Activation and nuclear translocation of mitogen-activated protein kinases by polyomavirus middle-T or serum depend on phosphatidylinositol 3-kinase.

Several cellular signal transduction pathways activated by middle-T in polyomavirus-transformed cells are required for viral oncogenicity. Here we focus on the role of phosphatidylinositol 3-kinase (PI 3-kinase) and Ras and address the question how these signaling molecules cooperate during cell cycle activation. Ras activation is mediated through association with SHC.GRB2.SOS and leads to increased activity of several members of the mitogen-activated protein (MAP) kinase family, while activation of PI 3-kinase results in the generation of D3-phosphorylated phosphatidylinositides whose downstream targets remain elusive. PI 3-kinase activation might also ensue as a direct consequence of Ras activation. Oncogenicity of middle-T requires stimulation of both Ras- and PI 3-kinase-dependent pathways. Mutants of middle-T incapable to bind either SHC.GRB2.SOS or PI 3-kinase are not oncogenic. Sustained activation and nuclear localization of one of the MAP kinases, ERK1, was observed in wild type but not in mutant middle-T-expressing cells. Wortmannin, an inhibitor of PI 3-kinase, prevented MAP kinase activation and nuclear localization in middle-T-transformed cells. PI 3-kinase activity was also required for activation of the MAP kinase pathway in normal serum-stimulated cells, generalizing the concept that signaling through MAP kinases requires not only Ras-but also PI 3-kinase-mediated signals.

Proteins expressed early in the virus life cycle of polyomavirus, the tumor antigens (T antigens), are responsible for tumor formation in virus-infected animals and virus-mediated transformation of cells in culture (1). Large tumor antigen (large-T) is a nuclear protein known to immortalize primary cells in culture (2) while middle tumor antigen (middle-T) causes phenotypic changes associated with malignant cell growth (3). The activity of middle-T results from its association with intracellular signal-transducing proteins like members of the Src family of tyrosine kinases (c-Src, Fyn, and c-Yes) (4), the 85-and 110-kDa subunits of a phosphatidylinositol 3-kinase (PI 3-kinase) 1 (5), the catalytic and regulatory subunits of protein phosphatase 2A (PP2A) (6,7), and the SH2 domain-containing protein SHC (8,9) whose putative role is to activate the Ras signaling pathway (10,11). More recently, middle-T immunoprecipitates have been found to contain a member of the 14-3-3 family of proteins, some of which are involved in stimulating ADP-ribosylation (12).
Middle-T activates intracellular signal transduction pathways mediated by PI 3-kinase and Ras, respectively. The latter becomes activated upon association of the SHC⅐GRB2⅐SOS complex with middle-T. Middle-T-transformed cells show an increase in the fraction of the GTP-bound form of Ras (13) and transfection with genes suppressing Ras activity results in reversion to a more normal phenotype (14). Activated Ras stimulates cell growth and differentiation through a kinase cascade involving Raf and MEK culminating in the activation and nuclear translocation of several members of the MAP kinase family (15)(16)(17). Middle-T has also been shown to activate transcription factors of the AP1 family like Jun and Fos (13,18) or Myc (19), the former being direct targets of MAP kinases (20 -24). The ability to activate transcription of cellular genes through MAP kinases is a prerequisite for cell transformation and delineating the underlying mechanisms is therefore of pivotal importance in understanding virus-mediated cell transformation.
In this study we investigate the signals activated by middle-T through SHC⅐GRB2⅐SOS and PI 3-kinase, respectively, and evaluate their importance for T antigen oncogenicity. Activation and translocation of ERK1 to the nucleus was observed in cells expressing wild type (WT) middle-T. Cells expressing T antigen mutants unable to bind SHC and PI 3-kinase, respectively, showed neither activation nor nuclear translocation of MAP kinases, suggesting that both pathways feed into the MAP kinase cascade. Similarly, we found that PI 3-kinase was also required for MAP kinase activation and translocation in untransformed cells stimulated with growth factors.

EXPERIMENTAL PROCEDURES
Materials-Restriction enzymes were purchased from Boehringer Mannheim. Pepstatin, aprotinin, myelin basic protein, and sodium vanadate were from Sigma. Wortmannin was purchased from Sigma. ␤-Glycerophosphate and protein A-Sepharose CL-4B were obtained from Pharmacia Biotech Inc., G418 (Geneticin) was from Life Technologies, Inc., and Micro BCA protein assay reagents were from Pierce. Luciferin was obtained from Chemie Brunschwig AG, and [␥-32 P]ATP was from ICN.
Cell Lines-NIH 3T3 and Fisher rat F111 fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% calf serum at 37°C in a humidified CO 2 incubator. 3T3 cell lines expressing wild type (mT4 and mT8) or mutant forms of middle-T were generated by stable transfection as described earlier (25). The cell line mT7 was obtained by infection of NIH 3T3 cells with the middle-T carrying retrovirus N-TKmT (26).
Plasmid Constructs-All plasmid construction steps were carried out using standard molecular cloning techniques. The plasmids * 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.
pcDNA1mT, pcDNAY250FmT, pcDNANG59mT, and pcDNA1387TmT were derived from the pcDNA1 expression vector (Invitrogen) by inserting the coding region of the respective polyomavirus middle-T mutants at HindIII/EcoRI downstream of the cytomegalovirus promotor. pcDNA1178TmT was cloned into the HindIII/BglI sites of the pcDNA1 vector. pcDNAdl1015mT was cloned into EcoRI/BamHI.
The reporter plasmids pGl2muPA-8.2 and pGl2muPA-35 were constructed by inserting the murine uPA gene promoter (from Ϫ8.2 kilobase pairs to ϩ398 base pairs with respect to the transcription initiation site) upstream of the luciferase-coding region of the promoterless plasmid pGL2-basic (Promega). In p3xAP1-tk-luc, 2 three consensus AP1 elements were inserted upstream of the minimal promoter of the tymidine kinase gene (-46 to ϩ52) containing the TATA box and the transcription initiation site, which is linked to the luciferase gene. The control plasmid, pGl2-control (Promega) contains the SV40 enhancer and promoter. pfos-luc contains a human c-fos promotor (Ϫ711 to ϩ45) linked to the luciferase gene (27). pSRA⌬mSOS1 contains the coding region of a dominant negative form of SOS under the control of the human T-cell lymphotrophic virus (type I)-long terminal repeat promoter (28). The plasmid pCMV N⌬raf encodes a dominant negative form of Raf under the control of the cytomegalovirus promotor (29).
Determination of ERK1 and ERK2 Activity-Cell lysates were prepared and ERK activity measured as described (30). Rabbit polyclonal antibodies against ERK1 (F15P) and ERK2 (F13S) used for immunoprecipitation have been described before (30). ERK1 and ERK2 activity were quantified in a PhosphorImager system (Molecular Dynamics).
Transient Transfection and Analysis of Reporter Gene Expression-NIH 3T3 cells (0.2 ϫ 10 6 /well) were plated in six-well tissue culture plates with DMEM containing 10% calf serum and transfected 17 h later by the calcium phosphate precipitation method (Pharmacia). 1.0 g of reporter gene and the indicated amounts of the plasmid to be coexpressed were diluted in 40 l of solution A (500 mM CaCl 2 , 100 mM Hepes/NaOH, pH 6.95). The solution was left for 10 min at room temperature, mixed with 80 l of solution B (280 mM NaCl, 50 mM Hepes NaOH, pH 6.95, 750 M NaH 2 PO 4 , 750 M Na 2 HPO 4 ), vortexed, and incubated for 15 min at room temperature to allow precipitation. The mixture was added to the cells without changing the medium, and the cells were kept at 37°C in a humidified incubator for 4.5 h. The medium was removed and the cells treated for 3 min with 1 ml of 15% glycerol in 20 mM Hepes/NaOH, pH 7.4, at room temperature and washed twice with phosphate-buffered saline containing 5 mM MgCl 2 and 9 mM CaCl 2 . The incubation was continued after addition of 2 ml of fresh DMEM containing 10% calf serum. After 18 h the cells were harvested in 200 l of luciferase lysis buffer (25 mM glycylglycine NaOH, pH 7.8, 1 mM dithiothreitol, 15% glycerol, 8 mM MgSO 4 , 1 mM EDTA, and 1% Triton X-100). Total protein (30 -60 g) was diluted with 150 l of dilution buffer (25 mM glycylglycine/NaOH, pH 7.8, 10 mM MgSO 4 , 5 mM ATP) and luciferase activity was measured in a luminometer (Autolumat LB 953, Berthold) after injecting 300 l of luciferin solution (25 mM glycylglycine NaOH, pH 7.8, 330 M luciferin). The protein concentration was determined using a Micro BCA protein assay system (Pierce).
Western Analysis-24 l of the lysates prepared for ERK activity assays were added to 6 l of 5-fold ϫ concentrated SDS sample buffer, separated on 12.5% SDS-polyacrylamide gel electrophoresis, and subsequently transferred to a polyvinylidene difluoride membrane. The membrane was immunodecorated with 1 g/ml anti-ERK1/ERK2 antibody (Santa Cruz) and developed with an alkaline phosphatase-linked second step antibody (Southern Biotechnology Association) using a colorimetric detection system (5-bromo-4-chloro-3-indolyl phosphate/ nitro blue tetrazolium, Boehringer Mannheim).
Transformation Assay-Focus formation in NIH 3T3 mouse fibroblasts and F111 rat fibroblasts was determined as described earlier (25).
Immunofluorescence-Localization of ERK1 in 3T3, F111, or microinjected REF-52 cells by immunofluorescence was performed as described (60) 8 -15 h after introduction of the corresponding DNA. The cellular distribution of ERK1 and T antigens was examined on a Leica TCS 4D confocal laser scanning microscope using a 40ϫ NA 1.00 -0.5 oil immersion or a 63ϫ NA 1.4 oil immersion objective. For detection of ERK1 we used a rabbit polyclonal anti-ERK1 antibody (Upstate Biotechnology Inc.). Middle-T was labeled with Pab762. 3 Wortmannin Treatment of NIH 3T3, , and F111 Cells-In microinjection experiments, wortmannin was added to 100 nM immediately after injection. After 4 h the same amount of the drug was added to compensate for loss of activity due to hydrolysis followed by incubation for another 4 h. in the activation of several members of the MAP kinase family. We determined the activity of two members of this family, ERK1 and ERK2, in normal and middle-T-transformed NIH 3T3 cells. As shown in Fig. 1A, stimulation of growth-arrested 3T3 cells with serum resulted in approximately 20-fold activation of ERK1 and ERK2. Maximal activity was reached within 15 min after serum addition and returned to basal levels of asynchronously growing cells within 5-7 h. In serum-starved middle-T-expressing cells serum activated ERK1 and ERK2 to the same maximal level as in control cells. The relative increase in ERK activity was, however, only 5-fold due to increased basal activity (Fig. 1B). Basal MAP kinase activity measured with myelin basic protein as substrate in vitro in asynchronously growing middle-T-expressing cell lines was higher than in control cells as shown in Fig. 2A. Western blots performed with an ERK1/ERK2-specific polyclonal antibody showed that in normal 3T3 cells most of ERK1 and ERK2 was shifted toward higher M r upon serum stimulation indicative of phosphorylation by MEK (Fig. 2B). In unsynchronized cells expressing WT middle-T, a barely detectable fraction of ERK1 and ERK2 showed the typical M r shift while the bulk of the protein remained in the low M r form (Fig. 2B). Cells expressing non-transforming mutants of middle-T-like 1178T (31), displaying dramatically reduced binding of PI 3-kinase and Y250F (8,9,32), deficient in associating with SHC and thereby unable to initiate signaling through Ras, respectively, showed a serum response similar to control cells (Fig. 1, C and D, Table I). The typical M r shift observed in control 3T3 cells upon serum stimulation was also detected in mutant T antigen-expressing cells (Fig. 2B). These data suggest that ERK activity was constitutively high in WT middle-T-expressing cells irrespective of the serum content in the growth medium ( Figs. 1 and 2). Interestingly, the constitutively high MAP kinase activity in middle-Ttransformed cells is not reflected by a corresponding increase in the M r of these kinases (Fig. 2B). The slightly delayed activation of ERK1 and ERK2 observed in Fig. 1 (C and D) was not of statistical significance.

