Regulation of the tyrosine kinase substrate Eps8 expression ... - Nature

Oncogene (1997) 15, 1929 ± 1936  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Regulation of the tyrosine kinase substrate Eps8 expression by growth factors, v-Src and terminal dierentiation Rita Gallo1, Claudia Provenzano1, Roberta Carbone2, Pier Paolo Di Fiore,2,3, Loriana Castellani4, Germana Falcone1 and Stefano AlemaÁ1 1

Istituto di Biologia Cellulare, C.N.R., 00137 Roma, 2European Institute of Oncology, 20141 Milano, 3Istituto di Microbiologia, UniversitaÁ di Bari, 70124 Bari, and 4Dipartimento di Medicina Sperimentale e Scienze Biochimiche, UniversitaÁ di Roma Tor Vergata, 00133 Roma, Italy

SH3-containing proteins are involved in signal transduction by a number of growth factor receptors and in the organization of the cytoskeleton. The recently identi®ed Eps8 protein, which contains an SH3 domain, is coupled functionally and physically to the EGFR and is tyrosine phosphorylated by this receptor and other receptors as well. Here, we examined the regulation of eps8 expression in response to mitogenic or dierentiative signals. We show that Eps8 is expressed at low levels in resting ®broblasts, but its expression is strongly induced during activation by serum, phorbol esters and the v-src oncogene. Conversely, expression of Eps8, but not of other EGFR substrates such as Shc or Eps15, is virtually extinguished in non-proliferating, terminally dierentiated murine myogenic cells. The putative role of Eps8 protein as a v-Src substrate was analysed in murine ®broblasts and in quail myogenic cells expressing a temperature-sensitive variant of the tyrosine kinase. Tyrosine phosphorylation of Eps8 was detected only at the permissive temperature. A non-myristylated, transformation-defective mutant of v-Src did not phosphorylate Eps8, whereas it phosphorylated Shc. Together, these ®ndings indicate that Eps8 may be a critical substrate of v-Src. They further establish Eps8 as an example of a signal transducer whose expression senses the balance between growth and dierentiation and might, therefore, be involved in the determination of the phenotype. Keywords: tyrosine kinase; Eps8; Shc; dierentiation; signal transduction

Introduction Eps8 is a recently identi®ed substrate for the epidermal growth factor receptor (EGFR) and other receptor tyrosine kinases (RTKs) (Fazioli et al., 1993a; Wong et al., 1994). The product of the eps8 gene exhibits several unique structural features (Fazioli et al., 1993a; Wong et al., 1994). The carboxy terminal half of the protein contains an SH3 domain that interacts with other signalling proteins, including Shb (Karlsson et al., 1995), Shc (Matoskova et al., 1995) and RN-tre (Matoskova et al., 1996). An eector role for the SH3 domain of Eps8 is suggested by the ®nding that it is able to induce maturation of Xenopus oocytes or transcription directed by a fos promoter in mammalian Correspondence: S AlemaÁ Received 3 March 1997; revised 13 June 1997; accepted 16 June 1997

cells (Fazioli et al., 1993a; G Scita and PP Di Fiore, unpublished observations). The amino terminal half comprises several proline-rich regions, which are potential sites for interaction with SH3 domains. At variance with other signal transducers which bind in a ligand-dependent manner through their SH2 domains to phosphotyrosine-containing motifs, Eps8 binds stably to the juxtamembrane region of EGFR, regardless of activation of the latter (Castagnino et al., 1995). Other lines of evidence have implicated Eps8 in the regulation of cell proliferation, since its overexpression in EGFR-expressing cells enhances mitogenic responsiveness to EGF (Fazioli et al., 1993a) and Eps8 is constitutively tyrosine-phosphorylated in human tumour cell lines (Matoskova et al., 1995). There is currently a resurgence of interest in elucidating the downstream eectors of tyrosine kinases of the src family. Much recent evidence has made clear that, despite the fact that receptor and nonreceptor tyrosine kinases share several signalling pathways, there are nonetheless signi®cant dierences in the modalities of recruitment of common transducers. In addition, a number of substrates for tyrosine phosphorylation appear to be uniquely recruited by non-receptor tyrosine kinases (reviewed in Brown and Cooper, 1996). The v-src oncogene encodes a protein tyrosine kinase which is necessary and sucient to transform target cells infected with Rous sarcoma virus (RSV) (Wyke and Stoker, 1987; Parsons and Weber, 1989). Moreover, v-src, when expressed under a strong promoter, is one of the few oncogenes capable of inducing eciently and in a single-step the complete conversion of a normal into a transformed cell (Hjelle et al., 1998). The sustained activation of the v-Src tyrosine kinase results in the tyrosine phosphorylation of numerous intracellular targets, many of which are associated with the cellular cytoskeletal network (Brown and Cooper, 1996). The transforming function of v-src also depends on speci®c regulators of its enzymatic activity as well as on other gene products acting as downstream transducers of the oncoprotein activity (Brown and Cooper, 1996). The function of these not yet fully de®ned activities is important in determining the susceptibility to transformation of dierent cell types, as demonstrated by the observation that some cells are easily transformed by v-src (AlemaÁ and TatoÁ, 1987; Boettiger, 1989; Falcone et al., 1992), others are resistant (Lipsich et al., 1984) while others undergo terminal dierentiation (AlemaÁ et al., 1985). In the present study, we address the question of whether the Eps8 protein is a biologically relevant

