Nerve growth factor promotes the activation of phosphatidylinositol 3-kinase and its association with the trk tyrosine kinase.

We investigated the involvement of phosphatidylinositol 3-kinase (PtdIns 3-kinase) in the initiation of signal transduction by nerve growth factor (NGF) in the rat pheochromocytoma PC12 cell line. PtdIns 3-kinase catalyzes the formation of phosphoinositides with phosphate in the D-3 position of the inositol ring and previously has been found to associate with other activated protein tyrosine kinases, including growth factor receptor tyrosine kinases. Anti-phosphotyrosine immunoprecipitates had PtdIns 3-kinase activity that reached a maximum (9 times the basal activity) after a 5-min exposure of PC12 cells to NGF (100 ng/ml). Since NGF activates the tyrosine kinase activity of gp140trk, the protein product of the trk proto-oncogene, we also examined the association of PtdIns 3-kinase with gp140trk. Anti-gp140trk immunoprecipitates from NGF-stimulated PC12 cells had increased PtdIns 3-kinase activity compared to that of unstimulated cells, and larger increases were detected in cells overexpressing gp140trk, indicating that PtdIns 3-kinase associates with gp140trk. NGF produced large increases in [32P]phosphatidylinositol 3,4-bisphosphate and [32P]phosphatidylinositol 3,4,5-trisphosphate in PC12 cells labeled with [32P]orthophosphate, indicating an increase in PtdIns 3-kinase activity in intact cells. Using an anti-85-kDa PtdIns 3-kinase subunit antibody, we found that NGF promoted the tyrosine phosphorylation of an 85-kDa protein and two proteins close to 110 kDa. These studies demonstrate that NGF activates PtdIns 3-kinase and promotes its association with gp140trk and also show that NGF promotes the tyrosine phosphorylation of the 85-kDa subunit of PtdIns 3-kinase. Thus, PtdIns 3-kinase activation appears to be involved in differentiation as well as mitogenic responses.


