(Received for publication, July 25, 1996, and in revised form, October 2, 1996)
From the Laboratory of Cell Signaling, the
§ Laboratory of Cell Biology, and the
Laboratory of
Biochemistry, NHLBI, National Institutes of Health,
Bethesda, Maryland 20892
Recent evidence indicates that reactive oxygen
species (ROS) may function as intracellular messengers in receptor
signaling pathways. The possible role of ROS in epidermal growth factor (EGF) signaling was therefore investigated. Stimulation of A431 human
epidermoid carcinoma cells with EGF resulted in a transient increase in
the intracellular concentration of ROS, measured with the
oxidation-sensitive fluorescent probe 2,7
-dichlorofluorescin diacetate and laser-scanning confocal microscopy. The predominant ROS
produced appeared to be H2O2, because the
EGF-induced increase in fluorescence was completely abolished by
incorporation of catalase into the cells by electroporation. The
elimination of H2O2 by catalase also inhibited
the EGF-induced tyrosine phosphorylation of various cellular proteins
including the EGF receptor and phospholipase C-
1. The dependence of
H2O2 production on the intrinsic tyrosine kinase activity of the EGF receptor and the autophosphorylation sites
located in its COOH-terminal tail was investigated. EGF failed to
induce H2O2 generation in cells expressing a
kinase-inactive EGF receptor. However, normal
H2O2 generation was observed in cells
expressing a mutant receptor from which the 126 COOH-terminal amino
acids had been deleted to remove four (out of the total of five)
autophosphorylation sites. These results suggest that EGF-induced
H2O2 formation requires the kinase activity but
probably not the autophosphorylation sites of the EGF receptor and that inhibition of protein tyrosine phosphatase activity by
H2O2 may be required for EGF-induced protein
tyrosine phosphorylation to be manifested.
Reactive oxygen species (ROS)1 are
generally considered cytotoxic, because of the oxidative damage they
can cause to cellular components. However, at low concentrations, ROS
may function as physiological mediators of cellular responses (1). For
example, hydrogen peroxide (H2O2) mimics the
stimulatory effects of insulin on glucose transport and lipid synthesis
in adipocytes (2, 3). Recently, the production of ROS has been detected
in a variety of cells stimulated with cytokines such as transforming growth factor-1 (4, 5), interleukin-1 (6), and tumor necrosis
factor-
(6, 7), with peptide growth factors such as platelet-derived
growth factor (PDGF) (8, 9) and basic fibroblast growth factor (7, 9),
with agonists of receptors with seven transmembrane spans such as
angiotensin II (10) and lysophosphatidic acid (11) or with phorbol
ester (12).
The term ROS encompasses many species including singlet oxygen, the
superoxide anion radical (O2), H2O2,
lipid peroxides, nitric oxide, peroxynitrite (ONOO
), the
thiyl peroxyl radical (RSOO·), the ferryl radical
(FeO2+) and the hydroxyl radical (OH·) (13, 14, 15, 16).
However, the chemical nature of ROS generated in response to the
activation of various receptors has not been well characterized.
H2O2 was shown to be a major component of ROS
in cells activated by transforming growth factor-
1 or PDGF (4, 8).
The generation of ROS in response to various external stimuli has been
related to the activation of transcription factors such as NF-
B (17)
and AP-1 (7, 18), mitogen-activated protein (MAP) kinases (8, 11), and
phospholipase A2 (19) to the triggering of apoptosis (20),
and to the inhibition of protein tyrosine phosphatases (PTPases) (21,
22). H2O2 is a small, diffusible, and
ubiquitous molecule that can be synthesized, as well as destroyed,
rapidly in response to external stimuli. As such it fulfills the
important prerequisites for an intracellular messengers. We have now
investigated the role of ROS in epidermal growth factor (EGF) signal
transduction by the EGF receptor (EGFR) protein.
Bovine catalase was obtained from Boehringer
Mannheim; Dulbecco's modified Eagle's medium (DMEM), modified
Eagle's medium without phenol red, fetal bovine serum (FBS),
penicillin, and streptomycin were from Life Technologies, Inc.;
enhanced chemiluminescence (ECL) reagents were from Amersham Corp.;
antibodies to phosphotyrosine and the EGFR were from Upstate
Biotechnology; antibodies to catalase and -tubulin were from
Calbiochem and Oncogene Science, respectively; protein A-Sepharose
beads were from Pharmacia Biotech Inc.; and 2
,7
-dichlorofluorescin
diacetate (DCFH-DA) was from Molecular Probes. A mixture of monoclonal
antibodies that was used for immunoprecipitation of phospholipase
C-
1 (PLC-
1) was prepared as described (23).
Human A431 epidermoid carcinoma cells were maintained at 37 °C under an atmosphere of 5% CO2 in 150-mm dishes containing DMEM supplemented with 10% FBS. At 80-90% confluency, cells were deprived of serum for 24 h and then harvested by trypsin treatment. Harvested cells were suspended in serum-free DMEM at a density of ~1 × 107 cells/ml, and 100-µl portions of the cell suspension were placed in an electroporation chamber in the absence or the presence of catalase (8 mg/ml). Electroporation was performed by subjecting cells to six pulses, at intervals of 1 or 2 s, at a field strength of 850 ± 50 V/cm and a single-pulse width of 250 µs. Cell viability, assessed by trypan blue exclusion, was typically 70-80% after electroporation under these conditions. The construction, operation, and efficiency of the electroporation apparatus have been described previously (24). The electroporated cells were transferred to DMEM supplemented with 1% FBS and the same concentration of catalase as was present in the electroporation chamber and were placed in an incubator for 18 h.
For analysis of EGF-induced tyrosine phosphorylation, cells were stimulated with EGF (500 ng/ml) for 10 min and then exposed to lysis buffer (20 mM Hepes-NaOH (pH 7.2), 1% Triton X-100, 10% glycerol, 50 mM NaF, 1 mM Na3VO4, leupeptin (5 µg/ml), aprotinin (5 µg/ml), and 1 mM phenylmethylsulfonyl fluoride). The lysates were incubated on ice for 5 min and then centrifuged for 5 min at 10,000 × g. Protein concentration of the supernatant was measured with the Bio-Rad detergent-compatible assay kit and bovine serum albumin as standard.
Immunoprecipitation and Immunoblot AnalysisThe lysate
supernatants were incubated with monoclonal antibodies to PLC-1 or
to phosphotyrosine for 8 h, after which protein A-Sepharose beads
were added, and the incubation was continued for an additional hour.
The beads were pelleted at 10,000 × g for 5 min,
washed three times with ice-cold lysis buffer, and subjected to
SDS-polyacrylamide gel electrophoresis on an 8% gel. The separated
proteins were transferred to a nitrocellulose membrane and probed with
antibodies to PLC-
1, to phosphotyrosine, or to EGFR. Immune
complexes were detected with appropriate secondary antibodies and ECL
reagents.
Intracellular ROS production was
measured by the method of Bass et al. (25) as modified for
confocal microscopy by Ohba et al. (4). Briefly, dishes of
confluent cells at various times after stimulation with EGF were washed
with modified Eagle's medium without phenol red and incubated in the
dark for 5 min in Krebs-Ringer solution containing 5 mM
DCFH-DA. DCFH-DA is a nonpolar compound that readily diffuses into
cells, where it is hydrolyzed to the nonfluorescent polar derivative
DCFH and thereby trapped within the cells (25). In the presence of a
proper oxidant, DCFH is oxidized to the highly fluorescent
2,7
-dichlorofluorescein (DCF). Culture dishes were transferred to a
Zeiss Axiovert 135 inverted microscope, equipped with a ×20 Neofluor
objective and Zeiss LSM 410 confocal attachment, and ROS generation was
detected as a result of the oxidation of DCFH (excitation, 488 nm;
emission, 515-540 nm). The effects of DCFH photo-oxidation was
minimized by collecting the fluorescent image with a single rapid scan
(line average, 4; total scan time, 4.33 s) and identical
parameters, such as contrast and brightness, for all samples. The cells
were then imaged by differential interference contrast microscopy. Five
groups of 20-30 cells each were randomly selected from the image in
the digital interserence contrast (DIC) channel for each sample, the
fluorescence intensity was then measured for each group from the
fluorescence image, and the relative fluorescence intensity was taken
as the average of the five values. Therefore, the relative fluorescence
intensity (given in arbitrary units) reflects measurements performed on
a minimum of 100 cells for each sample. All experiments were repeated
at least three times.
Intracellular generation of ROS in A431 cells was measured with
DCFH-DA and laser-scanning confocal microscopy. Exposure of quiescent
A431 cells to EGF (500 ng/ml) resulted in a rapid increase in DCF
fluorescence, with the maximal, 2-fold increase apparent 5 min after
stimulation (Fig. 1); fluorescence had returned to the
baseline value after 20 min.
Introduction of catalase, an enzyme that specifically catalyzes the
dismutation of H2O2 to O2 and
H2O into A431 cells by electroporation, prevented
EGF-induced DCFH oxidation (Fig. 2), suggesting that the
latter is mainly mediated by H2O2. The amount
of catalase incorporated into cells was about five times that of the
endogenous enzyme (see below). The introduction of catalase also
reduced DCFH oxidation in cells not exposed to EGF.
We next assessed the effect of incorporated catalase on EGF-induced
tyrosine phosphorylation. EGF induced a rapid increase in tyrosine
phosphorylation of several proteins in control cells as revealed by
immunoblot analysis of cell lysates with antibodies to phosphotyrosine
(Fig. 3A). Increased tyrosine phosphorylation of a broad band with an apparent molecular mass of ~160 kDa that was
recognized by antibody to the EGFR was particularly prominent. However,
in cells containing exogenous catalase, EGF had no apparent effect on
tyrosine phosphorylation of the EGFR or other proteins. Phosphorylation
of PLC-1, a well characterized target of the EGFR kinase, was
studied further. Immunoblot analysis, with antibodies to PLC-
1, of
PLC-
1 immunoprecipitated from control cells treated with EGF
revealed a broad immunoreactive band (Fig. 3B); the
increased breadth of the band relative to that apparent with
unstimulated cell is indicative of increased tyrosine phosphorylation
(26). The effect of catalase was investigated by subjecting PLC-
1
immunoprecipitated with antibodies to PLC-
1 to immunoblot analysis
with antibody to phosphotyrosine (Fig. 3C). Tyrosine
phosphorylation of PLC-
1 was apparent from control EGF-treated cells
but not with that from EGF-treated cells containing exogenous catalase.
As demonstrated previously (26, 27), coprecipitation of
autophosphorylated EGFR with tyrosine-phosphorylated PLC-
1 was
observed. The requirement for H2O2 of the
tyrosine phosphorylation of PLC-
1 was also demonstrated by
precipitating proteins with antibody to phosphotyrosine and immunoblot analysis with antibodies to PLC-
1 (Fig.
3D).
Because the exogenous bovine catalase and endogenous human catalase can
be separated on a SDS-polyacrylamide gel (Fig. 3E), the
amount of exogenous catalase incorporated to cells was estimated from
immunoblot analysis to be approximately five times the amount of
endogenous enzyme. Equal application of lysate protein among gel lanes
for all immunoblot experiments was confirmed by immunoblot analysis
with antibodies to -tubulin (Fig. 3F).
To test whether the number of EGFR on the cell surface is affected by electroporation or by the addition of catalase concurrent with electroporation, we measured the number of EGFR using 125I-EGF. The binding of 125I-EGF was saturable and inhibited in a concentration-dependent manner by the addition of unlabeled EGF (figure not shown). The calculated binding site densities per cell were 4 × 106, 6 × 106, and 5 × 106, respectively, for control cells, electroporated cells assessed after 18 h of recovery, and cells electroporated in the presence of catalase and assessed after 18 h of recovery, suggesting that receptor number is not affected by the treatments. We also found that the autophosphorylation activity of immunoprecipitated EGFR is not affected by the presence of catalase (8 mg/ml) or H2O2 (1 mM) (data not shown). These results suggest that the inhibition of EGF-induced tyrosine phosphorylation in cells electroporated in the presence of catalase is not attributable to either the reduction in EGFR number, the inhibition of EGFR kinase by catalase, or the requirement of H2O2 for the activation of EGFR kinase.
To assess the role of the intrinsic kinase activity and
autophosphorylation sites of the EGFR in the EGF-induced generation of
H2O2, we studied cell lines that express either
the wild-type EGFR, a tyrosine kinase-negative EGFR in which
Lys721 is replaced with Met, or a truncated EGFR (CD-126)
lacking the COOH-terminal 126 amino acids (and therefore the four
tyrosine phosphorylation sites at positions 1173, 1148, 1086, and
1068). These cell lines were generated previously by Margolis et
al. (28, 29) by expression of human EGFR cDNAs in NIH 3T3
cells that lack endogenous EGFR (2.2 cells). Receptor density was
3 × 105, 3 × 105, 1.5 × 105 receptors/cell, respectively, for cells expressing
wild-type, kinase-negative, and CD-126 mutant EGFR (30, 31). Treatment with EGF increased the concentration of H2O2 in
cells expressing the wild-type or CD-126 mutant EGFR but not in the
cells expressing the catalytically inactive mutant (Fig.
4). These results suggest that the intrinsic tyrosine
kinase activity, but not the four autophosphorylation sites, is
essential for the EGF-induced H2O2 generation.
Extracellular release of ROS is a well characterized response of
phagocytic cells to a variety of agonists. However, recent observations
indicate that nonphagocytic cells also generate ROS (4, 5, 6, 7, 8, 9, 10, 11, 12). To
identify the ROS responsible for the intracellular oxidation of DCFH,
Ohba et al. (4) added catalase to the culture medium of
osteoblastic cells. The addition of catalase completely inhibited the
transforming growth factor-1-induced increase in DCF fluorescence,
suggesting that H2O2 was important in DCFH
oxidation in response to transforming growth factor-
1. Because the
cells are not permeable to catalase, the researchers proposed that
H2O2 was released into the medium from the
plasma membrane, the site of its production, and then diffused into the
cells. Whether this diffusion was promoted by the trapped DCFH is not
clear. Catalase was also used to characterize the ROS generated in
PDGF-treated rat vascular smooth muscle cells (8). Unlike most other
cells, these muscle cells incorporated, by an unknown mechanism,
catalase that was added to the culture medium. Catalase incorporation
completely blocked the PDGF-stimulated increase in
H2O2 production, suggesting that
H2O2 is also the predominant ROS induced by
PDGF in these cells.
We have now shown that EGF induces ROS production in A431 cells and that the increase in DCF fluorescence was primarily attributable to H2O2 on the basis of its sensitivity to catalase introduced into the cells by electroporation. With incorporation of catalase by electroporation, no ambiguity arises as to whether H2O2 is released first into the medium and is then taken up by the cells. Unlike rat vascular smooth muscle cells, the simple addition of catalase to the culture medium did not result in uptake of the enzyme and a consequent decrease in DCF fluorescence (data not shown). The incorporation of catalase also reduced DCFH oxidation in A431 cells not exposed to EGF, indicating that substantial amounts of H2O2 are generated in the absence of EGF, probably as a result of respiratory activity and the presence of various growth factors in 1% FBS. Overnight incubation of electroporated cells in medium containing 1% FBS was necessary for cells to recover from the electroporation procedure and to exhibit a tyrosine phosphorylation response to EGF. A high background of DCF fluorescence was apparent even in cells containing exogenous catalase. Whether this background DCFH oxidation is caused by residual H2O2 not degraded by catalase, by cellular oxidants other than H2O2, or by oxidants introduced by the experimental procedure (for example, by photo-oxidation) is not clear. Therefore, we took great care to manipulate the cells under identical conditions, with the exception of the indicated additions.
The EGF-induced tyrosine phosphorylation of various cellular proteins
was completely blocked in A431 cells containing exogenous catalase.
Furthermore, detailed experiments with PLC-1 indicated that its
tyrosine phosphorylation in response to EGF requires the EGF-induced
increase in H2O2 concentration. Because
tyrosine phosphorylation is essential for the activation of PLC-
1 in
growth factor-treated cells (32), an increase in
H2O2 would appear to be required for growth
factor-induced hydrolysis of phosphatidylinositol 4,5-bisphosphate.
Treatment of smooth muscle cells with H2O2 was previously shown to release Ca2+ from intracellular stores
that are sensitive to inositol 1,4,5-trisphosphate (33). We also
observed a rapid increase in intracellular Ca2+ following
addition of 1 mM H2O2 to A431 cells
in a Ca2+-free medium (figure not shown). These
observations are consistent with the notion that
H2O2 inhibits PTPases and thereby causes activation (tyrosine phosphorylation) of PLC-
1 (see below).
