An Antibody Directed against Residues 100–119 within the α-Helical Domain of Gαs Defines a Novel Contact Site for β-Adrenergic Receptors*

A polyclonal antiserum that recognizes residues 100–119 within the α-helical domain of Gαs(K-20) caused a dissociation of Gs into its component subunits and activated a cholera toxin-sensitive high affinity GTPase. Consistently, the antibody mimicked the stimulatory effects of the β-adrenergic agonist, isoproterenol, on adenylyl cyclase, which is mediated by Gαs, and its inhibitory action on NADPH-dependent H2O2 generation, a Gβγ-mediated response. A peptide corresponding to the target sequence of K-20 not only neutralized the receptor-mimetic effects of the antibody but inhibited the whole spectrum of isoproterenol action as well, including its antagonistic effects on adenylyl cyclase and NADPH-dependent H2O2 generation. By contrast, COOH-terminal anti-Gαs selectively inhibited the stimulatory effect of isoproterenol on cAMP formation without affecting its inhibitory effect on NADPH-dependent H2O2 generation. The data are consistent with the concept that β-adrenergic receptors interact with multiple sites on Gαs each playing a distinct role, and strongly suggest that antibody K-20 defines a novel contact site for β-adrenergic receptors that localizes to the α-helical domain and is essential for eliciting the complete spectrum of β-adrenergic responses.

Heterotrimeric G proteins composed of ␣and ␤␥-subunits transduce signals from cell surface receptors to downstream effectors and regulate intracellular membrane transport events (1)(2)(3)(4)(5)(6). Interaction of ligand-occupied receptors with heterotrimeric G proteins triggers the exchange of GTP for GDP on the ␣-subunit, leading to a sequential release of G␣-GTP and the stable ␤␥-complex from the receptor. The released G protein subunits are then able to interact with distinct effector enzymes and ion channels. G protein activation is terminated by hydrolysis of GTP by the intrinsic GTPase activity of the ␣-subunit, leading to reassociation of G␣ and G␤␥. The cycle is then complete, and the heterotrimeric G protein is able to be activated again.
The structural determinations for several members of the family of heterotrimeric G proteins have shown that their ␣-subunits are composed of two domains (1)(2)(3)(4)(5)(6). The core domain contains regions with sequence similarity to other GT-Pases and has a structure very similar to Ras and elongation factor Tu (1)(2)(3)(4)(5)(6). The ␣-helical domain is unique to ␣-subunits of heterotrimeric proteins and not present in other GTPases. It is therefore tempting to assume that the latter domain may be important for specific functions of heterotrimeric G proteins that are not shared by other members of the GTPase superfamily, such as coupling to heptahelical receptors. Surprisingly, current modeling suggests that receptor-G protein coupling is exclusively by the core domain of G␣ and segments of the ␤and ␥-subunits, however (1)(2)(3)(4)(5)(6). The function of the ␣-helical domain is still under investigation. Evidence has been presented to suggest that it may influence the spontaneous rate of GDP release (7), and it has been proposed that it may function as a GTPase-activating protein (8), or may be involved in effector regulation (9).
A screening of different G␣ s antibodies for their applicability in studying insulin receptor-NADPH oxidase coupling revealed that one of the commercially available antibodies (K-20), which recognizes residues 100 -119 within the ␣-helical domain of G␣ s , mimicked the effects of inhibitory ligands, such as the ␤-adrenergic agonist isoproterenol, on NADPH-dependent H 2 O 2 generation that are transduced by G s . We therefore explored whether the epitope recognized by K-20 may define a new contact site for activated ␤-adrenergic receptors or be involved in an alternative route of G s activation that may be utilized for a receptor-independent activation of G s by intracellular pseudoreceptors or accessory proteins, for example (10 -15).

Materials
The characteristics and sources of antibodies and G s -derived peptides used in the current experiments are listed in Table I. Forskolin, 7␤-deacetyl-7␤-(␥-(morpholino)butyryl)hydrochloride, was from Research Biochemicals International and cholera toxin (A-subunit) from Calbiochem.

