Identification of Amino Acid Residues of Gs a Critical to Repression of Adipogenesis*

Gs a regulates the differentiation of 3T3-L1 mouse embryonic fibroblasts to adipocytes, a process termed adipogenesis. Through the expression of chimera created by substituting regions of Gs a with corresponding regions of the G protein Gi a 2, the domain of Gs a involved in repression of adipogenesis was localized to sequence 146–235 of the molecule J. Chem. 271, As a prelude to alanine-scanning mutagenesis, chimeras in Gs a constructed from trisection of the sequence 125– 213 of Gi a 2 were expressed stably, and clones were evaluated for the ability of the chimera to repress adipogenesis in response to the inducers, dexamethasone and methylisobutylxanthine, in combination. The chimera containing sequence 150–177 of Gi a 2 repressed adipogenesis, whereas the chimeras with either sequence 125–149 or 178–213 of Gi a 2 failed to repress induction of adipogenesis. Alanine-scanning mutagenesis of these two critical domains was performed first in clusters and then confirmed by analysis of single mutations. Six residues unique to Gs a were identified as critical to repression of adipogenesis, Asn 167 , Cys 200 , Leu 203 , Ser 205 , Val 214 , and Lys 216 . Leu 203 and Ser 205 are required in tan-dem, as mutagenesis to alanine of either one alone was without effect on repressor activity. The remaining four residues are

nation (see Wang et al. (6), and Refs. therein). In addition to mediating signaling from a populous class of plasma membrane receptors to a less populous group of effectors that includes adenylylcyclases, phospholipase C␤, and various ion channels (7)(8)(9), heterotrimeric G proteins participate in more complex biological responses, including oncogenesis (9), early (10) and neonatal (11) mouse development, as well as cellular differentiation (6,12).
The present study focuses on the ability of the G protein Gs␣ to control adipogenic conversion of the 3T3-L1 fibroblasts to adipocytes. Gs␣ has been shown to play a key role in the differentiation of 3T3-L1 cells, as evidenced by the following observations. Gs␣ expression declines dramatically within 48 h of induction of differentiation; constitutive expression of Gs␣ in 3T3-L1 cells blocks induction of cell differentiation by known inducers; suppression of Gs␣ expression by antisense oligodeoxynucleotides (mimicking the inducer-driven decline in Gs␣) accelerates the cell differentiation from a 10-day to a 3-day process in the presence of inducers; and treatment with oligodeoxynucleotides antisense to Gs␣ alone provokes adipogenesis in the absence of the classical inducers (1,6,(13)(14)(15). Overexpression of the G protein that antagonizes many Gs␣ effects, Gi␣2, provokes adipogenesis in either the absence or the presence of the inducers (16).
The central question remains how Gs␣ controls cell differentiation. The ability of Gs␣ to repress adipogenesis is not thought to involve adenylylcyclase based upon the following observations. Elevation of intracellular cAMP concentrations by treating cells with either the diterpene forskolin or pertussis toxin does not affect the differentiation process, direct addition of dibutyryl cAMP itself to the cultures does not alter differentiation, and treatment of cells with 2Ј,5Ј-dideoxyadenosine to reduce intracellular cyclic AMP concentrations likewise does not alter differentiation. Treatment of cells with cholera toxin does block adipogenesis, through activation of Gs␣, much like expression of the constitutively active mutant form of Gs␣ (G225T). Although both cholera and pertussis toxins elevate intracellular cAMP, only cholera toxin blocks adipogenesis (6). Recently, we found that expression of the chimeric G protein in which the sequence 145-235 of Gs␣ is substituted for the corresponding region of Gi␣2 (Gi␣2 1-122/Gs␣ 145-235/Gi␣2 236 -394) inhibited cell differentiation as effectively as wild-type Gs␣ (1). These data indicate that the sequence harboring residues 146 -235 of Gs␣, which is not the region interacting with adenylate cyclase (17)(18)(19), is critical in controlling cell differentiation. The region 146 -235 of Gs␣ includes Switch I and Switch II (20), which are involved both in contact with ␤␥ complex, binding of guanine nucleotides, as well as the Gap region (20,21).
