Dissociation and activation of adenosine 3',5'-monophosphate-dependent and guanosine 3',5'-monophosphate-dependent protein kinases by cyclic nucleotides and by substrate proteins.

Abstract Two adenosine 3',5'-monophosphate (cyclic AMP)-dependent protein kinases, with sedimentation coefficients of 5.4 S (mol wt 84,000) and 7.7 S (mol wt 140,000), respectively, have been purified from bovine brain. Each enzyme migrated as a single band at three different pH values on polyacrylamide disc gel electrophoresis. The enzymes had similar properties. Catalytic subunits of cyclic AMP-dependent protein kinases from bovine brain, and of cyclic AMP-dependent and guanosine 3',5'-monophosphate (cyclic GMP)-dependent protein kinases from lobster muscle were prepared by means of Enzite CM-cellulose-protamine affinity chromatography, and their properties were studied. The catalytic subunits of the two bovine brain enzymes and of the lobster muscle cyclic GMP-dependent enzyme each had a sedimentation coefficient of 3.6 S (mol wt 40,000), whereas that of the lobster muscle cyclic AMP-dependent enzyme sedimented at 4.5 S (mol wt 60,000). These results agree with kinetic studies, indicating that cyclic GMP-dependent protein kinases of arthropods have certain similarities to cyclic AMP-dependent protein kinases of vertebrates. Histone and cyclic AMP caused the dissociation of the cyclic AMP-dependent protein kinase from lobster muscle (5.7 S; mol wt 90,000) and those from bovine brain into subunits; histone and cyclic GMP caused the dissociation of the lobster muscle cyclic GMP-dependent protein kinase (7.7 S; mol wt 140,000) into subunits. The dissociation of the enzymes by histone was accompanied by a concomitant increase in cyclic nucleotide-independent activity. Iso-electric points of catalytic and regulatory subunits of the 7.7 S component of the brain enzyme were about pH 7.8 and 3.8, respectively. Addition of isolated regulatory subunit, derived from bovine brain cyclic AMP-dependent enzyme, inhibited the activity of catalytic subunits obtained from bovine brain enzymes. Interestingly, the regulatory subunit of the cyclic AMP-dependent enzyme from mammalian brain also inhibited the activity of the catalytic subunit prepared from the lobster muscle cyclic GMP-dependent enzyme, converting it into a cyclic AMP-dependent form. Thus, an interaction between catalytic and regulatory subunits occurred, although the subunits were from different tissues, different phyla, and different classes (with respect to cyclic nucleotide specificity) of protein kinase.

Each enzyme migrated as a single band at three different pH values on polyacrylamide disc gel electrophoresis. The enzymes had similar properties.
Catalytic subunits of cyclic AMP-dependent protein kinases from bovine brain, and of cyclic AMP-dependent and guanosine 3',5'-monophosphate (cyclic GMP)-dependent protein kinases from lobster muscle were prepared by means of Enzite CM-cellulose-protamine affinity chromatography, and their properties were studied.
The catalytic subunits of the two bovine brain enzymes and of the lobster muscle cyclic GMP-dependent enzyme each had a sedimentation coefficient of 3.6 S (mol wt 40,000), whereas that of the lobster muscle cyclic AMP-dependent enzyme sedimented at 4.5 S (mol wt 60,000).
These results agree with kinetic studies, indicating that cyclic GMP-dependent protein kinases of arthropods have certain similarities to cyclic AMP-dependent protein kinases of vertebrates.
Histone and cyclic AMP caused the dissociation of the cyclic AMP-dependent protein kinase from lobster muscle (5.7 S; mol wt 90,000) and those from bovine brain into subunits; histone and cyclic GMP caused the dissociation of the lobster muscle cyclic GMP-dependent protein kinase (7.7 S; mol wt 140,000) into subunits.
The dissociation of the enzymes by histone was accompanied by a concomitant increase in cyclic nucleotide-independent activity. Isoelectric points of catalytic and regulatory subunits of the 7.7 S component of the brain enzyme were about pH 7.8 and 3.8, respectively. * This work was supported by Grants NS-08440, MH-17387, and HL-13305

Addition
of isolated regulatory subunit, derived from bovine brain cyclic AMP-dependent enzyme, inhibited the activity of catalytic subunits obtained from bovine brain enzymes.
Interestingly, the regulatory subunit of the cyclic AMP-dependent enzyme from mammalian brain also inhibited the activity of the catalytic subunit prepared from the lobster muscle cyclic GMP-dependent enzyme, converting it into a cyclic AMP-dependent form.
Thus, an interaction between catalytic and regulatory subunits occurred, although the subunits were from different tissues, different phyla, and different classes (with respect to cyclic nucleotide specificity) of protein kinase.
Subsequent to the discovery of adenosine 3',5'-monophosphate (cyclic AMP)-dependent protein kinase in skeletal muscle (1) and then in liver (2), brain (3), and a wide variety of other sources (4,5), it was proposed (4,5) that the diverse actions of cyclic AMP' are mediated through regulation of this class of enzymes, Much recent evidence, obtained in numerous laboratories, is at least consistent with and, in many cases, supports this postulate for the mechanism of action of cyclic AMP in eukaryotic organisms.
