w, or · 10468 sh-ptp2 sh2 domain specificity mains unclear. sh-ptp2 (p64) has been linked to ras...

THE JOURNAL OF BIOUXICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 14, Issue of April 8, pp. 10467-10474, 1994 Printed i n U.S.A.

SH-PTP2Byp SH2 Domain Binding Specificity Is Defined by Direct Interactions with Platelet-derived Growth Factor P-Receptor, Epidermal Growth Factor Receptor, and Insulin Receptor Substrate-1-derived Phosphopeptides"

(Received for publication, November 10, 1993, and in revised form, January 10, 1994)

Randi D. Case$, Elizabeth PiccioneSI, Gert Wolfsn, Anton M. Benettll, Robert J. Lechleiderfl**, Benjamin G. NeelllSS, and Steven E. ShoelsonS00 From the &Joslin Diabetes Center & Department of Medicine, Harvard Medical School, and the lwolecular Medicine Unit, Beth Israel Hospital, Harvard Medical School, Boston, Massachusetts 02215

Signaling by tyrosine kinases involves direct associations between proteins with Src homology 2 (SH2) domains and sites of tyrosine phosphorylation. Specificity in signaling pathways results in part from inherent se- lectivity in interactions between particular SH2 domains and phosphopeptide sequences. The cytoplasmic phosphotyrosine phosphatase SH-PTP2 (Syp, PTF' lD, PTPdC) contains two SH2 domains (N and C) which me- diate its association with and activation by the platelet- derived growth factor (PDGF) and epidermal growth factor receptors and IRS-1. We have developed a com- petitive phosphopeptide binding assay to analyze specificity of the SH-PTP2 N-SH2 domain for phosphorylation sites of these phosphoproteins. The sequence surrounding %'" bound with greatest affinity (ID, = 14 p ~ ) of eight PDGF receptor-derived phosphopeptides tested. No peptides corresponding to known epidermal growth factor receptor phosphorylation sites bound with high affinity. However, an alternative sequence surrounding -bound tightly (ID, = 21 p ~ ) . Of the 13 IRS-1-related peptides analyzed, sequences surrounding w, " y P a , and Tyr"" bound with highest affinity (ID, = 11,4, and 1 p ~ , respectively). Alternative phosphopeptides generally bound with much weaker affinity (ID, > 150 p ~ ) . These findings are consistent with recent mutational analyses of the PDGF receptor and predict site-specific interactions between SH-FTP2 and each of these phosphoproteins. Comparisons between peptide sequences suggest that the N-terminal SH2 domain of SH-PTP2 binds with highest affinity to phosphotyrosine (pY) followed by a P-branched residue (Val, ne, T h r ) at pY+1 and a hydrophobic residue (Val, Leu, ne) at pY+3 positions. Peptide truncation studies also indicate that residues

* This work was supported by grants from the National Science Foun- dation (to S. E. S.), National Institutes of Health ( to B. G. NJ, and the Juvenile Diabetes Foundation, International (to S. E. S.). The Biochem- istry Facility at the Joslin Diabetes Center is supported by National Institutes of Health Diabetes and Endocrinology Research Center Grant DK36836. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

8 Recipient of a student fellowship from the Juvenile Diabetes Foun- dation.

ll Supported in part by Boehringer Mannheim. ** Damon-Runyon Fellow. $$ Recipient of a Junior Faculty Research Award from the American

Cancer Society.

betes Foundation. To whom correspondence should be addressed: Joslin $0 Recipient of a Career Development Award from the Juvenile Dia-

Diabetes Center, One Joslin Place, Boston, MA 02215. "el.: 617-732- 2528; Fax: 617-732-2593.

outside of the pY-1 to pY+4 motif are required for high affinity interactions.

Activated tyrosine kinase receptors elicit many cellular effects, including changes in differentiation, growth rate, me- tabolism, and morphology. The initiating events in these important signaling pathways are understood in some detail, including ligand-induced receptor dimerization, tyrosine kinase activation, autophosphorylation of particular tyrosine residues within the receptors themselves, and the intermolecu- lar phosphorylation of tyrosines on cellular substrates (1). Un- til recently many of the more distal signaling events that con- nect these surface receptors with their intracellular, physiological targets were unknown. However, the last few years have witnessed rapid expansion in our understanding of how receptor and non-receptor tyrosine kinases function, largely related to the realization that particular tyrosine phos- phorylations serve as molecular switches to facilitate direct associations between kinases or kinase substrates and additional cytoplasmic proteins having SH2 (Src homology 2)' domains (2-4).

Tyrosine kinase activation leads to the formation of multi- component complexes comprising one or more phosphoprotein(s) associated with additional SH2 domain-containing proteins. For example, activated PDGF receptors associate with phosphatidylinositol (PI) 3-kinase, phospholipase C-y (PLC-y), Ras-associated GTPase-activating protein (RasGAP), c-Src and related kinases, an SH2 domain-containing protein tyrosine phosphatase (SH-PTPZ), and additional proteins (5-16). EGF receptors associate with PLC-y, RasGAP, SH-FTP2, and Grb2, a protein with both SH2 and SH3 domains thought to link tyrosine kinase receptors to Ras signaling pathways (6, 16-21). Fewer SH2 proteins associate directly with the insulin receptor and the complexes appear to be less stable (22). However, the insulin receptor phosphorylates IRS-1 (231, a cytoplasmic protein which associates avidly with an overlapping collection of SH2 proteins, including PI 3-kinase, SH-PTPS, and Grb2 (22- 29).

SH-PTP2 (also referred to as Syp, PTP lD, and pTP2C) in- teracts with each of these important phosphoproteins and likely contributes a fundamental function to the corresponding signaling complexes. However, its physiological function re-

The abbreviations used are: SH2, src homology 2; PDGF, platelet- derived growth factor; PI, phosphatidylinositol; PLC, phospholipase C; EGF, epidermal growth factor; GST, glutathione S-transferase; HPLC, high performance liquid chromatography; cpm, countdmin; PDGFR, PDGF receptor; IRS-1, insulin receptor substrate-1.

~______ ~~

10467

10468 SH-PTP2 SH2 Domain Specificity

mains unclear. SH-PTP2 (p64) has been linked to Ras activation but not mitogenesis per se (15). In contrast to SH-PTP1 (also named PTPlC, HCP, or SHP, a related SH2-containing phosphatase found in hematologic tissues), SH-PTP2 is widely expressed in many mammalian tissues throughout development (6). SH-FTP2 is phosphorylated following growth factor stimulation (6, 16, 30), and its catalytic activity is potently stimulated by SH2 domain occupancy (11)2 Certainly it must dephosphorylate phosphotyrosine residues, but it is not known whether the targets of SH-PTP2 catalysis reside within the SH2-bound phosphoprotein or adjacent phosphoproteins. A closely related SH2-containing PTPase in Drosophila, the cork- screw (csw) gene product (321, acts in concert with a c-raf ho- mologue to transmit signals downstream of the torso receptor tyrosine kinase (32, 33). SH-PTP2 may play an analogous role in mammals.

