Thomas H. Hudson, James F. Roeber, and Gary L. Johnson. From the Section of Physiological Chemistry, Division of Biology and Medicine, Brown University, ...
THEJOURNALOF BIOLOGXCAL CHEMISTRY Vol. 256.No.3. Issue of February 10, pp. 1459-1465, 1981 Prrnied m U S.A.
Conformational Changesof Adenylate Cyclase Regulatory Proteins Mediated by Guanine Nucleotides* (Received for publication, August 6, 1980)
Thomas H. Hudson, James F. Roeber, and Gary L. Johnson From the Section of Physiological Chemistry, Division of Biology a n d Medicine, Brown University, Providence, Rhode Island 02912
and N ( 3 , plays a pivotal role in the activation of adenylate cyclase byhormones. The effect ofguanine nucleotides on the system appearsto be mediated by the G-protein,’which couples receptor-hormone interactions with activation of adenylate cyclase (1,4). Catecholamine-sensitive adenylate cyclase of avian erythrocytes is characterized by a relatively tight binding of GDP to the guanine nucleotide regulatory site (5). The hydrolysisresistant GTP analogs, Gpp(NH)p and GTPyS, are significantly less effective in activating adenylate cyclasewhen hormone is absent. If the membranes are first incubated with isoproterenol and GMP (6, 7) and then washed thoroughly, Gpp(NH)p or GTPyS is capable of markedly activating cyclase in the absence of hormone. Using labeled guanine nucleotides, it was demonstrated that release of bound GDP from the regulatory site occurred during hormone treatment, so that after washing of the membranes, guanine triphosphate can occupy the site. From these studies and others (8-11), Cassel and Selinger (5, 10, 11) have proposed that hormone stimulates adenylate cyclase by causing the release of GDP bound to the G protein, allowing guanine triphosphate to bind, resulting in activation of the enzyme. Activation is terminated by hydrolysis of the bound GTP by a hormonestimulated GTPase. Cholera toxin modifies guanine nucleotide regulation of adenylate cyclase apparently by inhibiting the hormone-stimdated GTPase activity associated with the enzyme (5). The inhibition of GTPase activity by cholera toxin treatment makes GTP behave similar to Gpp(NH)por GTPyS (3,5,12). The action of cholera toxin is due to its catalyzing ADPribosylation of a M , = 42,000 membrane protein (13-15). This M, = 42,000 peptide is a subunit of an oligomeric protein of M, = 126,000 (2, 16). The M, = 126,000 protein from pigeon erythrocyte membranes binds to a GTP affiity column (13, 17),and confersguanine nucleotide and fluoride ion regulation on adenylate cyclase activity (2, 17). A similar M , = 126,000 protein from human erythrocyte membranes reconstitutes hormone-sensitive adenylate cyclase activity when mixed with membranes from cyc- S49 cells (16), which are deficient in guanine nucleotide regulatory components (4) and the cholera toxin substrate (3, 15). The ability to radiolabel specifically the M , = 42,000 subunit of the G-protein using cholera toxin and [3ZP]NADpreviously The guanine nucleotide regulatory component of the d e - allowed us to study its structure and orientation in the memnylate cyclase system, variously referred to as G/F (I),G (2), brane (18). This communication extends this approach to
Cholera toxin, in thepresence of r3’P]NAD+, catalyzes the specific radiolabeling of a M, = 42,000 subunit of a protein (G protein) which mediates guanine nucleotide regulation of hormone-sensitive adenylate cyclase in pigeon erythrocyte membranes. Treatment of labeled membranes with GMP and isoproterenol followed by extensive washing presumably releases tightly bound GDP from the G-protein regulatory site. In the absence of added guanine nucleotide, treatment of these membranes with trypsin results in the partial digestion of the M, = 42,000 radiolabeled peptide and production of several minor trypsin-specific fragments which remain associated with the membrane. Incubation of memguanosine-5’-0-(3-thiotriphosphate) branes with (GTPyS) alters the partial tryptic digestion of the M, = 42,000 radiolabeled peptide, resulting in a loss of the minor fragments and the quantitative generation of a M, = 41,000 peptide. GTPyS exposes a new site to tryptic cleavage which releases from the membrane an inactive fragment of the G-protein containing the Mr = 41,000 labeled peptide. Guanosine-5’-0-(2-thiodiphosphate) (GDPBS) fails to activate adenylate cyclase and does not cause the generation of the M, = 41,000 tryptic fragment. Incubation of membranes with G D P P does, however, decrease tryptic digestion of the M, = 42,000 labeled peptide. The effects of GTPyS and GDPBS on the partial tryptic digestion of the G-protein are competitive suggesting their action is mediated by a common binding site. Generation of the GTPyS-specificM, = 41,000 tryptic fragment from the M, = 42,000 toxin labeled peptide exhibits a time-dependent increase upon incubation of membranes with G T P y S which is similar to that for cyclase activation. Addition of isoproterenol to the GTPyS incubation diminishes the lag time for cyclase activation and the generation of the M, = 41,000 fragment. Human erythrocyte membranes possess the G-protein but functionally lack both the P-adrenergic receptor and catalytic adenylate cyclase. Incubation of cholera toxin-labeled human erythrocyte membranes with GTPyS followed by exposure to trypsin also generates a specific M, = 41,000 fragment.
