Gsα Repression of Adipogenesis via Syk*

Gsα regulates the differentiation of 3T3-L1 mouse embryonic fibroblasts to adipocytes, a process termed adipogenesis. Inducers of adipogenesis lead to a loss of Gsα and derepress differentiation to adipocytes. The broad spectrum tyrosine kinase inhibitor genistein is shown to block induction of adipogenesis, suggesting an early role of tyrosine phosphorylation in adipogenesis. Staining of phosphotyrosine identified prominent staining of a ∼70-kDa protein, hypothesized to be the tyrosine kinase Syk. Reverse transcription and polymerase chain reaction amplification established the expression of Syk mRNA in these embryonic fibroblasts. Immunoprecipitations with Syk-specific antibodies demonstrated the presence of Syk in fibroblasts and a rapid increase in the amount of phospho-Syk, peaking at 24 h post induction. Clones constitutively expressing Gsα, which can no longer be induced to differentiate, no longer display increased phospho-Syk levels in response to inducers. The linkage between Gsα and Syk was probed by immunoprecipitations revealing association of Syk with Gsα in the absence of induction. Upon induction of adipogenesis, Gsα levels decline and phospho-Syk levels as well as Syk kinase activity increase. Expression of wild-type Syk both potentiates the ability of inducers to act as well as induces adipogenesis itself. Expression of the kinase-deficient Syk had no such effects on adipogenesis. These data provide a new insight into the control of adipogenesis, suggesting that Gsα represses adipogenesis via Syk. Treatment with the inducers promotes a decline in Gsα, increases in levels of phospho-Syk, and adipogenesis.

Green and co-workers isolated a clone of Swiss mouse 3T3-L1 cells that provides a unique model for insulin-sensitive primary fat cells (1)(2)(3)(4). The 3T3-L1 cells differentiate over 7-10 days from an embryonic fibroblast-like state to an adipocyte phenotype when treated with inducers, such as insulin (5) or dexamethasone and methylisobutylxanthine in combination (see Ref. 6 and references therein). Heterotrimeric G-proteins participate in cell signaling via effectors that include adenylylcyclases, phospholipase C␤, and various ion channels (7)(8)(9) as well as in more complex biological responses, including oncogenesis (9), early (10) and neonatal (11) mouse development, as well as differentiation (6,12). G s ␣ plays a key role in regulating differentiation of 3T3-L1 cells, as evidenced by the following data: G s ␣ expression declines dramatically within 24 h of induction of differentiation; constitutive elevation of G s ␣ in 3T3-L1 cells blocks induction of cell differentiation by known inducers; reduction of G s ␣ levels by antisense oligodeoxynucleotides both mimics the inducerdriven decline in G s ␣ and accelerated the cell differentiation from a 10-day process to a 3-day event in the presence of inducers, and oligodeoxynucleotides antisense to G s ␣ alone provoke adipogenesis in the absence of the classical inducers (1,6,(13)(14)(15). Overexpression of the G i ␣2, the G-protein that antagonizes many G s ␣ effects, stimulates adipogenesis in either the absence or the presence of the inducers (16).
How does G s ␣ control differentiation of 3T3-L1 cells? Changes in adenylylcyclase activity and/or cyclic AMP levels do not appear to play a role in the ability of G s ␣ to repress adipogenesis based upon the following observations: direct addition of dibutyryl cAMP itself to the cultures does not alter differentiation; elevation of intracellular cAMP concentrations by either the diterpene forskolin or pertussis toxin does not affect the differentiation process; and challenging cells with 2Ј,5Јdideoxyadenosine to reduce intracellular cyclic AMP concentrations likewise does not alter differentiation. Cholera toxin does block adipogenesis through activation of G s ␣, much like expression of the constitutively active mutant form of G s ␣ (G225T), yet both cholera and pertussis toxins elevate intracellular cAMP (6). Expression of the chimeric G-protein in which the sequence 145-235 of G s ␣ is substituted within G i ␣2 at the corresponding region of this homologous protein (G i ␣2 1-122/ G s ␣ 145-235/G i ␣2 215-355) blocks cell differentiation as effectively as wild-type G s ␣ (1). The domain of G s ␣ including residues 146 -235, which displays several contact regions with adenylate cyclase (17)(18)(19), is critical in controlling cell differentiation. This region of G s ␣ includes Switch I and Switch II (20), which participate in contact with ␤␥ complex, binding of guanine nucleotides, and adenylylcyclases (17)(18)(19)(20)(21)(22)(23). The repressor domain of G s ␣ has been subjected to further analysis by first being trisected into smaller sequences that were substituted with the corresponding domains of G i ␣2 and the chimeras stably expressed in 3T3-L1 cells (24). Sequences 147-171 and 200 -235, but not 172-199, of G s ␣ 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 G s ␣ to repress differentiation of 3T3-L1 cells (24).