Eps8 and v-Src transformation R Gallo et al

+TPA

FCS

39° 4D

39°

35° 2D

c

39° 2D

Low serum

b

35°

a

2D+FCS 2D

Eps8 protein has been shown to be expressed in a variety of cell types, including epithelial and fibroblastic lines and some hematopoietic cells (Fazioli et al., 1993a; Wong et al., 1994; Matoskova et al., 1995). To determine whether the levels of expression of eps8 could be modulated by mitogenic stimuli, we examined the expression of the Eps8 protein by Western blot analysis of cell lysates derived from growing and quiescent Balb/c and NIH3T3 cells. It has been previously shown that an anti-Eps8 serum speci®cally recognises 97 kDa and 68 kDa forms in mouse ®broblasts (Fazioli et al., 1993a). Figure 1 shows that indeed the eps8 gene product presents two forms of 97 and 68 kDa in Balb/c 3T3 cells kept in 10% serum (Figure 1a, lane 1). Upon reduction of serum to 0.5% for 2-4 days, the expression of the Eps8 protein is highly reduced in Balb/c 3T3 cells (Figure 1a) or in NIH3T3 cells (not shown). Under the same culture conditions the rate of entry into S phase of Balb/c 3T3 cells, assayed by a 5 h pulse with bromo-deoxyuridine (BrdU), was 64% in 10% FCS and 3.6% in cells cultured for 2 days in 0.5% FCS. The levels of accumulation of the Eps8 protein could be restored to those observed in growing ®broblasts by stimulating resting cells with 10% serum for 2 days (70% BrdUpositive cells) (Figure 1a, lane 4). Exposure of growing cells to the tumour promoter phorbol 12-0-tetradecanoyl phorbol 13-acetate (TPA) for 2 days resulted in a pronounced up-regulation of Eps8 protein in Balb/c 3T3 (Figure 1a, lane 5) and NIH3T3 ®broblasts (Figure 1c). The speci®c eect of culture conditions on the levels of expression of Eps8 was con®rmed by monitoring the expression levels of both Shc (Pelicci et al., 1992) and Eps15 (Fazioli et al., 1993b) proteins, which remained mostly unchanged (Figure 1). In an attempt to demonstrate that the up-regulation of Eps8 is a phenomenon observable in growing cells and possibly common to transformed cells, lysates of various oncogene-transformed NIH3T3 cells were analysed by Western blot for relative levels of the Eps8 protein. Whereas in growing NIH3T3 cells, the predominant form of Eps8 is the 97 kDa species, the

To determine whether accumulation of Eps8 proteins is directly regulated by v-src and is not an indirect consequence of transformation, we assayed the steadystate levels of Eps8 following the activation of a ts-vsrc mutant in a representative clonal strain of Balb/c 3T3 ®broblasts transformed with the MR31 retrovirus. When Balb/c 3T3-MR31 cells grown as subcon¯uent cultures at the restrictive temperature (398C) were shifted for two days to the permissive temperature (358C) for v-Src kinase activity, we observed a progressive increase in Eps8 protein accumulation up to levels comparable to those displayed by cells continuously kept at 358C (Figure 1b). Expression of the 68 kDa form was particularly up-regulated (about tenfold) both in Balb/c and NIH3T3 cells. This dierential accumulation was not in¯uenced by reactivation of v-Src in cells kept in low serum (not shown). Steady-state levels of Shc, Eps15 and cortactin proteins, which are known substrates of tyrosine kinases (McGlade et al., 1992; Fazioli et al., 1993b; Wu et al., 1991), were not aected signi®cantly under the same culture conditions (Figure 1b, cortactin not shown). This modulation occurred with a time course similar to that resulting in morphological transformation. Thus, v-Src aects the levels of Eps8 protein accumulation. Our observation that expression of Eps8 was upregulated in v-Src-transformed or TPA-treated cell

TPA 2D

Expression of eps8 is regulated by serum, tumour promoters and oncogenes

The Eps8 gene products are up-regulated by the v-src oncogene

4D

Results

accumulation of the 68 kDa form is strongly upregulated in cells transformed by erbB2, ras, raf and v-src oncogenes (not shown).

2D

intermediate of v-src-induced uncontrolled growth and block of dierentiation. We ®nd that Eps8 protein becomes tyrosine-phosphorylated at the permissive temperature in cells transformed by temperaturesensitive variants of v-Src. Tyrosine phosphorylation of Eps8 protein did not occur in cells expressing a nonmyristylated mutant of v-Src. We show further that the expression of Eps8 protein is directly regulated by vSrc and other mitogenic signals. The accumulation of Eps8 in dividing or transformed cells and the contrasting inhibition of its expression in postmitotic, terminally dierentiated muscle cells, strengthen the notion that Eps8 is a positive regulator of mitogenesis and transformation. Thus, Eps8 is an example of a signal transducer whose activity may be controlled by its levels of expression according to the growth or dierentiation conditions.

FCS

1930

Eps8

— 97 — 68

Shc

— 66 52 — — 46 — 142

Eps15 1

2 3 4 5 Balb/c 3T3

6

7

8

9

Balb/c 3T3-MR31

10 11 NIH 3T3

Figure 1 Regulation of Eps8 protein expression levels in normal and v-src-transformed Balb/c 3T3 cells. (a) Balb/c 3T3 cells were cultured on collagen-coated dishes in 10% FCS (lane 1) or in 0.5% FCS for 2 (lane 2) and 4 days (lane 3). Lane 4 refers to cells that had been maintained in 0.5% FCS for 2 days and then exposed to 10% FCS for two more days. Lane 5 refers to cells cultured in 10% FCS in the presence of 1077 M TPA for 2 days. (b) Balb/c 3T3 cells transformed by a temperature-sensitive mutant of v-Src (ts-MR31) were propagated in 10% FCS at either 358C (lane 6) or at 398C for 2 (lane 7) and 4 days (lane 9). Lane 8 refers to cells kept at 398C for 4 days and then shifted to 358C for 2 days. (c) NIH3T3 cells cultured in 10% FCS in the absence (lane 1) or the presence (lane 2) of 1077 M TPA for 2 days. Expression levels of Eps8, Shc and Eps15 proteins were determined by immunoblotting of total lysates


lines prompted us to investigate whether the regulation of Eps8 was exerted at the transcriptional level. To this end, total RNAs extracted from control and v-srctransformed Balb/c 3T3-MR31 cells grown at 358C and 398C and from control and TPA-treated NIH3T3 cells were analysed by Nothern blot (Figure 2). Two transcripts of approximately 4.7 and 4.0 kb were found in all cell types (Fazioli et al., 1993a). No signi®cant eect of temperature was evident in control Balb/c 3T3 cells, whereas both v-Src-transformation of Balb/c 3T3 cells and TPA-treatment of NIH3T3 cells greatly augmented the accumulation of both transcripts, in close agreement with that observed for the Eps8 proteins. A poor correlation between the ratios of the two transcripts and of the two protein forms, however, was observed in BalB/c 3T3 cells, possibly resulting from dierential rates of translation or processing. Silencing of eps8 upon terminal dierentiation of murine myogenic cells

NIH3T3+TPA

Tyrosine phosphorylation of Eps8 proteins in ts-v-src transformed ®broblasts and myoblasts

NIH3T3

Balb/c-MR31 39°

Balb/c-MR31 35°

Balb/c 39°

Balb/c 35°

Skeletal muscle terminal dierentiation entails the coordination of muscle-speci®c gene expression and withdrawal from the cell cycle. Because expression of eps8 is controlled by signals leading to mitogenesis or transformation we sought to determine whether expression of eps8 was regulated during differentiation. To this purpose we employed two systems of dierentiating murine myoblasts. Skeletal myogenesis in the C2C12 cell line and in primary satellite cells (MSC) can be induced by depriving cycling myoblasts of mitogens. This leads, within three days in culture, to the formation of multinucleated myotubes and to expression of a vast array of muscle-speci®c proteins. Expression of both the 97 kDa and 68 kDa proteins was strongly down-regulated during terminal differentiation of C2C12 cells, both in the absence and in the

— 28S Eps8

GAPDH

1

2

3

4

presence of the DNA synthesis inhibitor cytosine b-Darabinofuranoside (Ara-C) to eliminate undifferentiated cells, as well as in primary murine satellite cells (MSC) (Figure 3a and b). The decrease in the amount of both forms of Eps8 paralleled the increase in accumulation of the muscle dierentiation marker myosin heavy chain (MHC). Levels of Eps15 and Shc proteins were not signi®cantly aected by muscle dierentiation, with the exception of the 66 kDa form of Shc which was markedly up-regulated concomitantly with the attainment of the post-mitotic state. Because dierentiated myotubes do not re-enter the cell cycle in response to mitogen stimulation, we investigated whether eps8 silencing persisted under these conditions. Indeed, the down-regulation of Eps8 levels was maintained in fully dierentiated MSC after readdition of serum for 2 days (Figure 3b, lane 4). We also monitored the expression of Eps8 in a transformed line of C2C12 murine myoblasts that stably expresses a ts-v-src mutant gene (C2C12 tsMR31) and exhibits temperature-dependent transformation and block of dierentiation (MC Gauzzi and SA., unpublished). At high serum concentrations (GM) transformed C2C12 tsMR31 myoblasts expressed both forms of Eps8 in larger amounts than proliferating parental C2C12 cells (Figure 3c, compare lane 1 and 2). In the absence of serum (DM) at the permissive temperature the accumulation of Eps8 was in part reduced (Figure 3c, lane 3), although it remained higher than in parental cells kept in GM (Figure 3c, lane 1), consistent with the upregulation of Eps8 exerted by v-Src. Upon terminal dierentiation at the restrictive temperature, however, expression of Eps8 was virtually extinguished (Figure 3c, lane 4), and could only be resumed marginally by the addition of mitogens to post-mitotic myotubes or activation of v-Src upon shift to 358C (Figure 3c, lanes 5 and 6).

5

6

Figure 2 Accumulation of eps8 gene transcripts is modulated by TPA and transformation by v-src. RNAs (12 mg/lane) extracted from the following cells were examined by Northern analysis: Balb/c 3T3 (lanes 1 and 2) and Balb/c 3T3-MR31 (lane 3 and 4) cells grown at 358C (lanes 1 and 3) or 398C (lanes 2 and 4); control (lane 5) or TPA-treated (lane 6) NIH3T3 cells. Note that the 4.0 kb mRNA species in control cells only becomes visible with longer exposures. Total RNAs were harvested from cells in semicon¯uent cultures in 10% FCS. Filters were hybridised with probes for eps8 and GAPDH

To investigate tyrosine phosphorylation of Eps8 proteins, Balb/c 3T3-MR31 or NIH3T3-MR31 cells were kept either at the permissive temperature or at the restrictive temperature for 2 days. The tyrosine kinase v-Src was activated in cultures kept at 398C by temperature shift to 358C for 2 and 24 h. When lysates from these cells were immunoprecipitated with anti-Eps8 serum, bands of 97 kDa and 68 kDa were detected in blots that reacted with anti-phosphotyrosine antibody as early as 2 h after v-Src activation (Figure 4a), suggesting that both forms of Eps8 can be tyrosine-phosphorylated in the presence of an active vSrc. Comparable levels of tyrosine phosphorylation of Eps8 were detected both in high and low serum concentrations (not shown). We next estimated the percentage of Eps8 molecules that are tyrosinephosphorylated in vivo in both NIH3T3- and Balb/c 3T3-MR31 cells. To this end we adopted a previously described methodology (Matoskova et al., 1995). We estimated that about 10% of the total Eps8 pool was constitutively tyrosine-phosphorylated at the permissive temperature in both cell types, whereas about 5% of the total pool was phosphorylated following 24 h of v-Src activation in cells previously kept at the restrictive temperature for 2 days (not shown)

