|
|
|
|||||||
|
|||||||||
(Received for publication, October 18, 1994; and in revised form, December 21,
1994) From the
Fibroblasts transformed by the v-Src oncoprotein exhibit
elevated activity of the enzyme phosphatidylinositol 3`-kinase (PI
3-kinase), which binds to, and is activated by, a wide range of
receptor tyrosine kinases as well as v-Src and transforming polyoma
middle T/c-Src complexes. Here we consider the role of the v-Src
homology (SH) domains, SH3 and SH2, and the tyrosine kinase catalytic
domain, in the stimulation of v-Src-associated PI 3-kinase activity in
response to rapid activation of the oncoprotein. As shown by others, we
find that the v-Src SH3 domain tightly binds the PI 3-kinase p85
regulatory subunit in normal growing chicken embryo fibroblasts.
However, we also find that in transformed cells there is additional
efficient binding of PI 3-kinase to the v-Src SH2 domain in a
catalytically active form. Furthermore, the binding of p85 to the SH2
domain, which is almost undetectable in quiescent cells, is rapidly
stimulated upon activation of temperature-sensitive v-Src and
consequent cell cycle entry, demonstrating that binding is a target for
regulation. We also show that v-Src-associated PI 3-kinase differs
considerably from PDGF receptor-associated enzyme by a different mode
of binding, a lack of substantial allosteric activation, and a
dependence on the tyrosine kinase activity of v-Src. The rapidly
induced binding and activation of PI 3-kinase thus provides sensitive
regulation of recruitment of PI 3-kinase to its substrates and into
other signaling complexes at the cell membrane, which involves all the
Src homology domains. The c-src proto-oncogene, the cellular version
of the v-src oncogene of Rous sarcoma virus, encodes
the prototype of a family of cytoplasmic protein tyrosine kinases
associated with the plasma membrane (reviewed by Courtneidge et
al.(1993) and Fincham et al. (1994)). Src is also the
prototype of proteins containing Src homology (SH) ( There is
no detailed understanding of the crucial signaling events that lead to
transformation by the v-Src oncoprotein but it is clear that v-Src,
transforming variants of c-Src, and activated tyrosine kinase growth
factor receptors, such as the one for platelet-derived growth factor
(PDGF), share common binding partners like PI 3-kinase, and known
substrates such as Shc (Fukui and Hanafusa, 1989; McGlade et
al., 1992a, 1992b). Much recent attention has focused on PI
3-kinase as an important binding partner and target for v-Src (reviewed
by Cantley et al., 1991, Downes and Carter, 1991, Parker and
Waterfield, 1992, and Stephens et al., 1993). This lipid
kinase can phosphorylate phosphatidylinositol (PI), PI 4-phosphate, and
PI 4,5-bisphosphate in position D-3 of the inositol ring, giving rise
to products whose cellular role is still little understood. The only
clue as to their function comes from in vitro experiments in
which PI 3,4,5-trisphosphate was found to activate protein kinase
C PI 3-kinase is a
heterodimer of a 110-kDa catalytic subunit (p110) and an 85-kDa
regulatory subunit (p85) (reviewed by Fry and Waterfield, 1993). p110
is a dual specificity enzyme, able to act not only as a lipid kinase,
but also as a protein serine kinase. The phosphorylation on serine 608
in the inter-SH2 domain of p85 by p110 negatively regulates PI 3-kinase
(Dhand et al., 1994). PI 3-kinase was first identified in
anti-v-Src immunoprecipitates from lysates of Rous sarcoma
virus-infected chicken fibroblasts (Sugimoto et al., 1984). It
is also found associated with members of the protein tyrosine kinase
family of growth factor receptors (Kaplan et al., 1987; Bjorge et al., 1990) and activated by proteins of the hemopoietic
growth factor receptor family (Gold et al., 1994). The
widespread activation of PI 3-kinase by growth factor receptors points
to a common, fundamental function of this enzyme in growth factor
signaling. There is strong evidence implicating PI 3-kinase in cell
transformation. The transforming ability of the polyoma middle T
antigen/c-Src complex depends on its association with PI 3-kinase
(Whitman et al., 1985) and PI 3-kinase binds to v-Src as well
as all transforming variants of c-Src (Fukui and Hanafusa, 1989).
Mutations in the v-Src SH3 domain decrease the binding of PI 3-kinase
and result in a partially transformed, fusiform morphology (Wages et al., 1992), suggesting that the v-Src SH3 domain may
mediate the association with PI 3-kinase. More recently, PI 3-kinase
was shown to bind directly to the v-Src SH3 but not the SH2 domain
expressed as glutathione S-transferase (GST) fusion proteins
(Liu et al., 1993). In successive rounds of
anti-phosphotyrosine immunoprecipitations from cell extracts, however,
the PI 3-kinase activity able to bind v-Src is depleted (Fukui and
Hanafusa, 1991). This implies that tyrosine phosphorylation of the
lipid kinase is required for efficient association with v-Src,
presumably mediated by the v-Src SH2 domain. Thus, the exact role of
the v-Src SH domains in binding and activation of PI 3-kinase and the
mechanism of regulation of binding remain unclear. Here, we investigate
in detail the role of the v-Src homology domains, SH2 and SH3 alone and
in combination, and the tyrosine kinase domain, in the regulation of
v-Src-associated PI 3-kinase. We conclude that all of these domains
contribute and that the regulation of PI 3-kinase bound to v-Src
differs from that associated with activated PDGF receptor.
