c-Crk, a Substrate of the Insulin-like Growth Factor-1 Receptor Tyrosine Kinase, Functions as an Early Signal Mediator in the Adipocyte Differentiation Process*

Differentiation of 3T3-L1 preadipocytes into adipocytes is induced by a combination of inducers, including a glucocorticoid, an agent that elevates cellular cAMP, and a ligand of the insulin-like growth factor-1 receptor. Previous studies have implicated protein-tyrosine phosphatase (PTPase) HA2, a homologue of PTPase 1B, in the signaling cascade initiated by the differentiation inducers. Vanadate, a potent PTPase inhibitor, blocks adipocyte differentiation at an early stage in the program, but has no effect on the mitotic clonal expansion required for differentiation. Exposure of preadipocytes to vanadate along with the inducing agents led to the accumulation of pp35, a phosphotyrosyl protein that is a substrate for PTPase HA2. pp35 was purified to homogeneity and shown by amino acid sequence and mass analyses of tryptic peptides to be c-Crk, a known cytoplasmic target of the insulin-like growth factor-1 receptor tyrosine kinase. Transfection of 3T3-L1 preadipocytes with a c-Crk antisense RNA expression vector markedly reduced c-Crk levels and prevented differentiation into adipocytes. Studies with C3G, a protein that binds to the SH3 domain in c-Crk, showed that phosphorylation of c-Crk rendered the SH3 domain inaccessible to C3G. Taken together, these findings indicate that locking c-Crk in the phosphorylated state with vanadate prevents its participation in the signaling system that initiates adipocyte differentiation.

Adipocytes serve an important function in the energy economy of higher organisms, providing a large energy reserve that can be mobilized when needed. Thus, when caloric intake exceeds expenditure, metabolite flux is diverted into triglyceride synthesis for storage in adipocytes. Conversely, when caloric expenditure exceeds intake, this triglyceride reserve is mobilized as free fatty acids to provide physiological fuel for use by other tissue/cell types. The need for an energy reserve begins at birth when the newborn must be prepared to survive periods of energy deprivation. Adipocytes, which provide this reserve, develop late in embryonic life, with major expansion of this cell population occurring after birth. The adipose lineage arises from the same multipotent stem cells of mesodermal origin that give rise to the muscle and cartilage lineages.
Established preadipocyte cell lines, e.g. the 3T3-L1 preadipocytes, which can be induced to differentiate into adipocytes in cell culture, provide faithful models with which to investigate the adipocyte differentiation program (1)(2)(3)(4)(5)(6). When exposed to the appropriate differentiation inducers, including IGF-1 1 (or insulin at a non-physiologically high concentration), dexamethasone (a glucocorticoid), and 1-methyl-3-isobutylxanthine (MIX; a cAMP phosphodiesterase inhibitor that increases intracellular cAMP), 3T3-L1 preadipocytes differentiate into cells that express the adipocyte phenotype (6). Induction of the adipocyte differentiation program involves at least three different signal transduction systems, including those mediated by the glucocorticoid receptor, the cAMP-dependent protein kinase, and the IGF-1 receptor (6 -10). Activation of these pathways triggers the sequential expression of a group of transcription factors (9 -16), including members of the C/EBP family (C/EBP␣, -␤, and -␦) and peroxisomal proliferator-activated receptor-␥, leading to the coordinate transcriptional activation of a large number of adipocyte genes that produce the differentiated phenotype. Once adipocyte gene expression has been initiated, further stimulation with differentiation inducers is no longer required.
Although the transcriptional activation of adipocyte-specific genes during differentiation has been studied intensively (17)(18)(19)(20)(21)(22)(23)(24), far fewer studies have been conducted on the signal transduction pathways by which the differentiation inducers act. Upon exposure to these inducers, confluent growth-arrested preadipocytes synchronously reenter the cell cycle and undergo approximately two rounds of mitosis (25), referred to as "mitotic clonal expansion." Mitotic clonal expansion is required for progression through subsequent steps in the differentiation program (26,27). It appears that DNA replication and chromatin remodeling during mitotic clonal expansion render cis-ele-ments accessible to the trans-acting factors that activate (or derepress) transcription of genes critical to progression of the differentiation program (9).
The requirement of IGF-1 (or a high level of insulin, which can also activate the IGF-1 receptor) as a differentiation inducer for 3T3-L1 preadipocytes implicated tyrosine phosphorylation in the induction process, as the IGF-1 receptor is a ligand-activated tyrosine kinase (8). Previously, we showed that the expression of PTPase HA2 is both regulated during and required for differentiation of 3T3-L1 preadipocytes (28). The expression of PTPase HA2 increases dramatically for 2 days following induction of differentiation and then decreases (28). Furthermore, constitutive overexpression of PTPase HA2 by 3T3-L1 preadipocytes transfected with a PTPase HA2 expression vector blocks differentiation. Importantly, however, exposure of the transfected preadipocytes to vanadate (a potent PTPase inhibitor) at the time when the endogenous PTPase is normally down-regulated during differentiation, i.e. following clonal expansion, fully restores their capacity to differentiate into adipocytes (28). Exposure to vanadate at any other time during the differentiation program, however, fails to restore differentiation. Moreover, inhibition of PTPase HA2 activity with vanadate in untransfected 3T3-L1 cells (between days 0 and 2 of the standard differentiation protocol) also blocks differentiation (28). Taken together, these findings suggest that fluctuation of PTPase HA2 activity early in the differentiation program is both regulated during and required for adipocyte differentiation. Thus, it appears that a protein, phosphorylated by the IGF-1 receptor, is generated early in the program and is subsequently dephosphorylated by PTPase HA2. Conceivably, the coordinated sequential actions of the IGF-1 receptor tyrosine kinase and PTPase HA2 generate a signal required for the induction process.
