MoLEcuLAR AND CELLULAR BIOLOGY, Dec. 1992, p. 5824-5833 Vol. 12, No. 120270-7306/92/125824-10$02.00/0Copyright © 1992, American Society for Microbiology

The SH2 and SH3 Domain-Containing Nck Protein IsOncogenic and a Common Target for Phosphorylation

by Different Surface ReceptorsW. LI,' P. HU,' E. Y. SKOLNIK,1 A. ULLRICH,2 AND J. SCHLESSINGER'*

Department ofPharmacology, New York University Medical Center, 550 First Avenue, New YorkNew York 10016, and Max Planck Institute fir Biochemie, Am Klopferspitz 1&4,

8033 Martinsried, Germany2

Received 16 July 1992/Returned for modification 26 August 1992/Accepted 15 September 1992

Signalling proteins such as phospholipase C-y (PLC-y) or GTPase-activating protein (GAP) of ras containconserved regions of approximately 100 amino acids termed src homology 2 (SH2) domains. SH2 domains wereshown to be responsible for mediating association between signalling proteins and tyrosine-phosphorylatedproteins, including growth factor receptors. Nck is an ubiquitously expressed protein consisting exclusively ofone SH2 and three SH3 domains. Here we show that epidermal growth factor or platelet-derived growth factorstimulation of intact human or murine cells leads to phosphorylation of Nck protein on tyrosine, serine, andthreonine residues. Similar stimulation of Nck phosphorylation was detected upon activation of rat basophilicleukemia RBL-2H3 cells by cross-linking of the high-affinity immunoglobulin E receptors (FceRI). Ligand-activated, tyrosine-autophosphorylated platelet-derived growth factor or epidermal growth factor receptorswere coimmunoprecipitated with anti-Nck antibodies, and the association with either receptor molecule wasmediated by the SH2 domain of Nck. Addition of phorbol ester was also able to stimulate Nck phosphorylationon serine residues. However, growth factor-induced serine/threonine phosphorylation of Nck was not mediatedby protein kinase C. Interestingly, approximately fivefold overexpression of Nck in NIH 3T3 cells resulted information of oncogenic foci. These results show that Nck is an oncogenic protein and a common target for theaction of different surface receptors. Nck probably functions as an adaptor protein which links surface receptorswith tyrosine kinase activity to downstream signalling pathways involved in the control of cell proifferation.

Growth factors elicit a variety of immediate responses bybinding to and activating specific cell surface receptorswhich possess intrinsic protein-tyrosine kinase activity (1,47). These responses include stimulation of cellular proteinphosphorylation, intracellular [Ca2"] rise, cytoplasmic alka-linization, and early gene expression, leading to DNA syn-thesis and cell growth and differentiation. Growth factorbinding is followed by receptor dimerization, which is re-sponsible for activation of the intrinsic protein-tyrosinekinase activity and receptor autophosphorylation (40). Theintrinsic protein-tyrosine kinase activity was shown to beessential for the biological function of receptor tyrosinekinases (47). Several cellular proteins which directly interactwith and are tyrosine phosphorylated by receptors withtyrosine kinase activity have been identified. These includephospholipase C--y (PLC-y), ras GTPase-activating protein(GAP), and pp60csn` (5, 47). Recent studies have demon-strated that these proteins bind to specific tyrosine-auto-phosphorylated regions in growth factor receptors via aconserved stretch of 100 amino acids called src-homology 2(SH2) domains (20). SH2 domains were initially described asnoncatalytic sequence motifs found in pp6c-sr and in othercytoplasmic tyrosine kinases (13, 20, 28). It has been shownthat SH2 domains of signalling molecules bind to phosphor-ylated tyrosine residues flanked by specific short sequences(5, 10, 15, 18, 19, 34, 38). Different SH2 domains bind todifferent sequence motifs; SH2 domains of p85 bind to aYMXM motif (5), while the SH2 domains of PLC--y bind toa L/V XXXX EYL motif found in epidermal growth factor

* Corresponding author.

(EGF) and fibroblast growth factor (FGF) receptor (EGFRand FGFR) carboxy-terminal domains (34, 39). The bindingof SH2 domains to tyrosine-phosphorylated regions ofgrowth factor receptors is thought to provide a commonmechanism by which diverse regulatory proteins can interactspecifically with growth factor receptors, thereby couplinggrowth factor stimulation to multiple intracellular signallingpathways.SH2 domains are usually present in one or two copies per

molecule and are often accompanied by a shorter conservedstretch of 50 amino acids, called the SH3 domain, whosefunction is still unknown (20). SH2 domains have been foundin proteins with diverse enzymatic activities such as cyto-plasmic tyrosine kinases (20), a tyrosine-specific phos-phatase (41, 50), PLC--y (44), GAP (46, 49) and PI-3 kinase-associated protein p85 (GRB1) (11, 35, 42). SH2 domainswere also found in protooncogene vav (4, 16, 17, 29) and inthe cytoskeletal protein tensin (9). At least three proteins,crk (30), Nck (22), and GRB2/sem-5 (7, 26), consist almostexclusively of SH2 and SH3 domains.

In this report we show that Nck mRNA and proteins arewidely expressed in various murine tissues and in cell linesfrom human, murine, and rat origins. Nck is phosphorylatedon tyrosine, serine, and threonine residues in response tostimulation of EGF and platelet-derived growth factor(PDGF) in A431 and NIH 3T3 cells, respectively, andcross-linking of immunoglobulin E (IgE) receptors (FcsRI)in the rat basophilic leukemia cell line RBL-2H3. Like otherSH2-containing proteins, Nck is associated with tyrosine-autophosphorylated EGFR or PDGFR via its SH2 domain.Moreover, overexpression of Nck leads to transformation ofNIH 3T3 cells. These results suggest that Nck plays a role in

5824

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

Nck IS ONCOGENIC AND A TARGET FOR SURFACE RECEPTORS 5825

mitogenic signal transduction pathways stimulated by differ-ent receptors which activate protein-tyrosine kinase activity.

MATERIALS AND METHODS

A431 and NIH 3T3 cells were grown in Dulbecco's mod-ified Eagle's medium (DMEM) supplemented with 10% fetalcalf serum containing 50 IU each of penicillin and strepto-mycin per ml in a humidified CO2 (5%) incubator at 37°C.RBL-2H3 cells were grown in MEM a medium supple-mented with 15% Hybrimax (Sigma). Human recombinantEGF and PDGF type b homodimer were purchased fromINTERGEN (Purchase, N.Y.). Monoclonal anti-dinitrophe-nol (DNP) IgE antibodies (25) and DNP-bovine serum albu-min (BSA) were generously provided by J. Oliver (Univer-sity of New Mexico, Albuquerque). 32Pi (specific activity:314.5 to 337.5 TBq/mmol; carrier-free) was purchased fromNEN (Wilmington, Del.). Anti-EGFR anti-peptide antibod-ies were previously described (14). Anti-PDGFR antibodieswere kindly provided by A. Zilberstein (RPR Inc., King ofPrussia, Pa.). Monoclonal anti-phosphotyrosine (anti-PY)antibodies (1G2) covalently coupled to agarose were pur-chased from Oncogene Science (Manhasset, N.Y.). Otherreagents were obtained from Sigma (St. Louis, Mo.).

