¹Institute of Biochemistry, Chung Shan Medical and Dental College, No. 113, Section 2, Ta-Ching Street, Taichung 402, Taiwan, R.O.C.

²Department of Pharmacology, Medical College, National Cheng Kung University, Tainan, Taiwan 70101, R.O.C.

Correspondence to: T-H Leu, Department of Pharmacology, Medical College, National Cheng Kung University, Tainan, Taiwan 70101, R.O.C.

Two isoforms of Eps8, p97^Eps8 and p68^Eps8, have been identified as the substrates for receptor tyrosine kinases. Our previous studies indicated that both tyrosyl phosphorylation and protein expression of Eps8 were elevated in v-Src transformed cells. In an attempt to examine the role played by p97^Eps8 in tumorigenesis, we have first obtained cells overexpressing p97^Eps8 and its pleckstrin homology (PH)-truncated variant. We then demonstrated that cells overexpressing p97^Eps8 not only exhibited the ability of focus formation in cell culture but also promoted the tumor formation in mice as compared to controls. Furthermore, elevated serum-induced extracellular responsive kinase (ERK) activation was observed in p97^Eps8 overexpressors. This enhanced ERK activation was sensitive to a MEK1 specific inhibitor PD98059 and was important for p97^Eps8-mediated transformation, since transfection of vectors expressing dominant negative MEK1 and p97^Eps8 abrogated focus formation by p97^Eps8. In contrast, PH-truncated p97^Eps8 failed to localize at the plasma membrane and that the truncated variant also did not elevate ERK activation and cellular transformation in response to serum stimulation. Our results thus indicated that: (i) the gene encoding p97^Eps8 was an oncogene; (ii) p97^Eps8-induced oncogenesis was partly mediated by ERK activation; and (iii) the PH domain of p97^Eps8 was critical for its cellular localization, ERK activation and its ability to transform cells. Oncogene (2001) 20, 106-112.

p97^Eps8; pleckstrin homology domain; transformation; ERK

Introduction

Intracellular signaling proteins play important roles mediating a variety of cellular processes including growth, differentiation and apoptosis in response to extracellular stimuli. Proteins with conserved domains such as Src-homology-2 (SH2), Src-homology-3 (SH3) and PH usually act as key regulators in signal transduction (Pawson and Scott, 1997). While proteins with SH2 were explicitly bound to pTyr and involved in tyrosine kinase signaling pathways, SH3 had high affinity to proline-rich peptides. Interestingly, through the binding of PH to either phospholipids (Harlan et al., 1994; Ferguson et al., 1995; Kavran et al., 1998) or membrane-associated proteins (Touhara et al., 1994; Tsukuda et al., 1994), the association between PH-containing proteins and cell membrane was promoted (Falasca et al., 1998). To date, the prominent PH-containing molecules include GAPs, GRFs, Akt/PKB, PLC, and -adrenergic receptor kinase (ARK) (Musacchio et al., 1993).

Eps8 was initially identified as a substrate for EGF receptor (EGFR) whose tyrosyl phosphorylation was increased in response to EGF (Fazioli et al., 1993). Further studies indicated that Eps8 could also be phosphorylated by other receptor tyrosine kinases (Fazioli et al., 1993) as well as Src tyrosine kinase (Gallo et al., 1997; Maa et al., 1999). Proteins with molecular weight of 97 and 68 kDa recognized by Eps8 antibody have been referred to as the two Eps8 isoforms (p97^Eps8 and p68^Eps8) (Fazioli et al., 1993). Although p68^Eps8 has been speculated as a proteolytic or an alternatively spliced product of p97^Eps8, its exact coding sequences remains to be established (Fazioli et al., 1993).

Protein sequence analysis revealed the following domains of p97^Eps8 that might contribute to its cellular functions: an SH3 domain, a putative nuclear targeting sequence, a split PH domain, several proline-rich regions, and a degenerated SH2 domain at the N-terminus ((Wong et al., 1994; Fazioli et al., 1993). Previously, the proline-rich Eps8 had been demonstrated to interact with the SH3 domain of Src (Maa et al., 1999) and the SH3 of Eps8 was reported to associate with Shc (Matoskova et al., 1995), Shb (Karlsson et al., 1995), RN-tre (Matoskova et al., 1996), and E3b1 (Biesova et al., 1997). Unlike the canonical X-proline-X-X-proline (XPXXP) consensus, the SH3 domain of Eps8 was bound to a consensus sequence of proline-X-X-aspartate-tyrosine (PXXDY), implicating the existence of a novel family of SH3-containing proteins (Mongiovi et al., 1999). In addition to those Eps8-interacting proteins just described above, EGFR was also observed to associate with Eps8 basic region via its juxtamembrane region (Castagnino et al., 1995). Moreover, two Eps8 proteins were shown to form an interwined dimer with their SH3 domains being located at the interface (Kishan et al., 1997).

