Feb 25, 1994 - activator Arg259-Lys273 domain of the &-adrenergic receptor. (20) (data .... Storm for informing us of results before publication; Yoshiomi Tamai.
Communication
THEJOURNAL OF BIOLCG~CAL CHEMISTRY Vol. 269, No. 19, Issue of May 13, pp. 13756-13759, 1994 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.
An Intrinsic Guanine Nucleotide Exchange Inhibitor in Gi2a
G proteins are apy forms bound to GDP. Upon receptor stimulation they dissociate into GTP-bound Ga and nucleotide-free GPy, leading to the activationof membrane effectors. The major process in receptor activation of G proteins is the stimulaSIGNIFICANCE OF G-PROTEIN SELF-SUPPRESSION tion of this GDP/GTP exchange; receptors activate G proteins by promoting the guanine nucleotide exchange of Ga without WHICH ANTAGONIZES RECEPTOR SIGNAL* (Received for publication, February 1, 1994,and in revised form, altering its intrinsicGTP hydrolysis activity. However, the accomplish this process is February 25, 1994)mechanism by whichreceptors poorly understood. Takashi Okamoto,“b, e Yoshitake Murayama: Gi2ais an a subunit of heterotrimeric G proteins that inhibits Stephen M. Strittmatter? f , Toshiaki Katada,” adenylate cyclase. Among three known G,a proteins, only this Shigetaka hano: Etsuro Ogata,’ and version is demonstrated to be a proto-oncogene product (2,3). Ikuo Nishimoto“b * j Accordingly, normal G,a has beenimplicated in cell and organ From the “CardiovascularResearch Center and growth, motility, differentiation, and development (4-7). The ‘Developmental Biology Laboratory, Massachusetts present study was conducted to clarify the mechanismwhereby General Hospital and the Departmentsof bMedicine and receptors stimulate guanine nucleotide exchange of Gi2a by fNeurology, Harvard Medical School, Charlestown, Massachusetts 02129, the dFourth Department of identifying the regulatory domains inthis G protein. Internal Medicine, Tokyo University School of Medicine, The first key for the regulatory domains in Gi2ais in its C 3-28-6 Mejirodai, Bunkyo-ku, Tokyo 112, Japan, the terminus, which should contain plural subdomains with differhDepartment of Life Sciences, Tokyo Institute of to a receptor contact site (8, 9). The ent functions, in addition Technology, 4259 Nagatsuta, Midori-ku, Yokohama 227, next key would be Gia activator sequences in receptors. IGFJapan, and the ‘Cancer Research Institute, 1-37-1 IIR activates Giz through the receptor’s Arg2410-Lys2423 region Kami-Zkebukuro, Ibshima-ku, Tokyo 170, Japan which by itself is able to activateG , p (12, (peptide 14) (10-121, The a subunit of G , ( G i z a )is a member of the hetero- 13). The M, mAChR possesses the GJG, activator region (MIII) trimeric G protein family, which transduces receptor at residues 382-400 (14), which functions as a G, activator at signals as a proto-oncogene product. We have found a higher concentrations(15).We (12)have defined the structural novel self-suppressiveregion in Gaa near its C terminus. determinants for the G protein-activating function of polypepA polypeptide consisting of residues 338-352 of Gaa tides as (i) 2 basic residues at the N-terminal side of amino acid (G,a-338-352) antagonizes receptor- and receptor peptide-stimulated G,a activation, without affecting basal sequence and (ii)a C-terminal motif of B-B-X-B or B-B-X-X-B activity.Antagonismby G,a-338-352 is attributableto an (where B is a basic residue, andX is a nonbasic residue). Thus, interaction with activated Giza,which is not competitive receptor contact sites of Ga may share a structure matching with receptor polypeptides.Combined with the reports this motif. The third key has been provided by Heideman and suggestingthe presence of self-suppressive domains in a Bourne (1).They proposed that the C-terminal 36 residues of juxta-C-terminal portionof Gizaand Goa, this study sup- Gizacontain not only a receptor contactsite buta domainwhich ports the hypothesis that Giza-338-352constitutes an in- suppresses GDP/GTP exchange by the 01 subunit, based on the trinsic guanine nucleotide exchange inhibitor, which in study of Woon et al. (16). It would be important t o clarify the turn antagonizes receptor stimulation,suggesting that mechanism of the C terminus that suppresses guanine nucleotide exchange of Gi2a. G proteins are activated by receptors through relaxation of a self-suppressiveconformation. These prompted us to focus on the region Asp338-Cys352 (DAVTDVIIKNNLKDC) in Gi2a to examine the effect of this domain in suppressing guanine nucleotide exchange of this G Guanine nucleotide-binding proteins (G proteins)’are a fam- protein. We initially hypothesized that this might serve