We are interested in learning more about the functions of SH-FTP2 and its various domains. In vitro, SH-PTP2 does not dephosphorylate intact EGF or PDGF receptors efficiently (16). Phosphopeptides corresponding to sites of phosphorylated tyrosines in the EGF and PDGF receptors are poor substrates, in general. Nonetheless, SH-PTP2 exhibits a 10-fold range of se- lectivity for phosphopeptide sequences, and the better substrates exhibit non-linear double-reciprocal plots (34). Catalyti- cally, SH-PTP2 is more active toward the artificial, phosphorylated protein substrate RCM-lysozyme (341, which suggests that the PTPase domain recognizes proteins more efficiently than short peptides. Potent stimulation of catalytic activity by SH2 domain occupancy ( 11)2 and the lack of Michae- lis-Menten kinetics observed with some phosphopeptide substrates (34) suggest a functional interplay between SH2 and PTPase domains. We previously used a degenerate phosphopeptide library to map SH-PTP2 N-SH2 domain specificity (35); the preferred sequence included Ile or Val at pY+1, no selection a t pY+2, and Val, Ile, Leu, or Pro at pY+3. Although the library method provides a sense of residues which are preferred at a given position, by design it gives no information about sequence specificity within a putative tetrapeptide motif or the relative importance of residues N- or C-terminal to the motif. We have now developed an SH2 domain-phosphopeptide binding assay which allows us to analyze relative affinities of individual sequences and predict site(s) within phosphoproteins which interact with SH-PTP2.

MATERIALS AND METHODS Phosphopeptide Synthesis and SH2 Domain Expression-

Phosphopeptides were prepared following a stepwise, Nu-Fmoc SF- thetic strategy using N"-Fmoc-0-(0,O-dimethoxyphosphory1)-L-tyrosine (Fmoc-Tyr(OP(OCH,),)) as described (36, 37). The N- and C-terminal SH2 domains corresponding to human SH-PTP2 residues 1-105 and 101-267 were expressed as glutathione S-transferase (GST) fusion proteins (30) and purified by affinity chromatography on glutathione-agarose. After passage of bacterial extracts over the affinity ma- trix, the columns were washed extensively with buffer A (10 mM sodium phosphate, 140 mM NaCl, 10 mM dithiothreitol (pH 7.4)), and proteins were eluted with 8.0 M urea (38). The urea was removed and the fusion proteins were allowed to refold by dialysis against buffer A. Purified GST/SH2 domain fusion proteins were stored at -20 "C until needed.

'*51-RadioZabeled Phosphopeptide-Benzene was removed from 2.0 mCi (1.0 pmol) of monoiodo-'251-Bolton-Hunter reagent (ICN) using a gentle stream of nitrogen gas. Phosphopeptide IRS-1 pY1172 (0.5 mg) in 20 pl of neat dimethyl sulfoxide was added to the dried residue. The apparent pH was adjusted to 7.0 as necessary with N-methylmorpho- line, and the reaction was allowed to proceed at room temperature for 24 h. The peptide was precipitated with excess ice-cold acetone and reaction products were separated by reversed-phase HPLC. The radiolabeled peptide eluted significantly later than the unmodified peptide

S. Sugimoto, T. Wandless, S. E. Shoelson, B. G . Neel, and C. T. Walsh, manuscript submitted.

due to loss of charge and addition of the hydrophobic rn-['251]iodo-p- hydroxyhydrocinnamoyl moiety of the Bolton-Hunter reagent. The ma- jor fraction of '251-phosphopeptide that exhibited SH2 binding typically represented 1@40% of the original amount of '251-Bolton-Hunter reagent (0.2-0.8 mCi; specific activity -300 Ci/mol). Bovine serum albumin (50 pg/aliquot) was added, the mixtures were lyophilized, and the radiolabeled peptide was stored at -20 "C until use.

Phosphopeptide-SH2 Domain Binding Assay"GSTISH2 domain fusion proteins (0.5-1.0 p, estimated by Bradford assay), 35 fmol of HPLC-purified, '251-Bolton-Hunter-treated phosphopeptide (67 nCi), and varying concentrations of unlabeled peptides were combined in 200-1.11 total volume of 20 mM Tris-HC1, 250 mM NaCl, 0.1% bovine serum albumin, 10 mM dithiothreitol, pH 7.4, and vortexed. Glutathi- one-agarose (25 pl of a 1:4 aqueous slurry, Molecular Probes) was added, and the samples were incubated overnight at 22 "C with constant mix- ing. Following centrifugation for 5 min at 12,000 x g, supernatant solutions were removed by aspiration, and '251-radioactivity associated with the unwashed pellets was determined with a y-counter.

The amount of GST/SH2 domain was titrated to bind 10-20% (10,000-20,000 cpm) of added '251-radioactivity (100,000 cpm; 35 fmol). This value was chosen because it provides high sensitivity and a high ratio of signal to noise, yet minimizes GSTISH2 domain use. In the absence of GST/SH2 domain fusion protein or in the presence of GSTI SH2 domain and excess (1.0 m ~ ) , unlabeled IRS-1 pY1172 phosphopeptide 1.5-3% of the '251-phosphopeptide remained associated with the aspirated glutathione-agarose beads. This value was taken to be non- specific binding (NSB). In competition experiments the amount of radioactivity associated with the pellets varied according to concentration of unlabeled peptides. These values were plotted as: Bound (% maximum) uersus log peptide concentration (Figs. 2-5), where Bound (% maximum) = (Bound - NSB)/(Maximum bound - NSB); data were best fit to the sigmoid dose-response equation, y = [a + (1 + e"")l + d, using the program SigmaPlot. Results demonstrate that inhibition of lZ5I-

radiolabeled peptide binding depended on phosphopeptide concentration (Figs. 2-51, with no inhibition at the low concentrations, 100% inhibition at the highest concentrations (of high affinity phosphopeptide), and partial inhibition in between. Half-maximal inhibition (ID5,, ?

S.D.) values are listed in Table I. In homologous competition assays (where the same peptide is used as the radiolabeled tracer and competitor), IRS-1 pY1172 exhibited potent inhibition of '251-IRS-1 pY1172 binding (ID, = 1.0 PM); affinities of additional phosphopeptides are also reported relative to IRS-1 pY1172 (Table I).

Phosphopeptide Blockade of Intact EGF and PDGF p ReceptorlSH- PTP2 SH2 Domain Association-ATWT and A431 cells were grown in Dulbecco's modified Eagle's medium medium containing 10% fetal bovine serum and antibiotics; 2 4 4 8 h prior to an experiment the cells were deprived of serum. Quiescent cells were stimulated for 10 min either with 50 ng/ml PDGF ( A m ) or 100 ng/ml EGF (A431) and lysed in non-ionic detergent; resulting solutions were clarified by centrifugation as described (11). ATWT orA431 cell lysates (250 or 500 pg proteid ml, respectively) were incubated with GSTN-SH2 domain (0.5-1.0 PM) on glutathione-agarose beads for 2 h at 4 "C. For phosphopeptide competition experiments, peptides were preincubated with the GSTRV-SH2 domain 30 min prior to addition of cell lysates. The agarose pellets were washed extensively and SH2 domain-bound proteins were separated by SDS-polyacrylamide gel electrophoresis on 7.5% gels, transferred to PVDF membranes (Immobilon, Millipore), and identified by immuno- blotting with the monoclonal anti-phosphotyrosine antibody 4G10 (the generous gift of Brian Drucker, Dana Farber Cancer Institute) as described (11).