* This work wassupported by National Institutes of Health Grants GM 26776 and AM 007817, American Cancer Society Grant IN 45-T, and BRSG funds from Brown University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
’
The abbreviations used are: G-protein, guanine nucleotide regulatory protein; Gpp(NH)p, guanyl-5”yl imidodiphosphate; GTPyS, guanosine-5”0-(3-thiotriphosphate); GDPBS, guanosine-5”0-(2thiodiphosphate); EGTA, ethylene glycol bis(P-aminoethyl ether)N,N,N’,N’-tetraaceticacid; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate.
1459
Components Coupling
1460
examine differences in proteolytic susceptibility of the M, = 42,000 peptide after incubation of membranes with guanine nucleotide. Our results indicate that guanine triphosphates cause a specific conformational change of the G-protein, and that hormone-receptor interactionsaffect the rateof guanine triphosphate-induced change in conformation.
of Cyclase Adenylate
m m I m M m 110Ibr
EXPERIMENTALPROCEDURES
Methods Isolation of Plasma Membranes-Freshly collected pigeon erythrocytes were lysed in the presence of DNase and membranes were prepared as described previously for turkey erythrocytes (19). Human erythrocytes were prepared similarly without the DNase treatment. CholeraToxinTreatment of Membranes-Membranes were washed by centrifugation in 500 mM potassium phosphate (pH 7.0), and resuspended (5mg of protein/ml) in 250 mM potassium phosphate (pH 7.0), 20 mM thymidine, 5 mM ADP-ribose, 20 mM arginine-HCI, 100 p~ GTP, 400 p~ ATP, 10 pg/ml of cholera toxin (activated with 20 mM dithiothreitol (20)), and 25 p~ NAD'. The reaction mixture was incubated for 30 min a t 30°C and terminated by dilution and centrifugation in10 ml of ice-cold 20 mM Hepes (pH8.0), 2 mM MgClz, 1 mM EDTA, and 1 mM 2-mercaptoethanol (membrane buffer). For radiolabeling, the specific cholera toxin substrate ['12P]NAD' was added to the above incubationa t a specific activity of 5 Ci/mmol. Protease Digestions-Digestionswereperformed in membrane buffer a t 22OC. The membrane concentrationwas 4mg of protein/mL The reaction was inhibited by the addition of aprotinin followed by the additionof SDS and2-mercaptoct.hanol to final concentrations of 1 and 5%, respectively. Peptide mapping using two-dimensional gels was performed by a modification of the method of Cleveland et al. (21) as described previously (18). SDS-Polyacrylamide Gel Electrophoresis and Autoradiography-Membranes in 1% SDS and 5% 2-mercaptoethanol were boiled for 5 min and 708 glycerol was added to a final concentration of12%. SDS-polyacrylamide gel electrophoresis was performed by the method of Laemmli (22) as described by O'Farrell (23) using a 12%% acrylamide separating gel. Gels were stained and fixed in 2% Coomassie blue, 50% acetic acid, 10% ethanol, and destained in 10% acetic acid, 10% ethanol. Dried gels were used to expose Kodak XOmat H film and DuPont lightning-plus intensifying screens a t -7OOC. Adenylate Cyclase Assay-Adenylatecyclasewas assayed in a mixture containing 0.4 mM ['"PIATP (10 to 25 cpm/pmol), 50 mM Na Hepes (pH8.0). 6 mM MgClz, 0.2 mM EGTA. 2 mM 2-mercaptoethanol, 0.1 mg/ml of bovine serum albumin, 10 mM creatine phosphate, 10 units/ml of creatinephosphokinase (inafinalvolume of 100 pl). Effectors of adenylate cyclase included in the reaction mixturewhen appropriate were: 10 PM DL-iSOprOterenOl, 100 p~ GTPyS. 100 p~ GDPPS. The assay was initiated by adding 25 pg of membrane protein to the reaction mixture and incubated a t 30°C. The reaction was terminated by theaddition of 1 ml of 1% SDS. ['"'PICAMP was purified by sequential chromatography on Dowex and alumina. Chromatographic Analysis of GDPpS-GDPBS (1 mM final concentration) was incubatedeither (a) for 30 min at 30°C in the adenylate cyclase reactionmixturecontaining pigeon erythrocyte membranes (0.4 mg of membrane protein/ml), or ( b ) for 30 min a t 22°C in membrane buffer containing pigeon erythrocyte membranes (0.5 mg of membrane protein/ml). The membraneswere removed by centrifugation and 15 pl of the supernatant was chromatographed on polyethyleneimine cellulose sheets (polygram cell 300) as described by Eckstein etal. (25). The developing solution contained 0.5 M NaCl and 0.75 M Tris-HCI (pH 7.5). The nucleotides were visualized by UV absorption. One major spot of RF = 0.23 corresponding to GDPpS (25) and a minor spot (less than 10% of total) of RI; = 0.07 were observed in incubation a. Only the major GDPpS spotwas observed in the supernatant of incubation b.