In the present study, we sought to define more precisely the domain and amino acid residues of Gs␣ that are responsible for the control of adipogenesis, the differentiation of 3T3-L1 embryonic fibroblasts to adipocytes. The repressor domain of Gs␣ was trisected into smaller sequences that were substituted with the corresponding domains of Gi␣2 and the chimeras stably expressed in 3T3-L1 cells. Sequences 147-171 and 200 -235, but not 172-199 of Gs␣ are critical in control of cell differentiation. Alanine scanning mutagenesis of sequences 147-171 and 200 -235 identified four amino acids (Asn 167 , Cys 200 , Val 214 , and Lys 216 ) and one cluster (Leu 203 and Ser 205 ) that are critical to the ability of Gs␣ to repress differentiation of 3T3-L1 cells.
Site-directed Mutagenesis-Mutations were introduced into pAl-terGs␣ by oligonucletide-directed in vitro mutagenesis using a kit purchased from Promega. All mutations and chimeras were verified by restriction enzyme digestion and DNA sequencing using Sequenase Version 2 Kit (U. S. Biochemical, Cleveland, OH). To subclone all the chimeras and mutants, pCW1Gs␣Q227L was partially digested with HindIII, filled in by Klenow fragment, and then ligated. Plasmids lacking the HindIII site at the 3Ј end of Gs␣ were selected. All chimeras and mutants were digested with HindIII and BglII and the isolated fragments of interest inserted into the HindIII-BglII site of the plasmid. Direct dideoxy sequencing was employed to verify the sequence of the chimeras and mutations. All mutants and chimera were constructed from wild-type versions of Gs␣ and Gi␣2 with normal intrinsic GTPase activity.
Stable Expression of Chimeras and Mutants in 3T3-L1 Cells-Mouse embryo fibroblast 3T3-L1 cells were obtained from the American Type Culture Collection (Rockville, MD). Cells were maintained in culture in 100-mm Petri dishes in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum. The protocols for stable transfection of 3T3-L1 cells employed in these studies were described previously (6,16). Stably transfected clones were selected (400 g/ml) and then maintained (100 g/ml) in the presence of the active form of the gentamicin The membranes were subjected to SDSpolyacrylamide gel electrophoresis and the resolved proteins transferred to nitrocellulose blots. The blots were stained with anti-Gs␣ antibody (CM129) raised against the peptide corresponding to C-terminal sequence Gs␣-(384 -394). Immune complexes were made visible by a second goat anti-rabbit IgG to which calf alkaline phosphatase was coupled. Panel B, for cyclic AMP accumulation, cells were incubated in the absence or presence of isoproterenol (10 M) for 15 min at 37°C in medium supplemented with Ro compound (0.1 mM). Data are the mean values Ϯ S.E., representing four independent experiments. Labeling: EV, empty vector; Gs, Gs␣; chimera sisEM; chimera sisMA; and chimera sisAB.
Immunoblotting-Aliquots of crude membrane fractions (50 g of protein/SDS-polyacrylamide gel electrophoresis/lane) from aliquots of each subclone were subjected to SDS-polyacrylamide gel electrophoresis. The separated proteins were transferred to nitrocellulose and the blots were stained with a rabbit polyclonal antibody specific for Gs␣ (CM129). The immune complexes were made visible by staining with a second antibody (goat anti-rabbit IgG) coupled to calf alkaline phosphatase (6).
Cyclic AMP Accumulation-Aliquots (0.5 ϫ 10 5 cells) of 3T3-L1 cells were washed and incubated in Kreb's phosphate buffer (pH 7.5, 10 6 cells/ml) containing the cyclic AMP phosphodiesterase inhibitor Ro 20-1724 (0.1 mM) in either the absence or the presence of the ␤-adrenergic agonist (Ϫ) isoproterenol (10 M) for 15 min at 37°C. The reaction was terminated by the addition of ethanol. Measurements of cyclic AMP accumulation were made in triplicate from separate aliquots of cells. Cyclic AMP accumulation was determined by the competitive protein binding assay, using the bovine adrenal cyclic AMPbinding protein (23).
