If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by a Medical Research Council project grant (to M. J. W.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We have observed previously the co-immunoprecipitation of the p85 subunit of phosphatidylinositol-3 kinase (PI3K) and SHP2 in murine lymphohemopoietic cells after stimulation with interleukin-3. We have investigated this interaction in more detail and now report the identification of a potentially novel 100-kDa protein (termed p100), which is inducibly phosphorylated on tyrosine after interleukin-3 treatment and which co-immunoprecipitates with both p85 PI3K and SHP2. The Src homology region 2 domains of both p85 and SHP2 appear to mediate their interactions with p100. Sequential precipitation analyses suggest that these interactions are direct and do not involve Grb2, and that the same p100 protein, or a portion of it, interacts with both p85 and SHP2, implying that p100 may serve to link these two proteins. Far Western blotting with both full-length p85 and isolated p85 Src homology region 2 domains supports this view. Interestingly, p100 also appears to be a substrate for the SHP2 phosphatase activity. In addition, p100 is precipitated by Grb2-glutathione S-transferase fusion proteins, an interaction largely mediated by the Grb2 SH3 domains. p100 appears to be distinct from JAK2, Vav, STAT5, and c-Cbl. Although largely cytosolic, p100 can be detected associated with SHP2 and PI3K in crude membrane fractions after interleukin-3 stimulation. We propose that p100 plays a role as an adaptor molecule, linking PI3K and SHP2 in IL-3 signaling.
). However, IL-3 activates a number of signaling pathways known to be controlled by tyrosine phosphorylation and dephosphorylation events. These include the p21 ras /mitogen-activated protein kinase pathway (
)). Activation of these enzymes occurs after ligation of a broad range of receptors and leads to phosphorylation of the inositol ring at the D3 position, resulting in the transient in vivo production of phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-trisphosphate, which are thought to act as novel second messengers (reviewed in Ref.
). SHP2 has two SH2 domains that allow it to interact with tyrosine-phosphorylated proteins, and it has been reported that ligation of SHP2 SH2 domains contributes to activation of its phosphatase activity (
). SHP2 is itself tyrosine-phosphorylated in response to cell stimulation by a number of different growth factors, including PDGF, EGF, Steelfactor, erythropoietin, IL-3, and GM-CSF, which also lead to its activation (
), suggesting SHP2 tyrosine phosphorylation is not essential for its activation. In addition to its catalytic activity, it has been suggested that SHP2 may perform an additional function as an adaptor molecule. Tyrosine-phosphorylated SHP2 has been shown to interact with the small adaptor protein Grb2, via the Grb2 SH2 domain (
). In this role, SHP2 acts as a positive mediator of signal transduction, as it serves to recruit the Grb2-Sos complex to the activated PDGF receptor, facilitating activation of the p21 ras /mitogen-activated protein kinase cascade (
). We report here that the major tyrosine-phosphorylated protein that co-precipitates with both PI3K (p85) and SHP2 in murine cells is a protein of 100 kDa. The same p100 protein, or a portion of it, appears to interact directly with both SHP2 and PI3K and can be dephosphorylated by SHP2. We show that p100 is not JAK2, STAT5, Vav, or c-Cbl and that, although p100 is mainly cytosolic, it can be detected at the membrane in association with PI3K and SHP2 after IL-3 stimulation. Therefore, we propose that p100 is a novel adaptor protein that serves to link PI3K and SHP2 in IL-3 signaling.
We have reported previously that both PI3K and SHP2 are activated in response to IL-3 (
). No such association has been observed for PI3K; thus, the mechanism used by IL-3 to locate PI3K to the membrane, and so to its lipid substrates, remains undefined but is of obvious importance. In light of the fact that we have previously reported co-immunoprecipitation of p85 and SHP2 from extracts of IL-3-treated cells (
), it is possible that PI3K is coupled to the IL-3 receptor via SHP2-IL-3 receptor β-chain interactions. Therefore, we were interested to further investigate the interaction between PI3K and SHP2 in an attempt to identify any potential adaptor proteins involved in linking them after IL-3 treatment.