For PI
3-kinase activation studies, biotinylated PDGF receptor peptide
(donated by Zeneca) or GST fusion proteins on glutathione-agarose beads
were added to CEF lysates and incubated at 4 °C for 90 min. PDGF
receptor peptide-bound PI 3-kinase was precipitated using
streptavidin-agarose beads. Beads were washed and PI 3-kinase assays
performed as described below.
Figure 1:
Binding of p85 to the
non-catalytic domains of v-Src. A, p85 from 500 µg of
protein prepared from normal growing CEF binding 2.5 nmol of GST (lane1) or 0.5, 1.0, and 2.5 nmol of SH3 (lanes2-4), SH2 (lanes 5-7), and fused
SH3-SH2 (lanes 8-10) was detected by immunoblotting. Two
different exposures are presented. B, densitometric
quantitation of p85 binding using the longer exposure. C,
immunoblot showing inhibition of p85 binding to 2.5 nmol of SH3 (lanes 1-3), SH2 (lanes 4-6), and SH3-SH2 (lanes 7-9) by 0, 100, and 250 µM proline-rich peptide from p85, respectively. Two different
exposures are shown. D, densitometric quantitation of p85
binding using the longer exposure.
The enhancing effect
of fusing the SH2 domain to SH3 could be either the result of direct
binding of a population of p85 to SH2 or stimulation of binding to SH3
as a result of some conformational change. Thus, we examined the effect
of the proline-rich peptide (amino acid residues 84-98 of bovine
p85-
Figure 2:
PI 3-kinase activity binding to 0.2 nmol
of SH3, SH2 and SH3-SH2 was carried out as described under
``Materials and Methods,'' in the presence (
Figure 3:
A, activation of PI 3-kinase as a
consequence of equal volumes of excess CEF protein binding to 0.5 nmol
of PDGF receptor phosphopeptide and v-Src SH3, SH2, and SH3-SH2
domains. B, comparison of PI 3-kinase activity, relative to
amount of associated p85, bound by the PDGF receptor (PDGFR)
peptide and v-Src SH3 domain. Agarose beads (control) and
GST-agarose beads (GST) were included as controls. p85
immunoblots for bound proteins are shown in lanes
1-4.
Figure 4:
A, PI 3-kinase activity
associating with ts v-Src proteins. v-Src was
immunoprecipitated using 2 µg of EC10 from equal volumes of lysates
from CEF expressing RCAN vector only, RCAN-LA 29 (R29), RCAN-LA 29A2 (R29A2), and RCAN-LA 32 (R32) at restrictive and permissive temperatures. PI
3-kinase assays were performed on washed immunoprecipitates. B, PI 3-kinase activity associating with the kinase-defective (KD) LA 32 v-Src protein and the wild type (WT) positive control grown at 35 °C were similarly
determined. As negative control, lysates from CEF were also
immunoprecipitated with EC10. C, part A,
total cell lysates were immunoprecipitated with anti-p85, separated by
7.5% SDS-PAGE and immunoblotted with anti-phosphotyrosine; part
B, p85 immunoblot of total cell lysates. Shown are lysates from
cells expressing the following v-Src proteins: kinase-defective (lane1, KD), RCAN vector at permissive (P) and restrictive (R) temperatures (lanes2 and 3, respectively), RCAN-LA 29 (R29) permissive and restrictive (lanes4 and 5), RCAN-LA 29A2 (R29A2) permissive
and restrictive (lanes6 and 7), and
RCAN-LA 32 (R32) permissive and restrictive (lanes7 and 8).
Figure 5:
A, stimulation of
v-Src-associated PI 3-kinase activity in response to activation of ts v-Src kinase by temperature shift. The PI 3-kinase activity
associated with immunoprecipitated v-Src was determined from quiescent
CEF expressing ts LA 29 at restrictive temperature and
stimulated by shift to permissive temperature for 10, 20, 30, and 60
min. B, immunoblot of total p85 (upperpanel) and tyrosine-phosphorylated p85 (lowerpanel) in CEF upon activation of v-Src for 0, 10, 20, 30,
and 60 min.