Determination of Adipogenesis-Clones transfected with vector, wild type Gs␣, chimeras or mutants were maintained in 24-well plates for propagation. The differentiation protocol was described previously (1). Protocols for histochemical staining techniques are described in detail elsewhere (6,16). Adipogenesis was established via staining of accumulated lipid with oil red O.

RESULTS
The strategy adopted to the analysis of the sequence of Gs␣ responsible for the repression of adipogenesis was based upon our previous analysis in which the sequence of 145-235 of Gs␣ was identified as critical (1). Gs␣ sequence 145-235 embedded into the corresponding region of Gi␣2 retains the full capacity to block adipogenesis (1). The primary goal was to analyze by alanine-scanning mutagenesis the fine detail of the region of Gs␣ repressing adipogenesis. To make the task manageable, the region(s) of Gs␣ to be subjected to alanine-scanning mutagenesis had to be smaller than the parent sequence 145-235. To accomplish the task, the "repressor" region of Gs␣ was trisected and regions 125-149, 150 -177, and 178 -213 of Gi␣2 substituted individually for the corresponding region of Gs␣. The chimeras constructed in this fashion are displayed in Fig.  1. The inability of a chimera with the Gi␣2 region embedded in Gs␣ to repress adipogenesis would identify region(s) as candidate(s) for further analysis by alanine-scanning mutagenesis.
The cDNAs encoding each of the chimeras ( Fig. 1) were inserted individually in the pCW1 mammalian expression vector. The pCW1 vector is driven by the SV40 early promoter and harbors a selectable marker, the neomycin resistance gene (neo r ). Mouse embryonic fibroblasts were stably transfected and the derivative clones were selected in the presence of the gentamicin analogue G418. Each of the clones selected displays expression of Gs␣ well in excess of that observed in the cells stably transfected with the empty vector alone ( Fig. 2A). Immunoblots of cell membranes prepared from the stably transfected clones were stained with a rabbit polyclonal antibody raised against the decapeptide C terminus of Gs␣ (CM129). Expression of Gs␣ in clones harboring the expression vector for a chimera minimally was approximately 1-fold greater than the endogenous levels of Gs␣ expressed by 3T3-L1 clones stably transfected with the empty vector pCW1 alone. Constitutive expression of Gs␣ at a level 50% greater or more that endogenous Gs␣ is readily able to repress the ability of dexamethasone and MIX to induce adipogenesis (1). Chimeras of Gs␣ in which a smaller region of Gi␣2 (125-149, 150 -177, or 178 -213) FIG. 3. Expression of chimera sisEM and chimera sisAB lose the ability to repress induction of adipogenesis: analysis by oil red O staining. Stably transfected clones with empty vector (pCW1), as well as those transfectants stably expressing Gs␣ (␣s), sisEM, sisMA, and sisAB chimeras, as well as the D229S mutation of Gs␣ (␣sD229S), were plated on coverslips and propagated in 24-well culture plates. At confluence (day 0), one set of cells were treated with dexamethasone and methylisobutylxanthine (ϩD/M). Dexamethasone ϩ MIX were removed after incubation for 2 days, and the cells were maintained in DMEM containing 10% fetal bovine serum for 10 days. A second replicate set of cells was maintained in DMEM containing 10% fetal bovine serum in the absence (ϪD/M) of dexamethasone ϩ MIX. At day 10, cells were fixed by 3% paraformaldehyde for 5 min and stained with oil red O for 10 min. Hematoxylin (1%) was used to stained nuclei. Adipogenesis was examined under a Zeiss Axiophot microscope. The darkly stained bodies of the cytosol are oil droplets. Bar, 100 m. Data represent six different clones and at least two independent series of transfection experiments.
was substituted for the corresponding region in Gs␣ had little effect on the cAMP response of the clones to stimulation by isoproterenol, when compared with cells transfected with wildtype Gs␣ alone. Basal cAMP accumulation of these clones in comparison to that of the transfectants expressing wild-type Gs␣ alone was reduced slightly, perhaps reflecting the predominant Gi␣2 nature of the expressed protein (Fig. 2B). Earlier studies have demonstrated that Gs␣-Gi␣2 chimera with these types of substitutions retain their capacity to bind GTP and transduce signaling (17,24).