A 100-kDa phosphotyrosine-containing Protein Co-precipitates with SHP2 and p85
To investigate the major tyrosine-phosphorylated proteins interacting with SHP2 and p85 in response to IL-3, murine Ba/F3 cells were treated with IL-3 or left untreated as a control. Immunoprecipitates were prepared using either anti-p85 or anti-SHP2 antibodies, and immunoblotting was performed with the anti-phosphotyrosine antibody 4G10. The major co-precipitating species in the IL-3-treated samples in both p85 and SHP2 immunoprecipitates was a tyrosine-phosphorylated protein of 100 kDa (p100; see Fig.1A). Additional proteins of 60 and 70 kDa were also observed in p85 precipitates. As we have described previously, the p70 protein was SHP2 (
). Tyrosine phosphorylation of the 60-kDa protein was variably observed in unstimulated samples. Additional proteins of 70 and 135 kDa were observed in the SHP2 immunoprecipitates from IL-3-treated samples. SHP2 corresponds to the 70-kDa protein, and the 135-kDa protein is the β-chain of the IL-3 receptor (
). These results demonstrate that a 100-kDa phosphotyrosine-containing protein is the major protein that co-precipitates with p85 PI3K and SHP2 after IL-3 treatment of Ba/F3 cells. Similar results were obtained when FD-6 cells were subjected to the same analyses (data not shown). It is interesting to note that we do not observe co-immunoprecipitation of tyrosine-phosphorylated p145SHIP with SHP2 from lysates of IL-3-treated Ba/F3 or FD-6 cells. This is in agreement with our previous work (
), who describe co-precipitation of p145SHIP and SHP2 after IL-3 stimulation of murine B6SUtA1 cells, the reasons for these differences are not clear.
p100 Binds Directly to the SH2 Domains of p85
The interaction between p100 and p85 was investigated further to establish whether binding was direct and, if so, which region of p85 was involved. Ba/F3 were treated with IL-3 or left untreated as a control, and precipitations were performed using GST fusion proteins of full-length p85 (FLp85-GST), the N-terminal SH2 domain of p85 (NSH2-GST), the C-terminal SH2 domain of p85 (CSH2-GST), or the p85 SH3 domain (p85SH3-GST), the results are shown in Fig. 1B. A 100-kDa phosphotyrosine-containing protein was precipitated from extracts of IL-3-treated cells with FLp85-GST, NSH2-GST and CSH2-GST, but not with p85SH3-GST. These results suggest that the interaction between p100 and p85 PI3K is mediated by the p85 SH2 domains. A small amount of SHP2 was also precipitated by these fusion proteins, as we have described previously (
), but the relative stoichiometry suggests that the interaction is unlikely to be direct, there being far more p100 present. Similar results were observed in FD-6 cells (data not shown). Interestingly, the p85SH3-GST fusion protein precipitated a 64-kDa protein from IL-3-treated cell extracts, the identity of which is unknown, but may be related to the p60 protein observed in anti-p85 immunoprecipitates (see Fig. 1A).
The identity of the 145-kDa protein precipitated by the p85 SH2-domain GST fusions is not known. One candidate is the recently described SH2-containing inositol phosphatase p145SHIP. To investigate the possibility that p145 could be SHIP, the immunoblot in Fig. 1B was reprobed with anti-p145SHIPantibodies. These antibodies detected SHIP in the pre-immunoprecipitation samples, but they did not cross-react with any proteins co-precipitating with the p85-GST fusion proteins, suggesting that the p145 protein is not SHIP (data not shown). p145 is not observed in p85 immunoprecipitations (see Fig. 1A) using a number of different p85 antibodies,
). It may therefore represent a non-physiological association.
