig. 5, A and B, insets). Our data seem to be compatible with a model in which the heparinbinding region within the C4BP ␣-chains overlaps with an interactive site for LRP. DISCUSSION A first step in resolving the pathways that mediate the clearance of proteins is the identification of cellular receptors that control the endocytosis of the protein of interest. In the present study, we provide evidence that the endocytic pathway of the complement regulatory protein C4BP involves LRP, a hepatic receptor established to be responsible for the clearance of various other plasma proteins (14,18). First, in a system employing purified components, C4BP and a recombinant derivative thereof (rC4BP␣) both associated to immobilized LRP in a reversible and dose-dependent manner (Fig. 1). In complementary experiments, we show that soluble, recombinant fragments of LRP associate to immobilized rC4BP␣ and that this association is inhibited in the presence of GST-RAP, an antagonist of ligands binding to LRP (Fig. 2). Further, LRP-deficient mouse fibroblasts display diminished capacity compared with LRP-expressing fibroblasts in degrading rC4BP␣, and cellular degradation of rC4BP␣ by LRP-expressing cells was markedly reduced in the presence of the LRP-antagonist GST-RAP (Fig.  3). Finally, in the presence of GST-RAP, the initial clearance of C4BP was significantly delayed (Fig. 4). C4BP is the third component of the complement pathway that has been reported to be a ligand for LRP. Storm and co-workers (37) reported that C1 inhibitor/C1s complexes are cleared from the circulation in a process that involves LRP. In addition, Meilinger et al. (38) showed that the catabolism of the activated complement component C3 is mediated by LRP. C4BP, however, seems to be different from both these ligands in its interaction with LRP. Whereas C1 inhibitor/C1s and C3 are recognized by LRP after complex formation and proteolysis, respectively, C4BP is recognized by LRP in its native circulating form. As such, C4BP is the first identified member of the complement protein family whose plasma levels and function may be directly regulated by LRP.
Analysis of the SPR data suggested that the interaction between rC4BP␣ and immobilized LRP could be described employing a two-site binding model (Table I). The presence of multiple binding sites within LRP may result from heterogeneity of LRP because of its immobilization onto the biosensor chip. Alternatively, LRP may comprise multiple regions, which are able to interact with rC4BP␣. Indeed, recombinant fragments of LRP comprising distinct regions of the LRP molecule (i.e. clusters II and IV, respectively) proved able to associate with rC4BP␣ ( Fig. 2A) This would allow a continuous formation of the C4BP/LRP complex in vivo and subsequent removal of C4BP from the circulation. It should be mentioned, however, that, to calculate the apparent rate constants, molar concentrations of C4BP based on protein levels were used. Because each molecule contains seven identical ␣-chains (1), the actual number of potential binding sites is 7-fold the molar C4BP concentration. For binding of C4b protein to C4BP, it has been reported that, after four molecules of C4b have bound to C4BP, the binding of additional C4b molecules is sterically hindered (39). In view of this observation, it seems likely that only part of the seven potential binding sites are available for complex formation between C4BP and LRP. In our opinion, it is therefore reasonable to assume that the actual affinities are severalfold less stringent compared with the affinities that are summarized in Table I. The in vivo survival experiments, however, have been performed employing C4BP plasma levels of 190 nM, which is ϳ25% below the levels reported for human plasma. Nevertheless, C4BP was rapidly cleared (Fig. 4), indicating that physiological conditions allow complex formation between C4BP and LRP. It should be mentioned, however, that the clearance experiments do not completely rule out the possibility that other receptors contribute to C4BP clearance as well. First, GST-RAP is not a selective inhibitor of LRP, but recognizes also other members of the low density lipoprotein family, like megalin or apolipoprotein E receptor 2. Second, the notion that C4BP comprises a binding site for heparin suggests that also HSPG may be involved in the cellular uptake of C4BP. Whether this proceeds via direct internalization of C4BP by HSPG or via HSPG-mediated sequestration at the cellular surface before delivery to LRP is currently under investigation.
Complement protein C4b and heparin have been reported to share overlapping binding sites within a region located at the interface between the so called complement control protein domains 1 and 2 of the C4BP ␣-chains (9,40). This region is particularly enriched in positively charged amino acids, and residues Arg-39, Lys-63, and Arg-64 appear to be especially critical for C4b and heparin binding. Apparently, binding of these components to C4BP is mainly mediated by electrostatic interactions, which is supported by the observation that complex formation between C4BP and C4b or heparin is sensitive to NaCl (9, 40). The interaction between rC4BP␣ and LRP displayed a strikingly similar NaCl dependence as described for the C4b/C4BP and heparin/C4BP interactions ( Fig. 5; Refs. 9 and 40). Further, binding of rC4BP␣ to LRP was inhibited effectively by heparin, as well as by an antibody known to inhibit binding of C4BP to both heparin and C4b (Fig. 6). With regard to these observations, it seems conceivable that a LRP interactive site is similar or close to the reported binding site for C4b within the ␣-chains of C4BP. As such, LRP may interfere with C4BP-dependent down-regulation of C4b and the C3 convertase complex.
Apart from its role in the complement system, C4BP is functionally related to the hemostatic system as well. In plasma, C4BP may serve as a carrier protein for the anticoagulant component Protein S, and ϳ60% of the Protein S molecules are noncovalently bound to C4BP (41). When bound to the ␤-chain of C4BP, Protein S is unable to exert its cofactor activity to the Protein C anticoagulant pathway (42), indicating that C4BP contributes to the regulation of Protein S cofactor activity. Although speculative, our finding that the catabolism of C4BP could involve a LRP-dependent process may imply that LRP indirectly affects catabolism of Protein S as well. This would provide an extra link between the hemostatic process and LRP. Previously, it has been reported that LRP contributes to the clearance of thrombin/antithrombin and factor Xa/␣ 2 -macroglobulin complexes (43,44), and the cellular uptake of coagulation proteins factor VIII (33,45) and activated factor IX (26). In addition, LRP contributes to the tissue factor pathway inhibitor-dependent down-regulation of tissue factor/factor VIIa complex at the surface of monocytes and fibroblasts (46,47). The notion that LRP is involved in the cellular uptake of both pro-and anticoagulant proteins may indicate that LRP serves a so far unrecognized role in the regulation of the coagulation process.