Concentrations of cyclic AMP-dependent protein kinase subunits in various tissues.

The concentrations of the regulatory (R) and catalytic (C) subunits of adenosine 3':5'-monophosphate (cyclic AMP)-dependent protein kinase(s) were measured in extracts of skeletal muscle, heart, liver, kidney, and brain. These concentrations were also estimated for the particulate fraction from brain, the only tissue in which a major part of the total activity was not readily extracted in a soluble form. Values for R were determined by measuring the amount of cyclic [3H]amp bound to protein in these tissue fractions under specified conditions; it was assumed that 1 mol of cyclic AMP binds to 1 mol of R. Values for C were determined from measurements of the specific protein kinase activity of the fractions utilizing the turnover number of pure C in the calculations. Turnover numbers for C were found to be identical for this subunit obtained in the pure form from rabbit skeletal muscle, rabbit liver, and beef heart. The methods used for measuring C were evaluated by kinetic studies and through the use of the specific heatstable protein inhibitor of cyclic AMP-dependent protein kinase(s). R and C were found to exist in a 1:1 molar ratio in all of the tissue fractions that were studied. the absolute concentrations of R and C ranged from 0.23 mumol/kg wet weight for liver to 0.78 mumol/kg wet weight for brain. For brain this value was based on the amount of each subunit in the particulate as well as the soluble fraction. For other tissues the values were based solely on the subunit content of the latter fraction. It was noted that the molar concentrations of R are close to those of cyclic AMP under basal conditions in the various tissues.

The concentrations of the regulatory (R) and catalytic (C) subunits of adenosine 3':5'-monophosphate (cyclic AMP)dependent protein kinase(s) were measured in extracts of skeletal muscle, heart, liver, kidney, and brain. These concentrations were also estimated for the particulate fraction from brain, the only tissue in which a major part of the total activity was not readily extracted in a soluble form. Values for R were determined by measuring the amount of cycliQH1 AMP bound to protein in these tissue fractions under specified conditions; it was assumed that 1 mol of cyclic AMP binds to 1 mol of R. Values for C were determined from measurements of the specific protein kinase activity of the fractions utilizing the turnover number of pure C in the calculations.
Turnover numbers for C were found to be identical for this subunit obtained in the pure form from rabbit skeletal muscle, rabbit liver, and beef heart. The methods used for measuring C were evaluated by kinetic studies and through the use of the specific heatstable protein inhibitor of cyclic AMP-dependent protein kinase(sL R and C were found to exist in a 1:l molar ratio in all of the tissue fractions that were studied. The absolute concentrations of R and C ranged from 0.23 pmollkg wet weight for liver to 0.78 pmol/kg wet weight for brain. For brain this value was based on the amount of each subunit in the particulate as well as the soluble fraction. For other tissues the values were based solely on the subunit content of the latter fraction. It was noted that the molar concentrations of R are close to those of cyclic AMP under basal conditions in the various tissues. Cyclic AMP-dependent' protein kinases (EC 2.7.1.37; ATP:protein phosphotransferase) from several tissues are known to contain two types of subunits: catalytic subunits (C), * This work was supported by a grant from the Muscular Dystrophy Association of America, Inc. and by Grant AM-12842 from the National Institutes of Health.
0 Recipient of a Postdoctoral Fellowship from the Muscular Dystrophy Association of America, Inc.
which catalyze the transfer of the y-phosphate of ATP to certain proteins, and regulatory subunits (R), which, in the absence of cyclic AMP, inhibit the activity of the catalytic subunits (see Ref. 1). Cyclic AMP activates these enzymes by causing dissociation of the inactive holoenzyme (R,C,) to yield free C subunits and a cyclic AMP regulatory subunit complex (R,.cyclic AMP,) according to the following equation established for the enzyme purified from skeletal muscle (2) or heart (3, 4): R,C, + 2 cyclic AMP % R,.cyclic AMP, + 2C (1) At least two types of the cyclic AMP-dependent protein kinase exist in various mammalian tissues. These have been referred to as type I and type II according to their order of elution from DEAE-cellulose with increasing salt concentration (5). It is not known, however, whether "type I" and "type II" enzymes from different tissues are necessarily identical. Both types of protein kinase have been purified to near homogeneity from rabbit skeletal muscle and were referred to as Isozyme I and Isozyme II (6). In this instance it was found that each form contained identical C subunits but differed in the properties of R. In general, work from several laboratories has suggested that variability in the properties of the different cyclic AMP-dependent protein kinases resides in differences in their R subunits (5-8). The cyclic AMP-dependent protein kinase from bovine heart, which has been characterized extensively (9), is a type II enzyme.