Adipogenesis is readily detected by staining the cultures for lipid accumulation with oil red O (Fig. 3). The cellular nuclei were made visible by counterstaining with hematoxylin. For clones stably transfected with the empty expression vector, no differentiation was observed in the absence of the inducers dexamethasone ϩ MIX (Fig. 3, panel A). When exposed to dexamethasone ϩ MIX, cultures harboring the empty expression vector pCW1 display robust differentiation (panel B). Marked lipid accumulation, the hallmark of adipocytes, was evidenced throughout the cultures, as shown earlier (1, 6). Clones constitutively expressing wild-type Gs␣, in contrast, do not respond to induction by dexamethasone ϩ MIX, failing to differentiate into adipocytes (panel D). Clones expressing either chimera sisEM (panel F) or chimera sisAB (panel J) displayed phenotypes identical to clones transfected with empty vector, fully differentiating in response to dexamethasone ϩ MIX and replete with accumulated lipid stained by the oil red O. Chimera sisMA, in sharp contrast, continues to repress dexamethasone ϩ MIX-induced adipogenesis (panel H) much like expression of wild-type Gs␣ (panel D). These data demonstrate that regions 147-171 (sisEM) and 200 -235 (sisAB) of Gs␣ are critical in the control of adipogenesis, when substituted with the corresponding region of Gi␣2 the chimera lose the ability to repress adipogenesis. Region 172-199 of Gs␣, to the contrary, appears to be dispensable and can be replaced by the corresponding region of Gi␣2 without altering the ability of chimera to repress adipogenesis.
Having identified two smaller regions of Gs␣ critical for expression of its ability to repress adipogenesis, we developed a strategy to perform alanine-scanning mutagenesis within the sisEM domain and analyze the influence of these mutations on the repressor function of the expressed molecule. Initial mu- FIG. 4. Alignment of sequences of Gs␣ and Gi␣2 with regions of Gs␣ subjected to alanine-scanning mutagenesis to identify amino acid residues of Gs␣ critical to the repression of adipogenesis. Protein sequence is provided using the three-letter symbol and the number of each residues commencing from the N terminus is provided above the sequence for Gi␣2 and below for the sequences for Gs␣. Residues in boxes are conserved between Gs␣ and Gi␣2 and were not subjected to mutation analysis, since Gs␣ represses adipogenesis whereas Gi␣2 does not. tagenesis created both single (Gs␣M3, Gs␣M5, and Gs␣M6) and multiple (Gs␣M1, Gs␣M2, Gs␣M4, and Gs␣M7-M10) alanine substitutions in protein sequences 147-171 and 200 -235 of Gs␣ (Fig. 4, A and B), presuming that alanine substitution of a critical residue would abolish the ability of the mutant Gs␣ to repress adipogenesis. Clones stably transfected with pCW1 harboring the cDNA of Gs␣ with one or more mutations displayed increased immunoreactive Gs␣, reflecting the expression of the mutant forms in excess of the endogenous level of Gs␣. Clones were selected that expressed the mutant Gs␣ molecules at levels approximately 1-fold greater than the staining observed in clones harboring empty vector alone (Fig. 5A). All of the clones expressing Gs␣ mutants display elevated levels of basal cAMP accumulation compared with clones harboring empty expression vector alone and each displayed isoproterenol-stimulated cAMP accumulation (Fig. 5B).
