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

J. Biol. Chem., Vol. 282, Issue 18, 13561-13572, May 4, 2007

This Article Abstract Full Text (PDF) All Versions of this Article: 282/18/13561    most recent M610184200v1 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 Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gårdsvoll, H. Articles by Ploug, M. Search for Related Content PubMed PubMed Citation Articles by Gårdsvoll, H. Articles by Ploug, M. Social Bookmarking                      What's this?

Mapping of the Vitronectin-binding Site on the Urokinase Receptor

INVOLVEMENT OF A COHERENT RECEPTOR INTERFACE CONSISTING OF RESIDUES FROM BOTH DOMAIN I AND THE FLANKING INTERDOMAIN LINKER REGION^*

From the Finsen Laboratory, Rigshospitalet, Copenhagen Biocenter, DK-2200 Copenhagen N, Denmark

Received for publication, October 31, 2006 , and in revised form, January 23, 2007.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The urokinase-type plasminogen activator receptor (uPAR) has been implicated as a modulator of several biochemical processes that are active during tumor invasion and metastasis, e.g. extracellular proteolysis, cell adhesion, and cell motility. The structural basis for the high affinity interaction between the urokinase-type plasminogen activator (uPA) and uPAR, which focuses cell surface-associated plasminogen activation in vivo, is now thoroughly characterized by site-directed mutagenesis studies and x-ray crystallography. In contrast, the structural basis for the interaction between uPAR and the extracellular matrix protein vitronectin, which is involved in the regulation of cell adhesion and motility, remains to be clarified. In this study, we have identified the functional epitope on uPAR that is responsible for its interaction with the full-length, extended form of vitronectin by using a comprehensive alanine-scanning library of purified single-site uPAR mutants (244 positions tested). Interestingly, the five residues identified as "hot spots" for vitronectin binding form a contiguous epitope consisting of two exposed loops connecting the central fourstranded -sheet in uPAR domain I (Trp³², Arg⁵⁸, and Ile⁶³) as well as a proximal region of the flexible linker peptide connecting uPAR domains I and II (Arg⁹¹ and Tyr⁹²). This binding topology provides the molecular basis for the observation that uPAR can form a ternary complex with uPA and vitronectin. Furthermore, it raises the intriguing possibility that the canonical receptor and inhibitor for uPA (uPAR and PAI-1) may have reached a convergent solution for binding to the somatomedin B domain of vitronectin.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
The urokinase-type plasminogen activator receptor (uPAR²/CD87) is a modular, glycolipid-anchored membrane protein (1, 2). Through the specific binding of the urokinase-type plasminogen activator (uPA), uPAR assists in regulating and focalizing cell surface-associated plasminogen activation as demonstrated both in vitro (3, 4) and in vivo (5–7). The high affinity interaction between uPAR and uPA (K_D < 1 nM) has been extensively characterized (2, 8). Although all determinants required for this tight binding are contained within the small growth factor-like domain of uPA (GFD residues 1–48) (9), it critically depends on maintenance of a three-domain, modular structure of uPAR (10). Accordingly, site-directed mutagenesis and photoaffinity labeling studies have shown that elements located in distinct domains of uPAR are involved in the interactions with both uPA (8, 11–13) and a potent 9-mer peptide antagonist (9, 14–16). Two crystal structures solved for uPAR in complex with either a linear peptide antagonist (17) or the amino-terminal fragment (ATF) of uPA (18, 19) have provided the structural basis for the existence of a composite ligand-binding site. The assembly of the three homologous domains in uPAR creates a large and deep central ligand-binding cavity, where aliphatic side chains, provided by uPAR domain I, establish a hydrophobic binding site on one side of the cavity. The high affinity for both uPA and the linear peptide antagonists is achieved by an intimate interaction with this cavity (8, 17, 18), providing a well defined target site for rational drug design (20).

As opposed to this, the binding sites mediating the interactions between uPAR and its auxiliary binding partners, e.g. vitronectin (21, 22) and integrins (23, 24), must reside outside this cavity, but the molecular mechanisms underlying these interactions are largely unknown. As these interactions are dependent on or modulated by receptor occupancy with uPA (22, 25–27), uPAR may orchestrate the assembly of these ternary complexes on the cell surface and thus assist in the regulation of cell adhesion and migration. In accordance with this proposition, it has been reported that cell lines with low endogenous uPAR expression adhere poorly to vitronectin, but this adhesion is markedly promoted by receptor saturation with exogenously added uPA (21, 27). Increased uPAR expression in vitro by transfection or cytokine treatment may, however, uncouple this ligand dependence enabling adhesion to vitronectin by a local high density of unoccupied uPAR (27). Partitioning into lipid rafts may generate such a local high density of uPAR and could thus play a regulatory role for uPAR-mediated vitronectin adhesion (28).

