Multiple inhibitory and activating effects of nucleotides and magnesium on adrenal adenylate cyclase.

Adenylate cyclase in particulate fractions from rat adrenal glands is subject to regulation by purine nucleotides, particularly guanine nucleotides. While GTP activates the enzyme, this effect is not evident in all particulate fractions. Following dialysis of the refractory fractions activation by GTP is observed, an indication that endogenous nucleotides may obscure the effects of added GTP. The analog, guanyl-5'-yl imidodiphosphate (Gpp(NH)p gives considerable more activity than does GTP. GDP, on the other hand, is inhibitory, an effect revealed only in the absence of a nucleotide-regenerating solution. GDP blocks the action of both GTP and Gpp(NH)p. These results show that the gamma-phosphate of the nucleotide is required for but need not be metabolized in the activation process. At low substrate concentration (0.1 mM ATP or adenyl-5'-yl imidodiphosphate) stimulation of the enzyme by ACTH occurs only in the presence of added guanine nucleotide (GTP or Gpp(NH)p); the hormone and nucleotide act synergistically. While both GTP and Gpp(NH)p inhibit fluoride-stimulated activity, the level of fluoride required to demonstrate such inhibition appears not to be related to the level of fluoride required for activation of the enzyme. In the presence of GTP, or GTP plus ACTH, the enzyme exhibits normal Michaelis-Menten kinetics with respect to substrate utilization (K-m equal to 0.16 mM). In the activated state, produced with ACTH plus GTP, the enzyme is less susceptible to inhibition by a species of ATP uncomplexed with Mg2+, but is more susceptible to inhibition by Mg2+. These results demonstrate that fundamental differences exist between different states of the adenylate cyclase. The difficulties in describing kinetically the regulation of adenylate cyclase systems in view of the multiple actions of nucleotides and magnesium are discussed.

Adenylate cyclase represents a class of a complex regulatory enzyme systems that are responsible for the production of cyclic AMP,' the so-called "second messenger" of hormone action. Poorly defined in physicochemical terms, the enzyme systems are present in plasma membranes and are thought to consist of regulatory units bearing hormone-specific receptors and catalytic units containing the active site of the enzyme (1). Little is known of the mechanism by which hormones increase enzyme activity, although it is generally agreed that hormones induce maximal activity without changing the affinity for the substrate, MgATP (cf. Ref. 2). This concept has been contested by de Haen in a report on the kinetic analysis of two adenylate cyclase systems in which he concludes that hormones decrease the affinity for substrate (3).
Studies of the glucagon-sensitive hepatic adenylate cyclase system revealed that a nucleotide, preferentially GTP, acti- vates the enzyme system by a concerted or interdependent manner with glucagon (4, 5). The nucleotide also changes the binding of glucagon to its receptor from a slowly reversible to a rapidly reversible process (6). GTP appears to act on the catalytic unit at a site distinct from the active site of the enzyme (5) and inhibits the stimulatory effects of fluoride ion, a potent stimulator of adenylate cyclase systems in eukaryotic cells (7). The stimulatory effects of GTP are mimicked by ITP and ATP and have been observed with a number of adenylate cyclase systems irrespective of the nature of the receptor coupled to these systems (8-16).
We have examined the ACTH-sensitive adenylate cyclase system from rat adrenal glands to determine whether this system also exhibits a nucleotide requirement, and for comparison with other adenylate cyclase systems currently under investigation in this laboratory. The results show that GTP or, at much higher concentrations, ATP, is required for hormone action. Gpp(NH)p, an analogue of GTP that is resistant to nucleotide phosphohydrolase action. also activates the enzyme system in the absence or presence of ACTH, yielding activities higher than observed with GTP. The guanine nucleotides Twenty adrenal glands were enucleated, and the pooled capsular and decapsulated portions were processed separately. ATP concentration was 0.5 mM and GTP was 10 1~. Membrane protein per assay for Fraction A was 12 pg; B, 11 pg; C, 10 pg; D, 18 pg.
