Ligand binding regions in the receptor for urokinase-type plasminogen activator.

The interaction between urokinase plasminogen activator (uPA) and its cellular receptor (uPAR) is a key event in cell surface-associated plasminogen activation, relevant for cell migration and invasion. In order to define receptor recognition sites for uPA, we have expressed uPAR fragments as fusion products with the minor coat protein on the surface of M13 bacteriophages. Sequence analysis of cDNA fragments encoding uPA-binding peptides indicated the existence of a composite uPA-binding structure including all three uPAR domains. This finding was confirmed by experiments using an overlapping 15-mer peptide array covering the entire uPAR molecule. Four regions within the uPAR sequence were found to directly bind to uPA: two distinct regions containing amino acids 13--20 and amino acids 74--84 of the uPAR domain I, and regions in the putative loop 3 of the domains II and III. All the uPA-binding fragments from the three domains were shown to have an agonistic effect on uPA binding to immobilized uPAR. Furthermore, uPAR-(154--176) increased uPAR-transfected BAF3-cell adhesion on vitronectin in the presence of uPA, whereas uPAR-(247--276) stimulated the cell adhesion both in the absence or presence of uPA. The latter fragment was also able to augment the binding of vitronectin to uPAR in a purified system, thereby mimicking the effect of uPA on this interaction. These results indicate that uPA binding can take place to particular part(s) on several uPAR molecules and that direct uPAR-uPAR contacts may contribute to receptor activation and ligand binding.

nized as performing a central role in these processes (1). Experimental evidence of their involvement has come from studies in a variety of model systems, in which the ability of tumor cells to invade and metastasize could be down-regulated by uPA inhibitors (2), anti-uPA antibodies (3), and uPAR antagonists (4). Several reviews have provided details on the molecular and functional properties of the uPA-uPAR system and its role in extracellular matrix degradation (5)(6)(7). Recent investigations have indicated that uPAR also takes part in other protein interactions, relevant not only to proteolysis but also to cell adhesion and signal transduction, where uPAR serves as a pleiotropic interactive cell surface molecule (8 -10).
Human uPAR is a glycolipid-anchored membrane glycoprotein encoded as a 335-residue non-processed polypeptide (11). The entire sequence of the fully processed and completely extracellular uPAR (amino acids 1-283) is divided into three homologous domains primarily defined by a conserved pattern of cysteine residues and linked to a glycosylphosphatidylinositol moiety (12)(13)(14)(15). Although the N-terminal domain I of uPAR has the primary role in uPA binding (16), the integrity of the multidomain structure of uPAR is required for the maintenance of high affinity binding (17,18). The specific, high affinity interaction between uPA and uPAR is governed by the growth factor-like region of uPA, which also contains a kringle module and a serine protease domain (19). Besides uPA, uPAR is also able to interact with Vn, and uPA promotes uPAR-dependent cell adhesion on Vn. Multimeric Vn, rather than monomers, serves as predominant high affinity ligand for uPAR. In addition, uPAR has been reported to interact with several other soluble and membrane proteins, such as high molecular weight kininogen and integrins (10).
In an approach to uncover the uPA recognition regions on the uPAR molecule and to identify epitopes responsible for promotion of uPAR-dependent cell adhesion on Vn, we have employed two techniques in the current investigation, i.e. phage display and peptide array. The phage display technique was first reported in 1985 by Smith and co-workers (20,21) and relied on the ability to display a peptide of interest on the surface of a bacteriophage capsid. Goodson et al. (22) reported the identification and characterization of peptide antagonists with nanomolar affinity for the human uPAR by using a 15-mer phage display library. Utilizing the second method (23)(24)(25), the uPAR peptides were directly synthesized as spots on the cellulose membrane and used in solid phase binding tests. Together, four uPA-binding regions with distinct sequences on the uPAR molecule were identified and partially characterized, two of which could promote uPAR-mediated cell adhesion on Vn reminiscent of uPA itself. These results indicate that ligation with uPA can induce conformational changes to uPAR allowing direct uPAR-uPAR contacts and, thereby, may influence its functional activity in cell adhesion.