The interaction between p100 and p85 was analyzed further using the technique of Far Western blotting, whereby a recombinant protein is used in place of the primary antibody when probing an immunoblot. This technique is a useful way of detecting interactions that occur directly between particular proteins. Anti-p85 immunoprecipitates were prepared from control and IL-3-treated Ba/F3 cells and immunoblotted in triplicate with 4G10, to assess tyrosine phosphorylation (Fig.1A) or with NSH2-GST or FLp85-GST (Fig. 1C). Both of these p85 fusion proteins bound to a species of 100 kDa on the Far Western blot (see Fig. 1C), in both whole cell extracts (Pre IP samples) and p85 immunoprecipitates from IL-3-treated cells. This supports our previous data and suggests that a direct interaction occurs between the SH2 domains of p85 and tyrosine-phosphorylated p100. The NSH2-GST, CSH2-GST, and FLp85-GST fusion proteins also bound a 100-kDa protein on Far Western blots of whole cell extracts from IL-3-treated FD-6 cells (data not shown). No other species were detected in these Far Western blots; hence, it is unlikely that the association of the p85-GST fusion proteins with the p145 protein observed in Fig. 1B is of high affinity or direct.
The Same p100 Interacts with Both PI3K and SHP2
We were interested to determine whether p100, or a portion of it, which we observed in SHP2 precipitates, was the same as the p100 protein, which we have demonstrated above to interact directly with the SH2 domains of p85. To investigate this, sequential immunoprecipitation analyses were performed using FD-6 cells. Anti-p85 antibodies were used to prepare a primary immunoprecipitate that was re-precipitated using the SHP2SH2-GST fusion protein. Fig.2A shows the results of immunoblotting these two sets of precipitates with 4G10. The SHP2SH2-GST fusion protein reprecipitated a portion of the p100 protein that had initially co-precipitated with the p85 antibodies. Similar results were obtained from Ba/F3 cells (data not shown). These data suggest that the SHP2 SH2 domains can bind directly to p100 and that p85 PI3K interacts with a portion of the same p100 protein that interacts with SHP2. The sequential precipitations were repeated using SHP2 antibodies to prepare the primary precipitates, which were then re-precipitated using NSH2-GST (Fig. 2B) or CSH2-GST (Fig.2C). It can be seen that both of the p85 SH2 domain fusion proteins are able to reprecipitate p100 that originally co-precipitated with SHP2. These data support the conclusion that SHP2 and p85 bind to the same p100 protein and indicate that p100 can interact with either of the SH2 domains of p85. In addition, on Far Western blots, NSH2-GST and FLp85-GST bind to p100 co-precipitated with SHP2 antibodies (data not shown).
p100 Is a Substrate for the SHP2 Phosphatase Activity
We were interested to investigate the possibility that p100 was a substrate for SHP2 phosphatase activity. Control or IL-3-treated Ba/F3 cells were extracted in IP buffer containing or lacking the phosphatase inhibitors sodium orthovanadate, sodium molybdate and sodium fluoride and SHP2 immunoprecipitated in each case. After extensive washing the precipitates were resuspended in phosphatase buffer, again containing or lacking phosphatase inhibitors and incubated at 20 °C for 50 min. To specifically examine the effects of SHP2 phosphatase activity on SHP2-associated proteins, we analyzed the proteins that remained bound in the SHP2 precipitates by immunoblotting with 4G10 (Fig.3, upper panel). In the absence of phosphatase inhibitors, there was a significant dephosphorylation of all of the tyrosine-phosphorylated proteins that co-immunoprecipitated with SHP2. This suggests that p100 and p135 (Aic2A) are substrates for SHP2 and also that SHP2 can auto-dephosphorylate. The evidence that p135 is a substrate for SHP2 is supported by previous work indicating that phosphopeptides based on the sequence of the IL-3 receptor β chain surrounding tyrosine 612 can act as substrates for the phosphatase activity of SHP2 (
). Reprobing the blot with anti-SHP2 antibodies demonstrated equivalent amounts of SHP2 were present in the immunoprecipitation samples (Fig. 3, lower panel). In the presence of phosphatase inhibitors the SHP2 precipitated from IL-3-treated cells is significantly broadened, reflecting its phosphorylation. This is decreased in the samples incubated in the absence of phosphatase inhibitors, correlating with the observed SHP2 dephosphorylation.
p100 Associates with Grb2
The results presented above are consistent with the notion that p100 is an adaptor molecule playing a role similar to that of IRS-1/IRS-2 or the recently discovered Gab-1 or DOS proteins (
). We were therefore interested to investigate whether p100 would bind to other signaling molecules implicated previously in IL-3 signaling, such as the small adaptor protein, Grb2.