Low level of Eps8 was detected in the resting cell. However, the cellular level of Eps8 can be greatly enhanced in response to stimulations with serum, phorbol esters and the expression of v-Src (Gallo et al., 1997; Maa et al., 1999). Though Eps8 had been identified as a substrate for receptor and nonreceptor tyrosine kinases, its involvement in cell signaling was still unclear. Recent studies conducted with Eps8 -/- fibroblasts revealed that Eps8 was a mediator between Ras and Rac (Scita et al., 1999). Through its SH3 domain, Eps8 was demonstrated in vivo to form a ternary complex with E3b1 and Sos-1. Furthermore, this ternary complex exhibited Rac-specific guanine nucleotide exchange factor (GEF) activity in vitro. Since both membrane ruffling and actin reorganization were defective in Eps8 -/- fibroblasts stimulated by growth factors (i.e. PDGF and EGF) but not by 12-O-tetradecanoylphorbal-13-acetate, Eps8 was thus shown to act specifically on the growth factor-dependent actin remodeling (Scita et al., 1999). However, no abnormality was detected in Eps8-null mice, suggesting the existence of Eps8-related gene(s) (Scita et al., 1999).

Both the expression and tyrosyl phosphorylation of Eps8 are enhanced in v-Src transformed cells (Maa et al., 1999). Since p97^Eps8 has been implicated in EGF-induced mitogenesis and tumorigenesis (Fazioli et al., 1993; Matoskova et al., 1995), the enhancement of its expression and tyrosyl phosphorylation observed in v-Src transformed cells may contribute to the transformation mechanisms mediated by Src. To substantiate the hypothesis that overexpression of p97^Eps8 could lead to cellular transformation, murine C3H10T1/2 fibroblasts overexpressing p97^Eps8 or its PH-truncated variant were generated by retroviral infection method in this study. Compared to normal parental cells, cells overexpressing p97^Eps8 became transformed, since they could cause focus formation in culture dish and tumors in mice. By further studying the mechanisms underlying the p97^Eps8-induced transformation, we observed an increased serum-induced ERK activation in cells overexpressing p97^Eps8 relative to control cells. In addition, coexpression of dominant negative MEK1 was able to block the p97^Eps8-induced transformation. In contrast, PH-truncated p97^Eps8 was unable to localize at the cellular membrane; it also failed to potentiate ERK activation and to cause cell transformation. With these findings, we concluded that p97^Eps8 was indeed an oncogenic protein whose abnormal expression could lead to ERK activation that, in turn, could give rise to cellular transformation.

Results

Generation of cells overexpressing p97^Eps8 or N-terminal PH truncated variant 261-p97^Eps8

In studies aimed at showing p97^Eps8 to be an oncogenic protein and assessing the significance of its PH domain in p97^Eps8-induced transformation, control cells (Vec) and cells that individually overexpressing p97^Eps8 (Wt) or N-terminal PH-truncated variant 261-p97^Eps8 (261) were generated by retroviral infection method as described in Materials and methods (Figure 1a). To determine their Eps8 expression levels, the detergent lysates were immunoprecipitated with Eps8 antibody (C-eps8) that specifically recognized the C-terminal portion of p97^Eps8. The immunoprecipitated proteins were resolved by SDS-PAGE and immunoblotted with Eps8 antibody. The results showed that 261-p97^Eps8 migrated as a 68 kDa protein, and relative to control cells (Vec), significantly increased amounts of p97^Eps8 and 261-p97^Eps8 were detected in Wt and 261 cells, respectively (Figure 1b).

