Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Scaffolding protein Gab2 mediates fibroblast transformation by the SEA tyrosine kinase


Transformation of fibroblasts by V-SEA involves activation of the ERK and phosphatidylinositol 3-kinase (PI3K) pathways. Effector proteins that are key mediators of the ERK and PI3K pathways, namely Grb2, the tyrosine phosphatase, SHP2 and PI3K, interact with the two phosphotyrosines found in the bidentate motif in the carboxy-terminal region of V-SEA. Genetic analysis demonstrated that while Y557 was a primary binding site and thus activator of the PI3K-Akt pathway, Y564 also contributed to the activation of this pathway. Y564 was located within a Grb2-binding motif, this raised the possibility that a protein that associated with Grb2 might be important for this PI3K activation. The scaffolding proteins Gab1 and/or Gab2 were candidates for this role. In this report, we demonstrate that V-SEA preferentially interacts with Gab2. Furthermore by using Gab2 null fibroblasts, we demonstrate that Gab2 is essential for fibroblast transformation by V-SEA. Using mutant forms of Gab2, we show that activation of the PI3K-Akt pathway via Gab2 is required for V-SEA-induced transformation. However, efficient fibroblast transformation also requires the SHP2 interaction site on Gab2.


Oncogenes derived from cellular receptor tyrosine kinases (RTKs) are frequently constitutively activated by structural changes such as point mutation, truncation or rearrangement. The V-Sea oncogene of the avian retrovirus S13 is one such oncogene and it is derived from a cellular protein tyrosine kinase, C-SEA, whose extracellular and transmembrane domains have ‘been replaced by viral envelope sequences (Smith et al., 1989). Previous studies showed that the V-SEA tyrosine kinase activity is necessary for transformation by the V-SEA protein and that cell surface localization, oligomerization by viral envelope domains and autophosphorylation are all necessary for the transforming activity of the V-SEA protein (Crowe and Hayman, 1991; Crowe and Hayman, 1993a, 1993b; Morimoto and Hayman, 1994).

The SEA protein is a member of the receptor subfamily that contains the MET and STK/RON receptors. The STK gene was isolated from early mouse hematopoietic cells (Wang et al., 1995, 1994; Suda et al., 1997) and RON from human cells (Ronsin et al., 1993). The identification of the macrophage-stimulating protein (MSP), as the ligand for SEA, STK and RON (Wahl et al., 1999), together with their sequence similarities, has led to the hypothesis that they represent the avian, murine and human versions of the same receptor (Wahl et al., 1999). All members of this family have a bidentate motif at the carboxy-terminus of the receptor comprising two tandemly arranged tyrosine residues, which are phosphorylated upon receptor activation. These residues then serve as SH2- or phosphotyrosine binding, PTB-domain binding sites. Thus this motif functions as a multisubstrate interaction site. In the MET protein, it has been demonstrated that the tyrosine closest to the carboxy-terminus within this motif, tyrosine 1356, plays a major role in signaling by this receptor (Ponzetto et al., 1994; Zhu et al., 1994). Although this motif is conserved in SEA, the sequences surrounding the tyrosines are quite different from those in MET (Smith et al., 1989). However, this motif is also essential to signaling by SEA since mutation of both of these tyrosines (Y557 and Y564) to phenylalanine residues completely abolished SEA's ability to transform fibroblasts (Park and Hayman, 1999).

Analysis of protein–protein interactions involving these two phosphorylated tyrosines in V-SEA demonstrated that Grb2, PI3K and SHP2 can all bind to this motif (Park and Hayman, 1999; Agazie et al., 2002;). Functional analysis of these interactions demonstrated that Y557 primarily mediates the activation of PI3 K and to a lesser degree ERK1/2; and vice versa for Y564. Furthermore, when the PI3K-Akt and the Ras-ERK signaling pathways were inhibited, it was determined that activation of these pathways was essential for V-SEA-induced transformation of Rat-1 and NIH-3T3 cells (Agazie et al., 2002).

The Grb2-associated-binding protein 2, Gab2 (Gu et al., 1998), was tyrosine phosphorylated in V-SEA-transformed cells (Agazie et al., 2002). Gab2 interacted with V-SEA and PI3 K in vivo and mediated formation of a multimeric signaling complex that involved V-SEA, Gab2 and PI3 K (Agazie et al., 2002). The interaction of the Gab2 protein with V-SEA and its subsequent tyrosine phosphorylation and association with PI3K presumably leads to the activation of the PI3K pathway. Since PI3K can also be activated via direct interaction with V-SEA through the Y557 residue, the involvement of Gab2 in PI3K activation would on the face of it appear to be redundant. Similarly, the tyrosine phosphatase SHP2, a known mediator of ERK activation, can interact both with V-SEA and with tyrosine phosphorylated Gab2. Therefore, we decided to investigate the role of Gab2 in transformation by V-SEA. In this report, we used mouse embryo fibroblasts that are null for the expression of Gab2 to demonstrate that Gab2 is an essential mediator of cell transformation by V-SEA.


Involvement of Gab2 in signaling by the V-Sea oncogene

To extend our analysis of the role of Gab2 in signaling by V-Sea, we examined NIH-3T3 cells transformed by V-SEA, for tyrosine phosphorylation of Gab2 and the association of other signaling molecules. NIH-3T3 cells were infected with murine retroviruses expressing V-SEA, RON or empty vector. Cell lysates were prepared and immune precipitated with normal rabbit serum or antibodies against either Gab1 or Gab2. Figure 1a shows that in V-SEA-transformed NIH-3T3 cells there is a large increase in the tyrosine phosphorylation of Gab2 when compared to vector infected NIH-3T3 cells. In contrast, there is hardly any change in the level of Gab1 tyrosine phosphorylation. When the cells expressing the Ron receptor were activated with MSP and analysed, it was clear that in these cells Gab1 is preferentially tyrosine phosphorylated in comparison to Gab2. This preference was seen regardless of the time of activation of Ron (data not shown). Reprobing of this blot with antibodies against Gab1 or Gab2 revealed that equivalent amounts of Gab proteins were precipitated from all three cell types, Figure 1a, middle and lower panels.

Figure 1

V-SEA primarily induces tyrosine phosphorylation of Gab2. Lysates were prepared from NIH-3T3, V-SEA-transformed NIH-3T3 or MSP-treated Ron-expressing NIH-3T3 cells. Equal aliquots of lysates were immunoprecipitated with normal rabbit sera (NRS), anti-Gab2 or anti-Gab1 sera and the immune precipitated proteins run on an SDS polyacrylamide gel and analysed by Western blotting. (a) Western blot probed with antiphosphotyrosine monoclonal antibody (top). The blot was then stripped and reblotted with antibodies against Gab 1 (middle) and Gab2 (lower). (b) Same blot stripped and reprobed with antisera against the p85 subunit of PI3K, equal amounts of total cell lysates are also shown to demonstrate that the levels of PI3K were equivalent in the different lysates. (c) Same blot stripped and reprobed with antibodies against SHP2

Since Gab2 is an adaptor protein, its tyrosine phosphorylation should lead to the association of several signaling molecules. Therefore, we reprobed the above Western blot filters for the presence of PI3K and SHP2. As can be seen, expression of V-SEA in the NIH-3T3 cells leads to the increased association of the p85 PI3K subunit with Gab2, whereas there was no detectable increase in its association with Gab1 (Figure 1b). In contrast, activation of RON did result in a very small increased association of p85 with Gab1. Analysis of the association of SHP2 with Gab2 also revealed that expression of V-SEA increased the association of these two proteins and in this case there was also an increase in the association of SHP2 with Gab1 in the V-SEA expressing cells (Figure 1c). Comparison of NIH-3T3 cells with those expressing activated RON showed that there was an increase of SHP2 association with Gab1 but no major change in its association with Gab2. Analysis of total cell lysates for the expression of the p85 subunit of PI3K and SHP2 indicated that all the cells expressed equivalent levels of these two proteins (Figure 1b and c). These data indicate that the activated receptors induced the phosphorylation of the Gab adaptor proteins on tyrosine leading to the association of the PI3K and SHP2 proteins. Furthermore, V-Sea seemed to prefer to signal to Gab2, whereas Ron preferentially coupled with Gab1.

Gab2 mediates transformation by V-SEA

To examine the functional significance of the induction of tyrosine phosphorylation of Gab2 by V-SEA, we determined if V-SEA, could induce transformation of Gab2−/− mouse embryo fibroblasts that had been immortalized by multiple passages in tissue culture. The first parameter of transformation we used was growth and contact inhibition. We measured the growth of V-SEA expressing Gab2−/− cells in comparison to vector alone Gab2−/− cells over a 4-day period. As can be seen in Figure 2a, there was no difference in the growth of these two cell types and they reached confluence and contact inhibition after 3 days in culture. We introduced Gab2 back into the cells and determined if this affected their growth. Figure 2a shows that introduction of Gab2 back into the Gab2−/− cells had no effect on their growth. In contrast, when Gab2 was introduced into the V-SEA expressing cells, the cells grew to much higher densities indicating they were not as contact inhibited as the other cells. These data demonstrate that Gab2 is necessary for the increased cell densities that are associated with V-SEA cell transformation.

Figure 2

Transformation of Gab2−/− fibroblasts by V-SEA. Gab2−/− cells were infected with murine retroviruses expressing either Gab2 or V-SEA and assayed for changes in growth parameters. (a) Shows growth curves over 4 days of the different cells. (b) A representative focus assay of the cells and in addition of Ras-infected cells. Similar data were obtained in three separate experiments (c) Western blots of cell lysates taken from the cells used in (b) showing the levels of V-SEA-in the V-SEA infected cells, right two lanes compared to vector-infected cells, left two lanes (upper panel): and RAS in the Ras-infected cells, right two lanes, again compared to vector-infected cells, left two lanes (lower panel)

We extended this analysis to look at another parameter of cell transformation, namely focus formation. Figure 2b shows the results of a representative focus assay using Gab2−/− cells infected with a murine retrovirus expressing V-SEA or vector alone infected cells (similar results were seen in three experiments). As a positive control, we introduced an activated H-RAS(V12), RAS, oncogene into the Gab2−/− cells. Vector alone infected Gab2−/− cells did not form any foci, (Figure 2b, plate 1), and infection of the Gab2−/− cells with V-SEA gave rise to very few transformed foci (Figure 2b, plate 3). In contrast, introduction of RAS into the cells did transform the cells as judged by their ability to produce large numbers of foci (Figure 2b, plate 5). This indicates that V-SEA cannot readily transform Gab2−/− cells. To determine if this was a direct consequence of the lack of expression of Gab2, we reintroduced Gab2 back into the Gab2−/− cells and repeated the experiment. As can be seen, introduction of Gab2 into Gab2−/− cells did not give rise to any foci (Figure 2b plate 2). However, cells expressing both V-SEA and Gab2 became transformed and produced large numbers of foci, similar in number to what we saw using the RAS oncogene (Figure 2b, plate 4). In contrast, introduction of Gab2 did not increase the RAS transformation of the cells (Figure 2b, plate 6). These data indicate that Gab2 expression is important for transformation of fibroblasts by V-SEA, whereas it is not important for RAS transformation.

We analysed the cells used in the above focus assays for the levels of RAS, V-SEA and Gab2 to make sure that we were expressing these proteins from the retroviruses as expected. Figure 2c shows that the expression of V-SEA was equivalent in both the poorly transformed Gab2−/− cells and the transformed Gab2−/− cells that also expressed Gab2 from the retroviral construct. Gab2 was readily detected in the cells infected with the Gab2 retrovirus, whereas as expected it was undetectable in the Gab2−/− cells (data not shown and see Figure 4). Ras levels were also similar in the Gab2−/− cells and the Gab2−/− + Gab2-expressing cells, Figure 2c. Thus it appears that the ability of V-SEA to transform the Gab2−/− cells requires the expression of Gab2.

Figure 4

Interaction of the Gab2 mutants with PI3K and SHP2. Lysates were prepared from the Gab2−/− cells that were infected with viruses expressing the various Gab2 mutants and then, where indicated, superinfected with viruses expressing V-SEA. (a) Western blots were prepared with antisera against the M45-tag on SEA and the HA-tag on Gab2 to demonstrate the cells that were infected. Immune precipitates were prepared from the above lysates using antisera against Gab2 and then probed with: (b) antibodies against the HA-tag on Gab2, (c) antibodies against phosphotyrosine; (d) antisera against PI3K and (e) antibodies against SHP2

Activation of PI3K via Gab2 is important for transformation by V-SEA

Previous studies have demonstrated that V-SEA transformation requires the activation of the PI3K pathway as well as the Erk MAP kinase pathway (Agazie et al., 2002). Since the association of PI3K and SHP2 with Gab2 can influence these two pathways, we looked at their activation by V-SEA in the presence and absence of Gab2. As a control, we used the RAS-infected cells. We performed western blots on lysates from the different cells using antisera that were specific for the activated forms of either AKT (as a downstream readout of PI3K activation) or ERK kinase. Figure 3a shows that introduction of V-SEA into Gab2−/− cells gives rise to a small increase in the levels of activated AKT, pAKT. Presumably, this reflects the level of activation of the PI3K pathway via direct interaction with the V-SEA protein as shown previously (Agazie et al., 2002). However when Gab2 is also introduced into the cells there is a large increase in the level of pAKT, indicating that Gab2 is an important adaptor protein that links V-SEA with PI3K activation. The Gab2−/− cells had a significant basal level of ERK activation, even when grown in the absence of serum for several hours. Introduction of either Gab2 alone or together with V-SEA had no major affect on this level of activation. In contrast, introduction of RAS did cause increased levels of both pAKT and pERK irrespective of the presence of Gab2 (see Figure 3a). Reprobing of this blot with antisera against AKT and ERK2 shows that all cells express equivalent levels of these proteins, Figure 3a.