Generation of glutathione-S-transferase (GST)-Nck fusionproteins and polyclonal rabbit antibodies against Nck. Oligo-nucleotides flanking either the entire coding sequence or thecarboxy-terminal half (amino acids 199 to 377) of Nck andcontaining engineered restriction sites were used to amplifythe appropriate DNA fragments by polymerase chain reac-tion. These fragments were cleaved with BamHI and EcoRIand ligated into pGEX3X (43). Competent Escherichia coliHB101 cells were transformed, and recombinant clones werescreened by restriction enzyme analysis and sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE)analysis of overproduced Nck-containing fusion proteins.The fusion proteins were purified by using affinity chroma-tography with immobilized glutathione beads as describedpreviously (43). In addition, a peptide corresponding toamino acid residues 80 to 100 of human Nck was synthesizedand conjugated to keyhole limpet hemocyanin for antibodyproduction. Both the keyhole limpet hemocyanin-boundsynthetic peptide and the fusion proteins were injected intorabbits to raise antibodies (Cocalico Biological Inc., Ream-stown, Pa.). Rabbit polyclonal antibodies were generatedagainst GST-Nck (Ab65, Ab66), GST-NckC' (Ab67, Ab68),and the peptide (Ab57, Ab58). The various antisera recog-nize a 45-kDa protein as revealed either by immunoprecipi-tation from [35S]methionine-labeled A431 cells (see Fig. 2A)or by immunoblotting analysis (see Fig. 2B and data notshown).

Northern analysis. RNA was prepared from the humanembryonic kidney cell line 293, human epidermoid carci-noma cell line A431, human glioblastoma cell line 1242,HeLa cells, NIH 3T3 cells, and the rat bladder carcinomacell line NBT2, or from mouse tissues according to publishedprocedures (27). Equal amounts of RNA were size fraction-ated by electrophoresis in a 1.2% agarose-2.2 M formalde-hyde gel, transferred onto a nylon membrane by capillaryaction, and baked at 80°C for 2 h. Following prehybridiza-tion, the blot was hybridized with a 32P-labeled nick-trans-lated DNA probe (27). Mouse Nck cDNA was recentlyisolated in our laboratory (unpublished data). AcDNA probeof murine Nck composed of one SH3 and one SH2 domainwas used in the blotting experiment. Hybridization wascarried out overnight at 42°C in the presence of 50% form-

amide-5 x SSC (lx SSC is 0.15 M NaCl plus 0.015 M sodiumcitrate)-0.1% SDS-5x Denhardt's solution. The membranewas then washed in O.lx SSC-0.1% SDS at 420C andexposed to Kodak XAR film at -70°C.

Cell metabolic labeling. Subconfluent A431 and NIH 3T3cells were incubated in DMEM containing 1% fetal calfserum for 16 h, and RBL-2H3 cells were incubated in MEMcontaining 5% Hybrimax and 1.0 pg of anti-DNP IgE per mlfor 4 to 12 h. Cells were labeled with 32Pi as previouslydescribed (24). Briefly, cells were washed twice with phos-phate-free medium and incubated in the same medium con-taining 0.5% dialyzed fetal calf serum for 2 h. The cells werethen washed twice with phosphate-free medium again priorto incubation with the same medium containing 0.5% dia-lyzed fetal calf serum and 32Pi (1 mCi/ml) for 2 to 3 h. A431and NIH 3T3 cells were incubated in the presence orabsence of EGF (250ng/ml), PDGF (50 ng/ml), phorbolmyristate acetate (PMA) (100 nM), or vanadate (200 ,uM) at37°C for the indicated times. In certain experiments the cellswere preincubated with 10 nM PMA for 15 h before stimu-lation with EGF, PDGF, or PMA. Fc8RI receptors on thesurface of anti-DNP IgE-primed RBL 2H3 cells were stim-ulated by incubating the cells with DNP-conjugated BSA(DNP-BSA) (0.1 jig/ml). Subsequently, the cells were rap-idly washed three times with ice-cold phosphate-bufferedsaline, solubilized immediately on ice in lysis buffer (10 mMTris base, 50 mM NaCl, 30 mM sodium pyrophosphate, 50mM sodium fluoride, 100 ,uM sodium orthovanadate, 2 mMiodoacetic acid, 5 ,uM ZnCl2, 0.1% BSA, 1 mM freshlyprepared phenylmethylsulfonyl fluoride, 0.5% Nonidet P-40,pH 7.1, at room temperature), and centrifuged for 20 min at4°C at 8,000 x g in an Eppendorf microcentrifuge. Similaramounts of cellular proteins as determined by the method ofBradford (3) were used in subsequent immunoprecipitationexperiments.The cell lysates were incubated with anti-Nck antiserum

(Ab68) for 90 min at 4°C, and the immune complexes wereprecipitated with protein A-Sepharose for 45 min at 4°C. Thebeads were washed eight times with RIPA buffer (20 mMTris, pH 7.6,300 mM NaCl, 2 mM EDTA, 1% Triton X-100,1% sodium deoxycholate, 0.1% SDS) and heated in lxelectrophoresis sample buffer at 95°C for 5 min. The super-natants were resolved on SDS-PAGE (10% polyacrylamide)and visualized by autoradiography. To isolate tyrosine-phosphorylated proteins, the cell lysates were incubatedwith anti-PY antibody beads for 2 h at 4°C. The anti-PYbeads were washed five times with lysis buffer without BSAand PY-containing proteins eluted with the lysis buffercontaining 1 mM phenyl phosphate and analyzed by SDS-PAGE as described previously (24).

In vivo and in vitro binding of Nck to EGFR and PDGFR.The lysates of EGF- or PDGF-stimulated or unstimulatedcells were immunoprecipitated with Ab68. The immunecomplexes were washed three times with lysis buffer withoutBSA, boiled in lx sample buffer, and resolved by SDS-PAGE. The proteins were transferred to nitrocellulose mem-brane and blotted with anti-PY (14) or anti-receptor antibod-ies.For in vitro binding to EGFR, lysates of EGF-stimulated

or unstimulated A431 cells (_107) were incubated withGST-Nck fusion proteins (25 j,g) coupled to glutathionebeads for 2 h at 4°C. The beads were washed three times withlysis buffer, heated in sample buffer containing 3-mercapto-ethanol, analyzed by SDS-PAGE, transferred to nitrocellu-lose membrane, and blotted with anti-EGFR antibodies. The