RESULTS Phosphopeptide Design and Binding Assay Development-

Phosphopeptides listed in Table I correspond to known or putative phosphorylation sites within the PDGF receptor, EGF receptor, and IRS-1. Peptide residues at the +1, +2, and +3 positions relative to phosphotyrosine are important for conferring SH2 domain specificity (3, 5, 35, 37, 39). Crystallographic structural analyses of Lck and Src SH2 domaidphosphopeptide complexes suggest that important contacts may occur with peptide positions ranging from py-2 to pY+4, as well (40,41). In the design of peptides for these studies we attempted to include all residues which might be important for SH2 domain interactions. Most of the peptides contain 11 amino acids, with 3 residues N-terminal and 7 residues C-terminal to phosphoty-

EGF receptor

IRS-1

SH-PTP2 SH2 Domain Specificity 10469 TABLE I

Affinities between phosphopeptides and nonphosphorylated peptides and the N-terminal SH2 domain of SH-PTP2 Peptide sequences derived from the PDGF and EGF receptors and IRS-1 are numbered according to the position of tyrosine in the intact protein.

ID, and standard deviation (SD) values were calculated by Sigmaplot using the sigmoid dose response equation (see Results Section).

Parent protein Peptide name (pY position) Peptide sequence ID, k S.D. Relative

affinity"

W PDGF receptor PDGFR pY579 G-H-E-PY-I-Y-V-0-P-V-Q 49 t 6 0.022

PDGFR pY581 G-H-E-Y-I-pY-V-D-P-V-Q 224 t 42 0.004 PDGFR pY579181 G-H-E-pY-I-pY-V-D-P-V-Q 49 t 5 0.022 PDGFR pY740 D-G-G-pY-M-D-M-S-K-D-E >300 <0.003 PDGFR pY751 S-V-D-pY-V-P-M-L-D-M-K >300 <0.003 PDGFR pY771 S-S-N-pY-M-A-P-Y-D-N-Y 152 t 73 0.007 PDGFR pY1009* S-V-L-PY-T-A-V-Q-P-N-E 14 t 2 0.08 PDGFR Y 1009 S-V-L- Y-T-A-V-Q-P-N-E >300 <0.003 PDGFR pY1021 D-N-D-pY-I-I-P-L-P-D-P-K >300 0.004

EGFR pY920 I-D-V-pY-M-I-M-V-K-A 239 t 27 0.004 EGFR pY954 P-Q-R-pY-L-V-I-Q-G-D 21 ? 2.4 0.05 EGFR pY992 D-A-D-E-pY-L-I-P-Q-Q-G-F-F >300 <0.003 EGFR pY1068 V-P-E-PY-I-N-Q-S-V-P-K 249 t 31 0.004 EGFR pY1086 N-P-V-pY-H-N-Q-P-L-N >300 <0.003 EGFR pY1114 N-P-E-pY-L-N-T-V-Q-P-T 124 t 25 0.009 EGFR pY1148 N-P-D-pY-Q-Q-D-F-F-P-K >300 <0.003 EGFR pY1173 N-A-E-pY-L-R-V-A-P-Q-S >300 <0.003

IRS-1 pY147 E-D-L-S-pY-D-T-G-P-G-P-A 261 t 32 0.004 IRS-1 pY460 L-S-N-pY-I-C-M-G-G-K-G 48 t 16 0.021 IRS-1 pY546 I-E-E-pY-T-E-M-M-P-A-A 11 t 1.0 0.091 IRS-1 pY608 D-D-G-pY-M-P-M-S-P-G-V >300 <0.003 IRS-1 pY628 G-N-G-D-pY-M-P-M-S-P-K-S >300 <0.003

IRS-1 pY727 T-G-D-pY-M-N-M-S-P-V-G >300 <0.003 IRS-1 pY895 S-P-G-E-pY-V-N-I-E-F-G-S 4.8 t 1.0 0.21 IRS-1 pY939 S-E-E-pY-M-N-M-D-L-G >300 <0.003 IRS-1 pY987 R-G-D-pY-M-T-M-Q-I-G >300 <0.003 IRS-1 pYlOl0 P-V-S-pY-A-D-M-R-T-G-I 110 t 23 0.009 IRS-1 pY1172 S-L-N-pY-I-D-L-D-L-V-K 1.1 2 0.5 1.0" IRS-1 Y1172 S-L-N- Y-I-D-L-D-L-V-K >300 <0.003 IRS-1 pY1222 L-S-T-pY-A-S-I-N-F-Q-K 25 t 4.2 0.044

IRS-1 pY658 P-N-G-pY-M-M-M-S-P-S-G 96 t 13 0.01

a Relative affinities were determined by dividing ID,, values by the ID,, of IRS-1 pY1172, the peptide that binds with highest affinity. Demonstrated site of SH-PTP2 interaction with the PDGF receptor (10, 24, 26).

rosine. In some cases an extra residue was added or deleted for synthetic purposes. For example, for IRS-1 pY939 and IRS-1 pY987 C-terminal Pro was avoided to prevent low yields associated with diketopiperazine formation.

The recently developed PI 3-kinase p85 SH2 domain assay (36, 37) along with information from phosphopeptide library studies (35) provided a starting point for development of the SH-PTP2 SH2 domain assay. To maximize differences in relative affinities between a variety of ligands in a competition assay, it is generally best to radiolabel the highest affinity ligand. PDGFR pY1009 and IRS-1 pY1172 were initially chosen based on phosphopeptide library predictions (35). Since Bolton- Hunter reagent reacts preferentially with €-amino groups uersus a-amino groups (data not shown), an extra Lys was added to the N terminus of PDGFR pY1009. Each peptide was 12'1-

radiolabeled with Bolton-Hunter reagent and HPLC-purified. In titration studies lZ5I-IRS-l pY1172 was bound by the GST/ N-SH2 domain fusion protein more efficiently than "'I-PDGFR pY1009 (Fig. l) , suggesting that the affinity of the N-terminal SH2 domain of SH-FTP2 was greater with IRS-1 pY1172 than PDGFR pY1009. Therefore, lZ5I-IRS-l pY1172 was used in sub- sequent assays with the N-terminal SH2 domain.

Binding of PDGF Receptor-derived Phosphopeptides-Three regions of the PDGF 0-receptor appear to be involved in SH2 domain recognition. A juxtamembrane region surrounding Tyr'" and ljT5'l represents the putative c-Src-binding site (12), the kinase insert domain contains the well established sites of PI 3-kinase recognition (Tyr740 and Tyr'") and a potential site of RasGAP binding ( T y P ) (5, 8, 9), and Tyr1009 and Tyr"" in the

30u 1 u N-SHP vs. IRS-1 pY1172

c /

I I I I I

0 2 4 6 8 10

ISH2 Domain], pM

FIG. 1. Titration of radiolabeled peptides. lz5I-IRS-l pY1172 or lZ5I-PDGFR pY1009 (lo5 cpm), the indicated concentrations of N- or

together as described under "Material and Methods." Following cen- C-terminal SH2 domain, and glutathione-agarose beads were incubated

trifugation, supernatant solutions were removed and radioactivity associated with agarose pellets was determined with a y-counter. Bound values were determined as the amount associated with the pellet in the presence of SH2 domain minus the amount in the absence of SH2 domain.