In this work, we provide evidence that c-Crk (cellular CT10 regulator of kinase), a bona fide substrate of the IGF-1 receptor tyrosine kinase, functions "early" in the induction of adipocyte differentiation. Our results suggest that signaling, initiated by IGF-1 and mediated by c-Crk, involves tyrosine phosphorylation, followed by dephosphorylation. It appears that in the phosphorylated state, the SH3 domain of c-Crk is blocked due to an intramolecular interaction of the SH2 domain with the phosphotyrosyl group, thereby preventing its interaction with a putative downstream signaling molecule. Presumably, dephosphorylation by PTPase HA2 would allow this interaction to occur at the appropriate time in the differentiation program.
Cell Culture, Differentiation, and Vanadate Treatment of 3T3-L1 Preadipocytes-3T3-L1 preadipocytes were cultured in DMEM supplemented with 10% calf serum and allowed to reach confluence. Differentiation of 2-day post-confluent preadipocytes (designated as day 0) was initiated with 1 g/ml insulin, 1 M dexamethasone, and 0.5 mM MIX in DMEM supplemented with 10% fetal bovine serum (6,29). After 48 h (day 2), the culture medium was replaced with DMEM supplemented with 10% fetal bovine serum and 1 g/ml insulin, and the cells were then fed every other day with DMEM containing 10% fetal bovine serum. Cytoplasmic triglyceride droplets were visible by day 4, and cells were fully differentiated by day 8.
For vanadate treatment, 20 M sodium vanadate was added to the culture medium along with MIX, dexamethasone, and insulin on day 0. After 48 h (day 2), the cells were shifted to normal medium containing 10% fetal bovine serum and 1 g/ml insulin, and then the normal differentiation protocol was followed. For re-induction of differentiation after vanadate treatment, day 6 3T3-L1 cells that had been exposed to the differentiation protocol with vanadate were re-induced with MIX, dexamethasone, and insulin following the normal differentiation protocol. For delayed vanadate treatment, 20 M sodium vanadate was added to the cells after induction with MIX, dexamethasone, and insulin for the indicated time periods. On day 2, the medium was changed, and the cells were cultured as usual.
Cell Counting and Oil Red O Staining-3T3-L1 cells (6-cm plate) were trypsinized from the culture dishes and collected by centrifugation. An aliquot was subjected to cell counting using a hemocytometer plate. For oil red O staining, 3T3-L1 adipocyte monolayers (usually on day 8) were washed three times with phosphate-buffered saline (PBS) and then fixed for 2 min with 3.7% formaldehyde in PBS. Oil red O (0.5%) in isopropyl alcohol was diluted with 1.5 volumes of water, filtered, and added to the fixed cell monolayers for 1 h at room temperature. Cell monolayers were then washed with water, and the stained triglyceride droplets in the cells were visualized and/or photographed.
Isolation of RNA and Analysis of Expression of C/EBP␣ and 422/ aP2 mRNAs-Total cellular RNA was isolated by the guanidine isothiocyanate method (30) from day 0 (2-day post-confluent) cells and cells at various times after induction of differentiation in the presence or absence of vanadate. For Northern blot analysis of C/EBP␣ and 422/aP2 mRNAs, 20 g of total RNA were denatured with glyoxal and dimethyl sulfoxide and resolved by electrophoresis on 1% agarose gels in 10 mM sodium phosphate buffer (pH 7.0) as described (31,32). After transfer to Hybond-N membrane (Amersham Pharmacia Biotech), UV cross-linking, and removal of glyoxal, RNA was stained with methylene blue to locate 28 S and 18 S rRNAs and to verify equal loading. Blots were then hybridized overnight at 42°C with C/EBP␣ and 422/aP2 cDNA probes in 50% formamide, 4ϫ SSC, 1ϫ Denhardt's solution, 50 mM sodium phosphate (pH 7.0), 1% SDS, 100 g/ml denatured salmon sperm DNA, and 0.5 mg/ml sodium pyrophosphate. The blots were washed twice with 0.1ϫ SSC and 0.1% SDS at 50°C and twice with 0.1ϫ SSC and 0.1% SDS at 65°C and then visualized with a PhosphorImager.
Preparation of Cell Extracts and SDS-PAGE-For analysis of tyrosine-phosphorylated proteins, cell monolayers from cells treated as described in the figure legends were washed three times with cold PBS supplemented with 0.1 mM sodium vanadate. Cells were then scraped from the plates into hypotonic buffer containing 10 mM Hepes (pH 7.0), 2 mM MgCl 2 , 15 mM KCl, 0.1 mM phenylarsine oxide (PAO; a PTPase inhibitor) (33, 34), 1 mM sodium vanadate, 2 l/ml protease inhibitor mixtures 1 and 2 (PIC1 and PIC2, respectively) (35), and 1 mM PMSF and then homogenized with a glass homogenizer. Cellular membranes and cytosol were separated by centrifugation of the cell lysate at 150,000 ϫ g for 45 min at 4°C. Membranes were then extracted with 10 mM Hepes (pH 7.0), 1% Triton X-100, 0.3 M NaCl, 1 mM sodium vanadate, 0.1 mM PAO, 1 mM PMSF, and 2 l/ml PIC1 and PIC2 for 1 h at 4°C. The extract was centrifuged at 150,000 ϫ g for 45 min at 4°C. The supernatant is referred to as the solubilized membrane extract, and the pellet as the insoluble membrane fraction. The total cell extract was prepared by washing the cell monolayers with cold PBS as described above, followed by lysis with 1ϫ boiling Laemmli SDS sample buffer (36) containing 20 mM dithiothreitol, 0.1 mM PAO, and 1 mM sodium vanadate. The cell lysate was then heated at 100°C for 5 min.
For analysis, cell extracts (usually containing 10 -50 g of protein) were subjected to SDS-PAGE and then transferred to Immobilon-P membrane (Millipore Corp.). After blocking with 2% nonfat dried milk in 1ϫ Tween/Tris-buffered saline containing 25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.05% Tween, and 0.001% Merthiolate for 2 h at room temperature, membranes were incubated with primary antibody for 2 h at room temperature, followed by horseradish peroxidase-conjugated secondary antibody for 45 min. Target proteins were visualized by enhanced chemiluminescence (ECL).