In the present study we show the ability of a tyrosine kinase inhibitor to block adipogenesis in response to dexamethasone and methylisobutylxanthine and provide compelling evidence for a central role of the tyrosine kinase Syk in the G s ␣-mediated regulation of adipogenesis. The data suggest that Syk associates with G s ␣ and that inducers of adipogenesis promote Syk phosphorylation/activation. Blockade of the decline in G s ␣ or of tyrosine kinase activity with genistein blocks the adipogenic response to inducers, whereas overexpression of Syk itself promotes adipogenesis.

EXPERIMENTAL PROCEDURES
Stable Expression of Syk, Syk mutants, and G s ␣ Chimeras in 3T3-L1 Cells-Mouse embryo fibroblast 3T3-L1 cells were obtained from the American Type Culture Collection (Manassas, VA). Plasmids pRc/Syk and pRc/Syk(KϪ), harboring the cDNA for the kinase-dead version of Syk, were kindly provided by Dr. Anthony L. DeFranco (Department of Microbiology and Immunology, University of California, San Francisco, CA). The expression vector for G s ␣ (pCW1 G s ␣) and its empty counterpart (pCW1) were kindly provided from Dr. Gary L. Johnson (Basic Sciences, National Jewish Center for Immunology, Denver, CO). Cells were maintained in culture in 100-mm Petri dishes in Dulbecco's modified Eagle's medium 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 analogue, G418 sulfate (Life Technologies, Inc.).
Immunoprecipitations and Immunoblotting-Whole cell lysates were prepared in a buffer composed of 130 mM NaCl, 5 mM EDTA, 10 mM NaF, 20 mM sodium pyrophosphate, 10 mM sodium molybdate, 2 mM Na 2 V 3 0 4 , 6 mM dithiothreitol, 1.0% Triton X-100, 0.5% Nonidet P-40, 20 mM Tris-HCl, pH 7.5, 5 mg/ml aprotinin, 5 mg/ml leupeptin, and 0.2 mM freshly prepared phenylmethylsulfonyl fluoride. Washed cells were lysed directly in the Petri dishes, and the mixture was collected by aspiration and subjected to centrifugation, with the resultant supernatant referred to as the whole cell lysate. Aliquots of whole cell lysates (1-3 mg of protein) from each subclone were subjected immunoprecipitation with antibodies to Syk, to PY20, or to G s ␣. The antibodies were coupled to protein A/G-agarose (Santa Cruz Biotechnology). To create the proper control protein A/G-agarose preparations, normal mouse/ rabbit serum was coupled to the matrix, under the same conditions following the protocol of the commercial supplier. Immunoprecipitates and aliquots of whole cell lysates were subjected to SDS-polyacrylamide gel electrophoresis (PAGE). 1 The separated proteins were transferred to nitrocellulose, and the blots were stained with polyclonal antibodies (Syk(LR), Santa Cruz Biotechnology, Santa Cruz, CA; BR13, DNAX, Palo Alto, CA) and a mouse monoclonal antibody (Anti-Syk, clone 4D10.1, catalog no. 05-439, Upstate Biotechnology, Inc., Lake Placid, NY) specific for Syk, a murine monoclonal antibody specific for phosphotyrosine (PY20 from ICN Pharmaceuticals, Inc., Costa Mesa, CA) or G s ␣ (CM129, see Ref. 6). The immune complexes were made visible by staining with a second antibody (goat anti-rabbit or goat anti-mouse IgG) coupled either to calf alkaline phosphatase or to horseradish peroxidase (6).
Determination of Adipogenesis-Clones transfected with vector, wild-type G s ␣, wild-type Syk, or the kinase-deficient Syk mutant 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.
Reverse Transcription-Polymerase Chain Reaction Amplification of Syk-Mouse 3T3-L1 wild-type cells were grown in 100-mm culture dishes, lysed directly on the plates with 1 ml/plate of RNA Stat-60 (TEL-TEST, Inc., Friendswood, TX) and repeated reaspiration of the lysate through a small bore pipette. The extraction of the total cellular RNA was performed according to the directions provided by the commercial supplier of the RNA Stat-60. The RNA was washed and employed as a template for reverse transcription, according to the manufacturer's protocol (Promega, Madison, WI). Polymerase chain reaction amplification was performed using primers ATGGCGGGAAGTGCTGT-GGACAG and TGAGAGTGGTAATGGCAGAGGT specific for mouse Syk (GenBank TM ). The amplification was performed in a GeneAmp PCR 2400 apparatus (Perkin-Elmer Applied Biosystems, Foster City, CA). The amplification products were subjected to electrophoresis on 1.2% agarose gels and made visible in ethydium bromide by UV irradiation.
Syk Kinase Assay-Whole cell lysates were prepared from the 3T3-L1 clones of interest and immune precipitations performed with polyclonal anti-Syk antibodies (BR13 from DNAX). Syk kinase assay was performed with 5 g of the purified glutathione S-transferase and the cytoplasmic domain of human band 3 fusion protein (GST-CDB3) as substrate. The GST-CDB3 plasmid was a generous gift of Dr. Xin-Yun Huang (Cornell University Medical College, New York, NY). The kinase assay was conducted in 50 mM HEPES, pH 7.4, 10 mM MnCl 2 , 1 mM phenylmethylsulfonyl fluoride, 10 M ATP, 10 Ci of [␥-32 P]ATP for 30 min at 30°C. The samples were then subjected to SDS-PAGE on 10% acrylamide gels. The activity was quantified in the dried gel by use of a PhosphorImager.