EXPERIMENTAL PROCEDURES
Materials-Vn was purified from human plasma and converted to the multimeric form as described previously (26,27). The plasmid pBSW87 containing the full-length uPAR cDNA was obtained from Dr. N. Behrendt and the monoclonal antibodies (mAbs) R3 and R4 to uPAR (28) were a gift from Dr. G. Høyer-Hansen (both from the Finsen Laboratory, Copenhagen, Denmark). The phagemid vector pComb3B (29) was obtained from Dr. H. Pannekoek (University of Amsterdam, Amsterdam, The Netherlands). The expression vector pGEX-6P-1 and the protease-deficient expression strain Escherichia coli BL 21 were purchased from Amersham Pharmacia Biotech (Freiburg, Germany). The VCSM13 interference resistance helper phage and the electroporation competent E. coli XL-1 Blue MRFЈ were purchased from Stratagene (Amsterdam, The Netherlands). The plasmid pCDNA3 and BstXI linker were from Invitrogen (Groningen, The Netherlands). Human high molecular weight uPA, antibody 3936 to uPAR, and polyclonal rabbit antibodies against uPAR were from American Diagnostica (Greenwich, CT). Recombinant human soluble uPAR (suPAR) from insect cells was a gift from Dr. D. Cines (University of Pennsylvania, Philadelphia, PA). Recombinant bacterial uPA (Saruplase) was from Grü nenthal (Stolberg, Germany). All other chemicals and reagents were of analytical grade.
Production of uPA-Fc Fusion Protein-In previous studies, the growth factor domain of uPA was fused to the constant region of human IgG and this construct was used in in vivo models as a uPAR antagonist (30). In order to obtain sufficient amounts of chromogenically detectable uPA, we have produced the uPA-Fc fusion protein for the present investigation. The plasmid pSVscuPA-Gly 158 (31) containing the cDNA encoding full-length single chain uPA mutant (replacement of Lys 158 by Gly) was from Dr. R. Lijnen (University of Leuven, Leuven, Belgium). The cDNA of the constant region of human IgG ␥-heavy chain (CH 2 CH 3 ) was obtained from Genentech (San Francisco, CA). The construct pCDNA3-Gly 158 scuPA(N)-CH 2 CH 3 (C) was transfected into Chinese hamster ovary cells, and stable clones were chosen for uPA-Fc production. Purification of uPA-Fc was performed with Immuno-Pure Fab kit from Pierce, and the purity was controlled by SDS-polyacrylamide gel electrophoresis and amino acid sequencing.
Construction and Characterization of a uPAR Random Epitope Phage Library-To generate random fragments encoding epitopes of the uPAR molecule, a 1250-bp DNA fragment containing the full-length uPAR cDNA was obtained from plasmid pBSW87 and partially digested with DNase to produce random fragments of 50 -100 bp. To facilitate the cloning and expression of random DNA fragments, we chose phagemid vector pComb3B, which contained two non-self-complementary BstXI sites separated by a 350-bp replaceable segment that allowed the cloning of DNA fragments using BstXI adaptors. Once cloned in the correct orientation and translation reading frame, fragments were expressed as fusion proteins composed of a N-terminal pelB leader peptide, a fusion peptide F N , the epitope peptide, a second fusion peptide F C , a glycine-rich flexible GGGGS peptide linker, and the C-terminal half of the bacteriophage M13 gene III-encoded minor coat protein pIII. Upon superinfection with a helper phage VSCM13, the epitope peptides fused to a part of pIII were displayed on the surface of newly produced phagemid particles. Eighteen randomly picked phagemid particles from this library were subjected to plasmid extraction and subsequent digestion with XhoI. Released uPAR cDNA fragment inserts were visualized after agarose gel electrophoresis. The inserts were also sequenced as described below.