-CHOLERA TOXIN
+CHOLERA TOXIN
*-
-
.
4
.
FIG. 1. Effects of guanine nucleotides on tryptic digestion of the M, = 42.000 cholera toxin substrate. Pigeon erythrocyte membranes were labeled with ["'PINAD' in the presence (Lanes I to 4 ) orabsence(Lanes 5 to 7) of cholera toxin as detailed under "Experimental Procedures." Labeled membranes were resuspended (1 of mg protein/ml) in membrane buffer including 1 mM GMP and Fjg IIM Dr.-isoproterenol. After a 20-min incubation a t 37OC, the IIWIIIbranes were pelleted and washed three timesby dilution with ice-cold membrane buffer and centrifugation. The final pelletwas resuspended (5 mg of protein/ml) in membrane buffer and incubated for 45 min a t 22OC with either no additions, or100 PM GDPpS or GTPyS.Aliquots from these incubations were then incubated in the absence or presence of 25 pg/ml of trypsin for 12 min as detailed under"Experimental Procedures." The samples were then subjected to SDS-acrylamide electrophoresis and autoradiographed. The large arrows mark the M , = 42,000 cholera toxin substrate, and the small arrows designate the M, = 41,000 and 30.000 peptide fragments of the M , = 42.000 peptide. Lanes 1 and 5, no guanine nucleotide, no trypsin; Lane 2,no guanine nucleotide plus trypsin; Lanes 3 and 6, GDPBS plus trypsin; Lanes 4 and 7, GTPyS plus trypsin. Autoradiographs of Lanes 1 to 4 were exposed for 36 h; whereas Lanes 5 to 7 were exposed for 72 h in order to bring out anylow abundance fragmentswhich might interfere with the analysis of fragments generated specifically from the M , = 42,000 cholera toxin substrate. No difference in Coomassie blue staining patterns of high abundance proteins was observed in the presence or absence of GDPpS and GTPyS, or with or without cholera toxin treatment when comparing control (LanesI and 5) or trypsin-treated (Lanes 2 to 4,6and 7) membranes.
of GDPpS or GTPyS after treatment of membranes with isoproterenol and GMP. The thio analogs of GDP and GTP areresistantto cellularnucleotide phosphatases,GTPyS being an activator of adenylate cyclase (24-26) and GDPpS being a competitive inhibitor or partial activator (25, 26). Incubation of membranes with guanine nucleotide does not affect thebehavior of the M , = 42,000 peptide on SDS acrylamide gels (not shown). In the absence of guanine nucleotide (Fig. l, Lane Z),trypsin treatment of the membranes results in the partial digestion of the M , = 42,000 labeled peptide. Thelabeled proteins at the top of the gel in Lanes I to 4 other than the M , = 42,000 peptide are nonspecifically labeled proteins thatgenerally migrate asbroad diffuse bands. These nonspecifically labeled proteins are seenalso incontrol Materials GTPyS and GDPpSwere from Boehringer Mannheim. DL-ISOprO- preparations which are treated in exactly the same manner but without the addition of cholera toxin to the reaction terenol HCI and aprotinin were purhcased from Sigma. Trypsin and mixture. When control membranes are partially digested with chymotrypsin were purchased from Worthington. [."PINAD' and [y-'"PIATPwere purchased from New England Nuclear. Cholera trypsin (Lanes 5 to 7), these nonspecifically labeled bands are toxinwas obtained from Schwarz/Mann. All other materials were partially digested but do not contribute any of the fragments reagent grade. designated by arrows in Fig. 1, and do not impair the analysis RESULTS of digest products of the toxin-specific M , = 42,000 radiolaFig. 1 shows limited tryptic digestion patterns of the M , = beled peptide. 42,000 radiolabeled toxin substrate in the presence or absence Both GDPpS and GTPyS affect the digest pattern produced
Coupling Components of Adenylate Cyclase by trypsin. Incubation of membranes with GDPPS (Lane 3) causes a decrease in tryptic digestion of the M , = 42,000 labeled peptide with a corresponding loss of the other tryptic fragments observed in the absence of guanine nucleotide (Lane 2). Incubation with GTPyS results in the generation of a specific M , = 41,000 tryptic fragment (Lane 4). The effects of GTPyS are also observed with Gpp(NH)p, indicating that the formation of the M, = 41,000 tryptic fragment is not a GTPyS artifact. Similar effects of GDPpS and GTPyS on tryptic digestion of the M , = 42,000 labeled peptide are observed with Lubrol PX-solubilized G-protein only if the membranes are preincubated with isoproterenol and GMP (data not shown). Fig. 2 shows the densitometric tracings of the autoradiographs (Lanes 1 to 4J of Fig. 1. It is clearly shown that in the absence of added guanine nucleotide, trypsin treatment of membranes reduces the label in the M, = 42,000 band by approximately 50%, resulting in the generation of several minor fragments. Addition of GDPpS to the incubation mix-
11.7
NO TRYPSIN
-J I
9.9
TRYPSIN
+ GDPBS
7.6
I
TRYPSIN
+ GTPYS
M,xIO-~
35
41 4 2
FIG. 2. Densitometry of tryptic digestion patterns of the M. = 42,000 cholera toxin substrate. The autoradiographs shown in Densicord recording Fig. 1, Lanes 1 to 4, werescannedusinga electrophoresis densitometer. The relative intensity of each band was quantitated by determining the areaunder each peak. Identical results were obtained with autoradiographs exposed for different times ranging from 18 to 72 h.