as a ily of signal transducers of receptors (1). They are oligomers receptor contact site. The dataunexpectedly indicate that this consisting of a (Gal and py subunits (GPy). In the basal state, juxta-C-terminal region is an intrinsicGDP/GTP exchange inhibitor of Gi,a which antagonizes receptor stimulation. The * This work was supported in part by Bristol-Myers Squibb and by receptor-antagonizing, self-suppressive functionof the intrinsic grants from the Ministry of Education, Science, and Cultureof Japan, sequence in Gizasuggests a novel mechanism for receptor acChugai, and the Uehara Foundation. The costs of publication of this tivation of the G protein. article were defrayed in part by the payment of page charges. This EXPERIMENTAL.PROCEDURES article must thereforebe hereby marked “aduertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Syntheticpolypeptides were produced by the solid-phase method and Recipient of fellowships fromthe Byotai-Taisha Foundation andthe purified by high performance liquid chromatography to >95% purity. G,, Mochida Memorial Foundation. was trimeric Gi2purified from bovine brain to near homogeneity (171, Present address: Department ofNeurology, Yale University School of which was stored in buffer A (20 mM Hepes/NaOH (pH 7.4), 0.1 mM Medicine, New Haven, CT 06520-8018. EDTA, and 0.7% CHAPS) and diluted 210-fold for assays. G,, purified ’ To whom correspondence should be addressed: CVRC, MGH-East, from bovine spleen to homogeneity (18)(provided by Dr. Tomiko Asano, 149 13th St., Charlestown, MA02129.Tel.: 617-726-4348; Fax: 617-726Institute for Developmental Research, Aichi Prefectural Colony, Aichi, 5806. The abbreviations used are: G protein, guanine nucleotide-binding Japan) was also used. They yielded similar results.The Gpy-free, moprotein; Ga, a subunit of G protein; GPy, By subunit of G protein; nomeric Giza purified to homogeneity from bovine lung (19)was also provided by Dr. Tomiko Asano. TrimericG, and Go purified from bovine IGF-IIR,insulin-likegrowthfactor I1receptor;mAChR,muscarinic acetylcholinereceptor;GTP-yS,guanosine 5’-0-(3-thiotriphosphate); brain were described in the same reference as G, and were stored in CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-l-propanesulfonic buffer A. GTPyS binding to G proteins was assayed at 37 c in buffer C (20mM Hepes/NaOH (pH7.4),130p~ MgCI,, 110 p~ EDTA, and 60 nM acid; GDI, GDP dissociationinhibitor. ~~~
~~~
E
13756
G Protein
A GDPfGTP Exchange Inhibitor Region Dimeric in a
13757
[%]GTPyS) as described (14). The method used for calculation of the rate constantfor GTPyS binding(k,,,), which is equalt o the slope of the tangent to the GTPyS-binding curve at time zero and represents the actual GTPyS binding rate, was described (14), using t = 5 min for trimeric G,, 3 min for Gpy-freeG,,a, 10 minfor trimeric G,, and 2 min for trimeric Go.GDP release was measured a s described (13). The reaction was terminated by adding 100 m~ Tris/HCl (pH 8.0), 25 m~ MgCl,, 100 mM NaCl, 10 m~ NaF, and 100 PM AlC1,. Steady state GTP hydrolysis was measured a s described (20). The mAChR purified from porcine brain t o near homogeneity (21) was kindly provided by Dr. Tatsuya Haga (Institute for Brain Research,Tokyo University, Japan) and was reconstituted in phospholipid vesicles using gel filtration a s at least three times, described (10). The experiments were repeated unless otherwisespecified, and yielded similar results. RESULTS AND DISCUSSION
The polypeptide corresponding to the region Asp338-Cys352 (referred to as Gi,a-33%352) suppressed GTPyS binding to trimeric G,, promoted by peptide 14 witha n IC,, of -6 PMand did so completely at 30 PM (Fig. lA).Peptide 14corresponds to the Giza-specific activator Ar$41n-Lysz423domain (12) in IGFIIR, which is a single-spanning G,-coupled receptor (11).Giza338-352 did not affect the basal binding of GTPyS. This Giza polypeptide inhibited peptide 14-stimulated GDP release and turnover number of GTP hydrolysis by G,, with the same potency as observed for GTPyS binding (data not shown). As controls, we tested severalGa-derived polypeptides. First, we synthesized the extreme C-terminal10-residue polypeptide of Giza,Giza-346-355 (KNNLKDCGLF). As antibodies against these polypeptides selectively block receptor-G protein interactions, Gjza-34&355 might well affect peptide 14-G,, coupling. However, G,,a-34&355 did not alterG,, activation. Second, G p 377-391, a G,a polypeptide comparable to Giza-338-352, which has the same length, showed no inhibition or rather weak stimulation at high concentrations. We also tested 20-residue Ga-derived polypeptides, G,,a-1-20 and G,a-1-20. Both