Adrenal fraction -ACTH + ACTH -GTP +GTP -GTP +GTP pmol cyclic AMP formedlmg protein/5 min inhibit fluoride-stimulated activity in a noncompetitive manner. ACTH appears to enhance the adrenal adenylate cyclase activity by a process that increases the formation of states of the enzyme that are susceptible to activation by GTP, Gpp(NH)p, and at much higher concentrations, ATP. The major effects of the regulatory ligands (hormones and nucleotides) appear to be an increase in u,,., a decrease in the inhibitory effects of a species of ATP uncomplexed with Mgz+, and an increased sensitivity to inhibition by Mg*+. For some experiments the supematant fraction of the 600 x g centrifugation was centrifuged at 10,000 x g, the pellet saved, and the supernatant centrifuged at 100,000 x g to obtain a microsomal pellet. All pellets were suspended and stored as described above. The enzyme was stable indefinitely when stored under liquid N,. When in solution the enzyme was stabilized with 1 mM dithiothreitol.
With the exception of the experiments described in Table  I

Mammalian
cells are known to contain GTP in concentrations up to 10% of that of ATP (20-22), and the latter is thought to be in the millimolar range (21, 23). Since GTP in the micromolar range produces maximal effects on adenylate cyclase activity (see below), low levels of contaminating endogenous nucleotide may be expected to obscure the effects of added nucleotide.
An example of this phenomenon is seen in Table I, which shows the distribution of adenylate cyclase activity in particulate fraction of adrenal glands divided into capsular and decapsulated portions. Assays were performed with 0.5 mM ATP, a substrate level which permitted assessment of the effects of both GTP and ACTH (see Fig. 2). While the hormone enhanced activity in all fractions, the response to GTP was variable; the nucleotide considerably enhanced both basal and ACTH-stimulated activities in the 600 x g fraction of the capsular gland, but had no effect on activity in the 10,000 x g fraction of the decapsulated gland. In order to test whether this lack of a GTP effect may have resulted from contamination by endogenous nucleotide, the 10,000 x g fraction of the decapsulated gland was dialyzed against 300 volumes of 20 mM Tris-HCl, (pH 7.6)-l mM dithiothreitol, at 0" for 4 hours. Following such treatment, the dialyzed fraction exhibited a sensitivity to GTP comparable to that observed with the 600 x g fraction of the capsular gland which served as the control and was unaffected by dialysis. Thus, contaminating endogenous GTP (or other activating nucleotides) may affect the apparent distribution of enzyme activity among the various particulate fractions, as well as obscure the effects of added nucleotide.
Since we wished to investigate the role of GTP in the actions of ACTH on the adrenal system, it was obvious that the 10,000 x g fraction was unsuitable for such purposes. Accordingly, the 600 x g fraction, which displayed marked sensitivity to nucleotides and the highest specific activity of the various fractions investigated was used in all studies. Since preliminary studies revealed no qualitative differences between the 600 x g fractions of capsular and decapsulated glands, experiments were conducted with 600 x g fractions derived from homogenates of whole adrenal glands.* This avoided the *As with any tissue the possibility exists that the adrenal contains more than one adenylate cyclase system. A problem in interpretation of data would result if, in the absence of hormone, the activity measured were that of an enzyme different from the ACTH-stimulated enzyme.
We  Substrate was 0.1 mM App(NH)p. GTP, GDP, and Gpp(NH)p were present at 10 PM. Membrane protein was 9.5 pg per assay. of fluoride-stimulated activity is described below. Note the minimal activation by ACTH in the absence of guanine nucleotides.
The lowest concentration of GTP that produced an observable effect was 10 nM, and half-maximal activation of basal and ACTH-stimulated activities occurred with 0.2 PM GTP.
At concentrations above 10 pM the nucleotide inhibited activity in the absence (see also Table IV) and presence of ACTH.