Control and IL-3-treated Ba/F3 or FD-6 cell extracts were subjected to precipitations using GST fusion proteins of full-length Grb2 (FLGrb2-GST), the SH2 domain of Grb2 (Grb2SH2-GST), or full-length Grb2 with mutations in both SH3 domains (Grb2N+CSH3*-GST). Precipitates were analyzed using immunoblotting with the 4G10 antibody. The results for Ba/F3 cells are shown in Fig. 4A; the same results were also observed for FD-6 cell extracts (data not shown). A tyrosine-phosphorylated protein of 100 kDa precipitated with FLGrb2-GST but not with either of the other GST fusions. This suggests that the SH3 domains of Grb2 are capable of interacting with a 100-kDa protein.
To investigate this interaction further and determine if the p100 protein precipitated by the FLGrb2-GST fusion protein, or a portion of it, was the same p100 protein that had been observed to bind to SHP2 and p85, sequential precipitations were performed on control or IL-3-treated Ba/F3 cells. Anti-p85 (Fig. 4B) or anti-SHP2 (Fig. 4C) antibodies were used for the primary precipitations, while secondary precipitations were performed using either FLGrb2-GST or Grb2SH2-GST. The primary and secondary precipitations were then analyzed by immunoblotting with the 4G10 antibody.
It can clearly be seen that the FLGrb2-GST reprecipitated p100 that had first been precipitated by either p85 or SHP2 antibodies (see Fig. 4,B and C). This was not seen for Grb2SH2-GST, although this protein did reprecipitate SHP2 that had been immunoprecipitated by anti-SHP2 antibodies (Fig. 4C). On overexposure of the blot shown in Fig. 4C, a very small amount of p100 was seen to reprecipitate with Grb2SH2-GST. These data suggest that Grb2 interacts directly with p100 via its SH3 domains, with possibly some engagement of the Grb2 SH2 domain after stimulation of the cells with IL-3.
p85 (PI3K) and Grb2 Do Not Interact in These Cells
Although the evidence presented so far strongly suggests that p100 is mediating the interaction between p85 and SHP2, the possibility remained that the interaction could also be mediated by the small adaptor protein Grb2 (
). To investigate whether p85 and Grb2 interact directly in our cells, the blot shown in Fig. 4B (upper panel) was reprobed with a monoclonal antibody against p85α (Fig. 4B, lower panel). It can be seen that p85 is visible in the preimmunoprecipitation samples and greatly enriched in the primary precipitation samples, but is not reprecipitated by either FLGrb2-GST or Grb2SH2-GST. In contrast, it can be seen in Fig.4C that SHP2 precipitated using anti-SHP2 antibodies reprecipitated with both full-length Grb2-GST and Grb2SH2-GST. It is thus highly unlikely that Grb2 is mediating the interaction between p85 and SHP2 in IL-3-dependent murine hemopoietic cells.
p100 Is Not JAK2, Vav, or STAT5
A number of proteins of similar molecular weight to p100 have been reported previously to be tyrosine-phosphorylated in response to IL-3. These include JAK2, STAT5, and the proto-oncogene product Vav (
) to run as a sharp band at around 120 kDa. This is in contrast to p100, which migrates as a rather diffuse band at around 100 kDa. To investigate the possibility of p100 being STAT5 or Vav, murine Ba/F3 or FD-6 cells were treated with IL-3 or left untreated as a control. Immunoprecipitates were prepared using anti-SHP2, anti-Vav, or anti-STAT5 antibodies, and immunoblotting was performed using the anti-phosphotyrosine antibody 4G10. Fig. 5shows the data for Ba/F3 cells. No tyrosine-phosphorylated proteins precipitated with the anti-Vav antibody. The anti-STAT5 antibody precipitated a heavily tyrosine-phosphorylated protein of around 95 kDa that migrated slightly faster than the p100 protein that co-precipitates with SHP2. To confirm that p100 is not STAT5, this same immunoblot was stripped and reprobed with anti-STAT5 (Fig. 5,lower panel). STAT5 is clearly visible in the preimmunoprecipitation and anti-STAT5 immunoprecipitation samples but does not co-precipitate with SHP2. In addition, p100 precipitated with the anti-SHP2 antibodies did not react with anti-Vav antibodies upon immunoblotting (data not shown). The same results were seen for analyses using FD-6 cells (data not shown).
