Comparative studies on cyclic AMP binding and protein kinase in cyclic AMP-responsive and -unresponsive Walker 256 mammary carcinomas.

N”,02’-Dibutyryl cyclic adenosine 3’:5’-monophosphate (B&CAMP) treatment in uiuo inhibits growth of one type of Walker 256 mammary carcinoma (W256) transplants (B&CAMP-responsive) but does not affect growth of other W256 transplants (Bt,cAMP-unresponsive). Cyclic AMPbinding proteins and protein kinases are present in the cytosol of both responsive and unresponsive W256 but have qualitative and quantitative differences. Scatchard plots for the binding of CAMP have been compared at pH 4.5 and 6.5. At both pH values, the responsive tumor shows two types of CAMP binding, one with a higher affinity (K, < 10es M) for the nucleotide, the other with a lower affinity (K, > 10e7 M). The unresponsive tumor shows mainly lower affinity binding (K, > 10m7 M) at pH 4.5 and higher affinity binding (K, < lo-* M) at pH 6.5. The maximum binding is significantly greater in the responsive tumor at pH 6.5 and in the unresponsive tumor at pH 4.5. Under all conditions tested, the unresponsive tumor shows a decreased binding compared to the responsive tumor. The binding equilibrium is reached faster and the peak binding lasts longer in the responsive tumor cytosol than in the unresponsive tumor cytosol. The greater instability of binding by the cytosol of the unresponsive tumor is also shown in CAMP exchange reaction carried out at 0” and 23”, respectively. Preincubation at 50” for 15 min decreases the higher affinity binding of the unresponsive tumor by about 50%, but has no effect on the binding of the responsive tumor. Protein kinase activity in the cytosol of the responsive tumor is stimulated about 5-fold by CAMP, but the enzyme in the unresponsive tumor is stimulated only 2-fold. Cyclic AMP causes a decrease in the K,,, of the responsive tumor enzyme for ATP but has no effect on the affinity of the unresponsive tumor enzyme for ATP. The catalytic properties of the kinase from both responsive and unresponsive W256, however, are similar in the absence of CAMP.

. Cyclic AMPbinding proteins and protein kinases are present in the cytosol of both responsive and unresponsive W256 but have qualitative and quantitative differences. Scatchard plots for the binding of CAMP have been compared at pH 4.5 and 6.5. At both pH values, the responsive tumor shows two types of CAMP binding, one with a higher affinity (K, < 10es M) for the nucleotide, the other with a lower affinity (K, > 10e7 M). The unresponsive tumor shows mainly lower affinity binding (K, > 10m7 M) at pH 4.5 and higher affinity binding (K, < lo-* M) at pH 6.5. The maximum binding is significantly greater in the responsive tumor at pH 6.5 and in the unresponsive tumor at pH 4.5. Under all conditions tested, the unresponsive tumor shows a decreased binding compared to the responsive tumor. The binding equilibrium is reached faster and the peak binding lasts longer in the responsive tumor cytosol than in the unresponsive tumor cytosol. The greater instability of binding by the cytosol of the unresponsive tumor is also shown in CAMP exchange reaction carried out at 0" and 23", respectively. Preincubation at 50" for 15 min decreases the higher affinity binding of the unresponsive tumor by about 50%, but has no effect on the binding of the responsive tumor. Protein kinase activity in the cytosol of the responsive tumor is stimulated about 5-fold by CAMP, but the enzyme in the unresponsive tumor is stimulated only 2-fold. Cyclic AMP causes a decrease in the K,,, of the responsive tumor enzyme for ATP but has no effect on the affinity of the unresponsive tumor enzyme for ATP. The catalytic properties of the kinase from both responsive and unresponsive W256, however, are similar in the absence of CAMP. and 10 /*l of cytosol (-70 pg of protein).
The mixture -c 1 ELM CAMP was incubated at 30 for 5 min in a shaking water bath. The reaction was stopped by the addition of 0.5 ml of ice-cold 20% trichloroacetic acid. After standing in an ice bath for 30 min, the samples were passed through Millipore filters (0.45 pm) which had been premoistened with cold 5% trichloroacetic acid. The filters were washed six times with 5 ml of cold 5% trichloroacetic acid and dried, then 5 ml of Econofluor was added and radioactivity was counted in a Beckman LS-355 liquid scintillation spectrometer with a window setting of 400 to 700. A unit of enzyme activity was defined as that amount of enzyme which transferred 1 pmol ofS"P from [yJ3PlATP to recovered protein in 5 min at 30" in the standard assay system.  (Fig. 3). Since the CAMP exchange was completed within a few hours at 23" by the cytosols of both responsive and unresponsive tumors (Fig. 21, the total CAMP binding (binding to free sites as well as to those sites endogeneously bound) was measured at 23" and compared with the binding at 0". The binding equilibrium was tested over a wide range of CAMP concentrations (lo-" -10-l" M) and at pH 4.5 and 6.5 (see Fig. 1 pared to 50% (of the peak binding) for the responsive tumor. The difference in the binding between responsive and unresponsive tumors based on pH dependence, shown in Fig. 1, was reproduced in the binding equilibrium at both temperatures. The binding in the unresponsive tumor was higher at pH 4.5 than at pH 6.5 regardless of the CAMP concentration.