As the data in Fig. 1A show, the hormone and GTP acted synergistically to increase cyclic AMP production, i.e. with maximal ACTH and GTP, activity was twice the predicted sum of the activities observed with each agent alone.
While the 3-fold stimulation of activity by GTP ( Fig. 1A) may have resulted from interaction of the nucleotide with membranes that bore residual native ACTH, studies with membranes from animals bearing pituitary tumor transplants indicate otherwise.
The blood levels of ACTH in the tumorbearing animals may be 1000 times normal (17), and pairs of adrenals from these animals weighed over 500 mg, which is 10 times normal. However, when compared directly, both normal and hypertrophied adrenal preparations exhibited equivalent sensitivity to ACTH; half-maximal activity for both was achieved with 80 nM ACTH.
If retention of ACTH on the membranes were a problem, one would have expected an apparent decrease in sensitivity to added ACTH with membranes from the hypertrophied adrenals. Evidence in support of the notion that GTP acts as a regulatory ligand and not as a phosphorylating agent has been reported for the glucagon-sensitive hepatic adenylate cyclase In Table II it is seen'also that the addition of the regenerating system resulted in considerable inhibition under all conditions. Further studies have revealed that this inhibition is due primarily to creatine phosphate, as has been reported for the fat cell enzyme (13 Nucleotides-In order to test the effects of GDP on the adrenal adenylate cyclase system, assays were conducted in the absence of the nucleotide-regenerating system and with the use of App(NH)p as substrate. This was necessary since it has been shown that the regenerating system under conditions employed in these experiments efficiently converts GDP to GTP.3 The data in Table II show that GDP, at 10 PM, inhibited basal activity by approximately 40%, and, in contrast to GTP and Gpp(NH)p, permitted only minimal activation by ACTH. However, with the addition of the regenerating system no differences were seen between the effects of GDP and GTP, reflecting the rapid conversion of GDP to GTP. These data, showing that the regulatory site of the adrenal enzyme system exhibits a requirement for the y-phosphate of the activating nucleotide, are in contrast to information published with nearly all other adenylate cyclase systems (see "Discussion").
analog, Gpp(CH,)p, mimicked the actions of GTP (5). A with substrate. comparison of the effects of GTP and another analog, Cyclic GMP and guanosine over the concentration range of 1 nM to 10 pM had no effect on the enzyme. With App(NH)p as from the zona fasciculata. In comparative studies, no detectable substrate, 5'-GMP was also without effect. On the other hand, differences between normal adrenal and zona fasciculata fractions were with ATP as substrate, and in the presence of the regenerating observed with regard to the relative effects of the different guanine nucleotides or on the differential effects of Mg'+ and ATP in the system, 10 PM 5'-GMP stimulated both basal and ACTHabsence or presence of ACTH. 'Y. Salomon, unpublished observations. stimulated activitn to the extent seen with a half-maximal level of GTP. These effects of 5'-GMP may be explained, however, by the presence of contaminating enzymes in the creatine phosphokinase preparation which convert GMP to GDP in the presence of ATP, and the subsequent conversion of GDP to GTP.L Effects of Combinations of Different Guanine Nucleotides-It was of interest to determine the effects of combinations of the different guanine nucleotides (GDP, GTP, Gpp(NH)p) on the adrenal enzyme system. The effects of GDP on activation by GTP could not be determined since in the presence of the regenerating system GDP is converted to GTP However, the effects of GDP on stimulation by Gpp(NH)p were studied with App(NH)p as substrate (no regenerating system, Table III), as were the effects of GTP on stimulation by Gpp(NH)p with ATP as substrate (with regenerating system, Table IV). The data in Table IV show that submaximal concentrations of GTP inhibited the effects of Gpp(NH)p (10 PM), and that when the two nucleotides were present at equimolar concentration the Gpp(NH)p action was completely inhibited; activity was equivalent to that seen with GTP in the absence of Gpp(NH)p.
