(Received for publication, May 27, 1997)
From the Departments of Medicine, Immunology and
Medical Genetics and Microbiology, University of Toronto and the Samuel
Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario
M5G 1X5, Canada, the
Department of Molecular Oncology, M.D.
Anderson Cancer Center, Houston, Texas 77030, and the
Toronto Hospital Research Institute and the
Canadian Red Cross Society, Toronto, Ontario M5G 2M1, Canada
Activation of the cellular Src tyrosine kinase depends upon dephosphorylation of the carboxyl-terminal inhibitory tyrosine phosphorylation site. Herein we show that Src isolated from human platelets and Jurkat T cells is preferentially dephosphorylated at its inhibitory phosphotyrosine site by the SHP-1 tyrosine phosphatase. The data also revealed association of Src with SHP-1 in both platelets and lymphocytes and the capacity of Src to phosphorylate SHP-1 and interact with the SHP-1 NH2-terminal SH2 domain in vitro. Analysis of Src activity in thymocytes from SHP-1-deficient motheaten and viable motheaten mice revealed this kinase activity to be substantially lower than that detected in wild-type thymocytes, but to be enhanced by in vitro exposure to SHP-1. Similarly, immunoblotting analysis of thymocyte Src expression before and after selective depletion of active Src protein indicated that the proportion of active relative to inactive Src protein is markedly reduced in motheaten compared with wild-type cells. These observations, together with the finding of reduced Src activity in HEY cells expressing a dominant negative form of SHP-1, provide compelling evidence that SHP-1 functions include the positive regulation of Src activation.
Both the kinase and transforming activities of the
pp60c-src (Src) protein-tyrosine kinase
(PTK)1 are repressed by
phosphorylation of a tyrosine residue within the Src carboxyl terminus
(1-5). The negative regulatory effect of this phosphotyrosine (Tyr-530
and Tyr-529 in human and murine Src, respectively) has been ascribed to
its association with the Src SH2 domain and consequent SH2, as well as
SH3 domain-mediated intramolecular interactions that repress activity
of the kinase domain (6-8). The catalytic activities of the other Src
family members are similarly inhibited by phosphorylation of this
conserved C-tail tyrosine (9), and activation of each of these PTKs
thus depends on dephosphorylation at this site. However, while several lines of evidence implicate the Csk tyrosine kinase (10-12) as well as
Src-mediated autophosphorylation (13) in the phosphorylation of the
COOH-terminal regulatory tyrosine, relatively little is known about the
mechanisms whereby this inhibitory phosphotyrosine residue is
dephosphorylated and Src activation achieved. In contrast to the
Src-related Lck and possibly Fyn PTKs, the Src COOH-terminal phosphotyrosine does not appear subject to dephosphorylation by the
transmembrane protein-tyrosine phosphatase (PTP), CD45 (14-16). Moreover, while Src activity has been found to be increased in rodent
cell lines overexpressing the receptor-like protein-tyrosine phosphatase, RPTP (17, 18), a direct role for this PTP in dephosphorylating Tyr-529 in vivo has not been established.
However, recent data from studies of the SHP-1 tyrosine phosphatase, a cytosolic, dual SH2 domain-containing protein expressed primarily in
hemopoietic and epithelial cells (19-22), have raised the possibility that this PTP plays a role in the dephosphorylation and activation of
Src. Thus, for example, thrombin stimulation of platelets, which is
associated with an enhancement of Src activity as well as its rapid
translocation to the cytoskeletal fraction (23), has recently been
shown to induce increases in SHP-1 tyrosine phosphorylation,
association of SHP-1 with Src and redistribution of this PTP to the
cytoskeleton with kinetics similar to those observed for Src (24, 25).
These observations, together with data implicating the viral
transforming v-Src protein in the phosphorylation of SHP-1 (26),
suggest that, in at least some cell lineages, SHP-1 interacts with Src
and may serve to activate Src by dephosphorylating the COOH-terminal
regulatory tyrosine.
