The Constitutively Active Mutant Gα13Transforms Mouse Fibroblast Cells Deficient in Insulin-like Growth Factor-I Receptor*

Insulin-like growth factor-I (IGF-I) receptor plays an important role in normal cell cycle progression and tumor growth, and it is thought to be essential for cellular transformation. To test this hypothesis, we stably transfected a GTPase-deficient mutant human Gα13, which is highly oncogenic when overexpressed in vitro, into R− fibroblasts derived from IGF-I receptor-deficient mice. Northern blots of multiple clones revealed the expression of a 1.8-kilobase pair mutant Gα13 transcript in transfected cells, in addition to the 6-kilobase pair endogenous mRNA. The transfection resulted in a doubling of the expression of Gα13 protein in these cells as assessed by Western blot analysis. The transforming ability of the mutant Gα13 was tested using the soft agar assay. Nontransfected R− cells cultured with 10% fetal bovine serum failed to form colonies after 3 weeks. Most of the mutant Gα13-expressing clones formed significant numbers of colonies (11–50 colonies/1000 cells plated). Overexpression of the IGF-I receptor enabled R− cells to form colonies (27 colonies), and co-transfection of the mutant Gα13 caused a further increase in colony formation (117–153 colonies) in three of five clones analyzed. Apparently Gα13 works through pathways other than mitogen-activated protein kinase and c-Jun N-terminal kinase in transforming R− cells, because their activities were not significantly altered by the mutant Gα13 expression. These results demonstrate that Gα13 can induce cellular transformation through pathways apparently independent of the IGF-I receptor and that activation of the IGF-I receptor signaling pathways, although not essential for the transforming phenotype, enhances the effect of other pathways.

The insulin-like growth factor-I (IGF-I) 1 receptor, a trans-membrane heterotetrameric tyrosine kinase, is expressed by most normal and transformed cells (1). Upon ligand binding, IGF-I receptor undergoes autophosphorylation on intracellular tyrosine residues and activation of its intrinsic tyrosine kinase. Mice without the IGF-I receptor gene invariably die of respiratory failure at birth and exhibit severe intrauterine growth retardation, general organ hypoplasia, and delayed ossification, demonstrating the essential role of the IGF-I receptor in normal growth and development (1)(2)(3). Overexpression of the IGF-I receptor can transform mouse fibroblasts (4). The presence of the IGF-I receptor was found to be necessary for proteins such as SV40 large T antigen, Ha-Ras, EWS/FLI-1 (a fusion protein produced by Ewing's family of tumors), v-Src, and bovine papilloma virus to transform target cells (5)(6)(7)(8). Furthermore, decreasing the number of IGF-I receptors by antisense RNA technology causes a reversal of the transforming phenotype, including arrest of in vivo tumor growth (7), suggesting that the IGF-I receptor is also essential for oncogene-induced cellular transformation.
Heterotrimeric (␣, ␤, and ␥ subunits) guanine nucleotidebinding proteins (G proteins) couple to hundreds of different receptors in mammals and are central to the signaling processes of multicellular organisms (9). The ␣ subunits of G proteins belong to a large group of GTPases, which are classified into four sub-families based on amino acid homology: G ␣s , G ␣i/o , G ␣q/11 , and G ␣12/13 (10). Stimulatory (G s ) and inhibitory (G i ) G protein subtypes have been implicated in the regulation of adenylyl cyclase and the gating of certain ion channels (11). The G q family of proteins couple to the phospholipase C pathway (12). The ubiquitously expressed G ␣12/13 proteins are involved in mitogenesis and transformation probably through activation of c-Jun N-terminal kinase (JNK) and activation or inhibition of mitogen-activated protein kinase (MAPK) pathways (13)(14)(15)(16)(17)(18). Overexpression of wild type G ␣12/13 and the expression of GTPase-deficient active mutants of G ␣12/13 transform fibroblast cell lines (15, 18 -22).
To test the hypothesis that IGF-I receptor is essential for cellular transformation, we stably transfected a GTPase-deficient mutant human G ␣13 (Q226L) into RϪ fibroblasts derived from IGF-I receptor-deficient mice. These cells cannot be transformed by SV40 large T antigen and other oncoproteins (5)(6)(7)(8).
In contrast, we found that activated human G ␣13 (hG ␣13) is able to transform RϪ cells independent of the IGF-I receptor and that overexpression of both G ␣13 and IGF-I receptor caused synergistic enhancement in cellular transformation.

EXPERIMENTAL PROCEDURES
Mutant Human G ␣13 Expression Vector-A GTPase-deficient mutation was introduced into hG ␣13 cDNA (Ref. 14; GenBank accession number L22075) by polymerase chain reactions. The codon corresponding to glutamine (Q) at residue 226 was mutated to express leucine (L) and thereby generated a new restriction site Bsu361. The mutant cDNA (BglII-EcoRI) was then subcloned into the vector pcDNAIIIHA (Invitrogen, Carlsbad, CA). The resulting construct pcDNAIIIHAG13QL(A) (276) contains the mutant hG ␣13 cDNA (1.5 kb) downstream of a HA tag (derived from influenza protein hemagglutinin) sequence, driven by the human cytomegalovirus promoter, and genes whose products confer resistance to neomycin and ampicillin. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Northern Blot Analysis-The expression of the mutant hG ␣13 was evaluated by Northern blot analysis (25). The specific bands of expected sizes were illustrated, together with ethidium bromide staining of ribosomal RNAs, demonstrating equal loading of total RNA on the gel.
Western Blot Analysis-The expression of the G ␣13 protein was further confirmed by Western blot analysis using standard procedures (26) and rabbit anti-G ␣13 antiserum (1:1000, Calbiochem, San Diego, CA), which is specific to an 11-amino acid (LHDNLKQLMLQ) C-terminal region of hG ␣13 . The level of MAPK activity was determined by using phospho-specific MAPK and p44/42 MAPK antibodies (New England Biolabs, Beverly, MA). Rabbit polyclonal Crk-L (C20) and mouse monoclonal JNK1 (F3) antibodies (Santa Cruz Biotech, Santa Cruz, CA) were also employed in Western blots.
Colony Formation in Semisolid Media-For assays of anchorageindependent growth, suspensions of 1000 cells in 1 ml of DMEM with 10% FBS, 20 mM HEPES, pH 7.5, and 0.2% agarose were lain onto 1.5 ml of 0.4% agarose-containing medium in 35-mm tissue culture dishes (27). Colonies Ͼ0.05 mm in diameter were scored after 3 weeks of incubation at 37 C, 5% CO 2 and humidified (90%) atmosphere. Three dishes were plated each time for each clone.
IGF-I Receptor Binding Assay-To determine the level of IGF-I receptor expression, displacement binding of 125 I-IGF-I was done as described previously (28). The Scatchard analysis was performed using Scapre v4.42 and MacScafit v4.94 computer programs (National Institutes of Health, Bethesda, MD).
In Vitro Kinase Assay-Confluent cells in 100-mm diameter dishes was incubated with serum-free medium for 20 h and stimulated with 10% FBS for 10 min. Whole cell lysate (1 mg of protein) was immunoprecipitated with 1 g of JNK1 (C17) antibody and 2.5% protein A-Sepharose (Santa Cruz) overnight at 4°C, washed twice in ice-cold RIPA buffer (phosphate-buffered saline containing 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM NaV 3 O 4 , 3% aprotinin, and 1 mM phenylmethylsulfonyl fluoride), and washed twice with kinase buffer (50 mM HEPES, pH 7.5, 100 mM NaCl, 5 mM MgCl 2 , 0.1% Tween 20, 0.1% bovine serum albumin, 0.1 mM NaV 3 O 4 , and 0.1 mM phenylmethylsulfonyl fluoride). The beads were resuspended in 100 l of kinase buffer containing 3 g of GST-Jun (Santa Cruz), 5 M ATP, and 5 Ci of [␥-32 P]ATP (NEN Life Science Products). The reactions were terminated after 15 min at room temperature by adding 5 ϫ Laemmli buffer (29). Samples were separated by 9% SDS-PAGE and blotted onto Protran membrane (Scheicher & Schuell). After blocking with 3% bovine serum albumin, the blot was first exposed to a Kodak X-Omat film and then blotted with mouse monoclonal JNK1 antibody (Santa Cruz) to determine the amount of JNK1 present in the reaction.
