Purification and mechanism of activation of a nerve growth factor-sensitive S6 kinase from PC12 cells.

A nerve growth factor (NGF)-sensitive S6 kinase was purified by alkaline lysis of PC12 cells. The activity in lysates from NGF-treated cells was 10-20-fold higher than that from controls. Half-maximal stimulation of the S6 kinase by NGF treatment occurred in approximately 5 min, and the activity returned almost to basal levels by 2 h. A rapid purification method was devised in which crude extract was applied directly to a PBE 94 column after buffer exchange on a PD-10 column (Sephadex G-25 M). The activated S6 kinase was purified at least 673-fold with a recovery of approximately 70%. The S6 kinase has an apparent molecular weight of 45,000 and is highly specific for S6. It is not inhibited by the specific inhibitor of cAMP-dependent protein kinases, or by chlorpromazine or sodium vanadate, nor is it activated by Ca2+/calmodulin. It was inhibited by EGTA, beta-glycerophosphate, or NaF. Phosphorylation occurred solely on serine residues. The S6 kinase activity from control cells and from NGF-treated cells eluted at pH 5.69 and 5.58, respectively, during PBE 94 column chromatography. Pretreatment of crude extract from NGF-stimulated cells with alkaline phosphatase resulted in an elution of the enzyme at the position of S6 kinase from control cells and a concomitant decrease in activity. These results indicate that phosphorylation is involved in the mechanism of S6 kinase activation.

protease-activated kinase I1 (13). An epidermal growth factor (EGF)-stimulated S6 kinase has been observed in extracts of Swiss 3T3 cells (14), a transformation-sensitive S6 kinase (15) and a new S6-recognizing kinase (16) have been seen in chick embryo fibroblasts, and an S6-specific kinase has been found in extracts of Xenopus eggs (17). Indeed, S6 has been shown to be a substrate for a number of kinases in uitro; these include CAMP-dependent protein kinase (18-20), cGMP-dependent protein kinase (19,20), protease-activated kinases (20-23), protein kinase C (24,25), and calmodulin-dependent protein kinases (26,27). But there is as yet no information as to which kinase is involved in the nerve growth factor-stimulated phosphorylation of S6.
The mechanism by which S6 kinase is activated is not known. Novak-Hofer and Thomas (14,28) have studied the EGF-mediated activation of an S6 kinase in Swiss 3T3 cells and suggested that phosphorylation is involved in S6 kinase activation. Tabarini et al. (12) have suggested that the insulinstimulated S6 kinase from 3T3-Ll cells is activated by phosphorylation by tyrosine protein kinase or protein kinase C. Blenis and Erikson (29) studied a transformation-sensitive S6 kinase seemingly regulated by the action of tyrosine kinase or protein kinase C . Erikson and Maller (30) also suggested that S6 kinase is regulated by protein tyrosine kinases. These reports are consistent in that they indicate that phosphorylation of S6 kinase is involved in its activation.
In a previous paper (31) we described a soluble S6 kinase from PC12 cells, the action of which is stimulated by prior treatment of the cells with nerve growth factor. In this study, we have purified this S6 kinase, explored its properties, and tried to understand the mechanism by which it is activated.

Culture Conditions
PC12 cells were cultured as monolayers in 150-cm2 culture flasks in Dulbecco's modified Eagle's medium supplemented with 7% fetal bovine serum, 7% horse serum, and 100 pg of streptomycin and 100 units of penicillin/ml. They were kept at 37 "C in an atmosphere containing 6% COz. The cells were split in a 1:4 or 1% ratio each week, and the medium was changed once during the week. The cells were treated in the culture flasks with one or more of the following factors, usually for 60 min: NGF (50 ng/ml), 5'-N-ethylcarboxamideadenosine (NECA) (1 X M), dibutyryl CAMP (5 X M).
Control cultures were kept under similar conditions. After treatment, the cells were collected and washed by centrifugation (1100 X g, 5 min) as described previously (10).

Preparation of Cell-free Extracts
A lysis buffer (20 mM sodium borate, pH 10.2, containing 0.2 mM EDTA, 0.1 mM NazMoO,, and 1 mM phenylmethylsulfonyl fluoride) was added to the washed cell pellets and the mixture agitated on a Vortex mixer for 5 s to lyse the cells. The volume of lysis buffer used was 0.5 ml for cells from one flask. Since phenylmethylsulfonyl fluoride is labile in aqueous solution, it was added to the lysis buffer from a stock solution made in n-propyl alcohol (1/100 dilution) just before the lysis buffer was added to the cells. The lysate was immediately neutralized with 1 M MOPS, pH 3.0, containing 100 mM MgC1, and 10 mM dithiothreitol to give pH 7.3 (10% (v/v) volume was used for neutralization). The preparations were centrifuged at 2 "C for 2.5 h at 125,000 X g. The supernatants typically contained about 0.5 mg of protein/ml. Protein concentration was determined by the dye-binding method of Bradford (32) using Bio-Rad reagents.

Preparation of 40 S Ribosomal Subunits from Rat Liuer
Ribosomal 40 S subunits were prepared from livers of adult Sprague-Dawley rats according to the method of Thomas et al. (33). They were frozen and stored at -70 "C in buffer C (33) containing 50% glycerol.

