(Received for publication, February 4, 1997)
From the Laboratory of Immunogenetics, NIAID,
National Institutes of Health, Rockville, Maryland 20852 and the
¶ Department of Cancer Biology, Cleveland Clinic Foundation
Research Institute, Cleveland, Ohio 44195
SHP-1 is a protein-tyrosine phosphatase associated with inhibition of activation pathways in hematopoietic cells. The catalytic activity of SHP-1 is regulated by its two SH2 (Src homology 2) domains; phosphotyrosine peptides that bind to the SH2 domains activate SHP-1. The consensus sequence (I/V)XYXX(L/V) is present in the cytoplasmic tails of several lymphocyte receptors that interact with the second SH2 domain of SHP-1. In several of these receptors, there are two or three occurrences of the motif. Here we show that the conserved hydrophobic amino acid preceding the phosphotyrosine is critical for binding to and activation of SHP-1 by peptides corresponding to sequences from killer cell inhibitory receptors. The interaction of most SH2 domains with phosphopeptides requires only the phosphotyrosine and the three residues downstream of the tyrosine. In contrast, the shortest peptide able to bind or activate SHP-1 also included the two residues upstream of the phosphotyrosine. A biphosphopeptide corresponding to the cytoplasmic tail of a killer cell inhibitory receptor with the potential to interact simultaneously with both SH2 domains of SHP-1 was the most potent activator of SHP-1. The hydrophobic residue upstream of the tyrosine was also critical in the context of the biphosphopeptide. The contribution of a hydrophobic amino acid two residues upstream of the tyrosine in the SHP-1-binding motif may be an important feature that distinguishes inhibitory receptors from those that provide activation signals.
Receptor-mediated activation of cellular responses is often initiated by the activation of tyrosine kinases. The signal is propagated by the sequential recruitment of proteins to the phosphorylated targets of the kinases. SH2 (Src homology 2) domains are protein modules that specifically bind to phosphotyrosine residues and are found in a variety of proteins such as protein kinases and adapter molecules. A general feature of all SH2 domains is a conserved pocket that binds the phosphotyrosine moiety. The specificity of the interaction between individual SH2 domains and particular phosphoproteins is generally determined by three to six residues following the phosphotyrosine and often incorporates a hydrophobic residue at the third position (reviewed in Ref. 1). This specificity is provided by specific pockets that bind the phosphotyrosine and +3 residues as revealed by structural studies of several SH2 domains complexed with phosphopeptides (1).
A small family of protein-tyrosine phosphatases contain SH2 domains.
This family is composed of mammalian SHP-1 and SHP-2 and the
Drosophila homologue of SHP-2, corkscrew. These
protein-tyrosine phosphatases are important regulators of many cellular
signaling processes (reviewed in Ref. 2). SHP-2 is broadly expressed and is important for activation signals through several different growth factor receptors. In contrast, SHP-1 is expressed predominantly in hematopoietic cells and has been implicated in inhibition of signaling through growth factor, cytokine, and antigen receptors. These
protein-tyrosine phosphatases contain two SH2 domains in tandem and a
single catalytic domain. The protein-tyrosine phosphatase activity is
negatively regulated by the SH2 domains in that their removal or
occupancy by phosphopeptides increases the phosphatase activity of the
catalytic domain (3-8). The SH2 domains of SHP-1 and SHP-2 are more
closely related to each other than to any other SH2 domains known, with
their next closest relatives being the SH2 domains of
phosphatidylinositol 3-kinase and phospholipase C-1 (9). It is of
particular interest to understand the specificity of these SH2 domains
because they control both the localization and activation of these
phosphatases.
The optimal binding sequences for SH2 domains have been defined with degenerate phosphopeptide libraries and pooled sequence analysis. The NH2-terminal SH2 domain (SH2N domain)1 of SHP-2 binds preferably to pY(I/V)X(V/I) (9), and the SH2N domain of SHP-1 binds preferably to pYFXF (10). The predictive value of these motifs was supported by the presence of appropriate sequences in several receptors that bind to the SH2N domain of SHP-1, such as c-Kit, and erythropoietin and interleukin-3 receptors. Due to a strong sequence similarity to the SH2N domains of both SHP-1 and SHP-2, the second SH2 domain (SH2C domain) of SHP-1 is predicted to have a similar preference for hydrophobic residues at positions +1 and +3 (10).