VOL. 12, 1992

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

5826 LI ET AL.

results were visualized by 125I-protein A and autoradiogra-phy.For binding to PDGFR in vitro, the anti-PY-reactive

fractions isolated from untreated or PDGF-treated, 32p_labeled NIH 3T3 cells by anti-PY affinity column werewashed three times in Centricon 30 with lysis buffer withoutNonidet P-40 to remove phenyl phosphate prior to theaddition of GST-Nck fusion protein beads (10 ,ug) as de-scribed above. The GST-Nck beads were washed, and thebound PY proteins were analyzed by SDS-PAGE and auto-radiography.Phosphoamino acid analysis and proteolytic peptide map-

ping. Phosphorylated protein bands were excised from thedried SDS gel, swollen in 50 mM NH4HCO3, minced in anall-glass Dounce homogenizer, and boiled in 0.1% SDS-5%,3-mercaptoethanol for 5 min (final volume, 1 ml). Themixture was incubated at 37°C with agitation for at least 4 hand centrifuged for 10 min in an Eppendorf microcentrifuge.The supernatants were transferred to SARSTEDT 1.5-mltubes with rubber ring-containing screw caps. The phosphor-ylated proteins were precipitated in 15% trichloroacetic acidin the presence of 50 ,ug of BSA per ml on ice for 90 min,centrifuged for 8 min at 4°C, and washed twice with coldacetone. Acetone was removed, and the protein pellets weredried in a Speed-Vac for 15 min. The protein pellets werehydrolyzed in 5.7 N HCI at 110°C for 45 min. One ml of H20was added to dilute the acid, and the sample was dried in aSpeed-Vac. The washing was repeated three times with 300,100, and 50 pA of H20. The washed and dried samples weredissolved in 5 ,ul of electrophoresis buffer (pH 1.9) contain-ing 0.6 jig of phosphoamino acid standards (phosphoserine,phosphothreonine, and phosphotyrosine), spotted on a thin-layer chromatography plate (100-,um diameter; Merck) in0.25-,ul aliquots, and analyzed by two-dimensional electro-phoresis as described by Cooper et al. (8). Two-dimensionaltryptic peptide mapping was carried out as previously de-scribed (48).

Transformation of NIH 3T3 cells. pLXSN, pLXSNNCK,and pIBW-3ras expression vectors were assayed by trans-fection with the calcium phosphate precipitation method. Ineach experiment, 5 ,ug of test DNA (along with 5 ,ug ofsalmon sperm DNA as carrier) was used to transfect one10-cm plate of NIH 3T3 cells (5 x 105 cells) cultured inDMEM containing 10% calf serum (Colorado Serum Co.,Denver, Colo.). After 18 h of incubation with the DNAprecipitates, the transfected cells were trypsinized andseeded into three 10-cm tissue culture dishes with DMEMcontaining 5% calf serum. The medium was changed every 3days, and foci were scored after 2 weeks.

RESULTS

Ubiquitous expression of Nck mRNA and protein. We haveexamined the expression of Nck mRNA and protein by usingeither human or murine cDNA probes and rabbit polyclonalantibodies against Nck, respectively. Figure 1A shows that amajor Nck transcript of 2.2 kb was found in different humancell lines. The nature of the other minor Nck transcripts of1.7, 3, and 4 kb is currently unknown. These transcripts mayrepresent alternatively spliced forms of Nck or Nck-relatedgenes. Using a human Nck probe, we were unable to detecteither murine (lane 6) or rat (lane 1) Nck transcripts.However, a Nck transcript of 2.4 kb was detected (Fig. 1B)in RNA preparations from all murine tissues tested uponhybridization with a murine Nck probe. Nck transcriptswere observed in lanes 2 and 4 upon longer exposure of the

A zz

a) NI - c CN

I

z

4.4 -

128-

1 2 3 45 6

cIn

C,

_4_) -L.. CD E

43)

4) .) (/,

C a)4) -J 1 cl

Ir ~ LL .; - m

-2 8 S

-1 8S

1 2 3 4 5 6 7 8 9 10FIG. 1. Analysis of Nck transcripts in different cell lines. Ten

micrograms of total RNA was electrophoresed on a 1.2% agarose-2.2 M formaldehyde gel. The blot was hybridized with a 32P-labelednick-translated DNA probe of either human (A) or mouse (B) Nckand analyzed by autoradiography.

blot. Immunoprecipitation analysis of [35S]methionine-la-beled cells with polyclonal antibodies generated againsteither a synthetic peptide or against a recombinant Nckfusion protein confirmed that Nck protein is widely ex-pressed in different cell lines including A431, rat RBL-2H3,and NIH 3T3 cells, and that its apparent molecular mass inSDS gels is 45,000 Da (Fig. 2 and data not shown).

Phosphorylation of Nck is augmented upon activation ofdifferent cell surface receptors. Since Nck contains a SH2domain, we have examined the possibility that it is a targetfor the action of growth factor receptors with tyrosine kinaseactivity such as EGFR or PDGFR. Hence, 32Pi-labeled A431or 3T3 cells were treated with or without EGF or PDGF,respectively, solubilized, and immunoprecipitated with anti-Nck antibodies (Ab68). The immunoprecipitates were ana-lyzed by SDS-PAGE and autoradiography. Figure 3 showsthat Nck is phosphorylated in unstimulated cells and thatboth PDGF (Fig. 3A, lanes 1 and 2) and EGF (Fig. 3B, lanes1 and 2) induced stimulation of Nck phosphorylation in living

MOL. CELL. BIOL.

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

Nck IS ONCOGENIC AND A TARGET FOR SURFACE RECEPTORS 5827

A.

_-a p a p a p

4 8.5-

-w _~ 41t -itm _.. -NC K

B. A.

P D G Fpa

_ + - +

4 8.5 -

, W-NCK NIH3T3

E 36.5-Om NNCK

3 6.5 -

1 2 3 4 5 6 1 2

FIG. 2. Immunoprecipitation and immunoblotting analysis ofNck protein. A431 cells were either labeled with [35S]methionine (A)or unlabeled (B) and solubilized in lysis buffer. (A) Lysates of[35S]methionine-labeled cells were incubated with either preimmuneserum (p) or antisera (a) (Ab #58, #65, or #68), and precipitatedwith protein A-Sepharose. Samples were analyzed by SDS-PAGEand fluorography. (B) Cell lysates (80 Ill; 200 pg of protein) wereresolved by SDS-PAGE, transferred to nitrocellulose membrane,and blotted with anti-Nck antibody #58 (lane 2) or preimmuneserum (lane 1). Similar results were obtained when blotting wasperformed with antibodies #65 and #68.

B.E G F

A431

C.DNP -BSA

1 2 3 4

_ + - +

-- 3NCK

1 2 3 4

_ +-

+

cells. Quantitation of the radioactivity of Nck protein ex-cised from SDS gels showed an approximately two- tofourfold increase in response to either growth factor receptoractivation. Lanes 3 and 4 of Fig. 3 show that similar amountsof Nck protein were recovered from both unstimulated(lanes 3) and stimulated (lanes 4) cells. Phosphoamino acidanalysis (Fig. 4) indicated that basal phosphorylation was onserine residues and that growth factor-induced phosphory-lation occurred on tyrosine, serine, and threonine residues.