In order to determine whether this increase was
attributable to the SH3 or SH2 domain of v-Src, we examined the PI
3-kinase activity associated with GST-SH3 or GST-SH2 in response to
activation of the v-Src kinase by temperature shift. The increase in
v-Src-associated PI 3-kinase activity observed is almost certainly due
to stimulated binding to the SH2 domain, since binding to GST-SH2 was
similarly enhanced while binding to GST-SH3 was invariant (Fig. 6A). To test our hypothesis that tyrosine
phosphorylation of p85 provides a mechanism for association with v-Src
via the SH2 domain, lysates of CEF expressing ts LA 29 v-Src
at permissive temperature were depleted of phosphotyrosine-containing
proteins by immunoprecipitation with anti-phosphotyrosine antibody
before incubation with GST-SH2 or GST-SH3. We found that p85 binding to
SH2 but not SH3 requires tyrosine phosphorylation of p85 (Fig. 6B). In this experiment, the SH3 domain bound a
small amount of the faster migrating form of p85 (Fig. 6B, top, short exposure). This was seen
in some of our binding experiments and may reflect variations in the
state of the cells or the experimental conditions.
Figure 6:
A, PI 3-kinase activity
associated with GST-SH3 and GST-SH2 over a 60-min time course of v-Src
stimulation. B, p85 binding to GST-SH2 or GST-SH3 without
(-) or with (+) depletion of phosphotyrosine-containing
proteins by anti-phosphotyrosine immunoprecipitation of lysates from
CEF expressing ts LA 29 at permissive temperature. Two
exposures are shown.
To establish
whether activation of the v-Src kinase directly regulates binding to
the SH3 and SH2 domains, we prepared protein extracts from CEF
expressing ts LA 29 v-Src, which had been made quiescent at
restrictive temperature and stimulated by shift to permissive
temperature for various times. The ability of p85 to bind to GST-SH3,
GST-SH2, and GST-SH3-SH2 was determined by incubation of the extracts
with the bacterial fusion proteins, followed by elution and detection
by immunoblotting (Fig. 7). The ability of p85 to bind the SH3
domain was detectable in quiescent cultures (Fig. 7, lowerpanel (long exposure), lane1) and
largely unaffected by the activation of the v-Src kinase (closedarrow; Fig. 7, lanes4, 7, and 10), consistent with no regulation of the
proline-rich sequence-mediated binding upon cell stimulation. However,
p85 binding to the SH2 domain, which was detectable from lysates of
growing CEF, was almost undetectable in quiescent cultures (Fig. 7, lowerpanel (long exposure), lane2), was rapidly stimulated by 30 min after shift to
permissive temperature (Fig. 7, lane5) and
remained elevated at 60 min and 120 min (Fig. 7, lanes8 and 11). In particular, the faster migrating
form of p85 (arrow) was the more abundant species associating
with the SH2 domain early after induction (Fig. 7, lowerpanel (long exposure), lanes5 and 8). Although the slower migrating form of p85 (closedarrow) from unstimulated cells bound the SH3-SH2 fusion,
p85 binding to SH3-SH2 was responsive to activation of the v-Src kinase
by stimulation of binding of the faster migrating form of p85 (arrow; Fig. 7, upperpanel (short
exposure), lanes6, 9, and 12). We
have shown that this species preferentially, but not exclusively, binds
the v-Src SH2 domain (Fig. 1, 6B, and 7), implying that
the increase in its binding to the SH3-SH2 fusion in response to
activation of the v-Src kinase is mediated by the SH2 domain. Thus, the
contribution of the SH2 domain to the binding of the regulatory subunit
of PI 3-kinase to the non-catalytic domains of v-Src varies according
to the physiological growth state of the cells; in quiescent (Fig. 7) and exponentially growing normal cells (Fig. 1),
binding is largely mediated by the SH3 domain, while in transformed
cells soon after stimulation with an activating tyrosine kinase, the
SH2 domain contributes significantly to p85 binding.
Figure 7:
p85 from quiescent CEF. Immunoblot shows
p85 from quiescent CEF expressing ts LA 29 at restrictive
temperature (41 °C) (lanes 1-3) and stimulated by
shift to permissive temperature (35 °C) for 30 (lanes
4-6), 60 (lanes7-9) and 120 min (lanes 10-12), associating with the v-Src SH3, SH2, and
SH3-SH2 (SH3/2) domains. Two exposures are
shown.
Studies to date have implicated either the SH3 (Wages et
al., 1992; Liu et al., 1993) or the SH2 domain (Fukui and
Hanafusa, 1989) of v-Src in the interaction with PI 3-kinase. This led
us to examine in detail the requirements for the binding and activation
of PI 3-kinase by v-Src and to compare them to those proposed for the
PDGF receptor. In the experiments presented, p85 from growing normal
CEF bound to the v-Src SH3 domain and, although less strongly, to the
SH2 domain. The SH2 domain bound a faster migrating form of p85 in
addition to a species that co-migrated with the SH3-bound form. Both
species were phosphorylated on tyrosine in cells transformed by v-Src.