In contrast to the binding studies that considered only residues
downstream of the phosphotyrosine, we identified the motif (I/V)XYXXL by sequence alignment of inhibitory
receptors known to recruit SHP-1 (11). These include the B lymphocyte
receptors CD22 (12) and FcRIIB (13) and the family of killer cell
inhibitory receptors (KIR) expressed in natural killer and T
lymphocytes (11). Peptides derived from KIR and Fc
RIIB have been
reported to bind to the SH2C domain of SHP-1 (11, 13), whereas intact CD22 has been reported to interact with the SH2N domain (14). The motif
is also found in NKG2A (11, 15), a molecule associated with a receptor
complex recently shown to be involved in inhibition of human natural
killer cells (16), and in gp49, a molecule with inhibitory potential
expressed in mouse mast cells (17) and in natural killer cells (18). A
related motif, VXYXXV, was found in Ly-49
molecules (11), another family of inhibitory receptors that are
expressed in mouse natural killer cells (19). Phosphopeptides derived
from Ly-49 interact with the SH2 domains of both SHP-1 and SHP-2 (20).
The inhibitory receptors expressed in natural killer cells, including
human KIR and NKG2, as well as the mouse inhibitory receptors Ly-49 and
gp49 also share the sequence QEVT just upstream of the tyrosine.
Several recent observations have raised questions about the specificity
of the interaction of receptors with the
(I/V)XYXX(L/V) motif and SHP-1. SHP-1 was
implicated in the inhibitory signals of FcRIIB because of a defect
in the receptor's signaling in B cells from SHP-1-deficient mice (13).
Recently, it has been reported that Fc
RIIB functions in mast cells
independently of SHP-1 and associates with the inositol phosphatase
SHIP, which also contains an SH2 domain (21). Catalytically inactive
SHP-1 acts as a dominant-interfering molecule for the inhibitory
signals delivered by KIR (11), and SHP-1 associates with the receptor upon tyrosine phosphorylation (11, 20, 22, 23). However, it remained
possible that the dominant-interfering mutant of SHP-1 exerted its
effect by preventing the association of a different protein with
tyrosine-phosphorylated KIR, which was itself important in providing
the inhibitory signal. In addition, studies with metabolically labeled
cells show the association of KIR with a protein that corresponded in
size to SHP-1 as well as several unidentified proteins (23). To address
this issue, we have analyzed proteins that bind to synthetic
phosphotyrosine peptides corresponding to cytoplasmic tail sequences of
KIR and Fc
RIIB.
To determine whether the amino acids identified in the consensus sequence (I/V)XYXX(L/V) are important for interaction with SHP-1, we have examined the specificity of the interaction of sequences containing this motif with SHP-1 and the contribution of these residues to the activation and binding of SHP-1. The data established a critical role for the amino acid two residues upstream of the tyrosine in SHP-1 binding and activation.
All peptides were purchased from Quality Controlled Biochemicals, Inc. (Hopkinton, MA) and supplied at a purity of >98%. The phosphotyrosyl residue was incorporated during peptide synthesis. The peptides were synthesized by Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry, purified by reversed-phase high pressure liquid chromatography, and analyzed for purity by ion spray mass spectrometry and 31P NMR. Phosphopeptides were dissolved at 0.057 mM and coupled to Affi-Gel-10 beads as instructed by the manufacturer (Bio-Rad) at a ratio of 1 ml of peptide to 1 ml of settled beads. The sequence of peptide Y1Y2 is EQDPQEVTYAQLNHSVFTQRKITRPSQRPKTPPTDIIVYTELPNA. pY1Y2, Y1pY2, and pY1pY2 have the same sequence as Y1Y2 with the first, second, or both tyrosines phosphorylated, respectively. The sequence of pY1pY2-2A is EQDPQEAT(pY)AQLNHSVFTQRKITRPSQRPKTPPTDIAV(pY)TELPNA.
