regions involved in binding of urokinase-type-1 inhibitor complex

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 41, Issue of October 14, pp. 25668-25676, 1994 Printed in U.S.A.

Regions Involved in Binding of Urokinase-Type-1 Inhibitor Complex and Pro-urokinase to the Endocytic a,-Macroglobulin Receptor/ Low Density Lipoprotein Receptor-related Protein EVIDENCE THAT THE UROKINASE RECEPTOR PROTECTS PRO-UROKINASE AGAINST BINDING TO THE ENDOCYTIC RECEPTOR*

(Received for publication, March 10, 1994, and in revised form, June 16, 1994)

Anders NykjaerSO, Lars Kj~llerfl, Robert L. Cohenll, David A. Lawrencell, Beth Ann Garni-Wagner**, Robert E Todd III**, Anton-Jan van ZonneveldSS, J ~ r g e n GliemannS, and Peter A. Andreasenn From the $Department of Medical Biochemistry and Wepartment of Molecular Biology, University of Aarhus, DK-8000 Aarhus C, Denmark, the IlCancer Research Institute, University of California, Sun Francisco, California 94143, the **Department of Internal Medicine, University of Michigan Medical Center, Ann Arbor, Michigan 48109, and the Department of Biochemistry, University of Amsterdam, Amsterdam 1105 A2 The Netherlands

The a,-macroglobulin receptorflow density lipoprotein receptor-related protein (aJMRLRP) binds several ligands, including complex between the two chain urokinase-type plasminogen activator (uPA) and type-1 plasminogen activator inhibitor (PAI-l), and the single chain zymogen pro-urokinase (pro-uPA). We have used truncated variants of uPA and PAI-1 as well as Fab fragments of monoclonal antibodies with known epitopes to identify regions in the uPA-PAI-1 complex and in pro- uPA involved in binding to c+R/LRP. uPA.PAI-1 complex bound with high affinity (EC, about 0.4 m) via contacts in the PAI-1 moiety as well as the uPA serine proteinase domain and the uPAA chain. Pro-uPA bound with lower affinity (EC,, about 10 m), and efficient binding to a,MR/LRP was dependent on contact with both the A chain and the serine proteinase domain. We analyzed the effect of complex formation with the urokinase receptor since this is the primary target for binding of uPA.PAI-1 and pro-uPA at the cell surface, and since it has been demonstrated that urokinase receptor-bound uPA.PAI-1 complex is internalized following interaction with a,MR/LRP (Nykjaer, A, Petersen, C. M., M~ller, B., Jensen, P. H., Moestrup, S. K., Holtet, T. L., Etzerodt, M., Th~gersen, H. C., Munch, M., Andreasen, P. A, and Gli- emann, J. (1992) J. Biol. Chem. 267,14543-14546). Soluble recombinant urokinase receptor blocked the binding of pro-uPA to aJMRLRP but caused only a slight reduction in the affinity for binding of uPA-PAI-1. Moreover, pro- uPA bound to the urokinase receptor at the cell surface was not internalized and degraded unless activated to uPA and complexed with PAI-1. We conclude that pro- uPA is protected against degradation via a,MR/LRP when bound to uPAR due to shielding of a binding contact in the A chain, whereas the affinity of uPAR- bound uPA-PAI-1 complex for binding to a,MR/LRP remains sufficient to allow rapid internalization and degradation.

* This work was supported by grants from the Danish Cancer Society, the Danish Medical Research Council, the Danish Biotechnology Pro- gram, and Aarhus University Research Foundation. 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.

5 To whom correspondence should be addressed: Dept. of Medical Biochemistry, University of Aarhus, Ole Worms Ale, Bldg. 170, DK- 8000 Aarhus C, Denmark. Tel.: 45-89422884; Fax: 45-86131160.

The urokinase-type plasminogen activator (uPA)l catalyzes the conversion of plasminogen to plasmin, a proteolytic enzyme of broad specificity (for review, see Ref. 1). The secreted form is the single chain zymogen pro-urokinase (pro-uPA) which binds with high affinity to the glycolipid-anchored uPA receptor (uPAR) (for review, see Refs. 2, 3). Pro-uPA is activated by cleavage of a single peptide bond, and the cell-bound two-chain active uPA mediates pericellular proteolysis and matrix degradation, e.g. in cell migration, tissue remodeling, and tumor invasion (for review, see Refs. 2-4). The major inhibitor is plasminogen activator inhibitor type-1 (PAI-1) which forms stable complexes with soluble as well as uPAR-bound uPA (5; for review, see Ref. 6). uPA bound to cellular uPAR is only internalized and degraded to a minor degree, whereas uPAR-bound complex between uPA and PAI-1 is readily degraded (7-10). The reason is that the uPA.PAI-1 complex, in contrast to uncomplexed two-chain uPA, is internalized following binding to the multiligand receptor a,-macroglobulin receptorflow density lipoprotein receptor-related protein (a,MR/LRP) (11, 12). Re- cently, pro-uPA was reported to bind to a,MR/LRP, although with lower affinity than the uPA.PAI-1 complex (13). Other ligands of a,MRLRP include receptor-active a,M, e.g. a2M- methylamine (a,M-me), tissue-type plasminogen activator (tPA), tPA.PAI-1 complex, and the 39-40-kDa a,MR/LRP-associated protein (RAP) which inhibits binding of the other estab- lished ligands (for review, see Ref. 14).

uPA has an NH,-terminal A chain comprising amino acids (aa) 1-158 and containing a growth factor domain (aa 1-49) responsible for binding to uPAR, a kringle domain (aa 50-1311, and a linker region (aa 132-158) (1). The B chain comprises aa 159-411 and contains the serine proteinase domain. PAI-1 be- longs to the serpin superfamily, and the three-dimensional structure of latent PAI-1 was recently determined (15). In ad-

, The abbreviations used are: uPA, urokinase-type plasminogen activator; aa, amino acids; a,I,.ct, a1-inhibitor-3-chpotrypsin; alPI, a,- proteinase-inhibitor; a,M-me, a,-macroglobulin-methylamine; a,MW LRP, a,-macroglobulin receptorLDL receptor-related protein; ATIII, antithrombin 111; ATF, amino-terminal fragment of urokinase-type plasminogen activator; DFP, diisopropylflourophosphate; DFP-uPA, DFP- inhibited urokinase-type plasminogen activator; PAI-1, type-1 plasminogen activator inhibitor; PVDF, polyvinylidene fluoride; RAP, a,-macroglobulin receptor-associated protein; tPA, tissue-type plasminogen activator; uPA.PAI-1, complex between urokinase-type plasminogen activator and type-1 plasminogen activator inhibitor; uPAR, urokinase receptor; LRP, low density lipoprotein-receptor-related protein; PAGE, polyacrylamide gel electrophoresis.

25668

Binding of uPA-PAI-1 Complex and Pro-uPA to aflRILRP 25669

dition, the structure has been determined for a few reactive center-cleaved serpins, including the closely homologous al- proteinase inhibitor (alPI) (16). uPAR is an MR 55,000 glycosyl phosphatidylinositol-anchored protein containing three homologous domains of which the most NH,-terminal binds the growth factor domain of uPA (2, 3, 17).