We also observed enhanced binding of p85 to fused SH3 and SH2 domains,
as present in the intact v-Src protein. In line with these findings, an
apparent synergism has been proposed for p56 The proposed mode of interaction between PI 3-kinase and
v-Src is quite distinct from that suggested for the binding of p85 to
the PDGF receptor, which is mediated by interaction of the p85 SH2
domains with phosphotyrosine residues in the receptor tail. In addition
to the different modes of binding, the apparent regulation of PI
3-kinase activity as a consequence of binding was also different. In
the case of binding to a tyrosine phosphopeptide encompassing a PDGF
receptor PI 3-kinase binding site, the enzyme was substantially
activated, an effect that is obviously independent of tyrosine
phosphorylation. Activation of PI 3-kinase upon binding to the Src
homology domains was much weaker. Thus, in contrast to PI 3-kinase
binding to the PDGF receptor (Shoelson et al., 1993),
allosteric activation of PI 3-kinase does not occur to any great extent
upon binding to v-Src. We therefore addressed the role of the v-Src
catalytic domain in stimulation of v-Src-associated PI 3-kinase
activity. Using a kinase-defective and three ts mutants of
v-Src, we demonstrated that v-Src tyrosine kinase activity is required
for maximal v-Src-associated PI 3-kinase activity in transformed cells
and correlates with phosphorylation of p85 on tyrosine. Since cellular
PI 3-kinase bound directly to isolated v-Src SH domains, it seems
likely that association of the small amount of PI 3-kinase activity
consistently observed with the kinase-defective v-Src occurs through
binding of active enzyme to its SH domains. These findings are
consistent with the requirement for phosphorylation of PI 3-kinase by
another Src family member, p56 Association with PI 3-kinase is likely to be of
considerable importance for the biological activity of v-Src. Here we
have refined the understanding of the regulation of v-Src-associated PI
3-kinase. We have demonstrated that all Src homology domains cooperate
to positively regulate the binding of PI 3-kinase in response to
activation of the oncoprotein. Before ts v-Src kinase activity
is switched on, PI 3-kinase is bound predominantly by the SH3 domain.
Upon activation of the v-Src tyrosine kinase, additional efficient
binding to the SH2 domain occurs in a catalytically activated form. The
rapid nature of the induced binding of activated PI 3-kinase thus
provides a very sensitive means of recruiting the lipid kinase to the
vicinity of its substrates and signaling complexes at the cell
membrane. Any role for the SH3 and SH2 domains of p85 in the binding of
PI 3-kinase to v-Src remains unknown. The activation of PI 3-kinase in
response to the v-Src oncoprotein may also involve the catalytic
subunit p110, a possibility that we have not addressed here.
Volume 270,
Number 14,
Issue of April 7, 1995 pp. 7937-7943
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
A ROLE FOR THE Src HOMOLOGY 2 DOMAIN (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
)domains, i.e. SH1, the tyrosine kinase domain;
SH2, which binds to phosphotyrosine residues; and SH3, which binds to
proline-rich sequences in proteins. SH2 and SH3 domains have been found
either alone, or in combination, in many proteins involved in signal
transduction (reviewed by Pawson and Schlessinger, 1993).
(Nakanishi et al., 1993).
Tissue Culture
Chicken embryo fibroblasts (CEF)
uniformly infected with RCAN encoding the temperature-sensitive (ts) mutant v-Src proteins LA 29, LA 29A2,
or LA 32 were cultured in Dulbecco's minimum essential
medium supplemented with 10% newborn calf serum, 1% heat-inactivated
chick serum, 10% tryptose phosphate broth, 0.375% sodium bicarbonate, 1
mM sodium pyruvate, 2 mM glutamine (all supplied by
Life Technologies, Inc.). Cells were made quiescent by serum
deprivation (0.2% newborn calf serum, no tryptose phosphate, no chick
serum) for 4 days.Preparation of Fusion Proteins
v-Src homology
domains SH3 (amino acids 85-141), SH2 (amino acids
142-247), and combined SH3 and SH2 (amino acids 85-247)
were amplified by polymerase chain reaction. BamHI and EcoRI restriction sites were generated at the ends to
facilitate cloning into pGex2t (Pharmacia Biotech Inc.). The cloning
products were checked by sequencing. Bacterially expressed glutathione S-transferase fusion proteins were isolated and bound to
glutathione-agarose beads according to the method of Smith and
Johnson(1988). Proteins were assayed using Coomassie reagent (Pierce).Binding Assays
Semiconfluent cultures of CEF were
washed twice with ice-cold phosphate-buffered saline solution and lysed
in modified PLC lysis buffer (50 mM Tris, pH 7, 1% Triton, 10%
glycerol, 1.5 mM MgCl
, 1 mM EGTA, 150
mM NaCl containing 1.8 µg/ml aprotinin, 0.1 mM sodium orthovanadate, 0.5 mM NaF, 10 mM 
-glycerophosphate, 10 mM sodium pyrophosphate, and
1.25 mM phenylmethylsulfonyl fluoride). The lysates were
clarified by centrifugation and protein concentrations determined by
the BCA method (Pierce). 300 to 500 µg of cell protein (with or
without inhibiting proline-rich peptide) were preabsorbed with 200
µg of GST for 60 min at 4 °C and the supernatant incubated with
fusion protein bound to glutathione-agarose beads for 60 min at 4
°C in a volume of 1 ml. The beads were then washed four times with
cold lysis buffer, proteins resolved in a 7.5% SDS-polyacrylamide gel,
and Western blotting analysis performed as described below.Western Blotting Analysis
Equal amounts of
clarified lysates were incubated with glutathione agarose-bound
GST-fusion proteins for 2 h. The precipitates were then resolved on 10%
SDS-polyacrylamide gels, transferred to a nitrocellulose membrane, and
probed with monoclonal anti-p85 antibody. Detection was by ECL
(Amersham Corp.) following manufacturer's instructions.