Silver Stain-Cell lysates were prepared from bulk human T cell populations cultured in the presence of 100 units/ml recombinant interleukin-2 (gift of Hoffmann-La Roche). The cells were lysed at 4 × 107/ml in lysis buffer (1% Triton X-100, 0.15 M NaCl, and 20 mM Tris, pH 8, with 1 mM NaVO3, 1 mM iodoacetamide, 5 µg/ml pepstatin A, and 1 mM phenylmethylsulfonyl fluoride). For each sample, a suspension of beads corresponding to 10 µg of peptide was added to 1 ml of lysate and mixed at 4 °C for 5 h. The samples were washed six times with lysis buffer, separated by 9% SDS-polyacrylamide gel electrophoresis under reducing conditions, and stained with Silver Stain Plus according to the manufacturer's instructions (Bio-Rad).
GST Fusion ProteinsThe fusion protein of GST and mouse SHP-1 or the SH2C domain of SHP-1 has been described (24). To generate a construct of GST and the catalytic domain of SHP-1, GST-catSHP-1, a cDNA fragment encoding amino acids 247-595 of SHP-1 was derived from a murine SHP-1 cDNA in the pGEX vector (Pharmacia Biotech Inc.) using polymerase chain reaction primers (forward primer, GAATTCGAGATGGAGTTTGAGAGTCTACAA; and reverse primer, TGTCAGAGGTGGGCACGTCA). The amplified cDNA fragment was digested with the restriction enzyme EcoRI and cloned into the EcoRI site of the pGEX vector. The fusion proteins were purified on glutathione-Sepharose 4B (Pharmacia) and eluted with 15 mM reduced glutathione. Purified GST-SHP-1 was dialyzed into phosphate-buffered saline, pH 7.4; GST-catSHP-1 was dialyzed into 0.15 M NaCl and 10 mM Tris-HCl, pH 8. The concentration of the fusion proteins was determined with the micro-BCA assay kit (Pierce).
Phosphatase AssayAssays using soluble GST-SHP-1 and the substrate p-NPP were performed essentially as described (11). Peptides were dissolved in a small volume of Me2SO; diluted in phosphatase buffer (100 mM NaCl, 1 mM dithiothreitol, 0.5 mM EDTA, and 100 mM Hepes, pH 7.4) to a final concentration of 285 µM peptide and 5% Me2SO; and then serially diluted in a microtiter plate. 25 µl of 20 mM p-NPP (Sigma) and 5 µl of GST-SHP-1 in phosphatase buffer were added and incubated at 37 °C for 10 min. The reaction was stopped by addition of 100 µl of 2 N NaOH.
The assay to measure release of inorganic phosphate from phosphopeptides was based on a malachite green detection system (25). The standard curve was determined with KH2PO4 (dried overnight at 80 °C prior to weighing). A stock solution for the standard was made by first dissolving in water and then diluting in assay buffer (100 mM Tris-HCl, pH 7.5, and 5 mM 2-mercaptoethanol). The peptides were dissolved in a small volume of Me2SO and then diluted in assay buffer to 5% Me2SO and 0.057 mM peptide. 11.4 nmol of peptide were incubated in the absence or presence of 0.1 µg of GST-catSHP-1 (in a final volume of 25 µl) at 22 °C for 10 min. 50 µl of the malachite green detection solution was added, and the absorbance at 650 nm was measured after 5 min.
SHP-1 Binding AssayCell lysate was prepared at a concentration of 107 Jurkat cells in 1 ml of lysis buffer (1% Triton X-100, 0.15 M NaCl, 2 mM NaVO3, 5 mM iodoacetamide, 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 20 mM Tris-HCl, pH 8). Peptide stocks were prepared at 0.57 mM peptide in 10% Me2SO and were diluted 1:1 in 2 × 0.3 M NaCl and 40 mM Tris-HCl, pH 8, and then serially diluted in 0.15 M NaCl and 20 mM Tris-HCl, pH 8. A 100-µl volume of peptide was added to 100 µl of lysate or GST-SH2C(SHP-1) diluted in lysis buffer. Beads bearing KIR-pY1 were added, and the mixture was incubated at 4 °C for 90 min. Samples were washed three times in lysis buffer and analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting with anti-SHP-1 antibodies or with anti-GST antibodies (Upstate Biotechnology, Inc., Lake Placid, NY) using the chemiluminescence detection system (Pierce).