It has been shown that activation of rat basophilic leuke-mia RBL-2H3 cells by cross-linking of the high-affinity IgEreceptors (FceRI) leads to activation of tyrosine kinase(s)which phosphorylate SH2-containing proteins such asPLC-y (23, 36) and Vav (4, 29). Since Nck is expressed inRBL-2H3 cells, we have examined its phosphorylation statein response to cross-linking of FcsRI. Figure 3C shows thatactivation of RBL-2H3 cells leads to a nearly three-foldstimulation of Nck phosphorylation. Phosphoamino acidanalysis indicated that IgE receptor cross-linking-stimulatedNck phosphorylation occurred on serine and tyrosine resi-dues (Fig. 4).Thus, our data show that Nck is a target for tyrosine

kinase(s) and serine/threonine kinase(s) which are activatedby EGF and PDGF in A431 and NIH 3T3 cells and FceRIcross-linking in RBL-SH3 cells. Comparison of the amountof Nck isolated by anti-PY antibodies and by anti-Nckantiserum from [35S]methionine-labeled EGF-stimulatedA431 cells indicated that 5 to 8% of cellular Nck was boundto the anti-PY column (data not shown). Studies in A431cells showed that the kinetics of EGF-stimulated Nck phos-phorylation (Fig. 5A) is similar to the kinetics of EGF-stimulated tyrosine phosphorylation of other cellular sub-strates (Fig. 5B). EGF-induced Nck phosphorylation wasmaximal after 1 to 3 min and lasted for at least 60 min.Similar results were observed in PDGF-stimulated NIH 3T3cells and in FcsRI cross-linked RBL-2H3 cells (data notshown).Nck associates with tyrosine-autophosphorylated EGFR or

PDGFR via its SH2 domain. It has been shown that SH2-containing proteins interact with activated growth factorreceptors (13, 20, 28). We have, therefore, examined thecapacity of Nck to interact with activated PDGFR or EGFR.

RBL-2H3 __ 4W- CK

1 2 3 4FIG. 3. EGF, PDGF, and FceRI cross-linking stimulate phos-

phorylation of Nck in living cells. A431 (A), NIH 3T3 (B), oranti-DNP IgE-primed RBL-2H3 (C) cells labeled with (lanes 1 and 2)or without (lanes 3 and 4) 32p; were incubated in the presence (+) orabsence (-) of EGF (250 ng/ml) or PDGF (50 ng/ml) for 5 min orwith DNP-BSA (0.1 pg/ml) for 3 min at 37°C Anti-Nck immunopre-cipitates of the radiolabeled cell lysates were analyzed by SDS-PAGE and autoradiography (lanes 1 and 2). Duplicate unlabeled celllysates were subjected to immunoblotting analysis with anti-Nckantibodies and visualized by binding to 1"'I-protein-A and autoradi-ography (lanes 3 and 4). Exposure times: lanes 1 and 2, 30 to 40 min;lanes 3 and 4, 12 to 15 h.

Thus, either NIH 3T3 or A431 cells were treated with orwithout PDGF or EGF, respectively. Cells were solubilized,and their lysates were immunoprecipitated with anti-Nck antiserum. The immune complexes were subjected toSDS-PAGE, transferred to nitrocellulose membrane, andimmunoblotted with anti-PY, anti-EGFR, or anti-PDGFRantibodies. Figure 6A shows that a 170-kDa protein coim-munoprecipitated by anti-Nck antibodies from EGF-stimu-lated (lane 2) but not from unstimulated (lane 1) cells wasrecognized by anti-PY antibodies. In comparison to the totalamount of tyrosine-phosphorylated EGFR in the lysate (lane4), only a small fraction (-1%) of EGFR was coimmunopre-cipitated from A431 cells by the anti-Nck antiserum. Thesame protein was recognized in a control duplicate blot withanti-EGFR antibodies (data not shown). A similar coimmu-noprecipitation of PDGFR by anti-Nck antibodies was alsoobserved in lysates from NIH 3T3 cells (Fig. 6C). In thisexperiment a 180-kDa protein corresponding to PDGFR wasdetected in PDGF-stimulated (lane 2) but not unstimulated(lane 1) cells by using either anti-PDGFR antibodies (Fig.6C) or anti-PY antibodies (data not shown). Approximately

VOL. 12, 1992

en

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

5828 LI ET AL.

GF/Ag

A431

N I H3T3

R B L - 2 H 3

s~~

Y(\

T--'$--T

6fi

.- % P.' *-

, I ¶

FIG. 4. Phosphoamino acid analysis. 32P-labeled Nck bandswere excised from the SDS gels, and the proteins were extracted,dried, and hydrolyzed in 5.7 N HCI. Samples mixed with phos-phoamino acid standards were analyzed by two-dimensional elec-trophoresis. The plates were stained with 0.5% ninhydrin, dried, andsubjected to autoradiography. S, phosphoserine; T, phosphothreo-nine; Y, phosphotyrosine. Exposure times: 5 to 15 days.

15 to 20% of the tyrosine-phosphorylated PDGFR wascoimmunoprecipitated by anti-Nck antibodies in PDGF-stimulated NIH 3T3 cells.To test whether Nck directly associates with activated

growth factor receptors we have used GST fusion proteins ofwild-type Nck, the SH2 domain of Nck (NckSH2), the SH3domains of Nck (NCKSH3), or GST alone coupled toagarose beads according to a previously described bindingassay (39). Lysates of EGF-stimulated or unstimulated A431cells were incubated with the various Nck-containing GSTfusion proteins immobilized on beads, washed, and boiled insample buffer containing P-mercaptoethanol. Samples weresubjected to SDS-PAGE, transferred to nitrocellulose mem-brane, and immunoblotted with anti-EGFR antibodies. Fig-ure 6B shows that Nck-GST (lanes 1 and 2) but not GSTalone (lanes 7 and 8) bound to EGFR only from lysates ofEGF-stimulated cells. The experiment presented in Fig. 6Balso shows that the binding of Nck to tyrosine-phosphory-lated EGFR is mediated via the SH2 (lane 6) and not the SH3(lane 4) domains of Nck. Similar results were obtained forthe binding of the same Nck GST fusion proteins to tyrosine-autophosphorylated PDGFR (Fig. 6D). Namely, the SH2domain ofNck is also responsible for binding to the tyrosine-phosphorylated PDGFR.PMA induces the phosphorylation of Nck in lving cells. In

addition to tyrosine phosphorylation, EGF, PDGF, andFceRI cross-linking stimulate phosphorylation of Nck onserine and threonine residues (Fig. 4). We have thereforeexamined the capacity of phorbol ester (PMA), which acti-vates protein kinase C (PKC), to stimulate Nck phosphory-

A.TIME 0EGF -

1 3 10 30+ -I -4 -+

60

* _4 -NCK

B.E G F + + -1- -4- +

-E G F R

FIG. 5. Kinetics of EGF-stimulated Nck phosphorylation. 32Pi-labeled A431 cells were treated with (+) or without (-) EGF for theindicated times at 37C. After solubilization, the anti-Nck (A) oranti-PY (B) immunoprecipitates of the cell lysates were analyzed bySDS-PAGE and autoradiography. Exposure times: panel A, 40 min;panel B, 20 min.