Quantitation of p85 binding to the fusion proteins was done by
densitometry using a Molecular Dynamics computing densitometer. To
examine phosphotyrosine levels in the p85 subunit of PI 3-kinase from
cells expressing v-Src mutants, lysates were precipitated with anti-p85
antibody and then treated as above. Blots were then stripped following
ECL instructions and reprobed with monoclonal anti-phosphotyrosine
antibody PY20 (Upstate Biotechnology Inc.). To deplete cell lysates of
phosphotyrosine-containing proteins, cell lysates were incubated for 6
h with excess (1:50) polyclonal rabbit anti-phosphotyrosine antibody
(Affiniti) immobilized on protein A-Sepharose beads (Sigma).Phosphatidylinositol 3-Kinase Assay
Cells were
grown to around 90% confluence at the appropriate temperature, washed,
rapidly lysed in PLC lysis buffer (50 mM Hepes, pH 7, 10%
glycerol, 1% Triton, 1.5 mM MgCl, 1 mM EGTA, 150 mM NaCl containing 1.8 µg/ml aprotinin, 0.1
mM sodium orthovanadate, 0.5 mM NaF, 10 mM -glycerophosphate, 10 mM sodium pyrophosphate, and
1.25 mM phenylmethylsulfonyl fluoride) and protein
concentrations determined by the BCA method. Equal amounts of clarified
lysates were incubated either with protein A-Sepharose beads (Sigma)
coated with anti v-Src monoclonal antibody EC10 (Upstate Biotechnology
Inc.) or control antibody (normal mouse IgG (ICN) or monoclonal
anti-p85 (Upstate Biotechnology Inc.)) for 16 h or with
glutathione-agarose-bound GST fusion proteins for 2 h. Precipitates
were washed once with lysis buffer, twice with phosphate-buffered
saline, 1% Nonidet P-40 (Sigma), twice with 100 mM Tris/HCl,
500 mM LiCl, pH 7.6, and twice with 20 mM HEPES, 1
mM EGTA, pH 7.4, and then incubated in assay buffer (50 mM HEPES, 10 mM NaCl, 1 mM EGTA, pH 7.4) with a 3:1
sonicated mixture of phosphatidylinositol and phosphatidylserine (0.02
mg/ml) and 50 µCi of [
-
P] ATP, 10
mM MgCl as substrates for 30 min at 30 °C.
Reaction products were extracted and separated by thin layer
chromatography. After autoradiography the phosphatidylinositol
3-phosphate band was scraped off the plate, counted, and presented as
counts/minute. The identity of the lipid product was verified by HPLC
analysis. The PI 3-kinase assay was linear under these conditions.HPLC Analysis of Lipid Products
Lipid products of
PI 3-kinase assays were deacylated and HPLC analysis performed as
described previously (Carter and Downes, 1992).
The Binding of p85 to Fused v-Src SH3 and SH2 Domains
Is Enhanced
Since mutations in SH2 (Fukui and Hanafusa, 1989),
as well as SH3 (Wages et al., 1992), have been shown to
influence the v-Src association with PI 3-kinase activity, we tested
the ability of isolated GST-SH3 and GST-SH2 domains of v-Src, as well
as a GST-SH3-SH2 fusion, to bind p85 from extracts of growing chicken
embryo fibroblasts. Bacterial fusion proteins coupled to
glutathione-agarose beads were incubated with extracts of growing CEF
and bound proteins immunoblotted with p85 antiserum (Fig. 1A shows a short and a longer exposure of the immunoblots). As
previously reported (Liu et al., 1993), the SH3 of v-Src
domain bound p85 (Fig. 1A, lanes 2-4).
In contrast to this previous work, we found that the SH2 domain also
bound, although much more weakly (less than 10% of the amount) than SH3 (Fig. 1, panel A, lanes 5-7 and panel B). Furthermore, the SH2 domain bound both a species
that co-migrated with the SH3-binding form of p85 and a faster
migrating species (Fig. 1, A and C). Whether
the lower species is a different gene product or an alternatively
modified form of the higher molecular weight species is not clear. When
the v-Src SH3 and SH2 domains were fused, enhanced binding
(approximately twice the quantity bound to SH3 and SH2 combined) of p85
was observed (Fig. 1, panel A, lanes8-10 and panel B).