100

80

60

40

20

0

- 100 2 80 r 3 60

E 40

.-

U 5 20

g o

100

80

60

40

20

0

F\ - * PDGFR Y5791581

A. JUXTAMEMBRANE

'-"PDGFR pY740 -PDGFR pY751 - PDGFR Y771

B. KINASE INSERT

"PDGFR Y1009 d P D G F R pY1009

C. C-TERMINUS

-9 -8 -7 -6 -5 -4 -3 log [Phosphopeptide, (M)]

FIG. 2. SH-PTP2 SH2 domain interactions with PDGF receptor peptides. Eight PDGF p receptor-derived phosphopeptides were analyzed by competition assay (see "Materials and Methods"). Maximum bound represents the amount of '251-phosphopeptide bound in the absence of competing ligand (10,000-20,000 cpm). Results shown are with phosphopeptides from the receptor juxtamembrane (A), kinase insert ( B ) , and C-terminal tail domains (C). Data presented are mean 2 S.D. values; lines are derived according to the sigmoid dose-response equation. The effect of peptide phosphorylation is shown in panel C; compare PDGFR pY1009 with PDGFR Y1009 (peptide sequences are otherwise identical).

C terminus of the receptor provide docking sites for SH-PTP2 and PLC-y (7, 10, 11, 13, 14). Individual phosphopeptides corresponding to these phosphorylation sites were synthesized; for the TyPg/Tyr5'' site each monophosphorylated sequence and a diphosphorylated peptide were prepared (Table I).

Of these eight peptides, PDGFR pY1009 competed most effectively for SH2 domain binding (Fig. 2C yielding a calculated ID,, value of 14 PM (Table I). This peptide contains a pYTAV motif. Although Thr at the pY+1 position was not selected by the library method, Ile and Val were predicted to bind with high affinity (35). Since Thr is the only P-branched residue besides Ile and Val, taken together these findings suggest that P-branched residues are preferred at the pY+1 position. Nota- bly, mutation of PDGF receptor Tyr'009 + Phe markedly dimin- ishes binding to SH-PTP2 (10, 11, 14). In parallel fashion, re- moval of phosphate from peptide PDGFR pY1009 abolishes binding (PDGFR Y1009, Fig. 2C).

Peptide PDGFR pY579 also bound with relatively high affinity, having an ID,, value of 49 PM (3-fold weaker than PDGFR pY1009). PDGFR pY581 is the same peptide sequence which is phosphorylated at a different Tyr position (Table I). PDGFR pY581 binds with significantly lower affinity. PDGFR pY579/ 581 (the same sequence phosphorylated at both Tyr positions)

bound with the same affinity as PDGFR pY579. We interpret these findings as follows. The phosphotyrosine binding site of the SH2 domain is occupied by pY579 for peptides PDGFR pY579 and PDGFR pY5791581, with Ile a t pY+1 and Val a pY+3; these residues were predicted by library studies to be ideal for SH-PTP2 SH2 domain interactions (35). We presume that the reason PDGFR pY579 and PDGFR pY579/581 do not bind to the SH-PTP2 SH2 domain with even greater affinity relates to the presence of an additional steric or structural constraint like the ring side chain of Tyr pY+2. For PDGFR pY581 the pY-binding site of the SH2 domain is occupied by pY581, which aligns Val at pY+1 and Pro at pY+3. Pro at pY+3 is not favored (Table I). Alternative phosphorylation sites of the PDGF receptor bound with much lower affinity.

Binding of EGF Receptor-derived Phosphopeptides- Detailed mutational analyses of EGF receptor tyrosines to map sites of SH2 protein interactions have not been as successful as those described above with the PDGF receptor. Sites of autophosphorylation have been mapped biochemically to Tyr106',

Tyr1OE6, Tyr1148, and Tyr'173 (42-44) and by mutagenesis to T y r g g 2

(1). Each Tyr has been mutated to Phe and notably none of the mutations blocks the interaction of the EGF receptor with SH- PTP2 (data not shown). An isolated PLC-y SH2 domain blocks PTPase-mediated dephosphorylation at Tyrgg2 and binds with highest affinity to Tyrgg2 and Tyr1173 (20). Although additional SH2 proteins interact with the EGF receptor, to our knowledge no additional sites of interaction have been mapped. We prepared phosphopeptides corresponding to each of the reported phosphorylation sites and tested them in the SH-PTP2 SH2 domain assay. Phosphopeptides corresponding to sequences surrounding Tyrgg2, TyrlOm, Tyr10a6, Tyr114', and Tyr1173 bound ineffectively to the SH-PTP2 SH2 domain (Fig. 3), consistent with the mutational analyses.

Therefore, we re-examined the EGF receptor sequence to identify additional Tyr-containing sequences which might bind SH2 proteins if phosphorylated. One such site surrounds T y r g Z 0

within a Y" motif. This peptide was prepared and found to bind with full affinity to PI 3-kinase p85 SH2 domain^.^ Addi- tional sites surrounding Tyr954 and TY~"'~ appeared as though they might conform to the specificity of SH-PTP2. Both bound to the SH2 domain (Fig. 31, and the sequence surrounding Tyr954 (within a pYLVI motif) bound with high affinity (ID,, = 21 p ~ ; Table I). From these analyses we predict that Tyr954 in the intact EGF receptor might direct SH-PTP2 interactions (providing that this site becomes phosphorylated).

Binding of IRS-1 -derived Phosphopeptides-Although IRS-1 is extensively phosphorylated following insulin stimulation, in vivo sites of tyrosine phosphorylation have not reported. How- ever, IRS-1 contains six Y" motifs which as peptides are readily phosphorylated by the insulin receptor (45) and repre- sent prime recognition motifs for PI 3-kinase p85 SH2 domains (23,37,46). We have prepared phosphopeptides corresponding to many possible phosphorylation sites in IRS-1, including all Y" and YXXM motifs, and additional sites which fit predicted motifs for SH2 domain interaction (35, 39). We have shown previously that the Y" motifs bind with high affinity and specificity to p85 SH2 domains (37, 46). In contrast, none of the Y"-pept ides bound tightly with the SH-PTP2 N-SH2 domain (Fig. 4, B and C ) , although IRS-1 pY658 with Met at the pY+2 position bound more tightly (ID,, = 96 PM) than the rest (ID,,, > 300 w; Table I).

However, the SH-PTP2 N-SH2 domain did exhibit substan- tial affinity for two of the YXXM peptides (Fig. 4A ). ID,, values for IRS-1 pY460 and IRS-1 pY 546 binding were 48 and 11 p ~ , respectively. These peptides contain Ile or Thr, respectively, at

E. Piccione, R. Case, and S. E. Shoelson, unpublished observation.

SH-PTP2 SH2 Domain Specificity 10471

100

80

60

40

20 0

- 100 2 80

.s 60 2 E 40 -0 5 20

$ 0

100

80

60

40

20 0

EGFR pY954

G I -o- EGFR pY1068

I * EGFR pY1173 1 -9 -8 -7 -6 -5 -4 -3 log [Phosphopeptide, (M)]

with the SH-F'TP2 SH2 domain. Binding studies were conducted as FIG. 3. EGF receptor-derived phosphopeptide interactions

described under "Materials and Methods" and the legend to Fig. 2.

pY+1 positions (both being P-branched residues). All three YXXM sequences (the third is Tyr''" with Ala at pY+1) bind with high affinity to PI 3-kinase p85 SH2 domains (data not shown). Therefore, phosphorylation sites surrounding Tyr460

and Tyr54fi can make high affinity complexes with the SH2 domains of at least two different proteins.

