YXXL Motifs in SH2-Bβ Are Phosphorylated by JAK2, JAK1, and Platelet-derived Growth Factor Receptor and Are Required for Membrane Ruffling*

SH2-Bβ binds to the activated form of JAK2 and various receptor tyrosine kinases. It is a potent stimulator of JAK2, is required for growth hormone (GH)-induced membrane ruffling, and increases mitogenesis stimulated by platelet-derived growth factor (PDGF) and insulin-like growth factor I. Its domain structure suggests that SH2-Bβ may act as an adapter protein to recruit downstream signaling proteins to kinase·SH2-Bβ complexes. SH2-Bβ is tyrosyl-phosphorylated in response to GH and interferon-γ, stimulators of JAK2, as well as in response to PDGF and nerve growth factor. To begin to elucidate the role of tyrosyl phosphorylation in the function of SH2-Bβ, we used phosphopeptide mapping, mutagenesis, and a phosphotyrosine-specific antibody to identify Tyr-439 and Tyr-494 in SH2-Bβ as targets of JAK2 bothin vitro and in intact cells. SH2-Bβ lacking Tyr-439 and Tyr-494 inhibits GH-induced membrane ruffling but still activates JAK2. We provide evidence that JAK1, like JAK2, phosphorylates Tyr-439 and Tyr-494 in SH2-Bβ and that PDGF receptor phosphorylates SH2-Bβ on Tyr-439. Therefore, phosphorylated Tyr-439 and/or Tyr-494 in SH2-Bβ may provide a binding site for one or more proteins linking cytokine receptor·JAK2 complexes and/or receptor tyrosine kinases to the actin cytoskeleton.

SH2-B 1 is a putative adapter protein (see Fig. 1 below) implicated in the actions of multiple cytokines and growth factors.
The ␤ isoform of SH2-B was originally identified as a JAK2binding protein that is tyrosyl-phosphorylated in response to GH and interferon-␥, stimulators of the tyrosine kinase JAK2 (1). SH2-B␤ binds preferentially via its SH2 domain to the tyrosyl-phosphorylated, active form of JAK2. It increases dramatically the activity of JAK2 and enhances the tyrosyl phosphorylation of downstream targets of JAK2 such as STAT5 (2). Thus, SH2-B␤ would be predicted to act in a positive feedback manner to increase signals by cytokines that activate JAK2. SH2-B␤ also binds to and is tyrosyl-phosphorylated by JAK1, suggesting that SH2-B␤ may also act as an adapter protein in signaling through cytokines that activate JAK1 (3).
Four splice variants of SH2-B have been identified (1, 4 -6), ␣, ␤, ␥, and ␦. They are identical for the first 631 amino acids and diverge just past the SH2 domain (6). Initial studies suggest that they have overlapping signaling properties. One or more isoforms have been shown to be tyrosyl-phosphorylated by activated forms of the receptors for platelet-derived growth factor (PDGF) (6 -8), insulin-like growth factor-I (6,8,9), nerve growth factor (NGF) (10,11), insulin (5,9,12,13), and fibroblast growth factor (14). SH2-B␤ has been implicated in GHand PDGF-induced changes in the actin cytoskeleton and/or cell motility (15,16). SH2-B␤ and SH2-B␣ have both been implicated in NGF-induced neurite outgrowth (10,11) and SH2-B␣ has been shown to increase the NGF-induced tyrosyl phosphorylation of TrkA, the receptor tyrosine kinase for NGF (17). All four isoforms of SH2-B have been shown to increase DNA synthesis and cellular proliferation stimulated by PDGF or insulin-like growth factor-I (6,8), although to varying degrees. Lastly, SH2-B␤ has been shown to increase the phosphorylation and nuclear translocation of STAT5B mediated by a constitutively active form of fibroblast growth factor receptor 3 (14). These findings suggest that SH2-B mediates or regulates signaling pathways leading to cell motility, growth, and/or differentiation induced by multiple growth factors. However, the mechanisms by which SH2-B␤ mediates its effects on cell signaling are largely unknown.
The domain structure of SH2-B␤ suggests that it may act as an adapter protein to recruit proteins to cytokine receptor⅐JAK complexes and receptor tyrosine kinases. SH2-B␤ contains an SH2 domain, a pleckstrin homology domain, three proline rich regions, and nine tyrosines (see Fig. 1) (1). We have shown that the second proline-rich region in SH2-B␤ is required for maximal cell motility induced by GH and for binding to the small GTPase, Rac (16). Rac is a member of the Rho family of small GTPases and has been implicated in the formation of lamellipodia and other structures required for cell motility (18,19). Phosphorylated tyrosines in SH2-B␤ are likely to be binding sites for downstream molecules containing SH2 or phosphotyrosine binding domains.
To gain further insight into the mechanisms by which SH2-B␤ elicits its effects, we sought to identify which tyrosines in SH2-B␤ are phosphorylated and the role tyrosyl phosphorylation plays in the function of SH2-B␤. Phosphopeptide mapping determined that Tyr-439 and Tyr-494 in SH2-B␤ are targets of JAK2 in vitro and in vivo. We also show that, even though SH2-B␤ lacking these tyrosines still activates JAK2, SH2-B␤ lacking tyrosines 439 and 494 acts as a dominant negative to inhibit GH-induced membrane ruffling demonstrating a critical role for Tyr-439 and Tyr-494 in the ability of SH2-B␤ to regulate membrane ruffling. Furthermore, evidence is provided that JAK1 and the receptor tyrosine kinase, PDGFR, also phosphorylate SH2-B␤ on Tyr-439. Thus, phosphorylation of Tyr-439 (pY439) in response to PDGF as well as cytokines that activate JAK1 and JAK2 may provide a binding site for proteins important in the regulation of the actin cytoskeleton.