, SPPTPKPRPPRPLPV in single-letter amino acid code) on the
binding of p85 to isolated and combined SH3 and SH2 domains (Fig. 1C). This peptide is an extended version of a
peptide previously used by Liu et al. (1993) to demonstrate
direct binding of the v-Src SH3 domain to this proline-rich region of
p85. The peptide efficiently inhibited the SH3-p85 interaction (Fig. 1C, lanes2 and 3) and
had no effect on the SH2-p85 interaction (Fig. 1C, lanes5 and 6). The peptide also inhibited
the binding of p85 to the SH3-SH2 fusion (Fig. 1C, lanes8 and 9) albeit to a lesser extent.
Inhibition of p85 binding by the proline-rich peptide, in the case of
the SH3-SH2 fusion, did not reduce p85 binding to the level observed
for the isolated SH2 domain (Fig. 1D). One possible
explanation for this is that the combined SH3 and SH2 domains bind p85
with an increased affinity, which is greater than that for the peptide. Enhanced PI 3-Kinase Activity Associated with
SH3-SH2
The PI 3-kinase activity, which can associate with the
non-catalytic domains of v-Src, was measured by incubating the GST-SH3,
GST-SH2, and GST-SH3-SH2 fusion proteins with extracts prepared from
growing CEF, followed by isolation of fusion proteins on
glutathione-agarose. The ability of 1 mM proline-rich peptide
to inhibit was also tested. This concentration was approximately 4
times that required for half-maximal inhibition of binding of PI
3-kinase activity to GST-SH3 (data not shown). We found PI 3-kinase
activity associated with the SH2 domain similar to that bound to the
SH3 domain (Fig. 2), despite the consistently lower binding of
p85 found by immunoblotting (Fig. 1A). Possible
explanations for this are that SH2-bound PI 3-kinase has a higher
specific activity or that free p85 binds preferentially to SH3, while
the catalytically active p85/p110 dimer binds both domains with similar
efficiency. In addition, the activity bound to the SH3-SH2 fusion was
considerably higher than that bound to either domain alone (Fig. 2). This is consistent with the observed increase in
binding of p85 by the GST-SH3-SH2 fusion protein (Fig. 1A). The PI 3-kinase activity bound to SH3 and,
to a lesser extent, SH3-SH2, but not isolated SH2, was inhibited by 1
mM proline-rich peptide from p85 (Fig. 2).
) and
absence (
) of 1 mM proline-rich peptide from p85. A
representative experiment from three repeats is
shown.
Binding to the v-Src SH3 and SH2 Domains Results in
Weaker Direct Activation of PI 3-Kinase than Binding to a PDGF Receptor
Phosphopeptide
We next addressed possible activation of PI
3-kinase by direct binding to the Src homology domains, and we compared
this with the activation which occurs upon binding to a
tyrosine-phosphorylated peptide corresponding to a PDGF receptor
sequence that contains the PI 3-kinase binding site. This peptide is
GGYMDMSKDESVDY*VPMLDM (in single-letter amino acid code, where *
denotes a phosphorylated residue) and corresponds to residues
738-757 of the human PDGF receptor (Kazlauskas and Cooper, 1989).
We measured PI 3-kinase activity present in extracts of growing CEF,
which bound the GST-SH fusion proteins and the biotinylated receptor
peptide bound to streptavidin-agarose. In addition, we measured the PI
3-kinase activity present in the extract before and after binding. The
ratio of ``activity bound'' to the beads to the
``activity removed'' from the extract provided a measure of
the enzyme activation that occurred as a direct result of binding.
Binding to the receptor phosphopeptide resulted in activation of PI
3-kinase, considerably greater than that observed upon binding to the
SH3, SH2, or SH3-SH2 domains (Fig. 3A). This implies
that the substantial conformational activation, which occurs upon
binding PI 3-kinase to the tyrosine phosphorylated PDGF receptor,
mediated by the SH2 domain of p85 (Shoelson et al., 1993),
does not occur to the same extent upon binding PI 3-kinase to the Src
homology domains. Consistent with the activation as a direct result of
binding, the PDGF receptor phosphopeptide-associated PI 3-kinase
exhibited greater specific activity than that bound by the v-Src SH3
domain (Fig. 3B). A comparison of bound enzyme activity
with amount of p85 associated with the agarose beads demonstrated
clearly that the receptor phosphopeptide bound considerably more PI
3-kinase activity, despite binding only about 27% of the p85 associated
with the v-Src SH3 domain (Fig. 3B). In contrast to our
observations with the v-Src homology domains, including SH3, it has
been shown recently that peptides containing the SH3 domains of Lyn and
Fyn bind to a proline-rich region of p85 and augment PI 3-kinase
activity 5-7-fold (Pleiman et al., 1994). This may
reflect genuine differences between the v-Src and Lyn/Fyn SH3 domains
or could be due to differences in the methodology used.