Four additional peptides were also studied. Results from phosphopeptide library studies (35,39) predict that Tyr147 with an Asp at pY+1 might bind to the Nck SH2 domain and T y r S g 5

with Asn at pY+2 might interact with Grb2. T y r S g 5 with Val and Ile at pY+1 and pY+3 is also predicted to bind SH-PTP2, as is T~I-"~' , with Ile and Leu at pY+1 and pY+3. Peptides IRS-1 pY895 and IRS-1 pY1172 bind to the SH-PTP2 SH2 domain with higher affinity than any other peptide tested (ID5o = 5 p~ and 1 VM, respectively). These findings show that Tyr1172, Tyrsg5,

Ty154fi , and perhaps T y r 4 " can create high affinity sites of recognition for the N-SH2 domain of SH-PTP2 if they are exposed on the surface of IRS-1 and phosphorylated. Many of these and an additional site (Tyrl'") appear to be phosphorylated by the insulin receptor in vitro (28). While peptide IRS-1 pY1222 binds to the SH-PTP2 N-SH2 with lower relative affinity (ID5,, = 25 p~), this peptide stimulates catalytic activity suggesting possible association with the C-SH2 domain.,

Binding Studies with Intact Receptors-To further validate the use of phosphopeptides and SH2 domains in the prediction of interactions between intact proteins, we analyzed the ability of phosphopeptides to block PDGF receptor binding to the SH- PTP2 SH2 domain. At a 50 p~ concentration, the PDGFR pY1009 phosphopeptide effectively blocked the interaction be-

100

80

60

40

'

'

$ 20 .E 0 '

Y

. - 8 IRS-1 pY460 1 - A IRS-1 pY608 *IRS-1 pY546 "s-IRS-1 pY628

' "klRS-1 PYlOlO *IRS-I PY658

OIRS-1 Y1172

I . . * . . . . I . . ' . . . . . I

-9 -8 -7 -6 -5 -4 -3 -9 -8 -7 -6 -5 -4 -3 log [Phosphopeptide, (M)]

FIG. 4. Binding between IRS-1 phosphopeptides and the SH- pTp2 SH2 domain. Competition assays were performed with IRS-1 sequences containing YXXM motifs (panel A ), Y" motifs (panels B and C) and motifs predicted to interact with alternative SH2 domains (panel D, see text).

tween the autophosphorylated PDGF receptor and the SH2 domain (Fig. 5 A ) . In contrast, the same concentration of PDGFR pY771 had little effect in this assay. We have shown previously that of PDGF receptor-derived peptides, only PDGFR pY1009 is an effective competitor (11).

Similar studies were conducted with intact, autophosphorylated EGF receptors. Phosphopeptides corresponding to reported sites of EGF receptor autophosphorylation were ineffec- tive at inhibiting EGF receptor/SH-PTP2 SH2 domain binding (Fig. 5B). However, as predicted by the phosphopeptideBH2 domain assay (Fig. 3), phosphopeptide EGFR pY954 did compete effectively for binding. Therefore, results obtained by ana- lyzing direct interactions between phosphopeptides and the SH-PTP2 SH2 domain correlate perfectly with analogous results obtained by blocking interactions between intact phosphoproteins (EGF and PDGF receptors) and the SH2 domain.

The Effect of Phosphopeptide Length-As discussed at the beginning of "Results," most phosphopeptides in these studies contain 11 residues, with 3 residues N-terminal and 7 residues C-terminal to phosphotyrosine. For binding to PI 3-kinase p85 SH2 domains, we have previously showed that six amino acid peptides bind with high affinity, especially when N and C termini are b10cked.~ Similarly, hexapeptides having the general configuration Acetyl-X-pY-X-X-X-X(NH,) bind with high affinity with SH2 domains of Grb2, Src, and PLC-Y.~ We now show that the N-SH2 domain of SH-PTP2 behaves quite differently. Eleven- and 9-residue peptides containing the correct sequence bind tightly, whereas corresponding six amino acid peptides bind weakly (Fig. 6).

Binding Studies with the C-terminal SH-PTP2 SH2 Domain-Studies of binding specificity were also attempted with the C-terminal SH2 domain of SH-PTP2. 'T-Peptide binding and BIAcore analyses (conducted in collaboration with S. Sugimoto and C. Walsh) indicated that the C-terminal SH2 domain bound to IRS-1 pY1172 and PDGFR pY1009 with significantly lower affinity (data not shown). However, point mu-

E. Piccione, R. Case, and S. E. Shoelson, manuscript in preparation.


-1 B. A431 LYSATE

tions with SH-PTP2 SH2 domain. A, PDGF receptor is precipitated FIG. 5. Phosphopeptide inhibition of intact receptor interac-

from stimulated ATWT cell lysates by SH-PTP2 SHS/GST protein. Phosphopeptide PDGFR pY1009 blocks association. B , the EGF receptor from stimulated A431 cells is precipitated by SH-PTP2 SH2/GST protein. Association is blocked by phosphopeptide EGFR pY954 but not alternative EGF receptor-related peptides.

tations in either SH2 domain of the intact protein (Arg -> Lys substitutions of the conserved BB5 FLVR arginine) reduce the ability of phosphopeptides to stimulate catalytic activity; suggesting that both SH2 domains are functionally important. Proper constructs of the C-SH2 domain have been prepared recently and related analyses of specificity are underway.

DISCUSSION Associations between the SH2 domains of SH-PTP2 and

phosphorylation sites of proteins like the PDGF and EGF receptors and IRS-1 are specific and physiologically relevant, as they (i) occur in cells as a direct response to ligand stimulation, (ii) require tyrosine-phosphorylation, and (iii) can be blocked by the appropriate Tyr mutation. With potential for hundreds of related interactions in a given cell, specificity is crucial to main- taining order. We have developed an assay system which uses isolated SH2 domains and series of phosphopeptides to determine SH2 domain specificity directly. Since the sites of SH2 domain interaction with the PDGF receptor have been well characterized (7-14), we have validated the assay with the appropriate peptide sequences (Fig. 2) and binding analyses with the intact receptor (Fig. 5). For example, phosphopeptides corresponding to sequences surrounding PDGF receptor Tyr74n and Tyr7" bind tightest to PI 3-kinase SH2 domains, whereas PLC-y SH2 domains bind with greatest affinity to a peptide corresponding to PDGF receptor TyrIn2' (37). In the current study we have extended these observations to show that the SH-PTP2 SH2 domain binds tightest to PDGF receptor T y P o 9 .

In each of these examples and in every additional case tested specificity has been preserved and found to parallel what is known to be important for interactions between intact proteins and in vivo biologic effects.

Therefore, we can use this phosphopeptide/SH2 domain assay system to analyze determinants of SH2 domain specificity and to predict sites of interaction with additional phosphoproteins. In addition to the PDGF receptor, SH-PTP2 forms complexes with activated EGF receptors and insulin receptor-phosphorylated IRS-1. Although 5 or 6 tyrosines in the EGF receptor have been shown to be phosphorylated (42-44), none of these sites binds tightly to the SH-PTP2 N-SH2 domain (Fig. 3) or competes effectively for N-SH2 domain binding with the intact receptor (Fig. 5B ). We inspected the sequence of the EGF receptor to determine whether a previously unmapped phosphorylation site might provide the necessary motif for SH- PTP2 recognition. No intracellular Tyr occurs within an ideal motif. T y P 4 is followed by L-V-I, which looked like a reasonable alternative. The corresponding peptide bound with essentially the same affinity as PDGFR pY1009. Unlike peptides corresponding to known EGF receptor phosphorylation sites, EGFR pY954 inhibits binding of the autophosphorylated receptor

with the SH2 domain, therefore we predict that this site might be phosphorylated and act as a recognition motif. Although the nearest acidic residue is at the pY-4 position, relative to Tyrgs4, the requirement for acidic residues N-terminal to EGF receptor sites of autophosphorylation is questionable (e.g. there are no acidic residues near TyP" (431, a known site of phosphorylation). It is also worth noting that like Tyrgs4, the phosphorylation of T y r g g 2 has not been mapped biochemically. Nevertheless T y r g g 2 directs PLC-y interactions with the EGF receptor (2,20), suggesting that i t is either transiently phosphorylated in cells or phosphorylation is labile and lost during isolation.

