CAMP mediated cell growth regulation. I. Characterization of protein kinases in CHO cells

Protein kinases in wild-type CHO cells have been characterized. Cells cultured on MEM were collected, homogenized and the extract assayed for protein kinase activity. DEAE cellulose chromatography of 30.000xg extract yields 2 peaks of protein kinases activity, PKI and PKII. The two peaks when analyzed for the binding of 8-N3-(32P)cAMP show two subunits RI and RII and a RI not associated with the enzymatic activity, named RF. This characterization allows us to discuss the meaning of protein kinases in the modulation of the growth regulating effects of cAMP.     

The causal relationship between mutations in cAMP-dependent protein kinase and the loss of adrenocorticotropin-regulated adrenocortical functions.

The Y1 adrenocortical tumor cell mutants, Kin-7 and Kin-8, harbor point mutations in the regulatory subunit (RI) of the type 1 cAMP-dependent protein kinase (cAMPdPK) that render the enzyme resistant to activation by cAMP. These mutants also are resistant to many of the regulatory effects of ACTH and cAMP. In order to examine the causal relationships between the mutations in cAMPdPK and the resistance to ACTH and cAMP, the Kin mutants were transfected with expression vectors encoding wild type subunits of cAMPdPK in order to restore cAMP-responsive protein kinase activity. The transformants then were screened for the concomitant recovery of cellular responsiveness to ACTH and cAMP. In the mutant Kin-7, cAMP-responsive protein kinase activity was recovered after transfection with an expression vector encoding wild type mouse RI. Protein kinase activity in the mutant Kin-8 remained largely cAMP-resistant after transfection with the RI expression vector but could be rendered cAMP-responsive by transfection with an expression vector encoding the wild type catalytic subunit. The recovery of cAMP-responsive protein kinase activity was accompanied by the recovery of steroidogenic and morphological responses to ACTH and cAMP, suggesting that the cAMP-dependent signaling cascade plays an obligatory role in these actions of ACTH. The growth-regulatory effects of cAMP were not reversed with the recovery of cAMP-responsive protein kinase activity, suggesting that cAMP-resistant growth regulation results from second-site, adaptive mutations either in the original Kin mutant population or in the transformants. Studies on the conversion of 22(R)-hydroxycholesterol into steroid products in parent and mutant cells indicate that the Kin mutations reduce the steroidogenic capacity of the cell as well as inhibit the hormone- and cyclic nucleotide-dependent mobilization of substrate cholesterol.     

Identification and characterization of monoclonal antibodies specific for macrophages at intermediate stages in the tumoricidal activation pathway

Macrophage activation for tumor cell killing is a multistep pathway in which responsive macrophages interact sequentially with priming and triggering stimuli in the acquisition of full tumoricidal activity. A number of mediators have been identified which have activating capability, including in particular IFN-gamma and bacterial LPS. Although the synergistic functional response of normal macrophages to sequential incubation with these activation signals has been well-established, characterization of the intermediate stages in the activation pathway has been difficult. We have developed a model system for examination of various aspects of macrophage activation, through the use of the murine macrophage tumor cell line, RAW 264.7. These cells, like normal macrophages, exhibit a strict requirement for interaction with both IFN-gamma and LPS in the development of tumor cytolytic activity. In addition, these cells can be stably primed by the administration of gamma-radiation. In the studies reported here, we have used RAW 264.7 cells treated with IFN-gamma alone or with IFN-gamma plus LPS to stimulate the production of rat mAb probes recognizing cell surface changes occurring during the activation process. In this way we have identified three Ag associated with intermediate stages of the activation process. One Ag, TM-1, is expressed on RAW 264.7 cells primed by IFN-gamma or gamma-radiation. This surface Ag thus identifies cells at the primed cell intermediate stage of the tumoricidal activation pathway regardless of the mechanism of activation. A second Ag, TM-2, is expressed on IFN-treated RAW 264.7 cells but not on RAW 264.7 cells primed with gamma-radiation alone. Expression of this Ag can be induced by treatment of irradiated cells with IFN-gamma, but is not induced by IFN-gamma treatment of a noncytolytic cell line, WEHI-3. This Ag thus appears to be an IFN-inducible cell surface protein associated specifically with macrophage activation for tumoricidal activity. Finally, Ag TM-3 is detectable on RAW 264.7 cells primed by either IFN-gamma or gamma-radiation, after subsequent triggering of the primed cells with LPS. The addition of the mAb recognizing this antigen to the function assay of tumor cell killing can inhibit they lytic activity of both triggered cells. Thus, this Ag may play a role in the antitumor effector functions of activated macrophages. Overall, the results suggest that these mAb can serve as useful tools for identification of molecules associated with the process of macrophage activation for tumor cell killing.     

Activation of cyclic AMP-dependent kinase is required but may not be sufficient to mimic cyclic AMP-dependent DNA synthesis and thyroglobulin expression in dog thyroid cells.

Thyrotropin (TSH), via a cyclic AMP (cAMP)-dependent pathway, induces cytoplasmic retractions, proliferation, and differentiation expression in dog thyroid cells. The role of cAMP-dependent protein kinase (PKA) in the induction of these events was assessed by microinjection into living cells. Microinjection of the heat-stable inhibitor of PKA (PKI) inhibited the effects of TSH, demonstrating that activation of PKA was required in this process. Overexpression of the catalytic (C) subunit of PKA brought about by microinjection of the expression plasmid pC alpha ev or of purified C subunit itself was sufficient to mimic the cAMP-dependent cytoplasmic changes and thyroperoxidase mRNA expression but not to induce DNA synthesis and thyroglobulin (Tg) expression. The cAMP-dependent morphological effect was not observed when C subunit was coinjected with the regulatory subunit (RI or RII subunit) of PKA. To mimic the cAMP-induced PKA dissociation into free C and R subunits, the C subunit was coinjected with the regulation-deficient truncated RI subunit (RIdelta1-95) or with wild-type RI or native RII subunits, followed by incubation with TSH at a concentration too low to stimulate the cAMP-dependent events by itself. Although the cAMP-dependent morphology changes were still observed, neither DNA synthesis nor Tg expression was stimulated in these cells. Taken together, these data suggest that in addition to PKA activation, another cAMP-dependent mechanism could exist and play an important role in the transduction of the cAMP signal in thyroid cells.     

Post-translational modifications of the regulatory subunits of cAMP-dependent protein kinases during G1/S progression

During the G1/S transition of the cell cycle variations in the labelling by 8-N3-[32P]cAMP of the protein kinase A regulatory subunits RI and RII, used as a probe to monitor post-translational modifications that may regulate cAMP binding, were observed in synchronized HeLa cells. A decrease in 8-N3-[32P]cAMP labelling of RI, RII and RII phosphorylated by the catalytic subunit of PKA was correlated with the increased percentage of cells in phases G1. An increase in 8-N3-[32P]cAMP incorporated into the 54-kDa RII subunit during progression from G1 to S was correlated with an increase in intracellular cAMP. A transient increase in Mn-SOD activity was detected in cells arrested at the G1/S transition using two different techniques, suggesting that oxidative modulation of regulatory subunits by free radicals may modify cAMP binding sites during the cell cycle. Decreased photoaffinity labelling by 8-N3-[32P]cAMP of RI, RII and autophosphorylated RII subunits was found to be an inherent characteristic of PKA in the G1/S transition.     

The expression of cAMP-dependent protein kinase subunits is differentially regulated during liver regeneration

The levels of the regulatory (RI and RII) and catalytic (C) components of cAMP-dependent protein kinase and of their messages were studied during the first 36 h of liver regeneration after 70% hepatectomy. Both RI alpha mRNA and RII alpha mRNA started to increase 4 h after the resection, reaching peak levels after 9 h. RI mRNA decreased abruptly 9-12 h after resection, whereas RII mRNA stayed elevated. C alpha mRNA was rather constant during the period of study. In accordance with the mRNA data the level of C was constant while RI and RII increased during the prereplicative phase of liver regeneration. RI increased rapidly when its message became elevated. RII, however, increased noticeably only 6-8 h after its mRNA had become elevated. The increased expression of R led to a disproportion between R and C that was most pronounced 14 h after resection, i.e. coinciding with the prereplicative cAMP burst. The increased R/C ratio at that time of regeneration diminished the concentration of active C subunit during the cAMP burst. In that way the otherwise inhibitory effect of high concentrations of active C on the DNA replication may have been decreased. The fractional saturation of RI and RII by endogenous cAMP fluctuated in parallel as a function of liver cAMP levels, although there was a tendency that RI was more highly saturated than RII at high concentrations of cAMP.     

Molecular basis for isoform-specific autoregulation of protein kinase A

Protein kinase A (PKA) isozymes are distinguishable by the inhibitory pattern of their regulatory (R) subunits with RI subunits containing a pseudophosphorylation P(0)-site and RII subunits being a substrate. Under physiological conditions, RII does not inhibit PrKX, the human X chromosome encoded PKA catalytic (C) subunit. Using a live cell Bioluminescence Resonance Energy Transfer (BRET) assay, Surface Plasmon Resonance (SPR) and kinase activity assays, we identified the P(0)-position of the R subunits as the determinant of PrKX autoinhibition. Holoenzyme formation only takes place with an alanine at position P(0), whereas RI subunits containing serine, phosphoserine or aspartate do not bind PrKX. Surprisingly, PrKX reversibly associates with RII when changing P(0) from serine to alanine. In contrast, PKA-Calpha forms holoenzyme complexes with all wildtype and mutant R subunits; however, holoenzyme re-activation by cAMP is severely affected. Only PKA type II or mutant PKA type I holoenzymes (P(0): Ser or Asp) are able to dissociate fully upon maximally elevated intracellular cAMP. The data are of particular significance for understanding PKA isoform-specific activation patterns in living cells.     

Cyclic AMP-dependent protein kinase type I mediates the inhibitory effects of 3',5'-cyclic adenosine monophosphate on cell replication in human T lymphocytes.

Human T lymphocytes were used as a model system to study the expression and roles of cAMP-dependent protein kinase isozymes (cAKI and cAKII) in cAMP-induced inhibition of cell replication. Human peripheral blood T lymphocytes expressed mRNA for the alpha-subforms (RI alpha and RII alpha) of the regulatory subunits of cAKI and cAKII and for the alpha- and beta-subforms (C alpha and C beta) of the catalytic subunits of cAK. At the protein level, RI alpha represented approximately 75% of the total R subunit activity, whereas RII alpha (phospho and dephospho forms) accounted for the remaining 25%. RII beta was not detected at either the mRNA or the protein level. The RI alpha protein was mainly (greater than 75%) cytosolic, whereas RII alpha was almost exclusively (greater than 90%) particulate associated. Treatment of proliferating T lymphocytes (activated through the CD3 cell surface marker) with 10 different cAMP analogs demonstrated that all inhibited cell replication in a concentration-dependent manner. The potency (as measured by the concentration giving 50% inhibition, IC50) of the cAMP analogs ranged from 30 microM for 8-chlorophenylthio-cAMP to 1100 microM for 8-piperidino-cAMP. A cAMP analog pair directed to activate cAKI (8-aminohexylamino-cAMP and 8-piperidino-cAMP) synergized in the inhibition of T lymphocyte proliferation, whereas a cAKII-directed cAMP analog pair (8-chlorophenylthio-cAMP and N6-benzoyl-cAMP) did not. We conclude that activation of cAKI is sufficient to inhibit T lymphocyte proliferation. The membrane-bound cAKII may mediate cAMP actions not related to cell replication.     

Evidence for lipopolysaccharideinduced differentiation of RAW264.7 murine macrophage cell line into dendritic like cells

Effect of lipopolysaccharide (LPS) on RAW264.7 macrophage cell line was studied. LPS-treated RAW264.7 cells increased in cell size and acquired distinct dendritic morphology. At the optimal dose of LPS (1 mg/ml), almost 70% RAW264.7 cells acquired dendritic morphology. Flow cytometric studies indicate that the cell surface markers known to be expressed on dendritic cells and involved in antigen presentation and T cell activation (B 7.1, B 7.2, CD40, MHC class II antigens and CD1d) were also markedly upregulated on LPS-treated RAW 264.7 cells. Our results suggest the possibility that LPS by itself could constitute a sufficient signal for differentiation of macrophages into DC-like cells.     

Association of type II cAMP-dependent protein kinase with p34cdc2 protein kinase in human fibroblasts

Previous independent studies suggested that type II cAMP-dependent protein kinase and the p34cdc2 protein kinase cell cycle regulator co-localize at centrosomes. In order to investigate whether there is an association of type II cAMP-dependent protein kinase with p34cdc2 in human fibroblasts, we used three different approaches. First, the regulatory subunits RI and RII were photoaffinity-labeled with 8-N3-[32P]cAMP, and anti-p34cdc2 immunoprecipitates were screened for the presence of either RI or RII regulatory subunits by one- or two-dimensional gel electrophoresis. Second, anti-RII alpha immunoprecipitates were screened for the presence of p34cdc2 by Western blot using three different affinity-purified antibodies recognizing different domains of human p34cdc2. Conversely, anti-p34cdc2 immunoprecipitates (three different antibodies), as well as the material retained on p13suc1-Sepharose Bio-Beads, which binds specifically p34cdc2, were screened for the presence of RII alpha. Finally, we have looked for cAMP-dependent protein kinase activity specifically inhibited by PKI in immunoprecipitates obtained from extracts treated with different anti-p34cdc2 antibodies. All these experiments gave concordant results and demonstrate that at least at G0/G1, human fibroblasts contain a complex of active type II cAMP-dependent protein kinase associated through its RII alpha subunit with p34cdc2.     

Modulation of macrophage function by gamma-irradiation. Acquisition of the primed cell intermediate stage of the macrophage tumoricidal activation pathway

Macrophage activation for tumor cell killing is a multistep pathway in which responsive macrophages interact sequentially with priming and triggering stimuli in the acquisition of full tumoricidal activity. Although this synergistic response of normal macrophages to sequential incubation with activation signals has been well established, characterization of the intermediate stages in this pathway has been difficult, due in large measure to the instability of the intermediate cell phenotypes. We have developed a model system for examination of macrophage-mediated tumor cell lysis, with the use of the murine macrophage tumor cell line RAW 264.7. These cells, like normal macrophages, exhibit a strict requirement for interaction with both interferon-gamma (IFN-gamma, the priming signal) and bacterial lipopolysaccharide (LPS, the triggering signal) in the development of tumor cytolytic activity. In this system, the priming effects of IFN-gamma decay rapidly after withdrawal of this mediator and the cells become unresponsive to LPS triggering. We have recently observed that gamma-irradiation of the RAW 264.7 cells also results in development of a primed activation state for tumor cell killing. The effects of gamma-radiation on the RAW 264.7 cell line are strikingly similar to those resulting from incubation with IFN-gamma, with the exception that the irradiation-induced primed cell intermediate is stable and responsive to LPS triggering for at least 24 hr. Treatment with gamma-radiation also results in increased cell surface expression of major histocompatibility complex-encoded class I antigens; however, class II antigen expression is not induced. Irradiation-induced development of the primed phenotype is not solely the result of cytostatic effects as treatment of the cells with a radiomimetic drug, mitomycin C, results in decreases in [3H]thymidine incorporation that are similar to those observed after irradiation, without concomitant development of cytolytic potential. In addition, priming by gamma-radiation does not appear to be mediated by the release of soluble autoregulatory factors. This alternate pathway for induction of the primed macrophage activation state should serve as a useful tool for identification of molecules important to the functional potential of primed cells, and for elucidation of the biochemical mechanisms of the priming event in tumoricidal activation.     

Cyclic AMP-dependent protein kinases from Balb 3T3 cells and other 3T3 derived lines

Cyclic AMP-dependent protein kinase and 3H-cAMP-binding activities were determined in normal Balb 3T3 cells and compared with the same preparations from SV40, chemical, and spontaneous transformants of 3T3 cells. The cytosolic protein kinase activities and protein kinase activity ratios were similar in all cell lines, although when the normal 3T3 cytosol was prepared by homogenization it contained less 3H-cAMP binding activity than the transformed 3T3 cytosols. The Triton X-100 treated particulate fractions from the normal and transformed 3T3 cells contained similar protein kinase and binding activities. The isozymic profile of cAMP-dependent protein kinases was examined by DEAE-chromatography. The 3T3 cells contained only type II isozyme in either cytosolic or membrane fractions. All transformants of the 3T3 cells contained both type I and type II isozymes. Other cell cultures, including chicken embryo fibroblasts, rat kidney cells, and human or calf endothelial cells contained type I and type II isozymes. Binding of the photoaffinity analogue of cAMP, 8-N3 cAMP, to the regulatory subunits of protein kinases in sonicates obtained from Balb 3T3 and SV 3T3 cells followed by separation on SDS polyacrylamide electrophoresis showed that the amount of RII subunit was approximately equal in the two cell lines. RI in Balb 3T3 cells was detectable but in a much lower quantity than in SV 3T3 cells. The cyclic AMP dependent-protein kinases from Balb 3T3 cells appears to be different from SV 3T3 cells by three criteria: 3H-cAMP binding in homogenates, DEAE chromatographic separation of isozymes, and 8-N3 cAMP binding.     

Coelution of the type II holoenzyme form of cAMP-dependent protein kinase with regulatory subunits of the type I form of cAMP-dependent protein kinase

The types and subunit composition of cAMP-dependent protein kinases in soluble rat ovarian extracts were investigated. Results demonstrated that three peaks of cAMP-dependent kinase activity could be resolved using DEAE-cellulose chromatography. Based on the sedimentation of cAMP-dependent protein kinase and regulatory subunits using sucrose density gradient centrifugation, identification of 8-N3[32P]cAMP labeled RI and RII in DEAE-cellulose column and sucrose gradient fractions by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and Scatchard analysis of the cAMP-stimulated activation of the eluted peaks of kinase activity, the following conclusions were drawn regarding the composition of the three peaks of cAMP-dependent protein kinase activity: peak 1, eluting with less than or equal to 0.05 M potassium phosphate, consisted of the type I form of cAMP-dependent protein kinase; peak 2, eluting with 0.065-0.11 M potassium phosphate, consisted of free RI and a type II tetrameric holoenzyme; peak 3, eluting with 0.125 M potassium phosphate, consisted of an apparent RIIC trimer, followed by the elution with 0.15 M potassium phosphate of free RII. The regulatory subunits were confirmed as authentic RI and RII based upon their molecular weights and autophosphorylation characteristics. The more basic elution of the type II holoenzyme with free RI was not attributable to the ionic properties of the regulatory subunits, based upon the isoelectric points of photolabeled RI and RII and upon the elution location from DEAE-cellulose of RI and RII on dissociation from their respective holoenzymes by cAMP. This is the first report of a type II holoenzyme eluting in low salt fractions with free RI, and of the presence of an apparent RIIC trimer in a soluble tissue extract.     

Isozymes of cAMP-dependent protein kinase present in the rat corpus luteum.

Regulatory (R) subunits and their association with catalytic subunits to form cAMP-dependent protein kinase holoenzymes were investigated in corpora lutea of pregnant rats. Following separation by DEAE-cellulose chromatography, R subunits were identified by labeling with 8-N3[32P]cAMP and autophosphorylation on one and two-dimensional gel electrophoresis and by reactivity with antisera. DEAE-cellulose elution of R subunits with catalytic subunits as holoenzymes or without catalytic subunits was determined by sedimentation characteristics on sucrose density gradient centrifugation and by cAMP-stimulated kinase activation characteristics on Eadie-Scatchard analysis. We identified the presence of a type I holoenzyme containing RI alpha (Mr 47,000) subunits, a prominent type II holoenzyme containing RII beta (Mr 52,000) subunits, and a second more acidic type II holoenzyme peak containing both RII beta and RII alpha (Mr 54,000) subunits. However, the majority of total R subunit activity was associated with a catalytic subunit-free peak of RI alpha protein which on elution from DEAE-cellulose was associated with cAMP. This report establishes the more basic elution position from DEAE-cellulose of the prominent rat luteal RII beta holoenzyme in very close proximity to free RI alpha and presents one of the few reports of a normal tissue containing a large percentage of catalytic subunit-free RI alpha.     

Sphingosine kinase protects lipopolysaccharide-activated macrophages from apoptosis

Lipopolysaccharide (LPS) signaling is critical for the innate immune response to gram-negative bacteria. Here, evidence is presented for LPS stimulation of sphingosine kinase (SPK) in the RAW 264.7 murine macrophage cell line and rat primary hepatic macrophages (HMs). LPS treatment of RAW 264.7 cells resulted in a time- and dose-dependent activation of SPK and membrane translocation of SPK1. Further, LPS-induced SPK activation was blocked by SPK1-specific small interfering RNA (siRNA). Overexpression of Toll-like receptor 4 and MD2, the receptor and coreceptor of LPS, in HEK 293 cells activated SPK activity in the absence of LPS treatment. Inhibition of SPK by the pharmacological inhibitor N,N-dimethylsphingosine (DMS) or SPK1-specific siRNA blocked LPS stimulation of extracellular signal-regulated kinase 1/2 and p38 but enhanced LPS-induced c-Jun N-terminal kinase activation. The SPK inhibitor DMS and dominant-negative SPK1 also blocked LPS activation of Elk-1 and NF-kappaB reporters in RAW 264.7 cells. Inhibition of SPK sensitized RAW 264.7 cells and HMs to LPS-induced apoptosis. These data demonstrate the critical role of SPK1 in LPS signaling in macrophages and suggest that SPK1 is a potential therapeutic target to block hyperimmune responses induced by gram-negative bacteria.     

Lipopolysaccharide and Raf-1 kinase regulate secretory interleukin-1 receptor antagonist gene expression by mutually antagonistic mechanisms.

Lipopolysaccharide (LPS) treatment of monocytic cells has been shown to activate the Raf-1/mitogen-activated protein kinase (MAPK) signaling pathway and to increase secretory interleukin-1 receptor antagonist (sIL-1Ra) gene expression. The significance of the activation of the Raf-1/MAPK signaling pathway to LPS regulation of sIL-1Ra gene expression, however, has not been determined. This study addresses the role of the Raf-1/MAPK signaling pathway in regulation of sIL-1Ra gene expression by LPS. Cotransfection of the murine macrophage cell line RAW 264.7 with a 294-bp sIL-1Ra promoter/luciferase construct (pRA-294-luc) and a constitutively active Raf-1 kinase expression vector (pRSV-Raf-BXB) resulted in induction of sIL-1Ra promoter activity, indicating that Raf-1, like LPS, can regulate sIL-1Ra promoter activity. An in vitro MAPK analysis indicated that both LPS treatment and pRSV-Raf-BXB transfection of RAW 264.7 cells increases p42 MAPK activity. An in vitro Raf-1 kinase assay, however, failed to detect LPS-induced Raf-1 kinase activity in RAW 264.7 cells, suggesting that in RAW 264.7 cells, Raf-1 kinase is not an activating component of the LPS signaling pathway regulating MAPK activity or sIL-1Ra promoter activity. This observation was supported by results from transfection studies which demonstrated that expression of a dominant-inhibitory Raf-1 mutant in RAW 264.7 cells does not inhibit LPS-induced MAPK activity or sIL-1Ra promoter activity, indicating that LPS-induced sIL-1Ra promoter activation occurs independent of the Raf-1/MAPK signaling pathway. In additional studies, cotransfection of RAW 264.7 cells with pRA-294-luc and increasing amounts of pRSV-Raf-BXB caused a dose-dependent inhibition of LPS-induced sIL-1Ra promoter activity, indicating that the role of the Raf-1 pathway in the regulation of sIL-1Ra promoter activity by LPS is as an antagonizer. Interestingly, LPS treatment of RAW 264.7 cells, cotransfected with pRA-294-luc and pRSV-Raf-BXB, also inhibited pRSV-Raf-BXB-induced sIL-1Ra promoter activity, suggesting that inductions of sIL-1Ra promoter activity by LPS and Raf-1 actually occur by mutually antagonistic mechanisms. In support of this conclusion, sIL-1Ra promoter mapping studies indicated that LPS and Raf-1 responses localized to different regions of the sIL-1Ra promoter. Further studies demonstrated that mutual antagonism between the LPS and Raf-1 kinase pathways is not promoter specific, as the same phenomenon is observed in assays using a c-fos enhancer/thymidine kinase promoter/luciferase construct (pc-fos-TK81-luc). Additionally, mutual antagonism with regard to sIL-1Ra promoter activity also was observed between the LPS and MEK kinase pathways, indicating that mutual antagonism can occur in more than one MAPK activation pathway.     

Cyclic AMP-dependent protein kinase in canine pancreatic rough endoplasmic reticulum

A canine pancreas homogenate was subfractionated by several differential centrifugation steps. The distribution of cAMP-dependent protein kinase in the various fractions was monitored by assaying [3H]cAMP binding and photo-cross-linking of the regulatory subunits of the enzyme (RI and RII) with radiolabeled 8-azido-cAMP. The distribution of the kinase was also compared to that of markers for the plasma membrane, the endoplasmic reticulum and the cytosol. While our results confirm previous studies suggesting the presence of cyclic AMP-dependent protein kinase in the cytosol and Golgi, a significant amount of the total [3H] cAMP binding and photolabeled R subunits (both RI and RII) were found in rough microsomes (RM). The association is relatively resistant to extraction with EDTA, low and high ionic strength solutions. These extractions unmasked several new phosphorylation substrates in the "stripped" RM that were inaccessible in the RM, possibly because they were covered by ribosomes or peripheral membrane proteins. RII with a molecular mass of 52 kDa (RII-52 kDa) was the predominant RII found in the cytosolic fraction, whereas RII-52 kDa and RII with a molecular mass of 54 kDa (RII-54 kDa) were approximately equally enriched in the RM fraction. The mobility of the RII-52 kDa-photolabeled band could be shifted to the mobility of the RII-54 kDa band by phosphorylation with purified catalytic subunit and ATP, indicating that they represent "dephospho" and "phospho" forms of RII, respectively. A more precise localization to the rough endoplasmic reticulum was accomplished by isopycnic floatation in sucrose gradients. The enzyme cobanded at the density of rough microsomes and shifted to the lower density of "stripped" microsomes after treatment with puromycin/high salt, which specifically removes ribosomes.