are theN termini of the cognate Ga chains, and exhibited no inhibition, even at 100 PM. These results suggest that Giza-338-352 carries a polypeptide-specific function. Combined with the result that Giza-338-352 had no effect on G, activation a t 100 p~ (see below), it is thereforeunlikely that blockade of G,, activation by Giza-338-352 is through nonspecific interference by high peptide concentrations on stimulus-G protein communication. In an effort to demonstrate thespecific function of Giza-338352, we examined the function of the residue-altered variants (Fig. 1B). Even when Ala was substituted for Asp at position 342, the variant DAVTAVIIKNNLKDC, referred to as A342, showed unaltered action potency. In contrast, when Ala was substituted for Asp at position 338, which is the N-terminal residue of Giza-338-352, the A338 variant indicated more than 10 timeslower potency. The AC352 variant lacking the Cys352at the C terminus of Gi,a-338-352 also showed lower potency. , Althoughboth showed significantinhibition a t 100 p ~ the difference in potency is notablebecause these polypeptides are 93% identical to Giza-338-352. It should be further noted here that another variant,E351, which has anGlu substitution for Asp at position 351, did not inhibit G,, activation at up to 300 p ~ The . same E351 inhibited peptide 14-stimulated k,,, of GTPyS binding toG, by 42% a t 30 PM under the same condition (data notshown). These resultsclearly indicate thatGiza-338352 has a sequence-specific and thereby a polypeptide-specific function. Promotion of GTPyS binding to G,, by MI11 was also attenuIC,, ated by Giza-338-352 in a dose-dependent manner with an of 5 p~ (data not shown). MI11 is a polypeptide that corredomain (14) in the sponds to the GJG, activator Ar$8z-Ar&"'' M, mAChR, which is a typical multi-spanning G,-coupled receptor. G,,a-338-352 also inhibited Gi, activation induced by purified mAChR (Fig. 1C). InmAChR-G,, vesicles, the turnover
-
0
1
10
(G&38"216"
lW
M"1 AIF;
w.-ryI
- + + + - +
frep -1OmM-
-
-
FIG. 1.Effect of G,a-338462 on G protein activation.A, effect of G,a-33&352 and control polypeptides on GTPyS binding to G,, stimulated by peptide 14; E , effect of structural variantsof G,,a-338-352; C, effect of G,a-338-352 on steady state GTP hydrolysis of G , in mAChRGiz vesicles; D , inability of Gj2a-338-352 to antagonize the action of AlFq. A, trimeric G,, a t 10 nM was incubated with 30 PM peptide 14 (RVGLVRGEKARKGK, the G, activator domain of IGF-IIR) in the presence of Gi,a-338-352 (DAVTDVIIKNNLKJIC, El), the C-terminal 10-residue polypeptide of Giza(G,,a-346355, KNNLKDCGLF, O),G p 377-391, a 15-residue polypeptide in G p corresponding to Giza-338352 (NDCRDIIQRMHLRQY, V), ortwo20-residuecontrol polypeptides, Giza-1-20 (MGCTVSAEDKAAAERSKMID, and Goa-1-20 (MGCTLSAEERAALERSKAIE, 0) for 5 min and GTPyS binding toG proteins was measured. The effect of G,,a-33&352 on Gi, in theabsence of peptide 14 is indicated by A. In panels A and B, the extentof stimu" lation was expressed as the degree of stimulation of the rate constant k,,, relative t o the basal k,,,. The basal k,,, value for trimeric G,, was 0.10 2 0.01 min" in panel A and was simllar in B . All values in the 2 S.E. of three experiments. B, figure and its legend represent the mean G, was incubated with 30 pv peptide 14 in the presence of various concentrations of G,,a-338-352 or its structural variants, and GTPyS binding to Gi, was measured under the conditions described in the legend. Theeffect of Gjza-338-352 is indicatedby m. Structural variants are DAVTAVIIKNNLKDC (A342, ), AAVTDVIIKNNLKDC (A338, E), DAVTDVIIKNNLKD (AC352, 0 ), and DAVTDVIIKNNLKEC (E351,O). GTPyS binding to each of these synthetic peptides was negligible in the concentration range where they were used. The experiment with E351 was repeated six times and yielded similar results. C , the vesicles reconstituting mAChR and G, were incubated with 100 PM carbachol for 20 min in thepresence of various concentrationsof G,a-338-352 (El) or Gi,a-346-355 (O), and the GTP hydrolysis activity was measured. The effect of G,,a-338-352 in theabsence of carbachol is indicated by . The GTP hydrolysis activity was expressed as a percentage of the basal activity. The basal turnover number was 0.05 min". In phospholipid vesicles used in this experiment, the concentration of mAChR reconstituted was3 n~ and thatof G,2was 10n ~E ., Gi, was incubated with 10 m~ NaF and 100 pv AlCl, in the absence or presence of 30 p~ G,a338-352 at free or 10 m~ Mg2' a t 37 "C for 5 min, and GTPyS binding to G proteins was measured. In this panel, the absolute values of the rate constant k,,, were indicated.