Cells overexpressing p97^Eps8, but not 261-p97^Eps8, were capable to cause cellular transformation

By comparison of the growth property of Vec, Wt, and 261 cells, Figure 2 shows that Wt cells can be transformed as characterized by the focus formation in the culture dish, while the control Vec cells can not (Figure 2). The transformation efficiency of Wt cells was estimated to be greater than 0.006%. Interestingly, 261 cells did not exhibit such a transformed focus (Figure 2). To further analyse the tumorigenic ability of p97^Eps8 overexpressors, the cells recovered from p97^Eps8-induced foci were inoculated into mice. In the meantime, cells expressing v-Src (IV5) were also inoculated into mice as a control. As presented in Table 1, all mice inoculated with p97^Eps8 overexpressors have developed tumors with diameter of 0.5 cm within 20-30 weeks (Table 1) and similar results have also been observed in mice inoculated with other independent clonal p97^Eps8 overexpressors (data not shown). It should be noted that animals inoculated with cells expressing v-Src (IV5) can also produce tumors of similar size in less than two weeks (Table 1). These findings indicate that overexpression of p97^Eps8 can lead to cellular transformation in vivo, though it appears to be far less potent than that of v-Src.

Potentiation of serum-induced ERK activation in cells overexpressing p97^Eps8, but not 261-p97^Eps8

Since ERK activation was implicated in cell proliferation and tumorigenesis (Troppmair et al., 1994), we wondered whether p97^Eps8-induced tumor formation was mediated by this mechanism. To test this hypothesis, the basal and the serum-induced ERK activities in control cells and cells overexpressing p97^Eps8 or 261-97^Eps8 were determined. Since MEK-mediated ERK phosphorylation on residues Thr-202 and Tyr-204 was known to increase the enzymatic activity of ERK (Sturgill et al., 1988; Payne et al., 1991), we first applied monoclonal antibody specifically recognized these phosphorylated residues of ERK in Western immunoblotting. As shown in Figure 3A, a significantly increased level of phosphorylated ERK was detected in Wt cells as compared to Vec and 261 cells when the expression of ERK in these cell lines was normalized. Consistently, measurements of radioisotope activity of ³²P-incorporated myelin basic protein (³²P-MBP) demonstrated that there was about twofold increase of serum-induced enzymatic activation of ERK in Wt cells relative to that found in Vec and 261 cells (Figure 3b). However, no significant difference of serum-induced enzymatic activation of JNK or p38 MAPK was detected in Wt, Vec and 261 cells as assayed by Western immunoblotting with antibody specifically recognized the active form of JNK or p38 MAPK (data not shown). Based on these findings, we speculated that the p97^Eps8-induced transformation could, at least in part, be attributed to the activation of ERK.

Expression of dominant negative MEK1 abrogated p97^Eps8-induced transformation

Since ERK activation was speculated to be involved in the cellular transformation as induced by p97^Eps8, we therefore transfected plasmid expressing Myc-tagged p97^Eps8 (pCMV-myc-eps8) into C3H10T1/2 fibroblasts with control plasmid (pCMV) or plasmid encoding dominant negative MEK1 [pCHA-MEK1(K97A)]. As shown in Table 2, in the presence of dominant negative MEK1, the number of transformed focus was significantly decreased in p97^Eps8 overexpressing cells. To further dissect the signaling pathways that potentiate serum-induced ERK activation in cells overexpressing p97^Eps8, levels of ERK activation in serum-stimulated Vec, 261 and Wt cells that were pretreated with or without PD98059 or LY294002 (a PI3K inhibitor) were determined. As shown in Figure 4, we found that PD98059 blocked serum-dependent ERK activation in a dose dependent manner in all the cells tested. However, the reduced level of activated ERK was far more apparent in Wt cells than Vec or 261 cells. Interestingly, LY294002 could also abolish serum-induced ERK activation in Wt cells as compared to Vec and 261 cells (Figure 4). These results can be taken as evidence to support the notion that p97^Eps8-induced ERK activation and transformation were mediated by MEK1. Moreover, our results also imply the possible participation of PI3K in this process.

PH truncated-p97^Eps8 failed to localize at the plasma membrane

Since PI3K might be involved in p97^Eps8-induced ERK activation and since the products of PI3K, phosphatidylinositol-3,4-bisphosphate (PtdIns-3,4-P₂) and phosphatidylinositol-3,4,5-triphosphate (PtdIns-3,4,5-P₃), were known to facilitate the PH-containing proteins to associate with plasma membrane, we therefore speculated that the inability of PH-truncated p97^Eps8 to activate ERK might be due to its failure to localize at the plasma membrane. To address this point, cells expressing Myc-tagged p97^Eps8 (myc-Wt) or 261-p97^Eps8 (myc-261) were examined. In the absence of serum stimulation, both Myc-p97^Eps8 and Myc-261-p97^Eps8 were mainly localized in the cytoplasm especially at the perinuclear region (data not shown). However, in the presence of serum, Myc-p97^Eps8 was able to translocate to the plasma membrane while Myc-261-p97^Eps8 remained in the cytosol (Figure 5). These findings indicated that PH domain might be important for the attachment of p97^Eps8 to the plasma membrane as well as ERK activation in response to serum stimulation.

Discussion

In this communication, we attempted to study the biological consequence resulting from p97^Eps8 overexpression. By subjecting C3H10T1/2 fibroblasts to viral infection, we obtained p97^Eps8 overexpressors. Interestingly, we observed that p97^Eps8 overexpressors exhibited the ability to produce focus formation in culture dish and the tumor in mice. These findings supported the notion that an abnormal expression of p97^Eps8 could lead to the cellular transformation. To dissect the mechanisms underlying this effect, the activity of ERK, JNK and p38 MAPK in response to serum stimulation was further analysed in both control cells and p97^Eps8 overexpressors. Our data indicated that only the ERK kinase activity was significantly enhanced in serum-treated p97^Eps8 overexpressing cells. Interestingly, cells overexpressing PH-truncated p97^Eps8 failed to develop the transforming phenotype; in addition, they did not exhibit serum-induced ERK activation. Since overexpression of ERK was necessary for the cellular transformation (Troppmair et al., 1994), these results implicated that an absolute increase of p97^Eps8 leading to transformation might be associated with serum-induced ERK activation. Indeed, expression of dominant negative MEK1 blocked p97^Eps8-induced focus formation. Since serum contains multiple mitogens (i.e. lysophosphatidic acid (LPA), insulin, PDGF) and cells treated with these factors all exhibit ERK activation, a clear identification of the responsible component(s) in serum and the elucidation of the ERK activation mechanisms in p97^Eps8 overexpressors become important issues.

PH domain has been demonstrated to mediate the protein-protein interaction and the protein-phospholipid interaction, thus contributing to the process of membrane anchorage that is important in signal transduction. It is now clear that through the binding of its PH domain to PtdIns-3,4-P₂, Akt/PKB is recruited to the plasma membrane where it becomes phosphorylated and enzymatically active (Franke et al., 1997; Frech et al., 1997; Klippel et al., 1997). Similarly, studies carried out in Schlessinger's laboratory also indicate that in addition to tyrosyl phosphorylation, PtdIns-3,4,5-P₃ generated by PI3K activation can interact with the PH domain of PLC and target it to the plasma membrane in response to growth factor stimulation (Falasca et al., 1998). However, it should be noted here that unlike the intact PH present in Akt/PKB, PH in PLC is a split one that seems to retain the ability to mediate membrane targeting. And indeed like PLC, the PH of p97^Eps8 is split and does not deteriorate its linkage to the plasma membrane (Provenzano et al., 1998). Data present in this communication show that PH-truncated p97^Eps8 fails to localize at the plasma membrane in serum-stimulated cells. These results thus imply that the PH domain of p97^Eps8 is indeed critical for membrane association and serum-induced ERK activation. Since LY294002 can also reduce serum-induced ERK activation in cells overexpressing p97^Eps8 (Figure 4), it seems that PI3K is most likely involved in the cellular transformation caused by p97^Eps8 overexpression. Further study is needed to address this point.

Enrichment of p97^Eps8 in membrane ruffles and lamellipodia in serum stimulated Swiss 3T3 fibroblasts implied that p97^Eps8 might play a role in cytoskeletal reorganization (Provenzano et al., 1998). Indeed, by studying p97^Eps8 +/+ and -/- immortalized cell lines, Scita et al. (1999) provided evidence indicating that p97^Eps8 was involved in actin reorganization and membrane ruffles in response to growth factor stimulation. And interestingly, expression of activated Ras and PI3K could induce ruffling in +/+ cells, but not in -/- cells. Since ruffling could be restored by microinjection of a DNA construct encoding activated Ras or PI 3-kinase together with p97^Eps8 into -/- cells, p97^Eps8 therefore participated in Ras/PI3K-induced Rac activation. In this pathway, p97^Eps8 has been shown to form a ternary complex with E3b1 and Sos1, and this complex can exhibit Rac-specific GEF activity. According to all the published p97^Eps8-related papers and the findings demonstrated in this study, it seemed clear that p97^Eps8 might be involved in at least two signaling pathways, one leading to mitogenesis and the other leading to membrane ruffling.

