Involvement of a mitogen-activated protein kinase signaling pathway in the regulation of urokinase promoter activity by c-Ha-ras.

The expression of the urokinase-type plasminogen activator, which plays a crucial role in tissue remodeling by controlling the synthesis of the broadly acting plasmin serine protease, is regulated by several tyrosine kinases. Since the actions of these tyrosine kinases is dependent on the activation of ras proteins, we undertook a study to identify signaling events downstream of ras responsible for the stimulation of urokinase promoter activity. Transient expression of an activated c-Ha-ras in OVCAR-3 cells, which do not harbor the mutated oncogene, led to a dose-dependent trans-activation of the urokinase promoter. A sequence residing between -2109 and -1964 was critical for the stimulation of the urokinase promoter by c-Ha-ras. Mutation of an AP-1 and a PEA3 site at -1967 and -1973, respectively, or the co-expression of a transactivation domain-lacking c-jun substantially impaired the ability of c-Ha-ras to stimulate urokinase promoter activity. The induction of the urokinase promoter by ras was completely blocked by expression of a dominant negative c-raf expression vector and substantially reduced in cells made to co-express a catalytically inactive mitogen-activated protein kinase kinase. Further, the expression of an ERK1/ERK2-inactivating phosphatase (CL100) abrogated the stimulation of the urokinase promoter by c-Ha-ras. These data argue for a role of a mitogen-activated protein kinase-dependent signaling pathway in the regulation of urokinase promoter activity by ras.

The expression of the urokinase-type plasminogen activator, which plays a crucial role in tissue remodeling by controlling the synthesis of the broadly acting plasmin serine protease, is regulated by several tyrosine kinases. Since the actions of these tyrosine kinases is dependent on the activation of ras proteins, we undertook a study to identify signaling events downstream of ras responsible for the stimulation of urokinase promoter activity. Transient expression of an activated c-Ha-ras in OVCAR-3 cells, which do not harbor the mutated oncogene, led to a dose-dependent transactivation of the urokinase promoter. A sequence residing between ؊2109 and ؊1964 was critical for the stimulation of the urokinase promoter by c-Ha-ras. Mutation of an AP-1 and a PEA3 site at ؊1967 and ؊1973, respectively, or the co-expression of a transactivation domain-lacking c-jun substantially impaired the ability of c-Ha-ras to stimulate urokinase promoter activity. The induction of the urokinase promoter by ras was completely blocked by expression of a dominant negative c-raf expression vector and substantially reduced in cells made to co-express a catalytically inactive mitogen-activated protein kinase kinase. Further, the expression of an ERK1/ERK2-inactivating phosphatase (CL100) abrogated the stimulation of the urokinase promoter by c-Ha-ras. These data argue for a role of a mitogen-activated protein kinase-dependent signaling pathway in the regulation of urokinase promoter activity by ras.
The urokinase-type plasminogen activator has been implicated in a variety of physiological and pathological processes including prostatic involution, cytotrophoblast implantation, and tumor cell invasion, all of which require extensive extracellular matrix proteolysis (1)(2)(3). Urokinase facilitates this process by mediating the conversion of the abundant zymogen, plasminogen, into the widely acting serine protease plasmin, the latter which degrades several extracellular matrix components including laminin, fibronectin, and possibly type IV collagen (4,5). Transcription of the urokinase gene gives rise to a 2.5-kilobase mRNA (6), which is subsequently translated into a single chain Ӎ 50-kDa proenzyme (7,8) and secreted into the extracellular space. The steady state level of urokinase mRNA is determined at two levels. Firstly, the expression of the gene is regulated by 2.1 kilobases of the 5Ј-flanking sequence (9). Secondly, recent studies have indicated that 3Ј-untranslated sequences play a role in determining the stability of the urokinase mRNA (10).
Several studies have indicated that urokinase expression is regulated by growth factors (HGF/SF, EGF), 1 which bind to transmembrane receptor protein tyrosine kinases, and by nonreceptor protein tyrosine kinases including v-src and v-yes (11)(12)(13). One of the early events in this pathway involves the coupling of the ligand-activated receptor protein tyrosine kinase to ras through Grb2.sos or Shc Grb2.sos, this increasing the exchange of GTP for GDP and culminating in an activated ras (14,15). The pivotal role of ras in mediating the signal generated by tyrosine kinases has been deduced from several experimental observations. First and foremost, inhibitory ras mutant proteins (16) block the stimulation of growth and transformation of NIH 3T3 cells by oncogenic tyrosine kinases. Secondly, revertant clones isolated from activated ras-transformed NIH 3T3 cells are resistant to retransformation by protein tyrosine kinase-encoding oncogenes (17). Thirdly, ras, like protein tyrosine kinases, elevates urokinase mRNA levels in chick embryo fibroblasts (13).
However, the events connecting the ras signal to altered gene expression are still not entirely clear. Certainly, the involvement of c-raf, mitogen-activated protein kinase kinase (MAPKK also referred to as MEK) and mitogen-activated protein kinases (MAPK), which form successive tiers in a ras-dependent signal transduction pathway, has been suggested from several reports. c-raf is a protein serine/threonine kinase which acts as a cytoplasmic signal transducer downstream of ras (18) and is an immediate upstream activator of MAPKK (MEK1) (19,20). MEK1, in turn, activates the extracellular signalregulated kinase (ERK) group of MAPK (21) which regulates AP-1 activity via increased transcription of the c-fos gene (22) as a consequence of p62 tcf /elk phosphorylation. Alternatively, modulation of gene expression by ras can also be achieved in a c-raf-independent pathway involving the c-jun amino-terminal kinase (JNK) (23,24).
Although ras increases the steady state level of urokinase mRNA (13), the mechanism by which this is accomplished has not been elucidated. We therefore undertook a study to determine the role of a MAPK pathway in the regulation of urokinase expression by c-Ha-ras. We report, herein, that the stim- ulation of urokinase promoter activity by ras, is dependent on one, or multiple, MAPKs and is mediated via the stimulation of an AP-1 and PEA3 sequence located at Ϫ1967 and Ϫ1973, respectively, as well as a region residing between Ϫ1956 and Ϫ1905 in the urokinase promoter.

EXPERIMENTAL PROCEDURES
Vectors-The 6.6-kilobase BamHI fragment from the activated c-Haras EJ oncogene from the T24 bladder carcinoma cells cloned in a pSV2neo plasmid (25) was used for ras transfections. The TAM 67 vector, encoding a c-jun protein lacking the transactivation domain (amino acids 3-122 absent) of the molecule, has been described elsewhere (26,27). The CL100 expression vector (CL100 SG5), kindly provided by S. M. Keyse, University of Dundee, United Kingdom, contains the CL100 cDNA (encoding residues 1-314) of the CL100 protein (28) fused at the carboxyl terminus to sequences encoding a single copy of the myc epitope tag EQKLISEEDL and ligated into a pSG5 vector (Stratagene). The K97M mutant expression vector encodes a catalytically inactive MEK1 in which a methionine is substituted for a lysine in the ATP binding site (29,30). The c-raf C4 expression vector encodes c-raf lacking the conserved regions 2 and 3 and therefore lacks the kinase domain contained in the latter (18). The 3 ϫ AP-1 pBLCAT construct consists of 3 AP-1 tandem repeats upstream of a thymidine kinase minimal promoter-CAT reporter (pBLCAT) (31). The mutated urokinase promoter CAT reporter constructs have been described elsewhere (32).
Cell Lines-The OVCAR-3 cell line was derived from a malignant ascites of a patient with progressive adenocarcinoma of the ovary (33). Codons 12,13, and 61 of c-Ha-ras are not mutated in the OVCAR-3 cells. 2 All cells were maintained in McCoy's medium 5A supplemented with 10% fetal bovine serum. For the collection of conditioned medium, cells were plated in the presence of serum-containing medium and 24 h later changed to serum-free medium (McCoy's medium 5A supplemented with 4 g/ml transferrin, 5 g/ml of insulin, and 10 ng/ml EGF). Conditioned medium was collected 24 h later, and cells were counted.
Measurement of Secreted Urokinase by Western Blotting-The amount of urokinase in the conditioned medium was determined by Western blotting. Briefly, conditioned medium from equal numbers of cells was denatured in the absence of reducing agent and electrophoresed in a 12.5% SDS-polyacrylamide gel electrophoresis gel. The resolved proteins were transferred to a nitrocellulose filter. The filter was blocked with 3% bovine serum albumin and incubated sequentially with a rabbit polyclonal antibody to human urokinase (product 389, American Diagnostica, Greenwich, CT). Reactive proteins were visualized by ECL as described by the manufacturer (Amersham).
Chloramphenicol Acetyltransferase (CAT) Assays-OVCAR-3 cells were transfected by a calcium phosphate method with CAT reporter constructs fused to the wild type (2109 bp), mutated, or 5Ј-deleted fragments of the human urokinase promoter (9, 32) without, or with, the c-Ha-ras expression vector. All transient transfections were performed in the presence of 5 g of a ␤-galactosidase expression vector to correct for differences in transfection efficiencies. Briefly, DNA precipitate formed in the presence of 124 mM calcium chloride in a buffer containing 15 mM Hepes (pH 7.1), 280 mM NaCl, 1.5 mM Na 2 HPO 4 was added to 50% confluent OVCAR-3 cells. After 6 h, the cells were changed to fresh 10% fetal bovine serum-containing medium and cultured for an additional 36 h. The cells were harvested and lysed by repeated freeze-thaw cycles in a buffer containing 0.25 M Tris-Cl (pH 7.8). Transfection efficiencies were determined by assaying for ␤-galactosidase activity. CAT activity was measured by incubating cell lysate (normalized for transfection efficiency) at 37°C for 6 -8 h with 4 M [ 14 C]chloramphenicol and 1 mg/ml acetyl coenzyme A. After 3 h, the acetyl coenzyme A was replenished. The mixture was extracted with ethyl acetate, and acetylated products were subjected to thin layer chromatography using chloroform:methanol (95:5) as a mobile phase. The amount of acetylated [ 14 C]chloramphenicol was determined using a 603 Betascope.

Expression of c-Ha-ras Increases Urokinase
Promoter Activity-The OVCAR-3 cell line, which does not contain an activated ras, secretes an undetectable level of urokinase, as determined by Western blotting (Fig. 1). We therefore elected to use the OVCAR-3 cells to study the inductive effect of ras on urokinase promoter activity. Transient expression of c-Ha-ras in OVCAR-3 cells led to a dose-dependent increase in urokinase promoter activity (Fig. 2). Addition of 4 g of c-Ha-ras caused a 10-fold stimulation of urokinase promoter activity whereas 50 ng of the oncogene increased CAT activity three times. The induction of the urokinase promoter by ras could not be accounted for by a vector effect since 4 g of the pSV2neo expression vector did not elevate CAT activity over that achieved without the oncogene.

Stimulation of Urokinase Promoter Activity by ras Requires a Region of the Promoter Residing between Ϫ2109 and
Ϫ1964 -To determine the region of the urokinase promoter required for its stimulation by ras, we co-transfected OVCAR-3 cells with a CAT reporter driven by 5Ј deletion fragments of the urokinase promoter and the ras expression vector (Fig. 3). A dramatic stimulation of the urokinase promoter was evident with 2109 bp of the 5Ј-flanking sequence. This stimulation was almost entirely abolished when the CAT reporter was fused to 1964 or to 1870 bp of the 5Ј-flanking sequence. These data suggest that a sequence residing between Ϫ2109 and Ϫ1964 is critical for the stimulation by the ras oncogene.
Stimulation of the Urokinase Promoter by ras Requires Binding Sites for PEA3 and AP-1-The region of the urokinase promoter residing between Ϫ2109 and Ϫ1964 contains binding sites for AP-1 and PEA3 shown to be required for the stimulation of several inducible genes (34,35). Mutation of the AP-1 binding site at Ϫ1967 (mutant B) resulted in a dramatic reduction in the ability of ras to stimulate the urokinase promoter (Fig. 4). Likewise, mutation of a PEA3 (mutant A) motif at Ϫ1973 also impaired the ability of ras to stimulate the urokinase promoter, and this may be related to the observations that c-ets family members play a key role in the regulation of gene expression by ras (36,37). Mutation of two regions of the urokinase promoter (mutant I and F) shown to be required for the maximal stimulation by phorbol ester (32) reduced the stimulation by ras between 40 and 60%. In contrast, an AP-1 sequence residing at Ϫ1885 did not appear to be critical for the stimulation by ras since [ 14 C]chloramphenicol conversion was unchanged when this site was mutated (mutant G).
Stimulation of Urokinase Promoter Activity by ras Is Countered by the Expression of a Transactivation Domain-lacking c-jun-The observation that mutation of an AP-1 binding site could diminish the stimulation of the urokinase promoter ac-2 G. Mills, personal communication. Conditioned medium collected from OVCAR-3 cells and from the high urokinase-producing OC-7 cells (Positive Control) was harvested, clarified by centrifugation, and the cells were enumerated. Conditioned medium, normalized to cell number, and authentic, recombinant, urokinase (rSC-uPA) was denatured, in the absence of reducing agent, and electrophoresed in a 12.5% acrylamide gel. The resolved proteins were transferred to a nitrocellulose filter. The filter was blocked with a solution containing 3% bovine serum albumin and incubated with a polyclonal antibody to human urokinase. Reactive proteins were visualized by ECL. tivity by ras suggested the involvement of transcription factors which bind to this site. To further address this possibility, we transiently transfected OVCAR-3 cells with the urokinase promoter-driven CAT reporter and ras with, or without, an expression vector bearing a c-jun mutant (TAM 67). This c-jun mutant lacks amino acids 3-122 and is unable to transactivate AP-1containing genes (26,27). The mutant protein inhibits AP-1mediated processes through a quenching mechanism by inhibiting the function of endogenous fos and/or jun proteins (26).
Lysates of OVCAR-3 cells transfected with the urokinase promoter-CAT reporter, the c-Ha-ras expression vector, and the empty TAM 67 vector (cytomegalovirus vector) resulted in 92% [ 14 C]chloramphenicol conversion up from 2% in the absence of ras (Fig. 5). In contrast, expression of 1 and 10 g of the c-jun mutant attenuated the ability of ras to stimulate the urokinase promoter (Fig. 5). Thus, conversion of [ 14 C]chloramphenicol was reduced to 29% in cells co-transfected with 10 g of the c-jun mutant. These data suggest that AP-1-binding transcription factors are necessary for the ras-mediated stimulation of the urokinase promoter. The inability of the c-jun mutant to repress the stimulation of the urokinase promoter by ras to the same extent as the urokinase promoter mutated at the AP-1 binding site (at Ϫ1967-mutant B) (Fig. 4) may very well reflect an incomplete quenching brought about by expression of the TAM 67 construct.
Induction of Urokinase Promoter Activity by ras Is Abolished by the Expression of an ERK1-and ERK2-inactivating Phosphatase (CL100)-It is becoming increasingly apparent that the activity and/or synthesis of a number of transcription factors including fos and c-ets family members is modulated by MAPKs including ERK1 and ERK2 (22,38). Since our data suggest that transcription factors, which bind to the AP-1 and PEA3 sequences in the urokinase promoter, are important for the induction by ras, we examined the role of these MAPKs in the stimulation of urokinase promoter activity by the oncogene. Toward this end, we used an expression vector which contains the human CL100 cDNA (28). This encodes a dual specificity (Tyr/Thr) protein phosphatase which inactivates both ERK1 and ERK2 (39). The ability of CL100 to inactivate both ERK1 and ERK2 is of importance to these studies since the simultaneous block of both of these MAPKs is necessary to abrogate the stimulation of AP-1 by phorbol ester and by ras (40). Expression of the vector encoding the CL100 phosphatase completely abrogated the stimulation of urokinase promoter activity by c-Ha-ras (Fig. 6A) whereas the vector lacking the coding sequence of CL100 (pSG5) had only a minimal effect on this stimulation.
Since AP-1-binding transcription factors are required for the stimulation of the urokinase promoter by ras, we determined whether the expression of CL100 could also repress the induction of an AP-1-driven promoter by ras. OVCAR-3 cells were transiently transfected with a CAT reporter driven by 3 AP-1 tandem repeats (3 ϫ AP-1 pBLCAT) upstream of a thymidine kinase minimal promoter (pBLCAT) and the ras expression vector with, or without, the CL100-encoding expression vector. Expression of ras caused a marked stimulation of the AP-1driven CAT reporter (Fig. 6B) in OVCAR-3 cells, whereas the activity of the thymidine kinase minimal promoter-CAT reporter (pBLCAT), which was negligible, was unaffected by the oncogene. Co-expression of the CL100-encoding vector (CL100 SG5) countered the ras-dependent induction of the AP-1-driven CAT reporter. The highest level of the CL100-encoding vector (10 g), but not the empty vector (pSG5), completely abolished the stimulation of the AP-1-driven CAT reporter by ras. Although the CL100-encoding vector also repressed the basal activity of the 3 ϫ AP-1 pBLCAT vector, the amount of this repression was insufficient to account for the MAPK phosphatase-mediated suppression of the ras response. These data indicate that the induction of AP-1 activity by ras, which is critical for the stimulation of the urokinase promoter by the oncogene, is effectively countered by co-expression of a MAPKinactivating phosphatase.
Expression of a Catalytically Inactive MAPKK (MEK1) Inhibits the Stimulation of the Urokinase Promoter by ras-The ability of a MAPK-inactivating phosphatase (CL100) to counter

FIG. 2. Transient expression of c-Ha-ras increases urokinase promoter activity in a dose-dependent manner. OVCAR-3 cells
were transiently transfected at 50% confluency with 10 g of a CAT reporter driven by the wild type (2109 bp) urokinase promoter in the absence, or in the presence, of varying amounts of a vector bearing the c-Ha-ras sequence or 4 g of the vector alone (pSV2neo). All transfections were performed in the presence of 5 g of a ␤-galactosidaseexpressing vector. After 5 h, the medium was changed and cells were cultured for an additional 36 h. The cells were harvested and assayed for ␤-galactosidase activity. Cell extracts, corrected for differences in transfection efficiency, were incubated with [ 14 C]chloramphenicol for 8 h. The mixture was extracted with ethyl acetate and subjected to thin layer chromatography. The conversion of [ 14 C]chloramphenicol to acetylated derivatives was determined with a 603 Betascope.
FIG. 3. Activation of the urokinase promoter by c-Ha-ras requires a sequence residing between ؊2109 and ؊1964 in the 5-flanking sequence of the urokinase gene. OVCAR-3 cells were transiently co-transfected, as described in the legend to Fig. 2, with 4 g of the c-Ha-ras expression vector or an equimolar amount of the pSV2neo vector, 10 g of a CAT reporter driven by the indicated 5Ј-deleted fragment of the urokinase promoter, and 5 g of a vector encoding the ␤-galactosidase gene. After 48 h, the cells were harvested and transfection efficiencies were determined by ␤-galactosidase activity. Cell extract, normalized for transfection efficiency, was incubated with [ 14 C]chloramphenicol for 8 h, and acetylated products were extracted with ethyl acetate and subjected to thin layer chromatography. The conversion of [ 14 C]chloramphenicol to acetylated derivatives was determined with a 603 Betascope. the stimulation of the urokinase promoter by ras was consistent with the notion of ERK involvement. However, since there are conflicting reports concerning the specificity of this MAPKinactivating phosphatase toward the ERK-subgroup of MAPKs (41,42), we asked whether the expression of a functionally inactive upstream activator of ERKs (29) would abrogate the stimulation of the urokinase promoter by ras. MEK1 is a specific activator of the ERKs (21,43), and we made use of a mutant MEK1 expression vector (MEK K97M) in which a methionine is substituted for a lysine at the ATP binding site of the enzyme rendering it catalytically inactive (29,30). We transiently transfected (Fig. 7) OVCAR-3 cells with the urokinase promoter-driven CAT reporter and ras in the presence of an expression vector encoding the MEK K97M mutant. The stimulation of the urokinase promoter by the oncogene ([ 14 C]chloramphenicol acetylation increased from 4 to 17%) was almost entirely abrogated by co-expression of the catalytically inactive MAPKK (6% [ 14 C]chloramphenicol conversion was ev- FIG. 6. Abrogation of the c-Ha-ras-dependent induction of urokinase promoter activity by an ERK1/ERK2-inactivating phosphatase. OVCAR-3 cells were transiently transfected and assayed for CAT activity as described in the legend to Fig. 2. Cells were transfected with a CAT reporter driven by either the wild type urokinase promoter (7 g) (A) or 3 AP-1 tandem repeats (3 ϫ AP-1 pBLCAT) upstream of a thymidine kinase minimal promoter (pBLCAT) (1 g) (B) with (ϩ), or without (Ϫ), 3 g of the c-Ha-ras expression vector (c-Haras) or a vector control (pSV2neo) in the presence of the indicated amount of the CL100-encoding expression vector (CL100 SG5) or the empty vector (pSG5 vector equimolar with 10 g of CL100 SG5). Cell extract, normalized for differences in transfection efficiency, was incubated with [ 14 C]chloramphenicol, and the reaction products were extracted with ethyl acetate and subjected to thin layer chromatography. The conversion of [ 14 C]chloramphenicol was determined with a 603 Betascope.
FIG. 4. The stimulation of urokinase promoter activity by c-Ha-ras requires AP-1 and PEA3 binding sites in the 5-flanking sequence of the urokinase gene. OVCAR-3 cells were transiently co-transfected, as described in the legend to Fig. 2, with 8 g of a CAT reporter driven by the wild type (WT) or the point-mutated (A, B, I, F, G) urokinase promoter with (ϩ) or without (Ϫ) 4 g of c-Ha-ras expression vector or the empty expression vector alone (pSV2neo). CAT activity is indicated numerically as the percent of [ 14 C]chloramphenicol converted. The schematic to the right shows a representation of the mutated region (solid boxes) of the urokinase promoter. The numbers above the arrows refer to the position relative to the urokinase transcription start site. Mutated urokinase promoter constructs are described in more detail elsewhere (32). ident with 10 g of the MEK K97M). In contrast, expression of the wild type MAPKK (MEK wt) had little effect on either the basal or the ras-dependent activation of the urokinase promoter activity. The inability of the wild type MAPKK to stimulate the basal activity of the urokinase-driven CAT reporter (i.e. in the absence of ras) in the OVCAR-3 cells is consistent with a report by Mansour et al. (29) showing that this expression vector does not activate an AP-1-driven CAT reporter nor does it transform NIH 3T3 cells. These data indicate that a MAPKK, which functions as an ERK activator, is necessary (but possibly by itself insufficient) for the regulation of urokinase promoter activity by ras.
Co-expression of a Dominant Negative c-raf Expression Vector Abrogates the Stimulation of the Urokinase Promoter by ras-Since c-raf is an upstream activator of MAPKK (19,20,44), we then determined the sensitivity of urokinase expression to a dominant negative c-raf expression vector. OVCAR-3 cells were transiently transfected with the urokinase promoterdriven CAT reporter and the ras expression vector in the presence of increasing amounts of a dominant negative c-raf expression vector (raf C4). This vector encodes the aminoterminal 257 amino acids of c-raf (18) and is therefore kinasedeficient. Expression of the raf C4 led to a dose-dependent reduction in the ability of the ras expression vector to stimulate the urokinase promoter (Fig. 8). Thus, while a conversion of [ 14 C]chloramphenicol of 43% was apparent in OVCAR-3 cells transiently transfected with ras and the urokinase promoter-CAT reporter, this was reduced to 7% in the presence of 0.1 g of the dominant negative raf C4 expression vector. In contrast, the empty vector lacking the c-raf coding sequence failed to repress the induction of the CAT reporter by ras. Although the raf C4 expression vector also reduced the basal activity of the urokinase promoter (i.e. in the absence of ras), the magnitude of this response could not account for the reduction observed in the presence of the activated oncogene. These data are consistent with the hypothesis that the activation of the urokinase promoter by ras in the OVCAR-3 cells is dependent on c-raf. However, since the mutant protein encoded by the raf C4 expression vector binds to the effector domain of ras (45), we cannot completely rule out the possibility that the raf C4 dom-inant negative vector is functioning in the OVCAR-3 cells by interfering in the activation of ras effectors other than c-raf. DISCUSSION Urokinase gene expression is regulated by a diverse set of growth factors which interact with transmembrane receptor tyrosine kinases including HGF/SF and EGF as well as by non-membrane tyrosine kinases including v-src (11)(12)(13). Although there is ample evidence that these tyrosine kinases mediate their effects via ras-dependent pathways (16), the downstream event(s) by which ras stimulates urokinase gene expression has not been determined, and it now appears that ras can stimulate at least 2 distinct pathways, one which is c-raf-dependent, the other c-raf-independent (24). Stimulation of the c-raf-dependent pathway by ras leads to the sequential activation of MAPKK and the ERK group of the MAPKs (19 -21). The ERKs phosphorylate several transcription factors including p62 tcf /elk, this leading to increased AP-1 activity as a consequence of enhanced fos gene transcription (22). Several observations in this, and other, studies suggest that the induction of urokinase promoter activity by ras occurs through this pathway. Firstly, expression of the ERK1/ERK2-inactivating CL100 phosphatase efficiently blocked the ras signal in OVCAR-3 cells. Secondly, an expression construct encoding a catalytically inactive MEK1, which is specific for the ERKs, substantially reduced the ability of ras to stimulate the urokinase promoter. Thirdly, the block of c-raf, which is an upstream effector of MEK1 (19,44), with a dominant negative expression vector, completely abrogated the ras-dependent stimulation of the urokinase promoter. Thus, together these data would suggest a pivotal role for one, or multiple, ERKs in the stimulation of the urokinase promoter by ras and are consistent with the hypothesis that this is achieved by the c-raf-dependent pathway described by Minden et al. (24). Additionally, since dominant negative mutants to ERK1 and ERK2 also block AP-1 activation (an event necessary for the induction of the urokinase promoter by ras) as well as the proliferation response to growth factors including EGF (40,46), it may very well be that the ability of these stimuli to elevate urokinase expression (47) FIG. 7. Expression of a catalytically inactive MAPKK attenuates the ability of ras to stimulate the urokinase promoter. OVCAR-3 cells were transiently transfected and assayed for CAT activity as described in the legend to Fig. 2. Cells were co-transfected with 7 g of a CAT reporter driven by the wild type urokinase promoter (u-PA) with (ϩ) or without (Ϫ) 4 g of the c-Ha-ras expression vector (c-Ha-ras) in the presence of the indicated amount of an expression vector encoding either the wild type (MEK wt) or point-mutated (MEK K97M) MEK1 sequence (29). Cell extract, normalized for differences in transfection efficiency, was incubated with [ 14 C]chloramphenicol, and the reaction products were extracted with ethyl acetate and subjected to thin layer chromatography. The amount of acetylated [ 14 C]chloramphenicol was determined on a 603 Betascope. is mediated through this pathway.
Further evidence that the regulation of urokinase expression by ras is mediated through a c-raf-dependent pathway comes from studies with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate. TPA, via protein kinase C, is an activator of c-raf (48), and treatment of cells with this agent leads to a strong activation of urokinase gene transcription (49). The transcriptional requirements for urokinase promoter stimulation by TPA (32) and ras are similar in that mutation of the PEA3 and AP-1 motifs (Ϫ1973 and Ϫ1967, respectively), as well as a region shown to be required for the efficient induction of urokinase gene expression by the phorbol ester (32), substantially impaired the ability of the oncogene as well as TPA to stimulate the promoter. Presumably, these findings for ras reflect the well characterized observations of increased AP-1 and PEA3 activity brought about by the oncogene as a consequence of increased synthesis, or activation by phosphorylation, of c-jun, c-fos, and c-ets family members (50,51).
While these data suggest the involvement of ERKs in the regulation of urokinase expression by ras, we cannot currently exclude a role for the JNKs in this process. ras, via MEKK, is an inducer of JNK (24), and this leads to increased AP-1 activity subsequent to c-jun phosphorylation (23). However, any speculation as to the role of JNKs in the regulation of urokinase expression by ras must take into account our observations that interfering with the ERK pathway with a MEK1 mutant, or with a dominant negative c-raf expression vector, effectively suppresses the ras stimulus. Thus, either the JNKs play only a minor role in the regulation of urokinase expression by the oncogene or alternatively both ERKs and JNKs are concurrently required for the stimulation of urokinase promoter activity by ras. Notwithstanding these considerations, our findings favor a key role for the ERKs in the stimulation of urokinase promoter by ras.
In conclusion, we have shown that the induction of urokinase promoter activity by an activated c-Ha-ras requires a PEA3/AP-1 sequence in the urokinase promoter and is mediated through a c-raf and MAPK-dependent signal transduction pathway. Since several tyrosine kinases including the receptors for EGF (c-erbB1), and HGF/SF (p190 met ), as well as the non-receptor tyrosine kinase v-src, mediate their effects through ras, it is interesting to speculate that c-raf and/or the MAPKs may represent novel therapeutic targets for repressing the elevated synthesis of the plasminogen activator in invasive cancer which is driven by both autocrine and paracrine mechanisms.