In the context of region 147-171 (sisEM domain), alaninescanning mutations of Gs␣M1 through Gs␣M4 displayed no effect on the ability of the mutant form of Gs␣ to repress adipogenesis in response to the inducers dexamethasone ϩ MIX (Fig. 6). Clones constitutively expressing Gs␣M1, 2, 3, and 4 (panels C, D, E, and F, respectively) were refractory to induction of adipogenesis. In contrast, expression of Gs␣M5, a single alanine substitution for asparagine at position 167 abolishes the ability of the mutant form of Gs␣ to repress adipogenesis (panel G). Clones expressing N167A Gs␣ no longer repressed adipogenesis, displaying robust lipid accumulation in response to inducers (panel G), much like the wild-type cultures of  6. Expression of Gs␣M1, Gs␣M2, Gs␣M3, Gs␣M4 and Gs␣M8, but not Gs␣M5-7, Gs␣M9, nor Gs␣M10, blocks induction of adipogenesis: analysis by oil red O staining. Stably transfected clones with empty vector (pCW1), Gs␣ (␣s), and 10 Gs␣ mutants were plated on coverslips and propagated in 24-well culture plates, respectively. At confluence (day 0), one set of cells were treated with dexamethasone and methylisobutylxanthine (ϩD/M). Dexamethasone ϩ MIX were removed after incubation for 2 days, and the cells were maintained in DMEM containing 10% fetal bovine serum for 10 days. A second replicate set of cells was maintained in DMEM containing 10% fetal bovine serum in the absence of dexamethasone ϩ MIX (not differentiated and not shown). At day 10, cells were fixed by 3% paraformaldehyde for 5 min and stained with oil red O for 10 min. Expression of Gs␣M1, Gs␣M2, Gs␣M3, Gs␣M4, and Gs␣M8 but not Gs␣M5, Gs␣M6, Gs␣M7, Gs␣M9 nor Gs␣M10, mutants blocks induction of adipogenesis. The darkly stained bodies of the cytosol are oil droplets. Bar, 100 m. Data represent six different clones and at least two independent series of transfection experiments.

3T3-L1 or the clones stably transfected with empty expression vector alone (panel A).
Analysis of the sisAB region (Gs␣ residues 200 -235) initially focused upon aspartic acid residue 229 which was the only residue within the region of 221-235 of Gs␣ that varied from that of Gi␣2 (Fig. 4B). The D229S mutant form of Gs␣ retains its ability to repress adipogenesis in clones challenged with dexamethasone ϩ MIX (Fig. 3, panel L). With the C-terminal region of the AB region shown to be unimportant in the repressor activity of Gs␣, alanine-scanning mutagenesis was focused upon the N-terminal sequence from residue 200 -220 of this region (Fig. 4B). Five mutations, one a single alanine substitution (Gs␣M6) and the others multiple substitutions (Gs␣M7-M10), were constructed and shown to be expressed in stably transfected clones at Ͼ1-fold over endogenous Gs␣ (Fig. 5A). These clones displayed increased basal cAMP accumulation over that observed for clones expressing the endogenous Gs␣ (Fig. 5B) and were examined further for their ability to respond to induction with dexamethasone ϩ MIX.
Expression of Gs␣M6 (C200A) abolished the ability of Gs␣ to repress adipogenesis in the clones (Fig. 6, panel H). Clones expressing C200A Gs␣ now stain prominently for accumulated lipid following treatment with dexamethasone ϩ MIX. Expression of Gs␣M7 (L203A and S205A), likewise, abolished the ability of Gs␣ to repress adipogenesis in response to dexa- methasone ϩ MIX, resulting in robust lipid accumulation in the clones (Fig. 6, panel I). In contrast, alanine substitution for Phe 208 and Lys 211 of Gs␣ in the clones expressing Gs␣M8 had no discernible effect on the ability of Gs␣ to repress adipogenesis, as demonstrated by the absence of oil red O staining of lipid in these clones (Fig. 6, panel J). The triple mutation of Gln 213 , Val 214 , and Asp 215 to alanine resulted in a loss of repressor activity for the mutant Gs␣M9 (Fig. 6, panel K). In the presence of the inducers dexamethasone ϩ MIX, clones expressing the Gs␣M9 mutant form of Gs␣ fully differentiated into adipocytes, staining prominently by oil red O. Substitution of Lys 216 , Val 217 , Asn 218 , and His 220 with alanine also abolished the ability of the mutant Gs␣ to repress adipogenesis in response to dexamethasone ϩ MIX (Fig. 6, panel L). Thus, the mutations found in Gs␣M6, 7, 9, and 10 render the Gs␣ unable to repress adipogenesis as does the wild-type Gs␣ (Fig. 6,  panel B).
