SH2-B, a Membrane-associated Adapter, Is Phosphorylated on Multiple Serines/Threonines in Response to Nerve Growth Factor by Kinases within the MEK/ERK Cascade*

SH2-B has been shown to be required for nerve growth factor (NGF)-mediated neuronal differentiation and survival, associate with NGF receptor TrkA, and be tyrosyl-phosphorylated in response to NGF. In this work, we examined whether NGF stimulates phosphorylation of SH2-B on serines/threonines. NGF promotes a dramatic upward shift in mobility of SH2-B, resulting in multiple forms that cannot be attributed to tyrosyl phosphorylation. Treatment of SH2-B with protein phosphatase 2A, a serine/threonine phosphatase, reduces the many forms to two. PD98059, a MEK inhibitor, dramatically inhibits NGF-promoted phosphorylation of SH2-B on serines/threonines, whereas depletion of 4β-phorbol 12-myristate 13-acetate-sensitive protein kinase Cs does not. ERKs 1 and 2 phosphorylate SH2-Bβ primarily on Ser-96 in vitro. However, NGF still stimulates serine/threonine phosphorylation of SH2-Bβ(S96A). SH2-Bβ(S96A), like wild-type SH2-Bβ, enhances NGF-induced neurite outgrowth. In contrast, SH2-Bβ(R555E) containing a defective SH2 domain blocks NGF-induced neurite outgrowth and displays greatly reduced phosphorylation on serines/threonines in response to NGF. SH2-Bβ(R555E), like wild-type SH2-Bβ, associates with the plasma membrane, suggesting that the dominant negative effect of SH2-Bβ(R555E) cannot be explained by an abnormal subcellular distribution. In summary, NGF stimulates phosphorylation of SH2-B on serines/threonines by kinases downstream of MEK, which may be important for NGF-mediated neuronal differentiation and survival.

Neurotrophins, including NGF, 1 brain-derived neurotrophic factor, neurotrophin-3, neurotrophin-4/5, and neurotrophin-6, play a crucial role in differentiation, survival, and plasticity of developing neurons. NGF is essential for development and survival of sympathetic neurons and a subpopulation of sensory neurons (1)(2)(3)(4). NGF binds with high affinity to TrkA, a member of the Trk family of receptor tyrosine kinases, and with low affinity to p75NTR, a member of the tumor necrosis factor receptor family (5). TrkA appears to be the major mediator of NGF signaling in developing and adult neurons (6). In rat pheochromocytoma (PC12) cells, NGF promotes neuronal differentiation (e.g. extension of neurite outgrowth and expression of neuronal specific genes) through activation of TrkA (7)(8)(9)(10).
NGF stimulates the dimerization of TrkA (11), resulting in the activation of the intrinsic tyrosine kinase of TrkA and autophosphorylation of multiple tyrosines within the cytoplasmic domain of TrkA (12). The phosphorylated tyrosines recruit to the TrkA complex signaling molecules containing Src homology 2 (SH2) or phosphotyrosine-interacting domains, including Shc (12)(13)(14), phospholipase C␥ (12,13), and SH2-B (15,16). These signaling molecules then initiate the activation of multiple signaling pathways that mediate the biological responses to NGF.
One such pathway required for NGF-induced neuronal differentiation is the Ras/Raf/MEK/ERK pathway. NGF stimulates Shc binding to phosphorylated Tyr-490 in TrkA and the tyrosyl phosphorylation of Shc by TrkA (12,14,17,18), which then enables Shc to recruit Grb2-SOS complexes to the plasma membrane. This results in the activation of the Ras/Raf/MEK/ ERK pathway (19). Microinjection of antibody against Ras (20) or overexpression of dominant negative mutant Ras (21) or dominant negative mutant MEK (22) abrogates NGF-promoted neuronal differentiation of PC12 cells. Furthermore, overexpression of oncogenic Shc (23), oncogenic Ras (24,25), oncogenic Raf (26), or constitutively active MEK (22) is sufficient to promote neuronal differentiation of PC12 cells similar to that induced by NGF. These observations suggest that serine/threonine phosphorylation of proteins by the Ras/Raf/MEK/ERK pathway plays an essential role in NGF signaling.