1931


35°DM 39°

39° DM+FCS

39° DM

DM 3D+FCS

DM 4D

DM 3D

GM

3D+FCS

5D

3D

DM 3D

GM

DM/AraC

35° DM

c

b

35° GM

a

C2C12-35° GM

1932

— 97

Eps8

— 68

MHC

— 200 1

2

3

4

MSC

— 66 — 52 — 46

Shc

Eps15

— 142 1

2

3

4

5

C2C12

1

2

3

4

5

6

C2C12 tsMR31

Figure 3 Expression of Eps8 is extinguished during muscle dierentiation. (a) Western blot analysis of total lysates from C2C12 myoblasts grown at 378C in GM (lane 1), dierentiated in DM for 3 (lanes 2, 3 and 5) or 5 days (lane 4) in the absence (lane 2) or in the presence of 50 mM Ara C (lanes 3, 4 and 5) to eliminate residual undierentiated cells. Lane 5 refers to cells that had been maintained in DM for 3 days and then exposed to 10% FCS for 48 h. (b) Western blot analysis of total lysates from primary mouse satellite cells (MSC) maintained in GM (lane 1) or in DM for 3 (lane 2) and 4 (lane 3) days. Lane 4 refers to MSC maintained in DM for 3 days and then exposed to 10% FCS for 24 h. (c) Western blot analysis of total lysates from control C2C12 myoblasts at 358C in GM (lane 1) and from C2C12 tsMR31 cells (lanes 2 ± 6) maintained at 358C for 2 days in GM (lane 2) or in DM (lane 3) and dierentiated at 398C in DM (lanes 4 ± 6). Lanes 5 and 6 refer to cells dierentiated in DM at 398C for 3 days and then either challenged with 10% FCS for 24 h (lane 5) or shifted to 358C for 24 h (lane 6). Amounts of Eps8, myosin heavy chain (MHC), Shc and Eps15 proteins were determined

To further test the ability of vSrc to induce tyrosine phosphorylation of the Eps8 protein and since the available antibodies do not recognise avian Eps8 proteins, Eps8 was constitutively or transiently expressed in quail myoblasts transformed with ts-mutants of v-Src (QMb-LA29). We have previously shown that in these cells the level of P-Tyr-containing cellular proteins is temperature-dependent both in myoblasts and in myotubes shifted to the permissive temperature for v-Srs (AlemaÁ and TatoÁ, 1994; Castellani et al., 1995, 1996). Since structural modi®cations of speci®c domains of v-Src have been shown to alter the tyrosine phosphorylation of v-Src substrates (Kamps and Sefton, 1986; Wu et al., 1991; Fincham et al., 1995; Verderame et al., 1995; Brown and Cooper, 1996) we also sought to determine whether a membrane localisation of v-Src was required for tyrosine phosphorylation of Eps8 and Shc proteins. To this end we employed quail myoblasts infected by a non-myristylated, transformation-defective version of the ts-LA29 v-Src protein (QMb-LA29A2) (Catling et al., 1993). QMb-LA29 and -LA29A2 were transiently transfected with an eps8 expression vector and then either kept at 358C or allowed to dierentiate at 418C for 2 days. A third set of cells after dierentiation at 418C was shifted to 358C for 4 h. Both QMb-LA29 and -LA29A2 expressed v-Src mutant proteins at levels easily detectable by immuno¯uorescence analysis with the anti-Src mAb 327 and

sucient to induce high levels of phosphorylation of a number of cellular substrates (not shown). Cell lysates were immunoprecipitated with anti-Eps8 or anti-Shc antibodies and the resolved immunoprecipitates were blotted with anti-p-Tyr antibody. Phosphorylation of Eps8 was strictly temperature-dependent in QMb-LA29transformed myoblasts irrespective of Eps8 being expressed transiently (Figure 4b, left panel) or constitutively (not shown). Tyrosine phosphorylation of Eps8 was detected as early as 30 min after shifting cells to the permissive temperature and was very evident after 4 h (Figure 4b), when major morphological alterations become visible (Castellani et al., 1995). Under the same conditions, tyrosine phosphorylation of the 52 kDa Shc protein was also observable after activation of v-Src (Figure 4c, left panel). Note that although quail Shc proteins migrate faster than their mammalian counterparts and the antibody used identi®ed preferentially the 52 kDa form, we have retained the canonic designation (McGlade et al., 1992; see Verderame et al., 1995). As shown in Figure 4b (right panel), Western blot analysis showed no detectable tyrosine-phosphorylated Eps8 protein in extracts from cells expressing the nonmyristylated form of v-Src. In contrast, tyrosinephosphorylation of Shc was observed under all conditions whereby the tyrosine kinase was active, regardless of its membrane association (Figure 4c, right panel).


