The Hepatic Adenylate Cyclase System II. SUBSTRATE BINDING AND UTILIZATION AND THE EFFECTS OF MAGNESIUM ION AND pH

SUMMARY The kinetic characteristics of substrate utilization by hepatic adenylate cyclase were investigated under a variety of incubation conditions, including variations in pH, [sub-strate], [Mg”+], and in the absence or presence of glucagon. Activities were compared with ATP and 5’-adenylylimidodi-phosphate (App(NH)p) as substrates. The K, for both substrates was about 50 PM; V,,, given with App(NH)p was about 40% lower than obtained with ATP as substrate. In the presence of a saturating concentration of substrate (1 mu), basal activity was increased 4-fold by increasing [Mg’+] from 5 to 50 mM. The stimulatory effect of Mg2+ was not due to an allosteric action since basal activity was only marginally enhanced (40 %) when the substrate concentration was reduced to 10 PM. As suggested by deHaen ((1974) 249, 2’756), it is likely that Mg2+ increases enzyme activity by decreasing the concentration of an inhibitory, unchelated form of substrate that competes with the productive magnesium-substrate complex

, and in the absence or presence of glucagon. Activities were compared with ATP and 5'-adenylylimidodiphosphate (App(NH)p) as substrates. The K, for both substrates was about 50 PM; V,,, given with App(NH)p was about 40% lower than obtained with ATP as substrate.
In the presence of a saturating concentration of substrate (1 mu), basal activity was increased 4-fold by increasing [Mg'+] from 5 to 50 mM. The stimulatory effect of Mg2+ was not due to an allosteric action since basal activity was only marginally enhanced (40 %) when the substrate concentration was reduced to 10 PM. As suggested by deHaen ((1974) J. Biol. Chem. 249, 2'756), it is likely that Mg2+ increases enzyme activity by decreasing the concentration of an inhibitory, unchelated form of substrate that competes with the productive magnesium-substrate complex at the active site. Activity-pH profiles differed with ATP and App(NH)p as substrates; a shit in pH optimum was observed which correlated with the different pK, of the terminal phosphate groups of ATP and App(NH)p, and which reflect the concentration of protonated substrate (ATPH3-) present in the incubation medium. Accordingly, protonated substrate is the predominant inhibitory species of unchelated substrate and probably has a considerably higher affinity for the active site than does the magnesium-substrate complex. Glucagon-stimulated activity was less susceptible to inhibition by protonated substrate than is the basal state as evidenced by lower stimulatory effect when the [Mg'+] was increased from 5 to 20 mu. However, increasing the [Mg'+] from 20 to 50 mu resulted in marked inhibition of glucagonstimulated activity, particularly in the presence of 10 pM substrate.
Conversely, at a fixed [Mg"+], concentrations of substrate at least 20-fold higher than the K, were required to achieve maximal hormone-stimulated activity. These findings suggest that the unchelated, fully ionized form of substrate serves as an activating ligand, as has been observed with guanine nucleotides at considerably lower concentra- (cu,@Methylene)adenosine 5'-triphosphate and (a,@ methylene)adenosine tetraphosphate, which are not converted to cyclic adenosine 3',5'-monophosphate, competitively inhibited adenylate cyclase with a Ki of 40 pM and 10 paa, respectively.
The closeness of the Ki for (a, &methylene)adenosine 5'-triphosphate with the K, for ATP suggests that the Km approximates KeQ. If this is the case, a major effect of activating ligands (hormone and guanine nucleotide) is to engender a state of the enzyme that has an increased kcat, the rate constant for product formation by adenylate cyclase; this effect is accompanied by a decrease in affinity for protonated substrate.
In the preceding report (I), it was shown that guanosine 5'-triphosphates bind in their free form to a site, termed nucleotide regulatory site, on the hepatic adenylate cyclase system. Binding of the nucleotides to this site is fundamental to the process of activation of adenylate cyclase; glucagon appears to facilitate this process through its binding and action at the receptor component (l-3).