In contrast to the well characterized uPAR-ATF interaction, the structural elements in the uPAR·uPA complex that are responsible for the interaction with matrix-deposited vitronectin are largely undefined. Two important functional relationships have nonetheless been established for this interaction. First, the binding sites for uPA and vitronectin on uPAR must be completely nonoverlapping because of the ability to form tri-molecular complexes. Second, the integrity of the intact three-domain structure of uPAR is mandatory for this complex formation (27, 29). So far, only one study (12) has sought to define the functional epitope on uPAR for this interaction. Using homologue-scanning mutagenesis, by which segments of the individual domains of uPAR were swapped with the corresponding sequences in the single domain homologue CD59, Li et al. (12) reported that two segments in uPAR domain II (Asn¹⁷²–Lys¹⁷⁵ and Glu¹⁸³–Asn¹⁸⁶)³ were important for vitronectin binding but not for uPA binding. To gain further insight into the molecular details of this interaction, we have therefore probed the interaction between purified pro-uPA·uPAR complexes and vitronectin by using an almost complete alanine-scanning library of single-site uPAR mutants. We find that the functional epitope for this interaction encompasses a small region located outside the central uPA-binding cavity and involves residues in two loops connecting the central four-stranded -sheet in uPAR domain I as well as residues in the linker peptide connecting uPAR domains I and II. Intriguingly, this functional epitope on uPAR for vitronectin binding resembles the corresponding binding site on PAI-1 for the somatomedin B-like (SMB) domain of vitronectin as both interactions employ arginine as the critical hot spot residue.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Chemicals and Reagents—All reagents for time-resolved fluorescence measurements (DELFIA®) as well as the stable Eu³⁺-chelate used for specific protein conjugation (i.e. N¹-(p-isothiocyanatobenzyl)diethylenetriamine-N¹,N²,N³,N³-tetraacetic acid) were purchased from PerkinElmer Life Sciences.

Protein Preparations—Monomeric human vitronectin purified from plasma was obtained from Molecular Innovations (Southfield, MI). Recombinant active PAI-1 produced in Escherichia coli was a kind gift from Dr. M. D. Andersen (NOVO-Nordisk, Denmark). The monoclonal anti-uPAR antibody R2 (30), which also recognizes uPAR-vitronectin complexes (29), was produced in-house and was labeled with Eu³⁺ using the stable Eu³⁺-chelate of N¹-(p-isothiocyanatobenzyl)diethylenetriamine-N¹,N²,N³,N³-tetraacetic acid at pH 9.3 and a molar ratio of 20:1. This procedure yielded a labeling density of 2–4 molecules of Eu³⁺ per R2. The monoclonal anti-uPA clone 5 as well as a mixture of four monoclonal anti-PAI-1 antibodies (clones 2, 3, 5, and 7) also prepared in-house were labeled with Eu³⁺ following the same procedure.

Expression and Purification of Soluble Recombinant Human uPAR—Soluble forms of human uPAR were expressed and secreted by Chinese hamster ovary cells (CHO) or Drosophila melanogaster Schneider 2 (S2) cells, which were stably transfected with either suPAR (residues 1–277) (33) or pMTC/uPAR (residues 1–283) (31). These proteins are secreted to the conditioned medium because of a deletion of the carboxyl-terminal signal sequence that is required for glycolipid anchoring (1, 2). Single-site alanine replacements were introduced into pMTC/uPAR by site-directed mutagenesis using a previously designed three-gene cassette approach (31), and the corresponding soluble uPAR mutants (244 in all) were expressed by S2 cells and immunoaffinity-purified as described (8). Two additional uPAR variants with multiple mutations were generated in which the sequences ¹⁷²NTTK¹⁷⁵ and ¹⁸³ELEN¹⁸⁶ were swapped with the corresponding sequences found in the glycolipid-anchored, single domain homologue CD59, i.e. ⁶⁵KKDL⁶⁸ and ⁷⁹GTSL⁸². In addition, selected positions were mutated to residues other than alanine. The following mutants were produced: R30E, R30D, R30K, W32F, W32Y, E37K, R58K, R58E, R58D, I63L, I63V, R91D, R91E, R91K, Y92F, and Y92W. All constructs were verified by DNA sequencing using an ABI Prism 310 Genetic Analyzer (Applied Biosystems, Foster City, CA), and the identities of the individually purified uPAR mutants were also verified at the protein level by peptide-mass mapping using in-gel digestion of the reduced and alkylated protein followed by mass assignment and MS/MS-sequencing using an Autoflex^TM TOF-TOF II (8). As judged from SDS-PAGE of 5 µg of reduced and alkylated sample, the purity of these uPAR mutant preparations was >95%. Concentrations of purified uPAR^WT were accurately quantified by amino acid composition analyses (32). Protein concentrations of the purified uPAR mutants were determined using an extinction coefficient of 9.2 (33) except for those mutants involving Trp³² or Trp¹²⁹ for which an estimated of 6.8 was used.

Expression and Purification of Recombinant Human Pro-uPA^S356A and Human Vitronectin Fragments—Human pro-uPA was produced by Drosophila S2 cells stably transfected with a pMTB/uPA (residues 1–411) vector. Expression of pro-uPA^WT or the active site-mutated pro-uPA^S356A was induced by 0.5 mM Cu₂SO₄ as described previously (31), but in this particular case the S2 media also contained 10 µg/ml aprotinin to prevent activation of pro-uPA. The pro-uPA was isolated by immunoaffinity chromatography using an immobilized anti-uPA monoclonal antibody (clone 6). The purity of this pro-uPA was >95%, and the preparation showed no evidence of conversion as judged by SDS-PAGE. The identity of the purified pro-uPA was verified by MS as described (8). Protein concentration of a stock solution of pro-uPA^S356A was accurately determined by amino acid composition analysis (32).

Vitronectin fragment 1–121 was expressed in S2 cells using the expression vector pMTC/Vn-(1–121)/uPAR-DIII (34). This construction covers vitronectin residues 1–121 followed by an enterokinase cleavage sequence (DDDDK) and finally a carboxyl-terminal tag comprising uPAR domain III (uPAR-DIII residues 182–283) to assist subsequent immunoaffinity purification (34). The following single-site alanine replacements (F13A, V15A, D22A, E23A, L24A, S26A, Y27A, Y28A, and Q29A) as well as the double substitutions (T50A/T57A and T50E/T57E) were introduced into pMTC/Vn-(1–121)/uPAR-DIII as described previously (35). All constructs were verified by DNA sequencing. These fusion proteins were purified by immunoaffinity chromatography using an immobilized anti-uPAR antibody (R2), as outlined previously for the purification of uPAR, and used as uncleaved fusion proteins.

Human GFD-(1–48) and SMB-(1–47) domains fused to a carboxyl-terminal His₆ tag were expressed by stably transfected Pichia pastoris strain X-33 using the yeast expression vector pPICZ. The secreted domains were purified by Ni²⁺-nitrilotriacetic acid affinity chromatography followed by reversed-phase chromatography as described previously (8).

Assessing uPAR Binding to Immobilized Vitronectin by Dissociation-enhanced Time-resolved Fluorescence—White Maxisorb fluoroplates (Nunc, Roskilde, Denmark) were coated overnight at 4 °C with 1 µg/µl human vitronectin in 0.5 M carbonate buffer, pH 9.6. Excess binding sites were blocked by a brief incubation with SuperBlock^TM (Pierce) diluted to 50% (v/v) in 0.04 M NaH₂PO₄ and 0.3 M NaCl, pH 7.4. Complexes between 10 nM pro-uPA^S356A and 100 nM uPAR were preformed by incubation for 30 min at room temperature, and the interaction with immobilized vitronectin was subsequently allowed to proceed for 60 min at room temperature on an orbital shaker. All incubations and dilutions were performed in DELFIA® assay buffer (50 mM Tris-buffered saline, pH 7.8, containing 0.5% (w/v) BSA, bovine immunoglobulin, 0.04% (v/v) Tween 40, and 20 µM diethylenetriaminepentaacetic acid). After washing the fluoroplates six times, they were incubated for 60 min with 0.6 µg/ml of the Eu³⁺-labeled R2 monoclonal anti-uPAR antibody in DELFIA® assay buffer followed by six additional washes. Finally, the Eu³⁺ attached to the receptor-bound R2 was dissociated from its chelator by incubation in DELFIA® enhancement solution for 5 min. The fluorescence of the free Eu³⁺ was measured by time-resolved fluorescence using a Fluostar Galaxy fluorometer (PerkinElmer Life Sciences) with excitation set at 340 nm and reading emission at 615 nm with a 400-µs delay and a 400-µs acquisition window.

In a modified version of this binding assay, fusion proteins containing derivatives of vitronectin linked to uPAR domain III (Vn-(1–121)/uPAR-DIII) were immobilized on white Maxisorb fluoroplates using a 2-fold dilution series in coating buffer covering 0.12 to 1 µg/ml. After blocking excess binding sites as described above, the coated wells were either incubated with preformed pro-uPA·uPAR complexes (100 nM uPAR preincubated with 10 nM pro-uPA^S356A as described above), 10 nM active PAI-1, or with buffer alone. In these experiments the DELFIA® assay buffer included an additional 0.1% (v/v) Tween 20 and 0.25 M NaCl. The level of immobilized fusion protein was assessed by adding 0.6 µg/ml of the Eu³⁺-labeled monoclonal anti-uPAR antibody R2 to wells incubated with buffer only. Quantification of the pro-uPA-uPAR interaction with the immobilized vitronectin fusion proteins was accomplished by incubating the relevant wells with the Eu³⁺-labeled monoclonal anti-uPA antibody clone 5 and measuring the released Eu³⁺ as described above. PAI-1 binding was measured by a similar approach after incubation with a Eu³⁺-labeled mixture of four monoclonal anti-PAI-1 antibodies.

Surface Plasmon Resonance Studies—All real time interaction studies were carried out on a Biacore 3000^TM instrument (Biacore, Uppsala, Sweden) using 10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.005% (v/v) surfactant P-20 at pH 7.4 as running buffer. Human recombinant pro-uPA^S356A (0.25 to 1 µg/ml) was immobilized covalently on a carboxymethylated dextran matrix (CM5 sensor chip) using N-hydroxysuccinimide/N-ethyl-N'-[3-(diethylamino)propyl]carbodiimide as described previously (11). To obtain comparative evaluations of the kinetics for the uPA-uPAR interaction, 2-fold serial dilutions of the various uPAR mutants (2–200 nM in running buffer) were analyzed in parallel for uPA binding at a flow rate of 50 µl/min at 20 °C. Compared with previous studies (8, 11), in this study we have used a very high flow rate to minimize effects from mass transport limitations on the assessment of the kinetic rate constants. After each run the sensor chip was regenerated by two consecutive injections of 0.1 M acetic acid in 0.5 M NaCl. The kinetic rate constants, k_on and k_off, were derived from these real time interaction analyses by fitting the association and dissociation phases to a bimolecular interaction model using the BIAevaluation 4.1 software (Biacore, Uppsala, Sweden), as described in detail previously (8).