Even less is known about IRS-1. At the time of these studies no phosphorylation sites had been published. Therefore, we prepared a panel of IRS-1 derived phosphopeptides to determine which sites might interact with SH2 domains. Three of 12 IRS-1 related sequences exhibited high affinity binding with the SH-PTP2 N-SH2 domain: IRS-1 pY546, IRS-1 pY895, and IRS-1 pY1172. In fact two of these peptides bound with significantly greater affinity (3- to 14-fold) than PDGFR pY1009, the only previously known site of SH-PTP2 recognition. Studies in progress in M. White's laboratory suggest that both of these high affinity peptides are phosphorylated in vitro (28). Our results suggest that both are potential sites interaction with

The sequences of the PDGF receptor-, EGF receptor- and IRS-1-derived peptides provide general information about SH2 domain specificity. IRS-1 pY895 (E-pY-V-N-I-E-F) and IRS-1 pY1172 (N-pY-I-D-L-D-L) bind with highest affinity and likely define optimal motifs. These sequences have much in common. Notably, both have a hydrophobic P-branched residue a t pY+1 and a hydrophobic residue at pY+3. Both also have residues with either carboxylic acid or amide side chains a t pY-1, pY+2 and pY+4 and a hydrophobic residue a t pY+5. PDGFR pY1009 (L-pY-T-A-V-Q-P), EGFR pY954 (R-pY-L-V-I-Q-G), and IRS-1 pY546 (E-pY-T-E-M-M-P) each bind with somewhat lower affinity and contain distinct sequences. For example, PDGFR pY1009 and IRS-1 pY546 both have Thr at pY+1. Although Thr is /3-branched it is significantly less hydrophobic than Val or Ile and may provide a less stable interaction with a hydrophobic pocket or patch on the SH2 domain. EGFR pY954 contains Leu a t pY+1 which, although hydrophobic, contains a y- rather than a p-branch. We cannot tell whether these differences or varia- tions at pY-1, pY+2, pY+4, or pY+5 positions explain the lower affinities, although we are now in a position to test these and many additional questions about SH-PTP2 SH2 domain specificity.

Recent findings in our laboratory show that particular phosphopeptide sequences have dual specificity. In the current study, SH-PTP2 binds with reasonable affinity to PDGFR pY579 and PDGFR pY579/581, sites of Src SH2 domain recognition (12). Under certain situations it may be possible foF SH-PTP2 and Src to compete for binding at this site, although the location of the Tyr579 and near the membrane may favor Src binding over SH-PTP2. Additional sites of dual specificity are even more striking. For example, IRS-1 pY895 (YVNI) binds to both Grb2 and SH-PTP2 SH2 domains with equally high affinity.s Moreover, Grb2 and p85 SH2 domains bind to IRS-1 pY727 and IRS-1 pY939 sites (both are within YMNM motif^).^ These results predict that selected phosphorylation sites act as recognition motifs for multiple SH2 proteins. Con- firmation awaits the appropriate experiments with intact proteins in cells.

As an alternative to the l2'1-phosphopeptide competition assay described here, phosphopeptide/SH2 domain interactions

SH-PTP2.

' G. Wolf, E. Piccione, R. Case, and S. E. Shoelson, unpublished observation.

SH-PTP2 SH2 Domain Specificity 10473

FIG. 6. Effects of phosphopeptide length. Competition assays were performed with 11-residue PDGFR pY1009 and the corresponding 6-mer (panel A ) , and 11-residue IRS-1 pY1172 and the corresponding 9- and 6-mers (panel B ) . Pep- tide sequences are listed in the figure; 11- and 9-mers contain free N and C termini, whereas 6-mers are N"-acetylated and contain amide groups at the C termini.

100 E 80

f 60

E 40

3 20 m O

.-

-0 c 0

[A. PDGFR Tyrl009 I B. IRS-1 Tyrl 172

0 1 1 -mer: SVLpYTAVQPNE

0 6-mer: (Ac)LpYTAVQ(NH,) k SLNpYlDLDLVK

0 9-mer: LNpYlDLDLV

. 0 6-mer: (Ac)NpYIDLD(NH,)

-9

have been analyzed using biosensor technology (e.g. BIAcore, Pharmacia LKB Biotechnology Inc.) (46-48). The BIAcore is most useful in determining association and dissociation rates, and estimating actual Kd values, as has been shown previously for analyses of p85, Src, and Lck SH2 domain interactions with corresponding phosphopeptides. We have observed similar rapid kinetics with the SH2 domains of SH-PTP2.' However, each de novo calculation of an affinity constant is time consum- ing and subject to some error (46-48). Therefore, comparing relative affinities of series of related ligands competition experiments should be significantly more accurate and are certainly more convenient. While competition experiments can be performed on a BIAcore instrument (46, 47), the '251-phosphopeptide competition assay described here provides an inex- pensive method which requires no special instrumentation.

The importance of peptide residues at pY+1 and pY+3 positions suggests that the mode of peptide binding to the SH-PTP2 SH2 domain is similar to Src or Lck SH2 domain interactions with a high affinity YEEI peptide (35, 47). Crystallographic structures illustrate how phosphotyrosine anchors the peptide to the SH2 domain through an elaborate hydrogen bonding network within a discrete binding pocket (40, 41, 49). Interac- tions between phosphotyrosine and the SH2 domain provide the majority of free energy for phosphopeptide/SH2 domain interactions.6 Members of this family of SH2 domain-containing PTPases (csw, SH-PTP1, and SH-PTP2) lack the Arg (&) present in most SH2 domains which provides an amino- aromatic contact with the ring of phosphotyrosine (Fig. 7). Whether this contact is unnecessary for phosphotyrosine binding to SH-PTP2 or replaced by an alternative type of interaction is unknown.

Additional interactions which are unique to particular phosphopeptide/SH2 domain pairs provide specificity. For example, p Y " or pWXM motifs confer specificity to PI 3- kinase SH2 domain interactions. Accumulating evidence suggests that residues C-terminal to phosphotyrosine (pY+l, pY+2, and pY+3) confer specificity to most phosphoproteidSH2 domain interactions. The side chain of hydrophobic residue Ile at the pY+3 position of a pYEEI peptide inserts itself into a second cavity on the surfaces of Src and Lck SH2 domains. This hydrophobic cavity is formed by residues from the EF and BG loops (Fig. 7) (40, 41). Peptide residues at positions pY+1 and pY+2 (both Glu) do not fit into grooves or pockets in on Src or Lck SH2 domains, although their side chains may interact with basic residues on the SH2 domain surfaces.

Differential sensitivity to peptide length represents a signifi- cant difference between the N-terminal SH-PTP2 SH2 domain

S. Shoelson, unpublished observation.

-8 -7 -6 -5 -4 -3 -9 -8 -7 -6 -5 -4 -3 log [Phosphopeptide, (M)]

and previously studied SH2 domains. We have found that for several SH2 domains a capped hexapeptide binds with essentially the same high affhity as corresponding 9- or 11-mers. Based on these findings we prepared capped hexapeptides corresponding to sequences surrounding PDGF receptor %Irng

and IRS-1 Tyr"72. However, these 6-mers bound the SH-PTP2 SH2 domain with markedly reduced affinity (Fig. 61, suggesting that additional backbone and perhaps side chain interactions are necessary for high affinity binding with this SH2 domain. In fact, the sequences that bind with highest affinity to the N-SH2 domain all contain a hydrophobic residue (L, F, or P) at the pY+5 position; recent crystallographic analyses indicate that this side chain may participate in high affinity interac- t i o n ~ . ~

Since peptide side chains that help to confer specificity come in direct contact with SH2 domain EF and BG loops, it is likely that these loops are important in determining specificity on the part of the SH2 domain. EF loops are typically 2-3 residues in length and differ in sequence between SH2 domains (Fig. 7). The BG loops from different SH2 domains vary markedly both in length and composition. For example, SH2 domains from Src and Src-family members, Ab1 and RasGAP contain 9-residue BG loops, whereas PI 3-kinase N- and C-p85 domains have 14-residue BG loops. Cross-linking studies may help to explain the difference. The BG loop of the N-p85 SH2 domain is cross- linked by the peptide pY+1 position, which has led us to pos- tulate the presence of more extensive contacts between the longer loop of p85 and specificity-conferring positions of the peptide (31). Like PI 3-kinase p85, the BG loop of the SH-PTP2 N-SH2 domain contains 14 residues. Therefore, we would predict that BG loops of SH-PTP2 SH2 domains contact hydrophobic @-branched) residues at the peptide pY+1 position and, as previously predicted for p85, that these interactions participate in conferring specificity.

Comparison of N- uersus C-terminal SH-PTP2 SH2 domain sequences may also help explain why we have observed dis- similar protein stability and binding characteristics for the two SH2 domains. As discussed above, in general BG loops segre- gate by function. For example, Src and Lck SH2 domains both bind to YEEI motifs (35, 47), whereas both N- and C-terminal p85 SH2 domains bind best with Y" motifs (3, 5,35, 37, 46).4 Notably, the BG loops of Src and Lck exhibit high sequence homology, and BG loops of N-p85 and C-p85 are similar to one another (Fig. 7). In contrast, BG loops from N- versus C-terminal SH-PTP2 SH2 domains are unequal in length and show less homology, particularly within their N-terminal halves (Fig. 7). N- and C-SH2 domains of SH-PTP2 differ dramatically a t ad-

' J. Kuriyan, personal communication.


20 struc. BA ClA BB PC BD PE BF aB BG LOOP AA AB BC CD DE EF BG N-SHPTP2 WFH PNI TGVEAENLLL TRGVCG SFLARPS KSNP G DFTLSVRR NG AVTHIKIQNT GD YYDL YGG EKF A TLAELVQYYMEH HGQLKEKNGDVIELK YPL NCA C-SHPTP2 WFH GHL SGKEAEKLLT EKGKHG SFLVRES QSHP G DFVLSVRT GDDKGESNDGKS KVTWVMIRCQ EL KYDV GCG ERF D SLTDLVEHYKKN PMV ETLGTVLQLK QPL NTT N-pE5 WYW GDI SREEVNEKLR DTADG TFLVRDA STKMHG DYTLTLRK C-pa5 WNV GSS NRNKAENLLR GKRffi TFLVRES SKN G CACSTVW

GG NNKLIKIFHR DG KYGF SDP LTF N SVVELINHYRNE SLAQYNPKLDVKLL YPV SKY f f i GVKHCVINKT ATG YGFA EPY NLY S SLKELVLHYQHT SLVQHNDSWLA YPV YAN

Src Lck WFF KNL SRKDAERQLL AXNTHG SFLVRES EST AG SFSLSVRD WQNQGE VVKHYKIRNL DNG GFYI SPR ITF P GLHELVRWYTNA SDG LCTRLS RPC QTQ

WYF GKI TRRESERLLL NPENPRG TFLVRES ETT KG AYCLSVSD FDNAKGL NVKHYKIRKL DSG GFYI TSR TQF S SLQQLVAYYSKH A X LCHRLT NVC PTS

FIG. 7. Sequence alignment of SH2 domains from SH-PTP2, PI 3-kinase p85, Src, and Lck. Criteria for alignment and nomenclature are as described in Ref. 34. Secondary structures based on crystallographic analyses (40, 41, 49) are indicated at the top; assignments of intervening loops are one row down

ditional positions, as well. Most noticeably, the C-SH2 contains a long insert between P-strands C and D. Although we do not know how this insert affects function, it is worth noting that Src-family SH2 domains contain a similar but shorter CD in- sertion (Fig. 7).

In conclusion, phosphoprotein interactions with the SH2 domains of SH-PTP2 can be mimicked with corresponding phosphopeptides and expressed SH2 domains. Specificity in the interaction is preserved, so affinities of the SH2 domain for various peptide sequences can be analyzed directly. We have shown that of known PDGF receptor phosphorylation sites 5r1Oo9 binds with highest affinity. Since this is the actual site of PDGF receptor recognition by SH-PTP2, we have concluded that the assay accurately reflects known phosphoproteidSH2 domain interactions. Using related approaches we have predicted sites of SH-PTP2 interaction with the EGF receptor and IRS-1. These methods facilitate identification of site-specific interactions between SH-PTPZ and various phosphoproteins and should prove invaluable in further analyses of SH-PTPZ specificity and in developing potential pharmacophores and related agents for use in intact cells.

Acknowledgments-We thank L. C. Cantley, Z. Songyang, S. Sugi- moto, C. Walsh, and K. P. Williams for helpful discussions.

REFERENCES 1. Ullrich, A,, and Schlessinger, J. (1990) Cell 61, 203-212 2. Anderson, D., Koch, C. A., Grey, L., Ellis, C., Moran, M. F., and Pawson, T.

3. Cantley, L. C., Auger, K. R., Carpenter, C., Duckworth, B., Graziani, A., Kapel-

4. Pawson, T., and Gish, G . D. (1992) Cell 71,359-362 5. Fantl, W. J., Escobedo, J. A., Martin, G. A., Turck, C. W., del Rosario, M.,

6. Feng, G.-S., Hui, C.-C., and Pawson, T. (1993) Science 259, 1607-1611 7. Kashishian, A., and Cooper, J. A. (1993) Mol. Cell. Biol. 13, 49-57 8. Kashishian, A., Kazlauskas, A., and Cooper, J. A. (1992) EMBO J. 11, 1373-

9. Kazlauskas, A.. Kashishian, A,, Cooper, J. A,, and Valius, M. (1992) Mol. Cell.

(1990) Science 250, 979-982

ler, R., and Soltoff, S. (1991) Cell 64, 281-302

McCormick, F., and Williams, L. T. (1992) Cell 69, 413-423

1382

10. Kazlauskas, A., Feng, G.-S., Pawson, T., and Valius, M. (1993) Proc. Natl. Biol. 12, 2534-2544

Acad. Sci. U. S. A. 90.6939-6942 11. Lechleider, R. J., Sugimoto, S., Bennet, A. M., Kashishiair, A. S., Cooper, J. A.,

Shoelson, S. E., Walsh, C. T., and Neel, B. G. (1993) J. Biol. Chem. 268, 21478-21481