Figure 3

Activation of AKT and ERK. (a) Lysates were prepared from the cells characterized in Figure 2 and analysed by Western blotting using antisera specific for the activated forms of AKT and ERK (top). Reprobing with antisera against ERK or AKT showed equal expression (lower). (b) Gab2−/− cells and Gab2−/− cells infected with the Gab2-expressing retrovirus were probed with antibodies against Gab1 to show the level of Gab1 expression. (c) The same cells used in (b) were treated with 100 ng/ml FGF-1 for 10 min and lysates were analysed by Western blotting using antisera specific for the activated forms of AKT. Reprobing of the blot demonstrates equal levels of AKT in all lysates

Since V-SEA activation of Akt was weak in the Gab2−/− cells, it was of interest to determine if the Gab2−/− cells were expressing either of the other two Gab proteins, Gab1 and Gab3. Using both monoclonal and polyclonal antisera against Gab3 kindly provided by Dr L Rohrschneider, we were unable to detect clearly expression of this protein (data not shown). In contrast, the Gab2−/− cells did express Gab1 and the level of Gab1 was unaffected by the introduction of wt Gab2, Figure 3b. To demonstrate that the Gab1 in the Gab2−/− cells was functional, we looked at the ability of FGF to activate the PI3K pathway by Western blotting for pAKT. Activation of PI3K by FGF has been shown previously to utilize Gab1 (Ong et al., 2001). As shown in Figure 3c, AKT is strongly activated by FGF and furthermore the activation is independent of the presence of Gab2. Reprobing of the gel with antisera against AKT shows that all cells express equivalent levels of protein, Figure 3c. These data clearly demonstrate that although Gab2−/− cells express functional levels of Gab1, V-SEA cannot utilize this adaptor protein for transformation.

Gab2 is phosphorylated on tyrosine and the phosphorylated residues serve as binding sites for several effector molecules. The sites for SHP2 and PI3K binding have been identified. Therefore, to determine if the activation of signaling pathways involving either SHP2 or PI3K are important for V-SEA fibroblast transformation, we decided to introduce mutated versions of Gab2 back into the Gab2−/− cells to see if they could also complement V-SEA transforming ability. These were Gab2-3YF, in which all three tyrosines for the binding of the p85 subunit of PI3K have been converted into phenylalanine residues: Gab2-DM, which no longer has the two tyrosines that mediate SHP2 binding; and Gab2-ΔPH, in which the pleckstrin homology domain has been deleted. The Gab2−/− cells were infected with murine retroviruses expressing HA-tagged forms of Gab2 and then subsequently infected with the M45-tagged V-SEA expressing retrovirus. The cells were assayed initially for the expression of V-SEA and Gab2 by Western blot using a mixture of M45 and HA monoclonal antibodies. Figure 4a shows that the infected Gab2−/− cells expressed high levels of wt-, 3YF-, and Gab2-DM and lower levels of Gab2-ΔPH in both the V-SEA infected and uninfected cells. V-SEA was expressed at equivalent levels in the cells (Figure 4a).

To check whether the mutant forms of Gab2 were behaving as predicted, we assayed the association of the Gab2 protein with either PI3K or SHP-2 by immune precipitating Gab2 and assaying for the association of these proteins by Western blot. Figure 4b shows that equivalent amounts of Gab2 were precipitated from the cells regardless of the expression of V-SEA. However, reprobing of this blot with antibodies against phosphotyrosine demonstrated that Gab2 was only tyrosine phosphorylated if V-SEA was expressed in the cells, Figure 4c. The level of Gab2-3YF tyrosine phosphorylation was lower than that of wt- and Gab2-DM, which in part reflects the loss of three major tyrosine phosphorylation sites. To determine if PI3K was associated with Gab2, the blots were probed with antisera against the p85 subunit of PI3K. As can be seen, wt- and Gab2-DM efficiently coprecipitated PI3K only if V-SEA was expressed in the cells. The Gab2-ΔPH coprecipitated a small amount of PI3K and as expected the Gab2-3YF did not coprecipitate any PI3K. The blots were then probed for the presence of SHP-2, Figure 4e. Although only weakly tyrosine phosphorylated, the Gab2-3YF molecules were associated with SHP-2, as was wt and Gab2-ΔPH, and this increased association was dependent on the presence of V-SEA. In contrast, the Gab2-DM protein only very weakly coprecipitated SHP-2 in the presence of V-SEA and this level of association was marginally higher than that seen in the cells that did not express V-SEA.

The association of PI3K with Gab2 should lead to activation of AKT. To determine this directly, we performed Western blots on the various cells to determine pAKT levels. As can be seen in Figure 5, only wt-Gab2 and the Gab2-DM gave rise to a V-SEA-induced increase in pAKT levels. Reprobing of the gel indicated that the expression levels of AKT were similar in all cases. As shown previously, the Gab2−/− cells have a high basal level of activated ERK, pERK, and the expression of Gab2 or V-SEA had no demonstrable effect on this level.

Figure 5

Activation of AKT and ERK. Lysates were prepared from the cells characterized in Figure 4 and analysed by Western blotting using antisera specific for the activated forms of AKT and ERK. The blot was reprobed with antisera against AKT to demonstrate equal levels of expression among the different lysates

The above data are in agreement with the predictions of the consequences of the mutations and allow us to determine the biological consequences of activation of the PI3K by V-SEA via the Gab2 molecule. Figure 6 shows a representative focus assay of cells that have been infected with wt-Gab2, Gab2-3YF, Gab2-DM, Gab2-ΔPH, or all of the above plus V-SEA. (Similar results were obtained in three separate experiments.) As can be seen introduction of wt-Gab2 rescued V-SEA's ability to transform the cells efficiently. The cells infected with the Gab2 mutant that can still activate PI3K, Gab2-DM, also gave rise to foci, albeit to a significantly lesser extent than cells expressing wt-Gab2. In contrast, there were no foci formed when the two other forms of Gab2 were used. This was to be expected in the case of the Gab2-ΔPH protein, since previous analysis had demonstrated that deletion of the PH domain from any of the Gab proteins rendered them physiologically inactive. However, given that the levels of Gab2-ΔPH were significantly lower than the other mutant this conclusion has to be tempered to a degree. The inability of the Gab2-3YF protein to rescue efficient focus formation by V-SEA-indicates that the interaction of SHP2 with Gab2 is not sufficient for V-SEA induced cell transformation. However the Gab2-DM mutant, which lacked SHP2 association, was only able to give rise to foci with 5–15% of the efficiency seen with wt-Gab2 (Figure 6). Therefore, it is possible that while SHP2 association with Gab2 may not be sufficient for V-SEA-induced transformation, it could play a role in the efficient transformation associated with V-SEA. These data indicate that PI3K activation via Gab2 is important for V-SEA transformation.

Figure 6

The Gab2 mutant that can bind PI3K can rescue transformation of Gab2−/− cells by V-SEA. Gab2−/− cells were infected with viruses expressing the various Gab2 mutants and then where indicated superinfected with viruses expressing V-SEA and the resultant cells assayed for transformation in the focus assay as described in Materials and methods


Previous analysis of the transformation of fibroblasts by V-SEA had determined that activation of both the ERK and PI3K pathways was important (Agazie et al., 2002). Furthermore, it was determined that the two tyrosines in the bidentate motif in the carboxy-terminal region of V-SEA were important for signaling and transformation (Park and Hayman, 1999; Agazie et al., 2002). Effector proteins that are key mediators of the ERK and PI3K pathways, namely Grb2, SHP2 and PI3K, interact with the phosphotyrosines found within this bidentate motif. Therefore, it was reasonable to assume that these direct interactions should be sufficient to activate these two pathways sufficiently to cause transformation. However, mutational analysis of the two tyrosine residues within the bidentate motif demonstrated that while Y557 was a primary binding site and activator of the PI3K-Akt pathway, Y564 also contributed to the activation of this pathway (Park and Hayman, 1999; Agazie et al., 2002). Since Y564 was located within a Grb2-binding motif, this raised the possibility that a Grb2-associated protein was important for this PI3K activation and Gab2 was a candidate for this role. In this report, we demonstrate that the activation of the PI3K-Akt pathway by V-SEA occurs via Gab2 and this activation is necessary for transformation.

Our analysis of the involvement of proteins in signaling by V-SEA determined that V-SEA interacts more efficiently with Gab2 than with Gab1. This specificity was also seen in erythroid cells as well as the fibroblasts transformed by V-SEA (data not shown). At the present time it is not clear what determines this preference. V-SEA is a member of the tyrosine kinase family that includes MET, and Gab1 has been shown to interact with MET via a Met-binding domain (MBD), in addition to its ability to interact via Grb2. However, there is no evidence for a similar Grb2-independent binding of Gab2 with SEA or with MET (Lock et al., 2002). Gab1 is the mediator of several important biological effects induced by MET both in vitro and in vivo (Weidner et al., 1996; Bardelli et al., 1997; Nguyen et al., 1997; Ingham et al., 1998; Itoh et al., 2000; Sachs et al., 2000; Schaeper et al., 2000) and analysis of MET signaling shows that MET preferentially interacts with Gab1 rather than Gab2 (Schaeper et al., 2000; Lock et al., 2002). The MBD found in Gab1 seems to be responsible for this difference, but a mutant form of Gab2 that contained the MBD of Gab1 could not substitute for Gab1 function even though it interacted with MET as efficiently as Gab1. These results suggest that Gab1 and Gab2 were not redundant in MET signaling (Lock et al., 2002). These in vitro studies demonstrate that these two highly related scaffolding proteins have distinct roles in signaling and are in accord with genetic analysis in vivo (Gu et al., 1998; 2001; Itoh et al., 2000; Sachs et al., 2000). In the case of the interaction of V-SEA with the Gab proteins, further studies will be required to address the underlying reason for this specificity. Nevertheless, it is clear that V-SEA preferentially induces the tyrosine phosphorylation of Gab2 and this gives rise to the association of the effector molecules SHP2 and PI3K. This association leads to the activation of these two effector molecules, and in the case of PI3K, this is clearly important for fibroblast transformation by V-SEA.

V-SEA has two tyrosine residues in its bidentate motif that become phosphorylated. One of these is a perfect consensus site for PI3K association and the p85 PI3K subunit has been shown to bind to this site (Park and Hayman, 1999; Agazie et al., 2002). Furthermore, this association leads to activation of the PI3K pathway. Therefore, it is interesting that activation of the PI3K pathway via Gab2 is essential for fibroblast transformation by V-SEA, since one might have thought that activation via the direct interaction of V-SEA with PI3K would have been sufficient. It is not clear at this time why activation of PI3K via these two interactions seems to lead to different outcomes. It is possible that activation via interaction with V-SEA does not overcome some threshold of activation needed for transformation, since the activation via Gab2 is more robust than that from the V-SEA–PI3K interaction. Alternatively, it is possible that while both the signaling complexes formed with PI3K are membrane located they may be in different intracellular locations, which may lead to different signaling outcomes. It is also not clear which of the many downstream effectors of PI3K are important for V-SEA transformation (Cantley, 2002). In this regard, it is interesting to note that the transformation of chicken embryo fibroblasts by V-SEA was not inhibited by rapamycin, which inhibits the downstream target of PI3K, mTor. Rapamycin did effectively block oncogenic transformation induced by Akt (Aoki et al., 2001; Vogt, 2001). Further studies will be necessary to resolve these issues.

Previous analysis of signaling pathways has determined that Gab2 is a key intermediate in signaling via several different receptors (Liu and Rohrschneider, 2002). It has been shown to mediate differentiation signals downstream of the Fms receptor that are important for macrophage differentiation (Liu et al., 2001). In this case, it was the interaction of Gab2 with SHP2 and the subsequent ERK activation and inhibition of proliferation that were important. In our analysis, there was no indication that ERK activation via Gab2 was essential for V-SEA transformation. However, this conclusion is compromised by the fact that the Gab2−/− cells have a relatively high basal level of ERK signaling, and thus if the activation of this pathway via Gab2 was weak we may not have been able to measure it above the basal level. Nevertheless, it should be noted that activation of ERK by RAS was clearly discernible in these cells, so any contribution via Gab2 in V-SEA signaling cannot be that robust. Furthermore, activation of mast cells from Gab2−/− mice demonstrated that the absence of Gab2 had no effect on ERK activation, whereas it did on PI3K activation (Gu et al., 2001). This indicates that Gab2 is not always an important mediator of ERK activation. However it is important to note that the Gab2-DM mutant, which could no longer interact with SHP2, was only 5–15% as efficient as wt-Gab2 in the development of V-SEA-induced foci. This indicates that SHP2 activation via Gab2, while insufficient in its own right to cause transformation, may play a role in the efficient transformation of fibroblasts by V-SEA.

Gab2 can clearly function as an important mediator and relayer of signals by various receptors and nonreceptor tyrosine kinases as reviewed in Liu and Rohrschneider (2002). Interestingly these interactions lead to different responses depending on the cells under study or the receptors being activated. In some cases, Gab2 relayed signals affect differentiation and lead to a reduction in cell proliferation, for example, FMS signaling in macrophages (Liu et al., 2001). In other cases, Gab2 mediates inhibitory signals, for example, via PI3K in T-cell receptor signaling (Pratt et al., 2000). Recently, it has also been identified as an important mediator in transformation by the BCR-ABL oncogene (Sattler et al., 2002). Clearly, this protein is a multifunctional adaptor protein that can play various roles in signal transduction. The analysis reported here identifies yet another role for Gab2, this time in mediating cell transformation by the V-sea oncogene.

Materials and methods

Cells, cell culture, antibodies, etc

The cell types used in this study were NIH-3T3, Ron-NIH-3T3 and V-SEA-NIH-3T3, which have been described previously (Agazie et al., 2002). Primary mouse embryo fibroblasts isolated from E13 Gab2−/− embryos and upon passage became immortalized. Analysis of these immortalized cells revealed they no longer expressed functional p53 (data not shown). All cell lines were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (Fcs) at 37°C with 7% CO2. Anti-phospho- or pan- (Akt or ERK2) was purchased from Cell Signaling Laboratories, and protein-A or protein-G sepharose beads were from Pharmacia. Antibodies against PI3K and Gab2 were from Upstate Biotechnology and the monoclonal antibody against SHP2 was from Transduction Laboratories. Polyclonal and monoclonal antibodies against Gab3 (Seiffert et al., 2003) were kindly provided by Dr L Rohrschneider, Fred Hutchison Cancer Center, Seattle, WA, USA. The M45 and HA monoclonal antibodies were a kind gifts from Drs P Hearing and D Bar-Sagi, Stony Brook, NY, respectively, and the 4G10 anti-phosphotyrosine antibody was a kind gift from Dr D Morrison, NCI, MD, USA.