Methods
Subjects, Preparation of Fat Cells, and Fat Cell Ghosts-Experimental details have been described in detail elsewhere (16). Briefly, adipose tissue was from nondiabetic subjects undergoing elective abdominal or cosmetic breast surgery. The specimens were cut into small pieces, and fat cells were isolated in a HEPES-buffered Krebs-Henseleit solution, pH 7.4, containing 20 mM HEPES, 10 mM NaHCO 3 , 5 mM glucose, 20 g/liter albumin, and 1 mg/ml collagenase (CLS, Worthington). After 30 min, fat cells were washed and resuspended in 10 volumes of an ice-cold lysing medium containing 20 mM MES, 1 pH 6.0, 2 mM MgCl 2 , 1 mM CaCl 2 , 5 mM KCl, and 100 mg/liter soybean trypsin inhibitor. Cell lysis was completed by mechanical shaking, and fat cell ghosts were collected by low speed centrifugation (1,000 ϫ g, 4°C, 20 min).
Receptor-mediated Modulation of NADPH-dependent H 2 O 2 Generation-A two-step procedure was used, as reported elsewhere (16). Plasma membranes from adipocytes were first exposed to hormones and were then assayed for NADPH oxidase activity. The activation step was carried out in 30 mM MOPS, pH 7.5, containing 120 mM NaCl, 1.4 mM CaCl 2 , 2.5 mM MgCl 2 , 10 mM NaHCO 3 , and 0.1% human albumin. Membranes were first incubated with 5 nM insulin in absence or presence of 5 M isoproterenol for 5 min to allow receptor occupation. Thereafter, 50 M GTP␥S was added. After 20 min, ghosts were collected by centrifugation, washed, and then resuspended in 30 mM MES, pH 5.8, containing 120 mM NaCl, 4 mM MgCl 2 , 1.2 mM KH 2 PO 4 , 1 mM NaN 3, 250 M NADPH, and 10 M FAD for determination of NADPHdependent H 2 O 2 generation.
To assess the effects of G protein antibodies and peptides corresponding their target sequences on NADPH-dependent H 2 O 2 generation, membranes were exposed to both types of agents at 4°C for 45 min, as indicated in the legends to figures, and were then subjected to the two-step procedure described above.
Determination of Adenylyl Cyclase Activity-Adenylyl cyclase activity of human fat cell plasma membranes was determined in 30 mM Tris-HCl, pH 7.5, containing 1 mM ATP, 2.5 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM 3-isobutyl-1-methylxanthine, 10 M GTP, 10 mM creatine phosphate, 0.1 mg of creatine kinase, and 0.1% bovine serum albumin in a final volume of 100 l. Reactions were initiated by addition of 5-8 g of membrane protein and were continued for 15 min at 37°C in the absence or presence of isoproterenol, as indicated. Reactions were terminated by 100 l of ice-cold perchloric acid (5%). cAMP was determined by radioimmunoassay (Amersham Pharmacia Biotech) after neutralization. To assess the effect of G␣ s -derived peptides and antibody K-20 on adenylyl cyclase activity, membranes were pretreated with the K-20 antibody or peptides, as described above.
GTPase Activity-GTP hydrolysis was determined essentially as described by Jakobs and Aktories (20). Untreated or CTX-treated membranes were preincubated (20 min, 4°C) in 20 mM Tris-HCl, pH 7.6, containing 2.5 mM MgCl 2 , 0.5 mM EDTA, and 100 M N-ethylmaleimide (NEM) to inactivate G i . After washing, NEM-treated membranes were exposed to antibody K-20 or the peptide corresponding to its target sequence for 40 min at 0°C, as indicated. For determination of GTPase activity membranes (5-10 g of protein) were incubated in 0.1 ml Tris-HCl (20 mM, pH 7.6) containing 0.2 M GTP, 0.5 mM ATP, 0.5 mM AMP(PNP), 2.5 mM Mg Cl 2 , 1 mM EDTA, 1 mM dithiothreitol, 5 mM creatine phosphate, 0.025 mg of creatine kinase, and [␥-32 P]GTP (0.1-0.2 mol) at 37°C for 10 min. Reactions were terminated by addition of 25% (w/v) activated charcoal. Release of 32 P i was determined by counting aliquots of the supernatants for radioactivity.
Low affinity GTPase activity was determined by measuring the rate of GTP hydrolysis at 50 mM GTP. Less than 16% of total GTP hydrolysis was due to low affinity GTPases under the conditions used.
K-20-induced Dissociation of G s -In contrast to other G proteins, G␣ s does not dissociate in the presence of GTP␥S alone under the conditions used in the present experiments (17,18). Plasma membranes were therefore suspended in activation buffer containing 50 M GTP␥S, as described above, and incubated in the absence or presence of various concentrations K-20 for 25 min at ambient temperature, as indicated in the legend to Fig. 1.
Proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred to Hybond polyvinylidene difluoride membranes. Western blotting was performed using anti-G␣ s (RM/1) and anti-G␤ (SW/1 from NEN Life Science Products) antibodies. Bands were visualized by chemiluminescence using the ECL kit from Amersham Pharmacia Bio-tech and were quantified by densitometric analysis. The amount of G␤ was normalized to the amount of G␣ s immunoprecipitated for each condition.

RESULTS
Antibody K-20 selectively precipitated G␣ s ; neither G␣ i nor G␣ o was detectable after precipitation of solubilized membrane proteins by K-20 (data not shown). Fig. 1 shows that antibody K-20 was not only capable of specifically recognizing G␣ s but also promoted dissociation of G s into its component subunits in the presence of GTP␥S. Membranes were pretreated with different K-20 concentrations and monitored for dissociation of G s by a immunoblot technique utilizing antibody RM/1, which does not discriminate between heterotrimeric and monomeric G␣ s (17,18). Fig. 1 demonstrates that the amount of G␤ recovered in RM/1 immunoprecipitates was selectively reduced in membranes that had been pretreated with K-20 in the presence of 50 M GTP␥S. In the absence of GTP␥S, the antibody had no influence on the subunit composition of G s , indicating that K-20 caused a guanine nucleotide triphosphate-dependent dissociation of G s , which is characteristic for a receptor-mediated activation (1)(2)(3)(4)(5).
The latter observation suggested that K-20 may bind to and activate G s in a manner similar to activated heptahelical receptors. To corroborate this latter hypothesis, untreated and cholera toxin-treated membranes were incubated with increasing concentrations of K-20 and assayed for high affinity GTPase activity in the presence of NEM, which inactivates G i (20). The antibody caused a concentration-dependent increase in GTP hydrolysis, which was abolished after treatment with cholera toxin ( Fig. 2A). At a maximal concentration (1:1,000), the antibody increased GTPase activity by approximately onethird, which is similar to the extent of activation seen in the presence of a maximal concentration of the ␤-adrenergic agonist isoproterenol (Fig. 2B).
Consistent with its stimulatory effects on GTP hydrolysis by G s , K-20 activated adenylyl cyclase activity in human fat cell plasma membranes over the same range of concentrations that were effective in stimulating GTPase activity (Fig. 3A); the maximal effect of the antibody was similar to that seen in the presence of isoproterenol (16). Activation of adenylyl cyclase was not the sole isoproterenol-like effect of K-20 (Fig. 3B). The antibody mimicked the inhibitory effect of the ␤-adrenergic agonist on NADPH-dependent H 2 O 2 generation as well, which is mediated by ␤␥-subunits (16). At maximal concentrations, both isoproterenol and K-20 inhibited the stimulatory effect of insulin on NADPH-dependent H 2 O 2 generation completely, and their effects could be reversed by G␤-antibodies or agents that specifically bind G␤␥ 2 . Thus, K-20 mimicked the action of activated ␤-adrenergic receptors in every aspect studied, suggesting that the epitope recognized by K-20 may either define a new contact site for ␤-adrenergic receptors on the ␣-helical domain of G s or be involved in an alternative route of G s activation that may be utilized for a receptor-independent activation of G s .