In this report we consider the interaction of substrate with the catalytic site of the hepatic adenylate cyclase system, the kinetic parameters that may be affected by the activation process, and the problems encountered in evaluating these parameters when guanine nucleotides, essential to the activation process, are not added to the assay medium.
Magnesium ion is required for catalysis of ATP to cyclic AMP' by adenylate cyclase and it is generally considered that the true substrate of the enzyme is the magnesium-substrate complex (for review see Perkins (4)). If it is assumed that the two ligands (Mg2+ and ATP) bind independently to the catalytic site, then to the extent that they compete with the productive magnesium-substrate complex, they will be inhibitory. deHaen (5) tested these assumptions and by a computer-fitting process obtained reasonably close fits with data obtained from the fat cell and ventricular adenylate cyclase systems. He concluded that free ATP (ATP4-) is a competitive inhibitor of the enzyme with binding constants approximately loo-fold higher than the productive magnesium substrate complex. He concluded also that hormones induce a conformational change in the enzyme which manifests itself in an increase of V,,,,, (about Z-fold) and a parallel release of the enzyme from a state of high to low sensitivity toward the inhibitory action of ATP*-. This was considered to be a first approximation since the theoretical curves for the hormone-stimulated states deviated significantly from the experimental points. Interestingly, the computed values for the binding constants of the true substrate were 3-to 5-fold lower for the hormone-stimulated compared to those of the basal state of the enzyme. According to these computations, hormones also engendered a decrease in the affinity of the enzyme for the magnesium-ATP complex. Two possibilities were not evaluated in deHaen's model. One possibility is that ATPH3-, referred to here as protonated substrate, is a competitive inhibitor.
The concentration of protonated substrate is determined not only by the concentrations of Mg2f and substrate, but also by pH. The other possibility is that free ATP may bind to the nucleotide regulatory site and thus serve as an activator of the enzyme, as seen with guanosine 5'-triphosphates (1). Since the concentration of free ATP is determined by the concentrations of both ATP and Mg2+, it follows that the activating effects of free ATP will be correspondingly affected by variations in the relative concentrations of substrate and Mg2+.
We have investigated the effects of varying pH and the concentrations of substrate and Mg2+ on hepatic adenylate cyclase activity in the absence and presence of glucagon. Activities were compared with ATP and App(NH)p as substrates; the latter is not hydrolyzed by nucleotide phosphohydrolases in plasma membranes (6) and has a different pK, from that of ATP at the terminal phosphate group (7). Advantage was taken of this difference in evaluating the possible role of protonated substrate as inhibitor. In what, follows it will be shown that protonated substrate is a potent inhibitor, that basal activity is more susceptible to inhibition by protonated substrate than is hormonestimulated activity, and that free substrate mimics the activating effects of guanine nucleotides at the nucleotide regulatory site. A complex kinetic pattern develops as a function of pH and the concentrations of substrate and Mg2+. Shown here are the appropriate conditions for obtaining V,,, and K, for the basal and hormone-induced states of the enzyme system. clear Corp. ATP, GTP, cyclic AMP, creatine phosphate, creatine phosphokinase, and dithiothreitol were purchased from Sigma. Ap(CHz)p, Ap(CHz)pp, and Ap(CH,)ppp were from Miles Laboratories. Crystalline porcine glucagon was a gift from Eli Lilly and co.

Preparation of Hepatic Plasma
Membranes-Partially purified plasma membranes, prepared from rat liver by a modification of the procedure of Neville (8) as previously described (9), were stored in liquid nitrogen. The adenylate cyclase activities of the plasma membrane preparations varied; therefore, all experiments for direct comparison were performed with the same preparation on the same day. Protein concentration was determined according to Lowry et al. (10) with bovine serum albumin as standard.