Immunoprecipitation and immunoblotting

All immunoprecipitation experiments were performed by overnight incubation at 4°C with an additional incubation for 2-h after addition of sepharose beads. Unless otherwise specified, immunocomplexes captured on sepharose beads were washed three times with cell lysis buffer and bound proteins eluted by boiling with Laemmli sample buffer and then run on 10% denaturing polyacrylamide gel. After transfer onto nitrocellulose, blocking was performed by incubation with 3% bovine serum albumin. Primary antibody staining was carried out overnight at 4°C or for 2 h at room temperature and secondary antibody staining was for 30 min at room temperature. In all cases, the chemiluminescence method was used for detecting bands. The same gel separation and detection procedures were used for the analysis of total cell lysates.

Generation of the V-SEA-expressing cells

The V-SEA-expressing cells were produced by infecting cells with a murine retrovirus expressing V-SEA. Briefly, V-SEA was inserted into a retroviral vector termed REBNA/IRES/GFP at the NotI and XhoI sites and transfected into packaging cells using the Lipofectamine (Invitrogen) transfection reagent. After 48 h of incubation, 2 μg/ml puromycin was added to the cells to remove untransfected cells, since this viral vector contains the puromycin resistance gene. After 48 h under puromycin, resistant cells were split and incubated in fresh medium until confluent at which time the supernatant containing the virus was harvested and used to infect NIH-3T3 cells in the presence of 5 μg/ml polybrene (Sigma). Titration of the virus on NIH-3T3 cells gave an MOI of 98% per ml per 105 cells in a 3.5 cm plate as determined by fluorescent activated cell sorter and as well by examining for green fluorescence under a fluorescent microscope. This population of cells was termed V-SEA-3T3 and used for the indicated experiments. Gab2−/− cells were superinfected with the same V-SEA-expressing retrovirus and selected in a similar manner.

Gab2-expressing retroviruses were generated by cloning Not1 to HindIII fragments containing the HA-tagged Gab2 genes (WT, DM (Gu et al., 1998) and 3YF and Gab2-ΔPH, (Gu et al., 2000)) into a modified version of the murine retrovirus pLPC (Serrano et al., 1997). The original pLPC vector was modified to include a puromycin resistance marker.

Focus-Formation assay

To assess the ability of the cells for focus formation, 1.5 × 104 virus infected cells were mixed with 105 uninfected cells and plated in duplicate onto 6 cm dishes. Cells were maintained in DMEM supplemented with 5% FCS and the growth media were changed every 3 days. After 10 days, the colonies were stained with Giemsa and photographed.

Proliferation assay

Fibroblasts (2 × 105) were plated at equal density into 3 cm plates in duplicate. Cell number was assessed daily following trypsinization and counting in a Coulter counter.


  1. Agazie Y, Ischenko I and Hayman M . (2002). Oncogene, 21, 697–707.

  2. Aoki M, Blazek E and Vogt PK . (2001). Proc. Natl. Acad. Sci. USA, 98, 136–141.

  3. Bardelli A, Longati P, Gramaglia D, Stella MC and Comoglio PM . (1997). Oncogene, 15, 3103–3111.

  4. Cantley LC . (2002). Science, 296, 1655–1657.

  5. Crowe AJ and Hayman MJ . (1991). J. Virol., 65, 2533–2538.

  6. Crowe AJ and Hayman MJ . (1993a). Cell. Growth Differ., 4, 403–410.

  7. Crowe AJ and Hayman MJ . (1993b). Oncogene, 8, 181–189.

  8. Gu H, Maeda H, Moon JJ, Lord JD, Yoakim M and Nelson BH Neel BG . (2000). Mol. Cell. Biol., 20, 7109–7120.

  9. Gu H, Pratt JC, Burakoff and SJ Neel BG . (1998). Mol. Cell., 2, 729–740.

  10. Gu H, Saito K, Klaman LD, Shen J, Fleming T, Wang Y, Pratt JC, Lin G, Lim B, Kinet JP and Neel BG . (2001). Nature, 412, 186–190.