RESULTS AND DISCUSSION
Analysis of the role of protein kinase action in the regulation of adipogenesis revealed a key transient role for calcium, calmodulin-dependent protein kinase II but little role for either cyclic AMP-dependent protein kinase (protein kinase A) or protein kinase C (25). To probe a possible role for tyrosine kinase(s) in adipogenic conversion of 3T3-L1 cells, the broad spectrum, isoflavanoid tyrosine kinase inhibitor genistein was employed (Fig. 1). Treating the cells with genistein (100 M) blocks the ability of dexamethasone and methylisobutylxanthine (D/M) to induce adipogenesis in 3T3-L1 cells (Fig. 1A). Staining of fat droplets with oil red O makes visible the accumulation of lipid, the hallmark of adipogenic conversion. Genistein blocks adipogenesis, whereas the dimethyl sulfoxide vehicle itself (0.5%) has no effects, i.e. inducers promote robust lipid accumulation.
The dose dependence with respect to the effects of genistein on D/M-induced adipogenesis was explored (Fig. 1B). At 100 M, genistein effectively abolishes adipogenic conversion. Both 10 and 33 M concentrations of the tyrosine kinase inhibitor 1 The abbreviations used are: PAGE, polyacrylamide gel electrophoresis; GST, glutathione S-transferase; D/M, dexamethasone and methylisobutylxanthine; RT, reverse transcription; PCR, polymerase chain reaction. days, and the cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum for 10 days. A second replicate set of cells was maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum in the presence of D/M but exposed to 100 M genistein (panel c), a broad spectrum tyrosine kinase inhibitor for 2 h prior to and during the addition of the inducers. The genistein was dissolved in dimethyl sulfoxide (0.5%), and this vehicle was added to all of the cultures lacking the genistein, as a control (panel d). The presence of the dimethyl sulfoxide has no effect on the growth of the culture or their ability to undergo adipogenic conversion in response to D/M. At day 10, cells were fixed by 3% paraformaldehyde for 5 min and stained for lipid accumulation with oil red O for 10 min. Hematoxylin (1%) was used to stained nuclei. B, a dose-response of adipogenic conversion in the 3T3-L1 cells to genistein (3.3-100 M) was explored. Adipogenesis was examined under a Zeiss Axiophot microscope. The darkly stained bodies of the cytosol are oil droplets. Data presented are representative of three experiments performed on separate occasions with essentially identical data. provoke a marked inhibition of adipogenesis. At 3.3 M, the lowest concentration tested, genistein displays only a weak inhibitory influence on the differentiation of 3T3-L1 cells in response to the inducers D/M.
To probe further tyrosine kinase action in adipogenesis, whole cell lysates of 3T3-L1 cells treated for varying times with and without the inducers D/M were prepared and subjected to immunoprecipitation with either PY20 antibody coupled to protein A/G-agarose or the protein A/G-agarose coupled to normal rabbit serum as a control. The immunoprecipitates were subjected to SDS-PAGE, transferred to nitrocellulose blots, and stained for phosphotyrosine content using the PY20 antibody ( Fig. 2). Several phosphotyrosine-containing proteins with molecular masses greater than 50 kDa (120, 84, and 70 kDa) were made visible by the staining. The phosphotyrosine content of the ϳ70-kDa band was observed to increase within 24 h and thereafter declined to the base-line value. The p120 and p84 species displayed a temporal pattern of tyrosyl phosphorylation differing from that of p70.