The data in Table III show that GDP, while inhibiting basal activity, inhibited Gpp(NH)pstimulated activity to a greater extent, i.e. inhibition by GDP was proportionally greater with Gpp(NH)p than in the absence of the stimulating nucleotide.
These results suggest that the three guanine nucleotides act at the same site on the enzyme system, and that the y-phosphate of the nucleotide is required for but need not be metabolized in the process of enzyme activation.
Other Nucleotides-A comparison of the effects of GTP and of several other nucleoside triphosphates on basal and ACTHstimulated activities is shown in Table V. The two pyrimidine nucleotides tested, CTP and UTP, had little or no effect at 10 PM. lTP, a purine nucleotide, was stimulatory, but to a lesser extent than was GTP. As is shown below (Fig. 2), ATP in the Inhibition of Gpp(NH)p effect by GTP Substrate was 0.1 mru ATP, and the nucleotide-regenerating system was present. Membrane protein was 12 ag per assay.
No additions 10 PM GPPCWP millimolar range appeared also to act at the regulatory site of the enzyme system. These findings are consistent with other studies on adenylate cyclase systems that show, in general, that purine nucleotides are considerably more effective than are pyrimidine nucleotides (5, 12, 16). Actions of Fluoride-In the presence of 5 mM MgCl,, half-maximal activation was seen with approximately 5 mM NaF. If, however, MgCl, and NaF were combined 4 hours prior to the initiation of the reaction with membranes, fluoride was inhibitory at concentrations above 5 mM. Such inhibition may be accounted for by the removal of MgZ+ by fluoride; the solubility product constant of these two ions is 1O8.2 (25). As is seen in Fig. 1, both GTP and Gpp(NH)p inhibited fluoride-stimulated activity.
From studies on the hepatic enzyme it was concluded that GTP and fluoride acted noncompetitively since increasing the fluoride concentration did not overcome inhibition by GTP (5). We have tested the ability of GTP to inhibit activity in the presence of varying amounts of fluoride and found that, in general, inhibition was more prominent at high concentration of fluoride (20 mM) than at concentrations below 10 mM. Moreover, with a submaximal fluoride concentration (5 mM), GTP actually enhanced fluoride-stimulated activity. These data support the notion that fluoride and GTP act noncompetitively, and show that the concentration range of fluoride required to activate the enzyme is unrelated to the fluoride concentration required to elicit inhibition by GTP. High levels of fluoride may increase susceptibility to inhibition by the nucleotide by forming a complex with Mg*+, and, thus, changing considerably the assay medium composition (see "Discussion").

Effects of GTP on Substrate
Utilization-In order to determine whether GTP affects interaction of enzyme with substrate, basal and hormone-stimulated activities were tested with and without the nucleotide over the ATP concentration range of 0.025 to 5.0 mM in the presence of 10 mru Mg*+ (Fig. 2). This Mg*+ concentration was employed to minimize levels of uncomplexed ATP. In addition to serving as substrate, ATP appeared to fulfill a second role in this system, in that, like GTP, it permitted activation by the hormone when present in sufficient concentrations; at the lowest ATP concentrations tested activation by ACTH was minimal in the absence of GTP, but at the higher ATP concentrations activation by the hormone no longer required GTP. Such results probably reflect interaction of ATP (or contaminating GTP, Ref. 4) at the nucleotide-sensitive regulatory site on the enzyme system, a phenomenon observed also with the hepatic adenylate cyclase system (5).
Lineweaver-Burk plots of the data from Fig. 2 show that in the presence of GTP, basal and ACTH-stimulated activities exhibited no difference in their affinity for substrate (K, = 0.16 mM), and that the primary action of ACTH was to increase V max (Fig. 3). Curves derived from experiments conducted in the absence of GTP were nonlinear, being concave downward, and did not permit calculation of a K, value. A possible explanation for the nonlinear curves would be activation by substrate, and, as noted, ATP did appear to serve as a regulator as well as substrate.