Assay for Adenylate Cyclase Activity-Assay medium consisted of [aJ2P]ATP or [aJ2P]App(NH)p (30 to 500 cpm/pmol) at concentrations indicated in figures and tables, 5 mM MgC12 or as indicated, cyclic AMP (2070 of the substrate concentration), 40 mM Tris-HCl, pH 7.6, or as stated, 3 mg/ml of bovine serum albumin, and membranes, in a final volume of 100 ~1. Before addition to the incubation medium, hepatic membranes were suspended in Tris-HCI buffer containing 1 mM dithiothreitol and equilibrated for 1 min to the assay temperature (30"). Where ATP was used as substrate, an ATP-regenerating system consisting of 5 mM creatine phosphate and 50 units/ml of creatine phosphokinase was included. For the pH study, Tris-acetate buffers were used. When indicated, 1 PM glucagon was included in the assay medium.
Cyclic AMP was determined by a recently developed procedure (11). All data were obtained from linear initial rates of cyclic AJvIP formation determined by following the time course of the reaction. All experiments were repeated at least twice, the values reported are the average of duplicate samples which agreed within 5yo in a single experiment.
Calculations-The concentrations of protonated substrate ( Fig.  1) were determined from the stability constants reported for AT?' (12) and App(NH)p (7), and the pK, of the terminal phosphates of these nucleotides (7). These values for the two forms of substrate (8) were incorporated into a computer program designed to calculate the concentrations of MgS2-, Sa-, and HS3-formed over a pH range of 6.0 to 9.0 in the presence of nucleotide (ATP or App(NH)p) over a concentration range of 10~~ to Lomb, and with Mgz+ concentrations ranging from 1 to 100 mM. RESULTS pH Dependence of Hepatic Adenylate Cyclase Activity-Changes in concentrations of hydronium ion can influence several factors which may affect the activity of hepatic adenylate cyclase, including substrate ionization, the catalytic process, environment of the active site, glucagon binding and action in the case of hormone response, and enzyme structure. The fact that the pK, of the terminal phosphate differs markedly for ATP and App-(NH)p (7.1 and 7.7, respectively (7)) presented an opportunity to study the effect of substrate ionization on enzyme activity at a given pH such that all other factors are affected equally. In solution with moderate excess of Mg*f, substrate (S) is distributed mainly among three species: MgS2-, HS3-, and S4-. Based on the stability constants for MgATP" and MgApp(NH)p2- (7,12), it was calculated that in the presence of 100 pM substrate and 5 mM Mg*f, the concentration of MgS2-is essentially constant (86 to 99 PM) as is the concentration of S4-(1.0 to 1.16 PM) over a pH range of 6.0 to 9.0. By contrast, the concentration of HS3-changes by about lOOO-fold (from 13 PM to 15 nM) over the same pH range (see dashed line in Fig. 1). At a given pH, the concentration of App(NH)pH3-is about twice that of ATPH3-. Thus, by using either ATP or App(NH)p as substrate, it is possible to introduce different concentrations of protonated substrate into the incubation medium under otherwise identical conditions (same enzyme preparation, pH, and concentrations of Mg2+ and substrate). This procedure is justified further by the fact that ATP and App(NH)p have the same affinity for the active site (see Table II site of the enzyme can be compared primarily as a function of the concentration of protonated substrate. Fig . 1A illustrates the pH activity profiles of basal and glucagon-stimulated activities with ATP as substrate (pH range, 6.0 to 8.0) ; the pH activity profiles with App(NH)p are shown in Fig. 1B over the pH range 7.0 to 9.0. Basal activity with either substrate did not display a definitive pH optimum; activities were essentially constant over the pH range 8.0 to 9.0. The major difference in pH effects on basal activity with the two substrates was observed below pH 7.5. Thus, at pH 7.0 basal activity was reduced from its maximum by 50% with ATP but by about 85% with App(NH)p as substrate. A more dramatic difference in the effects of pH was observed with glucagonstimulated activity which displayed a more definitive pH optimum. The optimum pH with ATP was about 7.5; with App-(NH)p, the pH optimum was shifted to around 8.3. The shift is more obvious when activities with the two substrates are compared at a pH giving 50% that observed at the respective pH optimum for ATP and App(NH)p; 50% of maximum is given at pH 6.5 and 7.5 with ATP and App(NH)p, respectively. The shift in pH optima is somewhat greater than the difference in pK, of ATP (7.1) and App(NH)p (7.7). However, inaccuracies in the determination of the precise pH optima could account for the difference.