The monoclonal anti-Src 327 (27), anti-Src GD11 (28), and anti-SHP-1 (recognizing the COOH-terminal portion of SHP-1) antibodies were obtained from Dr. J. Brugge (ARIAD Pharmaceuticals, Cambridge, MA), Upstate Biotechnology Inc. (UBI, Lake Placid, NY), and Transduction Laboratories (Lexington, KY), respectively. The monoclonal anti-Src antibody, clone 28, which selectively recognizes the active, carboxyl-terminal-dephosphorylated form of Src (29), was a gift from Drs. J. Yano and K. Owada (Kyota Pharmaceutical University, Kyoto, Japan). The SRC2 rabbit polyclonal anti-Src antibody, which recognizes a similar epitope as clone 28 and was demonstrated in comparative studies to also selectively immunoprecipitate the active form of Src (data not shown), was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal anti-SHP-1 antibody recognizing the tandem SH2 domains of SHP-1 was generated in our laboratory as described previously (30). Rabbit anti-p56lck antibody was produced in our laboratory by immunizing rabbits with a bacterial TrpE fusion protein containing amino acids 7-144 of Lck (31). Rabbit muscle enolase (Boehringer Mannheim, Lavel, Canada) was acid-treated before use as described previously (32). Fyn and Lck enzyme purified from bovine thymus were obtained from UBI, while purified Src was purchased from UBI as a recombinant protein expressed in SF9 insect cells by baculovirus containing the human c-src gene.
Cells and Cell LinesThe human Jurkat leukemia and the MOLT 4 T cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells were cultured in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% (v/v) fetal bovine serum (Sigma) and 10 µM gentamycin (Life Technologies, Inc.). Cells were grown at 37 °C in a fully humidified incubator containing 5% CO2. Fresh human platelets and resting T cells were obtained from normal healthy peripheral blood buffy coats (Toronto Canadian Red Cross). Platelets were isolated from plasma after low speed centrifugation (PBL). T cells were isolated as peripheral blood lymphocytes (PBL) after purification of buffy coat blood on density gradients as described previously (33). Proliferating T cell lymphoblasts were prepared from PBL by stimulation with PHA (10 µg/ml; Difco) for 2 days followed with 100 units/ml recombinant interleukin-2 (Cetus) for an additional 2 days. The HEY ovarian adenocarcinoma cell line was obtained from ATCC and maintained in complete medium.
MiceSingle-cell suspensions of thymocytes were obtained from 10-14-day-old C3HeBFeJ-me/me (motheaten), C57BL6-mev/mev (viable motheaten), and congenic wild-type (+/+) mice derived at the Samuel Lunenfeld Research Institute facility by mating C3HeBFeJ me/+ and +/+ and C57BL/6J mev/+ and +/+ breeding pairs.
Immunoprecipitation and Western BlottingCells were lysed in 1% Nonidet P-40 lysis buffer (1% Nonidet P-40, 50 mM HEPES, pH 7.23, 150 mM NaCl, 50 µM NaF, 50 µM O-phosphate, 50 µM ZnCl2, 2 mM EDTA, 2 mM Na3VO4, 2 mM phenylmethylsulfonyl fluoride). Cell lysates were centrifuged at 14,000 × g for 10 min at 4 °C, and protein concentrations then determined by means of the bicinchoninic acid (BCA) assay (Pierce). Immunoprecipitation and Western immunoblots were performed as described previously (33). For protein detection, blots were blocked overnight in 5% nonfat milk for detection of Src, washed, incubated with the appropriate secondary antibody, and subjected to enhanced chemiluminescence (ECL) or incubation with 125I-Protein A. Where indicated, the immunoblots were stripped and reprobed with anti-Src or anti-SHP-1 antibody.
In Vitro Kinase AssayTyrosine kinase activity of Src was
assayed by immunoprecipitation of lysates as described above. The
immunoprecipitates were then washed in kinase wash buffer (50 mM HEPES, pH 7.23, 150 mM NaCl, 1 mM MgCl2, 1 mM MnCl2,
0.5% Nonidet P-40). The precipitates were then incubated in 20 µl of
kinase buffer containing 5 µCi of [-32P]ATP (ICN)
with or without 2 µg of acid-treated rabbit muscle enolase. The
mixture was incubated for 10 min at 37 °C, reduced SDS-gel sample
buffer added, and the samples boiled for 10 min. The samples were
centrifuged at 14,000 × g for 10 min prior to loading
supernatants onto SDS-PAGE gels containing 10-12% polyacrylamide. The
32P-labeled proteins were electrophoretically transferred
to Immobilon-P membranes (Millipore Corp., Bedford, MA) and the labeled
proteins visualized by autoradiography. Src quantitation was also
performed by anti-Src immunoblotting of the membranes using ECL. For
assays of Src kinase activity toward SHP-1, 10 ng of recombinant Src protein was incubated at 37 °C for 10 min in 20 µl of kinase
buffer containing 5 µCi of [
-32P]ATP in the presence
of acid-denatured enolase (2 µg), GST (10 µg), or GST-SHP-1 (2 µg) fusion protein. Samples were then boiled for 10 min in reduced
SDS-gel sample buffer, centrifuged at 14,000 × g for
10 min, and the supernatant proteins fractionated by SDS-PAGE. After
electrotransfer to Immobilon-P, Src was visualized by
autoradiography.
In vitro [-32P]ATP-labeled
or in vivo 32PO4-labeled Src was
immunoprecipitated with anti-pp60c-src 327 and
isolated by SDS-PAGE. After electrotransfer to nitrocellulose membranes, the 60-kDa Srccontaining band was excised from the membranes and subjected to CNBr cleavage as described previously (34).
The excised Src protein was incubated with 60 mg/ml CNBr in 70% formic
acid for at least 2 h at room temperature. Samples were then
washed and dried and the CNBr-generated peptide fragments resuspended
in Tricine SDS sample buffer, resolved by separation on 10-20%
gradient Tricine SDS-PAGE (Novex, San Diego, CA), transferred to
Immobilon-P membranes, and visualized by autoradiography. For phosphoamino acid analysis, CNBr-generated protein fragments were excised from Immobilon-P membranes, hydrolyzed in 6 M HCl
at 110 °C for 1 h and separated by one-dimensional phosphoamino
acid analysis on a cellulose plate as described previously (34-36).
Labeled phosphoamino acids were detected by autoradiography and
identified by comparison to ninhydrin-stained standards.
In vivo labeled Src for CNBr mapping studies was derived by incubating 4 × 108 human platelets or 5 × 107 Jurkat or MOLT 4 T cells for 2 h at 37 °C with 32PO4 (1 mCi/ml) in phosphate-free RPMI as described previously (23, 35, 36) followed by lysis in Nonidet P-40 lysing buffer and subsequent isolation of the labeled Src proteins as described above.
GST Fusion ProteinsThe glutathione
S-transferase (GST) fusion proteins used in this study were
derived by subcloning the following cDNA- or PCR-amplified fragments into pGEX2T: the full-length murine SHP-1 cDNA
(GST-SHP-1), a full-length murine SHP-1 cDNA containing a Cys-453
Ser mutation (GST-SHP-1 (C453S)), the SHP-1
NH2-terminal SH2 domain (amino acids 1-95), the SHP-1
COOHterminal SH2 domain (amino acids 110-205), and the SHP-1
NH2- and COOH-terminal SH2 domains (amino acids 1-221).
These GST expression plasmids were transfected into Escherichia coli JM101, and the fusion proteins purified from
isopropyl-1-thio-
-D-galactopyranoside-induced bacteria
with glutathione-conjugated Sepharose beads (Pharmacia Biotech Inc.).
GST-SHP-1 fusion protein was also obtained from UBI. For binding
studies, 1 µg of each GST fusion protein and GST beads was incubated
with 7 ng of in vitro 32P-labeled recombinant
Src protein at 4 °C for 2 h. The beads were then washed and
resuspended in SDS-sample buffer, and the proteins subjected to
fractionation over 10% SDS-PAGE, transferred to nitrocellulose, and
the 32P-labeled Src protein visualized by
autoradiography.
For phosphatase reactions using GST fusion proteins, proteins were eluted from equal amounts of GST-SHP-1 and GST-SHP-1 (C453S) bead complexes into 300 µl of elution buffer (10 mM reduced glutathione, 50 mM Tris-HCl, pH 8.0). Equal amounts of these eluted proteins were then incubated with immunoprecipitated 32P-labeled Src (in vitro or in vivo labeled) at 37 °C in 200 µl of phosphatase buffer (10 mM Tris-HCl, 1.0 mM EDTA, 1 mg/ml bovine serum albumin, 0.1% 2-mercaptoethanol, 0.01% NaN3, pH 7.34). To confirm the activity of the eluted proteins, activities of the PTPs were assayed in parallel using 10 mM p-nitrophenyl phosphate as substrate (37).