In addition to its action as substrate and as possible regulatory ligand, a third effect of ATP may be observed in Fig.  2; at the highest ATP concentration tested (5 mM), activity in the absence and presence of GTP (no hormone) was inhibited, while ACTH-stimulated activity seemed not to be affected. The differential effects of excess ATP on activity in the presence and absence of ACTH are discussed below. Effects of ATP on Basal and ACTH-stimulated Activities-de Haen has reported recently (3) that ATP uncomplexed with Mg2+ serves as a competitive inhibitor of the binding of the productive form of substrate, MgATPZ-, to the active site of the fat cell and ventricular adenylate cyclase systems. He concluded also that the hormone-stimulated states of the enzyme were less susceptible to inhibition by uncomplexed ATP than the basal states of these enzyme systems. Since Mg*+, at a fixed concentration of substrate, alters the concentration of uncomplexed ATP, it should follow from de Haen's findings that decreasing the concentration of Mg2+ from 10 mM, as in Fig. 2, to 2 mM should increase the concentration of uncomplexed ATP and should cause greater inhibition of basal activity than of ACTH-stimulated activity over a comparable range of substrate concentrations. As shown in Fig. 4, activities obtained in the presence of ACTH (with or without GTP) did not differ substantially with 2 mM Mg2+ compared with 10 mM Mg*+ (Fig. 2) over the range of 0.1 to 1.0 mM ATP. By contrast, basal activity with 2 mM Mg2+ was reduced substantially compared to that observed with 10 mM Mg2+ over the same range of substrate concentrations. It should be noted also that the inhibitory effects of decreasing Mg2+ concentration cannot be related simply to the complexing of Mg2+ by ATP since, as shown in Fig. 2, inhibition of activity was evident even when Mg*+ was present in 5 mM excess over ATP. These data provide, therefore, qualitative support for the hypothesis put forth by de Haen; namely, that hormonal activation of adenylate cyclase results in the formation or an enzyme state that is less susceptible to inhibition by uncomplexed ATP.
Also seen in Fig. 4 is the inhibition of all activities when the ATP concentration approached, but not necessarily exceeded, that of magnesium. However, even under conditions of excess ATP, inhibition of basal or GTP-stimulated activity was proportionally greater than that seen in the presence of ACTH.
Another interesting aspect of the experiments depicted in Fig. 4 was the wide variation, dependent upon ATP concentration, of the ratios of ACTH-stimulated to basal activities. For example, at 0.1, 1.0, and 2.0 mM ATP, the ratios of hormonestimulated to basal activities were, respectively, 3, 7, and 11 in the presence of GTP, and 1.5, 7, and 15 in the absence of the guanine nucleotide. Such data indicate that ratios of hormone to basal activities do not serve as reliable indicators of receptor function, or of the ability of a system to respond to hormone. In most of the experiments presented in this report, activities with ACTH were usually only 2-to 4-fold greater than basal activities. However, experiments were conducted under conditions of relatively high Mg*+:ATP ratios, and under such conditions basal activities may be relatively high due to relief from ATP inhibition.
Effects of Mg*+-The above data indicate that basic differences exist between basal and hormone-activated states of the adrenal enzyme with respect to inhibition by a form of ATP uncomplexed with Mg*+. Studies on the adenylate cyclase system from liver (26) and thyroid (27) indicate that when activated by hormones these enzymes may be more sensitive to inhibition by high Mg2+ concentrations than they are in the basal state. The adrenal adenylate cyclase system may be affected similarly by Mga+. Fig. 5 shows the effects of varying the MgZ+ concentration from 1 to 40 mM on ACTH-stimulated activity when tested at three ATP concentrations in the presence and absence of GTP. Under conditions where a minimal nucleotide effect was manifested (no GTP, 0.1 mM ATP), varying the Mg*+ concentration from 1 to 40 mM resulted in a small elevation of activity; activity under these conditions was essentially basal activity despite the fact that ACTH was included in the assay medium. Increasing ATP to 0.5 mM partially fulfilled the nucleotide requirement for ACTH action (see Fig. 2), and under these conditions increasing the Mgz+ concentration produced a noticeable inhibition (Fig. 5). At 2.5 mM ATP, a level sufficient to satisfy the nucleotide requirement, inhibition by Mgl+ was marked. In Panel B of Fig. 5 it is seen that in the presence of GTP, Mg*+ concentrations above 10 mru were inhibitory at all concentrations of ATP tested.
