Low Density Lipoprotein Receptor-related Protein and Factor IXa Share Structural Requirements for Binding to the A3 Domain of Coagulation Factor VIII*

Low-density lipoprotein receptor-related protein (LRP) is an endocytic receptor that binds multiple dis-tinct ligands, including blood coagulation factor VIII (FVIII). FVIII is a heterodimeric multidomain protein that consists of a heavy chain (domains A1, a1, A2, a2, and B) and a light chain (domains a3, A3, C1, and C2). Both chains contribute to high-affinity interaction with LRP. One LRP-interactive region has previously been located in the C2 domain, but its affinity is low in com-parison with that of the entire FVIII light chain. We now have compared a variety of FVIII light chain derivatives with the light chain of its homolog FVa for LRP binding. In surface plasmon resonance studies employing LRP cluster II, the FVa and FVIII light chains proved different in that only FVIII displayed high-affinity binding. Because the FVIII a3-A3-C1 fragment was effective in associating with LRP, this region was explored for structural elements that are exposed but not conserved in FV. Competition studies using synthetic peptides suggested that LRP binding involves the FVIII-specific region Lys 1804 –Ala 1834 in the A3 domain. In line with this observation, LRP binding was inhibited by a recombinant FVIII binding for empty microtiter to wells containing immobilized LRP A patient with inhibitory antibodies against FVIII (32). It is directed against the FVIII C2 domain and competes with monoclonal antibody CLB-CAg117 for FVIII binding (32, 40). scFv EL14 inhibits neither FVIII procoagulant activity nor FVIII light chain binding to LRP. This is in agreement with the observation that the affinity of LRP cluster II for the directly immobilized FVIII light chain ( K d (cid:8) 56 (cid:9) 12 n M ) or the scFv EL14-captured FVIII light chain ( K d (cid:8) 58 (cid:9) 7 n M ) was similar upon SPR analysis.

Coagulation factor VIII (FVIII) 1 serves its role in the intrinsic coagulation pathway as a cofactor for factor IXa (FIXa) in the proteolytic activation of factor X (for reviews, see Refs. 1 and 2). Functional absence of FVIII is associated with the bleeding disorder hemophilia A. The cofactor is a 300-kDa glycoprotein that comprises a discrete domain structure (A1-a1-A2-a2-B-a3-A3-C1-C2) (2,3). The A and C domains share 30 -40% homology with the A and C domains of the structurally related protein factor V (FV), whereas the B domain and the short acidic regions a1, a2, and a3 are unique to FVIII (4).
In plasma, FVIII circulates as a metal ion-linked heterodimer consisting of a 90 -220-kDa heavy chain (A1-a1-A2-a2-B) and an 80-kDa light chain (a3-A3-C1-C2) (5,6). The inactive protein is tightly associated with its carrier protein, von Willebrand factor (7). Limited proteolysis by either thrombin or factor Xa (FXa) converts the FVIII precursor into its activated derivative (8,9). The B domain and the acidic region that borders the A3 domain are then removed from the molecule (10), which leads to the loss of high-affinity binding to von Willebrand factor (7). The resulting FVIIIa molecule consists of a heterotrimer comprising the A2-a2 domain that is noncovalently associated with the metal ion-linked A1-a1/A3-C1-C2 moiety (10).
Within the heavy and light chains of FVIII, several regions have been identified as FIXa-interactive sites (11)(12)(13). A2 domain residues Arg 484 -Phe 509 , Ser 558 -Gln 565 , and Arg 698 -Asp 712 contribute to binding of the heavy chain to FIXa (11,12,14). Within the FVIII light chain, the A3 domain region Glu 1811 -Lys 1818 has been identified as a FIXa-interactive site (13). In addition, FVIII regions Arg 484 -Phe 509 and Lys 1804 -Lys 1818 have also been identified as target epitopes for antibodies that may occur in hemophilia A patients. Such antibodies inhibit FVIII activity by interfering with the complex assembly of FVIIIa and FIXa (15)(16)(17).
Recently, it has been demonstrated that FVIII interacts with the multifunctional endocytic receptor low-density lipoprotein receptor-related protein (LRP) (18,19). It is suggested that this receptor plays a role in the clearance of FVIII from the circulation (19,20). LRP is a member of the low-density lipoprotein receptor family, which also includes the low-density lipoprotein receptor, the very low-density lipoprotein receptor, apoE receptor-2, and megalin (for reviews, see Refs. 21 and 22). It is expressed in a variety of tissues, including liver, lung, placenta, and brain (23). The receptor consists of an extracellular 515-kDa ␣-chain that is noncovalently linked to a transmembrane 85-kDa ␤-chain (24). The ␣-chain contains four clusters of a varying number of complement-type repeats that mediate the binding of many structurally and functionally unrelated ligands (25)(26)(27). The FVIII light chain has been demonstrated to interact with recombinant LRP clusters II and IV, whereas no binding was observed to LRP clusters I and III (27).
