regulate neu tyrosine kinase and cell transformation - NCBI - NIH

13 downloads 0 Views 2MB Size Report
Oct 27, 1988 - The neu oncogene,characterized by Weinberg and colleagues, is a ...... We thank Drs Robert Weinberg, Mien-Chie Hung and Towia Libermann.
The EMBO Journal vol.8 no. 1 pp. 1 59 - 166, 1989

A chimeric EGF-R neu proto-oncogene allows EGF regulate neu tyrosine kinase and cell transformation

Heikki Lehvislaiho1' 2, Laura Lehtola', Lea Sistonen1 and Kari Alitalol 'Department of Virology and Pathology and 2Transplantation Laboratory, University of Helsinki, Haartmaninkatu 3, SF-00290 Helsinki, Finland Communicated by A.Valheri

The neu oncogene, characterized by Weinberg and colleagues, is a transforming gene found in ethylnitrosourea-induced rat neuro/glioblastomas; its human protooncogene homologue has been termed erbB2 or HER2 because of its close homology with the epidermal growth factor receptor (EGF-R) gene (c-erbBl). Expression of the rat neu oncogene is sufficient for transformation of mouse NIH 3T3 fibroblasts in culture and for the development of mammary carcinomas in transgenic mice, but the neu proto-oncogene has not been associated with cell transformation. We constructed a vector for expression of a chimeric cDNA and hybrid protein consisting of the EGF-R extracellular, transmembrane and protein kinase C-substrate domains linked to the intracellular tyrosine kinase and carboxyl terminal domain of the rat neu cDNA. Upon transfection with the construct, NIH 3T3 cells gave rise to EGF-R antigen-positive cell clones with varying amounts of specific EGF binding. Immunofluorescence and immunoprecipitation using neu- and EGF-receptor specific antibodies demonstrated a correctly oriented and positioned chimeric EGF-R- neu protein of the expected apparent mol. wt on the surface of these cells. EGF or TGFa induced tyrosine phosphorylation of the chimeric receptor protein, stimulated DNA synthesis of EGF-R-neu expressing cells and led to a transformed cell morphology and growth in soft agar. In contrast, the neu proto-oncogene did not show kinase activity or transforming properties when expressed at similar levels in NIH 3T3 cells. These results suggest that the neu proto-oncogene possesses mitogenic and transforming properties only in the presence of a ligand which stimulates its tyrosine kinase activity and provides the first model for studies of the function of the neu tyrosine kinase. Key words: c-erbBl/erbB2/epidermal growth factor/signal transduction

Introduction A transforming gene in ethylnitrosourea (ENU)-induced rat neuro/glioblastomas first described by Shih et al. (1981) was named neu and characterized by Weinberg and colleagues

(Padhy et al., 1982; Schechter et al., 1984; Bargmann et al., 1986a; Hung et al., 1986; Stern et al., 1986). The neu proto-oncogene encodes a protein very similar to the epidermal growth factor receptor, and its human homologue has been termed erbB2 (King et al., 1985; Semba et al., IRL Press

to

1985; Yamamoto et al., 1986) or HER2 (Coussens et al., 1985). The factor binding to this putative growth factor receptor is unknown, a fact which has hampered progress in understanding the biology of the frequent amplifications of the erbB21neu gene in human mammary carcinomas and some other adenocarcinomas (Yokata et al., 1986; Slamon et al., 1987; Venter et al., 1987; Van de Vijver et al., 1987; Yokata et al., 1988). The neu gene is characteristically activated by an ENU-induced point mutation changing glutamic acid for valine in the transmembrane domain of the receptor protein in cells of the developing rat nervous system (Bargmann et al., 1986b). The human erbB2/HER2 gene is a potent oncogene when sufficiently overexpressed in NIH 3T3 cells, even in the absence of an identifiable ligand (Di Fiore et al., 1987a), but its normal rat neu counterpart may lack a similar transforming activity (Bargmann et al., 1986b; Bargmann and Weinberg, 1988a). On the other hand, overexpression of the EGF-R (c-erbBl) gene causes transformation of NIH 3T3 cells only in the presence of EGF (Di Fiore et al., 1987b; Velu et al., 1987; Riedel et al., 1988). We report here a model system to study the function of the neu tyrosine kinase. We were interested in the possibility that the cytoplasmic domain of the neu receptor might have a mitogenic and transforming function in the presence of a ligand that would stimulate its activity. We also wanted to confirm that the structural analogies and predictions about the biological similarities between the EGF-R and the neu receptor proteins would indeed be manifested in functional collaboration by their ligand binding and signal transducing domains. For these reasons, we constructed a hybrid EGF-R-neu cDNA in the pSV2 vector which has previously been shown to lack transforming potency in NIH 3T3 cells when equipped with the normal neu coding sequence (Bargmann et al., 1986b; Bargmann and Weinberg, 1988a). We then tested several properties of the NIH 3T3 cells carrying the chimeric construct.

Results Construction of the EGF-R - neu expression vector We constructed a vector for expression of a chimeric cDNA and hybrid protein consisting of the EGF-R extracellular, transmembrane and protein kinase C-substrate domains linked to the intracellular tyrosine kinase and carboxyl terminal domain of the rat neu cDNA. The chimeric receptor construct was made starting from the expression vector pSV2neuNT which contains the neu oncogene cDNA under the SV40 early promoter (Bargmann et al., 1986b). The coding sequences of neu upstream from the shared PstI site at nucleotide 2107 were removed and the EGF-R cDNA sequences upstream of the corresponding PstI site (at 2233; Ullrich et al., 1984) were ligated to the downstream neu cDNA sequences to give the plasmid pSV2EGFR/neu (Figure 1). Thus the site of the activating point mutation in 1 59

H.Lehvaslaiho et al.