Transfection StudiesFor stable transfections of HEY cells, the plasmids pCMV4Neo, pCMV4Neo-SHP-1, and pCMV4Neo-SHP-1(C453S) were introduced by lipofection into HEY cells and the cells selected for Geneticin resistance. The resultant lines were assayed for expression of SHP-1 by immunoblotting analysis with rabbit anti-SHP-1 antibody and found to overexpress at similar levels either wild-type SHP-1 or phosphatase-inactive SHP-1 (C453S).
To determine whether the COOH-terminal or other phosphotyrosine
residues within Src are subject to dephosphorylation by SHP-1, the
effects of this phosphatase on Src tyrosine phosphorylation were
initially examined using CNBr cleavage analysis. As illustrated in Fig.
1A, CNBr treatment of
32P-labeled human Src has been shown previously to yield
phosphorylated cleavage fragments of about 31, 9.7, and 4.7 kDa, which,
respectively, contain the Src NH2-terminal region
encompassing the major sites for serine phosphorylation on Src, Ser-12
and Ser-17 (31-kDa fragment), the inhibitory tyrosine phosphorylation
site, Tyr-530 (4.7-kDa fragment), and a key site for
autophosphorylation on activated Src, Tyr-419 (9.7-kDa fragment; Refs.
14-16, 23, and 38-41). As is consistent with these data, Tricine
SDS-PAGE and subsequent phosphoamino acid analysis of the cleavage
fragments derived by CNBr hydrolysis of Src immunoprecipitates from
32P-labeled human T cells (Molt 4 and Jurkat) and
freshly-isolated platelets also revealed a predominantly
phosphoserine-containing 31-kDa and predominantly
phosphotyrosine-containing 9.7- and 4.7-kDa fragments (Fig. 1,
B and C). However, in addition to these
fragments, the CNBr cleavage products generated in this analysis also
included a phosphorylated fragment of about 3.3 kDa (Fig.
1B), which was shown by phosphoamino acid analysis to
contain primarily phosphotyrosine (Fig. 1C). Based on the
apparent molecular weight of this fragment and the locations of
methionine residues within Src (Fig. 1A), the 3.3-kDa
cleavage fragment most likely corresponds to amino acids 315-341
within the Src catalytic domain, a possibility that suggests Tyr-338, a
residue recently shown to become phosphorylated in the context of
in vitro phosphate labeling of Src (42), represents another
site of Src tyrosine phosphorylation in vivo and as such may
also be of relevance to the expression of Src kinase activity. By
contrast, although one other site of tyrosine phosphorylation (corresponding to Tyr-216 in human Src) has recently been identified within the 31-kDa CNBr fragment (43), this phosphotyrosine was not
detected under the conditions used in this current analysis.
As CNBr cleavage analysis under the conditions used in this study
distinguished the Src region containing Tyr-530 as well as the other
major known sites for tyrosine phosphorylation on Src, this approach
was next used to determine whether any of these sites represent targets
for SHP-1-mediated dephosphorylation in vitro. To this end,
32P-labeled recombinant Src was incubated with GST fusion
proteins carrying SHP-1 or, as a control, a mutated (C453S),
catalytically inert form of SHP-1, and the effects of the latter
proteins on the profile of phosphorylated CNBr Src cleavage fragments
then examined. As shown in Fig.
2A, the phosphorylated 9.7-, 4.7-, and 3.3-kDa CNBr fragments known to represent
phosphotyrosine-containing regions of Src were again detected following
CNBr hydrolysis of Src pretreated with inactive SHP-1 (C453S). By
contrast, intensities of the bands representing each of these
32P-labeled species, but most notably the 4.7- and 3.3-kDa
phosphorylated fragments, were markedly reduced in the context of Src
pretreatment with SHP-1, a result which suggests that the
phosphotyrosine residues located in these fragments can be
dephosphorylated in vitro by SHP-1. Along similar lines,
SHP-1 treatment of in vitro 32P-labeled Src
immunoprecipitates from human platelets was also associated with
dramatic decreases in the signal intensities and, by extension,
phosphorylation states of the three Tyr-containing CNBr-generated
cleavage fragments (Fig. 2B). As shown in Fig. 2B, phosphorylation of the Src regions represented by the
4.7- and 3.3-kDa cleavage fragments again appeared to be more sensitive to SHP-1 pretreatment than was phosphorylation of the 9.7-kDa fragment.
A further assessment of these changes by densitometric analysis
revealed the reduction in phosphorylation levels following a 30-min or
60-min incubation of Src with SHP-1 to be about 8-10-fold and 2-fold,
respectively, greater for the 4.7- and 3.3- than for the 9.7-kDa
fragment (Fig. 2C). Thus, while these data suggest the
capacity of SHP-1 to dephosphorylate in vitro Tyr-419,
Tyr-338, and Tyr-530, the major sites for tyrosine phosphorylation on
the 9.7-, 3.3-, and 4.7-kDa cleavage fragments, respectively, these findings also indicate that SHP-1 exerts a more pronounced effect on
the phosphorylation of Tyr-338 and Tyr-530 than on Tyr-419.
To determine whether SHP-1 also dephosphorylates Src phosphotyrosine residues labeled in vivo, additional CNBr cleavage studies were performed on SHP-1-treated Src immunoprecipitates derived from 32P-biosynthetically labeled human platelets and Jurkat T cells. As shown in Fig. 2D, pretreatment of the in vivo labeled Src with catalytically inert SHP-1 (C453S) was again associated with the generation of the three phosphorylated CNBr cleavage fragments known to contain key residues for Src tyrosine phosphorylation. However, in contrast to the data obtained by analysis of in vitro labeled Src protein, SHP-1 treatment of the in vivo labeled Src immunoprecipitates was associated with a reduction in the phosphorylation of the 4.7-kDa CNBr-derived cleavage fragment, but had little effect on phosphorylation status of the 9.7-and 3.3-kDa fragments. As indicated by the phosphoamino acid analysis of these respective segments shown in Fig. 1C, this difference in the susceptibility of in vivo versus in vitro 32P-labeled Src to SHP-1 catalytic activity cannot be ascribed to differences in the phospholabeling of serine/threonine residues and may therefore reflect in vivo labeling of additional tyrosine residues not efficiently phosphorylated under in vitro conditions. These residues might include, for example, tyrosines 329 and 343, which are also contained within the 3.3-kDa fragment and which are not susceptible to the action of SHP-1. While this latter issue requires further investigation, these data are consistent with the hypothesis that Src represents a substrate for SHP-1 and raises the possibility that SHP-1 mediates dephosphorylation of the COOH-terminal inhibitory phosphotyrosine site.
In view of the data suggesting SHP-1 involvement in the
dephosphorylation and potentially regulation of Src, the possibility that these two enzymes physically associate with one another in activated lymphocytes was next investigated. As shown in Fig. 3A (upper panel),
anti-Src immunoblotting analysis of SHP-1 immunoprecipitates prepared
from interleukin-2-stimulated peripheral blood lymphocytes, revealed an
association of SHP-1 with Src in these cells, although comparison of
the amount of Src present in total cell lysates versus the
amount coprecipitated with SHP-1 suggests that the fraction of total
cellular Src associated with SHP-1 is very low. SHP-1-Src association
was also observed in reciprocal experiments involving anti-SHP-1
immunoblotting analysis of anti-Src immunoprecipitates (Fig.
3A, lower panel) and in similar studies of Jurkat
T lymphocytes and human platelets (data not shown), but was
consistently not detected when cell lysates for these studies were
prepared in buffer lacking orthovanadate, a potent tyrosine phosphatase
inhibitor (Fig. 3A). Together these observations link the
association of SHP-1 with Src to a phosphotyrosine-dependent
interaction and, as is consistent with previous data showing that
thrombin stimulation induces both association and increased tyrosine
phosphorylation of platelet SHP-1 and Src (24), suggest that the
interaction of these enzymes may be mediated through the binding of one
or both of the SHP-1 SH2 domains with Src phosphotyrosine residues. To
address this issue, GST fusion proteins carrying full-length SHP-1 and
SHP-1 (C453S) or the tandem or single SHP-1 SH2 domains were evaluated
for their capacities to interact with phosphorylated recombinant Src.
The results of this in vitro analysis revealed specific
binding of phosphorylated Src (Fig. 3B), but not
phosphorylated Fyn or Lck to the SHP-1 (C453S) fusion protein.
Furthermore, as shown in Fig. 3C, Src was also precipitated
by the fusion proteins containing both SH2 or the
NH2-terminal SHP-1 SH2 domain, but not by the fusion
protein containing only the COOH-terminal SHP-1 SH2 domain. Thus, as
has also been described in relation to SHP-1 interactions with other
putative substrates, such as the erythropoietin and c-Kit receptors
(44, 45), these observations suggest that a single (in this instance,
the NH2-terminal) SH2 domain mediates the binding of SHP-1
to phosphotyrosine(s) on Src. Conversely, GST fusion proteins
containing the Src SH2 domain have been shown to precipitate
phosphorylated SHP-1 from thrombin-activated platelets (24), and it is
therefore possible that SHP-1 association with Src involves multiple
sites of interaction on these respective proteins. However, in view of
our results indicating that GST-SHP-1 (C453S) fusion proteins also
precipitate Src from peripheral blood lymphoblast lysates containing
orthovanadate, but not from lysates lacking orthovanadate (data not
shown) as well as data showing SHP-1 activity to be substantially
increased by the binding of its NH2-terminal, but not
COOH-terminal SH2 domain to tyrosine-phosphorylated ligands (46), the
interaction of the SHP-1 NH2-terminal SH2 domain with Src
appears to represent a physiologically relevant molecular association,
which may facilitate SHP-1 recruitment to and ultimately
dephosphorylation of Src.
While the analysis of SHP-1 effects on Src tyrosine phosphorylation identify Src as a potential substrate for this PTP, previous data showing that SHP-1 can be tyrosine-phosphorylated by v-Src (26), the product of the transforming gene of Rous sarcoma virus, suggest that SHP-1 may also represent a substrate of the cellular Src homologue. To address this possibility, the capacity of recombinant Src to phosphorylate GST-SHP-1 fusion proteins was investigated using an in vitro kinase assay. As shown in Fig. 3D, the results of this analysis revealed the induction of SHP-1 tyrosine phosphorylation by Src, thereby identifying SHP-1 as a potential substrate for this kinase. SHP-1 has also been shown to be a substrate for Lck (47) and ZAP-70 (48), and the respective relevance of these PTKs to SHP-1 phosphorylation in vivo thus requires further analysis. However, in view of data showing SHP-1 activity to be enhanced by tyrosine phosphorylation (49), the current data raise the possibility that a reciprocal functional relationship exists between SHP-1 and Src, the physical association of these proteins allowing each enzyme to activate the other in either a coordinate or sequential fashion. Along similar lines, binding of SHP-1 to another putative PTK substrate (ZAP-70) has been shown to induce changes in the catalytic activities of both enzymes, although in contrast to Src, ZAP-70 kinase activity appears to be diminished in conjunction with increases in SHP-1 phosphatase activity (48).
In addition to SHP-1, a number of other PTPs, including most recently
the structurally similar SHP-2 protein, have been shown to
dephosphorylate Src in vitro at the COOH-terminal regulatory tyrosine (17, 18, 50, 51), but the extent to which these respective
enzymes contribute to Src activation in vivo remains unclear. To specifically ascertain the biologic relevance of SHP-1 to
the regulation of Src, activity of this PTK was next investigated in
unstimulated thymocytes from motheaten (me) and viable
motheaten (mev) mice, animals that express
negligible SHP-1 catalytic activity consequent to loss-of-function
mutations in the SHP-1 gene (30). As is consistent with previous data
from our group (33), the results of anti-Src immunoblotting analysis
confirmed the presence of Src in thymocytes from both mutant and
normal, congenic mice. However, while levels of Src protein appeared
equivalent in the SHP-1-deficient and normal thymocytes, the in
vitro kinase activity of Src as evaluated by autophosphorylation
following immunoprecipitation was markedly lower in the me
and mev cells compared with normal thymocytes (Fig.
4A). By contrast, levels of
Lck autophosphorylation and protein after immunoprecipitation were no
different in the me and wild-type cells (Fig.
4B). These findings are consistent with the in
vitro data demonstrating the capacity of SHP-1 to dephosphorylate
the Src COOH-terminal inhibitory tyrosine and, although the biologic
functions of Src in thymocytes are currently unclear, these data
strongly suggest a role for SHP-1 in the activation of Src in
vivo. This contention is also supported by the finding that SHP-1
incubation with Src immunoprecipitates from me thymocytes
substantially increased the extent of Src autophosphorylation in
vitro (Fig. 4C), a result that again links SHP-1 to not
only the dephosphorylation, but also the activation of Src.
To confirm the association of the me mutation with reduced Src activity, Src protein was immunoprecipitated from me and wild-type thymocytes using antibodies (clone 28 and SRC 2) raised against the tyrosine dephosphorylated Src COOH terminus and shown previously to selectively recognize the active form of the enzyme (29). As illustrated in Fig. 4D (panel II), the results of an in vitro kinase assay using these respective immunoprecipitates revealed Src activity to be detectable in the wild-type, but not in the me-derived immunoprecipitates. As immunoblotting analysis of the me cell lysates with the GD11 (Fig. 4D, panel I), or 327 (data not shown) anti-Src antibodies, which react with both the inactive and active forms of Src, demonstrated comparable levels of total Src in me and wild-type cells, these data indicate that me thymocyte Src protein is not immunoprecipitable by the clone 28 and SRC 2 antibodies and therefore imply that the Src present in me thymocytes is largely inactive. As is consistent with this contention, re-immunoprecipitation of Src from supernatants of clone 28 or SRC2 immunoprecipitates using the 327 anti-Src antibody revealed the me, but not the wild-type immunoprecipitates, to contain a substantial level of Src, presumably representing the inactive protein (Fig. 4D, panel III). These findings, together with the demonstration that the me mutation does not alter expression of other prominent thymocyte PTPs, such as CD45 and SHP-2 (Fig. 4E), provide strong evidence that SHP-1 deficiency impairs Src activation in vivo.
As SHP-1 is expressed in epithelial as well as hemopoietic lineages, the role for SHP-1 in Src activation was also investigated by evaluating Src activity in HEY ovarian cancer cells transfected with expression constructs containing cDNAs for either wild-type SHP-1 or SHP-1 (C453S), the latter of which has been shown to function in a dominant negative fashion in some cell types (48, 52). Compared with HEY cells transfected with vector alone, levels of SHP-1 protein in the cells stably transfected with SHP-1 and SHP-1 (C453S) were found to be increased by about 2-fold (data not shown). As shown in Fig. 4F, analysis of the in vitro kinase activity of Src protein immunoprecipitated from these transfectants revealed the phosphorylation of both Src as well as exogenous enolase substrate to be relatively reduced in the SHP-1 (C453S) overexpressing cells compared with cells overexpressing wild-type SHP-1 or transfected with vector alone. Following epidermal growth factor stimulation of these cells, Src activity again appeared lower in the SHP-1 (C453S) expressing cells than in the empty vector transfectants, a comparison of enolase phosphorylation with Src protein levels suggesting about 1.5-2 fold reduction in Src kinase activity in the former relative to latter cells both before and after stimulation. By contrast, phosphorylation of a number of species coprecipitated with Src from the epidermal growth factor-treated cells was markedly increased in the wild-type SHP-1-expressing relative to empty vector-transfected cells, suggesting that Src kinase activity on endogenous substrates is enhanced in the context of SHP-1 overexpression. Together these findings strongly suggest that the effects of SHP-1 on Src tyrosine phosphorylation are relevant to the activation of Src not only in hemopoietic, but also in epithelial cell lineages.
The data presented herein, revealing the capacity of SHP-1 to dephosphorylate Src preferentially at the COOH-terminal negative regulatory tyrosine as well as a positive correlation between the expression of SHP-1 and Src activity in vivo, provide compelling evidence that SHP-1 plays a role in the activation of Src in vivo. The identification of Src as a target for positive modulation by SHP-1, a PTP previously implicated in the negative regulation of a spectrum of signaling effectors, including the ZAP-70 and JAK2 tyrosine kinases (48, 53), indicates that SHP-1 effects on cell behavior may be realized through enhancement as well as inhibition of intracellular signaling cascades. Thus, while the structural basis for and physiologic sequelae of SHP-1-Src association require further investigation, these data suggest SHP-1 regulatory effects on Src may be relevant to the genesis of human epithelial and potentially other cancers associated with increases in Src activity.
We thank Drs. J. Brugge, A. Eberhard, P. Johnson, M. Kon-Kozlowski, W. J. Muller, K. Owada, and J. Yano for provision of reagents.