Many factors can affect the equilibrium reaction shown in Equation 1 and could thus influence the activation of the protein kinases by cyclic AMP in vivo. Some of the factors that have been studied include: (a) the binding of MgATP to type I protein kinase and its effect on recombination of the subunits (2,4,10-12); (b) autophosphorylation of the regulatory subunit of type II protein kinase and the effect of this reaction on cyclic AMP binding and dissociation of the enzyme (3, 4); (c) dissociation and activation of the protein kinases by basic protein substrates (13,14); and (d) the presence of an inhibitory protein other than R that is capable of binding to C (15). In addition to the above, it has been stressed that it is necessary to take total enzyme concentration into account when attempting to estimate the degree of kinase activation by a given concentration of cyclic AMP (12, 16). Finally, it has been shown (lo), and would be anticipated from Equation 1, that the relative concentrations of R and C subunits affect the response to cyclic AMP in vitro.
The present study was undertaken in an attempt to gain 1442

Concentrations of Protein
Kinase Subunits further knowledge as to the concentration of the two types of cyclic AMP-dependent protein kinase subunits in different animal tissues. It was thought that such information would be useful with respect to our understanding of the activation of the enzyme by cyclic AMP, as discussed above, and would also provide a basis for considering possible alternative functions of the regulatory subunit. In addition, knowledge concerning the relative concentrations of the catalytic and regulatory subunits would have a bearing on work related to their biosynthesis or degradation or both.

MATERIALS AND METHODS
Preparation of Tissue Fractions -Rabbits, 2.5 to 3 kg, (New Zealand Whites) were anesthetized deeply with an overdose of sodium pentobarbital and bled from the jugular veins. Tissues were quickly removed and immediately chilled on ice. All subsequent steps were carried out at 4". Skeletal and heart muscle were passed through the coarse disc of a meat grinder but the preliminary mincing of the softer glandular tissues was done with scissors. The ground of minced tissues were then placed in either 2l12 or 5 volumes per kg wet weight of 6 mM EGTA (pH 7.0) containing 15 mM 2-mercaptoethanol and homogenized for 45 s at high speed in a Waring Blendor; these homogenates will be identified as 3:l or 6:l. The homogenates were centrifuged at 10,000 x g for 25 min and the supernatants decanted through glass wool. The sediment obtained from centrifugation of the brain homogenates was extracted two times with 2 volumes of homogenization buffer and the resulting supernatants were combined.
The final pellet was resuspended in 2i/z volumes of homogenization buffer. For the various types of assays (see below), extracts were diluted in 25 mM Mes buffer, pH 6.9, containing 0.5 mg/ml of bovine serum albumin.
Assays were carried out the same day the extracts were prepared, usually within 2 to 3 h after death of the animals.
Assam for Cvclic AMP-dmendent Protein Kinase Activitv-Protein kinase"&tivi<y was measured at pH 6.9 in reaction mixtures with a final volume of 0.1 ml containing 2 pmol of Mes, 0.2 pmol of EGTA, 0.75 Fmol of magnesium acetate, 0.01 kmol of [r-32P]ATP (100 cpm/ nmol). 100 us of histone f2b, ? 100 pmol of cyclic AMP, and 20 ~1 of diluted enzyme (see above for method of dilution) to start the reaction. Incubations were carried out for 2 min at 30". Reactions were terminated and the amount of 32P incorporated into the protein substrate determined as described (Method B) by Reimann et al. (17). Inasmuch as the holoenzyme forms of the cyclic AMP-dependent protein kinases are known to dissociate completely at the cyclic AMP concentration used in the assay (l), it was assumed that the activity expressed in these determinations was that of the free catalytic subunit.