To assess the relative specificity of the cytoplasmic tail
sequences of KIR and FcRIIB for SHP-1, proteins that associate with
phosphopeptides coupled to beads were analyzed by silver staining after
separation by SDS-polyacrylamide gel electrophoresis. The
phosphopeptide derived from the Fc
RIIB sequence bound two major
proteins from T cell lysates, corresponding in size to SHIP and SHP-1,
respectively (Fig. 1). In contrast, phosphopeptides derived from the first tyrosine in the KIR cytoplasmic tail or corresponding to the biphosphorylated KIR sequence each bound to a
single major protein that corresponded in size to SHP-1. Western
blotting confirmed the presence of SHIP at 155 kDa in the Fc
RIIB
peptide-associated proteins and of SHP-1 at 62 kDa in both the
Fc
RIIB and KIR peptide-associated proteins (data not shown). A
control tyrosine-phosphorylated peptide corresponding to a sequence in
the T cell receptor
-chain did not detectably bind to proteins in
cell lysates (Fig. 1). By this method, the tyrosine-phosphorylated
sequence from Ly-49 did not appear to bind to any proteins. Similar
data were obtained with lysate prepared from human natural killer cells
or from a human B cell line (data not shown). In experiments using
mouse cell lines as a source of lysate, weak bands were detected for
Ly-49 at 62 and 66 kDa (data not shown). Therefore, whereas all
sequences with the motif (I/V)XYXX(L/V) interact
with SHP-1, only Fc
RIIB was able to interact with both SHP-1 and
SHIP.
Activation of SHP-1 by Phosphotyrosine Peptides Derived from Lymphocyte Receptors That Bind SHP-1
To evaluate the importance
of the (I/V)XYXX(L/V) motif in the interaction
with SHP-1, synthetic phosphopeptides derived from the primary sequence
of several lymphocyte receptors (Table I) were tested
for their ability to activate SHP-1 in vitro. The peptides
activated SHP-1 with the hierarchy CD22-pY5 > KIR-pY1 Fc
RIIB > KIR-pY2
Ly-49-pY1 (Fig.
2A). The
-pY5 peptide derived from the
human T cell receptor-associated
-chain corresponds to part of the
immune receptor tyrosine-based activation motif that binds the SH2
domains of the ZAP-70 kinase (26). This peptide, containing the
sequence GLpYQGL, failed to activate SHP-1 (Fig. 2A),
suggesting that pYXXL is not sufficient for SHP-1
activation. These data correlate well with reported binding
measurements that indicated a higher affinity of SHP-1 for KIR-pY1 than
for KIR-pY2 and Ly-49-pY1 and a lack of binding to another
-chain-derived peptide (20).
|
To establish the core sequence necessary for activation of SHP-1 by the KIR-pY1 peptide, a set of shorter peptides based on the sequence of KIR-pY1 were tested for their ability to activate SHP-1. Whereas the four-residue peptide pYAQL was insufficient to activate SHP-1, the two amino acids upstream of the tyrosine in VTpYAQL conferred almost all of the activity of the original KIR-pY1 peptide (Fig. 2B).
Residues Upstream of Phosphotyrosine Are Important for Activation of SHP-1To directly test the role of residues at positions 1
and
2 in SHP-1 activation, a set of KIR-pY1 analogues were
synthesized with substitutions at these positions (Table
II). Substitutions of threonine at position
1 with
histidine (found in CD22-pY5) or leucine (found in
-pY5) did not
substantially affect SHP-1 activity (Fig. 2C). However, an
alanine in this position partially compromised the peptide's activity.
The effect of substitutions at position
2 provided clear evidence for
the importance of this residue in SHP-1 activation (Fig.