Nerve Growth Factor Promotes the Activation of Phosphatidylinositol 3-Kinase and Its Association with the trk Tyrosine Kinase*
Stephen P. SoltoffSO, Stuart L. Rabinq, Lewis C. Cantley$, and David R. Kaplann From the $Department of Physiology,Tufts University,Boston,Massachusetts 021 11,YEukaryotic Signal Transduction Group,Molecular Mechanisms of Carcinogenesis Laboratory,Frederick,Maryland 21 702 We investigated the involvement of phosphatidylinositol 3-kinase (PtdIns 3-kinase) in the initiation of signal transduction by nerve growth factor (NGF) in the rat pheochromocytoma PC12 cell line. PtdIns 3kinase catalyzes the formation of phosphoinositides with phosphate in the D-3 position of the inositol ring and previously has been found to associate with other activated protein tyrosine kinases, including growth factor receptor tyrosine kinases. Anti-phosphotyrosine immunoprecipitates had PtdIns 3-kinase activity that reached a maximum (9 times the basal activity) after a 6-min exposure of PC12 cells to NGF (100 ng/ml). Since NGF activates the tyrosine kinase activity of gp140trk, the protein product of the trk proto-oncogene, we also examined the association of PtdIns 3kinase with gp140trk. Anti-gp140trk immunoprecipitates from NGF-stimulated PC12 cells had increased PtdIns 3-kinase activity compared to that of unstimulated cells, and larger increases were detected in cells overexpressing gp140trk, indicating that PtdIns 3-kinase associates with gp140trk.
NGF produced large increases in [32P]phosphatidylinositol 3,4-bisphosphate and [32P]phosphatidylinositol 3,4,5-trisphosphate in PC12 cells labeled with [32P]orthophosphate, indicating an increase in PtdIns 3-kinase activity in intact cells. Using an anti-85-kDa PtdIns 3-kinase subunit antibody, we found that NGF promoted the tyrosine phosphorylation of an 85-kDa protein and two proteins close to 110 kDa. These studies demonstrate that NGF activates PtdIns 3-kinase and promotes its association with gp140trk and also show that NGF promotes the tyrosine phosphorylation of the 85-kDa subunit of PtdIns 3-kinase. Thus, PtdIns 3-kinase activation appears to be involved in differentiation as well as mitogenic responses.
factor to be purified and characterized. NGF is required for the survival and differentiation of sympathetic and some sensory neurons in the peripheral nervous system, and of certain populations of cholinergic neurons in the central nervous system (reviewed in Levi-Montalcini, 1987;and Barde, 1989). NGF also stimulates the differentiation of the pheochromocytoma PC12 cell line into cells that resemble sympathetic neurons (Greene and Tishler, 1976;reviewed in Halegoua et al., 1991). The differentiation of PC12 cells is initiated by the binding of NGF to high affinity cell surface receptors (Sutter et al., 1979;Landreth and Shooter, 1980;Schechter and Bothwell, 1981). In these cells, the high affinity receptor may consist of at least two components, gp140trk and p75NGFR (Johnson et al., 1986;Radaeke et al., 1987;Large et al., 1989; al., 1991;Kaplan et al., 1991aKaplan et al., , 1991bKlein et al., 1991). While no catalytic activity has been assigned to p75NGFR, gp140trk contains an intrinsic tyrosine kinase activity which is activated by NGF binding (Kaplan et al., 1991a;Klein et al., 1991). The ligand-activated gp140trk tyrosine kinase is thought to initiate a signalling cascade involving the tyrosine phosphorylation of several cellular proteins, such as PLC-yl (Maher, 1988;Kaplan et al., 1991a;Kim et al., 1991;Vetter et al., 1991). NGF induces several other early events in PC12 cells, including activations of serine/threonine kinases (MAP kinase (erk), S6 kinase, and protein kinase C) (Mutoh et al., 1988;Tsao and Green, 1990;Gotoh et al., 1990;Boulton et al., 1991;Blenis and Erikson, 1986;Hama et al., 1986;Heasley and Johnson, 1989), elevation of intracellular calcium (Pandiella-Alonso et al., 1986), increases in phosphoinositol turnover (Contreras and Guroff, 1987), and the stimulation of the transcription of certain early response genes, including c-fos (Greenberg et al., 1985).
Within hours of NGF treatment, the transcription of genes encoding proteins important to the neuronal phenotype is stimulated (reviewed in Halegoua et al., 1991). Full differentiation of PC12 cells, characterized by the outgrowth of neurites, typically occurs after 3 to 4 days of NGF treatment (Green and Tischler, 1976).
gp140trk is required for the NGF-induced differentiation of PC12 cells. PC12 mutant cell lines which lack gp140trk fail to differentiate when exposed to NGF, however, introduction of the trk gene into these cells restores the ability of NGF to induce neurite outgrowth (Loeb et al., 1991). Despite the importance of gp140trk to the NGF differentiative responses, the signal transduction pathways utilized by gpl40'" to stimulate PC12 cell differentiation have not been extensively characterized. Co-immunoprecipitation studies indicate that gp140trk interacts with at least one cellular protein involved in the generation of second messenger molecules, PLC-y1 (Vetter et al., 1991). PLC-yl also interacts with oncogenically activated trk (p70) expressed in NIH-3T3 cells (Ohmichi et al., 1991a) and with gp140bk in PC12 cells (Ohmichi et a&, 1991b). The ligand-induced association of receptor tyrosine kinase activity with PLC-yl has been observed with several other growth factor receptors, including the PDGF, EGF, and c-kit tyrosine kinases (Kumjian et ai., 1989;Margolis et al., 1989;Morrison et al., 1990;Rottapel et at., 1991;Herbst et at., 1991). In addition, p2lC-"-specific GTPase activating protein (ras-GAP), pp60'"", pp62e-yes, pp59c-/y" tyrosine kinase, and pp74'"0f serine kinase have been found to associate with various growth factor receptors following ligand activation (reviewed in Cantley et al., 1991).
PtdIns 3-kinase is of particular interest in protein-tyrosine kinase signalling cascades since it was the first cytosolic enzyme to be found associated with activated tyrosine kinases and since it is almost ubiquitously associated with activated protein tyrosine kinases . This enzyme, which consists of 85-and 110-kDa subunits (Carpenter et al., 1990), is not in the canonical PtdIns turnover pathway, but catalyzes the formation of a family of phosphoinositides with phosphate at the D-3 position of the inositol ring (Carpenter and Cantley, 1990). While the role of these phosphoinositides in signal transduction processes is not known, the activation of PtdIns 3-kinase activity and the association of this activity with tyrosine kinases correlates well with the ability of several tyrosine kinases to induce mitogenesis or cellular transformation (reviewed in . Previous studies have focused upon the activation of PtdIns 3-kinase activity during cell proliferation or the activation of neutrophils and platelets (reviewed in Auger and Cantley, 1991). In the present study, we investigated the activation of PtdIns 3-kinase by NGF and the association of PtdIns 3kinase with gp140trk, a tyrosine kinase that is associated with the initiation of cellular differentiation. NGF produced an increase in PtdIns-3,4-Pz and PtdIns-3,4,5-P3, the products of PtdIns 3-kinase, and promoted the physical association of this kinase with gp140trk. NGF also stimulates the tyrosine phosphorylation of multiple cellular proteins, including the 85-kDa protein subunit of PtdIns 3-kinase. Thus, PtdIns 3kinase appears to be directly activated by this NGF receptor.