1461
ture results in a marked protection of the radiolabeled peptide from tryptic digestion. In contrast to the effect of GDPPS, the presence of GTPyS in the incubation results in the generation of a M , = 41,000 fragment with a corresponding quantitative decrease in the M, = 42,000 radiolabeled peptide, suggesting M , = 41,000 fragment is generated by tryptic digestion that the of the M , = 42,000 peptide. Fig. 3 provides further evidence that theM , = 41,000 peptide is a tryptic digest product of the M , = 42,000 toxin-specific band. Partial proteolytic digestion using Staphylococcus aureus V8 protease indicates that the M, = 41,000 peptide generated by trypsin digestion of GTPyS-treated membranes produces similar labeled peptides as the M, = 42,000 toxinspecific peptide. The presence of one or two apparently unique peptides in the M , = 41,000 peptide digest pattern is not unexpected, since this peptide is a tryptic digestion product of the M, = 42,000 peptide. Significantly, none of the nonspecifically labeled peptides in other regions of the gel generate peptide fragments that correspond to any of the fragments in the maps of the cholera toxin-specific tryptic peptides. These results substantiate the findings that the nonspecifically labeled proteins do not contribute any of the peptide fragments designated by arrows in Fig. 1. Previously, it was demonstrated that GDPpS is either a competitive inhibitor or a weak partial activator of adenylate cyclase(25, 26). Table I confirms these findings with the pigeon erythrocyte system. Membranes pretreated with isoproterenol and GMP followed bythorough washing are markedly stimulated by GTPyS, whereas GDPPS isineffective. In addition, GDPPS is able to inhibit GTPyS activation of adenylate cyclase in a concentration dependent manner. Fig. 4 shows that GDPpS also is able to inhibit GTPyS-mediated formation of the M, = 41,000 tryptic fragment. When membranes are incubated with a 2-fold excess of GDPpS relative to GTPyS, there is a marked decrease in the ability of trypsin to generate the M, = 41,000 peptide, relative to GTPyS alone. GDPPS alone causes little or no generation of the M , = 41,000 tryptic fragment. The M, = 41,000 fragment is therefore a specific marker for guanine triphosphate-specific changes in conformation of the G-protein. In order to observe an effect of hormone on the susceptibility of the M, = 41,000 labeled peptide, we used membranes which were not preincubated with isoproterenol and GMP. Fig. 5 shows the adenylate cyclase activity of pigeon erythrocyte membranes incubated for different periods of time in the presence or absence of GTPyS and isoproterenol, followed by washing of the membranes and subsequent assay of cyclase activity. As demonstrated previously (12), hormone accelerates the rate at which maximalactivation of adenylate cyclase is achieved by hydrolysis-resistant GTP analogs. With sufficient time, the cyclase activity obtained by incubation with GTPyS alone is similar to that achieved in the presence of isoproterenol and GTPyS. Furthermore, the activation is stable to washing. The addition of GTPyS or isoproterenol to the adenylate cyclase assay mixture has littleeffect on enzyme activity once maximal activation is achieved. Fig. 6 shows the tryptic digestion products of the M, = 42,000 radiolabeled peptide during the time course of cyclase activation depicted in Fig. 5. The autoradiographs in Fig. 6A were exposedfor only 18 h. It is readily apparent that there is a time-dependent GTPyS-mediated increase in the generation of the M, = 41,000 tryptic fragment (Fig. 6 A , Lanes 2,6,and 10)and that isoproterenol accelerates the rateof formation of this digest product (Lanes 4, 8,and 12).Densitometry of the autoradiographs shown in Fig. 6A indicates a quantitative relationship between the formation of the M , = 41,000 fragment and a corresponding decrease in the M , = 42,000 peptide.