Affinity Selection of uPA-binding Phagemid Particles-Dynalbeads M-280 (tosyl-activated) (Dynal, Hamburg, Germany) have been designed as a solid phase for biomagnetic separation, which have been applied among others also for screening of phage display libraries (32,33). High molecular weight human uPA was coated onto tosyl-activated Dynalbeads M-280 according to the manufacturer's instructions, and the uPA-coated beads (ϳ5 g of uPA/10 8 beads in 1 ml) were stored in phosphate-buffered saline (PBS) containing 0.1% (w/v) bovine serum albumin (BSA) and 0.02% sodium azide at 4°C. After washing three times in PBS, 100 l of uPA-coated magnetic beads were blocked with 1.5 ml of PBS containing 3% BSA and 0.05% Tween 20 (PBST-3% BSA) for 1.5 h at room temperature. The beads were then resuspended in 50 l of PBST-3% BSA and mixed with 50 l of the non-diluted or in PBST-3% BSA 1:10, 1:100 or 1:1000 diluted uPAR phage display library, containing 1.6 ϫ 10 11 , 1.6 ϫ 10 10 , 1.6 ϫ 10 9 , or 1.6 ϫ 10 8 phagemid particles, respectively, at room temperature for 2 h with gentle agitation. After binding, the bead suspension was transferred into a new 1.5-ml Eppendorf tube to avoid plastic-bound phagemid particles. With the help of a magnetic particle concentrator (Dynal MPC-E-1), the beads carrying uPA-binding phages were separated from the supernatant containing uPA-non-binding phages, which, after proper dilution, were used to infect cells of E. coli XL-1 Blue MRFЈ at 37°C for 20 min. The paramagnetic beads were then washed 10 times with PBST-3% BSA and resuspended in 200 l of 0.3 M NaAc buffer, pH 5.2. After a short vortex, the supernatant was removed and immediately neutralized by adding 12 l of 2 M Tris buffer, pH 10.5. Eluted phagemid particles were used to infect 100 l of freshly cultured cells of E. coli XL-1 Blue MRFЈ at 37°C for 20 min. Subsequently, the paramagnetic beads were then incubated in 200 l of 0.1 M glycine-HCl buffer, pH 2.2, and the eluted phagemid particles were used to infect cells of E. coli. Finally, the beads were resuspended in 100 l of LB broth and incubated at 37°C for 20 min with 100 l of E. coli suspension, where the remaining, tightly bound phagemid particles were directly adsorbed by the bacteria. The phage-infected E. coli cells from the above steps were plated onto LA plates containing 50 g/ml ampicillin for overnight culture at 37°C.
Sequence of cDNA Fragment Inserts-Individual clones, harboring recombinant phagemids encoding uPA-binding or uPA-non-binding peptides, were picked and cultivated with gentle shaking at 37°C overnight in a 5-ml LB broth containing 50 g/ml ampicillin. Phagemids containing a uPAR cDNA fragment were purified using Miniprep kits from Qiagen (Hilden, Germany), and the sequences of the inserts were determined (SEQLAB, Göttingen, Germany). The following synthetic oligonucleotides were used as primer: 5Ј-GCC CAG GTG AAA CTG CTC G-3Ј and 5Ј-CAA ACG AAT GGA GAG CCA CC-3Ј.
Screening of uPA-binding Regions Using a uPAR Peptide Array Synthesized on Cellulose Membranes (SPOT-Membrane)-Overlapping 15mer peptides with a three amino acid shift, covering the entire uPAR molecule, were directly synthesized as spots on cellulose membranes (Dr. R. Frank, Molecular Biology Section, GBF-National Research Center for Biotechnology, Braunschweig, Germany). The chemical and technical performance of this type of simultaneous parallel solid phase synthesis has been optimized for the preparation of peptide sequences up to a length of ϳ20 amino acid residues (23)(24)(25). The uPAR peptide array contained a total of 108 uPAR peptides. After extensive washing with PBS, the SPOT-membrane was blocked with 5% skim milk in PBST (with 0.1% sodium azide) overnight at room temperature, followed by incubation with uPA-Fc (4 g/ml) in fresh blocking buffer for 3.5 h at room temperature and washing twice with PBST (10 min each). Under stringent conditions, the membrane was additionally washed (1 min each) with 0.3 M NaAc buffer, pH 5.2, and 0.1 M glycine-HCl buffer, pH 2.2. After PBST washing, the membrane was incubated for another In addition, the peptide array membranes were designed for reuse up to 10 times after stripping, which was performed accordingly as described previously (24,25).
Expression and Purification of uPA-binding Fragments as GST Fusion Proteins-The uPAR fragments selected from the phage library were cloned into expression vector pGEX-6P-1 and expressed as GST fusion proteins. The fusion proteins as well as GST were purified through a GSTrap prepacked column. Cloning, expression, and purification were performed according to the manufacturer's instructions (Amersham Pharmacia Biotech). The fusion proteins were recognized by polyclonal rabbit antibodies against uPAR (1.6 g/ml) as well as antibodies against GST (1 g/ml) in Western blots.
Microtiter Plate Binding Assay-The microtiter plate binding assay was performed as described previously elsewhere (34). Briefly, high molecular weight human uPA at a concentration of 5 g/ml in PBS was coated onto Maxisorb 96-well microtiter plates (Nunc, Roskilde, Denmark) overnight at 4°C. After the wells were blocked with 3% BSA-PBST for 1.5 h at 37°C, various concentrations of GST/GST-uPAR fragments (0 -2 M) and suPAR (90 nM) in blocking buffer were added and incubated for another 1 h. Bound suPAR was detected by the antibody R4 (10 g/ml), which was further detected by 1:1000 diluted peroxidase-conjugated rabbit anti-mouse immunoglobulins (Dako, Hamburg, Germany). Bound peroxidase was quantified using 2,2Јazino-di-(3-ethyl-benzthiazoline-6-sulfonic acid) as substrate (Roche Molecular Biochemicals, Mannheim, Germany) with a microplate reader at 405 nm (Molecular Devices, Munich, Germany). Similarly, suPAR was coated at a concentration of 2.5 g/ml in PBS overnight at 4°C. After BSA blocking, various concentrations of GST/GST-uPAR fragments (0 -2 M) and uPA-Fc (8 g/ml) or Vn (2 g/ml), in the absence or presence of uPA (50 nM), in blocking buffer were added and incubated for another 1 h. Bound uPA-Fc was detected by 1:5000 diluted peroxidase-conjugated antibody against human IgG (Fc), and bound Vn was detected by antibody Vn-7 (125 ng/ml) against Vn (35). Bound peroxidase was quantified as above. Nonspecific binding to BSAcoated wells was used as a blank and was subtracted to calculate specific binding. Experiments were performed in duplicate and repeated at least three times.
Cell Culture and Cell Adhesion Assay-BAF-3 (interleukin-3dependent mouse B-cell line) cells were from the American Type Culture Collection (ATCC, Rockville, MD), and cultured in RPMI 1640 medium containing 10% fetal calf serum and 2 ng/ml interleukin-3. BAF-3 cells were transfected by electroporation with uPAR cDNA in the sense and antisense orientation using the expression vector pCDNA3. Cells were selected in the presence of G418 (1.2 mg/ml) (Calbiochem, San Diego, CA) and determined to express uPAR by fluorescence-activated cell sorter analysis, Northern blotting, and uPAR-enzyme linked immunosorbent assay. Cell adhesion of uPAR sense-transfected BAF-3 cells to Vn-coated plates was performed according to a previously described protocol (34,36). Briefly, multiwell plates were coated with 2 g/ml Vn and blocked with 3% BSA. Cells were washed in serum-free RPMI and plated onto the precoated wells for 60 min at 37°C in the absence or presence of competitors in serum-free RPMI. After the incubation period, the wells were washed and the number of adherent cells was quantified by crystal violet staining at 590 nm.

Construction and Characterization of a uPAR Random
Epitope Phage Library-The phagemid library contained a total of 1.67 ϫ 10 4 clones with a phagemid titer of 3.2 ϫ 10 12 phagemids/ml. Eighteen randomly picked phagemid clones from this library contained a uPAR cDNA fragment with a size of 30 -120 bp from either strand, which were distributed over the entire uPAR cDNA sequence (data not shown).
Affinity Selection, Analysis, and Alignment of uPA-binding uPAR Fragments-Affinity selection of uPA-binding phagemid particles was performed as described, and the results are summarized in Table I. From the eluted phagemid particles, 30 Bottom, effect of the uPA-binding fragments from domains II and III on the binding of uPA to immobilized uPAR. As a control, both GST and uPA-non-binder uPAR-(198 -213) were used in the assay. clones from the pH 5.2 elution, 30 clones from the pH 2.2 elution, and 120 clones from the E. coli adsorption were subjected to plasmid extraction and DNA insert sequencing. All clones were found to contain a DNA insert. Only one of the clones eluting at pH5.2, uPAR-(26 -38), had a partial uPAR peptide sequence (Figs. 1 and 3A), presumably due to a recombination of the cDNA fragments during the construction of the phage library. None of the 30 clones eluted at pH 2.2 expressed a uPAR peptide, whereas 9 of the 120 clones eluted by E. coli adsorption, designated as, e.g., uPAR-(5-20), had a correct insert orientation and a correct open reading frame expressing a uPAR peptide. Results of the analysis and alignment of uPA-binding uPAR fragments is presented in Fig. 1, with details of the sequences in Fig. 3. In control experiments, a phage display library containing only vector pComb3B and BstXI adaptor was mixed with uPA-coated beads. After washing 10 times with PBST-3% BSA, no phagemid particles were obtained using the elution steps described above (Table I). In addition, by analyzing uPA-non-binding clones, a uPA-nonbinding uPAR peptide was also identified, which is located at the beginning of the domain III entailing amino acids 198 -213.
Screening of uPA-binding Regions Using a uPAR Peptide Array-Binding of uPA-Fc to uPAR peptide array was performed as described, and a total of 13 uPAR peptides were found to bind uPA-Fc under non-stringent conditions ( Figs 2B). This observation was in accordance with the results from the phage display affinity selection, since nearly all of the uPA-binding uPAR peptides were obtained after completion of the pH 5.2 and pH 2.2 elutions. Furthermore, SPOT-59, CPG-SNGFHNNDTFHF, which overlapped largely with uPAR-(154 -176), remained bound to uPA-Fc to a much lesser extent in the presence of a 10-fold excess of Saruplase uPA (40 g/ml) (Fig. 2C). Finally, SPOT-92, LGDAFSMNHIDVSCC, contained the overlapping sequence from uPAR-(236 -283), uPA-(247-291), uPAR-(241-262), and uPAR-(247-276). Since the stringent washing condition was applied after uPA-Fc binding, enhanced intensity at SPOT-23, -24, and -34 might be due to increased accessibility of bound molecules by anti-Fc antibodies. Similar results were obtained from several repetitions using different uPAR SPOT-membranes. In control experiments, human IgG (4 g/ml) was used instead of uPA-Fc to bind uPAR peptides, and no signal was obtained (Fig. 2D). In summary, the two independent screening procedures for uPA-binding uPAR peptides have produced largely congruent results.
Interference of the GST Fusion uPAR Fragments with the uPA-uPAR Interaction-When uPA binding to immobilized uPAR was performed, all the uPA-binding uPAR fragments from its three domains, and especially the two fragments from the domains II and III were shown to have an agonistic effect, whereas the uPA-non-binding uPAR fragment uPAR-(198 -213) and GST had no effect (Fig. 4). The influence of uPAR fragments was also studied in a reverse system utilizing binding of suPAR to immobilized uPA. Interestingly, whereas the fragments uPAR-(5-20) and uPAR-(13-33) as well as uPAR-(64 -84) and uPAR-(74 -122) from the domain I could partially inhibit suPAR binding to immobilized uPA, the fragments uPAR-(154 -176) and uPAR-(247-276) from the domains II and III again increased the association between suPAR and uPA (data not shown). Additionally, in the converse experiment, the fragment uPAR-(198 -213) and GST had no effect on suPAR binding to immobilized uPA.
Role of the GST Fusion uPAR Fragments in uPAR-dependent Cell Adhesion to Vn-As established previously, the adhesion of leukocytes to immobilized Vn is predominantly mediated by uPAR, and uPA augments this adhesion by increasing the affinity of the Vn-uPAR interaction (38). Only if BAF-3 cells were transfected with full-length uPAR-cDNA did they adhere to Vn and display uPA-augmented adhesion (36). Consequently, the influence of various uPAR fragments at different concentrations on uPAR-transfected BAF-3 cell adhesion to Vn in the absence or presence of uPA (50 nM) was tested. Fragment uPAR-(247-276) from the domain III increased Vn adhesion both in the absence or presence of uPA, whereas uPAR-(154 -176) from the domain II stimulated adhesion only in the presence of uPA (Fig. 5, A and B). Neither peptide stimulated adhesion of non-transfected BAF-3 cells to Vn. The uPAR-(247-276)-induced cell adhesion to Vn was blocked by anti-uPAR mAb R3 and mAb 3936, respectively, which are known to inhibit uPAR-mediated adhesion to Vn (Fig. 5C). A similar reactivity of both peptides was observed in a purified system when the uPAR fragments (1 M) were tested in the binding of Vn to immobilized uPAR. Although uPAR-(247-276) enhanced the Vn-uPAR interaction in a dose-dependent manner in the absence or presence of uPA (50 nM), uPAR-(154 -176) stimulated only the uPA-induced Vn-binding to uPAR (Fig. 6), whereas the other uPA-binding uPAR peptides had no effect on the Vn-uPAR interaction.

DISCUSSION
The urokinase plasminogen activator has been implicated in angiogenesis, fibrinolysis, wound healing, and tumor metastasis. The ability of uPA to participate in cell invasion and the remodeling of connective tissue through its enzymatic activity and to promote cell adhesion through non-enzymatic processes is modulated by interaction with its cellular receptor uPAR (10). The unique topography of uPAR appears to be essential for multiple binding contacts with uPA and other uPAR ligands, since isolated domains or proteolytic fragments of the receptor show by far less efficiency and affinity in ligand binding. The requirement of uPAR domain I as well as the contribution of domains II and III in ligand recognition has been reported, particularly in high affinity binding to uPA (16 -18, 39 -41). Gårdvoll et al. (37) have performed an alanine-scanning analysis of the putative loop 3 of uPAR domain I and have identified four positions (Arg 53 , Leu 55 , Tyr 57 , and Leu 66 ) exhibiting significant changes in the contribution to the free energy of uPA binding upon single-site substitutions to alanine. Furthermore, a recent report suggested that a region in domain II of uPAR (Arg 137 -Arg 145 ) plays a role in uPA binding (42).
Utilizing two different analytical approaches, the binding of uPA to several distinct regions in uPAR, including the above described uPA-binding determinants was observed. The regions in the putative loop 3 of domains II and III, uPAR-(154 -176) and uPAR-(247-276), were shown to have an agonistic influence on the uPA-uPAR interaction independent of the experimental setting, whereas the fragments from domain I could either augment the binding of uPA to immobilized suPAR or partially inhibit suPAR binding to immobilized uPA. However, the physiological relevance of the observed inhibition caused by the fragments from the domain I in the latter experimental setting remains to be further explored. Additional contact sites between uPA and its receptor cannot be excluded, since the constraints imposed within the receptor may well be different from those displayed by phage or immobilized on nitrocellulose membrane. Shliom et al. (43) reported novel interactions between uPA and its receptor and postulated that suPAR in solution is present in equilibrium between oligomer/dimer/monomer forms, and that the mixed population of suPAR might alter the affinity as well as the kinetics with regard to binding of uPA. Our results are the first direct experimental evidence for the fact that direct uPAR-uPAR contacts may contribute to the functional activities of the receptor. This contention is based on the observation that the fragment uPAR-(154 -176) in the domain II and the fragment uPAR-(247-276) in domain III could promote uPAR-dependent cell adhesion on Vn either in the presence of uPA or independently of uPA stimulation in the latter case.