Chemic&
Ail chemicals were reagent grade or better. [3'P]ATP (specific activity, 3000 Ci/mmol) and [32P]orthophosphate (8500 Ci/mmol) were purchased from Du Pont-New England Nuclear (Boston, MA). Dulbecco's modified Eagle's medium (DMEM) was obtained from GIBCO Laboratories. NGF (2.5s) was obtained from Boehringer-Mannheim or Genentech, Inc. Insulin was purchased from Sigma, and PDGF (BB) was obtained from Upstate Biochemicals Inc. Trk antisera (203) was generated in rabbits to the 14 carboxyl-terminal amino acids of human gp140ek as previously described ( M~i n -Z~c a et Kaplan et al., 1991a). This trk antibody recognizes trk molecules from mouse, rat, and human (Kaplan et al., 1991a(Kaplan et al., , 1991b. The anti-phosphotyrosine antibody, or anti-P-Tyr, was a murine monoclonal antibody generated using phosphotyramine as an immunogen and was kindly supplied by Dr. Brian Drucker (Dana Farber Cancer Institute, Boston, MA) and Dr. Deborah Morrison (National Cancer Institute-Frederick Cancer Research and Development Center). Antibody (anti-p85) to the 85-kDa subunit of PtdIns 3-kinase was raised in rabbits by Dr. Brian Schaffhausen (Tufts University) against recombinant p85 that was cloned from rat liver by Dr. Christopher Carpenter (Tufts University and Massachusetts General Hospital), and is commercially available from Upstate Biochemicals Inc.

Cell Culture
PC12 cells (from Dr. Larry Feig (Tufts University) or Dr. R. Kelly (University of California, San Francisco)) were grown in 100-or 150mm dishes in DMEM plus 5% horse serum and 5% calf serum at 37 "C in a 95/5% air/C02 mixture. Where noted, in some experiments cells were switched to serum-free medium (DMEM plus 0.1% BSA) overnight prior to stimulation with NGF. PC12 cells overexpressing gp140frk (line 6-15) were obtained by electroporation of cells with a plasmid containing the human trk and neomycin resistance genes under control of the CMV promoter. The generation of this cell line was as described.'

Assay of PtdIns 3-Kinase Activity
NGF was added to intact cells at 37 "C. To assess the association of PtdIns 3-kinase with anti-P-Tyr, cells were treated with NGF (100 ng/ml) for the indicated time, washed twice with ice-cold buffer A (137 mM NaCl, 20 mM Tris, 1 mM MgC12, 1 mM CaC12, 0.1 mM vanadate, pH 7.5), and lysed in lysis buffer (buffer A plus 10% glycerol (v/v), 1% Nonidet P-40 (v/v), and 1 mM phenylmethyisulfonyl fluoride). The lysates were vortexed and centrifuged at 14,000 RPM (Eppendorf 5414 centrifuge). The cleared supernatant was transferred to a fresh microcentrifuge tube, and incubated with anti-P-Tyr antibody (35 pg/ml lysate) for 2 h at 4 "C. To assess the association of PtdIns 3-kinase with gp140trk, PC12 cells (2 X 1 0 ' ) were treated for 5 min with 100 ng/ml NGF at 37 "C. Cells were washed twice in icecold Tris-buffer saline (TBS) (20 mM Tris, pH 8.0,0.15 M NaC1) and lysed in buffer A without MgC12 and CaClz as described . Lysates were clarified by centrifugation as described above and incubated with 4 pl of anti-trk antibody (203) for 3 h at 4 "C.
Protein A-Sepharose (4 mg/ml lysate) was then added for 2 h at 4 "C to the lysates containing the anti-trk antibody or anti-P-Tyr antibody.
The immunoprecipitates were washed three times in PBS/I% Nonidet P-40, two times in 0.1 M TrislO.5 M LiCI, and two times in TNE (10 mM Tris, 100 mM NaCl, 1 mM EDTA, pH 7.5). All wash solutions contained 100 p~ vanadate. The PtdIns kinase assay of the immunoprecipitate was performed by adding sonicated Ptdlns (in 10 mM HEPES, 1 mM EDTA, pH 7.5; 0.5 mg/ml final concentration) and [T-~'P]ATP (20-40 pCi/sample) to the immunoprecipitates for 10 min at room temperature. The reaction was stopped by the addition of 80 p1 of HCl(1 M) and 160 pl of methano1:chloroform (1:l mixture). The lipid-containing organic phase was resolved on oxalate-coated thin-layer chromatography plates (Silica Gel 60, MCB reagents, Merck, Rahway, NJ) developed in ch1oroform:methano~ water:ammonium hydroxide (604011.3:2). Radiolabeled spots corresponding to authentic PtdIns-4-P standards were excised and quantified by scintillation counting or Cerenkov radiation. In some experiments the PtdInsP spot was deacylated and subjected to HPLC analysis to determine lipid identity and quantification, as previously described .