Coupling Components of Adenylate Cyclase
1462
NO TRYPSIN
TRYPSIN
42
42
FIG.3. Peptide mapping of the Mr = 42,000 cholera toxin substrate and the M. = 41,000 tryptic fragment.
F
'
tGDPj3S
tGTP8S
42
s.
e
TRY PSlN
TRYPSIN
Samples shown in Fig. 1, Lanes 1 to 4, were electrophoresed in 10% SDS-acrylamide gels. The fmt dimension SDS gel was then placed on top of a discontinuous SDS gel systemcontaining 15% acrylamide in the separating gel. A 1% agarose gel was used to keep the fust dimension SDS gel in place and contained 125 p g / d of aureus V8 protease. Electrophoresis was performed in thenormalmanner, except thatthe power was turned off for 30 min when the bromphenol blue tracking dye neared the bottom of the stacking gel. The powerwas thenturnedon a t 20 rnA/gel and the peptide fragments generated were then separated in the 158 acrylamide separating gel. Autoradiographs represent a &day exposure.
4 2 41
I
I GDPpS competition
TABLE I GTPyS activation
adenylate cyclase Pigeon erythrocyte membranes were incubated with 10 PM isoproterenol and 100 PM GMP for 20 min at 30°C. Membranes were then washed three times by dilution with 10 ml of membrane buffer and centrifugation for 10 min at 30,000 X g. Membranes were resuspended in membrane buffer and adenylate cyclase activity wa.. measured during a 30-min assay in the presence of various concentrations of GTPyS or GDPPS.Eachvaluerepresentsthemean of triplicate determinations and is representativeof three separate experiments. GTP$ (PMJ
50 0 0 0 0 0
50 50 50 50 50 0
of
GDPflS
fWJ 0
10 25 50 100 250 10 50 100 250 0
of
1
2
3
7 " ic
Adenylate activity cyclase
pmol/min/mgprotein
138 39 40 39 39 39 132 118 112
89 71 46
These results aresimilar to thoseobserved when GTPyS was included in the incubation as shown in Fig. 2. The GTPySmediatedchange in tryptic digest patterns follows a time course similar to thatfor adenylate cyclase activation. Isoproterenol accelerates both effects with maximal activation of cyclase, the ability to generate the M , = 41,000 tryptic fragment being virtually complete after a 2-min incubation. With FIG.4. GDPm competition of GTPyS-mediated conformaGTPyS alone in the absence of isoproterenol, maximal acti- tional change of the G protein. Pigeon erythrocyte membranes M , were labeled with [:"I']NAD+ and cholera toxin, preincubated with vation of adenylate cyclase and the ability to generate the = 41,000 tryptic fragment are coordinately slower in onset and GMP and isoproterenol, and washed as described in the legend to reach maximal levels only after the15-min preincubation with Fig. 1 and under "Experimental Procedures." The membranes were GTPyS. Thus, there is a close parallel in the time course of then incubated for 45 min a t 22OC with 100 p~ CDPPS (Lane I), 200 GTPyS activation of cyclase and the time course of confor- p~ GDPPS + 100 PM CTPyS (Lane 21, or 100 PM CTPyS (Lane 3). and then treated with 25 pg/ml of trypsin for 10 min at 22°C. The mational change of the G-protein resulting in the ability of samples were then electrophoresed on 12.5% SDS-acrylamide gels trypsin to generate theM , = 41,000 fragment. andautoradiographed. The largearrows mark the M , = 42.000 Fig. 6B represents 72-h autoradiographs of the gels shown cholera toxin substrate, and the small a r r o m mark the specific M , in Fig. 6A. Several new lower abundance fragments are ob= 41,000 tryptic fragment.
Components Coupling
of
Iy)
ISOPROTERENOL
ADOITION
24 1
241
I6
I6
P
2
\
I
s
2
IS
s
IS
._E E '
C
1463
served whose formation is affected by either GTPyS or isoproterenol. A new M , = 30,000 fragment is observed whose intensity is decreased by GTPyS. In addition, isoproterenol in the absenceof GTPyS causesa time-dependent enhancement of a t least two specific bands (designated by arrows in Lanes 3, 7, and 2 2 ) . Theaddition of GTPyS or the removal of isoproterenol causes a marked reduction in the intensity of these bands. In other experiments (not shown), propranolol, a /I-adrenergic antagonist, does not cause the appearance of these fragments and inhibits effect the of isoproterenol. These results indicate that the effect of isoproterenol is mediated through the 8-adrenergic receptor. We demonstrated previously (18) that tryptic fragments M , = 42,000 peptide using membranes not generated from the incubated with GTPyS remain bound to the membrane. This
B
A
Adenylate Cyclase
D
_~
"
2
s
I5
buffer a t 5 mg/ml containing the following additions: A , none; B, 10 p~ isoproterenol; C, 50 PM GTPyS; D,10 p~ isoproterenol and 50 p~ G T P y S T h emembranes were incubated a t 3OoC for 2,5, and 15 min, and were then washed three times by dilution in ice-cold membrane buffer and centrifugation. The membranes were then resuspended in membrane buffer and adenylate cyclase activity was measured during a 30-min incubation in the presence of no addition (0).50 p~ GTPyS (O), or 50 p~ GTPyS and 10 p~ isoproterenol (A). Values are the means of triplicate determinationswhich varied by less than 10% and are representative of four separate experiments.