The purpose of the present work was to identify regions involved in the binding of uPA.PAI-1 complex and pro-uPA to the endocytic receptor a,MR/LRP, and to analyze the functional consequences. The effect of binding to uPAR via the uPAA chain was examined in view of the importance of this reaction at the cell surface. We demonstrate that following binding to uPAR, the binding of pro-uPA to a,MFULRP is blocked, whereas the affinity of the uPA.PAI-1 complex is only slightly reduced. As a result, pro-uPA bound to uPAR at the cell surface is protected against the a,MWLRP-mediated internalization and degradation.

MATERIALS AND METHODS Receptors-Human a,MWLRP was purified from solubilized placen-

tal membranes by a,M-me affinity chromatography as described else- where (18). Rabbit gp330 (19) was a gift from Dr. S. K. Moestrup, Department of Medical Biochemistry, University of Aarhus, Denmark. Recombinant human RAP (11) was provided by Dr. H. C. Thogersen, Laboratory of Gene Expression, University of Aarhus, Denmark. Re- combinant soluble human uPAR (amino acids 1-281) was produced in Chinese hamster ovary cells and purified by affinity chromatography on uPA-Sepharose as described (20). The protein migrated on SDS-PAGE as a single broad band centered at approximately 50-55 kDa and appeared >98% pure as assessed by SDS-PAGE and silver-staining. The purified uPAR reacted with both polyclonal anti-uPAR antisera and a clustered monoclonal antibody against uPAR (21). Recombinant soluble uPAR was similar to authentic uPAR in terms of affinity for uPA, extent of glycosylation, and susceptibility to chymotrypsin cleavage' (17).

Proteinases, Inhibitors, and Proteinase-inhibitor Complexes-Human two-chain active uPA was from Serono, Switzerland. Diisopropylfluoro- phosphate inactivated uPA (DFP-uPA) was prepared as described previously (22). Recombinant human pro-uPA, ATF, and an NH, terminally truncated variant of human uPA with NH, terminus at leucine 144 (Leu"-uPA) were kindly donated by Dr. J. Henkin, Abbott Company, Abbott Park, IL. The amino-terminal fragment of uPA (ATF, aa 1-136) contained less than 0.5% two-chain uPA or L Y S ' ~ ~ - ~ P A a s judged by its ability to hydrolyze a chromogenic substrate. Lysine 136-uPA (LYS'~~- uPA) was a gift from Dr. W. Gunzler, Griinenthal Company, Aachen, Germany. Protein sequencing confirmed the NH, termini of the truncated uPA forms. Human PAI-1 in latent, active, and reactive center- cleaved forms were prepared as described previously (23, 24). tPA was from Boehringer-Ingelheim, Germany. Human thrombin was purchased from Sigma. Human antithrombin I11 (ATIII) was a gift from the late Dr. Staffan Magnusson, Department of Molecular Biology, University of Aarhus.

In order to prepare uPA.PAI-1, active PAI-1 and uPA were incubated overnight at 4 "C. The complexes were isolated from unreacted proteins by sequential immunoaffinity chromatography (8, 11) on columns with monoclonal anti-uPA from hybridoma clone 6 and monoclonal anti- PAI-1 from hybridoma clone 2 (cf. legend to Fig. 1). Similar procedures were used for preparing tPA.PAI-1 complexes and complexes between PAI-1 and the NH, terminally truncated forms of uPA. uPA.ATIII complexes were prepared by incubating uPA with a 6-fold molar excess of ATIII and a 36-fold molar excess of heparin for 48 h at 20 "C, followed by immunoaffinity purification of the complexes on a column with monoclonal anti-uPA IgG from hybridoma clone 6. Human a,M-me and a,-inhibitor-,-chymotrypsin complex (a,I,.ct) were kindly donated by Drs. K. Dolmer and L. Sottrup-Jensen, Department ofMolecular Biology, University of Aarhus (25). All ligands were labeled with about 1 mol of 1261/mol protein following previously published procedures (11, 18, 22).

Antibodies-Three clones of murine hybridomas producing monoclonal anti-human uPAIgG, termed clones 2,6, and 12, were described earlier (2628). Anti-human PAI-1 IgG clones 2, 3, 5, and 6 were described previously (23) as well as the monoclonal anti-PAI-l antibodies CLB 16, CLB 2C8, CLB 10, CLB 15, CLB 33, and CLB 1C3 (29). Fab fragments of the mouse monoclonal IgGs were prepared according to a method described previously (30). The monoclonal anti-CD3 antibody, orthoclone OKT3, was from Cilag, Zug, Switzerland. Anti-a,MFULRP

D. A. Lawrence and R. L. Cohen, unpublished observations.

rabbit serum was directed against a recombinant fragment (aa 2500- 2922) (31) of the a-chain.

Assignment of the epitopes of monoclonal anti-uPA antibodies to different structural domains of uPA was undertaken by an enzyme- linked immunosorbent assay technique. The epitopes for monoclonal anti-PAI-1 antibodies from hybridoma clones 1-3 have been localized to certain stretches of the polypeptide chain (29). By the use of the three- dimensional structure of reactive center-cleaved a,PI (16), and the sequence alignment of different serpins given by Huber and Carrel1 (321, these epitopes may be assigned to different structural elements (a- helices or strands in p-sheets) which, due to the strong sequence ho- mology, are expected to exist in both PAI-1 and a,PI. The ability of monoclonal anti-uPA and anti-PAI-1 antibodies to bind uPA.PAI-1 was determined by an enzyme-linked immunosorbent assay technique in which the various monoclonal antibodies were immobilized, followed by incubation with uPA.PAI-1, and detection by rabbit polyclonal antibodies directed against uPA or PAI-1. In addition, the ability of the clones to bind the complex was tested using antibodies immobilized on Sepha- rose (Pharmacia, Sweden). Monoclonal antibodies of irrelevant specificity were used as negative controls.

(American Type Culture Collection (ATTC) CRL 1650) and murine NIW Cell Binding Experiments-COS-1 cells from African green monkey

3T3 cells (ATCC CRL 1658) were cultured as described (33). NIW3T3 cells were cotransfected with the cDNA encoding the human uPAR (34) and that for neomycin resistance using the calcium phosphate method as described previously (35). Neomycin-resistant clones were screened for uPAR expression by indirect immunofluorescence flow cytometry using clustered anti-uPAR monoclonal antibody (3B10f; 21) binding relative to the nonspecific binding of an isotype-identical negative control monoclonal antibody (anti-CDllb) (35). Among the clones screened, clone NIW3T3-3 demonstrated consistent, stable expression of human uPAR, while the wild-type nontransfected cell line showed no uPAR expression. Binding of radioactive ligands to adherent COS-1 cell mono- layers or NIW3T3 cells in suspension was assayed essentially as described previously (8, 22). Degradation of labeled ligands was assessed by measuring radioactivity soluble in 12% trichloroacetic acid (9).

Binding to Purified a$fRLRP and uPAR-Microtiter wells from NUNC (Denmark) were coated with purified a,MWLRP or uPAR, and binding of lZ6I-labeled ligands to the immobilized receptors was measured as described before (36). In the absence of immobilized receptor, the apparent binding (blank values) amounted to 0.1% for uPA,PAI-1 and 0.2% for pro-uPA, a,M-me, and alIs.ct. The measurements were corrected accordingly.

Binding to the receptor in solution was assayed by incubating 2 nM a,MWLRP, 10 PM '261-uPA.PAI-1, and various concentrations of unlabeled ligand at 4 "C for 16 h. Aliquots were then partitioned in an aqueous dextran T500 polyethylene glycol 6000 two-phase system by stand- ard methods. The partition coefficients of the radioactive species was used to calculate the fractions of free and receptor-bound ligand (37).

Western and Ligand Blotting-SDS-PAGE was performed on 4 or P 1 6 % polyacrylamide gradient gels using 5% SDS or, in the case of ligand blotting, 0.1% SDS. Proteins were transferred to polyvinylidene fluoride (PVDF) membranes (Millipore, Bedford, MA) by electroblot- ting, and analyzed by immuno- and ligand blotting techniques as described earlier (11).

RESULTS

Regions Involved in Binding of uPA.PAI-1 Complex-We have previously demonstrated that binding of uPA.PAI-1 to purified a,MR/LRP is inhibited by monoclonal antibodies directed against PAI-1 as well as against uPA (ll) , suggesting that both moieties might be involved in binding of the complex. To test this hypothesis, we incubated about 10 PM lZ6I-labeled uPA.PAI-1 with a,MR/LRP immobilized in microtiter wells in the presence of increasing concentrations of various unlabeled compounds. As shown in Table I, cleaved PAI-1, active PAI-1, and latent PAI-1 inhibited binding of the labeled complex by at least 95%, and the concentrations causing half-maximal inhibition (EC,,) of i251-uPA.PAI-1 binding were approximately 50 nM as compared with 0.4 nM for uPA.PAI-1 itself. tPA.PAI-1 competed with an EC,, of 3 nM even though tPA alone was a poor competitor. tPA alone binds to a,MFULRP (38) and is not competed for by pro-uPA(39). This suggests that the lower EC,, value for the tPA.PAI complex than for PAI-1 itself is caused by a higher affinity of the complex and therefore a more efficient

25670 Binding of uPA*PM-1 Complex and Pro-uPA to (r&RILRP

Competition for binding of 1251-uPA.PAI-l to immobilized ol@R/LRP TABLE I

by unlabeled compounds a,MR/LRP (about 40 fmol) was immobilized in microtiter wells and

incubated (100 pl, 16 h at 4 "C) with 10 PM '251-uPA*PAI-1 and increasing concentrations of the indicated compounds. Binding of the tracer alone ranged between 22 and 25%. The EC,, values were estimated as the

tion of '251-uPA.PAl-1 binding. When EC,, values are given, the inhibi- concentration of non-radioactive compound giving half-maximal inhibi-

tions were at least 95% when using the highest concentration of competitor. The mean values, ranges, and numbers of experiments are indicated.

Competitor EC,,

uPA*PAI- 1 0 . 4 n ~ (n=8, 0.2-0.51") Leu'"-uPA-PAI-1 LYS'~~-UPA.PAI-~

8nM (n=3, 7-8n~) 8nM (n=2,7-9n~)

tPA.PAI-1 3 n ~ (n=5,2-8n~) uPA-ATIII 3 2 n ~ (n=4, & 5 0 n ~ ) Latent PAI-1 Active PAI-1 Cleaved PAl-1 ATIII >1800nM (n= 2) uPA

pro-uPA ATF Leu'@-uPA LYS'~~-UPA

tPA Le~ '~~-pro-uPA 925n~ (n=2,900-950n~)

DFP-thrombin

5 7 n ~ (n=3,30-90n~) 5 5 n ~ (n=2,30-80n~) 4 1 n ~ ( n = 6 , 3 0 4 5 n ~ )

DFP-uPA 5 0 n ~ (n=4,30-70n~)

1901" (n=4, 100-2501") 1 2 n ~ (n=4,8-20n~)

10501" (n=5,950-1300n~) 7 2 0 n ~ (n=7,600-1000n~) 9 5 0 n ~ (n=4,800-1000n~)

1130n~ (n=3,1000-1300n~) >2000nM (n= 2)

inhibition of '251-uPA.PAI-1 via its PAI-1 moiety. Taken together, the results strongly suggest that regions in PAI-1 are involved in the binding contact.

The uPA moiety also took part in binding of the uPA.PAI-1 complex as judged from the inhibitions observed with pro-uPA, uPA, and uPA.ATIII complex (Table I). The A chain appeared to be involved since the EC,, for complexes between PAI-1 and L ~ u ' ~ ~ - u P A or LYS'~~-UPA was 8 nM as compared with 0.4 nM for PAI-1 complexed with full-length uPA, and since ATF at a high concentration could inhibit the binding of '251-uPA.PAI-1. The B chain also appeared to be involved as judged by the difference in the EC,, values between intact uPA and ATF as well as the ability of L~u'*~-uPA to compete for binding. This was further supported by the lower EC,, of L~u '~~-uPA.PAI-~ complex than the PAI-1 species alone. However, there was no indication that the linker region was involved since LYS'~~-UPA.PAI-~ and L~U'~~-LIPA.PAI-~ showed the same inhibitory potencies.

To further map the areas important for receptor interaction, we measured the effect of adding increasing concentrations of monoclonal anti-PAI-1 and anti-uPA Fab fragments. All the monoclonal antibodies tested were able to bind to the uPA.PAI-1 complex (not shown). The maximal inhibition re- flects the ability of the Fab fragment to interfere with the binding of uPA.PAI-1 to the receptor, whereas the concentration causing half-maximal inhibition is a reflection of its binding affinity. Fig. lA shows the effects of monoclonal anti-PAI-1 Fab fragments, with the epitopes of the mapped antibodies superimposed on the three-dimensional structure of a,PI (in- sets). The 80% inhibition caused by Fab fragments from hybridoma clone 6 antibody confirms the importance of the PAI-1 moiety for binding of uPA.PAI-1, but the epitope is unfortu- nately unknown. Fab fragments from the clone 3 antibody with binding epitope near the reactive center loop, and from another antibody (CLB 16, not shown) known to bind to the same aa sequence, were non-inhibitory, Fab fragments from clones 2 and 5 , having epitopes on the sides of the central P-sheet A and distal to the reactive center loop, both inhibited binding. Data

similar to those obtained with clone 2 were obtained with five antibodies (CLB 10, -15, -33, CLB 1C3, and CLB 2C8, not shown) known to bind in the same area. The results suggest that the part of PAI-1 distal to the reactive center loop makes contact with a,MFULRP.

Fig. 1B shows the results obtained with anti-uPA antibodies. The anti-uPA clone 6 Fab fragments, directed against the kringle domain, had little effect as compared with anti-CD3 fragments used as a control. Another antibody against the kringle domain, anti-uPA clone 12, caused about 50% inhibition. Fab fragments from anti-uPA hybridoma clone 2, directed against the serine proteinase domain, inhibited about 80% and thus confirmed the importance of the uPA B chain within the uPA.PAI-1 complex for contact with a,MFULRP. When taken together, the results indicate that the affinity of uPA.PAI-1 for a,MRLRP is the result of contact with binding sites in the PAI-1 moiety as well as both chains of the uPA moiety.

The Kinetics of uPA.PM-1 Binding to a$R/LRP-We have previously shown that receptor-active a2M can bind in high and low affinity modes to purified (18) as well as cellular (40) a,MW LRP. In addition, the dissociation of labeled a,M-me from the receptor is accelerated by the addition of unlabeled a,M-me. The basis for this pattern is the presence of four independent binding sites in the a2M tetramer. High affinity is due to contact with at least two sites in two receptor molecules (i.e. cross- linking of two receptors), and acceleration of dissociation by unlabeled ligand is caused by blocking of rebinding after "mi- crodissociation" from one of the sites (18). In view of these results, we analyzed the consequences of the presence of multiple binding sites in the uPA.PA1-1 complex for binding to purified as well as cellular a,MFULRP.

Fig. 2A demonstrates the concentration dependence of the EDTA and heparin-sensitive uPA.PAI-1 binding to purified a,MR&RP with the data plotted according to Scatchard shown in the inset. There were at least two binding affinity modes. Moreover, as shown in Fig. 2B, dissociation of bound ',,I- uPA.PAI-1 was markedly accelerated in the presence of excess unlabeled uPA.PAI-1 as compared with buffer alone. We next analyzed whether the different binding contacts in uPA.PAI-1 might interact with the same or with different receptor molecules. As shown in Fig. 2C, binding of '251-uPA.PAI-1 was proportional to the amount of immobilized a,MFULRP in contrast to the previously reported near-exponential relationship obtained when using the tetrameric '251-a,M-me because high receptor density favors the interaction of this ligand with two or more receptors (18).

Parallel control experiments (not shown) strongly suggested that the pattern was a characteristic of uPA.PAI-1 binding to purified a,MFULRP and not related to the technique per se. For control of the ligand, '251-uPA.PAI-1 was incubated with uPAR immobilized in microtiter wells. This binding reaction is known to occur via one region in uPA.PAI-1 (the growth factor domain of the uPAA chain) and one region in uPAR (the NH,-terminal domain). The competition curve was adequately described by assuming a single affinity (Kd = 73 PM), and dissociation (t1,, > 24 h) was not accelerated by unlabeled uPA.PAI-1 in agreement with previously published data (22). For control of the receptor preparation, immobilized a,MFULRP was incubated with rat alIB.ct complex, a homologue of the human a,M monomer containing one binding site for a,MR/LRP (18). The results were in accordance with a single affinity ( ICd = 450 PM), and unlabeled aI13.ct did not accelerate the dissociation of labeled ligand (t,,, -3 and 3.2 h, respectively) in agreement with previous results (18). In addition, binding of uPA.PAI-1 to a,MFULRP in solution, assayed by aqueous-polymer two-phase partitioning, were closely similar (EC,, about 0.4 nM) to that obtained with recep-

Binding ofuPA.PAl-1 Complex and Pro-uPA to a#fR/LRp 25671

FIG. 1. Effects of Fab fragments from monoclonal anti-PAI-1 and anti- uPA antibodies on binding of '%I- uPA.PAI-1 to purified (uJKIULRP. Mi- crotiter wells coated with 40 fmol of a,MR&RP were incubated with 10 PM '251-uPA~PAI-1 as described in the legend to Table I, with the indicated concentrations of monoclonal anti-PAI-1 (A) or anti- uPA ( B ) Fab fragments. The binding of Iz5I-uPA.PAI-1 alone (control) ranged between 22 and 25% of the added tracer. The points are mean values of triplicates 1 S.D. Panel A, left, anti-PAI-1 from hybridoma clone 3 (01, clone 2 (O), clone 5 (A), and clone 6 (A). Panel A, right, the epitopes are depicted in. black on the three-dimensional structure of reactive center-cleaved a,PI (adapted from Ref. 23). The epitopes are: anti-PAI-1 clone 3, aa 235-283 corresponding to p-sheet strands ( s ) 3B, 2C, 6A and a-helices (h) G and H; anti-PAI-1 clone 2, aa 110-145

295 (s6A, hI). The epitope for anti-PAI-1 (hE, slA, hF); anti-PAI-1 clone 5, aa 284-

from hybridoma clone 2 (0) with the clone 6 is unknown. Panel B, anti-uPA

epitope in the serine proteinase domain, and clones 6 (0) and 12 (A) with epitopes in the kringle domain. Fab fragments from a monoclonal anti-CD3 antibody (x) are included as a control.

A 120

100

80

60

40

20

0 0.01 0.1 I 10 100

Fab-fragments (pg1ml)

120 aCD3

100 -

80 -

60 -

40 -

20 - auPA cl. 2

tor immobilized in the microtiter wells. We used murine NIW3T3 cells to measure direct binding of

uPA.PAI-1 to cellular a,MR/LRP since murine uPAR is reported not to bind human uPA (2, 3). As a control, the murine cells were incubated with human '251-~PAinactivated with DFP to prevent the formation of complex with endogenous PAI-1, and binding was found not detectable. Western and ligand blot analysis of solubilized SDS-PAGE resolved and electroblotted NIW3T3 membranes (Fig. 3B) showed that uPA.PAI-1 binding occurred to a,MR/LRP. As shown in Fig. 3A, the concentration dependence of uPA.PAI-1 binding to the murine cells was closely similar to that observed with purified receptor, and unlabeled uPA.PAI-1 markedly accelerated the dissociation of the labeled species (inset). This pattern was not changed in the presence of 100 n~ human ATF (not shown), further supporting that uPA.PAI-1 bound to a,MR/LRP within the displayed concentration range.

The results are compatible with the presence of multiple independent sites as the basis for the high affinity of uPA.PAI-1 binding to purified and cellular o(,MR/LRP. The model re- sembles that reported for binding of the receptor active a,M (18). However, whereas a,M-me binds in the low and high affinity modes by interacting with either one or more a,MRnRP molecules, we propose that different regions in the uPA.PAI-1 complex make contacts with one a,MFUI,RP molecule.

01 I I I I I

0.01 0.1 1 10 7 00

Fab-fragments (pglml)

Binding ofPro-uPA-Fig. 44 shows the concentration dependence of pro-uPA binding to purified, immobilized receptor, and the inset shows accelerated dissociation of the labeled ligand in the presence of unlabeled pro-uPA compatible with at least two independent binding contacts with the receptor. As shown in Fig. 4 B , Fab fragments from the anti-uPA clone 2 antibodies were able to essentially block the binding of pro-uPA, confirm- ing the importance of the serine proteinase domain. The clone 12 antibody, directed against the kringle domain, inhibited of pro-uPA binding to an extent similar to that observed for uPA.PAI-1 binding. Interestingly, the clone 6 anti-uPA kringle antibody caused a marked inhibition of pro-uPA binding in spite of its marginal inhibition of uPA.PAI-1 binding ( c c Fig. 1B). This suggests that regions corresponding to the uPA A chain are particularly important for the affinity of pro-uPA as compared with uPA.PAI-1 complex.

Effect of Binding of uPA.P'-1 and Pro-uPA to uPAR-We first performed control experiments demonstrating that uPAR alone did not bind to immobilized a,MR/LRP (not shown) and that uPAR did not inhibit binding of '251-Le~144-~PA~PAI-1 de- void of the growth factor domain (Fig. 5). uPAR inhibited the binding of i251-uPA.PAI-l to purified a,MWLRP by about 65%, and EC,, for uPA.PAI-1 binding in the presence of a 200-fold molar excess of uPAR (i.e. uPAR:uPA.PAI-l complex) increased

25672

B

Binding of uPA-PM-1 Complex and Pro-uPA to aflRILRP A

I

35

30

- -

25 2

-

x 2 0 - 1 5 -

10

n -

5 -

~

0 250 500

Bound tpmolfll

P . ioo 10' lo2 10' 10' lo6 lo6

uPA:PAI-1 Iprnolnl

C

Buffer

uPA:PAI-1

0

0" 0 2 4 8 8 24

Hours

0 . 5 w I 0.0 0 50 100 150 200

a2MR/LRP Ifmoll

FIG. 2. Characterization of UPA-PAI-1 binding to purified m R P . Panel A, a,MWLRP (40 fmol) immobilized in microtiter wells was incubated (100 pl) for 16 h a t 4 "C with 3-10 PM 1251- uPA.PAI-1 in the absence or presence of unlabeled uPA.PAI-1. The abscissa shows the concentration of free uPA.PAI-1, and the ordinate shows the ratio of receptor bouncUfree ligand (BIF). The B/F value obtained in the presence of 200 nM uPA.PAI-1 has been subtracted from all other values. The curue represents the best least-square fit to the model BIF = Rhigh/(Kd.high + F) + R,, J(Kd.low + F), where R,, and R,, represent the apparent concentrations of receptor binding a t high affinity (K,,,,) and low affinity (Kd.low) modes, respectively. Panel A, inset, shows data plotted according to Scatchard. The calculated values are: Kd.,,. , 130 PM; Kd.low. 3.1 nM; Rhigh, 24 PM and R,,, 308 PM. Binding of 3 PM @'I-uPA.PAI-l in the presence of 10 mM EDTA (0) or 100 IU/ml heparin (A) are indicated. Panel B, dissociation at 4 "C of 1251-uPA.PAI-1 prebound to 40 fmol of immobilized a,MWLRP. BJBo is the ratio of tracer bound a t time t and a t time 0. The filled symbols indicate dissociation without, and the open symbols with 100 nM unlabeled uPA.PAI-1. Panel C, 0-180 fmol of a,MWLRP was immobilized in microtiter wells and assayed for binding of 10 PM 1251-uPA.PAI-l (0) and 1251-a,M-me (0) by incubation for 16 h a t 4 "C. The points are the mean values of triplicates with t 1 S.D. shown in panel A.

from 0.4 to about 0.8 m (Fig. 5, inset). Thus, binding of uPAR to the growth factor domain of the uPA moiety of uPA.PAI-1 causes a minor reduction in the affinity for binding of the complex to a,MFULRP.

By contrast, the binding of 1251-pro-uPA to a,MFULRP was essentially blocked by uPAR (Fig. 5). Additional experiments (not shown) demonstrated that preincubation of immobilized a,MFULRP with saturating concentrations of uPA-PAI-1, but not pro-uPA, resulted in binding of 1251-labeled uPAR. It therefore appeared that only the uPA-PAI-1 complex could bind to both receptors at the same time. This property might be nec- essary for efficient internalization via a,MFULRP since the ligands dissociate slowly from uPAR (22, and present experiments).

To test whether pro-uPA bound to cellular uPAR was protected against internalization and degradation, we used murine NIW3T3 fibroblasts transfected with human uPAR (clone NIW3T3-3), as well as simian COS-1 cells. Western blotting analysis (not shown) demonstrated a,MR/LRP expression in COS-1 cells as reported previously (411, and the expression in the NIW3T3-3 cells was as in the non-transfected cells (cfi Fig.

A

15

N

2 10 X

.I"_i_

Hours

0 100 101 lo2 lo3 lo4 lo5 io6 10'

uPA:PAI-l lprnol/ll

B '251-~PA:PAI-1 anti-a2MR/LRP

a-chain

FIG. 3. Binding of UPA-PAI-1 to QTRLRFj of murine W 3 T 3 fibroblasts. Panel A, 4.0 x lo6 NIW3T3 cells/ml were incubated for 16 h at 4 "C with 10 PM 1251-uPA.PAI-1 with or without unlabeled ligand. The points are mean values of triplicates 1 S.D. from one represent- ative experiment out of a total of three. The curue is computed according to the equation for high and low binding affinity modes as explained in the legend to Fig. 2. The calculated values for the three experiments are: Kd.high, 638 28 PM; Kd.lew. 6.3 t 1.7 m; Rhi ,40 = 4 PM; and R,,,, 336 = 86 PM. Panel A, inset, dissociation at 4 "C O ~ ~ ~ ~ I - U P A . P A I - ~ prebound to 4 x lo6 NIW3T3 celldml. The filled symbols indicate dissociation without, and the open symbols with 100 nM unlabeled uPA.PAI-1. The points are the mean values of triplicates. Panel B, left, binding of lZ5I- uPA.PAI-1 to 100 pg NIW3T3 cell membranes resolved by 6 1 6 % SDS- PAGE and electroblotted to PVDF membranes. Binding to the a-chain of 1 pg of purified a,MWLRP is also shown and, for comparison, binding to 2 pg of purified gp330. Panel B, right, Western blotting analysis of the filter shown in the left panel using a polyclonal rabbit antiserum raised against aa 2500-2922 of the a-chain of a,MR/LRP.

3B). As shown in Fig. 6B, human ATF bound to the human uPAR of the NIW3T3-3 cells and to the endogenous simian uPAR of COS-1 cells. To measure degradation of uPAR-bound ligands, 1251-pro-uPA, 1251-uPA, or 1251-uPA.PAI-1 were prebound to the two cell types at 4 "C in the presence of 200 nM RAP known to block any binding to a,MFULRP (11). The cells were washed and transferred to 37 "C in the absence or presence of 200 nM RAP, and degradation of prebound ligand to trichloroacetic acid-soluble products was measured. The minor degradation a t 37 "C in the presence of RAP was subtracted to yield the a,MFULRP-mediated degradation of uPAR-bound ligand.

Binding of uPA-PAl-I Complex and Pro-uPA to a&R ILRP 25673

A ~

15 -

N

2 1 0 - x Y m

5 - =\ 0 2 4 6 8 Hours

I 24

O B 100 10' lo2 lo3 lo4 l o 5 io6 10'

pro-uPA (pmolll)

B

100 -

80

60 8 c

m

n z 40

20

0 0.01 0.1 1 10 100

. .

Fab-fragments fpg/ml)

LRP. Panel A, lZ5I-pro-uPA (2-20 PM) with or without unlabeled pro-uPA FIG. 4. Characterization of pro-uPA binding to purified ( y 2 M R J

was incubated for 16 h a t 4 "C in microtiter wells coated with 250 fmol of a,MlULRP. The measurements have been corrected for binding in the presence of 1 PM pro-uPA. The points represent the mean values of triplicates & 1 S.D. from one of three experiments. Binding of 3 PM tracer in the presence of 10 mM EDTA (0) or 100 IU/ml heparin (A) are indicated. The curue is computed as described in the legend to Fig. 2. The calculated constants are: Kd.hlgh, 956 PM; Kd.,ow, 25.0 nM; Rhighr 40 PM; and R,,,, 2140 PM. The inset shows dissociation at 4 "C of prebound pro-uPA in the absence (filled symbols) or presence (open symbols) of 800 nM unlabeled pro-uPA. Panel B, immobilized a,MFULRP (250 fmol) was incubated with 10 PM lZ5I-pro-uPA in the presence of increasing concentrations of Fab fragments from anti-uPA hybridoma clone 2 (0) with the epitope in the serine proteinase domain, and clones 6 (0) and 12 (A) with epitopes in the kringle domain. Fab fragments from a monoclonal anti-CD3 antibody (x) are included as a control. Mean values of triplicates -c 1 S.D.

Fig. 6A shows negligible degradation of prebound pro-uPA in the NIW3T3-3 cells as compared with prebound uPA.PA1-1. Prebound uPA was degraded to a larger extent suggesting some conversion to labeled uPA.PA1-1. In COS-1 cells that produce high levels of pA1-1,~ prebound pro-uPA and uPA were degraded to the same extent as uPA.PAI-1. In contrast to the results with the NIW3T3-3 cells, more than half of the radioactivity associated with the COS-1 cells after prebinding of 1251-pro-uPA was recovered as uPA.PAI-1 when RAP was present at the 37 "C step to block the degradation of complex (Fig. 6C). The results strongly suggest that effective degradation of pro-uPA occurs only after activation, the formation of uPA.PAI-1, and binding of the complex to a,MlULRP. This conclusion was further strengthened by experiments using Fab fragments of the anti- PAI-1 hybridoma clone 6 antibody which was found not to interfere with binding of pro-uPA to a,MlULRP or the formation of uPA.PAI-1 complex. As shown in Table 11, the Fab fragments inhibited the degradation of pro-uPA in COS-1 cells by about 73%. The inhibition was the same for lZ5I-uPA and preformed

P. A. Andreasen, unpublished observation.

100

80

60

40

20

uPA:PAI-1 Ipmollll

0 ' 100 10' 1 0 2 1 o3 1 o4

uPAR (prnol/l)

FIG. 5. Effect of uPAR on binding of uPA.PAI-1 and pro-uPA to purified (u,MFf/LRP. lZ5I-labeled uPA.PAI-l(O), L~U'~~-UPA.PAI-I (A), and pro-uPA (A) were incubated (10 PM) in microtiter wells with 40 fmol (uPA.PAI-1) or 250 fmol (L~U '~~-UPA.PAI-~ and pro-uPA) of immobilized a,MR/LRP for 16 h at 4 "C in the absence (control) or presence of increasing concentrations of uPAR. The inset shows the concentration dependence of uPA.PAI-1 binding to 40 fmol of immobilized a,MRLRP in the absence ( 0 , EC,, -0.4 nM) and presence of a 200-fold molar excess of uPAR ( 0 EC,, -0.8 nM). The values are means of triplicates f 1 S.D. when indicated.

lZ5I-uPA.PAI-1 complex and corresponded closely to the inhibition of '251-uPA.PAI-1 binding to a,MWLRP shown in Fig. l A , whereas degradation of a2M-me, included as a control, was not inhibited by the antibody.

DISCUSSION

This report shows that the pattern of uPA.PAI-1 binding to a,MlULRP is compatible with the presence of multiple contacts within the uPA.PAI-1 complex for binding to a,MWLRP. Inhi- bition studies using uPA and PAI-1, truncated variants of uPA and uPA.PAI-1 complex, and Fab fragments of monoclonal antibodies provided evidence for contacts in the PAI-1 moiety, the uPA A chain, and the uPA serine proteinase domain. The in- volvement of the uPAA chain may be analogous to recent results with tPA and tPA.PAI-1 complex, indicating that a,MW LRP-mediated endocytosis does not occur with tPA mutants lacking the growth factor and finger domains (42-44).

The results of the present inhibition studies cannot be ac- counted for by impurities. For instance, the inhibitions by L ~ u ' ~ ~ - u P A and ATF were not due to contaminating uPA since the preparations contained less than 1% uPA, and the EC,, values differed only 15-20-fold. The inhibition caused by L~u '~~-uPA.PAI-~ was not due to contaminating uPA.PAI-1 since the truncated complex bound to a,MlULRP and was not inhibited by soluble uPAR. The binding of pro-uPA was not caused by contamination by uPA.PAI-1 as seen from the differ- ential effect of monoclonal antibodies (Fig. 1B uersus Fig. 4B ).

The results indicate that each of the sites providing contact has a low affinity and that a "bonus" effect on affinity (18) is obtained because multiple sites are available in the same complex. Predictions from such a multipoint attachment model are in agreement with the experimental results. Thus, the afini- ties of uPA and PAI-1 are much lower than that of the complex. Shielding or deletion of one of the multiple binding contacts in the uPA.PAI-1 complex causes a decrease in affinity. In pro- uPA, which presumably has only two binding contacts, shielding of one of them decreases the affinity to an insignificant level.

The finding that association with uPAR blocks efficient binding of pro-uPA to a,MlULRP has important implications. When 1251-pro-uPA was incubated with the NIW3T3-3 fibroblasts and COS-1 cells, we found that degradation was dependent on the

25674 Binding of uPA-PM-1 Complex and Pro-uPA to aflRILRP

NIH1313-3 cos-1

I Prebound ligand:

0 pro-uPA

NIH13T3-3 C O S 1 "

RAP + - + - uPA: PA1 -1 +

pro-uPAluPA- -I

FIG. 6. Degradation of pro-UPA in NIW3T3-3 cells and COS-1 cells depends on activation and the formation of complex with inhibitor. Panel A, NIW3T3 fibroblasts (2.0 x lofi cells/ml) transfected with the human uPAR cDNA (clone NIW3T3-3) or subconfluent simian COS-1 cells were incubated with 10 PM "'I-pro-uPA, 12sI-uPA, or 12511-uPA.PAI in the presence of 200 nM RAP for 6 h a t 4 "C. Following washing, cells with prebound ligands were incubated in a buffer with or without 200 nM a,MRAP a t 37 "C, and the amount of radioactivity released into the medium as trichloroacetic acid-soluble products was measured after 3 h. The cY,MR/LRP-mediated degradation was assessed as the difference in acid-soluble products appearing in the absence (total degradation) and the presence of 200 nM RAP after incubation a t 37 "C. The degradation in the presence of RAP a t 37 "C ranged between 4 and 6% of the prebound tracer. The columns (means of triplicate values 1 S.D. from one of three experiments) show the percent of prebound tracer degraded by the cY,MWLRP-mediated pathway. Panel B, cell membranes corresponding to 100 pg of protein from murine NIW3T3 cells (control), NIW3T3 cells transfected with human uPAR (NIW3T3-3), and simian COS-1 cells, as well as 0.25 pg of purified human uPAR, were applied to 4-16% SDS-PAGE followed by blotting to PVDF membranes, incubation with 50 PM '251-labeled human ATF (16 h a t 4 "C), and autoradiography. The arrow indicates the position of uPAR. Panel C, autoradiography of 6 1 6 % SDS-PAGE (non-reduced) of cell lysates from 3 x lofi NIW3T3-3 cells and COS-1 cells after preincubation a t 4 "C for 6 h with 10 PM 12'II-pro-uPA in the presence of 200 nM RAP followed by washing and incubation a t 37 "C for 3 h in the absence or presence of 200 n~ RAP.