SH2-B, a recently described adapter protein containing SH2 and pleckstrin homology (PH) domains (27,28), has been shown to be required for NGF-induced neuronal differentiation of PC12 cells (16) and implicated in NGF-mediated axonal growth and survival of primary sympathetic neurons (15). NGF stimulates association of SH2-B with TrkA in PC12 cells (16) as well as in primary sympathetic neurons (15). NGF promotes tyrosyl phosphorylation of SH2-B in primary sympathetic neurons (15) and PC12 cells overexpressing TrkA or SH2-B (16). NGF-induced tyrosyl phosphorylation of SH2-B was also observed in untransfected PC12 cells but only in the presence of a tyrosine phosphatase inhibitor (16), suggesting that any phosphorylated tyrosines in SH2-B are rapidly dephosphorylated.
In addition to its 9 tyrosines, SH2-B␤ has a large number of serines (82 serines) and threonines (29 threonines) including many that lie within consensus sequences for phosphorylation sites for protein kinase C, ERKs 1 and 2, cAMP-and cGMPdependent protein kinases, and casein kinase II. We previously reported that platelet-derived growth factor stimulates phosphorylation of SH2-B on serines/threonines as well as on tyrosines (28). However, the kinase(s) that phosphorylate SH2-B on serines/threonines are unknown. In this study, we provide strong evidence that NGF stimulates phosphorylation of SH2-B on multiple serines/threonines by MEK or kinases downstream of MEK. ERKs 1 and 2 phosphorylate SH2-B␤ primarily on Ser-96 in vitro, but SH2-B␤(S96A) is still phosphorylated on multiple serines/threonines in cells in response to NGF and enhances NGF-induced neurite outgrowth. These findings indicate that kinase(s) downstream of MEK, other than ERKs 1 and 2, phosphorylate SH2-B and that phosphorylation of Ser-96 by ERKs 1 and 2 is not required for its action in NGF-mediated neurite outgrowth. A dominant negative SH2-B␤(R555E), which is unable to bind to TrkA, exhibits a profound defect in its serine/threonine phosphorylation, suggesting that association with TrkA and/or subsequent tyrosyl phosphorylation by TrkA may be a prerequisite for serine/threonine phosphorylation of SH2-B in response to NGF. Antibodies-Antibodies to rat SH2-B␤ (␣SH2-B) were raised against a GST fusion protein containing the C-terminal portion of SH2-B␤ as described previously (27) and used at a dilution of 1:100 for immunoprecipitation and 1:15,000 for immunoblotting. Monoclonal anti-phosphotyrosine antibody 4G10 (␣PY) was purchased from Upstate Biotechnology Inc. and was used at a dilution of 1:7,500 for immunoblotting. Anti-active mitogen-activated protein kinase was from Promega and used at a dilution of 1:20,000 for immunoblotting. Anti-ERK2 was from Santa Cruz Biotechnology and used at a dilution of 1:100 for immunoprecipitation.
PC12 cells were transfected with plasmids (pEGFP-C1) encoding GFP, GFP-SH2-B␤, GFP-SH2-B␤ (S96A), or GFP-SH2-B␤ (R555E), using LipofectAMINE Plus (Life Technologies, Inc.) according to the protocol recommended by the manufacturer. After 72 h growth in regular medium, the transfectants were cultured for 50 additional days in medium supplemented with 1 mg/ml G418. The G418-resistant transfectants were pooled, and the top 2% of cells in terms of expression of GFP fluorescence was selected by flow cytometry.
