Compartmentalization of adenosine 3':5'-monophosphate and adenosine 3':5'-monophosphate-dependent protein kinase in heart tissue.

In rabbit heart homogenates about 50% of the cAMP-dependent protein kinase activity was associated with the low speed particulate fraction. In homogenates of rat or beef heart this fraction represented approximately 30% of the activity. The percentage of the enzyme in the particulate fraction was not appreciably affected either by preparing more dilute homogenates or by aging homogenates for up to 2 h before centrifugation. The particulate enzyme was not solubilized at physiological ionic strength or by the presence of exogenous proteins during homogenization. However, the holoenzyme or regulatory subunit could be solubilized either by Triton X-100, high pH, or trypsin treatment. In hearts of all species studied, the particulate-bound protein kinase was mainly or entirely the type II isozyme, suggesting isozyme compartmentalization. In rabbit hearts perfused in the absence of hormones and homogenized in the presence of 0.25 M NaCl, at least 50% of the cAMP in homogenates was associated with the particulate fraction. Omitting NaCl reduced the amount of particulate-bound cAMP. Most of the particulate-bound cAMP was probably associated with the regulatory subunit in this fraction since approximately 70% of the bound nucleotide was solubilized by addition of homogeneous catalytic subunit to the particulate fraction. The amount of cAMP in the particulate fraction (0.16 nmol/g of tissue) was approximately one-half the amount of the regulatory subunit monomer (0.31 nmol/g of tissue) in this fraction. The calculated amount of catalytic subunit in the particulate fraction was 0.18 nmol/g of tissue. Either epinephrine alone or epinephrine plus 1-methyl-3-isobutylxanthine increased the cAMP content of the particulate and supernatant fractions. The cAMP level was increased more in the supernatant fraction, possibly because the cAMP level became saturating for the regulatory subunit in the particulate fraction. The increase in cAMP was associated with translocation of a large percentage of the catalytic subunit activity from the particulate to the supernatant fraction. The distribution of the regulatory subunit of the enzyme was not significantly affected by this treatment. The catalytic subunit translocation could be mimicked by addition of cAMP to homogenates before centrifugation. The data suggest that the regulatory subunit of the protein kinase, at least that of isozyme II, is bound to particulate material, and theactive catalytic subunit is released by formation of the regulatory subunit-cAMP complex when the tissue cAMP concentration is elevated. A model for compartmentalized hormonal control is presented.

From the Department ofPhysiolog.y, School of Medicine, Vanderbilt University, Nashville, Tennessee 3 7232 In rabbit heart homogenates about 50% of the CAMPdependent protein kinase activity was associated with the low speed particulate fraction. In homogenates of rat or beef heart this fraction represented approximately 30% of the activity. The percentage of the enzyme in the particulate fraction was not appreciably affected either by preparing more dilute homogenates or by aging homogenates for up to 2 h before centrifugation.
The particulate enzyme was not solubilized at physiological ionic strength or by the presence of exogenous proteins during homogenization. However, the holoenzyme or regulatory subunit could be solubilized either by Triton X-100, high pH, or trypsin treatment.
In hearts of all species studied, the particulate-bound protein kinase was mainly or entirely the type II isozyme, suggesting isozyme compartmentalization.
In rabbit hearts perfused in the absence of hormones and homogenized in the presence of 0.25 M NaCI, at least 50% of the CAMP in homogenates was associated with the particulate fraction. Omitting NaCl reduced the amount of particulate-bound CAMP. Most of the particulate-bound cAMP was probably associated with the regulatory subunit in this fraction since approximately 70% of the bound nucleotide was solubilized by addition of homogeneous catalytic subunit to the particulate fraction. The amount of CAMP in the particulate fraction (0.16 nmol/g of tissue) was approximately one-half the amount of the regulatory subunit monomer (0.31 nmol/g of tissue) in this fraction.
The calculated amount of catalytic subunit in the particulate fraction was 0.18 nmol/g of tissue. Either epinephrine alone or epinephrine plus 1-methyl-3-isobutylxanthine increased the CAMP content of the particulate and supernatant fractions. The CAMP level was increased more in the supernatant fraction, possibly because the CAMP level became saturating for the regulatory subunit in the particulate fraction. The increase in CAMP was associated with translocation of a large percentage of the catalytic subunit activity from the particulate to the supernatant fraction. The distribution of the regulatory subunit of the enzyme was not significantly affected by this treatment. The catalytic subunit translocation could be mimicked by addition of CAMP to homogenates before centrifugation.
The data suggest that the regulatory subunit of the protein kinase, at least that of isozyme II, is bound to particulate material, and the active catalytic subunit is released by formation of the regulatory subunit. CAMP complex when the tissue cAMP concentration is elevated. A model for compartmentalized hormonal control is presented.
The adenosine 3':5'-monophosphate-dependent protein kinase has been found in both soluble and particulate fractions of mammalian tissues (l-6). The latter fraction of protein kinase activity comprises a large percentage of the total in some tissues. Rubin et al. (3) found more than 70% of the total erythrocyte CAMP-dependent protein kinase in the particulate fraction, presumably associated with the plasma membrane. Menon (4) reported that 50% of the total enzyme in bovine corpus luteum was associated with the particulate fraction. The presence of CAMP-dependent protein kinase has been noted in preparations of sarcoplasmic reticulum (7-9) and in membrane-enriched fractions (10-12) of heart tissue. Heart Perfusion-Fed male rats (150 to 200 g) or rabbits (500 to 700 g) were used. Removal of hearts and perfusion were carried out as described earlier (15,16). Perfused and unperfused hearts, and beef hearts which were obtained from a local slaughterhouse were frozen and powdered before use unless indicated otherwise. Since in rabbit hearts the amount of particulate-bound protein kinase was found to increase with increasing age of the animal, the size of the rabbits used is indicated in the text. Homogenization-Powdered tissue was suspended at 4" in the indicated buffer and homogenized with three up and down turns of a motor-driven Teflon pestle at high speed in a glass tube. In some cases a plastic tube was used without noticeable difference in results. Protein Kinase Assay -The protein kinase assay was based on the phosphorylation of histone (Sigma type II-A) and was carried out essentially as described earlier (17). The assay reaction was started by adding 10 ~1 of the diluted (60 ml per g of tissue) sample to 50 ~1 of reaction mixture in the presence or absence of 2 +LM CAMP. The incubation was carried out at 30" for 5 min unless indicated otherwise. Under these conditions the assay was linear with time and concentration of tissue extract. In some cases the kinase activity is expressed as the protein kinase activity ratio, i.e. the ratio of activity in the absence of CAMP to that in the presence of CAMP (2 PM).
Adenylate Cyclase and Phosphodiesterase Assays -Adenylate cyclase activity was determined essentially as described by Drummond and Duncan (18) After the initial l-ml sample was applied, the column was washed with 4 ml of 20 mu ammonium formate (pH 7.51, and all 5 ml was collected and counted in a toluene/Triton X-100 scintillant. Cyclic AMP Assay -The buffers used for preparation of supernatant and particulate fractions always contained 10 mM EDTA and 0.5 rn~ 1-methyl-3-isobutylxanthine. These agents inhibit the hydrolysis of CAMP by phosphodiesterase (20). To l-ml aliquots of the supernatant or resuspended particulate fraction was added 0.1 pmol of 13HlcAMP (50,000 cpmipmol) to estimate the CAMP recovery. Immediately thereafter, the suspension was deproteinized by addition of 100 ~1 of 50% trichloroacetic acid and centrifuged at 12,000 X g for 20 min. The supernatant fraction was then pipetted onto a Dowex 50 column (0.9 x 10 cm, 100 to 200 mesh) equilibrated with 0.1 N HCl. The column was then eluted with 30 ml of 0.1 N HCl, and the lo-to 30-ml fraction was collected and lyophilized to dryness. The residue was resuspended in 1 ml of 50 mM sodium acetate (pH 4.0). Cyclic AMP was then determined by the protein-binding assay of Gilman (21).
CAMP-binding Protein (Regulatory Subunit) Assay -Binding of CAMP to proteins was routinely assayed at neutral pH by a method similar to that of Gilman (21). An aliquot (20 ~1) suitably diluted (60 ml/g of tissue) in buffer was added to 55 ~1 of a solution of 50 mu potassium phosphate (pH 6.81, 1 rnM EDTA, 0.5 mg/ml of histone (Sigma type II-A), 2 M NaCl, and 1 PM [RHlcAMP (6000 cpmlpmol1. The mixture was incubated for 60 to 90 min at 20", and then 1 ml of ice-cold 10 rn~ potassium phosphate, 1 mu EDTA at pH 6.8 was added. The mixture was filtered through a Millipore filter (HA 0.45 +n) previously moistened with the same buffer. The reaction tube was rinsed with another 1 ml of the same buffer. The filter was rinsed with 8 ml of buffer and dried in an oven at 150". Radioactivity retained by the filter was estimated by counting in 10 ml of a toluene-based scintillant.
The binding assay was done under saturating conditions and was linear with concentration of tissue extract. Complete exchange of labeled for unlabeled CAMP occurred under the conditions of the assay.
Partial Purification of Regulatory Subunits -The regulatory subunits of rat heart protein kinase were purified on DEAE-cellulose columns essentially as described earlier (22). The method is similar to that for purification of the supernatant fraction holoenzyme as described in Fig. 4. A homogenate (4 ml/g of tissue) of 15 fresh rat hearts (no perfusion or freezing) was prepared in 10 rn~ potassium phosphate (pH 6.8) containing 1 rnM EDTA. After centrifugation at 27,000 x g for 20 min, the supernatant fraction (57 ml) was incubated at 0" in the presence of 10 PM CAMP for 30 min and then applied to a DEAE-cellulose (Whatman No. DEll) column (2.5 x 25 cm) equilibrated with the homogenization buffer.
The column was washed with 200 ml of buffer containing 10 PM CAMP, followed by 1 liter of buffer containing 1 /LM CAMP, and finally with 1 liter of buffer. The column was eluted with a 500-ml linear (0 to 0.5 M) gradient of NaCl, and the fractions were assayed for CAMP-binding activity. Two peaks of CAMP-binding activity eluted at 0.1 and 0.25 M NaCl, and were presumably derived from type I and type II protein kinases, respectively (22, 23). The peak fractions were pooled separately, concentrated by (NH&SO, precipitation (65% saturation), and resuspended in and dialyzed against 10 rn~ potassium phosphate (pH 6.8).
Sucrose Density Gradient Centrifugation -Sucrose gradients were prepared by the method of Martin and Ames (24) as described earlier (23) (Table II). Histone slightly increased the activity in this fraction. Addition of excess regulatory subunits which had been partially purified from rat heart homogenates did not alter the amount of particulate-bound CAMP-binding activity. The latter experiment suggested either that the particulate-binding sites for the regulatory subunits were saturated or that the protein kinase and regulatory subunit bound to particulate material were different from the soluble forms.

Solubilization
of Regu1ator.y Subunit and Protein Kinase -As will become clearer later (see Table VI), when CAMP is added to homogenates or particulate fractions, the particulatebound holoenzyme dissociates into its regulatory and catalytic subunits. In media of physiological ionic strength the regulatory subunit remains bound to particulate material, but the catalytic subunit is released into the supernatant fraction. This property can be used to prepare particulate fractions containing the regulatory subunit but essentially free of the catalytic subunit. Comparisons can then be made between the solubilization of the particulate-bound regulatory subunit (CAMP-washed) and holoenzyme. Triton X-100 solubilized the regulatory subunit and holoenzyme of the rabbit heart particulate fraction assayed by CAMP binding (Fig. 2). At a concentration of 1% almost all (>90%) of the particulate fraction regulatory subunit or holoenzyme was solubilized. The disappearance of either the regulatory subunit or holoenzyme from the particulate fraction was accompanied by a corresponding increase of activity in the supernatant fraction. There was no significant difference in solubilization of the regulatory subunit as compared with the holoenzyme. Triton X-100 treatment also solubilized the CAMP-dependent protein kinase activity (see Fig. 4). The regulatory subunit and holoenzyme could also be solubilized by partial proteolysis with the use of low concentrations of trypsin (Fig. 3). Regardless of whether the starting material was the particulate-bound holoenzyme or the regulatory subunit, trypsin treatment effectively solubilized the CAMP-binding activity. As before, disappearance of the activity from the particulate fraction was associated with its appearance in the supernatant fraction. Trypsin treatment also solubilized the CAMP-dependent protein kinase activity when the starting material was particulate-bound holoenzyme (not shown). It was also found that, using the holoenzyme as the starting material, most of the CAMP-binding and kinase Substrate Specificity- Uno et al. (30) have reported that soluble and membrane-bound kinases of various tissues exhibit different substrate preferences for histone and protamine. Using the protein kinase assay conditions described under "Experimental Procedures," we have found no significant difference in the substrate specificity of the rabbit heart supernatant protein kinase as compared with the particulatebound enzyme. The particulate-bound enzyme was solubilized with 0.2% Triton X-100 before testing. Using histone (Sigma Type II-A), protamine, and casein as substrates, the relative rates of phosphorylation were ll:l:l, respectively, for the supernatant enzyme and 12:l:l for the particulate enzyme.
DEAE-cellulose Chromatography of Supernatant and Solubilized Particulate-bound CAMP-dependent Protein Kinase -The existence of two classes (isozymes I and II) of CAMPdependent protein kinases in supernatant fractions of extracts from various tissues, which can be separated by DEAF,-cellulose chromatography, is well established (231. The type I isozyme elutes from the column at CO.1 M NaCl, and type II elutes at ~0.1 M NaCl (23). The isozyme distribution of CAMPdependent protein kinase in supernatant and solubilized particulate fractions of hearts from different species is shown in Fig. 4. The particulate enzymes tiere solubilized by Triton X-100 before DEAE-cellulose chromatography.
It can be seen that although the isozyme pattern varied considerably in the supernatant fractions, the solubilized particulate enzymes chromatographed as mainly or entirely the type II isozyme in all cases. Rat heart exhibited a small amount of particulate fraction type I enzyme. As pointed out earlier (271, the activity eluting in the flow-through fraction is probably the free catalytic subunit. The enzymes eluted from the column were CAMP-dependent in all cases. Triton X-100 treatment of supernatant fractions did not affect the profiles (27). Hormonal Effects on Supernatant and Particulate Protein Kinase and CAMP -It can be seen in Table III that a large amount of the total CAMP of perfused rabbit heart was associated with the particulate fraction. The concentration of CAMP in the particulate fraction was less than the concentration of its receptor, the regulatory subunit of the protein kinase. If the hearts were perfused with either epinephrine alone or epinephrine plus 1-methyl-3-isobutylxanthine, the CAMP level increased in the supernatant and particulate fractions. In the presence of the two agents together, the cAMP level increased more in the supernatant than in the particulate fraction. This could be explained by the saturation of the regulatory subunit in the latter fraction at the high concentrations of tissue CAMP. The measured level of the regulatory subunit monomer (0.31 nmol/g) was about the same as the level of CAMP (0.33 nmol/g) in this fraction. The concentrations of regulatory subunit (RC + RcAMP) shown in Table III were calculated from the CAMP-binding activity, assuming binding of 1 mol of CAMP per regulatory subunit monomer (31). The concentrations of total catalytic subunit (RC + C) were calculated from the specific activity of homogeneous catalytic subunit (26). The total amount of regulatory subunit (supernatant plus particulate) was not always exactly the same as the total amount of catalytic subunit, probably because of the errors inherent in calculation of catalytic subunit amounts from specific activity of pure enzyme. Perfusion of hearts either with epinephrine alone or epinephrine plus 1-methyl-3-isobutylxanthine did not significantly alter the regulatory subunit level in either frac- Homogenates (60 ml/g of tissue for Experiment 1 and 5 ml per g of tissue for Experiment 2) were prepared from frozen hearts from 0.5 kg rabbits which had been perfused for 2 min in the absence or presence of epinephrine (Epi) or epinephrine plus 1-methyl-3-isobutylxanthine (MIX) as indicated.
The homogenization buffer was 10 rnM potassium phosphate (pH 6.81, 10 mM EDTA, 0.5 rnM l-methyl-3isobutylxanthine, and 0.25 M NaCl. In Experiment 1 the particulate fraction was washed three times in the same volume and buffer in the absence of NaCl. In Experiment 2 the homogenate was centrifuged once at 20,000 x g for 10 min at 4". The particulate fraction was resuspended in the same volume of the buffer above, and the super-natant and particulate fractions were diluted 12 times i n the same buffer. The protein kinase activity (2 ELM CAMP) was determined immediately, and the CAMP-binding protein (R subunit) was determined the following day. Aliquots (0.4 ml) were placed in a boiling water bath for 1 min and then centrifuged for CAMP determination.
The concentration of catalytic subunit (C) was calculated from the specific activity and molecular weight of pure liver catalytic subunit monomer (3 x 10" unitsimg)  Perfusion and homogenization (5 ml/g of tissue) were done as described in Table III. The homogenization and wash buffer was 10 mM potassium phosphate (pH 6.81, 10 rnM EDTA, 0.5 rnM 1-methyl-3isobutylxanthine (MIX) in the presence or absence of 0.25 M NaCl as indicated.
The homogenates were centrifuged at 20,000 x g for 10 min at 4". The supernatant fraction was discarded, and the particulate fraction was resuspended in the respective buffer indicated and centrifuged again. All particulate fractions were resuspended (10 ml/ g of tissue) in the buffer above in the presence of 0.25 M NaCl and placed in a boiling water bath for 1 min. After centrifugation, the clear supernatant fraction was assayed for CAMP as described under "Experimental Procedures." Results are mean f S.E. for six determinations.

Removal
of CAMP from pnrtmdate fraction by addition of homogeneous bovi ne liver catalytic subunit As described in Fig. 1, approximately 600 mg of rabbit heart which had been perfused with epinephrine, frozen, and powdered was homogenized (60 ml/g of tissue) in 10 rnM potassium phosphate (pH 6.81, 10 mM EDTA, 0.5 rnM 1-methyl-3.isobutylxanthine, and 250 mM NaCl. The particulate fraction was washed three times as described in Fig. 1