The stimulatory effect of K-20 on GTP hydrolysis could be neutralized by a peptide corresponding to its target sequence, as expected (Fig. 4A). Surprisingly, the peptide not only neutralized the effect of the antibody on GTPase activity but inhibited isoproterenol action on GTP hydrolysis as well (Fig.  4B). At a concentration of 2.5 M, the peptide suppressed the stimulatory effect of isoproterenol over the whole range of concentrations tested, suggesting that the epitope recognized by K-20 may in fact be essential for interaction of activated ␤-adrenergic receptors with G s (Fig. 4B).  (Table I) on isoproterenol-stimulated rates of cAMP formation and its inhibitory effect on NADPH-dependent H 2 O 2 generation. At the concentrations used, both peptides decreased isoproterenol induced cAMP production to basal levels (Fig. 5A). The inhibitory effect of isoproterenol on NADPH-dependent H 2 O 2 generation was also reversed by peptide 100 -119 (Fig. 5B). Surprisingly, the COOH-terminal decapeptide was only effective in inhibiting cAMP formation, a G␣ s -mediated response, but failed to influence the inhibitory effect of isoproterenol on NADPH-dependent H 2 O 2 generation, which is mediated by ␤␥-subunits (Fig. 5B). The effect of peptide 100 -119 was specific, inasmuch as this fragment had no effect on forskolin-stimulated rates of cAMP production (Fig. 6). DISCUSSION This report demonstrates that polyclonal antibodies directed against a segment of the ␣-helical domain encompassing residues 100 -119 of G␣ s caused a dissociation of G s into its component subunits, stimulated a cholera toxin-sensitive high affinity GTPase, and were as efficient in stimulating adenylate cyclase activity as the ␤-adrenergic agonist isoproterenol, suggesting that this antibody binds to and stimulates G s in manner similar to that of ligand-occupied receptors. Consistently, the antibody not only stimulated adenylate cyclase but mimicked the inhibitory action of isoproterenol on NADPH-dependent H 2 O 2 generation as well, which is mediated by G␤␥, indicating that activation of a single G protein, G s , can provide enough G␤␥ for eliciting a G␤␥-mediated response, which has been questioned (21). With one possible exception (22), K-20 is the first example of an antibody exhibiting receptor-mimetic effects, a property that should be extremely useful in elucidating the role of G s in processes where receptors have not yet been identified, such as membrane traffic (11,23) or complex cellular responses, including cell differentiation (24).
The antigenic site recognized by antibody K-20 encompasses the distal end of helix A and the beginning of the following loop, a region of general sequence diversity among G␣-subunits that is freely accessible and seems to be poorly ordered in crystals of G␣ s -GTP␥S (25). Interestingly, in G␣ i2 , an adjacent region, e.g. the helix B-helix C segment, undergoes substantial structural changes upon GTP hydrolysis resulting in an opening of the nucleotide binding pocket (26). However, current modeling suggests that the antigenic site recognized by K-20 may be too distant from the plasma membrane (Ͼ35 Å) to be involved in direct physical contact with activated ␤-adrenergic receptors (5,6), raising the interesting possibility that helix A and the beginning of the following loop may be utilized for a receptorindependent activation of G s . Indeed, it is well established that the activity of G proteins may be directly modulated by a diverse group of proteins, including terminal complement complexes (13), presenilin (15), neuromodulin (11), tubulin (27), caveolins (28,29), not yet identified proteins (12), or amphiphilic small molecular weight compounds, such as mastoporans (10), carbolins (30), and taste substances (14). Surprisingly, the peptide corresponding to the antigenic site of K-20 (residues 100 -119) not only neutralized the action of the antibody but impaired ␤-adrenergic receptor signaling via both component subunits of G s as well, which resulted in inhibition of all effects of isoproterenol tested, including its antagonistic effects on adenylyl cyclase and NADPH-dependent H 2 O 2 generation. Peptides that are effective in blocking receptor-G protein interaction are believed to mimic interfacial contact sites between the proteins. Indeed, this latter type of peptide competition now serves as a standard approach to identifying protein interaction pairs and their recognition sequences (31). Although provocative, it therefore appears safe to conclude that peptide 100 -119 competed with G␣ s for activated ␤-adrenergic recep-tors, implying that the epitope recognized by antibody K-20 within the ␣-helical domain is directly accessible for this class of heptahelical receptors. Consistently, the peptide did not influence the stimulatory effect of forskolin which activates adenylyl cyclase directly.
As pointed out above, the conclusion that peptide 100 -119 competed with G␣ s for activated ␤-adrenergic receptors implies that current ideas regarding the orientation of heterotrimeric G s relative to the membrane and/or receptors may require careful reevaluation (1,2,4,6). Much current thinking is based on the idea of freely mobile receptors, G proteins, and effectors (1)(2)(3)(4)(5)(6). Based on this concept, current models predict that the heterotrimer face that interacts with receptors comprises the NH 2 and COOH termini of G␣, exposed residues on helix ␣ 5 and strand ␤ 6 , as well as the COOH terminus of the ␥-subunit and the sixth or seventh WD repeat of G␤. The ␣-helical domain is thought to form the cytosolic pole of G␣ (1, 2, 4 -6). However, heterotrimeric G proteins are enriched in highly organized vesicular invaginations of the plasma membrane termed caveolae, which may represent sites of assembly of a signal transducing complex that could include receptors, G proteins, effectors, and even intracellular targets of the second messengers generated (28,29). It has been reported that G proteins bind to chief structural proteins of these organelles, the caveolins, via a sequence that lies between the switch-I and switch-II regions of the ␣-subunit (28,29). An incorporation of this latter site into structural models of the receptor/G protein complex changes the predicted orientation of the G protein in a way that brings the ␣-helical domain in close proximity to the membrane, which is consistent with the present data predicting a direct contact of activated ␤-adrenergic receptors with the distal end of helix A and the beginning of the following loop.