IS
M~nutas
L
FIG. 5. Persistent activation by GTPyS of adenylate cyclase activity inpigeon erythrocyte membranes. Membranes were treated with cholera toxin and NAD and then washed three times in membrane buffer. Membranes were then resuspended in the same A
2 MIN I
2
15 M I N
3
I
4
9
IO
I1
12
I FIG. 6. Tryptic digestion products of the M , = 42,000 radiolabeled cholera toxin substrate during thetime course of adenylate cyclase activation. Pigeon erythrocytemembranes
5 MIN
2MIN 1
1
2
3
4
1
5
6 7 8
15 M I N 9 IO II
12
r
4
were labeled with [:"PINAD and cholera toxin as described under "Experimental Procedures."Labeled membranes were washed three times in membrane buffer and then treated exactly as described in the legend to Fig. 5. After the appropriate preincubationtime, membranes were washed three times and treated with 25 pg/ml of trypsin for 12 min a t 22'C. The samples were then electrophoresedon 12.5% SDS-acrylamide gels and autoradiographed. The samples in each lane correspond to membranes that were used for the assay of adenylate cyclase activity with no further additions in the secondincubation (Fig. 5, open circles). Lanes represent the following additions in the preincubation: I, 5, and 9, none; 2, 6,and 10.50 p~ GTPyS; 3, 7, and 11, 10 p~ isoproterenol; 4, 8, and 12, 50 p~ GTPyS plus 10 p~ isoproterenol. A , an 18-h exposure; B, an autoradiograph of the same gel exposed for 72 h. Large arrows without tails designate an approximate M , = 30,000 tryptic fragment of the M , = 42,000 peptide. Small arrows with tails designate two fragments that appear to be specific for the interaction of receptor. hormone complex with the G protein in the absence of guanine nucleotide.
Components Coupling
1464 I 2
3 4
t
t
FIG. 7. Release of a proteolytic fragment of the G-protein from pigeon erythrocyte membranes. Pigeon erythrocyte membranes were labeled with cholera toxin and [”PINAD as described under “Experimental Procedures.” Membranes were incubated for 20 min at 3OoC with (Lanes 2 and 2) or without (Lanes 3 and 4 ) GTPyS (100 PM) and isoproterenol (10 p ~ and ) subsequently incubated for 12 min at 22OC with 25 p g / d of trypsin. The membranes were pelleted by centrifugation at 135,000 X g for 30 min. Supernatants (Lanes 2 and 4 ) and resuspended pellets (Lanes 1 and 3) were solubilized and analyzed by SDS-acrylamide gel electrophoresis. The arrows indicate the M , = 42,000 and 41,000 toxin-labeled peptides.
NO ADDITION
GTPXS GDPBS
1 3 1 E
1-
l
”
r
L
n
‘-
’
1)“
4-
;:*1 8.
c:,
.,.
l
aw
FIG. 8. Tryptic digestion of the labeled M , = 42,000 cholera toxin substrate of human erythrocyte membranes. Membranes (2.5mg/ml) were labeled with [:“PINAD’ and cholera toxin as detailed under “Experimental Procedures,” and were washed three times by dilution in ice-cold membrane buffer and centrifugation. The final pellet was resuspended (5 mg of protein/ml) in membrane buffer and incubated for 30 min at 30°C with either no additions (Lanes I and 2). 100 PM GDPpS (Lanes 3 and 4 ) , or 100 p~ GTPyS (Lanes 5 and 6). Aliquots from these incubations were then incubated in the absence (Lanes I , 3,and 5 ) or presence (Lanes 2 , 4 , and 6) of 25 pg/ ml of trypsin for 12 min. Digestion was stopped with aprotinin and the samples were solubilized and subjected to SDS-acrylamide electrophoresis and autoradiographed for 56 h. Nodifference in Coomassie blue staining patterns of high abundance proteins was observed with control or trypsin-treated membranes in the presence or absence of GTPyS or GDPPS. Thesmall arrows M , = 42,000 toxin substrate, and the large arrow indicates the M , = 41,000 tryptic fragment.