Within FVIII, both the heavy and light chains contain LRPinteractive sites. Both the A2 domain region Arg 484 -Phe 509 and a so far unidentified region within the light chain are involved in the high-affinity interaction with LRP (18,19). Previously, we showed that an anti-C2 domain antibody inhibits highaffinity binding of the FVIII light chain to LRP, suggesting a major role for the C2 domain in the interaction (18). However, the isolated recombinant C2 domain demonstrates low-affinity binding to LRP compared with the intact FVIII light chain (18). In the present study, we investigated the apparently paradoxical role of the C2 domain in the interaction of the FVIII light chain with LRP. To this end, the interaction between the FVIII light chain and LRP is addressed using purified recombinant FVIII light chain fragments, synthetic peptides, recombinant antibody fragments, and a chimeric FVIII light chain variant. This approach allowed us to identify the FVIII light chain region Glu 1811 -Lys 1818 as a sequence that contributes to the interaction with LRP.
Recombinant Proteins-The plasmid pCLB-BPVdB695, encoding the FVIII B domain deletion variant FVIII-⌬(868 -1562), has been described previously (37) and was used as a template to construct the plasmid coding for the FVIII-(1811-1818)/FV chimera. Oligonucleotide primers derived from the FVIII light chain sequence containing the FVIII/FV codon replacements (see Table II) were employed to construct the plasmids using the overlap extension PCR mutagenesis method (38). Sequence analysis was performed to verify the presence of the mutations in the plasmid. Transfection of FVIII-encoding plasmids into murine fibroblasts (C127) cells was performed as described previously (37). Stable cell lines expressing wild-type FVIII or the FVIII-(1811-1818)/FV chimera were maintained in cell factories in RPMI 1640 medium supplemented with 5% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, 1 g/ml amphotericin B, and 0.8 g/ml deoxycholate. FVIII-containing medium was harvested three times/week. The medium was subsequently filtered to remove cell debris and concentrated ϳ10-fold employing a hollow fiber cartridge (Hemoflow F5, Fresenius, Bad Homburg, Germany). Benzamidine was added to a final concentration of 10 mM, and concentrates were stored at Ϫ20°C. FVIII was purified from the concentrated medium by immunoaffinity chromatography employing antibody CLB-CAg117 and Q-Sepharose chromatography according to an established procedure (37). FVIII light chains were prepared by incubating the purified FVIII-(1811-1818)/FV chimera and wild-type FVIII in buffer containing 40 mM EDTA, 100 mM NaCl, and 50 mM Tris (pH 7.4) for 4 h at 25°C. Subsequently, the FVIII-(1811-1818)/FV and wild-type FVIII light chains were purified by Q-Sepharose chromatography. Recombinant proteins were eluted in buffer containing 1 M NaCl and 50 mM Tris (pH 7.4), dialyzed against 150 mM NaCl and 50 mM Tris (pH 7.4), and stored at 4°C. The construction of the plasmid encoding the recombinant C2 domain (i.e. Ser 2173 -Tyr 2332 ) has been described previously (16). The plasmid pACgp67b-His-a3-A3-C1, encoding the FVIII a3-A3-C1 fragment (i.e. Glu 1649 -Asn 2172 ), was constructed by PCR employing the oligonucleotide primers 5Ј-TTACTCGAGGAAATAACTCGTACTACTC-3Ј (sense) and 5Ј-AATGCGGCCGCTTCAATTTAAATCACAGCCCAT-3Ј (antisense) using pCLB-BPVdB695 as a template (37). The amplified DNA fragment was purified, digested with XhoI and NotI, and ligated into pBluescript. The resulting construct was verified by sequencing. Subsequently, pBluescript-a3-A3-C1 was digested with EspI and NotI, and the obtained fragment was purified and ligated into the EspI/NotIdigested pACgp67b-80K plasmid (39). A DNA fragment encoding a polyhistidine tag (5Ј-ATTGGATCCGGCCATCATCATCATCATCATGG-CGGCAGCCCCCGCAGCTTTCAAAAGCCCGGGGCCATGGGA-3Ј) was digested with BamHI and NcoI and cloned into the BamHI/NcoIdigested pACgp67b-a3-A3-C1 plasmid. Using the baculovirus expression system, recombinant a3-A3-C1 and C2 fragments were obtained by infection of insect cells as described (16). The a3-A3-C1 fragment was purified from Insect-XPRESS medium by immunoaffinity chromatography using the anti-A3 domain antibody CLB-CAgA coupled to CNBr-Sepharose 4B as an affinity matrix. CLB-CAgA-Sepharose was incubated with medium containing the a3-A3-C1 fragment for 16 h at 4°C. After binding, the immunoaffinity matrix was collected; washed with buffer containing 1 M NaCl and 50 mM Tris (pH 7.4); and eluted with 150 mM NaCl, 55% (v/v) ethylene glycol, and 50 mM lysine (pH 11). Elution fractions were immediately neutralized with 1 M imidazole (pH 6); dialyzed against 150 mM NaCl, 50% (v/v) glycerol, and 50 mM Tris (pH 7.4); and stored at Ϫ20°C. The recombinant C2 domain was purified employing the same immunoaffinity chromatography technique, except that the anti-C2 domain antibody CLB-CAg117 was used instead of CLB-CAgA.
