Activation of cyclic AMP-dependent protein kinase isoenzymes: studies using specific antisera

A novel method for rapidly determining the amount and degree of association-dissociation of the Type I and Type II cAMP-dependent protein kinases has been developed and validated. Antibodies directed against the regulatory subunits of Type I and Type II cAMP-dependent protein kinases were used. The antibodies formed complexes with holoenzymes and regulatory subunits which were precipitated by goat anti-rabbit IgG (immunoglobulin G). These complexes bound [3H]cAMP with an apparent Kb of 20 nM for protein kinase I and 80 nM for protein kinase II. Immunoprecipitated protein kinases I and II were catalytically active when incubated with cAMP, [gamma-32P]ATP, and histone H2B. When mixtures of the two kinase isoenzymes or cytosol were incubated with various amounts of [3H]cAMP and the isoenzymes were separated by precipitation with antisera specific for each isoenzyme, the amount of [3H]cAMP associated with immunoprecipitates was proportional to the concentration of [3H]cAMP. In contrast, the catalytic activity that was immunoprecipitated varied inversely with the concentration of [3H]cAMP, showing that the activation of protein kinase could be assessed by the disappearance of catalytic activity from the immunoprecipitates. In the absence of MgATP protein kinase I was activated by a 10-fold lower concentration of cAMP than protein kinase II. However, when MgATP was added to the incubation, there was no significant difference in the binding of [3H]cAMP or dissociation of catalytic subunits of the two isoenzymes. The anti-R antibodies were also used to rapidly quantitate the concentration of regulatory subunits and the relative ratio of protein kinases I and II in tissue cytosols.     

Protein kinases in developing rat brain.

The levels of cAMP-dependent protein kinases were measured in developing rat brain by a variety of methods. The regulatory subunit (R) was measured both by [3H]cAMP binding and by 8-N3-[32P]cAMP incorporation. The catalytic subunit (C) was measured by an assay of histone kinase activity. Data were calculated per mg protein. Neither R nor C levels changed significantly in either membranes or cytosol during development. The ratio of R to C was essentially unity in the cerebra of both newborn (2-day-old) and adult (40-day-old) rats. Polyacrylamide-gel electrophoresis resolved two regulatory subunits (R-I) and (R-II) which were derived from the Type I and Type II cAMP-dependent protein kinases, respectively. 8-N3-[32P]cAMP incorporation into Proteins R-I and R-II indicated that the amounts of Proteins R-I and R-II did not change significantly in either membranes or cytosol during development.     

Induction of cAMP-dependent protein kinase I during human monocyte differentiation

Differentiation of human peripheral blood monocytes into macrophages was accompanied by induction of the regulatory subunit of cAMP-dependent protein kinase I as determined by photoaffinity labeling of cytosol proteins with 8-N3-[32P]cAMP and DEAE-Sephacel chromatography. The appearance of cAMP-dependent protein kinase I in macrophages was not due to translocation from the particulate fraction of monocytes. The regulatory subunit of cAMP-dependent protein kinase II was present in both monocytes and in vitro-differentiated macrophages. Protein kinase I in macrophages demonstrated higher affinity for 8-N3-cAMP (KD = 0.7 nM) than did protein kinase II from either monocytes (KD = 14.5 nM) or macrophages (KD = 4.9 nM). These studies demonstrate induction of the regulatory subunit of cAMP-dependent protein kinase I during the differentiation of a normal human cell and support the hypothesis that cAMP may regulate some stages of differentiation.     

Streptozotocin-induced diabetes decreases the cyclic AMP binding activity of the regulatory subunit of type I cAMP-dependent protein kinase from rat liver

Liver post-mitochondrial supernatant from diabetic rats showed a decrease in the [3H] cAMP binding activity which was associated with a decrease in the number of cAMP binding sites. On the other hand, the cAMP binding activity of nuclear fractions from diabetic rat liver was not significantly different than that of control. The cAMP binding activity of post-mitochondrial supernatant was further analyzed by using 8-azido-[32P] cAMP, a photoaffinity probe for cAMP binding sites. The diabetic supernatants showed a selective reduction in the photolabeling of a protein band representing the regulatory subunit of type I cAMP-dependent protein kinase without any appreciable change in the photolabeling of regulatory subunit of type II cAMP-dependent protein kinase.     

Human T lymphocyte cAMP-dependent protein kinase: subcellular distributions and activity ranges of type I and type II isozymes.

The role of the type I and type II protein kinase A isozymes in the regulation of human T lymphocyte immune effector functions has not been ascertained. To approach this question, we first characterized the distribution and enzyme activities of the type I and type II protein kinase A (PKA) isozymes in normal, human T lymphocytes. T cells possess both type I and type II isozymes with an activity ratio of 5.0:1 +/- 0.71 (mean +/- SD). The type I isozyme associates predominately with the plasma membrane whereas the type II isozyme localizes primarily to the cytosol. Analyses of isozyme activities demonstrated that T cells from approximately one-third of 16 healthy donors exhibited significantly higher type II isozyme activities (higher type II, type IIH) than the remaining donors (lower type II, type IIL) (mean = 605 +/- 75 pmol.min-1.mg protein-1, P less than 0.001). Scatchard analyses of [3H]cAMP binding in the cytosolic fraction demonstrated similar Kd values (type IIH, 1.1 x 10(-7) M; type IIL, 9.0 x 10(-8) M); however, the Bmax (maximal binding) of the type IIH was 400 fmol/mg protein compared to the Bmax of the type IIL of 126 fmol/mg protein. Scatchard analysis of [3H]cAMP binding to the type I isozyme associated with membrane fragments had a Kd of 5.6 x 10(-8) M and a Bmax of 283 fmol/mg protein. Eadie-Hofstee plots of type IIH and type IIL gave a Km and Vmax of 2.3 mg/ml and 1.5 nmol.mg-1.min-1, and 2.1 mg/ml and 1.6 nmol.mg-1.min-1, respectively. The 3.2-fold higher maximal binding of the type II isozyme in one-third of healthy donors may reflect a greater amount of isozyme protein. The compartmentalization of type I PKA isozyme to the plasma membrane and type II PKA isozyme to the cytosol may serve to localize the isozymes to their respective substrates in T lymphocytes.     

Characterization and subcellular localization of GnRH analog binding in rat brain

The binding of a degradation-resistant analog of gonadotropin-releasing hormone, [D-Phe6]GnRH, to rat brain crude particulate preparation was studied. The binding of this analog at 0 degrees C was saturable and Scatchard analysis revealed the presence of 2 binding sites: one with KD = 1.39 x 10(-7) M and Bmax = 265 pmole/mg protein, and another of lower affinity but higher capacity with KD = 5.58 X 10(-6) M and Bmax = 1734 pmoles/mg protein. The binding at 0 degrees C was substantially higher than that obtained at 37 degrees C, due to binding site-inactivation processes occurring at 37 degrees C. The binding sites exhibited a considerable degree of specificity for GnRH as unrelated peptides (with the exception of ACTH) display a much weaker affinity than GnRH and GnRH analogs. Subcellular fractionation demonstrated that most of the binding was associated with the mitochondrial fraction.     

A cAMP receptor from mouse liver cytosol whose binding capacity is enhanced by Mg++-ATP.

A receptor with a dissociation constant of 2·10^?6M for cyclic 3′,5′-AMP (cAMP) has been found in mouse liver cytosol. This cAMP binding activity can be differentiated from the cAMP-dependent protein kinase holoenzymes and the free regulatory subunits also found in the cytosol. Mg⁺⁺-ATP increases the number of binding sites for cAMP several fold. This increased capacity for cAMP binding persists after Sephadex G-25 filtration, and incubation for 14 hours in the presence of 5 mM EDTA. Among several adenosine- and guanosine-derivatives tested, only AMP, ADP and ATP compete efficiently with [³H] cAMP for the cAMP binding site.     

Presynaptic Cholinergic Mechanisms in the Rat Cerebellum: Evidence for Nicotinic, but Not Muscarinic Autoreceptors

The present study shows that N-[3H]methylcarbamylcholine ([3H]MCC) binds to a single population of high-affinity/low-density (KD = 5.0 nM; Bmax = 8.2 fmol/mg of protein) nicotinic binding sites in the rat cerebellum. Also, there exists a single class of high-affinity binding sites (KD = 4.8 nM; Bmax = 24.2 fmol/mg of protein) in the cerebellum for the M1 specific muscarinic ligand [3H]pirenzepine. In contrast, the M2 ligand, [3H]AF-DX 116, appears to bind to two classes of binding sites, i.e., a high-affinity (KD = 3 nM)/low-capacity (Bmax = 11.7 fmol/mg of protein) class, and a second class of lower affinity (KD = 28.4 nM) and higher capacity (Bmax = 36.3 fmol/mg of protein) sites. The putative M3 selective ligand [3H]4-diphenylacetoxy-N-methylpiperidine also binds to two distinct classes of binding sites in cerebellar homogenates, one of high affinity (KD = 0.5 nM)/low capacity (Bmax = 19.5 fmol/mg of protein) and one of low affinity (KD = 57.5 nM)/high capacity (Bmax = 140.6 fmol/mg of protein). In experiments which tested the effects of cholinergic drugs on acetylcholine release from cerebellar brain slices, the nicotinic agonist MCC enhanced spontaneous acetylcholine release in a concentration-dependent manner, and the maximal increase in acetylcholine release (59.0-68.0%) occurred at 10(-7) M. The effect of MCC to increase acetylcholine release was Ca2+-dependent and tetrodotoxin-insensitive, suggesting an action on cholinergic terminals. Also, the MCC-induced increase in acetylcholine release was effectively antagonized by dihydro-beta-erythroidine, d-tubocurarine, and kappa-bungarotoxin, but was insensitive to either atropine or alpha-bungarotoxin.(ABSTRACT TRUNCATED AT 250 WORDS)     

Evidence for a non-opioid sigma binding site in the guinea-pig myenteric plexus

The presence of a binding site to (+)-(3H)SKF 10,047 was demonstrated in a guinea-pig myenteric plexus (MYP) membrane preparation. Specific binding to this receptor was saturable, reversible, linear with protein concentration and consisted of two components, a high affinity site (KD = 46 +/- 5 nM; Bmax = 3.4 +/- 0.5 pmole/g wet weight) and a low affinity site (KD= = 342 +/- 72 nM; Bmax = 22 +/- 3 pmole/g wet weight). Morphine and naloxone 10(-4) M were unable to displace (+)-(3H)SKF 10,047 binding. Haloperidol, imipramine, ethylketocyclazocine and propranolol were among the most potent compounds to inhibit this specific binding. These results suggest the presence of a non-opioid haloperidol sensitive sigma receptor in the MYP of the guinea-pig.     

Effect of calcium on the binding in vitro of [3H]cAMP to synaptosomal membranes from rat brain

1. The binding of [3H]cAMP in vitro to synaptosomal membranes from rat brain was resolved in two components; one saturable at 20 nM cAMP with dissociation constant (KD) of 4.7 nM, and another nonsaturable within the 5-133 nM cAMP concentration range with an estimated KD value of 0.26 microM. 2. MgATP at concentration of 0.4 mM effected complete inhibition of the binding of [3H]cAMP to synaptosomal membranes throughout the used concentration range. This and the above finding indicate that the studied binding was focused on to the cAMP kinase on the membrane. 3. Calcium at concentrations of 0.1 and 10 mM stimulated a transient 20-30% increase of [3H]cAMP binding to the membranes which was influenced, as regards its time of appearance, by the concentration of cAMP. 4. The stimulation by calcium of the binding of [3H]cAMP to the membranes was inversely related to the phosphorylation of an Mr = 80,000 membrane protein, indicating stimulation of a negative effector function of cAMP--through cAMP-mediated phosphorylation--in the phosphorylation by calcium of this substrate. Moreover, the temporal displacement by cAMP of the peak of [3H]cAMP binding, produced similar temporal displacement of the inhibitory effect of cAMP on the Mr = 80,000 substrate phosphorylation. 5. These results suggest interaction in vitro of calcium and cAMP in modulation of the activity of cAMP kinase on the synaptosomal membranes.     

Parietal cell protein kinases. Selective activation of type I cAMP-dependent protein kinase by histamine

cAMP-dependent protein kinases have been characterized in parietal cells isolated from rabbit gastric mucosa. Both Type I and Type II cAMP-dependent protein kinase isozymes are present in these cells. Type II isozymes were detected in 900, 14,000, and 100,000 X g particulate fractions as well as 100,000 X g cytosolic fractions; Type I isozymes were found predominately in the cytosolic fraction. When parietal cells were stimulated with histamine, an agent that elevates intracellular cAMP content and initiates parietal cell HCl secretion, cAMP-dependent protein kinase activity was increased in homogenates of these cells as measured by an increase in the cAMP-dependent protein kinase activity ratio. Histamine activation of cAMP-dependent protein kinase was correlated with parietal cell acid secretory responses which were measured indirectly as increased cellular uptake of the weak base, [14C]aminopyrine. These results suggest that cAMP-dependent protein kinase(s) is involved in the control of parietal cell HCl secretion. The parietal cell response to histamine may be compartmentalized because histamine appears to activate only a cytosolic Type I cAMP-dependent protein kinase isozyme, as determined by three different techniques including 1) ion exchange chromatography; 2) Sephadex G-25 to remove cAMP and allow rapid reassociation of the Type II but not the Type I isozyme; and 3) 8-azido-[32P]cAMP photoaffinity labeling. Forskolin, an agent that directly stimulates adenylate cyclases, was found to activate both the Type I and Type II isozymes. Several cAMP-dependent protein kinases were also detected in parietal cell homogenates, including a Ca2+-phospholipid-sensitive or C kinase and two casein kinases which were tentatively identified as casein kinase I and II. At least two additional protein kinases with a preference for serine or lysine-rich histones, respectively, were also detected. The function of these enzymes in parietal cells remains to be shown.     

Effects of chronic nicotine treatment on nicotinic receptors in the rabbit urinary bladder

Nicotine induced a phasic contraction in the rabbit urinary bladder. The response was abolished by hexamethonium and partially reduced by atropine and capsaicin. Simultaneous atropine and capsaicin treatment did not abolish the contraction. These findings suggest that the response to nicotine is due to acetylcholine, tachykinins, and unknown mediator release. In contrast, nicotine-induced contraction diminished following the chronic nicotine treatment without a change of its pharmacological properties. These results suggest the possibility that chronic nicotine treatment causes a decrease in nicotinic receptor numbers. Therefore, the binding properties of (-)-[3H]nicotine on rabbit urinary detrusor muscle membrane fractions were studied to evaluate the effects of chronic nicotine treatment on nicotinic receptors. Specific (-)-[3H]nicotine binding reached saturation and Scatchard plots were curvilinear, suggesting the existence of two different affinity sites for (-)-[3H]nicotine. Dissociation constants (KD) and maximum binding sites (Bmax) were KD1 = 4.91 +/- 1.88 nM, Bmax1 = 2.42 +/- 0.22 fmol/mg protein and KD2 = 263 +/- 56 nM, Bmax2 = 25.0 +/- 4.3 fmol/mg protein. In urinary bladder membrane fractions from chronic nicotine-treated rabbits, KD and Bmax values were KD1 = 3.96 +/- 0.38 nM, Bmax1 = 1.07 +/- 0.25 fmol/mg protein and KD2 = 249 +/- 12 nM, Bmax2 = 10.8 +/- 1.5 fmol/mg protein. Dissociation constants for both sites following chronic nicotine treatment did not change but maximum binding site numbers for both sites significantly decreased (p less than 0.05). These results suggest that the decrease in contractile response evoked by nicotine after chronic nicotine treatment in rabbit urinary bladder is due to a decrease in numbers of nicotinic receptors.     

Comparison of adenylate cyclase and cAMP-dependent protein kinase in gametocytogenic and nongametocytogenic clones of Plasmodium falciparum

Adenylate cyclase and cAMP-dependent protein kinase activities in gametocytogenic (LE5) and nongametocytogenic (T9/96) clones of Plasmodium falciparum were compared to explore the role of cAMP in sexual differentiation of the parasite. Basal adenylate cyclase levels were equivalent in the 2 clones. However, cAMP-dependent histone II-A kinase activity was significantly higher in LE5 than in T9/96 over a range of cAMP concentrations. This difference was due to a decreased Vmax for the enzyme in the nongametocytogenic clone and not to an increased Ka for cAMP. Examination of parasite cAMP-binding proteins, likely to be kinase regulatory subunits, by both photoaffinity labeling with [32P]8-N3-cAMP and affinity chromatography of metabolically [35S]methionine-labeled cytosol of cAMP-agarose revealed a 53-kDa cAMP binding protein in both clones and a 49-kDa cAMP-binding protein in T9/96 that was absent in LE5. Our results suggest that T9/96 has lost the ability to undergo gametocytogenesis due to a substantial decrease in cAMP-dependent protein kinase activity rendering the parasite unable to respond to increased intracellular cAMP levels. Moreover, the reduction in cAMP-dependent protein kinase activity may be due to the presence of an alternative regulatory subunit of the kinase.     

Quantitative labeling of the regulatory subunit of type II cAMP-dependent protein kinase from bovine heart by a photoaffinity analog.

Several methods were compared for estimating the amount of regulatory subunit of an 800-fold purified Type II cAMP-dependent protein kinase from bovine heart. These methods included a reversable binding assay using either cAMP, or 8-N3-[32P]cAMP, photoaffinity labeling with 8-N3-[32P]cAMP, and autophosphorylation of the regulatory subunit of the enzyme. Although the regulatory subunit had a slightly lower affinity for 8-N3-cAMP than for cAMP, the total amount of regulatory subunit could be determined by each of the procedures examined. The results indicate that the photoaffinity analog 8-N3-[32P]cAMP is able to label quantitatively all cAMP-binding sites of the regulatory subunit of this cAMP-dependent protein kinase.     

Gastric H+ secretion: histamine (cAMP-mediated) activation of protein phosphorylation

Activation of H+ secretion by the gastric parietal cell involves major changes in morphology, metabolic activity and ion pathways of the secretory membrane. These changes are elicited by histamine binding to the H2 receptor, raising cAMP levels and presumably activating cAMP-dependent protein kinase. Concomitantly, the intracellular free Ca2+ concentration, [Ca2+]i, increases. Studies were performed to determine whether cAMP-mediated protein phosphorylation accompanies histamine activation of H+ secretion and to catalogue the major protein species serving as substrates for cAMP-dependent protein kinase in the parietal cell. 80% pure rabbit parietal cells, prepared by Nycodenz bouyant density centrifugation, were used. To investigate only cAMP-mediated effects, histamine-dependent changes in [Ca2+]i in these cells were abolished by depleting intracellular Ca2+ stores and performing experiments under Ca2+-free conditions. Acid secretion and steady-state levels of protein phosphorylation were then measured in unstimulated (cimetidine-treated) and histamine-stimulated cells. In intact parietal cells, concommitant with histamine stimulation of H+ secretion, increases in the level of protein phosphorylation were observed. Significantly changing phosphoproteins found in supernatant fractions showed apparent subunit sizes of approx. 148, 130, 47 and 43 kDa, and in microsomal fractions included those at approx. 130, 51 and 47 kDa. In parietal cell homogenates, using [gamma-32P]ATP, cAMP elicited significant phosphorylation of eight supernatant proteins and twelve microsomal proteins, which included the histamine-dependent phosphoproteins found in the intact parietal cell, except for the 51 kDa microsomal protein. As a working hypothesis, these proteins are involved in stimulus-secretion coupling in the parietal cell.     

Determination of the Equilibrium Dissociation Constants and Number of Glycine Binding Sites in Several Areas of the Rat Central Nervous System, Using a Sodium-Independent System

Parameters affecting the binding of [3H]glycine to membrane fractions isolated from the cerebral cortex, midbrain, cerebellum, medulla oblongata, and spinal cord of the rat were investigated in a Na+-free medium. A [3H]glycine binding assay was established in which the binding was specific, saturable, pH-sensitive, and reversible. Conditions were chosen in an effort to minimize binding to glycine uptake sites. From data on specific [3H]glycine binding Scatchard plots were prepared and the KD and Bmax values were calculated. Two glycine binding sites (high and low affinity) were identified only in the medulla (KD: 44, 211 nM; Bmax: 361, 1076 fmol/mg protein) and spinal cord (KD: 19, 104 nM; Bmax: 105, 486 fmol/mg protein). The ranges of the KD and Bmax values for the other three areas studied were 59 to 144 nM and 882 to 3401 fmol/mg protein, respectively. When the glycine content of each area, expressed as fmol/neuron, was plotted against the respective KD (high affinity), a negative correlation was found (r = --0.90; p less than 0.05). A similar negative correlation was found between the glycine content and Bmax (r = --0.88; p less than 0.05). Hill plots indicated a slope of essentially 1.0 for all areas. GABA, taurine, strychnine, diazepam, bicuculline, and imipramine had little or no effect on [3H]glycine binding.     

Cyclic AMP-dependent protein kinase in mitochondria and cytosol from different-sized follicles and corpora lutea of porcine ovaries

cAMP-dependent protein kinase was examined in mitochondria and cytosol prepared from different-sized antral follicles and corpora lutea of porcine ovaries. In all ovarian tissues examined except small follicles, protein kinase-specific activity was significantly higher in mitochondria than in cytosol, with the highest to lowest activities being found in medium (4-6 mm) follicles, large (7-12 mm) follicles, corpora lutea, and small (1-3 mm) follicles, respectively. Using the photoaffinity analogue [32P]8-N3cAMP, two major cAMP binding proteins with Mr = 47,000 (the apparent regulatory subunit of protein kinase Type I) and 54,000-56,000 (Type II) were found in all ovarian preparations. Type II was predominant in the cytosol of all ovarian samples, with the cytosolic Type I to Type II ratio increasing from approximately 0.05 in small and medium follicles top approximately 0.20 in large follicles and corpora lutea. In contrast, ovarian mitochondrial preparations contained relatively more Type I than did cytosol, with the mitochondrial Type I to Type II ratio increasing from approximately 0.50 in small and medium follicles to 0.88 in large follicles and 2.96 in corpora lutea. Also, mitochondrial [4-14C]cholesterol conversion and 3 beta-hydroxysteroid dehydrogenase/isomerase activities increased with follicle size and luteinization. These results suggest that Type I may play a role in the regulation of ovarian mitochondrial steroidogenesis.     

Activity of cAMP-dependent protein kinases and cAMP-binding proteins from rat renal cytosol upon dehydration

The activity of cAMP-dependent protein kinases, cAMP binding and the spectrum of cAMP-binding proteins in renal papillary cytosol of intact rats and of rats kept on a water-deprived diet for 24 hours were investigated. It was found that the stimulation of protein kinases by 10(-6) M cAMP in the experimental group was significantly higher than in the control one. On DEAE-cellulose chromatography, the position of peaks of the specific cAMP binding corresponded to those of the regulatory cAMP-dependent protein kinases type I and II. Under these conditions, more than 80% of the binding activity in intact animals was localized in peak II, whereas in rats kept on a water-deprived diet over 60% of the binding activity was localized in peak I. The total binding activity of cytosol in experimental animals remained unchanged is compared to intact rats. It is suggested that in renal papilla dehydration is accompanied by the induction of synthesis of regulatory subunits of cAMP-dependent protein kinase type I.     

The function of Mg-ATP in interactions between the regulatory and catalytic subunits of type I cAMP-dependent protein kinase from rabbit skeletal muscle

The regulatory subunit of Type I cAMP-dependent protein kinase from rabbit skeletal muscle can bind [3H]cAMP to form the R-[3H]cAMP complex, and the slow phase of the enhanced exchange of free cAMP with [3H]cAMP from the R-[3H]cAMP complexes was studied under various conditions using the equilibrium isotope exchange technique. Results indicate that Mg-ATP and the catalytic subunit are absolutely required for the enhanced exchange reaction to occur, but phosphorylation of the regulatory subunit by Mg-ATP does not play a determining role in the slow rate of the dissociation/association of the Type I protein-kinase in the presence of cAMP and the catalytic subunit. We interpret the role of Mg-ATP as being one in which it may provide the structural attributes required for formation of a stabilized transient state of the cAMP-regulatory subunit-catalytic subunit ternary complex, an obligatory intermediate involved in the dissociation/association of Type I cAMP-dependent protein kinase.     

Retinoic acid effect on cyclic AMP-dependent protein kinases in embryonal carcinoma cells: studies with differentiation-defective sublines

Retinoic acid induces the differentiation of PCC4.aza 1R and Nulli-SCC1 embryonal carcinoma (EC) cells. In response to retinoic acid treatment, the levels of cyclic AMP (cAMP)-dependent protein kinases are enhanced in the plasma membrane within 17 hours and in the cytosol fractions of these cells within 2 to 3 days, as determined by phosphotransferase activity and by 8-azido-cyclic [32P]AMP binding to the RI and RII regulatory subunits. PCC4 (RA)-1 and Nulli (RA)-1 are mutant EC lines that fail to differentiate in response to retinoic acid. The former line, but not the latter, lacks cellular retinoic acid-binding protein (cRABP). Basal levels of cAMP-dependent protein kinase activities are elevated in PCC4 (RA)-1 cells. When these cells are treated with retinoic acid, neither cAMP-dependent protein kinase activities nor cAMP binding activities are enhanced; rather, there is a decrease in cytosolic kinase activity and RI subunit. On the other hand, Nulli (RA)-1 cells exhibit increases both in cAMP-dependent protein kinase activities and cAMP binding in response to retinoic acid. These results raise the possibility that cRABP mediates the enhancement of regulatory and catalytic subunits of cAMP-dependent protein kinases in both the membrane and the cytosolic fractions of the teratocarcinoma cells. There also might be some effects of retinoic acid on the cAMP-dependent protein kinase that are unrelated to differentiation and to the presence of cRABP.