The Role of Acidic Phospholipids in Glucagon Action on Rat Liver Adenylate Cyclase

We have examined the effects of phospholipase C from Bacillus cereus (Bc) and from Clostridium perfringens (Cp) on various parameters involved in the activity and response of adenylate cyclase to glucagon in rat liver plasma membranes. A crude preparation of Cp-phospholipase C hydrolyzes “neutral” phospholipids (phosphatidylcholine, phosphatidylethanolamine, sphingomyelin) in these membranes. In contrast, highly purified Bc-phospholipase C hydrolyzes acidic phospholipids (phosphatidylserine, phosphatidylinositol) but not sphingomyelin. Treatment of the membranes with either type of phospholipase does not alter basal adenylate cyclase activity or the stimulatory effects of fluoride ion on the enzyme system. Bc-phospholipase C abolishes the effects of glucagon on adenylate cyclase whereas Cp-phospholipase C causes only partial loss of glucagon response even after hydrolysis of 60% of the membrane phospholipids. These findings provide evidence that acidic phospholipids are more specifically involved in glucagon activation of adenylate cyclase. Acidic phospholipids are not directly involved in the binding of glucagon to its receptor. Treatment with Bc-phospholipase C results in a lo-fold reduction in the affinity but not the quantity of specific binding sites for glucagon. Binding of Des-His-glucagon, a competitive, inactive analogue of glucagon is unaffected by Bc-phospholipase C treatment and displays the same apparent affinity as does glucagon for the binding sites in phospholipase-treated membranes. These findings suggest that acidic phospholipids are involved in the liganding of the histidine residue of glucagon to a regulatory site responsible for glucagon action. GTP, which is required for glucagon action on adenylate cyclase and which increases the rate of dissociation of glucagon from its receptor, does not exert these effects in Bc-phospholipase C-treated membranes. GTP also does not alter the rate of dissociation of Des-His-glucagon from the binding sites in either untreated or treated membranes, indicating that the histidine residue of glucagon is required for the effects of GTP to be expressed. It appears, therefore, that GTP and the histidine residue of glucagon bind to a common site involved both in the activation of adenylate cyclase and in the dissociation of glucagon

(Bc) and from Clostridium perfringens (Cp) on various parameters involved in the activity and response of adenylate cyclase to glucagon in rat liver plasma membranes.
Treatment of the membranes with either type of phospholipase does not alter basal adenylate cyclase activity or the stimulatory effects of fluoride ion on the enzyme system.
Bc-phospholipase C abolishes the effects of glucagon on adenylate cyclase whereas Cp-phospholipase C causes only partial loss of glucagon response even after hydrolysis of 60% of the membrane phospholipids. These findings provide evidence that acidic phospholipids are more specifically involved in glucagon activation of adenylate cyclase.
Acidic phospholipids are not directly involved in the binding of glucagon to its receptor. Treatment with Bc-phospholipase C results in a lo-fold reduction in the affinity but not the quantity of specific binding sites for glucagon.
Binding of Des-His-glucagon, a competitive, inactive analogue of glucagon is unaffected by Bc-phospholipase C treatment and displays the same apparent affinity as does glucagon for the binding sites in phospholipase-treated membranes. These findings suggest that acidic phospholipids are involved in the liganding of the histidine residue of glucagon to a regulatory site responsible for glucagon action. GTP, which is required for glucagon action on adenylate cyclase and which increases the rate of dissociation of glucagon from its receptor, does not exert these effects in Bc-phospholipase C-treated membranes.
GTP also does not alter the rate of dissociation of Des-His-glucagon from the binding sites in either untreated or treated membranes, indicating that the histidine residue of glucagon is required for the effects of GTP to be expressed. It appears, therefore, that GTP and the histidine residue of glucagon bind to a common site involved both in the activation of adenylate cyclase and in the dissociation of glucagon * Recipient of United States Public Health Service International Fellowship (F05 TWO 1743). from its receptor.
Acidic phospholipids are required for the concerted effects of glucagon and GTP at this site.
Several studies have suggested that lipids participate in the actions of hormones on adenylate cyclase, a membrane-bound enzyme system (l-9).
A notable example of the role of specific phospholipids has been reported recently by Levey (10,11) who found that addition of phosphatidylserine and phosphatidglinositol to "solubilized" nonresponsive preparations of cat heart adenylate cyclase restored the response of the enzyme to glucagon and epinephrine, respectively.
In other studies (7), treatment of hepatic plasma membranes containing a glucagon-sensitive adenylate cyclase with either purified phospholipase A, or with neutral detergents resulted in concomitant losses of both glucagon action and of binding of the hormone to specific sites having the characteristics of the receptor for glucagon (12). Partial restoration of binding and action of glucagon was observed upon addition of membrane lipids or purified phospholipids, the most effective being phosphatidylserine (7). Extraction of liver membranes with organic solvents also leads to reduction in glucagon action; phosphatidylinositol, but not phosphatidylserine, was reported to partially restore hormone action (9). Although t,hese studies indicate that phospholipids, probably of the acidic type, are required for hormone action, the precise role of these lipids remains unknown.
Adenylate cyclase systems are complex regulatory enzymes (13). The glucagon-sensitive system in rat liver plasma membranes consists of a specific receptor for glucagon which binds the hormone specifically and reversibly (12) and a component that catalyzes the conversion of ATP to cyclic AMP. ' The system also contains a regulatory site which binds specific nucleotides (GTP, GDP, ATP, and ADP) and which is involved in glucagon action (14). These nucleotides increase the rate of dissociation of glucagon from the receptor (15) suggesting that the nucleotide regulatory site is somehow coupled to the sites of glucagon binding on the receptor.
Thus, the loss of glucagon binding and ac-tion that occurs upon treatment of the membranes with surfactants or phospholipase A may result from alterations in the receptor, the catalytic component, the nucleotide regulatory site, or the "coupling" process, referred to as the "transducer" (13), between the receptor and catalytic component.
We report here the effects of phospholipase C (EC 3.1.4.3) from Bacillus cereus (Bc) and from Clostridium perfringens (Cp) on various parameters involved in the activit,y and hormonal regulation of the hepatic adenylate cyclase system. The crude preparation of C. perfringens phospholipase C used in this study contains a sphingomyelinase (16) and a phospholipase C that hydrolyzes phosphatidylcholine and phosphatidylethanolamine, but not acidic phospholipids such as phosphatidylserine and phosphatidylinositol (17)(18)(19).
In contrast, B. cereus phospholipase C hydrolyzes acidic phospholipids but not sphingomyelin (20-23). The relative specificities of these preparations of phospholipase C provides a means of distinguishing between the various classes of phospholipids that may be involved in glucagon action. An additional advantage of their use is that the products of phospholipase C action (diglyceride and phospho-residue) are not potent surfactants as is the case of the products of phospholipase A action (lysophospholipids and fatty acids).

Materials
Phospholipase C from C. perfringens, purchased from Worthington, was used without further purification.
Phospholipase C from B. cereus was kindly supplied in partial and highly purified forms by Dr. J. Shiloach of the Department of Biological Chemistry, Hebrew University (Jerusalem). Dr. P. R. Vagelos (Department of Biological Chemistry, Washington University, St. Louis) also generously provided a highly purified preparation of Be-phospholipase C that is free of proteolytic enzyme (24) (for purification procedures see Ref. 22). Oxoid filters were obtained from Amersham-Searle (Chicago). Highly purified phospholipids used as standards were purchased from Applied Sciences, Inc. Bovine serum albumin was obtained from Pentex. Glucagon was supplied by Lilly.
Des-His-glucagon (DH-glucagon) was a gift from Dr. Finn Sundby (Novo Research Institute, Copenhagen, Denmark), and was contaminated with 5 to 10% glucagon as estimated from adenylate cyclase assays. The sources of all other materials have been specified previously (12).

Methods
Partially purified plasma membranes from rat liver were prepared as described previously (25).
Determination of Phospholipase C Activity-The activities of the various preparations of phospholipase C were determined by a previously described titrimetric procedure (26) using egg yolk lipoprotein phospholipids (27) as substrate. Incubations were carried out at room temperature in 1 ml of medium containing 154 mM NaCl and 4 mrvr CaC&. The pH was maintained at 7.35. One unit of activity is defined as the amount of enzyme that causes the liberation of 1 pmole of H+ per min.