Enhanced tyrosyl phosphorylation level of p97^Eps8 was detected in cells stimulated with growth factors (i.e. EGF, PDGF) or cells transformed with v-Src. And studies carried out by Matoskova et al. (1995) further revealed the constitutive tyrosyl phosphorylation of p97^Eps8 in a panel of human tumor cell lines. However, to date, the implication of tyrosyl phosphorylated p97^Eps8 in mitogenesis and tumor formation is still obscure. In this study, we found that the expression level of p97^Eps8 was an important determinant in tumorigenic process. And consistent with this hypothesis was our observation of p97^Eps8 overexpression in a variety of breast tumor cell lines (data not shown). Thus it becomes an important issue to understand how the regulation of p97^Eps8 expression is achieved. Since overexpression of p97^Eps8 could contribute to the malignant phenotype, one could expect that p97^Eps8 has the potential to become a tumor marker in diagnosis and intervening target in cancer therapy.

Materials and methods

Plasmid DNA

The entire open reading frame of the mouse eps8 cDNA (i.e. nt -246 to -2708) was generated by PCR amplification method utilizing the pBS-eps8 DNA (Maa et al., 1999) as the template and the oligonucleotides tagged with either BamHI (P5) or EcoRI (P4) as the primers. P5 and P4 corresponded to nt -246 to -263 and nt -2705 to -2720 respectively. Then the BamHI- and EcoRI-digested PCR product was subcloned into the mammalian expression vector pBabe (Morgenstern et al., 1990) that was pre-treated with BamHI and EcoRI to give rise to pBabe-eps8. The eps8 DNA sequence in pBabe-eps8 was further confirmed by automated DNA sequencing. In addition, this PCR product was also subcloned into pCMV-Tag3 vector (Stratagene) to generate Myc-tagged p97^Eps8-expressing plasmid, pCMV-myc-eps8. For construction of the expressing plasmids encoding N-terminal PH-truncated p97^Eps8 and its Myc-tagged derivative, the same PCR reaction as described above was carried out except the EcoRI-tagged P4 and BamHI-tagged P6 (that corresponded to nt -1026 and -1042) were used as primers. And the resulting plasmids that encoded N-terminal PH-truncated p97^Eps8 and its Myc-tagged derivative were designated as pBabe-eps8 (261) and pCMV-myc-eps8 (261) respectively.

Cell lines and drug treatment

To generate control cells (Vec), and cells overexpressing wt p97^Eps8 (Wt) or N-terminal PH-truncated p97^Eps8 (261), the retrovirus transfection method was utilized. Briefly, the ecotropic packaging cells ( NX-Eco cells, generously provided by Dr Garry P Norlan at Stanford University; Costa et al., 2000) were either transfected with pBabe, pBabe-eps8 or pBabe-eps8 (261) DNA by calcium phosphate transfection method. The cultured medium that contained high titer of virus particles was collected and applied to murine C3H10T1/2 fibroblasts followed by 2 week puromycin (2 mg/ml) selection. These cells were maintained in Dulbecco modified essential medium (DMEM) supplemented with 10% fetal bovine serum and puromycin (1 mg/ml). For generation of cells expressing Myc-N-terminal PH-truncated p97^Eps8 (Myc-261-p97^Eps8) or Myc-p97^Eps8 with or without dominant negative MEK1, C3H10T1/2 cells were transfected with plasmids of either pCMV-myc-eps8 (261) or pCMV-myc-eps8 plus or minus pCHA-MEK1(K97A) (kindly provided by Dr Michael Weber) by Lipofectamine method (BRL).

For serum stimulation, Vec, Wt and 261 cells were grown to confluence on 100 mm diameter dishes. Then, cells were cultured in serum-free DMEM for 2 h followed by 10 min incubation with 10% fetal bovine serum without changing the medium. For drug treatments, the serum-starved cells were further incubated with or without PD98059 (3 and 10 M) or LY294002 (10 M) for another 30 min prior to the addition of optimal amount of fetal bovine serum. These serum-stimulated or -nonstimulated cells were then harvested and analysed.