+
number of GTP hydrolysis of G,, maximally stimulated by carbachol was attenuated by Giza-338-352 dose dependently with . at 100 p~ abolished the effect a n IC,, of 5-10 p ~ Giza-338-352 of 100 p~ carbachol, whereas thispolypeptide did not affect the basal turnover number. GiZa-34W55 againfailed to antagonize carbachol-stimulated GTP hydrolysis activity. All of these findings observed in mAChR-G, vesicles were also evident in the GTPyS binding assay (data notshown). Incontrast, G,,a-338-352 maynotinhibitnon-receptor stimulation of G,,. M e stimulates GTPyS binding toGiz(Fig. I D , left two columns), whereas G,,a-338-352 had no effect on this stimulationof G,, (Fig. I D , middle column). It is also well
13758
A GDPIGTP Exchange Inhibitor RegionDimeric in a
G Protein
known that AIF, activates GDP-bound G proteins by mimick400 a non-receptor ing they-phosphoryl group of GTP, that is, with mechanism (22). AlF; effects on guanine nucleotide binding to Gi, were not blocked by Giza-338-352 (Fig. I D , right two columns). This is consistent with the finding that Giza-338-352 did not inhibit the basal GTPyS binding to Gi,. Therefore, the antagonizing action of Giza-338-352 should be specific for receptor-related stimulation. Giza-338-352 wassignificantly less potent in blocking Go activation with little effect at 5 1 0 VM on GTPyS binding to Go stimulated by MIII, and with the IC,, (=20 VM) about 4-fold 350, 1 higher than for Gi, activation (data notshown). Gi2a-338-352 had little effect on GTPyS binding to G, promoted by the G, activator Arg259-Lys273domain of the &-adrenergic receptor (20) (data notshown). These findings suggest that this inhibitory polypeptide of Giza has restricted G protein selectivity. The three simplest mechanisms whereby Gi2a-338-352 inhibits receptor stimulation ofG,, are as follows: (i) Giza-338352 binds to receptor polypeptides and preventsreceptor interaction with Giz;(ii) Giza-338-352 binds to Gi, at the same site as 0 0.5 1.0 1.5 2.0 0 1 10 100 receptors and is a competitive antagonist of receptor activation; ratio (RecepiorIG) [+3363521 W) and (iii) Giza-338-352 binds to G,, at a site distinct from the FIG.2. The mechanism of G,a-338352 antagonism. A, dose-rereceptor contact site and non-competitively inhibits receptor sponse curves of Gi, activation by peptide 14 in the presence of G,,astimulation. 338-352; B , dose effect of Gi,a-338-352 on mAChR-Gi, coupling; C , Multiple lines of evidence indicate that Giza-338-352 binds replot of B to show the dose effect of mAChR on GI, in the presence of to Gi2 and not receptor polypeptides. The IC,, values of the Giza-338-352;D , effect of Gi,a-338-352 on monomeric Giza.A, by incuantagonizing effect of Giza-338-352 were 1 order of magnitude bating 10 nM Gi, with various concentrations of peptide 14 in the absence ( I 3 or presence of 1 PM ( ), 3 PM (0).10 PM ( 0 ), or 30 PM).( lower than the concentration of receptorpeptides used for Gi,a-338-352, GTPyS binding for 5 min was assayed under the same stimulation, which fits poorly with the idea that Gi2a-338-352 conditions described in legend to Fig. 1 . The extent of stimulation was binds to receptor peptides in a 1:l molar stoichiometry. The expressed as the degree of stimulation of the rate constantk,,, relative significant difference between the potencies of GiZa-338-352on to the basal k,,,. The basal k , ,value was 0.10 min" in this condition. All values shown represent t i e mean t S.E. of three experiments. In MIII-induced activation of Go and that of Gi, supports the no- panel A, the error bars representing S.E. were not indicated, which were tion that Giza-338-352 acts on G proteins rather than on a