Based upon its electrophoretic mobility and phosphotyrosine content, the p70 species was suspected to be Syk, or a homolog of Syk. The Syk/ZAP-70 nonreceptor protein-tyrosine kinases are major elements in signal transduction of T-and B-cells, regulating cellular processes that include antibody production, lymphokine production, proliferation, differentiation, and apoptosis (26 -29). A role for a Syk-like tyrosine kinase has not been proposed in other cell lines, so it was important to deter-mine whether Syk was expressed in the 3T3-L1 cells and whether its phosphorylation/activation was altered by adipogenesis. Reverse transcription and polymerase chain reaction (RT-PCR) amplification with Syk-specific primers reveal the presence of Syk mRNA in the 3T3-L1 cells (Fig. 3A) with the FIG. 2. Tyrosine phosphorylation of cellular proteins of mouse 3T3-L1 clones during early adipogenesis: analysis by immunoblotting (IB). Crude cell lysates were prepared from wild-type 3T3-L1 clones challenged with dexamethasone and methylisobutylxanthine to induce adipogenesis over 48 h post-confluence. The samples (3 mg of cellular protein of each whole cell extract) were subjected to immunoprecipitation reactions employing antibodies to phosphotyrosine (PY20) that were first chemically coupled to protein A/G-agarose and the immunoprecipitates (IP) subjected to SDS-polyacrylamide gel electrophoresis. Protein A/G-agarose prepared in the absence of the PY20 antibodies was coupled to normal rabbit serum to provide a control for the immune precipitations. The resolved proteins from the immunocomplexes were transferred to nitrocellulose blots. Then the blots were stained with the PY20 antibody. Immune complexes were made visible by chemiluminescence and the use of a second goat anti-mouse IgG to which horseradish peroxidase was coupled. The phosphotyrosine content of three prominent proteins (p120, p84, and p70) increased within 48 h of treatment with the inducers D/M. The data shown are representative of more than three independent experiments performed on as many separate cultures of 3T3-L1 cells. pRc/Syk plasmid positive control (P); lanes 3-6, amplification from RT products from cells treated with D/M for 0, 12, 24, and 48 h, respectively. PCR primers specific for glyceraldehyde-phosphate dehydrogenase (GAPDH) were employed as internal controls for loading equivalence. B, for immunoblotting, crude cell lysates (50 g of cellular protein/lane) were prepared from wild-type 3T3-L1 clones challenged with dexamethasone and methylisobutylxanthine to induce adipogenesis over 48 h post-confluence. The samples of cellular protein were subjected to SDS-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose blots. The blots were stained with a Syk-specific antibody (BR13). The p70 phosphotyrosine-containing phosphoprotein was established as authentic Syk by staining with three different polyclonal antibodies and one monoclonal antibody, obtained from independent sources. Immune complexes were made visible by chemiluminescence and the use of a second goat anti-rabbit IgG to which horseradish peroxidase was coupled. Note that the expression of Syk is not altered during adipogenesis. C, the blots were reprobed with antibodies to the ␤2-subunit of heterotrimeric G-proteins, which confirmed uniform sample loading. Three individual antibodies specific for Syk provided the same data. The data shown are representative of RT-PCR and immunoblotting from three separate cultures of 3T3-L1 cells.
same mobility as the signal obtained from pRc/Syk plasmid positive control (P). RT-PCR amplification with primers specific for glyceraldehyde-phosphate dehydrogenase establishes equivalent sample loadings. PCR performed using 3T3-L1 cell RNA as template was devoid of signal, demonstrating the absence of contamination of the RNA samples with DNA.
Immunoblots of proteins from whole cell lysates of 3T3-L1 cultures subjected first to SDS-PAGE and then stained with Syk-specific antibodies make visible a prominent 70-kDa species (Fig. 3B). This same 70-kDa species was stained with antibodies from several independent sources, including a mouse monoclonal and several rabbit polyclonal antibodies. Treatment of the cultures with the inducers D/M for up to 48 h does not appear to influence the steady-state level of expression of Syk in the 3T3-L1 cells. The level of G␤2 subunit of heterotrimeric G-proteins does not change during adipogenesis. Reprobing the blots with antibodies to G␤2 established the equivalence of sample loading of SDS-PAGE gels (Fig. 3C). The immunoblotting data are in agreement with the analysis of Syk mRNA by RT-PCR. Taken together, the immunoblotting and RT-PCR results establish the presence of the tyrosine kinase Syk and its mRNA in the mouse 3T3-L1 embryonic fibroblasts.
The initial observation of a p70 phosphoprotein whose phosphotyrosine content increased upon adipogenic induction by D/M (Fig. 2) was explored by use of antibodies capable of immunoprecipitating Syk from whole cell lysates. 3T3-L1 cells were challenged with inducers for 30 min to 48 h, and immunoprecipitation reactions were performed using the cell lysates followed by SDS-PAGE of the immunoprecipitates (Fig. 4A). Immunoblots of the immunoprecipitates of anti-Syk antibodies were stained with antibodies to phosphotyrosine (PY20) as well as with the anti-Syk antibodies itself. The phosphotyrosine content of Syk increased ϳ2-fold from 12 to 24 h post-induction. Within the 24 -48-h period post-induction, phosphotyrosine content of Syk returns to the initial, base-line value, in good agreement with the earlier observations (Fig. 2). The levels of phosphotyrosine were found to decline by 50 -60% within 30 min of induction (not shown) but then increased reproducibly within 24 h to ϳ2-fold over the zero time values.
The staining of the blots with anti-Syk antibodies revealed some variation in the amount of Syk immunoprecipitated over the time course of induction (Fig. 4A). Stripping and then restaining of the PY20 blots with anti-Syk antibodies demonstrates a modest decrease in the amount of immunoprecipitable Syk when measured at 12 h. The ratio of PY20 (phosphotyrosine content) signal to Syk signal ranged from 0.5 (at 30 min post-induction) to Ͼ2 (at 24 h post-induction) when compared with those for clones not treated with inducers. Control studies performed with increasing amounts of the cell lysate for immune precipitations from the untreated clones and the clones treated with inducer further establish, once corrected for the amount of immunoprecipitable Syk, that the phosphotyrosine content of Syk increases ϳ2-fold at 24 h post-induction of adipogenesis (Fig. 4B). The conditions employed for immunoprecipitation reactions performed with 1-4 mg of cell lysate protein demonstrate a linear relationship up to and including 3 mg of protein (Fig. 4B). The ratio of PY20 (phosphotyrosine content) to immunoreactive Syk ranged from 2.5 to 3.1 for the immunoprecipitation reactions performed with 1-3 mg of lysate protein/immunoprecipitation reaction for the clones stimulated with the inducers for 24 h. The results displayed earlier (Figs. 2 and 3) and all other immunoprecipitations were performed at a protein content within this linear range of the assay.