Characterization
In contrast, the binding in the responsive tumor was higher at pH 6.5 than at pH 4.5 at higher CAMP concentrations of lo-" M and lo-' M; whereas at CAMP concentrations lower than lo-* M, the binding was higher at pH 4.5. Scatchard Plots for CAMP Binding-The binding components in the responsive and unresponsive tumors were further analyzed by Scatchard plots (21). Results of such experiments, compared at two temperatures and pH values, respectively, are shown in Fig. 4. In the responsive tumor, the Scatchard plots at both pH 4.5 and 6.5 showed two major types of binding sites: a higher affinity binding (K, = 1.5 x 10mH M) was predominant at pH 4.5 and a lower affinity binding (K,( = 1.3 x 10m7 M) was predominant at pH 6.5. This pH optimum for high and low binding affinity was revealed in the maximum binding capacity. The amount of higher affinity binding, estimated by extending the upper curve of the Scatchard plot to the "bound" axis, was greater at pH 4.5 while the amount of lower affinity binding (the intercept on the bound axis minus the extended intercept due to higher affinity binding) was greater at pH 6.5. The unresponsive tumor, unlike the responsive tumor, showed mainly a single type of binding. The major binding component revealed at pH 4.5 exhibited a K,, of 1.3 x lo-' M which was identical with that of the lower affinity by guest on July 8, 2020 http://www.jbc.org/ Downloaded from binding present in the responsive tumor. The upper binding curve, however, showed a small upward concavity limb which indicated a higher affinity binding (& = 3.0 x lo-' M) that became more apparent at pH 6.5. The maximum binding by both higher and lower affinity sites in the unresponsive tumor was greater at pH 4.5 than at pH 6.5. This reverse pH optimum of the lower affinity binding in the unresponsive tumor might be due to the instability of the binding at pH 6.5 (see Fig. 3). The temperature of the binding reaction did not influence the shape of the binding curves but did influence markedly the intercept on the bound axis, the maximum binding. The maximum binding in both tumors at both pH values was significantly greater at 23" than at 0".
Heat Stability of CAMP Binding-The difference in the binding stability between responsive and unresponsive tumors was further evidenced when tumor cytosols were heated at several temperatures.
The binding activities of the heated cytosols were compared with unheated controls (Fig. 5). At low (lOms M) and high (10m6 M) CAMP concentrations, binding by the responsive tumor was more stable than binding by the unresponsive tumor. Greater heat stability was shown by the responsive cytosol, particularly at 50" for 15 min, when 95% of the binding activity of the unheated control was retained; under the same conditions, the unresponsive tumor reduced its binding capacity to 50%. This occurred at the low CAMP concentration which detects high affinity binding (16). The heat stability test on mixed cytosols from both responsive and unresponsive tumors gave additive results. The heat stability of CAMP-binding proteins can be influenced by the intracellular CAMP concentrations which may have a protective effect on the binding sites during heating. However, the heat lability of the unresponsive tumor binding was not due to a lower CAMP concentration since the CAMP concentrations in both responsive and unresponsive tumors were shown to be similar (2). Moreover, 8 mM P-mercaptoethanol decreased the heat lability of the binding in the unresponsive tumor but had little or no effect on the heat stability of the responsive tumor binding.
The Scatchard plots of the binding data from heated and unheated tumor cytosols are shown in Fig. 6. In the responsive tumor, heating resulted in about a 50% loss of the maximum binding (the intercept on the bound axis), while the amount of higher affinity binding (estimated by extending the upper curve to the bound axis) showed less than a 10% loss. Thus selective stability of the higher affinity sites from the heat inactivation was found in the responsive tumor. Heat treatment in the unresponsive tumor resulted in a decrease of the maximum binding to 30% of the unheated control. The slope of the binding curve of the heated cytosol was similar to that of the upper limb of the control curve and the apparent K,l was 4.2 x 10-s M. Thus the unresponsive tumor also showed a selective stability of the higher affinity sites from the heat inactivation. Protection of the higher affinity sites from heat denaturation, however, was significantly lower in the unresponsive tumor which lost 50% of the estimated amount of higher affinity sites after heat treatment.