12. Mori, S., Ronnstrand, L., Yokote, K., Engstrom, B., Courtneidge, S. A., Claes- son-Welsh, L., and Heldin, C:H. (1993) EMBO J. 12, 2257-2264

13. Ronnstrand, L., Mori, S., Arridsson, A,-K., Eriksson, A., Wernstedt, C., Hell- man, U., Claesson-Welsh, L., and Heldin, C.-H. (1992) EMBO J. 11, 3911-

14. Valius, M., Bazenet, C., and Kazlauskas, A. (1993) Mol. Cell. Bid. 13, 133-143 3920

15. Valius, M., and Kazlauskas, A. (1993) Cell 73, 321-334 16. Vogel, W., Lammers, R., Huang, J., and Ullrich, A. (1993) Science 259, 1611-

17. Egan, S. E., Giddings, B. W., Brooks, M. W., Laszlo, B., Sizeland, A. M., and

18. Laszlo, B., and Downward, J. (1993) Cell 73, 611420 19. Li, N., Batzer, A., Daly, R., Yajnik, V., Skolnik, E., Chardin, P., Bar-Sagi, D.,

1614

Weinberg, R. A. (1993) Nature 363,45-51

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

32. 31.

33.

34.

35.

36.

37.

38.

39.

40. 41.

42. 43.

44.

45.

46.

47.

48.

49.

Rotin, D., Margolis, B., Mohammadi, M., Daly, R., Daum, G., Li, W., Burgess, Margolis, B., and Schlessinger, J. (1993) Nature 363, 8 5 8 8

W., Fischer, E. H., Ullrich, A., and Schlessinger, J. (1992) EMBO J. 11, 559-567

Rozakis-Adcock, M., Fernley, R., Wade, J., Pawson, T., and Botwell, D. (1993) Nature 363, 83-85

Backer J. M., Myers, M. G., Sun, X. J., Chin, D. J., Shoelson, S. E., Miralpeix, M., and White, M. F. (1993) J. Biol. Chem. 268,820443212

Sun, X. J., Rothenberg, P., Kahn, C. R., Backer, J. M., Araki, E., Wilden, P. A., Cahill, D. A., Goldstein, B. J., and White, M. F. (1991) Nature 352, 73-77

Baltensperger, K., Kozma, L. M., Cherniack,A. D., Klarlund, J. K., Chawla,A., Banerjee, U., and Czech, M. P. (1993) Science 260,1950-1952

Kuhne, M. R., Pawson, T., Lienhard, G. E., and Feng, G:S. (1993) J. Biol. Chem. 268, 11479-11481

Skolnik, E., Batzer, A,, Li, N., Lee, C.-H., Lowenstein, E., Mohammadi, M., Margolis, B., and Schlessinger, J. (1993) Science 260, 1953-1955

Skolnik, E., Lee, C.-H., Batzer, A., Vincentini, L. M., Zhou, M., Daly, R., Myers, M., Backer, J., Ullrich, A,, White, M. F., and Schlessinger, J. (1993) EMBO J. 12, 1929-1936

Sun, X. J., Crimmins, D. L., Myers, M. G., Miralpeix, M., and White, M. F.

Tobe, K., Matuoka, K., Tamemoto, H., Ueki, K., Kaburagi, Y., Asai, S., Noguchi, (1993) Mol. Cell. Biol. 13, 7418-7428

T., Matsuda, M., Tanaka, S., Hatom, S., Fukui, Y., Akanuma, Y., Yazaki, Y., Takenawa, T., and Kadowaki, T. (1993) J. Biol. Chem. 268, 11167-11171

Lechleider, R. J., Freeman, R. M., and Neel, B. G. (1993) J. Biol. Chem. 268, 13434-13438

Williams, K. P., and Shoelson, S. E. (1993) J. Biol. Chem. 268, 5361-5364 Perkins, L. A,, Larsen, I., and Pemmon, N. (1992) Cell 70, 225-236 Lu, X., Chou, T.-B., Williams, N. G., Roberts, T., and Pemmon, N. (1993) Genes

7,621432 Sugimoto, S., Lechleider, R. J., Shoelson, S. E., Neel, B. G., and Walsh, C. T.

(1993) J. Biol. Chem. 268,22771-22776 Songyang, Z., Shoelson, S. E., Chaudhuri, M., Gish, G., Pawson, T., King, F.,

Roberts, T., Ratnofsky, S., Lechleider, R. J., Neel, B. G., Birge, R. B., Fa- jardo, J. E., Chou, M. M., Hanafusa, H., Schamausen, B., and Cantley, L. C. (1993) Cell 72,767-778

Domchek, S., Auger, K., Chatterjee, S., Burke, T., Jr., and Shoelson, S. E. (1992) Biochemistry 31,9865-9870

Piccione, E., Case, R. D., Domchek, S. M., Hu, P., Chaudhuri, M., Backer, J. M., Schlessinger, J., and Shoelson, S. E. (1993) Biochemistry 32,3197-3202

Shoelson, S. E., Sivaraja, M., Williams, K. P., Hu, P., Schlessinger, J., and Weiss, M. A. (1993) EMBO J. 12,795-802

Songyang, Z., Shoelson, S. E., McGlade, J., Olivier, P., Pawson, T., Bustelo, R. X., Hanafusa, H., Yi, T., Ren, R., Baltimore, D., Ratnofsky, S., Feldman, R. A., and Cantley, L. C. (1994) Mol. Cell. Biol., in press

Eck, M., Shoelson, S. E., and Hamson, S. C. (1993) Nature 362, 87-91 Waksman, G., Shoelson, S. E., Pant, N., Cowburn, D., and Kuriyan, J. (1993)

Margolis, B. L., Lax, I., Kris, R., Dombalagian, M., Honegger, A. M., Howk, R., Downward, J., Parker, P., and Waterfield, M. D. (1984) Nature 311,483-485

Givol, D., Ullrich, A,, and Schlessinger, J. (1989) J. Biol. Chem. 264,10667- 10671

Walton. G. M.. Chen, W. S., Rosenfeld, M. G., and Gill, G. N. (1990) J. Biol.

Cell 72, 779-790

Shoelson S. E., Chatterjee, S., Chaudhuri, M., and White, M. F. (1992) Proc. Chem. 265, 1750-1754

Felder S., Zhou, M., Hu, P., Urena, J., Ullrich, A,, Chaudhuri, M., White, M. F., Natl. Acad. SCL. U. S. A. 89, 2027-2031

Payne, G., Shoelson, S. E., Gish, G., Pawson, T., and Walsh, C. T. (1993) Proc. Shoelson, S. E., and Schlessinger, J. (1993) Mol. Cell. Biol. 13, 1449-1455

Panayotou, G., Gish, G., End, P., Truong, O., Gout, I., Dhand, R., Fry, M. J., Natl. Acad. Sci. U. S. A. 90, 49024906

Hiles, I . , Pawson, T., and Waterfield, M. D. (1993) Mol. Cell. Biol. 13, 3567-3576

Waksman, G., Kominos, D., Robertson, S. C., Pant, N., Baltimore, D., Birge, R., Cowburn, D.,Hanafusa, H., Mayer, B. J., Overduin, M., Resh, M. D., Rios, C. B., Silverman, L., and Kuriyan, J. (1992) Nature 358,646-653

w, or · 10468 sh-ptp2 sh2 domain specificity mains unclear. sh-ptp2 (p64) has been linked to ras...

Documents