Physiological Phosphorylation of Protein Kinase A at Thr-197 Is by a Protein Kinase A Kinase

Abstract

Phosphorylation of the catalytic subunit of cyclic AMP-dependent protein kinase, or protein kinase A, on Thr-197 is required for optimal enzyme activity, and enzyme isolated from either animal sources or bacterial expression strains is found phosphorylated at this site. Autophosphorylation of Thr-197 occurs in Escherichia coli and in vitro but is an inefficient intermolecular reaction catalyzed primarily by active, previously phosphorylated molecules. In contrast, the Thr-197 phosphorylation of newly synthesized protein kinase A in intact S49 mouse lymphoma cells is both efficient and insensitive to activators or inhibitors of intracellular protein kinase A. Using [³⁵S]methionine-labeled, nonphosphorylated, recombinant catalytic subunit as the substrate in a gel mobility shift assay, we have identified an activity in extracts of protein kinase A-deficient S49 cells that phosphorylates catalytic subunit on Thr-197. The protein kinase A kinase activity partially purified by anion-exchange and hydroxylapatite chromatography is an efficient catalyst of protein kinase A phosphorylation in terms of both a low K_m for ATP and a rapid time course. Phosphorylation of wild-type catalytic subunit by the kinase kinase activates the subunit for binding to a pseudosubstrate peptide inhibitor of protein kinase A. By both the gel shift assay and a [γ-³²P]ATP incorporation assay, the enzyme is active on wild-type catalytic subunit and on an inactive mutant with Met substituted for Lys-72 but inactive on a mutant with Ala substituted for Thr-197. Combined with the results from mutant subunits, phosphoamino acid analysis suggests that the enzyme is specific for phosphorylation of Thr-197.

Catalytic (C) subunit of cyclic AMP (cAMP)-dependent protein kinase (protein kinase A [PKA]) requires phosphorylation at Thr-197 for expression of full activity, and this residue is found phosphorylated in both the enzyme isolated from animal tissues and in recombinant C subunit expressed in Escherichia coli (26, 33, 38). In addition to lowering the K_m values for both ATP and peptide substrates, the Thr-197 phosphate causes a distinctive reduction in the mobility of the protein in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (33). Although C subunit is also phosphorylated at Ser-338 in both bacteria and mammalian cells and can be phosphorylated on additional Ser residues, these phosphorylations do not appear to affect C-subunit activity and have only minor effects on the SDS-PAGE mobility of the protein (6, 26, 33, 38).

Thr-197 falls in the activation loop region contained within subdomain VIII that also is associated with activating phosphorylation sites in many other protein kinases, including CDC2 kinase, the mitogen-activated protein (MAP) kinases, the MAP kinase kinases, and most protein tyrosine kinases (12, 13, 38). The sequence in this region is fully conserved in mammalian C subunits, including Cα, Cβ, and Cγ isoforms (3, 27, 37). Activation of protein tyrosine kinases by phosphorylation in this region appears to be by autophosphorylation (13), while that of CDC2, MAP kinases, and MAP kinase kinases is by heterologous enzymes (8, 12). C-subunit phosphorylation in E. coli is apparently an intermolecular autophosphorylation reaction, and the purified recombinant protein is capable of autophosphorylation with concomitant activation (33, 38). In the present report, we present evidence that the phosphorylation of C subunit in intact mammalian cells is catalyzed by a heterologous PKA kinase. Furthermore, we describe an activity from extracts of a PKA-deficient mutant of S49 mouse lymphoma cells that appears to phosphorylate C subunit specifically at Thr-197.

MATERIALS AND METHODS

Expression and radiolabeling of recombinant C subunits.

Wild-type and mutant forms of recombinant murine Cα subunit were expressed from the pET-8c expression vector in E. coli BL21(DE3) as described previously (33). Construction of the wild-type and Thr-197→Ala plasmids has been described elsewhere (33). The Lys-72→Met mutation was introduced by replacement of an NcoI-BstEII restriction fragment from sequences amplified from pMT-CαK72M-EV (17), using an upstream PCR primer modified to introduce an NcoI restriction site overlapping the C-subunit initiation codon. For the experiment represented in Fig. 1, the wild-type C subunit was coexpressed with yeast N-myristoyltransferase from plasmid pBB131 to generate the N-terminally myristoylated form (9).

FIG. 1.

The Thr-197 phosphorylation of C subunits expressed in E. coli is limited by the intracellular activity of C subunit and inhibitable with H-89. E. coli BL21(DE3) containing both a wild-type C-subunit expression plasmid and the yeast N-myristoyltransferase expression plasmid pBB131 were induced in minimal medium at about 24°C for 45 (lanes a and b), 67 (lanes c and d), or 90 (lanes e and f) min before addition of rifampin and incubation an additional hour as described in Materials and Methods. Samples of 250 μl were then shifted to 37°C and incubated for 5 min without (lanes a, c, and e) or with (lanes b, d, and f) 100 μM H-89 before addition of 10 μCi of [³⁵S]methionine and labeling for 10 min at this temperature. Additional samples without H-89 were incubated in parallel but without the addition of radioactivity for determinations of C-subunit specific activity (see text). Cells were extracted by sonication, and about 5,000 cpm of soluble extract protein was subjected to SDS-PAGE. Shown are C-subunit patterns from a 4-day autoradiographic exposure of the resulting gel. Positions of the Thr-197-phosphorylated (C_P) and nonphosphorylated (C_N) forms of C subunit are indicated.

Recombinant C subunits were labeled with [³⁵S]methionine in the presence of 200 μg of rifampin per ml as described by Studier et al. (34). Expression cultures at an optical density at 550 nm of 0.7 in minimal A medium (10.5 g of monobasic potassium phosphate, 4.5 g of dibasic potassium phosphate, 0.39 g of sodium citrate, 1 g of ammonium sulfate, and 2 g of glucose per liter) plus 100 μg of ampicillin per ml (and 100 μg of kanamycin per ml for pBB131-containing cells) were induced with isopropylthiogalactoside for 1.5 h at 24°C before addition of the rifampin. After 15 min or 1 h of incubation, [³⁵S]methionine was added to 100 μCi per ml and labeling was allowed to proceed for an hour. (The shorter rifampin treatment time improved incorporation but also allowed somewhat more background labeling of bacterial proteins.) To prevent autophosphorylation of wild-type C subunit, 100 μM N-[2-(p-bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide (H-89) was added to wild-type cultures at the time of induction. This was unnecessary for the Ala-197 and Met-72 mutant C subunits, since they did not autophosphorylate. (The presence of H-89 during labeling had no effect on subsequent phosphorylation of the Met-72 C-subunit preparation by either wild-type C subunit or the cellular PKA kinase activity [6].) Bacteria were harvested, washed, and extracted by indirect sonication in EB (10 mM Tris-HCl [pH 7.5], 2 mM dithiothreitol, 0.1 mM EDTA) as described previously (33). Supernatant fractions after 15 min of centrifugation at 11,000 × g were dialyzed against two changes of C-subunit storage buffer (100 mM 2-[N-morpholino]ethanesulfonic acid [MES; pH 6.5], 100 mM potassium phosphate, 2 mM dithiothreitol, 0.1 mM EDTA, 50% glycerol) and stored at −20°C. For the phosphate labeling experiments represented in Fig. 8 and 9, unlabeled, nonphosphorylated preparations of wild-type and mutant C subunits were made as for the labeled preparations except that the rifampin and [³⁵S]methionine were omitted.