One explanation for these data would hold that nucleotide (ATP or GTP) uncomplexed with Mg2+ is required by the regulatory site, and that inhibition by Mga+ was the result of a lowering of the concentration of the unbound nucleotide. However, in further experiments we found the apparent K, for Mga+ not to vary over a wide range of GTP concentrations. Further studies have revealed that the GTP-or Gpp(NH)pactivated enzyme, even in the absence of ACTH, is more cation must be taken into consideration in any formal description of the adrenal adenylate cyclase system. DISCUSSION Our results show that the y-phosphate of the guanine nucleotide is important for activation of the adrenal adenylate cyclase system. These data are in contrast to those published for several other adenylate cyclase systems showing that nucleotide mono-and diphosphates, as well as the triphosphates, serve as enzyme activators (5,12,(14)(15)(16). However, such studies were done in the presence of both ATP and a nucleotide-regenerating system, conditions which permit conversion of the mono-and diphosphate forms to the nucleotide triphosphates. In studies done in the absence of a nucleotideregenerating system, and with App(NH)p as substrate, the hepatic glucagon-sensitive enzyme was reported to be activated by GDP and GTP (5). More recent studies, however, have revealed that the liver enzyme is inhibited by GDP in the absence of glucagon and only weakly activated by GDP in the presence of hormone. Moreover, as is shown herein for the adrenal enzyme, the actions of Gpp(NH)p on the hepatic enzyme are blocked by GDP.' While the fat cell adenylate cyclase appears to discriminate somewhat between GDP and GTP, a direct comparison with the adrenal enzyme is difficult given the multiple effects of guanine nucleotides on the fat cell enzyme (13).
Although the y-phosphate of the guanine nucleotide is important for enzyme activation, the evidence does not indicate that a phosphorylation reaction occurs. This is supported by the finding that Gpp(NH)p is more active than GTP and by previous studies with the hepatic adenylate cyclase system showing that Gpp(CHJp substitutes for GTP (5). The possibility that GDP inhibits by interacting with the nucleotide regulatory site suggests a possible explanation for the differences between levels of activation achieved with GTP and Gpp(NH)p. If the regulatory site were to possess phosphohydrolase activity, the action of GTP, as opposed to that of Gpp(NH)p, would be limited by its hydrolysis to GDP. Studies on the mode of GTP interaction must await separation of the enzyme from nucleotide phosphohydrolases present in plasma membranes.
In providing a kinetic analysis of two adenylate cyclase systems de Haen has contributed significantly to the understanding of adenylate cyclase regulation (3). Our data support qualitatively his hypothesis that hormones render these enzymes less susceptible to inhibition by a species of uncomplexed ATP. de Haen concluded also that hormones decrease the affinity for substrate, MgATP; in the case of the fat cell enzyme ACTH shifted the K, upward by greater than 5-fold. We were able to calculate the K, for the adrenal enzyme by adding GTP to satisfy the nucleotide requirement at the regulatory site, and by performing experiments in the presence of a relatively high magnesium concentration (10 mM) to overcome inhibitory effects of uncomplexed ATP (Fig. 3). These experiments produced no evidence for an ACTHinduced decrease in the affinity for substrate. However, in the absence of GTP, ATP in excess of 3 mM was required to satisfy the nucleotide requirement of the adrenal enzyme; the calculated increase in K, of the ACTH-activated fat cell enzyme may reflect the need for higher ATP to fulfill the regulatory susceptible to inhibition than is the basal enzyme (data not 'Y. Salomon, M. C. Lin, C. Londos, M. Rendell, and M. Rodbell, shown). As is discussed below, such inhibitory effects of the J. Bid. Chem., in press. tides. As has been shown for the fat cell (13) and adrenal enzymes, the creatine phosphate, a constituent of the regenerating system, is an inhibitor. Thus, further studies would require a determination of the inhibitory species of creatine phosphate (free or magnesium-complexed). The availability of AP~WUP and GPPWWP, nucleotide phosphohydrolaseresistant nucleotides which serve as substrate and regulatory nucleotide, respectively, permit studies in the absence of a regenerating system. An analysis of the adrenal enzyme with the use of these compounds will be forthcoming.