Polyomavirus Middle-T Activates MAP Kinases-Middle-T
Middle-T Stimulates Transcription from Various Promoters-MAP kinases phosphorylate several transcription factors such as Jun and Fos resulting in the activation of a variety of promoters. Activation of MAP kinases was therefore measured as the increase in transcription of a series of luciferase reporter gene constructs carrying the uPA promoter, the Fos promoter, and an artificial promoter containing three AP1 sites, respectively. The effect of middle-T on these promoters was assessed in NIH 3T3 cells transiently transfected with a reporter plasmid together with a middle-T expression vector. Fig. 3A shows that middle-T expression resulted in approximately 30-fold activation of the uPA and 20-fold activation of the Fos and AP1 promoter while the SV40 promoter was only slightly induced. Similar activation of the uPA promoter was observed when the reporter construct was expressed in middle-T-transformed cell lines (mT7, Fig. 3A). These data establish that the uPA promoter and the PEA3/AP1 sites present in the Fos and AP1 promoter are regulated by signaling pathways activated by middle-T. The most likely candidates mediating this activation are members of the MAP kinase family.
Middle-T Activates ERK1 and ERK2 via the Ras, Raf, MEK Cascade-In order to identify signaling intermediates responsible for activation of MAP kinases in middle-T-transformed cells, we introduced dominant negative SOS or Raf together with the luciferase reporter plasmid into 3T3 fibroblasts. Fig.  3B shows that dominant negative Raf or SOS block T antigenmediated activation of the uPA promoter and suggests that Ras acts as signaling intermediate. , an ERK-specific dual specificity phosphatase required for down-regulation of MAP kinases, also blocked the response to middle-T in reporter gene-expressing cells, further demonstrating that middle-T activates the MAP kinase cascade (Fig. 3B).
So far we have shown that middle-T activates MAP kinases. Earlier reports suggest that constitutively activated MEK, the kinase acting upstream of ERK1 and ERK2, is sufficient for cell transformation (34 -36). This implies that Ras-mediated signaling is sufficient for ERK activation. An analysis of various middle-T mutants, on the other hand, suggests that both SHCand PI 3-kinase-initiated pathways are required for T antigen oncogenicity (4,37). To address this discrepancy, we investigated a series of non-oncogenic middle-T mutants (Table I).
1178T (31) shows dramatically reduced binding of PI 3-kinase, while Y250F (8,9,32) is deficient in associating with SHC and thereby unable to initiate signaling through Ras. NG59 (38) binds none of the molecules characterized so far, 1387T (a truncated mutant protein lacking the membrane anchor sequence) only binds PP2A (39,60), and dl1015 associates with all cellular enzymes described so far, but is unable to activate PI 3-kinase (40,41). Mutant genes were introduced into NIH 3T3 cells and activation of MAP kinase measured in the luciferase reporter assay. Fig. 3B shows that only WT middle-T fully stimulated the reporter gene. NG59, 1387T, and dl1015 were almost completely inactive, while 1178T and Y250F were about 50% as effective as WT in inducing the uPA promoter.  Residual stimulation of the uPA promoter by mutant middle-Ts was totally blocked by dominant negative SOS and Raf or by MKP-1. These findings further support the view that both SHC-and PI 3-kinase-initiated signaling pathways are required for efficient stimulation of the MAP kinase cascade.
In Middle-T-transformed Cells, ERK1 Is Localized in the Nucleus-It is well established that in order to accomplish mitogenic stimulation, MAP kinases have to translocate to the nucleus upon growth factor stimulation (15,(42)(43)(44). We therefore studied the localization of a representative member of this family of kinases, ERK1, in a variety of normal and middle-Ttransformed cell lines. Serum stimulation of resting 3T3 cells resulted in accumulation of ERK1 in the nucleus as expected (Fig. 4, A-C). Panels D-F show that most of ERK1 was nuclear in middle-T-expressing cells and, most importantly, that nuclear localization of ERK1 was only slightly influenced by the growth conditions of the cells. Similarly, translocation of ERK1 to the nucleus was observed in REF-52 cells microinjected with plasmids carrying a WT middle-T-specific cDNA (Fig. 4, G and  H). None of the non-oncogenic mutant T antigens stimulated translocation of ERK1 to the nucleus. Y250F middle-T, the mutant impaired in SHC binding, was completely defective (Fig. 4, I and K), while 1178T and dl1015, mutants unable to activate PI 3-kinase, had dramatically reduced potential to relocalize ERK1. A quantification of the data obtained with several hundred microinjected cells is shown in Fig. 5. Our data show that both PI 3-kinase and Ras activation are required for stimulation of the MAP kinase cascade by middle-T.

Wortmannin, an Inhibitor of PI 3-Kinase, Prevents Accumulation of MAP Kinases in the Nucleus and Blocks Entry into S Phase-We have shown so far that in WT middle-T-expressing cells, basal activity of MAP kinases is increased concomitantly
with relocalization of ERK1 to the nucleus. This activity requires both SHC and PI 3-kinase-dependent pathways. Middle-T-expressing cells were treated with wortmannin, a specific inhibitor of phosphatidylinositol 3-kinases (45), as shown in Fig. 5. The number of cells expressing nuclear ERK1 dropped to less than 10% in drug-treated WT or mutant middle-T-injected cells, indicating that PI 3-kinase activation was essential for initiation of the MAP kinase cascade by polyomavirus.
To test the relevance of these findings for growth factormediated signaling in normal cells, we studied the localization of ERK1 in serum-stimulated fibroblasts in the absence and presence of wortmannin (Fig. 4, L and M). Nuclear accumula-  tion of ERK1 was completely blocked by the drug, in agreement with the data shown for microinjected cells expressing middle-T. Similarly, the M r shift indicative of activation of ERK1 and ERK2 was abolished when G 0 -arrested cells were stimulated with serum growth factors (Fig. 2C).
We also investigated the effect of PI 3-kinase on mitogenstimulated induction of S phase in mouse NIH 3T3 and F111 rat fibroblasts. Table II shows that wortmannin blocked seruminduced initiation of DNA synthesis measured as bromodeoxyuridine incorporation into cellular DNA, while control cells efficiently entered S phase.

Oncogenic Transformation of Cells by Middle-T Requires Signals Elicited by SHC and PI 3-Kinase-
The oncogenic potential of various middle-T mutants was determined in focus assays as shown in Fig. 6. WT middle-T efficiently transformed both cell lines, while mutant T antigens defective in either SHC or PI 3-kinase binding were totally inactive. Since sustained activation of MAP kinases by middle-T required both the SHC and the PI 3-kinase pathway, we were interested in the question whether these pathways could complement each other when activated through separate T antigen mutants introduced into cells simultaneously. As shown in Fig. 6, foci also arose in cells concomitantly transfected with two defective middle-T mutants, 1178T and Y250F, disabled in binding PI 3-kinase and SHC, respectively. The number of foci, 20 -30% of that observed in WT middle-T-transfected cells, suggests that the two mutants created independent signals synergistically activating pathways required for oncogenic transformation. The high rate of focus formation rules out the possibility that the two mutant genes recombined giving rise to an oncogenic WT gene. (4,37) and activates intracellular signal transduction pathways mediated by the PI 3-lipid kinase and Ras, respectively. It is well established that Ras stimulates the MAP kinase pathway (15)(16)(17). Owing to its ability to induce the phosphorylation of cellular transcription factors like Jun and Fos, middle-T has been suggested to activate the MAP kinase cascade (13,18,46). It has also been shown that middle-T activates genes coding for various transcription factors (19,47) like Fos (18) and Jun (46). Investigating the effect of middle-T on the MAP kinase pathway we studied several parameters. (i) We measured the activity of MAP kinases in vitro using myelin basic protein as substrate; (ii) we determined the shift in the apparent M r of MAP kinases upon cell stimulation indicative of activation upon phosphorylation by MEK; (iii) we measured the activity of MAP kinaseregulated promoters in reporter plasmid-transfected cells using a luciferase reporter gene; and (iv) we studied the intracellular localization of a representative member of the MAP kinase family, ERK1, by immunofluorescence microscopy.

Middle-T transforms cells through association with a variety of proteins involved in cell signaling
Expression of WT middle-T, but not of transformation-defective mutant proteins, resulted in high basal MAP kinase activity in asynchronously growing or serum-starved cells. WT middle-T-expressing cells showed increased ERK1 and ERK2 activity, as determined in the myelin basic protein phosphorylation assay but showed no corresponding shift in the apparent M r of these kinases typical for MAP kinase activation in growth factor-stimulated cells. This might be explained by the fact that middle-T-transformed cells do not accumulate in G 0 upon serum starvation and are refractory to further stimulation by growth factors. The M r shift in MAP kinases might only arise in cells entering the cell cycle from G 0 but not in cycling cells. Alternatively, the M r shift of ERK1 and ERK2 might be transient preceding translocation to the nucleus. Since MAP kinases are constitutively localized in the nucleus of middle-Ttransformed cells, transient phosphorylation by MEK might escape detection on Western blots. Earlier data obtained with cells overexpressing mutant forms of ERK1 and ERK2 suggest that the shift in M r resulting from phosphorylation by MEK as well as enzymatic activity of ERK1 and ERK2 are not required for translocation to the nucleus (15,17,43). A change in activity and/or specificity of PP2A upon association with T antigens  might further contribute to altered phosphorylation and activity of ERK1 and ERK2 in polyomavirus-transformed cells. In agreement with this idea, recently published papers show that SV40 small-T activates the MAP kinase pathway by blocking PP2A-mediated down-regulation (48,49). A detailed analysis of the pathways targeted by middle-T was performed in cells co-transfected with middle-T and a reporter gene consisting of the coding region derived from the firefly luciferase gene under the control of the uPA, Fos, or AP1 promoters, respectively. Emphasis was on the uPA promoter, since it has been shown previously that expression of uPA is increased in middle-T-induced endotheliomas. This protease has been shown to be one of the major determinants in T antigen-induced morphological transformation (26). Signaling through the MAP kinase pathway was dramatically reduced in cells expressing transformation-defective middle-T mutants. Dominant negative Raf, dominant negative SOS, and overexpression of MKP-1, a phosphatase involved in down-regulation of MAP kinases, blocked T antigen-mediated activation of the uPA promoter reminiscent of experiments performed earlier with tyrosine kinase growth factor receptors (33,50).
While short term treatment of cells with growth factors is sufficient to transiently activate MAP kinases, sustained activation accompanied by nuclear translocation of ERK1 and ERK2 are required for mitogenic stimulation of cells through this pathway (17,51,52). A variety of T antigen mutants was used to address the question which of the pathways initiated by middle-T were required for sustained activation and nuclear translocation of MAP kinases. Our data show that only WT middle-T efficiently stimulates nuclear translocation of ERK1 establishing that SHC-mediated activation of Ras was not sufficient for activation of the MAP kinase cascade. Thus PI 3-kinase activation seems to be an important factor in T antigenmediated MAP kinase activation and mitogenic signaling. To corroborate these findings, we treated middle-T-expressing cells with wortmannin, an inhibitor of PI 3-kinase. Relocalization of ERK1 upon middle-T expression was completely blocked by the drug, confirming the results obtained with middle-T mutants unable to activate PI 3-kinase. Nuclear translocation of MAP kinase observed in a small fraction of 1178T middle-T-expressing cells is therefore most likely the consequence of residual PI 3-kinase activity and not due to a PI 3-kinaseindependent pathway (5,31,53). This explanation is consistent with earlier studies showing that mutation of tyrosine 315, the major binding site for the 85-kDa subunit of PI 3-kinase, to phenylalanine in the 1178T mutant, reduced but did not completely abolish oncogenicity and PI 3-kinase activity (5,31). Remaining activity might be the consequence of residual binding of p85 to this mutant protein. Alternatively, elevated D3phosphorylated PIP 3 levels might arise from SHC-mediated Ras activation, resulting in stimulation of PI 3-kinase (54). Our interpretation of these data was confirmed with another mutant, dl1015, still able to associate with this enzyme yet unable to activate PI 3-kinase activity (41).
The fact that Y250F middle-T was totally defective in causing nuclear localization of ERK1 and ERK2 yet only 2-fold reduced in inducing the uPA gene can be explained in three ways. (i) Detection of nuclear localization of ERK1 by immunostaining depends on accumulation of a significant fraction of the enzyme in the nucleus while only a small amount of nuclear ERK1 might be sufficient to activate the uPA promoter; (ii) Transient transfections of Y250F middle-T together with the reporter plasmid had to be performed in the presence of serum that might compensate for the defect of this mutant in initiating the Ras pathway; (iii) activation of the uPA promoter might also ensue after phosphorylation of transcription factors in the cytoplasm followed by their translocation to the nucleus. Oncogenicity of WT middle-T is most likely the result of the induction of a complex set of cellular genes upon phosphorylation of various transcription factors by MAP kinases. Our data suggest that activation and translocation of MAP kinases to the nucleus upon expression of middle-T best correlates with mitogenicity and oncogenicity of this protein. Partially defective mutants unable to cause relocalization of MAP kinases do not initiate the cell cycle, suggesting that some of the crucial substrates of MAP kinases are nuclear.
Wortmannin blocked nuclear translocation of MAP kinases in middle-T-expressing cells and efficiently prevented phosphorylation and relocalization of these kinases to the nucleus in serum-stimulated control cells, establishing our findings as a general phenomenon during mitogenic stimulation of cells. Support for our observation also comes from the fact that wortmannin has been shown by others to reduce the efficiency of signaling through the MAP kinase pathway upon insulin or serum treatment (55,56).
To test the hypothesis that SHC⅐GRB2⅐SOS-and PI 3-kinase-initiated pathways operate independently on MAP kinases, we performed focus assays with cells transfected with two plasmids encoding 1178T and Y250F middle-T, respectively. While each mutant alone was unable to induce foci, a combination of both efficiently transformed cells indicating that the two pathways can be initiated from separate middle-T complexes and efficiently cooperate to activate the MAP kinase cascade.
In summary, we have shown here that both SHC⅐GRB2⅐SOS as well as PI 3-kinase-induced signaling pathways are required for full stimulation of the MAP kinase cascade by polyomavirus middle-T or serum growth factors. It remains the goal of further studies to identify the level at which these pathways crosstalk. Recently published data demonstrate that a constitutively activated PI 3-kinase activates Ras, Raf, and MAP kinases and stimulates transcription of Fos, suggesting a role for this enzyme upstream of Ras (57). Studies with mutant growth factor receptors point to a role of PI 3-kinase in mitogenesis, cell migration, and receptor internalization (58). Whether localization of MAP kinases is affected by mutations preventing binding of PI 3-kinase to activated growth factor receptors remains to be determined. Other signaling molecules such as S6 kinase and the proto-oncogene akt1 have been identified as putative downstream targets of PI 3-kinase (59). It will be interesting to investigate whether these kinases are necessary for middle-T-mediated transformation and nuclear translocation of MAP kinases.