lation. Hence, 32Pi-labeled cells were treated with eitherPMA or EGF, washed, solubilized, and subjected to immu-noprecipitation with anti-Nck antibodies, followed by SDS-PAGE and autoradiography. Figure 7 shows that either PMA(lane 5) or EGF (lane 2) alone were able to induce more thantwo-fold stimulation of Nck phosphorylation in living A431cells. However, unlike EGF-induced phosphorylation ofNck, which took place on tyrosine, threonine, and serineresidues, PMA-induced Nck phosphorylation occurred onlyon serine residues (Fig. 7B, panel 2). We have also examinedthe ability of EGF to stimulate Nck phosphorylation in cellswhich were pretreated with PMA for 15 h in order todown-regulate PKC activity. As expected, this treatmentprevented PMA-induced Nck phosphorylation (Fig. 7, lane6). However, the ability ofEGF to stimulate Nck phosphor-ylation was not influenced by chronic pretreatment withPMA (Fig. 7, lane 3). Similar results were obtained forPDGF-induced Nck phosphorylation in NIH 3T3 cells whichwere pretreated with PMA in order to down-regulate PKCactivity (data not shown). Taken together, these resultssuggest that EGF- or PDGF-induced serine and threoninephosphorylation of Nck is not mediated via stimulation ofPKC activity.Vanadate induces phosphorylation of NcL Tyrosine phos-

phorylation of Nck can be due to activation of receptortyrosine kinases such as EGFR or PDGFR or activation ofreceptor-regulated cytoplasmic protein tyrosine kinases

MOL. CELL. BIOL.

Ak

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

Nck IS ONCOGENIC AND A TARGET FOR SURFACE RECEPTORS 5829

A.

EGF - + - +

:.

1 2 3 4

B.-+ - 4 - + - +

* EGFR

1 2 3 4 5 6 7 8

APMA

Pretreat - - +I - - +

EGF - + + - + PMA

180 -

16 ...

8 4 -

5 8 -

4 8 -

.,_404Wm m ---- N C K

C.PDGF - +

D.3 6-

_ + + + +

* "_ _~~~~~PDGFR0~~

1 2 1 2 3 4 5

FIG. 6. Association of Nck with EGFR and PDGFR in livingcells and in vitro. A431 (A and B) and NIH 3T3 (C and D) cells wereincubated with (+) or without (-) EGF or PDGF at 37°C, respec-tively, for 5 min. Anti-Nck immunoprecipitates of cell lysates wereresolved by SDS-PAGE, transferred to nitrocellulose membranes,and blotted with either anti-PY (A) or anti-PDGFR (C) antibodies,followed by labeling with "2I-protein A and autoradiography. (B)Lysates of A431 cells were incubated with 25 p.g of GST (lanes 7 and8), GST-Nck (lanes 1 and 2), GST-Nck-SH3 (lanes 3 and 4) orGST-Nck-SH2 (lanes 5 and 6) immobilized on glutathione-agarosebeads. Samples were analyzed by SDS-PAGE, transferred to nitro-cellulose membrane, and blotted with anti-EGFR antibodies. (D)32Pi_labeled and anti-PY-reactive fractions were isolated by anti-PYantibody affinity column from lysates of NIH 3T3 cells stimulatedwith or without PDGF. One part of the PY fractions was analyzedby SDS-PAGE (lanes 1 and 2) as a control, and the other sampleswere incubated with 10 1.g of GST-Nck (lane 3), GST-Nck-SH2(lane 4), or GST-Nck-SH3 (lane 5) for 30 min at room temperature,washed, and analyzed by SDS-PAGE and autoradiography.

2 6 -

2 3 4 5 6

B -PMA + PFMA1.2.

T

FIG. 7. PMA induces Nck phosphorylation. (A) 32Pi-labeledA431 cells were incubated at 37°C in the presence (+) or absence (-)ofEGF (250 ng/ml) for 5 min (lanes 1 to 3) or with (+) or without (-)PMA (100 nM) for 30 min at 37°C (lanes 4 to 6) prior to solubilizationand anti-Nck immunoprecipitation. One group of cells (lanes 3 and6) was pretreated with 10 nM PMA for 15 h before EGF (lane 3) orPMA (lane 6) stimulation. Samples were analyzed by SDS-PAGEand autoradiography. (B) Phosphoamino acid analysis of Nck phos-phorylation from dimethyl sulfoxide-treated cells (panel 1) or di-methyl sulfoxide- and PMA-treated cells (panel 2). S, phosphoser-ine; T, phosphothreonine; Y, phosphotyrosine. Exposure times:panel A, 30 min; panel B, 7 days.

augment the effect of EGF (panel 2), indicating that thesetwo treatments did not exert an additive effect on tyrosinephosphorylation of Nck. Comparison of tryptic phosphopep-tide maps of Nck revealed a similar phosphorylation pattern

such as pp6csrC. Moreover, tyrosine phosphorylation couldbe also regulated by activation or inhibition of proteintyrosine phosphatases. To study the factors which controltyrosine phosphorylation of Nck, we have examined theeffect of vanadate, an inhibitor of many tyrosine phos-phatases, on the state of phosphorylation of Nck in livingA431 cells. For this purpose 2Pi-labeled cells were incu-bated with vanadate at 37°C for 10 min prior to incubationsin the absence or presence of EGF. Cell lysates wereimmunoprecipitated with anti-Nck antibodies, followed bySDS-PAGE and autoradiography. Surprisingly, vanadatealone was able to stimulate Nck phosphorylation to aboutthe same extent as the stimulation observed with EGF aloneor when added together with EGF (Fig. 8A, lane 3 versuslanes 2 and 4). Vanadate-induced Nck phosphorylation doesnot seem to be mediated via the EGFR kinase, as autophos-phorylation of EGFR was not influenced by vanadate treat-ment (Fig. 8B, lane 3). Phosphoamino acid analysis (Fig. 8C)of phosphorylated Nck from vanadate-treated cells (panel 1)revealed phosphotyrosine as well as phosphoserine andphosphothreonine residues. Moreover, vanadate did not

V04 - - + +

E GF - + - +

- - NCK

B

t-tl _ia

2 3 4

C-

FIG. 8. Effect of vanadate on Nck phosphorylation. 32Pi-labeledA431 cells were incubated in the presence (+) or absence (-) ofvanadate (200 I1M) at 37C for 10 min, followed by treatment with(+) or without (-) EGF for 5 min. The anti-Nck (A) or anti-PY (B)immunoprecipitates were analyzed by SDS-PAGE and autoradiog-raphy. (C) Phosphoamino acid analysis of Nck from vanadate-treated (panel 1) or vanadate and EGF-treated cells (panel 2).Exposure times: A and B, 30 min; C, 10 days.

1. 2..- s

y

fta TI I

VOL. 12, 1992

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

5830 LI ET AL.

r,ia

0)0

t-

o

-C

U

--- lectiophor-esis

FIG. 9. Comparison of tryptic phosphopeptide maps of Nck. Phosphorylated Nck protein was extracted from excised gel bands (Fig. 8A,lanes 1, 2, and 3), and digested with folylsulfonyl phenylalanyl chloromethyl ketone-treated trypsin. The trypsin-digested phosphopeptideswere analyzed by electrophoresis on a thin-layer chromatography plate at pH 1.9 in the first dimension, followed by chromatography in thesecond dimension in n-butanol:puridine:acetic acid:H20 (187.5:125:37.5:150). The plates were air dried and subjected to autoradiography.Panels: Con, Nck from unstimulated cells; EGF, Nck from EGF-stimulated cells; V04, Nck from vanadate-treated cells; Mix, mixture of thesamples from EGF and V04 panels. or., origin where samples were spotted. Exposure time, 50 h.

induced by either EGF or vanadate treatment (Fig. 9),suggesting that similar sites were phosphorylated in re-sponse to the two different stimuli.

Transfection of Nck into NIH 3T3 cells causes focus forma-tion. To study the possible cellular function of Nck, weexamined the oncogenic potential of Nck by expressing NckcDNA in NIH 3T3 cells by using the retrovirus-expressingvector pLXSN (33) and performing a focus assay. As apositive control, the NIH 3T3 cells were transfected withpIBW-3 vector (2) containing the N-ras gene. Figure 10shows that foci of NIH 3T3 cells were detected in N-ras(panel C) and Nck (panel D)-transfected cells. The efficiencyof focus formation induced by Nck is approximately 20-foldlower than focus formation induced by N-ras. Immunopre-cipitation of lysed [35S]methionine-labeled cells with anti-Nck antibodies indicated that the transformed NIH 3T3 cellsexpress approximately fivefold more Nck proteins as com-pared with endogenous Nck expressed in 3T3 cells (data notshown). It appears therefore that Nck is an oncogenicprotein, and its overexpression leads to cellular transforma-tion.

DISCUSSIONThe interaction between tyrosine-autophosphorylated re-

gions in growth factor receptors and SH2 domains of signal-ling molecules appears to be crucial for determining theselectivity of signal transduction pathways (reviewed inreferences 13, 20, and 28). Therefore, an interesting problemis the identification of SH2-containing proteins and investi-gation of their interaction with growth factor receptors. Herewe show that Nck, a protein containing one SH2 and threeSH3 domains, is phosphorylated in response to EGF,PDGF, and FceRI cross-linking in various cells. Moreover,

Nck is able to interact with tyrosine-autophosphorylatedEGFR and PDGFR, and the association is mediated via theSH2 domain of Nck. Finally, overexpression of Nck resultsin transformation of 3T3 cells. Studies presented in theaccompanying reports show that Nck is phosphorylated inresponse to EGF, PDGF, and nerve growth factor stimula-tion, in response to activation of T-cell receptor or mem-brane IgM receptor and in v-src-transformed cells (6, 32, 37).These studies show that Nck is a target for many receptorand nonreceptor tyrosine kinases.The fact that Nck binds to tyrosine-phosphorylated EGFR

or PDGFR suggests that Nck is a substrate of these growthfactor receptors in the context of living cells. This notion issupported by recent experiments showing that tyrosinephosphorylation of PLC-,y depends upon the interactionbetween its SH2 domains and the tyrosine-phosphorylatedcarboxy-terminal tail of EGFR (39). Moreover, replacementof Tyr-766 of the FGFR by a phenylalanine residue abolishesassociation with and tyrosine phosphorylation of PLC--y(34), thus preventing FGF-induced phosphatidylinositol hy-drolysis. In these two cases, association of PLC--y withEGFR or FGFR is crucial for tyrosine phosphorylation andactivation of PLC--y. However, other SH2-containing pro-teins such as p85 or GRB2 do not seem to be tyrosinephosphorylated in spite of being strongly bound to tyrosine-phosphorylated PDGFR or EGFR (14, 26).The fact that vanadate treatment alone leads to tyrosine

phosphorylation of Nck similar to tyrosine phosphorylationdetected in EGF- or PDGF-stimulated cells suggests that, inthe context of living cells, EGF- or PDGF-induced tyrosinephosphorylation of Nck could be indirect and mediated bythe activation of cytoplasmic tyrosine kinase or the inhibi-tion of a specific protein tyrosine phosphatase. It is known

1 2 122 2

5 44 AM

I*7

con or. EGF or.

1 2 3 24w 3 4, 3

4 5 4

6 6~~9

10

0 oV04 otr. Mi r

MOL. CELL. BIOL.

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

Nck IS ONCOGENIC AND A TARGET FOR SURFACE RECEPTORS 5831

-4b

4 '-6~~~~~~~~~~~~~~~~~~~~P

FIG. 10. Transfected Nck causes cell transformation. Carrier (panel A), pLXSN (panel B), pIBW-3ras (panel C), and pLXSNNCK (panelD) were used in a transfection assay. Microphotographs of the foci produced by pIBW-3ras and pLXSNNCK or control cells were taken after2 weeks. Magnification, x 150.

that PDGFR activation leads to association with and activa-tion of pp60cs?C (12, 21). It is therefore possible that cyto-plasmic protein tyrosine kinases, which may be activated byPDGFR or EGFR, are in fact responsible for the phosphor-ylation of Nck on tyrosine residues. Hence, EGF- (orPDGF-) induced tyrosine phosphorylation of Nck could be aresult of activation of tyrosine kinase(s) and/or inactivationof specific protein tyrosine phosphatase(s).Our data suggest that EGF- or PDGF-induced serine/

threonine phosphorylation of Nck is not caused by PKC.Yet, PMA treatment is able to stimulate Nck phosphoryla-tion on serine residues. Thus, other growth factor-activatedserine/threonine kinases such as raf or MAP kinase (45)could be responsible for phosphorylation of Nck. It isnoteworthy that vanadate treatment leads to tyrosine as wellas serine and threonine phosphorylation of Nck, suggestingthat vanadate treatment also leads to activation of serine/threonine protein kinase(s) responsible for Nck phosphory-lation.One class of SH2-containing proteins, including crk, Nck,

and GRB2, is composed virtually of only SH2 and SH3domains. It was proposed that these proteins function asadaptors or regulatory components of their specific catalyticsubunits. For example, it is thought that PI3-kinase-associ-ated p85 is the regulatory subunit of pllO which is likely tofunction as the catalytic subunit of PI3-kinase (5). Similarto Nck, the oncogene product of the avian CT1O virusp47gagcr` is composed of only SH2 and SH3 domains. v-crkhas been shown to be associated with tyrosine-phosphory-lated proteins, and its expression leads to enhancement in

total cellular tyrosine phosphorylation (30). Moreover, mu-tations in the SH2 domain of p47gag9- decrease tyrosinephosphorylation and abolish virus-induced transformation(31). We have recently shown that GRB-2 protein, which iscomposed of one SH2 domain flanked by two SH3 domains,exhibits striking sequence homology to the Caenorhabditiselegans protein Sem-5, which together with let-23 (EGFR-like) is crucial for signal transduction pathway controllingvulval induction and sex myoblast migration in C. elegans(7, 26). On the basis of the resemblance between Nck, crk,and GRB2, it is likely that Nck is a member of the sameprotein family representing a subunit or regulatory compo-nent of an unknown downstream signalling protein.