FIG. 8.

PKA kinase can catalyze transfer of [³²P]phosphate from [γ-³²P]ATP to wild-type or Met-72 C subunit but not to Ala-197 C subunit. Bacterial extracts containing unlabeled, nonphosphorylated wild-type (lanes a to d), Met-72 (lanes e to h), or Ala-197 C (lanes i to k) subunit were incubated with [γ-³²P]ATP and active C subunit (lanes a, e, and i), no enzyme (lanes b, f, and j), PKA kinase (lanes c, g, and k), or PKA kinase and H-89 (lanes d and h) as described in Materials and Methods. Samples were subjected to SDS-PAGE and visualized by either Western immunoblot detection using an anti-C subunit antibody (A) or autoradiography (B). Only portions of gel patterns containing C subunits are shown, and positions of the Thr-197-phosphorylated (C_P) and nonphosphorylated (C_N) forms of wild-type C subunit are indicated as for Fig. 1, 3, and 5.

FIG. 9.

PKA kinase-dependent phosphorylation of C subunit is specific for threonine. Wild-type (lanes a to c) or Met-72 (lanes d and e) C subunit was incubated with active C subunit (lane a) or with partially purified PKA kinase in the absence (lanes b and d) or presence (lanes c and e) of H-89 as for Fig. 8 but with 10 times the amounts of [γ-³²P]ATP (Materials and Methods). The C subunits were resolved by SDS-PAGE, excised from dried gels, and hydrolyzed with hydrochloric acid to allow analysis of their labeled phosphoamino acids by thin-layer electrophoresis. Ser-P and Thr-P indicate positions of unlabeled phosphoserine and phosphothreonine markers in the resulting autoradiographic patterns.

Culture and radiolabeling of S49 cells.

Adenylyl cyclase-negative (subline 94.15.1) and kinase-negative (subline 24.6.1) S49 mouse lymphoma cells were grown in suspension culture in Dulbecco’s modified Eagle’s medium (DMEM) with 2.24 g of sodium bicarbonate per liter, 3 g of glucose per liter, and 10% heat-inactivated horse serum as described previously (29, 31, 32). For labeling, cells were centrifuged, washed, and resuspended at 5 × 10⁷ per ml in low-methionine medium (methionine-free DMEM supplemented with 2.5 μM l-methionine, 10 mM HEPES, and 10% dialyzed, heat-inactivated horse serum [28]). After 5 min of preincubation without or with 100 μM 8-(2-chlorophenylthio)-cAMP (CPT-cAMP) and/or 100 μM H-89, [³⁵S]methionine was added to 1 mCi per ml, and labeling was allowed to proceed for 10 or 15 min at 37°C. For chase experiments, the cells were then diluted 40-fold with conditioned medium (the filtered supernatant fraction after centrifugation of a mid-log-phase culture of S49 cells) and incubated for an additional 30 min at 37°C. Cells were added to 2.5 volumes or more of ice-cold phosphate-buffered saline (137 mM sodium chloride, 2.7 mM potassium chloride, 4.3 mM dibasic sodium phosphate, 1.5 mM monobasic potassium phosphate) containing 2 mM methionine and harvested by centrifugation (5 min at 1,000 × g for experiments involving a chase or 10 s at 10,000 × g for experiments with only pulse-labeled samples). After aspiration of medium, cell pellets were frozen on dry ice and stored at −70°C. Cells for PKA kinase preparations were harvested in mid-log phase by centrifugation, washed twice with phosphate-buffered saline by resuspension and recentrifugation, resuspended to 2 × 10⁸ per ml in EB, and stored frozen at −70°C.

Assays of protein and C-subunit activity.

Protein was assayed by the method of Lowry et al. (21), using bovine serum albumin as a standard. C-subunit activity was measured as the transfer of [³²P]phosphate from [γ-³²P]ATP to kemptide (the heptapeptide Leu-Arg-Arg-Ala-Ser-Leu-Gly [synthesized and purified by the Molecular Biology Resource Facility of the University of Oklahoma Health Sciences Center]) with both substrates at 100 μM as described previously (35). Cell extracts were assayed both without and with 100 μM cAMP. A unit of activity is the amount that transfers 1 nmol of phosphate per min at 30°C.

Immunoadsorption.

Cell pellets were thawed on ice and extracted with HEB (10 mM Tris-HCl [pH 7.5], 10 mM 2-mercaptoethanol, 5 mM EDTA). Supernatant fractions after 15 min of centrifugation at 11,000 × g were diluted 1 to 1 with twice-concentrated NaF-radioimmunoprecipitation assay (RIPA) buffer (10 mM Tris-HCl [pH 7.4], 158 mM sodium fluoride, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS). Under these conditions, dephosphorylation of endogenous C subunit was not detectable (19). In several experiments, replicate samples were extracted with EB and diluted with twice-concentrated RIPA buffer (NaF-RIPA buffer with sodium chloride instead of sodium fluoride) to permit Thr-197 dephosphorylation of the subunits by endogenous protein phosphatase(s) during the immunoadsorption procedure (19) (e.g., Fig. 2, lane e).

FIG. 2.

Thr-197 phosphorylation of newly synthesized C subunit is rapid in cyclase-negative S49 cells whether or not they are stimulated with a cAMP analog. Cyclase-negative S49 cells were pulse-labeled for 10 min with [³⁵S]methionine in the absence (lanes a, c, and e) or presence (lanes b and d) of 100 μM CPT-cAMP and either harvested immediately (lanes a, b, and e) or chased for 30 min after dilution with conditioned medium without or with drug (lanes c and d). Samples for lanes a to d were extracted with HEB and immunoadsorbed in NaF-RIPA buffer, while that for lane e was extracted in EB and immunoadsorbed in RIPA buffer, as described in Materials and Methods. The radiolabeled C subunit species were resolved by SDS-PAGE and visualized by fluorography. Positions of the Thr-197-phosphorylated (Cα_P and Cβ_P) and nonphosphorylated (Cα_N and Cβ_N) forms of Cα and Cβ subunits are indicated.

Radiolabeled C subunits were purified from the diluted cell extracts by immunoadsorption with an affinity-purified goat antibody to the insoluble form of recombinant murine Cα subunit (19). The extracts were preadsorbed with activated Pansorbin (Calbiochem), and C subunit was immunoadsorbed as described previously (29), using about 2 μg of anti-C subunit per 30 μl of supernatant fractions from the preadsorbed extracts and 5 μl of a 20% suspension of activated Pansorbin. Immunocomplexes were collected by centrifugation for 3 min at 11,000 × g, washed twice by resuspension and recentrifugation in 150 μl NaF-RIPA buffer, and solubilized with 25 μl of 1% SDS containing 1 M 2-mercaptoethanol and 10 mM Tris-hydrochloride (pH 7.4). After 10 min on ice, the solubilized samples were centrifuged as described above, and 20-μl portions of the supernatant fractions were diluted with 80 μl of a mixture containing 50 μl of twice-concentrated NaF-RIPA buffer without SDS and 30 μl of water for readsorption with a second 5.8-μg portion of anti-C subunit and 5 μl of 20% activated Pansorbin (19). After again collecting and washing with NaF-RIPA buffer, the immunocomplex was disrupted by boiling in SDS-gel sample buffer (22) and centrifuged, and the supernatant fractions were saved for scintillation counting and gel electrophoresis.

Gel electrophoresis, Western immunoblotting, fluorography, and quantitation of radioactivity in phosphorylated and nonphosphorylated forms of the radiolabeled C subunit.

SDS-PAGE was carried out as described by Laemmli (18), using 10% polyacrylamide gels. Gels were either dried for direct autoradiography, impregnated with 2,5-diphenyloxazole in dimethyl sulfoxide and dried for fluorography as described by Bonner and Laskey (4), or electroblotted onto Immobilon-P membranes (Millipore Corp.) for immunodetection using goat anti-C subunit (above), an alkaline phosphatase-conjugated rabbit anti-goat immunoglobulin (Cappel Products Division/Organon Teknika Corp.), and Rad-Free Lumi-Phos 530 substrate sheets (Schleicher & Schuell) as described elsewhere (19). Autoradiograms and fluorograms were scanned with a Molecular Dynamics model 300A computing densitometer and quantified by using the IQ software package (Molecular Dynamics) in area scan mode with manual baseline and peak selection. The relative labeling in Cα and Cβ protein species was estimated from samples that had been dephosphorylated as described above.

Assay of PKA kinase activity.

Standard reaction mixtures (4 to 10 μl) contained about 5,000 cpm of [³⁵S]methionine-labeled recombinant C subunit per μl and either kinase-negative cell extract or partially purified fractions of PKA kinase in buffer containing 10 mM 1,3-bis(tris[hydroxymethyl]methylamino)propane (Bis-Tris propane; pH 7.0), 10 mM 2-mercaptoethanol, 0.1 mM EDTA, 6 mM magnesium sulfate, 1 mM ATP, 100 mM potassium chloride, and 0.5 mg of bovine serum albumin per ml. Unless indicated otherwise, samples were incubated for 1 h at 30°C before reactions were stopped by the addition of 19 volumes of SDS-gel sample buffer. Samples of 20 μl (5,000 cpm) were subjected to SDS-PAGE, and the proportion of C subunit shifted to the slower-migrating, Thr-197-phosphorylated form was determined by densitometry of an autoradiogram of the dried gel. Early experiments (Fig. 3 and 5) used a slightly different buffer formulation that included sodium fluoride instead of potassium chloride, but the fluoride was found later to be somewhat inhibitory to the PKA kinase (6).

FIG. 3.

An extract of kinase-negative S49 cells can apparently phosphorylate recombinant C subunit. Soluble protein from kinase-negative S49 cells was mixed with [³⁵S]methionine-labeled recombinant C subunit to give about 3.0 (lanes a and f), 1.5 (lane b), 0.75 (lanes c and g), 0.38 (lane d), or 0.19 (lanes e and h) mg of extract protein per ml in 15 mM Tris-HCl (pH 7.5)–10 mM 2-mercaptoethanol–0.15 mM EDTA–11 mM magnesium sulfate–1 mM ATP–150 mM sodium fluoride–0.15 mg of bovine serum albumin per ml. Samples for lanes a to e were incubated for 1 h at 30°C before mixing with SDS-gel sample buffer, while those for lanes f to h were stopped immediately by addition of the SDS-containing buffer. Samples were analyzed as described in Materials and Methods, and positions of Thr-197-phosphorylated (C_P) and nonphosphorylated (C_N) forms of C subunit in the resulting autoradiographic patterns are indicated as for Fig. 1.

FIG. 5.

The PKA kinase activity shifts the SDS-PAGE mobility of wild-type and Lys-72→Met mutant C subunits but not that of a Thr-197→Ala mutant C subunit. [³⁵S]methionine-labeled wild-type (lanes a to c), Lys-72→Met (lanes d to f), or Thr-197→Ala (lanes g to h) C subunit were either incubated for 1 h at 30°C without PKA kinase (lanes a, d, and g) or mixed with a partially purified preparation of PKA kinase (first peak from Accell Plus QMA purified further by chromatography on ceramic hydroxylapatite and heparin-Sepharose) and incubated for 0 (lanes b, e, and h) or 1 (lanes c, f, and i) h at 30°C under standard conditions (Materials and Methods). Samples were analyzed by SDS-PAGE, and the positions of Thr-phosphorylated (C_P) and nonphosphorylated (C_N) forms of wild-type C subunit are indicated in autoradiographic patterns as in Fig. 1 and 3).

Extraction and partial purification of PKA kinase activity.

Frozen kinase-negative cells (see above) were extracted by thawing and centrifuged for 1 h at 100,000 × g. The supernatant fraction (about 100 mg of protein) was dialyzed overnight against QMA buffer (10 mM Bis-Tris propane [pH 7.0], 10 mM 2-mercaptoethanol, 0.1 mM EDTA) and loaded onto a 20-ml column of Accell Plus QMA (Millipore/Waters) in QMA buffer. After being washed with 200 ml of QMA buffer, the column was eluted with a 200-ml linear gradient from 0 to 300 mM potassium chloride in this buffer. Fractions were assayed for absorbance at 280 nm, conductivity, and kinase activity against [³⁵S]methionine-labeled C subunit (using 2 μl of fractions in 4-μl reactions). The activity peaks were either concentrated with a Centricon-10 concentrator (Amicon, Inc.) and stored frozen in small aliquots at −70°C or applied to 0.6-ml columns of hydroxyapatite (Bio-Gel HTP; Bio-Rad Laboratories), which were washed with QMA buffer and eluted with linear gradients from 0 to 500 mM in potassium phosphate (pH 6.5) before concentration further with Centricon-10 concentrators and freezing. The PKA kinase activity eluted from hydroxylapatite between about 0 and 200 mM potassium phosphate in a broad peak containing most of the applied protein (data not shown). Twofold dilutions of the partially purified preparations were assayed under standard conditions (see above) to determine the minimum amount that would give maximal phosphorylation. This amount was used in assays to characterize the kinase (e.g., Fig. 5 to 9). For the experiments represented in Fig. 5, 8, and 9, the PKA kinase was purified by several hundred-fold by optimizing the conditions for anion-exchange and hydroxylapatite chromatography (switching to macroprep ceramic hydroxyapatite [Bio-Rad Laboratories]) and subjecting the material to further purification by chromatography on heparin-Sepharose and Sephacryl S300HR (Pharmacia LKB Biotechnology) (6).

Phosphate labeling of C subunit in autophosphorylation and PKA kinase reactions, and phosphoamino acid analysis.

For autophosphorylation, bacterial extracts containing unlabeled, nonphosphorylated wild-type or mutant C subunits (see above) were diluted to about 0.15 mg/ml in buffer containing 10 mM Bis-Tris propane (pH 7.0), 10 mM 2-mercaptoethanol, 0.1 mM EDTA, 5.5 mM magnesium sulfate, and 0.5 mM ATP (with either 1.5 or 15 μCi of [γ-³²P]ATP per 10-μl reaction, respectively, for gel analysis or phosphoamino acid analysis) and incubated for 6 h at 30°C with 200 μg of purified, phosphorylated wild-type C subunit per ml. For labeling with partially purified PKA kinase, the bacterial extracts were diluted as described above but in buffer containing 10 mM Bis-Tris propane (pH 7.0), 10 mM 2-mercaptoethanol, 0.1 mM EDTA, 5.1 mM magnesium sulfate, 100 mM potassium chloride, 0.5 mg of bovine serum albumin per ml, and 0.1 mM ATP (with 0.5 or 5 μCi of [γ-³²P]ATP per 5-μl reaction, respectively, for gel analysis or phosphoamino acid analysis). Incubations (without or with partially purified PKA kinase or PKA kinase plus 100 μM H-89) were for 1 h at 30°C. For gel analysis, reactions were stopped by diluting with SDS-gel sample buffer. Autophosphorylation reaction mixtures were initially diluted 5-fold; 10 μl of 200-fold further dilutions were used for Western blot analysis (Fig. 8A), and 2-μl aliquots of the 5-fold-diluted samples were used for autoradiographic analysis (Fig. 8B). PKA kinase reaction mixtures were diluted 20-fold; 10-μl samples were used for Western blot analysis, and 20-μl samples were used for autoradiographic analysis. For phosphoamino acid analysis, samples were mixed with 5 μg of bovine serum albumin, brought up to 15 μl with water, and precipitated by adding 15 μl of trichloroacetic acid. The precipitates were collected by centrifugation, washed twice with 100-μl portions of 95% ethanol, allowed to air dry, and dissolved with 20 μl of SDS-gel sample buffer. After SDS-PAGE of the entire samples, the C-subunit bands were excised from dried gels by using a tracing of an autoradiogram as a guide and hydrolyzed for 5 h at 110°C with 2 N hydrochloric acid (33). Hydrolysates were dried, redissolved in a mixture of phosphoserine and phosphothreonine markers, and analyzed by thin-layer electrophoresis at pH 1.9 as described previously (33).

Affinity purification of active C subunit on PKIP-Sepharose columns.

The 20-amino-acid protein kinase inhibitor peptide (PKIP; Thr-Thr-Tyr-Ala-Asp-Phe-Ile-Ala-Ser-Gly-Arg-Thr-Gly-Arg-Arg-Asn-Ala-Ile-His-Asp) was coupled to epoxy-activated Sepharose 6B (Pharmacia LKB Biotechnology) to give a concentration of about 1.9 mg of peptide per ml of resin as described elsewhere (29). Columns (50 μl) of this material were poured, blocked with 5% nonfat dry milk, loaded, washed, and eluted at room temperature as described previously (29), with the following modifications: Bis-Tris propane (pH 7.0) was used in place of MES (pH 6.6) in all buffer solutions; column buffer was modified by increasing concentrations of magnesium sulfate and ATP to 5 mM and 100 μM, respectively; samples (∼21.5 μl containing 215,000 cpm of [³⁵S]methionine-labeled wild-type C subunit or 430,000 cpm [³⁵S]methionine-labeled Met-72 mutant C subunit) in 10 mM Bis-Tris propane (pH 7.0)–10 mM 2-mercaptoethanol–0.1 mM EDTA–5 mM magnesium sulfate–100 μM ATP were rinsed into columns and then incubated for 1 h; columns were washed 12 times with 100 μl of column buffer and then 6 times with column buffer lacking sodium chloride, methionine, and bovine serum albumin; and columns were eluted with 150 μl of arginine buffer (10 mM Bis-Tris propane, 10 mM 2-mercaptoethanol, 0.1 mM EDTA, 200 mM l-arginine [pH ∼7.5]). The eluates were precipitated by addition of 50 μg of bovine serum albumin and 10 μl of 100% (wt/vol) trichloroacetic acid and incubation for 30 min on ice. Precipitates were collected by centrifugation, washed once with 100 μl of ice-cold 95% ethanol, air dried, and dissolved in SDS-gel sample buffer for SDS-PAGE analysis. (Supernatant fractions were aspirated.) Coomassie blue staining of the bovine serum albumin carrier protein confirmed that protein recoveries were similar for all samples (not shown).

RESULTS

The autophosphorylation of C subunit at Thr-197 in overexpressing bacteria is inefficient.

In a previous study, we showed that accumulation of Thr-197-phosphorylated C subunit lags behind expression of C-subunit protein in bacteria induced to express C subunit at 24 to 25°C (33). The rate of phosphorylation of newly synthesized C subunit was faster late in induction, when there was more active C subunit, suggesting that the phosphorylation was an intermolecular autophosphorylation. Consistent with this interpretation, mutations that inactivated C subunit also prevented phosphorylation of bacterially expressed C subunit (38). Nonphosphorylated C subunit could be phosphorylated in vitro on Thr-197 by active C subunit in an intermolecular reaction, but even with concentrations of C subunit activity hundreds of times higher than found in mammalian cells, the half-time for the reaction was greater than an hour at 30°C (33). The experiment represented in Fig. 1 was designed to provide further perspective on the rate of the autophosphorylation reaction under physiological conditions. To mimic the C-subunit modification observed in mammalian cells, the C subunit was coexpressed in bacteria with yeast N-myristoyltransferase; such coexpression results in essentially complete myristoylation of C subunit (9). The bacteria were preinduced at 24°C to accumulate various levels of intracellular kinase activity, treated with rifampin to inhibit synthesis of most bacterial proteins, and then pulse-labeled with [³⁵S]methionine at 37°C. The relative phosphorylation of the newly synthesized C subunit at Thr-197 was estimated after SDS-PAGE as the proportion of C-subunit radioactivity in the slower-migrating, Thr-197-phosphorylated form. Based on assays of parallel cultures incubated without [³⁵S]methionine, the samples for Fig. 1, lanes a, c, and e, had PKA activity levels of about 11, 16, and 26 U per mg of protein, respectively, at the time of labeling. The proportion of C subunit label in the Thr-197-phosphorylated form varied from about 31% in the sample with lowest activity to 59% in that with the highest. Lanes b, d, and f, in Fig. 1 show that regardless of kinase activity level, the large mobility shift resulting from Thr-197 phosphorylation was inhibited fully by treatment with H-89, a selective inhibitor of PKA activity (7). The splitting of the nonphosphorylated C-subunit band into a closely spaced doublet in samples labeled in the presence of H-89 was observed reproducibly but only in cells expressing myristoyltransferase (6). Its underlying cause is unknown.

Phosphorylation of C subunit in intact S49 mouse lymphoma cells is efficient and independent of endogenous C-subunit activity.

Figure 2 and Table 1 present results from [³⁵S]methionine pulse-labeling experiments assessing the phosphorylation of newly synthesized C subunits in intact cyclase-negative S49 cells. Extracts of these cells had kinase specific activities of about 0.36 U per mg of protein when assayed without cAMP and 4.6 U per mg of protein when assayed with saturating cAMP. These cells were chosen over wild-type cells, because they have lower than normal basal activity of C subunit as judged by two-dimensional gel analysis of endogenous C-subunit substrates (31). For the experiment of Fig. 2, lanes a to d, cells were pulse-labeled for 10 min and either harvested immediately or chased for an additional 30 min. C subunits were immunoadsorbed, and the various C subunit forms were resolved by SDS-PAGE. Because both Cα and Cβ isoforms of C subunit are expressed in S49 cells and Cα migrates faster in SDS-PAGE than does Cβ, these patterns were more complex than those of the recombinant Cα subunit (23, 29). In the pulse-labeled samples, there was a small amount of label in the fast-migrating band that corresponds to the Thr-197-nonphosphorylated form of Cα subunit (29, 33). This species was not detected after the 30-min chase, suggesting that phosphorylation of Thr-197 was complete. Figure 2, lane e, shows a pulse-labeled sample that was extracted under conditions that promote C-subunit dephosphorylation by an endogenous protein phosphatase (19, 20). Densitometry of the patterns from both the chase samples (Fig. 2, lanes c and d) and the dephosphorylated sample (Fig. 2, lane e) revealed an about 4-to-1 ratio of Cα to Cβ subunit labeling. From this ratio and the proportion of total C-subunit label in the fastest-migrating, Thr-197-nonphosphorylated Cα subunit, it was possible to estimate the fraction of Cα subunit phosphorylated in pulse-labeled samples. This varied from about 70 to 90% in three experiments including that shown in Fig. 2 (Table 1). Inclusion of CPT-cAMP at a concentration sufficient to maximally activate endogenous cAMP-dependent protein kinase inhibited marginally the Thr-197 phosphorylation of newly synthesized C subunit and shifted the Thr-197-phosphorylated Cα subunit to slightly lower mobility (Fig. 2 and data not shown). This small shift was consistent with further phosphorylation on one of the three Ser residues known to be sites for C-subunit autophosphorylation (6, 38). Table 1 also shows that, in contrast to its effect on bacterially expressed C subunit, H-89 had no effect on Thr-197 phosphorylation of C subunit in S49 cells whether or not CPT-cAMP was present. (Two-dimensional gel analysis of replicate samples from experiment 2 confirmed that H-89 had inhibited endogenous cAMP-dependent phosphorylations known to be dependent on C-subunit activity [as described in reference 30] [data not shown]).

TABLE 1.

Effect of CPT-cAMP and/or H-89 on phosphorylation of newly synthesized Cα subunit in cyclase-negative S49 cells^a

Expt	Phosphorylated Cαb
	−Drugs	+CPT-cAMP	+H-89	+CPT-cAMP and H-89
1	0.91	0.81	ND	ND
2	0.69	0.60	0.70	0.59
3	0.74	0.70	0.73	∼0.7c

^aCells were preincubated without or with drugs as indicated for 5 min and then pulse-labeled with [³⁵S]methionine for 10 (experiments 1 and 3) or 15 min (experiment 2) as described in Materials and Methods. After immunoadsorption, SDS-PAGE, and fluorography as for Fig. 2, the proportion of total C-subunit radioactivity in the fastest-migrating form (the nonphosphorylated form of Cα) was determined by densitometry. This value was divided by the fraction of total C-subunit radioactivity in the Cα isoform (determined from dephosphorylated samples such as that shown in Fig. 2, lane e) to calculate the proportion of Cα subunit in the nonphosphorylated form. 

^bProportion of Cα subunit in the phosphorylated form (calculated by subtracting the nonphosphorylated fraction from 1). ND, not done. 

^cBecause of poor incorporation in this experiment and the combined inhibitory effects of the two drugs on incorporation, this pattern was too faint for accurate quantitation. 

Evidence for a PKA kinase in S49 cells.

The results described above suggested that C subunit itself was not the catalyst for Thr-197 phosphorylation of newly synthesized C subunit in intact cells. To test whether S49 cells contained another activity capable of phosphorylating C subunit on Thr-197, we used extracts of an S49 cell mutant that lacks functional C-subunit protein (32). For a substrate we used recombinant, wild-type C subunit labeled with [³⁵S]methionine in bacteria that had been treated with H-89 at the time of induction to prevent any accumulation of phosphorylated C subunit (labeled or unlabeled). This labeled substrate behaved identically with unlabeled, nonphosphorylated C subunit in autophosphorylation reactions (6). Figure 3 shows that an activity in crude extracts of kinase-negative S49 cells could shift the electrophoretic migration of a portion of this labeled C subunit protein to that characteristic of the Thr-197-phosphorylated form. The shift activity was sensitive to dilution of the extract (Fig. 3, lanes a to e). Labeled C subunit in control samples in which substrate and extract were mixed but not incubated retained the high mobility characteristic of the Thr-197-nonphosphorylated protein (Fig. 3, lanes f to h).

Figure 4 shows the result of fractionating a kinase-negative cell extract on an anion-exchange column. Two peaks of the C subunit shift—or PKA kinase—activity were resolved from the bulk of cell protein. A major peak of activity eluted at about 90 mM potassium chloride, and a smaller peak eluted at about 150 mM potassium chloride. The relative sizes of the two activity peaks were unaffected by treatment of cell extracts with a cocktail of protease inhibitors, suggesting that the peaks might represent distinct protein species, but they appeared to behave identically in their reactions on C subunit (6). Wild-type S49 cell extracts gave amounts of PKA kinase activity and chromatographic patterns similar to those of the kinase-negative cell extracts (6). In experiments using 5-min incubations at various temperatures, the crude or partially purified enzyme was stable up to 40°C, completely inactivated at 55°C, and inactivated to intermediate extents at 45 or 50°C (6).

FIG. 4.

Partial purification of PKA kinase activity by anion-exchange chromatography. An extract of kinase-negative cells was fractionated by chromatography on Accell Plus QMA, and fractions were assayed for absorbance at 280 nm (solid line), conductivity (dotted line), and PKA kinase activity (•) as described in Materials and Methods. PKA kinase activity is expressed as the percentage of total C-subunit radioactivity in the phosphorylated form after incubation under standard conditions (Materials and Methods).

The results in Fig. 5 and 6 provided evidence that the C-subunit mobility shift catalyzed by partially purified PKA kinase fractions indeed resulted from phosphorylation of Thr-197. The experiment of Fig. 5 compared the substrate activity of wild-type C subunit with those of mutant subunits with Met substituted for Lys-72, a critical residue in the ATP-binding site, or Ala substituted for Thr-197. For all three preparations, control samples that were not incubated or that were incubated without added PKA kinase gave single bands of C subunit that migrated with the nonphosphorylated form of the wild-type protein (Fig. 5, lanes a, b, d, e, g, and h). The mobilities of both wild-type and Met-72 mutant C subunits were shifted by incubation with the PKA kinase (Fig. 5, lanes c and f), but the Ala-197 C subunit was unaffected (Fig. 5, lane i). The PKA kinase-shifted band of the wild-type C subunit comigrated with autophosphorylated wild-type C subunit, and there was no further shift observed after incubation of the autophosphorylated protein with PKA kinase (data not shown). The smaller apparent shift of the Met-72 C subunit compared with the wild-type protein was also observed with protein phosphorylated by C subunit itself (6) and presumably reflects a difference in conformation, SDS binding, or Ser phosphorylation of the mutant protein (see also Fig. 8 and 9).

FIG. 6.

The PKA kinase activity has an apparent K_m for ATP of about 12 μM. The ATP dependence of partially purified PKA kinase (first peak) was assessed under the standard assay conditions described in Materials and Methods but with 5 mM free magnesium sulfate and various concentrations of an equimolar mixture of ATP and magnesium sulfate. Data shown are the results from densitometric analysis of SDS-PAGE patterns and are expressed as the percentage of total C-subunit radioactivity in the phosphorylated form (C_P) as for Fig. 4.

Figure 6 shows that the gel shift activity was dependent on ATP as a phosphate donor with an apparent K_m of about 12 μM. The activity was unaffected by the chelator EDTA or EGTA (at 2 mM), calcium (1 mM) with or without calmodulin (at 1 μM), or the nucleotide cAMP, cGMP, or 5′-AMP at 0.1 mM (6). Furthermore, the activity was effective on the myristoylated as well as on the nonmyristoylated C-subunit protein (6). Figure 7 shows kinetics of the PKA kinase reaction at 30 and 37°C. The initial reaction was faster at 37 than at 30°C and led to phosphorylation of more than 60% of the labeled C subunit within 15 min. At both temperatures, the reaction appeared to continue slowly for at least 30 min. Incubation for up to 3 h caused only slightly more phosphorylation (data not shown). The fraction of C subunit ultimately phosphorylated appeared to be limited not by the enzyme but rather by some as yet undefined property of the substrate that varied somewhat between preparations: preincubation of the enzyme without substrate did not diminish its activity; and addition of fresh enzyme to a complete reaction after 30 min of incubation did not stimulate significant further reaction (6).

FIG. 7.

The initial phase of C-subunit phosphorylation by the PKA kinase is quite rapid. Partially purified PKA kinase (first peak) was mixed with [³⁵S]methionine-labeled C subunit in standard assay buffer (Materials and Methods) and incubated for various times at 30 (•) or 37°C (○) before analysis of C-subunit phosphorylation as for Fig. 6.

A purified preparation enriched for the Thr-197 nonphosphorylated form of C subunit from bacteria induced for 50 min in the absence of H-89 (33) neither was phosphorylated by the PKA kinase activity nor inhibited phosphorylation of the labeled C subunit (6). Furthermore, even crude, labeled preparations of the Thr-197-nonphosphorylated wild-type protein prepared in the absence of H-89 were poor PKA kinase substrates (6). On the other hand, a purified preparation of the Met-72 mutant C subunit could be phosphorylated by the PKA kinase (6). These observations suggest the possibility, now under investigation, that Ser phosphorylation in the wild-type C subunit inhibits its phosphorylation by the PKA kinase.

The results in Fig. 5 left open the possibility that the Ala-197 C subunit was phosphorylated by the PKA kinase but did not undergo a detectable mobility shift. The results in Fig. 8 and 9 ruled out this possibility and provided evidence that the PKA kinase is specific for phosphorylation of C subunit on Thr-197. Figure 8 shows results from an experiment in which crude, unlabeled preparations of nonphosphorylated wild-type or mutant C subunits were incubated with either purified active C subunit or PKA kinase in the presence of [γ-³²P]ATP. Figure 8A shows Western immunoblots of the samples to verify that C subunits were present in comparable amounts in the crude bacterial extracts. (Because of the high concentrations of wild-type C subunits added for the C-subunit-catalyzed reactions [Fig. 8A, lanes a, e, and i], these samples were diluted back for Western blots and show only the added, fully phosphorylated, wild-type C subunit.) Although not so clearly as for the labeled samples of Fig. 5, PKA kinase-dependent shifts of the wild-type and Met-72 C subunits could be observed (Fig. 8A, lanes c, d, g, and h). Figure 8B shows the patterns of phosphate labeling from these same reactions. C-subunit-dependent labeling of C subunit in these samples was distributed between both slow- and fast-migrating forms of the protein (Fig. 8B, lanes a, e, and i), and the added C subunit also stimulated incorporation into several bacterial proteins in the extracts (not shown). No C-subunit labeling was observed in the absence of added C subunit or PKA kinase (Fig. 8B, lanes b, f, and j). Incubation with PKA kinase resulted in phosphate labeling of both wild-type and Met-72 C subunits (Fig. 8B, lanes c and g) but not of Ala-197 C subunit (Fig. 8B, lane k). Although H-89 inhibited somewhat the labeling of wild-type C subunit, it had no apparent effect on that of the Met-72 C subunit (Fig. 8B, lanes d and h). In contrast to C subunit, the PKA kinase preparation did not stimulate labeling of any bacterial proteins in the extracts (not shown).

Figure 9 shows phosphoamino acid analysis of C subunits labeled in reactions with either C subunit or PKA kinase as the catalyst. Hydrolysis of wild-type C subunit labeled by incubation with active C subunit released both labeled phosphothreonine and phosphoserine, with the majority of the label in phosphoserine (Fig. 9, lane a). The wild-type C subunit incubated with PKA kinase contained about equal amounts of labeled phosphothreonine and phosphoserine (Fig. 9, lane b), but the phosphoserine labeling was completely suppressed by including H-89 in the reaction (Fig. 9, lane c). Met-72 C subunits were labeled only on threonine by incubation with the PKA kinase in the presence or absence of H-89 (Fig. 9, lanes d and e).

It has not yet been possible to show that Thr-197 phosphorylation of wild-type C subunit by the PKA kinase increases the enzymatic activity of the protein, because the purified Thr-197 phosphorylated protein is not phosphorylated by the kinase kinase and residual H-89 and/or other substances in the crude bacterial extracts interfere with the phosphotransferase reaction (6). Figure 10 shows the results of an experiment using a surrogate assay—binding of C subunit to a PKIP-Sepharose affinity resin (29)—to demonstrate that PKA kinase not only phosphorylates the wild-type C subunit but also activates it. Preliminary experiments with [³⁵S]methionine-labeled, autophosphorylated C subunit showed that pretreatment of the protein with H-89 did not prevent its binding to the affinity matrix (data not shown). Samples of [³⁵S]methionine-labeled wild-type or Met-72 mutant C subunits were incubated without or with a partially purified preparation of PKA kinase and then loaded onto PKIP affinity columns. After incubation and extensive washing, the bound fractions were eluted with buffer containing 200 mM arginine. Gel patterns of the material loaded onto the columns show that both the wild-type and mutant C subunits were phosphorylated by incubation with the PKA kinase preparation, the wild type somewhat more efficiently than the mutant (Fig. 10, lanes a, b, e, and f). Only the slower-migrating, Thr-197-phosphorylated form of the wild-type C subunit incubated with PKA kinase was bound to the affinity column (Fig. 10, lane d); the faster-migrating form of the wild-type C subunit in samples incubated without or with the PKA kinase did not bind (Fig. 10, lanes c and d). Consistent with the inactive phenotype of the Met-72 mutant C subunit, neither its faster- nor its slower-migrating form bound effectively to the column.

FIG. 10.

PKA kinase-mediated phosphorylation of wild-type, but not of Met-72 mutant, C subunit activates the protein for binding to a pseudosubstrate inhibitor peptide. [³⁵S]methionine-labeled preparations of wild-type (lanes a to d) or Met-72 mutant (lanes e to h) C subunit were incubated for 1 h at 30°C without (lanes a, c, e, and g) or with (lanes b, d, f, and h) a partially purified preparation of PKA kinase under standard conditions but with only 100 μM ATP. Samples were either diluted immediately with SDS-gel sample buffer for SDS-PAGE analysis (lanes a, b, e, and f) or purified on columns of PKIP-Sepharose as described in Materials and Methods (lanes c, d, g, and h). For the samples taken before affinity column purification, equal amounts of protein radioactivity (∼10,000 cpm) were loaded onto the SDS-polyacrylamide gel. For the samples bound to—and subsequently eluted from—the affinity columns, equal proportions (∼23%) were subjected to gel analysis, although twice as much mutant protein radioactivity had been loaded onto the columns (Materials and Methods). Portions of autoradiographic patterns containing C subunits are shown, and positions of the Thr-197-phosphorylated (C_P) and nonphosphorylated (C_N) forms of wild-type C subunit are indicated as in Fig. 1, 3, 5, and 8).

DISCUSSION

Our results with intact S49 cells demonstrate that Thr-197 phosphorylation of newly synthesized C subunit is not the result of autophosphorylation: the reaction was rapid despite the very low basal activity of PKA in unstimulated, adenylate cyclase-deficient cells; the reaction was not accelerated by stimulating cells with CPT-cAMP and thereby activating fully the endogenous C subunit; and the reaction was resistant to the PKA-selective inhibitor, H-89. In contrast, Thr-197 phosphorylation of newly synthesized C subunit in bacteria was sensitive to H-89 and dependent on preexisting PKA activity. Even with C-subunit activity levels more than fivefold higher than those of cAMP-stimulated mammalian cells, the phosphorylation of newly synthesized C subunits in bacteria was inefficient.

Using [³⁵S]methionine-labeled recombinant C subunit synthesized in the presence of H-89 as a substrate, we found an activity in S49 cell extracts that could phosphorylate C subunit on Thr-197 as detected by the characteristic shift to lower mobility in SDS-PAGE. The reaction required incubation in the presence of ATP and was eliminated by a mutation that substituted Ala for Thr-197. The latter observation seems not to be attributable simply to improper folding of inactive, nonphosphorylated mutant subunits (e.g., as suggested by Yonemoto et al. [38]): bacterially expressed Ala-197 enzyme has activity similar to that of nonphosphorylated wild-type C subunit (33); and, although a Lys-72→Met mutation likewise prevented autophosphorylation during bacterial expression, it did not prevent phosphorylation by the PKA kinase. Furthermore, although the PKA kinase could have acted as a cofactor for autophosphorylation of wild-type C subunit, its ability to phosphorylate the inactive Met-72 C subunit argues strongly that its role is catalytic.

Since the PKA kinase activity was extracted from an S49 cell mutant that lacks detectable C-subunit activity (32), it appeared unlikely that the activity could be that of C subunit itself. This was amply confirmed by the properties of the activity revealed by the results in Fig. 4, 6, and 7. C subunit elutes in the flowthrough of Accell Plus QMA columns (33), and the partially purified PKA kinase fractions had only a weak, high-K_m phosphorylating activity on the C-subunit substrate, kemptide (6). Where the PKA kinase reaction had an apparent K_m for ATP of about 12 μM (Fig. 6), the autophosphorylation reaction had an apparent ATP K_m of several hundred micromolar (6) (perhaps artificially high because of the intrinsic ATPase activity of C subunit and the high concentrations of C subunit required to promote reasonable rates of autophosphorylation). While with even high concentrations of C subunit and ATP the autophosphorylation reaction proceeded over a time course of 4 to 6 h at 30°C (6, 33), the PKA kinase reaction was virtually complete within 15 to 30 min. Furthermore, although the kemptide phosphorylation and autophosphorylation activities of C subunit were inhibited completely by 100 μM H-89, the PKA kinase reaction was inhibited by less than 50% by this concentration of the inhibitor (6) (Fig. 8).

Although purified wild-type C subunit could not be phosphorylated with the PKA kinase, crude unlabeled preparations of nonphosphorylated wild-type or Met-72 C subunits could be phosphorylated and concomitantly labeled with [³²P]phosphate from ATP. In contrast to active C subunit, which promoted labeling of numerous proteins in the bacterial extracts, the PKA kinase appeared to stimulate labeling only of C subunit itself. Furthermore, the C subunits labeled by the PKA kinase migrated as single bands in SDS-PAGE. This result suggested specificity not only for the C subunit but also for Thr-197 over Ser phosphorylation sites within C subunit. This latter specificity was verified by phosphoamino acid analysis, which revealed only labeled phosphothreonine in Met-72 C subunits phosphorylated by the PKA kinase and in wild-type C subunits phosphorylated by PKA kinase in the presence of H-89. In contrast, the labeling catalyzed by added C subunit was mostly of phosphoserine (Fig. 9, lane a); the diffuse labeling extending through the positions of both slow and fast forms of C subunit in the autophosphorylation reactions of Fig. 8 (lanes a, e, and i) can be explained by Ser phosphorylation both of the faster-migrating substrate C subunits and of the active, slower-migrating, catalytic C subunit at multiple sites (38). The moderate labeling of phosphoserine in wild-type C subunits incubated with PKA kinase in the absence of H-89 (Fig. 9, lane b) probably reflects autophosphorylation by C subunits activated in the PKA kinase reaction. Activation of wild-type C subunit by the PKA kinase was confirmed by the retention of its slow-migrating, Thr-197-phosphorylated form on an affinity resin containing a covalently linked pseudosubstrate inhibitor peptide (Fig. 10). The Thr-197-phosphorylated form of the inactive Met-72 mutant subunit did not bind to the affinity resin.

Figure 11 shows subdomain VIII sequences from several protein kinases with strong sequence homology in this region to PKA that require phosphorylation of a Thr residue homologous to Thr-197 in PKA for maximal activity (11, 14, 16, 24, 25). For protein kinase Cβ (PKCβ), calcium/calmodulin-dependent protein kinases (CaMKs) I and IV, and AMP-activated protein kinase (AMPK), the phosphorylation has been shown to be by heterologous protein kinases (14, 16, 24, 25), where for the cGMP-dependent protein kinase (CGK) the mode of phosphorylation is unknown (11). Eight of the 11 residues immediately downstream of the target Thr are identical in PKA, CGK, PKCβ, CaMK I, and CaMK IV, and 7 of these conserved residues are also conserved in AMPK (Fig. 11). This finding suggests that these kinases could be phosphorylated by related protein kinase kinases, using the conserved downstream motif in part for recognition. Two CaMK kinases and an AMPK kinase have been purified and characterized, and the gene for one of the CaMK kinases has been cloned and expressed (10, 16, 25, 36). Phosphorylation of CaMK by the CaMK kinases is stimulated by calcium and calmodulin acting allosterically on both the CaMK substrate and the CaMK kinase itself (10, 14). In similar ways, the AMPK kinase reaction is stimulated by 5′-AMP (15, 16). The CaMK and AMPK kinases appear to be related, since at least one of the CaMK kinases can activate AMPK, and the AMPK kinase can activate CaMK I in reactions stimulated by the combination of calcium, calmodulin, and 5′-AMP (15). We speculate from the homologies of subdomain VIII target sequences that our PKA kinase belongs to a new family of protein kinases that includes these activators of CaMK and AMPK as well as activators for PKCβ and CGK. Nevertheless, the PKA kinase appears to be distinct from the CaMK and AMPK activators in that its reaction was not stimulated by low-molecular-weight ligands including calcium and calmodulin or 5′-AMP.

FIG. 11.

A number of protein kinases related to PKA have targets for Thr phosphorylation in the activation loop region in subdomain VIII that are followed by highly conserved sequences. Subdomain VIII sequences and the residues immediately upstream are shown for PKA, CGK, PKCβ, CAMK I and IV, and AMPK as described by Hanks et al. (13), using additional sequence data from references 5 and 15. Thr-197 of PKA C subunit is indicated with an asterisk, and residues conserved among PKA, CGK, PKCβ, CAMK I, and CAMK IV are shown in boldface.

Combined with our recent demonstration that PKA can be dephosphorylated by protein phosphatase 2A under near-physiological conditions (20), the present evidence for a PKA kinase responsible for physiological phosphorylation of PKA raises the possibility of regulation of PKA activity by phosphorylation/dephosphorylation. PKA is regulated acutely by intracellular levels of cAMP, but control of Thr-197 phosphorylation might serve to modulate a cell’s responsiveness to cAMP-dependent regulation. Because regulatory subunit does not bind with high affinity to the nonphosphorylated C subunit (1), the PKA holoenzyme should contain only Thr-197-phosphorylated C subunit. Also, by docking over the Thr-197 region of C subunit (1), regulatory subunit probably blocks access of protein phosphatase to the phosphate on Thr-197. cAMP-induced dissociation of the complex should increase the susceptibility of C subunit to phosphatase-mediated dephosphorylation, so that rephosphorylation by the PKA kinase would be necessary to preserve a high stoichiometry of Thr-197 phosphorylation during extended periods of kinase activation. If the PKA kinase activity were insufficient to keep up with the phosphatase activity and/or Ser phosphorylation inhibited rephosphorylation of the protein, the result would be net dephosphorylation of C subunit. In an ongoing series of experiments, we have used Western immunoblot analysis to investigate the steady-state level of C-subunit phosphorylation in a number of cell lines treated with a variety of effectors including insulin, phorbol ester, retinoic acid, and CPT-cAMP, but we have not yet observed any significant effector-dependent change in C-subunit phosphorylation (2).