Measurement of Polyphosphoinositides in Intact Cells
free DMEM (0.1% BSA). The cells were washed once and incubated Cells that were nearly confluent were cultured overnight in serumwith phosphate-free DMEM/O.l% BSA for 15 min then exposed to 5 ml of the same medium plus carrier-free [32P]orthophosphate (100 pCi/ml) at 37 'C for 4 h. NGF (100 ng/ml) was added to the cells for the designated time. The cells were washed twice with ice-cold buffer A and lysed in 1 M HC1:methanol (l:l, v/v), and the lipids were extracted with chloroform. The lipids were deacylated and analyzed by HPLC as described previously .

Immunoprec~pita~ion and Western Blot Assays
Anti-p85 Immunoprecipitations-A total of 2 X lo7 PC12, trk-PC12, or NIH-3T3 cells transfected with a rat trk gene (Kaplan et at., 1991b) were lysed in Lysis buffer, and cleared lysates were immunoprecipitated with 2 pl of anti-p85 for 3 h at 4 "C. In some experiments, cells were lysed in RIPA buffer (NaC1,150 mM; Nonidet P-40, 1%; deoxycholate, 0.5%; Tris, 50 mM; pH 8.0) containing 0.2% SDS. The immunoprecipitates were collected with 20 g1 of protein A-Sepharose, washed three times with buffer A and once with distilled water, and electrophoresed on 8.5% SDS-polyacrylamide gels, Proteins were transferred to 0.2-pm nitrocellulose filters, the filters were blocked using TBS/2% BSA for 1 h at room temperature, and the filters were incubated with anti-P-Tyr for 16 h at 4 "C.

NGF Stimulates PI 3-Kinase and Its Association with gp140trk
Amersham) with secondary antibodies from Boehringer-Mannheim.
Anti-P-Tyr Immunoprecipitations of p85-A total of -10' PC12 cells were lysed in Lysis buffer, and cleared lysates were immunoprecipitated with anti-P-Tyr (35 pglml) for 2 h at 4 "C. The immunoprecipitates were collected with protein A-Sepharose, washed three times in PBS/l% Nonidet P-40, two times in 0.1 M Tris/Od M LiCl, and two times in TNE. All wash solutions contained 100 p~ vanadate. The samples were boiled for 5 min, and the supernatants were electrophoresed on 7% SDS-polyacrylamide gels. Proteins were transferred to 0.2-pm nitrocellulose filters. The filter was blocked for 1 h in TBS/5% nonfat dry milk, washed two times with TTBS, and exposed to anti-p85 antibody (1:3000 dilution in TTBS/5% nonfat dry milk) for 1 h at room temperature. Filters were washed three times with TTBS. Filters were exposed to secondary anti-rabbit-AP antibodies (3 pl per 20 ml of TTBS/5% nonfat dry milk) from Promega and washed three times using TTBS, and proteins were identified by the alkaline phosphatase technique.

RESULTS
Anti-phosphotyrosine Immunoprecipitates of NGF-stimulated Cells Have PtdIns 3-Kinase Activity-Preliminary experiments demonstrated that exposure of PC12 cells to NGF produced an increase in anti-P-Tyr-immunoprecipitable PtdIns 3-kinase activity. To examine this more closely, we studied the time course of the effects of NGF on PC12 cells. As shown in Fig. 1, NGF (100 ng/ml) produced a timedependent increase in PtdIns 3-kinase activity. The maximum stimulation occurred after 5 min of NGF exposure. A t this time PtdIns 3-kinase activity was elevated to 9.2 f 1.2 (n = 3) times the basal level of activity. After longer exposures of

Exposure of PC12 cells to NGF increases anti-P-
Tyr-immunoprecipitable PtdIns 3-kinase activity. PC12 cells were serum-starved overnight and then were exposed to 100 ng/ml NGF for 0-60 min. Lysates were prepared in 1% Nonidet-P4O and immunoprecipitated for 2 h at 4 "C with anti-P-Tyr. PtdIns 3-kinase activity was measured in the anti-P-Tyr immunoprecipitate using exogenous PtdIns as a substrate. See "Materials and Methods" for further details. The increase in PtdIns 3-kinase activity at different times is shown as the percentage of the maximal level. Bars indicate standard error of the mean ( n = 3). Basal levels measured 964 cpm ( n = 2) by scintillation counting. Inset, HPLC analysis of the deacylation product of [3'P]PtdInsP generated from an anti-P-Tyr immunoprecipitate of cells stimulated by NGF (100 ng/ml) for 5 min.
[:"P]PtdInsP was separated by thin-layer chromatography and was deacylated with methylamine as described . The positions of the main product, glycero-PtdIns-3-P (gPZ-3-P), and glycero-PtdIns-4-P (gPI-4-P) which was run simultaneously as a standard, are indicated by the arrows. The vast majority of ["PI corresponded to gPI-3-P. the cells to NGF, the activity declined to a lower level that remained stimulated above the basal level for at least 60 min of NGF exposure. At this time, PtdIns 3-kinase activity was 4.1 +. 0.9 (n = 3) times the basal activity. HPLC analysis of the deacylated product of the PtdInsP product confirmed that it is primarily (>go%) PtdIns-3-P (Fig. 1, inset). Thus, PtdIns 3-kinase was either phosphorylated on tyrosine or associated with a tyrosine phosphorylated protein in NGF-stimulated cells.
Association of PtdIns 3-Kinase with gp14Pk--We next investigated whether gp140trk associated with PtdIns 3-kinase activity in wild-type PC12 cells and in PC12 cells that overexpressed gp140trk (trk-PC12 cells). Trk-PC12 cells express 20-fold more gp140trk than do wild-type cells, and show 10fold higher levels of NGF-activated gp140trk tyrosine kinase activity.2 The association of PtdIns 3-kinase activity with gpl40trk was measured in anti-trk immunoprecipitates of lysates prepared from untreated cells or cells treated for 5 min with NGF ( Fig. 2A). In wild-type cells, NGF stimulated a slight increase (%fold) in anti-trk-immunoprecipitable PtdIns 3-kinase activity. In contrast, NGF stimulated a much larger (10-fold) increase in anti-trk-immunoprecipitable PtdIns 3-kinase activity in trk PC12 cells. There was only a small amount of PtdIns 3-kinase activity present in precipitates prepared with the preimmune serum (Fig. 2B). Incuba were treated for 5 min with NGF (100 ng/ml). Cell lysates were immunoprecipitated with 4 pl of anti-trk antibody 203 for 3 h at 4 "C as described under "Materials and Methods." PtdIns 3-kinase activity of anti-trk immunoprecipitates was assayed using exogenous PtdIns as a substrate. E , the immunoprecipitation of PtdIns 3-kinase activity by anti-gp14Otrk (anti-trk) is specific. Trk PC12 cells were treated for 5 min with NGF (100 ng/ml), and immunoprecipitates were prepared with anti-trk serum (4 pl/ml), pre-immune serum (4 pl/ml), or antitrk serum (4 pl/ml) plus the peptide (10 pg/ml) used to generate antitrk.
tion of anti-trk immunoprecipitates with a peptide (10 pg/ml) used to generate the trk antibody reduced significantly the association of PtdIns 3-kinase activity (Fig. 2B). In NGFtreated cells the amount of PtdIns 3-kinase activity observed in anti-trk immunoprecipitates represented less than 2% of the PtdIns 3-kinase activity in anti-P-Tyr immunoprecipitates when wild-type PC12 cells were assayed and less than 10% in assays of trk-PC12 cells (not shown). These results indicate that PtdIns 3-kinase associates with pg140trk. In conjunction with the observation that anti-P-Tyr antibody immunoprecipitates PtdIns 3-kinase activity (Fig. 1) and that NGF activates PtdIns 3-kinase activity in vivo (below), these results suggest that the activation of PtdIns 3-kinase by NGF occurs via stimulation of the gp140trk tyrosine kinase activity.

Analysis of Phsphoinositides in Intact PC12 Cells after NGF
Stimulation-To investigate alterations in the levels of the various phosphoinositide products by NGF, we measured the in vivo levels of the D-3-phosphorylated polyphosphoinositides. Cells were prelabeled with [32P]orthophosphate and exposed to NGF for various times. The lipids were extracted and deacylated, and the glycerophosphoinositol polyphosphates were analyzed by HPLC (as in Fig. 1, inset). Unlike quiescent human vascular smooth muscle cells, which did not contain detectable amounts of PtdIns-3,4-P2 or PtdIns-3,4,5-P3 unless stimulated by growth factors , PC12 cells contained measurable radioactivity incorporated into PtdIns-3,4,5-P3 and PtdIns-3,4-P2 prior to stimulation by NGF. Small changes in these products were observed 1 min after NGF (100 ng/ml) exposure (Fig. 3A). Five min after NGF exposure, 32P incorporation into PtdIns-3,4,5-P3 was increased to levels that were five times the basal levels. After 15 min, the level of [32P]PtdIns-3,4,5-P3 had declined to near basal levels and was maintained at this level after 30 min. The [32P]PtdIns-3,4-Pz level increased more than ["PI PtdIns-3,4,5-P3 and was maintained at a greater elevated level. After 5 min of NGF exposure, [32P]PtdIns-3,4-Pz increased to 13-fold the basal level, and further increased to a slightly greater level after 15 min. After 30 min, the level of [32P]PtdIns-3,4-Pz declined to a value that remained 8-fold that of the basal amount.
In contrast to the marked changes produced in [32P]PtdIns-3,4-P2, much smaller relative changes were produced in [32Ppl PtdIns-4,5-Pz following the addition of NGF (Fig. 3A). The largest change was observed at 30 min after NGF exposure, when the [32P]PtdIns-4,5-Pz level declined to about 60% of the basal amount. Relatively small changes were observed in the in vivo levels of [32P]PtdIns-3-P and [32P]PtdIns-4-P over the course of the NGF stimulation (Fig. 3B).

Tyrosine Phosphorylation of 85-kDa Protein in NGF-stimulated
CelLs-We next examined whether a component of PtdIns 3-kinase was phosphorylated on tyrosine in response to NGF treatment of cells. PtdIns 3-kinase activity co-purifies with proteins of 85 and 110 kDa (Carpenter et aL, 1990). PC12 cells, trk-PC12 cells, or NIH-3T3 cells transfected with a rat trk cDNA (trk-3T3) were treated for 5 min with NGF (100 ng/ml), and cell lysates were immunoprecipitated with anti-p85 antibody. Protein blots of the immunoprecipitates were then probed with anti-P-Tyr antibody. In PC12 and trk-PC12 cells, NGF induced the tyrosine phosphorylation of an 85-kDa protein and of two proteins of 100-120 kDa (Fig. 4A).
When the filter was stripped and reprobed with anti-p85, a protein of 85 kDa was recognized which migrated identically with the tyrosine phosphorylated p85 species (Fig. 4B). A band at 115 kDa that was recognized in blots by anti-p85 did not co-localize with the tyrosine-phosphorylated llO/ll5-kDa bands. Tyrosine phosphorylation of the 85-kDa protein was

FIG. 3. Effects of NGF on the levels of phosphoinositides of intact PC12 cells labeled in vivo using 82P-orthophosphate.
NGF (100 ng/ml) was added at t = 0. The lipids were extracted and deacylated as described under "Materials and Methods" and analyzed by HPLC. Levels are presented in terms of percentage of maximum levels. The average results from two experiments are presented. The data were normalized to the total cpm at t = 0. For PtdIns-3,4,5-P3, PtdIns-3,4-P2, and PtdIns-4,5-P2, the mean differences between the two experiments were 8.7 f 2.9%, 16.0 * 10.4%, and 24.0 f 7.4%, respectively (n = 5 time points for each lipid). For PtdIns-3-P and PtdIns-4-P these values were 40.2 f 7.6 (n = 10, 5 time points from each). A, the maximum levels for PtdIn~-3,4,5-~, PtdIns-3,4-P~, and PtdIns-4,5-Pz were 4152, 9260, and 175,277 dpm, respectively. For PtdIns-4,5-Pz, the maximum values were reached at 1 min in one experiment and at 5 min in the other. B, the maximum levels for PtdIns-4-P and PtdIns-3-P were 98,615 and 9102 dpm, respectively. For PtdIns-4-P, the maximum values were reached at 1 min in one experiment and at 30 min in the other experiment. In terms of the amounts of radioactivity incorporated into the different polyphosphoinositides in unstimulated cells, PtdIns-4,5-Pz made up the largest fraction. Relative to this compound, PtdIns-3,4,5-P3, PtdIns-3,4-Pz, PtdIns-3-P, and PtdIns-4-P made up 0.51,0.34,3.22, and 28.02%. also observed in NGF-treated PC12 cells or PDGF-treated NIH-3T3 cells that were lysed in a more stringent buffer (RIPA) that contained 0.2% SDS or in lysis buffer that contained 1% SDS. When anti-p85 immunoprecipitates of these lysates were transferred to nitrocellulose and probed with anti-P-Tyr, we detected a phosphorylated 85-kDa protein in growth factor-treated cells. An 85-kDa protein that co-localized with this protein was detected when the filter was stripped and reprobed with anti-p85 (not shown).
In trk-3T3 cells, NGF stimulated the tyrosine phosphorylation of an 85-kDa protein and of two proteins at approximately 110 kDa (Fig. 4A) that were immunoprecipitatedusing anti-p85. NGF also stimulated PtdIns 3-kinase activity by about %fold in anti-P-Tyr immunoprecipitates prepared from trk-3T3 cells (not shown). PDGF treatment of the trk-3T3 cells induced the tyrosine phosphorylation of 85-and 110-kDa proteins and a third protein of 185 kDa. The 185-kDa protein is most likely the PDGF receptor (Escobedo et al., 1991), and the 110-kDa protein may be a second component of PtdIns 3-kinase (Carpenter et al., 1990;Escobedo et al., 1991;Otsu et al., 1991).
Immunoprecipitations and Western blot assays were also  p85. A and B, cells were treated for 5 min at 37 "C with NGF (100 ng/ml), insulin (100 ng/ml), or PDGF (2 nM). Lysates from PC12 cells, trk-PC12 cells, or NIH-3T3 cells expressing gp140trk were prepared in 1% Nonidet-P40, cleared, and proteins were immunoprecipitated with anti-p85 (2 pl/ml). Proteins were electrophoresed on 8.5% gels, transferred to nitrocellulose filters, and probed with anti-P-Tyr ( A ) . The filters were stripped and reprobed with anti-p85 as described under "Materials and Methods" ( B ) . Molecular mass markers in kilodaltons are indicated on the left. The positions of the 85-and 110-kDa proteins are indicated on the right. C, PC12 cells and NIH-3T3 cells were treated for 5 min at 37 "C with NGF (100 ng/ml) and PDGF (10 ng/ml), respectively. Lysates were prepared in 1% Nonidet-P40 and cleared, and proteins were immunoprecipitated using anti-P-Tyr. After electrophoresis (7% gel) and transfer to nitrocellulose filters, proteins were probed with anti-p85 (1:3000 dilution). The position of p85 is indicated on the right.

FIG. 4. NGF induces the tyrosine phosphorylation of
performed using these antibodies in the reverse order from that described above. In these experiments, 85-kDa proteins were detected using anti-p85 antibody to blot the anti-P-Tyr immunoprecipitates from cell lysates of NGF-treated PC12 cells and PDGF-treated NIH-3T3 cells (Fig. 4C). p85 may appear as a doublet due to the lower percentage of acrylamide (7%) used in the resolving gel in this experiment. Taken together, these experiments indicate that the 85-kDa subunit of PtdIns 3-kinase is phosphorylated on tyrosine in growth factor-stimulated cells.

DISCUSSION
The results presented in this paper demonstrate that 1) PtdIns 3-kinase is activated by NGF in PC12 cells, 2) PtdIns 3-kinase associates with the trk proto-oncogene product, gp140trk, and 3) an 85-kDa protein is phosphorylated on tyrosine in NGF-(and PDGF-) treated cells. Immunoprecipitation and Western blot assays using an antibody raised against the p85 subunit of PtdIns 3-kinase demonstrate that this protein is the 85-kDa PtdIns 3-kinase subunit.
While enhancements in inositol phosphate levels have been previously demonstrated in NGF-treated PC12 cells, the types of phosphoinositides generated were not characterized. We show that NGF treatment of cells induces an increase in the levels of PtdIns-3,4-P2 and PtdIns-3,4,5-P3, the expected products of PtdIns 3-kinase. Increases in the levels of PtdIns-4,5-P2, a substrate for PLC-yl activities, were not observed in these studies. The association of PtdIns 3-kinase with gp140trk and the increased cellular levels of PtdIns-3,4-P2 and PtdIns-3,4,5-P3 occurred within 5 min of NGF addition to cells, suggesting that PtdIns 3-kinase activation is a primary response to gp140trk activation.
Our results demonstrate an activation of PtdIns 3-kinase activity in cells responding to NGF, an anti-mitogenic and differentiative factor. Increases in PtdIns 3-kinase activity previously have been observed only in growth factor-treated and -transformed cells and in neutrophils and platelets exposed to activating agents (reviewed in Auger and Cantley, 1991;. It is unlikely that activation of PtdIns 3-kinase by NGF alone is sufficient to induce differentiation of PC12 cells. Both EGF and insulin, which are mitogenic agents for PC12 cells, stimulate the appearance of PtdIns 3-kinase activity in anti-P-Tyr immunoprecipitates from these cells.3 EGF also stimulates the production of PtdIns-3,4-P2 and PtdIns-3,4,5-P3 in PC12 cells labeled with [32P]P04.3 Neither EGF nor insulin promotes the morphological differentiation of PC12 cells. Therefore, although the activation of PtdIns 3-kinase is one of the first events involved in NGF-mediated signal transduction in PC12 cells, its role in differentiation remains unclear. The distinction between the involvement of PtdIns 3-kinase in differentiation from that of growth processes could conceivably involve the relative amounts of or temporal response of PtdIns-3,4-P2 and PtdIns-3,4,5-P3 to the activating agent. In fact, the kinetics and relative levels of the in uiuo production of PtdIns-3,4-P2 and PtdIns-3,4,5-P3 in NGF-treated cells are different from those measured in EGF-treated cells.3 These differences may denote a distinction between a growth response and a differentiative response and are the focus of current studies. Alternatively, the promotion of cell growth and transformation (and differentiation) may require multiple cellular responses of which the activation of PtdIns 3-kinase may be but one. This may involve the recruitment of multiple cytosolic enzymes to the plasma membrane and the generation of multiple signals in the cell (reviewed in Cantley et al., 1991).
The association of PtdIns 3-kinase with receptor proteintyrosine kinases has been a hallmark observation of this family of tyrosine kinases (Cantley et al., 1991). PtdIns 3kinase is a heterodimer consisting of 85-and 110-kDa proteins (Carpenter et al., 1990). The 85-kDa subunit has two src homology-2 (SH-2) domains, a src homology-3 (SH-3) domain, and a bcr homology domain (Otsu et al., 1991). SH-2 domains have recently been shown to confer special association with tyrosine-phosphorylated proteins (Koch et al., 1991;Cantley et al., 1991). The tyrosine phosphorylation domains of some receptor and nonreceptor proteins that bind PtdIns 3-kinase are highly conserved and have a phospho-Tyr-Met/ Val-Asp/Pro-Met consensus sequence . Phosphopeptides based on this sequence can block the in vitro association between PtdIns 3-kinase and the PDGF receptor (Escobedo et al., 1991) or polyoma middle T/pp60c'" (Auger et al., 1992). Certain ligand-activated tyrosine kinases such as the insulin receptor do not contain a good consensus PtdIns 3-kinase association sequence, yet bind PtdIns 3-kinase (Ruderman et al., 1990). The amount of PtdIns 3-kinase binding to insulin receptors represents only a small proportion of anti-P-Tyr immunoprecipitable PtdIns 3-kinase activity (Ruderman et al., 1990), suggesting that the association between the receptor and PtdIns 3-kinase is of a low affinity.
The characteristics of the association of PtdIns 3-kinase S. P. Soltoff, unpublished results.
with gp140trk are similar to that of the insulin receptor. The amount of PtdIns 3-kinase activity observed in anti-trk immunoprecipitates represent only a small fraction (~2 % ) of the PtdIns 3-kinase activity in anti-P-Tyr immunoprecipitates, indicating that as in the case of the insulin receptor, the association of PtdIns 3-kinase activity with gp140trk in detergent-containing buffer is unstable. Consistent with this result, we did not observe the presence of s i~i f i c a n t quantities of gp140trk in anti-p85 immunoprecipitates prepared from NGF-treated cells.4 The association of these proteins may be below the level of sensitivity of the anti-gp140trk and anti-p85 antibodies. Another possibility is that gp140trk may interact with other subunits of PtdIns 3-kinase. However, we do observe the association of baculovirus-expressed trk with SH2 domains of p85/PtdIns 3-kinase in uit~0,5 indicating that trk is capable of forming complexes with p85. Alternatively, PtdIns 3-kinase may become activated in NGF-treated PC12 cells by interacting with other receptor binding sites or other proteins, such as the recently identified insulin receptor substrate IRS-1 (Sun et al., 1991). In insulin-treated PC12 cells, IRS-1 directly binds PtdIns 3-kina~e.~ However, IRS-1 does not co-immunoprecipitate with PtdIns &kinase from NGFtreated PC12 cells," and IRS-1 is not a substrate of gp140trk,6 indicating that the activation of PtdIns 3-kinase by ligandactivated gp140trk receptors involves mechanisms distinct to that utilized by insulin receptors.