p100 Is Not c-Cbl
One other molecule of a similar molecular weight that is a candidate for p100 is the proto-oncogene product c-Cbl. Although the exact function of c-Cbl is not clear, it has been suggested that it may play an important role in recruiting PI3K to receptor complexes at the plasma membrane and hence bring it within close proximity of its substrates. c-Cbl has been shown to be a major tyrosine-phosphorylated protein that co-precipitates with PI3K in response to ligation of both T- and B-cell antigen receptors and engagement of the FcγR receptor on macrophages (
To investigate the possibility of a role for c-Cbl in PI3K-SHP2 interactions, FD-6 were stimulated with IL-3 for various times, immunoprecipitates prepared using a polyclonal antibody against c-Cbl and immunoblotted using the 4G10 antibody (Fig.6A, upper panel). A low basal level of tyrosine phosphorylation of c-Cbl was observed in unstimulated cells, which was not significantly elevated in response to IL-3. Reprobing with the anti-c-Cbl antibody verified the immunoblots were evenly loaded (Fig. 6A, middle panel). Reprobing the same immunoblot with anti-p85 antibodies, showed that, while p85 was clearly visible in the preimmunoprecipitation samples, it could not be detected in the c-Cbl immunoprecipitates from control or IL-3-stimulated cells (Fig. 6A, lower panel). We also examined c-Cbl tyrosine phosphorylation in response to GM-CSF, IL-4, and insulin in FD-6 cells and in Ba/F3 cells in response to IL-3. In none of these analyses could we detect significant alterations in the state of tyrosine phosphorylation of c-Cbl, nor could we detect any interaction between p85 and c-Cbl (data not shown). Therefore, it would appear that c-Cbl does not play a significant role in coupling to the PI3K pathway in FD-6 or Ba/F3 cells. As further confirmation of this, we could not detect c-Cbl in p85 immunoprecipitates prepared from Ba/F3 cells, in the presence or absence of IL-3. While we could detect significant amounts of p100 in these precipitates (Fig. 6B,upper panel), when the immunoblot was reprobed with anti-c-Cbl antibodies, despite c-Cbl being clearly visible in the preimmunoprecipitation samples (Fig. 6B, lower panel), it was not detectable in the anti-p85 immunoprecipitates, even after lengthy exposure of the immunoblot. These results suggest that p85 and c-Cbl do not interact to a great extent in FD-6 or Ba/F3 cells in response to IL-3. c-Cbl clearly migrates as a sharper band of approximately 116 kDa on SDS-PAGE, whereas p100 migrates as a diffuse band between 97 and 105 kDa on the same gels. These differences, together with the data we have presented on their respective interactions, strongly suggest that tyrosine-phosphorylated p100 is not c-Cbl and that the role played by c-Cbl in other cells may be performed in our hemopoietic cells by p100 instead.
Subcellular Localization of p100
We have demonstrated previously that SHP2 can interact directly with the β-chain of the IL-3 receptor (
), and we now propose that p100 could be serving as an adaptor protein linking p85 (PI3K) to SHP2 in response to IL-3. Association of a complex of p85 and p100 with receptor-bound SHP2 would thus act to translocate PI3K to the membrane within the locality of its substrates. To test this hypothesis, we examined the subcellular localization of PI3K-p100 and SHP2-p100 complexes after IL-3 stimulation. Crude membrane and cytosol fractions were prepared from control and IL-3-stimulated Ba/F3 and FD-6 cells. The extracts were subjected to immunoprecipitations using either anti-SHP2 or anti-p85 antibodies and immunoblotted with the anti-phosphotyrosine antibody 4G10. The data for Ba/F3 cells are shown in Fig.7. The major tyrosine-phosphorylated proteins that precipitated with SHP2 antibodies from the cytosolic fraction of IL-3-stimulated cells were p100 and SHP2 (p70). Complexes between SHP2 and p135 (Aic2A) and SHP2 and p100 were also detected in the membrane fraction after IL-3 treatment. As expected for a transmembrane protein, Aic2A (p135) co-precipitated with SHP2 exclusively from the membrane fraction. These results suggest that, although p100 and SHP2 have a predominantly cytosolic localization, after IL-3 stimulation a portion of each tyrosine-phosphorylated protein is present at the membrane.