One unit of activity is taken as that amount of enzyme catalvzinz the transfer of 1 firno of nhosnhate to histone f2b ner min.
Assay For Regulatory Subunit &ncentration -The concentration of the reaulatorv subunits of cvclic AMP-dependent protein kinasets) -" was determined by measuring the cyclic AMP binding capacity in tissue fractions using the method of Gilman (18). Binding reactions were carried out at DH 4.0 in a O.l-ml reaction mixture containing 5 pmol of sodium acetate buffer, 10 pg of bovine serum albumin, cyclic [3H] AMP (28 Ci/mmol), and aliquots of the extracts. After incubation for at least 60 min, the reaction mixtures were diluted with 2 ml of cold 20 mM potassium phosphate buffer (pH 6.2) and poured through cellulose ester filters: The filters were washed twice with 5 ml of cold diluting buffer, dissolved in 2 ml of ethoxyethanol, and counted in a toluene-based scintillator. Assays for Cyclic AMP Phosphodiesterase and Histone Phosphatase Activities -Assavs for cvclic AMP phosphodiesterase and histone phosphatase were carried out according to Beavo et al. (19) and England et al. (20), respectively, except for the following changes. The reaction mixtures had the composition of those used for determining protein kinase activities except that [r -:'*P]ATP was replaced by ATP, cyclic AMP was replaced by cyclic["H] AMP for the phosphodiesterase determinations, and histone f2b was replaced by 32Plabeled histone l2b in the phosphatase determinations. Reactions were run for 2 min at 30" in-both cases.
Preparation of Pure Cyclic AMP-dependent Protein Kinase Catalytic Subunit from Rabbit Liver and Beef Heart -The catalytic subunit (Cl of cyclic AMP-dependent protein kinase(s) was isolated from rabbit liver and beef heart by adapting the skeletal muscle procedure (Method B) of Beavo et al. (21) to these tissues. The following modifications were required: (a) the amount of DEAE-cellulose was increased to 1 liter of settled resin per kg wet weight of heart muscle and to 2 liters for the same amount of liver; (b) the CM50 batch absorption steps were carried out twice at pH 6.8, twice at pH 6.1, and once at pH 6.5 before cyclic AMP was added; and (c) the active fractions from the CM50 cellulose column were pooled and diluted with distilled H,O containing 0.15 mg/ml of bovine serum albumin to a final conductivity of 2.5 mmhos. The catalytic subunit was then absorbed on a 1 to 2-ml CM50 column. After extensive washing of the column, the subunit was eluted using a high salt buffer. This last step prevented inactivation of C from liver which otherwise occurred durina dialvsis.
Treatment of Skeletal Muscle Extract with Antibody to R type I and to R type II-Antibody to purified bovine skeletal muscle R type I was made in rabbits and antibody to bovine heart muscle R type II was made in goats.* In each case, the antibody was precipitated at 40% saturation with ammonium sulfate, dissolved in 100 mM NaCl, 1 mM EDTA, 10 mM Mes buffer, pH 6.5, and dialyzed to remove excess salt. Varying amounts of antibody were incubated with 20 ~1 of bovine skeletal muscle extract for 60 min at 23" and the mixture was centrifuged at 1600 x g for 15 min. necessary to know the specific activity and molecular weight of the subunit for each tissue to be examined. To isolate and characterize C from all tissues was not feasible, but it was possible to do this for three representative tissues. The tissues selected for this purpose included rabbit skeletal muscle, rabbit liver, and beef heart. In this connection, it should be noted that prior to this study it was already known that C prepared from skeletal muscle or from liver interacts with R from the opposite tissue (28). In addition, it was shown (6) that C derived from each of two different skeletal muscle cyclic AMPdependent protein kinase isozymes appeared to be identical in substrate specificity and physicochemical properties. * Details concerning the production and characterization of antisera to the two tvoes of R subunit will be nublished elsewhere (P. J. Bechtel The purification procedure designed for the isolation of C from rabbit skeletal muscle (21) was found to be applicable to rabbit liver or beef heart after slight modification, and essentially homogenous preparations were obtained from each of the latter tissues. For skeletal muscle this required a 3,600fold enrichment, for liver an ll,OOO-fold enrichment, and for heart a 2,800-fold enrichment i.e. calculating on the basis of the specific activity starting with crude extracts. The sodium dodecyl sulfate-gel electrophoresis patterns of the purified preparations are shown in Fig. 1. No detectable difference in the molecular weight was found for the subunits from the three sources; values ranged from 39,700 to 40,500. The specific activity was also the same for each catalytic subunit, all determinations falling in the range of 10.1 to 10.5 pm01 of phosphate incorporated into histone f2b per min mg of protein.
Extrapolating from these data it appeared reasonable to assume that C from other tissues would probably have similar properties, and that as a first approximation a molecular weight of 40,000 and a specific activity of 10.3 could be assumed in all cases; these values are used in subsequent calculations. It was recognized, however, that the possibility exists that one tissue or another could contain a catalytic subunit with different properties.

Development of Methodology for Determining
the Concentration of Catalytic Subunit in Tissue Fractions by Activity Measurements Dilution of Fractions-In order to carry out meaningful enzyme assays using crude tissue fractions it was essential that the assays be performed at high dilution. For example, with rabbit skeletal muscle extract specific activities did not become constant until at least a 50-fold dilution had been carried out (Fig. 2). Similar behavior was exhibited by fractions from tissues other than skeletal muscle, but in all cases it was possible to dilute out apparent inhibitory factors in the crude fractions. With proper dilution of the fraction, reaction rates were linear and were proportional to the amount of enzyme used. The choice of purified histone f2b as a substrate, rather than mixed histones, was essential in order to achieve workable reaction rates at these high enzyme dilutions.
Competing Reactions -Possible interference with the protein kinase assay due to competing reactions was examined systematically under conditions chosen to resemble as closely as possible those of the phosphoryl transfer reaction itself. Phosphoprotein phosphatase activity was estimated using 32Plabeled histone f2b as a substrate at a concentration equal to that maximally obtained in the protein kinase reaction. Under t-l M FIG. 1. Sodium dodecyl sulfate-gel electrophoresis of purified catalytic subunits. Proteins were heated at 60" for 60 min in 2% sodium dodecyl sulfate containing 5% 2-mercaptoethanol and examined by electrophoresis using 7.5% polyacrylamide gels containing 0.1% sodium dodecyl sulfate. Gels were stained with Coomassie blue for 12 h and destained by diffusion in 40% methanol, 7% acetic acid. L = C from rabbit liver, 2 pg; H = C from beef heart, 3 pg; and M = C from rabbit skeletal muscle, 3.5 pg. these conditions less than 1% of the phosphate was released at the high dilution of tissue fractions ordinarily employed, although as much as 50% of the substrate was hydrolyzed at low dilutions with some of the fractions. Cyclic AMP phosphodiesterase assays showed that negligible amounts (2 to 5%) of this component were degraded under assay conditions. The phosphorylation of endogenous protein, determined by omitting histone f2b, did not contribute significantly to the reaction product at the dilutions used.
Recovery of Added Enzyme-The validity of the enzyme assay was further evaluated by adding homogenous catalytic subunit to diluted extract in order to see whether or not it could be recovered quantitatively.
In these experiments either the amount of extract or the concentration of added catalytic subunit was varied. The added activity was recovered completely when amounts of extract corresponding to the linear part of the dilution curve were used or when increasing concentrations of catalytic subunit were added. The latter type of experiment is illustrated in Fig. 3. Identical results were obtained, whether catalytic subunit prepared from rabbit skeletal muscle, liver, or beef heart were used. Recovery measurements as described were carried out routinely for all the tissue examined. and particulate fraction from the brain homogenate resuspended in 2.5 volumes of buffer per kg wet weight of tissue and used at a 1:50 dilution.
The protein concentrations of the undiluted extracts were 15.8 mg/ml for skeletal muscle, 5.7 mg/ml for heart, 10.1 mg/ml for kidney, 5.1 mglml for brain supernatant, 22.0 mg/ml for liver, and 13.6 mg/ml for the brain particulate fraction. were assayed for cyclic AMP binding as described.

Concentrations
of Protein Kinase Subunits 1445 AMP-dependent protein kinase regulatory subunit(s) that might already be complexed with endogenous cyclic AMP, it was necessary to make certain that any bound cyclic AMP could exchange with the added cyclic ["HIAMP used in the assay. This exchange would require displacement of cyclic AMP from R, which could be achieved by reversing the reaction of Equation 1 i.e. by recombination of the protein kinase subunits.
Experiments were carried out in which crude tissue fractions were incubated in the presence of added MgATP and protein kinase catalytic subunit prior to carrying out a cyclic AMP binding assay. MgATP was added because it is known to facilitate recombination of protein kinase subunits (2, 4, lo-12, 34), at least with one of the isozyme forms of the enzyme. Additional catalytic subunit was provided since it would favor subunit recombination by mass action (see Equation 1). Two different results were obtained depending on the tissue examined. When skeletal muscle extracts were used, little or no increase in binding capacity was obtained regardless of whether or not the extracts were preincubated in the absence or presence of the various components. Similar results were obtained with extracts from liver and brain and with the particulate fractions from the latter tissue. In contrast, when heart or kidney extracts were used, up to a 2-fold increase in binding capacity was observed. The extent of the increase was variable from extract to extract and usually smaller in kidney than in heart. The binding capacity also increased in the kidney and heart extracts in the absence of added catalytic subunit, but at a lower rate whereas no change was seen in the absence of additions. As a result of these experiments, the preincubation of heart and kidney extracts with added MgATP and catalytic subunit for 30 min was adopted as a routine procedure. Inasmuch as no augmentation of cyclic AMP binding was observed for the other tissue fractions, they were assayed without preliminary incubation. A preliminary examination of the distribution of cyclic AMP-dependent protein kinase activity showed that in the tissues selected for study; namely, rabbit skeletal muscle, heart, kidney, liver, and brain, nearly all of the enzyme appeared to reside in the soluble fraction, obtained from homogenates except for brain. For that tissue approximately 60% of the enzyme appeared to be soluble and 40% was sedimented at 10,000 x g. Other workers have previously noted the high concentration of insoluble enzyme in brain (35). It was arbitrarily decided, therefore, to limit the study of subunit concentration to the soluble fraction except for brain in which the particulate as well as the soluble fraction were examined. The decision to limit the investigation to the soluble fraction of tissue homogenates was not meant to infer that the cyclic AMP-dependent protein kinase present in other fractions might not be important functionally, but only that its contribution to the total cellular concentration of enzyme subunits would be very small.
Protein kinase activities in the presence of cyclic AMP (catalytic subunit activities) and cyclic AMP binding capacities were determined for various tissue fractions. The results of a series of measurements are shown in Tables I and II. It will be noted that the specific catalytic subunit activity varied considerably from fraction to fraction, being highest in the brain supernatant and lowest in liver extract (Table I). The total activity per wet weight of tissue (Table II) showed less variation. The total activity for whole brain i.e. the sum of the soluble and particulate fractions, was appreciably higher than that for the other tissues (see above). The specific cyclic AMP binding capacities showed a similar variability from fraction to fraction (Table I), but again, binding capacities based on the wet weight of tissues (Table II) were more nearly constant. Molar concentrations of catalytic subunit in the different tis-

Concentrations of Protein
Kinase Subunits sues, calculated from the enzyme activity data, and molar concentration of regulatory subunit based on the binding capacity measurements, were found to be nearly identical. The R/C ratio was close to unity for all tissues (Table II).
The protein kinase activities recorded in Tables I and II were all performed in the presence of cyclic AMP but measurements on the various fractions were also carried out in the absence of the cyclic nucleotide. The ratio, activity (-cyclic AMP)/activity (+cyclic AMP), was on the average 0.1 for skeletal muscle, 0.15 for heart, 0.1 for kidney, 0.2 for brain extract as well as for the particulate fraction from this tissue, and 0.3 for liver. These results alone would require that the ratio, R/C, be as high as 0.7 to 0.9, assuming that Equation 1 accurately depicts the mechanism of activation of the cyclic AMP-dependent protein kinase.

DISCUSSION
This study describes methods that can be used to provide estimates of the molar concentrations of catalytic and regulatory subunits of cyclic AMP-dependent protein kinase(s) in crude tissue fractions. For skeletal muscle, heart, kidney, and liver, measurements carried out on the cytosol fraction alone give values that reflect the total tissue concentrations, since only a small proportion of either subunit is present in the particulate fraction. For brain, however, an appreciable amount of each subunit is present in the particulate as well as in the soluble fraction, and it is necessary to take both fractions into account in estimating the subunit content of the cell.
The methods used for measuring concentrations of the subunits depended upon (a) determinations of protein kinase activity using histone f2b as the substrate in the presence of sufficient cyclic AMP to completely dissociate the holoenzyme (see Equation 1) and (6) measurements of cyclic AMP binding capacity using the Millipore method. A necessary requirement for the validity of the catalytic subunit measurements was that the turnover number of the enzyme and that its molecular weight be identical for different tissues. This was established for three of the tissues examined and assumed to hold for the others. The reliability of the regulatory subunit determinations, as well as that of the catalytic subunit measurements, required that the methods used be specific for these proteins. Lending credence to the idea that the Millipore assay was specific for R, was the finding that essentially all of the assayable cyclic AMP binding protein present in muscle extract could be precipitated by antibody to pure type I and type II regulatory subunits. Similar experiments were not carried out, however, for the other tissue studied. Evidence that the only cyclic AMP binding protein detected in the Millipore assay was R was also derived from the results themselves. The molar ratio of R/C was found to be approximately 1 (Table II). Corroborating evidence for the presence of equimolar concentrations of C and R in crude tissue extracts can be taken from work with antibodies directed against the R subunit of bovine heart muscle protein kinase (type II) (36). In this case the ratio between catalytic activity versus either cyclic AMP binding capacity or immunoreactive R subunit remained constant throughout purification of cyclic AMP-dependent protein kinase from bovine heart muscle suggesting a similar ratio of C and R in extracts and purified preparations of holoenzyme. This latter work was also the first to document the immunological similarity of type II R values from a number of tissues and the lack of cross-reactivity of type I and type II R values even from the same species and tissue. Since the minimal ratios of R/C, as obtained from ratios of activity (-cyclic AMPYactivity (+cyclic AMP), were in the range of 0.7 to 0.9, it is clear that the tissue fractions could not have contained appreciable amounts of some other binding protein that was retained by the Millipore filter. The specificity of the catalytic subunit determination was supported by showing that essentially all of the detectable activity in the different tissue fractions could be inhibited by the heat-stable protein inhibitor of cyclic AMP-dependent protein kinases. The molar concentrations of regulatory subunit determined for the various tissues were surprisingly close to the reported basal levels of cyclic AMP i.e. to levels present under nonstimulated conditions in these tissues. Thus, in micromoles/kg wet weight, the basal level for cyclic AMP in skeletal muscle has been reported as 0.25 (371, whereas the regulatory subunit concentration determined here was 0.30 pmol/kg wet weight. Basal cyclic AMP levels of 0.36 wmol/kg wet weight have been reported for heart (371, 0.46 for liver (371, 0.67 for kidney (381, and 0.92 for brain (39). The range of regulatory subunit concentrations for these latter tissues was 0.25 to 0.76. The physiological implications of a situation in which the concentration of ligand and enzyme (or in this case regulatory protein) are the same order of magnitude have been discussed earlier (12).
It would appear that the relative rates of synthesis and degradation of the two types of subunits present in the cyclic AMP-dependent protein kinase are linked in such a manner that equimolar amounts are maintained within the cell. This finding makes it seem improbable that the regulatory subunit is involved in multiple regulatory events i.e. that it functions to regulate enzymes other than the protein kinase catalytic subunit.