2B). Substitution of valine with isoleucine, the residue
found at this position in the CD22-pY5, Fc
RIIB, and KIR-pY2
peptides, increased the activation of SHP-1. A conservative
substitution with leucine did not significantly alter the activity of
the peptide. All other substitutions tested (alanine, serine,
aspartate, and arginine) seriously compromised the ability of the
peptide to activate SHP-1 (Fig. 2D).
|
To control for the possibility that position 2 had an influence on
dephosphorylation of the peptide by SHP-1, we tested the various
peptides in an assay that compared them as substrates of the isolated
phosphatase domain of SHP-1 (Fig. 3). Although differences in the ability of SHP-1 to dephosphorylate the various peptides were obvious, there was no correlation between resistance to
dephosphorylation by GST-catSHP-1 and the ability to activate full-length SHP-1. For example, pY1-2A, which failed to activate SHP-1,
is not a better substrate than KIR-pY1. Also of interest, pYAQL was a
worse substrate than KIR-pY1 (data not shown). Therefore, the influence
of position
2 is most likely at the level of interaction with the SH2
domains of SHP-1.
The hierarchy in the ability of peptides from different receptors to
activate SHP-1 and the conservation of the sequence QEVTYXXL in the receptors expressed in natural killer cells suggested that positions other than position 2 could be important for achieving optimal activation of SHP-1. Therefore, another series of peptides with
various substitutions in the KIR-pY1 sequence were generated (Table II)
and tested for their ability to activate SHP-1. Substitutions with
alanine at position
4 caused a partial reduction of SHP-1 activity
(data not shown), but not below levels obtained with the KIR-pY2 or
Ly-49 peptide. Individual substitution at position
6 and double
substitutions at positions
4 and
3 did not modify the ability of
the peptide to activate SHP-1 (data not shown).
The reduced activation of SHP-1 by Ly-49 as compared with
that by KIR-pY1 can be explained by their difference at position +3.
Substitution of leucine at position +3 in KIR-pY1 with valine (found in
Ly-49) diminished the activation of SHP-1 to the same level as that
obtained with the Ly-49 peptide (Table III). These data
indicate that the SHP-1 SH2 domains prefer leucine over valine at
position +3. Two chimeric peptides, initially designed to test whether
residues upstream of the YXXL motif were sufficient to dictate recognition by SHP-1, revealed that residues +1 and +2 and
perhaps residues +4 and +5 may also contribute to the interaction with
SHP-1. The chimeric peptides pY1/ and pY1/Y2 (Table II) behaved
similarly to
-pY5 and KIR-pY2, respectively (Table III). Therefore,
residues at positions +1, +2, +4, or +5 could interfere with the
interaction of SHP-1 with the motif VXYXXL.
|
The relationship between activation of SHP-1
and peptide binding to SHP-1 was examined. The majority of the peptides
listed in Table II were also tested for their ability to compete with the binding of SHP-1 to KIR-pY1 coupled to beads. Representative experiments are shown in Fig. 4. The nonactivating
peptide -pY5 did not compete with KIR-pY1, whereas soluble KIR-pY1
competed efficiently (Fig. 4A, top and
middle panels). Similar to the weaker activation of SHP-1 by
KIR-pY2 relative to that by KIR-pY1, the ability of KIR-pY2 to compete
with KIR-pY1 for binding to SHP-1 was below that of KIR-pY1 (Fig.
4A, middle panel). The ability of KIR-pY1
peptides with substitutions at position
2 to compete for binding also
correlated with their ability to activate SHP-1. These results
reiterated the importance of position
2 in the consensus motif
(I/V/L)XYXX(L/V). The effect of position
2 was also observed for interaction with the isolated SH2C domain of SHP-1
(Fig. 4B).