Acknowledgments-We gratefully thank our colleagues Drs. M. C. Lin, B. R. Martin, M. Schramm, and Y. Salomon for their participation in the discussions of this study and for their many helpful suggestions.
function. de Haen noted systematic deviations between the theoretical curves and experimental points in his report (3). It would appear that the failure to consider the regulatory action of ATP contributed to these deviations.
In formulating a kinetic model it is essential, of course, to take into account all possible effects of Mg*+. de Haen has considered two actions: formation of substrate (MgATP) and alteration of the level of uncomplexed ATP, an inhibitor. Our data reveal another effect of magnesium, inhibition by the cation per se. The latter effect was particularly evident when the enzyme was activated by hormone. It is likely that yet another effect of magnesium requires attention, that is, the effect of the cation on the level of activating nucleotide. We have shown recently that it is the free, and not the magnesiumcomplexed form of the nucleotide, that activates the glucagonsensitive hepatic adenylate cyclase system. l Thus, by decreasing the level of uncomplexed nucleotide, increasing the Mg2+ concentration removes a factor (free nucleotide) which is both an inhibitor and an activator. However, with the adrenal enzyme the inhibition resulting from the removal of activating nucleotide is obscured by the inhibition of the cation per se; the latter occurred even in the presence of an excess of GTP.
The current literature provides little opportunity for comparing data on substrate utilization by the adrenal enzyme with other adenylate cyclase systems. In none of the published reports on the kinetics of substrate utilization were experiments performed in the presence of nucleotides other than ATP (cf. Refs. 2 and 3). As we have shown, only in the presence of GTP does the adrenal enzyme exhibit typical Michaelis-Menten kinetics. Since it may be assumed that most, if not all, eukaryotic adenylate cyclase systems are subject to regulation by purine nucleotides (5,(8)(9)(10)(11)(12)(13)(14)(15)(16)28), the question arises as to how assessments of K, were made in view of the likelihood that ATP served both as substrate and activator. One possible explanation is provided by our finding that as little as 0.2 mg/ml of a 10,000 x g adrenal fraction appeared to contain a sufficient level of contaminating nucleotides to satisfy the regulatory site. It is reasonable to suggest that some studies have been conducted, unknowingly, in the presence of activating nucleotides, and that the nucleotide source was the enzyme fraction itself. For instance, in studies on the renal medullary vasopressin-sensitive enzyme, Neer found no evidence for a regulatory role of ATP (2). However, those studies were performed with a membrane fraction enriched in both plasma membranes and mitochondria. The latter may have been a nucleotide source in Neer's studies, as well as in our studies on the 10,000 x g adrenal fraction.
The kinetic model of de Haen (3) is, as he states, a useful first approximation in describing the behavior of adenylate cyclase. However, it is clear that the multiple effects of magnesium and ATP, described above, as well as the effects of guanine nucleotides, constitute the minimal elements that should be incorporated into any further kinetic models for the enzyme.
Experiments designed to provide data for kinetic analysis of membrane-bound adenylate cyclase systems with the use of ATP and GTP are complicated by the need to add a nucleotide-regenerating system, as membrane-bound nucleotide