Assay for S6 Kinase Activity
Rat liver 40 S ribosomal subunits were used as substrate for measuring S6 kinase activity. The incubation was carried out in a total volume of 100 pl using a reaction mixture containing 50 mM MOPS, pH 7.2, 10 mM MgCl,, 1 mM dithiothreitol, 60 pM ATP, 2 pCi of [ Y -~~P I A T P , 0.34 unit of 40 S ribosomal subunits, and an S6 kinase preparation. The reaction was started by the addition of 40 S ribosomal subunits or a mixture of ATP and [ Y -~~P ] A T P and the incubation continued for 30 min at 30 "C. The reaction was stopped by the addition of 100 p1 of a sodium dodecyl sulfate (SDS) sample buffer containing 125 mM Tris-HC1 buffer, pH 6.7, 2% SDS, 20% glycerol, and 5% (3-mercaptoethanol. The samples were boiled for 30 min and analyzed by SDS-gel electrophoresis on 10% polyacrylamide gels according to the method of Laemmli (34). The gels were stained with Coomassie Brilliant Blue, and autoradiograms were prepared by exposing the dried gels to Kodak XAR film at -70 "C using an intensifying screen (Du Pont Cronex Lightning Plus XA).
The protein band corresponding to S6 was excised from the dried gels, and the 32P incorporated was quantitated by liquid scintillation spectroscopy. Autoradiograms were scanned with a Zeineh soft laser densitometer (LKB Instruments, Inc.).

Purification of 5' 6 Kinase from PC12 Cells
All procedures were carried out at 0-5 "C; the samples were never frozen. At each step of purification, the activity was assayed by the phosphorylation of 40 S ribosomal proteins from rat liver.
Method I-Cell-free extracts were prepared from cells from 50 flasks (13.8 mg of protein) as described above in the absence or presence, during the preparation, of 80 mM &glycerophosphate and 20 mM EGTA. The extracts from control and from NGF-treated cells were filtered through Millex-GS (0.22 pm, Millipore) and applied directly to heparin-Sepharose CL-GB columns (12.5 X 0.9 cm) equilibrated with 50 mM Tris-HCl buffer, pH 8.0. After application of the sample, the column was washed with 1 bed volume of the same buffer at a flow rate of 0.52 ml/min. The column was eluted at the same flow rate with a linear gradient made up of 70 ml each of 0 and 0.75 M NaCl in the same buffer. Fractions of 2.1 ml were collected and assayed for S6 kinase activity. Fractions 3-20 were pooled and concentrated in an Amicon ultrafiltration cell equipped with a PM-10 membrane. The samples were then subjected to gel filtration on a Sephadex G-200 column (90 X 1.5 cm) equilibrated with 25 mM imidazole HCl buffer, pH 7.0. The column was eluted at a flow rate of 0.17 ml/min with the same buffer. Fractions of 1.85 ml were collected and assayed for S6 kinase activity. The active fractions (50-70) were pooled and applied directly to a PBE 94 column (12.5 X 0.9 cm) equilibrated with 25 mM imidazole HC1 buffer, pH 7.0. The column was eluted at a flow rate of 0.33 ml/min with elution buffer (0.0075 mmol/pH unit/ml of Polybuffer 74, pH 4.0). Fractions of 1.0 ml were collected and assayed immediately for S6 kinase activity.
Method 2-Cell-free extracts were prepared from 4 flasks (1.12 mg of protein) as described above from control or ligand-stimulated cells, filtered with Millex-GS (0.22 pm, Millipore), and applied to a PD-10 column (Pharmacia) equilibrated with 25 mM imidazole HCI buffer, pH 7.0. The column was eluted with the same buffer and fractions of 20 drops (0.7 ml) each collected. After measuring the AZw of each fraction, the peak fractions (4-6) were pooled and applied to a PBE 94 column as described under "Method 1." All procedures were done as quickly as possible.

S ribosomal subunits from rat liver were incubated for 2 h with
[T-~*P]ATP and the purified S6 kinase in the assay mixture described above. The reaction was terminated by the addition of 950 pg of unlabeled 40 S ribosomal subunits, 0.1 volume of 1 M MgCl,, and 2 volumes of glacial acetic acid. The ribosomal proteins were extracted, precipitated with acetone cooled to -20 "C, and subjected to twodimensional electrophoresis as described by Thomas et al. (35) except that the first-dimension electrophoresis was carried out with a slab gel instead of a glass tube for 30 min at 70 V and then for 17 h at 180 V. The gel was cut with a thin knife blade to give a strip (1.5 mm X 12 cm x 6 mm) for the second-dimension gel electrophoresis. Gels were stained with Coomassie Blue, dried, and autoradiographed as described above.

Phosphoamino Acid Analysis
The analysis of amino acids phosphorylated in S6 by S6 kinase was performed by thin-layer cellulose electrophoresis as previously described (4) except that the electrophoresis buffer had a pH of 3.5 (acetic acid/pyridine/water, 50:5:945) (36) and the electrophoresis itself was carried out at 1000 V for 45 min.

Treatment with Alkaline Phosphatase
Alkaline phosphatase treatment was carried out for 30 min at 30 'C. One-half unit (1 unit is the activity that hydrolyzes 1 pmol of 4-nitrophenylphosphate in 1 min at 37 "C in 1 M diethanolamine buffer, 10 mM 4-nitrophenylphosphate, 0.25 mM MgCl,, pH 9.8) of calf intestinal alkaline phosphatase was added (0.01 volume) to 2.5 ml of the crude extract (1.1 mg of protein). The crude extract was prepared from 4 flasks by alkaline lysis and passage through PD-10 columns eluted with 25 mM imidazole HCI buffer, pH 7.0, as described above. The activity of alkaline phosphatase under the conditions adopted here (pH 7.0,30 "C) was only 1-2% of that determined under the standard conditions (pH 9.8, 37 "C). Control reactions were conducted in a similar manner, but without alkaline phosphatase. The reaction was terminated by the addition of sodium vanadate (250 p~) .
This preparation was applied to a PBE 94 column as described above.