The only peptide for which the binding and activation data did not correlate perfectly was CD22-pY5. CD22-pY5 was the most potent monophosphopeptide for activation of SHP-1, whereas its efficiency in the competitive binding assay was equivalent to that of KIR-pY1. Given that KIR-pY1 binding was detectable with the isolated SH2C domain and not with the SH2N domain (11), the competition assay likely measures the relative affinity of peptides for the SH2C domain. However, CD22 has been reported to interact with the SH2N domain (14). An interaction of CD22-pY5 with both SH2C and SH2N domains of SHP-1 could explain why it is a better activator of SHP-1 because activation of SHP-1 by phosphopeptides has been reported to be more efficient through the SH2N domain (8, 27). On the other hand, the relative resistance of the CD22 peptide to the phosphatase activity of SHP-1 (see Fig. 3) may explain its higher performance in the protein-tyrosine phosphatase assay with p-NPP as a substrate.
Interaction of the KIR-derived peptides and the isolated SH2N domain of
SHP-1 was not readily detectable (11). However, biphosphopeptides
spanning both tyrosines in the KIR sequence bind SHP-1 with a greater
affinity than the corresponding monophosphopeptides (22), suggesting
that these biphosphopeptides may interact simultaneously with the SH2C
and SH2N domains. The ability of such long biphosphopeptides to
activate SHP-1 in vitro as compared with the mixture of the individual phosphopeptides was measured (Fig. 5). The
leftward shift of the curve demonstrated that the biphosphopeptide
pY1pY2 is the best activator of SHP-1. The much lower level of activity of the peptide pY1pY2-2A confirmed that position 2 is also important in the context of the biphosphopeptide. The activity of pY1Y2 and Y1pY2
was similar to that of the corresponding 14-amino amino peptides pY1
and pY2, respectively. Therefore, the high activity of pY1pY2 is due to
the phosphate moiety on both tyrosines and not to the intervening
sequence.
Phosphopeptides corresponding to KIR-pY1 or KIR-pY1pY2 bound a
single major protein in lymphocyte lysates that comigrated with SHP-1.
These observations support the conclusion that a dominant-interfering form of SHP-1 prevents KIR function by specifically competing with
SHP-1 (11, 28). In contrast, a phosphopeptide corresponding to the
FcRIIB cytoplasmic tail bound two major proteins, one of which
comigrated with SHP-1 and one with the inositol phosphatase SHIP. These
results support the reports of Fc
RIIB-mediated inhibition by either
SHP-1 or SHIP (13, 21). More important, these results indicate that the
(I/V)XYXXL consensus motif contains elements that
are important for binding and activation of SHP-1, but not SHIP.
Short peptides corresponding to the KIR-pY1 sequence showed that the
minimal consensus sequence (I/V)XYXXL was
sufficient to confer most of the activity of the KIR-pY1 peptide. The
short peptide VTpYAQL activated SHP-1, whereas pYAQL did not,
indicating a need for the upstream residues in the peptide to interact
with SHP-1. Studies with analogues of KIR-pY1 indicated that position 2 is the most influential position upstream as only KIR-pY1 analogues with Val/Ile/Leu hydrophobic residues at this position were able to
activate the SHP-1 protein-tyrosine phosphatase. Position
2 also
influences binding of the KIR-pY1 sequence to the isolated SH2C domain,
indicating that it is involved in the peptide binding to this SH2
domain. The utilization of upstream residues in a phosphotyrosine-binding motif is reminiscent of phosphotyrosine-binding domains, which are not related to SH2 domains, but similarly bind phosphotyrosine in a pocket created by conserved amino acids. The fine
specificity of phosphotyrosine-binding domain binding is controlled by
amino acids exclusively upstream of the phosphotyrosine, e.g. the motif PNXpY (reviewed in Ref. 1). The
(I/V/L)XYXXL motif for activation of SHP-1 is
unique in that it involves residues on both sides of the tyrosine.
The importance of residue 2 was clear in the context of the highly
active biphosphopeptide pY1pY2. This result is particularly relevant
because a long peptide is more likely than a short one to adopt the
conformation of the actual protein. In fact, the peptide tested
included 45 of the 76 amino acids of the KIR cytoplasmic tail. Although
short peptides rarely exhibit a fixed conformation in aqueous
environments (29), it is possible that the upstream residues invoke a
conformation favorable for the interaction of the peptide and the SH2
domain. However, such an explanation is unlikely because pYAQL was not
able to activate SHP-1.
The crystal structure of the SH2 domains of the related
protein-tyrosine phosphatase SHP-2 complexed with phosphopeptides suggest how position 2 may be involved in peptide binding to the SH2
domains of SHP-1. These structures revealed that the peptide-binding groove accommodates residues
2 to +5 relative to the phosphotyrosine (30, 31). In contrast to a peptide complexed with the SH2 domain of Src
(32), residue
2 is in contact with the SH2 domains of SHP-2 (30, 31).
Based on their sequence similarity to SHP-2, the SH2 domains of SHP-1
are likely to share the feature of an extended peptide-binding groove
relative to prototypic SH2 domains. The peptide from the
platelet-derived growth factor receptor crystallized with SHP-2 has a
valine at position
2. In the structure, the side chain of the valine
caps the phosphotyrosine-binding pocket. Although not all the peptides
that bind to SHP-2 contain a similar residue at position
2, the
valine at position
2 in the platelet-derived growth factor receptor
peptide has been shown to be important for binding and activation of
SHP-2 (33). It is possible that the interaction with the tyrosine ring
and the side chain of residue
2 provides an entropic contribution for
binding. Such an interaction may be required in general for the binding
of peptides to the SH2C domain of SHP-1. The phosphotyrosine position
in the SH2 domains of SHP-2 is different than in typical SH2 domains
(30) and may contribute to the need for the extra interactions.
Alternatively, a hydrophobic pocket in the SH2C domain of SHP-1, but
not present in SHP-2, could explain how Ile/Val/Leu at position
2
contributes to the binding of peptides to the SHP-1 SH2C domain.
Perhaps these interactions compensate for a lack of interaction with
position +1 and become necessary for the sequences from the lymphocyte receptors that lack a hydrophobic residue at position +1. Of note, there are interactions reported for KIR- and Ly-49-derived peptides with SHP-2, but these interactions are quite weak (20).
A long groove for peptide binding can impose stringent requirements for
many of the residues downstream of a phosphotyrosine. In the structural
studies of SHP-2, interactions of both crystallized peptides with the
protein occurred for side chains +3 and +5 (30). Such a long groove may
exclude the majority of peptides due to steric interference. In support
of this notion, the upstream portion of pY1 did not confer activity to
the downstream portions of KIR-pY2 and -pY5 in terms of SHP-1
activation. This suggests that
-pY5 and KIR-pY2 could lack other
necessary residues downstream of the tyrosine. The lack of consensus at
any of these positions in the sequences that do bind suggest that
-pY5 and KIR-pY2 possess residues that are detrimental to the
interaction. Whereas the two residues between the tyrosine and the
leucine are likely culprits, residues beyond position +3 have also been
implicated for high affinity binding of phosphopeptides to the related
SH2N domain of SHP-2 (34).
The issue of which SH2 domain is involved with recognition of residue
2 is intriguing. A specific interaction of the peptides with the SH2N
domain alone was not observed. Thus, it is tempting to presume that the
SH2C domain is the only relevant domain. However, the ability of the
biphosphopeptide to produce such strong activation as compared with the
monophosphopeptides suggests that both tyrosines and both SH2 domains
are involved in the interaction. The interaction with the SH2N domain
may be of low affinity and therefore is not detected in isolation.
However, when the two phosphotyrosines are linked, a high avidity
interaction is generated. An even more dramatic example of enhanced
activation by biphosphopeptides has been observed for SHP-2 (7). More
important, substitution with alanine at position
2 drastically
decreased the ability of the biphosphopeptide to activate SHP-1.
We have provided evidence that the consensus residues
(I/V/L)XYXX(L/V) are critical amino acids for
binding and activation of SHP-1. Ile/Val/Leu at position 2
differentiates the YXXL sequences of inhibitory receptors
from the YXXL sequences in the immune receptor
tyrosine-based activation motifs of lymphocyte-activating receptors.
This motif differs from the usual binding motifs of SH2 domains by the
inclusion of upstream residues.
We thank Alan Laur for technical assistance and D. McVicar and V. Flamand for comments on the manuscript.