Importance of the v-Src Kinase for Associated PI 3-Kinase
Activity
The weaker activation of PI 3-kinase upon binding to
the Src homology domains SH3, SH2 and SH3-SH2 compared with the PDGF
receptor phosphopeptide suggests that PI 3-kinase activation in
response to v-Src is mediated differently from the allosteric mechanism
proposed for direct binding to receptor tyrosine-phosphorylated
peptides (Shoelson et al., 1993). Therefore, we examined the
PI 3-kinase activity associated with various ts mutants of
v-Src in CEF at restrictive (41 °C) and permissive (35 °C)
temperature. The mutants used were ts LA 29, which has a
temperature-sensitive kinase, ts LA 29A2, a
myristylation-defective version of ts LA 29, and ts LA 32, which has functional mutations in both the SH3 and catalytic
domains, which, in combination, produce a fusiform transformed
morphology at permissive temperature (Stoker et al., 1984,
1986; Catling et al., 1994). CEF were uniformly infected with
the replication-competent RCAN vector or RCAN-29, RCAN-29A2, or RCAN-32
as described previously (Catling et al., 1993). v-Src was
isolated from lysates of cells grown at restrictive and permissive
temperatures by immunoprecipitation using the monoclonal antibody EC10,
and associated PI 3-kinase activity determined. The PI 3-kinase
activity bound to v-Src was temperature-dependent in each case (Fig. 4A), although the degree of temperature
sensitivity varied between mutant proteins. In the case of ts mutants LA 29A2 and LA 32, v-Src-associated PI
3-kinase activity at restrictive temperature was not substantially
above the background ``vector only'' control. The higher
level of activity associated with LA 29 at restrictive
temperature has probably arisen due to its relatively higher frequency
of back mutation as cells are passaged at restrictive temperature, a
consequence of which is the expression of non-ts v-Src
proteins in some cultures. This implies that PI 3-kinase activity
associated with v-Src in transformed cells is dependent on activity of
the v-Src kinase. In order to test this directly, we expressed a
kinase-defective version of the mutant protein LA 32, LA 32KD, derived by mutation of Lys to Arg in the ATP binding site
(described by Catling et al., 1994). The PI 3-kinase activity
associated with this transformation-defective v-Src protein, although
consistently above the vector alone control, was substantially reduced (Fig. 4B). These results indicate a requirement for the
v-Src kinase for maximal associated PI 3-kinase activity.
Phosphorylation of p85 on Tyrosine by ts and
Kinase-defective v-Src
We also examined the tyrosine
phosphorylation of p85 in cells expressing ts and
kinase-defective v-Src mutant proteins by immunoprecipitating p85 from
lysates of growing CEF at restrictive and permissive temperatures,
blotting to nitrocellulose, and probing with both anti-p85 and
anti-phosphotyrosine antibodies. Although the amount of both p85 forms
immunoprecipitated was relatively invariant in the uniformly infected
cultures, tyrosine phosphorylation of the p85 proteins was
temperature-dependent in the case of cells expressing ts mutants (Fig. 4C). In cells expressing the
kinase-inactive v-Src protein (lane1), there was no
detectable tyrosine phosphorylation of p85. Thus, association of active
PI 3-kinase requires the v-Src tyrosine kinase in transformed cells and
correlates with tyrosine phosphorylation of p85. Both species of p85
detected from CEF cell extracts by immunoblotting, which co-migrated
with those bound to the SH domains (Fig. 1, A and B), are phosphorylated on tyrosine in response to v-Src.
Phosphorylation of p85 on tyrosine has also been demonstrated in 3T3
cells in response to PDGF stimulation, although the way in which this
tyrosine phosphorylation regulates its activity or interaction with
other proteins remains to be established (Kavanaugh et al.,
1992).The v-Src Tyrosine Kinase Rapidly Regulates the Binding
of PI 3-Kinase to the SH2 Domain in Quiescent Cells
In addition
to its transforming activity, v-Src can act as an intracellular
mitogen. In cells made quiescent by serum deprivation, activation of ts v-Src by temperature shift results in transition from
G
to G and on to S phase and mitosis in the
absence of exogenous growth factors (Bell et al., 1975; Durkin
and Whitfield, 1984; Welham et al., 1990; Catling et
al., 1993). The temperature-dependent association of PI 3-kinase
in cells transformed with ts v-Src implicates the lipid kinase
as a potential downstream intermediate in the transduction of v-Src
tyrosine kinase activity into its biological consequences. We therefore
examined the kinetics of association of active PI 3-kinase with ts
LA 29 v-Src in CEF that had been made quiescent at restrictive
temperature and stimulated for various times by shift to permissive
temperature. Fig. 5A demonstrates that there was a very
rapid stimulation of v-Src-associated PI 3-kinase activity upon
temperature shift. The kinetics of stimulation parallels the activation
of the LA 29 v-Src tyrosine kinase, which is detectable at 10
min and continues to rise at 30 min and 1 h after shift to permissive
temperature (shown by Catling et al., 1993). Thus, association
of PI 3-kinase activity with v-Src is a rapid consequence of activating
the v-Src kinase, suggesting that this may be an early event in the
pathways leading to v-Src-induced mitogenesis and transformation. In
addition, p85 tyrosine phosphorylation was increased upon temperature
shift over the same time course (Fig. 5B), suggesting
that this was responsible for the increase in v-Src-bound PI 3-kinase
activity.
, where the
presence of an adjacent SH2 domain facilitated binding of PI 3-kinase
from T cells to the p56 SH3 domain (Prasad et
al., 1993). In cells fully transformed by v-Src, binding of p85 to
the SH2 domain was considerably increased. Consistent with this,
altered specificity has been observed for PI 3-kinase binding to
p56, mediated by the SH3 domain in normal CEF and the SH2
domain in Rous sarcoma virus-transformed CEF (Vogel and Fujita, 1993).
The SH2 domain-associated PI 3-kinase had a higher specific activity,
possibly as a consequence of tyrosine phosphorylation of the p85
regulatory subunit. Upon stimulation of CEF by activation of the v-Src
tyrosine kinase, p85 binding to the SH2 domain was rapidly increased.
In particular, the faster migrating form of p85 from transformed cells,
which preferentially but not exclusively binds the SH2 domain,
displayed enhanced binding to both SH2 and the SH3-SH2 fusion. This
enhanced binding was reflected in an increase in the PI 3-kinase
activity, which could associate with the SH2 domain in vitro and v-Src protein in vivo, implying that p85 binding to
the non-catalytic domains of v-Src, in particular SH2, contributes to
the formation of active v-Src-PI 3-kinase complexes at the cell
periphery. The rapid increase in v-Src-associated PI 3-kinase activity
demonstrates that the increased ability of p85 to bind SH2 is an early
event that correlates with phosphorylation of p85 on tyrosine. This is
supported by the observation that depletion of cell lysates of
phosphotyrosine-containing proteins results in a considerable reduction
in SH2- but not SH3-bound p85. Furthermore, these observations suggest
that both SH3 and SH2 domains contribute to the binding of PI 3-kinase
and that binding to the SH2 domain is responsive to intracellular
signals., in Jurkat T cells (von
Willebrand et al., 1994). Thus, in addition to the differences
in mode of binding and allosteric activation, v-Src-bound and PDGF
receptor-associated PI 3kinase also differ in their tyrosine kinase
requirement.
)
We thank Andy Catling and Nigel Carter for their input
to this work in the early stages, John Wyke and Philip Cohen for
critical reading of the manuscript, and Sam Crouch for help with
photography.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. K. Bentley, D. C. Newcomb, A. M. Goldsmith, Y. Jia, U. S. Sajjan, and M. B. Hershenson Rhinovirus Activates Interleukin-8 Expression via a Src/p110{beta} Phosphatidylinositol 3-Kinase/Akt Pathway in Human Airway Epithelial Cells J. Virol., February 1, 2007; 81(3): 1186 - 1194. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. C. Newcomb, U. Sajjan, S. Nanua, Y. Jia, A. M. Goldsmith, J. K. Bentley, and M. B. Hershenson Phosphatidylinositol 3-Kinase Is Required for Rhinovirus-induced Airway Epithelial Cell Interleukin-8 Expression J. Biol. Chem., November 4, 2005; 280(44): 36952 - 36961. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Platek, M. Mettlen, I. Camby, R. Kiss, M. Amyere, and P. J. Courtoy v-Src accelerates spontaneous motility via phosphoinositide 3-kinase, phospholipase C and phospholipase D, but abrogates chemotaxis in Rat-1 and MDCK cells J. Cell Sci., September 15, 2004; 117(20): 4849 - 4861. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kagawa, L. Varticovski, Y. Sai, and I. M. Arias Mechanism by which cAMP activates PI3-kinase and increases bile acid secretion in WIF-B9 cells Am J Physiol Cell Physiol, December 1, 2002; 283(6): C1655 - C1666. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. J. Fincham, V. G. Brunton, and M. C. Frame The SH3 Domain Directs Acto-Myosin-Dependent Targeting of v-Src to Focal Adhesions via Phosphatidylinositol 3-Kinase Mol. Cell. Biol., September 1, 2000; 20(17): 6518 - 6536. [Abstract] [Full Text]
|
|
|
|
|
|
R. J. Lee, C. Albanese, R. J. Stenger, G. Watanabe, G. Inghirami, G. K. Haines III, M. Webster, W. J. Muller, J. S. Brugge, R. J. Davis, et al. pp60v-src Induction of Cyclin D1 Requires Collaborative Interactions between the Extracellular Signal-regulated Kinase, p38, and Jun Kinase Pathways. A ROLE FOR cAMP RESPONSE ELEMENT-BINDING PROTEIN AND ACTIVATING TRANSCRIPTION FACTOR-2 IN pp60v-src SIGNALING IN BREAST CANCER CELLS J. Biol. Chem., March 12, 1999; 274(11): 7341 - 7350. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Y. Kanemitsu, L. W. M. Loo, S. Simon, A. F. Lau, and W. Eckhart Tyrosine Phosphorylation of Connexin 43 by v-Src Is Mediated by SH2 and SH3 Domain Interactions J. Biol. Chem., September 5, 1997; 272(36): 22824 - 22831. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |