Nerve Growth Factor-induced Phosphorylation

Vol. 267, No. 24, Issue of August 25, PP. 17369-17374,1992 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Nerve Growth Factor-induced Phosphorylation Cascade in PC12 Pheochromocytoma Cells ASSOCIATION OF S6 KINASE I1 WITH THE MICROTUBULE-ASSOCIATED PROTEIN KINASE, ERK1’ (Received for publication, February 7, 1992)

Jean-Claude ScimecaS, Tien T. NguyenQ,Chantal Filloux, and Emmanuel Van Obberghen From the Znstitut National de la Sante etde la Recherche MedicaleU145, Facult6 de Medecine, 06107 Nice, Cedex 2, France

Microtubule-associated protein (MAP) kinases form a group of serinelthreonine kinasesstimulated by various growth factorssuch as nerve growth factor(NGF) and hormones such as insulin. Interestingly, MAP kinases are thought to participate in a protein kinase cascade leading to cell growth as they have been shown to phosphorylate and activate ribosomal protein S6 kinase. To further evaluate the interactions between the differentcomponents of this cascade, we looked at the possible coprecipitation of MAP kinase activator(s) or MAP kinase substrate(s) with MAP kinase. Using antipeptides to the C terminus of the M, 44,000 MAP kinase, ERK1, and cell extracts from unstimulated or NGF-treated PC12 cells, we obtained in addition to MAP kinase itself coprecipitation of a protein with a M , in the 90,000 range. We further show that this protein is protein a kinase since it becomes phosphorylated on serine residues, after sodium dodecyl sulfatepolyacrylamide gel electrophoresis and transfer to a polyvinylidene difluoride membrane. In vitro phosphorylation performed before sodium dodecylsulfatepolyacrylamide gel electrophoresis demonstrates NGFsensitive phosphorylation of this 90-kDa protein on both serine andthreonine; the serinephosphorylation is likely to be due to autophosphorylation, and the threonine phosphorylation due to phosphorylation by the copurifying MAP kinase. Furthermore, immunoprecipitation of this 90-kDaprotein was obtained with antibodies to S6 kinase11. Finally, using in situchemical cross-linking, we were able to demonstrate in intact cells the occurrence of an anti-ERK1immunoreactive species with a molecular mass of approximately 125,000 compatible with a complex between ERKl and a 90-kDa S6 kinase. Taken together, our observations demonstrate that the 44-kDa MAP kinase is associated, in intact PC12 cells, with a protein kinase which is very likely to be S6 kinase 11. In conclusion, our data represent strongevidence for a physiological role of the MAP kinase436 kinase cascade in PC12 cells. Finally, our antipeptidesprovide us with a pow-

* This workwas supported in part by funds from the Institut National de la Santi et de la Recherche Midicale; the Universite de Nice-Sophia-Antipolis; Grant 6760 from the Association pour la Recherche contre le Cancer; and theLigue Nationale Fran~aisecontre le Cancer, Fediration des Comites Dipartementaux, Comiti Dipartemental du Var.The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ Recipient of a fellowship from the Ligue Nationale Francaise Contre le Cancer, Federation des Comitis Dipartementaux, Cornit6 Departemental du Var. § Recipient of a postdoctoral fellowship from the Medical Research Council of Canada.

erful tool to search for additional physiologically relevant substrates for MAP kinase, a key integrator enzyme for growth factors and hormones.

During the last 5 years much research in the area of cell signaling has focussed on the serine/threonine MAP kinases, which are thought to play a central role in metabolic and mitogenic effects induced by various extracellular stimuli. In response to most growth-promoting and mitogenic factors tested, as well as tomany other signals such as phorbol esters or phosphatase inhibitors, MAP’ kinase phosphorylation is increased on both threonine andtyrosine residues, leading to the stimulation of its kinase activity (for review see Ref. 1). While tyrosine protein kinase receptors have been implicated in a wide variety of physiological functions, the understanding of the molecular mechanisms of their actions continues to represent a major challenge. Dissection of the transductional cascades induced by insulin (2), EGF (3-5), and NGF (6, 7) has identified MAP kinases as a link between the tyrosine kinase receptors of these polypeptides, and serine/threonine phosphorylations. Although the precise role(s) of MAP kinases remain(s) to be clarified, they have emerged as key “switch kinases,” i.e. serine/threonine kinases which are capable of converting a tyrosinephosphorylation signal coming from a tyrosine kinase receptor into a serine/threonine phosphorylation. In 1990Cobb et al. (8, 9) obtained an almost complete sequence of an “extracellular signal-regulated kinase 1”(ERKl), which was found to be closely related to two yeast kinases involved in cell cycle control, and whose properties strongly indicated identity with previously identified MAP kinases. Moreover, their studypresented evidence for the existence of a family of at least 4 ERK proteins with molecular weights ranging from 41,000 to 62,000 (10, 11). Using immunoprecipitation with antipeptides to the carboxyl terminus of ERK1, we have shown that mouse fibroblast ERKl was phosphorylated in vitro on both threonine and tyrosine. Moreover, this dual phosphorylation was concomitant with an enhanced ERKl kinase activity as measured by myelin basic proteinphosphorylation (12). However, this activation remained small compared to the level of ERKl The abbreviations used are: MAP, microtubule-associated protein; BSA, bovine serum albumin; C peptide, ERKl peptide/sequence 356-367; DTT, dithiothreitol; EGF, epidermal growth factor; ERK, extracellular signal-regulated kinase; Hepes, N-2-hydroxyethylpiperazine-N’-2-ethanesulfonic acid ISPK, insulin-stimulated protein kinase; NGF, nerve growth factor; PBS, phosphate-buffered saline; PMSF, phenylmethylsulfonyl fluoride; PVDF, polyvinylidene difluoride; Rsk, ribosomal protein S6 kinase; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TLC, thin-layer chromatography; DSS, disuccinimidyl suberate.

17369

17370

NGF-induced Phosphorylation of ERKl

activity observed after in uiuo treatment of cells with growth factors or hormones such as insulin. At the same time three groups reported independently that ERKl and ERK2 obtained by expression in Escherichia coli was phosphorylated in vitro on both tyrosine and threonine and that this was concurrent with a slow and moderate activationof the enzyme (13-15). To summarize, the available data suggest that, although ERKl/ERK2 dual self-phosphorylation can occur and might be physiologically relevant, the mechanism of ERKl activation is likely to be more complex and probably involves other proteins such as activator(s). The latter could be protein address the kinases or allosteric modulators. In an attempt to phosphorylation cascade issue, we investigated here the possibility of a coprecipitation of protein(s) with ERK1. Using PC12 cells and an antipeptide to the carboxyl terminus of ERK1, we immunopurified ERKl from unstimulated or NGFtreated cells. After elution from antipeptide precipitates, we performed in vitro phosphorylation experiments. The following key observations were made: (i) NGF induced a stimulation of 32Pincorporation into a protein in the 90-kDa range; (ii) thisphosphorylation occurred to the same extent on both threonine andserine residues. We next addressed the question of a possible autophosphorylation of this protein. To thisend we performed a denaturation/renaturation procedure after transfer to a PVDF membrane, followed by a kinase assay reaction. We observed phosphorylation of both ERKl and the 90-kDa protein, and phosphoamino acid analysis of the 90kDa phosphoprotein identified solely serine residues. Furthermore, using anti-S6 kinase I1 antibodies, we obtained specific precipitation of a 90-kDa protein from anti-ERK1 antipeptide eluates. Finally, using insitu chemical crosslinking, we observed by immunoblotting with anti-ERK1 antipeptides, a protein with an electrophoretic mobility consistent with a complex between ERKl and S6 kinase 11. As a whole, our data indicate that ERKl is associated in living PC12 cells with a phosphoprotein of 90 kDa, which corresponds very likely to the mammalian homologue of the frog S6 kinase 11. EXPERIMENTALPROCEDURES

Materials-C peptide (356-367: TARFQPGAPEAP), corresponding to the C terminus of ERKl (8), was produced by Neosystem (Strasbourg, France). Antipeptides against this region of ERKl were obtainedas previously described (12). Immobilon PVDFtransfer membrane was from Millipore. Bovine serum albumin (BSA; type 7030) for cell culture, and dimethyl sulfoxide were from Sigma. Disuccinimidyl suberate (DSS) was from Pierce. BSA for immunoblot experiments was from Intergen Company (Providence,RI). Anti-Rsk serum (Rsk: ribosomal protein S6 kinase), from a rabbit injected with bacterially produced Rsk protein, was a generous gift of Drs. E. Erikson and J. Maller (Department of Pharmacology, University of Colorado Health Sciences Center,Denver,CO). NGF was kindly provided to us by Dr. P. Kitabgi (Nice-Sophia-Antipolis, France). Cell Culture-PC12 cells were cultured in RPMI medium containing 10% horse serum and 5% fetal calf serum. Cells were plated at 2.5 X lo4 cells/cm* and grown to confluence. Before incubation with the effectors, cells were starved overnight in RPMI/O.2% BSA. NIH 3T3 cells, clone HIR 3.5 (HIR cells), transfected with a human insulin receptor cDNA construct and expressing lo6 receptors/cell, were provided to us by Dr. J. Whittaker (Stony Brook, New York, NY) (16). HIR cells were cultured to confluence in H21/10% fetal calf serum and starved overnight in H21/0.2% BSA. I n Vivo Stimulation of ERKl Actiuity and ERKl Immunopurification-After the starvation period, HIR or PC12 cells were incubated for 5 min in presence of effectors: insulin M) for HIR cells, and NGF(6 x lo-' M ) or EGF M ) for PC12 cells. All the following steps were performed at 4 "C. Cells were washed twice with ice-cold PBS (phosphate-buffered saline; 140 mM NaC1, 3 mM KC1, 6 mM NaZHP04,1 mM KH2P04, pH7.41, and once with buffer A (50 mM Hepes, pH 7.5, 150 mM NaCl, 10 mM EDTA, 10 mM Na4P207,2 mM sodium orthovanadate, 100 mM NaF, 100 units/ml aprotinin, 20 p M

leupeptin, and 0.18 mg/ml PMSF). Cells were scraped, and proteins were solubilized for 15 min in buffer A supplemented with 1%Triton X-100 (solubilization buffer). Cell extracts were then submitted to centrifugation (18,000 X g for 15 min), and samples were incubated for 90 min at 4 "C with anti-ERK1 antibodies adsorbed on protein A-Sepharose pellets. 32PLabeling in Intact PC12 Ceh-Confluent PC12 cells, growing in 145-mm culture dishes, were starved overnight in RPMI, 0.2% BSA medium (BSA 7030 from Sigma). Cells were then washed twice with Dulbecco's modified Eagle's medium withoutphosphate and incubated for 3.5 h in this medium containing 500 pCi of["PI orthophosphate (1.7 mCi/ml). At the end of labeling, NGF (6 X lo-@ M) was added for 5 min. After three washes with ice-cold PBS, proteins were solubilized for 15 min on ice in 50 mM Hepes, pH 7.5, containing 150 mM NaCl,10 mM EDTA, 10 mM Na4P,07, 2 mM sodium orthovanadate, 100 mM NaF, 1% TritonX-100,100 units/ml aprotinin, 20 p M leupeptin, and 0.18 mg/ml PMSF. Cell extracts were then incubated for 90 min with antibodies to ERKl adsorbed on a protein A-Sepharosepellet. After five washes with solubilization buffer, Laemmli buffer (3% sodium dodecyl sulfate) was added to driedpellets, andproteins were submitted toSDS-PAGE under reducing conditions. I n Vitro Phosphorylation of Immunopurified ERKl-At the end of incubation with anti-ERK1 antipeptides, protein A-Sepharose pellets were washed five times in solubilization buffer. After an additional wash in HNTG buffer (50 mM Hepes, pH 7.5, 150 mM NaCI, 0.1% Triton X-100, 10% glycerol, 20 PM leupeptin, 100 units/ml aprotinin, 0.18 mg/ml PMSF), ERKl was eluted by a 30-min incubation at room temperature in HNTG buffer containing 10 p~ C peptide and 0.2 mM sodium orthovanadate. I n vitro phosphorylation was performed for 1 h at room temperature in the presence of 5 mM MnC12, 10 mM MgAc, and [y-32P]ATP (5 pM, 33 Ci/mmol). The reaction was terminated by addition of 4-fold concentrated Laemmli buffer and samples were submitted to SDS-PAGEunder reducing conditions on a 10% acrylamide resolving gel. Phosphoamimacid Analysis-For phosphoamino acid analysis, phosphorylation was performed as described above, and samples were submitted to SDS-PAGE under reducing conditions. After electrophoresis, 32P-labeledproteins were localized by autoradiography, and gel pieces corresponding to the phosphoproteins of interest were excised. Labeled proteins were then eluted from the gel by an overnight incubation at 37 "C in 50 mM NH4HC03, pH 8,0.1% SDS, and 5% 0-mercaptoethanol. Eluted proteins were precipitated for 30 min on ice in the presence of 10% trichloroacetic acid and 25 pg of bovine y-globulin as a carrier. Pellets were washed once with 100% EtOH and once with EtOH/ethylic ether (1:l).Proteins were hydrolyzed for 90 min at 110 "C in 6 N HCl. Phosphoamino acids were then separated on cellulose thin-layer plates by electrophoresis at pH 3.5 for 2 h at 1000 V and analyzed by autoradiography as previously described (17). Protein Transfer, Renaturation Procedure, and Kinase ReactionAfter immunopurification and elution from anti-ERK1 antibodies, eluates were submitted to SDS-PAGE under reducing conditions on a 10% acrylamide resolving gel. Proteins were transferred to a PVDF membrane, and the membrane was incubated for 60 min at 4 "C in a denaturation solution (6 M guanidinium chloride, 50 mM Tris-HC1, pH 8.3, 50 mM DTT, 2 mM EDTA).ThePVDF membrane was washed twice with renaturation buffer (20 mM Tris-HC1, pH 7.5, 150 mM NaC1,2 mM DTT, 2 mM EDTA, 0.1% Nonidet P-40,2% glycerol), and incubated overnight at 4 "C in renaturation buffer. The membrane was saturated by incubation for 60 min at room temperature in renaturation buffer containing 0.2% polyvinylpyrrolidone, 0.2% ficoll, and 20 mM sodium pyrophosphate. The kinase assay was then performed in HNTG buffer containing 10 mM magnesium acetate, 5 mM manganese chloride, and 5 p~ [y-32P]ATP(50 &i/ml). After 1 h of phosphorylation at room temperature,the blot was washed extensively with PBS containing 1% Triton X-100 prior to autoradiography. Concerning the ERKlimmunoblot after kinase assay, the experiment was performed as previously described (121, except for a step of rehydration of the PVDF membrane (after autoradiography) in the saturating buffer. I n Situ Cross-linking-After a 5-min incubation in the absence or presence of NGF, PC12 cells were washed twice with ice-cold PBS, and incubated for 20 min at 4 "C in PBS/1.2% (v/v) dimethyl sulfoxide as a controlor in PBS/l.2% (v/v) dimethylsulfoxide containing 1.2 mM DSS as previously described (18). Cells were then washed with ice-cold PBS, and cellular proteins were solubilized before immunoprecipitation as described above.

17371

NGF-induced Phosphorylationof ERKl NGF ADDED TO PC12 CELLS:

RESULTS

I n Vivo and in Vitro Phosphorylation of a 90-kDa Protein Coprecipitatingwith ERK1-To search for coprecipitation with ERKl of cellular activators and/or substrates,confluent HIR and PC12 cells were incubated for 5 min with buffer or with insulin M) for HIR cells, and with buffer, NGF (6 X lo-* M) or EGF ( M) for PC12 cells. After solubilization, proteins were submitted to precipitation by antipeptides to ERK1, and thewashed pellets were incubated with C peptide (10 p M ) . I n vitro phosphorylation was then performed on eluates, and samples were analyzed by SDS-PAGE under reducing conditions (Fig. 1). In both cell types, we observed two major phosphoproteins, one in the range of 90 kDa, and another one at 44 kDa corresponding to ERK1. Furthermore, we found that insulin, NGF or EGF strongly stimulated the :IpPincorporation into the 90-kDa species. The band with an estimated molecular weight of 46,000, and whose phosphorylation is significantly stimulated by insulin in HIRcells, could be also detected on shorter exposures of phosphoproteins in eluates from NGF- and EGF-treatedPC12 cells. In additional experiments, we performed a 32Plabeling of intact PC12 cells, followed byNGF treatmentfor 5 min. After solubilization and immunoprecipitation with anti-ERK1 antipeptides, the phosphoproteins adsorbed on the pellet were eluted by incubation with C peptide, and submitted to SDSPAGE analysis under reducing conditions. As shown on Fig. 2, in addition to ERKl which showed increased phosphorylation afterNGF treatment, coprecipitation of an in vivo labeled 90-kDa phosphoprotein was revealed (upper arrow). Moreover, NGF induced the stimulation of 82Pincorporation intothis coprecipitatingprotein. Taken together, we can conclude that in living PC12 cells ERKl was associated with a NGF-sensitive 90-kDa phosphoprotein. Phosphoarnino Acid Analysisof the in Vitro Phosphorylated 90-kDa Protein-We next determinedthe phosphoamino acid content of the 90-kDa phosphoprotein obtained from PC12 cells incubated with NGF. After a 5-min treatment of PC12 cells with NGF, precipitation,elution, and in vitro phosphorylation were performed as described above. The 90-kDa phosphoprotein was eluted from the gel, andafter acidic hydrolysis, phosphoamino acids were analyzed by TLC. As shown in Fig. 3, a basal ”P incorporation into threonine and CELLS: OR

-

HIR

PC12

”

?Mi-

111

-

71

-

48.5

-

”

2 9 MrxlO‘ ADDEDTOCELLS:

BUFFER

INSULIN

BUFFER

NGF

EGF

FIG.1. In vitro phosphorylation of a 90-kDa phosphoprotein coprecipitating with ERK1. Serum-depleted HIR or PC12 cells were incubated for 5 min in presence of insulin (lo-’ M), NGF (6 X lo-” M),or EGF (lo-’ M). Solubilized proteins were submitted t o precipitation by anti-ERK1 antibodies, and after elution by C peptide eluates were phosphorylated in uitro in presence of [-p”P] A T P a sdescribed under “Experimental Procedures.” Phosphorylated proteins were analyzed by SDS-PAGE ona 10% acrylamide resolving gel under reducing conditions.

200

-

116

-

97

-

66

-

45

-

OR

NO

YES

-

ERKl

Mr x 10”

FIG. 2. Immunoprecipitation of a 90-kDa phosphoprotein by antibodies to ERKl after “P labeling of living PC12 cells. Confluent PC12 cells growing in 145-mm Petri dishes were starved overnight, and the ‘”P labeling was performed as described under “Experimental Procedures.” NGF (6 X lo-’ M) was added for 5 min, and solubilized proteins were submitted to precipitationby antipeptides to ERK1.After washes, phosphoproteins adsorbed on thepellet were eluted by incubation with C peptide, and sampleswere analyzed by SDS-PAGE on a 10% acrylamideresolving gel under reducing conditions. The upper arrow indicates the position of the 90-kDa phosphoprotein.

-

IN V m O PHOSPHORYLATION OF 9OkD PROTEINS

ADDED70 PC12 CELLS

PHOSPHOAMINOACIOANALYSIS

T

FIG. 3. Phosphoamino acid analysis of the in vitro phosphorylated 90-kDa phosphoprotein. PC12 cells were incubated in presence of NGF (6 X lo-” M ) for 5 min and in vitro phosphorylation of the coprecipitating 90-kDa protein was performed as described in Fig. l (left panel). The phosphorylated90-kDa protein was eluted from the gel, submitted to hydrolysis in 6 N HCI, and phosphoamino acids were analyzed by TLC (right panel).

serine residues was detected. Furthermore, NGF induced a strong stimulation of the labeling on the same 2 residues. Antipeptides toERKl Coprecipitate an Autophosphorylating Protein Kinase withM, 90,000-We were interested in determining whether this 90-kDa protein was a substrate protein for ERKl and/or whether it was a protein kinase endowed with the capacity to undergo autophosphorylation. To do so, a kinase renaturation assay was performed with solubilized proteins from unstimulated PC12 cells, which were submitted to precipitation with preimmune or antibodies to ERKlprotein. After transfer to a PVDF membrane, renaturation and incubation with [T-~~PIATP, blot the was extensively washed and exposed to an autoradiographic film (Fig. 4, right panel). In these conditions,we observed with anti-ERK1 antipeptides a specific immunoprecipitation of two phosphoproteins: (i) ERKl at 44,000, as shown by the Western blot experiment

NGF-induced Phosphorylationof ERKl

17372 WSTERN BLOT WITH ANnERKI AHnPEPTlDES

OR

-

, _"

KINASE R M A N R A T K ) H ASSAY

..

F

2 "

contrast to the eluate phosphorylation experiments, this"P incorporation occurred exclusively on serineresidues, indicating that the 90-kDa coprecipitating protein displayed the capacity to undergo autophosphorylation with a specificity for serine residues.

Precipitation of the in VitroPhosphorylated 90-kDa Protein by Antibodies to S6 Kinuse ZI-The experiments described above indicated to us that an autophosphorylating 90-kDa 71 in vitro proteinkinasecoprecipitatedwithERK1.When phosphorylation was performed in eluates containing E R K l 48.5 and the 90-kDa protein kinase, phosphoamino acid analysis showed that the 90-kDa proteinwas phosphorylated on both serine and threonine residues. In contrast, this phosphorylation occurred exclusively onserine residueswhen it was 2 9 performed on the membrane after transfer and renaturation. Mr x 10' As it has been previously described that MAP kinase in in uitro reconstitution experiments is able to phosphorylate and SERVY. activate the 90-kDa S6 kinase I1 from Xenopus oocyte, we FIG. 4. Kinase renaturation assay of the 90-kDa coprecip- next tested the hypothesis that the 90-kDa phosphoprotein itating protein. Proteins from unstimulated PC12 cells were sub- we described here could in fact correspondto the mammalian equivalent of the frog protein. To doso, after coprecipitation mitted to precipitation by preimmuneandimmuneantibodiesto ERK1. After SDS-PAGE under reducing conditions and tranfer toa of the 90-kDa protein and in vitro phosphorylation, samples PVDF membrane, a kinase renaturation assay was performed (right from unstimulatedor NGF-treated PC12cells were submitted panel), followed by a Western blot with anti-ERK1 antibodies (left to precipitation by nonimmune antibodies or antibodies dipanel). The upper arrow indicates the position of the 90-kDa phosrected against bacterially produced Xenopus Rsk protein. As phoprotein. shown in Fig. 6, we observed a specific immunoprecipitation of a 90-kDa phosphoprotein, while no signal was detected KINASE RENATURATION ASSAY PHOSPHOAMINOACID ANALYSIS using the nonimmune antibodies. We interpret these data to mean that the 90-kDa coprecipitating protein kinase is very likely to be the mammalian equivalent of the frog S6 kinase 11. More important, our data indicate that this S6 kinase I1 is associated with ERKl in living PC12 cells, as shown by the coprecipitation of the two proteins. I n S i t u Cross-linking of ERKl and a 90-kDa Phosphoprotein-In order to establish that ERKl and the 90-kDa copre18.5 cipitating protein were associated in intact cells before solubilization, we performed a 5-min NGF treatment of PC12 cells followed by a n incubation with DSS in the presence of dimethyl sulfoxide which permeabilizes cells. Then cellular proteins were solubilized and submitted to immunoprecipita111

-

OR ADDED TO PC12 CELLS

BUFFER

NGF

-

PRECIPITATION WITH: BUFFER

NGF

FIG. 5. Phosphoamino acid analysis of the 90-kDa protein phosphorylated on transfer membrane.Kinase renaturation assay was performed withthecoprecipitating 90-kDa protein from unstimulated (buffer) or NGF-treated PC12 cells (6 X lo-* M for 5 min) (leftpanel). After autoradiography,membrane pieces corresponding to phosphorylated proteins were submitted tohydrolysis in 6 N HCI, andphosphoaminoacids were analyzed by TLC (right panel ) .

OR

-

NON-IMMUNE ANTIBODY

ANTIBODY TO 56 KINASE II

W

zoo116 97

-

45

-

c1

performed on the same membrane with anti-ERK1 antibodies of 90 kDa(upper (Fig. 4,left panel);(ii) a protein in the range arrow). Note that in some experiments this 90-kDa protein appeared as a doublet after short exposure of the autoradiNOFADDED "+ ographic film. "+ T O PC12 CELLS: As these data supported the notion that the 90-kDa copre- FIG. 6. Immunoprecipitation with antibodies to S6 kinase cipitating proteinwas a n autophosphorylating protein kinase, I1of in vitro phosphorylated eluates from anti-ERK1 antibody pellets. In vitro phosphorylation of the coprecipitating 90-kDaprothe samekinaserenaturationassay was performedusing extracts from unstimulated and NGF-treated PC12 cells (Fig. tein was performed with eluates from unstimulated or NGF-treated M for 5 min). The phosphorylation reaction was 5, left panel), and phosphoamino acid content of this 90-kDa PC12 cells (6 X stopped by the addition of NaF/EDTA (100 mM and 20 mM, respecprotein was determined by TLC (Fig. 5, right panel). Similar tively), and samples were submitted to precipitation by nonimmune to the observations made in the eluate phosphorylation ex- antibodies and antibodies to S6 kinase 11. Phosphorylated proteins periments, NGFwas found to induce a small, butreproducible, adsorbed on the washed pellets were analyzed by SDS-PAGE on a stimulation of "P incorporation into the 90-kDa protein. In 10%acrylamide resolving gel under reducing conditions.

NGF-induced Phosphorylation of ERKl

17373

ERKl activity also promoted the 32Pincorporation into this 90-kDa protein; (iii) in vitro phosphorylation of this 90-kDa protein, in presence of ERK1, occurred on both serine and threonine residues, while phosphorylation on transfer membranes revealed an incorporation into serine residues exclusively; (iv) specific immunoprecipitation of this 90-kDa protein could be obtained with antibodies to the frog S6 kinase 11. In additional experiments, we performed in vivo 32P-labeling of PC12 cells followed by a 5-min incubation with NGF. After immunoprecipitation by anti-ERK1 antibodies and elution with the C peptide, coprecipitation of the in vivo labeled 90-kDa protein could be revealed. Moreover, phosphorylation of this protein was stimulated by exposure of PC12 cells to NGF. Our in situ cross-linkingexperiments show that in living PC12 cells, ERKl is associated with a protein with a molecular weight in the range of 90,000, and that this complex formation is stimulated by NGF. These experiments also indicate that this association is not generated by the cell solubilization procedure itself. As a whole, our data strongly DISCUSSION support thenotion of the physiological relevance of an assoOf great importancefor the idea of a ligand-activated phos- ciation in living PC12 cells between ERKl anda mammalian phorylation cascade involved in hormone and growth factor form of S6 kinase 11. Concerning the specificity of this interreceptor tyrosine kinase action, was the original demonstra- action, it would be interesting todetermine which domains of tion by Sturgill et al. (19) that 42 kDa/MAP kinase purified the two proteins are involved in this molecular association. from insulin-stimulated 3T3-Ll cells was able to phosphoryl- Note that thecoprecipitation we describe here is reminiscent ate and to activate, i n vitro, the S6 kinase I1 from Xenopus of recent reports by several groups describing association and oocytes. Similarresults were thereafter obtained by other coprecipitation of tyrosine kinase receptorsand some of their groups using progesterone-stimulated Xenopus oocyte ex- primary targets, such as phospholipase C-y and phosphatidyltracts (20), and EGF-stimulated Swiss 3T3 fibroblast extracts inositol-3-kinase (21-23). I n vitro phosphorylation experiments shown in Figs. 1 and (3, 4). Therefore, our findingthat ERKl isassociated with a 90-kDa phosphoprotein under various conditions and in dif- 3 suggest that the activated ERKl is capable of in vitro ferent cell types led 1:s to test the hypothesis that thisprotein interaction with the 90-kDa protein. According to this view the activated ERKl leads to threoninephosphorylation of the could correspond to a S6 kinase I1 or to a related protein. Several lines of evidence are in fact infavor of this view, and 90-kDa protein, which then becomes competent to undergo are as follows: (i) antipeptides to ERKl coprecipitated a 90- autophosphorylation on serineresidues. The high level of 32P kDa phosphoprotein; (ii) effectors capable of stimulating incorporation seen in cell-free systems after immunopurification might be explained by the loss, during immunopurification, of endogenous protein phosphatase(s) involved in regNGF ulation of these protein kinases. This issue is particularly r - 4+I intriguing, since we observed in our previous work (12) that OR in vitro ERKl phosphorylation is enhanced by sodium ortho221 vanadate addition to the phosphorylation mixture. While we cannot exclude at present that thiseffect could be accounted for by a direct interactionbetween orthovanadate and ERK1, we favor the idea that these data might reflect the copurifi106 cation of one or several sodium orthovanadate-sensitive protein phosphatase(s). 75 It should be noted that the antibodies to Xenopus oocyte S6 kinase I1 precipitate specifically only a small fraction of the 90-kDa protein when compared to the totalamount submitted to immunoprecipitation (data not shown). This could 46 be explained by the specificity of this antibody raised against bacterially produced Rsk protein and by the possible limited conservation of the amino acid sequence between frog and mammalianproteins. However, we have also observed, in NO YES some renaturation experiments, two phosphoproteins in the molecular species corresponding to the 90-kDa band. Hence, DSS FIG. 7. Immunoblot of anti-ERK1 immunoprecipitates after it is possible that one protein canbe recognized as a genuine in situ cross-linking with DSS. After a 5-min incubationof PC12 S6 kinase, while the other one could in fact be a proteolytic cells with NGF, in situ cross-linking by DSS was performed as degradation product. If the protein with the fastest electrodescribed under “Experimental Procedures.” Samples were then sol- phoretic mobility represents a proteolytic degradation produbilized, submitted to immunoprecipitation by anti-ERK1 serum, uct, the enzyme renaturation experiments indicate that this and separated by SDS-PAGEon a 10% acrylamide resolving gel proteolysis does not impair in a significant way the ability of under reducing conditions. Immunoprecipitated proteins were transferred to a PVDF membrane, incubated in presence of anti-ERK1 the proteolytic fragment to undergo autophosphorylation. Recently, purification and characterization of an insulinantibodies, and revealed with ’251-proteinA. The arrow indicates the position of ERKl/-gO-kDa complex. stimulated protein kinase (ISPK1) from rabbit skeletal mus-

tion by anti-ERK1 serum. After SDS-PAGE, samples were transferred to a PVDF membrane and finally blotting with anti-ERK1 serum was used to search for ERKl alone, and ERKl cross-linked to other protein(s). As shown in Fig. 7, we found that anti-ERK1antibodies detect in solubilisates from cells incubated in the presence of DSS, ERK1, and addition, in a molecular species with a molecular weight of approximately 125,000. This is compatible with a complex between ERKl and a 90-kDa protein. Moreover, NGF-treatment resulted in a n enhanced appearance of the 125-kDa species. In a control experiment, we have performed cell lysate precipitation using preimmuneserum before SDS-PAGE and immunoblotting with anti-ERK1 serum (datanot shown). No signal was detected compatible with ERKl or in the range of 125 kDa. In summary, our in situ cross-linking experiments strongly suggest that a complex between ERKl and the90-kDa phosphoprotein occurs in living PC12 cells and that thisassociation is increased upon stimulation of ERKl by NGF.

-

-

-

-

17374

NGF-induced Phosphorylation of ERKl

cle was reported by the laboratory of P. Cohen (Department of Biochemistry, University of Dundee, Dundee, UK) (24). The authorsdemonstrated that ISPKlwas closely related, if not identical, to the frog S6 kinase 11. ISPKl appears to play a major role in glycogen metabolism as it phosphorylates the G subunit of phosphatase 1, that in turn leads to dephosphorylation and activation of glycogen synthase (25). Together thesedata demonstrate that besides its effect onprotein synthesis via the S6 kinase II/S6 proteinpathway,MAP kinase is a key kinase in the regulation of major metabolic responses such as glycogen synthesis. It remains to be shown whether such an ISPKl exists in PC12 cells, and whether NGF plays a role in glycogen metabolism in these cells as well. The nature of the insulin-, EGF-, and NGF-sensitive 46kDa phosphoprotein we observed in Fig. 1 is unknown. It might correspond to a copurifying ERKl substrate or to an activator. Recently, Ahn et al. (26) and Gomez et al. (27) have reported the first stepsof purification and characterization of MAP kinase activators. Using extracts from EGF-treated Swiss 3T3 cells or from NGF-treated PC12 cells, both groups found two MAP kinase activators with a molecular weight in the range of50,000-60,000. A t the present time, it is not known whether these MAP kinase activators are themselves protein kinases. Ahn et al. (26) are leaning toward the idea that these activatorsare notkinases mainly based on the fact that neither proteins nor peptides representing a broadrange ofknown specificities for phosphorylation could be phosphorylated. Hence the action of these activating molecules could be limited to theinduction of a conformational change in MAP kinase, with aresulting increase in the rate and extent of the autophosphorylation. However, it remains possible that they are protein kinases with a virtually unique specificity for MAP kinases. In summary, we have shown that antipeptides to ERKl precipitate a functional ERKl/SO-kDa phosphoprotein complex, in which the 90-kDa species undergoes in vitro phosphorylation on serine and on threonine residues; the latter phosphorylation likely to be directly mediated by ERK1. Moreover, this 90-kDa phosphoprotein is specifically precipitated by antibodies to Xenopus oocyte S6 kinase 11. As a whole, our results provide strong evidence for the idea that in intact PC12cells a mammalian equivalent of the frog S6 kinase I1 is associated with the MAP kinase/ERKl.The essence of the whole concept of a kinase cascade triggered by a ligand-activated receptor tyrosine kinase and involving MAP kinase and S6 kinase rests for the major part on clever reconstitution experiments. Our data clearly show that MAP

kinase and S6kinase form functional complexes in intact cells and are therefore solid arguments in favor of an important physiological role of these two kinases in growth factor and hormone signaling pathways. Acknowledgments-We express our sincere thanks to Dr. Y. Le Marchand-Brustel, Dr. E. Van Obberghen-Schilling, Dr. V. Baron, and Dr. R. Ballotti for reviewing the manuscript and constructive comments. G. Visciano is acknowledged for illustration work.We thank Dr. E. Erikson and Dr. J. Maller (Department of Pharmacology, University of Colorado Health Sciences Center, Denver, CO) for the gift of anti-Rsk antibody, Dr. J. Whittaker (Stony Brook, New York, NY) for the HIR cells, and Dr. P. Kitabgi (Nice-SophiaAntipolis, France) for the gift of NGF. REFERENCES 1. Thomas, G. (1992) Cell 6 8 , 3 - 6 2. Ray, L. B., and Sturgill, T. W. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 1502-1506 3. Ahn, N. G., and Krebs, E. G. (1990) J. Biol. Chem. 265,11495-11501 4. Ahn, N. G., Weiel, J. E., Chan, C. P., and Krebs,E. G. (1990)J.Biol. Chem. 265,11487-11494 5. Hoshi, M., Nishida, E., and Sakai, H. (1988) J. Biol. Chem. 2 6 3 , 53965401 6. Gomez N Tonks N. K. Morrison, C., Harmar, T., and Cohen, P. (1990) FEB'S i k t t . 2 7 i , 119-i22 7. Mi asaka, T., Chao, M. V., Sherline, P., and Saltiel, A. R. (1990) J. Biol. $hem. 265,4730-4735 8. Boulton, T. G., Yancopoulos, G.D., Gregory, J. S., Slaughter, C., Moomaw, C., Hsu, J., and Cobb, M. H. (1990) Science 249,64-67 9. Boulton, T. G., Gregory, J. S., and Cobb, M. H. (1991) Biochemistry 3 0 ,

-."

-"

97Q-9QC

10. Boulton, T. G., and Cobb, M. H. (1991) Cell Regul. 2,357-371 11. Boulton. T. G.. Nve. S. H.. Robbins. D. J.. ID. N. Y.. Radzieiewska. E.. Morgenbesser, S: D., DePinho, R. A., Panay&tos, N:, Cobb,". H.,'and Yancopoulos, G. D. (1991) Cell 65,663-675 12. Scimeca, J.-C., Ballotti,R., Nguyen, T. T., Filloux, C., and Van Obberghen, E. (1991) Biochemistry 30,9313-9319 13. Crews, C. M., Alessandrini, A. A,, and Erikson, R. L. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,8845-8849 14. Seger, R., Ahn, N. G., Boulton, T. G., Yancopoulos, G. D., Panayotatos, N., Radzie'ewska, E., Ericsson, L., Bratlien, R. L., Cohb, M. H., and (1991) Proc. Natl. Acad. Sci. U. S. A . 88, 6142-6146 Krebs, E. 15. Wu, J., Rossomando, A. J., Her, J.-H., Vecchio, R. D., Weber, M. J., and Sturgill, T. W. (1991) Proc. Natl. Acad. Sci. U. S. A. 88,9508-9512 16. Whittaker, J., Okamoto, A. K., Thys, R., Bell, G. I., Steiner, D. F., and Hofmann, C. A. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,5237-5241 17. Cooper, J. A., Sefton, B. A,, and Hunter, T. (1983) Methods Enzymol.9 9 , 387-402 18. Hari, J., Shii, K., and Roth, R. A. (1987) Endocrinology 120,829-831 19. Sturgill, T. W., Ray, L. B., Erikson, E., and Maller, J. L. (1988) Nature 334,715-718 20. Haccard, O.,Jessus, C., Cayla, X., Goris, J., Merlevede, W., and Ozon, R. (1990) Eur.J. Biochem. 192,633-642 21. Margolis, B., Rhee, S. G., Felder, S., Mervic, M., Lyall, R., Levitzki, A., Ullrich, A,, Zilberstein, A., and Schlessinger, J. (1989) Cell 67, 1101-

d.

11n7

22. Morrison, D. K., Kaplan, D. R., Rhee, S. G., and Williams, L. T. (1990) Mol. Cell Biol. 10, 2359-2366 23. Kaplan, D. R., Morrisson, D. K., Wong, G., McCormick, F., and Williams, L. T. (1990) Cell 6 1 , 125-133 24. Lavoinne, A., Erikson, E., Maller, J. L., Price, D. J., Avruch, J., and Cohen, P. (1991) Eur.J. Biochem.1 9 9 , 723-728 25. Dent, P., Lavoinne, A,, Nakielny, S., Caudwell, F. B., Watt, P., and Cohen, P. (1990) Nature 348,302-308 26. Ahn, N. G., Seger, R., Bratlien, R. L., Diltz, C. D., Tonks, N. K., and Krebs, E. G. (1991) J . Biol. Chem. 266,4220-4227 27. Gomez, N., and Cohen, P. (1991) Nature 353,170-173