  11. Ingham RJ, Holgado-Madruga M, Siu C, Wong AJ and Gold MR . (1998). J. Biol. Chem., 273, 30630–30637.

  12. Itoh M, Yoshida Y, Nishida K, Narimatsu M, Hibi M and Hirano T . (2000). Mol. Cell. Biol., 20, 3695–3704.

  13. Liu Y, Jenkins B, Shin JL and Rohrschneider LR . (2001). Mol. Cell. Biol., 21, 3047–3056.

  14. Liu Y and Rohrschneider LR . (2002). FEBS Lett., 515, 1–7.

  15. Lock LS, Maroun CR, Naujokas MA and Park M . (2002). Mol. Biol. Cell., 13, 2132–2146.

  16. Morimoto AM and Hayman MJ . (1994). J. Virol., 68, 1837–1842.

  17. Nguyen L, Holgado-Madruga M, Maroun C, Fixman ED, Kamikura D, Fournier T, Charest A, Tremblay ML, Wong AJ and Park M . (1997). J. Biol. Chem., 272, 20811–20819.

  18. Ong SH, Hadari YR, Gotoh N, Guy GR, Schlessinger J and Lax I . (2001). Proc. Natl. Acad. Sci. USA, 98, 6074–6079.

  19. Park CY and Hayman MJ . (1999). J. Biol. Chem., 274, 7583–7590.

  20. Ponzetto C, Bardelli A, Zhen Z, Maina F, dalla Zonca P, Giordano S, Graziani A, Panayotou G and Comoglio PM . (1994). Cell, 77, 261–271.

  21. Pratt JC, Igras VE, Maeda H, Baksh S, Gelfand EW, Burakoff SJ, Neel BG and Gu H . (2000). J. Immunol., 165, 4158–4163.

  22. Ronsin C, Muscatelli F, Mattei MG and Breathnach R . (1993). Oncogene, 8, 1195–1202.

  23. Sachs M, Brohmann H, Zechner D, Muller T, Hulsken J, Walther I, Schaeper U, Birchmeier C and Birchmeier W . (2000). J. Cell. Biol., 150, 1375–1384.

  24. Sattler M, Mohi MG, Pride YB, Quinnan LR, Malouf NA, Podar K, Gesbert F, Iwasaki H, Li S, Van Etten RA, Gu H, Griffin JD and Neel BG . (2002). Cancer Cell., 1, 479–492.

  25. Schaeper U, Gehring NH, Fuchs KP, Sachs M, Kempkes B and Birchmeier W . (2000). J. Cell. Biol., 149, 1419–1432.

  26. Seiffert M, Custodio JM, Wolf I, Harkey M, Liu Y, Blattman JN, Greenberg PD and Rohrschneider LR . (2003). Mol. Cell. Biol., 23, 2415–2424.

  27. Serrano M, Lin AW, McCurrach ME, Beach D and Lowe SW . (1997). Cell, 88, 593–602.

  28. Smith DR, Vogt PK and Hayman MJ . (1989). Proc. Natl. Acad. Sci. USA, 86, 5291–5295.

  29. Suda T, Iwama A, Hashiyama M, Sakano S and Ohno M . (1997). Leukemia, 11 (Suppl 3), 451–453.

  30. Vogt PK . (2001). Trend Mol. Med., 7, 482–484.

  31. Wahl RC, Hsu RY, Huff JL, Jelinek MA, Chen K, Courchesne P, Patterson SD, Parsons JT and Welcher AA . (1999). J. Biol. Chem., 274, 26361–26368.

  32. Wang MH, Iwama A, Skeel A, Suda T and Leonard EJ . (1995). Proc. Natl. Acad. Sci. USA, 92, 3933–3937.

  33. Wang MH, Ronsin C, Gesnel MC, Coupey L, Skeel A, Leonard EJ and Breathnach R . (1994). Science, 266, 117–119.

  34. Weidner KM, Di Cesare S, Sachs M, Brinkmann V, Behrens J and Birchmeier W . (1996). Nature, 384, 173–176.

  35. Zhu H, Naujokas MA, Fixman ED, Torossian K and Park M . (1994). J. Biol. Chem., 269, 29943–29948.

Download references


We thank Drs D Bar-Sagi, P Hearing, D Morrison, B Neel and L Rohrschneider for kindly providing the reagents that made this work possible. We also thank the members of our laboratory for their helpful advice and constructive criticisms. This work was supported by NIH Public Service Grant CA28146 and the Carol Baldwin Breast Cancer Award to MJH, and American Association of Cancer Research-Susan G Komen Breast Cancer Foundation Career Development Award to HG.

Author information



Corresponding author

Correspondence to Michael J Hayman.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ischenko, I., Petrenko, O., Gu, H. et al. Scaffolding protein Gab2 mediates fibroblast transformation by the SEA tyrosine kinase. Oncogene 22, 6311–6318 (2003).

Download citation


  • V-SEA
  • Gab2
  • transformation

Further reading


Quick links