EXPERIMENTAL PROCEDURES
Cells and Reagents-The stocks of COS-7 and 293T cells were provided by Drs. M. D. Uhler (University of Michigan, Ann Arbor, MI) and E. R. Fearon (University of Michigan), respectively. The stock of 3T3-F442A murine fibroblasts was provided by H. Green (Harvard University, Cambridge, MA). Aprotinin, leupeptin, and Triton X-100 were from Roche Molecular Biochemicals. Recombinant protein A-agarose was from Repligen. Enhanced chemiluminescence (ECL) detection system was from Amersham Biosciences. Anti-JAK2 antiserum (␣JAK2) was raised in rabbits against a synthetic peptide corresponding to amino acids 758 -766 of murine JAK2 (20,21) and was used at a dilution of 1:500 for immunoprecipitation and 1:15,000 for immunoblotting. Antibody to rat SH2-B␤ (␣SH2-B␤) was raised against a glutathione Stransferase fusion protein containing amino acids 527-670 of SH2-B␤ as described previously (1) and was used at a dilution of 1:100 for immunoprecipitations and 1:15,000 for immunoblotting. Monoclonal anti-phosphotyrosine antibody (␣PY, clone 4G10) from Upstate Biotechnology, Inc was used at a dilution of 1:7,500 for immunoblotting. Polyvinylpyrrolidone and phosphoamino acid standards were from Sigma. Methylated trypsin was from Promega. Thin-layer chromatography plates were from EM Science.
Cell Culture and Transfection-COS-7, 293T, or 3T3-F442A murine fibroblast cells were grown in Dulbecco's modified Eagle medium (DMEM) supplemented with 1 mM L-glutamine, 100 units of penicillin per ml, 100 g of streptomycin per ml, 0.25 g of amphotericin per ml (supplemented DMEM), and 9% fetal calf serum (COS-7) or 9% calf serum (293T and 3T3-F442A). COS-7 and 293T cells were transiently transfected using calcium phosphate precipitation (23). Transfected cells were assayed 24 (293T) or 48 (COS-7) h after transfection. COS-7 cells overexpressing either PDGFR or JAK2 were incubated overnight in serum-free medium containing 1% bovine serum albumin before lysis. For imaging experiments, 3T3-F442A cells were plated on glass coverslips and transfected with 2.5 g of cDNA expression vector using Transfast (Promega) according to the protocol recommended by the manufacturer. Approximately 36 h after transfection, cells were incubated overnight in serum-free medium containing 1% bovine serum albumin, treated with ligands at 37°C, and processed for imaging as described below.
In Vitro Kinase Assay-In vitro kinase assays were performed as described previously (24). SH2-B␤ was immunoprecipitated with ␣SH2-B␤ and immune complexes were collected using protein A-agarose. Bound proteins were washed twice with lysis buffer (see above) and once with kinase buffer (50 mM Hepes, pH 7.6, 5 mM MnCl 2 , 0.5 mM dithiothreitol, 100 mM NaCl, 1 mM Na 3 VO 4 ). Immunoprecipitates were incubated at 30°C for 30 min in 50 l of kinase buffer containing 0.5 mCi of [␥-32 P]ATP, 10 g/ml aprotinin, and 10 g/ml leupeptin. Immunoprecipitates were washed five times with 500 l of lysis buffer. Proteins were eluted by boiling in a mixture (80:20) of lysis buffer and SDS-PAGE sample buffer. Proteins were then resolved by SDS-PAGE (5-12% gradient), transferred to nitrocellulose membrane, and visualized by autoradiography or phosphorimaging (Bio-Rad model 505).
In Vivo Labeling-293T cells were transfected as described above. Twenty-four hours after transfection, cells were washed with phosphate-free DMEM containing 1% bovine serum albumin. Cells were treated with 1 mCi of [ 32 P]orthophosphate (ICN) for 4 h followed by pervanadate for 6 or 30 min as indicated. Pervanadate was prepared by mixing 430 l of 100 mM Na 3 VO 4 with 10 l of 30% H 2 O 2 and incubating at room temp for 20 min. The solution was cooled on ice and added to cell medium to yield a final concentration of 100 M Na 3 VO 4 , 200 M H 2 O 2 . Cells were lysed. SH2-B␤ was immunoprecipitated, resolved by SDS-PAGE, and transferred to nitrocellulose as described above.
Phosphopeptide Mapping and Phosphoamino Acid Analysis-Twodimensional phosphopeptide mapping and phosphoamino acid analysis were performed as described (25). Briefly, nitrocellulose containing 32 P-labeled SH2-B␤ labeled in vivo or in vitro (see above) were washed twice with deionized H 2 O, soaked in 500 l of 0.5% polyvinylpyrrolidone in 100 mM acetic acid at 37°C for 30 min, washed five times with deionized H 2 O, and digested with 10 g of methylated trypsin for 4 h at 37°C. Approximately 85-90% of counts were recovered. Next, digested peptides were lyophilized, oxidized with performic acid, and re-lyophilized. Peptides were separated by thin-layer electrophoresis at pH 8.9 (in vitro) or pH 3.5 (in vivo) followed by a second dimension in thin-layer chromatography using phosphochromatography buffer (25). For phosphoamino acid analysis, 32 P-labeled peptides were scraped from the cellulose plates and eluted from the cellulose with pH 1.9 buffer. Eluted peptides or full-length SH2-B␤ were subjected to acid hydrolysis in 6 N HCl at 110°C for 60 min and resolved by thin layer electrophoresis in buffer at pH 3.5 containing 0.5 mM EDTA. Phosphoamino acid standards were visualized by ninhydrin, and radioactive spots were visualized by autoradiography or using a PhosphorImager (Bio-Rad model 505).