Overall, it thus appears that antibody K-20 defines an as yet unrecognized contact site for ␤-adrenergic receptors on G␣ s that localizes to the ␣-helical domain. This latter segment is the first example of a receptor binding region outside the core GTPase domain of G␣ and is essential for eliciting the responses transmitted by both G␣ s and G␤␥ s .
As yet, the most precisely defined site of receptor contact involves the extreme carboxyl terminus, which is thought to be of fundamental importance for receptor recognition and G protein activation (1)(2)(3)(4)(5)(6)(32)(33)(34)(35). Indeed, COOH-terminal anti-G␣ s and the peptide corresponding to its target sequence were as efficacious as peptide 100 -119 in blocking the isoproterenolinduced stimulation of adenylyl cyclase, indicating that the extreme carboxyl terminus is in fact necessary for ␤-adrenergic receptor signaling via G␣ s . However, in contrast to peptide 100 -119, COOH-terminal anti-G␣ s failed to prevent the isoproterenol-induced suppression of NADPH-dependent H 2 O 2 generation, a G␤␥ s -mediated response. Thus, a contact of activated ␤-adrenergic receptors with the extreme carboxyl terminus of G␣ s appears to be required solely for receptor signaling via G␣ s but is apparently not critical for receptor recognition and dissociation of the G s heterotrimer. The latter finding is difficult to reconcile with current models of G protein activation but is consistent with the observations of others, indicating that COOH-terminal G␣ s antibodies act as reliable and specific inhibitors of receptor signaling via G␣-subunits only. By contrast, their effects on G␤␥-mediated responses are variable and may depend on the G protein, receptors, or cell type studied. Thus, antibodies directed against the COOH termini of G␣ q and G␣ 13 have been shown to block signaling by m 3 muscarinic and AT 1A angiotensin II receptors via both types of component subunits of G q and G 13 , while anti-G␣ i1 , and anti-G␣ o , or anti-G␣ s had no effect on G␤␥-mediated responses elicited by sstr 3-somatostatin receptor stimulation or during transcytosis, respectively (23, 36 -38).
Together with the essential role of the epitope recognized by antibody K-20, the observation that COOH-terminal anti-G␣ s selectively impaired ␤-adrenergic receptor signaling via G␣ s without affecting the G␤␥ s -mediated response is consistent with the concept that ␤-adrenergic receptors interact with multiple sites on G␣ s , each playing a distinct role, as has also been proposed by others (39). More importantly, this latter finding implies that the activation and/or release of G␤␥ s by ligandoccupied ␤-adrenergic receptors may not require the simultaneous activation of G␣ s , which is in marked contrast to current concepts of G protein regulation. A subunit-specific activation of a G protein has in fact been described for the insulin-like growth factor II receptor, which may selectively turn on G␣ 2mediated pathways (40). However, this latter type of selectivity is achieved through sequestration of free ␤␥-subunits by the activated receptors, which is not comparable with the mechanism outlined above. Whether receptors exist that selectively activate G␣ without affecting the function of G␤␥ or vice versa remains to be established.
In conclusion, the current findings show that a commercially available polyclonal antiserum directed against residues 100 -119 within the ␣-helical domain of G␣ s (K-20) acts as a ␤-adrenergic receptor-mimetic agent, and defines a novel contact site for activated ␤-adrenergic receptors that localizes to the ␣-helical domain. In contrast to the COOH terminus of G␣ s , which seems to be critical for ␤-adrenergic receptor signaling via G␣ s only, the epitope recognized by antibody K-20 appears to be essential for eliciting the complete spectrum of ␤-adrenergic responses mediated by both G␣ s as well as G␤␥ s .  6. Lacking effect of peptide 100 -119 on forskolin-induced cAMP formation. Membranes were preincubated in the absence or presence of 2.5 M peptide 100 -119 and were stimulated by forskolin for 10 min and then assayed for adenylyl cyclase activity as described under "Methods." Values are means Ϯ S.D. of three separate experiments carried out in duplicate.