It will be noted in Fig. 1 that activities given with App(NH)p at its optimal pH were about 50% given with ATP as substrate (see also Table II). If it is assumed that the catalytic process is pa-dependent and that the pH optimum is around pH 7.5, then the shift to a higher pH optimum with App(NH)p due to the higher concentration of protonated substrate could result in lower apparent catalytic efficiency over the entire pH range. Another possible explanation for the different apparent efficiencies in catalysis is that PPi and P(NH)Pi, formed by catalysis of ATP and App(NH)p, respectively, have different binding constants at the catalytic site; a slower rate of dissociation of P(NH)Pi from this site could result in apparent differences in catalyt'ic efficiency. However, this would not explain the differing pH profiles given with the substrates.
The data in Fig. 1 suggest that binding of protonated substrate at the catalytic site is an important controlling factor in the activity of the hepatic adenylate cyclase system. In addition to pH, the concentration of protonated substrate is determined by the relative concentrations of magnesium ion and substrate present in the incubation medium.
E$ects of Mg2+-It has been postulated that magnesium ion serves as a regulatory ligand at an allosteric site (13) in addition to its role in forming the productive magnesium-substrate complex. This possibility was tested by increasing the concentration of Mg2+ in the incubation medium and examining the effects of the cation on hepatic adenylate cyclase activity as a function of increasing substrate concentration.
It was reasoned that stimulatory effects of Mg2f at a putative regulatory site should be observed independently of the concentration of substrate. However, as shown in Table I, the stimulatory effects of magnesium ion seen for basal activity were dependent markedly on substrate concentration.
Thus, increasing the concentration of Mg2+ lo-fold in the presence of 10 PM App(NH)p as substrate resulted only in a 40% increase in basal activity whereas activities were increased 3-and 4-fold when the concentration of substrate was increased to 0.1 and 1.0 mM, respectively. Therefore, an allosteric action of Mg2+ IS unlikely. Since the stimulatory effects of the cation are seen at concentrations far in excess of that required to form the productive Mg2+-substrate complex, it is likely that stimulation of basal activity is due to increased chelation of uncomplexed, inhibitory forms of substrate, as suggested by deHaen for other adenylate cyclase systems (5). AS shown above, the inhibitory species is protonated substrate, the concentration of which is reduced lo-fold by increasing [Mg'+] from 5 to 50 mM.
The effects of [Mg2+] on glucagon-stimulated activity were more complex (Table I). For exa.mple, at 5 mM Mg2+, increasing the concentration of substrate from 10 to 1000 PM increased the foldness of hormone-stimulated activity over that of basal activity; foldness was reduced by increasing the concentration of Mg2+ to 50 mM which, furthermore, resulted in inhibition of hormone-stimulated activity with 10 PM substrate. Inhibition alter the concentration of free Mg2+ at 50 mM, resulted in a 7-fold increase in glucagon-stimulated activity. It is likely, therefore, that glucagon-stimulated activity is dependent on the concentration of free substrate and that the inhibitory effects of Mg2+ are due to titration of free ATP. In this regard, the preceding study (I) showed that the free form of guanine nucleotides is required for binding to the nucleotide regulatory site; activation of the hepatic system by guanine nucleotides is facilitated by glucagon.
From the above findings, the effects of Mg2+ on adenylate cyclase activity can be attributed to chelation of substrate to form the productive Mg2+-substrate complex, and the titration of the inhibitory protonated substrate and the activating free form of substrate. Titration of the protonated substrate by Mg2+ appears to have a greater effect on basal activity than on the hormone-stimulated state as evidenced by the marked enhancement of basal activity with increasing concentrations of Mgzf in the presence of 1.0 mM ATP; this is not seen with glucagon (Table I).