EGF-R c-DNA

A

x P

P

P

I

I

I

I

I

i

.......

s P

x

I

I

I

pSV2 neuNT

.......

H p

(M)P

pEGFRXS

I

p

l

I

I

I

--

.......

im m

s

p

p

P P

PP

P

mmim

11

I

P I .......

....

.......

HPP

P

P

PP P 11 I

M I I EFI....... ...... _ p

.......

pSV2EGF-R/ neu

B

SV40 early prom.

CH

'3I

........... .

CH

....

dr o

PA

TK

TM

x . 'K%V %//////I...

I IE

EGF-R/nI

PstII CTC TTC ATG CGA AGG CGC CAC ATC GTT CGG AGG CGC ACG CTG CGG AGG CTG CTG CAG GAA ACT Leu Phe Met Arg Arg Arg His Ile Val Arg Lys Arg rLeu Arg Arg Leu LeuGln Glu Thr

nIr

PstI (2107) CTA ATC AAA CGA AGG AGA CAG AAG ATC CGG AAG TAT ACG ATG CGT AGG CTG CTG CAG GAA ACT Leu Ile Lys Arg Arg Arg Gln Lys Ile Arg Lys Tyr TrMet Arg Arg Leu Leu Gln Glu Thr

rT

**

EGF-R

*

*

*

*

I

CTC TTC ATG CGA AGG CGC CAC ATC GTT CGG AGG CGC ACG CTG CGG AGG CG CTG CAG GAG AGG Leu Phe Met Arg Arg Arg His Ile Val Arg Lys Arg Leu Arg Arg Leu Gin Glu Arg (2233)

Fig. 1. A. The construction of the EGF-R-neu expression vector. The ends of the full-length c-erbB1 cDNA in a XhoI fragment (a kind gift from Towia Libermann) were filled using the Klenow enzyme and recut with Sacl. This 3124 bp fragment was ligated to SniaI-SacI opened pGEM-3 blue vector (Promega Biotec). The resulting plasmid (pEGFRXS) was opened from the HindII site in the linker and digested partially with PstI to obtain a 2090 bp fragment that contained the extracellular, membrane-spanning and protein kinase C-substrate regions of the EGF-R cDNA. The corresponding region from plasmid pSV2neuNT (Bargmann et al., 1986b) was removed by opening with HindIH and partial PstI digestion. The remaining 6893 bp fragment was joined with the 2090 bp EGF-R 5' fragment to give plasmid pSV2EGF-R/neu. The correct joining of the ends was confirmed by sequencing. Abbreviations: H, HindIll; M, SniaI: P, Pstl: S, Sacl; X, XhoI. B. Schematic drawing of the EGF-R-neu expression vector showing the nucleotide and amino acid sequences from the transmembrane to the joining regions of the hybrid and parental constructs. Marked are the PstI sites used to cut and ligate the parent sequences. The cytoplasmic sides of the transmembrane domains of both proteins are marked on the left. The threonine encoded at residue 654 (boxed) has been shown to be the major substrate for protein kinase-C in the EGF-R (Hunter et al., 1985). Abbreviations: CH, cysteine-rich domain; TM, transmembrane domain; TK, tyrosine kinase; pA, polyadenylation signal.

the transmembrane domain of neu was excluded from the construct. The correct joining of the ends was confirmed

by sequencing using

a

synthetic oligonucleotide primer.

Expression of the chimeric EGF-R neu protein in transfected cells Initial testing of the construct by its expression in the COS-1 and CHO cells indicated the presence of the predicted hybrid protein (data not shown). For functional studies involving analysis of cell transformation, NIH 3T3 mouse cells were transfected with the construct together with the pSV2neo DNA (Southern and Berg, 1982) conferring neomycin resistance for the selection of the transfectants. Of 57 independent neomycin-resistant clonal cell lines tested, 36 proved to bind [ 5I]EGF significantly more than control cells. Three of these clones, NEN7, NEN16 and NEN49, were verified as separate clones by Southern blotting and hybridization analysis of their DNA with the neu cDNA, chosen for further studies and compared to parental NIH 3T3 cells and to a clone transfected only with the selection marker -

160

(NN). Besides the determination of specific [ '25I]EGF binding (Figure 2A), we measured the expression of the EGF-R antigenic determinants on the surface of the cells by indirect immunofluorescence with a monoclonal antibody against the EGF-R and fluorescence-activated cell sorting (FACS; Figure 2B). Most of the fluorescence for the EGF-R antigenic determinants was cell surface-associated also in detergent-permeabilized cells (Figure 2B, inset in the bottom panel), in accordance with the correct localization of the chimeric receptor. In addition, we stained the cells with our rabbit antibodies against the carboxyl terminal domain of the neu protein (Figure 2C). Control staining studies were carried out for A431 cells expressing 2 x 106 EGF-R per cell (Haigler et al., 1978), and with BT-474 (Lasfargues et al., 1978) and DHFR/G8 cells (Hung et al., 1986) expressing amplified erbB2 and neu genes respectively (data not shown). These analyses were consistent with the expression of the EGF-R-neu hybrid protein on the surface of the transfected cells. Furthermore, [125I]EGF competition experiments with nonradioactive EGF showed that the N7

EGF Regulation of neu tyrosine kinase

B

C NN

A

NE 6.. 300

L.