For direct affinity measurements of the SMB-uPAR interaction, purified human uPAR^WT, uPAR^W32A, and uPAR^R91A (2.5 µg/ml) were immobilized by conventional amine chemistry in separate flow cells on a single CM5 sensor chip with coupling yields ranging from 1,500 to 2,000 RU. 2-Fold serial dilutions of either GFD-(1–48)-His₆ (0.5–125 nM) or SMB-(1–47)-His₆ (0.02–80 µM) were analyzed in parallel for uPAR binding at a flow rate of 50 µl/min at 4 °C. Kinetic rate constants for GFD binding were derived by fitting the data to a single binding site model using BIAevaluation 4.1 software (8). K_D and R_max values for the SMB-uPAR interaction were calculated from the equilibrium binding isotherm by nonlinear curve fitting assuming a model with saturation of a single binding site as shown in Equation 1,

(Eq. 1)

The equilibrium binding (R_eq) and binding capacity (R_max) of the sensor chip are determined in resonance units (RU), but these can be transformed into binding densities using the approximation that 1000 RU equals 1 ng/mm².

For affinity measurements of the interaction between SMB and preformed pro-uPA·uPAR complexes, the immobilized uPAR mutants were initially saturated by 200-µl injections of 200 nM solutions of pro-uPA^S356A, ATF-(1–143), or GFD-(1–48) before 67 µl of 2-fold serial dilution series of SMB were injected (up to 10 µM).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Quantification of the Interaction between Immobilized Vitronectin and Pro-uPA·uPAR Complexes Using Time-resolved Fluorescence—Cellular attachment and adhesion studies to vitronectin are often conducted with immobilized protein in microtiter plates because the adsorbed vitronectin in such studies is considered a suitable in vitro surrogate for the natural, matrix-embedded protein. Several experiments employing either cell cultures (21, 27) or purified components (25, 29) clearly demonstrate that the interaction between uPAR and vitronectin is largely dependent on the degree of saturation of uPAR with uPA. In this study we have used a time-resolved immunofluorescence assay in microtiter plates to facilitate the accurate quantification of uPAR-vitronectin complexes. This method detects binding of preformed pro-uPA·uPAR complexes to immobilized vitronectin using a high affinity monoclonal anti-uPAR antibody (R2), which is conjugated with chelated Eu³⁺ as a time-resolved fluorescence reporter. In this experimental setting, unoccupied uPAR shows virtually no interaction with immobilized vitronectin at concentrations of up to 800 nM (Fig. 1, upper panel). In contrast, formation of pro-uPA·uPAR complexes by addition of the active site-mutated pro-uPA^S356A results in a high affinity interaction between these soluble complexes and immobilized vitronectin. Saturation occurs roughly at equimolarity, as shown for the titration of various fixed concentrations of pro-uPA with uPAR (Fig. 1, upper panel) and vice versa (Fig. 1, lower panel). Studies conducted in parallel with pro-uPA^WT rather than the active site mutant yielded comparable binding isotherms (data not shown). The presence of a molar excess of pro-uPA relative to uPAR has an inhibitory impact on the pro-uPA-uPAR-vitronectin interaction, which is evident from both dose-response profiles outlined in Fig. 1. Nevertheless, using the correct conditions this assay is ideally suited to the large scale screening of uPAR mutants for vitronectin binding. If, as illustrated by the dose-response profiles for the pro-uPA-uPAR-vitronectin interactions (Fig. 1, upper panel), a molar excess of uPAR (100 nM) is analyzed in the presence of limiting amounts of pro-uPA (10 nM), the method becomes relatively insensitive to variations in the concentration of the tested uPAR. Next, the use of high concentrations of both uPAR and pro-uPA relative to the K_D for their binding will minimize the impact of uPAR mutations that moderately affects uPA binding rather than having a direct effect on vitronectin binding. Under these conditions the amount of preformed uPA·-uPAR complexes is thus primarily controlled by the concentration of pro-uPA, which is kept constant.

Mapping the Functional Epitope on uPAR for Vitronectin Binding by Alanine-scanning Mutagenesis—To probe the relative contribution of the individual amino acid side chains in uPAR to the vitronectin binding properties of the corresponding pro-uPA·uPAR complexes, we have individually substituted 244 of the 283 residues in a truncated, soluble uPAR with alanine. Residues defining the consensus sequence of the Ly6/uPAR/-neurotoxin protein domain family (2, 36) were excluded on structural grounds. In cases where the authentic uPAR residue is an alanine, it was replaced by serine, and the five asparagine residues accounting for all the possible N-linked glycosylation-sites were replaced individually by glutamine. The interactions of preformed pro-uPA·uPAR complexes (10 nM) with immobilized vitronectin were measured for all individual uPAR mutants relative to uPAR^WT (Figs. 2, 3, 4). Evaluation of this complete alanine-scanning mutagenesis of uPAR reveals that the functional epitope for vitronectin binding of pro-uPA·uPAR complexes consists of three hot spot residues in uPAR domain I, Trp³², Arg⁵⁸, and Ile⁶³ (Fig. 2); two hot spots in the linker region between domains I and II, Arg⁹¹ and Tyr⁹² (Fig. 3); and two charged residues of moderate importance in uPAR domain I, Arg³⁰ and Glu³⁷ (Fig. 2). None of these seven residues are involved in the high affinity interaction with uPA, because their alanine mutants did not have a significant effect on the kinetic rate constants for this interaction as assessed by surface plasmon resonance (Table 1) (8). In a parallel experiment, the alanine mutants of uPAR were tested for vitronectin binding in the presence of a higher concentration of pro-uPA (25 nM), and an identical profile for the pro-uPA-uPAR-vitronectin interaction was observed demonstrating that formation of pro-uPA·uPAR complexes is not a limiting factor (data not shown).

