THE Joualva~ OF Bro~cmn~ C~ewsrsr Vo1.246. No.6.Iylue of March 25, pp. lsS'I-1869,1971 Printed in U.S.A. The Glucagon-sensitive Adenyl Cyclase System in Plasma Membranes of Rat Liver II. COMPARISON BETWEEN GLUCAGON- AND FLUORIDE-STIMULATED ACTIVITIES* (Received for publication, October 7. 1970) LUTZ BIRNBAUMER, STEPHEN L. POHL, AND MARTIN RODBELL From the Section on Membrane Regulation, National Institute of Arthritis and Metabolic Diseases, National In- stitutes of Health, Bethesda, Maryland 2001.$ SUMMARY Glucagon and fluoride ion stimulate the activity of a com- mon adenyl cyclase system in plasma membranes isolated from rat liver. Their actions are noncompetitive indicating that they act at separate sites in this system. Manganous ion, above 5.0 mM, inhibited selectively the response of the enzyme to glucagon. Inorganic pyrophosphate (1.5 mu), on the other hand, inhibited the response of the enzyme to fluoride but enhanced the response to glucagon. The in- hibitory effect of pyrophosphate was noncompetitive with fluoride ion. Highly purified phospholipase A caused a selec- tive loss of the glucagon response and enhanced the stimula- tory effect of fluoride ion. Treatment of liver membranes with digitonin also caused a selective loss of hormone re- sponse. Inactivation of glucagon response by digitonin was not restricted to the liver membrane adenyl cyclase system; incubation of ghosts of fat cells with digitonin resulted in loss of response of the adenyl cyclase system to glucagon, adreno- corticotropin, secretin, and epinephrine at very low concen- trations of the detergent. Digitonin enhanced the response of the fat cell system to fluoride ion. Sodium dodecyl sulf- ate, over a narrow range of concentrations, inhibited the re- sponse of liver membrane adenyl cyclase to glucagon, and enhanced the enzyme's response to fluoride ion. Other detergents caused a parallel loss of the response to both fluoride ion and glucagon. The present findings suggest that glucagon and fluoride ion activate adeny cyclase by different mechanisms. It is pos- sible that these agents react through different molecular entities in this complex enzyme system. It has been postulated (2) that, animal adenyl cyclase. systenls contain distinct molecular components, termed "discriminators," which react spectifically with hormones and through which the catalytic component, adeny cycls~e, is activated. This pos- tulate is based in part from the observation that a single adenyl * Some of the studies have been reported in preliminary form (1). cyclase enzyme in fat cells (or their "ghosts") of the rat is AC- tivated by glucagon, secretin, adrenocorticotropin, epinephrinr, thyrotropin, and luteinizing hormone through sites that are specific for each of the hormones (3, 4). Fluoride iou, which activates all adenyl cyclase systems in eucnr-otic* cells (5, 6), also activates the hormone-sensitive adenyl ryclase system in fat cells (7). Since fluoride ion appears to act nonspecifically on adenyl cyclase systems, it is likely that the halide ion does not activate the system through the homlont,-discrinlillators but possibly through some direct action on thr catalytic cow ponent. The purpose of this paper is to esta.blish that glucagon and fluoride ion activate the same adenyl cyclase system in liver plasma membranes and do so by different mechanismr. Such information not only helps bo establish the complexity of the adenyl cyclase system in these membranes, but also assists in efforts to isolate and characterize the fluoride-seusitivr com- ponent and thediscriminator for glucagon. EXPERIIUESTAL PROCEDURE All experiments using plasma membrane> were carried out with the partially purified preparation described in the preceding report (8). Preparation of isolated fat cells and fat cell ghosts from rat adipose tissue are described elsewhere (7, 0). Highly purified phospholipasc .4 (EC 3.1,1.4), (Y type (IO), 1400 units per mg, was generously supplied by Dr. Xchael .\. Wells (Lni- versity of Arizona, Tucson). Digitonin, from LIann, ww recrystallized txvice from ethanol before it was used. EGTh* was obtained from Sigma. The sources of :111 other chrrniwls have been described previously (7, 8). The incubation mixture for determining adeny cycla-c :lc- tivity in liver membranes and fat cell ghosts contained, iu 0.05 ml volume, the following: 3.2 mhf .%TP-&*P (30 to 50 cpm per pmole), 5.0 mnf MgCl*, 25 nw Tris-HCI, pH 7.6, 20 mN c'wntine phosphate, 1 mg per ml of creatine kinase (20 to 50 unit: pcl mg), 1 nlM EDT& and between 20 and 40 pg of liver membranes or' 50 pg of fat cell ghost protein. The reaction wan stolqn~d as described before (7) and cyclic 3',5'--4111' formed was isolated and determined according to the method of IGGhn:L~ Wei+, and Brodie (11). `The abbreviations used are: EGTA, ethyleue glycol his@- aminoethyl ether)-iV,A:`-tetraacetic acid; cyclic AMP, cyclic 3',5'-monophosphate. 1857 1858 Glucagon-sensitive Adenyl Cyclase System. II Vol. 246, Xo. 6 In all figures and tables, adenyl cyclase activity refers to nanomoles of cyclic 3',5'-AMP formed in 10 min per mg of protein. RESULTS Glucagon and fluoride ion act on the same adenyl cyclase system in rat liver plasma membranes. This was established by the experiment shown in Table I in which maximal stimu- lating concentrations of fluoride ion (15 rn$ and glucagon (10 pg per ml) were added to the adenyl cyclase incubation medium either alone or in combination. Combination of the two agents failed bo stimulate the activity of the enzyme more than glucagon alone. The lack of competitive interaction between the two agents suggested that fluoride ion and glucagon activate the enzyme through different sites. This finding permitted an evaluation of the possible different characteristics of the fluoride- and glucagon-responsive components contained within the same enzyme system. were stimulated. However, at concentrations higher than 5.0 mM, bin++ inhibited selectively the response of the enzyme to glucagon. Selective stimulation of the fluoride response has been observed also in the fat cell ghost adenyl cyclase system (7). Xlg+f acted somewhat differently than JIn++ in that it stimulated, in the presence of 1 mn< EDT;& the response to glucagon over a narrow range of concentrations, and caused parallel loss of basal, fluoride-, and glucagon-stimulated nc- tivities at higher concentrations (see Fig. 6 in the previous report 03)). Efecfs of Inorganic Yyrophospha!e-Pgrophosphate, at 1.5 mM, inhibited the response of the liver enzyme to fluoride ion by 76y0 whereas it stimulated the response to glucagon by 30';, as shown in Fig. 2. The selective inhibitory effect of pyropho- phate on the fluoride response was noncompetitive since doubling the concentration of fluoride ion failed to reduce the inhibitor? effect of pyrophosphate. k'fleects of Divalent Zons--Adenyl cyclase systems require for their activity a divalent cation, and both hlg++ and Jln++ support enzymatic activity. This was first shown for dog brain ndenyl cyclase (5) and later for a variety of adenyl cyclase sys- tcnis (6). The effect of varying concentrations of &In++ on .glucagon- and fluoride-stimulated activities in liver plasma membranes is shown in Fig. 1. At concentrations of Mn++ below 5.0 rnhl, both the response to glucagon and to fluoride ion T.IBLE I j? EBpct of glucagon andjluoride ion on adenyl cyciase activity in liver s plasma membranes - Additions Adenyl cyclase activitp I - .-- 0 0.1 0.2 0.3 DIW TONIN ( Ye ) Glucagon (10 pg per ml). . . . . . , . . . 4.28 f 0.23 NO. 4. r;ttect of digitonin on response of adenyl cyclase activity NaF (15 mM) . . . . . . . . . . . . . . 2.47 f 0.12 in liver plasma membranes to glucagon and fluoride ion. Liver Glucagon (10 pg per ml) i- NaF (15 mM) . . . 4.10 f 0.15 plasma membranes (0.6 mg per ml) were incubated for adenyl cyclase activity in the presence of either 20 rg per ml of glucagon a Values are mean + half the range of triplicate determinations. (0) or 10 mM VaF (A) and the indicated concentrations of digi- tonin. Incubations were carried out for 10 min at. 30". PPi (mM1 PHOSDHOLIPASE A t U/m,, FIG. 1 (161). Effect of varying concentrations of MnClr on adenyl cyclase activit.y determined in the presence of either 10 rg per ml of glucagon or 10 rnM NaF. EDTA and MgClz were omitted from the standard assay medium. FIG. 2 (center). Effect of varying concentrations of inorganic pyrophosphate on adenyl cyclase activities determined in the pres- ence of 1Org per ml of glucagon (0) or 10 rnM NaF (A). EDTA was omitted from the standard assay medium. FIQ. 3 (right). Effect of treat,ment of liver plasma membranes with phospholipase A on the response of adenyl cyclase to glu- cagon and fluoride ion. Liver plasma membranes (3.5 mg per ml) were incubated in a medium containing 1 mM CaCl*, 50 rnM Tris- HCl, pH 7.6, and the indicated concentrations of phospholipase -4. Incubations were carried out for 10 min at 30" and were terminated by addition of 2.0 mM EGTA. Then, lo-p1 aliquots of these mis- tures were analyzed for adenyl cyclase activitr in the presence of either 10 rg per ml of glucagon (0) or 10 ml! kaF (A). Issue of March 25, 1971 L. Birnbaumer, S. L. Pohl, and 211. Rodbell 1859 XO- 6.0 - 0.16 0.60 Di& sow ( % 1 FIG. 5. Effect of digitonin on response of adenyl cyclase in fat cell ghosts to hormones and fluoride ion. Fat cell ghosts (1.0 mg per ml) were incubated in adenyl cyclase assay incubation medium (7) containing the indicated concentrations of digitonin. Ac- tivity was determined in the absence (Control, C) and the presence of either 10 rn~ NaF (F) or 10 rg per ml of epinephrine (E), secre- tin (S), ACTH (A), and glucagon (G). Incubations were carried out for 10 min at 30". Ejfects of Phospholipnse A-We have reported previously (12) that t,reatment of liver plasma membranes with heat-treated snake venom, rich in phospholipase A activity, results in selective loss of the glucagon response and stimulation of the response of the enzyme to fluoride ion. In Fig. 3, it is seen that highly purified phospholipaae .4 has the same effects. In these esperi- merits, membranes were treated for 5 min at 30" with phos- pholipase A, followed by the addition of EGT.4, a calcium chelator, t,o stop the action of the calcium-dependent enzyme (13). Additionof the ehelatorpriorto that of the enzyme resulted in complete inhibition of the effects of phospholipase A4 on the response of adenyl cgclase to glucagon and fluoride ion. This is additional evidence that the observed effects were due to the enzymatic action of phospholipase A. Eflects of Detergents--The effects of phospholipase -4 suggested that lipids play a role in the act,ivation of adenyl cyclase by glucagon and fluoride ion. This possibility was investigated further by testing the effects of a variety of detergents. As shown in Fig. 4, digitonin (a neutral detergent) caused selective loss at concentrations less than O.l%, of the response of liver adenyl cyclase to glucagon. Selective inact,ivation by digitonin of the response of adenyl cyclasc to glucagon was not restricted to the plasma membranes of rat liver. In Fig. 5, it can be seen that digitonin also inhibited t,he response of the fat cell ghost ndenyl cyclase system to glu- cagon as well as to secretin, adrenocorticotropin, and epineyh- rine . As was observed with liver membranes treated with phospholipase A, digitonin stimulated the response of the fat cell ghost system t.o fluoride ion. This effect of digitonin was not observed with t,le liver membrane system. In experiments not shown, sodium dodecyl sulfate, at O.OOS%l,, stimulated the fluoride response in liver adenyl cyclase system by 2-fold and caused the loss of 40c0 of the respon,x t.o glucagon. Higher concentrations of the detergent caused complete loss of both activities. Triton X-100 and sodium deosycholate in- hibited the response to both fluoride ion and glucagon in a parallel fashion. The present studies show that glucagon and fluoride ion :IC- tivate a common adenyl cyclase system in liver l)lasma men,- branes through processes that have markedly different ch:tr:lc- teristics. It is likely that the two activation processr-; reprexlnt different molecular components in this complex tnzymr syst.em, as has been suggested previously from studies with the adcny1 cyclase system in fat cells (2). L'nderstanding of the III&C&II basis for the selective effects of divalent cations, l)yro])hosl)hate, phospholipase A, and detergents must await isolation and ch:ir- acterization of the components of the adenyl c~.clasc system. Selective inactivation by detrrgent,s of the glucngoll-res]Jollse in liver membranes or of the respon.se of the fat. cell ghost system to several hormones suggests t.hat this is a general characteristic of the processes through which hormones activate aden. C~&ISP. Other esamples of selec.tive inactivation of hormone response by detergents or agents thought to act as surfactnnts, such as IJhcno- thiazines, have been reported for a variety of adrnyl c~yc*lase systems (14). The catalytic component of the adenyl cyclase s!-stem, ;LP reflected by fluoride activation, is either unaffected or cnh:mcetl by agents that alter lipids, whereas the Iiorn~olle-acti\;ltcrl processes are inactivated by these agents. It would :I~~w;II' that the structures of the catalytic comlxment and the coml)oncnt or comport&s involved in hormone activation :II'V nlodifie(I differently by removal or modifiration of lipids. Thr> scll