Pretreatment of Cell-free Extract with CAMP-dependent Protein Kinase
Cyclic AMP-dependent protein kinase was added for 10 min at 30 "C in 0.1 ml containing 50 mM MOPS buffer, pH 7.0, 10 mM MgCl,, 1 mM dithiothreitol, 0.05 ml of extract (5.0 pg of protein), 5 units of CAMP-dependent protein kinase, and 10 p~ ATP. The pretreatments were terminated by the addition of the inhibitor of CAMP-dependent protein kinase at concentrations sufficient to inhibit the transfer of 10 pmol of 3zP to histone (Type V-S)/min by CAMP-dependent protein kinase. The 40 S ribosomal subunits were added and S6 kinase activity of this treated preparation measured. The CAMP-dependent kinase was reconstituted in deionized water containing 50 mg/ml dithiothreitol 10 min prior to use. Control reactions were conducted in a similar manner but with 50 mg/ml dithiothreitol without CAMP-dependent protein kinase.

Pretreatment of Cell-free Extract with Protein Kinase C
Protein kinase C was added for 3 min at 30 "C in 0.1 ml containing 20 mM Tris-HC1 buffer, pH 7.5, 50 mM 2-mercaptoethanol, 5 mM phenylmethylsulfonyl fluoride, 5 mM magnesium acetate, 0.5 mM CaCl,, 12.5 pg/ml phosphatidylserine, 1.25 pg/ml diolein, 0.05 ml of extract (5.0 pg of protein), 1.5 units of protein kinase C, and 10 p~ ATP according to Takai et al. (37). Control reactions were conducted in a similar manner without kinase C. The pretreatments were terminated by the addition of 0.5 mM chlorpromazine (38). Then 40 S ribosomal subunits were added and S6 kinase activity of this treated preparation measured.

Materials
NGF was prepared by the method of Bocchini and Angeletti (39). NECA was a gift of Dr. Dean Londos, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. [Y-~'P]ATP was purchased from New England Nuclear. Purified S6 was kindly provided by Drs. Alan Lin and Ira Wool, University of Chicago. Protein kinase C purified from rat brain was a gift from Dr. Kuo-Ping Huang, National Institute of Child Health and Human Development, Be-thesda, MD. Medium and serum were purchased from Gibco. Heparin-Sepharose CL-GB, Sephadex G-200, PBE 94, Polybuffer 74, and PD-10 (Sephadex G-25M) columns were purchased from Pharmacia. @-Glycerophosphate, chlorpromazine HCI, dibutyryl CAMP, bovine heart CAMP-dependent protein kinase catalytic subunit (20,000 units/mg), bovine heart CAMP-dependent protein kinase inhibitor (Type 11, inhibits 750 units of CAMP-dependent protein kinase/mg), p-nitrophenyl phosphate, imidazole (Grade I), casein (5% solution), histones Type V-S(f,), VI-S(f,a), VIII-S(f3), and phosvitin were purchased from Sigma. Sodium vanadate was purchased from Fisher. Calf intestinal alkaline phosphatase (2000 units/mg), histones HI, H2B, mixed histones, and calmodulin were purchased from Boehringer Mannheim. Rabbit and porcine myelin basic proteins were kindly provided by Dr. K.-T. Jesse Chan, National Institute of Neurological and Communicative Disorders and Stroke, Bethesda, MD. Adult male Sprague-Dawley rats were from Zivic-Miller. All other chemicals were analytical grade.

NGF-sensitiue S6
Kinase-Cell-free lysates were prepared by an alkaline lysis method which has been used to disrupt cells gently for the preparation of plasma membranes (40), as described under "Experimental Procedures." Fig. lA (lanes 1-5 ) shows the results obtained when S6 kinase activity in a lysate from control cells was compared to the S6 kinase activity in a lysate from NGF-treated cells. Increased phosphorylation of S6 (M, = 31,000) was consistently detected and was generally 10-20-fold greater in lysates from NGF-treated cells. Several proteins were phosphorylated in the cell lysates in the presence or absence of 40 S ribosomal subunits; no increase in the phosphorylation of any other protein was observed. This observed difference in specific activity was maintained even after S6 kinase activity from control or NGFtreated cells was purified by heparin-Sepharose CL-GB, gel FIG. 1. Activation of nerve growth factor-sensitive S6 kinase. 40 S ribosomal subunits from rat liver were incubated with [r-:"P]ATP and crude cell-free extracts ( A ) or purified S6 kinase ( R ) from PC12 cells prepared as described under "Experimental Procedures." After 30 min a t 30 "C, the reactions were stopped with SDSsample buffer, and the samples were analyzed by SDS-gel electrophoresis on 10% acrylamide gels. The gel was then stained, dried, and exposed to x-ray film. A: lone I, 40 S ribosomal subunits (0.34 AZW unit); lone 2, extract from control cells (5.0 pg of protein/assay); lone 3, 40 S ribosomal subunits plus extract from control cells; lone 4, extract (5.0 pg of protein) from NGF-treated cells (50 ng/ml, 60 min); lone 5,40 S ribosomal subunits plus extract from NGF-treated cells.
R: cell-free extracts from control and NGF-treated PC12 cells were further purified, in parallel, by heparin Sepharose CL-GB column chromatography, gel filtration, and chromatofocusing column chromatography. Assuming equal recovery on all columns, equal amounts (less than 13 ng of protein/assay) of purified S6 kinase activity from NGF-treated (lane I ) or control cells ( l o n e 2) were incubated with 40 S ribosomal subunits and analyzed as above. Arrows indicate the position of ribosomal protein S6. filtration chromatography, and PBE 94 column chromatography (Fig. lB, lanes I and 2) as described below.
Purification of S6 Kinase-S6 kinase was purified extensively from control and from NGF-treated PC12 cells. Soluble cell-free extracts, prepared as described under "Experimental Procedures," were subjected to heparin-Sepharose CL-GB column chromatography and elution with a linear gradient of 0-0.75 M NaCI. Two peaks of S6 kinase activity, designated I and I1 in their order of appearance from the column, were eluted by 30 and 130 mM NaCl (Fig. 2, A-D). Peak I is higher in extracts from NGF-treated cells; peak I1 is not. When the alkaline lysis of cells was carried out in the presence of 80 mM P-glycerophosphate and 20 mM EGTA, additions said to be required during the preparation of cell extracts in order to obtain fully active S6 kinase(s) in growth factor-stimulated Swiss 3T3 cells (14), the activity in peak I virtually disappeared (Fig. 2, E and F); the activity in peak 11, it should be noted, was significantly stimulated. The further purification of the NGF-sensitive S6 kinase in peak I, then, was done in the absence of P-glycerophosphate or EGTA; in all subsequent steps this activity fractioned as a single peak. Peak I from the heparin-Sepharose CL-GB column was chromatographed on Sephadex G-200 (Fig. 3). The S6 kinase activity eluted with an apparent molecular weight of 45,000. The final step of purification was chromatography on a PBE 94 chromatofocusing column and elution with a pH gradient of 7-4 using Polybuffer 74. The amount of protein present at this point was too low to obtain a protein profile by the Bradford method (32). The S6 kinase activities from control cells and from NGF-treated cells were eluted at pH 5.69 and 5.58, respectively (Fig. 4). This step of column chromatography (chromatofocusing) has been repeated a t least 20 times; the shift of elution position (pH difference about 0.1) between S6 kinases from control and NGF-treated cells was seen each time. Furthermore, if the pH curve was designed to give a more mild slope, for example pH 6.4-4.5, the peaks of activity could be separated more completely (6 fractions difference in Fig. 4A); if the pH curve was steeper, for example pH 7.4-3.5, the peak positions were closer to each other (data not shown). Assuming equal recovery of activity at all steps, the difference in specific activity between S6 kinase activities from control cells and from NGF-treated cells was maintained throughout the purification.
It was impossible to add 40 S subunits at a concentration sufficient to saturate the enzyme, and thus, the specific activity values obtained may underestimate the activity. In any case, the final enzyme preparation from NGF-treated PC12 cells was purified approximately >15-fold from the crude extract with a very low recovery. The enzyme was very unstable; a great deal of activity was lost during the purification procedure, and the purified enzyme lost its residual activity completely within 15 h. The addition of dithiothreitol (1 mM) and M$+ (10 mM) did not restore the lost activity and was not effective in preserving activity. Storage a t -80 "C in the presence of 50% glycerol also did not prevent inactivation.
The peak in Fig. 4A was pooled, concentrated, electrophoresed on a 10% SDS gel, and silver stained by the procedure of Oakley et al. (41), but no protein band was detected.
In order to obtain an active purified preparation of S6 kinase, a rapid and improved purification method was developed (Purification Method 2). T o provide a shorter handling time, generally, four cultured flasks were used to prepare one sample as described under "Experimental Procedures." For this improved method, the cell-free lysate was passed through a PD-10 column to exchange the buffer for that used in the subsequent PBE 94 column. In this one-step purification, the activity of S6 kinase was eluted at the same position as in Purification Method 1 as a single peak that was well resolved from the bulk of the protein in the preparation (Fig. 5). The S6 kinase was purified a t least 673-fold (specific activity, >9500 pmol/min/mg of protein) with a recovery of activity of approximately 70%. Furthermore, the S6 kinase activity from NGF-treated cells was again eluted a t a more acidic pH than was that from control cells. The specific activity of the Sf3 kinase in preparations from NGF-treated cells is 3-10-fold higher than that from control cells. This improved one-step purification method (Purification Method 2) was used to prepare enzyme for the following experiments.

Nerve Growth Factor-sensitive S6 Kinase from PC12 Cells
Effect of Protein Concentration and Incubation Time on S6 Phosphorylation in Vitro-In order to choose optimal conditions under which to measure S6 kinase activity in vitro, the effect of an increasing concentration of an enzyme purified Nerve Growth Factor-sensitiue S6 Kinase from PC12 Cells  Fig. 2, A and € 3 , were pooled separately and concentrated in an Amicon ultrafiltration cell equipped with a PM-10 membrane. The samples were applied to a Sephadex G-200 column (90 X 1.5 cm) equilibrated with 25 mM imidazole HCl buffer, pH 7.0. The column was eluted at a flow rate of 0.17 ml/rnin with the same buffer. Fractions of 1.85 ml were collected, and portions (10 pl) of each fraction were incubated with [Y-~*P]ATP for 30 min at 30 "C according to the routine assay method. Samples were subjected to SDS-gel electrophoresis and autoradiography. The elution profiles of S6 kinase activity from control cells (0) and from NGF-treated cells (0) were determined by liquid scintillation spectroscopy of the 32P incorporated into the gel band corresponding to S6. The elution profiles of protein (---) from control and from NGF-treated cells were measured at 280 nm. Aldolase (168,000), bovine serum albumin (68,000), and chymotrypsinogen A (25,000) (Boehringer Mannheim) were used as molecular weight standards, W, indicates the elution of blue dextran 2000 (Pharmacia P-L Biochemicals).
by Purification Method 2 from cells treated for 60 min with NGF was examined. The results in Fig. 6A show that the amount of 32P incorporated into S6 in 30 min increases linearly up to the strength of an undiluted preparation (the protein concentration is less than 13 ng/assay) from a PBE 94 column peak. The reaction is linear for almost 60 min (Fig.  6B). Therefore, in experiments described below, assays were carried out for 30 min with undiluted enzyme from the PBE 94 column peak.
Substrate Specificity-The S6 kinase preparations obtained by both Purification Methods 1 and 2 were tested for their ability to catalyze the phosphorylation of S6, casein, phosvitin, histones fia, f3, H2B, and H1, mixed histone, and myelin basic protein. Of the substrates tested, the NGF-sensitive S6 kinase exhibited activity only toward S6 ( Table I). The catalytic subunit of CAMP-dependent protein kinase also phosphorylates S6 strongly (specific activity, 11 nmol/min/mg of protein) under the assay conditions used in these experiments. However, in contrast, the CAMP-dependent protein kinase also catalyzed the phosphorylation of casein, phosvitin, and histone H1 at least as well as it did S6; its activity was inhibited by the protein inhibitor of CAMP-dependent protein kinase (data not shown).
Properties of the Purified Enzyme-The purified enzyme was found to be inhibited by several substances. As shown in Table 11, EGTA, @-glycerophosphate, and NaF were potent inhibitors. The enzyme was also inhibited slightly by calmodulin. No effect was observed with chlorpromazine, sodium vanadate, or Ca2+. These results coincide with those obtained with crude extract (31). In addition, the inability of the protein inhibitor of CAMP-dependent protein kinase to block phosphorylation by the S6 kinase provides further evidence that this preparation is free of catalytic subunits of CAMP-dependent protein kinase.
Phosphorylation of S6 in Vitro-In order to confirm that the phosphorylated protein M, = 31,000 shown in Fig. 1 was indeed S6 and to determine the distribution of radiolabel among the derivatives, 40 S ribosomal subunits were phosphorylated in vitro, extracted, analyzed by two-dimensional gel electrophoresis, and the gel subjected to autoradiography (Fig. 7). By superimposing the autoradiogram and the stained gel (Fig. 7A), phosphorylation by S6 kinase purified from NGF-treated cells can be seen as an increase in the radioactivity associated with S6 and corresponding shift in the mobility of phosphorylated S6 during two-dimensional electrophoresis (35); we can estimate that 4-5 phosphate groups were incorporated per S6 molecule, as shown by the spots designated d and e in Fig. 7B. With purified enzyme from control cells, only minor phosphorylation of S6 was observed (Fig.  7C). Phosphorylation of other ribosomal proteins was negligible. The phosphorylation of S6 in vitro occurred solely on serine residues (Fig. 8).
Time Course of Nerve Growth Factor Effect-To determine the time course of the NGF effect, cells were harvested at various times after NGF treatment, and cell-free extracts were prepared and purified by PBE 94 column chromatography (Purification Method 2) as described under "Experimental Procedures." The active fractions were pooled, and the pooled activity was measured. As shown in Fig. 9, halfmaximal 32P incorporation into ribosomal S6 protein occurred in approximately 5 min. After 1 h of exposure to NGF, the S6 kinase activity was beginning to decrease and by 2 h had almost returned to the basal level.
Effect of Pretreatment with Alkaline Phosphatase-It is reasonable to suggest that the effects of NGF on S6 kinase shown in Figs. 4 and 5, that is a higher specific activity and a shift of the elution position t o a more acidic pH on the chromatofocusing PBE column, are brought about by phosphorylation of the kinase. To explore this possibility, the effect of pretreatment of the S6 kinase preparation with alkaline phosphatase was examined according to the method described under "Experimental Procedures." As shown in Fig.  10, pretreatment of cell-free extracts from NGF-treated cells with alkaline phosphatase causes a decrease in activity and a shift in the elution of the kinase to the position of S6 kinase from control cells. On the other hand, the S6 kinase in extracts prepared from control cells showed no change in activity or subsequent elution pattern on the PBE 94 column due to pretreatment with alkaline phosphatase. These results strongly suggest that phosphorylation is involved in the activation of S6 kinase from PC12 cells by NGF. This experiment has been repeated five times with qualitatively similar results. When the S6 kinase purified by Method 1 under "Experimental Procedures" was used, qualitatively similar results were also obtained (data not shown).
Effect of Dibutyryl CAMP, NECA, and Nerve Growth Factor on S6 Kinase Activity in PC12 Cells-In a previous paper (31) we showed that treatment of intact PC12 cells with dibutyryl cAMP or NECA stimulated the subsequent cell-free phosphorylation of S6. It was suggested that although S6 kinase is not CAMP-dependent protein kinase per se, the action of a CAMPdependent protein kinase is involved in S6 kinase activation.
Figs. 11C and 12C show the elution patterns of S6 kinase from extracts prepared from the cells treated with dibutyryl cAMP or NECA on PBE 94 column chromatography. Each fraction was assayed in the presence of the inhibitor of CAMPdependent protein kinase to block the direct action of CAMPdependent protein kinase on S6. The elution patterns were indistinguishable from that produced by the treatment of PC12 cells by NGF alone (Figs. 1lB and 12B). That is, the specific activity of S6 kinase is much higher in preparations  Fig. 3 were pooled and applied to PRE 94 columns (12.5 X 0.9 cm) equilibrated with 25 mM imidazole HCI buffer, pH 7.0. The column was eluted a t a flow rate of 0.33 ml/min with elution huffer (0.0075 mmol/pH unit/ml of Polybuffer 74, pH 4.0). Fractions of 1.0 ml were collected, and portions (10 pl) of each fraction were incubated with [r-""P]ATP for 30 min at 30 "C according to the routine assay method. Samples were subjected to SDS-gel electrophoresis and autoradiography. A, the elution profiles of S6 kinase activity from control cells (0) and from NGF-treated cells (0) were determined by liquid scintillation spectroscopy of the "P incorporation into the gel band corresponding to S6. The gradient of pH was measured by pH meter (---). Readings of the absorption a t 280 nm were all a t background levels. Autoradiograms from 10% SDS-polyacrylamide gels of extracts from control cells ( B ) and NGF-treated from NGF-treated cells, and the elution position of the enzyme was shifted to a more acidic pH on the PBE column as compared to that of control cells (Figs. 1 l A and 12A). Furthermore, the extents of stimulation of phosphorylation achieved by the individual ligands were compared with those seen when the ligands were added in combination. The results of these experiments are shown in Figs. l l D and 1 2 0 . NGF and dibutyryl cAMP were not additive in the phosphorylation of S6 (Fig. l l D ) . The combination of NGF and NECA also was not additive (Fig. 12D), pointing to cAMP as a convergence point in the actions of NGF and NECA on S6 kinase stimulation. These results suggest that the enhanced phosphorylation of S6 in cells treated with NGF or cAMP occurs through a common mechanism. These experiments have been repeated three times with qualitatively similar results.
The Effect of Pretreatment with CAMP-dependent Protein Kinase or Protein Kinase C-Experiments were designed to see if an involvement of CAMP-dependent protein kinase or protein kinase C in the NGF-dependent activation of S6 FRACTION kinase could be demonstrated directly. Cell-free extracts prepared by alkaline lysis from control or from NGF-treated cells were pretreated with the catalytic subunit of CAMP-dependent protein kinase or with protein kinase C as described under "Experimental Procedures." The pretreatments were terminated by the addition of the inhibitor of CAMP-dependent kinase or chlorpromazine, respectively. Then 40 S ribosomal subunits were added and the S6 activity of the treated preparation measured. As shown in Fig. 13, pretreatment of cellfree extracts from control cells with the catalytic subunit of CAMP-dependent protein kinase caused an increase in S6 kinase activity (Fig. 13A, lanes 1 and 2), but it also elicits an increase in extracts from NGF-treated cells (Fig. 13A, lanes 3  and 4 ) . More detailed experiments were carried out using PBE 94 column chromatography in order to clarify the involvement of CAMP-dependent protein kinase in S6 kinase activation (Fig. 14, A-D). In this experiment, the assay of S6 kinase was carried out in the presence of the inhibitor of CAMP-dependent protein kinase to avoid the direct involvement of CAMP-

Nerve Growth
Factor-sensitive S6 Kinase from PC12 Cells Crude extracts (1.12 mg of protein from 4 flasks of cells) were prepared as described under "Experimental Procedures" from control cells or NGF-treated (50 ng/ml, 60 min) cells, filtered through Millex-GS (0.22 pm, Millipore) filters, and applied to PD-10 columns (Pharmacia) equilibrated with 25 mM imidazole HCI buffer, pH 7.0. The columns were eluted with the same buffer and fractions of 20 drops each (-0.7 ml) collected. The absorption of each fraction was read a t 280 nm, and the peak fractions (4-6) were pooled and applied to PBE 94 columns (12.5 X 0.9 cm) and analyzed as described in the legend to Fig. 4. Fractions of 1.0 ml were collected, and portions of each fraction were incubated with [r-"P]ATP for 30 min a t 30 "C according to the routine assay method. Samples were subjected to SDS-gel electrophoresis and autoradiography. A, the elution profiles of S6 kinase activity from control cells (0) and from NGF-treated cells (0) were determined by liquid scintillation spectroscopy of the a2P incorporated into the gel hand corresponding to S6. The elution profiles of protein (-) were measured a t 280 nm. The gradient of pH was measured by pH meter (---). Autoradiogram from 10% dependent protein kinase in S6 phosphorylation. Under these conditions, no activity of the catalytic subunits of CAMPdependent protein kinase in the direct phosphorylation of S6 was observed (Fig. 14E). As shown in Fig. 14F, pretreatment of cell-free extracts from control cells with the catalytic 4 . ATP and the purified S6 kinase from NGF-treated PC12 cells (50 ng/ml, 60 min), and "P incorporated into S6 was measured as described under "Experimental Procedures." In A, increasing concentrations of the purified protein were incubated for 30 min a t 30 "C. The relative concentration 1 indicates undiluted samples from the highest active fraction from PBE 94 column chromatography (Fig.   5A, less than 13 ng of protein/assay). In R, aliquots of the undiluted fraction from A (relative concentration, 1) were incubated for increasing times a t 30 "C. subunit of CAMP-dependent protein kinase produces an increase in S6 kinase activity and causes it to elute at a more acidic pH region, but the elution profile of S6 kinase was not the same as that of S6 kinase from NGF-treated cells. Pretreatment of cell-free extracts from NGF-treated cells with the catalytic subunit of CAMP-dependent protein kinase elicited a comparable activation and shift of elution peak of S6 kinase on the PBE 94 column. But two peaks of S6 kinase activity were observed in the elution profile of pretreated extracts. One coincides with the activated S6 kinase from NGF-treated cells, and the other is eluted at an even more acidic region. This experiment has been repeated five times with qualitatively similar results. When S6 kinase prepared by Purification Method 1 under "Experimental Procedures" was used, qualitatively similar results were obtained (data not shown). Protein kinase C treatment had no effect on the activity of S6 kinase under the present conditions (Fig. 1323).

DISCUSSION
The results presented here describe a procedure for purifying an S6 kinase from PC12 cells highly specific for S6, the major phosphorylated protein in 40 S ribosomal subunits. The homogeneity of the purified enzyme could not be judged by SDS-polyacrylamide gel electrophoresis and silver staining because of the low amount of enzyme available. Furthermore, the NGF-sensitive S6 kinase is very unstable and loses most

PC12 cells
Kinase reactions were carried out for 30 min a t 30 "C as described under "Experimental Procedures" with the indicated substrate. The radioactivity in S6 was quantified by SDS-gel electrophoresis on 15% acrylamide and excision of the S6 band. Portions (10 pl, less than 13 ng of protein/assay) of fractions from the PRE 94 column were assayed with the indicated substrate. Two different enzyme preparations were used: incubations in Experiment 1 contained the enzyme prepared from NGF-treated PC12 cells by Purification Method 1 (Fig. 4A), while incubations in Experiment 2 contained the enzyme prepared from NGF-treated PC12 cells by Purification Method 2 (Fig. 5A).  Effect of various agents on nerve growth factor-sensitive S6 kinase from PC12 cells Kinase reactions were carried out for 30 min a t 30 "C as described under "Experimental Procedures" with the indicated additions. The radioactivity in S6 was quantified by gel electrophoresis and excision of the S6 band. Portions (10 pl, less than 13 ng of protein/assay) of fractions from the PRE 94 column were assayed in the presence of each drug at the indicated final concentration. Two different enzyme preparations were used; incubations in Experiment 1 contained the enzyme prepared from NGF-treated PC12 cells by Purification Method 1 (Fig. 4A), while incubations in Experiment 2 contained the enzyme prepared from NGF-treated PC12 cells by Purification Method 2 (Fig. 5A).

Agents
Concen- of its activity during handling. It is noteworthy that the addition of 8-glycerophosphate and EGTA, a combination required during the preparation to obtain fully activated S6 kinase(s) in Swiss 3T3 cells (14), to the lysis buffer used in this paper caused an almost complete inhibition of the action )e S6 Kinase from PC12 Cells  Fig. 5A) as described under "Experimental Procedures" except that the reaction was carried out for 2 h. The reaction was terminated by the addition of 950 pg of unlabeled 40 S ribosomal subunits, 0.1 volume of 1 M MgCI,, and 2 volumes of glacial acetic acid. The ribosomal proteins were extracted, precipitated with -20 "C acetone, and subjected to two-dimensional gel electrophoresis as described under "Experimental Procedures." A, the Coomassie Bluestained gel (S6 kinase from NGF-stimulated cells). R and C, the corresponding autoradiograms of the region containing ribosomal protein S6 phosphorylated by the purified S6 kinase from NGFstimulated cells ( R ) and control cells (C). In R, d and e represent the positions of S6 phosphorylated with 4 and 5 mol of phosphate/mol of S6, respectively. The arrows indicate the position of unphosphorylated S6.

A FIRST DIMENSION
of NGF-sensitive S6 kinase. Indeed, the purified S6 kinase from PC12 cells is also quite sensitive to these compounds.
It is important to compare the properties of this S6 kinase with the properties of other S6 kinases described by various investigators and especially to the properties of those shown to be stimulated by treatment of the cells with various peptide effectors. A significant feature of the present enzyme is its high degree of specificity for S6 relative to other substrates frequently used for protein kinases in uitro. The enzyme did not phosphorylate casein, phosvitin, various histones, or myelin basic protein. The S6 kinase is clearly distinct from CAMPdependent protein kinase because it is unaffected by the specific inhibitor of CAMP-dependent protein kinases and because CAMP-dependent protein kinase can phosphorylate a number of the substrates described above. The S6 kinase studied here is not inhibited by chlorpromazine, a potent kinase C inhibitor (38). Ca'+/calmodulin does not stimulate the S6 kinase activity. Thus, the properties of the NGFsensitive S6 kinase from PC12 cells also distinguish it from such well-defined kinases as protein kinase C, casein kinase, from rat liver were incubated with [y-"P]ATP and portions (10 pl; less than 13 ng of protein) of the pool of active fractions of S6 kinase for 30 min a t 30 "C according to the routine assay method. The reactions were stopped with SDS sample buffer and subjected to SDSgel electrophoresis, and "P incorporation into S6 was measured by liquid scintillation spectroscopy. Results are the means of three independent experiments and are expressed in percentage of maximal activity; maximal activity was 6500 cpm of "P incorporated into S6.  40 S rihosomal subunits (0.34 A,,,, units) from rat liver were incubated with [y-''"PIATP and purified S6 kinase (10 pl of the active fraction shown in Fig. 5A) from NGF-treated cells as described under "Experimental Procedures." After 30 min a t 30 "C, the reaction was stopped with SDS sample buffer and the sample was analyzed by SDS-gel electrophoresis on a 10% acrylamide gel. The S6 hand was excised from the gel and hydrolyzed with 6 N HCI for 2 h a t 110 "C in a nitrogen-filled sealed tube. The hydrolysate was analyzed by thin layer cellulose electrophoresis as described under "Experimental Procedures." The autoradiogram from such a plate is shown. Appropriate standards were added to each sample, and their positions are indicated by the dotted outlines. SerlP), phosphoserine; ThrfP), phosphothreonine; Tyr(P), phosphotyrosine. and calmodulin-dependent protein kinase. Lubben and Traugb (22) have reported that a partially purified preparation of protease-activated kinase I1 is an S6 protein kinase. The protease-activated kinase I1 proenzyme migrates on gel filtration with a M , = 80,000, and the proteaseactivated species has a M , = 45,000-55,000 (22). But, unlike the protease-activated species, the S6 kinase from PC12 cells does not phosphorylate histone H1, and our previous studies do not support a protease activation mechanism for the S6 kinase from PC12 (31). Erikson and Maller (30) reported the purification of S6-specific kinase from Xenopus eggs, but its of each fraction for 30 min a t 30 "C according to the routine assay method. The reactions were stopped with SDS sample buffer and subjected to SDS-gel electrophoresis and autoradiography. Untreated extract from control cells, untreated extract from NGF-treated cells, alkaline phosphatase-treated extract from control cells, and alkaline phosphatase-treated extract from NGF-treated cells are arranged in order. The fraction numbers are shown at the bottom. molecular weight was found to be 92,000. Based on substrate specificity and molecular size alone, the NGF-sensitive S6 kinase is most similar to an insulin-stimulated S6 kinase from 3T3 cells; the latter has M , = 50,000-60,000 (12).
A glia-derived neurite-promoting factor with protease inhibitory activity (42) has properties quite similar to those of

FRACTION NUMBER
FIG. 11. Effect of dibutyryl (dB) cAMP and nerve growth factor on S6 kinase activity in PC12 cells. PC12 cells were treated with NGF, with dibutyryl CAMP, or with both for 60 min. The soluble cell-free extracts (1.12 mg of protein) were prepared, passed through PD-10 columns, and subjected to PBE 94 column chromatography as described for Purification Method 2 under "Experimental Procedures." 40 S ribosomal subunits (0.34 A,, units) from rat liver were incubated with [y3'P]ATP, and portions (10 pl) of each fraction were assayed in the presence of an amount of CAMPdependent protein kinase inhibitor which is sufficient to inhibit the transfer of 10 pmol of 32P to histone (Type V-S)/min by CAMPdependent protein kinase for 30 min at 30 "C according to the routine assay method. The reactions were stopped with SDS sample buffer and subjected to SDS-gel electrophoresis, and 32P incorporation into S6 ( U ) was measured by liquid scintillation spectroscopy. The pH gradients (---) were measured with a pH meter. In all cases, the ligands were added at concentrations which are maximal for the stimulatory effect: A , control; B, NGF (50 ng/ml); C; dibutyryl cAMP (0.5 mM); D, NGF (50 ng/ml) + dibutyryl cAMP (0.5 mM). this S6-specific kinase from PC12 cells; these include molecular size (Mr = 43,000 for the factor) and chromatographic behavior on both heparin-Sepharose CL-GB and Affi-Gel blue columns (data not shown). But neither urokinase nor plasminogen activator-dependent caseinolysis was inhibited by the S6-specific kinase from PC12 cells (data not shown).
It is also important to know how S6 kinase is activated by prior treatment of the cells with nerve growth factor. Novak-Hofer and Thomas (14,28) have studied the EGF-mediated activation of an S6 kinase in Swiss mouse 3T3 cells and have suggested the possibility that modification of the enzyme by phosphorylation might be involved in its activation. Tabarini et al. (12) have investigated the insulin-stimulated S6 kinase of 3T3-Ll cells and suggested that the activation of tyrosine protein kinases (including the insulin receptor) and protein kinase C trigger discrete biochemical pathways that ultimately converge to generate the same S6 protein kinase. Blenis and Erikson (29) studied a transformation-sensitive S6 kinase and presented evidence that it is regulated by tyrosine phosphorylation or, alternatively, by protein kinase C. Erikson and Maller (30) studied an S6-specific kinase in Xenopus oocytes and also suggested that the S6 kinase is capable of being regulated by protein tyrosine kinases, either directly by tyrosine phosphorylation or indirectly through the action of one or more intermediates between protein tyrosine kinases and protein serine kinases. Thus, several previous reports have suggested that phosphorylation is involved in the mechanism of S6 kinase activation. The data presented here support these previous hypotheses. These results include the obser-  vations that 1) activated S6 kinase from NGF-treated cells is tivation and a shift of the elution position to that of S6 kinase eluted at a more acidic p P H from a chromatofocusing column from control cells. In this regard it is reasonable to mention than the enzyme from control cells and 2) pretreatment of that in some experiments alkaline phosphatase treatment of activated S6 kinase with alkaline phosphatase causes a deac-enzyme from NGF-treated cells produced a kinase that seemed somewhat less stable than comparably treated enzyme from controls. It may be that a simple phosphorylationdephosphorylation mechanism is not sufficient to explain these transitions; some activity-influencing conformational changes may accompany the phosphorylation.
If S6 kinse from PC12 cells is activated by phosphorylation, it is of interest to know what kinase is responsible for the activation. We have reported recently (31) that treatment of intact PC12 cells with dibutyryl cAMP or NECA also increases the subsequent cell-free phosphorylation of S6 in uitro, suggesting that NGF may stimulate protein phosphorylation through an elevation of the intracellular cAMP level. This suggestion is supported by the present data that show that stimulation of protein phosphorylation elicited by saturating amounts of both NGF and NECA (or dibutyryl CAMP) is not additive. These results are consistent with previous data indicating that NGF raises intracellular cAMP levels and that cAMP can mimic some of the effects of NGF on the cells (43) and also support the results of Halegoua and Patrick ( 5 ) indicating that NGF, cholera toxin, and cAMP stimulate the phosphorylation of some proteins (such as S6) through a common mechanism.
The simplest interpretation of our results would be a model in which NGF binds to receptors which then activate adenylate cyclase to elevate intracellular cAMP levels. CAMP-dependent protein kinase would subsequently be activated to phosphorylate its substrates, among them S6 kinase. To gain support for this model, we asked whether or not CAMPdependent protein kinase directly phosphorylates and activates S6 kinase in uitro. Our data indicate that pretreatment of cell-free extracts from control cells with CAMP-dependent protein kinase results in an activation of S6 kinase activity, but the elution profile of the pretreated enzyme on PBE 94 column was not the same as that of activated kinase from NGF-treated cells. Furthermore, the pretreatment also elicits some activation of $36 in extracts from NGF-treated cells. To interpret the effect of CAMP-dependent protein kinase, there are two possibilities. 1) CAMP-dependent protein kinase does cause a phosphorylation-activation of S6 kinase in vivo but catalyzes a further incorporation of phosphate in uitro to produce a hyper-phosphorylated enzyme. In this case, if we choose more appropriate conditions, such as incubation time and enzyme concentration, we may be able to get an elution pattern on PBE 94 column comparable to that of activated S6 kinase from NGF-treated cells. 2) Pretreatment with CAMP-dependent protein kinase is artificial and does not reflect events in vivo.
Treatment of cell extracts from control or NGF-treated cells with protein kinase C had no effect on the activation of S6 kinase under the present conditions. This result is consistent with the fact that 12-0-tetradecanoylphorbol 13-acetate treatment of intact PC12 did not elicit an increase of S6 kinase activity (31). It should be noted that NGF activates a different pathway in S6 phosphorylation in PC12 cells than that of NsplOO phosphorylation, in which protein kinase C is involved (44).
In conclusion, we can say from the results presented here that NGF causes an activation of CAMP-dependent protein kinase following elevation of intracellular cAMP level, and this CAMP-dependent protein kinase directly or indirectly causes the phosphorylation-activation of S6 kinase in PC12 cells. To confirm these results more experiments, including 1) an evaluation of which amino acid residue in S6 kinase is involved in its activation and 2) an inquiry into the relationship between the degree of phosphorylation and the degree of activation of S6 kinase are needed.