Expression and Purification of Recombinant LRP Fragments-Recombinant LRP clusters II and IV were expressed in baby hamster kidney cells using Opti-MEM I medium supplemented with 100 units/ml penicillin and 100 g/ml streptomycin (27). After harvesting of the medium, CaCl 2 was added to a final concentration of 10 mM. Purification of LRP clusters II and IV from the conditioned medium was performed by a single purification step using glutathione S-transferasefused receptor-associated protein coupled to CNBr-Sepharose 4B as an affinity matrix. The matrix was collected in a column; washed with buffer containing 150 mM NaCl, 5 mM CaCl 2 , and 50 mM Hepes (pH 7.4); and eluted with 150 mM NaCl, 20 mM EDTA, and 50 mM Hepes (pH 7.4). Subsequently, purified LRP cluster preparations were concentrated in Centricon 10 concentrators (Millipore Corp., Bedford, MA) by successive rounds of centrifugation at 4000 ϫ g for 1 h at 4°C. Finally, the preparations were dialyzed against 150 mM NaCl, 2 mM CaCl 2 , and 20 mM Hepes (pH 7.4) and stored at 4°C.
Solid-phase Binding Assays-Recombinant LRP cluster II or IV (1 pmol/well) was adsorbed onto microtiter wells in 50 mM NaHCO 3 (pH 9.8) in a volume of 50 l for 16 h at 4°C. Wells were blocked with 2% (w/v) HSA, 150 mM NaCl, 5 mM CaCl 2 , and 50 mM Tris (pH 7.4) in a volume of 200 l for 1 h at 37°C. Subsequently, the FVIII light chain was incubated at various concentrations in a volume of 50 l of buffer containing 150 mM NaCl, 5 mM CaCl 2 , 1% (w/v) HSA, 0.1% (v/v) Tween 20, and 50 mM Tris (pH 7.4) for 2 h at 37°C. After three rapid washes (Ͻ5 s each) with 150 mM NaCl, 5 mM CaCl 2 , 0.1% (v/v) Tween 20, and 50 mM Tris (pH 7.4), bound ligand was detected by incubation with peroxidase-conjugated monoclonal antibody CLB-CAg12 in the same buffer for 15 min at 37°C. During this latter incubation period, one would expect that the FVIII light chain would completely dissociate from the immobilized LRP clusters. However, subsequent rebinding of the FVIII light chain to the LRP clusters allows the formation of new FVIII⅐LRP cluster complexes, which can be detected by this sensitive method. In surface plasmon resonance (SPR), this is prevented by a continuous buffer flow. Because of these differences in experimental approach, the solid-phase binding assay is compared only with SPR analysis in a qualitative manner. Antibody CLB-CAg12 did not interfere with binding of FVIII fragments to LRP or its clusters (data not shown). In competition experiments, the FVIII light chain (25 nM) was incubated with wells containing immobilized LRP clusters either in the presence or absence of serial dilutions of competitor in a volume of 50 l for 2 h at 37°C. Residual FVIII binding was detected as described above. Data were corrected for binding to empty microtiter wells, which was Ͻ5% relative to binding to wells containing immobilized LRP clusters.
Surface Plasmon Resonance-The kinetics of protein interactions was determined by SPR analysis using a BIAcore TM 2000 biosensor system (BIAcore AB, Uppsala). LRP (16 fmol/mm 2 ), the FVIII light chain (71 fmol/mm 2 ), the a3-A3-C1 fragment (67 fmol/mm 2 ), the FVa light chain (76 fmol/mm 2 ), or scFv EL14 (67 fmol/mm 2 ) was covalently coupled to the dextran surface of an activated CM5 sensor chip via primary amino groups using the amine coupling kit (BIAcore AB, Uppsala, Sweden) as recommended by the supplier. One control flow channel was routinely activated and blocked in the absence of protein.