The central role of G s ␣ in adipogenesis and increases in phosphotyrosine content of Syk during adipogenesis provoked a test of a possible linkage between G s ␣ and Syk in this unique system in which G s ␣ represses adipogenesis and inducers such as D/M act to derepress the actions of G s ␣ (1, 6, 24, 25). This possible involvement of a nonreceptor tyrosine kinase with the ␣-subunit of a heterotrimeric G-protein was strengthened further by the recent discovery of the association of nonreceptor tyrosine kinase Btk with the ␣-subunit of the heterotrimeric G-proteins G q (30, 31) and G12 (32). Immunoprecipitations of Syk were performed using whole cell lysates from 3T3-L1 sta-

FIG. 4. Increased amounts of tyrosine-phosphorylated Syk within 24 h of the induction of adipogenesis by dexamethasone and methylisobutyl xanthine in mouse 3T3-L1 embryonal fibroblasts.
A, at confluence (day 0), cultures of mouse 3T3-L1 embryonic fibroblasts were treated without (zero time) or with D/M for 12, 18, 24, 30, 36, and 48 h. Immunoprecipitation (IP) of Syk was performed with whole cell lysates and either anti-Syk antibodies coupled to protein A/G-agarose or the protein A/G-agarose coupled to normal rabbit serum as a control. The immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose blots. The blots were stained with the PY20 antibody that specifically recognizes phosphotyrosine as well as with the anti-Syk antibody. Immune complexes were made visible by chemiluminescence and the use of a second goat anti-rabbit IgG (for anti-Syk immune complexes) or second goat anti-mouse IgG (for anti-phosphotyrosine immune complexes) to which horseradish peroxidase was coupled. The blots were scanned by densitometry. The ratio of absorbance of the immunoblotting (IB) of PY20 to anti-Syk was calculated to provide a means of evaluating the net increase in phosphorylation versus immunoprecipitated Syk. Zero time values were set as 1 for both the display of the phosphotyrosine content and Syk content of the immunoprecipitations. The data included on the figure were then calculated from the zero time values as -fold change. B, parallel experiments were performed with whole cell lysates from control cells and cells induced with D/M for 24 h. The immunoprecipitation reactions were performed with anti-Syk antibody coupled to protein A/G-agarose or the protein A/Gagarose coupled to normal rabbit serum as a control. The amount of whole cell lysate subjected to immunoprecipitation was varied from 1 to 4 mg of protein. The resultant blots were scanned, and the ratio of absorbance of the immunoblotting of PY20 to anti-Syk was calculated to provide a means of evaluating the increase in phosphorylation versus immunoprecipitated Syk. Note that the amounts of immunoprecipitated samples were proportional to the amounts of whole cell lysate protein employed over the range of at least 1-3 mg of lysate protein. The assays throughout this study were performed at 1-3 mg of protein, well within this linear range of performance. The data shown are representative of more than three independent experiments performed on as many separate cultures of 3T3-L1 cells.
bly transfected either with an empty expression vector (pCW1) or with the same vector constitutively expressing G s ␣ (pCW1-G s ␣). 3T3-L1 clones that constitutively express G s ␣ fail to undergo adipogenesis in response to inducers by virtue of their maintenance of G s ␣ levels (1). Much like wild-type 3T3-L1 cells, the clones stably transfected with the empty vector alone display increased levels of phosphorylated Syk (Fig. 5) and robust adipogenesis in response to D/M (not shown, see Ref. 1). The clones expressing G s ␣ constitutively, in sharp contrast, show no increase in phosphorylated/activated Syk (Fig. 5) and fail to undergo adipogenic conversion in response to the inducers (not shown; see Ref. 1). These data establish an additional linkage between G s ␣ and both Syk phosphorylation and adipogenesis.
We tested directly the possible association of G s ␣ with Syk via immunoprecipitations with anti-Syk antibodies followed by SDS-PAGE of the immunoprecipitates, immunoblotting and staining for G s ␣ (Fig. 6). In the lysates prepared from stable transfectants harboring an empty expression vector pCW1 (Fig. 6) or those from wild-type 3T3-L1 cells (not shown), staining of the anti-Syk immunoprecipitations also identify G s ␣. Staining of anti-Syk immunoprecipitations with antibody to G i ␣2, in contrast, displays no association with Syk (not shown). In the absence of inducers D/M (0 h time), G s ␣ is immunoprecipitated with anti-Syk antibodies from the whole cell lysates of the 3T3-L1 clones. The regions of the blots in which Syk migrates were stained with anti-Syk antibody as well (Fig. 6). When induced to adipogenesis, G s ␣ levels decline as previously noted (6,16), whereas the amounts of Syk in the immunoprecipitations increase (Figs. 4 -6). These data demonstrate an association of G s ␣ and Syk in the 3T3-L1 embryonic fibroblasts. Upon induction of adipogenesis, the amount of Syk available for immunoprecipitation and more importantly the tyrosine phosphorylation and activity of Syk increase. Immune precipitations performed with protein A/G-agarose coupled with normal rabbit serum, in contrast, display no signal for either G s ␣ or Syk.