View this table: [in this window] [in a new window]   TABLE 1 Kinetic rate constants for the uPAR-pro-uPA interaction using uPAR mutants that affect either uPA or vitronectin binding The kinetic rate constants for interactions between immobilized uPA and the selected purified uPAR mutants produced in S2 cells were measured at 20 °C for serial 2-fold dilutions ranging from 2 to 200 nM uPAR using surface plasmon resonance. The mean values and standard deviations for these rate constants are shown for experiments, including 6–35 separate determinations. The use of high flow rates in these experiments (50 µl/min) minimizes the impact of mass transport limitations resulting in somewhat faster rate constants than published previously (8). The KD values were calculated from the mean values determined for the corresponding rate constants KD = koff/kon, and changes in the free energy of binding were calculated as G = RT ln(KD(mut)/KD(WT)), where R is 1.99 cal/mol K, and T is 293 K. Positions showing a >10-fold increase in the KD value upon the specified mutation (i.e. having a G >1.3 kcal/mol) are highlighted by boldface numbers. The surface accessibility of the residues stated has been calculated from the crystal structure of uPAR after removal of a bound peptide antagonist (17).

View larger version (27K): [in this window] [in a new window]   FIGURE 1. Dose-response of the interaction between pro-uPA·uPAR complexes and immobilized vitronectin. In the upper panel different concentrations of pro-uPAS356A were allowed to equilibrate for 30 min with serial dilutions of uPARWT (produced in CHO cells) before the preformed pro-uPA·uPAR complexes were transferred to vitronectin-coated wells, and binding proceeded for 60 min at room temperature. The bound uPAR was subsequently reacted with a Eu3+-labeled monoclonal anti-uPAR antibody (R2) and finally quantified by time-resolved fluorescence of the dissociated Eu3+. The following pro-uPA concentrations were tested: 0 nM (), 5 nM (), 10 nM (), 20 nM (•), and 40 nM (). The inset shows the linear dose-response curve for the vitronectin interaction in the presence of 100 nM uPAR. In the lower panel the inverse experiment has been performed in which fixed concentrations of uPARWT (10 nM (), 20 nM (), and 40 nM ()) were allowed to equilibrate with serial dilutions of pro-uPAS356A before reaction with immobilized vitronectin. The inset shows the uPA dependence for the binding of 40 nM uPAR to immobilized vitronectin. Raw fluorescence data are shown in both cases, but generally the nonspecific signal from noncoated wells treated in parallel never exceeded 10%. Arrows indicate the concentration of uPAR (upper panel) or pro-uPA (lower panel) causing saturation of the vitronectin binding, which occurs at a roughly 1:1 stoichiometry in both cases. Each data point represents the mean of two replicate measurements. The concentrations of the proteins used in these experiments were carefully determined by amino acid analysis.

View larger version (31K): [in this window] [in a new window]   FIGURE 2. Defining the functional epitope in uPAR domain I for the interaction between pro-uPA·uPAR complexes and immobilized vitronectin by alanine-scanning mutagenesis. The relative binding of preformed pro-uPA·uPAR complexes to immobilized vitronectin is shown for 78 single-site uPAR mutants in domain I compared with that of uPARWT. The error bars for the individual mutants represent the standard deviation for data derived from 4 to 10 separate determinations measured relative to a uPARWT standard analyzed in parallel on all plates. Binding of uPAR to vitronectin was quantified by the Eu3+ fluorescence released from bound Eu3+-labeled R2 as specified in Fig. 1, but in this case the fluorescence from noncoated wells treated in parallel has been subtracted. Arbitrary cut-off levels for vitronectin reactivity of these mutants are set at 25 and 50% of that obtained with uPARWT. Three positions (Trp32, Arg58, and Ile63) are considered "hot spots" for the vitronectin interaction (black bars), whereas two positions exhibit an intermediate reactivity (Arg30 and Glu37, shown as dark gray bars).

View larger version (34K): [in this window] [in a new window]   FIGURE 3. Defining the functional epitope in uPAR domain II for the interaction between pro-uPA·uPAR complexes and immobilized vitronectin by alanine-scanning mutagenesis. These data were generated as outlined in Fig. 2 for mutants in uPAR domain I. Of the 80 single-site mutants tested in uPAR domain II, we found only three positions (Arg91, Tyr92, and Leu113) that exhibited a reactivity to immobilized vitronectin that was <25% of that for uPARWT, whereas one position presented an intermediate reactivity (Glu120). The effect of alanine substitutions of Leu113 and Glu120 is most likely caused by structural perturbations as these residues have a limited surface exposure in the x-ray structure of uPAR, and furthermore uPARL113A is the only mutation found that seriously affects vitronectin as well as uPA binding.