Association of analyte was assessed in 150 mM NaCl, 2 mM CaCl 2 , 0.005% (v/v) Tween 20, and 20 mM Hepes (pH 7.4) for 2 min at a flow rate of 20 l/min at 25°C. Dissociation was allowed for 2 min in the same buffer flow. Sensor chips were regenerated by several pulses of either 100 mM H 3 PO 4 or 20 mM EDTA, 1 M NaCl, and 50 mM Hepes (pH 7.4) at a flow rate of 20 l/min. The association (k on ) and dissociation (k off ) rate constants were determined using BIAevaluation Version 3.1 software (BIAcore AB). Data were corrected for bulk refractive index changes and fitted by nonlinear regression analysis according to a oneor two-site binding model. Equilibrium dissociation constants (K d ) were calculated from the ratio k off /k on . The K d value for low-affinity interactions was estimated by steady-state affinity analysis using BIAevaluation software. In competition experiments, the FVIII light chain (50 nM) was incubated with immobilized LRP (16 fmol/mm 2 ) either in the presence or absence of serial dilutions of competitor for 2 min at a flow rate of 20 l/min at 25°C.
Binding of the FVIII-(1811-1818)/FV Light Chain to LRP Cluster II-The recombinant FVIII-(1811-1818)/FV or recombinant wild-type FVIII light chain was coupled to immobilized scFv EL14 to a density of 20 fmol/mm 2 in buffer containing 150 mM NaCl and 50 mM Tris (pH 7.4). scFv KM36 (100 nM), scFv KM41 (40 nM), or LRP cluster II (25-125 nM) was passed over separate channels with the immobilized FVIII-(1811-1818)/FV or recombinant wild-type FVIII light chain, respectively, and one control channel (scFv EL14-coated) in 150 mM NaCl, 2 mM CaCl 2 , 0.005% (v/v) Tween 20, and 20 mM Hepes (pH 7.4) for 2 min at a flow rate of 20 l/min at 25°C. scFv EL14 was previously isolated from a hemophilia A patient with inhibitory antibodies against FVIII (32). It is directed against the FVIII C2 domain and competes with monoclonal antibody CLB-CAg117 for FVIII binding (32, 40). scFv EL14 inhibits neither FVIII procoagulant activity nor FVIII light chain binding to LRP. This is in agreement with the observation that the affinity of LRP cluster II for the directly immobilized FVIII light chain (K d ϭ 56 Ϯ 12 nM) or the scFv EL14-captured FVIII light chain (K d ϭ 58 Ϯ 7 nM) was similar upon SPR analysis.

Interaction between LRP and FVIII Light Chain Fragments-
We have previously shown that the isolated FVIII C2 domain (i.e. Ser 2173 -Tyr 2332 ) associates with LRP less effectively than the intact FVIII light chain (18). In this study, we explored the possibility that additional sites in the FVIII light chain contribute to LRP binding. To this end, we monitored the interaction of four FVIII derivatives with immobilized LRP by SPR analysis. These derivatives included the FVIII light chain, the a3-A3-C1 moiety (i.e. Glu 1649 -Asn 2172 ), the C-terminal C2 domain, and a FVIII light chain fragment with the N-terminal acidic region deleted by cleavage at Arg 1721 by FXa (i.e. Ala 1722 -Tyr 2332 ).
As shown in Fig. 1, all FVIII fragments displayed time-dependent association with immobilized LRP, followed by dissociation, which appeared to be dose-dependent, as the highest response was observed at the highest LRP density (data not shown). The data showed complex binding behavior in which multiple components were involved. This is the result of not only the inherent biological properties of both LRP (i.e. clusters II and IV) (27) and FVIII, but also the immobilization procedure of LRP. The latter may result in partial blocking of FVIII interaction sites. Therefore, a heterogeneous two-site binding model was required to approximate the binding behavior of the FVIII light chain, the FXa-cleaved light chain, and the a3-A3-C1 fragment with immobilized LRP. The calculated association (k on ) and dissociation (k off ) rate constants that followed from this model were in the same order of magnitude for these fragments (Table I). This resulted in comparable K d values describing a high-affinity and a slightly lower affinity interaction with immobilized LRP, viz. 18 and 59 nM for the FVIII light chain, 22 and 60 nM for the FXa-cleaved light chain, and 26 and 74 nM for the a3-A3-C1 derivative (Table I). In contrast, no LRP binding could be observed by SPR analysis employing C2 domain concentrations Ͻ250 nM, indicating that the affinity for this interaction is low. This is in agreement with our previous study, in which we also demonstrated inefficient binding of the isolated C2 domain to LRP (18). Because of the low affinity, the K d value for this interaction was obtained by two calculation methods: from k on and k off and from steady-state affinity analysis using the response calculated at equilibrium. The affinities that followed from this analysis were similar, viz. 3.6 Ϯ 1.7 M (Table I)  results show that there is a high-affinity LRP-binding site in the A3-C1 region (i.e. Ala 1722 -Asn 2172 ) and a low-affinity site in the C2 domain.