The pp35 protein for use as substrate was prepared from day 2 cells after hormonal stimulation and vanadate treatment. In brief, 2-day post-confluent 3T3-L1 preadipocytes were treated with 1 g/ml insulin, 1 M dexamethasone, 0.5 mM MIX, and 20 M sodium vanadate for 2 days, after which cell monolayers were washed twice with PBS; scraped from the culture dishes into hypotonic buffer containing 10 mM Hepes (pH 7.0), 15 mM KCl, 2 mM MgCl 2 , 0.1 mM PAO, 1 mM PMSF, and 2 l/ml PIC1 and PIC2; and homogenized. The cytosolic fraction was separated by centrifugation at 150,000 ϫ g for 45 min. Vanadate was omitted from the hypotonic buffer since it interferes with the PTPase assay; the high concentration of DTT in the dephosphorylation assay neutralized the PAO.
Dephosphorylation of pp35 by PTPase HA2 was followed using the PTPase HA2 activity assay described previously (33). The PTPase HA2 preparation was incubated with pp35 for the indicated times at 30°C in a reaction mixture containing 50 mM PIPES (pH 6.5), 1 mM EDTA, and 5 mM DTT. The reaction was terminated by adding 3-fold concentrated Laemmli SDS sample buffer. The amount of pp35 protein was quantitated by SDS-PAGE and Western immunoblotting with anti-phosphotyrosine antibody as described above.
Purification of pp35 from 3T3-L1 Cells-650 10-cm cell monolayers, treated with 1 g/ml insulin, 1 M dexamethasone, 0.5 mM MIX, and vanadate (20 M on day 1 and an additional 15 M on day 2 to maximize pp35 accumulation) for 2 days, were washed twice with cold PBS containing 0.1 mM vanadate; scraped from the culture dishes; resuspended in hypotonic buffer containing 10 mM Hepes (pH 7.0), 2 mM MgCl 2 , 15 mM KCl, 0.1 mM PAO, 1 mM vanadate, 1 mM PMSF, and 2 l/ml PIC1 and PIC2; and homogenized using a glass homogenizer. After centrifugation at 150,000 ϫ g for 45 min, the supernatant (ϳ500 ml) was retained, and proteins were precipitated by bringing the solution to 60% saturation with solid ammonium sulfate. After stirring slowly for 30 min on ice, the mixture was centrifuged at 12,000 ϫ g for 15 min, and the supernatant was discarded. The protein pellet was redissolved in 80 ml of 20 mM Tris-HCl (pH 7.5), 10 mM NaCl, 0.1 mM vanadate, 0.1 mM PAO, 0.1 mM PMSF, 1 mM EDTA, and 2 l/ml PIC1 and PIC2 (Buffer A). The solubilized proteins were then dialyzed against the same buffer for 3 h at 4°C. After removal of undissolved proteins by centrifugation, the supernatant was used for subsequent chromatographic purification. With this cytosolic fractionation and the ammonium sulfate precipitation steps, 1400 mg of total cellular proteins were reduced to 350 mg, and the recovery of pp35 was Ͼ90%.
The first chromatographic step was Bio-Scale DEAE 10 FPLC. The above solubilized proteins (ϳ90 mg of proteins were loaded onto the column each time) were applied to a 10-ml column equilibrated with Buffer A at a flow rate of 0.5 ml/min. The column was washed with 30 ml of Buffer A and eluted with a two-step elution: 15 ml of 200 mM NaCl in Buffer A, followed by a linear gradient of 200 -500 mM NaCl (85 ml) in the same buffer at a flow rate of 2 ml/min. Twenty-five fractions (4 ml/fraction) were collected. The eluted fractions were monitored for pp35 by Western blotting with anti-phosphotyrosine antibody as described above. Protein concentration was determined by the Lowry method. After this step, proteins were reduced to ϳ3.6 mg, with a pp35 yield of ϳ23%.
Following Bio-Scale DEAE 10 FPLC, HiTrap heparin FPLC and Resource-Q FPLC were used to further separate pp35. pp35-containing fractions from the Bio-Scale DEAE 10 chromatography were pooled; buffer-exchanged to a column loading buffer containing 20 mM NaCl, 20 mM Tris-HCl (pH 7.5), 0.1 mM vanadate, 0.1 mM PAO, 0.1 mM PMSF, and 2 l/ml PIC1 and PIC2; and then applied to a HiTrap heparin affinity column pre-equilibrated with the same buffer at flow rate of 0.25 ml/min. The flow-through fraction containing all of the pp35 was immediately applied to a Resource-Q column equilibrated with the same buffer. After washing the column with 5 ml of the equilibration buffer, proteins were eluted with a 20-ml linear gradient of 20 -500 mM NaCl in the loading buffer (1 ml/fraction) at a flow rate of 0.5 ml/min. Fractions containing pp35 were pooled and concentrated with a Centricon C-10 microconcentrator (Millipore Corp.) to ϳ200 l.
The final purification step was gel filtration by Superose 12 FPLC. The concentrated pooled fraction (containing pp35) from the Resource-Q FPLC was applied to a Superose 12 HR 10/30 column and then eluted with buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.1 mM vanadate, 0.1 mM PAO, and 0.1 mM PMSF at flow rate of 0.4 ml/min. The first 8-ml eluate representing the void volume of the column was collected in one pool, and then 25 0.4-ml fractions were collected. After the chromatographic purification, the overall pp35 yield was ϳ10%, and the pp35 enrichment was 40,000-fold. The pp35-containing fractions were pooled and stored at Ϫ70°C.
To verify that the purified protein was truly pp35, a sample from the Superose 12 FPLC was mixed with an equal volume of isoelectric focusing gel loading buffer containing 9.5 M urea, 1.8% CHAPS, 20 mM DTT, and 2% ampholyte 3-10. This sample mixture was loaded onto an isoelectric focusing tube gel (2 mm ϫ 6.5 cm) that had been preelectrophoresed at 200 V for 10 min, 300 V for 20 min, and 400 V for 20 min. The isoelectric focusing gel was then run at 500 V for 10 min and 750 V for 4 h. After equilibration in 1ϫ Laemmli electrophoresis buffer for 15 min, the tube gel was cut into 14 fractions, each of which were divided into two parts and loaded onto two identical 10% SDS-polyacrylamide gels. After electrophoresis, one gel was fixed and silverstained to reveal the protein, and the other was transferred to Immobilon-P membrane for immunoblotting with anti-phosphotyrosine antibody to reveal pp35. The resulting gels revealed that a single silver-stained protein band and the anti-phosphotyrosine antibody-detected pp35 band were at an identical place on the gels.
Identification of pp35 as the Proto-oncogene Product c-Crk by Amino Acid Sequence and Mass Analyses of Tryptic Peptides Derived from pp35-For amino acid sequence analysis, the peak pp35-containing fractions from the Superose 12 FPLC were pooled and concentrated using the Centricon C-10 microconcentrator. The sample was subjected to 10% SDS-PAGE and then transferred to an Immobilon-P membrane in 10 mM CAPS (pH 11.0) and 10% methanol at 270 mA for 10 h. Protein on the membrane was stained with 0.1% Ponceau S in 1% acetic acid solution, and the pp35 band was cut out and destained in PBS. The protein on the membrane was extracted twice for 3 h at 37°C with 200 l of 40% acetonitrile and once for 20 min at 50°C with 200 l of 40% acetonitrile and 0.05% trifluoroacetic acid. All eluates were pooled and lyophilized. The dried protein was dissolved in 50 l of 100 mM Tris-HCl (pH 8.5) and 2% acetonitrile solution and then digested with trypsin (one-twentieth the amount of pp35 protein) for 48 h at 37°C. After digestion, the sample was dried in a Speed-Vac, dissolved in 100 l of Buffer A (0.06% trifluoroacetic acid solution), and injected into a preequilibrated microbore C-18 reverse-phase high-performance liquid chromatography column (2.1 ϫ 100 mm). After washing with Buffer A for 15 min at a flow rate of 0.2 ml/min, peptides were eluted at a flow rate of 0.2 ml/min with a two-step gradient: first, 100% Buffer A to 90% Buffer A and 10% Buffer B (0.052% trifluoroacetic acid and 80% acetonitrile solution) in 5 min; and second, 90% Buffer A and 10% Buffer B to 20% Buffer A and 80% Buffer B in 100 min. The effluent was monitored for UV absorbency at 214 nm, and the fractions were collected manually. Three peptide peaks (eluting at 42, 52, and 56 min) were subjected to matrix-assisted laser desorption ionization mass spectral analysis and Edman amino acid sequencing.
Four tryptic peptide sequences were obtained from these three peaks, i.e. peptide 1, SSWYWGR; peptide 2, DKPEEQWWNAEDSEGK; peptide 3, GMIPVPYVEK; and peptide 4, LLDQQNPDEDFS. Peptides 2 and 4 were each from a single peak. The amino acid assignments for these two peptides were verified by the mass spectral analysis. Peptides 1 and 3 were from one reverse-phase chromatographic peak. Owing to the large differences in the amounts of these two peptides, the amino acid sequences for both peptides could be readily deduced from each Edman degradation cycle. The amino acid sequences of the four peptides were used to search the GenBank TM protein sequence data base. The computer search for sequence similarities of these four peptides revealed a 100% amino acid sequence match with the mouse protooncogene product c-Crk, an adapter molecule. The results from matrixassisted laser desorption ionization mass spectral analysis of these four tryptic peptides corresponded exactly with the masses of predicted tryptic cleavage products of mouse c-Crk. The sequence of peptide 4 matched that of the C terminus of c-Crk, explaining why this peptide did not end with Arg or Lys. This purified pp35 was also immunoblotted by anti-Crk antibody on a Western blot (see Fig. 7C, panel II). Taken together, these results provide strong evidence that pp35 is the tyrosine-phosphorylated form of mouse c-Crk, a known substrate of the IGF-1 receptor tyrosine kinase (39,40).
Construction of and Transfection with c-Crk Antisense and Sense RNA Expression Vectors-Two pBCMGneo expression vectors (18,28) were constructed: pBCMGneo-Crk/Antisense, with a 200-base pair Crk cDNA fragment (from the BamHI site in the 5Ј-untranslated region to the XhoI site in the coding region) (37) inserted in the antisense orientation, and pBCMGneo-Crk/Sense, with the same fragment inserted in the sense orientation. These vectors were transfected into 30% confluent low-passage 3T3-L1 preadipocytes using the calcium phosphate precipitation method (38) to generate stably transfected cell lines. Briefly, 20 g of pBCMGneo-Crk/Antisense or pBCMGneo-Crk/Sense vector DNA in a 250 mM CaCl 2 solution were added to an equal volume solution containing 280 mM NaCl, 50 mM Hepes (pH 7.12), and 1.5 mM Na 2 HPO 4 to form DNA/calcium phosphate coprecipitates, and the mixture was then added directly to the culture medium. After 8 h at 37°C in the CO 2 incubator, cells were shocked with 10% dimethyl sulfoxide and PBS for 3 min and then returned to the incubator for 24 h in fresh DMEM containing 10% calf serum. G418 was added to select the neomycin-resistant cells. The antibiotic-resistant foci were isolated and propagated to generate stable cell lines for further analysis.
Immunoprecipitation-3T3-L1 cell monolayers (10-cm plate) treated as described in the figure legends were washed twice with ice-cold PBS and lysed in 1 ml of 1% Triton X-100 buffer containing 50 mM Hepes (pH 7.4), 2.5 mM EDTA, 150 mM NaCl, 30 mM ␤-glycerophosphate, 1 mM sodium orthovanadate, 1 mM PMSF, and 2 l/ml PIC1 and PIC2. The lysed cells were homogenized and extracted at 4°C for 1 h. Insoluble material was removed by centrifugation at 12,000 ϫ g for 15 min at 4°C. Cell extract (1000 g of total protein) was mixed with 1 g of anti-Crk antibody for 2 h at 4°C, and then protein A-agarose was added to the mixture at 4°C overnight. After incubation, the agarose beads were collected by centrifugation at 1000 ϫ g for 5 min at 4°C. The pellet was then washed four times with the Triton X-100 lysis buffer and resuspended in 1ϫ Laemmli SDS sample buffer with 20 mM dithiothreitol. After heating at 100°C for 5 min, the sample was subjected to SDS-PAGE and Western immunoblot analysis.