Statistical Analysis-Unpaired t test was performed with InStat 2.03 computer program (GraphPad Software, San Diago, CA)

RESULTS
Characterization of the Cell Lines-Three cell lines were used in transformation studies. Previously, IGF-I receptor gene-deficient RϪ cells have been demonstrated to not express IGF-I receptor using immunoprecipitation, IGF-I-cross-linking, and RNase protection assay, in contrast to wild type mouse fibroblasts (5,6). Indeed, RϪ cells demonstrate a total lack of binding to 125 I-IGF-I as determined by Scatchard analysis (data not shown). W6ϩ cells that overexpress IGF-I receptors demonstrate a level of IGF-I receptor expression (730,000 receptors/cell, K d of 0.16 nM) comparable with that of NWTb3 (550,000 receptors/cell, K d of 0.55 nM), which are NIH 3T3 fibroblasts overexpressing the IGF-I receptors (30).
Expression of Human Mutant G ␣13 in IGF-I Receptor-defi-cient RϪ Cells-To study the role of the IGF-I receptor in G ␣13 -induced cellular transformation, we transfected a constitutive active mutant hG ␣13 into RϪ cells. All G418 resistance transfected RϪ clones express a 1.8-kb mRNA hybridized to an antisense hG ␣13 riboprobe (Fig. 1, upper panel). This band is absent in untransfected cells, despite equal amounts of total RNA loaded as shown by staining of the rRNA bands (middle panel). A 6-kb endogenous mouse G ␣13 mRNA has been observed in all clones (data not shown). At the protein level, G ␣13 (wild type and mutant forms) are present mainly as a 40-kDa band (lower panel), which is increased 2-fold upon transfection in all the clones selected (clones 5-26), as analyzed by densitometry. G ␣13 Is Capable of Transforming IGF-I Receptor-deficient RϪ Cells-G ␣13 transfection caused no morphological change in RϪ cells. The effect of the constitutively active mutant hG ␣13 on cellular transformation was tested using the soft agar assay (Table I). In three experiments each with triplicate dishes, 5 of 6 mutant hG ␣13 -expressing clones (except 12) formed significant numbers of colonies (11-50 colonies/1000 cells plated, p Ͻ 0.05) as compared with untransfected RϪ cells, which formed only 1-3 tiny colonies as shown in Fig. 2. Cells expressing mutant hG ␣13 protein formed either significant number albeit small colonies (clones 17 and 20) or many colonies comparable in size with the ones seen in NWTb3 control cells (clones 7 and 26) (Table I and Fig. 2). This result suggests that hG ␣13 can transform RϪ cells in the absence of the IGF-I receptor.
Expression of Human Mutant G ␣13 in IGF-I Receptor-overexpressing W6ϩ Cells-To determine if overexpression of IGF-I receptor can further enhance the transforming ability of hG ␣13 , W6ϩ cells were transfected and analyzed for the expression of hG ␣13 using Northern and Western blot analysis (Fig. 3). Abundant hG ␣13 mRNA is expressed in all the clones transfected with the construct. In transfected W6ϩ cells, hG ␣13 mRNA is present in two molecular forms (2.  protein) with anti-G ␣13 antibody. A 40-kDa band is present in all cells, and the density of it is doubled following G ␣13 transfection (densitometric analysis, data not shown). The same blot was subsequently stripped and reblotted with an unrelated Crk-L antibody, which demonstrates equal loading of total protein (data not shown).
tion-The transforming ability of the transfected IGF-I receptor and hG ␣13 was tested using the soft agar assay. As shown in Table I and Fig. 2, overexpression of the IGF-I receptor enabled RϪ cells to form colonies (27 versus 2, p Ͻ 0.001). Transfection of W6ϩ cells with hG ␣13 caused no phenotypical change except a further increase in colony formation (117 to 153, p Ͻ 0.05) in 3 of 5 clones analyzed (although without further increase in the colony size formed by W6ϩ cells, Fig. 2). This effect is much greater than simple addition of the effects of hG ␣13 and IGF-I receptor alone, suggesting a synergism between them.
G ␣13 -induced Cellular Transformation Does Not Work through MAPK or JNK-In an attempt to identify signaling pathways involved in G ␣13 -activated cellular transformation, MAPK and JNK activities were analyzed. As shown in Fig. 4A, a 10-min incubation with 10% FBS dramatically induced MAPK phosphorylation in both untransfected RϪ and W6ϩ cells. Although basal level of MAPK phosphorylation was elevated to various degree in clones that form colonies (RG7, RG20, WG18, and WG21), transfection of G ␣13 caused no obvious change in serum-activated MAPK activity. In Fig. 4B, JNK activity was studied by in vitro GST-Jun phosphorylation. As in the case of MAPK, serum incubation-activated JNK activity dramatically in both untransfected cells. With 10% serum, G ␣13 transfection caused no significant change in maximum JNK activation. Only in one clone (RG20, of six tested) G ␣13 increased basal JNK activity. Therefore, we conclude that G ␣13induced cellular transformation is apparently not accompanied by changes in MAPK and JNK activities. DISCUSSION In this study, we have confirmed that IGF-I receptor-deficient RϪ fibroblasts are incapable of forming colonies in soft agar and that active mutant G ␣13 is highly oncogenic. The unique finding is that G ␣13 is able to transform fibroblasts independent of IGF-I receptor.
The role of G ␣12/13 , a novel class of G␣ proteins, in signal transduction is under intense investigation. The G ␣12/13 proteins are known to be coupled to two receptors: thyrotropin receptors in the thyroid gland, which stimulate adenylyl cyclase and phospholipase C pathways (31), and the thrombin receptor in platelets, which stimulates platelet aggregation (32). It is relatively clear that constitutively active G ␣12/13 activates Ras, which recruits the small G protein Rac, and leads to activation of JNK/stress-activated protein kinase (SAPK) and enhanced transcriptional activity of c-Jun (14,16). The effect of G ␣12/13 on MAPK activity is not yet fully elucidated. Although the G ␣12/13 proteins do not activate MAPK directly, there have been reports that these activated G proteins can enhance epidermal growth factor (EGF)-induced MAPK activation in Rat-1 cells (18) or inhibit EGF-stimulated extracellular signal-regulated kinase expression in COS-7 cells (17). In Swiss 3T3 cells, G ␣12/13 regulate Rho-dependent responses in actin polymerization (28). Constitutively active mutants of G ␣12/13 have also been known to activate the Na ϩ /H ϩ exchanger (22,33) and to potentiate serum-stimulated phospholipase A 2 activity but do not stimulate inositol phospholipid hydrolysis (19).