The discovery (6, 7) of a class of protein kinases activated specifically by guanosine 3') 5'-monophosphate (cyclic GMP), rather than by cyclic AMP, has raised the possibility that the effects of cyclic GMP may be mediated through this class of enzymes. However, in contrast to the case of the cyclic AMP-dependent protein kinases, no further studies have been reported on the enzymological properties of cyclic GMPdependent protein kinases. The fact that cyclic AMP-and cyclic GMP-dependent protein kinases are found in the same tissues raises a number of questions concerning the biochemical and physiological relationship of these two types of enzyme. In view of the importance of these two classes of protein kinase, it seemed desirable to study and compare their enzymological properties.
In the present investigation, we have compared some of the properties of the holoenzyme and catalytic subunit of cyclic GMP-dependent protein kinase from lobster muscle, cyclic AMPdependent protein kinase from lobster muscle, and cyclic AMPdependent protein kinases from mammalian brain.
Brain was selected as the source of the mammalian enzymes, since recent evidence suggests that cyclic AMP-dependent protein kinase may play an important role in the functioning of the mammalian nervous system (S-10).
The present report includes a description of the preparation of two purified enzymes from mammalian brain.
Studies in several laboratories, using a variety of tissues (ll-17), have provided evidence that the mechanism by which cyclic AMP activates cyclic AMP-dependent protein kinases involves dissociation of the holoenzyme into a regulatory subunit, which binds cyclic AMP and is inhibitory to the enzyme, and a catalytic subunit.
In the present study, it has been found that cyclic GMP-dependent protein kinase can be dissociated and activated by cyclic GMP.
In a preliminary study (16), certain substrate proteins caused the partial dissociation and activation of cyclic AMP-dependent enzymes. In the present study, both cyclic GMP-dependent and cyclic AMP-dependent protein kinases were completely dissociated by certain proteins capable of acting as substrates.
The recombination of subunits has also been studied, including the formation of a "hybrid" enzyme composed of the regulatory subunit of a cyclic AMPdependent protein kinase and the catalytic subunit of a cyclic GMP-dependent protein kinase.

Materials
Frozen bovine brains were obtained from Pel-Freez. They were stored at -20' until the enzyme purification was started. Histone mixture (calf thymus) obtained from Schwarz-Mann was the histone preparation used throughout this study, unless otherwise specified.
One unit of enzyme activity was defined as that amount of enzyme which transferred 1 pmole of azP from [y-32P]ATP to recovered protein in 5 min at 30" in the standard assay system.
Analytical Gel Electrophoresis-Analytical gel electrophoresis was performed at pH 8.1, pH 9.5, and pH 12.0 in accordance with the procedures of Ornstein (19) and Davis (Xl), with a Canalco model 66 electrophoresis apparatus without stacking gel. Gels were stained with aniline blue-black and destained electrophoretically.
Assay for Cyclic AMP-binding Activity-The cyclic AXPbinding activity of protein was determined by the method of Gilman (21), except for omission of the heat-stable inhibitor. The incubation mixture contained, in a final volume of 0.2 ml, 10 pmoles of sodium acetate buffer (pH 4.0) and 3 pmolcs of cyclic [G-3H]AMP (16.3 Ci per mmole). The reaction was initiated by the addition of binding protein.
Sucrose Density Gradient Centrifugation-Ultracentrifugation of protein kinases was carried out at 37,500 rpm in a Beckman SW 39 L rotor for 16 hours at 3". In addition to a 5 to 200/, sucrose gradient, the centrifuge tube contained 50 mM acetate buffer (pH 6.0), 0.3 mM EGTA, and 2.5 rnr+f 2mercaptoethanol. At the termination of the centrifugation, the bottom of the tube was punctured and 7-drop fractions, containing about 0.14 ml, were collected. Sedimentation coefficients and molecular weights were determined by the method of Martin and Ames (22), with catalase (11.6 S, mol wt 232,000) (23) which was assayed by the method of Beers and Sizer (24), glyceraldehyde 3-phosphate dehydrogenase (7.7 S, mol wt 140,000) (25) which was assayed by the method of Velick (26)) and horse liver alcohol dehydrogenase (5.4 S, mol wt 84,000) (27) which was assayed by the method of Vallee and Hoch (28), as internal markers in each experiment.
Isoelectrk Focusing-Isoelectric focusing was carried out by a slight modification of the procedure described previously (29), by using a 1 lo-ml column (LKB Instrument).
The amount of carrier ampholyte used, with a pH range of 3 to 10, was 2% (w/v) in a 0 to 477, (w/v) sucrose gradient.
Electrofocusing was initiated at 400 volts; as the current decreased, the voltage was increased to and maintained at 800 volts. The duration of electrofocusing was 24 hours, and the current was constant (0.9 ma) for the last 4 hours. Upon completion of the isoelectric focusing, 3-ml fractions were collected.
Other XethodsCyclic GMP-dependent and cyclic AMPdependent protein kinases from lobster muscle were prepared by a method described earlier (6). Enzit'e CM-cellulose-protamine was prepared by the method of Mitz and Summaria (30).
[y-32P]ATP was prepared by the procedure of Post and Sen (31 Preparation of Brain Holoenzyme-Four frozen bovine brains, Assay for Protein K&se-The activity of cyclic AMP-with a combined weight of 1,450 g, were thawed, cut into small dependent and cyclic GMP-dependent protein kinases and of pieces, homogenized with 3 volumes of 5 mM potassium phostheir catalytic subunits was assayed in a standard incubation phate buffer (pH 7.0), containing 2 m&t EDTA, and the homogsystem by a procedure described previously (18). The incuba-enate was centrifuged at 13,000 x g for 20 min. This and all tion mixture contained, in a final volume of 0.2 ml, sodium ace-succeeding steps of purification were performed at 4", and all tate buffer (pH 6.0), 10 kmoles; histone mixture, 40 pg; [T-"~P]-buffers used contained 2 InM EDTA. Protamine sulfate (15 mg per 100 ml) was added to the crude extract; after 10 min of gentle stirring, the resultant nucleic acid precipitate was removed by centrifugation at 13,000 x g for 20 min. Solid ammonium sulfate (27.7 g per 100 ml) was added to the supernatant solution.
After stirring for 30 min, the solution was centrifuged; the protein precipitate was dissolved in about 450 ml of 5 mM potassium phosphate buffer (pH 7.0); and the enzyme solution was dialyzed overnight against 10 volumes of the same buffer, with two changes of buffer. Precipitate formed during the dialysis was removed by centrifugation.
The enzyme solution was then applied to a DEAE-cellulose column (4 x 30 cm) which had been previously equilibrated with 5 mM potassium phosphate buffer (pH 7.0), and the column was washed with 2 volumes of the same buffer. The enzyme was then eluted from the column with a linear gradient of phosphate buffer (0.095 to 0.3 M), pEI 7.0, in a total volume of 4.0 liters.
Two active enzyme peaks, eluted at 0.03 and 0.2 M phosphate, respectively, were obtained.
The enzyme activity eluted at 0.03 M phosphate was less than 5y0 of that eluted at 0.2 BI phosphate and was discarded.
The act,ive fractions cluted at 0.2 M phosphate were pooled and dialyzed estensively against 5 rnlr potassium phosphate buffer.
After dialysis, the enzyme solution was ayp!ied to a column (2.6 x 12 cm) of hydroxylapatite which had been previously equilibrated with 5 m&r potassium phosphate buffer at pH 7.0. The enzyme was then eluted from the column with a linear gradient of potassium phosphate buffer (0.005 to 0.2 M), pH 7.0, in a total volume of 2.0 lit.ers. Two enzyme peaks, designated Peak 1 and Peak 2, respectively, were obtained (Fig. 1). The active fractions from each peak were pooled, dialyzed overnight against 5 InM potassium phosphate buffer at pH 7.0, and then concentrated separately by means of ultrafiltration on Diaflo membranes (PLVI-10, Amicon Corp.) under pressure of nitrogen gas. The resultant enzyme preparations were designated protein kinase I and protein kinase II, respectively.
About 5 mg of protein kinase I, in a 2.0.ml volume, containing 50 m;\l potassium phosphate buffer (pH 7.0), 1.2 rnnl cysteine, 570 sucrose, and O.O015ol, bromphenol blue (as tracking dye), was applied to a column (2.1 x 8.0 cm) of 770 acrylamide gel, and preparative gel electrophoresis was carried out at 20 ma and 4" for 11 hours. The gel column had been subjected to a "blank" electrophoresis run at 20 ma and 4' for 3 hours before 1. Chromatography of bovine brain cyclic AMP-dependent protein kinase activity on hydroxylapatite column. Experimental details are described in the text. An aliquot (10 ~1) from each fraction was assayed for protein kinase activity in the presence of 5 PM cyclic AMP. 181 the application of the enzyme sample; this prerun was found to increase recovery of enzyme activity in t,he gel purification step. The upper buffer was 0.05 M Tris-glycine (pH 8.9) containing 1.2 rnM cysteine, and the lower buffer was 0.1 M Tris-Cl (pH 8.1). After completion of electrophoresis, the gel was removed from the glass tube and soaked in S-anilino-naphthalene-sulfonate (magnesium salt) ; the protein bands were visualized under ultraviolet light, according to the method of Hartman and Udenfriend (33). Five protein bands were observed. The gel was cut into 2-mm slices, each slice was homogenized in a Teflon homogenizer with 10 ml of 10 mM potassium phosphate buffer at pH 7.0, and the homogenate was centrifuged at 27,000 x g for 5 min. The gel was then resuspended in 5 ml of the same buffer and again removed by centrifugation.
The supernatant fluids (containing enzyme eluted from the gel) were combined and dialyzed overnight against 5 mM potassium phosphate buffer at pH 7.0. The protein kinase activity stimulated by cyclic AMP was fouud to be associated with two of the five protein bands. The slower moving of the two active bands was designated protein kinnse IA, and the faster moving was designntcd protein kinase In.
About 4 ml (4 to 5 mg) of protein kinase II, the enzyme from I'cak II of the hydrosylnpntite chromatography, was applied to a column (1.3 x 90.0 cm) of Sephnroae 613, which had been previously equilibrated with 5 mM potassium phosphate buffer at pII 7.0. The enzyme was then eluted from the column with about 100 ml of the same buffer. The active fractions were pooled and concentrated by means of Diaflo membrane ultrafiltration.
Two milliliters of this concentrated enzyme solution, containing about 4 mg of protein, was then subjected to preparative gel electrophoresis, and the protein TX-as eluted from the gel by the procedure described above for protein kinase I. After exposure to 8-anilino-naphthalene-sulfonate, three bands (and in some preparations, a fourth band) were visualized.
The cyclic AMP-dependent enzyme activity was found to be associated with only one band on the gel. This activity is designated protein kinase IIA.
The procedure used for the preparation of the various brain protein kinascs, together with their S values, is given in Scheme 1. The results of the enzyme purification are summarized in Table 1. The low recovery of the purified enzyme was because only the chromatographic fractions with highest activity were used for subsequent purification steps.
Protein kinases IA, 113, aud IIA were subjected to analytical gel elcctrophorcsis.
Protein kinases 113 and IIA each migrated as a single band at each of three pH values, namely pI1 8. 1,9.5,and 12.0 (Fig. 2); in each case, enzyme activity was found only in the region of the gel corresponding to the protein band. Protein kinase IA contained two protein bands. The one staining more intensely contained all of the enzyme activity.

Preparation of Regulatory and Catalytic Subunits from Bovine Brain Cyclic AXP-dependent
Protein Kinases--Subunits mere prepared from bovine brain cyclic AMP-dependent protein kinases by affinity chromatography on columns containing covalently bound protamine.
Protein kinase II (6.3 mg) from bovine brain was preincubated at 0" for 10 min with 5 X 1OV M cyclic [G-3H]AMP (25 ,uCi) in 7.0 ml of 10 mM 2-(N-morpholino)ethane sulfonic acid buffer (pH 6.0), containing 1 ml{ EDTA. The solution was then applied to a column (0.9 X 7.0 cm) of Enzite CM-cellulose-protamine which had been previously equilibrated with the same buffer, and the column was washed with 7 ml of the buffer containing 5 x lO-'j M of the radioactive cyclic AMP.
The subunits were eluted from the column with 170 ml of a linear gradient of sodium chloride (0 to 1 M), present  Summary of purification of cyclic AMP-dependent protein lcinases from bovine brain Four whole bovine brains (1450 g) were used as the starting material.
Protein kinase activity was determined in the presence of 5 pM cyclic AMP.
The purification procedure used is described in the text. in the same buffer, in the absence of cyclic AMP (Fig. 3). The protein peak containing bound radioactive cyclic AMP was not retained by the column, and there was no catalytic activity associated with this peak. The catalytic activity retained by the column was eluted at approximately 0.2 M sodium chloride. This catalytic activity was not associated with protein-bound radioactive cyclic AMP.
Protein kinase IC (3.5 mg), prepared from protein kinase I, as described in the next section, was also chromatographed on Enzite CM-cellulose-protamine and gave results similar to those shown in Fig. 3.
Regulatory subunit free of bound cyclic AMP was prepared from brain protein kinase by taking advantage of the ability (16) of histone to cause the dissociation of the enzyme into subunits. For this purpose, 20 mg of the enzyme preparation from the DEAE-cellulose column was preincubated at 30" for 10 min in the presence of 100 pg of histone per ml. After preincubation, the solution was applied to a column (2.1 x 1.5 cm) of hydroxylapatite, which had been previously equilibrated with 5 mM pot,assium phosphate buffer (pH 7.0), containing 2 mM EDTA. The column was then washed with 10 ml of the same buffer, and the protein was eluted with 0.02 M potassium phosphate buffer (pH 7.0), containing 2 mM EDTA.
Cyclic AMP-binding activity was assayed on an aliquot of each fraction.
The peak of cyclic AMP-binding protein appeared in the 7th through the 12th ml. Active fractions were combined and used as cyclic AMP-binding protein in the recombination experiments described below. Protein kinase catalytic activity of this preparation was negligible.
Previous methods of preparing regulatory subunits of protein kinase have used cyclic AMP to dissociate the holoenzyme and have suffered from the disadvantage that it is extremely difficult to free the regulatory subunit from bound cyclic nucleotidc.
Dissociation of holoenzymes by protein substrates provides an effective means of preparing regulatory subunits without cyclic nucleotides being attached.
The solution was then applied to a column (0.9 X 5.0 cm) of Enzite CM-cellulose-protamine, which had been previously equilibrated with the same buffer, and the column was then washed with 5 ml of the buffer containing 5 x 10Ve M of the radioactive cyclic GMP.
The elution of protein from the column was carried out with 170 ml of a linear gradient of sodium chloride (0 to 1 M) present in the same buffer, in the absence of cyclic GMP.
About 737, of the total enzyme activity, and all of the protein-bound radioactive cyclic GMP and the free radioactive cyclic GMP, passed straight through the column.
Attempts to purify further the cyclic GMP-binding protein from this fraction have not been successful, The remaining 27% of the enzyme activity retained by the column was eluted at approsimately 0.2 M sodium chloride; this activity was independeut of added cyclic GMP. The results suggested that only about one-fourth of the holoenzyme had been dissociated into subunits by this procedure.
The catalytic subunit from lobster muscle cyclic AMP-dependent enzyme was prepared by the procedure described for the cyclic GMP-dependent enzyme, except that the enzyme was preincubated with 5 x IO-6 M cyclic [G-3H]AMP (15 &i) instead of cyclic [GJH]GMP.
The yield of the catalytic subunit was 32%. Its activity was found to be independent of added cyclic AMP, and it was free from protein-bound radioactive cyclic AMP.
The cyclic nucleotide-independent catalytic subunits of the two classes of protein kinase from lobster muscle were separately concentrated on PM-10 Diaflo membranes.

Sucrose Density Gradient Centrifugation of Holoenzyme and
Catalytic Subunit of Cyclic AMP-dependent and Cyclic GMP-dependent Protein Kinases Protein kinase I, the enzyme from Peak I of the hydroxylapatite column, was found on sucrose density gradient centrifugation to consist of two distinct components, with sedimentation coefficients of 7.7 S (mol wt 140,000) and 5.4 S (mol wt 84,000) (Fig. 4). Both components of the enzyme were activated by cyclic AMP.
The 5.4 S component of protein kinase I was designated protein kinase IC, and the 7.7 S component was designated protein kinase ID.
Protein kinase II consisted of only one active component, with a sedimentation coefficient of 7.7 S (Fig. 4). The two active enzyme bands, protein kinase IA and IB, eluted from the gel following preparative gel electrophoresis of protein kinase I, were found to have sedimenta- tion coefficients of 7.7 S and 5.4 S, respectively. Protein kinase IIA, prepared from protein kinase II by Sepharose 6B chromatography and preparative gel electrophoresis, as described above, sedimented at a position corresponding to 7.7 S. The peak of catalytic activity (catalytic subunit) and that of cyclic AMP-binding protein (regulatory subunit), derived from protein kinase II by chromatography on Enzite CM-celluloseprotamine (Fig. 3), were centrifuged in a sucrose density gradient; protein kinase II holoenzyme was included for purposes of comparison.
The 2 subunits were found to sediment in positions corresponding to 3.6 S (mol wt 40,000) and 6.1 S (mol wt lOO,OOO), respectively (Fig. 5). The cyclic AMP-binding activity peak of the holoenzyme sedimented in a position corresponding to 7.7 S (mol wt 140,000) (Fig. 5) as did its catalytic activity peak (Fig. 4).
The peak of catalytic activity (catalytic subunit) and that of cyclic AMP-binding protein (regulatory subunit), derived from protein kinase IC by chromatography on Enzite CM-celluloseprotamine, were also analyzed by sucrose density gradient centrifugation, along with the holoenzyme (5.4 S) from which they were obtained.
The catalytic subunit sedimented in a position corresponding to 3.6 S (mol wt. 40,000), whereas the protein-bound radioactive cyclic AMP sedimented in positions corresponding to 6.1 S (mol wt 100,000) and 2.0 S (mol wt 20,000), respectively.
The 6.1 S component of the regulatory subunit may be an aggregate form of the 2.0 S component.
The isolated catalytic subunit of lobster muscle cyclic GMPdependent enzyme sedimented in a sucrose density gradient at a position corresponding to 3.6 S (mol wt 40,000), and the holoenzyme from which it was derived sedimented at a position of 7.7 S (mol wt 140,000) (Fig. 6A). The isolated catalytic subunit and the holoenzyme of the lobster muscle cyclic AMPdependent protein kinase, on the other hand, sedimented at positions corresponding to 4.5 S (mol wt 60,000) and 5.7 S (mol wt 90,000), respectively (Fig. 6B). With bot.11 classes of enzyme, the cat:tlytic actiT-itI\-of the holoenzyme was dependent upon added cyclic nucleotide, wl1crea.s t,he cat.alytic activity of t.he subunit was indepcndc~nt of added cyclic nucleotide.
It is interrsting that the cnt'alytic subunits derived from lobster muscle cyclic (XII'-dependent prot,ein kinase and from bovine brain cyclic ARWdcpendent protein kinascs had the same s&ncntation coefficient (3.6 S) which differed from that for the cat:llgtic subunit (4.5 S) from lobster muscle cyclic XKPdependent protein kinase. These findings correlate with our earlier observations (6, i) that, l&h respect t,o some kinetic properties, cyclic GRIP-dependent enzymes from art'hropods resembled mammalian cyclic AMP-dependent enzymes and were dissimilar to arthropod cyclic AMP-dependent, enzymes.

Dissociation and &iz~ation of Protein Iiinases by Histone and bv Q&c Nucleotides
Bovine Brain Cyclic LIJIP-dependent Protein Kineses-When protein kinase II was preincubated Gth historic (1 mg per ml) and t,hen centrifuged in the presence of the ame concentration of l&tone, the enzyme became dissociated.
Thus, the catalytic activity of this enzyme preparation, which hnd had a peak of i.i S (mol wt 140,000) and had been cyclic XW-dependent, shifted in the presence of histone to a position of 3.6 S (mol wt 40,000) and was no~v cyclic BRIP-ilrdepcndent. (Fig. 7). Studies of the effect of l&one on protein kinase II-1 which had been further purified, through the stages of Sepharose 6l3 chromatography and preparative gel electrophoresis, gave results similar to those observed with protein kinase II.
Ilnder the same csperimental conditions as were used in the experiment of Fig. 7, hist.onc also caused the dissociation and act.ivation of protein kinase IR (Fig. 8)  FIG. (i. Sucrose density gradient centrifugation of the holoenzymes and catalytic subunits of lobster muscle cyclic GMPdependent and cyclic BMP-dependent protein kinases. The catalytic subunits were prepared by column chromatography on Enzite CN-cellulose-protamine, as described in the text. Roloenzyme (2.0 mg) and catalyt.ic subunit (50 rg) of cyclic G;\lPdependent protein kinase (A), and holoenzyme (2.9 mg) and catalytic subunit (57 pg) of cyclic AMP-dependent protein kinase (B), each in 0.22 ml of solution containing 50 mn sodium acetate buffer (pII G.O), 0.3 mnr EGTA, and 2.5 mM 2-mercaptoethanol, were separately layered onto 4.8 ml of a 5 to 20% sucrose density gradient, cont.aining the same concent.ration of acetate buffer, EGTA, and 2.mercaptoethanol.
Protein kinase activity in fractions was amayed in the absence (.+.e) or presence (--) of 5 FM cyclic GNP (A) or 5 pM cyclic AMP (B), respect.ively. The data for holoenzymes are represented by circles, and those for the catalytic subunits are represented by triangles. activity shifted, in the presence of 1 mg of histone per ml, from a cyclic AMP-dependent peak at 5.4 S (mol wt 84,000) to a cyclic ilMP-independent peak at 3.6 S. The effect of cyclic AMP, alone and in combination with histone, was studied on the dissociation of protein kinase IC and protein kinase II from bovine brain.
In these experiments, preincubation a,nd centrifugation in the presence of cyclic AMP, with or without histone, were carried out in a manner analogous to that used for the studies of histone a,lone (Fig. 7). In high concentration (5 x 10-* M), cyclic AMP caused compIete dissociation of each of these enzymes. In some experiments, the effects of suboptimal concentrations of histone and of cyclic AMP were examined. h low concentration of histone (200 pg per ml) caused the partial dissociation of protein kinase IC, so that two peaks of activity were observed, one at 5.4 8 which was cyclic LMP-dependent, and one at 3.6 S which was cyclic AXIPindependent.
A low concentration of cyclic AMP (5 x 1OV &I) also caused the partial dissociation of protein kinase IC: two peaks of catalytic activity appeared, one at 5.4 S and one at 3.6 S. In the combined presence of these low concentrations of histone and cyclic AMP, protein kinase IC was completely dissociated: all catalytic activity appeared in a position corresuond-ing to 3.6 S. A low concentration of histone (200 pg per ml) also caused t,he partial dissociation of protein kinase II. In this case, however, the peak of xtivity at 7.7 S, observed with the holoenzyme, was converted quantitatively to two peaks of lower molecular IT-eight. One peak of catalytic activity at 5.4 S was cyclic AMP-dependent, and another peak of catalytic activity at 3.6 S was indcpendeut of cyclic AMP.
A low concentration of cyclic ,1?\IP (5 X lop6 RI) also caused the dissociation of protein kinase II: two peaks of catalytic activity appeared, one at 5.4 S and one at 3.6 S. In the combined presence of these low concentrations of hist.one and cyclic AMP, protein kinase II was completely dissociated, all catalytic activity appearing in a position corresponding to 3.6 S. Thus, the effects of histone and of cyclic AMP were additive in bringing about the dissociation of each of the bovine brain protein kinase preparations . At the end of the preincubation, t,he entire vol~m~e of the solution was layered onto 4.8 ml of a 5 to 20fo sucrose density gradient containing the same concentrations of acctntc buffer, I<:GTA, and Zmercaptoethanol, with or without histone, as were present in the preincubation solution.
After centrifngation, t.he fract,ions obtained from each t,ube were assayed for enzyme activity under standard conditions in the absence (+ . . . ) or presrncc (--) of 5 piv cyclic AMP. studied.
The value of 3.6 S for the catalytic subunit of protein kinase IC and of protein kinase II agrees with the corresponding data, obtained by sucrose density gradient centrifugation, for the catalytic subunits of these enzymes isolated by column chromatography on Enzite CM-cellulose protamine. Several proteins, other than histone, were examined for their cffectivencss in dissociating the brain enzymes, protein kinase II and protein kinase IC. Protamine (500 pg per ml) caused the dissociation of both protein kinases with formation of the 3.6 S cata.lytic component, whose activity was no longer stimula.ted by cyclic Ah!tP. Casein (3 mg per ml) caused protein kinase II to dissociate partially into the 5.4 S and 3.6 S components, and caused partial dissociation of protein kinase IC into the 3.6 S component.
Neither protein kinase was dissociated when preincubated with a variety of other proteins, such as bovine serum albumin (500 ,LL~ per ml), fructose 6-phosphate kinase (200 pg per ml), poly-L-serine (200 pg per ml), cat,alase (200 fig per ml), glyceraldehyde 3-phosphate dehydrogenase (200 pg per ml), or horse liver alcohol dehydrogenase (200 pg per ml). Tao (34), using a cyclic AMP-dependent protein kinasc from rabbit ergt'hrocytes, was able to confirm our earlier observations (16) on the dissociation of brain cyclic AMP-dependent protein kinases by protamine, but could not demonstrate dissociation of the e&hrocyte enzyme by histone.
His somewhat different result,s may be a reflection of t,issue or species differences in the properties of protein kinases.
The conversion of protein kinase in the presence of histone to a cyclic AMP-independent form, observed in the sucrose density gradient esperimeuts (Figs. 7 and S), could also be demonstrated simply by preincubation of the enzyme with histone, followed by assay of protein kinase activity in the absence of cyclic AWL'. Thus, preincubation of protein kinascs IB, IC, II, or IIA in the presence of histone (5 to 200 pg per ml) caused a substantial itlcrease in the activity of the enzyme in the absence of cyclic AhIP, with a corresponding decrease in the stimulatory effect of cyclic AMP.
The effect. on the activity of protein kinasc II of preincubation in the presence of various concentrations of histone is shown in Fig. 9. It is clearly possible that substratek. Isoelect.ric focusing of protein kinase II. The enzyme (6.1 mg) was preincubated at 30' for 5 min in a volume of 3.0 ml cont,aining 50 mM acetate buffer (pH 6.0), 0.3 mM EGTA, 2.5 mM Z-mercaptoethanol, 5 X 10-C M radioactive cyclic AMP (25 pCi), and 200 rg of histone per ml. The entire solution was then applied to a 110-ml isoelectric focusing column. Isoelectric induced dissociation is in part responsible for the basal (i.e. noncyclic AMP-dependent) protein kinase activity. Certain other proteins, including Iysine-rich histone, slightly lysine-rich histone, arginine-rich histone, protamine, and, to a lesser extent, casein, were also effective in activating both protein kinase IC and protein kinase II. Most other proteins tested were not effective in activating either of these enzymes; the inactive proteins included acetylcholinesterase, L-arginine decarboxylase, lactic dehydrogenase, avidin, lysozyme, and ,&galactosidase, as well as those proteins previously mentioned which were unable to cause the dissociation of the brain enzymes. In contrast to the ability of some substrate proteins to cause dissociation and activation of the kinases, preincubation of protein kinase IC or II in the presence of ATP plus Mg"f did not cause dissociation or activation of these enzymes.
Isoelectric Points of Catalytic and Regulatory Xubunits from Protein Kinase II-Protein kinase II was dissociated into its catalytic and regulatory subunits by preincubation in the presence of histone and cyclic AMP; it was then applied to a 110-ml isoelectric focusing column which did not contain added histone or cyclic AMP. By using these experimental conditions, it was possible to measure the pI for cyclic AMP-dependent enzyme and for the separate catalytic and regulatory subunits in the same experiment. A typical isoelectric focusing pattern of this enzyme and its subunits is illustrated in Fig. 10. A p1 of pH 4.6 was observed for the cyclic AMP-dependent activity. The peak of cyclic AMP-independent protein kinase activity (catalytic subunit) appeared at a pH value of 7.8, and the peak of protein-bound radioactive cyclic AMP (regulatory subunit) appeared at a pH value of 3.8. Chen and Walsh (35) reported earlier that the cyclic AMP-independent activity of hepatic protein kinascs had p1 values of pH 7.6 to 8.9.
Lobster Muscle Cyclic GMP-dependent and Cyclic AMPdependent Protein Kinases-Lobster muscle cyclic GMP-dependent enzyme, either not preincubated or preincubated in the absence of histone and cyclic GMP, was found to have a sedi-focusing was carried out as described under "Experimental Procedure." The pH, absorption at 280 nm, protein-bound radioactive cyclic AMP, and protein kinase activity in the presence (0-O) or absence (0.e.. 0) of 5 PM cyclic AMP were then determined on each fraction. mentation coefficient of 7.7 S (Fig. 1lA). When this enzyme was preincubated and then centrifuged in the presence of histone (1 mg per ml), a partial dissociation of the enzyme was observed (Fig. UA). The peak of catalytic activity at 7.7 S which had been cyclic GMP-dependent decreased in size, and a new peak of catalytic activity which was cyclic GMP-independent appeared in a position of 3.6 S. When this enzyme was preincubated and then centrifuged in the presence of 5 x 10m5 M cyclic GMP (Fig. llB), the enzyme was only partially dissociated. Two peaks of catalytic activity again appeared at positions corresponding to 7.7 S and 3.6 S. Cyclic GMP was more effective than cyclic AMP in dissociating and in activating the cyclic GMP-dependent enzyme. When the cyclic GMP-dependent enzyme was preincubated with both histone (1 mg per ml) and cyclic GMP (5 X 10U5 M), an almost complete dissociation of the enzyme into the 3.6 S component was observed (Fig. 11B).
Histone and cyclic nucleotides also caused dissociation of the cyclic AMP-dependent protein kinase from lobster muscle. Thus, the holoenzyme, which sedimented at a position of 5.7 S (mol wt 90,000), dissociated about 50% to a cyclic AMP-independent 4.5 S component (mol wt SO,OOO), when the enzyme was preincubated and centrifuged in the presence of 200 pg of histone per ml. Complete dissociation of the enzyme to the 4.5 S component was observed when it was preincubated with 5 x 10d6 M cyclic AMP. Cyclic GMP was less effective than cyclic AMP in dissociating or activating the cyclic AMP-dependent enzyme. When the enzyme was preincubated with both histone and cyclic AMP at the concentrations used separately, the enzyme also dissociated into the 4.5 S component. The values of 3.6 S and 4.5 S, respectively, for the catalytic subunits of cyclic GMP-dependent and cyclic AMP-dependent protein kinases of lobster muscle agree with the corresponding data, obtained by sucrose density gradient centrifugation, for the catalytic subunits from column chromatography on Enzite CM-cellulose-protamine.
In addition to histone, protamine (500 pg per ml) was also effective in causing dissociation of lobster muscle cyclic GMPdependent and cyclic SMP-dependent protein kinases into the corresponding cyclic nucleotide-independent catalytic subunits. Those proteins which were ineffective in causing the dissociation 12 - of the brain enzymes were also ineffective against the two lobster enzymes.
It remains to be determined whether the dissociation and activation of protein kinases by substrate proteins may represent a physiological mechanism for the regulation of biological processes. The level of cyclic AMP constitutes a mechanism for regulating protein kinase activity in response to extracellular signals; conceivably, availability of substrate proteins might constitute an intracellular mechanism for regulating the activity of this class of enzyme.
It was found in the present study that regulatory subunit prepared from bovine brain cyclic AMP-dependent protein kinase was able to inhibit the enzymatic activity of catalytic subunits derived from bovine brain cyclic AMP-dependent protein kinases, or, interestingly, from lobster muscle cyclic GMP-dependent protein kinase, accompanied in each case by a concomitant restoration of cyclic nucleotide dependence of the resultant holoenzyme (Fig. 12). For each enzyme, the activity of the free catalytic subunit, measured in the presence or absence of added cyclic AMP, and that observed with the reconstituted holoenzyme in the presence of added cyclic AMP were comparable.
The relatively high amount of regulatory subunit required to inhibit catalytic activity may have been due either to partial denaturation of the isolated regulatory subunit, or to an incomplete association of regulatory and catalytic subunits under the experimental conditions used. The reconstituted 'Lhomologous" holoenzyme, obtained by combining regulatory and catalytic subunits both derived from the cyclic AMPdependent class of protein kinase, was found to be activated preferentially by cyclic AMP rather than by cyclic GMP (Fig.  13A), as seen for the brain kinases before dissociation.
Cyclic AMP was also more effective than cyclic GMP in activating the reconstituted "hybrid" holoenzyme consisting of the regulatory subunit from a cyclic AMP-dependent protein kinase and the catalytic subunit from a cyclic GMP-dependent protein kinase (Fig. 13B). The latter results indicate that the cyclic nucleotide specificity of the regulatory subunit was not affected by its combination with the "heterologous" catalytic subunit. It would be interesting to study the effect of the regulatory and of isolated catalytic subunits of protein kinases IC and II were all in the range from 3.8 X lo+ to 5.6 X lo+ M. Double reciprocal plots of enzyme activity versus histone concent.ration showed a bipha.sic response for all protein kinase preparations studied.
The concentration of histone required to gire half-maximal activity of the catalytic subunits of lobster muscle cyclic AMP-dependent and cyclic GMP-dependent protein kinascs were 95 pg per ml and 105 pg per ml, respectively, and the activity reached a plateau at a histone concentration of about 500 pg per ml with both enzymes. The relationship between enzyme activity and histone concentration was also studied by using the holoenzymes of protein kinases IB, IC, II, and IIA and the catalytic subunits derived from protein kinases IC and II. In all six cases, the result,s were similar to those found with the catalytic subunits of the two lobster enzymes.
Meetal Ion Requirement--A comparison of the effects of divalent metal ions was carried out with the holoenzymes of protein kinases IB, IC, II, and IIA, the catalytic subunits derived from protein kinases IC and II, and the catalytic subunits derived from the lobster muscle cyclic AMP-dependent and cyclic GMPdependent protein kinases. All eight enzyme preparations had an absolute requirement for a divalent metal ion: Co2+, Mg2+, and hIn2f supported protein kinase activity, but C!a*+ did not. In fact Ca2f antagonized the stimulatory effect of the other metal ;'ons. For all eight enzyme preparations, enzyme activity was highest in the presence of 10 niM Co2+. Inhibition by iiDP and Related Compounds-It was observed in earlier studies that ADI', adenosine, and 2'-deosyadenosine inhibit,ed the activity of cyclic AMP-dependent protein kinases from brain, liver, and a variety of other tissues (5,29,38). In the present study, it was found that the activity of the holoenzymes of protein kinnses IB, IC, II, and IIA, as well as of the catalyt'ic subunits derived from protein kinases IC and II, were each inhibited 50 to 60% by 50 PM ADP. Sdenosine and 2'dcosyndenosine were about half as i)otent as ADP. These results are similar to those reported earlier with holoenzymes and indicate that the inhibitory action of these substances occurs on the catalytic subunits.
The inhibition by ADP, adenosine, and 2'-deosyadenosine of the catalytic subunits of the brain enzymes was readily overcome by increasing the ATP concentration, suggesting that these inhibitors compet,e with ATP for a common site on the catalytic subunit. Results similar to those obtained with the brain enzymes, i.e. inhibition by the three compounds mentioned above and removal of the inhibition by ATP, were also obtained with the catalytic subunits of lobster muscle cyclic AMP-dependent and cyclic GNP-depcndcnt protein kinases.