was interpreted to indicate that the ADP-ribosylated site probably in the domainof the peptideclosely associated with themembrane. Fig. 7 illustratesthatafterincubation of M , = 41,000 tryptic digestprodmembranes with GTPyS, the uct is released from the membrane and can be recovered in the supernatant after centrifugation. This finding indicates
of Adenylate Cyclase
that GTPyS hasincreased trypsin accessibility of at least one site of the G-protein. This site could be either a small tail of the M , = 42,000 peptide involved inanchoringittothe membrane or a separate anchor peptide which binds the M , = 42,000 labeledpeptide. The G-proteintrypticfragment appears to be inactive and fails to reconstitute adenylate cyclase activity whenmixed with cyc- S49 membranes. Nonetheless, the release by trypsin of a G-protein fragment from the membrane only after GTPyS treatment indicates that guanine triphosphate changes the site of tryptic digestion resulting froma conformational changeof the G-protein. Fig. 8 shows that with human erythrocytes, GTPyS but not GDPPS is capable of mediating the conformational change of the G-protein, resultingin a M , = 41,000 fragment when the membranes are subsequently exposed to trypsin. This finding is significant, becausehumanerythrocytes possess the Gprotein of adenylate cyclase but functionally lack both hormonereceptorsandcatalyticadenylate cyclase (27-29). Therefore, the ability of guanine triphosphates to cause a conformational change of the G-protein does not require its interaction with either hormone receptor or catalytic cyclase. DISCUSSION
We have demonstrated guanine nucleotide-specific changes in susceptibility of the M , = 42,000 cholera toxin substrate to tryptic digestion. A unique M , = 41,000 peptide is generated in the presence of GTPyS. In some experiments, a very small amount of the M , = 41,000 fragment could be detected in the presence of GDPPS. However, this is always less than 5% of that generated in the presence of GTPyS. The formation of the M , = 41,000 fragment can therefore be used to measure guanine triphosphate interactions with the G-protein. Hormone-receptor interaction was shown toaffect the rate of the specific conformational changeinduced by GTPyS, i.e. the generation of the M , = 41,000 tryptic fragment. This finding supports the Cassel and Selinger hypothesis (5) that hormone acts by making the G-protein regulatory site available to guanine triphosphate. In addition to its effects on guanine nucleotide accessibility, isoproterenol in the absence of guanine nucleotide results in a partial trypticdigest pattern of the M , = 42,000 labeled peptide distinguishable from that when GTPyS is present,or when both hormone and guanine nucleotide are absent (Fig. 6 R ) . The simplest explanationfor this finding is that hormone-receptor complex physically interacts with the G-protein. This notion is substantiated by the recent findings of Limbird et al. (30),which demonstrate that the binding of the /3-adrenergic agonist hydroxybenzylisoproterenol to membranes followed by detergent solubilization resulted in an increase in the Stokes radiusof the soluble M , = 42,000 toxin-labeled peptide on gel filtration columns, indicating the association of receptor with the G-protein. The addition of guanine triphosphate resulted in the dissociation of receptor from the G-protein. Our results extendfinding, this in that we have directly demonstrated that GTPyS alters the conformation of the G-proteinso that itis capable of activating adenylate cyclase in a quasi-irreversible manner, and greatly diminishes the influence of hormone-receptor interaction on G-protein tryptic digestion. The simplest interpretation of our results is that guanine triphosphates causea conformational changeof the G-protein resulting in altered partial trypticdigest patterns. Pfeuffer (2) was and Howlett andGilman (31) have recently reported changes in the sedimentation behavior of the G-protein in sucrose gradients following incubation with guanine triphosphates, supporting the notion of conformational changes of the Gprotein mediated by GTPyS. The fact that Pfeuffer (2) was able to observe this effect of guanine triphosphates using
Coupling Components of Adenylate Cyclase pigeon erythrocyte G-protein resolved from catalytic adenylate cyclase suggests that theconformational change is a result of guanine nucleotide interaction with the G-protein, and not interaction of the protein with the catalytic unit. Our results showing the change in tryptic digest pattern mediated by GTP# of the G-protein from human erythrocytes confirms this hypothesis (Fig. 8). The finding that incubation of membranes with GTPyS results in the release from the membrane of a tryptic fragment of the M, = 126,000 G-protein supports the hypothesis fmt suggested by Rosa et al. (1) and substantiated by others (2, 16, 18) that the G-protein has a small hydrophobic domain and is probably a stalked intrinsic membrane protein (16, 18). It is tempting to speculate that thesmall fragment lost from the M, = 42,000 toxin-labeled peptide after tryptic digestion of the GTPyS-activated G-protein is part of the stalk that binds the protein to themembrane. The fact that theprotein released from the membrane is inactive is not surprising, based on our previous observation (18) that there appears to be a second peptide, probably a subunit of the G-protein, which is more sensitive to trypsin thanthe M, = 42,000 peptide which wasrequired for reconstitution with cyc- membranes. A major question still to be answered is how hormone receptors interact with the G-protein. The work of Limbird et al. (30) suggests that /3-adrenergic receptors can physically interact with the G-protein. The findings we have presented here indicate that guanine triphosphates cause a specific conformational change of the G-protein, hormone increases the rate of guanine triphosphate-mediated conformational change, and the receptor-hormone interactions induce changes in tryptic digest patterns of the toxin-labeled M, = 42,000 peptide in the absence of guanine triphosphate. This last observation indicates that receptor-mediated events alter the conformation of the G-protein. The ability to measure conformational changes of the G-protein mediated by guanine nucleotides and hormone-receptor interactions by partial tryptic digestion and peptide mapping provides a new approach to studythe regulation of hormone-sensitive adenylate cyclase. We are now using this technique of partial proteolytic digestion together with reconstitution procedures to combine /3-adrenergic receptors and the G-protein into defined membrane environments in an attempt to gain insight into the molecular mechanisms of coupling. REFERENCES 1. Ross, E. M., Howlett, A. C., Ferguson, K. M., and Gilman, A. G.