View larger version (35K): [in this window] [in a new window]   FIGURE 4. Defining the functional epitope in uPAR domain III for the interaction between pro-uPA·uPAR complexes and immobilized vitronectin by alanine-scanning mutagenesis. These data were generated as outlined in Fig. 2. None of the 86 single-site mutants tested in uPAR domain III were found to significantly impair the reactivity with immobilized vitronectin as compared with uPARWT. The lack of reactivity presented by suPARD275A is not a consequence of an impaired reactivity with immobilized vitronectin but is caused by its impaired reactivity with the Eu3+-labeled R2 monoclonal anti-uPAR antibody used for detection.

Differential Mutagenesis of Residues in uPAR Important for Vitronectin Binding—To investigate the contribution of the individual amino acid side chains of the functional epitope on uPAR for vitronectin binding, they were mutated separately into residues other than alanine. The results obtained from this differential scanning mutagenesis are shown in Fig. 5. Charge reversal of Arg³⁰ or Glu³⁷ did not have any further effect on vitronectin binding as compared with the corresponding side chain deletions beyond the respective C atoms. Long range ionic interactions involving these residues are consequently not considered important for the uPA-uPAR-vitronectin interaction. By contrast, charge reversals of Arg⁵⁸ had a strong impact on vitronectin binding. However, these radical changes also severely affect uPA binding (Table 1), despite Arg⁵⁸ not being part of the uPA-binding site (8). Therefore, these mutations most likely compromise the interface between domains I and II rather than affect vitronectin binding directly. Charge reversal of Arg⁹¹ did not further impair vitronectin binding compared with uPAR^R91A, but the actual charge contribution from this hot spot residue cannot be established in this experiment because the alanine mutation of Arg⁹¹ on its own reduces the interaction with immobilized vitronectin to a level that is very close to the base-line reaction observed in the absence of uPAR (Fig. 5). None of the differential mutations replacing the surface-exposed Arg⁹¹ have any significant impact on uPA binding (Table 1), which is concordant with its location outside the central uPA-binding cavity (17, 18).

We next addressed the importance of the aromatic side chains of Trp³² and Tyr⁹². Substitutions of Trp³² with other aromatic side chains only marginally improved the vitronectin binding compared with the alanine mutant, and binding was still far below that of uPAR^WT (Fig. 5) highlighting the importance of the indole side chain at this position. By contrast, the impaired vitronectin binding of uPAR^Y92A was almost completely restored to wild-type levels in uPAR^Y92F and uPAR^Y92W. This observation argues for a role of the planar aromatic side chain rather than the involvement of specific hydrogen bonding of the hydroxyl group of Tyr⁹². None of the mutations of aromatic side chains have any measurable effect on uPA binding (Table 1).

Finally, conservative replacements of Ile⁶³ with Leu or Val revealed differing effects on vitronectin binding with the clear preference of Ile > Val >> Leu and Ala for optimal binding (Fig. 5). This demonstrates the limited spatial freedom in the geometry of the C and C atoms that is allowed at this position.

A previous study (12) reported that residues 172–175 and 183–186 of uPAR are involved in the interaction with vitronectin based on homology-scanning mutagenesis, where short, selected peptide sequences in uPAR were exchanged individually with the corresponding sequences in CD59, a single domain member of the Ly6/uPAR/-neurotoxin protein domain family (2, 36). However, our present alanine-scanning mutagenesis of uPAR does not implicate any single residue in these two regions as being of importance for the interaction with vitronectin, when mutated individually to alanine (Figs. 3 and 4). We therefore constructed the two relevant uPAR mutants (uPAR^{N172-K175/CD59} and uPAR^{E183-N186/CD59}) and measured the kinetic rate constants for their interaction with uPA by surface plasmon resonance as well as their capacity for vitronectin binding in the time-resolved fluorescence assay. Both of these tetra-mutants exhibited unaltered uPA binding (Table 1), but we found that they also retained close to normal vitronectin binding (Fig. 5) in accordance with the data obtained in the single-site alanine-scanning mutagenesis.

