HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 24, 22063-22072, June 14, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/24/22063    most recent M201327200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Nakayama, T. Articles by Saito, H. Search for Related Content PubMed PubMed Citation Articles by Nakayama, T. Articles by Saito, H. Social Bookmarking                      What's this?

Identification of the Regulatory Elements of the Human von Willebrand Factor for Binding to Platelet GPIb

IMPORTANCE OF STRUCTURAL INTEGRITY OF THE REGIONS FLANKED BY THE CYS¹²⁷²-CYS¹⁴⁵⁸ DISULFIDE BOND^*

Takayuki Nakayama, Tadashi Matsushita^§, Zhengyu Dong^¶, J. Evan Sadler^¶, Sylvie Jorieux^**, Claudine Mazurier, Dominique Meyer^§§, Tetsuhito Kojima^¶¶, and Hidehiko Saito

From the  First Department of Internal Medicine, Nagoya University School of Medicine, the ^¶ Howard Hughes Medical Institute, Departments of Medicine and of Biochemistry and Molecular Biophysics, Washington University School of Medicine, St. Louis, Missouri 63110, the ^** Centre Regional de Transfusion Sanguine de Lille 19, 21, rue Camille-Guèrin, Bôite Postale 2018, Lille, Cedex 59012, France,  De Diveloppement Pri-Clinique, LRB, 59, Rue De Trivise, BP2006, Lille, Cedex 59011, France, ^§§ INSERM U.143, Hôpital Bicêtre, 94275 le Kremlin-Bicêtre Cedex 94276, France, the ^¶¶ Department of Medical Technology, Nagoya University School of Health Sciences, 1-1-20, Daiko-Minami, Higashi-ku, Nagoya 461-8673, Japan, and the  Nagoya National Hospital, Sannomaru-4-1-1, Naka-ku, Nagoya 461-0001 and Aichi Blood Disease Research Foundation, Moriyama-ku, Nagoya 463-0074, Japan

Received for publication, February 8, 2002, and in revised form, April 6, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In vitro platelet glycoprotein Ib (GPIb) binding of the human von Willebrand factor (VWF) increases markedly by exogenous modulators such as ristocetin or botrocetin, and the binding does not occur in normal circulation. GPIb binding sites have been assigned in the VWF A1 domain, which consists of a disulfide loop Cys^1272(509)-Cys^1458(695) where amino acid residues are numbered from the starting methionine as +1. The previous numbering from the N-terminal Ser of the mature processed VWF is indicated in parentheses. In contrast, several gain-of-function mutations have been found in two regions comprised of the disulfide loop and its N- and C-terminal flanking regions. In this study, Cys^1222(459)-Tyr^1271(508), Gln^1238(475)-Tyr^1271(508), Glu^1260(497)-Tyr^1271(508), and Asp^1459(696)-Asp^1472(709) were sequentially deleted of full-length multimeric recombinant VWF. Deletions at either side resulted in normal GPIb binding, indicating that the flanking regions are not GPIb binding sites. However, the addition of a mutation at Arg^1308(545) on each deletion mutant resulted in spontaneous GPIb binding without requiring modulators, suggesting that both regions are important for the inhibition of GPIb binding. Spontaneous binding was completely inhibited by monoclonal antibodies that recognize the GPIb binding sites. Interestingly, mutant proteins with N-terminal but not C-terminal deletions lost binding to monoclonal antibodies B328, B710, and 23C7, which selectively inhibit ristocetin-induced GPI binding. Their epitopes were found at His^1268(505) or Asp^1269(506). The crystallographic structure of the A1 domain suggests that GPIb binding is influenced by the molecular interface between the two regions and that the antibody binding to the interface inhibits binding.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Von Willebrand factor (VWF)¹ is a multimeric glycoprotein that plays an important role in primary hemostasis. VWF circulates in the blood as multimers with heterogeneous molecular sizes that are assembled from subunits of ~250 kDa. The multimer sizes range from dimers of ~500 kDa to >10,000 kDa. Multimeric VWF binds to the  chain of platelet glycoprotein Ib (GPIb) (1-3). Damage to the endothelium allows VWF to bind to subendothelial connective tissues, and this enables VWF to bind platelets at the site of injury. An additional hemostatic function of VWF is the stabilization of coagulation factor VIII, which is deficient in patients with hemophilia A.

The primary translation product consists of 2813 amino acids and includes a signal peptide of 22 residues, a large propeptide of 741 residues, and a mature subunit of 2050 residues. The GPIb binding site on VWF corresponds approximately to the first of three repeated A domains in the VWF subunit (4). Domain A1 extends from Glu^1260(497) to Gly^1479(716) and contains an intrachain disulfide loop that is defined by the disulfide bond Cys^1272(509)-Cys^1458(695) (5, 6). In vitro, VWF aggregates fixed platelets, indicating that signaling-dependent platelet function is not required for this process. Binding is induced in the presence of small bacterial glycopeptide antibiotic ristocetin or snake venom protein botrocetin. Ristocetin apparently binds both to platelets and to VWF (7), whereas botrocetin binds to VWF domain A1 but not to GPIb (8). Binding sites of botrocetin and GPIb to the A1 domain have been studied by scanning mutagenesis (9-11).

Without stimulation to bind GPIb, VWF interacts weakly with platelets. Thus, VWF binding to platelets appears to require the high affinity binding state that can be influenced by natural or artificial mutations of the VWF A1 domain. Von Willebrand disease (VWD) type 2B is characterized by enhanced platelet aggregation and the consequent consumption of circulating platelets and VWF. Mutations that cause this phenotype are in three distinct regions relative to the Cys^1272(509)-Cys^1458(695) disulfide bond within domain A1: in the N- and C-terminal flanking regions (e.g. P503L, H505D, L697V, A698V) and in a cluster of mutations between Met^1303(540) and Arg^1341(578) in the disulfide loop of Cys^1272(509)-Cys^1458(695) (type 2B region, Human Gene Mutation data base, uwcm.ac.uk/uwcm/mg/hgmd0.html). Alanine-scanning mutagenesis had also indicated that several mutations in flanking regions resulted in the gain-of-function phenotype for GPIb binding that is similar to type 2B (9).

To further characterize the role of these regions in regulation of GPIb binding, we have produced deletion mutants of flanking regions with or without an additional mutation in the type 2B region. These mutants were tested for ristocetin- and botrocetin-induced GPIb binding and for spontaneous GPIb binding. We have also identified the epitopes of several monoclonal antibodies for the VWF A1 domain that inhibits VWF activation.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Constructs-- The expression plasmid pSVvWF1.1 (9) contains a unique NgoMI and KpnI site in the full-length coding sequence of human VWF. The NgoMI/KpnI fragment encodes amino acid residues 442-821 (9) and was subcloned into pGEM-4Z/vWFa-2. Methods for deletion mutagenesis are based on the PCR-based technique by Tessier and Thomas (12). To add the substitution at Arg^1308(545), a BbsI-Bpu1102I fragment containing the mutation from a mutant R545A (9) was subcloned into each deletion mutant. DNA sequence analysis was performed by dideoxy sequencing by Sequenase 2.0 (USB) for the amplified region by PCR. Each fragment was further digested by NgoMI and KpnI and then inserted into pSVvWF1.1.

Expression of Recombinant VWF-- Human 293T cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum. Cells were transfected by the lipofection method using the TransFast transfection kit (Promega) according to the manufacturer's instructions. Twenty-four hours after transfection, cells were washed with phosphate-buffered saline (PBS) and then incubated with a serum-free medium (Optimem-1, Invitrogen). After 48 h, recombinant VWF (rVWF) secreted in the medium was concentrated using Centriprep-30 and Centricon-100 devices (Millipore). The VWF antigen was measured by ELISA using polyclonal anti-human VWF antibody 082 (DAKO) and peroxidase-conjugated rabbit anti-human VWF antibody P226 (DAKO) (13). Both antibodies recognize the dimer of VWF as well as fully multimeric VWF (13). VWF multimer analysis was performed as previously described (14) with minor modifications.

Binding of Monoclonal Antibodies to rVWF-- Monoclonal antibody (mAb) 6D1 against human platelet GPIb (15) was provided by Dr. Barry Coller (Mt. Sinai Medical Center, New York). The mAb NMC-4 was provided by Dr. Midori Shima, Nara Medical University, Nara, Japan and Avw3 was from Dr. Philip Kroner, The Blood Center of Southeastern Wisconsin, Milwaukee, WI. All antibodies were purified from ascites fluid by standard chromatographic methods using protein A beads as described earlier (11). The binding of mAbs to rVWF was studied by ELISA as previously described (11). Diluted antibodies (7.5 µg/ml) in bicarbonate buffer (pH 9.6) were coated onto microtiter plates with U-shape bottoms (Costar, Cambridge, MA) at 4 °C overnight. The wells were washed five times with PBS containing 0.1% Tween 20 (PBST) and then incubated for 120 min at room temperature with 15 µl of various concentrations (62.5-500 ng/ml) of wild type or mutant rVWF diluted in PBS containing 3% BSA. The wells were washed again and incubated with P226 for 90 min at room temperature followed by color development of the absorbance at 490 nm. Negative control assays were performed using concentrated media from mock transfection. Binding of rVWF mutants was determined at a fixed concentration of rVWF (500 ng/ml) and normalized to the value obtained for wild type rVWF assayed at the same time.

Platelet Binding Assays of rVWF-- Modulator-induced binding assays of rVWF were performed as described (9) with minor modifications. For ristocetin-induced binding assays, rVWF (550 ng/ml) was mixed with 1.02 × 10⁸/ml human lyophilized platelets (Helena, Beaumont, TX) and various concentrations (0-2.0 mg/ml) ristocetin (Helena), in TBS (Tris-buffered saline: 50 mM Tris-HCl, pH 7.5, 150 mM NaCl). BSA was added at the final 3% concentration. The reaction mixtures were incubated for 30 min at room temperature and centrifuged for 10 min at 10,000 × g; the VWF antigen in the supernatant was then measured by ELISA. Reactions without platelets were tested simultaneously to verify the absence of nonspecific VWF flocculation and sedimentation in the presence of ristocetin. The values were normalized to those obtained for control assays of wild type rVWF in the presence of 1.5 and 2.0 mg/ml ristocetin.

For botrocetin-induced binding assays, each reaction mixture containing 667 ng/ml rVWF, 2.0 × 10⁸/ml human lyophilized platelets, and various concentrations (0-0.5 µg/ml) of botrocetin was incubated at room temperature. The BSA concentration was 0.2%. Highly purified two-chain botrocetin was provided by Dr. Yoshihiro Fujimura (Nara Medical University, Nara, Japan). After 30 min, the mixture was centrifuged, and VWF in the supernatant was measured by ELISA. Control assays were performed in the absence of platelets. For both assays, the unbound VWF was determined as the percentage of the values obtained with no modulators, and the bound VWF was calculated by subtracting from 100%. The percentage of bound VWF was normalized to the value obtained for wild type rVWF assayed at the same time and compared in the presence of 0.0631 and 0.25 µg/ml botrocetin.

For spontaneous platelet binding assays of rVWF, the 30-µl reaction mixture contained 1.0 × 10⁸/ml human lyophilized platelets and varying concentrations of rVWF (0-1.0 µg/ml) in TBS. BSA concentration was 0.2%. After 30 min at room temperature, the mixtures were centrifuged, and then the VWF concentration in the supernatant was measured by ELISA. The unbound VWF was determined as the percentage of the values obtained without platelets; bound VWF was calculated by subtracting from 100%. Finally, values obtained with 500 ng/ml rVWF were compared.

Biotinylated Botrocetin Binding Assay-- The assay was modified from a method described previously (16) using anti-VWF mAb 33E12, which binds to the C-terminal region of the VWF subunit. Antibody 33E12 has no effect on VWF binding to platelets in the presence of either ristocetin or botrocetin (17). In brief, purified botrocetin (200 µg) was dissolved in 1 ml of HEPES-buffered saline (HBS) and dialyzed against 0.1 M NaHCO₃ at 4 °C overnight. An equal amount (w/w) of Sulfo-NHS-LC-Biotin (Pierce) dissolved in dimethyl sulfoxide was added to dialyzed botrocetin and incubated at room temperature for 4 h in the dark. Biotinylated botrocetin was dialyzed against HBS at 4 °C overnight and then stored at 80 °C until use. Microtiter plates were coated with 100 µl of 33E12 (10 µg/ml) for 16 h at 4 °C followed by washing with HBS that contained 0.1% Nonidet P-40 and blocking with 4% BSA. 120 µl of each rVWF mutant (5 µg/ml) and 0.167 µg/ml biotinylated botrocetin were incubated for 2 h at room temperature, and subsamples were added to the washed wells and incubated for 30 min. Following washing, 50 µl of peroxidase-conjugated streptavidin (VECTOR, Burlingame, CA) was added and incubated for 20 min. The wells were washed again, and color was developed and measured at A_490 nm. Nonspecific binding was zero for several rVWF concentrations (data not shown). Therefore, total absorbance was used as the specific binding. To compare the binding data, absorbance values were normalized to wild type rVWF assayed at the same time.

Effect of Monoclonal Antibodies for the Spontaneous Platelet Binding of rVWF-- The varying concentrations (0.25-10 µg/ml) of anti-VWF mAbs NMC4, Avw3, B328, B710, and 23C7 with control monoclonal anti-human IgG antibody (DAKO) were mixed with 0.25 µg/ml mutant rVWF displaying spontaneous platelet binding. After incubation for 30 min at room temperature, 2 × 10⁸/ml lyophilized platelets were added and incubated for 30 min. After centrifugation, the unbound rVWF in the supernatant was measured by ELISA. Bound VWF was calculated by subtracting the percent of unbound VWF from 100%, and the percent of inhibition was expressed by normalizing to the values obtained without antibodies.

Crystallographic Structure Representations-- The crystal structure was built based on coordinates by Emsley et al. (18). Stereo ribbon diagrams of the VWF A1 domain were prepared with the program INSIGHT II (Molecular Simulations, San Diego, CA). Helices are numbered according to Celikel et al. (19).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Design and Expression of VWF A1 Domain Mutants-- Within domain A1 of human VWF, at least four natural mutations, P503L, H505D, L697V, and A698V have been found in patients with VWD type 2B, which is characterized by mutant VWF with increased affinity for platelet GPIb (von Willebrand factor data base). A previous study using charged-to-alanine-scanning mutagenesis indicated that several mutations at residues between Glu^1260(497) and Asp^1269(506) and between Arg^1450(687) and Glu^1452(689) resulted in enhanced GPIb binding (9), proposing possible involvement of these residues for regulation of GPIb binding. Here flanking regions were targeted for deletion mutagenesis between amino acids Cys^1222(459) and Tyr^1271(508) (N-terminal) and between Asp^1459(696) and Thr^1472(709) (C-terminal).

Four segments, Cys^1222(459)-Tyr^1271(508), Gln^1238(475)-Tyr^1271(508), Glu^1260(497)-Tyr^1271(508), and Asp^1459(696)-Asp^1472(709) were deleted, and the full-length multimeric rVWF was expressed in 293T cells. Fig. 1 illustrates schematic diagrams of each mutant. In this strategy, amino acids flanking the deleted segments are joined. Mutants were named according to the deletion range of the residue numbers. Amino acid positions of mutant rVWF were numbered from the N-terminal Ser of the mature processed VWF subunit as +1. For example, a mutant deleted with residues between Cys^1222(459) and Tyr^1271(508) was named 459-508. When both Cys^1222(459)-Tyr^1271(508) and Asp^1459(696)-Asp^1472(709) were deleted, it was named (459)+(709). To evaluate the binding of anti-VWF mAbs, a mutant 509-695 was produced that had complete deletion of the Cys^1272(509)-Cys^1458(695) disulfide loop. Deleted residues of 509-695 were replaced by a glycine-alanine dipeptide (Fig. 1).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Recombinant VWF mutants. Deleted or mutated amino acid segments in rVWF mutants are schematically presented. Segments indicated by thick straight lines are the flanking regions of the human VWF A1 domain, which is composed of residues Cys1222(459)-Tyr1271(508) and Asp1459(696)-Asp1472(709). In each deletion mutant of the flanking regions, removed segments are depicted below each diagram. The mutants were named according to the residue numbers of deleted parts, and the name is indicated at the upper left corner of each diagram. Additional R545A mutations are indicated by R before the deleted range. For example, when Cys1238(475)-Tyr1271(508) was deleted, the mutant was named 475-508. When Arg1308(545) of 475-508 was converted to Ala, it was named R475-508. Mutant 509-695 lacked amino acid residues from Cys1272(509) to Cys1458(695), and the Cys1272(509)-Cys1458(695) disulfide loop was replaced by a glycine-alanine dipeptide. The deleted disulfide loop is indicated by a dotted circle. Mutant (459)+(709) lacked both Cys1222(459)-Tyr1271(508) and Asp1459(696)-Asp1472(709). Plasmids for R459-508 and R459+709 were constructed, but proteins were not expressed.

Arg^1308(545) is mutated in patients with VWD type 2B (von Willebrand factor data base), and alanine-scanning mutagenesis had indicated that a mutation at Arg^1308(545) showed the increased GPIb binding (9). In the current study, Arg^1308(545) of several deletion mutants was also converted to Ala, and such mutants were named according to the original name with an R added (Fig. 1). For example, when Arg^1308(545) of 475-508 was changed, the mutant was named R475-508. We constructed 11 expression plasmids, and the DNA sequence analysis indicated that the coding sequence for human VWF was normal except for introduced mutations. Despite several transfection experiments, conditioned media of the cells transfected with plasmids R459-508 and R (459)+(709) produced no VWF, whereas nine other plasmids yielded detectable VWF antigen by ELISA (data not shown). The remaining nine mutants listed in Fig. 1 were thus subjected to further studies.

Binding of the Panel of Conformation-dependent Monoclonal Antibodies for VWF-- In the previous study of alanine-scanning mutagenesis (11), folding of mutants was evaluated with a panel of mAbs for the VWF A1 domain including NMC4 (20, 21), Avw3 (22), CLB-RAG34, CLB-RAG35 (23), and 211A6 (24). Recognition by these antibodies was impaired by reduction or denaturation of VWF. Here we have also studied binding of the mutants to the mAb panel by using the specific ELISA. The absorbance obtained for each mutant rVWF was expressed as a percentage of the value obtained for wild type rVWF (Fig. 2). Mutant 509-695 abolished binding to five of six mAbs, suggesting that binding of these antibodies was dependent on the Cys^1272(509)-Cys^1458(695) disulfide loop (Fig. 2). The mAb 211A6 had 38% binding of wild type rVWF, suggesting that 211A6 binding is not fully dependent on VWF residues between 1222(459) and 1472(709). In contrast, all other mutants displayed more than 60% binding of wild type rVWF to the mAbs, excluding the significant misfolding of the A1 domain. Concentrated conditioned medium from mock-transfected cells gave no signal (data not shown).

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Binding of rVWF to the panel of conformation-dependent monoclonal antibodies against human VWF A1 domain. Each monoclonal antibody (indicated at the top of each histogram) was coated onto a plastic microtiter well. The binding was compared for varying VWF concentrations, and the mutant binding was determined at a fixed concentration of rVWF (500 ng/ml) and normalized to the value obtained from wild type rVWF assayed at the same time, as described under "Experimental Procedures." The mutant proteins are indicated at the left. Each column represents the mean and range of values obtained from duplicate assays. In each experiment, wild type VWF was assayed simultaneously, and it gave similar results.

VWF Multimer Analysis-- Fig. 3 shows the multimer distribution of nine recombinant mutants. Although binding of the mAb panel was normal, high molecular size multimers were decreased in 459-508 and 475-508. Only 8-9 multimer bands were visible, whereas wild type or other mutant rVWF had at least 12 multimers (Fig. 3).

View larger version (64K): [in this window] [in a new window]   Fig. 3.   VWF multimer analysis. The multimer analysis was performed in four different experiments, and the distributions of multimers are visualized along with wild type rVWF studied at the same time. Samples of rVWF (0.5 µg) were analyzed by SDS-1.5% agarose gel electrophoresis as described under "Experimental Procedures." The name of each mutant is indicated above the corresponding lane.

Ristocetin-induced Binding to Platelets-- Binding to GPIb was assessed for the deletion mutants by quantitating rVWF that bound to fixed platelets in the presence of ristocetin or botrocetin. In either case, binding to platelets was blocked completely by mAb 6D1 to platelet GPIb (15) as reported previously (25).

Fig. 4 summarizes the results of ristocetin-induced GPIb binding for eight deletion mutants with or without R545A substitution. Fig. 4A indicates the dose-response binding of rVWF for increasing concentrations of ristocetin. Ristocetin causes precipitation and sedimentation of the rVWF in conditioned medium during centrifugation (7), and this nonspecific agglutination was prevented by addition of 3-4% BSA (9). When the reaction mixtures contained no platelets, there was no nonspecific flocculation of VWF (Fig. 4A). Under these conditions, binding of wild type rVWF to lyophilized platelets was dependent on the ristocetin concentration. Mutant binding was determined at several ristocetin concentrations, but abnormal phenotypes were distinguished clearly at the higher ristocetin concentrations (Fig. 4A). Values were normalized to those obtained for control assays of wild type rVWF in the presence of 1.5 and 2.0 mg/ml ristocetin (Fig. 4B).

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Ristocetin-induced binding of rVWF to platelets. The ristocetin-induced GPIb binding of eight deletion mutants is summarized. For comparison, the results for R545A are also included. The binding was assessed by quantitating rVWF that bound to fixed platelets. Each mutant was incubated with human lyophilized platelets and varying concentrations of ristocetin as described under "Experimental Procedures." The mixtures were centrifuged, and VWF antigen present in the supernatant was measured by ELISA. A, binding of rVWF to platelets with increasing concentrations of ristocetin. Platelet binding is expressed as the percentage of unbound VWF antigen compared with the values obtained with no ristocetin and no platelets. For each mutant (), binding is shown with wild type rVWF () assayed at the same time. The names of mutants are indicated at the top of the panel. Ristocetin causes precipitation and sedimentation of the rVWF in conditioned medium during centrifugation (7); this nonspecific agglutination was prevented by the addition of 3-4% BSA (9). Results obtained with the no-platelets condition are shown in the same panel for wild type () and each mutant () rVWF. B, histogram of ristocetin-induced binding to platelets. The value for each mutant rVWF including R545A is expressed relative to that for wild type rVWF performed at the same time. Binding to platelets was determined in the presence of 2 mg/ml ristocetin (filled column) and 1.5 mg/ml ristocetin (hatched column). For panels A and B, each data point or column represents the mean ± range of values obtained in duplicate assays except that data for the no-platelet control were obtained in one assay.

Mutants 459-508, 497-508, 696-709, (459)+(709), and R475-508 essentially showed normal ristocetin-induced binding (Fig. 4). In contrast, binding of R497-508 and R696-709 was greatly enhanced (Fig. 4, B and C). A previously studied mutant, R545A, was assayed and showed slightly increased ristocetin-induced GPIb binding. Indeed, maximum binding of wild type rVWF varied (25~50%) among several assays, resulting in a variety of relative binding of R545A (Fig. 4A) (9). In either case, the degree of binding enhancement was lower than R497-508 and R696-709, suggesting that deletions at residues 1260(497)-1271(508) or 1459(696)-1472(709) resulted in enhancement of binding of R545A.

Fig. 4 indicates that 475-508 showed 28% binding of wild type rVWF in the presence of 2.0 mg/ml ristocetin. However, 459-508 and 497-508 showed normal binding, suggesting that residues 1238(475)-1271(508) are not necessarily required for GPIb binding. The relative decrease in higher molecular size multimers may underlie the decreased GPIb binding of 475-508 (Fig. 3). Lower molecular size forms appear to have decreased GPIb affinity, and those forms are prominent in 475-508. No other mutants showed reduced ristocetin-induced GPIb binding, suggesting that the flanking regions are not the GPIb binding sites. Mutant 509-695, which lacks the Cys^1272(509)-Cys^1458(695) disulfide loop, completely lost GPIb binding (data not shown).

Spontaneous Binding of rVWF to Platelets-- Varying concentrations of rVWF were tested for binding to fixed human platelets in the absence of ristocetin and botrocetin (Fig. 5, A-C). Four deletion mutants without the R545A substitution, 459-508, 475-508, 497-508, and 696-709, did not show spontaneous binding (Fig. 5A). When Arg^1308(545) was changed to Ala, spontaneous binding was induced. Fig. 5B indicates that two mutants, R497-508 and R696-709, showed marked spontaneous binding to platelets. At 0.5 µg/ml rVWF concentration, ~75-86% R497-508 and R696-709 bound platelets without adding modulators (Fig. 5, B and C). In contrast, R545A showed spontaneous binding by 29%, which is lower than those of R497-508 and R696-709. Mutant (459)+(709) also showed spontaneous binding, and 36.5% of added rVWF bound platelets. However, as shown in Fig. 4, A and B, R545A and (459)+(709) did not show spontaneous binding under the conditions of the ristocetin-induced binding assay, apparently because of different BSA concentrations used in the two assays.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Spontaneous binding of rVWF to platelets. The binding of eight deletion mutants is summarized. For comparison, the results for R545A are also included. A and B, each reaction mixture contained human lyophilized platelets and varying concentrations of rVWF as described under "Experimental Procedures." Binding is expressed as the percentage of unbound VWF antigen compared with the values obtained with no platelets. The mutant proteins are: panel A, 459-508 (), 475-508 (), 497-508 (), and 696-709 (); panel B, (459)+(709) (), R475-508 (), R497-508 (), R696-709 (), and R545A (). Wild type rVWF is not shown because the spontaneous binding was undetectable. C, histogram of spontaneous binding of rVWF to platelets. For each mutant, the percent of VWF binding was compared by subtracting values obtained at 0.5 µg/ml VWF from 100%. Each data point or column represents the mean ± range of values obtained in duplicate assays.

Binding of mAbs B328, B710, and 23C7 to rVWF-- mAb B328, also known as B322 (26), selectively inhibits ristocetin-induced GPIb binding but not botrocetin-induced GPIb binding. mAb B710 (27) and 23C7² have the same effect as B328. To specify the role of the flanking regions in ristocetin-induced GPIb binding, the reactivity of these mAbs was tested for our deletion mutants. Fig. 2 indicates that B328, B710, and 23C7 had reduced binding to 509-696, 459-508, 475-508, 497-508, (459)+(709), R475-508, and R497-508. In contrast, these antibodies bound normally to 696-709 and R696-709 (Fig. 6). These findings indicate that the epitopes of the antibodies are localized within two regions: the Cys^1272(509)-Cys^1458(695) disulfide loop and the residues Glu^1260(497)-Tyr^1271(508) in the N-terminal flanking region.

View larger version (49K): [in this window] [in a new window]   Fig. 6.   Epitope mapping of mAbs B328, B710, and 23C7. Binding of the mAbs was determined for the deletion mutants and for the mutant panel of charged-to-alanine-scanning mutagenesis of VWF A1 domain. Alanine mutants were produced as previously described (9, 11) and are indicated at the left. The names of clustered mutants are abbreviated with the range of residue numbers and the number of alanine substitutions. For example, (505-506)2A contains alanine substitutions at the two charged residues His1268(505) and Asp1269(506). Binding assays were performed as described in the legend to Fig. 2. Single or clustered alanine mutations at residues Asp1277(514), Asp1283(520), Arg1315(552), and Arg1374(611) had shown reduced binding to conformation-dependent mAbs, suggesting significant misfolding of those mutants (11); they are excluded from the histogram. Each column represents the mean and range of values obtained for at least duplicate assays. In each experiment, wild type VWF was assayed simultaneously and gave similar results.

To further characterize the epitopes of the mAbs, we studied binding to a set of charged-to-alanine mutants of the A1 domain (11). In this set, mutations at Asp^1277(514), Asp^1283(520), Arg^1315(552), and Arg^1374(611) had shown reduced binding to all conformation-dependent mAbs, suggesting that these substitutions cause significant misfolding of the A1 domain (11). Binding of these four mutants to B328, B710, and 23C7 was also reduced (data not shown). Fig. 6 indicates that among 54 single or clustered alanine mutants, (505-506)2A specifically reduced binding to B328, B710, and 23C7. Therefore, the epitopes in the N-terminal flanking region are either His^1268(505) or Asp^1269(506). Mutant (539-543)3A reduced binding to 23C7, indicating that binding of 23C7 is also dependent on Asp^1302(539), Glu^1305(542), or Arg^1306(543) (Fig. 6). Therefore, the two regions spanning the disulfide bond between Cys^1272(509) and Cys^1458(695) appear to be close spatially. For B328 and B710, no critical residues were identified within the Cys^1272(509)-Cys^1458(695) disulfide loop. Because these three mAbs selectively inhibit ristocetin-induced binding, the two regions appear to be important for GPIb binding by ristocetin but not by botrocetin.

Effect of mAbs on the Spontaneous Binding of rVWF to Platelets-- We studied the effects of several mAbs for VWF on the spontaneous GPIb binding of two mutants R497-508 or R696-709. NMC4 and AvW3 are known to interfere with both ristocetin- and botrocetin-induced GPIb binding (21, 22, 28). The epitope of NMC4 is included in residues Arg^1395(632) and Arg^1399(636) (19), which are closely located to the GPIb binding site (11). The epitope of Avw3 has not been determined. Fig. 7, A and B indicates that NMC4 and Avw3 completely blocked spontaneous GPIb binding, suggesting that the mutant binding occurred through the GPIb binding site. Control IgG did not interfere with spontaneous binding (data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Effect of monoclonal antibodies on the spontaneous GPIb binding of mutants R497-508 and R696-709. Mutants were incubated with anti-VWF monoclonal antibodies as described under "Experimental Procedures." Lyophilized human platelets were added to the mixture, and unbound VWF in the supernatant was assayed by ELISA. Spontaneously bound VWF was determined by subtracting the percent of unbound VWF from 100%, and the residual binding was determined to values obtained without antibodies. The mutant name is indicated at the top of each panel. NMC4 () or Avw3 () were tested for the spontaneous binding of R497-508 (A) and R696-709 (B). C, B328 (), B710 (), and 23C7 () were evaluated for inhibition of spontaneous GPIb binding of mutant R696-709. For panels A-C, each data point represents the mean ± range of values obtained in duplicate assays.

We also tested the inhibitory effect of B328, B710, and 23C7 on spontaneous platelet GPIb binding. Because these antibodies did not bind R497-508 (Fig. 2), the inhibition was tested using R696-709. Fig. 7C indicates that the antibodies partially inhibited binding, suggesting that the two epitope regions of the mAbs are also involved in the spontaneous binding of R696-709.

Botrocetin-induced Binding of rVWF to Platelets and Binding of Biotinylated Botrocetin to rVWF-- Botrocetin interacts with its specific binding sites that include amino acids Arg^1382(629), Arg^1395(632), Arg^1399(636), and Lys^1430(667), which reside within the Cys^1272(509)-Cys^1458(695) disulfide loop (11). To further characterize the botrocetin-VWF interaction, we performed botrocetin-induced platelet binding assays and biotinylated botrocetin binding assays of our deletion mutants. The optimal concentration of purified botrocetin for the assays will vary among preparations of venom from Bothrops jararaca. First we tested several different concentrations of the purified botrocetin used in this study. Unlike the previous optimal range (0-17.6 µg/ml) (9, 11), we found that 0-0.5 µg/ml purified botrocetin gave an optimal signal for both the botrocetin-induced platelet binding assay and the biotinylated botrocetin binding assay (data not shown). Unlike ristocetin, botrocetin does not cause platelet-independent precipitation of VWF.

Fig. 8, A and B indicates the dose-responsiveness of botrocetin for the platelet GPIb binding of the mutants. Fig. 8C compares relative binding in the presence of 0.25 and 0.0631 µg/ml botrocetin. Fig. 8C also compares the result of direct biotinylated botrocetin binding. At 0.0631 µg/ml botrocetin, the relative binding of mutants R497-508, R696-709, and R545A was enhanced, whereas none of the deletion mutants displayed increased GPIb binding at 0.25 µg/ml botrocetin. The increased binding at lower botrocetin concentrations appeared to be because of spontaneous binding of these mutants. These observations may indicate that stimulation by botrocetin does not enhance GPIb binding of the mutants that display increased ristocetin-induced binding or spontaneous GPIb binding.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Botrocetin-induced binding of rVWF to platelets. The binding of eight deletion mutants is summarized. For comparison, the result for R545A is also included. The binding was determined as described under "Experimental Procedures." Botrocetin concentrations ranged from 0 to 0.5 µg/ml. A and B, binding of rVWF to platelets with varying concentrations of botrocetin. Platelet binding is expressed as the percentage of unbound VWF antigen compared with the values obtained with no botrocetin and no platelets. The symbol of each mutant is described in the legend to Fig. 5. Wild type rVWF is shown by . C, histogram of botrocetin-induced VWF binding and biotinylated botrocetin binding to VWF. Botrocetin-induced binding of each rVWF mutant was compared in the presence of 0.25 µg/ml (filled column) and 0.0631 µg/ml (open column) as described under "Experimental Procedures." Direct binding of biotinylated botrocetin (0.167 µg/ml) to each mutant rVWF (hatched column) was measured as described under "Experimental Procedures." The values for mutants were normalized relative to that determined for wild type rVWF performed at the same time. For each panel, bars represent the mean ± range of values obtained in quadruplicate assays.

In contrast, three mutants, 459-508, 475-508, and R475-508 showed decreased botrocetin-induced GPIb binding in the presence of 0.25 µg/ml botrocetin, whereas they appeared to be normal in the presence of 0.0631 µg/ml botrocetin (Fig. 8C). The same mutants plus one more, (459)+(709), showed decreased response to increasing concentrations of botrocetin (Fig. 8, A and B) and also showed lowered binding of biotinylated botrocetin (Fig. 8C). These data suggest that the lowered botrocetin-induced platelet binding is due to decreased binding to botrocetin. Therefore, at least amino acid sequences Cys^1222(459)-Glu^1260(497) may be involved in the binding of botrocetin to VWF.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The crystallographic structure of human VWF was determined recently for two forms of the fragment of the A1 domain (18, 19). The structure by Emsley et al. (18) is composed of 208 amino acids between Asp^1261(498) and Thr^1468(705) that include a part of the flanking regions targeted in this study. This structure provides a framework for interpreting the effects of mutations on VWF-GPIb binding (Fig. 9).

View larger version (47K): [in this window] [in a new window]   Fig. 9.   Location of the flanking regions and amino acid residues required for VWF-GPIb regulation. Schematic drawing of the VWF A1 domain based on the coordinates of Emsley et al. (18). Positions of the amino acids are indicated for residues involved in the binding of platelet GPIb and botrocetin. These are shown by large sized spheres, including Lys1362(599) (green), Arg1382(629) (red), Arg1395(632) (purple), Arg1399(636) (blue), and Lys1430(667) (orange). The disulfide bridge between 509 and 695 is shown as a thick cyan bar and links the N- and C-terminal flanking sequences. The disulfide loop is the "body" of the A1 domain, and N- and C-terminal flanking sequences of 1261(498)-1271(508) (purple ribbon) and 1459(696)-1468(705) (red ribbon) are the arms of the body. Nine type 2B sites lie along the interface formed between the body and the N-terminal arm and are indicated by small spheres or sticks. These are Pro1266(503) (black sphere), His1268(505) (orange sticks), Arg1308(545) (red sticks), Ile1309(546) (yellow sphere), Ser1310(547) (green sphere), Trp1313(550) (gray sphere), Val1314(551) (cyan sphere), Leu1460(697) (pink sticks), and Ala1461(698) (blue sticks). Point mutations at His1268(505) and Arg1308(545) also showed increased GPIb binding in the alanine-scanning mutagenesis study(9). Along with His1268(505), Asp1269(506) possibly includes the epitopes of mAb B328, B710, and 23C7; it is indicated by cyan sticks. Other epitopes of mAb 23C7, Asp1302(539), Glu1305(542), or Arg1306(543) are indicated by sticks colored by purple, yellow, and green, respectively. Arg1306(543) is also mutated in patients with type 2B VWD. See the "Discussion" for detail.

Increased GPIb Binding of the Deletion Mutants-- Deletion mutants with R545A substitution, R497-508 and R696-709, resulted in increased ristocetin-induced GPIb binding and in marked spontaneous GPIb binding (Figs. 4 and 5). The observed spontaneous binding appears to depend on the GPIb binding sites of the A1 domain because it is completely inhibited by NMC4 or AvW3, which specifically blocks the GPIb binding sites (Fig. 7).

If sequences within the flanking regions interacted with the GPIb binding sites and blocked the access of GPIb, the GPIb binding sites of deletion mutants might be exposed on the surface and result in a gain-of-function phenotype. This model, however, was not true, because only mutants (459)+(709), R497-508, and R696-709 displayed increased ristocetin-induced GPIb binding or spontaneous GPIb binding (Figs. 4 and 5), and the increases in GPIb binding were not dependent on the length of the deleted amino acid sequences. Another model indicates that the flanking regions will participate in inhibition of GPIb binding in terms of protein conformational changes and thus deletion mutants will show gain-of-function phenotype. Our previous study (9) and Fig. 5 show that a single alanine mutation at Arg^1308(545) increased GPIb binding. In this context, additional deletions of 1260(497)-1271(508) or 1459(696)-1472(709) might affect the R545A phenotype and cause marked spontaneous binding (Fig. 5). These facts suggest that conformational integrity between Arg^1308(545) and the flanking regions is important for regulating VWF GPIb binding.

By adopting loss-of-function mutants to the crystallographic model of the A1 domain, GPIb binding sites had been localized in helices 3 and 4, among which Lys^1362(599), Arg^1382(629), and Arg^1395(632) play central roles for GPIb binding (11). In Fig. 9, these sites are on one side of the A1 domain. The Cys^1272(509)-Cys^1458(695) disulfide bond is located on the opposite side, and it links the N- and C-terminal flanking sequences. Here the disulfide loop is defined as the "body" of the A1 domain, and the N- and C-terminal flanking sequences of 1261(498)-1271(508) and 1459(696)-1468(705) are defined as the arms of the A1 body. The interface is formed between the lower surface of the body and the N-terminal arm. Several amino acids mutated in VWD type 2B, including Pro^1266(503), His^1268(505), Arg^1308(545), Ile^1309(546), Ser^1310(547), Trp^1313(550), Val^1314(551), Leu^1460(697), and Ala^1461(698) are located at the interface (Fig. 9). In particular, the side chain of Arg^1308(545) is buried downward toward the pocket surrounded by the hydrophobic body base and the N- and C-terminal arms. Emsley et al. (18) reported that Arg^1308(545) forms a hydrogen bond to Cys^1272(509) and may participate in stabilizing VWF structure. Thus disruption of the salt bridge may cause destabilization of the interface, implying the gain-of-function phenotype of mutations at Arg^1308(545).

The deletions in the flanking regions enhanced the gain-of-function phenotype of R545A, leading to the marked deregulation of VWF for GPIb binding (Figs. 4 and 5). In the case of R497-508, amino acid sequences of the N-terminal arm between 1260(497) and 1271(508) are removed and the neighboring N-terminal sequences shifted in the position, leading to destabilization of the folded conformation of the N-terminal arm. Therefore the equilibrium could be shifted toward the extended conformation, explaining the possible mechanism for the strong gain-of-function of R497-508 (Fig. 5). In contrast to R497-508, R475-508 did not display the gain-of-function phenotype (Figs. 4 and 5). In R475-508, a different amino acid segment is shifted/inserted and might not affect the conformation of the interface.

In Fig. 9, the C-terminal arm extends downwards, whereas the proximal 3-4 amino acids from Cys^1458(695) are located near the interface. The proximal amino acid sequences are deleted in R696-709, and its gain-of-function phenotype may be related both to alterations of Arg^1308(545) and the proximal amino acids. In fact, natural or artificial mutations at Leu^1460(697) and Ala^1461(698) had resulted in the gain-of-function phenotype (Human Gene Mutation data base) (9), indicating their effects at the structure of the interface. These observations strongly suggest that the increased GPIb binding of R497-508 and R696-709 appears to be due to disruption of the conformational integrity of the VWF body and N- or C-terminal arms.

Although there have been no good assays for direct interaction between ristocetin and VWF, several studies have suggested the importance of a unique proline-rich sequence in the distal portion of the C-terminal arm (29, 30). Azuma et al. (29) used a dimeric recombinant VWF fragment harboring mutations at three proline residues between 1465(702) and 1467(704), and those mutants displayed no ristocetin-induced GPIb binding. De Luca et al. (30) found a mAb that specifically interacts with peptides containing Glu^1463(700)-Asp^1472(709) and inhibits only ristocetin-induced binding. Our assay, however, indicated that recombinant multimeric VWF deleted with the sequence between 1459(696) and 1472(709) resulted in normal or increased ristocetin-induced binding (Fig. 4). Moreover, mutants with selective loss of ristocetin-induced binding have been found in the disulfide loop between Cys^1272(509)-Cys^1459(695) but not in the N-terminal flanking region (9, 11) (Human Gene Mutation data base). Therefore, it is possible that the distal portion of 1459(696)-1472(709) may participate in the interaction with ristocetin but that this interaction is not specific for the amino acid sequences. Such a hypothesis should await further new assays that can quantitate direct ristocetin-VWF interaction.

Bovine VWF has been described as spontaneously binding to platelet GPIb in the absence of any modulators (31). Such a phenotype, however, has not been found in human VWF. Mutants R497-508 and R696-709 will be a good model for studying the molecular basis of VWF-GPIb interaction, and development of such recombinant proteins may help establish a simple assay system without need for exogenous modulators.

Epitopes of B328, B710, and 23C7 and Their Inhibition of GPIb Binding-- The mAb binding study of the deletion mutants indicated that B328, B710, and 23C7 had the epitopes both in the N-terminal flanking region and in the Cys^1272(509)-Cys^1458(695) disulfide loop (Fig. 6). The binding study using the mutant panel of the alanine-scanning mutagenesis (9, 11) clarified the epitope in His^1268(505)-Asp^1269(506) (Fig. 6). Fig. 6 also indicates that the epitope of 23C7 is restricted within Asp^1302(539), Glu^1305(542), or Arg^1306(543). In Fig. 9, His^1268(505) and Asp^1269(506) are located at the corner of the loop of the N-terminal arm, and the side chain of His^1268(505) is projected toward the end of helix 1 of the VWF-A1 body. His^1268(505) forms a salt bridge to Glu^1305(542) (18), possibly providing a recognition site for 23C7. Other epitopes of B328 and B710 are not determined in the Cys^1272(509)-Cys^1458(695) disulfide loop, but amino acids other than charged residues will include their epitopes.

Ristocetin induces VWF activation by different mechanisms from botrocetin. These three mAbs inhibit ristocetin-induced GPIb binding but not botrocetin-induced GPIb binding (26, 27). Although the possibility remains that the mAbs may block the nonspecific ristocetin-VWF interaction, the mAbs inhibit the conformational change of VWF by binding to both the body and the N-terminal arm (Fig. 9). It is therefore possible that in the presence of ristocetin, GPIb binding is accomplished via conformational events occurring in the molecular interface.

In mutants R696-709, VWF appears to be already activated. Binding of the mAbs may lock the molecular interface between the N-terminal arm and the lower body surface (Fig. 9) and thus may partially inhibit activation of R696-709 (Fig. 7). On the other hand, structural interpretation of R696-709 implies that its conformational change is also through the proximal sequences in the C-terminal arm (Fig. 9). These observations suggest that both the N- and C-terminal arms are cooperatively involved in the regulation of VWF-GPIb binding.

The above two examples suggest that VWF may induce multiple molecular events to accomplish GPIb binding, and each of them cooperatively regulates. Botrocetin has its own binding site on the VWF-A1 domain and bypasses the VWF activation induced by ristocetin or by several kinds of VWF mutations. A study on the crystal structure of the recombinant type 2B mutant I546V (32) indicated that different types of molecular changes occur in I546V, including water molecule internalization and positional or major conformational changes. These facts suggest that blocking of any one of the activation steps will lead to the loss of ristocetin-induced binding or spontaneous binding. Indeed, our previous study (11) indicated that eight VWF mutations caused a selective decrease in ristocetin-induced binding with normal botrocetin-induced binding, although mutated sites do not appear to have a consistent relationship to the GPIb binding sites. Therefore, such mutations appear to prevent conformational changes induced by ristocetin and may not represent the direct ristocetin interaction sites. Further studies are required, and structural analysis of VWF mutants will explain each activation step for GPIb binding.

Botrocetin Binding of the Deletion Mutants-- Fig. 8 indicated that mutants R497-508, R696-709, and R545A displayed enhanced platelet GPIb binding in the presence of lower concentrations of botrocetin. However, at the higher botrocetin concentrations, the enhancement was not observed, which appears to be because of spontaneous binding (Fig. 8B). In contrast, mutations deleted between residues 1222(459) and 1271(508) but not between 1260(497) and 1271(508) decreased binding to biotinylated botrocetin (Fig. 8C), implicating the involvement of residues 1222(459)-1260(497) for botrocetin binding. The botrocetin binding sites, Arg^1382(629), Arg^1395(632), Arg^1399(636), and Lys^1430(667) are located in helices 4b and 5 in the crystallographic modeling of the A1 domain (Fig. 9) (11). However, the known A1 structures do not cover the amino acid sequences beyond Glu^1260(497), and interactions between these two sites are not interpreted. mAb B724 inhibits the binding of botrocetin to VWF (33), and B724 epitopes have been mapped at Lys^1423(660) and Arg^1426(663), which overlap with the botrocetin binding site on the surface of the A1 domain (11). We studied the binding of B724 to four deletion mutants, 459-508, 475-508, 497-508, and 696-709. Only 475-508 showed 50% binding of wild type, but the binding of the other three mutants was normal (data not shown), indicating that the critical B724 binding site is not found in the flanking regions. The above observations, therefore, do not fully explain the requirement of the flanking regions for binding to botrocetin. The correct information awaits further studies that include the structural analysis of the botrocetin-VWF complex.

	ACKNOWLEDGEMENTS

We thank Chika Wakamatsu, Yukako Yamamoto, and Yuka Nomura for excellent technical assistance. We also thank Drs. Shuji Miura and Tomoki Naoe for helpful discussions.

	FOOTNOTES

^* This work was supported in part by Grants-in-Aid for Scientific Research (12670983) from the Ministry of Education, Science, Sports, and Culture and the Welfide Medicinal Research Foundation (to T. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed: The First Dept. of Internal Medicine, Nagoya University School of Medicine, 65 Tsuruma-cho, Showa-ku, Nagoya 466-8550, Japan. Tel.: 81-52-744-2145; Fax: 81-52-744-2161; E-mail: tmatsu@med.nagoya-u.ac.jp.

Present address: Merck Research Laboratories, WP17-301, West Point, PA 19486.

Published, JBC Papers in Press, April 9, 2002, DOI 10.1074/jbc.M201327200

² T. Nakayama, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: VWF, von Willebrand factor; rVWF, recombinant VWF; VWD, VW disease; HBS, HEPES-buffered saline; mAb, monoclonal antibody; PBS, phosphate-buffered saline; GPIb, platelet glycoprotein Ib; BSA, bovine serum albumin; ELISA, enzyme-linked immunosorbent assay.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES