(Received for publication, March 11, 1997, and in revised form, June 24, 1997)
From the Department of Experimental Pathology, United Medical and Dental School, Guy's Hospital, London SE1 9RT, United Kingdom
The activation of the mitogen-activated protein kinase (MAPK) cascade by a variety of growth factors and other agents is central to a mitogenic response. In the case of polypeptide growth factors such as the epidermal growth factor (EGF) and platelet-derived growth factor (PDGF), the steps leading to activation of MAPK require the function of the adaptor protein Grb2 (growth factor receptor binding protein 2), which can bind either directly or indirectly via its Src homology 2 domain to activated receptor tyrosine kinases. A cell-permeable mimetic of the EGF receptor Grb2 binding site has been investigated for its ability to inhibit biological responses stimulated by a variety of growth factors. Pretreatment of cells with this peptide results in the accumulation of the peptide in cells and its association with Grb2. This is associated with a complete inhibition of the mitogenic response stimulated by EGF and PDGF. In contrast, the peptide has no effect on the mitogenic response stimulated by fibroblast growth factor. The peptide could also inhibit the phosphorylation of MAPK stimulated with EGF and PDGF in the absence of an effect on the fibroblast growth factor response. These data demonstrate that cell-permeable mimetics of Src homology 2 binding sites can selectively inhibit growth factor-stimulated mitogenesis, and also directly demonstrate specificity in the coupling of activated receptor tyrosine kinases to the MAPK cascade.
Growth factors such as EGF,1 PDGF, and FGF stimulate cell proliferation by binding to and activating membrane-spanning receptors that have cytoplasmic tyrosine kinase domains (1, 2). Ligand binding induces receptor dimerization, and this is associated with autophosphorylation on tyrosine residues both within and outside the catalytic domain in the receptor dimer (e.g. see Refs. 3 and 4). Whereas the former are required for catalytic activity, the latter can serve as high affinity binding sites for effector molecules or adaptor molecules that recruit effectors to the receptor. Grb2 is a key adaptor molecule that couples some receptor tyrosine kinases to the Ras/mitogen-activated protein kinase (MAPK) cascade, and this is central to the mitogenic response stimulated by many growth factors (5). Grb2 binds to defined motifs, containing phosphorylated tyrosine (designated Y in the single-letter amino acid code), via its SH2 domain, and these motifs are found in a wide range of molecules including the EGF and PDGF receptors and the Shc (Src homology and collagen protein) adaptor (6-8). In this context, when phosphorylated, Tyr1068 in the EGF receptor acts as a docking site for the SH2 domain of Grb2 and synthetic phosphopeptides corresponding to this site inhibit Grb2 binding to the activated EGF receptor in vitro (8). These peptides can also inhibit Grb2 function in intact cells, e.g. a 9-amino acid phosphopeptide corresponding to this site can inhibit the ability of EGF to activate Ras in permeabilized cells (9). Moreover, when made in tandem with a peptide that allows for translocation across the cell membrane, this phosphopeptide retains its ability to bind to Grb2 and can inhibit Grb2 binding to the activated EGF receptor in live cells (10). This was associated with a competitive inhibition of EGF stimulation of the MAPK cascade, with the peptide being able to partially (~60%) inhibit the effects of EGF at 10 ng/ml but being ineffective against EGF at 50 ng/ml. In both systems phosphorylation of Tyr1068 was required for inhibitory activity.
We have recently shown that a cell-permeable peptide mimetic of the
PLC binding site on the FGF receptor (the FGFR766Y-P
peptide) can inhibit phosphatidylinositol hydrolysis stimulated by FGF
in intact cells in the absence of any effect on the same response
stimulated by PDGF or neurotrophin-3 (11). This level of specificity
was unexpected and might be related to lower levels of FGF receptors as
compared with the other receptors and/or differences in the affinity of
PLC
binding to the various receptors. Moreover, we were able to show
that this peptide could inhibit the neurite outgrowth response
stimulated by FGF, but had no effect on neurite outgrowth stimulated by
agents that act downstream from PLC
activation in the FGF receptor
signal transduction cascade, or other growth factors such as nerve
growth factor (11, 12). Again, phosphorylation of Tyr766
was required for inhibitory activity. Thus, cell-permeable
phosphopeptides that mimic receptor binding sites for specific SH2
domain-containing proteins have the potential to be developed as tools
that will help us understand how growth factor receptors couple to
signaling cascades and biological responses.
In the present study we report on the effects of cell-permeable
mimetics of the EGF receptor Grb2 binding site and FGF receptor PLC
binding site for their ability to inhibit mitogenic responses stimulated by EGF, PDGF, and FGF. All three responses were inhibited by
PD 098059, which inhibits MAPK kinase (13), and this is consistent with
activation of the MAPK cascade being required for these responses. We
find that the EGF receptor Grb2 binding site mimetic was able to bind
to Grb2 in cells and inhibit the mitogenic response stimulated by EGF
and PDGF, but had no effect on the response stimulated by FGF. This
peptide could also inhibit the phosphorylation of MAPK stimulated by
EGF and PDGF, but not FGF. In contrast, the distinct phosphopeptide
corresponding to the FGF receptor PLC
binding site had no effect on
the mitogenic response stimulated by EGF, PDGF, or FGF. These data
demonstrate for the first time that cell-permeable peptide mimetics of
SH2 domain binding sites can selectivity inhibit activation of the MAPK
cascade and mitogenic responses stimulated by polypeptide growth
factors.
Peptides were synthesized on a 431A
Applied Biosystem peptide synthesizer using
p-hydroxymethylphenoxymethyl polystyrene resin and standard
Fmoc chemistry. The EGFR Grb2-binding peptide sequence was as follows,
Biot-Ahx-Arg-Arg-Trp-Arg-Arg-Trp-Trp-Arg-Arg-Trp-Trp-Arg-Arg-Trp-Arg-Arg-Pro-Val-Pro-GluTyr(P)-Ile-Asn-Gln-Ser, where the underlined sequence corresponds to the 9-amino acid sequence containing phosphotyrosine 1068 of the EGF receptor that has
been established to bind to the SH2 domain of Grb2 (the full peptide is
called the EGFR1068Y-P peptide throughout the study). A
phosphorylated Fmoc tyrosine (Novabiochem) was used in this sequence.
The rest of the peptide sequence corresponds to a functional analogue
of a sequence derived from the third helix of the Antennapedia protein,
which has been established to act as a vector that can carry other
peptides across the cell membrane into the cytosol (for full details of
internalization properties of Antennapedia-based peptides, see Refs. 11
and 14). This peptide also contained an aminohexanoic acid (Ahx) spacer
(Novabiochem) and a biotin at the N terminus. Synthesis of the
cell-permeable Antennapedia-FGFR PLC-binding peptide (called the
FGFR766Y-P peptide throughout the study) was as described
previously with the exception that this peptide also contained an
N-terminal biotin (11). After synthesis and deprotection, the peptides were desalted on a Sephadex G-10 column and lyophilized. The peptides were analyzed for purity by analytical high pressure liquid
chromatography, and in each case a single peak was observed. Peptides
were made as stock solutions at 25 mg/ml in dimethyl sulfoxide
(Me2SO) and stored at
20 °C before use.
All surface plasmon resonance
experiments were carried out using a BIAcore XTM biosensor (Pharmacia
Ltd., Biacore, Uppsala, Sweden), the principle of which has been
described elsewhere (15). Parallel flow cells of CM 5 sensor chips were
derivatized, using the amine coupling method, with streptavidin (200 µg/ml) in 10 mM sodium acetate, pH 4.0, according to the
manufacturer's protocol. Approximately 2100-2600 resonance units of
ligand were immobilized, corresponding to a concentration of 2.1-2.6
ng/mm2. Biotinylated peptides were coupled via streptavidin
to one flow cell only (the other remaining as a control) by injection
at 1 µg/ml in running buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 3.4 mM EDTA, 0.005% P20) at a flow
rate of 5 µl/min and at 25 °C. Five short (4 µl) pulses of 0.1%
sodium dodecyl sulfate (SDS) were carried out to remove any
nonspecifically bound peptide and also to test regeneration conditions.
Glutathione S-transferase (GST) fusion proteins were
subsequently injected simultaneously over both flow cells at 100 nM in running buffer at a flow rate of 5 µl/min. After a
period of free buffer flow, any analyte remaining bound to the surface
was removed with a pulse of 0.1% SDS bringing the signal back to
baseline. Specific binding of the GST fusion proteins to each peptide
was determined automatically by the system by subtraction of test from
control flow cell responses, measured in arbitrary units called
resonance units. The GST fusion proteins used in the present study
included the full sequence and SH2 domain of Grb2, the SH2 domain of
PLC, and the SH2 domain of the p85 subunit of PI 3-kinase (these
were a kind gift of Dr. Pam Ganju, Sandoz Institute, UCL, London).
Newt A1 myoblast cells were grown in gelatin-coated plastic flasks in medium composed of 60% minimum essential medium with Earle's salts (ICN), 27% distilled water (AMEM), 10% fetal calf serum (FCS) supplemented with insulin, glutamine, and antibiotics as described previously (16). Cells were maintained at 25 °C in a humidified atmosphere of 2% CO2. Mammalian C2 myoblast cells (17) were cultured in Dulbecco's modified Eagle's medium (DMEM, Life Technologies, Inc.) supplemented with 15% FCS and glutamine at 37 °C in 8% CO2. Mitogenic assays were set up by seeding either 2000 A1 cells/well of 48-well plates (Nunc), previously coated with 0.75% gelatin, or 4000 C2 cells/well of 48-well plates. Cells were initially serum-starved by culture in serum-free medium for 3 days. Serum-free medium was then replaced with medium containing 0.5% FCS alone or further supplemented with basic FGF (bFGF, 5 ng/ml), EGF (10 ng/ml), or PDGF (10 ng/ml), (all from Collaborative), in the presence or absence of the EGFR1068Y-P peptide (10 µg/ml) or the MAPK kinase inhibitor PD 098059 (100 µM; Calbiochem). In some experiments cultures were treated with the EGFR1068Y-P peptide for 2 h before being washed three times and the medium replaced with growth factors in the absence of additional peptide (see "Results and Discussion"). After 24 h, cultures were labeled for another 24 h with 1 mg/ml 5-bromo-2-deoxyuridine/5-fluoro-2-deoxyuridine (BrdUrd) using the Amersham proliferation kit according to instructions. Cells were fixed for 5 min with ice cold methanol and stained for BrdUrd as described previously (18). Briefly, following fixation cells were treated for 10 min with 1 M HCl followed by 5 min with 100 mM Tris (pH 7.8), then blocked for 10 min with phosphate-buffered saline (PBS) containing 10% goat serum (GS); the primary and secondary antibodies were mouse anti-BrdUrd and anti-mouse fluorescein isothiocyanate (FITC), respectively (diluted in PBS/10% GS). Cell nuclei were counterstained with 1 mg/ml Hoechst dye and visualized with UV epifluorescent optics.
MAPK PhosphorylationA1 cells were grown to confluence on gelatin-coated 6-cm plates (Nunc) and then serum-starved for 2 days. Medium was then replaced with 0.5% FCS/AMEM alone or supplemented with either the PD 098059 compound or the EGFR1068Y-P peptide and individual growth factors (see text for treatment schedules and concentrations). The cultures were then washed with ice-cold PBS before being lysed and collected in 200 µl of sample buffer (62.5 mM Tris-HCl at pH 6.8, 2% SDS, 10% glycerol, 50 mM dithiothreitol, 0.1% (w/v) bromphenol blue), again at the indicated time points. Samples were boiled for 5 min and sonicated for 30 s before being resolved on 12.5% SDS-polyacrylamide gels and Western blotted onto polyvinylidene difluoride membranes (Millipore). Membranes were blocked for 1 h at room temperature with gentle agitation in blocking buffer (5% nonfat dried milk, 0.1% Tween 20 in PBS), and then incubated for 24 h at 4 °C with a polyclonal rabbit anti-active MAPK antibody (1:20,000 from Promega) in PBS (containing 5% BSA, 0.05% Tween 20). Membranes were then washed three times with blocking buffer, incubated with HRP-conjugated donkey anti-rabbit secondary antibody (1:4000 in blocking buffer; Amersham) for 1 h at room temperature, and finally washed three times with blocking buffer before probing with enhanced chemiluminescence (ECL; Amersham).
Peptide Uptake into CellsCultures of A1 cells were
established overnight on gelatin-coated wells of an eight-chamber
tissue culture slide. The cultures were washed and media replaced with
AMEM containing 10 µg/ml biotinylated EGFR1068Y-P
peptide. After 2 h the cultures were washed three times with AMEM
and then fixed and permeabilized with ethanol:acetic acid (9:1 by
volume) for 5 min at 20 °C. Nonspecific protein binding sites were
blocked for 30 min with 10% GS in PBS. The cultures were then
incubated for 1 h with FITC-conjugated streptavidin (1:500;
Amersham). The cultures were washed another three times with PBS before
being mounted for viewing by confocal microscopy.
C2 myoblast cells were grown to 70% confluence on collagen-coated plates, serum-starved in DMEM supplemented with L-glutamine (2 mM), insulin (10 µg/ml), transferrin (5 µg/ml), and sodium selenite (10 nM) for 4 h and then treated with or without peptide (50 µg/ml) for 1 h at 37 °C. The cell monolayers were washed extensively with ice cold PBS (three times), 2 M NaCl at pH 7.4 (three times), and 2 M NaCl/20 mM sodium acetate at pH 4 (two times) to remove extracellular peptide before being lysed in 20 mM Tris-HCl (pH 7.5) containing 1 M NaCl, 1% CHAPS, and complete protease inhibitors from Boehringer Mannheim (according to manufacturer's instructions). Lysates were clarified by centrifugation and the biotinylated peptides precipitated from the supernatant with streptavidin-agarose following gentle mixing at 4 °C for 2 h. Precipitates were washed three times with ice-cold Tris-HCl (20 mM, pH 7.5) containing 1 mM CaCl2. Sample buffer (see above) was added to both the precipitates, and depleted lysates and samples were boiled for 5 min, separated on 10% SDS-polyacrylamide gels, and then transferred to nitrocellulose membranes. Membranes were blocked for 1 h in blocking buffer (4% nonfat dried milk, 0.05% Tween 20 in PBS) and then incubated for 1 h with anti-Grb2 antibody diluted 1:1000 in blocking buffer (Transduction Laboratories). Membranes were washed three times in wash buffer (0.05% Tween 20 in PBS) before being incubated with HRP-conjugated horse anti-mouse secondary antibody (Vector) diluted 1:3000 in blocking buffer for 1 h, prior to another five washes in wash buffer. Membranes were then probed with ECL (Amersham).
We have synthesized a biotinylated peptide consisting of a
functional analogue of the 16-amino acid sequence derived from the
third helix of the Antennapedia protein that acts as an internalization vector for other peptide sequences (see Ref. 14 for a detailed description of the internalization properties of Antennapedia-related peptides and Ref. 11 for an example) in tandem with the well characterized Grb2 docking site on the activated EGF receptor (EGFR1068Y-P peptide; see "Experimental Procedures" for
details). Fig. 1 shows characterization
of the binding properties of this peptide using BIAcore technology. The
peptide readily bound a GST fusion protein containing the full Grb2
molecule and also one containing only the single SH2 domain. The
peptide did not bind a control GST fusion protein or a GST fusion
protein containing the N-terminal p85 SH2 domain, and very poorly bound
a GST fusion protein containing the N-terminal SH2 domain of PLC
(Fig. 1A). In contrast, a peptide mimetic of the FGF
receptor PLC
binding site (the FGFR766Y-P peptide)
specifically bound PLC
, and showed only background binding of Grb2
and p85 (Fig. 1B). In addition, a range of control phosphopeptides that do not share the consensus sequence for PLC
or
Grb2 binding bound neither molecule (data not shown). Thus the peptides
show the predicted specificity in terms of SH2 domain interactions (see
Ref. 6 for details) despite the presence of the internalization
sequence.
The uptake of the biotinylated EGFR1068Y-P peptide into
cells was monitored by confocal microscopy. Cells were treated with the biotinylated peptide at 10 µg/ml for 60 min before being washed, fixed, and permeabilized (see "Experimental Procedures"). Optical sections through the cell confirmed the presence of the peptide within
the cytoplasm (Fig. 2, left
panel). Similar results were obtained with the other
phosphopeptides (e.g. see Ref. 11). To determine if the
EGFR1068Y-P peptide could complex with Grb2 inside cells,
cultures of C2 myoblasts were treated with the peptide at 50 µg/ml
for 60 min before being extensively washed to remove noninternalized peptide. The cells were then lysed and the biotinylated peptide precipitated with streptavidin-agarose. Grb2 co-precipitated with the
EGFR1068Y-P peptide and the peptide essentially depleted
all of the Grb2 from the cell lysate (Fig. 2, right panel).
Other phosphopeptides, including the FGFR766Y-P peptide,
failed to precipitate Grb2 or deplete it from the cell lysates (Fig. 2,
right panel, and data not shown).
EGF, PDGF, and FGF all stimulate proliferation of newt A1 myoblasts as
indexed by BrdUrd incorporation. It was established that a maximal
response was elicited by 10, 10, and 5 ng/ml EGF, PDGF, and FGF,
respectively, and that these responses could be fully inhibited by the
MAPK kinase inhibitor PD 098059 (Fig. 3). At the lower concentrations of growth factors (1 ng/ml), the peak activation of MAPK was seen at 30 min, and this was inhibited by
treatment with PD 098059 (Fig. 4,
top). However, at the higher concentrations of growth
factors (5-10 ng/ml) activation of the MAPK cascade was apparent even
after 4 h of treatment with the growth factors (shown for EGF and
FGF in Fig. 4 (bottom), and EGF in Fig. 6 (top)).
Addition of PD 098059 2 h after the addition of the growth factors
brought the MAPK activity down to control levels (Fig. 4,
bottom) and also fully inhibited the mitogenic response
(data not shown), suggesting that this sustained activation of the MAPK
cascade is required for the full mitogenic response.
The effect of the EGFR1068Y-P peptide on the mitogenic
activity of EGF, PDGF, and FGF on A1 cells was also determined. Results pooled from three independent experiments showed that the continued presence of 10 µg/ml peptide over the 2-day assay period completely inhibited the mitogenic response stimulated by EGF and PDGF, but had no
effect on the response to FGF (Fig.
5A). The peptide also had no
obvious effect on the viability or morphology of control cells (data
not shown), suggesting that its effects on mitogenesis are the result
of a specific rather than toxic action, and this is in accord with the
failure of the peptide to affect the FGF response. The uptake studies
show that the peptide will accumulate in cells following a 60-min
treatment period (Fig. 2) and in this respect we find that a 2-h
pretreatment of cells with the peptide is sufficient to fully inhibit
the mitogenic responses stimulated by EGF and PDGF (Fig.
5B). Pretreatment of C2 myoblasts with 50 µg/ml peptide
was also associated with a full inhibition of the mitogenic response
stimulated by EGF and PDGF in the absence of any effect on the response
stimulated by FGF (data not shown). This result also suggests that the
peptide has a relatively long effective half-life in cells, and this
might reflect the slow dissociation of the peptide from Grb2, which
would be expected to protect it from dephosphorylation and degradation;
it also confirms that the peptide is effective in both mammalian and
newt cells.
The Grb2-binding peptide was initially tested for its ability to inhibit the activation of MAPK stimulated by EGF. At the maximally active concentration of EGF (10 ng/ml) 10 µg/ml peptide substantially inhibited the activation of the MAPK cascade, and this was seen at all time points tested (Fig. 6, top). The competitive nature of the inhibition was demonstrated by showing that 100 µg/ml peptide could fully inhibit the activation of MAPK stimulated by 1 ng/ml EGF (measured at 30 min; see Fig. 6, bottom). Under the same conditions the peptide also fully inhibited the activation of MAPK by PDGF, but had no effect of the response stimulated by FGF (Fig. 6, bottom).
The specificity of action of the EGFR1068Y-P peptide was
confirmed by showing that the FGFR766Y-P peptide, which
inhibits FGF-stimulated inositol phosphate accumulation and axonal
growth responses (11), had no effect on the FGF, EGF, or PDGF mitogenic
response (Table I). In addition, we have
tested three other cell-permeable phosphopeptides that do not bind to
the SH2 domain of Grb2 as measured by BIAcore, and found that they also
have no effect on the mitogenic response stimulated by EGF and PDGF
(data not shown). Thus the inhibitory activity of the
EGFR1068Y-P is very specific and can most probably be
accounted for by its ability to bind to the SH2 domain of Grb2 and
inhibit activation of the MAPK cascade. This is best explained by the
known ability of peptide mimetics of the Grb2 binding site on the EGF
receptor to compete out binding of Grb2 to the activated receptor
in vitro (8), as well as in live cells (10). Although we
have demonstrated that our peptide will bind to Grb2 in intact cells,
and that it does not interact with PLC, p85 (this study), or Shc
(data not shown), we cannot exclude the possibility that the peptide
interacts with other targets. However, it is clear from the results
that its effects are highly specific and that they are best explained by the peptide's ability to bind to the SH2 domain of Grb2.
|
The fact that the EGF1068Y-P peptide can inhibit both EGF and PDGF responses is perhaps not surprising, as the SH2 domain of Grb2 is required for both the direct and indirect interactions of this adaptor with the EGF receptor (see Introduction) and an indirect association via Syp with the activated PDGF receptor (19, 20). The possibility that the failure of the peptide to inhibit the FGF response reflects the use of a completely different mitogenic pathway is unlikely for a number of reasons. For example, FGF activated the MAPK cascade, and both this and the FGF-stimulated mitogenic response were inhibited by the PD 098059 MAPK kinase inhibitor. The alternative possibility, that the selectivity of the Grb2-binding peptide reflects differences in the coupling of the EGF/PDGF receptors and the FGF receptor to the MAPK cascade, is a more probable and interesting explanation. An emerging view is that a novel ~90-kDa protein might couple the FGF receptor to the Ras-MAPK pathway via Grb2 (21-24), and in this context a new lipid anchored FGF receptor substrate, called FRS2, has recently been cloned; following tyrosine phosphorylation, FRS2 can bind Grb2 and link the FGF receptor to MAPK activation (25). Shc can also act as an adaptor molecule that can link Grb2 to the activated FGF receptor (21, 25). Thus, the inability of the EGFR1068Y-P peptide to inhibit the FGF response might be accounted for by its inability to inhibit Grb2 binding to FRS2 and/or Shc. In fact, it has been reported previously that the EGFR1068Y-P peptide is relatively ineffective at inhibiting the association of Grb2 with Shc (10).
A number of control peptides that do not bind to the Grb2 SH2 domain
had no effect on the mitogenic response stimulated by EGF and PDGF. The
most conspicuous of these is a cell-permeable peptide mimetic of the
PLC binding site on the FGF receptor that inhibits phospholipid
hydrolysis and axonal growth responses stimulated by FGF (11). It is
interesting that this peptide had no effect on the mitogenic response
stimulated by FGF, and this is in accord with a large body of work that
has reported that whereas mutation of Tyr766 can prevent
the FGF receptor from activating PLC
, it has no effect on the
mitogenic response stimulated by this factor (4).
In summary, we have shown that cell-permeable peptides that have unique binding profiles for SH2 domain-containing molecules can be used to inhibit the mitogenic activity of polypeptide growth factors. This confirms our prediction (11) that the ability of cell-permeable phosphopeptides to interact with high affinity and relative specificity to the SH2 domains of adaptors and effectors can be readily exploited to design novel tools for the dissection of signaling pathways in living cells.
The functional analogue of the internalization sequence derived from the third helix of the Antennapedia protein was developed by Drs. Alain Prochiantz and Gerard Chassaing and their colleagues, and we thank them for sharing this sequence with us.