Binding of the FVIII Light Chain to Immobilized LRP Clusters II and IV-A previous study showed that LRP ligandbinding clusters II and IV mediate the interaction with the FVIII light chain (27). In the present study, we used a solidphase binding assay to address the question of whether or not LRP clusters II and IV can compete for binding to the FVIII light chain. As demonstrated in Fig. 2 (inset), the FVIII light chain was able to bind immobilized LRP cluster II in a dosedependent manner. This observation is in agreement with a previous study in which SPR analysis was used to monitor the interaction between LRP cluster II and the immobilized FVIII light chain (27). Competition studies revealed that LRP clusters II and IV competed for binding to the FVIII light chain (Fig. 2). LRP cluster II displayed a dose-dependent inhibition of FVIII light chain binding to immobilized LRP cluster IV. These data imply that LRP clusters II and IV share a similar binding region within the FVIII light chain.
Binding of LRP Cluster II to the Immobilized FVa Light Chain-In view of the known homology between FVIII and FV (4), the question may arise as to whether or not the light chains of FVIII and FVa share LRP cluster II-binding properties. To this end, serial dilutions of LRP cluster II were incubated with immobilized FVIII and FVa light chains upon SPR analysis. As shown in Fig. 3, the light chains of FVa and FVIII proved different in that only FVIII displayed high-affinity binding to LRP cluster II. These observations indicate that the a3-A3-C1 domains of the FVIII light chain contain a high-affinity LRP cluster II-interactive region that is not conserved in the FVa light chain.
Effect of Synthetic Peptides on FVIII Light Chain Binding to LRP Cluster II-We constructed a panel of synthetic peptides that mimic the surface loops of the a3-A3-C1 domains. The observation that the FVa light chain does not efficiently associate with LRP cluster II was used as a selection criterion for construction of synthetic peptides that are unique to FVIII. The solvent accessibility of these loops was verified by hydropathy analysis (13) and by studying the three-dimensional model of the intact FVIII heterodimer (41). The synthetic peptides com-  (Table II). Subsequently, these peptides were tested for their ability to interfere with the interaction between the FVIII light chain and immobilized LRP cluster II. As shown in Table II is corrected for nonspecific binding, which was Ͻ5% relative to coated channels.

TABLE I
Kinetic parameters for binding of the FVIII light chain and its derivatives to immobilized LRP Association and dissociation of various concentrations of the FVIII light chain (10 -250 nM), the 67-kDa fragment (10 -250 nM), the a3-A3-C1 fragment (10 -250 nM), or the C2 domain (500 -2000 nM) to immobilized LRP (16 fmol/mm 2 ) were assessed as described under "Experimental Procedures." The data obtained were analyzed to calculate association (k on ) and dissociation (k off ) rate constants using a one-or two-site binding model. Each class of binding sites is referred to as 1 and 2, respectively. Affinity constants (K d ) were calculated from the ratio k off /k on . Data are based on three to six measurements using at least five different concentrations for each measurement. Data represent the means Ϯ S.D.  (31). These scFv fragments were evaluated for their ability to interfere with the interaction between the FVIII light chain and LRP or LRP cluster II. The first scFv, referred to as scFv KM36, is directed against a region within Gln 1778 -Asp 1840 , but does not require Arg 1803 -Lys 1818 for FVIII binding (31). The second scFv, designated as scFv KM41, is directed against region Arg 1803 -Lys 1818 and inhibits FVIII procoagulant activity (31). As shown in Fig. 4A, scFv KM36 did not affect the interaction between the FVIII light chain and immobilized LRP cluster II. Accordingly, scFv KM36 did not interfere with binding of the FVIII light chain to full-length LRP upon SPR analysis (Fig. 4B, inset). scFv KM36 slightly increased the response of FVIII light chain binding to LRP, which was due to a small increase in mass as the result of FVIII light chain (80 kDa)⅐scFv KM36 (30 kDa) complex formation. In contrast, the presence of scFv KM41 inhibited the binding of the FVIII light chain to LRP cluster II (Fig. 4A). The effect of scFv KM41 on the interaction between the FVIII light chain and LRP was further studied by SPR analysis. As shown in Fig. 4B, association of the FVIII light chain with immobilized LRP was inhibited in the presence of scFv KM41. These data suggest a role for the FVIII A3 domain region Arg 1803 -Lys 1818 in the interaction with LRP.