formation of labeled UPAePAI-1 complex. It should be noted that cellular accumulation of '251-uPA.PAI-1 required the con- tinuous presence of RAP to inhibit the a,MR/LRP-mediated

uPA*PAI-1 by the hybridoma clone 6 anti-PAI-1 antibody degradation. This may explain the apparent contradiction be- Subconfluent COS-l cells were incubated (250 pl) in 1.9 cm2 culture tween Our data and those of Kounnas et (13). The Protection

dishes with 10 PM '*'I-labeled pro-uPA, uPA, uPA.PAI-1 or a,M-me for 3 of the zymogen from degradation should generate a pool of h a t 37 "c with or without 25 Pdml Fab fragments from the hybridoma upm-bound pro-UpA at cell surfaces before the initiation ofthe clone 6 anti-PAI-1 antibody or from an anti-CD3 antibody (control), and with or without 200 nM RAP. Degradation to trichloroactic acid-soluble cascade starting with activation Of the 'Fogen* products in the presence of RAP (3.0-3.5% of the added tracer) was Fig. 7 shows a possible model for the events at the surface of subtracted to yield the a2MR/LRP mediated degradation (40.142.6% cells expressing both uPAR and a,MR/LRP. The secreted zymo- for all tracers). The a,MWLRP-mediated degradation in the absence of gen pro-up~ may bind to up^ or to a , ~ ~ ~ p . men Fab fragments is set at loo%, and the results are the means of triplicate values r 1 S.D. cupied uPAR is available, the first alternative is by far the most

likely due to the high affinity for binding to cellular uPAR with Unlabeled Labeled ligand a Kd determined previously a t about 50 PM (9, 22, 45). Direct competitor '251-Pro-UpA "'I-UPA '2sI-upA4'AI-I '2SI-a,M-me binding of pro-uPA to a,MR/LRP (inset A ) may function as a

No addition 100.0 * 0.5 100.0 11.0 100.0 * 5.1 100.0 * 12.0 backup system for clearance when uPAR is saturated, e.g. as Anti-PM-1 Clone 6 26.4 2 6.0 29.9 f 3.6 26.9 f 8.3 101.8 f 10.1 the result of pharmacological infusions. m e n bound to u p m , Anti-CD3 103.2 * 6'1 'OL8 * 2'4 95'6 * 5'0 96'0 * pro-uPA does not interact with a,MWLRP due to shielding of

the binding contact corresponding to the A chain. uPAR-bound pro-uPA is subsequently activated by plasmin to the catalyti-

TABLE I1 Inhibition of aJMRILRP mediated degradation of pro-uPA, uPA and

Binding of uPA.PM-1 Complex and Pro-uPA to a$lR/LRP 25675

FIG. 7. A possible model for the interaction between the uPA system and c#R/LRP. Regions of uPA.PAI-1 involved m binding to a,MFULRP are sym- bolized by a square (PAI-l), a triangle (uPA B chain), and a semicircle (uPA A chain). A filled semicircle indicates capa-

EC50 : 0.4 nM

.,. . ..: : a2 MFULRP n

bility for interaction with a,MFULRP, and an open semicircle indicates a shielding for this interaction. The conversion of pro- uPA to uPA is indicated by an indentation in the uPA B chain. The regions in a,MW LRP containing 2, 8, 10, and 11 complement-type repeats (31) are indicated by

YWTD repeats by dotted lines. The appar- open ellipses, and EGF repeats and

ent affinities (EC,, values) are indicated for the binding steps at the cell surface, and for direct binding of pro-uPA (inset A and uPA.PAI-1 complex (inset B ) to the region in a,MFULRP containing eight complement-type repeats. For further explanation, see text.

PAI-1

pro-uPA

uPAR

GPI-anchor

* ECm : 50 pM **ECm : 0.8 nM

cally active uPA, which in turn is inactivated by the formation of complex with PAI-1 (Fig. 7). The tight binding to uPAR is maintained since pro-uPA, uPA, and uPA.PAI-1 bind to cellular uPAR with similar affinities (5,9,22,45) in agreement with the presently determined Kd of 73 PM for binding of uPA.PAI-1 to immobilized recombinant uPAR. Finally, uPA.PAI-1 associated with uPAR binds within the region of a,MR/LRP containing eight complement-type repeats (31) and is internalized by the endocytic receptor.

Alternatively, pro-uPA may be activated in solution, and the resulting uPA appears to be cleared slowly due to its poor binding to a,MFULRP. PAI-1, which is secreted by several cell types (61, is also likely to be cleared slowly by (r,MFULRP due to its low affinity, and binding to vitronectin (46) may provide further protection. uPA.PAI-1 complexes are therefore readily formed, they will bind to a,MR/LRP with high affinity due to the full bonus effect of the multiple binding contacts (Fig. 7, inset B ) , and clearance will proceed rapidly.

In conclusion, the multipoint attachment model for binding to a,MR/LRP provides a simple and plausible explanation for the selective clearance of complexes between uPA and PAI-1 relative to its uncomplexed counterparts and for the lack of degradation of pro-uPA when bound to uPAR.

Acknowledgments-We thank S. Andersen, T. M~ller, and L. Nielsen for excellent technical assistance. C. Munck Petersen is thanked for valuable discussions, H. R~igaard for culturing cells, and K. E. J~rgensen for making computer drawings. S. K. Moestrup, H. C. Th~gersen, K. Dolmer, L. Sottrup-Jensen, J . Henkin, and W. Gunzler are gratefully acknowledged for providing us with reagents.

REFERENCES 1. Dane, K., Andreasen, P. A,, Grendahl-Hansen, J., Kristensen, P., Nielsen, L. S.,

2. Vassalli, J. D., Sappino, A. P., and Belin, D. (1991) J. Clin. Inuest. 88, 1067-

3. Fazioli, F., and Blasi, F. (1994) Dends Pharmacol. Sei. 15, 25-29 4. Pollanen, J., Stephens, R. W., and Vaheri, A. (1991) Adu. Cancer Res. 57,

5. Cubellis, M. V., Andreasen, P. A,, Ragno, P., Mayer, M., Dane, K., and Blasi, F.

6. Andreasen, P. A,, Georg, B., Lund, L. R., Riccio, A., and Stacey, S. (1990) Mol.

7. Cubellis, M. V., Wun, T. C., and Blasi, F. (1990) EMBO J. 9, 1079-1085 8. Jensen, P. H., Christensen, E. I., Ebbesen, P., Gliemann, J., and Andreasen, P.

9. Nykjaer, A,, Petersen, C. M., Meller, B., Andreasen, P. A,, and Gliemann, J. A. (1990) Cell Reg. 1, 1043-1056

(1992) FEBS Lett. 300, 13-17 10. Olson, D., Pollanen, J., Heyer-Hansen, G., Ronne, E., Sakaguchi, K., Wun, T.

C.,Appella, E., Dane, K., and Blasi, F. (1992) J. Bid. Chem. 267,9129-9133

and Skriver, L. (1985) Adu. Cancer Res. 44, 139-266

1072

273-328

(1989) Proc. Natl. Acad. Sci. U. S. A. 86, 4828-4832

Cell. Endocrinol. 68, 1-19.

11. Nykjer, A,, Petersen, C. M., Meller, B., Jensen, P. H., Moestrup, S. K., Holtet, T. L., Etzerodt, M., Thegersen, H. C., Munch, M., Andreasen, P. A., and Gliemann, J. (1992) J. Biol. Chem. 267, 14543-14546

12. Herz, J., Clouthier, D. E., and Hammer, R. E. (1992) Cell 71,411421 13. Kounnas, M., Henkin, J., Argraves, W. S., and Strickland, D. (1993) J. Bid.

Chem. 268,21862-21867 14. Gliemann, J., Nykjaer, A,, Petersen, C. M., Jergensen, K. E., Nielsen, M.,

Andreasen, P. A,, Christensen, E. I., Lookene, A., Bengtsson-Olivecmna, G.,

15. Mottonen, J., Strand,A., Symersky, J., Sweet, R. M., Danley, D. E., Geoghegan, and Moestrup, S. K. (1994) Ann. N. Y Acad. Sci. 737, 20-39

16. Lobennann, H., Tokuoka, R., Deisenhofer, J., and H u h , R. (198415. Mol. Biol. K. F., Gerard, R. D., and Goldsmith, E. J. (1991) Nature 366, 270-273

17. Behrendt, N., Ploug, M., Patthy, L., Houen, G., Blasi, E, and Dane, K. (1991) 177,531-556

J. B i d . Chem. 266, 7842-7847

19. Moestrup, S. K., Nielsen, S., Andreasen, P. A,, Jergensen, K. E., Nykjaer, A,, 18. Moestrup, S. IC, and Gliemann, J. (1991) J. Biol. Chem. 266, 14011-14017

RBigaard, H., Gliemann, J., and Christensen, E. I. (1993) J. Biol. Chem.

20. Bianchi, E., Cohen R. L., Thor, A. T., Todd, R. F., 111, Mizukami, I. F., Lawrence, 268, 16564-16570

D. A,, Ljung, B. M., Shuman, M. A,, and Smith, H. S. (1994) Cancer Res. 54, 861466

21. Todd, R. F., 111, Barnatban, E. S., Bohuslav, J., Chapman, H. A,, Cohen, R. L., Felez, J., Howell, A., Knapp, W., Kramer, M., Miles, L. A,, Nykjaer, A,,

Differentiation Antigens (Schlossman, S. F., Boumsell, L., Gilks, W., Harlan, Ralfkiaer, E., and Schuren E. (1994) in Leukocyte Typing V: White Cell

J. M., Kishimoto, T., Morimoto, C., Ritz, J., Shaw, S., Silverstein, R. L., Springer, T. A,, Tedder, T. F., and Todd, R. F., 111, eds) Oxford University

22. Nykjer, A., Petersen, C. M., Christensen, E. I., Davidsen, O., and Gliemann, J. Press, Oxford, in press

23. Munch, M., Heegaard, C. W., Jensen, P. H., and Andreasen, P. A. (1991) FEBS (1990) Biochim. Biophys. Acta 1062, 399-407

24. Munch, M., Heegaard, C. W., and Andreasen, P. A. (1993) Biochim. Biophys. Lett. 281, 181-184

25. Sottrup-Jensen, L., Sand, O., Kristensen, L., and Georgh, H. F. (1989) J. Bid. Acta 1202.29-37

26. Grendahl-Hansen, J., Ralkiaer, Nielsen, L. S., Kristensen, P., Frentz, G., and Chem. 264, 15781-15789

27. Pollanen, J., Saksela, O., Salonen, E."., Andreasen, P., Nielsen, L., Dane, K., Dane K. (1987) J. Znuest. Dermatol. 88, 2&32

28. Stephens, R. W., Bokman, A. M., Myohinen, H. T., Reisberg, T., Tapiovaara, H., and Vaheri, A. (1987) J. Cell B i d . 104,1085-1096

Pedersen, N., Grendahl-Hansen, J., Llinas, M., and Vaheri, A. (1992) Bio- chemistry 31, 7572-7579

29. Keijer, J., Linders, M., van Zonneveld, A,-J., Ehrlich, H. J., de Boer, J. P., and Pannekoek, H. (1991) Blood 78, 401409

30. Johnstone, A., and Thorpe, R. (1987) in Immunochemistry in Practice, pp. 58-60, Blackwell Scientific Publications, Oxford

31. Moestrup, S. K., Holtet, T. L., Etzerodt, M., Thegersen, H. C., Nykjaer, A,, Andreasen, P. A,, Rasmusaen, H. H., Sottrup-Jensen, L., and Gliemann, J. (1993) J. Biol. Chem. 268, 13691-13696

32. Huber, R., and Carrell, R. W. (1989) Biochemistry 28,89514966 33. Andreasen, P. A., Nielsen, L. S., Kristensen, P., Grendahl-Hansen, J., Skriver,

L., and Dane, K. (1986) J. Bid. Chem. 261, 7644-7651 34. Min, H. Y., Semnani, R., Mizukami, I. F., Watt, K., Todd, R. E, 111, and Liu, D.

Y. (1992) J. Zmmunol. 148,363&3642 35. Mizukami, I. F., Garni-Wagner, B. A,, DeAngelo, L. M., Liebert M., Flint, A.,

Lawrence, D. A., Cohen, R. L., and Todd, R. F., I11 (1994) Clin. Immunol.

36. Nykjer, A,, Bengtsson-Olivecrona, G., Lookene, A., Moestrup, S. K., Petersen, Immunopathol. 71, 96-104

25676 Binding of uPA.PAI-1 Complex and Pro-uPA to a&RILRP C. M., Weber, W., Beisigel, U., and Gliemann, J. (1993) J. Biol. Chem. 268, 15048-15055

37. Backman, L. (1985) in Partitioning in Aqueous Ituo-Phase Systems: Theory, Methods, Uses, andApplications to Biotechnology (Walter, H., Brooks, D. E.,

38. Bu, G., Williams, S., Strickland, D. K., and Schwartz, A. L. (1992) Proc. Natl. and Fisher, D., eds) pp. 267-314, Academic Press, Orlando

39. Nykjaer, A,, Kjeller, L., Cohen, R., Lawrence, D. A., Gliemann, J., and An- Acad. Sci. U. S. A. 89, 7427-7431

40. Gliemann, J., and Davidsen, 0. (1986) Biochim. Biophys. Acta 885, 49-57 dreasen, P. A. (1994) Ann. N . Y. Acad. Sei. 737,483-486

41. Orth, K., Madison, E. L., Gething, M. J., Sambrook, J. F., and Herz, J. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 7422-7426

42. Bassel-Duby, R., Jiang, N. Y., Bittick, T. Madison, E. McGookey, D., Orth, K., Shohet, R., Samhrook, J., and Gething, M.-J. (1992) J. Biol. Chem. 267,

43. Grobmyer, S. R., Kuo, A., Orishimo, M., Okada, S. S., Cines, D. B., and Bar-

44. Camani, C., Bachmann, F., and Kruithof, E. K. 0. (1994) J. Biol. Chem. 269, nathan, E. S. (1993) J. Biol. Chem. 268,13291-13300

45. Nykjrer, A,, Maller, B., Todd 111, R. F., Christensen, T., Andreasen, P. A., Gli- 5770-5775

46. Salonen, E."., Vaheri, A,, Pollanen, J., Stephens, R.,Andreasen, P. A., Mayer, emann J., and Munck Petersen, C. (1994) J. Zmmunol. 152, 505-516

M., Dana, K., Gailit, J., and Rouslahti, E. (1989) J. Biol. Chem. 264,6339- 6343

9668-9677

regions involved in binding of urokinase-type-1 inhibitor complex

Documents