SMB Exploits the Same Hot Spot as Vitronectin for Its Low Affinity Binding to uPAR—As the engagement of vitronectin in uPAR and PAI-1 binding is largely governed by the small amino-terminal SMB domain (37, 38), we next investigated this interaction by surface plasmon resonance using comparable levels of immobilized uPAR^WT, uPAR^W32A, and uPAR^R91A. Initially, the binding capacity (B_max) and kinetic rate constants for the high affinity interaction with purified GFD were established for each flow cell containing one of these uPAR mutants (Fig. 6, upper panel, and Table 2). The corresponding kinetic rate constants for the interaction between immobilized uPAR^WT and SMB were too fast, however, to be quantified reliably by the Biacore3000^TM (Fig. 6, middle panel). This particular interaction was consequently analyzed under conditions yielding equilibrium binding. As illustrated in Fig. 6 (lower panel), this revealed a specific interaction between uPAR^WT and SMB, whereas the linear and parallel dose-response profiles recorded for uPAR^W32A and uPAR^R91A are indicative of nonspecific binding (solid lines). Taking further advantage of this observation, we used the dose-response profile recorded for uPAR^R91A as an approximation for the subtraction of nonspecific binding to uPAR^WT. The resultant binding isotherm for the SMB-uPAR^WT interaction was subsequently fitted to saturation of a single binding site (Fig. 6, lower panel, broken line). This yields a K_D of 1.9 µM for the SMB-uPAR^WT interaction and a binding capacity for the chip of 8.5 fmol of SMB/mm² (Table 2). Reassuringly, the SMB and GFD modules bind with a similar stoichiometry to the immobilized uPAR^WT (i.e. comparable binding capacities) despite their widely differing K_D-values (Table 2). These data clearly demonstrate that SMB exploits the same hot spot binding site on uPAR as intact immobilized vitronectin. Unfortunately, a similar quantitative interaction analysis using intact vitronectin is precluded because of the instability and heterogeneous state of this purified protein preparation.

View larger version (22K): [in this window] [in a new window]   FIGURE 5. Differential mutagenesis of the functional hot spots for vitronectin binding in uPAR. Residues in uPAR considered important for vitronectin binding were differentially mutated to several different amino acids and tested for vitronectin binding in the presence of 10 nM pro-uPA as described in the legend to Fig. 2. Also shown is the vitronectin binding of two purified uPAR mutants derived by homology "scanning" mutagenesis during which defined uPAR sequences were exchanged with the corresponding sequences found in CD59 as published previously (12). uPARN172-K175, NTTK175 replaced by KKDL and uPARE183-N186, ELEN186 replaced by GTSL.

The present alanine-scanning mutagenesis of SMB largely confirms data obtained from a previous mutagenesis study (38), with the exceptions that we did not identify Glu²³ as a functionally important residue for binding uPAR·pro-uPA complexes, but we did find Phe¹³ to be important.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
In this study we have characterized the complete functional epitope on the uPAR that is responsible for the binding of uPA· uPAR complexes to immobilized vitronectin. This interaction is considered functionally important as several studies in cell culture clearly show that the level of uPAR expression regulates vitronectin-dependent cell adhesion and motility (21, 22, 25, 27, 40). However, in cells with a low to moderate level of uPAR expression, the uPAR-dependent adhesion to vitronectin is also controlled by the level of receptor saturation with its cognate protease ligand uPA (21, 27). This emphasizes the importance of the tri-molecular pro-uPA·uPAR·vitronectin complex in the regulation of this process.

The identification of the functional epitope on uPAR for vitronectin binding combined with our previous identification of the functional epitope for uPA binding (8, 11) has enabled us to assemble a low resolution model for these interactions as represented by the cartoon shown in Fig. 9. This model allows for the formation of a tri-molecular complex, in which the small SMB module of vitronectin is positioned on the "apical side" of uPAR distal to the "basal" attachment site for the glycolipid membrane anchor. This topology is particularly compelling considering that the glycolipid-anchored uPAR should be able to bind concomitantly to both the opposing matrix-embedded vitronectin and to pro-uPA. Importantly, the identified functional epitope for the uPAR-vitronectin interaction forms a coherent surface-exposed patch on uPAR with a size comparable with the corresponding epitope identified on the SMB module (Fig. 9). This observation is excellently aligned with previous biochemical studies highlighting the SMB domain of vitronectin as the primary interaction site on vitronectin for uPAR (38).

View larger version (27K): [in this window] [in a new window]   FIGURE 6. SMB and GFD interaction with immobilized uPARWT, uPARW32A, and uPARR91A as assessed by surface plasmon resonance. Comparable levels of purified uPARWT (1665 RU), uPARW32A (1708 RU), and suPARR91A (2003 RU) were immobilized on a CM5 sensorchip by conventional amine coupling. 2-Fold serial dilutions of purified GFD-(1–48) (from 125 nM) and SMB-(1–47) (from 80 µM) were injected after 248 s and allowed to interact with the immobilized uPAR preparations for 300 s at a flow rate of 50 µl/min at 4 °C. The data recorded by Biacore3000TM for the GFD-uPARWT interaction are shown in the upper panel and that for SMB-uPARWT interaction in the middle panel. The two insets show the corresponding interaction profiles recorded for SMB-uPARW32A and SMB-uPARR91A at the same scale. The lower panel shows the equilibrium binding isotherms for the interactions of SMB with immobilized uPARWT (•), uPARW32A (), and uPARR91A () measured at 500 s (arrows in middle panels). The equilibrium-binding isotherm for the SMB-uPARWT interaction "corrected for nonspecific" binding by subtracting the signal recorded for the flow cell with immobilized uPARR91A is shown () along with the fit to a one binding site saturation model by nonlinear regression (broken line).