Treatment of Membranes with Phospholipase C-Plasma membranes containing 1.9 to 2.7 mg of protein were incubated at 15" for 5 min (with gentle agitation) in 3.0 ml of medium consisting of 10 mM Tris-HCI, pH 7.5, 1 mM KHCOs, and 1 mM CaC& in the absence or presence of Bc-phospholipase C or Cp-phospholipase C. The concentrations of added phospholipases were adjusted according to the units of activity determined as described above in order to give essentially identical rates of hydrolysis of membrane phospholipids.
The amount of phospholipase C used in this study was 0.01 to 0.7 unit.
The results are supplied in the figures and table as the percentage of membrane phospholipid hydrolyzed in 5 min of incubation.
After incubation, the samples were diluted to 10 ml with icecold buffer (10 mM Tris-HCl, 1 mM KHCOI, pH 7.5) and centrifuged for 10 min at 4" at 10,000 X g in a Sorvall centrifuge.
The supernatant fluid was discarded and the pellet was suspended in 10 ml of cold buffer and centrifuged as above. This washing procedure was repeated once more and the pellet was suspended in 0.5 ml of the buffer. Various assays, described below, were carried out on aliquots of these preparations.
Solutions of creatine kinase and phosphocreatine were prepared freshly for each experiment since it was found that diluted preparations stored even in the frozen state lost considerable ATP-regenerating capacity and did not maintain constant levels of ATP during incubation.
Response of adenylate cyclase to glucagon was measured, unless stated otherwise, in the presence of lo+ M glucagon; fluoride response was to 20 mM NaF.
Incubations were carried out for 10 min at 30". The method of Krishna et al. (28) was employed for measuring the amount of cyclic AMP formed in the reaction.
Binding of i251-Glucagon fir 1251-DH-Glucag~n-1251-Glucagon (specific activity 200,000 to 660,000 cpm per pmole) and *Y-DH-glucagon (specific activity 259,000 cpm per pmole) were prepared as described previously (8). Binding of the labeled peptides to the membrane preparations was carried out as follows. Liver membranes (40 to 80 pg of protein) were incubated in 100 ~1 of medium containing 20 mM Tris-HCl, pH 7.6, lye bovine serum albumin, and the labeled peptides for 15 min (unless specified otherwise) at 25". The samples were diluted rapidly with 1 ml of ice-cold solution of 1 y0 albumin in 20 mM Tris-HCl, pH 7.6, and immediately filtered on oxoid filters. The oxoid filters were soaked in 10% albumin for 30 min prior to use and then washed with 1 ml of 1% albumin in 10 mM Tris-HCI, pH 7.6; this procedure reduced adsorption of the labeled peptides to the filter.
As a control for nonspecific adsorption of the peptides to the liver plasma membranes, incubation of membranes in the presence of 4 X 10e6 M unlabeled glucagon was carried out concurrently with membranes incubated with labeled peptide alone. The difference in bound radioactivity was used to determine the amount of glucagon bound to its specific binding sites (12) in liver membranes.
Radioactivity was determined in a well-type scintillation counter. Phospholipid Extraction and Analyses-Lipids were extracted from washed membranes by the method of Ways and Hanahan (29). Phospholipid phosphorus was determined by the method of Bartlett (30). The types of phospholipids present in the lipid extracts were evaluated by thin layer chromatography (Adsorbosil-5 standard Prekotes, Applied Sciences Laboratories) using a developing solvent system containing (v/v) chloroform-acetonemethanol-acetic acid-water (100:40:20:20: 10). The lipid-containing spots were first located under ultraviolet light after spraying the plates with 0.1% 1-anilinonaphthalene-8-sulfonate (magnesium salt) in water (31). After marking the lipid spots, the chromatograms were sprayed with 0.3y0 ninhydrin in a solution of 3% acetic acid in ethanol (32) ; ninhydrin-positive lipids were visualized after heating the plates for 5 min at 120". Plasma membranes (2.5 mg of protein) were incubated in the absence or presence of 0.i unit of either type of phospholipase. Other incubation conditions and the procedures for extraction, chromatography, and phospholipid detection are described under "Experimental Procedure." Protein-Protein was measured by the procedure of Lowry et al. (33) using bovine serum albumin as st,andard.

RESULTS
In all experiments described in this study, incubations of hepatic plasma membranes were kept to 5 min at 15" in the absence and presence of Bc-phospholipase C or Cp-phospholipase C in order to minimize changes in the stability of adenylate cyclase that occurred with longer times and higher temperatures of incubation.
The concentrations of the enzymes were adjusted, based on their capacity to hydrolyze egg yolk phospholipids (see "Methods"), to give approximately equal rates of hydrolysis of membrane-bound phospholipids. Fig. 1 shows the types of membrane phospholipids hydrolyzed by Bc-phospholipase C and Cp-phospholipase C under standard incubation conditions with sufficient concentrations of the enzymes to cause, respectively, 57 and 58% hydrolysis. Four major phospholipid spots, representing phosphatidylcholine, phosphatidylethanolamine, sphingomyelin, and phosphatidylserine, were detected in chromatograms of lipids extracted from untreated membranes.
Phosphatidylinositol was not separated from phosphatidylserine with the solvent system used. Bc-phospholipase C hydrolyzed phosphatidylserine (phosphatidylinositol), phosphatidylethanolamine, and phosphatidylcholine, but not sphingomyelin.
In contrast, Cp-phospholipase C did not hydrolyzephosphatidylserine (phosphatidylinositol). This qualitative pattern of hydrolysis by the two phospholipases was confirmed by Dr. William Weglicki2 who analyzed the phosphorus  (0) and Cp-phospholipase C(0) on response of adenylate cyclase to glucagon (10-G M) and fluoride ion (20 mrvr). Incubation conditions and procedures for assaying adenylate cyclase activity are described under "Experimental Procedure." content of the individual phospholipids chromatographed with a solvent system that separated completely all of the phospholipids present in hepatic membranes (34-37).
The specificities of the two types of phospholipases confirm the observations of others (16-23, 38) using either purified phospholipids or membranebound phospholipids as substrates. Fig. 2 illustrates the effects of Bc-phospholipase C and Cpphospholipase C on the response of adenylate cyclase to glucagon and fluoride ion; the effects are expressed relative to the percentage of phospholipids hydrolyzed with increasing concentrations of the phospholipases.
Fluoride ion stimulates adenylate cyclase by a different process from hormonal activation and is thought to act directly on the catalytic component (14). The effects of fluoride ion are useful, therefore, for distinguishing the characteristics of the catalytic component from those of the components through which hormonal stimulation is expressed. It can be seen that the stimulatory effects of fluoride ion were not affected after hydrolysis of 55y0 of the phospholipids by either of the phospholipases.
Basal activity (not shown) was also not altered by the phospholipases.
In contrast, both phospholipases, at concentrations sufficient to hydrolyze more than 20% of the membrane phospholipids, caused increasing losses of glucagon response. Bc-phospholipase C abolished the hormone response completely when 45 to 50% of the membrane phospholipids were hydrolyzed.
However, in twelve experiments of this type, Cpphospholipase C did not cause more than 50% loss of hormone phosphatidylserine, phosphatidylinositol, and sphingomyelin hydrolyzed by the two phospholipases under considerations described in Fig. 1 3. Effects of Bc-phospholipase C and Cp-phospholipase C on the response of adenylate cyclase to glucagon (10-C M) and on the binding of *2SI-glucagon (10-g M) to liver plasma membranes. Incubation conditions and procedures for assaying adenylate cyclase activity and binding of labeled glucagon are described under "Experimental Procedure." response even when 60% of the phospholipids were hydrolyxedP Since the major differences in phospholipid specificity between Bc-phospholipase C and Cp-phospholipase C are their ability, respectively, to hydrolyze acidic phospholipids and sphingomyelin, we attribute the quantitative difference in their effects on hormone response to the hydrolysis of the acidic phospholipids by Bc-phospholipase C. The basis of the relatively small effects of Cp-phospholipase C onhormoneresponse isunknownand requires further experimentation with purified preparations of the phospholipases in C. perfringens.
We have observed that omission of calcium ion from the incubation medium reduces correspondingly both hydrolysis of membrane phospholipids and the effects of Cp-phospholipase C on hormone response, in keeping with the stimulatory effects of calcium ion on Cp-phospholipase C activity (38). The effects of Bc-phospholipase C, a zinc-requiring enzyme (21), were unchanged by the omission of calcium ion; the three preparations (partially purified and highly purified) of Bc-phospholipase C used in this study gave identical effects on hormone response when equivalent units of activity were added.
We have shown previously (5,7) that phospholipase A2 causes the loss of both glucagon action and binding of the hormone to specific sites having the characteristic of the receptor for glucagon. As shown in Fig. 3, Bc-phospholipase C and Cp-phospholipase C also caused decreases in the binding of 1251-glucagon to these sites. However, although the loss of binding bore some relationship to the effectiveness of the two phospholipases on glucagon action, glucagon bindingwas not diminished to the same extent as was glucagon action.
In these experiments the concentration of glucagon used for binding was 10Mg M whereas that for action was 10-G M. The same differences were obtained when equimolar ( 1OB8 M) concentrations of the hormone were employed 3 In subsequent experiments we found that inclusion of 1 mM p-mercaptoethanol in the media used for incubating the membranes with phospholipase resulted in a significant decrease in loss of glucagon response due to Cp-phospholipase C treatment. In the presence of the sulfhydryl agent but under otherwise identical conditions (see "Experimental Procedure"), Cp-phospholipase C-treated membranes displayed (in three experiments) 52 f 2% phospholipid hydrolyzed and 20 f 2% loss of glucagon response rather than the 40 to 500/o loss noted in Figs. 2 and 3. Inclusion of the sulfhydryl reagent during treatment with Bc-phospholipase C did not alter the extent of phospholipid hydrolysis or the loss of glucagon response. The results are expressed as the percentage of labeled hormone bound to relative control, untreated membranes.
for both binding and action.
These findings suggested that binding per se of glucagon was not uniquely involved in loss of glucagon action and raised the possibility that the affinity of glucagon for its specific binding sites was altered by phospholipase C treatment.
As illustrated in Fig. 4, Bc-phospholipase C, at concentrations sufficient to catalyze 60% hydrolysis of membrane phospholipids, caused only a slight (lOyO) decrease in glucagon binding when the concentration of 1251-glucagon was increased to 10e7 M, indicating that the sites available for glucagon binding are not diminished significantly by hydrolysis of acidic phospholipids. Shown in Fig. 4 and consistently observed, is increased binding of glucagon (at concentrations higher than lOwE M) after treatment with low concentrations of Be-phospholipase C. The basis of the increased binding is not understood but may reflect either unmasking of cryptic binding sites by removal of phospholipids or by the introduction of "pores" in "inside-out" membrane vesicles (39) through which glucagon may penetrate the membrane and bind to glucagon binding sites on the interior face of the membrane. The latter possibility is raised since the receptor for glucagon is thought to be located exclusively on the outer surface of the plasma membrane (40).
Since Bc-phospholipase C did not diminish the specific binding sites for glucagon, it was apparent that a major change introduced by phospholipid hydrolysis is a reduction in the affinity of the binding sites for glucagon. This is illustrated in Fig. 5 by Lineweaver-Burk plots of the binding of 1251-glucagon to untreated and Bc-phospholipase C-treated membranes. 4 The apparent 4 It should be emphasized that the dissociation constants for glucagon or DH-glucagon binding are not provided by the Lineweaver-Burk plots. The free concentrations of glucagon or DH-glucagon change rapidly during incubation owing largely to conversion of the peptides to products which do not bind to liver membranes or, in the case of glucagon, activate adenylate cyclase (12). For this reason, Lineweaver-Burk (or Scatchard) plots of 0 affinity of the binding sites for glucagon was reduced from 5 X 10-g M in untreated membranes to about 5 x lo-* M in Bc-phospholipase C-treated membranes.
In these experiments, we also compared the binding of 1251-DH-glucagon, a biologically inactive, competitive inhibitor of glucagon binding and action (41), to that of native glucagon in treated and untreated membranes. The apparent affinity of DH-glucagon for the glucagon binding sites was unaltered by Bc-phospholipase C treatment and was similar (about 6 X lo-* M) to that observed for native glucagon binding to the treated membranes.
We have shown in earlier studies (42) that binding of glucagon to its receptor is not sufficient per se to activate adenylate cyclase in hepatic membranes.
GTP and other purine nucleotides (GDP, ATP, ADP) bind to regulatory sites reauired for the actions of glucagon on adenylate cyclase (42) and increase the rate of dissociation of bound glucagon from its receptor (15). It was of interest, therefore, to determine whether the hydrolysis of acidic phospholipids by Bc-phospholipase C would alter the effects of GTP on the rate of dissociation of bound lz51-glucagon. As shown in Table I, GTY increased the rate of dissociation of bound 1251-glucagon in untreated plasma membranes in the absence or presence of excess (low5 M) unlabeled hormone.
In contrast, GTP did not alter the rates of dissociation of labeled glucagon in Uc-phospholipase C-treated membranes. Since the affinity of the binding sites for glucagon in treated mrmbranes is reduced to that of DH-glucagon, it seemed possible that the loss of the nucleotide effect on binding of glucagon to treated membranes is related to the reduction in affinity and that the histidine residue of the hormone is required for the GTP effect to be expressed. Accordingly, we examined the effects of the binding data yield, in the experiments described in Fig. 5, only relative changes in apparent affinities of the binding sites for glucagon and DH-glucagon.
The apparent increase in binding sites for DH-glucagon relative to glucagon in treated and untreated membranes (note the intercepts on the ordinate) may be explained in part by the lower rates of inactivation of DH-glucagon relative to glucagon (8). Lower rates of inactivation might yield, in a kinetic sense, an apparent increase in the binding sites for DH-glucagon since the concentration of DH-glucagon would not change as rapidly as that of glucagon. GTP on the rates of dissociation of *2"I-DH-glucagon from cithel untreated or Bc-phospholipase C-treated membranes. As shown in Table I, GTP did not alter the rates of dissociation of bound 1251-DH-glucagon in either untreated or treated membranes. Moreover, the amount of 12SI-DH-glucagon bound was essentially the same as that of 1251-glucagon bound to treated membranes. The amount of labeled glucagon or DI-I-glucagon bound was reduced by approximately 50% after addition of excess unlabeled glucagon, in contrast to a 30% reduction (in the absence of GTP) of bound labeled glucagon to untreated membranes. This difference is expected since the affinity of the binding sites in treated membranes is lo-fold less than that of the affinity of untreated membranes for glucagon.

DISCUSSIOiY
The present studies confirm and extend previous reports that acidic phospholipids play a selective role in glucagon action on adenylate cyclase and show the value of using phospholipase C preparations having different affinities for the classes of phospholipids found in biological membranes. As stated in the introduction, another advantage of using phospholipase C for probing the functions of phospholipids in membrane processes is that the products of its action are not potent surfactants as is the case of phospholipase A action on membrane phospholipids. Nonspecific surfactant effects may have been the basis of our previous findings (7) that low concentrations of phospholipase Az or digitonin increased both basal and fluoride-stimulated activities, whereas higher concentrations of these agents produced a loss of these activities in either liver or fat cell membranes Phospholipase C treatment of the liver membranes did not alter either basal or fluoride-stimulated activities; only the hormone response was affected. It would appear, therefore, that phospholipids have little to do with the structure and activity of the catalytic component or with the process by which fluoride ion stimulates enzyme activity.
The precise nature of the acidic phospholipid (or phospholipids) involved in glucagon action on the hepatic system has not been established.
The studies of Levey (10) suggest that phosphatidylserine, which is required for glucagon action on a "solubilized" preparation of myocardial adenylate cyclase, may be specifically involved.
Studies of the restoration of glucagon action by addition of phospholipids to membranes treated with Bc-phospholipase C may help to resolve this question.
Our principle objectives are to establish at what stage of hormonal regulation phospholipids may participate and to assign a specific role for lipids in this process. The following possibilities are considered.
(a) The phospholipids form a complex with the receptor and participate in the binding of glucagon, possibly with the intensely hydrophobic COOH-terminal region of the hormone (41); (b) phospholipids transmit or modulate the physicochemical signals that may be produced by hormone-receptor interaction; (c) phospholipids play a strictly structural role possibly by maintaining the receptor-enzyme complex in its hormonally responsive state.
The possibility that phospholipids participate directly in the binding of glucagon to its receptor can be ruled out. Hydrolysis of acidic phospholipids by Bc-phospholipase C did not diminish the number of specific binding sites for glucagon.
The major change observed was a lo-fold decrease in the affinity of the binding sites for glucagon.
Most importantly, the affinity of glucagon for the binding sites in the treated membranes became the same as that for DH-glucagon which binds with affinity identical with the receptors in either untreated or phospholipase C-treated membranes.
The NHS-terminal histidine residue of glucagon plays a unique role in both the binding and action of glucagon on adenylate cyclase systems (41). Histidine is required for biological action and contributes substantially to the forces involved in the binding of glucagon to its receptor.
The lo-fold difference in the binding affinity of DH-glucagon compared to native glucagon observed here and in a previous report from this laboratory (8) calculates to be (at 30") about 1.4 Cal per mole. This difference in binding energy is consistent with the histidine residue binding through hydrogen bonds to a special site involved in the regulation of adenylate cylase activity by glucagon.
Since DH-glucagon specifically competes with glucagon at the receptor (41), stereospecificity of glucagon binding resides in the region encompassing the 2 to 29 amino acid residues of the hormone.
Removal of the acidic phospholipids did not change the stereospecificity or the affinity of the 2 to 29 region for the receptor binding sites (cf. Fig. 5). Thus, phospholipids play an important role in the binding of the histidine residue to a regulatory site involved in hormone action.
Another function lost upon removal of the acidic phospholipids is the ability of nucleotides such as GTP to enhance the rate of dissociation of glucagon from its receptor.
We have shown previously (14) that purine nucleotides (GTP, GDP, ATP, and ADP) are required for glucagon action on adenylate cyclase and, over the same range of concentrations, stimulate the rate of dissociation of bound hormone.
Similar requirements for purine nucleotides in the actions of several hormone-sensitive adenylate cyclase systems have been reported subsequently (43-46) and have been extended recently in this laboratory5 to the adrenocorticotropic hormone-sensitive adrenal adenylate cyclase system and the multireceptor system in rat fat cells (47). It has been 5 C. Londos and M. Rodbell, unpublished observations. suggested that the nucleotidrs bind to a regulatory site involved in the binding and action of glucagon as one means of accounting for the dual effects of the nucleotides (14). The present study showed that GTP does not alter the rate of dissociation of bound DH-glucagon from either phospholipase C-treated or untreated membranes and affects the dissociation rate of glucagon only when the hormone is bound to untreated membranes (Table I). Thus, GTP only affects the rate of glucagon dissociation from the receptor when the histidine residue can bind to the regulatory site involved in glucagon action. This finding can be taken as strong evidence that GTP (or other nucleotide modifiers) and the histidine residue of glucagon bind to a common regulatory site which controls the state of activity of adenylate cyclase as well as the rate of dissociation of the hormone from its receptor.
The results suggest that acidic phospholipids are involved in the action of glucagon (i.e. the histidine residue) and the nucleotides at the regulatory site.
The nature of the regulatory site remains unknown. It is not possible, therefore, to assign a strictly structural or functional role to phospholipids in hormone action.
The negatively charged polar group of the acidic phospholipids might participate functionally in hormone regulation if, for example, a metal ion, with which acidic phospholipids form strong ligands (48, 49), were a component of the regulatory site. With regard to a metal ion as a component, it should be noted that histidine forms complexes with metal ions, particularly those of the transition series (50). As for a structural role, phospholipids may serve to maintain the quaternary structure of the receptor-enzyme complex such that the regulatory site is oriented in a configuration necessary for interaction with glucagon and nucleotides.
In conclusion, acidic phospholipids are required for the interrelated effects of glucagon and nucleotides on the hepatic adenylate cyclaee system. A relationship is clearly indicated between these lipids and the liganding of the histidine residue of glucagon and of nucleotides to a site that controls the activity of this complex regulatory enzyme system.