Cyclic AMP-binding proteins and protein kinase during regression of Walker 256 mammary carcinoma.

Cytoplasmic and nuclear CAMP-binding and protein kinase activities in NG,@‘-dibutyryl cyclic AMP (Bt,cAMP)responsive and -unresponsive Walker 256 mammary carcinomas (W256) have been studied during Bt,cAMP treatment in z&o. Total nuclear CAMP binding in the responsive tumor is increased 3-fold within 1 day after Bt,cAMP treatment. This increase in nuclear binding is accompanied by a 50% decrease in total cytoplasmic CAMP binding. The same treatment produces no change in the binding by unresponsive tumor cytosol or nuclei. The predominant species of CAMPbinding proteins which is increased in the nuclei of the responsive tumor has a sedimentation constant of 3.6 S. This peak also shows protein kinase activity which is not stimulated by CAMP but is inhibited by the inhibitor protein of CAMP-dependent protein kinase. The protein substrate specificity of this increased protein kinase is similar to that of cytosol CAMP-dependent protein kinase and distinct from that of kinase present in control nuclei. Scatchard analysis of CAMP-binding data shows that the increased binding proteins in Bt,cAMP-stimulated nuclei exhibit two types of binding: one is identical with those present in the unstimulated nuclei, and the other is similar to one of the two types of binding found in the cytosol. Cytosol from the responsive tumor exhibits three major peaks of CAMP-binding activity, sedimenting at 4.3 S, 5.6 S, and 6.9 S, respectively. The cytosol also contains CAMPindependent and -dependent forms of protein kinase, sedimenting at 3.3 S, 5.6 S, and 6.9 S, respectively. In the unresponsive tumor cytosol, the binding component of 4.3 S and the protein kinase fraction of 3.3 S decreased while the higher molecular weight species of these proteins (>7 S) increased. Dibutyryl CAMP treatment results in a shifting of the heavier binding and kinase components to their respective lighter components in the responsive tumor but not in the unresponsive tumor. The major CAMP-binding components of the responsive tumor cytosol show electrophoretic mobilities distinctive from those in the unresponsive tumor cytoso1.

YOON SANG CHO-CHUNG,+ TIMOTHY CLAIR, AND ROCHELLE PORPER From the Laborator-y of Pathoph.ysiology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 2b014 --Cytoplasmic and nuclear CAMP-binding and protein kinase activities in NG,@'-dibutyryl cyclic AMP (Bt,cAMP)responsive and -unresponsive Walker 256 mammary carcinomas (W256) have been studied during Bt,cAMP treatment in z&o. Total nuclear CAMP binding in the responsive tumor is increased 3-fold within 1 day after Bt,cAMP treatment. This increase in nuclear binding is accompanied by a 50% decrease in total cytoplasmic CAMP binding. The same treatment produces no change in the binding by unresponsive tumor cytosol or nuclei. The predominant species of CAMPbinding proteins which is increased in the nuclei of the responsive tumor has a sedimentation constant of 3.6 S. This peak also shows protein kinase activity which is not stimulated by CAMP but is inhibited by the inhibitor protein of CAMP-dependent protein kinase. The protein substrate specificity of this increased protein kinase is similar to that of cytosol CAMP-dependent protein kinase and distinct from that of kinase present in control nuclei. Scatchard analysis of CAMP-binding data shows that the increased binding proteins in Bt,cAMP-stimulated nuclei exhibit two types of binding: one is identical with those present in the unstimulated nuclei, and the other is similar to one of the two types of binding found in the cytosol.
Cytosol from the responsive tumor exhibits three major peaks of CAMP-binding activity, sedimenting at 4.3 S, 5.6 S, and 6.9 S, respectively.
The cytosol also contains CAMPindependent and -dependent forms of protein kinase, sedimenting at 3.3 S, 5.6 S, and 6.9 S, respectively.
In the unresponsive tumor cytosol, the binding component of 4.3 S and the protein kinase fraction of 3.3 S decreased while the higher molecular weight species of these proteins (>7 S) increased. Dibutyryl CAMP treatment results in a shifting of the heavier binding and kinase components to their respective lighter components in the responsive tumor but not in the unresponsive tumor. The major CAMP-binding components of the responsive tumor cytosol show electrophoretic mobilities distinctive from those in the unresponsive tumor cytoso1. * This is Paper 5 in a series on the role of cyclic AMP in neoplastic cell growth and regression from the Laboratory of Pathophysiology, National Cancer Institute, National Institutes of Health, Bethesda, Md. 20014. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom all correspondence should be addressed.
These data suggest that the nuclear accumulation of CAMP-binding proteins and protein kinase may play an important role in the CAMP-mediated control of growth and that this event may be related to the characteristics of the binding proteins and the kinase present in the cytosol. of 10 mM Tris/HCl, pH 7.5, and boiled for 10 min. After removal of particulate material by centrifugation at 16,000 x g for 10 min, activity was precipitated with 'is volume of 50% trichloroacetic acid. The precipitate was collected by centrifugation at 16,000 x g for 10 min, dissolved in water, and the pH adjusted to 7.0 with 1 N NaOH.
This fraction was dialyzed against distilled water at room temperature, and the precipitate which formed was discarded. The preparation was used at this stage of purity. Other Methods -Protein concentration was determined by the method of Lowry et al. (10). Glucose-6-phosphate dehydrogenase was assayed according to the method of Glock and McLean (11). Maintenance of Bt,cAMP-responsive and -unresponsive W256 in Sprague-Dawley female rats (3 to 4 months old, 200 g average body weight) was the same as described in the preceding paper (1). Dibutyryl CAMP (10 mg10.2 ml of 0.85% NaCl solution per rat subcutaneously1 was injected daily as previously described (12).

Effect of B&CAMP Treatment in Vivo on CAMP-binding and Protein Kinase Activities
Distribution of CAMP-binding Activity between Cytoplasmic and Nuclear Compartments -Previous studies (13) have shown that CAMP-binding activity in the cytosol of responsive W256 decreased sharply to 50% of the basal activity 30 min after a single injection of Bt,cAMP into the host animals at which time the CAMP level of the tumor reaches its peak. This depression of binding activity of the responsive cytosol could be the result of an increase in endogenous binding of unlabeled CAMP which decreases the exogenous ["HIcAMP binding (131, or may be the result of an actual loss of cytosol-binding activity due to the intracellular translocation following Bt,cAMP treatment. This possibility was examined by measuring cytoplasmic and nuclear CAMP-binding activities of tumors during Bt+AMP treatment (Fig. 1). One day after Bt,cAMP treatment total cytoplasmic CAMP-binding activity of the responsive tumor decreased appreciably, and by the 3rd day the binding activity decreased to 50% of the untreated control value. On the other hand, total nuclear CAMP binding of the responsive tumor increased markedly 1 day after treatment and binding activity was 3-fold higher than the control value after 3 days." However, no appreciable change was observed following Bt,cAMP treatment in either cytoplasmic or nuclear CAMP-binding activities of the unresponsive tumor which grows during the treatment.3 The data in Fig. 1 are given on the basis of per mg of protein in the subcellular fraction. If the binding data are expressed per unit wet weight of tumor, the total cytoplasmic CAMP binding at 3 days after the treatment remains as 50% of the control value, but the binding becomes 30% of the control value after 6 days when the protein concentration of tumor cytosol decreases to 70% that of the control value. The protein concentration of the nuclei did not decrease for 6 days after treatment. The decrease in the volume of the responsive tumor at 1,3, and 6 days post-Bt,cAMP treatment was 0, 15, and 40% of the original size, respectively (12).
Effect on Sedimentation Characteristics of CAMP-binding Proteins and Protein Kinase -The sedimentation patterns of CAMP-binding and protein kinase activities from responsive and unresponsive tumor cytosols are shown in Fig. 2. Cytosol from the responsive tumor exhibited three major fractions of CAMP-binding activity, sedimenting at 4.3 S, 5.6 S, and 6.9 S, respectively. The cytosol also contained CAMP-independent and -dependent forms of protein kinase, sedimenting at 3.3 S, 5.6 S, and 6.9 S, respectively. The latter two sedimenting constants are identical with those of heavier CAMP-binding components. Cytosol from the unresponsive tumor showed a different sedimentation pattern of CAMP-binding and protein kinase activities: the binding components sedimenting at 4.3 S and the CAMP-independent protein kinase sedimenting at 3.3 p The nuclear preparations from tumors, both control and Bt,cAMP-treated, apparently were not contaminated with nonspecifically adsorbed soluble proteins as judged by the absence of glu-. case-6-phosphate dehydrogenase activity in these preparations. s The nuclei prepared from Bt,cAMP-responsive and -unresponsive tumors, both control and Bt,cAMP-treated, were indistinguishable by phase-contrast microscopy, suggesting that the difference observed between responsive and unresponsive tumors is not due to the difference in the degree of purity of the nuclei. S decreased, while the higher molecular weight species of the binding and the kinase components (>'7 S) which were not apparent in the responsive cytosol increased. Sucrose density gradient centrifugation of the cytosols was also performed in the presence of lo-" M CAMP to examine the effect of CAMP on the sedimentation patterns ( Fig. 2, lower portion). In the presence of CAMP, the heaviest binding fraction (6.9 S) in the responsive cytosol decreased and the lightest binding peak (4.3 S) increased. Simultaneously, the heavier fraction of CAMPdependent protein kinase (6.9 S) decreased and the fraction of CAMP-independent kinase (3.3 S) increased. The shifting of the heavier binding and kinase fractions to their respectively lighter fractions was not perceptible in the unresponsive cytosol that sedimented in the presence of CAMP. These data suggest that CAMP-dependent dissociation of the holoenzyme or the inactive form of protein kinase, CR, into an active, catalytic subunit, C, and a regulatory component, R (the CAMP-binding protein) (15-21) decreased in the unresponsive tumor.
The effect of Bt,cAMP treatment on the sedimentation profiles of cytoplasmic and nuclear CAMP-binding proteins and protein kinase in the responsive and unresponsive W256 is shown in Fig. 3. Dibutyryl CAMP treatment resulted in a marked decrease in the major fractions of CAMP-binding and protein kinase activities in the cytoplasm of the responsive tumor (C). Moreover, the sedimentation profiles of the binding and kinase activities in the treated responsive cytosols (C) were different from those of the control cytosol (A), i.e. an increase in the lighter components of the binding protein and protein kinase and a decrease in the heavier fractions. These changes in the sedimentation profiles of the binding and ki- nase activities indicate a possible conversion in viva of heavier binding and kinase components to their respective lighter components (15-21) due to the increased endogenous CAMP following Bt,cAMP treatment (see Fig. 2 for CAMP effect). However, B&CAMP treatment did not affect the sedimentation profiles of CAMP-binding and protein kinase activities in the unresponsive tumor cytosol (compare E and G), despite the increased endogenous CAMP concentration in the tumor (12). The decrease in cytoplasmic binding and kinase activities in the treated responsive tumor was accompanied by an increase of these protein species into the nuclei. The nuclear extract from the treated responsive tumor (D) gave a sharp sedimentation profile composed mainly of a 3.6 S species containing both CAMP-binding and CAMP-independent protein kinase activities4 This nuclear accumulation of CAMP-binding and 4 In order to examine whether the increased nuclear peak of CAMP-binding and kinase activity may be due to an experimental artifact, the following experiments were performed: the isolated nuclei (see "Experimental Procedures") from Bt,cAMP-treated responsive tumors were added to 10 volumes of cytosol from Bt,cAMPtreated responsive and unresponsive tumors, respectively; the mixtures were rehomogenized in *lo-" M CAMP, and the nuclei reisolated, then the specific activities of CAMP-binding and protein kinase in the original nuclear extracts were compared to those in the reisolated nuclear extracts. protein kinase components was not found in the treated unresponsive tumor (H), i.e. the sedimentation profiles of these protein species in the nuclear extract were the same as those in the untreated control tumor. The sedimentation profiles of CAMP binding and protein kinase of control nuclei for both responsive and unresponsive tumors showed small broad peaks (B and F). This result presumably reflects the several species of nuclear protein kinase that have been described (22)(23)(24)(25).
The differences observed between BtcAMP-treated responsive and unresponsive tumor nuclei with respect to CAMPbinding and protein kinase activities may be related to the presence of an activator in the responsive tumor nuclei or, alternatively, an inhibitor in the unresponsive tumor nuclei. To investigate this possibility, CAMP binding and protein kinase were assayed in the mixtures of responsive and unresponsive nuclei. When mixed in a ratio of l:l, the resulting nuclear fraction exhibited these CAMP-binding and protein kinase activities predicted from the activities of the individual fractions (Fig. 4). When mixed in various ratios, the nuclear fractions showed the same binding and kinase activities as expected from an additive effect. These data, therefore, provided no evidence of the presence of an inhibitor in the unresponsive nuclei nor of an activator in the responsive nuclei.

Polyacrylamide
Gel Electrophoresis of Cytoplasmic and Nuclear CAMP-binding Proteins -The electrophoretic resolution of CAMP-binding components in the cytosol of both responsive and unresponsive W256 is shown in Fig. 5. It shows three major CAMP-binding components in the responsive cytosol and four major binding components in the unresponsive cytosol. Moreover, the two binding components separated in the unresponsive cytosol showed distinctively slower mobilities toward anode than those of the responsive cytosol. Thus cytoplasmic CAMP-binding components from responsive and unresponsive tumors are distinctive with respect to their electrophoretic mobilities.
The comparison between cytoplasmic and nuclear CAMPbinding components of Bt,cAMP-treated responsive W256 is shown in Fig. 6. The CAMP-binding components in Bt+zAMPstimulated nuclei consist of a single major species which migrates toward the anode faster than those of the cytoplasmic species. These data together with the data of sucrose density gradient centrifugation suggest the physicochemical difference between CAMP-binding proteins of cytoplasm and those of Bt,cAMP-stimulated nuclei. Scatchard Analysis of Cytoplasmic and Nuclear CAMP Binding in Control and B&CAMP-treated Responsive W256 -To compare the CAMP-binding components in the cytosol and nucleus in terms of binding affinity toward CAMP, the binding activity was measured as a function of CAMP concentration and analyzed by Scatchard plots (26). The total binding activity was measured at CAMP exchange conditions (1, 3-5) and the results are shown in Fig. 7. The cytosols from Bt*cAMPtreated and untreated tumor exhibited two major types of original and reisolated nuclear extracts suggesting that there was no acquisition of protein kinase subunits in the nuclei under the experimental conditions used. The recovery of CAMP-binding and protein kinase activities following reisolation of the nuclei was -90% in all cases. The nuclear peak of CAMP-binding and kinase activity, therefore, does not seem to be due to nonspecific adsorption of cytoplasmic protein kinase subunits. binding sites: a higher affinity binding (Kd -2 x lo-@ M) and a lower affinity binding (K, -1 x 10m7 M). The amount of total binding in both cytosols was similar. The nuclear extract from untreated tumor exhibited a single class of binding with a high affinity (K,* -3.5 x lOmy M), but the amount of total binding was about 2.5% of that in untreated tumor cytosol. Dibutyryl CAMP treatment resulted in the appearance of a new type of the binding (Kd -2.5 x 10mR M) in the nuclei. The affinity of the newly appeared binding sites in the nuclei of treated tumor is similar to that of the higher affinity binding sites present in the cytosol. The binding data suggest indirectly the transfer of cytoplasmic binding components into the nucleus following B&CAMP treatment.  control nuclear kinase activity, whereas the casein phosphorylation decreased to 50% of the control activity. Thus histone is the preferred protein substrate for both the nuclear kinase and the cytoplasmic kinase in the responsive tumor following B&CAMP treatment. Cytoplasmic and nuclear protein kinase in the control unresponsive tumor showed a similar substrate specificity as that of the control responsive tumor but CAMP stimulation of histone phosphorylation was appreciably lower than that in the responsive tumor. Furthermore, Bt&MP treatment produced no change in the protein substrate specificity of either cytoplasmic or nuclear protein kinase in the unresponsive tumor. in the nuclei of the responsive tumor, but the phosphorylation was not stimulated by CAMP. Whether this increased protein kinase is the catalytic subunit derived from CAMP-dependent protein kinase was examined using a protein kinase inhibitor prepared (2, 9) from rat skeletal muscle. Table II shows that protein kinase activity from the nuclei of BtcAMP-treated responsive tumor was significantly inhibited but the kinase of the control nuclei was not affected. The data suggest a difference in the protein kinase of the control and BbcAMP-treated nuclei of responsive W256, the former being CAMP-independent kinase and the latter CAMP-dependent kinase.

DISCUSSION
The present studies have shown that the growth response of W256 to cyclic nucleotide treatment is closely related to the response of CAMP binding and protein kinase of tumors. Little or no salt-extractable nuclear CAMP-binding and protein kinase activities can be detected in W256 nuclei before their exposure to exogenous B&CAMP. The binding protein and protein kinase appears to be present exclusively in the cytoplasm of untreated tumors. After the administration in vivo of BtcAMP a progressive increase occurs in the nuclear CAMPbinding and protein kinase activities with a concomitant decrease of these protein species in the cytoplasm. It appears, therefore, that CAMP promotes the nuclear accumulation of CAMP-binding protein and protein kinase, originally present in the cytoplasm, in much the same way as steroid hormones promote the nuclear accumulation of receptor proteins. This event of nuclear accumulation of CAMP binding and protein kinase with Bt+AMP treatment occurs only in responsive W256 which regresses following treatment. Dibutyryl CAMPunresponsive W256 which possesses a significant amount of cytoplasmic CAMP-binding and protein kinase activities fails to accumulate these protein species into the nuclei after in vivo administration of cyclic nucleotide. Although the biological relevance of the nuclear accumulation of CAMP binding and protein kinase can only be speculated at the present time, the correlation between the nuclear accumulation of these macromolecules and B&CAMP-responsive growth arrest is striking.
The same correlation was also found with other BQAMP-responsive rat tumors, i .e . dibutyryl CAMP-responsive FE3230 AC mammary tumor and 5123 hepatoma (27) and DMBA mammary tumor.s Nuclei-bound protein kinase in responsive W256 after BtcAMP treatment was found to be different from the enzymes in control (untreated) nuclei but similar to those in the cytosol of treated responsive tumor. The nuclear kinase found in the B&CAMP-treated responsive W256 showed a preference for histones as the protein substrate just as the cytoplasmic kinase, but the nuclear kinase of the control tumor (untreated) preferentially utilized casein as the substrate. The increased histone kinase of responsive tumor nuclei following Bt+AMP treatment was not stimulated further by CAMP but was strongly inhibited by the protein inhibitor of protein kinase, suggesting that the kinase is probably the catalytic subunit derived from the CAMP-dependent protein kinase (9, 28-30). Scatchard analysis (26) of CAMP-binding data shows that the increased binding proteins in the nuclei of BtcAMP-treated responsive W256 exhibit two types of binding sites; one is identical with that present in the unstimulated nuclei and the other is similar to one of the two binding sites found in the cytosol. This further suggests that CAMP stimulation causes translocation of cytoplasmic CAMP-binding proteins into the nucleus. Sucrose density gradient centrifugation and electrophoresis on polyacrylamide gel, however, showed nonidentity between cytoplasmic CAMP-binding proteins, and protein kinase and these protein species in the stimulated nuclei. One possible explanation of this apparent physicochemical difference is that the binding proteins and the kinase in the cytoplasm were modified at some step during the nuclear translocation.
Translocation of cytoplasmic CAMP-binding protein and protein kinase into the nucleus has also been suggested in other tissues. Jungmann et al. (31,32) showed that CAMP formed in a complex with a calf ovary cytosol protein (CAMP-binding protein) binds to ovary nuclei and chromatin and that concomitantly with the nuclear association of CAMP-binding protein, CAMP-independent protein kinase activity increases in the nuclei. Bhalla et al. (33) found that as a consequence of hormonal stimulation of uterine adenylate cyclase, a significant decrease of cytosol CAMP-binding and protein kinase activities occurs. This hormone-induced decrease of cytosol kinase was associated with a large increase in protein kinase activity in the nuclear 500 x g pellet. Palmer et al. (34) and Castagna et al. (35) showed that glucagon or Bt,cAMP elicit a redistribution of the protein kinase catalytic subunit from the cytoplasm to the nuclear fraction in rat liver.
The failure of CAMP-binding proteins to accumulate into the nuclei in BtcAMP-unresponsive tumors correlated with their altered physicochemical properties of the binding proteins As shown in Fig. 6

Page 3530
In the last paragraph before "Discussion," references to Fig.  6, E and F are interchanged.
Thus, line 11 from the end of the paragraph should read phoresis.
As shown in Fig. 6