Assessment of Membrane Ruffling-To measure the effect of SH2-B␤ on membrane ruffling, cells expressing GFP-SH2-B␤ or GFP-SH2-B␤(Y439F,Y494F) were deprived of serum overnight and treated with GH as indicated in the figure legends. Cells were rapidly rinsed three times with PBS (10 mM sodium phosphate, pH 7.4, 150 mM NaCl) and fixed for 30 min at room temperature in 4% formaldehyde in PBS. Cells were permeabilized with 0.1% Triton X-100 in PBS for 15 min at room temperature and rinsed three times in PBS. Filamentous actin was stained by incubating samples with Texas Red-phalloidin (1:60) for 30 min at room temperature. Coverslips were then rinsed three times with PBS, mounted on slides, and imaged the same day. Transfected cells expressing GFP-tagged forms of SH2-B␤ were located with a fluorescein isothiocyanate filter set using a Nikon TE200 microscope. The number of ruffles, assessed as a concentration of F-actin at a plasma membrane protrusion, per transfected cell was determined. Each transfection was repeated three times with similar results. Between 30 and 119 untransfected cells or cells positive for GFP, GFP-SH2-B␤, or GFP-SH2-B␤(Y439F,Y494F) from the combined three experiments were scored for the presence of ruffles for each experimental condition.

JAK2 Phosphorylates Tyrosines 439 and 494 in SH2-B␤ in
Vitro-Two-dimensional peptide mapping was used to identify which of the one or more tyrosines in SH2-B␤ are phosphorylated by JAK2. Each of the nine tyrosines in SH2-B␤ was individually mutated to phenylalanine within the context of full-length, myc-tagged SH2-B␤ ( Fig. 1). Wild type and mutant forms of myc-tagged SH2-B␤ were co-expressed with JAK2 in 293T cells. SH2-B␤ was immunoprecipitated with ␣SH2-B␤. The immobilized SH2-B␤⅐JAK2 complex was incubated with [␥-32 P]ATP in an in vitro kinase assay. Radiolabeled proteins were separated by SDS-PAGE, transferred to nitrocellulose, and visualized by autoradiography. As reported previously (2, 3), JAK2 is constitutively active when overexpressed in 293T cells and co-precipitates with SH2-B␤ ( Fig. 2A). Wild type SH2-B␤ was phosphorylated in vitro ( Fig. 2A, lane 1), presumably by JAK2, because SH2-B␤ is not phosphorylated when JAK2 is not overexpressed (data not shown). Each of the individual tyrosine to phenylalanine mutants of SH2-B␤ was also phosphorylated in the kinase assay ( Fig. 2A, lanes 2-10), suggesting that SH2-B␤ is phosphorylated on more than one tyrosine. SH2-B␤ with tyrosine 439 mutated to phenylalanine (SH2-B␤(Y439F)) migrated faster than wild type or other mutant forms of SH2-B␤ ( Fig. 2A, lane 5). This faster migration could be due to the loss of phosphorylation of Tyr-439.
The region of the nitrocellulose containing 32 P-labeled SH2-B␤ was excised and digested with trypsin. Tryptic peptides were oxidized in performic acid and then separated first by thin layer electrophoresis and then in the second dimension by ascending chromatography. 32 P-Labeled peptides were visualized by autoradiography. Wild type SH2-B␤ contained two highly phosphorylated peptides (Fig. 2B, panel i, Spots 1 and 2) and several more peptides that are phosphorylated less robustly. Darker exposures of the map of wild type SH2-B␤ revealed at least five more spots (data not shown). The phosphopeptide map generated from SH2-B␤ overexpressed in COS-7 cells resembles that seen with SH2-B␤ expressed in 293T cells (Fig. 2B, panel i versus panel ii). Thus, two-dimen-sional maps of SH2-B␤ from both 293T and COS-7 cells reveal that JAK2 phosphorylates SH2-B␤ in vitro on at least two peptides.
To determine if peptides 1 and 2 generated from two-dimensional mapping of SH2-B␤ are phosphorylated on tyrosine(s), serine(s), or threonine(s), we performed phosphoamino acid analysis. Peptides 1 and 2 were isolated from Fig. 2B (panel i).
As a comparison, we also performed phosphoamino acid analysis on the tryptic digest of full-length, wild type SH2-B␤ used for this panel. Phosphoamino acid analysis revealed that fulllength SH2-B␤ is phosphorylated primarily on tyrosine residues in vitro (Fig. 2C, lane 1). SH2-B␤ was minimally phosphorylated on serines and threonines. The 32 P incorporated into Spots 1 (Fig. 2C, lane 2) and 2 (Fig. 2C, lane 3) contained exclusively phosphotyrosine. No 32 P co-migrating with phosphoserine or phosphothreonine was detected. We also performed phosphoamino acid analysis on the five lighter spots that are visible in darker exposures (48 h) of maps of SH2-B␤ (data not shown). Each of these spots contained 32 P-labeled phosphoserine and phosphothreonine as well as phosphotyrosine (data not shown).
Two-dimensional phosphopeptide mapping of SH2-B␤ in which individual tyrosines were mutated to phenylalanine (Fig. 2B, panels iii-xi) yielded maps similar to wild type SH2-B␤ with two exceptions. In the maps of SH2-B␤(Y439F), Spot 1 completely disappears (Fig. 2B, panel vi) suggesting that in wild type SH2-B␤, this spot corresponds to a peptide containing phosphorylated Tyr-439. Spot 2 disappears when Tyr-494 is mutated suggesting that Tyr-494 is phosphorylated by JAK2 in vitro (Fig. 2B, panel vii). Neither the presence nor the migration of any of the lighter spots reproducibly changed when maps of wild type and all mutant forms of SH2-B␤ were compared, suggesting that these lighter spots are not derived from SH2-B␤. Taken together, the data from phosphopeptide mapping of wild type and mutant forms of SH2-B␤ as well as phosphoamino acid analysis reveal that JAK2 phosphorylates SH2-B␤ on tyrosines 439 and 494 in vitro.
JAK2 Phosphorylates Tyrosines 439 and 494 in SH2-B␤ in Vivo-To provide insight into whether Tyr-439 and Tyr-494 are also phosphorylated in vivo, JAK2 was co-expressed with wild type SH2-B␤, SH2-B␤(Y439F), SH2-B␤(Y494F), or the double mutant SH2-B␤(Y439F,Y494F) in 293T (Fig. 3A) or COS-7 ( Fig. 3B) cells. SH2-B␤ was immunoprecipitated and Western blotted with ␣PY. SH2-B␤ is not tyrosyl-phosphorylated when expressed alone (see Fig. 9B, lanes 1 and 2, below). As reported previously (3), SH2-B␤ is tyrosyl-phosphorylated by JAK2 when they are co-expressed in either 293T or COS-7 cells (Fig.  3, A and B, lane 2). When Tyr-439 is mutated to phenylalanine, the tyrosyl phosphorylation of SH2-B␤ decreased compared with wild type in both cell types tested (Fig. 3, A and B, lane 3 versus lane 2). SH2-B␤(Y439F) migrates faster than wild type SH2-B␤, consistent with a lower level of phosphorylation. SH2-B␤(Y494F) has, at best, only a modest reduction in tyrosyl phosphorylation compared with wild type SH2-B␤ (Fig. 3, A and B, lane 4 versus lane 2). When both Tyr-439 and Tyr-494 were mutated to phenylalanine within the context of fulllength SH2-B␤ (SH2-B␤(Y439F,Y494F)) (Fig. 3, A and B, lane  5), the tyrosyl phosphorylation of SH2-B␤ was similar to or less than in cells expressing wild type SH2-B␤ or cells expressing SH2-B␤ lacking Tyr-439 or Tyr-494 individually. These data indicate that Tyr-439 in SH2-B␤ is phosphorylated by JAK2 in vivo. Whether or not Tyr-494 is also phosphorylated is more difficult to discern. The lack of a reproducible, substantial decrease in the ␣PY signal when Tyr-494 is mutated may indicate that 4G10 ␣PY does not recognize pY494, that phosphorylation at Tyr-494 is very labile or undergoes dephospho- rylation after the cells are lysed, that Tyr-494 is not phosphorylated to as great an extent as Tyr-439, that the relatively small decrease in overall phosphorylation upon mutation of Tyr-494 cannot be reproducibly detected by Western blotting with ␣PY, or that Tyr-494 is not phosphorylated by JAK2 in intact cells.
To determine more definitively whether Tyr-494 in SH2-B␤ is phosphorylated by JAK2 in vivo, and to confirm that Tyr-439 in SH2-B␤ is phosphorylated by JAK2 in vivo, we used [ 32 P]orthophosphate to label SH2-B␤ in vivo and subjected 32 P-labeled SH2-B␤ to phosphopeptide mapping. To maximize the amount of 32 P-labeled SH2-B␤ recovered, cells were treated prior to lysis with the phosphatase inhibitor pervanadate for 6 min (Fig. 4, A-C) or as long as 30 min (Fig. 4, D and E). We anticipated that SH2-B␤ might be phosphorylated on serines and threonines as well as tyrosines in vivo (1). Therefore, the first dimension (thin layer electrophoresis) was run at pH 3.5 rather than pH 8.9 as seen in Fig. 2, because we predicted that pH 3.5 would provide greater resolution of multiply phosphorylated peptides from SH2-B␤. The map of 32 P-labeled SH2-B␤ expressed in the absence of JAK2 contained two detectable phosphopeptides after a 12-h exposure (Fig. 4A, Spots a and b). A longer exposure (21 h) of this map reveals one additional peptide (Fig. 4B, Spot c). When SH2-B␤ was co-expressed with JAK2, the map of SH2-B␤ contained at least two additional phosphopeptides (Fig. 4, C and D, Spots 1 and 2). Migration of Spots 1 and 2 is similar to migration of peptides containing pY439 and pY494, respectively, observed in maps of SH2-B␤ phosphorylated in vitro. To confirm that Spots 1 and 2 comigrated with peptides containing pY439 and pY494 labeled in vitro, the tryptic peptides from in vivo labeled SH2-B␤ were mixed with the tryptic peptides from in vitro labeled SH2-B␤ and run on the same thin layer chromatography plate. Peptides 1 and 2 from in vitro labeled SH2-B␤ comigrated (Fig. 4E) with Spots 1 and 2 in the in vivo map of 32 P-labeled SH2-B␤. These  2 and 3). As a comparison, fulllength, wild type SH2-B␤ (panel A, lane 1) was also acid hydrolyzed. Amino acids were separated by thin layer electrophoresis and visualized by autoradiography. Migration of phosphoserine (pSer), phosphothreonine (pThr), and phosphotyrosine (pTyr) standards is indicated. data indicate that Tyr-439 and Tyr-494 in SH2-B␤ are the primary tyrosines within SH2-B␤ phosphorylated by JAK2 in vivo. We next asked if SH2-B␤ is phosphorylated on any other tyrosine residues in vivo. Phosphoamino acid analysis revealed that Spots a, b, and c contained 32 P exclusively in one or more phosphorylated serines (Fig. 5). These data are consistent with earlier reports using 3T3-F442A cells, which indicated that SH2-B␤ is Ser/Thr-phosphorylated in the absence of ligand stimulation (1). Alternatively, Spots a, b, and c may represent peptides from one or more other proteins that co-precipitates and co-migrates with SH2-B␤. Taken together, Figs. 1-5 demonstrate that overexpressed, constitutively active JAK2 phosphorylates SH2-B␤ on Tyr-439 and Tyr-494 both in vivo and in vitro.
Tyrosyl Phosphorylation of SH2-B␤ Is Not Required for the Ability of SH2-B␤ to Activate JAK2-We next asked if tyrosyl phosphorylation of SH2-B␤ by JAK2 is required for the ability of SH2-B␤ to increase the activity of JAK2. We have shown previously that overexpression of SH2-B␤ increases the kinase activity of JAK2 up to 20-fold (2), establishing SH2-B family members as the only known cytoplasmic activators of JAK2. Consistent with the ability of SH2-B␤ to activate JAK2, SH2-B␤ also increases the tyrosyl phosphorylation of JAK2 and its downstream target, STAT5B, in response to GH, a potent stimulator of JAK2. To determine if phosphorylation of SH2-B␤ by JAK2 is required for the ability of SH2-B␤ to activate JAK2, JAK2 was expressed in 293T cells alone or co-expressed with wild type or mutant forms of SH2-B␤ lacking Tyr-439 and/or Tyr-494, or all nine tyrosines. Proteins from cell lysates were Western-blotted with ␣PY. When JAK2 was expressed alone, JAK2 as well as other cellular proteins were tyrosyl-phosphorylated (Fig. 6, lane 1). When SH2-B␤ was co-expressed with JAK2, the phosphorylation of JAK2 and other cellular proteins was increased compared with when JAK2 was expressed alone, indicating that the activity of JAK2 was increased (Fig. 6, lane  1 versus lane 2). The phosphorylation of JAK2 and other cellular proteins was similarly increased when JAK2 was co-expressed with SH2-B␤(Y439F), SH2-B␤(Y494F), SH2-B␤ (Y439F,Y494F), or SH2-B␤(9YF) (Fig. 6, lane 1 versus lanes  3-6). The phosphorylation of JAK2 was slightly greater when JAK2 was co-expressed with SH2-B␤(Y494F) or SH2-B␤(9YF) (Fig. 6, lanes 4 and 6, respectively) compared with wild type SH2-B␤ (Fig. 6, lane 2). This difference was not seen in other experiments and most likely reflects higher levels of JAK2 in those samples (Fig. 6, bottom panel). Taken together, these data indicate that tyrosyl phosphorylation of SH2-B␤ does not affect the ability of SH2-B␤ to increase JAK2 activity.
SH2-B␤(Y439F,Y494F) Inhibits GH-induced Membrane Ruffling-GH, a potent activator of JAK2, stimulates actin reorganization and membrane ruffling (15,16). Membrane ruffles are found on the leading edge of motile cells and are thought to be required for cell motility (26). SH2-B␤ has been shown to be required for maximal GH-induced membrane ruffling and cell motility (15,16). To determine if Tyr-439 and Tyr-494 in SH2-B␤ are required for GH-stimulated membrane ruffling, 3T3-F442A murine fibroblasts were transfected with cDNA encoding GFP-SH2-B␤, GFP-SH2-B␤(Y439F,Y494F), or GFP as a control. Cells were deprived of serum overnight and treated with vehicle or 50 ng/ml of GH for 15 min. Filamentous actin was stained with Texas Red-phalloidin, and the number of ruffles per cell was counted. In the absence of GH, most of the cells had either no ruffles or one ruffle (Fig. 7, A, E, and I).
Overexpression of GFP alone, GFP-SH2-B␤, or GFP-SH2-B␤(Y439F,Y494F) had no effect on the number of ruffles in these unstimulated cells. Addition of GH stimulated ruffling of cells (Fig. 7, C, G, and I) as shown previously (15). The average number of ruffles per cell (ϳ2) was similar in untransfected cells versus cells overexpressing GFP or GFP-SH2-B␤. These data are consistent with earlier results indicating that at 50 ng/ml GH, ruffling is maximal and overexpression of GFP-SH2-B␤ fails to increase ruffling further (15). In contrast to  5. SH2-B␤ is phosphorylated on serines in vivo. Phosphorylated peptides a, b, and c scraped from the two-dimensional map of SH2-B␤ in vivo (Fig. 4D) were eluted and subjected to acid hydrolysis. Amino acids were separated by thin layer electrophoresis and visualized by autoradiography. Migration of phosphoserine (pSer), phosphothreonine (pThr), and phosphotyrosine (pTyr) standards is indicated. these data, overexpression of GFP-SH2-B␤(Y439F,Y494F) blocked GH-induced cell ruffling (Fig. 7, E-H and I). Thus, GFP-SH2-B␤(Y439,494) acts as a dominant negative suggesting that Tyr-439 and/or Tyr-494 within SH2-B␤ are required for GH-induced membrane ruffling.
To determine if any additional tyrosines in SH2-B␤ are phosphorylated by JAK1, all nine tyrosines in SH2-B␤ were mutated to phenylalanine. SH2-B␤(9YF) was co-expressed in 293T cells with JAK1, or for comparison, JAK2. The small amount of phosphorylation associated with SH2-B␤(Y439F,Y494F) was no longer detected when SH2-B␤(9YF) was co-expressed with JAK1 (Fig. 8, lane 5 versus lane 4) suggesting that JAK1 may phosphorylate one or more tyrosines in SH2-B␤ in addition to Tyr-439 and Tyr-494. However, when SH2-B␤(9YF) was coexpressed with JAK2, tyrosyl phosphorylation of SH2-B␤(9YF) was similarly reduced (when the ␣ PY signal was normalized to amount of expressed SH2-B␤) (Fig. 8, lane 8 versus lane 7). The inability to detect phosphorylated tyrosines other than Tyr-439 and Tyr-494 in the more definitive two-dimensional phosphopeptide maps of SH2-B␤ from cells co-expressing JAK2 (Figs. 2 and 4) suggest that the reduced amount of ␣PY binding to the region corresponding to SH2-B␤ when all 9 tyrosines are mutated may reflect low level phosphorylation of tyrosines other than Tyr-439 and Tyr-494. It is also useful to note that when SH2-B␤(9YF) is co-expressed with JAK2, a band that co-migrates with SH2-B␤ is detected by immunoblotting with ␣PY (Fig. 8, lane 8). Given that all 9 tyrosines are mutated in the overexpressed SH2-B␤(9YF), it is unlikely that this tyrosylphosphorylated protein is overexpressed SH2-B␤(9YF). This band may be endogenous SH2-B␤ that is expressed at very low levels in 293T cells. Alternatively, this band may represent a comigrating protein that is tyrosyl-phosphorylated by JAK2.
SH2-B␤ Is Phosphorylated on Tyr-439 by PDGFR-In addition to JAK1 and JAK2, other tyrosine kinases have been reported to tyrosyl-phosphorylate SH2-B␤, including the receptor tyrosine kinase, PDGFR (6,7). We next sought to determine whether PDGFR phosphorylates Tyr-439 and/or Tyr-494 in SH2-B␤. SH2-B␤ was expressed alone or with PDGFR (␤ subunit) in COS-7 cells. Cells were deprived of serum overnight and stimulated with 25 ng/ml PDGF BB for 15 min. Proteins from cell lysates were Western-blotted with ␣PY. In the ab- sence of overexpressed PDGFR, minimal tyrosyl phosphorylation of proteins was detected in either the absence of presence of PDGF (Fig. 9A, lanes 1 and 2). After transfection of as little as 0.1 g of cDNA encoding PDGFR, a slight increase in the phosphorylation of cellular proteins was detected even in the absence of PDGF (Fig. 9A, lanes 3 and 4). In particular, a ligand-independent band that co-migrates with PDGFR is visible when Western-blotted with ␣PY, suggesting that PDGFR, like JAK1 and JAK2, is autophosphorylated and constitutively active when overexpressed. These data are consistent with other studies in which PDGFR is overexpressed (2,27). Stimulation with PDGF failed to increase further the phosphorylation of cellular proteins or PDGFR (Fig. 9A, lane 4). When 2.0 g of PDGFR cDNA was transfected, the phosphorylation of PDGFR and cellular proteins was significantly increased (Fig.  9A, lanes 5 and 6). In particular, a band that co-migrates with SH2-B␤ is detected with ␣PY even in the absence of PDGF stimulation, suggesting that SH2-B␤ is phosphorylated by the constitutively active PDGFR (Fig. 9A, lanes 5 versus 6).
To more closely examine the phosphorylation of SH2-B␤ by PDGFR, SH2-B␤ was immunoprecipitated with ␣SH2-B␤ and Western blotted with ␣PY. In the absence of PDGFR, SH2-B␤ was not detectably tyrosyl-phosphorylated whether PDGF was present or not (Fig. 9B, lanes 1 and 2). When SH2-B␤ was co-expressed with PDGFR, SH2-B␤ was tyrosyl-phosphorylated in the absence of PDGF (Fig. 9B, lane 3). Tyrosyl phosphorylation of SH2-B␤ was detected even when as little as 0.1 g of PDGFR cDNA was transfected into the cells (data not shown). Stimulation with PDGF modestly increased the tyrosyl phosphorylation of SH2-B␤ (Fig. 9B, lanes 3 versus 4).

JAK2 Phosphorylates Tyr-439 and Tyr-494 in SH2-B␤-We
have demonstrated previously that SH2-B␤ is tyrosyl-phosphorylated by constitutively active, overexpressed JAK2, as well as by endogenous, GH-stimulated JAK2 (1, 3, 24). In the current study, we have identified Tyr-439 and Tyr-494 in SH2-B␤ as targets of JAK2 both in vitro and in vivo. This conclusion is based primarily on two-dimensional phosphopeptide mapping of SH2-B␤ labeled in vitro and two-dimensional phosphopeptide mapping of SH2-B␤ labeled in vivo. Phosphopeptide maps of SH2-B␤ overexpressed with JAK2 in 293T cells and phosphorylated by co-precipitating JAK2 in an in vitro kinase assay show that SH2-B␤ is predominantly phosphorylated by JAK2 on two peptides. These peptides are also phosphorylated by JAK2 when SH2-B␤ and JAK2 are co-expressed in COS-7 cells. Phosphoamino acid analysis confirmed that these two peptides were phosphorylated exclusively on tyrosines. Mutation of Tyr-439 to phenylalanine eliminated one of these spots in phosphopeptide maps of SH2-B␤ and mutation of Tyr-494 eliminated the other spot. Mutation of any of the remaining tyrosines did not affect the phosphopeptide map. The simplest explanation for these results is that wild type SH2-B␤ is phosphorylated on Tyr-439 and Tyr-494.
Darker exposures of the maps of wild type and mutant forms of SH2-B␤ labeled in vitro show that each of the maps contained at least five other spots. If any of these spots were peptides from SH2-B␤ containing phosphotyrosine, we would expect the spots to either disappear or shift when tyrosines in SH2-B␤ are mutated to phenylalanine. However, none of these spots were reproducibly altered by mutations of tyrosines in SH2-B␤ indicating that these spots do not contain phosphotyrosine(s) from SH2-B␤. Digestion of SH2-B␤ with protein phosphatase 2A, a serine/threonine phosphatase, reduces the broadly migrating SH2-B␤ to a tighter migrating band indicating that SH2-B␤ is highly phosphorylated on serines and threonines in 293T cells as well as in 3T3-F442A and PC12 cells (1,7,28). Thus, one possible explanation would be that some or all of the other spots correspond to peptides containing 32 P-labeled phosphoserine or phosphothreonine. However, phosphoamino acid analysis of these spots indicated that they contain phosphotyrosine and, to a lesser degree, phosphoserine and phosphothreonine (data not shown). Thus, these spots are likely to be peptides from other protein(s) that co-migrate with SH2-B␤.
Maps of SH2-B␤ labeled in vivo indicate that both Tyr-439 and Tyr-494 in SH2-B␤ are phosphorylated by JAK2, whereas results obtained from examining the phosphorylation of SH2-B␤(Y494) with ␣PY were variable. The tyrosyl phosphorylation of SH2-B␤(Y494) was less than that of wild type SH2-B␤, and the phosphorylation of SH2-B␤(Y439F,Y494F) was decreased compared with that of SH2-B␤(Y439F) in 293T cells overexpressing JAK2 but not reproducibly so in COS-7 cells. This variability in detection of pY494 may reflect its relatively lower amount of phosphorylation compared with Tyr-439 observed in the maps. JAK2 may have a lower affinity for Tyr-494 than for Tyr-439, or the phosphorylation of Tyr-494 may be more transient than that of Tyr-439, perhaps indicating that Tyr-494 is targeted by a phosphatase. Alternatively, pY494 may be poorly recognized by ␣PY. Anti-phosphotyrosine antibodies have been observed to exhibit variability in their ability to detect phosphotyrosines residing in different amino acid environments (29). For example, Tyr-343 in the cytoplasmic domain of the receptor for erythropoietin has been reported to be phosphorylated but not recognized by ␣PY (30,31). Thus, for whatever reason, the phosphorylation at Tyr-494 in SH2-B␤, which is readily detectable in two-dimensional peptide maps from in vivo and in vitro labeled SH2-B␤, is difficult to discern on Western blots using 4G10 ␣PY.
SH2-B␤ Is Phosphorylated on Serine(s) in Vivo-Phosphoamino acid analysis of peptides from SH2-B␤ labeled in vivo support the hypothesis that SH2-B␤ is phosphorylated on serines in addition to being phosphorylated on tyrosines. SH2-B␤ contains 82 serines. Many of these serines lie within consensus target sequences for ERKs 1 and 2, protein kinase C, and other Ser/Thr kinases. NGF stimulates the Ser/Thr phosphorylation of SH2-B␤ in PC12 cells. An inhibitor of MEK, the kinase that phosphorylates ERKs 1/2, prevents most of the NGF-induced Ser/Thr phosphorylation of SH2-B␤ consistent with Ser/Thr kinases downstream of MEK phosphorylating SH2-B␤. SH2-B␤ is phosphorylated on Ser-96 by ERKs 1 and 2 in vitro. However, mutation of Ser-96 to alanine failed to abolish Ser/Thr phosphorylation of SH2-B␤ in response to NGF. Activation of PKC by PMA also stimulates Ser/Thr phosphorylation of SH2-B␤ in PC12 cells. However, down-regulation of PKC by prolonged treatment with PMA did not affect NGFpromoted phosphorylation of SH2-B␤ on Ser/Thr, suggesting that activation of PKC is not required for NGF-induced Ser/Thr phosphorylation of SH2-B␤ (28). These data suggest that SH2-B␤ is phosphorylated on more than one Ser/Thr, consistent with our observation here of more than one SH2-B␤ peptide containing 32 P-ser. Further investigation will be required to determine whether ERKs 1 and 2, PKC, or other Ser/Thr kinases phosphorylate SH2-B␤ constitutively or in response to activation of JAK2 as well as to identify which serine(s) in SH2-B␤ is(are) phosphorylated.
SH2-B␤ Is Phosphorylated on Tyr-439 by JAK1 and PDGFR-We examined the tyrosyl phosphorylation of SH2-B␤ by another Janus family kinase, JAK1, as well as by a receptor tyrosine kinase, PDGFR, to see whether the different kinases phosphorylated the same or different tyrosines within SH2-B␤ as JAK2. One can envision that different tyrosines in SH2-B␤ phosphorylated by JAK2 or JAK1 or PDGFR could form binding sites for different proteins and lead to the activation of signaling pathways specific to PDGF or JAK1 or JAK2. Western blotting of mutant SH2-B␤ with ␣PY suggests that the primary site in SH2-B␤ phosphorylated by JAK1 and PDGFR in vivo is Tyr-439 as it is for JAK2. As previously discussed, phosphorylation at Tyr-494 is not easy to detect with ␣PY, however, Tyr-494 is also likely to be phosphorylated by JAK1, because phosphorylation of SH2-B␤(Y439F,Y494F) was decreased relative to SH2-B␤(Y439F). Similar results were not obtained with PDGFR, suggesting that Tyr-494 may not be as good a target of PDGFR. Other tyrosines may also be phosphorylated by JAK2, JAK1, and PDGFR based on the further decrease in the ability of ␣PY to recognize a protein corresponding in migration to SH2-B␤ when these tyrosine kinases are co-expressed with SH2-B␤ lacking all nine of its tyrosines. At least in the case of JAK2, such phosphorylation is not likely to be substantial, because it was not observed in the phosphopeptide maps.
The Function of Tyrosyl Phosphorylation of SH2-B␤-Presumably, tyrosyl phosphorylation of SH2-B␤ on Tyr-439 and Tyr-494 serves some function. We have shown that SH2-B␤ increases the kinase activity of JAK2 thereby increasing the autophosphorylation of JAK2 as well as the phosphorylation of targets of JAK2 such as STAT5B (2). Here we demonstrate that tyrosyl phosphorylation of SH2-B␤ is not required to activate JAK2. SH2-B␤ with individual tyrosines mutated or all nine tyrosines mutated to phenylalanine increase the phosphorylation of JAK2 to levels similar to those seen when JAK2 is co-expressed with wild type SH2-B␤. These data are consistent with previous results suggesting that the C terminus of SH2-B␤ (amino acids 504 -670), which lacks Tyr-439 and Tyr-494, is necessary and sufficient for SH2-B␤ to increase the activity of JAK2 (24).
Previously, we have shown that overexpression of SH2-B␤ increases GH-stimulated membrane ruffling. Overexpression of SH2-B␤ lacking an arginine in the SH2 domain critical for binding to phosphotyrosines (SH2-B␤(R555E)) inhibited membrane ruffling. Further investigation revealed that the second proline-rich region in the N terminus of SH2-B␤ (Fig. 1, amino acids 85-106) is required for binding of SH2-B␤ to the small GTPase, Rac, and for cell motility. These data suggest that multiple domains in SH2-B␤ are required for the role of SH2-B␤ in membrane ruffling. Here, we demonstrate that Tyr-439 and/or Tyr-494 in SH2-B␤ are also required for GHin-duced membrane ruffling. Expression of SH2-B␤-(Y439F,Y494F) in 3T3-F442A cells inhibits the ruffling normally observed in response to GH but not basal ruffling observed in the absence of GH. Mutating Tyr-439 and Tyr-494 to phenylalanine had no effect upon the ability of SH2-B␤ to activate JAK2, thus the dominant negative effect of SH2-B␤(Y439F,Y494F) suggests that GH-dependent phosphorylation of SH2-B␤ at either Tyr-439 and/or Tyr-494 forms binding site(s) for one or more signaling molecules that play a critical role in the formation of cell ruffles.
Interestingly, both Tyr-439 and Tyr-494 reside within the consensus sequence of YXXL. Thus, candidate interacting proteins should bind phosphorylated tyrosines in the sequence of YXXL. One such candidate protein is CrkII. CrkII is an adapter protein implicated in the regulation of the actin cytoskeleton and known to be tyrosyl-phosphorylated in response to stimulation by GH (32,33). The preferred binding motif for the SH2 domain of CrkII is YXXP, however, the Crk SH2 domain has some affinity for pYXXL-containing sequences (34,35). We overexpressed hemagglutinin-tagged CrkII with JAK2 in the presence and absence of SH2-B␤, immunoprecipitated SH2-B␤, and Western-blotted against hemagglutinin. Our preliminary data suggest that CrkII does not associate with SH2-B␤. Further studies will be required to identify other cytoskeletal proteins that may bind pY439 and/or pY494 in SH2-B␤.
Another protein candidate binding partner for tyrosyl-phosphorylated SH2-B␤ is the protein-tyrosine phosphatase, SHP-2. SHP-2 has not yet been implicated in the regulation of the actin cytoskeleton; however, in response to GH, SHP-2 has been found to associate with GH receptor and signal-regulatory protein-1␣. SHP-2 requires pY449 and pY473 in SIRP for association. These two tyrosines are found within the sequence of YXXL (Tyr-449) and the related sequence of YXXV (Tyr-473) (36). Similarly, SHP-2 requires pY487 and pY595 in GH receptor for maximal association with GH receptor. Both of these tyrosines are found within the similar sequence of YXXV (37). To determine if SHP-2 also associates with SH2-B␤, SH2-B␤ was co-expressed with JAK2, SH2-B␤ was immunoprecipitated with ␣SH2-B␤, and immunoprecipitated proteins were Western-blotted with ␣SHP-2. This approach did not detect association of SHP-2 with SH2-B␤. Because Tyr-439 and Tyr-494 are both within the consensus sequence of YXXL, one can envision that they bind proteins that recognize ITAMs (immunoreceptor tyrosine-based activation motif). However, the traditional definition of an ITAM is two YXXL sequences separated by 8 -10 amino acids (38). Tyr-439 and Tyr-494 are separated by too many amino acids to form a traditional ITAM binding site. The three-dimensional structure may place Tyr-439 and Tyr-494 within a close enough proximity to form an ITAM. Further studies will be needed to identify proteins that associate with pY439 and/or pY494 in SH2-B␤.
In summary, using two-dimensional phosphopeptide mapping, we have determined that Tyr-439 and Tyr-494 in SH2-B␤ are phosphorylated by JAK2 in vivo and in vitro. Further, Tyr-439 and Tyr-494 in SH2-B␤ are shown to play a significant role in the regulation of GH-dependent membrane ruffling. Finally, we demonstrate that JAK1 and PDGFR also phosphorylate Tyr-439. Therefore, Tyr-439 in SH2-B␤ may serve as a binding site for a similar set of signal proteins following stimulation by PDGF as well as by cytokines that activate JAK1 and/or JAK2.