Immunofluorescence-3T3-L1 cells were cultured and induced on glass coverslips. At the indicated times, the coverslips were rinsed with PBS and fixed for 10 min in 3.7% formaldehyde and 0.18% Triton X-100 in PBS solution. The fixed cells were incubated in blocking buffer (1% bovine serum albumin in Tween/Tris-buffered saline) for 2 h at room temperature, with primary anti-Crk antibody at 4°C overnight, and with tetramethylrhodamine B isothiocyanate-conjugated secondary antibody for 1 h at room temperature, after which the cells were visualized by fluorescence microscopy.

Vanadate Inhibits an Early
Step in Adipocyte Differentiation without Blocking Mitotic Clonal Expansion-Mitotic clonal expansion is an essential step that occurs early in the differentiation program (7,9,10). A previous study (28) showed that vanadate, a potent PTPase inhibitor, blocks differentiation of 3T3-L1 preadipocytes when added along with differentiation inducers during the first 48 h, during which mitotic clonal expansion occurs. To locate the time window during this period when vanadate exerts its inhibitory effect, preadipocytes were induced to differentiate in the presence or absence of 20 M vanadate either from 0 to 48 h or from 24 to 48 h after induction. Differentiation was assessed initially by the extent of staining of accumulated cytoplasmic triacylglycerol with oil red O. As illustrated in Fig. 1A, the presence of vanadate during the first 24 h completely prevented triacylglycerol accumulation. When vanadate addition was delayed for 24 h after the induction of differentiation, however, the preadipocytes differentiated normally. Moreover, when preadipocytes with differentiation blocked by exposure to vanadate for 0 -48 h were later subjected to the differentiation protocol (i.e. on day 6), the cells differentiated into adipocytes normally. Thus, inhibition by vanadate is reversible. Other experiments in which the addition of vanadate was delayed for differing periods of time following induction revealed that the period of greatest susceptibility to inhibition by vanadate is between 0 and 20 h after induction of differentiation (Fig. 1C). To verify that the effect of vanadate on cytoplasmic triacylglycerol accumulation truly reflects inhibition of differentiation per se rather than merely an inhibition of lipogenesis, the effect of vanadate on the expression of two adipocyte gene markers, i.e. the 422/aP2 and C/EBP␣ mRNAs, was assessed by Northern analysis. C/EBP␣ is a key transcription factor that coordinately transactivates numerous adipocyte-specific genes during differentiation, in-cluding the 422/aP2 gene (17). As illustrated in Fig. 1B, vanadate inhibited the expression of both of these marker genes.
During the first 48 h after induction of differentiation, 3T3-L1 preadipocytes synchronously undergo mitotic clonal expansion progressing through approximately two rounds of cell division. This process does not appear to be affected by vanadate; thus, the increase in cell number following induction of differentiation was the same whether vanadate was present or not (Fig. 2). To ascertain whether vanadate might cause a more subtle change in cell cycling during mitotic clonal expansion,

FIG. 1. Effect of vanadate on the accumulation of cytoplasmic triglyceride and expression of adipocyte genes by 3T3-L1 cells after induction of differentiation.
A, 3T3-L1 preadipocytes were induced to differentiate under different conditions in the absence and presence of vanadate. Cytoplasmic triglyceride was visualized by staining with oil red O on day 8. control, cells were subjected to the standard differentiation protocol; ϩvanadate, cells treated with 20 M sodium vanadate during the period of hormonal stimulation (first 48 h); ϩvanadate 24-h delay, cells treated with vanadate after they were induced 24 h after induction of differentiation was initiated; ϩvanadate reinduced, cells treated as described for ϩvanadate, after which the medium was replaced on day 2 following the normal differentiation protocol, and the cells were cultured to day 6, when cells were reinduced, i.e. subjected again on day 6 to the differentiation protocol. B, total cellular RNA was isolated from the cells on the days indicated (after induction of differentiation) and subjected to Northern blot analysis using C/EBP␣ and 422/aP2 cDNA probes. Contr. and ϩVan. are as described for A. Numbers 0 -3 refer to days after induction of differentiation. C, 3T3-L1 preadipocytes were induced to differentiate on day 0; 20 M sodium vanadate was added to the medium at the indicated times after the addition of the differentiation inducers.
fluorescence-activated cell sorter analysis was conducted on 3T3-L1 cells during the first round of differentiation-induced mitotic clonal expansion. Based on the fluorescence-activated cell sorter analysis (data not shown), the first round of the cell cycle was completed ϳ28 h after induction. Moreover, vanadate had no detectable effect on progression through any phase of the cell cycle. It can be concluded that vanadate blocks the induction of differentiation without affecting mitotic clonal expansion and that the blockade is reversible.

Accumulation of a 35-kDa Phosphotyrosyl Protein during Induction of Differentiation in the Presence of Vanadate-Since
vanadate appears to block a PTPase-dependent signaling event required for induction of adipocyte differentiation, the possibility was considered that this blockade might cause accumulation of a phosphotyrosyl intermediate in the signaling pathway. To test this possibility, cell lysates from 3T3-L1 preadipocytes induced to differentiate in the presence or absence of vanadate were subjected to Western blot analysis with anti-phosphotyrosine antibody. As illustrated in Fig. 3, several phosphotyrosyl proteins were detected. The most prominent of the vanadatedependent phosphoproteins, whose level was markedly decreased when exposure to vanadate was delayed for 24 h, was a 35-kDa phosphotyrosyl protein (pp35). This phosphoprotein did not accumulate in cells that had not been induced to differentiate, but that had been treated with vanadate (data not shown), and was particularly prominent in the short ECL exposure shown in Fig. 3C.
Cell fractionation revealed that pp35 is located primarily (Ͼ95%) in the cytosol (Fig. 3B). The slightly faster moving phosphoprotein of ϳ33 kDa evident in Fig. 3B appears to be a proteolytic fragment of pp35 and was not always detected (data not shown). pp35 is not mitogen-activated protein kinase or Rab3, which also have molecular masses in the 35-kDa range, as pp35 was not immunoprecipitated by antibodies directed against either of these proteins (data not shown). A time course study showed that during the normal differentiation process (days 0 -6) in the absence of vanadate, pp35 did not accumulate (Fig. 3D). Moreover, the accumulation of pp35 was transient and present only on days 1 and 2. When added on days 3 and 4 after the hormonal induction, vanadate did not cause the accumulation of pp35 (Fig. 3D), nor did it block differentiation (28). Maximal accumulation of pp35 (in the presence of vanadate with preadipocytes that had been induced to differentiate) occurred between days 1 and 2; by day 3, virtually no pp35 remained. This was likely due to dephosphorylation after the removal of vanadate by the medium change on day 2 because there was a higher amount of the non-phosphorylated form of pp35 (c-Crk) in day 3 cells than in day 2 cells (data not shown).
Dephosphorylation of pp35 by PTPase HA2 in Vitro-The time window during which vanadate is capable of blocking adipocyte differentiation (Fig. 1C) and causing accumulation of pp35 (Fig. 3) coincides with the time of maximal PTPase HA2 expression (28). It was therefore of interest to verify that the PTPase activity in preadipocytes is sufficient to hydrolyze in vitro the maximal amount of pp35 contained in vanadatetreated cells. It should be noted that in previous studies, it was shown that the only PTPase present in preadipocytes at a significant level is PTPase HA2, a homologue of PTPase 1B (33). Thus, pp35 was partially purified and used as substrate for PTPase present in lysates of "induced" 3T3-L1 preadipocytes. As shown in Fig. 4, ϳ10 min was required for 50% hydrolysis of pp35 by a cell equivalent amount of cell lysate PTPase (measured at 30°C). Consistent with PTPase HA2 being the responsible phosphatase activity, this enzymatic activity was inhibited by phosphotyrosine, vanadate, and Zn 2ϩ , with vanadate being the most potent inhibitor (data not shown). Similar results were obtained with pp35 and PTPase HA2 purified from 3T3-L1 preadipocytes by methods described previously, i.e. hydroxylapatite and DEAE-cellulose chromatography (33).
Effect of Expressing c-Crk Antisense RNA on Adipocyte Differentiation-Activation of the IGF-1 receptor by IGF-1 or a high concentration of insulin was previously shown to be involved in the induction of differentiation of 3T3-L1 preadipocytes (8). Since pp35 was identified to be the tyrosine-phosphorylated form of mouse c-Crk (see "Experimental Procedures"), a known substrate of the IGF-1 receptor tyrosine kinase (39,40), it became important to verify that c-Crk plays an essential role in the differentiation process. Thus, the role of c-Crk in this 3T3-L1 preadipocytes (6-cm monolayers) were treated with or without vanadate as described for Fig. 1A. Cell number was determined at different times following the induction of differentiation.

FIG. 3. Analysis of phosphotyrosyl proteins in 3T3-L1 preadipocytes following induction of differentiation in the absence or presence of vanadate.
A, lysates from 3T3-L1 preadipocytes treated with vanadate (ϩVanadate or Vanadate delayed) or not (Control) as described for Fig. 1A were prepared 2 days after induction of differentiation and subjected to Western blot/ECL analysis with anti-phosphotyrosine antibody. The arrow indicates the 35-kDa phosphotyrosyl protein. B, intracellular distribution of pp35 in 3T3-L1 preadipocytes treated or not with vanadate as described for A. Cellular fractionation was performed as described under "Experimental Procedures." C, shorttime ECL exposure of anti-phosphotyrosine antibody Western blots of lysates from cells treated or not with vanadate as described above. D, kinetics of the appearance of pp35 during differentiation of 3T3-L1 preadipocytes in the presence and absence of vanadate added during days 1 and 2 or days 3 and 4. Loading of gels was based on plate equivalents. Membr. extr., membrane extract; Memb. pel., membrane pellet. process was investigated. A 200-base pair fragment of c-Crk cDNA corresponding to the 5Ј-untranslated region and the first 117 base pairs of the coding region (37) was engineered into the pBCMGneo expression vector in either the antisense or sense orientation. Expression of the pBCMGneo vector was shown previously to have no adverse effect on the differentiation of 3T3-L1 preadipocytes (18,28). 3T3-L1 preadipocyte lines stably transfected with the vectors were selected and subjected to the differentiation induction protocol. As shown in Fig. 5, 3T3-L1 preadipocytes transfected with the c-Crk "antisense" vector differentiated very poorly. By reducing the cellular level of c-Crk protein in 3T3-L1 preadipocytes, antisense c-Crk inhibited their differentiation induction. In contrast, preadipocytes transfected with the "sense" vector or an empty vector differentiated normally. The amount of c-Crk protein expressed by cells harboring the antisense vector was markedly reduced compared with the control 3T3-L1 cells (Fig. 6). These results suggest that c-Crk functions in the induction of the adipocyte differentiation program, presumably by mediating the signal from the IGF-1 receptor. Taken together with the vanadate results ( Fig. 1) (28), these findings provide compelling evidence for the involvement of the IGF-1 receptor, c-Crk, and PTPase HA2 in a tyrosine phosphorylation/dephosphorylation signaling cascade in the initiation of the adipocyte differentiation program.
Phosphorylation of c-Crk at Tyrosine Prevents Its Interaction with C3G, a Downstream Signaling Molecule-It has been demonstrated that the SH2 domain of c-Crk can interact intramolecularly with Tyr(P) 221 , which is phosphorylated by IGF-1 receptor tyrosine kinase (39 -42). This interaction might be expected to obscure the SH3 domain of c-Crk because it lies between the SH2 domain and Tyr 221 (39,40). The fact that vanadate, a PTPase inhibitor, causes accumulation of phosphoc-Crk (pp35, presumably at Tyr 221 because it is phosphorylated by IGF-1 receptor kinase in 3T3-L1 preadipocytes) and blocks the induction of differentiation suggested that phosphorylation and dephosphorylation of c-Crk are involved in the induction mechanism essential for signal transduction. Thus, by locking c-Crk in the phosphorylated state with vanadate, the SH3 domain would be unable to interact with its downstream signaling partner, e.g. C3G (43)(44)(45)(46), which is a signaling molecule that possesses an SH3 domain-binding proline-rich motif that has been shown to bind to the SH3 domain in c-Crk.
To test the hypothesis that interaction of c-Crk with C3G occurs during differentiation induction and is prevented by phosphorylation of c-Crk, we determined whether C3G could be coprecipitated with antibody against c-Crk at different times following induction of differentiation when c-Crk was in either the phospho or dephospho state. Fig. 7A shows that both c-Crk and C3G were expressed by 3T3-L1 preadipocytes at an early stage of the differentiation program. Following induction of differentiation, C3G protein was coprecipitated with antibody directed against c-Crk (Fig. 7B). Within 15 min after induction, a substantial amount of C3G was co-immunoprecipitated with anti-Crk antibody. Even after 24 h, there was still some C3G binding to c-Crk. However, in non-induced 3T3-L1 preadipocytes, c-Crk did not associate with C3G, i.e. C3G was not co-immunoprecipitated with c-Crk prior to hormonal induction of differentiation (Fig. 7B). Thus, induction of differentiation led to rapid association of C3G with c-Crk. At this moment, it is not clear whether the activation of c-Crk occurs before phosphorylation of c-Crk or after dephosphorylation of phospho-c-Crk.
Since vanadate causes the accumulation of phospho-c-Crk, and phosphorylation may shield the SH3 domain from C3G (see above), we investigated the effect of vanadate on the interaction of c-Crk, i.e. phospho-c-Crk, with C3G. Immunoprecipitation of c-Crk from lysates of day 1 cells (treated or not with vanadate) revealed virtually no association between phospho-c-Crk and C3G in vanadate-treated cells (Fig. 7C). Significant association of C3G with c-Crk did occur, however, in day 1 cells not treated with vanadate, at which time almost all of the c-Crk was in the dephospho form. Moreover, there was little change in the amounts of either c-Crk or C3G protein per se in cells that were treated or not with vanadate (Fig. 7C). Taken together, these results indicate that C3G associates only with differentiation inducer-activated (presumably through the IGF-1 receptor) c-Crk and does not associate with either phospho-c-Crk or "unactivated" c-Crk. Vanadate treatment prevents the turnover of phospho-c-Crk to c-Crk, thereby preventing downstream signaling molecules from binding to the SH3 domain of Crk.
To prevent c-Crk signaling with vanadate, a significant fraction of c-Crk would have to be locked in the phosphorylated form. As illustrated in Fig. 7C, most of the c-Crk from vanadate-treated preadipocytes migrated more slowly upon SDS-PAGE than c-Crk from control preadipocytes. Since vanadate blocked differentiation only when added during the first 24 h after induction (Fig. 1), we focused our efforts on this time interval even though there was more phospho-c-Crk in day 2 cells (Fig. 3D). To verify that the slow-moving form of c-Crk is indeed phospho-c-Crk, purified phospho-c-Crk (pp35) was treated with alkaline phosphatase. It is clear that the phosphatase treatment completely converted the slow-moving form of c-Crk to the fast-moving form (Fig. 7C). Because phospho-c-Crk was purified by following the phosphotyrosyl group in pp35 with anti-phosphotyrosine antibody, (see "Experimental Procedures"), all of the purified phospho-c-Crk (pp35) must be in the phosphorylated form. We conclude that the majority of c-Crk is locked in the phosphorylated state in vanadate-treated cells.
To function as a mediator for the induction of differentiation, c-Crk must be present in the preadipocyte at the time differentiation is induced, i.e. on day 0 prior to treatment with the hormone inducers. To verify that c-Crk is present at the appropriate time in the differentiation program, the expression of c-Crk was assessed just before (days Ϫ1 and 0) and during (days 1 and 2 after induction) the early phase of the program. As shown in Fig. 8, c-Crk was not expressed by proliferating preadipocytes, and its expression was induced just after the cells became confluent (which occurred on day Ϫ2). In proliferating cells or cells just reaching confluence, c-Crk expression was not detectable by immunofluorescence with anti-Crk antibody (Fig. 8A). c-Crk began to appear at 1 day after confluence (day Ϫ1). After induction, expression of c-Crk was further  (1 day) indicate the times at which cells were harvested. MDI induction refers to MIX, dexamethasone, and insulin induction; IP refers to immunoprecipitation; and IB refers to immunoblotting with antibody against c-Crk or C3G. The Ϫ lane is the 0-min cell lysate. C, c-Crk phosphorylation and protein-protein association. Panel I, whole cell lysate of day 1 3T3-L1 preadipocyte was subjected to the differentiation protocol in the presence (ϩ) or absence (Ϫ) of vanadate, followed by Western immunoblotting with anti-Crk antibody. Panel II, purified phospho-c-Crk protein (pp35) was treated with (ϩ) or without (Ϫ) alkaline phosphatase (AP). An equal amount of purified protein was loaded onto the gel. Anti-Crk antibody (␣-Crk) was used to immunoblot the transferred membrane. Panel III, cell extracts prepared from day 1 3T3-L1 preadipocytes induced in the presence (ϩ) or absence (Ϫ) of vanadate were immunoprecipitated with anti-Crk antibody, and then the immunoprecipitated samples were subjected to Western blotting with anti-C3G antibody (␣-C3G). Panel IV, aliquots of cell extracts used for immunoprecipitation in panel III were subjected directly to Western blotting with anti-C3G antibody.
FIG. 8. Expression of c-Crk occurs after 3T3-L1 preadipocytes achieve confluence. A, immunofluorescence of 3T3-L1 preadipocytes stained with anti-Crk antibody. proliferating refers to 3T3-L1 preadipocytes prior to reaching confluence; confluent refers to 3T3-L1 preadipocytes just reaching confluence. 0h, 2h, 24h, and 48h indicate the time after induction of differentiation. The yellow bar represents 50 m. B, Western blot of c-Crk in 3T3-L1 preadipocyte lysates before induction and in the early stages of differentiation. C refers to 3T3-L1 preadipocytes just reaching confluence. Ϫ1 is 1 day post-confluent; 0 is 2 days post-confluent; and 1 and 2 refer to 1 and 2 days after induction of differentiation, respectively.
increased. Thus, consistent with its putative role as signal mediator, c-Crk was present at the time of exposure of preadipocytes to the differentiation inducers. DISCUSSION When treated with differentiation inducers (on day 0 of the differentiation protocol), confluent growth-arrested 3T3-L1 preadipocytes synchronously reenter the cell cycle, undergo mitotic clonal expansion (days 0 -2), and then coordinately express genes that produce the terminally differentiated adipocyte phenotype (days 3-6) (7,9,47). Immediately following induction, PTPase HA2 (a homologue of PTPase 1B) is expressed, reaching a maximal level during clonal expansion and then declining (28). If vanadate, a potent PTPase inhibitor, is added to the cells during this period of maximal PTPase level, differentiation induction is blocked (28). The time window during which vanadate can block differentiation is relatively short, i.e. between 0 and 20 h following induction (Fig. 1C). After that, exposure to vanadate has no effect on differentiation (Fig. 1C) (28). Thus, vanadate blocks a tyrosine dephosphorylation event required for terminal differentiation, but has no effect on clonal expansion per se (Figs. 1-3). Since the PTPase-catalyzed tyrosine dephosphorylation process appears to be required at an early stage of adipocyte differentiation, the phosphotyrosine protein will be an important intermediate in the induction of adipocyte differentiation. In this investigation, we have identified phosphotyrosyl-c-Crk, a known substrate of IGF-1 receptor kinase (39,40), as the intermediate for the PTPase-catalyzed tyrosine dephosphorylation process during 3T3-L1 preadipocyte differentiation.
That the IGF-1 receptor tyrosine kinase is involved in the induction process was first recognized by Rubin and co-workers (8). IGF-1, rather than insulin, was found to be the physiological inducer of adipocyte differentiation, along with cAMP and glucocorticoid. Because of its low binding affinity for the IGF-1 receptor, insulin can be used to replace IGF-1 as a differentiation inducer only at non-physiologically high concentrations (5,6). In contrast, IGF-1 induces differentiation at in vivo concentrations. It should also be noted that preadipocytes possess numerous IGF-1 receptors, ϳ30,000/cell (48), but only a small number of insulin receptors. The number of insulin receptors begins to increase only after induction of differentiation (49). Thus, activation of the IGF-1 receptor tyrosine kinase by IGF-1 is one of the signals (along with activation of protein kinase A by cAMP and of the glucocorticoid receptor by glucocorticoid) that triggers adipocyte differentiation. Thus, taken together with the findings that vanadate, a potent PTPase inhibitor, blocks hormone (insulin through the IGF-1 receptor)-induced 3T3-L1 preadipocyte differentiation and PTPase HA2-catalyzed dephosphorylation of phospho-c-Crk (pp35) both ex vivo and in vitro (28, Fig. 3 and 4), c-Crk serves both as a substrate of the IGF-1 receptor tyrosine kinase and as a substrate of PTPase HA2. These results support the view that the IGF-1 receptor tyrosine kinase, c-Crk, PTPase HA2, and, most likely, C3G (see below) function in a signaling cascade early in the adipocyte differentiation program.
In view of the apparent role of c-Crk in the signaling cascade, it became important to verify that it is actually required for adipocyte differentiation. Results of the antisense RNA experiments (Figs. 5 and 6) confirmed our observation. As an adapter molecule having both SH2 and SH3 domains, c-Crk has been suggested to be involved in growth factor-mediated tyrosine phosphorylation signaling cascades (50 -52). Based on our findings and those of others, we postulate the following sequence of events summarized in Fig. 9. The interaction of IGF-1 with cell-surface IGF-1 receptors activates receptor autophosphorylation and thereby activation of the receptor tyrosine kinase. c-Crk, which possesses both one SH2 domain and two SH3 domains (only the first of which is functional) (51,52), binds to a phosphotyrosine on the receptor through the SH2 domain. We suggest that the proline-rich motif (amino acids 67-84) (37,42), between the SH2 and functional SH3 domains of c-Crk, may interfere with binding of the SH3 domain by downstream signaling proteins through their proline-rich motif and that binding of the phosphotyrosine of the IGF-1 receptor to the SH2 domain induces a conformational change in c-Crk that renders its SH3 domain-accessible to an SH3 domain-binding signaling protein such as C3G (Fig. 7B) (43)(44)(45). Next, the receptor tyrosine kinase catalyzes the phosphorylation of c-Crk at Tyr 221 (39,40). Studies by Koval et al. (40) showed that the phosphorylation at Tyr 221 by the ligand-activated IGF-1 receptor tyrosine kinase is SH2 domain-dependent. This phosphorylation of c-Crk at tyrosine leads to a competitive intramolecular interaction between the phosphotyrosine and the SH2 domain of phospho-c-Crk (41,42), causing the dissociation of phospho-c-Crk from the receptor. This is also supported by the observation that mutation of Tyr 221 in c-Crk not only prevents phosphorylation of c-Crk, but also increases its association with the IGF-1 receptor (40). A consequence of such an intramolecular interaction is to obscure the SH3 domain that is juxtaposed between the SH2 domain and Tyr 221 (Fig. 7C) (41,42) and to cause the release of the "activated" downstream molecule, e.g. C3G. PTPase HA2 then dephosphorylates phospho-c-Crk to release c-Crk from its internally locked phosphorylated state. Our results support this hypothesis in that dephosphorylated c-Crk interacts with C3G, whereas phosphorylated c-Crk does not (Fig. 7). Because of the rapid turnover of phospho-c-Crk to form c-Crk, little phospho-c-Crk would be detectable in normally induced 3T3-L1 cells. If vanadate is present, however, this turnover is blocked; phospho-c-Crk accumulates; and c-Crk is unavailable for the signal-transmitting phosphorylation/dephosphorylation cascade and consequently blocks the induction of differentiation. At present, the signaling molecule(s) lying downstream is unknown. It should also be noted that SH3 domain-binding signaling molecules, other than C3G, might be involved.