The alanine-scanning mutagenesis of single residues had revealed that N167A Gs␣ and C200A Gs␣ mutants were devoid of repressor activity. Mutagenesis of multiple residues as a single cassette identified regions critical for repressor activity, whereas establishing the precise residue(s) necessary for repressor activity of Gs␣ would require finer detailed analysis. Single point alanine substitutions were created for each of nine residues implicated as critical to the repressor activity of Gs␣ (Fig. 4C). Mutant forms of Gs␣ were stably transfected at levels approximately 1-fold greater than that of endogenous Gs␣ (Fig.   7). Although Gs␣M7 with L203A and S205A double mutation has lost the ability of Gs␣ to repress adipogenesis in response to dexamethasone ϩ MIX (Fig. 6, panel I, and Fig. 8, row A,  panel c), single point mutations of L203A (Fig. 8, row A, panel  d) or S205A (Fig. 8, row A, panel e) were without effect on the repressor activity. These data argue persuasively that Leu 203 and Ser 205 together play a critical role in the control of adipogenesis exerted by Gs␣.
The expression of Gs␣M9, a mutant form of Gs␣ with alanine substitution of Gln 213 , Val 214 , and Asp 215 , resulted in the loss of repressor activity (Fig. 6, panel K, and Fig. 8, row B, panels  b-d). The single point mutation of G213A had no discernable effect on the ability of the mutant Gs␣ to repress adipogenesis. Similarly, mutation of D215A resulted in a mutant form of Gs␣ that fully repressed the ability of dexamethasone ϩ MIX to induce differentiation, just as observed with constitutive expression of wild-type Gs␣. In sharp contrast to the G213A and D215A mutations, alanine substitution of Val 214 abolished the ability of Gs␣ to repress adipogenesis of the clones in response to dexamethasone ϩ MIX. These data reveal the basis of the Gs␣M9 cassette of mutations to abolish the repressor activity of Gs␣ resides solely in the V214A mutation.
The Gs␣M10 cassette of mutations includes K216A, V217A, N218A, and H220A (Fig. 4B). This cluster of alanine substitutions was analyzed further by creation of stably transfected clones of 3T3-L1 cells that expression of each of the individual mutations of the sisAB region of Gs␣ (Fig. 4, panel C). Although expression of the Gs␣M10 cassette eliminates the ability of Gs␣ to repress adipogenesis in the clones, single alanine substitutions of V217A, N218A, and H220A were without effect, i.e. constitutive expression of each of the single mutant forms of Gs␣ repressed the ability of dexamethasone ϩ MIX to induce adipogenesis equally well (Fig. 8, row C, panels b-e). Only one of the mutations, K216A, was found to mimic the effects of the Gs␣M10 cassette on loss of repressor activity by Gs␣.
Single mutations of N167A, C200A, V214A, and K216A as well as the double mutation L203A,S205A abolished the ability of expressed Gs␣ to repress adipogenesis in 3T3-L1 cells. Mutant forms of Gs␣ were expressed at levels approximately 1-fold greater than that of endogenous Gs␣ (Fig. 9A) and retained the ability to elevate both basal cyclic AMP accumulation and mediate stimulation of cyclic AMP accumulation in response to isoproterenol (Fig. 9B).