Immunoprecipitation and Immunoblotting-Cell lysates were incu-bated with the indicated antibody on ice for 2 h. The immune complexes were collected on protein A-agarose (50 l) during 1-h incubation at 4°C. The beads were washed 3 times with washing buffer (50 mM Tris, pH 7.5, 0.1% Triton X-100, 150 mM NaCl, 2 mM EGTA) and boiled for 5 min in a mixture (80:20) of lysis buffer and SDS-PAGE sample buffer (250 mM Tris-HCl, pH 6.8, 10% SDS, 10% ␤-mercaptoethanol, 40% glycerol, 0.01% bromphenol blue). The solubilized proteins were separated by SDS-PAGE (5-12% gradient or 7.5% gels). Proteins on the gel were transferred to nitrocellulose membrane (Amersham Pharmacia Biotech) and detected by immunoblotting with the indicated antibody using ECL. Some membranes were then incubated at 55°C for 30 -60 min in stripping buffer (100 mM ␤-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7). The membranes were then immunoblotted with the desired antibody. In some cases, the blots were reprobed with a second antibody without stripping. Dephosphorylation-PC12 cells were treated with 100 ng/ml NGF for 10 min, and the cell lysates were immunoprecipitated with ␣SH2-B. The immunoprecipitates were incubated at 37°C for 60 min with 40 units of alkaline phosphatase in 100 l of dephosphorylation buffer (50 mM Tris-HCl, 0.1 mM EDTA, pH 8.5, 10 g/ml aprotinin, and 10 g/ml leupeptin) in the presence or absence of 5 mM Na 3 VO 4 . The reaction was terminated, and proteins were eluted by boiling in a mixture (80:20) of lysis buffer and SDS-PAGE sample buffer. As controls, the immunoprecipitates were treated identically except no alkaline phosphatase was added. The resultant dephosphorylated proteins were resolved by SDS-PAGE and immunoblotted with ␣SH2-B.
Separation of the Plasma Membrane from the Cytosol-The confluent PC12 cells were rinsed three times with ice-cold PBSV, scraped from plates in lysis buffer (50 mM Tris-HCl, pH 7.5, 50 mM ␤-mercaptoethanol, 2 mM EGTA, 0.1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin), and subjected to sonication on ice for 50 s. The cell lysate was centrifuged at 120,000 ϫ g for 45 min at 4°C. The supernatant contained the cytosol. The pellet (designated the membrane fraction) was resuspended in lysis buffer.
Confocal Fluorescence Microscopy-Confocal imaging was performed with a Noran OZ laser scanning confocal microscope equipped with a ϫ 60 Nikon objective. GFP was excited at 488 nm by a krypton-argon laser, and fluorescence above 500 nm was captured. Cells were grown on collagen-coated glass coverslips attached to the bottom of a 60-mm culture dish and imaged at room temperature in Krebs-Ringer phosphate buffer (128 mM NaCl, 7 mM KCl, 1 mM CaCl 2 , 1.2 mM MgSO 4 , 1 mM NaHPO 4 , 10 mM glucose, pH 7.4) containing 0.1% bovine serum albumin. The contribution of cellular autofluorescence was judged to be less than 1%. The presented images are representative of at least three separate experiments.

NGF Stimulates Phosphorylation of SH2-B on Multiple
Serines/Threonines-We have shown previously that NGF stimulates a large shift in mobility of SH2-B as well as tyrosyl phosphorylation of SH2-B when PC12 cells overexpress TrkA or SH2-B or are pretreated with Na 3 VO 4 , a tyrosine phosphatase inhibitor (16). To test whether serine/threonine phosphorylation of SH2-B contributes to the NGF-induced shift in mobility of SH2-B, PC12 cells were treated with NGF in the absence of Na 3 VO 4 . SH2-B was immunoprecipitated with ␣SH2-B and immunoblotted with ␣PY. No detectable tyrosyl phosphorylation of SH2-B was observed using this experimen-tal paradigm in either control or NGF-treated cells (Fig. 1A,  lanes 1 and 2), as reported previously (15,16). Surprisingly, NGF still stimulated a dramatic shift in mobility of SH2-B, generating multiple forms of SH2-B with different migrations (Fig. 1A, lanes 3 and 4). These data suggest that the NGFinduced shift in mobility of SH2-B is due to phosphorylation of SH2-B on multiple serine(s)/threonine(s).