To further test the linkage between G s ␣ and Syk association, immunoprecipitations with anti-Syk antibodies and anti-G s ␣ antibodies were performed using the whole cell lysates from stable transfectants of 3T3-L1 that fail to undergo adipogenesis by virtue of their constitutive expression of G s ␣ (1). When performed with 3T3-L1 clones constitutively expressing G s ␣, immunoprecipitations with anti-Syk antibodies demonstrate the co-immunoprecipitation of G s ␣ (Fig. 6), as observed in the untreated wild-type clones as well as clones expressing the empty expression vector. The wild-type clones and clones expressing the empty vector display a loss of G s ␣ (1) and an increase in the amount of Syk available for activation and immunoprecipitation in response to adipogenic inducers (Figs. 4 and 5). The 3T3-L1 clones constitutively expressing exogenous G s ␣ display, in contrast, persistent levels of G s ␣ content in the anti-Syk immunoprecipitations in the presence of inducers (Fig. 6). In keeping with the inability of the G s ␣-expressing clones to undergo adipogenic conversion in response to the inducers D/M, the amount of Syk available for immunoprecipitation remained unchanged in response to inducers and equivalent to untreated G s ␣-expressing clones and untreated wildtype 3T3-L1 clones.
Immune precipitations performed with antibodies to G s ␣ (Fig. 7A) confirm the observations obtained with anti-Syk immune precipitations, i.e. G s ␣ associates with Syk, and induction of adipogenesis results in loss of Syk associated with G s ␣. Expression of G s ␣ blocks adipogenesis and the decline in G s ␣-Syk association. Immune precipitations performed with anti-G i ␣2 antibodies, in contrast, reveal no association with Syk (not shown). Analysis of the activation of Syk under these same conditions using the phosphotyrosine-specific antibody PY20 in cell lysates first depleted of G s ␣ similarly reveals Syk activation only in the cells capable of adipogenic conversion in response to D/M (Fig. 7B). Measurement of Syk kinase activity was performed directly in the anti-Syk immunoprecipitates from clones expressing the empty vector (pCW1(EV)), or pCW1 (G␣ s ), using the band 3 GST fusion protein (GST-CDB3) as a substrate. A sharp increase in Syk activity was observed at 24 h of adipogenic induction (Fig. 7C). No such activation of Syk was observed in anti-Syk immunoprecipitations from the clones constitutively expressing G s ␣ (pCW1(G s ␣)) (Fig. 7C). The activity of Syk increased 3-fold at 24 h post-treatment with the inducers only for the clones expressing the empty vector (Fig. 7D). The clones stably expressing G s ␣, in contrast, fail to display any increase in Syk activity following the challenge with inducers (Fig. 7D). Immunoprecipitations performed with protein A/G-agarose coupled with normal rabbit serum, in contrast, displayed no Syk kinase activity (Fig. 7C). At confluence (day 0), cultures of mouse 3T3-L1 embryonic fibroblast clones stably transfected with either empty vector alone (pCW1(EV)) or the expression vector for G s ␣ (pCW1(G s ␣)) were treated without (0 time) or with D/M for 0.5, 12, 24, and 48 h. Immunoprecipitation (IP) of Syk was performed with whole cell lysates (1.2 mg of total cellular protein) and anti-Syk antibody. The immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose blots. The immunoblots (IB) were stained with the PY20 antibody that specifically recognizes phosphotyrosine as well as the anti-Syk antibody. Immune complexes were made visible by chemiluminescence and the use of a second either goat anti-rabbit IgG or anti-mouse IgG to which horseradish peroxidase was coupled. The data shown are representative of more than three independent experiments performed on as many separate cultures of 3T3-L1 cells.
FIG. 6. Co-immunoprecipitation of G s ␣ and Syk in whole cell lysates from mouse 3T3-L1 embryonic fibroblasts: immunoprecipitation analysis with anti-Syk antibodies. At confluence (day 0), cultures of mouse 3T3-L1 embryonic fibroblasts stably transfected with either empty vector alone (pCW1(EV)) or the expression vector for G s ␣ (pCW1(G s ␣)) were treated without (0 time) or with D/M for 24 h. Immunoprecipitation (IP) of Syk was performed with whole cell lysates and anti-Syk antibodies coupled to protein A/G-agarose and protein A/G-agarose coupled to normal rabbit serum as a control. The immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose blots (IB). The duplicate blots were stained with antibodies either to G s ␣ or to Syk. Immune complexes were made visible by a second goat antirabbit IgG to which calf alkaline phosphatase is coupled. The data shown are representative of more than three independent experiments performed on as many separate cultures of 3T3-L1 cells.
If Syk acts as a mediator for some feature of G s ␣ action in the derepression of adipogenic conversion by D/M, expression of the Syk itself might influence adipogenesis in these cells. Syk action has been studied largely in T-and B-cells and their signal transduction. A possible role for Syk in other differentiation pathways, such as adipogenesis, would be rather novel. The presence of Syk in immunoprecipitations of G s ␣ as well as the presence of G s ␣ in Syk immunoprecipitations are important observations, but they do not establish a biological role for the interaction in adipogenesis. The increase in both the amount of phosphorylation and kinase activity of Syk accompanying adipogenesis is, however, quite revealing. The inhibition of adipogenesis in response to D/M by the tyrosine kinase inhibitor genistein and the block of Syk activation and adipogenesis by the constitutive expression of G s ␣ also are quite compelling and lend greater weight to the immunoprecipitation data.
To further test the possible role of Syk more directly in the cells, 3T3-L1 cells were stably transfected with pCDNA3 expression vector harboring either Syk or a kinase-deficient Syk (Syk(KϪ)), and the stable transfectants were studied. Clones were selected, and the expression of the wild-type and kinasedeficient mutants was established by immunoblotting of whole cell lysates (Fig. 8A). Expression of Syk and the kinase-deficient mutant form of Syk were approximately 2-fold greater than the level of endogenous expression of Syk. In the presence of the inducers D/M, only the clones overexpressing Syk demonstrated marked oil red O staining (45% stained positive) at day 3 post-induction ( Fig. 8B and Table I). Because the adipogenic conversion typically is a 7-10-day process in response to inducers, the appearance of lipid droplets at day 3 reveals early differentiation in the clones overexpressing Syk. The wild-type 3T3-L1 clones as well as the Syk kinase-deficient clones, in contrast, displayed much less positive staining with oil red O at day 3 post-induction. At day 7 post-induction with D/M, the clones overexpressing Syk still demonstrate enhanced oil red O staining (Ͼ60% stained positive), although the wild-type clones now display a positive staining also (Table I). The clones expressing the Syk kinase-deficient mutant generally displayed much less oil red O staining than the clones overexpressing Syk, suggesting a need for kinase-active Syk in adipogenesis. Expression of kinase-deficient Syk, however, does not block the ability of the cells ultimately to differentiate. The endogenous level of wild-type Syk may well be sufficient to permit adipogenesis in the presence of the expressed levels of mutant, kinase-deficient Syk obtained in these clones. Thus, overexpression of Syk facilitates the ability of D/M to induce adipogenic conversion in 3T3-L1 cells.
The ability of Syk to influence adipogenesis in the absence of inducers was explored. At 10 days after confluence, cultures of 3T3-L1 begin to display some adipogenic conversion in the absence of inducers (Fig. 8C). The wild-type clones reveal oil red O staining (14% positive staining) by day 10 ( Table I). The clones overexpressing Syk display more marked adipogenic conversion, reflected by enhanced positive staining of lipid by oil red O (Ͼ30% positive staining; see Table I). Expression of the kinase-deficient Syk resulted in a modest but clear reduction in adipogenic conversion, displaying oil red O staining that is less than that observed in the wild-type cells. When overexpressed at levels approximately 2-fold over endogenous levels, Syk itself appears capable of promoting adipogenic conversion. the protein A/G-agarose coupled to normal rabbit serum as a control. The Syk kinase activity was measured in the immunoprecipitates using a GST-band 3 fusion protein as the substrate. D, Syk kinase activity assayed as described in C. The data displayed are the mean values of replicate experiments with deviation from the mean of less than 10% (n ϭ 2).

FIG. 7.
Co-immunoprecipitation of G s ␣ and Syk in whole cell lysates from mouse 3T3-L1 embryonic fibroblasts: immunoprecipitation analysis with anti-G s ␣ antibodies and Syk kinase determination. At confluence (day 0), cultures of mouse 3T3-L1 embryonic fibroblasts stably transfected with either empty vector alone (pCW1(EV)) or the expression vector for G s ␣ (pCW1(G s ␣)) were treated without (0 time) or with D/M for 24 h. A, immunoprecipitation of G s ␣ was performed in whole cell lysates with anti-G s ␣ antibodies coupled to protein A/G-agarose. The immunoprecipitates (IP) were subjected to SDS-polyacrylamide gel electrophoresis, and the resolved proteins were transferred to nitrocellulose blots. The blots were stained with antibodies either to Syk or to G s ␣. B, the whole cell lysate depleted with anti-G s ␣ antibodies was then treated with antibodies to Syk, and the immunoprecipitates were subjected to SDS-PAGE. The resolved proteins were subjected to immunoblotting (IB) and stained with antibodies either to Syk or to PY20. Note that in the wild-type cells, the loss of G s ␣ induced by adipogenesis leads to the appearance of phosphorylated Syk, whereas the cells constitutively expressing G s ␣ fail to display activated Syk. Immune complexes were made visible by a second goat anti-rabbit IgG to which calf alkaline phosphatase is coupled or a goat anti-mouse IgG to which horseradish peroxidase is coupled. The data shown are representative of more than three independent experiments performed on as many separate cultures of 3T3-L1 cells. C, Syk kinase activation occurs in response to induction of adipogenesis but not in 3T3-L1 clones in which the induction of adipogenesis and decline in G s ␣ is prevented by constitutive expression of G s ␣. Whole cell lysates were prepared from clones constitutively expressing G s ␣ and those harboring the empty vector, both treated with or without dexamethasone and methylisobutylxanthine for 24 h. Immunoprecipitations were performed with antibodies to Syk coupled to protein A/G-agarose or with Induction of adipogenesis with D/M increased phosphotyrosine content of Syk in the wild-type cells as well as in the clones stably overexpressing Syk (Fig. 9). The amount of Syk increased 2-fold in the Syk-overexpressing clones as compared with wild-type cells, as determined by immunoblotting. Thus, Syk is phosphorylated and activated in response to induction of adipogenesis. Overexpression of Syk is adipogenic itself, enhances adipogenesis in response to D/M, and results in in-  At confluence (day 0), cultures of mouse 3T3-L1 embryonic fibroblasts were treated without (ϪD/M) or with (ϩD/M) dexamethasone and methylisobutylxanthine. Dexamethasone and methylisobutylxanthine were removed after incubation for 2 days, and the cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum for the number of days indicated. At the day of post-confluent growth indicated (day 1, 3, 7, or 10), cells were fixed by 3% paraformaldehyde for 5 min and stained for lipid accumulation with oil red O for 10 min. Hematoxylin (1%) was used to stained nuclei. Adipogenesis of four individual stably transfected clones was examined under a Zeiss Axiophot microscope. The red-stained bodies of the cytosol are oil droplets. Bar, 100 m. Data presented are representative of experiments performed on each of the four separate clones assayed on at least two occasions, providing essentially identical data. The percentage of the cells of each clone that scored positive (oil red O staining) were quantified for each of the clones and conditions described (Table I).
FIG. 9. Tyrosine phosphorylation of Syk in wild-type mouse 3T3-L1 clones and clones stably overexpressing Syk: response to induction of adipogenesis. Crude cell lysates were prepared from wild-type 3T3-L1 clones (WT) and clones stably expressing Syk that were challenged with dexamethasone and methylisobutylxanthine to induce adipogenesis. Left-hand panel, the samples (3 mg of cellular protein of each whole cell extract) were subjected to immunoprecipitation reactions employing antibodies to phosphotyrosine (PY20) that were first chemically coupled to protein A/G-agarose and the immunoprecipitates (IP) subjected to SDS-polyacrylamide gel electrophoresis. Protein A/G-agarose coupled with normal rabbit serum provided a control for the immunoprecipitations. The resolved proteins from the immunocomplexes were transferred to nitrocellulose blots. Then the blots were stained with a second, anti-phosphotyrosine antibody (PY69). The p70 phosphotyrosine-containing phosphoprotein was established as authentic Syk by staining with three different polyclonal antibodies and one monoclonal antibody, obtained from independent sources. The data shown are representative of three independent experiments performed on as many separate cultures of 3T3-L1 cells. Right-hand panel, a sample of whole cell lysate was employed for SDS-PAGE, immunoblotting (IB), and staining with anti-Syk antibody to measure the level of Syk expression. creased phosphorylation and activation in response to treatment with D/M. Several observations herein support the notion that G s ␣ represses adipogenesis in 3T3-L1 cells via the nonreceptor tyrosine kinase Syk, including: 1) the ability of the broad action tyrosine kinase inhibitor genistein to block adipogenic conversion; 2) phosphorylation/activation of Syk in 3T3-L1 cells early in adipogenesis; 3) increased levels of immunoprecipitable and tyrosine-phosphorylated Syk in response to inducers that is blocked by overexpression of G s ␣; 4) co-precipitation of G s ␣ with immunoprecipitation of Syk; 5) co-precipitation of Syk with G s ␣ in the naive but not the clones induced to adipogenic conversion; and 6) overexpression of Syk enhances adipogenic conversion in response to inducers as well as displays a modest capacity to induce adipogenic conversion itself in the absence of inducers like D/M. The association between Syk and G s ␣ may be fortuitous, but recent discovery of the association between the nonreceptor tyrosine kinase Btk and the G-protein ␣ subunits Gq␣ (30,31) and G12␣ (32) provide an example in which a tyrosine kinase acts as an effector for a G-protein.
For adipogenesis, the sharp reduction in G s ␣ expression that accompanies adipogenic conversion, the ability of oligodeoxynucleotides antisense to G s ␣ to suppress G s ␣ and to promote adipogenesis, and the ability of overexpression of G s ␣ to block adipogenesis in response to inducers provide an interesting model for G-protein regulation of differentiation, not unlike the case of signaling in the T-and B-cells for Gq␣ via Btk (30,31). Clearly overexpression of Syk promotes a robust, earlier adipogenic conversion. Based upon these observations by others and the data presented herein, we propose that G s ␣ represses aspects of adipogenesis via Syk, a nonreceptor tyrosine kinase heretofore shown to participate in signaling and differentiation of T-and B-cells.