Recent reports by Baserga and co-workers (5, 7, 8, 23, 34 -36) have demonstrated that the IGF-I receptor is required for the establishment and maintenance of the transformed phenotype in vivo and in vitro. Indeed, oncogenes including SV40 T antigen and Ha-Ras and overexpressed EGF receptors all failed to transform RϪ cells (5, 8, 23, 34, 35); the resistance of RϪ cells  to transformation can be abolished by expressing IGF-I receptors (5,23). Furthermore, a decrease in the number of IGF-I receptors caused a reversal of the transformed phenotype (7). C6 rat glioblastoma cells failed to grow into tumors in syngeneic rats when antisense IGF-I receptor mRNA is expressed (36). Our finding that RϪ cells can be transformed by G ␣13 in contrast to other oncogenes suggests that IGF-I receptor is not required for all the transformation processes. Nevertheless, G ␣13 is not the only signaling protein capable of transforming RϪ cells. When transfected with constitutively active Ras or Ras plus the SV40 T antigen, RϪ cells were partially transformed (23). Most recently, Valentinis et al. (37) demonstrated that oncoprotein v-Src can also bypass the requirement for a functional IGF-I receptor in transforming mouse fibroblasts. However, c-Src was unable to do so. Taken together, the results of these studies suggested that G ␣13 acts independently of the IGF-I receptor in cellular transformation.
Another interesting observation from this study is that G ␣13 and the overexpressed IGF-I receptor act in concert in transforming RϪ cells. This is a very similar scenario with the reported synergism between G ␣12 and c-Raf-1 in transforming NIH 3T3 cells (38). In that case, the synergism seemed to occur between two parallel pathways: MAPK activated by c-Raf-1 and JNK/SAPK activated by G ␣12 . The IGF-I receptor is known to signal through the Ras-Raf-MAPK cascade of serine/threonine and tyrosine kinases. Therefore, our observation of similar synergism may well be explained by cooperation between a G ␣13 -activated pathway and overexpressed IGF-I receptor-induced MAPK activation. Analysis of MAPK and JNK activities upon serum stimulation revealed no change caused by mutant G ␣13 expression, although in some clones their basal activities were elevated. Because the cell transformation was studied using 10% serum and the increase in basal kinase activities does not correlate with transforming ability of each clone, we propose that these two pathways are not apparently involved in G ␣13 -induced cellular transformation. Investigation is underway to explore the specific targets activated by G ␣13 in RϪ cells.
In summary, we have found that 1) IGF-I receptor-deficient RϪ fibroblasts did not form colonies in soft agar assay; 2) oncogenically active mutant hG ␣13 was able to transform the RϪ cells; 3) overexpression of the IGF-I receptor also enabled RϪ cells to form colonies; 4) the effects on transformation of hG ␣13 and the overexpressed IGF-I receptor were synergistic; and 5) G ␣13 -induced cellular transformation does not apparently work through MAPK or JNK pathways. Therefore, G ␣13 induces cellular transformation through pathways apparently independent of the IGF-I receptor. Furthermore, whereas IGF-I receptor is not essential for the transforming phenotype, it can enhance the effect of G ␣13 -activated pathways. FIG. 4. Effect of G ␣13 transfection on MAPK and JNK activities in R؊ and W6؉ cells. A, Western blotting of phosphorylated MAPK isoforms. Whole cell lysates (50 g) from cells incubated with or without serum, as indicated, were separated by 9% SDS-PAGE and transferred onto nitrocellulose membranes. Blots were incubated with a phosphospecific antibody that recognizes phosphorylated tyrosine 204 on both p44 and p42 forms of MAPK and developed using standard procedures. The same blots were subsequently stripped and reblotted with polyclonal p44/42 MAPK antibody to confirm equal amount of total MAPK present in each lane by densitometry (data not shown). B, in vitro GST-Jun phosphorylation induced by JNK1 immunoprecipitated from cells untreated or treated with serum. JNK1 immunoprecipitates were washed, and the pellets were incubated in an in vitro JNK assay as described under "Experimental Procedures." Samples were separated by 9% SDS-PAGE, transferred onto a nitrocellulose membrane, and exposed to a x-ray film. [ 32 P]Phosphorylated GST-Jun is illustrated. Equal loading of JNK1 was confirmed by blotting with a monoclonal JNK1 antibody (data not shown). Both panels are representative of at least two experiments.