NEN16

C)

-0

400*

-

NE i-'q

) _ 3000

NEN49 2

200

NE .-

.n J

11(00

A431

NN4

NEN7 N

EN16 N

i

Fluorescence intensity

EN 49 t ErJEN7

Fig. 2. A. Analysis of expression of EGF-R-neu hybrid protein in transfected cells by ['25I]EGF binding. Neomycin-resistant (NN) cells and A431 human epidermoid carcinoma cells were used as control negative and positive cell lines respectively. The NIH 3T3 cells gave values similar to the NN cells. The results represent mean values from three parallel determinations. Variation between individual determinations was less than 10%. B. Analysis of expression of EGF-R-neu hybrid protein by indirect immunofluorescence and FACS. Cells grown to near confluency were detached with EDTA, stained in indirect surface immunofluorescence with the monoclonal antibody against the EGF-R (solid line) or control mouse and monoclonal antibodies (dotted line) and subjected to analysis in the FACS. The inset in the bottom panel shows NP-40 detergent-permeabilizedcell nonpermeabilized NEN7 cells stained with the EGF-R antibody (left and right-hand panels respectively). Note that virtually all fluorescence is surface-associated. C. Immunofluorescence photomicrographs of fixed, permeabilized cells stained in indirect immunofluorescence with the rabbit antibodies against the carboxyl terminal domain of the neu protein. Note that most of the fluorescence is associated with the surface of the cells. Similar results were obtained with antibodies against synthetic carboxyl terminal neu peptides (Gullick et al., 1987). The DHFR/G8 cells had a fluorescence intensity similar to the NEN49 cells. No fluorescence was obtained without permeabilization of the cells. The sides of the panels correspond to 280 gm.

cells express 2 x 106 specific EGF-binding sites per cell, whereas the N16 and N49 cells express -4 x 105 and 5 x 105 receptors per cell respectively (data not shown). -

Immunoprecipitation of the EGF-R neu receptor Antibodies against the neu carboxyl terminus precipitated a specific 190 000 mol. wt polypeptide from [35S]methionine-labelled NEN7, NEN16 and NEN49 cells (Figure 3A and data not shown). The p190 polypeptide was also precipitated with the anti-EGF-R antibodies, but this precipitation could not be blocked with the bacterial neu protein. These experiments confirmed the hybrid nature of the p190 as the protein product of the EGF-R-neu construct, and quantitative estimates of the levels of the chimeric protein produced by the different clones expressing p190 were consistent with the amounts of the chimeric receptor detected on the cell surface by EGF binding and FACS (compare data shown for clones NEN7 and NEN16 in Figures 2B and 3A). Parallel immunoprecipitations from the DHFR/G8 cells (Hung et al., 1986) which contain an experimentally amplified neu gene and protein showed the 185 000 mol. wt neu protein at levels comparable to the levels of the -

chimeric EGF-R -neu receptor in the NEN7 cells (Figure 3A). EGF induces tyrosyl phopshorylation of the chimeric EGF-R - neu protein Further immunoprecipitation experiments were carried out from [35S]methionine-labelled cells using affinity-purified anti-phosphotyrosine antibodies (Seki et al., 1986). Cells expressing either the chimeric receptor protein (NEN16, NEN49 and NEN7) or correspondingly amplified amounts of the rat neu protein (DHFR/G8) were labelled for 16 h. Then some of the labelled cell cultures were treated with 25 nM EGF while other duplicate cultures received solvent. Equal amounts of neu polypeptides were precipitated from the cells whether or not they were treated with EGF (Figure 3B). In contrast, anti-phosphotyrosine antibodies recognized the chimeric protein only from EGF- or TGFa-treated cells, except for the highly expressing NEN7 cells which contained small amounts of tyrosyl-phosphorylated neu polypeptides even in the absence of EGF (Figure 3B and data not shown). Recently, two groups have reported on the phosphorylation of the neu protein in trans by the EGF-R in rat-1 cells

161

H.Lehvaslaiho et al.

A

NENi p

antibody: 1CieJ

NEN1.6 G&

B

b

EGF-R

1fJLi 5

P .-T yr 0 5

__

_

0

GB-

NEN7

NEN16

antibody: E:GFF min)n

lei 0 1 5

P-lTyr 0

1

P-Tyr

!CL, 5

01

5

0

1

5

---I 190

-

"-_o

_MM

l

* -

_

2 00

6j9

01-iw

;3 10

.16

TriWV

10

Fig. 3. SDS-PAGE analysis of anti-neu (A) and anti-phosphotyrosine immunoprecipitates (B) from metabolically labelled cells. [35S]methioninelabelled cells were lysed and proteins were immunoprecipitated with the monoclonal anti-EGF-R antibodies, with the anti-neu antiserum or with the anti-phosphotyrosine antibodies (P-Tyr) as shown. Preimmune serum (p) and the antiserum against neu carboxyl terminus blocked with the f-galactosidase-neu protein (b) were used as specificity controls. The stronger signal obtained in anti-neu immunoprecipitation from the NEN7 cells as compared with the anti-EGF-R precipitation may depend on the use of monoclonal versus polyvalent antibodies and direct versus indirect immunoprecipitation. This signal intensity varied slightly from experiment to experiment. No neu polypeptides could be precipitated from NN or NIH 3T3 cells. The precipitation of the 220 kd polypeptide by anti-phosphotyrosine antibodies was independent of EGF and a similar band frequently also appeared in long exposures of anti-neu and control immunoprecipitates of various cells. We therefore interpret this band as part of the nonspecific background. Lanes G8: DHFR/G8 cells. Approximate mol. wt of the receptor proteins are given in kd.

EGF

EGF

NIH3T3

NEN16

NN

N EN49

G8

NEN7

Fig. 4. Effects of EGF on cell morphology. The cells were grown for 2 days in the absence (-) or presence (+) of 3 nM EGF and photographed in phase contrast microscopy. Note refractile cell bodies and thin, long processes in the EGF-treated NEN cells. The slight tendency of the NN cells to assume a spindle-shaped morphology was a consistent finding in our neomycin-selected cell cultures. However, both the parental NIH 3T3 cells and NN cells were unaffected in gross morphology by EGF in the conditions used. The sides of the panels correspond to 280 /Am.

(Stern and Kamps, 1988) and in SK-BR-3 breast carcinoma cells (King et al., 1988). However, after EGF stimulation no neu polypeptides were seen in anti-phosphotyrosine immunoprecipitates from the DHFR/G8 cells (Figure 3B). Thus it is unlikely that the observed phosphorylation of the chimeric neu receptor occurs in trans by the small amounts of endogenous EGF-R of NIH 3T3 cells. 162

Tranisforming properties of the EGF-R

-

neu

receptor

in the presence of EGF The cells expressing the hybrid protein, in contrast to control cells, responded to EGF with a transient membrane ruffling observed with time-lapse videography followed by partial loss of the flattened cell morphology within a few hours after the administration of EGF. EGF-dependent DNA synthesis

EGF Regulation of neu tyrosine kinase

EGF

Table I. Efficiency of soft-agar colony fonnation in the presence and absence of EGF or TGFa

+

Cells NN DHFR/G8 NEN16 NEN49 NEN7 E4

NN

No. of colonies -EGF 0 7.8 2.3 14 34 7200

± ± ± ± ±

±850

-TGFa NEN7

0 3.4 2.6 18 29

28 + 16

+EGF

7.3± 7.0 80 650 980 7700

13 ± 2.4 ± 83 ± 130 ± 16

±600

+TGFa 910 + 66

Only colonies of >50 cells were counted after 14 days. Mean values of three independent experiments are given with SD. The E4 cells (Sistonen et al., 1987) are c-Ha-ras-transformed NIH 3T3 cells selected for growth in soft agar.

G8

practically identical to that of EGF (Table I). The c-Ha-rasoncogene-transformed E4 cells (Sistonen et al., 1987) were used as a positive control for agar growth in this experiment. The E4 cells gave rise to colonies of the size of the NEN7 colonies two weeks after transfection. However, the growth of the E4 cells in soft agar was also independent of EGF (Table I).

t

I. e~~~~~~~~~~~~~~~~~~~~~~~~~~~

NEN49

0i a

Discussion t.

.WIA

9

NEN7 c

I*

Fig. 5. Growth in soft agar in the presence and absence of EGF. The cells were suspended and grown in soft agar for 14 days. The sides of the panels correspond to 1.5 mm.

increased 7- to 10-fold in cells expressing the chimeric receptor, in comparison to the effect of EGF on DNA synthesis in NIH 3T3 control cells (data not shown). When treated with EGF for 2 days the NEN cells became elongated, spindle-shaped and piled in groups oriented in a crisscross manner, unlike similarly treated control cells (Figure 4). This morphology was reminiscent of the changes seen in neu oncogene-expressing cells and suggested that the EGFtreated cells acquired transformed features when grown in the presence of EGF. To analyse this possibility further, anchorage-independent growth of the cells was assessed in was

soft agar. The cells were trypsinized, plated in soft agar and observed for 14 days. Only few colonies developed from cells grown in normal media during this time period. In contrast, the provision of EGF allowed for the growth of sizeable colonies in NEN16, NEN49 and NEN7 agar cultures, but did not significantly improve the inability of the other cells, including the NN and DHFR/G8 cells, to grow in agar (Figure 5 and Table I). The effect of TGFca on colony formation was

Our data indicates that expression of a chimeric EGF-R-neu receptor provides transfected NIH 3T3 cells with the ability to respond to EGF by neu-catalysed tyrosyl phosphorylation and growth stimulation. In addition, they suggest that abundant expression of the chimeric protein can lead to a transformed phenotype, but only in the presence of EGF. This means that the ligand binding and tyrosine kinase domains of EGF-R and neu receptor proteins are sufficiently similar to be able to provide functional complementation for each other in signal transduction through the plasma membrane. Riedel et al. (1986) were the first to combine ligand binding and signal transducing domains of two different growth factor receptors. They used the insulin receptor extracellular domain linked to the intracellular domain of the EGF-R. This construct, expressed transiently in COS-7 cells, allowed for signal transduction from insulin to the EGF-R tyrosine kinase, but the effect of this signal on cell phenotype was not analysed. In contrast, the EGF-R kinase was not stimulated by IL-2 in NIH 3T3 cells expressing an interleukin-2-receptor-EGF-R hybrid, and the cells assumed a ligand-independent transformed phenotype (Bernard et al.,

1987). Our immunoprecipitation experiments with anti-phosphotyrosine antibodies showed that EGF stimulation of the EGFR-neu receptor in NIH 3T3 cells led to its rapid phosphorylation in tyrosine residues. No phosphorylation of neu was obtained in control experiments with DHFR/G8 cells expressing an amplified neu gene and protein, thus excluding the possibility that the EGF-R of the NIH 3T3 cells was

responsible for the phosphorylation. The phosphorylation of neulerbB-2 in trans by the EGF-R in rat-I (Stern and Kamps, 1988) cells and human SK-BR-3 (King et al., 1988) cells may depend on a considerably higher ratio of EGF-R to 163

H.Lehvaslaiho et a!.

neulerbB-2 (- 1 x 104 and 9 x 104 EGF-R and approximately equal amounts of neulerbB-2 protein per cell respectively) expressed by these cells as compared with the NIH 3T3 cells ( - 3 x 103 receptors/cell, see Di Fiore et al., 1987b). The small receptor content of the NIH 3T3 cells is also consistent with immunoprecipitation and our EGF binding studies. Our in vivo and in vitro kinase experiments strongly favour the possibility that the chimeric receptor possesses autophosphorylating activity (the present data and unpublished data of Dr William Gullick and the authors). Small amounts of the chimeric receptor protein were precipitated also from unstimulated NEN7 cells with the antiphosphotyrosine antibodies, suggesting a partial loss of control of this tyrosine kinase acitivity by receptor overexpression. However, even this activity was stimulated by EGF. Treatment of the EGF-R-neu expressing cells with EGF led also to cell membrane changes similar to those seen in cells overexpressing the EGF-R. The delayed effects of EGF on receptor-expressing cells included a change to a rounded cell morphology reminiscent of the morphology of cells expressing the neu oncogene. Starved, EGF-treated cells carrying the chimeric receptor also respond by internalization and degradation of the receptor by activation of immediate early and early growth response genes, the glucose transporter and ornithine decarboxylase and an EGF dosedependent increase in DNA synthesis as measured by incorporation of [3H]thymidine or bromodeoxyuridine into DNA (L.Lehtola et al., L Sistonen et al., in preparation). Finally, the EGF-dependent growth in soft agar confirmed that the cells acquired a transformed phenotype. In fact, even the NEN7 cells expressing the highest levels of the chimeric receptor failed to form colonies in soft agar without EGF. On the basis of our results, the transformed phenotype of cells expressing the EGF-neu construct seems to be strictly controlled by the presence of receptor stimulating ligand. We are at present testing whether the same holds true in the presence of more excessive overexpression of the chimeric receptor from an amplified DNA template. On the other hand, the function of neu as a normal cellular protooncogene requires that its binding of a ligand does not lead to cell transformation. There may thus exist a carefully controlled relationship between amounts of receptor expressed and availability of the ligand for normal neu signal transduction and growth stimulation. This balance could be disrupted by erbB2/HER2 or EGF-R amplifications in human mammary carcinomas and some squamous cell carcinomas respectively. The signals mediated by the intracellular domains of the neu protein may be very similar to those of the EGF receptor. It has recently been shown that overexpression of the EGFR transforms cells only in the presence of its ligand (Di Fiore et al., 1987b; Velu et al., 1987; Riedel et al., 1988). In contrast, overexpression of the human erbB2 gene is sufficient to transform NIH 3T3 cells in the absence of an added ligand (Hudziak et al., 1987), while rat neu protooncogene is incapable of transformation (Bargmann et al., 1986b; Bargmann and Weinberg, 1988a). If a putative erbB2 ligand is present in NIH 3T3 cell cultures, it would therefore not efficiently recognize the homologous rat neu receptor since it does not stimulate its tyrosine kinase activity or transform neu-overexpressing cells. Careful receptor and ligand dose-transformation dependence titrations have to 164

be performed simultaneously for the neu and erbB2/HER2 proteins in order to clarify these differences in the biological consequences of heterologous receptor overexpression. For such studies involving the EGF-R-neu receptor, we have constructed mouse mammary tumor virus promoter - EGFR-neu vectors for regulated expression of the chimeric receptor protein in cells in culture and in vivo. According to recent results of neu expression in transgenic mice, the pl85neu oncoprotein is sufficient for carcinogenesis in the mammary gland (Muller et al., 1988). This indicates that expression of the activated neu in the mammary epithelium has particularly strong transforming properties. Tumorigenesis was not obtained in other organs, despite expression of the construct in various other tissues. So far it has not been reported whether overexpression of protoneu or erbB2 causes mammary carcinomas in transgenic mice. Recent results (Bargmann and Weinberg, 1988a; Stern et al., 1988) show that activation of neu as an oncogene is associated with a substantial increase in its tyrosine kinase activity in cell membrane preparations. This activity, however, was not retained after detergent solubilization of the receptor (Stern et al., 1986; Bargmann and Weinberg, 1988a). The neu proto-oncogene is > 100 times less active in cell transformation than the corresponding oncogene, which has an activating point mutation exchanging glutamic acid for valine (residue 664) in its transmembrane domain (Bargmann and Weinberg, 1986b). The correlation of increased tyrosine kianse activity and cell transformation associated with the neu oncoprotein and its deletion mutants (Bargmann and Weinberg, 1988a,b) and our present results support the importance of this biochemical function as a mode of cell transformation used by a number of oncogenes of this class (Hunter and Cooper, 1985; Hanks et al., 1988; Yarden and Ullrich, 1988). Although the biochemical difference between the normal and mutationally activated neu oncogene has thus been shown as increased receptor tyrosylphosphorylation in membrane preparations, experimental manipulations of the cellular neu kinase has previously not been possible. The present findings show for the first time the ligand-dependent stimulation of the neu tyrosine kinase and the associated cell growth activation. Such a dependence has not been obtained for chimeric receptors involving cytoplasmic tyrosine kinase domains of the non-receptor type (see Yarden and Ullrich, 1988). Although the biological functions of transmembrane tyrosine kinases may be associated with diverse phenomena from cell differentiation and organogenesis to metabolic processes (Hanks et al., 1988), our data favours the view that the neu protein is indeed a growth factor receptor. The present model provides the first system to study the regulation, properties and cellular function of the neu receptor kinase, its substrates, and various other aspects of its biological signalling activities, a line of research which has been frustrated for a long time by the lack of a neu receptor-binding ligand.

Materials and methods Reagents Receptor-grade EGF from mouse submaxillary glands were obtained from Collaborative Research; TGF-a was from Peninsula Laboratories; 1251 labelled mouse EGF (100 tCiltg; 1 pCi = 37 kBq), [_y-32P]ATP (> 5000 Ci/mmol), [35S]methionine (1000 Ci/mmol) and methyl-[3H]thymidine (84 Ci/mmol) from Amersham; protein A-Sepharose from Pharmacia; cell

EGF Regulation of neu tyrosine kinase culture reagents including the neomycin analogue geneticin (G418 sulphate) from Gibco Laboratories.

Construction of the prokaryotic neu-expression vector The eukaryotic expression vector pSV2neuNT (a kind gift from Robert Weinberg; Bargmann et al., 1986) containing the full length cDNA clone of the neu-oncogene was linearized with the BstEUi restriction endonuclease, the 5' overhangs were filled with the Klenow enzyme (Boehringer Mannheim) to give blunt ends and recut with BamHI. A 1086 bp fragment was isolated from gel and ligated to SmaI-BamHI opened prokaryotic expression vector pEX3 (Stanley and Luzio, 1984). The resulting plasmid pEXneu3' expresses a hybrid 3-galactosidase-neu fusion protein with 142 carboxy-terminal amino acids from the carboxyl terminus of p185neu*

Cells and transfections NIH 3T3 (Jainchill et al., 1969), COS-1 (Gluzman, 1981), A431 human squamous carcinoma cells which express an amplified c-erbB gene (Giard et al., 1973), BT-474 mammary carcinoma cells with an amplified c-erbB2 gene and protein (Lasfargues et al., 1978; Kraus et al., 1986) and E4 cHa-ras oncogene-transformed NIH 3T3 (Sistonen et al., 1987) cell lines have been previously described. Dihydrofolate reductase-deficient CHO cells were a kind gift from Dr Randal Kaufman. The DHFR/G8 cells expressing an amplified neu gene were a kind gift from Dr Mien-Chie Hung and Robert Weinberg (MD Anderson Cancer Center, University of Texas, Houston, TX and The Whitehead Institute, Boston, MA respectively). NIH 3T3 cells were grown in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal calf serum (FCS). 20 tig of linearized plasmid with 2 Ag of selection marker plasmid were used to transfect 9-cm plates using the calcium phosphate precipitate method (Graham and Van der Eb, 1973). Isolation of cellular DNA, Southern blotting and hybridization followed standard procedures. The cells were routinely checked for Mycoplasma contamination using the Hoechst fluorochrome 33 258 (Russell et al., 1975), with negative results. Time-lapse videography was carried out with Olympus IMT-2 inverted microscope and National NV-8050 time lapse recorder.

Antibodies Plasmid pEXneu3' was expressed in Escherichia coli strain pop 2136 (Boehringer Mannheim) and hybrid protein isolated as described (Stanley and Luzio, 1984). The fusion protein was purified by preparative PAGE and used to immunize rabbits. Two adult rabbits were immunized s.c. at multiple sites on day zero with mashed gel strip in PBS and complete Freunds adjuvant (-5 ml of emulsion containing 250 Mg of the hybrid protein). Animals were given booster immunizations containing - 150 Mg of the hybrid protein in incomplete Freunds adjuvant on days 18, 33, 47, 74 and 88, and test bleeds taken on days 53, 67, 78 and 97. Antisera were characterized for immunoprecipitation and specificity using the A431 and BT-474 cell lines as sources of the EGF-R and the c-erbB2 proteins respectively. The anti-phosphotyrosine antibodies were a kind gift from Dr H.Fujio (Osaka University, Osaka, Japan;Seki et al., 1983) and the neu-specific antipeptide antibodies 20N and 21N that were used as initial controls of the specificity of our antisera were a kind gift from Dr Gullick (Imperial Cancer Research Fund Laboratories, London, UK; Gullick et al., 1987). Mouse monoclonal antibodies against the EGF receptor extracellular domain were from Amersham (RPN.513).

EGF-binding studies EGF-binding assay was performed essentially as described by Nestor et al. (1985). The cells were seeded in 24-well dishes (Limbro) at 5 x I04 cells per well and grown for 1-2 days to 80% confluence. The cells were then washed twice with cold DMEM-20 mM Hepes, pH 7.4-0.1% BSA and incubated at +4°C for 60 min in 500 A1 of same medium supplemented with 1 nM [125I]EGF (Amersham) (specific activity 29 uCi/Mg). Duplicate wells were incubated in the presence of 100 MM EGF to determine nonspecific binding. The cells were washed three times with the medium, solubilized with 0.8 ml of 0.1 N NaOH and radioactivity was counted. Cell protein was measured from duplicate wells treated as above but without addition of radioactive isotope. Protein was assayed using the Bio-Rad protein assay kit with BSA standards.

Immunoprecipitation For metabolic labelling of cell proteins, the cultures were incubated in methionine-free medium for 30-60 min and then labelled for 14-16 h with [35S]methionine (100 ACi/mil, Amersham). For immunoprecipitation, the cells were lysed in 0.1% SDS, 0.5% Triton X-100, 0.5% sodium desoxycholate, 20 mM Tris-HCI, pH 7.5 and sonicated for 1 min at 300 Won ice. The lysates were centrifuged for 30 min at 10 000 r.p.m. in a HB4 rotor in a Sorvall centrifuge, and the supernatant was divided into

1-ml aliquots to which - 1 M1 of antibody was added and allowed to bind at 4°C for 1 h. For the blocking experiments, 2 M1 of antibody was preincubated with 2 Mg of bacterial neu protein for 20 h. About 30 Ml of a 50% v/v solution of protein A-sepharose (Pharmacia) was added and the tubes mixed gently for 1 h at 4°C. For the EGF-R immunoprecipitation with mouse monoclonal antibodies, rabbit anti-mouse immunoglobulincoated, washed protein A-sepharose particles were used. The immune complexes were washed five times with the immunoprecipitation buffer lacking SDS, twice with PBS and once with 20 mM Tris-HCI, pH 7.0, and dissolved in the electrophoresis sample buffer containing 2 % SDS, 5 % ,B-mercaptoethanol, 10% glycerol and 50 mM Tris-HCI (pH 6.8), boiled for 5 min, and analysed in a 7.5% SDS-polyacrylamide gel according to Laemmli (1970). After electrophoresis the gels were fixed in 10% acetic acid, the 35S-labelled samples were impregnated with Amplify (Amersham), dried onto filter paper and fluorographed using X-Omat R film (Eastman Kodak). Immunofluorescence For indirect surface immunofluorescence, unfixed cells were stained on ice in PBS containing 0.1 % FCS. For indirect immunofluorescence and photomicrography, the cells were trypsinized and transferred onto coverslips. Fixation was carried out at room temperature after rinsing the cells on coverslips with PBS, incubation in 4% (w/v) paraformaldehyde in PBS for 30 min and again washed with PBS. To visualize intracellular antigens, the cells were permeabilized with 0.2% (v/v) Nonidet P-40 (BDH Chemicals Ltd, Poole, UK) in PBS for 5 min. After another three PBS washes the cells were labelled using the rabbit antibodies and stained with TRITCconjugated goat anti-rabbit IgG (Cappel Products, Organon Teknika, NV, Tumhout, Belgium) or goat anti-mouse IgG (Cappel). Stained samples were mounted in 50% glycerol and photographed on Agfa Pan 400 film using a Leitz Dialux immunofluorescence microscope.

Analysis of DNA synthesis Cells were seeded in small plates at 105 cells per plate. After overnight incubation in DMEM containing 10% FCS, cells were washed twice and then incubated in medium containing 1 % FCS for 24 h. DNA synthesis was stimulated by addition of 10 nM EGF or TGF-ce and was monitored 12-24 h later by measurement of BrdU incorporation using the cell proliferation kit (Amersham RPN.20).

Soft-agar assay Subconfluent cells were trypsinized and plated at 105 and 104 cells per 6-cm dish in the presence or absence of 3 nM EGF in 3 ml DMEM containing 10% FCS and 0.25% agar (Difco) on a bottom layer of 0.5% agar in 5 ml DMEM. Colonies of 50 or more cells were counted 14 days later.

Acknowledgements We thank Drs Robert Weinberg, Mien-Chie Hung and Towia Libermann for molecular clones and DHFR/G8 cells, Drs William Gullick and Hajime Fujio for antibodies and Michael Schlossmacher and Dr Joseph Schlessinger for encouragement and discussions, Elina Roimaa for expert technical assistance and Monna Schoulz for the FACS analysis. Our studies were supported by the Finnish Academy and the Finnish Cancer Organizations and carried out under a research contract with the Finnish Life and Pension Insurance Companies.

References Bargmann,C.I. and Weinberg,R.A. (1988a) EMBO J., 7, 2043-2052. Bargmann,C.I. and Weinberg,R.A. (1988b) Proc. Natl. Acad. Sci. USA, 85, 5394-5398. Bargmann,C.I., Hung,M.-C. and Weinberg,R.A. (1986a) Nature, 319, 226-230. Bargmann,C.I., Hung,M.-C. and Weinberg,R.A. (1986b) Cell, 45, 649-657. Bernard,O., de St Groth,B.F., Ullrich,A., Green,W. and Schlessinger,J. (1987) Proc. Natl. Acad. Sci. USA, 84, 2125-2129. Coussens,L., Yang-Feng,T.L., Liao,Y.-C., Chen,E., Gray,A., McGrath,J., Seeburg,P.H., Libermann,T., Schlessinger,J., Francke,U., Levinson,A. and Ullrich,A. (1985) Science, 230, 1132-1139. Di Fiore,P.P., Pierce,J.H., Kraus,M.H., Segatto,O., King,C.R. and Aaronson,S.A. (1987a) Science, 237, 178-182. Di Fiore,P.P., Pierce,J.H., Fleming,T.P., Hazan,R., Ullrich,A., King,C.R., Schlessinger,J. and Aaronson,S.A. (1987b) Cell, 51, 1063-1070. Giard,D.J., Aaronson,S.A., Todaro,G.J., Arnstein,P., Kersey,J.H., Dosik,H. and Parks,P. (1973) J. Natl. Cancer Inst. (US), 51, 1417-1423.

165

H.Lehvaslaiho et al.

Gluzman,Y. (1981) Cell, 23, 175-182. Graham,F.L. and Van der Eb,A.J. (1973) Virology, 52, 456-467. Gullick,W.J., Berger,M.S., Bennett,P.L.P., Rothbard,J.B. and Waterfield,M.D. (1987) Int. J. Cancer, 40, 246-254. Haigler,H., Ash,J.F., Singer,S.J. and Cohen,S. (1978) Proc. Natl. Acad. Sci. USA, 75, 3317-3321. Hanks,S.K., Quinn,A.M. and Hunter,T. (1988) Science, 241, 42-52. Hudziak,R.M., Schlessinger,J. and Ullrich,A. (1987) Proc. Natl. Acad. Sci. USA, 84, 7159-7163. Hung,M.-C., Schechter,A.L., Chevray,P.-Y., Stern,D.F. and Weinberg, R.A. (1986) Proc. Natl. Acad. Sci. USA, 83, 261-264. Hunter,T. and Cooper,J.A. (1985) Annu. Rev. Biochem., 54, 897-930. Hunter,T., Ling,N. and Cooper,J.A. (1985) Nature, 311, 480-483. Jainchill,J.L., Aaronson,S.A. and Todaro,G.J. (1969) J. Virol., 4, 549-553. Muller,W., Sinn,E., Pattengale,P.K., Wallace,R. and Leder,P. (1988) Cell, 54, 105-115. Nestor,J.J., Newman,S.R., DeLustro,B., Torada,G.J. and Schreiber,A.B. (1985) Biochem. Biophys. Res. Commun., 129, 226-232. King,C.R., Kraus,M.H. and Aaronson,S.A. (1985) Science, 229, 974-978. King,C.R., Borello,I., Bellot,?, Comoglio,P. and Schlessinger,J. (1988) EMBO J., 7, 1647-1651. Kraus,M.H., Popescu,N.C., Amsbaugh,S.C. and King,C.R. (1986) EMBO J., 6, 605-610. Laemmli,U.K. (1970) Nature, 227, 680-685. Lasfargues,E.Y., Coutinho,W.G. and Redfield,E.S. (1978) J. Natl. Cancer Inst., 61, 967-978. Padhy,L.C., Shih,C., Cowing,D., Finkelstein,R. and Weinberg,R.A. (1982) Cell, 28, 865-871. Riedel,H., Dull,T.J., Schlessinger,J. and Ullrich,A. (1986) Nature, 324, 68-70. Riedel,H., Massoglia,S., Schlessinger,J. and Ullrich,A. (1988) Proc. Natl. Acad. Sci. USA, 85, 1477-1481. Russell,W.C., Newman,C. and Williamson,D.H. (1975) Nature, 253, 461 -462. Schechter,A.L., Stern,D.F., Vaidyanathan,L., Decker,S.J., Drebin,J.A., Greene,M.I. and Weinberg,R.A. (1984) Nature, 312, 513-516. Seki,J., Owada,M.K., Sakato,N. and Fujio,H. (1986) Cancer Res., 46, 907-916. Semba,K., Kamata,N., Toyoshima,K. and Yamamoto,T. (1985) Proc. Natl. Acad. Sci. USA, 82, 6497-6501. Shih,C., Padhy,L.C., Cowing,D., Finkelstein,R. and Weinberg,R.A. (1981) Nature, 290, 261-264. Sistonen,L., Keski-Oja,J., UlmanenI., Holtta,E., Wikgren,B.-J. and Alitalo,K. (1987) Exp. Cell Res., 168, 518-530. Slamon,D.J., Clark,G.M., Wong,S.G., Levin,W.J., Ullrich,A. and McGuire,W.L. (1987) Science, 235, 177-182. Southern,P. and Berg,P. (1982) In Gluzman,Y. (ed.), Eukaryotic Viral Vectors. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 41-45.

Stanley,K.K. and Luzio,J.P. (1984) EMBO J., 3, 1429-1434. Stern,D.F. and Kamps,M.P. (1988) EMBO J., 7, 995-1001. Stern,D.F., Kamps,M.P. and Cao,H. (1988) Mol. Cell. Biol., 8, 3969-3973.

Stern,D.F., Heffernan,P.A. and Weinberg,R.A. (1986) Mol. Cell. Biol., 6, 1729-1740. Ullrich,A., Coussens,L., Hayflick,J.S., Dull,T.J., Gray,A., Tam,A.W., Lee,J., Yarden,Y.,Liberman,T.A., Schlessinger,J., Downward,J., Mayes,E.L.V., Whittle,N., Waterfield,M.D. and Seeburg,P.H. (1984) Nature, 309, 418-425. Velu,T.J., Beguinot,L., Vass,W.C., Willingham,M.C., Merlino,G.T., Pastan,I. and Lowy,D.R. (1987) Science, 238, 1408-1410. Van de Vijver,M., van de Bersselaar,R., Devilee,P., Cornelisse,C., Peterse,J. and Nusse,R. (1987) Mol. Cell. Biol., 7, 2019-2023. Venter,D.J., Kumar,S., Tuzi,N.L. and Gullick,W.J. (1987) Lancet, ii, 69-71.

Yamamoto,T., Ikawa,S., Akiyama,T., Semba,K., Nomura,N., Miyajima,N., Saito,T. and Toyoshima,K. (1986) Nature, 319, 230-234. Yarden,Y. and Ullrich,A. (1988) Annu. Rev. Biochem., 57, 443-478. Yokota,J., Yamamoto,T., Toyoshima,K., Terada,M., Sugimura,T., Battifora,H. and Cline,M.J. (1986) Lancet, i, 765-767. Yokota,J., Yamamoto,T., Miyajima,H., Toyoshima,K., Nomura,N., Sakamoto,H., Yoshida,T., Terada,M. and Sugimura,T. (1988) Oncogene, 2, 283 -288. Received on September 15, 1988; revised on October 27, 1988

166