When anti-p85 IPs were carried out on subcellular fractions of Ba/F3 cells, the major tyrosine-phosphorylated protein co-precipitating with p85 from both membrane and cytosolic fractions was p100 (Fig. 7). Here again, p100 is mainly cytosolic with some membrane localization of the tyrosine-phosphorylated protein after IL-3 stimulation. A small amount of tyrosine-phosphorylated SHP2 (p70) can also be seen co-precipitating with p85 from the cytosol; on close inspection of a longer exposure, this band is also present in the immunoprecipitation of the membrane fraction of IL-3-treated cells, although it is rather faint. These data thus support the notion of a complex forming between Aic2A, SHP2, p100, and p85 (PI3K) at the membrane after IL-3 stimulation, and the potential role of p100 as an adaptor molecule linking PI3K and SHP2.
In this study, we provide evidence that the major tyrosine-phosphorylated protein that co-precipitates with both the p85 subunit of PI3K and SHP2 in IL-3-treated murine lymphohemopoietic cells is a 100-kDa protein, which we have termed p100. The same p100 protein, or a portion of it, appears to interact directly with both p85 and SHP2 via their SH2 domains, and evidence presented suggests that p100 is also a substrate for the phosphatase activity of SHP2. p100 also interacts with Grb2, an association that appears to be mediated predominantly by the Grb2 SH3 domains, suggesting that p100 either contains proline-rich motifs or an as yet undescribed SH3 binding domain.
With respect to proteins of similar size shown previously to be tyrosine-phosphorylated in response to IL-3, we have shown that p100 is not Vav or STAT5. p100 is also highly unlikely to be JAK2, since this protein, although tyrosine-phosphorylated in response to IL-3, runs as a sharp band at 120 kDa (
) while p100 migrates as a more diffuse band at around 100 kDa.
We also investigated the possibility that p100 could be the proto-oncogene product c-Cbl, since this protein had been shown previously to be the major tyrosine-phosphorylated protein that co-precipitates with p85 (PI3K) in response to a number of different stimuli in a variety of different cell lines. However, a number of lines of evidence that we have presented suggest that p100 is not c-Cbl. First, there is a difference in molecular mass; p100 migrates as a diffuse protein, ranging in size from 97 to 105 kDa, whereas c-Cbl migrates as a sharp band of 116 kDa in total cell lysates and as a closely migrating doublet in c-Cbl immunoprecipitates. Second, c-Cbl is not appreciably tyrosine-phosphorylated in response to treatment with a number of different cytokines in either Ba/F3 or FD-6 cells, whereas p100 is one of the major proteins phosphorylated in response to treatment of these same cells with IL-3. Finally, we did not detect any interaction between c-Cbl and p85 in response to IL-3, GM-CSF, IL-4, or insulin in the cell lines used in this study. These data are in contrast to a recent study by Anderson et al. (
), who reported that IL-3 induces the association of p85 with c-Cbl in the cell line 32D cl3. It should be noted that these investigators did not demonstrate a direct interaction between p85 and c-Cbl and also did not demonstrate significant co-immunoprecipitation of these two molecules (
). However, it is of interest that they did obtain good co-precipitation of a tyrosine-phosphorylated protein of around 100 kDa with p85, which could well be the p100 protein discussed here. Nonetheless, p100 does share certain characteristics with c-Cbl. For instance, p100 interacts with the SH2 domain(s) of p85 and with the SH3 domain(s) of Grb2 so it is possible that the two proteins are related or perform a similar function. Hartley et al. (
) have suggested that c-Cbl associates specifically with the β-isoform of p85. However, this isoform appears either to be expressed at very low levels or not at all in our cells.2
A number of phosphotyrosine-containing proteins, with molecular masses ranging from 105 to 115 kDa, have been shown previously to associate with SHP2, although from the data we have presented here, p100 appears to be a distinct species. A 115-kDa protein is phosphorylated in response to insulin in CHO-IR cells and 3T3-L1 adipocytes or in response to nerve growth factor or EGF in PC12 cells, binds SHP2, and appears to be a major substrate for the SHP2 phosphatase (
) report that EGF also induces association of SHP2 with a 115-kDa phosphoprotein in NIH3T3ER and HepG2 cells. Additionally, in Jurkat T-cells, the major tyrosine-phosphorylated protein that associates with SHP2 is a 105-kDa protein (
). Recently, a family of transmembrane phosphotyrosine-containing glycoproteins have been identified that bind to SHP2 in response to treatment with different growth factors, and that appear to act as negative regulators of signaling (
). However, p100 is again distinct in that it has a predominantly cytosolic localization.
We propose that p100 functions, at least in part, by linking p85 and SHP2. Other investigators have suggested that Grb2 performs this linking function, which would require Grb2 to interact with both SHP2 and p85. We and others (
). In monocytes, M-CSF induces an association between Grb2 and p85 that appears to be direct. However, this association is dependent upon M-CSF-induced tyrosine phosphorylation of p85 creating a binding site for the SH2 domain of Grb2 (
) demonstrated such an interaction using the yeast two-hybrid system and also using GST fusion proteins of various domains of p85 and Grb2. In contrast, we could not detect any association of p85 and Grb2 in Ba/F3 or FD-6 cells, suggesting that Grb2 is not involved in mediating the interaction of p85 and SHP2. We used a monoclonal antibody specific for p85α in these analyses and our cells appear to contain little or no p85β.2 However, the p85β bcr region is more proline-rich than the same region of p85α; hence, p85β may participate in interactions with SH3-containing proteins, such as Grb2. Alternatively, since p85 has not been separately cloned from hemopoietic cells, it is formally possible that, although recognized by anti-p85α specific antibodies, the form of p85 present in our cells could differ to the extent that it lacks the particular proline-rich regions required for Grb2 binding. Indeed, the region of p85α and -β that has the lowest identity is the bcr/proline-rich region.
In its characteristics, p100 most closely resembles the recently reported adaptor proteins Gab-1 (
). DOS binds to CSW, theDrosophila homologue of SHP2, and seems to be a substrate for the phosphatase activity of this molecule. DOS also has a potential binding site for the SH2 domain(s) of p85. Antibodies against DOS do not cross-react with p100.3 Gab-1, first described as a Grb2-associated docking protein for EGF and insulin signaling, co-precipitates with Grb2, SHP2, and PI3K. In addition, Gab-1 is detected on Far Western blots of EGF-stimulated cell extracts by PI3K SH2 domain fusion proteins (among others) and associates with Grb2 via the Grb2 SH3 domains. We cannot rule out the possibility that p100 is in some way related to Gab-1. However, the p100 protein that co-precipitates with PI3K and SHP2 from FD-6 and Ba/F3 cells is not detected on an immunoblot using anti-Gab-1 antibodies (data not shown), suggesting that p100 is distinct from Gab-1.
During revision of this manuscript, two other groups have characterized proteins that may well be identical to the p100 protein described here (
) characterized a tyrosine-phosphorylated protein of 100 kDa that associates with SHP2 and p85 (PI3K) in response to M-CSF in FDC-P1 cells. This cell line is related to the FD-6 cells used in this study, and we have reported previously observing a tyrosyl-phosphorylated protein of around 100 kDa in response to CSF-1 stimulation in these cells (
). This p100 protein is similar in many ways to the p100 protein we describe here. Both interact with SHP2 and p85 (PI3K) via the SH2 domains of these proteins, and both appear to be substrates for the SHP2 phosphatase activity. Carlberg and Rohrschneider suggest that their p100 protein acts by directly binding to p85 (PI3K), which is bound to c-Fms (the M-CSF receptor), thus allowing for recruitment of SHP2. This was not addressed directly by experimentation.
) report the characterization of two SHP2-binding proteins from hemopoietic cells. Given that the authors used Ba/F3 cells, the p97 protein that they describe is likely to be identical to the p100 protein discussed here. p97 associated inducibly with SHP2 in response to IL-3 via the SH2 domains of SHP2 and appeared to be a substrate for the phosphatase activity of SHP2. In contrast to the studies reported here, Gu et al. did not detect any p97 in anti-SHP2 IPs from membrane fractions although tyrosine-phosphorylated SHP2 was present in small amounts. This apparent discrepancy with our results may be a question of sensitivity, because p97 was quite difficult to discern in the cytosolic fractions from IL-3-stimulated cells in the data presented by these authors (
). In contrast, we demonstrated good co-precipitation of tyrosine-phosphorylated p100 with SHP2 from the cytosolic fraction of IL-3-stimulated Ba/F3 cells and a contrastingly small amount co-precipitating from the membrane fraction.
Gu et al. also describe characterization of a p135 transmembrane glycoprotein that interacts with SHP2. They conclude that this protein is not the β-chain of the IL-3 receptor and that SHP2 does not interact with IL-3R. However, we have shown previously that SHP2 does interact with βc in response to IL-3, and that this interaction is mediated directly by the SH2 domains of SHP2 (
). Such an interaction is difficult to show in Ba/F3 cells due to the low numbers of IL-3 receptors thought to be present on these cells and the lack of good antibodies against Aic2A (the murine IL-3R β-chain). Based on our previous work, we therefore believe that the tyrosine-phosphorylated p135 protein that we see co-precipitating with SHP2 in response to IL-3 is Aic2A and use this to suggest a model for the role of p100 in localizing PI3K to the membrane after IL-3 stimulation.
We propose that p100 is an adaptor protein in IL-3 signaling that may serve to link SHP2 and PI3K pathways in response to IL-3. One important role of the protein-protein interactions of PI3K mediated by the p85 subunit is its translocation from the cytosol to the plasma membrane, or membrane-bound organelles, where its lipid products are located. Recent reports suggest that targeting of the p110 subunit of PI3K to membranes is sufficient for activation (
). For other factors that can activate PI3K, including IL-3, the situation is likely to be more complex. The IL-3 receptor α and β subunits do not possess tyrosines that conform to the consensus for p85 binding, i.e. YXXM motifs, that could serve as binding sites for p85 (
). We suggest that p100 could play a role in the translocation of PI3K to the plasma membrane in response to IL-3 in the following way. Direct binding of SHP2 to the IL-3 receptor β-chain, which we have reported previously (
), would allow for recruitment of a p100-p85 (PI3K) complex to the membrane by SHP2 binding to the p100 component, and hence facilitate translocation of PI3K to the vicinity of its substrates. This model is illustrated schematically in Fig. 8. We demonstrate, via subcellular fractionation, that a small amount of tyrosine-phosphorylated p100 can be co-precipitated from crude membrane fractions of IL-3-stimulated cells using both anti-SHP2 and anti-p85 antibodies. These data support our proposed model that p100-p85 (PI3K) translocates to the membrane after IL-3 stimulation. SHP2 may act to regulate these interactions via dephosphorylation of p100. With respect to this last point, it is of interest to note that expression of dominant negative SHP2 in Rat-1 fibroblasts overexpressing the human insulin receptor attenuates the activation of PI3K in response to insulin. This is thought to be mediated via an effect on the tyrosine phosphorylation of the adaptor molecule IRS-1 (
). If p100 plays a similar role to IRS-1 in IL-3 signaling, then modulation of SHP2 activity could also affect PI3K activation in this system.
It will obviously be important in the future to identify p100 at the molecular level and determine the significance of the interaction between p85 and SHP2. Work is currently under way to achieve these aims. Further work is also required to confirm the role of p100 in IL-3 signaling, but given that it is the major tyrosine-phosphorylated protein that co-precipitates with SHP2 and p85 (PI3K) in response to IL-3, it is likely to be an important one.