Discussion

35°/2 hrs

35°/24 hrs

39°

39°

35°/24 hrs 39°

39°

35°/2 hrs 39°

39°

35°

a

35°

The Eps8 protein is a recently identi®ed tyrosine kinase substrate that carries an SH3 domain and binds to

— 97

Blot: α Eps8

— 68 — 97 — 68 2

3

4

6

7

41°

1

41°DM

Blot: α P-Tyr 5

8

35°/4 hrs

35°GM

41°

41°DM

35°GM

b

35°/4 hrs

IP: α Eps8

Blot: α Eps8

— 97

Blot: α P-Tyr

— 97 1

2

3

4

5

6

35°/4 hrs 41°

41°DM

35°GM

41°

41°DM

35°GM

c

35°/4 hrs

IP: α Eps8

Blot: α Shc

— 52 — 46

Blot: α P-Tyr

— 52 — 46 1

2

3

4

5

6

IP: α Shc Figure 4 Phosphorylation of Eps8 and Shc proteins in v-srctransformed cells. (a) ts-MR31-transformed Balb/c 3T3 (left panel) and NIH3T3 cells (right panel) were grown in 10% FCS at 358C (lanes 1 and 5) or at 398C for 2 days (lanes 2 and 6) and subsequently shifted to 358C for 2 (lanes 3 and 7) and 24 h (lanes 4 and 8). Cells were lysed and immunoprecipitated with anti-Eps8 serum. Immunoprecipitates were immunoblotted with anti-Eps8 or anti-phosphotyrosine (P-Tyr) antibodies. (b and c) Quail myoblasts transformed by the ts-LA29 strain of Rous sarcoma virus (QMb-LA29) (lanes 1 ± 3, left panels) or expressing a nonmyristylated, transformation-defective version of the v-Src kinase (QMb-LA29A2) (lanes 4 ± 6, right panels) were transiently transfected with pCEV-eps8 vector, propagated at 358C in GM (lanes 1 and 4) or dierentiated at 418C in DM for 2 days (lanes 2 and 5) and then shifted to 358C in DM for 4 h (lanes 3 and 6). Cell were lysed and immunoprecipitated with anti-Eps8 (b) or anti-Shc (c) antibodies. Immunoprecipitates were immunoblotted with anti-phosphotyrosine (P-Tyr) and with anti-Eps8 (b) or antiShc antibodies (c). Mobilities of the Eps8 and Shc proteins are indicated. Note that in b, left panel, the 97 kDa Eps8 protein appears as a closely-spaced doublet

EGFRs in a ligand-independent fashion. Overexpression of Eps8 in ®broblasts enhances EGF-dependent mitogenic signals (Fazioli et al., 1993a) and promotes a transformed phenotype (Matoskova et al., 1995). Furthermore, Eps8 and the signal transducer Shc are constitutively tyrosine-phosphorylated in many tumorigenic cell lines, suggesting a key role in neoplastic growth for these intracellular transducers (Pelicci et al., 1995; Matoskova et al., 1995). Here, we have demonstrated that Eps8 expression is highly regulated in murine mesenchymal cells. Whereas resting 3T3 cells express very low levels of Eps8, we found that mitogenic activators such as serum, TPA and selected oncoproteins, notably the v-Src tyrosine kinase, all induce Eps8 expression. We have shown that eps8 transcripts accumulate in TPA-treated or v-Src-transformed 3T3 ®broblasts at levels comparable to those attained by Eps8 proteins. Possible mechanisms for the overall increase in mRNA accumulation include increased rate of transcription or enhanced transcript stability. We also observed a striking relative increase of the 4.0 kb mRNA species, which correlated with increased levels of expression of the 68 kDa protein form in the same cells. Whereas at present we can only speculate as to how the two Eps8 proteins are generated, a satisfactory explanation is that the two mRNA species arise from alternative splicing and code separately for the 97 kDa and 68 kDa forms. The alternative possibility that the 68 kDa species represents a speci®c cleavage product of the 97 kDa species was addressed by forced expression of a 97 kDa Eps8-myc isoform in Balb/c 3T3-MR31 cells. Eps8-myc levels were not up-regulated by activation of the ts-allele of v-src, nor was a 68 kDa species detected (unpublished data), and hence, in all probability, the 68 kDa species is unlikely to result from post-translational modi®cations. In addition, we demonstrate that expression of Eps8 in both primary and established murine myoblasts is turned o following mitogen removal and onset of terminal dierentiation, suggesting that Eps8 may function in myoblasts to prevent them from commitment to permanent withdrawal from the cell cycle. Murine myogenic cells and the signals that induce their dierentiation are, so far, unique in their ability to suppress expression of Eps8. It will be interesting to see whether this kind of regulation operates in differentiated cell types belonging to other lineages. Whereas levels of expression of the 52 and 46 kDa Shc proteins were not in¯uenced by growth conditions or transformation, the 66 kDa isoform was markedly upregulated in muscle cells concomitantly with the attainment of the post-mitotic state. The sequence of the 66 kDa form overlaps that of the 52 kDa isoform and contains a unique N-terminal region. Unlike the other two protein species the 66 kDa form does not increase EGF activation of MAP kinases and inhibits fos promoter activation (Migliaccio et al., 1997). Regulation of this isoform expression is, therefore, expected to in¯uence the cellular response to growth factors in dierent cell types. Eps8 appears to belong to the class of v-Src substrates that are also phosphorylated by activated RTKs and are most likely involved in mitogenic signalling pathways. A second class of substrates includes proteins that are speci®cally phosphorylated

1933


1934

in v-src-transformed cells and these are often associated to focal adhesions and the cytoskeleton (Brown and Cooper, 1996). Regardless of its belonging to either class, several lines of evidence support the contention that Eps8 may be a critical substrate of vSrc. First, tyrosine-phosphorylated Eps8 is a relatively abundant protein in v-Src-transformed cells. The extent of Eps8 tyrosine-phosphorylation by v-Src (5 ± 10% of total Eps8 protein pool) is comparable to that found in human tumour cell lines (Matoskova et al., 1995), or after activation of overexpressed RTKs by ligands (Fazioli et al., 1993a; Matoskova et al., 1995). Second, Eps8 proteins are phosphorylated in a transformationspeci®c manner. The tyrosine-phosphorylation of a number of identi®ed v-Src substrates, notably talin, p36, vinculin and cortactin, is poorly correlated with transformation by various mutants of v-Src kinase. Other substrates, however, such as FAK and pp120cas satisfy the criterion of not being phosphorylated in cells expressing the transformation-defective, nonmyristylated mutant of the kinase (Brown and Cooper, 1996). Eps8 was not phosphorylated in avian cells expressing non-myristylated v-Src but was phosphorylated in both murine and avian cells transformed by ts-v-src. Moreover, Eps8 proteins were readily phosphorylated after a temperature shift to the permissive temperature. Shc proteins, on the other hand, were tyrosine-phosphorylated in cells expressing the transformation-defective non-myristylated allele of v-Src, thus suggesting that phosphorylation of this substrate is not sucient for cell transformation, although it may be required. Accordingly, in chicken cells expressing host-range or transformation-defective v-src alleles, p52shc was highly phosphorylated on tyrosine whenever a kinase-active src is expressed, regardless of whether the cells are transformed or not (Verderame et al., 1995). Is Eps8 directly phosphorylated by v-Src? The observation that phosphorylation of Eps8 takes place with at least the same kinetics of other substrates may be relevant to this point. Tyrosine phosphorylation, however, is apparently uncoupled from direct binding of v-Src to Eps8 proteins. In immunoprecipitation experiments we could neither isolate a v-Src/Eps8 complex, nor did v-Src and Eps8 interact in vitro via their SH3 domains (our unpublished results). While the lack of a stable detergent-resistant association between Eps8 and v-Src does not preclude their contact in vivo, it also leaves the possibility open that v-Src does not directly phosphorylate Eps8 but does so through other cellular tyrosine kinases which are activated in v-srctransformed cells, perhaps through autocrine secretion of growth factors. However, we note that Eps8 is phosphorylated by v-Src with comparable eciency in low-serum, arguing against the possibility that its phosphorylation requires the cooperation of kinases activated by mitogens present in serum. p190, a GTPase activator of Rho, and Shc are examples of substrates that display high levels of tyrosine phosphorylation in v-src-transformed cells, yet have not been shown to construct stable detergent-resistant interactions with the kinase (Brown and Cooper, 1996). Alternatively, it is possible that Eps8 is made accessible to v-Src phosphorylation following interaction with other intermediate proteins. A signi®cant fraction of the Eps8 protein pool is found associated to the plasma

membrane, co-localising with F-actin and the F-actin binding protein cortactin. Furthermore, both Eps8 and cortactin are readily re-located to podosomes and to the actin bodies (our unpublished results) generated upon activation of v-Src in quail myotubes (Castellani et al., 1995, 1996). These observations and previous work (Wu and Parsons, 1993; Okamura and Resh, 1995) suggest a similar subcellular localisation of Eps8, cortactin and v-Src in cytoskeleton-associated structures. Although it is still not clear whether Eps8 is a direct substrate of v-Src, the proximity of the two proteins at speci®c cellular districts suggests that this may be the case. The ®nding that Eps8 protein was not phosphorylated in non-myristylated mutant-infected cells, is therefore expected on the basis of the subcellular localisation of the mutant v-Src protein. The potential link between phosphorylation of Eps8 and the morphological changes associated with cell transformation is currently under investigation. In conclusion, our results indicate that Eps8 may be one of the critical signal transduction molecules associated with and possibly required in the transmission of the signals generated by v-Src, that lead to cell transformation as well as to block of myogenic dierentiation. The reported ability of Eps8 in modulating cell proliferation, the association of its tyrosine phsphorylation with the transformed state and the induction of its expression in response to growth promoting stimuli, converge to present it as a highly suitable candidate to link a variety of growthregulatory pathways, including cell dierentiation.

Materials and methods Materials 12-0-tetradecanoyl phorbol 13-acetate (TPA) and a monoclonal antibody to phosphotyrosine (PT-66) were from Sigma; anti-Src mAb 327 was kindly provided by J Brugge and anti-cortactin mAb 4F11 by T Parsons; antiMyc mAb 9E10 was a gift from A Cattaneo; a polyclonal anti-Shc was from Transduction Labs; polyclonal sera speci®c for the eps8 and eps15 gene products were generated as previously described (Fazioli et al., 1993a and b); a polyclonal anti-skeletal muscle myosin serum was developed in house using puri®ed chicken muscle myosin as immunogen. Cell culture and transfection NIH3T3 cells (provided by E Westin), Balb/c 3T3 cells and al v-src transformants derived from them were maintained in Eagle's Minimal Essential Medium (MEM) supplemented with 10% FCS and grown in 5% CO2 at 378C. NIHerbB2, NIH-ras, and NIH-raf cells (kindly provided by O Segatto) were maintained in Dulbecco's modi®ed Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS). C2C12 myoblasts were maintained in DMEM supplemented with 20% FCS (growth medium, GM). Primary murine satellite cells (MSC) were provided by M Crescenzi and were maintained in DMEM containing 10% FCS and 20 ng/ml ®broblast growth factor. Muscle dierentiation was induced by incubating cultures at 70% con¯uence in DMEM containing 10 mg of insulin and 5 mg of transferrin per ml (dierentiation medium, DM) for three days. Clonal strains of C2C12 myogenic cells and NIH3T3 and Balb/c 3T3 ®broblasts expressing v-src were obtained by infection with the retroviral construct MR31,


carrying the ts-src mutant gene from LA31-RSV (kindly provided by A La Rocca), followed by selection of transformed colonies in soft agar at the permissive temperature. A clonal strain of Balb/c-MR31 cells stably expressing an Eps8-myc-tagged protein was established by transfection of an eps8-myc expression vector derived by myc-epitope tagging of the Eps8 protein encoded by the previously described pCEV-eps8 vector (Fazioli et al., 1993a) and selection with G418 (0.8 mg/ml). Polyclonal population of ts-src-transformed quail myoblasts were established from primary cultures infected with viral stocks of ts-LA29, a temperature-sensitive mutant of RSV, as previously described (QMb-LA29) (Falcone et al., 1991; AlemaÁ and TatoÁ, 1994). Quail myoblasts infected by a non-myristylated, transformation-defective version of the tsLA29 v-Src protein (RCAN-29A2) (Catling et al., 1993) were kindly provided by F TatoÁ. Transformed myoblasts stably expressing Eps8 protein (QMb-LA29-eps8) were obtained by transfection of QMb-LA29 cells with pCEV-eps8 and selection in G418 (0.8 mg/ml). Infected quail myoblasts were propagated on collagen-coated dishes in DMEM supplemented with 10% FCS, 10% tryptose phosphate broth, 1% chicken serum, referred to as growth medium (GM) at 358C (permissive temperature). Dierentiation was induced by plating 105 cells on 35 mm collagen-coated dishes in GM and, the following day, by shifting the cultures to 418C (restrictive temperature) for 2 ± 3 days in F14 medium, supplemented with 2% foetal calf serum (referred to as dierentiation medium, DM); for reactivation of ts-Src in dierentiated myotubes, a parallel set of plates kept at 418C was shifted to 358C for the appropriate lengths of time. Transient transfections in v-Src-transformed quail myoblasts (1 mg of DNA per 35 mm plate or 5 mg per 100 mm plate) were carried out by using an optimised calcium phosphate transfection procedure (Chen and Okayama, 1987). The proliferation rate of growing and serum-derived Balb/ c 3T3 cells was assayed by measuring the percentage of nuclei which incorporated BrdU. 105 cells were plated in duplicate on 35 mm dishes in MEM-10% FCS and the day after shifted to fresh MEM-10% FCS or MEM-0.5% FCS for 2 days. Cells were then fed for 5 h with the same media also containing 10 mM BrdU. BrdU-labelled cultures were ®xed for 15 min with ice-cld 95% ethanol/5% acetic acid, treated for 10 min with 1.5 N HC1, rinsed in PBS and stained with an anti-BrdU monoclonal antibody (Amersham) for 1 h.

ing a 1.2 Kb avian GAPDH (glyceraldehyde-3-phosphatedehydrogenase) cDNA (obtained from C Schneider). Quantitative analysis was carried out using a Molecular Dynamics PhosphorImager 400A with Image Quant version 3.2 software. Whole-cell extracts and Western blot analysis Cells were brie¯y rinsed with phosphate-buered saline (PBS), containing 0.5 m M orthovanadate and collected with 0.2 ml/35 mm plate of SDS sample buer (8 M Urea, 0.14 M b-mercaptoethanol, 0.04 M DTT, 2% SDS, 0.075 M Tris-Cl, pH 8.0). SDS ± PAGE (2 ± 20 mg of total proteins per well) was carried out according to Laemmli 1970. Western blots were carried out as described by Towbin et al., (1979) using horseradish peroxidaseconjugated goat anti-rabbit and anti-mouse antibodies and revealed using the ECL chemiluminescence detection system (Amersham). Analysis of Eps8 and Shc tyrosine phosphorylation Immunoprecipitations of Eps8 and Shc proteins were performed as described (Fazioli et al., 1993a; Matoskova et al., 1995). Brie¯y, cultures were washed twice with PBS and stored frozen at 7808C. Lysis was performed in 1.5 ml/100 mm dish of lysis buer (30 mM Tris-Cl pH 7.5, 0.1 M NaC1, 1% Triton X100, 2 mM MgCl2, 1.5 mM EGTA, 10 mM NaF, 1 mM orthovanadate, 40 mg/ml leupeptin, 40 mg/ml aprotinin, 40 mg/ml soybean trypsin inhibitor, 1 mM PMSF, 10% glycerol) and cell extracts were centrifuged at 15 000 g for 15 min at 48C. Cell extracts (0.5 ± 2 mg of total proteins), normalised for protein content by the BCA colorimetric method, were incubated with either polyclonal anti-Eps8 serum or polyclonal anti-Shc serum for four hours and 50 ml of 50% protein A-agarose slush for 1 h at 48C. Immunoprecipitates were washed twice with lysis buer, once with 20 mM Tris-Cl pH 7.5 and boiled in 50 ml of SDS sample buer. Immunoprecipitates were analysed on SDS ± PAGE, and blotted with speci®c antibodies as indicated. Sequential anti-phosphotyrosine immunoprecipitations for quantitation of phosphorylation were carried out as described in Matoskova et al., 1995)

RNA isolation and Northern blot ananlysis Total RNA was prepared by the Ultraspec RNA isolation system (Biotecx). 12 mg aliquots of the obtained RNA were resolved in 0.9% agarose/2.2 M formaldehyde gels. Transfer to nitrocellulose membranes and high stringency hybridisation were carried out according to standard procedures. Probes were labelled with a random primed DNA labelling kit. For detection of transcripts, inserts of the following plasmids were cut with the appropriate restriction enzymes and used as probes: pCEV-eps8, containing a 3.6 Kb mouse eps8 cDNA a plasmid contain-

Acknowledgements We would like to thank Oreste Segatto, Giuliana Pelicci, Thomas Parsons, Joan Brugge, Marco Crescenzi, Anna La Rocca and Franco TatoÁ for generous gifts of reagents, and Delio Mercanti for skilful assistance in antibody production. RG was supported by fellowships from CNR and Fondazione A Buzzati-Traverso. CP is an AIRC fellow. This work was supported by grants from Comitato Promotore Telethon, CNR (PF-ACRO), AIRC and European Community (BIOMED-2)

References AlemaÁ S and TatoÁ F. (1987) Adv. Cancer Res., 49, 1 ± 28. AlemaÁ S and TatoÁ F. (1994). Sem. Cancer Biol., 5, 147 ± 156. AlemaÁ S, Casalbore P, Agostini E and TatoÁ F. (1985). Nature, 316, 5579 ± 559. Boettiger D. (1989). Curr. Topics Microbiol. Immunol., 147, 31 ± 78. Brown MT and Cooper JA. (1996). Biochim Biophys. Acta, 1286, 121 ± 149. Castagnino P, Biesowa Z, Wong WT, Fazioli F, Gill GN and Di Fiore PP. (1995). Oncogene, 10, 723 ± 729.

Castellani L. Reedy MC. Gauzzi MC, Provenzano C, AlemaÁ S and Falcone G. (1995). J. Cell Biol., 130, 871 ± 885. Castellani L, Reedy M, Airey JA, Gallo R, Ciotti MT, Falcone G and AlemaÁ. (1996). J. Cell Sci., 109, 1335 ± 1346. Catling AD, Wyke JA and Frame MC. (1993). Oncogene, 8, 1875 ± 1886. Chen C and Okayama H. (1987). Mol. Cell. Biol., 7, 2745 ± 2752.

1935


1936

Falcone G, AlemaÁ S and TatoÁ F. (1991). Mol. Cell. Biol., 11, 3331 ± 3338. Falcone G, Provenzano C, AlemaÁ S and TatoÁ F. (1992) Oncogene, 7, 1913 ± 1920. Fazioli F, Minichiello L, Matoska V, Castagnino P, Miki T, Wong WT and Di Fiore PP. (1993a). EMBO J., 12, 3799 ± 3808. Fazioli F, Minichiello L, Matoskova B, Wong WT and Di Fiore PP. (1993b) Mol. Cell. Biol., 13, 5814 ± 5828. Fincham VJ, Wyke JA and Frame MC. (1995). Oncogene, 10, 2247 ± 2252. Hjelle B, Liu E and Bishop JM. (1988). Proc. Natl. Acad. Sci. USA, 85, 4355 ± 4359. Karlsson T, Songyang Z, Landgren E, Lavergen C, Di Fiore PP, Ana® M, Pawson T, Cantley LC, Claesson-Welsh L and Welsh M. (1995). Oncogene, 10, 1475 ± 1483. Kamps MP, Buss JE and Sefton BM. (1986) Cell, 45, 105 ± 112. Laemmli UK. (1970). Nature, 227, 680 ± 685. Lipsich L, Brugge JS and Boettiger D. (1984). Mol. Cell. Biol., 4, 1420 ± 1424. Matoskova B, Wong WT, Salcini AE, Pelicci PG and Di Fiore PP. (1995). Mol. Cell. Biol., 15, 3805 ± 3812. Matoskova B, Wong WT, Nomura N, Robbins KC and Di Fiore PP. (1996). Oncogene, 12, 2679 ± 2688. McGlade J, Cheng A, Pelicci G, Pelicci PG and Pawson T. (1992). Proc. Natl. Acad. Sci. USA, 89, 8869 ± 8873.

Migliaccio E, Mele S, Salcini AE, Pelicci G, Lai K-M, Superti-Gurga G, Pawson T, Di Fiore PP, Lanfrancone L and Pelicci PG. (1997). EMBO J., 16, 706 ± 716. Okamura H and Resh MD. (1995). J. Biol. Chem., 270, 26613 ± 26618. Pelicci G, Lanfrancone L, Grignani F, McGlade J, Cavallo F, Forni G, Nicoletti I, Grignani F, Pawson T and Pelicci PG. (1992) Cell, 70, 93 ± 104. Pelicci G, Lanfrancone G, Salcini AE, Romano A, Mele S, Borrello MG, Segatto O, Di Fiore PP and Pelicci PG. (1995). Oncogene, 11, 899 ± 907. Parsons JT and Weber MJ. (1989). Curr. Top. Microbiol. Immunol., 147, 79 ± 127. Towbin H, Staehelin T and Gordon J. (1979). Proc. Natl. Acad. Sci. USA, 76, 4350 ± 4354. Verderame MF, Guan J-L and Woods Ignatoski KM. (1995). Mol. Biol. Cell, 6, 953 ± 966. Wong WT, Carlomagno F, Druck T, Barletta C, Croce CM, Huebner K, Kraus MH and Di Fiore PP. (1994). Oncogene, 9, 3057 ± 3061. Wu H and Parsons JT. (1993). J. Cell Biol.,120, 1417 ± 1426. Wu H, Reynolds AB, Kanner SB, Vines RR and Parsons JT. (1991). Mol. Cell. Biol., 11, 5113 ± 5124. Wyke JA and Stoker AW. (1987). Biochim. Biophys. Acta., 907, 47 ± 69.