Our study with EGF, together with previous studies with basic fibroblast growth factor (7, 9) and PDGF (8, 9), suggests that the generation of H2O2 is a common signaling event for peptide growth factors. However, the role of H2O2 in growth factor signaling is not clear. Exogenously added H2O2 was previously shown to elicit tyrosine phosphorylation in several cell types (34, 35), whereas inhibition of the PDGF-induced increase in H2O2 blocked various steps in signaling by this growth factor, including tyrosine phosphorylation of MAP kinase (8). Furthermore, H2O2 directly inhibits PTPase activity in vitro, and this inhibition is completely reversed by incubation with dithiothreitol (21). All PTPases contain one essential sulfhydryl group at their active site that is susceptible to oxidation because of its unusually low pKa (<5) (36). These observations suggest that PTPases may be targets of intracellularly generated H2O2. Inactivation of PTPases would result in increased tyrosine phosphorylation. Furthermore, the specific activities of PTPases in vitro are 10-1000 times those of protein tyrosine kinases (37). Therefore, in most cells, the activation of a receptor tyrosine kinase by the binding of a growth factor may not be sufficient to increase the steady-state level of protein tyrosine phosphorylation; concurrent inhibition of PTPases might be necessary, and this inhibition may be achieved through H2O2.
Binding of various peptide growth factors to their cognate receptors
activates multiple signaling pathways, including those mediated by
PLC-1, phosphatidylinositol 3-kinase, signal transducer and
activator of transcription protein (STAT), and MAP kinase (38, 39). The
ligand-bound receptors dimerize and transphosphorylate each other at
several tyrosine residues, thereby creating binding sites for cellular
proteins that contain Src homology 2 (SH2) domains, including
PLC-
1, GTPase-activating protein of RAS, the 85-kDa subunit of
phosphatidylinositol 3-kinase (p85), and SH2-containing collagen
protein. Studies with autophosphorylation site mutants of the receptors
for PDGF (40), colony-stimulating factor (41), fibroblast growth factor
(42), and nerve growth factor (43) have shown that elimination of
specific individual sites selectively abrogates the association of one
or two SH2-containing proteins with the receptors, suggesting that
individual autophosphorylation sites mediate the binding of specific
SH2-containing protein. However, the association of PLC-
1,
GTPase-activating protein, p85, or SH2-containing collagen protein with
the EGFR, which contains five autophosphorylation sites (residues 992, 1068, 1086, 1148, and 1173) in the COOH-terminal region, does not
appear to stringently require individual autophosphorylation sites, but
decreases gradually as the sites are removed one by one by
COOH-terminal truncation (44).
The intrinsic tyrosine kinase activity and autophosphorylation sites of the EGFR are not required for all signaling pathways activated by EGF. The activation of MAP kinase can occur independently of EGFR kinase activity (45, 46), and the activation of signal transducer and activator of transcription proteins requires none of the autophosphorylation sites (47). Our data with the kinase-inactive mutant indicate that EGFR-dependent H2O2 generation requires the intrinsic kinase activity of the receptor. The CD-126 mutant receptor, which lacks four of the five autophosphorylation sites, was as effective as the wild-type receptor in inducing H2O2 generation, suggesting that the tyrosine phosphorylation sites of the EGFR are not required for this effect. However, the possibility that the remaining autophosphorylation site (Tyr992) in CD-126 specifically serves as the binding site for a signaling molecule that mediates H2O2 generation cannot be excluded.
The mechanism by which H2O2 is generated has
been studied extensively in neutrophils and macrophages (48). In these
phagocytic cells, O2 is first produced from the reduction of
O2 by a multicomponent NADPH oxidase system, which consists
of a membrane-bound cytochrome b, a 22-kDa subunit
(p22-phox), and a 90-110-kDa glycoprotein subunit (gp91-phox). In
addition to these membrane-bound components, NADPH oxidase activity
requires various cytosolic factors including SH3-containing proteins
(p47-phox and p67-phox) and the small GTP-binding protein Rac.
The mechanism of O2 and H2O2
generation in nonphagocytic cells is unknown. Neither the xanthine
oxidase system nor the mitochondrial respiratory chain appears to
mediate receptor-triggered ROS generation (6). Evidence suggests
involvement of an NADPH oxidase-like system, which might be linked to
p22-phox, p47-phox, and p67-phox proteins, in fibroblast cells (6, 7,
9, 10, 49). However, the cytochrome b in fibroblasts appears
to be structurally and genetically distinct from that in phagocytic
cells (50). The mechanism by which this cytochrome b might
be modulated by the kinase activity of the EGFR requires further
investigation.
We thank Drs. J. Schlessinger and I. Lax for the NIH 3T3 2.2 cell lines expressing the wild-type, kinase-negative, and CD-126 EGFR proteins.