To verify that the NGF-reduced change in migration of SH2-B is caused by phosphorylation, ␣SH2-B precipitates were treated with alkaline phosphatase, a nonspecific protein phosphatase that dephosphorylates phosphotyrosines, phosphoserines, and phosphothreonines. Alkaline phosphatase treatment reduced the multiple forms of SH2-B observed in NGFtreated cells to primarily a faster migrating form (Fig. 1B, lane 2 versus 4) that co-migrated with the band seen in cells not incubated with NGF (Fig. 1B, lane 1). Sodium vanadate, an inhibitor of alkaline phosphatase, abolished the effect of alkaline phosphatase on SH2-B migration (data not shown), indicating that the change of migration of SH2-B by alkaline phos-phatase is caused by dephosphorylation.
To confirm that the NGF-dependent phosphorylation of SH2-B is primarily on multiple serines/threonines, ␣SH2-B precipitates were treated with protein phosphatase 2A (PP2A) that dephosphorylates only phosphoserines and phosphothreonines (28 -31). PP2A treatment of SH2-B from NGFtreated cells reduced the multiple forms of SH2-B to two distinct forms (Fig. 1B, lane 2 versus 6), with the majority comigrating with SH2-B from control cells (Fig. 1B, lanes 5 and  6). The small portion of the slower migrating SH2-B (Fig. 1B, lane 6) may represent tyrosyl-phosphorylated SH2-B whose phosphotyrosine(s) cannot be detected with ␣PY or SH2-B phosphorylated on specific serine(s)/threonine(s) that are resistant to PP2A. Okadaic acid, a potent inhibitor of PP2A, blocked the effect of PP2A on the NGF-induced mobility shift of SH2-B (data not shown). These results indicate that the NGFinduced shift in mobility of SH2-B is primarily caused by phosphorylation of SH2-B on serines and/or threonines. The presence of multiple forms of SH2-B in response to NGF suggests that SH2-B is phosphorylated at multiple sites in response to NGF.
The upward shift in mobility of SH2-B was detected within 1 min of NGF treatment, was maximal within 15 min, and decreased after 30 min of NGF treatment (Fig. 1C, upper panel). A residual upward shift in mobility of SH2-B was maintained through 2 h (Fig. 1C, upper panel). A mobility shift of SH2-B␤ was detectable at concentrations of NGF as low as 1 ng/ml and reached a maximum at 25 ng/ml (Fig. 1C, lower panel). These data suggest that NGF-stimulated serine/threonine phosphorylation of SH2-B is rapid and transient.
MEK Is Critical for NGF-induced Phosphorylation of SH2-B on Serines/Threonines-NGF activates MEK, a kinase that plays an essential role in NGF-induced neuronal differentiation of PC12 cells (22). PD98059, a MEK inhibitor (32), inhibited NGF-induced neurite outgrowth (Fig. 2b) consistent with previous work (33). Interestingly, overexpression of SH2-B␤, which has been shown to enhance NGF-induced neurite outgrowth (16), was unable to overcome the inhibition of NGFinduced neuronal differentiation by the MEK inhibitor (Fig.  2d), raising the possibility that phosphorylation of SH2-B by kinases within the MEK/ERK cascade may be involved in regulation of neurite outgrowth in response to NGF.
To determine whether MEK plays a role in NGF-induced serine/threonine phosphorylation of SH2-B, PC12 cells were  1 and 2). The same blot was reprobed with ␣SH2-B (lanes 3 and 4). B, PC12 cells were stimulated with 100 ng/ml NGF for 10 min, and SH2-B was immunoprecipitated with ␣SH2-B. The precipitated SH2-B was incubated with dephosphorylation buffer in the absence (lanes 1 and 2)   pretreated for 30 min with 100 M PD98059 prior to stimulation with 50 ng/ml NGF for 10 min. SH2-B was immunoprecipitated and immunoblotted with ␣SH2-B. PD98059 substantially, but not completely, inhibited the NGF-induced shift in mobility of SH2-B (Fig. 3A, top panel, lane 2 versus 3), suggesting that MEK is required for a substantial portion of the NGF-induced phosphorylation of SH2-B. Similarly, inhibition of the NGF-induced shift in mobility of SH2-B by PD98059 was also observed when cell lysates were immunoblotted with ␣SH2-B (Fig. 3A, middle panel, lane 2 versus 3). As expected, NGF stimulated activation of ERKs 1 and 2 (Fig. 3A, bottom panel, lane 2) as revealed by immunoblotting cell lysates with an antibody that recognizes the dual phosphorylated, active form of ERKs 1 and 2. PD98059 treatment inhibited substantially, but not completely, NGF-induced activation of ERKs 1 and 2 (Fig. 3A, bottom panel, lane 2 versus 3). Therefore, the residual serine/threonine phosphorylation of SH2-B after PD98059 treatment may be dependent on MEK.
Because multiple isoforms of PKC are proposed to play a role in NGF signaling (34 -37), we also examined whether PMAsensitive PKCs are involved in NGF-induced phosphorylation of SH2-B. PMA, a robust activator of multiple isoforms of conventional PKCs (38), stimulated a mobility shift of SH2-B in PC12 cells (Fig. 3B, upper panel, lane 1 versus 2). PMA also stimulated activation of ERKs 1 and 2 (Fig. 3B, lower panel,  lanes 1 and 2), consistent with the previous observations that PKC activates the Raf/MEK/ERK pathway (39 -41). Therefore, it is unclear whether PKC activated by PMA phosphorylates SH2-B directly or indirectly via MEK or kinases downstream of MEK. Chronic PMA pretreatment abolished the ability of PMA to activate ERKs 1 and 2 (Fig. 3B, lower panel, lanes 4 and 5) and to stimulate a mobility shift of SH2-B (Fig. 3B, upper  panel, lanes 4 and 5), consistent with chronic treatment of cells with PMA depleting PMA-sensitive PKCs (36). In contrast, chronic treatment of PC12 cells with PMA did not alter the NGF-induced mobility shift of SH2-B (Fig. 3B, upper panel, lane 3 versus 6) or activation of ERKs 1 and 2 (Fig. 3B, lower panel, lane 3 versus 6), suggesting that PMA-sensitive PKCs are not responsible for the NGF-stimulated serine/threonine phosphorylation of SH2-B that is responsible for its shift in mobility.
SH2-B␤ Is Phosphorylated at Ser-96 in Vitro by Activated ERKs 1 and 2-Because ERKs 1 and 2 lie downstream of MEK and SH2-B has a consensus phosphorylation site (Pro-Leu-Ser 96 -Pro) for mitogen-activated protein kinase, we tested whether Ser-96 is a target of activated ERKs 1 and 2. ERKs 1 and 2 from either control or NGF-stimulated cells were immunoprecipitated with ␣ERK2 (␣ERK2 recognizes both ERKs 1 and 2) and used for an immunocomplex kinase assay with GST-SH2-B␤ fusion protein as an exogenous substrate. ERKs 1 and 2 precipitated from NGF-stimulated but not control cells phosphorylated SH2-B␤ in this in vitro kinase assay (Fig. 4A,  upper panel, lanes 3 and 4). When Ser-96 was mutated to Ala (S96A), phosphorylation of SH2-B␤ (S96A) by activated ERKs 1 and 2 in vitro was reduced dramatically (Fig. 4A, upper panel,  lane 2 versus 4), suggesting that Ser-96 is the primary phosphorylation site within SH2-B␤ for ERKs 1 and 2. Similarly, ERKs 1 and 2 from EGF-treated cells phosphorylated SH2-B␤ in vitro (Fig. 4B). Surprisingly, the mobility of SH2-B␤ did not detectably change after its phosphorylation by ERKs 1 and 2 (Fig. 4A, lower panel, lanes 3 and 4), suggesting that phosphorylation of Ser-96 in SH2-B by ERKs 1 and 2 does not account for the NGF-induced mobility shift of SH2-B observed in PC12 cells. Combined with data from the MEK inhibitor experiments, these results suggest that in addition to ERKs 1 and 2, MEK or kinases downstream of MEK phosphorylate SH2-B on serines/threonines in response to NGF.
Phosphorylation of Ser-96 in SH2-B by ERKs 1 and 2 Is Not Required for NGF-induced Neurite Outgrowth-To investigate whether phosphorylation of SH2-B on Ser-96 by ERKs 1 and 2 plays a role in NGF-induced neuronal differentiation, GFPtagged SH2-B␤(S96A) was stably overexpressed in PC12 cells as described previously (16). In agreement with our previous observation, overexpression of GFP-SH2-B␤ significantly enhanced NGF-induced neurite outgrowth, whereas overexpression of GFP-SH2-B␤(R555E) blocked neurite outgrowth induced by NGF (Fig. 5), indicating that SH2-B is an essential signaling molecule for NGF-induced neurite outgrowth. SH2-B␤(R555E) is a dominant negative form of SH2-B that inhibits the function of endogenous SH2-B in neuronal differentiation induced by NGF. Interestingly, overexpression of GFP-SH2-B␤(S96A) enhanced NGF-induced neurite outgrowth of PC12 cells to a similar extent as wild-type GFP-SH2-B␤ (Fig. 5), indicating that phosphorylation of SH2-B on Ser-96 by ERKs is not required for NGF-induced neurite outgrowth.

FIG. 5. Phosphorylation of Ser-96 in SH2-B by ERKs is not required for its action in NGF-induced neurite outgrowth.
PC12 cells were stably transfected with plasmids encoding GFP, GFP-tagged SH2-B␤ (WT), GFP-SHZ-B(S96A) (S96A), or GFP-SH2-B␤(R555E) (R555E) as described previously (16). Cells were stimulated with 25 ng/ml NGF for 2 days and visualized using phase contrast microscopy. The scale bar represents 20 m. tionated by centrifugation. Cytosolic or membrane proteins were immunoblotted with ␣SH2-B. The majority of SH2-B was in the membrane compartment (Ͼ85%, Fig. 7A, upper panel). In contrast, all three isoforms of Shc were present primarily in the cytosolic compartment (Fig. 7A, lower panel). The association of SH2-B with the plasma membrane was also observed in 3T3-F442A cells (data not shown).
To verify the association of SH2-B with the plasma membrane, wild-type SH2-B␤, SH2-B␤(S96A), or SH2-B␤(R555E) was tagged with GFP, stably expressed in PC12 cells (16), and visualized using confocal microscopy. As reported previously (42), GFP alone was evenly distributed throughout the cell (Fig. 7B). In contrast, GFP-SH2-B␤ was excluded from the nucleus and accumulated at the plasma membrane (Fig. 7B). SH2-B␤(S96A) and SH2-B␤(R555E) were also present predominantly at the plasma membrane (Fig. 7B). This observation suggests that phosphorylation of Ser-96 does not affect its subcellular localization nor is the dominant negative effect of SH2-B␤(R555E) attributable to an abnormal subcellular distribution. DISCUSSION SH2-B has been shown to be tyrosyl-phosphorylated by TrkA and required for NGF-mediated neuronal differentiation and survival (15,16). In this work, we provide strong evidence that SH2-B is phosphorylated on multiple serines/threonines in response to NGF. In support of NGF promoting phosphorylation of SH2-B on multiple sites, NGF stimulates a dramatic shift in mobility of SH2-B, resulting in multiple forms with different migrations. When treated with alkaline phosphatase that dephosphorylates phosphotyrosines, phosphoserines, and phosphothreonines, these multiple forms revert to a single form of SH2-B migrating similarly to SH2-B from control cells. In support of SH2-B being phosphorylated on multiple serines/ threonines, the multiple forms of SH2-B induced by NGF in the absence of a tyrosine phosphatase inhibitor are not recognized by anti-phosphotyrosine. Furthermore, treatment of SH2-B from NGF-stimulated cells with PP2A, which specifically dephosphorylates phosphoserines and phosphothreonines, dramatically reduces the migration of SH2-B.
Our results indicate that kinase(s) within the MEK/ERK cascade play a critical role in this NGF-induced phosphorylation of SH2-B on serines/threonines. PD98059 inhibits the NGF-induced shift in mobility of SH2-B. NGF promotes prolonged phosphorylation of SH2-B, whereas EGF stimulates transient phosphorylation (data not shown), consistent with sustained activation of the MEK/ERK cascade by NGF and transient activation of this pathway by EGF. It is unlikely that MEK phosphorylates SH2-B, because there is no consensus sequence in SH2-B for phosphorylation by MEK. Obvious candidates downstream of MEK are ERKs 1 and 2, since ERKs 1 and 2 are known to be required for NGF-induced neuronal differentiation of PC12 cells (22), and there is a consensus sequence for ERKs 1 and 2 within SH2-B. In vitro studies identified Ser-96 of SH2-B as a phosphorylation site for ERKs 1 and 2, although whether ERKs phosphorylate Ser-96 in intact cells remains to be determined. However, phosphorylation of SH2-B␤ on Ser-96 by ERKs 1 and 2 in vitro did not detectably change its migration in SDS-PAGE gels, and SH2-B␤(S96A) ectopically expressed in PC12 cells exhibited a large NGF-induced shift in its mobility indistinguishable from that of wild-type SH2-B␤. This suggests that besides ERKs 1 and 2 another kinase downstream of MEK or an as yet unidentified PD98059-sensitive kinases phosphorylate SH2-B on serines/ threonines in response to NGF. A PD98059-sensitive kinase(s) other than ERKs 1 and 2 has been reported to phosphorylate SOS in insulin signaling (43). One candidate kinase is p90 rsk , which lies downstream of MEK and phosphorylates cAMPresponse element binding protein and Fos in response to NGF in PC12 cells (44 -48).
In contrast, depletion of PMA-sensitive isoforms of PKCs does not affect NGF-promoted phosphorylation of SH2-B on serines/threonines, although SH2-B has multiple potential phosphorylation sites for PKC. These results indicate that PMA-sensitive isoforms of PKC may not play a significant role in the NGF-induced phosphorylation of SH2-B.
Interestingly, NGF-induced serine/threonine phosphorylation of the dominant negative SH2-B␤(R555E) was dramatically reduced, although the activity of kinases that phosphorylate endogenous SH2-B in PC12 cells overexpressing SH2-B␤(R555E) appeared to be normal. These findings indicate that the SH2 domain of SH2-B is necessary for SH2-B to be phosphorylated on serines/threonines. We observed previously that SH2-B␤(R555E) is unable to bind to TrkA and be tyrosyl-phosphorylated by TrkA (16). We speculate that association with TrkA or/and tyrosyl phosphorylation of SH2-B is required for NGF-induced phosphorylation of SH2-B on serines/threonines. SH2-B has been proposed as an adapter for a variety of hormones, cytokines, and growth factors (15,16,27,28,49). In addition to its SH2 domain, it has multiple proline-rich motifs, a PH domain, and multiple potential serine, threonine, and tyrosine phosphorylation sites. PH domains and proline-rich motifs are present in many signaling molecules and are thought to target these proteins to the plasma membrane by binding to phospholipids (50 -52) and constitutively associate with SH3 or WW domain-containing proteins, respectively (53)(54)(55)(56)(57)(58)(59). Therefore, we think it likely that its PH domain targets SH2-B to the plasma membrane, and its proline-rich motifs interact constitutively with signaling molecules containing SH3 domains. NGF-induced phosphorylation of SH2-B on tyrosines may recruit downstream effectors containing SH2 domain. Thus, one could envision that SH2-B acts as an adapter or a scaffold protein that assembles a large protein complex (signalingsome) of multiple signaling proteins. The interaction of the SH2 domain of SH2-B with phosphorylated tyrosine(s) in the cytoplasmic domain of TrkA recruits this signalingsome to TrkA in response to NGF. Phosphoserines and phosphothreonines have also been shown to form binding sites for various signaling molecules (60 -65). One domain that has been shown to bind specifically to phosphoserines and phosphothreonines is the WW domain that is present in many signaling molecules (66). Thus, phosphoserines and phosphothreonines in SH2-B may recruit to this signalingsome downstream effectors in response to NGF. Alternatively, NGF-induced serine/threonine phosphorylation of SH2-B may change the conformation of SH2-B, thereby regulating the composition of this signalingsome and/or activity of some components of this signalingsome. Consistent with these ideas, SH2-B␤(R555E), which is defective in its association with TrkA and its phosphorylation on tyrosines and serines/threonines, blocks NGF-induced neurite outgrowth, presumably by constitutively binding and sequestering critical downstream effectors away from endogenous SH2-B. The C-terminal part of SH2-B also acts as a dominant negative mutant in NGF signaling (15). This mutant contains the entire SH2 domain but lacks most proline-rich motifs, tyrosines, and serines/threonines. It would be expected to compete with endogenous SH2-B for TrkA but not to bind all of the signaling molecules needed for the actions of NGF.
SH2-B has been reported to associate constitutively with Grb2 and to mediate NGF-stimulated activation of ERKs 1 and 2 via a mutant TrkA lacking its Shc-binding site (15). However, our previous data indicate that overexpression of neither wildtype SH2-B␤ nor a dominant negative SH2-B␤(R555E) alters NGF-induced activation of ERKs 1 and 2 in PC12 cells (16), suggesting that SH2-B does not play a significant role in NGFinduced activation of ERKs via endogenous TrkA, at least in PC12 cells. When the primary phosphorylation site Ser-96 for ERKs was mutated, SH2-B␤(S96A) was still able to enhance NGF-induced neurite outgrowth, suggesting that phosphorylation of SH2-B by ERKs is also not required for its action in promoting NGF-induced neurite outgrowth.
In summary, we show that SH2-B resides at the plasma membrane, which presumably positions it to bind rapidly to TrkA in response to NGF. We also show that upon NGF stimulation, SH2-B is phosphorylated on multiple serines/threonines by kinase(s) downstream of MEK. ERKs are known to be activated by NGF and, like SH2-B, are required for neuronal differentiation of PC12 cells. However, although SH2-B is phosphorylated on Ser-96 by ERKs 1 and 2 in vitro, phosphorylation of Ser-96 does not play a significant role in NGF-promoted neurite outgrowth. Kinases downstream of MEK (or of other as yet unidentified PD98059-sensitive kinases) other than ERKs 1 and 2 phosphorylate SH2-B on multiple sites. NGF-induced phosphorylation of SH2-B on serines/threonines requires the SH2 domain of SH2-B, as does the NGF-induced association with TrkA, phosphorylation of SH2-B on tyrosine(s), and the action of SH2-B on NGF-induced neurite outgrowth. These findings raise the possibility that phosphorylation of SH2-B on serines/threonines by kinases downstream of MEK other than ERKs 1 and 2 is important in mediating the role of SH2-B in NGF-induced neurite outgrowth.