Increase in urokinase plasminogen activator mRNA synthesis in human carcinoma cells is a primary effect of the potent tumor promoter, phorbol myristate acetate.

The effect of tumor promoters and growth factors on the synthesis of urokinase and urokinase mRNA in human carcinoma cells has been investigated. In urokinase-producing human carcinoma cells (A1251), a 20-40-fold increase in urokinase mRNA level is obtained after treatment with 10 nM phorbol myristate acetate (PMA), a smaller effect (two- to fourfold) with 2 ng/ml platelet-derived growth factor (PDGF) and no effect with epidermal growth factor (EGF) (up to 50 nM). After treatment with PMA, urokinase mRNA level increases already at 30 min peaking 2-4 h thereafter. Cell line A431, which has an abnormally high number of EGF receptors, shows the same response to PMA, but also responds to EGF (two- to fourfold increase in mRNA). The kinetics are similar to those of A1251. Nuclear transcription experiments show that the PMA-induced increase in urokinase mRNA is due to increased synthesis. The protein synthesis inhibitor, cycloheximide (10 micrograms/ml), also increases the level of urokinase mRNA. When both cycloheximide and PMA are used, super-induction is observed. This result may indicate that a short-lived protein negatively regulates the level of urokinase. The different efficiency of the effectors (PMA and PDGF better than EGF) and their kinetics, as well as the effect of cycloheximide on urokinase mRNA synthesis, (a) are reminiscent of the effect of PDGF and PMA on competence phase genes (Kelly, K., B.H. Cochran, C.D. Stiles, and P. Leder, 1983, Cell, 35: 603-610), (b) demonstrate that the synthesis of urokinase is part of the early cellular response to these factors, and (c) provide a preliminary insight in the overproduction of urokinase by primary malignant tumors and transformed cells in culture.

An insulin epidermal growth factor-binding protein from Drosophila has insulin-degrading activity.

We have recently described the purification and characterization of an insulin-degrading enzyme (IDE) from Drosophila melanogaster that can cleave porcine insulin, is highly conserved through evolution and is developmentally regulated. We now report that the IDE is, in fact, an insulin EGF-binding protein (dp100) that we had isolated previously from Drosophila using an antihuman EGF receptor antiserum. This conclusion is based upon the following evidence. (a) dp100, identified by its ability to cross-link to labeled insulin, EGF, and transforming growth factor-alpha (TGF-alpha), and to be immunoprecipitated by anti-EGF receptor antisera, copurifies with the IDE activity. Thus, the purified IDE can be affinity labeled with either 125I-insulin, 125I-EGF, or 125I-TGF-alpha, and this labeling is specifically inhibited with unlabeled insulin, EGF, and the insulin B chain. (b) The antiserum to the human EGF receptor, which recognizes dp100, is able to specifically immunoprecipitate the insulin-degrading activity. (c) The purified IDE preparation contains a single protein of 110 kD which is recognized by both the anti-EGF receptor antiserum and anti-Drosophila IDE antiserum. (d) Polyclonal antiserum to the purified IDE, which specifically recognized only the 110-kD band in Drosophila Kc cells, immunoprecipitates dp100 cross-linked to 125I-TGF-alpha and dp100 cross-linked to 125I-insulin from the purified IDE preparation. (e) EGF, which competes with insulin for binding to dp100, also inhibits the degradation of insulin by the purified IDE. These results raise the possibility that a functional interaction between the insulin and EGF growth factor families can occur which is mediated by the insulin-degrading enzyme.

Characterization of a Drosophila protein that binds both epidermal growth factor and insulin-related growth factors.

The identification of a novel protein from Drosophila melanogaster that binds both mammalian epidermal growth factor (EGF) and insulin has been reported (Thompson, K. L., S. J. Decker, and M. R. Rosner, 1985, Proc. Natl. Acad. Sci. USA., 82:8443-8447). This 100-kD protein (designated dp100) is also recognized by an antiserum against the human EGF receptor. To further characterize the properties of this protein, we have determined the binding spectrum, glycosylation state, and cellular distribution of dp100. Our results indicate that dp100 binds to other insulin-like and EGF-like growth factors with dissociation constants ranging from 10(-6) to 10(-9) M, and these ligands compete with each other for binding to dp100. All other ligands tested, including platelet-derived growth factor, transforming growth factor-beta, nerve growth factor, and glucagon, either did not bind or bound with a Kd greater than 10(-6) M. Unlike the Drosophila insulin receptor, dp100 does not bind to wheat germ agglutinin and is present in a cytoplasmic as well as a membrane-bound form that cannot be differentiated by two-dimensional PAGE. Further, dp100 is the sole transforming growth factor-alpha-binding protein detected by affinity labeling in Drosophila Kc cells. These results indicate that dp100 shares properties in common with, but distinct from, the Drosophila homologues of the insulin and EGF receptors.

Phosphorylation of human pro-urokinase on Ser138/303 impairs its receptor-dependent ability to promote myelomonocytic adherence and motility.

Serine phosphorylation of human pro-urokinase (pro-uPA) by A431 human carcinoma cells results in a catalytically active molecule with reduced sensitivity to plasminogen activator inhibitor type 1. We mapped the phosphorylated seryl residues by analyzing the in vivo phosphorylation state of engineered pro-uPA variants carrying a COOH-terminal poly-histidine tag. Stably transfected A431 cells do not incorporate radioactive phosphate into tagged pro-uPA in which the serines 138 and 303 have been replaced with glutamic residues, although endogenous nontagged pro-uPA is 32P-labeled on A and B chains. Moreover, the catalytic-independent ability of the mono- and di-substituted "phosphorylation-like" variants to bind to the GPI-anchored urokinase receptor (uPAR) and promote adherence of differentiating U937, HL-60, and THP-1 myelomonocytic cells was examined. We found that glutamic residues as well as the naturally occurring phosphoserines at positions 138 and 303 abolish proadhesive ability, although they do not interfere with receptor binding. In addition, pro-uPA carrying Glu138/303 lacks the capability to induce a chemotactic response of THP-1 cells. The exclusive presence of Glu138 reduces pro-uPA proadhesive and chemotactic ability by 70-80%, indicating that a phosphoserine residue at the same position plays a major inhibitory role of myeloid cell response to pro-urokinase. The di-substitution does not affect pro-uPA ability to interact with vitronectin or to enhance binding of urea-denatured vitronectin to uPAR. However, unlike wild-type tagged pro-uPA, the di-substituted variant does not induce receptor polarization in pre-adherent U937 cells. Taken together, the data support the possibility that pro-uPA phosphorylation on Ser138/303 can modulate uPAR transducing ability.

Regulation of urokinase receptors in monocytelike U937 cells by phorbol ester phorbol myristate acetate.

A specific surface receptor for urokinase plasminogen activator (uPA) recognizes the amino-terminal growth factor-like sequence of uPA, a region independent from and not required for the catalytic activity of this enzyme. The properties of the uPA receptor (uPAR) and the localization and distribution of uPA in tumor cells and tissues suggest that the uPA/uPAR interaction may be important in regulating extracellular proteolysis-dependent processes (e.g., invasion, tissue destruction). Phorbol myristate acetate (PMA), an inducer of U937 cell differentiation to macrophage-like cells, elicits a time- and concentration-dependent increase in the number of uPAR molecules as shown by binding, cross-linking, and immunoprecipitation studies. The effect of PMA is blocked by cycloheximide. Overall, the data indicate that PMA increases the synthesis of uPA. PMA treatment also causes a decrease in the affinity of the uPAR for uPA, thus uncovering another way of regulating the interaction between uPA and uPAR. In addition, the PMA treatment causes a modification of migration of the cross-linked receptor in mono- and bidimensional gel electrophoresis.

Upregulation of epidermal growth factor receptor induced by alpha-interferon in human epidermoid cancer cells.

Unregulated or increased expression of epidermal growth factor receptor (EGF-R) is a common event in neoplastic transformation; modulation of such a receptor by physiological agents could be, therefore, of clinical interest. We have studied the binding ability, the availability at cell surface, and the synthesis of EGF-R in the A431 and KB human epidermoid cancer cell lines after treatment with recombinant alpha-interferon (IFN-alpha). After 48 h of treatment, IFN-alpha induces, in both cell lines, growth inhibition and enhances class I major histocompatibility HLA complex expression, which is a common marker of IFN action. [125I]EGF total binding assessed after 48 h of treatment with IFN-alpha shows a dose-dependent upregulation of EGF-R binding capacity. Saturation plots of the binding data show that IFN-alpha treatment does not dramatically alter the affinity of the EGF-R and indicate that IFN-alpha only increases the number of low affinity receptors. We show that this effect is due to a specific increase in the synthesis of the receptor protein, as assessed by immunoprecipitation of [35S]methionine-labeled cell extracts. Electron microscopy analysis has confirmed an increase of EGF-R proteins at cell surface without major changes in the morphology of the cells. Taken together, these results indicate that IFN-alpha consistently induces both the binding capacity and the synthesis of EGF-R in human epidermoid cancer cells and suggest the use of such a mechanism for new anticancer therapies.

Human urokinase receptor concentration in malignant and benign breast tumors by in vitro quantitative autoradiography: comparison with urokinase levels.

We measured the tissue concentration of human urokinase receptor (uPAR) in 22 breast carcinomas and 9 benign breast lesions using in vitro quantitative autoradiography. Tissue sections were incubated with increasing concentrations of 125I-pro-urokinase in the presence or absence of unlabeled competitor. Breast carcinomas were found to contain 5 times more uPAR than benign breast lesions with respect to their protein content [523 +/- 72 versus 108 +/- 20 (SE) fmol/mg (P < 0.001)]. Simultaneous quantitation of urokinase (uPA) by immunoenzymatic assay on tissue extracts from the same specimens showed that breast carcinomas also contain 19 times more uPA than benign tumors (611 +/- 134 versus 32 +/- 8 fmol/mg) (P < 0.01). The reliability of quantitative autoradiography measurements was confirmed by uPAR cross-linking assay on membrane fraction from either U937 histiocytic lymphoma cells or breast carcinomas and immunoperoxidase staining with an anti-uPAR antibody on tumor sections. Also, immunoperoxidase staining with an anti-uPA monoclonal antibody showed that uPA is indeed localized on the plasma membrane of epithelial tumor cells in confined areas of breast carcinomas whereas cells from benign breast lesions were devoid of uPA under the same experimental conditions. In conclusion, our findings support the hypothesis that uPAR plays a central role in the acquisition of an invasive phenotye and favor its potential use as a prognostic factor in patients with breast carcinoma.

Tissue distribution of soluble and receptor-bound urokinase in human breast cancer using a panel of monoclonal antibodies.

Current evidence regarding the regulation of urokinase-dependent extracellular proteolysis indicates that plasminogen activation is a surface-associated process. We have compared the histological localization of urokinase plasminogen activator (uPA) and urokinase plasminogen activator receptor (uPAR) in breast cancer sections using a panel of monoclonal antibodies. First, the ability of six different anti-uPA monoclonal antibodies to recognize pro-uPA, uPA, and in vitro-formed complexes of uPA with either soluble uPAR or with plasminogen activator inhibitor type 1 was compared. Then the reactivity of the anti-uPAR antibodies was tested, and the occurrence of an uPA receptor of about M(r) 55,000 in samples from breast carcinoma was assessed by immunoprecipitating the uPA receptor from an in vitro 125I-labeled tumor extract. Immunocytochemical data from adjacent sections of 10 tumor specimens showed that antibodies recognizing free and bound uPA mostly stain the cytoplasm and the membrane of epithelial tumor cells in confined areas of the tumor and some fibroblast-like stromal cells. Acid pretreatment of tumor sections, which removes receptor-bound uPA, causes a strong reduction of the immunocytochemical reactivity of epithelial tumor cells, whereas staining of fibroblast-like cells is not considerably affected. Consistent with these results, epithelial tumor cells were mostly unreactive to anti-uPAR antibodies unless pretreated with acidic buffer, whereas fibroblast-like stromal cells showed a faint but acid-resistant staining with all anti-uPARs. In conclusion, these results show that occupied uPA receptors are definitely present on the membrane of epithelial tumor cells and suggest the occurrence of uPA-uPAR-dependent proteolytic activity on the surface of breast cancer cells.

Coordinate up-regulation of Sp1 DNA-binding activity and urokinase receptor expression in breast carcinoma.

The regulatory mechanisms underlying the overexpression of urokinase-type plasminogen activator (uPA) and its receptor (uPAR) in highly invasive breast carcinomas remain poorly understood. In this study, we have simultaneously determined the level of uPAR and the activity of the transcription factor Sp1 in 14 breast carcinomas and 5 benign lesions. We found that uPAR levels and Sp1-binding activity are coordinately elevated in malignant tumors (r, 0.94; P < 0.001). On the contrary, undetectable or only barely detectable levels of uPAR and Sp1 activity were found in benign breast lesions. Finally, the engagement of uPAR by catalytically inactive uPA in the MDA-MB-231 breast carcinoma cell line results in a rapid up-regulation of Sp1-binding activity followed by an increase of uPAR protein. These results, taken together, suggest the existence of a uPA-dependent positive regulatory loop that may progressively enhance malignant breast cell invasiveness.

Urokinase receptor interacts with alpha(v)beta5 vitronectin receptor, promoting urokinase-dependent cell migration in breast cancer.

Perturbation of adhesive interactions at cell-substratum and cell-cell contact sites is a critical event in the multistep process of cancer invasion. Recent studies indicate that the urokinase receptor (uPAR) is associated in large molecular complexes with other molecules, such as integrins. To test the possibility that uPAR may physically and functionally interact with vitronectin (Vn) receptors, we determined the expression level of uPAR, alpha(v)beta3, and alpha(v)beta5 Vn receptors in 10 human breast carcinomas. Here, we show the ability of uPAR to physically associate with alpha(v)beta5 in the breast carcinomas examined. The functional effects of this interaction were studied using HT1080 human fibrosarcoma and MCF-7 human breast carcinoma cell lines, both exhibiting a urokinase-dependent physical association between uPAR and alpha(v)beta5. Both cell lines respond to urokinase or to its noncatalytic amino-terminal fragment by exhibiting remarkable cytoskeletal rearrangements that are mediated by alpha(v)beta5 and require protein kinase C activity. On the contrary, binding of Vn to alpha(v)beta5 results in the protein kinase C-independent formation of F-actin containing microspike-type structures. Furthermore, alpha(v)beta5 is required for urokinase-directed, receptor-dependent MCF-7 and HT1080 cell migration. These data show that uPAR association with alpha(v)beta5 leads to a functional interaction of these receptors and suggest that uPAR directs cytoskeletal rearrangements and cell migration by altering alpha(v)beta5 signaling specificity.

Vitronectin binding to urokinase receptor in human breast cancer.

Functional assembly of the plasminogen-dependent proteolytic system on the cell surface requires multiple interactions involving urokinase (uPA), urokinase receptor (uPAR), plasminogen activator inhibitors, and other molecules that mediate cell migration and adhesion. We analyzed the in vitro interaction of uPAR-containing particulate cell fractions with the amino-terminal fragment (ATF) of human urokinase and the matrix-like form of vitronectin. Binding and cross-linking of 125I-labeled ATF to crude membrane extracts from LB6-19 mouse cells overexpressing human uPARs in the presence of 25 nM urea-denatured vitronectin led to the formation of Mr 137,000, 92, 000, and 82,000 covalent complexes. Immunoprecipitation of the preformed cross-linked 125I-labeled complexes with anti-vitronectin, anti-uPA, or anti-uPAR antibodies revealed that the Mr 82,000 and 92, 000 species do contain ATF and vitronectin and identified the Mr 137, 000 species as a ternary complex formed by ATF, uPAR, and vitronectin. A similar electrophoretic pattern was displayed by acid-pretreated membranes extracted from MCF-7 breast carcinoma or HT1080 fibrosarcoma cell lines, as well as a ductal breast carcinoma specimen; the latter exhibited complex formation at concentrations of vitronectin lower than 10 nM. Finally, uPAR-vitronectin interaction was further documented by the decreased reactivity of an anti-uPAR polyclonal antibody to acid-pretreated sections of 10 breast carcinomas that had been preincubated with vitronectin. Our findings highlight the ability of uPAR to interact simultaneously with vitronectin and uPA in breast cancer, supporting a dynamic coupling of the molecular mechanisms underlying plasminogen-dependent matrix degradation and cell adhesion.

Serine phosphorylation of biosynthetic pro-urokinase from human tumor cells.

Phosphorylation is a potent mechanism regulating the activity of many intracellular enzymes. We have discovered that the product of the human urokinase plasminogen activator gene, pro-uPA, is phosphorylated in serine in at least two human cell lines. Phosphorylation occurs within the cell during biosynthesis, and phosphorylated intracellular pro-uPA is secreted into the medium. Of the secreted pro-uPA molecules, 20-50% are phosphorylated in serine, thus representing a meaningful fraction of the total biosynthetic pro-uPA. Although the sites of phosphorylation have not yet been determined, at least two such sites must exist; in fact plasmin cleavage of phosphorylated single chain pro-uPA yields a two chain uPA in which both chains are phosphorylated. A specific function for pro-uPA phosphorylation has not yet been identified; however, it is tempting to speculate that, as in many other cases, phosphorylation may affect the activity of the enzyme, its response to inhibitors or the conversion of pro-uPA zymogen to active two-chain uPA. This would represent an additional way of regulating extracellular proteolysis, an important pathway involved in both intra- and extravascular phenomena like fibrinolysis, cell migration and invasiveness.

In vitro receptor imaging for characterization of human solid tumors.

Since many of the factors involved in tumor growth and progression act through a receptor-mediated mechanism, we applied in vitro receptor imaging techniques to study the intratumoral distribution and concentration of receptor-molecules having experimental biological relevance in such processes. Here we summarize the results of a study concerning the role of urokinase receptor (uPAR) in the acquisition of an invasive phenotype by tumor cells. Independently of the system studied, we demonstrated that in vitro receptor imaging techniques can be used to define the biological characteristics of human solid tumors and can contribute to clarify the complex network of ligand/receptor interactions leading to tumor spread.

The soluble form of urokinase receptor promotes angiogenesis through its Ser88-Arg-Ser-Arg-Tyr9² chemotactic sequence.

BACKGROUND: The urokinase plasminogen activator receptor (u-PAR) focuses the proteolytic activity of the urokinase plasminogen activator (u-PA) on the endothelial cell surface, thus promoting angiogenesis in a protease-dependent manner. The u-PAR may exist in a glycophosphatidylinositol-anchored and in a soluble form (soluble u-PAR [Su-PAR]), both including the chemotactic Ser88 -Arg-Ser-Arg-Tyr9² internal sequence. OBJECTIVE: To investigate whether Su-PAR may trigger endothelial cell signaling leading to new vessel formation through its chemotactic Ser88 -Arg-Ser-Arg-Tyr9² sequence. METHODS AND RESULTS: In this study, the formation of vascular-like structures by human umbilical vein endothelial cells was assessed by using a matrigel basement membrane preparation. First, we found that Su-PAR protein promotes the formation of cord-like structures, and that this ability is retained by the isolated Ser(88) -Arg-Ser-Arg-Tyr9² chemotactic sequence, the maximal effect being reached at 10 nmol L⁻¹ SRSRY peptide (SRSRY). This effect is mediated by the α(v) ß3 vitronectin receptor, is independent of u-PA proteolytic activity, and involves the internalization of the G-protein-coupled formyl-peptide receptor in endothelial cells. Furthermore, exposure of human saphenous vein rings to Su-PAR or SRSRY leads to a remarkable degree of sprouting. Finally, we show that Su-PAR and SRSRY promote a marked response in angioreactors implanted into the dorsal flank of nude mice, retaining 91% and 66%, respectively, of the angiogenic response generated by a mixture of vascular endothelial growth factor and fibroblast growth factor type 2. CONCLUSIONS: Our results show a new protease-independent activity of Su-PAR that stimulates in vivo angiogenesis through its Ser88 -Arg-Ser-Arg-Tyr9² chemotactic sequence.

The receptor for urokinase-plasminogen activator.

Many human cells and cell lines possess a specific receptor that binds urokinase plasminogen activator (uPA) with an affinity of about 10(-10) M. Bound enzyme is not internalized, is slowly dissociated, and retains its enzymatic activity. The amino acid sequence of uPA responsible for receptor binding is located within the first 35 aminoterminal residues, ie, in the growth factor domain. Binding, however, is not competed for by other proteins that contain the growth factor domain (including epidermal growth factor). Cells that produce uPA secrete the pro-uPA form, which subsequently binds to the receptor. A431 cells, in fact, have their receptors completely saturated with pro-uPA. It is proposed that uPA:uPA-receptor interaction plays a direct role in physiological and pathological processes that require cell migration.

Developmental regulation of an insulin-degrading enzyme from Drosophila melanogaster.

The precise mechanism by which insulin is degraded in mammalian cells is not presently known. Several lines of evidence suggest that degradation is initiated by a specific nonlysosomal insulin-degrading enzyme (IDE). The potential importance of this insulin protease is illustrated by the fact that there is an IDE in Drosophila melanogaster Kc cells that shares both physical and kinetic properties with its mammalian counterpart. We now demonstrate that the IDE is present in other Drosophila cell lines and in the embryo, the larvae, the pupae, and adult tissues of the fruit fly. Further, the level of the IDE is developmentally regulated, being barely detectable in the embryo but elevated approximately 5-fold in the larvae and pupae and approximately 10-fold in the adult fly. The IDE levels in the cell lines are particularly high, at least 10-fold greater than in the adult fly. Analysis of Schneider L3 cells indicates that the addition of the Drosophila hormone ecdysone, which induces differentiation of the cells, causes a small but reproducible increase in the level of the IDE and the insulin-degrading activity. These results demonstrate that the IDE is evolutionarily conserved and that its expression is tightly regulated during differentiation of Drosophila. The particular pattern of developmental regulation suggests that the IDE plays a specific and critical role in the later stages of the life cycle of the fly.

Differentiation-enhanced binding of the amino-terminal fragment of human urokinase plasminogen activator to a specific receptor on U937 monocytes.

The purified amino-terminal fragment (ATF) of human urokinase plasminogen activator (residues 1-135), which is not required for activation of plasminogen, binds with high affinity to specific plasma membrane receptors on U937 monocytes. Intact urokinase efficiently competes for 125I-labeled ATF binding; 50% competition occurs with 1 nM urokinase. A large part of receptor-bound urokinase remains on the cell surface for at least 2 hr at 37 degrees C. Differentiation of U937 monocytes into macrophage-like cells specifically increases ATF binding 10- to 20-fold. These results suggest an important role for urokinase in monocyte/macrophage biology: the native enzyme binds to the cells with the amino-terminal domain; the catalytic, carboxyl-terminal domain remains exposed on the cell surface to stimulate localized proteolysis and facilitate cell migration.

Identification and primary sequence of an unspliced human urokinase poly(A)+ RNA.

Human urokinase cDNA clones have been identified from a cDNA library prepared from total RNA of human fibroblasts transformed by simian virus 40 [Okayama, H. & Berg, P. (1983) Mol. Cell. Biol. 3, 280-289]. Synthetic oligonucleotides, corresponding to urokinase protein sequence, were used as probes. The cloned cDNA covers most of the coding sequence and the entire 3' untranslated region. The nucleotide sequence of one of the clones identifies this as a copy of a partially spliced polyadenylylated precursor to urokinase mRNA. The introns separate functionally different domains of the enzyme. Human urokinase mRNA has been identified by RNA blot and its size was estimated at 2500 nucleotides.

The receptor-binding sequence of urokinase. A biological function for the growth-factor module of proteases.

Previous studies have shown that the region of human urokinase-type plasminogen activator (uPA) responsible for receptor binding resides in the amino-terminal fragment (ATF, residues 1-135) (Stoppelli, M.P., Corti, A., Soffientini, A., Cassani, G., Blasi, F., and Assoian, R.K. (1985) Proc. Natl. Acad. Sci. U.S. A. 82, 4939-4943). The area within ATF responsible for specific receptor binding has now been identified by the ability of different synthetic peptides corresponding to different regions of the amino terminus of uPA to inhibit receptor binding of 125I-labeled ATF. A peptide corresponding to human [Ala19]uPA-(12-32) resulted in 50% inhibition of ATF binding at 100 nM. Peptides uPA-(18-32) and [Ala13]uPA-(9-20) inhibit at 100 and 2000 microM, respectively. The human peptide uPA-(1-14) and the mouse peptide [Ala20]uPA-(13-33) have no effect on ATF receptor binding. This region of uPA is referred to as the growth factor module since it shares partial amino acid sequence homology (residues 14-33) to epidermal growth factor (EGF). Furthermore, this region of EGF is responsible for binding of EGF to its receptor (Komoriya, A. Hortsch, M., Meyers, C., Smith, M., Kanety, H., and Schlessinger, J. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 1351-1355). However, EGF does not inhibit ATF receptor binding. Comparison of the sequences responsible for receptor binding of uPA and EGF indicate that the region of highest homology is between residues 13-19 and 14-20 of human uPA and EGF, respectively. In addition, there is a conservation of the spacings of four cysteines in this module whereas there is no homology between residues 20-30 and 21-33 of uPA and EGF. Thus, residues 20-30 of uPA apparently confer receptor binding specificity, and residues 13-19 provide the proper conformation to the adjacent binding region.