The results of sucrose gradient centrifugation further demonstrated the greater loss of binding components due to heat treatment in the unresponsive tumor (data not shown mammalian tissues the CAMP-binding protein is a regulatory subunit that controls the activity of a protein phosphokinase (4-10). Experiments were performed to determine whether this is the case in W256. A homogenate of responsive W256 was shown to contain specific CAMP-binding and CAMP-dependent protein kinase activities, both of which were localized predominantly in the high speed supernatant fraction (Table I)  Cell fractionation was carried out essentially by the procedure of Hogeboom (22). All fractions were washed once with homogenizing buffer and once with the buffer minus sucrose. The assays for CAMP binding (at 0", 75 min, pH 4.5) and protein kinase were performed as described in the text.

Fraction
Homogenate 755 x g pellet 15,000 x g pellet 105,000 x g pellet 105,000 x g supernatant CAMP binding Protein kinase activity pmolimg pmollfrac-, 10-Z pmoli 10-3 pmoll protein tion mg protein fraction 6.5 190 6.7 67. of responsive and unresponsive cytosols also showed the identity between the CAMP-binding proteins and the CAMP-activated kinase (24). It appears, therefore, that in both tumors, the CAMP-binding protein is associated with CAMP-dependent protein kinase. and pH -In the presence or absence of CAMP, the enzyme activity was proportional to the reaction time for 7 min and proportional to the amount of protein up to 120 pg per reaction mixture (data not shown). The pH optimum of the enzyme was about 7.5, although the activity decreased only slightly at more alkaline pH values up to 8.5 (data not shown). The effect of incubation time, amount of enzyme, and pH optimum for the kinase was similar for both responsive and unresponsive tumors. Apparent K,n for CAMP -The relationship between enzyme activity and CAMP concentration is shown in Fig. 7. From this plot (Fig. 7) and from other similar experiments, the concentrations of CAMP required to give half-maximal stimulation were 4.0 x 10mR M for the kinase of the responsive tumor and 1.5 x 10m7 M for the enzyme in the unresponsive tumor. This nucleotide exerted a cooperative effect on the kinase from the responsive tumor but not on the enzyme from the unresponsive tumor. With the kinase in the responsive tumor, reciprocal plot, as shown, is nonlinear and a Hill coefficient of 1.4 was obtained (Fig. 7). Apparent K,,, for ATP -The effect of ATP concentrations on histone phosphorylation in the presence and absence of CAMP is shown in Fig. 8. It is clear that CAMP causes a great decrease in the K,,h of the responsive tumor enzyme for ATP. Double reciprocal plots showed that the apparent K,, of the enzyme for ATP in the presence and absence of CAMP was 5.0 x 10e6 M and 1.8 x 10m4 M, respectively. In addition to the pronounced effect on the K,,, of the enzyme for ATP, CAMP also caused a 5-fold increase in the V,,,,, of the responsive tumor enzyme. On the contrary, however, CAMP did not affect the affinity of the unresponsive tumor enzyme for ATP. The K,R of the enzyme for ATP was 2.0 x 10e4 M either in the presence or absence of CAMP. The V,,, of the enzyme was increased a-fold by CAMP.  Table II. In general, the ability of a compound to inhibit the activity in the presence and absence of added CAMP was found to be parallel. ADP proved to be the most potent inhibitor of enzyme activity, causing approximately 50% inhibition at a concentration of 0.5 mM. The inhibitory action of adenosine was somewhat less and that of 5'-AMP was much weaker; ribose showed no inhibitory effect. Since the degree of inhibition in the presence and absence of CAMP was the same, the inhibitory effect could be accounted for solely by the binding of the inhibitor to the catalytic site. A similar inhibitory effect by the compounds tested was exerted on the enzymes from both responsive and unresponsive tumors.

DISCUSSION
Our present studies have revealed several differences in the properties of CAMP-binding proteins and CAMP-dependent protein kinase between Bt,cAMP-responsive and -unresponsive W256. An analysis of the CAMP binding by the two tumor cytosols produced Scatchard plots of different shapes. The responsive tumor showed higher and lower affinity binding at both pH values of 4.5 and 6.5. The unresponsive tumor showed mainly lower affinity binding at pH 4.5 and higher affinity binding at pH 6.5. Our preliminary studies (3) showed the presence of only the higher affinity binding in the responsive tumor under standard conditions (16) of binding assay. At O", the binding equilibrium in the responsive tumor reached its maximum after 16 h of assay (see Fig. 3); therefore, it seems that the lower affinity binding was not detected after 75 min of binding assay. A different interpretation is possible for the binding curve of the responsive tumor which is curvilinear with an upward concavity. Rather than attribute it to the presence of two classes of binding sites with different fixed affinities, it might be due to the presence of a single class of binding sites showing negative cooperativity.
The binding sites might be present as a homogenous class with a high affinity for CAMP ligand when unoccupied, but are then switched to a conformation with a low affinity for the ligand when occupancy increases. Such a model has been proposed for insulin-receptor interaction (25). On the other hand, both negative cooperativity and the dissociation model can give rise to a hyperbolic shape with upward concavity in the Scatchard plot, so that these two models become indistinguishable (26). Thus, despite its sensitivity, the Scatchard plot may result in an ambiguous interpretation. The difference in the binding curves between responsive and unresponsive tumors, therefore, seems to be due to the apparent difference in pH optimum or pH stability of the binding proteins rather than to the difference in the type of binding sites. The difference in pH stability of the binding proteins from responsive and unresponsive tumors was also shown by their maximum binding capacity. The maximum binding in the responsive tumor was higher at pH 6.5 than at pH 4.5; whereas the maximum binding in the unresponsive tumor was higher at pH 4.5 than at pH 6.5. Altered physical properties of CAMP-binding proteins for cultured mouse lymphosarcoma cells have also been reported by Daniel et al. (27). Cells resistant to the killing effect of B&CAMP possessed CAMP-binding proteins that have different CAMP binding and pH optimum than those in sensitive cells.
The difference of stability of the binding proteins between responsive and unresponsive tumors became quite substantial after tumor cytosol was heated. The binding by the unresponsive tumor was more sensitive to temperature than that by the responsive tumor. This correlation between the heat stability of binding proteins and "responsiveness" in vivo to exogenous B&CAMP was also noted recently in other tumor models (28). The observation that the presence of P-mercaptoethanol decreased the heat lability of the binding in the unresponsive tumor but had no effect on the stability of responsive tumor binding suggests a difference in the structure or conformation of the binding proteins between responsive and unresponsive tumors. A different temperature sensitivity of CAMP-binding proteins for cultured neuroblastoma cells has also been reported by Simantov ant! Sachs (29). Cells resistant to the cytotoxic effect of Bt,cAMP had CAMP-binding proteins more sensitive to temperature than nonresistant cells. The heat stability of the binding proteins from normal rat tissues is shown in Fig. 9. Binding proteins from brain and heart are the most heat-stable; those from intestine, uterus, and skeletal muscle are the least stable; and liver and mammary gland appear to have moderately stable binding proteins. The stabil-ity of the binding proteins of normal tissues is not related either to the amount or to the capacity of binding, with the exception of intestine which showed a low binding capacity and heat instability of the binding. Moreover, the heat instability was not due to a low CAMP concentration in the tissues nor to the enzymatic degradation of the binding protein during heating. Thus, it seems that the difference in heat stability of the CAMP-binding proteins among normal tissues is probably related to the tissue specificity of the binding proteins. The binding proteins from the responsive tumor were stable, as those of brain and heart, and the binding proteins from the unresponsive tumor were unstable, as those of intestine, uterus, and skeletal muscle.
Cyclic AMP-binding proteins are the regulatory subunits of CAMP-dependent protein kinases. The other subunits of these enzymes are catalytic ones that phosphorylate various proteins, including membrane proteins (30, 311, histones (32, 331, ribosomal proteins (34), and various enzymes (35,36). The present studies show that W256 cell populations which are unresponsive to Bt,cAMP had 2-fold lower CAMP-dependent protein kinase activity than the responsive tumor. On the other hand, the catalytic properties of the enzyme from both responsive and unresponsive tumors in the absence of CAMP were similar. The modification of CAMP-binding proteins is, therefore, probably associated with the decrease in CAMPdependent protein kinase activity of unresponsive W256.
The mechanism by which CAMP induces regression of B&AMP-responsive W256 is unknown. Dibutyryl CAMP treatment produced an increase in .cAMP concentration in both responsive and unresponsive W256, but tumor regression occurred in only one of the cell populations (2). Thus a simple endocellular change in the CAMP concentration is probably not a determining factor in the regulation of tumor growth in ho. Kuo and Greengard (5) proposed that all CAMP effects in animal cells are modified through protein kinase. Our results are consistent with this hypothesis. Yet the possibility that differences observed in kinase may be due to differences in cell populations must also be considered. Further elucidation of the causal relationship between BbcAMP responsiveness and the behavior of CAMP-binding proteins will be presented in two subsequent papers (24, 37).