(1978) J.Biol. Chem. 253,6401-6412
1465
2. Pfeuffer, T.(1979) FEBS Lett. 101,85-89 3. Johnson, G. L., Kaslow, H. R., and Bourne, H.R. (1978) Proc. Natl. Acad.Sci. U. S. A . 75,3113-3117 4. Ross, E. M., and Gilman. A. G. (1977) Proc. Natl. Acad. Sci. U. S. A. 74,3715-3719 5. Cassel, D., and Selinger, Z. (1978) Proc. Natl. Acad.Sci. U. S. A. 75,4155-4159 6. Spiegel, A. M.,Downs, R. W., Jr., and Aurbach, G. D. (1979) J. Cyclic Nucleotide Res. 5, 3-17 7. Lad, P. M.,Nielsen, T. B., Preston, M. S., and Rodbell, M.(1980) J. Biol. Chem. 255,988-995 8. Rodbell, M., Bimbaumer, L., Pohl, S. Z., andKraus, H. M. J. (1971) J.Biol. Chem. 246, 1877-1882 9. Rodbell, M.,Lm,M. C., and Salomon, Y. (1974) J.Biol. Chem. 249,59-65 10. Cassel, D., and Selinger, Z. (1976) Biochim. Biophys. Acta 452, 538-551 11. Cassel, D., and Selinger, Z. (1979) Biochem. Biophys. Res. Commun. 77,868-873 12. Ross, E. M., Maguire, M. E., Sturgdl, T. W., Biltonen, R. L., and Gilman, A. G. (1977) J.Biol. Chem. 252,5761-5775 13. Cassel, D., and Pfeuffer, T.(1978) Proc. Natl. Acad.Sci. U. S. A . 75,2669-2673 14. Gill, D. M., and Meren, R. (1978) Proc. Natl. Acad. Sci. U. S. A. 75, 3050-3054 15. Johnson, G. L.,Kaslow, H. R., and Bourne, H. R. (1978) J.Biol. Chem. 253,7120-7123 16. Kaslow, H. R., Johnson, G. L., Brothers, V. M., and Bourne, H. R. (1980) J. Biol. Chem. 255, 3736-3741 17. Pfeuffer, T.(1977) J.Biol. Chem. 252, 7224-7234 18. Hudson, T.H., and Johnson, G . L. (1980) J. Biol. Chem. 255, 7480-7486 19. Cassel, D., and Selinger, Z. (1977) Proc. Natl. Acad.Sci. U. S. A. 84,3307-3310 20. Johnson, G. L., and Bourne, H. R. (1977) Biochem. Biophys. Res. Commun. 78,792-798 21. Cleveland, D. W., Fischer, S. J., Kirschner, M. W., and Laemmli, U. K. (1977) J.Biol. Chem. 252, 1102-1106 22. Laemmli, U. K. (1970) Nature (London) 228,680-685 23. O’Farrell,P. H. (1975) J.Biol. Chem. 250,4007-4021 24. Pfeuffer, T., and Helmreich, E. J. M.(1975) J.Biol. Chem. 250, 867-876 25. Eckstein, F., Cassel, D., Levkovitz, H., Lowe, M., and Selinger, Z. (1979) J. Bwl. Chem. 254.9829-9834 26. Cassel, D., Eckstein, F., Lowe, M., and Selinger, Z. (1979) J. Biol. Chem. 254,9835-9838 27. Tsukamoto, T., and Sonenberg, M. (1977) Endocrinology 100 (suppl.), 285 28. Kaslow, H. R., Farfel,Z., Johnson, G. L., and Bourne, H. R. (1979) Mol. Phannacol. 15,472-483 29. Rasmussen, H., Lake, W., and Allen, J. E. (1975) Biochim. Biophys. Acta411.68-73 30. Limbird, L. E., Gill, D. M., and Lefkowitz,R. J. (19RO)Proc. Natl. Acad. Sei. U. S. A. 77,775-779 31. Howlett, A. C., andGilman, A. G . (1980) J. Biol. Chem. 255, 2861-2866