The FVIII Light Chain Sequence Glu 1811 -Lys 1818 Contains a Binding Site for LRP-The Lys 1804 -Lys 1818 and Tyr 1815 -Ala 1834 synthetic peptides and scFv KM41 are effective inhibitors of FVIII procoagulant activity by interfering with the assembly of the FVIIIa⅐FIXa complex (13,31). As the FVIII A3 domain region Glu 1811 -Lys 1818 contributes to the interaction with FIXa (13), this particular FVIII light chain region was investigated with respect to its role in the interaction with LRP. As the FVa light chain did not interact with LRP cluster II, a FVIII chimera was constructed in which residues Glu 1811 -Lys 1818 were replaced with the corresponding residues of FV (i.e. 1704 SSYTYVWH 1711 ). This chimera is referred to as FVIII-(1811-1818)/FV. The purified chimera was previously evaluated for its ability to mediate the FIXa-dependent activation of factor X. At saturating concentrations of FIXa, replacement of the FVIII region Glu 1811 -Lys 1818 with the corresponding residues of FV abolished Ͼ80% of FVIII activity (42). This indicates that the FVIII A3 domain region Glu 1811 -Lys 1818 is indeed indispensable for FVIII cofactor function.
The light chain of the FVIII-(1811-1818)/FV chimera was isolated and first tested for binding to scFv fragments KM36 and KM41 by SPR analysis. As shown in Fig. 5A, scFv KM36 displayed similar binding to both the FVIII-(1811-1818)/FV and wild-type FVIII light chains, suggesting that the chimera does not suffer from significant structural defects. In contrast, scFv KM41 did not recognize the immobilized FVIII-(1811-1818)/FV light chain, whereas it readily reacted with the immobilized recombinant wild-type FVIII light chain (Fig. 5B). These observations indicate that the FVIII A3 domain region FIG. 4. Effect of scFv fragments on the interaction between the FVIII light chain and LRP. A, the FVIII light chain (LCh; 25 nM) was incubated with immobilized LRP cluster II (1 pmol/well) in a volume of 50 l of 150 mM NaCl, 5 mM CaCl 2 , 1% (w/v) HSA, 0.1% Tween 20, and 50 mM Tris (pH 7.4) in the presence of various concentrations (0 -100 nM) of scFv KM41 (q) or scFv KM36 (E) for 2 h at 37°C. After washing with the same buffer, bound FVIII light chain was quantified by incubation with peroxidase-conjugated anti-FVIII antibody CLB-CAg12 for 15 min at 37°C. Residual binding is expressed as the percentage of binding in the absence of competitor and is corrected for nonspecific binding (Ͻ5% relative to binding to LRP cluster II-immobilized wells). Data represent the means Ϯ S.D. of three experiments. B, the FVIII light chain (50 nM) was incubated with immobilized LRP (16 fmol/mm 2 ) as described in the legend of Fig. 1. Binding was assessed in the absence (curve 1) or presence of increasing concentrations of scFv KM41 (20, 60, 300, and 500 nM (curves 2-5, respectively)). Complexes were allowed to form for 30 min before SPR analysis. Inset, the FVIII light chain (50 nM) was incubated with immobilized LRP (16 fmol/mm 2 ) as described above. Binding was assessed in the absence (solid line) or presence (dashed line) of scFv KM36 (500 nM). RU, resonance units.

TABLE II
Effect of FVIII a3-A3-C1 fragment-derived synthetic peptides on the interaction between the FVIII light chain and LRP cluster II The FVIII light chain (25 nM) was incubated with immobilized LRP cluster II (1 pmol/well) in a volume of 50 l of 150 mM NaCl, 5 mM CaCl 2 , 1% (w/v) HSA, 0.1% Tween 20, and 50 mM Tris (pH 7.4) in the presence or absence of various concentrations of synthetic peptide (0 -1 mM) for 2 h at 37°C. After washing with the same buffer, bound FVIII light chain was quantified by incubation with peroxidase-conjugated anti-FVIII antibody CLB-CAg12 for 15 min at 37°C. Half-maximum inhibition constants (IC 50 ) represent the mean Ϯ S.D. of three experiments. To further address the role of the FVIII light chain region Glu 1811 -Lys 1818 in the interaction with LRP, we studied the interaction between the FVIII-(1811-1818)/FV light chain and LRP cluster II by SPR analysis. To this end, the chimera was compared with the recombinant wild-type FVIII light chain in the interaction with LRP cluster II in terms of affinity. Although some residual binding could be observed, LRP cluster II displayed a reduced binding signal to the immobilized FVIII-(1811-1818)/FV light chain (Fig. 6B) compared with the intact FVIII light chain (Fig. 6A). This suggests that a substantial part of the LRP-binding site was lost as a result of the mutagenesis. To investigate the importance of the A3 domain region Glu 1811 -Lys 1818 in LRP binding in more detail, the kinetic parameters that describe both interactions were addressed. LRP cluster II bound the wild-type FVIII light chain with an affinity of 58 Ϯ 7 nM (k off ϭ (1.9 Ϯ 0.2) ϫ 10 Ϫ2 s Ϫ1 and k on ϭ (3.3 Ϯ 0.3) ϫ 10 5 M Ϫ1 s Ϫ1 ). Strikingly, LRP cluster II interacted with the FVIII-(1811-1818)/FV light chain with 4 -5-fold lower affinity compared with the wild-type FVIII light chain (k off ϭ (4.3 Ϯ 0.4) ϫ 10 Ϫ2 s Ϫ1 , k on ϭ (1.6 Ϯ 0.4) ϫ 10 5 M Ϫ1 s Ϫ1 , and K d ϭ 262 Ϯ 62 nM). These data demonstrate that the A3 domain region Glu 1811 -Lys 1818 contributes to the assembly of the FVIII light chain⅐LRP complex. DISCUSSION In this study, we have demonstrated that the A3 domain region Glu 1811 -Lys 1818 of the FVIII light chain contributes to the high-affinity interaction with LRP. Several lines of evidence support this conclusion. First, the A3 domain-derived synthetic peptides Lys 1804 -Lys 1818 and Tyr 1815 -Ala 1834 affected the interaction between the FVIII light chain and LRP cluster II (Table II). Second, a recombinant scFv directed against region Glu 1811 -Lys 1818 inhibited binding of the FVIII light chain to LRP or its cluster II fragment (Fig. 4). Third, a chimeric FVIII light chain variant in which residues Glu 1811 -Lys 1818 were replaced with the corresponding residues of FV displayed a reduction in affinity for LRP cluster II compared with the wild-type FVIII light chain (Fig. 6).
To date, no clear consensus sequence that mediates the interaction with LRP has been identified. However, for a number of LRP ligands, including receptor-associated protein, lipoprotein lipase, and ␣ 2 -macroglobulin, it has been established that positively charged residues at the ligand surface are involved in the interaction with LRP (43)(44)(45)(46). Interestingly, the FVIII A3 domain region Glu 1811 -Lys 1818 (i.e. 1811 ETKTYFWK 1818 ) also contains two exposed positively charged lysine residues at positions 1813 and 1818. Based on the sequence of the inhibitory peptides Lys 1804 -Lys 1818 and Tyr 1815 -Ala 1834 (Table II), a possible involvement of the overlapping 1815 YFWK 1818 region in the interaction with LRP is further suggested. However, compared with the homolog part within the A3 domain of FV (i.e. 1704 SSYTYVWH 1711 ), the lysine residues appear to be unique to the FVIII A3 domain (4). Replacement of FVIII residues Glu 1811 -Lys 1818 with the corresponding residues of FV resulted in impaired binding to LRP cluster II (Fig. 6). These results suggest that positively charged residues within the FVIII region Glu 1811 -Lys 1818 mediate an electrostatic interaction with LRP.