The display of proteins or their functional domains provided a system for the analysis of structure-function relationships, and the potential to generate proteins with altered binding characteristics or novel catalytic properties (44 -46). The phagemid pComb3B was employed here to construct a uPAR phage display random epitope library, which required a library of only limited size in contrast to random peptide epitope mapping (29). Through affinity selection, phagemid particles displaying a functional epitope, e.g. uPA binding, could be selected from the library and by sequence analysis of the corresponding DNA fragment, the primary structure could be determined. The reasons that, in our study, the uPAR phage library was panned for only one round against immobilized uPA were as follows: first, to cover all possible uPA-binding regions; second, to minimize identical clones; and third, to avoid the risk of losing uPAbinding uPAR fragments. Although the peptide array screening (see Fig. 2) was performed under physiological conditions, the uPA-binding phage-displayed uPAR fragments were obtained after pH 5.2 and 2.2 treatment, which suggested considerable affinity for the interactions. The two independent screening procedures for uPA-binding uPAR-derived peptides have generated largely congruent results. Both the SPOT membrane uPAR peptide array as well as the uPAR phage display library have proved useful in determination of binding motifs, whereas the latter has the advantage over immobilized linear peptides that the fused uPAR fragments have a greater potential to adapt to a structural conformation. It is worth noting that both experimental approaches in this study share technical limitations, and that not all possible linear sequences were taken into account, nor were the non-attainable discontinuous sequences. In addition, some of these sequences may not be expressed within the intact receptor, whereas other undetected discontinuous sequences may dominate binding. Antibody and peptide inhibition studies with intact cellular receptor are currently being carried out in order to achieve a complete understanding of the physiological relevance of specific parts of the uPAR molecule.
Our results confirm the existence of a composite uPAbinding structure including all three uPAR domains and that an interdomain cooperation involving all three domains appears to be required to achieve an appropriate functional conformation of uPAR (Fig. 7). Furthermore, our data also support the notion that, when uPA binds to uPAR, both components of the complex undergo conformational changes required for cell adhesion, signal transduction, or the induction of enzymatic activity (47)(48)(49)(50). The former aspect was analyzed in more detail using uPAR-transfected BAF3 cells. The results from adhesion experiments to Vn indicated that uPAR-(247-276) could stimulate uPAR-dependent adhesion by increasing both uPAR-uPA and uPAR-Vn interactions, whereas uPAR-(154 -176) could only affect uPA binding to uPAR and thereby augmented cell adhesion. The fragment uPAR-(247-276) could possibly by itself induce a conformational change and expose Vn-binding site(s) of the receptor, acting in a manner similar to that of uPA. In the presence of uPA, the fragment of domain III would in a non-competitive manner enhance the uPA effect synergistically. The fragment uPAR-(154 -176), on the other hand, could either, in a noncompetitive manner, directly enhance the effect of uPA in terms of exposing Vn-binding site(s) of the receptor or, in the presence of uPA, unfold uPAR molecule exposing domain III and augment the effect induced by the fragment uPAR-(247-276).
In agreement with our findings, Trigwell et al. (51) observed that incubation of macrophages with either full-length or twodomain (domains II ϩ III) soluble uPAR could increase the adhesion of THP-1 cells to vitronectin and fibronectin, and this effect was independent of the epitop SRSRYLE. Similarly, a synthetic peptide comprising the central heparin binding region of Vn (residues 348 -361) not only could bind to plasma Vn but also induced multimerization of Vn (27). Moreover, Sitrin et al. (52) demonstrated that clustering of uPAR induces proinflammatory signaling in human polymorphonuclear neutrophils. Taken together, these results indicate that uPAR activation can occur (a) by direct contact between several receptor molecules and (b) when multiple binding sites on different uPAR molecules serve to form a composite ligand-receptor complex.
These aspects can well be related to recent observations on cluster formation of uPAR within lipid rafts or caveolae (48) and the cross-talk of uPAR with other receptors such as integrins (50). Based on the present data, clustering of uPAR could occur through contacts involving uPAR-(154 -176) and uPAR-(247-276) regions, followed by interactions of these clusters with integrins, both in cis or trans interactions. Additionally, such uPAR-uPAR contacts might play a role in cell surface-associated plasminogen activation and this hypothesis is currently under investigation. Whether fragments of uPAR derived from shedding and cleavage of the receptor in vivo (53) have any potential biological activities similar to those observed in our study remains to be analyzed.