Kinetic Constants with ATP and App(NH)p-The activating and inhibitory effects of protonated and free forms of substrate obviously may complicate evaluation of the kinetic constants for Mg*f-substrate complex at the catalytic site. Fig. 2 shows double reciprocal plots of kinetic data obtained when the hepatic enzyme system was incubated with ATP as substrate in medium containing 5 mM Mg*f, ATP-regenerating system, at pH 7.5. The curves obtained were nonlinear, particularly that for basal activity. These curves were obviously too complex for determining K, and V,,,,, for the basal and hormone-stimulated states of the enzyme.
Kinetic studies were conducted subsequently with the enzyme incubated in the presence of 50 mM Mg2+ at pH 8.5 (Fig. 3B).  I  I  I  I1  I  I  I  I  I  2  4  6  6  IO  12  14  16 16 20 A+p mM+ At pH 8.5, the concentration of protonated substrate is lo-fold less than that at pH 7.5 (Fig. 1); the concentrations of free and protonated substrate are decreased lo-fold by increasing the Mg2+ from 5 to 50 mM. We used App(NH)p rather than ATP as substrate in these experiments since the presence of the ATPregenerating system could be avoided. Creatine phosphate in the regenerating system, although forming weak complexes with Mg2+ (14), nevertheless can alter the concentration of Mg2+ and, therefore, the concentrations of unchelated species of substrate. Indeed, creatine phosphate inhibits adenylate cyclase activity (15), possibly because of resultant increases in concentration of protonated substrate. Since increasing the [Mg"+] also reduces the concentration of free activating nucleotide. evaluation of the kinetic parameters for the activated state of the enzyme was carried out in the presence of saturating concentrations of glucagon (1 FM) and 0.1 rnM Gpp(NH)p as activator. conditions which cause maximal activation of the hepatic adenylate cyclase system (1). Linear double reciprocal plots were obtained for both basal activity and the activated state (Fig. 3B). Fig. 34 shows for comparison the same membrane preparation incubated under identical conditions except that the [Mg"+] was reduced to 5 mM as a means of increasing the concentration of protonated and free forms of the substrate. The basal curve was nonlinear which possibly reflects the stimulating effects of free substrate at the nucleotide regulatory site. Addition of both glucagon and Gpp(NH)p abolished this deviation from linearity. Extrapolation of the linear curves (Fig. 3B) to the abscissa revealed that the K, for the productive Mg2+ -substrate complex was the same (50 PM) for both the basal and activated state; V,,,,, was increased 4-fold by glucagon and Gpp(NH)p. Thus, the activating ligands engender a state having a lower affinity for protonated substrate, an increase in Vmax, without changing the apparent affinity for the productive form of the substrate.
EfSects of ~ho~phonu~e Analogues of ATP-AptCH,)pp cannot be converted to cyclic AMP by adenylate cyclase and inhibits competitively interactions of either ATP (A) or App(NH)p (B) with the enzyme as shown in Fig. 4. The fat cell adenylate cyclase system is also competitively inhibited by Ap(CHJpp (16). In the presence of 50 mM Mg2+, linear double reciprocal plots were obtained with either ATP or App(NH)p at their pH optimum (pH 7.5 and 8.5, respectively). The kinetic constants are summarized in Table II for the productive substrates and the inhibitory analogues. As discussed previously, the V,,,,, was about 40% lower with ApptNH)p compared with ATP as substrate; both forms of substrate gave essentially the same K,. The methylene analogue of ATP, Ap(CH*)pp, gave a K; nearly equivalent to the K, for the two substrates. Surprisingly, the analogue of adenosine 5'-tetraphosphate, Ap(CHJppp, was the strongest inhibitor; it appears that the active site can accommodate an extra phosphate without reduction in affinity. The inhibition constants were calculated on the assumption that the analogues have the same stability constants with Mg2+ as those of the productive substrates (ATP and App(NH)p) and that the Ki reflect the binding of the magnesium complexes rather than the unchelated forms of the nucleotide. Inhibition by the analogues is not likely due to their contribution as protonated sub- basically different from that drawn by deHaen (5), the exception being that his calculations were made for free ATP which cannot be distinguished from protonated substrate with data obtained from assays conducted at a fixed pH.
Free ATP appears to be an activator of adenylate cyclase as judged by the marked dependence of glucagon-stimulated activity on the concentration of ATP and the inhibitory effects of Mg2+ on this activity when the concentration of the cation is far in excess of substrate. Similar inhibitory effects of Mg2f on glucagon-stimulated activity have been reported previously for the hepatic adenylate cyclase system (9). In the presence of 5 mM Mg2+, 2 mM ATP was required for maximal hormonestimulated activity; this concentration is at least 3 orders of magnitude higher than that required with GTP or Gpp(NH)p (1 PM) which also activate the enzyme in their free form (1). We have suggested (2) Fig. 2) as the substrate concentration groups is required for strong binding to the catalytic site or that was increased; this is due to the relatively high concentration of Ap(CH2)p forms a much weaker complex with Mg2+ at the unchelated substrate introduced into the incubation medium.
catalytic site.
Since basal and hormone-stimulated activities are preferentially affected by protonated and free substrate, respectively, it was not feasible to determine the kinetic parameters (K, and Vmax) DISCUSSION for the two states of the enzyme system. In this regard, it is of Just as Gpp(NH)p has proven to be a useful analogue of GTP interest that deHaen (5) calculated that the binding constants for investigating the role of the nucleotide regulatory site in the for Mg2+-substrate complex at the adrenocorticotropic hormoneactivation of adenylate cyclase (1,17), App(NH)p has proven induced and basal states of the fat cell adenylate cyclase system to be valuable for investigating the characteristics of the catalytic were about 1.0 mM and 0.17 mM, respectively; a similar reduction site. In addition to its resistance to hydrolysis by potent nucleo-in substrate affinity was calculated for the epinephrine-stimulated tide phosphohydrolases in plasma membranes (6), which allows state of the ventricular enzyme system. Our studies would sugassays to be conducted in the absence of ATP-regenerating sys-gest that these calculated values are erroneous and that the terns, App(NH)p, with its different pK, from that of ATP, pro-apparent lowering of the affinity of substrate for the hormonevided a means of showing that the protonated form of substrate induced states actually reflects the high concentrations of actiis a potent inhibitor of the hepatic adenylate cyclase. The same vating free substrate required for hormone-stimulated activities conclusion has been drawn from similar studies of the fat cell in the absence of guanine nucleotides. As reported elsewhere adenylate cyclase system.2 Since the concentration of protonated (17), the nucleotide regulatory site is an ubiquitous feature of substrate is very low when the [Mg2+] is in excess of substrate, adenylate cyclase systems in eucaryotic cells irrespective of the its affinity for the active site must be several orders of magnitude hormone-receptor component; this site is fundamental to the higher than that of the productive magnesium complex of sub-process of activation of adenylate cyclase in the presence or abstrate, which the present study shows to be about 50 PM. In an sence of hormones (1,3). accompanying report (3)) computer-fitting analysis of the various In summary, this study shows that activating nucleotides and kinetic parameters of substrate binding and utilization indicates glucagon engender a presumed conformational change at the that the Ki for protonated substrate is in the nanomolar range. active site which manifests itself in a decrease in affinity for This extraordinary sensitivity to inhibition by protonated sub-protonated substrate and a 4-fold increase in I',,,. Quantitative strate explains why changes in pH and Mg2+ have profound analysis of the kinetic parameters at the active site and evidence effects on enzyme activity even when the concentration of Mg2f for a three state model are provided in an accompanying report is far in excess of substrate concentration.
. It is perhaps significant with regard to the increase in Vmax One of the conclusions of this study is that the basal state of that the apparent K, for Mg2+-substrate complex is close to the the enzyme is more sensitive to the inhibitory effects of proton-Ki for the nonproductive phosphonate analogue, Ap(CH2)pp; ated substrate than is the activated state formed by the actions this finding suggests that the K, is equivalent to K,,. If this of glucagon and activating nucleotides. This conclusion is not should prove to be the case from direct binding studies, then activation of adenylate cyclase involves an increase in koat, the * M.  J. Biol. Chem., in press. rate constant for product formation.