Antibodies

p97^Eps8 specific antibody (C-eps8) was raised in rabbits against the bacterially expressed GST fusion protein containing the C-terminal p97^Eps8 (residues 355-668; Maa et al., 1999). The ERK2-specific polyclonal antibody TR2 was a generous gift from Dr Michael Weber in University of Virginia (Reuter et al., 1995). The E10 monoclonal antibody recognized phosphorylated Thr-202 and Tyr-204 of ERK was purchased from New England Biolabs, Inc. (Beverly, MA, USA). The monclonal antibody recognized Myc (9E10) and the rabbit polyclonal antibodies recognized unphosphorylated ERK2 were purchased from Santa Cruz.

Focus formation analysis

Approximately 5´10⁵ cells of control (Vec) and cells overexpressing wt p97^Eps8 (Wt) or its N-terminal PH-truncated mutant (261) were cultured in 60 mm dishes in triplicate. The cells were grown to confluence for around 30 days with complete medium changed every 2 days. The transformed foci were counted after cells were stained with modified Giemsa Stain (Sigma).

Tumorigenicity

Approximately 10⁷ cells in 0.2 ml of sterile PBS were injected subcutaneously in each hip of 28 to 42 day old mice (i.e. C3H/HeN). The animals were checked twice every week and the tumors were measured with the aid of a micrometer.

Immunoprecipitation and immunoblot analysis

Lysis of the cells was carried out with modified RIPA buffer as described before (Maa et al., 1999) and protein concentration was determined by protein assay kit (Bio-Rad). The Eps8 expression level of these cell lines was determined by Western immunoblotting with polyclonal anti-Eps8 antibody (C-eps8) or mouse monoclonal anti-Eps8 antibody (Transduction). Methods for immunoprecipitation and immunoblotting analysis have been described (Maa et al., 1999). For detection of p97^Eps8, ERK and its phosphorylated derivatives, western immunoblots were performed with respective antibody followed by ¹²⁵I-protein A or horseradish peroxidase-conjugated protein A and detected by autoradiography or Enhanced Chemiluminescence method (Amersham, Arlington Heights, ILL, USA) respectively.

ERK kinase activity assay

Endogenous ERK kinase activity was determined in the ERK immunocomplexes as described previously (Reuter et al., 1995). Briefly, near 1 mg of cellular lysates were immunoprecipitated by rabbit polyclonal antiserum TR2. After three washes with RIPA buffer and two washes with kinase reaction buffer (50 mM HEPES (pH 7.5), 10 mM magnesium acetate), two-fifths of the immunocomplexes were resolved in an SDS-PAGE and Western immunoblotted with monoclonal antibody against ERK. The other three-fifths were aliquoted into three parts and performed in vitro kinase reaction in the presence of 40 l kinase reaction buffer supplemented with 10 g myelin basic protein (MBP; Sigma), 20 M ATP and 10 Ci [-³²P]-ATP for 20 min at 30°C. The reaction was stopped by adding into equal volume of 2´Lamli sample buffer followed by 5 min boiling. The ³²P-incorporated MBP was resolved by SDS-PAGE, dried down, and visualized by autoradiography. Then, the ³²P-labelled MBP excised from the gel was measured by Cherenkov counting.

Confocal microscopy

Cells overexpressing Myc-tagged wt- and N-terminal PH-truncated p97^Eps8 were seeded on glass coverslips and grown overnight. In preparation for microscopy, monolayers were treated as described previously (Maa et al., 1992) and immunostained with primary monoclonal antibody against Myc (9E10) followed by the secondary FITC-conjugated goat anti-mouse IgG (Jackson Immuno Research Laboratories). Confocal analysis was carried out with MRC-1000 confocal imaging system (Bio-Rad) at the Department of Pharmacology, National Cheng Kung University.