Results presented in our report and in the accompanyingreports (6, 32, 37) provide evidence that Nck is phosphory-lated by different types of tyrosine kinases and that overex-pression of Nck leads to transformation of mammalianfibroblasts and tumor formation in nude mice. These resultsprovide further evidence that Nck plays a role in signallingpathways crucial for the control of mitogenesis. We showthat fivefold overexpression of Nck is sufficient for transfor-mation of NIH 3T3 cells. The mechanism of Nck-inducedtransformation is not clear. It is noteworthy that the avail-able cDNA clone of Nck was originally isolated from amelanoma expression cDNA library (22) and therefore maycontain a mutation(s) which activates its oncogenic poten-tial. It is therefore not clear at this time whether transfor-mation is caused by overexpression of a normal or a mutatedform of Nck.

Nevertheless, it is clear from the data presented in this and

VOL. 12, 1992

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

5832 LI ET AL.

accompanying reports (6, 32, 37) that the elucidation of thebiological role of Nck and its mechanism of action willprovide additional clues concerning the regulatory compo-nents responsible for the transduction of postreceptor sig-nals controlling mitogenesis and transformation.

ACKNOWLEDGMENTSWe thank Judith P. Johnson for human Nck cDNA and Janet M.

Oliver for RBL-2H3 cells, DNP-BSA, and monoclonal anti-DNPIgE. We thank Kiki Nelson for oligonucleotide and peptide synthe-sis and Gilad Barnea for RNA preparation.

This work was supported by a grant from SUGEN (J.S.) andHFPSO (J.S.).

REFERENCES1. Aaronson, S. A. 1991. Growth factors and cancer. Science

254:1146-1153.2. Bond, V. C., and B. Wold. 1987. Poly-L-ornithine-mediated

transformation of mammalian cells. Mol. Cell. Biol. 7:2286-2293.

3. Bradford, M. M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72:248-254.

4. Bustelo, X. R., J. A. Ledbetter, and M. Barbacid. 1992. Productof vav proto-oncogene defines a new class of tyrosine proteinkinase substrates. Nature (London) 356:68-71.

5. Cantley, L. C., K. R. Auger, C. Carpenter, B. Duckworth, A.Graziani, R. Kapelier, and S. Soltoff. 1991. Oncogenes andsignal transduction. Cell 64:281-302.

6. Chou, M. M., J. E. Fajardo, and H. Hanafusa. 1992. The SH2-and SH3-containing Nck protein transforms mammalian fibro-blasts in the absence of elevated phosphotyrosine levels. Mol.Cell. Biol. 12:5834-5842.

7. Clark, S. G., M. Stern, and H. R. Horvitz. 1992. C eleganscell-signalling gene sem-5 encodes a protein with SH2 and SH3domains. Nature (London) 356:340-344.

8. Cooper, J. A., B. M. Sefton, and T. Hunter. 1983. Detection andquantification of phosphotyrosine in proteins. Methods En-zymol. 99:387-405.

9. Davis, S., M. L. Lu, S. H. Lo, S. Lin, J. A. Bulter, B. J. Druker,T. M. Roberts, Q. An, gnd L. B. Chen. 1991. Presence of an SH2domain in the actin-binding protein tensin. Science 252:712-715.

10. Escobedo, J., D. R. Kaplan, W. M. Kavanaugh, C. W. Truck,and L. T. Williams. 1991. A phosphatidylinositol-3 kinase bindsto platelet-derived growth factor receptors through a specificreceptor sequence containing phosphotyrosine. Mol. Cell Biol.11:1125-1132.

11. Escobedo, J. A., S. Navankasattusas, W. M. Kavanaugh, D.Milfay, V. A. Fried, and L. T. Williams. 1991. cDNA cloning ofa novel 85 kd protein that has SH2 domains and regulatesbinding of P13-kinase to the PDGF 1-receptor. Cell 65:65-82.

12. Gould, K. L., and T. Hunter. 1988. Platelet-derived growthfactor induces multisite phosphorylation of pp60C-sPr and in-creases its protein-tyrosine kinase activity. Mol. Cell. Biol.8:3345-3356.

13. Heldin, C.-H. 1991. SH2 domains: elements that control proteininteractions during signal transduction. Trends Biochem. Sci.16:450-452.

14. Hu, P., B. Margolis, E. Y. Skolnik, R. Lammers, A. Ulirich, andJ. Schlessinger. 1992. Interaction of phosphatidylinositol 3-ki-nase-associated p85 with epidermal growth factor and platelet-derived growth factor receptors. Mol. Cell. Biol. 12:981-990.

15. Kashishian, A., A. Kazlauskas, and J. A. Cooper. 1992. Phos-phorylation sites in the PDGF receptor with different specifici-ties for binding GAP and P13 kinase in vivo. EMBO J. 11:1373-1382.

16. Katzav, S., J. L. Cleveland, H. E. Heslop, and D. Pulido. 1991.Loss of the amino-terminal helix-loop-helix domain of vavproto-oncogene activates its transforming potential. Mol. Cell.Biol. 11:1912-1920.

17. Katzav, S., D. Martin-Zanca, and M. Barbacid. 1989. vav, anovel human oncogene derived from a locus ubiquitously ex-

pressed in hematopoietic cells. EMBO J. 8:2283-2290.18. Kazlauskas, A., and J. A. Cooper. 1990. Phosphorylation of the

PDGF receptor 1 subunit creates a tight binding site for phos-photidylinositol 3 kinase. EMBO J. 9:3279-3286.

19. Kazlauskas, A., A. Kashishian, J. A. Cooper, and M. Valius.1992. GTPase-activating protein and phosphatidylinositol 3-ki-nase bind to distinct regions of the platelet-derived growthfactor receptor 1B subunit. Mol. Cell. Biol. 12:2534-2544.

20. Koch, C. A., D. Anderson, M. F. Moran, C. Ellis, and T.Pawson. 1991. SH2 and SH3 domains: elements that controlinteractions of cytoplasmic signaling proteins. Science 252:668-674.

21. Kypta, R. M., Y. Goldberg, E. T. Ulug, and S. A. Courtneidge.1990. Association between the PDGF receptor and members ofthe src family of tyrosine kinases. Cell 62:481-492.

22. Lehmann, J. M., G. Riethmuller, and J. Johnson. 1990. Nck, amelanoma cDNA encoding a cytoplasmic protein consisting ofthe src homology units SH2 and SH3. Nucleic Acids Res.18:1048.

23. Li, W., G. C. Deanin, B. Margolis, J. Schlessinger, and J. M.Oliver. 1992. FceRI-mediated tyrosine phosphorylation of mul-tiple proteins, including phospholipase C-yl and the receptor pY2complex, in RBL-2H3 rat basophilic leukemia cells. Mol. CellBiol. 12:3176-3182.

24. Li, W., Y. G. Yeung, and E. R. Stanley. 1991. Tyrosinephosphorylation of a common 57-kDa protein in growth factor-stimulated and transformed cells. J. Biol. Chem. 266:6808-6814.

25. Liu, F. T., J. W. Bohn, E. L. Ferry, H. Yanamoto, C. A.Molinaro, L. A. Sherman, N. R. Klinman, and D. H. Katz. 1980.Monoclonal dinitrophenyl-specific murine IgE antibody: prepa-ration, isolation, and characterization. J. Immunol. 124:2728-2736.