DISCUSSION
Comparisons of the sequences of the domains in Gs␣ (Tyr 147 -Leu 171 and Cys 200 -Ile 235 ) that repress adipogenesis in 3T3-L1 cells with the corresponding sequences of other G protein ␣ subunits reveals significant levels of non-identity. Many residues within these two protein sequences, however, such as Ile 180 , Glu 182 , Arg 193 , Gly 199 , Gln 200 , Arg 201 , Glu 203 , Lys 206 , and Trp 207 involved in binding of ␤␥ complex by Gt␣ are conserved in Gs␣ also (21). The protein sequence Met 221 -Ile 235 of Gs␣ is highly conserved in virtually all of the heterotrimeric G protein ␣ subunits (21). In an effort to define more precisely the residues within the putative repressor domains that are required for the control of adipogenesis, the amino acid sequences of Gs␣ and Gi␣2 were compared (22), as displayed in Fig. 4  (panels A-C). The displays identify 35 identical residues within these domains. Of the 35 residues noted, 28 are conserved in all heterotrimeric G protein ␣ subunits (21). Of the 26 nonidentical residues, 21 were mutated to alanine in clusters targeting the regions of interest in chimeras sisEM and sisAB (Fig. 4). The remaining five amino acids (Ala 150 , Lys 151 , Ala 152 , Asp 229 , and Arg 232 ) include naturally occurring alanine at positions 150 and 152, and therefore were not studied further. Asp 229 is the only non-conserved residue of Gs␣ found in the sequence from 221 to 235, and D229S mutant form of Gs␣ is shown in the present work to retain the ability to repress adipogenesis (Fig.  3). Mutagenesis of either Lys 151 or Arg 232 was not undertaken, simply because other mammalian heterotrimeric G protein ␣ subunits have the same amino acids at these positions (21), and the lysine for arginine substitution is considered a conservative substitution.
The focus of the analysis was to evaluate the structural basis for repression of adipogenesis by Gs␣, exploiting our knowledge of the repressor domain by alanine-scanning mutagenesis. Repressor activity of protein sequence 146 -235 of Gs␣ was identified (6) through the construction and then expression of various chimeras between Gs␣ (repressor) and corresponding regions of Gi␣2 (activator). The ability of these Gs␣/Gi␣2 chimeras to block adipogenesis in 3T3-L1 clones challenged with well-known inducers of differentiation was evaluated (1). Since Gi␣2 and Gi␣1 display more than 95% identify in their primary sequence (22), the crystal structure of Gi␣1-GTP␥S was adapted for a first-approximation description of the regions of Gi␣2 analyzed in the present work (Fig. 10). The structure displayed is that of Gi␣1 in which domains EM (147-171), MA (172-199), and AB (200 -235) are projected as regions I, II, and III, respectively. The major region implicated in control of adenylylcyclase (AC) is displayed in yellow (9), while that for regions I (EM), II (MA), and III (AB) are rendered cyan, green, and blue, respectively (Fig. 10, panels A and B). Maintaining the same landmarks of conserved regions, the projection of Gs␣ residues upon the corresponding structure of Gi␣2 would place Leu 203 and Ser 205 at positions of Gi␣1 residues Lys 180 and Thr 182 , respectively (Fig. 10, panel C). In Gi␣1, Lys 180 and Thr 182 appear as exposed residues, available for protein-protein contact. It is of interest that Lys 180 and Thr 182 of Gi␣1 have been shown to participate in the binding of the GTPase activator for Gi␣1, RGS4 (24). Both Leu 203 and Ser 205 are essential residues to the repressor activity of Gs␣, substitution of both to alanine abolishes the ability of the mutant Gs␣ to block adipogenesis. The results predict that this domain (Fig.  10, panel D) is an important contact site for Gs␣ with the effector controlling differentiation in these cells or that these residues, when altered via mutagenesis, interrupt some extended conformation of Gs␣ critical to the repressor activity of the molecule.
The Cys 200 residue is a unique feature of Gs␣, not shared by any other heterotrimeric G protein ␣ subunit (Fig. 4). Cys 200 is a critical residue for Gs␣ function with respect to repression of adipogenesis. Alanine substitution of Cys 200 effectively abolished the ability of Gs␣ to exert its repressor activity on adipogenesis. Projection of this information on the structure of Gi␣1 identifies Thr 177 (Fig. 10, panels C and D), a residue located in the C-terminal end of Switch I region near the junction with linker 2 (19,20). Cys 200 of Gs␣ is embedded in a sequence highly conserved among G protein ␣ subunits Leu Arg Cys Arg Val Xaa Thr, including Gi␣1, Gi␣2, and Go␣. The projections suggest that Leu 203 and Ser 205 (both in Switch I region) as well as Cys 200 are in close proximity and likely exposed residues important in Gs␣ repressor function (Fig. 10,  panel D).
Three of the remaining residues critical to repressor activity, Asn 167 , Val 214 , and Lys 216 , appear to be exposed when projected upon the corresponding area of the Gi␣1 structure (Fig.  10, panels C and D). Asn 167 , located at the flex region between ␣ helices D and E (19,20), is embedded in the sequence Arg Ser Asn Glu Tyr Gln Leu that is highly conserved among G protein ␣ subunits including Gs␣, Gi␣1, Gi␣2, and Go␣. Val 214 and Lys 216 , located in the C-terminal reach of Switch I region (19,20), are embedded in a unique region of six residues in Gs␣ The mutagenesis data are project onto the structure of Gs␣. To facilitate the discussion, the domain of Gs␣ implicated in the control of adenylylcyclase (AC) is displayed in yellow, while the GTP organizing elements Switch 1 (Sw1), Switch 2 (Sw2), and Switch 3 (Sw3) are rendered in cyan, blue, and green, respectively (panel A). The GTP molecule has been eliminated from the image to allow greater clarity of the switch regions. Please see "Discussion" for further details. flanked both by 10 residues N-terminal as well as 20 residues C-terminal that display a high degree of conservation among several G protein ␣ subunits, including Gi␣1, Gi␣2, and Go␣. The protein sequence 213-218 of Gs␣ would appear to play some unique role(s) in its function, including a critical role in repressor activity, since alanine substitution of either Val 214 or Lys 216 abolishes the ability of Gs␣ to block adipogenesis.
The recent elucidation of the crystal structure of Gs␣ at 2.5 A in a complex with GTP␥S (25) affords the opportunity to relate the mutagenesis data to the structure of Gs␣ (Fig. 11). To facilitate the discussion, the domain of Gs␣ implicated in the control of adenylylcyclase (AC) is displayed in yellow, while the GTP organizing elements Switch 1 (Sw1), Switch 2 (Sw2), and Switch 3 (Sw3) are rendered in cyan, blue, and green, respectively (Fig. 11, panel A). The Leu 203 , Ser 205 cluster, essential for represser activity of Gs␣, is displayed in white (Fig. 11, panel A), as are Asn 167 , Val 214 , and Lys 216 residues (Fig. 11,  panel B). Cys 200 is far less visible in the Gs␣-GTP␥S structure than in the Gi␣1-GTP␥S structure (Fig. 11, panel A). A stick model of Gs␣, in which the GTP molecule has been purposely deleted from the structure, illuminates all of the residues essential for the repressor activity of Gs␣ (Fig. 11, panel C). Inspection of the ribbon and coil diagram of Gs␣ highlights the critical placement of the Leu 203 , Ser 205 cluster in Switch 1 (Sw1), the proximity of Asn 167 to Switch 3 (Sw3), and the exposure of Val 214 and Lys 216 (Fig. 11, panel D).
The current study adds to our expanding understanding of the structure-activity relationships in the Gs␣ molecule. Constitutive expression of Gs␣ blocks adipogenesis and induction of adipogenesis occurs through a rapid loss of Gs␣ in 3T3-L1 adipocytes. Using construction of chimeric ␣ subunits as well as alanine-scanning mutagenesis we provide an insight into the regions of Gs␣ that are required for repressor activity. The regions do not overlap with those implicated in the control of adenylylcyclase (Figs. 10 and 11). In concert, alterations in intracellular cyclic AMP accumulation fail to influence either the induction or the course of adipogenesis (6). The nature of the effector through which Gs␣ repressed adipogenesis remains unsolved. The structural information garnered from the cur-rent analysis will assist in efforts aimed at identifying this additional and novel effector for Gs␣.