Acknowledgements

We thank Dr Sally Parsons and Dr Michael Weber for the pBabe (puro) expression vector and polyclonal ERK antibody (TR2) respectively, Dr Garry P Norlan and Dr Sue Lin-Chao for the virus-packaging cell line NX-Eco and Dr Ching-Hsine Huang and Dr Hsin-Fang Yang-Yen for their critical comments on this manuscript. We appreciate the technical help provided by Ya-Chun Chuang, Hsaio-Hwai Chen and Hong-Ming Lin. We are also grateful to the animal center at NCKU for the care of animals. This work was supported by National Science Council Grant to M-C Maa (NSC 89-2320-B-006-007) and grants of NHRI to T-H Leu (DOH-87-HR-625, NHRI-GT-EX89S932L).

Biesova Z, Piccoli C and Wong WT. (1997). Oncogene 14, 233-241. MEDLINE

Castagnino P, Biesowa Z, Won WT, Fazioli F, Gill GN and Di Fiore PP. (1995). Oncogene 10, 723-729. MEDLINE

Falasca M, Logan S.K, Lehto VP, Baccante G, Lemmon MA and Schlessinger J. (1998). EMBO J. 17, 414-422. Article MEDLINE

Fazioli F, Minichiello L, Matoska V, Castagnino P, Miki T, Wong WT and Di Fiore PP. (1993). EMBO J. 12, 3799-3808. MEDLINE

Ferguson KM, Lemmon MA, Schlessinger J and Sigler PB. (1995). Cell 83, 1037-1046. MEDLINE

Franke TF, Kaplan DR, Cantley LC and Toker A. (1997). Science 275, 665-668. Article MEDLINE

Frech M, Andjelkovic M, Ingley E, Reddy KK, Falck JR and Hemmings BA. (1997). J. Biol. Chem. 272, 8474-8471. MEDLINE

Gallo R, Provenzano C, Carbone R, Di Fiore PP, Castellani L, Falcone G and Alema S. (1997). Oncogene 15, 1929-1936. MEDLINE

Harlan JE, Hajduk PJ, Yoon HS and Fesik SW. (1994). Nature 371, 168-170. MEDLINE

Karlsson T, Songyang Z, Landgren E, Lavergne C, Di Fiore PP, Anafi M, Pawson T, Cantley LC, Claesson-Welsh L and Welsh M. (1995). Oncogene 10, 1475-1483. MEDLINE

Costa GL, Benson JM, Seroogy CM, Achacoso P, Fathman CG and Nolan GP. (2000). J. Immunol. 164, 3581-3590. MEDLINE

Kavran JM, Klein DE, Lee A, Falasca M, Isakoff SJ, Skolnik EY and Lemmon MA. (1998). J. Biol. Chem. 273, 30497-30508. Article MEDLINE

Kishan KV, Scita G, Wong WT, Di Fiore PP and Newcomer ME. (1997). Nat. Struct. Biol. 4, 739-743. MEDLINE

Klippel A, Kavanaugh WM, Pot D and Williams LT. (1997). Mol. Cell. Biol. 17, 338-344. MEDLINE

Maa M-C, Wilson LK, Moeyers JS, Vines RR, Parsons JT and Parsons SJ. (1992). Oncogene 7, 2429-2438. MEDLINE

Maa M-C, Lai J-R, Lin R-W and Leu T-H. (1999). Biochim. Biophys. Acta 1450, 341-351. MEDLINE

Matoskova B, Wong WT, Salcini AE, Pelicci PG and Di Fiore PP. (1995). Mol. Cell. Biol. 15, 3805-3812. MEDLINE

Matoskova B, Wong WT, Nomura N, Robbins KC and Di Fiore PP. (1996). Oncogene 12, 2679-2688. MEDLINE

Mongiovi AM, Romano PR, Panni S, Mendoza M, Wong WT, Musacchio A, Cesareni G and Di Fiore PP. (1999). EMBO J. 18, 5300-5309. Article MEDLINE

Morgenstern JP and Land H. (1990). Nucl. Acids Res. 18, 3587-3596.

Musacchio A, Gibson T, Rice P, Thompson J and Saraste M. (1993). TIBS 18, 343-348. MEDLINE

Pawson T and Scott JD. (1997). Science 278, 2075-2080. Article MEDLINE

Payne DM, Rossomando AJ, Martino P, Erickson AK, Her JH, Shabanowitz J, Hunt DF, Weber MJ and Sturgill TW. (1991). EMBO J. 10, 885-892. MEDLINE

Provenzano C, Gallo R, Carbone R, Di Fiore PP, Falcone G, Castellani L and Alema S. (1998). Exp. Cell Res. 242, 186-200. Article MEDLINE

Reuter CWM, Catling AD and Weber MJ. (1995). Meth. Enzymol. 255, 245-256. MEDLINE

Scita G, Nordstrom J, Carbone R, Tenca P, Glardina G, Gutkind S, Bjarnegard M, Betsholtz C and Di Fiore PP. (1999). Nature 401, 290-293. Article MEDLINE

Sturgill TW, Ray LB, Erikson E and Maller JL. (1988). Nature 334, 715-718. MEDLINE

Touhara K, Inglese J, Pitcher JA, Shaw G and Lefkowitz RJ. (1994). J. Biol. Chem. 269, 10217-10220. MEDLINE

Troppmair J, Bruder JT, Munoz H, Lloyd PA, Kyriakis J, Banerjee P, Avruch J and Rapp UR. (1994). J. Biol. Chem. 269, 7030-7035. MEDLINE

Tsukuda S, Simon MI, Witte ON and Katz A. (1994). Proc. Natl. Acad. Sci. USA 91, 11256-11260. MEDLINE

Wong WT, Carlomagno F, Druck T, Barletta C, Croce CM, Huebner K, Kraus MH and Di Fiore PP. (1994). Oncogene 9, 3057-3061. MEDLINE

Figure 1 The expression of wt- and N-terminal PH-truncated p97^Eps8 in C3H10T1/2 fibroblasts. (a) Schematic representation of p97^Eps8 and its N-terminal PH truncated derivative (261-p97^Eps8). NLS: nuclear localization signal. (b) p97^Eps8 was immunoprecipitated from equal amounts of lysates harvested from Vec, Wt or 261 cells by C-eps8 antibody, resolved in an SDS-PAGE and analysed by Western immunoblotting with monoclonal Eps8 antibody purchased from Transduction Laboratories (Vec: control cells; Wt: cells overexpressing wt p97^Eps8; 261: cells overexpressing 261-p97^Eps8). The position of 261-p97^Eps8 was marked with an asterisk

Figure 2 Enhanced foci formation in cells overexpressing p97^Eps8. Equal numbers of cells overexpressing p97^Eps8 (Wt) and its N-terminal PH truncated mutant (261) and control cells (Vec) were cultured for around 4 weeks. The transformed foci were observed as described in Materials and methods

Figure 3 Enhanced serum-induced ERK activation in cells overexpressing p97^Eps8. (a) Equal amounts of lysates prepared from nonstimulated or serum-stimulated Vec, 261, and Wt cells as described in Materials and methods were resolved in an SDS-PAGE and examined by Western immunoblotting with ERK2 or phosphorylated ERK1/2 antibody (E10) respectively. (b) The same lysates as described above were immunoprecipitated with ERK polyclonal antiserum (TR2). Two-fifths of the immunocomplexes were resolved in an SDS-PAGE, transferred to nitrocellulose membrane and probed with ERK2 antibody (upper panel). The other three-fifths were aliquoted into three parts and subjected to an in vitro kinase assay in the presence of myelin basic protein (MBP) and [-³²P]ATP. The ³²P-labeled MBP was resolved by SDS-PAGE and detected by autoradiography (bottom panel)

Figure 4 Abolishment of serum-induced ERK activation in PD98059 and LY 294002 treated p97^Eps8 overexpressors. Vec, 261 and Wt cells were treated as above except that PD98059 (3 or 10 M) or LY 294002 (10 M) was added and incubated for 30 min prior to serum stimulation. Equal amounts of lysates were resolved by SDS-PAGE and examined by immunoblot analysis with ERK2 or phosphorylated ERK1/2 antibody (E10)

Figure 5 PH truncated-p97^Eps8 fails to localize at the plasma membrane in response to serum stimulation. Cells overexpressing Myc-tagged wt- and N-terminal PH truncated-p97^Eps8 (designated as myc-Wt and myc-261 respectively) were grown on coverslips. After serum stimulation, cells were sequentially immunostained with Myc monoclonal antibody (9E10) and FITC-conjugated Goat anti-mouse IgG as described in Materials and methods. Bar, 25 m. At least 20 cells were observed and the representative one was shown

Table 1 Tumorigenicity of p97^Eps8 overexpressors recovered from p97^Eps8-induced foci

Table 2 Transforming activity of the eps8 cDNA