26. Lowenstein, E., R Daly, A. G. Batzer, W. Li, B. Margolis, R.Lammers, A. Ullrich, E. Y. Skolnik, D. Bar-Sagi, and J. Schles-singer. 1992. The SH2 and SH3 domain-containing proteinGRB2 links receptor tyrosine kinases to ras signaling. Cell70:431-442.

27. Maniatis, T., E. F. Fritsch, and J. Sambrook 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

28. Margolis, B. 1992. Proteins with SH2 domains: transducers inthe tyrosine kinase signalling pathway. Cell Growth Differ.3:73-80.

29. Margolis, B., P. Hu, S. Katzav, W. Li, J. M. Oliver, A. Ullrich,A. Weiss, and J. Schlessinger. 1992. Tyrosine phosphorylation ofvav proto-oncogene product containing SH2 domain and tran-scription factor motifs. Nature (London) 356:71-74.

30. Mayer, B. J., M. Hamaguchi, and H. Hanafusa. 1988. A novelviral oncogene with structural similarity to phospholipase C.Nature (London) 332:272-275.

31. Mayer, B. J., and H. Hanafusa. 1990. Association of the v-crkoncogene product with phosphotyrosine-containing proteinsand protein kinase activity. Proc. Natl. Acad. Sci. USA 87:2638-2642.

32. Meisenhelder, J., and T. Hunter. 1992. The SH2/SH3 domain-containing protein Nck is recognized by certain anti phospholi-pase C-yl monoclonal antibodies, and its phosphorylation ontyrosine is stimulated by platelet-derived growth factor orepidermal growth factor treatment. Mol. Cell. Biol. 12:5843-5856.

33. Milier, A. D., and G. J. Rosman. 1989. Improved retroviralvectors for gene transfer and expression. BioTechniques 7:980-986.

34. Mohammadi, M., A. M. Honegger, D. Rotin, R. Fischer, F.Bellot, W. U, C. A. Dionne, M. Jaye, M. Rubinstein, and J.Schlessinger. 1991. A tyrosine-phosphorylated carboxy-terminalpeptide of the fibroblast growth factor receptor (Flg) is a bindingsite for the SH2 domain of phospholipase C-yl. Mol. Cell. Biol.11:5068-5078.

35. Otsu, M., I. Hiles, I. Gout, M. J. Fry, F. Ruiz-Larrea, G.Panayotou, A. Thompson, R. Dhand, J. Hsuan, N. Totty, A. D.Smith, S. J. Morgan, S. A. Courtneidge, P. J. Parker, and M. D.Waterfield. 1991. Characterization of two 85 kd proteins that

MOL. CELL. BIOL.

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from

Nck IS ONCOGENIC AND A TARGET FOR SURFACE RECEPTORS 5833

associate with receptor tyrosine kinases, middle-T/pp6Oc-srccomplexes, and P13 kinase. Cell 65:91-104.

36. Park, D. J., H. K. Miin, and S. G. Rhee. 1991. IgE-mediatedtyrosine phosphorylation of phospholipase Cyl in rat basophilicleukemia cells. J. Biol. Chem. 266:24237-24240.

37. Park, D. J., and S. G. Rhee. 1992. Phosphorylation of Nck inresponse to a variety of receptors, phorbol myristate acetate,and cyclic AMP. Mol. Cell. Biol. 12:5816-5823.

38. Reedijk, M., X. Liu, P. van der Geer, K. Letwin, M. D.Waterfield, T. Hunter, and T. Pawson. 1992. Tyr 721 regulatesspecific binding of the CSF-1 receptor kinase inert to PI 3'-kinase SH2 domains: a model for SH2-mediated receptor-targetinteractions. EMBO J. 11:1365-1372.

39. Rotin, D., B. Margolis, M. Mohammadi, R. J. Daly, G. Daum, N.Li, E. H. Fisher, W. H. Burgess, A. Ulirich, and J. Schlessinger.1992. SH2 domains prevent tyrosine dephosphorylation of theEGF receptor: identification of Tyr 992 as the high-affinitybinding site for SH2 domains of phospholipase Cy. EMBO J.11:559-567.

40. Schiessinger, J. 1988. Signal transduction by allosteric receptoroligomerization. Trends Biochem. Sci. 13:443-447.

41. Shen, S.-H., L. Bastein, B. I. Posner, and P. Chretien. 1991. Aprotein-tyrosine phosphatase with sequence similarity to theSH2 domain of the protein-tyrosine kinases. Nature (London)352:736-739.

42. Skolnik, E. Y., B. Margolis, M. Mohammadi, E. Lowenstein, RFischer, A. Drepps, A. Ullrich, and J. Schlessinger. 1991. Clon-ing of PI3 kinase-associated p85 utilizing a novel method forexpression/cloning of target proteins for receptor tyrosine ki-nases. Cell 65:83-90.

43. Smith, D. B., and K. S. Johnson. 1988. Single-step purification of

polypeptides expressed in Escherichia coli as fusions withglutathione S-transferase. Gene 67:31-40.

44. Suh, P.-G., S. H. Ryu, K. H. Moon, H. W. Suh, and S. G. Rhee.1988. Inositol phospholipid-specific phospholipase C: completecDNA and protein sequences and sequence homology to ty-rosine kinase-related oncogene products. Proc. Natl. Acad. Sci.USA 85:5419-5423.

45. Thomas, G. 1992. Map kinase by any other name smells just assweet. Cell 8:3-6.

46. Trahey, M., G. Wong, R Halenbeck, B. Rubinfield, G. Martin,M. Ladner, C. M. Long, W. J. Crosier, K. Katt, K. Koths, andF. McCormick 1988. Molecular cloning of two types of GAPcomplementary DNA from human placenta. Science 242:1696-1700.

47. Ulhrich, A., and J. Schiessinger. 1990. Signal transduction byreceptors with tyrosine kinase activity. Cell 61:203-212.

48. van der Geer, P., and T. Hunter. 1991. Tyrosine 706 and 807phosphorylation site mutants in the murine colony-stimulatingfactor-1 receptor are unaffected in their ability to bind orphosphorylate phosphatidylinositol-3 kinase but show differen-tial defects in their ability to induce early response genetranscription. Mol. Cell. Biol. 11:4698-4709.

49. Vogel, U. S., R A. F. Dixon, M. D. Schaber, R. E. Diehl, M. S.Marshall, E. M. Scolnick, I. S. Sigal, and J. B. Gibbs. 1988.Cloning of bovine GAP and its interaction with oncogenic ras.Nature (London) 335:90-93.

50. Yi, T., J. L. Cleveland, and J. N. IhIe. 1992. Protein tyrosinephosphatase containing SH2 domains: characterization, prefer-ential expression in hematopoietic cells, and localization tohuman chromosome 12pl2-pl3. Mol. Cell. Biol. 12:836-846.

VOL. 12, 1992

on February 5, 2018 by guest

http://mcb.asm

.org/D

ownloaded from