Two additional residues located outside uPAR domain I contribute significantly to vitronectin binding (Arg⁹¹ and Tyr⁹²), with Arg⁹¹ in particular representing a dominant hot spot for this interaction. Both residues are located in the flexible linker region between uPAR domains I and II, which is not well defined by the electron densities in either of the two different crystal structures solved for uPAR (17, 18). Interestingly, this linker region has been implicated in the chemotactic effect of the receptor that is elicited either by uPA binding (41) or by proteolytic cleavage of the linker region (41, 42) and that involves activation of the FPRL1 receptor (43). The present assignment of the vitronectin-binding site to this particular region in uPAR suggests that the engagement of uPA·uPAR complexes in vitronectin binding during cell adhesion would most likely preclude a concomitant ligand-induced chemotactic effect (41). Chemotaxis is thought to be mediated by a pentapeptide Ser⁸⁸-Arg-Ser-Arg-Tyr⁹² situated in the linker region (42, 44), which would presumably be shielded by vitronectin binding. Our data clearly show that the two last residues of this pentapeptide provide an indispensable contribution to the functional epitope on uPAR for vitronectin binding (Figs. 3, 6, and 9). The proposed interaction between this chemotactic peptide in uPAR and the FPRL1 receptor (43) can therefore occur only under conditions where uPAR is not engaged in vitronectin binding.

View larger version (19K): [in this window] [in a new window]   FIGURE 7. SMB binding to uPARWT saturated with GFD, ATF, or pro-uPA as assessed by surface plasmon resonance. Purified uPARWT (1571 RU), uPARW32A (1831 RU), and suPARR91A (2355 RU) were immobilized on a CM5 sensorchip. In some experiments the immobilized uPAR was saturated by a 240-s injection of 200 nM GFD-(1–48), ATF-(1–143), or pro-uPARS356A before a 2-fold serial dilution of purified SMB-(1–47) (from 10 µM) was injected after 690 s and allowed to interact with the immobilized uPAR for 80 s. The sensorgrams recorded for the interaction of SMB with ATF-uPARWT complexes at 4 °C and a flow rate of 50 µl/min are shown in A. The resulting equilibrium binding isotherms for the interactions of SMB with immobilized uPARWT (•), GFD-(1–48)-uPARWT (), ATF-(1–143)-uPARWT (), and pro-uPAS356A-uPARWT () measured at 700 s are shown in B. Curve fittings to a one binding site saturation model for SMB are shown for unoccupied uPAR (broken line) and ATF-(1–143)-uPARWT (solid line).

View larger version (17K): [in this window] [in a new window]   FIGURE 8. Defining the functional epitope in vitronectin for uPAR-pro-uPA. Equal levels of the indicated vitronectin derivatives were immobilized on microtiter plates and incubated with 10 nM preformed pro-uPA·uPAR complexes. Specific binding was quantified by subsequent incubation with a Eu3+-labeled anti-uPA clone 5 monoclonal antibody and is shown relative to that obtained with immobilized recombinant vitronectin-(1–121). Insignificant binding of 10 nM pro-uPA to vitronectin-(1–121) was detected in the absence of uPAR (-). Nonspecific binding from uncoated wells has been subtracted, but this was always less than 5% of wild-type binding. Each bar represents the mean value of 6–10 measurements.

View larger version (39K): [in this window] [in a new window]   FIGURE 9. Cartoon of the tri-molecular ATF·uPAR·SMB complex based on alanine-scanning mutagenesis. The proposed model of the complex between uPAR, ATF, and SMB is shown as a ribbon diagram. The SMB domain of vitronectin is positioned manually on the uPAR-ATF structure without any energy minimization attempts. The homologous three-finger fold domains of uPAR are highlighted by yellow (domain I), blue (domain II), and red (domain III), and the position of the glycolipid anchor, which tethers uPAR to the cell membrane, is also indicated. The ATF of uPA is shown in light blue, whereas the SMB domain of vitronectin is shown in green. -Helices are colored red. The side chains identified in this study by alanine-scanning mutagenesis to be important for the interaction between uPAR·pro-uPA complexes and immobilized full-length vitronectin are shown as white sticks with preservation of the orientations defined in the original x-ray structures. Note that the key residue Arg91 is not shown in this illustration as the electron densities in neither of the two solved uPAR structures (17, 18) define this residue. The Protein Data Bank coordinates used to construct this model are 2FD6 (uPAR-ATF) (18) and 1OC0 (SMB-PAI-1) (46) using the PyMOL software (DeLano Scientific).

	FOOTNOTES

 
^* This work was supported by grants from The John and Birthe Meyer Foundation, The Lundbeck Foundation, the Danish Cancer Society, and European Union Contract LSHC-CT2003-503297. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

¹ To whom correspondence should be addressed: Finsen Laboratory, Copenhagen Biocenter Bldg. 3, 3rd Floor, Ole Maaløes Vej 5, DK-2200 Copenhagen N, Denmark. Fax: 45-35453797; E-mail: m-ploug{at}finsenlab.dk.

² The abbreviations used are: uPAR, uPA receptor; uPA, urokinase-type plasminogen activator; ATF, amino-terminal fragment of uPA; PAI-1, plasminogen activator inhibitor 1; SMB, somatomedin-B like domain of vitronectin; Vn, vitronectin; WT, wild type; RU, resonance units; GFD, growth factor-like domain of uPA.

³ The numbering of amino acid residues in PAI-1, uPA, uPAR, and vitronectin refers to the cDNA-derived sequences omitting the signal sequences. Nomenclature for the secondary structure elements in uPAR follows the conventions established for snake venom -neurotoxins and are more explicitly clarified for the modular