To date, two amino acid regions within the FVIII light chain that contribute to the assembly of the FVIII light chain⅐LRP complex have been identified. Besides a role for the A3 domain region Glu 1811 -Lys 1818 found in this study, also the C-terminal C2 domain is known to contribute to the interaction with LRP (18). The LRP-interactive site in the A3 domain seems more predominant than the one in the C2 domain, as the isolated C2 domain exhibited a low-affinity interaction with LRP (K d Ϸ 3.6 M) (Table I). This is in agreement with a previous study in which the isolated C2 domain showed only modest association with LRP (18). In addition, the affinity for FVIII light chain binding to LRP was not affected upon deletion of the C2 domain (Table I). However, it has been demonstrated that an anti-C2 domain monoclonal antibody (ESH4) completely inhibits the interaction between the FVIII light chain and LRP (18). The mechanism by which antibody ESH4 inhibits LRP binding is not yet elucidated. Because the anti-C2 antibody does not require region Glu 1811 -Lys 1818 for its interaction with the FVIII light chain (47), it is unlikely that ESH4 competes with LRP for binding to the same site in the A3 domain. Therefore, one of the  mechanisms that could contribute to the inhibition includes steric interference. In contrast, scFv KM41 only partially inhibited the interaction between the FVIII light chain and LRP (Fig. 4). This might be due to the relative small size of a scFv (Ϸ30 kDa) compared with a complete antibody (Ϸ150 kDa). These observations suggest that, besides region Glu 1811 -Lys 1818 , other surface-exposed structural elements within the A3-C1 domains (i.e. Ala 1722 -Asn 2172 ) contribute to the assembly of the FVIII light chain⅐LRP complex. This is in line with the observation that the FVIII-(1811-1818)/FV light chain demonstrated residual binding to LRP cluster II (Fig. 6). The fact that replacement of FVIII residues Glu 1811 -Lys 1818 with residues of FV affected the affinity by 4 -5-fold and as such had relatively limited impact on the binding energy of the FVIII light chain-LRP interaction also suggests that LRP binding involves an additional or more extended binding site. In this context, it should be mentioned that region Glu 1811 -Lys 1818 within the A3 domain of the FVIII light chain is part of a larger segment that is exposed to the protein surface (i.e. Glu 1804 -Lys 1818 ) (13). Besides the lysine residues at positions 1813 and 1818, this region contains two additional unique FVIII lysine residues at positions 1804 and 1808, which might play a role in the interaction with LRP.
Besides the high-affinity LRP-interactive site within the FVIII light chain, the A2 domain region Arg 484 -Phe 509 within the FVIII heavy chain is also thought to contain residues critical for the interaction with LRP (19). Interestingly, an antibody that is directed against region Arg 484 -Phe 509 completely inhibits the interaction between intact FVIII and LRP (19). This observation suggests a minor role of the light chain in the interaction between the intact FVIII heterodimer and LRP. However, similar to the inhibitory mechanism of ESH4, steric interference cannot be excluded. Both LRP-interactive regions within the FVIII A2 and A3 domains are positioned at the same flank of the FVIII molecule, as judged from the three-dimensional model of the intact FVIII heterodimer (41). As such, both regions might cooperate in the high-affinity interaction with LRP. Therefore, it seems conceivable that full-length antibodies directed against this flank of the FVIII protein are able to completely inhibit FVIII binding to LRP. Similar inhibitory mechanisms have been reported for the assembly of the FVIIIa⅐FIXa complex (15)(16)(17)31). Further studies, preferably employing site-directed mutagenesis, are required to establish the relative importance of both FVIII subunits in the assembly of the FVIII⅐LRP complex.
The complementary FVIII-interactive regions on the LRP molecule remain to be elucidated. It is well established that the ligand-binding regions within LRP are located within its clusters (27). The FVIII light chain is known to be equally effective in its association with recombinant LRP clusters II and IV, whereas no interaction was observed using LRP cluster I or III (27). Moreover, both LRP clusters II and IV were demonstrated to compete for binding to the FVIII light chain (Fig. 2). It is therefore unclear which of the two LRP clusters is involved in the interaction with the FVIII light chain. However, as judged from the three-dimensional model of the membrane-bound FVIII heterodimer, region Glu 1811 -Lys 1818 is positioned in close proximity to the membrane surface (41). It is therefore conceivable that LRP cluster IV, being located closest to the membrane surface, is more important in the interaction with FVIII compared with LRP cluster II. The preferential binding of the FVIII A2 domain to individual LRP clusters remains unknown. This leaves the possibility that one or more LRP clusters accommodate the binding to intact FVIII. The latter model is compatible with the recent observation that two separate LRP clusters cooperate to mediate high-affinity binding to ␣ 2 -macroglobulin (48).
Intriguingly, both LRP-interactive sites within the FVIII A3 and A2 domains were previously identified as FIXa-interactive regions (13,14,19,49). This suggests that LRP binding to these domains involves the same structural elements that also contribute to the interaction between FVIII and FIXa. Therefore, one interesting question is whether LRP plays a regulatory role in the assembly of the FVIIIa⅐FIXa complex (50). Such a modulating role for LRP is further suggested by the observation that LRP also recognizes FIXa as a ligand (51). Whether or not LRP is capable of binding both FVIIIa and FIXa simultaneously and to what extent the FVIIIa⅐FIXa complex is dissociated by LRP remain speculative and need further study.